A dynamic three-step mechanism drives the HIV-1 pre-fusion reaction

Abstract

Little is known about the intermolecular dynamics and stoichiometry of the interactions of the human immunodeficiency virus type 1 (HIV-1) envelope (Env) protein with its receptors and co-receptors on the host cell surface. Here we analyze time-resolved HIV-1 Env interactions with T-cell surface glycoprotein CD4 (CD4) and C-C chemokine receptor type 5 (CCR5) or C-X-C chemokine receptor type 4 (CXCR4) on the surface of cells, by combining multicolor super-resolution localization microscopy (direct stochastic optical reconstruction microscopy) with fluorescence fluctuation spectroscopy imaging. Utilizing the primary isolate JR-FL and laboratory HXB2 strains, we reveal the time-resolved stoichiometry of CD4 and CCR5 or CXCR4 in the pre-fusion complex with HIV-1 Env. The HIV-1 Env pre-fusion dynamics for both R5- and X4-tropic strains consists of a three-step mechanism, which seems to differ in stoichiometry. Analyses with the monoclonal HIV-1-neutralizing antibody b12 indicate that the mechanism of inhibition differs between JR-FL and HXB2 Env. The molecular insights obtained here identify assemblies of HIV-1 Env with receptors and co-receptors as potential novel targets for inhibitor design.

Figure 1. HIV-1 Env – CD4 and CCR5 or CXCR4 stoichiometry super-resolution imaging and fluorescence fluctuation spectroscopy in live cells.

a, The CD4 and CCR5 or CXCR4 receptors were labelled with mOrange and mTFP1. R5 and X4 tropic labelled HIV virions were produced using JR-FL and HXB2 Env and Gag-iCherry. The pre-fusion reaction of individual HIV-1 virions was assessed multiplexing real-time single virus tracking with two color number and brightness. b, Labelled CD4 receptors and co-receptors diffuse through a confocal volume generating fluorescence fluctuation traces that are informative of the oligomeric state of the labelled receptors as described in the NandB method. When cross-correlating both traces, the one coming from CD4-mOrange and the one coming from the co-receptor (e.g. CCR5-mTFP1) one can detect protein-protein interactions, as described in ccNandB theory. c, Flow diagram depicting the overall strategy from image acquisition until analysis. COS7 cells co-expressing CD4-mOrange and CCR5 or CXCR4-mTFP1 were exposed to HIV_{JR-FL/Gag-iCherry} or HIV_{HXB2/GagiCherry} virions and imaged using a confocal system equipped with two HyD detectors and one PMT (1, first column from the left). Number and Brightness analysis was performed on time-stacks of images for both CD4-mOrange Channel and CCR5-mTFP1 or CXCR4-mTFP1; in parallel single virus tracking (SVT) was performed on the third Gag-iCherry channel. Pixel-by-Pixel Brightness images containing information on the oligomeric state for CD4 and CXCR4 were produced using a software developed in house in R (2, second column from the left)). Finally, co-localization analysis was carried out between the x-y coordinates for the labelled virions (in this example HIV_{HXB2/GagiCherry}) and the cross-variance (Bcc) (3, third column from the left) coming from CD4 and CXCR4 channels (4, fourth column from the left). The computation of Bcc was also carried out with a second R package developed in house.

Figure 2. Visualization of HXB2 based HIV-1 virion receptor stoichiometry in live cells.

a, Fast time-resolved, three-color imaging was performed on a COS7 cell co-expressing CXCR4-mTFP1 (green micrographs) and CD4-mOrange (red micrographs) exposed to HIV_{HXB2 –Gag-iCherry} virions (cyan micrographs), scale bar 1 μm. b, Time-resolved brightness analysis was performed for both CXCR4-mTFP1 (first column from the left) and CD4-mOrange (second column from the left). Time-resolved cross-variance (Bcc) analysis was also performed (third column from the left) rendering a small region of interest in which CD4-CXCR4 interactions occurred (green). Real-time single virus tracking was performed in parallel and the micrographs show co-localization of HIV_{HXB2 –Gag-iCherry} (cyan pixels) and positive Bcc (green pixels). c, A second region of interest (dashed square) shows a HIV_{HXB2-Gag-iCherry} virus that did not induce positive Bcc (d). These viruses most likely were immature and constitute robust built-in controls for our N&B and Bcc analysis. e, Time-resolved stoichiometry for CD4 (red dots) and CXCR4 (green dots) upon addition of HIV_{HXB2/GagiCherry} (left panel). The circles indicate the average (n = 10) and the error bars indicate the standard error for each time point. The time-resolved homotypic interactions for CD4 (red dots) and CXCR4 (green dots) upon addition of HIV_{HXB2/GagiCherry} in the presence of inhibitory concentrations of b12. CD4 and CXCR4 did not interact in this case. Eror bars indicate standard error for each time point wich is the average for n = 12.

