Adaptive Multi-Epitope Targeting and Avidity-Enhanced (AMETA) Nanobody Platform: Leveraging Diverse Mechanisms for Ultrapotent, Durable Antiviral Therapy

Summary

Pathogens, known for their high genetic diversity, constantly develop escaping mutations that evade host immunity and treatment. Mitigation of pathogen’s mutational escape requires targeting multiple evolutionarily conserved vulnerabilities, as changes often incur penalties to pathogen fitness. We introduce AMETA, a modular multivalent nanobody platform. AMETA conjugates potent and bispecific nanobodies to the human IgM scaffold. It can be armed with 20 (or more) nanobody warheads, enabling superior avidity binding against multiple conserved and neutralizing epitopes. Using multi-epitope SARS-CoV-2 nanobodies and structure-guided design, we have crafted AMETA constructs that exponentially improve antiviral potency, surpassing nanobody monomers by over one million-fold. These constructs demonstrate ultrapotent, broad and durable effectiveness against pathogenic sarbecoviruses, including highly evolved Omicron sublineages, showing high preclinical efficacy. Structure analysis by single-particle cryo-electron microscopy, cryotomography, and modeling has unveiled multiple unprecedented antiviral mechanisms within a single construct. At picomolar to low nanomolar concentrations, it can efficiently promote inter-spike and inter-virus cross-linking, induce the spike’s post-fusion state and strikingly disarm the virus. By simultaneously targeting multiple conserved epitopes, AMETA achieves enhanced avidity and maximizes antiviral effectiveness and shows promise in mitigating future escapes. Its modularity harnesses swift, cost-effective nanobody production and engineering, enabling rapid adaptation to pathogen evolution through updated nanobody designs.

Introduction

Pathogens such as viruses and bacteria often have high genetic diversity and share the ability to develop escaping mutations that undermine the effectiveness of host immunity and therapeutics(Carabelli et al., 2023; Harvey et al., 2021). The development of durable countermeasures requires targeting multiple neutralizing and ideally evolutionarily conserved epitopes(Chen et al., 2023; Klein et al., 2012; Walker and Burton, 2018). Few mutations are observed on these sites as they may compromise pathogen fitness(Starr et al., 2020). Monoclonal antibodies provide high specificity but require frequent updates against evolving variants (Liu et al., 2022a; Wang et al., 2023). Broad spectrum and long-lasting therapeutics are critically needed (Chen et al., 2023; Duan et al., 2023).

IgM is a pentameric antibody that circulates in the mucosal membrane and blood. IgM can recognize pathogens by using multivalency to overcome their lack of affinity maturation(Gronwall et al., 2012; Schroeder and Cavacini, 2010). It is effective in agglutination and complement activation, making it crucial in early defense against pathogens(Blandino and Baumgarth, 2019; Ehrenstein and Notley, 2010; Gupta and Gupta, 2017; Keyt et al., 2020). However, IgM’s low abundance and complex, flexible structure pose obstacles to discovery, recombinant production, and engineering (Gautam and Loh, 2011). The lack of specificity of natural IgM is another challenge for therapeutic development.

Nanobodies (Nbs) have emerged as a promising solution for antiviral therapy(Huo et al., 2020; Li et al., 2022; Mast et al., 2021; Nambulli et al., 2021; Pymm et al., 2021; Schoof et al., 2020; Xiang et al., 2020; Xu et al., 2021). As a clinically validated modality, these minimal antibody fragments (15 kDa), derived from camelid heavy-chain antibodies, are readily produced and engineered (Hamers-Casterman et al., 1993; Shen et al., 2022). Affinity-matured Nbs can specifically bind a variety of neutralizing epitopes, including conserved and cryptic sites, employing diverse mechanisms to thwart viral infections(Koenig et al., 2021; Sun et al., 2021; Xiang et al., 2022). Here, we report the development of AMETA (Adaptive Multi-Epitope Targeting with enhanced Avidity), a modular platform that integrates the unique antiviral properties of Nbs with the marked multivalency of IgM to combat drug-resistant microbes. We selected pandemic sarbecoviruses as a model system for developing AMETA, recognizing their profound clinical impact, extensive genetic diversity and numerous escaping variants(Boni et al., 2020; Carabelli et al., 2023). We covalently conjugated a versatile repertoire of specific, potent, and bi-epitope antiviral Nbs to a robust human IgM core scaffold, resulting in a class of therapeutic agents armed with dozens of Nb warheads. Fueled by unparalleled avidity and multiple epitopes targeting capabilities, our AMETA designs maximize the efficacy and breadth against pandemic sarbecoviruses, including a diverse array of highly evolved variants. Their stability in the lower respiratory tract and high preclinical efficacy were demonstrated through intranasal administration in mouse models. Hybrid structure analyses further revealed a variety of unprecedented antiviral mechanisms, such as efficient viral cross-linking and disarming of viral spikes. Altogether, these results underscore AMETA as a modular, effective and enduring strategy against challenging pathogens.

A Robust System to Produce Avidity-Enhanced Nanobodies

Addressing significant challenges in IgM production and engineering, we have developed a modular system that leverages SpyTag003-SpyCatcher003(Keeble et al., 2019) to directly and covalently conjugate Nbs to a stable human IgM scaffold. The core IgM scaffold encompasses the constant regions of the heavy (μ) chain spanning Cμ2-Cμ4, and the J-chain (Keyt et al., 2020), replacing the Fab domain with SpyTag003 (13 amino acids). To mitigate potential immunogenicity, we engineered a small and truncated SpyCatcher003 variant by deleting the immunogenic N-terminal domain (NTD) (Liu et al., 2014). This resulted in a compact structure that retains a single immunoglobulin (Ig) fold (Ig_{ΔNSpyCatcher003}), yet exhibits full activity comparable to the full-length SpyCatcher003. Nb-Ig_{ΔNSpyCatcher003} fusion constructs were produced with high solubility from E.coli while the SpyTag003-IgM scaffold was expressed in high yields from mammalian cell secretion. Nb-Ig_{ΔNSpyCatcher003} were efficiently conjugated to the SpyTag003-IgM scaffold through incubation at a high molar ratio (e.g., 30:1). The resulting Nb-IgM-Fc conjugates were highly stable and produced abundantly, effectively addressing the issues associated with poor expression of the genetic Nb-IgM-Fc fusions (Figure S1). Fully conjugated proteins were rapidly purified and concentrated using a 100 kDa molecular weight cut-off centrifugal filter (Figure 1A). Analyses through denatured SDS and native protein gels confirmed the complete covalent attachment of Nbs across all ten arms of the IgM core (Figure 1B). Size exclusion chromatography (SEC) and mass photometry further corroborated these conjugates’ high purity and expected masses (Figure 1C).

Figure 1. A modular system to produce multivalent nanobodies with enhanced avidity.

(A) Schematics of conjugation and purification of Nb-IgM-Fc constructs. Nanobodies and recombinant IgM-Fc were incubated with a 30:1 molar ratio at 4°C overnight. Excessive nanobodies were removed through centrifugation using a 100 kDa molecular weight cutoff filter. Nb-IgM-Fc conjugates were efficiently recovered from the filter.

(B) Denatured SDS and native protein gel analyses showing the complete conjugation and high purity of the conjugates.

(C) Mass photometry analysis of a conjugated Nb-IgM-Fc. The total measurement mass is 731±143 kDa, including the predicted molecular weight of 708 kDa from amino acids and glycosylations on the IgM-Fc.

(D) Summary of five major epitope classes targeted by RBD nanobodies, based on high-resolution cryo-EM structures.

(E) Correlation analysis of neutralization potency (EC₅₀s) of Nb-IgM-Fc and the corresponding Nb monomers against the pseudotyped SARS-CoV-2 Wuhan-Hu-1 (D614G) strain. Each dot represents an Nb.

