INTRODUCTION

The regulation of intestinal epithelial chloride transport is under the control of neural, humoral, and immune-related mechanisms (1). As the active transepithelial movement of chloride into the gastrointestinal lumen is an important mechanism governing the passive movement of water, breakdown in the regulation of this process can lead to excessive secretory responses, resulting in diarrhea. Our laboratory has focused on understanding the intracellular mechanisms responsible for the regulation of chloride secretion.

Chloride secretion is positively regulated via two predominant pathways, utilizing cyclic nucleotides or calcium as second messengers, respectively (1). Calcium-dependent secretion can be evoked experimentally by agents such as the muscarinic agonist, carbachol, histamine, or the calcium ATPase inhibitor, thapsigargin. All of these stimuli elevate cytoplasmic calcium concentrations, which in turn evokes secretion. Moreover, we have shown that calcium-dependent chloride secretion is also subject to a number of negative regulatory influences. Thus carbachol, while itself serving as an initial agonist of chloride secretion, can also subsequently inhibit calcium-dependent chloride secretion, such as that stimulated by thapsigargin (2). This inhibition occurs at a step distal to the rise in intracellular calcium and has been attributed to a putative negative messenger, inositol(3,4,5,6)tetrakisphosphate (D-Ins(3,4,5,6)P₄)¹ (3). This inositol polyphosphate may directly block calcium-activated chloride channels (4), thereby inhibiting the overall process of transepithelial chloride secretion. The peptide growth factor, epidermal growth factor (EGF), also exerts an inhibitory effect on calcium-dependent chloride secretion but does not itself act as a secretagogue (5). The inhibitory effect of EGF displays some similarities to that evoked by carbachol. Thus, EGF inhibits calcium-activated secretion without affecting the rise in intracellular calcium. EGF also causes a comparatively small increase in DL-Ins(3,4,5,6)P₄ (5). However, we recently demonstrated that this likely does not account for the inhibitory effects of EGF on secretion.² Moreover, simultaneous addition of maximally inhibitory concentrations of EGF and carbachol had a greater inhibitory effect on chloride secretion compared with either agent alone (5). These data suggest that EGF is activating a novel mechanism responsible for the inhibition of calcium-dependent chloride secretion that is distinct from that evoked by carbachol.

Our laboratory has demonstrated a relationship between tyrosine phosphorylation and inhibition of chloride secretion (6). Therefore, we hypothesized that the activation of tyrosine kinase-dependent pathways by EGF is involved in the inhibition of chloride secretion. One signaling pathway regulated by tyrosine phosphorylation in response to EGF, in other cell types, involves the recruitment of phosphatidylinositol 3-kinase (PI 3-kinase) (7). This lipid kinase is responsible for the production of D-3-phosphorylated phosphoinositides. Its activation involves recruitment of the SH2 domains of its 85-kDa regulatory subunit to tyrosine-phosphorylated residues of receptor proteins (erbB3, p120cbl) (8, 9), thus bringing the associated 110-kDa catalytic subunit closer to its membrane substrates. Furthermore, in vitro binding of p85 to tyrosine-phosphorylated residues enhances the activity of p110 (10). The activation of PI 3-kinase and consequent generation of its lipid products has been implicated in cell growth (11), movement (12), vesicular transport (13), glucose uptake (14), and oxidant production (15). Recently, PI 3-kinase has been reported to regulate the activity of the Na⁺/H⁺ exchanger (16, 17). This led us to speculate that PI 3-kinase might also regulate other epithelial transport proteins, such as those involved in chloride secretion. Thus, we sought to determine if EGF activates PI 3-kinase in T₈₄ cells, and, if so, its involvement in inhibition of calcium-dependent chloride secretion by EGF. Furthermore, carbachol has also been shown to increase the amount of tyrosine-phosphorylated proteins in T₈₄ cells (6). We therefore additionally wanted to determine if carbachol also activates PI 3-kinase and mediates any of its inhibitory effects through activation of this enzyme. Finally, a constitutively active phosphatidylinositol-specific 3-kinase (PtdIns 3-kinase) has been detected in yeast (18), and a homologue of this protein may also be present in mammalian cells (19). Thus, we finally wanted to determine if PI 3-kinase activity plays any role in regulating basal chloride secretion.

