TMEM16 Proteins Produce Volume-regulated Chloride Currents That Are Reduced in Mice Lacking TMEM16A*

All vertebrate cells regulate their cell volume by activating chloride channels of unknown molecular identity, thereby activating regulatory volume decrease. We show that the Ca2+-activated Cl− channel TMEM16A together with other TMEM16 proteins are activated by cell swelling through an autocrine mechanism that involves ATP release and binding to purinergic P2Y2 receptors. TMEM16A channels are activated by ATP through an increase in intracellular Ca2+ and a Ca2+-independent mechanism engaging extracellular-regulated protein kinases (ERK1/2). The ability of epithelial cells to activate a Cl− conductance upon cell swelling, and to decrease their cell volume (regulatory volume decrease) was dependent on TMEM16 proteins. Activation of ICl,swell was reduced in the colonic epithelium and in salivary acinar cells from mice lacking expression of TMEM16A. Thus TMEM16 proteins appear to be a crucial component of epithelial volume-regulated Cl− channels and may also have a function during proliferation and apoptotic cell death.

Regulation of cell volume is fundamental to all cells, particularly during cell growth and division. External hypotonicity leads to cell swelling and subsequent activation of volume-regulated chloride and potassium channels, to release intracellular ions and to re-shrink the cells, a process termed regulatory volume decrease (RVD) 3 (1). Volume-regulated chloride currents (I Cl,swell ) have dual functions during cell proliferation as well as apoptotic volume decrease (AVD), preceding apoptotic cell death (2). Although I Cl,swell is activated in swollen cells to induce RVD, AVD takes place under normotonic conditions to shrink cells (3,4). Early work suggested intracellular Ca 2ϩ as an important mediator for activation of I Cl,swell and volume-regulated K ϩ channels (5), whereas subsequent studies only found a permissive role of Ca 2ϩ for activation of I Cl,swell (6), reviewed in Ref. 1. In addition, a plethora of factors and signaling pathways have been implicated in activation of I Cl,swell , making cell volume regulation an extremely complex process (reviewed in Refs. 1, 3, and 7). These factors include intracellular ATP, the cytoskeleton, phospholipase A2-dependent pathways, and protein kinases such as extracellular-regulated kinase ERK1/2 (reviewed in Refs. 1 and 7). Previous approaches in identifying swelling-activated Cl Ϫ channels have been unsuccessful or have produced controversial data. Thus none of the previous candidates such as pI Cln , the multidrug resistance protein, or ClC-3 are generally accepted to operate as volume-regulated Cl Ϫ channels (reviewed in Refs. 8 and 9). Notably, the cystic fibrosis transmembrane conductance regulator (CFTR) had been shown in earlier studies to influence I Cl,swell and volume regulation (10 -12). The variable properties of I Cl,swell suggest that several gene products may affect I Cl,swell in different cell types.
The TMEM16 transmembrane protein family consists of 10 different proteins with numerous splice variants that contain 8 -9 transmembrane domains and have predicted intracellular N-and C-terminal tails (13, 16 -18). TMEM16A (also called ANO1) is required for normal development of the murine trachea (14) and is associated with different types of tumors, dysplasia, and nonsyndromic hearing impairment (13,15). TMEM16A has been identified as a subunit of Ca 2ϩactivated Cl Ϫ channels that are expressed in epithelial and nonepithelial tissues (16 -18). Interestingly, members of the TMEM16 family have been suggested to play a role in osmotolerance in Saccharomyces cerevisiae (19). Here we show that TMEM16 proteins also contribute to I Cl,swell and regulatory volume decrease.
