CK2 is a key regulator of SLC4A2-mediated Cl−/HCO3− exchange in human airway epithelia

Transepithelial bicarbonate secretion by human airway submucosal glands and surface epithelial cells is crucial to maintain the pH-sensitive innate defence mechanisms of the lung. cAMP agonists stimulate HCO3− secretion via coordinated increases in basolateral HCO3− influx and accumulation, as well as CFTR-dependent HCO3− efflux at the luminal membrane of airway epithelial cells. Here, we investigated the regulation of a basolateral located, DIDS-sensitive, Cl−/HCO3− exchanger, anion exchanger 2 (AE2; SLC4A2) which is postulated to act as an acid loader, and therefore potential regulator of HCO3− secretion, in human airway epithelial cells. Using intracellular pH measurements performed on Calu-3 cells, we demonstrate that the activity of the basolateral Cl−/HCO3− exchanger was significantly downregulated by cAMP agonists, via a PKA-independent mechanism and also required Ca2+ and calmodulin under resting conditions. AE2 contains potential phosphorylation sites by a calmodulin substrate, protein kinase CK2, and we demonstrated that AE2 activity was reduced in the presence of CK2 inhibition. Moreover, CK2 inhibition abolished the activity of AE2 in primary human nasal epithelia. Studies performed on mouse AE2 transfected into HEK-293T cells confirmed almost identical Ca2+/calmodulin and CK2 regulation to that observed in Calu-3 and primary human nasal cells. Furthermore, mouse AE2 activity was reduced by genetic knockout of CK2, an effect which was rescued by exogenous CK2 expression. Together, these findings are the first to demonstrate that CK2 is a key regulator of Cl−-dependent HCO3− export at the serosal membrane of human airway epithelial cells.


Introduction
The secretion of bicarbonate (HCO 3 − ) by epithelial cells is essential for maintaining the normal function of many epithelial tissues primarily due to its ability to act as a biological buffer and therefore a key regulator of extracellular pH [22]. In the human airways, HCO 3 − plays a major role in the innate defence of the lungs to inhaled pathogens. Transepithelial secretion of HCO 3 − and Cl − from serous cells of the submucosal glands, as well as surface epithelial cells, drives isosmotic water secretion, and this HCO 3 − -rich fluid is an important component of the perciliary layer (PCL) of the airway surface liquid (ASL) which lines the conducting airways [4]. The depth of the PCL is important for efficient mucociliary clearance [5,62] while maintaining the correct ASL pH is vital for efficient bacterial killing by pH-dependent antimicrobials as well as reducing mucus viscosity [17,44,53]. In addition, HCO 3 − allows for efficient solubilization and transportation of airway mucus [47]. Therefore, HCO 3 − secretion enables pH-sensitive components of the innate defence mechanisms of the lung to function efficiently. Transepithelial HCO 3 − secretion is a two-stage process that involves (i) import across the basolateral membrane and accumulation above electrochemical equilibrium and (ii) exit across the apical (serosal) membrane down its electrochemical gradient. It is well established that cAMP-dependent stimulation of the cystic fibrosis transmembrane conductance regulator (CFTR) is crucial for HCO 3 − exit across the luminal membrane in human airway epithelia [11,14,18,29,54,56]. Loss of functional CFTR causes cystic fibrosis (CF), an autosomal recessive disorder characterized by the accumulation of thick, sticky mucus, bacterial colonization and chronic inflammation of the airways [34]. Impaired cAMP-stimulated HCO 3 − secretion has been reported in a wide range of CF tissues [6], and the ASL of CF patients is more acidic than that of healthy patients [10,13] which is thought to contribute to the pathology of the disease. Although functional CFTR expression is clearly important for airway HCO 3 − secretion, members of the SLC26 family of Cl − /HCO 3 − exchangers, SLC26A3, SLC26A4 and SLC26A6, also transport HCO 3 − [57,61] and previous studies from our laboratory revealed that SLC26A4 (pendrin) participates in HCO 3 − secretion by Calu-3 cells [14].
We proposed that a functional interaction existed between the phosphorylated R domain of CFTR, and the STAS domain of pendrin as described by Ko et al. [25] which enabled CFTR to regulate HCO 3 − secretion via modulation of pendrin-mediated Cl − /HCO 3 − exchange [14].