EXPERIMENTAL PROCEDURES

The following were obtained from the sources indicated: thapsigargin and the PI 3-kinase inhibitor, wortmannin (LC Laboratories, Woburn, MA); carbachol and histamine (Sigma); EGF (Genzyme, Cambridge, MA); polyclonal rabbit anti-p85 and monoclonal mouse antiphosphotyrosine antibodies (UBI, Lake Placid, NY); polyclonal rabbit anti-p110 antibody (Santa Cruz Biotechnology, Santa Cruz, CA); soybean phosphatidylinositol (Avanti Polar Lipids, Alabaster, AL); [-³²P]ATP and ³²PO₄ (DuPont NEN). All other agents were of at least reagent grade and were obtained commercially.

Methods for the maintenance of T₈₄ cells for use in transepithelial electrolyte transport studies have been described previously (20). In brief, T₈₄ cells were grown in Dulbecco's modified Eagle's/F12 media (JRH, Lenexa, KS) with the addition of 5% newborn calf serum. Cells were passaged by trypsinization. For the measurement of chloride secretion, 10⁶ cells were seeded onto collagen-coated polycarbonate filters (Nuclepore, Pleasanton, CA) glued onto Lexan rings as described previously (20). For experiments involving immunoprecipitation and Western blotting, PI 3-kinase assays, or the measurement of 3phosphorylated lipids, 10⁶ cells were seeded onto 30-mm, Millicel-HA Transwells (Millipore, Bedford, MA). Cells seeded onto polycarbonate or Millicel filters were cultured for 7-10 days prior to use.

Chloride secretion was measured as short circuit current (I_sc) across monolayers of T₈₄ cells, mounted in Ussing chambers (window area = 2 cm²) modified for use with cultured cells (20). I_sc (normalized to µA/cm²) was used to quantitate both basal transepithelial Cl secretion and that induced by calcium-dependent secretagogues. T₈₄ cells secrete chloride in response to various calcium-mobilizing agonists, and the resulting changes in I_sc are wholly reflective of chloride secretion (21). I_sc measurements were carried out in Ringer's solution containing (in mM): 140 Na⁺, 5.2 K⁺, 1.2 Ca²⁺, 0.8 Mg²⁺, 119.8 Cl, 25 HCO₃, 2.4 H₂PO₄, 0.4 HPO₄², and 10 glucose.

On the day of the experiment, T₈₄ cells were washed three times in Ringer's solution and allowed to equilibrate for 30 min at 37 °C. Cells were then stimulated as noted. The reaction was stopped by a wash with ice-cold phosphate-buffered saline (PBS). Ice-cold lysis buffer was then added (consisting of PBS, 1% Nonidet P-40, 1 mM NaVO₄, 1 µg/ml leupeptin, and 100 µg/ml phenylmethylsulfonyl fluoride), and the cells were incubated at 4 °C for 30 min. Cells were then scraped into microcentrifuge tubes, centrifuged at 10,000 rpm for 10 min to remove insoluble material, and an aliquot was removed from each sample to determine protein content (Bio-Rad protein assay). Samples were then adjusted such that each contained an equal amount of protein. 5 µg of monoclonal antiphosphotyrosine antibody was then added to each sample and allowed to incubate on ice for 60 min. This was followed by the addition of 50 µl of a 1:1 mixture of Protein A-Sepharose and water, and samples were placed on a rotating platform at 4 °C for 60 min. Samples were then centrifuged to pellet the protein A-Sepharose-antibody-antigen complex, and the complex was washed three times with cold lysis buffer, followed by three more washes with cold PBS. The beads were then resuspended in gel loading buffer (50 mM Tris, pH 6.8, 2% SDS, 100 mM dithiothreitol, 0.2% bromphenol blue, 20% glycerol). Samples were placed in boiling water for 5 min and then loaded onto a 7.5% polyacrylamide gel to resolve proteins. The proteins were transferred from the gel onto a polyvinylidene difluoride membrane (DuPont NEN). The membrane was then blocked with a 1% solution of skim milk in PBS for 30 min, followed by further incubation of the membrane with a 1% skim milk solution including either 5 µl of polyclonal anti-p85 or 10 µg of polyclonal anti-p110 subunits of PI 3-kinase, for 60 min. This was followed by three, 15-min washes with wash buffer (1% skim milk, 0.5% bovine serum albumin, 0.2% Tween 20, in PBS). Following washes, 2 µl of secondary antibody (goat-anti-mouse IgG conjugated to alkaline phosphatase (Clontech, Palo Alto, CA)) was added to the membrane, in the presence of 1% skim milk, and allowed to incubate for an additional 30 min. This was followed by three more washes with wash buffer. The membrane was then treated with a chemiluminescent solution according to the manufacturer's directions (Clontech, Palo Alto, CA) and exposed to film. Densitometric analysis of the blots was performed using a digital imaging system.

