Programmable antibody-based chimeric entry receptors for sarbecoviruses

Abstract

Despite frequent spillover of sarbecoviruses, most SARS-related viruses discovered in animals fail to engage human ACE2 (hACE2), limiting mechanistic insight and risk assessment. Here we developed antibody-based chimeric entry receptors (ABCERs) that reprogram antibody-antigen recognition into a synthetic, cell-anchored receptor interface. By replacing the extracellular protease domain of hACE2 with single-chain variable fragments (scFvs) from broadly neutralizing antibodies, ABCERs mimic viral receptor engagement while preserving the intracellular architecture required for cathepsin L-dependent endocytic fusion. This modular design converts antibody specificity into a programmable entry module, supporting efficient infection and replication of diverse sarbecoviruses from both clinical and animal sources. Among the tested scFvs, E7 exhibited exceptional breadth, recognizing conserved epitopes shared across representative sarbecoviruses from all clades. Sera from Pfizer-BioNTech mRNA-vaccinated individuals potently blocked E7 binding to SARS-CoV-2 but showed limited cross-inhibition of E7 interactions with RBDs from hACE2-independent sarbecoviruses, revealing a substantial gap in current vaccine-induced humoral immunity. Together, our findings establish E7-based ABCERs as a programmable synthetic receptor platform that bridges antibody recognition and viral propagation, offering a universal tool for isolating, studying, and surveying sarbecoviruses beyond the hACE2-dependent paradigm.

Engineered antibody-based chimeric receptors enable entry and propagation of diverse sarbecoviruses, providing a versatile platform to assess spike compatibility across clades

Introduction

The repeated emergence of highly pathogenic coronaviruses such as SARS-CoV and SARS-CoV-2 underscores the continual risk of zoonotic spillover from animal origin to humans¹. Both viruses are believed to have originated from bats, with spillover events likely involving intermediate hosts such as palm civets for SARS-CoV and possibly pangolins for SARS-CoV-2². Serological evidence from communities near bat caves further suggests that direct bat-to-human transmission may occur more frequently than previously recognized^3–5. Each such event provides an evolutionary opportunity for viral adaptation and cross-species transmission, underscoring the importance of understanding receptor usage among pre-spillover sarbecoviruses.

The coronavirus spike glycoprotein dictates host range and tissue tropism by mediating receptor recognition and membrane fusion⁶. Structurally, the spike comprises an S1 subunit containing the receptor-binding domain (RBD) and an S2 subunit mediating membrane fusion¹. Among sarbecoviruses, RBD-based phylogeny defines receptor usage and divides them into four established clades: clade 1a (SARS-CoV-related), clade 1b (SARS-CoV-2-related), clade 2 (ACE2-independent), and clade 3 (European/African)⁷. While human ACE2 (hACE2) functions as the canonical receptor for clade 1a and 1b, comparative genomics and binding assays reveal that many sarbecoviruses, particularly those in clades 2 and 3, do not interact with hACE2⁸. These viruses with unknown receptors remain largely uncharacterized because they cannot be propagated in conventional hACE2-dependent culture systems, leaving their entry mechanisms and zoonotic potential poorly understood.

The COVID-19 pandemic led to the discovery of numerous broadly neutralizing antibodies (bNAbs) that recognize conserved epitopes across divergent sarbecoviruses, including those unable to bind hACE2⁹. These antibodies have revealed that the spike possesses alternative binding interfaces that could, in principle, substitute for canonical receptor interactions. Following receptor attachment, viral entry proceeds via either TMPRSS2-mediated plasma membrane fusion¹⁰ or cathepsin L-dependent endocytosis¹¹, the latter providing an alternative route in cells lacking surface proteases. This mechanistic redundancy implies that engineered membrane receptors capable of binding spike and engaging the endocytic pathway could enable controlled viral entry independent of native receptors.

Recent advances in receptor engineering have established conceptual foundations for reconstructing viral entry through modular design^12–14. Early proof-of-principle studies showed that chimeric receptors combining antibody fragments with hACE2 could act as surrogate entry factors for SARS-CoV-2¹⁵, and that membrane-tethered hACE2 protease domain (PD) could function as a “cell-anchored soluble receptor” mediating spike-triggered fusion¹⁶. Building on this concept, customized viral receptors (CVRs) were developed by integrating artificial binding modules into transmembrane scaffolds to expand receptor flexibility¹⁷. Despite these advances, replication of most ACE2-independent sarbecoviruses has still relied on the trypsin-supplemented conditions¹⁸, underscoring the lack of a generalizable receptor system for their propagation. Moreover, most engineered receptors fail to reproduce the broad receptor compatibility observed among ACE2-dependent sarbecoviruses (clade 1a and 1b), nor do they provide a framework to explain how distinct binding modules and membrane architectures cooperate to mediate cross-lineage viral entry.

Here, we extend this concept by unifying antibody specificity and endocytic functionality within a single synthetic receptor framework, antibody-based chimeric entry receptors (ABCERs). In this design, the extracellular PD of hACE2 is replaced with single-chain variable fragments (scFvs) derived from pan-sarbecovirus antibodies, while the hACE2 neck, transmembrane, and cytoplasmic domains are retained to preserve receptor orientation and endocytic trafficking. This architecture integrates the precision of antibody recognition with the cellular machinery required for cathepsin L-dependent fusion, creating a programmable platform for receptor-mediated viral entry across all sarbecoviruses clades, advancing pandemic preparedness.

Results

Design and optimization of the synthetic antibody-based chimeric entry receptors

Broadly neutralizing antibodies (bNAbs) that cross-react with diverse sarbecoviruses have become indispensable for anticipating and countering future coronavirus outbreaks⁹. Here, we harnessed the broad reactivity of bNAbs to overcome the constraints of natural receptor tropism by re-engineering hACE2 into a synthetic antibody-based chimeric entry receptor (ABCER). The resulting ABCERs were designed to mediate viral entry across both hACE2-dependent (clade 1a and 1b) and hACE2-independent (clade 2 and 3) sarbecoviruses.

In this study, we used four classes of pan-sarbecovirus monoclonal antibodies for the ABCER design (Fig. 1a). The extracellular protease domain (PD) of hACE2 was replaced with single-chain variable fragments (scFvs) derived from representative bNAbs spanning major epitope classes: S2K146¹⁹ (class I), E7²⁰ (class I/II/IV), S309²¹ (class III), and CR3022²² (class IV), as well as the S2-targeting antibody S2P6²³ (Fig. 1a; Tables S1 and S2).

