Structure and function analysis of a potent human neutralizing antibody CA521LALA against SARS- CoV-2


 Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent of the ongoing coronavirus disease 2019 (COVID-19) pandemic, which has resulted in ~1,119,431 deaths. There is currently no approved vaccines or therapeutics for treating COVID-19. The SARS-CoV-2 Spike protein promotes entry into host cells and is considered a key therapeutic target by many researchers. Here we describe the identification of several monoclonal antibodies that target the SARS-CoV-2 Spike protein. One human antibody, CA521LALA, demonstrated neutralization potential by immunizing human antibody transgenic mice. CA521LALA showed potent SARS-CoV-2-specific neutralization activity against SARS-CoV-2 pseudovirus and authentic SARS-CoV-2 infection in vitro. The LALA mutation introduced to CA521 abrogates the binding with Fc receptors or complement receptors reducing antibody-dependent enhancement seen with anti-SARS-CoV antibodies. CA521LALA also demonstrated having a long half-life of 9.5 days in mice and 9.3 days in rhesus monkeys. CA521LALA inhibited SARS-CoV-2 infection in SARS-CoV-2 susceptible mice at a therapeutic setting with the virus titer of the lung reduced by 4.5 logs. Structural analysis by cryo-EM revealed that CA521LALA recognizes an epitope overlapping with angiotensin converting enzyme 2 (ACE2)-binding sites in SARS-CoV-2 receptor binding domain (RBD) in the Spike protein. CA521LALA blocks the interaction by binding all three RBDs of one SARS CoV-2 spike trimer simultaneously. These results demonstrate the importance for antibody-based therapeutic interventions in the treatment of COVID-19 and identifies CA521LALA a promising antibody that reacts with SARS-CoV-2 Spike protein to strongly neutralize its activity.


Introduction
World Health Organization as of October 21th has reported that the coronavirus disease 2019  caused by the virus severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused 1,119,431 deaths globally (https://www.who.int/). SARS-CoV-2 is closely related severe acute respiratory syndrome coronavirus (SARS-CoV) and belongs to the lineage B of the genus Betacoronavirus in the Coronaviridae family. There are no vaccines approved for the treatment of COVID-19. Monoclonal antibodies (MAbs) are promising candidates to combat emerging viruses. For example, in the case of Ebola virus, MAbs MAb114 and REGN-EB3 have shown striking treatment bene ts, reducing the mortality rate from ~67% to 33.5-35.1% for all patients and to 9.9-11.2% for patients with low viral loads 1 . MAbs are also being considered as promising therapeutics for COVID-19 patients 2 .
Sharing an amino acid sequence identity of ~80% in the envelope-located spike glycoprotein (Sike protein), both SARS-CoV-2 and SARS-CoV use human angiotensin-converting enzyme 2 (hACE2) to enter host cells. Cellular entry is achieved by the homotrimeric S-mediated virus-receptor engagement through the receptor-binding domain (RBD) followed by virus-host membrane fusion 3,4,5,6,7 . The Spike protein is a viral factor that mediates attachment to cells and fusion of the viral and cellular membrane, functions associated with the S1 and S2 subunits respectively 3,4,5,6,7 . The primary amino acid sequence of the S1 subunit of coronaviruses determine host receptors. The structure and function of the SARS CoV-2 Spike protein has been determined 8 . S protein trimer of coronavirus S structures always have some RBDs in an ''up'' conformation and some in the ''down'' conformation 9,10,11,12 . 3D structures of SARS-CoV-2 S protein also reveal some RBDs in ''up'' conformations and some in the ''down'' conformation 8,13,14,15 . The crystal structure of the RBD of the spike protein of SARS-CoV-2 bound to the cell receptor ACE2 has also been determined. This structural analysis identi ed the residues of the RBD that are essential for ACE2 binding 16 .
Potent neutralizing antibodies including S230, m396, 80R, CR3022 have been shown to target the SARS-CoV RBD in S1, disabling receptor interactions. Unfortunately, none demonstrate potent neutralizing activity against SARS-CoV-2 12,14,17,18,19,20 . Additionally, previous studies have shown that pre-existing serum antibodies associate with poor outcomes in patients with the 2009 in uenza infection or SARS-CoV infection 21,22,23 . This may be due in part to antibody-dependent enhancement (ADE) which could lead to acute respiratory injury and is a potential risk for antibody therapeutic developed against SARS-CoV infection 24 . ADE has been showed in part to be induced mainly through engaging viruses and monocytes by anti-virus antibody binding with Fc Receptor or complement receptors on monocytes 24,25,26,27,28,29,30 . Recent studies show that sera from the COVID-19 patients with severe disease has higher NAb titer than sera from mild or asymptomatic patient 31 . Given the highly phylogenetic relationship between SARS-CoV and SARS-CoV-2, the risk of ADE will likely need to be mitigated for anti-SARS-CoV-2 development.
Here we report the isolation of several highly potent neutralizing MAbs against SARS-CoV-2 from human antibody transgenic mice. Given that ADE is a potential risk for antibody therapeutics against SARS-CoV-2 infection we introduced LALA mutation to potentially abrogate the effect in these neutralizing MAbs.
One human MAb designated CA521 LALA demonstrated potent SARS-CoV-2-speci c neutralization activity in vitro and in vivo and had no risk of ADE. Pharmacokinetics revealed that CA521 LALA is also stable in serum of mice and rhesus monkeys. Cryo-electron microscopy characterization of the SARS-CoV-2/CA521 LALA shows that it recognizes an epitope overlapping with angiotensin converting enzyme 2 (ACE2)-binding sites in SARS-CoV-2 receptor binding domain (RBD) in the Spike protein. Structural analysis reveals that the CA521 LALA can block the interaction by binding all three RBDs of one SARS CoV-2 spike simultaneously weather it was in 'up' or 'down' conformations.

