- Modeling Pattern Formation in Biopolymer Systems induced by Reaction Kinetics and Molecular Motors (2006)
- In my thesis I studied pattern formation in nonequilibrium (NE) polymer systems motivated from cell biology. Actin and microtubules (MTs) can be met in a state of continuous de-/polymerization (D/P), which is used by the cell e.g. during locomotion. This is a NE state since the polymerization is actively coupled to ATP or GTP hydrolysis. A second NE state of biological relevance is caused by motor proteins. These are mobile crosslinkers that walk on the filaments whereby creating forces and reorienting or transporting the latter. The cell displays filament-related ordered structures like aster patterns in the mitotic spindle, bundles in actin stress fibers and also oscillating structures e.g. in muscle bundles. The question is to what extent these structures inside the cell are governed by the physics of active polymers. In part I of my work, we proposed a pattern forming mechanism in a filament solution at high density that is subject to a D/P state. Since actin and MTs are rod-like objects, at high filament concentration a transition to lyotropic nematic order occurs. This transition is first order and thus accompanied by a phase separation. In the absence of D/P kinetics, the solution will thus tend to decompose into an isotropic domain with low density and a domain of high density and nematic order, i.e. the filaments preferentially aligned in one direction. To highlight that the D/P process interplays with this transition, we assumed that filaments are generated and decaying with some specific rates, implying a finite lifetime for the filaments. Accordingly the latter can only diffuse a finite length during their lifetime, which competes with the tendency of the system to phase separate and gives rise to a finite wavelength instability towards a pattern with alternating isotropic and nematic regions with a wavelength of the order of 10 microns. The model developed to describe these patterns is also interesting since it allows a feasible linear stability analysis of the homogeneous nematic state. Part II is devoted to the NE interaction of motor proteins with the filaments. As the starting point of our modeling efforts we chose a mesoscopic approach, namely a Smoluchowski equation which can be coarse-grained to obtain equations for the density and the orientation of the filaments. The main difference to a passive solution of rods are active motor-mediated currents caused by a motor density assumed sufficiently high and homogeneously distributed. These active contributions can be determined to leading order, introducing phenomenological motor transport rates containing details like active motor density, duty ratio, etc. After a thorough linear analysis of the model we obtained a rich instability diagram with an orientational finite wavelength instability which is either stationary or oscillatory and a demixing instability similar to spinodal decomposition but also motor-mediated. The finite wavelength instability has been analyzed by perturbative techniques and numerical simulations of the model equations. In the stationary case, we calculated the existence and stability regions of stripes and squares, which could be related to bundle-like structures and regular lattices of asters respectively. In the oscillatory case, there is competition between traveling and standing waves in one dimension and between traveling and alternating waves in two dimensions, the latter being a four mode solution built from two standing waves in perpendicular directions with a phase shift of 90 degrees. The long-wavelength demixing instability has also been investigated, showing coarsening aster-like structures. Experiments on MT-motor solutions display dissipative patterns in the NE state. Recent experiments on actin filaments and myosin oligomers show a rather different behavior, namely cluster patterns do not appear until ATP is nearly depleted. We proposed two mechanisms to explain these patterns: first, motors lacking ATP form rigor bonds with actin inducing small bundles, which through a combination of reduced diffusivity and enhanced interaction cross-section can be transported more efficiently, allowing the system to cross one of the instabilities discussed above. A second important feature is the presence of crosslinking proteins in the experiments. We propose that these can be interpreted as a parametric disorder. Assuming in the model a random contribution to the active current, a Ginzburg-Landau equation with multiplicative stationary noise could be derived leading to a threshold reduction. To conclude, it seems to be fruitful to apply and combine methods from statistical physics and pattern formation to NE problems in cell biology to foster the understanding of actively polymerizing filament and motor proteins in their different NE states.