Potential Drugs Targeting Early Innate Immune Evasion of SARS-Coronavirus 2 via 2'-O-Methylation of Viral RNA

Abstract

The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) causing the COVID-19 respiratory disease pandemic utilizes unique 2'-O-methyltransferase (2'-O-MTase) capping machinery to camouflage its RNA from innate immune recognition. The nsp16 catalytic subunit of the 2'-O-MTase is unusual in its requirement for a stimulatory subunit (nsp10) to catalyze the ribose 2'-O-methylation of the viral RNA cap. Here we provide a computational basis for drug repositioning or de novo drug development based on three differential traits of the intermolecular interactions of the SARS-CoV-2-specific nsp16/nsp10 heterodimer, namely: (1) the S-adenosyl-l-methionine-binding pocket of nsp16, (2) the unique "activating surface" between nsp16 and nsp10, and (3) the RNA-binding groove of nsp16. We employed ≈9000 U.S. Food and Drug Administration (FDA)-approved investigational and experimental drugs from the DrugBank repository for docking virtual screening. After molecular dynamics calculations of the stability of the binding modes of high-scoring nsp16/nsp10-drug complexes, we considered their pharmacological overlapping with functional modules of the virus-host interactome that is relevant to the viral lifecycle, and to the clinical features of COVID-19. Some of the predicted drugs (e.g., tegobuvir, sonidegib, siramesine, antrafenine, bemcentinib, itacitinib, or phthalocyanine) might be suitable for repurposing to pharmacologically reactivate innate immune restriction and antagonism of SARS-CoV-2 RNAs lacking 2'-O-methylation.

Keywords: COVID-19; computational screening; drug repurposing; methylation; methyltransferases; molecular docking; molecular dynamics.

Figure 1

(A) Therapeutic relevance of targeting the nsp16/nsp10 2′-O-methyltransferase (2′-O-MTase) protein complex. Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) encodes two S-adenosyl-l-methionine (SAM)-dependent methyltransferases (MTases), which sequentially methylate the RNA cap at the guanosine-7 and ribose-2′-O-positions, and are catalyzed by nsp14 N7-MTase and nsp16 2′-O-MTase, respectively. The molecular traits of the coronaviral 2′-O-MTase relative to other eukaryotic MTases make the nsp16/nsp10 protein complex an attractive target for inhibitor design. A unique feature of SARS-CoV-2 is that nsp16 obligatorily requires the non-structural protein nsp10 as a stimulatory factor for exerting its 2′-O-MTase activity by stabilizing the co-substrate SAM-binding pocket and extending the substrate RNA-binging groove. Drugs capable of inhibiting the formation of the nsp16/nsp10 complex (by targeting the nsp16/nsp10 activating interface) and the 2′-O-MTase activity (by targeting the SAM-binding site and the RNA-binding groove) can directly contribute to the suppression of genome replication. The pharmacological reduction of nsp16 2′-O-MTase activity can disrupt the 2′-O-methylation in the RNA cap structure and promote the accumulation of cap-0 structures with lower efficiency of translation, thereby suppressing the synthesis of viral proteins, while the viral RNA lacking a cap-1 structure can be recognized by innate RNA sensors and stimulate an early type I interferon response. (B) A computational approach for uncovering SARS-CoV2 2′-O-MTAse-targeting drugs. We performed a structure-based virtual screening (VS) procedure by employing molecular docking of almost 9000 U.S. Food and Drug Administration (FDA)-approved investigational and experimental (discovery-phase) drugs to predict candidates targeting one or many of the three differential traits of the intermolecular interactions of the SARS-CoV-2-specific nsp16/nsp10 heterodimer, namely: a SAM cofactor-binding site slightly different to that of main eukaryotic non-viral MTases; the interface between the stimulatory subunit nsp10 and the catalytic subunit nsp16, which is uniquely different from all other partner-independent 2′-O-MTases; and the (nsp10) allosterically regulated extension and narrowing of the substrate RNA-binding groove. First, we selected the top 20 high-scoring compounds (which therefore have a potentially higher affinity) for each site in a first filtering step; second, based on root mean square deviation (RMSD) values and the molecular mechanics Poisson–Boltzmann surface area (MM/PBSA) parameter, we selected a final list of 20 candidates that were predicted to strongly and stably interact with the abovementioned binding sites of the nsp16/nsp10 complex.

