Decision letter: Antigenic mapping and functional characterization of human New World hantavirus neutralizing antibodies

Abstract

Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Hantaviruses are high-priority emerging pathogens carried by rodents and transmitted to humans by aerosolized excreta or, in rare cases, person-to-person contact. While infections in humans are relatively rare, mortality rates range from 1 to 40% depending on the hantavirus species. There are currently no FDA-approved vaccines or therapeutics for hantaviruses, and the only treatment for infection is supportive care for respiratory or kidney failure. Additionally, the human humoral immune response to hantavirus infection is incompletely understood, especially the location of major antigenic sites on the viral glycoproteins and conserved neutralizing epitopes. Here, we report antigenic mapping and functional characterization for four neutralizing hantavirus antibodies. The broadly neutralizing antibody SNV-53 targets an interface between Gn/Gc, neutralizes through fusion inhibition and cross-protects against the Old World hantavirus species Hantaan virus when administered pre- or post-exposure. Another broad antibody, SNV-24, also neutralizes through fusion inhibition but targets domain I of Gc and demonstrates weak neutralizing activity to authentic hantaviruses. ANDV-specific, neutralizing antibodies (ANDV-5 and ANDV-34) neutralize through attachment blocking and protect against hantavirus cardiopulmonary syndrome (HCPS) in animals but target two different antigenic faces on the head domain of Gn. Determining the antigenic sites for neutralizing antibodies will contribute to further therapeutic development for hantavirus-related diseases and inform the design of new broadly protective hantavirus vaccines. Editor's evaluation Antibodies perform a critical function in host defense against viruses and have emerged as major therapeutic tools in modern medicine, as evidenced by the large-scale use of antibody-based therapies during the COVID-19 pandemic. This paper describes the characterization of human antibodies to hantaviruses that have the potential to create devastating epidemics. The results teach us about the viral structures that are targets for neutralization and the results are relevant for vaccine development and antibody therapeutic design. The evidence provided is convincing and the results are important and should be of interest to immunologists, virologists, and those working on antibody engineering and therapeutic antibodies. https://doi.org/10.7554/eLife.81743.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Hantaviruses are emerging zoonotic pathogens that are endemic worldwide and classified into two categories based on the geographic distribution of their reservoir hosts and pathogenesis in humans (Jonsson et al., 2010). Almost 60 virus species have been identified in rodents or shrews, and over 20 species cause disease in humans (Laenen et al., 2019). Old World hantaviruses, including Hantaan virus (HTNV), Puumala virus (PUUV), Dobrava -Belgrade virus (DOBV), and Seoul virus (SEOV), mainly occur in Eastern Europe and China and cause hemorrhagic fever with renal syndrome (HFRS). New World hantaviruses (NWHs), including Sin Nombre virus (SNV) and Andes virus (ANDV), are endemic in North and South America and cause hantavirus cardiopulmonary syndrome (HCPS). Person-to-person transmission of ANDV has been reported, including in a recent outbreak in Argentina resulting in 34 confirmed cases and 11 fatalities (Martinez et al., 2005; Martínez et al., 2020). There are no current FDA-approved medical countermeasures to prevent or treat hantavirus-related disease. The viral glycoproteins, designated Gn and Gc, form a hetero-tetrameric spike on the surface of the hantavirus virion and facilitate attachment and entry. Previous studies have reported crystal structures for the Gn (Li et al., 2016; Rissanen et al., 2017) and Gc (Guardado-Calvo et al., 2016; Willensky et al., 2016) ectodomains of PUUV and HTNV, and recent work has described the molecular organization of Gn/Gc on the virion surface (Serris et al., 2020). Gc is a class II fusion protein and undergoes conformational