Results
MAb CA521 LALA can block binding of SARS-CoV-2-RBD to hACE2 receptor and speci cally bind the Spike protein of SARS-CoV-2 Potential SARS-CoV-2 targeting antibodies were identi ed by screening immunized human antibody transgenic mice by phage display (Supplementary Fig. 1). For this seven human antibody transgenic mice were immunized with recombinant Spike ectodomain (Cat 40589-V08B1, Sino Biological), RBD protein (Cat 40592-V05H, Sino Biological), Spike S1 protein (Cat 40591-V02H, Sino Biological) and Spike S2 protein (Cat 40590-V08B, Sino Biological) of SARS-CoV-2, respectively. These antigens were also used as the baits for screening the phage antibody library. 60 potential clones were identi ed for in vitro evaluation. Given that ADE is a potential risk for antibody therapeutics against SARS-CoV-2 infection, LALA mutation was introduced into various targets to abrogated or eliminate the binding with Fc receptors or complement receptors during production of human monoclonal antibody. This phage display screen, in vitro evaluation and mutagenesis identi ed an excellent candidate in the novel human anti-Spike antibody CA521 LALA (Supplementary Fig. 1).
Further analysis demonstrated that CA521 LALA blocked the binding of SARS-CoV-2-RBD with recombinant hACE2 receptor strongly compared to hACE2 protein, with IC50 as 0.343 nM, 8.887 nM respectively (Fig.   1a). CA521 LALA also showed the ability to block the binding of recombinant ACE2 to SARS-CoV-2 Spike expressing 293F cells and the binding of SARS-CoV-2 RBD to ACE2 expressing 293T cells ( Supplementary Fig. 2). Flow cytometry (FACS) experiments revealed that CA521 LALA could also speci cally bind to SARS-CoV-2 Spike protein transfected CHO-K1 cells (Fig. 1b). CA521 LALA could bind SARS-CoV-2-Spike speci cally and not cross-react with SARS-CoV and MERS-CoV Spike ( Fig. 1c-e). Biolayer Interferometry (BLI) using the Octet RED96 system (FortéBio) assessed the binding kinetics of CA521 LALA . The measured equilibrium constant (KD) of CA521 LALA with SARS-CoV-2-RBD, S1 and spike trimer (Shuimu BioSciences Ltd.) were 0.698 ± 0.028 nM, 6.508 ± 0.655nM and 1pM (below detection limit), respectively ( Fig. 1f-h). CA521 LALA shows modi ed binding a nity to various Fc or complement receptors in BIAcore or cell based assay As mentioned previously, ADE is a potential risk for neutralizing MAbs agninst SARS-CoV infection. (19,20). The LALA mutation was introduced to CA521 potentially abrogates the binding with Fc receptors or complement receptors. To con rm the result of the mutation, the binding a nity of CA521 LALA and CA521(IgG1) to various human Fc or complement receptors were examined by Surface Plasmon Resonance (SPR) or Elisa assay. The binding a nity of CA521 LALA with Gama Fc Receptor 1 (CD64) was below the detection limit and in the case of FcγRIIA R167 and FcγRIIB/C it was >20 μM (Fig. 2a-f). A nity of CA521 LALA to C1q was also signi cantly reduced compared with CA521(IgG1) and slightly reduced compared with Opdivo (IgG4) (Supplementary Fig. 3).
To further con rm abrogated binding in cell based assay, Antibody-Dependent Cellular Phagocytosis (ADCP) was conducted with SARS-CoV-2 Spike protein transfected CHO-K1 cells as target cells and macrophages derived from CD14 + monocytes as effector cells. Target cells were stained with CFSE and macrophages were stained with APC-anti-CD206. Double stained macrophages were considered to be macrophages that have phagocytized target cells. Phagocytosis rate of CA521 LALA was similar to the isotype control and lower than that of CA521(IgG1), as 3.33%, 3.76% and 12.74% respectively (Fig. 2g-i).
For CA521 LALA , phagocytosis by macrophages was almost completely avoided consistent with the notation that the binding of CA521 LALA with Fc receptors was almost abrogated.

