Role of Akt and Ca2 + on cell permeabilization via connexin43 hemichannels induced by metabolic inhibition

Biochimica et Biophysica Acta 1852 (2015) 1268–1277

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbadis

Role of Akt and Ca2 + on cell permeabilization via connexin43 hemichannels induced by metabolic inhibition Daniela Salas a,b,⁎, Carlos Puebla b, Paul D. Lampe c, Sergio Lavandero a, Juan C. Sáez b,d,⁎⁎ a Advanced Center for Chronic Diseases (ACCDIS) & Centro Estudios Moleculares de la Célula (CMEC), Facultad Ciencias Químicas y Farmacéuticas & Facultad Medicina, Universidad de Chile, Santiago, Chile b Departamento de Fisiología, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Santiago, Chile c Translational Research Program, Human Biology and Public Health Sciences Divisions, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Seattle, WA 98109, USA d Instituto Milenio, Centro Interdisciplinario de Neurociencias de Valparaíso, Valparaíso, Chile

a r t i c l e

i n f o

Article history: Received 20 October 2014 Received in revised form 25 February 2015 Accepted 6 March 2015 Available online 14 March 2015 Keywords: Surface hemichannels Intracellular Ca2+ Oxygen-glucose deprivation Akt activation

a b s t r a c t Connexin hemichannels are regulated under physiological and pathological conditions. Metabolic inhibition, a model of ischemia, promotes surface hemichannel activation associated, in part, with increased surface hemichannel levels, but little is known about its underlying mechanism. Here, we investigated the role of Akt on the connexin43 hemichannel's response induced by metabolic inhibition. In HeLa cells stably transfected with rat connexin43 fused to EGFP (HeLa43 cells), metabolic inhibition induced a transient Akt activation necessary to increase the amount of surface connexin43. The increase in levels of surface connexin43 was also found to depend on an intracellular Ca2+ signal increase that was partially mediated by Akt activation. However, the metabolic inhibition-induced Akt activation was not significantly affected by intracellular Ca2+ chelation. The Akt-dependent increase in connexin43 hemichannel activity in HeLa43 cells also occurred after oxygen-glucose deprivation, another ischemia-like condition, and in cultured cortical astrocytes (endogenous connexin43 expression system) under metabolic inhibition. Since opening of hemichannels has been shown to accelerate cell death, inhibition of Akt-dependent phosphorylation of connexin43 hemichannels could reduce cell death induced by ischemia/reperfusion. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Connexins (Cxs) constitute a protein family ubiquitously expressed in chordates that form two types of structures: hemichannels (HCs), pore-like structures in the cell surface that communicate the intracellular and extracellular media, and gap junction channels (GJCs), the result of the docking of two HCs from adjacent cells that communicate the cytosol of contiguous cells [1]. Once HCs are opened, free movement of ions and small molecules occurs down their concentration gradients. Because of the big size and low selectivity of HC pores, for many years it was believed that they were always closed, but currently it is known that they open under specific physiological and pathological conditions [2], including ischemia [3].

Abbreviations: Cx43, connexin43; Cx, connexin; HC, hemichannel; GJC, gap junction channel; OGD, oxygen-glucose deprivation; MI, metabolic inhibition; I/R, ischemia/ reperfusion; IP, ischemic preconditioning; Etd, ethidium. ⁎ Correspondence to: D. Salas, Advanced Center for Chronic Diseases (ACCDIS) & Centro Estudios Moleculares de la Célula (CMEC), Facultad Ciencias Químicas y Farmacéuticas & Facultad Medicina, Universidad de Chile, Santiago, Chile. Tel.: +56 2 29782903. ⁎⁎ Correspondence to: J.C. Sáez, Departamento de Fisiología, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Santiago, Chile. Tel.: + 56 2 26862862; fax: +56 2 22225515. E-mail addresses: [email protected] (D. Salas), [email protected] (J.C. Sáez).

http://dx.doi.org/10.1016/j.bbadis.2015.03.004 0925-4439/© 2015 Elsevier B.V. All rights reserved.

Ischemia increases the open probability of connexin43 (Cx43) HCs [4,5]. This has been associated with dephosphorylation or nitrosylation of the protein subunit and accelerated cell death [3,6–8]. Moreover, Cx43 seems to be essential for ischemic preconditioning (IP), a mechanism in which an organ is exposed to a non-lethal ischemia/reperfusion (I/R) insult, resulting in protection against a subsequent more harmful I/ R episode [9]. Metabolic inhibition (MI), a controlled ischemic-like condition, increases cell membrane permeability through Cx43 HCs in rat cardiomyocytes [10–12], astrocytes [5,7] and kidney epithelial cells [13]. Elevated Cx HC activity may be accomplished by either increasing the open probability or the number of HCs at the cell surface. MI has been proved to activate both of them. Some of the mechanisms that lead to the opening of Cx43 HCs and are activated during MI are: decreased phosphorylation status [3,7,8], reduced redox potential [8], increased intracellular [Ca2+] [14,15] and decreased extracellular [Ca2+] [16,17]. MI also increases the cell surface levels of Cx32 and Cx43 [8, 18,19], but the mechanism behind this phenomena is still unknown. The structure of Cx43 has multiple phosphorylation sites that are dynamically modified and affect membrane insertion and degradation [20] as well as the activity of Cx43 GJCs and HCs [1,21,22]. Akt is a Ser/Thr kinase activated during I/R with an important role during IP [23]. It can phosphorylate Cx43 on Ser369 and Ser373 [24], affecting the size of Cx43 gap junctions during hypoxia/reperfusion [25] and Cx43 HC activity induced by mechanical stress [26], but a similar mechanism operating

