In silico design and validation of high-affinity RNA aptamers for SARS-CoV-2 comparable to neutralizing antibodies
- Institute of Quantitative Biology, College of Life Sciences, Zhejiang University, China
- State Key Laboratory of Plant Environmental Resilience, College of Life Sciences, Zhejiang University, China
- The First Affiliated Hospital, College of Medicine, Zhejiang University, China
Figures
Figure 1 with 1 supplement
The workflow of the CAAMO framework for designing high-affinity RNA aptamers and its application in the development of novel RNA aptamers targeting the receptor binding domain (RBD) of the SARS-CoV-2 spike protein.
(A) Schematic diagram of the spike protein (trimer) of SARS-CoV-2 virus with one of the three RBD regions highlighted in gold. (B) The sequence and Mfold-predicted secondary structure of a SELEX-derived RNA aptamer termed Ta targeting the RBD of SARS-CoV-2 spike protein, which was the initial input information of the CAAMO framework. (C) Schematic diagrams show that high-affinity RNA aptamers optimized through the CAAMO framework can complement existing antibody-neutralizing treatments for COVID-19. Antibodies (purple) and RNA aptamers (yellow) that bind to the RBDs of the SARS-CoV-2 spike protein can neutralize viral infection by blocking its interaction with the human ACE2. The designed aptamers (marked by red dots) with binding affinities comparable to antibodies can further strengthen the neutralizing treatment when antibody escape occurs. (D) Illustrative workflow of the CAAMO framework by integrating computational techniques with experimental validation. (E) A representative output after CAAMO framework optimization (the aptamer TaG34C) and its comparison with the original Ta sequence. RNA aptamer is colored pink and RBD is shown in dark blue surface. The mutated nucleotides (G34 in wild type Ta and C34 in optimized TaG34C) are highlighted in pink sticks; atoms oxygen and nitrogen are colored red and blue, respectively. (F) Competitive binding experiments to compare the binding capabilities of RNA aptamers Ta or TaG34C and a commercial monoclonal SARS-CoV-2 neutralizing antibody (SinoBiological, Cat: 40592-R001) to the RBD of SARS-CoV-2 spike protein. Gradient amount of the commercial neutralizing antibody (0–1.67 μM) was titrated into the buffer containing 0.5 μM aptamer (Ta or TaG34C) and 40 μM RBD. The intensity of aptamer-RBD bands was quantified with Image J and normalized to that of the mixture without antibody, which was set to 100%. Data were collected from the images of electrophoretic mobility shift assay (EMSA) shown in Figure 5D (n=1).
Figure 1—figure supplement 1
The workflow of CAAMO.
The workflow is to design high-affinity aptamers with physics-based simulations. It consists of four phases, including (i) constructing an aptamer conformational ensemble, (ii) identifying a proper aptamer binding mode, (iii) free energy perturbation (FEP)-based rational design, and (iv) experimental validation.
Figure 2 with 5 supplements
Determination of the binding model of the aptamer Ta and the receptor binding domain (RBD) of the SARS-CoV-2 spike protein via a multi-strategy approach.
(A) Flowchart illustrating the combination of RNA 3D structure prediction, ensemble docking and clustering, and binding capacity assessment by molecular mechanics generalized Born surface area (MM/GBSA) and steered molecular dynamics (SMD) to determine the most probable binding conformation of the aptamer Ta to the RBD. (B) Six main binding modes, that is conformations 01–06, with clearly distinct binding conformations after ensemble docking and clustering for further binding ability assessment. The RBD and the aptamer Ta are shown in dark blue and dark red, respectively. For clarity, the surface of RBD is also displayed. For the aptamer Ta, conformations 01–06 are named in a descending order of their respective cluster sizes and their colors are lightened gradually. (C) The binding energies between the aptamer Ta and RBD estimated by MM/GBSA method for six binding candidates shown in plane (B). Data represent mean ± SD collected from three independent calculations (mean ± SD, n=3). (D) The rupture works required to separate the bound aptamer Ta from the RBD for different binding conformation candidates. Data were collected from four independent steered MD simulations. Limited by the available computing resources, only the first four binding conformations were assessed. The orange line is the median, boxes extended from lower to upper quartiles, whiskers showing the range of nonoutlier data (n=4). (E) Overview of the most probable binding model (conformation 01) of the aptamer Ta and RBD. The RBD is shown in cyan cartoon representation, the apical loop, bulge part, and end stem of the aptamer Ta are colored plum, pink, and sandy brown, respectively. The key residues of the RBD and nucleotides of the aptamer Ta for binding interactions are shown in sticks, all nitrogen atoms colored blue, and all oxygen atoms red. (F) Electrophoretic mobility shift assay (EMSA) results of the aptamer Ta (upper panel) and the weakly binding aptamer Tc (negative control, lower panel) bound to the RBD of the SARS-CoV-2 spike protein and (G) the resultant binding curve for the aptamer Ta. The dissociation constant (Kd) was calculated from the EMSA image quantification from three independent experiments (mean ± SD, n=3).
