Expression and purification of hACE2 peptidase domain

For alpaca immunization, the human ACE2 peptidase domain (residues 19-615, Uniprot: Q9BYF1) was cloned in pcDNA3.1. The signal peptide from the IgK chain (METDTLLLWVLLLWVPGSTG) was introduced at the N-terminus to ensure the protein secretion to the extracellular medium, and His(x8) and Strep tags were added in-tandem at the C-terminus, following a Thrombin cleavage site. For structural studies, the Strep tag was exchanged for the ybbR tag (DSLEFIASKLA).

The plasmid coding for hACE2-His-Strep was transiently transfected into Expi293F^TM cells (Thermo Fischer) using FectoPro® DNA transfection reagent (PolyPlus) according to the manufacturer’s instructions, and 5 μM kifunensine was added to the culture medium immediately after transfection. The cells were incubated 5 days at 37 ° C and harvested by centrifugation. The supernatant was concentrated and recombinant hACE2 was purified by affinity chromatography on a Streptactin column (IBA) and by size-exclusion chromatography (SEC) on a Superdex200 column (Cytiva) equilibrated with 10 mM Tris, 100 mM NaCl, pH8. hACE2 was digested over-night at room temperature with Thrombin (1 U per 100 μg of recombinant protein) to remove the His and Strep tags. The proteolysis was stopped with 1 mM PMSF (phenylmethylsulfonyl fluoride) and the untagged hACE2 was recovered in the flow-through after injection in a Streptactin column. A final SEC was performed, after which untagged hACE2 was concentrated and stored at -80 ° C. For VHH epitope mapping experiments using BLI, hACE2 peptidase domain fused to a human Fc from IgG³⁶ was used (sACE2-Fc). For the measurement of the affinity between VHH B07-Fc and sACE2 by interferometry (BLI), we used a His- tagged sACE2 (sACE2-His) from human or hamster species (R&D systems).

For structural studies, the hACE2-His-ybbR plasmid was transfected into Expi293F^TM GnTI- cells (Thermo Fischer A14527) using FectoPro® DNA transfection reagent (PolyPlus) according to the manufacturer’s instructions. The cells were incubated 5 days at 37 ° C and harvested by centrifugation. The supernatant was concentrated and recombinant hACE2 was purified by IMAC on a His-Excel column. The eluted protein was concentrated, the buffer was exchanged (10 mM Tris-HCl, 100 mM NaCl, pH 8.0) and hACE2 was digested over-night at room temperature with Thrombin (1 U per 100 µg of recombinant protein). The cleavage was stopped with 1 mM PMSF (phenylmethylsulfonyl fluoride) and untagged hACE2 was recovered in the flow-through fraction after injection on the His-Excel column. The protein was concentrated and incubated over-night at room temperature with 1 U of endoglycosidase H in acidic buffer (NaAc 100 mM, NaCl 100 mM, pH 5.5). A final SEC was performed using a column equilibrated with 10 mM Tris-HCl, 100 mM NaCl, pH 8.0.

Plasmids, site-directed mutagenesis, and transfection

Mouse ACE2 plasmids was purchased from addgene (pscALPSpuro-MmACE2, #158087). Hamster ACE2 (haACE2) pcDNA3.1 hygro plasmid was synthetized by GeneArt (Thermo Fisher). The construct human ACE2 (hACE2) was obtained by substitution of the sequence encoding mCardinal from mCardinal-C1 plasmid (addgene, #54799) with hACE2 from the pLenti6-attB-hACE2-BSD³². The construct myc-hACE2 was obtained by substitution of the sequence encoding mCardinal from mCardinal-C1 plasmid (addgene, #54799) with myc-hACE2 from the pCEP4-myc-ACE2 (addgene, #141185)⁵⁹. Polymorphism substitutions and alanine substitutions were generated by site-directed mutagenesis using the Q5-site directed mutagenesis kit (New England Biolabs) according to the manufacturer’s instructions. The hACE2 and haACE2 mutants were all verified by sequencing (Eurofins) using the primers: CMV-Fw 5’CGCAAATGGGCGGTAGGCGTG and either hACE2-300-Rv 5’CTGTGCATCCCAGGCCTG or haACE2-275-Rv 5’CTGTGCATCCCAGGCCTG. For transient expression, HEK293 (2 x 10⁵) cells were cotransfected (lipofectamine 2000) with WT or mutant ACE2 vectors and with eGFP-N1 (Clontech, Palo Alto, Calif.), in a 5:1 ratio. The eGFP positive cells were analyzed by flow cytometry after immunostaining.

