The new England Journal of Medicine published a recent article How to Discover Antiviral Drugs Quickly by Jerry M. Parks, Ph.D., and Jeremy C. Smith, Ph.D.
We urgently need effective drugs for coronavirus disease 2019 (Covid-19), but what is the quickest way to find them? One approach that sometimes seems akin to a “Hail Mary” pass in American football is to hope that drugs that have worked against a different virus (such as hepatitis C or Ebola) will also work against Covid-19. Alterna-tively, we can be rational and specifically target proteins of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) so as to interrupt its life cycle.The SARS-CoV-2 genome encodes approxi-mately 25 proteins that are needed by the virus to infect humans and to replicate (Fig. 1). Among these are the notorious spike (S) protein, which recognizes human angiotensin-converting en-zyme 2 in the initial stage of infection; two proteases, which cleave viral and human pro-teins; the RNA polymerase, which synthesizes viral RNA; and the RNA-cleaving endoribonucle-ase. Finding drugs that can bind to the viral proteins and stop them from working is a logi-cal way forward and the priority of many re-search laboratories.One approach toward this goal involves mim-icking nature with the use of computational structure-based drug discovery (Fig. 2). In this process, computers “dock” trial compounds into binding sites in three-dimensional models of the protein targets. The binding affinities of the com-pounds are calculated with the use of physics-based equations that quantify the interactions between the drug and its target. The top-ranked compounds are then tested experimentally to see if they do indeed bind and have the required downstream effects (such as stopping viral in-fectivity) on cells and in animal models.Structure-based drug discovery has been im-portant in finding antiviral drugs, an example being nelfinavir, discovered in the 1990s, to treat human immunodeficiency virus (HIV) infection. Unfortunately, though, at that time the process was relatively inefficient: calculations were inac-curate and computers so feeble that only about 100 compounds could be docked at a time. Moreover, both the target and the drug had to be held rigid in the docking process in a lock-and-key approach. Rigid docking does not often take place in real life, because proteins undergo thermally driven internal motions that lead to fluctuating binding-site shapes.Since the 1990s, the power of supercomput-ers has increased by a factor of a million or so. Rigid docking of over a billion compounds can now be performed in a few days. Thus, virtual high-throughput screening is outperforming equivalent experimental high-throughput screen-ing and can rapidly identify very tightly binding compounds.1 Furthermore, molecular-dynamics simulations can be performed to calculate inter-nal protein motions, and candidate drugs can be screened through a process that uses the differ-ent shapes formed by the binding site in a pro-cedure known as “ensemble docking.”2 This ap-proach is more realistic than rigid docking and has been successful, for example, in serving the HIV drug-discovery efforts from the 2000s on-ward. In our own laboratory, ensemble docking has produced experimentally validated hits against each of the 16 protein targets presented to us over the past few years.Modern supercomputers such as the Summit supercomputer at Oak Ridge National Laboratory, which is currently the world’s most powerful, perform massively parallel processing in which many calculations are performed at the same time. This enables molecular-dynamics simula-tions of many replicas of the target to be run in parallel, each exploring a slightly different con-formational space. Thus, a comprehensive simu-lation model of a SARS-CoV-2 protein drug tar-get can be obtained with the use of Summit in a day, whereas it would take months with the use of a typical computer cluster. Supercomputers are also used in rapid parallel docking of large databases of compounds. The structure-based drug-discovery field is thus primed for quick results.
Figure 1. The SARS-CoV-2 Virion and Its Proteins. Although all the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) proteins are potential drug targets, some are likely to be more easy to find a drug against, in part because they play principal roles in the viral lifecycle and also lack human protein homologues. Examples include the spike glycoprotein, the papain-like protease, the chymotrypsin-like main protease, and the RNA-dependent RNA polymerase. A list of the Worldwide Protein Data Bank identifiers of the structures shown is provided in the Supplementary Appendix, available with the full text of this article at NEJM.org. ACE2 denotes angiotensin-converting enzyme 2, NSP nonstructural protein, ORF open reading frame, and RdRP RNA-dependent RNA polymerase.
Ref and Source of Information
Download full article :