EAVK segment “c” sequence confers Ca²⁺-dependent changes to the kinetics of full-length human Ano1

Calcium-activated chloride channel anoctamin1 (Ano1) is necessary for normal slow waves in the gastrointestinal interstitial cells of Cajal. Exon “0” encodes the NH₂ terminus of full-length human Ano1 [Ano1₍₀₎], while exon 13 encodes residues EAVK on its first intracellular loop. Splice variants lack EAVK more often in the stomach than other tissues. Ano1₍₀₎ without EAVK [Ano1₍₀₎ΔEAVK] has reduced sensitivity for intracellular calcium, attributable to slower kinetics. Differential expression of EAVK may function as a calcium-sensitive switch in the human stomach.

Abstract

Anoctamin1 (Ano1 and TMEM16A) is a calcium-activated chloride channel specifically expressed in the interstitial cells of Cajal (ICC) of the gastrointestinal tract muscularis propria. Ano1 is necessary for normal electrical slow waves and ICC proliferation. The full-length human Ano1 sequence includes an additional exon, exon “0,” at the NH₂ terminus. Ano1 with exon 0 [Ano1₍₀₎] had a lower EC₅₀ for intracellular calcium ([Ca²⁺]_i) and faster chloride current (I_Cl) kinetics. The Ano1 alternative splice variant with segment “c” encoding exon 13 expresses on the first intracellular loop four additional amino acid residues, EAVK, which alter I_Cl at low [Ca²⁺]_i. Exon 13 is expressed in 75–100% of Ano1 transcripts in most human tissues but only 25% in the human stomach. Our aim was to determine the effect of EAVK deletion on Ano1₍₀₎ I_Cl parameters. By voltage-clamp electrophysiology, we examined I_Cl in HEK293 cells transiently expressing Ano1₍₀₎ with or without the EAVK sequence [Ano1₍₀₎ΔEAVK]. The EC₅₀ values of activating and deactivating I_Cl for [Ca²⁺]_i were 438 ± 7 and 493 ± 9 nM for Ano1₍₀₎ but higher for Ano1₍₀₎ΔEAVK at 746 ± 47 and 761 ± 26 nM, respectively. Meanwhile, the EC₅₀ values for the ratio of instantaneous to steady-state I_Cl were not different between variants. Congruently, the time constant of activation was slower for Ano1₍₀₎ΔEAVK than Ano1₍₀₎ currents at intermediate [Ca²⁺]_i. These results suggest that EAVK decreases the calcium sensitivity of Ano1₍₀₎ current activation and deactivation by slowing activation kinetics. Differential expression of EAVK in the human stomach may function as a switch to increase sensitivity to [Ca²⁺]_i via faster gating of Ano1.

NEW & NOTEWORTHY Calcium-activated chloride channel anoctamin1 (Ano1) is necessary for normal slow waves in the gastrointestinal interstitial cells of Cajal. Exon 0 encodes the NH₂ terminus of full-length human Ano1 [Ano1₍₀₎], while exon 13 encodes residues EAVK on its first intracellular loop. Splice variants lack EAVK more often in the stomach than other tissues. Ano1₍₀₎ without EAVK [Ano1₍₀₎ΔEAVK] has reduced sensitivity for intracellular calcium, attributable to slower kinetics. Differential expression of EAVK may function as a calcium-sensitive switch in the human stomach.

the interstitial cells of Cajal (ICC) generate electrical slow waves that drive rhythmic contractions in the gastrointestinal tract (11, 34, 37). Several ion channels generate and pattern the electrical slow wave, and among these is a Ca²⁺-activated Cl⁻ current (10, 12, 15, 21, 22, 42). Anoctamin1 (Ano1, TMEM16A, and DOG1) is an 8-transmembrane Ca²⁺-activated Cl⁻ channel that, along with the receptor tyrosine kinase Kit, is selectively expressed in ICC in the gastrointestinal tract (2, 8, 9, 14, 27, 33, 41). Ano1 is localized at the plasma membrane (26, 35) and functions in ICC to regulate proliferation (18, 30) and contribute to slow wave generation (13, 25, 42). In tissues other than gastrointestinal smooth muscle, Ano1 contributes to cellular excitability in vascular smooth muscle (1) and peripheral nerves (3), functions that require rapid gating and can involve movement of small numbers of ions. In contrast to Ano1 functions in electrically excitable cells, Ano1 also contributes to Cl⁻ secretion in epithelia of the intestine (20), lung (2), and salivary gland acinar cells (24). These functions require bulk movement of Cl⁻ ions but do not necessarily depend on rapid gating.

