Purification and Application of Genetically Encoded Potassium Ion Indicators for Quantification of Potassium Ion Concentrations within Biological Samples

Vital cells maintain a steep potassium ion (K+) gradient across the plasma membrane. Intracellular potassium ion concentrations ([K+]) and especially the [K+] within the extracellular matrix are strictly regulated, the latter within a narrow range of ~3.5 to 5.0 mM. Alterations of the extracellular K+ homeostasis are associated with severe pathological alterations and systemic diseases including hypo- or hypertension, heart rate alterations, heart failure, neuronal damage or abnormal skeleton muscle function. In higher eukaryotic organisms, the maintenance of the extracellular [K+] is mainly achieved by the kidney, responsible for K+ excretion and reabsorption. Thus, renal dysfunctions are typically associated with alterations in serum- or plasma [K+]. Generally, [K+] quantifications within bodily fluids are performed using ion selective electrodes. However, tracking such alterations in experimental models such as mice features several difficulties, mainly due to the small blood volume of these animals, hampering the repetitive collection of sample volumes required for measurements using ion selective electrodes. We have recently developed highly sensitive, genetically encoded potassium ion indicators, the GEPIIs, applicable for in vitro determinations of [K+]. In addition to the determination of [K+] within bodily fluids, GEPIIs proved suitable for the real-time visualization of cell viability over time and the exact determination of the number of dead cells.


INTRODUCTION
Potassium ions (K + ) are fundamentally involved in multiple cellular and systemic functions (Ceccarelli, Fesce, Grohovaz, & Haimann, 1988;Larkin, Brown, Goldstein, & Anderson, 1983;Page & Di Cera, 2006;Toda, 1969). The maintenance of extracellular K + concentrations ([K + ] ex ) is especially essential and, thus, is tightly regulated in a range between 3.5 and 5.0 mM in humans (Thier, 1986). Consequently, it is not surprising that disturbances in K + homeostasis are associated with severe pathological alterations, including hypo-or hypertension, heart rate alterations, heart failure, neuronal damage, or skeletal muscle dysfunction (Antunes et al., 2014;Entz et al., 2016;Haddy, Vanhoutte, & Feletou, 2006;Kardalas et al., 2018;Lehnhardt & Kemper, 2011;Sica et al., 2002). Typically, the measurement of K + concentrations within human serum or plasma samples represents a standard clinical procedure determined by ion-selective electrodes (ISEs; Rastegar, 1990). Such measurements allow tracking and tightly controlling [K + ] ex in individuals, and often serve as an indicator of renal damage (Kunis & Charney, 1981). However, [K + ] measurements using ISEs generally require sample volumes of several milliliters. Although this is not critical in humans, such sample volumes become a critical parameter when working with small laboratory animals such as mice, which possess very small blood volumes of less than 2 ml (Riches, Sharp, Thomas, & Smith, 1973). To develop pharmacologically active substances that might have an impact on blood K + levels in mammals, tracking extracellular K + dynamics in small laboratory animals is inevitable and requires alternative applicable approaches for K + quantifications.
In addition to [K + ] ex , the intracellular [K + ] ([K + ] in ) is also very strictly controlled (Checchetto, Teardo, Carraretto, Leanza, & Szabo, 2016;Palmer, 2015). Vital cells maintain a steep K + gradient across the plasma membrane in order to keep the electrochemical gradient, which is important for numerous cellular functions (Ceccarelli et al., 1988;Larkin et al., 1983;Page & Di Cera, 2006). It has recently been demonstrated that [K + ] ex within the tumor microenvironment is elevated due to K + release from necrotic cells (Eil et al., 2016). The measurement of [K + ] ex might represent a valuable alternative to available cell viability assays, allowing the visualization of cell death with high spatial and temporal resolution in vitro or in vivo. Here, we show that quantification of [K + ] ex in the supernatant of cultured cells allows time-resolved visualization of cell death.
We have recently developed a highly sensitive Förster resonance energy transfer (FRET)based, genetically encoded K + indicator referred to as GEPII 1.0 ( Fig. 1A; Bischof et al., 2017). Subcloning of the GEPII 1.0 coding sequence into a vector for bacterial expression (pETM11; Fig. 1B) allowed purification of the recombinant protein. The use of the recombinant purified protein enables quantification of [K + ] within various biological samples. Our data highlight the suitability of GEPIIs for repetitive [K + ] measurements within samples of small laboratory animals that are as precise as the gold-standard instrument, the ISE, and use only a small drop of blood from the animal. In addition, the probe proved suitable for online visualization of cell viability within the supernatant of mammalian cells (Bischof et al., 2017).
Basic Protocol 1 describes the transformation, expression, and purification of recombinant GEPII 1.0 from Escherichia coli, yielding functional K + probes in a K + -free solution. Basic Protocol 2 deals with the collection of murine blood samples, the preparation of serum from the blood, and the application of recombinant purified GEPII 1.0 for quantification of [K + ] within tiny volumes of these samples. Basic Protocol 3 describes the use of the recombinant purified protein for online visualization of cell viability and growth. Finally, Basic Protocol 4 demonstrates how to estimate the number of dead cells using GEPII 1.0.
Bischof et al.

