The sodium-phosphate cotransporter 2a (NPT2a) is the principal phosphate transporter expressed in the brush border of renal proximal tubules and is downregulated by parathyroid hormone (PTH) through an endocytic mechanism. Apical membrane expression of NPT2a is dependent on interactions with the sodium-hydrogen exchanger regulatory factor 1 (NHERF-1). An LLC-PK1 renal cell line stably expressing the PTH receptor (PTH1R) and NHERF-1, termed B28-N1, fails to functionally express NPT2a. In B28-N1 cells, NHERF-1 and NPT2a are inappropriately localized to the cytoplasm. Ezrin, in the activated state, is capable at linking NHERF-1-assembled complexes to the actin cytoskeleton. Early-passage LLC-PK1 cells stably transfected with either empty vector or wild-type ezrin express a comparable level of the active, T567 phosphorylated form of ezrin and are capable of functionally expressing NPT2a. Colocalization of the PTH1R, NPT2a, and ezrin exists and is prominently associated with actin-containing microvilli in apical domains of these cells. Upon PTH treatment, the PTH1R, NPT2a, NHERF-1, and ezrin colocalize to endocytic vesicles and NPT2a-dependent phosphate uptake is markedly inhibited. LLC-PK1 cells expressing the constitutively active ezrin (T567D) display enhanced NPT2a functional expression and PTH-mediated regulation of phosphate. Expression of a dominant-negative ezrin, consisting of the NH2-terminal half of the protein, markedly disrupts NPT2a-dependent phosphate uptake. PTH does not appear to alter ezrin phosphorylation at T567. Instead, PTH perhaps initiates NPT2a endocytosis by inducing reorganization of the actin-containing microvilli in a process that is blocked by the actin-stabilizing compound jasplakinolide.
- proximal tubule
- hormonal regulation
the type 2a sodium-phosphate cotransporter (NPT2a) is the principal transporter expressed in the proximal tubule of the kidney and is responsible for the reabsorption of phosphate from the urine. When expressed in the membrane compartment, NPT2a is constitutively active. Hormonal regulation of phosphate reabsorption, therefore, occurs via compartmentalization of NPT2a to either the apical membrane or to intracellular vesicles (reviewed in Ref. 17). Parathyroid hormone (PTH) is a primary regulator of mineral ion homeostasis. To maintain free calcium ions in the blood, PTH induces phosphate wasting in the urine by promoting endocytosis of NPT2a from the brush-border membranes of the proximal tubule. The PTH receptor (PTH1R), a G protein-coupled receptor, directs NPT2a endocytosis principally through coupling to Gαs and activation of the cAMP/PKA pathway, but also via coupling to Gαq/11 and activation of the PLC/PKC pathway (10, 33, 34). The mechanisms through which these pathways evoke NPT2a endocytosis are poorly understood.
Clues as to possible regulatory mechanisms stem from the fact that the PTH1R and NPT2a share a common interaction with a scaffold protein called the sodium-hydrogen exchanger regulatory factor 1 (NHERF-1) (45). NHERF-1 contains two PDZ domain-binding motifs and directly interacts with the COOH termini of both the PTH1R (27) and NPT2a (22). In the opossum kidney cell line (OK), NHERF-1 is required for apical localization (22) and PTH-mediated regulation of NPT2a (26). NHERF-1 knockout mice display phosphate wasting and decreased expression of NPT2a in the brush border of kidney proximal tubules (38). Furthermore, reintroduction of NHERF-1 into primary cultures of proximal tubule cells isolated from NHERF-1-null mice promotes phosphate uptake and PTH-mediated regulation of NPT2a (13).
As noted above, OK cells readily display a robust reduction of phosphate uptake upon PTH treatment. However, many continuous kidney cell lines exist that fail to display PTH-mediated inhibition of phosphate uptake despite the presence of PTH-stimulated increases of cAMP (29). For example, the porcine kidney cell line, LLC-PK1, is a popular line for the analysis of PTH1R signaling because it lacks endogenous PTH receptors and thus allows for the introduction of mutant receptors through generation of stable lines (6). Paradoxically, certain LLC-PK1 lines expressing the PTH1R display an increase in phosphate uptake, a process that is dependent on PKC and presumably mediated by a transporter other than NPT2a (20). Quabius et al. (35) reported that LLC-PK1 cells do not endogenously express NPT2a. Attempts to exogenously express NPT2a in LLC-PK1 cells, however, resulted in a transporter that localized to intracellular sites and failed to promote the uptake of phosphate (35). These findings reveal that a membrane targeting and/or anchoring mechanism for NPT2a is absent in some LLC-PK1 cell lines, and thus suggests that the NHERF-1-assembled scaffold is defective.
