Protein Kinase A-independent Activation of ERK and H,K-ATPase by cAMP in Native Kidney Cells

This study aimed at determining the signaling pathways underlying calcitonin- and isoproterenol-induced stimulation of H,K-ATPase in rat renal collecting duct. H,K-ATPase activity was determined in microdissected collecting ducts preincubated with or without either specific inhibitors or antibodies directed against intracellular signaling proteins. Transient cell membrane permeabilization with streptolysin-O allowed intracellular access of antibodies. The stimulation of H,K-ATPase by calcitonin and isoproterenol was mimicked by cAMP analogues and was abolished by adenylyl cyclase inhibition. Protein kinase A inhibition abolished isoproterenol but not calcitonin effect on H,K-ATPase. Calcitonin increased the phosphorylation of extracellular signal-regulated kinase (ERK) in a protein kinase A-independent manner, and the inhibition of the ERK phosphorylation prevented the stimulation of H,K-ATPase by calcitonin. Antibodies directed against either the cAMP-activated guanine-nucleotide exchange factor Epac I, the monomeric G protein Rap-1 or the kinase Raf-B, curtailed the stimulation of H,K-ATPase by calcitonin, whereas antibodies against the related monomeric G protein Ras or kinase Raf-1 had no effect. out a contribution of cyclic nucleotide-gated channels and increased [Ca 2 (cid:1) i in the mediation of Sct on H,K-ATPase activity. demonstrate the presence of a functional H89-sensitive cAMP/PKA signaling pathway in I (cid:1) cells.

The rat kidney cortical collecting duct (CCD) 1 consists of three intermingled cell types expressing specific hormone receptors positively coupled to adenylyl cyclase via G proteins (1): principal cells with vasopressin V 2 receptors, ␣-intercalated (I␣) cells with calcitonin receptors, and ␤-intercalated (I␤) cells with ␤-adrenergic receptors (1). Although it is difficult to physically separate pure populations of these different cell types, one can determine the cellular origin of cAMP-mediated responses on the basis of this hormone selectivity. Using this approach, we have previously reported that in rat CCD, the H,K-ATPase is present in both I␣ and I␤ cells in which it is stimulated by salmon calcitonin (Sct) and isoproterenol (Iso), respectively (2). The stimulation of H,K-ATPase, along with that of H-ATPase reported previously (3), probably participates in the regulation of proton transport by these two hormones (3,4).
The signaling mechanisms underlying the regulation of H,K-ATPase in I␣ and I␤ cells remain unclear for two reasons. Firstly, besides activating the Gs/adenylyl cyclase/cAMP/protein kinase A (PKA) pathway, calcitonin and isoproterenol also increase intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) in rat CCD (5)(6)(7). Secondly, several cAMP-binding proteins other than PKA have now been described: the cAMP receptor of Dictyostelium discoideum, which participates in the regulation of the development of this unicellular eukaryote (8), S-adenosyl homocysteine hydrolase, which participates in the regulation of protein methylation (9), cyclic nucleotide-gated channels involved in transduction of olfactory and visual signals (10,11), and cAMP-activated guanine-nucleotide exchange factors (GEF) (Epac I or cAMP-GEF I and CNrasGEF), which activate specifically the monomeric G proteins of the Ras family, Rap-1 and Ras, respectively (12)(13)(14). Thus, depending on the nature of cAMP-binding proteins expressed in different cell types and on their supramolecular organization, the effects of cAMP may be mediated by PKA-dependent or independent mechanisms. Interestingly, the kidney is one of the tissues displaying the highest levels of expression of mRNAs encoding Epac I (12,13) and the cyclic nucleotide-gated channel 3 (15), making possible an important role of these two proteins in the signalization of cAMP actions in kidney cells.
Thus, the aim of this study was to determine the signaling pathways underlying the stimulation of H,K-ATPase activity by Sct and Iso in the rat CCD, in particular the involvement of the cAMP-binding proteins PKA, Epac I, and cyclic nucleotidegated channels.
