Skip to main content
. 2020 Feb 4;9:e50571. doi: 10.7554/eLife.50571

Figure 5. FynSensor reveals integrin-dependent Fyn activity to be spatially compartmentalized.

(A) Spontaneous formation of zones enriched in active Fyn is seen in serum starved cells plated on fibronectin (FN). (I) Representative confocal fluorescence micrographs showing sensitized emission (total FRET: FRETT) levels indicative of active Fyn in serum starved-U2OS cells expressing FynSensor (Scale bar = 10 μm). For quantitative image analysis of compartmentalized Fyn activity, we have divided the cell into quadrants labelled as Q1-Q4 as shown. (II) Bar graph comparing FynSensor FRET levels in distinct cellular zones. The mean of Max-FRETT-HFQ (HFQ: high-FRET quadrant, Q4 in panel I) and Max-FRETT-LFQ (LFQ: low-FRET quadrant, Q2 in I) values (see Materials and methods) are plotted. Values are mean ± s.e.m. Student’s paired one-tailed t-test has been used to determine the p-value (n = 37 cells). (B) FynSensor FRET readout is dependent on F29 binding Fyn, with the non-binding control mutant of F29 (P41A) showing no or significantly reduced FRETT signal compared to FynSensor. (I) Confocal fluorescence micrographs showing sensitized emission (total FRET: FRETT) levels for cells expressing mCer-Fyn and non-binding control binder mVenus-F29 P41A. (II) Comparison of FRETT levels for FynSensor with non-binding control. The bar graph shows the average maximum FRETT obtained (Max-FRETT-HFQ) for non-binder (n = 5 cells) and FynSensor (n = 37 cells). Values are mean ± s.e.m. Student’s unpaired one tailed-t-test was used to determine the p-value. (C) Fyn activity levels are sensitive to inhibition of focal adhesion kinase (FAK), a known mediator of integrin signaling. (I) Confocal fluorescence micrographs showing FynSensor FRETT levels before and after treatment with 10 μM of FAK inhibitor (PF 562271). Loss of basal Fyn activity is observed in cells when treated with inhibitor but not with DMSO (vehicle control). (II) Time course of Fyn activity in response to FAK inhibition. Average fluorescence intensity profile (FRETT) of FynSensor cells (n = 5) over time, before and after addition of FAK inhibitor. (Scale bar = 10 μm). (III) Quantifying FynSensor FRET on FAK inhibition. Bar graph shows the average FRETT intensity of cells before FAK inhibitor treatment (-FAKi), five mins after FAKi treatment and 15 mins after FAKi treatment (n = 5 cells). Values are mean ± s.e.m. Student’s paired one tailed-t-test was used to determine the p-value. (IV) FAK-inhibitor induced reduction in FynSensor FRET levels is NOT seen with vehicle control. Bar graph shows time-averaged, mean cell FRETT intensity before and after treatment with the vehicle control (n = 8 cells). Values are mean ± s.e.m. Student’s paired one tailed-t-test was used to determine the p-value.

Figure 5—source data 1. Quantification of FynSensor FRET levels in low and high activity zones in U2OS cells.
Figure 5—source data 2. Quantification of FynSensor non-binding control FRET levels in low and high activity zones in U2OS cells.
Figure 5—source data 3. Quantification of FynSensor FRET levels in cells in treated with FAK inhibitor.

Figure 5.

Figure 5—figure supplement 1. FynSensor FRET readout is highly sensitive to F29 binding to Fyn kinase in cells. Non-binding mutant of FynSensor fails to show significant FRET response.

Figure 5—figure supplement 1.

FynSensor FRET readout is dependent on F29 binding Fyn, with the non-binding control mutant of F29 (P41A) showing no or significantly reduced FRETT signal. Confocal fluorescence micrographs showing Donorex-Donorem, Acceptorex-Acceptorem and corresponding FRETT (sensitized emission) images for U2OS cells expressing mCer-Fyn and non-binding control binder mVenus-F29 P41A (marked as ‘CFP ch’, ‘YFP ch’ and ‘FRETT’ respectively). These images confirm that mCer-Fyn and non-binding control constructs are co-expressed in these cells and show expected ‘CFP’ and ‘YFP’ fluorescence on direct excitation. However, for non-binding control cells little or no FRET (sensitized emission) is seen. This lack of FRET shows that the FynSensor response is indeed specific and is dependent on F29 binding to activated Fyn. Also, plotted here is the average difference in FRETT levels between intracellular zones showing maximum and minimum FRETT (bottom right panel). Unlike the cells expressing FynSensor, non-binder expressing U2OS cells (n = 5 cells) do not show any discernible spatially compartmentalized kinase activity pattern. Values are mean ± s.e.m. Student’s paired one tailed-t-test was used to determine the p-value.
Figure 5—figure supplement 1—source data 1. Quantification of F29 non-binding mutant FRET levels across cell zones.
Figure 5—figure supplement 2. Localized Fyn activity (spatially localized FynSensor FRET) is not an artifact of binder localization.

Figure 5—figure supplement 2.

