Human coronavirus OC43 nanobody neutralizes virus and protects mice from infection

ABSTRACT Human coronavirus (hCoV) OC43 is endemic to global populations and usually causes asymptomatic or mild upper respiratory tract illness. Here, we demonstrate the neutralization efficacy of isolated nanobodies from alpacas immunized with the S1B and S1C domain of the hCoV-OC43 spike glycoprotein. A total of 40 nanobodies bound to recombinant OC43 protein with affinities ranging from 1 to 149 nM. Two nanobodies WNb 293 and WNb 294 neutralized virus at 0.21 and 1.79 nM, respectively. Intranasal and intraperitoneal delivery of WNb 293 fused to an Fc domain significantly reduced nasal viral load in a mouse model of hCoV-OC43 infection. Using X-ray crystallography, we observed that WNb 293 bound to an epitope on the OC43 S1B domain, distal from the sialoglycan-binding site involved in host cell entry. This result suggests that neutralization mechanism of this nanobody does not involve disruption of glycan binding. Our work provides characterization of nanobodies against hCoV-OC43 that blocks virus entry and reduces viral loads in vivo and may contribute to future nanobody-based therapies for hCoV-OC43 infections. IMPORTANCE The pandemic potential presented by coronaviruses has been demonstrated by the ongoing COVID-19 pandemic and previous epidemics caused by severe acute respiratory syndrome coronavirus and Middle East respiratory syndrome coronavirus. Outside of these major pathogenic coronaviruses, there are four endemic coronaviruses that infect humans: hCoV-OC43, hCoV-229E, hCoV-HKU1, and hCoV-NL63. We identified a collection of nanobodies against human coronavirus OC43 (hCoV-OC43) and found that two high-affinity nanobodies potently neutralized hCoV-OC43 at low nanomolar concentrations. Prophylactic administration of one neutralizing nanobody reduced viral loads in mice infected with hCoV-OC43, showing the potential for nanobody-based therapies for hCoV-OC43 infections.

was first isolated in 1967 from patients who presented with upper respiratory disease, though molecular clock analysis has pointed toward its original emergence in humans from a rodent reservoir via a livestock species in the 1890s (4)(5)(6).Generally, hCoV-OC43 presents with mild upper-respiratory tract symptoms; however, the virus has been shown to have neuroinvasive properties as well as being capable of causing more severe disease and fatal pneumonia in rare cases in children, the elderly, and the immunocompromised (7)(8)(9).
On the virion surface, hCoV-OC43 expresses a trimeric spike protein which is critical for viral entry, host range, and tissue tropism and is a major target of the host immune response (10).Spike proteins undergo major conformational changes in their transition from prefusion to post-fusion states.Each monomer consists of two major functional subunits-S1 and S2.The S1 subunit is further divided into four distinct domains (S1 A-D ) (Fig. S1A).The S1 subunit is responsible for host receptor recognition, whereas the S2 subunit is involved in membrane fusion.Structural characterization shows how OC43 S1 A (also known as the N-terminal domain) mediates viral attachment to host 9-O-acetylated sialic acids (11).Interestingly, the S1 B domain (also known as the C-terminal domain), which among other human coronaviruses is typically responsible for binding to protein receptors, has no such role yet elucidated for hCoV-OC43.
While the human infective betacoronaviruses share a conserved fold in the S1 B core, the end distal from spike in the open conformation is largely structurally unique to each hCoV, with the exception of similar distal end structures in SARS-CoV-1 and SARS-CoV-2 (12,13).This distal section contains the receptor-binding motif (RBM) in all other human infective betacoronaviruses that directly engages with the host receptor for cell entry (12)(13)(14)(15)(16)(17).Despite a relative lack of sequence information compared to SARS-CoV-2, and epidemic hCoVs SARS-CoV-1 and MERS-CoV, there is considerable variation in the OC43 S1 B domain.This is particularly the case in regions which overlap the receptor-binding motifs in other hCoVs (Fig. S1B and C).
Monoclonal antibodies (mAbs) are important anti-viral treatments in the global effort against the COVID-19 pandemic (18).Given the essential role of S1 B in receptor binding for viral entry, it is the major antigenic site recognized by neutralizing mAbs across infective human coronaviruses.Over 5,000 mAbs have been described against SARS-CoV-2 S1 B (19), including ones with therapeutic potential (20,21).S1 B -binding neutralizing antibodies have also been described for all human endemic coronaviruses (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32).Although a collection of mAbs against hCoV-OC43 has been described, none are currently in clinical use (33)(34)(35)(36)(37)(38)(39)(40).Anti-hCoV-OC43 mAbs have been generated by either immunizing mice with hCoV-OC43 (33), MERS-CoV (34), or SARS-CoV-2 spike (35,36), or derived from SARS-CoV-2 convalescent patients (37)(38)(39).SARS-CoV-2-and MERS-CoV-derived mAbs, which showed cross-recognition of hCoV-OC43, bind to conserved epitopes on the S2 subunit, specifically the stem helix region (34)(35)(36)(37)(38)(39).However, the majority of these cross-reactive mAbs have little to no neutralizing activity against hCoV-OC43 S vesicular stomatitis virus pseudotypes (35).Recently, over 22 chimeric neutralizing mAbs obtained from mouse immunization were mapped to the S1 subunit of hCoV-OC43 (33).Four neutralizing S1 A mAbs compete with sialic acid for binding to the OC43 spike.Using cryo-EM, S1 A -targeting mAb 46C12 was confirmed to bind to the sialoglycan-binding site and represents one of the more potent neutralizing antibodies against hCoV-OC43.Cryo-EM also revealed two unique epitopes on the S1 B subunit where 43E6 bound close to the threefold symmetry axis and 37F1 bound distal to the threefold symmetry axis.A further three mAbs were found to bind to the S1 B , but cryo-EM only showed the apo-form of the prefusion spike trimer.It is speculated that the binding sites were inaccessible in the prefusion spike conformation and required opening of the S1 B domain.Thus, hydrogen-deuterium exchange mass spectrometry and single-substitution mutagenesis were performed, revealing residues 538 and 404 as crucial for binding of the neutralizing mAb.Collectively, binding of S1 B -directed mAbs did not compete in sialoglycan-binding assays and are thought to neutralize infection through an alternative mechanism (33).None of these neutralizing antibodies have been tested in an in vivo model of hCoV-OC43 infection.Nanobodies contain the smallest natural antigen-binding fragment consisting of only a heavy chain domain and are present in camelids (41).They maintain high-affinity recognition of antigen while having improved stability across pH and temperature ranges compared with conven tional mAbs.Multiple nanobodies have been identified as potential therapeutics against SARS-CoV-2 and other human coronaviruses, and their small size allows a wide variety of multivalent display formats to increase potency (42)(43)(44)(45)(46)(47)(48).Their increased stability also allows for alternative delivery mechanisms to be explored, including inhaled delivery for respiratory diseases (49,50).To the best of our knowledge, no nanobodies have been characterized against hCoV-OC43.Here, we identified nanobodies by screening a phage display library generated from an alpaca immunized with hCoV-OC43 S1 B and S1 C domains.Two high-affinity nanobodies showed potent neutralization against hCoV-OC43 and reduced viral loads in a mouse model of respiratory infection.In addition, we provide structural characterization of a neutralizing nanobody against hCoV-OC43.