CA521 LALA inhibited SARS-CoV-2 infection in vitro and in vivo
The in vitro neutralization abilities of CA521 LALA against SARS-CoV-2 infection was evaluated using a pseudoviruses system expressing Spike protein of SARS-CoV-2 and plaque-reduction neutralization test (PRNT) against an authentic SARS-CoV-2 infection of Vero cells. CA521 LALA can inhibit pseudoviruses transduction into Huh-7 and hACE2 expressing HEK293T cells with IC50 at 0.121 nM and 0.104 nM, respectively (Fig. 3a, b). Additionally, CA521 LALA also exhibited strong neutralizing activity against an authentic SARS-CoV-2 strain with a PRNT50 of 0.73 nM (Fig. 3c).
Next, we sought to assess the correlation between in vitro neutralization and in vivo protection. To in ltration, alveolar septal thickening and distinctive vascular system injury developed in BALB/C mice belonging to the PBS control group at 3 dpi (Fig. 3f). In contrast, the lungs in mice from the CA521 LALA treated group only showed very mild in ammatory cell in ltration, and no obvious lesions of alveolar epithelial cells or focal hemorrhage (Fig. 3g). These results demonstrate that CA521 LALA is a potent neutralizing antibody, which is effective in conferring protection on mice against SARS-CoV-2.

