Scientists generate molecular models of compounds relevant for COVID-19

By Dr. Ananya Mandal, MDApr 30 2020

The discovery of suitable drugs that could help treat diseases is a long process. However, computer-aided drug design and simulation could speed up the process and also increase the possibilities say experts. This form of drug design and simulation could soon be the gold standard in drug development.

Frontera and Longhorn – leading the way

A supercomputer called the Frontera is one of the fastest there is. It has been used to predict the characteristics of novel drugs. Leading researcher Thomas Cheatham, professor of medicinal chemistry and director of the Center for High-Performance Computing at the University of Utah and Rodrigo Galindo, a professor on his team are working with Frontera. Frontera is assisted by Longhorn, an IBM/NVIDIA system at the Texas Advanced Computing Center (TACC). Longhorn’s task is to generate new molecules and compounds that could be used for the treatment of the deadly COVID-19 infection caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Medicinal chemists are using the team and the machines.

What happens during drug development?

An array of possible molecules and compounds are selected for the treatment of a disease. The researchers explain that at the molecular level, they have a potential energy field that helps them bind and interact with the atoms and organisms around them. These molecules twist and bend to alter their shape when they interact with the host or organism proteins around them. These force fields between the molecules and the proteins around the cells can be challenging to predict and may often influence the final utility of the drug molecule.

Amber

Amber (Assisted Model Building with Energy Refinement) is one of the significant tools that experts use in simulating the force fields that influence a drug molecule within its cellular environment. Cheatham is one of the primary researchers on the team developing Amber. Amber has been evolving to its present form since 1978. It has undergone a sea change from what it was at the start. At present, it is relatively accurate in predicting the cellular environment force fields that the potential drug molecule would encounter in real life.

Experts say that Amber is capable of matching experimental results with an accuracy of less than half an angstrom (Å). An angstrom is one hundred millionth of a centimeter or 10^-10 meters. Cheatham and Galindo have used Amber on the molecules that are candidate drugs and simulated the biomolecular environments for the candidates so as to see their applicability in medicine. They explained, “The goal is to understand the structure, function, dynamics, and energetics of biomolecular systems in their native environment, with water and other ligands.”

Representation of the coronavirus main protease with a peptide inhibitor. [Credit: Cheatham Lab]

Progress in drug development

At the start of their work, the duo was working on two lead molecules that were principally copper-containing compounds and were being tried to fight cancer. These molecules were experimentally modified using computer-aided design and simulations to see if their alterations in the DNA could allow them to remain protected from degradation within the body. During that time, the world was hit by the COVID-19 pandemic. The NSF RAPID was supporting the initial work that started in 2015.

Cheatham and Galindo now started working on potential molecules that could kill the novel coronavirus that was fast reaching almost all corners of the world, infecting millions and killing thousands. During the initial phase, the duo was working on funded research to bring out a possible drug candidate that could treat Ebola virus infection. The team was using crystal structure studies using the Rosetta software suite to select the best possible candidate with optimum amino acid side chains stuck to the basic peptide backbone template. Once the molecule was found, they used Amber to simulate the biomolecular environments and optimize the structures of the candidate molecules.

Fight against COVID-19 and hope for the future

Galindo explained that they had in hand over 2,000 molecular models that could be used against the COVID-19 infection using Longhorn and Frontera supercomputers at the TACC. They then applied for 2.7 million node hours on Blue Waters. This is a GPU-based system found in the National Center for Supercomputing Applications (NCSA). They were granted their requirements through the COVID-19 HPC Consortium. This consortium is a public-private one that is working to match researchers with the resources so as to accelerate research against the coronavirus.

For this effort, they have identified a COVID-19 main protease crystal structure. The protease is an enzyme that can break down proteins and peptides. This structure is in complex with a peptide inhibitor N3. The team explains that they would be working with the Ebola peptide design to see if their COVID-19 protease holds up.

Once their molecule is ready and selected, they would be made into circular modified peptides at the Schmidt lab in the Medical Chemistry department at the University of Utah.

Cheatham said, “Our hope is that we find a new peptide inhibitor that can be experimentally verified in the next couple of weeks. And then, we will engage in further design to make the peptide cyclic to make it more stable as a potential drug. The hope is we can, in the next few months, find and experimentally verify a better peptide inhibitor for the COVID main protease.”