A Feasible Alternative Strategy Targeting Furin Disrupts SARS-CoV-2 Infection Cycle

Figures

• 

  FIG 1. Design of furin competitive FuG1 strategy. (A, B) Schematic of SARS-CoV-2 and SARS-CoV show RBD domain, S1/S2, and S2' sites. (C) Ribbon structure of SARS-CoV-2 spike monomer (PDB: 6ZGI). Red spheres: S1/S2 represent furin substrate residues. Yellow spheres: S2' residues. RBD domain is in green. (D) 22-IgG1 (PDB: 6W41) half-body schematic inserted with Fc extendable linkers harboring competitive furin engaging residues (pink, cartoon only). (E) Schematic showing competitively engaged active site cleft of furin (PDB: 6HZD) with Fc extended peptide (pink) of 22-FuG1 antibody. VH, gold; VL, gray; furin, blue ribbon with key substrate-binding pocket residues (H194, S253, P256, N295, S368) shown as green stick model (16). (F) Sequences of FuG1 (optimal competitive lead) and cFuGI (control cleavable) Fc-extended linkers. (G) Cleavage score of FuG1 and cFuG1 Fc-extended linkers based on PiTou algorithm. (H) The amino acid sequence of human IgG4-Fc tagged recombinant DR5. The sequence in blue represents 101 amino acids (∼12KDa) at the N-terminal of DR5 MCRKCR (P1-P6) residues (red), and the sequence in black is 261 amino acids (∼30KDa), C-terminal to MCRKCR. (I) Amino acids in blue…

• 

  FIG 2. Confirmation of in vivo stability, furin inhibitory function, and flexibility of Fc-extended FuG1 peptide strategy. (A, B) Genetic construction schematic of various recombinant antibodies in (A) were run under reducing conditions in (B). Anti-FOLR1 antibody, farletuzumab is shown as a regular IgG1 (lane 1 in B). A 50 kDa and 25 kDa bands corresponding to heavy and light chains are evident in reduced sample in lane 1. Lane 2 is a bispecific antibody where a single chain variable fragment (scFv) was covalently generated in continuous of CH3 domain of VH, hence 75 kDa and 25 kDa bands corresponding to heavy and light chains are evident in reduced sample. Lane 3 is a linkered farletuzumab IgG1, where 3′ end of light chain is covalently linked by G4S linkers to the 5′ end of heavy chain. As a result, a single band around ∼75 kDa is evident. Lane 4 is exactly the same as lane 3 except a G4S linker and 18 amino acid cFuG1 peptide is engineered after the CH3 domain only in one chain (hole). Because the cFuG1 (Fig. 1F and G) peptide is an optimal PiTou furin substrate, it was cleaved in CHO cells during expression. As a result, size of lane 4 is similar to linkered IgG in lane 3. Lane 5…

• 

  FIG 3. TGN independent cleavage of ectopically expressed spike. (A) 293-ACE2 cells were treated with indicated Brefeldin-A concentrations at the time of spike transfection. Lysates were analyzed for spike and DR5 24 h later using immunoblotting. (B to D) Cells grown in 150 cm2 dishes (divided in groups as indicated on top) were transfected with the spike. ER and Golgi fractions were enriched 24 h later using Minute ER and Golgi enrichment kits (see Materials and Methods section) as indicated on top. Lysates were subjected to immunoblotting with indicated ER and Golgi resident markers along with spike. (E, F) Same as B-D, except along with spike, GFP (with N-terminal membrane sorting peptide) was used as a positive control for traditional TGN sorting. (G) Schematic of crude endosome (CE) enrichment assay for data shown in (H) (see Materials and Methods section). (H) Along with spike and spike+GFP total lysates, enriched CE from spike+GFP co-transfected cell were analyzed for indicated endosomal, lysosomal marker using immunoblotting.

