Florescent ligand-directed co-localization of the parathyroid hormone 1 receptor with the brush-border scaffold complex of the proximal tubule reveals hormone-dependent changes in ezrin immunoreactivity consistent with inactivation

Abstract

Through binding to parathyroid hormone (PTH), PTH1R interacts with kidney-specific scaffold proteins, including the sodium hydrogen exchanger regulatory factors 1 and 2 (NHERFs), and ezrin. To facilitate in vivo localization, tetramethylrhodamine-labeled PTH (PTH-TMR) was used as a florescent probe. In mice, PTH-TMR localizes to luminal surfaces of tubular S1 segments that overlap PTH1R immunostaining, but does not directly overlap with megalin-specific antibodies. PTH-TMR staining directly overlaps with Npt2a in nascent, endocytic vesicles, marking the location of transporter regulation. PKA substrate antibodies display marked staining increases in segments labeled with PTH-TMR, demonstrating a functional effect. In the presence of secondary hyperparathyroidism, PTH-TMR staining is markedly reduced and shifts to co-localizing with megalin. At 15 minutes post-injection, PTH-TMR-labeled vesicles do not co-localize with either NHERF or ezrin, suggesting PTH1R dissociation from the scaffold complex. At the 5 minute time point, PTH-TMR stains the base of microvilli where it localizes with both NHERF2 and ezrin, and only partially with NHERF1. Strikingly, the bulk of ezrin protein becomes undetectable with the polyclonal, CS3145 antibody, revealing a PTH-induced conformational change in the scaffold. A second ezrin antibody (3C12) is capable of detecting the altered ezrin protein. The CS3145 antibody only binds to the active form of ezrin and fails to recognize the inactive form, while the 3C12 reagent can detect either active or inactive ezrin. Here we show that the PTH1R is part of the ezrin scaffold complex and that acute actions of PTH suggest a rapid inactivation of ezrin in a spatially defined manner.

1. Introduction

The parathyroid hormone 1 receptor (PTH1R) directs a remarkably complex set of physiological processes [1]. At the root of this complexity is that two physiologically distinct peptide hormones activate the PTH1R, namely parathyroid hormone (PTH) and parathyroid hormone related protein (PTHrP). PTH, an endocrine factor, regulates mineral ion homeostasis by targeting bone and kidney. Conversely, PTHrP is an auto/paracrine factor that plays an important role during the development of many tissues [2]. The paradigm of only one receptor binding to two hormones combined with these diverse physiological responses likely requires secondary levels of protein-protein interactions beyond the receptor and ligand. Consistent with this paradigm, the PTH1R not only couples to every major heterotrimeric G protein sub-family, including G_s, G_q/11, G_i and G_12/13, but also direct interactions with many non-G proteins have been described [3].

As a primary regulator of mineral ion homeostasis, PTH directly targets the proximal convoluted tubules (PCT) of the kidney where it mediates the down-regulation of sodium-dependent phosphate co-transporters, NPT2a, NPT2c and PiT2, and enhances the expression of 25-hydroxyvitamin D3 1α-hydroxylase, leading to an increase in the bioactive form of vitamin D. Considering that PTH is a blood-borne endocrine factor, one would assume that the PTH1R would reside on basolateral surfaces of the proximal tubule. However, several groups have reported that the PTH1R is both physically and functionally located on the apical membranes of this target tissue in mouse kidney [4–7], characteristics that have also been demonstrated in well-polarized kidney cell models [8–10]. Notably, physical and functional interactions between the PTH1R and prominent brush-border-specific scaffold proteins have been described.

