Phosphatases control PKA-dependent functional microdomains at the outer mitochondrial membrane

Abstract

Evidence supporting the heterogeneity in cAMP and PKA signaling is rapidly accumulating and has been largely attributed to the localization or activity of adenylate cyclases, phosphodiesterases, and A-kinase-anchoring proteins in different cellular subcompartments. However, little attention has been paid to the possibility that, despite homogeneous cAMP levels, a major heterogeneity in cAMP/PKA signaling could be generated by the spatial distribution of the final terminators of this cascade, i.e., the phosphatases. Using FRET-based sensors to monitor cAMP and PKA-dependent phosphorylation in the cytosol and outer mitochondrial membrane (OMM) of primary rat cardiomyocytes, we demonstrate that comparable cAMP increases in these two compartments evoke higher levels of PKA-dependent phosphorylation in the OMM. This difference is most evident for small, physiological increases of cAMP levels and with both OMM-located probes and endogenous OMM proteins. We demonstrate that this disparity depends on differences in the rates of phosphatase-dependent dephosphorylation of PKA targets in the two compartments. Furthermore, we show that the activity of soluble phosphatases attenuates PKA-driven activation of the cAMP response element-binding protein while concurrently enhancing PKA-dependent mitochondrial elongation. We conclude that phosphatases can sculpt functionally distinct cAMP/PKA domains even in the absence of gradients or microdomains of this messenger. We present a model that accounts for these unexpected results in which the degree of PKA-dependent phosphorylation is dictated by both the subcellular distribution of the phosphatases and the different accessibility of membrane-bound and soluble phosphorylated substrates to the cytosolic enzymes.

Keywords: PKA; cAMP; mitochondria; phosphatases; signaling.

Fig. 1.

Small cAMP increases induce higher AKAR4 signals at the OMM of HeLa cells. (A) Calibration and comparison of H187 and OMM-H187 sensors. Intracellular buffer containing H187 or OMM-H187 (Insets) was complemented with increasing concentrations of cAMP, and FRET ratios were acquired after 10 min incubation. Each data point is the average of four independent experiments. (B) Calibration and comparison of AKAR4 and OMM-AKAR4 sensors. Intracellular buffer containing AKAR4 or OMM-AKAR4 (Insets) was complemented with increasing doses (units) of PKA catalytic subunit. FRET values were measured after 10–15 min incubation. Each data point is the average of five independent experiments. (C) Increases in cAMP produced in response to Fsk were not different at the cytosol or OMM of HeLa cells expressing the cAMP sensors H187 or OMM-H187. Data shown are the average ± SEM of 17 H187-expressing cells or 28 OMM-H187-expressing cells in four independent experiments. ns, not significant. (D) PKA-dependent phosphorylation in response to Fsk was higher at the OMM than at the cytosol in mixed populations of HeLa cells. Data are shown as the average ± SEM of 11 AKAR4-expressing cells or 13 OMM-AKAR4–expressing cells in four independent experiments.

Fig. 2.

Iso induces distinct PKA-dependent patterns in the cytosol and OMM without measurable differences in cAMP levels. (A) Coculture of cells expressing H187 or OMM-H187 subjected to increasing doses of Fsk. cAMP increases in response to Fsk were not different in the two compartments. (Inset) Average ± SEM of 21 H187-expressing cells and 12 OMM-H187-expressing cells in three independent experiments. F/I, Fsk 20 µM combined to IBMX 100 µM; ns, not significant. (B) Coculture of cells expressing AKAR4 or OMM-AKAR4 subjected to increasing doses of Iso. AKAR4 signals in response to Iso were significantly higher at the OMM than in the cytosol. (Inset) Average ± SEM of 23 AKAR4-expressing cells and eight OMM-AKAR4–expressing cells in three independent experiments. ***P < 0.001; ** P < 0.006. (C) NRVMs expressing H187 or OMM-H187 were challenged with increasing doses of Iso. cAMP increases in response to Iso were not different in the two compartments. (Upper Inset) Photomicrograph of a representative field (Magnification: 40×.) (representative ROIs: dotted lines, OMM-H187; solid lines, H187). (Lower Inset) Average ± SEM of 28 H187-expressing cells and 18 OMM-H187–expressing cells in three independent experiments. (D) NRVMs expressing AKAR4 or OMM-AKAR4 were subjected to increasing doses of Iso. AKAR4 signals in response to Iso were significantly higher at the OMM than in the cytosol. (Upper Inset) Photomicrograph of a representative field (Magnification: 40×.) (representative ROIs: dotted lines, OMM-AKAR4; solid lines. AKAR4). (Lower Inset) Average ± SEM of 26 AKAR4-expressing cells and 18 OMM-AKAR4–expressing cells in three independent experiments. ***P < 0.001; ** P < 0.005. (E) Cytosolic and mitochondrial fractions of NRVMs treated with Iso or Fsk combined with IBMX or vehicle control (DMSO). The phosphorylation status was assessed by a phospho-PKA substrate antibody, RRX(S/T)^P. The levels of the OMM marker Tom20 indicated efficient Proteinase K digestion. Purity and loading of the mitochondrial fractions were tested using an antibody mixture against the rodent OXPHOS subunits; GAPDH was used as a cytosolic marker. (F) Intensities of phospho-bands normalized to GAPDH for the cytosol or to OXPHOS for the mitochondria (five independent NRVM cultures). ***P < 0.001; ** P < 0.002; *P < 0.02).