-
Figure 2—source data 1
PDF showing the original uncropped electrophoretic mobility shift assay (EMSA) gel for Figure 2F, with the aptamer-receptor binding domain (RBD) complex and free aptamer bands labeled.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig2-data1-v1.pdf
-
Figure 2—source data 2
Original image files for the electrophoretic mobility shift assay (EMSA) gel displayed in Figure 2F.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig2-data2-v1.zip
Figure 2—figure supplement 1
Simulation results for receptor binding domain (RBD) of the SARS-CoV-2 spike protein.
(A) Heavy-atom RMSDs as functions of simulation time for receptor binding domain (RBD) protein. Three 500-ns independent runs (Run1, Run2, and Run3) are performed (n=3). (B) Superposition of the molecular dynamics (MD) refined structure (blue) and crystallographically resolved structure (Protein Data Bank [PDB] id: 6LZG, gray).
Figure 2—figure supplement 2
3D conformations of Ta and Ta-RBD (receptor binding domain) complexes.
(A) 25 Ta 3D structures were predicted by five RNA 3D modeling tools (IsRNA, FARFAR2, SimRNA, iFoldRNA, and RNAcomposer). (B) The Ta-RBD complexes were predicted by four docking tools (HADDOCK, HDOCK, ZDOCK, and RosettaDock). (C) Representative complex conformations from the top six clusters are shown. The number of conformations contained within each cluster is indicated in parentheses.
Figure 2—figure supplement 3
RMSDs of heavy atoms are plotted as functions of simulation time across six selected conformations (conformation 01–06) of Ta aptamer and receptor binding domain (RBD) protein.
Three independent 500-ns runs (Run1, Run2, and Run3) are performed (n=3). The initial structures are shown in white, while the molecular dynamics (MD) refined structures are shown in red and blue for the aptamer and protein, respectively.
Figure 2—figure supplement 4
The binding interfaces between the Ta aptamer and receptor binding domain (RBD).
(A) Details of interactions between RBD and nucleotides C24, G25, C29, U30, and U31 in the apical loop part are presented. (B) Details of interactions between RBD and nucleotides A14 in the bulge loop. (C) Details of interactions between RBD and nucleotides C3, A42, and G43 in the end stem part. Hydrogen bonds between RNA nucleotides and RBD residues are indicated by thin black lines. Electrostatic interactions between RNA nucleotides and RBD residues are indicated by thin yellow lines.
Figure 2—figure supplement 5
The modeling of the negative control Tc-RBD (receptor binding domain) complex.
(A) The Mfold-predicted secondary structure of Tc. (B) The binding energies (ΔG) of four representative Tc-RBD conformations calculated by the molecular mechanics generalized Born surface area (MM/GBSA) method (mean ± SD, n=3). (C) The putative binding mode of the Tc aptamer and RBD complex. (D) The works required to pull the aptamer lead Ta and the negative control Tc away from RBD, respectively. Ta-1 through Ta-4 and Tc-1 through Tc-4 represent four parallel replicates for the Ta and Tc simulations, respectively (n=4).
Figure 3 with 1 supplement
The aptamer Ta exhibits a comparable binding capability to the receptor binding domain (RBD) compared to neutralizing antibodies.
(A) Contact ratios of residues on the RBD by ACE2 (derived from molecular dynamics [MD] simulations), the aptamer Ta (derived from MD simulations), and neutralizing antibodies (derived from all available SARS-CoV-2 RBD–antibody complex structures curated in CoV-AbDab). For reference, the electrostatic potential distribution on the RBD surface generated by the PyMOL (version 2.3.5) program was also shown. (B) Key residues on RBD (with contact ratio larger than 0.5 in panel [A]) contacted by ACE2, the aptamer Ta, and neutralizing antibodies were displayed as a Venn diagram. (C) Binding energies estimated by MM/GBSA calculations for ACE2, the aptamer Ta, and three representative antibodies (P2C-1F11, 2H2, and S2E12) binding to RBD. Data represent mean ± SD collected from three independent calculations (mean ± SD, n=3). (D) Binding ability of the commercial antibody (40592-R001) to RBD was assessed by native-PAGE. The reaction mixture (10 μL) contained 8.57 μM RBD protein and 2 μM antibody, incubated individually or combined, followed by Coomassie brilliant blue staining. (E) The RBD binding abilities of the aptamer Ta and commercial antibody 40592-R001 were compared by EMSA competitive binding experiments. The aptamer-RBD complex bands were shown after running on an agarose gel following the incubation of 40 μM RBD protein, 0.5 μM aptamer Ta, and 0.5 μM antibody 40592-R001 (final concentrations in the reaction mixture).
