ClC-3 Is a Fundamental Molecular Component of Volume-sensitive Outwardly Rectifying Cl− Channels and Volume Regulation in HeLa Cells and Xenopus laevis Oocytes*

Volume-sensitive osmolyte and anion channels (VSOACs) are activated upon cell swelling in most vertebrate cells. Native VSOACs are believed to be a major pathway for regulatory volume decrease (RVD) through efflux of chloride and organic osmolytes. ClC-3 has been proposed to encode native VSOACs in Xenopus laevis oocytes and in some mammalian cells, including cardiac and vascular smooth muscle cells. The relationship between the ClC-3 chloride channel, the native volume-sensitive osmolyte and anion channel (VSOAC) currents, and cell volume regulation in HeLa cells andX. laevis oocytes was investigated using ClC-3 antisense. In situ hybridization in HeLa cells, semiquantitative and real-time PCR, and immunoblot studies in HeLa cells and X. laevis oocytes demonstrated the presence of ClC-3 mRNA and protein, respectively. Exposing both cell types to hypotonic solutions induced cell swelling and activated native VSOACs. Transient transfection of HeLa cells with ClC-3 antisense oligonucleotide or X. laevis oocytes injected with antisense cRNA abolished the native ClC-3 mRNA transcript and protein and significantly reduced the density of native VSOACs activated by hypotonically induced cell swelling. In addition, antisense against native ClC-3 significantly impaired the ability of HeLa cells and X. laevis oocytes to regulate their volume. These results suggest that ClC-3 is an important molecular component underlying VSOACs and the RVD process in HeLa cells and X. laevis oocytes.

A basic homeostatic function of all cells is to regulate their cell volume when exposed to either intracellular or extracellular anisosmotic conditions. Most vertebrate cells respond to hypoosmotically induced cell swelling by actively decreasing their volume, a process known as regulatory volume decrease (RVD) 1 (1). The RVD response occurs by efflux of K ϩ , Cl Ϫ , and organic osmolytes accompanied by osmotically obliged water loss. Approximately 70% of the osmolyte loss during the RVD response is accounted for by loss of KCl via separate conductive pathways (2,3). Cell volume regulatory mechanisms are fundamentally important in physiological processes such as cell differentiation, cell growth, apoptosis, and cellular metabolism (4).
Volume-sensitive osmolyte and anion channels (VSOACs) are activated upon cell swelling in most vertebrate cells. Blockade of VSOACs were found to suppress RVD in a variety of cell types (5). Macroscopic outwardly rectifying VSOAC currents are characterized by activation after cell volume increase (6), a SCN Ͼ I Ͼ Br Ͼ Cl Ͼ F Ͼ gluconate permeability sequence (6), time-dependent inactivation at positive potentials (7), inhibition by tamoxifen, 1,9-dideoxyforskolin, and stilbene derivatives (8), dependence on intracellular ATP (6), and a single channel conductance in the range of 20 -40 pS (9). These biophysical and pharmacological characteristics are considered to represent classic characteristics of VSOACs. However, variations in the pharmacological characteristics, voltage dependence, and signaling mechanisms responsible for channel activation have been described in different cell types (5,10). These findings have led to the current debate as to whether a group of channel proteins rather than just one protein may comprise the overall biophysical and pharmacological characteristics of VSOACs described in different cell types (11,12).
The protein or proteins conclusively responsible for the native VSOAC current have yet to be identified. Several proteins have been implicated as candidates for VSOAC, including Pglycoprotein (13), pICln (14), ClC-2 (15), and ClC-3 (16,17). P-glycoprotein is now thought to be a regulator of this channel (18,19) and pICln is also thought to possibly regulate VSOAC; however, its precise role is still under investigation (20,21). ClC-2, when expressed, has been shown to display biophysical and pharmacological characteristics that differ from native VSOACs and has also been shown not to contribute to RVD (15,22). ClC-3 was originally cloned from rat kidney by Kawasaki et al. (23,24), and, when expressed, currents showed basal activation and inhibition by phorbol esters and Ca 2ϩ (25,26). ClC-3 cloned from guinea pig ventricular myocytes was proposed to be the molecular candidate responsible for cardiac VSOACs (16).
When expressed in NIH3T3 cells ClC-3 gave rise to a basally active chloride conductance that was modulated by cell volume, strongly inhibited by protein kinase C, displayed outward rectification with a unitary slope conductance of 40 pS, an anion selectivity sequence of I Ͼ Cl Ͼ Asp, voltage-dependent inactivation at positive potentials, and similar pharmacological characteristics as native VSOACs (27). Furthermore, Wang et al. (17) showed in bovine epithelial cells that ClC-3 antisense treatment delayed the rate of activation of native VSOAC and reduced its amplitude by up to 60% in a dose-dependent manner. Also, a polyclonal anti-ClC-3 antibody (Ab) was shown to functionally inhibit native VSOACs in guinea pig cardiac cells, canine pulmonary arterial smooth muscle cells, and Xenopus laevis oocytes (28). Although the ClC-3 hypothesis has received additional experimental support from other laboratories, (29 -32) its role as a volume sensitive chloride channel has recently become controversial (33)(34)(35)(36).