The carboxy-terminal tail of the PTH1R, which is located in the cytoplasm, contains two independent interaction domains that mediate binding to sodium-hydrogen exchange factors (NHERFs) and ezrin. NHERF1 and NHERF2 possess two PDZ interaction modules that direct binding to the C-terminal four amino acids of the PTH1R (ETVM) [11, 12]. Using various cell models, NHERFs promote PTH-mediated signaling via the PLC pathway [11, 12] and anchor the PTH1R to the membrane, interactions that influence receptor mobility and internalization [13–16]. NHERFs contain a C-terminal domain that also directs interactions with ERM (ezrin, radixin, moesin) proteins, such as ezrin. Ezrin is a membrane-actin cytoskeleton cross-linking scaffold protein, interactions mediated by an amino-terminal FERM domain and a conserved actin-binding module, respectively. Within the juxtamembrane region of the PTH1R C-terminal, a lysine-arginine-lysine (KRK) motif directs interactions with the FERM domain of ezrin in vitro and in cells [10]. Consistent with the binding mechanisms of other membrane proteins, such as CD44 [17], the PTH1R C-tail binds to the FERM domain of ezrin at a site that is mutually exclusive from the binding groove occupied by the C-terminus of NHERF1. At least in vitro, therefore, the PTH1R, ezrin and NHERF1 are capable of forming a ternary complex [10].

As noted above, the PTH1R appears to be predominantly expressed on apical surfaces of the proximal convoluted tubule (PCT) of the kidney. Importantly, both NHERF1 and ezrin are abundantly expressed and NHERF2 to a lesser extent on the apical, brush-border of the PCT [18, 19]. Consistent with the in vivo expression patterns, ezrin, NHERF1 and the PTH1R readily co-localize to the apical microvilli of the LLC-PK1 cell line, a well-polarized kidney cell model [10]. In LLC-PK1 cells, apical and basolateral PTH1R sub-populations exist at a 3 to 1 ratio, respectively. Notably, apical expression of the PTH1R is dependent upon the active form of ezrin [10]. The focus of the current research is to analyze the localization patterns of the PTH1R and the ezrin-NHERF scaffold complex in the PCT of the mouse kidney. Due to the poor quality of PTH1R antibodies, receptor localization is facilitated by the use of a fluorescently labeled PTH ligand. Here we show that upon subcutaneous injection of hormone, the ligand-bound PTH1R is spatially localized to the brush-border of the PCT where it directs Npt2a endocytosis, enhances PKA activity and likely induces dynamic regulation of the ezrin-NHERF scaffold complex.

2. Materials and Methods

2.1 Reagents

Human parathyroid hormone (1–34) was synthesized at the MGH Biopolymer Core facility, a gift from Dr. Gardella (Massachusetts General Hospital). The primary amine on the side chain of lysine 13 was labeled with tetramethylrhodamine to generate the florescent ligand probe (PTH-TMR). The polyclonal antibody to the PTH1R was from Covance (Princeton Township, NJ). The polyclonal ezrin antibody (CS3145) and the PKA substrate antibody (CS9621) were purchased from Cell Signaling Technology (Beverly, MA). The 3C12 monoclonal ezrin antibody, Alexa fluor 488-labeled phalloidin, and secondary antibodies labeled with Alexa fluor 488 and 546 were purchased from Life Technologies Corporation (Carlsbad, CA). Primary antibodies to panGβ (sc-378), Gα_s (sc-383), Gα_i3 (sc-262), Gα_q/11 (sc-392) and PLCβ1 (sc-205) and HRP-labeled secondary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal megalin antibodies were a generous gift from Dr. Daniel C. Biemesderfer, Yale School of Medicine. Antibodies directed towards NHERF1 and NHERF2 were generous gifts from Dr. Chris Yun, Emory University School of Medicine. The aquaporin 1 antibody was a gift from Dr. Dennis Brown, Massachusetts General Hospital. The primary Npt2a antibody was a gift from Dr. Jurg Biber, Univeristy of Zurich, Switzerland. Western Lightning enhanced chemiluminescence reagent was from Perkin-Elmer (Boston, MA). All general chemicals were purchased from Sigma (Saint Louis, MO).

