点击显示 收起
【摘要】 The function of the NaPiIIa renal sodium-phosphate transporter is regulated through a complex network of interacting proteins. Several PDZ domain-containing proteins interact with its COOH terminus while the small membrane protein MAP17 interacts with its NH 2 end. To elucidate the function of MAP17, we identified its interacting proteins using both bacterial and mammalian two-hybrid systems. Several PDZ domain-containing proteins, including the four NHERF proteins, as well as NaPiIIa and NHE3, were found to bind to MAP17. The interactions of MAP17 with the NHERF proteins and with NaPiIIa were further analyzed in opossum kidney (OK) cells. Expression of MAP17 alone had no effect on the NaPiIIa apical membrane distribution, but coexpression of MAP17 and NHERF3 or NHERF4 induced internalization of NaPiIIa, MAP17, and the PDZ protein to the trans -Golgi network (TGN). This effect was not observed when MAP17 was cotransfected with NHERF1/2 proteins. Inhibition of protein kinase C (PKC) prevented expression of the three proteins in the TGN. Activation of PKC in OK cells transfected only with MAP17 induced complete degradation of MAP17 and NaPiIIa. When lysosomal degradation was prevented, both proteins accumulated in the TGN. When the dopamine D1-like receptor was activated with fenoldopam, both NaPiIIa and MAP17 also accumulated in the TGN. Finally, cotransfection of MAP17 and NHERF3 prevented the adaptive upregulation of phosphate transport activity in OK cells in response to low extracellular phosphate. Therefore, the interaction between MAP17, NHERF3/4, and NaPiIIa in the TGN could be an important intermediate or alternate path in the internalization of NaPiIIa.
【关键词】 phosphate transport PDZ PDZK protein interaction opossum kidney cells
RENAL REABSORPTION OF INORGANIC phosphate (P i ) and its regulation involve molecular mechanisms more complicated than initially thought. After expression cloning of NaPiIIa ( 24 ), the transporter responsible for most renal P i reabsorption, a complex network of interacting proteins that modulate the activity of this transporter has been described (for reviews, see Refs. 4, 5, 15, 35 ).
Most of the proteins that interact with this transporter contain one or more PDZ domains (acronym of the postsynaptic density protein PSD-95, the Drosophila junctional protein Disc-large, and the tight junction protein ZO1). These domains consist of 80-90 amino acids that very often (but not always) scaffold specific proteins into supramolecular complexes ( 17 ). Several PDZ domain-containing proteins interact with the COOH end of NaPiIIa ( 12 ), namely NHERF1 EBP50 (NHE3 Regulatory Factor 1) ( 34 ) and NHERF2 E3KARP ( 36 ), PDZK1 CAP70/NaPi-Cap1/NHERF3 (PDZ Protein of Kidney 1) ( 19 ) (now renamed NHERF3) ( 11 ), and PDZK2 NaPi-Cap2/NHERF4 ( 12 ) (now renamed NHERF4) ( 11 ), CAL PIST/GOPC/FIG ( 5, 9 ), and Shank2E ( 26 ). In addition, several proteins interact with NaPiIIa independently of PDZ domains. For example, MAP17 DD96/SPAP is a small membrane protein that interacts with the NH 2 terminus of the phosphate transporter ( 28 ). MAP17 was first identified by differential display because its RNA was overexpressed in several carcinomas ( 18 ). In addition, it was also isolated during the expression cloning of a renal Na/ D -mannose cotransport in Xenopus laevis oocytes, and it was considered to be an activator of a mannose transporter endogenous to the oocyte ( 6 ). Yeast two-hybrid experiments using the COOH-terminus end of MAP17 showed that this protein interacts specifically with NHERF3. These assays failed to find interactions with other PDZ-containing proteins ( 6, 13, 28 ).
Overexpression of MAP17 in the mouse liver provided the first clues toward the understanding of its function ( 29 ). Hepatic overexpression of MAP17 resulted in liver deficiency of both NHERF3 and the high-density lipoprotein receptor SR-BI, with a consequent increase in plasma HDL. The reduction in NHERF3 levels was due to an increase in its degradation through a proteasome-independent mechanism. However, expression of MAP17 in opossum kidney (OK) cells did not modify the pattern of expression of NaPiIIa, nor the magnitude or nature of its adaptations to low or high phosphate concentrations, nor its response to parathyroid hormone ( 28 ).
In the present work, we aimed to clarify the role of MAP17 and its protein interactions. To achieve this, we performed an extensive protein-protein interaction screening using a bacterial two-hybrid system and a mouse kidney library with two baits derived from the mouse MAP17 sequence. The identified interactions were confirmed by additional biochemical methods and colocalization of the proteins in intact kidney and in transfected OK cells was shown by fluorescence microscopy. The significance of the MAP17 protein interactions in OK cells was further analyzed by studying the effect of various pharmacological treatments on the protein localization in coexpressing OK cells.
MATERIALS AND METHODS
Cloning of MAP17-interacting proteins. A mouse kidney cDNA library, cloned into pTRG plasmid of the Bacteriomatch II Two-Hybrid System, was obtained from Stratagene (La Jolla, CA). Propagation and screening of the library were done in histidine auxotrophic mutant Escherichia coli XL1-Blue MRF' Kan cells containing a HIS3-aadA reporter cassette, following strictly the manufacturer's instructions. Two baits were generated from MAP17, both of them corresponding to the cytosolic domain of the protein. The first bait consisted of the last 63 amino acids of mouse MAP17 (accession number NM_026018 ), and it was prepared into pBT by directional cloning of PCR amplicons using Platinum Taq DNA Polymerase High Fidelity (Invitrogen S.A., Barcelona, Spain) and standard PCR-cloning strategies. All DNA analysis was done using Vector NTI Suite for Mac software (Invitrogen). The following primers were used, and contained, respectively, Eco RI and Xho I sites (both underlined) for directional cloning: sense 5'-CCG GAA TTC CCA CTT CTG GTG CCA GGA GGA-3', antisense 5'-CCG AT C TCG AG T CAC ATG GGT GTG CTG CGG AC-3'. The second bait corresponded to a truncated version of the previous bait, in that the last four amino acids (the PDZ binding site) were eliminated by PCR (Map STPM). The primers also contained Eco RI and Xho I sites for directional cloning into pBT: sense 5'-CCG GAA TTC CCA CTT CTG GTG CCA GGA GGA-3', antisense 5'-CCG AT C TCG AG T CA G CGG ACC CTG CCC TCT TCC T-3' (in bold antisense stop codon).
Screenings of the library were performed by cotransformation of E. coli with pools of library-containing pTRG plasmids and one of the MAP17 baits in pBT. Two consecutive selections of the cotransformant clones were performed with 3-amino-1,2,4-triazole (3-AT; a competitive inhibitor of His3 enzyme) and streptomycin.
Mammalian two-hybrid system. Studies of two-hybrid interactions in mammalian OK cells were done using the BD Matchmaker Mammalian Assay kit from BD Clontech (Mountain View, CA). Baits of MAP17 were generated by standard PCR amplification with primers harboring Eco RI and Mlu I sites: MAP17 full-length COOH bait, sense 5'-CG G AAT TC C ACT TCT GGT GCC AGG AGG A-3'; antisense 5'-CG A CGC GT A GCT TGT CAC ATG GGT GTG CTG-3'. For the Map STPM COOH bait, sense 5'-CG G AAT TC C ACT TCT GGT GCC AGG AGG A-3'; antisense 5'-CG A CGC GT T CAG CGG ACC CTG CCC TCT TCC T-3'. The amplicons were cloned directionally into pM. The cDNA preys (accession numbers are indicated in Tables 1 and 2 ) were inserted as full-length open reading frames into pVP16. The individual PDZ domains of the four NHERF proteins were inserted with primers having the following restriction sites. Mouse NHERF1 domains were cloned with Eco RI- Hin dIII: aa 1-100 for PDZ1, and aa 149-234 for PDZ2; mouse NHERF2 domains were cloned with Eco RI- Hin dIII: aa 1-98 for PDZ1, and 145-242 for PDZ2; mouse NHERF3 was cloned with Bam HI- Hin dIII: aa 1-95 for PDZ1, aa 114-217 for PDZ2, aa 224-322 for PDZ3, aa 346-461 for PDZ4; mouse NHERF4 domains were cloned with Eco RI- Sal I: aa 38-149 for PDZ1, aa 145-240 for PDZ2, aa 243-365 for PDZ3, and aa 407-487 for PDZ4. Complete open reading frames (orf) of mouse NaPiIIa and NHE3 (including its signal peptide) were also inserted in pVP16 with primers containing Eco RI and Sal I restriction sites. The activity of the reporter enzyme, secreted alkaline phosphatase (SEAP), was measured with a chemiluminescent BD Great EscAPe SEAP kit (BD Clontech) and recorded on a X-ray film from white opaque 96-well, flat transparent bottom microtiter plates (Corning, Acton, MA). The nonsaturated signals were quantified with a Fluor-S MAX MultiImager (Bio-Rad, Hercules, CA).