High-affinity hCoV-OC43 RBD-binding nanobodies that neutralize virus
To identify nanobodies that are effective at neutralizing hCoV-OC43, we immunized an alpaca with recombinant hCoV-OC43 protein which encompassed the S1 B and S1 C domains (OC43 S1 B+C ) (Fig. S1A).We generated a nanobody phage display library from the alpaca and performed two rounds of phage display panning.Sequence alignment and phylogenetic analyses of positive phage supernatants from ELISA-based screen ing identified 45 distinct nanobody clonal groups based on at least one amino acid difference in the complementary determining region 3 (CDR3), which we named WNb 276 to WNb 320 (Fig. 1A; Table 1).The CDR3 lengths varied between 11 and 19 residues.The 45 nanobodies were expressed and purified with yields ranging from 0.7 to 2.9 mg (Fig. S2; Table 2).We tested the potential of the 45 nanobodies to neutralize hCoV-OC43 virus using a microneutralization assay (MNV) with viral cytopathic effect as a measure ment of infectivity (51).Two nanobodies WNb 293 and WNb 294 neutralized virus at 0.21 and 1.79 nM, respectively (Fig. 1B; Table 2).Compared to the other 43 nanobodies, the CDR3s of WNb 293 and WNb 294 are more like each other, with four amino acid differences in their CDR3 between them (Fig. 1B; Table 1).
Using bio-layer interferometry, we confirmed that 40 out of 45 nanobodies bound to recombinant OC43 S1 B+C with K D ranging from 1.55 to 148.70 nM (Fig. 1C; Table 2; Fig. S3).The irrelevant nanobody controls WNb 15 and WNb 36 which have specificity to SARS-CoV-2 had no detectable binding to OC43 S1 B+C (42).WNb 293 and WNb 294 bound with K D 4.60-26.04nM, respectively (Table 2).To determine if WNb 293 and WNb 294 recognize distinct antigenic sites, we performed a limited nanobody competition experiment with WNb 293 and WNb 294 together with the top 7 candidates based on nanobody affinities to OC43 S1 B+C (Fig. 1D).We observed that the nine nanobodies bound to the OC43 S1 B+C in two major groups with WNb 317 bound to a distinct epitope bin, and the other cluster containing WNb 293, WNb 294, and the other six nanobodies.