CA521 LALA is stable and has a long half-life in mice and rhesus monkeys
The pharmacokinetics of the human antibody CA521 LALA was studied in mice and rhesus monkeys. To achieve this a single intravenous injection dose of CA521 LALA was given to C57BL/6 mice (N=4) and rhesus monkeys (N=3) at 10 mg/kg and 50 mg/kg. ELISA (enzyme-linked immunosorbent assay) was used to determine the concentration of CA521 LALA in serum. Following a single-dose 10 mg/kg intravascular injection in C57BL/6 mice, CA521 LALA showed a bi-exponential serum concentration-time pro le with a short distribution phase followed by a long elimination phase, with a terminal half-life (t1/2, λz) of 9.5 ± 4 days, C max of 96±5 μg/mL and AUC (0-t) of 647 ± 43 day*μg/mL (Fig. 4a). Following a singledose 50 mg/kg intravascular injection in rhesus monkeys, CA521 LALA showed a bi-exponential serum concentration-time pro le with a short distribution phase followed by a long elimination phase, with a terminal half-life (t 1/2, λz ) of 9.3 ± 5 days, C max of 975 ± 110 μg/mL and AUC (0-t) of 5101 ± 2020 day*μg/mL (Fig. 4b).
Cryo-electron microscopy analysis of the SARS-CoV-2 Spike protein-CA521 LALA complex To investigate the interaction between CA521 LALA and the SARS-CoV-2 Spike protein we determined the structure of this complex using cryo-EM. Details of sample preparation, data collection and EM analysis are in the methods and extended data(Extended Data Table 3). In brief, puri ed extracellular domain of S protein (S-ECD) was incubated with full-length CA521 LALA and then the mixture was applied to prepare cryo-EM grids. After 2D classi cation good particles were selected out ( Supplementary Fig. 4) and then subjected to 3D classi cation. In the good class corresponding to 614,999 particles (Fig. 5a) all three RBDs of the Spike trimer are "up" with three Fabs bound asymmetrically. Particles in this class were subjected to further 3D re nement with C1 symmetry to generate the consensus map of the complex. The consensus map at lower threshold showed that the two Fabs bound to RBD-1 and RBD-2 connected with each other, which suggests the bivalent binding of one CA521 LALA IgG molecule with two RBDs (Fig. 5b). In the consensus map the quality of RBD region is poor, therefore focused 3D classi cation and re nement of RBD region were performed to improve the local resolution of RBD-Fab interface. Finally, the local resolution of RBD region was improved to near 5 Å. The improved maps show that three RBDs bound with Fabs possess different conformations (Fig. 5c). Besides, the 5 Å resolution allows us to deduce the binding sites of CA521 LALA on RBD.
The residues of RBD mediating interaction with Fab were determined by the recognizable electron cloud density of Fab on the RBD surface (Fig. 5d). Two patches at RBD contributed to binding with CA521 LALA . The rst patch includes residues 483-489(VEGFNCY). The other patch includes residues 446-451 (GGNYNY) and 493-497 (QSYGF). All these residues are located in the receptor-binding motif (RBM, 438-506) that interacts directly with ACE2.
We superimposed the crystal structure of RBD-ACE2 (cyan model) with three RBDs (red density) of S-CA521 LALA complex. The density of all three Fabs are in clash with ACE2, so MAb CA521 LALA may block the interaction of SARS-Cov-2 with human ACE2 by occupying the binding site directly (Fig. 5e). To compare the CA521 LALA and ACE2-interaction residues on the SARS-CoV-2 RBDs, we mapped them to the RBD sequences (supplementary Fig. 5). Seven amino acid positions were shared by CA521 LALA and ACE2 for the interaction with RBD, including GLY 446, Tyr 449, Phe 486, Asn 487, Tyr 489, Gln 493, Gly 496.
Some CA521 LALA interaction residues on the SARS-CoV-2 RBDs are no more than 3 amino acids away from the direct contact residues of ACE2 on RBD, such as Val 483, Cys 488, Gly 447 and Phe 497. Main H014 and CB6-interaction residues on the SARS-CoV-2 RBDs were also mapped to the RBD sequences ( Supplementary Fig. 5). Epitope of CA521 LALA is signi cantly different from CB6 and H014.
One IgG binding with two adjacent RBDs in one Spike trimer simultaneously may contribute to the potent blocking and neutralization ability of CA521 LALA . Among the residues of SARS-CoV-2 RBD mediating the interaction with MAb CA521 LALA , mutation variants G446V, V483I, V483A have been found in nature (reported in GISAID database up to May 6, 2020), and percentage for V483A is higher than 0.1%. The other three variants of L452R, F490L, P491R whose percentage is below 0.1% are adjacent to the interaction residues of CA521-Spike. We detected whether V483A could affect the a nity between RBD and CA521 LALA . Recombinant RBD proteins harboring V483A mutation and several other mutations exhibited an basically indistinguishable binding a nity to CA521 LALA (Supplementary table 4).

Discussion
There is an urgent need for therapeutic interventions to combat the COVID-19 caused by SARS-CoV-2.
Neutralizing antibodies are an important tool that can effectively ght the coronavirus 1  In summary, our study identi es CA521 LALA as an excellent neutralizing antibody against SARS-CoV-2 with three major advantages: direct competitive binding with ACE2, binding all three RBDS of one spike simultaneously and bivalent binding of one IgG. CA521 LALA shows promise as an effective intervention to the COVID-19 pandemic caused by SARS-CoV-2. Potent neutralizing ability, the low risk of ADE and long half-life make it an ideal candidate to move on to clinical trial.