Here, we show that ezrin, a membrane-actin cytoskeleton cross-linking protein, is required for the functional expression of NPT2a in LLC-PK1 cells and in doing so establishes the regulatory pathway for PTH-mediated inhibition of phosphate uptake.
MATERIALS AND METHODS
The primary antibody specific for NHERF-1 is from Abcam Limited (Cambridge, UK). NPT2a antibodies are from Alpha Diagnostics International (San Antonio, TX). Ezrin and phospho-ezrin (T567) antibodies are from Cell Signaling Technology (Beverly, MA). Secondary antibodies conjugated with horseradish peroxidase (HRP) are from Santa Cruz Biotechnology (Santa Cruz, CA). Secondary antibodies conjugated to Alexa Fluor 488 and 546, Alexa Fluor 647 phalloidin, and jasplakinolide are from Molecular Probes/Invitrogen (Carlsbad, CA). DNA ligase and restriction enzymes were from either New England Biolabs (Beverly, MA) or from Promega (Madison, WI).
DNA constructs and adenoviral vectors.
Replication-incompetent adenoviral vectors were generated using the ViraPower System from Invitrogen, following the manufacturer's procedures. Briefly, cDNAs were first cloned into the pENTR vector, followed by recombination into the adenoviral genome (pAd/CMV/V5-Dest) using LR Recombinase. The recombinant adenoviral genomes were transfected into 293A cells using FuGene 6 (Roche, Indianapolis, IN) to generate viruses for the transduction experiments.
The DNA construct expressing the PTH1R fused to YFP is as previously reported (28). Briefly, YFP was cloned in-frame using a native Not1 site located within the COOH-terminal tail of the rat PTH1R, which maintains a functional PDZ interaction motif located on the COOH terminus required for NHERF-1 interactions. Full-length human NPT2a (a gift from Dr. H. Juppner) was cloned into the pENTR vector using PCR with the proofreading PfuTurbo polymerase (Stratagene, La Jolla, CA). The NH2-terminal half of ezrin, corresponding to amino acids 1-297, was ligated into pENTR using PCR-mediated cloning. All sequences were confirmed using the DNA Core Facility at Massachusetts General Hospital.
Cell lines, transfection, and immunoblotting.
A complete description of all of the cell lines used is described in Table 1. All LLC-PK1-based cell lines were cultured in DMEM supplemented with 10% fetal bovine serum and antibiotics. Subcloning of the LLC-46 line and generation of the PTH1R stable B28 cell line are as previously described (6). For stable transfections, full-length NHERF-1 and ezrin-T567D (a gift from Dr. M. Arpin) were subcloned into pcDNA3.1/Hygro (Invitrogen) and transfected into B28 and LLC-46 cells, respectively, using FuGene 6. Several B28-N1 and LLC-46-TDEZ cell lines were cloned through selection using hygromycin (Calbiochem, San Diego, CA). Development of the LLC-3.1, LLC-WEZ, and LLC-TDEZ cell lines was as previously described (18). These cell lines are a generous gift from Dr. M. Arpin.
For the isolation of cellular extracts, cells were scraped in PBS and pelleted in a tube. For analysis of NHERF-1, cells were treated with whole cell extraction buffer comprising 25 mM HEPES (pH 7.4), 10% glycerol, 150 mM NaCl, 0.5% Triton X-100, and cocktails containing protease and phosphatase inhibitors from Sigma (St. Louis, MO). Ezrin was extracted from the particulate fraction. Briefly, cells were incubated in hypotonic buffer (25 mM HEPES, pH 7.4), 10% glycerol, and cocktails containing protease and phosphatase inhibitors for 20 min on ice. Cells were passed through a homogenizer (HGM Lab Equipment, Heidelberg, Germany) and nuclei were pelleted by low-speed centrifugation at 500 g. Supernatants were spun down at high-speed to obtain the membrane/particulate fractions. Ezrin was extracted from the particulate fraction using SDS-PAGE loading buffer. For the isolation of soluble and insoluble ezrin, pelleted cells were first lysed with the whole cell extraction buffer containing Triton X-100. Cellular debris was pelleted and the Triton X-100-soluble supernatant was collected for analysis. The remaining pellet was then sequentially extracted with whole cell extraction buffer containing 1% SDS. Samples were vortexed to reduce viscosity, centrifuged, and the Triton X-100-insoluble supernatant was collected for analysis.
