We describe here an improved Luminescence Resonance Energy Transfer (LRET) method where we introduce a protease cleavage site between the donor and acceptor fluorophore sites. This modification allows us to obtain specific LRET signals arising from membrane proteins of interest, allowing for the study of membrane proteins without protein purification.
Luminescence Resonance Energy Transfer, or LRET, is a powerful technique used to measure distances between two sites in proteins within the distance range of 10-100 Å. By measuring the distances under various ligated conditions, conformational changes of the protein can be easily assessed. With LRET, a lanthanide, most often chelated terbium, is used as the donor fluorophore, affording advantages such as a longer donor-only emission lifetime, the flexibility to use multiple acceptor fluorophores, and the opportunity to detect sensitized acceptor emission as an easy way to measure energy transfer without the risk of also detecting donor-only signal. Here, we describe a method to use LRET on membrane proteins expressed and assayed on the surface of intact mammalian cells. We introduce a protease cleavage site between the LRET fluorophore pair. After obtaining the original LRET signal, cleavage at that site removes the specific LRET signal from the protein of interest allowing us to quantitatively subtract the background signal that remains after cleavage. This method allows for more physiologically relevant measurements to be made without the need for purification of protein.
Luminescence Resonance Energy Transfer (LRET) is a derivative of the well-known Fluorescence Resonance Energy Transfer (FRET) technique1. Similar to FRET, LRET can be used to measure distances and distance changes between donor and acceptor fluorophores attached to specific sites on the protein of interest within the range of 10-100 Å1-3. The principles of LRET are also similar to FRET in that resonance energy transfer occurs between two proximal fluorophores when the emission spectrum of the donor fluorophore overlaps with the absorption spectrum of the acceptor fluorophore. The efficiency of this transfer is related to the distance between the two fluorophores by the following equation:
Eq. 1
where R is the distance between the two fluorophores, E is the efficiency of energy transfer, and R0, discussed below, is the Förster radius for the fluorophore pair, i.e. the distance at which efficiency of transfer is half-maximal. From this equation, one can see that efficiency is related to the magnitude of the distance raised to the inverse sixth power1. It is this inverse sixth power dependence that allows for FRET and LRET measurements to be exquisitely sensitive even to small distance changes when near the R0 of the FRET pair. The ability to specifically label desired sites on proteins or other macromolecules allows one to take advantage of this sensitivity to monitor conformational changes.
When compared to FRET, which uses conventional organic dye molecules, LRET offers additional advantages. In LRET, instead of using an organic dye as the donor fluorophore, a lanthanide series cation, typically Tb3+ or Eu3+, is used1,4-6. Fluorophores that fall under this category, e.g., terbium chelate, are also very versatile in that they can be used with a wide range of acceptor fluorophores. This flexibility is made possible because the emission spectra of chelated lanthanides contain multiple sharp emission peaks, allowing for a single species of donor fluorophore to be used with one of a wide variety of acceptor fluorophores. Thus, sensitized acceptor emission can be detected without any fear of contaminating bleed-through from donor emission5. The experimenter selects the specific acceptor based on the expected distance between the two fluorophores (Figure 1 and Table 1). In these chelated lanthanide fluorophores, the metal ion is chelated by a molecule that contains an antenna group that sensitizes the normally poorly-absorbing lanthanide to excitation as well as a bioreactive functional group to tether the ion to a specific functional group on the macromolecule1,5,6. Once excited, lanthanides relax to the ground state via the release of photons with a decay rate in the millisecond range. Because the decay is neither a singlet-to-singlet relaxation nor a triplet-to-singlet relaxation, the emission of photons cannot properly be called fluorescence or phosphorescence, but is more properly termed luminescence1. The long decay of lanthanide luminescence greatly helps in lifetime measurements. Lifetime measurements can then be used to determine efficiency via the following relation:
Eq. 2
where, E is the efficiency of transfer, τD is the lifetime of the donor (chelated lanthanide) when not participating in energy transfer, and τDA is the lifetime of the donor when participating in energy transfer with the acceptor. With LRET, τDA can also be measured as the lifetime of the sensitized acceptor emission because terbium’s lifetime is so much greater than an organic acceptor fluorophore. The acceptor emits with the same lifetime as its inciting excitation (donor lanthanide), and any contribution to the lifetime from the acceptor’s own intrinsic fluorescence lifetime is relatively negligible. By measuring the sensitized emission rather than donor emission, we also eliminate the need to ensure labeling at exactly a 1:1 ratio of donor to acceptor. Protein can instead be labeled simultaneously with both acceptor and donor fluorophores. A heterogeneously labeled population will result, but double-donor labeled proteins will not emit in the acceptor wavelength and double-acceptor labeled proteins will not be excited. Moreover, the distance between fluorophores should be the same, regardless of which cysteine site a given fluorophore attaches to, especially when using the isotropic lanthanides as a donor, so the need to specify a given site to receive either the donor or acceptor is unnecessary. Intensity may be affected with a heterogeneous population, but should still be more than sufficient to be detected.
When planning experiments, the choice of fluorophores should be dictated by the R0 value of the pair as well as the expected distance range being measured. The R0 value is defined by the following equation:
Eq. 3
where, R0 is the Förster radius in Angstroms, κ2 is the orientation factor between the two dyes (usually assumed to be 2/3), ϕD is the quantum yield of the donor, J is the spectral overlap integral between the donor's emission spectrum and the acceptor's absorbance spectrum in M–1cm-1nm4, and n is the refractive index of the medium1.
Our laboratory has added a modification to the conventional LRET technique by introducing a protease recognition site between the donor and acceptor label sites on the protein being probed. This modification allows for investigation in non-purified systems such as whole mammalian cells7. This technique is particularly useful when using cysteines as sites for labeling, since in the process of labeling with maleimide-conjugated dyes that bind to cysteine sulfhydryl groups, other proteins on the cells that have cysteines are also labeled. However, by including protease cleavage sites on the protein of interest and measuring lifetimes before and after cleavage, the experimenter can quantitatively subtract the background signal after protease cleavage from the raw signal. This subtraction isolates the specific signal arising from the protein of interest (Figure 2). Using the modification described above, LRET can be used to measure distance changes between the terbium chelate donor and the acceptor probe on a protein, and thus monitor conformational changes in the protein’s near physiological state without the requirement for purification.
Figure 1.The absorption and emission spectra of chelated terbium in black, as well as a representative acceptor, Alexa 488, in red. Notice the multiple emission peaks and the sharp, narrow emission range for each peak of terbium chelate. This pattern allows for terbium to be used with a variety of acceptor fluorophores and facilitates the measurement of sensitized emission within those ranges where terbium shows no emission. Terbium’s emission peak at 486 nm overlaps quite well with the absorption peak of Alexa 488, allowing for resonance energy transfer to occur between the two fluorophores. A wavelength of 515 nm is an excellent choice to detect sensitized emission for this pair as it is in the valley between the terbium emission peaks, and quite near Alexa 488’s emission peak of 520 nm. Note that being near the acceptor peak, though desirable, is not required—565 nm is still able to detect Alexa 488 emission without also detecting terbium emission.
Acceptor Fluorophore | R0 (Å) | Emission wavelength (nm) |
Atto 465 | 36 | 508 |
Fluorescein | 45 | 515 |
Alexa 488 | 46 | 515 |
Alexa 680 | 52 | 700 |
Alexa 594 | 53 | 630 |
Alexa 555 | 65 | 565 |
Cy3 | 65 | 575 |
Table 1. A list of commonly used acceptor fluorophores for LRET using terbium chelate as the donor11. The R0 values were measured when the donor and acceptor were attached to the soluble agonist binding domain of AMPA receptors. It is ideal to measure the R0 value again for each new system being studied.
