Prominin-1 (alias CD133) has received considerable interest because of its expression by several stem and progenitor cells originating from various sources, including the neural and hematopoietic systems. As a cell surface marker, prominin-1 is now used for somatic stem cell isolation. Its expression in cancer stem cells has broadened its clinical value, as it might be useful to outline new prospects for more effective cancer therapies by targeting tumor-initiating cells. Cell biological studies of this molecule have demonstrated that it is specifically concentrated in various membrane structures that protrude from the planar areas of the plasmalemma. Prominin-1 binds to the plasma membrane cholesterol and is associated with a particular membrane microdomain in a cholesterol-dependent manner. Although its physiological function is not yet determined, it is becoming clear that this cell surface protein, as a unique marker of both plasma membrane protrusions and membrane microdomains, might reveal new aspects of the cell biology of rare stem and cancer stem cells. The aim of this review is to outline the recent discoveries regarding the dynamic reorganization of the plasma membrane of rare CD133+ hematopoietic progenitor cells during cell migration and division. / Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG-geförderten) Allianz- bzw. Nationallizenz frei zugänglich.

Cytokinesis involves constriction of the cell at the equator. Without decreasing in volume, a spherical cell requires a net increase in the surface area during this constriction. The constriction is driven by formation of an actomyosin contractile ring, and the surface increase by addition of membrane during the formation of the cleavage furrow. Both events depend on the central spindle microtubules at the midzone of the spindle and, in particular, on the centralspindlin protein complex. The communication between the central spindle microtubules and the actomyosin ring involves binding of a GAP and a GEF for RhoA to the centralspindlin kinesin Pavarotti/MKLP1. However, it is still unclear which molecular machinery connects the mitotic spindle to membrane trafficking during cleavage furrow ingression. ARF6 is a member of the ARF family of small GTPases, and previous studies suggest that it is an important regulator of membrane trafficking through the endocytic pathway, and cortical Actin remodelling. I generated an arf6 null mutant in Drosophila. arf6 null mutants survive to adulthood without obvious morphological defects, indicating that ARF6 is not required for Drosophila somatic development. However, ARF6 is required for cytokinesis in Drosophila spermatocytes. The centralspindlin kinesin Pavarotti, identified as an ARF6 interactor in a Yeast-2-Hybrid assay, binds ARF6 in GST pulldowns, and interacts genetically with the arf6 mutant. ARF6 localizes to the plasma membrane and a population of early and recycling endosomes. During cytokinesis, ARF6 is enriched on recycling endosomes at the central spindle. arf6 mutants form a cleavage furrow during cytokinesis, which later regresses. Cytokinesis in arf6 mutant spermatocytes lacks the rapid plasma membrane expansion observed during normal divisions. The results of this study suggest that ARF6 might promote rapid recycling of endosomal membrane stores at the central spindle to the plasma membrane during cytokinesis. ARF6 might be recruited to the central spindle via its interaction with Pavarotti, and act as part of the molecular link between the central spindle cytoskeleton and the rapid plasma membrane addition necessary for cytokinesis. Für die Ansicht der quick-time-Movies mit der Endung "avi" ist die Installation des "Apple QuickTime-Players" erforderlich.