The purpose of the present study was to investigate the role of endogenous ClC-3 in native VSOAC activation and the RVD mechanism intrinsic to HeLa cells and X. laevis oocytes. HeLa cells or X. laevis oocytes were transiently treated with an antisense oligonucleotide or injected with antisense cRNA against ClC-3, respectively, to abolish endogenous protein expression. Our results demonstrate that antisense treatment abolished both mRNA and protein expression of ClC-3 in HeLa cells and X. laevis oocytes, significantly reduced native VSOAC current density, and significantly diminished the ability of cells to undergo RVD.

EXPERIMENTAL PROCEDURES
All experiments on HeLa cells were performed at the Instituto de Ciencias Biomédicas, Facultad de Medicina Universidad de Chile, and all oocyte experiments were conducted at the University of Nevada, Reno, School of Medicine.
In Situ Hybridization Studies-733-bp in situ hybridization probes were obtained from HeLa cells cDNA, complementary to the human ClC-3 mRNA initiation codon region (37). An antisense probe was obtained using conventional procedures. HeLa cells were grown up to 70% confluence and fixed in 4% paraformaldehyde at 4°C for 30 min. Excess fixative was washed with cold phosphate-buffered saline containing 0.1% Triton X-100. Fixed HeLa cells were pre-hybridized with Denhardt's solution for 30 min and hybridized overnight with 10 ng/l ClC-3 antisense or sense RNA probes at 65°C. After washing in 2ϫ SSC/0.1% Triton X-100 and 0.2ϫ SSC/0.1% Triton X-100, cells were incubated overnight with a rabbit anti-digoxigenin (alkaline phosphatase-conjugated) Ab at 4°C. After excess Ab was washed away, cells were incubated with NTP/5-bromo-4-chloro-3-indolyl phosphate substrates, mounted with Permount (Fisher Scientific, NJ) and examined under a light microscope.
Acute Transfection of HeLa Cells-The antisense (ASO) or mismatched (MMO) oligonucleotides used for HeLa cell transfection were second generation chimeras high-performance liquid chromatographypurified (Oligos Etc. Inc., Wilsonville, OR). The ASO and MMO sequence was 5Ј-CGTCCCTCTTTAACTGGTT-3Ј and 5Ј-CTGCCTCCAT-TTGTCATTG-3Ј, respectively. The ASO sequence was complementary to the human ClC-3 mRNA initiation codon region (37), starting from the second base before the initiation codon (ATG corresponding to the third open reading frame). For electrophysiological and volume measurement experiments, the transfection efficiency was assessed with a fluorescein isothiocyanate (FITC)-or a rhodamine derivative (TAMRA)labeled oligonucleotides at bases 1 and 19, or 1 and 2, respectively. Confluent HeLa cell cultures were acutely transfected with 1 M ASOor MMO-ClC-3 in the presence of LipofectAMINE Plus reagent (Invitrogen, Paisley, UK). Cells were then grown for another 24, 48, or 72 h until electrophysiological recordings, cell volume measurements, and ClC-3 mRNA and protein content determinations were conducted.
ClC-3 Antisense and Sense cRNA Synthesis and Injection into Oocytes-Antisense cRNA was used to abolish ClC-3 mRNA transcription in X. laevis oocytes. X. laevis oocyte ClC-3 (xClC-3) cDNA, corresponding to a 211-bp fragment (15-225 bp, GenBank TM accession number Y09941) in either the sense or antisense direction, was subcloned into a TA cloning vector PCR 2.1 (Invitrogen). The sequence includes the start codon of xClC-3 mRNA and is not homologous to other known ClC channels. The constructs were linearized by digestion with the restriction enzyme BamHI and complementary RNA (cRNA) was synthesized by the bacteriophage RNA polymerase T7 using the mMessage mMachine in vitro transcription kit (Ambion). Injection pipettes with tips of ϳ1 m in diameter were baked at Ͼ150°C overnight to destroy RNases and were mounted in a Drummound Nanoject autoinjector (Drummound Scientific, Broomall, PA). Oocytes were injected with 9 ng of either sense or antisense cRNA in 50 nl of diethyl pyrocarbonate water.
RNA Extraction and Reverse Transcription-Total RNA was isolated from either HeLa cells transfected with ASO-or MMO-ClC-3 or X. laevis oocytes injected with sense or antisense ClC-3 cRNA by the use of a TRIzol total RNA isolation reagent (Invitrogen, Gaithersburg, MD) following the manufacturer's instructions. The total RNA was quantified by absorbance at 260 nm in a spectrophotometer, and the 260/280 absorbance ratio was determined for purity. Total RNA was incubated with RNase-free DNase (Promega, Madison, WI) for 20 min at 25°C followed by heat inactivation at 65°C. cDNA was synthesized using 1 g of RNA reverse-transcribed with 200 units of Superscript II TM reverse transcriptase (Invitrogen) in a 20-l reaction containing 25 ng of oligo(dT) [12][13][14][15][16][17][18] primer (HeLa cells) or random hexamer primer (oocytes), 500 M each of dNTP, 50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl 2 , and 10 mM dithiothreitol.
