Institutional approval
This study has received institutional regulatory approval. All recombinant DNA (rDNA) and biosafety work were performed under the guidelines of the Yale Environment, Health and Safety (EHS) Committee with approved protocols (Chen-15-45, 18-45, 20-18, 20-26; Xiong-17302; Wilen-18/16-2). All animal work was performed under the guidelines of Yale University Institutional Animal Care and Use Committee (IACUC) with approved protocols (Chen-2018-20068; Chen-2020-20358; Wilen-2018-20198).
Animal immunization
Standard 28-day repetitive immunization protocol was utilized for immunization. M. musculus (mice), 6–12 weeks old females, of C57BL/6 J and BALB/c strains, were purchased from Jackson laboratory and used for immunization. First, all mice are ear-marked and around 200 μl blood was taken as a pre-immunization sample, where serum was collected from the blood by centrifugation (1000×g for 10 min). Two days later (day 0), for each mouse, 20 μg SARS-CoV-2 RBD-his tag protein (Sino biological) in 100 μl PBS was mixed with 100 μl Complete Freund’s Adjuvant (CFA) with three-way stop Cock. The fully emulsified mixture was subcutaneously injected into the back of each mouse. On day 7, a second immunization was performed, where each mouse was injected subcutaneously with 20 μg RBD-his tag protein fully emulsified with Incomplete Freund’s Adjuvant (IFA). On day 13, around 50 μl of blood from each mouse was obtained for serum preparation as first bleeds. On day 14, a third immunization is performed, where all the procedures were similar to the second immunization. On day 20, second bleeds were taken. On day 21, the fourth immunization is performed, where all the procedures were similar to the second immunization. On day 24, each mouse receives 20 μg RBD-his tag protein in 200 μl PBS intraperitoneally as final immunization. On day 28, mice with strong serum conversion detected by ELISA were sacrificed. Spleen, lymph nodes, and bone marrow were collected for B cells isolation and purification for single-cell BCR sequencing. Serums from pre, first, and second bleeds were subjected to ELISA for anti-RBD tilter determination.
Mouse B cell isolation and purification
Primary B cells from the spleen, draining lymph nodes, bone marrow of RBD-his tag protein immunized mice were isolated and purified with mouse CD138 MicroBeads (Miltenyi Biotec, 130-098-257) following the standard protocol provided by the manufacturer. Spleens and draining lymph nodes were homogenized gently. Bone marrows were fragmented, rinsed with PBS containing 2% FBS, and filtered with a 100 μm cell strainer (BD Falcon, Heidelberg, Germany). The cell suspension was centrifuged for 5 min with 400 × g at 4 °C. Erythrocytes were lysed briefly using ACK lysis buffer (Lonza) with 1 mL per spleen for 1–2 min before adding 10 mL PBS containing 2% FBS to restore iso-osmolarity. The single-cell suspensions were filtered through a 40 μm cell strainer (BD Falcon, Heidelberg, Germany). CD138 positive B cells were isolated using magnetic cell sorting by positive selection according to the manufacturer’s instructions. Cell samples post-magnetic selection were counted and prepared for single-cell BCR sequencing.
Single-cell BCR sequencing
The enriched CD138+ plasma cells and progenitor B cells were loaded on a 10X Chromium Next GEM Chip G. The target cell number was 10,000 cells per sample. Single-cell lysis and RNA first-strand synthesis were performed using Chromium Next GEM Single Cell 5′ Gel Bead V2 according to the manufacturer’s protocol. The following RNA and V(D)J library preparation was performed according to the manufacturer’s protocol (Chromium Next GEM Single Cell V(D)J reagent kit, mouse BCR). The resulting VDJ-enriched libraries were sequenced following the reading mode recommended by 10× Genomics. Sequencing was performed on a NovaSeq targeted for 10,000 reads/cell, with a total of 100 million reads.
Single-cell VDJ sequencing data analysis
Raw sequencing data were processed using Cell Ranger v3.1.0 with default settings, aligning the reads to the GRCm38 mouse VDJ reference. Outputs from Cell Ranger were then visualized using the Loupe V(D)J Browser for quality control assessment and to identify the top enriched clonotypes. The consensus amino acid sequences for the top-ranked heavy/light chain pairs in each sample were then extracted and codon-optimized for human expression.
Plasmid construction
The cDNA sequences of the paired variable heavy and light chain region of anti-RBD antibody clones were synthesized as gBlocks (IDT) and cloned by the Gibson assembly (NEB) into human IgG1 heavy chain and light chain expression plasmids, pFUSEss-CHIg-hG1(InvivoGen, pfusess-hchg1) and pFUSE2ss-CLIg-hK (InvivoGen, pfuse2ss-hclk), respectively. pFUSEss-CHIg-hG1 plasmid is a cloning plasmid that expresses the constant region of the human IgG1 heavy chain and includes multiple cloning sites to enable cloning of the heavy chain (CH) variable region. Parallelly, pFUSE2-CLIg-hK is a cloning plasmid that expresses the constant region of the human kappa light chain and contains multiple cloning sites to enable cloning of the light chain variable region. For anti-RBD antibody clones’ heavy chain plasmid cloning, gBlocks, containing cDNA sequence of the variable region of the heavy chain of anti-RBD antibody clones and the regions overlapping with corresponding flanking sequences of EcoRI and NheI restriction sites pFUSEss-CHIg-hG1, were ordered from IDT. pFUSEss-CHIg-hG1 were digested with EcoRI and NheI restriction enzyme (Thermo Fisher). These synthesized gBlocks were cloned into gel-purified restriction enzyme digested backbone by the Gibson assembly (NEB). For anti-RBD antibody clones’ light chain plasmid cloning, gBlocks, containing cDNA sequence of the variable region of the light chain of anti-RBD antibody clones and the regions overlapping with corresponding flanking sequences of EcoRI and BsiWI restriction sites pFUSE2ss-CLIg-hK, were ordered from IDT. The gBlocks were then cloned into the pFUSE2ss-CLIg-hK backbone, which was digested with EcoRI and BsiWI restriction enzyme (Thermo Fisher).
