Simple methods to detect the selective activation of G proteins by G protein-coupled receptors remain an outstanding challenge in cell signaling. Here, Fӧrster resonance energy transfer (FRET) biosensors have been developed by pairwise tethering a GPCR to G protein peptides to probe conformational changes at controlled concentrations in live cells.
Fӧrster resonance energy transfer (FRET)-based studies have become increasingly common in the investigation of GPCR signaling. Our research group developed an intra-molecular FRET sensor to detect the interaction between Gα subunits and GPCRs in live cells following agonist stimulation. Here, we detail the protocol for detecting changes in FRET between the β2-adrenergic receptor and the Gαs C-terminus peptide upon treatment with 100 µM isoproterenol hydrochloride as previously characterized1. Our FRET sensor is a single polypeptide consisting serially of a full-length GPCR, a FRET acceptor fluorophore (mCitrine), an ER/K SPASM (systematic protein affinity strength modulation) linker, a FRET donor fluorophore (mCerulean), and a Gα C-terminal peptide. This protocol will detail cell preparation, transfection conditions, equipment setup, assay execution, and data analysis. This experimental design detects small changes in FRET indicative of protein-protein interactions, and can also be used to compare the strength of interaction across ligands and GPCR-G protein pairings. To enhance the signal-to-noise in our measurements, this protocol requires heightened precision in all steps, and is presented here to enable reproducible execution.
G-protein-coupled receptors (GPCRs) are seven-transmembrane receptors. The human genome alone contains approximately 800 genes coding for GPCRs, which are activated by a variety of ligands including light, odorants, hormones, peptides, drugs and other small molecules. Nearly 30% of all pharmaceuticals currently on the market target GPCRs because they play a large role in many disease states2. Despite decades of extensive work done on this receptor family, there remain significant outstanding questions in the field, particularly with regards to the molecular mechanisms that drive GPCR-effector interactions. To date, only one high-resolution crystal structure has been published, providing insight into the interaction between the β2-adrenergic receptor (β2-AR) and the Gs protein3. Together with extensive research in the last three decades, it reiterates one specific structural component that is critical in this interaction: the Gα subunit C-terminus. This structure is important for both G protein activation by the GPCR4 and G protein selection5-6. Hence, the Gα C-terminus provides a crucial link between ligand stimulation of the GPCR and selective G protein activation.
Research over the last decade suggests that GPCRs populate a broad conformational landscape, with ligand-binding stabilizing subsets of GPCR conformations. While several techniques, including crystallography, NMR and fluorescence spectroscopy, and mass spectrometry are available to examine the GPCR conformational landscape, there is a paucity of approaches to elucidate their functional significance in effector selection7. Here, we outline a Fӧrster resonance energy transfer (FRET)-based approach to detect G protein-selective GPCR conformations. FRET relies on the proximity and parallel orientation of two fluorophores with overlapping emission (donor) and excitation (acceptor) spectra8. As the donor and acceptor fluorophores come closer together as a result of either conformational change in the protein or a protein-protein interaction, the FRET between them increases, and can be measured using a range of methods8. FRET-based biosensors have been employed extensively in the GPCR field9. They have been used to probe conformation changes in the GPCR by inserting donor and acceptor in the third intracellular loop and GPCR C-terminus; sensors have been designed to probe GPCR and effector interactions by separately labeling the GPCR and effector (G protein subunits/arrestins) with a FRET pair10; some sensors also detect conformational changes in the G protein11. These biosensors have enabled the field to ask a multitude of outstanding questions including conformational changes in the GPCR and effector, GPCR-effector interaction kinetics, and allosteric ligands12. Our group was particularly interested in creating a biosensor that could detect G protein-specific GPCR conformations under agonist-driven conditions. This biosensor relies on a recently developed technology named SPASM (systematic protein affinity strength modulation)13. SPASM involves tethering interacting protein domains using an ER/K linker, which controls their effective concentrations. Flanking the linker with a FRET pair of fluorophores creates a tool which can report the state of the interaction between proteins12. Previously1 the SPASM module was used to tether the Gα C-terminus to a GPCR and monitor their interactions with FRET fluorophores, mCitrine (referred to in this protocol by its commonly known variant, Yellow Fluorescent Protein (YFP), excitation/emission peak at 490/525 nm) and mCerulean (referred to in this protocol by its commonly known variant Cyan Fluorescent Protein (CFP), excitation/emission peak 430/475 nm). From N- to C-terminus, this genetically encoded single polypeptide contains: a full length GPCR, FRET acceptor (mCitrine/YFP), 10 nm ER/K linker, FRET donor (mCerulean/CFP), and the Gα C-terminus peptide. In this study, sensors are abbreviated as GPCR-linker length-Gα peptide. All components are separated by an unstructured (Gly-Ser-Gly)4 linker which enables free rotation of each domain. The detailed characterization of such sensors was previously performed using two prototypical GPCRs: β2-AR and opsin1.
