|Home | About | Journals | Submit | Contact Us | Français|
In mammalian brain, neurons and astrocytes are reported to express various chloride and anion channels, but the evidence for functional expression of Ca2+-activated anion channel (CAAC) and its molecular identity have been lacking. Here we report electrophysiological evidence for the CAAC expression and its molecular identity by mouse Bestrophin 1 (mBest1) in astrocytes of the mouse brain. Using Ca2+-imaging and perforated-patch clamp analysis we demonstrate that astrocytes displayed an inward current at holding potential of -70 mV that was dependent on an increase in intracellular Ca2+ following upon Gαq-coupled receptor activation. This current was mediated mostly by anions and was sensitive to well-known anion channel blockers such as niflumic acid, NPPB, and flufenamic acid. To find the molecular identity of the anion channel responsible for the CAAC current, we analyzed the expression of candidate genes, and found that the mRNA for mouse Bestrophin 1 (mBest1) is predominantly expressed in acutely dissociated or cultured astrocytes. Whole-cell patch clamp analysis using HEK293T cells heterologously expressing full-length mBest1 showed a Ca2+-dependent current mediated by mBest1, with a complete impairment of the current by a putative pore mutation, W93C. Furthermore, mBest1-mediated CAAC from cultured astrocytes was significantly reduced by expression of mBest1-specific shRNA, suggesting that the CAAC is mediated by a channel encoded by mBest1. Finally, hippocampal CA1 astrocytes in hippocampal slice also showed mBest1-mediated CAAC as it was inhibited by mBest1-specific shRNA. Collectively, these data provide molecular evidence that the mBest1 channel is responsible for CAAC function in astrocytes.
In the brain, anions including Cl−, bicarbonate, and other organic anions regulate various cellular events. A combination of transporters and ion channels are highly expressed in both neurons and glia to move these anions across membranes (Scott et al., 1995; Lambert and Oberwinkler, 2005; Kimelberg et al., 2006). Anion channels are classified as several families including cystic fibrosis transport regulator (CFTR), Ca2+-activated anion channel (CAAC), volume-regulated anion channel (VRAC), and ClC channel, according to how they are activated (Kimelberg et al., 2006). The function of CAACs has been studied extensively because of their important role in excitable tissues, which require an increase in intracellular Ca2+ concentration ([Ca2+]i) for channel activity (Hartzell et al., 2005; Kunzelmann et al., 2007). CAAC is directly or indirectly activated by an increase in [Ca2+]i and produces anion efflux or influx in mammalian cells and Xenopus laevis oocytes. This anion flux results in membrane depolarization or hyperpolarization, depending on the equilibrium potential for permeant anions (Hartzell et al., 2005). Whereas CAACs are previously known to be involved in the regulation of olfaction, taste, phototransduction, and excitability in the nervous system, the molecular identity and functional role of CAAC in the brain have not been well established.
At least seven different Cl− or anion-mediated currents have been functionally characterized in astrocytes (Walz, 2002; Kimelberg et al., 2006), but direct evidence for the expression of Ca2+-activated Cl− current in astrocytes has been lacking. Astrocytes are actively involved in communication between neurons, and this role is mainly mediated by an increase in [Ca2+]i secondary to activation of G protein-coupled receptors (GPCR) activated by neurotransmitter release from synaptic activity (Volterra and Meldolesi, 2005; Halassa et al., 2007). Thus, it is expected that astrocytic CAAC has an important role in astrocyte-neuron intercellular communication provided that the channel is functionally expressed in astrocytic membrane.
In this study, we investigated the functional expression of CAAC in astrocytes and identified a candidate gene encoding astrocytic CAAC current. Recently, bestrophin-1 (Best1), which was initially cloned as the gene linked to Best’s vitelliform macular dystrophy, a juvenile form of macular degeneration, has been suggested to encode a functional CAAC in nonneuronal tissue and peripheral neurons with an anion selective pore and single channel activity (Marmorstein et al., 2000; Sun et al., 2002; Eggermont, 2004; Qu et al., 2004; Chien et al., 2006; Pifferi et al., 2006; Kunzelmann et al., 2007; Hartzell et al., 2008). However, the expression profile and the function of bestrophin in the central nervous system (CNS) have not been explored. To address this issue, we examined both the expression pattern of bestrophin family members, and the possibility that these gene products might function as astrocytic CAAC. Our results demonstrate for the first time that astrocytes both in primary cell culture and in situ express functional CAAC encoded by Best1.
