Stimulation of Cl- secretion via membrane-restricted Ca2+ signaling mediated by P2Y receptors in polarized epithelia.

Extracellular nucleotides such as ATP have been shown to regulate ion transport processes in a variety of epithelia. This effect is mediated by the activation of plasma membrane P2Y receptors, which leads to Ca(2+) signaling cascade. Ion transport processes (e.g. activation of apical calcium-dependent Cl(-) channels) are then stimulated via an increase in [Ca(2+)](i). Many polarized epithelia express apical and/or basolateral P2Y receptors. To test whether apical and basolateral stimulation of P2Y receptors elicit polarized Ca(2+) signaling and anion secretion, we simultaneously measured the two parameters in polarized epithelia. Although activation of P2Y receptors located at both apical and basolateral membranes evoked an increase in [Ca(2+)](i), only apical P2Y receptors-coupled Ca(2+) release stimulated an increase in anion secretion. Moreover, the calcium influx evoked by apical and basolateral P2Y receptor stimulation is predominately via the basolateral membrane domain. It appears that the apical P2Y receptor-regulated Ca(2+) release and activation of apical Cl(-) channels is compartmentalized in polarized epithelia with basolateral P2Y-stimulated Ca(2+) release failing to activate anion secretion. These data suggest that there may be two distinct ATP-releasable Ca(2+) pools, each coupled to apical and basolateral membrane receptor but linked to the same calcium influx pathway located at the basolateral membrane.

Extracellular nucleotides such as ATP and UTP have been documented in a large number of cell types to elicit a variety of biological responses through binding to specific membrane receptors termed P2Y receptors (1,2). P2Y receptor stimulation results in the activation of Ca 2ϩ -dependent Cl Ϫ secretion in various types of epithelia, such as human airway epithelia (3), rat sertoli cells (4), mouse cortical-collecting duct cells (5), and both human and equine sweat gland epithelia (6,7). In many cases, the P2Y receptors are located on the apical membranes. It has been proposed that ATP and possibly other nucleotides may be released across the apical membrane of stimulated epithelia (8 -10). The released nucleotides activate the P2Y receptors either in the membranes of the nucleotide-releasing cell itself or of adjacent cells. Therefore, it is possible that these apical P2Y receptors may act as autocrine and/or paracrine regulators of epithelial transport (11,12).
There is, however, a paucity of information on the mechanisms of stimulus-secretion coupling in polarized epithelia that ex-pressed both apical and basolateral P2Y receptors. It is now emerging that certain types of P2Y receptor (e.g. P2Y2) are coupled to multiple signal transduction pathways in which changes in [Ca 2ϩ ] i is not obligatory (13;14). More importantly, the functional expression of at least some P2Y receptor subtypes appears to be a feature of the polarized phenotype. These receptors may not be detected in isolated non-polarized cells that have been grown on glass coverslips. Therefore, simply using the methods traditionally applied to single cells (e.g. microspectrofluorimetry, patch clamping) may fail to reveal how extracellular nucleotides regulate anion secretion in polarized epithelia.
In this study, a technique that allows us to monitor nucleotide-evoked short-circuit current (I SC ) 1 and [Ca 2ϩ ] i concurrently in a polarized epithelium was employed. The Ca 2ϩmediated Cl Ϫ secretion activated by various nucleotides was explored in an equine sweat gland epithelial cell line with emphasis on the polarized Ca 2ϩ mobilization and influx in response to apical versus basolateral P2Y receptor stimulation. Previous studies have shown that the cultured cells will form polarized monolayer on permeable support (15) and express both apical purine and pyrimidine receptors (7,16) that are linked to internal Ca 2ϩ release and capacitative calcium entry (17).

