Publication
Biomolecular condensates are cellular organelles without a membrane, which have come into the scientific focus about 15 years ago. They are made from a mix of proteins and RNAs. It is hypothesized that they maintain their shape and function through liquid-liquid phase separation. The functions of condensates are diverse and include the catalysis of chemical reactions and the sequestering of specific proteins and RNAs in moments of cellular stress. This suggest that there are active, energy-consuming processes involved in the maintenance of biomolecular condensates, which is confirmed by a growing number of experimental observations.
Active processes in condensates have already been modelled to a certain extent in the framework of non-equilibrium thermodynamics, starting from ad hoc assumptions about the underlying microscopic structure. In this thesis, we propose a more complex mathematical model which derives mesoscopic equations from a microscopic description of the particles in the system. This model enables us to describe diffusion and chemical reactions as a function of space for arbitrary chemical networks involving activity. Although restricted to the vicinity of equilibrium, our model provides some insights into the interplay between activity, diffusive fluxes and spatial distributions across the condensate-cytosol interface. We also find a set of minimal conditions necessary for the system to exhibit diffusive fluxes in the presence of active driving.
We apply our model to two different chemical networks, using both analytical and numerical methods. Firstly, we study a simple system exhibiting a chemical cycle, such as might be involved in the maintenance of the condensate in general. We show that in this system, the diffusive fluxes are maximal at the condensate-cytosol interface and are further augmented for larger interface widths. We discuss implications of these observations for condensates that function as chemical factories and point out similarities to experimental observations.
Secondly, we study a more complex chemical network which includes a non-linear aggregation step. This network allows us to investigate how active processes might be involved in the maintenance of biomolecular condensates in the liquid state and the prevention of a transition into a pathologic fibrillar state. Using our model, we show how different system parameters influence the concentration distributions and fluxes in the system. We construct settings in which turning on the active reaction is bringing the system from an aggregated state into a liquid state. Our model can also be applied to investigate the behaviour and distribution of proteins that partition into condensates, but are not themselves prone to condensation.
Although our model has several constraints and points for future improvement, it is a powerful tool that offers new insights into the interplay between chemical reactions, activity and spatial fluxes at condensate interfaces. A