mTOR inhibition causes a reduction in HIV-1 NL4–3 Gag, Vif, and Vpr

Our previous study showed that treatment with Rapamycin, an mTOR inhibitor commonly used to induce autophagy, caused a significant defect in HIV-1 NL4–3 replication and virion production in many cell types, including primary CD4⁺ T cells, but the virus uses its accessory protein Nef to counteract this restriction [44]. These findings indicate that autophagy acts as an important barrier for HIV-1 replication. In a follow-up study, we revealed that this defect in virion production is caused by the autophagy-dependent clearance of HIV-1 Gag [43]. Gag is the major structural protein of the virus, it drives virion assembly, and it is also the precursor of the virus capsid protein, so degradation of Gag causes major defects in HIV-1 replication [48–52]. These observations prompted us to analyze the effect of autophagy on other HIV-1 proteins. For this, we infected Jurkat CD4⁺ T cells with lab adapted HIV-1 NL4–3 or NL4–3 Δnef in the presence of increasing amounts of Rapamycin, using a range of concentrations that we have already validated triggers autophagy without impacting cell viability (1.3-13 μM) [43,44]. Eighteen hours later, the levels of HIV-1 proteins were measured by western blot and quantified by densitometry analyses. Their expression was normalized to ACTB/β-actin and their levels relative to the DMSO treatment (D on the x axis) were calculated. Whereas Rapamycin had no major impact on HIV-1 proteins for the nef-competent virus, Gag, Vpr and Vif were reduced in a dose-dependent manner in the absence of Nef (Fig 1A-1C, red arrows). Consistent with our previous study showing that Nef inhibits LC3-I-to-LC3-II conversion [43,44], autophagy progression was impaired in cells infected with the nef-competent virus. Rather than the expected decrease in LC3-I with its corresponding increase in LC3-II (hallmark for autophagy activation), an accumulation of LC3-I was detected, despite increasing amounts of Rapamycin, which reflects an impairment in the lipidation of LC3 (Fig 1C, left blots). However, in cells infected with the nef-defective virus, autophagy proceeded as expected: with a dose-dependent accumulation of LC3-II and an associated reduction in LC3-I (Fig 1C, right blots). Similar observations were obtained in primary CD4⁺ T cells. However, unlike in the Jurkat cells, Nef’s ability to inhibit autophagy was more limited. This is in line with previous reports showing that upon cellular activation, primary CD4⁺ T cells exhibit greater autophagy flux at steady state than T cell lines like Jurkat cells [43,44,53–56]. Thus, additional autophagy induction by Rapamycin overwhelms Nef’s ability to counteract autophagy (Fig 1C and 1F, compare LC3 blots). Also, because of their inherently high autophagy activation at steady state, Rapamycin treatment of primary cells often does not lead to a striking increase in LC3-I-to-LC3-II conversion, yet its effect on autophagy targets remains apparent [54]. In fact, Rapamycin caused a dose-dependent reduction in Gag, Vif and Vpr, which was more prominent in cells infected with NL4–3 Δnef (Fig 1D-1F). Overall, these results suggest that besides Gag, the HIV-1 accessory proteins Vif and Vpr are also downregulated when mTOR is inhibited by Rapamycin, but Nef rescues their levels.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Rapamycin causes a reduction in HIV-1 NL4-3 Gag, Vif and Vpr.

The effect of mTOR inhibition by Rapamycin on HIV-1 was examined in Jurkat CD4⁺ T cells (A-C) and primary CD4⁺ T cells (D-F) for NL4-3 and NL4-3 Δnef. The levels of HIV-1 proteins were measured by densitometry assays for NL4-3 (A and D) and NL4-3 Δnef (B and E) infected cells. Protein levels were normalized to ACTB/β-actin, and their expression relative to the DMSO treatment (D on the x axis) was calculated. Data correspond to the mean and the SEM of 3 independent experiments. Dotted lines represent 2-fold threshold. Representative western blots from infections in Jurkat CD4⁺ T (C) and primary CD4⁺ T cells (F). Membranes were probed for gp120, Gag/p55, Vif, Nef, Vpr, Vpu and LC3. ACTB was used as a loading control. Red arrows indicate defects in HIV-1 proteins.

https://doi.org/10.1371/journal.ppat.1014020.g001

HIV-1 NL4–3 Vif downregulation by mTOR-mediated inhibition is due to increased proteasomal function

To understand the interplay between Vif and autophagy, we first examined whether Vif expression impacts autophagy flux. For this, and to eliminate any potential confounding effects caused by interactions of other HIV-1 genes with autophagy (i.e., nef), we used a plasmid encoding for Vif (Flag-Vif) to transfect HEK293T cells, a cell line that we previously demonstrated recapitulates the antagonistic phenotype of autophagy in HIV-1 infection observed in primary CD4⁺ T cells [43,44]. HEK293T cells were transfected with either an empty vector or the Flag-Vif construct. Forty-four hours later, cells were treated with increasing concentrations of Torin2, another mTOR inhibitor [57], in the presence or absence of BafilomycinA1 (BafA1) for 4 hours. BafA1 is a lysosomal inhibitor commonly used to measure autophagy flux [8,58], which was quantified by the LC3-II/I and SQSTM1 levels relative to the vector-transfected and DMSO-treated cells. No differences in the LC3-II/I and SQSTM1 ratios were observed between Vif-expressing and vector-transfected cells, even under conditions of BafA1 treatment, confirming that Vif does not impact autophagy turnover (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The inhibition of mTOR causes the proteasomal degradation of Vif.

(A) HEK293T cells were transfected with 2,000 ng of an empty vector (V) or a NL4-3 Flag-Vif construct. 44 h later, cells were treated with Torin2 (5-10 nM) with or without BafA1 (500 nM) for 4 h. Next, cells were harvested, analyzed by western blot, and autophagy flux was measured by determining the LC3-II/I ratios and SQSTM1 levels relative to the vector-transfected and DMSO-treated cells. Data correspond to the mean and the SEM of 4 independent experiments. Blots are representative of 3 independent experiments. (B) Parental (wild type; WT) and ATG5KO HEK293T cells were transfected with 2,000 ng of a NL4-3 Flag-Vif construct. 44 h later, cells were treated with DMSO or increasing concentrations of Rapamycin (1.3-6 μM) or Torin2 (2.5-10 nM) at two-fold increments. 4 h later, cells were harvested for their analysis by western blotting. Blots are representative of 3 independent experiments confirming ATG5 depletion and its effects on SQSTM1 levels, the lipidation of LC3 as well as on Flag-Vif. ACTB was included as a loading control. (C) The LC3-II/I ratios, the SQSTM1 and Vif levels across the Rapamycin treatment range relative to DMSO were determined by densitometry analyses of western blots from 3 independent assays. SQSTM1 and Vif levels were normalized over ACTB prior to normalization to DMSO. Data represent the mean and the SEM of 3 independent experiments. Dotted lines represent 2-fold threshold. (D) The LC3-II/LC3-I ratios, the SQSTM1 and Vif levels across the Torin2 treatment range relative to DMSO were determined by densitometry analyses of western blots from 3 independent assays. Data represent the mean and the SEM of 3 independent experiments. Dotted lines represent 2-fold threshold. (E) ATG5KO HEK293T cells were transfected with NL4-3 Flag-Vif. 44 h later, cells were treated with DMSO, Torin2 (2.5 nM), Torin2 (2.5 nM) with Chloroquine (60 μM), or Torin2 (2.5 nM) with MG132 (20 μM) for 4 h. Cells were then harvested for their analysis by western blot for Flag-Vif, LC3, ATG5 and SQSTM1. ACTB was included as a loading control. Blots are representative of 3 independent experiments.

https://doi.org/10.1371/journal.ppat.1014020.g002

Next, to examine whether Vif is a genuine autophagy target, the implication of other pathways modulated by mTOR needed to be ruled out. This is relevant as mTOR inhibition not only induces autophagy but also proteasomal and lysosomal function [59–62]. Therefore, we generated ATG5 knockout cells. ATG5 was selected because it is a critical factor for autophagy progression, particularly important for the transition of LC3-I into LC3-II and is necessary for autophagosome biogenesis [63,64]. ATG5 knockout was verified by western blot from pooled cells, showing low or undetectable levels of the protein, a defect in the emergence of LC3-II, and, despite the presence of mTOR inhibitors, the accumulation of the autophagy cargo receptor SQSTM1 (Fig 2B-2D), which are all a consequence of deleting ATG5 and impairing autophagy [65,66]. Besides Rapamycin, we also included Torin2 to validate our findings. Next, parental (wild type; WT) and ATG5KO HEK293T cells were transfected with the Flag-Vif construct. Forty-four hours later, cells were treated with either DMSO, Rapamycin or Torin2 for 4 hours before harvesting them for analysis. As expected, both Rapamycin and Torin2 caused the transition from LC3-I to LC3-II in WT cells. However, at their highest concentrations we frequently observed a reduction in LC3-II, which was often accompanied by a decrease in SQSTM1, indicating that at these concentrations, autophagy culminated in maturation, which causes degradation of LC3-II and autophagy receptors (i.e., SQSTM1) (Fig 2B-2D). By contrast, these effects of Rapamycin and Torin2 were not observed in ATG5KO cells, as LC3-II/I ratios and SQSTM1 levels did not fluctuate across the assay (Fig 2B-2D), confirming that autophagy is impaired in ATG5KO cells. Remarkably, Rapamycin and Torin2 treatments caused a dose-dependent reduction of Vif regardless of ATG5 expression (Fig 2B-2D), indicating that Vif downregulation is autophagy-independent. To elucidate how mTOR inhibition causes a reduction in Vif levels, subsequent experiments were performed in Torin2-treated and autophagy-deficient (ATG5KO) HEK293T cells in the presence of lysosomal (Chloroquine) or proteasomal (MG132) inhibitors. This assay revealed that proteasomal inhibition rescues Vif levels when mTOR is repressed (Fig 2E). Hence, these results confirm that Vif is not an autophagy target and that its proteasomal turnover is likely enhanced under conditions of mTOR repression.

