Materials
Quinine and Tetraethylammonium chloride (TEA) were purchased from TargetMol. Di-4-ANEPPS and nicotine were purchased from GlpBio. Glycine (gly), L-Glutamic acid (glu), Calcium chloride (CaCl2) and 4-aminopyridine (4-AP) were purchased from Sigma-Aldrich. Ethylenediaminetetraacetic acid (EDTA) was purchased from Sangon Biotech. N-2-Hydroxyethylpiperazine-N’-2-ethanesulfornic acid (HEPES), Glucose, Sodium chloride (NaCl) were purchased from Macklin. Potassium chloride (KCl), N-methyl-D-glucamine (NMDG), Para formaldehyde were purchased from Aladdin. Indium tin oxide (ITO)-coated glass slides (coating thickness, ~ 23 nm; resistance, 80-100 Ω/square) were purchased from South China Science & Technology Company Limited, Shenzhen, China. All the chemicals were used directly without further purification, and all the buffers were prepared with sterile deionized water (for cell culture, Solarbio) and filtered through a 0.22 μm pore size syringe filter prior to use.
Sample preparation
Cell culture. PC-12 cells were cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS), HEK-293 cells were cultured in MEM medium supplemented with 10% FBS, SH-SY5Y cells were cultured in MEM/F12 medium supplemented with 15% FBS. All cells were kept in a humidified atmosphere at 37 ⁰C and 5% CO2 / 95% air. For measurement, the cells were plated onto ITO electrode and grown to 80% confluency. Before experiments, the culture medium was discarded and cells were rinsed twice with 200 μl phosphate buffer saline (PBS).
Preparation of electrochemical cell. ITO-coated glass slides (coating thickness, ~23 nm; resistance, 80-100 Ω/square) were cleaned by sonication sequentially in acetone, ethanol, and DI water, each for 20 min, and then dried with N2. A 3 mm thick polydimethylsiloxane (PDMS) with a circular hole of 10 mm at its center was attached to ITO surface, assembled into a culture chamber for cells. Both ITO slides and PDMS were irradiated under UV for 30 min before cell seeding.
Expression of NMDAR in HEK-293 cell. HEK-293 cells were cultured in a 10 mm PDMS chamber in MEM supplemented with 10% FBS at 37 ⁰C in a 5% CO2 atmosphere. Gene expression vectors encoding Human GluN1 (pCAG-mNR1-GFP) and GluN2A (pCAG-mN2A-mcherry) were purchased from Vigenebio (Shandong, China). Vectors were used for expression without further amplification or purification. GluN1 and GluN2A victors (0.5 µg: 0.5 µg) were transfected into the HEK-293 cells, when the cell confluence was about 80%. Lipofectamine 3000 (Invitrogen, Carlsbad, CA) transfection reagent was used according to the manufacturer’s protocol. The expression of NMDAR in HEK-293 cells was examined by confocal fluorescence imaging (TCS SP8 STED, Leica, Supplementary Fig. 1), immediately after EM-iSCAT detection and fixation in 4% PFA in PBS.
Optical setup
The experimental setup is depicted schematically in Supplementary Fig. 2. The imaging setup was built on an inverted microscope (Nikon Ti2-U). The collimated output of a 532 nm laser (MW-GLN-532, Changchun Laser Optoelectronics Technology Co., Ltd.) is passed through a 5× achromatic Galilean beam expander (GBE05-A, Thorlabs). A 300 mm focal length len focuses the beam onto the back focal plane of the microscope objective (S Plan Fluor ELWD 60×/0.7 NA, Nikon). The signal collected by the objective is transmitted through a partial reflector (PR) and focused onto a sCMOS camera (Dhyana 400D, Tucsen). The PR consists of a 30 nm thick circular gold layer of 2 mm diameter (transmissivity of 10%) evaporated onto a 10:90 (R:T) plate beam splitter. An AC-modulated voltage was applied to the ITO slide with a potentiostat (CH150, Corrtest) using a three-electrode configuration, where the ITO, a Ag/AgCl wire and a Pt coil served as the working, reference and counter electrodes, respectively. The size of the working electrode and the distance between working and counter electrodes were maintained by a double-sided tape (thickness: 89µm, 3M Tape 665) with a 2 mm hole. A digital signal generator (DG1000Z, RIGOL) was used to synchronize the applied potential and image acquisition. All data were recorded at 25 ℃ by a home-built sample stage heater.
Data acquisition
For the experiment presented in Figs. 2 and 4, images were acquired with an exposure time of 300 µs, at a frame rate of 300 Hz, and the voltage modulation frequency was set to 30 Hz, with an amplitude of 1V. Experiments presented in Fig. 3 were carried out using the same exposure time at a frame rate of 1.5 kHz, and the voltage modulation frequency was set to 150 Hz, with an amplitude of 1.2 V.
Data analysis
Unless otherwise stated, analysis was performed using custom written code in MATLAB (MathWorks). To remove the static scattering background from the ITO surface and cell membrane structure, ratiometric images, I, were calculated as \(I=\left({I}_{i}-{I}_{bg}\right)/{I}_{bg}\), where \({I}_{bg}\) is the temporal median intensity background calculated from first 300 frames of image sequence. We performed a STFT analysis on each pixel of ratiometric image sequence with a Hann window length of 128 frames and rolling step length of 1 frame. Then we extracted the amplitude averaged over 1 s, and constructed the STFT amplitude image (EM-iSCAT image).
For whole-cell detection (Supplementary Fig. 3a), the Hann window’s overlap length was set to 64 frames, and the mean intensity within the profile of the cell, according to the bright field image at the same location, was used to determine the whole cell signal (\({{\Phi }}_{i}\)). Thus, the change of charge density on the whole cell membrane along with the cell electrical activity can be calculated by \(\varDelta {\Phi }={{\Phi }}_{i}-{{\Phi }}_{0}\), where \({{\Phi }}_{0}\) is the whole cell signal at t = 0. Considering that the electrical activity of the cell is dominated by ion transport across the membrane, the summation of the activity of all ion channels on the membrane can be obtained in this mode.
For single channel detection (Supplementary Fig. 3b), we introduce the concept of an accumulated temporal amplitude map (ATAP), wherein we sum all the STFT images into one image. The pixel value of ATAP is related to the change of surface charge density within observation time. Considering the blinking behavior of single ion channel and its continuously changing position due to membrane fluidity, active regions with larger pixel value in ATAP indicated a greater change of surface charge density, which may be caused by the activity of one or more ion channels. A region of interest (ROI) with 10 × 10 pixels was selected for each hot point that represents the ion channels active zone. The single channel signal, \(\varDelta \text{A}={\text{A}}_{i}-{\text{A}}_{0}\), was determined by the amplitude of a 2D Gaussian function (\({\text{A}}_{i}\)) fitting frame by frame to the spot at the corresponding ROI position in STFT image sequence (snaps shown in Fig. 3 and Supplementary Fig. 3b), and the mean intensity within the ROI was used to determine the baseline (\({\text{A}}_{0}\)). Dynamics related to channel opening and closing were determined based on the single channel signal from EM-iSCAT images.