Glutathione Quantification in Live Cells with Real-Time Imaging and Flow Cytometry

Summary Glutathione (GSH) is a highly dynamic, high abundance molecule regulating redox homeostasis in most mammalian cells. Traditional methods could not achieve quantification of glutathione in live cells with high spatial and temporal resolution. Here, we provide protocols on how to use reversible reaction-based ratiometric fluorescent probes, RealThiol (RT) and its derivatives, to quantify GSH globally or in specific organelles. The protocols are applicable to cultured or harvested cells through confocal imaging and flow cytometry. For complete details on the use and execution of this protocol, please refer to Chen et al. (2017) and Jiang et al. (2015, 2017, 2018a).

HIGHLIGHTS Quantitative realtime imaging of glutathione in live cells Subcellular quantification of glutathione using chemical-or proteinbased targeting Quantitative flow cytometry of glutathione in live single-cell suspension SUMMARY Glutathione (GSH) is a highly dynamic, high abundance molecule regulating redox homeostasis in most mammalian cells. Traditional methods could not achieve quantification of glutathione in live cells with high spatial and temporal resolution. Here, we provide protocols on how to use reversible reaction-based ratiometric fluorescent probes, RealThiol (RT) and its derivatives, to quantify GSH globally or in specific organelles. The protocols are applicable to cultured or harvested cells through confocal imaging and flow cytometry. For complete details on the use and execution of this protocol, please refer to Chen et al. (2017) and Jiang et al. (2015Jiang et al. ( , 2017Jiang et al. ( , 2018a.

BEFORE YOU BEGIN Cell Culture
Timing: 1-7 days before the experiment All cell lines are purchased through ATCC unless otherwise specified.
For imaging, cells should be cultured on a glass-bottom dish or equivalent; for flow cytometry, cells should be prepared as single-cell suspension.

MATERIALS AND EQUIPMENT Stock Solutions
To prepare the stock solution, dissolve the lyophilized RT-AM/RT probe powder in 10 mL of anhydrous DMSO to make a 5 mM stock solution.
For HaloRT ligand, dissolve the lyophilized HaloRT ligand powder in 10 mL anhydrous DMSO to make a 5 mM stock solution.
For MitoRT probe, dissolve the lyophilized MitoRT probe powder in 10 mL of deionized water (DI water) to make a 5 mM stock solution.
CRITICAL: DMSO is hygroscopic. DMSO with significant amount of water can hydrolyze RT-AM. It is highly recommended to dissolve RT-AM in anhydrous DMSO.
CRITICAL: DMSO can react with MitoRT and cannot be used as the solvent for MitoRT.
Stock solution is stable at À80 C for up to 6 months. Make small aliquots if necessary.

Staining Solutions
To prepare the staining solution, typically, the stock solution is diluted with its corresponding solvent (DMSO for RT-AM and HaloRT ligands, DI water for MitoRT) 50 times to 100 mM. Staining solution can be stored at À20 C for 1 week. Make small aliquots if necessary.

Working Solutions
To prepare working solution, typically, the staining solution is further diluted with desired imaging buffer (fresh cell culture medium or other appropriate buffers) 100 times to a final probe concentration of 1 mM.
Note: For RT-AM and HaloRT ligand, make sure to add the imaging buffer (aqueous phase) into the staining solution (organic phase), otherwise, precipitation of the fluorescent probes may happen. If precipitation occurs, centrifuge at 3,000 3 g for 3 min and keep the supernatant. This may lead to inconsistent probe concentration but should not interfere with most measurements.
Note: PBS is generally not advised for live cell experiments as the GSH level usually changes when cells are exposed to PBS for more than 5 min.

Other Solutions Used in This Protocol
GSH standard solution can be prepared by dissolving GSH powder in PBS or DI water. Typical concentration of GSH standard solution is between 0. Adjust gain and offset of each channel to satisfy the whole dynamic range of the experiment by imaging a positive (e.g., transient 200 mM GSH-ester treated cells) and negative control (e.g., 48-72 h 500 mM BSO treated cells) first.
Note: The setting for Zeiss confocal microscopes provides the optimal signal to noise ratios on our hands.