The bispecific antibody with the same Fab regions of clone 2 and clone 6 was generated by using the CrossMab-KiH bispecific constructs69. The CrossMab-KiH bispecific constructs were designed and generated based on pFUSEss-CHIg-hG1 and pFUSE2ss-CLIg-hK. The bispecific antibody consists of two hetero-half IgG1, one is knob IgG1, and the other is hole IgG1 (Knob-in-Hole conformation). Four plasmids were employed: pFUSE2ss-knobLight-hK, pFUSE2ss-knobheavy-hG1, pFUSE2ss-HoleLight-hK, and pFUSE2ss-HoleHeavy-hG1. The pFUSE2ss-knobLight-hK is pFUSE2ss-CLIg-hK with no further editing. The pFUSE2ss-knobheavy-hG1 contains two knob mutations (T366W and S354C) in the CH3 region when compared with pFUSEss-CHIg-hG1. The gBlock (pPR024), containing constant region of heavy chain with two knob mutations and the regions overlapping with corresponding flanking sequences of NsiI and NheI restriction sites in pFUSEss-CHIg-hG1 was ordered from IDT and then cloned into NsiI and NheI restriction enzymes digested pFUSEss-CHIg-hG1 backbone by the Gibson assembly (NEB). The pFUSE2ss-HoleLight-hK was generated by replacing the constant region of light chain (CL) in pFUSE2ss-CLIg-hK with CH1 region of heavy chain in pFUSEss-CHIg-hG1 vector. The CH1 region were PCR amplified from pFUSEss-CHIg-hG1vectors with a forward primer (oPR81-F) and a reverse primer (oPR82-R) containing regions overlapping with corresponding flanking sequences of the NcoI and NheI restriction sites in the pFUSE2ss-CLIg-hK. CH1 PCR amplified fragments were gel-purified and cloned into restriction enzyme digested pFUSE2ss-CLIg-hK by the Gibson assembly (NEB). The pFUSE2ss-HoleHeavy-hG1 possesses three “hole” mutations (T366S, L368A, and Y407V) in the CH3 region and a Y349C on the “hole” side to form a stabilizing disulfide bridge. In addition, to get the correct association of the light chain and the cognate heavy chain, the CH1 region in the pFUSE2ss-HoleHeavy-hG1 was exchanged with the constant region of the light chain (CrossMab conformation). The gBlock (pPR023), containing cDNA sequence of the constant region of light chain, CH2 and CH3 with “hole” mutations, and regions overlapping with corresponding flanking sequences of NsiI and NheI restriction sites in pFUSEss-CHIg-hG1 was ordered from IDT and cloned into NsiI and NheI restriction enzymes digested pFUSEss-CHIg-hG1 backbone through Gibson assembly (NEB). All plasmids were sequenced and Maxiprepped for subsequent experiments.
A list of oligos used for plasmid construction is provided in Supplemental Table 1 in the Supplemental Information.
Cloning of SARS-CoV-2 spike variants
The construct of wild-type (WT) SARS-CoV-2 ectodomain of spike trimer is a gift from Dr. Jason S. McLellan at the University of Texas at Austin34. The recently emerged SARS-CoV-2 spike SA variant B.1.3517 and Indian variant B.1.61711 was generated by standard cloning. The pVP21-SA variant includes four mutations in the N-terminal domain (L18F, D80A and D215G, R246I), three mutations at key residues in the RBD (N501Y, E484K, and K417N), and one is in loop 2 (A701V). The pVP28-Indian variant includes seven mutations in Spike G142D, E154K, L452R, E484Q, D614G, P681R, and Q1071H. The pVP21-SA and pVP28-Indian were generated based on pcDNA3.1-pSARS-CoV-2-S, which was derived by insertion of a synthetic human codon-optimized cDNA (Geneart) encoding a WA1 SARS-CoV-2 S protein. For pVP21-SA-variant, two gBlocks, contain mutations in SA variant regions overlapping with corresponding flanking sequences of NheI and BsrGI restriction sites pcDNA3.1-pSARA-CoV-2. The gBlocks were then cloned into the pcDNA3.1-pSARA-CoV-2 backbone, digested with NheI and BsrGI restriction enzyme (Thermo Fisher) through Gibson assembly. For pVP28-Indian, four gBlocks, contains mutations in Indian variant regions overlapping with corresponding flanking sequences of NheI and BamHI restriction sites pcDNA3.1-pSARA-CoV-2. The gBlocks were then cloned into the pcDNA3.1-pSARA-CoV-2 backbone, digested with NheI and BamHI restriction enzyme (Thermo Fisher) through Gibson assembly. For the HIV-1-based SARS-CoV-2 spike pseudotyped virus generation, WT pcDNA3.1-pSARS-CoV-2-S, pVP21-SA-variant, and a pVP28-Indian variant lacking the C-terminal 19 codons were employed. A pair of forward and reverse primers were utilized to amplify fragments lacking the C-terminal 19 codons with the pVP21-SA variant and pVP28-Indian variant as templates separately. The amplified fragments were gel-purified and cloned into the pVP21-SA variant backbone and pVP28-Indian variant backbone, digested with BbvCI and BamHI.
