A potently neutralizing and protective human antibody targeting antigenic site V on RSV and hMPV fusion glycoprotein

Summary

Human respiratory syncytial virus (RSV) and human metapneumovirus (hMPV) are frequent drivers of morbidity and mortality in susceptible populations. The primary target of neutralizing antibodies is the fusion (F) glycoprotein on the surface of the RSV and hMPV virion. As a result of the structural conservation between RSV and hMPV F, three antigenic regions are known to induce cross-neutralizing responses: sites III, IV, and V. Leveraging LIBRA-seq, we identify five RSV/hMPV cross-reactive human antibodies. One antibody, RM 5-1, potently neutralizes all tested viruses from the major subgroups of RSV and hMPV and provides protection against RSV and hMPV in a mouse challenge model. Structural analysis reveals that RM 5-1 utilizes an uncommon genetic signature to bind an epitope that spans sites Ø, II, and V. These findings highlight the molecular and structural elements influencing RSV and hMPV cross-reactivity as well as the potential of antibody RM 5-1 for translational development.

Graphical abstract

Highlights

• •Characterized 5 RSV and hMPV cross-reactive mAbs identified via LIBRA-seq
• •Monoclonals display potent neutralization across the major genotypes of RSV and hMPV
• •RM 5-1 targets sites Ø, II, and V
• •RM 5-1 protects against RSV and hMPV in vivo

Abu-Shmais et al. isolate and characterize respiratory syncytial virus (RSV) and human metapneumovirus (hMPV) cross-neutralizing antibodies. Structural analysis reveals one antibody, RM 5-1, binds an epitope that spans sites Ø, II, and V on the fusion glycoprotein. RM 5-1 has potential for translational development.

Introduction

Human respiratory syncytial virus (RSV) and human metapneumovirus (hMPV) are worldwide, endemic respiratory pathogens of the Pneumoviridae family.¹ Representing non-segmented negative-strand RNA viruses, RSV and hMPV induce severe and lethal bronchiolitis and pneumonia among particularly susceptible populations, most notably infantile, geriatric, and immunocompromised,^2,3 with RSV being a leading cause of lower respiratory-tract-infection-associated hospitalization and mortality in children under 5 years of age.^4,5 Similar to RSV, infants younger than 1 year have disproportionately high risks of severe hMPV infection, particularly in low- and middle-income countries, making hMPV an important pathogen in the first year of life.⁶ A turbulent history of disease enhancement following RSV vaccination⁷ has only recently been met with clinical success in the advancement of effective prophylactic strategies leveraging structure-based vaccine design^8,9,10 and neutralizing antibodies with extended half-lives.^11,12 Currently there are no approved therapeutic or prophylactic options against hMPV infection.

The major target of neutralizing antibodies in human sera against RSV and hMPV infection is the fusion (F) glycoprotein on the surface of the virion.^13,14,15,16 RSV/hMPV F is a trimeric type I transmembrane fusion protein responsible for mediating viral entry into host cells of the airway epithelium.¹⁷ Substantial conformational changes occur in F, as it transitions from the metastable prefusion form to the stable postfusion form, and understanding of these structural rearrangements has enabled engineering of prefusion-stabilized F antigens.^{18,19,20,21,22} Stabilization of RSV and hMPV F in the prefusion state induces high neutralizing titers in experimentally inoculated animals, and prefusion-stabilized RSV F serves as the backbone of the recently approved human RSV vaccines. Importantly, differential glycosylation patterns on the apex of RSV and hMPV F result in conformationally specific contributions toward the induction of neutralizing responses: RSV prefusion F epitopes are exceptionally immunogenic and invoke potently neutralizing antibodies,^14,23 whereas pre- and postfusion hMPV F stimulate comparable neutralizing responses.^21,24 Antibody isolation and characterization efforts against RSV and hMPV have enabled extensive definition of the antigenic landscapes of RSV and hMPV F. The antigenic topology of RSV and hMPV F follows a synonymous nomenclature, with the major sites represented as site Ø through site V, as well as the more recently described site VI on RSV F.²⁵ Antigenic sites Ø, V, and VI are preserved exclusively on the prefusion conformations of the proteins,^23,26,27 whereas sites I, II, III, and IV are exposed on the pre- and postfusion conformations.

Broadly reactive and neutralizing antibodies that recognize both RSV and hMPV have been described with varied breadth and potency of virus neutralization.^{23,28,29,30,31,32,33,34,35} The frequency of such cross-reactive B cells in human repertoires is extremely low and nearly undetectable, suggesting natural infection does not efficiently elicit B cells of this phenotype.³⁶ However, due to the structural homology between RSV and hMPV F glycoproteins, three shared epitopes on F, sites III, IV, and V, can induce cross-reactive antibody responses, despite low overall amino acid sequence identity between RSV and hMPV F monomers (∼35%).³⁷ Residues within site III are relatively conserved between both viruses and a common target of cross-neutralizing antibodies encoded by IGHV3-11/IGHV3-21:IGLV1-40,^29,33,34,35 a germline gene pairing reported to be enriched in infant and adult anti-RSV antibody repertoires recognizing site III.³⁸ Low- and high-resolution structural analyses of site III and IV cross-reactive antibodies provide evidence that binding pose may influence cross-reactivity; however, the mode of antigenic recognition of a site V cross-neutralizing antibody remains unknown.

Leveraging LIBRA-seq (linking B cell receptor sequence to antigen specificity by sequencing), we identified from human peripheral blood mononuclear cell (PBMC) samples five RSV/hMPV cross-reactive antibodies that showed high neutralization potencies against both RSV and hMPV that were comparable to virus-specific (RSV- or hMPV-only) as well as cross-reactive antibodies in the literature, with one monoclonal antibody (mAb) RM 5-1 potently neutralizing the major subgroups of RSV and hMPV. We determined the epitope of RM 5-1 by single-particle cryo-EM using a prefusion-stabilized hMPV F with inter- and intra-protomer disulfide bonds and found that the binding site of RM 5-1 spans antigenic sites Ø, II, and V on an individual protomer. Analysis of the interface identified residues that are important for RSV and hMPV cross-neutralization. Finally, RM 5-1 showed robust protection in a mouse challenge model against both RSV and hMPV, therefore establishing this antibody as a prime candidate for further translational development.

Results

Isolation of RSV/hMPV cross-reactive monoclonal antibodies by LIBRA-seq

To identify RSV/hMPV cross-reactive antibodies, we mined previously reported LIBRA-seq datasets^39,40 that included prefusion-stabilized F glycoproteins from RSV A, RSV B, hMPV A, hMPV B, as well as control antigens. These B cells were bulk sorted from healthy donor PBMC samples, based on the expression of several markers: CD19⁺, IgG⁺, and antigen⁺. After sequencing and computational filtering, we isolated a total of 27 B cells with positive signal (defined as a minimum LIBRA-seq score of 1) for at least one of the F glycoproteins belonging to both RSV and hMPV, while exhibiting low signal (defined as an LIBRA-seq score less than 1) for binding to control antigens.