Table 1. MAP17-interacting proteins identified with Bacteriomatch II and their interaction strength in a mammalian two-hybrid system
Table 2. MAP17 protein interactions tested directly using the mammalian two-hybrid system
Expression of GST-fusion constructs and pull-down assays. The last 75 amino acids of rat MAP17 were fused to glutathione- S -transferase using plasmid pGEX-2T from GST Gene Fusion System (Amersham Biosciences, Barcelona, Spain) and standard high-fidelity amplification with primers that included restriction sites for Bam HI and Eco RI. Expression of soluble fusion proteins or inclusion bodies was obtained by transformation of BL21-Gold (DE3)pLysS cells (Stratagene) grown to 0.4 OD 600 and stimulation with 0.5 mM isopropyl-1-thio- - D -galactoside for 5 h at 30°C. Cells were spun and lysed for 20 min at room temperature with 5 ml Bugbuster Protein Extraction Reagent and 1 µl Benzonase (Novagen, Madison, WI) per gram of bacterial pellet in the presence of protease inhibitors (Complete-Mini, Roche Applied Science, Barcelona, Spain). Lysates were cleared at 16,000 g for 20 min at 4°C and fusion proteins were bound to Glutathione Uniflow Resin (BD Clontech). The resin was previously washed out of ethanol, equilibrated in 50 mM Tris·HCl, pH 8, and incubated with lysate for 30 min at 4°C. Fusion proteins were eluted with 10 mM glutathione and analyzed by Coomassie staining of polyacrylamide electrophoresed gels and Western blotting with anti-MAP17 polyclonal antibodies.
For pull-down assays, OK cells were grown in 35-mm dishes and transfected with 5 µg of cDNA encoding the MAP17-interacting protein of interest. These cDNAs were cloned into different plasmids with standard PCR-cloning technology: NHERF1 and NHERF2 were cloned into pEYFP-C1 plasmid (BD Clontech), while NHERF3 and NHERF4 were into pCMV-Myc (BD Clontech). After 48 h, the cells were lysed in a buffer containing 20 mM Tris, pH 8, 200 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40 and protease inhibitors. Lysates corresponding to 3 million OK cells were mixed with 1 ml GST-MAP17 fusion protein-loaded resin for 1 h with continuous mixing at 4°C, in 50 mM Tris·HCl, pH 8.0. The resin was washed in the same buffer, and the proteins were eluted with 20 mM glutathione and detected by Western blotting with a mouse anti-green fluorescent protein (Roche Applied Science), or a goat anti-c- myc antibody (Santa Cruz Biotechnology, Santa Cruz, CA).
Site-directed mutagenesis. Tagging of MAP17 with the 9 aa hemagglutinin antigen using the plasmid pCMV-HA from BD Clontech was not possible, most likely due to a posttranscriptional folding of the RNA that interferes with translation initiation. Instead, the HA-tag was introduced in the 23rd amino acid position by site-directed mutagenesis as explained ( 6 ), using the Quick-Change Site-Directed Mutagenesis kit (Stratagene) and the following sense and antisense primers: sense: 5'-GTC AAC AAT ACC CAT ACG ACG TCC CAG ACT ACG CGG GTC TAG GAA ACC-3', antisense: 5'-GGT TTC CTA GAC CCG CGT AGT CTG GGA CGT CGT ATG GGT ATT GTT GAC-3'.
Cell culture, transfections, and P i uptake assays. OK cell culture and P i uptake assays were performed exactly as reported ( 6, 25, 31 ) with [ 32 P]orthophosphoric acid as radiotracer. Transfections and cotransfections were achieved with Lipofectamine 2000 (Invitrogen) and cells at 90% confluency, following manufacturer's instructions. Maximal cotransfection was achieved in the following conditions: cells grown on eight-well Lab-Tek chamber slides (Nunc, Naperville, IL) were cotransfected with 2 µg MAP17 cDNA in pCMV-Script, 1 µg PDZ cDNA plasmid in pCMV-Myc, and 2 µl lipofectamine 2000 per well. Assays were done 24-48 h posttransfection.
SDS-PAGE, Western blots, and preparation of polyclonal antibodies. SDS-polyacrylamide gel electrophoresis and Western blots were performed as previously described ( 1, 22 ). For the preparation of peptide antibodies, a 15-aa peptide common to rat and mouse MAP17 and corresponding to the cytosolic, COOH-terminal region (aa 71-85, kadgvlvgmdgryss), was conjugated to keyhole limpet hemocyanin, mixed in Freund's complete adjuvant, and injected into New Zealand White rabbits (Davids Biotechnologie, Regensburg, Germany). Four booster injections in incomplete Freund's adjuvant were given monthly to the rabbits, and 35 ml of immune sera were collected, affinity-purified, and tittered. The antibody was used at 1 ) 1:200 dilution for immunohistochemistry and 2 ) 1:5,000 dilution for Western blotting.
Fluorescence microscopy. OK cells expressing HA-tagged MAP17, myc -tagged PDZ proteins, or the endogenous NaPiIIa (NaPi-4) transporter were grown on poly- L -lysine-coated coverglasses (Nunc), grown to confluency, fixed for 20 min with 3% paraformaldehyde in PBS supplemented with Ca 2+ /Mg 2+ (PBS 2+ ), and quenched for 10 min with 20 mM glycine before staining ( 6, 26 ). Immunohistochemistry was performed as described ( 23 ), in rat Wistar (Harlan, Barcelona, Spain) kidneys perfused with 4% paraformaldehyde in PBS. After being blocked with 10% goat serum in PBS, kidney slices were incubated overnight with primary antibodies as indicated in the text, and the next day rinsed and incubated with Alexa Fluor-conjugated secondary antibody (Molecular Probes, Eugene, OR) against the specific IgG/IgY of the primary antibody. HA epitope was detected with a rabbit anti-HA antibody (Sigma) and the myc -tagged proteins with a monoclonal or a goat polyclonal anti- myc antibody (Oncogene Research Products). For structure and organelle identification, several proteins were used as markers: brush-border membrane (BBM) was localized with a goat anti- -actin antibody (Santa Cruz Biotechnology); trans -Golgi network (TGN) with a monoclonal anti-TGN38 antibody (Oncogene Research Products), a monoclonal anti- -COP antibody (Sigma), or by transfection of OK cells with pEYFP-Golgi encoding the NH 2 -terminal 81 amino acids of human 1,4-galactosyltransferase (BD Clontech); early endosomes with a goat anti-early endosomal antigen A1 (EEA1) antibody (Santa Cruz Biotechnology); and lysosomes with a monoclonal anti-lysosome-associated membrane glycoprotein-1 (LAMP-1; Santa Cruz Biotechnology). NaPiIIa from OK cells (NaPi4) ( 30 ) was immunodetected with a rabbit antiserum against the COOH-terminal amino acid sequence ( 20 ). In both immunocytochemistry and immunohistochemistry, samples were mounted in 90% glycerol (Merck, Darmstadt, Germany), 10% PBS containing 2.5% 1,4-diazabicyclo(2.2.2)octane (Sigma). As indicated in text, some proteins were visualized by fusing them to fluorescent proteins. Immunodecorated preparations were imaged and analyzed using a laser-scanning confocal microscope (either an LSM510 or a LSM 5 PASCAL, Carl Zeiss, Thornwood, NY) with a x 40 water immersion objective and the corresponding postacquisition software.