Fig. 1. Design and characterization of synthetic antibody-based chimeric entry receptors (ABCERs).

a Structural model of the SARS-CoV-2 spike trimer in the RBD-up conformation overlaid with representative pan-sarbecovirus single-chain variable fragments (scFvs). Antibody epitopes are shown as colored surface patches, and the human ACE2 (hACE2) binding footprint is outlined in black. Arrows indicate binding classes of neutralizing antibodies targeting the receptor-binding domain (RBD): class I, class I/II/IV, class III, and class IV. A schematic of the spike subunit architecture is shown below, with annotated domains—N-terminal domain (NTD), RBD, subdomains 1/2 (SD1, SD2), heptad repeats (HR1, HR2), central helix (CH), connector domain (CD), stem helix (SH), fusion peptide (FP), transmembrane domain (TMD), and cytoplasmic tail (CT); the S1/S2 and S2′ cleavage sites are indicated by arrows. b Domain architecture of ABCERs. In native hACE2 (UniProt Q9BYF1), the ectodomain comprises the protease domain (PD; residues 18–615) and neck domain (ND; 616–740), followed by the TMD (741–761) and CT (762–805). In ABCER constructs, the PD is replaced by spike-binding scFvs while retaining the ND, TMD, and CT to preserve membrane orientation and trafficking. Three variants—ΔECD1, ΔECD2, and ΔECD3—were generated by progressively shortening the ND segment. c Schematic illustrating ABCER-mediated viral attachment and entry. When expressed on the plasma membrane, ABCERs engage the spike via antibody-derived recognition and trigger endocytic entry independent of native hACE2 tropism. d Pseudovirus entry into HEK293T cells expressing indicated receptors. HEK293T cells were transiently transfected with the indicated constructs using FuGENE® 6 (Promega, E2692), then infected with the indicated pseudoviruses. Luciferase signals (RLUs) are normalized to the control (HEK293T cells transfected with an irrelevant plasmid and infected with a matched dose of pseudovirus). Colored circles indicate the spike epitope classes targeted by the parental antibodies: class I (green), class II (magenta), class III (orange), and class IV (purple). Data in d are shown as a heatmap; the color scale represents the mean normalized luciferase signal from n = 3 independent experiments.

To preserve correct membrane topology and trafficking, the neck (ND), transmembrane (TMD), and cytoplasmic tail (CT) domains of hACE2 were retained, while the PD was substituted with spike-binding scFvs to generate receptors of defined architecture. Three structural variants, ΔECD1, ΔECD2, and ΔECD3, were engineered with progressively shortened ND segments to fine-tune the spatial geometry between the antibody module and membrane anchor (Fig. 1b). When expressed on the cell surface, ABCERs engage viral spikes through antibody-mediated recognition and trigger entry independent of native receptor usage (Fig. 1c).

This modular framework yielded eleven ABCERs based on hACE2ΔECD1-construct, incorporating both S1- and S2-directed epitopes^{19,20,22–29} (Fig. 1d). While all constructs were robustly expressed, protein levels varied despite identical transfection conditions (Fig. S1). Notably, highly expressed variants like CR3022 and S2P6 still failed to support entry, indicating that receptor specificity, rather than density, is the primary driver of efficiency (Fig. 1d). In spike-pseudotyped virus entry assays, HEK293T cells transiently expressing ABCERs supported measurable entry of sarbecoviruses from clades 1a (SARS-CoV, WIV1) and 1b (SARS-CoV-2, XBB.2.3, GX-P5L, RaTG13) (Fig. 1d). Two constructs, hACE2ΔECD1-CR3022 and hACE2ΔECD1-S2P6, showed minimal activity (Fig 1d), consistent with the inaccessibility of the CR3022 epitope in prefusion spike^22,30, and the fusion-blocking activity of S2P6²³. These data confirm that productive receptor engagement must trigger the prefusion-to-postfusion transition of spike, a prerequisite for viral entry. Among all constructs, hACE ΔECD1-E7 exhibited the broadest tropism, mediating efficient entry of sarbecoviruses across clade 1 and outperforming native hACE2 in several cases (Fig. 1d).

To further optimize receptor performance, we progressively shortened the spacer linking the scFv to the membrane, from ΔECD1 (125 aa) to ΔECD2 (14 aa) and finally ΔECD3 (direct fusion), guided by evidence that membrane proximity critically determines receptor potency¹⁶ (Fig. 1d). The intermediate ΔECD2 variant achieved the highest pseudovirus entry across all tested clades, whereas complete spacer removal (ΔECD3) reduced activity (Fig. 1d). Together, these results establish a synthetic receptor architecture that enable entry of SARS-like pseudoviruses into the cells. This ABCER platform exhibits programmability through its modular architecture, allowing for systematic tuning of receptor specificity via scFv substitution and entry efficiency through spacer length optimization.

Functional characterization of synthetic antibody-based chimeric entry receptors

To validate the functionality of the optimized synthetic receptor, stable HEK293T cell lines expressing E7-based ABCERs (hACE2ΔECD1/2/3-E7 variants) were established via lentiviral transduction and blasticidin selection. Flow cytometry and fluorescence imaging confirmed robust surface expression of all constructs (Fig. S2). Notably, the polyclonal E7-based ABCERs stable cell lines exhibited a broader distribution in RBD binding compared to the hACE2 population (Fig. S2b). Nevertheless, the overall binding levels and surface expression across the stable cell lines remained robust and sufficient for evaluating viral engagement (Fig. S2c, d). These results indicate that hACE2ΔECD1/2/3-E7 variants maintain proper membrane localization and high-affinity spike interaction, providing a functional platform to dissect antibody-driven viral entry.

We next examined whether these synthetic receptors could mediate spike-induced membrane fusion using a split-GFP cell-cell fusion assay³¹. Co-culture of spike-expressing (GFP1-10) and receptor-expressing (GFP11) cells reconstituted fluorescence, demonstrating successful membrane fusion (Fig. S3). Both hACE2ΔECD1-E7 and hACE2 ΔECD2-E7 induced robust fusion with spike proteins from clade 1a (SARS-CoV and WIV1) and clade 1b (SARS-CoV-2 and GX-P5L), reaching levels comparable to hACE2, whereas fusion with RaTG13 spike was less efficient (Fig. S3b, c). Notably, both hACE2ΔECD1-E7 and hACE2 ΔECD2-E7 supported substantially higher fusion efficiency with RaTG13 spike than hACE2, underscoring that E7-mediated recognition can overcome lineage-specific entry barriers (Fig. S3b, c).

In pseudovirus infection assays, hACE2ΔECD1-E7 and hACE2ΔECD2-E7 supported dose-dependent entry of sarbecoviruses from clade 1a and 1b (Fig. 2). hACE2ΔECD2-E7 consistently achieved the highest entry efficiency, exceeding hACE2 for SARS-CoV (Fig. 2b). Notably, our data demonstrated that hACE2ΔECD1/2-E7, but not hACE2, supports RaTG13 pseudovirus entry (Fig. 2f). In contrast, hACE2ΔECD3-E7 exhibited minimal activity, confirming that removal of the neck domain disrupts receptor functionality. Because absolute VSV*ΔG titers inherently vary between different spike pseudotypes due to differences in particle production and infectivity, these values should not be used as a direct comparative measure of receptor compatibility. Taken together, these data demonstrate that E7-based ABCERs could mimic hACE2 function while expanding viral tropism beyond canonical hACE2 usage.