(F) Plots summary of EC₅₀s of Nb monomers and their corresponding Nb-IgM-Fc against pseudotyped SARS-CoV-2 variants and SARS-CoV. Nanobodies targeting different epitope classes were plotted separately.

(G) Summary of neutralization potencies (EC₅₀s) of Nb-IgM-Fc conjugates against pseudotyped SARS-CoV-2, variants and SARS-CoV. Welch’s t-tests were used for comparison of EC₅₀s of Nb-IgM-Fc targeting different epitopes. *: p < 0.05, ***: p < 0.001, ****: p < 0.0001.

(H) A plot of average solvent-accessible surface areas (SASA) of four major epitopes of RBD Nbs during the transition of an RBD from the “close” to fully open state. SASA of individual Nb can be found in Figure S4. The transition was modeled by Molecular Dynamics (MD) simulation (Methods). The exposed area was presented as a function of the distance between the centroids of Subdomain 1 (SD1, amino acids 531–592) and the RBD (amino acids 336–518) across different protomers of the Spike protein. All calculations of accessible surface area were conducted using a 7-Å probe.

To assess the potential avidity effects of the IgM scaffold on Nbs, we conjugated a diverse collection of high-affinity anti-SARS-CoV-2 Nbs (Table S1)(Li et al., 2021; Xiang et al., 2022; Xiang et al., 2020). These Nbs specifically target at least five neutralizing epitopes on the receptor-binding domain (RBD) (Figure 1D). When compared to monomeric Nbs, conjugation to the IgM-Fc significantly enhanced their neutralization activities against SARS-CoV-2 (Wuhan-Hu-1 D614G), with an average improvement of approximately 500-fold, ranging from 4 to 7,000-fold (Figures 1E, S2, S6B, Table S2). These Nb-IgM-Fc conjugates also outperformed their Nb-IgG-Fc counterparts, indicating the enhanced avidity afforded by the pentameric IgM scaffold (Figure S3, Table S3) (Ku et al., 2021; Liu et al., 2023a; Liu et al., 2023b; Singh et al., 2022). A notable positive correlation was observed between the neutralization potency (EC₅₀s) of individual Nbs and their corresponding conjugates (Figure 1E).

Given these monomeric Nbs can be evaded at different extents by highly mutated variants, we investigated whether the Nb-IgM-Fc conjugates had elevated avidity and activity against variants such as Omicron BA.2, BA.4/5, XBB.1.5, XBB.1.16, EG.5, BA.2.86, and the more antigenically distant SARS-CoV. For the Nbs that showed weak activities (i.e., 2.5 micromolar/μM), the avidity-enabled conjugates drastically enhanced efficacy against these highly evolved variants. The most pronounced improvements were observed for epitopes I and III Nbs (Figures 1F, S2, S7B), which strongly bind to variable receptor binding sites and semi-conserved epitopes, respectively. Epitope I and III Nb-IgM-Fc conjugates generally potently neutralize the viruses in the sub-nanomolar (nM) range (Figure 1G). In contrast, conjugates containing Nbs targeting more conserved epitopes, such as II, IV, and V, displayed moderate improvements, with median EC₅₀ values ranging from low nM to μM concentrations (Figures 1G, S2, S7B). Of note, Epitope II Nbs and their conjugates most efficiently neutralized the more antigenically distant SARS-CoV (Figure S2). Interestingly, more substantial improvement in antiviral potency was observed in more solvent-accessible epitopes such as I and III, implying favorable cross-linking by the Nb-IgM-Fc conjugates (Figures 1H, S4). However, Nbs lacking initial neutralization activity against these viruses did not benefit from conjugation to the IgM scaffold, as evidenced by the absence of conferred protection (detailed in Table S2, Figures S2, S7B).

Optimizing Avidity and Multi-Epitope Targeting of AMETA

We selected nine highly soluble and high-affinity RBD Nbs that bind four distinct neutralizing epitopes to design multi-epitope targeting dimers with improved activities(Xiang et al., 2022; Xiang et al., 2020). Dimers that show improved neutralization indicate cooperative binding and are preferred warheads for AMETA design to further enhance avidity. These Nbs specifically target four neutralizing epitopes including conserved and cryptic sites. Guided by structure information, we generated six distinct epitope combinations, predominantly heterodimers and a homodimer of epitope II Nb (118). A flexible GS peptide linker was introduced between the Nbs to facilitate cooperative binding. These dimers span epitopes containing 11–44% of conserved surface residues on the SARS-CoV-2 RBD (Figure 2A) and can be rapidly produced from E.coli with high expression. The majority of these dimers (11/14) showed improved neutralization potency (Figures 2A, S5, S6, Table S4), indicating successful designs for cooperative binding, either within a single RBD or across two RBDs on the spike glycoprotein (Figure 2B). From these, we selected lead dimers from each combination based on yield and potency improvements, to act as a warhead in constructing AMETA multimers. To bolster valency and epitope coverage further, we conjugated two unique dimer warheads (in a 1:1 molar ratio) to the IgM-Fc scaffold. Using seven lead dimers, we generated four distinct AEMTA constructs (Figures 2C, 2D). They were efficiently conjugated and produced in high purity, with intact masses consistent with the predictions (Figures 2E, 2F, S7A). The resulting AMETA constructs are each armed with 20 Nbs for superior avidity binding. They target three or four RBD epitopes, covering a substantial portion (between 42%- 53%) of conserved, solvent-exposed residues (Figure 2D).

Figure 2. Engineering and production of AMETA constructs against sarbecoviruses.

(A) Structure-guided design of Nb dimers to improve cooperative binding and neutralization activities. Sequence conservation is based on 19 sarbecovirus RBD sequences (Methods). The fold improvement of a dimer over the corresponding monomers was based on their average EC₅₀s against the SARS-CoV-2 Omicron variants and SARS-CoV (Table S4).

(B) Models of the dimeric nanobodies on the SARS-CoV-2 spike glycoprotein. Upper panel: cooperative binding within a single RBD. Lower panel: cooperative binding between two RBDs.

(C) Schematic presentation of a representative AMETA construct.

(D) Design of four AMETA constructs and their potential coverage on the conserved RBD surface residues.

(E) Denatured SDS and native protein gel analyses showing the complete conjugation and high purity of the AMETA constructs.

(F) Mass photometry analysis of AMETA constructs.

Next, we used highly mutated SARS-CoV-2 variants, including various Omicron lineages (BA.2, BA.4/5, XBB.1.5, XBB.1.16, EG.5, BA.2.86) and the original SARS-CoV to assess the broad-spectrum activities of AMETA constructs. Specific variants, such as BA.2.86, have been shown to resist all clinical antibodies and can dramatically evade passive immunity by vaccines (Qu et al., 2024; Zhang et al., 2024). Remarkably, all AMETA constructs potently neutralized every virus we tested (Figures 3A, S7B, S8E). AMETA1 and AMETA2 were particularly effective against earlier Omicron variants and SARS-CoV, with EC₅₀s ranging from 10 to 50 picomolar (pM). Their potencies against later Omicron variants (i.e., XBB.1.5, XBB.1.16, EG.5, BA.2.86) were only moderately decreased, ranging from 150 to 600 pM. In contrast, AMETA3 and AMETA4 showed consistently low EC₅₀s within the 10– 100 pM range. Their in vitro potencies may be restrained only by diffusion limits. Moreover, we conducted side-by-side comparisons of AMETA constructs (targeting multi-epitope) with the corresponding Nb-IgM-Fc conjugates (targeting a single-epitope) (Figures 3B–E, Table S2). Interestingly, 182-IgM-Fc, which targets non-conserved epitope I, was completely evaded by the most recent variants and SARS-CoV. This complete loss of activity was also seen in the monomer. Other single-epitope conjugates were either partially evaded or exhibited inferior potencies, typically in the nM range. Multi-epitope targeting with high valency appeared to be crucial for neutralizing highly mutated variants, as shown by AMETA’s unprecedented potencies and broader activities. Notably, AMETA3 and AMETA4 outperformed the corresponding Nb-IgM-Fc conjugates by a factor of 10–10,000 against later Omicron variants and SARS-CoV (Figures 3B–E, Table S2). Such enhancements were more dramatic–exceeding one million-fold – when compared to specific Nb monomers (Figures S7B, S8A–D, Table S2).

