Cyclic GMP Kinase II (cGKII) Inhibits NHE3 by Altering Its Trafficking and Phosphorylating NHE3 at Three Required Sites

Background: cGKII acutely inhibits brush-border NHE3, but the mechanism is unknown. Results: cGMP/cGKII phosphorylates NHE3 at three sites. All are necessary for NHE3 inhibition. One of these is also phosphorylated by SGK1 to stimulate NHE3. Conclusion: cGKII inhibits NHE3 by phosphorylating and reducing NHE3 surface amount. Significance: Phosphorylation of the same site in a protein can alter function differently based on phosphorylation of additional sites. The epithelial brush-border Na+/H+ exchanger NHE3 is acutely inhibited by cGKII/cGMP, but how cGKII inhibits NHE3 is unknown. This study tested the hypothesis that cGMP inhibits NHE3 by phosphorylating it and altering its membrane trafficking. Studies were carried out in PS120/NHERF2 and in Caco-2/Bbe cells overexpressing HA-NHE3 and cGKII, and in mouse ileum. NHE3 activity was measured with 2′,7′-bis(carboxyethyl)-S-(and 6)carboxyfluorescein acetoxy methylester/fluorometry. Surface NHE3 was determined by cell surface biotinylation. Identification of NHE3 phosphorylation sites was by iTRAQ/LC-MS/MS with TiO2 enrichment and immunoblotting with specific anti-phospho-NHE3 antibodies. cGMP/cGKII rapidly inhibited NHE3, which was associated with reduced surface NHE3. cGMP/cGKII increased NHE3 phosphorylation at three sites (rabbit Ser554, Ser607, and Ser663, equivalent to mouse Ser552, Ser605, and Ser659), all of which had to be present at the same time for cGMP to inhibit NHE3. NHE3-Ser663 phosphorylation was not necessary for cAMP inhibition of NHE3. Dexamethasone (4 h) stimulated wild type NHE3 activity and increased surface expression but failed to stimulate NHE3 activity or increase surface expression when NHE3 was mutated to either S663A or S663D. We conclude that 1) cGMP inhibition of NHE3 is associated with phosphorylation of NHE3 at Ser554, Ser607, and Ser663, all of which are necessary for cGMP/cGKII to inhibit NHE3. 2) Dexamethasone stimulates NHE3 by phosphorylation of a single site, Ser663. The requirement for three phosphorylation sites in NHE3 for cGKII inhibition, and for phosphorylation of one of these sites for dexamethasone stimulation of NHE3, is a unique example of regulation by phosphorylation.

phorylation of one of these sites for dexamethasone stimulation of NHE3, is a unique example of regulation by phosphorylation.
The Na ϩ /H ϩ exchanger NHE3 is expressed in the intestinal brush border. NHE3 is regulated by second messengers during digestion and inhibited in diarrheal diseases. cAMP and cGMP both inhibit NHE3 activity when added directly to intestinal cells and when intracellular cAMP and cGMP levels are increased by agents that produce diarrhea, for example, cholera toxin for cAMP and heat stable Escherichia coli enterotoxin for cGMP. cAMP acts by phosphorylating NHE3 primarily at two amino acids in the mouse intestine and kidney: Ser 552 and Ser 605 (our studies are with rabbit NHE3 and we refer to these amino acids as Ser 554 and Ser 607 ). Other cAMP consensus sequences in NHE3 have been reported, including Ser 330 , Ser 514 , Ser 576 , Ser 662 , Ser 691 , Ser 692 , and Ser 805 . However, none of these have been shown to affect NHE3 activity (1)(2)(3). cAMP also reduces the amount of plasma membrane NHE3, which appears to occur by an additional process that follows phosphorylation, and is likely to be stimulation by endocytosis (4). In contrast, how elevated cGMP, which is known to act via brushborder cGKII, inhibits NHE3 activity is unknown. Information lacking includes whether cGMP/cGKII directly phosphorylates NHE3 in vivo and what are the consequences of such phosphorylation on NHE3 trafficking.
In some cases, cGMP regulates intracellular events by mechanisms analogous to those demonstrated for cAMP. However, the effects of cGMP in the small intestine are not fully understood. The intrinsic ileal peptide guanylin and the E. coli heatstable enterotoxin (STa) both bind to the same brush-border receptor, guanylate cyclase C, and subsequently increase intracellular cGMP content within minutes (5). STa, guanylin, and cGMP all rapidly inhibit small intestinal NaCl-linked absorp-tion, principally at the level of NHE3, which is an essential component of this sodium-absorptive process (5). This effect on NHE3 is specific because other brush-border transporters, including NHE2 and SGLT1, are not acutely altered during this process. The downstream effect of cGMP on ion and fluid transport in the small intestinal enterocytes appears to occur entirely via activation of the type II isoform of cGMP-dependent protein kinase in the brush border (6,7), which we showed previously was part of a NHE3 signaling complex (8). Moreover, previous studies of cGKII identified a number of its phosphorylated substrates, all of which were also phosphorylated by PKA (9).