-
Figure 3—source data 1
PDF showing the original uncropped gels for Figure 3D (native-PAGE) and Figure 3E (electrophoretic mobility shift assay [EMSA]), with the relevant bands labeled.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig3-data1-v1.pdf
-
Figure 3—source data 2
Original image files for the gels displayed in Figure 3D and E.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig3-data2-v1.zip
Figure 3—figure supplement 1
Comparative analysis of receptor binding domain (RBD) binding from ACE2, Ta, and antibodies.
Molecular dynamics (MD) simulations are performed on the ACE2-RBD complex and the contact ratios of RBD residues are obtained (upper panel). Subsequently, the contact ratios are determined for RBD residues in the Ta-RBD complexes (middle panel). Additionally, the contact ratios are accessed for RBD residues among all antibodies that bind to the ACE2-RBD interface, as recorded in the RCSB database (lower panel).
Figure 4 with 2 supplements
Structure-based rational design of Ta analogues with improved binding affinities and their experimental validation.
(A) Flowchart illustrating the combination of rational mutation scanning, secondary structure analysis (SSA), and free energy perturbation (FEP) to optimize aptamer binding affinities. (B) Contact ratios of nucleotides on the aptamer Ta bound to receptor binding domain (RBD; dark blue surface). Data were collected from three independent molecular dynamics (MD) simulations. (C) SSA based on base-pair (BP) similarity for 16 selected nucleotides mutated to other three bases. Definition of BP similarity was given in the main text. Only mutations with BP similarity greater than 0.9 were subjected to further FEP calculations. (D) The binding free energy changes assessed by FEP calculations for selected single mutations. Data represent mean ± SD collected from five independent simulations (mean ± SD, n=5). (E) EMSA results of Ta, Tc, and six designed candidate sequences bound to the RBD of SARS-CoV-2 spike protein. The aptamer-RBD complex bands were detected by running an agarose gel after incubation of 40 μM of RBD protein with 0.5 μM indicated aptamer variant. (F) Comparison of the binding free energy changes derived from FEP calculations (ΔΔG, left scale) and electrophoretic mobility shift assay (EMSA) experiments (ΔΔGexp, right scale) for six designed candidate aptamers. Data represent mean ± SD collected from five (FEP)/three (EMSA) independent replications (mean ± SD, n=5 for FEP and n=3 for EMSA).
-
Figure 4—source data 1
PDF showing the original uncropped electrophoretic mobility shift assay (EMSA) gel for Figure 4E, with the relevant bands labeled.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig4-data1-v1.pdf
-
Figure 4—source data 2
Original image files for the electrophoretic mobility shift assay (EMSA) gel displayed in Figure 4E.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig4-data2-v1.zip
Figure 4—figure supplement 1
The secondary structures of Ta, TaG34C, and TaU30G, predicted by Mfold.
The mutated bases are highlighted in orange and marked with red pentagrams.
Figure 4—figure supplement 2
Electrophoretic mobility shift assay (EMSA) results of the designed Ta analogues.
(A) Electrophoretic mobility shift assay (EMSA) results of six Ta mutations (TaG34C, TaG34U, TaG34A, TaC23G, TaC23A, and TaC23U). (B) Binding curves and Kd values for Ta, TaG34C, TaG34U, TaG34A, TaC23G, TaC23A, and TaC23U. The Kd was calculated from the EMSA image quantification from three independent experiments (mean SD, n=3).
-
Figure 4—figure supplement 2—source data 1
PDF showing the original uncropped electrophoretic mobility shift assay (EMSA) gels for the six Ta mutants, with the relevant bands labeled.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig4-figsupp2-data1-v1.pdf
-
Figure 4—figure supplement 2—source data 2
Original image files for the EMSA gels.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig4-figsupp2-data2-v1.zip
Figure 5 with 2 supplements
Designed aptamer TaG34C showed superior binding ability to WT Ta or antibody.
