key: cord-0854798-xfkeqx4r
authors: Haga, Kei; Takai-Todaka, Reiko; Matsumura, Yuta; Takano, Tomomi; Tojo, Takuto; Nagami, Atsushi; Ishida, Yuki; Masaki, Hidekazu; Tsuchiya, Masayuki; Ebisudani, Toshiki; Sugimoto, Shinya; Sato, Toshiro; Yasuda, Hiroyuki; Fukunaga, Koichi; Sawada, Akihito; Nemoto, Naoto; Song, Chihong; Murata, Kazuyoshi; Morimoto, Takuya; Katayama, Kazuhiko
title: Nasal delivery of single-domain antibodies improves symptoms of SARS-CoV-2 infection in an animal model
date: 2021-04-09
journal: bioRxiv
DOI: 10.1101/2021.04.09.439147
sha: f8269bd8890100bc8936efed7caf4c77568fcb05
doc_id: 854798
cord_uid: xfkeqx4r

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that causes the disease COVID-19 can lead to serious symptoms, such as severe pneumonia, in the elderly and those with underlying medical conditions. While vaccines are now available, they do not work for everyone and therapeutic drugs are still needed particularly for treating life-threatening conditions. Here, we showed nasal delivery of a new, unmodified camelid single-domain antibody (VHH), termed K-874A, effectively inhibited SARS-CoV-2 titers in infected lungs of Syrian hamsters without causing weight loss and cytokine induction. In vitro studies demonstrated that K-874A neutralized SARS-CoV-2 in both VeroE6/TMPRSS2 and human lung-derived alveolar organoid cells. Unlike other drug candidates, K-874A blocks viral membrane fusion rather than viral attachment. Cryo-electron microscopy revealed K-874A bound between the receptor binding domain and N-terminal domain of the virus S protein. Further, infected cells treated with K-874A produced fewer virus progeny that were less infective. We propose that direct administration of K-874A to the lung via a nebulizer could be a new treatment for preventing the reinfection of amplified virus in COVID-19 patients. Author summary Vaccines for COVID-19 are now available but therapeutic drugs are still needed to treat life-threatening cases and those who cannot be vaccinated. We discovered a new heavy-chain single-domain antibody that can effectively neutralize the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that causes COVID-19. Unlike other drug candidates, which prevent the virus from attaching to the receptor on the host cell, this new antibody acts by blocking the virus membrane from fusing with the host cell membrane. We studied the behavior of the new antibody in vitro using VeroE6/TMPRSS2 cells and human lung organoids. When delivered through the nose to infected Syrian hamsters, we found that this antibody could prevent the typical symptoms caused by SARS-CoV-2. Our results are significant because delivering simple drugs directly to infected lungs using a nebulizer could increase the potency of the drugs while reducing the risk of immune reaction that could occur if the drugs escape or are delivered through the blood stream.

From the frequency distribution of VHH clones appearing in the selected library, clone K-874A 120 appeared most frequently (39.5%) after three rounds of in vitro selection, indicating that K-874A 121 has a high affinity to S1 proteins ( Fig 1B) . Biolayer interferometry assay showed the binding 122 affinity of K-874A to SARS-CoV-2 S1 protein is 1.4 nM (Ka (1/Ms) = 6.72E+04, Kd (1/s) = 9.42E-123 05) (Fig 1C) . When a direct antigen enzyme-linked immunosorbent assay (ELISA) was performed 124 using FLAG-tagged K-874A and immobilized recombinant His-tagged S1 protein, we found that 125 K-874A bound with high affinity to the S1 protein of SARS-CoV-2 but not to other coronaviruses 126 including human CoV (HCoV-229E, HCoV-NL63, HCoV-HKU1, HCoV-OC43), MERS-CoV 127 and SARS-CoV-1 ( Fig 1D) . Together, these results indicate that K-874A binds strongly and 128 specifically to the S1 protein of SARS-CoV-2. With its strong binding affinity and specificity to SARS-CoV-2, we investigated K-874A as a 133 potential therapeutic drug for COVID-19. We infected African green monkey kidney (VeroE6) 134 cells expressing transmembrane protease serine 2 (VeroE6/TRMPSS2) with SARS-CoV-2 and 135 measured how well K-874A inhibited SARS-CoV-2 infection in the cells using quantitative real-136 time polymerase chain reaction (qRT-PCR). Half-maximal inhibitory concentration (IC50) 137 calculated from RNA copies was 1.40 µg/mL (Fig 2A) . IC50 indicates the concentration of K-874A 138 required to inhibit 50% of SARS-CoV-2. 139 140