For immunoblotting, proteins were separated by standard SDS-PAGE and blotted to PVDF membranes. Membranes were probed with specific antibodies in phosphate-buffered saline containing 0.1% Tween 20 and 5% nonfat dry milk. Visualization of the proteins was achieved with the use of secondary antibodies conjugated to HRP and detected by autoradiography with enhanced chemiluminescence using Western Lightning (Perkin-Elmer, Boston, MA).
Viral transduction and immunofluorescence analysis.
Cells were cultured in four-well chamber slides coated with collagen Type I (BD Biosciences, Bedford, MA) at an initial density of 80,000 cells per well. The early-passage LLC-PK1 cells were generally resistant to viral transduction, which is perhaps due to the polarized phenotype and tight cell junctions that exclude the virus from its receptor. To enhance transduction, cells were washed with Hank's balanced salt solution without calcium and magnesium and incubated with DMEM without calcium supplemented with 1% feta bovine serum for 1 h in the incubator. Cells generally rounded up without detaching from the plate and the resulting increase in surface area enhanced transduction. Adenoviruses were diluted in the calcium-free DMEM and added to cells in the range of 500 to 1,000 multiplicity of infection for 6 h. After the incubation period, the virus-containing media were aspirated from the cells and replaced with complete media. The cells were cultured an additional 72 h before analysis. Adenoviral transduction efficiencies among the LLC-PK1 cell lines were comparable, as determined by the use of an adenovirus expressing YFP alone.
Cells were fixed with 4% parafomaldehyde in PBS for 20 min and then thoroughly washed with PBS alone. Cells were then blocked and permeabilized with PBS containing 0.1% Triton X-100 and 5% nonfat dry milk for 30 min. Specific primary antibodies were added to the cells in the blocking solution at dilutions of 1:500 to 1:1,000 for 1 h followed by washing. Host-specific secondary antibodies conjugated to either Alex Fluor 488 (green) or 546 (red) were used to stain the antigens. Alexa Fluor 647 phalloidin (far red) was used to detect actin at a concentration of 2.5 μl/ml blocking solution. Immunostained cells were analyzed using a Radiance 2100 confocal microscope and the associated LaserSharp 2000 software (Bio-Rad, Hercules, CA). Images were cropped using Adobe Photoshop 7.
Sodium-phosphate cotransport assay.
The LLC-PK1 cell lines were split into 48-well plates at an initial density of 45,000 cells per well. The following day, cells were transduced with adenoviruses and cultured an additional 72 h before analysis. NPT2a-dependent phosphate uptake was determined by the difference in phosphate uptake between cells transduced with just the PTH1R and cells transduced with the PTH1R and NPT2a. The phosphate uptake assay is as described previously (26). All phosphate uptake data are presented as a mean percent of vehicle-treated cells ± SD (n = 6) and statistical analysis using the Student's t-test for at least three independent experiments.
B28 cells are a stable line expressing the PTH1R that display robust activation of adenylyl cyclase (39, 40), but fail to display PTH-mediated inhibition of phosphate uptake, a process that is readily demonstrated by opossum kidney cells (8, 11, 26, 33, 34). B28 cells were generated from a clonal line of wild-type LLC-PK1 cells called LLC-46. Compared with OK cells, B28 cells express low levels of NHERF-1 (Fig. 1A), suggesting that a deficiency of this scaffold protein is responsible for the absence of PTH-mediated inhibition of phosphate uptake. Several double-stable lines overexpressing NHERF-1 were generated and are referred to as B28-N1 (Fig. 1A). B28-N1 cells fail to display PTH-mediated inhibition of phosphate uptake (Fig. 1B), suggesting that NHERF-1 is not required or sufficient to establish this regulatory pathway. However, based on a report by Quabius et al. (35), LLC-PK1 cells do not endogenously express NPT2a. To express NPT2a, an adenoviral vector expressing the full-length cotransporter was generated. Despite the inclusion of NPT2a through adenoviral transduction, PTH promotes an increase of phosphate uptake in both the B28 and B28-N1 cell lines (Fig. 1B), which is consistent with findings reported by Guo et al. (20). Compared with nontransduced B28 cells, expression of NPT2a generated an insignificant increase in phosphate uptake (data not shown), revealing a lack of functional expression of the cotransporter. To assess localization, immunofluorescence analysis was performed. B28-N1 cells transduced with NPT2a display basolateral localization of the PTH1R, cytoplasmic localization of both NHERF-1 and NPT2a, and the absence of detectable actin-containing microvilli (Fig. 1C). Cytoplasmic staining of NPT2a and a lack of enhanced phosphate uptake in LLC-PK1 cells are consistent with findings reported by Quabius et al. (35). Attempts to stably express NPT2a in LLC-PK1 cells by Quabius et al. (35) resulted in the absence of phosphate transport. In these stable lines, NPT2a localized to the cytoplasm in a pattern that is remarkably similar to that demonstrated by NPT2a in the B28-N1 cell lines (Fig. 1C).