Cell cultures of HEK293T cells and mouse astrocytes were performed as previously described. For cultured astrocytes, mouse brain tissues containing cortex and hippocampus were isolated and dissociated mechanically. For Ca2+ imaging, cultured astrocytes were incubated with 5 μM Fura2-AM in 1 μM Pluronic acid (Molecular Probes, Eugene, OR, USA) for 30 min at room temperature, and subsequently transferred to a microscope stage for imaging using external solution contained 150 mM NaCl, 10 mM HEPES, 3 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 5.5 mM glucose, at pH 7.3. Intensity images of 510 nm wavelength were taken at 340 nm and 380 nm excitation wavelengths, and the two resulting images were taken for ratio calculations. Imaging Workbench software (INDEC BioSystems, Santa Clara, CA, USA) was used for acquisition of intensity images and conversion to ratios. The extracellular recording solution for simultaneous Ca2+ imaging and perforated patch clamp recording was comprised of (in mM) 150 NaCl, 10 HEPES, 3 KCl, 2 CaCl2, 2 MgCl2, and 5.5 glucose at pH 7.3. For perforated patch clamp recordings, the internal solution contained 25 μg/ml gramicidin D and (in mM) 75 Cs2SO4, 10 NaCl, 0.1 CaCl2, and 10 HEPES, at pH 7.1. Pipette resistances ranged from 5 to 8 MΩ. It took 20 to 30 min to achieve acceptable perforation, with final series resistances ranging from 15 to 40 MΩ. In whole-cell patch clamp, patch pipettes which have 3–6 MΩ of resistance are filled with the standard intracellular solution (see below). Current voltage curves were established by applying 100-, 200-, or 1000 ms duration voltage ramps from −100 to +100 mV. Data were acquired by an Axopatch 200A amplifier controlled by Clampex 9.2 via Digidata 1322A data acquisition system (Axon Instruments, Union City, CA, USA). Experiments were conducted at room temperature (20 ~ 24 °C). The pipette solution for whole-cell patch clamp in astrocytes was comprised of (in mM) 135 CsCl, 5 MgCl2, 5 EGTA, 10 HEPES, and 10 glucose at pH 7.3, adjusted with CsOH. To activate CAAC or mBest1 directly, we applied high Ca2+-containing intracellular patch pipette solution to cultured astrocytes or mBest1-expressing HEK293T cells, which are comprised of (in mM) 146 CsCl, 5 (Ca2+)-EGTA-NMDG, 2 MgCl2, 8 HEPES, and 10 Sucrose at pH 7.3, adjusted with NMDG. For control experiments, we used Ca2+-free intracellular solution comprised of (in mM) 146 CsCl, 5 EGTA-NMDG, 2 MgCl2, 8 HEPES, and 10 Sucrose at pH 7.3, adjusted with NMDG. The concentration of free [Ca2+]i in the solution was determined as described. The extracellular solution was comprised of (in mM) 150 NaCl, 10 HEPES, 3 KCl, 2 CaCl2, 2 MgCl2, and 5.5 glucose at pH 7.3 with NaOH (~320 mOsm). In some experiments, the bath and pipette solutions for mBest1-expressing HEK293T cells to activate Cl− current were prepared as described (O’Driscoll et al., 2009). The extracellular solution contained (in mM) 126 NaCl, 10 HEPES, 20 glucose, 1.8 CaCl2, 1.2 MgCl2 and 10 TEA-Cl, and pH was set to 7.35, adjusted with NaOH. The pipette solution contained (in mM) 20 TEA-Cl, 106 CsCl, 8.7 CaCl2, 5 HEPES, 10 BAPTA, 3 MgATP, 0.2 GTP-Li and 0.5 MgCl2, at pH 7.2 adjusted with CsOH. Relative permeability of the channels was determined by measuring the shift in Erev upon changing the bath solution from one containing 161 mM Cl− to another with 150 mM X and 11 mM Cl−, where X is the substituted anion. We used an agar bridge to prevent any nonspecific holding potential shift due to a replacement of Cl− with different anions. The permeability ratio was estimated using the modified Goldman-Hodgkin-Katz equation as previously reported (Qu et al., 2004).
In slice patch experiments, transverse brain slices (300~400 μm) containing hippocampus (from at least three 6~8 week old B6 mice) were prepared as previously described (Lee et al., 2007). SR-101 (Sigma, St. Louis, MO, U.S.A; 1 μM) dye was loaded into acute brain slice as previously described (Kafitz et al., 2008). Prepared slices were left to recover for at least 1 h before recording in oxygenated (95 % O2 and 5 % CO2) artificial cerebrospinal fluid (ACSF; in mM, 130 NaCl, 24 NaHCO3, 3.5 KCl, 1.25 NaH2PO4, 1 CaCl2, 3 MgCl2 and 10 glucose, pH 7.4.). Since SR-101 loaded cells from hippocampal CA1 stratum radiatum region showed typical properties of astrocytes such as passive current response and absence of depolarization-evoked action potential responding to voltage step command (from −100 to +100 mV; Fig. 7B), we used this dye for identifying hippocampal astrocytes. The standard ACSF recording solution was composed of (mM): 130 NaCl, 24 NaHCO3, 3.5 KCl, 1.25 NaH2PO4, 1.5 CaCl2, 1.5 MgCl2 and 10 glucose saturated with 95 % O2–5 % CO2, at pH 7.4. For hypertonic ACSF, sucrose was added to increase the osmolarity of standard recording ACSF up to ~15 % (Takano et al., 2005). Pipette solution contained (in mM) 60 KCl, 73 K-gluconate, 1 MgSO4, and 10 HEPES, at pH 7.3 adjusted with KOH. For Cs-containing pipette solution, an equimolar concentration of potassium was replaced with cesium (CsCl and Cs-gluconate). Experiments with a holding current more than −100 pA were rejected.