Figure 2

Predicted potential inhibitors of the nsp16/nsp10 2′-O-MTAse complex. Structures of the potential drugs were predicted to target: (a) the (nsp10-stabilized) S-adenosyl-l-methionine (SAM)-binding pocket of nsp16 (green code), (b) the unique “activating surface” between nsp16 and nsp10 (brown code), and (c) the (nsp10-extended) RNA-binding groove of nsp16 (pink code).

Figure 3

Incorporation models of the predicted SAM-binding-site-targeting drugs in SARS-CoV-2 2′-O-MTAse. The best poses of the predicted candidates coupled to the SAM-binding site of the nsp16/nsp10 complex before (0 ns) and after (100 ns) the molecular dynamics (MD) simulation are shown. The protein is represented as a function of the hydrophobicity of its surface amino acids and the Na⁺ and Cl⁻ ions have been eliminated to facilitate visualization. Each inset shows the detailed interactions of each drug candidate docked to the SAM-binding site of nsp16, indicating the participating amino acids involved in the interaction and the type of interaction (hydrogen bonds, hydrophilic interactions, salt bridges, Π-stacking, etc). The root mean square deviation (RMSD, Å) of each drug’s heavy atoms over the simulation time, measured after superposing the protein on its reference structure, is shown. In the context of the MD presented here, the RMSD incorporates traces of SAM in two independent control simulations (i.e., nsp16/nsp10 complex and nsp16 alone).

Figure 4

Incorporation models of the predicted nsp16-activating-surface-targeting drugs in SARS-CoV-2 2′-O-MTAse. The best poses of the predicted candidates coupled to the nsp16 activating surface before (0 ns) and after (100 ns) the molecular dynamics (MD) simulation are shown. The protein has been represented as a function of the hydrophobicity of its surface amino acids and the Na⁺ and Cl⁻ ions have been eliminated to facilitate visualization. Each inset shows the detailed interactions of each drug candidate docked to the nsp16 activating surface, indicating the participating amino acids involved in the interaction and the type of interaction (hydrogen bonds, hydrophilic interactions, salt bridges, Π-stacking, etc). The root mean square deviation (RMSD, Å) of each drug’s heavy atoms over the simulation time, measured after superposing the protein on its reference structure, is shown. In the context of the MD presented here, the RMSD incorporates traces of SAM alone and in two additional control simulations (i.e., nsp16/nsp10 complex and nsp16 alone).

Figure 5

Incorporation models of predicted RNA-binding-groove-targeting drugs in SARS-CoV-2 2′-O-MTAse. The best poses of predicted candidates coupled to the RNA activating surface of nsp16 before (0 ns) and after (100 ns) the molecular dynamics (MD) simulation are shown. The protein has been represented as a function of the hydrophobicity of its surface amino acids and the Na⁺ and Cl^- ions have been eliminated to facilitate visualization. Each inset shows the detailed interactions of each drug candidate docked to the RNA-binding groove of nsp16, indicating the participating amino acids involved in the interaction and the type of interaction (hydrogen bonds, hydrophilic interactions, salt bridges, Π-stacking, etc). The root mean square deviation (RMSD, Å) of each drug’s heavy atoms over the simulation time, measured after superposing the protein on its reference structure, is shown. In the context of the MD presented here, the RMSD incorporates traces of SAM alone and in two additional control simulations (i.e., nsp16/nsp10 complex and nsp16 alone).

Figure 6

Potential repurposed drug candidates against SARS-CoV2 2′-O-MTAse: a multi-layer network framework. Upon the selection of strong candidates based on the stability of high-scoring nsp16/nsp10–drug complexes (Table S2; Figure 3), we re-evaluated their pharmacological overlapping with functional modules of the virus–host interactome relevant for the viral lifecycle, as well as with clinical and laboratory features of COVID-19. The proposed multi-layer network framework might help to select the narrowest list of candidates that can be rapidly tested experimentally before evaluating their in vivo efficiency and side-effects.