Introduction

Calcium activated chloride channels (CaCCs) are anionic channels with both calcium and voltage dependence. CaCCs play a variety of physiological roles in many organs and tissues, including sensory transmission, regulation of neuronal and cardiac excitability, regulation of vascular tone and transepithelial Cl− secretion (see [14] for review). The molecular identity of CaCCs remained controversial until recent years when two families of proteins were shown to function as CaCC or at least to be major components of CaCC: bestrophins and TMEM16 (anoctamin) proteins [4, 19, 20, 43, 36, 2].

Bestrophins are generally believed to form CaCC channels [27, 2, 34] and/or to regulate ion channels [35, 44]. Bestrophin-1 (Best1) is highly expressed in human basolateral membrane of retinal pigment epithelial (RPE) cells and may form the CaCC channels there [24, 37, 15]. Also in a population of sensory neurons, bestrophin has been suggested to form CaCC [2]. Moreover, hBest1 was recently found to be involved in Ca2+ handling in endoplasmic reticulum stores [10, 46, 1]. TMEM16-proteins form a family of transmembrane proteins with ten members some of which act as Cl− channels. Among the TMEM16 family, TMEM16A (ANO1) has been the most extensively studied; it underlies the CaCC in epithelial tissue [14, 43], neurons [18, 22, 5] and smooth muscle [23, 38]. When expressed in the expression systems, both TMEM16A and Best1 show the characteristics of endogenous CaCC: Cl− currents with Ca2+ and voltage dependency [9, 16, 4, 36]. Both channels are sensitive to broad-spectrum Cl− channel blockers [43, 34]. The facts that (1) both TMEM16A and Best1 form functional CaCCs in expression systems; (2) both proteins are often expressed in tissues displaying native CaCC currents; (3) both channels are sensitive to conventional Cl− channel blockers make it difficult to reveal molecular identity of endogenous CaCCs.

The broad-spectrum Cl− channel blockers used in CaCC studies include niflumic acid (NFA), 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS), 5-nitro- (3-phenylpropylamino)-benzoic acid (NPPB) and flufenamic acid (FFA). Numerous studies report that these compounds block endogenous CaCCs in Xenopus laevis oocytes [30, 33, 40, 41]. However, these blockers are known to be non-specific and also block other channels. For example, NFA and DIDS also block volume-regulated anion channel (VRAC) in some cell types [42, 11, 29]. NFA and FFA are shown to active Ca2+-dependent K+ channels [12, 31] whereas NFA and DIDS are found to block Kv4 channels [39] and NPPB has a blocking effect on endogenous K+ currents in Xenopus oocytes [30, 29]. NFA, FFA and NPPB also cause an increase in intracellular Ca2+ concentration ([Ca2+]i) in several cell types, which could elicit other cellular responses [30, 29]. In addition, NFA and FFA are described to modulate glutamate and glycine transporters [13]. DIDS has also been shown to increase TRPV1 current induced by capsaicin [45].

Tannic acid [26], CaCCinh-A01 [6] and T16Ainh-A01 [25] are recently reported to inhibit CaCCs. Tannic acid which is present in green tea and red wine is identified as an inhibitor of TMEM16A and TMEM16B, but has little effect on CFTR Cl− conductance or on ENaC Na+ conductance [26]. CaCCinh-A01 and T16Ainh-A01 are found to inhibit TMEM16A and native CaCC. CaCCinh-A01 is shown to target human intestinal CaCC channel directly rather than upstream signaling mechanisms, and has no effects on intracellular Ca2+ and Ca2+-calmodulin kinase II or CFTR [6]. T16Ainh-A01 is an aminophenylthiazole which potently inhibits TMEM16A directly without interfering with upstream processes such as agonist binding or Ca2+ signaling [25].

Up to now, a systematic comparison study of the effects of these blockers on TMEM16A currents is lacking, and more importantly many of these compounds have not been tested for their effect on Best1 currents. In this study, we compared the efficacy and potency of seven described CaCC inhibitors (Fig. 1) on TMEM16A and Best1 currents. Profiling of different modulators of TMEM16A and Best1 will benefit our use of these modulators in studying CaCC and promote development of more specific modulators of CaCC.

