Ethics statement for animal care and informed consent
Two young alpacas (Vicugna pacos) half-siblings—a 19-month-old male named “Puta” and a 19-month-old female named “Christy”—were immunized. Veterinarians of the KYODOKEN Institute for Animal Science Research and Development (Kyoto, Japan) bred, maintained health, recorded conditions, and performed the immunization studies by adhering to the published Guidelines for Proper Conduct of Animal Experiments by the Science Council of Japan. The KYODOKEN Institutional Animal Care and Use Committee approved the protocols for these studies (approval number 20200312) and monitored health conditions. The veterinarians immunized the alpacas with antigens and collected blood samples under anaesthesia. Nasal swab specimens were collected from patients who had agreed to use them for these studies. The Committee of Shizuoka City Hospital approved the protocols (approval number 20220128).
Immunization and library generation
Immunized antigens were purified recombinant SARS-CoV-2 spike complexes: the extracellular domain of the original Wuhan-1 protein (GenBank: QHD43416) with or without the D614G mutation that carried a maintained or mutated furin cleavage site—N679SPRRA or IL680. The protein mixture emulsified in complete Freund’s adjuvant was subcutaneously injected into the two alpacas up to 9 times at 2-week intervals. Blood samples were collected from the jugular vein; peripheral blood mononuclear cells (PBMCs) were obtained with a sucrose density gradient using Ficoll69 (Nacalai Tesque, Kyoto, Japan). The PBMC samples were washed with PBS and suspended in RNAlater solution (Thermo Fisher Scientific K.K., Tokyo, Japan). Total RNA was isolated from the PBMC samples (Direct-Zol RNA MiniPrep: Zymo Research, Irvine, CA).
Complementary DNA was synthesized from 1 μg of the total RNA as a template with random hexamer primers and using SuperScript II reverse transcriptase (Thermo). Coding regions of the heavy-chain variable domains were amplified using LA Taq polymerase (TAKARA Bio Inc., Shiga, Japan) with two PAGE-purified primers (CALL001: 5′-GTCCTGGCTGCTCTTCTACAAGG-3′ and CALL002: 5′-GGTACGTGCTGTTGAACTGTTCC-3′). The amplified coding gene fragments of heavy-chain variable domains were separated on a 1.5% low-temperature melting agarose gel (Lonza Group AG, Basel, Switzerland). Approximately 700 base pair bands corresponding to the heavy-chain only immunoglobulin were extracted (QIAquick Gel Extraction Kit: Qiagen K.K., Tokyo, Japan). Nested PCR was performed to amplify coding genes of VHH domains using the VHH-PstI-For and VHH-BstEII-Rev primers and subcloned into the pMES4 phagemid vector12,70 (GeneArt DNA Synthesis: Thermo). Electroporation-competent Escherichia coli TG1 cells (Agilent Technologies Japan, Ltd., Tokyo, Japan) were transformed with the ligated plasmids under chilled conditions (Bio-Rad Laboratories, Inc., Hercules, CA). Colony-forming units of the libraries were checked with limiting dilution to maintain >107 per microlitre. Colonies from 8 ml of cultured cells were harvested, pooled, and reserved in frozen glycerol stocks as parent libraries.
Plasmid construction for protein expression
The gene encoding the extracellular domain of the SARS-CoV-2 spike protein (residues 31–1213) was codon-optimized and synthesized into the pcDNA3.1(+) vector (Thermo) with an N-terminally modified IL-2-derived signal peptide71 (ILco2: MRRMQLLLLIALSLALVTNS); proline substitutions at residues K986P and V987P; and the C-terminal T4-phage fibritin trimerization domain (foldon) following a 6×His-tag34,72,73,74. The expression vector of the SARS-CoV-2 S2 domain (residues 744–1213) was constructed by removing an N-terminal part of the extracellular domain of the SARS-CoV-2 spike (residues 31–743) and subcloned into the pcDNA3.1(+) vector. The expression vectors of the full-length SARS-CoV-2 spike including variants and the human serine protease TMPRSS2 with a C-terminal C9-tag (TETSQVAPA) were acquired from AddGene (Summit Pharmaceutical International, Tokyo, Japan). The genes encoding the extracellular domain of human ACE2 (residues 1–614) and the endemic human coronavirus HCoV-OC43 (residues 1–1322) were codon-optimized, synthesized with a C-terminal 6×His-tag and subcloned into the pcDNA3.1(+) vector. The whole gene of the human apolipoprotein B messenger-RNA-editing enzyme catalytic polypeptide-like (APOBEC) 3G (A3G)26 was also codon-optimized, synthesized with a C-terminal 6×His-tag, and subcloned into the pcDNA3.1(+) vector.