HIV-1 NL4–3 Vpr is a bona fide autophagy target

Next, we studied the interplay between Vpr and autophagy by assessing the impact of Vpr on autophagy flux. HEK293T cells were transfected with an empty vector or a NL4–3 Vpr-myc construct. Forty-four hours later, cells were treated with increasing concentrations of Torin2 in the presence or absence of BafA1 for 4 hours. No significant differences were observed between vector and Vpr-expressing cells, particularly under BafA1 treatment, reflecting no major impact on autophagy flux. However, in cells expressing Vpr and treated only with Torin2 we noticed a trend (yet insignificant) towards increased flux, which is reflected in the turnover of LC3-II and SQSTM1 (Fig 3A), suggesting that Vpr might induce autophagy. Importantly, Vpr levels were restored in cells treated with BafA1, indicating that either autophagy or lysosomal activity drive the degradation of Vpr (Fig 3A, blots). To assess whether autophagy mediates its degradation, we transfected WT and ATG5KO HEK293T cells with the Vpr-myc construct and treated them with increasing concentrations of Rapamycin or Torin2. The effects of these mTOR inhibitors on LC3-II/I and SQSTM1 were quantified as in the Vif experiments for both WT and ATG5KO cells (Fig 3B-3D). Both Rapamycin and Torin2 treatments caused a dose-dependent downregulation of Vpr in WT cells. In contrast, Vpr levels were restored in the ATG5KO cells (Fig 3B-3D), supporting the notion that Vpr is an autophagy target.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. NL4-3 Vpr levels are rescued in ATG5KO cells despite mTOR inhibition.

(A) HEK293T cells were transfected with 2,000 ng of an empty vector (V) or a NL4-3 Vpr-myc construct. 44 h later, cells were treated with Torin2 (5-10 nM) with or without BafA1 (500 nM) for 4 h. Next, cells were harvested, analyzed by western blot, and autophagy flux was measured by determining the LC3-II/I ratios and SQSTM1 levels relative to the vector-transfected and DMSO-treated cells. Data represent the mean and the SEM of 3 independent experiments. Blots are representative of 3 independent experiments. (B) Parental (WT) and ATG5KO HEK293T cells were transfected with 2,000 ng of a NL4-3 Vpr-myc construct. 44 h later, cells were treated with DMSO or increasing concentrations of Rapamycin (1.3-6 μM) or Torin2 (2.5-10 nM) at two-fold increments. 4 h later, cells were harvested for their analysis by western blotting. Blots are representative of 3 independent experiments from parental and ATG5KO cells confirming ATG5 depletion and its effects on SQSTM1 levels and the lipidation of LC3 as well as its impact on Vpr-myc. ACTB was included as a loading control. (C, D) Vpr and SQSTM1 levels were normalized to ACTB and their expression relative to DMSO (D on the x axis) was calculated. LC3-II and LC3-I levels were determined and provided as the LC3-II/I ratio relative to DMSO. Data represent the mean and the SEM of 3 independent experiments. Dotted lines represent 2-fold threshold.

https://doi.org/10.1371/journal.ppat.1014020.g003

The role of autophagy in the degradation of Vpr was corroborated in primary CD4⁺ T cells infected with NL4–3 Δnef and treated with increasing amounts of Torin2 in the presence of either DMSO or VPS34IN, an autophagy inhibitor that targets the VPS34 subunit in both class III PtdIns3K complex I (autophagy initiation) and class III PtdIns3K complex II (autophagy maturation) [67]. The effect of VPS34IN on autophagy flux was confirmed by western blot, where a delay in LC3-I-to-LC3-II transition was observed as well as an accumulation of LC3-II and SQSTM1 (Fig 4A and 4B). In line with the effect of Rapamycin on Vpr in primary cells (Fig 1E), Torin2 also caused the downregulation of Vpr in DMSO-treated primary CD4⁺ T cells. However, Vpr levels were rescued in the presence of VPS34IN (Fig 4A and 4B). To confirm that Vpr is targeted to autophagosomes, its subcellular distribution was examined by fluorescence microscopy in HEK293T cells constitutively expressing EGFP-LC3B and treated with Torin2 to induce autophagy. As endogenous LC3-I, EGFP-LC3-I also displays a dispersed cytosolic distribution. However, when converted into EGFP-LC3-II, its association with autophagosomes forms a characteristic puncta distribution in the cytosol [8,43,44,68] (Fig 5A). Consistent with a role for autophagy in targeting Vpr for elimination, extensive overlap between Vpr and LC3 puncta was observed (Fig 5A), indicating that Vpr is present in autophagosomes. Since we and others have previously reported that Nef blocks autophagy [35,38,40,44], we assessed whether providing Nef in trans could rescue Vpr levels. For this, we co-transfected HEK293T cells with the NL4–3 Vpr-myc construct and increasing concentrations of a plasmid encoding for NL4–3 Nef-HA in the presence of incremental doses of Torin2. In line with its role as an autophagy antagonist, Nef rescued Vpr expression in a dose-dependent manner (Fig 5B and 5C). Accordingly, Nef displaced Vpr from LC3 puncta, causing a significant reduction in the Pearson’s correlation coefficient (R) for the co-localization of Vpr with LC3 (Fig 5D). Overall, these findings confirm that Vpr is an autophagy target.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Vpr levels are rescued by autophagy inhibitors in primary cells.

(A) Primary CD4⁺ T cells were infected with NL4-3 Δnef in the presence of DMSO or 10 μM VPS34IN. 18 h later, cells were treated with increasing amounts of Torin2 (2.5-20 nM) for 4 h. Cells were then harvested and analyzed by western blot for LC3, SQSTM1 and Vpr. ACTB was added as a loading control. Blots are representative of 3 independent experiments. (B) Vpr and SQSTM1 levels were normalized to ACTB and their expression relative to DMSO (D on the x axis) was calculated. LC3-II and LC3-I levels were determined and provided as the LC3-II/I ratio relative to DMSO. Data represent the mean and the SEM of 3 independent experiments. Dotted lines represent 2-fold threshold.

https://doi.org/10.1371/journal.ppat.1014020.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Vpr localizes to autophagosomes, and its autophagy-mediated elimination is counteracted by Nef.

(A) The subcellular distribution of Vpr relative to autophagosomes (LC3) was examined by fluorescence microscopy in HEK293T cells constitutively expressing EGFP-LC3B and treated with Torin2 (10 nM) 4 h before staining. Images are representative of 3 independent experiments. White scale bar: 10 μm. (B) HEK293T cells were co-transfected with 2,000 ng NL4-3 Vpr-myc and increasing amounts of a plasmid encoding NL4-3 Nef-HA (0-2,000 ng). 44 h later, DMSO or increasing concentrations of Torin2 (5-10 nM) were added. 4 h later, cells were harvested and analyzed by western blot. Membranes were probed for Vpr, Nef, LC3, and SQSTM1. ACTB was included as a loading control. Blots are representative of 3 independent experiments. (C) Vpr protein levels were measured by densitometry, normalized to ACTB and their expression relative to the DMSO and no Nef treatment was calculated. Data correspond to the mean and the SEM of 3 independent experiments. Dotted lines represent 2-fold threshold. (D) The degree of co-localization between Vpr and LC3 puncta was examined by fluorescence microscopy in HEK293T cells constitutively expressing EGFP-LC3B in the presence of Nef and treated with Torin2 (10 nM) 4 h before staining. Images are representative of 3 independent experiments. White scale bar: 10 μm. Graph: The Pearson’s correlation coefficient (R) for the co-localization of Vpr and EGFP-LC3B in the presence and absence of Nef was calculated from 17 randomly selected fields. Data correspond to the raw values, the mean and SEM. *: p < 0.05; **: p < 0.01; ****: p < 0.0001. ns: not significant.