The quantification of X. laevis oocyte ClC-3 mRNA content was done using real-time PCR. The specific primers used to quantitate xClC-3 expression in X. laevis oocytes were as follows: xClC-3 (GenBank TM accession number Y09941), forward 5Ј-CCCAATGGATATCTCTTCAG-ATC-3Ј (sense, nt 87-109) and reverse 5Ј-CAATAGGTGAGTCGTGCT-GT-3Ј (antisense, nt 225-206), amplicon 138 bp; 18 S rRNA (Gen-Bank TM accession number K01373), forward 5Ј-ACAGTGAAACTGCG-AATGGCT-3Ј (sense, nt 77-97) and reverse 5Ј-GCTCGTCGGCATGTA-TTAGCT-3Ј (antisense, nt 183-163), amplicon 107 bp. The amplification profile for these primer pairs was as follows: 95°C for 10 min to activate the AmpliTaq polymerase, then 40 cycles of 95°C for 15 s and 60°C for 1 min, performed in a GeneAmp 2400 thermal cycler (PE Applied Biosystems, Foster City, CA). Real-time quantitative PCR was performed with the use of SYBR Green chemistry on an ABI 5700 sequence detector (PE Applied Biosystems, Foster City, CA). Standard curves were generated for xClC-3 and the constitutively expressed 18 S rRNA from regression analysis of the mean values of RT-PCR values for the log 10 diluted cDNA. Unknown quantities relative to the standard curve for the xClC-3 primers were calculated, yielding transcriptional quantitation of xClC-3 cDNA relative to the endogenous standard (18 S rRNA). Each cDNA sample was tested in triplicate, and cDNA was obtained from at least five different frogs. The reproducibility of the assay was tested by analysis of variance (ANOVA) comparing repeat runs of samples, and mean values generated at individual time points were compared by Student's t test.
The PCR products were resolved in 1.5 and 2.5% agarose/ethidium bromide gels, for the HeLa and X. laevis ClC-3 amplification products, respectively. Bands were observed under an UV transilluminator and analyzed with Molecular Analyst software (Bio-Rad, Richmond, CA).
Net intensity values were obtained for each band resolved in the agarose gels.
Immunoblotting-Crude and membrane protein from HeLa cells and X. laevis oocytes were extracted as described previously (28). Protein content in HeLa cells and X. laevis oocytes was determined by the Bradford method (38) or the bicinchoninic acid method (39), respectively.
Total protein was dissolved in sample buffer (40) and subjected to electrophoresis in parallel to prestained low range molecular weight standards on 8% SDS-PAGE gels. Gels were then blotted to nitrocellulose, and proteins were electrically transferred (41). Nitrocellulose nonspecific binding was blocked with 5% dried nonfat milk, 0.05 M Trisbuffered saline (TBS), pH 7.2. Blots were then initially incubated with the commercially available anti-ClC-3 Ab (Alomone Laboratories, Jerusalem, Israel) raised against a C terminus epitope (amino acids 592-661) at a final concentration of 7.5 g. Due to recent concerns (35) regarding the specificity of this Ab, some blots were incubated with a new rabbit polyclonal affinity-purified anti-ClC-3 Ab (42) raised against a different C terminus epitope (amino acids 670 -687 of the mouse sequence) at a final concentration of 25 g. The peptide used to generate the C 670 -687 Ab is similar to the ClC-3 epitope previously used to generate a specific C terminus anti-ClC-3 Ab, which exhibited no crossreactivity with human ClC-1, rat ClC-2, or human ClC-4 (43). Ab excess was removed with 0.1% Tween 20-TBS, and specific ClC-3 reactivity was determined with a goat-anti rabbit IgG conjugated to peroxidase or alkaline phosphatase (Jackson ImmunoResearch Laboratories, West Grove, PA). The peroxidase activity was detected by the ECL chemiluminescence reagent (PerkinElmer Life Sciences, Boston, MA) and revealed with Kodak films. Immunoreactive bands present in the films were observed under a visible light transilluminator and digitized with a Kodak DC40/DC120 camera. Images were analyzed with the Kodak digital science 1D program (KdS 1d, Kodak). Phosphatase activity was detected with Nitro BT/5-bromo-4-chloro-3-indolyl phosphate tablets. Net intensity values for each band were analyzed with Molecular An-alyst software (Bio-Rad, Richmond, CA).
The isotonic solution (ND72) for X. laevis oocyte current recordings contained, in mM: 72 NaCl, 2 KCl, 1 CaCl 2 , 1 MgCl 2 , 5 Hepes, and 55 mannitol, pH 7.5. 100 M niflumic acid was added to all solutions to eliminate contamination by endogenous I Cl.Ca . All whole cell recordings were in either isotonic (220 mosM) or hypotonic (165 mosM, by exclusion of mannitol) bath solutions. For cell volume measurements, NaCl was replaced by N-methyl-D-glucamine chloride at a concentration of 100 mM adjusting the solution to be 50% hypotonic. All solutions were tested with a freezing point depression osmometer.