Prophylactic administration of WNbFc fusions reduces viral loads in mice
The two lead nanobodies WNb 293 and WNb 294 were fused to the Fc domain of human IgG1 to allow bivalent binding and to prevent rapid clearance in vivo.The resulting WNbFc 293 and WNbFc 294 fusions were purified as dimers (Fig. 2A).WNbFc 293 and WNbFc 294 bind to OC43 S1 B+C with K D 7.51 and 36.88 nM, respectively (Fig. 2B; Table 2).WNbFc 293 and WNbFc 294 neutralize hCoV-OC43 with increased potency compared to their monomer counterpart at 0.04 and 0.57 nM, respectively (Fig. 2B).To determine neutralizing activity in vivo with topical administration in the upper respiratory tract, a mouse model of OC43 infection was developed (52) based on prior work (53,54).In this model, the virus replicates 1-2 logs above input virus, peaking at day 2 post infection and is therefore suitable for studying interventions that aim to block infection.Mice were either dosed with WNbFc 293 intranasally at the same time as hCoV-OC43 infection (Fig. 2C, left) or intraperitoneally 24 hours before infection (Fig. 2C, right) to determine if systemic administration inhibits hCoV-OC43 replication in the upper respiratory tract.Infected controls were untreated (OC43 control) or dosed either intranasally or intraperi toneally with a non-specific nanobody-Fc fusion (Nb-cntrl).Nasal turbinates and nasal washes were collected at indicated times after infection to assess viral load (Fig. 2C).
Intranasal treatment with WNbFc 293 rapidly reduced viral load in nasal turbinates (84-fold) compared with both control groups by 24 hours post infection with the peak effect observed at 48 hours post infection (115-fold reduction) (Fig. 2D).Systemic intraperitoneal WNbFc 293 administration also reduced turbinate viral load (approx.competition experiments using bio-layer interferometry using immobilized nanobodies indicated on the left column incubated with nanobodies indicated on the top row pre-incubated with OC43 S1 B+C using a 10:1 molar ratio.Binding of RBD premixed with nanobody was calculated relative to RBD binding alone, which was assigned to 100%.The green and white boxes represent non-competing and competing nanobodies, respectively.10-fold) compared to Nb-cntrl-treated mice at 48 hours post infection (Fig. 2E).To gain further insight into effect on suppressing infection viral load in nasal wash, virus shedding was also assessed.Mice treated intranasally had reduced viral shedding from 24 hours post infection which was significant when compared to Nb-cntrl-treated mice (approx.84-fold) by 48 hours (Fig. 2F).Dosing intraperitoneally also reduced virus shedding in the upper respiratory tract, and this was significant (18-fold) by 96 hours post infection compared to OC43-infected control group (Fig. 2G).

Structural characterization of neutralizing hCoV-OC43 nanobody
To visualize how WNb 293 bound to OC43 S1 B , we incubated WNb 293 and OC43 S1 B at a ratio of 1.2:1 for 1 hour at room temperature and purified the complex by size exclusion chromatography (SEC).Unfortunately, none of our initial crystallization hits were suitable for high resolution diffraction.To improve crystallization success, we added WNb 317 to the WNb 293-OC43 S1 B complex.WNb 317 is a non-neutraliz ing nanobody that binds to OC43 S1 B with K D 3.15 nM and to a non-overlapping epitope compared to WNb 293 (Fig. 1B and D; Table 2).We determined the crystal structure of WNb 293-WNb 317 OC43 S1 B to 2.90 Å (Fig. 3A; Fig. S4).The complex was solved in the P2 1 2 1 2 1 space group with three complexes in the asymmetric unit (Table 3).We found that WNb 293 contacts the OC43 S1 B at the top left of the subunit, while WNb 317 contacts toward the top center of S1 B when oriented in the open spike conformation with the S1 B connections to the S1 C and S1 D subunits as "down"(Fig.3B and C).Overlaying the OC43 S1 B with that of SARS-CoV-2 reveals that both of our nanobodies bind in OC43s structurally unique RBM region.WNb 293 binds primarily via the CDR2 and CDR3 loops, with a buried surface area of 732 Å 2 (Fig. 3D and E).A total of 19 WNb 293 residues form the binding interface, including 19 hydrogen bonds (Table S1) (Fig. S4A and B).S1 B contacts WNb 293 with binding loop 1 (CDR2 and surrounding residues from residues 52 to 60) and the binding loop 2, which consists of residues 100-109 within the CDR3 (Fig. 3D; Table S1).These WNb 293 contact residues bind a continuous footprint on the S1 B surface covering residues 497-499, 503-508, 524-533, and 546-548 (Fig. 3E; Table S1).
Using SEC, we were able to obtain a WNb 317 OC43 spike complex but unable to complex WNb 293 with OC43 spike (Fig. S5A and B).However, there is no obvious steric hinderance for WNb 293-OC43 spike binding when overlayed on a cryo-EM map of the closed conformation (12-15, 17, 55) (Fig. S5C).There are three substitutions in WNb 293 footprint on OC43 between the isolates used here for OC43 S1 B and OC43 spike (GenBank AAT84362.1 and GenBank AIL49484.1,respectively).Of these, the D531 sidechain in the OC43 S1 B is involved in multiple hydrogen bonds to the WNb 293 CDR2 in the crystal structure, which may be abrogated by its substitution to histidine.

