The role of calmodulin in the activation of Ca2+-activated Cl− channels (CaCCs) has been an important topic over the past three decades of CaCC research. Recently, the discovery of ANO1/TMEM16 as a component of CaCC raised the question of whether calmodulin modulates ANO1 activity (Tian et al., 2011; Terashima et al., 2013; Vocke et al., 2013; Yu et al., 2014a). Apart from its activation by calmodulin, our group reported that HCO3− permeability of human ANO1 (hANO1) can be dynamically modulated by Ca2+/calmodulin (Jung et al., 2013). We read with interest the article titled “Calcium-calmodulin does not alter the anion permeability of the mouse TMEM16A calcium-activated chloride channel” (Yu et al., 2014b) that challenges the conclusions of our study (Jung et al., 2013). The two major concerns raised by Yu et al. are: (1) they were unable to reproduce the calmodulin-dependent ion permeability changes in inside-out patch recordings of mouse ANO1/TMEM16A (mANO1) channel, and (2) whole-cell recordings used in our study to measure the bi-ionic potentials are not suitable to obtain the reversal potential (Erev) because of the series resistance and/or ion accumulation problems. However, we believe that the conclusions of Yu et al. were based on several inappropriate assumptions and technical issues.
First, our conclusion of Ca2+/calmodulin-induced regulation of ANO1 HCO3− permeability is based on an integrated approach of biochemical and electrophysiological methods. For example, in whole-cell recordings we used (a) the calmodulin-binding inhibitor J-8, (b) calmodulin knockdown by siRNAs, and (c) mutation of calmodulin-binding domains (CBDs) in hANO1 to examine the involvement of calmodulin in the high Ca2+-induced regulation of hANO1 HCO3− permeability. In addition, inclusion of calmodulin to the cytoplasmic side of outside-out and inside-out patches reproduced the results of the whole-cell recordings. On the other hand, Yu et al. (2014b) used exclusively inside-out patches expressing mANO1 and recombinant tagged-bovine calmodulin, and concluded that calmodulin does not alter the anion permeability of ANO1. Yu et al. did not examine the effects of calmodulin using other approaches. Suspecting that the differences in calmodulin and approach used by Yu et al. may have led to the disparate findings, we attempted to reproduce their findings using similar inside-out patch recordings and, importantly, the same His6-tagged recombinant bovine calmodulin used by Yu et al. (C4874; Sigma-Aldrich) and the calmodulin purified from human brain (208698; EMD Millipore). Although the human calmodulin reproduced our results of calmodulin-induced regulation of ANO1, the effect of the tagged-bovine recombinant calmodulin was much smaller than that observed with human calmodulin (Fig. 1, A–C). In Fig. 1 A, a strong blockade or reduction in the current magnitude during calmodulin treatment might shift Erev to 0. When we analyzed the I-V relationship during zero-current clamping, the channel conductance decreased time dependently as a result of the rundown of ANO1 current in excised patches (Fig. 1 B). However, only 5.6% of the initial conductance at point (4) was enough to maintain Erev, indicating that the ANO1 channel conductance (gC) is far greater than the background conductance (gB) during the entire Erev measurement, and that the reduction in gC cannot account for the elevation in Erev.
The reason for the discrepancy between the two calmodulins is unclear at the present time. Although the amino acid sequence of human and bovine calmodulins is identical, the His6 tag attached to the recombinant bovine calmodulin seems to affect its properties and the effect of calmodulin on ANO1 HCO3− permeability. In our previous study (Jung et al., 2013), we used recombinant human calmodulin after removing the GST-tag by thrombin digestion for patch-clamp experiments. Calmodulin is a strongly negatively charged molecule and commonly binds to an amphipathic α-helical segment that is positively charged. The CBDs of hANO1 belong to an α-helical “1–8–14 motif” with a weak net positive charge (+1) between 1 and 14 residues (Jung et al., 2013). Therefore, because of the relatively weak electrostatic interaction, it is conceivable to speculate that high fidelity calmodulin structure is required to modulate ANO1 activity, and that especially positive charges from the His6 tag might hamper the proper protein–protein interaction between calmodulin and ANO1.