Pretreatment with Inhibitors and Hormones-Microdissected CCDs were preincubated with specific inhibitors of different intermediates of signaling pathways or with their solvent before incubation with hormones. Pools of CCDs dissected from collagenase-treated kidneys (2-4 mm in length) were pretreated with or without the inhibitors at 30°C for 45-120 min before treatment with hormones (10 min at 37°C). The hormone treatment was stopped by cooling the samples at 4°C. H89 dihydrochloride, myristoylated protein kinase A peptide inhibitor 14 -22 amide, and bisindolylmaleimide I were from Calbiochem. U0126 (1,4-diamino-2,3-dicyano-1,4-bis-(O-aminophenylmercapto)butadiene) was from Sigma. All inhibitors were prepared from aqueous solutions with the exception of U0126, which was dissolved in dimethyl sulfoxide (Me 2 SO). The corresponding control groups contained the same concentration of Me 2 SO (Ͻ0.1% v/v).
When specific inhibitors were not commercially available, we evaluated the effect of antibodies directed against several signaling proteins. Intracellular entry of antibodies was made possible by transient permeabilization of CCD cells by streptolysin-O (18,19). For this purpose, pools of CCDs were first preincubated at 37°C for 8 min and then for 90 min at 4°C with or without the antibody in a medium containing (in mM) 137 NaCl, 3 KCl, 5 glucose, 20 Pipes, 1 mg/ml bovine serum albumin, and 0.2 IU/ml streptolysin-O (Sigma) at pH 6.8. CCDs were then transferred into the usual dissection medium and incubated with hormones as described above. All antibodies used were from Santa Cruz Biotechnology (Tebu, Le Perray en Yvelines, France): affinity-purified rabbit polyclonal antibody against a peptide mapping at the C terminus of the ␣-catalytic subunit of human PKA (sc-903); affinity-purified goat polyclonal antibodies against a peptide mapping near the C terminus (catalytic domain) or at the N terminus (cAMP-binding domain) of human Epac I (sc-8880 and sc-8879, respectively); affinity-purified rabbit polyclonal antibody against a peptide mapping at the C terminus of human B-Raf (sc-166); rabbit polyclonal IgG against an epitope mapping near the C terminus of human Rap-1A (sc-65); a monoclonal IgG 1 antibody raised against a peptide mapping at the C terminus of Raf-1 p74 of human origin (sc-7267); and an affinity purified rat monoclonal antibody derived by fusion of spleen cells from a rat immunized with Y3Ag 1.2.3. rat myeloma cells (sc-35). Although all of these antibodies were directed against the active portion of the proteins (12,20,21), the anti-Ras antibody was the only one with reported inhibitory properties (22).
H,K-ATPase Activity Assay-H,K-ATPase activity was determined using the radiochemical microassay described previously (23) and adapted for microplate assay. To avoid contamination with Na,K-ATPase, CCDs were rinsed in a cold Na ϩ -free solution containing (in mM) 0.8 MgSO 4 , 1 MgCl 2 , 0.5 CaCl 2 , 100 Tris-HCl, 1 mg/ml bovine serum albumin, and mannitol up to 400 mosmol/kg at pH 7.4. After transfer within 0.5 l of rinsing solution into a 96-well flat bottom plastic microplate, samples were permeabilized by adding 2 l of hypotonic solution (10 mM Na ϩ -free Tris-HCl, pH 7.4, with or without 10 Ϫ4 M Sch-28080) to each sample and freezing on dry ice. After thawing and adding 10 l of assay medium (see composition below), the microplate was incubated at 37°C for 15 min. The incubation was stopped by cooling and adding 300 l of an ice-cold suspension of 15% (w/v) activated charcoal. After centrifugation, aliquots of 50 l of each supernatant were transferred to a 96-well sample microplate for Cerenkov counting (Trilux microbeta 1450, Wallac, Finland).