NHE3 and cAMP-dependent protein kinase type II (PKAII) are part of the same signaling complex that is scaffolded by either NHERF1 or NHERF2, which are multi-PDZ domain scaffolding proteins (10,11). Based on the cell type, cAMP inhibition of NHE3 requires NHERF1 or NHERF2 (12), both of which bind ezrin, which is currently thought to act as an A-kinase anchoring protein (AKAP) to position PKAII so it can phosphorylate NHE3 (11,13,14). Nonetheless, the role of ezrin in NHE3 phosphorylation has been questioned recently (15). We previously reported that cGMP inhibition of NHE3 requires NHERF2, an effect not duplicated by NHERF1, and that NHERF2 links cGKII into a NHE3 signaling complex (8).
Sites of NHE3 phosphorylation by cGMP/cGKII have not been identified. Therefore, the current study tested the hypothesis that cGMP regulates the brush-border Na ϩ /H ϩ exchanger NHE3 by phosphorylating it at specific sites to reduce its plasma membrane expression.
Cell Culture-For transport assays, stably transfected PS120 cells were grown on glass coverslips to ϳ70% confluence and studied after serum starvation for 3 h. For cGKII virus or empty vector infection, 50 -60% confluent cells were incubated with cGKII viral particles or empty vector in serum-free medium at 37°C for 6 h and then replaced with normal growth medium. OK proximal tubule cells, which lack endogenous NHERF2 and cGKII, were stably transfected with rabbit NHE3 with the HA 3 epitope tag at the N terminus and human NHERF2. OK cells were maintained in high glucose Dulbecco's modified Eagle's medium supplemented with 25 mM NaHCO 3 , 10 mM HEPES, 1 mM sodium pyruvate, 10% fetal bovine serum, 100 units/ml of penicillin, and 50 g/ml of streptomycin, 400 g/ml of G418 (Neomycin) and/or 600 g/ml of hygromycin in a 5% CO 2 , 95% O 2 humidified incubator at 37°C. For cGKII virus or empty vector infection, OK cells were treated with 6 mM EGTA in serum-free OK medium for 2 h at 37°C before viral infection to allow virus exposure to the basolateral in addition to the apical surface. Appropriate amounts of cGKII viral particles (10 9 -10 10 particles/ml) were diluted in serum-free OK medium and added to EGTA-treated cells. Cells were infected by incubating at 37°C for 6 h and then the medium was replaced with normal growth medium. For transport assays, OK cells were grown to confluence on glass coverslips and polarized cells were used ϳ48 h after infection.
The Caco-2/Bbe cell line, originally derived from a human adenocarcinoma, was obtained from M. Mooseker (Yale University, New Haven, CT) and J. Turner (University of Chicago, Chicago, IL) (20). Cells were maintained in DMEM ϩ 25 mM NaHCO 3 , 0.1 mM nonessential amino acids, 10% fetal bovine serum, glutamine (4 mM), 50 units/ml of penicillin, and 50 g/ml of streptomycin, pH 7.4, at 37°C in 5% CO 2 . To achieve polarity, cells were cultured on 6-well collagen-coated dishes (Transwell filters, 0.02 m, Corning) for 12 days after becoming confluent. For virus infection, cells were first exposed to 6 mM EGTA in serum-free medium for 2 h at 37°C to increase virus access to the basolateral surface; then 10 9 -10 10 viral particles/ml in serum-free medium were added to both Transwell chambers. After incubating for more than 6 h, cells were rinsed and incubated further in normal growth medium.
Mass Spectrometry Analysis and Database Searching-Total membrane proteins were extracted from mouse ileum pretreated with an equal concentration of DMSO as a control or 100 M 8-pCPT-cGMP. Each protein extract was digested with trypsin and the peptides were labeled with iTRAQ reagent according to the manufacturer's instructions. iTRAQ-labeled peptides from different samples (up to 8) were mixed and separated into 24 fractions by strong cation exchange chromatography. Ten percent of the peptides in each strong cation exchange fraction were analyzed by reverse phase chromatography coupled to a tandem mass spectrometer (LC-MS/MS) using a Waters Nanoacquity UPLC interfaced with a Thermo Scientific LTQ Orbitrap Velos with 0.03 Da mass accuracy on fragment ions. The remaining 90% of the strong cation exchange fractions were enriched for phosphopeptides using TiO 2 before LC-MS/MS analysis. The MS/MS spectra were extracted and searched for phosphorylation on Ser, Thr, and Tyr against the Refseq database using Mascot (Matrix Science) through Proteome Discoverer software (version 1.2, Thermo Scientific) with extract, specifying species, enzyme allowing one missed cleavage, fixed cysteine methylthiolation and variable methionine oxidation, and 8-plex-iTRAQ labeling of N-terminal Lys and Tyr. Peptide identifications from Mascot searches were processed within Proteome Discoverer to identify peptides with a confidence threshold 1% false discovery rate, based on a concatenated decoy database search to calculate the median protein and peptide ratios. Technical variation was Ͻ10% as determined from four iTRAQ 8-plex experiments using 32 technical replicates of immuno-depleted human plasma samples.