Viruses

SARS-CoV-2 Delta (B.1.617.2; GISAID ID: EPI_ISL_2029113), D614G (hCoV-19/France/GE1973/2020; GISAID ID: EPI_ISL_414631), Omicron BA.1 (GISAID ID: EPI_ISL_6794907), BQ.1.1 (GISAID ID: EPI_ISL_15731523), and JN.1.1 (hCoV-19/France/HDF-IPP21391/2023) were previously described^5,40,60,61. XBB.1.5 was isolated from a nasopharyngeal swab of an individual of Hôpital Européen Georges Pompidou (HEGP; Assistance Publique, Hôpitaux de Paris) (GISAID ID: EPI_ISL_16353849)⁶². The laboratory of Virology of HEGP sequenced the swabs. The patient provided written informed consent for the use of the biological materials. The XBB.1.16.1 strain (hCoV-19/France/GES-IPP07712/2023), the EG.5.1.3 strain (hCoV-19/France/BRE-IPP15906/2023), and the BA.2.86.1 strain (hCoV-19/France/IDF-IPP17625/2023) were supplied by the National Reference Centre for Respiratory Viruses hosted by Institut Pasteur (Paris, France) and headed by Dr Etienne Simon-Lorière / Marianne Rameix-Welti⁵. The human sample from which strain hCoV-19/France/GES-IPP07712/2023 was isolated has been provided by Dr Vanessa COCQUERELLE from Laboratory Deux Rives. The human sample from which strain hCoV-19/France/BRE-IPP15906/2023 was isolated has been provided by Dr F. Kerdavid from Laboratoire Alliance Anabio, Melesse. The human sample from which hCoV-19/France/IDF-IPP17625/2023 was isolated by Dr Aude LESENNE from Cerballiance, Lisses (France). All strains were previously collected. Titration of viral stocks was performed on Vero E6 or IGROV-1 cells, with a limiting dilution technique enabling the calculation of the median tissue culture infectious dose or on S-Fuse cells. KP.3.3 strains (hCoV-19/France/BFC-IPP06087/2024) were isolated and amplified by the Reference center for Respiratory Viruses hosted by Institut Pasteur⁷.

In replication experiments, XBB.1.16.1 (hCoV-19/France/GES-IPP07712/2023), XBB.1.5 (hCoV-19/France/PDL-IPP58867/2022), BQ.1.1 (hCoV-19/France/IDF-IPP50823/2022) and BA.1 (hCoV-19/France/PDL-IPP46934i/2021; GISAID ID: EPI_ISL_8353353) were supplied by the National Reference Centre for Respiratory Viruses hosted by Institut Pasteur (Paris, France). The human samples from which strains were isolated has been provided from the Laboratoire Deux Rives, CH Laval and Laboratoire des centres de santé et d’hôpitaux d’IDF. All strains were previously collected. Viral productions were performed on Vero-TMPRSS2 cells and titration of viral stocks was performed on Vero E6 cells. Viral concentrations were determined through Plaque Assay and the titers were quantified as plaque forming units (pfu) per milliliter.

The sequence of the viral stocks was verified by RNAseq by the National Reference Centre for Respiratory Viruses hosted by Institut Pasteur (Paris, France). All work with infectious virus was performed in biosafety level 3 containment laboratories at Institut Pasteur.

Cell lines

HEK293-hACE2 (human embryonic kidney cell line, ATCC CLR 1573) were generated by transfection and cultured in Blasticidin (10 µg/mL). U2OS (human osteosarcoma cell line, ATCC Cat# HTB-96) and A549 (human lung epithelial cell line, ATCC cat# CCL-185) cells stably expressing hACE2, and U2OS-hACE2 cells stably expressing the GFP split system (GFP1-10 and GFP11; S-Fuse cells) were previously described³². These cells were cultured in DMEM or in F-12K Nutrient Mixture Media (A549) supplemented with 10% fetal bovine serum (FBS), and 1% penicillin/streptomycin. Blasticidin (10 μg/mL) and puromycin (1 μg/mL) were used to select for hACE2 and GFP split transgenes expression, respectively.

Vero E6 (Vero 76, clone E6, Vero E6, ATCC® CRL-1586TM, simian kidney cell line) was obtained from ATCC (USA) and cultured in DMEM medium (Gibco) containing 10 % (v/v) FBS (Gibco) and penicillin/streptomycin (Thermo Fisher Scientific). IGROV-1 cells (Human ovarian cancer cell line) were from the NCI-60 cell line panel⁴⁰ and cultured in RPMI medium (Thermo Fisher) containing 10% FBS and penicillin/streptomycin. Vero E6 and IGROV-1 endogenously express ACE2.

Absence of mycoplasma contamination was confirmed in all cell lines with the Mycoalert Mycoplasma Detection Kit (Lonza). All cell lines were cultured at 37° C and 5 % CO₂.

