Fig 1.
Vpu-APEX2 fusion protein design and activity.
(A) A schematic representation of C-terminally tagged Vpu constructs; Vpu-FLAG and Vpu-FLAG-APEX2. A GGGS linker lies between the FLAG epitope and APEX2. (B) HeLa P4.R5 cells were transfected with Vpu constructs bearing C-terminal FLAG or FLAG and APEX2. Protein expression was analysed by western blot. (C) Cell-surface levels of BST-2 and CD4 were measured in the presence of Vpu-FLAG or Vpu-FLAG-APEX2 by flow cytometry. Representative two-color flow plots are shown, GFP was used as a marker for transfection. Surface levels of BST-2 and CD4 on cells expressing FLAG- or FLAG-APEX2-tagged Vpu was expressed as the % of control cells not expressing Vpu. Error bars represent standard deviation of n = 3 experiments.
Fig 2.
Vpu-APEX2 localizes to enlarged juxtanuclear endosomes and the limiting membranes of multi-vesicular bodies.
(A) Schematic depicting APEX2 staining protocol for visualization by electron microscopy. (B) HeLa P4.R5 cells were transfected with the codon-optimized Vpu constructs bearing a C-terminal APEX2 tag. 24 hr later the cells were fixed before APEX2-dependent polymerization of DAB in the presence of hydrogen peroxide. The cells were then stained with OsO4, processed, and imaged by transmission electron microscopy (TEM). Left panel: Cells expressing Vpu-APEX2 contain juxtanuclear accumulations of osmium-highlighted enlarged vesicles (EV) likely derived from the Golgi and endosomes. Right panel: a higher magnification image of the juxtanuclear region of the cell shown at left. The limiting membranes of vesicles resembling multivesicular bodies (MVB) are highlighted by osmium. (C) Cells expressing Vpu (without an APEX2 tag) also contain enlarged juxtanuclear vesicles. (D) A control image showing Golgi (G) stacks in non-transfected HeLa P4.R5 cells.
Fig 3.
Light and electron microscopic distributions of Vpu-APEX2 and mutants A18H, AAA/F, and S52,56N.
(A) HeLa P4.R5 cells were transfected to express Vpu-FLAG or Vpu-FLAG-APEX2, either wild type or encoding the mutations A18H, AAA/F, or S52,56N. The cells were fixed and stained for FLAG (shown in red) and TGN-46 (shown in green). (B) HeLa P4.R5 cells were transfected to express Vpu WT and the indicated mutants tagged with APEX2; 24 hr later the cells were reacted with DAB in the presence of hydrogen peroxide. The cells were stained and osmiophilic DAB polymer visualized in whole-cells by brightfield microscopy. (C) Thin-section electron microscopy of cells expressing Vpu WT-, A18H-, AAA/F-, or S52,56N-APEX2. Arrows indicate concentrations of osmiophilic polymer stain. N = nucleus, ER = endoplasmic reticulum, PM = plasma membrane.
Fig 4.
Biotinylation of Vpu-APEX2 proximal proteins.
(A) Schematic representation of APEX2-mediated biotinylation reaction; in the presence of hydrogen peroxide and biotin-phenol, APEX2 catalyzes biotinylation of nearby proteins, which can be captured by streptavidin. (B) Detection of biotinylated proteins by immunofluorescent-streptavidin. HeLa P4.R5 cells expressing Mito-V5-APEX2 or Vpu-FLAG-APEX2 were incubated with biotin phenol and hydrogen peroxide for 1 min before quenching, fixation, and staining with streptavidin-Alexa Fluor 594 (red) and anti-V5 or anti-FLAG (green). (C) Biotinylation pattern of protein species visualized by western blot; HeLa P4.R5 cells transfected to express Mito-V5-APEX or WT Vpu-, S52,56N-, A18H-, or AAA/F-FLAG-APEX2 were incubated with biotin-phenol, lysed and proteins separated by SDS-PAGE and western blot. Biotinylated proteins were detected using streptavidin-HRP. The control cells were transfected to express Vpu-FLAG only; streptavidin staining is absent in the absence of APEX2.
Fig 5.
Pair-wise comparisons of the proximity-omes of wild type Vpu relative to the ER-retained Vpu-A18H and the plasma-membrane-enriched Vpu-S52,56N mutants.