https://doi.org/10.1371/journal.ppat.1014020.g005

Vpr proteins from ten HIV-1 group M transmitted/founder viruses are resistant to autophagy

Despite the high degree of variability within the pool of HIV-1 variants in an individual, only specific clones can establish an infection in a new person. These clones are referred to as transmitted/founder viruses (TFVs) [69]. Unlike other virus variants, TFVs can circumvent the transmission bottleneck due to their high resistance to innate immune barriers [69]. Since autophagy is considered an innate defense response, we investigated the susceptibility of Vpr proteins from different TFVs to autophagy. The following 10 HIV-1 group M TFV clones were tested: CH077, Z3567, CH106, Z1123, 3137M, N31F, CH042, CH198, A190184, and CH470 (Table 1). These clones were selected because their vpr sequences are fully annotated, allowing us to individually assess their sensitivity to autophagy. Their vpr genes were cloned into pcDNA5 harboring either a Myc, Flag or a 6xHis tag in their C-terminus. As for Fig 3, these constructs were transfected into HEK293T cells in the presence of increasing concentrations of Torin2, and the effect of autophagy on their expression levels was examined by western blot. Unlike NL4–3 Vpr, the Vpr proteins from all 10 HIV-1 TFVs were not impacted by autophagy induction, even at the highest concentration of Torin2 (Fig 6A and 6B). These observations were confirmed by fluorescence microscopy, showing poor co-localization between one of the TFV Vpr proteins (CH077) and LC3 puncta (Fig 6C). Collectively, these findings indicate that the Vpr proteins of these ten group M HIV-1 TFVs are resistant to autophagy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Group M HIV-1 transmitted/founder viruses used in this study.

https://doi.org/10.1371/journal.ppat.1014020.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Vpr proteins from group M HIV-1 TFVs are resistant to autophagy.

(A) HEK293T cells were transfected with 2,000 ng of constructs coding for the indicated Vpr proteins. 44 h later, either DMSO or Torin2 (at 5 and 10 nM) were added. 4 h later, cells were harvested and analyzed by western blotting. Membranes were probed for Vpr, LC3 and ACTB (as a loading control). Blots are representative of 3 independent experiments. (B) The levels of Vpr relative to the DMSO treatment (D on the x axis) were calculated by densitometry analyses after normalizing with ACTB. Data correspond to the mean and the SEM of 3 independent experiments. Dotted line represents 2-fold threshold. (C) The degree of co-localization between one of the TFV Vpr proteins (CH077) and LC3 puncta was examined by fluorescence microscopy in HEK293T cells constitutively expressing EGFP-LC3B and treated with Torin2 (10 nM) 4 h before staining. Images are representative of 3 independent experiments. White scale bar: 10 μm. Graph: The Pearson’s correlation coefficient (R) for the co-localization of CH077 Vpr and EGFP-LC3B was calculated from 25 randomly selected fields. Data correspond to the raw values, the mean and SEM.

https://doi.org/10.1371/journal.ppat.1014020.g006

Residues responsible for Vpr’s sensitivity to autophagy map to the second alpha-helix and the unstructured C-terminal tail

Vpr is conserved across HIV-1, HIV-2 and SIV (Simian Immunodeficiency Virus) [70–72]. However, the fact that the TFV Vpr proteins we tested here were resistant to autophagy suggests that any amino acid differences between TFV Vprs and NL4–3 Vpr account for their distinct susceptibility to autophagy. To identify the residue(s) in NL4–3 Vpr that are responsible for its autophagy sensitivity, we generated chimeras between NL4–3 Vpr and one of the TFV Vpr: CH077, as they both belong to subtype B (Table 1). By aligning both Vpr proteins, we noticed that most of their differences clustered either in the first half of the protein or in the last twenty amino acids (Fig 7A). Therefore, the two chimeras we generated consisted of NL4–3 Vpr with either the N-terminal half of CH077 (residues 1–45) or the last twenty amino acids of CH077 Vpr (residues 77–96). Because there is still one amino acid difference between NL4–3 and CH077 Vpr at position 61, which is not found at these N-terminal or C-terminal regions, an additional NL4–3 Vpr mutant was engineered in which the residue present in CH077 was introduced into NL4–3 Vpr (I₆₁T) (Fig 7B, left panel, red dots). These chimeras, the point mutant and wild type NL4–3 Vpr were engineered to harbor a Myc tag in their C-terminus. These constructs were transfected into HEK293T cells in the presence of increasing concentrations of Torin2. While the substitution introduced at position 61 did not render NL4–3 Vpr resistant to autophagy, Vpr proteins harboring either the N-terminus or C-terminus of CH077 Vpr became resistant to autophagy (Fig 7B, blots), indicating that determinants for autophagy susceptibility reside within the N-terminus and C-terminus of the protein, and that changes at either of these domains are sufficient to render the protein immune to autophagy. To narrow down the specific amino acids, we constructed additional mutants. Here, differences between NL4–3 and CH077 Vpr were clustered in smaller groups, and these amino acid group differences were individually introduced into NL4–3 Vpr (Fig 7C, left panel, red dots). Despite some variability in their susceptibility to Torin2, none of these mutants became fully resistant to autophagy (Figs 7C and S1 Fig), suggesting that resistance to autophagy is conferred by several clustered mutations. In view of these results, we generated new NL4–3 Vpr mutants in which these groups of amino acids were introduced in combination. We found that substitution of residues I₃₇ and H₄₅ by the corresponding amino acids in CH077 Vpr (P₃₇ and Y₄₅) rendered NL4–3 Vpr fully resistant to autophagy (Fig 7D), revealing these two residues in CH077 Vpr are sufficient to bypass autophagy restriction. A similar strategy was carried out to identify the amino acids within positions 77–96 in the C-terminal region of CH077 Vpr that confer resistance to autophagy. After generating multiple Vpr mutants with different combinations of amino acids within that region (S1 Fig), we found that all the residues that differ between NL4–3 and CH077 Vpr in their C-terminus are required to confer resistance to autophagy. Therefore, our results indicate that Q₇₇I₈₃I₈₄Q₈₅T₈₆S₉₃R₉₄ in CH077 Vpr are responsible for the resistance to autophagy of chimera II (Fig 7D). Next, we located positions 37, 45, 77, 83–86, 93 and 94 onto the NL4–3 Vpr NMR structure and found that residues 37 and 45 locate within the second alpha-helix while the rest cluster at the unstructured C-terminal tail (Fig 7E), potentially creating a pocket that might facilitate recognition by autophagy factors. Finally, we examined the subcellular distribution of a NL4–3 Vpr protein harboring mutations that render it resistant to autophagy (P₃₇Y₄₅) and found that the co-localization between this Vpr protein and LC3 puncta was drastically reduced (Fig 7F, white arrowhead shows area of co-localization), confirming that mutations P₃₇Y₄₅ are sufficient to render NL4–3 Vpr immune to autophagy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Residues responsible for Vpr’s susceptibility to autophagy cluster in the second alpha helix and C-terminus of the protein.

(A) Protein alignment of the Vpr proteins of the lab-adapted clone NL4-3 and the TFV clone CH077. Consensus sequence is shown underneath. Red squares indicate the residues that account for their different susceptibility to autophagy. (B) Left panel. Design of chimeras I and II and the I₆₁T point mutant. Red dots are representative of amino acids that differ between NL4-3 (green background) and CH077 (orange background). Right panel: HEK293T cells were transfected with the chimeras and point mutant. As a control, cells were transfected with NL4-3 Vpr-myc. The susceptibility of these Vpr proteins to increasing amounts of Torin2 (5-10 nM) was examined by western blot. DMSO was added as a control. Membranes were probed for Myc, LC3 and ACTB. Data are representative of 3 independent experiments. The Vpr levels were normalized to ACTB and their expression relative to DMSO was calculated. The mean values from the 3 replicates are provided underneath the blots. (C-D) Design of the subsequent Vpr mutants is shown in the left panels. The right panels show the susceptibility of these Vpr mutants to autophagy. The levels of Vpr-myc over ACTB relative to the DMSO treatment is shown underneath the blots. Data are representative of 3 independent experiments. (E) Different views of the NL4-3 Vpr NMR structure where positions 37, 45, 77, 83-86 and 93-94 are highlighted in red. NTD: N-terminal domain. CTD: C-terminal domain. (F) The degree of co-localization between the autophagy-resistant Vpr_P37Y45 mutant and LC3 puncta was examined by fluorescence microscopy in HEK293T cells constitutively expressing EGFP-LC3B and treated with Torin2 (10 nM) 4 h before staining. Images are representative of 3 independent experiments. White scale bar: 10 μm. White arrowhead shows area of co-localization. Graph: The Pearson’s correlation coefficient (R) for the co-localization of Vpr and EGFP-LC3B was calculated from 25 randomly selected fields. Data correspond to the raw values, the mean and SEM.

