High affinity nanobodies block SARS-CoV-2 spike receptor binding domain interaction with human angiotensin converting enzyme

Abstract

There are currently few approved effective treatments for SARS-CoV-2, the virus responsible for the COVID-19 pandemic. Nanobodies are 12-15 kDa single-domain antibody fragments that can be delivered by inhalation and are amenable to relatively inexpensive large scale production compared to other biologicals. We have isolated nanobodies that bind to the SARS-CoV-2 spike protein receptor binding domain and block spike protein interaction with the angiotensin converting enzyme 2 (ACE2) with 1-5 nM affinity. The lead nanobody candidate, NIH-CoVnb-112, blocks SARS-CoV-2 spike pseudotyped lentivirus infection of HEK293 cells expressing human ACE2 with an EC₅₀ of 0.3 µg/mL. NIH-CoVnb-112 retains structural integrity and potency after nebulization. Furthermore, NIH-CoVnb-112 blocks interaction between ACE2 and several high affinity variant forms of the spike protein. These nanobodies and their derivatives have therapeutic, preventative, and diagnostic potential.

Figure 1

Overview of the therapeutic potential and isolation of nanobodies targeting the SARS-CoV-2 RBD:ACE2 interaction. (a) Illustration of the structure of SARS-CoV-2 spike protein, with receptor binding domain in contact with the human ACE2 receptor on the surface of a lung epithelial cell. A major therapeutic goal is to develop inhibitory agents that disrupt the interaction between the spike protein and the human ACE2 receptor. (b) Isolation of nanobodies binding SARS-CoV-2 spike protein. An adult llama was immunized 5 times over 28 days with purified, recombinant SARS-CoV-2 spike protein. On day 35 after first immunization (7 days after last immunization), llama blood was obtained through a central line, B-cells were isolated, the single heavy-chain variable domains (nanobodies) of the llama antibodies were amplified and cloned to construct a recombinant DNA library containing more than 10⁸ clones. The library of clones was expressed in a phage display format, in which each phage expresses between 1–5 nanobody copies on its surface and contains the DNA sequence encoding that nanobody. Immuno-panning was performed to isolate candidate nanobodies for expression and validation studies. (c) Selection strategy for isolation of nanobody candidates which bind to the RBD:ACE2 interaction surface. Using a phage display library from an adult llama immunized with full length S1 spike protein, nanobodies were isolated which block the interaction between RBD and ACE2. (c_i) In a standard radioimmunoassay tube, ACE2 was immobilized and the surface blocked with non-specific protein. Biotinylated-RBD, was incubated with the nanobody phage library and then added to the immuno-tube and allowed to interact with the immobilized ACE2. Biotinylated-RBD with no associated nanobodies, or with nanobody associations which do not block the ACE2 binding domain, bound the immobilized ACE2. (c_ii) Biotinylated-RBD with associated nanobodies that blocked the ACE2 binding domain remained in solution and were recovered using streptavidin-coated magnetic particles that bind to the biotin. (c_iii) Nanobodies which did not bind to RBD were removed during washing of the magnetic beads. This method allowed for specific enrichment of nanobodies which both bind to the RBD and compete for the RBD-ACE2 binding surface. (Figure generated using BioRender.com).

Figure 2

Protein sequences for novel nanobodies that bind to the SARS-CoV-2 spike protein receptor binding domain. Single letter amino acid codes. Clustal Omega algorithm used for alignment. Blue highlights indicate sequence diversity with NIH-CoVnb-112, highlighted in gray, set as the reference sequence for comparison. For comparison, seven previously reported nanobody sequences have clearly distinct sequences: Ty1, VHH72, H11-D4, MR3, Sb#14, Sb23, and W25UACh and possess shorter CDR3 domains (represented in NIH-CoVnb-112 by amino acids 99–120).

