Relationship between Phosphatidic Acid Level and Regulation of Protein Transit in Colonic Epithelial Cell Line HT29-cl19A*

Colonic epithelial HT29-cl19A cells are polarized and secrete proteins among which α1-antitrypsin represents about 95%. Secretion occurs via a constitutive pathway, so that the rates of secretion directly reflect the rates of protein transit. In this paper we have demonstrated that: 1) in resting cells phospholipase D (PLD) is implicated in the control of apical protein transit; 2) phorbol esters stimulate apical protein transit (stimulation factor 2.2), which is correlated with a PLD-catalyzed production of phosphatidic acid (PA) (2.45-fold increase); 3) the stimulation of cholinergic receptors by carbachol results in an increase (stimulation factor 1.45) of apical protein transit which is independent of protein kinase C and PLD activities, but related to PA formation (1.7-fold increase) via phospholipase(s) C and diacylglycerol kinase activation; 4) an elevation of the cAMP level enhances apical protein transit by a PA-independent mechanism; 5) a trans-Golgi network or post-trans-Golgi network step of the transit is the target for the regulatory events. In conclusion, we have shown that PA can be produced by two independent signaling pathways; whatever the pathway followed, a close relationship between the amount of PA and the level of secretion was observed.

The protein machinery underlying protein transport from the endoplasmic reticulum to the cell surface has been extensively studied (for a review, see Ref. 1). However, interest in the regulatory mechanisms that control the different steps of the intracellular traffic and more particularly the role of the lipids in this processes is rapidly growing at the moment.
The possible involvement of different signal transduction pathways in the control of intracellular protein transit has recently been investigated; the first evidence indicating that anterograde transport along the secretory pathway is regulated by protein phosphorylation was provided by the use of the phosphatase inhibitor okadaic acid (2)(3)(4)(5).
The involvement of protein kinase A (PKA) 1 in regulating membrane traffic was suggested when it had been found that the regulatory RII subunit of PKA was partially associated with Golgi membranes (6 -8). Pimplikar and Simons (9) and Hansen and Casanova (10), respectively, showed that PKA activators increase the apical transport of influenza hemagglutinin and the secretion of the endogenous glycoprotein gp80 in Madin-Darby canine kidney cells. More recently, Jilling and Kirk (11) showed that, in cultured colonic epithelial cells, cAMP stimulated constitutive membrane traffic from the TGN to the apical cell surface. Muniz et al. (12) demonstrated that, in NRK cells, the inhibition of PKA by the PKA inhibitor H89 decreased protein transit from the ER to the Golgi and blocked the exit of the vesicular stomatitis virus G glycoprotein from the Golgi. In lacrimal glands Robin et al. (13) showed that the PKA inhibitor H89 affected the intracellular transit of regulated secretory proteins. As far as a regulatory effect of protein kinase C (PKC) on intracellular traffic is concerned, few data are available. Membrane trafficking between the ER and the Golgi apparatus of NRK cells was found to be regulated by a diacylglycerol/phorbol ester-binding protein (14), whereas in Madin-Darby canine kidney cells Pimplikar and Simons (9) and Cardone et al. (15), respectively, described activation of apical transport of influenza hemagglutinin and stimulation of transcytosis of polymeric IgA receptor, by phorbol esters. In cell-free systems, PKC has also been shown to play a role in the formation of vesicles from the TGN (16 -21).
Finally, enzymes that modify membrane lipids may also regulate constitutive membrane traffic. Evidence for the participation of phosphoinositides and phosphatidylinositol kinases in the regulation of membrane traffic has been reported (for a review, see Ref. 22). Lipases are also implicated in these phenomena; as long ago as 1993, Stutchfield and Cockcroft (23) showed that phospholipase D (PLD) activity was correlated with secretion in HL60 cells. More recently, PLD and phosphatidic acid (PA) produced by the hydrolysis of phosphatidylcholine (PC) by PLD, were shown to be required for the formation of Golgi secretory vesicles (24 -26) and for the regulation of transport from the endoplasmic reticulum to the Golgi complex (27).
To study the effect of the signal transduction pathways on the intracellular transit of proteins more precisely, the human colonic cell line HT29-cl19A appeared to be a good model system. These epithelial cells, like the HT29-cl16E, emerged from the parental HT29 cells after the induction of differentiation by treatment with butyrate (28) and are stably differentiated. HT29-cl19A cells have mostly been used to study chloride secretion and cystic fibrosis transmembrane conductance regulator expression, whereas HT29-cl16E cells were selected to study mucin secretion (29 -31). HT29-cl19A cells are unable to secrete mucins but synthesize and secrete different proteins * 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.