2.2 Immunoflorescence analysis of mouse kidneys

Eight to ten week old C57BL/6 mice were purchased from Charles River Laboratories (Wilmington, MA) and housed in the MGH animal facility using accepted ethical guidelines. For experiments aimed at inducing secondary hyperparathyroidism, mice were fed a low-calcium diet (0.02% calcium from Harlan Teklad, Madison, WI) for ten days prior to analysis. PTH-TMR was injected subcutaneously at a dose of 40 nmol/kg. After the times indicated, mice were anaesthetized and the left ventricle perfused first with 10 to 15 mls of phosphate buffered saline (PBS), followed by perfusion with freshly prepared 4% formaldehyde solution (from paraformaldehyde) in PBS. Kidneys were cut on the sagittal plane and incubated in 4% formaldehyde for 1 to 2 hours on ice. Kidneys were washed in PBS and incubated overnight in 30% sucrose mixing end-over-end. Frozen sections (8 μ) were prepared by the Histo Core Facility at MGH using standard procedures.

Frozen sections were thawed and washed in PBS. In order to unmask epitopes, sections were treated with 0.4% SDS for 4 minutes depending on the antibody used, followed by washing in PBS. Samples were blocked using Super Block from Pierce (Rockford, IL) in PBS containing 0.1% Triton X-100 (PBST) for 30 minutes. Primary antibodies, typically at 1:200 dilutions, were incubated on the sections for 3 hours at room temperature, followed by washing in PBST. The appropriate species-specific secondary antibodies labeled with either Alexa fluor-488 or -546 were incubated with the sections for 2 hours at room temperature, followed by washing in PBST. Confocal micrscopic images were obtained using a Radiance 2000 confocal microscopy system (Bio-Rad Laboratories, Hercules, CA) using LaserSharp 2000. Images were imported into Adobe Photoshop software as TIFF files for cropping, merging and level adjustment.

2.3 Isolation of brush-border membranes

Kidney brush-border membranes were prepared as described [20]. Briefly, mice were fed either a normal or a calcium-deficient diet, as described above, for ten days prior to isolation. Cortices were rapidly isolated and homogenized with a Polytron homogenizer for 90 s in a buffer containing 300 mM mannitol, 25 mM HEPES, pH 7.4, 5 mM EGTA and a protease inhibitor cocktail (Sigma, Saint Louis, MO). While on ice, 1.3 volumes of water was added to the homogenate, followed by the addition of MgCl₂ to a final concentration of 12 mM. Samples were incubated on ice for 15 min with occasional mixing. Aggregated membranes were removed by a low-speed spin (3,000×g), followed by isolation of brush-border membranes via a high-speed spin (30,000×g) of the supernatant. Membrane pellets were resuspended in a HEPES buffer using a 28-guage needle, protein content was determined by the Bradford assay and 60 μg loaded onto an SDS-PAGE gel for PTH1R immunoblot analysis.

2.4 Recombinant ezrin production and analysis of activation

An adenoviral vector expressing a rhodopsin-tagged ezrin was generated, as previously described [10]. Early passage LLC-PK1 cells were seeded in 15 cm dishes and transduced with the ezrin-rho adenovirus. Forty-eight hours post-transduction, cells were extracted with a solution containing 25 mM HEPES, pH 7.4, 10% glycerol, 150 mM NaCl, 1.0% Triton X-100, 0.1% SDS and protease inhibitor cocktails (Sigma, Saint Louis, MO). Ezrin-rho was immunopurified using rho-tag-specific 1D4 antibodies (National Cell Culture Center, Minneapolis, MN) and protein A agarose (Roche, Indianapolis, IN), followed by elution with the 9 amino acid rho peptide (TETSQVAPA) at 10 μM for 1 hour at room temperature. Ezrin-rho was cleaved with recombinant calpain protease (EMD Biosciences, Darmstadt, Germany) in a HEPES buffer supplemented with 1 mM CaCl₂ for 30 minutes at room temperature. Uncleaved and cleaved ezrin were analyzed using standard SDS-PAGE and immunoblotting with 1:1000 dilutions of the CS3145 and 3C12 antibodies, followed by visualization using enhanced chemiluminescence.