Preparation of BBM and Golgi apparatus membranes. Wistar rats were killed by CO 2 narcosis, and the blood was drawn from the aorta. BBM were prepared by Mg 2+ precipitation as described ( 21 ), after homogenization of cortical slices in 300 D -mannitol, 5 mM EGTA, pH 8.0, 0.5 mM PMSF, and 16 mM HEPES-Tris, pH 7.5. The final pellet was resuspended in a buffer of 300 mM mannitol and 16 mM HEPES-Tris, pH 7.5. Golgi membranes from kidney cortex were obtained by adaptation of the dextran homogenization method ( 27 ). Rat kidney slices were resuspended in 20 ml of ice-cold dextran buffer (500 mM sucrose, 50 mM Tris-maleic acid, pH 6.4, 1% dextran MW 225,000 and protease inhibitors) and homogenized for 40 s at 10,000 rpm and 4°C in a Disperser DIAX 600 (Heidolph, Kelheim, Germany), using 10-s pulses and 10-s rests. The homogenate was then centrifuged at 5,000 g for 15 min at 4°C, and the upper layer of the pellet was resuspended in 1 ml of residual supernatant by gentle agitation. This suspension was then layered over 2 vol of sucrose barrier buffer (1.2 M sucrose, 50 mM Tris-maleic acid, pH 6.4) and centrifuged 30 min at 120,000 g and 4°C. The Golgi membranes (stacks) were recovered from the interface and washed with 3 vol of dextran buffer for 20 min at 10,000 g and 4°C. BBM and Golgi membrane protein were quantified by the Bradford method ( 7 ).
Pharmacological treatment of OK cells. All pharmacological reagents were from Sigma. OK cells, 6 h after transfection, were incubated with 10 µM H-89, 1 µM chelerytrine, or 100 µM PD098059 for 3 h to inhibit PKA and PKC, and extracelular signal-regulated kinase, respectively. To activate PKA, transfected OK cells were incubated for 2 h with 100 µM 8-bromo-cAMP and 100 nM forskolin. To activate protein kinase C, the transfected cells were incubated for 1 h with 10 nM phorbol 12-myristate 13-acetate (PMA). To inhibit lysosomal and proteasome-mediated protein degradations, cells were incubated for 2 h with 1 µM bafilomycin A1 and 10 µM MG132 (Z-Leu-Leu-Leu-al), respectively. To inhibit protein translation, OK cells were incubated with 10 µM cycloheximide from 8 to 48 h, and the culture medium was changed every 12 h with fresh medium containing the inhibitor at the same concentration. To activate the dopamine D1-like receptor, 24 h after transfection with HA-tagged MAP17, OK cells were incubated with 10 µM fenoldopam for 1 h.
Phosphate adaptation experiments. In OK cells these assays were performed 24-48 h posttransfection of the respective plasmidic cDNAs. For acute adaptation of the cells to low and high phosphate in the cell culture medium, cells were made quiescent by overnight exposure to DMEM/Ham's F12 without serum and then exposed for 1 h to either 0.1 or 2 mM potassium phosphate in DMEM (Sigma, St. Louis, MO) as described before ( 8 ).
In Wistar male rats (Harlan) adaptation to changes in dietary P i was done as published before ( 22 ) with minor modifications. In short, the animals were trained to consume their diet between 8 AM and 1000 AM (2 h). The animals were then pair-fed either a high-P i (1.2%) or a low-P i (0.1%) diet for 5 days. On day 6 the following three groups of four rats in each experimental group were studied: group 1, rats that were chronically fed a 1.2% P i diet and on the day of the experiment were also fed a 1.2% P i diet for 2 h; group 2, rats that were chronically fed 0.1% P i and on the day of the experiment were also fed a 0.1% P i diet for 2 h; group 3, rats that were chronically fed 0.1% P i and on the day of the experiment were acutely fed a 1.2% P i diet for 2 h. On the day of the experiment at the end of the feeding period, the rats were anesthetized with intraperitoneal pentobarbital sodium, and the kidneys were rapidly removed for immunofluorescence as explained above.
Statistics and data analysis. Analysis of uptake data was done as previously described ( 6 ). Comparison of means and determination of significances were done using unpaired t -test.
RESULTS
Identification of proteins that interact with MAP17. A mouse kidney cDNA library from the Bacteriomatch II Two-Hybrid System (Stratagene; see MATERIALS AND METHODS ) was screened with a bait corresponding to the last 63 amino acids of mouse MAP17. A total of 1,080,000 clones were screened in 9 cotransformations, from which 728 corresponded to NHERF3, 108 to NHERF4, and 14 to the chaperone HSP40 DnaJ/Tid56. These and other minor cDNAs are indicated in Table 1. To increase the chance of baiting non-PDZ proteins, we used a truncated bait (Map STPM), without the last four amino acids of the COOH tail of MAP17. With this bait, we screened 803,000 clones in 8 cotransformations, obtaining very few and heterogeneous interacting proteins, whose pTRG clones were retested for interaction with the full-length bait. Three of them were confirmed and are included in Table 1.
In addition to the different interaction confirmation methods of the Bacteriomatch system (3AT and streptomycin selection), we tested the interactions identified with the Bacteriomatch system with a mammalian two-hybrid system (BD Clontech). OK cells were cotransfected with the bait (full-length or truncated MAP17 variants), prey (mouse kidney library cDNA identified as interactive in the Bacteriomatch II system), and reporter plasmids. The intensities of the interaction signals derived from the activity of the reporter enzyme (secreted alkaline phosphatase, SEAP) are indicated in Table 1 and show that most of the clones obtained with the Bacteriomatch system also exhibited interacting activity in the OK cell system. In particular, NHERF4 was confirmed as a novel and strong interacting partner of MAP17. In addition, the truncated bait failed to interact with the PDZ protein preys. Some of the interaction signals with both baits are shown in Fig. 1. Cotransfection of MAP17 bait with the empty prey plasmid (pVP16) did not activate the expression of SEAP.
Fig. 1. In vitro interaction assays between the indicated proteins. Left : mammalian 2-hybrid interaction intensities obtained after exposure of the phosphatase alkaline reaction to an X-ray film. As indicated, 2 baits were assayed from the MAP17 sequence: complete COOH terminus and a truncated bait in which the last three amino acids were eliminated by site-directed mutagenesis (MAP·-bait). Right : GST pull-down of the indicated PDZ proteins using similar baits as fusion proteins with glutathione transferase. Only the tax-interacting protein (TIP) was negative in both assays.
Because the previously reported interaction between MAP17 and NaPiIIa ( 28 ) was not detected during the screening of the mouse kidney cDNA library, we directly tested the NaPiIIa cDNA as well as several other relevant cDNA preys for interaction with MAP17 baits in the OK cell two-hybrid system ( Table 2 and Fig. 1 ).
All of the PDZ domain-containing fusion proteins of the NHERF family as well as the nuclear Tax Interacting Protein TIP1 ( 16 ) were found to interact with the MAP17 COOH bait, NHERF3 and NHERF1 yielding the strongest interaction signals. Each of the PDZ domains of the PDZ proteins was also tested individually for interaction with the MAP17 COOH bait. We found interactions with the second and fourth PDZ domains of NHERF3, with the first and third domains of NHERF4, and with the first PDZ domain of NHERF1 and of NHERF2. All interactions with the PDZ-domain baits faded when the truncated version of the MAP17 COOH end bait was used, i.e., when the last four amino acids of the COOH terminus were omitted.
We also confirmed interaction of the MAP17 COOH bait with two non-PDZ-containing proteins, the endogenous OK cell NaPiIIa (NaPi-4) ( 28 ), and the sodium-hydrogen exchanger NHE3 ( Table 2 ). In both cases, the interactions showed a very weak intensity.
MAP17 interactions with PDZ domain-containing proteins were verified by pull-down assays 48 h after transfection of the myc -tagged versions of the proteins ( Fig. 1 B ). With the exception of TIP, all PDZ proteins were successfully precipitated with the GST-MAP17 COOH-terminus fusion protein, therefore confirming the presence of the interactions observed with both the bacterial and the mammalian two-hybrid systems. GST alone was unable to precipitate any of these PDZ proteins ( Fig. 1 B ).
