A FRET biosensor reveals spatiotemporal activation and functions of aurora kinase A in living cells

Abstract

Overexpression of AURKA is a major hallmark of epithelial cancers. It encodes the multifunctional serine/threonine kinase aurora A, which is activated at metaphase and is required for cell cycle progression; assessing its activation in living cells is mandatory for next-generation drug design. We describe here a Förster's resonance energy transfer (FRET) biosensor detecting the conformational changes of aurora kinase A induced by its autophosphorylation on Thr288. The biosensor functionally replaces the endogenous kinase in cells and allows the activation of the kinase to be followed throughout the cell cycle. Inhibiting the catalytic activity of the kinase prevents the conformational changes of the biosensor. Using this approach, we discover that aurora kinase A activates during G1 to regulate the stability of microtubules in cooperation with TPX2 and CEP192. These results demonstrate that the aurora kinase A biosensor is a powerful tool to identify new regulatory pathways controlling aurora kinase A activation.

Figure 1. The AURKA biosensor detects the autophosphorylation of AURKA on Thr288 in vitro.

(a) Model illustrating the mode of action of the AURKA biosensor. The complete sequence of AURKA is located between the donor (D, EGFP) and the acceptor (A, mCherry) fluorophores. When AURKA is autophosphorylated on Thr288, the kinase undergoes a conformational change bringing the donor and the acceptor in proximity and allowing FRET detection. Of note, the real three-dimensional orientations of the two fluorescent proteins are not known. (b) (Left panels) Representative fluorescence (GFP channel) and lifetime images from in vitro FLIM analysis of purified GFP-AURKA and GFP-AURKA-mCherry proteins. (Right panel) The graph illustrates a time-lapse analysis of the fluorescence lifetime of EGFP for both proteins. Images were acquired every 5 min. Data represent means±s.e.m. of three independent experiments. (c) (Left panels) Representative fluorescence (GFP channel) and lifetime images taken at selected time points, and (right panel) corresponding quantification of the in vitro FLIM analysis of GFP-AURKA and GFP-AURKA-mCherry following λPP and ATP treatments. All treatments were performed at 30 °C and images were acquired every 5 min. The addition of λPP and ATP is indicated by an arrow on the graph. Data represent means±s.e.m. of three independent experiments. The pseudocolour scale in b,c represents pixel-by-pixel lifetimes; conditions and/or time points are indicated in italics. Scale bar, 5 μm. ***P<0.001 against each time point in the corresponding ‘GFP-AURKA' condition; NS, not significant. Statistical tests: two-way ANOVA. (d) Representative in vitro kinase assay and corresponding western blot analysis showing the proportion of AURKA undergoing autophosphorylation on Thr288 (AURKA pThr288) in samples incubated at 30 °C with λPP for 1 h and then treated or not with ATP for 1 h at the same temperature. Loading control: total AURKA. (e) Representative in vitro kinase assay illustrating the presence of a Ser10-positive band on histone H3 after the incubation of GFP-AURKA and GFP-AURKA-mCherry with ATP for 30 min at 37 °C. Loading control: phosphoThr288 AURKA.

Figure 2. Autophosphorylation on Thr288 is sufficient for a catalytically active AURKA biosensor in the presence of ATP.

(a) (Left panels) Representative fluorescence (GFP channel) and lifetime images taken at the indicated time points and (right panel) corresponding quantification of EGFP lifetime images taken every 5 min from GFP-AURKA, GFP-AURKA-mCherry or GFP-AURKA Lys162Met-mCherry samples treated λPP for 1 h at 30 °C, and then incubated with ATP for 1 h at 30 °C. Data represent means±s.e.m. of three independent experiments. (b) In vitro kinase assay and western blot analysis showing the abundance of AURKA pThr288 in samples containing GFP-AURKA-mCherry or the Lys162Met variant following treatment with λPP for 1 h at 30 °C and then incubated or not with ATP for 1 h at the same temperature. (c) (Upper panels) Representative fluorescence (GFP channel) and lifetime images, and (lower panels) corresponding quantification of EGFP lifetime from GFP-AURKA or GFP-AURKA-mCherry samples treated with λPP for 1 h at 30 °C, and imaged following the addition of ATP together with dimethylsulfoxide (DMSO), MLN8237 or MLN8054 for 20 min at 37 °C. Images were acquired every 2 min. Data represent means±s.e.m. of three independent experiments. (d) In vitro kinase assay and corresponding western blot illustrating the abundance of a Ser10-positive band on histone H3 and of autophosphorylated AURKA after the incubation of GFP-AURKA and GFP-AURKA-mCherry with ATP and DMSO, MLN8237 or MLN8054. Scale bar, 5 μm. Arrows: addition of λPP/ATP. Pseudocolour scale: pixel-by-pixel lifetime. *P<0.05, ***P<0.001 against each time point in the corresponding ‘GFP-AURKA' condition in a or in the corresponding ‘Phosphatase+DMSO' condition in c. NS, not significant. Statistical tests: two-way ANOVA.