The underlying mechanisms for the differences in Ano1-gating and ionic conductances that are needed to fulfill the disparate Ano1 functions are not clear. However, it is evident that regulation by protein kinases (6), lipids (23), and intracellular Ca²⁺ (2, 27, 41) can alter the biophysical properties of Ano1. In addition to the impact of these molecules on Ano1 function, several transcriptional variants of Ano1 are generated from the mouse and human Ano1 gene (Fig. 1). The properties of these segments vary with respect to anion selectivity, inhibitor sensitivity, intracellular Ca²⁺ regulation, and activation and deactivation kinetics (31, 32, 40). The many possible transcriptional variants of Ano1 result from either alternative splicing of exons or the use of differing NH₂-terminal sequences (7, 19, 29). The splice variants can be generated by in-frame deletion of three exons, referred to as segments “b,” “c,” and “d” of the channel protein and coding for 22, 4, and 26 amino acids (exons 6b, 13, and 15, respectively, in the human ANO1 gene) (7). At least three segments with differing NH₂-terminal amino acid sequences have been reported (17–19, 29). Thus tissue-specific expression of any of the multiple isoforms of Ano1 proteins can confer differing biophysical properties on the expressed channels. Furthermore, altered expression of specific Ano1 segments can have pathological effects (17, 31).

Fig. 1.

Sequence alignment of human and mouse anoctamin1 (Ano1) isoforms. The human Ano1 isoforms Ano1(0abcd) (19, 31), Ano1(abcd) (19, 31), hAno1(Δ1,2bcd) (19, 29), hAno1(Δ1,2,3bcd) (19), and hAno1(bcd) (29) were previously characterized by electrophysiology and are aligned using sequences from the current human annotation (#108). The mouse Ano1 sequence (annotation #106) is aligned to the human sequences using the transcript XM_006508458.3 that contains the commonly used full-length sequence. This alignment does not include the recent mouse transcript XM_006508463.3 (that has an alternative 5′-end with an alternative exon 1 that joins to exon 2 by skipping the current exon 1) that has not been previously studied by electrophysiology. In colors are highlighted the segments “a,” “b,” “c,” and “d” with the exon correspondence between mouse and human genes. The putative transmembrane domains are highlighted in gray.

Our studies on Ano1 variants in human gastric muscularis propria identified differential expression of the three spliced exons as well as three alternative NH₂-terminal sequences (17, 29). One of the NH₂-terminal segments is generated using an exon 93-kb upstream of, and in frame with, the annotated first exon in the human ANO1 gene (19). This upstream exon “0” [hAno1(0abcd), Fig. 1] is homologous to the annotated first exon of the mouse Ano1 gene (mAno1, Fig. 1), and its expression is driven by a functional promoter sequence upstream of exon 0 (19). The functional effect of including the sequence from exon 0 in heterologously expressed full-length clones of Ano1 (including exons 6b, 13, and 15) was a significant increase in the current density, a greater sensitivity to [Ca²⁺]_i, and changes to the Ca²⁺ dependence of activation and deactivation kinetics (31).