For the analysis of ezrin activation-dependent antibody binding, enzymatic depletion of phosphatidylinositol-4,5-bisphosphate (PIP2), which inactivates intracellular ezrin, was employed. Generation of the membrane-anchored, phosphatidylinositol-4,5-phosphatase (PIPase) construct was as previously described [21]. HEK293 cells were sparsely seeded on gelatin-coated chamber slides. Cells were transiently transfected with ezrin-rho using FuGene 6 (Roche, Indianapolis, IN) in combination with either LacZ (control) or PIPase. Forty-eight hours post-transfection, cells were washed in PBS and fixed with 4% formaldehyde (from paraformaldehyde) in PBS. Cells were immunostained with either CS3145 or 3C12 ezrin antibodies and secondary antibodies labeled with Alexa fluor-488, followed by confocal microscopic analysis, as described above.

2.5 Data analysis

A minimum of three independent experiments was performed for each condition. Representative confocal microscopic images are presented.

3. Results

3.1 PTH-TMR marks the location of the PTH1R in PCTs

Localization of the PTH1R is hampered by the lack of quality antibodies capable of detecting receptors expressed in native tissues using immunoflorescence-based analyses. To overcome this limitation, PTH was labeled with the florescent dye, tetramethylrhodamine (PTH-TMR), which displays a remarkable level of sensitivity and specificity for the detection of the PTH1R in cell culture-based assays [21, 22]. Analysis of the PTH-TMR staining pattern in the mouse kidney was undertaken. As shown in Figure 1A, PTH-TMR readily localizes to the luminal surface of PCT within vesicle-like structures. PCT segments were easily identified by the thick, actin-rich brush-borders (Figure 1B). PTH-TMR staining is also evident in parietal cell layers of the glomeruli (Figure 1A). PTH-TMR-labeled segments were generally localized near glomeruli and did not co-localize with aquaporin1, which is predominantly expressed in S2/S3 segments (Figure 1B), demonstrating that S1 segments of the PCT are the primary binding sites for the ligand post subcutaneous injection.

Figure 1.

PTH-TMR labels S1 regions of the proximal tubules, co-localizes with PTH1R-specific antibodies and marks segments that display PTH1R mediated actions. A. Mice were injected with PTH-TMR (40 nmol/kg) for 5 minutes, followed by cardiac perfusion and analysis of frozen sections, as described in the Materials and Methods section. Confocal microscopic images of PTH-TMR (red; left panel) and phalloidin-stained actin (green; right panel) from the cortex are shown. B. A merged confocal image of aquaporin-1 (Aqp-1; green), which primarily labels S2 and S3 proximal tubule segments, and PTH-TMR (PTH; red) is shown. C. Mice were injected with PTH-TMR for 15 minutes, followed by immunoflorescent analysis of megalin expression. Confocal images displaying localization patterns of megalin (green; left panel), PTH-TMR (red; middle panel) and the merge of the two (right panel) are shown. D. Five minutes post-ligand injection, proximal tubule images displaying immunoflorescent analysis of PTH1R localization (green; left panel), PTH-TMR staining (red; middle panel) and the merge of the two (right panel) are shown. E. Five minutes post-ligand injection, localization patterns of Npt2a (green; left panel), PTH-TMR (red; middle panel) and the merge of the two (right panel) are shown. In the lower left panel, Npt2a staining in the absence of PTH-TMR labeling on the same kidney section is shown. F. Localization of substrates phosphorylated by PKA (PKA-sub; green) and PTH-TMR (PTH; red) is shown in this merged image of the kidney cortex five minutes post-ligand injection.

Although PTH-TMR staining on the luminal surfaces is consistent with previous studies localizing the PTH1R to the PCT brush-border, as described in the Introduction, experiments were performed to verify that the florescent ligand was in fact marking the location of the receptor. Megalin or LRP2 is a multi-ligand, endocytotic receptor that binds to PTH with micromolar affinity in vitro [23]. Although some general overlapping staining exists between megalin and PTH-TMR, the unique, internalized vesicular staining pattern generated by the ligand does not co-localize with the megalin antibody (Figure 1C), indicating that the PTH-TMR is likely labeling the PTH1R. Although PTH-TMR co-endocytosis with megalin cannot be fully ruled out for all vesicles. Consistent with this data, immunostaining with a commercially available PTH1R antibody and PTH-TMR co-localize to the apical membranes of the PCT (Figure 1D), albeit with a high degree of background staining from the antibody.