The aim of this study was to investigate in intact cells, which normally secrete proteins, the different pathways controlling protein transit and to determine the nature of the second messenger(s) connecting the signaling events, evoked by the binding of receptor-directed agonists, to protein secretion.
In this work we quantified the amounts of ␣ 1 -antitrypsin, the major protein secreted. We showed that protein transit can be regulated at a TGN or post-TGN step of the pathway and confirmed that agents that increase the cAMP level enhanced apical protein transit. We demonstrated that other mechanisms also controlled secretion: (i) phorbol esters, which we showed to activate PLD, enhanced the PA amount and the apical transit; (ii) carbachol, which we showed to activate PLC and produce an elevation of the PA level (but without activating PLD), also increased apical transit, but to a lower extent.
We can conclude that in HT29-cl19A cells, an increase of PA level, either directly by PLD action or by a two-step mechanism involving the action of both PLC and DAG kinase is correlated to an enhancement in the rate of apical protein transit. Cell Culture-HT29-cl19A cells (kindly donated by Dr. C. L. Laboisse) were a cloned line derived from the parental HT29 cells after treatment by butyrate (28). The cells were grown in Falcon culture flasks (25 cm 2 ) in a humidified atmosphere of 95% air, 5% CO 2 , at 37°C. The cells were fed every day with Dulbecco's modified Eagle's medium (DMEM; 4.5g/liter glucose) supplemented with 10% heat-inactivated fetal bovine serum. Cells were subcultured after trypsin treatment every 7 days when they had reached about 90% confluence. The passage number of the cells used in this study varied between 16 and 30. For all experiments, cells were seeded (about 0.25 ϫ 10 6 cells/filter) onto Falcon cell culture inserts (10.5-mm membrane diameter, 0.4-m pore size, 1.0 ϫ 10 8 pores/cm 2 ). Cells were used between 14 and 20 days following seeding, i.e. at least 1 week after they had become confluent.

Materials-Tissue
Metabolic Labeling and Measurement of Protein Secretion-For metabolic labeling, cells were incubated for 30 min in leucine-free medium supplemented with 1% BSA. They were then pulsed for 10 min with [ 3 H]leucine (38.5 Ci/mmol) in 1% BSA leucine-free medium; [ 3 H]leucine was added to the basolateral side (65 Ci/380 l volume). Cells were washed three times with 1 mM leucine in PBS to stop the labeling, and 1 ml of 1% BSA in DMEM was placed in both the apical and the basolateral compartments. Secretagogues, when used, were added to the basolateral medium and the cells were incubated for 3-6 h at 37°C. Medium samples (duplicate aliquots of 80 l) were collected at various time periods, centrifuged at 2,000 ϫ g to remove any detached cells, and kept at 4°C. Cells were lysed for 15 min in 1 ml of 0.5 N NaOH. The media samples and duplicate aliquots (80 l) of cell lysates were subjected to two cycles of 20% trichloroacetic acid -0.1% phosphotungstic acid precipitation, and the insoluble radioactivity was counted in a 1212 Rackbeta liquid scintillation counter (Amersham Pharmacia Biotech, Saint-Quentin en Yvelines, France).
When secretagogues were solubilized in Me 2 SO, the final concentra-tion of this solvent in the incubation medium was below 0.2%. Protein secretion was expressed as the percentage of 3 H-labeled proteins released into either the apical or basolateral medium (i.e. 100 ϫ total amount of 3 H-labeled proteins released in one medium/total amount of 3 H-labeled proteins in the two media and in the cells). The secretion stimulation factor represents the ratio of protein secretion in the presence of a secretagogue to control protein secretion.
SDS-PAGE Analysis and Fluorography of Secreted Proteins-For these experiments, proteins were radiolabeled for 30 min with [ 3 H]leucine (47.5 Ci/mmol) and then incubated with 300 l of DMEM in both compartments for 6 h. BSA was absent from all the culture media. The radiolabeled proteins secreted in the incubation media were denaturated by adding 5ϫ denaturating sample buffer to each sample and heating to 100°C. Samples were separated by SDS-PAGE on 10% acrylamide gels. After fixation and staining with Coomassie Blue, the gels were soaked in Amplify (Amersham Pharmacia Biotech), dried, and exposed to Amersham Hyperfilms-MP for 1 to 5 days. Films were then quantified by laser scanning densitometry with a Personal Densitometer SI and ImageQuaNT software (Molecular Dynamics, Evry, France).
