Drug Discovery Assays Support—Troubleshooting

The single most common reason that a TR-FRET assay fails is that an incorrect choice of emission filters was made. Unlike other fluorescent assays, the filters used in a TR-FRET assay must be exactly those that are recommended for your instrument as the emission filter choice can make or break the assay. The excitation filter has more of an impact on the assay window. Please refer to our instrument setup guides in our instrument compatibility portal. If your instrument is not listed there, please contact Drug Discovery Technical Support at drugdiscoverytech@thermofisher.com.

Note: Test your microplate reader’s TR-FRET setup before you begin any work with your assay. Using reagents that you have already purchased for your assay, you can test the microplate reader’s TR-FRET setup. Please refer to the Terbium (Tb) Assay and Europium (Eu) Assay Application Notes for plate reader setup.

The primary reason for differences in EC₅₀ (or IC₅₀) between labs is differences in the stock solutions, typically at 1 mM, that the different labs have prepared.

• The compound may not be able to get across the cell membrane or is being pumped out.
• The compound in the cell-based assay may be targeting an inactive form of the kinase or an upstream or downstream kinase. Kinase activity assays must use the active form of the kinase and as such, the inactive form of the kinase cannot be studied. Please note however that a binding assay, such as the LanthaScreen Eu Kinase Binding Assay, can be used to study the inactive form of a kinase.

Taking a ratio of the two emission relative fluorescence units (RFUs) represents the best practice in data analysis for TR-FRET assays, including the LanthaScreen assay. This ratio, referred to as the emission ratio, is calculated by dividing the acceptor signal by the donor signal, 520 nm/495 nm for Terbium (Tb) and 665 nm/615 nm for Europium (Eu). The primary signal is in the acceptor channel with the donor serving as an internal reference.

Most Tb or Eu donors are not involved in producing TR-FRET signal and so the donor signal serves as a useful internal reference in the assay. Dividing by the donor signal helps account for small variances in the delivery (i.e., the pipetting) of the LanthaScreen reagents and lot-to-lot variability of the reagents. TR-FRET is a distance-dependent phenomenon between the donor and the acceptor, and small lot-to-lot differences in the number of labels or their positioning (donor or acceptor) can lead to differences in the 520 nm (for Tb) or 665 nm (for Eu) output. However, the ratio is not affected. As a point of interest, TR-FRET benefits from larger distances when compared to FRET, since molecular vibrations bring the donor and acceptor into the correct distance.

Emission ratios look quite small compared to RFU because an acceptor/donor ratio has been taken. Since the donor counts are typically significantly higher than the acceptor counts, the top of the assay curve will generally be less than 1.0. The numerical values of RFU are typically in the thousands or tens of thousands and these digits are factored out when the ratio is taken. Some instruments arbitrarily multiply the ratio by 1,000 or 10,000 to make the units on the output look more familiar. The statistical significance of the data is not affected by this multiplication.

Below is an example of two different manufacturing lots of LanthaScreen reagents for a tyrosine kinase assay. On the left is the RFU from the acceptor signal (520 nm) for each sample. In this example, sample #1 has a lower acceptor signal then sample #2. However, when the acceptor signal (520 nm) is divided by the donor signal (495 nm) to determine the emission ratio (the right graph), the variation between the samples is negated.

The RFU values of donor and acceptor signals are dependent on instrument settings, such as instrument gain. There is no typical RFU value as instruments can differ significantly depending on how they were designed. Even RFU values between instruments of the same type can differ significantly. The values are essentially arbitrary.

When plotting the emission ratio against the log of the compound concentration the resulting “top” and “bottom” of an assay window will depend on RFU values, and will vary from instrument to instrument.   What matters is the change in the ratio between the top and the bottom of the curve in combination with the noise associated with the data, the Z’-factor.  To get a quick assessment of the assay window, divide the ratio at the top of the curve by the ratio obtained for the bottom of the curve.

For convenience, titration curves can be normalized by dividing all values in the curve by the average ratio obtained at the bottom of the curve. This is called the response ratio and the top and bottom of this curve is the same thing as the assay window. In the below example, each data point from the emission ratio graph (left graph) was divided by the average emission ratio from the bottom of the curve (in this case 0.05) and then graphed on the right as the response ratio. The response ratio allows one to quickly determine the assay window and assess assay performance; assay window always begins at 1.0. Graphing the data as the response ratio has no effect on IC₅₀ values or Z’-factor.

No, according to the Z’-factor, assay window alone is not a good measure of assay performance. Assay window depends on instrument type and settings. What is important in determining the robustness of an assay is not only the size of the window but also the size of the errors (standard deviation) in the data relative to the difference in the maximum and minimum values. The “Z’-factor”, proposed by Zhang and colleagues, takes both of these factors into account, and is a key matrix to assess data quality in an assay. Please take a look at this publication. A large assay window with a lot of noise may have a Z’-factor that is lower than the Z’-factor for an assay with a small window but with little noise. Assays with Z’-factor > 0.5 are considered suitable for screening.

