Structural insights into VRC01-class bnAb precursors with diverse light chains elicited in the IAVI G001 human vaccine trial

Abstract

The development of germline-targeting vaccines represents a potentially transformative strategy to elicit broadly neutralizing antibodies (bnAbs) against HIV and other antigenically diverse pathogens. Here, we report on structural characterization of vaccine-elicited VRC01-class bnAb precursors in the IAVI G001 Phase 1 clinical trial with the eOD-GT8 60mer nanoparticle as immunogen. High-resolution X-ray structures of eOD-GT8 monomer complexed with Fabs of five VRC01-class bnAb precursors with >90% germline identity revealed a conserved mode of binding to the HIV CD4-binding site via IGHV1-2-encoded heavy chains, mirroring mature bnAb interactions. The light-chain V-gene diversity emulated VRC01 bnAbs and stabilized antigen engagement, while their conserved five-residue LCDR3 motifs prevented steric clashes. Notably, the VRC01-class bnAb precursors accommodated the N276 glycan, a key barrier in HIV Env recognition, through structural rearrangements in HCDR3 or LCDR1, despite its absence in the immunogen. Surface plasmon resonance analysis showed that 87% of elicited antibodies retained glycan binding capacity, albeit with reduced affinity. These findings validate the ability of eOD-GT8 60mer nanoparticles to prime VRC01-class bnAb precursors with native-like paratopes but with intrinsic glycan adaptability. Structural mimicry of mature bnAbs was observed even with limited somatic hypermutation, indicating that critical features are encoded in the germline repertoire. The structures highlight how germline-encoded features drive bnAb-like recognition at early stages. This work provides molecular evidence supporting germline targeting in humans and provides guidance for designing booster immunogens to shepherd affinity maturation toward broad neutralization.

Keywords: HIV CD4-binding site; N276 glycan accommodation; VRC01-class antibodies; X-ray crystallography; germline-targeting vaccine.

Fig. 1.

Structural characterization of G001 trial-derived antibody precursors bound to eOD-GT8. (A) X-ray crystal structures of five vaccine-elicited antibody precursors (heavy chains in orange, light chains in yellow) in complex with eOD-GT8 (gray). (B) Structures of bnAb Fab/HIV gp120-core complexes encoded by four different light-chain V genes matching those found in the G001 trial-derived antibody precursors. The PDB IDs from left to right are 3NGB, 5CD5, 4J6R, and 5TE6. (C) Germline sequence amino acid identity (%) of heavy and kappa chain variable regions for precursors (Top) and mature bnAbs (Bottom). (D, Left) Residue-specific contact frequency (heatmap) on eOD-GT8 for antibody precursors, calculated from Fab paratope interactions (color scale: 0 to 5 contacts/residue). (D, Right) Contact residues of bnAb VRC01 (PDB: 3NGB) mapped onto HIV gp120 core, highlighting conserved CD4bs engagement. Residue positions contacted by all five precursors in (D, Left) are highlighted with a mesh. Contact residues were calculated using PDBePISA.

Fig. 2.

Conserved heavy chain–mediated recognition of the CD4-binding site. (A) Structural alignment of bnAb VRC01 (PDB: 3NGB) and precursor heavy chains (colored: VRC01 [lightblue surface], G001-0087 [red], G001-58 [yellow], G001-59 [blue], G001-179 [pink], G001-14 [green]), performed by superimposing the eOD-GT8 antigen (gray) onto HIV gp120 core (lightcyan surface), highlighting conserved IGHV1-2 interactions. Interacting residues on eOD-GT8 are shown in dark gray to emphasize regions of contact. (B) Zoomed-in view of the IGHV1-2 interaction of G001-0087 with bnAb VRC01 as a comparison. Residues involved in hydrogen bonds are shown in sticks, and conserved hydrogen bonds in both antibody precursor and bnAb VRC01 are indicated by black dashed lines. All other hydrogen bonds are marked by yellow dashed lines. (C) Antibody (Top) and antigen (Bottom) contact residues involved in the interaction are compared in sequence, with hydrogen bonds indicated by stars and surface buried area in a red scale. Interfaces were calculated using PDBePISA.

Fig. 3.

Light-chain diversity and structural constraints in antibody precursors. (A) Frequency of light-chain V gene usage among G001-derived VRC01-class bnAb precursors (n = 230). Asterisks highlight the light chains used by precursors in structural analyses. (B) Structural alignment of bnAb VRC01 (PDB: 3NGB) and precursor light chains (colored as in Fig. 2A), performed by superimposing eOD-GT8 (gray) onto HIV gp120 core (lightcyan surface), showing divergent LCDR1 conformations driven by V gene usage. Interacting residues on eOD-GT8 are shown in dark gray to emphasize regions of contact. (C) Zoomed-in view of the light-chain interaction of G001-0087, with bnAb VRC01 for comparison. (Top: hydrogen bonds between the precursor light chain and antigen; Bottom: steric constraints on LCDR3 imposed by antigen-proximal regions.) The five-amino acid LCDR3 is shown in sticks to illustrate how it avoids steric clashes with adjacent regions including LCDR1, HCDR3, and the D and V5 loops of the antigen, which are shown in surface representation. (D) Antigen (Left) and antibody (Right) contact residues involved in interaction are compared in sequence, with hydrogen bonds in stars and surface buried area in a red scale. Interfaces were calculated using PDBePISA. (E) LCDR3 sequence alignment of precursors and VRC01, highlighting conserved motifs (stars).

Fig. 4.

Differential accommodation of the N276 glycan in antibody precursors. (A) X-ray crystal structure comparison of how VRC01-class bnAb precursors G001-59 (Left) and G001-14 (Middle) engage the N276 glycan in complex with eOD-GT8-N276 (gray), compared to mature bnAb VRC01 (Right) bound to a high-mannose glycan variant of eOD (PDB: 5KZC). Heavy chains are colored in orange, and light chains are colored in yellow. In the cartoon view (Top row), the N276 glycan is positioned between the LCDR1 and HCDR3 loops in precursors, whereas VRC01 accommodates the N276 glycan near LCDR1. The surface view (Bottom row) highlights differences in specific availability for possible N276 glycan position. (B) Binding affinities (K_D) of antibody precursors (n = 230) for eOD-GT8-N276 (x-axis) and eOD-GT8 (y-axis, previously published in ref. 19). Most data points fall below the line of no K_D change (dashed line), indicating that binding affinity is generally reduced in the presence of the N276 glycan. The geometric mean difference among binders to both antigens corresponds to an approximately 58-fold reduction in affinity in the presence of the N276 glycan. (C) The frequency of heavy chain D gene usage among trial-derived precursors (n = 230) reflects diverse HCDR3 sequences. (D) Distribution of HCDR3 and LCDR1 lengths among trial-derived precursors (n = 230), demonstrating substantial variability.