Multiscale simulations of cell adhesion receptor clustering

After the onset of multicellularity, organisms have evolved various strategies for regulating and maintaining the tissue integrity throughout their lives. Classical cadherins are the most widespread family of transmembrane proteins mediating cell-cell adhesion and their downregulation is one of the hallmarks of cancer metastasis. Cadherins need to act as sensitive receptors to recognise nascent contacts with neighboring cells, but also as robust mechanical linkers to withstand strong tissue-scale forces. They meet these demands by switching between a dispersed state and nano-to-microscale clusters upon cell-cell contact formation. Clustering was shown to require a slow nucleation step, only after which the energetically expensive cellular pathways leading to contact expansion and maturation are initiated. Knowledge of how cells initiate, maintain and downregulate contacts with their neighbours is an indispensable prerequisite for understanding most of the tissue remodelling processes, with implications on early development and on potential strategies of cancer therapies. Details of adhesion formation have been mostly resolved on the cell and tissue length and time scales, as well as at sub- nanometer level where high resolution structures are available. The mesoscopic structural and dynamic details of this process, however, remain unknown, precluding efficient cross-talk between molecular level studies and cell biology experiments. Aim 1 of this project is to resolve structural and kinetic details of the initial steps of cadherin clustering. To this end, theoretical coarse grained model of the adhesion receptors will be built and evolved using Monte Carlo and Molecular Dynamics simulations to identify key factors influencing the clustering of cadherin molecules. Aim 2 is to translate the outcome of the simulations into macroscopic observables, directly accessible in the cell biology experiments, with the help of path based simulation and analysis methods. The theoretical predictions will be tested in the Zebrafish gastrulation model system, where fast cellular rearrangements are crucial for correct tissue differentiation and positioning. As an expected result, the nucleation of cadherin will be elucidated, but also a practical workflow will be created that brings to bear state of the art coarse grained simulation methods to the problems of cell biology.

Multicellular organisms rely on adhesion proteins to connect their cells and maintain integrity of the tissues. Pathogens also developed specialized adhesion proteins to anchor themselves to our cells and complete their life cycle. These proteins usually do not act alone but rather form multi-protein assemblies to perform their function. The main objective of the project J4332-B28 entitled "Multiscale simulations of cell adhesion receptor clustering" was the understanding of how interactions between adhesion molecules lead to large functional assemblies. In the course of the project, I combined computer simulations with electron microscopy data to solve, in an innovative way, the structure of a desmosome, strong type of cell-cell adhesion found in e.g. skin and heart, where mechanical resistance is important. I found that a unique truss-like structure and overall large stickiness of adhesion molecules is responsible for the desmosome stability. Implications of this study range from methodological (novel approach to solving intrinsically flexible structures, new method to study protein-protein interactions) to biomedical (skin and heart diseases). Beyond that, I was involved in the urgent research to combat the SARS-CoV-2 pandemic. In collaboration with structural biology groups, we have obtained the first electron microscopy images of intact viruses and solved the structure of the spike protein (virus uses it to attach to our cells) in the native context. This study revealed that spike proteins look like lollipops with very flexible sticks attached to the surface of spherical viruses, which could be used by the virus to adjust to uneven surface of cells in our bodies. We have also found, that sugar molecules which are attached to the spike proteins are constantly moving and act like windshield sweepers in a car, sweeping away the antibodies we produce thus protecting the virus. We detected several hidden spots where these sweepers cannot reach and are difficult for the virus to change. These sites can be used to design vaccine constructs that are resistant to future virus variants (one such construct is being tested by our collaborators). These results are important for the biomedical industry but also show how computer simulations can be used in the future emergencies. Knowing the importance of sugars in adhesion molecules and beyond, we developed a webtool to add these sugars to protein models without having to run very expensive simulations. It will be of use for structural biology, where these sugars have been largely ignored to date. In summary, the impact of the project was very interdisciplinary, particularly important in the fields of cell biology, electron microscopy and antiviral medicine. Results of the project further attracted general media interest including New York Times, National Geographic or Frankfurter Allgemeine Zeitung.