Distinctive Membrane Accommodation Traits Underpinning the Neutralization Activity of HIV-1 Antibody against MPER

Abstract

The membrane-proximal external region (MPER), located in the carboxy-terminal section of HIV's envelope glycoprotein (Env) ectodomain, which is essential for viral entry into host cells, has gained considerable attention as a target for HIV vaccine development due to the exceptional neutralization breadth of antibodies against MPER epitopes. A distinctive feature of broadly neutralizing antibodies (bnAbs) targeting MPER is their requirement to accommodate the viral membrane into the surface of the antigen-binding fragment, or Fab moiety, to optimize antigen recognition. In this study, we sought to elucidate the molecular mechanism behind this interaction and its relevance to the antiviral function of bnAb 10E8. We conducted all-atom molecular dynamics simulations of three systems: (i) Fab 10E8 positioned on the surface of a viral-like lipid bilayer (VL-LB), (ii) Fab 10E8 in complex with an MPER helix anchored to the VL-LB via the Env glycoprotein transmembrane domain (TMD), and (iii) a Fab/MPER-TMD complex similarly embedded in the VL-LB but with a chemically optimized Fab 10E8 variant showing enhanced potency. Comparing these systems enabled us to derive atomic-scale Fab-membrane accommodation profiles pertinent to 10E8's neutralizing function. Our findings support that Fab adaptation to the viral membrane interface following epitope binding is crucial for developing MPER-targeted neutralizing activity. This analysis also provides insights into pathways for strengthening lipid interactions, which may prove valuable in designing MPER-based biologics and vaccines to prevent or treat HIV infection.

Keywords: HIV-1 neutralization mechanism; MD simulations; MPER antibody; antibody engineering; antibody–membrane interaction; site-selective chemical modification; therapeutic antibody.

Figure 1

Probing the membrane-accommodating surface of Fab 10E8 through chemical modification with Fus4. Left: location within the HIV Env glycoprotein and the mode of recognition of the 10E8 MPER epitope. Right, top: solvent-exposed residues facing the membrane from the CDRL1 (LC.S30), FRL3 (LC.S65), and CDRH3 (HC.S100c), which were considered for single site-directed modification, are indicated in orange. Right, bottom: antiviral activity tested in cell entry assays. Reduction of the IC₅₀ values (μg/mL) denoted higher neutralization potency of the antibody. A Fab modified after introducing a Cys at the N-terminus of the HC (C216-Fus4) was assayed as a negative control. Plotted values are means ± SD of three independent experiments (see also Supporting Information Figure S1A).

Figure 2

MD simulations of CDRL1, FRL3, and CDRH3 loops of the Fab 10E8 implanted into the VL-LB. (a) RMSD of three loop regions for each of the three systems. CDRL1 (LC Cys23-Ser34) in red, CDRH3 (HC Tyr98-Gly100h) in black, and FRL3 (LC Asp60-Ala78) in blue. The analysis was performed on frames extracted every 100 ps from each trajectory. (b) Specific Fab-MPER helix contacts in the “Fab/MPER” and “Fab-Fus4/MPER” systems were defined based on atomic proximity, where any heavy atom of an antibody residue (paratope/Fab) within 3.5 Å of any heavy atom in MPER (epitope) was considered a contact. The analysis was performed on frames extracted every 1 ns from each trajectory.

Figure 3

Interactions with VL-LB lipids of CRL1, FRL3, and CDRH3 loop residues in MD simulations of Fab 10E8. (a) Right: lipid components of the VL-LB mixture. Left: density profiles of the different VL-LB components. (b) Contacts with VL-LB lipids established by CDRL1, FRL3, and CDRH3 loop residues selected from maps displayed in Figure S2. Plotted values are means ± SD. Means were calculated after registering in frames collected at 1 ns intervals the number of contacts established by a given residue with a given class of lipid (heavy atom distance <3.5 Å) in Fab (gray), Fab/MPER (blue), and Fab-Fus4/MPER (green) simulations. N values amounted to 2662, 3480, and 2873 for Fab, Fab/MPER, and Fab-Fus4/MPER simulations, respectively. When contacts were detected, one-way ANOVA tests confirmed significant differences among the three simulations (p < 0.05). Horizontal bars indicate samples subjected to pairwise t-Student analysis (***p < 0.0005).