Normalization of Mass Spectrometry Data-Quantitative mass spectrometry data from total membrane proteins from the ileum with and without exposure to 8-pCPT-cGMP underwent normalization to quantitate differential phosphorylation at three sites in NHE3. Peptides from tryptic digests of protein extracted from control or cGMP-treated ileum were labeled with unique iTRAQ isobaric mass tags and identified using mass spectrometry before and after TiO 2 enrichment for phosphopeptides. Fragmentation spectra revealed the peptide sequences and reporter ions corresponding to each iTRAQ tag. The relative intensities of the reported ions reflect the relative amounts of the peptide present in each sample. To normalize for variations in sample preparation, the median reporter ion signal per sample was first determined for all peptide sequences in a sample. All sample values were then normalized to minimize technical variations between samples, such as amount of loading. The median value for each sample was divided by the normalized median of the total sample to determine their relative concentrations, and thus the normalization factor per sample. Because both samples from control and the 8-pCPT-cGMP samples were mixed prior to MS analysis and TiO 2 enrichment of phosphopeptides, the same sample-specific normalization factor was applied to each phosphopeptide identified after TiO 2 enrichment. This yielded the relative fold-changes in the three NHE3 phosphopeptides as shown in Fig. 3, A-C.
Measurement of Na ϩ /H ϩ Exchange Activity-Na ϩ /H ϩ exchange activity in cells expressing HA-NHE3 was determined fluorometrically using the intracellular pH-sensitive dye BCECF-AM (10 M). Polarized OK cells on glass coverslips were studied 3 days after confluence. PS120 cells were studied when ϳ70% confluent as described previously (21 7.4), which was then switched to Na ϩ solution (130 mM NaCl instead of TMA-Cl) for the Na ϩ -dependent pH i recovery. At the end of each experiment, the fluorescence ratio was calibrated to pH i using the high potassium/nigericin method (17,19). For PS120 cells, Na ϩ /H ϩ exchange activity data were calculated as the product of the change in pH i (⌬pH i /dt) ϫ the intracellular buffering capacity at each pH i at multiple pH i values using at least three coverslips per condition in a single experiment. Kinetics of Na ϩ /H ϩ exchange were analyzed by a Hill plot using Origin (Microcal Software) to estimate V max and KЈ(H ϩ ) i in an individual experiment (19). Mean Ϯ S.E. were determined from at least three experiments. For OK epithelial cells, initial rates of Na ϩ -dependent intracellular alkalinization were calculated starting at pH i 6.3 over the initial 1-min Na ϩ exposure (linear phase of pH i change) and expressed as ⌬pH/⌬T in an individual experiment. Mean Ϯ S.E. were determined from at least three experiments.
Buffering Capacity (mM/pH) was calculated for each pH interval according to the following equation (22), where the intracellular NH 4 ϩ concentration was calculated according to the Henderson-Hasselbach equation, [NH 3 ] i was ignored in the physiological pH range (pH Ͻ 7.5) and the pH for each buffering capacity interval was determined as the average value over that pH interval. Quantitation of Surface NHE3 by Biotinylation-NHS-SSbiotin was used to determine the percentage of total NHE3 on the surface of PS120 cells, as described previously (23). PS120 cells grown on 10-cm dishes were infected with cGKII virus or empty vector and studied 48 h later. PS120 cells were serumstarved for 3 h, rinsed three times in cold PBS (150 mM NaCl, 20 mM Na 2 HPO 4 , pH 7.4), and once in borate buffer (154 mM NaCl, 1.0 mM boric acid, 7.2 mM KCl, and 1.8 mM CaCl 2 , pH 8.0), then 1 mg/ml of NHS-SS-biotin in borate buffer was added at 0 and 30 min. Cells were then washed three times with quenching buffer (120 mM NaCl, 20 mM Tris, pH 7.4) to remove unreacted biotin, washed three times in PBS and solubilized with 1 ml of N ϩ buffer (60 mM HEPES, pH 7.4,150 mM NaCl, 3 mM KCl, 5 mM Na 3 EDTA, 3 mM EGTA, and 1% Triton X-100). The cell lysate was centrifuged at 5000 rpm for 10 min and the postnuclear supernatant was incubated with streptavidin-agarose beads from 3 h to overnight at 4°C. After precipitation, the supernatant was retained as the intracellular fraction and the avidin-agarose beads were washed five times in N Ϫ buffer (60 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM KCl, 5 mM Na 3 EDTA, and 3 mM EGTA) with 0.1% Triton X-100 to remove nonspecifically bound proteins. The avidin-agarose bead-bound proteins, which represent plasma membrane NHE3, were solubilized in loading buffer (5 mM Tris-HCl, pH 6.8, 1% SDS, 10% glycerol, and 1% 2-mercaptoethanol), boiled for 10 min, size fractionated by SDS-PAGE (10% gel), and then electrophoretically transferred to nitrocellulose. After blocking with 5% nonfat milk in PBS, the blots were probed with monoclonal anti-HA antibody. Western analysis and quantification of the surface fractions were performed using the Odyssey system and software (LI-COR). Multiple volumes of the total and surface samples were used, and the signal intensity was derived by linear regression to obtain a single value for each sample. The percentage of surface NHE3 was calculated from a ratio ((surface NHE3 signal/total NHE3 signal) ϫ dilution factor of surface and total NHE3 samples) and expressed as percentage of total NHE3 with normalization to GAPDH.