For structural analyses, the sequence encoding the spike ectodomain (residues 1–1208) with proline substitutions, a “GSAS” substitution at the furin cleavage site (residues 682–685)75, and the C-terminal foldon trimerization motif followed by an 8×His-tag was cloned into the pcDNA3.1(+) expression vector. Furthermore, the D614G mutation in the spike protein was introduced by the inverse PCR method. For cryo-EM, recombinant spike proteins were transiently expressed in Expi293-F cells (Thermo) maintained in HE400AZ medium (Gmep, Inc., Fukuoka, Japan). The expression vector was transfected using a Gxpress 293 Transfection Kit (Gmep) according to the manufacturer’s protocol. The culture supernatants were harvested 5 days post-transfection. The C-terminally 6×His-tagged spike proteins were purified using a nickel Sepharose 6 FF column (Cytiva) and size exclusion chromatography using a Superdex200increase 10/300 GL column (Cytiva) with buffer containing 50 mM HEPES (pH 7.0) and 200 mM NaCl.
The purified proteins were (i) the extracellular domain of the SARS-CoV-2 spike; (ii) only the S2 domain of the SARS-CoV-2 spike; (iii) the extracellular domain of the seasonal cold coronavirus spike of HCoV-OC43; (iv) A3G; and (v) homemade IgM coupled to N-hydroxysuccinimide (NHS)-activated magnet beads (Dynabeads, Thermo). One round of biopanning was performed using different protein-coated magnet beads in 50 mM phosphate buffer (pH 7.4) containing 1% n-dodecyl-β-D-maltopyranoside (DDM: Nacalai), 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propane sulfonate (CHAPS: Nacalai), 0.001% cholesterol hydrogen succuinate (CHS: Tokyo Chemical Industry Co., Ltd. (TCI), Tokyo, Japan), 0.1% LMNG (Anatrace, Maumee, OH) and 500 mM NaCl. After 3 washes with the same buffer, the remaining phages bound to the washed beads were eluted with a trypsin-ethylenediaminetetraacetic acid (EDTA: Nacalai) solution at room temperature for 30 min. The elution was neutralized with a PBS-diluted protein inhibitor cocktail (cOmplete, EDTA-free, protease inhibitor cocktail tablets: Roche Diagnostics GmbH, Mannheim, Germany) and used to infect electroporation-competent cells. The infected cells were cultured and selected in LB broth Miller containing 100 μg ml–1 ampicillin (Nacalai) at 37 °C overnight. The selected phagemids were collected using a QIAprep Miniprep Kit (Qiagen).
Sequence analysis of nanobody libraries
The VHH-coding regions within parent libraries and 1-round target-enriched sublibraries were PCR amplified and purified using AMPure XP beads (Beckman Coulter, High Wycombe, UK). Then, dual-indexed libraries were prepared and sequenced on an Illumina MiSeq (Illumina, San Diego, CA) using a MiSeq Reagent Kit v3 with paired-end 300 bp reads76 (Bioengineering Lab. Co., Ltd., Kanagawa, Japan). Approximately 100,000 paired reads of each library were generated. The raw data of reads were trimmed of the adaptor sequence using cutadapt v1.1877, and low-quality reads were subsequently removed using Trimmomatic v0.3978. The remaining paired reads were merged using fastq-join79 and then translated to the amino acid sequences using EMBOSS v6.6.0.080. Finally, unique amino acid sequences in each library were counted using a custom Python script combining seqkit v0.10.181 and usearch v.1182. Enrichment scores of each clone were analysed by calculating the P-value of χ2 tests between the existing ratios among the different sublibraries.
We clustered nanobody sequences with no more than six Damerau–Levenshtein distances83,84. We found eight and ten clusters—from Cristy’s and Puta’s libraries, respectively—significantly enriched in only the SARS-CoV-2 panned sublibraries. We synthesized the top read clones in each family as linked dimers. Among them, four (C17, C49, C116 and C246) out of eight and five (P17, P86, P158, P334 and P543) out of ten dimer clones were expressed and further analysed.
Each selected amino acid sequence was connected with a (GGGGS)4 linker as a tandem dimer; coding genes of these and of the previously reported SARS72 dimer, mNb6 dimer and Ty1 monomer were codon-optimized and synthesized (Eurofins Genomics Inc., Tokyo, Japan). The synthesized genes were subcloned in the pMES4 vector to express N-terminal PelB signal peptide-conjugated and C-terminal 6×His-tagged or Flag-tagged nanobodies into the bacterial periplasm. These gene constructs were transformed into BL21(DE3) E. coli cells (BioDynamics Laboratory Inc., Tokyo, Japan) and plated on LB agar with ampicillin, which were incubated at 37 °C overnight. Colonies were picked and cultured at 37 °C to reach an OD of 0.6 AU; the cells were cultured at 37 °C for 3 h or at 28 °C overnight with 1 mM IPTG (isopropyl-β-D-thiogalactopyranoside: Nacalai). Cultured cells were collected by centrifugation. Nanobodies were eluted from the periplasm by soaking in a high osmotic buffer containing 200 mM Tris, 0.5 mM EDTA and 500 mM sucrose (pH 8.0) at 4 °C for 1 h. They were incubated with 2× volumes of a diluted buffer containing 50 mM Tris, 0.125 mM EDTA, and 125 mM sucrose (pH 8.0) with a trace amount of benzonase nuclease (Merck KGaA, Darmstadt, Germany) at 4 °C for 45 min, and the supernatants were centrifuged (20,000 × g, 4 °C for 10 min). The supernatants were sterilized with the addition of gentamicin (Thermo) and passed through a 0.22-μm filter (Sartorius AG, Göttingen, Germany). The filtered supernatants of Flag-tagged trimer were used for ELISA and His-tagged dimers were applied to a HisTrap HP nickel column (Cytiva) equipped on an ÄKTA purifier HPLC system (Cytiva) and washed; the bound His-tagged protein was eluted with 300 mM imidazole. The elution was concentrated with a VIVAspin 3000-molecular weight cut off (MWCO) filter column (Sartorius) and applied to a Superdex75increase 10/300 GL gel-filtration column (Cytiva) equipped on an ÄKTA pure HPLC system (Cytiva) to obtain the dimer fractions and exclude cleaved monomers and imidazole. Purity was measured via Coomassie Brilliant Blue (CBB) staining.