Almost all animal cells maintain a thin layer of actin filaments and associated proteins underneath the cell membrane. The actomyosin cortex is subject to internal stress patterns which result from the spatiotemporally regulated activity of non-muscle myosin II motors in the actin network. We study how these active stresses drive changes in cell shape and flows within the cortical layer, and how these cytoskeletal deformations and flows govern processes such as cell migration, cell division and organelle transport. Following a continuum mechanics approach, we develop theoretical descriptions for three different cellular processes, to obtain - in collaboration with experimental groups - a detailed and quantitative understanding of the underlying cytoskeletal mechanics. We investigate the forces and cortex flows involved in adhesion-independent cell migration in confinement. Many types of cell migration rely on the extension of protrusions at the leading edge, where the cells attach to the substrate with specific focal adhesions, and pull themselves forward, exerting stresses in the kPa range. In confined environments however, cells exhibit migration modes which are independent of specific adhesions. Combining hydrodynamic theory, microfluidics and quantitative imaging of motile, non-adherent carcinosarcoma cells, we analyze the mechanical behavior of cells during adhesion-independent migration. We find that the accumulation of active myosin motors in the rear part of these cells results in a retrograde cortical flow as well as the contraction of the cell body in the rear and expansion in the front, and we describe how both processes contribute to the translocation of the cells, depending on the geometric and mechanical parameters of the system. Importantly, we find that the involved propulsive forces are several orders of magnitude lower than during adhesive motility while the achieved migration velocities are similar. Moreover, the distribution of forces on the substrate during non-adhesive migration is fundamentally different, giving rise to a positive force dipole. In contrast to adhesive migration modes, non-adhesive cells move by exerting pushing forces at the rear, acting to expand rather than contract their substrate as they move. These differences may strongly affect hydrodynamic and/or deformational interactions between collectively migrating cells. In addition to the work outlined above, we study contractile ring formation in the actin cytoskeleton before and during cell division. While in disordered actin networks, myosin motor activity gives rise to isotropic stresses, the alignment of actin filaments in the cortex during cell division introduces a preferred direction for motor-filament interactions, resulting in anisotropies in the cortical stress. Actin filaments align in myosin-dependent shear flows, resulting in possible feedback between motor activity, cortical flows and actin organization. We investigate how the mechanical interplay of these different cortical properties gives rise to the formation of a cleavage furrow during cell division, describing the level of actin filament alignment at different points on the cortex with a nematic order parameter, in analogy to liquid crystal physics. We show that cortical anisotropies arising from shear-flow induced alignment patterns are sufficient to drive the ingression of cellular furrows, even in the absence of localized biochemical myosin up-regulation. This mechanism explains the characteristic appearance of pseudocleavage furrows in polarizing cells. Finally, we study the characteristic nuclear movements in pseudostratified epithelia during development. These tissues consist of highly proliferative, tightly packed and elongated cells, with nuclei actively travelling to the apical side of the epithelium before each cell division. We explore how cytoskeletal properties act together with the mechanics of the surrounding tissue to control the shape of single cells embedded in the epithelium, and investigate potential mechanisms underlying the observed nuclear movements. These findings form a theoretical basis for a more detailed characterization of processes in pseudostratified epithelia. Taken together, we present a continuum mechanics description of the actomyosin cell cortex, and successfully apply it to several different cell biological processes. Combining our theory with experimental work from collaborating groups, we provide new insights into different aspects of cell mechanics.

Das stäbchenförmige Bakterium Escherichia coli teilt sich in zwei gleich große Tochterzellen. Dies ist nur möglich, wenn sich die Zelle in der Mitte teilt. Bei E. coli wird die Zellteilung durch den Zusammenschluss der FtsZ-Proteine an der Membran zum Z-Ring eingeleitet. Topologische Regulierung des Z-Ringes erfolgt durch räumlich-zeitliche Oszillationen von Min-Proteinen zwischen den beiden Zellpolen. MinC, MinD und MinE binden an und lösen sich von der Membran unter Hydrolyse von ATP und in antagonistischer Art und Weise, was zu einer alternierenden Ansammlung von MinC und MinD an den Zellpolen führt. Gemittelt über die Zeit ergibt sich somit ein MinD-Verteilungsprofil, das maximale Konzentration an den Zellpolen und ein Minimum in der Zellmitte aufweist. MinC bindet an MinD und folgt somit seiner Verteilung. Der Zusammenschluss von FtsZ-Proteinen wird durch MinC unterbunden, und somit kann sich der Z-ring nur an einer Position herausbilden, die ein Minimum an MinC aufweist - der Zellmitte. Das Min-system wurde in der Vergangenheit auch mit einem in-vitro-Ansatz untersucht, indem Min-Proteine in künstliche, aufliegende Lipiddoppelschichten (supported lipid bilayers, SLB) rekonstitutiert