Immunoprecipitation-Secretion media were obtained as described for fluorography experiments. Cells were lysed in 300 l of 0.1% triton X-100 in DMEM in the presence of protease inhibitors (antipain, chymostatin, leupeptin, pepstatin, phenylmethylsulfonyl fluoride, benzamidine). Aliquots of identical volumes of media and lysates were then incubated at 4°C with polyclonal antibody against ␣ 1 -antitrypsin (Sigma; dilution 1:250) for 2.5 h followed by incubation with immobilized protein A (Amersham Pharmacia Biotech) for 30 min. Immunoprecipitates were washed twice with DMEM and solubilized in sample buffer for SDS-PAGE. In some experiments, immunoprecipitations were performed in the absence of antibody in order to determine the aspecific binding of radiolabeled proteins on protein A-Sepharose.
ELISA-Secretion media were obtained as described for fluorography experiments, and cells were lysed as described for immunoprecipitation experiments. The quantification of ␣ 1 -antitrypsin in the media was performed using a competition ELISA. Human ␣ 1 -antitrypsin (ICN) was used to standardize this assay, in the concentration range 3 ϫ 10 Ϫ10 to 10 Ϫ7 M. The samples containing the antigen were incubated, first for 2 h with the anti-␣ 1 -antitrypsin peroxidase-labeled polyclonal antibody (ICN; dilution 1:10000) and then for 2 h on a titration plate coated with ␣ 1 -antitrypsin. The peroxidase activity of the antigenantibody complex bound to the plate was determined using 3,3Ј,5,5Јtetramethyl benzidine as a substrate.
Lipid Analysis-For the measurement of PA and phosphatidylethanol (PEt) production, HT29-cl19A cells grown for 14 -20 days on cell culture inserts were washed in serum-free DMEM and then labeled for 2 days with [ 3 H]myristic acid in DMEM containing 1 mg/ml lipid-free BSA. [ 3 H]Myristic acid (49 Ci/mmol) was added to the basolateral side (2 Ci/1 ml of medium). Before beginning the experiments, the cells were washed with PBS and incubated in 1 ml of DMEM in the presence of the drugs to be tested. At the end of the experiment, the cells were collected by cutting the filters from the filter cups and the cellular lipids were extracted by vortex-mixing in 1.8 ml of methanol/chloroform/HCl (100:50:1) at 4°C. 30 min later 0.5 ml of water, 0.6 ml of chloroform, and 0.6 ml of 2 M KCl were successively added to produce an aqueous and an organic phase. The phases were separated by centrifugation at 1000 ϫ g for 10 min; the aqueous phase was discarded, and the organic phase containing the lipids was evaporated. The dried samples were dissolved in chloroform/methanol (19:1), loaded on TLC silica plates (Whatman LK6D), and resolved by developing with the organic phase of the solvent system prepared by mixing isooctane/acetic acid/ethyl acetate/water (2:3:13:10). The radiolabeled lipids were detected and quantified with a Tracemaster 20 automatic TLC-linear analyzer (Berthold, EGG, Evry, France).
When PLD activity was determined through the transphosphatidylation reaction, experiments were performed as described above except that the cells were preincubated 15 min in the presence of 3% ethanol prior to the addition of the drugs to be tested.
The identity of PA, PEt, and PC were confirmed by comparison of their R F values with PA and PEt standards revealed after iodine sublimation, and [ 3 H]choline-labeled PC radiodetected.
The amounts of PA and PEt formed were expressed as percentages of the amount of PC detected.
