Conditioning the heart induces formation of signalosomes that interact with mitochondria to open mitoKATP channels

Abstract

Perfusion of the heart with bradykinin triggers cellular signaling events that ultimately cause opening of mitochondrial ATP-sensitive K+ (mitoKATP) channels, increased H2O2 production, inhibition of the mitochondrial permeability transition (MPT), and cardioprotection. We hypothesized that the interaction of bradykinin with its receptor induces the assembly of a caveolar signaling platform (signalosome) that contains the enzymes of the signaling pathway and that migrates to mitochondria to induce mitoKATP channel opening. We developed a novel method for isolating and purifying signalosomes from Langendorff-perfused rat hearts treated with bradykinin. Fractions containing the signalosomes were found to open mitoKATP channels in mitochondria isolated from untreated hearts via the activation of mitochondrial PKC-epsilon. mitoKATP channel opening required signalosome-dependent phosphorylation of an outer membrane protein. Immunodetection analysis revealed the presence of the bradykinin B2 receptor only in the fraction isolated from bradykinin-treated hearts. Immunodetection and immunogold labeling of caveolin-3, as well as sensitivity to cholesterol depletion and resistance to Triton X-100, attested to the caveolar nature of the signalosomes. Ischemic preconditioning, ischemic postconditioning, and perfusion with ouabain also led to active signalosome fractions that opened mitoKATP channels in mitochondria from untreated hearts. These results provide initial support for a novel mechanism for signal transmission from a plasma membrane receptor to mitoKATP channels.

Fig. 1.

Schematic representation of the experimental protocol. Each experiment required mitochondria from a perfused donor heart and mitochondria from untreated assay hearts. Mitochondrial fractions from both hearts were purified on a 24% Percoll density gradient. Light layers (LL) from donor hearts were assayed for their ability to open mitochondrial ATP-sensitive K⁺ (mitoK_ATP) channels in assay mitochondria. In separate experiments, LL were further purified [purified LLs (PLLs)] and used for light scattering assays, immunoblot analysis, and electron microscopy.

Fig. 2.

Bradykinin perfusion causes a persistent open state of mitoK_ATP channels. Shown are changes in mitoK_ATP channel activity (in %) measured in mitoplasts prepared from mitochondria isolated from perfused rat hearts. Mitoplasts from sham-perfused (no treatment) and bradykinin-perfused hearts were compared for the response of mitoK_ATP channels to ATP (0.2 mM), ATP + protein phosphatase 2A (PP2A; 11 ng/ml), ATP + diazoxide (Dzx; 30 μM), and ATP + PP2A + Dzx. mitoK_ATP channels from bradykinin-perfused hearts were open and unresponsive to ATP unless PP2A was added. This indicates that the persistence of the open state was due to phosphorylation. Data are averages ±SD of 3 experiments from 3 separate perfusions. *P < 0.05 vs. bradykinin-treated mitochondria without PP2A treatment.

Fig. 3.

LLs and PLLs from bradykinin (brady)-treated hearts open mitoK_ATP channels in isolated mitochondria. Mitochondrial matrix volume is plotted versus time. Aliquots of LLs (10–15 μg/ml), PLLs (0.5 μg/ml), and residual LLs (RLLs; 10–15 μg/ml) were added to the medium 1 s after mitochondria. ATP (0.2 mM) was present in each run. KT-5823, a PKG inhibitor, was added at 0.5 μM. Traces are representative of at least 5 independent experiments. Six separate traces were plotted, and some of them are superimposed.

Fig. 4.

Properties of LL-induced mitoK_ATP channel opening. Shown are changes in mitoK_ATP channel activity (in %). Standard mitoK_ATP channel activity was observed in the absence of LLs (no LLs). LLs from sham-perfused hearts (sham LLs) had no effect on mitoK_ATP channel activity (n = 3). Caveolae isolated from untreated hearts (50) did not open mitoK_ATP channels at amounts up to 700 μg/mg mitochondrial protein. (Shown are data using 300 μg/mg caveolae; n = 3.) Bradykinin treatment of LLs induced mitoK_ATP channel opening that was inhibited by 5-hydroxydecanoate (5-HD; 1.0 mM) and ɛV_1-2 (0.5 μM). Note that a higher concentration of 5-HD was required to block mitoK_ATP channels after they had been opened by phosphorylation (8). Bradykinin-treated LLs opened mitoK_ATP channels to the same extent as Dzx, and Dzx had no additional effect in the presence of bradykinin-treated LLs. Bradykinin data are averages ±SD of at least 4 independent experiments. *P < 0.05 vs. sham LLs or caveolae.