Subcellular distribution of MAP17 and the NHERF proteins. Full-length MAP17 tagged with HA was found in both the apical membrane and in the TGN ( Fig. 2 A ) of OK cells as described ( 6, 28 ). MAP17 immunostaining of the apical membrane showed a typical patched distribution pattern also observed with NaPiIIa and other membrane proteins ( 8, 20, 28 ). Some cells only exhibited TGN staining and corresponded to nondifferentiated cells. Expression of a truncated MAP17, in which the last three amino acids were eliminated by introduction of a stop codon at position 2, was not restricted to the TGN but was distributed throughout the cytoplasm, with reduced expression in the apical membrane (data not shown) ( 28 ). On the other hand, each of the four NHERF proteins was expressed at the apical membrane and exhibited a patched pattern ( Fig. 2 A ). None of them was detected in TGN, but all four proteins showed additional variable expression throughout the cytosol, especially NHERF4.
Fig. 2. Effect of MAP17 on the cellular distribution of PDZ proteins. A : confocal images of opossum kidney (OK) cells expressing the indicated proteins. MAP17 is in green, decorated with a rabbit anti-HA polyclonal antibody and an anti-rabbit IgG secondary antibody conjugated with Alexa 488 (Molecular Probes). PDZ proteins are in red: NHERF1 and NHERF2 were fused to EYFP protein, while NHERF3 and NHERF4 were myc -tagged, detected with a monoclonal anti- myc antibody and labeled with a goat anti-mouse antibody conjugated with Alexa 568 (Molecular Probes). All 5 proteins exhibit an apical pattern of patched expression. Intracellular expression is more heterogeneous, with NHERF proteins having a diffusive expression of variable intensity, while MAP17 is the only protein showing an exclusive staining in the trans -Golgi apparatus. B : confocal microscopy colocalization of the indicated coexpressed proteins, with the same antibodies as in A. Only merged images are shown of coronal and side views of the cells obtained through the indicated sectioning lines. Whereas NHERF1 and NHERF2 apical expression was not altered by MAP17, NHERF3 and NHERF4 apical staining showed a reduced intensity up to a complete annulment depending on the cell. Intracellular expression also was nonhomogeneous: NHERF1 and NHERF2 expression was not altered by MAP17, whereas NHERF3 and NHERF4 showed a new intracellular staining in the trans -Golgi that colocalized completely with MAP17 (yellow). C : localization of expressed MAP17 ( pseudocolor green ) and NHERF3 ( pseudocolor red ) to the Golgi apparatus ( pseudocolor blue ). The trans -medial region of the Golgi apparatus was labeled by transfection with a fusion of enhanced yellow fluorescent protein (EYFP) and the NH 2 terminal 81 amino acids of human 1,4-galactosyltransferase (see MATERIALS AND METHODS ). The white pseudocolor in the merged image indicates complete colocalization of the 3 peptides (MAP17, NHERF3, and EYFP-Golgi marker).
Effect of MAP17 on the subcellular distribution of its interacting proteins. Coexpression of MAP17 with either NHERF1 or NHERF2 did not alter the expression of any of the three proteins, as shown in the coronal and side views of the cells in Fig. 2 B. They colocalized in the apical membrane of OK cells (yellow in the merged image) but not at intracellular sites such as the TGN where only MAP17 was found. However, in most cells, cotransfection of MAP17 with NHERF3 or with NHERF4 resulted in nearly completely abolished apical expression of MAP17, NHERF3, or NHERF4. Apical localization of MAP17 remained only in a few NHERF3- or NHERF4-cotransfected cells and with much reduced intensity. In addition, MAP17, NHERF3, and NHERF4 were all strongly expressed in the TGN (yellow in the merged image) of cotransfected cells. Whereas we previously showed localization of MAP17 in the Golgi by costaining with dye-labeled ceramide ( 6 ), we now confirmed the morphological evidence for colocalization in the TGN by triple transfection with an EYFP-tagged Golgi protein fragment, HA-MAP17, and myc -NHERF3 (see MATERIALS AND METHODS ). The white color in the merged image ( Fig. 2 C ) indicates perfect colocalization of the three proteins in the TGN.
The selectivity of the effect of MAP17 in the NHERF family of proteins suggests a specificity of the mechanism of MAP17 action under conditions of exogenous transfection. Even so, we studied the endogenous expression of native MAP17 in rat kidney slices, with the aim to analyze its presence in the trans -Golgi. As reported, most of the MAP17 staining was at the BBM of proximal tubular cells where it colocalized with -actin filaments ( Fig. 3 A, top, arrow) ( 18, 28 ). However, MAP17 staining was also present intracellularly, although with a lesser intensity, and colocalized significantly with the trans -Golgi marker TGN38, in agreement with our previous work in OK cells ( Fig. 3 A, second row, arrow) ( 6 ). Therefore, expression of MAP17 in the TGN is not simply a consequence of its overexpression. In the normal kidney in steady-state conditions, we were unable to find MAP17 in endosomes or in lysosomes ( Fig. 3 A, third and fourth rows). Expression of MAP17 at the TGN of rat kidney cortex was also evaluated by Western blot using purified BBM and membranes isolated from the Golgi apparatus. Signals of identical size were obtained with both types of membranes, confirming expression of MAP17 in the TGN of the normal rat kidney ( Fig. 3 B ).
Fig. 3. Expression of MAP17 in the rat kidney cortex. MAP17 was identified with a polyclonal antibody made in rabbit. A, top : MAP17 (green) colocalizes with -actin (red) in the brush-border membrane (BBM) of proximal tubular cells (arrow). Row 2 : MAP17 (red) is also detected in the trans -Golgi apparatus using an antibody marker against the trans -Golgi network (TGN) membrane protein 38, TGN38. Rows 3 and 4 : MAP17 does not localize to endosomes ( row 3, using an early endosome antigen 1 antibody, EEA1) or to lysosomes ( row 4, using an antibody against lysosome-associated membrane glycoprotein-1, LAMP-1). In all images, green colors were obtained using a secondary antibody conjugated with Alexa 488; red colors using an Alexa 568-conjugated antibody. B : Western blot of MAP17 expression in BBM and Golgi membrane preparations. A polyacrylamide gel was loaded with 20 µg of BBM proteins and 50 µg of Golgi membrane proteins and blotted with a rabbit anti-MAP17 primary antibody and a horseradish peroxidase-conjugated goat anti-rabbit IgG.
Effects of MAP17/NHERF on NaPi trafficking. Next, we investigated whether coexpression of MAP17 and NHERF3/4 also induced redistribution of the endogenous OK cell NaPi4 (OK cell variant of NaPiIIa), from the apical membrane to the TGN. In untransfected OK cells, NaPi4 is mainly expressed in apical membrane patches ( Fig. 4 A ). In agreement with previous reports, single transfections of either MAP17 or any of the four NHERF cDNAs did not alter the steady-state distribution of NaPi4 (data not shown) ( 13, 14, 28 ). However, when MAP17 and NHERF3 were coexpressed ( Fig. 4 B ), NaPi4, MAP17, and NHERF3 were all detected in the TGN as indicated by the white arrows and the white colocalization signal was observed in the merged image. The same effect was obtained with NHERF4 ( Fig. 4 C ). Coexpression of MAP17 with NHERF1 or NHERF2 did not alter the localization of NaPi4 (data not shown), as expected since the distribution of these two PDZ proteins was not altered when they were individually coexpressed with MAP17 ( Fig. 2 ). This again supports the specificity of the MAP17 effect.
Fig. 4. Effect of MAP17-PDZ interaction on the distribution of NaPiIIa in OK cells. A : in steady-state conditions, NaPiIIa is exclusively expressed in the apical membrane, in a patched distribution, as shown in the coronal and side confocal views of the cell. B : simultaneous expression of MAP17 and NHERF3 induces a change in NaPiIIa expression from the brush border to an intracellular compartment, the TGN (white arrows); colocalization of green, red, and blue pseudocolors provides a white merge as seen in both the coronal and side confocal views of the cells. C : coexpression of NHERF4 and MAP17 induces the same distribution change in NaPi4 as in B (white arrow). As a control, arrowhead indicates a nontransfected cell in the same field, showing NaPi4 in the original patched pattern of expression. MAP17 is labeled with a monoclonal anti-HA and visualized with an anti-mouse antibody conjugated with Alexa 488. NHERF3 and NHERF4 are immunodecorated with a goat polyclonal anti- myc, and a secondary anti-goat antibody conjugated with Alexa 633. NaPi4 is detected with a polyclonal rabbit antibody and with a secondary anti-rabbit antibody conjugated with Alexa 568.
