Expression cloning of TMEM16A as a calcium-activated chloride channel subunit

Abstract

Calcium-activated chloride channels (CaCCs) are major regulators of sensory transduction, epithelial secretion, and smooth muscle contraction. Other crucial roles of CaCCs include action potential generation in Characean algae and prevention of polyspermia in frog egg membrane. None of the known molecular candidates share properties characteristic of most CaCCs in native cells. Using Axolotl oocytes as an expression system, we have identified TMEM16A as the Xenopus oocyte CaCC. The TMEM16 family of "transmembrane proteins with unknown function" is conserved among eukaryotes, with family members linked to tracheomalacia (mouse TMEM16A), gnathodiaphyseal dysplasia (human TMEM16E), aberrant X segregation (a Drosophila TMEM16 family member), and increased sodium tolerance (yeast TMEM16). Moreover, mouse TMEM16A and TMEM16B yield CaCCs in Axolotl oocytes and mammalian HEK293 cells and recapitulate the broad CaCC expression. The identification of this new family of ion channels may help the development of CaCC modulators for treating diseases including hypertension and cystic fibrosis.

Figure 1. Expression cloning of Xenopus oocyte CaCC and functional expression of xTMEM16A and its mouse homologs in Axolotl oocytes

(A–J) Two-electrode voltage clamp recording from Axolotl- and Xenopus oocytes injected with various RNAs. With the exception of (A) all oocytes have been injected with caged IP₃ at least 1 h before the experiment. The red bar indicates time of light flashes used for photo-release of IP₃. (A) Uninjected Axolotl oocyte exhibiting endogenous voltage-gated proton currents. Oocyte was clamped to voltages between −80 mV and +80 mV in 20 mV steps. Holding potential was −60 mV. (B, C) Axolotl oocyte expressing the SK2 Ca²⁺ activated K⁺ channels yielded current after Ca²⁺ increase due to flash photolysis, but not after depolarization alone (C). No calcium-induced current was found in control oocytes injected with water (B). Holding potential was −100 mV. (D and E) Unlike water injected control oocytes (D), oocytes injected with Xenopus oocyte mRNA revealed new current after uncaging of IP₃. Holding potential was −60 mV. (F) Peak amplitude of IP₃-induced currents at −80 mV and +20 mV in water injected control oocytes and xTMEM16A cRNA-injected Axolotl oocytes (mean ± SEM, n=10). Error bars correspond to SEM. (G) Endogenous Ca²⁺-activated chloride currents from Xenopus oocyte measured at +20 mV and −80 mV after photo-release of IP₃. Holding potential was −60 mV. (H–J) Ca²⁺-activated currents similar to those in (G) are found in Axolotl oocytes injected with Xenopus TMEM16A cRNA (H), mouse TMEM16A (mTMEM16A-GFP) (I) and mouse TMEM16B (J). Note the difference in kinetics for outward and inward currents. Holding potential was −60 mV. (K) Phylogenetic tree of human TMEM16 members generated using MAFFT multiple sequence program (Katoh and Toh, 2008).

Figure 2. Calcium dependent outward rectification of IP₃-induced currents in Axolotl oocytes expressing Xenopus TMEM16A

(A) Two-electrode voltage clamp recording from a water-injected Axolotl oocyte. Voltage steps range from −140 mV to +60 mV in 20 mV increments. The UV flash (indicated by red bar) lasted 250 msec. Holding potential was −60 mV. (B–D) Axolotl oocyte injected with xTMEM16A cRNA (same voltage clamp as in (A)), with UV flash time of 50 msec (B), 100 msec (C), and 250 msec (D). Holding potential was −60 mV. (E) IV curves of xTMEM16A-induced peak currents from the same oocytes ((B), black; (C), red; (D), green), showing that the outward rectification diminishes as the UV flash is lengthened causing increasing internal calcium concentration. (F) Dependence of the xTMEM16A-induced outward and inward current amplitude on the duration of UV flash for photo-release of caged IP3. Current was normalized against the current induced by 500 msec UV flash (saturation condition). The correlation between UV flash duration and internal calcium level is not linear. For simplicity, we used the Hill curve for fitting data. UV flash durations necessary for half maximal current at +40 mV and −80 mV were 76 msec and 106 msec, respectively. (G). Summary of recordings from several oocytes showing in each case the UV flash duration for half maximal activation as determined from individual Hill plots is larger for the inward current at −80 mV (170 ± 16 msec) than for the outward current at +40 mV (95 ± 9 msec) (mean ± SEM). This voltage dependence is highly significant (Wilcoxon matched-pairs signed ranks test, p <0.002). Error bars correspond to SEM.