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277

under pathological conditions remains unknown. The aim of this work was to investigate the role of Akt on Cx43 HC activity in metabolically inhibited cells. Our results show that MI-induced membrane permeabilization due to enhanced Cx43 HC activity is mediated by a rapid and transient activation of Akt, leading to an increase in Cx43 HC levels at the cell surface by a mechanism dependent on intracellular Ca2+ signal. 2. Materials and methods 2.1. Cell cultures Parental HeLa cells (HeLa-p) were obtained from ATCC (CCL-2; ATCC, Rockville, MD) and HeLa cells stably transfected with cDNA encoding rat Cx43 fused to EGFP in its C-terminus (HeLa43 cells) were kindly provided by Dr. Felix Bukauskas (Department of Neuroscience, Albert Einstein College of Medicine, NY, USA). Construction of the pIRES vector containing Cx43 with Ser373 mutated by Ala (Cx43S373A) was previously described [25]. All cell lines were cultured in DMEM supplemented with 10% fetal bovine serum (GIBCO-Invitrogen), 100 μg/ml streptomycin sulfate, 100 U/ml penicillin and 250 μg/ml G418. The latter was used to select transfected cells. Untransfected HeLa-p cells were used as negative control. For ethidium (Etd) uptake experiments, cells were seeded at a density of 60,000 cells/plate in 60mm-diameter dishes (Nunc, Roskilde, Denmark), containing several #1 glass coverslips. For biotinylation experiments (see below), cells were seeded at a density of 120,000 cells/plate in 100-mm-diameter dishes (Nunc). In all experiments, cells were fed 24 h prior to the experiment. Cells for primary cultures were obtained using protocols approved by the Ethics and Biosecurity Committee of the Faculty of Biological Sciences, Pontificia Universidad Católica de Chile, Santiago, Chile. Primary astrocyte cultures were obtained from the cortex of newborn (1–2 d of age) Sprague–Dawley rats as described previously [5]. Briefly, meningeal tissue was removed, and the neocortex was minced and incubated at 37°C for 20 min in PBS containing trypsin (0.5%) and EDTA (5 mM). Cells were then removed by centrifugation and resuspended in MEM medium supplemented with 10% horse serum and 1 mg/ml bovine pancreas DNase I and triturated using a Pasteur pipette. Dissociated cells were centrifuged. The pellet was resuspended in MEM supplemented with 10% FBS, 100 μg/ml streptomycin sulfate and 100 U/ml penicillin and cells were finally plated in 25 ml plastic tissue culture bottles (Nunc Clone). For Etd uptake experiments, confluent cell cultures were replated on 60 mm plastic culture dishes (Nunc Clone), each one containing several #1 glass coverslips at a density of 2.5 × 103 cells per culture dish. Cells were grown at 37°C in a 5% CO2/95% air atmosphere at nearly 100% relative humidity. 2.2. Transfection protocol HeLa-p were seeded on 24 well plates (Nunc, Roskilde, Denmark) at a density of 4 × 104 cells/well, containing one #1 glass coverslip and cultured in DMEM-10%FBS for 24 h. When cells reached 60% confluence, they were washed 3 times with PBS and cultured in Opti-MEM (GIBCO-Invitrogen). After 1 h, a mixture of lipofectamine® 2000 (Invitrogen) and 500 ng of Cx43-S373A cDNA (within vector pIRES) was added to each well and incubated for 3 h at 37°C in a 5% CO2/95% air atmosphere at nearly 100% relative humidity. Finally, the complex lipofectamine® 2000-cDNA was replaced by fresh medium and cells were cultured for 48 h until they were used for Etd uptake experiments. 2.3. Metabolic inhibition and ischemia/reperfusion Metabolic inhibition (MI) was induced by blocking the two main pathways for ATP production through inhibition of the enzyme glyceraldehyde-3-phosphate dehydrogenase (0.3 mM iodoacetic acid dissolved in saline), and the complex III of the mitochondrial respiratory

1269

chain (10 ng/ml antimycin A, dissolved in absolute ethanol) [7,10,27]. Both inhibitors (Sigma, St. Louis, MO) were simultaneously applied to cultures. Ischemic-like conditions (called oxygen-glucose deprivation, or OGD) were induced by cultivating cells in a solution mimicking changes that occurred in an ischemic environment [in mM: NaCl (139), KCl (12), MgCl2 (0.5), CaCl2 (1.3), 2-deoxy-D-glucose (5), lactic acid (20), HEPES (10), pH 6.8] under 100% nitrogen (O2 b 1%) for 8 h. Reperfusion was achieved by taking the cells out of the hypoxic environment and placing them in a 37°C chamber until Etd uptake experiments were performed (approximately 30 min). Parallel controls were simultaneously incubated in medium containing [in mM: NaCl (139), KCl (4.7), MgCl2 (0.5), CaCl2 (1.3), D-glucose (5), HEPES (10), 10% fetal bovine serum, pH 7.4] in a 37°C chamber for 8 h. Etd uptake experiments were performed identically in both groups of cells. 2.4. Dye uptake HC activity was assessed by using the Etd uptake method [2]. Briefly, a coverslip with approximately 80% confluent HeLa cells was transferred to a 30-mm dish (Nunc Clone) coated with a thin vaseline layer to adhere and immobilize the coverslip. Then, cells were washed twice with recording solution [in mM: NaCl (150), KCl (5.4), HEPES (5), CaCl2 (2.3), MgCl2 (1.5)] containing 5 μM Etd. Basal fluorescence intensity (before MI) from selected regions of interest was recorded for 10 min, after which, MI was induced by adding chemicals to the recording solution. After 50 min of MI, 200 μM lanthanum ion (La3+), a Cx HC blocker, was added to the recording solution in order to confirm that Etd uptake was mediated by HCs. Dye uptake was recorded in an Olympus BX51WI upright microscope with a 40× water immersion objective (Melville) and equipped with the image acquisition system Q Imaging, model Retiga 13001, fast-cooled monochromatic digital camera (12-bit) (Qimaging, Burnaby, BC, Canada). Images were captured every 1 min (exposure time = 30 ms, gain 0.5). For off-line image analysis and fluorescence quantification, Metafluor software (version 6.2R5, Universal Imaging Co., Downingtown, PA, USA) was used. For graphical data representation and slope calculation, the average of three independent background fluorescence intensity measurements was subtracted from the fluorescence intensity of each cell. Results from this calculation, including at least 30 cells per independent experiment, were averaged and plotted against time (in min). Slopes, calculated with Microsoft Excel software, were expressed in arbitrary units/minute: A.U./min. Microscope and camera settings remained constant in all experiments. For experiments with divalent cation-free solution (DCFS, without Ca2+ and Mg2+), basal Etd uptake was first recorded for 10 min in a solution containing normal divalent cation concentrations. Then, cells were rinsed five times with DCFS and supplemented with ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA, Merck) before Etd uptake was recorded for additional 10 min. At the end of the experiment, 200 μM La3+ was added to the recording solution. In experiments where Akt inhibitor (Akti) was used, cells were pretreated with 10 μM Akti for 1 h and maintained during Etd uptake measurements. For experiments with BAPTA, cells were pre-incubated with 5 μM BAPTA-AM for 30 min and then rinsed twice with recording solution before the experiment. 2.5. Determination of the amount of surface Cx43 HCs Biotinylation of cell surface proteins was performed as described previously [8]. In brief, cell cultures were grown in 100 mm-dishes (Nunc Clone), washed three times with ice-cold Hank's saline solution pH 7.4 and incubated for 30 min at 4°C in 3 ml of 0.5 mg/ml sulfo-NHS-SSbiotin (Pierce, Rockford, IL). Then, cells were washed three times with a saline solution containing 15 mM glycine pH 8.0 to neutralize unreacted biotin and finally harvested in saline solution with a cocktail of protease

1270

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277

and protein phosphatase inhibitors [in mM: trypsin inhibitor from soybean (200 μg/ml), benzamidine (6.4), ε-aminocaproic acid (7.6), EDTA (20), phenylmethanesulfonyl fluoride (3.2), sodium pyrophosphate (100) and NaF (100)]. After this, cells were sonicated on ice (Ultrasonic Processor, GEX130PB, Cole-Parmer) and protein concentration was determined by using the Bio-Rad protein assay (Bio-Rad, Richmond, CA). Volumes containing the same amount of proteins were taken and an excess of immobilized NeutrAvidin (Pierce) was added (1 ml of NeutrAvidin per 3 g of biotinylated protein). The mixture was incubated for 1 h at 4°C. Then, 1 ml of washing buffer (saline solution, pH 7.4, 0.1% SDS and 1% NP-40) was added, and the mix was then centrifuged for 2 min at 14,000 rpm in a bench centrifuge at 4°C. The supernatant was discarded. After performing this procedure three times, 40 μl Hanks solution buffer at pH 2.8 containing 0.1 M glycine was added to the pellet to release the proteins bound to biotin. The pellet was resuspended and centrifuged at 14,000 rpm for 2 min at 4°C. The supernatant was transferred to a new 1.5 ml Eppendorf tube, and pH was adjusted immediately by adding 10 μl of 1 M Tris pH 7.4. Finally, relative levels of protein were evaluated by Western blot analysis as described below. The density of the bands stained with specific Cx43 antibody (previously described [8]) was measured, and the results were expressed in arbitrary units (A.U.). These relative levels were compared to total protein extracts from the same biotinylated fractions stained with Ponceau red before the biotinylation process was performed. Changes in cell surface Cx43-EGFP fluorescence were evaluated on subconfluent HeLa43 cells seeded on glass coverslips and mounted on a hermetic camera suitable for an inverted confocal microscope. EGFP fluorescence (excitation wave length 488 nm and emission at 530 nm) was used as indicative of Cx43 signal because of the fusionprotein. Confocal images of living cells under control conditions and after exposure to 30 min of MI were acquired using the Spectral Imaging Confocal Microscope Digital Eclipse C2 (Nikon, Japan) equipped with a 60 × Plan apochromat CFI NA 1.4 oil objective and analyzed using the NIS-Elements AR software provided by Nikon. Cells treated with BAPTA or Akti were stimulated as described before (“Dye uptake”). Changes in continuous EGFP fluorescence of four rectangular sections per cell (from unopposed regions of cells) were analyzed between control and MI conditions. Data were corrected considering photobleaching.