Slides and Imaging Dish Preparation
Coat glass slides with poly-D-lysine and dry under 18 C-25 C. The coated slides can then be used for beads (for calibration) and non-adherent cell imaging. Alternatively, coated glass slides can be used to culture adherent cells and used for imaging similar to glass-bottom dishes. For glass-bottom dish or chamber slides, no additional steps are needed before imaging. Clean the bottom of imaging container with appropriate tools right before imaging to ensure the imaging quality.
Note: DAPI channel usually uses 380 nm excitation and 450/50 nm emission, which is not optimal but still feasible. FITC channel is generally not advised because the emission filters do not match well.

OPEN ACCESS
STAR Protocols 1, 100170, December 18, 2020 Adjust gain of each channel (including forward scatters, or FSC, side scatters, or SSC, green and blue channel) to satisfy the whole dynamic range of the experiment by running positive and negative control samples first.
Alternatives: Other commercially available confocal microscopes and flow cytometers with the appropriate optical filters can also be used for this protocol.

STEP-BY-STEP METHOD DETAILS Cell Sample Preparation
Timing: 2-3 days In this step, sample staining and pretreatment for different scenarios are described in details. For best quantitative results, prepare control samples as described. For HaloRT-based probes, transfect cells with the appropriate plasmid is required.

Cell Transfection (for HaloRT-Based Experiment Only)
Timing: 2-3 days 1. Seed the cells 24 h before transfection at desired density in an appropriate container (2.0310 6 cells per sample in a glass-bottom dish for imaging; 1.0310 5 cells per well in a 12-well plate for flow cytometry). 2. Transfect cells with desired HaloTag plasmids using Lipofectamine 3000 transfection reagent or other suitable reagents (e.g., for nuclear GSH measurement, use Halo-NLS plasmid for transfection, 500 ng DNA per glass-bottom dish/per well in a 12-well plate). Culture the cells for 24-48 h.
Alternatives: a stable cell line expressing desired HaloTag protein can be generated.

Drug Treatment
Timing: 1 min to 3 days 3. For GSH synthase inhibitor BSO treatment, add the appropriate amount of BSO to the cell culture medium while seeding the cells. Culture for 48-72 h. Cysteine free culture medium should be adopted for cell culture and for the following measurements.
Note: BSO will partially affect cell growth when the seeding density is low. To best maintain the viability of cells, use a higher seeding density.

OPEN ACCESS
Note: To minimize changes to the culture environment, it is advised to add the drug at a relatively high concentration, e.g., 1003 or 103. It is also advised to avoid directly adding DMSO containing stock solution into water phase during signal acquisition as the release of heat from solvent mixing may instantly kill the cells and create artifacts.
Note: Some cells are highly capable of defending against external stimulants. In this case, the changes of the GSH concentrations may be subtle, or the changes may be reversed within a short period of time. Thus, real-time imaging can be advantageous over other signal acquisition methods for recording such transient events.

Cell Staining
Timing: 8-60 min per sample 6. Replace cell culture medium with working solution containing RT-AM/MitoRT/HaloRT ligand. For adherent cells, this step could be done through aspiration of culture medium and then addition of the working solution; for non-adherent cells, a centrifugation step is required before removing the culture medium, and a resuspension step after addition of working solution. 7. For flow cytometry experiments, all cells need to be in single-cell suspension prior to staining. For adherent cells, use trypsin to digest cells and resuspend into fresh culture medium. The staining process follows the non-adherent cell staining protocol with centrifugation and resuspension steps. 8. Incubate the cells in the working solution for 8-10 min at 18 C-25 C for RT-AM stained cells; 15-30 min at 37 C for MitoRT and HaloRT ligand stained cells. Alternatively, longer incubation (30-60 min) at room temperature for MitoRT and HaloRT ligand stained cells may be used in cases where heating devices are unavailable.
CRITICAL: Temperature of staining will influence the distribution homogeneity of RTbased probes. Generally, a lower temperature will induce less aggregation or clearance of probe thus lead to more homogeneous probe loading. For organelle targeting probes, the targeting mechanism will lock desired probes in place with the protein, a higher loading temperature will facilitate clearing of non-specific probes in undesired organelles.
9. For non-adherent cells (including polystyrene bead-based calibration), place 10 mL of suspension containing stained cells onto a coated glass slides, and then cover with cover glass. The sample is then ready for imaging.

