2025-08-29_144124_nature_nature25453

Structure of the mechanically activated ion channel Piezo1 | Nature

URL: https://www.nature.com/articles/nature25453

Saved: 2025-08-29T21:41:29.527Z

Article Published: 20 December 2017 Structure of the mechanically activated ion channel Piezo1 Kei Saotome, Swetha E. Murthy, Jennifer M. Kefauver, Tess Whitwam, Ardem Patapoutian & Andrew B. Ward

Nature volume 554, pages 481–486 (2018)Cite this article

40k Accesses

546 Citations

123 Altmetric

Metrics details

An Author Correction to this article was published on 04 July 2022

This article has been updated

Abstract

Piezo1 and Piezo2 are mechanically activated ion channels that mediate touch perception, proprioception and vascular development. Piezo proteins are distinct from other ion channels and their structure remains poorly defined, which impedes detailed study of their gating and ion permeation properties. Here we report a high-resolution cryo-electron microscopy structure of the mouse Piezo1 trimer. The detergent-solubilized complex adopts a three-bladed propeller shape with a curved transmembrane region containing at least 26 transmembrane helices per protomer. The flexible propeller blades can adopt distinct conformations, and consist of a series of four-transmembrane helical bundles that we term Piezo repeats. Carboxy-terminal domains line the central ion pore, and the channel is closed by constrictions in the cytosol. A kinked helical beam and anchor domain link the Piezo repeats to the pore, and are poised to control gating allosterically. The structure provides a foundation to dissect further how Piezo channels are regulated by mechanical force.

You have full access to this article via your institution.

Download PDF Similar content being viewed by others Crowding-induced opening of the mechanosensitive Piezo1 channel in silico Article Open access 19 January 2021 Intrinsically disordered intracellular domains control key features of the mechanically-gated ion channel PIEZO2 Article Open access 15 March 2022 Optical control of PIEZO1 channels Article Open access 07 March 2023 Main

The conversion of mechanical force into cellular signals, or mechanotransduction, is an essential biological function that is conserved throughout evolution. Mechanically activated channels confer force sensitivity to cells and organisms by allowing the passage of ions across the membrane in response to a mechanical stimulus1. Piezo1 and Piezo2 were identified in 2010 as a conserved class of eukaryotic mechanically activated ion channels2,3 that are involved in various mechanotransduction pathways, including touch sensation4, proprioception5, nociception6, vascular development7 and breathing8. In humans, gain-of-function (GOF) mutations in Piezo1 are linked to dehydrated hereditary stomatocytosis9,10,11,12, a disease in which red blood cells are dehydrated and have increased permeability to cations, and GOF mutations in Piezo2 are associated with distal arthrogryposis13. Loss-of-function mutations in Piezo1 and Piezo2 are linked to congenital lymphatic dysplasia14 and defective touch perception and proprioception15,16,17, respectively. Piezo proteins are structurally unrelated to other proteins, and are uniquely large compared to other ion channels (more than 2,000 amino acids). Electrophysiological studies have demonstrated that Piezo1 ion channels are gated directly by changes in membrane tension18,19,20, but the structural bases of mechanosensitivity and ion conduction in Piezo channels remain unknown. Previous crystallographic21,22 and medium-resolution cryo-electron microscopy (cryo-EM)22 studies have suggested the overall architecture of Piezo proteins. However, a mechanistic understanding of Piezo function requires high-resolution structural information of the channel.

Cryo-EM sample and structure determination

A previous cryo-EM reconstruction of mouse Piezo122 lacked amino acid side-chain definitions and connectivity of resolved regions, precluding building of an accurate structural model. Here, we optimized Piezo1 sample preparation parameters to make Piezo1 amenable to high-resolution single particle cryo-EM structure determination (Extended Data Fig. 1). After computationally sorting particle heterogeneity (Extended Data Fig. 2a), we identified a subset of particles that were refined into both asymmetric and C3-symmetry-imposed maps of the mouse Piezo1 trimer, with overall resolutions of 4.5 and 3.9?Å, respectively. We conducted masked refinement of the central ‘core’ (Extended Data Fig. 2b–d) and C3-symmetry expansion followed by masked classification and refinement (Extended Data Fig. 3) to improve densities proximal to the central three-fold axis and the protein periphery, respectively. Both the 3.8?Å resolution core-masked map and a 4.1?Å resolution symmetry-expanded map (blade class 1) were used to build and refine a molecular model of mouse Piezo1 that spans structured regions of residues 1134–2544, including 14 transmembrane (TM) helices per protomer and other conserved C-terminal domains (Extended Data Fig. 4, Extended Data Table 1, Supplementary Data 1).

Piezo1 core architecture

The homotrimeric Piezo1 complex exhibits a three-blade propeller shape with an extracellular ‘cap’ domain at the top of the central axis (Fig. 1), consistent with the previous medium-resolution cryo-EM reconstruction22 (Extended Data Fig. 5a–c). The transmembrane region can be approximated by the weak density corresponding to the detergent micelle (Fig. 1a, b), in agreement with the hydrophobicity of the membrane-exposed amino acids in our model (Fig. 1c). Interestingly, the transmembrane region has a pronounced bend, suggesting Piezo1-induced local distortion of the membrane. Although we cannot rule out the possibility that the bending is a product of detergent solubilization, protein-induced membrane curvature has also been proposed for mechanosensitive channels of the two-pore potassium23 and bacterial MscS24,25 families.

Figure 1: Architecture and domain arrangement of the Piezo1 core.

a, b, Side (a) and top (b) views of the Piezo1 cryo-EM map refined with C3 symmetry imposed. Density corresponding to modelled regions is coloured green. Less resolved density that was not modelled is coloured yellow. At lower thresholds (transparent grey map), scattered density probably originating from a detergent micelle can be observed. c, Surface electrostatics of the Piezo1 core model. Note the bent hydrophobic stripe representing the transmembrane region. Dotted lines approximate the extracellular (black) and cytosolic (yellow) membrane boundaries. d, e, Side (d) and top (e) views of the Piezo1 core in cartoon representation, with each domain coloured differently. In e, the cap domain is removed to highlight domain swapping between Piezo repeat A and the outer helix. f, Schematic of the Piezo1 core domain arrangement. Dotted lines represent flexible regions that were not clearly resolved in the density maps. TM, transmembrane domain; rep., repeat; C-term, C terminus.

PowerPoint slide

Full size image

The Piezo1 core (Fig. 1d–f) defines an architecture that is unique from other membrane proteins of known structure. The inner and outer transmembrane helices lie closest to the central axis, and are intervened by the extracellular cap, the structure of which was determined previously by X-ray crystallography21,22. In isolation, the trimeric arrangement of inner and outer helices and cap domains bear structural resemblance to acid-sensing ion channels26 and P2X receptors27, as noted previously28. The 12 transmembrane helices peripheral to the inner and outer transmembrane helices can be grouped into three four-transmembrane-helix-containing bundles, each with the same topology, that we term Piezo repeats A, B and C, according to proximity to the central axis. Piezo repeat A of one protomer is adjacent to the outer helix of a different protomer, producing a ‘domain-swapped’ arrangement analogous to some tetrameric ion channels29,30 (Fig. 1e). The ‘anchor’ domain forms a triangular three-helix wedge between the inner and outer transmembrane helices and Piezo repeat A. The C-terminal domain (CTD) connects to the intracellular end of the inner helices, collaring the central axis. The intracellular ‘beam’, is an approximately 64-residue (~95?Å) kinked helix that runs nearly parallel to the membrane, directly underneath Piezo repeats A to C. Although the N terminus of each beam is tethered directly to Piezo repeat C, its C terminus projects towards the central axis and is connected via a roughly 37-residue disordered linker to a short, coiled peptide that we term the ‘latch’, which contacts the intracellular face of the CTDs. The latch then connects back to Piezo repeat B via a relatively non-conserved 240-residue stretch of peptide (Supplementary Data 1) that is not resolved in our maps. The beam and latch are positioned to transmit conformational changes from the propeller blade periphery to the central ion channel. The beam may also be influenced by a nearby domain that we term the ‘clasp’ (Fig. 1d, f), which was modelled without amino acid assignment owing to lower resolution and discontinuity with other domains.

