Data for: PIEZO1-HaloTag hiPSCs: bridging molecular, cellular and tissue imaging

PIEZO1 channels play a critical role in numerous physiological processes by transducing diverse mechanical stimuli into electrical and chemical signals. Recent studies underscore the importance of endogenous PIEZO1 activity and localization in regulating mechanotransduction. To enable physiologically and clinically relevant human-based studies, we genetically engineered human induced pluripotent stem cells (hiPSCs) to express a HaloTag fused to endogenous PIEZO1. Combined with super-resolution imaging, our chemogenetic approach allows precise visualization of PIEZO1 in various cell types. Further, the PIEZO1-HaloTag hiPSC technology allows non-invasive monitoring of channel activity via Ca2+-sensitive HaloTag ligands, with temporal resolution approaching that of patch clamp electrophysiology. Using lightsheet imaging of hiPSC-derived neural organoids, we also achieve molecular scale PIEZO1 imaging in three-dimensional tissue samples. Our advances offer a novel platform for studying PIEZO1 mechanotransduction in human cells and tissues, with potential for elucidating disease mechanisms and development of targeted therapeutics.

Methods
Micropatterned neural rosettes (MNRs) of 140 µm in diameter grown on 19 mm square Arena A CYTOO chips (Cat. No.10-020-00-18) were imaged on a modified adaptive optical lattice light-sheet microscope (AO-LLSM). Prior to all imaging, the microscope was calibrated to correct for optical aberrations from the system. The Cytoo chip was mounted on a custom designed sample holder and immersed in 40 ml of phenol-free MNR culture medium. The excitation (Thorlabs 0.6 NA, TL20X-MPL) and detection (Zeiss 1.0 NA, 421452-9800-000) objectives were also immersed into the imaging medium. The MNRs were imaged with a multiBessel square lattice light sheet with the NAsq of 0.35 and a 0.3/0.4 NA annular mask. The 488 nm, 560 nm and 642 nm lasers were used to visualize SPY505-DNA (DNA, Spirochrome Cat# CY-SC101), SPY555-actin (actin, Spirochrome Cat# CY-SC202), and Janelia Fluor® 635-HaloTag Ligand (Non-Bapta) or Janelia Fluor® 646-BAPTA-AM- HaloTag ligand (PIEZO1), respectively, with power at the back pupil of the excitation objective of 45 µW for 488 nm, 50 µW for 560 nm and ranging between 2.1-2.9 mW for 642 nm. To balance volumetric imaging speed, signal to noise and photobleaching, images were acquired using camera exposure between 30–50 milliseconds. All data from the AO-LLSM were collected on two Hamamatsu ORCA-Fusion sCMOS cameras. Emission light from actin, DNA, and JF635 or JF646 was separated by a dichroic mirror (Chroma T600DCRB) and passed to two cameras equipped with either Semrock FF01-600/37-25 emission filter for actin or Semrock FF01-538/685-25 filter for DNA and JF635 or JF646-BAPTA .

The 3D volumetric imaging of MNRs was performed by tiling AO-LLSM across the sample. Each tile (approximately 200 x 100 x 15 µm3) was scanned by moving the sample stage at 400 nm step sizes. Each neighboring tile had a 5 µm overlap between each adjacent tile. Mismatch between the excitation LLS and the detection focal plane caused by the MNR were corrected prior to every acquisition and was essential to ensure optimal imaging of PIEZO1, DNA and actin structures. The AO-LLSM data was processed using MATLAB versions R2022b and R2023a. The large 3D MNR volumes were stitched in skewed space, deconvolved, deskewed and rotated on Advanced Bioimaging Center’s computing cluster at UC Berkeley using the computational pipelines published on GitHub (https://github.com/abcucberkeley/LLSM5DTools). The skewed space deconvolution was performed with experimentally measured point spread functions obtained from 200 nm fluorescent beads (Invitrogen FluoSpheres Carboxylate-Modified Microspheres, 505/515 nm, F8811). The nuclei were denoised using Content-Aware Image Restoration (CARE). The training data for the denoising model was collected using lattice light-sheet microscopy by volumetrically scanning LLC-PK1 cells expressing nuclear marker (H2B) to record low SNR and corresponding high SNR 3D stacks. The AOLLSM instrument was controlled using a custom LabVIEW based image acquisition software (National Instruments, Woburn, MA) licensed from Janelia Research Campus, HHMI.

To observe the dynamics of PIEZO1-HaloTag in MNRs, 3-plane stack videos were acquired. For this, the sample stage (using the SmarAct MLS-3252 Electromagnetic Direct-Drive) was rapidly scanned across a 1 µm sample range (comprised of three image planes spaced 400 nm along the sample stage scan axis, corresponding to 215 nm along the optical z axis) at an interval between 160-210 ms per stack for 30 time points. The dynamic time series datasets were deskewed and maximum intensity projected prior to analysis. The corresponding DNA and actin volumes were recorded at the same location prior to acquiring the PIEZO1-HaloTag time series. A new dataset of 30 time points was collected every ~10 µm throughout the MNR and was repeated on multiple samples.