NHERF-1 is normally localized to the apical membrane compartments of both OK cells (26, 28) and in renal proximal tubules (43). NPT2a also localizes to apical domains through PDZ domain-mediated interactions with NHERF-1 in OK cells (22) and proximal tubules (38). Analysis of PTH1R function and expression patterns reveal that apical and basolateral subpopulations exist in OK cells (28, 37) and in renal proximal tubules (1, 2, 23). Analogous to NPT2a, targeting and/or anchoring of the PTH1R to apical membranes are likely mediated by interactions with NHERF-1 (28). Combined, NHERF-1 assembles a multiprotein complex at the apical membrane, which includes the PTH1R and NPT2a. Therefore, cytoplasmic localization of NPT2a and strict basolateral localization of the PTH1R suggest that the scaffolding function of NHERF-1 is defective in the B28 clonal line.
Evidence from the OK cell model reveals that the cytoskeleton is also a vital component of the NHERF-1-assembled complexes, as demonstrated by the prominent, actin-containing apical patches or microvilli displayed by this cell line (22, 28). The absence of actin-containing microvilli on the B28-N1 cell lines points to a defect in ezrin function. Ezrin is a membrane-cytoskeleton cross-linking protein and essential microvilli component belonging to the ERM (ezrin, radixin, moesin) family of proteins (4). In the activated state, ezrin is capable of simultaneously binding to NHERF-1 via an NH2-terminal domain and to actin via a COOH-terminal domain and thus indirectly links NHERF-assembled complexes to the cytoskeleton (36). Ezrin expression is not detectable in the B28 cell lines as determined by immunofluorescence analysis (data not shown). Conversely, early-passage wild-type LLC-PK1 cells and the same cells supplemented with ezrin, referred to as LLC-WEZ, display abundant levels of actin-containing microvilli that colocalize with ezrin (18). Using adenoviral-mediated expression, the PTH1R tagged with YFP displays prominent localization to microvilli in the apical compartment (Fig. 2A), an expression pattern that is evident in OK cells (28) and renal proximal tubules (1, 2). A second subpopulation of the PTH1R also localizes to the basolateral membrane (data not shown). Contrary to the B28 cell lines, NHERF-1 is abundantly expressed endogenously and colocalizes with the PTH1R in the apical compartment in LLC-WEZ cells (Fig. 2A) and in early-passage LLC-PK1 cells (data not shown). Presumably through direct binding to NHERF-1, ezrin colocalizes with the PTH1R in the apical membranes, thus providing an indirect interaction with the actin cytoskeleton (Fig. 2B). As noted above, NPT2a contains a consensus PDZ domain-binding motif and interacts with NHERF-1. Consistent with these findings, NPT2a expressed through adenoviral transduction colocalizes with the PTH1R and hence is a component of the NHERF-1/ezrin-assembled complex associated with apical microvilli (Fig. 2C). Similar NHERF-assembled complexes also exist in OK cells (26, 28). Treatment with PTH for 30 min induces internalization of the PTH1R as expected (Fig. 2A). Unexpectedly, PTH also promoted internalization of the entire NHERF-1-assembled complex, including ezrin and NPT2a (Fig. 2, B and C). Distinct colocalization with the ligand-bound PTH1R in submembranous vesicles reveals that the assembled complex remains intact at least during the early stages of internalization. Consistent with these findings, Cant and Pitcher (7) reported that ezrin localizes to ligand-induced endocytic vesicles containing β2-adrenergic and M1 muscarinic receptors. Importantly, hormone-induced endocytosis of NPT2a in the LLC-WEZ cell line suggests that the PTH-mediated phosphate regulatory pathway exists, a process that is completely absent in the B28 clonal lines.