Electrophysiological Experiments-HeLa cells transfected with FITC-ASO-ClC-3 or FITC-MMO-ClC-3 were plated on 25-mm round coverslips and placed in a microchamber mounted on a fluorescent inverted microscope. The bath was grounded via an agar bridge. Standard whole-cell patch-clamp recordings were performed using the Axopatch 200 B (Axon Instruments, Foster City, CA) amplifier. Patchclamp pipettes were made from thin borosilicate (hard) glass capillary tubing with an outside diameter of 1.5 or 1.7 mm (Clark Electromedical, Edenbridge, UK) using a BB-CH puller (Mecanex, Geneva, Switzerland) and had resistances of 3-5 M⍀. Voltage and current signals from the amplifier were recorded on a digital tape recorder (DTR-1204, Biologic, France), digitized using a computer equipped with a Digidata 1200 (Axon Instruments) analog-to-digital/digital-to-analog interface and analyzed with Axon software. Changes in liquid junction potentials, which occurred as a result of bath solution changes during an experiment, were calculated (44), and current-voltage relations were corrected accordingly. HeLa cells were clamped at a 0-mV holding potential, and voltage steps were applied from Ϫ80 to ϩ120 mV for 500 ms in 40-mV increments unless otherwise stated.
Membrane currents from X. laevis oocytes 0 -3 days post surgery were recorded using a two-microelectrode voltage clamp system (GeneClamp 500B, Axon Instruments). Double-blind recordings were performed in all experiments. Microelectrodes, filled with 3 M KCl, had resistances of 0.5-3 M⍀. Voltages are reported with reference to the bath. Membrane currents were filtered at 1.0 kHz, digitized on-line, and stored on a computer. Data acquisition and analysis were performed using pClamp 6 software (Axon Instruments). Oocytes were clamped at a Ϫ30mV holding potential and voltage steps were applied from Ϫ100 to ϩ120 mV for 400 ms in ϩ20-mV increments unless otherwise stated.
Cell Volume Measurements-Cell water volume was assessed in single HeLa cells by measuring changes in concentration of an intracellularly trapped fluorescent dye (45). HeLa cells transfected with TAMRA-ASO-ClC-3 or TAMRA-MMO-ClC-3 were loaded with 5 M calcein-AM/ 10% pluronic acid for 5 min at 22°C, and volume measurements were performed using a confocal laser imaging system (LSM, Carl Zeiss). Excitation light was 488 and 523 nm, for calcein and TAMRA, respectively, and emitted light was measured at wavelengths longer than 515 nm. Images of transfected HeLa cells were obtained at 10-s intervals, and fluorescence of a ϳ10-m 2 area in the cell center was measured and subsequent correction for fluorescence decay was depicted. The data are presented as V t /V 0 , the ratio between the cell water volume in isoosmotic solution at time ϭ 0 and time ϭ t, calculated from F 0 /F t (F ϭ fluorescence intensity) as described (45).
Cell volume was estimated in oocytes by diameter measurements (46). Oocyte diameters were measured using a digital video line measurement system (Ionoptix, Milton, MA) with contrast analysis of digitized image data. The measurement system was mounted to a stereooptic microscope (Cambridge Instruments). Oocytes were illuminated with focused broad-spectrum light emitted from a halogen lamp. Specifically, the left and right edges of the oocyte were measured independently using separate video lines placed at the peripheral edges of the oocyte. These lines were stripped out of the image, and intensity was displayed following online smoothing with digital filters to improve the signal-to-noise ratio. A threshold value and a crossing condition were then specified; crossing the threshold value was treated as the edge location. Left and right edges were scanned from the outside toward the inside of the cell to ensure proper edge detection. Edge data were acquired at 1 Hz and stored on either compact disc or magnetic media for later analysis. Changes in cell volume were calculated as the ratio of V t /V t Ϫ V o , where V t is the volume at time t and V o is the basal volume. Cell volumes were estimated using the equation, V ϭ 4/3r 3 , assuming an oocyte is a perfect sphere.
HeLa cells and X. laevis oocytes were superfused with isoosmotic solution for 20 min before starting each experiment. After the isoosmotic period, a hypotonic solution containing 10 M gramicidin was superfused. The rationale of the experiment was to make the rate of change in cell volume proportional to the anion (i.e. Cl Ϫ ) permeability instead of the K ϩ conductance (47). For HeLa cells and oocytes, this was achieved by replacing all monovalent cations in the medium with the impermeant cation NMDG and exposing them to 10 M gramicidin, causing the Cl Ϫ conductance to be rate-limiting.
Reagents-Calcein was obtained from Molecular Probes Inc. (Eugene, OR). All other reagents were of analytical grade and were purchased from Sigma Chemical Co. (St. Louis, MO) and Merck (Darmstadt, Germany).
Statistics-Data are expressed as means Ϯ S.E. For the electrophysiological and cell volume studies, n ϭ the number of cells. For the quantification of ClC-3 RNA and protein expression levels in oocytes, n ϭ 2 and n ϭ 8 oocytes were used at each time point, respectively. Statistical analyses were by means of Student's t test and one-way ANOVA followed by the Newman-Keuls multiple comparison test where appropriate. All statistical analyses were performed using GraphPad Prism version 3 (GraphPad Software Inc., San Diego, CA). Differences were considered to be significant when p Ͻ 0.05.