Isolation of Ileal Total Membranes from Mice and Lysate from Cells and Immunoblotting-Ileum was rinsed with icecold 0.9% saline and opened along the mesenteric border. Ileum harvested as above was preincubated for 15 min at 37°C in Ringer's HCO 3 , 25 mM glucose, gassed at 95% O 2 , 5% CO 2 , and then exposed at 37°C to 20 M forskolin for 15 min or 100 M 8-pCPT-cGMP or an equal concentration of DMSO (as control) for 30 min. Then ileal villus cells were scraped and placed into homogenization buffer A (in mM: 60 mannitol, 2.4 Tris, pH 7.1, 1 EGTA, 1 ␤-glycerol-PO 4 , 2 Na 3 VO 4 , and 1 phenylalanine). Protease inhibitor mixture (1:100; Sigma) and phosphorhamidon (1:1,000) were added to the buffer. Cells were then homogenized at 4°C with a Polytron (10 times for 10 s at speed 5 with a 20-s interval between each burst) followed by homogenizationofsamplesinaglass-Teflonhomogenizer.Thehomogenates were centrifuged at 5,000 ϫ g for 10 min at 4°C to remove cell debris and nuclei and designated as total lysates. Lysates were then centrifuged at 40,000 ϫ g for 60 min, and total membrane (TM) pellets were collected. The resulting TMs were resuspended in buffer B (in mM: 300 mannitol, 20 HEPES, pH 7.4, 5 Mg-gluconate, 1 Na 3 VO 4 , 1 ␤-glycerol-PO 4 , 1 phenylalanine, and protease inhibitors added as in buffer A). The protein concentrations of the TM samples were measured with Bradford protein assay (Sigma) and subsequently analyzed by SDS-PAGE and Western blotting using primary antibodies for NHE3, phospho-NHE3 (Ser 554 and Ser 607 ), and actin with fluorescently labeled secondary goat anti-mouse or anti-rabbit IRDye TM 700 or TM 800 antibodies (Rockland). The fluorescence intensity of detected protein bands was visualized by the Odyssey system (LI-COR).
Statistics-Results were expressed as mean Ϯ S.E. Statistical evaluation was by analysis of variance or Student's t test.

cGMP/cGKII Inhibits NHE3 in PS120 and Polarized OK Cells in a Time-dependent Manner and Reduces NHE3 Surface
Expression-Our previous studies done in PS120, OK, or Caco-2/Bbe cells, and mouse ileum showed that cGMP inhibits NHE3 activity using a NHERF2 and cGKII-dependent process (8,12,24). To identify the mechanism for this cGMP regulation, we first defined the time course of cGMP inhibition of NHE3. Studies were performed in PS120/NHERF2/NHE3 and OK/NHERF2/NHE3 cells, both transiently infected with replication-deficient adenovirus containing the coding sequence of rat cGKII (25). OK/NHERF2/NHE3 cells were exposed to 100 M 8-pCPT-cGMP for 15, 30, 45, and 60 min and then Na ϩ /H ϩ exchange activity was measured. cGMP inhibition of NHE3 activity occurred rapidly, was significant at 15 min after addition of 8-pCPT-cGMP, maximum at 45 min, and lasted for at least 60 min (Fig. 1, A and B). cGMP inhibition of NHE3 also occurred rapidly in PS120/NHERF2/NHE3 cells, being significant by 15 min and maximum by 60 min (Fig. 1, C and D). Buffering capacity was not changed by cGMP, cGKII, or both (Fig. 1E). As a negative control, the inhibition of NHE3 by 8-pCPT-cGMP did not occur in cells infected with empty vector and cells without NHE3 (Fig. 1F).
