Dynamic Clustering Regulates Activity of Mechanosensitive Membrane Channels

Experiments have suggested that bacterial mechanosensitive channels separate into 2D clusters, the role of which is unclear. By developing a coarse-grained computer model we find that clustering promotes the channel closure, which is highly dependent on the channel concentration and membrane stress. This behaviour yields a tightly regulated gating system, whereby at high tensions channels gate individually, and at lower tensions the channels spontaneously aggregate and inactivate. We implement this positive feedback into the model for cell volume regulation, and find that the channel clustering protects the cell against excessive loss of cytoplasmic content.

Figure 1

Single channel properties. (a) MSC is modelled as five rods connected by springs. Each rod consists of hydrophobic (gray) and hydrophilic heads (cyan) and is $\sim 10\mathrm{nm}$ long. Channel are lined with hydrophilic beads (silver). Explicit inter-channel attractions can be turned on via a hydrophobic patch (dark blue). Under osmotic shock the MSC naturally opens. (b) Pore size of a single MSC. The vertical red line marks the application of an instantaneous osmotic shock yielding membrane tension of $\gamma =1.2\mathrm{mN}/\mathrm{m}$. The solid red line is the moving time average (window: ${10}^5$ time steps). (c) The variation of the pore size and solute flux though the channel versus membrane tension.

Figure 2

Multiple channels interacting only via membrane-mediated interactions gate independently. The probability of channel pore opening versus the number of channels present in the system. The error bars correspond to the standard deviation of five different runs. Inset: total flux of solutes through the channels versus the number of channels in the system.

Figure 3

Channels interacting via explicit attractive interactions gate cooperatively. (a) Average pore size of two MSCs as a function of distance between them ($\gamma =1.30\mathrm{mN}/\mathrm{m}$). (b) Opening probability per channel versus cluster size ($\gamma =1.70\mathrm{mN}/\mathrm{m}$, MSC fraction is 0.16, ${ϵ}_{\mathrm{protein}-\mathrm{protein}}=0.9\mathrm{kT}$). The dashed line is a guide to the eye. Inset: The average pore size within a single cluster of 12 MSCs. (c) Upon osmotic shock the average distance between the channels in a cluster increases and is larger for higher membrane tension (orange: $\gamma =1.70\mathrm{mN}/\mathrm{m}$, blue: $\gamma =0.70\mathrm{mN}/\mathrm{m}$). The shading shows a 95% confidence interval of five simulation runs. Far right: Cluster configuration in the last time frame. Inset: MSC opening increases as they disperse.

Figure 4

Clustering regulates channel closing to prevent leaky cell membrane. (a) Normalized cell volume $V_n$ (gray line) at 0.96 Osmol hypoosmotic shock and 0.5% channel packing fraction. Color bars: probability of finding an isolated channel (blue) and a cluster of five channels (orange). (b) Effect of increasing the extent of clustering (here by changing interprotein attractions) on cell volume dynamics (grey and black) in comparison to dispersed channels (red). Conditions as in (a). (c) The average MSC pore size at zero tension for 12 MSCs in a dispersed (blue) and aggregated (orange) states. Red and green lines show the moving time averages (${10}^5$ steps).