changes triggered by low pH to mediate the fusion of the viral and host endosomal membranes. Gn is proposed to play a role in receptor attachment and stabilize and prevent the premature fusogenic triggering of Gc (Mittler et al., 2019; Bignon et al., 2019). Structural studies have identified a capping loop on Gn that shields the fusion loop on Gc, and the glycoprotein complex is thought to undergo dynamic rearrangements between a closed (or capped) and open (or uncapped) form of the spike (Serris et al., 2020; Bignon et al., 2019). Recent efforts have identified features of the molecular basis of neutralization by some antibodies targeting HTNV Gn (Rissanen et al., 2021) or PUUV Gc (Rissanen et al., 2020). A rabbit-derived antibody, HTN-Gn1, targets domain A on Gn and overlaps with the putative binding sites for other murine-derived HTNV (Arikawa et al., 1992) and ANDV mAbs (Duehr et al., 2020). An antibody isolated from a bank vole, P-4G2, targets a site spanning domain I and II on Gc that is occluded in the post-fusion trimeric form, suggesting that the antibody may neutralize through blocking conformational changes required for fusion (Willensky et al., 2016; Rissanen et al., 2020). Although mAb P-4G2 neutralizes PUUV and ANDV, this activity has only been tested in pseudovirus neutralization assays, and it is unknown if the antigenic site on Gc is occluded on the surface of the authentic virus. Also, these antibodies were not derived from human B cells and were not induced by natural infection, but rather were elicited following inoculation of animals immunization with recombinant protein, recombinant VSV constructs, or virus-infected tissues. It is unclear what sites are accessible to antibodies during a natural infection, and if human antibodies target antigenic sites that differ from those recognized by rodents. The first clues toward the sites of vulnerability on the Gn/Gc spike for the human antibody response were recently described by Mittler et al., 2022. A panel of 135 antibodies were isolated from convalescent PUUV donors, and two distinct neutralizing sites were determined by negative stain electron microscopy (nsEM); one at the Gn/Gc interface and one on prefusion exposed surface of Gc domain I. ADI-42898, a quaternary-site mAb, demonstrated cross-clade neutralizing activity and protected in both PUUV bank vole and ANDV hamster post-exposure challenge models. However, mAbs targeting Gn were not described, and, due to its level of surface exposure, the N-terminal domain of Gn likely represents a major site of the neutralizing human antibody response (Serris et al., 2020; Engdahl and Crowe, 2020). Gn exhibits a higher degree of sequence variability than Gc, likely indicating that Gn is under more immune pressure than Gc (Li et al., 2016). Previously, we characterized a panel of human mAbs isolated against Gn/Gc from survivors of SNV or ANDV infection (Engdahl et al., 2021). We demonstrated that NWHs antibodies target at least eight distinct sites on the ANDV Gn/Gc complex, four of which contained potently neutralizing antibody clones, but the location of those sites on the Gn/Gc spike was not known. Here, we define four distinct neutralizing antigenic sites on the Gn/Gc complex. Two potently neutralizing species-specific antibodies, ANDV-5 and ANDV-34, map to two non-overlapping epitopes in the Gn ectodomain and neutralize the virus by blocking attachment. We also show that these antibodies are similar to two human antibody clones previously isolated, MIB22 and JL16 (Garrido et al., 2018). Broadly neutralizing antibodies, ANDV-44 and SNV-53, map to the interface of Gn/Gc, while a less potently neutralizing but broadly reactive antibody, SNV-24, targets domain I of Gc. Both classes of broad antibodies function by inhibiting the triggering of fusion. These data shed light on the basis for both ANDV-specific and broad hantavirus recognition by these antibody clones and suggest that hantavirus vaccine designs should focus on effectively eliciting antibodies to these sites of vulnerability. Results Neutralizing antibodies target at least four sites on ANDV GnH/Gc spike The hantavirus glycoprotein spike forms a complex hetero-tetrameric structure on the surface of