Rapid regulation of NHE3 activity in epithelial cells usually involves, at least in part, changes in the amount of NHE3 at the plasma membrane by changes in regulated endocytosis and/or exocytosis (26,27). We therefore tested the hypothesis that cGMP inhibition of NHE3 activity might occur by a reduction in the amount of NHE3 on the cell surface. The surface NHE3 amount was determined by cell surface biotinylation in PS120 cells after exposure to 100 M 8-pCPT-cGMP for 45 min. As shown in Fig. 1, G and H, cell surface NHE3 was reduced ϳ35% by 8-pCPT-cGMP exposure (p Ͻ 0.05) without a change in total NHE3 expression. Thus, cGMP inhibition of NHE3 activity includes a decrease in NHE3 surface amount, which is consis-cGKII Phosphorylates NHE3 tent with a role for trafficking in this effect. In these studies, normalization of total NHE3 was to GAPDH with different conclusions if actin were used for normalization. We speculate that this might be due to changes in cytoskeleton caused by cGMP/cGKII. cGMP/cGKII Increases NHE3 Phosphorylation at Ser 554 and Ser 607 , Which Does Not Require PKA-Multiple studies have implicated that two sites of NHE3 phosphorylation are neces-sary for inhibition of NHE3 in response to stimuli, which act by elevating cAMP/PKA, including dopamine (28 -30). In many cases, cGMP regulates intracellular events by mechanisms analogous to those demonstrated for cAMP. Therefore, we tested the hypothesis that cGMP/cGKII phosphorylates NHE3 at the same sites as cAMP. Studies were undertaken by determining NHE3 phosphorylation using well characterized phospho-specific antibodies, anti-Ser(P) 554 and anti-Ser(P) 607 , in  Fig. 2, A-C, both Ser 554 and Ser 607 were significantly phosphorylated by cGMP/cGKII (p Ͻ 0.05) in the same samples in which cAMP/PKA phosphorylated these amino acids. The cGMP/cGKII-, plus cAMP-induced phosphorylations were slightly but not significantly greater than either alone.

JOURNAL OF BIOLOGICAL CHEMISTRY 1957
In some cases cGMP acts via PKA (31,32). Thus, we determined whether the phosphorylation of NHE3 by cGMP required PKA. H89 at low concentrations is a relatively specific inhibitor of PKA (30,(33)(34)(35). PS120 and Caco-2/Bbe cells were preincubated with 10 M H89 for 30 min, 20 M forskolin for 15 min, or 100 M 8-Br-cAMP or 100 M 8-pCPT-cGMP for 30 min at 37°C. Immunoblotting was done with anti-HA and anti-Ser(P) 554 antibodies. As shown in Fig. 2, D and E, H89 significantly prevented the phosphorylation of NHE3 at Ser 554 by 8-Br-cAMP and forskolin in PS120 (Fig. 2D) and Caco-2/Bbe cells (Fig. 2E), respectively (p Ͻ 0.05). However, H89 did not prevent the phosphorylation of NHE3 at Ser 554 by 8-pCPT-cGMP/cGKII (Fig. 2, D and E). These results support that phosphorylation of NHE3 at Ser 554 by cGMP/cGKII does not involve PKA and that both cGKII and PKA stimulate the phosphorylation of NHE3-Ser 554 .

Quantitative Mass Spectrometry and Truncation Studies Identify a Novel cGMP/cGKII-induced NHE3 Phosphorylation
Site, Ser 663 -Previous studies of cGKII have identified a number of its phosphorylated substrates, all of which have also been phosphorylated by PKA (9,37,38). To identify cGMP-specific phosphorylation sites in NHE3, if they exist, total membrane proteins were extracted from mouse ileum by ultracentrifugation and digested by trypsin. The generated tryptic peptides were then analyzed by quantitative mass spectrometry. As shown in Fig. 3, A-C, three of the identified phosphopeptides from mouse ileum containing Ser 659 , Ser 789 , and Ser 808 , had the amount of phosphorylation increased by cGMP. These are equivalent to Ser 663 , Ser 797 , and Ser 816 in rabbit NHE3.
Truncation mutations were performed to determine whether any of these sites was functionally important and required for cGMP inhibition of NHE3 activity. As shown in Fig. 3D, unlike cGMP inhibition of wild type NHE3 activity, cGMP had no effect on NHE3 truncated to amino acid 660 (NHE3/⌬660). However, cGMP inhibited the activity of NHE3 truncated to amino acid 690 (NHE3/⌬690V) to a similar extent as that in wild type cells. These results are consistent with NHE3 phosphorylation at Ser 663 but not Ser 797 and Ser 816 being necessary for cGMP/cGKII inhibition of NHE3 activity.