Figure 2. An overview of the LRET method presented. (A) The AMPA receptor is a membrane protein which undergoes conformational changes upon ligand-binding. The clamshell-shaped ligand-binding domain is circled here in red. (B) The ligand-binding domain of AMPA when not bound to protein exists in an open conformation (left). When bound to ligand glutamate, the protein closes around its ligand (right). By placing fluorophores at probative sites on the LBD, the nature of this conformational change can be seen as the distance between the fluorophores changes, which will then affect fluorescence lifetime. (C) When labeling whole cells, labeling of both the protein of interest as well as background membrane proteins may occur (left). After protease cleavage, LRET signal from the protein of interest will disappear due to the release of a soluble fragment, leaving background signal intact (right). This background signal can then be subtracted from the raw signal.
1. Create the Construct Containing the Protein of Interest
Clone the gene expressing the protein of interest into a suitable vector. Use vectors such as the pcDNA series or pRK5 as they are well suited for expression in mammalian systems like HEK293 and CHO cells.
2. Select the Sites on the Protein to be Tagged with the Fluorophores
3. Test the Expression and Functionality of the Protein
4. Select the Fluorophores to be Used
Select fluorophores based on the expected distance range being measured such that the range is between 0.5-1x the R0 of the fluorophore pair.
NOTE: This allows for an easier subtraction of the background, which typically has much longer lifetimes. For example, if the expected distance range being measured is around 35 Å, an appropriate fluorophore pair to use would be terbium chelate as the donor and Alexa 594 as the acceptor, because the R0 for this pair is 53 Å (Table 1).
5. Express the Protein by Transiently Transfecting the Required Amount of Mammalian Cells
Transiently transfect the protein of interest into the chosen mammalian cells using any of the common transfection reagents. Typically, use four 10-cm dishes per LRET experiment for HEK and CHO cell lines; however, this amount may vary depending on protein expression, stability, etc. Allow the cells to express the protein for 36-48 hr before harvesting.
6. Label the Proteins
7. Set up the LRET Experiment
8. Analyze the Data Obtained
A successful LRET measurement with a lanthanide donor should have a donor only lifetime in the millisecond time range. The lifetime of sensitized LRET emission for the protein labeled with both the donor and the acceptor will be notably shorter, with a lifetime in the microsecond range after background subtraction Figure 3. Protease cleavage results in an increase in lifetime that should become stable over time (i.e. no longer changing), showing that the protease cleavage is complete Figure 4. If the LRET signal comes from only one set of sites, the resulting emission lifetime after subtraction of the background should give a single exponential lifetime.
When measuring conformational changes, the specific LRET signal should show a change in lifetime outside of the error of the measurement Figure 3. The error of the measurement can be calculated by the propagation of the errors associated with the fit of the lifetimes. After fitting the data to the minimum number of exponential decay functions described in equation 4, the lifetime, τ, can be determined. Using equation 5, the donor-acceptor and donor-only lifetimes can be used along with the R0 value for the LRET pair to calculate the distance between the two fluorophores under the conditions tested. Relating the distance change resulting from changing these conditions to the overall structure and function of the protein is now the job of the experimenter.
Figure 3. LRET measurements of the Acid-Sensing Ion Channel 1a (ASIC1a). Fluorescence intensity has been plotted on a logarithmic scale to improve the ease of visual interpretation. (a) Donor-only samples show a single-exponential decay that does not change with pH. (b) In donor-acceptor labeled samples, a decrease in the lifetime of sensitized emission is seen upon a decrease in pH from 8 to 6 (black to red). This lifetime decrease indicates a decrease in the distance between the finger and thumb domains of ASIC1a. This figure has been modified from Ramaswamy, et al, 20138.
