CFTR and TMEM16A are Separate but Functionally Related Cl- Channels

Previous reports point out to a functional relationship of the cystic fibrosis transmembrane conductance regulator (CFTR) and Ca2+ activated Cl- channels (CaCC). Recent findings showing that TMEM16A forms the essential part of CaCC, prompted us to examine whether CFTR controls TMEM16A. Inhibition of endogenous CaCC by activation of endogenous CFTR was found in 16HBE human airway epithelial cells, which also express TMEM16A. In contrast, CFBE airway epithelial cells lack of CFTR expression, but express TMEM16A along with other TMEM16-proteins. These cells produce CaCC that is inhibited by overexpression and activation of CFTR. In HEK293 cells coexpressing TMEM16A and CFTR, whole cell currents activated by IMBX and forskolin were significantly reduced when compared with cells expressing CFTR only, while the halide permeability sequence of CFTR was not changed. Expression of TMEM16A, but not of TMEM16F, H or J, produced robust CaCC, which that were inhibited by CaCCinh-A01 and niflumic acid, but not by CFTRinh-172. TMEM16A-currents were attenuated by additional expression of CFTR, and were completely abrogated when additionally expressed CFTR was activated by IBMX and forskolin. On the other hand, CFTR-currents were attenuated by additional expression of TMEM16A. CFTR and TMEM16A were both membrane localized and could be coimmunoprecipitated. Intracellular Ca2+ signals elicited by receptor-stimulation was not changed during activation of CFTR, while ionophore-induced rise in [Ca2+]i was attenuated after stimulation of CFTR. The data indicate that both CFTR and TMEM16 proteins are separate molecular entities that show functional and molecular interaction.


Introduction
The cystic fibrosis transmembrane conductance regulator (CFTR) is regarded as a cAMP/PKA/ATPregulated Clchannel that also controls the function of other membrane proteins [1]. Apart from the well examined effects of CFTR on epithelial Na + conductance, it has also been reported earlier that CFTR inhibits endogenous Ca 2+ activated Clcurrents (CaCC) in Xenopus oocytes, bovine pulmonary artery endothelium and isolated parotid acinar cells, by an yet unknown mechanism [2][3][4][5]. Moreover, also volume-regulated anion channels were inhibited by expression of CFTR [6]. Additional studies demonstrated that upregulation or downregulation of CFTR resulted in a parallel up-and downregulation of cAMP-, Ca 2+ -, and volume regulated Clconductance [7]. Although these data suggested a functional and even molecular relationship of different anion conductances, they are regarded as separate molecular entities [8].
Another line of evidence, indicating the relationship of CFTR and CaCC, came from the observation that CaCC is augmented in airways of patients suffering from cystic fibrosis. Thus purinergic peak secretory response was enhanced in cultured CF airway epithelial cells as well as freshly excised nasal tissues from CF patients [9,10]. Also in nasal epithelial of CF mice, enhanced Ca 2+ dependent Clsecretion was detected [11]. It was concluded that expression levels for alternative Ca 2+ dependent Clchannels are an important determinant of the severity of CF [12]. Also in the intestinal epithelium of CFTR-deficient mice that were generated by backcrossing and intercrossing of progenies with different inbred strains, prolonged survival was observed that was paralleled by up-regulation of CaCC [13].
In the previous search for modifiers of disease severity of cystic fibrosis, the CLCA gene locus was identified as a modulator of the gastrointestinal basic defect in CF [14]. Although the role of CLCA for Ca 2+ dependent Clconductance remains illusive, its role during inflammatory airway disease is clearly emerging [15,16]. However, CLCA proteins are probably not independent CaCCs, but may contribute to the activity of the actual membrane localized CaCC. Notably, hClCa1 has been shown to elevate the single channel conductance of endogenous Ca 2+ dependent Clchannels by lowering the energy barriers for ion translocation through the pore [17].
Recently TMEM16A has been identified as CaCC by three independent groups [18][19][20]. TMEM16A knockout animals have severe transport defects in various epithelial tissues, including the airways [21,22]. Since Ca 2+ -dependent Clchannels, or at least a major component of it, have now been identified, we examined the functional relationship between TMEM16A and CFTR and aimed to identify the mechanism by which CFTR may affect TMEM16A. Our experiments suggest that CFTR controls Ca 2+ dependent Clsecretion by inhibiting TMEM16A.