Fig. 2. Functional characterization of synthetic antibody-based chimeric entry receptors (ABCERs) in HEK293T cells.

a Schematic of pseudovirus infection assays using ABCER-expressing HEK293T cells challenged with sarbecovirus spikes from clades 1a and 1b. b-f, Dose-dependent pseudovirus entry mediated by hACE2ΔECD1-E7, hACE2ΔECD2-E7, or hACE2ΔECD3-E7. Luminescence was quantified following infection with VSV*ΔG-luciferase pseudoviruses bearing spikes from SARS-CoV (b, clade 1a), WIV1 (c, clade 1a), SARS-CoV-2 (d, clade 1b), GX-P5L (e, clade 1b), and RaTG13 (f, clade 1b). Data in b–f represent mean ± SD from three independent experiments (n = 3).

Isolation of authentic SARS-related coronaviruses using ABCERs

Isolation of authentic sarbecoviruses from clinical and animal sources is essential for understanding viral transmissibility and pathogenesis. Having established that antibody-based chimeric entry receptors (ABCERs) enable efficient pseudovirus infection across clades 1a and 1b, we next tested their ability to mediate isolation and propagation of live viruses (Fig. 3a). Nasopharyngeal swabs positive for SARS-CoV-2 variants XBB (n = 1) and JN.1 (n = 3) were inoculated onto hACE2-, hACE2ΔECD1-E7, and hACE2ΔECD2-E7-expresing HEK293T cells. Engineered hACE2ΔECD1-E7-293T cells supported robust viral replication, exhibiting marked increases in viral genomic RNA and infectious titers, whereas hACE2ΔECD2-E7-293T cells showed only minimal replication relative to hACE2-expressing HEK293T cells (Fig. 3b, c).

Fig. 3. ABCER-mediated entry and replication of clinical SARS-CoV-2 isolates and zoonotic-potential sarbecoviruses.

a Workflow for virus isolation using ABCER-expressing HEK293T cells. Nasopharyngeal swabs or SARS-related coronavirus samples were inoculated onto engineered cells in virus transport medium (VTM). Following ABCER-mediated endocytic entry, viruses underwent replication, assembly, and egress. Viral replication was quantified by qPCR and infectious titers by TCID50. b, c Replication of clinical SARS-CoV-2 variants in ABCER-engineered cell lines. Parental HEK293T, hACE2 (positive control), hACE2ΔECD1-E7, and hACE2ΔECD2-E7 cells were infected with SARS-CoV-2 XBB (n = 1) or JN.1 (n = 3). Viral RNA in culture supernatants was measured by qPCR (b), and infectious titers were determined by TCID₅₀ (c) at 7 days post-inoculation (d.p.i.). d, e ABCERs enable productive infection by zoonotic-potential sarbecoviruses. The same panel of cell lines was infected with recombinant SARS-related viruses, including rSARS-CoV, rWIV1, rRsSHC014, rSARS-CoV-2, rPgCoV-GD1, and rBANAL-20-236. Viral RNA in supernatants was quantified at 1, 2, and 3 d.p.i. (d), and corresponding infectious titers were determined by TCID₅₀ (e). Data represent mean ± SD (n = 3 independent experiments). Statistical significance was assessed using two-tailed Student’s t-tests; n.s., not significant. For each time point, p-values (top to bottom) compare hACE2, hACE2ΔECD1-E7, and hACE2ΔECD2-E7 to HEK293T. Identical p-values are shown once. Exact p-values are indicated in the figure.

We further extended infection assays to sarbecoviruses with zoonotic potential, including recombinant SARS-CoV³², WIV1³³, RsSCH014^33,34, SARS-CoV-2³⁵, PgCoV-GD1³⁶ and BANAL-20-236³⁷ (Fig. 3d,e). Both hACE2ΔECD1-E7 and hACE2ΔECD2-E7 cells readily supported productive infection by all six viruses tested. Quantitative RT-PCR (Fig. 3d) and TCID50 analyses (Fig. 3e) confirmed sustained replication across multiple cycles without the need for exogenous proteases, demonstrating that antibody-mediated receptor replacement alone is sufficient to support authentic viral propagation. Replication kinetics in hACE2ΔECD1-E7 cells were comparable to those in hACE2 cells, indicating that antibody-driven receptor engagement preserves intrinsic infection dynamics of sarbecoviruses.

Across all infections, hACE2ΔECD1-E7 consistently outperformed hACE2ΔECD2-E7, in contrast to pseudovirus entry assays (Fig. 2b–f), highlighting the limitations of pseudotyped systems in recapitulating authentic sarbecovirus infection. The observed shift in optimal ABCER configuration between pseudotyped and authentic viruses likely reflects differences in viral particle morphology and spike density/distribution. Collectively, these findings establish E7-based ABCERs as a broadly compatible and physiologically relevant receptor platform for isolating and propagating both clinical and zoonotic sarbecoviruses.

Binding breadth of E7-mediated ABCER across sarbecovirus RBDs

We next investigated whether E7 can recognize and facilitate entry of hACE2-independent sarbecoviruses from clades 2 and 3 (Fig. 4a). Surface plasmon resonance (SPR) analysis revealed high-affinity interactions between E7 and a broad spectrum of sarbecovirus RBDs, encompassing representatives from both clades (Fig. 4b–g). The equilibrium dissociation constants (KD) indicated that E7 binds strongly not only to the representatives from clade 2 RBDs, ZC45 (KD = 2.80 × 10^-7 M), JTMC15 (KD = 1.15 × 10^-10 M), and Rf1 (KD = 4.15 × 10^-8 M), but also to the more divergent clade 3 viruses Khosta-2 (KD = 1.35 × 10^-8 M), BM48-31 (KD = 1.74 × 10^-11 M) and BtKY72 (KD = 1.46 × 10^-11 M). Indirect ELISA further confirmed robust binding of E7 to RBDs from both clades (Fig. S4). Collectively, these results indicate that E7 exhibits exceptional breadth and affinity across phylogenetically distant sarbecoviruses, providing a mechanistic basis for its function as a universal recognition module within the ABCER framework.

Fig. 4. Broad sarbecovirus entry mediated by antibody-based chimeric entry receptors (ABCERs).

a Phylogenetic tree of representative sarbecovirus spike proteins. Clades are color-coded: clade 1a (black), clade 1b (gray), clade 2 (red), and clade 3 (blue). Scale bar indicates amino-acid substitutions per site. b–g Surface plasmon resonance (SPR) analysis of E7 binding to sarbecovirus RBDs from clades 2 and 3. Fc-tagged E7 was immobilized on a Protein A sensor chip, except for JTMC15. Serial dilutions of RBD proteins were injected as analytes: ZC45 RBD (clade 2) (b), Rf1 RBD (clade 2) (d), Khosta-2 RBD (clade 3) (e), BM48-31 RBD (clade 3) (f), and BtKY72 RBD (clade 3) (g). For JTMC15 (clade 2) (e), His-tagged JTMC15 RBD was captured on an NTA sensor chip, and serial dilutions of E7 (0-6.5 nM) were used as analytes. Representative sensorgrams corresponding to different analyte concentrations (0–25/50/100 nM) are shown, with global fits to a 1:1 Langmuir binding model. The equilibrium dissociation constants (KD) are indicated. h-m, Dose-dependent entry of pseudotyped sarbecoviruses from clade 2 and clade 3 mediated by ABCERs. Luminescence was measured in HEK293T cells stably expressing hACE2ΔECD1-, hACE2ΔECD2- or hACE2ΔECD3-E7 following infection with serial dilutions of VSV*ΔG luciferase pseudoviruses: ZC45 (clade 2) (h), JTMC15 (clade 2) (i), Rf1 (clade 2) (j), Khosta-2 (clade 3) (k), BM48-31 (clade 3) (l), and BtKY72 (clade 3) (m). Data in h-m represent mean ± SD from n = 3 independent experiments.