wurden. Dabei bildeten die Min-Proteine kein oszillierendes Muster aus, sondern organisierten sich vielmehr in parallelen und propagierenden Wellen (Loose, 2008, Science, 320). In diesen in-vitro-Experimenten war das Membransubstrat wesentlich größer als die Wellenlänge der Min-Proteinwellen. In vivo hingegen ist die Länge der Zelle in der gleichen Größenordnung wie die charakteristische Länge des Oszillationsmusters der Min-Proteine. Daher war es das Ziel dieser Arbeit, den Einfluß einer beschränkten Fläche und geometrischer Formgebung der künstlichen Lipiddoppelschichten auf die Wellenpropagation der Min-Protein zu untersuchen. Flächige Beschränkung künstlicher Membranen erfolgte durch Mikrostrukturtechnologie. Deckglässchen wurden mit einer Goldschicht und mikroskopischen Aussparungen unterschiedlicher geometrischer Formen strukturiert. Funktionale SLBs bildeten sich nur auf Glasflächen ohne Goldbeschichtung aus. Nach der Rekonstitution der Min-Proteine, organisierten sich diese auf den Membranstücken in parallele Wellen. Dabei bestimmte die flächige Beschränkung der künstlichen Membranen die Ausbreitungsrichtung der Min-Proteinwellen. Min-Proteinwellen konnten entlang gekrümmter Membranstreifen, in Ring- und sogar in Slalomstrukturen geleitet werden. In geraden, länglichen Strukturen richteten sich die Wellen entlang der längsten Achse aus. Kopplung von Proteinwellen auf räumlich getrennten Membranstücken in Abhängigkeit des Abstandes und des sogenannten Molecular Crowdings in der wässrigen Lösung konnte ebenfalls beobachtet werden. Diese Kopplung ist ein Indiz für inhomogene Proteinverteilungen in der Lösung oberhalb der Membran. Desweiteren konnten Min-Proteinwellen auch in diversen dreidimensionalen künstlichen Membranen rekonstitituiert werden. Im Wildtyp von E. coli ähneln die Min-Proteindynamiken der einer Oszillation mit einer charakteristischen Länge von 5 µm. Auf SLBs, bilden Min-Proteine Wellen mit einer Wellenlänge aus, die ca. zehnmal größer ist als in vivo. Dieser Unterschied zwischen der in-vivo- und der in-vitro-Welt wurde untersucht und diskutiert. In vitro konnte die Wellenlänge um 50 % durch Erhöhung des Molecular Crowding in der Lösung sowie um 33 % durch Temperaturerhöhung verkleinert werden. Das oszillierende Muster könnte dahingegen eine Folge der Kompartimentierung sein. Erste Versuche, das Min-System in geschlossene Membrankompartimente zu rekonstitutieren, wurden getestet. / Escherichia coli, a rod-like bacterium, divides by binary fission. Cell division into two daughter cells of equal size requires that fission takes place at a midcell position. In E. coli, cell division is initiated by assembly of the FtsZ-proteins at the inner membrane to the Z-ring. Topological regulation of the Z-ring is achieved by spatiotemporal pole-to-pole oscillations of Min-proteins. MinC, MinD and MinE bind to and detach from - under hydrolysis of ATP - the membrane in an antagonistic manner leading to an alternating accumulation of MinC and MinD at the cell poles. Averaged over time, the distribution profile of MinD exhibits maximal concentration at the cell poles and a minimum at the cell center. MinC binds to MinD and thus follows its distribution. FtsZ assembly is inhibited by MinC and therefore the Z-ring can only form at a cell position low in MinC - at the cell center. In the past, the Min-system was also investigated in an in vitro approach by reconstitution of Min-proteins into a supported lipid bilayer (SLB). Here, Min-proteins did not self-organize into an oscillatory pattern but into parallel and propagating waves (Loose, 2008, Science, 320). In this in vitro assay, the membrane substrate was infinitely large compared to the wavelength. However, in vivo, the cell length is on the same order of magnitude as the respective length scale of the oscillatory pattern of Min-proteins. Therefore, we wished to investigate the effect of lateral confinement and geometric structuring of artificial lipid bilayers on the Min-protein wave propagation. Lateral confinement of artificial membranes was achieved by microfabrication technology. Glass slides were patterned by a gold coating with microscopic windows of different geometries, and functional SLBs were only formed on uncoated areas. Upon reconstitution, Min-proteins organized into parallel waves on the geometric membrane patches. Confinement of the artificial membranes determined the direction of propagation of Min-protein waves. Min-protein waves could be guided along curved membrane stripes, in rings and even along slalom-geometries. In elongated membrane structures, the protein waves always propagate along the longest axis. Coupling of protein waves across spatially separated membrane patches was observed, dependent on gap size and level of molecular crowding of the aqueous media above the bilayer. This indicates the existence of an inhomogeneous and dynamic protein gradient in the solution above the membrane. Furthermore, reconstitution of Min-protein waves in various three-dimensional artificial membranes was achieved. In wild-type E. coli, Min-protein dynamics resemble that of an oscillation with a characteristic length scale of 5 µm. On supported lipid bilayers, Min-proteins self-organize into waves with a wavelength approximately 10-fold larger than in vivo. These discrepancies between the in vivo and in vitro world were investigated and discussed. In vitro, the wavelength could be decreased by a factor of 50 % by increase of the molecular crowding in solution and by 33 % through temperature increase. The oscillatory pattern is thought to be a consequence of compartmentalization and first attempts to encapsulate the Min-system in closed bilayer compartments are presented.