Cell culture
HEK293FT (Thermo Fisher) and 293T-hACE2 (gifted from Dr. Bieniasz’ lab) cell lines were cultured in the complete growth medium, Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher) supplemented with 10% Fetal bovine serum (FBS, Hyclone),1% penicillin-streptomycin (Gibco) (D10 media for short). Cells were typically passaged every 1–2 days at a split ratio of 1:2 or 1:4 when the confluency reached 80%. Expi293FTM (Thermo Fisher) cells were cultured in Expi293™ Expression Medium (Thermo Fisher) in a 125-mL shaker flask in a 37 °C incubator with 8% CO2 on an orbital shaker rotating at 125 rpm. For routine maintenance, Expi293FTM cells were grown to 3–5 × 106 cells/mL, then split to 0.3–0.5 × 106 cells/mL every 3 days.
Expression and purification of WT SARS-CoV-2 ectodomain of spike trimer
The WT ectodomain of the SARS-CoV-2 spike trimer was expressed in Expi293F cells. For 100 mL expression scale, 100 μg construct DNA was mixed with 400 μg polyethylenimine in 10 mL Opti-MEM® I Reduced-Serum Medium (Thermo Fisher) for 30 min, and then added into 90 mL Expi293F cells at a density of 2.5–3 × 106 cells/mL for incubation, shaking at 125 rpm in a 37 °C incubator with 8% CO2. After 5 days, the medium with the secreted protein was harvested and loaded onto an ion-exchange column. Fractions containing the target protein was pooled and further purified using a Ni-NTA affinity column, followed by size exclusion chromatography using a Superose 6 10/300 column (GE Healthcare) with a buffer of 30 mM Tris, pH 8.0, 100 mM NaCl. The monodispersed peak containing the ectodomain of the spike trimer was pooled and concentrated for subsequent analysis.
Recombinant antibody generation
The top-ranked enriched IgG clones were selected and cDNAs of a relative variable region of paired heavy- and light-chain were codon-optimized and cloned separately into human IgG1 heavy chain and light chain expression vectors, containing the human IgG1 constant regions (pFuse plasmids). IgG1 antibodies were expressed in Expi293FTM cells. ExpiFectamine 293 transfection kit (Thermo Fisher) was utilized for heavy and light chain plasmids transfection following the manufacturer’s instruction. After 5 days, the antibody-containing supernatants were collected. A suitable amount of rProtein A Sepharose® Fast Flow beads (Cytiva) was prewashed and added into supernatants. After overnight incubation at 4 °C, antibody-bound protein A beads were collected with Poly-Prep® Chromatography Columns (BIO-RAD). After three times wash with DPBS, mAbs were eluted with Fab elution buffer, then neutralized with Tris-HCl. Buffer exchange was performed with Amicon Ultra-4 Centrifugal Filter (MilliporeSigma) to keep mAbs in PBS for the following assays. The numbering of mAbs was based on the order of mouse immunization and cloning. Clones 1–4 were mAbs chosen from enriched clones from RBD-his tag protein immunized C57BL/6 J mice. Clone 5-11 were mAbs chosen from RBD-his tag protein immunized BALB/c mice.
Bispecific antibody generation
Clone 16 (Clone 6-KiH-Clone 2) bispecific antibody is a human IgG1-like bispecific antibody, generated based on CrossMab-KiH bispecific constructs, including pFUSE2ss-knobLight-hK, pFUSE2ss-knobheavy-hG1, pFUSE2ss-HoleLight-hK, and pFUSE2ss-HoleHeavy-hG1. The design and generation of CrossMab-KiH bispecific constructs was described in the above plasmid constructs parts. The variable region of the Clone 6 heavy chain was cloned into pFUSE2ss-knobheavy-hG1 vector. The variable region of the Clone 6 light chain was cloned into pFUSE2ss-knobLight-hK vector. Clone 6-KiH-Clone 2 bispecific antibody was expressed in vitro in Expi293FTM cells by co-transfecting four plasmids (Clone 6 knob heavy chain plasmid, Clone 6 knob light chain plasmid, Clone 2 hole heavy chain plasmid, and Clone 2 hole light chain plasmid) with ExpiFectamine 293 transfection kit (Thermo Fisher). The expression and antibody purification protocol was similar to the recombinant antibody expression described above. The bispecific antibody was efficiently purified by using rProtein A Sepharose Fast Flow antibody purification resin (Cytiva, Cat:#17127901).
Antibody humanization
In order to humanize the antibody, we first determine the six CDR loops from murine variable domains by using the online free program “IGBLAST” (https://www.ncbi.nlm.nih.gov/igblast/). Followed by applying the CDR-grafting technique and grafting six CDR loops onto human acceptor frameworks. The framework template selection was based on sequence similarity to close human germline sequence, as well as homology to clinically validated germline sequences. Thereafter, we identify Vernier zone residues through Cryo-EM structure between Clone 2 and trimeric S protein of SARS-CoV-2 from parent antibody (FR residues of Clone 2 within 5 Ǻ of trimeric S protein) and substitute the key residues into the human acceptor framework of Clone 13 A.
