Biophysical Society Thematic Meeting | Hamburg 2022

Biophysics at the Dawn of Exascale Computers

Poster Abstracts

48-POS Board 48 UNDERSTANDING THE EFFECTS OF POST-TRANSLATIONAL MODIFICATIONS ON THE PHASE BEHAVIOUR OF TDP-43 AND ITS ROLES IN DISEASE Lukas S. Stelzl 1,2 ; Lisa M Pietrek 3 ; Gerhard Hummer 3,4 ; Dorothee Dormann 1,2 ; 1 Johannes Gutenberg University, Mainz, Germany 2 Institute of Molecular Biology (IMB), Mainz, Germany 3 Max Planck Institute of Biophysics, Theoretical Biophysics, Frankfurt am Main, Germany 4 Goethe University, Institute of Biophysics, Frankfurt am Main, Germany The discovery of liquid-liquid phase separation is revolutionizing our understanding of cellular biophysics. Dysregulation of phase separation is implicated in neurodegenerative diseases. According to the "stickers and spacers" model from polymer science, some residues - the "stickers" - engage in favorable interactions with other chains but most residues form few stabilizing interactions - the "spacers". However, these interactions are not understood at the molecular scale and thus we lack a predictive understanding of how mutations and post- translational modifications affect phase behavior and how phase-separated condensates recognize other molecules. With multi-scale simulations, we can simulate the phase behavior of proteins and zoom in on biomolecular condensates with atomic resolution to understand protein- sequence specific interactions and molecular recognition. Investigating TDP-43, which is involved in neurodegenerative disease, we find that aromatic residues act as "stickers". Phosphorylation of TDP-43 is a hallmark of neurodegenerative disease. TDP-43 phosphorylation is thought to drive condensation and subsequent pathological aggregation, but recent experiments have suggested the opposite: Phosphorylation protects rather than harms cells. Indeed, we find that in coarse-grained simulations of TDP-43 phase behaviour that hyper-phosphorylation can prevent TDP-43 condensation. To better understand how TDP-43 condensates are shaped by sequence-specific interactions we use atomistic molecular dynamics simulations. We find that phosphomicking mutations increase interactions with solvent, which suggests a mechanistic basis for the protective effect of phosphorylation. Chain-growth Monte Carlo enables us to generate meaningful and independent start configurations and we envisage that this could provide a way forward for large-scale atomistic molecular dynamics simulations on the emerging next generation of high-performance computing resources.

103