Two Segments of the Human PIEZO1 Mechanosensitive Ion Channel Can Reassemble into a Functional Unit

Abstract

We created clones that covalently linked a fluorescent protein to PIEZO1 in order to image PIEZO1 on the cell surface. Inserting mCherry1 in internal positions of the PIEZO1 protein, we found a construct at position 1591 that yielded a protein whose channel kinetics were nearly identical to wild type. Through cotransfection, we imaged PIEZO1 (mCherry) and TREK-1 (GFP) on HEK cells. These two mechanosensitive channels were at the cell surface in completely different physical domains. PIEZO1 appeared as random clusters but TREK tended to follow cytoskeleton tracks.Given the stability of the PEIZO1 protein with a fluorescent group at 1591, we asked if the PIEZO1 protein could be split at position 1591 and still function. Fragment 1 (1-1591) was fused to mCherry, and fragment 2 (1592-2520) to GFP. Co-expression of the two segments showed currents in whole cell recording, cell-attached patches, and outside-out patches. The biophysical characteristics were nearly identical to the wild type channel, having a reversal potential around 0 mV and voltage dependent inactivation in whole cell and cell attached mode.We next asked if either fragment alone was functional. The C-terminal fragment produced mechanosensitive currents in outside-out patches but the N terminal segment (1-1591) did not in any recording mode. We then shortened the C terminal fragment from its C- and N- termini using the outside-out patches to screen for current. We obtained mechanosensitive currents from a fragment of ∼ 450 amino acids, so the fundamental properties of the giant PIEZO1 channel are contained in this small fragment. This protein segment appears to contain the mechanosensor and the gate, and establishes that eight putative transmembrane domains are required for pore formation. None of the shortened fragments displayed whole cell currents.