EXPERIMENTAL PROCEDURES
Culture of Cells-Experiments were undertaken using a spontaneously transformed cell line (E/92/3) derived from the secretory epithelium of the equine sweat gland (18). Cells were grown as described previously (15) on Transwell-col membranes (Costar, Cambridge, MA) with 0.4-m pore diameter (culture area 0.1 cm 2 ). Cells reach confluence after 4 days with a resistance greater than 150 ohm cm 2 .
Simultaneous Measurements of [Ca 2ϩ ] i and I SC -Nucleotide-evoked calcium signals and anion secretion were measured simultaneously as described previously (6,13,14,16). Cells grown on Transwell-col membrane were loaded with Fura-2 by incubation (45 min at 37°C) in bicarbonate-buffered saline containing 3 M Fura-2-acetoxymethylester and 1.6 M pluronic F127. The membranes bearing Fura-2-loaded epithelia were then mounted in a miniature Ussing chamber (19) attached to the stage of an inverted microscope (Nikon TE300). The cells were then viewed with a ϫ40 extra long working distance objective (Nikon CFI Plan Fluor ELWD, 0.6 numerical aperture), and the Fura-2 fluorescence ratios were recorded (PTI Ratio-Master fluorescence system, Photon Technology International, NJ) from an optical field containing 30 -40 cells in the center of the epithelium. The potential difference was clamped to 0 mV, and short-circuit current (I SC ) was simultaneously measured using a voltage-clamp amplifier (VCC 600, Physiologic Instruments, San Diego, CA). Both signals were digitized and recorded to computer hard disc. Positive currents are displayed as upward deflections of the traces and were defined as those carried by anions moving from the basolateral to the apical compartments. A transepithelial potential difference of 1 mV was applied periodically, and the resultant * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed. Before use, this nucleotide (10 mM) was incubated (1 h at 37°C) in Hepes-buffered saline containing hexokinase (10 IU ml Ϫ1 , Roche Molecular Biochemicals) and 22 mM D-glucose to remove contaminating nucleotide triphosphates. The resulting solution was then aliquoted and stored at Ϫ20°C (20). Uridine triphosphate and adenosine triphosphate were obtained form Pharmacia as ultrapure solutions. Thapsigargin was from Calbiochem (La Jolla, CA). All other general laboratory reagents were from Sigma, and cell culture reagents were from Invitrogen.
Data Analysis-Experimentally-induced changes (⌬) in Fura-2 fluorescence ratio and I SC were quantified by measuring each parameter at the peak of a response and subtracting the equivalent values measured immediately prior to stimulation. Pooled data are presented as the means Ϯ S.E., and values of n refer to the number of experiments in each group. The significance of any differences between mean values was tested using Student's t test with p Ͻ 0.05 considered significant.

Basolateral ATP Activates an Increase in [Ca 2ϩ ] i but Not I SC -
Our previous studies demonstrated that the epithelia express apical purine and pyrimidine receptors (7). ATP elevates [Ca 2ϩ ] i and anion secretion through the activation of IP 3 -mediated calcium signaling pathway, apical CaCC, and basolateral Ba 2ϩ -sensitive K ϩ channels (15). In contrast, basolateral application of ATP could only elicite a small and variable increase in I SC (7). Therefore, it was concluded that the P2Y receptors are essentially confined to the apical membrane. In this study, however, when [Ca 2ϩ ] i and I SC were monitored simultaneously, basolateral ATP was capable of increasing [Ca 2ϩ ] i (Fig. 1A). Fig. 1 shows the representative tracings depicting simultaneous measurements of [Ca 2ϩ ] i (Fig. 1A) and I SC (Fig. 1B). The 340/380 nm fluorescence ratio has been used to represent changes in [Ca 2ϩ ] i . The transient upward deflections shown in each I SC tracing were produced by intermittently clamping the potential difference at 1 mV as described under "Experimental Procedures." When ATP was first added to the apical solution, it led to an increase in both [Ca 2ϩ ] i and I SC . Subsequently, apical ATP was removed, and basolateral ATP was applied. Although basolateral application of ATP elicited an increase in [Ca 2ϩ ] i , it was only accompanied by a small increase in I SC (n ϭ 8). Similar results were obtained when ATP was administered in the reverse sequence (n ϭ 8, data not shown).