Measurement of Inositol Phosphates Production-For the measurement of inositol phosphates (IPs) production, HT29-cl19A cells grown for 14 -20 days on cell culture inserts were washed in serum-free DMEM and then labeled for 2 days with [ 3 H]inositol in DMEM containing 1 mg/ml lipid-free BSA. [ 3 H]Inositol (17.6 Ci/mmol) was added to the basolateral side (5 Ci/1 ml of medium). Before beginning the experiments, the cells were washed with PBS and incubated in 1 ml of DMEM, supplemented with 10 mM LiCl, in the presence of the drugs to be tested. The reaction was stopped by addition of 1 ml of cold 7% trichloroacetic acid and the cells were scrapped from the inserts. Cell homogenates were centrifuged for 10 min at 15,000 ϫ g and 4°C. The IPs-containing supernatants were treated (4 ϫ 6 ml) with diethylether to remove trichloroacetic acid, neutralized with Tris base, and then applied to anion-exchange columns containing about 1 g of Dowex 1 (AG 1-X8, 200 -400-mesh, formate form, Bio-Rad). Columns were washed with 10 ml of water to remove the excess of myo-inositol; glycerophosphoinositols were eluted with 10 ml of 60 mM ammonium formate, 5 mM disodic tetraborate. Total IPs were eluted with 12 ml of 1 M ammonium formate, 0.1 M formic acid. The radioactivity of aliquots of all the eluates was determined.

Characterization of Protein Secretion in HT29-cl19A
Cells-In order to understand the regulation of protein secretion in these filter-grown cells, we first analyzed the unstimulated secretion of metabolically pulse-labeled proteins by different methods.
When we followed kinetics of secretion of leucine-labeled, trichloroacetic acid-precipitated proteins in apical medium, we observed a lag time of about 1 h before the beginning of protein release; this may correspond to the minimal duration of intracellular transit. The rate of protein release then increased linearly for the next 3 h (data not shown). The amount of radiolabeled proteins released in the apical medium after 3 and 6 h of incubation is shown in Table I. In the basolateral medium, the secretion was lower ( Table I).
The [ 3 H]leucine-labeled proteins released by HT29-cl19A cells were analyzed by SDS-PAGE and fluorography. Fig. 1a shows the fluorographic patterns of the proteins secreted in the two media. In the apical medium from unstimulated cells (lane 4), a 55-kDa polypeptide was observed. When the film was overexposed, minor polypeptides were detected (lane 4Ј); the 55-kDa polypeptide was found to represent about 95% of the secreted radiolabeled proteins. In the basolateral medium from unstimulated cells (lane 5), the 55-kDa polypeptide was also present, but its amount was lower than in the apical medium. Minor polypeptides with molecular weights different of those of the polypeptides found in the apical medium were seen when the film was overexposed (lane 5Ј). The nature of the proteins secreted and the relative proportion of the 55-kDa band are not identical in the two media, indicating a polarized secretion.
Jilling and Kirk (11) identified the 55-kDa polypeptide as ␣ 1 -antitrypsin. The immunoblots presented in Fig. 1b confirm the presence of this protein in the two media and the cell lysate.
We thus analyzed the secretion of ␣ 1 -antitrypsin by immunoprecipitation. Fig. 1c (lanes 1-5) shows the result of the fluorography of the immunoprecipitates obtained from the medium and from the cell lysate. The intensity of the band of radiolabeled ␣ 1 -antitrypsin from the apical medium was higher than that from the basolateral medium.
Finally, we quantified by ELISA the amount of ␣ 1 -antitryp- H]leucine and treated or not with 1 M PDBu immediately after. Radiolabeled proteins were quantified, using the trichloroacetic acid assay, after 3 or 6 h of incubation; ␣ 1 -antitrypsin was quantified by ELISA after 6 h of incubation. Values are means Ϯ S.E., n Ͼ 10 for trichloroacetic acid assay and n ϭ 4 for ELISA. The total ␣ 1 -antitrypsin (secreted ϩ non-secreted) was estimated as 250 Ϯ 50 pmol/insert (n ϭ 2).   (Table I). We also found a higher amount of ␣ 1 -antitrypsin in the apical medium than in the basolateral one.
Effect of Phorbol Esters on Protein Secretion-PKA activators have been described to selectively stimulate apical protein secretion in HT29-cl19A cells (11), but no data were available concerning the possible involvement of PKC in the regulation of protein transit in these polarized cells. We thus tested the hypothesis that phorbol esters, potent PKC activators, might also stimulate apical protein secretion.
We investigated the dose-dependent effect (3 nM to 3 M) of PDBu, one phorbol ester, on radiolabeled protein secretion and showed that it was able to stimulate the protein release in both compartments but with different efficiencies. In the apical medium, a maximal effect on protein release was observed for PDBu concentrations Ն1 M; the 1 M concentration was used for further experiments.
The time course of apical radiolabeled protein secretion stimulated by PDBu was followed; the lag time was shortened to about 50 min, and the rate of protein release was increased compared with unstimulated cells (data not shown). The effect of PDBu on radiolabeled protein secretion into both media was quantified, and the results are given in Table I. Under the same experimental conditions as those used for PDBu, PMA, another phorbol ester, used at a concentration of 1 M, yielded identical results (data not shown).