H,K-ATPase activity was distinguished from other ATP-hydrolyzing activities on the basis of its sensitivity to Sch-28080, a specific inhibitor of the renal H,K-ATPase activities (23). Thus, for each experimental condition, eight samples were divided into two groups, one for measuring total ATPase activity and the other for measuring the ATPase activity in presence of Sch-28080. H,K-ATPase activity was calculated as the difference between these two mean ATPase activities.
Western Blot Analysis of Phospho-ERKs-Pools of 30 CCDs dissected in medium supplemented with leupeptin (25 g ml Ϫ1 ) and aprotinin (25 g ml Ϫ1 ) were preincubated at 37°C for 1 h with or without H89 and then at 37°C for 10 min with or without hormones. Samples were then transferred with 0.5 l of incubation medium into 14.5 l of lysis buffer containing (in mM) 100 NaCl, 1.5 MgCl 2 , 2 sodium pyrophosphate, 2.5 glycerophosphate, 30 NaF, 1 EGTA, 20 Hepes, 1 g ml Ϫ1 leupeptin, 10 g ml Ϫ1 aprotinin, 10 g ml Ϫ1 aminoethylbenzenesulfonyl fluoride, and 10 g ml Ϫ1 antipain at pH 7.4, vortexed, and kept on ice for 1 h. Cell lysate was centrifuged at 15,000 ϫ g for 10 min, and the supernatant was removed and stored at Ϫ80°C until use.
In each experiment, the amounts of proteins loaded to each lane of the electrophoresis gel corresponded to the same initial length of isolated CCDs (Ϯ5%). The chemiluminescence of each lane was quantified by densitometry (arbitrary units) and expressed in each experiment as the percent of the control lane (no hormone treatment). Results are expressed as means Ϯ S.E. from several animals.
Expression of Epac I and Rap-1 mRNAs-RNAs were extracted from pools of 20 -50 microdissected glomerulus or nephron segments as previously described (17) and summarized below. The pools of tubules were transferred with 10 l of microdissection solution into 400 l of denaturing solution (4 M guanidium thiocyanate, 25 mM sodium citrate, pH 7.0, 0.1 M ␤-mercaptoethanol, and 0.5% sarcosyl) and 20 g of yeast RNA (Amersham Biosciences) used as a carrier. After phenol/chloroform extraction and isopropyl alcohol precipitation, the final RNA pellet was dried under vacuum and dissolved in RNA dilution buffer (10 mM Tris, pH 7.6, 1 mM EDTA, 2 mM dithiothreitol, 40 units/ml RNasin (Promega, Madison, WI)). The yield of this RNA extraction procedure is Ն90% (17).
Expression of mRNAs encoding Epac I and Rap-1 was evaluated by reverse transcription-PCR using the following specific primers: Epac I sense (5Ј-CGAGCAGGAACGCAGCACCTACATC-3Ј) and antisense (5Ј-TCACTTCCCTCACGGAGGCTGTCAC-3Ј), bases 1429 -1453 and 1811-1836 from the initiating ATG, respectively; and Rap-1 sense (5Ј-GCTCTGACAGTTCAGTTTGTTCAGGGAA-3Ј) and antisense (5Ј-AACATCTTCCGTGTCCTTAACCGGTAAA-3Ј), bases 6 -34 and 254 -282, respectively. Reverse transcription was carried out for 45 min at 45°C in a final volume of 50 l in the presence of RNAs from 1 glomerulus or 1 mm of tubular length, of antisense primer (6.25 pmol), dNTP (200 M), and Moloney murine leukemia virus reverse transcriptase (200 units). PCR was subsequently carried out in the same tube in a final volume of 100 l after the addition of the sense primer (6.25 pmol), 5 Ci [␣-32 P]dCTP (6000 Ci mmol Ϫ1 ), and Taq polymerase (1.25 units). Samples were submitted to 30 cycles of three temperature steps: 95°C for 30 s; 60°C for 30 s; and 72°C for 1 min with the exception of the last cycle in which the elongation lasted 10 min. The DNA fragments were separated by electrophoresis on 2% agarose slab gels, which were analyzed by PhosphorImager (Molecular Dynamics) for determining the intensity of the signals.