the virus and is composed of two proteins, Gn and Gc (Guardado-Calvo and Rey, 2021a). Gn is separated into two main domains: an N-terminal, membrane-distal Gn head domain (GnH), and a C-terminal Gn base domain (GnB) (Li et al., 2016; Serris et al., 2020). GnH likely functions in viral attachment, but also has a ‘capping loop’ region that interfaces with the Gc fusion loop to prevent premature exposure of the hydrophobic loop (Bignon et al., 2019). Due to the surface exposure and sequence variability of GnH, it is likely immunodominant compared to Gc and GnB17. GnB promotes the tetramerization of the complex and is shielded from immune recognition in the Gn/Gc complex. We previously isolated a panel of 36 mAbs from human survivors who were naturally infected with SNV or ANDV, and we demonstrated that the mAbs target eight sites on Gn/Gc, several of which are sites of vulnerability for neutralization. We selected five representative mAbs from this panel for further study to determine the features of the neutralizing antigenic sites. We also produced IgG1 forms of two previously reported human mAbs based on publicly available cDNA sequences, MIB22 or JL16 (Garrido et al., 2018), and cloned them into a human IgG1 expression vector for recombinant expression (hereafter we designate these IgGs as rMIB22 or rJL16). To determine which subunits of the spike were targeted by neutralizing antibodies, we expressed and purified hantavirus antigens in Drosophila S2 cells and tested for antibody binding reactivity to the GnH, GnB, or Gc monomeric proteins or the GnH/Gc heterodimer (Figure 1a). We also generated and purified a ‘stabilized’ form of GnH/Gc by introducing a H953F mutation that prevents the Gc protein from making conformational changes to the post-fusion form (Serris et al., 2020). All the mAbs tested demonstrated reactivity to the linked ANDV GnH/Gc construct by ELISA, as well as to the ‘pre-fusion’ stabilized form (ANDV GnH/Gc_H953F). Most of the mAbs, apart from ANDV-34 and rMIB22, demonstrated reactivity to Maporal (MAPV) GnH/Gc, a closely related species to ANDV. ANDV nAbs (ANDV-34, ANDV-5, rMIB22, and rJL16) all displayed binding reactivity to GnH, with EC50 values less than 100 ng/mL except for rMIB22 (EC50 value 5.6 µg/mL). SNV-24 was the only antibody that displayed reactivity to Gc, and we did not detect binding reactivity to GnB for any mAbs tested. Notably, bnAbs ANDV-44 and SNV-53 did not have detectable binding to GnH, GnB, or Gc alone, and only bound to the linked GnH/Gc antigen, suggesting these antibodies bind a quaternary site only present on the Gn/Gc heterodimer. Figure 1 Download asset Open asset Hantavirus neutralizing antibodies target four distinct regions on the glycoprotein spike. (a) Binding potency of mAbs to recombinant hantavirus antigens, ANDV GnH/Gc, ANDV GnH/Gc_H953F, MAPV GnH/Gc, ANDV GnH, ANDV GnB, ANDV Gc, expressed in S2 cells. Binding curves were obtained using non-linear fit analysis, with the bottom of curve constrained to 0, using Prism software. The data shown are representative curves from 3 independent experiments. Mean ± SD of technical duplicates from one experiment are shown. (b) Competition binding analysis of neutralizing antibodies to ANDV GnH/Gc recombinant protein measured using BLI. % Competition is designated by the heatmap, where black boxes indicate complete competition, gray boxes indicate intermediate competition, and white boxes indicate no competition. The data are shown are representative from two independent experiments. We also performed competition-binding studies utilizing biolayer interferometry and tested the seven mAbs for binding to the linked GnH/Gc antigen (Figure 1b). As previously shown with a cell surface-displayed version of Gn/Gc, bnAbs ANDV-44 and SNV-53 bin to similar competition groups while SNV-24 bins to a distinct site on Gc. We also determined that rJL16 bins in a group with ANDV-5, while rMIB22 bins with ANDV-34. However, rMIB22 and rJL16 also asymmetrically compete for binding, indicating that all four mAbs likely bind in close spatial proximity to each other on the GnH domain. Low likelihood of viral escape for neutralizing hantavirus antibodies Determining escape mutations is helpful