Stimulus-specific distinctions in spatial and temporal dynamics of stress-activated protein kinase kinase kinases revealed by a fluorescence resonance energy transfer biosensor

Abstract

The stress-activated protein kinases (SAPKs), namely, p38 and JNK, are members of the mitogen-activated protein kinase family and are important determinants of cell fate when cells are exposed to environmental stresses such as UV and osmostress. SAPKs are activated by SAPK kinases (SAP2Ks), which are in turn activated by various SAP2K kinases (SAP3Ks). Because conventional methods, such as immunoblotting using phospho-specific antibodies, measure the average activity of SAP3Ks in a cell population, the intracellular dynamics of SAP3K activity are largely unknown. Here, we developed a reporter of SAP3K activity toward the MKK6 SAP2K, based on fluorescence resonance energy transfer, that can uncover the dynamic behavior of SAP3K activation in cells. Using this reporter, we demonstrated that SAP3K activation occurs either synchronously or asynchronously among a cell population and in different cellular compartments in single cells, depending on the type of stress applied. In particular, SAP3Ks are activated by epidermal growth factor and osmostress on the plasma membrane, by anisomycin and UV in the cytoplasm, and by etoposide in the nucleus. These observations revealed previously unknown heterogeneity in SAPK responses and supplied answers to the question of the cellular location in which various stresses induce stimulus-specific SAPK responses.

FIG. 1.

FRET reporter for SAP3K activity. (A) Schematic model of the SAP3K activity reporter. The amino acid sequence around the phosphorylation sites (red dots) in wild-type (WT) MKK6 and the modified sequence in L225 are shown. L1, L2, and L3, linkers 1, 2, and 3, respectively. (B) Emission spectra of the SAP3K reporter L225 in the presence (red) or the absence (black) of cotransfected MTK1ΔN. AU, arbitrary units. (C) YFP/CFP emission ratios (means ± standard errors of the means [SEM]) of the SAP3K reporter L225 cotransfected with various MAP3Ks and/or their activators. G45β, Gadd45β. (D and E) Dose-dependent changes in the L225 YFP/CFP emission ratio (D) and phosphorylation of the L225 reporter (E) with increasing amounts of activating MTK1ΔN. K/R, kinase-dead MTK1ΔN K1371R mutant; IB, immunoblot.

FIG. 2.

Effect of various mutations in L225 on the MTK1ΔN-induced FRET change. In vivo YFP/CFP emission ratio changes (means ± SEM) of the SAP3K reporter L225, or its mutant derivatives, cotransfected without (gray) or with (white) the constitutively active MTK1ΔN kinase are shown. FHA1-R70A is a non-phosphothreonine-binding mutant of the FHA1 domain. MKK6-S207A and MKK6-T211A are nonphosphorylating mutants of the MKK6 activation loop. MKK6-V328G is a nondocking mutant of the MKK6 DVD domain. WT, wild type.

FIG. 3.

Detection of SAP3K activation in individual cells. (A and B) Cell images (A) and time courses (B) of the YFP/CFP emission ratios of L225 reporters in response to TNF-α using HeLa cells. T/A, T211A; V/G, V328G. (C) Phosphorylation of endogenous MKK3/6 following stimulation of HeLa cells by TNF-α. Total cell lysates were probed either with anti-phospho-MKK3/6 antibody (top) or with anti-MKK3/6 antibody (bottom). IB, immunoblot. (D) Phosphorylation of L225 reporters following stimulation of the transfected HeLa cells by TNF-α as described above for B. Anti-GFP antibody reacts with YFP and CFP in L225 (bottom). (E and F) Expression of the dominant negative TAK1-K63W mutant (E) or TAK1 shRNAi (F) inhibits TAK1 activation by TNF-α (10 ng/ml for 30 min), as revealed by a decrease in the change in the YFP/CFP ratio [Δ(YFP/CFP)] of L225. Vec, vector; Cont, control. (G) shRNAi inhibition of TAK1 expression. Immunoblot analysis of the effect of TAK1-specific shRNAi on TAK1 protein expression is shown. The control is luciferase-specific shRNAi. A nonspecific band serves as a loading control.

FIG. 4.

Reversibility of L225 reporter activity. HeLa cells expressing the L225 reporter were stimulated with IL-1β (10 ng/ml) at time a, washed out at time b, and restimulated with IL-1β at time c. The time course of the YFP/CFP emission ratio was monitored for two separate cells (black and gray curves).

FIG. 5.

Cells stably expressing YL225. (A) Images of the YFP/CFP ratio of HEK293A-YL225 cells before and after TNF-α (10 ng/ml) stimulation. (B) Time course of the change in YFP/CFP ratios of cells shown in panel A. The means ± SEM (n = 9) are shown. (C) Immunoblot analysis of the anisomycin (10 ng/ml)- and UV-C (80 J/m²)-induced phosphorylation of endogenous p38, JNK, MKK3, MKK6, MKK4, and the stably expressed YL225 probe.

FIG. 6.