Patch Clamp-Cells were grown on coverslips that were mounted in a perfused bath on the stage of an inverted microscope (IM35, Zeiss) and kept at 37°C. Cells and acini isolated from mouse glands were allowed to settle onto poly-L-lysinecoated coverslips. The bath was perfused continuously with Ringer solution at about 10 ml/min. For activation of volumedependent Cl Ϫ currents, isotonic Ringer bath solution (mM: NaCl 145, KH 2 PO 4 0.4, K 2 HPO 4 1.6, D-glucose 6, MgCl 2 1, calcium gluconate 1.3, pH 7.4) was changed to isotonic control solution in which 50 mmol/liter of NaCl was replaced by 100 mmol/liter of mannitol. Subsequently, mannitol (50, 75, and 100 mmol/liters) was removed to produce extracellular hypotonicity of Ϫ50, Ϫ75, and Ϫ100 mosmol/liter, i.e. 17, 25, and 33% hypotonicity. Patch clamp experiments were performed in the fast whole cell configuration. Patch pipettes had an input resistance of 2-4 M⍀, when filled with an intracellular like "physiological" solution containing (mM) KCl 30, potassium gluconate 95, NaH 2 PO 4 1.2, Na 2 HPO 4 4.8, EGTA 1, calcium gluconate 0.758, MgCl 2 1.034, D-glucose 5, ATP 3, pH was 7.2. The Ca 2ϩ activity was 0.1 M. We choose this solution because it enabled normal swelling/shrinkage behavior and allowed for direct comparison of results from patch clamping and volume measurements. The access conductance was measured continuously and was 30 -140 nS. Currents (voltage clamp) and voltages (current clamp) were recorded using a patch clamp amplifier (EPC 7, List Medical Electronics, Darmstadt, Germany), the LIH1600 interface and PULSE software (HEKA, Lambrecht, Germany) as well as Chart software (AD Instruments, Spechbach, Germany). Data were stored continuously on a computer hard disc and analyzed using PULSE software. In regular intervals, membrane voltage (V c ) was clamped in steps of 10 mV from Ϫ50 to ϩ50 mV relative to resting potential. Membrane conductance G m was calculated from the measured current (I) and V c values according to Ohm law.
Animals and Ussing Chamber Experiments-Generation of a null allele of Tmem16a and TMEM16A knock-out animals has been described previously (14). Pups (1-4 days) were sacrificed with Isofluran (Baxter, Germany). The pancreas and submandibular glands were removed and epithelial cells were dispersed in phosphate-buffered saline composed of Collagenase VIII (Sigma). Tracheas were dissected, opened longitudinally on the opposite side of the cartilage-free zone, and transferred immediately into an ice-cold buffer solution. Stripped colon was put into ice-cold Ringer bath solution containing amiloride (10 M) and indomethacin (10 M). For activation of volume-dependent Cl Ϫ currents, isotonic Ringer bath solution was changed to isotonic control solution in which 50 mmol/liter of NaCl was replaced by 100 mmol/liter of mannitol. Subsequently, mannitol was removed to produce extracellular hypotonicity of Ϫ100 mosmol/liter, i.e. 33% hypotonicity. Tissues were mounted into a microperfused Ussing chamber with a circular aperture of 0.785 mm 2 . Luminal and basolateral sides of the epithelium were perfused continuously at a rate of 5 ml/min. Bath solutions were heated to 37°C, using a water jacket. Experiments were carried out under open circuit conditions. Data were collected continuously using PowerLab (AD Instruments, Australia). Values for transepithelial voltages (V te ) were referred to the serosal side of the epithelium. Transepithelial resistance (R te ) was determined by applying short (1 s) current pulses (⌬I ϭ 0.5 A). R te and equivalent short circuit currents (I SC ) were calculated according to Ohm law (R te ϭ ⌬V te /⌬I, I SC ϭ V te /R te ).
