The spike protein of SARS-CoV-2 has been the focus of a huge collective effort in computational research to fight the COVID-19 pandemic. It is crucial to the virus’s function, and also represents the best target for treating the disease. Gerhard Hummer of the Max Planck Institute of Biophysics has been leading a project that aims to elucidate the structure and dynamics of this infamous protein.
Gerhard Hummer and his team at the Max Planck Institute of Biophysics use computer simulations in various forms and flavours to look at the biomolecular processes that make the natural world tick. From the precise movements of protons in an enzyme, all the way up to the large-scale remodelling processes that lead to complex cellular structure, the group’s work provides fascinating insights into the layers of nature that lay just beyond our perception.
Before the outbreak of COVID-19, some members of Hummer’s group had already been investigating viral proteins involved in membrane interactions and membrane remodelling. Specifically, they had been actively working on HIV-1 Env – the HIV equivalent of the SARS-CoV-2 spike protein – trying to understand how it carries out its function of breaking cellular defence systems and getting the virus inside.
Combining their expertise in this field with their broad toolset of molecular simulations, Hummer’s group has been using its resources to help fight the pandemic at the frontline of computational science. Their first task was to build a molecular model of the SARS-CoV-2 spike protein – the protein responsible for getting the virus inside of cells which is also the target of many potential medicines and vaccines. Collaborating with scientists from across the world, this significant piece of work was completed in record time.
“The model provides a picture of the 3D structure of the protein, but also other aspects that are less well understood,” says Hummer. “For instance, post-translational modifications can add sugar molecules to the protein’s surface and change the structure of the viral membrane.”
The first thing that strikes anyone studying the spike protein is that it is a huge molecule. Its other defining characteristic, which became apparent after the creation of the 3D model, is its flexibility. The implications of this were that to carry out molecular dynamics simulations of the spike interacting with other molecules would require very large computational resources. “The spike protein can move and bend like a flower in the wind,” says Hummer. “To see how this affects its interactions with other molecules, you have to have a large space surrounding it in your simulations.”
Model of the spike protein. The three individual chains of the spike are shown in shades of red, N-glycosylation in blue, lipids of the endoplasmic reticulum–like membrane in grey, and phosphates in green. “Hip,” “knee,” and “ankle” mark positions of the three flexible hinges.
Image source: Science, In situ structural analysis of SARS-CoV-2 spike reveals flexibility mediated by three hinges
Biological systems cannot be understood as static structures. The dynamics – the way that the molecules move and interact with other molecules – are extremely important. The spike protein in SARS-CoV-2 has to undergo a series of large conformational changes to latch onto the cell and open up a passageway into the cell.
As well as this, the sugar-like molecules that coat its surface play a large role in its function. “Investigating how the movement of these sugar molecules play a role in the spike’s function, as well as making parts of the spike more accessible to antibodies, has been one of the main focuses of our PRACE-supported project.”
Another interesting aspect of the spike protein is the way it is connected to the virus. The bulbous end of the spike that carries out the main function of latching on to human cells is connected to the rest of the virus by a long, slender, and very flexible stalk. This stalk has multiple “hinges” that allow the active part of the spike to connect with our cells in various different ways.
“This is an aspect of the spike protein that only emerged through the simulations of our PRACE project,” says Hummer. “While doing this work, we learned from our experimental collaborators that they had created some extremely detailed tomographic images of the live virus which confirmed much of what we had found in high resolution. This provided our project with fresh impetus, as we were able to see whether our models were consistent with their data. Our collaboration has enriched each other’s work in a really nice way.”
Hinge flexibility in the MD simulation illustrated through backbone traces (grey) at 75-ns intervals with different parts of the spike protein fixed (red).
The project has not been without its share of obstacles. The widely publicised cyber-attacks on supercomputers across Europe delayed the start of the project, but work is now well underway and Hummer and his team are nearing the end of their PRACE allocation. Looking forwards, they are now pushing their research efforts in several directions. “One really interesting avenue is to do with the engagement of the virus with human receptor proteins,” says Hummer. “Experimental data has produced quite a lot of information about this subject, but the dynamics of it – how it is orchestrated through space and time – is still very poorly understood. Our models and our previous work have put us in an excellent position to study this in more detail, which can hopefully shed some more light on possible ways to interfere with this interaction for therapeutic purposes.”
Another avenue for possible further work is understanding exactly how antibodies interact with the spike protein. “We want to look at how spike proteins and so-called neutralizing antibodies interact in large systems,” says Hummer. “As well as this, we want to further investigate the structure of the viral membrane. SARS-CoV-2 has quite an unusual membrane composition, but more experimental data will probably be needed before we can proceed with any work in that direction.”
This article was also published in PRACE Digest 2020.
20 million core-hours on SuperMUC-NG, hosted by GCS at LRZ, Germany
B. Turoňová, M. Sikora, Ch. Schürmann, W.J.H. Hagen, S. Welsch, F.E.C. Blanc, S. von Bülow, M. Gecht, K. Bagola, C. Hörner, G. van Zandbergen, J. Landry, N. Trevisan Doimo de Azevedo, S. Mosalaganti, A. Schwarz, R. Covino, M.D. Mühlebach, G.Hummer, J.K. Locker and M. Beck. In situ structural analysis of SARS-CoV-2 spike reveals flexibility mediated by three hinges, Science 370 (2020)