Acute Regulation of Na/H Exchanger NHE3 by Adenosine A1 Receptors Is Mediated by Calcineurin Homologous Protein*

Adenosine is an autacoid that regulates renal Na+ transport. Activation of adenosine A1 receptor (A1R) by N6-cyclopentidyladenosine (CPA) inhibits the Na+/H+ exchanger 3 (NHE3) via phospholipase C/Ca2+/protein kinase C (PKC) signaling pathway. Mutation of PKC phosphorylation sites on NHE3 does not affected regulation of NHE3 by CPA, but amino acid residues 462 and 552 are essential for A1R-dependent control of NHE3 activity. One binding partner of the NHE family is calcineurin homologous protein (CHP). We tested the role of NHE3-CHP interaction in mediating CPA-induced inhibition of NHE3 in opossum kidney (OK) and Xenopus laevis uroepithelial (A6) cells. Both native and transfected NHE3 and CHP are present in the same immuno-complex by co-immunoprecipitation. CPA (10-6 M) increases CHP-NHE3 interaction by 30 - 60% (native and transfected proteins). Direct CHP-NHE3 interaction is evident by yeast two-hybrid assay (bait, NHE3C terminus; prey, CHP); the minimal interacting region is localized to the juxtamembrane region of NHE3C terminus (amino acids 462-552 of opossum NHE3). The yeast data were confirmed in OK cells where truncated NHE3 (NHE3Δ552) still shows CPA-stimulated CHP interaction. Overexpression of the polypeptide from the CHP binding region (NHE3462-552) interferes with the ability of CPA to inhibit NHE3 activity and to increase CHPNHE3Full-length interaction. Reduction of native CHP expression by small interference RNA abolishes the ability of CPA to inhibit NHE3 activity. We conclude that CHPNHE3 interaction is regulated by A1R activation and this interaction is a necessary and integral part of the signaling pathway between adenosine and NHE3.

Adenosine is an ubiquitous nucleoside generated from AMP by 5Ј-nucleotidase or from S-adenosyl-L-homocysteine (SAH) 1 by SAH hydrolase (1,2). Renal adenosine is derived exogenously from the systemic circulation through mainly neuromuscular sources or endogenously via the intrarenal AMP and SAH pathways (2). Adenosine modulates many functions in mammals via binding to A 1 , A 2A , A 2B , and/or A 3 adenosine receptors (1,3). In the kidney, adenosine regulates glomerular filtration rate, renin release, erythropoietin production, cellular proliferation, and tubular water and Na ϩ transport (2,4).
Renal adenosine generation is markedly increased in response to hypoxia, ischemia, or inflammation (2,3). It is well documented that adenosine can exert its protective effect against acute renal ischemia by containing inflammatory damage inflicted by circulating immune cells (2,3). A concomitant protective effect during ischemia can be achieved by reducing oxygen expenditure of the renal tubules through inhibition of Na ϩ transport, the principal oxygen-consuming process (5,6). In the renal proximal tubules, which constitute over 70% of renal cortical mass, adenosine decreases the activity of the Na ϩ /H ϩ exchanger NHE3 on the apical membrane of the proximal tubule (7)(8)(9). The inhibition of apical Na ϩ entry decreases the activity of the Na ϩ /K ϩ -ATPase and oxygen requirement of the highly aerobic proximal tubule (10,11).
NHE3 belongs to a superfamily of electroneutral mammalian Na ϩ /H ϩ exchangers that has eight members documented to date (12)(13)(14)(15)(16)(17). With the exception of NHE5, all NHE isoforms are found in the kidney. All NHEs have a predicted N-terminal hydrophobic ion-translocating domain and a variable C-terminal hydrophilic domain that harbors regulatory sequences (15)(16)(17)(18). NHE3 is the isoform in the apical membrane based on antigenic (19,20) and functional data (21)(22)(23)(24). Non-NHE3mediated Na ϩ /H ϩ exchange activity in the proximal tubule apical membrane (24) can theoretically be caused by NHE8 (14), but definitive data are still forthcoming. The current study focuses on NHE3. Using a heterologous system, we found that acute A 1 R and A 2 R stimulation inhibits NHE3 via primarily PKC-coupled and PKA-coupled pathways, respectively (7). In a recent report (8), we showed that early A 2 R activation is associated with a change in intrinsic transport activity of NHE3, whereas sus-tained inhibition of NHE3 was accompanied by a reduction of surface NHE3 protein. In addition, we showed that regulation of NHE3 activity by A 2 R activation is dependent on phosphorylation of NHE3 at two primary PKA target sites (8). Although there are some parallels between PKA-and PKC-dependent control of NHE3 activity, it is not known whether phosphorylation supports the A 1 R ability to regulate NHE3.