Cells were stimulated, lysed, and immunoprecipitated with antiphosphotyrosine as described for immunoprecipitation and Western blotting with the following modifications to the washing of the Protein A-Sepharose pellet. Following centrifugation of the Sepharose-antibody-antigen complex, the complex was washed twice with cold PBS containing 200 µM adenosine, twice with 0.1 M Tris containing 0.5 M LiCl and 200 µM adenosine, and finally, twice with kinase assay wash (20 mM Hepes pH 7.4, 10 mM MgCl₂, and 10 µM cold ATP). The complex was mixed with 20 µg of soybean phosphatidylinositol and 40 µl of kinase buffer consisting of kinase assay wash components and 0.8 mCi/ml [-³²P]ATP. The reaction was allowed to continue for 10 min. The reaction was stopped by the addition of 80 µl of a 1:1 mixture of 2.4 M HCl and methanol. Lipid products were extracted with chloroform and spotted onto thin layer chromatography (TLC) plates impregnated with potassium oxalate, and developed in one dimension in a mobile phase consisting of chloroform, acetone, methanol, acetic acid, and water (80:20:26:24:14:, v/v). Spots corresponding to phosphatidylinositol 3-phosphate were detected using autoradiography and identified on the basis of their co-migration with a known standard. The corresponding regions of the plates were scraped into vials and radioactivity assessed by liquid scintillation counting.

Phospholipid products of PI 3-kinase were measured using a modification of the method of Traynor-Kaplan et al. (22). Cells were washed three times with phosphate-free buffer (in mM: 140 NaCl, 5 KCl, 1.2 CaCl₂, 0.8 MgCl₂, 10 glucose, 20 Hepes, pH 7.4) and allowed to equilibrate for 30 min at 37 °C. Fresh buffer with 0.5 mCi/ml ³²PO₄ was then added to the basolateral surface, and the cells were incubated for 3 h at 37 °C. Cells were then washed and stimulated with agonist in phosphate-free buffer for various times. The reaction was stopped by aspirating the phosphate-free buffer and adding 1 ml of cold methanol to the basolateral and apical surfaces of the cells. The cells were scraped from the filters and transferred to polypropylene tubes using borosilicate Pasteur pipettes. Inserts were washed twice with 1 ml of 2.4 N HCl, and these washes were pooled with the cell lysate in methanol. Phospholipids were extracted using chloroform, dried down under nitrogen, and resuspended in 90 µl of chloroform. Samples were then spotted onto TLC plates and developed in the mobile phase as described in the measurement of PI 3-kinase activity (above). Autoradiography was used to detect PIP, PIP₂, and PIP₃ in comparison with known standards. These lipids were then scraped from the TLC plate, deacylated using 2 ml of methylamine reagent (40% methylamine in water/methanol/butanol, 5:4:1, v/v), and sonicated. The glycero-inositol phosphates thus generated were separated via high performance liquid chromatography using a Whatman Partisil 5 SAX high pressure anion exchange column. Elution was performed using a three-step gradient of 0.01 M NH₄H₂PO₄, pH 3.8, for 15 min, 0.01-0.25 M for 60 min, and 0.25-1 M for 45 min. ³²P-Labeled glycerophospholipids were identified on the basis of their retention time in comparison to known [³H]inositol phosphate standards prepared from [³H]PIP₂ and [³H]PIP (Du Pont NEN) and quantitated by automated integration (Beckman System Gold software).

