Understanding protein interactions to develop COVID-19 treatments
Dr. Alex Dickson is an assistant professor at Michigan State University in the Department of Biochemistry and Molecular Biology. He along with his group of researchers are using computer simulations to gain greater insights into the structure, dynamics and interactions of biological molecules. One interaction of interest is between the ACE2 enzyme in human cells and the S1 subunit of the spike protein in COVID-19. The spike protein is on the viral capsid and its binding to the ACE2 enzyme is the primary interaction that leads to the virus entering the human cell and the replication of infected cells. Dr. Dickson's group is studying the structure and dynamics of the ACE2 enzyme and how its structure can be impacted by different small drug molecules.
One researcher in Dr. Dickson's group, Dr. Arzu Uyar, specifically uses molecular dynamics simulations and machine learning methods to understand the changes in the ACE2 enzyme and spike protein during the binding process. Dr. Uyar is a postdoctoral research associate whose expertise is in studying protein motions.
This work is both challenging and vital as the protein structures are highly complex. Even the extracellular domain (outer membrane) of ACE2 contains over 10,000 atoms and cataloging and analyzing the structures of these complex proteins takes a lot of effort. That's why Uyar has been developing a computational analysis technique to see which specific structures will be compatible or incompatible with the viral spike protein binding. Dr. Dickson states, "While a human can look at the structure of ACE2 and say 'that's open' or 'that's closed' or 'that's straight' or 'that's bent' we need to find ways to capture those details in an automated way." These methods allow a program to "quickly analyze hundreds of thousands of these structures that come from simulations."
This structure analysis technique can then be used in drug development, as drug molecules that bind to ACE2 will each alter ACE2 structure in a certain way. By seeing which alterations will render the ACE2 unable to bind with the spike protein, researchers could gain insight as to which drugs are viable candidates for treating COVID-19. This can also provide deeper insight into the connection between structure and drug molecule binding, which could be used going forward to design new drugs for a wide range of diseases.
Dr. Dickson's group stated it's important to be aware of how although they're working on developing effective simulations, they're still just that: simulations. Dr. Dickson states that in order to see what is truly effective, we still need clinical trials and experiments. However, accurate computational simulations are vital as they generate hypotheses that can be tested experimentally. Essentially, they are generating a list of possible effective compounds that can be tested in clinical experiments. They also provide us with a detailed view of the atomic world, and a greater understanding of exactly how these small drug molecules can affect the dynamics of their targets in the cell.
While they've made excellent strides in their research efforts, Dr. Dickson and Dr. Uyar also discussed some of the challenges they've faced during the span of their project. Dr. Dickson stated that their lab had shut down in early March due to COVID-19 and though they're still able to conduct their research from home, it's difficult to keep things at the same speed. He said that though they're still able to communicate and collaborate at home, "there's something lost in the interactions." Despite this, they've still been able to run their simulations at a very fast rate and get the resulting data out and available for other researchers to use in their projects and efforts towards developing treatments for COVID-19.
Dr. Uyar largely attributes this success to the computational resources they've acquired in addition to the data other groups have already shared online. Dr. Uyar stated that typically with these types of projects "the greatest challenge is getting more computational resources." She expressed many thanks to ICER and MSU for providing the group with high powered GPUs that enable them to run the simulations much faster. Dr. Uyar said that "Without HPCC, our simulations would take several months but with them we finished them in two weeks or even less than two weeks."
Additionally, both Dr. Uyar and Dr. Dickson expressed praise and thanks to the D.E. Shaw Research group in New York. Both researchers stated that D.E. Shaw Research "performed several MD simulations on ACE2 and the virus, with 78 potential inhibitor drugs bound, and shared all of their data." After Dr. Dickson found out about this data being available on twitter, he contacted Dr. Uyar and they were able to download the data almost instantly. Dr. Uyar is grateful that she was able to use this data to test her classifier algorithm and develop effectiveness scores for each compound in order to predict their potential to block viral entry. Furthermore, with that data Dr. Dickson's group was able to obtain results in only a day or two. Both Dr. Dickson and Dr. Uyar place a heavy emphasis on their support for the importance of collaboration and making any data pertaining to COVID-19 available for public use.
Dr. Dickson emphasizes that "In these days when everyone is stuck at home, having a project and a place to dedicate your effort and thoughts to can help cope with the world around us. Some people see work as an obligation but work is also a motivating force and a reason for being so that's something that I think about."
Provided by Michigan State University