This sensor is transiently transfected into HEK-293T cells and fluorometer-based live cell experiments measure fluorescence spectra of the FRET pair in arbitrary units of counts per second (CPS) in the presence or absence of ligand. These measurements are used to calculate a FRET ratio between the fluorophores (YFPmax/CFPmax). A change in FRET (ΔFRET) is then calculated by subtracting the average FRET ratio of untreated samples from the FRET ratio of ligand treated samples. ΔFRET can be compared across constructs (β2-AR-10 nm-Gαs peptide versus β2-AR-10 nm-no peptide). Here, we detail the protocol to express these sensors in live HEK-293T cells, monitor their expression, and the setup, execution, and analysis of the fluorometer-based live cell FRET measurement for untreated versus drug treated conditions. While this protocol is specific for the β2-AR-10 nm-Gαs peptide sensor treated with 100 µM isoproterenol bitartrate, it can be optimized for different GPCR-Gα pairs and ligands.
1. DNA Preparation
2. Cell Culture Preparation
3. Transfection Conditions
4. Reagent and Equipment Preparation
Figure 2. Microcentrifuge Tube Set Up and Position Reference in Heat Block. Cuvette for untreated samples is in position 1; cell aliquot tubes are in positions 2 – 6. Cuvette for drug treated samples is in position 7; cell aliquot tubes are in positions 8 – 12. Please click here to view a larger version of this figure.
5. Experiment & Data Collection
Figure 3. Experimental Schematic. A detailed step-wise guide for experimental set up and execution. Please click here to view a larger version of this figure.
6. Data Analysis
A generalized schematic of the experiment set up and execution is detailed in Figure 3.
In order to detect a FRET change in the narrow dynamic range of the sensor, it is critical to adhere to the nuances of the system be adhered to. Cell quality is imperative to protein expression as well as consistency in sampling. Figure 1 features images of cultured cells growing in a consistent monolayer (10X) that is optimal for six-well plating and transfection Figure 1 (a) and cells growing in clumped patterns that lead to dendritic shapes Figure 1 (b) which is not recommended for consistent plating. Transfection conditions can also be optimized in order to achieve reproducible expression. Several conditions have been optimized for the recommended transfection reagent used here. Table 1 details these conditions for reduced serum media. DNA, and transfection reagent ratios.
Once all data has been collected and data entered into the CSV file for analysis (see sample in Table 3), the generated results will resemble the Raw FRET spectra data shown in Figure 4 (a) and the normalized, mean FRET spectra shown in Figure 4 (b). In Figure 4 the red spectra are the untreated samples and the blue are the drug treated samples. All of the Raw FRET spectra in Figure 4 (a) have sufficient signal-to-noise range for consistent data analysis (i.e., CPS at 450 nm compared to CPS at 475 nm). With this range, the spectra are smoother and the water peak from the sample (Raman peak is at 500 nm) is also minimal. Low expression levels result in a prominent water peak that interferes with data analysis. Data is normalized at 475 nm, setting the CPS of this value to 1.0 (Figure 4 (b)). In this data set for the β2-AR-10 nm-Gαs peptide sensor, there is a significant change at the 525 nm reading between untreated (red) and treated (blue) samples. The ΔFRET change is calculated from the FRET ratios (525 nm/475 nm) of these data sets and are accessible through the OUTPUT file.