The fluorographic pattern of the proteins secreted into the apical medium after PDBu stimulation (Fig. 1a, lane 1 compared with lane 4) showed that the intensity of the 55-kDa band was increased. When the film was overexposed, the minor polypeptides were always detected (data not shown); their relative amounts toward the 55-kDa band were found to be identical to that of the control.
␣ 1 -Antitrypsin was quantified by ELISA after PDBu stimulation. The results, given in Table I, indicate that about 80% of the ␣ 1 -antitrypsin is secreted from the cells after 6 h of stimulation with PDBu.
The fluorographic patterns of the immunoprecipitates (Fig.  1c) confirmed these results. After PDBu stimulation, the intensity of the 55-kDa band was increased and the amount of ␣ 1 -antitrypsin in the cell lysate greatly reduced.
The results obtained by the two kinds of quantification (tri-chloroacetic acid precipitation assay and ELISA) are coherent. After 6 h of incubation, phorbol esters enhanced the apical secretion about 2-fold and the basolateral one only 1.3-fold (Table I). Consequently, for subsequent experiments, the trichloroacetic acid precipitation assay was mainly used because it was easier than ELISA to perform routinely. We verified that the stimulation of secretion by PDBu was neither a consequence of lysis of the cells nor the result of a PDBu-induced elevation of the rate of protein synthesis (data not shown).
To evaluate the requirement of a phosphorylating activity of PKC in the pathway enhanced by phorbol esters, we tested the effect of a competitive inhibitor for the ATP-binding site of PKC, BIM (33). The dose-dependent effect of BIM (0.15-10 M) on the apical secretion of radiolabeled proteins enhanced by PDBu was investigated. Under our experimental conditions, the maximal effect of the drug was obtained for concentrations Ͼ 2.5 M and a plateau was reached; for these concentrations BIM was found to decrease the PDBu stimulation factor of about 60%. These results indicate that a part of the transit regulated by phorbol esters is dependent on the phosphorylating activity of PKC.
Implication of PLD and PA in Protein Secretion-Recent data obtained from experiments performed with cell-free systems have pinpointed the role of PLD in vesiculation (24,25). Since this enzyme can be activated by PKC (34,35), we wondered whether, in HT29-cl19A cells, the PDBu-activated PKC could regulate the PLD activity and consequently the secretory process.
In the presence of primary alcohols, PLD catalyzes a transphosphatidylation reaction; the production of PA is inhibited, and a phosphatidyl alcohol product is formed instead. We tested the effect of different alcohols on protein secretion. The unstimulated secretion in the apical medium was reduced by ethanol in a dose-dependent way (Fig. 2a). After 3 h of incubation in the presence of 3% ethanol, the secretion rate decreased by 36%. The effect of 0.8% butanol-1 (another primary alcohol) was close to that of 3% ethanol, whereas 0.8% butanol-3 (a tertiary alcohol, which is not a substrate for the transphosphatidylation reaction) was unable to reduce secretion (data not shown). It is also noteworthy that ethanol had no effect on basolateral protein secretion (data not shown). To investigate the relationship between secretion and PLD activity, we measured the activity of this enzyme under different experimental conditions (Fig. 2, b and c). In control cells, a small amount of PEt was detected, indicating a basal activity of PLD. PDBu enhanced the amount of PEt formed by about 9-fold, and BIM reduced the effect of PDBu by about 65%. This effect of BIM is in good correlation with its 60% inhibitory effect on protein secretion enhanced by PDBu.
Since very recent data have suggested a role for PA, the physiological product of PC hydrolysis by PLD, in vesiculation (26,27), we wondered whether this phospholipid would be able to regulate protein secretion. We thus determined the amounts of PA formed under different experimental conditions and showed that they were accurately correlated with the percentages of secretion (Fig. 3); in the presence of 3% ethanol the amount of PA decreased by 35% and the secretion rate by 36%, whereas in the presence of PDBu the amount of PA increased by 2.45-fold and the secretion rate by 2.2-fold.