We next evaluated whether E7-based ABCERs could substitute for hACE2 to mediate entry of these hACE2-independent sarbecoviruses. HEK293T cells expressing hACE2ΔECD1-E7 supported dose-dependent entry of VSV*ΔG luciferase pseudoviruses bearing spikes from ZC45 (clade 2; h), JTMC15 (clade 2; i), Rf1 (clade 2; j), Khosta-2 (clade 3; k), BM48-31 (clade 3; l), and BtKY72 (clade 3; m), even in the absence of exogenous protease (Fig. 4h–m). Remarkably, hACE2ΔECD1-E7 outperformed both hACE2ΔECD2-E7 and hACE2, underscoring the ability of antibody-derived modules to bypass canonical receptor constraints. Interestingly, hACE2ΔECD3-E7, which lacks the neck domain, exhibited selective susceptibility to ZC45, indicating that receptor geometry modulates lineage-specific entry. When benchmarked against the recently reported Nb27-CVR platform¹⁷, hACE2ΔECD1-E7 mediated substantially higher infection and retained functionality toward multiple lineages, including Rf1 (Fig. S5).

Together, these findings establish E7-based ABCERs as a synthetic receptor system with exceptional binding breadth and functional adaptability, enabling entry of both hACE2-dependent and hACE2-independent sarbecoviruses.

Limited blocking of E7-binding to clade 2 and 3 sarbecovirus RBDs by sera from SARS-CoV-2 vaccinated individuals

To determine whether current vaccine-induced antibodies can interfere with E7 recognition of divergent sarbecoviruses, we established an E7-RBD competitive blocking assay as a surrogate for evaluating antibody breadth in human sera³⁸ (Fig. 5a). Sera from individuals vaccinated with current SARS-CoV-2 vaccines efficiently blocked E7 binding to the SARS-CoV-2 RBD (inhibition ratio = 97.8 ± 0.6%) but showed minimal cross-inhibition against RBDs from clade 2 and 3 sarbecoviruses (inhibition ratio ≤26.4%) (Fig. 5b and Fig. S6). Our data indicate limited competition by sera from vaccinated individuals against E7 binding to clade 2/3 sarbecovirus spikes under the conditions tested, likely reflecting weak or low-titer responses to the E7-epitope.

Fig. 5. Limited vaccine-induced neutralizing antibody protection against hACE2-independent sarbecoviruses.

a Schematic of the competition ELISA using E7 and sarbecovirus RBDs. Human sera from healthy donors (pre-pandemic) or Pfizer-BioNTech mRNA-vaccinated individuals were incubated with His-tagged RBDs, followed by exposure to E7 pre-coated on ELISA plates. Neutralizing antibodies that blocked the interaction between RBDs and E7 reduced the horseradish peroxidase (HRP) signal after TMB substrate addition. b Human serum reactivity against sarbecovirus RBDs. Individual sera from pre-pandemic donors (n = 3) or Pfizer–BioNTech mRNA-vaccinated individuals (n = 7) were diluted 20-fold and tested in the competition ELISA against the indicated RBDs. Data in (b) are shown as mean ± SD, derived from individual serum samples (dot plot; n = 3 for pre-pandemic donors and n = 7 for Pfizer–BioNTech mRNA-vaccinated individuals). Statistical significance of vaccinated versus pre-pandemic sera was assessed for each RBD using a two-sided Wilcoxon rank-sum (Mann–Whitney U) test; exact p-values are indicated in the figure.

Together, these findings indicate that humoral immunity elicited by current SARS-CoV-2 vaccines provides limited cross-protective antibodies against non-hACE2-using sarbecoviruses, while E7 itself exhibits a rare cross-clade binding capacity across all major sarbecovirus lineages.

Discussion

Our study establishes a programmable synthetic receptor framework that redefines how antibody recognition can be repurposed to mediate sarbecovirus entry. By integrating the bNAb E7 into the extracellular domain of hACE2, we generated antibody-based chimeric entry receptors (ABCERs) capable of engaging sarbecoviruses across multiple evolutionary clades. This design extends beyond previous receptor-engineering efforts that reconstructed hACE2-like or antibody-anchored entry systems for SARS-CoV-2 alone, offering a unified mechanism that directly links antibody-defined recognition to functional viral entry, even among hACE2-independent lineages.

Our binding and infection data demonstrate that E7 targets a conserved epitope shared among sarbecoviruses spanning clades 1–3 (Figs. 2–4). The affinities observed for clade 2 and 3 RBDs reveal that E7 recognizes a cross-clade conversed epitope distinct from the canonical ACE2-binding site, setting it apart from most known neutralizing antibodies^9,20. Building on our previous findings that E7 neutralizes clade 1a and 1b sarbecoviruses, the current results further reveal its capacity to engage RBDs from clade 2 and 3 viruses, underscoring its potential as an antibody bridging hACE2-dependent and hACE2-independent sarbecoviruses.

Mechanistically, ABCER-mediated entry proceeds via a clathrin-dependent endocytic route, as shown by strong inhibition with clathrin inhibitor chlorpromazine (IC50 = 2.86 μM), endosomal acidification inhibitor chloroquine (IC50 = 2.88 μM) and the cathepsin L inhibitor E64d (IC50 = 0.065 μM), but not by the caveolae pathway inhibitor Filipin (Figs. S7 and S8). These results confirm that viral particles bound by hACE2△ECD1-E7 are internalized via clathrin-mediated endocytosis followed by endosomal acidification and lysosomal proteolysis^39–41. Neutralization assays further verified that soluble E7 monoclonal antibody effectively blocked pseudovirus (RaTG13, ZC45, BtKY72) entry mediated by hACE2ΔECD1- or hACE2ΔECD2-E7 receptors (Fig. S9), reinforcing the functional equivalence of E7-driven entry and its susceptibility to antibody inhibition.

Our data suggest that the key prerequisites for selecting bnAbs as effective ABCER ectodomains are as follows: i) Broad recognition of conserved regions across various sarbecoviruses with high binding affinity (Figs. 1d and 2b–f). ii) Epitope bound by bnAbs overlaps with multiple major epitope classes (Fig. 1a, d). iii) Epitopes bound by bnAbs lie in the RBD region, as demonstrated by the failure of the S2-targeting SH antibody despite targeting a conserved region (Fig. 1d). The ABCER platform currently relies on motifs recognized by specific scFvs (e.g., E7). E7 recognizes a quaternary epitope on the trimer but still binds RBDs if the interface is accessible. Its breadth across the sarbecovirus genus or other Coronaviridae still remains to be fully determined; however, the modular design allows for the exchange of scFv components of newly discovered bNAbs.