Recently, we presented evidence that inactivation of NHE3 activity in response to A 1 R stimulation depends on elevation of intracellular Ca 2ϩ (9). Calcineurin homologous protein (CHP) is a recently described widely expressed Ca 2ϩ -binding EF-hand protein of ϳ22 kDa (25,26) with significant similarity to the regulatory B subunit of the heterodimeric protein phosphatase, calcineurin (26). CHP inhibits calcineurin phosphate activity (27) and interacts with NHEs (26,28,29). CHP attenuates the stimulation of NHE1 by serum and a mutationally activated GTPase in CCL39 cells (26). In PS120 fibroblasts, CHP binding to NHE1 does not affect surface NHE1 abundance but was necessary for surface NHE1 to function (28). The ability of CHP to alter NHE1 transport independent of surface NHE1 protein (28) renders it an attractive candidate for supporting adenosine-mediated change in intrinsic transport activity of NHE3. CHP associates with microtubules via an N-myristoylation-dependent mechanism that does not involve conventional microtubule-associated proteins (30) and is involved in vesicular transport (25). CHP also interacts with kinesin-related protein KIF1B␤2, a molecular motor in neurons (31). These properties of CHP make it a plausible mediator in changes in trafficking between surface and intracellular NHE3 proteins. Although CHP clearly binds NHE3 (28), the functional effect of CHP-NHE3 interaction has not been studied in the context of hormonal regulation of Na ϩ and H ϩ transport. In this manuscript, we described the role of CHP in mediating the regulation of NHE3 by adenosine.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Experiments were performed on OK and A6/C1 cells. A subclone of OK cells selected on the basis of expression of parathyroid hormone-sensitive Na ϩ /phosphate co-transport (32) was cultured in a mixture of Dulbecco's modified Eagle's medium Ham's F-12 (Invitrogen, Carlsbad CA) supplemented with 10% fetal calf serum, 2 mM glutamine, and 50 IU ml Ϫ1 penicillin and 50 g ml Ϫ1 streptomycin in a humidified 95%/5% air/CO 2 atmosphere at 37°C and sub-cultured weekly. Cells generally reached confluence within 3-4 days, and experiments were conducted 1-2 days after confluence. Studies on OK cells were performed between passages 23 and 57. A6/C1 cells are a subclone of A6-2F3 cells, functionally selected on the basis of high transepithelial resistance and responsiveness to aldosterone (33). These cells express endogenous basolateral Na ϩ /H ϩ exchange activity (7,34) and transfected NHE3 is targeted to the apical membrane (7,8). A6/C1 cell lines were grown in 0.8ϫ concentrated Dulbecco's modified Eagle's medium, supplemented with 25 mM NaHCO 3 , 10% heat-inactivated fetal bovine serum, 50 IU ml Ϫ1 penicillin and 50 g ml Ϫ1 streptomycin in a humidified 95%/5% air/CO 2 atmosphere at 28°C. For A6 cells transfected with NHE3, medium was supplemented with 450 g ml Ϫ1 FIG. 1. Effect of the A 1 R agonist CPA on native and transfected NHE activities in OK cells and A6 cells. NHE activity was assayed microfluorometrically as the rate of Na ϩ -dependent intracellular pH (pH i ) recovery. CPA (10 Ϫ6 M) was applied to the apical cell surface for 15 or 30 min as indicated. In OK cells stimulation of A 1 R was from the apical or basolateral site. In A6 cells where A 1 R are located in the apical membrane, CPA was added only to the apical cell surface. Data are expressed as percentage of activity of untreated cells. The number of experiments, each consisting of a pair of control and agonist-treated cells, is given in parentheses. Bars and error bars represent means and standard errors, respectively. Asterisks indicate statistical significance from control measurements (*, p Ͻ 0.05; **, p Ͻ 0.01, ANOVA). A, effect of CPA on apical NHE3 activity in opossum kidney (OK) cells. B, effect of CPA on opossum NHE3 stably transfected into A6 cells. C, effect of CPA on rat NHE3 stably transfected into A6 cells. D, effect of CPA on endogenous basolateral Xenopus NHE (XNHE) in untransfected (A6/ C1) and transfected A6 cell lines. hygromycin B. Cells were subcultured weekly by trypsinization (Ca 2ϩ / Mg 2ϩ -free salt solution containing 0.25% (w/v) trypsin and 2 mM EGTA). Cells generally reached confluence between 7 and 8 days after seeding, and studies were performed between passages 6 and 43.
For transient transfection, cells were grown to ϳ70% confluence in culture dishes, and 1.3 g of cDNA was introduced into cells using LipofectAMINE (Invitrogen). Transfection efficiency was monitored by co-transfection of the pEGFP plasmid (Clontech, Palo Alto, CA) and was approximately 70%. For stable transfection, A6 cells were grown to 20 -25% confluence in 35-mm tissue culture dishes, and cDNA was introduced into cells using FuGENE 6 TM (Roche Applied Science, Mannheim, Germany). The construct of interest (1.5 g) was co-transfected with the p3ЈSS⌬LacI vector (0.5 g) to allow selection by hygromycin B resistance (7). Clonal populations of transfected cell lines obtained by ring cloning were maintained in hygromycin. Hygromycinresistant clones were further selected for NHE3 transcripts by reverse transcription-PCR (data not shown) as previously described (8). Cells used for experiments were treated for 6 -7 days with 1 M dexamethasone known to accelerate maturation and differentiation (35). To ensure true segregation of the apical and basolateral surface, we established that A6/C1 cell lines exhibited reproducible high transepithelial resistances. Untransfected A6/C1 cells transepithelial resistance was comparable with NHE3-transfected ones (data not shown).