Analysis of Cell Volume and FACS-For cell volume measurements, HEK293 cells were loaded with 2 g of calcein-AM (Molecular Probes) and 0.025% pluronic in a standard bath solution (Ringer) for 30 -60 min at 37°C. Analysis was done at an excitation wavelength of 500 nm and and emission wavelength of 520 -550 nm. Cell swelling and RVD were observed for 10 -15 min after applying hypotonic bath solution. To produce extracellular hypotonicity, isotonic Ringer bath solution was changed to isotonic control solution in which 50 mmol/ liter of NaCl was replaced by 100 mmol/liter of mannitol. Subsequently, mannitol was removed to reduce extracellular hypotonicity by 100 mosmol/liter (33% hypotonicity). Experiments were performed 48 h after transfection of HEK293 cells with cDNAs encoding P2Y2 and hTMEM16A and 72 h after transfection with siRNA targeting TMEM16 family members or scrambled. All the compounds were dissolved in bath solutions and their effect on cell swelling was analyzed after 5-10 min incubation. Cell volume of HEK293 cells was also assessed by flow cytometry (Prof. Mathias Mack, University Hospital Regensburg) before (t ϭ 0) and 30 s (t ϭ 0.5 min) and 5 min (t ϭ 5 min) after swelling, which was induced by reduction of the solution osmolarity to 1 ⁄ 3. In flow cytometry experiments, HEK293 cells were resuspended in phosphate-buffered saline and the average volume of 30,000 cells per condition determined.
Materials and Statistical Analysis-All compounds (ionomycin, carbachol, ATP, tamoxifen, DIDS, suramin, 1,2bis(2-aminophenoxyl)ethane-N,N,NЈ,NЈ-tetraacetic acid-AM, U0126, staurosporine, CFTR inh-172 , apyrase, MTSET, and cyclopiazonic acid) were of highest available grade of purity and were from Sigma or Merck. The anti-hTMEM16A was a generous gift from Prof. van de Rijn (Dept. of Pathology, Stanford University). All cell culture reagents were from Invitrogen. For activation of volume-dependent Cl Ϫ currents, isotonic Ringer bath solution was changed to isotonic control solution in which 50 mmol/liter of NaCl was replaced by 100 mmol/liter of mannitol. Subsequently, mannitol (50, 75, and 100 mmol/liter) was removed to produce extracellular hypo-tonicity of Ϫ50, Ϫ75, and Ϫ100 mosmol/liter, i.e. 17, 25, and 33% hypotonicity. RVD due to swelling activation of ion channels is an initial event that was assessed by analysis of the initial volume decrease. RVD indicates the initial rate of regulatory volume decrease or initial RVD rate, RVDi(%) ϭ (20). Initial rate of RVD in hepatocytes is controlled by extracellular ATP (21). Student's t test (for paired or unpaired samples as appropriate) and analysis of variance was used for statistical analysis. p Ͻ 0.05 was accepted as significant.

RESULTS AND DISCUSSION
Suppression of I Cl,swell by siRNA-TMEM16A in Three Different Cell Lines-Because TMEM16A has been shown to be a major component of the Ca 2ϩ -activated Cl Ϫ channel, we examined the contribution of TMEM16A to swelling-activated whole cell currents (I Cl,swell ) in a human pancreatic cell line (CFPAC) that lacks expression of the CFTR Cl Ϫ channel (22). CFPAC cells express TMEM16A, -F, and -H and show pronounced activation of whole cell Cl Ϫ currents when exposed to hypotonic bath solution (22) (Fig. 1, A and B). Activation of whole cell currents was suppressed by siRNA for TMEM16A, but not by siRNA knockdown of CLC3, ClCA1, and CLCA4, which are also expressed in CFPAC cells (Fig. 1C). This result was confirmed in the human colonic epithelial cell line HT 29 , which also expresses TMEM16A, -F, -H, and -J ( Fig. 2A). I Cl,swell was activated upon exposure to increasing extracellular hypotonicity (17,25, and 33%, Fig. 2B, arrows). Partial removal of Cl Ϫ from the extracellular bath solution (Fig. 2, white bars, 30 mM Cl Ϫ ) reduced I Cl,swell and depolarized V m by 9.1 Ϯ 0.8 mV  OCTOBER 16, 2009 • VOLUME 284 • NUMBER 42 (n ϭ 7), indicating activation of Cl Ϫ conductance by hypotonic bath solution. Importantly, activation of I Cl,swell was reduced by siRNA-TMEM16A (Fig. 2, B and C). Finally, the role of TMEM16A for I Cl,swell was examined in HEK293 cells, which also express endogenous TMEM16 proteins (Fig. 3A). Equal to HT 29 and CFPAC cells, I Cl,swell was activated upon exposure to hypotonic bath solution (Fig. 3B). Partial removal of Cl Ϫ from the extracellular bath solution reduced I Cl,swell and depolarized V m by 8.2 Ϯ 0.9 mV (n ϭ 7). Most important, I Cl,swell was inhibited by siRNA-TMEM16A (Fig. 3, B and C).