Total RNA was prepared from whole brain of adult mice (C57BL/6, 3~4 week) or cultured astrocyte from P0~P3 postnatal mice using Trizol reagent (Invitrogen, Carlsbad, CA, USA). cDNA was synthesized using Super Script III reverse transcriptase (Invitrogen). The RT-PCR primers used to check expression of mBest1, 2, and 3 from brain or cultured astrocyte cDNA were as followings: mBest1-F : 5 ′-AGGACGATGATGATTTTGAG-3′, mBest1-R: 5′-CTTTCTGGTTTTTCTGGTTG-3′; mBest2-F : 5 ′-TCGTCTACACCCAGGTAGTC-3′, mBest2-R : 5 ′-GAAAGTTGGTCTCAAAGTCG-3′; mBest3-F: 5′-AAAGGCTACGTAGGACATGA-3′, mBest3-R: 5′-GAAAGGACGGTATGCAGTAG-3′. To test the presence of other CAAC candidate in mouse brain or astrocyte, following primer sets were used: mCLCA1, 2, 4-F: 5′-TTCAAGATCCAAAAGGAAAA-3′, mCLCA1, 2, 4-R: 5′-GCTCAGTCTGGTTTTGTTTC-3′, mCLCA5-F: 5′-TAAGATTCCAGGGACAGCTA-3′, mCLCA5-R : 5 ′-AAAGGAGGAAAAATACCTGG-3′, mTtyh1-F : 5 ′-AGACACCTATGTGCTGAACC-3′, mTtyh1-R: 5′-AGAAAAGAGCATCAGGAACA-3 ′, mTtyh2-F : 5 ′-CCAGCTTCTGCTAAACAACT-3 ′, mTtyh2-R : 5 ′-AATCTCTGTCCCTGTTGATG-3′, mTtyh3-F: 5′-CAGTACTGAGTGGGGACATT-3′, mTtyh3-R: 5′-CTGTGACAAAGGAGAAGAGG-3′. To perform single cell RT-PCR, a single astrocyte and neuron was acutely and mechanically dissociated from cortex of adult mouse brain. The cDNA of single, dissociated cell was amplified using Sensiscript RT kit (Qiagen, Valencia, CA, USA). Neuron-specific enolase (NSE, 300 bp) and glial fibrillary acidie protein (GFAP, 360 bp) were used to identify the harvested cell type. PCR amplification was performed using the following primers: mBest1 forward outer primer: 5′-AGGACGATGATGATTTTGAG, mBest1 forward inner primer: 5′-ACCTTCAACATCAGCCTAAA-3′, mBest1 reverse common primer: 5′-CTTTCTGGTTTTTCTGGTTG-3 ′, NSE forward common primer: 5′-GCTGCCTCTGAGTTTTACCG-3′, NSE reverse outer primer: 5′-GAAGGGGATCACAGCACACT-3′, NSE reverse inner primer: 5′-CTGATTG ACCTTGAGCAGCA-3′, GFAP forward outer primer: 5′-GAGGCAGAAGCTCCAAGATG-3′, GFAP forward inner primer: 5′-AGAACAACCTGGCTGCGTAT-3′, GFAP reverse common primer: 5′-CGGCGATAGTCGTTAGCTTC-3′. The first PCR amplification was performed as described below. Samples were heated to 94 °C for 5 min. Each cycles consisted of denaturation at 94 °C for 30 sec, annealing at 50 °C for 30 sec, and elongationat 72 °C for 30 sec. Forty-two cycles were performed with a programmable thermocycler (Eppendorf, Westbury, NY, USA). The second PCR condition consisted of denaturation at94 °C for 30 sec, annealing at 55 °C for 30 sec, elongationat 72 °C for 30 sec for forty-two cycles, and subsequently cooled to 4 °C until analysis.
For the cloning of full-length mouse bestrophin 1 (mBest1) cDNA, total RNA was purified from cultured astrocytes or testis from adult male mice (C57BL/6), and cDNA was synthesized. Using 21 base primer pair spanning the open-reading frame based on NCBI database cDNA [GenBank accession numbers NM_011913.2], PCR was performed to acquire full-length ORF of mBest1. Resulting PCR products were cloned into a pGEM-T easy vector (Promega, Madison, WI, USA) and sequenced to be verified. In order to express mBest1 in mammalian cell, mBest1 full-length fragment from pGEM-T easy plasmid (6.65 kb) was subcloned into pIRES2-dsRED (Invitrogen) by using XbaI and XmaI sites. The pIRES2-mBest1-dsRED plasmid was transfected into HEK293T cells using Effectene transfection reagent (Qiagen). ~12 hours after transfection, cells were replated onto glass coverslips for electrophysiological recording. Transfected cells were used for patch clamp experiments within 12 hours. mBest1 pore mutant (mBest1 W93C) and shRNA-insensitive mutant constructs (mBest1 shRNA insensitive mutant) were generated by using PCR-based site-directed mutagenesis kit (Stratagene, Cedar Creek, TX, USA). PCR primers for mutagenesis of mBest1 W93C were followings: mBest1 g279c-sense, 5′-GGTGAGCCGCTGCTGGAGCCAGTAC-3′; mBest1 g279c-antisense. 5′-GTACTGGCTCCAGCAGCGGCTCACC-3′. PCR primers for mutagenesis of mBest1 shRNA-insensitive mutant constructs were followings: c567t_c570t_t571c_a576c-sense, 5′-GTGCCCTGGGTGTGGTTTGCTAATCTGTCCATGAAGGCCTATCTTGGAGG-3′; c567t_c570t_t571c_a576c-antisense, 5′-CCTCCAAGATAGGCCTTCATGGACAGATTAGCAAACCACACCCAGGGCAC-3′.