If protein expression is low, there is poor transfection efficiency, or low cell density in the cuvette for fluorescence reading, spectra may appear noisier, as shown in Figure 5. Compared to Figure 4 (a) the signal-to-noise range for this data set is not ideal, at approximately 1.0 – 1.6, with CPSmax of 4 x 105. This low expression level contributes to the jagged spectra seen in the raw data Figure 5 (a), however the data is tight and able to be normalized Figure 5 (b). Though the water peaks (~ 500 nm) does not line up completely between data sets Figure 5 (b) this experiment is still usable for analysis. Figure 6 is representative of an experiment that is inadequate for analysis. While the signal-to-noise of the sample is approximately 2 – 3 for treated (blue) and untreated (red) samples, cell density in the cuvette across samples is too low (450 nm value of 1 x 105) Figure 6 (a). This creates an issue in background subtraction and normalization Figure 6 (b) and the spectra do not align. Subtraction can be adjusted for samples by increasing or decreasing OD in Table 3. However, even with low cell density, the water peak (~ 500 nm) becomes much larger and noisier Figure 6 (c) and inhibits this data set from further analysis.
Figure 1. Example of Cultured Cell Growth. Cells growing in a monolayer (a) are ideal for consistent plating and transfection. Cells that appear to grow in clumps or with dendritic patterns (b) may not plate consistently in six-wells, may display poor transfection efficiency, and may create inconsistencies in amplitude of the FRET spectrum. Scale bar = 100 µm. Please click here to view a larger version of this figure.
Figure 4. Representative Data Analysis for β2-AR-10 nm-Gαs Peptide ± Drug. Representative image of raw data (a) and normalized, averaged data (b) collected with the β2-AR-10 nm-Gαs peptide sensor with untreated (red) samples and treated (blue) samples after 5-minute incubation with 100 µM isoproterenol bitartrate. CFPmax emission at 475 nm, YFPmax emission at 525 nm. Please click here to view a larger version of this figure.
Figure 5. Analysis of Poor Protein Expression with Raw and Normalized Data. Noisy raw data spectra (a) are a result of low expression levels, low cell density per sample, and/or poor transfection efficiency of construct. This data set is still interpretable as normalized (b), although it is not ideal. Please click here to view a larger version of this figure.
Figure 6. Analysis of Low Cell Density in Fluorometer Cuvette with Raw and Normalized Data Sets. The low cell density per sample, seen in raw data (a) complicates water/background subtraction (b, c) and makes this data set uninterpretable. Please click here to view a larger version of this figure.
Condition | 0.5x | 0.7x | 1x | 2x | 2x* |
DNA | 1 μg | 1.4 μg | 2 μg | 4 μg | 4 μg |
Reduced serum media | 100 μl | 70 μl | 100 μl | 100 μl | 200 μl |
Transfection reagent | 3 μl | 4.2 μl | 6 μl | 6 μl | 12 μl |
*Note: Recommended for exceptionally difficult construct. Caution must be taken, as this condition induces high cell death. |
Table 1. Transfection Conditions. This table includes the optimized conditions of reagents for transfections into 2 ml wells of HEK-293T cells using the recommended transfection reagent.
Cell Buffer | ||
Volume | Reagent | Comments |
9 ml | ultrapure DNase/RNase free water | |
1 ml | HBS (10x), pH 7.4, store at 4 °C: | |
200 mM HEPES | ||
50 mM KCl | ||
450 mM NaCl | ||
20 mM CaCl2 – H2O | ||
10 mM MgCl2 – H2O | ||
100 μl | 20% D-glucose | |
15 μl | aprotinin (1 mg/ml in dH2O) | prevent degradation |
15 μl | leupeptin (1 mg/ml in dH2O) | prevent degradation |
100 μl | ascorbic acid (100 mM in dH2O)* | stabilize agonist |
*add immediately before beginning assay | ||
Drug Buffer | ||
Volume | Reagent | Comments |
9 ml | ultrapure DNase/RNase free water | |
1 ml | HBS (10x), pH 7.4, store at 4 °C: | |
200 mM HEPES | ||
50 mM KCl | ||
450 mM NaCl | ||
20 mM CaCl2 – H2O | ||
10 mM MgCl2 – H2O | ||
100 μl | ascorbic acid (100 mM in dH2O)* | stabilize agonist |
*add immediately before beginning assay |
Table 2. Buffer Constituents. This table details the reagents used to make both Cell Buffer and Drug Buffer for use in the experiment. Make both buffers fresh each day of the experiment; store Cell Buffer at 37 °C, and Drug Buffer at room temperature.
