Structural Comparison and Drug Screening of Spike Proteins of Ten SARS-CoV-2 Variants

Abstract

SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) has evolved many variants with stronger infectivity and immune evasion than the original strain, including Alpha, Beta, Gamma, Delta, Epsilon, Kappa, Iota, Lambda, and 21H strains. Amino acid mutations are enriched in the spike protein of SARS-CoV-2, which plays a crucial role in cell infection. However, the impact of these mutations on protein structure and function is unclear. Understanding the pathophysiology and pandemic features of these SARS-CoV-2 variants requires knowledge of the spike protein structures. Here, we obtained the spike protein structures of 10 main globally endemic SARS-CoV-2 strains using AlphaFold2. The clustering analysis based on structural similarity revealed the unique features of the mainly pandemic SARS-CoV-2 Delta variants, indicating that structural clusters can reflect the current characteristics of the epidemic more accurately than those based on the protein sequence. The analysis of the binding affinities of ACE2-RBD, antibody-NTD, and antibody-RBD complexes in the different variants revealed that the recognition of antibodies against S1 NTD and RBD was decreased in the variants, especially the Delta variant compared with the original strain, which may induce the immune evasion of SARS-CoV-2 variants. Furthermore, by virtual screening the ZINC database against a high-accuracy predicted structure of Delta spike protein and experimental validation, we identified multiple compounds that target S1 NTD and RBD, which might contribute towards the development of clinical anti-SARS-CoV-2 medicines. Our findings provided a basic foundation for future in vitro and in vivo investigations that might speed up the development of potential therapies for the SARS-CoV-2 variants.

Figure 1

Spike structure prediction of ten SARS-CoV-2 strains. The structures of full length, receptor-binding domain (RBD), and S1 N-terminal domain (NTD) of spike protein were shown, respectively. Monomer spike proteins were displayed in the cartoon model. RBD and NTD were shown in the electrostatic surface. Red and blue indicate negative and positive charges, respectively. Red arrows indicate the significantly changed sites on RBD and NTD among ten strains. Red boxes indicate the major different parts of RBD and NTD in the Delta variant compared with the original spike protein.

Figure 2

Validation of the predicted S protein structures. (a–c) The mean pLDDT values of full-length S proteins (a), S1 NTD (b), and S1 RBD (c) in different SARS-CoV-2 variants. (d) Comparison of spike protein between AlphaFold predicted and experimental structures. The experimental structures were downloaded from the protein data bank (PDB), and their accession numbers were labeled. (e) Alignment of AlphaFold prediction and experimental structure (PDB: 7DDN). The prediction and experimental structures are colored in green and red, respectively.

Figure 3

Cluster analysis of SARS-CoV-2 strains based on protein sequences and structures. (a) Four kinds of clusters of SARS-CoV-2 strains are based on protein sequences. The genome cluster is based on the genome sequences. The spike, spike RBD, and spike NTD clusters are based on their protein sequences. (b) Three kinds of clusters of SARS-CoV-2 strains are based on protein structures. Structural similarities are evaluated by RMSD related to Table S1.

Figure 4

Comparison of spike protein substrate 1 NTD between SARS-CoV-2 Delta and original strains. (a) Structural changes of Delta strain S1 NTD compared to the original strain. Black arrows and boxes indicate the structures with and without changes, respectively. (b) Mutations of amino acids on Delta S1 NTD. Triangle represents deletion of amino acid. (c, d) Comparison of five loops of S1 NTD. The RMSD values are shown in (c). The structural comparisons are shown in (d). (e) The interaction between NTD and 4A8 antibody (PDB: 7C2L). (f) The differences of three sites on loop N3 and N5 between Delta and original NTD. Yellow lines indicate the distance between corresponding amino acids.

Figure 5

Comparison of spike protein RBD between SARS-CoV-2 Delta and original strains. (a, b) Comparison of three loops of spike RBD. The RMSD values are shown in (a). The structural comparisons are shown in panel (b). Red arrows indicate the structures with significant changes. (c) Interactions between spike RBD and ACE2. Spike RBD and ACE2 are downloaded from PDB (6LZG).

Figure 6

The effects of Delta strain mutations on interactions between RBD and antibodies. (a) Comparison of two mutated amino acids on RBD. (b) The interactions between RBD and antibodies. Antibodies and original RBD are downloaded from PDB (7JX3).

Figure 7

Complexes of top 10 drugs targeting Delta S1 NTD domain. (a) Structures of top ten drugs combined with Delta NTD. (B1 and B2) Two major pharmacophores on Delta NTD. Red and blue colors on the surface indicate negative and positive charges, respectively. Yellow circles indicate the pharmacophores. Drugs used for virtual screening are downloaded from the ZINC database and filtered by “world subset.” The values are related to table S5.

Figure 8

Chemical structures of top 10 drugs targeting Delta S1 NTD. The drug interacts with the red amino acids on the Delta S1 NTD domain based on the blue interactions, including hydrophobic interaction, hydrogen bond, pi-stacking, pi-cation, and salt bridge related to Figure 7 and Table S5.

Figure 9

Complexes of top 10 drugs targeting Delta S1 RBD domain. (a) Structures of top ten drugs combined with Delta S1 RBD and RBM. The top nine drugs are commonly identified in S1 RBD and RBM. Red and blue boxes indicate the unique drugs screened in S1 RBD and RBM, respectively. (B1–3) Three major pharmacophores on Delta S1 RBD and RBM. Red and blue colors on the surface indicate negative and positive charges, respectively. Yellow circles indicate the pharmacal cavities. Drugs used for virtual screening are the same as Figure 7.

Figure 10

Chemical structures of drugs targeting Delta S1 RBD and RBM. Chemicals A1 to A9 are common drugs in S1 RBD and RBM. Chemicals A10 and A11 are unique for RBD and RBM, respectively. The drug interacts with the red amino acids on the Delta S1 NTD domain based on the blue interactions, including hydrophobic interaction, hydrogen bond, pi-stacking, pi-cation, and salt bridge related to Figure 9 and Table S5.

Figure 11

The five compounds exhibit high binding affinity to S1 RBD of the Delta variant. The estimated KD constants of dihydroergotoxine (a), trypan blue (b), irinotecan (c), biosone (d), and cepharanthine (e) are 49.9, 10.8, 243.5, 77.3, and 271.4 μM, respectively.