The reduction in the abundance of NaPi4 in the apical membrane of cells cotransfected with MAP17 and NHERF3 was accompanied by a simultaneous reduction in the Na-dependent phosphate transport rate. The maximal reduction in phosphate uptake achieved was 50%, a percentage similar to the measured cotransfection efficiency. No change in Na-Pi cotransport rate was detected with single transfections of MAP17, NHERF1, or NHERF3, nor with cotransfection of MAP17 and NHERF1 ( Fig. 4 D, and data not shown).
Time course of induced NaPi redistribution. The change in the distribution of NaPiIIa, MAP17, and NHERF3/4 could be due to internalization from the apical membrane or to inhibition of newly synthesized protein insertion into the apical membrane (exit from the trans -Golgi). To distinguish between these possibilities, the time course of the changes in expression of the proteins was evaluated in OK cells cotransfected with MAP17 and NHERF3 ( Fig. 5 A ). NHERF3 starts to be expressed (at the apical membrane) about 6 h posttransfection, while expression of MAP17 occurs between 6 and 12 h posttransfection. As soon as expression of MAP17 is observed, there is a simultaneous decrease in the apical membrane expression and an increase in TGN expression of both NHERF3 and NaPi4. This indicates that MAP17 affects the cellular distribution of NHERF3 and NaPi4 within 6 and 12 h posttransfection, i.e., in less than 6 h after its initial expression.
Fig. 5. Time course of the effect of MAP17 and NHERF3. A : OK cells were cotransfected with MAP17 and NHERF3 and visualized at different time points. NHERF3 (blue) was first expressed at 6 h, but MAP17 (green) was only observed after 12 h of transfection. NaPi4 disappears from the apical membrane between 6 and 12 h of the cotransfection, coincident with the simultaneous expression of NHERF3 and MAP17. This suggests that the effect on NaPi4 is achieved in less than 6 h. At 24 h, MAP17 and NHERF3 are still observed in the apical membrane, but the patched distribution has been lost. Intracellular expression at 0 and 6 h was negative in all cases (not shown). B : confocal analysis of the half life of NaPi4 in the apical membrane. OK cells were incubated with 10 µM cycloheximide for the indicated durations. NaPi4 signal intensity at the apical membrane starts to be reduced at 12 h of treatment. The maximal reduction in intensity is seen at 24 h, when the patch expression pattern has disappeared.
To block insertion of newly synthesized proteins in the apical membrane, OK cells were treated with 10 µM cycloheximide, which inhibits the translation machinery of the cells ( Fig. 5 B ). Eight hours after starting cycloheximide exposure there was no detectable change in the apical distribution of NaPi4. Only after 12 h was the intensity of the apical NaPi4 expression significantly reduced compared with controls, but the presence of an apical patched localization was still observable. Inhibition of the apical expression of NaPi4 was complete only at 24 h, when only a diffuse apical staining was still observed. Therefore, inhibition of protein synthesis begins to alter the apical distribution of NaPi4 only 12 h after initiation of cycloheximide exposure, whereas the cotransfection-induced redistribution of NaPi4 is already complete in less than 6 h. These observations strongly suggest that the redistribution of NaPi4 is mostly due to internalization of transporters already expressed in the apical membrane.
Role of PKA and PKC in MAP17/NHERF3/4-induced trafficking of NaPi. To determine whether the protein redistribution to TGN after cotransfection could be modulated by a signal transduction process, we cotransfected MAP17 and NHERF3 in the presence of inhibitors of PKA, PKC, and extracellular signal-regulated kinase (ERK/MAPKK). Inhibition of PKA with 10 µM H-89 and ERK with 100 µM PD098059 had no effect on the redistribution of NaPi4, MAP17, or NHERF3: all three proteins colocalized in the TGN ( Fig. 6 A and data not shown). However, inhibition of PKC with 1 µM chelerytrine completely prevented the effect of MAP17 and NHERF3 cotransfection on protein distribution, as none of the three proteins was detectable in the TGN and all remained in the apical membrane ( Fig. 6 B ). This result does not necessarily mean that PKC is activated by the cotransfection of MAP17 and NHERF3, but it indicates that at least a basal activity of PKC is necessary, which, presumably, phosphorylates some component of the machinery involved in the process.
Fig. 6. Involvement of protein kinase C on MAP17-NHERF3 coexpression effect. OK cells were cotransfected with MAP17 and NHERF3 and after 6 h the cells were treated pharmacologically for 2 h with inhibitors of protein kinases. A : inhibition of PKA with 10 µM H-89, showing that under this condition MAP17 and NHERF3 cotransfection was still affecting the distribution of NaPi4, i.e., NaPi4, MAP17, and NHERF3 are now mainly located in the TGN (white merge). B : inhibition of PKC with 1 µM chelerytrine abolishes the MAP17-NHERF3 coexpression effect, because all 3 proteins are still apically expressed and located in membrane patches, and the amount in TGN is minimal.
Next, we examined whether activators of protein kinases could mimic the redistribution effect observed with MAP17-NHERF3 cotransfection ( Fig. 7 ). OK cells transfected with MAP17 were treated with activators of PKA (100 nM forskolin and 100 µM 8-Br-cAMP) or of PKC (10 nM PMA). Activation of PKA for 2 h had a moderate inhibitory effect on the expression of NaPi4 in the apical membrane (second row), and this effect became stronger with 4 h of treatment (data not shown). The effect of PKA activation on MAP17 was negligible at either 2 or 4 h of treatment. PKC activation had a very strong effect on MAP17 expression, very similar to its effect on NaPi4 (third row). In both cases, the apical immunofluorescent signal was absent after 2 h of incubation with phorbol esters, and no evidence of intracellular staining was observed. These data are compatible with the well-known PKC-induced degradation of NaPi4. To confirm the involvement of degradation pathways in the disappearance of NaPi4 and MAP17 induced by PKC activation, we selectively inhibited the proteasome with 10 µM MG132 (Z-Leu-Leu-Leu-al), and the lysosomal route with 1 µM bafilomycin A, while simultaneously stimulating PKC with PMA (fourth and fifth rows, respectively). MG132 had no effect on the PKC-mediated degradation of NaPi4 and MAP17, and no immunoreactivity was observed either apically or at any intracellular locations. However, inhibition of lysosomal degradation with bafilomycin A1 in PMA-activated cells not only prevented the complete degradation of both NaPi4 and MAP17, but resulted in strong expression of both proteins in the TGN. Finally, inhibition of the lysosomal or proteasomal degradation in MAP17-NHERF3-cotransfected cells did not prevent the expression of NaPi4, MAP17, and NHERF3 in the TGN (data not shown).
Fig. 7. Role of protein degradation pathways. OK cells transfected with MAP17 were pharmacologically treated for 2 h to activate PKA (100 nM forskolin plus 100 µM 8-Br-cAMP, row 2 ) and protein kinase C (10 nM phorbol esters, row 3 ). In both cases, NaPi4 is eliminated from the apical membrane and sent for lysosomal degradation. Apparently, MAP17 only responds to PKC activation. Inhibition of proteasomal degradation route with 10 µM MG132 ( row 4 ) did not prevent the effect of PKC activation on either NaPi4 or MAP17. However, inhibition of lysosomal degradation with 1 µM bafilomycin A1 completely avoided the final degradation of both proteins and showed their accumulation in the TGN ( row 5 ). Three first rows are apical views, while the rows 4 and 5 are intracellular views.
Effect of dopamine on NaPi trafficking. Activation of dopamine D1-like receptor with fenoldopam induces internalization of NaPi4 toward intracellular locations ( 3 ). Because the effect and the pattern of expression of NaPi4 induced by fenoldopam are very similar to our findings with MAP17 and NHERF3 coexpression, we studied the effect of fenoldopam on MAP17 ( Fig. 8 ). As shown before, single transfection of MAP17 in OK cells did not modify the NaPi4 expression pattern ( Fig. 8 A; compare nontransfected cell, arrowhead, and MAP17-transfected cell, arrow). Twenty-four hours after transfection, incubation of MAP17-transfected OK cells with 10 µM fenoldopam for 1 h induced expression of both MAP17 and NaPi4 in the TGN of MAP17 single transfected cells ( Fig. 8 A ). Therefore, internalization of NaPi4 and MAP17 from apical membrane to the TGN is a common feature of activation of D1 receptors (a physiological process) and of NHERF3/4 and MAP17 overexpression. Very interestingly, the internalization induced by fenoldopam is likely independent of NHERF3 and 4. As shown in Fig. 8 B, while HA-tagged MAP17 colocalizes with a EYFP-Golgi marker after fenoldopam treatment, transfected myc -tagged NHERF3 or NHERF4 (data not shown) does not. Therefore, the internalization of NaPi4 and MAP17 induced by fenoldopam likely is independent of NHERF3.