https://doi.org/10.1371/journal.ppat.1014020.g007

As mentioned earlier, Vpr is present in all primate lentiviruses with a high degree of conservation [70–72]. Hence, we were curious to know whether the substitutions at positions 37, 45, 77, 83–86, 93 and 94 in NL4–3 Vpr emerged as a result of lab adaptation. To assess this, we compared the Vpr sequences of other commonly used lab adapted HIV-1 molecular clones, namely HBX2, BaL and JR-FL, with that of NL4–3 Vpr. Although Vpr in BaL is shorter, for the most part, the sequences of these Vpr proteins preserved the residues present in NL4–3 Vpr at positions 37, 45, 77, 83–86, 93–94 (S2A Fig). There was only one exception, though, in JR-FL. Here, JR-FL Vpr harbored Y₄₅, Q₇₇ and I₈₃ (S2A Fig, red boxes), which are three of the substitutions present in CH077 Vpr that contribute to its resistance to autophagy – although on their own they are not sufficient to render NL4–3 Vpr insensitive to autophagy. In view of these data, we decided to examine the sequence variation at positions 37, 45, 77, 83–86, 93–94 in Vpr of primary HIV-1 isolates. For this, we analyzed 212 group M HIV-1 sequences of the HIV Sequence Database from the Los Alamos National Laboratory (LANL) [73] and found that the majority of these isolates share the CH077 residues at positions 37, 45, 77 and 83 in Vpr. However, most of these primary isolates share the NL4–3 residues at positions 93 and 94. Interestingly, a high percentage of these strains harbor amino acids that differ from CH077 and NL4–3 at positions 85 and 86 (S2B Fig). Because our mapping data indicates that P₃₇Y₄₅ are sufficient to render NL4–3 Vpr resistant to autophagy, it is tempting to speculate that, similar to most TFV Vprs tested here (S2C Fig), the majority of primary group M HIV-1 isolates harbor Vpr proteins with amino acid changes that would confer resistance to autophagy (i.e., P₃₇Y₄₅), while most lab adapted Vprs harbor residues at those sites that would render the protein susceptible to autophagy.

Amino acid changes between autophagy-sensitive and -resistant Vpr account for their differential binding to autophagy receptors

As indicated earlier, residues 37, 45, 77, 83–86, and 93–94 map to the second alpha-helix and the unstructured C-terminal tail of the NL4–3 Vpr NMR structure (Fig 7E). We hypothesize that changes at these locations might affect the overall protein conformation and its interactions with autophagy receptors. This idea was tested by investigating NL4–3 Vpr (hereafter Lab-Vpr) and CH077 Vpr (hereafter TFV-Vpr) for their ability to interact with different autophagy receptors, such as NBR1, NDP52, OPNT, SQSTM1, and TAX1BP1, [74–77]. BNIP3L was used as a negative control, since this is a mitophagy receptor that targets malfunctioning mitochondria for degradation and as such, it is not expected to interact with Vpr [78]. For this, HEK293T cells were transfected with constructs coding for either Lab-Vpr or TFV-Vpr. Forty-eight hours later, cells were harvested. A small aliquot of the lysates was set aside to assess input levels of these proteins. Of note, the migration pattern of Lab-Vpr and TFV-Vpr is slightly different likely due to harboring different tags in their C-terminus. The rest of the lysates were used to immunoprecipitate the above-mentioned autophagy receptors using specific antibodies (Table 2) and their pulldown fraction was analyzed by western blot using antibodies for each of these receptors plus Vpr. As controls, lysates for each of these transfections were incubated with Protein G magnetic beads (but no antibody) to rule out any unspecific binding of Vpr and/or autophagy factors with the beads. Additionally, the IgG antibodies used in the immunoprecipitations were incubated with Protein G magnetic beads but no cell extract to account for any bands corresponding to the heavy or light chains of the antibodies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Antibody sources and conditions.

https://doi.org/10.1371/journal.ppat.1014020.t002

As expected, neither Lab-Vpr nor TFV-Vpr were found in the pulldown fraction of BNIP3L (Fig 8A). Additionally, no binding between these Vpr proteins and the receptors OPNT and NBR1 was detected either (Fig 8B and 8C). By contrast, Lab-Vpr but not TFV-Vpr was found in association with SQSTM1 and TAX1BP1 (Fig 8D and 8E). Mixed results were obtained, however, with the NDP52 receptor, where Lab-Vpr was detected in its pulldown fraction ~50% of the time (Fig 8F, representative blots for positive and negative interaction). To confirm the relevance of these autophagy receptors in recruiting Lab-Vpr to autophagosomes, SQSTM1 + TAX1BP1 double knockout as well as SQSTM1 + TAX1BP1+NDP52 triple knockout HEK293T cells were engineered by CRISPR/Cas9. While Lab-Vpr levels were noticeably rescued in the SQSTM1^KO+TAX1BP1^KO cells, its expression levels were fully restored in the SQSTM1^KO+TAX1BP1^KO+NDP52^KO cells (Fig 8G), supporting that these three autophagy receptors participate in the degradation of Vpr. These observations were corroborated by immunofluorescence microscopy, with noticeably more overlap between Lab-Vpr and these autophagy receptors than TFV-Vpr, which was confirmed by determining the Pearson’s correlation coefficient (R) for the co-localization between each Vpr protein and these autophagy factors (Fig 9, white arrowheads indicate areas of co-localization). These findings suggest that Vpr recognition by NDP52, SQSTM1 and TAX1BP1 contributes to its susceptibility to autophagy elimination, and that amino acid differences between Lab-Vpr and TFV-Vpr account for their differential interactions with these autophagy receptors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Lab-Vpr but not TFV-Vpr associates with NDP52, SQSTM1 and TAX1BP1 autophagy-cargo receptors.

HEK293T cells were transfected with 2,000 ng of either Lab-Vpr or TFV-Vpr. 48 h later, cells were harvested, and lysates were immunoprecipitated for BNIP3L (A), OPNT (B), NBR1 (C), SQSTM1 (D), TAX1BP1 (E), or NDP52 (F). The pulldown fraction was analyzed for each of these autophagy receptors and Vpr. The whole cell lysates were also analyzed for these autophagy factors, Vpr and ACTB. Blots are representative of 3 independent experiments. (G) Lab-Vpr was transfected in parental, SQSTM1^KO/TAX1BP1^KO and SQSTM1^KO/TAX1BP1^KO/NDP52^KO knockout cells. 44 h later, cells were treated with increasing amounts of Torin2 (5-10 nM) for 4 h. Cells were then harvested and analyzed by western blot for Vpr, SQSTM1, TAX1BP1, NDP52 and ACTB. Blots are representative of 3 independent experiments. Graph: Vpr levels were quantified by normalizing to ACTB, and their relative expression compared to WT cells treated with DMSO (D on the x axis) were calculated from 3 independent experiments. Data correspond to the mean and the SEM of 3 independent experiments. *: p < 0.05.

https://doi.org/10.1371/journal.ppat.1014020.g008

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Lab-Vpr but not TFV-Vpr co-localizes with autophagy receptors SQSTM1, TAX1BP1 and NDP52.

HEK293T cells were transfected with 2,000 ng of either Lab-Vpr (A) or TFV-Vpr (B). 44 h later, cells were treated with Torin2 (10 nM) for 4 h. Next, cells were analyzed by immunofluorescence for the subcellular distribution of Vpr (green) relative to SQSTM1, TAX1BP1 and NDP52 (red). Hoechst was used to stain the nuclei (blue). White scale bar: 10 μm. White arrowheads indicate co-localization between Vpr and the selected autophagy markers. Images are representative of 3 independent experiments. Graphs: The Pearson’s correlation coefficient (R) for the co-localization of Vpr and these receptors was calculated from 8 randomly selected fields. Data correspond to the raw values, their mean and SEM.