Animal immunization and library construction

All immunization processes were executed according to the French legislation and in compliance with the European Communities Council Directives (2010/63/UE, French Law 2013-118, February 6, 2013). The Animal Experimentation Ethics Committee of Pasteur Institute (CETEA 89) approved this study (2020-27412). We subcutaneously injected an adult alpaca at days 0, 21, and 36 with approximately 150 µg of sACE2 mixed with Freund complete adjuvant for the first immunization and with Freund incomplete adjuvant for the following immunizations. A blood sample of about 150 mL of the immunized animal was collected and Peripheral Blood Mononuclear cells (PBMCs) were isolated on a ficoll discontinuous gradient.

Briefly, the vhh-encoding genes isolated from PBMCs (approximately 2 x 10⁸) were then cloned into the phagemid vector pHEN6 by using primers contained enzymatic SfiI and NotI restriction sites at the 5’ and 3’ ends, respectively. The size of the library was estimated at about 8 x 10⁷ cfu (colony-forming units). Phage display protocol was performed as described in Lafaye et al.⁶³. About 10¹² phage-VHH diluted in PBS were used to perform three rounds of panning by using sACE2 coated on Nunc Immunotubes (Maxisorp) tubes (10 µg/mL). To increase the stringency, different blocking buffers for each panning were used: 2% skimmed milk as saturating agent for the first round; 5% BSA (Bovine Serum Albumin) and Odyssey blocking buffer (LI-COR Biosciences) diluted at 1/2 for the second and third rounds.

Following panning, individual clones were screened by standard phage-ELISA procedures using an HRP/anti-M13 monoclonal antibody conjugate (11973-MM05T). Clones were selected and their VHH-encoding DNA was sequenced by Eurofins Genomics (Ebersberg, Germany). Sequence analysis was conducted using SnapGene v8.1.0 and multiple sequence alignment with ClustalW2 (EMBL-EBI). Out of the 192 individual clones tested in the phage ELISA, 123 were positive. All 123 positive clones were sequenced, revealing that 80 belonged to a dominant cluster derived from the same IGHV3S53 gene as B07 and B09. Within this cluster: B07 was found 31 times; B09 was found 23 times. Given that B07 and B09 originate from the same IGHV3S53 gene and differ by only six amino acid mutations in their sequences and 2 in their CDRs, they likely represent variants of the same lineage. B10, in contrast, was found only once and originates from a different IGHV3-3 gene, indicating a distinct evolutionary origin.

VHH production

The pHEN6 vector allows the periplasmic expression and purification of selected VHHs with a c-myc tag and His-tag at the C terminal end in frame with the VHH. Transformed E. coli TG1 cells expressed VHHs in the periplasm after induction with IPTG (1 mM) for 16 h at 30° C. After centrifugation, cells were resuspended in PBS then lysed by polymixin B sulfate (1 mg/mL) for 1 h at 4° C, the periplasmic extract being obtained after centrifugation. Purified VHHs were obtained by His pure Cobalt sepharose beads (Thermo) according to the manufacturer’s instructions.

Dimeric VHH-Fc production

The coding sequences of the selected VHHs were digested with NcoI and NotI and subcloned into pFuse-huIgG-Fc2 digested with NcoI and NotI⁶⁴. E. coli XL1-Blue transformants were obtained on 2YT agar plates containing zeocin 25 µg/mL. The plasmids coding for the recombinant proteins were purified with Nucleobond Xtra Midi Plus EF (Macherey Nagel), transiently transfected in Expi293™ cells (Thermo Fisher Scientific) using Fectro PRO DNA transfection reagent (Polyplus), according to the manufacturer’s instructions. Cells were incubated at 37° C for 5 days and then the cultures were centrifuged. Proteins were purified from the supernatants by affinity chromatography using a HiTrap protein A HP (Cytiva), followed by Size Exclusion Chromatography on a Superdex 75 column (Cytiva) equilibrated in PBS. Peaks corresponding to the dimeric VHH-Fc proteins were concentrated and stored at -80° C until used.

Biolayer interferometry (BLI)

Equilibrium binding was performed via BLI-analysis using an Octet HTX instrument (Sartorius). hACE2-Fc (or VHH-His) diluted into Sartorius’s PBS working buffer containing a final concentration of 300 mM NaCl (pH 7.2) were immobilized onto Anti-hIgG Fc Capture (AHC) (or Ni-NTA) biosensors (Sartorius), which were pre-equilibrated in the same buffer. Following a baseline, ACE2-coated sensors (or VHH-coated sensors) were incubated with various concentrations of monomeric VHHs (or various concentration of soluble hACE2) for 10 min to reach equilibrium. Subsequently, biosensors were put in working buffer for 10 min to initiate dissociation. All incubations were performed at a temperature of 30° C under continuous shaking (1,000 rpm). To minimize classical surface-dependent avidity artefacts and avoid artificial multiphasic behavior, all binding experiments were performed using a range of protein loading concentrations. Optimal ACE2-Fc densities on the biosensor were determined to be 1 µg/mL for B07, B09, and B10. After minimizing avidity effects, binding data were analyzed using the Octet Software 13.0 and association/dissociation curves were globally fitted with a 1:1 binding model to determine K_D apparent values⁶⁵. For B07-Fc binding of human or hamster ACE2, sACE2-His (1 or 2 µg/mL, R&D Systems)-coated HIS2 sensors (Sartorius) were incubated with various concentrations of VHH-B07-Fc (as indicated) in 50 mM Tris-HCl buffer (pH 7.5), supplemented with 0.01% P20 and 0.1% BSA. Binding curves from the association and dissociation steps were fitted using a 1:1 binding model.