Figure 4

SPM filling of the binding site in the MD simulations of Fab 10E8. (a) Structure of the binding site (left) and number of contacts (heavy atom distance <3.5 Å) established by the LC.Tyr32 residue with the different VL-LB lipids (right). Conditions otherwise as in previous Figure 3b. (b) Contacts (heavy atom distance <3.5 Å) of SPM phosphocholine moiety with Fab residues from CDRL1, FRL3, and CDRH3 loops in simulations of the three systems. Black represents the maximum of 1 contact between single residue and phosphocholine headgroup. (c) Contacts established with VL-LB lipids of the same monolayer by the SPM molecule that remained in contact with Tyr32 in most of the trajectories. Number of contacts measured using frames at 1 ns intervals, between the SPM residue, with its headgroup occupying the cavity formed around the residue Tyr32 of the LC-CDR1 loop, and the other lipids in the upper leaflet of the membrane.

Figure 4

SPM filling of the binding site in the MD simulations of Fab 10E8. (a) Structure of the binding site (left) and number of contacts (heavy atom distance <3.5 Å) established by the LC.Tyr32 residue with the different VL-LB lipids (right). Conditions otherwise as in previous Figure 3b. (b) Contacts (heavy atom distance <3.5 Å) of SPM phosphocholine moiety with Fab residues from CDRL1, FRL3, and CDRH3 loops in simulations of the three systems. Black represents the maximum of 1 contact between single residue and phosphocholine headgroup. (c) Contacts established with VL-LB lipids of the same monolayer by the SPM molecule that remained in contact with Tyr32 in most of the trajectories. Number of contacts measured using frames at 1 ns intervals, between the SPM residue, with its headgroup occupying the cavity formed around the residue Tyr32 of the LC-CDR1 loop, and the other lipids in the upper leaflet of the membrane.

Figure 5

Phospholipid-binding sites in structures derived from cluster analyses. (a) Cluster analysis of “Fab”, “Fab/MPER”, and “Fab-Fus4/MPER”. (b) Position of the phospholipid-binding site with respect to the bilayer plane in the extracted structures (membrane phosphate level indicated by the orange line). Side-chains of HC.W100b, HC.S100c, LC.Tyr32, and LC.Arg29 are displayed for better appreciation of the displacement of the constituent loops. The bound SPM molecule depicted in magenta portraits choline and phosphate groups encircled (blue and orange circles, respectively). Position of the LC.Arg29 Cα is indicated by the gray circle. Intramolecular (HC.Ser100c-LC.Tyr32 and HC.Asp100-LC.Tyr32) and intermolecular (SPM-LC.Gly68 and SPM-LC.Arg29) H-bonds are highlighted by the lines in cyan. (c) Lipid nanoenvironment surrounding the SPM ligand. VL-LB lipids making contact at a distance <3.5 Å are depicted together with the Fab-bound SPM molecule (VW surface in magenta).

Figure 6

Somatically mutated residues in membrane-accommodating surfaces of the functional Fab 10E8. (a) Positions in the sequence and structures of the CDRL1 and FRL3 loops of somatically mutated residues (bold letters in red). IGLV3-19 sequence based on allele 01 (UNIPROT KB accession number: P01714). (b) Models depicting the position of LC.His31/LC.Ala66 in the SPM-binding site of the “Fab-Fus4/MPER” structure derived from the cluster analysis. Side chains of displayed residues established contacts with the phosphocholine moiety at a distance <3.5 Å. (c) Interactions with VL-LB lipids (distance <3.5 Å) of LC.Arg24 and LC.Arg70 in the same structure.