To determine whether K-874A VHH neutralizes SARS-CoV-2 by preventing the virus from 141

attaching to the host cell, we compared the early phases of infection in cells infected with K-874A-142 treated virus and untreated virus. VeroE6/TMPRSS2 cells were incubated with K-874A-treated 143 SARS-CoV-2 and untreated virus for 1 hr and the number of RNA copies that are attached or 144 incorporated into the cells were determined by qRT-PCR. K-874A treatment did not change the 145 levels of virus attachment or early incorporation in the cells (Fig 2B) . At 6 hours post-infection, 146 the number of viral RNA copies in cells infected with K-874A-treated virus remained unchanged 147 while the numbers were significantly higher in cells infected with untreated virus (Fig 2C) . 148

Additionally, expression of the viral protein, NSP8, and double stranded RNA (dsRNA), which is 149 generated during viral RNA replication, were suppressed in cells infected with K-874A-treated 150 viruses ( Fig 2D) . These results indicated that K-874A did not reduce viral attachment. 151

We further investigated whether K-874A acts by inhibiting the virus from fusing with the host cell. 153

VeroE6/TMPRSS2 cell expresses the ACE2 receptor and the TMPRSS2 protease on its surface. 154

To study viral fusion, we transduced the VeroE6/TMPRSS2 cells with a lentivirus coding for the 155 viral S protein into the cells. When the S protein from one VeroE6/TMPRSS2 cell binds to the 156 ACE2 receptor on an adjacent VeroE6/TMPRSS2 cell, the TMPRSS2 protease cleaves and 157 activates the S protein for fusion. To quantify this cell fusion, we used HiBiT technology, which 158 involves the binding of HiBiT (a small 11-amino acid peptide) with a larger subunit, called LgBiT, 159

to form a complex with luciferase activity. VeroE6/TMPRSS2 cells expressing HiBiT or LgBiT 160 were produced and co-cultivated. The HiBiT-LgBiT complex and the resulting luciferase activity 161 form only when the S protein from one cell binds to the ACE2 receptor on an adjacent cell and 162 their membranes are fused. We found that treating VeroE6/TMPRSS2 cells with K-874A 163 suppressed cell fusion in a dose-dependent way ( Fig 2E) . Cells without S protein (S ˗) did not fuse 164 while those expressing S protein but were not treated with K-874A (S+) did fuse. Optical 165 micrographs confirmed these results ( Fig 2F) . Direct binding of K-874A and recombinant ACE2 166 to immobilized recombinant S protein in ELISA further showed that both K-874A and ACE2 167 bound to S protein in a dose-dependent way ( Fig 2G) . However, K-874A did not block ACE2-S 168 protein interaction (right panel in Fig 2G) . Together, these findings suggest that K-874A does not 169 prevent the virus from attaching to the ACE2 receptor on the cell surface. Rather, K-874A prevents 170 the virus from entering the cell by blocking the viral membrane from fusing with the host cell. 171

To estimate the K-874A-binding region on the S protein, we compared the cryo-electron 173 microscopy (cryo-EM) structures of recombinant S trimer and the K-874A-S trimer complex. In 174 the prefusion state, the RBD is known to move upwards to bind ACE2 [13, 14] . We observed that 175 K-874A was located in the vacant space between the NTD and RBD of the S protein in the 176 prefusion conformation (Fig 2H) . CDR1 and CDR2 amino acids and the N-terminal of K-874A 177 formed polar bonds with the NTD of the S protein while CDR2 and CDR3 bonded hydrophobically 178 with the RBD. These findings demonstrate that K-874A neutralizes SARS-CoV-2 via a different 179 route that did not involve ACE2 binding. 180