Unlike prior customized viral receptors¹⁷ or membrane-anchored soluble binders16 that mimic hACE2 attachment, the ABCER platform reprograms antibody recognition into a modular receptor architecture while preserving native hACE2 transmembrane and cytoplasmic domains. This modularity allows precise control of receptor orientation and rapid swapping of scFv modules to redirect viral specificity without redesigning the cellular scaffold. Such flexibility facilitates systematic dissection of receptor–virus compatibility and supports live-virus isolation from both clinical and zoonotic sources. Notably, the ABCER platform serves as an engineered surrogate to isolate diverse sarbecoviruses and does not necessarily reflect natural receptor usage or physiological permissiveness.

Furthermore, the differential performance of ΔECD1 and ΔECD2 across different assay systems can arise from two key reasons. Firstly, the morphology of VSV-based pseudovirus virions is bullet-shaped, whereas authentic SARS-related viruses are spherical⁴². Secondly, spike protein density and distribution may be different between the pseudotyped virions and authentic SARS-related viruses, an observed issue of pseudotyping viruses^43,44. These two points likely affect the membrane contact interface between the virus and the ABCER platform, resulting in variation between pseudovirus assays and replication-competent virus infections. Such biophysical variations provide an explanation as to why authentic virus replication favors the ΔECD1 architecture over the pseudovirus model. It should be noted that while ABCERs facilitate entry, successful replication may still depend on host factors absent in HEK293T cells. Therefore, a lack of replication in this system may not definitively indicate non-permissiveness in native target cells.

Although next-generation sequencing remains central to viral discovery, isolation of live viruses from clinical samples is indispensable for functional characterization. Conventional systems such as Vero E6 or trypsin-assisted cultures often fail to propagate emerging variants^18,45, whereas organoid-based models, though physiologically relevant⁴⁶, remain limited in viral permissiveness. In this context, ABCERs provide a potential experimental platform for virus isolation and antiviral screening (Fig. 3 and Fig. S8). Integration of ABCERs into organoid or airway epithelial systems could further enhance translational utility, enabling studies of viral pathogenicity, antiviral testing, and vaccine antigen evaluation. For biosafety considerations, this study focused on known SARS-related coronaviruses (Fig. 3) to validate the reliability and controllability of the ABCER system before extending it to uncharacterized field isolates.

In summary, our work advances receptor engineering from mimicking natural interactions to rationally reprogramming them. By harnessing the precision of antibody-antigen recognition within a cellular receptor context, ABCERs bridge the conceptual gap between immunity and viral entry. This platform establishes a generalizable strategy for the synthetic reconstruction of virus-host interfaces, with implications that extend from coronavirus surveillance to the broader field of programmable synthetic virology.

Methods

Ethical statement

Clinical samples were collected under the following DSRB protocol: A Multi-centered Prospective Study to Detect Novel Pathogens and Characterize Emerging Infections (The PROTECT study group, DSRB No. 2012/00917).

Plasmid construction

Synthetic antibody-based chimeric entry receptors (ABCERs, PhCMV-hACE2▵1/2/3-×scFv-pA) were generated by standard cloning as described previously⁴⁷. Briefly, the ABCERs expression cassettes were designed using human ACE2 (UniProt ID: Q9BYF1) according to previously published protocols⁴⁷. Synthesized scFv sequences (Tables S1, S2) were inserted into pCAGGS by Gibson Assembly (NEB, cat. no. E2611L) using NotI-HF and NheI-HF sites (NEB, cat. no. R3189L, R3131L). Plasmid designs are listed in Table S4. All constructs were confirmed by DNA sequencing (Bio Basic Asia Pacific Pte Ltd).

Cell culture and transfection

Human embryonic kidney 293T (HEK293T, ATCC, CRL-3216), TMPRSS2-Vero E6 (made in-house) cells were maintained at 37 °C, 5% CO2 in DMEM (Gibco, cat. no. 11965092) supplemented with 10% (v/v) FBS (Gibco, cat. no. A5256701) and 1×Penicillin-Streptomycin-Glutamine (Gibco, cat. no. 10378016). For 24-well plates, 1.5 × 10⁵ HEK293T cells were seeded and transfected at 80%-90% confluence using FuGENE® 6 Transfection Reagent (Promega, cat. no. E2692). Briefly, plasmids were diluted in 50 μL Opti-MEM (Gibco, cat. no. 31985070) and mixed with FuGENE® 6 Transfection Reagent according to the manufacturer’s instructions.

Lentivirus production

Lentiviral particles were generated as described previously⁴⁷. Stable ABCER-expressing cell lines were generated using a third-generation HIV-1-based lentiviral packaging system. Briefly, HEK293T cells (6 × 10⁵ per well, 6-well plates) were co-transfected with the pLenti-puro-ABCER lentiviral vector (1.25 μg), pMDLg/pRRE (gag/pol) (0.65 μg), pRSV-ReV (Rev) (0.25 μg), and VSV G-encoding pMD2.G (0.35 μg) using FuGENE® 6. Plasmids were prepared with ZymoPURE II Plasmid Maxiprep Kit (Zymo Research, cat. no. D4202). Supernatants were harvested 48 h, titrated, and stored at −80 °C. Transduction efficiency was assessed three days post-transduction via flow cytometry.

Generation of ABCER-expressing cells

HEK293T cells were transduced with lentiviral constructs and, 24 h later, selected in 10 μg/mL blasticidin (Thermo Fisher Scientific, cat. no. A1113903) for 3 days. Stable populations were maintained in blasticidin-containing medium (10 μg/mL).

Western blot

HEK293T cells seeded in 24-well plates were transfected with 500 ng per well of expression plasmids encoding the indicated ABCER constructs. At 48 h post-transfection, cells were lysed in RIPA Lysis and Extraction Buffer (Thermo Scientific, cat. no. 89900) containing cOmplete™, EDTA-free Protease Inhibitor Cocktail (Roche, cat. no. 11873580001). Lysates were mixed with 1× SDS-PAGE loading buffer, boiled for 10 min, and separated by SDS-PAGE on 10% gels. Proteins were then transferred onto PVDF membranes (Bio-Rad, cat. no. 1620177), which were blocked with 5% (w/v) BSA in TBST (TBS containing 0.1% Tween-20) for 1 h at room temperature. Membranes were incubated for 1 h at room temperature with HRP-conjugated His tag antibody (1:10,000; SouthernBiotech, cat. no. 4603-05) or HRP-conjugated GAPDH antibody (1:10,000; GenScript, cat. no. A00192) for 1 h at room temperature, then washed three times with TBST. Signals were detected using Amersham ECL Prime Western blotting detection reagent (Cytiva, cat. no. RPN2232) on ChemiDoc™ MP Imaging System (Bio-Rad).