• 

  FIG 4. Fc-extended peptide does not interfere with target binding and activity FuG1 antibodies. (A) Genetic construction schematic of 22-IgG1, 22-FuG1. Fc extended FuG1 peptide is shown in green. (B) A reducing SDS-Gel of 22-IgG1 (lane 1) and 22-FuG1 (lane 2). Red arrows depict the difference in the sizes of IgG1 and FuG1 VH. (C) A reducing SDS-Gel of 6.30 and 12.25 IgG1 and FuG1 antibodies. Red arrows show the difference in the sizes of IgG1 and FuG1 VH. (D) Cell viability assay of KMTR2-IgG1, KMTR2-FuG1, and KMTR2-FuG1 with linkered Fab against ovarian cancer cells. (E) Cell viability assay of trastuzumab-IgG1, trastuzumab-FuG1 and trastuzumab-FuG1 with linkered Fab against breast cancer cells. (F) IgG4-Fc tagged rFOLR1 antigen was coated on 96-well plates overnight. Coated plates were treated with the increasing concentrations of either farletuzumab-IgG1 or farletuzumab-FuG1 and IgG1 control antibodies as indicated. Following numerous washes, the HRP conjugated secondary antibody that is specific to IgG1-Fc (but not IgG4-Fc) was used to measure the binding strength using TMB substrate and ELISA plate reader capable of reading at 450 nm (n = 3). (G, H) The binding kinetics of…

• 

  FIG 5. FuG1 strategy interferes with S1/S2 cleavage in spike transfected cells. (A) Top, Schematic of SARS-CoV-2 showing RBD (green) and non-RBD binding antibodies (blue). Middle, Schematic of SARS-CoV-2 showing wild type S1/S2 sequence and RRAR mutant sequence. Bottom, Sequence comparison of Fc-extended FuG1 peptide with Fc-extended random charge control (RCC) peptide. (B) 4 h post-WT spike DNA transfection, indicated antibodies were added onto the HEK-293 cells (70% to 80% confluent) and lysates were prepared after 24 h posttransfection. For control Furin mutant ΔRRAR spike plasmid was used. (C) Same as B, except ACE2-stable HEK-293 cells were used. (D) Same as C except Calu3 cells were tested. (E) Same as C except 6.30-IgG1 and 6.30-FuG1 antibodies were used. In lane 4, 5-fold KMTR2 was added along with 6.30-FuG1 and in lane 5, 5-fold 6.30-IgG1 was added to significantly out-compete with 6.30-FuG1 binding with spike. (F, G) Same as (E), except ACE2-stable HEK-293 cells were used, and instead of 22-FuG1, other spike RBD targeting (6.29-IgG1, 6.29-FuG1, 6.30-IgG1, 6.30-FuG1), and non-RBD targeting (12.19-FuG1, 12.25-FuG1) antibodies were used as indicated on the top. Murine DR5…

• 

  FIG 6. FuG1 antibodies destabilize S0 for particle incorporation in virus-producing cells. (A) Schematic of generation of WT and ΔRRAR spike expressing pseudovirus. (B) VSV-G, SARS-CoV and WT or furin mutant SARS-CoV-2 spike transfected BHK1 cells were added with replication-restricted G*ΔG-luciferase-rVSV particles. 48 h later lysates, harvested pseudovirions were analyzed for S0 and S2 fragments using immunoblotting. (C) Harvested pseudoviral particles bearing WT spike proteins were incubated with indicated antibodies for 1 h at 37, followed by transduction of 293-ACE2 cells. After 24 h, a virus-encoded luciferase signal was measured from the lysates. ΔRRAR pseudoviral particles were used as control. Error bar indicated SEM (n = 3). (D) Schematic workflow of experiments shown in E, F, and G. (E) Same as B, except similar to 293-ACE2 transfection experiments (Fig. 4B and C), spike transfected BHK1 cells were pretreated with indicated antibodies during pseudoviral production. (F) WT spike transfected BHK1 cells were treated with indicated antibodies (4 h posttransfection) during pseudoviral production. 48 h later lysates, harvested pseudovirions were analyzed for S0 and S2 fragments…

