HIV-1 Nef disrupts MHC-I trafficking by recruiting AP-1 to the MHC-I cytoplasmic tail

Abstract

To avoid immune recognition by cytotoxic T lymphocytes (CTLs), human immunodeficiency virus (HIV)-1 Nef disrupts the transport of major histocompatibility complex class I molecules (MHC-I) to the cell surface in HIV-infected T cells. However, the mechanism by which Nef does this is unknown. We report that Nef disrupts MHC-I trafficking by rerouting newly synthesized MHC-I from the trans-Golgi network (TGN) to lysosomal compartments for degradation. The ability of Nef to target MHC-I from the TGN to lysosomes is dependent on expression of the mu1 subunit of adaptor protein (AP) AP-1A, a cellular protein complex implicated in TGN to endolysosomal pathways. We demonstrate that in HIV-infected primary T cells, Nef promotes a physical interaction between endogenous AP-1 and MHC-I. Moreover, we present data that this interaction uses a novel AP-1 binding site that requires amino acids in the MHC-I cytoplasmic tail. In sum, our evidence suggests that binding of AP-1 to the Nef-MHC-I complex is an important step required for inhibition of antigen presentation by HIV.

Figure 1.

HIV-1 Nef redirects MHC-I from the TGN to lysosomes. (A and B) Nef does not affect the maturation of HLA-A2 through the TGN. CEM HA-HLA-A2 cells were transduced with a control adenovirus (nef ⁻) or an adenovirus expressing HIV-1 Nef (nef ⁺) and subjected to a pulse-chase metabolic labeling assay to follow proteins through the biosynthetic pathway. HA-HLA-A2 species with different gel mobilities were identified based on enzymatic digestion profiles: (a) N-glycosylated and sialylated HA-HLA-A2; (b) N-glycosylated HA-HLA-A2; and (c) core HA-HLA-A2 protein. (B) Quantitation of the relative percentage of endo H resistant and sialylated HA-HLA-A2. The mean percentage ± SD from two independent experiments is plotted over time. (C) Nef-induced degradation of mature MHC-I is blocked by inhibitors of acidic degradation. CEM HA-HLA-A2 cells were treated with adenovirus, pulsed with radioactive amino acids, and were either collected immediately (lanes 1 and 2) or chased for 4 h in media containing the indicated chemical inhibitor (lanes 3–12). Cellular lysates were divided equally, and HA-HLA-A2 or transferrin receptor was recovered by immunoprecipitation. All samples were digested with endo H before SDS-PAGE. The results are representative of three independent experiments. (D) Bafilomycin A1 treatment increases the degree of colocalization between HLA-A2 and LAMP-1. Cells transduced with the indicated adenovirus were treated with either solvent alone (DMSO) or bafilomycin A1 (Baf A1) for 4 h before staining. HLA-A2 and LAMP-1 were detected by indirect immunofluorescence using mAbs as described in Materials and methods. Arrows indicate colocalization between HLA-A2 and LAMP-1. Images were collected using a confocal microscope. Individual z-sections are shown. Bar, 5 μm.

Figure 2.

HIV-1 Nef targets MHC-I into an AP-1–dependent pathway. (A and B) Western blot analysis of adaptin and Nef expression in siRNA-treated cells. Astrocytic cells (A) or CEM T cells (B) were transfected with the indicated siRNA and transduced with control or Nef-expressing adenoviruses as described in Materials and methods. Lysates were harvested and immunoblotted for the indicated protein. (C and D) Flow cytometric analysis of siRNA-treated cells. Astrocytic cells (C) or CEM T cells (D) were stained for surface HLA-A2 expression (left column) or surface CD4 expression (right column); shaded curve, control adenovirus; black line, Nef-expressing adenovirus. (E) Depletion of AP-1 inhibits Nef-dependent HLA-A2 degradation. CEM T cells were transduced and treated with siRNAs as described above except that an additional siRNA transfection was included before transduction. Lysates were prepared and analyzed by immunoblot to examine the expression levels of the indicated proteins.

Figure 3.

Nef expression results in coprecipitation of AP-1 with HLA-A2. (A) HLA-A2⁺ primary human T cells or control CEM-SS cells that did not express HLA-A2 (Ctrl) were transduced with an HIV molecular clone expressing Nef (nef ⁺), or a matched control HIV that did not express Nef (nef⁻). Cells were harvested, and HLA-A2 was immunoprecipitated as described in Materials and methods. Proteins that coprecipitated with HLA-A2 were detected by Western blotting as indicated. Western blots of protein inputs are also shown as a control for relative protein levels in each sample before immunoprecipitation (Input Controls). Results are typical of two independent experiments. (B) AP-3 does not coprecipitate with HLA-A2. CEM-SS cells (lanes 1 and 4) or CEM HA-HLA-A2 cells (lanes 2, 3, 5, and 6) were infected with HIV, and HLA-A2 was recovered by immunoprecipitation as described in Materials and methods. Coprecipitating proteins were detected by Western blotting (lanes 1–3). A fraction of the protein input before immunoprecipitation was also analyzed to ensure similar protein expression levels (lanes 4–6). Western blots of δ-adaptin (AP-3) and γ-adaptin (AP-1) were performed simultaneously using the same antibody concentrations. Apparent molecular masses of protein standards are denoted in kilodaltons on the left.