(A) Binding curves and the resultant dissociation constants (Kd) for WT Ta and the designed TaG34C with receptor binding domain (RBD) protein. The Kd values were calculated from the electrophoretic mobility shift assay (EMSA) image quantification with SD from three independent experiments (mean ± SD, n=3). (B) Comparison of the Ta-RBD and TaG34C-RBD binding complexes. RNA is shown as a dark red ribbon while RBD is displayed as a dark blue surface. Zoom-in is the nucleotide G34 (WT Ta)/C34 (TaG34C) and its surrounding nucleotide U35 and residue PHE486 shown in sticks. Atoms oxygen and nitrogen are colored in red and blue, respectively, and carbon atoms are colored by their locations. Definitions of two related distances (D1 and D2) are also displayed. (C) Violin plots of distributions of two related distances (D1 and D2) between selected RNA nucleotide-protein residue in Ta-RBD and TaG34C-RBD binding complexes, respectively. Data were collected from three independent molecular dynamics (MD) simulations (n=3). (D) EMSA images of competitive binding experiments to characterize the RBD binding abilities of RNA aptamers (WT Ta and TaG34C) and the commercial monoclonal SARS-CoV-2 neutralizing antibody 40592-R001. The aptamer-RBD complex bands were showed by running an agarose gel after incubation of 40 μM of RBD protein and 0.5 μM indicated aptamer with varying concentrations of the antibody 40592-R001. Final antibody concentrations ranged from 0 to 1.67 μM in the reaction mixtures. Results showed that TaG34C, but not WT Ta, exhibited a higher binding affinity to the RBD protein than that of the antibody.
-
Figure 5—source data 1
PDF showing the original uncropped electrophoretic mobility shift assay (EMSA) gel for Figure 5D, with the relevant bands labeled.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig5-data1-v1.pdf
-
Figure 5—source data 2
Original image files for the electrophoretic mobility shift assay (EMSA) gel displayed in Figure 5D.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig5-data2-v1.zip
Figure 5—figure supplement 1
Electrophoretic mobility shift assay (EMSA) assays with BSA as a non-target protein control verify the target-specific binding of designed aptamers to receptor binding domain (RBD).
EMSA images of Ta, Tc, and TaG34C incubated with SARS-CoV-2 spike RBD or BSA protein. Aptamer-protein complex bands were visualized by agarose gel electrophoresis after incubation of 0.5 µM of the indicated aptamer with 40 µM RBD or BSA. Only weak, comparable background signals were observed with BSA for all three aptamers, whereas markedly stronger binding was detected between RBD and Ta or TaG34C, while no detectable binding was observed between Tc and RBD, confirming that the aptamer-RBD interactions are target-specific (n=2).
-
Figure 5—figure supplement 1—source data 1
PDF showing the original uncropped electrophoretic mobility shift assay (EMSA) gels with BSA as a non-target control, with the relevant bands labeled.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig5-figsupp1-data1-v1.pdf
-
Figure 5—figure supplement 1—source data 2
Original image files for the electrophoretic mobility shift assay (EMSA) gels.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig5-figsupp1-data2-v1.zip
Figure 5—figure supplement 2
SDS-PAGE analysis of the SARS-CoV-2 spike receptor binding domain (RBD) protein, neutralizing antibody (40592-R001), and BSA reference.
This gel validates the high purity and structural integrity of the commercially sourced RBD protein and neutralizing antibody used in this study.
-
Figure 5—figure supplement 2—source data 1
PDF showing the original uncropped SDS-PAGE gel, with the relevant lanes/bands labeled.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig5-figsupp2-data1-v1.pdf
-
Figure 5—figure supplement 2—source data 2
Original image files for the SDS-PAGE gel.
- https://cdn.elifesciences.org/articles/107785/elife-107785-fig5-figsupp2-data2-v1.zip
Additional files
-
Supplementary file 1
Supplementary tables for this study.
(A) The relative binding free energy changes for single nucleotide mutations on Ta binding (mean ± standard deviation, kcal/mol, from five independent FEP runs). (B) The aptamer sequences employed in this study and their binding energies with RBD. (C) Relative binding free energy changes (ΔΔG, kcal/mol) for Ta binding to SARS-CoV-2 RBD variants (Alpha and Beta) calculated by FEP/HREX. (D) Relative binding free energy changes (ΔΔG, kcal/mol) for TaG34C binding to SARS-CoV-2 RBD variants calculated by FEP/HREX. (E) Summary of simulation systems in this study.
- https://cdn.elifesciences.org/articles/107785/elife-107785-supp1-v1.docx
-
MDAR checklist
- https://cdn.elifesciences.org/articles/107785/elife-107785-mdarchecklist1-v1.docx
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
In silico design and validation of high-affinity RNA aptamers for SARS-CoV-2 comparable to neutralizing antibodies
eLife 14:RP107785.
https://doi.org/10.7554/eLife.107785.4