2.6. Western blot analysis Cell cultures were rinsed three times with cold PBS (pH 7.4) and harvested by scraping them with a rubber policeman in ice-cold solution containing protease and phosphatase inhibitors (detailed above). In aliquots of cell homogenates, proteins were measured using the Bio-Rad protein assay. Pelleted cells were resuspended in 100 μl of the protease and phosphatase inhibitor solution, placed on ice, and lysed by a brief sonication (Ultrasonic Processor, GEX130PB, Cole-Parmer). Subsequently, samples were analyzed by immunoblotting. Briefly, aliquots of cell lysates [100 μg of protein for anti-phospho-Akt in S473 (p-AktS473) or 50 μg for anti-Akt antibodies] were resuspended in Laemli's sample buffer, separated in 10% SDS-PAGE, and electro-transferred to nitrocellulose sheets. Nonspecific protein binding was blocked by incubation of membranes in 5% nonfat milk in TBS, 1% Tween-20 buffer for 60 min at room temperature. Blots were then incubated with primary polyclonal antiAkt or p-AktS473 (both from Cell Signaling, Danvers, MA) sera overnight at 4°C (each primary antibody was diluted in 5% nonfat milk in TBS, 1% Tween-20 buffer at 1:1000 dilution) followed by five 20-min TBS, 1% Tween-20 buffer washes. Membranes were then incubated with goat anti-rabbit secondary antibody, conjugated to horseradish peroxidase (1:5000 in 5% nonfat milk in TBS, 0.1% Tween-20). Finally, immunoreactivity was detected by ECL using the SuperSignal kit (Pierce, Rockford, IL) according to the manufacturer's instructions. Akt and p-AktS473 were detected in two different gels because of the different amount of

protein required for detection of each protein, but it was the same sample and the experiments were run in parallel. 2.7. Intracellular Ca2+ signal measurement Intracellular Ca2+ signal (F340/F380) measurements were performed as previously described [28]. Briefly, cells plated on glass coverslips were loaded with 5 μM Fura-2-AM (Molecular Probes, Eugene, OR) in DMEM (serum free) for 30 min at 37°C and washed three times in recording solution [in mM: NaCl (150), KCl (5.4), HEPES (5), CaCl2 (2.3), MgCl2 (1.5), pH = 7.4]. Fluorescence intensity was measured at excitation wavelengths of 340 and 380 nm, while the emission wavelength was 510 nm. Images were acquired by using an Olympus BX 51W1I upright microscope and the imaging system described above. Data were recorded every 1 min and analyzed with METAFLUOR software (Universal Imaging, Downingtown, PA) as described previously [29]. For experiments with Akti, cells were first incubated for 30 min with 10 μM Akti and then loaded with Fura2 by incubating them for additional 30 min in saline solution containing 5 μM Fura2-AM. 2.8. Statistical analysis For each group of data, results were expressed as mean ± SEM, and n represents the number of independent experiments or the number of cells analyzed as indicated. For Etd uptake or FURA-2 experiments, each mean corresponds to the average of at least 30 cells. Data sets were compared by one-way analysis of variance (ANOVA) followed by a Bonferroni's post-test. Differences were considered significant at p ≤ 0.05. The analyses were performed with GraphPad Prism 5 software for Windows (1992–2007, GraphPad Software). 3. Results 3.1. The activity of Cx43 HCs expressed by HeLa cells increases under metabolic inhibition Metabolic inhibition (MI) is used as a model of ischemia because it shares important features such as reduction in ATP levels and generation of free radicals despite the fact that it occurs in normoxic conditions [7,8]. To study the effect of MI on Cx43 HC mediated cell membrane permeability, HeLa-p (deficient in Cx and pannexin expression) and HeLaCx43-EGFP (called HeLa43 hereafter) were treated with antimycin A and iodoacetic acid to induce MI. Changes in the rate of Etd uptake were measured as an indicator of HC activity. No significant changes in Etd uptake rate were detected in HeLa-p under MI for 50 min, and a similar result was verified for HeLa43 cells under control conditions (“C”) during the same time period (Fig. 1a–b). Basal Etd uptake (“B”, first 10 min of Etd recording, before MI was performed) was similar in HeLa-p and HeLa43, but MI increased Etd uptake (5.1 ± 0.3 times basal Etd uptake) only in HeLa43 and was completely inhibited when La3+, a known Cx HC inhibitor [7,10,11], was added to the recording solution (Fig. 1b–c). Addition of La3+ to HeLa-p cells did not significantly affect the Etd uptake rate (Fig. 1b–c). 3.2. Increased Cx43 HC activity induced by MI is mediated by Akt activation Akt activation plays a critical role in models of hypoxia or I/R [25,30]. The evidence that Cx43 is a substrate of this Ser/Thr kinase [24] led us to investigate its involvement in Cx43 HC activity increase induced by MI. The Akt activity was measured by Western blot analysis as the increase in Ser473 phosphorylation (pAktS473) relative to total Akt. The latter is activated rapidly and transiently in HeLa43 cells metabolically inhibited, reaching maximum levels during the first 3 min and returning to values below the control conditions at 15 min of MI (Fig. 2a). Inhibition of Akt activity with Akti almost completely abolished the Cx43HC activity increase induced by MI (Fig. 2b–c, 1.7 ± 1.0 times basal Etd uptake

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277

1271

Fig. 1. Metabolic inhibition increases the membrane permeability in HeLa cells transfected with Cx43 fused to EGFP (HeLa43). Cultured HeLa cells were placed in a recording solution containing 5 μM Etd and dye uptake was recorded every 1 min as fluorescence emission of Etd binding to DNA (518 nm arbitrary units—A.U.—of fluorescence intensity). After 10 min, cells were metabolically inhibited and the Etd uptake was recorded during 50 min. At the end of the experiment, La 3+ (200 μM) was added to the recording solution. (a) Photomicrographs of Etd incorporation (upper panels) and bright field images (lower panels) of cells exposed to 5 μM Etd during 50 min of control conditions (C, left panels) or after 50 min of metabolic inhibition (MI, right panels) in HeLa-p or HeLa43. (b) Left graph: time course of Etd uptake in HeLa43 (black circle) or HeLa-p (open circle) exposed to MI (indicated by arrow) or HeLa43 exposed to control conditions (C, grey circle). The black line under La3+ indicates the period of time of La3+ treatment. Each point is the average ± SEM of fluorescence of at least 30 cells. (c) Rate of Etd uptake (A.U./min) of HeLa-p (open bar) or HeLa43 (black bar) during the first 10 min of recording (C), after 50 min of MI (MI) or after La3+ was added. Data are mean ± SEM of 5 (HeLa-p) or 9 (HeLa43) independent experiments. ***p ≤ 0.001, one-way ANOVA with Bonferroni's post-test.