Because SARS-CoV-2 targets lung tissues and induces severe respiratory disease, we evaluated 183 K-874A VHH in a human lung-derived alveolar organoid, which we have shown is susceptible to 184 SARS-CoV-2 and releases high levels of progeny virus into the culture supernatant[15] (Ebisudani 185 T et al. submitted) . To evaluate whether K-874A can block the production of virus progeny in 186 infected alveolar organoid, we incubated the organoid cells with SARS-CoV-2 for 1 hr before 187 treating the infected cells with K-874A-containing medium for 3 days. Untreated cells released 188 >1x10 5 RNA copies/µL into the culture supernatant at day 3, whereas VHH treatment reduced the 189 progeny production to 1x10 4 RNA copies/µL (Fig 3A) , indicating that K-874A can reduce the 190 production of virus progeny. To test the infectivity of virus progeny produced from K-874A-191 treated and untreated infected cells, we incubated the virus progeny with fresh VeroE6/TMPRSS2 192 cells. Untreated virus progeny showed 78.4 TCID50/µL, while K-874A-treated virus progeny 193 showed less than 1.2 TCID50/µl ( Fig 3B) . These data suggested that the K-874A-treated progeny 194 have a lower infectivity. This is likely due to K-874A VHH binding to the S protein of the virus 195 progeny, preventing them from infecting other cells. 196 197 Infected Syrian hamsters recover after K-874AVHH treatment 198 199 We further tested K-874A VHH as a potential therapeutic drug for COVID-19 using the Syrian 200 hamster animal model. When infected with SARS-CoV-2, Syrian hamsters lose weight but 201 spontaneously heal over time [16] . To determine whether K-874A can improve the symptoms of 202 COVID-19 infection, we infected 6 week-old male Syrian hamsters with SARS-CoV-2 (2 x 10 3 203 TCID50) and intranasally administrated different regimens of K-874A. The animals were treated 204 with either a single dose of K-874A just before the infection or at 1 day after the infection, or 205 multiple doses at 1 and 2 days after the infection (Fig 4A) . Each dose was 3 mg K-874A per animal. 206

All infected animals treated with K-874A did not lose weight whereas infected control animals 207 treated with bovine serum albumin in phosphate buffered saline (BSA/PBS) exhibited weight loss 208 3 days after the infection ( Fig 4B) . Further, all K874A-treated hamsters regardless of treatment 209 regimen had a statistically lower number of viral RNA copies in their lung tissues than BSA/PBS-210 treated hamsters 4 days after the infection (Fig 4C) . Because cytokines are known to be upregulated 211 after an infection [17, 18] , we compared the mRNA levels of several cytokines in the infected lung 212 tissues of the untreated and K-874A-treated animals. While untreated animals showed elevated 213 levels of IL6 and IL10 but not IL1b and TNF-α, in their lungs 4 days after the infection, all K-214 874A-treated animals regardless of the treatment regimen displayed cytokine levels that were 215 similar to uninfected animals ( Fig 4D) . 216 217 When different doses (0.12 mg, 0.6 mg or 3 mg per animal) of K-874A were administered in a 218 two-dose regimen (on Day 1 and Day 2 post-infection), animals receiving the lowest K-874A dose 219 displayed the highest weight loss and viral RNA copies in their lung tissues. Their IL-6 and IL-10, 220

but not IL1b and TNF-α, levels were also elevated compared to those receiving higher doses of K-221 874A ( Supplementary Fig 2) . These data demonstrate that K-874A can improve the symptoms of VHH are promising drug candidates because they are more stable and cheaper to produce than 228 human monoclonal antibodies. They are also amenable to nasal administration, allowing high 229 concentrations of drugs to reach directly to infected lungs and remain effective for longer. Indeed, 230 nasal administered VHH against RSV have been shown to be effective for 3 days [6]. However, 231 due to their low molecular weight, VHH monomers have very short half-lives in the blood stream 232

[4]. VHHs under development as antivirals need to be either made multivalent or modified with 233 human antibody fragments to enhance their antiviral effect or extend their half-life. A previous 234 report showed VHH that bound to the S protein of SARS-CoV-2 could neutralize the virus only 235 when VHH was fused with an Fc domain of a human antibody [7] . 236

Here, using in vitro selection, we identified a standalone anti-SARS-CoV-2 S1 VHH that binds to 238 the S protein of SARS-CoV-2 with a higher affinity than previous VHHs [7, 8] and displays 239 excellent neutralizing ability in VeroE6/TMPRSS2 cells and human normal alveolar-derived cells. 240