Yeast Two-hybrid-Bait vectors were transformed into Saccharomyces cerevisiae (strain L40) containing the genotype MATa trp1 leu2 his3 LYS2::lexA-HIS3 URA3::LexA-LacZ (38,41). Yeast cells were grown in YPDA medium (1% yeast extract, 2% Difcopeptone, 2% glucose, 0.003% adenine hemisulfate) or the synthetic minimal Trp dropout medium (SD-Trp) complemented with 0.003% adenine hemisulfate (42). For transformation, a YPDA culture from L40 (20 ml) grown to an A 600 of FIG. 3. Effect of CPA on apical membrane NHE3 protein abundance in A6 cells stably transfected with opossum NHE3. CPA (10 Ϫ6 M for 15 or 30 min, apical addition) or vehicle-treated cells on permeable support were labeled with biotin on apical cell surface, and biotinylated proteins were precipitated with streptavidin-bound agarose. The NHE3 abundance was measured by immunoblot using anti-OK-NHE3 (#5683) antisera. Total cellular NHE3 protein abundance was measured on cell lysate by immunoblot. A, representative experiment. B, summary of results of effect of CPA on biotin-accessible NHE3 surface antigen; the number of experiments is given in parentheses, and significance from control measurements is indicated by asterisks (*, p Ͻ 0.05, ANOVA). C, A6 cells were transiently transfected with the NHE3/eGFP construct. Forty-eight h after transfection, cells were visualized by fluorescent microscopy prior to and after exposure to CPA for 15 or 30 min.  Gietz and Schiestl (43). After an incubation at 30°C (30 min at 260 rpm) followed by a heat shock at 42°C (20 min), cells were harvested at 16,500 rpm for 15 s in a tabletop centrifuge, resuspended in 100 l of distilled H 2 O, and plated onto 14-cm SD-Trp Petri dishes.
A cDNA library (MATCHMAKER, Clontech) of whole adult mouse kidneys was screened as previously described (44) except mNHE3 cyto was used as bait. cDNA inserts were ligated to the C terminus of the GAL4 activation domain in vector pACT2, which delivers the leu2 nutritional gene for complementation. Putative positive colonies were isolated by consecutive colony lift assays where the permeabilized cells transferred onto Whatman filters were overlaid with 0.2 mg/ml 5-bro-mo-4-chloro-3-indolyl ␤-D-galactopyranoside (Alexis, Lausen, Switzerland), 150 mM NaCl, 50 mM Tris-HCl (pH 7.4), and 0.8% agarose (45,46). Yeast DNA of true positives was isolated (47), and prey plasmids were rescued by transformation into KC8 cells, which carry trpC, leuB, and hisB mutations (48). Positivity of purified preys was validated once again in yeast. Insertion sizes were checked by BglII digestion. After direct sequencing, identical sequences were grouped via ClustalW at Pôle Bio-Informatique (Lyonnais) or via Pileup from Genetics Computer Group (Oxford) and overlaps connected by the Contig assembly program (Baylor College of Medicine, Houston, TX). Searches for protein relationships were performed using BLAST (NCBI, Bethesda, MD) (49). A number of positive preys were identified as either partial or fulllength CHP. Truncated NHE3 baits were constructed as described above and tested against CHP. Briefly, overnight cultures of yeast previously transformed with selected bait vectors were washed and transformed using a reaction mixture of small-scale yeast transformation containing 10 g of cDNA from each prey as described. After 5 days on SD-Trp/Leu/His plates, prototrophies were tested for LacZ expression by a colony-lift assay (45,46).
For total protein extract, cells were disrupted by sonication and harvested at 16,500 rpm for 10 min in a tabletop centrifuge. 60 g of total lysate, fractionated by SDS-PAGE, was immunoblotted by anti-CHP antibody, anti-NHE3 antibody (#5683), or anti-Na/K ATPase antibody (␣2F) (Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA).
Confocal Fluorescent Imaging-OK cells were plated on glass cover-slips, and A6 cells were plated onto permeant filter supports (4.7 cm 2 , Transwell; Costar, Cambridge, MA) previously coated with a thin layer of rat collagen. Transiently transfected cells (48 h later) were fixed (4% formalin in PBS for 10 min), permeabilized (0.1% Triton X-100 in PBS for 10 min), blocked by 5% bovine serum albumin in PBS at 37°C for 30 min. Specimens were incubated with rabbit anti-HA polyclonal antibodies (1:200 dilution, Sigma) and mouse anti-FLAG monoclonal antibodies (1:100 dilution; Sigma) at 37°C for 1 h, followed by Rhodamine Red-X-conjugated anti-rabbit and Cy5-conjugated anti-mouse (all secondary polyclonal antibodies from donkey, Jackson ImmunoResearch Laboratories), respectively. Confocal fluorescence images were visualized through a Zeiss 100ϫ objective lens using a Zeiss LSM-410 laserscanning confocal microscope. Fluorescence of Rhodamine Red, GFP, and Cy5 were detected using excitation laser (wavelength in nanometers) of 568, 488, and 633 and emission filter of 590 long-pass filter, 510 -560 band-pass filter, and 670 -810 band-pass filter, respectively. In Fig. 10B (see below), the image of FLAG-tagged NHE3 462-552 was assigned pseudo-colors blue or green.