HCO 3 − transport across the serosal membrane is believed to be governed by members of the SLC4 family of transporters, including importers such as the Na + /HCO 3 − cotransporters SLC4A4, SLC4A5 and SLC4A7 [11,26,54] which act to accumulate HCO 3 − inside the cell. In contrast, SLC4A2, an electroneutral Cl − /HCO 3 − anion exchanger (AE2) that exports HCO 3 − , plays an important role in the regulation of intracellular pH during cell alkalinization, as well as in the control of cell volume by Cl − uptake [2,59]. Furthermore, it has been shown that AE2 is expressed in the airways [1,12,55] and immunostaining experiments performed on polarized Calu-3 cells have localized the expression of AE2 to the basolateral membrane [32]. Our laboratory [14,15] and others [24] have recently demonstrated that a functional Cl − /HCO 3 − exchanger is present at the basolateral surface of Calu-3 cells under resting conditions, with features consistent with AE2. In addition, Huang et al. [18] showed that intracellular alkalinization produced by basolateral Cl − removal was decreased by 80% in SLC4A2 knockdown Calu-3 cells. We found that this exchanger was almost completely inhibited by elevations in intracellular cAMP [14,15], which we proposed would enhance cAMPstimulated transepithelial HCO 3 − secretion, by reducing HCO 3 − export across the basolateral membrane. However, Kim et al. [24] demonstrated that inhibition of CFTR maintained AE2 activity in cAMP-stimulated Calu-3 cells, while Huang et al. [18] suggested that AE2, acting in concert with Na + /HCO 3 − cotransporters, was involved in basolateral Cl − loading and HCO 3 − recycling that helped maintain transepithelial Cl − secretion under cAMP-stimulated conditions in Calu-3 cells. Thus, the role of AE2 in airway HCO 3 − and fluid secretion is still controversial and not fully understood. Therefore, the major aim of the present study was to further understand the cellular pathways that regulate basolateral Cl − /HCO 3 − exchange activity in human airway epithelial cells to help provide a better understanding of its role in airways HCO 3 − secretion. Our results have uncovered a novel regulation of the Cl − /HCO 3 − exchanger by Ca 2+ /calmodulin signalling and by the master protein kinase, CK2 (casein kinase 2). Since CK2 has previously been implicated in regulating HCO 3 − transport by CFTR in secretory epithelia [65], our findings strongly suggest that CK2 is a key regulatory component of transepithelial HCO 3 − transport in the human airways.
Human embryonic kidney-293T (HEK-293T) cell culture HEK-293T cells were grown in T 75 Costar cell culture flasks (75 cm 2 ) with 30 ml of Dulbecco's Modified Eagle's Medium (DMEM), supplemented as for Calu-3 cells and incubated at 37°C in a humidified air containing 5% (v/v) CO 2 . Cells were initially seeded at 1 × 10 6 cells per flask and passaged every 7 days using 0.05% trypsin and 0.02% ethylenediaminetetraacetic acid (EDTA) in Earle's balanced salt solution. The knockout CK2 (both the alpha (α) and alpha prime (α prime) catalytic subunits) HEK-293T cells were generated by CRISPR/Cas9 gene editing. Genomic target sequences of guide RNA (gRNA) were 5′-TTACATGTATGAGA TTCTGA-3′ (CK2 α) and 5′-GGGTCTACGCCGAG GTGAAC-3′ (CK2 α prime). The absence of the catalytic subunits was confirmed by Western blotting (see supplementary Fig. 2 Transfection of HEK-293T cells Mouse AE2 cDNA was a kind gift from Beth Lee and Ron Kopito, and contained a human haemagglutinin (HA) tag; for more details, see Lee et al. [28]. Empty plasmid (pcDNA 3.1 myc/His), human WT-CK2 (CK2-alpha) and the double CK2 mutant (V66A and I174A; DM-CK2) cDNA were generated as previously described [50,51]. DNA sequencing analysis confirmed sequence identity of all the constructs. To transfect cDNA constructs into HEK-293T cells, cDNA was pre-complexed with Lipofectamine-2000 (Thermo Fisher) at a ratio of 1:2.28, respectively. Opti-MEM media with GlutaMax (Thermo Fisher) was then added for 15 min at room temperature, and then diluted in culture media to produce a final concentration of 1 μg DNA ml −1 . This complex media was added to the cells and incubated for 6 h at 37°C before the complex media was removed and cells were incubated with Opti-MEM plus 10% FBS overnight before being returned back to normal culture media.
Intracellular pH measurements For measurements in polarized Calu-3 cells and well-differentiated human nasal epithelia, cells were loaded with 10 or 40 μM, respectively, of the pH sensitive dye, 2′-7′-bis (carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM) in Na-HEPES and incubated for 60 min at 37°C. Cells were mounted on to the stage of a Nikon fluor inverted microscope and viewed at ×60 magnification using a long working distance objective (N.A 0.6). Cells were perfused with Krebs solution at 37°C gassed with 5% (v/v) CO 2 /95% (v/v) O 2 at a rate of 3 ml min −1 (apical) and 6 ml min −1 (basolateral). For measurements in HEK-293T cells, cells were loaded with 10 μM BCECF-AM in Na-HEPES and incubated for 10 min at 37°C. After dye loading, coverslips were placed in a perfusion chamber and then mounted onto the stage of a Nikon inverted microscope. Cells were viewed at ×40 magnification using an oil immersion objective (N.A 1.2) and perfused with Krebs solution at 37°C gassed with 5% CO 2 /95% O 2 at a rate of 3 ml min −1 . Intracellular pH (pH i ) was measured using a Life Sciences Microfluorimeter System in which cells were alternatively excited at 490 and 440 nm wavelengths every 1.024 s with emitted light collected at 510 nm. The ratio of 490 nm emission to 440 nm emission was recorded using the PhoCal 1.6b software and calibrated to pH i using the high K + / nigericin technique [16] in which cells were exposed to high K + solutions containing 10 μM nigericin, set to a desired pH, ranging from 6.6 to 8.4. Total buffering capacity (β tot ) was calculated by addition of the intrinsic buffering capacity (β i ) to the buffering capacity of the CO 2 -HCO 3 − buffer system (βHCO 3 − ) in which β i was calculated using the NH 4 + technique as described by Roos, Boron [49]. For analysis of pH i measurements, delta pH i (ΔpH i ) was determined by calculating the mean pH i over 60 s before, during and after treatment. Rate of pH i change (ΔpH i /Δt) was determined by performing a linear regression over a period of at least 30 s which was converted to a transmembrane HCO 3 − flux (−J(B)) by multiplying ΔpH i /Δt by β tot .