for designing vaccines and therapeutics to treat hantavirus-related diseases. We next sought to identify neutralization-resistant viral variants for each of the seven nAbs described in this study. To do this, we employed two different methods: (1) a high-throughput escape mutant generation assay using real-time cellular analysis based on similar assays previously described (Gilchuk et al., 2020; Greaney et al., 2021; Suryadevara et al., 2022) and (2) serial passaging of virus in increasing concentrations of neutralizing antibodies (Figure 2a). Previously, we demonstrated that our mAbs showed similar neutralization potencies for VSV-pseudotyped viruses and authentic viruses (Engdahl et al., 2021). The one class of mAbs that demonstrated a notable discrepancy was that of Gc-targeting mAbs, such as SNV-24, which showed an ~1000 fold lower neutralization potency for authentic viruses compared to VSV-pseudotyped viruses. Based on the general similarities between the two systems, we considered the VSV surrogate system appropriate to study escape mutation generation of escape mutant glycoprotein sequences. Thus, for all neutralization assays, we employed pseudotyped VSVs bearing the glycoproteins Gn/Gc from either the SNV or ANDV species. For RTCA-based escape mutant generation, we tested each antibody at saturating neutralization conditions and evaluated escape based on delayed cytopathic effect (CPE) in each individual replicate well. Thus, if a replicate well demonstrated a delayed drop in cellular impedance, this finding indicated the presence of an escape mutant virus and was noted out of the total number of replicate wells to give a percentage of escape. Notably, we did not detect escape mutants using this method for most neutralizing antibodies (e.g. SNV-53, ANDV-44, SNV-24, ANDV-34, or ANDV-5; Figure 2a). For example, we could not detect any escape mutants through this method for SNV-53, even though we attempted 368 replicates with the VSV/SNV virus and 184 replicates with the VSV/ANDV virus for selection (Figure 2a). Figure 2 with 2 supplements see all Download asset Open asset Escape mutant generation and mutagenesis mapping indicate critical binding residues for hantavirus mAbs. (a) Results from viral escape selection for indicated antibodies. Real-time cellular analysis escape mutant mapping shows the number of replicates with escape over the total number of replicates for each selection mAb against the indicated selection virus. Mutations from serial passaging were identified for each mAb, and escape was confirmed in the presence of saturating mAb concentrations. VSV/ANDV or VSV/SNV were used for escape selections. (b) Side and top view of escape mutants mapped to the ANDV Gn/Gc spike (PDB: 6ZJM). The colored spheres designate escape mutants for the indicated antibody. Gn is shown in white, and Gc is shown in grey. (c) Heatmap of mAb binding in the presence of ANDV mutant constructs. Dark blue boxes indicate loss of binding. The black boxes designate escape mutants for the indicated antibody. The percent binding (% WT) of each mAb to the mutant constructs was compared to the WT SNV or ANDV control. The data are shown as average values from three to four independent experiments. All numbering for ANDV sequences was based on GenBank AF291703.2 and SNV sequences were based on GenBank KF537002.1. (d) Heatmap of SNV mutant constructs as described in c. All numbering for SNV sequences were based on GenBank KF537002.1. Figure 2—source data 1 Percent mAb binding in the presence of ANDV mutants. https://cdn.elifesciences.org/articles/81743/elife-81743-fig2-data1-v2.docx Download elife-81743-fig2-data1-v2.docx Figure 2—source data 2 Percent mAb binding the presence of SNV mutants. https://cdn.elifesciences.org/articles/81743/elife-81743-fig2-data2-v2.docx Download elife-81743-fig2-data2-v2.docx In contrast, a CPE profile for resistance was noted for 92% or 100% of the replicates for rMIB22 or rJL16, respectively (Figure 2a). We confirmed resistance to either rMIB22 or rJL16 at a 10 µg/mL concentration of the selecting mAb and sequenced the M-segment gene of the escaped virus from six replicates. All sequenced rMIB22-selected viruses bore the mutation K76T. Due to the high