Fig. 3.

Phosphatases are responsible for the different AKAR4 responses between the cytosol and OMM in NRVMs. (A) Western blotting of PKA components in soluble (S) and mitochondrial (M) fractions from three independent primary NRVM cultures. An antibody mixture against the rodent OXPHOS subunits and GAPDH assessed purity of mitochondria and cytosol, respectively. (B) PKA-dependent phosphorylation kinetics measured by OMM-AKAR4 (red trace) or cytosolic AKAR4 (black trace) in NRVMs. Challenge with 20 µM Fsk and 100 µM IBMX (F/I) resulted in saturation of both sensors. Upon rinsing the stimuli, the termination kinetics of the two sensors (depending on phosphatases) was drastically different, with OMM-AKAR4 being the slower of the two. Shown is an experiment representative of at least three independent repeats. (C) Western blotting testing the presence of PP2B, PP2A, and PP1 in soluble and mitochondrial fractions from primary NRVM cultures. An antibody mixture against the rodent OXPHOS subunits and GAPDH assessed the purity of mitochondria and cytosol, respectively. Shown is an experiment representative of three independent experiments. (D) Coculture of NRVMs expressing AKAR4 or OMM-AKAR4 challenged with Fsk (20 nM) followed by CalA (50 nM) and CsA (200 nM) to block phosphatases. Data shown are the average ± SEM of 39 AKAR4-expressing cells and 21 OMM-AKAR4-expressing cells in six independent experiments. ***P < 0.001). (E) HeLa cells expressing AKAR4 or OMM-AKAR4 or coexpressing OMM-AKAR4 with PP2Aα 1.2 µg/mL or 2.4 µg/mL (light and dark blue traces, respectively) or OMM-PP2Aα 1.2 µg/mL or 2.4 µg/mL (yellow and orange traces, respectively) were treated with Fsk 20 µM combined with IBMX 100 µM (F/I) to saturate the sensors. The addition of H89 inhibited PKA, unveiling the phosphatase effect on OMM-AKAR4. Overexpression of PP2Aα 1.2 µg/mL (22 cells) only marginally affected the dephosphorylation rate of OMM-AKAR4, whereas PP2Aα 2.4 µg/mL (22 cells), OMM-PP2Aα 1.2 µg/mL (56 cells), and 2.4 µg/mL (16 cells) significantly accelerated dephosphorylation, bringing it near that of AKAR4. Data shown are the average ± SEM of 32 AKAR4-expressing and 68 OMM-AKAR4–expressing cells. (F) Statistical analysis of the OMM-AKAR4 dephosphorylation rate based on the initial slope (approximately the first 90 s) of the ratio decrease after inhibiting PKA using 30 μM H89. Data are expressed as mean ± SEM; cell numbers are as in E. ***P < 0.001; ns, not significant.

Fig. 4.

Phosphatases limit PKA-dependent CREB activation in response to GPCR signals. (A) Confocal photomicrographs (Magnification: 60×.) of NRVMs treated for 24 h with Iso (100 pM), Fsk (20 nM), or, as control, with the PKA inhibitor H89 (10 µM) and loaded with MitoTracker-Red. (B) Quantification of mitochondrial elongation in NRVMs. Bar graphs show the average ± SEM of four independent blinded experiments (DMSO, 85 cells; Iso 100 pM, 49 cells; Fsk 20 nM, 53 cells; H89 10 µM, 85 cells). *P < 0.03). (C) Western blot showing Drp1 phosphorylation at S⁶⁵⁶ in total cell lysates of NRVMs treated with 100 pM of Iso or left untreated (NT). Total Drp1 and GAPDH were used as loading controls. (Inset) Intensities of phospho-Drp1 bands normalized to total Drp1 (three independent NRVM cultures). (D) AKAR4 responses in the nuclear (blue trace) and cytosolic (light purple trace) fractions of cells in response to increasing Fsk concentrations. Shown are representative traces of one experiment (four cells) of three independent experiments (12 cells). (E) AKAR4 responses in the nuclear (blue trace) and cytosolic (light purple trace) fractions of cells in response to the addition of CalA and low Fsk (20 nM). Shown are representative traces of one experiment (six cells) of six independent experiments (30 cells). (F) Relative quantification of CREB activity using a CREB-Luciferase reporter in NRVMs. Cells were challenged for 18–24 h with Iso (100 pM), CalA (5 nM), or a combination of both. Differences are expressed relative to the vehicle control (DMSO). All results were normalized for total protein. Bar graphs show the average ± SEM of three to five independent experiments. (G) Real-time PCR analysis of NR4A3 expression in NRVMs. Shown are the levels of the CREB target NR4A3 in response to different stimuli; data shown are the average ± SEM of three to five independent experiments. (H) Competitive ELISA immunoassay for total cAMP levels. Bar graphs show the average ± SEM of four to six independent experiments.