Figure 3. In vitro neutralization activities of the AMETA constructs.

(A) Summary of neutralization potencies of AMETA against a panel of SARS-CoV-2 variants and SARS-CoV.

(B)-(E) Radar plots comparing the neutralization potencies (EC₅₀s) of four AMETA constructs and their corresponding single-epitope Nb-IgM-Fc. The highest tested concentration was 100 nM.

High Preclinical Efficacy of AMETA

To assess AMETA’s in vivo distribution and stability, we conjugated AMETA4 with ⁸⁹Zr and delivered this tracer-conjugate intranasally to C57BL/6 mice. Intranasal administration of therapeutics may offer protection against respiratory infection in both upper and lower airways(Duty et al., 2022; Imsuwansri et al., 2023; Ku et al., 2021). We performed whole-body PET/CT scans on various time points post-delivery (Figure 4A). The construct rapidly distributed to both upper and lower respiratory tracts. PET/CT scan intensities in the nasal cavity were initially abundant and dropped below detectable levels within 24 hours. However, the presence of the construct remained highly detectable in the lung throughout the entire 48-hour analysis period. Ex vivo gamma counting at 72 hours post-delivery reveals significant and specific enrichment in the lung. Minimal signals were detected in the bloodstream and most other organs, indicating a low systemic distribution. The exception to this was observed in the gastrointestinal tract, where PET signals were nonspecifically introduced during intranasal dosing (Figure 4B). Our data underscore the sustained presence of AMETA in the lung by targeted delivery, which could be pivotal for respiratory therapies.

Figure 4. Biodistribution and in vivo efficacy of AMETA4 in mouse models.

(A) Intranasal administration of ⁸⁹Zr-labeled AMETA4 (275.5 ± 43.5 mCi) in C57BL/6 mice (N=5) was followed by a 20-minute PET scan to monitor AMETA4’s distribution. Representative 3D reconstruction images are shown.

(B) Mice were euthanized 72 hours post-injection, and tissues were analyzed for radioactivity. Gamma-counting determined the residual tracer, corrected for decay, and expressed as a percentage of the injected dose per gram of tissue (%ID/g).

(C) Experimental design summary: mouse-adapted SARS-CoV-2 (1 × 10⁴ PFU) was intranasally administered to three groups of S/129 mice (N=8). AMETA4 (2.2 nmole/kg or 2 mg/kg) was delivered intranasally either 6 hours before (green squares) or 6 hours after (purple triangles) infection. A control group received isotype IgM (gray circles, 2.2 nmole/kg), and a group of uninfected animals served as additional controls. Daily monitoring of animal weight changes was conducted, and animals were euthanized for lung tissue viral titer analysis on day 3.

(D) Body weight changes over time, expressed as a percentage. Significant differences are marked with *** (P < 0.001).

(E) Viral titers in lung tissues at 3 days post-infection (d.p.i.), showing significant reductions in treated groups compared to controls. Significance levels are indicated as **** (P < 0.0001) and * (P < 0.05).

(F) Lung pathology scores, comparing treated and control groups.

We then evaluated the in vivo antiviral efficacy of AMETA4. We divided 32 mice into four groups. Three groups were infected intranasally with 10,000 PFU of mouse-adapted SARS-CoV-2 (USA-WA1/2020) strain and the fourth group was used as a non-infected control(Rathnasinghe et al., 2021). As for prophylaxis evaluation, we administered AMETA4 or IgM-Fc core (2 mg/kg) intranasally to the first two groups 6 hours before infection. As for therapeutic evaluation, AMETA4 was delivered to the third group at 6 hours post-infection. We monitored weight changes and clinical signs for three days before sacrificing the animals (Figure 4C). Lung viral titers were measured using the microneutralization assay. All infected animals with IgM-Fc core treatment experienced rapid weight loss (~ 10% by day 3 post-infection, Figure 4D) and high viral lung titers (median ~10⁸ PFU, Figure 4E). However, AMETA4 treatment significantly protected against these effects in both prophylaxis and treatment settings, indicated by a dramatic rebound in weight loss (Figure 4D). Viral lung titers were significantly reduced in all AMETA-treated animals (Figure 4E). Compared to the IgM-Fc control, prophylactic and therapeutic treatment by AMETA4 drastically reduced lung viral titers by 6-logs and 3-logs, respectively. Consistently, treatments significantly alleviated lung pathology and virus-induced inflammation (Figure 4F). These data demonstrated the high in vivo efficacy of AMETA in protecting against viral challenges. Given the accelerated virus replication in this model compared to humans, our findings indicate the potential benefits of the AMETA technology in preventing severe human symptoms.

Hybrid Structural Analysis Revealing AMETA’s Dynamic Architecture and Multifaceted Antiviral Mechanisms

To explore the enhanced avidity effects, we employed single-particle cryo-electron microscopy (cryo-EM) to image the structure of a representative AEMTA construct (AMETA3). Our analysis resolved most IgM-Fc core components, including the Cμ3-Cμ4 pentamer and J chain (Figure S9), aligning with recently published structures(Chen et al., 2022; Li et al., 2020). Other components such as Cμ2, the Ig_{ΔNSpyCatcher003} domain, Nbs, and linker peptides were unresolved due to their inherent flexibility. Utilizing the cryo-EM findings, we developed an in silico model of AMETA3 (Methods). Its overall architecture is reminiscent of a carousel, with 20 Nbs (akin to wooden horses) arranged out-of-plane around the rigid IgM-Fc core with their CDR loops highly accessible (Figure 5A). Replacing the natural IgM’s Fab region with Nb dimers and the Ig_{ΔNSpyCatcher003} domains may result in high flexibility, facilitating both in-plane and out-of-plane pivoting(Chen et al., 2022). Coupled with additional flexible linkers (between Cμ2 and the Ig_{ΔNSpyCatcher003}, between the Ig_{ΔNSpyCatcher003} and Nb, and between bi-epitope Nbs), AMETA’s arms are likely considerably more dynamic than the Fab regions. Molecular Dynamics (MD) simulations, corroborating the cryo-EM observations, have revealed significant structural flexibility in the conjugated Nbs (Figures 5B, 5C). Such flexibility may help different Nbs to effectively target a variety of viral epitopes including both conserved and cryptic ones, thereby improving avidity binding and antiviral efficacy.

Figure 5. The architecture of an AMETA construct.

(A) Structure model based on single-particle cryo-EM illustrating the full architecture of an AMETA construct (AMETA3). Bi-epitope Nb dimers S36–182 (targeting epitope III and I, respectively), as well as 132–118 (targeting epitope IV and II, respectively), are distinguished by varying colors. Additional components such as the Cμ domains, J chain, Ig_{ΔNSpyCatcher003} domain, and peptide linkers are annotated.

(B) Structural model of a single clamp of AMETA3 derived from a 50 ns full-atom Molecular Dynamics (MD) simulation, encompassing 100 frames. Models from all the frames were aligned to the initial conformations of Cμ3 and Cμ4 domains (amino acids 404–728 and 1159–1484) to generate the figure. The Nb sequences (S36, 182, 118, and 132) are colored in pink, green, yellow, and blue, respectively.

(C) MD simulation displaying the root-mean-square fluctuation (RMSF) of AMETA3 at the amino acid level, highlighting the notable flexibility of the Nb dimers, Ig_{ΔNSpyCatcher003} and the linkers.