Flexible blades consist of at least six Piezo repeats

Blade-focused classification and refinement of symmetry-expanded particles yielded two classes (blade class 1 and 2) with improved density at the periphery (Extended Data Fig. 3). Remarkably, in each of these maps, clear density for an additional three Piezo repeats is resolved (Fig. 2a, b). Thus, a total of six Piezo repeats (which we have named A–F ordered by proximity to the central axis), or 26 transmembrane helices per protomer (including the inner and outer helices) are apparent, in disagreement with a previous cryo-EM study22, which proposed 14 transmembrane helices. We note the possibility of additional transmembrane helices, and potentially Piezo repeats, in the Piezo1 protein that were not resolved in our maps, but have been predicted bioinformatically31.

Figure 2: Propeller blade composition and conformational heterogeneity.

a, b, Side (a) and top (b) views of blade class 1 cryo-EM map, with individual domains segmented and coloured uniquely. In b, the map is sliced along the dotted line in a to highlight the positions of the transmembrane helices. c, Schematic of the Piezo repeat topology, which includes a cytosolic membrane parallel helix and four transmembrane helices. d, Aligned models for Piezo repeats A (yellow), B (cyan) and C (pink). e, Sagittal slice of electron density corresponding to the transmembrane region of blade class 1 (blue) and blade class 2 (orange). Approximate trajectory of the beam is shown as pink dotted line. Individual Piezo repeats are circled by black dotted lines and labelled A–F. Red double arrow depicts direction of conformational change between the two classes. f, Superposed cryo-EM maps and cartoon models corresponding to beam domains of blade class 1 (blue) and blade class 2 (orange). Cα positions of the conserved Ala-Ser-Arg-Gly motif at the beam pivot are shown as spheres.

PowerPoint slide

Full size image

Although we were only able to place amino acids confidently for Piezo repeats A–C, the topology of the peripheral three repeats is resolved and indicates that all six have the same overall structure and arrangement of transmembrane helices (Fig. 2c, d). The termini of each repeat are cytoplasmic, with a pair of short linkers connecting the N-terminal transmembrane helices, and an extended extracellular linker connecting the third and fourth transmembrane helices. A notable feature of the Piezo repeats is the presence of cytoplasmic membrane-parallel helices tethered to their N termini. In addition to the Piezo repeats, the clasp domain also contributes a membrane-parallel helix, adjacent to that of Piezo repeat B (Fig. 2b). Amino acid assignment for the helices in Piezo repeats A, B and C demonstrates their amphipathic nature (Extended Data Fig. 6a–d). Interestingly, amphipathic helices are thought to sense or induce membrane curvature in various proteins32, and have been proposed to be crucial for membrane tension-induced gating in various classes of mechanically activated channels33. That the amphipathic helices dictate local membrane deformation, and perhaps mechanosensitivity of Piezo1 is an intriguing possibility. We speculate the array of amphipathic helices in each blade may work together to invoke a local membrane environment with continued curvature, thereby contributing to the puckered architecture.

Our blade-focused classification of the symmetry-expanded particles yielded two classes with distinct conformations demonstrating conformational flexibility of the blade. While the central regions are relatively static, the peripheral regions undergo horizontal movement in the plane of the membrane, not vertically in and out of the membrane (Fig. 2e, Extended Data Fig. 5f, g). Interestingly, associated with this lateral flexibility is an approximately 5° pivoting of the beam at a kink centred on a highly conserved Ala-Ser-Arg-Gly motif (A1316–G1319) (Fig. 2f, Supplementary Data 1). This pivot may be an important structural feature that regulates how membrane-tension-induced conformational changes at the periphery are transmitted through the beam to the central gating machinery.

Ion-conducting pore

The extracellular cap domain, inner helix and cytosolic CTD line the central pore axis of Piezo1. Extracellular cations could approach the pore entryway vertically through the internal cavity along the three-fold axis of the cap domain, or laterally through gaps between flexible linkers connecting the cap and inner and outer helices (Fig. 3a). The cap domain, which is flexibly linked to the transmembrane domain, might have a role in mediating ion accessibility to the pore, as suggested previously28. Indeed, a comparison of our current structure to the previous Piezo1 structure indicates distinct orientations of the cap domain and inner and outer helices (Extended Data Fig. 5d, e), illustrating the flexibility of the cap and the associated transmembrane helices.

Figure 3: Ion pore structure and electrophysiological characterization of M2493A and F2494A.

a, Cartoon model of Piezo1, with the inner helix and CTD coloured blue and purple, respectively. Flexible linkers between the cap domain and the transmembrane region are depicted as dotted lines. Green spheres represent ions, with possible entryways into the pore shown as black arrows. b, Surface representation of the transmembrane region of Piezo1 surrounding the central axis, highlighting lateral membrane openings to the pore pathway. c, van der Waals radii of the pore, plotted against distance along the pore axis. The radial distance between the pore axis and the protein surface is shown as a cyan surface in a, b and d. d, Expanded view of the Piezo1 pore, with residues lining the pore pathway shown in yellow and labelled. Electron density corresponding to a ‘cytosolic plug’ is shown as a yellow mesh. e, f, Expanded top-down views of constrictions formed by M2493 and F2494 (e) and P2536 and E2537 (f). Residues mentioned in the text are also labelled. g, Left, representative traces of probe displacement and mechanically activated whole-cell currents recorded (−80?mV) from one cell expressing wild-type, M2493A or F2494A mouse Piezo1 channels. Right, inactivation time constant for individual cells across different conditions (Piezo1 n?=?7 cells, M2493A n?=?8 cells and F2494A n?=?8 cells). ***P?=?0.0001, one-way analysis of variance (ANOVA) with Dunn’s multiple comparison test. h, Left, representative traces of stretch-activated macroscopic currents recorded (−80?mV) in response to −60?mm Hg pressure from one cell expressing wild-type, M2493A or F2494A Piezo1 channels. Right, inactivation time constant (at −60?mm Hg) for individual cells across different conditions. *P?=?0.008, one-way ANOVA with Dunn’s multiple comparison test. i, Percentage steady state current (at −60?mm Hg) for individual cells across different conditions. *P?=?0.012, one-way ANOVA with Dunn’s multiple comparison test. In h and i, Piezo1 n?=?7 cells, M2493A n?=?8 cells and F2494A n?=?8 cells. Data in g��i denote mean?±?s.e.m.