DISCUSSION
Here, immunization of an alpaca with recombinant OC43 protein enabled the identification of two potent neutralizing nanobodies against hCoV-OC43, WNb 293, and WNb 294.Furthermore, when fused to the Fc domain of human IgG1, the WNb 293-Fc fusion retained high-affinity binding K D 7.51 nM with potent neutralizing activity against hCoV-OC43 as low as 0.04 nM.To the best of our knowledge, the WNb 293-Fc fusion is  the first nanobody Fc fusion shown to prevent hCoV-OC43 virus infection in mice.Our crystal structure of WNb 293 bound to OC43 S1 B reveals a neutralizing epitope with unknown function that binds a separate domain to that interacting with sialoglycans in host cell entry (Fig. 3A; Fig. S1A).
We determined the effectiveness of WNbFc 293 as an anti-viral drug-delivered either topically intranasally to model a nasal spray-based treatment or systemically intraperitoneally to determine if indirect treatment could provide protection in the upper respiratory tract.Intranasal treatment at the time of infection was highly effective reducing viral load in nasal tissue and virus shedding in nasal wash and clearly supports the development of a nanobody-based nasal spray.Further work is now needed to determine the window of anti-viral treatment effect pre-and post-infection.Systemic delivery was also effective at suppressing viral replication in the upper respiratory tract.The magnitude of reduction (10-fold compared to controls) was less than that observed with intranasal treatment (>100-fold).Further studies will build on this proof-of-con cept data to determine the duration of anti-viral activity of CoV-targeting nanobody-Fc fusions in the upper respiratory tract following intranasal topical delivery versus systemic treatment intraperitoneally.
Two neutralizing nanobodies were identified from our immunized library, WNb 293 and WNb 294 (Fig. 1A; Table 1).These nanobodies share a high degree of homology with 92% sequence identity.Both the CDR1 and CDR2 sequences for these nanobodies are identical, with six changes in the framework regions and four substitutions in the CDR3 loop (Table 1).Despite this sequence conservation, the neutralizing potency of WNb 293 is more than eightfold higher than for WNb 294 (Fig. 1B).Our structural analyses of WNb 293 with OC43 S1 B do not provide definitive explanations for the difference in potency.The OC43 S1 B -WNb 293-WNb 317 structure reveals that none of the framework substitutions are close to the S1 B -binding interface, suggesting that these framework changes are not responsible for the difference in potency.Only two of the four CDR3 substitutions are involved in direct OC43 S1 B interactions, G100W and F106Y (WNb 294 residue, position, WNb 293 residue).The hydroxyl introduced by the F106Y substitution does not bind to OC43 S1 B and is unlikely to have any impact on the interaction.The WNb 293 W100 contacts K546 OC43 S1 B via the sidechain which would not be possible with G100 in WNb 294 and may impact WNb 293 binding to OC43 S1 B (Fig. 3D; Fig. S3A).The remaining two residues substituted in the CDR3 loop, E99L, and A111S do not directly contact OC43 S1 B .However, these CDR3 loop substitutions result in additional sidechain-mediated interactions within the CDR3 loop and between the surrounding nanobody framework regions.These may impact the conformation and stabilization of the CDR3 loop as a whole and alter the interaction between OC43 S1 B and the nanobody, though this requires further mutational analyses.
Though sequence information is extremely limited for hCoV OC43 compared with SARS-CoV-2, analysis of 213 unique OC43 S1 B sequences obtained from the Bacterial and Viral Bioinformatics Resource Center (https://www.bv-brc.org)revealed considerable variation within the domain (Fig. S1B and C).Within the WNb 293 footprint on the S1 B , several residues display considerable variation (Fig. S5E).In particular, the loop region from 503 to 508 contains several highly variable residues with the direct WNb 293 contact position 505 being conserved in only 57% of our 213 identified S1 B sequences (Fig. S1B and C).Along with the previously identified contact residue mutation D531H in strain 1783A_10 (GenBank AIL49484.1),this suggests that, similar to SARS-CoV-2, neutralization may be strain specific for some antibodies and nanobodies.
We identified WNb 317 as a non-neutralizing nanobody that binds to OC43 S1 B with K D 3.15 nM.Both WNb 293 and WNb 317 were found to bind to a structurally unique region of the OC43 S1 B that corresponds to the RBM in other hCoV S1 B domains (Fig. 3A through C).A large collection of RBM-targeting antibodies have been identified against SARS-CoV-2 and were grouped into class 1 and class 2 depending on the epitope targeted; high-affinity antibodies in these classes were found to be potently neutralizing (56)(57)(58).The majority of antibodies in these classes neutralized only a narrow range of SARS-CoV-2 VoC due to the rise of escape mutations in the RBM (58); however, there are some reports of broadly neutralizing antibodies targeting these epitopes (59,60).
Using cryo-EM, Wang et al. and Bangaru et al. structurally characterized five human antibodies that partly overlap with the WNb 317-binding site on OC43 (Fig. 3H through L) (33,40) as well as determining a low-resolution structure of an antibody which may overlap the WNb 293 site (Fig. S5D).Two of the five WNb 317 antibodies, 43E6 and 47C9, display in vitro neutralizing activity (IC 50 = 0.105 and 0.228 µg/mL, respectively) using virus neutralization assays, while the remaining three have not been tested (33).Unfortunately, direct structural comparison is somewhat hindered by difference in resolution of the neutralizing antibody structures, 47C9 and 43E6 at 3.9 and 3.7 Å, respectively, compared with 2.9 Å for WNb 317.However, their interfaces have a broadly similar surface area at 903, 766.5, and 714 Å 2 for WNb 317, 47C9, and 43E6 respectively (PDBePISA).The 47C9 and 43E6 heavy chains partly overlap with the binding epitope for WNb 317 (Fig. 3H and I; Table S1).However, 26 of 31 OC43 S1 B contact residues with WNb 317 are unique to the WNb 317 interaction with no neutralizing antibodies identified having a more similar epitope to WNb 317.
We speculate that the mechanism of neutralization for 47C9 and 43E6 might be due to the unique portions of their epitopes compared to WNb 317 which could incorporate part of a putative receptor-binding site.Differences in size and orientation between 47C9/43E6 antibodies at 150 kDa and the WNb 317 nanobody at just 14 kDa may also allow the larger antibodies to sterically hinder an OC43 S1 B interaction with a putative receptor at an adjacent site.
Alternatively, sequence variation between OC43 isolates may be responsible for the lack of neutralization in our MNV assay.Similarly to the WNb293 footprint, the WNb 317-binding interface includes regions containing considerable variation.The region 470-480 has several variable positions, particularly 480 which is only conserved in 59% of our 213 isolates (Fig. S1B and C), and forms part of a hydrophobic patch, which coordinates WNb 317 CDR2 and CDR3 binding (Fig. S4G; Fig. S5E).Screening for neutralization against a broader range of clinical isolates may provide additional information on the neutralizing capacity of this epitope.
Wang et al. also identified neutralizing antibodies 56E10, 45B9, and 65A11 that bind to a cryptic epitope which is unavailable in the closed spike conformation (33).Though they were unable to structurally characterize this epitope, HDMX studies showed that these antibodies bound to residues 399-406, 417-421, and 537-547 (33).Interestingly, WNb 293 contacts 546 in this proposed cryptic region, and we were unable to complex WNb 293 with using a prefusion-stabilized OC43 spike which is in the closed conforma tion (Fig. S3C and D), though this is likely due to sequence variation between viral isolates used for our recombinant spike and S1 B domains (GenBank AAT84362.1 and GenBank AIL49484.1,respectively).Further structural insight or alternative approaches are required to fully determine if WNb 293 occupies part of a cryptic epitope on OC43 spike.
Nanobody and antibody cocktails targeting non-overlapping epitopes on the coronavirus receptor-binding domains have shown promise in preventing occurrence of resistance mutation.Combining nanobodies and conventional antibodies against hCoV-OC43 may be advantageous for controlling more highly infectious variants and reducing the potential for virus escape mutations to develop.Proteinaceous receptors for all other hCoVs have been identified (12)(13)(14)(15)(16)(17)61) The identification of a proteinaceous entry receptor for OC43 could help to provide insight into the neutralizing mechanisms of these nanobodies that do not bind at the OC43 sialoglycan-binding site but do bind at the S1 B domain, which is the protein receptor-binding site in all other hCoV.