Note:
The sample may dry out in 15 min, please mind the timing. Alternatively, seal the sample with a specific O-ring to prevent drying.
CRITICAL: Washing steps should be generally avoided for RT-AM and MitoRT stained cells when confocal imaging is used for signal acquisition because confocal imaging can easily distinguish between intracellular signal and excess fluorophore in solution and a washing step may induce additional stress to cells.
Note: Washing step will influence cell state and lead to rapid changes in cytosolic redox conditions. We have observed 20% GSH concentration fluctuation due to washing. This is especially important for imaging experiment because 1) sampling in typically limited in confocal imaging experiment; 2) imaging experiment is typically designed for capturing fast yet subtle changes compared to other type of experiments thus imaging is more prone to such bias. optimal after one washing step. For confocal imaging experiments, the washing step may be skipped if the signal to noise is above 5.
Note: HaloRT ligand require long incubation for cells to clear non-specific binding of probes. The signal is also retained better than chemical targeting and thus allow possible longer recovery time to overcome the fluctuation caused by washing.
11. Wash cells with fresh culture medium can be skipped for flow cytometry experiments. If the signal to noise is below 5, washing steps may help.

Signal Acquisition
Timing: 2-15 min per sample In this step, quantitative fluorescent signal acquisition of individual samples using confocal microscope or flow cytometer are described. Signal acquisition steps should immediately follow staining steps for each sample. Generally, avoid staining and measuring in batches as GSH and the probes are highly dynamic, and that staining relies on fast reversible chemical reactions.

Confocal Imaging
Timing: 5-15 min per sample 12. Use sequential imaging on two channels. Use line-switching mode when possible, especially for imaging within smaller organelles such as mitochondria. Z-stack is generally not advised due to the slow speed. Alternatively, if sequential imaging is not available or too slow, a more complicated calculation should be performed when processing the data. Keep the imaging conditions exactly the same as the conditions used for the calibration curve when absolute quantification is needed.
Note: Generally, ratio between signals from blue and green channels is linear to the GSH concentration. However, if the imaging condition is not ideal, the linearity will not hold and thus a complex ratio (R-R min )/(R max -R) should be used, where R is the measured ratio, R max is the ratio with maximum signal (need 50 mM or higher of GSH for measurement), R min is the ratio with minimum signal (need 0.5 mM or lower combined GSH, cysteine, and other thiols for measurement).
13. For RT-AM based global GSH quantification in live cells, we advise a short imaging time due to the relatively fast probe clearance especially in cancer cell lines. In cases when long monitoring is needed, ensure applying an appropriate non-treated, stained cell sample as a control. Alternatively, the HaloRT strategy can be applied for long-term monitoring experiments. 14. For MitoRT based mitochondria GSH (mGSH) quantification in live cells, a higher magnification may be needed. Some super-resolution imaging method (such as stimulated emission depletion super-resolution imaging or airy-scan) may be used, but the quantification may be compromised. 15. For HaloRT-based organelle-specific GSH quantification in live cells, the imaging time can be extended.
CRITICAL: Stain the next sample while waiting for imaging to complete. Keep staining time the same for each sample.
16. For quantification or semi-quantification purpose, please refer to the calibration section for details about creating standard curve for calibration; for qualitative assessment of samples, standard curve can be skipped.

OPEN ACCESS
Note: Please keep laser intensity low to avoid light toxicity and fluorophore bleaching.
Note: Please have an appropriate probe only control for live monitoring experiments to test the probe clearance speed in your cells of interest.
Note: Trypan blue can be used to quench cell surface non-specific stain in cases where cell surface fluorescence is high and the signal to noise is below 5. For quality purpose, signal to noise ratio of 2 is considered low and signal to noise above 6 is considered high. Please refer to Troubleshooting for additional information regarding abnormal imaging results.