Concentration-dependent Effect of Apical or Basolateral ATP, UTP, and UDP on [Ca 2ϩ ] i and I SC -The effects of ATP, UTP, and UDP on [Ca ϩ2 ] i and I SC were compared between the stimulation of P2Y receptors located on apical and basolateral membrane. Brief pulse (30 s) 1-300 M ATP (n ϭ 8) or UTP (n ϭ 8) elicited concentration-dependent increases in [Ca 2ϩ ] i when applied to either apical or basolateral aspects of the epithelia ( In contrast, the I SC responses of the basolateral ATP and UTP were very small (Fig. 2, B and D), so it was not possible to estimate the EC 50 on basolateral side. These data indicate that both apical and basolateral membrane express functional P2Y receptors, and apical nucleotides were more potent in evoking a calcium response. The P2Y receptor number and/or efficiency may be greater on the apical surface as compared with the basolateral membrane. However, it is important to note that at maximal concentration of nucleotides used the increase in [Ca 2ϩ ] i , level were not distinguishable between apical and basolateral sides (Fig. 2, A and C). However, the increase in [Ca 2ϩ ] i evoked by basolateral P2Y receptors stimulation did not evoke any substantial increase in I SC (Fig. 2, B and D).
Apical and Basolateral ATP Activate Separate Releasable Ca 2ϩ Pools-The efficient stimulation of anion secretion by apical nucleotides could signify localized regulation of Ca 2ϩ release and activation of CaCC. To determine whether apical and basolateral P2Y receptor stimulation releases Ca 2ϩ from the same internal pool, experiments were undertaken in which external Ca 2ϩ was removed from the solution. Fig. 3 shows the effects of apical and/or basolateral ATP on [Ca 2ϩ ] i and I SC in monolayers perfused with calcium-free solution. ATP (100 M) was first added to the apical solution, and this led to an increase in both [Ca 2ϩ ] i (Fig. 3A) and I SC (Fig. 3B). After [Ca 2ϩ ] i and I SC had returned to basal level, ATP was applied to the basolateral side in the continuous presence of apical ATP. This led to a second increase in [Ca 2ϩ ] i but a minimal increase in I SC . In control experiments, after the first apical (or basolateral) ATP response, a second addition of 100 M ATP to the same side did not evoke any discernible increases in [Ca 2ϩ ] i (n ϭ 6). The identical protocol was used, but the sequence of ATP addition was reversed. Despite the internal store having been previously depleted by basolateral ATP under Ca 2ϩ -free condition, apical ATP stimulated an increase in both [Ca 2ϩ ] i (Fig. 3C) and I SC (Fig. 3D). Summarized data were shown in Fig. 3, E and F. A change in [Ca 2ϩ ] i (Fig. 3E) and I SC (Fig. 3F) was quantified in control monolayers and in monolayers pretreated with apical or basolateral ATP before the activation of the P2Y receptors localized in the contralateral membrane. When compared with the control responses of apical ATP (⌬ratio ϭ 0.86 Ϯ 0.13, ⌬ I SC ϭ 57.63 Ϯ 4.61 Acm Ϫ2 , n ϭ 9), pretreating the epithelia with basolateral ATP reduces the Ca 2ϩ and I SC responses to apical ATP (⌬ratio ϭ 0.55 Ϯ 0.08, ⌬ I SC ϭ 30.79 Ϯ 3.85 Acm Ϫ2 , n ϭ 8, p Ͻ 0.05). However, the magnitude of the basolateral ATP-induced [Ca 2ϩ ] i and I SC responses are not affected by a previous exposure of the apical membrane to ATP (Control, ⌬ratio ϭ 0.31 Ϯ 0.04, ⌬ I SC ϭ 3.70 Ϯ 0.73 Acm Ϫ2 , n ϭ 8 versus pretreated; control, ⌬ratio ϭ 0.39 Ϯ 0.11, ⌬ I SC ϭ 1.76 Ϯ 0.30 Acm Ϫ2 , n ϭ 8, p Ͼ 0.05). It appears that the apical ATPreleasable calcium pool and the basolateral ATP-releasable calcium pool are distinct but partially overlapped. However, ATP applied apically can release Ca 2ϩ from the basolateral pool but not vice versa.