Antibodies used for western blotting and cell staining were anti-His (rabbit polyclonal PM032: Medical and Biological Laboratories Co., Ltd. (MBL), Nagoya, Japan), anti-His (rabbit monoclonal EPR20547: ab213204, Abcam), anti-Flag (mouse monoclonal M2: Thermo) and anti-rhodopsin C9 (TETSQVAPA) (mouse monoclonal 1D4, sc-57432: Santa Cruz Biotechnology Inc., CA). Horseradish peroxidase (HRP)-linked secondary antibodies included anti-mouse IgG (sheep polyclonal, NA931: GE Healthcare, Buckinghamshire, UK), anti-rabbit IgG (sheep polyclonal, NA934: GE Healthcare), anti-alpaca IgG85 (goat polyclonal, 128-035-232: Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), Alexa Fluor 488 goat anti-mouse IgG (rabbit polyclonal, P0449: Dako, Glostrup, Denmark) and Alexa Fluor 594 goat anti-rabbit IgG (Dako).
Nasal swab specimens and western blotting
Nasal swab specimens were collected on 1 February 2022 (I and J), 3 February 2022 (K to P), 10 February 2022 (A to H) and 23 March 2022 (Q to V) from emergency fever patients. All specimens were tested to be SARS-CoV-2 positive by PCR. Sixteen samples (A to D, I to R, T and V) were confirmed to be SARS-CoV-2 positive and 6 samples (E to H, S and U) were not confirmed to be positive by the PCR test.
Samples were incubated at 37 °C for 30 min (for SARS-CoV-2 spike variants and swab specimens) or boiled for 2 min (for nanobodies) with Laemmli’s SDS sample buffer containing 2.5 mM Tris (pH 6.8), 2% SDS, 10% glycerol, 0.001% bromophenol blue and 13.3 mM dithiothreitol (DTT). Samples were electrophoretically separated on 5–20% or 15–25% gradient polyacrylamide gels and electrophoretically transferred onto polyvinylidene difluoride (PVDF) membranes (Immobilon-P: Millipore, Billerica, MA). Blotted membranes were incubated overnight at 4 °C with the C9 antibody (the dilution ratio was 1:5000) or the C-terminally 6×His-tagged homodimer of nanobodies in Tris-buffered saline (TBS, pH 7.4) containing 0.005% Tween 20 (TBST) and 5% skim milk—the concentrations of P158, P334 and P543 were 0.2, 1.0 and 0.4 μg ml–1, respectively. For testing them for the Omicron BA.1 variant, concentrations of P86 and C246 were 0.4 μg ml–1, and of the other nanobodies (C17, P17, C49 and C116) were 5.0 μg ml–1. In the case of nanobody-based blotting, after 3 washes with TBST, the membranes were incubated with 1:5000-diluted anti-His-tag antibody (MBL) in TBST containing 5% skim milk at room temperature for 1 h. The membranes were soaked with 1:5000-diluted HRP-conjugated anti-rabbit or anti-mouse IgG secondary antibodies (GE Healthcare) in TBST containing 5% skim milk for 30 min at room temperature. After 3 washes with TBST, reactive protein bands were visualized using an ECL Plus system (Cytiva). For CBB R-250 staining, a Rapid Stain CBB Kit (Nacalai) was used according to the manufacturer’s protocol.
