Figure 6. Working model: Local control of Piezo2 function by interdependent actions of Mtmr2 and PI(3,5)P₂.

Mtmr2 controls the abundance of PI(3,5)P₂ by dephosphorylation (please see Figure 4a). Mtmr2 and Piezo2 expression as well as PI(3,5)P₂ might be compartmentalized in membrane microdomains. Piezo2 localization in Mtmr2-negative microdomains would facilitate its access to local PI(3,5)P₂ and consequently potentiate Piezo2 RA-MA currents (left side). On the other hand, high Mtmr2 levels and its localization in the proximity of Piezo2 would augment PI(3,5)P₂ turnover, thereby decreasing local PI(3,5)P₂ availability and suppressing Piezo2 RA-MA currents (right side). One could further speculate that Mtmr2, via binding Piezo2, might recruit Piezo2 to membrane microdomains depleted of PI(3,5)P₂. This would provide an active mechanism to inhibit Piezo2 RA-MA currents in membrane compartments – may they be at the plasma membrane or intracellular membranes. Ultimately, Mtmr2 and PI(3,5)P₂ may contribute to dynamically tuning touch sensitivity of an organism in response to diverse conditions modulating Mtmr2 and PI(3,5)P₂ levels (e.g. osmotic stress as indicated by results shown in Figure 4).The following questions await further clarification: (i) How are Mtmr2/PI(3,5)P₂ regulated during (patho)physiological conditions in the somatosensory system, (ii) does the modulation of RA-MA currents require additional yet to be identified effector proteins, and (iii) are other PIPs also involved, for example PI(4,5)P₂ shown to modulate Piezo2 function (Borbiro et al., 2015)? As the structure of Piezo2 has not been resolved yet, Piezo2 is depicted after the recently solved structure of Piezo1 (Ge et al., 2015; Guo and MacKinnon, 2017; Saotome et al., 2018). If this structure holds also true for Piezo2, the PI(3,5)P₂ binding domain (depicted in green) would roughly be localized within the first third of the N-terminal blade.