OC43 recombinant protein expression and purification
For immunization of alpacas and in vitro assays, two recombinant OC43 constructs were designed consisting of the S1 B and S1 C domains [OC43 S1 B+C , amino acid (aa) 321-675, Uniprot P36334] and the S1 B domain alone (OC43 S1 B , aa 331-608), respectively.Recombinant OC43 S1 B+C contained a C-terminal tobacco etch virus (TEV) protease site to allow cleavage of the hexahistidine tag, while OC43 S1 B has a C-terminal hexahistidine tag only.
Both constructs were expressed in Expi293F suspension cells (ThermoFisher), which were maintained in suspension at 37°C, ≥80% relative humidity and 8% CO 2. Transfection was performed using the ExpiFectamine 293 Transfection Kit following the manufactur er's protocol (Thermo-Fisher/Gibco #A14525).After 6 days, the supernatant was collected by centrifugation and filtered through a 0.22-µm filter.The proteins were individually purified by loading the supernatant on a 5-mL HisTrap Excel HP column (Cytiva Cat# 17371206) and eluted over Ni-NTA IMAC chromatography gradient from 0% to 100% B (buffer A: 5 mM imidazole pH 7.5, 100 mM NaCl DPBS; buffer B: 400 mM imidazole pH 7.5, 100 mM NaCl DPBS).Elution fractions containing the respective proteins were pooled and concentrated using a 10K MWCO Amicon Ultra-15 centrifugal filter unit (Merck Millipore #UFC901096).A second step purification was performed on Superdex 200 10/300 Increase GL size exclusion chromatography column (Cytiva Cat# 28990944) pre-equilibrated with DPBS pH 7.5 buffer for functional studies or 20 mM HEPES 150 mM pH 7.5 for crystallographic studies.
To remove the hexahistidine tag from OC43 S1 B , TEV protease was added to purified protein and incubated overnight at 4°C.The solution was applied to an Ni-NTA IMAC column to capture any residual His-tagged OC43 S1 B and His-tagged TEV protease.The flow through containing untagged hCoV-OC43 S1 B was applied to Superdex 75 10/300 Increase size exclusion chromatography column (Cytiva #29148721) pre-equilibrated in 1× DPBS pH 7.5 for a final purification.

