Plasmids, antivirals, and antibodies
The chemically synthesized human codon-optimized S protein-coding cDNAs, which were designed based on the publicly available protein sequence in the National Center for Biotechnology Information database (ca. Wuhan-Hu-1: YP_009724390.1), were obtained from Twist Bioscience (South San Francisco, CA, USA) or through Bionics Inc. (Seoul, S. Korea). The synthetic DNA fragments were PCR-amplified and cloned into pcDNA3.1 vector (Invitrogen, Waltham, MA, USA; cat. no. V79020) using the In-Fusion HD cloning kit (Takara Bio, Kusatsu, Shiga, Japan; cat. no. 638947). Mutations were introduced into cDNAs using the QuikChange II XL Site-Directed Mutagenesis kit (Agilent, Santa Clara, CA, USA; cat. no. 200521) to construct expression vectors for S variants with a single amino acid change. The expression vectors for the S variants with multiple amino acid changes or deletions were generated using the In-Fusion HD cloning kit (Takara Bio). The pcDNA3.1-hACE2 used to express human ACE2 (hACE2) protein was constructed by inserting the PCR-amplified hACE2 cDNA (NP_001358344.1) between the Nhe I and Xho I sites. The plasmid expressing hTMPRSS2 with an N-terminal HA epitope tag was purchased from Sino Biological Inc. (Beijing, China; cat. no. HG13070-NY). The plasmids used for the dual split protein (DSP)-based cell-to-cell fusion assay were constructed as previously described39,40. Briefly, the Renilla luciferase (Rluc) gene was PCR-amplified from phRL-TK (Promega, Madison, WI, USA, cat. no. E6241), and the eGFP gene was PCR-amplified from a codon-optimized synthetic gene fragment obtained from Bionics Inc. (Seoul, S. Korea). The Rluc gene was split between the 229th and the 230th amino acid to make the two subunits, the N-terminal Rluc (nRluc) and the remaining C-terminal Rluc (cRluc). The GFP gene was split at the end of the 7th β sheet and the start of the 8th β sheet to make GFP1–7 (1–157 aa) and GFP8-11 (158–231 aa). pDSP1–7 was constructed by linking nRluc with GFP1–7 with a 4-amino acid residue linker (GLQG). pDSP8-11 was constructed by linking GFP8–11 and cRluc with a 2-aa residue linker (VD). Each gene fragment was PCR-amplified and cloned into a pcDNA 3.1 vector by using the In-Fusion HD cloning kit (Takara Bio). Lipofectamine 2000 (Invitrogen; cat. no. 11668019) was used for plasmid transfection.
Camostat (cat. no. SML0057) and E-64d (cat. no. E8640) were purchased from Sigma-Aldrich (St. Louis, MO, USA). A short synthetic dsRNA, poly(I:C) was obtained from Invivogen (San Diego, CA, USA; cat. no. TLRL-PICW).
Antibodies were obtained as follows: mouse monoclonal antibody targeting the C-terminal region of the SARS-CoV-2 S protein (GeneTex, Irvine, CA, USA; cat. no. GTX632604); rabbit polyclonal antibody targeting the N-terminal region of the SARS-CoV-2 S protein (GeneTex; cat. no. GTX135384); mouse monoclonal antibody against SARS-CoV nucleocapsid (N) protein (Sino Biological; cat. no. 40143-MM08); rabbit polyclonal antibody targeting N-terminus region of human ACE2 (Genetex; cat. no. GTX101395); rabbit polyclonal antibody against human GAPDH (Abcam, Cambridge, UK; cat. no. ab9485); mouse monoclonal antibody targeting the 872–891 amino acid sequence within the human α-actinin (Santa Cruz Biotechnology Inc, Dallas, TX, USA; cat. no. sc-166524); mouse monoclonal antibody against MLV p30 (Abcam; cat. no. ab130757).
