By Robert Sanders, Media Relations | November 12, 2021
UC Berkeley scientists have modeled the atomic-level interactions between the SARS-CoV-2 spike protein and the human ACE2 receptor, as well as antibodies that target the spike protein and prevent infection. (Credit: Jason McLellan Laboratory/University of Texas at Austin)
A supercomputer at the National Energy Research Scientific Computing Center (NERSC) helped researchers at the University of California, Berkeley, visualize and study how the spike protein on the surface of the SARS-CoV-2 virus attaches to a receptor protein on human cells and how various antibodies block that interaction.
The simulations showed, somewhat unexpectedly, that the spike protein of the highly infectious Delta variant — the dominant strain in the U.S. and much of the world at the time of the study — is particularly good at binding to the human ACE2 receptor, which serves as the gateway for the virus to enter human cells.
The researchers also observed, however, that antibodies elicited by both infection with SARS-CoV-2 and the current, mRNA-based vaccines, such as those manufactured by Pfizer, Moderna and BioNTech, neutralize the Delta variant just as well as they do other variants, including the original strain that emerged in Wuhan, China, in late 2019.
“People have assumed that the Delta variant is more infectious because the spike protein is better at binding to ACE2, and that may be true, but our simulations show that’s not the whole story,” said study leader Rommie Amaro, a UC Berkeley professor of chemistry and of bioengineering and a Howard Hughes Medical Institute investigator. “In addition to how the spike protein binds ACE2, you have to consider how antibodies and other parts of the immune system fight back.”
The team published their findings in the journal Nature Structural and Molecular Biology. The corresponding authors are Amaro and her former postdoctoral scholar, Jose Manuel Flores-Solis, now an associate research scientist at the University of Texas at Austin.
Simulations performed on DOE’s NERSC supercomputer Cori
To get an atomic-level view of how the spike protein interacts with the ACE2 receptor and antibodies, the researchers first simulated the molecular structures of the spike protein, the receptor and several classes of antibodies. They did this using the crystal structures of the spike protein, the ACE2 receptor and several of the antibodies, which had been determined by other researchers.
They then “soaked” the spike protein into a lipid bilayer, a fatty membrane like that which encapsulates the SARS-CoV-2 virus, and simulated how the spike protein moved around and interacted with water molecules and other molecules on the surface of the lipid bilayer. This allowed them to determine the different conformations that the spike protein could adopt and how tightly the spike protein bound to the ACE2 receptor.
“Spike protein flexibility is important, because the different conformations of the spike protein determine which antibodies can bind to it,” Amaro said.
The researchers performed many simulations, looking at how well different variants of the spike protein attach to the human ACE2 receptor, how different classes of antibodies bind to the spike protein, and how effective different antibody combinations are at binding to and neutralizing the spike protein.
To perform so many simulations, the team used powerful computing resources at NERSC in Berkeley, including the Cori supercomputer.
Delta strain binds better, but antibodies neutralize it just as well
One of the most important things they learned is that the spike protein of the Delta variant binds more tightly to the human ACE2 receptor than does the spike protein of the original SARS-CoV-2 strain. This is probably one of the reasons the Delta variant is so infectious.
The team also found that most classes of antibodies tested were equally effective at binding to and neutralizing the Delta variant as they were at neutralizing the original strain. The notable exceptions were antibodies that target a part of the spike protein called the N-terminal domain. These antibodies were less effective at binding to the Delta variant, and this weakened their ability to neutralize it. Fortunately, most antibodies in the body’s immune response and most of those induced by the COVID-19 vaccines target other regions of the spike protein that don’t vary much between the Delta variant and earlier variants.
“In general, our results support continued use of the currently authorized mRNA-based vaccines, even in the presence of the Delta variant, because they elicit antibodies that appear to neutralize the Delta variant with similar potency as they do other variants,” the researchers wrote.
Co-authors of the paper include former UC Berkeley postdoctoral scholars Alexander Pak, now an assistant professor in the department of molecular medicine and genetics at the University of South Florida, and Daniel Wrapp, now an assistant professor of biochemistry at the University of Arizona. Other co-authors are researchers at UC San Francisco and the University of Colorado, Boulder.
This work was supported by grants from the National Institutes of Health (R01AI093984-07S1, U54HG007013, and R01AI093984), the California Initiative on Precision Medicine at UC San Francisco, and the UC Berkeley Department of Bioengineering. Use of the computational resources at NERSC was supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02- 05CH11231.