To measure competitive binding to sACE2 protein between the different VHHs by BLI, we immobilized hACE2 fused to the human Fc domain (sACE2-Fc) onto Anti-hIgG Fc Capture (AHC) Biosensors (Sartorius) at 5 µg/mL for 10 minutes. A baseline step was realized by dipping sensors into the working buffer. Then, a first VHH was applied at 5 µg/mL, in order to bind the immobilized hACE2 and reach saturation. The sensor was then dipped into a well containing a mixture of the same VHH with the competitor one added at the same concentration (5 µg/mL).

Measurement of VHH binding or spike binding on ACE2 expressing cells by flow cytometry

Cells (2 x 10⁵) (transfected or not) were incubated for 60 min at 4° C with 10 µg/mL VHHs or 0.1 µg/mL VHH-Fc, washed twice with 2% FBS-PBS, incubated for 60 min at 4° C with anti-myc antibody (9E10, Abnova MAB0957), washed twice with 2% FBS-PBS, and incubated for 60 min at 4° C with AF488 (or AF647)-conjugated anti-mouse antibody (Invitrogen, A11029, A21236) (for VHHs detection), or with the AF488 (or AF647)-conjugated anti-human IgG antibody (Thermo Fischer A11013, A21445) (for VHH-Fc detection). The stained cells were washed twice, fixed in PFA 1.5% and analyzed on Attune flow cytometer NxT flow cytometer (Thermo Fisher) and Kaluza software v3.2.1. The parental cells (HEK293, A549) or a VHH control (anti-IgE or anti-IgM or R3VQ-Fc VHHs) were used as negative controls.

To perform competition experiments of soluble spike binding an increasing concentration of VHH (between 0.01 to 100 µg/mL) were incubated with HEK293-hACE2 cells in RPMI/10% FBS. After 30 min at 37 ° C, 5 µg/mL, or different doses (5, 8, 10 µg/mL), of soluble trimeric spike (ancestral spike)⁶⁶ was added for 30 min at 37° C. The “ancestral spike” referred here to SARS-CoV-2 Spike corresponding to the primary isolate hCoV-19/France/IDF-0372/2020 (GISAID ID: EPI_ISL_406596), Pango lineage B, Nextstrain Clade 19 A. Then, immuno-staining with 5 µg/mL anti-spike monoclonal antibody mAb48 (gift of Hugo Mouquet, Institut Pasteur)⁶⁷ and alexa-Fluor-488-conjugated anti-human IgG antibody (Thermo Fischer A11013) was performed at 4° C. Fixed cells (PFA 1.5%) were analyzed on Attune flow cytometer NxT flow cytometer (Thermo Fisher) and analyzed as described above.

ACE2 enzymatic activity

Human ACE2 activity was evaluated using SensoLyte 390 ACE2 Activity Assay Kit (AnaSpec) according to the manufacturer’s protocol. To measure hACE2 activity on purified sACE2, 10 nM of sACE2 was incubated with B07, B09, and B10 at different concentrations before adding the fluorogenic peptide substrate, MethoxyCoumarin acetic acid-Ala (Mc-Ala)/Dnp, where Mc-Ala is quenched by Dnp. The protease activity of sACE2 is indicated by an increase of fluorescence after peptide cleavage, and Mc-Ala recovery, monitored at Ex/Em = 330 nm/390 nm (Tecan SPARK).

To measure hACE2 activity on cells, 2.5 x 10⁴ HEK293-hACE2 cells per well (96-well black plate, Greiner-BioOne) were incubated with B07, B09, B10 (>20 µM) or the control inhibitor DX600 (1 µM) in 40 µL complete medium without phenol red (Gibco) before adding 40 µL of substrat (Mc-Ala/Dnp) diluted in PBS (SensoLyte 390 ACE2 Activity Assay kit). Peptide cleavage was measured in the supernatant on Tecan SPARK plate reader (SparkControl Magellan 3.2).