Figure 3. Visualization of HIV-1 virion receptor stoichiometry in cells with multi-colour dSTORM.

a, The labelled CD4 and CCR5 or CXCR4 receptors were labelled again with nanoboosters (Chromotek) specifically engineered for dSTORM imaging. b, HIV virions were exposed to COS7 cells and 10 minutes after fixed and imaged in a TIRF-dSTORM set up equipped with two EM-CCD cameras (Zeiss Elyra dual-cam, see Material and Methods section). The pre-fusion reaction of individual HIV-1 virions was assessed by co-localization analysis of single molecules with an average axial resolution of 30 nm per event. Colocalization positive images were produced to generate masks to recover their stoichiometry, defined as the normalized sum of photons per interaction event. CD4 and CCR5 or CXCR4 co-localization masks were used against the JR-FL or HXB2 labeled Env super-resolution image to recover the number of HIV-1 Env engaged with CD4 –CCR5 or CD4 –CXCR4 complexes. c, The total number of normalized events per interaction area of CXCR4 labeled with Atto 488 was plotted against the total number of events per interaction area of CD4 labeled with Atto 642 for COS7 cells exposed to HIV_HSB2 also labelled with Alexa 405 against the Env (as described in material and methods) (top chart, n = 103). Three different regions corresponding to the three steeps of the pre-fusion reaction are populated only for cells exposed to HIV_HXB2. In the bottom chart, the total number of HXB2 Env interacting with CD4 and CXCR4 labelled with Alexa 405 are plotted against the total number of CD4 – CXCR4 interacting complexes (n = 18).

Figure 4. Visualization of JR-FL based HIV-1 virion receptor stoichiometry in live cells.

a, Fast time-resolved, three-color imaging was performed on a COS7 cell co-expressing CCR5-mTFP1 (green micrographs) and CD4-mOrange (red micrographs) exposed to HIV_{JR-FL –Gag-iCherry} virions, scale bar 1 μm. The brightness histogram for each channel together with the pixel by pixel Brightnes maps are also presented (second row). In the last row, microsgraphs corresponding to the HIV_HXB2-iCherry (in red) togher with the Bcc map is shown. Green regions represent positive Bcc. The time-resolved co-localization map coming from the white squares is also presented (third row, right panels) and correspond to regions of 0.5 X 0.5 μm. b, Time-resolved stoichiometry for CD4 (red dots) and CXCR4 (green dots) upon addition of HIV_{JR-FL/GagiCherry} (right panel). The circles indicate the average value (n = 12) and the error bars indicate the standard error for each time point c, Time-resolved homotypic interactions for CD4 (red dots) and CCR5 (green dots) upon addition of HIV_{JR-FL/GagiCherry} in the presence of inhibitory concentrations of b12. CD4 and CCR5 did not interact in this case. Eror bars indicate standard error for each time point which represents the average (n = 14). c, The total number of normalized events per interaction area of CCR5 labeled with Atto 561 was plotted against the total number of events per interaction area of CD4 labeled with Atto 488 for COS7 cells exposed to HIV_JR-FL also labelled with Alexa 633 against the Env (as described in material and methods) (n = 23). The middle panel shows the total number of HXB2 Env interacting with CD4 and CXCR4 labelled with Alexa 405 are plotted against the total number of CD4 – CXCR4 interacting complexes (n = 12). The right panel shows the distribution of all CD4 – Env interactions and the relative frequency of their stoichiometry (n = 36).

Figure 5. A three-step stoichiometric model for HIV-1 Env-host receptor interactions.

a, The HIV-1 JR-FL Env glycoprotein (Blue) (PDB ID: 4ZMJ) is present as a trimer on the mature virion. b, Time-resolved stoichiometry pre-fusion reaction for CXCR4-CD4 (red dots) induced by HIV_{HXB2–Gag-iCherry} virions. (gp120: light blue, gp41: dark blue, CD4: orange, CXCR4: green, b12: yellow)

Figure 6. b12 disrupts CD4 co-receptors interactions.

a, Model showing how b12 totally impedes CD4 and CCR5 interactions inhibiting the fusion reaction from not allowing step 1 to occur. b, Model depicting one possibility for the targeting of mAb12 (yellow) to the CD4 binding region of HIV-1 Env, which prevents fusion into the host cell, but still allows CXCR4 oligomerisation, possibly due to incomplete coverage on the Env spike. Non-interacting time-resolved oligomeric states for CD4 and CXCR4 following HIV_{HXB2 –Gag-iCherry} virions in the presence of 100 μg/mL b12.