of 22
Figure 1 Functional principle and plasmid map of GEPII 1.0. (A) Schematic representation of the K + -sensing mechanism of GEPII 1.0. mseCFP (cyan), wild-type Kbp (grey), and circularly permuted Venus (yellow) are shown. (B) Plasmid map of GEPII 1.0 subcloned into pETM11 vector for bacterial expression. mseCFP (cyan), wild-type Kbp (wt Kbp, grey), and circularly permuted Venus (cp173Vens, yellow) as well as the most important features of the plasmid are indicated in the map. Single-cutting restriction enzymes (EcoRV, HindIII, KpnI NcoI, XbaI, and XhoI, all in bold) as well as internal restriction sites with multiple cutting sites (ClaI and EcoRI) are indicated. Locations of the 6× His-Tag and a TEV protease site located between the His-Tag and mseCFP are shown.

EXPRESSION AND PURIFICATION OF RECOMBINANT GEPII 1.0
The following protocol describes transformation of the plasmid encoding GEPII 1.0 into chemically competent E. coli cells and subsequent expression and purification of GEPII 1.0. Purification according to this protocol will yield 3-5 mg of highly pure and functional GEPII 1.0, suitable for quantification of [K + ] in various biological samples.
Prepared glycerol stocks can be used in the future to start directly with induction.

Induce protein expression
NOTE: From this point on, avoid excessive light exposure at all steps to prevent photobleaching of the fluorescent proteins.
10. Add 1 mM IPTG and incubate for 4-6 hr at 20°C with shaking to induce protein expression.
After 4-6 hr, bacteria should have a slightly green to yellow color caused by expression of GEPII 1.0.
11. Transfer bacterial suspension to centrifuge tubes and pellet cells at 7,800 × g (6,000 rpm in Sorvall LYNX Superspeed centrifuge) for 10 min at 4°C.

Remove supernatant and resuspend pelleted cells in 15 ml lysis buffer.
If desired, resuspended cells can be frozen at −80°C. Frozen cells should be thawed in a 37°C water bath before use.
Prepare cell lysate 13. Add 250 U Benzonase nuclease and 150 µl bacterial protease inhibitor cocktail (1:100 dilution) to the cell suspension and incubate on ice for 30 min.
14. Break up cells by sonicating for 20 min on ice. Perform 10 s of sonication followed by 10 s of pause over a period of 20 min. 20. Allow the lysate to flow through the column. Perform size-exclusion chromatography NOTE: Other FPLC systems may be used, but the protocol may vary depending on the system.
25. Choose the "inlet" and "column position" and define the "system flow". Set the column position at "Bypass" to avoid applying any impurities onto the column during washing.
Inlet defines the tube that is connected to the equilibration buffer (SEC buffer). System flow is set to 0.5 ml/min. 26. Wash the selected pump system with SEC buffer.
27. Under "monitors", select "wavelengths" to define the wavelengths that are shown on the screen. Set wavelength to 480 nm to screen for the yellow fluorescent GEPII 1.0 protein.
28. Set "alarms" according to the used column. Define the pre-column pressure and the delta column pressure.
29. Press "execute" to wash the system without the column.
30. Select the "manual load" option on the injection loop and apply 5 ml SEC buffer with a syringe onto the loop. 31. Switch from "manual load" to "inject" to wash the loop.
Every loop has a defined volume that is retained in the loop. Any excess volume will flow into the waste. For 500 µl protein solution, use a 1-ml loop to ensure that the protein is not lost during application.
Bischof et al.