Alpaca immunization
One alpaca was immunized six times with approximately 200 µg of recombinant OC43 S1 B+C protein on days 0, 14, 21, 28, 35, and 42.The adjuvant used was GERBU FAMA.Immunization and handling of the alpacas for scientific purposes were approved by Agriculture Victoria, Wildlife and Small Institutions Animal Ethics Committee, project approval No. 26-17.Blood was collected 3 days after the last immunization for the preparation of lymphocytes.Nanobody library construction was carried out according to established methods as described (62).Briefly, alpaca lymphocyte mRNA was extracted and amplified by RT-PCR with specific primers to generate a cDNA nanobody library.The library was cloned into a pMES4 phagemid vector containing 10 4 unique nanobodies, amplified in E. coli TG1 strain, and subsequently infected with M13K07 helper phage for recombinant phage expression.

Isolation of OC43 S1 B+C nanobodies
Biopanning for recombinant OC43 S1 B+C nanobodies using phage display was per formed as previously described with following modifications (42).Phages displaying OC43 S1 B+C -specific nanobodies were enriched after two rounds of biopanning on streptavidin-coated dynabeads (Invitrogen) coated with 200 pMol of biotinylated recombinant protein.After the second round of panning, we screened 376 clones from both round 1 and round 2 by ELISA and observed a 90% positive hit rate, and 676 clones were selected for further analysis.Positive clones were sequenced and annotated using the International ImMunoGeneTics database (IMGT) and aligned in Geneious Prime 2020.2.4.A total of 45 distinct nanobody clonal groups were identified based on at least one amino acid difference in the complementary determining region 3.

Nanobody expression and purification
Nanobodies were expressed in Escherichia coli WK6 cells (63).Bacteria were grown in 250 mL Terrific Broth supplemented with 0.1% (wt/vol) glucose and 100 ug/mL carbenicillin at 37°C to an OD 600 of 0.7, induced with 1 mM IPTG, and grown overnight at 28°C for 16 hours.Cell pellets were harvested and resuspended in 20% (wt/vol) sucrose, 10 mM imidazole pH 7.5, 150 mM NaCl, DPBS and incubated for 15 min on ice.Furthermore, 5 mM EDTA pH 8.0 was added and incubated on ice for 20 min.After this incubation, 10 mM MgCl 2 was added to prevent nickel ion-EDTA chelation, periplasmic extracts were harvested by centrifugation, and the supernatant was loaded onto a 1-mL HisTrap Excel HP column (Cytiva Cat# 17371205).The nanobody was eluted with 400 mM imidazole pH 7.5, 100 mM NaCl, DPBS.The appropriate fractions were concentrated, and buffer exchanged into sterile DPBS using 3K MWCO Amicon Ultra-15 centrifugal filter unit concentrators (Merck Millipore).

OC43 propagation and quantification
Original OC43 (ATCC Number VR-1558) passage history is unknown and propagated on HCT-8 cells.Initial stocks from ATCC were received at concentrations of 2.8 × 10 5 TCID50/mL and passaged two times in MRC-5 cells to generate a working stock.MRC-5 cell lines were used for TCID50 assays using the Karber method to quantify viral load.

OC43 microneutralization assay
The ability of nanobodies to neutralize the infectivity of 100 median tissue culture infectious doses (TCID50) of virus was assessed in a microneutralization assay as previously described (64) .Serial twofold dilutions of nanobodies starting at 1:20 were incubated with OC43 in MEM/0.5% BSA at room temperature for 1 hour.Residual virus infectivity was assessed in quadruplicate wells of MRC-5 cells, and viral cytopathic effect was read on day 7.The neutralizing antibody titer was calculated using the Reed-Muench method as previously described (64).