TMEM16A Is Activated by Cell Swelling
Members of the TMEM16 Family Participate in I Cl,swell -HEK293 cells that express low but detectable amounts of TMEM16A (Figs. 3A and 4A) were incubated with two different batches of siRNA and knockdown of TMEM16A was verified by real time RT-PCR (Fig. 4B). Activation of whole cell conductance by hypotonic bath solution (G Hypo ) was potently inhibited by the inhibitor of TMEM16A currents, DIDS (100 M) (16), the specific inhibitor of Ca 2ϩ -activated Cl Ϫ currents CACCinh-A01 (23) and tamoxifen (100 M) (Fig. 4C).
We further confirmed the role of TMEM16 proteins for swelling-activated Cl Ϫ currents by overexpression of exogenous TMEM16A or TMEM16B, another Ca 2ϩ -activated Cl Ϫ channel and member of the TMEM16 family (24). Overexpression of both TMEM16 proteins augmented swelling-activated Cl Ϫ currents in HEK293 cells (Fig. 4D). This suggests that several TMEM16 proteins may contribute to swelling-activated   Cl Ϫ currents. We therefore knocked down endogenous TMEM16F, -H, and -J individually or simultaneously, each by two different batches of siRNA, which in each case significantly reduced I Cl,swell (Fig. 4E). Because knockdown of each TMEM16 protein caused a non-additive reduction in I Cl,swell , this may suggest that TMEM16 proteins hetero-oligomerize to form swelling-activated Cl Ϫ channels. In preliminary studies we found that endogenous volume-activated Cl Ϫ channels in HEK293 cells were potently suppressed by expression of K610K-TMEM16A, a TMEM16A mutant that produced very little Cl Ϫ conductance. 4 Role of Purinergic Signaling for Activation of TMEM16A by Cell Swelling-Coexpression of G-protein-coupled receptors has been shown to increase Ca 2ϩ -activated Cl Ϫ currents produced by TMEM16A (16). Although P2Y 2 receptors are expressed endogenously in HEK293 cells, additional expression of exogenous purinergic P2Y 2 receptors further increased swelling-activated whole cell currents in HEK293 cells (Fig. 5A).
Intracellular ATP and swelling-induced ATP release from cells has been identified as an important factor for RVD (25). Moreover, it has been shown that ATP binds to purinergic P2Y 2 receptors and activates TMEM16A (16). We found that   OCTOBER 16, 2009 • VOLUME 284 • NUMBER 42 blockage of P2Y 2 receptors with 100 M suramin largely reduced swelling-activated whole cell currents (Fig. 5, B and  C), confirming the role of P2Y 2 signaling for I Cl,swell . Because ATP release is obviously a critical factor for activation of I Cl,swell , we exposed the cells to hypotonic bath solution in the presence of the ATP-hydrolyzing enzyme apyrase, or removed ATP from the pipette filling solution, which both significantly reduced I Cl,swell (Fig. 5C). Moreover, TMEM16A has been shown to be inhibited by the sulfhydryl reagent MTSET, which reacts with three cysteine residues in the putative pore forming loop of TMEM16A (16). MTSET also blocked swelling-activated whole cell currents in HEK293 cells. Remarkably, the ERK1/2 inhibitor U0126 completely inhibited I Cl,swell thus clearly implying a role of this kinase for activation of TMEM16 channels (Fig. 5C). Notably, also ATP, i.e. Ca 2ϩ -activated whole cell currents, were reduced by U0126 (data not shown).