Confocal microscopy to detect expression of HA-tagged mAE2 Control, untransfected and transfected HEK-293T cells expressing HA-tagged mAE2 were grown on glass coverslips for 2 days and fixed with 4% PFA for 10 min at room temperature. Cells were washed with PBS three times for 5 min, and then with 50 mM NH 4 Cl to quench any remaining PFA. After washing, fixed cells were permeabilized using 1% Triton X-100 for 5 min at room temperature and washed in PBS three times for 5 min. To block non-specific binding, cells were incubated with blocking buffer, consisting of 5% goat serum and 1% Na-azide in PBS, at room temperature for 30 min. Blocking buffer was removed and cells were incubated in diluted primary antibody (Anti-HA16B12, 1/1000 in blocking buffer, Abcam) overnight at 4°C on a shaker. Cells were then rinsed in PBS three times for 15 min to remove any unbound primary antibody, and then incubated with FITC-conjugated goat anti-mouse antibody (1/100 in blocking buffer) for 1 h at room temperature in the dark. Following this, cells were washed with PBS three times for 15 min to remove any unbounded secondary antibody. DAPI dye (1 μg ml −1 ) was added onto coverslips for 2 min, at room temperature, away from light, to stain the nucleus and gently washed in PBS to remove any remaining DAPI. Coverslips were mounted onto a microscope slide, using mounting medium (VectaShield , 115 NaCl, 5 KCl, 1 CaCl 2 , 1 MgCl 2 and 10 D-Glucose. In the Ca 2+ -free high Cl − -Krebs solution, the NaCl concentration was increased to 116 mM, and CaCl 2 was replaced with MgCl 2 and 0.5 mM EGTA was added to chelate any remaining Ca 2+ . The Cl − -free Krebs solution consisted of (in mM) 25 NaHCO 3 − , 115 Na-Gluconate, 2.5 K 2 SO 4 , 6 Ca-gluconate, 1 Mg gluconate and 10 D-Glucose. In the Ca 2+ -free Cl − -free Krebs solution, the Na-gluconate concentration was increased to 124 mM, and 0.5 mM EGTA was added to chelate any remaining Ca 2+ . The intracellular pH calibration solutions consisted of (in mM) 5 NaCl, 130 KCl, 1 CaCl 2 , 1MgCl 2 , 10 D-Glucose, 10 HEPES (for solutions set at pH 7.6 or below) or 10 TRIS (for solutions set at pH 7.8 or above) as well as 10 μM nigericin. Solutions were set to the desired pH by using either 1 M HCl or 1 M NaOH. The ammonium pulse solutions used to determine intracellular buffering capacity consisted of (in mM) 4.5 KCl, 1MgCl 2 , 2 CaCl 2 , 5 BaCl, 10 HEPES, 10 D-Glucose as well as varying concentrations of NH 4 Cl/NMDG-Cl, ranging from 0 NH 4 Cl/145 NMDG-Cl to 30 NH 4 Cl/115 NMDG-Cl. All solutions were titrated to pH 7. 4 at 37°C using 1 M CsOH.
Statistical analysis All results are presented as mean ± S.E.M. where n is the number of experiments. The GraphPad Prism 4 software (GraphPad Software, USA) was used for statistical analysis and either a Student's t test (paired or unpaired), oneway ANOVA (with Tukey's multiple comparison post-test) or two-way ANOVA (with Bonferroni's post-test), where applicable. p values of <0.05 were considered statistically significant.