The mBest1 nucleotides from 774 to 793 (5′-tttgccaacttgtcaatgaa-3′) was selected for target region of mBest1-shRNA. For lentivirus-based shRNA expression, mBest1 shRNA was synthethized as followings: 5′-tttgccaacttgtcaatgaattcaagagatcattgacaagttggcaattttttc-3′ a n d 5 ′-tcgagaaaaaatcgcatagcgtatgccgtttctcttgaaaacggcatacgctatgcgaa-3′. The annealed double stranded oligo was inserted into HpaI-XhoI restriction enzyme sites of pSicoR lentiviral vector (provided by Dr. T. Jacks; (Ventura et al., 2004)) and verified by sequencing. Scrambled shRNA-containing pSicoR construct was used as control. shRNA-containing pSicoR was electroporated into cultured astrocytes (MicroPorator; Digital Bio, Seoul, South Korea). For shRNA expression into hippocampal region, lentivirus (produced by Macrogen Inc. Seoul, South Korea) was delivered into hippocampal CA1 region by stereotaxic surgery method (Cetin et al., 2006).
To make specific riboprobe for mRNA of mBest1, we cloned partial cDNA fragments of mBest1. Primers were as follows: forward : 5 ′-ACCTTCAACATCAGCCTAAA-3; reverse: 5′-CTTTCTGGTTTTTCTGGTTG-3′. The plasmid was linearized and used for in vitro transcription (Roche Dignostics, Indianapolis, IN, USA) to label RNA probes with digoxigenin-UTP. In situ hybridization was performed as previously described with some modifications. Frozen brains of adult mice were sectioned at 20 μm thickness on a cryostat. The sections were then fixed in 4 % paraformaldehyde, washed with PBS, and acetylated for 10 min. The sections were incubated with the hybridization buffer (50 % formamide, 4X SSC, 0.1 % CHAPS, 5 mM EDTA, 0.1 % Tween-20, 1.25 × Denhartdt’s, 125 μg/ml yeast tRNA, 50 μg/ml Heparin) and digoxigenin-labeled probes (200 ng) for 18 h at 60 °C. Non-specific hybridization was removed by washing in 2X SSC for 10 min and in 0.1X SSC at 50 °C for 15 min. For immunological detection of digoxigenin-labeled hybrids, the sections were incubated with anti-digoxigenin antibody conjugated with alkaline phosphatase (Roche Diagnostics) for 1 h, and the colour reaction was carried out with NBT/BCIP (4-nitroblue tetrazolium chloride/bromo-4-chloro-3-indolyl phosphate; Sigma). Sections were dehydrated and mounted with Vectamount (Vector Laboratories, Burlingame, CA, USA).
Rabbit polyclonal mBest1 IgG was produced using antigen previously described (Barro Soria et al., 2006)(Ab Frontier, Seoul, South Korea). B6 adult mice were deeply anesthetized by 2 % avertin and perfused with room-temperature 0.1 M PBS followed by ice-cold 4 % paraformaldehyde. Brains were post-fixed in 4 % paraformaldehyde at 4 °C for 24 hr and 30 % sucrose 4 °C for 48 hr. Brains were then cut in coronal section of 30 μm on a cryosection. Sections were blocked in 0.1 M PBS containing 0.3 % triton X-100 (Sigma) and 2 % serum from species of the secondary antibody for 1 hr. Primary antibody was then applied at appropriate dilution (mBest1 IgG; 1:200), GFAP (Chemicon, Temecula, CA, USA; 1:200), NeuN (Chemicon; 1:200) and incubated overnight at 4 °C. After overnight incubation, the sections were washed three times in PBS and then incubated in secondary antibody (Alexa 488 goat anti rabbit IgG, Invitrogen; 1:400), TRITC conjugated goat anti chicken IgG (Jackson laboratory, Bar Harbor, ME, USA; 1:100), Alexa 555 goat anti mouse IgG (Invitrogen) for 2 hr. After three rinses in PBS, the sections were mounted on slide glass. Images were acquired on an Olympus Fluoview FV1000 confocal microscope and analyzed using Image J software.
Previous studies have provided ample experimental evidence for functional expression of anion channels in astrocytes (Walz, 2002), but there is only a limited number of reports describing functional expression or properties of CAAC in astrocytes of the CNS (Crepel et al., 1998; Walz, 2002; Parkerson and Sontheimer, 2004; Takano et al., 2005; Abdullaev et al., 2006; Ramos-Mandujano et al., 2007). Therefore, we examined whether astrocytes express functional CAAC. We simultaneously recorded whole cell currents and intracellular Ca2+ responses in cultured astrocytes from mouse brain under the gramicidin-D perforated patch configuration (Fig. 1A), which allowed us to minimize dialysis of intracellular anions. An astrocyte that was well isolated from other cells in culture was selected for recording to avoid gap-junction coupling that might compromise our ability to control the voltage. Since astrocytic G protein (Gq or G12/13)-coupled receptors such as protease activated receptor-1 (PAR-1) are known to induce a transient increase in the [Ca2+]i (Porter and McCarthy, 1997; Lee et al., 2007), we applied the PAR-1 selective peptide agonist TFLLR (TFLLR-NH2) (Lee et al., 2007) to test if a PAR1-stimulated increase in [Ca2+]i in astrocyte caused any changes in membrane conductance. Upon application of 30 μM TFLLR (3-fold EC50), we observed a large inward current that closely followed the time course of the Ca2+ response (154 ± 16 pA, mean ± s.e.m, n= 26; Fig. 1A). This result raised the possibility of functional expression of CAAC in astrocytes. When other types of Gq coupled receptors such as P2Y receptor, bradykinin receptor, lysophosphatidic acid (LPA) receptor, and prostaglandin E2 (PGE2) receptor were activated by corresponding selective agonists, similar concomitant increase of [Ca2+]i and inward current were observed (Fig. 1B), indicating that this current induction is a general mechanism shared by a host of astrocytic Gq coupled receptors.