Epitope mapping and in vitro functional properties

Five B cell receptor sequences from our analysis, corresponding to B cells with high LIBRA-seq scores of at least 1 for both RSV A/B and hMPV A/B, were produced recombinantly as IgG1 monoclonal antibodies (mAbs) (Figure 1A). Four of the five antibodies are encoded by gene segments belonging to the VH3 family, with two of the four using the archetypal IGHV3-11/3-21:IGLV1-40 of site III cross-reactive antibodies such as MPE8, 25P13, RSV199, and MxR.³⁵ In contrast, mAb RM 5-1 is encoded by IGHV1-18:IGKV1-5 (Table 1). LIBRA-seq predicted reactivity, as well as binding to more recently circulating strains of hMPV, was confirmed via enzyme-linked immunosorbent assay (ELISA) (Figures 1B and 1C). To assess autoreactivity, binding to permeabilized HEp-2 cells was performed. At 1 μg/mL and 10 μg/mL, none of the antibodies displayed binding to HEp-2 cells (Figure S1). To investigate the antigenic binding sites of the cross-reactive mAbs, we tested the antibodies for competition ELISA binding against site-specific published antibodies with prefusion-stabilized RSV F and hMPV F protein antigens. Antibodies RM 2-6, RM 9-1, and RM 1-2 displayed consistent competition binding profiles on RSV and hMPV F proteins, mapping to sites III (RM 2-6 and RM 9-1) and IV (RM 1-2). Intriguingly, mAb RM 5-1 strongly competed for binding to multiple sites on RSV prefusion F (sites Ø, II, and III) and hMPV prefusion F (sites II, III, and V). mAb RM 0-20 also strongly competed for binding to multiple sites on RSV prefusion F (sites Ø, II, and V) and hMPV prefusion F (sites II and V) (Figure 2A). Due to the unusual competition profiles of mAbs RM 5-1 and RM 0-20, we conducted epitope binning using competition biolayer interferometry (BLI). Similar to their competition ELISA binding profile, mAb RM 5-1 competed with site Ø, II, III, and V mAbs on RSV and II, III, and V on hMPV, while mAb RM 0-20 competed with sites Ø, II, and V mAbs on RSV and sites II and V mAbs on hMPV (Figure 2B).

Figure 1.

Identification and characterization of RSV/hMPV cross-reactive antibodies

(A) LIBRA-seq predicted RSV- and hMPV-specific B cells. Each dot indicates an individual B cell. Max RSV A/RSV B LIBRA-seq score on the x axis and max hMPV A/hMPV B LIBRA-seq score on the y axis. Dots colored in purple were selected for further characterization.

(B) ELISA binding of recombinantly produced antibodies against RSV and hMPV prefusion F trimers, calculated as absorbance at 450 nm. Data are representative. Experiments were performed in technical and biological duplicate.

(C) ELISA area under the curve (AUC) of recombinantly expressed antibodies against RSV and hMPV F trimers and BG505 Env negative control antigen. See also Figure S1.

Table 1.

Sequence characteristics of RSV/hMPV cross-reactive antibodies

mAb	Native isotype	VH gene	VH % identity	CDRH3	CDRH3 length	VL gene	VL % identity	CDRL3	CDRL3 length
RM 0-20	IGHG1	IGHV3-11	87.5	ARGNNLFDDRGLFDH	15	IGLV3-21	89.47	QVRDTGTFQHV	11
RM 5-1	IGHA1	IGHV1-18	88.89	ARGPCCSSPRPYDI	14	IGKV1-5	94.2	QQCYTYSQT	9
RM 2-6	IGHG2	IGHV3-11	93.4	ARISYTSTGPFYFDS	15	IGLV1-40	97.22	QSYDRSLSGYV	11
RM 9-1	IGHG1	IGHV3-21	89.58	ARDSGQQLDPFDY	13	IGLV1-40	94.79	QSYDKRLFGWV	11
RM 1-2	IGHG3	IGHV3-30	93.4	ARAAYDSLTYFEF	13	IGLV3-21	94.98	QVWDSTSDHWV	11

Native isotype, VH/VL gene, VH/VL % identity, CDRH3/L3, and CDRH3/L3 length are shown for each antibody. Percent identity is calculated at the nucleotide level, and sequences and VDJ/VJ length are displayed at the amino acid level.

Figure 2.

Binding characteristics of RSV/hMPV cross-reactive mAbs

(A) Antibody-antibody competition binding to RSV and hMPV prefusion F trimer against control site-specific antibodies. Percentage of binding of biotinylated antibody is shown as a heatmap from 0% (black) to 100% (white). Non-biotinylated competitor antibodies were coated first, and then biotinylated control mAbs were added to detect competition. Competition is calculated as the signal obtained for binding of the biotin-labeled reference antibody in the presence of the unlabeled antibody, expressed as a percentage of the binding of the reference antibody alone. Experiments were performed in technical and biological duplicate.

(B) Epitope binning via BLI for binding of mAbs RM 0-20 and RM 5-1 to RSV and hMPV prefusion F trimer. Data indicate the percent binding of the second antibody in the presence of the first antibody, as compared to the second antibody alone. Percentage of binding is shown as a heatmap from 0% (black) to 100% (white). Data are representative. Experiments were performed in technical and biological duplicate.

(C) ELISA area under the curve (AUC) of mutated (M) (sequence from sorted B cells) versus germline reverted (G) recombinantly expressed antibodies against RSV A and B and hMPV A and B prefusion F trimers. Data are representative. Experiments were performed in technical and biological duplicate. See also Figure S5.

To investigate whether cross-reactivity emerged as a result of somatic hypermutation, we reverted each candidate mAb to its germline sequence and tested binding to recombinant F antigens. While mAbs RM 9-1 and RM 2-6 both target site III, germline-reverted mAb RM 2-6 preferred binding to RSV F while germline-reverted mAb RM 9-1 preferred binding to hMPV F (Figure 2C). Binding to both RSV and hMPV F was abrogated for the germline-reverted mAb RM 0-20, while mAb RM 5-1 and mAb RM 1-2 displayed preferential binding to RSV F and hMPV F, respectively (Figure 2C).

Unlike other RNA viruses subject to high mutability within the immunodominant glycoprotein, such as with influenza hemagglutinin or SARS-CoV-2 spike, RSV and hMPV F are not documented to experience significant antigenic drift over time.^41,42 Homology analysis of publicly available full-length F glycoprotein sequences deposited between 01/01/2024 and 09/01/2025 exhibit a high degree of similarity (average 99.4%–99.7%) with representative viral strains of the major antigenic groups of RSV and hMPV (Table 2). Using these strains, as well as two more recently circulating strains of RSV, antibody-virus neutralization potency was determined by plaque reduction neutralization test (PRNT) using live virus to inoculate cells. All candidate mAbs exhibited neutralization against at least one of the tested viruses. As previously reported, Nirsevimab, the half-life extended and affinity matured version of RSV-specific mAb D25 that is now available in the clinic, potently neutralized all RSV viruses; MPE8, a well-characterized RSV and hMPV cross-reactive antibody, neutralized both RSV and hMPV with varying degrees of potency. Of the antibodies isolated here, and in comparison with MPE8, mAb RM 5-1 demonstrated the highest neutralization potencies against RSV viruses, including 2022 isolates, as well as hMPV TN 93-32 (B2). mAb RM 2-6 was the most potent against hMPV CAN/97-83 (A2) (Figures 3A and 3B). Of the antibodies able to bind both RSV and hMPV, mAb RM 5-1 generally neutralized the greatest proportion of viruses with high potency (Figure 3C).

Table 2.