Column chromatography for protein purification
HEK293T cells expressing the SARS-CoV-2 S2 domain or the extracellular domain of the HCoV-OC43 spike were cultured in serum-free Opti-MEM (modified Eagle’s medium: Thermo) containing 1% penicillin and streptomycin. After 48 h, the culture supernatants were centrifuged, filtered, and concentrated with VIVAspin 20 size exclusion columns (30,000-MWCO). Pellets of HEK293T cells expressing the extracellular domain of the SARS-CoV-2 spike were suspended in 50 mM phosphate buffer (pH 7.4) containing 1% DDM, 0.1% CHAPS, 0.001% CHS, 0.1% LMNG and 500 mM NaCl. The suspension was passed through a 21-gauge needle (Terumo Co., Tokyo, Japan) several times, trace amounts of benzonase nuclease (Merck) were added, and the lysates were incubated at 4 °C overnight. The cell lysate was cleared by centrifugation and filtration. The C-terminal 6×His-tagged spike protein was purified through a HisTrap HP 1 ml column equipped on an ÄKTA pure HPLC system (Cytiva) under step-by-step elution conditions using running (20 mM imidazole) and eluting (500 mM imidazole) phosphate buffers containing 20 mM phosphate, 500 mM NaCl, 0.5% DDM, 0.1% CHAPS, 0.001% CHS, and 0.1% LMNG (for spike proteins from the cell lysate) (pH 7.4). Elution fractions of 6×His-tagged proteins were identified via western blotting. Anti-6×His-tagged antibody-positive fractions were gathered and concentrated via VIVAspin 6 size exclusion columns (30,000-MWCO) to reach a volume under 0.5 ml. Size exclusion chromatography experiments were performed on a Superose6increase 10/300 GL column (Cytiva) with an ÄKTA pure HPLC system (Cytiva). Sample volumes were ~0.5 ml; injected samples were separated in PBS or PBS-LMNG (for SARS-CoV-2 spike protein) in a chilled chamber with a flow rate of 0.1 ml min–1. Chromatograms were monitored at 280 nm using a UV spectrophotometer. Elution fractions were identified via western blotting, gathered, and concentrated via VIVAspin 6 size exclusion columns (30,000-MWCO). Protein concentrations were measured with a NanoDrop 2000c spectrophotometer (1 mAU at 280 nm was equivalent to 1 mg ml–1) (Thermo).
Cell culture and transfection
HEK293T, K562 and HOS (human osteosarcoma) cells were grown in Dulbecco’s modified Eagle’s medium (DMEM: Invitrogen) supplemented with 10% foetal bovine serum (FBS) and antibiotics (1% penicillin and streptomycin). The cells were cultured in a humidified incubator with 5% CO2 at 37 °C. The human ACE2 and TMPRSS2 genes were transduced into K562 and HOS cells with a lentivirus. The lentiviral vector pWPI-IRES-Bla-Ak-ACE2-TMPRSS2 was acquired from AddGene (plasmid #154983). DT40 cells (from the Japanese Collection of Research Bioresources Cell Bank, Osaka, Japan) were grown in RPMI1640 medium (Invitrogen) supplemented with 10% FBS, 10% chicken serum, 1% penicillin and 1% streptomycin.
Measuring titres and nanobody-based sandwich ELISA
Two micrograms of the recombinant extracellular domain of the SARS-CoV-2 spike or 10 μg of the homodimer of nanobodies was diluted with 10 ml of 0.1 M carbohydrate buffer (pH 9.8). Each well of a MaxiSorp 96-well plate (Thermo) was coated with 100 μl of the diluent at 4 °C overnight. To measure the titres of the immunized two alpacas, the wells were washed 3 times with high-salt PBS-LMNG (500 mM NaCl and 0.001% LMNG) and blocked with the high-salt PBS-LMNG containing 1% FBS at room temperature for 1 h. For screening conditions with the nanobody-based sandwich ELISA for the SARS-CoV-2 spike, the wells were washed with PBST (0.05% Tween 20, pH 7.4) and were blocked at room temperature for 1 h with five kinds of blocking solutions: 1×Carbo-Free blocking solution (Vector Laboratories, Inc., Burlingame, CA); 5% bovine serum albumin (BSA) (Sigma-Aldrich) in PBST; 5% skim milk in PBST; 3% casein (Merck) in 20 mM TBS (pH 11.0); or 5% polyvinylpyrrolidone (PVP) average molecular weight 10,000 (Sigma-Aldrich) in PBST. For measuring titres, 10 μl of a 0.1% dilution of serum from immunized alpacas in PBST was added to the well and incubated at room temperature for 1 h. After 3 washes with PBST, HRP-conjugated anti-alpaca VHH antibody (Jackson) at a dilution of 1:5000 was reacted at room temperature for 30 min. For screening of nanobody-based sandwich ELISA conditions, 100 μl of 1% (w/v) of the extracellular domain of SARS-CoV-2 spike (1 μg 100 μl–1) in high-salt PBS-LMNG was added to each well and captured at room temperature for 1 h. The well was washed 3 times with the high-salt PBS-LMNG; each 30 ng of biotin-conjugated nanobody in 100 μl (0.03% w/v) of PBS-LMNG was soaked at room temperature for 30 min. After 3 washes with PBS-LMNG, 100 μl of HRP-conjugated streptavidin (TCI) at a dilution of 1:5000 in PBS-LMNG was incubated at room temperature for 30 min.
The amounts of remaining HRP conjugates after 3 washes with the buffer were measured with the addition of 100 μl of 50 mM phosphate-citrate buffer (pH 5.0) containing 0.4 μg of o-phenylenediamine dihydrochloride (OPD) (Sigma-Aldrich) to develop at room temperature for 30 min, after which the reaction was stopped with the addition of 10 μl of 5 M sulfuric acid (H2SO4). Each well was read at an optical density (OD) of 450 nm using a microplate reader (Bio-Rad).