Materials and Methods
Cell culture, RT-PCR CFBE41o-, 16HBE, and CFBE41o-cells stably expressing wild-type CFTR [23] were generous gifts from Dr. D. Gruenert (Univ. California San Francisco, USA) and Dr. J.P. Clancy (Univ. of Alabama at Birmingham, USA) and were grown in Minimum Essential Medium supplemented with 2 mM glutamine. Human embryonic kidney (HEK293) cells were grown in Dulbecco's modified Eagle's medium. All media were supplemented 10% fetal calf serum. Cells were incubated in 5% CO 2 at 37°C. Expression of mRNAs encoding all 10 human TMEM16 proteins was examined in CFBE41o-cells by standard RT-PCR using primers as described in [24]. cDNAs, siRNA, real time RT-PCR and transfection cDNA for human TMEM16A, F, and J were amplified from total RNA of 16HBE and CFBE cells by RT-PCR, as recently described [25]. In brief: Real-time PCR was performed in a Light Cycler (Roche), using the Quanti Tect SYBR Green PCR Kit (Qiagen, Hilden, Germany). Each reaction contained 2 µl Master Mix (including Taq polymerase, DNTPs, SYBR green buffer), 1 pM of each primer sense and anti-sense, 2.5 mM MgCl 2 , and 2 µl cDNA. After activation of the Taq polymerase for 10 min at 94° cDNA was amplified by 15 s at 94°C, 10 s at 55°C, and 20 s at 72°C, for 50 cycles. The amplification was followed by a melting curve analysis to control of the PCR products. The primer sequences were: hTMEM16A: (NM_018043. Analysis of the data was performed using Light Cycler software 3.5.3. Standard curves for TMEMs mRNA and β-actin mRNA were produced by using cDNA of transfected HEK cells at different dilutions. The ratio of the amount of TMEM to β-actin mRNA was calculated for each sample and analysis was performed in triplicates. TMEM16A-K610A was produced by PCR-based site-directed mutagenesis. Human TMEM16H was purchased from imaGenes (Clone:IMAGp998FF0411661Q, Berlin, Germany). The cDNAs were subcloned into pcDNA3.1 V5-His (Invitrogen, Karlsruhe, Germany).

Patch Clamping
Cells grown on cover slips were mounted in a perfused bath on the stage of an inverted microscope (IM35, Zeiss) and kept at 37°C The bath was perfused continuously with Ringer solution (mM: NaCl 145, KH 2 PO 4 0.4, K 2 HPO 4 1.6, d-glucose 6, MgCl 2 1, Ca-gluconate 1.3, pH 7.4) at about 10 ml/min. NaCl was replaced by NaBr, NaI or NaSCN. Patch-clamp experiments were performed in the fast whole-cell configuration. Patch pipettes had an input resistance of 4-6 MΩ, when filled with an intracellular like solution containing (mM) KCl 30, K-gluconate 95, NaH 2 PO 4 1.2, Na 2 HPO 4 4.8, EGTA 1, Ca-gluconate 0.758, MgCl 2 1.034, D-glucose 5, ATP 3. pH was 7.2, the Ca 2+ activity was 0.1 µM. The access conductance was measured continuously and was 90-140 nS (EPC 9 amplifier, List Medical Electronics, Darmstadt, Germany). In regular intervals, membrane voltages (V c ) were clamped in steps of 10 mV from -100 to +100 mV and the membrane conductance G m was calculated from the measured current (I) and V c values according to Ohm's law.