The Z’-factor is easy to picture. On the binding curve, imagine getting the mean value of the ratio at the top of the curve and determine the standard deviation about this mean. Go three standard deviations down from the top. Then at the bottom of the curve, calculate the mean value and standard deviation there. Go three standard deviations up from the bottom of the curve. If 50% of the assay window is left in between these points then the Z’-factor is 0.5. If 75% of the assay window is left, the Z’-factor is 0.75, etc. Mathematically, the formula is expressed as:

Below is a graph of Z’-factor (Z’) as a function of assay window assuming a standard deviation of 5%. As the assay window increases, the Z’-factor quickly reaches a plateau within the first 4-5–fold increase in assay window. Above a 5-fold assay window, large swings in the assay window result in only an incremental increase in the Z’-factor. For example, with a 5% standard error, a 10-fold assay window results in a Z’-factor of 0.82. With a 3-fold increase in the assay window to 30, only a 0.02 gain in the Z’-factor is observed. Statistically, the assay with the 10-fold window is just as suitable for screening as the 30-fold window.

Please keep in mind the following: Output is the blue/green ratio

Draw out your data and controls on a graph and take a quick look at your results. The drawing shown below is intended only for a quick assessment of initial results and not for determination of percent phosphorylation, percent inhibitions or IC₅₀s. The ratio obtained in Z’-LYTE assays is in fact not linear between 0% and 100% phosphorylation. Please refer to the protocol for details. 

In the actual example below, target phosphorylation is between 20 - 50%. The two Kinase Controls used different amount of kinase. The Kinase Controls do contain 1% DMSO which would come from the compunds added. The ratio for Kinase Control #1 is 1.5873 and the ratio for Kinase Control #2 is 0.8825.

Complete lack of an assay window can either be a problem with the instrument setup or with the development reaction. To determine whether or not the problem is with the development reaction or with the instrument setup, please do the following:

Using buffer to make up the volume from reagents that are not used, perform a development reaction as follows.

• 100% Phosphopeptide control:Do not expose the 100% phosphopeptide to any development reagent, this will ensure that it is not cleaved and will give the lowest value of the ratio.
• Substrate:Expose the 0% phosphopeptide, the substrate, to 10-fold higher development reagent than necessary according to the Certificate of Analysis (COA). This will ensure full cleavage after 1 hour and will give the highest value of the ratio.

Typically, for properly developed Z’-LYTE reactions, there is a 10-fold difference in the ratio of the 100% phosphorylated control and the substrate. If not, the dilution of the development reagent used needs to be checked. Please refer to the Certificate of Analysis (COA) for your kit and lot. Please note that the Ser/Thr 7 phosphopeptide is easy to over develop.

If no difference in ratios is observed, either the reagents are very over- or under-developed, or, more likely, it is an instrument problem. Please refer to our instrument setup guides in our instrument compatibility portal. If your instrument is not listed there, please contact Drug Discovery Technical Support at drugdiscoverytech@thermofisher.com.

We do a full titration of Development Reagent when the Quality Control for the development reagent is performed. You should not have to repeat this titration. However, here are typical curves:

The 0% phosphorylated peptide cleaves faster than the 100% phosphorylated peptide, but at high concentration of development reagent, both will cleave and yield a high ratio. With little development reagent, both will yield a low ratio. Usually, there is a range of development reagent concentrations that will work depending on the differential cleavage rate. With little difference in the cleavage rate, the two curves will be very close to each other (Ser/Thr7) so that there is little room for error in the dilution of the development reagent.

• No assay window: Check that your instrument is designed to do FP experiments.
• Poor discrimination between the minimum and maximum mP Control (both include the receptor): This may be an indication of the quality of the receptor. Do not make single use aliquots of the receptor or store diluted.
• The type of microplate is critical: Please see if the protocol specifies untreated polystyrene or NBS-coated plates. If the plate binds the fluorescent substrate, it cannot freely rotate and the mP reading will always be high. White plates cannot be used for FP assays.
• Avoid the use of silanized pipette tips and vials.
• The “No Receptor Control’ (free Fluormone/Tracer) should have much higher RFU values compared to buffer alone. If not, there is a problem with the instrument setup or less commonly with the Fluormone/Tracer.
• If the free Fluormone/Tracer mP value is greater than 100 mP or negative, please refer to the user manual for your instrument and reset the G factor.

Here are possible causes and solutions:

• Cells were not placed immediately either into culture or in liquid nitrogen. Cells are shipped overnight on dry ice, but cannot be left longer on dry ice, even for an additional day.
• The recommended medium/serum was not used. Cells have been FACS sorted and originate typically from a single cell and so are more sensitive to media and components. Please ensure that no substitutions are made.
• The DMSO was not removed from the medium. Please ensure that the DMSO is removed with a medium exchange as described in the protocol.

• Since the FRET data is expressed as a ratio with the larger and more constant emission put into the denominator, the ratio at the top of the curve will typically be less than 1.0. Take the ratio at the top of the curve and divide by the ratio at the bottom of the curve to quickly see your assay window.
• Please ensure that a bottom read was done. All instruments need to be programmed to do bottom read and some require manual movement of the dichroic mirror.
• Observe your cells under a fluorescent microscope. Cells loaded with BLA substrates such as LiveBLAzer will appear green. Stimulated cells will appear blue.
• Take any adherent cell line that you have in culture in your lab. Plate the cells and also have some cell free control wells as described in the customer protocol. Treat all wells with BLA substrate as described in the customer protocol. The cells will remain green. The green signal should be significantly higher from well with cells loaded with BLA substrate as compared to cell-free control wells. If so this is a good sign that your instrument is properly set up.