Temporal dynamics of SAP3K activation in single cells. (A and B) Immunoblot (IB) analyses of MKK3/6 phosphorylation in HeLa cells following stimulation with anisomycin (A) or MMS (B). The stars indicate nonspecific bands. (C and D) Cell images of the L225 YFP/CFP emission ratios in response to anisomycin (Aniso) (C) or to MMS (D). HeLa cells were transfected with the L225 expression plasmid. Numbers are minutes after stimulation. Three different fields from one picture were spliced together to save space (D). (E and F) Time course of the changes in the YFP/CFP ratio in HeLa cells stimulated with either anisomycin (E) or MMS (F). Black curves represent the six cells shown in panels C and D, respectively, while the red curves represent the average values.

FIG. 7.

Temporal dynamics of SAP3K activation by various stimuli in single cells. (A to D) Time course of the changes in the YFP/CFP ratio (ΔYFP/CFP) in individual HeLa cells expressing L225 (A to C) or YL225-pm (D) following stimulation (stim.) with either 80 J/m² UV-C irradiation (A), 0.4 M sorbitol (B), 10 ng/ml TNF-α (C), or 100 ng/ml EGF (D). Average traces are shown in red. For UV-C, image acquisition was started at 5 min after UV irradiation. (E) Distribution of the time required to reach 50% of the maximum ΔYFP/CFP increase in individual cells, calculated from the data sets shown in panels A to D and in Fig. 6E and F.

FIG. 8.

Etoposide, but not UV-C, activates SAP3Ks in the nucleus. (A) Schematic diagram of the nucleus-directed SAP3K probe L225-nuc, which contains tandem NLS sequences (2xNLS) at the C terminus. (B) Subcellular localization of YL225-nuc. The L225-nuc probe was transfected into COS-7 cells, and the blue-green fluorescence of SECFP was monitored by epifluorescence microscopy. An intense fluorescence of SECFP was observed only in the nuclear region. (C) Cell images of the L225-nuc YFP/CFP emission ratio. Two images from images that were continuously acquired every minute starting at 10 min after UV-C irradiation (80 J/m² UV-C) are shown. (D) COS-7 cells coexpressing L225-nuc and L225-cyt were treated with 100 nM etoposide (ETO) (100 nM; continuous exposure in minimal essential medium with 10% FCS). Cell images of the YFP/CFP emission ratio at the indicated times following the start of etoposide treatment are shown.

FIG. 9.

Epifluorescence FRET imaging discriminates plasma membrane-localized signals from cytoplasmic signals. (A) Schematic diagrams of the cytoplasm-directed SAP3K reporter YL225-cyt and the plasma membrane-directed SAP3K reporter YL225-pm. (B) Dose-dependent changes in the YFP/CFP ratios of YL225-cyt (closed circles) and YL225-pm (open circles). The means ± SEM (n = approximately 9 to 32) were plotted. (C) COS-7 cells triply coexpressing YL225-cyt, YL225-pm, and the active form of MTK1 (MTK1ΔN) were imaged by epifluorescence microscopy, and the images of the YFP/CFP ratios are shown. (D to F) Either YL225-cyt, YL225-pm, or both were substituted by the corresponding inactive reporter that contains the MKK6-T211A mutation (T/A). A gradient of FRET signals, higher in the peripheral zone (red) and lower in the central area (blue-green), was observed when the cytoplasm-directed reporter was inactive (D). The converse gradient with a centrally higher FRET signal was observed when the plasma membrane-directed reporter was inactive (E). When both reporters were inactive, no signal was detected (F). No significant signal was observed in the nucleus in all cases.

FIG. 10.

Activation of SAP3K at the cell peripheries by EGF. (A) COS-7 cells coexpressing both YL225-cyt and YL225-pm were stimulated with EGF. Images of the YFP/CFP ratio (Y/C) were taken immediately before (0 min) and at the indicated times after EGF (100 ng/ml) stimulation. (B and C) HeLa cells expressing either YL225-cyt (cyt) or YL225-pm (pm) were stimulated with either EGF (100 ng/ml) or TNF-α (10 ng/ml). Changes in the YFP/CFP ratio (ΔYFP/CFP) in individual cells were monitored by epifluorescence FRET imaging, and averages for approximately 9 to 19 cells were plotted. Error bars represent SEM.

FIG. 11.

Spatiotemporal dynamics of SAP3K activation in single cells. (A) Subcellular distributions of SAP3K activities in COS-7 cells challenged by hyperosmotic stress (0.4 M sorbitol [sorb], continuous exposure), ribotoxic stress (10 ng/ml anisomycin [aniso], continuous exposure), and UV irradiation (80 J/m² UV-C, irradiated at time zero) were imaged. Numbers in the panels are times (minutes) after stimulation. (B) Propagation of stress-induced SAP3K activities as visualized by montages of images of YFP/CFP (Y/C) ratios of the white rectangular region along the radius from the nucleus (bottom) to the cell periphery (top). Cells were stimulated at 0 min, and images were continuously obtained every minute.