Fig. 1
figure 1

Chemical structures of CaCC inhibitors used in this study

Full size image

Materials and methods

Cell culture and channel stable expression

The mouse TMEM16A cDNA clone was kindly provided by Prof. Uhtaek Oh (Seoul National University, Korea) and was subcloned to expression vector pEGFPN1. The human Best1 cDNA clone was kindly provided by Prof. Jeremy Nathans (Johns Hopkins Medical School, Baltimore,MD, USA) and was subcloned to expression vector pRK5. The human cDNA clone TMEM16A (abc) was kindly provided by Prof. Luis Galietta (Giannina Gaslini Institute, Italy) and was subcloned to expression vector pcDNA3.1. Chinese hamster ovary (CHO) cells were cultured in F-12 K with 10 % fetal calf serum and 1 % penicillin/streptomycin at 37 °C in a humidified atmosphere of 5 % CO2, and were passaged every 2 days. CHO cells were transfected with above channel cDNA constructs with Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer’s instructions. Twenty-four hours after transfection the cells were trypsinised and plated to 9-cm Petri dishes in media supplemented with the selection factor G418 (1,000 μg/ml) for 14 days with the medium being replaced every 3–5 days. The colonies were selected and transferred to a 24-well plate, cultured and the CHO cells stably expressing TMEM16A or Best1 were identified by Western blots and patch clamp technique. Stably transfected CHO cells were cultured in F-12 K supplemented with 10 % fetal calf serum, 600 μg/ml G418 and 1 % penicillin/streptomycin at 37 °C, 5 % CO2, and were removed from the culture flask by a 2-min digestion with 2.5 mg/ml trypsin (1:250) and plated at low density onto 13-mm-diameter glass coverslips in 24-well tissue culture plates for patch-clamp analysis. The cells were used for recording within 48 h.

Western blots analysis

Control and stably transfected CHO cells were harvested in lysis buffer (20 mM Tris-base, 137 mM NaCl, 10 % glycerol, 1 % Triton-X-100, 2 mM EDTA, and 1 μl/ml protease inhibitor) and lysates were centrifuged at 17,200 × g for 30 min; pellets were discarded. And supernatants were further centrifuged at 100,000 × g for an hour to precipitate membrane proteins. Protein samples were heated to 70 °C for 10 min in sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE) loading buffer and separated on a 10 % polyacrylamide gel. The separated proteins were transferred at 30 V to nitrocellulose membrane (Millipore, USA) overnight at 4 °C. After blocking the membrane in TBS containing 5 % non-fat milk, the blots were incubated with primary antibodies at RT for 4–6 h. Antibody dilutions were as follows: polyclonal anti-TMEM16A, 1:500 (rabbit; Aviva System Biology, USA); polyclonal anti-bestrophin, 1:500 (rabbit; Abcam, UK); monoclonal anti-β-tubulin III, 1:500 (mouse; Sigma, USA); monoclonal anti-GAPDH 1:500 (mouse; Abcam, UK). After incubation with primary antibodies, membranes were rinsed with TBST (150 mM NaCl, 20 mM Tris-base, 0.05 % Tween-20) three times for 10 min and then incubated with secondary antibodies conjugated with IRDye700DX and IRDye800CW (1:4,000 and 1:3,000, respectively; Rockland, USA) at room temperature for 1–2 h. Membranes were washed twice with TBST for 2 × 10 min and once with TBS for 10 min. Protein bands were detected and quantified on an Odyssey two-color infrared imaging system (LI-COR Biosciences, USA).

Electrophysiology

TMEM16A and Best1 currents in CHO cells were recorded using whole-cell recording technique with an EPC 10 amplifier (HEKA Electronic, Germany) and PULSE software (HEKA). The acquisition rate was 10 kHz and signals were filtered at 2.5 kHz. Patch electrodes were pulled with a horizontal micropipette puller (P-97, Sutter Instruments, USA) and fire polished. The recording electrodes had a resistance of 2–4 MΩ when filled with internal solution of composition (mM): 130 CsCl, 10 EGTA, 1 MgCl2, 10 HEPES and different concentrations of CaCl2 to obtain the desired free Ca2+ concentration (1 mM for 12 nM, 6 mM for 166 nM and 8 mM for 447 nM). On a day of the recording the intracellular solution was supplemented with ATP (to final concentration of 2 mM) and pH was adjusted to 7.3 with CsOH. Free Ca2+ concentration was calculated using the Webmaxc software (Stanford; http://www.stanford.edu/~cpatton/webmaxc/webmaxcS.htm). The bath solution contained (in mM): 140 NMDG, 1 CaCl2, 1 MgCl2, 10 glucose, 10 HEPES and pH 7.4 adjusted with HCl. All recordings were performed at room temperature.