number of replicates demonstrating a neutralization escape phenotype and the fact that all the escape viruses isolated had the K76T mutation, it is likely that this escape mutation was present at a high proportion in the viral preparation’s original stock. Five rJL16-selected virus sequences had an L224R mutation, and one had an L224P mutation. Overall, these results indicate that identifying viral escape from hantavirus neutralizing antibodies using this method is possible but suggests a low likelihood of in vitro escape for most of the potently neutralizing hantavirus nAbs targeting multiple distinct antigenic regions we selected for study. Mapping mutations in antibody escape variant viruses selected with ANDV-specific nAbs or bnAbs Although we could not select escape mutants in a high-throughput, single passage approach for five of the mAbs, we still wanted to map the critical binding residues and identify escape mutations for the neutralizing antibodies of interest. Thus, we identified escape mutants for mAbs through serial passaging in cell culture monolayers in the presence of antibody and confirmed neutralization-resistant phenotypes for each escape mutant virus (Figure 2—figure supplement 1). Potent species-specific neutralizing antibodies selected for mutations located in GnH. ANDV-34-selected viruses contained mutations S309Y and D336Y, two residues on the surface exposed face of Gn domain B (Figure 2b). rMIB22-resistant variant viruses contained a single mutation, K76T, located in domain A of Gn in close spatial proximity to S336, corresponding with binning analysis for these mAbs (Figure 2b). Neutralization-resistant viruses selected by ANDV-5 contained mutations E231G and A270D, mapping to α–3 and α–4 helices of the Gn head domain, respectively. Both residues are located near L224, identified in rJL16-resistant viruses, supporting previous data that ANDV-5 and rJL16 compete for a similar binding site (Figure 1b). For bnAbs SNV-53 and ANDV-44, we identified mutations in the interface region between Gn/Gc (Figure 2a). Variant viruses selected by ANDV-44 or SNV-53 contained mutations in both the Gn ectodomain (K86 on the VSV/SNV background and K356, S97, A96 on the VSV/ANDV background) and Gc domain II near the highly conserved fusion loop (K759, P772 on the VSV/SNV background and Y760 on the VSV/ANDV background). For SNV-24, we selected escape mutants D822E and K833N in the variant VSV/ANDV viruses and a single homologous mutation, K834N, in the escape-resistant VSV/SNV. Residues K86, Y760, D822, and K833 are highly conserved among members of the Orthohantavirus genus (Figure 2—figure supplement 2). Mapping of ANDV-specific nAb and bnAb epitopes by mutagenesis Since most of the neutralization-resistant viruses contained multiple mutations, we next sought to identify which mutant residues impacted antibody binding. We generated a panel of single- or double-point mutants in the SNV or ANDV M-segment genes based on escape mutants that we selected with mAbs and previously published VSV/ANDV escape mutants (Duehr et al., 2020; Garrido et al., 2018). We expressed each mutant M segment gene on the surface of Expi293F cells and used a flow cytometric binding assay to assess how each variant impacted mAb binding. The % wild-type (WT) values were generated by dividing the binding of the mutant by that of the WT construct, and values were normalized based on a positive control oligoclonal mix of mAbs. The expression levels of SNV and ANDV M-segment variants were comparable to that of the WT constructs, except for three single-point mutants: A270D, C1129F, and K759E (Figure 2—figure supplement 1b). All three mutations were identified in infectious VSVs, so it is not clear why we could not detect the expression of the mutated proteins. Notably, these residues all co-occurred with additional mutations, indicating that these residues may be found in functionally constrained sites. Most single-point mutants we identified partially impacted the corresponding mAb binding reactivity (Figure 2c and d). However, the loss-of-binding phenotype was most evident for the double-mutant M-segment genes, suggesting that multiple mutations are required to generate neutralization-resistant