Results
Calu-3 cells express a basolateral DIDS-sensitive, Cl − / HCO 3 − exchanger Our laboratory [14,15] and others [24] have previously reported that Cl − /HCO 3 − exchange occurs across the basolateral membrane in non-stimulated Calu-3 cells. In support of these findings, intracellular pH measurements showed that removal of basolateral Cl − caused an intracellular alkalinization of 0.36 ± 0.02 units (n = 8; Fig. 1A) and that readdition of Cl − caused pH i to reacidify at a rate of 0.57 ± 0.07 pH units min The anion exchange inhibitor 4,4′-Diisothiocyano-2,2′stilbenedisulfonic acid (DIDS) dose-dependently inhibited this basolateral Cl − /HCO 3 − exchange activity, with 500 μM DIDS causing complete inhibition (Fig. 1C). Analysis of the DIDS dose-response curves for inhibition of the change in pH i following basolateral Cl − removal, and rate of reacidification following basolateral Cl − reintroduction, gave a calculated IC 50 value of 16.5 ± 1.3 μM and 7.5 ± 1.2 μM, respectively (Fig. 1D, E). Therefore, these data demonstrate that Calu-3 cells have a DIDS-sensitive, basolateral Cl − /HCO 3 − anion exchanger with properties consistent with SLC4A2 (AE2), which support previous results from these cells [14,15,24]. The basolateral Cl − /HCO 3 − exchanger is inhibited by elevations in cytosolic cAMP The next series of experiments were designed to investigate the regulation of the basolateral Cl − /HCO 3 − exchanger. We first sought to assess the effect of elevated cAMP on Cl − /HCO 3 − exchanger activity, since cAMP is known to stimulate transepithelial HCO 3 − secretion from Calu-3 cells in a PKA-dependent manner [14]. As shown in Fig Importantly, although forskolin promoted HCO 3 − efflux across the apical membrane, it appeared to markedly reduce HCO 3 − efflux across the basolateral membrane by the anion exchanger (Fig. 2). Forskolin reduced the mean change in pH i following basolateral Cl − removal by 85.2 ± 2.6% compared to non-stimulated cells (p < 0.001 vs. control; n = 10; Fig. 2C) and also reduced the rate of reacidification following basolateral Cl − reintroduction by 98.4 ± 1.6% compared to non-stimulated cells (p < 0.001 vs. control; n = 10; Fig. 2D). Similar inhibition of the basolateral Cl − /HCO 3 − exchanger was also observed in cells stimulated with the cAMPelevating agonist adenosine (Fig. 2C), a key physiological regulator of cAMP-stimulated ion and fluid transport in human airways [48,63]. In addition, the non-specific phosphodiesterase inhibitor IBMX, and the membrane permeable cAMP analogue dibutryl-cAMP (db-cAMP), both induced almost complete inhibition of the basolateral Cl − /HCO 3 − exchanger (Fig. 2C, D). The multidrug resistance protein 4 (MRP4) inhibitor MK-571 also induced an inhibition of the exchanger, but this was less pronounced than for other cAMP agonists (Fig. 2C, D). Therefore, elevations of inhibition of PKA by two different inhibitors, RpcAMPs and H89, had no effect on the ability of forskolin to reduce AE activity (Fig. 2C, D). Note that this result is in marked contrast to the effect these PKA inhibitors have on the forskolin-activated, CFTR-dependent, apical anion exchanger pendrin, which was reduced by over 80% by H89 [14]. In addition, we also found that the exchange protein activated by cAMP (EPAC) and mTOR, two reported downstream targets of cAMP which are activated independently of PKA [23,52], were not involved in mediating the effect of cAMP on the basolateral Cl − /HCO 3 − exchanger (Fig. 2C, D). Activation of EPAC by 8CPT-2Me-cAMP-AM failed to mimic the effect of forskolin, with similar Cl − /HCO 3 − exchanger activity compared to control cells (n = 3; Fig. 2C, D). Inhibition of mTOR kinase by rapamycin, also had no effect on the forskolininduced inhibition of exchanger (n = 3: data not shown).
In another series of experiments, we tested whether forskolin-stimulated HCO 3 − secretion at the apical membrane could potentially mask any basolateral Cl − /HCO 3 − exchange activity, as suggested by Kim et al. [24]. Addition of the CFTR inhibitor, GlyH-101, partially, but not fully, relieved the forskolin-induced inhibition of the basolateral Cl − /HCO 3 − exchanger (Fig. 3). Given that the percent inhibition of the rate of reacidification in response to reintroduction of basolateral Cl − was 65.2 ± 14.3% (n = 4; Fig. 3C [9,27,38], it has also been demonstrated that the basolateral Cl − /HCO 3 − exchanger in mouse salivary acinar cells is positively regulated by intracellular Ca 2+ [39]. To this