rate). In HeLa-p cells, La3+ had no significant effect on Etd uptake, as well as in HeLa43 cells pre-incubated with Akti and then exposed to MI (1.1 ± 0.1 times basal Etd uptake rate) (Fig. 2b–c). The Akt inhibitor effect over kinase activation was corroborated by Western blot analysis in which 1 h Akti pretreatment completely abolished the Akt activation induced by MI (Fig. 2a). To determine whether Akti is a direct Cx43 HC inhibitor, HeLa43 cells were exposed to a divalent cation-free solution (DCFS), a condition used to increase the open probability of Cx HCs [31]. Akti did not significantly change the Etd uptake rate in cells exposed to DCFS (Suppl. Fig. 1), suggesting that its action on Etd uptake induced by MI was mainly through Akt inhibition. Akt is a pleiotropic kinase that affects many molecular targets downstream. To determine if Akt-dependent phosphorylation of Cx43 is responsible of the increased Cx43 HC activity induced by MI, we used HeLa-p cells transiently transfected with Cx43 with Ser373 (one of the Akt phosphorylation sites on Cx43) mutated to Ala. A positive response

to DCFS was used to select transfected cells with functional, cell-surface Cx43S373A HCs. For this, cells were first exposed to DCFS for 5 min, then returned to normal divalent cation levels and then MI was performed. Only cells that increased HC activity in response to DCFS were considered in the analysis. MI did not significantly affect Cx43 HC activity in cells transfected with Cx43-S373A (Fig. 3a–b). Thus, Ser373 phosphorylation induced by Akt may account for increased Cx43 HC activity induced by MI. 3.3. MI increases intracellular Ca2+ signal, which depends on Akt activation and is required for increased HC activity To determine if MI increases Ca2+ signal as was reported for HeLa cells expressing Cx32 [19], changes in Ca2+ signal were measured using the radiometric dye Fura-2. MI induced an increase in Ca2+ signal, which was slightly reduced by La3+ (Fig. 3a–b, black circles).

1272

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277

Fig. 2. Akt activation is necessary for Cx43 hemichannel activity induced by metabolic inhibition. (a) HeLa cells transfected with Cx43 fused to EGFP (HeLa43) were exposed to metabolic inhibition (MI) for different times in the presence or absence of 10 μM Akt inhibitor (Akti) as indicated at the bottom. Two top panels are representative Western blot analysis showing relative levels of total Akt (lower panel) or phosphorylated Akt on Ser473 (p-AktS473, upper panel). Graph: densitometric analysis represents levels of p-AktS473 relative to total Akt of 3 independent experiments. (b) Etd uptake rate of HeLa43 exposed to MI (indicated by arrow) in the presence of Akti (open circles) or in its absence (black circles). The black line under La3+ indicates the period of time of 200 μM La3+ treatment. Each point corresponds to the average ± SEM of fluorescence of at least 30 cells. (c) Rate of Etd uptake of HeLa43 exposed to 50 min of MI in the presence or absence of 10 μM Akti or after La3+ exposure, as indicated. Data are shown as Etd uptake mean ± SEM and the number of independent experiments is indicated in each column. **p ≤ 0.01, ***p ≤ 0.001, one-way ANOVA with Bonferroni's post-test.

Akt regulates cellular Ca2+ content by stimulating the translocation of L-type Ca2+ channels to the plasma membrane [32] and by increasing their activity [33]. For this reason, we investigated whether Akt is necessary for the intracellular Ca2+ signal increase induced by MI. Pretreatment with Akti drastically reduced the Ca2+ signal promoted by MI

(Fig. 4a–b, open circles) as compared to cells under MI but without Akt inhibition. The Ca2+ signal increase was necessary for the increased Cx43 HC activity induced by MI because pretreatment with BAPTA, a Ca2+-chelator, totally prevented this response (Fig. 4c, grey circles). Then, we

Fig. 3. Serine 373 phosphorylation is necessary for metabolic inhibition-induced Cx43 hemichannel activity. HeLa cells transfected with Cx43 fused to EGFP (HeLa43) or with Cx43 mutated at Ser373 to Ala (HeLa43-S373A) were exposed to metabolic inhibition (MI). (a) Etd uptake of HeLa43 (black circles) or HeLa43-S373A (open circles) exposed to MI (arrow).The black line under La3+ indicates the period of time of 200 μM La3+ application. Each point corresponds to the average ± SEM of fluorescence of at least 30 cells. (b) Rate of Etd uptake of HeLa43-S373A exposed to 50 min of MI or divalent cation free solution (DCFS). Data are shown as Etd uptake mean ± SEM of 3 independent experiments. **p ≤ 0.01, one-way ANOVA with Bonferroni's post-test.

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277

1273

Fig. 4. Metabolic inhibition increases Cx43 hemichannel-dependent membrane permeability by increasing intracellular Ca2+ levels. (a) Pseudo color images of representative experiments of Ca2+ signal measurements in HeLa cells transfected with Cx43 fused to EGFP (HeLa43) incubated during 1 h with 10 μM Akt inhibitor (Akti, middle panels) or not (upper and lower panels) and loaded with 5 μM Fura-2 in control conditions (C, left panel) or after 50 min metabolic inhibition (MI, right panel) alone (upper panels) or in the presence of 200 μM La3+ (lower panel). Changes in intracellular Ca2+ signals are expressed as Fura-2 fluorescence ratio F340/F380. (b) Representative experiment showing intracellular Ca2+ signal variations over time of cells treated with 10 μM Akti (open circles) or not (black circles) and exposed to MI (arrow) alone or in the presence of La3+ (triangle) shown in (a) of 3 independent experiments. Black line under La3+ represents the time period of La3+ treatment. (c) HeLa43 cells loaded with 5 μM BAPTA (grey circles) or not (black circles) were exposed to MI (arrow) and changes in Etd uptake were recorded over time. Black line under La3+ represents the time period of La3+ application. Each point corresponds to the Etd fluorescence mean ± SEM of at least 30 cells. (d) HeLa43 cells were loaded with 5 μM BAPTA or not (as indicated in the bottom) and metabolically inhibited during different periods of time. Representative Western blot analysis of total Akt (lower figure) and phosphorylated Akt on Ser473 (p-AKTS473, upper panel). Densitometric analysis represents the level of p-AktS473 relative to total Akt of 3 independent experiments. ***p ≤ 0.001, one-way ANOVA with Bonferroni's post-test.

tested if BAPTA affects the Akt phosphorylation state as was reported in cells stimulated with EGF [34]. BAPTA pretreatment tended to reduce Akt phosphorylation (pAktS473) status in HeLa43 cells under MI, but the effect was not statistically significant (Fig. 4d, 3 and 15 min). Considering that Cx43 HCs are permeable to Ca2+ [35], we tested whether their opening may explain the increase of intracellular Ca2+ signal induced by MI. To accomplish this, La3+ was added at the same time that MI was induced and changes of intracellular Ca2+ signal were measured with Fura-2. Inhibition of Cx43 HCs with La3+ abolished the increase of intracellular Ca2+ levels induced by MI (Fig. 4b, triangles).

and this increase can be positively correlated with the increased rate of Etd uptake in astrocytes [8] and HeLaCx32 cells under MI [19]. We used two experimental approaches to study whether MI increases the amount of Cx43 HCs at the plasma membrane in HeLa43 and the involvement of Ca2+ signal rise on Akt activation. Measurements of i) cell surface protein by biotinylation (Fig. 5a–b) and ii) EGFP fluorescence intensity at the cell membrane by confocal microscopy (Fig. 5c–d) were performed. Both experimental approaches revealed that MI increases the amount of Cx43 HCs at the cell surface, while this increase was absent in cells loaded with BAPTA (Fig. 5a, d) or treated with Akti (Fig. 5b, d).