Our results show that this VHH neutralizes SARS-CoV-2 by preventing the virus membrane from 241 fusing with the host cell membrane. Cryo-EM analysis of the S protein-VHH complex revealed 242 that the VHH binds between the RBD and NTD region on the S protein, rather than at the interface 243 of the RBD and ACE2. Studies in human lung-derived alveolar organoid indicated that the VHH 244 can reduce the production of virus progeny. Moreover, virus progeny produced from VHH-treated 245 infected cells had a lower infectivity than those that came from untreated infected cells, suggesting 246 that the VHH could prevent the virus from spreading to uninfected cells or persons. Intranasal 247 administration of our VHH to SARS-CoV-2-infected Syrian hamsters prevented weight loss, viral 248 replication in the lungs and the upregulation of cytokines typically caused by SARS-CoV-2 249 infection. 250

Our VHH has several advantages. A previous study showed that inhaling a human monoclonal 252 antibody against SARS-CoV-2 can inhibit virus replication in the lung and nasal turbinate [19] . 253

This suggests that the therapeutic benefits of our VHH could also be delivered using a nebulizer. 254

Nasal administration via a nebulizer is expected to lower the amount of VHH entering the blood 255 stream, which could reduce the risk of immunoreaction against VHH on repeated use. Further, 256 because our VHH displays striking antiviral effects without the additional Fc domain, its risk to 257

Fc-related antibody-dependent enhancement is likely to be low. its mRNA with puromycin linker. cDNA of linked mRNA was revers transcribed and VHH-cDNA 292 complex was produced. High-affinity VHH-cDNA complex to immobilized S1 protein was 293 isolated and its cDNA was amplified. Three rounds of selection were performed and cDNA 294 libraries from rounds 2 and 3 were sequenced and anti-SARS-CoV-2 VHH candidates were 295 translated. (B) Frequency distribution of amino acid sequences corresponding to VHH antibody 296 candidates targeting SARS-CoV-2 S1 subunits in the selected VHH libraries. Sample ID "1" with 297 the highest frequency (39.5%) is clone K-874A. (C) Binding affinity of K-874A to SARS-CoV-2 298 S1 subunits. Biolayer interferometry sensorgram measures the apparent binding affinity of K-299 874A-6xHis to immobilized SARS-CoV-2 S1 fused with sheep Fc. Binding curves for different 300 concentrations of K-874A are shown in different colors. Black curve is 1:1 fit of the data. (D) 301

Direct antigen ELISA measuring the apparent binding affinity of FLAG-tagged K-874A to 302 immobilized S1-6xHis subunits of alpha-(HCoV-229E and HCoV-NL63) and beta-coronaviruses 303 Tang T, Bidon M, Jaimes JA, Whittaker GR, Daniel S. Coronavirus membrane fusion 383 mechanism offers a potential target for antiviral development. Antiviral Res. 2020; 178:104792. 384 Epub 2020 HCl (pH 7.5) with 200 mM NaCl under the following annealing conditions: heating at 90°C for 2 530 min, followed by lowering the temperature to 70°C at a rate of 0.1°C/s, incubating for 1 min, then 531 cooling to 25°C at a rate of 0.1°C/s, incubating for 30 sec., and then stored at 4°C until use. The 532 mixture was irradiated with UV light at 365 nm using a handheld UV lamp (6W, UVGL-58, 533 254/365 nm, 100V; Analytik Jena, USA) for 5 min to obtain mRNA-linker complex. 534

The mRNA-linker complex was translated using a Rabbit Reticulocyte Lysate System, 535

Nuclease-Treated (Promega) at 37°C for 15 min. To synthesized mRNA-linker-VHH complex, 536