SARS-CoV-2 inhibition assay: S-Fuse

U2OS-ACE2 GFP1-10 and GFP11 cells, also termed S-Fuse cells, become GFP+ when they are productively infected by SARS-CoV-2³². Cells were mixed (ratio 1:1) (U2OS-ACE2 GFP1-10 (10⁴) and GFP11 cells (10⁴)) and plated in a μClear 96-well plate (Greiner Bio-One). Serial dilution of VHHs were incubated with the cells at 37° C and infected with a SARS-CoV-2 strain. Eighteen hours later, cells were fixed with 4% paraformaldehyde (PFA), washed and stained with Hoechst (dilution 1:10,000, Invitrogen). The extent of fusion was then quantified by measuring the number of objects GFP+ with an Opera Phenix high-content confocal microscope (PerkinElmer) and Harmony software version 4.9 (PerkinElmer). The percentage of inhibition was calculated using the number of syncytia as value with the following formula: 100 × (1 − (value with VHH −value in ‘non-infected’)/(value in ‘no VHH’ − value in ‘non-infected’)). Inhibition activity of each VHH was expressed as the IC50. We previously reported that the neutralization assay with the S-Fuse system is not affected by differences in fusogenicity between variants⁶⁸.

SARS-CoV-2 inhibition assay: Quantification of SARS-CoV-2 nucleoprotein positive cells

Sixteen hours before infection, 3 × 10⁴ IGROV-1 cells per well were seeded in a μClear black 96-well plate (Greiner Bio-One). Cells were incubated with serially diluted VHH-Fc for 1 h at 37° C and infected with the indicated SARS-CoV-2 strains. After 24 hours, cells were fixed with 4% PFA (Electron Microscopy Sciences, cat# 15714-S). The cells were then intracellularly stained with anti-SARS-CoV-2 nucleoprotein (N) antibody NCP-1 (0.1 μg/mL) as described⁷. The staining was carried out in PBS with 0.05% saponin 1% BSA, and 0.05% sodium azide for 1 hour. Cells were then washed twice with PBS and stained with anti-IgG Alexa Fluor 488 (dilution 1:500, Invitrogen; cat# A11029) for 30 minutes before being washed twice with PBS. Hoechst 33342 (Invitrogen, cat# H3570) was added during the final PBS wash. Images were captured using an Opera Phenix high-content confocal microscope (PerkinElmer). The N-positive area and the number of nuclei were quantified using Harmony Software v4.9 (PerkinElmer). The percentage of inhibition of infection was calculated using the N-positive area with the following formula: 100 × (1 – [value with VHH – value in “non-infected”]/[value in “no VHH” – value in “non-infected”]). The half-maximal inhibitory concentration (IC50) was calculated with a reconstructed curve using the percentage of inhibition at each concentration.

SARS-CoV-2 inhibition assay: Viral replication

For the evaluation of SARS-CoV-2 replication, the number of SARS-CoV-2 copies was measured by reverse transcription quantitative polymerase chain reaction (RT-qPCR). Individual VHHs were pre-incubated with Vero E6 cells (ten specified concentrations) 2 hours prior infection. Each plate included PBS (2 µL, negative control) (or PBS 0.25 % DMSO) and remdesivir (25 µM; SelleckChem) as a positive control. After the pre-incubation time, individual VHHs and remdesivir were removed, and the cells were exposed to the different variants of SARS-CoV-2 (at a pre-established multiplicity of infection: MOI of 0.07 (Delta, XBB.1.16.1), 0.27 (BQ.1.1, BA.1), and 0.33 (XBB.1.5) PFU/Vero E6 cells). After a one-hour adsorption at 37 ° C, the supernatant was aspirated and replaced with 2% FBS/DMEM media containing the respective VHHs or remdesivir at the indicated concentrations. The cells were then incubated at 37 ° C for 48 (B.1.617.2; BQ.1.1; XBB.1.16.1; BA.1) or 66 hours (XBB.1.5). The MOI and incubation time were defined previously to obtain the same rate of viral production after infection. Supernatants were collected and heat inactivated at 80° C for 20 minutes. The Luna Universal One-Step RT-qPCR Kit (New England Biolabs) was used for the detection of genomic SARS-CoV-2 RNA through RT-qPCR using a thermocycler (QuantStudio 6 thermocycler; Applied Biosystems). Specific primers targeting the N-region viral gene: Forward: 5’ TAATCAGACAAGGAACTGATTA Reverse: 5’ CGAAGGTGTGACTTCCATG were used. Cycle threshold (Ct) were determined by the second derivative maximum method in the QuantStudio software v2.6.0. The quantity of viral genomes is expressed as Ct and was normalized against the Ct values of the negative (PBS) and positive controls (remdesivir 25 µM). Curve fits and IC50 values were obtained in Prism using the variable Hill slope model.

Cell viability assay

Cell viability assays were performed in compound-treated cells using the CellTiter-Glo assay according to the manufacturer’s instructions (Promega) and a Berthold Centro XS LB960 luminometer (MikroWin 2000). 3,000 cells/well of Vero E6 were seeded in white with clear bottom 384-well plates. The following day, compounds were added at concentrations indicated. PBS only and 10 μM camptotecin (Sigma-Aldrich) controls were added in each plate. After 48 h incubation, 10 μL of CellTiter Glo reagent was added in each well and the luminescence was recorded using a luminometer (Berthold Technologies) with 0.5 sec integration time. Raw data were normalized against appropriate negative (0 %) and positive controls (100 %) and are expressed in % of cytotoxicity.