ClC-3 Transcript Expression in HeLa
Cells-To study endogenous ClC-3 expression in HeLa cells, ClC-3 mRNA was detected using in situ hybridization and semiquantitative PCR. HeLa cells in culture expressed the ClC-3 transcript. The in situ ClC-3 reactivity was confined to the cytoplasm and mostly associated to the perinuclear cell area (Fig. 1A, left). HeLa cells did not react with the ClC-3 sense probe (Fig. 1A, right).
ClC-3 Antisense Oligonucleotides Reduce Endogenous ClC-3 Expression in HeLa Cells-To evaluate endogenous ClC-3 mRNA and protein expression levels after treatment with ASO-ClC-3 or MMO-ClC-3, semiquantitative PCR and Western blot analysis were performed. Total RNA from HeLa cell was extracted at 0, 6, 12, 16, 24, and 48 h after ASO-or MMO-ClC-3 oligonucleotide treatment, and samples were analyzed by semiquantitative PCR. The ClC-3 amplicon was eliminated after 24 h of ASO-ClC-3 treatment as illustrated in the representative gel (Fig. 1B, lane 6, 24 h, left). A decrease in the 533-bp PCR ClC-3 product intensity was observed after transfection, which followed type 1 exponential kinetics (r ϭ 0.99) (Fig. 1B,  right). HeLa cells treated with either LipofectAMINE or MMO-ClC-3 did not significantly modify endogenous ClC-3 transcript expression (data not shown).
The relative level of ClC-3 protein expression was evaluated using two different anti-ClC-3 Abs in HeLa cells at 0, 24, 48, and 72 h after transfection with either ASO-ClC-3 or MMO-ClC-3 (Fig. 1C). Both C 592-661 and C 670 -687 Abs recognized two bands of ϳ90 and 113 kDa. In HeLa cells the 90-kDa band was confirmed to be ClC-3, because a progressive decrease in the densities of this band was detected after ASO-ClC-3 treatment.

FIG. 2. Inhibition of endogenous ClC-3 mRNA and protein expression by injected antisense cRNA in X. laevis oocytes.
A, real-time quantitative RT-PCR for endogenous ClC-3 expression relative to 18 S rRNA in X. laevis oocytes (n ϭ 10). Bars indicate mean data for antisense cRNA-injected oocytes as a percentage of control senseinjected oocytes. B, representative Western blot of native ClC-3 expression in membrane extracts from oocytes (left, sense; right, antisense). Immunoblots were labeled with C 670 -687 Ab to detect endogenous xClC-3. C, mean band densities for endogenous ClC-3 protein expression (n ϭ 24). Immunoblots were labeled with either the C 670 -687 Ab (left) or the C 592-661 Ab (right). Bars indicate mean data for antisense cRNA-injected oocytes as a percentage of control sense-injected oocytes. Error bars represent mean Ϯ S.E. **, p Ͻ 0.01 denotes mean significantly different from time point 0 for mRNA expression; ***, p Ͻ 0.001 denotes mean significantly different from 0 and 24 h for protein expression by a one-way ANOVA with a Newman-Keuls multiple comparison procedure.

xClC-3 cRNA Antisense Reduces Endogenous ClC-3 Expression in X. laevis Oocytes-
To verify reduction of endogenous ClC-3 protein expression by antisense xClC-3 cRNA injection, Western blot analysis was performed using the two different anti-ClC-3 Abs. Fig. 2B is a representative Western blot showing a decrease in endogenous ClC-3 expression by antisense (right) but not sense (left) cRNA injected X. laevis oocytes after 0, 24, 48, and 72 h using the C 670 -687 Ab. Two bands corresponding to ϳ86 -88 kDa and 113 kDa were evident, which were not observed in the secondary Ab controls (data not shown). The 86-to 88-kDa band is close to the predicted molecular mass for xClC-3 of 88 kDa (48). In immunoblots incubated with C 670 -687 Ab, antisense xClC-3 cRNA significantly decreased the density of this band by 64% at 48 h and 83% at 72 h when compared with time point 0 (Fig.   2C, left). In immunoblots incubated with the C 592-661 Ab, there also was a significant decrease in endogenous ClC-3 protein expression of 60% at 48 h and 94% at 72 h when compared with time point 0 (Fig. 2C, right).