KCl and MgCl2 were added to final concentrations of 900 and 75 mM, respectively, and the 537 mixture was incubated at 37°C for 20 min. In vitro selection of VHHs using cDNA display 557 SARS-CoV-2 S1 subunits-His tag (Sino Biological, China) were used as target molecules for 558 screening. Microtiter wells of Nunc-Immuno TM Plate II (ThermoFisher Scientific) were coated 559 overnight at 4°C with 100 µL of 100 µg/mL (R1), 10 µg/mL (R2), 1 µg/mL (R3) recombinant 560 SARS-CoV-2 S1 subunits-His tag which was prepared with PBS. The S1 subunits-coated wells 561

were blocked with 200 µL of 3% BSA in PBST for 2 hours at room temperature, and then washed 562 at three times with 200 µL of HBST. 100 µL of VHH-cDNA complex library was prepared as 563 follows: 50 µL of HEPES buffer containing 1% BSA was added to 50 µL of VHH-cDNA 564 conjugate library for R1, 50 µL of HBST containing 1% BSA, 25 µL of HBT was added to 50 µL 565 of cDNA display library for R2 and R3, respectively. The VHH-cDNA complex library was added 566 to the target-fixed well, incubated for 1 hour at room temperature. The residual VHH-cDNA 567 conjugates were removed by washing at 10 times with 200 µL of HBST. The target-binding VHH-568 cDNA complex was eluted by adding 100 µL of 100 mM Tris (hydroxymethyl) aminomethane 569 (pH 11), gently pipetting, and incubating for 10 min at 37°C. The elution was immediately 570 transferred to PCR cocktails prepared with KAPA HiFi HotStart ReadyMix (2x) (Kapa Biosystems, 571 USA), forward primer (5 ′ -572 GATCCCGCGAAATTAATACGACTCACTATAGGGGAAGTATTTTTACAACAATTACCA 573 ACA-3′), and reverse primer (5′-TTTCCACGCCGCCCCCCGTCCT-3′). The PCR cycle 574 conditions consisted of 95°C for 2 min, followed by 22 cycles (R1) or 24 cycles (R2 and R3) of 575 98°C for 20 sec, 68°C for 15 sec, and 72°C for 20 sec, and then 72°C for 5 min. The PCR products 576 were purified using Agencourt AMPure XP beads (Beckman Coulter Genomics), according to the 577 manufacturer's instruction. The purified DNA libraries were used for translation, as described 578 above, to prepare cDNA display libraries used at the next selection round. 579 580

DNA libraries obtained at R2 and R3 were used as templates in amplicon PCR for Illumina 582 sequencing. The DNA library was quantified using PicoGreen ® dsDNA reagent kit (ThermoFisher 583 Scientific) and prepared at 100 ng/mL with nuclease-free water. 1 st PCR was performed with 12.5 584 µL of KAPA HiFi HotStart ReadyMix (2x) (Kapa Biosystems), 0.5 µL of 10 µM forward primer 585 CCTCCCG-3′), 9 µL of nuclease-free water, and 2.5 µL of template DNA. The PCR cycle 590 conditions consisted of 95°C for 3 min, followed by 16 cycles of 98°C for 20 sec, 62°C for 15 sec, 591 and 72°C for 20 sec, and then 72°C for 5 min. PCR clean-up was performed with Agencourt 592

AMPure XP beads (Beckman Coulter Genomics). Subsequently, index PCR was performed with 593 12.5 µL of KAPA HiFi HotStart ReadyMix (2x) (Kapa Biosystems), each 1 µL of 10 µM forward 594 and reverse primers from Nextera XT Index Kit v2 (Illumina, USA), 8 µL of nuclease-free water, 595 and 2.5 µL of template DNA. The PCR cycle conditions consisted of 95°C for 3 min, followed by 596 8 cycles of 98°C for 20 sec, 55°C for 15 sec, and 72°C for 30 sec, and then 72°C for 5 min. After 597 magnetic bead-based purification of PCR products, the library concentration was measured with 598

PicoGreen ® dsDNA reagent kit (ThermoFisher Scientific) and prepared to 4 nM with nuclease-599 free water. The library was diluted to a final concentration of 7 pM and 5% of PhiX DNA (Illumina, 600

Epub 1997/11/14

cDNA display: 423 a novel screening method for functional disulfide-rich peptides by solid-phase synthesis and 424 stabilization of mRNA-protein fusions

Anti-survivin 428 single-domain antibodies derived from an artificial library including three synthetic random 429 regions by in vitro selection using cDNA display

Structures and distributions of 437 SARS-CoV-2 spike proteins on intact virions

Alveolar Stem Cell Culture Models Reveal Infection Response to SARS-CoV-2. Cell Stem Cell

