A FRET-based biosensor for measuring Gα13 activation in single cells

Abstract

Förster Resonance Energy Transfer (FRET) provides a way to directly observe the activation of heterotrimeric G-proteins by G-protein coupled receptors (GPCRs). To this end, FRET based biosensors are made, employing heterotrimeric G-protein subunits tagged with fluorescent proteins. These FRET based biosensors complement existing, indirect, ways to observe GPCR activation. Here we report on the insertion of mTurquoise2 at several sites in the human Gα13 subunit, aiming to develop a FRET-based Gα13 activation biosensor. Three fluorescently tagged Gα13 variants were found to be functional based on i) plasma membrane localization and ii) ability to recruit p115-RhoGEF upon activation of the LPA2 receptor. The tagged Gα13 subunits were used as FRET donor and combined with cp173Venus fused to the Gγ2 subunit, as the acceptor. We constructed Gα13 biosensors by generating a single plasmid that produces Gα13-mTurquoise2, Gβ1 and cp173Venus-Gγ2. The Gα13 activation biosensors showed a rapid and robust response when used in primary human endothelial cells that were exposed to thrombin, triggering endogenous protease activated receptors (PARs). This response was efficiently inhibited by the RGS domain of p115-RhoGEF and from the biosensor data we inferred that this is due to GAP activity. Finally, we demonstrated that the Gα13 sensor can be used to dissect heterotrimeric G-protein coupling efficiency in single living cells. We conclude that the Gα13 biosensor is a valuable tool for live-cell measurements that probe spatiotemporal aspects of Gα13 activation.

Fig 1. Insertion of a fluorescent protein at different positions in Gα13.

(A) The protein structure of human Gα13 (PDB ID: 1ZCB). The highlighted residues indicate the amino acid preceding the inserted fluorescent protein. Successful sites for inserting mTurquoise2-Δ9 into Gα13 in pink and unsuccessful sites in orange. (B) A partial protein sequence alignment (full alignment see S1 Fig) of different Gα classes. The highlighted residues indicate the amino acid preceding the inserted fluorescent protein (or luciferase). In bold, the sites that were previously used to insert Rluc [26]. Insertion of mTurquoise2-Δ9 in Gα13 after residue Q144 (black) was based on homology with previous insertions in Gαq and Gαi (black). Successful sites for inserting mTurquoise2-Δ9 (R128, A129 and R140) in pink and unsuccessful sites (L106 and L143) in orange. The numbers indicated below the alignment correspond with the Gα13 variant numbers, used throughout the manuscript. The colors under the alignment match with the colors of the αHelices shown in (A). (C) Confocal images of the tagged Gα13 variants transiently expressed in HeLa cells. The numbers in the left bottom corner of each picture indicate the number of cells that showed plasma membrane localization out of the total number of cells analyzed. The tagged Gα13 variants also localize to structures inside the cell, which are presumably endomembranes,. The width of the images is 76μm.

Fig 2. Capacity of the tagged Gα13 variants to recruit p115-RhoGEF.

(A) Confocal images of a representative HeLa cell expressing SYFP1-p115-RhoGEF, Gα13.2-mTurquoise2-Δ9 and LPA2-P2A-mCherry (here only SYFP1-p115-RhoGEF is shown, for the localization of the other constructs see S2 Fig) (before (t = 0s) and after (t = 100s) addition of 3μM LPA). The width of the pictures is 67μm. (B) The mean cytoplasmic fluorescence intensity of SYFP1-p115-RhoGEF over time. After 8s, 3μM LPA was added. All cells transiently expressed LPA2 receptor-P2A-mCherry. The number of cells imaged is p115-RhoGEF n = 5, Gα13 untagged + p115-RhoGEF n = 15, Gα13.1 + p115-RhoGEF n = 27, Gα13.2 + p115-RhoGEF n = 28, Gα13.3 + p115-RhoGEF n = 24, Gα13.5 + p115-RhoGEF n = 20. Data have been derived from three independent experiments. (C) Quantification of the fluorescence intensity at t = 50s for each Gα13 variant, relative to t = 0s. The dots indicate individual cells and the error bars show 95% confidence intervals. The numbers of cells analyzed is the same as in (B).