Figure 3. Carbachol- and calcium-induced currents with outward rectification and pharmacology of IP3-induced currents in Axolotl oocytes expressing Xenopus TMEM16A

(A–C) Two electrode voltage clamp traces of oocytes clamped at −60 mV and treated with 5 µM carbachol at time indicated. (A) Typical current trace of a water injected axolotl oocyte. (B) Axolotl oocyte injected with xTMEM16A cRNA. The shape of carbachol-induced inward currents from different oocytes varies. In some oocytes the fast component (first peak) was almost absent. (C) Axolotl oocytes injected with xTMEM16A and human m1AChR cRNA. The current response of all tested oocytes showed a large early component and a more variable and smaller slow response (n=10). (D–F) Current response of A23187 pre-treated oocytes to elevation of external calcium from 0 to 5 mM at time indicated by the horizontal bar. Oocytes were clamped at −60 mV. (D) Recording from uninjected Xenopus oocyte showing the two typical CaCC components. The slow component is more variable. (E) Typical trace from an uninjected Axolotl oocyte. (F) Recording from an Axolotl oocyte injected with xTMEM16A cRNA. Fast and slow components showed variability. (G–H) Voltage clamp recording from an A23187-treated Axolotl oocyte injected with xTMEM16A cRNA at pH6.2 in Ca²⁺-free solution (G) and 90 sec after elevating external calcium to 5 mM (H). Oocytes were clamped from −30 mV to voltages between −140 and +60 mV in 20 mV steps. (I) Voltage dependence of current amplitude induced by carbachol application for 90 s, or Ca²⁺ elevation subsequent to A23187 exposure in xTMEM16a cRNA or water injected oocytes. Amplitude in the IV curve (mean ± SEM, n=10 each) was measured at 0.75 s after the voltage step from a holding potential at −30 mV to the indicated potential. Error bars correspond to SEM. (J) Dose response curves for the xTMEM16A-induced current block by different CaCC blockers determined by measuring IP₃-induced peak current at +20 mV. Holding potential was −30 mV. For each oocyte, the current block as function of blocker concentration was fitted with a Hill curve to yield the dissociation constant (K) and Hill coefficient (N). Curves shown are based on the means of those values determined from individual fits. Values given below are mean ± SEM. NFA (black squares): K=29 ± 6 µM, N=1.1 ± 0.1 (n=10). DIDS (red circles): K=24 ± 2 µM, N=1.4 ± 0.2 (n=6). NPPB (blue triangle): K=77 ± 21 µM, N=1.8 ± 0.3 (n=6). DPC (turquoise diamond): K=155 ± 13 µM, N=2.5 ± 0.4 (n=6). Tamoxifen (green triangle): no dissociation constant or Hill coefficient could be determined due to limited solubility (n=8). Error bars correspond to SEM.

Figure 4. Xenopus oocyte CaCC and xTMEM16A-induced CaCC have multiple current components with different anion selectivity