HeLa P4.R5 cells were transfected to express Vpu constructs bearing C-terminal APEX2 tags, in duplicate. Following proximity biotinylation reactions, the biotinylated proteins were isolated and subject to quantitative mass spectrometry. Volcano plots of protein enrichment in the presence of Vpu mutants A18H or S52,56N relative to the wild-type Vpu are shown, n = 2 experiments. Significantly enriched proteins highlighted in red and blue are derived from the Student’s t-test (p < 0.05). (A) The A18H mutant was enriched in proteins derived from the biosynthetic pathway, including the ER (labeled), while WT Vpu was enriched for proteins associated with plasma and endosomal membranes (labeled). (B) The S52,56N mutant was enriched in proteins derived from the plasma membrane (labeled), while WT Vpu was again enriched in endosomal sorting proteins (labeled). Proteins enriched by the S52,56N mutation included the known targets CD4, CD81, and HLA-C, and possible targets EGFR and CD55. For both (A) and (B), the x-axis shows log2 fold change of proteins enriched by mutant/WT Vpu and the y-axis -log10 of p-value derived from Student’s t-test. The 10 most highly enriched gene ontology (cellular component) terms are shown on the right of each volcano plot, corresponding to significantly enriched proteins proximal to the mutants; p-value derived from Bonferroni test.
Fig 6.
Heat map and k-means clustering of proteins (384) for which any Vpu-mutant was significantly different from wild type.
(A) Heatmap of the relative abundance of proteins measured in Vpu-APEX2 WT and mutant samples. The heatmap was sorted into 6 clusters by k-means clustering analysis; the cluster profile is shown on the left. Data are presented as the fold change in protein abundance relative to WT. (B) Cluster 5 contains known and potential targets of Vpu, including CD4, CD81, and HLA-C. k-means cluster 6 contains potential serine-dependent cofactors of Vpu, including endosomal sorting proteins. Data are derived from duplicate samples per condition, from two independent experiments.
Fig 7.
Proteins enriched in the proximity-ome of Vpu-S52,56N relative to wild type: Vpu-mediated and serine-dependent decreases in steady-state expression.
(A) HEK293T cells were co-transfected with target cDNAs and plasmids expressing WT Vpu or Vpu-S52,56N. Levels of proteins were measured 48 hours later by immunofluorescent staining and automated quantification of fluorescence. Candidate Vpu targets whose expression is significantly decreased by Vpu are colored. (B) Vpu-mediated degradation of putative target proteins was tested by western blotting of exogenously-expressed cDNA targets.
Fig 8.
Potential Vpu cofactors: STRING relationships and co-localization with Vpu.
(A) The interactions of proteins in k-means cluster 6 (Fig 6) were visualized using the STRINGdb (Cytoscape) network analysis tool. The clusters of interrelated proteins identified represent candidate Vpu cofactors. (B) Immunofluorescence microscopy of Vpu-FLAG and candidate cofactors. HeLa P4.R5 transfected to express Vpu-FLAG were fixed and stained for the endogenous proteins STAM, PTPN23, and Vps35. Images are z-stack projections of full cell volumes; insets show single z-sections, with arrows indicating co-localized foci. (C) Proximity ligation assay (PLA) demonstrates close proximity of Vpu-FLAG to the endogenous proteins. P4.R5 cells transfected to express either Vpu-FLAG or empty plasmid (control) were fixed and stained with primary antibodies targeted to FLAG and the indicated endogenous proteins, followed by PLA procedure (see Materials and Methods). Red signal (over background) indicates close proximity of two proteins of interest (< 40nm). Images are z-stack projections of full cell volumes. Upper panels show PLA signal alone, lower panels are overlaid with DAPI nuclear stain. Scale bars are 10 μm.
Fig 9.
Cofactors required for Vpu activities.
Candidate cofactor proteins were transiently depleted using siRNAs in HeLa P4.R5 cells. The cells were transfected with pNL4-3 (an HIV proviral plasmid expressing the complete viral genome including Vpu) or pNL4-3ΔVpu 48 hours later. The cells were lysed 24 hr post-transfection and BST-2 was probed by SDS-PAGE and western blotting. (B) BST-2 signals were measured relative to loading control (GAPDH) and presented as fold signal over NL4-3 (negative control siRNA; Vpu-expressed). Data are the mean +/- SD of three independent experiments, the p-value was determined by Student’s t-test. (C) Cofactors affecting virus release. Candidate cofactors affect virus release. Candidate cofactor proteins were transiently knocked-down using siRNAs in HeLa P4.R5 cells 48 hr prior to transduction with VSV-G-pseudotyped ΔEnv HIV (dHIV) or dHIVΔVpu at MOI 0.1. Cell-free and cell-associated virus (p24) was measured 24 hr later. Data are expressed as p24 in the supernate proportional to cell-associated p24, i.e. proportional p24 released from the cells. Data are mean +/- SD of two independent experiments, with duplicate wells measured in each experiment. Data points are colored according to experiment. The p-value was determined by Student’s t-test. (D) Cofactors affecting surface BST-2 and CD4. HeLa P4.R5 cells transiently depleted of the candidate cofactors by siRNA-knockdown were transfected to express Vpu-FLAG or empty plasmid control, and GFP. BST-2 and CD4 levels at the cell surface were measured by immunofluorescence staining and flow cytometry 24 hr later. Surface levels of BST-2 and CD4 on cells expressing Vpu were expressed as the % of control cells not expressing Vpu. Error bars represent standard deviation of n = 2 experiments, p-value determined by Student’s t-test.