An alpaca nanobody neutralizes SARS-CoV-2 by blocking receptor interaction

Abstract

SARS-CoV-2 enters host cells through an interaction between the spike glycoprotein and the angiotensin converting enzyme 2 (ACE2) receptor. Directly preventing this interaction presents an attractive possibility for suppressing SARS-CoV-2 replication. Here, we report the isolation and characterization of an alpaca-derived single domain antibody fragment, Ty1, that specifically targets the receptor binding domain (RBD) of the SARS-CoV-2 spike, directly preventing ACE2 engagement. Ty1 binds the RBD with high affinity, occluding ACE2. A cryo-electron microscopy structure of the bound complex at 2.9 Å resolution reveals that Ty1 binds to an epitope on the RBD accessible in both the 'up' and 'down' conformations, sterically hindering RBD-ACE2 binding. While fusion to an Fc domain renders Ty1 extremely potent, Ty1 neutralizes SARS-CoV-2 spike pseudovirus as a 12.8 kDa nanobody, which can be expressed in high quantities in bacteria, presenting opportunities for manufacturing at scale. Ty1 is therefore an excellent candidate as an intervention against COVID-19.

Fig. 1. Nanobody discovery.

a Overview of the nanobody generation process. b Variant frequencies quantified by next generation sequencing (NGS) across successive enrichment steps. Identified using receptor-binding domain (RBD) bait, Ty1 exhibits the greatest total fold change of all nanobodies, increasing in proportion over 10,000-fold between initial and final libraries. c Sequence of Ty1. Complementarity-determining regions (CDRs) are indicated. Amino acid color labels; hydrophobic, blue; positive charge, red; negative charge, magenta; aromatic, cyan; polar, green; cysteine, pink; glycine, orange; proline, yellow.

Fig. 2. Ty1 neutralizes SARS-CoV-2 by binding to SARS-CoV-2 spike glycoprotein.

a VSV G or SARS-CoV-2 spike pseudotyped lentivirus (PSV) was incubated with a dilution series of Ty1, Ty1-Fc, or control nanobody (NP-VHH1). Infectivity relative to cells infected with pseudotyped virus in the absence of nanobody is shown. Neutralization by Ty1 was repeated in duplicate across six assays, neutralization by Ty1-Fc was repeated in duplicate across two assays, and the error bars represent the standard deviation. b Cells were transfected with a plasmid harboring the SARS-CoV-2 spike for 24 h. Cells were fixed, permeabilized, and stained with Ty1-AS635P (black and red) or left unstained (gray). Cells were analyzed by flow cytometry. Cell counts are presented as % of max (representative histogram). c Vero E6 cells were infected with SARS-CoV-2 at a MOI of 1 for 24 h. Cells were fixed, permeabilized, and stained for DNA (blue), dsRNA (green), and with Ty1-AS635P (red). Pictures were taken by fluorescence microscopy and representative examples are shown. Scale bar, 20 µm. d ACE2 expressing HEK293T cells were trypsinized, fixed, and stained with RBD-AS635P alone (blue), or preincubated with NP-VHH1 (green) or Ty1 (red). Cells were analyzed by flow cytometry.

Fig. 3. RBD binds to Ty1 with a K_D of 5–10 nM.

a RBD bound to surface-immobilized Ty1 (red curves), but not to NP-VHH1 (blue curves). Almost equal nanobody immobilization levels of about 0.7 nm were obtained by first loading Ty1 and then NP-VHH1. In bio-layer interferometry (BLI), binding of molecules over time is recorded as sensorgrams recording the shift in wavelengths (unit: nm) due to an increase in the optical thickness of the surface layer. b (1st panel) RBD titration sensorgrams obtained at high salt concentrations revealed concentration-dependent responses. Sensorgrams are color-coded according to the log2 RBD concentration scale. Standard and Bayesian 1:1 binding models are shown as gray and black solid lines. (2nd panel) Pseudo-equilibrium response values plotted against the logarithmic RBD concentration revealed sigmoidal binding curves that were fit to the single-site interaction model yielding K_D-values in the low nM range. K_D-values and standard deviations are shown as solid and dotted lines, respectively. (3rd panel) Sensor immobilization levels are shown as jittered box plots. (4th panel) Two-dimensional distribution of dissociation rate (k_off) and affinity (K_D) values obtained from the Bayesian and standard 1:1 model fits, visualized as densities according to depicted normalized height scale and single black crosses, respectively. See related Supplementary Fig. 1a. Plots for RBD/Ty1 titrations at normal salt condition with same legends and scales. c (1st panel) ITC demonstrated high-affinity binding of Ty1 to RBD with fitted K_D and binding enthalpy (ΔH) mean values of 9 nM (with estimated bounds of 1 and 70 nM) and −10 ± 0.5 kcal/mole (mean value ± standard deviation), respectively, for two binding experiments (red). Injection of NP-VHH1 into RBD yielded heat changes at background level (blue). (2nd panel) Baseline-corrected heat changes plotted for two Ty1/RBD, a single Ty1/HBS buffer, and NP-VHH1/RBD titration experiments. Negative and positive heat changes are colored according to the red-to-blue gradient. See related Supplementary Fig. 1b: Same figure on larger scale to highlight the Ty1 into buffer dilution spikes.

Fig. 4. Ty1 binds to the RBD in ‘up’ and ‘down’ conformation and prevents ACE2 engagement.

a Cryo-EM reconstruction to an overall resolution of 2.9 Å (0.143 FSC) of the spike trimer with three bound molecules of Ty1. b Atomic model (cartoon representation) of trimeric spike in complex with three molecules of Ty1. Three chains of spike are shown in three different colors. The RBD of chain A (light green) is present in ‘up’ conformation while the other two RBDs are captured in ‘down’ conformation. The ACE2 interaction surface of RBD1 and the Ty1 interaction surface is highlighted. Magnified view of RBD2 (in ‘down’ conformation) and Ty1 interaction is shown. CDR1, 2, and 3 of Ty1 are shown in blue, green, and red, respectively. c Ty1 shows a two-pronged inhibition of ACE2 receptor binding through binding the RBD in the ‘up’ conformation and by binding to the neighboring RBD in the ‘down’ conformation. Binding of Ty1 to RBDs (both ‘up’ and ‘down’) would make ACE2 interaction surface inaccessible for ACE2.