https://doi.org/10.1371/journal.ppat.1014020.g009

Role of ubiquitination in targeting Vpr to autophagosomes

In mammals, the most prevalent autophagy-targeting signal for cargo is the presence of ubiquitin chains [79]. Many autophagy receptors bind to ubiquitin on their cargo by means of a ubiquitin-binding domain, so their dual interaction with cargo and LC3 allows to recruit molecules to autophagosomes. Classic examples of this are NDP52, SQSTM1 and TAX1BP1 [74,75,77]. Because these three receptors bind and co-localize with Lab-Vpr, but not TFV-Vpr, we reasoned that ubiquitination of Lab-Vpr might be an important signal for its recruitment to autophagosomes. To test this, we used TAK-243, a small molecule inhibitor of ubiquitination [80]. Specifically, TAK-243 acts by blocking the E1 ubiquitin-activating enzymes by binding to free ubiquitin to form a TAK-243-ubiquitin adduct, thereby, restricting the amount of ubiquitin available for ubiquitination [80]. HEK293T cells were transfected with Lab-Vpr. Twenty-four hours later, cells were treated with either DMSO, as a negative control, or TAK-243. One day later, cells were incubated with increasing amounts of Torin2 for 4 hours before cells were harvested. Compared to the DMSO-treated cells, TAK-243 restored Lab-Vpr levels, even at the highest concentration of Torin2 (Fig 10A), which suggests that Vpr ubiquitination is necessary to target it for autophagy degradation. However, when we analyzed the lysates for autophagy status – by looking at the relative abundance of LC3-II over LC3-I – we realized that TAK-243 affected LC3 lipidation, since there was a strong accumulation of LC3-I, which in turn affected the LC3-II/I ratios (Fig 10A). Therefore, to better assess if Vpr ubiquitination causes its autophagy-mediated degradation, we introduced an amino acid substitution at the single Lysine residue in Vpr – which has remained conserved across primate lentiviruses [81] – as Lysine amino acids are common targets for ubiquitination [82]. Introduction of Alanine or Arginine substitutions at this location were not tolerated, as Vpr levels were undetectable for these mutants. Hence, we replaced this Lysine by a Methionine, which does not affect Vpr expression, as reported by others [81]. However, this mutant was still susceptible to autophagy elimination upon Torin2 treatment (Fig 10B). We then performed a ubiquitination assay where we compared the ubiquitin levels between Lab-Vpr and TFV-Vpr immunoprecipitated from HEK293T cells, as we have done in the past for other ubiquitinated substrates [43,83]. As expected for ubiquitinated proteins, a smear spanning from the top of the membrane was detected for Lab-Vpr. However, a similarly intense smear was also observed for TFV-Vpr (Fig 10C), indicating that the ubiquitination patterns of these two proteins are similar. Importantly, the ubiquitin smear for both Lab-Vpr and TFV-Vpr was only minimally above the background smear detected in the IgG control, suggesting that Vpr may not be ubiquitinated or, even if it is, its ubiquitination levels are very low.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 10. Role of ubiquitination in targeting Vpr for autophagy-mediated degradation.

(A) HEK293T cells were transfected with 2,000 ng of Lab-Vpr. 24 h later, DMSO or TAK-243 (200 nM) was added. 44 h later, cells were treated with increasing amounts of Torin2 (5-10 nM). 4 h later, cells were harvested and analyzed by western blot. Membranes were probed for Myc, LC3 and ACTB. The LC3-II/I ratio as well as the Vpr levels relative to DMSO (D on the x axis) are provided over the Torin2 treatment range for both DMSO- and TAK-243-treated cells. Dotted line represents 2-fold threshold. Data correspond to the mean and the SEM of 3 independent experiments. (B) HEK293T cells were transfected with Lab-Vpr or the Lab-Vpr K₂₇M mutant. 44 h later, cells were treated with DMSO or Torin2 (10 nM). 4 h later, cells were processed for western blot analyses as in panel A. Blots are representative of 3 independent experiments. (C) HEK293T cells were transfected with either Lab-Vpr or TFV-Vpr. 48 h later, cells were harvested, and lysates were immunoprecipitated (IP) for Vpr. The IP samples were analyzed by western blot by probing membranes for ubiquitin and Vpr. The whole cell lysates were analyzed for Vpr and ACTB. Blots are representative of 3 independent experiments. (D) Diagram of SQSTM1 and truncation mutants. (E) HEK293T cells were co-transfected with Lab-Vpr and the mCherry-SQSTM1 constructs. 48 h later, cells were harvested, and lysates were immunoprecipitated (IP) for Vpr. The IP samples were analyzed by western blot by probing membranes for mCherry and Myc. The whole cell lysates were analyzed for mCherry, Vpr-myc and ACTB. Blots are representative of 3 independent experiments. (F) HEK293T cells were co-transfected with Lab-Vpr and the SQSTM1 constructs shown in panel D. 44 h later, cells were treated with Torin2 (10 nM) for 4 h and were subsequently stained for the SQSTM1 constructs (mCherry, red), Vpr (green), and the nuclei (blue). White scale bar: 10 μm. Images are representative of 3 independent experiments.

https://doi.org/10.1371/journal.ppat.1014020.g010

We next used an indirect approach to examine whether ubiquitination of Lab-Vpr is necessary to route the protein to autophagosomes. For this, we performed co-immunoprecipitations with SQSTM1 mutants where motifs important for ubiquitin binding were truncated. Specifically, we co-transfected HEK293T cells with Lab-Vpr and either wild type (WT) SQSTM1, a SQSTM1 construct lacking the N-terminal PB1 domain (ΔPB1) that allows SQSTM1 oligomerization – which enhances its association with ubiquitinated substrates [84] – and a SQSTM1 mutant lacking the C-terminal ubiquitin-binding domain (ΔUBA) necessary to interact with ubiquitinated cargo [85,86] (Fig 10D). Remarkably, all SQSTM1 constructs were still detected in the Lab-Vpr pulldown fraction (Fig 10E), and Vpr was found co-localizing with all these SQSTM1 mutants (Fig 10F). Collectively, our assays indicate that the Vpr-SQSTM1 binding is ubiquitin-independent, but the role of ubiquitination in targeting Lab-Vpr to autophagosomes still remains unclear.

Autophagy-resistant Vpr proteins confer an advantage in virus spread in an in vitro setting

Because the Vpr proteins of the ten different group M TFVs were resistant to autophagy, we theorize that resistance to autophagy by Vpr might provide an advantage in HIV-1 infection and/or spread. To investigate this, we assessed cell-free as well as total (cell-free plus cell-to-cell) infection using three HIV-1 NL4–3-based clones: the wild type, encoding for NL4–3 Vpr (autophagy-susceptible, or Vpr_Lab), a NL4–3 clone harboring the Vpr mutations P₃₇Y₄₅, which render Lab-Vpr resistant to autophagy (Vpr_mut), and another NL4–3 clone encoding for the TFV CH077 Vpr (autophagy-resistant, or Vpr_TFV). We also generated variants of these three viruses carrying mutations in Nef at residues 48–49, which render the protein unable to counteract autophagy (Nef_mut) [43]. All viral clones were engineered to harbor mCherry-furin-P2A-T2A in the nef reading frame, so we could track infection, while allowing the expression of all the NL4–3 genes at comparable levels (S3 Fig). This is important, as Nef in the original CH077 molecular clone is barely detected (S3 Fig, white arrowhead). This low Nef signal in the CH077 virus is likely due to using an antibody raised against NL4–3 Nef, as CH077 and NL4–3 Nef only share 20% similarity. Additionally, CH077 harbors substitutions in other genes, which could also impact viral infectivity and spread. Therefore, comparing the impact of Vpr susceptibility to autophagy on infection by using the original molecular clones NL4–3 and CH077 would have complicated data interpretation due to all these differences.

For these in vitro infection assays, Jurkat LTR-GFP CCR5⁺ cells were used as recipient cells. These cells encode GFP under the HIV-1 LTR promoter, so in the presence of HIV-1 infection, or HIV-1 Tat, GFP is produced, which serves as an indicator of a successful infection event. Importantly, besides being GFP⁺, these cells are also expected to be mCherry⁺ upon infection with our NL4–3 mCherry reporter viruses, so any background level of GFP from a leaky LTR or due to non-productive infection of the mCherry clones can be compensated. As donor cells, we used MT4 cells because, unlike other CD4⁺ T cell lines, MT4 are more permissive and reach higher levels of infection [87]. MT4 cells were infected with 100 ng of p24 equivalents of each recombinant virus by spinoculation for 2 hours at 37 ºC. Cells were washed, resuspended with fresh media and kept for 2 days at 37 ºC. After this 2-day period, an aliquot of MT4 cells was analyzed by flow cytometry to measure the percentage of infected cells. Infectivity usually ranged between 85% and 90%. The rest of the MT4-infected cells were washed again to eliminate any virions from the supernatant and 10⁶ MT4-infected cells were seeded onto 24-well plates. For the 2D cell-free infection assay, 10⁶ Jurkat LTR-GFP CCR5⁺ cells were co-cultured on top of the MT4-infected cells using a transwell to physically separate them. For the 2D total infection assay, a similar procedure was performed, but here the transwells were omitted (Fig 11A). The co-cultures were treated with either DMSO or Torin2 at 10 nM for 18 hours and were subsequently analyzed by flow cytometry to identify the double GFP and mCherry population. Next, the infection efficacy was calculated by determining the percentage of GFP⁺ and mCherry⁺ cells in the presence of Torin2 over DMSO. This was measured for each individual virus, rather than one virus (i.e., Vpr_TFV) relative to another (i.e., Vpr_Lab) to account for potential differences in their replication kinetics. Remarkably, the Vpr_Lab virus achieved a ~ 77% infection efficacy in the total infection assay while the Vpr_mut and Vpr_TFV viruses achieved 85% and 96%, respectively, which was significantly higher. However, when these viruses harbored a Nef protein unable to counteract autophagy (Nef_mut), infection rates were between 52–65% (Fig 11B, left panel). By contrast, the susceptibility of Vpr to autophagy was irrelevant during infection in a 2D cell-free system, while the presence of autophagy-competent Nef proteins still provided a significant advantage during infection (Fig 11B, right panel).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 11. Vpr proteins from TFVs provide an advantage in virus spread.