For imaging of NHE3 in live cells, OK and A6 cells on glass coverslips were transfected with NHE3/eGFP and maintained for 48 h. Coverslips were placed on the stage of an inverted confocal fluorescence microscope, and living green fluorescent cells were selected before the treatment. CPA or vehicle was then added to the cell medium. After the stated period of incubation, images were acquired and the pattern of expression of transporters was compared with that before treatment.
Selection and Preparation of siRNA Duplexes from the Target CHP Sequence-Target region was selected from cDNA sequence of endogenous opossum CHP. By homology cloning we identified 300 nucleotides downstream of the CHP start codon (GenBank TM accession number: AY281349). The AA-N19 mRNA target was synthesized by Dharmacon Research Co. (Lafayette, CO), and sequences were as follows: 5Ј-agagcaaagaugugaaugg-3Ј (sense strand) and 5-ccauucacaucuuugcucu-3Ј (antisense strand). GL2 (firefly luciferase gene) siRNA was synthesized according to Elbashir et al. (50). The 21-nucleotide siRNA duplex targeting OK mRNA sequence of the protein tyrosine kinase, Pyk2 (5Јaaugcccuugauaagaagucc-3Ј) was kindly provided by Dr. Patricia Preisig (University of Texas Southwestern Medical Center, Dallas, TX). Oligonucleotides were provided at the concentration of 20 M, purified, annealed duplex, and were ready to use for transfection. Approximately 24 h prior to transfection OK cells were plated at an appropriate cell density so that they were ϳ70% confluent the following day. For the complex formation, we used 7 l of the TransIT-TKO transfection reagent (Mirus Corp., Madison, WI) in 100 l of serum-free medium. Diluted TransIT-TKO reagent was incubated at room temperature for 20 min. Then siRNA (70 nM) was added to the complex and incubated at room temperature for an additional 20 min. The TransIT-TKO reagent-siRNA complex mixture was given to OK cells, and cells were incubated with the mixture for 48 h. Assays for silencing were performed after 48 h of serum deprivation.
Statistical Analyses-Results are represented as means Ϯ S.E. Quantitative differences between control and test conditions were assessed statistically by analysis of variance. Probability (p) of Ͻ0.05 was considered statistically significant.

Activation of A 1 R Inhibits NHE3 Activity-We previously
showed that acute activation of A 1 R decreases NHE3 activity in renal epithelial cells (7,9). This study uses OK cells expressing endogenous NHE3 and A6 cells expressing transfected NHE3 to analyze the mechanism underlying A 1 R-induced inactivation of NHE3. As shown in Fig. 1A, activation of A 1 R in OK cells with 10 Ϫ6 M CPA (N 6 -cyclopentyladenosine) reduced the activity of the native apical NHE3 by 31 and 35% at 15 and 30 min, respectively. As shown in Fig. 1 (B and C), CPA had similar effects on opossum or rat NHE3 stably transfected into A6 cells. A 1 R activation in the apical membrane by 10 Ϫ6 M CPA for 15 or 30 min decreased the activity of transfected OK-NHE3 by 27 and 40%, respectively (Fig. 1B), and reduced the activity of transfected rat-NHE3 by 31 and 39%, respectively (Fig. 1C). In both transfected and untransfected A6 cells, exposure to CPA (10 Ϫ6 M for 15 min) inhibited the native basolateral Xenopus NHE (XNHE) activity (Fig. 1D).
CPA Decreases Surface NHE3 Protein-One mechanism that has been shown to acutely alter NHE3 activity is the change in FIG. 6. C-terminal murine NHE3 baits used in two-hybrid system screens. Various constructs derived form the C terminus of murine NHE3 were used as baits against full-length CHP as prey. Strength of interaction was determined semi-quantitatively by ␤-galactosidase assay and is indicated by "ϩ" (ϩϩϩ, strong; ϩϩ, moderate; 0, no binding). cell surface NHE3 protein abundance (8,40,(51)(52)(53)(54). We used surface biotinylation to determine whether the decrease in NHE3 activity in response to activated A 1 R is associated with reduction in surface NHE3 antigen. As shown in Fig. 2, there was no detectable change in surface or total native OK-NHE3 antigen after 15 min of 10 Ϫ6 M CPA. However, 30 min of 10 Ϫ6 M CPA induced a 35% reduction in surface NHE3. Similar results were obtained from studying A6 cells expressing transfected OK-NHE3 protein (Fig. 3, A and B). In both cell lines, CPA did not affect the amount of total cellular NHE3 protein (see Fig. 2A and Fig. 3A). In addition to biochemical biotin accessibility assays, we also studied CPA-induced changes of NHE3 distribution in live cells transiently expressing OK-NHE3 tagged with enhanced green fluorescent protein (OK-NHE3/eGFP). Fluorescent microscopy showed that CPA caused a visible decrease in surface NHE3 at 30 min of exposure to 10 Ϫ6 M CPA (Fig. 2C). Comparable findings were obtained in A6 cells transiently expressing OK-NHE3/eGFP (Fig. 3C). Exposure to vehicle did not cause any change in NHE3/eGFP distribution in either cell lines (not shown).
