A Reliable System for Quantitative G-Protein Activation Imaging in Cancer Cells

Abstract

Fluorescence resonance energy transfer (FRET) biosensors have proven to be an indispensable tool in cell biology and, more specifically, in the study of G-protein signalling. The best method of measuring the activation status or FRET state of a biosensor is often fluorescence lifetime imaging microscopy (FLIM), as it does away with many disadvantages inherent to fluorescence intensity-based methods and is easily quantitated. Despite the significant potential, there is a lack of reliable FLIM-FRET biosensors, and the data processing and analysis workflows reported previously face reproducibility challenges. Here, we established a system in live primary mouse pancreatic ductal adenocarcinoma cells, where we can detect the activation of an mNeonGreen-Gαi3-mCherry-Gγ2 biosensor through the lysophosphatidic acid receptor (LPAR) with 2-photon time-correlated single-photon counting (TCSPC) FLIM. This combination gave a superior signal to the commonly used mTurquoise2-mVenus G-protein biosensor. This system has potential as a platform for drug screening, or to answer basic cell biology questions in the field of G-protein signalling.

Keywords: FLIM; FRET; G-protein signalling; GPCR; LPA; LPAR; PDAC; fluorescence lifetime; live cell imaging.

Figure 1

Comparison of signal intensity and photobleaching between the mNeonGreen- mCherry and the mTurquoise2-mVenus G-Protein Biosensors. (A) Schematic of the mNeonGreen- mCherry biosensor plasmid expressing all three subunits of the heterotrimeric G-protein complex under the control of the CMV promoter to achieve similar expression levels. Representative signal intensity images of PDAC cells transiently transfected with the (B) mNeonGreen-mCherry and (C) mTurquoise2-mVenus biosensor taken with our 2-photon TCSPC FLIM system (TriMScope). In both (B,C), intensity scales are shown on the bottom right. (D) Quantification of signal intensity comparisons between the two biosensors. Signal intensity was measured on three different days for each biosensor (n = 3), with 3 cells per biosensor quantified each day (n = 9 datapoints per biosensor). (E) Quantification of photobleaching for the two biosensors. Solid lines represent the mean signal, while the shaded area represents the range of values of three replicates. In both (D,E), y-axes show arbitrary intensity units.

Figure 2

The mNeonGreen-mCherry G-protein biosensor is activated by serum (FCS; fetal calf serum), and LPA. For serum stimulation, a final concentration of medium with 3.3% FCS was used. The final concentration of the LPAR1/3 inhibitor (Ki16425) used was 10 μM. Final LPA concentration used was 1 μM. The independent t-test was used to calculate statistical significance with the statannot Python package (*: 1.00 × 10⁻² < p ≤ 5.00 × 10⁻²). The p values for the serum stimulation, LPA stimulation, and inhibitor vs. vehicle are 0.02, 0.03, and 0.04, respectively. The data visualisation web tool SuperPlotsofData [26] was used to present the data. The different colours used show separate replicates, with the larger dot indicating the mean. Distribution per replicate is also shown. Cohen’s d (Cd) was calculated with the numpy (v. 1.26.4) package in Python, and reports the effect size. LPA; lysophosphatidic acid. The inhibitor vehicle was dimethylsulfoxide (DMSO).

Figure 3

Examples of signal intensity (in grayscale) and false-coloured lifetime images. (A) PDAC cells expressing mNeonGreen-Gαi3 only. (B,C) PDAC cells expressing the full G-protein biosensor. (B) PDAC cells pre- and post-addition of starvation medium. (C) PDAC cells pre- and post-serum stimulation. Signal intensity colour bar from 0 to 450 arbitrary intensity units. Lifetime colour bar from 2000 to 3500 picoseconds. Scale bar, 50 μm. (Note that for quantification, only segmented cell membranes were used).

Figure 4

Different approaches to FLIM data visualisation. Data from this work were used here for demonstration purposes. (A) Paired point plot example. (B) Box plot example. (C) Bar plot example. The plots were created using Python and the pandas, matplotlib, statannot and seaborn packages. Note that statistical annotations are shown here for demonstration purposes only. (****: 1.00 × 10⁻⁴ < p ≤ 1.00 × 10⁻⁵); ns = not significant.

Figure 5

The LPA receptor inhibitor (Ki16425) cannot prevent G-protein activation after serum stimulation in PDAC cells even at high concentrations. Graph showing a lifetime increase after serum stimulation in cell samples treated with Ki16425 or vehicle (DMSO). Samples were treated with either 10 μΜ (left panel) or 100 μΜ (right panel) Ki16425. The independent t-test was used to calculate statistical significance with the statannot Python package (n.s: not significant). The data visualisation web tool SuperPlotsofData [26] was used to present the data. The different colours used show separate replicates, with the larger dot indicating the mean. Distribution per replicate is also shown. LPA; lysophosphatidic acid. DMSO; dimethylsulfoxide.