Figure 3. The AURKA biosensor displays intramolecular FRET in vitro.

(a) Representative fluorescence (GFP channel) and lifetime images taken at selected time points, and corresponding quantification of the in vitro FLIM analysis of the lifetime of EGFP from GFP-AURKA, GFP-AURKA-mCherry or GFP-AURKA and AURKA-mCherry samples incubated at 30 °C with λPP for 1 h, and then treated with ATP for 1 h. Images were acquired every 5 min. Data represent means±s.e.m. of three independent experiments. ***P<0.001 against each time point in the corresponding ‘GFP-AURKA' condition. Statistical test: two-way ANOVA. Arrows: addition of λPP/ATP. (b) (Left panels) Representative fluorescence (GFP channel) and lifetime images of GFP-AURKA-mCherry incubated with the indicated quantities of untagged AURKA. (Right panel) Histograms show the corresponding quantification of the lifetime of EGFP obtained from the GFP-AURKA-mCherry protein treated as in the left panels. Data represent means±s.e.m. of three independent experiments. Comparisons were made against the ‘+glycerol 50%' condition. Pseudocolour scale: pixel-by-pixel lifetime. Scale bar, 5 μm. NS, not significant. Statistical test: one-way ANOVA.

Figure 4. The AURKA biosensor rescues the phenotype induced by AURKA deficiency and it is active at discrete subcellular locations.

(a) (Left panels) Immunofluorescent micrographs of non-transfected, GFP-AURKA and GFP-AURKA-mCherry stable U2OS cells silenced for endogenous AURKA and synchronized in metaphase. The mitotic spindle and centrosomal defects in cells following AURKA depletion and the corresponding rescue by GFP-AURKA and by GFP-AURKA-mCherry were detected by labelling the mitotic spindle with TUBA1A and the centrosomes with TUBG1. DNA was stained with DAPI. (Right panel) Quantification of the proportion of cells with centrosomal defects in the three cell lines transfected with a control- or an AURKA-specific siRNA. n=100 cells per condition scored in each of three independent experiments. (b) Representative fluorescence (GFP channel) and corresponding lifetime images of GFP-AURKA and GFP-AURKA-mCherry U2OS stable cell lines not synchronized (left panels) or synchronized in mitosis (right panels) and illustrating the presence of both proteins at the centrosome, in the cytosol and at mitotic spindles. Graphs: corresponding quantification of the lifetime of EGFP in the two cell lines and in the indicated subcellular compartments. Centrosomes were labelled with CETN1-iRFP670. n=30–40 cells per condition from three independent experiments. (c) Quantification of the lifetime of EGFP GFP-AURKA and GFP-AURKA-mCherry cells synchronized at mitosis and treated with dimethylsulfoxide (DMSO) or with MLN8237. n=30–40 cells per condition from three independent experiments. (d) (Left panels) Fluorescence (GFP channel) and lifetime images of GFP-AURKA and GFP-AURKA-mCherry cells transfected with an AURKA-specific siRNA and synchronized at mitosis. (Right panel) Quantification of EGFP lifetime in GFP-AURKA or in GFP-AURKA-mCherry cells transfected with a control- or an AURKA-specific siRNA and synchronized at mitosis; n=30–40 cells per condition from three independent experiments. Data represent means±s.e.m. Pseudocolour scale: pixel-by-pixel lifetime. Scale bar, 10 μm. ***P<0.001 against the corresponding ‘GFP-AURKA' condition in a–d; ^aP<0.01 against the corresponding ‘GFP-AURKA-mCherry' condition at the centrosome in b and the corresponding MLN8327 condition in c. NS, not significant. Statistical tests: two-way ANOVA in a,c and d; Student's t-test in b.

Figure 5. The AURKA biosensor activates during the G1 phase and at mitosis.