TMEM16A Is Activated by Cell Swelling
We further examined the contribution of Ca 2ϩ and ERK1/2 for activation of I Cl,swell and Ca 2ϩ -dependent activation of TMEM16A by ATP, using a Ca 2ϩfree solution (in patch pipette and bath) and by mutating two potential C-terminal ERK1/2 consensus sides (Ser 967 and Ser 970 ) in TMEM16A. We found that eliminating Ca 2ϩ abolished activation of TMEM16A by ATP and inhibited I Cl,swell (Fig. 5D). Mutating Ser 967 to alanine inhibited ATPactivated currents, whereas mutating Ser 970 to alanine inhibited both ATP-activated Cl Ϫ currents (TMEM16A) and I Cl,swell (Fig. 5D). Thus both swelling activation and ATP-dependent activation of TMEM16A use Ca 2ϩ and ERK1/2 as intracellular messengers. Although I Cl,swell requires ERK1/2 for activation, Ca 2ϩ has only a permissive function (26,27). ATP release by swollen cells may activate ERK1/2 at very low local extracellular ATP concentrations, and may thereby induce swelling-activated Cl Ϫ conductance, as reported earlier (27).
To further demonstrate that ATP and cell swelling target the same Cl Ϫ channel, we applied ATP before and after exposure to hypotonic bath solution (paired experiments). We observed that after activation of I Cl,swell , ATP-induced (Ca 2ϩ dependent) whole cell currents were largely reduced (Fig. 5, E and F). Similar non-additive effects of ATP and cell swelling on whole cell Cl Ϫ conductance were observed in HT 29 cells (data not shown).
TMEM16 Proteins Are Essential for RVD in HEK293 Cells-To further confirm the role of TMEM16 channels for I Cl,swell and RVD we loaded HEK293 cells with the fluorescence dye calcein (28). Hypotonic bath solution induced cell swelling and decreased calcein fluorescence, which was followed by a 98% recovery after 20 min, indicating RVD (Fig. 6A). Recovery from swelling was reduced in cells treated with siRNA-TMEM16A (42% after 20 min). Similarly inhibition of P2Y 2 receptors by suramin, or application of the Cl Ϫ channel blockers tamoxifen and DIDS, or apyrase, all significantly reduced RVD (Fig. 6, A  and B).
These results obtained in single cell measurements were confirmed by FACS analysis of a larger number of cells. HEK293 cells treated with scrambled siRNA increased their cell volume, which returned to control values within 5 min of exposure to hypotonic bath solution (Fig. 6C). In contrast, cells treated with siRNA for TMEM16A did not return to their initial volume, demonstrating reduced RVD. Notably, overexpression of P2Y 2 FIGURE 7. Impaired I Cl,swell in epithelial tissues of mice lacking TMEM16 ؊/؊ . A and B, whole cell currents and corresponding I/V curves obtained in submandibular acinar cells of TMEM16A knock-out mice (Ϫ/Ϫ) and wild type littermates (ϩ/ϩ) before and after exposure to hypotonic (33%) bath solution. C, summary of the whole cell conductance induced by hypotonic bath solution in WT and knock-out animals. D and E, whole cell currents and corresponding whole cell conductance obtained in hepatocytes of TMEM16A knock-out mice (Ϫ/Ϫ) and wild type littermates (ϩ/ϩ), before and after exposure to hypotonic bath solution. F, recording of the change in transepithelial voltage (V te ) induced by hypotonic bath solution (25%) in WT (ϩ/ϩ) and knock-out (Ϫ/Ϫ) animals. G, summary of ion transport, i.e. the short circuit currents induced by hypotonic bath solution in WT (ϩ/ϩ) animals and knock-out (Ϫ/Ϫ) littermates. Mean Ϯ S.E., (n) ϭ number of cells measured. *, significant increase in whole cell conductance (paired t test). #, significant difference when compared with WT (ϩ/ϩ) (unpaired t test).
