Selection, biophysical and structural analysis of synthetic nanobodies that effectively neutralize SARS-CoV-2

Abstract

The coronavirus SARS-CoV-2 is the cause of the ongoing COVID-19 pandemic. Therapeutic neutralizing antibodies constitute a key short-to-medium term approach to tackle COVID-19. However, traditional antibody production is hampered by long development times and costly production. Here, we report the rapid isolation and characterization of nanobodies from a synthetic library, known as sybodies (Sb), that target the receptor-binding domain (RBD) of the SARS-CoV-2 spike protein. Several binders with low nanomolar affinities and efficient neutralization activity were identified of which Sb23 displayed high affinity and neutralized pseudovirus with an IC₅₀ of 0.6 µg/ml. A cryo-EM structure of the spike bound to Sb23 showed that Sb23 binds competitively in the ACE2 binding site. Furthermore, the cryo-EM reconstruction revealed an unusual conformation of the spike where two RBDs are in the 'up' ACE2-binding conformation. The combined approach represents an alternative, fast workflow to select binders with neutralizing activity against newly emerging viruses.

Fig. 1. Selection of sybodies against SARS-CoV-2 RBD.

a Sybody selection statistics for the three different libraries. b Radial phylogenetic tree of all unique binders identified in this study. Sequence alignment was performed using PROMAL3D and the phylogenetic tree was construct using the maximum likelihood (ML) analysis in MEGA. Two sybodies (76 and 88), expected to stem from the convex library, were found to belong to the loop library, presumably due to a spill-over during the selection procedure.

Fig. 2. Biophysical characterization of identified sybodies.

BLI sensorgrams of immobilized SARS-CoV-2 RBD with 2-fold serial dilution of 75 nM Sb23 (a), 188 nM Sb12 (b), 50 nM Sb42 (c), 250 nM Sb76 (d), 250 nM Sb95 (e), and 188 nM Sb100 (f). Binding curves are colored black and the global fit of the data to a 1:1 binding model is red. Resulting affinities are indicated. g Thermal unfolding data of isolated RBD and in complex with Sb23. h Resulting melting temperatures of RBD alone and in complex with Sb23, Sb76, Sb95, and a control sybody (NC), selected against the human peptide transporter (hPepT2). Data represent the mean ± SD of three replicate experiments.

Fig. 3. Sb23 neutralizes SARS-CoV-2 pseudoviruses and competes with ACE2.

a SARS-CoV-2 or VSV-G spike pseudotyped lentivirus was incubated with a dilution series of Sb23, Sb23-Fc, or a control sybody (specific for hPepT2). Neutralization by Sb23 is a representative of two independent experiments. Data are mean ± SD of six replicate experiments. Neutralization by Sb23-Fc represents two independent assays, performed in duplicates. Data are mean ± SD of two or four replicate experiments. b BLI sensorgrams of immobilized SARS-CoV-2 RBD with 2-fold serial dilution of 20 nM Sb23-Fc. Binding curves are colored black and the global fit of the data to a 1:1 binding model is red. c BLI sensorgrams of immobilized SARS-CoV-2 RBD with ACE2 in the presence (blue) or absence (black) of 150 nM Sb23. The assay was performed in a concentration range of 200-12.5 nM ACE2 and fit of the data to a 1:1 binding model is shown in red. d Microscale thermophoresis (MST) binding data of spike with fluorescently labeled Sb23, in the presence or absence of 200 nM ACE2. One representative measurement is shown. Three independent measurements were performed and affinities of spike to Sb23 in the absence of ACE2 ranged from 0.6 to 10 nM, while they were significantly lower in the presence of ACE2 (K_D = 58–200 nM).

Fig. 4. SAXS analysis of the Sb23–RBD complexes.

a Experimental SAXS data from Sb23, RBD, and its complex. b Distance distribution functions of Sb23, RBD, and their complex. c Two-phase MONSA shape of Sb23 (red beads) and RBD (blue beads). d Hybrid model of Sb23 in complex with RBD.

Fig. 5. Cryo-EM reconstruction of SARS-CoV-2 spike bound to Sb23.

a Locally sharpened Coulomb potential map and cartoon model of Sb23 bound to the spike protein in the “1-up” conformation and cartoon model of Sb23-bound spike. b Locally sharpened Coulomb potential map and cartoon model of Sb23 bound to the spike protein in the “2-up” conformation.

Fig. 6. Top view of cryo-EM reconstruction of SARS-CoV-2 spike bound to Sb23 and modeling of the structural basis for Sb23-based blockage of SARS-CoV-2 spike binding to ACE2.

a Top view of locally sharpened Coulomb potential map and cartoon model of Sb23 bound to the spike protein in the “1-up” conformation b Top view of locally sharpened Coulomb potential map and cartoon model of Sb23 bound to the spike protein in the “2-up” conformation. c Cartoon model of Sb23-bound Spike in the “1-up” (left) “2-up” (right) conformation showing how ACE2 binding is blocked by Sb23 bound to the RBD in the “up” conformation as well as Sb23 bound to the neighboring RBD in the down conformation.