As noted above, adenoviral-mediated expression of NPT2a in the B28 and B28-N1 cell lines fails to enhance phosphate uptake to any significant degree. The parental cell line for the B28 clones is a subclone of wild-type LLC-PK1 cells referred to as LLC-46. LLC-46 cells display the same expression patterns as the PTH1R-stable B28 cells transduced with the PTH1R, NHERF-1, and NPT2a (data not shown). Adenoviral transduction of NPT2a in the LLC-46 cells generates a 1.1-fold increase of phosphate uptake over basal (Fig. 3A), which is consistent with the lack of functional expression and cytoplasmic localization of the transporter in this cell line. In contrast, NPT2a expression in early-passage LLC-PK1 cells stably transfected with either empty vector (LLC-3.1) or ezrin (LLC-WEZ) yielded an equivalent increase of phosphate uptake of approximately threefold over basal (Fig. 3A), demonstrating functional expression of the cotransporter. Active phosphate uptake is consistent with the apical membrane expression patterns of NPT2a in these cell lines (data not shown and Fig. 2C). Equivalent functional expression of NPT2a in the vector only cells and the ezrin stable cells suggests that ezrin is not required for membrane localization of the cotransporter. However, these findings raise the possibility that perhaps the activation status of ezrin is the key element associated with NPT2a functional expression.
Ezrin assumes an inactive conformation through an intramolecular association between the NH2- and COOH-terminal regions of the protein resulting in the masking of interaction domains (5). Inactive ezrin is localized to the cytosol and is incapable of binding to either NHERF-1 or actin. Ezrin assumes an active, open conformation through PIP2 binding and phosphorylation of T567 (16, 18, 30). In the active state, ezrin is readily associated with actin-containing microvilli and NHERF-1. Mutating the threonine at 567 to aspartic acid generates a phosphorylation mimic, yielding a constitutively active ezrin protein when expressed in cells (18). Stable expression of the T567D mutant of ezrin into the early-passage LLC-PK1 cells (LLC-TDEZ) markedly enhances the functional expression of NPT2a upon adenoviral transduction to levels that are sixfold above basal phosphate uptake (Fig. 3A). Cells exposed to YFP alone adenovirus display comparable transduction efficiencies for all of the LLC-PK1-based cell lines (data not shown), demonstrating that the increased functional expression of NPT2a in the LLC-TDEZ cell line is not due to enhanced viral transduction. NPT2a expression in the LLC-TDEZ cells displays localization to qualitatively similar microvilli; however, the relative level of fluorescence intensity per cell in the apical domain appears greater in the LLC-TDEZ cells compared with the LLC-3.1 or LLC-WEZ cell lines (data not shown). Quantitative membrane expression of NPT2a in these cell lines using immunoblot analysis could not be done due to the ineffectiveness of the commercially available antibodies. Considering that NPT2a is constitutively active, however, analysis of phosphate uptake is a sensitive and quantitative method for assessing membrane expression. The T567D phosphorylation mimic of ezrin is not fully active in vitro (9); however, this mutant form promotes the formation of microvilli and other membrane structures when expressed in LLC-PK1 cells (18). This process may occur in conjunction with normal levels of phospho-T567 ezrin endogenously expressed in these cells.
To further corroborate the findings above, an adenoviral vector expressing a dominant-negative form of ezrin was generated. Expressing the NH2-terminal half of ezrin (EZNT) effectively disrupts the membrane-cytoskeleton cross-linking functions. When EZNT is stably expressed in early-passage LLC-PK1 cells, microvilli are markedly reduced (12). As shown in Fig. 3B, transduction of the adenoviral EZNT construct also effectively disrupts the formation of microvilli in LLC-WEZ cells. Increasing the dosages of the EZNT adenovirus relative to the NPT2a virus markedly decreases NPT2a-dependent phosphate uptake compared with equivalent dosages of an adenovirus expressing beta-glucuronidase (gus) in both the LLC-3.1 and LLC-WEZ cell lines (Fig. 3C). The same dosing regimens of the dominant-negative ezrin construct also block NPT2a-dependent phosphate uptake in the LLC-TDEZ cell line but to a lesser degree (Fig. 3C), an outcome that is likely due to the overexpression of the constitutively active form of ezrin. A similar dominant-negative ezrin construct expressed in OK cells inhibits NPT2a expression in the membrane/lipid raft compartment and prevents PTH-mediated regulation (32), thus providing strong evidence that ezrin is required for the functional expression of NPT2a in both cell models.