Ser 554 , Ser 607 , and Ser 663 Must All Be Present for cGMP/cGKII to Inhibit NHE3-Having shown that cGMP/cGKII stimulates the phosphorylation of NHE3 at three sites, Ser 554 , Ser 607 , and Ser 663 , we next studied how these three sites affect the cGMP regulation of NHE3 activity. All three sites were singly mutated from Ser to Ala and Ser to Asp, and triple mutated from Ser to Asp. The engineered constructs were transfected into PS120/ NHERF2 cells to make stable cell lines: NHE3-S554A and -S554D, NHE3-S607A and -S607D, NHE3-S663A and -S663D, and NHE3-S554D,S607D,S663D. As shown in Fig. 4, A and B, unlike wild type NHE3 containing cells, which had significant inhibition of NHE3 activity after being exposed to 8-pCPT-cGMP (p Ͻ 0.05), single mutant S554A, S607A, and S663A cells completely lost the cGMP/cGKII inhibition of NHE3 activity. Three phosphomimetic mutations, NHE3-S554D, NHE3-S607D, NHE3-S663D, however, were inhibited by cGMP similarly to wild type NHE3 (p Ͻ 0.05). These results indicate that the inhibition of NHE3 activity by cGMP requires phosphorylation of all three sites, Ser 554 , Ser 607 , and Ser 663 . To confirm these results, NHE3 activity was measured in NHE3-S554D,S607D,S663D triple mutant cells infected with cGKII. As shown in Fig. 4C, Ser to Asp triple mutant cells lost the inhibition of NHE3 by 8-pCPT-cGMP/cGKII. This result confirmed that changes in phosphorylation in all three sites are required for 8-pCPT-cGMP/cGKII inhibition of NHE3 activity.
These mutants also had changes in basal NHE3 activity, which was dramatically increased in the single mutant NHE3-S663A cell line (p Ͻ 0.05) and significantly decreased in single mutant NHE3-S554A, -S554D, -S607A, -S607D and triple mutant NHE3-S554D,S607D,S663D cell lines. To evaluate the explanation for the increase in basal activity of S663A and decrease in basal activity of S554A, S554D, S607A, and S607D, total NHE3 and the percentage of the surface amounts were determined. As shown in Fig. 4, D-I, there was no significant change in total NHE3 expression in NHE3-S554A, -S554D, -S607A, -S607D, -S663A, and -S663D mutant cells compared with wild type cells. However, the amount of surface NHE3 expressed as a percent of total NHE3 was significantly decreased in NHE3-S554A, -S554D, -S607A, and -S607D (p Ͻ 0.05) and significantly increased in S663A cells (p Ͻ 0.05), but there was no significant change in S663D cells (Fig. 4I). Surface NHE3 was significantly decreased by 8-pCPT-cGMP in NHE3-S554D and -S607D cell lines, but not in NHE3-S554A, -S607A, and -S663A cell lines, which is consistent with changes in NHE3 activity by 8-pCPT-cGMP. The surface amount of NHE3 was not reduced by 8-pCPT-cGMP (Fig. 4, H and I) in S663D mutant cells, which is different from the inhibition of activity of this mutant by 8-pCPT-cGMP. These data further support our hypothesis that cGMP regulates the brush-border Na ϩ /H ϩ exchanger NHE3 by stimulating the phosphorylation of NHE3 at specific sites to reduce its plasma membrane expression.
To determine whether Ser 663 is a specific site for cGMP/ cGKII inhibition of NHE3 activity and not for PKA, PS120/ NHERF2/NHE3-S663A cells were preincubated with FSK, and NHE3 activity was determined. FSK significantly inhibited NHE3 activity to a similar extent in both wild type cells and S663A-mutated cells (p Ͻ 0.05) (Fig. 4J). These data showed that Ser 554 , Ser 607 , and Ser 663 all are required for cGMP/cGKII to inhibit NHE3, but only Ser 663 is a specific site for cGKII inhibition of NHE3.
Dexamethasone Fails to Stimulate NHE3 Activity or Increase Surface Expression When NHE3 Is Mutated to S663A or S663D-NHE3-Ser 663 has been reported to be a specific phosphorylation site for SGK1 and to be necessary for dexamethasone to stimulate NHE3 over short times, not involving transcriptional regulation (39,40). Because our studies showed that cGMP/cGKII phosphorylation of NHE3 at this site inhibited NHE3 activity, we initially confirmed that dexamethasone stimulates NHE3 activity under basal conditions through phosphorylation of Ser 663 and increases NHE3 surface expression (Fig. 5, A and B). Further evidence of the role of NHE3-Ser 663 phosphorylation in dexamethasone stimulation of NHE3 was provided by studies of S663A and S663D mutants. Dexamethasone did not induce a change of activity in NHE3-S663A or NHE3-S663D (Fig. 5A) or increase the surface expression of these mutants (Fig. 5B).     The NHE3 Inhibition by cGMP/cGKII Involves Phosphorylation of NHE3 at Ser 554 and Ser 607 in Mouse Ileum-To confirm that inhibition of NHE3 activity by cGMP/cGKII involves phosphorylating NHE3 at Ser 554 , and Ser 607 , we used polyclonal specific anti-P-Ser 554 and anti-P-Ser 607 to examine the phosphorylation level of NHE3 in mouse ileum. Western blots were prepared from mouse ileum total membrane. In mouse ileum, NHE3 was phosphorylated at Ser 554 and Ser 607 by both FSK and cGMP (Fig. 6).