Fig. 8. Effect of activation of dopamine D 1 -like receptor. A : OK cells were transfected with MAP17 only, and after 24 h were incubated with 10 µM fenoldopam (an activator of dopamine D1-like receptors) for 1 h. In control cells, MAP17 (red) only colocalized with NaPi4 (green) in patches of the apical membrane and at the periphery of the cell (yellow), with no evidence of intracellular staining (white arrow). Compared with nontransfected cells (white arrowhead), the distribution of NaPi4 is not altered by single MAP17 transfection. The apical patched NaPi4 staining of the control cells appears in the intracellular view only because this cell is less high than the adjacent MAP17-transfected cell. These cells were still not seen in the top apical. Activation of dopamine D1-like receptors inhibited the apical expression of both NaPi4 and MAP17 (mainly), with disappearance of the patched pattern of expression. Instead, both NaPi4 and MAP17 colocalize intracellularly. B : OK cells were cotransfected with pEYFP-Golgi and either HA-tagged MAP17 or myc -tagged NHERF3. After 24 h, the cells were incubated with 10 µM fenoldopam for 2 h. While MAP17 is internalized to the TGN, NHERF3 is not and maintains its typical diffuse intracellular distribution. Only intracellular confocal sections of the cells are shown.
Effect of adaptation to P i concentrations on the expression of NaPiIIa and MAP17. We also sought for additional physiological relevance of the expression of MAP17 and NaPiIIa in the TGN by performing P i adaptation experiments both in OK cells and in rats.
In OK cells, we tested whether coexpressing MAP17 and NHERF3 blunted the well-known adaptive increase in NaPi4 activity induced by exposure of OK cells to low phosphate concentration in the incubation media ( 25 ). OK cells were transfected with several combinations of cDNA, and 24 h later were made quiescent by incubation in serum-free medium. The cells were then exposed for 1 h to culture medium containing either 2 or 0.1 mM potassium phosphate and assayed for 32 P i transport activity. Cells incubated in low phosphate medium exhibited a 66% increase in P i transport compared with cells incubated with 2 mM phosphate. However, OK cells cotransfected with MAP17 and NHERF3 at 50% efficiency failed to adapt to the low phosphate medium by increasing P i transport activity ( Fig. 9 A ). Immunofluorescence microscopy ( Fig. 9 B ) shows that cotransfected cells (arrowhead) exposed to low P i medium accumulated MAP17, NHERF3, and NaPi4 in an intracellular location. The same optical field shows as control a nontransfected cell growing in a different confocal level that exhibits an intense immunofluorescence signal for NaPi4 in apical membrane patches as a response to the low P i exposure (arrow).
Fig. 9. OK cells overexpressing MAP17 and NHERF3 do not adapt to low phosphate medium. A : OK cells transfected with the indicated cDNAs were incubated with 2 mM (open bars) or 0.1 mM phosphate (filled bars) in the culture medium for 1 h. Only the cells expressing MAP17 and NHERF3 do not show an increase in transport activity after exposure to the low phosphate medium. B : detail of 2 cells incubated with 0.1 mM phosphate medium. A nontransfected cell is indicated with an arrow and shows intense expression of NaPi4 in apical patches. An arrowhead indicates a cell that has been cotransfected with HA-tagged MAP17 and myc -tagged NHERF3 and shows both proteins and NaPi4 intracellularly, with no signal in the apical membrane, despite the low phosphate medium incubation.
In the rat we studied whether the renal adaptation to a low or high phosphate diet affected the renal expression of MAP17 and NaPi2 ( 22 ). Wistar rats were fed different P i diets as detailed in MATERIALS AND METHODS to study both chronic and acute adaptation mechanisms ( 22, 23 ). Figure 10 shows sections of representative proximal tubules immunodecorated with anti-NaPi2 ( left ) or anti-MAP17 ( right ). In all sections the Golgi apparatus was identified using a monoclonal anti- -COP antibody. Abundance of NaPi2 in the brush border was altered by P i diet as reported previously: it was maximal in rats chronically fed low P i diets, minimal in animals chronically fed a high P i diet, and low in rats acutely fed a high P i after having been on a chronic low P i diet. In contrast, changes in the abundance of MAP17 were not as dramatic as those observed for NaPi2, and only slight differences could be appreciated as published before ( 28 ). Animals which were fed a high P i diet, either chronically or acutely, showed an intense colocalization of -COP with NaPi2 and MAP17 (yellow in merged images). In contrast, for animals chronically fed a low P i diet, the colocalization of -COP with NaPi2 was negligible, and the colocalization of -COP with MAP17 was of low intensity.
Fig. 10. Adaptation of rat kidney to dietary phosphate. Rats were chronically fed 1.2 or 0.1% P i for 6 days. On day 6, one group of animals fed a 0.1% P i diet received a 1.2% P i diet instead (acute adaptation to high P i diet). Abundance of NaPi2 in the BBM of proximal tubular cells was maximal with the low P i diet, while the expression in the Golgi apparatus (merge with -COP immunodecoration) was more intense in the case of animals fed a high P i diet, either acutely of chronically. Changes in the expression of MAP17 in the brush border were of less intensity than NaPi2. However, the abundance of MAP17 in the Golgi mimicked that of NaPi2, i.e., it was of maximal intensity in animals fed a high P i diet.
DISCUSSION
In this work, we showed that MAP17, in addition to its well-known interaction with NHERF3 PDZK1 ( 13, 19, 28 ), also interacts with other PDZ proteins, namely NHERF1, NHERF2, and NHERF4 PDZK2 ( Fig. 1 ). Very interestingly, simultaneous expression of MAP17 and NHERF3 or of MAP17 and NHERF4 in OK cells changes the predominant expression of both PDZ proteins from the apical membrane to the TGN of the cells ( Fig. 2, B and C ). The effect is specific, because it is not observed when MAP17 is coexpressed with NHERF1 or NHERF2. However, when the proteins are expressed individually, MAP17 is the only protein that shows expression in the TGN, in addition to that in the brush border, either in OK cells ( Fig. 2, A and C ) or in the intact kidney ( Fig. 3 ). In addition, we observed that OK cells expressing MAP17 exclusively in the TGN correspond to nondifferentiated cells, as these cells are not stained with NaPi4 antibody (data not shown). The four NHERF proteins expressed individually in OK cells show a mostly apical expression with only a diffuse cytoplasmic pattern whose intensity depends on the particular NHERF protein ( Fig. 2 A ).
As shown before, single expressions of MAP17 or the NHERF proteins in OK cells did not cause any change in the expression of the endogenous sodium/phosphate cotransporter NaPi4 ( 13, 14, 28 ). However, when MAP17 was coexpressed with NHERF3 or NHERF4, NaPi4 also changed its expression pattern from the apical membrane ( Fig. 4 A ) to the TGN ( Fig. 4, B and C ). The effect on NaPi4 is accompanied by a reduction in the phosphate transport rate in OK cells ( Fig. 4 D ). Coexpression of MAP17 with NHERF1 or NHERF2 did not alter the expression of NaPi4, consistent with the lack of an effect of MAP17 on NHERF1 and NHERF2 expression. It is noteworthy that NaPi4 interacts with the third PDZ domain of NHERF3 and NHERF4 ( 14 ), while MAP17 interacts with other PDZ domains of NHERF3 and NHERF4. By contrast, MAP17 and NaPi4 use the same (first) NHERF1 and NHERF2 PDZ domain, and NaPi4 interacts more strongly with NHERF1 than with NHERF3 ( 14 ). This could mean that even if the interaction between MAP17 and NHERF1 or NHERF2 exists in vitro, the interaction in OK cells could be outcompeted by a stronger interaction between the COOH end of NaPi4 with NHERF1 and NHERF2, and therefore no effects of MAP17 are observed on these two proteins.