Regulation of Protein Secretion by Carbachol-Since phorbol esters stimulate protein transit partly via PKC activation, we wondered whether receptor-directed agonists, known to lead to PKC activation, would also be able to regulate this transit. HT29-cl19A cells possess muscarinic M 3 receptors (36); therefore, we tested the effect of carbachol, an agonist of this receptor type, on protein secretion. The dose-dependent effect of this agonist was investigated; carbachol (concentrations in the range of 30 -100 M) induced a maximal increase of protein secretion in the apical medium but was without effect on the basolateral secretion. The stimulation factors obtained for a carbachol concentration of 100 M were 1.45 Ϯ 0.04 and 1.32 Ϯ 0.03 after 3 and 6 h of incubation, respectively (n Ͼ 10). This carbachol concentration was used for further experiments.
The fluorographic pattern of the proteins secreted in the apical medium shows a small increase in the intensity of the 55-kDa band (Fig. 1a, lane 2 compared with lane 4).
The activation of muscarinic receptors can be coupled, via a G protein, to the hydrolysis of phosphatidylinositol bisphosphate (PIP 2 ) by phosphatidylinositol-phospholipase C (PI-PLC), which produces DAG and inositol trisphosphate (IP 3 ). IP 3 production then leads to Ca 2ϩ mobilization, and DAG activates PKC. In order to examine whether this Ca 2ϩ pathway is implicated in the secretory response to carbachol, we first quantified the amount of total IPs. We found that 10 min, 30 min, and 2 h after the stimulation by carbachol the amount of IPs was increased by about 3-, 7-, and 16-fold respectively. This result is in agreement with previous data indicating that, in HT29-cl19A cells, carbachol stimulates increases in cellular IP 3 and cytosolic Ca 2ϩ (37). Nevertheless, the Ca 2ϩ pathway may not be implicated in the regulation of protein secretion, since 1 M ionomycin (11) and 10 M of the ionophore A23187 (data not shown) were unable to enhance protein release.
Since DAG is the physiological activator of most PKCs, we tested the effect of the PKC inhibitor, BIM, on the secretory response triggered by carbachol. 5 M BIM was unable to inhibit the enhancement of protein release induced by carbachol (data not shown). These results suggest that PKC is not implicated in the regulation of protein transit when the cells are stimulated by carbachol.
Finally, as about 40% of the PLD activation triggered by PDBu was independent of the PKC activity, we wondered whether carbachol could activate PLD by a PKC-independent mechanism. Fig. 2c shows that carbachol was unable to trigger the activation of PLD since no PEt was formed in these conditions.
These results are not in agreement with the commonly admitted statement that PLD activation involves the phosphoinositide cycle and PKC (35).
Implication of PA in Protein Secretion Enhanced by Carbachol-Since Ca 2ϩ release, PKC, and PLD activations could not explain protein release stimulated by carbachol, we investigated the possible implication of another signaling pathway. DAG produced by PIP 2 hydrolysis can be converted into PA by means of a DAG kinase. To enhance the amount of PA, we used propranolol, an inhibitor of the phosphatidic acid phosphohydrolase which hydrolyzes PA into DAG (38), and tested the effect of this drug on protein secretion. 300 M propranolol increased the secretion obtained with carbachol; the stimulation factor obtained after 6 h of incubation was 1.55 Ϯ 0.06 (n ϭ 3).
In order to check whether the amount of PA was increased by carbachol treatment, we quantified the amounts of PA in different experimental conditions. The chromatograms presented in Fig. 4a show that the amount of PA formed after stimulation with carbachol was increased. Carbachol increased the amount of PA by about 1.7-and 1.4-fold in control cells or in cells pretreated with 300 M propranolol, respectively (Fig. 4, b and  c). It is noteworthy that this increase in the amount of PA is less important than after a PLD stimulation triggered by PDBu (2.45-fold increase) but is well correlated with a lower effect of carbachol on protein secretion (stimulation factor 1.45 with carbachol and 2.2 with PDBu). When cells were pretreated with 10 M DAG kinase inhibitor R59949 (39) prior to a stimulation by carbachol, the increase of PA was reduced by about 70% (Fig. 4b). These results confirmed the hypothesis of a production of PA, via the activation of both PLC and DAG kinase, which can explain the regulatory effect of carbachol on protein transit.
Nevertheless, DAG, and consequently PA, can also be produced by the carbachol-triggered hydrolysis of PC by a phosphatidylcholine phospholipase C (PC-PLC) (40,41). Since PC is the main phospholipid labeled with [ 3 H]myristic acid (data not shown), we wondered whether labeled PA might have also been produced by PC hydrolysis. To investigate the possible implication of the PC-PLC in the formation of PA, we tested the effect of a preincubation of the cells with 200 M D609, a competitive inhibitor of this enzyme (42). Fig. 4c shows that this inhibitor reduced by about 70% the amount of PA formed after incubation with carbachol in the presence of propranolol. This effect was not due to an inhibition of PI-PLC, since we verified that D609 did not reduce the production of IPs stimu- lated by carbachol (data not shown).