Measurement of Intracellular Ca 2ϩ Concentration-Intracellular calcium concentration was measured in single microdissected CCDs by using the calcium-sensitive fluorescence probe Fura-2/AM as described previously (5). After 1 h of loading with 10 M Fura-2/AM in presence or absence of 1 M PKA inhibitor H89, each CCD was transferred into a superfusion chamber on the stage of an inverted fluorescence microscope (Zeiss IM 35, Oberkochen, Germany). The tubules were superfused (3-4 ml min Ϫ1 ) at 37°C with a solution containing (in mM) 140 NaCl, 4 KCl, 1 MgCl 2 , 0.8 MgSO 4 , 0.44 NaH 2 PO 4 , 0.33 Na 2 HPO 4 , 1.0 CaCl 2 , and 5 glucose with or without 1 M H89 at pH 7.4. After 5-10-min equilibration period, calcitonin or ATP used as a positive control was added to the superfusate. The tubular portion selected for fluorescence measurements included approximately 30 cells. Double wavelength measurements of Fura-2/AM fluorescence were recorded every 2 s. The calculation of [Ca 2ϩ ] i was performed as described previously (5).
Statistics-Results are given as the means Ϯ S.E. from different animals. The data were compared according to either paired or unpaired Student's t test as appropriate or when comparing more than two groups together according to ANOVA test with PLSD Fisher. Preincubation with 1 M H89 (Fig. 1A) did not modify basal and Sct-stimulated H,K-ATPase activity but totally blocked the stimulation by Iso. Preincubation with 10 M myristoylated inhibitory peptide 14 -22 amide, another inhibitor of the PKA catalytic subunit, also blocked Iso-induced but not Sct-induced stimulation of H,K-ATPase (Fig. 1B). The effect of H89 and myristoylated inhibitory peptide 14 -22 amide were also mimicked by a polyclonal antibody directed against PKA catalytic subunit (Fig. 1C).

Role of Protein Kinase
In summary, these results demonstrate that Iso stimulates H,K-ATPase activity through the canonical PKA signaling pathway, whereas the effect of calcitonin is independent of PKA. Therefore, we attempted next to determine the signaling mechanism of calcitonin.
Role of cAMP in Sct-induced Stimulation of H,K-ATPase-To investigate whether or not cAMP is the second messenger that underlies Sct-induced stimulation of H,K-ATPase in I␣ cells, the requirement for adenylyl cyclase and cAMP was determined by three approaches. Firstly, CCDs were pretreated with 2.5 mM 2Ј,5Ј-dideoxyadenosine, an inhibitor of adenylyl cyclase. Results show that 2Ј,5Ј-dideoxyadenosine did not change basal H,K-ATPase activity but abolished its stimulation by Sct ( Fig. 2A). This finding demonstrates that adenylyl cyclase activity is essential for the stimulation of H,K-ATPase by Sct. Secondly, to further investigate the role of cAMP in Sct-induced enzyme activation, we determined the effects of the permeant cAMP analogue 8-Br-cAMP (100 M) and of forskolin (1 M). Because the results obtained with the two compounds were similar, they were pooled (Fig. 2B). The addition of 8-Br-cAMP or forskolin stimulated H,K-ATPase activity over 8-fold. The fact that this stimulatory factor (ϫ8) equals the sum of those observed in response to Iso (ϫ4) and Sct (ϫ4) suggests that 8-Br-cAMP and forskolin mimick the effects of the two hormones. This observation is further demonstrated by the finding that stimulation by 8-Br-cAMP or forskolin was inhibited by 50% with H89 (as expected from inhibition of ␤-intercalated cells). Finally, Sct-induced stimulation of H,K-ATPase was also mimicked by (R p )-cAMP (Fig. 2C), a diastereomer of cAMP that inhibits PKA (24).