Chemicals

NFA, DIDS, NPPB, T16Ainh-A01, tannic acid and FFA were purchased from Sigma (St Louis, MO, USA). Digallic acid was purchased from TRC (Toronto, Canada). CaCCinh-A01 was synthesized in-house. The purity of CaCCinh-A01 (>99.9 %) was determined by infrared nuclear MR and high-resolution mass spectrometry. These compounds were dissolved in DMSO to yield stock solutions of 100 mM. The stock solutions were kept at −20 °C. Stock solutions were diluted to final concentrations in bath solution. The vehicle at the final concentration did not affect the currents measured. All drug solutions were freshly made before each experiment and kept away from light.

Data analysis and statistics

Results were expressed as means ± SEM. Student’s test (unpaired) was used to assess statistical significance. P < 0.05 was considered significant. Concentration–response curves were fitted with the logistic equation: y = A 2 + (A 1 − A2)/(1 + (x/x 0)p); where y is the response; A 1 and A 2 are the maximum and minimum response, respectively; x is the drug concentration; and p is the Hill coefficient. One-way ANOVA test was used to assess statistical comparison of IC50s data. To analyze the kinetics of TMEM16A current deactivation, the tail currents deactivating from +80 to –100 mV were used to obtain the time constants. The deactivation traces were fitted to a single exponential function: I = A × [1 − exp(−t/τ)] + I 0, where I is the current, I 0 is the steady-state amplitude of the current at the end of the holding potential of −100 mV, A is the difference between the peak and steady-state current amplitudes, t is time, and τ is the time constant.

Results

Characterization of TMEM16A and bestrophin-1 currents stably expressed in CHO cells

All experiments were performed using CHO cells that were stably transfected with mouse TMEM16A and human Best1 (see Materials and methods). Figure 2a and b shows the Western blot results of TMEM16A and Best1 protein expression. Bands corresponding to TMEM16A and Best1 were detected in both transfected and control CHO cultures, indicating possible endogenous expression of TMEM16A and Best1 in these cells. Overexpression of TMEM16A and Best1 greatly increased corresponding protein expression level (Fig. 2a and b, upper panels). However, Western blot analysis using the membrane protein indicated that levels of both TMEM16A and Best1 at the cell membranes of un-transfected CHO cells were barely detectable while strong bands were detected in TMEM16A- and Best1-transfected cells (Fig. 2a and b, bottom panels). TMEM16A and Best1 currents were recorded by using the whole-cell patch-clamp technique. The cells were clamped from the holding potential of 0 mV to voltages between −100 and +100 mV in 20 mV steps followed by a step to −100 mV. TMEM16A currents were first characterized. CHO cells with stable expression of TMEM16A manifested membrane currents with typical voltage and intracellular Ca2+ dependence, and slow activation at high depolarizing voltages (Fig. 2c and d) [36]. The steady-state current–voltage relationship showed a strong outward rectification in a 447 nM free intracellular Ca2+ concentration (Fig. 2c and d). For a comparison, Best1 currents were recorded in same conditions. The CHO cells with stable expression of Best1 also presented currents with voltage and Ca2+ dependence similar to TMEM16A currents, and outward rectification. However, the activation kinetics of Best1 currents was much faster than that of TMEM16A currents (Fig. 2e and f). No significant currents were detected in CHO cells expressing Best1 or TMEM16A when intracellular Ca2+ concentration was low (12 nM). The untransfected CHO cells showed little currents even with high (447 nM) free intracellular Ca2+ (Fig. 2g and h). The properties of TMEM16A and Best1 currents recorded in this study were consistent with these reported before [36, 44].

Fig. 2
figure 2

Characterization of TMEM16A and Bestrophin-1 currents in stably-transfected CHO cells. a, b Western blots results of TMEM16A and Best1 protein levels in control or stably transfected CHO cell lines. Upper panels: Western blots results of TMEM16A and Best1 total protein levels in control or stably transfected CHO cell lines. Bottom panels: Western blots results of TMEM16A and Best1 membrane protein levels in control or stably transfected CHO cell lines. c Whole-cell currents recorded from CHO cells expressing TMEM16A at the indicated intracellular free Ca2+ concentrations. The currents were elicited by voltage pulses from a holding potential of 0 mV to voltages between −100 and +100 mV in 20-mV steps followed by a step to −100 mV as indicated at the bottom of panel g. d Mean current–voltage relationships of TMEM16A at different intracellular Ca2+ concentrations; n = 8–15. e Whole-cell currents recorded from CHO cell transfected with Bestrophin-1 (Best1). The currents were recorded using the same protocol as used with TMEM16A. f Mean current–voltage relationships for the experiments shown in panel e; n = 6–15. g Whole-cell currents recorded from untransfected CHO cell with 447 nM intracellular Ca2+. h Mean current–voltage relationships for the experiments shown in panel g; n = 5