Figure 3

Affinity binding curves of isolated anti-SARS-CoV-2 RBD nanobodies. Using Biolayer Interferometry on a BioForte Octet Red96 system, association and dissociation rates were determined by immobilizing biotinylated-RBD onto streptavidin coated optical sensors (a–e). The RBD-bound sensors were incubated with specific concentrations of purified candidate nanobodies for a set time interval to allow association. The sensors were then moved to nanobody-free solution and allowed to dissociate over a time interval. Curve fitting using a 1:1 interaction model allows for the affinity constant (K_D) to be measured for each nanobody as detailed in (f).

Figure 4

Competitive inhibition of ACE2 and RBD binding using anti-SARS-CoV-2 RBD nanobodies. (a) RBD coated ELISA plates were blocked with non-specific protein and incubated with serial dilutions of each candidate anti-SARS-CoV-2 RBD nanobody. Biotinylated-ACE2 was added to each well and allowed to bind to unoccupied RBD. The ELISA was then developed using a standard streptavidin-HRP and tetramethylbenzidine reaction. Unoccupied RBD allows for a positive reaction signal which is suppressed in the presence of bound competitive nanobody. NIH-CoVnb-112 produces the greatest inhibition of ACE2 binding with an EC₅₀ of 0.02 µg/mL (1.11 nM). (b) Comparable findings using the commercially available Genscript SARS-CoV-2 neutralization surrogate assay.

Figure 5

Interaction of NIH-CoVnb-112 with SARS-Cov-2 Spike Protein RBD variants. (a) Binding of NIH-CoVnb-112 to RBD “wild type” and 3 variant forms of the RBD had similar affinity, with half maximal binding at approximately 0.01 µg/mL (b) Competitive inhibition assay: RBD “wild type” and 3 variant forms of the RBD coated ELISA plates were blocked with non-specific protein and incubated with dilutions of the lead candidate anti-SARS-CoV-2 RBD nanobody NIH-CoVnb-112. Biotinylated-ACE2 was added to each well and allowed to bind to unoccupied RBD. The ELISA was then developed using a standard streptavidin-HRP and tetramethylbenzidine reaction. Unoccupied RBD allows for a positive reaction signal which is suppressed in the presence of bound competitive nanobody. NIH-CoVnb-112 inhibited ACE2 binding to each of the variants with a similar EC₅₀ of 0.02 µg/mL (1.11 nM).

Figure 6

NIH-CoVnb-112 stability and potent inhibition of SARS-CoV-2 pseudovirus following nebulization. (a) An Aerogen Solo High-Performance Vibrating Mesh nebulizer was placed in line with a custom glass bead condenser to allow for collection of the nanobody following nebulization. (Image element courtesy of Aerogen.) (b) A pre-nebulization and post-nebulization sample of purified NIH-CoVnb-112 was analyzed by SDS-PAGE gel. The dominant band for each sample remains at approximately 14 kDa indicating no detectable degradation or aggregation of NIH-CoVnb-112 following nebulization. (c) Size exclusion chromatography demonstrated a prominent peak eluting at 13.5 mL elution volume in both pre and post-nebulization samples. (d) A fluorescence reporter assay utilizing SARS-CoV-2 spike protein pseudotyped lentivirus was used to demonstrate potent inhibition of the RBD:ACE2 interaction. HEK293 cells overexpressing human ACE2 were cultured for 24 h with pseudotyped virus which was pretreated with NIH-CoVnb-112 at different concentrations. Inhibition of the spike RBD occurs when the virus is not able to transduce the HEK293-ACE2 cells and subsequently produce RFP reporter protein. e Following 48 hr incubation, HEK293-ACE2 cells were analyzed by flow cytometry to quantify the fluorescence level. NIH-CoVnb-112 potently inhibited viral transduction both pre and post-nebulization with an EC₅₀ of 0.323 µg/mL (23.1 nM) and 0.116 µg/mL (8.3 nM) respectively. (Figure elements generated using BioRender.com).