Large doses of recombinant growth factors formulated in solution form directly injected into the body is usual clinical practice in treating second-degree scald injuries, with promising results, but this approach creates side effects; furthermore, it may not allow appropriate levels of the factor to be sensed by the target injured tissue/organ in the specific time frame, owing to complications arising from regeneration. In this research, two delivery methods (infusion pumping and local topical application) were applied to deliver recombinant human erythropoietin (rHuEPO) for skin regeneration. First, rHuEPO was given in deep second-degree scald injury sites in mice by infusion pump. Vascularization was remarkably higher in the rHuEPO pumping group than in controls. Second, local topical application of rHuEPO gel was given in deep second-degree scald injury sites in rats. Histological analysis showed that epithelialization rate was significantly higher in the rHuEPO gel-treated group than in controls. Immunohistochemical studies showed that the rHuEPO gel-treated group showed remarkably higher expression of skin regeneration makers than the control group. An accurate method for visualization and quantification of blood vessel networks in target areas has still not been developed up to this point, because of technical difficulties in detecting such thin blood vessels. A method which utilizes a series of steps to enhance the image, removes noise from image background, and tracks the vessels edges for vessel segmentation and quantification has been used in this study. Using image analysis methods, we were able to detect the microvascular networks of newly formed blood vessels (less than 500 μm thickness), which participate in the healing process, providing not only nutrition and oxygen to grow tissues but also necessary growth factors to grow tissue cells for complete skin regeneration. The rHuEPO-treated group showed higher expression of stem cell markers (CD 31, CD 90, CD 71, and nestin), which actively contribute to in-wound-healing processes for new hair follicle generation as well as skin regeneration. Collectively, both rHuEPO group pumping into the systemic circulation system, and injection into the local injury area, prompted mice and rats to form new blood vessel networks in scald injury sites, which significantly participate in the scald healing process. These results may lead to the development of novel treatments for scald wounds.

Prominin-1 (alias CD133) has received considerable interest because of its expression by several stem and progenitor cells originating from various sources, including the neural and hematopoietic systems. As a cell surface marker, prominin-1 is now used for somatic stem cell isolation. Its expression in cancer stem cells has broadened its clinical value, as it might be useful to outline new prospects for more effective cancer therapies by targeting tumor-initiating cells. Cell biological studies of this molecule have demonstrated that it is specifically concentrated in various membrane structures that protrude from the planar areas of the plasmalemma. Prominin-1 binds to the plasma membrane cholesterol and is associated with a particular membrane microdomain in a cholesterol-dependent manner. Although its physiological function is not yet determined, it is becoming clear that this cell surface protein, as a unique marker of both plasma membrane protrusions and membrane microdomains, might reveal new aspects of the cell biology of rare stem and cancer stem cells. The aim of this review is to outline the recent discoveries regarding the dynamic reorganization of the plasma membrane of rare CD133+ hematopoietic progenitor cells during cell migration and division. / Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG-geförderten) Allianz- bzw. Nationallizenz frei zugänglich.

Almost all animal cells maintain a thin layer of actin filaments and associated proteins underneath the cell membrane. The actomyosin cortex is subject to internal stress patterns which result from the spatiotemporally regulated activity of non-muscle myosin II motors in the actin network. We study how these active stresses drive changes in cell shape and flows within the cortical layer, and how these cytoskeletal deformations and flows govern processes such as cell migration, cell division and organelle transport. Following a continuum mechanics approach, we develop theoretical descriptions for three different cellular processes, to obtain - in collaboration with experimental groups - a detailed and quantitative understanding of the underlying cytoskeletal mechanics. We investigate the forces and cortex flows involved in adhesion-independent cell migration in confinement. Many types of cell migration rely on the extension of protrusions at the leading edge, where the cells attach to the substrate with specific focal adhesions, and pull themselves forward, exerting stresses in the kPa range. In confined environments however, cells exhibit migration modes which are independent of specific adhesions. Combining hydrodynamic