Table 3. Sample CSV File for Analysis. This sample data file highlights the entry set after one experiment. Multiple experiments can be entered into the same CSV file and can be discerned under the 'Additive' column if necessary. Each row must be filled out with the following information:
File name – individual SPC graph files Receptor – designate which GPCR construct was tested (e.g., Β2)
Binder – designate which peptide variant of the construct was tested (e.g., S)
Agonist – designate untreated (N) or drug treated (D) conditions
Directory – the path folder in which SPC files are saved, usually organized by date
OD – recorded optical density of sample in spectrophotometer
Please click here to view a larger version of this table.
The tight dynamic range of FRET measurements in this system reinforces the necessity of sensitive quality control in every step of this protocol. The most important steps to ensure a successful FRET experiment are 1) cell culturing, 2) transfection 3) protein expression and 4) timely, precise coordination during the assay execution.
Cell health and maintenance/plating quality can have a significant impact on the signal-to-noise of the experimental system and poor cell health can make it impossible to detect any consistent change in FRET. Conservatively, cells are healthiest for approximately 20 passages, though this may vary based on cell line, handling, and culture conditions. Once cells have difficulty growing as a confluent monolayer, or begin growing consistently in more dendritic patterns (see Figure 1 (b)), experimental background noise will be adversely affected. Careful cell maintenance, including routine media changes and regularly removing non-adherent cells and debris from maintenance plates, will enhance the quality of cells for six-well plating and transfections. Cell clumps, which adversely affect transfection efficiency, can be effectively separated into individual cells by trypsinization of maintenance plates: treat 10 cm confluent dishes with 10 ml of 0.25% trypsin for 30 sec, remove trypsin but leave approximately 200 μl, place dish in 37 °C incubator for 2 – 3 min. Cells will come off dish very easily and are less susceptible to clumping.
It is critical to optimize the transfection step for this experiment. Six-wells must ideally be 60 – 80% confluent for efficient transfection and optimal expression. If too few cells have adhered (< 60%), wait approximately 2 – 6 hr to transfect, or until at least 70% of cells are adhered. Six-wells that are over-confluent (> 80%) will also reduce transfection efficiency. Transfecting at lower cell confluency increases cell death rate. DNA concentration and purity are also critical (see Step 1.2). Using low concentration and/or poor quality DNA preparations negatively affect transfection efficiency Transfection conditions can be adjusted per construct, refer to Table 1 for more information.
Accurate and consistent monitoring of protein expression using a tissue-culture fluorescence microscope is another crucial step in this process. Though this step is subject to individual judgement, it is possible to use other techniques, such as microscopy, to monitor expression over time quantitatively, though they are not detailed here. For constructs we have tested successfully in our system, expression takes approximately 18-36 hr to reach optimal expression. In our experience, constructs that display poor expression during this time window rarely improve after 40 hr. The constructs we have published with have not shown signs of degradation, however this may be an issue for some GPCRs. In our assays, sensor degradation is possible at transfection times over 30 hr. Sensor integrity can be tested using YFP/CFP ratios: 525 nm reading from the YFP-excited (490 nm) spectrum, and 475 nm reading from CFP-excited (430 nm) spectrum. For settings, see step 4.6. Recommended YFP/CFP ratios are in the range of 1.7 – 2.0, with an ideal ratio of 1.8. This value is dependent on the approximate two-fold greater brightness of YFP relative to CFP14. An integral sensor with minimal degradation will therefore contain both fluorophores and have YFP:CFP ratio of approximately 2:1. After sensor quality has been confirmed it is important to confirm protein localization and expression at the plasma membrane. Significant intracellular expression could be a result of either protein degradation, internalization, or ongoing trafficking of the protein. Monitor constructs over time to see if expression is enhanced at the plasma membrane. The next critical challenge in expression is transfection efficiency. Approximately 70% + transfection efficiency is necessary for adequate signal-to-noise detection in this fluorometer system. If fewer cells are transfected, the amount of signal-to-noise may still be detectable but will be much less consistent between samples in one FRET experiment. This will hinder accurate data analysis within the narrow dynamic range of the system. Expression levels also present a significant hurdle to achieving ample signal-to-noise during the experiment. For expression levels detectable by the fluorometer, the ratio of signal between 475 nm and 450 nm (cell scattering) of 1.5 is sufficient for detecting a change in FRET, however optimal expression will have a ratio of approximately 2.0+. For reference, the β2-AR data is collected in a signal-to-noise range of 4 x 105 CPS (450 nm) to 1 x 106 CPS (475 nm), signal-to-noise ratio of 2.5. This ratio will help reduce the amount of variability between samples, which can also affect data analysis. These expression levels are also subject to the sensitivity and optimal alignment of the fluorometer optics; other systems may require different parameters for adequate signal-to-noise optimization.