PowerPoint slide

Full size image

Below the cap, the ion conduction pathway is lined by the three inner transmembrane helices. Notably, we find two lateral openings between the inner helices separated by a ‘seal’ formed by K2479 and F2480 (Fig. 3b). We term these openings the upper crevice (approximately 11?Å wide and 16?Å tall) and lower fenestration (10?Å wide and 12?Å tall), which may provide pore access for lipids or other hydrophobic molecules as a means to regulate ion conduction, reminiscent of the proposed force-gating mechanism of mechanically activated two-pore potassium channels23. Molecular dynamics simulations of the homotrimeric ATP-gated P2X3 receptor34 revealed that a lower fenestration provides a passageway for ion flow into the cytoplasm. Although the possibility that the Piezo1 lower fenestration represents an ion permeation pathway warrants further study, we assume that, like most other ion channels, the pore is coincident with the central vertical axis. The pore-facing residues of the inner helices are mainly hydrophobic (Fig. 3d). The 35 Å-long membrane-spanning segment of the pore is relatively wide (more than 2?Å van der Waals radius) and does not house the narrowest constrictions of the conduction pathway (Fig. 3c, d). Instead, the ion conduction pathway narrows directly below the lower membrane boundary into a cytosolic ~11?Å long ‘neck’ (with a minimum van der Waals radius less than 1?Å) that defines the closed state of the pore in the current structure.

At the top of the cytosolic neck, the side chains of M2493 and F2494, which are conserved across Piezo orthologues (Supplementary Data 1), form a hydrophobic barrier (Fig. 3e). To examine the functional role of the apparent hydrophobic barrier, we recorded mechanically activated currents (Fig. 3g, h) from Piezo1-knockout HEK293T cells14 expressing the alanine substitution mutants M2493A, F2494A and M2493A/F2494A. The double mutant M2493A/F2494A resulted in a non-functional channel despite proper trimeric assembly and surface expression (Extended Data Fig. 7a–c), demonstrating the importance of these residues. Both of the single mutants expressed as functional channels, with F2494A having significantly slower inactivation kinetics than wild-type Piezo for mechanically activated macroscopic currents in cell-attached and whole-cell configurations (Fig. 3g, h). Furthermore, stretch-activated currents from F2494A had greater steady state currents relative to wild type (Fig. 3i, Extended Data Table 2). The increased channel activity of F2494A in these recordings is consistent with this residue having a role in gating, possibly by stabilizing a non-conducting state of the channel. The M2493A mutant did not show altered permeation properties, whereas the F2494A mutant only exhibited modestly increased single channel conductance (Extended Data Fig. 7d, e) and slightly altered ion selectivity (Extended Data Fig. 7f, g, Extended Data Table 2). Our results implicate the hydrophobic barrier as an important contributor to the gating properties of Piezo1, with a minor role in determining ion conduction and selectivity. In previous studies, mutation of acidic residues (E2495 and E2496) adjacent to M2493/F2494 (Fig. 3e) affected pore properties28, suggesting that they might coordinate permeating cations. In the current structure, these residues do not project into the pore (Fig. 3e), although structural changes associated with gating could result in this region adopting conformation(s) in which the side chains of E2495 and/or E2496 face the pore to form cation-binding site(s).

Directly below the hydrophobic barrier formed by M2493 and F2494, the pore pathway narrows into a second constriction delineated by P2536 and E2537 (Fig. 3f), which are located at the turn of a helix–turn–helix motif in the distal CTD. The side chain of E2537 forms a possible hydrogen bond with Y1412 of the latch domain (Fig. 3f), which may allosterically couple movement of the latch to pore structure in this region. Approximately 4?Å below E2537 a robust density that cannot be reliably modelled but probably originates from the Piezo peptide is present (Fig. 3d). This density, which forms a cytosolic ‘plug’ to the ion channel pore, could also have a role in the permeation pathway.

Gating machinery and lipid pocket

Our structure suggests that activation of Piezo1 may involve concerted movements of the inner helix, outer helix, CTD, anchor domain, latch and beam domains. Interdomain contacts that are likely to govern gating movements allosterically are shown in Fig. 4. The conserved anchor domain is poised to have a prominent involvement in gating through its interactions with the inner helix and CTD (Fig. 4b, c). In particular, anchor domain residue E2133 and R2482 of the adjacent inner helix are probably engaged in an electrostatic interaction. In previous studies, mutations at E2133 affected single-channel conductance and ion selectivity, and sensitivity to the pore blocker ruthenium red34. The equivalent mutation to R2482H in human PIEZO1 results in slower inactivation kinetics and is associated with dehydrated hereditary stomatocytosis10,12. In the context of the present structure, these data implicate the anchor domain–inner helix interaction in regulating gating and pore properties of Piezo1. The cytosolic membrane-lining helix of the anchor domain forms several contacts with the CTD (Fig. 4c), including potential electrostatic interaction(s) between R2169 and E2495/E2496, which are adjacent to pore-facing residues M2493 and F2494 (Fig. 3e). This interaction provides a possible explanation for why mutation of E2495 or E2496 affected the ion conduction properties of Piezo128. Overall, the proximity of the anchor domain membrane-lining helix to the cytosolic neck of the permeation pathway suggests that its movement would be strongly coupled to channel gating.

Figure 4: Interdomain interactions and lipid pocket.

a, Cartoon model of Piezo1, with colour-coded boxes around regions expanded in b–e. In a and d, the red mesh shows electron density corresponding to putative lipid. In a–e, residues that contribute to interdomain contacts (b, c and e) or putative lipid binding (d) are shown as sticks and coloured according to domain.

PowerPoint slide

Full size image

We observe a robust lipid-shaped density in an apparent ligand-binding pocket formed by the anchor domain on one side and Piezo repeat A on the other (Fig. 4d). Two conserved arginine residues (R2035 and R2135) project into the pocket, positioned to stabilize a negatively charged species. Notably, mutation of R2135 to alanine rendered Piezo1 non-functional34, suggesting that this ligand-binding pocket allosterically supports channel activity, and may therefore be a target for amphiphiles and other compounds aimed at modulating Piezo function. In addition, we note the presence of a separate density, potentially representing a lipid headgroup, that is sandwiched between Piezo repeats B and C (Extended Data Fig. 6e, f).

On the cytosolic side, the CTD interacts with the latch and beam (Fig. 4e). Latch residues Y1412, L1414 and F1415 protrude into the surface of the CTD, making contacts with K2528, F2531, E2537 and W2542 (Fig. 4e). Towards the C-terminal end of the beam, a series of basic and polar residues (R1351, R1353, Q1356 and R1360) face acidic CTD residues (D2509, E2518, E2520 and E2524). These electrostatic interactions probably promote allosteric coupling between the beam and CTD. Notably, the equivalent mutation to R1353P, which would presumably break or kink the helical beam at this position, is found in human patients with dehydrated hereditary stomatocytosis and results in delayed inactivation kinetics10. Thus, the distal beam end may be a structural determinant of gating behaviour, possibly through its interactions with the CTD.

Concluding remarks

The Piezo1 structure described here represents the first, to our knowledge, high-resolution view of this functionally and structurally novel class of ion channel. Importantly, most known human PIEZO1 and PIEZO2 GOF disease mutants35,36 can be mapped onto the structure, which indicates that our model encompasses many of the functionally important domains, and moreover, allows a structural interpretation of disease-causing mechanisms (Fig. 5a).

Figure 5: Mapping of disease mutants and schematic diagram of Piezo1 structure and conformational flexibility.

a, Model of the Piezo1 core, with Cα positions of Piezo1 and Piezo2 human GOF disease mutants shown as red spheres. b–d, Schematic diagrams of the various Piezo1 structural elements that contribute to its function as a membrane-tension gated channel. In c and d, bottom-up (c) and top-down (d) views of a single Piezo1 propeller blade are shown, with dotted lines and arrows depicting lateral conformational flexibility, which may be involved in gating.