theory, microfluidics and quantitative imaging of motile, non-adherent carcinosarcoma cells, we analyze the mechanical behavior of cells during adhesion-independent migration. We find that the accumulation of active myosin motors in the rear part of these cells results in a retrograde cortical flow as well as the contraction of the cell body in the rear and expansion in the front, and we describe how both processes contribute to the translocation of the cells, depending on the geometric and mechanical parameters of the system. Importantly, we find that the involved propulsive forces are several orders of magnitude lower than during adhesive motility while the achieved migration velocities are similar. Moreover, the distribution of forces on the substrate during non-adhesive migration is fundamentally different, giving rise to a positive force dipole. In contrast to adhesive migration modes, non-adhesive cells move by exerting pushing forces at the rear, acting to expand rather than contract their substrate as they move. These differences may strongly affect hydrodynamic and/or deformational interactions between collectively migrating cells. In addition to the work outlined above, we study contractile ring formation in the actin cytoskeleton before and during cell division. While in disordered actin networks, myosin motor activity gives rise to isotropic stresses, the alignment of actin filaments in the cortex during cell division introduces a preferred direction for motor-filament interactions, resulting in anisotropies in the cortical stress. Actin filaments align in myosin-dependent shear flows, resulting in possible feedback between motor activity, cortical flows and actin organization. We investigate how the mechanical interplay of these different cortical properties gives rise to the formation of a cleavage furrow during cell division, describing the level of actin filament alignment at different points on the cortex with a nematic order parameter, in analogy to liquid crystal physics. We show that cortical anisotropies arising from shear-flow induced alignment patterns are sufficient to drive the ingression of cellular furrows, even in the absence of localized biochemical myosin up-regulation. This mechanism explains the characteristic appearance of pseudocleavage furrows in polarizing cells. Finally, we study the characteristic nuclear movements in pseudostratified epithelia during development. These tissues consist of highly proliferative, tightly packed and elongated cells, with nuclei actively travelling to the apical side of the epithelium before each cell division. We explore how cytoskeletal properties act together with the mechanics of the surrounding tissue to control the shape of single cells embedded in the epithelium, and investigate potential mechanisms underlying the observed nuclear movements. These findings form a theoretical basis for a more detailed characterization of processes in pseudostratified epithelia. Taken together, we present a continuum mechanics description of the actomyosin cell cortex, and successfully apply it to several different cell biological processes. Combining our theory with experimental work from collaborating groups, we provide new insights into different aspects of cell mechanics.

Das stäbchenförmige Bakterium Escherichia coli teilt sich in zwei gleich große Tochterzellen. Dies ist nur möglich, wenn sich die Zelle in der Mitte teilt. Bei E. coli wird die Zellteilung durch den Zusammenschluss der FtsZ-Proteine an der Membran zum Z-Ring eingeleitet. Topologische Regulierung des Z-Ringes erfolgt durch räumlich-zeitliche Oszillationen von Min-Proteinen zwischen den beiden Zellpolen. MinC, MinD und MinE binden an und lösen sich von der Membran unter Hydrolyse von ATP und in antagonistischer Art und Weise, was zu einer alternierenden Ansammlung von MinC und MinD an den Zellpolen führt. Gemittelt über die Zeit ergibt sich somit ein MinD-Verteilungsprofil, das maximale Konzentration an den Zellpolen und ein Minimum in der Zellmitte aufweist. MinC bindet an MinD und folgt somit seiner Verteilung. Der Zusammenschluss von FtsZ-Proteinen wird durch MinC unterbunden, und somit kann sich der Z-ring nur an einer Position herausbilden, die ein Minimum an MinC aufweist - der Zellmitte. Das Min-system wurde in der Vergangenheit auch mit einem in-vitro-Ansatz untersucht, indem Min-Proteine in künstliche, aufliegende Lipiddoppelschichten (supported lipid bilayers, SLB) rekonstitutiert wurden. Dabei bildeten die Min-Proteine kein oszillierendes Muster aus, sondern organisierten sich vielmehr in parallelen und propagierenden Wellen (Loose, 2008, Science, 320). In diesen in-vitro-Experimenten war das Membransubstrat wesentlich größer als die Wellenlänge der Min-Proteinwellen. In vivo hingegen ist die Länge der Zelle in der gleichen Größenordnung wie die charakteristische Länge des Oszillationsmusters der Min-Proteine. Daher war es das Ziel dieser Arbeit, den Einfluß einer beschränkten Fläche und geometrischer Formgebung der künstlichen Lipiddoppelschichten auf die Wellenpropagation der Min-Protein