To better understand AMETA’s marked antiviral activities, we incubated our constructs with native SARS-CoV-2 (Victoria) at different concentrations and performed cryo-electron tomography (cryo-ET) to visualize their binding to the virus in situ. It has been shown that spike glycoproteins on isolated inactivated SARS-CoV-2 virus particles display a significant amount of post-fusion conformation, ranging from 25% to 75% (Ke et al., 2020; Liu et al., 2020; Turonova et al., 2020; Yao et al., 2020). However, spikes from intracellular and egressed virus particles in the context of infected cells are all in the pre-fusion conformation (Mendonca et al., 2021). Therefore we carried out the AMETA binding experiments directly with egressed SARS-CoV-2 in tissue culture in the BSL-3 containment lab prior to fixation and vitrification for subsequent cryo-ET analysis. In addition to blocking RBD binding to the receptor, our results indicate the presence of multiple other critical mechanisms by which AMETA can leverage to inhibit the virus efficiently. At an exceptionally low concentration close to the neutralization EC₅₀ values (i.e., 10 pM), AMETA constructs induced inter-spike cross-linking, which may impair virus-host interactions. Additional EM densities corresponding to AMETA constructs were visualized to cross-link multiple spikes on the same virus (Figures 6A, 6E, Supplemental Movie 1, 2). As we increased AMETA’s concentration to 100 pM, virus cross-linking was evident and became a predominant phenomenon at 0.1–1 nM concentration (Figures 6B, 6C, S10, Supplemental Movie 3–6, 9–14). Intriguingly, we found that AMETA can also lead to massive spike disarming from the virions. This is supported by tomography data that AMETA-treated virions drastically reduced their spike numbers (Figures 6F, 6G, S10). For the IgM-Fc control sample, the average number of spikes per viral particle was 20.6. However, the number dropped drastically to 3.4 in the AMETA-treated samples (1 nM) (Figures 6C, 6D, 6G, Table S6, Supplemental Movie 7, 8). The loss of spikes may be triggered, at least in part, by the presence of epitope III (and likely epitope IV) Nbs, which can uniquely destabilize the integrity of the recombinant spike trimer at high concentrations (e.g., > 100 nM) (Sun et al., 2021). The combined effects of enhanced avidity and multiple epitopes binding likely greatly exacerbated these destabilizing impacts, leading to the emergence of unusually “bald”, disarmed virions. In addition, a significant presence of post-fusion spikes was detected up to 90.5% (at 1 nM of AMETA4), in stark contrast to the isotype control (2.6%, Figure 6H, Table S6). The decrease in viral spike count and the rise in the prevalence of the post-fusion state both depend on the concentration of AMETA. This marks the first observation of essential viral structures being disarmed or deconstructed by a therapeutic molecule in the native context– strikingly, with such high efficiency. Collectively, our data have revealed a plethora of unprecedented antiviral mechanisms by AMETA, underscoring their unique properties and potential synergy in the development of ultrapotent, broad-spectrum and durable antiviral therapeutics.

Figure 6. In situ cryo-electron tomography (cryo-ET) analysis of AMETA binding to authentic SARS-CoV-2.

Various concentrations of AMETA4 (10 pM - 1nM, A to C) or an isotype control of recombinant IgM-Fc protein (D) were incubated with the egressed SARS-CoV-2 (Victoria) in tissue culture before fixation and vitrification for subsequent cryo-ET analysis. The tomographic slices presented are 3.6 nm in thickness. Blue arrows indicate pre-fusion spikes, red arrows post-fusion spikes, purple arrows the AMETA4 densities crosslinking the spikes, and yellow arrows dense AMETA surrounding the viruses. Scale bar: 50 nm.

(E) Segmentation of the slice (Figure A, lower panel), showing AMETA4 at a concentration of 10 pM. Pre-fusion spikes are in blue, and partially resolved AMETA densities are shown in purple.

(F) Segmentation of slice (Figure B, lower panel) corresponding to AMETA4 at 100 pM concentration. The virus ribonucleoproteins (RNPs) are in beige. Unassigned densities in cyan imply either dissociated spike or AMETA4.

(G) Quantification of the average number of identifiable spikes per viral particle under each condition (see details in Table S6).

(H) Quantification of the percentage of spike confirmation (pre-/post-fusion) per identifiable spike under each condition (see details in Table S6).

Discussion

Around a quarter of global mortality is attributed to infections caused by viruses, bacteria and fungi, both known and emerging(Diseases and Injuries, 2020). Pathogenic microbes have high genetic diversity and share the ability to develop mutations that undermine the effectiveness of host immunity, vaccines, and therapeutics. Mutational escape and its effects on countermeasures have been investigated through the unparalleled scientific surveillance of SARS-CoV-2, making it an ideal system to advance broad-spectrum and durable strategies against escaping variants (Carabelli et al., 2023; Harvey et al., 2021). Central to thwarting mutational escape is by targeting conserved epitopes. Alterations in these residues may incur functional or structural penalties to pathogen’s fitness and survival advantage. Strategies that achieve enhanced avidity, high binding affinity, and comprehensive coverage of multiple epitopes are likely instrumental in ensuring broad-spectrum activity against the ever-evolving challenge of infectious diseases(Walker and Burton, 2018).

With these considerations in mind, we have surmounted significant technical challenges to successfully develop a modular and multivalent nanobody platform. Our AMETA system combines the remarkable avidity of the IgM scaffold with the high specificity and exceptional bioengineering potential of Nbs. The miniature size and small footprints of Nbs enable avidity binding of AMETA constructs to an extensive range of pathogenic epitopes including small nooks, crevices and conserved sites (Desmyter et al., 1996; Spinelli et al., 1996; Wilson and Stanfield, 2021). Through targeting multiple epitopes—covering over 50% of conserved solvent-exposed residues on the RBD—and enhancing avidity, AMETA constructs have shown unparalleled antiviral potency (in the picomolar range) and breadth against sarbecoviruses, including SARS-CoV, ancestral SARS-CoV-2 and all highly mutated SARS-CoV-2 variants we have tested, a stark contrast to monoclonal antibodies and nanobodies. AMETA has significantly outperformed Nb-IgM-Fc conjugates, which exhibited limited potency and breadth against highly evolved variants. Utilizing a murine infection model, we have demonstrated the outstanding in vivo efficacy of a representative AMETA construct for both prophylaxis and early therapeutic applications. Our cryo-EM, MD simulation, and hybrid structural modeling have further elucidated AMETA’s architecture, highlighting the structural dynamics of 20 Nb warheads. The high valency and flexibility of AMETA may enable efficient engagement with a wide variety of pathogen geometries and epitope diversities, maximizing avidity binding and efficacy.

Through cryo-ET analysis, we discovered a plethora of antiviral mechanisms by AMETA beyond the classical mechanism of neutralization by competition with binding to the receptor. First, AMETA constructs can cross-link the viral spikes. A similar phenomenon has recently been observed in a rare anti-Zika human IgM(Singh et al., 2022). However, it is surprising that AMETA constructs can promote cross-linking at such low (pM) concentrations. Second, AMETA constructs can potently induce cross-linking of viruses. At sub-nM to low nM concentrations, we observed predominant inter-viral crosslinking. It is likely that virus cross-linking represents one of the major mechanisms accounting for the high preclinical efficacy of AMETA constructs. Cross-linking may facilitate fast virus clearance in vivo through the effective recruitment of the effector’s functions(Blandino and Baumgarth, 2019). Nevertheless, enhanced viral cross-linking corroborates strong preclinical evidence in which infectious particles were below the detection limit after intranasal delivery of AMETA for prophylaxis treatment. Third, and perhaps to our greatest surprise, we observed a drastic, almost complete loss of pre-fusion spikes in a significant portion of AMETA-treated viruses. This is associated with a substantial increase in the post-fusion state. It is possible that through multi-epitope targeting, avidity-enhanced binding towards specific epitopes may trigger spike destabilization, although the exact mechanisms remain to be fully understood.