Analysis of variance (ANOVA) with Student-Newman-Keul's post-hoc test was used to compare mean values. p values <0.05 were considered significant. All data are expressed as the mean for a series of (n) experiments ± S.E.

RESULTS

EGF inhibits calcium-dependent chloride secretion by ``uncoupling'' the rise in intracellular calcium from the downstream response of chloride secretion (5). We aimed to determine the signal transduction pathway involved in this inhibitory effect. Since phosphatidylinositol 3-kinase has been implicated in the control of ion transport in other cell types (16, 17), this enzyme was the focus of the current studies. Fig. 1, A and B, shows the effect of a PI 3-kinase inhibitor, wortmannin (50 nM), on the inhibitory effects of either EGF or carbachol on thapsigargin-induced chloride secretion by T₈₄ cells. By itself, wortmannin had a small effect on basal chloride secretion within 10 min (peak I_sc, 0.57 ± 0.1 µA/cm², n = 15, p < 0.01). Wortmannin also exerted differential effects on secretory responses to calcium agonists. It significantly potentiated the response to thapsigargin (Fig. 1, A and B) while having no effect on the secretory response to carbachol (Fig. 1B). As reported previously, both EGF (Fig. 1A) and carbachol (Fig. 1B) significantly reduced the secretory response to subsequently added thapsigargin (5). Fig. 1A also shows the effect of wortmannin on the ability of EGF to inhibit thapsigargin-induced chloride secretion. Pretreatment with wortmannin partially reversed EGF's inhibition of the thapsigargin effect at early time points and completely reversed the inhibitory response at later time points, when compared with the wortmannin plus thapsigargin control (compare the open symbols in Fig. 1A). Similar results were obtained using the calcium ionophore, A23817, as a stimulus of chloride secretion (not shown). These data suggest that EGF may exert at least part of its inhibitory effect on chloride secretion through the activation of PI 3-kinase. In contrast, wortmannin failed to reverse carbachol-induced inhibition of thapsigargin-induced chloride secretion (Fig. 1B). Carbachol alone inhibited thapsigargin-induced chloride secretion by 62%. In the presence of wortmannin, carbachol still inhibited thapsigargin to a similar degree (63%) (compare the open symbols in Fig. 1B). This suggests that the ability of carbachol to inhibit chloride secretion likely does not involve PI 3-kinase activity. These latter results are also consistent with our hypothesis that the inhibitory effects of carbachol on calcium-dependent chloride secretion are mediated by the second messenger, Ins(3,4,5,6)P₄ (3).

The results described above, where thapsigargin was used as the stimulus of chloride secretion, were not unambiguous, however, since wortmannin alone altered the response to this secretagogue. In contrast, the secretory response to carbachol was not potentiated by wortmannin (Fig. 1B) and therefore provided a better paradigm for testing effects of wortmannin on EGF-induced inhibition. Fig. 2, A and B, shows the effects of wortmannin on the ability of EGF to inhibit either carbachol or histamine-induced chloride secretion. Unlike its effects on thapsigargin-induced secretion, wortmannin alone did not alter secretory responses to carbachol or histamine. As expected, EGF significantly inhibited secretion induced by carbachol (Fig. 2A) or histamine (Fig. 2B) (5). Furthermore, wortmannin completely reversed the inhibitory effects exerted by EGF on responses to carbachol or histamine. These data suggest that EGF activates PI 3-kinase and that this enzyme is involved in mediating the inhibitory effect of EGF on calcium-dependent chloride secretion.