ELISA
ELISA for anti-serum titer determination
The antibody tilters in sera from pre, first, and second bleeds were determined using direct coating ELISA. The 384-well ELISA plates (Corning) were coated with 3 μg/mL SARS-CoV-2 RBD-his tag protein (Sino) in PBS at 4 °C overnight. After standard washing with PBST washing buffer (phosphate-buffered saline containing 0.05% Tween 20), ELISA plates were blocked with blocking buffer (2% bovine serum albumin dissolved in PBST and filtered) for 1 h at room temperature. Serial dilutions of pre-immune, first, and second immune anti-sera in blocking buffer were added into plates and for 1 h at room temperature. Plates were washed and incubated with relative goat anti-mouse IgG(H + L)/HRP (Thermo Fisher,1:5000) for 1 h at room temperature. Plates were washed and developed using TMB reagents as substrates (Biolegend) following the manufacturer’s recommended protocol. The reaction was stopped with a stop solution (1 M H3PO4) and absorbance at 450 nm was recorded by a microplate reader (Perkin Elmer).
ELISA for Anti-RBD antibody clones binding
SARS-CoV-2 RBD protein with 6 X Histidine (Sino) was coated at 3 μg/ml in PBS on a 384-well microtiter plate overnight at 4 °C. After standard washing with PBST and blocked with 2% (w/v) solution of BSA in PBST to remove the nonspecific binding, purified anti-RBD antibodies were diluted proportionally in PBST + 2% BSA and transferred to the washed and blocked microtiter plates. After 1 h of incubation at RT, plates were washed, and RBD-his tag protein-bound antibody was detected with goat anti-human IgG1 (H + L) with horseradish peroxidase (HRP) conjugated (Invitrogen,1:1000) The plates were washed and developed using TMB substrate solution (Biolegend) according to manufacturer’s recommendation and absorbance at 450 nm was measured on a microplate reader after the reaction was stopped by stop solution (1 M H3PO4).
Affinity determination via bio-layer interferometry (BLI)
Antibody binding kinetics for anti-spike mAbs were evaluated by BLI on an Octet RED96e instrument (FortéBio) at room temperature. Two types of measurements were performed. (1) HIS1K biosensors (FortéBio) were first loaded with his-tagged SARS-CoV-2 RBD protein to a response of about 1 nm, followed by a 60 s baseline step in the kinetic buffer (PBS, 0.02% Tween, pH 7.4). After that, the biosensors were associated with indicated concentrations of the antibodies (from 50 to 0.78125 nM with twofold dilutions, where the kinetic buffer was served as the negative control) for 200 s, then dissociated in the kinetic buffer for 1000 s. (2) 25 ng/ul of Clone 13A-IgG1 antibodies were captured on an AHC biosensor (ForteBio). The baseline was recorded for 60 s in a running buffer (PBS, 0.02% Tween 20, and 0.05% BSA, pH 7.4). Afterward, the sensors were subjected to an association phase for 500 s in wells containing RBD-his diluted in the buffer. In the dissociation step, the sensors were immersed in the running buffer for 1000 s. The dissociation constants KD, kinetic constants Kon and Koff, were calculated by using a 1:1 Langmuir binding model with FortéBio data analysis software. Octet data were analyzed with Octet® CFR software and Prism.
Affinity measurement by surface plasmon resonance (SPR)
Kinetics binding measurement for anti-spike mAbs in this study was performed using a Biacore T200 instrument (GE Healthcare). The system was flushed with filtered 1xHBS-P + running buffer (0.01 M HEPES, 0.15 M NaCl, and 0.05%v/v Surfactant P20, pH 7.4) and all steps were performed at 25 °C chip temperature.
Kinetics binding measurement on CM5 Chip (Series S sensor chip CM5)
For kinetic binding measurements, the CM5 chip surface was activated by injecting a solution of EDC/NHS (GE Healthcare). Mouse anti-human IgG (Fc) mAb (25 μg/ml) was immobilized on the sensor chip by amine coupling, followed by deactivation using 1 M ethanolamine. Afterward, anti-spike mAbs (0.1 μg/ml) were then flowed over and captured on anti-human IgG (Fc) mAb-coated surface. Subsequently, gradient diluted his-tagged SARS-CoV-2 RBD solutions (1.875–30 nM, twofold serial dilution) were injected individually in a single-cycle kinetic format without regeneration (30 μl/min, association:180 s, dissociation:60 s). The binding data were double referenced by blank cycle and reference flow cell subtraction. Processed data were fitted by a 1:1 interaction model using Biacore T200 Evaluation Software 3.1.
Kinetics binding measurement on NTA Chip
For kinetic binding measurements, the NTA chip was activated manually by loading a solution of NiCl2. Histidine-labeled SARS-CoV-2 RBD protein (0.075 μg/ml) was then flowed over the chip and captured on a nickel-coated surface. Subsequently, gradient diluted anti-spike mAbs solutions (0.9875–15 nM, twofold serial dilution) were injected individually in a single-cycle kinetic format without regeneration (30 μl/min, association:240 s, dissociation:90 s). The binding data were double referenced by blank cycle and reference flow cell subtraction. Processed data were fitted by a 1:1 interaction model using Biacore T200 Evaluation Software 3.1.