Sequence homology of F glycoproteins of RSV and hMPV representative viral strains and 2024–2025 isolates

	2024-2025 sequences	Average pairwise identity	Lowest pairwise identity
RSV A	1931	99.40%	96.27%
RSV B	2221	99.70%	96.52%
hMPV A	83	99.70%	98.70%
hMPV B	28	99.40%	97.21%

Sequences acquired from Pathoplexus and filtered on collection dates between 2024-01-01 and 2025-09-08. Incomplete sequences (any sequence with two or more consecutive “X” characters) were removed, and sequences were aligned with Geneious software using clustal omega. Accession numbers of F glycoproteins for representative strains are as follows: RSV A2 Genbank: ALS35589.1, RSV B1 Genbank: AAB82436.1, hMPV CAN 97-83 (A2) Genbank: AAQ67695.1, and hMPV TN 93-32 (B2) Genbank: AAK62968.2.

Figure 3.

Neutralization potency of RSV/hMPV cross-reactive mAbs

(A) Antibody neutralization against RSV A2, RSV B1, RSV A 2022 (NY-Wadsworth), RSV B 2022 (NY-Wadsworth), hMPV CAN 97-83 (A2), and hMPV TN 93-32 (B2) via PRNT.

(B) IC₅₀ values, expressed as a heatmap with stronger neutralization depicted in darker shades of purple and weak/non-neutralizing (>10 μg/mL) shown in light purple. Calculated by non-linear regression analysis by GraphPad Prism software. Neutralization assays were performed in technical triplicate and biological duplicate; data are represented as mean ± SD.

(C) Fraction of viruses neutralized from (A) with indicated IC₅₀ in μg/mL by RSV/hMPV cross-reactive antibodies.

Structure of mAb 5-1 complexed with hMPV F

As mAb RM 5-1 exhibited the greatest overall neutralization potencies compared to MPE8 and displayed an unusual competition profile that was not resolved by competition biolayer interferometry, we investigated the epitope of mAb RM 5-1 using negative stain electron microscopy (nsEM) and cryo-electron microscopy (cryo-EM). Efforts with a prefusion RSV F protein (DS-Cav1) and RM 5-1 antigen-binding fragment (Fab) were unsuccessful, as most of the trimers were observed in a splayed-open state (Figure S2). Therefore, we used a prefusion-stabilized hMPV F construct (hMPV F-DS-CavEs2-IPDS), which contains intra- and inter-protomer disulfide bonds to lock hMPV F in a closed prefusion trimer conformation.⁴³

Cryo-EM analysis of hMPV F and RM 5-1 Fab revealed a heterogeneous mixture of complexes composed of three Fabs per trimer, with the majority of the particles displaying flexibility at the membrane-proximal base of the F protein (Figure S3). However, a subset of particles retained after 2D classification were identified with a well-ordered base (∼23%), and further processing yielded a 3D reconstruction with a global resolution of 4.3 Å (Figures S4B and S4C). The cryo-EM map agrees very well with a model of the complex produced with AlphaFold3,⁴⁴ and only light refinement was required to obtain an excellent map-to-model fit.

The structure reveals that the RM 5-1 epitope is contained within the F1 subunit of a single protomer and primarily spans antigenic sites II and V, with some additional interactions with site Ø (Figures 4A and 4B). The RM 5-1 heavy and light chains bury 597 Ų and 303 Ų of surface area, respectively, with the complementarity-determining region (CDR) 1 and 2 of the light chain contributing to the interaction with site Ø and the top half of site V. The light chain primarily interacts with residues on α4 through an electrostatic interaction network formed by Asp31_CDRL1 and Arg50_CDRL2 with hMPV F residues Lys171 and Asp167 and with residues on the loop preceding β3 through the electrostatic interaction of Glu55_CDRL2 with Lys143 (Figure 4C). The heavy chain packs its CDR loops against the cleft between β3 and α6, with Tyr53_CDRH2 inserted into the cleft. Interestingly, the RM 5-1 CDRL3 only interacts with residues on the CDRH2 and CDRH3 loops rather than with hMPV F, which may be important for stabilizing the heavy-chain interactions (Figure 4C). In addition, there appear to be interactions between light-chain framework residues and the N-linked glycans attached to Asn172 on hMPV F, despite the low resolution and partially modeled glycan chains (Figures S4A–S4D).

Figure 4.

RM 5-1 Fab binds to prefusion hMPV F at sites II and V and the glycan at Asn172

(A) Front view of the fit of hMPV F complex into a DeepEMhanced EM map at the contour level of 0.432. The global DeepEMhanced EM map is shown as a white transparent map with a single hMPV F protomer and Fab variable domain colored (hMPV F, blue; heavy-chain variable domain, red; light-chain variable domain, orange).

(B) Overlay of the RM 5-1 epitope onto the defined antigenic sites of hMPV F reveals that RM 5-1 primarily interacts with residues in sites II and V, with additional contacts within site Ø.

(C) Atomic model of RM 5-1 and hMPV F interface with key residues highlighted as sticks. RM 5-1 and one hMPV F protomer are shown as cartoons. Oxygen atoms are colored red, and nitrogen atoms are colored blue. Partially modeled Asn-172 glycan is shown as dark-colored sticks.

(D) Sequence conservation of the RM 5-1 epitope between hMPV F and RSV F with the epitope of RM 5-1 delineated in white.

(E) Sequence alignment of the RM 5-1 epitope with two representative hMPV F sequences from A2 (NCBI accession: AAQ67695.1) and B2 (NCBI:accession: AAK62968.2) subgroups and two representative RSV F sequences from A2 (NCBI accession: ALS35589.1) and B1 (NCBI accession: AAB82436.1) subgroups. The conservation of each residue is indicated underneath, and the RM 5-1 interacting residues are highlighted in red. The glycosylation site at Asn-172 is shown as a branch. See also Figures S2, S3, S4, S5, S6 and S7, and Table S1.

The structural model obtained from cryo-EM analysis agrees well with the ELISA and BLI competition binding data. Superposition of the cryo-EM structure with previously determined structures of the antibodies used in the competition assays predicts that RM 5-1 would sterically clash with D25, motavizumab, MPE8, hRSV90, ADI-61026, and MPV467 (Figure S5). Further comparison to known hMPV and RSV F antibody complexes revealed that hRSV90 binds to a similar epitope on RSV F, except with an inverted arrangement of the heavy and light chains (Figure S6). However, hRSV90 is specific for RSV and does not bind or neutralize hMPV. Systematic comparison of the RM 5-1 binding pose and epitope with other site V antibodies, where structural information was available, revealed multiple different modes of site V binding, with notable differences in the contact residues of the respective antibody epitopes, as observed with site III^29,34,35 and IV binders.³⁰ While many of the antibodies discussed here engage site V contact residues that are conserved between hMPV and RSV, the majority of these antibodies retain specificity for RSV or hMPV alone, likely as a result of the structural difference between RSV and hMPV trimers (Figure S6).

The RM 5-1 epitope contains some amino acids that are not well conserved among RSV and hMPV F proteins, yet the antibody binding mode can accommodate these differences (Figures 4D and 4E). The substitutions will likely impact the affinity of RM 5-1 to different extents, but they do not introduce clashes that would prevent antibody binding. The region including the β3 strand is generally well conserved (hMPV F residues 142–150), as is the cleft between β3 and α6, into which Tyr53_CDRH2 inserts. Thus, the structure and AlphaFold3 models of RM 5-1 bound to hMPV F and RSV F provide a structural basis for how RM 5-1 can bind an epitope at the F apex that is thought to be under immune pressure and less conserved than other regions.