Figure 4. The effect of background subtraction on LRET measurements made on the Acid Sensing Ion Channel 1a (ASIC1a). The sites to be specifically labeled by fluorophores are separated by a protease cleavage site. After LRET measurements are made, the protease is introduced to the protein sample and subsequent cleavage of the protein of interest results in a loss of the specific signal. Any LRET that remains is background fluorescence from fluorophores bound to other cysteines present on other membrane proteins. Subtracting this background isolates the true LRET data for the protein of interest. This figure has been modified from Ramaswamy, et al, 20138.
LRET is a powerful technique that allows scientists to measure distances between domains within a single protein as well as between subunits in a multimeric protein. As such, LRET is well-suited to examining the conformational changes and dynamics of proteins or other macromolecules. The above protocol should allow the properly equipped lab to easily test their hypotheses; however, there are many common sources of error that may plague the new investigator. If little or no LRET signal is seen, first check the wavelength settings used. Excitation should be in the absorbance range of terbium (330-340 nm). For donor only measurements, in which sample has been labeled only with donor fluorophore and without acceptor fluorophore, the emission wavelength should be at one of the peaks shown in Figure 1, while for donor-acceptor measurements, the emission wavelength should match the acceptor fluorophore being used Table 1. If the wavelength settings are correct, check the compatibility of buffers. Some fluorophores may not be compatible with certain buffers or in certain pH ranges. Next, ensure that the choice of residues and fluorophores are compatible. If the experimental design is completely correct, then the problem likely lies either with the fluorophores or the protein itself. Over time, stock solutions of fluorophores may degrade and may result in inadequate labeling. Finally, check expression and functionality of the protein. Many mutations have been added, including the introduction of cysteines, the removal of native cysteines, and the introduction of at least one protease cleavage site. Thus, it is possible that, even under normal transfection conditions, the introduced mutations destabilize the protein and cause under-expression of the protein of interest, lowering the signal seen. If expressed, the mutations or labeling may cause denaturation of the protein, causing the residues to be placed differently from the expected distances seen in wild-type protein. Western blots can be used to verify expression of the protein. If there is any question about the trafficking of a surface membrane protein, biotinylation of the cell surface and pull-down of surface-exposed proteins, followed by a western blot for the protein of interest, will specifically demonstrate surface trafficking. For functional tests, there is no single assay to recommend for use specifically with LRET, since LRET can be used on a wide variety of protein types. Again though, possible examples of functional assays include enzyme activity assays, ligand-binding assays, and electrophysiology studies. If protein expression or function has been too adversely affected by the introduced mutations, then new labeling sites must be chosen.
If an LRET signal is seen but cannot be fitted by a single exponential when such a result is expected, first check that the background was correctly subtracted. If, after subtraction, a multi-exponential decay is seen, this signal could be an indication of LRET being observed from multiple interactions other than what was intended. Check to make sure all other accessible cysteines have been removed from the protein. If a crystal structure is available, it will be a very useful tool to check for these cysteines. Again, disulfide-bonded and buried, inaccessible cysteines do not need to be mutated away. To test the inaccessibility of these cysteines if there is a compelling reason not to mutate them away, introduce one non-native cysteine in the protein and ensure that there is no LRET signal in a donor-acceptor labeled sample. If all accessible cysteines have indeed been mutated away, and if the protein has multiple subunits or is part of a complex, then there may be confounding LRET signal due to dye attaching to those nearby proteins or subunits. Choosing a different protease cleavage site may help with these problems; otherwise, other labeling sites may need to be chosen. Finally, if the issue with fitting is simply a matter of signal-to-noise, the likely problem is due to low expression of the protein. Expression will then need to be optimized through different transfection conditions, a different vector, etc.