(A) C2C12 cells showing localized Fyn activity (spatially localized FynSensor FRET), show a spatially uniform binder localization. Therefore, localized FRETT patterns are not determined by binder localization (I) Fluorescent confocal micrograph shows that while labeled F29 binder (myr-mVenus-F29; ‘YFP ch’) localizes nearly uniformly in cells, the FynSensor FRET index image clearly reveals a spatially-localized pattern of Fyn activity. (II) Quantitative comparison of average FRETT and ‘YFP ch’ (labeled F29) intensities in two distinct intracellular zones (quadrants). Quadrant analysis performed on both the YFP channel image and the FRET index image for each cell (refer to Materials and methods section for details) yielded the maximum ‘fluorescence intensity’ values from the ‘high FRETT quadrant (HFQ)’ and the ‘low FRETT quadrant (LFQ)’ in FRET index and ‘YFP ch’ images. Bar graph shows the average of these values from n = 32 cells. Student’s paired one tailed-t-test was used to determine the p-value. While there is a clear and significant difference in the FRETT levels across two intracellular zones (quadrants), fluorescence due to the binder across the cell (‘YFP ch’) remains constant. (B) Illustration of uniform binder localization in regions showing localized FRETT. (I) Fluorescent intensity scan across FRET and ‘YFP ch’ confocal micrographs to determine uniformity of signal. (II) Plots of average intensity values scanned across an extended cellular area (white box) for FRET index and ‘YFP ch’/binder alone images. While FRETT is spatially modulated, binder localization is uniform.
Figure 5—figure supplement 2—source data 1. Quantification of FynSensor and F29 localization levels across cell.
Figure 5—figure supplement 3. Spatio-temporal modulations in Fyn activity are not simply due to changes in local Fyn protein concentration.

Figure 5—figure supplement 3.

(A) Spatially enhanced Fyn activity patterns are observed even when activation levels are normalized for any changes in local protein concentrations. This shows localized protein activation cannot simply be ascribed to increased protein accumulation. (I) FynSensor FRET index images (confocal micrographs) of 3 representative FynSensor expressing C2C12 cells showing localized Fyn activity. Specifically, intracellular zones with enhanced levels of active Fyn can be observed. (II) Bar graphs comparing the mean of maximum ‘CFP-normalized FRET’ intensity values (see Materials and methods) in both the high FRETT quadrant and low FRETT quadrant in serum starved. For mean, ‘CFP-normalized FRET’ values were averaged across the specified quadrant in a single cell and then across cells. ‘CFP normalized FRET’ quantifies Fyn activity normalized for changes in levels of labeled mCer-Fyn (protein concentration). For each cell, time point showing max FRETT was chosen. Values are mean ± s.e.m. Student’s one-tailed t-test has been used to determine the p-value (n = 32 cells). (B) Oscillations in Fyn activity oscillations are not due to fluctuations in local protein levels. (I) Normalized FynSensor FRET temporal profiles (FRETT normalized to mCerulean intensity/protein levels) also show the characteristic pulsatile signals as seen with FRETT. (II) While significant temporal oscillations are observed in Fyn activity, Fyn protein levels do not fluctuate similarly. Shown here is the extent of oscillations in FRETT (cell averaged differences between maximum and minimum FRETT levels observed over time, in the high-FRET cellular quadrant, HFQ). While significant changes in FRETT levels are observed, mCerulean (‘CFP ch’) fluorescence in the corresponding cellular zones remains largely constant over time. (n = 9 cells). Values are mean ± s.e.m. Student's paired one-tailed t-test has been used to calculate the p-value.
Figure 5—figure supplement 3—source data 1. Quantification of donor-normalized FRET levels in low and high activity zones.
Figure 5—figure supplement 3—source data 2. Comparison of FynSensor FRETT levels and Fyn kinase localization at different time points.
Figure 5—figure supplement 4. Localized Fyn activity patterns are distinct from patterns of Fyn localization.

Figure 5—figure supplement 4.

Fyn localization does not equate with Fyn activity. (A) Confocal micrograph of FynSensor FRET index (Fyn activity, I) and ‘CFP ch’ (mCer-Fyn localization, II). Images showing intracellular zones (zones 1 and 2, enlarged and marked) with differing levels of FRETT (Fyn activity) but similar levels of labeled Fyn protein. (B) Plots quantifying and comparing levels of FRETT (activity) and mCerulean fluorescence (‘CFP ch’/protein concentration) between the highlighted intracellular zones 1 and 2 (bar represent mean of the 5 ROIs drawn in each zone, see A). Values are mean ± s.e.m. Student's paired one-tailed t-test has been used to calculate the p-value. Intracellular zones have significantly different FRETT levels but similar local concentrations of Fyn.
Figure 5—figure supplement 4—source data 1. Quantification of FRETT levels and kinase localization across selected cellular zones.
Figure 5—video 1. FRET response in human osteosarcoma cell-line, U2OS cells expressing non-binder mutant of FynSensor.
Download video file (445.4KB, mp4)
Live cell imaging of U2OS cells transfected with non-binder mutant of FynSensor before and after PDGF stimulation. Panel shows FRET index image using Fire LUT (Fiji) scaled appropriately for ease of visualization. Individual images were acquired at 60X at an interval of 30 s and then stitched together to make a video (played at 5 frames/second). Time stamp indicates elapsed time, ‘+PDGF’ label shows the point at which PDGF was added to the cells. Color bar shows intensity of FRET signal. No/very little FRET signal is seen from these cells and this FRET signal is not ‘enhanced’ post PDGF stimulation. Scale bar = 10 µm.
Figure 5—video 2. Effect of FAK inhibition on Fyn activity in human osteosarcoma cell-line, U2OS cells.
Download video file (874.9KB, mp4)
Live cell imaging of U2OS cells transfected with FynSensor before and after inhibition of FAK. Panel shows FRET index image using Fire LUT (Fiji) scaled appropriately for ease of visualization. Individual images were acquired at 60X at an interval of 60 s as part of a multi-point-time-lapse imaging regime and then stitched together to make a video (played at 3 frames/second). Time stamp indicates elapsed time, ‘+FAK inhibitor’ label shows the point at which FAK inhibitor was added to the cells. Colour bar shows intensity of FRET signal. Basal level of Fyn activity (unstimulated state) is highly localized in space. Post addition of inhibitor the Fyn activity in the cell is drastically attenuated. Scale bar = 10 µm.