Inhibition of ClC-3 Protein Translation by Antisense Oligonucleotides Reduces Native VSOACs in HeLa Cells-VSOAC
currents from single HeLa cells, transfected with 1 M ASO-or MMO-ClC-3 ϩ 5 g/ml LipofectAMINE, were recorded after 0, 24, 48, and 72 h. Almost 50% of the cells took up the FITClabeled oligonucleotides, as detected using confocal fluorescence microscopy (Fig. 3, A-C; representative light, fluorescent and superimposed images of oligonucleotide-treated cells at 48 h with FITC-oligonucleotides). Fluorescence microscopy was used to detect the cells that had incorporated the FITC-oligonucleotides. The uptake process may occur by fluid-phase endocytosis (pinocytosis), perhaps mediated by receptor-like recognition, and it may depend on factors such as oligonucleotide chain length and class (49). Fig. 3D illustrates representative whole cell currents in response to voltage steps of 40-mV increments from Ϫ80 to ϩ 120 mV. Very little current was activated by voltage clamp pulses when MMO-ClC-3-transfected HeLa cells were exposed to isotonic bath solutions (Fig. 3D). When the isotonic bath solution was exchanged to a hypotonic solution, large VSOAC currents were activated in response to applied voltage clamp pulses in MMO-ClC-3-transfected HeLa cells (Fig. 3D), which resemble the endogenous VSOAC currents present in untransfected HeLa cells (50). The currents exhibited outward rectification and time-and voltage-dependent inactivation at positive potentials (Fig. 3D). Current activation was reversed by re-exposure of cells to isotonic bath solutions (data not shown). In contrast, hypotonic solutions induced very little VSOAC acti- vation in ASO-ClC-3-transfected HeLa cells (Fig. 3D). Fig. 3E shows normalized VSOAC current densities (difference currents recorded in hypotonic and isotonic solutions) at ϩ120 and Ϫ80 mV, from ASO-ClC-3-transfected HeLa cells after 24, 48, and 72 h. There was a significant decrease in VSOAC current density at ϩ120 and Ϫ80 mV after 48 h and at ϩ120 mV after 72 h in ASO-ClC-3-treated HeLa cells (Fig. 3E). ASO-ClC-3 treatment decreased native VSOAC currents at Ϫ80 mV, by 85.6% and 24.4% and at ϩ120 mV, by 88% and 67.3% at 48 and 72 h, respectively, compared with control, untransfected cells. VSOAC current densities for MMO-ClC-3-transfected cells were not significantly different from controls (data not shown).
xClC-3 cRNA Antisense Reduces Native VSOACs in X. laevis Oocytes-X. laevis oocytes were injected with antisense xClC-3 cRNA to examine whether or not endogenous xClC-3 contributes to native VSOACs in this cell type. Oocytes were injected with either 9 ng of sense or antisense xClC-3 cRNA. Currents were measured using the two-electrode voltage clamp technique at 0, 24, 48, and 72 h following injections. To obtain whole cell current voltage relations, cells were held at Ϫ30 mV, and voltage pulses were applied from Ϫ100 to ϩ120 mV in 20-mV increments. To elicit VSOAC currents, oocytes were exposed to a hypotonic solution (165 mosM; 25% hypotonic), which contained 100 M niflumic acid to prevent activation of endogenous I Cl.Ca (51,52). When sense xClC-3 cRNA-injected oocytes 72 h post-injection were exposed to isotonic solutions, only small background currents were observed. However, when the bath was replaced by a hypotonic solution, large native VSOAC currents were elicited that showed outward rectification and voltage-dependent inactivation at positive potentials (Fig. 4A). On the other hand, 72 h following antisense xClC-3 injection, oocytes exhibited little, if any, activation of native VSOAC currents when challenged with hypotonic solutions (Fig. 4B). Fig. 4C summarizes the results from a number of experiments. Antisense cRNA against xClC-3 decreased native oocyte VSOAC current densities at Ϫ80 mV, by 57 and 92%, and at ϩ80 mV, by 70 and 86% at 48 and 72 h, respectively, compared with control sense injected oocytes.
ClC-3 Antisense Oligonucleotides Inhibit Regulatory Volume Decrease in HeLa Cells-To study the role of ClC-3 in the RVD process, experiments were conducted to monitor cell volume using the intracellularly trapped fluorescent dye calcein. Cellular uptake of the TAMRA-labeled oligonucleotides and calcein was monitored by confocal fluorescence microscopy (Fig. 5, A-C, representative light, fluorescent, and superimposed images of oligonucleotide-treated cells at 48 h with TAMRA-oligonucleotides). Fig. 5D shows experiments that examined the role of Cl Ϫ permeability in RVD by monitoring changes in cell volume following exposure to hypotonic solutions after 48 h of transfection with either MMO-ClC-3 or ASO-ClC-3. HeLa cells transfected with MMO-ClC-3 (open circles) swelled when exposed to a hypoosmotic NMDG-Cl solution containing 10 M gramicidin (arrow, 220 mosM). Within minutes a rapid regulatory decrease in cell volume occurred, which involved the activation of a Cl Ϫ conductance pathway. The time course of the cell volume decrease, and an estimate of Cl Ϫ conductance could be fitted to a monoexponential decay function (11). In contrast, ASO-ClC-3-transfected HeLa cells (filled circles), although also swelling in response to hypotonic exposure, were unable to regulate volume (Fig. 5D). Thus, a decrease in endogenous ClC-3 and VSOAC current expression correlated with an impaired RVD in HeLa cells.
xClC-3 cRNA Antisense Inhibits Regulatory Volume Decrease in X. laevis Oocytes-We next tested whether or not a decrease in endogenous ClC-3 expression (Fig. 2) and native VSOAC currents (Fig. 4) in X. laevis oocytes would alter RVD. Using a digital cell-diameter measurement system, cell volume estimates were made 72 h after injection of 9 ng of sense or antisense xClC-3 cRNA into oocytes.