of 22
32. Change back to "manual load" on the loop and then remove the syringe.
Do not remove the syringe without changing back to "manual load" or air will be trapped in the system. 33. Change from the "bypass" position to the actual column position and allow 30 ml SEC buffer to flow through the column to wash and equilibrate it.
All solutions and buffers will now pass through the column.
34. With "manual load" selected in the software, use a small syringe to load protein solution from the concentrator tubes onto the loop. Be careful to avoid any air bubbles.
35. Change to "inject" to apply the protein onto the column. After ß6 ml, start the fractionation to collect fractions containing protein.
36. Combine fractions containing protein and determine the protein concentration using a method of choice.
37. Calculate the molarity of the protein solution. 42. Determine the average FRET ratio value from the blank wells and subtract it from the value for each well containing K + calibration solution. Plot the blank-corrected FRET/CFP ratios ( Ratio) against [K + ] both linearly and logarithmically and determine the EC 50 .
If GEPII 1.0 is functional, a concentration-dependent increase in the FRET ratio should be observed, with an EC 50 of ß500-600 µM at room temperature, as shown in Figure 2.
43. If GEPII 1.0 has proven functional, prepare aliquots as desired and shock-freeze in liquid nitrogen. Store up to 2 years at −80°C in the dark. If protein has been stored for >1 year, it is recommended to repeat the functionality tests (steps 38-42) before use.

COLLECTION OF MURINE SERUM SAMPLES AND DETERMINATION OF SERUM K + CONCENTRATION USING GEPII 1.0
The following protocol describes how to collect murine blood by phlebotomy from v. facialis and how to prepare murine serum samples. Furthermore, presents the generation of a K + calibration curve for GEPII 1.0 and the determination of [K + ] in murine serum samples using GEPII 1.0.
Generally, mice (e.g., wild-type C57BL/6) should be maintained in a clean environment with a regular light-dark cycle (12 hr/12 hr) and unrestricted access to food and water. The mouse strain as well as conditions can be modified as needed to suit the experimental design.

Collect blood and prepare serum samples
1. On the day of sample collection, remove mouse from its environment and restrain it securely by grasping the neck scruff with the thumb and index finger and holding the base of the tail between the palm and the ring finger.
2. Puncture the superficial temporal vein using a sterile single-use lancet.
4. Lightly press the puncture site with a sterile gauze for a few seconds to cause hemostasis. Release mouse from restraint and return to its home cage.
Bischof et al.

of 22
5. Repeat for the desired number of mice.
6. Allow blood samples to clot for 15 min at room temperature.
7. Centrifuge samples for 10 min at 400 × g to separate the blood clot from the serum. Carefully transfer serum to a new 1.5-ml microcentrifuge tube, taking care not to disturb the pelleted blood clot, and place on ice.  Figure 3A.
13. Calculate the FRET/CFP ratios of the blank wells, determine the average value, and then subtract that from each value of the calibration curve. Plot the blank-corrected FRET/CFP ratios ( Ratio) against [K + ] as demonstrated in Figure 3B, and fit the values using a one-phase decay. Create the equation of the curve and solve it for x as demonstrated in Figure 3C.  20. Determine the average for the blank wells and subtract it from each sample and standard.
21. Verify the accuracy of the internal standards (K + standard solutions) by calculating [K + ] in mM using the formula generated from the calibration curve, as shown in Figure 3C.   Example data from sera of six animals are presented in Figure 4.

MEASURING EXTRACELLULAR [K + ] WITH GEPII 1.0 FOR VISUALIZATION OF CELL VIABILITY OVER TIME
The following protocol describes how to apply recombinant GEPII 1.0 for online visualization of cell viability over time by measuring extracellular [K + ]. NOTE: All of the following steps should be performed in a sterile environment. All buffers and solutions used on cells should be prewarmed to 37°C.