Bio-layer interferometry
Nanobody affinities to recombinant OC43 S1 B+C were measured using an Octet RED96e instrument (ForteBio).Assays were performed at 25°C in solid black 96-well F-bottom plates (Greiner-Bio One #655209) agitated at 1,000 rpm.Kinetic buffer was composed of Gibco 1× DPBS (Life Technologies Cat# 14190144) supplemented with 0.1% (wt/vol) BSA, 0.05% (vol/vol) TWEEN-20.A 60-s biosensor baseline step was applied before nanobodies were loaded onto Ni-NTA (Octet NTA) (Sartorius) sensors by submerging sensor tips in 5 µg/mL of each nanobody until a response of 0.5 nm was obtained, then washed in kinetic buffer for 60 s.Association measurements were performed for 180 s using a twofold concentration gradient of untagged OC43 S1 B+C from 3 to 100 nM for higher-affinity nanobodies or 16 to 250 nM for low-affinity nanobodies.Dissociation was measured in kinetic buffer for 180 s.Sensor tips were regenerated using a cycle of 5 s in 500 mM imidazole pH 7.5 and 5 s in kinetic buffer repeated five times.Baseline drift was corrected by subtracting the shift of a nanobody-loaded sensor not incubated with cleaved OC43 S1 B+C .Curve fitting analysis was performed with Octet Data Analysis 11.1 software (ForteBio) using a global fit 1:1 model to determine K D values and kinetic parameters.Curves that could not be fitted were excluded from the analysis.
WNbFc antibody affinities to OC43 S1 B+C were measured using the above method with the following modifications.Anti-human IgG Fc capture sensor tips (Octet AHC) were used for affinity measurements.For measuring affinities against OC43 S1 B+C , WNbFc antibodies were loaded onto sensor tips by submerging in 5 µg/mL of WNbFc antibody for 200 s or until a signal shift of 0.5 nm.

Epitope binning
For WNb epitope binning experiment, 30 nM OC43 S1 B+C was pre-incubated with each nanobody at a 10-fold molar excess (300 nM) for 1 hour at RT.A 30-s baseline step was established between each step of the assay.NTA sensors were first loaded with 10 µg/mL of nanobody for 5 min.The sensor surface was then quenched by dipping into 10 µg/mL of an irrelevant nanobody (Pf12p-B9 nanobody) for 5 min (65).Nanobody-loaded sensors were then dipped into premixed solutions of OC43 S1 B+C and nanobody for 5 min.Nanobody-loaded sensors were also dipped into OC43 S1 B+C alone to determine the level of OC43 S1 B+C binding to immobilized nanobody in the absence of other nanobod ies.Percentage competition was calculated by dividing the maximum response of the premixed OC43 S1 B+C and nanobody solution binding by the maximum response of OC43 S1 B+C binding alone, multiplied by 100.

Inoculations
Six-to 8-week-old female BALB/c mice were obtained from Australian Bioresources (Moss Vale, Sydney, NSW).Intranasal administration of nanobodies and viral challenge were performed under light isoflurane anesthesia (66).Nanobodies were delivered at 5 mg/mL concentrations equating to 0.5 mg in a 100-µL intraperitoneal injection and 150 µg in a 30-µL intranasal administration of 1.42 × 10 6 TCID50 units of OC43.

RNA extraction and reverse transcription
Apical lung lobes were harvested in RNA later (Ambion) and stored at −80°C, lysed in buffer RLT (Qiagen) containing 1% betamercaptoethanol with tissue dissociation using a TissueLyser II (Qiagen).Debris was pelleted by centrifugation (10 min at 10,000 RCF), and supernatants stored at −80°C.Nasal turbinates were excised and vortexed for 30 s in RLT containing 1% beta-mercaptoethanol. Nasal turbinate debris was removed, and the lysate stored at −80°C.Apical lung lobe and nasal turbinate RNA were extracted using the miRNAeasy Kit (Qiagen) following the supplier's protocol.Viral RNA was manually extracted from the nasal wash fluid using the QIAamp Viral RNA Mini Kit (Qiagen) following the supplier's protocol.RNA was measured by spectrophotometry (Nanodrop), and 1,000 ng (lung), 500 ng (nasal turbinate), or 200 ng (nasal wash) of RNA was used in reverse transcription reactions with High-Capacity cDNA Reverse Transcription Kit (ABI) per manufacturers' recommendations.

Quantitative PCR
Quantitative PCR (qPCR) was performed on the QuantStudio 6 using TaqMan Gene Expression Master Mix (Thermo Fisher Scientific) and primer-probe combinations (Thermo Fisher Scientific) as outlined in Table 4 (67).Standards of known concentration were used for absolute quantification of genes of interest.18S was used as the reference gene to normalize the copy numbers of the genes of interest.

Statistical analyses
Statistical analyses were performed using GraphPad Prism software (version 9.1.2).Data spanning multiple timepoints were analyzed by two-way ANOVA or mixed-effect analysis using Holm-Sidak's correction or Tukey's correction for multiple comparisons unless stated otherwise in the figure legends.P < 0.05 was considered statistically significant.