We note that AMETA’s warhead Nbs were isolated during the winter of 2020. While these Nbs have shown broad effectiveness against a variety of SARS-CoV-2 variants, the relentless evolutionary pressure exerted by widespread global transmission in the ensuing three years has allowed highly mutated Omicron subvariants to evade most of them(Cao et al., 2022; Planas et al., 2022; Qu et al., 2024; Wang et al., 2023; Zhang et al., 2024). Nonetheless, the innovative AMETA design enables our multivalent Nbs to drastically improve their efficacy. They not only potently neutralize past and current variants but also show promising resilience against future mutations due to their broad spectrum neutralizing activities. Hence, our data demonstrate the exceptional and enduring efficacy of our technology, highlighting its potential for long-term viral defense despite rapid antigenic evolution.

AMETA’s versatility and modularity mark a substantial advancement over existing antibody technologies, addressing limitations of IgM therapy such as production challenges and bioengineering constraints. Its adaptability to other modalities (such as nucleotides, chemical compounds, antibody fragments and in silico designed proteins) positions AMETA as a promising tool against highly mutable pathogens like HIV and influenza, and its multispecific nature could have broad applications in oncology(Elshiaty et al., 2021; Keri et al., 2024). Our findings broaden the scope of therapeutic strategies against infectious diseases and drug-resistant systems, paving the way for innovative approaches to combat a wide array of pathogens.

Limitations of the study

Despite extensive efforts, our study does not include high-resolution structural determination of AMETA constructs in complex with the virus, due to the significant flexibility of AMETA constructs and their varied binding modes to the virus. Resolving these structures unambiguously could provide deeper insights into their antiviral mechanisms and inform the rational design of future interventions against other challenging pathogens. While we have conducted comprehensive evaluations of our technology against a broad spectrum of escape variants, we have not performed gain-of-function studies to identify potential escaping mutations (if any) under AMETA’s selection pressure. Such research is subject to regulatory scrutiny. Our investigations have unveiled several novel antiviral mechanisms; however, as AMETA constructs derive from the natural IgM scaffold, they may also efficiently harness IgM’s effector functions for in vivo viral clearance(Wang et al., 2016), meriting further study. Lastly, although AMETA incorporates a fully human IgM scaffold and Nbs–a clinically validated modality– future clinical trials are required to thoroughly assess its safety profile for human applications.

STAR Methods

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Yi Shi (wally.yis@gmail.com).

Materials availability

All plasmids generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

Atomic coordinate electron cryo-microscopy structures in this paper have been deposited to the Protein Data Bank (PDB). The accession numbers are available in Table S5.

The structure, MD simulation of AMETA, and open dynamics of an RBD of the SARS-CoV-2 spike analyses have been deposited to Zenodo. The accession numbers are available in the Key Resource Table.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
in-house mAb 1C7 against SARS-CoV-2 NP protein, provided by Dr. Andrew Duty	BioSS	bs-41408P
Bacterial and Virus Strains
One Shot™ BL21(DE3) Chemically Competent E. coli	Invitrogen	C600003
Subcloning Efficiency™ DH5α Competent Cells	Invitrogen	18265017
SARS-CoV-2 Victoria strain	Tao Dong’s group, CAMS-Oxford Institute	N/A
SARS-CoV-2 USA-WA1/2020	BEI Resources	NR-52281
SARS-CoV-2 Beta (B.1.351), Omicron (BA.5, XBB.1.5, XBB.1.16)	Mount Sinai Pathogen Surveillance program	IRB HS#13-00981
Biological Samples
SARS-CoV-2 and variants, SARS-CoV pseudovirus particles (luciferase reporter)	Integral Molecular	RVP-702L, 707L, 770L, 774L, 786L, 789L, 792L, 794L, 801L
Chemicals, Peptides, and Recombinant Proteins
LB Miller Broth	IBI Scientific	49030
Ampicillin, sodium salt, irradiated	Gibco	11593027
LB Miller’s Agar Powder	Growcells.com	MBPE-3020
Dulbecco’s phosphate buffer saline no Ca 2+, Mg 2+	Thermo Fisher	14190250
Pierce™ Protease Inhibitor Tablets, EDTA-free	Thermo Fisher	A32965
Tween® 20	Sigma	P9416-100ml
Puromycin	Gibco	A1113803
Penicillin-Streptomycin Solution, 100 x	Corning	30-002-CI
Critical Commercial Assays
Monarch® PCR & DNA Cleanup Kit	NEB	T1030L
Monarch® DNA Gel Extraction Kit	NEB	T1020S
EndoFree Plasmid Kits	Qiagen	12362
ExpiFectamine 293 kit	Thermo Fisher	A14635
Renilla-Glo luciferase assay system	Promega	E2720
Mycoplasma Detection Kit	InvivoGen	rep-mys-20
Deposited data
Structure, simulation and open dynamics analysis	Zenodo	10.5281/zenodo.10684642
Experimental models: Cell lines
Expi293F cells	Thermo Fisher	A14635
Vero-E6 cell	ATCC	CRL-1586™
Vero-E6-TMPRSS2	BPS Bioscience	78081
HeLa-ACE2	BPS Bioscience	79958
293T-hsACE2	Integral Molecular	C-HA101
Oligonucleotides
ST3-F: 5’-cgcggatcccgtggcgtgccgcatattgtgatggtggatgcgtataaacggtacaagggtggaggtggctccatcgatac-3’	This paper	N/A
ST3-R: 5’-gaggcagctcagcaat-3’	This paper	N/A
NSC3-F: 5’-aagaatgcggccgcagggggccatattaaattctcaaaacgtgatgag-3’	This paper	N/A
NSC3-R: 5’-ccgctcgagcccgggcccctggaacagaacttccagcccgccagtatgagcgtcacctt-3’	This paper	N/A
Recombinant DNA
Human IgM-Fc	Addgene (Li et al., 2020)	158214
Human IgM J-chain	Addgene (Li et al., 2020)	158215
SpyCatcher003-mi3	Addgene (Rahikainen et al., 2021)	159995
Multiple Nbs and dimers in pET-21b(+) vector, multiple Nb-IgG-Fc in p3BNC vector	Synbio Technologies	N/A
Software and Algorithms
CryoSPARC	(Punjani et al., 2017)	https://cryosparc.com
ChimeraX	(Pettersen et al., 2021)	https://www.cgl.ucsf.edu/chimerax/
FreeSASA	(Mitternacht, 2016)	https://freesasa.github.io
CHARMM-GUI webserver	(Lee et al., 2016)	https://www.charmm-gui.org
MotionCor2	(Zheng et al., 2017)	N/A
eTomo	(Mastronarde and Held, 2017)	N/A
IsoNet	(Liu et al., 2022b)	N/A
Others
HisPur™ Cobalt Resin	Thermo fisher	89966
PROTEINDEXTM rProtein G Agarose 4 Fast Flow	Marvelgent Biosciences, Inc.	11-0206-025

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

The 293T-hsACE2 stable cell line was purchased from Integral Molecular in the freezing media (50% cell culture media, 40% FBS, 10% DMSO). Cells were cultured in a cell culture media (DMEM, 10%FBS, 10 mM HEPES, 1x Penicillin-Streptomycin and 0.5 μg/ml Puromycin) at 37°C, 5% (v/v) CO₂. Cells were split once 80–90% confluency had been reached. After at least three passages, cells were used for the pseudovirus neutralization assay at 80–90% confluency and more than 95% viability. The cell line was authenticated by Integral Molecular using flow cytometry by an ACE2-specific monoclonal antibody (R&D system, cat# MAB9332–100).