If the LRET measurements produce an anomalous or seemingly physically meaningless result, there may be protein-specific issues that may not be readily obvious. For example, with acid-sensing ion channels, even the careful addition of an acid to change the pH might result in some cell death and protein denaturation. Thus, multiple samples need to be prepared, one for each pH to be tested. Also, in addition to resonance energy transfer, local environment changes can affect a fluorescence signal. Such a change, if significant, would be noted in the donor-only measurement as a double or multi-exponential decay. In these cases, the labeling sites need to be moved to different positions to make sure the change in conditions does not change the spectral properties of the fluorophores.
Even while keeping these sources of problems in mind, there are some caveats to and limitations of LRET of which an experimenter must be aware. First, conventional labeling technique relies on labeling cysteine residues. To reduce labeling of non-specific residues, other cysteines are typically mutated out; however, this method is not always practical. For example, if a protein has many non-disulfide-bonded cysteines native to its structure that are critical to the protein’s structure or function, then mutating them out will be impossible, greatly increasing the limitation of the technique and the interpretation of data. Also, the LRET technique is more suited to detecting changes in distance, rather than absolute distances, as any errors on absolute distance due to the effect of the orientation factor κ2 on the R0 value are likely to be reduced in distance change analysis because these errors affect measurements under all conditions equally. Alternative techniques can be done to overcome some of these limitations. For example, to avoid adding too many cysteines, one could affix a His-tag and label it with a fluorophore bound to nickel-NTA. Also, native tryptophans can be used as a one of the fluorophores with the understanding that if used as a donor, tryptophans have a much smaller lifetime than terbium, thus intensity based measurements may be more appropriate than lifetime measurements. If more exact atom to atom distances are required, techniques such as X-ray crystallography, molecular dynamics, or NMR are still more appropriate techniques to get these absolute distances.
Due to its exquisite sensitivity to distance changes, LRET can measure distance changes with Angstrom-level resolution for solution-phase proteins and can provide experimental data without the need for high-purity, isotopic labels or the size restriction that impairs both NMR and molecular dynamics. After learning and mastering the technique, examinations into conformational changes of proteins can be done much faster and with more ease than already available conventional techniques. LRET also provides an excellent foundation for further specialized resonance energy transfer techniques such as single molecule FRET (smFRET), which can examine the population distribution of conformational states of individual molecules, rather than the ensemble average.
The authors have nothing to disclose.
This work was supported by National Institutes of Health Grant GM094246, the American Heart Association Grant 11GRNT7890004, and the National Science Foundation Grant MCB-1110501.
Name of Reagent/ Equipment | Company | Catalog Number | Comments/Description |
LanthaScreen Thiol Reactive Tb Chelate | Life Technologies | PV3579 | |
Acceptor fluorophore-Fluorescein-5-Maleimide | Life Technologies | F-150 | Choice of acceptor depends on the specific experiment |
Acceptor fluorophore-Alexa Fluor 488 C5 Maleimide | Life Technologies | A-10254 | Choice of acceptor depends on the specific experiment |
Acceptor fluorophore-Alexa Fluor 555 C2 Maleimide | Life Technologies | A-20346 | Choice of acceptor depends on the specific experiment |
Acceptor fluorophore-Alexa Fluor 594 C5 Maleimide | Life Technologies | A-10256 | Choice of acceptor depends on the specific experiment |
Acceptor fluorophore-Alexa Fluor 680 C2 Maleimide | Life Technologies | A-20344 | Choice of acceptor depends on the specific experiment |
QuantaMaster 3-SS | Photon Technology International | Spectrofluorometer should have pulsed excitation with the ability to measure lifetimes in the ms range | |
FluoreScan 2.0 | Photon Technology International | Data Acquisition Software used in manuscript. Software provided with fluorescence instrument. | |
Origin 8.6 | OriginLab | Origin is the data analysis software used in protocol. Can use other similar data analysis software | |
Quartz Cuvette | Starna Cells, Inc | 3-Q-10 | |
Stir Bar | Bel-Art Products | F37119-0007 | Used in cuvette to keep cells in suspension. Can use any stir bar that fits the cuvette |