Immunohistochemistry
GFP tagged TMEM16A and CFTR-Cherry coexpressing HEK293 cells were grown on glass cover slips. Nuclei were stained with Hoe33342 (0.1 µg/ml Ringer, Aplichem, Darmstadt, Germany) and examined as life images with an ApoTome Axiovert 200M fluorescence microscope (Zeiss, Göttingen, Germany). His-tagged TMEM16J and CFTR-Cherry transfected HEK293 cells were grown on glass cover slips and fixed for 10 min with 4% (w/v) paraformaldehyde at room temperature. Cells were permeabilized and blocked with 2% (w/v PBS) BSA and 0.04% (v/v PBS) Triton X-100 and incubated for 1 h with primary antibody mouse anti His tag (1:500, Qiagen, Hilden, Germany) at 37°C. Binding of the primary antibody was visualized by incubation with a secondary donkey anti-mouse antibodies conjugated with AlexaFluor®488 (1:1.000, Molecular Probes, Invitrogen). Nuclei were stained with Hoe33342 (0.1 µg/ml PBS, Aplichem, Darmstadt, Germany). Cells were mounted on glass slides with fluorescent mounting medium (DAKO Cytomation, Hamburg, Germany) and examined with an ApoTome Axiovert 200M fluorescence microscope (Zeiss, Göttingen, Germany). As a measure of co-localization between the signals in the red and green channels of confocal images, we calculated the Spearman's rank correlation coefficient using ImageJ PlugIn PSC-colocalization after background subtraction Immunoprecipitation and Western blotting Protein was isolated from CFBE or transfected HEK293 cells in lysis buffer containing 10mM HEPES, 2mM EDTA, 1% TritonX-100, 1% proteinase inhibitor cocktail and 5U/ml Benzonase. Prior to the addition of mouse monoclonal CFTR antibody (no. 570; CF Foundation Therapeutics) or GFPantibody (SIGMA, Taufkirchen, Germany), protein lysates were precleared with protein G agarose beads (Pierce, Rockford, USA). Incubation of the precleared protein lysates with antibodies was performed overnight at 4°C, and the protein-antibody complex was immobilized with the addition of 20 µl of 50% slurry of protein G agarose beads for 2.5 h at 4°C. The beads were washed five times in wash buffer, and after the last washing step, beads were mixed in 2x Laemmli sample buffer for Western blot analysis. For Western blot analysis, the 10% SDS polyacrylamide gel was transferred to a polyvinylidene difluoride membrane (GE Healthcare EuropeGmbH, Munich, Germany) using wet transfer (BioRad, Munich, Germany). Membranes were incubated with anti-GFP rabbit antibodies (1:1,000, Molecular probe, Invitrogen) overnight at 4°C. Proteins were visualized using an anti-rabbit horseradish peroxidase (HRP)-conjugated antibody (1:10,000, Acaris) and blots were visualized using chemiliminescences (Pierce).

Measurement of the intracellular Ca 2+ concentration
Transfected HEK293 cells were loaded with 5 µM Fura2-AM in Ringer solution at 37°C for 2 h. Fluorescence was detected at 37°C, using an inverted microscope IMT-2 (Olympus, Nürnberg, Germany) and a high speed polychromator system (VisiChrome, Puchheim, Germany). FFP-18 was excited at 340/380 nm, and emission was recorded between 470 and 550 nm using a CCD camera (CoolSnap HQ, Visitron). The results were obtained at 340/380 nm fluorescence ratio (after background subtraction).

Materials and statistical analysis
All compounds used were of high grade of purity and were from Sigma (Taufkirchen, Germany) or Merck (Darmstadt, Germany). The siRNAs were from Invitrogen and Santa Cruz Biotechnology, respectively). All cell culture reagents were from Invitrogen. Student's t-test (for paired or unpaired samples as appropriate), and analysis of variance (ANOVA) was used for statistical analysis. P values <0.05 were accepted as significant.

CFTR inhibits TMEM16A/CaCC in human airways
Airway epithelial cell lines such as 16HBE14o- [27] and CFBE41o- [28] express endogenous CaCC. Similar to earlier experiments in Xenopus oocytes and mammalian cells [2][3][4], we found inhibition of CaCC after stimulation of CFTR in 16HBE14o-cells, which express both proteins endogenously [27,29] (Fig. 1). To analyze the correlation between CaCC and CFTR in more detail, we examined expression of TMEM16-mRNA in CFBE/ CFTR airway epithelial cells, which stably express CFTR as well as TMEM16A, F, H and J ( Fig. 2A). In fact, a small but significant CaCC could be activated by purinergic stimulation with extracellular ATP (100 µM) (Fig. 2B). CFBE (parental) cells do not express wtCFTR and therefore do not generate Clcurrents when stimulated with IBMX and forskolin (I/F) (Fig. 2B-D 2D-F). Thus expression of CFTR appears to inhibit TMEM16A, although expression of TMEM16-proteins was not changed (data not shown). We also found that siRNA-suppression of TMEM16A reduced ATPactivated whole cell currents from 28±3.8 to 5.3±1.1 nS (n = 5), indicating that TMEM16A is essential for CaCC in human airway epithelial cells. Conversely, siRNA knockdown of TMEM16-proteins in CFBE/CFTR cells by two independent siRNAs did not change CFTR conductance (Fig. 2G,H)     ( Fig. 2H). Taken together, our results show that i) TMEM16A forms CaCC in human airway epithelial cells, and ii) is inhibited CFTR.