The change in the expression of NHERF3/4 and NaPiIIa could be due to internalization from the apical location or to inhibition of the maturation and exocytosis of de novo synthesized proteins (or to a combination of both mechanisms). In the second case, the reduction in the apical abundance of NaPi4 would be explained as a consequence of protein turnover and paced by its half-life. To elucidate the mechanism involved, we could not perform traditional biochemical studies because the NaPi4 antibody was not functional under such conditions. Therefore, we studied the time course of the changes in expressions of MAP17, NHERF3, and NaPi4 ( Fig. 5 ). The fact that inhibition of protein synthesis only affects the apical abundance of NaPi4 after 12 h, while the effect of MAP17-NHERF3 coexpression takes place within 6 h, indicates that the cotransfection effect is mainly due to internalization of NaPi4. Nevertheless, some contribution of inhibition of the export of new translated transporters to the membrane cannot be completely excluded under these experimental conditions.
We have shown that the protein redistribution effect of the MAP17-NHERF3 interaction requires, at least in part, PKC, while PKA and extracellular signal-regulated kinase are not involved. As shown in Fig. 6, the expression of NHERF3 and NaPi4 in the TGN is abolished when the cells are treated with the PKC inhibitor chelerytrine; however, no effect is observed with H-89, an inhibitor of PKA, or with PD098059, an inhibitor of ERK. In addition, activation of PKC with phorbol esters induces the well-known degradation of NaPi4 (e.g., Ref. 20 ) as well as degradation of MAP17 ( Fig. 7 ). When PKC activation is accompanied by inhibition of lysosomal activity with bafilomycin A1, both MAP17 and NaPi4 are accumulated in the TGN ( Fig. 7 ). Therefore, the expression in the TGN can be considered either as an intermediate or alternate step to the lysosomal degradation of the NaPiIIa.
We also performed physiological studies to correlate the changes in NaPi4 distribution after MAP17/NHERF3 cotransfection with some well-known regulatory mechanisms of P i homeostasis. For example, because activation of the dopamine D 1 -like receptor with fenoldopam also changes the expression of NaPiIIa from the apical membrane to intracellular locations in isolated mouse proximal tubules and in OK cells ( 3 ), we determined the effect of fenoldopam on MAP17 in OK cells and found that MAP17, too, was internalized and localized to the TGN in response to fenoldopam ( Fig. 8 A ). Therefore, activation of the dopamine D 1 -like receptor is enough to send both proteins to the TGN, without need of exogenous expression of NHERF3 or NHERF4. In fact, fenoldopam does not modify the expression of NHERF3 and does not move it to the TGN ( Fig. 8 B ). This indicates that activation of the dopamine receptor and cotransfection of MAP17 with NHERF3 induce MAP17-NaPi4 internalization by different mechanisms. In addition, PKA was found to be necessary for the fenoldopam-induced internalization, because this effect was blunted by the specific PKA inhibitor H-89 ( 3 ). By contrast, H-89 did not inhibit the internalization of NaPi4 induced by MAP17-NHERF3 cotransfection ( Fig. 6 ), again suggesting that the two internalization processes are initiated by different mechanisms and correspond to, at least, two parallel processes.
We also studied the relationship between expression of these proteins in the TGN and the adaptation to P i concentration. In OK cells we observed that MAP17-NHERF3 cotransfection reduces the apical expression of NaPi4 and the basal NaPi cotransport rate. In addition, the cotransfection interferes with the acute (within 1 h) adaptive upregulation in response to conditions of low phosphate concentration in the culture medium ( Fig. 9 ), a powerful stimulus that in native OK cells results from posttranslational increases in the apical abundance of the NaPi4 transporter and mimics the renal adaptation that occurs in intact animals acutely deprived of dietary P i ( 22, 23, 25 ). In the rat kidney, we observed an increase of both NaPi2 and MAP17 in the Golgi apparatus of proximal tubular cells in response to a high P i diet administered either chronically of acutely ( Fig. 10 ). The expression of NaPi2 was already described in this subcellular location ( 23 ), and we extended the observations to MAP17. In high phosphate diet treatments as well as in PTH, phorbol esters, and dopamine treatment, phosphate transport by the apical membrane is reduced and, at least in the first two cases, NaPi2 is targeted to the lysosome for degradation. Therefore, Golgi expression of NaPi2 under these circumstances and treatments could be viewed as an intermediate step in its degradation.
A similar phenomenon has also been observed for the CFTR conductance regulator: the PDZ protein CAL (CFTR-associated ligand) induces the translocation of CFTR from the plasma membrane to the trans -Golgi, where it is retained for subsequent targeting to the lysosomes under the control of the Rho GTPase TC10 ( 10 ). We found that expression of CAL in OK cells has the same effect on NaPi4 (and on MAP17) as on CFTR, i.e., a change in the expression of NaPi4 (and MAP17) from the apical membrane to the TGN of the cells (Lanaspa MA, unpublished results). Therefore, the MAP17-NHERF3 complex, analogous to CAL, may behave as a Golgi adaptor protein complex that sorts membrane transporters either synergistically or redundantly to the TGN and then to the lysosomes.
It is noteworthy that the OK cells contain most of these interacting proteins. We amplified by PCR-RACE, fragments of MAP17, NHERF1, NHERF3, NHERF4, and CAL from OK cell RNA (Giral H, unpublished results; some of them are being communicated to GenBank). Since these native proteins are likely under regulatory controls, their overexpression could overwhelm these regulatory systems and lead to interactions that affect the cellular localization of membrane proteins. Nevertheless, taken together the specificity of the redistribution effect of cotransfection of MAP17 with NHERF3 or 4 (and not with NHERF1 or 2), the endogenous presence of NaPiIIa and MAP17 in the trans -Golgi of kidney cortex epithelial tubular cells, and the similar redistribution effect induced by activation of dopamine D1-like receptors in OK cells and by a high phosphate diet in the proximal tubule suggest that internalization of apical MAP17 and NaPiIIa to the TGN is not simply an undesirable consequence of overexpression of proteins by transfection. Further research on the native proteins of the proximal tubule cells including selective protein knock down with siRNAs and identification of the signal transduction mechanisms involved in their regulation are ongoing and needed to understand the complexity of these interacting protein networks.
In conclusion, MAP17 is a strong interacting partner of NHERF3 (and NHERF4). This complex favors internalization of other associated proteins (NaPi2a, NaPi4) from the apical membrane to the trans -Golgi. We studied the case of NaPiIIa in OK cells, but it is tempting to suggest that similar effects take place with other NHERF3 (and 4)-interacting membrane proteins, such as the hepatic high-density lipoprotein SR-BI receptor ( 29 ), the Na + /H + exchanger NHE3 ( 13 ), the chloride exchanger CFEX ( 32 ), the urate-anion exchanger URAT1 ( 2 ), and the CFTR chloride channel ( 33 ). The retention of NaPiIIa in the TGN could represent a physiological state attained in different experimental conditions. We still have to understand the details of the MAP17 interactions with PDZ proteins and the pre- and post-Golgi mechanisms affected by MAP17. The complexity of protein-protein interaction networks and the regulatory factors that modulate renal solute reabsorption require an integrative approach where several components must be analyzed simultaneously: over- or underexpression of single elements is often insufficient to fully understand the complexity of the system under study.
GRANTS
This work was supported by Grants BFI2003-06645 from the Spanish Minister of Education and Science (to V. Sorribas), 5 R01-DK-066029-02 from National Institutes of Health (to M. Levi), a predoctoral fellowship AP2001-4101 from the Spanish Minister of Education and Science (to M. A. Lanaspa), and an American Heart Association postdoctoral fellowship (to S. Breusegem).
ACKNOWLEDGMENTS
The authors thank Dr. M. Barac-Nieto for criticism and suggestions.
【参考文献】
Alcalde AI, Sarasa M, Raldúa D, Aramayona J, Morales R, Biber J, Murer H, Levi M, Sorribas V. Role of thyroid hormone in regulation of renal phosphate transport in young and aged rats. Endocrinology 140: 1544-1551, 1999.
Anzai N, Miyazaki H, Noshiro R, Khamdang S, Chairoungdua A, Shin HJ, Enomoto A, Sakamoto S, Hirata T, Tomita K, Kanai Y, Endou H. The multivalent PDZ domain-containing protein PDZK1 regulates transport activity of renal urate-anion exchanger URAT1 via its C terminus. J Biol Chem 279: 45942-45950, 2004.