SARS-CoV-2 pseudovirus reporter and neutralization assays
HIV-1 based SARS-CoV-2 S pseudotyped virions were generated according to a previous study70. Two plasmids are adopted to generate HIV-1 based SARS-CoV-2 S pseudotyped virions. HIV-1 dual reporter vector expressing mCherry and luciferase (NL4-3 mCherry Luciferase, plasmid#44965) was purchased from Addgene. Plasmid expression of a C-terminally truncated SARS-CoV-2 S protein (pSARS-CoV-2Δ19) was obtained from Dr. Bieniasz’s lab. In order to generate HIV-1 based SARS-CoV-2 S pseudotyped virions, 15×106 293FT cells were seeded in 150 mm plates one day before in 20 ml D10 media. The following day, after the cell density reaches 90%, the medium was discarded and replaced with a 13 mL serum-free Opti-MEM medium. 20 μg NL4-3 mCherry Luciferase reporter plasmids and 15 μg SARS-CoV-2 (pSARS-CoV-2Δ19) plasmids were mixed thoroughly in 225 μl serum-free Opti-MEM medium. Then 100 μl Lipofectamine 2000 (Invitrogen) were diluted in 225 μl serum-free Opti-MEM medium. Then the diluted plasmid mixture and Lipofectamine 2000 were mixed thoroughly and incubated for 10 min at RT before adding into cells. After 6 h, the culture medium was changed back to the completed growth medium, 20 mL for one 150 mm plate. At 48 h after transfection, the 20 mL supernatant was harvested and filtered through a 0.45-μm filter, aliquoted, and frozen in −80 °C.
Parallelly, the three plasmids-based HIV-1 pseudotyped virus systems were utilized to generate (HIV-1/NanoLuc2AEGFP)-SARS-CoV-2 particles and (HIV-1/NanoLuc2AEGFP)-SARS-CoV-2-SA variant particles. The reporter vector, pCCNanoLuc2AEGFP, and HIV-1 structural/regulatory proteins (pHIVNLGagPol) expression plasmid were gifts from Dr. Bieniasz’s lab70. Briefly, 293 T cells were seeded in 150 mm plates and transfected with 21 µg pHIVNLGagPol, 21 µg pCCNanoLuc2AEGFP, and 7.5 µg of a SARS-CoV-2 SΔ19 or SARS-CoV-2 SA SΔ19 plasmid utilizing 198 µl PEI. At 48 h after transfection, the 20-ml supernatant was harvested and filtered through a 0.45-μm filter, and concentrated before aliquoted and frozen at −80 °C.
The pseudovirus neutralization assays were performed on 293T-hACE2 cell line70. One day before, 293T-hACE2 cells were plated in a 96-well plate, 0.02 × 106 cells per well. The following day, serial dilution of monoclonal IgG from 40 μg/mL (fourfold serial dilution using complete growth medium, 55 μL aliquots) were mixed with the same volume of SARS-CoV-2 pseudovirus. The mixture was incubated for 1 h at 37 °C incubators, supplied with 5% CO2. Then 100 μL of the mixtures were added into 96-well plates with 293T-hACE2 cells. Plates were incubated at 37 °C supplied with 5% CO2. Forty-eight hours later, 1 μL d-luciferin reagent (Perkin Elmer, 33.3 mg/ml) was added to each well and incubated for 5 min. Luciferase activity was measured using a microplate spectrophotometer (Perkin Elmer). The inhibition rate was calculated by comparing the OD value to relative negative and positive control wells. For the three plasmids-based HIV-1 pseudotyped virus systems, 293 T cells were collected and the GFP + cells were analyzed with Attune NxT Acoustic Focusing Cytometer (Thermo Fisher). The 50% inhibitory concentration (IC50) was calculated with a four-parameter logistic regression using GraphPad Prism 8.0 (GraphPad Software Inc.).
Cell fusion assay
Vectors and plasmids
Plasmid encoding human ACE2 (hACE2) was obtained from Addgene (hACE2; catalog #1786). The hACE2 2.6 kbp ORF was also blunt-cloned into a third-generation HIV vector 3′ of the CMV promoter and 5′ of an IRES-puror cassette to generate pHIV-CMV-hACE2-IRES-Puro. It was inserted into a piggybac transposon (Matt Wilson of Baylor College of Medicine, along with the transposase plasmid pCMV-piggybac) that had been modified to encode a CMV-IRES-bsdr cassette; the resultant plasmid was named pT-PB-SARS-CoV-2 Spike-IRES-Blasti. This too was inserted into piggybac transposon to make pT-PB-SARS-CoV-2-UK Spike-IRES-Blasti.
Cell lines
The HOS cells were stably transduced with a third-generation HIV vector encoding tat, along with eGFP, mRFP, and bleomycin resistance gene; they were maintained in 200–400 μg/mL phleomycin (Invivogen) and were eGFP and mRFP-positive by flow cytometry. hACE2 was subsequently introduced by VSV G-mediated HIV-based transduction using pHIV-CMV-hACE2-IRES-Puro to produce HOS-3734, which cell lines maintained in selection using 10 μg/mL puromycin (Sigma-Aldrich). TZMbl cells (#JC53BL-13) were obtained from the NIH AIDS Reagent Program. TZMbl cells stably expressing wild-type S/UK variant S were created by co-transfecting TZMbl cells with pT-PB-SARS-CoV-2- Spike-IRES-Blasti or pT-PB-SARS-CoV-2-UK Spike-IRES-Blasti, respectively, along with pCMV-piggybac and resistant cells selected with 10 μg/mL blasticidin (Invivogen). The control TZMbl cell line not expressing S was generated by co-transfecting pCMV-piggybac with pT-pB-IRES-Blasti and selecting for blasticidin-resistant TZMbl cells.