Each well of a MaxiSorp 96-well plate was coated with 100 μl of the diluent containing 10 μg of P158 diluted with 10 ml of 0.1 M carbohydrate buffer, pH 9.8. The P158-coated wells were blocked with 3% BSA in PBST. Nasal swab specimens were dissolved with 1% DDM and benzonase nuclease. Lysed samples in PBS were diluted with water to contain the final 20 mM NaCl concentrations. Dissolved and diluted nasal swab samples containing 1% BSA were added to the wells. To increase sensitivity, the Flag-tagged P86 trimer nanobody was used as a primary antibody at the concentration of 30 ng 100 μl–1; the 1:5000-diluted anti-Flag antibody (secondary antibody) and the 1:5000-diluted HRP-conjugated anti-mouse antibody (final antibody) were used for detection.
Antigen test kits detecting the SARS-CoV-2 spike based on nitrocellulose lateral flow assays were developed by Yamato Scientific Co., Ltd. (Tokyo, Japan). Briefly, purified P158 (330 ng μl–1) was lined ~1 mm wide on an IAB90 nitrocellulose membrane (Advantech, Toyo Roshi Kaisha, Ltd., Tokyo, Japan). The membrane was soaked in 1×Carbo-Free blocking solution at room temperature for 1 h and air-dried. Purified P86 or P543 nanobodies (500 ng μl–1) were amine coupled to 30-nm diameter Estapore beads (Sigma-Aldrich) according to the manufacturer’s protocol. Glass fibre paper (GF/DVA: Cytiva) was soaked in 1×Carbo-Free blocking solution containing 0.005% (w/v) nanobody-coupled Estapore beads for saturation and then dried under vacuum conditions. The lined nitrocellulose membrane and the bead-absorbed glass fibre paper were overlaid on a backing sheet (Cytiva). CF4 paper (Cytiva) was used for both the sample pad and the absorbance pad; they were overlaid on the prepared backing sheet as well. The four-layered sheet was cut 5 mm wide and housed in black cases: the cassettes were stored in sealed packages with silica gels. When dried, 150 μl of sample was spotted onto the sample pad; the kits were photographed under a 315 nm UV lamp for an arbitrary amount of time.
Antigen test kits detecting the SARS-CoV-2 spike from nasal swab specimens have been further modified. Purified P158 (330 ng μl–1) was lined ~2 mm wide on the IAB90 nitrocellulose membrane; The membrane was soaked in 3% BSA blocking solution at room temperature for 1 h and air-dried; Purified P543 trimer (500 ng μl–1) was amine coupled to 30-nm diameter Estapore beads. Glass fibre paper was soaked in 3% BSA blocking solution containing 0.005% (w/v) nanobody-coupled Estapore beads for saturation and then dried under vacuum conditions. The nasal swab specimens were sampled with 1% DDM and nuclease and were diluted with water to contain the final 20 mM NaCl concentration. Approximately, 120 μl of the sample was spotted on the sample pad; the strips were photographed under a 385 nm UV lamp.
Kinetic assays via biolayer interferometry (BLI)
Real-time binding experiments were performed using an Octet Red96 instrument (fortèBIO, Pall Life Science, Portsmouth, NH). Each purified nanobody clone was biotinylated with EZ-Link Sulfo-NHS-LC-Biotin (Thermo) according to the manufacturer’s protocol; uncoupled biotin was excluded with a size exclusion spin column (PD SpinTrap G-25: Cytiva) in PBS (pH 7.4). Assays were performed at 30 °C with shaking at 1000 rpm. Biotin-conjugated clones at 10 μg ml–1 were captured on a streptavidin-coated sensor chip (SA: fortèBIO) to reach the signals at 4 nm. One uncoated sensor chip was monitored as the baseline; another biotin-conjugated anti-IL-6-coated sensor chip (anti-IL-6 nanobody: COGNANO Inc.) was monitored as the background. The remaining 6 channels were immobilized with biotinylated anti-SARS-CoV-2 spike clones, and real-time binding kinetics to the purified extracellular domain of the SARS-CoV-2 spike trimer complex were measured in sequentially diluted concentrations at the same time (8 channels per assay). The concentrations of the SARS-CoV-2 spike loaded varied between 1 and 32 μg ml–1, corresponding to 1–32 nM or less, when an average of the molecular weight of the SARS-CoV-2 spike trimer complex was estimated to be ~1000 kDa or more according to chromatograms of gel filtration column chromatography (see Supplementary Fig. 1b). Assays were performed with high-salt phosphate buffer containing 500 mM NaCl and 0.001% LMNG (stabilized spike: for P17, P86, P334 and C116) or with hypotonic phosphate buffer containing 25 mM NaCl and 0.00005% LMNG (fluctuated spike: for P158, P543, C17, C49 and C246). After baseline equilibration for 180 s in each buffer, the association and dissociation were carried out for 600 s each. The data were double subtracted before fitting was performed with a 1:1 fitting model in fortèBIO data analysis software. The equilibrium dissociation constant (KD), koff and kon values were determined with a global fit applied to all data.