PowerPoint slide

Full size image

A central question in the study of Piezo channels is how membrane tension results in channel activation. A recent study probing for mechanosensitive domains in Piezo137 implicated regions as disparate as the N terminus (which probably resides towards the outer ‘tip’ of the propeller blade) and the central extracellular cap. Thus, mechanical activation is likely to involve movements of various domains spanning the entirety of the protein complex. Our structure clarifies the features that contribute to this process (Fig. 5b, c). The arrayed Piezo repeats potentially induce a remarkable degree of local membrane deformation, conceivably through their amphipathic helices. Although free energy calculations38 have suggested that the energetic cost of local membrane deformation plays a key role in the mechanosensitivity of MscS25 and TRAAK23 channels, its contribution to Piezo structure and function in membrane bilayers will require further study. Nonetheless, our results allow us to speculate that membrane-tension-induced conformational changes of the Piezo repeats are transmitted through the beam and/or anchor domains to the inner helix and CTD, which in turn determine the activation state of the channel. The nature of these conformational transitions remains unclear, but the present structure of a closed state will guide further investigations of the process of Piezo mechanical activation.

Methods

No statistical methods were used to predetermine sample size. The experiments were not randomized and investigators were not blinded to allocation during experiments and outcome assessment.

Sample preparation

Protocols to purify mouse Piezo1 were adapted from a previous study22 with modifications. Suspension-adapted HEK293F cells (purchased and authenticated from ATCC, tested negative for mycoplasma) were grown at 37?°C to a cell density of approximately 1.5?million per ml and transfected with a pcDNA3.1-IRES-GFP plasmid22 containing mouse Piezo1 (Uniprot KB accession number E2JF22) fused to a Prescission protease cleavage site (LEVLFQGP) followed by a glutathione S-transferase GST coding sequence (this plasmid was a gift from B. Xiao). For transfection, DNA (300?μg) and polyethylenimine (900?μg) were added per litre of HEK cells. Sodium butyrate (4?mM) was added to the cells 12–24?h after transfection. Approximately 48?h after transfection, cells were collected and resuspended in buffer containing 25?mM PIPES pH 7.2, 150?mM NaCl, 1% CHAPS, 0.5% soy phosphatidylcholine, 0.1% C12E9, 2?mM dithiothreitol (DTT), and cOmplete protease inhibitor cocktail (Roche). The lysate was stirred for 1?h at 4?°C, clarified by centrifugation, and incubated with Glutathione Sepharose 4B resin (GE Healthcare) for 3?h. The resin was washed extensively with buffer A (25?mM HEPES pH 8.0, 150?mM NaCl, 100?μM glyco-diosgenin, 2?mM DTT). Prescission protease was added to a ~30% slurry of resin in buffer A at a ratio of approximately 100?μg per ml resin and incubated at 4?°C overnight. The flow-through was concentrated and subjected to size-exclusion chromatography on a Superose 6 increase column equilibrated to buffer A. The peak corresponding to Piezo1 trimer was collected and concentrated to ~5?mg ml−1. Protein (3.5?μl) was applied to a freshly plasma cleaned UltrAuFoil R1.2/1.3 grid, blotted for 3.5?s with a blot force of 0, and plunge-frozen in liquid ethane using a Vitrobot Mark IV (FEI).

Cryo-EM data collection

An initial dataset of mouse Piezo1 was collected in a Talos Arctica electron microscope (FEI) equipped with a K2 direct detector camera (Gatan) operated at 200?kV and having a nominal pixel size of 1.15?Å. Subsequent data were collected on a Titan Krios electron microscope (FEI) equipped with a K2 direct detector camera (Gatan) operated at 300?kV and having a nominal pixel size of 1.03?Å. Micrographs were automatically collected using Leginon software39. The total accumulated dose was ~60?e− Å−2 and the target defocus range was 0.8–2.2?μM. Micrograph movie frames were dose-weighted and aligned using MotionCor240. CTF values were estimated using Gctf41.

Image processing

Template-based particle picking, 2D and 3D classification, and 3D refinements were carried out in Relion 2.042. Unwanted particle picks due to ice contaminations or the grid hole edges were removed using EM Hole Punch43. An initial set of 527 micrographs was collected on the Talos Arctica microscope, from which approximately 96,000 particle images were extracted. After reference-free 2D classification and 3D classification using the published medium-resolution Piezo1 reconstruction22 as an initial reference (low-pass filtered to 60?Å), a ~26,000 particle subset was identified that refined to a ~5?Å resolution reconstruction with C3 symmetry applied.

Two separate data collection sessions were conducted on the Titan Krios microscope. 1,978 micrographs were collected in session 1, yielding approximately 281,000 particles. This particle stack was subjected to 2D classification to remove classes not displaying well-defined features of the Piezo1 trimer, resulting in 266,000 particles. In session 2, 1,086 micrographs were collected, yielding approximately 144,000 particles from which 112,000 particles were chosen after 2D classification. The post-2D classification particle stacks from sessions 1 and 2 were combined into a stack with approximately 378,000 particles. This stack was subjected to 3D classification requesting three classes, using a 60?Å low-pass-filtered map of mouse Piezo1 generated from the Talos Arctica dataset as an initial reference. The best class contained 106,000 particles, which were subjected to a round of 3D classification without alignment requesting six classes. The most populated class from this round of refinement contained 73,000 particles. Auto-refinement of these particles with and without C3 symmetry imposed resulted in reconstructions at 3.9 and 4.5?Å resolution, respectively. C3 symmetry imposed refinement with a soft mask around the Piezo1 core (cap, inner and outer helices, anchor, CTD, Piezo repeats A–C, latch and beam domains) requesting local searches resulted in a 3.8?Å resolution map with improved features.

Inspection of the asymmetrically refined map indicated a progressive break of symmetry towards the blade periphery (Extended Data Fig. 3a). To resolve the peripheral regions of the blade better, relionparticlesymmetry_expand was used to enlarge the particle stack threefold artificially, and add 120° or 240° to the first Euler angle determined in the initial C1 refinement. Alignment-free 3D classification of these approximately 220,000 ‘expanded’ particles requesting six classes with a soft mask around a single blade yielded two highly populated classes containing approximately 102,000 (blade class 1) and 74,000 expanded particles (blade class 2) with distinct blade conformations. Refinement with a soft mask around a single blade and the most symmetric parts of the molecule (cap, anchor, Piezo repeat A, inner and outer helices) and requesting local angular searches resulted in reconstructions at 4.1 (blade class 1) and 4.5?Å (blade class 2) resolution, respectively.

Model building and refinement

The 3.8?Å resolution map of the Piezo1 core sharpened with a B-factor of −118 automatically determined using relion_postprocess was primarily used for model building and refinement. Initially, the crystal structure of the extracellular cap domain22 (PDB code 4RAX) was docked as a rigid body, and a poly-alanine model was fitted into the rest of the Piezo1 core. Rough sequence assignment was ascertained from the position of the docked cap domain, as well as transmembrane topology prediction and previous immunostaining and mass spectrometry data34 demonstrating the positions of extracellular loops and intracellular phosphorylation sites, respectively. We were able to model amino acids for 14 out of the 26 transmembrane helices observed in our electron microscopy maps. We note here that the previous membrane topology analysis34 approximates the presence of 21 transmembrane helices in addition to the 14 we have modelled. The N and C termini of the channel are intracellular34, necessitating an even number of transmembrane helix segments. This would suggest a putative total of 34 or 36 transmembrane helices, and up to 8 Piezo repeats per protomer. Nonetheless, additional studies are needed to clarify the exact membrane topology of Piezo proteins.