We next determined if EGF does indeed activate PI 3-kinase in T₈₄ cells. In addition, the fact that wortmannin increased the level of basal chloride secretion and potentiated responses to thapsigargin (although not to carbachol or histamine) forced us to consider whether an activated PI 3-kinase was present in T₈₄ cells under basal conditions and/or if other agonists are capable of activating this enzyme. As tyrosine kinase-dependent PI 3-kinase is activated by recruitment of its 85-kDa regulatory subunit to tyrosine-phosphorylated proteins (6), we tested whether various agonists increased the amount of p85 found in antiphosphotyrosine immunoprecipitates. Fig. 3A shows that all agonists increased the levels of p85 in such immunoprecipitates but with slightly different kinetics (Fig. 3B). EGF increased the levels of this protein most rapidly, with peak levels seen at 1 min (which was the earliest time point examined). In contrast, carbachol, histamine, and thapsigargin all caused maximal recruitment of p85 at around 5 min. The different kinetics of recruitment for EGF compared with the calcium-dependent agonists suggest that different tyrosine kinase-dependent pathways may be activated. Moreover, neither histamine nor thapsigargin has been shown to inhibit chloride secretion, in contrast to the effects of carbachol. On the surface, therefore, these data, as well as the failure of wortmannin to reverse the inhibitory effects of carbachol on secretion, argued against a role for PI 3-kinase in inhibiting chloride secretion. Thus, we concluded that more direct assays of enzyme activity were needed.

The tyrosine kinase-dependent PI 3-kinase is a heterodimeric complex consisting of the regulatory p85 and the catalytic p110 subunit. Both subunits must be recruited for enzyme activity (7). We therefore determined whether the panel of agonists studied above were capable of increasing the levels of p110 found in antiphosphotyrosine immunoprecipitates. Unlike the results in Fig. 3, Fig. 4 shows that, at times of peak p85 accumulation, only EGF significantly increased the levels of p110. In some experiments, we observed a small increase in p110 in response to carbachol or histamine (but not thapsigargin). However, these responses were variable and were not found to be statistically significant. These results suggest that only EGF is able to significantly recruit the heterodimeric PI 3-kinase complex to a significant extent and thereby result in activation of the enzyme. These results also suggest that the calcium-mediated agonists recruit a pool of p85 that is independent of p110.

As various isoforms of p110 have been discovered (23, 24, 25), we performed further tests to verify that only EGF is able to increase PI 3-kinase activity. Fig. 5 shows the effects of the same panel of agonists on PI 3-kinase activity in vitro. In untreated cells, PI 3-kinase activity was detected, and this activity was sensitive to wortmannin (Fig. 5A). In keeping with the results in Fig. 4, only EGF increased PI 3-kinase activity above basal levels in antiphosphotyrosine immunoprecipitates (Fig. 5B). Thus, of the agonists tested, only EGF is able to increase PI 3-kinase activity through a tyrosine kinase-dependent pathway. Furthermore, pertussis toxin treatment had no effect on either thapsigargin-stimulated or carbachol-induced inhibition of calcium-dependent chloride secretion (not shown) suggesting that the pertussis toxin-sensitive, G-protein-mediated PI 3-kinase activity is also not involved in mediating inhibitory effects (26). These data suggest that, of the agonists tested, only EGF is able to activate PI 3-kinase in T₈₄ cells.

The products of PI 3-kinase activity include the 3-phosphorylated lipids PI(3)P, PI(3, 4)P₂, and PI(3,4,5)P₃. Thus, final studies examined the phospholipid profile in T₈₄ cells under various conditions, in order to correlate these findings with chloride secretory results. Fig. 6 shows the effects of EGF, carbachol, or wortmannin on levels of PI(3)P. A significant amount of PI(3)P was found in untreated cells, and levels were not significantly altered by either EGF or carbachol. However, wortmannin significantly reduced basal levels of PI(3)P. At the concentrations of wortmannin used (50 nM), no effects on other phospholipids were observed under unstimulated conditions (not shown). As a wortmannin-sensitive PI 3-kinase activity was present in antiphosphotyrosine immunoprecipitates of control cells (Fig. 5), and wortmannin alone had a small stimulatory effect on chloride secretion, we hypothesize that a PtdIns 3-kinase activity produces PI(3)P in unstimulated cells and that this lipid might regulate basal chloride secretion.

