HIV-1 Nef hijacks clathrin coats by stabilizing AP-1:Arf1 polygons

Abstract

The lentiviruses HIV and simian immunodeficiency virus (SIV) subvert intracellular membrane traffic as part of their replication cycle. The lentiviral Nef protein helps viruses evade innate and adaptive immune defenses by hijacking the adaptor protein 1 (AP-1) and AP-2 clathrin adaptors. We found that HIV-1 Nef and the guanosine triphosphatase Arf1 induced trimerization and activation of AP-1. Here we report the cryo-electron microscopy structures of the Nef- and Arf1-bound AP-1 trimer in the active and inactive states. A central nucleus of three Arf1 molecules organizes the trimers. We combined the open trimer with a known dimer structure and thus predicted a hexagonal assembly with inner and outer faces that bind the membranes and clathrin, respectively. Hexagons were directly visualized and the model validated by reconstituting clathrin cage assembly. Arf1 and Nef thus play interconnected roles in allosteric activation, cargo recruitment, and coat assembly, revealing an unexpectedly intricate organization of the inner AP-1 layer of the clathrin coat.

Fig. 1

Arf1, cargo, and Nef trimerize and unlock AP-1. (A) Size exclusion chromatography of AP-1 complexes with the partners indicated with color codes at right. Peak 1 “PK1” indicates the high molecular weight AP-1:Arf1:tetherin-Nef, whose central fraction was used for subsequent biophysical and cryoEM studies. The gel shown at right was loaded with samples from PK1 for AP-1:Arf1:tetherin-Nef, and the highest molecular weight peak eluted for each of the other samples tested. (B) Multi-angle light scattering (MALS) of AP-1:Arf1:tetherin-Nef (red), and free AP-1 (blue). (C) Emission spectra of free AP-1 and the complex with AP-1:Arf1:tetherin-Nef, with excitation at 532 nm. The increase in Cy3 emission and decrease of Cy5 emission shows the FRET efficiency decreases in the complex. The original intensity counts were normalized by the estimation of total Cy5 amount. (D) Distance changes anticipated between the locked and unlocked AP-1 complexes, obtained from the indicated PDB entries. The distance increase from 17 to 58 Å corresponds to a two-fold decrease in FRET efficiency difference for the Cy3-Cy5 pair ( ${\mathrm{R}}_0^6=60Å$) and $\mathrm{E}={\mathrm{R}}_0^6/({\mathrm{R}}_0^6+{\mathrm{R}}^6)$.

Fig. 2

CryoEM reconstruction of two types of AP-1:Arf1:tetherin-Nef trimers. (A) Representative 2D/3D classes of trimers. (B) Relative population of dimers and closed and open trimers. (C) Submasking and 7 Å reconstruction of a subassembly of the closed trimer. (D) Overall view of the docking of hyper-unlocked AP-1 structure, Arf1, and Nef into closed trimer reconstruction.

Fig. 3

Molecular interactions in the closed AP-1:Arf1:tetherin-Nef trimer. (A) Exploded view of subunit-by-subunit fits into the closed trimer reconstruction. (B) Details of interactions between Arf1 and β1 and γ subunits. (C) Interactions between Arf1 and μ1 CTD drive hyper-unlocking. (D) Interactions at the center of the Arf1 nucleus stabilize the trimer. (E) Interactions between Nef and the μ1-CTD (left) mirror those seen previously (21), while a distinct trimer-bridging interaction is observed between Nef and the N-terminus of β1(right). (F) The closed AP-1:Arf1:tetherin-Nef trimer. The reported membrane and cargo binding subunits including Arf1, μ1 and Nef are kept in their original colors, while the other parts are colored grey.

Fig. 4

The open AP-1:Arf1:tetherin-Nef trimer suggests a hexagonal assembly model. (A) Reconstruction and docking of the open AP-1:Arf1:tetherin-Nef trimer. (B) The open AP-1:Arf1:tetherin-Nef trimer. The reported membrane and cargo binding subunits including Arf1, μ1 and Nef are kept in their original colors, while the other parts are colored grey. The membrane is represented by a line. The C-terminus of the β1 core, which is the location of the clathrin binding region, is colored purple. (C) The open AP-1:Arf1 trimer formed in the presence of Nef (left) and the known Nef-free AP-1:Arf1 dimer (to the right of the plus sign) (9) were docked onto one another by overlaying one copy each of the β1 subunit and Arf1^β (outlined in black) to yield the composite structure at far right. (D) Iterating the docking operation shown in (C) generates the hexagon shown. Nef was docked onto all copies of AP-1 in the mode identified in the closed trimer, and is shown in yellow. (E) Juxtaposition of the modeled hexagon and one hexagonal segment of the cryoEM reconstruction of the clathrin D6-barrel (29). The membrane binding face is represented by a curved line. (F) A close-up of the hexagon model and the clathrin D6-barrel density shows that the clathrin binding site on β1 adjoins the AP-1 binding site on clathrin.

Fig. 5

Visualization and validation of AP-1 polygons. (A) Negative stain EM of the high molecular weight tail of the size exclusion separation of AP-1:Arf1:MHC-I-Nef. The top row shows representative particles. The second row shows 2D class averages, which were carried out to quantitate the relative frequency of different assemblies. The number of particles within each class is indicated. The class averages show blurring and loss of density at the edges, indicative of heterogeneity within the classes. (B) The corresponding docked model. The scale is as same as in (A), indicating good agreement with the single particle images. (C) Location of Arf1 mutants in interfaces. Arf1 is red, while the other parts are colored grey. (D) Size exclusion chromatography of AP-1:Arf1:Nef mixtures, showing that Arf1 mutants interfere with dimerization or trimerization of AP-1. In Arf1^Δ148-152GS, residues L₁₄₈RHRN₁₅₂ were replaced by AGSGS. (E) Negative stain EM of clathrin cages assembled in vitro from purified clathrin, AP-1 core including full-length β1 (“FLβ.AP-1”, following terminology used for AP-2 (38)), wild-type Arf1, Nef, and the non-hydrolyzable GTP analogue GppCp. (F) The relative percentage of clathrin cages were counted in 22 randomly chosen fields of view for the indicated mixtures of wild-type and/or mutant Arf1 proteins.

Fig. 6

The hexagonal inner coat promotes clathrin cage assembly. (A) Diagram of Nef mutants. Structures from AP-2-α-σ2:Nef (PDB ID 4NEE) and AP-1-μ1:MHC-I-Nef (PDB ID 4EMZ) were combined and homologs are colored as before. (B) Size exclusion chromatography of AP-1:Arf1:Nef mixtures, showing that Nef mutants interfere with dimerization or trimerization of AP-1. (C) The relative percentage of clathrin cages were counted in 20 randomly chosen fields of view for the indicated mixtures of wild-type and/or mutant Nef proteins. (D) Size exclusion chromatography of AP-1:Arf1:MHC-I-Nef mixtures, showing that MHC-I-Nef mutants interfere with dimerization or trimerization of AP-1. (E) The relative percentage of clathrin cages were counted in 20 randomly chosen fields of view for the indicated mixtures of wild-type and/or mutant MHC-I-Nef proteins. (F) Concept for the role of AP-1:Arf1 polygons in clathrin assembly with Nef. Nef (yellow) stabilizes the AP-1:Arf1 polygons increases such that closed polygons can form even without clathrin, creating a pre-formed template for rapid clathrin assembly.