(a) Representative immunocytochemical images of GFP-AURKA-mCherry U2OS cells synchronized in the indicated cell cycle phases. The fluorescence of EGFP was used to visualize AURKA, centrosomes were labelled with CETN1-iRFP670. N, nucleus. (b) Quantification of the fluorescence intensity of EGFP in GFP-AURKA and GFP-AURKA-mCherry cells synchronized in G1 phase, S phase and at mitosis. The fluorescence intensity of both proteins at mitosis was arbitrarily set at 100% and it was used to normalize the fluorescence intensities of EGFP in the other conditions. n=40–50 cells per condition from three independent experiments. Data represent means±s.e.m. (c) (Left panels) Representative fluorescence (GFP channel) and lifetime images of GFP-AURKA-mCherry cells and (right panel) corresponding quantification of the lifetime of EGFP in GFP-AURKA and GFP-AURKA-mCherry cells synchronized as in a. n=30–40 cells per condition from three independent experiments. Data represent means±s.e.m. (d) Quantification of the fluorescence recovery of EGFP in GFP-AURKA-mCherry cells synchronized as indicated and subjected to FRAP analysis. n=30 cells per condition from three independent experiments. Data represent means±s.e.m. (e) Degree of fluorescence recovery of EGFP after 60 s from the photobleaching procedure. Scale bar, 10 μm. *P<0.05, **P<0.01 and ***P<0.001 against the corresponding ‘mitosis' condition in b,d, the corresponding ‘GFP-AURKA' condition in c and against the ‘S' condition in e. ^aP<0.001 against the corresponding ‘S' condition within the GFP-AURKA-mCherry paradigm in c. NS, not significant. Statistical tests: one-way ANOVA in b,d and e; two-way ANOVA in c.

Figure 6. TPX2 and CEP192 are required for the activation of the AURKA biosensor during the G1 phase.

(a) (Left panels) Representative fluorescence (GFP channel) and lifetime images of GFP-AURKA and GFP-AURKA-mCherry cells synchronized in G1 and treated with dimethylsulfoxide (DMSO) or with MLN8237. (Right panel) Quantification of the lifetime of EGFP at the centrosomes of GFP-AURKA and GFP-AURKA-mCherry cells synchronized in G1 and treated as in the corresponding left panels. n=10–15 cells per condition from one experiment representative of three. (b) (Left and middle panels) Representative fluorescence (GFP channel) and lifetime images of GFP-AURKA or of GFP-AURKA-mCherry cells transfected with a control, a TPX2- or a CEP192-specific siRNA and synchronized at mitosis. (Right panel) Corresponding lifetime quantification. n=10–15 cells per condition from one experiment representative of three. (c) Images and quantification performed as in b, on cells synchronized in G1 and depleted for TPX2 or CEP192. n=10–15 cells per condition from one experiment representative of three. Data represent means±s.e.m. Pseudocolour scale: pixel-by-pixel lifetime. Scale bar, 10 μm. ***P<0.001 against the corresponding ‘GFP-AURKA' condition in a, *P<0.05 against the ‘CEP192 siRNA' condition within the GFP-AURKA-mCherry paradigm and ^aP<0.001 and ^bP<0.01 against the corresponding ‘GFP-AURKA' condition in b,c. NS, not significant. Statistical tests: two-way ANOVA in a; one-way ANOVA in b,c.

Figure 7. Microtubule dynamics of G1 cells is altered in the absence of TPX2 or CEP192.

(a) Representative immunofluorescence micrographs and corresponding quantification of centriole amplification in GFP-AURKA-mCherry cells depleted for TPX2 or CEP192, and synchronized in G1. Centrioles were labelled with CETN1-iRFP670. n=30 cells per condition from three independent experiments. (b) Immunofluorescence images of the TUBA1A network in cells transfected and synchronized as in a. The pseudocolour scale represents the measure of the intensity of the TUBA1A staining. (c) Representative time-lapse images and corresponding boxplots of the regrowth of the microtubule network at 0 and 150 s after nocodazole-mediated depolymerization in HeLa GFP-TUBA1A cells transfected with the indicated siRNAs. Line graph: complete time-lapse analysis of the microtubule regrowth. n=30 cells per condition from three independent experiments. N, nucleus. Scale bar, 10 μm in a,b, 5 μm in c. Data represent means±s.e.m. *P<0.05, *** P<0.001 against the ‘control siRNA' condition in a,c, ^aP<0.01 against the corresponding time-points in all other conditions. Statistical tests: one-way ANOVA.