Fig. 7, A and B, shows the effects of EGF, carbachol, and wortmannin on the levels of PI(3,4)P₂ and PI(3,4,5)P₃. Unlike PI(3)P, these phospholipids were barely detectable in unstimulated cells. Treatment of T₈₄ cells with EGF (Fig. 7A) led to a large, rapid, and sustained elevation in the levels of both phospholipids. Moreover, the kinetics of this response correspond to the time course for the inhibitory effect of EGF on calcium-dependent chloride secretion (5). In contrast, carbachol had no effect on the levels of PI(3,4)P₂ and PI(3,4,5)P₃ (Fig. 7A). Pretreatment with wortmannin completely blocked the increase in PI(3,4)P₂ and PI(3,4,5)P₃ attributable to EGF stimulation (Fig. 7B). These data are in agreement with those presented above and again imply that only EGF, and not carbachol, activates PI 3-kinase. These data are also in agreement with the thesis that the rapid appearance of PI(3,4)P₂ and PI(3,4,5)P₃ in response to EGF mediates a wortmannin-sensitive inhibitory effect on calcium-dependent chloride secretion.

DISCUSSION

PI 3-kinase is a lipid kinase responsible for the production of 3-phosphorylated lipids and has been implicated in cell proliferation (11), cell movement (12), and glucose transport (14). The activation of PI 3-kinase has also been shown to be involved in regulating the Na⁺/H⁺ exchanger (16) and perhaps in the activation of NaCl absorption (17). It is often the case that pathways involved in the activation of intestinal secretion simultaneously decrease absorption. By analogy, we hypothesized that PI 3-kinase might be involved in mediating the inhibitory effect of EGF on chloride secretion. In fact, the ability of wortmannin, a PI 3-kinase inhibitor, to reverse inhibitory actions of EGF indeed suggests involvement of PI 3-kinase in this process. Thus, wortmannin partially reversed EGF-induced inhibition of thapsigargin-stimulated chloride secretion and completely reversed EGF's inhibition of carbachol- and histamine-induced chloride secretion.

The fact that wortmannin only partially reversed the inhibitory effect of EGF on thapsigargin-induced secretion merits some comment. On the surface, these results are difficult to reconcile with the fact that carbachol, histamine, and thapsigargin all activate calcium-dependent chloride secretory responses. However, the secretory response to thapsigargin was potentiated by wortmannin alone while those to carbachol and histamine were not. Chloride secretion evoked by thapsigargin appears to involve only the elevation of intracellular calcium (27), whereas G-protein-mediated agonists such as carbachol and histamine likely activate additional signaling pathways (28). Both carbachol and histamine evoke responses that are more transient than that induced by thapsigargin. Thus, we hypothesize that they may be activating other inhibitory pathways that may override any potentiative effects wortmannin could exert.

EGF induced a rapid and sustained elevation of PI(3,4)P₂ and PI(3,4,5)P₃ in T₈₄ cells. This effect correlates well with the ability of EGF to inhibit calcium-activated chloride secretion. EGF inhibits carbachol-induced chloride secretory responses within 1 min, and the inhibition is then maintained for at least 60 min (5). In addition, the increase in these lipids induced by EGF was completely inhibited by wortmannin, further suggesting a role for these lipids in EGF inhibition of secretion. However, the mechanism whereby these lipids might inhibit secretion is currently not known. Preliminary data suggest that EGF exerts a wortmannin-sensitive inhibitory effect upon a basolateral potassium channel (29). Furthermore, previous studies have shown that the basolateral potassium channel involved in calcium-activated chloride secretion is inhibited by protein kinase C (PKC) (30). 3-Phosphorylated phospholipids have recently been demonstrated to be capable of activating both novel and atypical PKCs in both in vitro and in vivo systems (31, 32). Preliminary data from our lab also suggest that these novel and atypical isoforms of PKC are present in T₈₄ cells.³ Thus, 3-phosphorylated phospholipids could activate these PKCs in the process of inhibiting chloride secretion. As these isoforms of PKC are quite insensitive to available inhibitors (33), further experiments, utilizing more sophisticated molecular approaches, will be required to test this hypothesis.