Effect of Store Depletion by ATP and Thapsigargin-In most non-excitable cells, the emptying of internal stores would lead to opening of plasma membrane calcium channels and hence CCE. We next tested whether the depletion of internal stores by apical and basolateral ATP would lead to Ca 2ϩ influx pathway(s) that is confined to the same membrane domains containing the P2Y receptors. In the first series of experiments, the relationship between ATP-stimulated CCE and anion secretion was investigated. The experimental protocol used involved stimulating the epithelia with apical ATP in a Ca 2ϩ -free solution to deplete the internal Ca 2ϩ store after which 2.5 mM Ca 2ϩ was added to either apical or basolateral perfusing solution to detect Ca 2ϩ influx (Fig. 4). In Fig. 4, A and B . Fig. 4, C and D, summarizes the data. In Fig. 4C, the change in Fura-2 ratio due to apical calcium influx was 0.58 Ϯ 0.06, whereas that of basolateral calcium influx was 1.21 Ϯ 0.12. The mean values of I SC activated by apical calcium influx and basolateral calcium influx were 5.29 Ϯ 0.50 Acm Ϫ2 and 18.42 Ϯ 0.91 Acm Ϫ2 , respectively (n ϭ 8). Reversing the sequence of Ca 2ϩ , the readdition (i.e. basolateral followed by apical calcium) showed similar results (n ϭ 8, data not shown). These results indicate that in polarized epithelia, the stimulation with apical ATP leads to the activation of transmembrane Ca 2ϩ influx pathway, which is located mainly in the basolateral membrane. The Ca 2ϩ influx then stimulated anion secretion across the epithelia.
To test whether basolateral ATP also activated a similar Ca 2ϩ influx, an identical protocol was used but ATP (100 M) was delivered to the basolateral side. When extracellular Ca 2ϩ was restored in the apical side and then to the basolateral side with the continuous presence of basolateral ATP, the increase in Fura-2 ratios were 0.15 Ϯ 0.03 and 0.46 Ϯ 0.04 (n ϭ 8) for apical and basolateral Ca 2ϩ influx, respectively. Similar results were obtained when the sequence of Ca 2ϩ readdition to the perfusate was reversed (n ϭ 7, data not shown). It was clearly shown that basolateral Ca 2ϩ influx caused a larger increase in Fura-2 ratio than the apical Ca 2ϩ influx, similar to that of apical ATP. However, emptying the apical ATP-releasable pool produced three times more Ca 2ϩ influx through the basolateral membrane than that of basolateral ATP.
Thapsigargin (Tg) was used as another pharmacological tool to deplete the internal calcium pool and to bypass the receptors and the subsequent signaling cascade. Thapsigargin inhibits the endoplasmic reticular Ca 2ϩ -ATPase, which pumps back the calcium into the internal store and discharges the intracellular calcium store (21). In the absence of extracellular Ca 2ϩ , Tg (3 M) induced the emptying of intracellular Ca 2ϩ stores, and calcium influx occurred upon restoring extracellular Ca 2ϩ in the apical side followed by the basolateral side (Fig. 5, A and B). Basolateral Ca 2ϩ influx induced a larger increase in both Fura-2 ratio and I SC than the apical Ca 2ϩ influx. The increase in fluorescence ratios due to apical and basolateral calcium influx were 0.50 Ϯ 0.13 and 1.33 Ϯ 0.28, respectively (Fig. 5C). The mean value of I SC activated by apical and basolateral calcium influx were 7.39 Ϯ 1.06 Acm Ϫ2 and 27.23 Ϯ 3.35 Acm Ϫ2 (n ϭ 5), respectively (Fig. 5D). Similar results were obtained when Ca 2ϩ was added to the basolateral followed by the apical side after the store was depleted by Tg (n ϭ 5, data not shown). These results indicate that in these cells, the emptying of internal store by Tg leads to the activation of transmembrane Ca 2ϩ influx pathway, which is also located mainly in the basolateral membrane.