Full size image

Effects of Cl− channel inhibitors on TMEM16A currents expressed in CHO cells

We first investigated six Cl− channel inhibitors for their effect on TMEM16A currents. These inhibitors were applied at increasing concentrations from 0.1 to 3,000 μM. A voltage ramp from −100 to +100 mV in 5 s was used to record the whole-cell TMEM16A currents (Fig. 3a, right top). The effects of six blockers were quantified by measuring the changes in the current amplitude at both +80 and −80 mV induced by these drugs. Figure 3a–f (left panels) shows the time courses of concentration-dependent inhibition of the TMEM16A currents by the inhibitors recorded at −80 mV (below the dotted zero current line) and +80 mV (above the dotted zero current line); the TMEM16A currents were activated with a free intracellular Ca2+ concentration of 447 nM. The concentration dependencies of the effects of these blockers are shown in Fig. 3g and h. Among these blockers we tested, CaCCinh-A01 and NFA showed the highest potency in inhibiting TMEM16A currents at +80 mV, with half maximal inhibition (IC50) of 7.84 ± 0.62 and 7.40 ± 0.95 μM, respectively. Both inhibitors also showed high efficacy and at the maximal tested concentration (300 μM), inhibited the TMEM16A currents recorded at +80 mV by 90.67 ± 1.15 % and 93.65 ± 0.79 %, respectively. On the other hand, DIDS was much less potent in inhibiting TMEM16A currents, with an IC50 of 548.86 ± 25.57 μM. In summary, the six tested blockers inhibited TMEM16A currents with IC50 values in the following order: NFA (7.40 ± 0.95 μM, n = 4 − 16) ~ CaCCinh-A01 (7.84 ± 0.62 μM, n = 4 − 15) > FFA (14.21 ± 1.92 μM, n = 4–12) ~ tannic acid (25.68 ± 1.14 μM, n = 6 − 11) > NPPB (64.14 ± 1.56 μM, n = 4–16) > DIDS (548.86 ± 25.57 μM, n = 4–16) (Fig. 3, Table 1). Despite its low potency, DIDS almost completely inhibited the TMEM16A currents (93.01 ± 1.15 %) at the maximal tested concentration. In general, all blockers tested significantly (~80–95 %) inhibited TMEM16A currents at the maximal tested concentrations.

Fig. 3
figure 3

Effects of Cl− channel inhibitors on TMEM16A currents. Whole-cell currents of TMEM16A were induced by a voltage ramp from −100 to +100 mV. a–f Left panels: time courses of concentration-dependent modulation of the TMEM16A currents recorded at +80 mV (upper line, black squares) and −80 mV (lower line, open circles) with a free intracellular Ca2+ concentration of 447 nM. The dotted lines indicate the zero current level. The current amplitudes were measured every 5 or 10 s. Right panels: the representative current traces recorded when the effect of inhibitors has stabilized. The gray traces are control currents in the absence of modulators. g Concentration–response relationships for different modulators on TMEM16A currents recorded at +80 mV. Data were fitted with logistic function. The following IC50 values were obtained: FFA, 14.21 ± 1.92 μM (n = 4–12); CaCCinh-A01, 7.84 ± 0.62 μM (n = 4–15); DIDS, 548.86 ± 25.57 μM (n = 4–16); tannic acid, 25.68 ± 1.14 μM (n = 6–11); NPPB, 64.14 ± 1.56 μM (n = 4–16); NFA, 7.40 ± 0.95 μM (n = 4–16). h Concentration–response relationships for CaCCinh-A01 (18.39 ± 2.34 μM, n = 4–15), DIDS (699.37 ± 82.81 μM, n = 4–16) and tannic acid (29.05 ± 4.56 μM, n = 6–11) on the inward TMEM16A currents recorded at −80 mV. i Summary data for the effects of NPPB, FFA and NFA on the inward TMEM16A currents recorded at −80 mV. *P < 0.05, **P < 0.01 compared with basal currents in the absence of inhibitors recorded at +80 mV. # P < 0.05, compared with basal currents in the absence of inhibitors recorded at −80 mV

Full size image
Table 1 Effects of different inhibitors on TMEM16A and Best1 currents measured at +80 mV
Full size table