3.4. Intracellular free Ca2+ and activation of Akt are required for the MI-induced increase in the amount of surface Cx43 HCs

3.5. The increase in Cx43 HC activity induced by oxygen-glucose deprivation is dependent upon Akt activation

Previous work showed that different cellular stresses, including MI, induce an increased level of Cx43 HCs in the plasma membrane [8,36]

To determine whether the role of Akt on MI-induced increase in Cx43 HC activity was independent of the way the metabolic stress was

1274

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277

Fig. 5. Ca2+ and Akt are required to increase the number of Cx43 hemichannels at the cell surface. HeLa cells transfected with Cx43 fused to EGFP (HeLa43) were loaded with 5 μM BAPTA for 30 min or incubated 1 h with 10 μM Akt inhibitor (Akti) and then exposed to metabolic inhibition (MI) for different periods of time. (a and b) Surface proteins were biotinylated, precipitated and subjected to immunoblot for Cx43. Representative images showing the amount of Cx43 hemichannels at the cell surface (upper panel) in HeLa43 cells loaded with BAPTA (a) or Akti (b). Ponceau staining (lower panels in a and b) corresponds to the total protein fraction, before purification of hemichannels and was used as a loading control. C = control conditions. (c and d) Changes in EGFP fluorescence at the cell surface determined using confocal microscopy. (c) Representative confocal image of HeLa43 cells showing changes in surface Cx43-EGFP signal (fluorescence at 530 nm)when exposed to MI for 30 min without treatment (upper panels), treated with Akti (middle panels) or BAPTA (lower panels). Insert shows a zoom image of one cell from the field (asterisk). Changes in continuous Cx43-EGFP fluorescence in cells exposed to control conditions or after 30 min of MI were analyzed in rectangular sections, as shown in the insert. (d) Average Cx43-EGFP fluorescence intensity of membrane domains was measured under control conditions or after exposure to 30 min of MI in the presence of BAPTA or Akti, as indicated at the bottom. The number inside the bars represents the number of cells analyzed in each condition. ***p ≤ 0.001, one-way ANOVA with Bonferroni's post-test.

performed, we treated cells with OGD under ischemic condition (see Material and methods). During the first 30 min of reperfusion after 8 h of OGD, Cx43 HC activity was increased (Fig. 6a, 3.3 ± 0.2 times control Etd uptake rate), but this was not evident if Akti was present during the ischemia period in HeLa43 cells (Fig. 6a, 1.8 ± 0.1 times control Etd uptake rate) or HeLa-p cells (Fig. 6a, white bar; 1.9 ± 0.2 times control Etd uptake rate).

system that endogenously expresses Cx43, we chose astrocytes because in culture they express only Cx43 [37,38] and Cx HC activity increases during MI [7,8]. MI was performed as in HeLa cells and changes in Etd uptake rate were recorded as indicative of Cx HC activation. Pretreatment with Akti significantly reduced the activity of Cx HC induced by MI (Fig. 6b). 4. Discussion

3.6. Akt inhibition partially reduces Cx HC activity increase triggered by MI in cortical astrocytes To determine whether the Akt-dependent mechanism described above that regulates Cx43 HC activity in cells under MI occurs in a

In the present work, we show that MI, a chemically induced model of I/R, increases Cx43 HC activity in HeLa43 through an Akt and Ca2+ signal-dependent mechanism. In addition, Akt activation was found to be required for the increase in Ca2+ signal. The role of Akt activation

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277

1275

Fig. 6. Akt activation is involved in Cx43 hemichannel activity in cells expressing endogenous Cx43 and in HeLa43 cells exposed to ischemia–reperfusion. (a) HeLa parental cells (HeLa-p, open bar) and HeLa cells transfected with Cx43 fused to EGFP (HeLa43, black bar) were subjected to 8 h oxygen-glucose deprivation (OGD: high K+, pH 6.8, 2-deoxyglucose, without serum or O2) in the presence or absence of 10 μM Akt inhibitor (Akti, indicated at the bottom). Hemichannel activity was measured as the rate of Etd uptake over time after reperfusion. Data are presented as average ± SEM of 3 (HeLa-p) or 7 (HeLa43) independent experiments. (b) Primary cultured rat astrocytes were pre-incubated during 1 h with or without Akti metabolically inhibited and Etd uptake rate was measured to evaluate Cx43 HC activity. Data are shown as mean ± SEM of 3 independent experiments. The line under La3+ indicates the period of time cells were treated with 200 μM La3+. ***p ≤ 0.001, one-way ANOVA with Bonferroni's post-test.

on Cx43 HC activity was also observed under OGD conditions and in cultures of cortical astrocytes, an endogenous Cx43 expression system. Therefore, we suggest that Akt orchestrates part of the Cx43 HC activity increase induced by ischemic-like conditions. MI-induced cell permeabilization has been described in different cell types including cardiomyocytes [11,12] cultured astrocytes [7,8], HeLa cells expressing Cx32 [19] and HEK293 cells expressing Cx43 wild type or fused to GFP [39]. To study the mechanism of cell permeabilization induced by MI, we choose HeLa cells stably transfected with Cx43 fused to EGFP in its C-terminus because they behave similar to wild type Cx43 [40] and do not express other functional Cx HCs or pannexin channels in the cell surface [41]. As expected and based on previous electrophysiological and dye uptake studies [40], the basal Cx43 HC activity was low. However, under MI the cellular permeabilization mediated by Cx43 HCs increased despite the presence of extracellular divalent cations. This is further supported by the following findings: 1) MI did not permeabilize HeLa-p cells, which do not express functional Cx HCs or pannexin channels in the cell surface; 2) there is a positive correlation between EGFP expression and Etd uptake (Suppl. Fig. 2) and 3) MI-induced cell permeabilization to Etd is inhibited by La3+, a known Cx43 HC blocker. Other pathways for Etd entry such as P2X receptors or TRPV1 channels were discarded based on previous findings performed in the same HeLa-p clone [19]. Moreover, CALHM1 channels, a recently described pathway of Etd entry sensible to La3+ [42], are not expressed in HeLa cells [43]. During the first 3 min of MI, we found a rapid and transient activation of Akt, reminiscent of Akt changes described in cardiomyocytes exposed to I/R [25]. The rapid turnover of Akt phosphorylation may be a consequence of the fast drop in intracellular ATP levels [2], but we did not perform further experiments to address this issue. The increase in Cx43 HC-mediated membrane permeability induced by MI was prevented by the Akt inhibitor. Since Akt has many downstream targets that may explain the increase in Cx43 HC activity induced by MI, we tested whether Akt-dependent Cx43 phosphorylation mediates this response. The membrane permeability of HeLa cells transiently transfected with Cx43S373A was found to be insensitive to MI, indicating that MI-induced Akt activation leads to phosphorylation of Cx43 on Ser373 increasing Cx43 HC activity. Akt has been described as a protective kinase important for ischemic pre- and post-conditioning [44], but its role was attributed to the inhibition of the opening of the mPTP [45,46] and prevention of apoptosis [47]. An alternative mechanism of cell protection could imply that Cx43 HCs are activated via an Akt-dependent mechanism leading to