The rationale behind these experiments was exactly the same described for volume regulation by HeLa cells. In oocytes, the latency of current activation (ϳ5-10 min) varied with each group of oocytes (53) and correlated with the change in cell volume. In oocytes injected with sense or antisense xClC-3 cRNA, swelling was induced with a hypoosmotic NMDG-Cl solution containing 10 M gramicidin (110 mosM). Fig. 6 (A and  B) illustrates that hypoosmotic exposure initially increased cell volume in both groups of oocytes as determined by changes in cell diameter. Fig. 6A shows that the oocyte volume of sense xClC-3 cRNA-injected oocytes eventually returned to near control levels in the continued presence of the hypoosmotic solution. This indicates that RVD had occurred, and, because of the experimental paradigm, this was primarily dependent on chloride efflux. In contrast, oocytes injected with antisense xClC-3 cRNA also swelled when exposed to the same hypoosmotic NMDG-Cl solution in the presence of 10 M gramicidin (110 mosM) but were unable to regulate their cell volume (Fig. 6B).

FIG. 4. Inhibition of endogenous VSOAC currents by injected antisense xClC-3 cRNA in X. laevis oocytes.
A, representative chloride currents recorded from an oocyte injected with 9 ng of sense xClC-3 cRNA 72 h post injection. B, representative chloride currents recorded from an oocyte injected with 9 ng of antisense xClC-3 cRNA 72 h post injection. Oocytes were bathed in isotonic solution (ϳ10 min, 220 mosM) before exposure to a hypotonic solution (ϳ40 min, 165 mosM) to activate native VSOACs. Currents were evoked by 400-ms voltage step exposure, in 20-mV increments from a holding potential of Ϫ30 mV. Test potentials ranged from Ϫ100 to ϩ120 mV. C, bars indicate mean data for antisense cRNA-injected oocytes as a percentage of control senseinjected oocytes (n ϭ 6). Error bars represent mean Ϯ S.E. Peak current densities were measured after 40-min hypotonic exposure to ensure maximal current activation. *, p Ͻ 0.05; **, p Ͻ 0.001 denotes mean significantly different from time point 0 and 24 h by a one-way ANOVA with a Newman-Keuls multiple comparison procedure. Fig. 6C demonstrates the mean data for a number of experiments expressed as percent change in volume. Sense xClC-3 cRNA-injected oocytes had a mean volume increase of 8.86 Ϯ 1.97% when exposed to hypotonic solutions and a regulatory volume decrease (RVD) to 2.1 Ϯ 1.2% greater than control volume. In antisense xClC-3-injected oocytes hypotonic exposure induced a 5.9 Ϯ 0.94% volume increase, and there was no subsequent RVD, because the oocyte volume remained elevated at 5.8 Ϯ 1.01% greater than control volume levels. DISCUSSION VSOACs are activated upon cell swelling in most vertebrate cells studied to date, including X. laevis oocytes and HeLa cells (9,51,52,54). The physiological role of VSOACs has yet to be fully elucidated due to a lack of specific pharmacological inhibitors and its still unknown molecular identity (55). However, the majority of evidence indicates that VSOACs may be a major pathway for regulatory volume decrease (RVD) through efflux of both chloride and organic osmolytes in several cell types (53, 56 -59).
ClC-3, a ubiquitously expressed member of the ClC chloride channel superfamily, has been proposed to encode one type of native VSOAC in guinea pig ventricular myocytes, canine pulmonary arterial smooth muscle cells (pulmonary arterial smooth muscle cells), and X. laevis oocytes (16,27,28). Since ClC-3 was first cloned and expressed by Kawasaki et al. in 1994, contradictory data has been reported for the functional expression of the ClC-3 chloride channel. Investigation into the expression levels of both mRNA and protein in various tissues has led to conflicting results (24,33,36,60). Expression of cloned ClC-3 either from human (32,36,43), rat (24,33,34), or guinea pig (16,27,28) has led to differing biophysical and pharmacological properties or the inability to detect expressed ClC-3 chloride currents (36,37,61). There is also contradictory data as to whether the protein is an intracellular channel (35) or plasma membrane channel (33,43). These conflicting results have hindered determination of the physiological role of ClC-3 and whether or not it is the VSOAC channel or a component thereof in some types of cells.
Therefore, the purpose of the present study was to perform antisense experiments in two independent laboratories to determine if ClC-3 is a component of the native VSOAC and whether or not ClC-3 is involved in RVD. For this purpose, antisense oligonucleotides (ASO-ClC-3) and antisense RNA (antisense xClC-3 cRNA) against ClC-3 were used to block the expression of the protein in HeLa cells and X. laevis oocytes, respectively. Antisense (oligonucleotides or cRNA) has become a highly specific and effective way to inhibit protein translation in vivo (62,63). The use of antisense, targeted to transcripts encoding biologically active proteins, provides a novel and highly selective means to understand the function of native proteins (64). Among the various antisense strategies available, the most specific and widely used have been shown to be antisense oligodeoxynucleotides (ASO) and antisense RNA (64). Our results show that both antisense strategies significantly inhibited native ClC-3 protein, VSOAC currents, and RVD. These results confirm that ClC-3 channel activity participates in the RVD process in both HeLa cells and X. laevis oocytes and support the role of ClC-3 as a molecular candidate for native VSOACs in these cells.