When we studied the modulation of the inward TMEM16A currents by the inhibitors, it gave a different picture. Three of the six inhibitors tested, NFA, NPPB and FFA, had dual effects on the inward TMEM16A currents recorded at −80 mV (Fig. 3i): they increased the inward currents at lower concentrations and inhibited the currents at higher concentrations. NFA (Fig. 3f, right, dotted line) and FFA (Fig. 3a, right, dotted line) demonstrated a maximal activation of TMEM16A at 10 μM, which increased the inward currents by 119.4 ± 27.7 % and 30.58 ± 15.99 %, respectively. NPPB at the maximal activation concentration (30 μM) (Fig. 3e, right, dotted line) increased the inward currents by 51.8 ± 10.4 %. Moreover, a transient increase of inward currents was observed before the inhibition was relieved when the inhibitors were washed out (Fig. 3a, e and f). On the other hand, the other three inhibitors, CaCCinh-A01, tannic acid and DIDS only showed concentration-dependent inhibition of the inward TMEM16A currents recorded at −80 mV. However, these were less potent in inhibiting the inward currents than inhibiting the outward currents recorded at + 80 mV, with IC50s at −80 mV of 18.39 ± 2.34, 29.05 ± 4.56 and 699.37 ± 82.81 μM, respectively. In addition, these inhibitors were less efficacious in inhibiting inward current with maximal inhibitions of 84.49 ± 2.88 %, 78.7 ± 4.47 % and 73.03 ± 4.46 %, respectively. The effects of the six inhibitors on the TMEM16A inward current are summarized in Fig. 3i.

Ramp protocol does not allow estimating the effect of the drugs on current kinetics. To investigate effects of the drugs on the TMEM16A channel kinetics and also to test effects on currents in steady-state conditions step protocol was used (from a holding potential of −100 mV to voltages between −100 mV and +100 mV in 20-mV steps followed by a step to −100 mV). For this, we tested the effects of six inhibitors with concentrations of their respective IC50 obtained using the voltage ramp protocol (Fig. 3 and Table 1). Not surprisingly, these inhibitors at these concentrations inhibited steady-state TMEM16A currents at positive voltages by ~50 % (Fig. 4). Interestingly, we noted that the deactivation of the TMEM16A tail currents at −100 mV was significantly slowed by all modulators but CaCCinh-A01 (Fig. 4a–c). A detailed analysis of the time constants of the deactivating tail currents showed that effects of NPPB, FFA and NFA were most significant (Fig. 4m). As we noted above, NPPB, FFA and NFA potentiated the inward TMEM16A currents, and this slowing of channel deactivation could be a related mechanism.

Fig. 4
figure 4

Effects of Cl− channel inhibitors on steady-state TMEM16A currents and channel deactivation. a–f The cells were clamped from the holding potential of −100 mV to voltages between −100 and +100 mV in 20-mV steps followed by a step to −100 mV (voltage protocol is depicted above traces in a). The dotted lines indicate the zero current level. Left panels: whole cell currents recorded from CHO cells transfected with TMEM16A with a free intracellular Ca2+ concentration of 447 nM. Middle panels: TMEM16A currents when the indicated inhibitors were applied. Right panels: effects of CaCC blockers on the deactivation kinetics of TMEM16A currents from +80 to −100 mV. The gray line shows the deactivating currents in the presence of the blockers, which were scaled up to match the amplitude of the deactivating currents in the absence of the blockers. g–l Current–voltage relationships of TMEM16A currents in the absence or presence of the blockers; n = 4–6. m Summary of effects of CaCC blockers on the time constants of TMEM16A deactivating currents. **P < 0.01 compared with control in the absence of blockers; n = 4–7