end, we first tested whether elevations in intracellular Ca 2+ affected the activity of the basolateral Cl − /HCO 3 − exchanger. Therefore, Calu-3 cells were stimulated with either basolateral carbachol, a muscarinic receptor agonist, (Fig. 4A), or thapsigargin, an inhibitor of sarcoplasmic/endoplasmic reticulum Ca 2+ ATPase (SERCA), (Fig. 4B), which have both been shown to stimulate Ca 2+ -dependent anion transport in Calu-3 cells [37]. However, as shown in Fig. 4C, D, neither carbachol nor thapsigargin had any significant effect on basolateral Cl − /HCO 3 − exchanger activity. Therefore, these data suggest that the exchanger is not regulated by [Ca 2+ ] i . We next tested whether basal levels of [Ca 2+ ] i had any effect on anion exchanger activity. To do this, Calu-3 cells were preloaded with the calcium-chelator, BAPTA-AM, and then basolateral Cl − /HCO 3 − exchanger activity assessed ( Fig. 5A, B). BAPTA-AM loading did not affect resting pH i (control = 7.60 ± 0.06; n = 7 and BAPTA-AM-loaded cells = 7.70 ± 0.02; n = 5; p > 0.05), but did significantly reduce the mean change in pH i following basolateral Cl − removal by 56.2 ± 2.9% (p < 0.001 n = 8; Fig. 5C), as well as the rate of reacidification following basolateral Cl − reintroduction, by 51.3 ± 8.9% (p < 0.01; n = 8; Fig. 5D) compared to control cells. However, BAPTA-AM-loaded cells still showed normal forskolin-induced inhibition of the remaining anion exchange activity. These results suggest that resting levels of intracellular Ca 2+ play an important role in regulating tions, experiments were performed to assess the role of CK2, a protein kinase that is active under resting conditions in airway cells [65], and which is the main serine/threonine kinase both in vivo and in vitro that can phosphorylate CaM [3]. Furthermore, CK2 has been implicated in the regulation of other ion channels including CFTR [7,33,36] and M-type potassium channels, in which it has been reported that CK2dependent phosphorylation of CaM underlies its binding to the channel [21]. In addition, sequence analysis of SLC4A2 reveals the presence of several potential CK2 phosphorylation sites showing the canonical CK2 consensus sequence pS/ pTxxE/D (Supplementary Fig. 1). Calu-3 cells were therefore e x p o s e d t o t h e C K 2 i n h i b i t o r 4 , 5 , 6 , 7 -Tetrabromobenzotriazole (TBB) and then basolateral Cl − / HCO 3 − exchange activity measured (Fig. 6A). TBB exposure alone caused a small intracellular acidification (7.6 ± 0.02 to 7.5 ± 0.01 (p < 0.05, n = 6)), and caused a significant, but fully reversible, inhibition of the basolateral Cl − /HCO 3 − exchanger ( Fig. 6A-C). Since the mean pH i after TBB exposure was within the normal range for Calu-3 cells, it was unlikely that the fall in pH i caused the decrease in basolateral AE activity. However, to provide further support for a role of CK2 in regulating anion exchange activity, the effect of a more specific CK2 inhibitor, CX4945 [45] was tested. Figure 6D shows that exposure to CX4945 almost completely abolished Cl − /HCO 3 − exchanger activity, an effect which was fully reversible on washout of the inhibitor, and which did not involve a change in pH i (Fig. 6E, F). The use of two different pharmacological inhibitors provided strong evidence that CK2 was essential for basal Cl − /HCO 3 − exchanger activity in Calu-3 cells. The efficacy of a short-term exposure to the two inhibitors on the activity of CK2 was further verified by an in vitro kinase assay of whole cell lysates from Calu-3, as detailed in the 'Methods' section. Figure 13A shows that a 5-min treatment of Calu-3 cells with 10 μM TBB lowered CK2 catalytic activity by more than 40%, whereas the same concentration of CX4945 led to an even more dramatic inhibition. In order to investigate whether CK2 regulation of the basolateral Cl − /HCO 3 − exchanger involved CaM, Calu-3 cells were preincubated with the CaM inhibitor J-8 for 60 min, and then cells were acutely exposed to TBB, before basolateral Cl − /HCO 3 − exchanger activity was measured. Although TBB + J-8 caused a trend to further reduce AE activity compared to either drug alone, this effect was not statistically significant and, even in the presence of both drugs, a significant amount of AE activity still remained (Fig. 7A, B). Since there was no obvious additive effect of TBB and J-8 suggests that CK2 potentially controls the resting activity of the basolateral Cl − /HCO 3 − exchanger through the downstream target CaM, in Calu-3 cells.