zu untersuchen. Flächige Beschränkung künstlicher Membranen erfolgte durch Mikrostrukturtechnologie. Deckglässchen wurden mit einer Goldschicht und mikroskopischen Aussparungen unterschiedlicher geometrischer Formen strukturiert. Funktionale SLBs bildeten sich nur auf Glasflächen ohne Goldbeschichtung aus. Nach der Rekonstitution der Min-Proteine, organisierten sich diese auf den Membranstücken in parallele Wellen. Dabei bestimmte die flächige Beschränkung der künstlichen Membranen die Ausbreitungsrichtung der Min-Proteinwellen. Min-Proteinwellen konnten entlang gekrümmter Membranstreifen, in Ring- und sogar in Slalomstrukturen geleitet werden. In geraden, länglichen Strukturen richteten sich die Wellen entlang der längsten Achse aus. Kopplung von Proteinwellen auf räumlich getrennten Membranstücken in Abhängigkeit des Abstandes und des sogenannten Molecular Crowdings in der wässrigen Lösung konnte ebenfalls beobachtet werden. Diese Kopplung ist ein Indiz für inhomogene Proteinverteilungen in der Lösung oberhalb der Membran. Desweiteren konnten Min-Proteinwellen auch in diversen dreidimensionalen künstlichen Membranen rekonstitituiert werden. Im Wildtyp von E. coli ähneln die Min-Proteindynamiken der einer Oszillation mit einer charakteristischen Länge von 5 µm. Auf SLBs, bilden Min-Proteine Wellen mit einer Wellenlänge aus, die ca. zehnmal größer ist als in vivo. Dieser Unterschied zwischen der in-vivo- und der in-vitro-Welt wurde untersucht und diskutiert. In vitro konnte die Wellenlänge um 50 % durch Erhöhung des Molecular Crowding in der Lösung sowie um 33 % durch Temperaturerhöhung verkleinert werden. Das oszillierende Muster könnte dahingegen eine Folge der Kompartimentierung sein. Erste Versuche, das Min-System in geschlossene Membrankompartimente zu rekonstitutieren, wurden getestet. / Escherichia coli, a rod-like bacterium, divides by binary fission. Cell division into two daughter cells of equal size requires that fission takes place at a midcell position. In E. coli, cell division is initiated by assembly of the FtsZ-proteins at the inner membrane to the Z-ring. Topological regulation of the Z-ring is achieved by spatiotemporal pole-to-pole oscillations of Min-proteins. MinC, MinD and MinE bind to and detach from - under hydrolysis of ATP - the membrane in an antagonistic manner leading to an alternating accumulation of MinC and MinD at the cell poles. Averaged over time, the distribution profile of MinD exhibits maximal concentration at the cell poles and a minimum at the cell center. MinC binds to MinD and thus follows its distribution. FtsZ assembly is inhibited by MinC and therefore the Z-ring can only form at a cell position low in MinC - at the cell center. In the past, the Min-system was also investigated in an in vitro approach by reconstitution of Min-proteins into a supported lipid bilayer (SLB). Here, Min-proteins did not self-organize into an oscillatory pattern but into parallel and propagating waves (Loose, 2008, Science, 320). In this in vitro assay, the membrane substrate was infinitely large compared to the wavelength. However, in vivo, the cell length is on the same order of magnitude as the respective length scale of the oscillatory pattern of Min-proteins. Therefore, we wished to investigate the effect of lateral confinement and geometric structuring of artificial lipid bilayers on the Min-protein wave propagation. Lateral confinement of artificial membranes was achieved by microfabrication technology. Glass slides were patterned by a gold coating with microscopic windows of different geometries, and functional SLBs were only formed on uncoated areas. Upon reconstitution, Min-proteins organized into parallel waves on the geometric membrane patches. Confinement of the artificial membranes determined the direction of propagation of Min-protein waves. Min-protein waves could be guided along curved membrane stripes, in rings and even along slalom-geometries. In elongated membrane structures, the protein waves always propagate along the longest axis. Coupling of protein waves across spatially separated membrane patches was observed, dependent on gap size and level of molecular crowding of the aqueous media above the bilayer. This indicates the existence of an inhomogeneous and dynamic protein gradient in the solution above the membrane. Furthermore, reconstitution of Min-protein waves in various three-dimensional artificial membranes was achieved. In wild-type E. coli, Min-protein dynamics resemble that of an oscillation with a characteristic length scale of 5 µm. On supported lipid bilayers, Min-proteins self-organize into waves with a wavelength approximately 10-fold larger than in vivo. These discrepancies between the in vivo and in vitro world were investigated and discussed. In vitro, the wavelength could be decreased by a factor of 50 % by increase of the molecular crowding in solution and by 33 % through temperature increase. The oscillatory pattern is thought to be a consequence of compartmentalization and first attempts to encapsulate the Min-system in closed bilayer compartments are presented.