(A) Reversal potential of IP₃-induced conductance as function of the extracellular Cl⁻concentration (Cl⁻ has been substituted with gluconate). In Xenopus oocytes (red circles, n=10) and Axolotl oocytes injected with xTMEM16A cRNA (black squares, n=8), the slopes (53 ± 1 mV and 62 ± 3 mV per ten fold concentration change) are typical of Cl⁻channels. Replacing external sodium with calcium (2Na⁺ by 1Ca²⁺ + 1glucose, blue triangle) or NMDG (green triangle) had no significant effect on the reversal potential (1 ± 3 mV and 3 ± 3 mV per ten fold concentration change). Error bars correspond to SEM. (B) Voltage clamp traces of Xenopus oocyte in high SCN⁻ solution clamped in 5 mV steps from −85 mV to −60 mV. Whereas the CaCC currents were sustained for at least a couple seconds after the UV flash, clamping the membrane potential at −70 mV caused some of the current components to manifest as outward currents while others appeared as inward currents. The red bar indicates time of light flashes used for photo-release of IP₃. (C–F) Membrane potential traces from oocytes perfused with high Cl⁻, high Br⁻, high I⁻and high SCN⁻ solutions (see Experimental Procedures). Recordings were made with a single electrode. After photo-release of IP₃ (red bar), the Ca²⁺-activated chloride channels dominate the cellular conductance and the membrane potential can be used as estimation of the reversal potential of CaCCs. (C and D) Traces from Xenopus oocyte. In Xenopus oocytes exposed to external bromide, iodide or thiocyanate, IP₃-induced CaCC currents drove the membrane potential toward different levels at different times (C). Whereas Xenopus oocytes in isotonic chloride solution were driven toward the chloride equilibrium potential of around −20 mV upon CaCC activation, replacing 90 mM external chloride with thiocyanate revealed the presence of at least two CaCC current components with different permeability ratio, so that the membrane potential was first driven toward ~−80 mV (C) then toward ~−70 mV regardless whether the resting potential happened to be above (D) or below (C) −70 mV. The oocyte used in (C) is the same as that in (B). (E, F) Traces from axolotl oocytes injected with xTMEM16A cRNA. In an Axolotl oocyte with more depolarized membrane potential, exposure to external iodide or thiocyanate caused the IP₃-induced CaCC activity to drive the membrane potential first below and then above the resting potential (F). In oocytes with more hyperpolarized membrane potential, under bi-ionic conditions the CaCC activation caused the membrane potential to be driven first quickly toward one reversal potential and then slowly towards another, more depolarized, reversal potential (E). While traces shown in (C) and (E) are more typical, a fast and at least one slow component with different reversal potential in bi-ionic conditions are more obvious in (D) and (F). (G) Permeability ratios (mean ± SEM) calculated from changes of reversal potentials for different anions. The index 1 refers to the permeability ratio immediately after Ca²⁺ increase and channel opening. In a simple model the majority of the open channels might be in the same fast state. For the calculation of permeability ratios with index 2 the most positive reversal potential determined under bi-ionic conditions was used. At this point ion channels might occupy various slower states and differences in permeability ratios between Xenopus oocytes and Axolotl oocytes injected with xTMEM16A cRNA are likely due to differences in the occupation of these states.

Figure 5. Ca²⁺-activated currents in HEK293 cells expressing mTMEM16A-GFP

(A) Representative whole cell recording from a HEK293 cell transfected with mTMEM16A-GFP. The patch pipette contained 500 nM free calcium. The cell was clamped from the holding potential (0 mV) to voltages between −100 and +100 mV in 20 mV steps followed by a step to −100 mV (see (B), insert). The same protocol (see Experimental Procedures) was used in (B, C). (B) Whole cell recording from a HEK293 cell transfected with GFP. The patch pipette contained 500 nM free calcium. (C) Whole cell recording from a HEK293 cell transfected with mTMEM16A-GFP. The patch pipette contained 0 nM free calcium. (D) Bar graph showing the mean and SEM of whole cell currents measured 0.75 s after depolarization to +80 mV (n=10 each). All recordings were performed 3 to 5 minutes after break in. The typical TMEM16A current was observed in all 10 TMEM16A-GFP transfected cells under 500 nM Ca²⁺, but not in GFP transfected cells. Three of the TMEM16A-GFP transfected cells showed significant currents immediately after brake in with Ca²⁺ free pipette solution, but these currents disappeared within 3 min. Possible explanations include slow diffusion of calcium buffer, calcium leakage during break in, and slow channel closure conceivably involving calcium depending enzymes. (E) IV curve (mean and SEM) showing outward rectification for the same set of experiments. (F) Time constant of deactivation (mean and SEM, n=10) of Ca²⁺-activated currents as function of membrane potential, from whole-cell patch clamp recording of HEK293 cells expressing mTMEM16A-GFP. Error bars in (D–F) correspond to SEM.

Figure 6. Surface expression of TMEM16A-GFP and TMEM16A mRNA expression in mammary and salivary glands

(A, B) Confocal image of HEK293 cells transfected with mTMEM16A-GFP. A strong GFP signal is visible in the plasma membrane. We noticed that cells expressing mTMEM16A detach more easily from the surface and have a rounder appearance than GFP transfected cells. (C, D) In situ hybridization of mouse mammary gland (day 18 of pregnancy) with antisense probe (C) directed against mTMEM16A. A strong signal is visible in epithelial cells of the alveoli. No signal appears in control with a mTMEM16A sense probe (D). (E) In situ hybridization of mouse salivary glands. TMEM16A expression is high in all epithelial cells, but highest in acinar cells of parotid gland. Abbreviations: ln: lymph node, p: parotid gland, sl: sublingual gland, sm: submandibular gland. (F) In situ hybridization of mouse parotid gland using an mTMEM16A antisense probe.