These findings demonstrate that in I␣ cells, cAMP interacts with a cAMP-binding protein, different from PKA that mediates the stimulation of H,K-ATPase activity. Interestingly, this cAMP-binding protein is equally well activated by cAMP and by its analogue (R p )-cAMPs conversely to PKA regulatory subunit, which is inhibited by (R p )-cAMPs.
Lack of Implication of Cyclic Nucleotide-gated Channels in Sct-induced Stimulation of H,K-ATPase-Because cyclic nucleotide-gated channels 3 are expressed at a high level in the kidney (15) and because Sct increases [Ca 2ϩ ] i in rat CCD (5), cyclic nucleotide-gated channels and [Ca 2ϩ ] i might possibly mediate the stimulation of H,K-ATPase activity. However, Sctinduced increase in [Ca 2ϩ ] i was markedly reduced by pretreating CCDs with 1 M PKA-specific inhibitor H89 (Fig. 3). This finding demonstrates that these increases in [Ca 2ϩ ] i resulted from the activation of the canonical cAMP/PKA cascade. This

cAMP/ERK Interactions in Renal Epithelial Cells
conclusion is corroborated by the lack of stimulation of phospholipase C by calcitonin in CCDs (data not shown).
These findings rule out a contribution of cyclic nucleotidegated channels and increased [Ca 2ϩ ] i in the mediation of Sct on H,K-ATPase activity. They also demonstrate the presence of a functional H89-sensitive cAMP/PKA signaling pathway in I␣ cells.
Role of Epac I in Sct-induced Stimulation of H,K-ATPase-Epac I (also called cAMP-GEF I) is a guanine-nucleotide exchange factor specific of the monomeric G protein Rap-1, which is directly activated by cAMP (12,13). Because Epac I is expressed at high levels in kidneys (12,13), we evaluated whether it might be the target of cAMP involved in the stimulation of H,K-ATPase by Sct in I␣ cells. For this purpose, we determined (a) whether Epac I and its target Rap-1 are expressed in rat collecting duct and (b) the role of Epac I in Sct-induced stimulation of H,K-ATPase.
Reverse transcription-PCR on microdissected nephron segments revealed that mRNAs encoding Epac I (Fig. 4A) and Rap-1 (Fig. 4B) were expressed along the whole nephron. However, the highest levels of expression were found in CCDs, which is compatible with a possible involvement of Epac I and Rap-1 in Sct action in CCD I␣ cells. Indeed, the pretreatment of CCDs with polyclonal antibodies directed against either the C terminus (Fig. 5A) or the N terminus of Epac I (Fig. 5B) abolished Sct-induced stimulation of H,K-ATPase activity, whereas they did not alter the effect of Iso. Altogether, these results identify Epac I as the cAMP-binding protein involved in the stimulatory effect of calcitonin in CCD I␣ cells.
Signaling Pathway of Epac I in Sct-induced Stimulation of H,K-ATPase-Because Rap-1 activates B-Raf that in turn con-trols the ERK1/2 pathway (26), we evaluated (a) the roles of Rap-1 and B-Raf in Sct-induced stimulation of H,K-ATPase, (b) whether Sct activates ERK1/2 in CCDs, and (c) whether activation of ERK1/2 mediates the stimulation of H,K-ATPase.
The pretreatment of CCDs with a polyclonal antibody directed against an epitope close to the C terminus of Rap-1 curtailed the stimulatory effect of Sct (Fig. 6A). In contrast, the pretreatment with a monoclonal antibody that inhibits the activity of the related G protein Ras did not alter the effect of Sct on H,K-ATPase activity (Fig. 6B). Similarly, the pretreatment with an antibody directed against the C terminus of B-Raf abolished the stimulatory effect of Sct on H,K-ATPase, whereas a monoclonal antibody mapping a C terminus epitope of the related kinase Raf-1 had no effect (Fig. 6, C and D).
