Abstract :
[en] Life is based on energy conversion by which cells and organisms can adapt to the environment. The involved biological processes are intrinsically multiscale phenomena since they are based on molecular interactions on a small scale leading to the emerging behavior of cells, organs and organisms. To understand the underlying regulation and to dissect the mechanisms that control system behavior, appropriate mathematical multiscale models are needed. Such models do not only offer the opportunity to test different hypothesized mechanisms but can also address current experimental technology gaps by zooming in and out of the dynamics, changing scales, coarse-graining the dynamics and giving us distinct views of the phenomena. In this dissertation substantial efforts were done to combine different computational modeling strategies based on different assumptions and implications to model an essential system of eukaryotic life -- the energy providing mitochondria -- where the spatiotemporal domain is suspected to have a substantial influence on its function.
Mitochondria are highly dynamic organelles that fuse, divide, and are transported along the cytoskeleton to ensure cellular energy homeostasis. These processes cover different scales, in space and time, where on the more global scale mitochondria exhibit changes in their molecular content in response to their physiological context including circadian modulation. On the smaller scales, mitochondria show also faster adaptation by changing their morphology within minutes. For both processes, the relation between the underlying structure of either their regulating network or the spatial morphology and the functional consequences are essential to understand principles of energy homeostasis and their link to health and disease conditions.
This thesis focuses on different scales of mitochondrial adaptation. On the small scales, fission and fusion of mitochondria are rather well established but substantial evidence indicates that the internal structure is also highly variable in dependence on metabolic condition. However, a quantitative mechanistic understanding how mitochondrial morphology affects energetic states is still elusive. In the first part of this dissertation I address this question by developing an agent-based dynamic model based on three-dimensional morphologies from electron microscopy tomography, which considers the molecular dynamics of the main ATP production components. This multiscale approach allows for investigating the emergent behavior of the energy generating mechanism in dependence on spatial properties and molecular orchestration. Interestingly, comparing spatiotemporal simulations with a corresponding space-independent approach, I found only minor space dependence in equilibrium conditions but qualitative difference in fluctuating environments and in particular indicate that the morphology provides a mechanism to buffer energy at synapses.
On the more global scale of the regulation of mitochondrial protein composition, I applied a data driven approach to explore how mitochondrial activity is changing during the day and how food intake restrictions can effect the structure of the underlying adaptation process. To address the question if at different times of the day, the mitochondrial composition might adapt and have potential implications for function, I analyzed temporal patterns of hepatic transcripts of mice that had either unlimited access to food or were hold under temporal food restrictions. My analysis showed that mitochondrial activity exhibits a temporal activity modulation where different subgroups of elements are active at different time points and that food restriction increases temporal regulation.
Overall, this thesis provides new insights into mitochondrial biology at different scales by providing an innovative computational modeling framework to investigate the relation between morphology and energy production as well as by characterizing temporal modulation of the regulatory network structure under different conditions.