pre-conditioning [48]. It has been previously shown that MI increases the intracellular Ca2+ signal, which is important for activation of Cx32 HCs expressed in HeLa cells [19]. Here, we extend this observation to Cx43 HCs. In the current work, this response was found to depend to a great extent on Akt activation and Cx43 HC opening. Cx43 HCs have been shown to be permeable to Ca2+ in HeLa43 cells treated with FGF-1 [28] or upon exposure to alkaline pH in Cx43 HCs reconstituted in liposomes [35]. Taken the above together with our observation that La3+ prevented the MI-induced increase of intracellular Ca2+, we suggest that Cx43 HCs may be the pathway for Ca2+ inflow. It is worth mentioning that in other cell types, it remains possible that the Aktmediated Ca2+ increase induced by MI could be mediated by the increase of L-type Ca2+ channel expression and/or its activity [32,33]. In a model of mammary epithelial cells, it was shown that the increase in Akt phosphorylation induced by EGF depends on intracellular Ca2+ levels [34]. Accordingly, we found that Ca2+ chelation with BAPTA tends to decrease the MI-mediated Akt activation in HeLa cells. Future experiments are needed to identify the exact mechanism(s). Akt-dependent Cx43 phosphorylation on Ser373 has been associated with prevention of ZO-1 interaction, thus increasing the gap junction size [25,49,50] and Cx43 HC-dependent membrane permeabilization [26]. It is tempting to argue that a similar mechanism might explain the increase in cell surface Cx43 HC levels, but further research is needed to address this issue. In this work, we focused on the effects of Akt over the amount of Cx43 HCs at the cell surface. We showed, by two different techniques (surface protein biotinylation and changes in the surface EGPFdependent fluorescence by confocal microscopy), that MI increases the amount of Cx43 HCs on the cell surface. This effect is dependent upon intracellular Ca2+ increase and Akt activation. The increase in the amount of Cx HCs induced by increased Ca2+ signal is consistent with experiments performed in HeLaCx32 cells under MI and in HeLa cells transfected with Cx43 or Cx45 treated with a Ca2+ ionophore [28]. However, further experiments are needed to pinpoint the mechanism underlying this process. Considering that Akt activation occurs before the increase in Ca2+ signal, another possible interpretation is that Akt activation fulfills its role by increasing the levels of intracellular Ca2+. The role of Akt may be related to the regulation of Rab protein activity, as it occurs with GLUT4 traffic to the cell surface induced by insulin which is mediated by AS160-dependent Akt phosphorylation [51]. To our knowledge, there is no evidence that Rab proteins regulated by Akt may affect Cx43 HC trafficking to the cell surface, but we cannot discard this mechanism. Further research is required to investigate the mechanism by which Akt regulates the

1276

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277

amount of Cx43 HCs at the cell surface. Our data showed that Aktdependent increase in Cx43 HC activity is a mechanism also present in cells endogenously expressing Cx43 (e.g., astrocytes and osteoblasts). Moreover, they suggest that Akt can be added to those regulators of Cx43 HC activity [5,7,8,52]. Using a different experimental approach to deplete intracellular ATP consisting in hypoxia, nutrient deprivation, high K+ and acidic pH to mimic the extracellular milieu of ischemic cells, our results showed that Akt-mediated regulation of Cx43 HCs was also present in this experimental model. Considering that linoleic acid, a pro-inflammatory unsaturated fatty acid, also increase Cx43 HC activity by a Ca2+/Akt dependent pathway [29], we propose that this mechanism is activated in general inflammatory states regardless the stimuli. Considering our results and those reported previously, Aktdependent regulation of Cx43 HCs and gap junctions seems to be a general mechanism operating in physiological (mechanic stress) [26] as well as pathologic conditions (hypoxia/reoxygenation, MI) [25,45], but further experiments should be done to test this hypothesis. Taking our results together, we propose the following mechanism: MI activates Akt, which in turn increases intracellular Ca2+ levels. Both signals increase the amount of cell surface Cx43 HCs, thus increasing the cell membrane permeability and finally affecting cell survival. The biological consequence should depend on features of the insult such as intensity and duration. Accordingly, brief and/or low intensity (sublethal insult) could be associated to preconditioning possibly due to enhanced adenosine release via Cx43 HCs [53], whereas, long and/ or intense insult lead to cell death associated to Ca2+ overload possibly related to persistent Cx43 HC opening [7]. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.bbadis.2015.03.004. Transparency document

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17] [18]

The Transparency document associated with this article can be found, in the online version. [19]

Acknowledgments We would like to thank Ms. Teresa Vergara and Ms. Paola Fernández for their technical support. This work was supported by Fondo Nacional de Desarrollo Científico y Tecnológico (FONDECYT): Grant 1111033 (J.C.S.), Chilean Science Millennium Institute P09-022-F (J.C.S.), Fondo de Investigación Avanzada en Areas Prioritarias (FONDAP) Grant: 15130011 (S.L.) and a grant from the US National Insitutes of Health, GM55632 (P.D.L.). D.S. holds a PhD fellowship (21080284 and AT24100026) from the Comisión Nacional de Investigación Científica y Tecnológica (CONICYT), Santiago, Chile. The data of this paper are from a thesis submitted in partial fulfillment of the requirements for the degree of Doctor in Biochemistry, (D.S.) at University of Chile, Santiago, Chile.

[20]

References

[26]

[1] J.C. Sáez, M.A. Retamal, D. Basilio, F.F. Bukauskas, M.V. Bennett, Connexin-based gap junction hemichannels: gating mechanisms, Biochim. Biophys. Acta 1711 (2005) 215–224, http://dx.doi.org/10.1016/j.bbamem.2005.01.014. [2] K.A. Schalper, N. Palacios-Prado, J.A. Orellana, J.C. Sáez, Currently used methods for identification and characterization of hemichannels, Cell Commun. Adhes. 15 (2008) 207–218, http://dx.doi.org/10.1080/15419060802014198. [3] M.A. Beardslee, D.L. Lerner, P.N. Tadros, J.G. Laing, E.C. Beyer, K.A. Yamada, A.G. Kleber, R.B. Schuessler, J.E. Saffitz, Dephosphorylation and intracellular redistribution of ventricular connexin43 during electrical uncoupling induced by ischemia, Circ. Res. 87 (2000) 656–662, http://dx.doi.org/10.1161/01.RES.87.8.656. [4] X. Li, K. Boengler, R. Schulz, G. Heusch, Role of connexin43 in ischemic preconditioning does not involve intercellular communication through gap junctions, J. Mol. Cell. Cardiol. 36 (2004) 161–163, http://dx.doi.org/10.1016/j.yjmcc.2003.10.019. [5] J.A. Orellana, D.E. Hernández, P. Ezan, V. Velarde, M.V. Bennett, C. Giaume, J.C. Sáez, Hypoxia in high glucose followed by reoxygenation in normal glucose reduces the viability of cortical astrocytes through increased permeability of connexin 43 hemichannels, Glia 58 (2010) 329–343, http://dx.doi.org/10.1002/glia.20926. [6] R. Schulz, P. Gres, A. Skyschally, A. Duschin, S. Belosjorow, I. Konietzka, G. Heusch, Ischemic preconditioning preserves connexin 43 phosphorylation during sustained