Independence of the Pathways Regulating Protein Secretion-Since our results showed that the secretory responses to carbachol and PDBu occurs via different pathways, we wondered whether they were independent and additive. Fig. 5a shows that the effects of 100 M carbachol and 0.1 or 1 M PDBu were additive (the PDBu concentration of 0.1 M was tested, since 80% of the total ␣ 1 -antitrypsin being released after 6 h of incubation with 1 M PDBu, an additional increase in secretion was difficult to detect). PKA activators had also been shown to stimulate apical transport in HT29-cl19A cells (11). We thus asked the question of whether the secretory responses to secretagogues that increase the intracellular cAMP level are independent of, and consequently additive with, the secretory responses studied in this work. We first compared, under our experimental conditions, the effect of vasoactive intestinal peptide, a physiological stimulator of the HT29-cl19A cells, dibutyryl cAMP, a permeant cAMP analogue, as well as forskolin, a direct activator of adenylyl cyclase. The stimulation factors were close to 1.4 for the three secretagogues (data not shown) and in the fluorographic pattern of the proteins secreted in the apical medium after forskolin stimulation, the intensity of the 55-kDa band was slightly increased (Fig. 1a, lane 3 compared with lane 4). We also found that forskolin acted by a completely different mechanism from that of carbachol or phorbol esters, as it did not increase the production of PA (data not shown).
When 100 M carbachol and 10 M forskolin were added together, the secretion stimulation factor was increased and corresponded to a complete additivity of the effects of the two secretagogues (Fig. 5b). This result confirms that those two regulatory pathways are fully independent.

Regulation of Protein Transit at a TGN or Post-TGN
Step of the Secretory Pathway-We finally investigated whether the different modulators of protein secretion act at earlier or later steps of the transit. To answer this question, we used the "20°C block" (43). HT29-cl19A cells were pulse-labeled at 37°C and then incubated for 3 h at 20°C. At this temperature, the radiolabeled proteins should be entrapped within the TGN. The different drugs to be tested, PDBu, carbachol, and ethanol, were then added to the basolateral medium, and the incubation was continued at 37°C for 3 h. A typical experiment showing the kinetics of radiolabeled protein secretion in the apical me- dium is shown in Fig. 6. After a lag time of about 45 min, secretion started and exhibited linear kinetics. The stimulation factors calculated after 3 h were 1.88 Ϯ 0.20, 1.31 Ϯ 0.12, and 0.55 Ϯ 0.02 for PDBu, carbachol, and ethanol, respectively. They are in the same range as the stimulation factors calculated when the drugs were added just after the pulse.
Although these results do not allow us to exclude a role of the secretagogues at the ER-to-Golgi step, they clearly indicate that a TGN (or post-TGN) step in protein transit is regulated in HT29-cl19A cells.

DISCUSSION
In this work we present new data concerning the regulation of apical protein transit by several different signaling pathways. Colonic epithelial HT29-cl19A cells, 5-10 days after they have reached confluence, exhibit morphological cell polarity and functional differentiation (28). They secrete in a constitutive way ␣ 1 -antitrypsin, which represents about 95% of the apically secreted proteins.
We showed that phorbol esters reduce the duration of transit and enhance the rate of apical protein transit by 2.2-fold. This increase in protein transit is inhibited by BIM, a PKC inhibitor. Since BIM interacts specifically with the catalytic domain of PKC, we conclude that PKC activity is needed for the regulation of protein transit in HT29-cl19A cells. The requirement of the phosphorylating activity of a Golgi-associated PKC in the vesicle formation and release was described by Westerman et al. (21) and Simon et al. (19); our experiments in intact cells are in good agreement with these findings from cell-free systems.