(A) Experimental design for the 2D total (cell-to-cell plus cell-free) and cell-free infection assays. (B) Left panel: 2D total infection was examined by co-culturing infected MT4 cells with Jurkat LTR-GFP CCR5⁺ reporter cells in the presence of DMSO or Torin2 (10 nM). Right panel: 2D cell-free infection was measured in a similar manner. However, here donor (infected MT4) and recipient (Jurkat LTR-GFP) cells were separated by transwells. (C) Experimental design to measure 3D infection using the cervicovaginal epithelium model. (D) Left panel: 3D total infection was measured by infiltrating infected MT4 cells and recipient Jurkat LTR-GFP CCR5⁺ cells into the 3D epithelium. Right panel: 3D cell-free infection was determined by infiltrating Jurkat LTR-GFP CCR5⁺ into the epithelium and inoculating mCherry recombinant viruses onto the apical part of the cervicovaginal epithelium. %Infection efficacy was measured by determining the amount of mCherry⁺ and GFP⁺ cells in the Torin2 condition over the DMSO condition for each recombinant virus. Data correspond to the mean and the SEM of 3 biological replicates. *: p < 0.05; **: p < 0.01; ****: p < 0.0001; ns: not significant by One-way ANOVA. Experimental pipelines were created in BioRender (License: Serra-moreno, R. (2026) https://BioRender.com/fp56lo4).

https://doi.org/10.1371/journal.ppat.1014020.g011

We next used a 3D cervicovaginal epithelial model to assess how susceptibility to autophagy by Vpr impacts infection and spread in a more physiologically relevant system. For this, we used the model developed by Edwards and collaborators [88] (Fig S4) and we infiltrated the tissue with 10⁵ Jurkat LTR-GFP CCR5⁺ cells by making a wound in the epithelium (Fig 11C). For total infection, 2.5 x 10⁵ MT4 cells infected with each recombinant virus (donor cells) were juxtaposed to the Jurkat cells 2 hours after their infiltration, and cells were analyzed by flow cytometry 4 days later (Fig 11C, top). For cell-free infection, the different recombinant viruses were inoculated to the apical part of the 3D culture and cells were analyzed 4 days post-infection (Fig 11C, bottom). Regardless of the type of setup (total vs. cell-free), infection efficacy was significantly lower for viruses harboring autophagy-susceptible Vpr proteins and, once again, the ability of Nef to counteract autophagy was an important determinant for infection for all viruses (Fig 11D). In summary, these results indicate that Vpr proteins resistant to autophagy provide a fitness advantage in the context of autophagy-competent Nef proteins, at least in our 2D and 3D in vitro settings.

Plasmid DNA constructs

1. (i) Full-length proviral constructs.
   1. a. The following full-length proviral constructs were obtained through BEI resources. Wild-type HIV-1 NL4–3 (pNL4–3, #ARP-114) and NL4–3 Δnef (pNL4–3 Δnef, #ARP-12755) were obtained from Drs. Malcolm Martin and Olivier Schwartz, respectively [99,100]. HIV-1 viruses based on these plasmids were generated by transient transfection in HEK293T cells, as previously described [44,101,102].
   2. b. The proviral constructs for subtype A transmitted-founder virus clone A191084 (pBR322_HIV-1 M subtype A 191084) was provided by Dr. Konstantin Sparrer, University of Ulm, Germany. The subtype B transmitted-founder virus clones CH077 (pCH077.t/2627), and CH106 (pCH106.c/2633,) were obtained from Dr. John Kappes and Dr. Christina Ochsenbauer (University of Alabama at Birmingham). Clone CH470 was provided by Dr. Konstantin Sparrer, University of Ulm, Germany. The subtype C transmitted-founder virus clone Z3576 (pZ3576F) was obtained from Dr. Eric Hunter (Emory University) [103–105]. Clones CH042 and CH198 were provided by Dr. Konstantin Sparrer, University of Ulm, Germany. Clones Z1123M, 3137M, and N31F were obtained from Dr. Christina Ochsenbauer (University of Alabama at Birmingham).
   3. c. Full-length proviral constructs of HIV-1 NL4–3 harboring mCherry were generated following a previously described cloning approach [106]. Specifically, mCherry-furin-P2A-T2A was inserted between env and nef using the HapI and XhoI unique restriction sites. A version of this mCherry clone in which vpr harbored substitutions at positions 37 and 45 and another recombinant virus in which vpr was replaced by CH077 vpr was generated using the AgeI and NheI unique restriction sites (Fig S3). Virus versions harboring Alanine substitutions at residues 48 and 49 in Nef were also generated. These recombinant viruses were generated by transient transfection in HEK293T cells, as previously described [44,101,102]
2. (ii) Vpr constructs.

pALPS Vpr, a construct harboring codon optimized HIV-1 NL4–3 Vpr, was a gift from Jeremy Luban (Addgene #101329) [107]. This construct was later engineered to harbor a Myc tag in the C-terminus of the protein. HIV-1 transmitted-founder virus vpr genes were cloned into the expression vector pcDNA5 (ThermoFisher Scientific, V601020) using the KpnI and BamHI unique restriction sites. Vpr constructs harbored a Myc tag in their C-terminus. In the case that the Myc tag conflicted with the cloning restriction sites, either a Flag or a 6xHis tag was used. Vpr chimera I and chimera II were generated by overlapping PCR and were cloned into the expression vector pcDNA5 using the same strategy. Vpr mutants were generated by site-directed mutagenesis using QuickChange (Stratagene, #200518) and following the manufacturer’s instructions.

1. (iii) Vif constructs.

The HIV-1 NL4–3 Flag-Vif construct was a gift from Dr. Judd Hultquist (Northwestern University).

1. (iv) Lentivirus-based CRISPR-Cas9 genome editing constructs.

ATG5 knockout cell lines were generated using LentiCRISPRv2-ATG5, a gift from Dr. Edward Campbell (Addgene plasmid # 99573) [108]. SQSTM1, TAX1BP1 and NDP52 knockout cell lines were generated by CRISPR-Cas9. For this, the guide RNAs (gRNAs) that appear below were cloned into LentiCRISPRv2, a gift from Dr. Feng Zhang (Addgene plasmid# 52961) [109], using the BsmBI unique restriction site. Guide RNA sequences were generated using the CRISPR design tool (http://crispor.tefor.net). SQSTM1 gRNAs: 5’- CGACTTGTGTAGCGTCTGCG -3’, 5’- TCAGGAGGCGCCCCGCAACA-3’, and 5’- GCCCACGTCCTCCTCGTGCA-3’. TAX1BP1 gRNAs: 5’-GGCCGACTAAAGTTTCGAGC -3’, 5’- CAACGTACGAGACAGAACGA -3’, and 5’- TCCACGAATTTCACCCTTAT -3’. NDP52 gRNAs: 5’- GTGAGTATTACACCTTCATG -3’, 5’- GTATTACCAGTTCTGCTATG -3’, and 5’- GGAGGCGCAAGACAAAATCC -3’. The HIV-Gag-Pol packaging plasmid (psPAX2, Addgene plasmid# 12260; a gift from Dr. Didier Trono) was used to generate lentiviral particles harboring Cas9 and these gRNAs.

1. (v) VSV-G construct.

The plasmid for the expression of the Indiana variant of the Vesicular Stomatitis Virus (VSV) glycoprotein was a gift from Dr. David T. Evans (University of Wisconsin, Madison, WI).

1. (vi) Plasmids coding for autophagy proteins.

Human LC3B was cloned into the expression vector pQCXIP with an EGFP-tag attached to its N-terminus. MLV Gag-Pol, a gift from Dr. Patrick Salmon (Addgene plasmid # 35614) was used to generate retroviral particles encoding for EGFP-LC3B. The SQSTM1/p62 constructs SQSTM1 mCherry-WT, mCherry-ΔPB1, and mCherry-ΔUBA were gifts from Dr. Sascha Martens (Addgene plasmids # 187986, # 187985, and # 187984).