Cell culture and virus infection
HEK293T (human embryonic kidney-derived cells), Calu-3 (human lung adenocarcinoma cells), A549 (adenocarcinomic human alveolar basal epithelial cells), and VeroE6 (African green monkey kidney cells) cell lines were obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA). The VeroE6 cell line stably expressing hTMPRSS2 (Vero E6/TMPRSS2) was a gift from Dr. Seong-Jun Kim (Korea Research Institute of Chemical Technology, KRICT, Daejeon Korea). The VeroE6 cell line stably expressing hTMPRSS2 and hACE2 (Vero E6-TMPRSS2-T2A-ACE, referred to as VeroE6/A2-T2) was obtained through BEI Resources, NIAID, NIH (cat. no. NR-54970). The HEK293T/hACE2 and A549/hACE2 cell lines, which stably express human ACE2 receptor, were established by transducing HEK293T and A549 cells with an hACE2-expressing recombinant lentivirus and selecting a single cell clone by limiting dilution of the transduced cells in the presence of 2 or 3 µg ml− 1 puromycin. Third-generation lentiviral transfer vector encoding hACE2, the second-generation lentiviral packaging plasmid (Addgene plasmid #12260), and a VSV G-expressing vector (Addgene plasmid #12259) were used to prepare the lentiviruses used in transduction experiments.
The HEK293T, Vero E6, Calu-3, and A549 cells were cultivated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U ml− 1 of penicillin, and 100 µg ml− 1 streptomycin at 37°C in 5% CO2. The A549/hACE2 cells stably expressing human ACE2 receptors were maintained in a complete medium with 3 µg ml− 1 of puromycin. Vero E6/TMPRSS2 and VeroE6/A2-T2cells were maintained in a complete medium with 150 mg ml− 1 of hygromycin and 10 µg ml− 1 of puromycin, respectively.
All SARS-CoV-2 viruses used in this study (two S-clade strains, NCCP 43326 and NCCP 43331; Omicron BA.5 strain, NCCP 43426) were obtained from the National Culture Collection for Pathogens, Korean Center for Disease Control (Osong, S. Korea). SARS-CoV-2 was propagated in VeroE6 cells for S-clade and VeroE6/TMPRSS2 cells for Omicron BA.5. Cells were infected at a multiplicity of infection (MOI) of 0.01. Virus titer was determined by plaque assay on VeroE6 cells for S-clade SARS-CoV-2 and VeroE6/TMPRSS2 cells for Omicron BA.5 as previously described41.
S protein structure modeling
AlphaFold2, provided by Neurosnap Inc., was used to predict the monomeric structure of the S protein42, with the highest-ranking model selected as the representative image.
RNA sequencing analysis
Viral RNA from the original P0 (10 µl) or P1 sample (125 µl) was extracted using TRIzol LS reagent (Invitrogen; cat. no. 10296028). RNA was subjected to variant calling sequencing using barcode-tagged contiguous sequencing (BTseq) by Celemics Inc., Seoul, Korea.
Real-time reverse-transcription quantitative PCR (RT-qPCR)
Total RNA from cells or tissue homogenate was extracted using TRIzol reagent (Invitrogen; cat. no. 15596026). The SARS-CoV-2 genomic RNA copy number was determined by RT-qPCR using a specific set of primers and a probe targeting the ORF1ab along with the real-time PCR Master Mix (TOYOBO, Osaka, Japan; cat. no. QPK-101) as described previously41. Pseudovirus genome copy number was quantified using RT-qPCR. Packaged viral RNA was extracted using TRIzol LS reagent (Invitrogen) and subjected to RT-qPCR using primers targeting the 5′ LTR of MLV RNA43. Standard RNAs were prepared through in vitro transcription using a T7 MEGAscript kit (Invitrogen; cat. no. AM1334), followed by electrophoresis on an 8 M urea-5% polyacrylamide gel and purification via gel-extraction.
TNF-α and IFN-β mRNA expression levels were determined using the ΔΔCt method and were normalized to the expression level of GAPDH mRNA, as described44.
Immunoblotting
Cells were lysed with RIPA buffer (150 mM NaCl, 1.0% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris, pH 8,0, 1× protease inhibitor cocktail) and centrifuged at 13,000g for 15 min at 4 ℃. For the analysis of S proteins loaded onto pseudoviruses, the MLV-containing culture supernatant was subjected to ultracentrifugation (150,000g for 2 h at 4 ℃) over 20% (w/v) sucrose cushion in a TL-100 centrifuge using the TLA100.3 rotor to collect pseudovirus pellets, as described45. Subsequently, both pseudovirus samples and cell lysates were subjected to SDS-PAGE, and the proteins were transferred to a PVDF membrane (GE Healthcare Life Sciences, Chicago, IL, USA; cat. no. 10600023). The membrane was then blocked with 5% bovine serum albumin and probed with the appropriate set of primary and secondary antibodies.