Cell fusion inhibition by monoclonal antibodies
Producer cells (TZMbl-wild-type Spike/ Tzmbl-UK Spike) and target cells (HOS-3734) were generated as described above. Ten thousand S-expressing cells (TZMbl-wild-type Spike/TZMbl-UK Spike) in 100 µL of the medium in the absence of blasticidin were seeded in 96-well plates. After 24 h, 70 μL of fourfold serially diluted antibody was added into producer cells and incubated at 37 °C for 1 h. At that time 104 target cells (HOS-3734) in 50 μL medium were then added to the producer cells, and after another 24 h cells were lysed in 0.1 mL and RLU measured. Data were analyzed with nonlinear regression using GraphPad Prism to determine the neutralization curve and the IC50 values calculated.
In vitro neutralization against authentic SARS-CoV-2
SARS-CoV-1 (USA-WA1/2020) was produced in Vero-E6 cells and tittered as described previously71. SARS-CoV-2 neutralization was assessed by measuring cytotoxicity. About 5 × 105 Vero-E6 cells were plated per well of a 96-well plate. The following day, serial dilutions of antibodies were incubated with 2.5 × 103 plaque-forming units (PFU) SARS-CoV-2 for 1 h at room temperature. SARS-CoV-2 neutralization was assessed by measuring cytotoxicity. About 5 × 105 Vero-E6 cells were plated per well of a 96-well plate. The following day, serial dilutions of antibodies were incubated with 2.5 × 103 PFU SARS-CoV-2 for 1 h at room temperature. The medium was then aspirated from the cells and replaced with 100 µl of the antibody/virus mixture. After 72 h at 37 °C, 10 µl of CellTiter- Glo (Promega) was added per well to measure cellular ATP concentrations. Relative luminescence units were detected on Cytation5 (Biotek) plate reader. All conditions were normalized to uninfected control. Each condition was done in triplicate in each of three independent experiments.
Focus reduction neutralization test
Serial dilutions of mAbs or sera were incubated with 102 focus-forming units (FFU) of different strains or variants of SARS-CoV-2 for 1 h at 37 °C. Antibody-virus complexes were added to Vero-TMPRSS2 cell monolayers in 96-well plates and incubated at 37 °C for 1 h. Subsequently, cells were overlaid with 1% (w/v) methylcellulose in MEM supplemented with 2% FBS. Plates were harvested 24 h later by removing overlays and fixed with 4% PFA in PBS for 20 min at room temperature. Plates were washed and sequentially incubated with an oligoclonal pool of SARS2-2, SARS2-11, SARS2-16, SARS2-31, SARS2-38, SARS2-57, and SARS2-714,72. The anti-S antibodies and HRP-conjugated goat anti-mouse IgG (Sigma, 12-349) in PBS supplemented with 0.1% saponin and 0.1% bovine serum albumin. SARS-CoV-2-infected cell foci were visualized using TrueBlue peroxidase substrate (KPL) and quantitated on an ImmunoSpot microanalyzer (Cellular Technologies).
In vivo efficacy testing against authentic SARS-CoV-2
The efficacy of mAbs against replication-competent SARS-CoV-2 virus was evaluated in vivo, using both a prophylactic setting where the animals were treated with mAb prior to viral infection and a therapeutic setting where the animals were treated post-infection. These experiments were performed in an animal BSL3 (ABSL3) facility. The replication-competent SARS-CoV-2 (USA-WA1/2020) virus was produced in Vero-E6 cells, and the titer was determined by plaque assay using WT Vero-E6.
The K18-hACE2 mice (B6.Cg-Tg(K18-ACE2)2Prlmn/J) were purchased from the Jackson Laboratory and bred in-house using a trio breeding scheme. Mice were sedated with isoflurane and infected via intranasal inoculation of 2000 PFU (20x LD50) SARS-CoV-2 (USA-WA1/2020) virus administered in 50 uL of DPBS. Six to eight-week-old K18-hACE2 littermate-controlled mice, mixed-gender (male/female) mice were divided randomly into three groups and administered with 20 mg/kg (of mice body weight) Clone 2, Clone 6 or placebo/control, via intraperitoneal (IP) injection. For a prophylactic experiment, the mAb drug/placebo treatment was 24 h prior to the infection; for a therapeutic experiment, the treatment was 18 h post-infection. The control for the prophylactic experiment was DPBS, and the control for the therapeutic experiment was isotype control hIgG1, where both controls are similar (no effect on disease progression). Survival, body conditions, and weights of mice were monitored daily for 10 consecutive days.