CAAC is known to be activated by submicromolar physiological range of cytosolic Ca2+ and shows anion selectivity and outwardly rectifying current, which are blocked by anion channel blockers (Eggermont, 2004). To test whether the inward current we detected in astrocytes is mediated by CAAC, we firstly examined the Ca2+ dependency of TFLLR-induced current. TFLLR-induced inward current was intact in Ca2+ free bath (Fig. 1C), but eliminated by BAPTA-AM treatment (Fig. 1E). Impairment of the Ca2+ release from internal stores by prior application of either thapsigargin (Fig. 1D) or phospholipase C inhibitor, U73122 (Fig. 1F), also impaired both the TFLLR-induced [Ca2+]i increase and the inward current, indicating a dependency to Ca2+ release from intracellular stores.
We next tested whether this inward current is mediated by an anion channel. Perforated patch clamp measurement showed that [Ca2+]i-activated inward current was blocked by treatment of a series of anion channel blockers including 100 μM niflumic acid, 100 μM flufenamic acid, and 100 μM NPPB (Qu and Hartzell, 2001; Hartzell et al., 2005) (Fig. 2A and C). Niflumic acid-mediated block of the TFLLR-induced current was voltage-independent (Fig. 2D), with an IC50 of 9.8 μM (Fig. 2B), virtually identical to that reported for CAACs expressed in Xenopus laevis oocytes (IC50= 10.1 μM; (Qu and Hartzell, 2001)). In addition, we were able to directly induce an increase in conductance in astrocytes by a membrane rupture for whole-cell mode with internal solutions containing high free [Ca2+]i (~4.5 μM) (Qu et al., 2004). The current induced by this manipulation displayed a non-desensitizing time course, and was readily blocked by treatment of 100 μM niflumic acid (Fig. 2E and F) and 100 μM NPPB (n= 3; data not shown). On the contrary, carbenoxolone (100 μM ; n= 3) or chlorotoxin (1 μM; n= 2) treatment of astrocytes did not block this current, suggesting that gap-junction hemi-channels or chlorotoxin-sensitive chloride channels are not involved in this Ca2+-activated current (data not shown)(Dalton et al., 2003; Ye et al., 2003). These results demonstrated that astrocytes display Ca2+-activated current that is readily blocked by anion channel blockers.
We subsequently tested whether the astrocytic Ca2+-activated inward current recorded under gramicidin-D perforated patch configuration was carried by anions. We determined the current-voltage (I–V) relationship with a voltage ramp protocol for the TFLLR-induced current in astrocytes in the presence of extracellular Cl− or with extracellular Cl− replaced by the larger and presumably less permeable isethionate (Fig. 3A). The reversal potential in perforated patch recordings was significantly shifted to the right, from −13.2 ± 1.9 to +5.4 ± 1.5 mV (n= 8, and 5, respectively; p<0.05, unpaired t-test), suggesting that this current was carried by anions. In a separate experiment, the reversal potential of the TFLLR-induced current under whole-cell configuration with the intracellular solution containing 7 mM Cl− was −71 ± 1.5 mV (n= 4; Fig. 3B and C), which agreed well with the calculated reversal of −75 mV, according to the Nernst equation.
To determine the relative permeability to different anions, a series of ion substitution experiments was performed under whole cell patch clamp mode with internal solution containing a high concentration of Ca2+. The Cl− of external bath solution was substituted with I−, Br−, or F−, and the I–V relationship of astrocytic CAAC was determined. We found that astrocytic CAAC showed outwardly rectifying I–V relationship and displayed the relative permeability order of I− > Br− > Cl− > F− (Fig. 3D and E), which was identical to the previously known properties of other CAACs reported in Xenopus laevis oocytes and mammalian cells (Cliff and Frizzell, 1990; Clapp et al., 1996; Kuruma and Hartzell, 1999). The astrocytic CAAC also showed a significant permeability to large anions such as glutamate and isethionate (Fig. 3Dc and E) as previously reported for aspartate in Xenopus laevis oocytes (Qu et al., 2003). However, ion substitution of Na+ ion by an exchange of NaCl with NMDG-Cl did not change the reversal potential (n= 5; p= 0.66, unpaired t-test), indicating that Na+ ion is the minimal charge carrier of CAAC current in astrocyte. Taken together, these results are consistent with the idea that astrocytes express functional CAACs, which can be activated by an increase in cytosolic Ca2+ from internal Ca2+ store upon GPCR activation.