The Expi293F cells were obtained from Thermo Fisher. 1 ml of the cells was thawed at 37°C water bath and added to 29 ml of pre-warmed Expi293 Expression Media in a 125 ml non-baffled, vented flask. The cells were cultured at 37°C, 8% (v/v) CO₂ on an orbital shaker at 180 r.p.m. Cells were split when density reached 3 – 5 × 10⁶ cells/ml. At least three passages were performed, and more than 98% cell viability was ensured before the transfection. The authentication of the cell line was not performed in our hands.

All SARS-CoV-2 Victoria variant (SARS-CoV-2/human/AUS/VIC01/2020) cultures and infections were performed in the containment level 3 (CL3) lab at the Oxford Particle Imaging Centre (OPIC). 15 ml 6× 10⁶ Vero E6 cells were seeded into a T7 flask one day before infection. On the day of infection, the medium was replaced with 15 ml of DMEM (1% FBS with Glutamine supplement). 100 μl viruses at titer 10⁵ were seeded in the flask and incubated for 2–3 days. Cell debris was removed by centrifugation at 400 × g for 20 min at 4°C. The supernatant was aliquoted and stored in the −80°C freezer. The frozen virus was thawed, and the virus titer was tested by plaque assay using Vero E6 cells.

HeLa-ACE2 cells (BPS Bioscience, were maintained in DMEM (Corning) supplemented with 10% FBS, 0.5 μg/mL puromycin, and penicillin/streptomycin (Corning) at 37°C and 5% CO2.

All cell lines used in this study were regularly screened for mycoplasma contamination using the MycoStrip^™ - Mycoplasma Detection Kit. Cells were infected with SARS-CoV-2, isolate USA-WA1/2020, a previously described mouse-adapted SARS-CoV-2 strain (MA-SARS-CoV-2) and representative Beta (B.1.351), Omicron (BA.5), (XBB.1.16) and (XBB.1.5) SARS-CoV-2 variants. These variants were collected from nasopharyngeal swab specimens as part of the routine SARS-CoV-2 surveillance conducted by the Mount Sinai Pathogen Surveillance program (IRB approved, HS#13–00981). Viruses were grown in Vero-TMPRSS2 cells for 4–6 d; the supernatant was clarified by centrifugation at 4,000g for 5 min and aliquots were frozen at −80°C for long-term use. Expanded viral stocks were sequence-verified to be the identified SARS-CoV-2 variant and titered on Vero-TMPRSS2 cells before use in antiviral assays. Infections with viruses were performed under biosafety level 3 (BSL3) containment in accordance with the biosafety protocols developed by the Icahn School of Medicine at Mount Sinai.

All the antiviral animal studies were performed in an animal biosafety level 3 (BSL3) facility at the Icahn School of Medicine in Mount Sinai Hospital, New York City. All procedures performed on animals were following regulations and established guidelines and were reviewed and approved by an Institutional Animal Care and Use Committee (IACUC) or through an ethical review process. All work was conducted under protocols approved by the Mt. Sinai IACUC.

METHOD DETAILS

Nanobody (Nb) and Nb-dimer DNA synthesis and cloning

The cDNA encoding SpyCatcher003 was obtained from Addgene. Plasmids of the Nbs and Nb dimers were synthesized and cloned into the pET-21b(+) vector at EcoRI and HindIII sites from Synbio Technologies (Table S1). The Nb genes were codon-optimized and synthesized. To produce Nb/dimer-Ig_{ΔNSpyCatcher003}, the DNA fragment of ΔNSpyCatcher003 was first PCR amplified from the Addgene plasmid by using primers NSC3-F and NSC3-R to introduce a linker sequence and two restriction sites of XhoI and NotI that facilitate cloning. The PCR fragment was then inserted into the pET-21b(+) vector of different Nbs at the same restriction sites to produce the fusion constructs.

Purification of Nbs and Nb-dimers

Nb or Nb-dimer DNA constructs were transformed into BL21(DE3) cells and plated on Agar with 50 μg/mL ampicillin at 37°C overnight. Cells were cultured in an LB broth to reach an O.D. of ~0.6–0.8 before IPTG (0.5–1 mM) induction at 16°C overnight. Cells were then harvested, sonicated, and lysed on ice with a lysis buffer (1xDPBS, 150 mM NaCl, 0.2% TX-100 with protease inhibitor). After cell lysis, protein extracts were collected by centrifugation at 21,000 × g for 10 mins and the his-tagged Nbs were purified by the His-Cobalt resin and natively eluted with a buffer containing 150 mM imidazole buffer. Eluted Nbs were subsequently dialyzed in a dialysis buffer (e.g., 1x DPBS, pH 7.4 or 20mM Na-HEPES, 150mM NaCl, pH 7.45).

Production of recombinant Nb-IgG-Fc, IgM core, genetic fusion Nb-IgM-Fc, Nb-IgM-Fc conjugates and AMETA

The cDNA encoding human IgM-Fc and J-chain (his-tag) were obtained from Addgene. SpyTag003 was added to the N-term of IgM-Fc using primers ST3-F and ST3-R. Plasmids of the genetic fusion Nb-IgM-Fc and Nb-IgG-Fc were synthesized from Synbio Technologies. To express the proteins, Expi293F cells were transiently transfected with a mixture of IgM-Fc/Nb-IgM-Fc_genetic and J-chain (2:1 molar ratio) or Nb-IgG-Fc using the ExpiFectamine 293 kit. After 24 hrs of transfection, enhancers were added to further boost protein expression. Cell culture was harvested 3–4 days after transfection and the supernatant was collected by high-speed centrifugation at 21,000×g for 30 min. The secreted pentameric IgM or Nb-IgG-Fc in the supernatant were purified using His-Cobalt resin or protein G resin and natively eluted according to the manufacturer’s protocols. Eluted proteins were subsequently dialyzed in a dialysis buffer (e.g., 1x DPBS, pH 7.4 or 20mM Na-HEPES, 150mM NaCl, pH 7.45).

Nbs/dimers were then incubated with IgM core at a 30:1 molar ratio at 4°C overnight or 23°C for 2 hours. The final conjugates were further purified by 100kDa molecular weight cutter (for Nb-IgM-Fc) or size-exclusion chromatography using a Superose 6 Increase 10/300 (for AMETA) in SEC buffer (20 mM Hepes pH 7.5 and 150 mM NaCl).

Mass photometry of AMETA samples

Mass photometry experiments were carried out using a Refeyn TwoMP (Refeyn Ltd., Oxford, UK) MP system. AcquireMP and DiscoverMP software packages were used to record movies and analyze data, respectively, using standard settings. Ready-to-use sample carrier slides (Refeyn Ltd., Oxford, UK) were used for each measurement. Samples well cassettes (6 wells, Refeyn Ltd., Oxford, UK) were used to keep the sample in droplet shape. Contrast-to-mass calibration was carried out using SEC polished γ-globulins from human blood (Millipore-Sigma: G4386), giving molecular weights of 158, 316, 474, and 632 kDa. Each sample was first diluted in PBS to 100 nM at room temperature. Droplet dilution was initiated using 18 μL PBS. Upon focusing, 2 μL of diluted sample was added with quick mixing in the droplet. Movies were then recorded when all parameters were in blue.

The percentage of conserved RBD surface covered by Nbs

Relative solvent accessible surface area (rSASA) for each RBD residue is calculated by FreeSASA. Surface residue is defined using a cutoff of rSASA > 0.8. The conservation score of each RBD residue is calculated as described in (Xiang et al., 2022). An amino acid residue is defined as conserved if its conservation score is larger than 0.7. The percentage of conserved RBD surface covered by Nbs is calculated as the number of conserved residues on the epitope(s) of Nb(s) by the total number of conserved RBD surface residues.