receptors together with TMEM16A reduced the cell volume of HEK293 cells under control conditions. Overexpression of P2Y 2 receptors probably sensitizes HEK293 cells toward extracellular ATP, thus P2Y 2 /TMEM16A expressing cells activate Cl Ϫ channels more rapidly and therefore tend to reduce their cell volume. Moreover, their large RVD capacity counteracts any tendency to swell and therefore the cells show only a small increase in cell volume when exposed to hypotonic bath solution (Fig. 6C). We determined the rate of recovery from cell swelling and found that not only TMEM16A but also the other TMEM16 proteins expressed in HEK293 cells contribute to RVD, as demonstrated by siRNA knockdown of the individual TMEM16 proteins (Fig. 6D).
It has been reported earlier that cell shrinkage due to opening of ion channels not only occurs during hypotonic swelling, but also in normotonic solution, when cells go into apoptotic cell death (2,29). AVD is an initial step during apoptosis, which can be induced by compounds such as staurosporin (29). When we exposed HEK293 cells to staurosporin we observed a decrease in cell volume, indicating AVD (Fig. 6E). In contrast, knockdown of TMEM16A with two independent siRNAs significantly reduced AVD, suggesting that TMEM16A is important for both cell volume regulation (RVD) and AVD.
Impaired I Cl,swell in Epithelial Tissues of Mice Lacking TMEM16-TMEM16A is expressed at high levels in the submandibular gland and to a lower degree in hepatocytes (14,17) (Unigene data base). We examined activation of I Cl,swell in freshly isolated acinar cells from submandibular glands and in hepatocytes of 1-4-day-old WT (ϩ/ϩ) mouse pups and compared the results with those obtained in cells from mice lacking expression of TMEM16A (Ϫ/Ϫ). Exposure to hypotonic solution of submandibular acinar cells from ϩ/ϩ animals activated a whole cell current that was essentially absent in cells from Ϫ/Ϫ animals (Fig. 7, A and C). Thus the volume-activated whole cell conductance was almost abolished in submandibular gland cells from Ϫ/Ϫ mice (Fig. 7C). In hepatocytes I Cl,swell was of similar magnitude in both cells from ϩ/ϩ and Ϫ/Ϫ animals, however, currents activated in Ϫ/Ϫ cells showed a different time dependence, suggesting some contribution of TMEM16A to I Cl,swell also in hepatocytes (30).
Volume-activated Cl Ϫ currents have also been identified previously in the basolateral membrane of colonic epithelial cells (31). Opening of basolateral Cl Ϫ channels due to exposure to hypotonic bath solution reduced the lumen negative transepithelial voltage (V te ) in the distal colonic epithelium (Fig. 7F). In contrast the distal colon from TMEM16A Ϫ/Ϫ animals showed little changes when exposed to hypotonic solution, and thus swelling-induced short circuit currents were significantly reduced in the TMEM16A Ϫ/Ϫ colon (Fig.  7, F and G). Taken together the data clearly indicate that TMEM16A channels contribute to volume-regulated Cl Ϫ currents in native mouse epithelial cells. The present results suggest that volume-regulated Cl Ϫ channels may be composed of different members of the TMEM16 family, probably in a tissue-specific manner, thus giving rise to variable properties of I Cl,swell in different cell types and tissues (9). As these channels also control apoptotic cell shrinkage, they may affect cell survival, which could explain the role of these proteins in tumor development (13).