Additional evidence in support of PTH-TMR/PTH1R co-localization would be to demonstrate a correlative functional effect. As noted earlier, PTH mediates the acute down regulation of the type IIa sodium-phosphate co-transporter (NPT2a) in the PCT. During the initial stages of this process, PTH-TMR directly co-localizes with NPT2a at the base of the microvilli (Figure 1E), revealing that the PTH1R co-internalizes with the regulated transporter. Furthermore, the PTH1R readily couples to the stimulatory G protein (G_s), leading to the robust generation of cAMP and activation of PKA. As shown in Figure 1F, PTH-TMR-labeled tubular segments display a marked increase in staining with PKA substrate antibodies when compared to segments lacking the ligand, demonstrating a strong correlation between the presence of ligand and receptor activation.

Further experiments were done to verify that PTH-TMR is marking the PTH1R. When mice are placed on a calcium-deficient diet, circulating levels of PTH markedly increase, an adaptive process that leads to the maintenance of blood calcium levels through bone resorption and enhanced tubular reabsorption in the kidney. In the hyperparathryoid state, PTH-TMR labeling is markedly reduced and limited primarily to the parietal cell layer of Bowman’s capsule (Figure 2A). General megalin staining is largely unaffected by the low calcium diet when compared to mice fed a normal diet (Figure 2A). Notably, PTH-TMR labeling of the parietal cell layer and immediate tubular sections display co-localization with megalin (Figure 2B). Furthermore, PTH1R expression within brush-border membranes is markedly decreased in mice fed the low calcium diet when compared to normal diets, suggesting that the receptor is chronically internalized in the hyperparathyroid state (Figure 2C) and thus unable to bind to the PTH-TMR probe.

Figure 2.

Secondary hyperparathyroidism induced by a low-calcium diet down regulates the brush-border PTH1R and promotes PTH-TMR/megalin co-localization. A. Mice, fed either a normal diet (Norm Ca; left panels) or a low-calcium diet (Low Ca; right panels) for ten days, were subcutaneously injected with PTH-TMR for 5 min prior to cardiac perfusion and fixation. Confocal images of PTH-TMR labeling (upper panels) and megalin immunostaining (lower panels) of kidney cortices are shown. B. Confocal microscopic images of megalin immunostaining (green; left panel), PTH-TMR labeling (red; middle panel) and merged (left panel) from mice fed a low calcium diet are shown. C. Kidney brush-border membranes were isolated from mice fed the normal or low calcium diets and analyzed by immunoblotting with PTH1R-specific antibodies, as indicated. The arrow designates the location of the PTH1R on the blot.

As shown in Figure 3, many components of the heterotrimeric G protein-based signaling machinery are expressed on the luminal surfaces of the PCT, including Gβ, Gα_s, Gα_i, Gq/11 and PLCβ1. These findings demonstrate that essential components of GPCR signaling are present in the brush-border compartment of the PCT, overlapping PTH-TMR binding. It is unclear, however, if these components are directly associated with the scaffold complex. Combined with the data noted above, these findings strongly indicate that PTH-TMR is binding to and marking the location of the apically expressed PTH1R within the PCT and ligand in the primary filtrate is functionally active.

Figure 3.

Various effectors of G protein coupled receptors are prominently expressed in the brush-border membranes of proximal tubules. Immunoflorescent analyses of (A) panGβ, (B) Gα_s, (C) Gα_i3, (D) Gα_q/11 and (E) PLCβ1 in proximal tubule segments using confocal microscopy are shown.