Precise sequence registry assignment was determined by locating bulky Phe, Arg, Tyr, and Trp side chains, as well as the absence of side-chain density for Gly. Side-chain densities were not well resolved for many Asp and Glu side chains, probably owing to radiation damage44. Some portions of the beam and Piezo repeats B and C were built using the blade class 1 map (sharpened to an automatically determined B-factor of −152), then symmetrized three-fold to fit into the C3 symmetry core map. Iterative manual model building and automatic refinement were carried out in Coot45, and Phenix.realspacerefine46, respectively. Final structural refinement was carried out in the C3 symmetry core map, with strong restraints applied to portions built using the blade class 1 map, as well as non-crystallographic symmetry restraints for the entire model. Model geometry and fit to map were validated using MolProbity47 and EMRinger48. Electrostatic potential surfaces were calculated using APBS49. Structural figures were made in PyMOL50 or UCSF Chimera51.

Generation of mutants

Mouse Piezo1?mutants M2493A, F2494A and double mutant M2493A/F2494A were generated using QuickChange II XL site-directed mutagenesis kit according to the manufacturer’s instruction and confirmed by full-length DNA sequencing. All mouse Piezo1 constructs were inserted into an IRES-GFP pcDNA3.1 vector.

Cell culture and transfections

Wild-type and mutant mouse Piezo1 were transfected and tested in Piezo1-knockout (P1KO) HEK293T cells. HEK293T-P1KO cells were generated using CRISPR–Cas9 nuclease genome editing technique as described previously14 and not tested for mycoplasma contamination nor further authenticated. Cells were grown in DMEM containing 4.5?mg?ml−1 glucose, 10% fetal bovine serum, 50?U?ml−1 penicillin and 50?μg?ml−1 streptomycin. Cells were plated onto 12-mm round glass poly-D-lysine coated coverslips placed in 24-well plates and transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instruction. All plasmids were transfected at a concentration of 600?ng?ml−1. Cells were recorded from 24 to 36?h after transfection.

Electrophysiology

Patch-clamp experiments were performed in standard whole-cell or cell-attached mode using Axopatch 200B amplifier (Axon Instruments). Macroscopic currents were filtered at 2?kHz and sampled at 20?kHz. Leak currents before mechanical stimulations were subtracted off-line from the current traces. Voltages were not corrected for a liquid junction potential except for ion selectivity experiments. Liquid junction potential was calculated using Clampex 10.3 software. All experiments were done at room temperature.

Mechanical stimulation: For whole-cell recordings, mechanical stimulation was achieved using a fire-polished glass pipette (tip diameter 3–4?μm) positioned at an angle of 80° relative to the cell being recorded. Downward displacement of the probe towards the cell was driven by Clampex-controlled piezoelectric crystal microstage (E625 LVPZT Controller/Amplifier; Physik Instrumente). The probe had a velocity of 1?μm?ms−1 during the ramp phase of the command for forward movement and the stimulus was applied for 150 or 300?ms. To assess the mechanical sensitivity of a cell, a series of mechanical steps in 0.5 or 1?μm increments was applied every 20?s. For I–V relationship recordings, voltage steps were applied 700?ms before mechanical stimulation (150?ms) from a holding potential of −60?mV. Voltage steps were given from −80 mV to +80 in 20?mV increments.

Stretch-activated currents were recorded in the cell-attached patch-clamp configuration. Membrane patches were stimulated with a 5 s prepulse duration of +5 or 0?mm Hg followed by 500?ms negative pressure pulses through the recording electrode using Clampex-controlled pressure clamp HSPC-1 device (ALA-scientific)19. Because the single-channel amplitude is independent of the pressure intensity, the most optimal pressure stimulation was used to elicit responses that allowed single-channel amplitude measurements. These stimulation values were largely dependent on the number of channels in a given patch of the recording cell. Single-channel amplitude at a given potential was measured from trace histograms of 5–10 repeated recordings. Histograms were fitted with Gaussian equations using Clampfit 10.3 software. Single-channel slope conductance for each individual cell was calculated from linear regression curve fit to single-channel I–V plots.

Solutions: For whole-cell patch-clamp recordings, recording electrodes had a resistance of 2–3?MΩ when filled with internal solution composed of (in?mM) 133 CsCl, 1 CaCl2, 1?MgCl2, 5 EGTA, 10 HEPES (pH 7.3 with CsOH), 4?MgATP and 0.4 Na2GTP. The extracellular solution was composed of (in?mM) 133 NaCl, 3 KCl, 2.5 CaCl2, 1?MgCl2, 10 HEPES (pH 7.3 with NaOH) and 10 glucose. For ion-selectivity experiments, internal solution used was (in?mM) 150 CsCl, 10 HEPES (pH 7.3 with CsOH) and extracellular solution consisted of (in?mM) 100 CaCl2 and 10 HEPES (pH 7.3 with CsOH).

For cell-attached patch-clamp recordings, external solution used to zero the membrane potential consisted of (in mM) 140 KCl, 1?MgCl2, 10 glucose and 10 HEPES (pH 7.3 with KOH). Recording pipettes were of 1–3?MΩ resistance when filled with standard solution composed of (in mM) 130?mM NaCl, 5 KCl, 1 CaCl2, 1?MgCl2, 10 TEA-Cl and 10 HEPES (pH 7.3 with NaOH).

Surface immunostaining

Surface immunostaining of Piezo1 was carried out as described previously34, with slight modifications. In brief, pcDNA3.1 IRES-GFP vector encoding wild-type or mutant Piezo1 constructs containing the Myc tag (EQKLISEEDL) inserted at a previously determined extracellular amino acid position 89734 were expressed in HEK293T-P1KO cells on poly-D-lysine-treated glass coverslips. Two days after transfection, labelling of non-permeabilized cells was carried out by incubating the cells with 9E11 anti-Myc antibody (1:50; Santa Cruz Biotechnology). After six washes with warm medium, cells were incubated with secondary antibodies conjugated to Alexa Fluor 568 (1:200; Life Technologies) for 10?min at room temperature. Cells were washed six times with warm medium and once with PBS, and then fixed with 4% paraformaldehyde (PFA) in PBS for 30?min. For permeabilization, cells were fixed with 4% PFA for 10?min and then treated with 0.3% Triton X-100 and blocked with 10% normal goat serum in PBS before incubation with antibodies (primary: 1:200 for 2?h, secondary: 1:400 for 1?h, in block solution). Cells were imaged with a Nikon C2 confocal microscope with 40× oil immersion objective by illumination with the 543?nm laser. The live labelling and permeabilized staining were repeated in three separate experiments to confirm results.

Fluorescence detection size-exclusion chromatography

Suspension-adapted HEK293F cells were transfected with pcDNA3.1 vector containing wild-type Piezo1 or the M2493A/F2494A double mutant fused to tdTomato at the C terminus. Then, 48?h after transfection, 1?ml of the culture was pelleted and resuspended in 200?μl of buffer containing 25?mM HEPES pH 8.0, 150?mM NaCl, 1% CHAPS, 0.5% soy phosphatidylcholine, 1?mM PMSF, 2?μM pepstatin A, 2?μg ml−1 aprotinin, 2?μg ml−1 leupeptin. One-hundred microlitres of clarified supernatant was injected into a Superose 6 increase column in line with a fluorimeter tracking tdTomato fluorescence (excitation at 554?nM, emission at 581?nM).