In vivo protection against viral infection

Next, we investigated the protective efficacy of mAb RM 5-1 in both an RSV and an hMPV infection model in BALB/c mice. Fourteen-week-old female mice were mock-treated with PBS, an isotype control human mAb VRC01, or different doses of mAb RM 5-1 6 hours prior to intranasal RSV or hMPV challenge (Figures 5A and 5B). Lung viral titers of mice were determined by plaque assay on day 6 post-infection to assess mAb RM 5-1 prophylaxis against infection. At the highest mAb RM 5-1 dose of 10 mg/kg, viral lung titers were below the detection limits for both RSV and hMPV for all animals (Figure 5B). Even at the 10-fold lower dose of 1 mg/kg, 2/5 animals (40%) showed no detectable viral titers in the lung for both RSV and hMPV and were overall significantly lower than those observed in the control groups. Animals receiving the lowest dose of 0.1 mg/kg of mAb RM 5-1 showed significantly reduced lung viral titers for RSV (Figure 5A) and a 3.33-fold (though not statistically significant) reduction for hMPV. Together, these results showcase the in vivo protective ability of mAb RM 5-1 against RSV and hMPV challenge.

Figure 5.

Prophylaxis of RM 5-1 against RSV and hMPV challenge

Protective efficacy of RM 5-1 against (A) RSV A2 and (B) hMPV TN 93-32 (B2) replication in vivo. BALB/c mice were treated intraperitoneally with 10 mg/kg, 1 mg/kg, and 0.1 mg/kg of mAb RM 5-1 6 h prior to intranasal RSV and hMPV infection. Viral titers in the lung homogenates of BALB/c mice in each treatment group (n = 5 mice per group, 5 females) were determined by plaque assay. Statistical significance determined by one-way ANOVA where ∗∗∗∗p < 0.0001, ∗∗∗p = 0.0001, ∗p = 0.02 (A); and ∗∗∗∗p < 0.0001, ∗∗p = 0.005, ∗p = 0.01 (B). n.s., not significant, Limit of detection (LOD) is indicated with a dashed line.

Discussion

Respiratory illness associated with infection by either RSV and/or hMPV remains a public health threat, with the potential for severe disease in neonatal, geriatric, and immunocompromised patients such as those undergoing hematopoietic stem cell transplant and patients suffering from pulmonary co-morbidities. While strategies to prevent severe infection induced by RSV have advanced in the last year, there are currently no approved treatments for infection by hMPV. We and others have isolated rare RSV and hMPV cross-neutralizing antibodies that present an interesting alternative to mono-valent therapies, providing a protective regimen for the prevention or amelioration of disease caused by either mono- or co-infection of RSV and hMPV.

We discovered five antibodies targeting three previously reported epitopes on the F protein known to elicit cross-reactive humoral responses. Consistent with the enrichment of site-III-directed antibodies encoded by IGHV3-11/3-21:IGLV1-40, mAbs RM 9-1 and RM 2-6 display competition profiles indicative of binding at antigenic site III. Interestingly, germline-reverted mAbs RM 9-1 and RM 2-6 favored binding to F from different viruses, despite targeting the same site. Loss of antigenic binding of mAb RM 0-20 to both RSV and hMPV F in the germline state suggests cross-reactivity can be achieved through multiple antibody evolution pathways, i.e., through subsequent activation of either RSV- or hMPV-specific B cells.

All five RSV/hMPV antibodies displayed in vitro neutralizing activity against infection by at least one representative virus of each genotype, albeit some mAbs displayed preferential neutralization against RSV or hMPV alone. mAb RM 5-1 exhibited the greatest potencies against RSV A and B, including two recent isolates (from 2022), with comparable IC₅₀ values to Nirsevimab, approved for prophylaxis in infants against severe disease caused by RSV infection. A considerable strength of mAb RM 5-1 over Nirsevimab is RM 5-1’s expanded breadth to encompass hMPV. While mAb RM 5-1 is not the first reported RSV/hMPV cross-neutralizing antibody, it is one of only two other site-V-targeted mAbs (M1C7³¹ and M2D08⁴⁵) at the time of this publication. While difficult to directly compare reported neutralization due to differences in assay technique, of these site V cross-reactive antibodies, RM 5-1 demonstrates potent neutralization (0.01–0.86 μg/mL) against all of the major subgroups of RSV and hMPV, whereas M1C7 favors hMPV (RSV B IC₅₀ > 1 μg/mL) and M2D08 favors RSV (hMPV A and B IC₅₀ > 2 μg/mL). This neutralization correlated with a reduction in viral lung titers for both RSV and hMPV in a dose de-escalation mouse challenge model, where at the 10 mg/kg and 1 mg/kg doses, RM 5-1 showed significantly reduced lung viral titers for both RSV and hMPV, motivating further exploration into the translational efficacy of mAb RM 5-1. While all antibodies evaluated here were expressed with an IgG1 backbone, it is worth noting RM 5-1 is natively an IgA, which may provide enhanced protection against respiratory pathogens, particularly in dimeric form.

In this study, we included the well-characterized MPE8 as a benchmark for cross-neutralization. With the exception of hMPV CAN 97-83 (A2), RM 5-1 more potently neutralized all strains of RSV and hMPV, likely due to differences in epitope specificity. Studies leveraging recent and contemporaneously circulating strains of RSV and hMPV are lacking in the antibody development space, in part due to the accepted standard of utilizing representative strains of each genotype for measurements of neutralization potential and in part due to difficulties in resource procurement, especially for the lesser studied of these viruses, hMPV. Yet, a significant proportion of hMPV field strains contain amino acid substitution D280N,^46,47 which may impede binding of IGHV3-11/3-21:IGLV1-40 site III cross-reactive antibodies such as MPE8, 25P13, and RSV199. Indeed, our binding data demonstrate MPE8 loses binding to an engineered D280N mutant, while all five antibodies isolated here tolerate this mutation. Additionally, our antibodies retained binding to hMPV B Seattle, a D280N variant from 2015, suggesting these candidates may provide functionality against more recently circulating strains of hMPV.

A limitation of this work is that high-resolution structural information was obtained only for the complex of RM 5-1 with prefusion hMPV F. Detailed analysis of the binding modes of each of the antibodies isolated here would deepen our understanding of the complex interactions governing cross-reactivity between these two viruses. Efforts to structurally characterize RM 5-1 Fab with a prefusion RSV F protein were unsuccessful, as most of the trimers were observed in a splayed-open state (Figure S2); it is possible that binding of RM 5-1 may enhance prefusion trimer dissociation, as reported with another site-V-specific antibody, CR9501.⁴⁸ Structural analysis of RM 5-1 with prefusion hMPV F agrees with the ELISA and BLI competition assay data, with the exception for antibody DS7. The modeling indicates that DS7 is not predicted to clash with RM 5-1; however, some competition was observed (Figure 2B). This may be influenced by the ability of DS7 to bind a conformation of the hMPV F protomer that contains elements of both the prefusion and postfusion conformation.⁴⁹ The apex of hMPV F is shielded by glycans on Asn57 and Asn172 (Figure S7), reducing antigenic exposure and dampening the immune response, relative to that of RSV, against sites V and Ø.^23,24,45,50 However, despite this immune evasion technique, the human immune system has proven its ability to circumvent this obstacle through penetration of the glycan shield, as demonstrated with antibody ADI-61026,⁵⁰ where ADI-61026 positions itself into a pocket between two glycans and directly interacts with Asn57-glycan. Glycan-shield-penetrating antibodies have also been reported that bind to HIV-1 Env^51,52 and hepatitis C E2.⁵³ Here, we demonstrate that RM 5-1 is also able to breach the glycan shield at the apex of hMPV F (Figure S7).