DISCUSSION
This study identified three amino acids of NHE3, which were phosphorylated by cGMP/cGKII, an effect that inhibited NHE3 activity by decreasing the surface expression but not total expression over the short time (up to 1 h) of these studies. These results are consistent with an effect via trafficking, and established the requirement for all three phosphorylation sites at the same time in NHE3 for inhibition by cGKII. These results provide evidence of a unique form of regulation by phosphorylation of NHE3, because we identified a single multifunctional site that, when phosphorylated, causes both stimulation and inhibition of NHE3. In addition, the difference in magnitude of the cGMP/cGKII effect on NHE3 activity, which exceeds the effect on surface expression, suggests an effect on the turnover number as well as an effect on trafficking. These mechanisms of the cGKII/cGMP effects are very similar to the effects of cAMP elevation on NHE3 regulation, as reported in the past (1).
Previous studies of cGKII have identified a number of its phosphorylated substrates, all of which have also been phosphorylated by PKA (9,37,38). Serulle et al. (9) showed that cGKII phosphorylates GluR1 at Ser 845 in the brain, a site also phosphorylated by PKA. By quantitative mass spectrometry and further truncation studies, we identified Ser 663 as a novel cGKII phosphorylation site on NHE3 and showed by functional studies that it is not involved in NHE3 regulation by PKA. By using specific anti-phosphoantibodies, we showed that cGKII also increased the NHE3 phosphorylation at Ser 554 and Ser 607 . The inhibition of NHE3 activity by cGKII was prevented by the single mutation of each of these three sites from Ser to Ala and expression in a model in which WT NHE3 was inhibited by cGMP/cGKII. However, the Ser to Asp single mutation of Ser 663 , Ser 554 , and Ser 607 still preserved the cGKII inhibition of NHE3, which we suggest is due to the other two sites still being able to be phosphorylated by cGMP/cGKII, such that all three sites act as if they are phosphorylated. Triple mutation studies (Ser to Asp) confirmed that all three sites must be able to be phosphorylated at the same time (or show increases in phosphorylation) for NHE3 to be inhibited by cGMP/cGKII. These results indicate that all three sites, Ser 554 , Ser 607 , and Ser 663 , are . The activity of NHE3 in S554A, S607A, and S663A single mutants was not inhibited by 8-pCPT-cGMP/cGKII. A, NHE3 activity (V max , M/s) was determined in adenovirus cGKII-infected PS120/NHERF2/HA-NHE3 WT, and NHE3-S554A, -S607A, -S663A, and -S663D single mutant cells. NHE3 activity was decreased in wild type and S663D cells pretreated with 8-pCPT-cGMP for 45 min (p Ͻ 0.05), but not changed significantly by cGMP in all Ser to Ala single mutant cell lines (mean Ϯ S.E., n ϭ 3 except S663D in which n ϭ 5). Also noted is that basal NHE3 activity was significantly increased in S663A mutant cells and significantly reduced in S554A and S607A mutant cell lines (mean Ϯ S.E., n ϭ 3) (p Ͻ 0.05). B, NHE3 basal activity was decreased in S554D and S607D mutant cell lines compared with WT cells, and NHE3 activity was significantly further decreased in both cell lines by 8-pCPT-cGMP. C, NHE3 basal activity was decreased in the S554D,S607D,S663D triple mutant cell lines compared with WT cells, and triple mutant cells lost inhibition of NHE3 activity by 8-pCPT-cGMP. D-I, total and cell surface NHE3 expression in WT and mutant cell lines. There were no changes of total NHE3 in NHE3-S554A or -S554D, NHE3-S607A or -S607D, and NHE3-S663A or -S663D mutants compared with wild type cells. Surface NHE3 significantly decreased in NHE3-S554A and -S554D and NHE3-S607A and -S607D cell lines was significantly increased in S663A cells (mean Ϯ S.E., n ϭ 3) (p Ͻ 0.05), and not significantly changed in S663D cells. Surface NHE3 in NHE3-S554D and NHE3-S607D cells was significantly reduced by 8-pCPT-cGMP, but not in NHE3-S554A, NHE3-S607A, and NHE3-S663A and -S663D cells. In each experiment, surface expression of WT NHE3 was set to 100%. J, NHE3 activity (V max , M/s) was inhibited by FSK in PS120/NHERF2/HA-NHE3-S663A single mutant cells similarly to WT cells (mean Ϯ S.E., n ϭ 3) (p Ͻ 0.05). . Dexamethasone failed to stimulate NHE3 activity or increase the surface NHE3 amount when NHE3 was mutated to S663A or S663D. A, transport activity (initial rates) was measured in adenovirus-cGKII-infected PS120/NHERF2/HA-NHE3 WT and S663A and S663D single mutant cells. Dexamethasone increased the activity of NHE3 in control cells (p Ͻ 0.05), but failed to increase NHE3 activity in either S663A or S663D cells (mean Ϯ S.E., n ϭ 3). B, the surface NHE3 amount was not changed by dexamethasone treatment for 4 h in S663A and S663D mutant cells (mean Ϯ S.E., n ϭ 3). PS120 cells were grown in 10-cm dishes, and cells were infected with adeno-cGKII virus. 48 h after infection, cells were serum starved for 3 h and then treated with 2 M dexamethasone for 4 h at 37°C before biotinylation. Surface expression of NHE3 was determined as a percent of total NHE3.