Molecular identification of anion channels has been hampered by the lack of specific blockers. CAAC is the one of very few channels for which a cDNA has not yet been unequivocally identified (d’Anglemont de Tassigny et al., 2003; Eggermont, 2004). To identify the genes encoding CAAC channels in astrocytes, we performed reverse transcriptase polymerase chain reaction (RT-PCR) with primer sets for various CAAC candidate genes such as Cl− channel-Calcium Activated (CLCA), Drosophila tweety homolog (Ttyh), and bestrophin (Best) family genes (Eggermont, 2004; Suzuki and Mizuno, 2004). Our RT-PCR analysis demonstrated that mouse bestrophin 1 and 3 (mBest1 and 3), and mouse homologues for Ttyh family genes (mTtyh1, 2, and 3) were significantly expressed, while CLCA’s were absent in the brain and cultured astrocytes (Fig. 4A). Despite the significant expression in astrocytes, mouse Ttyh family genes were not considered as the astrocytic CAAC candidate due to the non-overlapping properties such as insensitivity to niflumic acid and the lack of outward rectified current-voltage relationship (Suzuki and Mizuno, 2004). Although the expression levels of mBest1 and 3 in cultured astrocytes were slightly lower than other tissues such as testis and heart (Supplementary Fig. 1) (Marmorstein et al., 2000; Pifferi et al., 2006), it was evident that mBest1 was significantly expressed with higher levels than mBest3 in astrocytes (Fig. 4A, Supplementary Fig. 1)). Therefore, mBest1 appeared to be an appropriate candidate gene for astrocytic CAAC.
We next analyzed the expression pattern of mBest1 within the brain by cell type and by brain region. By performing single-cell RT-PCR using a mBest1-specific primer set and cDNA of individual acutely-dissociated astrocytes or neurons from adult mouse cortex, we identified the expression of mBest1 in both GFAP (glial fibrillary acidic protein) and NSE (neuron specific enolase)-expressing cell types (Fig. 4B), indicating that mBest1 is expressed both in astrocytes and neurons. In situ hybridization analysis also showed a widely distributed expression pattern of mBest1 mRNA with higher level in olfactory bulb, hippocampus and cerebellum and significant expression in both neuron and astrocyte-like cells (Fig. 4C). To confirm the expression of mBest1 protein in mouse hippocampus, immunohistochemical analysis using mBest1-specific antibody (Barro Soria et al., 2006) was performed. Consistent with single cell RT-PCR and in situ hybridization analysis, mBest1 protein showed significant expression in both NeuN- and GFAP-positive cells (Fig. 4D and E). Taken together, these results demonstrate for the first time the astrocytic and neuronal expression of mBest1 in the brain.
Recently it has been reported that heterologously expressed mBest1 displays CAAC current by whole cell patch clamp recordings using submicromolar Ca2+ containing patch pipette internal solution. Consistent with this report, we observed that heterologously expressed mBest1 in HEK293T cells displayed Ca2+-induced outwardly rectifying current, which was largely inhibited by pretreatment with niflumic acid or by using a pipette internal solution lacking Ca2+ (Fig. 5A and B; current amplitude at Vh= 70 mV; mBest1 expression w/high Ca2+ 295 ± 71 pA, n= 7; mBest1 expression w/0 Ca2+ 41 ± 17 pA, n= 5; NFA pretreated mBest1 w/high Ca2+ 19 ± 11 pA, n= 9; **p<0.01 versus mBest1 expression w/high Ca2+ group; One way ANOVA with Dunnett’s multiple comparison test). To verify whether mBest1 forms an ion channel, we constructed an anion selective pore mutant of mBest1 by mutating tryptophan-93 to cysteine (W93C), which has been suggested to reside within the anion selective pore region in mouse bestrophin-2 (mBest2) channel. Since this presumed pore-forming residue is conserved between mBest1 and mBest2 (Qu and Hartzell, 2004; Qu et al., 2006b), we hypothesized that W93C mutation of mBest1 should impair the CAAC current when heterologously expressed in HEK293T cells. As expected, we found that mBest1-W93C expressing HEK293T cells did not show any significant CAAC current using a high Ca2+-containing pipette internal solution (Fig. 5A and B; current amplitude at Vh= −70 mV; mBest1-W93C expression w/high Ca2+ 45 ± 11 pA, n= 9; **p<0.01 versus mBest1 expression w/high Ca2+ group; One way ANOVA with Dunnett’s multiple comparison test). These data suggest that heterologously expressed mBest1 encodes a Ca2+-activated anion channel with a pore similar to mBest2.
