Endocytosis has been identified as a genuine multiscale process in which molecular reaction kinetics, supramolecular assembly, morphological dynamics of various membrane bound compartments and whole-cell transport process are coupled in a way that resists efficient description on a single length scale.
We have developed a simulation platform that takes account of this multiscale property by extending stochastic chemical kinetics with an enhancement of a well-established method from soft-matter physics, multipolar reactive dissipative particle dynamics (mpRDPD). DPD, as a scalable, particle-based method, allowed us to implement complex reaction kinetics (including combinatorial chemistry) in a way that can be directly interfaced with standard means of communication in systems biology
such as SBML. Endowing the basic constituents of the simulation (point particles) with additional structure (multipole moments) and appropriate interactive dynamics results in the emergent formation of membrane structures.
These membranes not only arise as metastable or equilibrium structures, but owing to the specific treatment of mechanical interactions in DPD (complete momentum conservation even with random forces) they also behave properly in dynamic situations. This includes proper mesoscopic behavior in the hydodynamic limit, allowing for example motor protein induced endosome transport to be modelled consistently. More importantly, morphological dynamics such as chemically controlled vesicle budding from large compartments or endosome fission and fusion caqn be treated physically. Besides implementing reaction kinetics and interactions enabling the emergence of extended complex structures, the link-up of chemistry and membrane dynamics requires a proper treatment of transport processes on different length scales. We have achieved significant progress towards a self-consistent treatment of reactions and mesoscale dynamics e.g. by making certain to keep the simulated membranes (despite their mechanical stability) in a fluid state. This is a necessary condition for studying dynamic Rab-domain formation, a key process in endocytosis, but also for membrane bound chemistry in general.
We are also applying this simulation tool to examine compartment differentiation by protein sorting in endocytosis, extending the purely chemical kinetics model proposed by Heinrich and Rapaport.
The fact that complex mechanical and structural processes can be simulated with entities interacting only via local interactions is of considerable relevance for studies in systems biology. Experimentally accessible are usually only local concentrations of molecules (e.g. via fluorescence imaging) or other local quantities. This means that an mpRDPD-based platform can be directly linked up with high throughput imaging data and therefore can serve as a tool that allows an automated connection of experiments with theoretical predictions, thereby opening new possibilities for control and analysis in biological research.