Identification of Structural Elements Required for Downregulation of NHE3 by CPA-According to the preceding results, inhibition of NHE3 activity by CPA at 15 min is exclusively due to a reduction of intrinsic NHE3 activity. We recently have shown that A 1 R-mediated inactivation of NHE3 is dependent on protein kinase C (PKC) (7,9). Phosphorylation of NHE3 is an early event in regulation of NHE3 by protein kinases (36,37,55,56). Elimination of phosphoserines on NHE3 may disrupt CPA regulation of NHE3 activity. To test this hypothesis, we assayed A6 cells stably expressing rat-NHE3 bearing either single or multiple serine substitutions for their response to CPA. As shown in Fig. 4A, substitution of serine residues at a single position (S605G) or multiple positions (S513G, S552A, S575A, S634A, S661A, and S690G) on NHE3 did not appreciably affect NHE3 inactivation by 10 Ϫ6 M CPA. Similar results were obtained in studies using the PKC agonist PMA (phorbol 12-myristate 13-acetate). This is in contradistinction to the previous data in fibroblasts where phosphoserines were necessary but not sufficient for alteration of NHE3 activity by PMA (37). Co-exposure to the PKC antagonist calphostin C (10 Ϫ8 M for 15 min) completely reversed NHE3 inhibition induced by either CPA or PMA (data not shown). CPA or PMA reduced the activity of basolateral XNHE in transfected cell lines to an extent that is similar to the effect of either of the agonist on XNHE in untransfected A6 cells (Fig.  4B). These findings suggest that direct phosphorylation of NHE3 is not necessary for A 1 R-induced down-regulation of NHE3.
In the following series of experiments we tested the functional behavior of truncated versions of OK-NHE3 to identify the region in the NHE3 C terminus that is essential for the A 1 R response. Fig. 5A gives an overview on the truncated NHE3s screened for the ability of A 1 R activation to regulate NHE3. Compared with wild-type NHE3, mutants lacking the carboxyl tail distal to position 462 (NHE3 ⌬462 ) were completely unresponsive to 15-min treatment with 10 Ϫ6 M CPA, whereas NHE3 truncated at position 552 (NHE3 ⌬552 ) retained full response to CPA (Fig. 5B). Similarly, NHE3 ⌬462 lost PMA responsiveness, whereas NHE3 ⌬552 and NHE3 ⌬640 were inhibited by 10 Ϫ7 M PMA (Fig. 5B). Co-exposure to calphostin C (10 Ϫ8 M for 15 min) prevented the response of OK-NHE3 ⌬552 to either CPA or PMA (data not shown). The endogenous XNHE responded to either CPA or PMA in A6 cell lines expressing wild-type or truncated versions of OK-NHE3 (Fig. 5C). Various truncations of NHE3 per se did not significantly affect whole cell NHE3 protein expression (data not shown). We conclude that A 1 R-induced inhibition of NHE3 requires regulatory elements within residues 462-552 of the C-terminal domain of NHE3.
Identification of CHP as a CPA-regulated NHE3-binding Protein-By analogy to the molecular mechanism allowing inactivation of NHE3 by PKA, elevated Ca 2ϩ , or PKC (53, 57-60), we hypothesized that PKC-mediated inactivation of NHE3 by activated A 1 R involves interaction of NHE3 with a regulatory protein. In a yeast-two-hybrid screen of an adult mouse kidney cDNA library using full-length murine-NHE3 C terminus as bait, full-length murine CHP (calcineurin B homologous protein) was repeatedly positive in interacting clones. CHP has previously been reported to bind to NHE3 near the vicinity of the region of aa 462-552 essential for A 1 R-induced regulation (28). We next examined whether CHP interacts with the subdomain of NHE3 that is critical for A 1 R-dependent control of NHE3 activity. To this end, we used truncated versions of the C-terminal domain of murine NHE3 as bait in the yeast twohybrid assay. As shown in Fig. 6, CHP specifically interacted with murine NHE3 aa 381-557, which is homologous to the CPA-sensitive domain on OK-NHE3 (aa 462-552). CHP interacted with a minimal version of the juxtamembrane C-terminal cytoplasmic domain of NHE3 comprising aa (murine) 338 -501 (Fig. 6). In contrast, baits devoid of this region of the NHE3 C terminus showed no interaction with CHP (Fig. 6).