In vivo efficacy testing of humanized Clone 13A to authentic SARS-CoV-2 virus
Ten-12-week-old littermate-controlled female and male K18hAce2Tg+ mice were pretreated with 20 mg/kg of either control hIgG1 (purchased from BioXCell) or clone 13A mAb (produced by the Chen lab) administered IP in 300 uL of DPBS. Twenty-four hours later, mice were anesthetized with isoflurane, and SARS-CoV-2 isolate USA-WA1/2020, or Delta variant (B.1.617.2), was inoculated intranasally at a dose of 2 × 103 PFU/mouse (determined using wild-type Vero-E6) in 50 μL of DPBS. Weights were obtained daily for 10 days following infection, and mice were euthanized when morbid.
Fab generation
The Fab fragments of Clone 2 and Clone 6 were generated from full-length IgGs of Clone 2 and Clone 6 using a commercial PierceTM Fab Preparation Kit (Thermo Fisher). All procedures were performed following the manufacturer’s instructions. Briefly, 2 mg of the whole IgGs of Clone 2 and Clone 6 were digested with immobilized papain at 37 °C for 4 h with rotation. Then protein A beads were applied to bind the Fc fragments and undigested IgG. Then Fab fragments were recovered in the flow-through fraction, and further purified by size exclusion chromatography using a Superdex 200 10/300 column (GE Healthcare) in 30 mM Tris pH 8.0, 100 mM NaCl. The monodispersed peak of Fab fragments was pooled and concentrated for subsequent analysis.
Cryo-EM sample preparation and data collection
The purified SARS-CoV-2 spike trimer at a final concentration of 0.3 mg/mL (after mixture) was mixed with Clone 2 or Clone 6 Fab at a molar ratio of 1:2 at 4 °C for 30 min. Then 3 μl of the protein mixture was applied to a Quantifoil-Cu-2/1-3 C grid (Quantifoil) pretreated by glow-discharging at 15 mA for 1 min. The grid was blotted at 4 °C with 100% humidity and plunge-frozen in liquid ethane using FEI Vitrobot Mark IV (Thermo Fisher). The grids were stored in liquid nitrogen until data collection.
Images were acquired on an FEI Titan Krios electron microscope (Thermo Fisher) equipped with a Gatan K3 Summit direct detector in super-resolution mode, at a calibrated magnification of 81,000× with the physical pixel size corresponding to 1.068 Å. Detailed data collection statistics for the Fab-spike trimer complexes are shown in a Supplemental Table. Automated data collection was performed using SerialEM 3.873.
Cryo-EM data processing
A total of 2655 and 1766 movie series were collected for Clone 2 Fab-S trimer complex and Clone 6 Fab-S trimer complex, respectively. The same data processing procedures were carried out for each complex as described below. Motion correction of the micrographs was carried out using RELION74 and contrast transfer function (CTF) estimation was calculated using CTFFIND475. Particles were picked automatically by crYOLO76, followed by 2D and 3D classifications without imposing symmetry. The 3D classes with different S trimer conformations were then processed separately by consensus 3D refinement and CTF refinement. Image processing and 3D reconstruction using cryoSPARC77 produced similar results. For each state of the Clone 6 Fab-S trimer complex, multibody refinements were then carried out in RELION by dividing the complex into individual rigid bodies (three refinements each with a rigid body containing a unique Fab, RBD, and the N-terminal domain (NTD) of spike S1 subunit, and another rigid body for the rest of the spike-ectodomain trimer). For each state of the Clone 2 Fab-S trimer complex, local masked 3D classification without image alignment was performed focusing on one Fab-RBD region, and the best class of particles was selected for consensus refinement of the whole complex. Subsequently, multibody refinement was performed as described above for the rigid body containing the focused region. The 3D reconstruction of the other Fab-RBD regions were obtained with the same procedure. The final resolution of each reconstruction was determined based on the Fourier shell correlation (FSC) cutoff at 0.143 between the two half maps78. The final map of each body was corrected for K3 detector modulation and sharpened by a negative B-factor estimated by RELION79, and then merged in Chimera for deposition. The local resolution estimation of each cryo-EM map is calculated by RELION74. See also Supplementary Fig. 7 and Table 1.
Model building and refinement
The structure of the ectodomain of SARS-CoV-2 spike trimer (PDB 6VSB) was used as an initial model and docked into the spike trimer portion of the cryo-EM maps using Chimera80. The initial models of Clone 2 and Clone 6 Fabs were generated by homology modeling using SWISS-MODEL81, and then docked into the Fab portions of the cryo-EM maps using Chimera80. The initial models were subsequently manually rebuilt in COOT82, followed by iterative cycles of refinement in Refmac583 and PHENIX84. The final models with good geometry and fit to the map were validated using the comprehensive cryo-EM validation tool implemented in PHENIX85. All structural figures were generated using PyMol (http://www.pymol.org/) and ChimeraX80.
Homology modeling of SARS-CoV-2 variants
The structural models of SARS-CoV-2 variants of RBD were generated by SWISS model81 using the wildtype / WA RBD Cryo-EM structure as a template. The generated structures were aligned with the wild-type RBD in complex with Clone 2, Clone 6, and/or other mAbs. The cryo-EM structures and homology models were analyzed in Pymol.
Replication, randomization, blinding, and reagent validations
Sample size determination was performed according to similar work in the field, e.g., (Wang et al. 2021 Nature).
Replicate experiments have been performed for key data shown in this study, as detailed in methods and/or legends. Replicate experiments were successful where applicable.
Biological or technical replicate samples were randomized where appropriate. In animal experiments, mice were randomized by cage, sex, and littermates.