Pseudovirus entry assay
SARS-CoV-2 S protein (with a deletion of the C-terminal 19-amino acids ER-retention signal)-pseudotyped murine leukemia virus (MLV) was produced as previously described41,46.
Cell-to-cell fusion assay
The DSP-based SARS-CoV-2 S fusogenic activity assay was performed as previously described47. Briefly, HEK293T cells co-transfected with pDSP1–7 and a full-length S protein-expression vector (donor cells) and HEK293T cells expressing hACE2 and pDSP8–11 (acceptor cells) were co-cultured in a 12-well plate with a total seeding density of 1.5 × 106, by mixing the donor and acceptor cells at a 1:1 ratio at 24 h post-transfection. After 12 or 24 h, the cells were harvested and subjected to a luciferase assay using the Rluc Assay System (Promega; cat. no. E2820). Luminance (ALU) measured in a GloMax Navigator Microplate Luminometer (Promega) was normalized to microgram lysate protein.
Brightfield microscopy was employed to visualize the fusion-induced multicellular syncytia using an inverted phase contrast microscope (OPTIKA, Ponteranica, Italy). Confocal microscopy was performed to quantity the reconstituted GFP fluorescence generated by fusion between donor and acceptor cells (1.8 × 105 each per well) co-cultured in a collagen (Thermo Fisher Scientific; cat. no. A1048301) pre-coated (5 µg cm− 2) 4-well chamber slide (Thermo Fisher Scientific, cat. no. 154526). A fluorescence image was captured using an LSM 880 confocal microscope (Carl Zeiss AG, Jena, Germany). Image reconstruction was conducted using the ZEN software (Zeiss).
mRNA-LNP preparation
The BA.4/5 S-encoding plasmid, which has a template for 3′-end A-tailing (approximately 120-nt), was linearized with BbsI, purified using the Gel Extraction kit (Bionics; cat. no. BNROP-0020), and used as a template for in vitro transcription with the MEGAscript T7 transcription kit. When indicated, ψ-UTP (TriLink, San Diego, CA, USA; cat. no. N-1019) was used to substitute for UTP to prepare ψ-UTP-incorporated mRNA. The CleanCap Reagent AG (3′ OMe) (TriLink; cat. no. N-7413) was added to the in vitro transcription reaction mixture for co-transcriptional 5′-end capping. The capped, A-tailed mRNAs were resolved on formaldehyde-denaturing agarose (0.7%) gel and stained with ethidium bromide to assess RNA integrity before use.
mRNA was transfected into cells using either Lipofectamine 2000 (Invitrogen) or LNPs. mRNA encapsulation within LNPs was carried out using the NanoAssemblr Spark (Precision Nanosystems, Vancouver, Canada). Briefly, a mixture of ionizable (ALC-0315; MedChemExpress, cat no. HY-138170), structural (cholesterol, Sigma-Aldrich; cat. no. C8667), and helper [distearoylphosphatidylcholine (DSPC), Avanti Polar Lipids, Inc., Alabaster, AL, USA; cat. no. 85030] lipids along with polyethylene glycol (PEG)-conjugated lipid (ALC-0159, MedChemExpress; cat. no. HY-138300) were combined with the mRNA in citrate buffer (pH 4) at a 1:6 ration (mRNA:lipid, molar ratio). The resulting encapsulated mRNA-LNP mixture was subjected to dialysis against PBS with 8% sucrose, as described48. mRNA encapsulation efficiency was calculated using the Quant-it RiboGreen RNA Assay kit (Invitrogen; cat. no. R11490) as described49. mRNA-LNP particle size and uniformity were evaluated by dynamic light scattering using a particle size analyzer (ELSZ-1000, Otsuka Electronics, Osaka, Japan).