To test for association of CHP with full-length NHE3 in vertebrate cells and to determine whether CPA modulates the association, we performed co-immunoprecipitation studies. Immunoprecipitation of native NHE3 (Fig. 7A) brought down native CHP in the immune complex (Fig. 7A). Importantly, exposure to CPA increased the amount of NHE3 bound to native CHP by ϳ30% (Fig. 7, A and D). Additional evidence for an association of CHP and NHE3 was obtained in studies using OK transiently expressing His 6 -tagged rat-NHE3 and HAtagged murine-CHP. As shown in Fig. 7B, anti-HA antibody co-precipitated His 6 -tagged rat-NHE3 only in double-transfected OK cells but not in OK cells that were not transfected (Fig. 7B) or solely transfected with His 6 -tagged rat-NHE3 (data not shown). Again, exposure of double-transfected OK cells to CPA increased the amount of NHE3 bound to CHP by 60% (Fig.  7, B and D).
If the critical interacting region on NHE3 is indeed aa 462-552, the association between CHP and NHE3 should be preserved with truncation of OK-NHE3 at aa 552 and disrupted by truncation of OK-NHE3 at aa 462. Co-transfection of myctagged OK-NHE3 ⌬552 and HA-tagged CHP in OK cells showed intact co-immunoprecipitation as well as increased NHE3-CHP association in response of CPA (Fig. 7C), whereas co-expression of myc-tagged OK-NHE3 ⌬462 and HA-tagged CHP in OK cells demonstrated no association of NHE3 with CHP (Fig. 7D). Fig.  7E summarized the CPA-mediated increases of association of CHP with NHE3 (native and transfected proteins).
Subcellular Localization of CHP and NHE3 in OK Cells and A6 Cells-In the next series of experiments we compared subcellular localization of NHE3 and CHP in OK cells and A6 cells transiently expressing HA-tagged CHP (HA/CHP) and NHE3/ eGFP by confocal fluorescent microscopy. HA-tagged CHP showed primarily intracellular distribution (Fig. 8A, upper two panels, xz section), whereas NHE3/eGFP localizes to both intracellular compartment as well as the apical membrane (Fig.  8A, bottom two panels, xz section). However, there is definite co-localization of HA/CHP and NHE3/eGFP in the apical membrane in both OK and A6 cells (Fig. 8, B and C, yellow signal of the merged composite).
Functional Importance of CHP-NHE3 Interaction for CPAdependent Control of NHE3 Activity-We used two approaches to define the functional significance of the CHP-NHE3 interaction in A 1 R-dependent control of NHE3 activity. First we attempted to create a "dominant negative" NHE3 polypeptide. We transiently transfected a construct expressing aa 462-552 of OK-NHE3 (NHE3 462-552 ) in OK cells to attempt to stoichiometrically bind endogenous CHP and competitively inhibit the endogenous CHP-NHE3 interaction. To validate this approach, we first established that transfected NHE3 462-552 actually competes with wild-type full-length NHE3 for CHP binding. To this end, OK cells were triple-transfected with HA-tagged CHP (HA/CHP), His 6 -tagged NHE3 (NHE3/6H), and FLAG-tagged NHE3 462-552 (FLAG-NHE3 462-552 ). Lysate from vehicle or CPA-treated cells were subjected to immunoprecipitation by anti-His 6 or anti-FLAG antibodies and immunoblotted with anti-HA antibody. As illustrated in Fig. 9, both NHE3 462-552 and full-length NHE3 interacts with CHP. Interestingly, in the presence of NHE3 462-552 , CPA failed to increase binding of full-length NHE3 to CHP (Fig. 9). To further substantiate that NHE3 and NHE3 462-552 compete for CHP binding, we examined OK cells transiently expressing HA/CHP, NHE3/eGFP, and FLAG/NHE3 462-552 by immunocytochemistry. Fig. 10A highlights expression of the three heterologous recombinant proteins in OK cells. FLAG/NHE3 462-552 is diffusely distributed in OK cells as compared with NHE3/eGFP, which has a visibly apical distribution. Fig. 10B highlights the co-localization of either NHE3 or NHE3 462-552 with CHP (yellow signal of the merged composite). Localization of CHP and NHE3 was comparable to results shown in Fig. 8 where apical targeting of both proteins is evident (Fig. 10B, left panel). These findings support the notion that expression of NHE3 462-552 competes for binding of NHE3 to CHP. Finally, we measured NHE3 activity in OK cells transiently expressing the NHE3 462-552 peptide. As shown in Fig. 11, expression of NHE3 462-552 does not influence baseline NHE3 activity but abolished the CPA-induced inhibition of NHE3 activity.