[21]

[22] [23]

[24]

[25]

[27]

[28]

[29]

[30]

[31]

ischemia in pig hearts in vivo, Faseb J. 17 (2003) 1355–1357, http://dx.doi.org/10. 1096/fj.02-0975fje. J.E. Contreras, H.A. Sánchez, E.A. Eugenín, D. Speidel, M. Theis, K. Willecke, F.F. Bukauskas, M.V. Bennett, J.C. Sáez, Metabolic inhibition induces opening of unopposed connexin 43 gap junction hemichannels and reduces gap junctional communication in cortical astrocytes in culture, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 495–500, http://dx.doi.org/10.1073/pnas.012589799. M.A. Retamal, C.J. Cortes, L. Reuss, M.V. Bennett, J.C. Sáez, S-nitrosylation and permeation through connexin 43 hemichannels in astrocytes: induction by oxidant stress and reversal by reducing agents, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 4475–4480, http://dx.doi.org/10.1073/pnas.0511118103. U. Schwanke, X. Li, R. Schulz, G. Heusch, No ischemic preconditioning in heterozygous connexin 43-deficient mice-a further in vivo study, Basic Res. Cardiol. 98 (2003) 181–182. S.A. John, S.Y. Wang, J.I. Goldhaber, J.N. Weiss, Connexin-43 hemichannels opened by metabolic inhibition, J. Biol. Chem. 274 (1999) 236–240, http://dx.doi.org/10. 1074/jbc.274.1.236. R.P. Kondo, S.Y. Wang, S.A. John, J.N. Weiss, J.I. Goldhaber, Metabolic inhibition activates a nonselective current through connexin hemichannels in isolated ventricular myocytes, J. Mol. Cell. Cardiol. 32 (2000) 1859–1872, http://dx.doi.org/10.1006/ jmcc.2000.1220. F. Li, K. Sugishita, Z. Su, I. Ueda, W.H. Barry, Activation of connexin-43 hemichannels can elevate [Ca(2+)]i and [Na(+)]i in rabbit ventricular myocytes during metabolic inhibition, J. Mol. Cell. Cardiol. 33 (2001) 2145–2155, http://dx.doi.org/10.1006/ jmcc.2001.1477. L. Vergara, X. Bao, M. Cooper, E. Bello-Reuss, L. Reuss, Gap-junctional hemichannels are activated by ATP depletion in human renal proximal tubule cells, J. Membr. Biol. 196 (2003) 173–184, http://dx.doi.org/10.1007/s00232-003-0636-9. D.W. Choi, Calcium-mediated neurotoxicity: relationship to specific channel types and role in ischemic damage, Trends Neurosci. 11 (1988) 465–469, http://dx.doi. org/10.1016/0166-2236(88)90200-7. N. Wang, M. De Bock, G. Antoons, A.K. Gadicherla, M. Bol, E. Decrock, W.H. Evans, K.R. Sipido, F.F. Bukauskas, L. Leybaert, Connexin mimetic peptides inhibit Cx43 hemichannel opening triggered by voltage and intracellular Ca2+ elevation, Basic Res. Cardiol. 107 (2012) 304, http://dx.doi.org/10.1007/s00395-012-0304-2. K. Ohta, R. Graf, G. Rosner, W.D. Heiss, Profiles of cortical tissue depolarization in cat focal cerebral ischemia in relation to calcium ion homeostasis and nitric oxide production, J. Cereb. Blood Flow Metab. 17 (1997) 1170–1181, http://dx.doi.org/10. 1097/00004647-199711000-00006. S.H. DeVries, E.A. Schwartz, Hemi-gap-junction channels in solitary horizontal cells of the catfish retina, J. Physiol. 445 (1992) 201–230. M.A. Retamal, K.A. Schalper, K.F. Shoji, J.A. Orellana, M.V. Bennett, J.C. Sáez, Possible involvement of different connexin43 domains in plasma membrane permeabilization induced by ischemia-reperfusion, J. Membr. Biol. 218 (2007) 49–63, http://dx.doi.org/10.1007/s00232-007-9043-y. H.A. Sánchez, J.A. Orellana, V.K. Verselis, J.C. Sáez, Metabolic inhibition increases activity of connexin-32 hemichannels permeable to Ca2+ in transfected HeLa cells, Am. J. Physiol. Cell Physiol. 297 (2009) C665–C678, http://dx.doi.org/10.1152/ ajpcell.00200.2009. D. Segretain, M.M. Falk, Regulation of connexin biosynthesis, assembly, gap junction formation, and removal, Biochim. Biophys. Acta 1662 (2004) 3–21. X. Bao, S.C. Lee, L. Reuss, G.A. Altenberg, Change in permeant size selectivity by phosphorylation of connexin 43 gap-junctional hemichannels by PKC, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 4919–4924, http://dx.doi.org/10.1073/pnas.0603154104. J.L. Solan, P.D. Lampe, Connexin43 phosphorylation: structural changes and biological effects, Biochem. J. 419 (2009) 261–272, http://dx.doi.org/10.1042/BJ20082319. E. Murphy, C. Steenbergen, Preconditioning: the mitochondrial connection, Annu. Rev. Physiol. 69 (2007) 51–67, http://dx.doi.org/10.1146/annurev.physiol.69. 031905.163645. D.J. Park, C.J. Wallick, K.D. Martyn, A.F. Lau, C. Jin, B.J. Warn-Cramer, Akt phosphorylates Connexin43 on Ser373, a “mode-1” binding site for 14-3-3, Cell Commun. Adhes. 14 (2007) 211–226, http://dx.doi.org/10.1080/15419060701755958. C.A. Dunn, P.D. Lampe, Injury-triggered Akt phosphorylation of Cx43: a ZO-1-driven molecular switch that regulates gap junction size, J. Cell Sci. 127 (2014) 455–464, http://dx.doi.org/10.1242/jcs.142497. N. Batra, M.A. Riquelme, S. Burra, R. Kar, S. Gu, J.X. Jiang, Direct regulation of osteocytic connexin 43 hemichannels through AKT kinase activated by mechanical stimulation, J. Biol. Chem. 289 (2014) 10582–10591, http://dx.doi.org/10.1074/jbc. M114.550608. G.J. Gores, C.E. Flarsheim, T.L. Dawson, A.L. Nieminen, B. Herman, J.J. Lemasters, Swelling, reductive stress, and cell death during chemical hypoxia in hepatocytes, Am. J. Physiol. Cell Physiol. 257 (1989) C347–C354. K.A. Schalper, N. Palacios-Prado, M.A. Retamal, K.F. Shoji, A.D. Martinez, J.C. Sáez, Connexin hemichannel composition determines the FGF-1-induced membrane permeability and free [Ca2+]i responses, Mol. Biol. Cell 19 (2008) 3501–3513, http://dx. doi.org/10.1091/mbc.E07-12-1240. V. Figueroa, P.J. Sáez, J.D. Salas, D. Salas, O. Jara, A.D. Martínez, J.C. Sáez, M.A. Retamal, Linoleic acid induces opening of connexin26 hemichannels through a PI3K/Akt/ Ca2 +-dependent pathway, Biochim. Biophys. Acta 1828 (2012) 1169–1179, http://dx.doi.org/10.1016/j.bbamem.2012.12.006. J.W. Mockridge, M.S. Marber, R.J. Heads, Activation of Akt during simulated ischemia/reperfusion in cardiac myocytes, Biochem. Biophys. Res. Commun. 270 (2000) 947–952, http://dx.doi.org/10.1006/bbrc.2000.2522. H. Li, T.F. Liu, A. Lazrak, C. Peracchia, G.S. Goldberg, P.D. Lampe, R.G. Johnson, Properties and regulation of gap junctional hemichannels in the plasma membranes of cultured cells, J. Cell Biol. 134 (1996) 1019–1030.