Western blot analysis using an antibody against phospho-ERKs p42 and p44 demonstrates that calcitonin increased the phosphorylation of ERK1/2 by Ͼ50% and that this effect was insensitive to the PKA inhibitor H89 (Fig. 7A). (R p )-cAMPs mimicked the effect of calcitonin on ERK1/2 phosphorylation (Fig. 7B). These findings are compatible with the involvement of ERK activation in the stimulation of H,K-ATPase by Sct. Indeed, the preincubation with 5 M U0126, an inhibitor of the ERK kinase MEK (27), abolished Sct-induced stimulation of H,K-ATPase (Fig. 7C). DISCUSSION This study shows that calcitonin-induced stimulation of H,K-ATPase in I␣ cells of rat collecting duct is mediated by cAMP but is independent of PKA. Instead, it involves Epac I, the guanine-nucleotide exchange factor of the monomeric G protein Rap-1, whose activation triggers B-Raf and the MAPK kinase ERKs.
Study of Signaling Pathways in Native Cells-A myriad of molecules participates in the intracellular control of cell functions. These molecules can constitute different signaling networks according to their level of expression, the presence of their various isoforms and their subcellular compartmentalization. Thus, the signaling pathways characterized in cultured and/or transfected cells may not be relevant in native cells under physiological conditions, especially in highly differentiated cells such as kidney epithelial cells. However, deciphering signaling pathways in native cells remains difficult because we lack specific pharmacological inhibitors for most intracellular signaling molecules.
In this study, we tentatively used antibodies directed against signaling proteins as specific inhibitors based on the hypothesis that the binding of an immunoglobulin on or close to the active site of a protein should alter its activity and/or its interaction with other proteins located upstream or downstream the signalization cascade. Although an inhibitory activity was previously demonstrated only for the anti-Ras antibody (22), the five antibodies, which were tested in this study, proved to display a biological effect. Furthermore, the inhibitory effect of an antibody may be as potent as that of a pharmacological inhibitor as illustrated in Fig. 1 for the anti PKA antibody. In addition, the biological effects of the antibodies were specific. (a) As shown on Fig. 1C, the anti-PKA antibody inhibited the effect of isoproterenol in I␤ cells but not that of calcitonin in I␣ cells, whereas two different anti Epac I antibodies curtailed calcitonin but not isoproterenol action (Fig. 5). (b) Antibodies directed against Rap-1 and B-Raf abolished calcitonin action, whereas antibodies directed against Ras and Raf-1, two related proteins of the same families, had no effect (Fig. 6).
Finally, transient permeabilization of cell membranes with streptolysin-O is a convenient method for integrating macromolecules into native epithelial cells, because it allows to maintain a cell integrity sufficient to preserve the activity of H,K- Afterward, samples were superfused at 37°C without or with H89 with a solution containing 1 mM Ca 2ϩ (instead of 0.5 mM for H,K-ATPase measurements) to amplify the magnitude of Ca 2ϩ signals secondary to Ca 2ϩ entry. After a 10-min equilibration period necessary to reach a stable [Ca 2ϩ ] i level, tubules were stimulated successively with Sct (10 nM) and ATP (100 M), a nucleotide known to stimulate phospholipase C in this segment through extracellular P 2 Y receptors (25). A and B, representative traces obtained in the same rat from a control and a H89-treated tubule respectively. C, mean [Ca 2ϩ ] i increases (peak minus basal) calculated for each agonist from seven controls (solid bars) and seven H89-treated tubules (open bars) isolated from four rats. *, p Ͻ 0.001 versus control (no H89) by unpaired Student's t test.

cAMP/ERK Interactions in Renal Epithelial Cells
ATPase as well as its stimulation by hormones (see control activities in Figs. 1C, 5, and 6). This method could be useful to integrate recombinant proteins with constitutive activating or inhibiting mutations (positive or negative dominants).
