The broadly neutralizing anti-human immunodeficiency virus type 1 4E10 monoclonal antibody is better adapted to membrane-bound epitope recognition and blocking than 2F5

Abstract

The broadly neutralizing 2F5 and 4E10 monoclonal antibodies (MAbs) recognize epitopes within the membrane-proximal external region (MPER) that connects the human immunodeficiency virus type 1 (HIV-1) envelope gp41 ectodomain with the transmembrane anchor. By adopting different conformations that stably insert into the virion external membrane interface, such as helical structures, a conserved aromatic-rich sequence within the MPER is thought to participate in HIV-1-cell fusion. Recent experimental evidence suggests that the neutralizing activity of 2F5 and 4E10 might correlate with the MAbs' capacity to recognize epitopes inserted into the viral membrane, thereby impairing MPER fusogenic activity. To gain new insights into the molecular mechanism underlying viral neutralization by these antibodies, we have compared the capacities of 2F5 and 4E10 to block the membrane-disorganizing activity of MPER peptides inserted into the surface bilayer of solution-diffusing unilamellar vesicles. Both MAbs inhibited leakage of vesicular aqueous contents (membrane permeabilization) and intervesicular lipid mixing (membrane fusion) promoted by MPER-derived peptides. Thus, our data support the idea that antibody binding to a membrane-inserted epitope may interfere with the function of the MPER during gp41-induced fusion. Antibody insertion into a cholesterol-containing, uncharged virion-like membrane is mediated by specific epitope recognition, and moreover, partitioning-coupled folding into a helix reduces the efficiency of 2F5 MAb binding to its epitope in the membrane. We conclude that the capacity to interfere with the membrane activity of conserved MPER sequences is best correlated with the broad neutralization of the 4E10 MAb.

FIG. 1.

Gp41 MPER sequence, derived peptides, and predicted membrane effects. (A) Membrane-proximal and transmembrane sequences of the HIV-1 gp41 integral subunit (sequence and numbering are derived from the prototypic HXBc2 isolate). The gray continuous line represents the average hydropathy plot calculated for a window of 5 amino acids using the WW hydrophobicity scale at membrane interfaces (81). The red dotted line is the mean WW moment (window of 11 amino acids) calculated for a fixed δ of 100° (helical periodicity) and hydrophobicity-at-interface scale (58). Full 2F5 and 4E10 epitopes are shown in red and green, respectively. Residues belonging to the transmembrane domain are shown in blue. The box below displays the peptide sequences used in this study. MPER designates the peptide used by Sun et al. (72). (B, left) Predicted membrane topology for the sequences depicted in panel A. The MPER 656-to-683 region is shown to consist of two segments: an amphipathic sequence (red) followed by a fully interfacial stretch (green). The tilt of the N-terminal amphipathic helix is based on the bipartite NMR structure observed in DPC micelles (72). Insertion of this interfacial element into the external monolayer of the virion membrane might be required to disrupt interactions between lipid molecules and be directly involved in the destabilization of membrane integrity. The differential-surface increase of the external membrane monolayer might also contribute to membrane deformation and thinning. (Right) Extraction of the epitope sequences by 2F5 or 4E10 would result in bilayer architecture stabilization, thereby interfering with fusion.

FIG. 2.

