Abstract

Spatiotemporal organization is key to transforming a “bag of molecules” into a functional cell capable of exerting complex tasks, such as chromosome segregation and cell division. In bacteria, molecular transport and positioning relies on protein systems centered around ParA-type ATPases, which serve as nucleotide-dependent molecular switches. Of these systems, the Escherichia coli MinCDE system, which self-organizes by a reaction-diffusion mechanism, has been studied the most extensively. Based on the ATPase MinD, its ATPase-activating protein MinE, the passenger protein MinC and the membrane as a reaction matrix, this minimal oscillator defines the midcell position in E. coli. In this thesis, I set out to further refine the understanding of the MinDE self-organization mechanism and to decipher additional roles of the MinDE system, by taking advantage of the established in vitro reconstitution assay of Min(C)DE self-organization. By providing a detailed protocol for the in vitro reconstitution assay and reviewing experimental approaches and insights on the MinCDE system in detail, I further established the MinCDE system as a model for reaction-diffusion systems and pattern formation in biology. While the general mechanism of MinDE self-organization has been extensively studied and is remarkably well understood, the nature of the autocatalytic step, the cooperative MinD membrane binding, had remained elusive. To shed light on the molecular mechanism of this step, I employed high-speed atomic force microscopy to visualize MinDE oscillations on nanometer-sized membrane patches. Analysis of the kinetics of MinDE oscillations dependent on protein concentrations and membrane patch size revealed the different oscillation phases. Based on these results, I proposed that MinD associates into higher order structures on the membrane, thereby enabling high attachment and detachment cooperativity. Moreover, in this thesis, I found that the ATP-consuming MinDE dynamics may play a role beyond regulating MinC/FtsZ localization. Prior to this work, the biological purpose of the MinDE oscillations had been exclusively attributed to the spatiotemporal regulation of MinC, that passively follows the oscillations by binding specifically to MinD. These oscillations result in a temporal concentration gradient of MinC, which is an inhibitor of FtsZ polymerization, thereby restricting FtsZ assembly to midcell. Taking advantage of the well-established in vitro reconstitution assay of MinDE self-organization on planar supported lipid bilayers, I showed that MinDE dynamics are able to spatiotemporally regulate membrane-bound molecules by a non-specific mechanism that does not require binding of the components to MinDE. MinDE self-organization induced patterns of membrane-bound molecules that were anti-correlated to MinDE accumulation on the membrane. Regulation occurred for a wide variety of different membrane-bound molecules, including peripheral membrane proteins, lipid-anchored proteins and DNA molecules, as well as for a wide range of MinDE patterns. Intriguingly, when the membrane-bound molecules exhibited a long dwell-time on the membrane, i.e. were anchored via cholesteryl or via streptavidin to biotinylated lipids, the regulation by MinDE resulted in large-scale gradients of these molecules on the membrane, indicative of a net transport. When co-reconstituted in rod-shaped microcompartments MinDE pole-to-pole oscillations drove counter-oscillations of these tightly attached molecules, resulting in time-averaged protein gradients with maximum concentration at the compartment middle. Heterologous expression of MinDE and model peripheral membrane proteins in the fission yeast Schizosaccharomyces pombe demonstrated that regulation of membrane proteins by MinDE also occurs in a physiologically more relevant context, with cytoplasmic and membrane crowding. These findings imply that MinDE is able to position a much larger set of proteins in E. coli than previously known, thereby contributing, independent of MinC, to division site selection by prepositioning divisome proteins to midcell. Applying this simplistic mechanism to transport a synthetic cargo, i.e. highly controllable, membrane-anchored DNA origami nanostructures, I characterized the phenomenon in more detail. By systematically modifying the nature and number of membrane anchors of the DNA origami cargo, I found that the membrane footprint of the DNA origami, akin to an interaction area, is likely to determine the extent of the transport by MinDE. Using this knowledge, I demonstrated that MinDE self-organization is able to spatially sort different DNA origami nanostructures. Single particle tracking revealed that diffusion of DNA origami in the presence of membrane-bound MinD was reduced, indicating that they experience friction by directly colliding with MinD. Finally, by harnessing the geometry sensitivity of the MinDE system and the unique properties of DNA origami nanostructures, I could show that MinDE-induced transport can be directed and applied to pattern membrane-bound molecules on the micron scale. Taken together, the experiments presented in this thesis suggest that MinDE is able to position, transport and sort membrane-bound molecules by a purely non-specific physical interaction, which modulates the membrane binding/unbinding and diffusion of these molecules. This previously unknown physicochemical transport mechanism is based on two proteins only and is remarkably simple. Thus, I speculate that also other bacterial or eukaryotic self-organizing systems are capable of regulating a large set of proteins by a similar non-specific effect. The controllability together with the non-specific nature of the effect highlight its applicability to bottom-up synthetic biology and nanotechnology, where it could be harnessed to position and transport membrane-bound molecular assemblies.