Full size image

We next investigated the effect of T16Ainh-A01, a newly developed inhibitor of CaCCs [25], on TMEM16A currents. When the effect of T16Ainh-A01 was tested against TMEM16A currents recorded using the voltage ramp protocol, to our surprise, T16Ainh-A01 only showed modest inhibition of TMEM16A currents. The maximal inhibition was observed at 3 μM, with an inhibition of the outward and inward currents of 28.74 ± 2.19 % and 27.55 ± 3.45 % at +80 and −80 mV, respectively. It did not further inhibit the currents when the concentration was increased up to 30 μM (Fig. 5a–c). When TMEM16A currents were recorded using the step voltage protocol, similar results were observed (Fig. 5d–f). T16Ainh-A01 also did not alter the deactivation of TMEM16A currents (Fig. 5g and h). TMEM16A channel exist in several splice variants. The mouse TMEM16A used in this study does not contain segment of amino acids coded by exon 6b. To test whether the absence of exon 6b in mTMEM16A is responsible for the low sensitivity to T16Ainh-A01, we investigated the effect of T16Ainh-A01 on hTMEM16A (abc) transiently expressed in CHO cells. T16Ainh-A01 (10 μM) inhibited hTMEM16A currents measured at +80 mV by about 50 % (Fig. 5i), which is larger than the inhibition of mTMEM16A currents but still somewhat less than previously reported for hTMEM16A ac or abc isoforms [3, 25]. Finally, we also tested the effect of digallic acid, which is chemically related to tannic acid and was also reported to inhibit TMEM16A [26, 25]. However, we found that similar to T16Ainh-A01, the inhibitory effect of digallic acid at TMEM16A was weak; the compound was without an effect at concentrations below 1 mM, and at 1 mM it inhibited both mTMEM16A (ac) and hTMEM16A (abc) by 70.44 ± 6.42 % (n = 5) and 62.93 ± 5.91 % (n = 3), respectively (at +80 mV; at −80 mV inhibition was weaker; data not shown).

Fig. 5
figure 5

Effects of T16Ainh-A01 on TMEM16A currents. a Time course for the effect of the increasing concentrations of T16Ainh-A01 on the TMEM16A currents recorded at +80 mV (upper line, black squares) and −80 mV (lower line, open circles) with a free intracellular Ca2+ concentration of 447 nM. The TMEM16A currents were induced using the voltage ramp protocol similar to that shown in Fig. 3a. b Representative current traces recorded in the presence of different concentrations of T16Ainh-A01. c Summary of the effects of T16Ainh-A01 on TMEM16A currents. *P < 0.05, **P < 0.01 compared with the control currents in the absence of T16Ainh-A01 at +80 mV. # P < 0.05, ## P < 0.01 compared with the control currents in the absence of T16Ainh-A01 at −80 mV; n = 10–20. d–f Effects of T16Ainh-A01 on TMEM16A currents recorded using voltage step protocol. g, h Effects of T16Ainh-A01 on the deactivation of TMEM16A currents. i Summary data for the effects of T16Ainh-A01 on currents of hTMEM16A (abc) transiently overexpressed in CHO cells. j Summary data for the effects of T16Ainh-A01 on currents of mTMEM16A (ac) transiently overexpressed in HEK-293 cells

Full size image

Effects of Cl− channel inhibitors on bestrophin-1 currents expressed in CHO cells

Next, we systematically compared all the inhibitors tested above for their effects on Best1 currents using the Best1 stably transfected CHO cells. Figure 6a–f (right panel) shows the current traces of Best1 recorded using voltage protocol shown at the top of the figure. Among the inhibitors we tested, DIDS showed the highest potency in inhibiting Best1 currents recorded at +80 mV, with an IC50 of 3.93 ± 0.73 μM, which was 160 folds lower than the IC50 for TMEM16A current inhibition (Tables 1 and 2). Maximal inhibition by DIDS was 82.57 ± 2.34 % at 300 μM (Table 1). NFA, on the other hand, was much less potent in inhibiting Best1 currents as compared to its effect on TMEM16A (IC50s: 102.19 ± 15.05 μM for Best1 vs. 7.40 ± 0.95 μM for TMEM16A). At maximal concentration of 3 mM, NFA inhibited outward Best1 currents by 87.62 ± 2.47 %. CaCCinh-A01 inhibited Best1 currents with IC50 of 7.15 ± 0.65 μM and maximal inhibition of 90.21 ± 2.4 %, which was very similar to its effects on TMEM16A currents (IC50 = 7.84 ± 0.62 μM, maximal inhibition is 90.67 ± 1.15 %; Tables 1 and 2). Tannic acid, NPPB and FFA also inhibited Best1 currents with similar potencies as compared to their effects on TMEM16A currents (Fig. 6, Tables 1 and 2). In summary, the six tested inhibitors inhibited Best1 currents with IC50 values in the following order: DIDS (3.93 ± 0.73 μM, n = 5–10) > CaCCinh-A01 (7.15 ± 0.65 μM, n = 5–14) ~ tannic acid (14.84 ± 0.86 μM, n = 5–14) ~ NPPB (20.76 ± 1.20 μM, n = 6–14) > FFA (63.64 ± 4.0 μM, n = 5–13) > NFA (102.19 ± 15.05 μM, n = 5–14).