Bacic D, Capuano P, Baum M, Zhang J, Stange G, Biber J, Kaissling B, Moe OW, Wagner CA, Murer H. Activation of dopamine D1-like receptors induces acute internalization of the renal Na + /phosphate cotransporter NaPi-IIa in mouse kidney and OK cells. Am J Physiol Renal Physiol 288: F740-F747, 2005.
Biber J, Gisler SM, Hernando N, Wagner CA, Murer H. PDZ interactions and proximal tubular phosphate reabsorption. Am J Physiol Renal Physiol 287: F871-F875, 2004.
Biber J, Gisler SM, Hernando N, Murer H. Protein/protein interactions (PDZ) in proximal tubules. J Membr Biol 203: 111-118, 2005.
Blasco T, Aramayona JJ, Alcalde AI, Catalán J, Sarasa M, Sorribas V. Rat kidney MAP17 induces cotransport of Na-mannose and Na-glucose in Xenopus laevis oocytes. Am J Physiol Renal Physiol 285: F799-F810, 2003.
Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248-254, 1976.
Breusegem SY, Halaihel N, Inoue M, Zajicek H, Lederer E, Barry NP, Sorribas V, Levi M. Acute and chronic changes in cholesterol modulate Na-Pi cotransport activity in OK cells. Am J Physiol Renal Physiol 289: F154-F165, 2005.
Cheng J, Moyer BD, Milewski M, Loffing J, Ikedaq M, Mickle JE, Cutting GR, Li M, Stanton BA, Guggino WB. A Golgi-associated PDZ domain protein modulates cystic fibrosis transmembrane regulator plasma membrane expression. J Biol Chem 277: 3520-3529, 2002.
Cheng J, Wang H, Guggino WB. Regulation of cystic fibrosis transmembrane regulator trafficking and protein expression by a Rho family small GTPase TC10. J Biol Chem 280: 3731-3739, 2005.
Donowitz M, Milgram S, Murer H, Kurachi Y, Yun C, Weinman E. Coming out of the NHERF family. J Physiol 567: 1, 2005.
Gisler SM, Stagljar I, Traebert M, Bacic D, Biber J, Murer H. Interaction of the type IIa Na/Pi cotransporter with PDZ proteins. J Biol Chem 276: 9206-9213, 2001.
Gisler SM, Pribanic S, Bacic D, Forrer P, Gantenbein A, Sabourin LA, Tsuji A, Zhao ZS, Manser E, Biber J, Murer H. PDZK1. I. A major scaffolder in brush borders of proximal tubular cells. Kidney Int 64: 1733-1745, 2003.
Hernando N, Déliot N, Gisler SM, Lederer E, Weinman E, Biber J, Murer H. PDZ-domain interactions and apical expression of type IIa Na/P(i) cotransporters. Proc Natl Acad Sci USA 99: 11962-11975, 2002.
Hernando N, Gisler SM, Pribanic S, Déliot N, Capuano P, Wagner CA, Moe OW, Biber J, Murer H. NaPi-IIa and interacting partners. J Physiol 567: 21-26, 2005.
Kanamori M, Sandy P, Marzinotto S, Benetti R, Kai C, Hayashizaki Y, Schneider C, Suzuki H. The PDZ protein tax-interacting protein-1 inhibits beta-catenin transcriptional activity and growth of colorectal cancer cells. J Biol Chem 278: 38758-38764, 2003.
Kim E, Sheng M. PDZ domain proteins of synapses. Nat Rev Neurosci 5: 771-781, 2004.
Kocher O, Cheresh P, Brown LF, Lee SW. Identification of a novel gene, selectively up-regulated in human carcinomas, using the differential display technique. Clin Cancer Res 1: 1209-1215, 1995.
Kocher O, Comella N, Gilchrist A, Pal R, Tognazzi K, Brown LF, Knoll JHM. PDZK1, a novel PDZ domain-containing protein upregulated in carcinomas and mapped to chromosome 1q21, interacts with cMOAT (MRP2), the multidrug resistance-associated protein. Lab Invest 79: 1161-1170, 1999.
Lederer ED, Sohi SS, Mathiesen JM, Klein JB. Regulation of expression of type II sodium-phosphate cotransporters by protein kinases A and C. Am J Physiol Renal Physiol 275: F270-F277, 1998.
Levi M, Baird BM, Wilson PW. Cholesterol modulates rat renal brush border membrane phosphate transport. J Clin Invest 85: 231-237, 1990.
Levi M, Lötscher M, Sorribas V, Custer M, Arar M, Kaissling B, Biber J, Murer H. Cellular mechanisms of acute and chronic adaptation of rat renal P i transporter to alterations in dietary P i. Am J Physiol Renal Fluid Electrolyte Physiol 267: F900-F908, 1994.
Lötscher M, Kaissling B, Biber J, Murer H, Levi M. Role of microtubules in the rapid regulation of renal phosphate transport in response to acute alterations in dietary phosphate content. J Clin Invest 99: 1302-1312, 1997.
Magagnin S, Werner A, Markovich D, Sorribas V, Biber J, Murer H. Expression cloning of human and rat renal cortex Na/Pi cotransport. Proc Natl Acad Sci USA 90: 5979-5983, 1993.
Markovich D, Verri T, Sorribas V, Forgo J, Biber J, Murer H. Regulation of opossum kidney (OK) cell Na/Pi cotransport by Pi deprivation involves mRNA stability. Pflügers Arch 431: 266-271, 1995.
McWilliams RR, Breusegem SY, Brodsky KF, Kim E, Levi M, Doctor RB. Shank2E binds NaPi cotransporter at the apical membrane of proximal tubule cells. Am J Physiol Cell Physiol 289: C1042-C1051, 2005.
Morré DJ, Cheetham RD, Nyquist SE. A simplified procedure for isolation of golgi apparatus from rat liver. Prep Biochem Biotechnol 2: 61-69, 1972.
Pribanic S, Gisler SM, Bacic D, Madjdpour C, Hernando N, Sorribas V, Gantenbein A, Biber J, Murer H. Interactions of MAP17 with the NaPi-IIa/PDZK1 protein complex in renal proximal tubular cells. Am J Physiol Renal Physiol 285: F784-F791, 2003.
Silver DL, Wang N, Vogel S. Identification of small PDZK1-associated protein, DD96/MAP17, as a regulator of PDZK1 and plasma high density lipoprotein levels. J Biol Chem 278: 28528-28532, 2003.
Sorribas V, Markovich D, Hayes G, Stange G, Forgo J, Biber J, Murer H. Cloning of a Na/Pi cotransporter from opossum kidney cells. J Biol Chem 269: 6615-6621, 1994.
Sorribas V, Markovich D, Verri T, Biber J, Murer H. Thyroid hormone stimulation of Na/Pi-cotransport in opossum kidney cells. Pflügers Arch 431: 266-271, 1995.
Thomson RB, Wang T, Thomson BR, Tarrats L, Girardi A, Mentone S, Soleimani M, Kocher O, Aronson PS. Role of PDZK1 in membrane expression of renal brush border ion exchangers. Proc Natl Acad Sci USA 102: 13331-13336, 2005.
Wang S, Yue H, Derin RB, Guggino WB, Li M. Accessory protein facilitated CFTR-CFTR interaction, a molecular mechanism to potentiate the chloride channel activity. Cell 103: 169-179, 2000.
Weinman EJ, Steplock D, Wang Y, Shenolikar S. Characterization of a protein cofactor that mediates protein kinase A regulation of the renal brush-border membrane Na + -H + exchanger. J Clin Invest 95: 2143-2149, 1995.
Weinman EJ, Cunningham R, Wade JB, Shenolikar S. The role of NHERF-1 in the regulation of renal proximal tubule sodium-hydrogen exchanger 3 and sodium-dependent phosphate cotransporter 2a. J Physiol 567: 27-32, 2005.
Yun CH, Oh S, Zizak M, Steplock D, Tsao S, Tse CM, Weinman EJ, Donowitz M. cAMP-mediated inhibition of the epithelial brush border Na + /H + exchanger, NHE3, requires an associated regulatory protein. Proc Natl Acad Sci USA 94: 10006-10011, 1997.
作者单位：1 Laboratory of Molecular Toxicology, University of Zaragoza, Zaragoza, Spain; and 2 Departments of Medicine, Physiology and Biophysics, Division of Renal Diseases and Hypertension, University of Colorado Health Sciences Center, Denver, Colorado