Crystallization and structure determination

Purified hACE2 was incubated overnight at 4 ° C with VHHs B07 and B10 at final concentrations of 23 µM, 129 µM, and 129 µM, respectively. Then, a SEC step was performed to isolate the complex peak from unbound fractions. The complex was purified to 7.9 mg/mL and used for an initial crystallization screening trials by flowing the procedure established at the Crystallography Core Facility of the Institut Pasteur⁶⁹. Briefly, crystallization drops were generated by mixing equal volumes (200 nL) of the protein complex and reservoir solution in the format of 96 Greiner plates, using a Mosquito robot, and monitored for crystal growth by a Rock-Imager. Crystals grew at 18° C in 10% (w/v) 2-propanol, 20% (w/v) Polyethylene glycol 4000, and 0.1 M HEPES pH 7.5. For cryoprotection, the crystals were soaked in crystallization solution supplemented with 20% glycerol before flash-frozen under liquid nitrogen. The crystals obtained were tested at beamlines PX1 and PX2 at SOLEIL synchrotron (Saint Aubin, France). The best dataset was collected at beamline PX2-A on a EIGER X 9 M area detector at a resolution of 2.7 Å, and was indexed, integrated, scaled, and merged using XDS⁷⁰ and AIMLESS⁷¹. The high-resolution limit was determined using CC1/2-based cutoffs of 30%⁷². The structures of molecules were then determined by molecular replacement with PHASER⁷³ using the search models of hACE2 from PDB code 1R42 and alpha-fold models for VHH B07 and VHH B10⁷⁴. The asymmetric unit (au) contained 3 complexes of hACE2-VHH B07-VHH B10. After reconstruction of subdomains of hACE2 and of CDRs of VHHs, the models were alternatively manually corrected using COOT⁷⁵ and refined using non-crystallographic symmetry and TLS parameterization (Translation/Libration/Screw)⁷⁶ with BUSTER-TNT⁷⁷. The final R and free R factors are 21.9% and 25.4%, respectively. The final refined models were analyzed with MolProbity⁷⁸. The data collection and refinement statistics are listed in Supplementary Table 2.

Structural data analysis

The figures of the structures were prepared using the PyMOL Molecular Graphics System, version 2.0.0 (Schrödinger LLC) (http://pymol.sourceforge.net). The polar interactions (hydrogen bonds and salt bridges), the contact residues (van der Waals) and the accessible (ASA) and buried (BSA) surface areas were assessed using ‘Protein interfaces, surfaces and assemblies’ service PISA at the European Bioinformatics Institute⁷⁹ (http://www.ebi.ac.uk/pdbe/prot_int/pistart.html). The intermolecular interactions were computed using a maximal cut-off distance of 3.9 Å for polar contacts (PISA distance cut-off) and 4.75 Å for van der Waals contacts (PyMol distance cut-off).

For the final analysis of contacts between molecules (au containing 3 complexes), we used the chain A for hACE2, chain D for VHH B10 and chain E for VHH B07 (chains D and E have lower B-factors compared to other chains of VHHs, indicative of a more orderly disposition of atoms within the crystal). However, for all the complexes the sites of interactions between hACE2 and VHHs are ordered.

Electrostatic surfaces were computed using the PDB2PQR⁸⁰ and APBS⁸¹calculations software. Sequence alignments were performed using clustal W omega⁸² and plotted using ESPript server⁸³. The accession numbers of ACE2 sequences are: Human (UniProt ACE2_HUMAN); Hamster (UniProt A0A1U7QTA1_MESAU); Mouse (UniProt ACE2_MOUSE). The VHH sequences were analysed by Abysis (http://www.abysis.org/abysis) and IMGT⁸⁴ (https://www.imgt.org) websites for mapping CDR/framework regions according to Kabat⁸⁵ and IMGT conventions, respectively. The analysis of the putative germline and somatic maturation events was done with the IMGT website.

Immunofluorescence imaging of primary human nasal epithelium cells and K18-hACE2 mouse tissues

MucilAir^TM, reconstructed human nasal epithelial cells (hNECs), previously differentiated for 4 weeks, were cultured in 700 µL MucilAir^TM media on the basal side of the air/liquid interface (ALI) cultures as described⁸⁶. For imaging, cells were fixed on the apical and basal sides with 4% PFA for 15 minutes. Cells were stained O/N at 4° C with anti-hACE2 VHH-B07-Fc at 6 µg/mL revealed with a Goat anti-Human Alexa Fluor-594 secondary Ab (Jackson Immuno Research, #109-587-003, 1:200), added with Phalloidin Atto-647 (Sigma-Aldrich 65906, 1:400) (F-actin staining), and DAPI (Thermo Fisher, 1 mg/mL) (Fig. 6d) or they were co-stained with anti-hACE2 VHH B07-Fc and rat anti-alpha-tubulin (ThermoFisher, # MA1-80017, 1:200) revealed with a Goat anti-Human Alexa Fluor-594 and a Donkey anti-Rat Alexa Fluor 488 (Thermo Fisher scientific, #A-21208, 1:200) secondary antibodies (Supplementary Fig. 9c).