Pseudotyped SARS-CoV-2 neutralization assay

The 293T-hsACE2 stable cell line and pseudotyped SARS-CoV-2 (Wuhan-Hu-1 strain D614G, Beta, Omicron variants and SARS-CoV) particles with luciferase reporters were purchased from the Integral Molecular. The neutralization assay was carried out according to the manufacturer’s protocols. In brief, 3- or 5-fold serially diluted Nbs/Nb-IgG-Fc/Nb dimer/Nb-IgM-Fc/AMETA was incubated with the pseudotyped SARS-CoV-2-luciferase for accurate measurements. At least seven concentrations were tested for each protein and at least two repeats were done. Pseudovirus alone in the culture media was used as a negative control. 100 μL of the mixtures were then incubated with 100 μL 293T-hsACE2 cells at 2.5×10e5 cells/mL in the 96-well plates. The infection took ~72 hrs at 37°C with 5% CO2. The luciferase signal was measured using the Renilla-Glo luciferase assay system with the luminometer at 1 ms integration time. The obtained relative luminescence signals (RLU) from the negative control wells were normalized and used to calculate the neutralization percentage at each concentration. Data was processed by Prism 9 (GraphPad) to fit into a 4PL curve and to calculate the EC₅₀s (half-maximal effective concentration).

Viral growth and neutralization assay

Four thousand HeLa-ACE2 cells were seeded into 96-well plates in DMEM (10% FBS) and incubated for 24 hours at 37°C, 5% CO2. Two hours before infection, the medium was replaced with 100 μL of DMEM (2% FBS) containing the AMETA proteins at concentrations 50% greater than those indicated. Plates were then transferred into the BSL3 facility and 100 PFU (MOI = 0.025) was added in 50 μL of DMEM (2% FBS), bringing the final protein concentration to those indicated. Plates were then incubated for 24 hours at 37°C. After infection, supernatants were removed and cells were fixed with 4% formaldehyde for 24 hours before being removed from the BSL3 facility. The cells were then immunostained for the viral N protein (an inhouse mAb 1C7, provided by Dr. Andrew Duty, Andrew.Duty@mssm.edu) with a DAPI counterstain. Infected cells (488 nm) and total cells (DAPI) were quantified using the Cytation 1 (Biotek) imaging cytometer. Infectivity was measured by the accumulation of viral N protein (fluorescence accumulation). Percent infection was quantified as ((Infected cells/Total cells) - Background) *100 and the PBS control was then set to 100% infection for analysis. Data was fit using nonlinear regression and IC₅₀s for each experiment were determined using GraphPad Prism version 10.0.0 (San Diego, CA). Cytotoxicity was also performed using the MTT assay (Roche), according to the manufacturer’s instructions. Cytotoxicity was performed in uninfected cells with the same protein dilutions and concurrent with viral replication assay. All assays were performed in biologically independent triplicates.

Preclinical evaluation of AMETA4 in the SARS-CoV-2 infection experiments

Animal studies with the Wuhan-like SARS-CoV-2/WA1 were performed using 12-week-old female S/129 (Jackson Laboratory strain 002448) mice. There are four groups in the experiment: non-infected control, isotype control, AMETA prophylaxis (−6 hrs), and AMETA therapeutic (+6 hrs) treatment. 8 mice were used per group. Mice were intranasally infected with 1 × 10⁴ PFU of SARS-CoV-2 in 50 μl of PBS. For prophylaxis evaluation, AMETA4 or IgM-Fc were administered intranasally (i.n.) at 2.2 nmole/kg (~ 2 mg/kg) 6 hours prior to infection. For therapeutic treatment evaluation, AMETA4 was administered i.n. 6 hours after infection. All dosing was performed through the intranasal (i.n.) route using PBS as a vehicle at a volume of 50 μl/mouse. Mice were weighed daily throughout the study as a measure of pathology. On day 3 post SARS-CoV-2 infection, animals were humanely euthanized, and lung tissues were harvested. Lungs were homogenized in PBS with silica glass beads and frozen at −80°C for viral titration by TCID₅₀. Briefly, infectious supernatants were collected at 48 hrs post-infection and frozen at −80°C until later use. Infectious titers were quantified by limiting dilution titration using Vero E6 cells. Briefly, Vero E6 cells were seeded in 96-well plates at 20,000 cells/well. The next day, SARS-CoV-2-containing supernatant was applied at serial 10-fold dilutions ranging from 10⁻¹ to 10⁻⁶ and, after 5 days, viral cytopathic effect (CPE) was detected by staining cell monolayers with crystal violet. TCID₅₀/ml was calculated using the method of Reed and Muench. The Prism software (GraphPad) was used to determine differences in lung titers using an unpaired T-test on log-transformed data.

Mouse lung histological analysis

Paraffin-embedded lung tissue blocks for mouse lungs were cut into 5μm sections. Sections were stained with hematoxylin and eosin (H&E) and analyzed by Histowiz (Brooklyn, NY). Digital light microscopic scans of the whole lung processed in toto were examined by an experienced veterinary pathologist. Hematoxylin Eosin stained sections of lungs were examined by implementing a semi-quantitative, 5-point grading scheme (0 - within normal limits, 1 - mild, 2 - moderate, 3 - marked, 4 - severe) that took into account four different histopathological parameters: 1) perivascular inflammation 2) bronchial or bronchiolar epithelial degeneration or necrosis 3) bronchial or bronchiolar inflammation and 4) alveolar inflammation. These changes were absent (grade 0) in the lungs of uninfected mice. All mice from AMETA and IgM-Fc treated infected groups exhibited multifocal pulmonary lesions.

Coupling AMETA with p-Bz-NCS-DFO

To a solution of AMETA in PBS (1.18 mg/mL, 1 mg) carbonate buffer 0.1 M was added to reach pH=8.3. A solution of deferoxamine (p-Bz-NCS-DFO) in DMSO (2.8 mL, 10 eq) was added. The reaction was stirred at 37 °C for 2 h. The reaction mixture was then purified (3 washes with PBS) and concentrated using 100k MWCO Vivaspin. DFO-AMETA was used for the radiolabeling without further purification.

Radiolabeling AMETA-DFO

A solution of ⁸⁹Zr in oxalic acid (20 mL, 3.3 mCi) was diluted in PBS (0.100 mL) and the solution was basified using a solution of Na₂CO₃ to pH=7.8. The solution of AMETA-DFO was added and the reaction was stirred for half an hour at 37 °C. The complete incorporation of ⁸⁹Zr was checked by radio TLC. The reaction was purified with 100k MWCO Vivaspin (3 washes with PBS). A radiochemical purity of >99% was reached and the radiolabeled protein was used for in vivo studies without further purification (2.6 mCi, RCY= 79%).

PET/CT imaging of AMETA biodistribution in mice

⁸⁹Zr-labeled AMETA (275.5 ± 43.5 mCi) was intranasal injected in C57BL/6 mice. Two, fourteen, twenty-four, and forty-eight hours later, mice were anesthetized using 1.0% isoflurane in O₂ at a flow rate of ~1.0 liter/min. PET/CT scans were performed using a Mediso nanoScan PET/CT (Mediso, Budapest, Hungary). A whole-body CT scan was performed (energy, 50 kVp; current, 180 mAs; isotropic voxel size, 0.25 mm) followed by a 20-min PET scan. Reconstruction was performed with attenuation correction using the TeraTomo 3D reconstruction algorithm from the Mediso Nucline software. The coincidences were filtered with an energy window between 400 and 600 keV. The voxel size was isotropic with 0.4-mm width, and the reconstruction was applied for four full iterations, six subsets per iteration.

Biodistribution studies

C57BL/6 mice were sacrificed 72 hours after injection of ⁸⁹Zr-labeled AMETA and perfused with PBS (20 ml). Tissues of interest were harvested, weighed, and gamma-counted using a Wizard² 2480 automatic gamma counter (PerkinElmer, Waltham, MA). Values were corrected for decay and expressed as a percentage of the injected dose per gram of tissue (%ID/g).