cGKII Phosphorylates NHE3
phosphorylated by cGKII and must be phosphorylated for cGKII to inhibit NHE3 activity.
The requirement for multiple sites of NHE3 or its homologues to be phosphorylated at the same time has been demonstrated for two other kinases/kinase cascades: Mos-MEK-MAPK-p90RSK and CaMKII. In starfish NHE3, the Mos-MEK-MAPK-p90RSK pathway stimulates NHE3 via Rsk phosphorylation of NHE3 at three distal C-terminal sites, all of which are necessary for stimulation of NHE3 activity (41). Also in mammalian NHE3, CaMKII inhibits basal NHE3 activity via phosphorylation of three sites, all of which must be present for this effect of CaMKII (42). In these examples, the phosphorylated sites are clustered relatively close to each other based on linear modeling of the NHE3 C terminus (within 84, 110, and 118 amino acids) (Fig. 7, A-C). We suggest that the total negative charges might alter the environment of the intracellular domain of NHE3, including by affecting protein-protein interactions that have been shown to form on the NHE3 C terminus and to be involved in NHE3 regulation (43). We speculate that this might indicate a more complicated regulation of NHE3 by kinases that involves changes in signaling complexes affected by NHE3 phosphorylation and/or interaction with components of the trafficking machinery. In this regard, CaMKII inhibition of NHE3 does not appear to involve changes in NHE3 surface expression (42), favoring a function of the multiple phosphorylations via NHE3 complex formation. Other recent studies have identified clustering of sites phosphorylated by single kinases and provided evidence that clustered phosphorylation sites tend to be activated by the same kinase and may act to increase the robustness of the phosphorylation-dependent response (36).
Ser 663 was previously reported to be phosphorylated by SGK1 in response to the short exposure to dexamethasone (ϳ4 h) to stimulate NHE3 by an effect that did not involve changes in transcriptional stimulation of NHE3 (20). Our studies confirmed the role of Ser 663 in dexamethasone stimulation of NHE3 activity over short intervals, but also showed that Ser 663 was phosphorylated by cGKII to inhibit NHE3. These results suggest that phosphorylation of Ser 663 involves a unique bifunctional regulation in which separate pools are affected, brush-border NHE3 is inhibited by endocytosis, and intracellular NHE3 is stimulated by exocytosis. The concept that a single site when phosphorylated alone stimulates and when phosphorylated along with several additional sites inhibits transport function has not been described previously. We suggest that there are separate changes in the C terminus of NHE3 when this site is phosphorylated alone versus with other amino acids, which accounts for the divergent regulation, although how that occurs is unknown.
The studies correlating the effects of NHE3 Ser to Ala and Ser to Asp mutants (NHE3-S554D,S607D,S663D) on NHE3 activity with surface expression under basal and after cGMP/cGKII exposure are summarized in Table 1. Mutating any of the three amino acids Ser to Asp eliminated cGMP/cGKII inhibition of NHE3 activity and the reduction of surface NHE3. The Ser to Asp mutations reconstituted cGMP inhibition of NHE3 activity. Of great interest was that only the S663D mutation failed to duplicate the reduction in surface NHE3 by cGMP/cGKII. We suggest that this finding indicates that multiple phosphorylation sites interact in a complex manner, perhaps involving differences in timing of the effect or with different kinetics.
In addition to the changes in cGMP and dexamethasone regulation of NHE3 related to phosphorylation of Ser 663 , mutation  cGKII Phosphorylates NHE3 of this amino acid to Ala stimulated basal NHE3 activity without changing total NHE3 expression. The effect of the S663A mutation to increase both NHE3 activity and surface expression is consistent with removing the inhibitory effect of a kinase, which acts to inhibit NHE3 under basal conditions. Please note that our studies did not emphasize identification of the kinase(s), which regulates basal NHE3 activity and we did not further explore the mechanisms of these changes in basal NHE3 activity. Thus we have described a new phenomenon in which phosphorylation of the same amino acid in a single protein can stimulate or inhibit function of the protein based on whether there is simultaneous phosphorylation of additional sites. Understanding in more detail how this occurs is likely to increase understanding of NHE3 regulation by signaling complexes, which form on its C terminus to dynamically regulate its activity (43).