Experiments were not blinded. It is unnecessary for animal immunization for antibody production to be blinded. Cryo-EM structure study can not be blinded.
Antibodies and dilutions
Commercial antibodies used for staining were the following, with typical dilutions listed:
Mouse anti-Human lgGl Fe Secondary Antibody, HRP Thermo Fisher Cat#A-10648, 1:2000
InVivoMAb human lgGl isotype control BioXcell Cat#BE0297.
Goat Anti-Mouse lgG H&L (HRP) Abcam ab6789, Abcam Cat#ab6789, 1:5000.
Recombinant monoclonal human lgGl antibody against Spike RBD lnvivogen Cat#srbd-mabl
Custom antibodies were generated in this study, where dilutions were often serial titrations (i.e., a number of dilutions as specified in each figure)
Anti-SARS-CoV-2 Spike mAbs:
Clone 1
Clone 2
Clone 3
Clone 4
Clone 5
Clone 6
Clone 7
Clone 8
Clone 9
Clone 11
Clone 12
Clone 13
Clone 13A
Clone 16 (bispecific)
Commercial antibodies were validated by the vendors, and re-validated in-house as appropriate. Custom antibodies were validated by specific antibody–antigen interaction assays, such as ELISA. Isotype controls were used for antibody validations.
Commercial antibody info and validation info where applicable:
https://www.thermofisher.com/antibody/product/Mouse-anti-Human-IgG1-Fc-Secondary-Antibody-clone-HP6069-Monoclonal/A-10648
https://bxcell.com/product/invivomab-human-igg1-isotype-control/
https://www.abcam.com/goat-mouse-igg-hl-hrp-ab6789.html
Eukaryotic cell lines
Cell line sources: Various, e.g., HEK293FT, Thermo Fisher Cat#R70007
HEK293T-hACW2, Dr Bieniasz’s lab
Vero-E6, ATCC, Cat#CRL-1586TM
Expi293FTM, Thermo Fisher Cat#A14527
HOS-3734, ATCC, Cat#CRL-1543TM
TZMbl, Dr. Sutton’s lab.
Cell lines were authenticated by original vendors, and re-validated in the lab as appropriate, by morphology and PCRs.
All cell lines tested negative for mycoplasma.
No commonly misidentified lines involved.
Animals and other organisms
Laboratory animals: M. musculus,
C57BL/6 J, Jackson laboratory, Cat#000664
B6.Cg-Tg(K18-ACE2)2Prlmn/J, Jackson laboratory, Cat#034860
BALB/c, Jackson laboratory, Cat#000651
Animals are maintained and bred in standard individualized cages with a maximum of five mice per cage, at regular room temperature (65–75 °F, or 18–23 °C), 40–60% humidity, and a 12 h:12 h light cycle for breeding, and 13 h:11 h or 14 h:10 h light cycle for experiments.
Wild animals: No wild animals were used in this study.
Field-collected samples: No field-collected samples were used in this study.
Software and codes
Data collection
ELISA data were recorded by a microplate reader (Perkin Elmer) (no version number).
Antibody binding kinetics for anti-spike mAbs were evaluated by BLI on an Octet RED96e instrument (FortéBio) at room temperature (version 12).
Affinity measurement by surface plasmon resonance (SPR): Kinetics binding measurement for anti-spike mAbs in this study was performed using a Biacore T200 instrument (GE Healthcare) (v3).
Cryo-EM data were acquired on an FEI Titan Krios electron microscope (Thermo Fisher) equipped with a Gatan K3 Summit direct detector in super-resolution mode, at a calibrated magnification of 81,000× with the physical pixel size corresponding to 1.068 Å. Detailed data collection statistics for the Fab-spike trimer complexes are shown in a Supplemental Table. Automated data collection was performed using SerialEM (v3.8).
Data analysis
Standard biological assays’ data were analyzed in Prism (v8 or v9).
SPR results were analyzed by using Biacore T200 Evaluation Software 3.0.
BLI data were analyzed by using Octet Analysis Studio Software 10.0.
Motion correction of the micrographs was carried out using RELION (v3.1.2) and contrast transfer function (CTF) estimation was calculated using CTFFIND4 (v4.1).
Particles were picked automatically by crYOLO (v1.8.1), followed by 2D and 3D classifications without imposing symmetry. The 3D classes with different S trimer conformations were then processed separately by consensus 3D refinement and CTF refinement.
Image processing and 3D reconstruction using cryoSPARC (v3.3.1) produced similar results.
The final map of each body was corrected for K3 detector modulation and sharpened by a negative B-factor estimated by RELION and then merged in Chimera for deposition.
The structure of the ectodomain of the SARS-CoV-2 spike trimer (PDB 6VSB) was used as an initial model and docked into the spike trimer portion of the cryo-EM maps using Chimera (v1.15).
The initial models of Clone 2 and Clone 6 Fabs were generated by homology modeling using SWISS-MODEL (online https://swissmodel.expasy.org) and then docked into the Fab portions of the cryo-EM maps using Chimera.
The initial models were subsequently manually rebuilt in COOT (0.9.7), followed by iterative cycles of refinement in Refmac5 (v5) and PHENIX (v1.19).
The final models with good geometry and fit to the map were validated using the comprehensive cryo-EM validation tool implemented in PHENIX. All structural figures were generated using PyMol (v1.3) (online http://www.pymol.org/) and ChimeraX (v1.2).
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