Pseudotyped virus production
HIV-1-based SARS-CoV-2 spike pseudotyped virus was prepared as follows: LentiX-HEK293T cells were transfected using a polyethyleneimine transfection reagent (Cytiva) with plasmids encoding the C-terminally C9-tagged full-length SARS-CoV-2 spike variants (original, Alpha, Beta, Delta and Omicron) and HIV-1 transfer vector encoding a luciferase reporter, according to the manufacturer’s protocol. The mutations of each variant are as follows: original (D614G); Alpha (H69del, V70del, Y144del, N501Y, A570D, D614G, P681H, T716I, S982A and D1118H); Beta (L18F, D80A, D215G, R246I, K417N, E484K, N501Y, D614G and A701Y); Delta (T19R, G142D, E156del, F157del, R158G, L452R, T478K, D614G, P681R and D950N); and Omicron (A67V, H69del, V70del, T95I, G142D, V143-Y145del, N211I, L212del, R214EPE, G339D, S371L, S373P, K417N, N440K, G446S, S477N, T478K, E484A, Q493R, G496S, Q498R, N501Y, Y505H, T547K, D614G, H655Y, N679K, P681H, N764K, D796Y, N856K, Q954H, N969K and L981F)86. Cells were incubated for 4–6 h at 37 °C with the medium that was then replaced with DMEM containing 10% FBS for the following 48-h culture. The supernatants were then harvested, filtered through a 0.45-μm membrane, concentrated with ultracentrifugation and frozen at −80 °C.
Pseudotyped virus neutralization assay
Fivefold sequentially diluted nanobodies (from 10 μg ml–1) were incubated with SARS-CoV-2 pseudotyped viruses for 1 h. K562 and HOS cells expressing human ACE2 and TMPRSS2 were subsequently infected with the antibody-virus mixture for 1 h at 37 °C and cultured for 2 days. The cells were lysed, and luciferase activity was measured using the Steady-Glo Luciferase Assay System (Promega KK, Osaka, Japan) with a microplate spectrophotometer (ARVO X3: PerkinElmer Japan Co., Ltd., Kanagawa, Japan). The obtained relative luminescence units were normalized to those derived from cells infected with the SARS-CoV-2 pseudotyped virus in the absence of nanobodies.
The ability of nanobodies to bind to the cell surface of the SARS-CoV-2 spike was studied by fluorescence-activated cell sorting (FACS)87. K562 cells expressing the SARS-CoV-2 spike were incubated with 1 μg ml–1 purified nanobody on ice for 30 min. After washing, the cells were incubated with the 1:400-diluted anti-His antibody (Abcam) on ice for 30 min and then the 1:400-diluted Alexa 647-conjugated anti-rabbit IgG (Dako). The cells were analysed with a Beckman-Coulter FC-500 Analyzer (Coulter Electronics, Hialeah, FL). The Ty1 and B988 nanobodies were used as controls. A region of positive signals for Ty1 was square-gated.
DT40 cells stably expressing the SARS-CoV-2 Omicron BA.1 variant spike were incubated with 1 μg ml–1 purified nanobodies and the 1:400-diluted antibodies. The cells were analysed with the BD Accuri flow cytometer (Becton, Dickinson and Company, Franklin Lakes, NJ). A homemade anti-Her-2 nanobody (COGNANO Inc.) was used as a control.
Microscopy analyses for cell staining
HEK293T cells were transiently transfected with plasmids encoding the C-terminally C9-tagged full-length SARS-CoV-2 spike variants using Lipofectamine 3000 (Thermo) according to the manufacturer’s instructions. The next day, the cells were seeded on collagen type I-coated culture plates (IWAKI, AGC TECHNO GLASS CO., LTD., Shizuoka, Japan) and cultured for 24 h before being fixed with 2% paraformaldehyde (PFA) at 4 °C overnight. After 3 washes with PBST (0.005% Tween), the cells were blocked with PBST containing 2% goat serum (blocking solution) at room temperature for 1 h. Each well was soaked with 100 μl of the blocking solution containing 30 ng of purified nanobody, except for 6 ng of C116, at 4 °C overnight. After washing with PBST, the 1:400-diluted anti-His-tagged antibody and the 1:400-diluted anti-C9-tagged antibody in blocking buffer were added and reacted at room temperature for 1 h. Finally, after washing, the fluorescently conjugated anti-rabbit IgG (594 nm emission) and anti-mouse IgG (488 nm emission) antibodies (Alexa Fluor: Thermo) were diluted with blocking buffer (1:400) and added to the wells, and the fixed cells were labelled at room temperature for 1 h before washing 3 times with PBST. Cellular nuclei were visualized with 4′,6-diamidino-2-phenylindole (DAPI). Stained cells were imaged with a 2-ms exposure time (594 nm emission), with a 10-ms exposure time (488 nm emission), or automatically adjusted exposure time (DAPI) using microscopy (IX71S1F-3: Olympus Corporation, Tokyo, Japan) with the cellSens Standard 1.11 application (Olympus). A 3 × 4 cm2 printed rectangle corresponds to a 165 × 220 μm2 observed field.