The effect of this extended NH₂-terminal sequence of Ano1 in the context of transcriptional variants generated by exon splicing is not known. Several groups have determined that shorter constructs without exon 0 exhibit differences in [Ca²⁺]_i sensitivity and gating kinetics when the exon usage is changed (4, 7, 36, 40). Studies showed that the variant segment c has significantly lower Cl⁻ current densities at all voltages and altered channel kinetics at low or no intracellular Ca²⁺ (7, 40). This exon is found in Ano1 transcripts isolated from human gastric muscularis propria at a much lower abundance (25% of Ano1 transcripts; Ref. 17) than that typically found in other tissues (>75% of transcripts; Ref. 7). Interestingly, segment c abundance was higher in the gastric muscularis propria of patients with diabetic gastroparesis compared with healthy controls, indicating an association between Ano1 function and disease (17). Due to the contribution of Ano1 to normal electrical slow wave activity (13, 28), it is important to understand how these properties of Ca²⁺ sensitivity and altered channel density and kinetics are affected by differential splicing of exon 13 in the context of exon 0-containing transcripts of the gene. Therefore, the aim of this study was to determine the effect that segment c expression has on voltage, time, and [Ca²⁺]_I dependence of full-length Ano1 variants containing exon 0.

METHODS

Cloning of Ano1₍₀₎ and Ano1₍₀₎ΔEAVK.

An expression construct for full-length Ano1 with the alternative 5′-transcriptional start site encoding exon 0 (Ano1₍₀₎) was described previously (19, 31). A second construct was made from Ano1₍₀₎ truncating exon 13, which encodes four amino acids, EAVK, to create a splice variant in the Ano1₍₀₎ genetic background: Ano1₍₀₎ΔEAVK.

HEK293 cell transfection.

Constructs for Ano1₍₀₎ or Ano1₍₀₎ΔEAVK were transiently cotransfected along with green fluorescent protein as a reporter (pEGFP-C1; Clontech, Palo Alto, CA) by the reagent Lipofectamine 2000 (Invitrogen, Carlsbad, CA) into HEK293 cells (American Type Culture Collection, Manassas, VA) as previously described (17, 18).

Solutions.

To record Ca²⁺-activated Cl⁻ current, intra- and extracellular solutions were made as described previously (31) with equimolar Cl⁻ in and outside of the cell to set the reversal potential of Cl⁻ (E_Cl) to an estimated 0 mV. Principal cations in the solutions were N-methyl-d-glucamine (NMDG⁺) or Cs⁺ so that visible current could be readily attributed to Cl⁻. The extracellular solution for whole cell patch-clamp experiments was as follows (in mM): 155 Cl⁻, 150 NMDG⁺ or 150 Cs⁺, 10 HEPES, 5.5 glucose, and 2.5 Ca²⁺ at pH 7.35 and 300 mmol/kg. The intracellular solution was as follows (in mM): 150.1–151.5 Cl⁻, 150 Cs⁺, 2 EGTA and 10 HEPES at pH 7.0 and 300 mmol/kg. Stock CaCl₂ was diluted in the EGTA-buffered intracellular solution at volumes calculated with Maxchelator software (http://www.stanford.edu/~cpatton/) to reach 30–3,000 nM free intracellular Ca²⁺. The predicted liquid junction potential was +1.0 mV.

Electrophysiology.

Patch-clamp electrodes were pulled from KG12 glass on a Sutter P97 puller (Sutter Instruments) to a resistance of 1–3 MΩ as described previously. Data were recorded with pClamp 10.4 software on a CyberAmp320, Digidata 1440A, and Axopatch 200B (Molecular Devices). Whole cell Cl⁻ currents from Ano1₍₀₎- or Ano1₍₀₎ΔEAVK-transfected HEK cells were elicited at 20 mV intervals from a holding potential of −100 mV by twelve 1-s steps from −100 through +120 mV. Families of current traces were averaged up to 10 times with an intersweep time of 5–10 s. Samples were taken at a rate of 10 kHz and filtered at 1–2 kHz.

Analysis.