Structural and repertoire analyses, in the context of antibodies elicited as a result of natural infection by RSV and hMPV,^23,27,43,54 have revealed the propensity of site V toward the induction of potently neutralizing humoral responses. Within the trimeric prefusion F protein, the fusion peptide is buried inside a hydrophobic cavity occluded by the site V epitope. As demonstrated with a previously reported antibody targeting site V on hMPV F,⁴³ one potential explanation for the potency of mAb RM 5-1, as compared to the other mAbs in our set, is that binding of mAb RM 5-1 may prevent extension of the fusion peptide from the F protein, thereby disrupting the conformational changes necessary for productive infection. Additional structural characterization may resolve this open-ended question.

Currently, no Food and Drug Administration (FDA)-approved prophylaxis or therapeutics against hMPV F are available, despite substantial efforts.^55,56 Recent progress, including structure-based RSV vaccines and anti-RSV antibody prophylaxis have been made. Nirsevimab, a site-Ø-specific antibody, offers sustained protection against hospitalization due to severe RSV-associated respiratory tract illness in infants,⁵⁷ yet an antibody that potently neutralizes RSV with a different antigenic footprint (site Ø, II, and V) may offer additional benefits when considering the potential for virus evolution. While infrequently documented, site-Ø-drifted variants have been identified from Nirsevimab recipients.⁵⁸ As the therapeutic landscape evolves to include pre-fusion stabilized F subunit vaccines as standard care, placing unprecedented pressure on the virus, the propensity for viral escape presents a growing concern. Furthermore, an antibody that provides cross-protection against both RSV and hMPV infection can be utilized to prevent severe disease from either of these viruses in at-risk populations, providing important logistical advantages over developing multiple virus-specific mAbs, increasing the intrinsic value of a single administration, and mitigating disparities in healthcare access, particularly among lower and middle-income countries where RSV, and to a lesser extent hMPV, cause disproportionate morbidity and mortality.^6,59,60 While these exceptionally valuable antibodies are not commonly induced by natural infection, it is possible to elicit them through vaccination. In a phase 1 clinical trial,³⁶ immunization with pre-fusion stabilized RSV F induced a transient increase in RSV and hMPV cross-reactive memory B cells, suggesting an RSV vaccine may provide enhanced protection against hMPV. However, hMPV neutralizing titers only increased modestly and varied greatly from person to person, making it difficult to interpret the contribution and quality of the RSV vaccine-induced cross-reactive antibody response. As an alternative to vaccination, a broadly cross-reactive and potently neutralizing antibody like mAb RM 5-1 presents an attractive target for further translational development for the prevention of both RSV- and hMPV-associated disease.

Limitations of the study

The present study investigates the genetic features, antigen-binding profiles, and neutralization breadth of RSV and hMPV cross-reactive antibodies. A limitation of this work is that neutralization against more recently circulating strains of hMPV was not shown, which could identify antibodies that are functionally relevant against current hMPV field strains. We note that structural analysis is limited to RM 5-1 and that high-resolution structures of each of the antibodies isolated here could provide additional insights into the mechanisms of RSV and hMPV cross-reactivity. Lastly, further in vivo modeling measuring weight loss and viral load in the upper, as well as lower, respiratory tract is needed to robustly demonstrate the protective efficacy of RM 5-1 on RSV and hMPV infection.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to the lead contact, Ivelin Georgiev (ivelin.georgiev@vanderbilt.edu).

Materials availability

All unique/stable reagents generated in this study are available from the lead contact with a completed Materials Transfer Agreement.

Data and code availability

• •Raw sequencing data used in this study are available on the Sequence Read Archive under BioProject accession numbers PRJNA1049361 and SRR27891916. The EM map and coordinates for the hMPV F and RM 5-1 Fab complex have been deposited into the Electron Microscopy DataBank (EMDB-45412) and the Protein DataBank (9CB1; DOI: https://doi.org/10.2210/pdb9CB1/pdb).
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We thank all members of the Georgiev laboratory for their support and feedback. We thank David Flaherty, Olivia Murfield, Emma McLaughlin, and Brittany Matlock from the VUMC Flow Cytometry Shared Resource for their help with cell sorting. The VUMC Flow Cytometry Shared Resource is supported by the Vanderbilt-Ingram Cancer Center (P30 CA68485) and the Vanderbilt Digestive Disease Research Center (DK058404). We thank Angela Jones, Jamie Roberson, and Latha Raju with the Vanderbilt Technologies for Advanced Genomics Core (VANTAGE) for providing technical assistance with library production and sequencing. VANTAGE is supported in part by CTSA (5UL1 RR024975-03), the Vanderbilt-Ingram Cancer Center (P30 CA68485), the Vanderbilt Vision Center (P30 EY08126), and NIH/NCRR (G20 RR030956). For work described in this manuscript, I.S.G., A.A.A.-S., A.K.J., and M.J.V. were supported in part by the G. Harold and Leila Y. Mathers Charitable Foundation (MF-2107-01851) and NIH R01AI175245 (to I.S.G.). L.G., S.A.R., and J.S.M. were supported in part by Welch Foundation (F-0003-19620604). A.A.A.-S. and R.M.W. were supported in part by NIH grant T32 (5T32AI112541-07). R.M.W. was supported by 1K01OD036063-01. L.E.B. was supported by NIH T32 AR059039 and F31 DK141224. R.H.B. was supported by NIH R01 DK131070. J.J.M was supported by NIH NIAID R01 AI143865 and R01 AI173416. The funders had no role in the conceptualization or execution of any studies or drafting of the manuscript.

Author contributions

Conceptualization and methodology, A.A.A.-S. and I.S.G.; investigation, A.A.A.-S., L.G., A.M.K., S.E.L., R.J.M., A.K.J., M.J.V., L.E.B., S.A.R., Y.P.S., R.M.W., N.K.; writing—original draft, A.A.A.-S. and I.S.G.; writing—review & editing: all authors; funding acquisition, A.A.A.-S., J.S.M., and I.S.G.; resources, J.S.M., J.J.M., R.H.B., R.H.C., J.E.C.Jr., and I.S.G.; supervision, A.A.A.-S., J.S.M., J.J.M., R.H.B., R.H.C., J.E.C.Jr., and I.S.G.