CK2 inhibition abolishes the activity of basolateral cl − / HCO 3 − exchange in primary human nasal epithelia Having demonstrated that human AE2 was regulated by CK2 in a human airway epithelial cell line, we next assessed whether AE2 activity showed similar CK2-dependent regulation in well-differentiated human nasal epithelial (HNE) cultures. AE2 mRNA expression has previously been identified in the proximal airways and in HNE cells [1,12,55], and HNE cells have also been shown to possess a basolateral, DIDS-sensitive Cl − /HCO 3 − exchanger, indicative of functional expression of AE2 [55]. To this end, intracellular pH measurements were performed on HNE monolayers and the effect of CK2 inhibition on AE2 activity was assessed. In control conditions, removal of basolateral Cl − increased pH i by 0.08 ± 0.01 pH i units, and this response was significantly reduced to 0.01 ± 0.03 pH i units in the presence of CX4945 (n = 5; p < 0.05; Fig. 8). The effect of CX4945 was reversible as the response to basolateral Cl − removal could be recovered after wash out of the drug (0.09 ± 0.01 pH i units; n = 4; p < 0.01 vs. CX4945; Fig. 8). Therefore, these data indicate that basolateral AE2 activity in primary human nasal epithelia is also positively regulated by CK2 which is consistent with our results from Calu-3 cells. . Therefore, to study the regulation of AE2 further, we expressed mouse AE2 (mAE2) in HEK-293T cells. Immunocytochemistry revealed successful transfection and expression of the protein at the cell surface of these cells (Fig. 9), although considerable intracellular expression was also observed. Intracellular pH measurements in nontransfected cells showed that HEK cells possessed an endogenous Cl − /HCO 3 − exchange activity (as previously reported by Sterling et al. [58]). However, this endogenous AE activity was significantly reduced by a low concentration of DIDS (Fig. 10A), which had little effect on exogenously expressed mAE2 activity (Fig. 10B). Therefore, in all further Cl − readdition, respectively. *Significant effect of TBB vs. untreated controls (p < 0.05). Data represents mean ± S.E.M., n = 6. d Shows a representative experiment in which Calu-3 cells were exposed to CX4945 (10 μM) for 5 min and Cl − /HCO 3 − exchanger activity measured. e and f summarize the effect of CX4945 treatment and CX4945 reversibility on the mean change in pHi caused by basolateral Cl − removal and the rate of reacidification upon Cl − readdition, respectively. *Significant effect of CX4945 vs. untreated controls (p < 0.05). Data represents mean ± S.E.M., n = 6 experiments, mAE2 activity was assessed in the continuous presence of 25 μM DIDS. Using this approach, after 2 days of transfection, mAE2 expression increased both the magnitude of alkalinization in response to removal of basolateral Cl − (0.13 ± 0.02 vs. 0.69 ± 0.06; p < 0.001; n = 10) and the rate of reacidification upon basolateral Cl − readdition (0.04 ± 0.01 pH units min −1 vs. 0.46 ± 0.07 pH units min −1 ; p < 0.001; n = 10), compared to non-transfected cells (Fig. 10C, D, respectively). We found mAE2 activity to be sensitive to DIDS, with 100 μM DIDS causing a 40.0 ± 4.3% inhibition of the mean pH i change in response to extracellular Cl − removal (p < 0.001; n = 3) and a 80.4 ± 4.5% inhibition of the rate of reacidification in response to reintroduction of extracellular Cl − (p < 0.001; n = 3).
We next investigated whether mAE2 transiently expressed in HEK-293T cells was regulated in a similar fashion to human AE2 expressed in Calu-3 and HNE cells. Figure 11A-D shows that mAE2 activity in HEK-293T cells was reduced by both BAPTA-AM and J-8, as well as TBB and CX4945, demonstrating that both Ca 2+ /CaM signalling and CK2 regulate mAE2 when exogenously expressed in HEK-293T cells. In addition, short-term exposure to CX4945 also inhibited CK2 catalytic activity in HEK-293T cell lysates, demonstrating that the effect of CX4945 on mAE2 activity was very likely a result of CK2 inhibition (see Fig. 13B).
Knockout of CK2 catalytic subunits reduces mAE2 activity in HEK-293T cells To further investigate CK2 regulation of mAE2, we studied the activity of mAE2 in genetically altered HEK-293T cells. CK2 is a tetramer consisting of two catalytically active α subunits and two β subunits. We therefore genetically knocked out either the α catalytic subunit or the α prime catalytic subunit of CK2 (αCK2-KO and αprimeCK2-KO HEK-293T cells; see the 'Methods' section and supplementary Fig. 2). Note that α and α prime isoforms are highly homologous proteins with large overlapping functions and we found that CK2 catalytic activity was reduced by~50% in αCK2-KO cells (Fig. 13B). In both types of CK2 KO cells, mAE2 activity was significantly reduced compared to WT HEK-293T cells, with the mean pH i increase in response to removal of extracellular Cl − reduced by 24.6 ± 6.4% in αCK2-KO cells (p < 0.05; n = 7; Fig. 12A) and 24.2 ± 4.0% in αprimeCK2-KO cells ( p < 0.05; n = 6; Fig. 12A) while the rate of reacidification after extracellular Cl − readdition was reduced by 42.0 ± 5.9% in αCK2-KO cells (p < 0.01; n = 7; Fig. 12B) and 49.4 ± 5.5% in αprimeCK2-KO cells (p < 0.01; n = 6; Fig. 12B). Furthermore, transfecting αCK2 KO HEK-293T cells with α-CK2 (to restore α CK2 activity) significantly recovered mAE2 activity in co-transfected cells, compared to 'control' co-transfected cells (i.e. mAE2 and empty plasmid) (Fig. 12C, D). Finally, co-transfecting normal HEK-293T cells with a double α catalytic CK2 mutant subunit with reduced sensitivity to TBB (DM CK2; see the 'Methods' section) [35,65] led to a significant decrease in TBB inhibition of Fig. 8 CK2 inhibition abolishes AE2 activity in primary human nasal epithelia: Well-differentiated, primary human nasal epithelia were isolated and cultured as described in the 'Methods' section. The activity of the basolateral Cl − /HCO 3 − exchanger was assessed by measuring pH i changes in response to replacement of basolateral Cl − with gluconate. The effect of CX4945 treatment (10 μM; 5 min) and reversibility on the mean change in pH i caused by basolateral Cl − removal is shown. Note that in these experiments, because the change in pH i induced by removal of basolateral Cl − was relatively small, it was difficult to obtain accurate rates of reacidification after Cl − readdition, particularly for CK2-treated cells, and therefore these data have not been included. *Significant effect of CX4945 treatment vs. control (p < 0.05); **Significant effect of CX4945 wash off (recovery) vs. CX4945 (p < 0.01). Data represents mean ± S.E.M., n = 4-5 from two donors mAE2 activity, compared to control cells (Fig. 12E, F). Taken together with the pharmacological data, our results clearly indicate that CK2 plays a critical role in the regulation of mAE2 activity under resting conditions. These data are also consistent with the findings from Calu-3 cells and primary HNE cells, which provides further support that the identity of the basolateral Cl − /HCO 3 − exchanger in airway epithelial cells is hAE2 (SLCA42). Thus, our results have uncovered a key regulatory mechanism for mammalian AE2 by CK2.