Ca²⁺-activated Cl⁻ currents were normalized to the whole cell capacitance dialed in during recording and then analyzed with Clampfit 10.4, Excel 2010, and SigmaPlot 12.3 as described previously (31). Families of Cl⁻ current densities were measured at three time points; instantaneous (I_INST) or steady-state (I_ACT) Cl⁻ current densities were measured, respectively, at 25 or 975 ms after activation by the voltage stimulus, while tail currents were measured 25 ms after cessation of each voltage stimulus. Activation (τ_ACT) or deactivation (τ_DEACT) time constants were calculated by fitting currents to a weighted exponential function, f(t) = K₀(f₁e^−t/τi) + C, as previously described (17, 31). The effect of [Ca²⁺]_i was analyzed by a four-parameter logistic function, y = min + (max − min)/[1 + (x/EC₅₀) − Hill slope], in which the min or max are the lower or upper asymptotes, the Hill slope is a reflection of the steepness of the curve, and the EC₅₀ is the concentration of half response to [Ca²⁺]_i. Data are expressed as means ± SE, and significance was assigned at P < 0.05 by an unpaired two-tailed Student’s t-test with Welch’s correction (see Figs. 2 and 3) or a two-way ANOVA with Bonferroni posttest (see Fig. 5).

Fig. 2.

Activating currents from the Ano1₍₀₎ΔEAVK splice variant have a higher EC₅₀ of Ca²⁺ dependence than Ano1₍₀₎. A: representative Cl⁻ current traces activated by a voltage step to +80 mV, recorded from HEK293 cells transfected with Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right), and dialyzed with intracellular Ca²⁺ concentrations ([Ca²⁺]_i) ranging from 100 to 3,000 nM as indicated by the spectrum. Dotted line; 0 pA/pF. B: Cl⁻ currents activated at the 1-s time point (I_ACT) from Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right), plotted vs. voltage stimulus with 100–3,000 nM [Ca²⁺]_i (spectrum). C: I_ACT from Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right), plotted as a function of [Ca²⁺]_i at the indicated voltage steps (spectrum). EC₅₀ values (dashed lines): Ano1₍₀₎, 438 ± 7 nM; Ano1₍₀₎ΔEAVK, 746 ± 47 nM; n = 11; *P < 0.05, compared with Ano1₍₀₎ control by an unpaired t-test with Welch’s correction. D–E: voltage of half-activation (V_1/2; D) or slope (δV; E) plotted as a function of [Ca²⁺]_i (P < 0.05, [Ca²⁺]_i as a source of variation of V_1/2 or δV values; P > 0.05, interaction between the EAVK sequence and [Ca²⁺]_i for either V_1/2 or δV by two-way ANOVA with Bonferroni posttest).

Fig. 3.

Deactivating currents from the Ano1₍₀₎ΔEAVK splice variant have a higher EC₅₀ of Ca²⁺ dependence than Ano1₍₀₎. A: representative Cl⁻ current traces deactivating from a conditioning voltage step to +80 mV, recorded from HEK293 cells transfected with Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right) and dialyzed with intracellular Ca²⁺ concentrations ([Ca²⁺]_i) ranging from 100 to 3,000 nM as indicated by the spectrum. Dotted line: 0 pA/pF. B: peak Cl⁻ currents deactivating at the holding voltage (I_DEACT) from Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right), plotted vs. the conditioning voltage with 100–3,000 nM [Ca²⁺]_i (spectrum). C: I_DEACT from Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right), plotted as a function of [Ca²⁺]_i after the indicated voltage steps (spectrum). EC₅₀ values (dashed lines): Ano1₍₀₎, 493 ± 9 nM; Ano1₍₀₎ΔEAVK, 761 ± 26 nM; n = 11, *P < 0.05, compared with Ano1₍₀₎ control by an unpaired t-test with Welch’s correction.

Fig. 5.