Pseudovirus assays

Sarbecovirus spike expression plasmids were cloned into pCAGGS. VSV*ΔG-based spike-pseudotyped viruses were produced as described previously³⁸. Briefly, HEK293T cells were transfected with spike plasmids (SARS-CoV, GenBank: AAP30030.1; WIV1, GenBank: KC881007.1; SARS-CoV-2, Accession, YP_009724390; GX-P5L, GenBank: QIA48632.1; RaTG13, GenBank: QHR63300.2; ZC45, GenBank: AVP78031.1; JTMC15, GenBank: ANA96027.1; Rf1, GenBank: ABD75323.1; Khosta-2, GenBank: QVN46569.1; BM48-31, Accession: YP_003858584; BtKY72, GenBank: APO40579.1), and 24 h later, infected with luciferase-encoding VSV*ΔG. After 1 h of adsorption, the inoculum was replaced with fresh medium containing anti–VSV-G antibody (I1-hybridoma, ATCC, cat. no. CRL-2700) to neutralize residual input virus. Supernatants were collected 20 h post-infection, clarified by centrifugation at 3000 × g for 10 min, and stored at −80 °C.

For infection, hACE2-293T or ABCER-293T cells were seeded in 96-well plates in DMEM supplemented with 2% FBS and exposed to serial dilutions of pseudoviruses. After 24 h incubation at 37 °C, 5% CO2, luciferase was quantified with ONE-Glo™ Luciferase Assay System (Promega, cat. no. E6110) on a BioTek Cytation 5 Cell Imaging Multimode Reader (Agilent). Background-subtracted relative luminescence intensity (a.u.) was normalized to that of mock-infected wells. For titers determination, serial dilutions of harvested supernatants were applied to hACE2-293T, and luciferase signals were quantified 24 h post-infection.

Cytotoxicity

HEK293T cells (3 × 10⁴ per well, 96-well plates) were treated for 24 h with serial dilutions of Filipin (Sigma-Aldrich, cat. no. F9765), chlorpromazine (Sigma-Aldrich, cat. no. C8138), chloroquine (Sigma-Aldrich, cat. no. C6628), or E64d (Sigma-Aldrich, cat. no. E8640). Viability was measured with CellTiter-Glo® Luminescent Cell Viability Assay (Promega, cat. no. G7570). Plates were incubated 15 min at room temperature before reading on a BioTek Cytation 5 Cell Imaging Multimode Reader (Agilent).

Neutralization assay

HEK293T cells (3 × 10⁴ per well) were seeded in 96-well plates and incubated overnight at 37 °C with 5% CO₂. Serial two-fold dilutions of E7 monoclonal antibody (0–32 μg/ml) were prepared in fresh growth medium (30 μL per dilution) and mixed 1:1 with pseudovirus suspensions. After incubation for 1 h at 37 °C, 50 μL of the antibody-virus mixture was added to pre-seeded cells.

Following 24 h incubation, under identical conditions, infection efficiency was quantified by measuring luminescence using the ONE-Glo™ Luciferase Assay System (Promega, cat. no. E6110) on a BioTek Cytation 5 Cell Imaging Multimode Reader (Agilent). Neutralization was calculated as: Inhibition (%) = [1 − (luminescence intensity_sample/luminescence intensity_virus-only)] × 100.

Protein purification

Recombinant proteins were expressed in Expi293 cells (Thermo Fisher Scientific), transfected with endotoxin-free plasmids using ExpiFectamine 293. Cells were cultured in FreeStyle 293 medium at 37 °C, 5% CO2, and harvested seven days post-transfection. Conditioned medium was clarified (4000 × g, 20 min) and filtered (0.22 µm).

For His-tagged RBD proteins (SARS-CoV RBD, GenBank: AAP30030.1, residues 306-527; WIV1 RBD, GenBank: KC881007.1, residues 307-528; SARS-CoV-2 RBD, Accession, YP_009724390, residues 319-541; GX-P5L RBD, GenBank: QIA48632.1, residues 317-539; RaTG13 RBD, GenBank: QHR63300.2, residues 319-541; ZC45 RBD, GenBank: AVP78031.1, residues 315-518; JTMC15 RBD, GenBank: ANA96027.1, residues 305-508; Rf1 RBD, GenBank: ABD75323.1, residues 310-513; Khosta-2 RBD, GenBank: QVN46569.1, residues 307-526; BM48-31 RBD, Accession: YP_003858584, residues 310-528; BtKY72 RBD, GenBank: APO40579.1, residues 309-530), clarified supernatants were applied to Nuvia™ IMAC Resin (BIO-RAD, cat. no. 7800801) pre-equilibrated with 20 mM Tris–HCl (pH 8.0), 300 mM NaCl, and 20 mM imidazole. Bound proteins were eluted using a linear 20–500 mM imidazole gradient, concentrated (Amicon Ultra, 10 kDa cutoff), and purified by gel filtration on a Superdex 200 Increase 10/300 GL column (Cytiva) in PBS or HEPES-buffered saline (pH 7.4).

For Fc-tagged constructs (E7 antibody, PDB: 7Y71), media were loaded onto Pierce™ Protein A Agarose (ThermoFisher, cat. no. 20334), washed with PBS, and eluted with 0.1 M glycine-HCl (pH 2.7). Fractions were immediately neutralized with 1 M Tris–HCl (pH 8.0), concentrated, and buffer-exchanged into PBS.

Protein purity was confirmed by SDS–PAGE under reducing and non-reducing conditions, and aliquots were stored at −80 °C.

Flow cytometry

Cells were harvested 48 h post-transfection and resuspended in PBS. Approximately 5 × 10⁵ live cells from each transfected well were incubated with 2 μg/mL biotinylated XBB RBD protein (made in-house) for 1 h, stained with streptavidin-Phycoerythrin (PE) secondary antibody for 1 h (Streptavidin-PE, R&D Systems, cat. no. F0040). Fluorescence was measured on a BD LSRFortessa™ Cell Analyzer flow cytometer (BD) using acquired with PE channel. Data were compensated and analyzed using FlowJo v10 software.

Confocal microscopy

Fluorescence imaging was performed 48 h after transfection using an Olympus IX73 inverted microscope equipped with a 10× objective and a Hamamatsu ORCA-Flash4.0 camera. EGFP signals were captured at 488 nm excitation with 90% intensity; exposure times were 300 ms. Images were processed with cellSens Dimension software (Olympus).

For high-resolution imaging, cells on coverslips were imaged using an Olympus SpinSR10 confocal microscope with a 100× oil objective at 14-bit depth. Images were analyzed using ImageJ (NIH).

Reverse genetics of SARS-CoV, Pangolin CoV (PgCoV) GD1, and BANAL-20-236

Recombinant viruses, including SARS-CoV (rSARS-CoV, Frankfurt-1 strain; GenBank AY291315.1), BANAL-20-236 (rBANAL-20-236, GenBank MZ937003.2), and PgCoV GD1 (rPgCoV GD1, GISAID EPI_ISL_410721 and GeneBank accession: MT121216.1), were generated by circular polymerase extension reaction (CPER) as previously described^48–50 with modifications⁵¹. In brief, rSARS-CoV was provided by Dr. Shutoku Matsuyama (National Institute of Infectious Diseases, Japan), while rPgCoV GD1⁵² and recombinant infectious clones of WIV1 (rWIV1) and RsSHC014 (rRsSHC014) were all provided by Dr. Ralph Baric (University of North Carolina at Chapel Hill). DNA fragments covering the rBANAL-20-236 genome were synthesized by Fasmac Co., Ltd. Viral RNA was extracted, reverse-transcribed to cDNA, and cloned into plasmid fragments for CPER assembly.