Declaration of interests

A.A.A.-S. and I.S.G. are listed as inventors on patents filed describing the antibodies discovered here. I.S.G. is listed as an inventor on patent applications for the LIBRA-seq technology. I.S.G. is a co-founder of AbSeek Bio. I.S.G. has served as a consultant for Sanofi. The Georgiev laboratory at VUMC has received unrelated funding from Merck and Takeda Pharmaceuticals. J.E.C. has served as a consultant for Luna Labs USA, Merck Sharp & Dohme Corporation, and Emergent Biosolutions; is a former member of the Scientific Advisory Boards of Gigagen (Grifols), Meissa Vaccines, and BTG International; is founder of IDBiologics; and receives royalties from UpToDate. The laboratory of J.E.C. received unrelated sponsored research agreements from AstraZeneca, Takeda Vaccines, and IDBiologics during the conduct of the study.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
APC-Cy7 mouse anti-human CD14	BD	Cat#561709; RRID:AB_10893806
PE-Cy5 mouse anti-human IgG	BD	Cat#551497; RRD: RRID: AB_394220
FITC anti-human CD3 (OKT3)	Tonbo Bioscience	Cat#35–0037; RRD: RRID:AB_2621662
BV711 mouse anti-human CD19	BD	Cat#563036; RRID: AB_2737968
Goat anti-human IgG (Fc specific) peroxidase	Sigma-Aldrich	Cat#A0170; RRID: AB_257868
Biotin Monoclonal Antibody (Z021), HRP	Invitrogen	Cat#03–3720; RRID:AB_2532266
RM 0-20	This paper	N/A
RM 1-2	This paper	N/A
RM 2-6	This paper	N/A
RM 5-1	This paper	N/A
RM 9-1	This paper	N/A
VRC01	NIH AIDS Reagent Program	Cat#12033; RRID: AB_2491019
MPE8	Corti et al.33	GenBank:KF314816, KF314817
D25	Mclellan et al.26	PDB ID: 4JHA
Nirsevimab	Vanderbilt University Medical Center	N/A
ADI-15651	Gilman et al.50	N/A
Motavizumab	Mclellan et al.61	PDB ID: 3IXT
101F	Mclellan et al.62	PDB ID: 3O45
CR9501	Gilman et al.48	PDB ID: 6OE4
SAN 32-2	Rush et al.24	PDB ID: 7TL0
MPV 458	Huang et al.63	N/A
MPV 482	Banerjee et al.43	N/A
MPV 467	Banerjee et al.43	N/A
CR9114	Dreyfus and Laursen et al.64	PDB: 4FQY
MPV 364	Bar-Peled et al.65	N/A
hRSV90	Mousa et al.27	N/A
ADI-61026	Gilman et al.23	N/A
germline RM 0-20	This paper	N/A
germline RM 1-2	This paper	N/A
germline RM 2-6	This paper	N/A
germline RM 5-1	This paper	N/A
germline RM 9-1	This paper	N/A
Bacterial and virus strains
hMPV CAN/97-83	Jarrod Mousa	N/A
hMPV TN/93-32	Jarrod Mousa	N/A
RSV A2	Jarrod Mousa	N/A
RSV B1	Jarrod Mousa	N/A
RSV A 2022 NY-Wadsworth	Zeptometrix	Cat#0810690CF
RSV B 2022 NY-Wadsworth	Zeptometrix	Cat#0810692CF
Chemicals, peptides, and recombinant proteins
hMPV-F A1(NL/1/00)	Jason McLellan	N/A
hMPV-F B2(TN99-419)	Jason McLellan	N/A
hMPV-F v3B-D12_D45C-V458C (D280N)	Ou et al.66	N/A
hMPV-F Seattle/USA/SC3259/2015	GenBank: MF045425.1	N/A
RSV-F A2 (DS-Cav-1)	Jason McLellan	N/A
RSV-F B9320 (DS-Cav-1)	Jason McLellan	N/A
hMPV-F DS-CavEs2-IPDS	Jason McLellan	N/A
HIV-1 gp140 SOSIP BG505	Ivelin Georgiev	N/A
Streptavidin R-phycoerythrin (SA-PE)	Invitrogen	Cat#S866
Protein A resin	GenScript	Cat# L00210
1-Step Ultra TMB-ELISA Substrate Solution	Thermo Fisher	Cat#34029
Pluronic Acid F-68	Fisher	Cat# 24040-032
Bovine serum albumin (BSA)	Sigma-Aldrich	A1470
Ghost dye red 780	Tonbo biosciences	Cat#13-0865
4 mM L-glutamine	Fisher	Cat# 25030-081
Opti-MEMTM I Reduced Serum Medium	ThermoFisher	Cat#31985070
Gibco™ Trypsin-EDTA	ThermoFisher	Cat#25200056
TrueBlue substrate	Seracare	Cat#5510-0030
Gibco™ Antibiotic-Antimycotic	ThermoFisher	Cat#15240062
Gibco™ FreeStyle™ F17 Expression Medium	ThermoFisher	Cat#A1383503
Critical commercial assays
ExpiFectamine™ 293 Transfection Kit	Thermo Fisher	Cat #A14526
EZ-link Sulfo-NHS-Biotin No-Weigh	Thermo Fisher	Cat #A39258
Plasmid Kits for Plasmid DNA Extraction	QIAGEN	Cat#12165
BION ENTERPRISES LTD ANA TEST KIT FLOURESCENT	FisherScientific	Cat# NC0001870
Deposited data
hMPV F and RM 5-1 complex EM map	This paper	EMDB-45412
hMPV F and RM 5-1 complex	This paper	PDB: 9CB1
Antibody Sequences	Abu-Shmais et al.39,40	PRJNA1049361 SRR27891916
Experimental models: Cell lines
Human: Expi293F cells	Thermo Fisher	Cat#A14527
LLC-MK2	ATCC	CCL-7
HEp-2	ATCC	CCL-23
Experimental models: Organisms/strains
BALB/c mice	The Jackson Laboratory	Strain #:000651 RRID:IMSR_JAX:000651
Software and algorithms
cryoSPARC v4.4.0	Punjani et al.67	https://cryosparc.com/
DeepEMhancer	https://github.com/rsanchezgarc/deepEMhancer	N/A
AlphaFold3	AlphaFold Server	https://alphafoldserver.com/welcome
Coot	Emsley et al.68,69	https://bernhardcl.github.io/coot/
PHENIX	Liebschner et al.70	https://www.phenix-online.org/
ISOLDE	Croll71	https://isolde.cimr.cam.ac.uk/
Privateer	Agirre, J., Iglesias-Fernández, J., Rovira, C. et al.72	http://www.ccp4.ac.uk/html/privateer.html
ChimeraX	Goddard et al.73	https://www.rbvi.ucsf.edu/chimerax/
Octet Data Analysis Software v11.1	ForteBio	N/A
GraphPad Prism 9.5.0	Dotmatics	https://www.graphpad.com/
Geneious 11.1.5	Dotmatics	https://www.geneious.com/
HMMER	Finn, R.D et al.74	http://www.ebi.ac.uk/Tools/hmmer/
Other
Superose 6 Increase 10/300 column	Cytiva	Cat#29091596

Experimental model and subject details

RSV and hMPV mouse challenge model

The mouse studies were approved by the Florida State University Animal Use and Care Committee under protocol IPROTO202300000025. BALB/c mice (Female, 14 weeks old; The Jackson Laboratory strain # 000651) were intranasally infected with RSV A2 (2.0E+6 PFU/mouse) or hMPV TN/93-32 (3.0E+5 PFU/mouse) and euthanized 6 days postinfection. Monoclonal antibody RM 5-1 was administered intraperitoneally at 10.0, 1.0, or 0.1  mg/kg. Control mice were intraperitoneally injected with PBS or VRC01 (isotype control) at 10.0 mg/kg. N = 5 for all groups. All injections occurred 6 h prior to infection. Lung homogenates were used for viral titration by plaque assay.

Method details

Cell lines

LLC-MK2 and HEp-2 cells were obtained from ATCC (CCL-7, CCL-23) and grown in growth media (Opti-MEM with 2% FBS) at 37°C, 5% CO₂. Expi293-F mammalian cells (ThermoFisher) were maintained in F17 expression medium (ThermoFisher) in a 37°C shaker supplied with 8% CO₂ and 80% humidity. Cell lines were authenticated by vendors and regularly tested for mycoplasma contamination.