Cryo-EM sample preparation and imaging

3.5 μl of 0.6 mg/mL AMETA were applied to ultrafoild R1.2/1.3 grids, and grids were blotted for 2.5 s with 0 blot force in the environmental chamber of a Vitrobot Mark IV (ThermoFisher) at 4 °C and 100% humidity. Blotted grids were plunged frozen in liquid ethane. Cryo grids were screened on a Glacios microscope (FEI/Thermo) at 200 kV and data collection was performed on a Titan Krios microscope (FEI/Thermo) at 300 kV using serialEM at PNCC. Movies were recorded on a K3 camera in super-resolution mode with a total dose of 38 electrons per Ų in a moving stack of 68 frames (dose rate 1 electron/pixel/frame) with a pixel size of 1.336 Å/pixel and a defocus range between −1 to −3 μm.

Cryo-EM data processing and model-building

Image processing was performed using CryoSPARC 4.3.1. The particles were automatically picked using the blob picker with a 300 Å diameter. The first round of reference-free 2D classification was performed with 200 classes and limited maximum resolution to 18 Å. Particles resembling IgM were selected to generate the initial volume using ab initio reconstruction. 3D refinement was first carried out using non-uniform refinement using ab initio volume as the reference without mask. A box size of 720 pixels was used to re-extract the IgM-like particle to visualize the AMETA-captured spike protein.

Models of IgM (PDB entry: 6KXS) were rigidly docked into the cryo-EM map and refined using res-space refinement in Phenix. The maps and models were displayed in ChimeraX. Details on data recording and processing of the IgM scaffold are summarized in Table S5.

Integrative Structure Modeling of AMETA

The structure model of AMETA3 was constructed using Modeller (Version: 10.3) and Alphafold2. The AMETA construct is composed of a fixed core region (five protomers of IgM-Fc homodimeric SpyTag003-Cμ2-Cμ3-Cμ4, one J-chain) and ten versatile warhead regions (Nb dimer-Ig_{ΔNSpyCatcher003}). The Cryo-EM structure of the IgM-Fc pentamer core, i.e Cμ3-Cμ4 and J-chain (PDB:8AE3) was used as a template for AMETA. The missing residues in hinge regions between Cμ3 and Cμ4 were fixed by Modeller. Missing structures in the J-chain were fixed by using Alphafold2. The structure of Cμ2 and SpyTag003 was predicted by Alphafold and grafted to the IgM core template with the restraint of one disulfide bond in the hinge region between Cμ2 and Cμ3 domain. The fusion protein of two Nb dimers: 1) S36–182-Ig_{ΔNSpyCatcher003} and 2) 132–118-Ig_{ΔNSpyCatcher003} was constructed by Modeller with the templates of Alphafold2 prediction of NbS36, RBD-182 complex structure (PDB:8CYJ), RBD-132 complex structure (PDB:8CYD), RBD-118 complex (PDB:8CWV) and SpyCatcher-SpyTag structure (PDB:8MLI). Ten Nb dimers were then loaded to the AMETA3 core guided by the SpyCatcher-SpyTag interaction.

Molecular Dynamics Simulation of one protomer of AMETA3

Amber 20.0(Case et al., 2005) was used to perform MD simulations. All input files for MD simulations of one protomer of AMETA3 and parameter files were generated using CHARMM-GUI server(Lee et al., 2016). The water-boxed AMETA3 protomer (including 450674 water molecules, 1283 chloride ions, and 1307 potassium ions) were simulated with CHARMM forcefield at 310K temperature with a step size of 2 femtoseconds using one A100 enabled node with 8 processors for a period of 50 ns, after minimization and equilibrations. During MD runs, the cut-off of 12 angstroms was used for Van der Waals and short-range electrostatic interactions, along with PME conditions. After simulation, water molecules and ions were removed from the trajectory. Frames were aligned to the one of the Cμ2-Cμ3-Cμ4 chain in the first frame. The root-mean-square-fluctuations (RMSF) was calculated by pytraj python package(Roe and Cheatham, 2013).

Analysis of Epitope Accessibility in RBD Opening Dynamics

The MD simulation of RBD opening dynamics was obtained from (https://covid.molssi.org/simulations/#pmf-calculations-of-sars-cov-2-spike-opening)(Pang et al., 2022). The trajectory without glycans was used for the analysis of Nb epitope accessibility. Nb epitope is defined by RBD residues within 8 angstrom Ca-distance to the Nb. Solvent accessible surface area (SASA) for each amino acid in each frame was calculated by the tool FreeSASA (Version 2.1.2)(Mitternacht, 2016). The epitope SASA was then calculated by summation of SASAs of all epitope residues. To smooth the curve, data points for each Nb were fitted to the quadratic function. Curves of different Nbs in the same epitope were averaged in Figure 1H.

Cryo-ET sample preparation and vitrification

EM grids (G300F1, R2/2 Quantifoil holey carbon, gold) were glow-discharged and placed in the wells of a 12-well plate. The grids were treated with bovine fibronectin (20 μg/ml) for 30 minutes then washed with PBS and UV-treated for 30 minutes. 6 × 10⁴ of Vero E6 cells resuspended in 1 ml of complete DMEM medium were seeded on top of the grids in each well of the 12-well plate, incubated for 24h at 37°C /5% CO2 to allow cell attachment to grid carbon. The medium was moved to the BL3 facility before being replaced with the SARS-CoV-2 Victoria variant diluted to MOI 0.5 in DMEM (1% FBS with Glutamine supplement). The cells were incubated at 37°C/5% CO₂ for 24h in the CL3 facility.

The AMETA or IgM-Fc were warmed up at 37°C and diluted to the desired concentration in warm PBS. The 24 well plates with grids of infected cells were washed with warm PBS once. 1 ml of diluted AMETA was added to each well, and returned to 37°C/5% CO₂ for 2h. PBS was used as a blank control. After incubation, the grids were washed with warm PBS and fixed with 4% PFA (EM grade) PBS solution for 1h at RT before being removed from the CL3 facility.

Grids were washed with PBS. 2 μl of PBS with 6 nm Au fiducial beads (EMS) in the carbon side and 1 μl from the Au side were added to the grids before blotting. The grids were blotted from the back for 5 seconds and plunge-frozen in liquid ethane at −183°C using the Leica GP2 plunger. Grids were clipped and stored in liquid nitrogen until imaging.

Cryo-ET data collection and processing

Tilt series of SARS-CoV-2 virus with AMETA or IgM-Fc bound were collected with Titan Krios equipped with a Selectris X energy filter and Falcon 4 detector, or a Gatan BioQuantum energy filter and K3 detector. The tilt angles are from −60 degrees to 60 degrees with a step of 3 degrees with a dose-symmetric scheme (Hagen et al., 2017). The pixel size is 1.50 A/pixel (falcon camera) or 1.34 A/pixel (K3 camera). Each image contains 10 movie frames. Periphery of the cells that contain the freshly egressed SARS-CoV-2 virus particles were targeted for cryo-electron tomography data collection under each experimental condition (See supplementary table). Raw movie frames were motion corrected by MotionCor2 (Zheng et al., 2017). The tilt-series were aligned by eTomo (Mastronarde and Held, 2017) based on the tracking of gold beads fiducial markers. IsoNet neural networks (Liu et al., 2022b) were trained and then used to correct the missing wedge artifact in the reconstructed tomograms. Virus morphology and spike number and conformation were analyzed. Segmentation is processed by manually docking spikes (emd-11651)(Tegunov et al., 2021) and AMETA (from this study) to corresponding areas (Figure 6E) or automatically segmented by EMAN2 (Chen et al., 2017) after training the neural network (Figure 6F).

QUANTIFICATION AND STATISTICAL ANALYSIS

GraphPad Prism (version 9.3.0) was used for all statistical calculations. Welch’s unpaired two- tailed t test was performed between groups. For statistical difference analysis, p values less than 0.05 were considered statistically significant. ns: no significant difference; *: p < 0.05; **: p < 0.01; ***: p < 0.001; ****: p < 0.0001.