Cells are extremely sensitive to time and temperature. Once the experimental procedure has begun, gentle handling is essential to avoid cell death. Specific logistical measures can be taken to expedite the process and avoid timely mistakes including preparing the work station, making sure all equipment is functioning properly, and planning out the goals of the experiment beforehand. It is ideal to use cells within 30 min of harvesting, and in our system, this protocol is executed within 20 min. Once the technique is mastered, specifically transfection optimization and the manual dexterity of the exercise itself, this experiment can be used to compare various constructs against each other, generate dose-response curves, and the sensor can be expanded to the full-length G protein.
Though the FRET sensor used here is a unique development in GPCR FRET sensors, this specific experimental setup is detailed as a well-characterized assay for the implementation of the sensor. The fluorometer-based assay allows a large population of cells to be assessed in each experiment and does not rely on purified protein or membrane preps, therefore maintaining an in vivo environment. This experimental design has also been optimized to detect very small changes seen in the system upon agonist stimulation of the GPCR.
The authors have nothing to disclose.
R.U.M was funded by the American Heart Association Pre-doctoral Fellowship (14PRE18560010). Research was funded by the American Heart Association Scientist Development Grant (13SDG14270009) & the NIH (1DP2 CA186752-01 & 1-R01-GM-105646-01-A1) to S.S.
B2-AR-10 nm- Gas peptide sensor | Addgene | 47438 | https://www.addgene.org/Sivaraj_Sivaramakrishnan/ |
GeneJET Plasmid Miniprep Kit | Fermentas/Fisher Sci | FERK0503 | Elute in 2 mM Tris elution buffer |
HEK-293T-Flp-n cells | Life Technologies | R78007 | |
Trypsin (0.25%) | Life Technologies | 25200056 | |
DMEM- high glucose | Life Technologies | 11960-044 | Warm in 37 °C water bath before use |
FBS, certified, Heat inactivated, US origin | Life Technologies | 10082147 | |
Glutamax I 100x | Life Technologies | 35050061 | |
HEPES | Corning | MT25060CL | |
Opti-MEM | Life Technologies | 31985-070 | Reduced serum media; Bring to room temperature before use |
XtremeGene HP transfection reagenet | Roche | 6366236001 | Highly recommended for its consistency. Bring to room temperature before use |
FluoroMax 4 | Horiba | Use with FluorEssence V3.8 software | |
3-mm path length quartz cuvette | Starna | NC9729944(16.45F-Q-3/z8.5) | May require cuvette holder/adaptor for use in Fluorometer, available from Starna |
Sc100-S3 Heated Circulating water bath pump | Fisher Scientific | 13-874-826 | Warm to 37 °C before use |
Thermomixer Heat Block | Eppendorf | 22670000 | Warm to 37 °C before use |
Ultrapure DNA/RNAse free water | Life Technologies | 10977015 | Use at room temperature |
D(+)-glucose, anhydrous | Sigma | G5767 | |
aprotinin from bovine lung | Sigma | A1153 | |
leupeptin hemisulfate | EMD | 10-897 | |
L-ascorbic acid, reagent grade | Sigma | A0278 | |
(-)-isoproterenol (+)-bitartrate | Sigma | I2760 | Use fresh aliquot each experiment |