Flow Cytometry
Timing: 2-5 min per sample 17. Typically, cells are dispersed into single-cell suspension (500 mL in a 5 mL polystyrene tube) and brought to the flow cytometer prior to staining. 18. The gain settings should allow signal to noise >3 in both channels for both positive and negative control samples compared with the non-stained control sample. 19. It is advised to acquire signals from at least 10,000 cells per sample with an optimal of 30,000 or more cells. 20. For quantification or semi-quantification purpose, please refer to the calibration section for details about creating standard curve for calibration; for qualitative assessment of samples, standard curve can be skipped.
Note: MitoRT may partially lose organelle targeting ability during single-cell suspension preparation from adherent cells, because part of the mitochondria loses activity and polarity during cell detachment process thus interrupts the charge based chemical targeting mechanism. To confirm, an imaging flow cytometer can be used. Please refer to Troubleshooting for additional information.  (Jiang et al., , 2018a Note: Image processing in Fiji ImageJ (steps 21-23) can be partially automated by scripts. However, ratio data collection and quantitative analysis still require manual work. Example script can be found on GitHub: https://github.com/jinwanglab/RT-Imaging-Analysis 28. Export the blue and green channel signal in the single-cell population to a .csv file. 29. Calculate the ratio in each cell by divide signal from the blue channel by that from the green channel. Remove outliers that are 4 standard deviations away from the mean. 30. Compare with standard curve to deduct the absolute GSH concentrations when performing quantification or semi-quantification.
Note: Statistical data analysis from step 29 can be automated by scripts. Example script can be found on GitHub: https://github.com/jinwanglab/RT-FACS-Analysis

Calibration against Lysate-Based Assay (Preferred Method)
Timing: 2-3 days 31. Prepare two identical groups of cells by treating with increasing concentrations of BSO (e.g., 0-500 mM for HeLa cells) for 48-72 h. One group of cells should be cultured in a 12-well or 6-well plate for lysate-based assay, the other group of cells should be cultured in the appropriate containers as described above. 32. Take the group of cells cultured for lysate-based assay, measure GSH using standard Ellman's assay (Rahman et al., 2006) or liquid chromatography-mass spectrometry-based assay Squellerio et al., 2012). 33. Measure the other group of cells using any of the probe with the corresponding protocol described above. Analyze and record ratio value for each sample. 34. Compare assay results with ratio and fit with an appropriate function.

Note:
The ratio and GSH concentration may not follow linear relationship depending on the actual experimental conditions. Please refer to the literature (Jiang et al., 2015 for detailed explanation and function deduction in cases where the fitting significantly deviates from linear. Note: Even for semi-quantitative/qualitative experiments, it is still recommended to perform this calibration experiment to assess GSH levels based on the ratio readouts. 36. Serial dilute GSH stock solution to desired concentration, for example, 20, 10, 5, 2.5, 1.25, 0.625 mM of GSH. 37. Mix equal amount (typically 4 mL) of GSH with RT solution, incubate at 18 C-25 C for 10 min. 38. Add the above solution onto a coverslip, cover with slide glass, and take images. Make sure to pipette the mixture to prevent beads from aggregating into one spot. 39. To measure R min , use 0.01 mM of final concentration of GSH; to measure R max , use 100 mM final concentration of GSH. 40. Calculate the ratio from each group, plot (RÀR min )/(R max ÀR) against GSH concentration, a linear relationship can be established as a standard curve for comparing results.
Note: All conditions for calibration should be kept the same with the actual imaging experiments, including objectives, temperature, magnification, Gain/PMT, offset, filter, laser intensity settings.
Note: This calibration method pairs with the RT-AM based imaging experiments.
Note: The GSH standard solutions can be oxidized in air. It is advised to perform this experiment as fast as possible or use nitrogen for protection. Always start with the lowest concentration of GSH because it is more prone to oxidation errors.
Note: This calibration method can also be used for HaloRT-based imaging experiments, in which the RT solution should be replaced with the HaloRT solution (made from recombinant HaloTag protein reaction with the HaloRT ligand).
Note: This calibration method is not suitable for MitoRT based imaging experiments, possibly because the microenvironment for MitoRT is mitochondria cannot be faithfully simulated using an aqueous solution.

Cell Volume Estimation
Timing: 5-60 min The cell volume data can be used to calculate the absolute intracellular GSH concentrations from bulk lysate-based measurements. The data is needed for lysate-based assay calibration purpose.

Protein Concentration-Based Estimation
Timing: 60 min 50. Use standard bicinchoninic (BCA) assay or other assays to measure the protein concentrations of cells. The cells should be prepared the same way as the calibration group. Other protein quantification methods such as Bradford assays can also be used as long as the dynamic range fits. 51. Estimate cell volume using available data on the corresponding protein concentration per unit volume.