Regulation of H,K-ATPase by Calcitonin and Isoproterenol in Rat CCD-intercalated Cells-Although calcitonin and isoproterenol receptors can couple to effectors other than adenylyl cyclase (28 -30), cAMP is responsible for the stimulation of H,K-ATPase in both I␣ and I␤ cells. Indeed, cAMP analogues or forskolin mimicked the hormone effects (Fig. 2, B and C), and inhibition of adenylyl cyclase prevented them ( Fig. 2A). In I␤ cells, cAMP effect on H,K-ATPase activity is mediated by PKA (Fig. 1) by a mechanism that has not been investigated. In contrast, in I␣ cells, the blockade of PKA did not alter calcitonin-induced stimulation of H,K-ATPase (Fig. 1) despite the presence of a functional enzyme attested by its involvement in Sct-induced rise in [Ca 2ϩ ] i (Fig. 3). Instead, the stimulatory effect of cAMP on H,K-ATPase activity in I␣ cells is mediated through its binding to Epac I (Fig. 5). Data from the literature indicate that the binding of cAMP to Epac I stimulates its guanine-nucleotide exchange activity and thereby activates the monomeric G protein Rap-1 (12, 13). Activated Rap-1 stimulates the kinase activity of B-Raf and the phosphorylation of the MAPK kinase MEK (26,31), which in turn phosphorylates and activates ERK. The results from this study demonstrate the key roles of Rap-1, B-Raf, and ERK (Figs. 6 and 7) in the signalization of calcitonin action on H,K-ATPase in I␣ cells (summarized in Fig. 8).
The involvement of ERKs in the regulation of H,K-ATPase activity in rat CCD I␣ cells is demonstrated by the following. (a) Calcitonin increased the phosphorylation of ERKs (Fig. 7A). (b) The regulation of ERK and H,K-ATPase by calcitonin were both independent of PKA ( Figs. 1 and 7A). (c) The inhibition of the ERK kinase MEK abolished the stimulatory effects of calcitonin on H,K-ATPase (Fig. 7C).
Because PKA was shown to stimulate ERK through direct activation of either Rap-1 (32) or B-Raf (31), the Epac I-and PKA-dependent mechanisms of ERK activation could coexist in the same cell. However, the finding that the anti-Epac I antibodies fully inhibited the effect of calcitonin on H,K-ATPase (Fig. 5) shows that PKA is not able to activate Rap-1 or B-Raf in I␣ cells. This result suggests that the PKA pathway and the Epac I/Rap-1/B-Raf pathway are segregated into distinct subcellular compartments in CCD I␣ cells.
In this study, the mechanism of H,K-ATPase activation beyond ERK has not been investigated. In most cells, the activation of ERK triggers its translocation into the nucleus where it activates transcription factors through direct phosphorylation (Elk-1, Myc, cAMP-response element-binding protein, Sap-1) or through phosphorylation of kinases such as p90RSK. In turn, these activated transcription factors control the expression of specific genes. The rapid time course of calcitonin-induced activation of H,K-ATPase activity (Ͻ10 min) rules out a possible mediation through de novo synthesis of its subunits and suggests an activation of preexisting ATPase units. Several observations support that calcitonin may stimulate H,K-ATPase activity through exocytotic insertion into the plasma membrane of H,K-ATPase units present in the membrane of cytoplasmic vesicles. (a) H,K-ATPase colocalizes with H-ATPase (33), an other proton pump that is controlled through exocytosis/endocytosis in CCD I␣ cells (34). (b) In gastric mucosa, functional expression of H,K-ATPase at the luminal membrane of parietal cells is achieved through exocytosis (35). (c) ERKs control exocytosis in several cell types (36 -38).
In conclusion, this study provides original data regarding several aspects of protein G-coupled receptor signalization in native kidney cells. (a) It outlines the important role of Epac I in mediating some effects of cAMP in collecting duct cells. (b) It demonstrates an intracellular compartmentalization of cAMP effects. (c) It indicates that ion transporters such as H,K-ATPase can be extranuclear effectors of ERK signalization. The expression of Epac I in other nephron segments (Fig. 4A) suggests that many other effects of cAMP-producing hormones, usually attributed to the activation of PKA, may be mediated in fact through the activation of Epac I and ERKs.