AISpreTM-induced membrane-restructuring in POPC-Chol (2:1, mol/mol) vesicles and its inhibition by the 2F5 and 4E10 MAbs. (A, left and center) Continuous lines correspond to the kinetic traces of leakage (left) or fusion (center) upon AISpreTM injection into stirred solutions of LUV (arrows) at a final peptide concentration of 1 or 4 μM, respectively. The dotted traces follow the incorporation of AISpreTM into the vesicles monitored through energy transfer from peptide tryptophans to membrane-residing d-DHPE. (Right) Percentages of AISpreTM-induced leakage (circles) and lipid mixing (squares) after a 30-min incubation with vesicles plotted as a function of the peptide concentration. The L B arrow indicates the condition selected to measure MAb-induced blocking of leakage and MAb binding to vesicles, and the F arrow indicates that selected to measure MAb-induced blocking of fusion. The final lipid concentration was 100 μM in all cases. (B) Inhibition of ANTS leakage (top) and mixing of vesicle lipids (bottom). LUV suspensions (100 μM lipid) were treated with 1 or 4 μM AISPreTM, respectively, and subsequently they were supplemented with 10 μg/ml of 2F5 or 4E10 MAbs (addition times indicated by 1 and 2, respectively). Dotted traces correspond to the controls in the absence of antibody. (C) Dose dependency of AISpreTM-induced ANTS leakage and fusion inhibition by the 4E10 MAb (circles) or 2F5 MAb (squares). The initial rates were determined as the increase in fluorescence during the 20 s that followed MAb addition. Inhibition percentages were plotted as a function of the MAb concentration, and the data are represented as the means ± standard deviations of three (leakage) and four (fusion) independent experiments.

FIG. 3.

Physical MAb-bilayer association as seen by electron microscopy in POPC-Chol (2:1) LUV. (A and B) Electron micrographs of control vesicles devoid of peptide and AISpreTM-containing vesicles (peptide-to-lipid mole ratio, 1:100), respectively. (D, E, G, and H) AISpreTM-containing vesicles incubated with 4E10 (D and E) or 2F5 (G and H) MAbs. These samples consist of vesicles (1 mM) incubated for 5 min with AISpreTM and then supplemented with 50 μg/ml MAb before staining with uranyl acetate. (C and F) Control samples devoid of the AISpreTM peptide. Bars = 100 nm.

FIG. 4.

Stable MAb-vesicle association as determined by flow cytometry. (A) The 4E10 and 2F5 MAbs (50 μg/ml) were incubated with a stirred solution of f-DHPE-labeled LUV (250 μM) under experimental conditions that were otherwise similar to those for leakage blocking (Fig. 2). After 10 min, the fluorescently labeled secondary antibody was added and the resulting mixtures were incubated for 5 min before being analyzed by flow cytometry. The incubation of AISpreTM-containing POPC-Chol vesicles with the 4E10 and 2F5 MAbs rendered double-labeled vesicles (center and top right panels, respectively). This pattern could not be observed in the absence of the MAbs (left panels) or when the MAbs were incubated with bare POPC-Chol vesicles (center and bottom right panels). (B) Differential association of the 4E10 and 2F5 MAbs with AISpreTM-containing POPC-Chol vesicles (1:100 peptide-to-lipid molar ratio) as determined by flow cytometry. The 4E10 (blue traces) and 2F5 (red traces) MAbs were incubated at the final concentrations (μM) indicated with AISpreTM-containing LUV under conditions that were otherwise similar to those in the panel A. Black traces correspond to control samples not incubated with MAb.

FIG. 5.

Dependence of MAb blockage of induced membrane restructuring and MAb-vesicle association on specific epitope recognition. (A) Binding of the 2F5 (red) and 4E10 (blue) MAbs to 1 μM AISpreTM(9,10)A (left) or AISpreTM(17,18)A (right) immobilized on ELISA plaques. (B) Blocking of leakage induced by the addition of 2 μM AISpreTM(9,10)A (left) or 1 μM AISpreTM(17,18)A (right) to POPC-Chol vesicles by the 2F5 (red) and 4E10 (blue) MAbs. The black trace corresponds to the control without MAb. The MAbs were added at a concentration of 50 μg/ml. The insets display the blocking induced by the same concentration of the 2F5 (left) or 4E10 (right) MAb on the leakage induced by parental AISpreTM. (C) Flow cytometry determination of the association of the 2F5 (red) and 4E10 (blue) MAbs with POPC-Chol vesicles bearing AISpreTM(9,10)A (left) or AISpreTM(17,18)A (right). MAbs were incubated with the vesicles at a concentration of 30 μg/ml.