Simulation of the clinical and 445 pathological manifestations of Coronavirus Disease 2019 (COVID-19) in golden Syrian hamster 446 model: implications for disease pathogenesis and transmissibility

Coordinate induction of 450 IFN-alpha and -gamma by SARS-CoV also in the absence of virus replication

Induction of 454 interferon-gamma-inducible protein 10 by SARS-CoV infection, interferon alfacon 1 and 455 interferon inducer in human bronchial epithelial Calu-3 cells and BALB/c mice

Epub 2010/03/17

Therapeutic activity of an inhaled potent SARS-CoV-2 neutralizing human monoclonal antibody 460 in hamsters

Cytokine storm induced by SARS-CoV-2

PEAR: a fast and accurate Illumina Paired-End 465 reAd mergeR

Molecular Evolutionary Genetics 469 Analysis across Computing Platforms

Bacillus subtilis AprX 473 involved in degradation of a heterologous protein during the late stationary growth phase

Functional 477 receptor molecules CD300lf and CD300ld within the CD300 family enable murine noroviruses to 478 infect cells

Validation of assays to 482 monitor immune responses in the Syrian golden hamster (Mesocricetus auratus)

MotionCor2: 486 anisotropic correction of beam-induced motion for improved cryo-electron microscopy

CTFFIND4: Fast and accurate defocus estimation from electron 490 micrographs

New tools for 494 automated high-resolution cryo-EM structure determination in RELION-3. Elife

Chimera--a visualization system for exploratory research and analysis

SWISS-502 MODEL: homology modelling of protein structures and complexes

Features and development of Coot

Sequencing was performed with the Miseq Reagent Nano kit V2 500 cycle using 601 a MiSeq

400 MHz DDR4 ECC SDRAM) was used to analyze sequencing 606 data. The raw Illumina paired-end reads that passed through the Q30 filter were merged using 607 PEAR software [21]. with default parameter on Linux. Nucleotide sequences encoding VHHs were 608 extracted using combination of basic Linux command lines specifying typical nucleotide 609 sequences conserved on VHHs frame regions. The VHHs-encoding sequences were

Bacillus subtilis 168, which is deficient for nine proteases (i.e., epr, wprA, mpr, nprB, bpr, nprE

23] and contains a sigma factor for a sporulation (sigF)-deficient mutant was 615 used (JP4336082B2). B. subtilis was precultured in L medium

5% NaCl) and cultured in 2xL-Mal medium (2% tryptone, 1% yeast extract, 1% NaCl, 7.5% 617 maltose hydrate, and 7.5 μg/mL MnSO4) with 15 μg/mL tetracycline at 30°C 72 h

Supernatants were subjected to SDS-PAGE and western blotting. SDS-PAGE used

Tricine Gel (Wako, Japan), and then proteins in the gel were transferred 620 to PVDF membrane using Trans-Blot Turbo Mini PVDF Transfer Packs

PVDF membranes were treated with 6xHis Tag Monoclonal 622

Antibody (3D5) HRP (Invitorogen, USA) for 6xHis-tagged proteins, and ANTI-FLAG M2-623

USA) for FLAG ® (DYKDDDDK)-tagged proteins 624 in iBind Western System (Invitrogen)

Blotting Solution

CoVHH1-6xHis and CoVHH1-CoVHH1-6xHis were purified from the supernatants of 627 culture medium using Ni-NTA agarose beads (Wako) and resuspended into PBS with 50 mM

The CoVHH1-FLAG fraction was collected from the supernatants using an Amicon ® 629

Ultra Centrifugal Filter Unit (Merck, USA) and resuspended into PBS with 50 mM imidazol. The 630 purified VHHs were stored at 4°C until use

4°C with 100 µL of recombinant human-CoV-229E, NL63, HKU1, and

SARS-CoV-2 S1 subunits-His tag (Sino Biological) and recombinant SARS-CoV S1 subunits-His 636 tag (The Native Antigen Company, UK) diluted in phosphate-buffered saline (PBS) containing 637

The S1 subunits-coated wells were blocked with 200 µL of 5% skim 638 milk in PBST for 1 hour at room temperature, and then washed at three times with 200 µL of PBST