variants. We identified a few single mutations that impacted the binding of multiple broadly-neutralizing mAbs. For example, Y760F (a residue near the fusion loop) ablates the binding of ANDV-44/SNV-53 (Gn/Gc interface) and SNV-24 (Gc domain I). Another single point-mutant, D822E (located in Gc domain I), also reduced the binding of all bnAbs. This finding may indicate that this altered residue promotes rearrangements of Gn/Gc that have allosteric effects on interface mAb binding. Additionally, one double mutant, S309Y/D336Y, exhibited a moderate reduction of the binding of the bnAbs and lost binding of ANDV-specific nAbs tested. Both residues are located on Gn domain B and could represent a critical evolutionary strategy to escape species-specific immunity. If amino acid changes in highly immunogenic epitopes on Gn domain B do not impact viral fitness, then genetic changes in these sites may occur over time or in the case of a large outbreak. BnAbs recognize two antigenic sites on Gn/Gc Next, we sought to determine the location of the antigenic sites targeted by bnAbs, SNV-53 and SNV-24, using negative-stain electron microscopy of MAPV GnH/Gc in complex with antigen-binding fragments (Fabs) of antibodies (Figure 3a). To identify and dock the right Fabs to the maps, we used the information from the binding groups (Figure 1) and escape mutations (Figure 2). The 3D reconstructions (EMD-26735) of SNV-53 and SNV-24 in complex with the MAPV GnH/Gc heterodimer showed that the Fabs localize to two distinct regions in proximity to the corresponding viral escape mutations selected (Figure 3b). SNV-24 bound to Gc domain I, an epitope that is only accessible in the pre-fusion state of Gc and is near the interface between inter-spike Gc heterodimers. P-4G2 (a bank vole-derived mAb) (Rissanen et al., 2021) and group II human PUUV mAbs described previously (Mittler et al., 2022) also map to a similar site on Gc. This site has limited accessibility in the full virus lattice structure due to neighboring interspike contacts and may results in the incomplete neutralizing activity to authentic virus we demonstrated by these mAbs previously (Engdahl et al., 2021). Based on escape mutants generated and nsEM reconstructions, SNV-53 may interact with the capping loop region on Gn, which interfaces with the cd and bc loops on domain II of Gc (Figure 3b and c; Serris et al., 2020). This 3D reconstruction supports our previous finding that bnAbs SNV-53 and ANDV-44 target a quaternary epitope only accessible on the Gn/Gc heterodimer and is consistent with the broadly protective site targeted by ADI-42898 (Mittler et al., 2022). Modeling SNV-53 Fab in the context of the (Gn-Gc)4 lattice structure shows possible clashes with adjacent spikes in the full virus assembly (Figure 3—figure supplement 1). However, the low resolution of the model may not fully recapitulate the binding angle and we previously demonstrated that this mAb was capable of complete neutralization of authentic viruses. We also performed hydrogen-deuterium exchange mass spectrometry (HDX-MS) with the Fab forms of SNV-53 or ANDV-44 in complex with ANDV GnH/Gc. In the capping loop region, a reduction in deuterium uptake was seen for peptides of ANDV-44 (peptide spanning amino acids 80–100 and peptide spanning 82–100) while the deuterium uptake for SNV-53 remained unchanged. (Figure 3—figure supplement 2). Figure 3 with 2 supplements see all Download asset Open asset BnAbs SNV-53 and SNV-24 target two sites on the GnH/Gc heterodimer. (a) Representative nsEM 2D-class averages of SNV-53 and SNV-24 Fabs in complex with MAPV GnH/Gc heterodimer. (b) Surface representations (light grey) of SNV-53 (blue) and SNV-24 (green) in complex with MAPV GnH/Gc. Escape mutations are indicated by the colored spheres. GnH is colored in purple, GnB is colored in light grey, and the capping loop is colored in pink. Domain I, II, and III of Gc are colored in red, yellow, and blue, respectively, and the fusion loop is colored in orange. (c) Model of bnAbs in complex with the (Gn-Gc)4 spike as colored in b, top and side view are shown. Structural basis for neutralization by Gn-targeting antibodies We were also interested in the epitopes targeted by ANDV-specific