Figure 4.

Analysis of MHC-I and Nef domains that contribute to AP-1 recruitment. (A) Recruitment of AP-1 requires an intact Nef–MHC-I complex and is independent of the dileucine motif in Nef. CEM-SS cells (lane 1) or CEM HA-HLA-A2 cells (lanes 2–5) were transduced with the indicated adenoviral vector. The cells were harvested at 72 h and subjected to the coimmunoprecipitation assay. Cells were transduced with each adenovirus to achieve equivalent Nef expression levels (Input Controls): wild-type NL4-3 Nef (MOI = 25); NL4-3 Nef with a NH₂-terminal deletion (V₁₀EΔ17-26, MOI = 300); HXB Nef with a mutation in the dileucine motif (LL_164,165AA, MOI = 25). (As shown, HXB Nef migrates more slowly on SDS PAGE than NL4-3 Nef. Wild-type HXB Nef and HXB NefLL_164,165AA are directly compared in Fig. S3.) All results are typical of at least two independent experiments. (B) HLA-A2 tyrosine 320 (Y₃₂₀) is required for efficient Nef binding, AP-1 recruitment, and targeting for degradation. CEM-SS (lanes 1 and 2), CEM HA-HLA-A2 (lanes 3 and 4), or CEM cells expressing the MHC-I mutant, HA-HLA-A2 Y₃₂₀A (lanes 5 and 6), were transduced with a control (−) or Nef-expressing adenovirus (+). At 72 h after transduction, HLA-A2 immunoprecipitations were recovered and analyzed by Western blotting as described in Fig. 3.

Figure 5.

Characterization of the A2/Nef fusion protein. (A) Schematic diagram of the A2/Nef fusion protein. (B) The addition of Nef to the COOH terminus of HLA-A2 results in a reduction of cell surface expression in cis but not trans. In the left panel, the effect of Nef in cis was measured by transducing CEM-SS cells with a bi-cistronic murine retrovirus encoding GFP and HLA-A2 (shaded curve), GFP and A2/Nef (black line), or GFP alone (gray line). HLA-A2 was detected by staining with mAb BB7.2 and using flow cytometry to gate on the GFP-positive cells. In the right panel, the effect of A2/Nef in trans was measured by examining the effect of the fusion protein (black line) or wild-type HLA-A2 (shaded curve) on stably expressed HA tagged HLA-A2 (HA-HLA-A2). The negative control is CEM-SS cells transduced with virus expressing GFP only (gray line). HA-HLA-A2 was detected by staining with an mAb directed against HA and using flow cytometry to gate on the GFP-positive cells. Results are representative of three independent experiments. (C) The A2/Nef fusion protein accumulates in the TGN. CEM-SS cells that stably expressed a tagged late Golgi resident protein (EYFP-Golgi) were transduced with retroviral constructs (A2 or A2/Nef). Cells were stained for HLA-A2 and analyzed by confocal microscopy as described in Materials and methods. Bar, 5 μm.

Figure 6.

The cytoplasmic domain of HLA-A2 and the NH ₂ -terminal α helix of Nef are both required for AP-1 binding. (A and B) The NH₂-terminal α-helix in Nef is necessary for AP-1 recruitment. CEM-SS cells were transduced with the murine retrovirus encoding HLA-A2 or the indicated A2/Nef fusion constructs. Coprecipitating proteins were detected by Western blotting of anti-HLA-A2 immunoprecipitations. (C) AP-1 binding to the A2/Nef fusion protein requires tyrosine 320 (Y₃₂₀) in the cytoplasmic tail of HLA-A2. The indicated mutations in the A2/Nef fusion protein were tested for involvement in AP-1 binding using the assay described above. A2ΔTail is an A2/Nef fusion protein that lacks the entire HLA-A2 cytoplasmic tail. All results are representative of at least three independent experiments. White lines indicate that intervening lanes have been spliced out.

Figure 7.

Model for MHC-I trafficking in Nef-expressing T cells. To disrupt MHC-I trafficking, Nef first binds to the MHC-I cytoplasmic tail in the secretory pathway. The formation of the Nef–MHC-I complex creates a binding site for AP-1 that is independent of the Nef dileucine motif. This tertiary complex leads to the recruitment of MHC-I into AP-1–positive clathrin-coated vesicles destined for degradation in the lysosomes.