Analysis of ezrin's activation status can be achieved using antibodies that recognize the phoshorylation of T567. LLC-46 cells express low levels of total ezrin and almost no phospho-ezrin (Fig. 4A), which is consistent with the lack of microvilli and the inability to functionally express NPT2a. Conversely, LLC-3.1 and LLC-WEZ cells express high levels of phospho-ezrin (Fig. 4A), suggesting a correlation between the presence of activated ezrin and the functional expression of NPT2a. Despite the higher levels of total ezrin in the LLC-WEZ cells compared with the vector only cells, the levels of phospho-ezrin are comparable and thus explain the equivalent functional expression of NPT2a shown in Fig. 3A. These findings also suggest that ezrin activation via phosphorylation of T567 is rate limiting because expressing more ezrin does not result in a concomitant increase of the active form. Combined, these data demonstrate that the activated form of ezrin and the processes that result in the phosphorylation of T567 are essential for the functional expression of NPT2a in LLC-PK1 cells. Surprisingly, no quantitative difference in membrane levels of NHERF-1 is evident between the LLC-3.1, LLC-WEZ, or LLC-TDEZ cell lines, suggesting that NHERF-1 deficiency is not a limiting factor in the early-passage LLC-PK1 cells (data not shown).
As previously described, LLC-46-derived cell lines, such as B28, are NHERF-1 deficient (Fig. 1A) and that exogenous expression of NHERF-1 fails to promote functional expression of NPT2a on the surface of the cell (Fig. 1, B and C). Data shown in Fig. 4A reveal that LLC-46 cells are also ezrin deficient with virtually undetectable amounts of the phosphorylated, active form. Combined, the absence of the NHERF-1/ezrin scaffold complex may be responsible for the lack of NPT2a surface expression. To investigate this possibility, LLC-46 cells were stably transfected with ezrin-T567D. The phosphomimetic, constitutively active form of ezrin was expressed to account for the possibility that the activation process may be defective in LLC-46 cells. As shown in Fig. 4C, stable expression of ezrin-T567D (LLC-46-TDEZ) alone fails to enhance NPT2a functional expression compared with the vector only control cell line (LLC-46-3.1). LLC-46-TDEZ cells transduced with NHERF-1 adenovirus, however, display an increase in the functional expression of NPT2a, thus demonstrating that both scaffold components are necessary for phosphate transport (Fig. 4C). A similar phenomenon was reported by Li et al. (25), revealing that both NHERF-1 and ezrin promote channel activities of the cystic fibrosis transmembrane conductance regulator and that ezrin controls the formation of the macromolecular complex.
Now that we established functional expression of NPT2a in the LLC-PK1 model, the regulatory effect of PTH on phosphate uptake was examined. Analogous to the results shown in Fig. 1 for B28 cells, PTH1R and cotransporter-transduced LLC-46 cells display an increase in phosphate uptake in response to PTH, presumably via a phosphate transport mechanism not associated with NPT2a (Fig. 5). Conversely, PTH inhibits NPT2a-dependent phosphate uptake to a similar degree in both the LLC-3.1 and LLC-WEZ cell lines (Fig. 5). The extent of PTH-mediated inhibition of phosphate uptake is statistically greater (P < 0.05) in the LLC-TDEZ cell line, suggesting that the activated form of ezrin promotes the formation of a regulatory complex (Fig. 5).