The Ano1₍₀₎ΔEAVK splice variant has slower activation kinetics than Ano1₍₀₎ at intermediate [Ca²⁺]_i. A: normalized Cl⁻ current traces activated by a step pulse to +80 mV from Ano1₍₀₎ (black) or Ano1₍₀₎ΔEAVK (gray) at 500 nM (left) or 3 µM (right) [Ca²⁺]_i. B: time constant of activation (τ_ACT) vs. voltage stimulus of Ano1₍₀₎ (black) or Ano1₍₀₎ΔEAVK (gray) with 500 nM (left) or 3 µM (right) [Ca²⁺]_i. P < 0.05, variation from EAVK. P > 0.05, variation from or interaction with voltage stimulus by a two-way ANOVA. *P < 0.05, by Bonferroni posttest. C: τ_ACT of Ano1₍₀₎ (black) or Ano1₍₀₎ΔEAVK (gray) vs. [Ca²⁺]_i. P < 0.05, variation from and interaction with [Ca²⁺]_i by a two-way ANOVA. *P < 0.05, by Bonferroni posttest; n = 5–16 cells per point.

RESULTS

We first aimed to determine whether Ano1₍₀₎ and Ano1₍₀₎ΔEAVK constructs had differences in calcium sensitivities of Cl⁻ currents. Ano1₍₀₎ΔEAVK channels were voltage clamped, and Cl⁻ current densities were compared with those of Ano1₍₀₎ across a range of intracellular Ca²⁺ concentrations ([Ca²⁺]_i). Both constructs produced robust Cl⁻ currents (Fig. 2A). The activation of both constructs was voltage and Ca²⁺ dependent (Fig. 2B). Ano1₍₀₎ activated at lower [Ca²⁺]_i than Ano1₍₀₎ΔEAVK (Fig. 2C). We compared plateau currents (the current at 1 s per step voltage) vs. [Ca²⁺]_i. These data showed that Ano1₍₀₎ΔEAVK currents had a significantly higher EC₅₀ for Ca²⁺ dependence of Cl⁻ currents than Ano1₍₀₎ controls (Ano1₍₀₎, 438 ± 7 nM; Ano1₍₀₎ΔEAVK, 746 ± 47 nM; n = 11; P < 0.05, compared with control by an unpaired t-test with Welch’s correction) (Fig. 2C). Interestingly, the voltage dependence was calcium sensitive, but this calcium sensitivity of voltage dependence was not significantly different between of these two constructs (P < 0.05, [Ca²⁺]_i as source of variation of V_1/2 or δV values; P > 0.05, interaction between the EAVK sequence and [Ca²⁺]_i for either V_1/2 or δV; Fig. 2, D and E).

To look at the calcium sensitivity of Cl⁻ current independent of the voltage dependence of each Ano1 variant, we tested whether Ano1₍₀₎ΔEAVK Ca²⁺ sensitivity was also different using peak tail currents at the holding potential (−100 mV) immediately after steps to the conditioning pulse (Fig. 3). The Ano1₍₀₎ tail current, like steady state, was activated at lower [Ca²⁺]_i than Ano1₍₀₎ΔEAVK (Fig. 3B). The EC₅₀ for [Ca²⁺]_i of deactivating Cl⁻ current was right-shifted for Ano1₍₀₎ΔEAVK compared with full length [Ano1₍₀₎, 493 ± 9 nM vs. Ano1₍₀₎ΔEAVK, 761 ± 26 nM; n = 11; P < 0.05, by an unpaired t-test with Welch’s correction; Fig. 3C]. Our data show that the calcium sensitivities of Ano1 tail currents sensitivity were similar to those obtained using voltage-dependent activation. These results suggest that calcium sensitivity and not voltage dependence is different between the Ano1₍₀₎ΔEAVK and Ano1₍₀₎ constructs.