Cytokine ELISAs
Serum cytokine levels were detected via ELISA one day after mRNA-LNP immunization. To measure lung tissue cytokine levels in mice challenged with SATS-CoV-2, lung homogenates were incubated with Triton X-100 at a final concentration of 1% for 1 h at room temperature to inactivate the virus. The concentrations of mIFN-α, mIL-1β, mTNF-α, mIFN-γ, and mIL-6 in sera were determined using the ELISA kits obtained from PBL Assay Science (Piscataway, NJ, USA; cat. no. 42120-1 for mIFN-α), Elabscience (Houston, Texas, USA; cat. no. E-EL-M0037 for mIL-1β), and Thermo Fisher Scientific (cat. no. 88-7324-88 for mTNF-α, cat. no. 88-7314-88 for mIFN-γ, and cat. no. 88-7064-88 for mIL-6) according to the manufacturer’s recommendation.
IgG ELISAs
SARS-CoV-2 S and RBD-specific IgGs were detected via ELISA. Plates were coated with 1 µg ml− 1 of the antigen protein (Sino Biological; cat. no. 40589-V08H33 for BA.5 S and cat. no. 40592-V08H131 for BA.5 RBD) in phosphate buffer saline (PBS) overnight at 4°C. Plates were then blocked for 6 h with PBST (PBS with 0.01% Tween-20) containing 2% BSA (Sigma-Aldrich; cat. no. 10735094001) at room temperature. After blocking, plates were washed with PBST and incubated with mouse sera diluted in the blocking buffer for 3 h at room temperature on a rocking stand. Following incubation with the sera, plates were washed four times with PBST, and α-S or α-RBD IgG bound to the plate was then detected using HRP-conjugated goat anti-mouse IgG (H + L) antibody (Invitrogen; cat. no. G21040), anti-mouse IgG1 antibody (Invitrogen; cat. no. A10661), or anti-mouse IgG2a antibody (Invitrogen; cat. no. M32207) at a 1:1500 dilution for 1.5 h at room temperature on a rocking stand. After washing the plates four times with PBST, 100 µl of TMB substrate (Thermo Fisher Scientific; cat. no. 34028) was added to each well for the detection of HRP-conjugated antibodies. The reaction was stopped by adding 1 M phosphoric acid, and the absorbance at 450 nm was read using the GloMax-Multi Detection System (Promega).
Endpoint titer was defined as the serum dilution fold that yielded an absorbance equal to the cut-off value. The cut-off optical density value was set to the plate background OD450nm detected at the highest dilution of the sample sera plus 10× standard deviation, as previously described15. Endpoint titer was determined by plotting a dose-response curve (OD450 nm vs. reciprocal serum-dilution fold) and four-parameter nonlinear regression analysis using the GraphPad prism 8·0 (GraphPad Software Inc., San Diego, CA, USA). The dose-response nonlinear regression curve equation and the equation used to determine S-binding IgG endpoint titer by extrapolation are as follows:
A nonlinear regression equation for the dose-response curve:
$$\text{Y}=\text{B}\text{o}\text{t}\text{t}\text{o}\text{m} +\frac{\text{T}\text{o}\text{p}-\text{B}\text{o}\text{t}\text{t}\text{o}\text{m}}{1+{\left(\frac{\text{I}\text{C}50}{\text{X}}\right)}^{\text{H}\text{i}\text{l}\text{l}\text{S}\text{l}\text{o}\text{p}\text{e}}}$$
An equation for endpoint titer determination:
$$\text{E}\text{n}\text{d}\text{p}\text{o}\text{i}\text{n}\text{t} \text{t}\text{i}\text{t}\text{e}\text{r}=\frac{\text{I}\text{C}50}{{\left(\left(\frac{\text{T}\text{o}\text{p}-\text{B}\text{o}\text{t}\text{t}\text{o}\text{m}}{{\text{C}\text{u}\text{t}}_{\text{o}\text{f}\text{f}} \text{O}\text{D}-\text{B}\text{o}\text{t}\text{t}\text{o}\text{m}}\right)-1\right)}^{\frac{1}{\text{H}\text{i}\text{l}\text{l}\text{S}\text{l}\text{o}\text{p}\text{e}}}}$$
Pseudovirus neutralization assay (PNA)