The hypothesis that PKC could play a role in vesicular transport through the activation of PLD, which has been found to be enriched in Golgi membranes (44,45) is attractive. The possible implication of PLD in the regulation of protein transit in HT29-cl19A was thus investigated. Ethanol was found to reduce unstimulated secretion in a dose-dependent way; moreover, in unstimulated cells incubated in the presence of ethanol, a small amount of PEt was detected, indicating a low basal PLD activity. The suggestion that a basal activity of this enzyme may mediate constitutive secretion was recently put forward by Bi et al. (27). To confirm the involvement of PLD and PA in protein transit, we determined the amounts of PA (or PEt) in different experimental conditions and observed a close correlation between these results and our data concerning se-cretion. (i) PDBu increased the amount of PA by 2.45-fold, whereas it stimulated secretion by 2.2-fold; (ii) ethanol decreased the amount of PA by 35%, whereas it lowered secretion by 36%; (iii) BIM inhibited the PDBu-stimulated PLD activity (65% inhibition) to the same degree as it reduced protein release triggered by PDBu (60% inhibition). Nevertheless, about 40% of both PLD activation and protein secretion triggered by phorbol esters bypasses the PKC step, and at present we do not have explanation for this pathway.
Taken together our results strongly suggest that, in intact cells, phorbol esters regulate protein transit via PLD-catalyzed PA production. They are in good agreement with the recent findings, obtained from experiments performed with a cell-free system, that PA increases vesicle budding from the TGN (26).
In this work, phorbol esters were shown to activate PKCs and PLD with a resulting increase in protein transit. However, although these drugs are useful pharmacological tools, they are not physiological modulators of the regulatory pathways. We took advantage of the fact that HT29-cl19A cells possess muscarinic M 3 receptors (36) to study the effect of carbachol, a cholinergic receptor-directed agonist, on protein transit. Activation of these receptors results in the triggering of a well established signal transduction pathway; the activation of a PI-PLC generates IP 3 , which controls cytosolic Ca 2ϩ levels, and DAG, which is an important regulator for PKC.
Carbachol was less efficient than phorbol esters in increasing protein secretion (stimulation factor 1.45 compared with 2.2 obtained with PDBu). Nevertheless, the results obtained with carbachol were rather unexpected. Although carbachol, via PI-PLC activation, was able to produce IP 3 (and consequently DAG), (i) Ca 2ϩ was not found to be involved in the control of protein transit; (ii) DAG was not found to activate PLD via a PKC activation as it was the case for PDBu (perhaps the PKC isoform(s) activated by DAG is not implicated in PLD stimula-FIG. 7. Proposed pathways for PA formation. PA can be produced by two reactions: phosphorylation of DAG by a DAG kinase and hydrolysis of PC by PLD. DAG is produced either from PIP 2 by a PI-PLC or from PC by a PC-PLC. These two enzymes are regulated by muscarinic receptors. PLD is activated by phorbol esters, partly via a PKC isoform that is not activated by muscarinic receptors. tion). Another mechanism explaining the secretory effect of carbachol had to be searched. A product derived from DAG may be involved. As DAG can be phosphorylated into PA by a DAG kinase, we analyzed the PA content of the cells and showed that the level of this phospholipid was increased by about 1.7-fold on carbachol stimulation. Thus, carbachol, on the one hand, enhanced the PA level in the cells and, on the other hand, increased protein secretion. These two results are well correlated and strengthen the conclusion that PA plays a role in the regulation of protein transit.
DAG (and consequently PA) was also reported to be produced by direct hydrolysis of PC by a PC-PLC; cholinergic and bradykinin receptors have been shown to activate this enzyme (40,41,46). In HT29-cl19A cells, we found that D609, a specific inhibitor of this enzyme (42), partly inhibited the production of PA triggered by carbachol. The two phospholipases may play a role in the physiological regulation of protein transit, as McKenzie et al. (41) showed that, in fibroblasts treated with carbachol, PI-PLC activation lead to a transient increase in DAG level, whereas PC-PLC activation would produce a sustained level of DAG. Since secretion is a phenomenon that lasts several hours, a sustained PC-PLC-dependent production of PA is not unlikely in our system.
Finally we show that PA production is not the sole mechanism controlling secretion, since we demonstrated that cAMP controls protein transit by a PA-independent pathway.
Our results allow us to describe a part of the regulation of protein secretion in HT29-cl19A cells according to the scheme presented in Fig. 7; the constitutive secretion depends on a basal PLD activity. A phorbol ester-triggered enhancement of the rate of protein transit is well correlated to a classical PLD-catalyzed PA production. Nevertheless, when the cells are stimulated via their cholinergic receptors, the increase in the rate of protein transit, also in good correlation with an enhancement of the amount of PA, implicates a non-classical pathway involving the activation of PLCs and DAG kinase.