To test whether apical ATP-activated and basolateral ATPactivated Ca 2ϩ influx through the basolateral membrane are additive, the epithelia were stimulated with apical and/or basolateral ATP under Ca 2ϩ -free condition, and Ca 2ϩ was added back to the basolateral perfusate subsequently. The increase in Fura-2 ratio due to Ca 2ϩ influx after the depletion of both apical and basolateral ATP-releasable pools was 0.37 Ϯ 0.06 (n ϭ 6), which is not statistically different from emptying either the apical or basoalteral ATP-releasable pool (p Ͼ 0.05). These data suggest that separate pools are coupled to a common CCE pathway(s). DISCUSSION As in other epithelia, P2Y receptors expressed in equine sweat gland epithelia are linked through G proteins to phospholipase C, resulting in the generation of IP 3 and hence the mobilization of Ca 2ϩ from internal stores (22). The emptying of internal stores activates a persistent Ca 2ϩ influx through plasma membrane (17). This type of Ca 2ϩ influx has been termed "capacitative Ca 2ϩ entry" or more recently "store-operated Ca 2ϩ entry" (23)(24)(25). Previous studies have identified the functional expression of apical purinoceptors (P2Y2) and pyrimidinoceoptors (P2Y4 and/or P2Y6) when the cells are grown on permeable supports. P2X receptors are not involved in mediating the nucleotide responses (16). This cell line also represents the simplest calcium-dependent secretory system in which cAMP-dependent Cl Ϫ secretion (i.e. cystic fibrosis transmembrane conductance regulator function) is absent (15). Transepithelial secretion can only be activated by calciummobilizing agonists, such as extracellular ATP, UTP, UDP, and bradykinin (16).
When grown on permeable support, epithelial cells usually adopt a polarized phenotype in which secretion is mediated by the complex coordination of various receptors, ion channels, Ca 2ϩ influx pathway, and others located on specialized cellular domains. Our previous findings suggest that it is important to measure both parameters, namely [Ca 2ϩ ] i and anion secretion simultaneously in a polarized epithelium. Firstly, it is because the expression of certain P2Y receptor subtypes does only occur when the epithelial cells are cultivated on permeable supports, which allows polarized differentiation (14,16,26). This is supported by other studies (27,28) showing prominent changes of P2Y receptor expression as a function of short term culturing of salivary gland cells. Down-regulation of UDP-activated receptor, probably P2Y6, during culture on glass coverslips has also been suggested in tracheal epithelial cells isolated from P2Y 2receptor-deficient mice (29). Secondly, the activation of Cl Ϫ secretion by extracellular ATP can occur without the involvement of [Ca 2ϩ ] i , i.e. the Ca 2ϩ -independent pathway (30 -32). Our recent study also suggests such a Ca 2ϩ -independent regulation of anion secretion by extracellular ATP exists in equine sweat gland and colonic epithelia (13,14). Therefore, to better correlate the role of intracellular Ca 2ϩ (both Ca 2ϩ release and influx) with secretory activity in polarized epithelia, we adopted a technique to monitor both parameters simultaneously.