Data availability

Electron microscopy maps have been deposited into the Electron Microscopy Data Bank (EMDB) under accession code EMD-7128. A structural model has been deposited to Protein Data Bank (PDB) under accession code 6BPZ. All other data can be obtained from the corresponding authors upon reasonable request.

Accession codes Primary accessions Electron Microscopy Data Bank

EMD-7128

Protein Data Bank

6BPZ

Change history 04 July 2022

A Correction to this paper has been published: https://doi.org/10.1038/s41586-022-04976-8

References

Ranade, S. S., Syeda, R. & Patapoutian, A. Mechanically activated ion channels. Neuron 87, 1162–1179 (2015)

Article CAS PubMed PubMed Central Google Scholar

Coste, B. et al. Piezo1 and Piezo2 are essential components of distinct mechanically activated cation channels. Science 330, 55–60 (2010)

Article ADS CAS PubMed PubMed Central Google Scholar

Coste, B. et al. Piezo proteins are pore-forming subunits of mechanically activated channels. Nature 483, 176–181 (2012)

Article ADS CAS PubMed PubMed Central Google Scholar

Ranade, S. S. et al. Piezo2 is the major transducer of mechanical forces for touch sensation in mice. Nature 516, 121–125 (2014)

Article ADS CAS PubMed PubMed Central Google Scholar

Woo, S. H. et al. Piezo2 is the principal mechanotransduction channel for proprioception. Nat. Neurosci. 18, 1756–1762 (2015)

Article CAS PubMed PubMed Central Google Scholar

Kim, S. E., Coste, B., Chadha, A., Cook, B. & Patapoutian, A. The role of Drosophila Piezo in mechanical nociception. Nature 483, 209–212 (2012)

Article ADS CAS PubMed PubMed Central Google Scholar

Li, J. et al. Piezo1 integration of vascular architecture with physiological force. Nature 515, 279–282 (2014)

Article ADS CAS PubMed PubMed Central Google Scholar

Nonomura, K. et al. Piezo2 senses airway stretch and mediates lung inflation-induced apnoea. Nature 541, 176–181 (2017)

Article ADS CAS PubMed Google Scholar

Zarychanski, R. et al. Mutations in the mechanotransduction protein PIEZO1 are associated with hereditary xerocytosis. Blood 120, 1908–1915 (2012)

Article CAS PubMed PubMed Central Google Scholar

Albuisson, J. et al. Dehydrated hereditary stomatocytosis linked to gain-of-function mutations in mechanically activated PIEZO1 ion channels. Nat. Commun. 4, 1884 (2013)

Article ADS PubMed CAS Google Scholar

Andolfo, I. et al. Multiple clinical forms of dehydrated hereditary stomatocytosis arise from mutations in PIEZO1. Blood 121, 3925–3935 (2013)

Article CAS PubMed Google Scholar

Bae, C., Gnanasambandam, R., Nicolai, C., Sachs, F. & Gottlieb, P. A. Xerocytosis is caused by mutations that alter the kinetics of the mechanosensitive channel PIEZO1. Proc. Natl Acad. Sci. USA 110, E1162–E1168 (2013)

ADS CAS PubMed PubMed Central Google Scholar

Coste, B. et al. Gain-of-function mutations in the mechanically activated ion channel PIEZO2 cause a subtype of Distal Arthrogryposis. Proc. Natl Acad. Sci. USA 110, 4667–4672 (2013)

Article ADS CAS PubMed PubMed Central Google Scholar

Lukacs, V. et al. Impaired PIEZO1 function in patients with a novel autosomal recessive congenital lymphatic dysplasia. Nat. Commun. 6, 8329 (2015)

Article ADS CAS PubMed Google Scholar

Chesler, A. T. et al. The role of PIEZO2 in human mechanosensation. N. Engl. J. Med. 375, 1355–1364 (2016)

Article CAS PubMed PubMed Central Google Scholar

Mahmud, A. A. et al. Loss of the proprioception and touch sensation channel PIEZO2 in siblings with a progressive form of contractures. Clin. Genet. 91, 470–475 (2017)

Article CAS PubMed Google Scholar

Haliloglu, G. et al. Recessive PIEZO2 stop mutation causes distal arthrogryposis with distal muscle weakness, scoliosis and proprioception defects. J. Hum. Genet. 62, 497–501 (2017)

Article CAS PubMed Google Scholar

Syeda, R. et al. Piezo1 channels are inherently mechanosensitive. Cell Reports 17, 1739–1746 (2016)

Article CAS PubMed Google Scholar

Lewis, A. H. & Grandl, J. Mechanical sensitivity of Piezo1 ion channels can be tuned by cellular membrane tension. eLife 4, e12088 (2015)

Article PubMed PubMed Central Google Scholar

Cox, C. D. et al. Removal of the mechanoprotective influence of the cytoskeleton reveals PIEZO1 is gated by bilayer tension. Nat. Commun. 7, 10366 (2016)

Article ADS CAS PubMed PubMed Central Google Scholar

Kamajaya, A., Kaiser, J. T., Lee, J., Reid, M. & Rees, D. C. The structure of a conserved Piezo channel domain reveals a topologically distinct β sandwich fold. Structure 22, 1520–1527 (2014)

Article CAS PubMed PubMed Central Google Scholar

Ge, J. et al. Architecture of the mammalian mechanosensitive Piezo1 channel. Nature 527, 64–69 (2015)

Article ADS CAS PubMed Google Scholar

Brohawn, S. G., Campbell, E. B. & MacKinnon, R. Physical mechanism for gating and mechanosensitivity of the human TRAAK K+ channel. Nature 516, 126–130 (2014)

Article ADS CAS PubMed PubMed Central Google Scholar

Pliotas, C. et al. The role of lipids in mechanosensation. Nat. Struct. Mol. Biol. 22, 991–998 (2015)

Article CAS PubMed PubMed Central Google Scholar

Phillips, R., Ursell, T., Wiggins, P. & Sens, P. Emerging roles for lipids in shaping membrane-protein function. Nature 459, 379–385 (2009)

Article ADS CAS PubMed PubMed Central Google Scholar

Jasti, J., Furukawa, H., Gonzales, E. B. & Gouaux, E. Structure of acid-sensing ion channel 1 at 1.9 Å resolution and low pH. Nature 449, 316–323 (2007)

Article ADS CAS PubMed Google Scholar

Kawate, T., Michel, J. C., Birdsong, W. T. & Gouaux, E. Crystal structure of the ATP-gated P2X4 ion channel in the closed state. Nature 460, 592–598 (2009)

Article ADS CAS PubMed PubMed Central Google Scholar

Zhao, Q. et al. Ion permeation and mechanotransduction mechanisms of mechanosensitive Piezo channels. Neuron 89, 1248–1263 (2016)

Article CAS PubMed Google Scholar

Long, S. B., Campbell, E. B. & Mackinnon, R. Crystal structure of a mammalian voltage-dependent Shaker family K+ channel. Science 309, 897–903 (2005)

Article ADS CAS PubMed Google Scholar

Liao, M., Cao, E., Julius, D. & Cheng, Y. Structure of the TRPV1 ion channel determined by electron cryo-microscopy. Nature 504, 107–112 (2013)

Article ADS CAS PubMed PubMed Central Google Scholar

Coste, B. et al. Piezo1 ion channel pore properties are dictated by C-terminal region. Nat. Commun. 6, 7223 (2015)

Article ADS PubMed Google Scholar

Drin, G. & Antonny, B. Amphipathic helices and membrane curvature. FEBS Lett. 584, 1840–1847 (2010)