A neutralization assay using SARS-CoV-2 S protein-pseudotyped MLV (SARS2pp) was performed to determine neutralizing antibody titers in the sera, as previously described46. Briefly, equal volumes (60 µl each) of diluted mouse sera and diluted SARS2pp, which yields about 1 × 106 ALU per well, were mixed (n = 3) and incubated for 1 h at 37 ℃. Then VeroE6/A2-T2 cells, which were grown for 24 h following seeding at 1.5 ×104 per well on 96-well white plates, were transduced with the pseudotyped virus and mouse sera mix (100 µl per well). Following a media change at 12 h post-transduction, cells were further cultivated for 36 h and harvested for luciferase assay using the Nano-Glo luciferase assay system (Promega; cat. no. N1120) and the GloMax Navigator Microplate Luminometer (Promega). The 50% neutralizing dilution (ND50), which was defined as the serum dilution causing a 50% reduction in luciferase activity compared to the wells displaying no inhibition, was determined by plotting a dose-response curve (log10 ALU vs. reciprocal serum-dilution fold) and four-parameter nonlinear regression analysis using the GraphPad prism 8·0 (GraphPad Software Inc). The dose-response nonlinear regression curve equation and the rearranged equation used to determine the ND50 by extrapolation are as follows:
A nonlinear regression equation for the dose-response curve:
$$\text{Y}=\text{B}\text{o}\text{t}\text{t}\text{o}\text{m} +\frac{\text{T}\text{o}\text{p}-\text{B}\text{o}\text{t}\text{t}\text{o}\text{m}}{1+{\left(\frac{\text{I}\text{C}50}{\text{X}}\right)}^{\text{H}\text{i}\text{l}\text{l}\text{S}\text{l}\text{o}\text{p}\text{e}}}$$
An equation for ND50 determination:
$$\text{N}\text{D}50=\frac{\text{I}\text{C}50}{{\left(\left(\frac{\text{T}\text{o}\text{p}-\text{B}\text{o}\text{t}\text{t}\text{o}\text{m}}{(\text{T}\text{o}\text{p}-{\text{l}\text{o}\text{g}}_{10}2)-\text{B}\text{o}\text{t}\text{t}\text{o}\text{m}}\right)-1\right)}^{\frac{1}{\text{H}\text{i}\text{l}\text{l}\text{S}\text{l}\text{o}\text{p}\text{e}}}}$$
Reporter virus neutralization assay (RVNA)
A nanoluciferase (Nluc)-expressing live, chimeric SARS-CoV-2 carrying the Omicron BA.5 S gene (derived from the BA.5 strain NCCP 43426) was generated using a full-length cDNA of the G-clade SARS-CoV-2 YS00623 that was cloned in a bacterial artificial chromosomal plasmid to obtain pBAC-SARS-CoV-2/YS006(BA.5_S)_Nluc, in which the ORF7 was replaced by the Nluc gene, as previously described50. The recombinant chimeric virus, rYS006(BA.5_S)_Nluc, was rescued by co-culturing of BHK-21 cells transfected with pBAC-SARS-CoV-2/YS006(BA.5_S)_Nluc with VeroE6/TMPRSS2 cells at 6 h post-transfection, as previously described50. The co-cultured cells were incubated until clear signs of cytopathic effects (CPE) were observed. Once significant CPE was observed in the coculture, supernatant (passage 0, P0 viral stock) was harvested and passaged onto VeroE6/TMPRSS2 cells to obtain the P1 viral stock used in the neutralization assay. Equal volumes (60 µl each) of diluted mouse sera and diluted rYS006(BA.5_S)_Nluc (95 PFU per well), which yields about 1 × 106 ALU per well, were mixed (n = 3) and incubated for 1 h at 37 ℃. Then, the mixture (100 µl) was used to infect VeroE6/TMPRSS2 cells, which were grown for 24 h following seeding at 1.5 × 104 per well on 96-well white plates. Growth media was changed 1 h post-infection and cells were further cultivated for 24 h. Harvested cells were subjected to luciferase assay. The ND50 was determined using the same method used in the PNA assay described earlier.