mAbs elicited during natural infection, thus, we performed 3D reconstructions of ANDV-5 and ANDV-34 in complex with the ANDV GnH protomer (EMD-26736), which showed that the mAbs bind to ANDV GnH at two distinct sites in concordance with the results from competition binning (Figure 1b). Fab ANDV-5 bound the Gn head domain α3 – α4 angled parallel to the membrane and may interact with a neighbor Gn protomer in the (Gn-Gc)4 complex (Figure 4a). Fab ANDV-34 bound to a distinct site on the opposite face of Gn corresponding to domain B and is angled perpendicular to the viral membrane in the (Gn-Gc)4 complex (Figure 4a).To understand the molecular level interaction of the ANDV-specific neutralizing antibodies, ANDV-5 and ANDV-34, and GnH, we collected a cryo-EM data set and reconstructed a 3D map at 4.1 Å resolution (Figure 4b and Figure 4—figure supplement 1, Supplementary file 1). The heavy chain solely drives the interactions between ANDV-5 and Gn with four hydrogen bonds and excessive hydrophobic interactions (Figure 4c and e). The CDRH3 loop interacts with a hydrophobic cavity on Gn (Figure 4—figure supplement 2). The escape mutation Ala270 is located in the center of the epitope. We could not detect expression of the Gn/Gc glycoproteins bearing the A270D mutation, indicating that A270D may solely ablate ANDV-5 binding but may incur fitness costs offset by mutations (i.e. E231G) that are not located in the epitope. The ANDV-34 interface with Gn is more extensive than the ANDV-5:Gn interface and includes both heavy and light chain interactions with 11 hydrogen bonds (Figure 4d and f). Both the escape mutations Ser309 and Asp336 are part of the interface, and Ser309 has one hydrogen bond with the epitope. Figure 4 with 2 supplements see all Download asset Open asset Cryo-EM structure of neutralizing antibodies ANDV-5 and ANDV-34 in complex with ANDV GnH. (a) Top view (right) and side view (left) of the heterotetramer Gn/Gc (Serris et al., 2020) (PDB: 6ZJM) with low resolution map and model of Gn(H) ANDV-5 and ANDV-34 (purple, orange and red, respectively) complex superimpose with GnH. (b) Cryo-EM map and model of the Gn-Fabs complex. Right, transparent EM map color by chain with Gn purple, ANDV-5 heavy chain in yellow, ANDV-5 light chain in orange, ANDV-34 heavy chain in red and ANDV-34 light chain in pink. (c) Zoom-in on the paratope/epitope interface of the ANDV-5:GnH. Fab ANDV-5 residues that is in close contact with Gn. Heavy chain residues (yellow stick), label with single letter and residue number. Gn is shown in purple with the contact residues shown in a surface representation. Asterisks correspond to residues found in the escape mutant viruses. Blue dashed line, H-bond. (d) Zoom-in on the paratope/epitope interface of the ANDV-34:GnH. Fab ANDV-34 residues that is in close contact with Gn. Heavy residues (red stick), light chain residues (pink stick), label with single letter and residue number. Gn is shown in purple with the contact residues shown in a surface representation. Asterisks correspond to residues found in the escape mutant viruses. Blue dashed line, H-bond. (e) Fab ANDV-5 paratope and epitope residues involved in hydrogen bonding (dashed lines) and hydrophobic interactions. Hydrophobic interactions residues are shown as curved lines with rays. Atoms shown as circles, with oxygen red, carbon black, and nitrogen blue. Interacting residues that belong to CDR loops are colored in different shade. Image was made with Ligplot+49 (Laskowski and Swindells, 2011). (f) Fab ANDV-34 paratope and epitope residues involved in hydrogen bonding (dashed lines) and hydrophobic interactions. Hydrophobic interactions residues are shown as curved lines with rays. Atoms shown as circles, with oxygen red, carbon black, and nitrogen blue. Interacting residues that belong to CDR loops are colored in different shade. Image was made with Ligplot+49. Neutralization potency is dependent on bivalent interactions We have previously demonstrated that all antibody clones described here neutralize ANDV or SNV as full-length IgG1 molecules (Engdahl et al., 2021). Here, we sought to determine if bivalency was required for NWH neutralization; therefore, we generated the five antibo