A linker fragment encoding a hepatitis delta virus ribozyme, bovine growth hormone poly A signal, and cytomegalovirus promoter was generated by PCR using KOD One® PCR Master Mix (TOYOBO, cat. no. KMM-201). Gel-purified fragments were assembled by CPER and transfected into TMPRSS2-Vero E6 cells using TransIT-X2 (Takara, cat. no. MIR6003) according to the manufacturer’s protocol. Upon observation of cytopathic effects (CPE), supernatants were collected and used to generate virus stocks, which were stored at −80 °C until use.

Preparation and titration of recombinant sarbecoviruses

Working stocks of rSARS-CoV, rWIV1, rRsSHC014, rSARS-CoV-2, rPgCoV GD1, and rBANAL-20-236, were prepared and titrated as previously described^50,53. In brief, seed viruses ( ~ 30 µL) were inoculated into TMPRSS2-Vero E6 cells (5 × 10⁶ cells in T-75 flasks) in DMEM (low glucose, Sigma-Aldrich, cat. no. D6046) supplemented with 10% FBS and 1% penicillin–streptomycin (PS). At 1 h adsorption, the inoculum was replaced with 2% FBS and 1% PS medium, and cultures were maintained until CPE was observed. Supernatants were harvested, clarified by centrifugation, and stored at −80 °C.

Viral titers were determined by 50% tissue culture infectious dose (TCID50) on TMPRSS2-Vero E6 cells. Cells (1 × 10⁴ per well, 96-well plates) were infected with tenfold serial dilutions of virus and monitored for 4 days at 37 °C. CPE were scored microscopically, and TCID50/mL was calculated using the Reed–Muench method.

For sequence verification, viral RNA was extracted from working stocks using the QIAamp Viral RNA Mini Kit (Qiagen, cat. no. 52906), and genomes were analyzed as described in the Viral genome sequencing section (Table S3).

Infection with clinical SARS-CoV-2 isolates and zoonotic sarbecoviruses

For isolation of clinical SARS-CoV-2, HEK293T (negative control), hACE2-293T (positive control), hACE2ΔECD1-E7-293T, and hACE2ΔECD2-E7-293T cells were seeded in T25 flasks (1 × 10⁶ cells per 3 mL) one day before infection. Nasopharyngeal swabs positive for SARS-CoV-2 were inoculated onto monolayers and incubated for 1 h at 37 °C. The inoculum was then replaced with fresh medium, and cultures were maintained for 7 days before supernatant collection.

For zoonotic sarbecoviruses, HEK293T, hACE2-293T, hACE2ΔECD1-E7-293T, and hACE2ΔECD2-E7-293T cells (4 × 10⁴ cells per well) were seeded on collagen-coated 96-well plates (IWAKI, cat. no. 4860-010). Cells were infected with rSARS-CoV, rWIV1, rRsSHC014, rSARS-CoV-2, rPgCoV GD1, or rBANAL-20-236 at 4 × 10³ TCID50 (MOI = 0.1) and incubated for 1 h at 37 °C. After washing, 180 µL of fresh medium was added, and 10 µL of supernatant was collected at the indicated time points for RT–qPCR (to quantify viral RNA copies) and TCID₅₀ assays (for infectious titers).

Authentic samples were processed under identical conditions, with incubation periods of 1–3 days, depending on replication kinetics.

RT-qPCR

For clinical SARS-CoV-2 isolates JN.1, qPCR was performed directly on harvested culture supernatants using the Alliance Biomed RESOLUTE V2.0 SARS-CoV-2 Direct Detection Kit (AMT, cat. no. CKFG0002), following the manufacturer’s instructions. For SARS-CoV-2 XBB, viral RNA was first extracted from cell culture supernatants using the QIAamp Viral RNA Mini Kit (Qiagen, cat. no. 52904), and qPCR targeting the XBB-specific region was performed using the Novaplex™ SARS-CoV-2 Variants VII assay (RUO) (Seegene), according to the manufacturer’s protocol.

For zoonotic sarbecoviruses, RT-qPCR was performed as previously described⁴⁹. In brief, 5 μL of culture supernatant was mixed with 5 μL of 2× RNA lysis buffer [2% Triton X-100 (Nacalai Tesque, cat. no. 35501-15), 50 mM KCl, 100 mM Tris–HCl (pH 7.4), 40% glycerol, 0.8 U/μl recombinant RNase inhibitor (Takara, cat. no. 2313B)] and incubated for 10 min at room temperature, diluted with 90 μL RNase-free water, and 2.5 μL was used as the template for RT-qPCR with the One Step TB Green PrimeScript PLUS RT-PCR kit (Takara, cat. no. RR096A). Primers were: Forward, 5′ GGAAAGGTTATGGCTGTAGTTGTG-3′ and Reverse, 5′-CCGCACGGTGTAAGACGG-3′⁵⁴. Viral RNA copy numbers were standardized using a SARS-CoV-2 direct detection RT-qPCR kit (Takara, cat. no. RC300A).

All reactions were performed on a Bio-Rad CFX96 Real-Time PCR System, and cycle threshold (Ct) values were analyzed using Bio-Rad CFX Manager software.

TCID50 assay

Infectious titers were determined by 50% tissue culture infectious dose (TCID50) assay using Vero E6 cells. Monolayers were inoculated with tenfold serial dilutions of virus or culture supernatant, incubated for 1 h at 37 °C, washed with PBS, and cultured in maintenance medium for 7 days. Cytopathic effects were monitored daily, and TCID50/mL values were calculated by the Reed–Muench method.

Viral genome sequencing

Viral genome sequencing was performed as previously described⁴⁹. Viral RNA was extracted using a QIAamp viral RNA mini kit (Qiagen, cat. no. 52906). The sequencing library employed for total RNA sequencing was prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs, cat. no. E7530). Paired-end 76-bp sequencing was performed using a MiSeq system (Illumina) with MiSeq reagent kit v3 (Illumina, cat. no. MS-102-3001). Sequencing reads were trimmed using fastp v0.21.072⁵⁵ and subsequently mapped to the viral genome sequences of a lineage A isolate (strain WK-521; GISAID ID: EPI_ISL_408667)⁵⁶, BANAL-20-236 (GISAID ID: EPI_ISL_4302647 and GenBank accession: MZ937003.2)⁵⁷, PgCoV GD1 (GISAID accession no. EPI_ISL_410721 and GeneBank accession: MT121216.1), WIV1 (GenBank accession: KC881006.1), RsSHC014 (GenBank accession: KC881005.1), and SARS-CoV (strain Frankfurt 1; GenBank accession: AY291315.1) using BWA-MEM v0.7.1749⁵⁸. Variant calling, filtering, and annotation were performed using SAMtools v1.973⁵⁹ and snpEff v5.0e74⁶⁰. Information on the unexpected substitutions detected is summarized in Table S3.