The existence of PTH-mediated regulation of phosphate uptake in the presence of the T567 phosphorylation mimic (LLC-TDEZ cell line) suggests that inactivation of ezrin through dephosphorylation of this site is not a mechanism associated with hormone-regulated NPT2a endocytosis. To examine this possibility further, immunofluorescence analysis using the phospho-ezrin antibodies was undertaken. Expectedly, the T567 phosphorylated, active form of ezrin distinctly localizes to the apical compartment associated with microvilli and colocalizes with the PTH1R complex (compare Fig. 2B with Fig. 6A). Upon PTH treatment, phospho-ezrin remains associated with the PTH1R within endocytic vesicles, a process that is similar to patterns displayed when using total ezrin antibodies (Fig. 6A). These findings demonstrate that at least qualitatively PTH does not alter the phosphorylation status of ezrin at position 567. Immunoblot analysis reveals that the active, phosphorylated form of ezrin exists in the Triton X-100-resistant fraction, which is consistent with cytoskeletal interactions (Fig. 6B). In the LLC-PK1 model, PTH treatment has no apparent effect on ezrin's phosphorylation status at T567 (Fig. 6B), which is consistent with the immunofluorescence analysis shown in Fig. 6A. However, PTH increases the Triton X-100 solubility of the active form of ezrin, suggesting that PTH signaling induces ezrin dissociation from the actin cytoskeleton via a mechanism that does not involve dephosphorylation at position 567. In the OK cell model, PTH also does not alter phosphorylation of ezrin at T567, as reported by Nashiki et al. (32). However, PTH-elicited changes in ezrin phosphorylation at other sites mediated by PKA and/or PKC may participate in mechanisms associated with NPT2a endocytosis (32).
Ezrin indirectly links NHERF-assembled complexes to the actin cytoskeleton. As shown in Fig. 7A, actin is a principal component of the apical complexes. Notably, the actin-containing microvilli appear to undergo reorganization upon treatment with PTH (Fig. 7A). PTH-mediated microvilli retraction has been shown in primary cultures of proximal tubule cells (19). PTH also induces retraction and cell shape changes in osteoblastic cell lines due to alterations in cytoskeletal structure (3, 31). A growing body of evidence strongly suggests that the endocytosis mediated by both caveolae and clathrin requires dynamic cytoskeletal reorganization because these processes are readily inhibited by compounds that either promote disassembly or assembly of actin polymers, such as latrunculin B and jasplakinolide, respectively (14, 24, 42, 47). These findings suggest that PTH targets the actin-based cytoskeleton leading to the endocytosis of NPT2a.
Targeting the actin cytoskeleton is a potential mechanism through which PTH initiates the endocytosis of NPT2a. As shown in Fig. 7B, actin polymer stabilization with jasplakinolide completely blocks the PTH-induced inhibition of phosphate uptake. Jasplakinolide also effectively inhibits PTH-mediated inhibition of the sodium-hydrogen exchanger 3 (NHE3) (41) and the cytoskeletal reorganization required for ligand-induced endocytosis of the thromboxane (24) and β-adrenergic receptors (42). This finding suggests that PTH initiates a pathway that results in actin reorganization and that alterations in cytoskeletal dynamics promote NPT2a endocytosis.
The phosphaturic actions of PTH are well-established; however, the mechanisms by which the PTH1R mediates NPT2a endocytosis are still poorly understood. In vitro, the bulk of knowledge has been generated from the OK cell model, which faithfully displays PTH-mediated regulation of phosphate uptake through endogenously expressed PTH1Rs and NPT2a (referred to as Na-Pi4 in the opossum). The use of pathway activators reveals PTH1R signaling through both cAMP/PKA and PLC/PKC direct regulation of NPT2a endocytosis (10, 33, 34). However, the downstream targets of these pathways are unknown. Analysis of protein-protein interactions using the yeast two-hybrid assay has revealed that the PTH1R (27) and NPT2a (22) possess COOH-terminal PDZ interaction motifs and share a common partner in NHERF-1. A subclone of the OK cell line, termed OKH, displays defective regulation of phosphate uptake and expresses low levels of NHERF-1. OKH cells replete with NHERF-1 display phenotypic characteristics that are indistinguishable from wild-type cells for both phosphate regulation (26) and PTH1R signaling via intracellular calcium (28). Furthermore, reintroduction of NHERF-1 into primary cultures of proximal tubule cells isolated from NHERF-1(−/−) mice restores both functional expression of NPT2a and PTH-mediated regulation of phosphate uptake (13). These findings reveal that NHERF-1 is a vital component for hormonal regulation of phosphate homeostasis. However, the salient finding in the current report demonstrates a critical role for ezrin and its ability to link NHERF-1-assembled complexes to the actin cytoskeleton.