Data mining

LIBRA-seq datasets generated from 2020 to 2023 that included prefusion RSV A-F, RSV B-F, hMPV A-F, and hMPV B-F in the antigen screening library were mined for B cells displaying a minimum LIBRA-seq score of one for at least one of the F antigens, while also displaying a score below one for a control antigen, in this case, recombinant HIV-1 envelope protein. LIBRA-seq experiments were performed on peripheral blood mononuclear cells (PBMCs) samples obtained from otherwise healthy adult individuals. The established LIBRA-seq pipeline was used for score generation.⁷⁵

Antibody expression and purification

For each antibody, variable genes were synthesized as cDNA and were inserted into bi-cistronic plasmids encoding for the constant regions of the heavy chain and either the kappa or lambda light chain (Twist BioScience). Antibodies were transiently expressed with Expifectamine transfection reagent (Thermo Fisher Scientific) in Expi293F cells in FreeStyle F17 expression media (Thermo Fisher) (0.1% Pluronic Acid F-68 and 20% 4 mM L-glutamine). Cells were cultured for 5 days at 8% CO₂ saturation and 37°C with shaking. Five days post transfection, cells were collected and centrifuged at a minimum of 6000 rpm for 20 min. Supernatant was filtered with Nalgene Rapid Flow Disposable Filter Units with PES membrane (0.45 or 0.22 μm) and purified over protein A equilibrated with PBS. Antibodies were eluted with 100 mM glycine HCl at pH 2.7 directly into a 1:10 volume of 1 M Tris-HCl pH 8 and then exchanged into PBS for storage at 4°C.

Enzyme-linked immunosorbent assay (ELISA)

Recombinant antigen was plated at 2 μg/mL overnight at 4°C. The next day, plates were washed three times with PBS supplemented with 0.05% Tween 20 (PBS-T) and coated with 1% bovine serum albumin (BSA) in PBS-T. Plates were incubated for one hour at room temperature and then washed three times with PBS-T. Primary antibodies were diluted in 1% BSA in PBS-T, starting at 10 μg/mL with a serial 1:5 dilution, plated, and then incubated at room temperature for one hour before washing three times in PBS-T. The secondary antibody, goat anti-human IgG conjugated to peroxidase, was added at 1:10,000 dilution in 1% BSA in PBS-T to the plates, which were incubated for one hour at room temperature. Plates were washed three times with PBS-T and then developed by adding TMB substrate to each well. The plates were incubated at room temperature for five minutes, and then 1 N sulfuric acid was added to stop the reaction. Plates were read at 450 nm. ELISAs were performed in technical and biological duplicate.

Competitive binding of mAbs with site-specific antibodies in the literature

Wells of 384-well microtiter plates were coated with 25 μL of 2 μg/mL purified F antigenic protein at 4°C overnight. Plates were blocked with 50 μL of 1% BSA in PBS-T for 1 h before washing three times with PBS-T. Primary antibodies at 10 μg/mL were added to wells (20 μL per well) in duplicate and incubated for 1 h at room temperature. A biotinylated preparation of recombinantly produced site-specific monoclonal antibodies were added to wells of each primary antibody at a concentration of 10 μg/mL in a volume of 5 μL per well, without washing of unlabeled antibody, and then incubated for 1 h at room temperature. Plates were washed three times with PBS-T and bound antibodies were detected using horseradish peroxidase (HRP) -conjugated anti-biotin 1:1000 (Invitrogen) and a TMB substrate. The signal obtained for binding of the biotin-labelled reference antibody in the presence of the unlabeled tested antibody was expressed as a percentage of the binding of the reference antibody alone after subtracting the background signal. Tested mAbs were considered competing if their presence reduced the reference antibody binding to less than 40% of its maximal binding and non-competing if the signal was greater than 71%. A level of 41–70% was considered intermediate competition.

Epitope binning using biolayer interferometry

An initial baseline in Octet buffer (PBS, 0.1% bovine serum albumin (BSA), and 0.02% Tween 20) was obtained for 60 s prior to loading the anti-Histidine biosensor tips (Gator Bio, Catalog #160009) with 50 μg/mL of His-tagged RSV DS-Cav1 or hMPV DS-CavEs2 F protein for 120 s. The baseline signal was measured again for 60 s before biosensor tips were associated with 50 μg/mL of primary antibody for 300 s. Biosensors were then placed directly into wells containing 50 μg/mL of a secondary mAb for 300 s for competition. Percent binding of the second mAb in the presence of the first mAb was determined by comparing the maximal signal of the second mAb after the first mAb was added to the maximum signal of the second mAb alone. Tested mAbs were considered competing if presence of the primary antibody reduced the second antibody binding to less than 40% of its maximal binding and non-competing if the signal was greater than 71%. A level of 41–70% was considered intermediate competition.

Germline reversion of BCRs

Nucleotide sequences for the heavy and light chains of the described antibodies were annotated using IMGT V-Quest. Mutations occurring outside of the CDR3 region were reverted to the residues present in the V and J genes and alleles that most closely aligned to the mature sequence.

Cell culture and virus CPE determination

LLC-MK2 cells were obtained from ATCC (CCL-7) and grown in growth media (Opti-MEM with 2% FBS) at 37°C, 5% CO₂. Propagated virus was grown in viral growth media (Opti-MEM with 5 μg/mL trypsin-EDTA and 1% antibiotic-antimycotic) in LLC-MK2 cells at a multiplicity of infection (MOI) of 0.01 for 3–5 days at 37°C, 5% CO₂ until CPE was observed. Virus was harvested using the freeze-thaw method into 25% sucrose solution and stored at −80°C until use.

Plaque reduction neutralization test with MPV (CAN/97-83 and TN/93-32) or RSV (A2, B1, A 2022 NY-Wadsworth, B 2022 NY-Wadsworth) virus

24 h prior to viral infection, LLC-MK2 (for hMPV) or HEp-2 (for RSV) cells were plated in growth media at 5 × 10⁴ cells per well in 24 well plates and incubated at 37°C, 5% CO₂. The day of viral infection, mAbs were serially diluted in Opti-MEM with a starting concentration of 40 μg/mL. hMPV (CAN/97-83 and TN/93-32) or RSV (A2, B1, A 2022 NY-Wadsworth, B 2022 NY-Wadsworth) virus was diluted in Opti-MEM to a final concentration of 2400 plaque-forming units (pfu)/mL and added to the mAb mixtures at a 1:1 volume ratio. The mAb/virus mixture incubated for 1 h at room temperature. Prior to adding the mAb/virus mixture to cells, confluent cells in 24 well plates were washed gently three times with PBS. mAb/virus mixture was added to each well (50 μL per well) and the plates rocked at 37°C, 5% CO₂ for 1 h. Warm overlay (0.75% methylcellulose in Opti-MEM, 5 μg/mL trypsin-EDTA and 1% antibiotic-antimycotic) was added to each well and the plates incubated for 4 days at 37°C, 5% CO₂. Following incubation, the cells were fixed with 10% neutral buffered formalin, washed with water three times, then blocked with milk blocking buffer (2% milk powder, 2% goat serum in PBS-T). Plates were washed three times with water and immunostained with human mAbs MPV364 (for hMPV) or 101F (for RSV) diluted to 5 μg/mL in milk blocking solution for 1 h at room temperature. Plates were washed three times with water before adding the secondary antibody, goat anti-human IgG Fc conjugated to horseradish peroxidase, at a dilution of 1:2000 in milk blocking solution and incubated for 1 h at room temperature. Plates were washed three times with water and developed with TrueBlue substrate by rocking for 10 min. After plaques were visibly stained by the substrate, the plates were washed once with water to stop the developing reaction. Immunostained plaques were counted and graphed on GraphPad Prism9.