Prepare cells
1. Cultivate cells of interest to a confluency of ß80% using appropriate culture medium and a humidified incubator at 37°C with 5% CO 2 . The initial number of cells needed depends on experimental need. Assuming one 10-cm dish at ß80% confluency has ß10 million cells, this is sufficient for ß200 wells (or two 96-well plates).
2. One day before the experiment, wash cells with PBS and trypsinize them. For a 10-cm dish, use ß3 ml trypsin solution and incubate 2-4 min in a humidified 37°C incubator.
3. Suspend floating cells in 10 ml medium to stop trypsinization.
5. Remove supernatant and wash cells with 10 ml PBS, centrifuging again at 200 × g for 10 min.
6. Remove PBS, carefully resuspend cell pellet in 10 ml supplemented RPMI 1640, and place on ice to prevent cell adherence.
7. Determine cell density (number/ml) using a Bürker-Türk counting chamber or any other method of choice. Adjust to ß250,000 cells/ml. 13. At the end of measurement, add 5 µl of 850 µM digitonin stock solution to all wells including blanks (final 50 µM).

Application of digitonin will permeabilize all cell membranes and is important in order to ensure the same cell numbers under all conditions.
14. Calculate the FRET ratio signal of GEPII 1.0 by determining the FRET/CFP fluorescence (525 nm/475 nm) for all wells including blanks.
15. Perform blank correction of FRET ratio signals from wells containing cells using the respective blanks. If several wells were used as a blank, calculate the average blank value and subtract it from each well of the respective condition.
16. Analyze and present the FRET ratio signal of GEPII 1.0 over time as demonstrated in Figure 5.

GENERATION OF A GEPII 1.0 CALIBRATION CURVE FOR ESTIMATING THE NUMBER OF DEAD CELLS
The following protocol describes how to quantify the number of dead cells using recombinant GEPII 1.0 for quantification of extracellular [K + ]. Refer to Basic Protocol 3 for all materials needed.

Prepare cells
1. Cultivate cells of interest to a confluency of ß80% using the appropriate culture medium and a humidified incubator at 37°C with 5% CO 2 .
It is critical to note that the determined regression curve and formula are cell-type specific and are not applicable for other cell lines than the one used to generate them. The sample data in Figure 6 were generated using HeLa cells cultured in DMEM (see recipe).
2. On the day of the experiment, wash cells with PBS and trypsinize them. For a 10-cm dish, use ß3 ml trypsin solution and incubate 2-4 min in a humidified 37°C incubator.
3. Suspend floating cells in 10 ml medium to stop trypsinization.
5. Remove supernatant and wash cells with 10 ml cell wash buffer, centrifuging again at 200 × g for 10 min.
Bischof et al.

of 22
6. Remove supernatant and carefully resuspend cells in 5 ml cell wash buffer.
7. Determine cell density (number/ml) using a Bürker-Türk counting chamber or any other method of choice.

Count cells as precisely as possible, as deviations in the assumed cell number will affect the estimated number of dead cells for Basic Protocol 3.
Perform assay 8. Dispense increasing cell numbers into the wells of a black 96-well cell culture plate with clear bottom and adjust the volume of each well to 40 µl using cell wash buffer. Include blank wells containing 40 µl cell wash buffer without cells.
We recommend performing at least triplicates for each cell number tested. Appropriate cell numbers range from 1,000 to 200,000, depending on cell type. For the example in Figure 6, we used 0,2,500,5,000,10,000,20,000,40,000,and 80,000 HeLa cells. 9. Add 35 µl cell assay solution to all wells.
10. Add 5 µl of 850 µM digitonin stock solution to all wells (final 50 µM) and incubate for 10 min for full permeabilization of all cells.
Application of digitonin will permeabilize all cell membranes and allow determination of the K + released from defined cell numbers.   13. Perform blank correction of FRET ratio signals from wells containing cells using the respective blanks. If several wells were used as a blank, calculate the average blank value and subtract it from each well of the respective condition.
14. Analyze and present the FRET ratio signal of GEPII 1.0 as demonstrated in Figure 6A.
15. Fit values using a proper regression, create the equation of the curve, and solve for x as demonstrated in Figure 6B.
The formula allows calculation of cell number from the [K + ] and can be applied to data obtained in Basic Protocol 3.