Expression and purification of monomeric P86 and P17
The monomeric C-terminally 6×His tagged nanobody genes were cloned into the pMES4 vector. The complete amino acid sequences are as follows (signal peptide sequence is underlined). P86: MKYLLPTAAAGLLLLAAQPAMAQVQLQESGGGLVQAGGSLRLSCVASGRTFSSLNIVWFRQAPGKERKFVAAINDRNTAYAESVKGRFTISRDNAKNTVHLQMNSLKPEDTAVYYCHSADVNGGMDYWGKGTQVTVSSHHHHHH. P17: MKYLLPTAAAGLLLLAAQPAMAQVQLQESGGGLVQAGGSLRLSCAASGRTSSVYNMAWFRQTPGKEREFVAAITGNGGTTLYADSVKGRLTISRGNAKNTVSLQMNVLKPDDTAVYYCAAGGWGKERNYAYWGQGTQVTVSSHHHHHH.
Bacterial BL21(DE3) E. coli cells were transformed with the plasmids and grown on an LB ampicillin-supplemented plate; colonies were picked and inoculated into 5 ml of LB medium containing 200 μg ml–1 ampicillin; and the cells were cultured in a shaking incubator overnight at 37 °C. The cultures were transferred to 1 L of LB medium containing 200 μg ml–1 ampicillin. At an optical density below 0.6, cells were cultured for 3 h at 37 °C with 1 mM IPTG.
The bacterial pellet was collected by centrifuging at 6000 × g for 30 min and suspended in 45 ml of lysis buffer containing 20 mM Tris (pH 8.0), 0.68 mM EDTA, 500 mM sucrose and a trace amount of benzonase nuclease (Merck). The solutions were mixed on a TR-118 tube rotator (AS ONE Corporation, Osaka, Japan); 90 mL of 20 mM Tris (pH 8.0) was added, and the mixtures were rotated for 45 min at 4 °C. The resulting solutions were centrifuged at 20,000 × g for 10 min at 4 °C. The supernatant was filtered and loaded onto a Ni-NTA column (Cytiva) equilibrated with 20 mM Tris buffer (pH 8.0). The column was washed several times with 20 mM Tris buffer (pH 8.0) containing 40 mM imidazole for P86 or 20 mM Tris buffer (pH 8.0) containing 150 mM NaCl and 40 mM imidazole for P17. Then, the nanobody was eluted with 20 mM Tris buffer (pH 8.0) containing 250 mM imidazole for P86 or 150 mM NaCl and 350 mM imidazole for P17.
The fractions containing nanobodies were collected and further purified using a HiLoad 16/60 Superdex 75 gel filtration column (Cytiva) equilibrated with 20 mM Tris buffer (pH 8.0) containing 150 mM NaCl for P86 or with 20 mM HEPES buffer (pH 8.0) containing 150 mM NaCl for P17. The peak fractions were collected and concentrated. The final concentrations of the P86 and P17 sample were 51 and 3.0 mg ml–1, respectively, based on the use of the absorption coefficient at 280 nm.
Cryo-EM specimen preparation and data collection
An epoxidized graphene grid (EG-grid)89 was used to increase the number of protein particles. The trimer complex of the SARS-CoV-2 spike (D614G) with the furin-resistant mutation (“GSAS”) at a concentration of 0.1 mg ml–1 was mixed with a 5-time molar excess of P86 or P17 monomer and incubated on ice for 10 min, and 3 μl of the spike-nanobody complex solution was applied to the EG-grid. After incubation at room temperature for 5 min, the grids were blotted with a force of –3 and a time of 2 s in a Vitrobot Mark IV chamber (Thermo) equilibrated at 4 °C and 100% humidity and then immediately plunged into liquid ethane. Excessive ethane was removed with filter paper, and the grids were stored in liquid nitrogen. All cryo-EM image datasets were acquired using a JEM-Z300FSC (CRYO ARM 300: JEOL, Tokyo, Japan) operated at 300 kV with a K3 direct electron detector (Gatan, Inc., Pleasanton, CA) in CDS mode90. The Ω-type in-column energy filter was operated with a slit width of 20 eV for zero-loss imaging. The nominal magnification was ×60,000, corresponding to 0.870 Å per pixel. Defocus varied between –0.5 and –2.0 μm. Each movie was fractionated into 60 frames (0.0505 s each, total exposure: 3.04 s) with a total dose of 60 e−/Å2.
Cryo-EM image processing and refinement
The images were processed using RELION 3.191. Movies were motion corrected using MotionCor292, and the contrast transfer functions (CTFs) were estimated using CTFFIND 4.193. Micrographs whose CTF max resolutions were beyond 5 Å were selected. Three-dimensional (3D) template-based autopicking was performed for all images, and the particles were extracted with 4× binning, which were subjected to two rounds of 2D classification. An initial model was generated and used as a reference for the following 3D classification in the P86 dataset. In the P17 dataset, a density map of spike trimers (our previous dataset) was used as a reference. Reference-based 3D classification (into 4 classes) without applying symmetry was conducted, and the selected particles were re-extracted without binning. 3D autorefinement without applying symmetry, soft mask generation, postprocessing, CTF refinement, Bayesian polishing, and another round of 3D autorefinement were performed. Then, focused 3D classification without alignment was performed to separate up- and down-states in one RBD. The selected particles were subjected to another round of 3D autorefinement, postprocessing, CTF refinement, 3D autorefinement. C3 symmetry was applied during the final round of 3D autorefinement for the 3-up+P86 dataset. The data were imported and further processed with non-uniform refinement in cryoSPARC v3.2.094. The final map resolutions (FSC = 0.143) were 3.03 and 2.70 Å in the 2-up and 3-up states in the P86 dataset and 3.20 and 3.29 Å in the 1-up and 2-up states in the P17 dataset, respectively. For the 2-up+P86 dataset, we tried local refinement to visualize the density of P86 bound to the down-RBD, but the resolution was limited to 5.13 Å (FSC = 0.143).