We also observed a small elevation in basal chloride secretion in response to wortmannin, which correlated with a reduction in the basal levels of PI(3)P. Recently, a mammalian PtdIns 3-kinase activity was found that is highly homologous to yeast Vps34p (19). The existence of this isoform of PI 3-kinase, and the presence of PI 3-kinase activity in antiphosphotyrosine immunoprecipitates from unstimulated cells, may explain the relatively high levels of PI(3)P in resting T₈₄ cells. In other systems, wortmannin has been shown to affect basal endocytotic processes and vesicular trafficking, with these effects attributed to actions on PI 3-kinase activity and 3-phosphorylated lipids (13). We can therefore speculate that wortmannin, by reducing basal levels of PI(3)P, might also regulate the trafficking of various membrane proteins involved in chloride secretion. Trafficking of membrane proteins clearly plays a role in regulating secretion, since cAMP-mediated secretory responses, for example, require the insertion of Na⁺/K⁺/2Cl cotransporters into the basolateral membrane for maximal activity (34). Thus, altering the levels of certain membrane proteins could also conceivably result in negative regulation of secretion.

Some studies suggest that wortmannin may not be wholly specific for the inhibition of PI 3-kinase. Inhibition of myosin light chain kinase (35), PI 4-kinase (36), phospholipase A₂ (37), and phospholipase D (38) have also been reported. However, at the concentrations of wortmannin used in our study (50 nM), inhibition of myosin light chain kinase is not likely, and no reduction in the level of PI(4)P was observed. It is also unlikely that wortmannin is inhibiting phospholipase A₂ or phospholipase D in our system since their products, arachidonic acid and phosphatidic acid, respectively, have pro-secretory effects (39, 40). Inhibition of these enzymes would therefore be expected to reverse a stimulatory rather than an inhibitory effect on secretion. We therefore believe that the actions of wortmannin on chloride secretion are likely reflective of its specific inhibition of PI 3-kinase.

The significance of the ability of carbachol, histamine, and thapsigargin to recruit p85, independent of p110, is currently unknown. p85 contains an SH3 domain, a proline-rich sequence, and a region homologous to the bcr gene product, in addition to its 2 SH2 domains. It is therefore likely that p85 can couple proteins to tyrosine kinase-dependent signaling pathways, in a manner similar to the adaptor proteins, Grb2 and Shc (6). Studies have in fact shown that p85 can be recruited to tyrosine-phosphorylated proteins independent of p110 (41). Other studies have demonstrated that p85 is able to couple to p120 Gap and Grb2 (42, 43), proteins involved in the regulation of Ras activation. As carbachol has been shown to activate Ras in other cell systems (44), it is possible that calcium agonists activate the Ras pathway via coupling to p85. However, a definitive answer to this question, as well as an understanding of its significance for cell function, will require further study.

In summary, we have shown that EGF can activate PI 3-kinase and that the lipid products of this enzyme may regulate calcium-activated chloride secretion. Furthermore, our data suggest that a PtdIns 3-kinase activity is present in untreated T₈₄ cells and that this activity, through the production of PI(3)P, may be involved in the regulation of basal chloride secretory tone. We speculate that, in addition to its mitogenic effects, the activation of PI 3-kinase may be involved in growth factor-induced effects on the epithelium that could reflect adaptive responses to epithelial damage and inflammation. By limiting active secretion, cellular energy resources could then be diverted to cell proliferation and epithelial restitution, while limiting inappropriate fluid and electrolyte loss. Finally, we have shown that the inhibitory effects of carbachol on chloride secretion likely do not involve PI 3-kinase activity. However, the ability of carbachol and histamine to recruit either p85 alone, or small amounts of the p85/p110 complex, without activating PI 3-kinase, may have alternate signaling consequences that have yet to be elucidated.