100 µL of 20 µg/mL CoVHH1-FLAG prepared with PBST were added to the wells, incubated for 640 1 hour, and then washed at three times with 200 µL of PBST. The binding of CoVHH1-FLAG to 641 S1 subunits-His tag was detected by incubating for 1 hour at room temperature with mouse 642 monoclonal ANTI-FLAG ® M2-Peroxidase (Merck) diluted with PBST. The wells were washed 643 three times with PBST, and then 100 µL of substrate buffer

After incubating for 30 min at room temperature, the 646 absorbance at 450 nm was immediately measured with Microplate Reader Infinite M1000 PRO 647

Biolayer interferometry

Binding affinities of CoVHH1 and S1 protein were determined using biolayer interferometry (BLI)

The kinetic buffer was PBST for BLI. The 652 temperature was fixed at 25°C. 50 μL of each measurement solution was added to a 384-well black 653

USA), and the measurements were performed as below. 1) The loading baseline 654 was measured in kinetic buffer for 30 sec. 2) CoVHH1-6xHis was immobilized on a HIS1K

3) The measurement baseline 656 was measured in kinetic buffer for 30 sec. 3) The loaded sensors were dipped into twofold serial 657 dilutions from 245.8 nM of the SARS-CoV-2 (2019-nCoV) Spike S1-Fc Recombinant Protein 658 (Sino Biological) to measure a 180-sec specific binding at the association step. 4) The dissociation 659 was obtained by dipping the biosensors once more time into the kinetic buffer for 240 sec. The 660 data were analyzed using ForteBio Octet analysis software

Fortebio) and kinetic parameters were determined using a 1:1 monovalent binding model

Cells and virus

VeroE6/TMPRSS2 were purchased from JCRB Cell Bank and preserved in our laboratory

CoV-2, 2019-nCoV JPN/TY/WK-521 were provided from National Institute of Infectious 666

) was incubated with serial diluted VHHs for 2 hours 670 at 37ºC and subsequently for 24 hours at 4ºC. VHH-treated virus was incubated with 5

VeroE6/TMPRSS2 cells for 1 hour at 37ºC, and the wells were washed to remove unbound viruses

At 24 hours after infection, the culture supernatant was used to determine the number of RNA 673 copies of virus by quantitative real-time PCR (SARS-CoV-2 Detection Kit, TOYOBO), according 674 to manufacturer's protocol

TCID50 (MOI=50) of SARS-CoV-2 was incubated with 150 µg/ml of VHHs for 2 hours 678 at 37ºC and then 24 hours at 4ºC and inoculated to 5

Similarly, to compare early RNA replication in the infected cells, infected cells were collected at 680 3 or 6 hours after infection. Total RNA was extracted from the collected cells with nucleospin 681 (Takara). Numbers of RNA copies of the virus were determined by quantitative real-time PCR

To quantify the cell fusion induced by S protein, HiBiT technology (Promega) was applied. HiBiT 686 was linked to the C-terminal of ZsGreen (TAKARA) and transduced by lentivirus vector LVSIN-687

and LgBiT was transduced by lentivirus vector LVSIN-IRES-Hyg [24]. Equal 688 numbers of ZsGreen-HiBiT-expressing cells and LgBiT-expressing cells

At 14 hours 690 after transduction, the culture supernatant was replaced to Opti-MEM

and 25 µL of diluted Nano-Glo live solution (Promega) was added into the each well. After mixing, 692 the relative luminescence was measured by Ensight

ELISA assay for direct interaction of recombinant ACE2 and S protein

#RP012383LQ, ABclonal) with PBS was incubated to 696 immobilize on the maxisorp plate (464718, Thermo). After blocking with 1% BSA/PBS, 1 µg of

Then, serial diluted 698 recombinant ACE2 (from 1 µg/ml to 0.06 µg/mL) (ab151852, abcam), subsequent rabbit anti 699

Cell 700 signaling) were incubated. To detect K-874A binding to immobilized S protein

874A (1 µg/ml to 0.002 µg/ml) in 1%BSA/PBS was incubated, and K-874A which bound to S 702 protein was detected by HRP-conjugated anti VHH antibody

Between each incubation, wells were washed with PBST. To color

phenylenediamine dihydrochloride (p6662, Merck) was used as a substrate of HRP and reaction 705 was stopped 2M H2SO4. Signaling was quantified by Ensight