The LLC-PK1 cell line is a popular model for the analysis of proximal tubule functions in vitro. Initial establishment of PTH1R stable LLC-PK1 cell lines by Bringhurst et al. (6) relied on the use of a clonal line termed LLC-46. LLC-46 cells upon stable transfection of the PTH1R display characteristic signaling pathways. However, LLC-46-based PTH1R stable lines, such as B28, display paradoxical increases of phosphate uptake in response to PTH. Conversely, the ability of LLC-PK1 cell lines derived from early-passage parents, such as the LLC-3.1, LLC-WEZ, and LLC-TDEZ, to fully display the PTH1R/NPT2a regulatory pathway highlights heterogeneity among LLC-PK1 subclones. In fact, early reports regarding LLC-PK1 cells demonstrate heterogeneity based on morphological characteristics related to the ability or inability to form dome-like structures in culture (46). Domes arise from cells grown on impermeable cell culture substrata and indicate vectorial, transcellular transport of fluid and solutes that accumulate between the cell monolayer and the culture dish. In contrast to the early passage-derived LLC-PK1 cell lines and OK cells, the LLC-46 clonal line does not display dome formation. Thus the capacity of a cell line to form domes in culture correlates with the ability to functionally express NPT2a. Furthermore, transient expression of the PTH1R and NPT2a using the adenoviral system greatly reduces variability due to clonal heterogeneity.
One potential discrepancy between the OK cell model and the LLC-PK1 model described herein exists. As shown in Fig. 2, a 30-min exposure to PTH induces NPT2a endocytosis in a complex that also contains the PTH1R, NHERF-1, and ezrin. In contrast, Deliot et al. (15) reported that PTH induces dissociation of NPT2a from NHERF-1 complexes in OK cells. This apparent discrepancy can be explained by the fact that the NPT2a/NHERF-1 dissociation was evident after 4 h of PTH treatment in the OK cell model. Therefore, it is possible that NPT2a/NHERF-1 complexes remain intact during early stages of endocytosis, followed by dissociation as NPT2a moves into the late endosomal compartments during the later stages of endocytosis. Furthermore, unlike the LLC-PK1 model described herein, PTH has no apparent effect on NHERF-1 localization in the brush border of the proximal tubule of the mouse kidney (15). NHERF-1 expression levels in the kidney have been estimated to be as high as 0.01% of the total protein (21). Considering the substantially lower level of endogenous transporter expression, only a very small fraction of NHERF-1 would undergo coendocytosis with NPT2a to levels that are possibly below the limit of immunological detection. Enhanced expression of NPT2a in the current model, in contrast, increases the relative amount of NHERF-1 present within the endocytic vesicles. In support of this contention, ezrin colocalizes to endocytic vesicles containing ligand-bound β2-adrenergic receptors presumably via interactions with NHERF-1, as reported by Cant and Pitcher (7).
In search of cell models that emulate renal phosphate handling, the OK cell line is the only model to display PTH responses coupled to both activation of adenylyl cyclase and inhibition of Na+-dependent phosphate uptake (29). Through exogenous expression of the PTH1R and NPT2a, we developed an LLC-PK1-based model, which has the added advantage of being able to manipulate protein components via mutational analysis. Several similarities with the OK cell model and our LLC-PK1 model exist. For instance, NHERF-1-assembled complexes consisting of the PTH1R, NPT2a, and the actin cytoskeleton colocalize to the apical domain. Ezrin's participation in this regulatory pathway is also essential for both models, as demonstrated herein and by Nashiki et al. (32) for OK cells. In summary, several lines of evidence demonstrate that activated ezrin is required for functional expression and PTH-mediated regulation of NPT2a: 1) LLC-PK1 clonal lines lacking activated ezrin do not functionally express NPT2a in the apical domain despite the presence of NHERF-1, 2) LLC-PK1 cells expressing a constitutively active form of ezrin (T567D) display enhanced levels of NPT2a activity and increased potency of PTH-mediated regulation of phosphate uptake, and 3) expression of the dominant-negative form of ezrin (EZNT) effectively disrupts NPT2a functional expression. The ability of ezrin to indirectly link NHERF-1-assembled complexes to actin may provide a mechanism(s) through which NPT2a is regulated because PTH appears to induce cytoskeletal reorganization. Since its initial cloning, many proteins have been shown to interact with NHERF-1 (44). Evidence presented herein demonstrates that the mere presence of NHERF-1 and a binding partner may not necessarily be sufficient for a full functional interaction to be evident and that activated ezrin and indirect links to the actin cytoskeleton may be required.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant P01-DK-11794.
I thank Dr. M. Arpin for the generous gift of supplying the LLC-PK1 cell lines and the technical assistance provided by J. Greenwood.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society