HeLa cells and X. laevis oocytes endogenously express ClC-3 mRNA transcripts and protein, and both generate Cl Ϫ currents after exposure to hypotonic solutions, whose activation and voltage dependence are similar to that previously reported for native VSOACs in these cell types (6, 50 -53, 65). Native VSOACs in X. laevis oocytes have a number of pharmacological and biophysical properties similar to that reported for expression of gpClC-3 in NIH3T3 cells (16,27) and rClC-3 (long form) expressed in Chinese hamster ovary cells (33).
In this study, native ClC-3 mRNA from HeLa cells and X. laevis oocytes was significantly reduced using two different antisense strategies to inhibit native ClC-3 protein expression. After 24 h of antisense treatment, the ClC-3 mRNA was significantly reduced when detected by semiquantitative and quantitative real-time PCR amplification (Figs. 1B and 2A), indicating the ClC-3 transcript had been successfully inhibited. The half-life of ClC-3 turnover is unknown, however, half-lives for chloride channels such as CFTR and ClC-2 have been reported. CFTR was found to have a biochemical half-life in the plasma membrane of ϳ4 h for ⌬F508, and wild-type CFTR was demonstrated to exceed 48 h (66). A half-life of ϳ2 h has been reported for ClC-2, however, a detailed biochemical analysis has yet to be done (67). Based on the time-dependent reduction of the native ClC-3 protein following both antisense treatments, the ClC-3 turnover half-life in these cells is estimated to be ϳ27-35 h (Figs. 1C, 2B, and 2C). The ClC-3 protein reached an apparent plateau in HeLa cells (Fig. 1C, right panel), which might be attributed to the possibility that the antisense oligonucleotides were not taken up uniformly into all cells.
Using an ASO-ClC-3 conjugated to FITC in HeLa cells, it was possible to monitor ASO-ClC-3 presence in the cell nucleus after 3 h of transfection, and the fluorescence was maintained for 24 and 48 h in at least 50% of the cells. We used X. laevis oocytes for antisense RNA experiments, because cRNA can be introduced by microinjection. A significant role of ClC-3 in native VSOAC function was indicated, because HeLa cells transfected with ASO-ClC-3 and X. laevis oocytes injected with antisense xClC-3 cRNA exhibited significantly reduced VSOAC current densities compared with controls (Figs. 3D, 3E, and 4). The decrease in VSOAC currents at ϩ120 and Ϫ80 mV (HeLa cells) and ϩ80 mV and Ϫ80 mV (oocytes) due to antisense treatment directly correlated with the reduced mRNA and protein expression data (Figs. 1B, 1C, and 2). Moreover, VSOAC current densities began to recover after 72 h of ASO-ClC-3 treatment of the HeLa cells, following a similar time-course demonstrated by ClC-3 protein expression ( Fig. 1C and 3E). The incomplete inhibition of the current could be due to another molecular entity making a minor contribution to the whole cell native VSOAC currents or, possibly, because less than 100% inhibition of the endogenous ClC-3 mRNA and protein expression levels occurred.
HeLa cells and X. laevis oocytes undergo RVD following cell swelling by hypotonic stress (11,68,69), and VSOACs have been proposed to be an important osmosensitive Cl Ϫ conductance responsible for RVD (11,53). We found that RVD was significantly impaired in HeLa cells acutely transfected with ASO-ClC-3 and X. laevis oocytes injected with antisense when evaluated after 48 h and 72 h, respectively. These data suggest that inhibition of native ClC-3 expression interferes with the development of the RVD process.
In conclusion, the parallel reduction in ClC-3 mRNA, protein, VSOAC current density, and RVD following ClC-3 antisense oligonucleotide and ClC-3 antisense cRNA treatment in HeLa cells and X. laevis oocytes, respectively, demonstrated here, strongly supports a fundamental role of endogenous ClC-3 in native VSOAC function and cell volume regulation. In contrast, mismatched oligonucleotides and sense RNA having antisense-injected oocytes (n ϭ 7, solid bars). Hypotonic indicates peak percent volume change from isotonic while RVD indicates subsequent volume recovery. Oocytes were injected with 9 ng of antisense xClC-3 cRNA and 9 ng of sense xClC-3 cRNA for 72 h. Oocytes were initially perfused with a standard ND72 isotonic solution (220 mosM). The solution was then switched to a hypotonic solution (110 mosM) in which all monovalent cations were replaced by NMDG and contained 10 M gramicidin (indicated by the bar). *, p Ͻ 0.01 denotes mean significantly different from hypotonic by a one-way ANOVA with a Newman-Keuls multiple comparison procedure.
hybridization characteristics similar to the antisense failed to significantly affect mRNA, protein, VSOAC current density, or RVD in these experiments. Future experiments are required to conclusively determine whether endogenous ClC-3 represents a molecular component of the VSOAC channel protein or, alternatively, whether ClC-3 is an important accessory subunit of VSOACs or may be involved in the volume-sensing mechanism, which regulates native VSOAC function. Site-directed mutagenesis experiments (16,27) support the former possibility.