The second approach we used to demonstrate the functional significance of CHP for A 1 R-induced regulation of NHE3 was to reduce the expression level of endogenous CHP. To this end we cloned a partial sequence of OK CHP to enable construction of AA(N19)TT duplex RNA sequence (siRNA), which upon transfection in OK cells acts to silence expression of OK-CHP. The opossum CHP is aligned against the murine, rat, and human homologue in Fig. 12A. siRNA sequence was selected from a region of the CHP third EF-hand motive. Fig. 12 (B and C) shows (by immunoblot and immunostaining, respectively) that CHP siRNA-transfected OK cells exhibited markedly reduced expression of endogenous CHP (by ϳ70%, see Fig. 12B). In contrast when siRNA targeting coding sequence for Pyk2 and GL2 were used, no changes in CHP protein expression were detected (Fig. 12B). Transfection of OK cells with siRNA targeting CHP reduced total NHE3 expression by 50% but did not modify Na/K-ATPase protein abundance (Fig. 12D). Interestingly, when CHP expression is reduced, baseline NHE3 activity is significantly decreased and CPA no longer regulates NHE3 function (Fig. 12E). Together, these findings strongly suggest that CHP is an essential cofactor required for inhibition of NHE3 by activated A 1 R. Furthermore, CHP may trigger other mechanisms involved in NHE3 regulation. DISCUSSION In addition to maintenance of sodium homeostasis under physiological conditions, a rise in adenosine after renal ischemia has protective effects against renal cell damage via a multitude of intrarenal and extrarenal mechanisms (2,(61)(62)(63). Regulation of NHE3 activity by adenosine may represent a versatile mechanism to adapt the energy supply/demand balance of the proximal tubule under physiological as well as certain pathophysiological conditions. Our recent findings support a role of A 1 R activation in modulation of proximal tubule Na ϩ transport via regulation of NHE3 (7,9). Using homologous as well as heterologous systems, we found that acute A 1 R stimulation inhibits NHE3 primarily via a Ca 2ϩ -sensitive, PKC-coupled pathway (7,9). In the present study we report that stimulation of A 1 R affects NHE3 in a mode similar to that of parathyroid hormone, dopamine, or A 2 R activation (8,40,51,52). Early effects (Ͻ30 min) of A 1 R regulate intrinsic transport activity of NHE3, and more sustained (Ͼ30 min) inhibition is associated with reduction of NHE3 protein in the plasma membrane. Reduction of NHE3 cell surface expression (within 30 min) was not accompanied by a comparable decreased of NHE3 activity at 30 min. The present findings are quite similar to that seen in response to parathyroid hormone or dopamine in OK cells (40,52). The molecular mechanisms underlining this observation are currently unknown. Two hypothetical models have been proposed that may explain the findings (52). In the first model, at 15 min only a fraction of surface NHE3 is down-regulated. This same fraction at 30 min undergoes endocytosis leaving the remaining surface transporters running at a rate similar to control conditions. In the second model, all surface NHE3 is inhibited (within 15 min), but only a restricted population of NHE3 becomes internalized at 30 min. Those transporters that are not targeted to be internalized recover their transport activity with time (30 min). Currently, there are no data to discern these two paradigms.
In fibroblasts, phosphorylation of NHE3 may have a permissive role in regulation of NHE3 activity by PKC (37), but NHE3 phosphorylation is insufficient to mediate the functional regulation (37,64). In A6 cells, rat NHE3 bearing either single or six substitutions of PKC phosphorylation sites retained functional regulation of transport activity by A 1 R activation. However, our present findings do not exclude the possibility that other phosphorylated sites on the NHE3 C terminus are involved in A 1 R-induced inactivation of NHE3.
A 1 R-induced NHE3 inhibition requires the integrity of a domain located between amino acids 462 and 552 of the opossum NHE3 sequence. This domain does not include classic PKC consensus sites compatible with the notion that activation of A 1 R does not involve changes in NHE3 phosphorylation. A cofactor recently identified to bind in the vicinity of this region is CHP (28). By screening for proteins that may function as regulators of NHE3 in a yeast two-hybrid study, we provided evidence that CHP interacts directly with a region of NHE3 corresponding to the domain that mediates A 1 R-induced inhibition. The juxtamembrane localization of the binding domain of CHP on NHE3 is consistent with a previous report (28). This prompted us to study the role of CHP in NHE3 regulation by adenosine.
Association of CHP with NHE3 in vivo was confirmed by co-immunoprecipitation in OK cells with either native or transfected NHE3 and CHP. OK cells have been shown recently to predominantly express the endogenous CHP2 isoform (65). However, our findings demonstrate that the OK cell CHP exhibits highest homology with human and murine CHP1 but not CHP2. There is considerable heterogeneity among different OK cells, and our cell line may express both CHP1 and CHP2 but the reverse transcription-PCR with CHP1-derived primers selectively picked out OK CHP1 rather than CHP2. Similar to reported data on direct interaction of cloned CHP from differ-ent species with NHE3 (28, 65), our results in OK cells indicate interaction of CHP with NHE3 under basal conditions. In addition, we showed that NHE3-CHP binding increases upon A 1 R stimulation with CPA, and we further provided functional data to support the role of this interaction in regulation of NHE3 by either CPA.