Regents, Mice, Cells and Viruses
Reagents, cell lines and viral strains used in this study are listed in Supplementary Table 1.
Phage display library construction RNA was extracted from spleen cells of immunized mice by Trizol method. cDNA synthesis was performed using Transcriptor First Strang cDNA Synthesis Kit. The construct of the phage library was carried out according to the method described in Phage Display: A laboratory manual 32 The variable regions of the heavy and light chains were obtained from the cDNA by PCR. Scfv were obtained by overlapping the variable regions of heavy and light chains by PCR, and then were digested with S I and ligated into plasmid pCOMB3x. The ligation products are electro transfected into E. coli TG1 competent cells. After adding helper phage and culturing overnight, the supernatant was precipitated with PEG8000 and NaCl. Libraries were got by resuspending the pellet.

Panning of phage libraries
Plates coated with immunogen (Spike S1+S2 or Spike S1 or Spike RBD) or streptavidin-magnetic beads loading biotin-protein were used to capture phages with interest ScFvs. After washing by PBST, captured phage was eluted and then used to infect E.Coli TG1. Scfvs was expressed and its binding and blocking activity was tested by ELISA. Positive hits were obtained and sequenced.

Recombinant Vector Construction
Recombinant antibody heavy chain variable region and light chain variable region were ampli ed (2 × Phanta Max Master Mix, Vazyme, P515-01) using the positive clones screened from the library as the template. Overlap PCR was conducted to assemble variable region and signal peptide. Puri ed gene fragments were separately fused (ClonExpress II One Step Cloning Kit, Vazyme, C112-01) into the linearized pcDNA3.4 vectors with constant regions. The recombinant plasmid was prepared for production. (For recombinant human mAb production, the cDNA's encoding the CA521 mAb variable regions of the heavy and light chains were cloned into expression.)

Antibody Expression and Puri cation
Candidate antibodies were expressed with Expi-CHO Expression system(gibco) for 12days and the supernatant was harvested and puri ed by protein A resin (GE healthcare). The antibodies were further puri ed by Q FF (GE healthcare) and Capto S ImpAct (GE healthcare) sequentially and then changed to buffer containing 10mM CH 3 COONa▪3H 2 O, 30mM NaCl, 5% sucrose, 0.03% tween-20, pH5.0/6.0.

ELISA -based receptor-binding inhibition assay
High binding ELISA plates were coated with 0.5 μg/mL recombinant soluble 2019-nCoV Spike RBD at 4 o C overnight, and then were blocked with 3% skim milk powder in PBST (PBS containing 0.05%Tween-20) at 37 o C for 1 hour, following two times washing with PBST. Serially diluted CA521 was mixed with 0.04 μg/mL ( nal concentration) biotinylated recombinant human ACE2 and then was incubated with coated RBD in the plates at 37 o C for 1 hour. After washing, the biotinylated ACE2 binding to coated RBD was detected by HRP-conjugated Strep second mAb. Inhibition rate%=(OD450 of no antibody -OD450)/ OD450 of no antibody*100%. Irrelevant mAb with the same constant region of CA521 was used as isotype. Experiments were performed in triplicate, value=Mean±standard error.

Cross-reactivity by Elisa analysis
Recombinant SARS-CoV-2 Spike S1+S2 protein (40589-V08B1, Sino Biological), SARS Spike protein (SPN-S52H5, Acro) and MERS-CoV Spike protein (40069-V08B, Sino Biological) were coated on high binding ELISA plates with 0.5 μg/mL at 4 o C overnight. Plates were blocked with 3% skim milk powder in PBST at 37 o C for 1 hour and then washed two times with PBST. Serially dilutions of mAbs were added following incubation at 37 o C for 1 h. Plates were washed two times and then HRP-goat anti-human IgG (H+L) mAb was used to detect antibodies binding to the Spikes. Irrelevant mAb with the same constant region of CA521 was used as isotype. Experiments were performed in triplicate, value=Mean±standard error.

Cell based binding for CA521
SARS-CoV-2 Spike protein transfected CHO cells were harvested and washed by FACS buffer (0.2% BSA in PBS) two times. 1E5 CHO-SARS-CoV-2-Spike cells were stained with isotype control IgG or CA521 at a concentration of 0.74 μg/mL at 4 o C for 1 h. After washing by FACS buffer two times, cells were incubated in dark with FITC-anti-human IgG Fc 2 nd mAb at 4 o C for 30 min and then analyzed by NovoCyte 2060R ow cytometry. Irrelevant mAb with the same constant region of CA521 was used as isotype.
Experiments were performed in triplicate, value=Mean±standard error.