Fig 3. The ability of the tagged Gα13 variants to report on dynamic Gα13 activation via ratiometric FRET imaging.

(A) Ratiometric FRET traces of HeLa cells expressing the LPA2 Receptor (untagged), Gβ (untagged), one of the Gα13 variants (as indicated in the title of the graphs) tagged with mTurquoise2-Δ9 and Gγ tagged with cp173Venus. The grey lines represent individual cells and the black graph represents the average of which the error bars indicate the 95% confidence intervals. LPA was added at t = 42-50s, indicated by the arrowhead. The number of cells analyzed is: Gα13.2-mTurquoise2-Δ9 n = 20, Gα13.3-mTurquoise2-Δ9 n = 16 and Gα13.5-mTurquoise2-Δ9 n = 11. The data from panel A are acquired on multiple days. (B) Ratiometric FRET traces of HeLa cells expressing the LPA2 Receptor (untagged), Gα13.2-mTurquoise2-Δ9 and either Gγ tagged with cp173Venus (n = 38) (and untagged Gβ) or Gβ tagged with mVenus (n = 25) (and untagged Gγ). LPA was added at t = 50s, indicated by the arrowhead. The data from panel B are acquired on the same day.

Fig 4. Development and characterization of Gα13 activation FRET based biosensors.

(A) Architecture of the Gα13 biosensor construct, encoding Gβ-2A-cp173Venus-Gγ₂-IRES-Gα13-mTurquoise2-Δ9, under control of the CMV promoter. (B) CFP and YFP emission was measured from individual cells expressing the Gα13.2 sensor from a single plasmid or from cells transfected with separate plasmids that encoded Gα13.2 and cp173Venus-Gγ₂. The r² is the correlation coefficient. (C) Confocal images showing the localization of the Gα13 in the sensor variants (upper, cyan) and cp173Venus-Gγ₂ (lower, yellow) in HeLa cells (for Gα13.2 sensor localization in Hek293T and HUVEC see S3 Fig). The width of the images is 75μm. (D) FRET ratio traces of HUVECs expressing the different Gα13 biosensors, stimulated with Thrombin at t = 100s (dotted lines depict 95% CI). For the corresponding YFP and CFP traces see S4 Fig. The number of cells analyzed is: Gα13.2 sensor n = 16, Gα13.3 sensor n = 11, Gα13.5 sensor n = 16.

Fig 5. Effects of the p115-RhoGEF RGS domain on Gα13.2 activity and cell morphology.

(A) Normalized ratiometric traces (upper graphs) and corresponding YFP and CFP traces (lower graphs) (dotted lines depict 95% CI) of HUVECs that were transfected with either the Gα13.2 FRET sensor and Lck-mCherry (Control, n = 11) or the Gα13.2 FRET sensor and Lck-mCherry-RGS (+ RGS, n = 13). Cells were stimulated at t = 110s. (B) Ratiometric images of representative cells measured in (A). Cool colors represent low YFP/CFP ratios, corresponding to emission ratios (ERs) on the right.(C) Cell area change of the cells measured in (B), visualized according to the LUT panel on the right. Dotplots on the right represent individual measurements (± 95% CI) of corresponding cells measured in (A). Image width = 54μm.

Fig 6. Direct observation of Gα13 and Gαq activation by different GPCRs.

Normalized ratio-metric FRET traces of HeLa cells transfected with the Gq sensor (grey line) or the G13.2 sensor (black line) (dotted lines depict 95% CI). (A) As a control, cells expressing only the Gq (n = 37) or G13.2 (n = 20) sensor were measured. Agonists were sequentially added after 50s, 150s and 230s of imaging. (B) Ratio traces of cells transfected with an untagged LPA2 receptor next to the Gαq (n = 13 (out of 60 in total)) or the Gα13.2 (n = 14 (out of 37 in total)), stimulated at t = 50s. (C) Ratio traces of cells transfected with AngiotensinII type 1 receptor-P2A-mCherry next to the Gαq (n = 22) or the Gα13.2 (n = 9) sensor, stimulated at t = 50s. (D) Ratio traces of cells transfected with an untagged kiss-receptor next to the Gαq (n = 13) or the Gα13.2 (n = 30) sensor, stimulated at t = 50s (indicated with the arrowhead).