EXPECTED OUTCOMES Expected Absolute GSH Concentration
The typical concentration range of intracellular GSH is 1-10 mM, where RT-based probes show optimal performance. In cancer cells or highly metabolic active cells, the GSH concentration is usually close to the high-end; while in certain drug treated cells and certain organelles such as in the ER, the GSH concentration is closer to the low-end. For cultured cells, low passage cells showed slightly lower overall GSH concentration. Cells may exhibit day-to-day or batch-to-batch differences in their concentration, sometimes the differences can be as high as 50%. Some example results that we measured previously using this protocol are summarized in the table: Stable Readout Time Window Before Probe Clearance The probes rely on fast reversible chemical sensing reaction for GSH quantification. Due to the nature of reaction and cell intrinsic clearance pathways, long-term incubation leads to probe clearance and unstable readouts. We recommend keeping the readout time short and consistent whenever possible. The stability of signals in different conditions that we have tested are summarized in the table: Cell Type Subcellular Location

Drug Treatment Expected Concentration
Long-term tracking of GSH using RT-based probes may not be ideal in certain cell lines due to the clearance of probes through ABC transporters, especially in some cancer cell lines. We recommend multiple short time-lapsed imaging for cells with fast probe clearance.
Certain cells may experience significant outer membrane accumulation of free probe causing strong background. Higher spatial resolution or certain washing and masking steps may be helpful.
Laser-based technology with high spatial resolution is generally required to achieve high signal to noise and good quantitative results. Generally, epi-fluorescence microscopy may not distinguish signals from the background, unless specific optical settings or algorithm are adopted.
Ratiometric quantification relies on the corresponding standard curve. In some cases, the standard curve is unavailable or difficult to obtain. Therefore, semi-quantitative or qualitative measurement can be used.

TROUBLESHOOTING Problem 1
Signals primarily accumulate on cell outer membrane.

Potential Solution
Please check if the cells have been passaged for too many times. Generally, this could be avoided using low passage cells and washing thoroughly each time during passage. For certain cells with extremely sticky membrane, additional washing steps may be helpful. Alternatively, adding nonpermeable fluorescent quenching reagent may also help, such as trypan blue (Fig. 1). However, it will lower the overall signal to noise and alter the quantification curve.

Problem 2
No signals can be detected. The two channels have different propensity of lingering signal to the outer membrane, thus will lead to artifacts. Trypan blue partially resolves this issue but will also change the raw ratio, resulting in the need for a new calibration curve for quantification.

Potential Solution
First, check the fluorescent channel selection to make sure the blue (Pacific Blue, 405 excitation) and green (PE, 488 excitation) channels are selected correctly; second, check the gain setting for each channel (including FSC and SSC) and adjust slowly across the whole dynamic range. Contact flow cytometry service personnel when necessary.

Problem 3
No signal differences between samples.

Potential Solution
Use positive control (200 mM GSH-ester treated cells) and negative control (500 mM BSO treated cells) to acquire data and compare ratios with non-treated cells. HeLa cells are advised to be used as a control cell line for pilot study, as we have extensively demonstrated different RT-based probes can be successfully applied for monitoring GSH changes under various conditions, and that HeLa cells are readily available in most biology labs. Prepare fresh staining solution for measurements. Check if cysteine free medium should be used for the low GSH group.

Problem 4
Inconsistent signals between repeats.

Potential Solution
Certain cell lines (especially cancer cell lines) suffer from fast probe clearance, resulting in loss of signal in a relatively short period. To test signal acquisition time window, try a different staining time between 5 and 30 min, then measure the same sample 0, 10, 20, and 30 min post staining with acquisition time <2 min for each measurement. If none of the conditions can provide stable signals, try lowering the temperature to prevent probe clearance, or by adding ABC transporter inhibitors such as GF120918 or probenecid.

Problem 5
Standard curve quantification does not correlate with biochemical/HPLC assay.
Potential Solution BSO may deplete GSH differently across different cell types, therefore, different dilution factors should be tested when testing on a new cell type. Cell volume estimation may introduce the biggest error in the calculation due to the relatively high variance with the current method.
Alternatively, a semi-quantitative standard curve may be used in which the RT-based measurement can be correlated with biochemical/HPLC direct readings.

RESOURCE AVAILABILITY Lead Contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact: Jin Wang, wangj@bcm.edu

Materials Availability
All materials are commercially available unless otherwise specified.

Data and Code Availability
All original data provided by this study is available through the lead contact.