To determine the molecular identity of astrocytic CAAC as mBest1, we designed an mBest1-specific, short hairpin RNA (mBest1-shRNA) to selectively knock-down the expression of mBest1 transcript. Our semi-quantitative RT-PCR analysis showed that the expression of mBest1-shRNA resulted in effective and selective knock-down of endogenous mBest1 mRNA in cultured astrocytes (Fig. 6A), suggesting that mBest1-shRNA could effectively knock-down the expression of CAAC current. We subsequently recorded astrocytic CAAC current in scrambled-shRNA or mBest1-shRNA expressing cultured astrocytes using perforated patch clamp mode to test whether mBest1 encodes astrocytic CAAC current. We found that CAAC current in cultured astrocyte was significantly suppressed by mBest1-shRNA expression (Fig. 6B; current amplitude at Vh= −70 mV; scrambled-shRNA expressing astrocytes 176 ± 46 pA, n= 4; mBest1-shRNA expressing astrocytes 15 ± 4 pA, n= 5; ***p<0.001 versus scrambled-shRNA group; unpaired t-test) without disrupting Ca2+ responses by PAR-1 activation (relative change % of Fura-2 340/380 ratio; scrambled-shRNA 100 ± 17 %; mBest1-shRNA 131 ± 16 %; p= 0.92; paired t-test). This mBest1-dependent CAAC current was also confirmed by performing whole cell patch clamp analysis using Ca2+ containing internal solution in cultured astrocytes (Fig. 6C; holding current at Vh= −70 mV; scrambled-shRNA expressing astrocytes 174 ± 7 pA, n= 10; mBest1-shRNA expressing astrocytes 50 ± 8 pA, n= 11; ***p<0.001 versus sc-shRNA group; unpaired t-test). These results indicate that mBest1 expression in astrocytes is necessary for functional CAAC, consistent with our working hypothesis that mBest1 encodes CAAC channels.
Since cultured astrocytes were shown to have different gene expression profile than astrocytes in tissue (Cahoy et al., 2008), it is possible that our observation of mBest1-mediated CAAC in cultured astrocytes might not be representative of the in vivo condition. To test whether astrocytes in situ display the CAAC current, we first characterized Ca2+-induced current from hippocampal CA1 astrocytes under whole cell patch clamp mode (Fig. 7A). SR-101, which has been used for labeling astrocytes in brain slice, was used for identifying astrocytes. To induce a Ca2+ increase in hippocampal CA1 astrocytes, we applied a brief pressure pulse to a TFLLR-containing micropipette (~500 μM; 1 sec) that was placed near a single SR-101-loaded astrocyte. PAR-1 activation by the pulse of TFLLR induced an inward current at a holding potential of −70 mV (59 ± 17 pA; n= 6), which was significantly blocked by pretreatment of anion channel blocker (NPPB, 100 μM; 14 ± 5 pA; n= 6) or in the presence of Ca2+ chelator in the patch pipette solution (BAPTA, 10 mM; 9 ± 2 pA; n= 6; Fig. 7D). It is possible that Ca2+-induced astrocytic volume increase and subsequent activation of a volume-regulated anion channel (VRAC) might be responsible for this PAR-1-induced Cl− current (Takano et al., 2005). However, the PAR-1-induced current was not affected by pretreatment of hypertonic ACSF (15 % higher osmolarity than isotonic ACSF for at least 20 min; 57 ± 17 pA; n= 3), which argues against the possibility of PAR-1-induced VRAC activation (Fig. 7D). Moreover, PAR-1 induced current was not significantly changed by using Cs-containing pipette solution, suggesting that the PAR-1-induced current is minimally influenced by a Ca2+-activated K conductance in our experimental condition (Cecchi et al., 1987; Hu et al., 1989)(K-pipette: 59 ± 17 pA, n= 6; Cs-pipette: 40 ± 16 pA, n= 8; p= 0.12, unpaired t-test). These results demonstrate that PAR-1 activation leads to the generation of a CAAC current in hippocampal CA1 astrocytes in acutely prepared brain slices, the properties of which are similar to cultured astrocytes. Next, using lentiviral shRNA expression system, we expressed scrambled- or mBest1-shRNA in hippocampal CA1 astrocytes to determine whether mBest1 mediates the channel-dependent CAAC current. Our immunohistochemical analysis showed that lentiviral expression of mBest1 shRNA significantly reduces the expression level of mBest1 protein in hippocampal slice (Supplementary Figure 2). Using this gene silencing tool, we measured CAAC current in scrambled or mBest1-shRNA expressing astrocytes. Similar to the measurements in cultured astrocyte, PAR-1-induced CAAC current was significantly reduced in mBest1-shRNA expressing astrocytes compared to scrambled-shRNA expressing astrocytes (Fig. 7C and D; scrambled-shRNA: 37 ± 4 pA; n= 12, mBest1-shRNA: 11 ± 4 pA; n= 7). All together, these results demonstrated that mBest1 is the functional CAAC both in in vitro and in vivo astrocytes.
In this study we have demonstrated that in cultured astrocytes the activation of a GPCR such as PAR-1 can induce a Ca2+-dependent current. This current was functionally similar to the previously characterized CAACs with overlapping properties such as the rank order of anion permeability, outwardly rectifying current-voltage relationship, and the sensitivity to anion channel blockers (Fig. 1~~3).3). This current was also observed in hippocampal CA1 astrocytes in hippocampal slices, which was similarly dependent on Ca2+ (Fig. 7).