• 

  FIG 7. Spike S0 destabilization by FuG1 strategy is proteasome dependent. (A) Workflow of data shown in B. (B) Spike transfected, non-FuG1 treated (lanes 1, 2) and spike transfected plus 22-FuG1 treated (lanes 4, 6) were subjected to immunoblotting analysis of spike ±MG-132 (10 μM). Lanes 3 and 5 were loaded with control ER-enriched lysates (5 μg) to confirm if equal amount of translating spike protein was present ±MG-132 proteasome inhibitor in FuG1 treated samples. (C) Spike, S0 signal intensity in total lysates (±MG-132 treated) after normalization with GAPDH is shown. (D) 4 h post-spike DNA transfection, indicated antibodies (on top) were added onto the 293-ACE2 cells ± increasing (20 μM and 50 μM) MG-132 concentrations. Lysates were prepared 36 h posttransfection following by immunoblotting against spike. (E) Experimental control to confirm that MG-132 treatment does not interfere with WT spike processing and non-processing of ΔRRAR spike. (F) WT spike transfected BHK1 cells were pre-treated with indicated antibodies (± 20 μM and 50 μM MG-132 or PBS alone) after 4 h. 12 h later replication-restricted G*DG-luciferase-rVSV particles were added on to the cells. 48 h later harvested…

• 

  FIG 8. Selectivity of FuG1 strategy against spike expressing cells. (A) Flow cytometry analysis of surface FOLR1 in HEK-293 cells. (B) Flow cytometry confirmation of similar level of FOLR1 expression in 293-ACE and HCC-1806 cells. (C) Schematic of the coculture experiment described in E and F. (D) Schematic showing preferred binding of 22-FuG1 (red) toward high spike expressing 293-ACE2 cells, while farletuzumab-FuG1 (Blue) distributes equally against both cell lines expressing similar levels of FOLR1 (see B). (E, F) Cocultured cells were treated with 22-IgG1, farletuzumab-FuG1, and 22-FuG1 antibodies 4 h after spike-EGFP construct transfection. Cells were imaged 30 h later at ×10 magnification using Evos fluorescence microscope. (G) Number of syncytia from F. Error bar indicated SEM (n = 3). (H) Same experiment as E to G, except additional internal transfection control HA-tagged murine-DR5 (HA-muDR5-pCDNA3.1) was co-transfected with spike-EGFP construct. Total lysates from spike-transfected 293-ACE2 cells alone, and cocultured-cells, were subjected to lysate preparation 48 h later, followed by immunoblotting analysis of S0 and S2. HA-muDR5 served an internally equivalent transfection…

• 

  FIG 9. Working mechanism of FuG1 strategy. (A) In SARS-CoV-2 infected (viral replicating) cells, after RNA packing and viral structural assembly in endoplasmic-reticulum Golgi-intermediate compartment (ERGIC), the spike harboring RNA capsid assembly complex is shuttled toward deacidified lysosomal, late, early, recycling or membrane endosomes route (12) rather than TGN secretory pathway (see Fig. 3). The endosomal furin mediates proteolytic processing of spike S1/S2 before viral release. If cell surface enriched furin also contributes to spike cleavage could not be ruled out. Nonetheless, furin cleavage generates fusion-ready S2 fragments, allowing spike to bind NRP1(4) effectively to enhance overall viral cellular entry in ACE2+ acceptor cells. (B) Membrane and endosomal recycling and spike targeting FuG1 antibodies make use of spike binding as an anchor to competitively disrupt furin proteolytic processing of spike S1/S2 in the membrane-endosomal network. Furthermore, large size complex aggregates (slower migrating on gel) are generated potentially due to tripartite SARS-CoV-2 spike-FuG1-furin multivalent interactions (see Fig. 6). The resulting higher-order aggregates are degraded by…