Table 1.
Ex = peak excitation wavelength (nm), Em = peak emission wavelength (nm), τ = fluorescence lifetime of donor and QY = quantum yield. Extinction Coefficients (ε) of acceptors are given in 10E3/M/cm.
Table 2.
Predicted FRET efficiencies fluorescent protein pairs.
The Overlap Integral (J) of the spectra for each pair was calculated using the freeware program a|e—UV-Vis-IR Spectral Software (http://www.fluortools.com/software/ae), which uses the normalised emission spectrum of the donor and the normalised excitation spectrum of the acceptor corrected for the published Extinction Coefficient (as detailed in Table 1) to determine the value of J in nm4/M.cm. The sources of each spectral dataset are detailed in the Supporting Information. The Forster Radius (r0) for the pair was then calculated according to the equation detailed in the Materials and Methods section. Once r0 was established an expected FRET Efficiciency at a given separation distance (r) was easily calculated according to the relationship E = r06/(r06+r6).
Fig 1.
Fluorescence lifetime measurements of FRET donors expressed in mammalian cells.
A. Fluorescence lifetimes of FRET donors expressed alone in HEK293 cells, measured in both the frequency domain (black bars) and the multiphoton time domain (white bars). Error bars represent the standard deviation in the lifetime measurements of a total 130 > n < 450 individual cells for each bar. B. Representative Frequency Domain fluorescence lifetime images for the 4 donors, acquired at 60X magnification on a Nikon TE2000 microscope equipped with the Lambert Instruments Fluorescence Attachment system. Images are false coloured with lifetime information; warm colours represent longer lifetimes and cool colours shorter lifetimes as per the colour scale provided. C. Representative Multiphoton Time Domain fluorescence lifetime images for the 4 donors, acquired at 20X magnification on a LaVision BioTec TRIMScope. Images are false coloured with lifetime information; warm colours represent longer lifetimes and cool colours shorter lifetimes as per the colour scale provided.
Fig 2.
Frequency domain fluorescence lifetime measurements of FRET pairs expressed in mammalian cells.
A. Frequency domain measurements of fluorescence lifetimes of FRET donors when fused to acceptors and expressed in HEK293 cells (dark grey bars) and the respective FRET efficiencies (yellow bars) calculated from the donor-alone lifetimes from Fig 1A (shown for reference as greyed-out bars). Error bars represent the standard deviation in the lifetime measurements of a total 90 > n < 155 individual cells for each bar. B. Representative Frequency Domain fluorescence lifetime images for the 11 FRET pairs, acquired at 60X magnification on a Nikon TE2000 microscope equipped with the Lambert Instruments Fluorescence Attachment system. Images are false coloured with lifetime information; warm colours represent longer lifetimes and cool colours shorter lifetimes as per the colour scale provided.
Fig 3.
Multiphoton time domain fluorescence lifetime measurements of FRET pairs expressed in mammalian cells.
A. Multiphoton time domain measurements of fluorescence lifetimes of FRET donors when fused to acceptors and expressed in HEK293 cells (dark grey bars) and the respective FRET efficiencies (yellow bars) calculated from the donor-alone lifetimes from Fig 1A (shown for reference as greyed-out bars). Error bars represent the standard deviation in the lifetime measurements of a total 105 > n < 145 individual cells for each bar. B. Representative Multiphoton Time Domain fluorescence lifetime images for the 11 FRET pairs, acquired at 20X magnification on a LaVision BioTec TRIMScope. Images are false coloured with lifetime information; warm colours represent longer lifetimes and cool colours shorter lifetimes as per the colour scale provided.
Fig 4.
Z-Factor and FRET efficiency as a measure of the single time point assay potential of FRET pairs expressed in mammalian cells.
Throughout this figure the donor involved is denoted by the shape of the point (EGFP = circle, Clv = square, mTFP = triangle and mTq2 = diamond) and the acceptor is denoted by the fill-colour of the symbol (mCh = red, mR2 = blue, YPet = yellow, Ven = green and sRCh = black) A. Plot of the mean Z’ of the frequency domain lifetime measurements in 3 separate experiments against their mean FRET Efficiency for those experiments. Error bars on both axes represent standard deviation. B. Plot of the mean Z’ of the multiphoton time domain lifetime measurements in 3 separate experiments against their mean FRET Efficiency for those experiments. Error bars on both axes represent standard deviation. C. Direct comparison of the Z’ values from the two different FLIM systems; pairs that respond similarly on both systems will fall on the x = y diagonal. Error bars represent standard deviation.
Table 3.
Assay potential (Z’) and FRET efficiency under frequency and time domain FLIM conditions at t = 0.
Fig 5.
Evaluating stability and variability in fluorescence intensity and fluorescence lifetime.