Discussion
SLC4A2 has been identified on the basolateral membrane of human airway epithelia [1,32] yet its regulation still remains poorly understood. Although the exact role of SLC4A2 in the conducting airways in general, including Calu-3 cells, still needs to be determined, it is likely to be important in regulating intracellular pH via its ability to transport HCO 3 − across the basolateral membrane, as shown in this study and by others [14,20,24,40]. Furthermore, because it transports Cl − under resting conditions, AE2 will also act to accumulate Cl − inside the cell, particularly when working in parallel with the sodium/potassium-dependent chloride (NKCC1) cotransporter that facilitates the influx of Na + , K + and 2Cl − ions across the basolateral membrane of Calu-3 cells [31]. Huang et al. [18] have shown AE2 knockdown in Calu-3 cells reduced transepithelial Cl − secretion by~60% whereas the NKCC1 inhibitor bumetamide reduced Cl − secretion bỹ 20%. These results demonstrate the importance of AE2 in transepithelial Cl − secretion. Here, we have explored the cellular mechanisms that regulate the activity of the basolateral Cl − /HCO 3 − exchanger in Calu-3 cells.
We demonstrated that the IC 50 for DIDS inhibition of the basolateral Cl − /HCO 3 − exchanger was approximately 17 μM, which is in good agreement with the IC 50 of~13 μM reported by Humphreys et al. [19] for DIDS inhibition of human AE2 heterologously expressed in Xenopus oocytes. However, given that 500 μM DIDS was required to fully inhibit the exchanger suggested that other DIDS-sensitive transporters might also be present in the basolateral membrane of Calu-3 cells. This could be other SLC4 family members, such as SLC4A9 (AE4) which has recently been demonstrated in mouse submandibular gland acinar cells [41], or even members of the SLC26 family, such as SLC26A7, which play an important role in HCO 3 − transport across the basolateral membrane of gastric parietal cells [42] and intercalated cells of the outer medullary collecting duct [43]. We observed that the basolateral Cl − /HCO 3 − exchanger was active in nonstimulated Calu-3 cells yet elevations in cAMP, using a variety of different mechanisms, (i.e. activation of transmembrane adenylyl cyclase, inhibition of phosphodiesterases or inhibition of MRP-dependent cAMP efflux) all significantly inhibited the exchanger, demonstrating that cAMP negatively regulates this anion transporter. This cAMP-dependent inhibition of AE2-dependent HCO 3 − efflux at the basolateral membrane would enhance the efficiency of cAMP-stimulated, CFTR-dependent HCO 3 − secretion in Calu-3 cells.
However, the downstream mechanism underlying AE inhibition by cAMP appeared not to involve PKA, as it was insensitive to two different PKA inhibitors, nor did it involve a number of other well characterized cAMP-dependent/PKAindependent signalling pathways, including EPAC and mTOR kinase. Although the present study supports our previous work [14], further investigations into the effect of cAMP on the basolateral Cl − /HCO 3 − exchanger activity demonstrated that some forskolin-induced inhibition of the exchanger was alleviated in the presence of the CFTR inhibitor GlyH-101. These findings are similar, but not identical, to those reported by Huang et al. [18,54] and Kim et al. [24] who failed to demonstrate any effect of cAMP on basolateral AE activity. These authors suggested that the apparent cAMP-dependent inhibition of the basolateral AE activity we observed was due to the activation of CFTR-dependent HCO 3 − efflux across the apical membrane, which 'swamped' the basolateral response. Given that we were able to detect basolateral Cl − /HCO 3 − exchanger activity in forskolin-stimulated cells when CFTR was inhibited lends support to this explanation. However, it is important to consider that the forskolin-stimulated inhibition of AE2 was not fully relieved, indicating there was still an effect of cAMP on the activity of the exchanger. In addition, our data, and those reported by others [18,24,54], do not explain how two different PKA inhibitors failed to affect the cAMPdependent inhibition of the basolateral AE activity, particularly as we have previously shown these PKA inhibitors markedly reduced CFTR-dependent HCO 3 − efflux in Calu-3 cells [14]. Clearly, further work is still required to unravel the role of cAMP in the regulation of AE2-dependent HCO 3 − transport across the basolateral membrane of Calu-3 cells. It was also interesting that GlyH-101 alone affected the rate of reacidification caused by basolateral anion exchanger activity in the absence of cAMP stimulation (Fig. 3B). This effect is likely due to a rise in intracellular [Cl − ], through an inhibition of Cl − efflux by CFTR at the apical membrane, reducing the driving force for Cl − entry at the basolateral membrane (on the reintroduction of external Cl − ), and thereby reducing Cl − /HCO 3 − exchanger activity.