X-ray crystallography of the OC43 S1 B -WNb 293-WNb 317 complex
Crystallization trials were undertaken at the Monash Macromolecular Crystallisation Facility at 20°C using 96-well sitting drop vapor diffusion plates (Greiner).Crystals of the OC43 S1 B -WNb 293-WNb 317 complex were obtained from a solution containing 25% PEG 1500, 0.1 M DL-malic acid, MES monohydrate, Tris (MMT) pH 4.0 (Molecular Dimensions) after 3 days.Crystals were reproduced in 24-well hanging drop trays, and crystals forming in 27% PEG 1500, 0.1 M MMT pH 4.0 were used to seed into fine screens.Seeded crystals were grown in 23% PEG 1500, 0.1 M MMT pH 4.0 and flash frozen in liquid nitrogen at 100 K following equilibration in a reservoir solution containing 15% glycerol as a cryoprotectant.The MX2 beamline at the Australian Synchrotron (Melbourne, Victoria) was used to collect a data set for the OC43 S1 B -WNb 293-WNb 317 complex at 2.9 Å.Data were recorded using an Eiger 16M detector (Dectris) and processed using the XDS package (68).Molecular replacement was undertaken using Phaser (69,70).A search model was generated for the OC43 S1 B using a portion of a previously determined cryo-electron microscopy OC43 spike structure (PDB ID 6OHW) corresponding to this domain (AA: 331-608).Nanobody search models were generated using Nanonet (71) to modify the previously determined nanobody structure (PDB ID 5LHR chain B) to better match the WNb 293 and WNb 317 sequences, respectively.

ACKNOWLEDGMENTS
We thank Geoffrey Kong from the Monash Macromolecular Crystallisation Facility (MMCF, Clayton, VIC, Australia) for assistance with setting up the crystallization screens.This

FIG 1
FIG 1 Identification and functional characterization of neutralizing anti-OC43 nanobodies.(A) Nanobodies are ordered based on their clonal lineage as determined by their CDR3 sequences.(B) MNV IC 100 values shown are the geometric mean of n = 2-4 biological replicates (top).Alignment of CDR3 WNb 293 and 294 nanobodies with amino acid differences highlighted in black (bottom).(C) Iso-affinity plot showing the dissociation rate constants (k d ) and association rate constants (k a ) of WNb nanobodies as measured by bio-layer interferometry.Symbols that fall on the same diagonal lines have the same equilibrium dissociation rate constants (K D ) indicated on the top and right sides of the plot.WNb 293 and 294 are highlighted in blue and orange, respectively.(D) Epitope

6 TABLE 1 7 TABLE 1 a
Nanobody sequences generated against OC43 S1 B+C a Nanobody sequences generated against OC43 S1 B+C a Amino acid sequences are shown for OC43 S1 B+C -binding nanobodies including the CDR1, CDR2, and CDR3 loops and full-sequence information.Full-Length Text Journal of Virology June 2024 Volume 98 Issue 6 10.1128/jvi.00531-248

FIG 2
FIG 2 Antibody affinities and neutralization potencies of bivalent nanobody-Fc fusions.(A) Purified WNbFc 293 and WNbFc 294 migration on an SDS-PAGE under non-reducing (NR) and reducing (R) conditions.Molecular weight in kilodalton is shown on the left.(B) Neutralization of OC43 using bivalent nanobody-Fc fusions.For MNV, the values are the geometric mean of n = 2-4 biological replicates.Bio-layer interferometry affinity measurements with immobilized WNbFc fusion and OC43 S1 B+C in solution.Corresponding mean + SEM k d values are indicated.(C) Mouse models of infection.Nasal tissue lysate viral load was determined by qPCR (normalized to 18s house-keeper) following (D) intranasal and (E) intraperitoneal dosing (n = 6).Viral load in nasal wash was determined by qPCR following (F) intranasal and (G) intraperitoneal dosing (n = 6).*P < 0.05, **P < 0.01 (WNbFc 293 compared to Nb-cntrl), ^P < 0.05, ^^^P < 0.001 (WNbFc 293 compared to untreated control) by two-way ANOVA with Holm-Sidak correction for multiple comparisons.

TABLE 2
Binding affinities, neutralization values, and expression yields for OC43 S1 B+C nanobodies a

× 10 5 M −1 s −1 ) k d (× 10 −5 s −1 ) Full X 2 Full R 2 MNV (mg/mL) MNV (nM) Yield (mg)
a Binding data including mean ± SD values for K D , k a , k d , X 2 , and R 2 values as measured by bio-layer interferometry.Neutralization values shown in milligram per milliliter and nanomolar for neutralization in MNV assays.Expression yields shown in mg for 250 mL periplasmic expression in E. coli.b NB, non-binding.Full-Length Text Journal of Virology June Volume 98 Issue 6 10.1128/jvi.00531-249

TABLE 3
Data collection and refinement statistics for OC43 S1 B WNb 293 WNb 317 complex b

OC43 S1 B WNb 293-WNb 317 (PDB 8TZU)
a Values in parentheses are for highest-resolution shell.b X-ray diffraction data were collected on single crystals.