3.2 Localization of NHERF1, NHERF2 and ezrin in the kidney cortex

With this new analytical tool, PTH1R co-localization with the ezrin-NHERF-based scaffold complex was assessed. First, immunological expression analysis of these scaffold proteins was examined using the available antibody reagents in the cortex of the mouse kidney. Ezrin and NHERF1 are predominantly expressed in the brush-borders of the PCT, as shown in Figures 4A and 4B, respectively. Ezrin is also expressed in the podocytes of the glomerulus, albeit to a lesser extent (data not shown). NHERF2 is abundantly expressed in the podocytes of the glomerulus and in the endothelial cells of the peritubular capillaries and major blood vessels, and to a lesser extent in the brush-border of the PCT (Figure 4C). As shown in Figure 4D, expression patterns for ezrin and NHERF1 directly overlap in the brush-border of the PCT. Conversely, NHERF2 appears to be expressed at the base of the PCT microvilli and only partially co-localizes with ezrin (Figure 4E). In general, these localization patterns are consistent with previously published findings [18, 19].

Figure 4.

Localization patterns of the ezrin-NHERF scaffold complexes in the proximal tubules of the kidney. Confocal microscopic images of immunoflorescent localization analysis of (A) ezrin (CS3145 polyclonal), (B) NHERF1 and (C) NHERF2 in the kidney cortex are shown. Co-localizations of (D) NHERF1 (green; left panel) and (E) NHERF2 (green; left panel) with ezrin (3C12 monoclonal, red; middle panels) and the merged images (right panels) are shown.

3.3 Co-localization of PTH1R/PTH-TMR with the ezrin-NHERF scaffold

As previously mentioned, both NHERF1 and 2 interact with the PTH1R in vitro and within cell models via a PDZ domain-mediated process. To assess these interactions in vivo, localization patterns of the PTH-TMR/PTH1R with the NHERF scaffold proteins in the PCT were analyzed. Two-time points were used post PTH-TMR injection at 5 and 15 minutes. A caveat exists in that these time points are relative due to the post-injection processing of the mice, including the process of cardiac perfusion with PBS and formaldehyde, which adds an additional 10 to 15 minutes to the time course prior to arrest via fixation. At the 5 minute time point, PTH-TMR appears in vesicle-like structures lined up close to the membrane, while at 15 minutes PTH-TMR labeled vesicles are generally further from the membrane in a more scattered pattern (Figure 5A). We interpret these patterns to represent early and latter stages of PTH1R endocytosis. At the 5-minute time point, PTH-TMR-labeled vesicles partially co-localize with NHERF1 at the innermost portion of the stain, while at 15 minutes the internalized vesicles do not co-localize with NHERF1 (Figure 5A). Regarding NHERF2, PTH-TMR co-localizes with this scaffold protein in a region that likely represents the base of the microvilli (Figure 5B). Similar to NHERF1, internalized vesicles do not readily co-localize with NHERF2 in the latter, 15-minute time point (Figure 5B). PTH-TMR co-localization with ezrin was then undertaken.

Figure 5.

Co-localization of PTH-TMR and NHERF proteins in the PCT. Mice were injected with PTH-TMR for either 5 (5m) or 15 (15m) minutes, as indicated. PTH-TMR staining (red; middle panels) and localization patterns for either (A) NHERF1 or (B) NHERF2 (green; left panels), as indicated, were analyzed using immunoflorescence techniques and confocal microscopy within labeled proximal tubular segments. Merged images are shown in right panels.

As previously mentioned, the PTH1R interacts with ezrin through a juxtamembrane-binding domain on the C-terminal tail of the receptor. Analysis of the PTH-TMR/ezrin co-localization patterns at the 5 and 15-minute time points revealed an unexpected effect. At 5 minutes, PTH-TMR-labeled vesicles directly co-localize with a small fraction of the ezrin protein (Figure 6A); however, the bulk of the ezrin protein appears absent. As shown in the same microscopic field of Figure 6A, marked reductions of total ezrin staining is evident in PTH-TMR-stained PCT segments, in contrast, those segment lacking the ligand display abundant levels of immunostained ezrin. At 15 minutes, the PTH-TMR-labeled vesicles have completely dissociated from ezrin and the bulk of the ezrin protein is now detectable (Figure 6B). A possible interpretation of these results is that PTH induces a conformational change that has rendered the bulk of ezrin undetectable by the polyclonal, CS3145 antibody and that this effect is rapidly reversed by the 15-minute time point. To investigate this possibility, a second ezrin antibody was used. As shown in Figure 6B, a monoclonal ezrin antibody (3C12) readily detects the bulk of the ezrin protein in PTH-TMR-labeled segments at the 5-minute time point, confirming that ezrin is still present in the same PCT segments and that this antibody can detect ezrin regardless of the PTH-elicited alternations in immunoreactivity. Staining of the basement membrane, shown in Figure 6B, occurs in the absence of the primary antibody and is due to the use of anti-mouse IgG secondary antibodies in the same species.