CFTR-currents are largely independent of TMEM16-proteins
Additional experiments were performed in HEK293 cells transiently overexpressing wtCFTR. HEK293 cells express TMEM16F,H,J, and traces of TMEM16A and thus show very little CaCC [24] (Fig. 3A). Using two independent siRNAs, we knocked down expression of TMEM16-proteins individually (Fig. 3B), and measured CFTR whole cell conductances after stimulation with IBMX and forskolin. For TMEM16A, successful knockdown was verified by Western blotting. Because antibodies were not available for other TMEM16 proteins, knockdown was also demonstrated by real time RT-PCR [24] (Fig. 3B). siRNA did not affect CFTRconductances except of siRNA for TMEM16J, which significantly increased the baseline conductance in HEK293 cells, and inhibited the IBMX/forskolin activated fraction, a result that is currently unexplained (Fig. 3C). However, the halide permeability sequence of I/F-activated CFTR currents remained essentially Ousingsawat/Kongsuphol/Schreiber/Kunzelmann  unchanged by siRNA-knockdown of the TMEM16 isoforms (Fig. 3D). Moreover, activation of CFTR and the halide-permeability sequence of CFTR-currents were not affected by coexpression of TMEM16-proteins together with CFTR. These results suggest that CFTRcurrents are largely independent of TMEM16-proteins (Fig. 3E).
Because receptor mediated Ca 2+ signaling and activation of TMEM16A might be compromised during stimulation and activation of CFTR, we also examined whether receptor-independent activation of TMEM16A, by increase of Ca 2+ through the Ca 2+ ionophore ionomycin, was also affected by CFTR (Fig. 5B,C). In fact TMEM16A could no longer be activated by ionomycin after activation of CFTR. Notably, activation of whole cell Clconductance by ATP (29.2±3.1, +I/F vs. 28.8±2.2, -I/F; n = 8) nS; or ionomycin (25.3±2.6, +I/F vs. 26.1±2.6, -I/F; n = 7) nS; were independent of I/F in cells expressing TMEM16A only. Interestingly, cells overexpressing TMEM16A show a higher baseline conductance that was inhibited by 20 µM of the CaCC inhibitor AO1 [30] or niflumic acid (NFA; 10 µm), but not by CFTR inh 172, indicating partially active TMEM16A even at baseline [Ca 2+ ] i levels (Fig. 6A). Stimulation by ATP or I/F only slightly further increased G (Fig. 6B). Notably, CFTRconductance (G I/F ) in HEK293 cells was attenuated by coexpression of TMEM16A (Fig. 6C). Taken together these data indicate an inverse relationship between CFTR and TMEM16A in HEK293 cells.

Membrane localized CFTR and TMEM16A may physically interact
The results suggest that CFTR and TMEM16A functionally interact. Both CFTR and TMEM16A are clearly membrane localized, which could allow for direct protein interaction (Fig. 7A) [18,25]. Colocalization of both proteins was suggested by the high Spearman rank correlation coefficient of 0.83±0.13 (n = 11; c.f. methods). We found that TMEM16A was coimmunoprecipitated with CFTR in HEK293 cells, which provides some evidence for a direct physical interaction of both proteins (Fig. 7B). However, CFTR and TMEM16F did not coimmunoprecipitate when expressed in HEK293 cells, providing hints for a specific interaction of CFTR and TMEM16A (Fig. 7C).

