Structural Basis for a Neutralizing Antibody Response Elicited by a Recombinant Hantaan Virus Gn Immunogen

Abstract

Hantaviruses are a group of emerging pathogens capable of causing severe disease upon zoonotic transmission to humans. The mature hantavirus surface presents higher-order tetrameric assemblies of two glycoproteins, Gn and Gc, which are responsible for negotiating host cell entry and constitute key therapeutic targets. Here, we demonstrate that recombinantly derived Gn from Hantaan virus (HTNV) elicits a neutralizing antibody response (serum dilution that inhibits 50% infection [ID₅₀], 1:200 to 1:850) in an animal model. Using antigen-specific B cell sorting, we isolated monoclonal antibodies (mAbs) exhibiting neutralizing and non-neutralizing activity, termed mAb HTN-Gn1 and mAb nnHTN-Gn2, respectively. Crystallographic analysis reveals that these mAbs target spatially distinct epitopes at disparate sites of the N-terminal region of the HTNV Gn ectodomain. Epitope mapping onto a model of the higher order (Gn-Gc)₄ spike supports the immune accessibility of the mAb HTN-Gn1 epitope, a hypothesis confirmed by electron cryo-tomography of the antibody with virus-like particles. These data define natively exposed regions of the hantaviral Gn that can be targeted in immunogen design. IMPORTANCE The spillover of pathogenic hantaviruses from rodent reservoirs into the human population poses a continued threat to human health. Here, we show that a recombinant form of the Hantaan virus (HTNV) surface-displayed glycoprotein, Gn, elicits a neutralizing antibody response in rabbits. We isolated a neutralizing (HTN-Gn1) and a non-neutralizing (nnHTN-Gn2) monoclonal antibody and provide the first molecular-level insights into how the Gn glycoprotein may be targeted by the antibody-mediated immune response. These findings may guide rational vaccine design approaches focused on targeting the hantavirus glycoprotein envelope.

Keywords: glycoprotein; hantavirus; neutralizing antibody; structure; zoonosis.

FIG 1

HTNV Gn immunization strategy. (A) (Upper) Schematic diagram illustrating the Gn and Gc glycoproteins encoded in the HTNV M segment. The construct of HTNV Gn (residues 18 to 371) used for immunization is highlighted and colored lilac (produced with DOG 4.0 [94]). Predicted N-linked glycosylation sites (NXT/S, where X≠P) are annotated with sticks. (Lower) Schematic diagram of the (Gn-Gc)₄ lattice (based upon EMD-4867), as revealed by previous cryo-ET and X-ray crystallography studies (35). Although the Gc may likely impinge, the N-terminal region of the hantaviral Gn is predicted to make up the majority of the membrane-distal region (lilac) of the (Gn-Gc)₄ lattice. (B) Timeline of rabbit immunization experiments. Rabbits were immunized with recombinant HTNV Gn and boosted at 4-week intervals. Seven days following the third immunization, HTNV Gn binding and neutralization titers were measured. mAbs were isolated through antigen-specific single B cell sorting of PBMCs (Fig. S2).

FIG 2

Immunization with HTNV Gn elicits a nAb response enabling isolation of neutralizing mAb HTN-Gn1 and non-neutralizing mAb nnHTN-Gn2. (A) Analysis of the IgG-specific response to HTNV Gn by ELISA in rabbit sera (rabbits 3946 to 3949) following the second (blue, indicated by _2) and third (black, indicated by _3) HTNV Gn immunizations. A prebleed serum control is shown in green. (B) Neutralization of live HTNV strain 76-118 by rabbit sera (rabbits 3946 to 3949) following the third HTNV Gn immunization. A prebleed serum control is shown in green. (C) Characterization of mAbs HTN-Gn1 and nnHTN-Gn2 binding to HTNV Gn by ELISA. (D) Neutralization of live HTNV strain 76-118 by mAbs HTN-Gn1 and nnHTN-Gn2. Despite mAb nnHTN-Gn2 exhibiting high binding to HTNV Gn, this mAb did not show neutralizing activity. Error bars represent the standard errors of the mean. In panels A and C, ELISA was carried out three times in duplicate. In panels B and D, the neutralization assay was carried out twice in duplicate. Representative graphs are shown.