Transfections

HEK293T cells were transfected using GenJet in vitro DNA transfection reagent (SignaGen Laboratories, SL100488), following the manufacturer’s instructions, including total DNA, dish size, incubation time, and the ratio of GenJet:DNA for optimal transfection efficiency. Cell viability was monitored for every transfection to evaluate potential toxicity. If viability was below 85%, cells were considered unsuitable for further analyses. Viabilities were usually above 90%. For autophagy studies, cells were transfected with 2,000 ng of the indicated plasmid DNAs. Forty-four h later, cells were treated with either DMSO, Rapamycin, Torin and/or VPS34IN or Bafilomycin A1 (see drug treatments section for details). Next, cells were analyzed by western blot, immunofluorescence and/or immunoprecipitation.

Cell culture

Human HEK293T cells (American Type Culture Collection [ATCC], CRL-11268) and BJ fibroblast cells (ATCC, #CRL 2522) were cultured in complete medium (Dulbecco’s Modified Eagle Medium [DMEM, ThermoFisher Scientific, 11885–084] supplemented with 10% fetal bovine serum [FBS, ThermoFisher Scientific, 26140–079], 1% Penicillin-Streptomycin [ThermoFisher Scientific, 15070–063] and 1% L-glutamine [ThermoFisher Scientific, 25030–081]). HEK293T cells stably expressing EGFP-LC3 [44], HEK293T ATG5KO [110], SQSTM1, TAX1BP1, and NDP52 knockout cells were cultured in complete medium supplemented with 1 ng/mL of puromycin. VK2-E6-E7 vaginal epithelial cells (ATCC, #CRL 2616) were maintained in VK2 complete medium (Keratinoctye-SFM with BPE and EGF, Gibco, #10725–018) supplemented with 0.4 M calcium chloride (Amresco, #E506).

CD4⁺ T Jurkat (ATCC, TIB-152), MT4 (BEI, #ARP-120) and Jurkat LTR-GFP CCR5⁺ (BEI, #ARP-11586) cells were cultured in complete medium (RPMI 1640 Medium [ThermoFisher Scientific, 11875–119] supplemented with 10% fetal bovine serum [FBS, ThermoFisher Scientific, 26140–079], 1% Penicillin-Streptomycin [ThermoFisher Scientific, 15070–063] 1% L-glutamine [ThermoFisher Scientific, 25030–081]), and 2.5% HEPES [ThermoFisher Scientific, 15630080].

Normal human peripheral blood cryopreserved naïve CD4⁺ T cells (SER-PBCD4⁺TH-N-F) were purchased from ZenBio. Cells were activated using 25 μL anti-CD3/CD28 beads (Invitrogen, #111.31D), 1 μg/mL of IL-4 (Peprotech, #500-P24), 2 μg/mL IL-12 (Peprotech, #500-P154G), 1 ng/mL TGF-B (Peprotech, #100–21) and expanded for 3 days in RPMI medium (RPMI 1640 Medium [ThermoFisher Scientific, 11875–119] supplemented with 10% fetal bovine serum [FBS, ThermoFisher Scientific, 26140–079], 1% Penicillin-Streptomycin [ThermoFisher Scientific, 15070–063] 1% L-glutamine [ThermoFisher Scientific, 25030–081]), 2.5% HEPES [ThermoFisher Scientific, 15630080] in the presence of 30 IU/ mL of IL-2 (BEI #ARP-136).

Generation of knockout cells using CRISPR-Cas9 genome editing

Five million HEK293T cells were seeded in 10 cm dishes. Twenty-four h later, cells were transfected with 3.45 μg HIV-Gag-Pol packaging plasmid, 1.725 μg VSV-G envelope expressing plasmid, and 6.9 μg of either LentiCRISPRv2-ATG5, SQSTM1, TAX1BP1 or NDP52 plasmids. The supernatant was collected 48 h post-transfection and centrifuged for 10 min at 931 x g to remove the cell debris. Supernatants containing VLPs for CRISPR/Cas9 engineering were then aliquoted into 1 mL cryotubes and stored at -80ºC. To CRISPR-engineer HEK293T cells, cells were seeded in 25 cm² flasks. 24 h later, cells were transduced with VLPs harboring Cas9 and the gRNA targeting ATG5, SQSTM1, TAX1BP1 and/or NDP52. 48 h later, the cell medium was replaced and supplemented with puromycin. Cells were cultured under puromycin for 10 days to allow for the selection of cells successfully transduced with Cas9 and the corresponding gRNAs. Cells were then harvested for knockout assessment by western blotting and phenotypic analyses.

Infections

1. (i) One million Jurkat CD4⁺ T cells were infected with 100 ng of p24 equivalents of HIV-1 NL4–3 or NL4–3 Δnef by spinoculation at 1650 xg for 2 h at 37 °C. After infection, cells were washed and re-suspended in 5 mL of RPMI medium supplemented with 10% of fetal bovine serum. Cells were then cultured with increasing concentrations of Rapamycin (Sigma- Aldrich, #R8781): 1.3 μM, 2.6 μM, 4 μM, 6.5 μM, and 13 μM for 18 h. Cell lysates were subjected to western blot analysis 18 h post-infection.
2. (ii) One million primary CD4⁺ T cells (Zen-Bio, Inc., SER-PBMCD4⁺TH-N-F) were activated and infected with 100 ng of p24 equivalents of HIV-1 NL4–3 or HIV-1 NL4–3 Δnef by spinoculation at 1650 xg for 2 h at 37 °C. After the infection, cells were washed and re-suspended in 5 mL of RPMI medium supplemented with 10% of fetal bovine serum and 30 IU/mL of IL-2. One d post-infection, cells were either treated with DMSO, Rapamycin (0–13 μM at 2-fold increments; Sigma- Aldrich, #R8781) or Torin2 (2.5-20 nM Torin2 [Selleckchem, #S2817]) for 18 h. Cell lysates were then subjected to western blot analysis.

2D HIV infection assay

1. (i) Total (cell-free + cell-to-cell) HIV infection.

One million MT4 cells were infected with 100 ng of p24 equivalents of HIV NL4–3 Vpr_Lab mCherry, Mutant Vpr (Vpr_mut; harboring P₃₇Y₄₅ substitutions) mCherry, or Vpr_TFV mCherry by spinoculation at 1650 xg for 2 h at 37 °C. Similar infections were performed with these recombinant viruses harboring mutations in Nef that render the protein unable to counteract autophagy (Nef_mut). After the infection, cells were washed and re-suspended in 1 mL of RPMI medium supplemented with 10% of fetal bovine serum and seeded in a well of a 24-well plate and further incubated at 37 ºC. 2 d post-infection, cells were analyzed by flow cytometry to examine infectivity. Infectivity usually ranged between 85–90%. Next, 10⁶ infected MT4 cells were washed and re-suspended in 1 mL of RPMI medium supplemented with 10% of fetal bovine serum. At that moment, 10⁶ Jurkat LTR-GFP CCR5⁺ cells were added to the culture, in the presence or absence of Torin2 at 10 nM and incubated at 37 ºC. 18 h later, cells were collected for flow cytometry analysis.

1. (ii) Cell-free HIV infection.

One million MT4 cells were infected with 100 ng of p24 equivalents of HIV-1 NL4–3 Vpr_Lab mCherry, Vpr_mut mCherry, or Vpr_TFV mCherry harboring autophagy-competent or -defective Nef proteins by spinoculation at 1650 xg for 2 h at 37 °C. After the infection, cells were washed and re-suspended in 1 mL of RPMI medium supplemented with 10% of fetal bovine serum and seeded in a well of a 24-well plate and further incubated at 37 ºC. 2 d post-infection, cells were analyzed by flow cytometry to examine infectivity. Infectivity usually ranged between 85–90%. Next, 10⁶ MT4 infected cells were washed, re-suspended in 1.6 mL of RPMI medium supplemented with 10% of fetal bovine serum, and seeded in the bottom of a well of a 24-well with a transwell insert (Fisher, #07-000-676). Next, 10⁶ Jurkat LTR-GFP CCR5⁺ cells were re-suspended in 0.4 mL of RPMI medium supplemented with 10% of fetal bovine serum, and seeded on the top of the transwell in the presence or absence of Torin2 at 10 nM. Cells were incubated at 37 ºC. 18 h later, cells were collected for flow cytometry analysis.

3D HIV infection assay

The 3D cervicovaginal epithelium model was established following the protocol described by Edwards et al., [88]. Briefly, VK2 epithelial cells were polarized following the pipeline in S4 Fig in which collagen and BJ fibroblast cells provide support and mimic the epithelial-air interface. After 8 d in culture, the cervicovaginal epithelium is stratified, closely resembling the 3D structure in the female reproductive tract.