Live virus plaque reduction neutralization (PRNT)
SARS-CoV-2 Omicron BA.5 variant (NCCP NCCP 43426) propagated in VeroE6/TMPRSS2 cells, which form ~ 40 plaques in each well of a 12-well plate, was mixed with serially diluted mouse sera and incubated for 1 h at 37°C. After incubating the mixture with VeroE6/TMPRSS2 cells for infection, the cells were overlaid with 1% agarose. When plaques became visible, cells were fixed with 10% formaldehyde and stained with 1% crystal violet. The ND50, which was defined as the serum dilution causing a 50% reduction in plaque number compared to the wells displaying no inhibition, was determined by four-parameter non-linear regression analysis of a dose-response curve (% neutralization vs. reciprocal serum-dilution fold) using the GraphPad prism 8·0 (GraphPad Software Inc). The % neutralization at each serum dilution = (plaque number formed w/o sera − plaque number formed w/ sera)/(plaque number formed w/o sera)
Interferon-γ (IFN-γ) enzyme-linked immunosorbent spot (ELISpot) assay
ELISpot assay was carried out using murine IFN-γ single-color enzymatic ELISpot assay kit (ImmunoSpot, Cleveland, OH, USA) according to the manufacturer’s instruction. Briefly, spleens from immunized mice were harvested two weeks after the last vaccination. The splenocytes of immunized mice were then plated on IFN-γ antibody-coated PVDF-backed plates and stimulated with 0.1 µg ml− 1 overlapping peptide pools spanning the full-length SARS-CoV-2 BA.4/5 S protein (JPT, Berlin, Germany; cat. no. PM-SARS2-SMUT-10). After overnight incubation at 37°C in a humidified atmosphere containing 5% CO2, the plates were washed with PBS and PBST and the secreted IFN-γ was detected with a set of detection antibodies. The number of spot-forming units (SFUs) was counted automatically using an ELISpot reader (Immunospot Analyzer, ImmunoSpot).
Animal experiments
All animal experiments were performed following the Korean Food and Drug Administration guidelines. Experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee of the Yonsei Laboratory Animal Research Center (Permit No: IACUC-A-202203-1427-02). At the termination of experiments, all mice were euthanized by CO2 inhalation.
BALB/c mice (7–8-week-old) were immunized by intramuscular injection of mRNA-LNP (50 µl) into the ipsilateral hind limb for prime-boost vaccination, using an insulin syringe (BD, Franklin Lakes, NJ, USA; cat. no. 328868). Blood samples were obtained through tail-clip bleeding or cardiopuncture. The collected blood samples were left to clot for 30 min at room temperature and were centrifuged at 6,000g for 25 min. Before use, all mouse sera were heat-inactivated by incubating for 30 min at 50°C.
Challenge infection
A vaccine efficacy study using BALB/c mice was carried out in a biosafety level 3 (BL3) facility at the Avison Biomedical Research Center (ABMRC), Yonsei University College of Medicine. The study was conducted under IBC permit number A-202009-260-01. The animal experiments involving SARS-CoV-2 were approved by the IACUC at Yonsei University College of Medicine (IACUC number: 2020 − 0227). Animals received two (D0 and D14) IM immunizations of mRNA-LNP vaccine or PBS. BALB/c mice were intranasally challenged with 1 ×105 PFU of SARS-CoV-2 Omicron BA.5 variant. BALB/c mice were euthanized 2 days post-infection and lung viral titer was determined by performing plaque assay using VeroE6/A2-T2 cell and qRT-PCR of the lung homogenate. Four right lung lobes (superior, middle, inferior, and post-caval) were homogenized in serum-free DMEM using a bead beater. The resulting homogenate was clarified by centrifugation at 4°C for 15 min. Supernatants were collected and subjected to viral titer analysis where viral RNA was detected via qRT-PCR and infectious viral load was determined via PRNT.
Immunohistochemical analysis
The left lung lobe was subjected to histological analysis. Paraffin blocks containing the left lung lobe were sectioned at approximately 4 microns and stained using hematoxylin and eosin (H&E) or immunostained using a rabbit monoclonal antibody against SARS-CoV-2 nucleocapsid protein (Sino Biological; cat. no. 40588-R002) and rabbit polyclonal antibodies against CD-3 protein (Dako, Glostrup, Denmark; cat. no. A0452).
Statistical analysis
Results are presented as the mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software Inc.). Differences between groups were considered statistically significant at P < 0.05. (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant.)
Material availability
All reagents generated in this study will be made available on request after completion of a Materials Transfer Agreement.