A previous study (7) using conventional Ussing chamber devices demonstrates that basolateral application of ATP could only elicit a small and variable I SC . Therefore, it was concluded that the P2Y receptors are essentially confined to the apical membrane (7). In this study in which both [Ca 2ϩ ] i and I SC were monitored simultaneously, it has been shown that both sides of the epithelium contained P2Y receptors coupled to internal Ca 2ϩ release. However, only apical ATP stimulation can exert an effect on anion secretion even though high concentrations of basolateral ATP could elicit an increase in [Ca 2ϩ ] i similar to that of apical stimulation. Other nucleotides (UTP and UDP) elicited the same asymmetric patterns. Differing effects of apical versus basolateral ATP on ion transport have been reported in other epithelia such as rat epididymis (33). In epididymal cells, apical but not basolateral ATP stimulated Cl Ϫ transport. However, the two parameters, namely [Ca 2ϩ ] i and Cl Ϫ secretion, were not measured simultaneously in most of the studies. Therefore, basolateral P2Y receptors might exist in these epithelia. However, it was not detected when secretory activity was used as the only functional read-out of receptor stimulation. Moreover, the receptor-coupled second messenger pathway such as [Ca 2ϩ ] i was not measured simultaneously.
With the simultaneous measurement technique, the differential effects of apical versus basolateral ATP on anion secretion and intracellular calcium release can be examined in detail. First, it is interesting to find out whether unilateral ATP administration stimulated P2Y receptors and tapped internal Ca 2ϩ pools that are associated with the plasma membrane ipsilateral but not contralateral to the stimulated receptors. The results show that even pretreated with apical ATP under Ca 2ϩ -free condition, basolateral ATP can still release Ca 2ϩ from internal pools, but this was not accompanied by a substantial increase in anion secretion. It appears that the apical and basolateral ATP-releasable internal Ca 2ϩ pools are separate. It may be that separate Ca 2ϩ pools are associated with apical or basolateral P2Y receptors. Thus, the signal generated in the apical membrane can access only the pool coupled to the apical membrane. The resultant localized release of Ca 2ϩ then activates adjacent Cl Ϫ channels that are also located on the same membrane domain. The data are conceptually in agreement with the explanation about the inability of basolateral ATP to activate anion secretion. Recently, similar results were obtained in normal and cystic fibrosis airway epithelia (34). By using similar technique, Paradiso and his co-workers (34) found that intracellular Ca 2ϩ regulation of the Ca 2ϩ -sensitive anion conductance via CaCC is compartmentalized in both cystic fibrosis and normal airway epithelia with basolaterally released Ca 2ϩ failing to activate CaCC in both epithelia (34). However, it is equally possible that ATP causes a localized release of Ca 2ϩ from the same Ca 2ϩ store. Further experiments have to be conducted to clarify this point. Nonetheless, the basolateral ATP-releasable Ca 2ϩ pool was not accessible by apical P2Y receptor stimulation. Pretreating the epithelia with Tg under Ca 2ϩ -free condition also fails to affect the mobilization of Ca 2ϩ by basolateral ATP. 2 In pancreatic and salivary gland cells, the internal Ca 2ϩ pool is highly compartmentalized and that compartmentalization is achieved in part by polarized expression of Ca 2ϩ channels (35). On the contrary, it has been demonstrated that the endoplasmic reticulum acts as one continuous Ca 2ϩ pool in pancreatic acinar cells, and Ca 2ϩ released from the apical endoplasmic reticulum terminals is quickly replenished from the bulk of the rough endoplasmic reticulum at the base (36,37).
Another possibility to explain the inability of basolateral ATP to activate onion secretion is that Ca 2ϩ may not be the "ultimate" signaling molecule that is solely responsible for the activation of Cl Ϫ channels. Once the P2Y receptor is stimulated by ATP, it will activate phospholipase C to breakdown phosphatidylinositol bisphosphate into IP 3 and diacylglycerol. Therefore, other potential signaling molecules such as diacylglycerol or protein kinase C along the phospholipase C pathway may also be involved in the stimulation of CaCC. Basolateral ATP, in contrast, may not be able to generate the signaling molecule(s) (or the concentration is too low) required for channel activation. Finally, basolateral ATP may activate an unidentified inhibitory pathway that suppresses Cl Ϫ secretion. It is important to undertake further experiments to establish the mechanism of this differential effect as it may be an important component of a poorly understood mechanism that permit ion transport.