A nity to SARS-CoV-2 Spike RBD mutants from different virus strain variants
The binding kinetics were assessed by Surface Plasmon Resonance (SPR) assay using the BIAcore 8K system. The measured equilibrium constant (KD) Measurements were performed at room temperature with CM5 chip, which was amino coupled by human antibody capture kit. HBS-EP+ buffer (150 mM NaCl, 10 mM HEPES, 3 mM EDTA and 0.05% (v/v) surfactant P20 pH 7.4) was used as running buffer. The blank channel of the chip served as the negative control. CA521 was captured on the chip at 400-500 response units. Serial dilutions of SARS-CoV-2-RBD mutants (from 50 nM to 3.125 nM with 2-fold dilution) were applied to ow over the chip surface which was regenerated with 3 M MgCl 2 after each cycle. The a nity was calculated using a 1:1 (Langmuir) binding t model with BIAevaluation software.
Fortebio analysis of antibody binding to CoV spike antigens RBD Antibodies to be tested were diluted to the concentration of 4 μg/mL with PBST and then immobilized onto Octet Fab2G biosensors for real-time association and dissociation. After arriving the Signal Change Threshold 1.1 nm and washed in PBST biosensor tips were immersed into the wells containing RBD protein (40592-V05H, Sino Biological) of serial dilutions and allowed to associate for 200 seconds followed by a dissociation step of 400 seconds. KD was calculated using a 1:1 binding model in Octet Data Acquisition 9.0.0.49 (Sartorius AG).
Fortebio analysis of antibody binding to CoV spike antigens Spike-Trimer Antibodies were diluted to the concentration of 3 μg/mL with PBST and then immobilized onto ProA biosensors. After arriving at the Signal Change Threshold 0.8 nm and washing with PBST, biosensor tips were immersed into the wells containing Spike-Trimer protein of serial dilutions and allowed to associate  Table 2) were applied to ow over the chip surface which was regenerated with 10 mM glycine-HCl (pH 1.5) after each cycle. The a nity was calculated using a 1:1 (Langmuir) binding t model with BIA evaluation software.
In this method, SARS-CoV-2(2019-nCoV) spike protein was used as the capture reagent, and goat antihuman IgG, monkey ads-HRP was detecting agent. Results are shown as mean ± standard error (N = 3).
The main PK kinetic parameters were calculated using Phoenix WinNonlin.
Neutralization assay Pseudoviruses (80033) purchased from Beijing SanYao Science & Technology Development Company were produced and titrated as described previously 33 .
SARS-CoV-2 pseudovirus were incubated with 3-fold serially diluted CA521 at 37 °C for 1 hour, and then cells suspension of Huh-7(Japanese Collection of Research Bioresources [JCRB], 0403) or HEK-293T-hACE2(CHENGDU NB BIOLAB CO., LTD) were added to the mixtures. After 24 h incubation at 37°C, neutralizations potencies of mAbs were evaluated in a luciferase assay. Luciferase activity was measured using Bio-Glo Assay reagent as a substrate (Promega). The percentage of infectivity was calculated as ratio of luciferase readout in the presence of mAbs normalized to luciferase readout in the absence of mAb. The half maximal inhibitory concentrations (IC50) were determined using 4-parameter logistic regression (GraphPad Prism). Experiments were performed in triplicate.

Infectious SARS-CoV-2 Neutralization assay
Neutralizing activity of mAbs was measured using a standard plaque reduction neutralization with Vero cells. Brie y, 5-fold serial dilutions of mAbs were added to approximately 100 PFU of SARS-CoV-2 and incubated for 1 h at 37 °C. Then, the mixture was added to Vero cell monolayers in a 24-well plate in duplicate and incubated for 1 h at 37 °C. The mixture was removed, and 1 ml of 1.0% (w/v) LMP agarose (Promega) in DMEM plus 4% (v/v) FBS was layered onto the infected cells. After further incubation at 37°C for 2 days, the wells were stained with 1% (w/v) crystal violet dissolved in 4% (v/v) formaldehyde to visualize the plaques. PRNT50 values were determined using non-linear regression analysis. All experiments were performed followed the standard operating procedures of the approved Biosafety Level-3 facility.
Infection and antibody treatment of mice