The [Ca²⁺]_i is known to alter the kinetics of Ano1 activation (steady state) currents, such that the current shape at low [Ca²⁺]_i is “box-like,” as reflected quantitatively by the ratio between total current (I_ACT) and instantaneous current (I_INST), I_INST/I_ACT of ~1. At high [Ca²⁺]_i, this ratio is ~0.3; rather, the current shape is no longer box like. Importantly, Ano1 splice variants missing segment c have been reported to have greater instantaneous current relative to steady-state current at low [Ca²⁺]_i (7). Indeed, our data show that despite a higher EC₅₀ for [Ca²⁺]_i (Figs. 2C and 3C), instantaneous currents of Ano1₍₀₎ΔEAVK were larger than Ano1₍₀₎ at low [Ca²⁺]_i (Fig. 4B). Thus we looked for whether Ca²⁺ dependence of the instantaneous currents was different between splice variants. The ratio of instantaneous-to-steady-state currents (I_INST/I_ACT) was voltage and [Ca²⁺]_I dependent with currents more box-like (approaching I_INST/I_ACT = 1) at lesser [Ca²⁺]_i and/or lower voltages (Fig. 4, C and D). Interestingly, although the steady-state Ano1₍₀₎ current was more Ca²⁺ sensitive (438 ± 7 nM EC₅₀) than Ano1₍₀₎ΔEAVK (746 ± 47 nM EC₅₀; Fig. 2C), instantaneous Cl⁻ currents of the isoforms had similar calcium sensitivities [EC₅₀: Ano1₍₀₎, 413 ± 15 nM; Ano1₍₀₎ΔEAVK, 395 ± 22 nM; n = 11, P > 0.05, compared with by an unpaired t-test with Welch’s correction; Fig. 4D].

Fig. 4.

The Ano1₍₀₎ΔEAVK splice variant has a higher ratio of instantaneous to steady-state Cl⁻ current at low [Ca²⁺]_i than Ano1₍₀₎. A: normalized Cl⁻ current traces activated by a voltage step to +80 mV, recorded from HEK293 cells transfected with Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right), and dialyzed with [Ca²⁺]_i ranging from 100 to 3,000 nM as indicated by the spectrum. I_ACT, steady-state Cl⁻ current at 1 s is normalized to 1; I_INST, instantaneous Cl⁻ current at 25 ms is indicated by the dashed line. B: instantaneous Cl⁻ current densities from Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right), plotted as a function of the voltage step and dialyzed with 100–3,000 nM [Ca²⁺]_i (spectrum). C: ratio of I_INST to I_ACT (I_INST/I_ACT), plotted vs. the voltage stimulus from Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right) with 100–3,000 nM [Ca²⁺]_i (spectrum), where I_INST/I_ACT ~1 indicates a “box-like” Cl⁻ current. D, I_INST/I_ACT of Ano1₍₀₎ (left) or Ano1₍₀₎ΔEAVK (right), plotted as a function of [Ca²⁺]_i at the indicated voltage steps (spectrum). EC₅₀ values (dashed lines): Ano1₍₀₎, 413 ± 15 nM; Ano1₍₀₎ΔEAVK, 395 ± 22 nM; n = 11; P > 0.05, compared with Ano1₍₀₎ control by an unpaired t-test with Welch’s correction.

Our data show that compared with Ano1₍₀₎, Ano1₍₀₎ΔEAVK has a significantly different [Ca²⁺]_i EC₅₀ of I_ACT but not [Ca²⁺]_i EC₅₀ of I_INST/I_ACT. Therefore, we predicted that the differences in I_ACT are due to faster time constants of activation (τ_ACT) for Ano1₍₀₎ than Ano1₍₀₎ΔEAVK. This is consistent with previous studies using an Ano1 construct without exon 0, where Ano1ΔEAVK activated slower than wild-type Ano1 (40). Here, we find that at 500–1,000 nM [Ca²⁺]_i, for which Ano1₍₀₎ current was larger than Ano1₍₀₎ΔEAVK, the latter channels opened significantly slower than full length {Fig. 5, A and B, left, and C; 500 nM [Ca²⁺]_i: Ano1₍₀₎, 547 ± 74 ms, n = 15; Ano1₍₀₎ΔEAVK, 968 ± 110 ms, n = 6; 1,000 nM [Ca²⁺]_i: Ano1₍₀₎, 705 ± 87 ms, n = 13; Ano1₍₀₎ΔEAVK, 1,422 ± 144 ms, n = 4; P < 0.05, variation from the EAVK sequence by a two-way ANOVA with Bonferroni posttest}. At the highest [Ca²⁺]_i tested (3 µM), The Ano1₍₀₎ΔEAVK current was sloped (τ_ACT, 892 ± 41 ms) and not different from Ano1₍₀₎ (τ_ACT, 886 ± 121 ms; Fig. 5, A and B, right, and C). These data suggest that the changes in activation kinetics of Ano1₍₀₎ΔEAVK compared with Ano1₍₀₎ lead to the shift in the Ca²⁺ dependence of steady-state I_ACT of Ano1₍₀₎ΔEAVK.