Glycerol solution, 50% (v/v)
Dilute an equal amount of glycerol (Carl Roth, cat. no. 6967.1) in distilled water and autoclave. Store at room temperature (stable at least 1 year if kept sterile).

of 22
Autoclave Store at room temperature (stable at least 6 months)

Serum assay solution
Dilution buffer (see recipe) 400 nM GEPII 1.0 (see Basic Protocol 1) Prepare fresh daily Keep on ice in the dark

Background Information
The first successful design of a FRETbased probe based on a yellow and cyan fluorescent protein (FP) variant was reported in 1997 by Miyawaki and colleagues, who introduced a Ca 2+ -sensitive FRET-based sensor (Miyawaki et al., 1997). Today, a wide variety of these FRET-based indicators is available, including sensors for various metal ions, pH, cell metabolites, or even small molecules, often having short half-lives Burgstaller et al., 2019;Imamura et al., 2009;Tanimura, 2011). All of these probes are based on diverse FP variants, enabling a FRET-based read-out, fused to a specific analyte-binding domain, undergoing a conformational rearrangement upon analyte binding. In 2016, Ashraf et al. unraveled the function of the bacterial protein Kbp, formerly known as YgaU (Ashraf et al., 2016). They demonstrated that Kbp represents a K + -binding protein in E. coli that is important to ensure normal growth of the bacteria under conditions of high extracellular [K + ]. Upon K + binding, Kbp undergoes a huge conformational rear-rangement, from an elongated conformation towards a spherical one (Ashraf et al., 2016).
Based on this protein, we recently developed a series of genetically encoded K + indicators, the GEPIIs (Bischof et al., 2017). These GEPIIs consist of a cyan and a yellow fluorescent protein, namely monomeric super enhanced CFP (mseCFP) and circularly permuted Venus (cpV), a well-characterized FP FRET pair, fused to the N and C terminus of Kbp, respectively. In the absence of K + , FRET efficiency is low, yielding high donor fluorescence. However, upon K + binding to the construct, the protein undergoes a conformational rearrangement, leading to increased FRET and decreased cyan emission (Bajar, Wang, Zhang, Lin & Chu, 2016). While several mutated GEPII variants showed K + affinities suitable for intracellular K + measurements, the GEPII variant referred to as GEPII 1.0, containing the wild-type Kbp, showed a very high affinity and specificity for K + . Based on this high sensitivity, we hypothesized that the recombinant GEPII 1.0 protein represents a valuable tool for quantification of extracellular [K + ] in various Bischof et al. biological samples. Our data emphasized that GEPII 1.0 is able to determine [K + ] in serum and urine samples of healthy and diseased human donors as precisely as the gold-standard method for [K + ] measurements, the ISE. The use of GEPII 1.0 for quantification of [K + ] in body fluids requires only a fraction of the sample volume required for determination by ISE. Thus, we exploited this high sensitivity and accuracy for quantification of [K + ] in murine serum, urine, and even bile samples, as mice possess very limited amounts of these biological fluids. Using GEPII 1.0 for these measurements will in future allow repetitive sample collection from one given animal over time, without need for its sacrifice. In this work, we provide scientists a detailed step-by-step manual for purifying the recombinant GEPII 1.0 protein, preparing murine serum samples, and quantifying serum [K + ] using GEPII 1.0.
Furthermore, our recent data demonstrated the suitability of GEPII 1.0 for online visualization of cell viability over time by measuring extracellular [K + ]. As vital cells maintain a steep K + gradient towards the plasma membrane, the measurement of extracellular [K + ] represents a facile method to visualize cell viability with high temporal resolution, without the need of expensive chemicals, which often allow only end-point measurements. We also demonstrate an example for estimating the number of dead cells using this K + -sensitive, FRET-based probe.
To our knowledge this is the first time that a genetically encoded, FRET-based biosensor has been applied as a recombinant, purified protein for the quantification of an analyte within biological samples.