Second, regarding the concerns related to series resistance and ion accumulation in whole-cell recording, we think that Yu et al. (2014b) assumed extreme conditions, which maximize the effect of those problems that are not relevant to our recordings. In our study, zero-current clamping mode was used to measure the membrane potentials, and these values were used for all mechanistic analyses, including statistical comparisons. During zero-current clamping, we occasionally applied ramp pulse to obtain I-V curves, which were used only to confirm ANO1 currents. Therefore, our ion permeability (Px/PCl) analysis was entirely based on the membrane potential obtained from zero-current clamping. Theoretically, in zero-current clamping, there will be no potential change problems caused by series resistance (Verr = Icmd × Rseries. If Icmd = 0 pA, then Verr = 0. Where Verr = voltage error, Icmd = command current, and Rseries = series resistance; Armstrong and Gilly, 1992; Sakmann and Neher, 1995). Consistent with this notion, the voltage drop was negligible when we analyzed the plots of membrane potential versus conductance that had been used for our PHCO3/PCl and PI/PCl measurements (Fig. 1, D and E). This result contradicts the findings in Fig. 4 of the paper by Yu et al. (2014b). The voltage drop caused by large current amplitude is problematic only in the voltage-clamp recording of reversal potential (Erev) measurements using I-V curves. In this case, Verr caused by the large currents (Rseries = ∼2 MΩ and Rm = ∼10 MΩ in our whole-cell recording) is usually within 20% of the Erev (a maximum of ∼4 mV in HCO3−/Cl− bionic potential). The value corresponds to ∼0.1 of PHCO3/PCl changes in our measurements, where PHCO3/PCl was increased from 0.38 to 1.07 by high Ca2+ (Jung et al., 2013). Therefore, even in the Erev measurements using I-V curves, Verr caused by series resistance would influence only 15% of the total PHCO3/PCl change induced by high Ca2+. In addition, Yu et al. (2014b) suspected that the absence of PHCO3/PCl change by high cytosolic Ca2+ in the CBD-mutated ANO1 might be caused by the smaller current size of the mutant ANO1, because the voltage drop would be smaller if the CBD mutation itself negatively affected ANO1 current. However, there was no difference between the current size of wild-type and CBD-mutated ANO1 (Fig. 1 F). Collectively, the above results indicate that the series resistance problems caused by large current amplitude do not account for the increase in PHCO3/PCl by high cytosolic Ca2+.
We are fully aware of the ion accumulation problem in bi-ionic potential measurements during whole-cell recording. However, this problem would be negligible in zero-current clamping experiments, as discussed by Yu et al. (2014b). Moreover, even in the Erev measurements using I-V relationships, we injected a ramp pulse of 250 ms, which is significantly shorter than that used by Yu et al. (3 s; 12 times longer than Jung et al., 2013), and which minimizes any potential ion accumulation. Therefore, the condition that Yu et al. used for demonstrating a potential ion accumulation problem was many folds more favorable than our conditions to induce ion accumulation. Another point that we would like to stress is that the accumulation of anion X− in the cytosolic side always reduces the Px/PCl value according to the Goldman–Hodgkin–Katz (GHK) flux equation. The main focus of our study is the narrowing of the Px/PCl intervals by Ca2+/calmodulin. Therefore, high cytosolic Ca2+ decreased ANO1 permeability to highly permeable anions such as I− (PI/PCl), whereas it increased ANO1 permeability to the poorly permeable ions such as HCO3− and F− (PHCO3/PCl and PF/PCl). Yu et al. attempted to address the ion accumulation problem and Px/PCl changes by only measuring anions highly permeable to ANO1, such as SCN− and I−, and showing a drop in the SCN−/Cl− bi-ionic potential and PSCN/PCl by the large current amplitude. However, ion accumulation in the cytosolic side cannot explain the increase in PHCO3/PCl by high cytosolic Ca2+, which is the main finding of our study. If the arguments by Yu et al. were valid and HCO3− was indeed accumulated during our whole-cell recordings, the PHCO3/PCl should have decreased by the high cytosolic Ca2+-induced large currents according to the GHK equation. Our findings show the opposite! One might argue that depletion of cytosolic Cl− during ANO1 activation in the whole-cell patch would increase PHCO3/PCl. However, because the pipette resistance (2 MΩ) is much lower than the membrane resistance (10 MΩ, even in the case of large whole-cell current), intracellular Cl− replenishment from the pipette solution will exceed the Cl− depletion through the channel, and thus we do not think that simple cytosolic Cl− depletion elevates PHCO3/PCl from 0.38 to 1.07 in zero-current clamp recordings. Therefore, we believe that all our inside-out, outside-out, and whole-cell patch-clamp data together with the results of biochemical analyses are valid to demonstrate the dynamic modulation of hANO1 anion permeability by Ca2+/calmodulin.
Acknowledgments
Angus C. Nairn served as editor.