1. (i) Total (cell-free + cell-to-cell) HIV infection. To prepare donor cells, 10⁶ MT4 cells were infected by spinoculation with 100 ng of p24 equivalents of HIV-1 NL4–3 Vpr_Lab mCherry, Vpr_mut mCherry, or Vpr_TFV mCherry harboring autophagy-competent or -defective Nef proteins and cells were kept in culture for 2 d at 37 ºC. After this period of time, the VK2 epithelial cells were infiltrated with 10⁵ Jurkat LTR-GFP CCR5⁺ cells in 10 μL RPMI supplemented with 10% of fetal bovine serum. Two h later, 2.5 x 10⁵ infected MT4 cells were inoculated at the same location where Jurkat cells were infiltrated, and cells were incubated at 37 ºC. Four d later, cells in the apical compartment were collected for flow cytometry analysis.
2. (ii) Cell-free HIV infection. For this, 10⁵ Jurkat LTR-GFP CCR5⁺ cells were infiltrated into the VK2 epithelial cells in an analogous manner as explained above. Next, 1,500 ng of p24 equivalents of either HIV-1 NL4–3 Vpr_Lab mCherry, Vpr_mut mCherry, or Vpr_TFV mCherry harboring autophagy-competent or -defective Nef proteins were inoculated to the apical part of the 3D tissue and kept at 37 ºC. Four d later, cells in the apical compartment were collected for flow cytometry analysis.

Drug treatments

HEK293T cells were transfected with 2,000 ng of Vpr or Vif constructs. Forty-four h later, cells were treated with DMSO, Rapamycin (1.3-13 μM Rapamycin [Sigma-Aldrich, #R8781]), Torin2 (2.5-10 nM Torin2 [Selleckchem, #S2817]) or Torin2 (5–10 nM) plus Bafilomycin A1 (BafA1; 500 nM [Fisher, # J67193XFPM]). Treatments were kept for 4 h. Then samples were collected. Jurkat and primary CD4⁺ T cells were treated with Rapamycin (1.3-13 μM) or Torin 2 (2.5-20 nM) 18 h before sample collection. For the 3D HIV transmission assay, 10 nM Torin2 was added on day 1 and day 3 post-infection. To inhibit autophagy, cells were treated with 10 μM VPS34IN (Selleckchem, S7980) 18 h before sample collection. To investigate how Vif is degraded, the lysosomal inhibitor Hydroxychloroquine (Sigma-Aldrich, #H0915, 60 μM) and the proteasomal inhibitor MG132 (Sigma-Aldrich, #474791, 20 μM) were added to ATG5KO cells 18 h before sample collection. To inhibit ubiquitination, cells were treated with 200 nM TAK-243 (Fisher, #50-187-1707) 4 h before sample collection.

Western blotting

Transfected and infected cells were washed with DPBS and harvested in 0.3 mL lysis IP buffer (ThermoFisher Scientific, #87787) supplemented with protease inhibitors (Roche, #04693116001) and phosphatase inhibitor cocktails 2 and 3 (Sigma-Aldrich, #P5726 and #P0044). Cell lysates were then incubated on ice for 30 min. Samples were centrifuged at 16,000 xg at 4 °C for 8 min to remove cell debris. Supernatants were collected, mixed with 2x SDS sample buffer (Sigma-Aldrich, S3401), and boiled for 10 min on a heat block. Proteins were then separated by electrophoresis on 12% SDS-PAGE polyacrylamide gels and transferred to a polyvinylidene difluoride (PVDF) membrane (BioRad, #1620177) using a Trans-Blot Turbo Transfer System (Bio-Rad, #1704150EDU). Membranes were then incubated for 1 h with blocking buffer (Bio-Rad, #1706404) at room temperature, followed by an overnight incubation at 4 ºC with a primary antibody against our protein of interest (antibody sources and dilutions are detailed in Table 2). The following day, membranes were washed with PBS-tween (Sigma-Aldrich, #P3563) 3 times for 15 min at room temperature, followed by a 60-min incubation with the corresponding secondary antibody (antibody sources and dilutions are detailed in Table 2) at room temperature. Next, the membranes were washed 3 additional times for 15 min in PBS-tween. Finally, membranes were imaged with SuperSignal West Femto maximum sensitivity substrate (ThermoFisher Scientific, #34095), and visualized in a ChemiDoc MP imaging system (Bio-Rad, #12003154). The expression level of proteins was quantified using Image Lab software (Bio-Rad, Hercules, CA).

Immunoprecipitation assays

One million HEK293T cells were transfected with either 2,000 ng of Vpr constructs or 2,000 ng of Vpr constructs plus 2,000 ng of SQSTM1 plasmids. Forty-eight h post-transfection, cells were washed with DPBS and incubated with lysis IP buffer on ice for 30 min. After cell debris was eliminated by centrifugation (16,000 xg at 4 °C for 8 min), the whole cell lysates were incubated for 1 h at room temperature with Protein G magnetic beads (New England Biolabs, #S1430S). This step was performed to remove proteins unspecifically bound to the beads. In parallel, fresh Protein G magnetic beads were coated with an antibody against SQSTM1, NDP52, TAX1BP1, BNIP3L, NBR1, OPNT, or Vpr (Table 2) for 1 h at room temperature. Next, pre-cleared lysates were incubated with the antibody-coated Protein G beads overnight at 4 °C. The next day, the beads were washed with lysis IP buffer 5 times and resuspended in 2x SDS sample buffer. Samples were analyzed by western blotting by probing membranes with antibodies for Vpr, SQSTM1, NDP52, TAX1BP1, BNIP3L, OPNT, or ubiquitin (Table 2). As controls, cell lysates incubated with beads (but not coated with antibody; beads control) as well as a sample consisting of IP lysis buffer mixed with beads and antibody (IgG control) were included. These controls helped rule out any bands detected by western blot that corresponded to proteins unspecifically bound to the beads, material from the magnetic beads as well as the IgG heavy or light chains.

Fluorescence microscopy

One hundred thousand HEK293T or HEK293T EGFP-LC3B cells were seeded in sterile tissue culture-treated 8-well slides. Eighteen h later, cells were transfected with Vpr and/or Nef or SQSTM1 constructs. Forty-four h later, cells were treated with DMSO or Torin2. Four h later, cells were washed with ice-cold DPBS, permeabilized, and fixed by incubating in 1:1 acetone-methanol (Sigma-Aldrich, #270725, #34860) at room temperature for 1 h. Cells were then blocked with antibody diluent solution (2% fish skin gelatin [Sigma-Aldrich, 67765] + 0.1% Triton X-100 [Sigma-Aldrich, #X100] + 10% goat serum [ThermoFisher Scientific, #500062Z] with 1 x DPBS) for 30 min at room temperature and incubated for 60 min with the following primary antibody cocktail: antibody diluent solution + anti-Vpr/anti-SQSTM1/anti-NDP52/anti-TAX1BP1/anti-Nef at a dilution of 1:200 (Table 2). Subsequently, cells were washed with wash buffer (2% fish skin gelatin + 0.1% Triton X-100 with 1 x DPBS) 3 times and incubated with a secondary antibody cocktail (Table 2). Next, cells were washed with wash buffer 3 times and incubated with Hoechst (ThermoFisher Scientific, #H3570; 1:5,000 dilution) for 3 min to visualize cell nuclei. After staining, the slides were washed and mounted using an anti-quenching mounting medium (Vector Laboratories, #3304770). The slides were visualized in a BioTek Lionheart FX automated microscope and a Nikon A1R HD with TIRF confocal microscope using 40x and 60x/1.49 oil objectives, respectively, and filter cubes or excitation diodes 350 nm, 488 nm, and 586 nm in order to excite DAPI, EGFP, and TexasRed, respectively. Images were processed and analyzed using the Gen5 software (BioTek Instruments, Winooski, VT). Proportional adjustments of brightness and contrast were applied. Co-localization between Vpr and autophagy markers (EGFP-LC3B, NDP52, SQSTM1 and TAX1BP1) was determined by calculating the Pearson’s correlation coefficient of 8–25 independent fields using ImageJ (NIH https://imagej.nih.gov/ij/).

Flow cytometry

Cells from the HIV-1 infection assays were collected, washed with DPBS (ThermoFisher Scientific, #14190144), and centrifuged at 600 xg for 5 min. Samples were then fixed with 2% paraformaldehyde in DPBS. Cells were analyzed using an Attune instrument (ThermoFisher Scientific, Waltham, MA). Data were processed with FlowJo software (version 10.5.3) using 50,000 events. Debris and doublets were excluded using FSC and SSC gating, and the percentage of double positive GFP and mCherry cells was calculated for every sample and treatment condition.

Statistical analysis

Statistical calculations for 2-group comparisons were performed with a two-tailed unpaired Student T test. All other statistical comparisons were performed with One-Way ANOVA with Dunnet post hoc testing. Analyses were performed using Graph Pad Prism version 10.4.2. P values ≤ 0.05 were considered statistically significant.