D. Salas et al. / Biochimica et Biophysica Acta 1852 (2015) 1268–1277 [32] P. Viard, A.J. Butcher, G. Halet, A. Davies, B. Nurnberg, F. Heblich, A.C. Dolphin, PI3K promotes voltage-dependent calcium channel trafficking to the plasma membrane, Nat. Neurosci. 7 (2004) 939–946, http://dx.doi.org/10.1038/nn1300. [33] D. Catalucci, D.H. Zhang, J. DeSantiago, F. Aimond, G. Barbara, J. Chemin, D. Bonci, E. Picht, F. Rusconi, N.D. Dalton, K.L. Peterson, S. Richard, D.M. Bers, J.H. Brown, G. Condorelli, Akt regulates L-type Ca2+ channel activity by modulating Cavalpha1 protein stability, J. Cell Biol. 184 (2009) 923–933, http://dx.doi.org/10.1083/jcb. 200805063. [34] T.B. Deb, C.M. Coticchia, R.B. Dickson, Calmodulin-mediated activation of Akt regulates survival of c-Myc-overexpressing mouse mammary carcinoma cells, J. Biol. Chem. 279 (2004) 38903–38911, http://dx.doi.org/10.1074/jbc.M405314200. [35] K.A. Schalper, H.A. Sánchez, S.C. Lee, G.A. Altenberg, M.H. Nathanson, J.C. Sáez, Connexin 43 hemichannels mediate the Ca2+ influx induced by extracellular alkalinization, Am. J. Physiol. Cell Physiol. 299 (2010) C1504–C1515, http://dx.doi.org/ 10.1152/ajpcell.00015.2010. [36] J.K. VanSlyke, L.S. Musil, Cytosolic stress reduces degradation of connexin43 internalized from the cell surface and enhances gap junction formation and function, Mol. Biol. Cell 16 (2005) 5247–5257, http://dx.doi.org/10.1074/jbc.M405314200. [37] R. Dermietzel, E.L. Hertberg, J.A. Kessler, D.C. Spray, Gap junctions between cultured astrocytes: immunocytochemical, molecular, and electrophysiological analysis, J. Neurosci. 11 (1991) 1421–1432. [38] C. Giaume, C. Fromaget, A. el Aoumari, J. Cordier, J. Glowinski, D. Gros, Gap junctions in cultured astrocytes: single-channel currents and characterization of channelforming protein, Neuron 6 (1991) 133–143, http://dx.doi.org/10.1016/08966273(91)90128-M. [39] S.A. John, R. Kondo, S.Y. Wang, J.I. Goldhaber, J.N. Weiss, Connexin-43 hemichannels opened by metabolic inhibition, J. Biol. Chem. 274 (1999) 236–240, http://dx.doi. org/10.1074/jbc.274.1.236. [40] J.E. Contreras, J.C. Sáez, F.F. Bukauskas, M.V. Bennett, Gating and regulation of connexin 43 (Cx43) hemichannels, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 11388–11393, http://dx.doi.org/10.1073/pnas.1434298100. [41] Y. Huang, J.B. Grinspan, C.K. Abrams, S.S. Scherer, Pannexin1 is expressed by neurons and glia but does not form functional gap junctions, Glia 55 (2007) 46–56, http://dx. doi.org/10.1002/glia.20435. [42] A.P. Siebert, Z. Ma, J.D. Grevet, A. Demuro, I. Parker, J.K. Foskett, Structural and functional similarities of calcium homeostasis modulator 1 (CALHM1) ion channel with connexins, pannexins, and innexins, J. Biol. Chem. 288 (2013) 6140–6153, http://dx. doi.org/10.1074/jbc.M112.409789.

1277

[43] A.J. Moreno-Ortega, A. Ruiz-Nuno, A.G. García, M.F. Cano-Abad, Mitochondria sense with different kinetics the calcium entering into HeLa cells through calcium channels CALHM1 and mutated P86L-CALHM1, Biochem. Biophys. Res. Commun. 391 (2010) 722–726, http://dx.doi.org/10.1016/j.bbrc.2009.11.127. [44] D.J. Hausenloy, D.M. Yellon, Survival kinases in ischemic preconditioning and postconditioning, Cardiovasc. Res. 70 (2006) 240–253, http://dx.doi.org/10.1016/j. cardiores.2006.01.017. [45] J. Feng, E. Lucchinetti, P. Ahuja, T. Pasch, J.C. Perriard, M. Zaugg, Isoflurane postconditioning prevents opening of the mitochondrial permeability transition pore through inhibition of glycogen synthase kinase 3beta, Anesthesiology 103 (2005) 987–995. [46] S.M. Davidson, D. Hausenloy, M.R. Duchen, D.M. Yellon, Signalling via the reperfusion injury signalling kinase (RISK) pathway links closure of the mitochondrial permeability transition pore to cardioprotection, Int. J. Biochem. Cell Biol. 38 (2006) 414–419, http://dx.doi.org/10.1016/j.biocel.2005.09.017. [47] R.M. Engelman, N. Maulik, D.K. Das, Role of Akt signaling in mitochondrial survival pathway triggered by hypoxic preconditioning, Circulation 109 (2004) 3042–3049, http://dx.doi.org/10.1161/01.CIR.0000130674.29030.90. [48] R. Schulz, G. Heusch, Connexin 43 and ischemic preconditioning, Cardiovasc. Res. 62 (2004) 335–344, http://dx.doi.org/10.1016/j.cardiores.2003.12.017. [49] C.A. Dunn, V. Su, A.F. Lau, P.D. Lampe, Activation of Akt, not connexin 43 protein ubiquitination, regulates gap junction stability, J. Biol. Chem. 287 (2012) 2600–2607, http://dx.doi.org/10.1074/jbc.M111.276261. [50] J.L. Solan, P.D. Lampe, Specific Cx43 phosphorylation events regulate gap junction turnover in vivo, FEBS Lett. 588 (2014) 1423–1429, http://dx.doi.org/10.1016/j. febslet.2014.01.049. [51] G.R. Peck, S. Ye, V. Pham, R.N. Fernando, S.L. Macaulay, S.Y. Chai, A.L. Albiston, Interaction of the Akt substrate, AS160, with the glucose transporter 4 vesicle marker protein, insulin-regulated aminopeptidase, Mol. Endocrinol. 20 (2006) 2576–2583, http://dx.doi.org/10.1210/me.2005-0476. [52] M.A. Retamal, K.A. Schalper, K.F. Shoji, M.V. Bennett, J.C. Sáez, Opening of connexin 43 hemichannels is increased by lowering intracellular redox potential, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 8322–8327, http://dx.doi.org/10.1073/pnas. 0702456104. [53] J.H. Lin, N. Lou, N. Kang, T. Takano, F. Hu, X. Han, Q. Xu, D. Lovatt, A. Torres, K. Willecke, J. Yang, J. Kang, M. Nedergaard, A central role of connexin 43 in hypoxic preconditioning, J. Neurosci. 28 (2008) 681–695, http://dx.doi.org/10.1523/ JNEUROSCI.3827-07.2008.