Mechanosensing and Symmetry of Potassium Channels Studied by Molecular Dynamics Simulations

Potassium channels are relevant in a variety of physiological processes in all three kingdoms of life. In mammals these processes including nerve signaling, heartbeat regulation, and osmotic regulation of cells. Thus a malfunction of these channels is the cause for multiple diseases making potassium channels an important target for drug design. Some of these potassium channels are mechanogated. This particular group and their gating mechanism, as well as quantification of asymmetry in homomultimeric proteins, are the focus of this thesis. Mechanogating is the response of channels to membrane stress through bending and stretching and is involved in many physiological processes including the sensation of touch, hearing, and pain perception. Crystal structures of mechanogated potassium channels can, in some cases, be directly interpreted as open (conductive) or closed (non conductive) states and the gating mechanism can directly be inferred. However, other channels of this group display different states as observed by crystallographic structure determination, but their configuration does not permit insight to the underlying gating mechanism. Two of these channels are the mechanogated two-pore domain potassium channels TREK-2 and TRAAK which are studied here. To elucidate their gating and mechanogating mechanism in particular, we applied molecular dynamics (MD) simulations, allowing insight down to atomistic scale. First we studied the hypothesis of the lipid block based gating mechanism, that suggests a lipid entering the channel blocking the ion permeation pathway. Our results challenge this mechanism as we find ion permeation even in the presence of the lipid that is proposed to be responsible for the lipid block. We suggest a gate located in the narrowest region of the channel called the selectivity filter. The open probability of this gate is influenced by the states of the channel. Two states are known from crystallographic structure determination termed ’up’ and ’down’. We find the up state increasing the open probability compared to the down state. As our main interest is on mechanogating of these channels, we varied the membrane tension to trigger this mechanism. Corresponding MD simulations display a transition towards the up state at increased membrane tension and by this the open probability is increased as corroborated by experimental results. Concluding this research we formulated an overall model by which the channel states as known from crystal structures influence selectivity filter stability and are influenced by membrane tension. Second we studied asymmetry of homomultimeric proteins of which potassium channels are a typical representative. Many computational analysis tools implicitly use ensemble averages to determine protein motions, e.g. Principal Component Analysis, and thus neglect asymmetry which can be important for protein function. Therefore, a solid understanding of asymmetry is required to correct for the neglect. A first step towards this aim is asymmetry quantification. To do so, we developed two algorithms, one evaluating the overall asymmetry and a second one restricted to a functional reference motion. We applied the algorithms to modified potassium channels TREK-2 and KcsA to demonstrate their ability to correctly quantify asymmetry. Furthermore, we applied it to the unfolding process of TTR demonstrating their ability to handle even complex functional properties. Both algorithms can therefore be used to quantify the asymmetry of homomultimeric proteins. In this thesis we provide evidence for a selectivity filter based gating mechanism and demonstrate how this mechanism can be influenced by membrane tension. Furthermore, we provide two algorithms able to quantify asymmetry and demonstrate their usage