We also demonstrated that the exchanger was Ca 2+sensitive, as BAPTA-AM and depletion of thapsigarginsensitive intracellular Ca 2+ stores all markedly reduced the activity of the exchanger. One possible explanation for the effect seen with Ca 2+ store depletion could involve a rise in cAMP, via the activation of storeoperated cAMP signalling (socAMPs) as described by Lefkimmiatis et al. [30], in which the endoplasmic reticulum Ca 2+ -sensor STIM couples ER Ca 2+ levels to cAMP production, via activation of transmembrane adenylyl cyclase 8. However, it is worth noting that in these intracellular Ca 2+ -depleted conditions, we observed that cAMP-stimulated Cl − /HCO 3 − exchange at the apical membrane was also reduced which argues against socAMPs being activated by these conditions in Calu-3 cells (unpublished observations    HEK-293T cells that were also co-transfected with WT CK2. Anion exchanger activity was assessed by measuring pH i changes in response to replacement of extracellular Cl − with gluconate in the presence of 25 μM DIDS to inhibit endogenous Cl − /HCO 3 − exchange activity. *Significant effect of WT CK2 expression (p < 0.05). Data represents mean ± S.E.M., n = 4. e and f display the percent inhibition of mAE2 activity by TBB in cells transfected with a CK2 double mutant (DM CK2). ***Significant difference between DM CK2 and WT CK2 (p < 0.001). Data represents mean ± S.E.M., n = 7-9 data obtained in Calu-3 cells, mAE2 activity in HEK-293T cells was also reduced by chelation of intracellular Ca 2+ and inhibition of CaM. This is consistent with findings from Chernova et al. [8] who showed that both BAPTA-AM and the CaM inhibitor, calmidazolium, significantly reduced Cl − transport by murine AE2, when expressed in Xenopus oocytes, findings also reported by Stewart et al. [60]. Collectively, these results indicate that under resting conditions, CaM maintains normal AE2 activity potentially through a Ca 2+ -dependent pathway. Finally, we also demonstrated that pharmacological inhibition, or genetic knockout of CK2, significantly reduced mAE2 activity, consistent with results from Calu-3 and HNE cells. Because CK2 catalytic activity was reduced by~50% in HEK-293T KO cells (Fig. 13B) and that mAE2 activity was reduced by a similar amount, (Fig. 12B) provides further evidence that CK2 plays a major role in the regulation of AE2 in airway epithelial cells. It was interesting to observe that the degree of inhibition by CX4945 was significantly less for mAE2 expressed in HEK-293T cells compared to that observed for the exchanger in Calu-3 cells (compare Figs. 6 with 11). This could be explained by the 'addiction' phenomenon (reviewed by Venerando et al. [66]) occurring in tumour cell lines, such as Calu-3, in which the level of CK2 is higher than in normal cells such as HEK-293T. Consequently, highly specific inhibition of CK2 by CX4945 in Calu-3 cells causes more pronounced downstream effects than those observed in HEK-293T cells. An interesting result in support of this hypothesis is shown in Fig. 13, where the percent inhibition of CK2 by CX4945 is significantly higher in Calu-3 cells compared to HEK-293T cells (69.6 ± 1.6% vs. 15.4 ± 3.0%; p < 0.001; n = 4; compare Fig. 13A, B), indicating Calu-3 cells have more active CK2 than HEK-293T cells.
In summary, we have demonstrated for the first time that human airway epithelial cells express a basolateral DIDS-sensitive, Cl − /HCO 3 − exchanger which is regulated by CK2. Since mouse slc4A2 was regulated in an identical way when studied in a heterologous expression system suggests that the identity of the transporter in human airway cell was SLC4A2. Therefore, our findings identify CK2 as a new regulator of SLC4A2dependent anion transport in human airway epithelia and suggest that CK2-dependent phosphorylation of SLC4A2, or an associated regulatory protein such as CaM, is essential for AE2 activity under resting conditions. We suggest that maintenance of basal AE2 activity will ensure efficient Cl − loading into the cell prior to cAMP-stimulated anion secretion, which will help support net transepithelial HCO 3 − and fluid secretion in airway epithelial cells [18]. Indeed, since CK2 also positively regulates the activity of CFTR suggests that this protein kinase plays an essential role in coordinating the activity of HCO 3 − transporters at both the apical and basolateral membranes of airway epithelial cells. Compliance with ethical standards Ethical approval for collection of paediatric human nasal epithelial cells was granted by the NRES Committee North East -Newcastle and North Tyneside 1 (REC reference: 15/NE/0215) and informed written parental consent was obtained.
Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.