Critical Parameters and Troubleshooting
One of the most critical parameters of these protocols is the purification of recombinant GEPII 1.0 from E. coli. It is essential to test the functionality of recombinant GEPII 1.0 after purification to demonstrate and ensure the functionality of the probe for reporting and responding in a concentration-dependent manner to increasing [K + ].
It is of utmost importance to use highquality distilled water and ultrapure graded salts when eluting the protein from the sizeexclusion column, as pre-saturation of the recombinant probe will drastically affect the dynamic range of the sensor. To quantitate and identify a possible pre-saturation with K + , one may include wells containing K + chelators such as poly(sodium 4-styrenesulfonate) . Typically, 50 µM concentrations of a given chelator are capable of buffering at least 20-30 mM of K + . In cases of a drastically reduced GEPII 1.0 FRET ratio signal in wells containing the chelator, GEPII 1.0 is pre-saturated with K + .
The primary reason for K + impurity of the recombinant protein solution might be the use of contaminated SEC buffer or impurities on the size-exclusion column itself. Under such circumstances, one can restart the protein purification, increase the buffer volume in the wash step prior to size exclusion, and control the recipe and chemicals used for preparing the SEC buffer. Another possibility is to perform desalting protocols after size-exclusion chromatography.
The exact determination of [K + ] within biological samples using this GEPII 1.0 from bacterial expression requires the generation of a K + calibration curve. Importantly, the calibration solutions and protein-containing solutions must be prepared as precisely as possible. Improper preparation of the calibration curve will cause incorrect calculations of [K + ] from all samples. Determining the [K + ] of diverse K + standard solutions can assist in the identification of pipetting, dilution, or calculation errors.
As the K + sensitivity of GEPII 1.0 appears temperature dependent in vitro, [K + ] measurements need to be performed at constant temperature settings of the fluorescence plate reader. In principle, the K + sensitivity decreases with increasing temperature and vice versa; thus, measuring at room temperature or below leads to lowered detection limits of GEPII 1.0 for K + .
For additional troubleshooting, see Table 1.

Understanding Results
Mammalian organisms tightly control their extracellular [K + ] to ensure proper function of all cell types and organs (Palmer, 2015). Alterations of the extracellular [K + ] are associated with severe pathological alterations and mostly require urgent medical treatment. Frequently, renal dysfunction and insufficiencies are associated with increased serum K + levels (Kunis & Charney, 1981), and thus the measurement of serum [K + ] often serves as an indication for renal disorders. Using GEPII 1.0, we have demonstrated drastic differences in the extracellular [K + ] of healthy control mice and mice suffering surgically inflicted ischemia reperfusion injury. Our data emphasize that the measurement  Because vital cells maintain a steep K + gradient across the plasma membrane, GEPII 1.0 can further be used for visualizing cell viability over time (Bischof et al., 2017;Palmer, 2015). Cells undergoing cell death release K + into their extracellular environment, which can be measured using GEPII 1.0. The use of the recombinant protein in cell supernatants thereby allows online visualization of cell death over time in the presence of different compounds or their respective vehicle controls, which is not easily feasible using standard cell viability assays that represent end-point measurements rather than online assays. Additionally, the signal received from different cells can be calibrated by using defined cell numbers, allowing a calculation of real cell numbers that underwent cell death.

Time Considerations
Basic Protocol 1 takes a total of ß29 hr. This includes an overnight incubation (ß12-hr) for bacterial transformation with GEPII 1.0, inoculation of LB medium and induction of protein expression (ß8 hr), protein purification from the cell lysate (ß8 hr), and functionality testing of the purified GEPII 1.0 (ß1 hr).
Basic Protocol 2 takes ß2 hr, including collection of murine blood samples and preparation of serum (ß30 min), generation of the K + calibration curve for GEPII 1.0 (ß30 min), and determination of [K + ] in murine serum samples (ß1 hr).
Basic Protocol 3 takes at least 26.5 hr. Trypsinization, counting, and seeding of cells in 96-well plates takes ß1.5 hr. Cultivation of cells in 96-well plates takes ß24 hr, and preparation of the cells for the cell viability assay takes ß1 hr. The duration of the experiment depends on the research question.