cAMP-dependent stimulation affects intracellular Ca 2+ signaling in HEK293 cells
Because CFTR may also affect activation of TMEM16A by changing intracellular Ca 2+ signaling, we measured ATP-and ionomycin-induced increase of intracellular peak and plateau [Ca 2+ ] using Fura2 (Fig.  8). We were unable to detect an effect of IBMX/forskolin (I/F) on ATP-induced [Ca 2+ ] i increase in cells coexpressing CFTR and TMEM16A (Fig. 8A-C). In Fig. 7. Membrane localized CFTR and TMEM16A may physically interact: A) Immunocytochemistry (life imaging) of CFTR-Cherry (red; Cherry) and GFP tagged TMEM16A (GFP, green) when transiently overexpressed in HEK293 cells. Both CFTR and TMEM16A are clearly membrane localized. B) Coimmunoprecipitation of CFTR and TMEM16A-GFP. TMEM16A-GFP was detected using a GFP-antibody (input). Immunoprecipitation of beats only (no conjugated antibody) served as a control. TMEM16A-GFP was coimmunoprecipitated after pull-down of CFTR using CFTR-AB 596 (right panel), while CFTR was not coimmunoprecipitated with TMEM16A-GFP (left panel). C) Coimmunoprecipitation of CFTR and TMEM16F-GFP expressed in HEK293 cells. TMEM16F-GFP was not coimmunoprecipitated after pull-down of CFTR (right panel), while CFTR was not coimmunoprecipitated with TMEM16F-GFP (left panel). Input represents the loading control, i.e. amount of protein used for immunoprecipitation. Ousingsawat/Kongsuphol/Schreiber/Kunzelmann contrast ionomycin induced peak [Ca 2+ ] i increase was significantly reduced in cells expressing CFTR or TMEM16A (Fig. 8D). Inhibition of [Ca 2+ ] i increase during activation of TMEM16A or CFTR is possibly due to depolarization of the membrane voltage [31]. Activation of CFTR may thereby reduce the driving force for Ca 2+ influx and attenuate activation of TMEM16A [31]. Taken together, the present results indicate that CFTR and TMEM16A are functionally related and maybe physically interact with each other.

Discussion
Enhanced CaCC in CF epithelial cells and inhibition of CaCC due to acute stimulation of CFTR implies that both Clchannels are somehow related [2][3][4]. Previously, a more detailed examination of this relationship was hampered by the fact that the molecular nature of CaCC was unknown. With identification of TMEM16A as a crucial component of CaCC [18][19][20] and demonstration of its intrinsic role for Ca 2+ dependent Cltransport [21,22], we are now able to re-examine this relationship. Interestingly, and similar to Namkung et al [32], we also found that siRNA-knockdown of TMEM16A in HT 29 colonic carcinoma cells suppressed only 20 % of CaCC. In contrast, siRNA-knockdown of TMEM16A in 16HBE, CFBE (airways), CFPAC (pancreas), MDCK (kidney), and Cal-33 (stroma) cells almost completely abolished Ca 2+ activated Clconductance. Therefore we assume that TMEM16A plays the dominating role for epithelial Ca 2+ activated Clchannels in most, but not all epithelial tissues. The present data now clearly show that Clcurrents generated by TMEM16A and activated by receptor stimulation or by Ca 2+ ionophores, are inhibited by CFTR. We found that similar to purinergic activation, also muscarinic activation of TMEM16A was inhibited during stimulation of CFTR (data not shown). The present data COIP-data supply some evidence for a direct molecular interaction of CFTR and TMEM16A.
The data obtained for ionophore-mediated activation of TMEM16A, also point to a functional interaction of Cell Physiol Biochem 2011;28:715-724 CFTR and TMEM16A. Thus ionophore-induced increase in [Ca 2+ ] i was significantly attenuated in the presence of CFTR and after stimulation with I/F, which may impede activation of TMEM16A (Fig. 8). Along this line it is interesting to note that in colonic epithelial cells the contribution of Ca 2+ activated SK4 K + channels to total cellular K + conductance is high under control (resting) conditions, but is reduced after cAMP-dependent stimulation of the cells [33]. This is due to the cellular depolarization, causing a reduced driving force for Ca 2+ influx and attenuated activation of TMEM16A, after cAMP-dependent stimulation. Surprisingly, attenuated increase in [Ca 2+ ] i after stimulation with IBMX/forskolin was also observed in cells expressing TMEM16A only. However, these attenuated Ca 2+ signals did not lower TMEM16A currents (data not shown). It is therefore uncertain whether attenuated [Ca 2+ ] i signals, as measured by Fura-2, really lead to attenuated TMEM16A-currents. We found earlier that local compartmentalized intercellular Ca 2+ concentrations control Cland K + channels, rather that global cytosolic Ca 2+ levels [34].
The team by Nilus and coworkers concluded that the C-terminal part of the R-domain, but not the PDZ binding motif of CFTR is involved in the interaction with endogenous Ca 2+ -activated Clchannels in endothelial cells [35]. We confirmed this result by expressing a PDZ binding motif-mutant in Xenopus oocytes, which was still able to inhibit CaCC (data not shown). It is known that CFTR interacts with numerous proteins, many of them are receptors or have a function in cellular signaling [36]. It is entirely possible that crosstalk of CFTR with these signaling proteins contributes to inhibition of TMEM16A by CFTR.