Article CAS PubMed Google Scholar

Bavi, N. et al. The role of MscL amphipathic N terminus indicates a blueprint for bilayer-mediated gating of mechanosensitive channels. Nat. Commun. 7, 11984 (2016)

Article ADS CAS PubMed PubMed Central Google Scholar

Mansoor, S. E. et al. X-ray structures define human P2X3 receptor gating cycle and antagonist action. Nature 538, 66–71 (2016)

Article ADS CAS PubMed PubMed Central Google Scholar

Wu, J., Lewis, A. H. & Grandl, J. Touch, tension, and transduction - the function and regulation of Piezo ion channels. Trends Biochem. Sci. 42, 57–71 (2017)

Article PubMed CAS Google Scholar

Glogowska, E. et al. Novel mechanisms of PIEZO1 dysfunction in hereditary xerocytosis. Blood 130, 1845–1856 (2017)

Article CAS PubMed PubMed Central Google Scholar

Wu, J., Goyal, R. & Grandl, J. Localized force application reveals mechanically sensitive domains of Piezo1. Nat. Commun. 7, 12939 (2016)

Article ADS CAS PubMed PubMed Central Google Scholar

Ursell, T., Kondev, J., Reeves, D., Wiggins, P. & Phillips, R. in Mechanosensitive Ion Channels Vol. 1 (eds Kamkin, A. & Kiseleva, I. ) 37–70 (Springer, 2008)

Article Google Scholar

Suloway, C. et al. Automated molecular microscopy: the new Leginon system. J. Struct. Biol. 151, 41–60 (2005)

Article CAS PubMed Google Scholar

Zheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods 14, 331–332 (2017)

Article CAS PubMed PubMed Central Google Scholar

Zhang, K. Gctf: Real-time CTF determination and correction. J. Struct. Biol. 193, 1–12 (2016)

Article ADS CAS PubMed PubMed Central Google Scholar

Scheres, S. H. RELION: implementation of a Bayesian approach to cryo-EM structure determination. J. Struct. Biol. 180, 519–530 (2012)

Article CAS PubMed PubMed Central Google Scholar

Berndsen, Z., Bowman, C., Jang, H. & Ward, A. B. EMHP: An accurate automated hole masking algorithm for single-particle cryo-EM image processing. Bioinformatics 33, 3823–3826 (2017)

Article CAS Google Scholar

Bartesaghi, A., Matthies, D., Banerjee, S., Merk, A. & Subramaniam, S. Structure of beta-galactosidase at 3.2-Å resolution obtained by cryo-electron microscopy. Proc. Natl Acad. Sci. USA 111, 11709–11714 (2014)

Article ADS CAS PubMed PubMed Central Google Scholar

Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D 66, 486–501 (2010)

Article CAS PubMed PubMed Central Google Scholar

Adams, P. D . et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D 66, 213–221 (2010)

Article CAS PubMed PubMed Central Google Scholar

Chen, V. B. et al. MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr. D 66, 12–21 (2010)

Article CAS PubMed Google Scholar

Barad, B. A. et al. EMRinger: side chain directed model and map validation for 3D cryo-electron microscopy. Nat. Methods 12, 943–946 (2015)

Article CAS PubMed PubMed Central Google Scholar

Unni, S. et al. Web servers and services for electrostatics calculations with APBS and PDB2PQR. J. Comput. Chem. 32, 1488–1491 (2011)

Article CAS PubMed PubMed Central Google Scholar

The PyMOL Molecular Graphics System. Version 1.8 Schrödinger, LLC; http://www.pymol.org/ (2015)

Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004)

Article CAS PubMed Google Scholar

Download references

Acknowledgements

We thank H. Turner and G. Ozorowski for assistance with electron microscopy data collection. We acknowledge early efforts to characterize Piezo1 using electron microscopy by E. Wilson-Kubalek and R. Milligan, and S. Kakuda for screening purification conditions. We thank A. Sobolevsky for critical reading of the manuscript, and members of the Ward and Patapoutian laboratories for discussion. This work was supported by a Ray Thomas Edwards Foundation grant to A.B.W. and National Institutes of Health (NIH) grants NS083174 and DE022358 to A.P. Computational analyses were performed using shared instrumentation funded by NIH 1-S10OD021634. A.P. is an investigator of the Howard Hughes Medical Institute. This is manuscript 29582 from The Scripps Research Institute.

Author information Authors and Affiliations

Department of Neuroscience, Howard Hughes Medical Institute, The Scripps Research Institute, La Jolla, California, 92037, USA

Kei Saotome, Swetha E. Murthy, Jennifer M. Kefauver, Tess Whitwam & Ardem Patapoutian

Department of Integrative Structural and Computational Biology, The Scripps Research Institute, La Jolla, California, 92037, USA

Kei Saotome, Jennifer M. Kefauver & Andrew B. Ward

Contributions

K.S. prepared electron microscopy samples, collected and processed electron microscopy data, built the structural model and conducted fluorescence-detection size-exclusion chromatography analysis. S.E.M. carried out electrophysiological experiments and data analysis. J.M.K. developed sample preparation protocols. T.W. carried out immunostaining experiments. A.P. and A.B.W. supervised the project. K.S. drafted the manuscript, which was edited by A.P. and A.B.W. with input from all authors.

Corresponding authors

Correspondence to Ardem Patapoutian or Andrew B. Ward.

Ethics declarations Competing interests

The authors declare no competing financial interests.

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables Extended Data Figure 1 Purification and electron microscopy analysis of Piezo1.

a, Preparative gel filtration chromatogram of mouse Piezo1 after affinity purification and proteolytic removal of the GST tag. b, SDS–PAGE analysis of Piezo1 following affinity purification (pre-SEC) and after subsequent size-exclusion chromatography (SEC) steps (SEC fractions). Purifications of Piezo1 have been repeated more than three times with similar results. c, Aligned micrograph of purified Piezo1 embedded in a thin layer of vitrified ice. Scale bar, 100?nm. d, Representative 2D classes of Piezo1 showing different particle orientations.

Extended Data Figure 2 Classification and refinement of Piezo1 core.

a, Data processing flow chart. b, Local resolution maps calculated by the locres program in RELION 2.0. Colour key for local resolution (in Å) is shown. c, Angular distributions of the particles after the final step of refinement in RELION. The radius of the sphere is proportional to the number of particles with a given orientation. Only one-third of the sphere is shown due to applied C3 symmetry. d, FSC plots calculated using relion_postprocess for unmasked maps and maps with a soft mask applied to remove the contribution of scattered detergent density.

Extended Data Figure 3 Asymmetry of Piezo1 and blade-focused classification and refinement.

a, Duplicate unsharpened maps of Piezo1 refined without symmetry imposed, with the blue map rotated approximately 120° around the central axis of pseudosymmetry relative to the pink map, then superimposed with the ‘fit in map’ function of UCSF Chimera. The left panel shows top view, while the right panel shows a horizontal slice through the transmembrane region. b, Flow chart of propeller blade-focused classification and refinement procedure. c, Local resolution maps calculated by the locres program in RELION 2.0. Colour key for local resolution (in Å) is shown and is the same as Extended Data Fig. 2b. d, Angular distributions of the particles after the final step of refinement in RELION. The radius of the sphere is proportional to the number of particles with a given orientation. e, FSC plots calculated using relion_postprocess for unmasked maps and maps with a soft mask applied to remove the contribution of scattered detergent density.

Extended Data Figure 4 Fit of molecular model to electron density.