Immunofluorescence staining of SARS-CoV-2-infected cells

After 709 treatment with 0.05% Triton-X and 3% BSA/PBS for permeabilization and blocking, cells were 710 incubated with anti-NSP8 (5A10, GeneTex) or anti-dsRNA (rJ2, Merck) for an hour and 711 subsequently with Alexa 568-conjugated goat anti-mouse IgG (A110301

Hoechest for an hour at room temperature. Images were taken by BZX800 (Keyence)

Infection to Alveolar organoid and VHH treatment

Human alveolar-derived organoids were established

The cells were infected with SARS-CoV-2 (MOI 5) for 1 hour at 37℃, 718 washed to remove unbound virus, and compounded into Matrigel. After the compete medium 719 solidified, VHH (15 µg/mL) was added or not

VHH treatment of SARS CoV-2-infected Syrian hamsters

2x10 3 TCID50 of SARS-CoV-2 was nasally inoculated to 7 weeks age of Syrian hamsters

mg of K-874A was nasally administrated at just before, 1 day or 1 day and 2 days after virus 725 inoculation. The weight of each hamster was measured every day up to 4 days after inoculation

Lung tissues were homogenized at 4 days after inoculation, and RNA was extracted with 727

Nucleospin RNA (TAKARA) for qRT-PCR

Quantitative real-time PCR for inflammatory cytokines

Primers and probes for hamsters were designed as reported[25]. The information of primers and 731 probes is shown in Table 1. qRT-PCR experiments were performed on

according to manufacturer's instructions. Briefly, 50 ng of RNA was used for each reaction as 734

The final concentrations of each test primer and probe set for target gene were 0.4 and 735 0.2 µM, respectively. The final concentrations of the internal control primer and probe set were 736 0.2 and 0.1 µM, respectively. Cycling conditions were as follows: 10 min at 50℃ (initial reverse 737 transcription), 20 sec at 95℃ (inactivation and initial denaturation), and 40 cycles of 3 sec at 95℃, 738 followed by 60 sec at 60℃ per cycle

S protein trimer of SARS-CoV-2 (# SPN-C52H9, Acrobiosystems, 600 μg/mL) and K-874A

Aliquot (2.5 μL) of the 743 sample was placed onto a holey-carbon copper grid (Quantifoil Micro Tools, R 1.2/1.3) previously 744 glow-discharged in a plasma ion bombarder (PIB-10

The cryo-EM data were acquired 747 with a Titan Krios at 300 kV (Thermo Fisher Scientific) and a Gatan K3 camera (Gatan Inc.) at a 748 nominal magnification of 64,000, corresponding to 1.11 Å per pixel on specimen

) was used with a slit width of 20 eV to remove inelastically 751 scattered electrons. Individual movies were subjected to per-frame drift correction by MotionCor2 752

The contrast transfer function parameters of each micrograph were estimated using 753

3D refinement, and local resolution 754 calculation were performed with RELION3.1 software [28]. Particles were selected from 6,984 755 micrographs and the final 3D reconstruction was computed with 115,297 particles. The final map 756 resolution was 3.9 Å (gold standard FSC criterion) by imposing C3 symmetry. Using the 757

Mask creation" in 761 RELION 3.1. The images containing K-874A, RBD, and NTD were subtracted from the 762 symmetry-expanded images using "Particle subtraction" in RELION3.1. The prepared mask was 763 used for subtraction. These sub-particles were subjected to 3D classification without shift and 764 rotation, and the sub-particle images (143,647 particles) of the selected good classes were used for 765 3D refinement by imposing C1 symmetry. The resolution of the subtracted map was estimated to 766 be 5.0 Å based on a gold standard FSC

The SWISS-MODEL [30] server was used to generate homology models of K-874A, and S protein 769 monomer of SARS-CoV-2 (SPN-C52H9) using the atomic model of VHH-72 (PDB ID: 6WAQ) 770 and 2019-nCoV S protein (PDB ID: 6VSB) as a template

The homology 772 models of K-874A, RBD and NTD were rigid-body-fitted into the map using the COOT [31] and 773 then refined using PHENIX

Animal experiments were approved by the President of Kitasato University through the 777

Institutional Animal Care and Use Committee of Kitasato University (20-011), and performed in 778 accordance with the Guidelines for