FIG 3

Crystal structures of Fab fragments from neutralizing mAb HTN-Gn1 and non-neutralizing mAb nnHTN-Gn2 in complex with HTNV Gn reveal that the antibodies target disparate epitopes on the HTNV Gn surface. (Upper left) Structure of the HTNV Gn−HTN-Gn1 complex. The heavy and light chains of Fab HTN-Gn1 are colored dark and light gray, respectively. The CDR loops of the Fab are colored shades of pink (heavy chain) and green (light chain), respectively, as defined in the upper right legend. HTNV Gn is colored according to domain, with domain A in light blue, domain B in dark purple, and the β-ribbon/E3-like domain in purple, as defined in the upper right legend. (Upper right) Structure of the HTNV Gn−nnHTN-Gn2 complex. Colored as described for panel A. The Fab nnHTN-Gn2 binds to domain A of HTNV Gn in an interaction that relies on the CDRs of both heavy and light chains and occludes 900 Ų of surface area. (Lower) Zoom-in of the HTNV Gn−HTN-Gn1 interface. The ordered glycan extending from Asn134 (green sticks) of HTNV Gn from the Fab HTN-Gn1 complex was likely protected from endoglycosidase F₁ during sample preparation and is stabilized by neighboring crystal contacts. Detailed representations of the interactions at the antibody-antigen interfaces are provided in Fig. S3 and S4 in the supplemental material.

FIG 4

Cryo-ET of HTNV VLPs in complex with Fab HTN-Gn1 provides a model for mAb-mediated obstruction of the (Gn-Gc)₄ lattice. (A) Side (left) and top (right) views of the HTNV VLP−Fab HTN-Gn1 reconstruction with the crystal structure of HTNV Gn−Fab HTN-Gn1 (cartoon representation and colored as described in the legend to Fig. 3) fit into the density as a single rigid body. The HTNV VLP is shown as a surface with density corresponding to Fab HTN-Gn1 colored white, the N-terminal ectodomain of HTNV Gn colored purple, the viral membrane colored light blue, and the expected ectodomain regions of the HTNV Gc colored gray. (B) Model of Fab HTN-Gn1 binding in the context of a HTNV VLP, prepared by mapping (Gn-Gc)₄ spike complexes onto the refined coordinates of a single VLP in the data set. For each position, one of the two possible overlapping binding sites was chosen randomly. Colored as described for panel A.

FIG 5

Epitope accessibility provides a rationale for neutralizing activity. (A) Mapping antibody-accessible surfaces onto the N-terminal region of HTNV Gn. The tetrameric assembly of HTNV Gn, formed upon fitting of the HTNV Gn−Fab HTN-Gn1 crystal structure into the HTNV-VLP reconstruction (Fig. 4), is shown in surface representation. Solvent-accessible surfaces of the HTNV Gn ectodomain are colored purple, occluded surfaces (with subunit contacts located within ≤10 Å) are colored dark gray, and mAb HTN-Gn1 and nnHTN-Gn2 epitopes are colored cyan and orange, respectively. Consistent with the non-neutralizing activity of mAb nn-HTN-Gn2, the nnHTN-Gn2 epitope is located at regions of the molecule expected to form intersubunit (i.e., Gn-Gn and Gn-Gc) interactions and is less immunogenically accessible than the neutralizing mAb HTN-Gn1 epitope in the context of the observed higher-order (Gn-Gc)₄ lattice. (B) Mapping of neutralization evasion (NE) mutation sites that indicate key residues for neutralizing antibody activity, reported in the Gn ectodomain across hantaviral species, reveals that NE sites cluster at two regions on the Gn (highlighted in subunit [i.] by dashed circles). The epitope of mAb HTN-Gn1 colocalizes with one of these sites. (C) A close-up of a single Gn ectodomain subunit (i). The region proximal to mAb HTN-Gn1 epitope is critical to the activity of HTNV-neutralizing antibodies mAb 3D5 (NE mutations H304Y and K76E), 16E6 (P303T), and 16D2 (K76E) and ANDV-neutralizing mAb KL-AN-4E1 (N108K) (25, 27, 32). Details of the mapped NE sites are presented in Table S3.