DISCUSSION

We report presently that for the full-length Ano1, including the recently discovered exon 0 [Ano1₍₀₎], omission of segment c (exon 13 or EAVK) leads to a decrease in the calcium sensitivity of Ano1 Cl⁻ currents at all examined voltages. Compared with Ano1₍₀₎, Ano1₍₀₎ΔEAVK had a slower rate of activation and a higher ratio of instantaneous to steady-state Cl⁻ current. The changes in kinetics were not uniform across all [Ca²⁺]_i; there was relatively more Cl⁻ current at low [Ca²⁺]_i early upon depolarization of Ano1₍₀₎ΔEAVK, while Ano1₍₀₎ΔEAVK had less Ca²⁺-activated Cl⁻ current late after depolarization. The codependence of the parameters of voltage, [Ca²⁺]_i, and time leads to a shift in the Ca²⁺ dependence of Cl⁻ current density of Ano1₍₀₎ΔEAVK compared with Ano1₍₀₎.

Previous studies on the splicing of segment c focused on both the EAVK sequence encoded by exon 13 and the adjacent amino acids, especially a repeat sequence of glutamate residues that together with exon 13 create the sequence motif, EEEEEAVK. This concentration of negatively charged residues has similarities to calcium-binding sites in other proteins, such as the large conductance Ca²⁺-activated K⁺ channels (BK) and bestrophin (5, 39), and might be expected to underlie calcium sensitivity of Ano1. However, Ferrera et al. (7) reported that ΔEAVK or ΔVK but not ΔEA in the human Ano1(a,b,c) construct (lacking segment d and exon 0) increased the relative instantaneous Ano1 current and ΔEAVK changed the voltage but not calcium dependence of Ano1. Xiao et al. (40) found that ΔEAVK from the mouse Ano1(a,c) clone (lacking exon 0 and segments b and d) resulted in higher current densities at 0 M Ca²⁺ and, therefore, concluded EAVK affects Ca²⁺-independent gating of Ano1. Yet, they found that EAVK was important for Ano1 activation in the presence of intracellular Ca²⁺. The affinity of mouse Ano1(a,c) for Ca²⁺ was 15-fold lower when segment c was deleted, which is a larger effect than the twofold change that we observed here in our studies on the full-length human Ano1(0,a,b,c,d) construct (40). Furthermore, compared with Ano1(a,c), Ano1(a)ΔEAVK had an increased τ_SLOW without change to τ_FAST, suggesting that EAVK sped up the slow-gating mode of Ano1 activation (38). These studies suggest that EAVK contributes to but is not necessary for voltage and calcium sensitivity of Ano1, but the effects are variable depending on the specific Ano1 construct used and intracellular Ca²⁺ concentration employed. The dependence of mouse Ano1(a,c) on EAVK for the dual-gating modes (fast in milliseconds and slow in seconds) exhibited by this channel is not evident when extracellular Cl⁻ is lowered to 30 mM (4), so we did these studies in physiological Cl⁻ concentrations. In our study, compared with Ano1₍₀₎, Ano1₍₀₎ΔEAVK increased τ_ACT, suggesting that EAVK speeds up activation time even in the exon 0 background in human Ano1(a,b,c,d). Our data are also consistent with other published studies, in that ΔEAVK in the human Ano1(0,a,b,c,d) background resulted in currents with a large, fast instantaneous component that is voltage sensitive and less dependent on Ca²⁺ than the less prominent slow outward component, due to the increased time constant for the slow component (τ_SLOW) as reported by Xiao et al. (40). The main difference observed by including exon 0 and segment d in the construct was a reduced shift in the EC₅₀ value for [Ca²⁺]_i when segment c is removed. One needs to also take into account that human Ano1(0,a,b,c,d) is already more Ca²⁺ sensitive than Ano1(a,b,c,d) (31).