Select regions of the molecular model are shown as yellow cartoon, with superimposed electron density as blue mesh. The density is derived from the C3-symmetry core masked map for the top two rows, and derived from blade class 1 for the bottom row.

Extended Data Figure 5 Structural comparisons of current and previous (EMDB 6343) mouse Piezo1 structures.

For each panel, the ‘fit in map’ function of UCSF Chimera was used to align uncropped maps. In a–e, unsharpened map of C3 refinement is shown. a–c, Top (a), bottom (b) and side (c) views of current Piezo1 (green) and EMDB accession EMD-6343 (grey) structurally aligned. d, Expanded top view of the cap domain to illustrate an approximately 15° rotation of the cap between the two maps. e, Expanded sliced top view of pore-proximal transmembrane region. Subtle conformational rearrangements are present between the current and previous structures in the inner and outer helices. f, g, Top and side views of EMDB accession EMD-6343 (grey), blade class 1 (blue), and blade class 2 (orange). The trajectories of the propeller blades are broadly similar across the maps in both the membrane-parallel and membrane-normal directions.

Extended Data Figure 6 Structural features of Piezo repeats.

a, Top view of Piezo1 depicted as cartoon, with amphipathic helices from Piezo repeats A, B and C shown in red. Extracellular cap domain is omitted for clarity. b–d, Ribbon models of amphipathic helices from Piezo repeat A (b), Piezo repeat B (c), and Piezo repeat C (d). Cα positions are shown as spheres and hydrophobic amino acids are coloured grey, basic amino acids are coloured blue, acidic amino acids are coloured red, and polar uncharged amino acids are coloured yellow. e, Cartoon model of a portion of the Piezo1 propeller blade, with Piezo repeats coloured separately. A strong density peak putatively representing a lipid head group is shown as a pink mesh. f, Expanded view of the putative lipid binding site sandwiched between Piezo repeats B and C (helices B4 and C1). Electron density (blue mesh) is superimposed onto the molecular model, and the putative lipid head group is shown as pink mesh. Side chains contributing to the binding pocket are shown as yellow sticks and labelled.

Extended Data Figure 7 Expression and functional properties of Piezo1 mutants.

a, HEK293T cells deficient in Piezo1 (P1KO) that express the Piezo1 double mutant M2493A/F2494A do not show discernible currents activated by indentation with blunt pipette (left) or stretch (right). b, Fluorescence-detection size-exclusion chromatography traces of untransfected HEK293F cells (grey) or cells expressing C-terminal tdTomato fusions of wild-type Piezo1 or double mutant M2493A/F2494A injected into Superose 6 increase column. Similar to the wild type, the double mutant displays a single dominant peak eluting shortly after void volume, indicating proper trimeric expression. Data are from one independent experiment. c, Representative images of Myc labelling in HEK293T-P1KO cells transfected with Piezo1–IRES-GFP (top), Piezo1–Myc–IRES-GFP (middle) or Piezo1(M2493A/F2494A)–Myc–IRES-GFP (bottom). GFP, green fluorescent protein. IRES, internal ribozyme entry site. Myc tags were inserted at position 897. Immunostaining was performed before (left) or after (right) cell permeabilization. Piezo1(M2493A/F2494A)–Myc is labelled at the surface in live cells similar to Piezo1–Myc, indicating that surface expression is preserved in the non-functional mutant. Scale bar, 10?μm. Experiments were repeated three times with reproducible results. d, Left, representative traces of stretch-activated single channel currents recorded (−80?mV) from HEK293T-P1KO cells expressing wild-type, M2493A or F2494A Piezo1 channels (traces displayed after applying 1?kHz digital filter). The corresponding pressure stimulus used to elicit the response is illustrated above the current trace. For each condition, the amplitude histogram for the corresponding trace is depicted on the right. The Gaussian fit for the closed (black curve) and open (red curve) components is overlaid on the histograms. e, Left, average I–V relationship of stretch-activated single-channel currents from wild-type (n?=?4), M2493A (n?=?5) or F2494A (n?=?4) Piezo1 channels. Amplitude was measured as a difference in Gaussian fits of full-trace histograms. Right, mean unitary conductance calculated from the slope of linear regression line fit to individual cells in each condition. **P?<?0.01, one-way ANOVA with Dunn’s comparison. f, Mean I–V relationship curves of mechanically activated currents recorded from wild-type (n?=?6), M2493A (n?=?8) or F2494A (n?=?9) Piezo1 channels with 150?mM CsCl-based intracellular solution and 100?mM CaCl2-based extracellular solution. Currents were elicited from −69.6 to 50.4?mV (Δ20?mV). Inset, expanded view of curves between 5 and 15?mV. Values are mean?±?s.e.m. g, Mean reversal potential from individual cells; Piezo1: 8.8?±?1.3?mV (n?=?6), M2493A: 10.2?±?1.2?mV (n?=?8) and F2494A: 13.5?±?0.5?mV (n?=?9). **P?=?0.0063, one-way ANOVA with Dunn’s multiple comparison test. N denotes X individual cells.

Source data

Extended Data Table 1 EM data collection, data processing, model refinement and validation Full size table Extended Data Table 2 Mechanically activated current properties of Piezo1 constructs Full size table Supplementary information Supplementary Data

Sequence alignment of Piezo orthologues. Amino acid sequence alignment of selected regions of mouse Piezo1 (mPiezo1; UniProt E2JF22), human Piezo1 (hPiezo1; UniProt Q92508), mouse Piezo2 (mPiezo2, UniProt Q8CD54), human Piezo2 (hPiezo2, UniProt Q9H5I5), fruit fly Piezo (dmPiezo, UniProt M9MSG8). Structural features are annotated above the mPiezo1 sequence. Unmodeled regions are depicted as dotted lines. (PDF 4081 kb)

Life Sciences Reporting Summary (PDF 68 kb) PowerPoint slides PowerPoint slide for Fig. 1 PowerPoint slide for Fig. 2 PowerPoint slide for Fig. 3 PowerPoint slide for Fig. 4 PowerPoint slide for Fig. 5 Rights and permissions

Reprints and permissions

About this article Cite this article

Saotome, K., Murthy, S., Kefauver, J. et al. Structure of the mechanically activated ion channel Piezo1. Nature 554, 481–486 (2018). https://doi.org/10.1038/nature25453

Download citation

Received 07 September 2017

Accepted 14 December 2017

Published 20 December 2017

Issue Date 22 February 2018

DOI https://doi.org/10.1038/nature25453

Share this article

Anyone you share the following link with will be able to read this content:

Get shareable link

Provided by the Springer Nature SharedIt content-sharing initiative

Subjects Cryoelectron microscopy Ion channels This article is cited by Piezo1 activation protects against sepsis-induced myocardial dysfunction in a pilot study Angwei GongJing DaiBing Xiao

Scientific Reports (2025)

TRPML1 ion channel promotes HepaRG cell differentiation under simulated microgravity conditions Huancai FanDongyuan LüSijin Liu

npj Microgravity (2025)

Lipid-mediated gating of a miniature mechanosensitive MscS channel from Trypanosoma cruzi Jingying ZhangAashish BhattPeng Yuan

Nature Communications (2025)

Targeted elimination of mesenchymal-like cancer cells through cyclic stretch activation of Piezo1 channels: the physical aspects Ivana Pajic-LijakovicMilan MilivojevicPeter V. E. McClintock

Biophysical Reviews (2025)

Iron-Mediated Regulation in Adipose Tissue: A Comprehensive Review of Metabolism and Physiological Effects Xinyu YangXianghong WangHongyun Lu

Current Obesity Reports (2025)