Fig. 5.

LL-induced mitoK_ATP channel opening after ischemic preconditioning (IPC) and ouabain treatment. Shown are changes in mitoK_ATP channel activity (in %) when LLs treated with bradykinin, IPC, ouabain, or ischemic postconditioning (PostCon) were added to assay mitochondria. Also shown are the effects of PP2A (11 ng/ml) and KT-5823 (KT; 0.5 μM) added to the assay. Note that PP2A cannot cross the mitochondrial outer membrane (MOM) of intact mitochondria, so the effect is on the MOM. Data are averages ±SD of 3–4 independent experiments. *P < 0.05 vs. LLs alone.

Fig. 6.

LL-induced inhibition of the mitochondrial permeability transition (MPT). LL inhibition of the MPT in isolated heart mitochondria and the effects of various antagonists are shown. Data are plotted as MPT inhibition (in %). The following additions were made: LLs (10–15 μg/ml), Dzx (30 μM), 5-HD (0.3 mM), KT (0.5 μM), and PP2A (11 ng/ml). Data are shown as averages ±SD of at least 4 independent experiments. *P < 0.05 vs. bradykinin-treated LLs.

Fig. 7.

Bafilomycin A or methyl-β-cyclodextrin (methyl-β-CD) abolish LL-induced mitoK_ATP channel opening; Triton X-100 does not. Shown are changes in mitoK_ATP channel activity (in %). These experiments examined the effects of bafilomycin A (50 nM), methyl-β-CD (5%), and Triton X-100 (1%) on mitoK_ATP channel opening induced by bradykinin-treated LLs. Bafilomycin A was added to the heart perfusate before the isolation of LLs. Methyl-β-CD, a cholesterol-binding agent, was incubated with LLs at 30°C for 5 min before the assay. Triton X-100 was incubated with purified LLs for 10 min at 4°C. Before the assay, vesicles were pelleted in a microcentrifuge (16,000 g for 15 min), washed twice in dilute salt solution, and resuspended in buffer without Triton X-100. Dzx (30 μM) was added at the same time as the aliquot of LLs (10–15 μg/ml). Data are averages ±SD of at least 3 independent experiments. *P < 0.05 vs. corresponding LLs.

Fig. 8.

Immunodetection analysis of bradykinin-treated signalosomes. A: representative immunoblots of 6 independent experiments. Mitochondria protein (MITO; 5 μg) from perfused untreated hearts or 5 μg of PLLs from untreated (sham perfusion) or bradykinin-treated hearts, obtained as described in materials and methods, were loaded. For each antigen, the signal shown was obtained on samples loaded on the same gel and processed under the same conditions. eNOS, endothelial nitric oxide synthase; VDAC, voltage-dependent anion channel. B: pixel density analysis performed on each band of n = 6 Western blots. BG, pixel density of a protein-free region of the blot. *P < 0.05 vs. sham perfusion; **P < 0.05 vs. mitochondria.

Fig. 9.

Immunogold staining of bradykinin-treated signalosomes. Transmission electron microscopic analysis of PLLs from bradykinin-perfused hearts is shown. Samples were applied to carbon-coated grids, fixed, and processed for the presence of immunogold labeling when treated either with the primary antibody to caveolin-3 (A) or without the primary antibody as a negative control (B). Magnification: ×100,000 in A and ×67,000 in B.

Fig. 10.

The signalosome hypothesis. It is proposed that the interaction of bradykinin with its receptor [bradykinin B₂ receptor (Bk2)] induces the formation of a vesicular caveolar signaling platform (signalosome) that phosphorylates a receptor (R1) on the MOM. The identity of R1 is unknown. The terminal kinase of the bradykinin-treated signalosome is PKG, which phosphorylates R1 (∼P) at a serine/threonine residue. After phosphorylation of the MOM receptor, the signal is transmitted across the intermembrane space to activate PKC-ɛ₁ on the mitochondrial inner membrane. PKC-ɛ₁, which is constitutively expressed in mitochondria (9) and localized in close association with the mitoK_ATP channel (23), causes mitoK_ATP channel opening (8) and a consequent increase in H₂O₂ production (1). H₂O₂ then activates a second PKC-ɛ, PKC-ɛ₂, which leads to a reduction in necrosis through inhibition of MPT (9).