For immunofluorescence experiments on whole-mount preparations of mouse organs, following euthanasia, samples of nasal turbinates and lungs of K18-hACE2 mice were collected and fixed in phosphate-buffered saline (PBS) containing 4% paraformaldehyde for 24 h at 4° C. Then, nasal respiratory epithelium, olfactory epithelium and lung bronchioli were finely excised and post-fixed in PBS containing 4% paraformaldehyde for 1 hour, rinsed in PBS and immersed in blocking buffer (20% normal goat serum (NGS) and 0.3% Triton X-100 in PBS) for 1 hour at room temperature, and then incubated overnight at 4 ° C in primary antibodies (in PBS containing 2% bovine serum albumin (BSA)): Mouse anti-SARS-CoV-2 nucleoprotein (N) antibody (Genetex GTX36802, 1:100) and Rat anti-alpha-tubulin (1:200). After rinsing in PBS, samples were incubated for 2 hours at room temperature in PBS-2% BSA with the appropriate secondary antibodies: Goat anti-mouse Alexa-Fluor 555 (Abcam; ab150114, 1:200); Donkey anti-Rat Alexa-Fluor 488, Phalloidin Atto 647 (Sigma-Aldrich) (F-actin staining) and DAPI (Thermo Fisher Scientific, 1 mg/mL) were added in secondary antibodies mix. Fluorescent immunolabellings were imaged with a LSM-700 confocal microscope and ZEN software 2012 (Zeiss) and analyzed using ImageJ v2.16.0 (Fiji) software as described^37,86.

SARS-CoV-2 infection and treatment in K18-hACE2 mouse model

B6.Cg-Tg(K18-ACE2)2Prlmn/J mice (stock #034860) were imported from The Jackson Laboratory and bred at the Institut Pasteur under strict specific pathogen-free conditions. Infection studies were performed on 10−13 week-old male and female mice, in animal biosafety level 3 (BSL-3) facilities at the Institut Pasteur. All animals were handled in strict accordance with good animal practice. Mice were housed in groups of up to 5 individuals, in negative pressure isolators, under 12 h:12 h light-dark cycle, at 24 + /-1.5 ° C temperature and 50-70% relative humidity. They were fed standard chow ad libitum, with autoclaved water. Animal work was approved by the Animal Experimentation Ethics Committee (CETEA 89) of the Institut Pasteur (project dap 210050), and authorized by the French Ministry of Research (under project 31816) before the experiments were initiated. Twenty-four hours before infection, anesthetized (ketamine/xylazine) mice were administered intra-nasally (i.n.) with either 0.9% NaCl (40 µL), control VHH (7 mg/kg or 5 mg/kg in 20-30 µL) or B07-Fc (7 mg/kg in 50 µL or 5 mg/kg in 30-50 µL). On the next day, they were again anesthetized (ketamine/xylazine) and inoculated i.n. with 1 × 10⁵ PFU of SARS-CoV-2 XBB.1.5 (hCoV-19/France/PDL-IPP58867/2022; 10 μL/nostril) or 1×10⁴ PFU of SARS-CoV-2 D614G (hCoV-19/France/GE1973/2020; GISAID ID: EPI_ISL_414631; 15 μL/nostril). Clinical signs of disease (ruffled fur, hunched posture, reduced mobility and breathing difficulties) and weight loss were monitored everyday (except weekend). Each of the four criteria received a 0–2 score and was added to a global score. Mice were euthanized when reaching 25% body weight loss or a global score of 8 or euthanized at specific time points for tissue collection. On day 3 or 6 post-infection, mice were euthanized by cervical dislocation. The trachea and upper lung lobes were fixed by submersion in 10% phosphate buffered formalin for 24-36 hours prior to removal from the BSL3 for processing and transfer in 70% ethanol. Lower lung lobes (and nasal turbinates in a subset of mice) were dissected and frozen at -80° C. Samples were homogenized in 500 μL of PBS using lysing matrix M (MP Biomedical) and a MP Biomedical FastPrep 24 Tissue Homogenizer. For lung viral load measurement, genomic viral RNA was extracted using an extraction robot IDEAL-32 (IDsolutions) and the NucleoMag Pathogen extraction kit (Macherey Nagel). Viral RNA quantification was performed by quantitative reverse transcription PCR (RT-qPCR) using the IP4 set of primers and probe (nCoV_IP4-14059 Fw GGTAACTGGTATGATTTCG and nCoV_IP4-14146 Rv CTGGTCAAGGTTAATATAGG giving a 107 bp product, and nCoV_IP4-14084 Probe(+)TCATACAAACCACGCCAGG [5’]Hex [3’]BHQ-1) and the Luna Universal Probe One-Step RT-qPCR Kit (NEB). Serial dilutions of a titrated viral stock were analyzed simultaneously to express viral loads as PFU equivalents (eqPFU) per gram of tissue.