The model was built for the 2-up+P86 dataset. The SARS-CoV-2 spike trimer D614G mutant (PDB entry: 7KRR)95 and the crystal structure of P86 were used initial models. After the models were manually fitted into the density using UCSF Chimera v1.1596 and modified in Coot v0.8.9.297, real-space refinement was performed in PHENIX v1.19.198. The model was validated using MolProbity99, and this cycle was repeated several times. Figures were prepared using UCSF Chimera96, ChimeraX100 and PyMOL v2.5.0 (Schrödinger, LLC, New York, NY). The parameters are summarized in Table 1.
Crystallization of P86
Crystallization was performed by the sitting drop vapour diffusion method. A mosquito crystallization machine (TTP LabTech Ltd., Melbourn, Hertfordshire, UK) was used to prepare drops on 96-well VIOLAMO plates (AS ONE). The reservoir solution was 60 μl in volume, and 0.1 μl of protein solution was mixed with 0.1 μl of reservoir solution. Crystals appeared under the condition of 51 mg ml–1 protein, 0.2 M ammonium sulfate, and 30% (w/v) polyethylene glycol 4000 at 20 °C. A crystal cluster was crushed, and a peeled single crystal was harvested by LithoLoop (Protein Wave, Nara, Japan). Before the crystal was frozen in liquid nitrogen, it was soaked in the crystallization solution supplemented with 5% (v/v) ethylene glycol.
X-ray data collection, processing, structure solution and refinement
An X-ray diffraction experiment was performed on the BL44XU beamline of SPring-8 (Hyogo, Japan). Diffraction images were collected at 100 K using an EIGER X 16 M detector (Dectris, Philadelphia, PA, USA). The beam size was 50.0 × 50.0 μm2 (h × w). A 0.8-mm Al attenuator was used to weaken the X-ray. The crystal-to-detector distance was 160 mm. The exposure time per frame and the oscillation angle were 0.1 s and 0.1°, respectively. A total of 1800 images were collected. The dataset was processed using XDS101 and scaled by Aimless102. Molecular replacement phase determination was performed by MOLREP103 with a nanobody structure (PDB code ID: 5IVO)104 as a search model. Initial model building was performed by PhenixAutoBuild implemented in PHENIX98. Manual model building was performed using Coot97. The programme refmac5105 in the ccp4 suite106 and the programme Phenix-refine98 were used for structural refinement. The stereochemical quality of the final model was checked by MolProbity99. Data collection and refinement statistics are summarized in Table 2.
Authentic virus neutralization assay
TMPRSS2-expressing Vero E6 (VeroE6 + TMPRSS2) cells (JCRB1819)107 were maintained in high-glucose DMEM containing 10% FBS, 1% penicillin and streptomycin and 1 mg ml–1 G418 (Nacalai).
All experiments with authentic SARS-CoV-2 viruses were performed in a biosafety level-3 (BSL-3) facility at the Research Institute for Microbial Diseases, Osaka University. A Wuhan variant (strain SARS-CoV-2/Hu/DP/Kng/19-020) was kindly provided by Dr. Sakuragi at the Kanagawa Prefectural Institute of Public Health. An Omicron variant BA.1 (strain hCoV-19/Japan/TY38-873/2021) was obtained from the National Institute of Infectious Diseases, Japan. All viruses were initially amplified in VeroE6 + TMPRSS2 cells, and the culture supernatants were harvested and stored at −80 °C until use. The infectious virus titre was determined as plaque-forming units (PFU) by plaque assay.
Antibodies and viruses were diluted with DMEM containing 1% penicillin and streptomycin, and 2% FBS. Serially diluted nanobodies (from 50 μg ml–1 for P86 and 25 μg ml–1 for C246) and antibodies (from 50 μg ml–1) were mixed with an equal volume of virus solution containing 100 PFU. After incubation for an hour, the mixture was inoculated into confluent VeroE6 + TMPRSS2 cells in a 6-well plate for 1 h at 37 °C with 5% CO2. After washing the cells with DMEM containing 2% FBS, the cells were overlaid with 2 mL of DMEM containing 1% agarose, 5% FBS, 10 mM HEPES (Invitrogen), 0.3% sodium hydrogen carbonate (Invitrogen). At a few days post-infection, 10% formalin neutral buffer solution (FUJIFILM Wako Pure Chemical Co., Osaka, Japan) was overlaid to fix the cells. After removing the formalin and solid overlay medium, the fixed cells were stained with 0.1% crystal violet (bioWORLD, Dublin, OH), and then the number of plaques was counted.
Statistics and reproducibility
The number of replicates is indicated in the figure legend and means ± standard deviation (SD) are shown in graphs.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.