Due to the complex and overlapping pharmacological targets of various anion channel blockers such as niflumic acid and NPPB (Greenwood and Leblanc, 2007), it is possible that other channels might be involved in PAR-1-induced current from cultured or hippocampal CA1 astrocytes. Our results show that activation of the gap-junction hemichannel (Ye et al., 2003) and chlortoxin-sensitive chloride channel are unlikely to mediate CAAC because of the insensitivity of the current by the treatment of corresponding blockers (see Results). Other possible contributions by either non-selective cation channels or Ca2+-activated K+ conductances (Gebremedhin et al., 2003) seemed unlikely given the good agreement between the observed reversal potential and calculated ECl (Fig. 3C–D), as well as given the use of Cs+-or TEA-containing internal solution for whole-cell patch clamp recordings (Fig. 5 and and7).7). Collectively, despite the caveats of using poorly selective anion channel blockers for identifying CAAC conductance, our results provide strong evidence that hippocampal astrocytes express functional CAAC. Finally, the possibility exists that a Ca2+-induced volume increase and subsequent VRAC activation could contribute in part to the PAR-1-induced current we observed. However, we found that the PAR-1-induced current in hippocampal astrocytes was unchanged in hypertonic ACSF condition, indicating that the involvement of VRAC is minimal in PAR-1 induced anion conductance (Fig. 7D).
The results from RT-PCR, single-cell PCR, in situ hybridization, and gene silencing experiments all indicate that the CAAC current observed in mouse brain astrocytes is mediated by the mouse Bestrophin-1 channel (mBest1). Among the previously suggested molecular candidates for CAAC, including ClC-3, CLCA, and bestrophin channels, bestrophin channel appears to be the best candidate for CAAC because the Cl− currents from bestrophin-expressing cells showed many of the properties of native CAAC properties (Sun et al., 2002; Qu et al., 2004; Barro Soria et al., 2006), such as Ca2+ activation, sensitivity to niflumic acid, and outwardly-rectifying current-voltage relationship (Fig. 5A). ClC-3 gene was excluded from the list of candidates genes, because it has been reported that ClC3-deleted mice displayed the impaired CaMKII-activated but intact Ca2+-activated anion conductance (Hartzell et al., 2005). Our results showed a lack of expression of CLCA in astrocytes, and CLCA possesses properties that are inconsistent with astrocytic CAAC (Fig. 4A; (Eggermont, 2004)). Despite their strong expression, Ttyh1 and Ttyh2 are insensitive to Ca2+ (Suzuki and Mizuno, 2004). Ttyh3, one of the Ttyh family genes showing Ca2+-dependent generation of maxi-anion current when expressed heterologously in CHO cells, was also unlikely to mediate CAAC because Ttyh3 response displays a linear current-voltage relationship, and is sensitive to DIDS, but not to niflumic acid (Suzuki and Mizuno, 2004). Recently, the transmembrane protein 16 A and B (TMEM16A and B) gene was suggested as a bona fide CAAC (Caputo et al., 2008; Schroeder et al., 2008; Yang et al., 2008; Rock et al., 2009), but due to the low expression level in brain TMEM16A or B seems unlikely to encode astrocytic CAAC (Schroeder et al., 2008; Yang et al., 2008).
The possible contribution of mBest3 to the CAAC current in astrocyte appears to be minimal due to the following reasons. First, mBest3 transcript is expressed at much lower level than that of mBest1 in brain. Second, mBest3 which was previously called mBest4, was reported to be non-functional for generating anion conductance unless a portion of the C-terminus is truncated (Qu et al., 2006). Finally, CAAC current was almost completely impaired by mBest1 gene-silencing in astrocytes (Fig. 6 and and7).7). Nevertheless, we cannot exclude the possibility of a regulatory function for mBest3 in mBest1-mediated CAAC current. We also cannot exclude the existence of C-terminus-truncated form of mBest3 in astrocytes. Future study should be directed to elucidate the exact role of mBest3 in astrocytes and neurons, possibly using mBest3-specific gene silencing tool. Collectively, these results strongly support the idea that Best1 channel is a functional CAAC in astrocytes.
In a recent report, it has been proposed that bestrophin can function not only as a CAAC but also as a regulator of L-type Ca2+ channels (Rosenthal et al., 2006; Yu et al., 2008). In astrocytes, the source of Ca2+ increase is mainly from intracellular Ca2+ stores following GPCR activation, rather than from extracellular Ca2+ entry through voltage-gated Ca2+ channels (Fig. 1). Despite the previous reports about functional expression of L-type Ca2+ channel in astrocytes (Latour et al., 2003; D’Ascenzo et al., 2004), we found that PAR-1 activation-induced Ca2+ responses were intact during the treatment astrocytes with the selective L-type Ca2+ channel blocker, nimodipine (2 μM, data not shown; n= 3). In addition, mBest1-shRNA expression in astrocytes did not affect the TFLLR-induced Ca2+ responses (see Results). Furthermore, mBest1 channels lack a functional domain for interactions with L-type Ca2+ channels. Therefore, at least in mouse astrocytes, it is unlikely that mBest1 acts as a Ca2+ channel regulator.
The present study provides compelling evidence for functional expression of CAAC in astrocytes. Our results revealed that Best1 is functionally expressed in astrocytes and encodes the channel that gives rise to most of the CAAC in astrocytes. The physiological function of astrocytic CAAC and Best1 channel remains to be explored, but the results and molecular tools described here should provide an opportunity to address many interesting questions regarding the physiological role of anion channels in astrocytes.
We thank P. Lyuboslavsky and P. Le for their excellent technical assistance. This work was supported by the Star Postdoc fellowship (H.P), Ente Cassa di Risparmio Firenze(G.M.), PRIN 2007 (G.M.), NIH NS39419 (S.F.T.), NS43875 (C.J.L.), Korea research foundation KRF-2005-070-C00096 (C.J.L.), and KIST Core Competency Program (C.J.L.).