Fig. 6
figure 6

Effects of Cl− channel inhibitors on Best1 currents. a–f Whole-cell currents of Best1 were induced by a voltage ramp from −100 to +100 mV. Left panels: time courses of concentration-dependent modulation of the Best1 currents by the increasing concentrations of blockers recorded at +80 mV (upper line, black circles) and −80 mV (lower line, open circles) with a free intracellular Ca2+ concentration of 447 nM. The dotted lines indicate the zero current level. The current amplitudes were measured every 5 or 10 s. Right panels: representative current traces recorded when the effect of modulators has stabilized. The gray traces are control currents in the absence of modulators. g Concentration–response relationships for different inhibitors on Best1 currents recorded at +80 mV. Data were fitted with logistic function. The following IC50 values were obtained: FFA, 63.64 ± 4.0 μM (n = 5–13); CaCCinh-A01, 7.15 ± 0.65 μM (n = 5–14); DIDS, 3.93 ± 0.73 μM (n = 5–10); tannic acid, 14.84 ± 0.86 μM (n = 5–14); NPPB, 20.76 ± 1.20 μM (n = 6–14); NFA, 102.19 ± 15.05 μM (n = 5–14). h Concentration–response relationships for different inhibitors on the inward Best1 currents recorded at −80 mV. Data were fitted with logistic function. The following IC50 values were obtained: FFA, 75.07 ± 2.37 μM (n = 5–13); CaCCinh-A01, 8.53 ± 0.78 μM (n = 5–14); DIDS, 5.18 ± 1.32 μM (n = 5–10); tannic acid, 22.55 ± 1.26 μM (n = 5–14); NPPB, 24.76 ± 0.72 μM (n = 6–14); NFA, 119.76 ± 19.15 μM (n = 5–14)

Full size image
Table 2 Effects of different inhibitors on TMEM16A and Best1 currents measured at −80 mV
Full size table

The effects of Cl− channel inhibitors on the inward currents of Best1 were also studied. Three inhibitors, NFA, NPPB and FFA, which showed dual effects of activation and inhibition of TMEM16A inward currents, did not show same pattern of effect on Best1 currents, and inhibited the inward Best1 currents recorded at −80 mV in a concentration-dependent manner. However, the potencies of the inhibitors in respect of the Best1 inward currents were slightly lower than these for the outward currents (Tables 1 and 2). Interestingly, the maximal inhibition of the inward Best1 currents by most of the inhibitors was more substantial (ranging from 93 % to 98 %), than the maximal inhibition on the outward currents (ranging from 88 % to 95 %). However, DIDS was an exception, with a 77.41 ± 2.63 % inhibition of the inward currents as compared with an inhibition of 81.91 ± 2.66 % of the outward currents.

Finally we tested the effect of T16Ainh-A01 and digallic acid on Best1 currents. Figure 6g and h shows that even at high concentrations, T16Ainh-A01 did not significantly affect Best1 currents recorded either at +80 mV or at −80 mV. Digallic acid (1 mM) inhibited Best1 currents by ~40 % at both voltages (data not shown).

Discussion

In this study, we first established the TMEM16A and Best1 stably transfected CHO cell lines. Although endogenous TMEM16A and Best1 proteins are expressed at low levels in untransfected CHO cells, no plasma membrane expression of TMEM16A or Best1was detected in untransfected CHO cells and no significant intracellular Ca2+-dependent currents were detected in these cells. On the other hand, overexpression of TMEM16A and Best1 not only resulted in prominent membrane expression of these channels, but more importantly produced large intracellular Ca2+-dependent currents with characteristics of CaCC as reported before. Thus our observed effects of the Cl− channel inhibitors should be related to the modulation of the exogenously expressed TMEM16A and Best1.