In polarized human nasal epithelia, it has been demonstrated that extracellular ATP stimulated a membrane-restricted regulation of Ca 2ϩ release and influx (38). Ca 2ϩ influx of similar magnitude was demonstrated on both sides of the epithelium. The activation of plasma membrane Ca 2ϩ influx by ATP was confined to the membrane ipsilateral to receptor stimulation. This finding demonstrates that the regulation of CCE could be a membrane-restricted phenomenon. Both internal Ca 2ϩ release and activation of the Ca 2ϩ influx pathway were confined to the membrane of receptor activation. In contrast, although the ATP-activated Ca 2ϩ release is membranerestricted in equine sweat gland epithelia, Ca 2ϩ influx is not. Experiments in which the epithelia were stimulated with apical or basolateral ATP and external Ca 2ϩ was removed and/or replaced confirmed that the nucleotide-evoked [Ca 2ϩ ] i signals were initiated by the mobilization of cytoplasmic Ca 2ϩ and followed by Ca 2ϩ influx (i.e. CCE). Interestingly, although the receptors controlling Ca 2ϩ entry were on the apical or basolateral membrane, the Ca 2ϩ influx occurred primarily across the basolateral membrane. In further experiments, thapsigargin was used to deplete internal Ca 2ϩ stores, and similar results were obtained.
As a result, in the equine sweat gland epithelia, the storeoperated Ca 2ϩ entry pathway is primarily located in the basolateral membrane and can be activated by apical and basolateral ATP or by receptor-independent depletion of intracellular Ca 2ϩ stores with thapsigargin. It is interesting that similar results were also found in polarized MDCK-C7 cells in which the Ca 2ϩ influx, which could be activated by extracellular ATP or thapsigargin, was exclusively located in the basolateral membrane (39). Recently, it has been suggested that the localization of CCE in the basolateral membrane is a more general characteristic of all polarized epithelial tissues. For example, recent data from human colonic T84, bronchial epithelial cells 16HBE14 Ϫ (40), and human airway epithelial Calu-3 cells (41) also support the finding that CCE entry pathway in polarized epithelial cells may be confined to the basolateral side. How does the activation of apical P2Y receptors induce a CCE located in the contralateral membrane? There are two leading hypotheses about the activation of CCE, namely "diffusible messenger" and "conformational coupling" (23)(24)(25). According to the former hypothesis, the CCE could be activated by a diffusible messenger, the so called "calcium influx factor" (42), that is generated by the emptying of apical or basolateral ATP-releasable calcium pools. Alternatively, both apical or basolateral receptor-activated Ca 2ϩ pools have to be in direct contact with the neighboring basolateral membrane in which the Ca 2ϩ influx channels are located. Recent findings strongly suggest that the control of CCE involves the interaction between IP 3 receptors in the internal stores and one of the candidates for plasma membrane calcium channels, the hTrp3 (human transient receptor potential 3) protein (43)(44)(45)(46).
In summary, this study indicates that P2Y receptor, calcium signals, and activation of Cl Ϫ channels are compartmentalized in the apical membrane domain. Recent studies also show a "membrane-delimited" cAMP signaling in the apical membrane of airway epithelial cells (47) and regulation of Na ϩ -H ϩ exchanger isoforms by P2 receptors in the rat submandibular gland duct (48). Therefore, without a global activation of intracellular signaling pathway(s), selective activation of apical P2Y receptors could lead to fluid and electrolyte secretion. This may be useful and important as a global increase in Ca 2ϩ (or other cellular messengers) is detrimental and affects many other cellular functions.