For plaque assay, 10-fold serial dilutions of samples in DMEM were added onto VeroE6 monolayers in 24 well plates. After one-hour incubation at 37° C, the inoculum was replaced with equivalent volume of 5% FBS DMEM and 2% Carboxymethylcellulose. Three days later, cells were fixed with 4% formaldehyde, followed by staining with 1% crystal violet to visualize the plaques.

Measurement of VHH concentration in lungs homogenate by ELISA

sACE2 (2 µg/mL) were coated on polystyrene plate (Thermo Fisher Scientific) over night at 4° C. After washing with PBS/Tween 0.1%, plates were incubated with B07-Fc (from 0.07 to 10 ng/mL) or homogenate of lower lung lobes in PBS 0.5% Gelatin, 0.1% Tween-20 (PGT solution) for 90 min at 37° C. A negative control (sACE2 alone) was prepared under the same conditions. Then, plates were incubated at 37° C with Peroxydase Affinipure rabbit anti Human IgG (Jackson ImmunoResearch, AB_2339647, 1/1000) in PGT for 45 min and treated with ortho-phenylenediamine solution containing 0.1% hydrogen peroxide (H2O2) for 5 min. The reaction was stopped by adding chlorohydric acid 3 M (HCl) and the absorbance of the wells was measured at 492nm in a spectrophotometer-microplate reader (Sunrise, Tecan, software XFLUOR4 v4.51).

SARS-CoV-2 infection and treatment in Syrian golden hamster

The animals used were male Syrian golden hamsters (Mesocricetus auratus, strain RjHan:AURA), 6 week of age (weight between 70−90 grams), obtained from Janvier Laboratories. Golden hamsters were housed and manipulated in class III safety cabinets in the Institut Pasteur animal facilities accredited by the French Ministry of Agriculture. Animal work was approved by the Animal Experimentation Ethics Committee (CETEA) of the Institut Pasteur (project dap 210011) and authorized by the French legislation (project #21045) in compliance with the European Communities Council Directives (2010/63/UE, French Law 2013-118, February 6, 2013) and according to the regulations of Institut Pasteur Animal Care Committees before the experiments were initiated. Hamster were housed in groups of up to 4 individuals, in negative pressure isolators, under 14 h:10 h light-dark cycle, at 22 + /-2° C temperature and 55 % ( + /- 10 %) relative humidity. Anesthetized animals received intranasal inoculation (5 mg/kg) of VHH-B07-Fc or VHH-ctl-Fc. Twenty-four hours after inoculation, hamsters were infected intranasally (i.n.) with 10⁴ PFU of SARS-CoV-2 XBB.1.16.1 as previously described⁶⁷.

All hamsters were followed up daily. At day 3 post-infection, animals were euthanized with an excess of anesthetics (ketamine, xylazine and laocaine) (AVMA Guidelines 2020), and samples collected. 100 mg lung fragments were ground in Lysing Matrix X tubes (MP Biomedicals) using a grinder at a speed of 6.5 m/s for 30 seconds and centrifuged at 2,000 g for 10 minutes (4° C) to clarify the supernatant.

For the evaluation of the number of SARS-CoV-2 copies in tissues by RT-qPCR, 100 µL of each grinding solution was mixed with 500 μL of Trizol to deactivate virulence. The total RNA was then extracted using the Direct-zol RNA MiniPrep Kit (Zymo Research) and quantified. The presence of genomic SARS-CoV-2 RNA in these samples was evaluated by one-step RT-qPCR in a final volume of 5 μL per reaction in 384-well PCR plates using a thermocycler (QuantStudio 6 thermocycler v2.6.0; Applied Biosystems). The primers used targeted N-region viral gene: Forward: 5’ TAATCAGACAAGGAACTGATTA Reverse: 5’ CGAAGGTGTGACTTCCATG. Viral load quantification (expressed as RNA copy number/g lung) was assessed by linear regression using a standard curve of seven known quantities of RNA transcripts containing the N-region viral gene. All collected data were plotted and analyzed using GraphPad Prism software. Mann-Whitney tests were performed for statistical analysis.

Statistical analysis

Figures were generated using GraphPad Prism 10 (GraphPad software). Statistical analysis was conducted using GraphPad Prism 10. Statistical significance between different groups was calculated using the tests indicated in each figure legend.

Reporting summary

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