Our studies are also consistent with a requirement for negatively charged amino acids on the COOH-terminal end of the Ano1 protein for effective activation of mouse Ano1(a,c) (lacking mouse exon 1 and segments b and d) by intracellular Ca²⁺ (36). For mouse Ano1(a,c), mutations in any one of six glutamate and aspartate residues to alanine residues between the fifth and seventh α-helical structures on the COOH-terminal end of the protein result in effectively blocking intracellular Ca²⁺ regulation (up to 1,000-fold increase in EC₅₀ value) of the Ano1 channels, indicating that these residues form a Ca²⁺-binding site on the intracellular side of the protein (36). Our data support the concept that exon 0 and segment c are intracellular domains that contribute to the effective coupling of the Ca²⁺-binding site to channel gating and that transcriptional variants with or without these sites alter the Ca²⁺ sensitivity within the physiological range of 100 to 1,000+ nM [Ca²⁺]_i.

The EAVK sequence expressed by exon 13 is a segment that is present in only 25% of all Ano1 transcripts from smooth muscle strips of the human stomach (17), where Ano1 is expressed in ICC and has a pacemaking function (8). The expression of EAVK in human stomach at 25% is low compared with >75% of Ano1 transcripts that include this exon in other tissues including intestinal mucosa, brain, and lung (7), where Ano1 is expressed in the epithelia and plays a role in secretion. In the present work, we used HEK293 cells for transient Ano1 transfection. This is an established and well-validated heterologous expression system that has minimal endogenous Cl⁻ currents and robustly expresses heterologous Ano1. Our findings will need to be verified in primary ICC that express Ano1 transcripts with and without the EAVK sequence, such as those from nondiabetic human gastric muscularis propria.

In the context of regulation of pacemaking by ICC in the gastrointestinal tract, the low expression of EAVK means that the predominant Ano1 channels in gastric smooth muscle that lack EAVK will exhibit a reduced sensitivity to Ca²⁺ but faster instantaneous Ca²⁺-activated Cl⁻ currents that deactivate more slowly. We propose that these properties contribute to the rapid activation of the pacemaker potential but only when intracellular Ca²⁺ reaches the necessary level for activation. In addition, there is evidence that Ca²⁺-activated Cl⁻ currents also contribute to the width of the plateau of the electrical slow wave (16); therefore, slower deactivation of Ano1 channels means that the channels will contribute to a sustained Cl⁻ current, broadening of the electrical slow wave, and more effective electrical coupling between adjacent ICC. Our studies on Ano1 knockout mice show a requirement for Ano1 in the synchronization of Ca²⁺ transients within pacemaker ICC of the mouse small intestine (28). We propose that this synchronization will be facilitated in Ano1 channels that express human exon 0 or mouse exon 1 and the EAVK sequence.

In conclusion, the human full-length Ano1 that includes the recently discovered exon 0 [Ano1₍₀₎], when lacking exon 13 (segment c or EAVK), has a decreased steady-state calcium sensitivity due to slower channel activation. These changes would alter Cl⁻ flux in tissues where there is physiological decreased expression of the EAVK segment such as gastric smooth muscle.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant DK-57061 and Mayo Clinic Center for Cell Signaling in Gastroenterology NIDDK Grant P30-DK-084567.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

P.R.S., S.J.G., A.M., and G.F. conceived and designed research; P.R.S., A.M., and C.E.B. performed experiments; P.R.S. and A.M. analyzed data; P.R.S., S.J.G., A.M., and A.B. interpreted results of experiments; P.R.S. and A.M. prepared figures; P.R.S. drafted manuscript; P.R.S., S.J.G., A.B., and G.F. edited and revised manuscript; P.R.S., S.J.G., A.M., C.E.B., A.B., and G.F. approved final version of manuscript.