HEp-2 cell immunofluorescence assay to detect mAb autoreactivity

HEp-2 cell coated slides (BION ENTERPRISES LTD ANA (Hep-2) Test System, ANK-120) were incubated with purified antibodies at 10.0 and 1.0 μg/mL or control sera in a moist chamber at room temperature for 30 min. Controls provided with the kit included anti-nuclear antibody (ANA)⁺ and (ANA)^- human sera. Slides were washed twice with PBS for 5 min. Cells were stained with FITC-goat anti-human Ig per the manufacturer’s instructions and incubated in a moist chamber at room temperature for 30 min. Slides were washed twice with PBS for 5 min, mounted with DAPI mounting medium (Southern Biotech 0100-20) and visualized by fluorescence microscopy (Olympus BX60 epifluorescence microscope coupled with a CCD camera and MagnaFire software Optronics International) at 40x magnification. Image brightness and contrast were optimized using Adobe Photoshop.

Recombinant protein production for binding assays, negative stain and cryo-EM

Stabilized ectodomains of hMPV-F A1(NL/1/00),²⁴ hMPV-F B2(TN99-419),²⁴ hMPV-F v3B-D12_D45C-V458C⁶⁶ (hMPV Canada) with a D280N substitution introduced via site directed mutagenesis, hMPV/Seattle/USA/SC3259/2015 (GenBank: MF045425.1), RSV-F A2 (DS-Cav-1),^18,76 and RSV-F B9320 (DS-Cav-1)¹² protein antigens were used for binding assays. Prefusion RSV-F strain A2 (DS-Cav-1) and was used for negative stain-EM. Prefusion hMPV-F construct DS-CavEs2-IPDS⁴³ protein was used for cryo-EM structural studies. In brief, plasmids encoding antigens were transfected into FreeStyle 293F cells (ThemoFisher) by PEI. Kifunensine and Pluronic F-68 (Gibco) were introduced 3 h post transfection. Six days later, the cell supernatant was filtered, and buffer exchanged into PBS by tangential flow filtration. Then, Step-TactinXT 4 Flow resin (IBA) was used to purify the protein from the filtered supernatant following the manufacturer’s instruction. The purified protein was then concentrated using a 10 kDa molecular weight cutoff Amicon Ultra-15 centrifugal filter unit (Millipore) and subject to a Superose 6 increase 10/300 column (Cytivia) in PBS running buffer (except for RSV A2 DS-Cav-1 which was pre-equilibrated with 2 mM Tris pH 8.0, 200 mM NaCl, and 0.02% NaN₃) for preparative size-exclusion chromatography. Peaks corresponding to trimeric species were identified based on elution volume and SDS-PAGE of elution fractions. Fractions containing pure fusion protein were pooled.

Negative stain-EM

For screening and imaging of negatively stained RM 5-1 Fab in complex with RSV-F A2 DS-Cav-1, sample was diluted to 100mg/mL with buffer containing 10 mM NaCl, 20 mM HEPES buffer, pH 7.4, and 5% glycerol and applied to glow-discharged grid with continuous carbon film on 400 square mesh copper EM grids (Electron Microscopy Sciences). The grids were stained with 2% uranyl formate (UF). Grids were examined on a 100 kV Morgagni microscope with a 1k x 1k AMT CCD camera.

Cryo-EM sample preparation and data collection

The purified hMPV-F-DS-CavEs2-IPDS was combined with RM 5-1 Fab in PBS buffer with a final concentration of 4.8 μM and 21.6 μM and incubated on ice for 3 min. Then, the 3 μL mixture was applied to a UltrAuFoil R1.2/1.3 300 mesh grid (Electron Microscopy Sciences) that had been glow-discharged with a PELCO easiGlow glow discharge cleaning system for 1 min. Grids were plunge-frozen using a Vitrobot Mark IV (ThermoFisher Scientific) at 4°C, 100% humidity. Blot settings were 4s of blotting with force 2. Movies (3,538) were collected from a single grid on a 200 kV Glacios microscope (ThermoFisher Scientific) equipped with a Falcon 4 direct electron detector (ThermoFisher Scientific). Data were collected at a 50-degree tilt and at a magnification of 150,000x, where the calibrated pixel size is 0.94 Å/pix and the total exposure is 48.6 e⁻/Ų.

Cryo-EM data processing

Movies were imported into cryoSPARC v4.4.0⁶⁷ for gain correction, motion correction, patch CTF estimation, micrograph curation, particle picking, and particle extraction with a 2X Fourier crop. After two rounds of particle curation through 2D class averaging, the generated 2D class averages were used as templates to perform another round of template-based particle picking. Then, the particles were curated by 2D class averaging and curated particles were subject to ab initio reconstruction, heterogeneous refinement, and homogeneous refinement with C3 symmetry applied. Due to the presence of flexibility at the bottom region of the homogeneous-refined EM map, a 3D variability analysis job was performed with a focused mask to explore alternative conformations. After 3D variability analysis, a 3D classification job with a focused mask on the hMPV F base region was executed to generate EM maps of different conformations, followed by heterogeneous refinement. As particles were processed with Fourier cropping in the procedure described above, we re-extracted the particles with raw pixel size, removed the duplicate particles and reconstructed one EM map with homogeneous refinement and reference-based motion correction. Finally, the map from the last round of homogeneous refinement was sharpened using DeepEMhancer.⁷⁷ For model building, an initial model was generated by AlphaFold3 server.⁴⁴ As the predicted model aligned well with our 3D EM map, the following iterative refinements were performed using this model in Coot,^68,69 PHENIX⁷⁰ and ISOLDE.⁷¹ The adjacent cystines in 5-1 Fab CDRH3 loop were modeled as a disulfide bond in the AlphaFold3 predicted model and were left unchanged during refinement. At the last round of refinement, glycans were built into the model, refined and validated using Coot and Privateer software.⁷² The EM processing workflow is shown as Figure S3 and EM validation results are shown in Figure S4. Refinement statistics are shown in Table S1.

Sequence conservation analysis and alignment

The glycoprotein sequence of hMPV F protein from strain NL/1/100 (A1 sub lineage, NCBI accession: YP_009513268.1) was uploaded into the HMMER web server⁷⁴ to search for homologous sequences against UniProtKB database with phmmer programs and default parameters. The searching results were then manually filtered based on species, similarity, coverage and hit position. To avoid potential bias, 250 sequences for both hMPV F and RSV F were extracted from the search results and aligned with Clustal Omega.⁷⁸ The output was imported into ChimeraX⁷³ to generate a sequence conservation map. For direct alignment of two representative hMPV F and two RSV F protein sequences, hMPV F lineage CAN97-83 (A2 subgroup, NCBI accession: AAQ67695.1), hMPV F TN93-32 (B2 subgroup, NCBI accession: AAK62968.2),and RSV F from A2 subgroup (NCBI accession:ALS35589.1) and B1 subgroup (NCBI accession: AAB82436.1) sequences were pooled and aligned with Clustal Omega. Related to Table 2: Full length RSV F and hMPV F sequences collected between 2024 and 01-01 to 2025-09-08 (acquired from Pathoplexus) were aligned with representative RSV and hMPV strains to provide average pairwise identities. Sequences were aligned with Clustal Omega using Geneious software.

Quantification and statistical analysis

Statistical analyses were performed using GraphPad Prism v9.5.0. ELISAs were analyzed by fitting a four-parameter logistic regression curve to determine area under the curve (AUC). Neutralization assays were analyzed by fitting a four-parameter logistic curve to interpolate IC₅₀ values. Viral titers in mice treatment groups were analyzed by one-way Anova. Quantification details can be found in the figure legends.

Published: January 16, 2026