ELISA assay

MaxiSorp plates (Nunc) were coated with 50 ng per well of full-length E7 mAb in coating buffer (100 µL; 28.6 mM Na2CO3, 71.4 mM NaHCO3, pH 9.55) and incubated overnight at 4 °C. Wells were washed three times and blocked for 30 min at room temperature. Samples, calibrators, and controls were diluted in OptEIA blocking buffer (BD, cat. no. 5552213) and incubated with 10 ng His-tagged spike RBD per well for 1 h at room temperature. Plates were washed and incubated with HRP-conjugated anti-His antibody (1:4,000; SouthernBiotech, cat. no. 4603-05) for 40 min. ZC45 RBDs (10 ng/well, His-tagged) were immobilized on plates and blocked. The wells were then incubated with 50 ng of E7 mAb, and binding was detected using HRP-conjugated anti-human IgG (1:4,000; Abcam, ab6759). After three washes, TMB substrate (3,3’,5,5’-Tetramethylbenzidine, BD, cat. no. 555214) was added (50 µL per well), and color development was stopped with TMB stop solution (Sera Care, cat. no. 5150-0020). Optical density at 450-530 nm was read using a Cytation 5 reader (Agilent), and data were processed in Gen5 software.

For serum blocking assays, human sera were diluted 1:20 and pre-incubated with spike RBDs for 1 h before transfer to E7-coated plates. For ZC45 binding inhibition, diluted sera were added to ZC45 RBD-pre-coated plates for 1 h, followed by incubation with HRP-conjugated E7 mAb. Subsequent steps were identical to the ELISA protocol above.

Surface plasmon resonance (SPR)

Binding kinetics were measured on a Biacore 1S+ instrument (Cytiva) at 25 °C. Full-length E7 mAb was immobilized on a Series S Protein A sensor chip (Cytiva, cat. no. 29127555), and serial two-fold dilutions of spike RBDs (0–25/50/100 nM) were injected in HBS-EP+ buffer (Cytiva, cat. no. BR100826).

For JTMC15 RBD, His-tagged protein was immobilized on a Series S NTA chip (Cytiva, cat. no. 28994951) and E7 analyte was injected at 0-6.5 nM in HBS-P buffer (Cytiva, cat.no. BR100827). All runs used multi-cycle kinetics. Sensorgrams were double-reference subtracted and fitted to a 1:1 Langmuir model using Biacore Evaluation Software (Cytiva).

Quantification and statistical analysis

Statistical analysis was performed using GraphPad Prism Version 10.2.0. The results are presented as the mean ± SD of at least 3 biological replicates, which were representative of two independent biological experiments. Graphs depict a bar diagram overlaid with a dot plot of individual data points. All statistical tests were multiple t-tests, and a p-value of 0.05 was considered statistically significant. Information regarding statistical significance is provided in the Figures, with n.s. (not significant), *p  <  0.05, **p  <  0.01, ***p  <  0.001, ****p  <  0.0001. All raw data are provided in the supplemental item.

Biosafety and biosecurity

Laboratory work with VSVΔG luciferase pseudoviruses in mammalian cell lines was performed under biosafety level 2 (BSL-2) conditions according to standard operating procedures approved by the Institutional Biosafety Committees (IBC) at the DUKE-NUS Medical School, National University of Singapore. Experiments involving authentic SARS-CoV-2 and recombinant SARS-related coronaviruses (including rSARS-CoV) were conducted in the biosafety level 2 (BSL-2) Infectious Disease Research Laboratory of the National Center for Infectious Diseases (NCID), Singapore, in accordance with approved SOPs and institutional biosafety oversight/risk assessment.

Experiments involving live zoonotic sarbecoviruses were performed in a biosafety level 3 enhanced (BSL-3) facility at The University of Tokyo (Tokyo, Japan) using a Class III biological safety cabinet. The laboratory was maintained under negative pressure ( − 30 Pa) with directional airflow and HEPA-filtered exhaust. All personnel completed institutional biosafety training and conducted procedures under approved SOPs. Work was performed by two trained researchers wearing appropriate PPE (protective suits, N95 respirators, face shields, dedicated footwear and double gloves), with on-site supervision by an experienced technician. All solid and liquid waste was decontaminated using validated sterilization procedures, with verification prior to disposal.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Transparent Peer Review file ^{(5.7MB, pdf)}

Supplementary information ^{(2.8MB, pdf)}

42003_2026_10417_MOESM3_ESM.pdf ^{(31.8KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data ^{(28.2MB, xlsx)}

Reporting summary ^{(2.5MB, pdf)}

Author contributions

L.-F.W., C.W.T., and H.Y. conceived the project. L.-F.W., C.W.T., and Y.W. designed the experiments. Y.W., C.C.E.Z., W.S., A.K.E.Z., Z.W.C., and B.L.L. performed the experiments. L.-F.W., C.W.T., H.Y., Y.W., C.E.Z.C., A.K.E.Z., W.N.C., K.S., F.Z., H.W., B.L.L., and B.Y. analyzed and interpreted the data. L.-F.W., C.W.T., H.Y., Y.W., C.E.Z.C., A.K.E.Z., and W.N.C. wrote the manuscript. All authors reviewed and approved the final version.

Peer review

Peer review information

Communications Biology thanks Stephan Steinke and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary handling editor: Ophelia Bu. A peer review file is available.

Data availability

All sequencing and structural data supporting the findings of this study are available in public databases. Specifically, the referenced viral RNA sequences are available in GenBank under the following accession numbers: SARS-CoV (AY291315.1), WIV1 (KC881006.1), RsSHC014 (KC881005.1), SARS-CoV-2 (strain WK-521; GISAID ID: EPI_ISL_408667), PgCoV-GD1 (MT121216.1), and BANAL-20-236 (MZ937003.2). Human ACE2 protein data were retrieved from UniProt (Q9BYF1). Accession numbers for full-length spike and RBD proteins (with residue ranges detailed in brackets) are available in GenBank: SARS-CoV (AAP30030.1; RBD: residues 306–527), WIV1 (KC881007.1; RBD: residues 307–528), SARS-CoV-2 (YP_009724390; RBD: residues 319–541), GX-P5L (QIA48632.1; RBD: residues 317–539), RaTG13 (QHR63300.2; RBD: residues 319–541), ZC45 (AVP78031.1; RBD: residues 315–518), JTMC15 (ANA96027.1; RBD: residues 305–508), Rf1 (ABD75323.1; RBD: residues 310–513), Khosta-2 (QVN46569.1; RBD: residues 307–526), BM48-31 (YP_003858584; RBD: residues 310–528), and BtKY72 (APO40579.1; RBD: residues 309–530). Published antibody structures were retrieved from the Protein Data Bank (PDB) under the following IDs: S2K146 (7TAS), F61 (7XMX), SA55 (7Y0W), SA58 (7Y0W), S309 (7XSW), S2X324 (8ERQ), CR3022 (6W7Y), S2X259 (7RAL), S2H97 (9ATM), E7 (7Y71), and S2P6 (7RNJ). Further details are provided in Supplementary Tables S1–S3 and the Materials and Methods section. Source data underlying graphs can be obtained from Supplementary Data. Genetic components can be made available upon reasonable request under a standard material transfer agreement by contacting linfa.wang@duke-nus.edu.sg or gmstanc@nus.edu.sg.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s42003-026-10417-3.