A. Example frequency domain time course intensity and lifetime data (in this case, for the green donor Clv). In both cases error bars represent the standard deviation in the normalised intensity and lifetime measurements (each individual cells time course normalised to the mean t = 0 value for that experimental day.) Full data set available in S4 Fig. B. Example multiphoton time domain time course intensity and lifetime data (in this case, for the green donor Clv). In both cases error bars represent the standard deviation in the normalised intensity and lifetime measurements (each individual cells time course normalised to the mean t = 0 value for that experimental day) Full data set available in S5C Fig. Summary of temporal stability and inter-sample variability in frequency domain data. Box plots represent the distribution of the change (i.e. the difference between t = 0 and t = 10 values) in each measurement (fluorescence lifetime = lime green box with blue mean symbol; fluorescence intensity = turquoise box with red mean symbol) over the timecourse, showing the range of the 10th to 90th percentiles (whiskers) and the 25th to 75th percentiles (boxes), for all the data sets contributing to the timecourse charts shown in S5 Fig. The median change is marked by an internal black line, the mean change by the centre of the circular points. The size of these circular data points is proportional to the scale of the error in the total data set, with smaller markets indicating a smaller standard deviation in the data. D. Summary of temporal stability and inter-sample variability in multiphoton time domain data. Box plots represent the distribution of the change (i.e. the difference between t = 0 and t = 10 values) in each measurement (fluorescence lifetime = lime green box with blue mean symbol; fluorescence intensity = turquoise box with red mean symbol) over the timecourse, showing the range of the 10th to 90th percentiles (whiskers) and the 25th to 75th percentiles (boxes), for all the data sets contributing to the timecourse charts shown in S6 Fig. The median change is marked by an internal black line, the mean change by the centre of the circular points. The size of these circular data points is proportional to the scale of the error in the total data set, with smaller markets indicating a smaller standard deviation in the data.
Fig 6.
Z-Factor and FRET efficiency as a measure of the dynamic assay potential of FRET pairs expressed in mammalian cells.
Throughout this figure the donor involved is denoted by the shape of the point (EGFP = circle, Clv = square, mTFP = triangle and mTq2 = diamond) and the acceptor is denoted by the fill-colour of the symbol (mCh = red, mR2 = blue, YPet = yellow, Ven = green and sRCh = black) A. Plot of the mean Z’ of the frequency domain lifetime measurements over a 10 minute time course in 3 separate experiments against their mean FRET Efficiency for those experiments. Error bars on both axes represent standard deviation. B. Plot of the mean Z’ of the multiphoton time domain lifetime measurements over a 10 minute time course in 3 separate experiments against their mean FRET Efficiency for those experiments. Error bars on both axes represent standard deviation. C. Direct comparison of the Z’ values from the two different FLIM systems; pairs that respond similarly on both systems will fall on the x = y diagonal. Error bars represent standard deviation.
Table 4.
Assay potential (Z’) and FRET efficiency under frequency and time domain FLIM conditions over 10 minute time course.
Fig 7.
Comparison of an existing GFP-RFP GTPase Biosensor with novel mTq2-sRCh GTPase biosensors.
A. Schematic representations of the GTPase biosensors tested: a GFP-RFP based RhoA-Raichu [43] (left); a new mTq2-sRCH based NRas-Raichu (right); and a modified version of an alternative geometry RalB probe [46] (centre). B. Fluorescensce lifetime responses of serum starved HEK 293 cells transiently expressing GTPase biosensors to 20% serum stimulation at t = 0. Measurements were taken 1 and 2 minutes before and after the central (t = 0) image at which the stimulus added. Each plot represents one repeat of the experiment consisting of the mean lifetimes from 10 < n > 30 cells spread over at least 3 dishes. Error bars represent standard deviations. As above, the GFP-RhoA data is in the left panel, the mTq2-RalB data in the centre and the mTq2-NRas on the right. C. FRET Efficiency of the time course data seen in 7B, in which each time point for each day was normalised to the equivalent time point of the average of the donor alone control time courses that underwent the same stimulation protocol on the same day. Error bars represent the percentage error calculated from the standard deviations in the FRET-lifetimes. As above, the GFP-RhoA data is in the left panel, the mTq2-RalB data in the centre and the mTq2-NRas on the right. D. Average difference in fluorescence lifetime before and after stimulation compared to the t = 0 value of each time course. Error bars represent standard deviations, and significance was determined using a Rank Sum Test. Statistically significant differences are marked with asterisks. One asterisk represents 0.05 < p > 0.01, two asterisks 0.01 < p > 0.005 and three asterisks p < 0.005. E. Average difference in FRET Efficiency before and after stimulation compared to the t = 0 value of each time course. Error bars represent standard deviations, and significance was determined using a Rank Sum Test. Statistically significant differences are marked with asterisks. One asterisk represents 0.05 < p > 0.01, two asterisks 0.01 < p > 0.005 and three asterisks p < 0.005. F. Sensitivity of biosensors in terms of picosecond lifetime change per percentage unit of FRET Efficiency after stimulation. Error bars represent standard deviations.