The CHP-interacting region on NHE3 derived from the yeast two-hybrid data was confirmed in OK cells by the observation that truncation of NHE3 at aa 552 still retained baseline and CPA-activated association of CHP with NHE3. In contrast, truncation of NHE3 at aa 462 showed no binding of NHE3 to CHP. The functional significance of this interaction was supported by additional findings. Truncation of NHE3 beyond aa 552 abolished CPA-induced regulation. This finding per se does not rule out regulatory functions other than CHP binding in this region. However, transient overexpression of an NHE3 peptide, aa 462-552, which competes for binding of CHP to NHE3, prevented CPA-induced early inhibition of NHE3 activity and CPA-mediated binding of CHP to NHE3. The functional role of CHP was further secured by the fact that reduction of native CHP by siRNA abolished the ability of CPA to inhibit NHE3 activity. Importantly, reduction of native CHP by siRNA influenced baseline NHE3 activity and total NHE3 protein abundance suggesting that CHP may influence NHE3 regulation by activating diverse mechanisms.
As noted above, A 1 R activation inhibits NHE3 by decreasing its intrinsic activity as well as surface protein. The mechanism whereby CHP binding to aa 462-552 affects intrinsic transport NHE3 activity is unknown. It has recently been suggested that Ca 2ϩ allows a conformational change of CHP and that conformational changes in CHP transduces cellular Ca 2ϩ signals to other cellular proteins, somewhat akin to the role of calmodulin (66). Furthermore, CHP has potential phosphorylation sites for PKC (26). It is conceivable that a finite number of NHE3 proteins are bound to CHP under basal conditions. After CPA stimulation, the Ca 2ϩ -induced conformational change in CHP stimulates further association of CHP to NHE3, although Ca 2ϩdependent binding of CHP to NHE1 could not be demonstrated with recombinant proteins in vitro (28). PKC upon activation moves from the cytoplasm to plasmalemma. Because CHP1 associates with microtubules (30) and the microtubule surface may provide a scaffold of two or more factors that otherwise do not directly interact (67), it is possible that the microtubule surface is responsible for bringing PKC into the vicinity of NHE3. Phosphorylation of NHE-bound CHP may then inactivate NHE3 by directly altering the transport subdomain of NHE3. It has been proposed that CHP attenuates the stimulation of NHE1 by serum and a mutationally activated GTPase in CCL39 cells by a change of CHP phosphorylation state (26,68).
Comparable to previous findings in PS120 fibroblasts and OK cells expressing exogenous CHP2 (28,65), CHP partly co-localized with NHE3 in the surface membrane in OK cells and A6 cells. Microtubule dynamics appear to play a role in membrane trafficking events (69), and EF-hand Ca 2ϩ -binding microtubule-interacting proteins, such as p22/CHP1, have 60 g of total lysate was immunoblotted by anti-CHP antibody. Cells were exposed to transfection reagent either without siRNA (Untransfected) or with siRNA against CHP (CHP siRNA), protein tyrosine kinase (Pyk2 siRNA) or non-mammalian protein (GL2 siRNA). The arrow indicates expected mobility of CHP. C, immunostaining of native CHP in untransfected (left image) and siRNA transfected OK cells (right image). Immunostaining was performed using anti-CHP antibody. XZ and XY indicate orthogonal optical sections. The apical membrane is indicated by an arrow on the XZ image. D, typical immunoblot and quantification of native CHP, NHE3, and Na/K ATPase total protein content in untransfected and CHP siRNA-transfected OK cells. 60 g of total lysate was immunoblotted by anti-CHP, anti-NHE3 (#5683), or anti-Na/K ATPase (␣2F), respectively. E, NHE3 activity of untransfected or CHP siRNA-transfected OK cells in response to a 15-min treatment with 10 Ϫ6 M CPA or vehicle. Data presented are means Ϯ S.E. The numbers in parentheses refer to the number of experiments performed under identical experimental conditions. The asterisk indicates the statistical difference compared with untreated cells (*, p Ͻ 0.05; **, p Ͻ 0.01 ANOVA). been suggested to influence membrane trafficking by effecting either directly the components of the vesicle transport machinery or indirectly the organization of the cytoskeleton (30). NHE3 is sensitive to the organization of the cytoskeleton (70). It is possible that CHP can exert its effect by altering the organization of the cytoskeleton. In the present study we also demonstrated that decreased surface expression of NHE3 is associated with inhibition by prolonged exposure of A 1 R to CPA. CHP1 is involved in targeting/fusion of transcytotic vesicles with the apical membrane (25). Although, NHE3 is endocytosed into clathrin-coated vesicles (40,71), exactly how CHP regulates NHE3 trafficking is unclear.
In summary, we present evidence for a novel function of the ubiquitous Ca ϩ -binding protein CHP. In epithelial cells of the kidney, we have shown that activation of A 1 R leads to increased interaction of NHE3 with CHP. Physical interaction of NHE3 with CHP is necessary for acute regulation of the intrinsic transport activity of NHE3.