Next, we characterized the effects of seven Cl− channel inhibitors on TMEM16A and Best1 currents. The results demonstrate that CaCCinh-A01 is the most efficacious inhibitor for both TMEM16A and Best1 currents. NFA and DIDS expressed inverse selectivity towards TMEM16A and Best1 currents, respectively. Thus, NFA inhibited TMEM16A currents with an IC50 around 7 μM and maximal inhibition of ~94 %, as compared with an IC50 around 100 μM and a maximal inhibition of ~ 88 % for Best1 currents. In contrast, DIDS inhibited Best1 currents with an IC50 around 3.5 μM, which is 160 folds lower than the IC50 (~550 μM) for TMEM16A currents. Our work provides a direct comparison of the effects of a number of conventional and novel Cl− channel inhibitors on two CaCC channels, TMEM16A and Best1. Although effects of some of these compounds on these channels are known [17, 7, 9, 27, 37], to our knowledge no direct comparison has been carried out before, therefore our data provide clear and distinct pharmacological signatures for TMEM16A and Best1 which will inform studies of native CaCCs in the future. More importantly, the differentiating effects towards these two proposed CaCCs subunits demonstrated by some of the inhibitors will provide valuable tools to dissect their contribution to native CaCCs expressed in different cells/tissues, and to help to understand their functions. Although the molecular mechanism for the selective effects of NFA and DIDS over TMEM16A or Best1 is not clear, this finding and the differentiation of effects of other inhibitors described above provide primary information for the future development of more selective and potent Cl− channel inhibitors.

It is interesting to note that different from their effects on the outward currents, NFA, NPPB and FFA increased the inward currents of TMEM16A at low concentrations (≤30 μM) and inhibited the currents at high concentrations (≥300 μM). This is a surprising result because in previous studies, NPPB and FFA have been described as inhibitors of CaCC currents [41, 33, 40, 30]. NFA, on the other hand, has been described previously to increase the inward currents and inhibit the outward currents of the CaCC recorded from rabbit pulmonary artery myocytes [32] and rabbit coronary arterial myocytes [21]. Piper et al. [32] and Ledoux et al. [21] postulated that this dual effect of NFA could be due to the multiple binding sites in CaCC accessible by NFA. Clearly NFA, FFA and NPPB are structurally more similar when compared with other inhibitors we used (Fig. 1). This implies that a specific mechanism exists for these three inhibitors for their action on TMEM16A protein. Interestingly, NFA, FFA and NPPB markedly slowed the deactivation of TMEM16A currents (Fig. 4), proposing a mechanism for the observed potentiation of TMEM16A currents. Clearly more experiments are needed to further characterize this dual effect of the NFA, FFA and NPPB on the TMEM16A channel. No activation of inward currents of Best1 was observed for any of the inhibitors we tested. These differential effects of the inhibitors on TMEM16A and Best1 further add the value of the present study in dissecting the contribution of these two channels to endogenous CaCC and the molecular mechanisms of CaCC modulation.

The human TMEM16A has multiple isoforms which are characterized by inclusion/skipping of four segments labeled as a (116 residues), b (22 residues), c (4 residues), and d (26 residues) [4]. The biophysical properties of hTMEM16A are regulated by alternative splicing and the alternative exon 6b may play an important role in the regulation of the TMEM16A channel by Ca2+ [8, 28]. Although T16Ainh-A01 and digallic acid have been described as a novel and high performance inhibitors of TMEM16A currents [25], in this study both compounds were found much less potent and efficacious. Thus, T16Ainh-A01 inhibited mTMEM16A (ac isoform) by about 30 % and hTMEM16A (abc isoform) by about 50 %. This is markedly lower than a value of 80 % reported in a recent patch-clamp study on heterologous hTMEM16A (ac isoform) [3] or almost complete block of hTMEM16A (abc isoform) reported by Namkung et al. [25]. It is possible that the presence of segment b can account for some of the observed difference; indeed, in both this study, and the previous investigations the (abc) isoforms appear to show some 20 % greater sensitivity to T16Ainh-A01 as compared to (ac) isoforms. Further difference can arise from the difference between human and mouse TMEM16A proteins. We tested the effect of T16Ainh-A01 with an intracellular free Ca2+ concentration of 447 nM, which is higher than Ca2+ concentration of 275 nM or 171 nM used in the previous studies [25, 3]. However, even at lower intracellular Ca2+ concentration (166 nM), we still did not see significant effect of T16Ainh-A01 on mTMEM16A (ac isoform; data not shown). It is possible that the poor efficacy of T16Ainh-A01 and digallic acid on mTMEM16A expressed in CHO is related to the expression system itself and potential contamination with other TMEM16 or associated proteins. However, we tested the effect of T16Ainh-A01 on mTMEM16A overexpressed in HEK293 cells, and found T16Ainh-A01 (10 μM) inhibited TMEM16A current by 28 % at +80 mV (Fig. 5j), which was similar to the magnitude obtained in CHO expression system. It is clear though that T16Ainh-A01 does not affect Best1 currents at all.

In conclusion, the results of the present study demonstrate that the Cl− channel inhibitors inhibit TMEM16A and Best1 currents differentially, and, thus, the information provided by this study will help in further studies of CaCC.