Figure 6.

Time-dependent co-localizations between PTH-TMR and ezrin reveal possible ligand-induced conformational changes. Ezrin localization patterns using (A and B) polyclonal CS3145 antibodies or (B) monoclonal 3C12 antibodies (green; left panels), as indicated, were analyzed in tubular segments stained with PTH-TMR (red; middle panels) for the times indicated, 5 or 15 minutes (5m or 15m). Confocal microscopic images and merged images (right panels) are shown.

Through intramolecular interactions between amino- and carboxy-terminal domains, ezrin can assume a closed, inactive confirmation, yielding a masked FERM domain incapable of binding to membrane-bound proteins [24]. A combination of phosphorylation at Thr567 and binding to phosphatidylinositol-4,5-bisphosphate (PIP2), generates an open, active form of ezrin [25, 26]. To identify the general location of the epitope used to develop the polyclonal, CS3145 antibody, recombinant ezrin was purified and cleaved with the calpain protease, which generates two fragments of 55 and 15 kDa (Figure 7A) [27]. As shown in Figure 7B, CS3145 antibodies recognize the 55 kDa fragment, revealing an epitope that is likely within the FERM domain. The monoclonal antibody (3C12) detects the 15 kDa fragment, confirming the presence of a C-terminal epitope (Figure 7B). Based on the hypothesis put forth from data in Figure 6, we predict that the polyclonal, CS3145 antibody can only recognize the active, open conformation of ezrin and that the monoclonal, 3C12 antibody can bind to both active and inactive conformations. When transiently transfected into HEK293 cells, ezrin is readily expressed to distinct, microvilli-like structures [21], a localization pattern consistent with the active conformation of this scaffold protein. Expectedly, both antibodies readily detect this active form of ezrin when co-expressed with the LacZ control protein, as shown in Figures 7C and 7D. Hydrolysis of PIP2 using a phosphatidylinositol-4,5-phosphatase (PIPase) inactivates ezrin when co-expressed in these cells [21]. As shown in Figure 7C, co-transfection of ezrin with PIPase markedly reduces immunoflorescent detection using the CS3145 antibody. In contrast, the 3C12 antibody readily detects the cytoplasmic, inactive form of ezrin when co-expressed with the PIPase enzyme (Figure 7D), as demonstrated by the presence of nuclear shadows in the image.

Figure 7.

The ezrin polyclonal antibody appears to detect only the active conformation, while the monoclonal antibody binds to ezrin regardless of the activation status. A. Schematic depicting the calpain cleavage site on ezrin, which yields 55 and 15 kDa proteolytic fragments. B. Ezrin containing a C-terminal rho-tag was expressed and purified as described in the Materials and Methods section. Ezrin was treated in the absence or presence of calpain protease, as indicated, for 30 minutes at room temperature, followed by immunoblot analysis using either the polyclonal (CS3145; left panel) or the monoclonal (3C12; right panel) antibodies. Representative immunoblots of the ezrin epitopes are shown. C and D. HEK cells were transiently transfected with ezrin and either LacZ (left panels) or PIPase (right panels), as indicated. Confocal microscopic images of ezrin expression patterns using either the polyclonal CS3145 (C) or the monoclonal 3C12 (D) antibodies are shown.

4. Discussion

It is well established that scaffold proteins assemble multi-faceted complexes that facilitate physiological processes by limiting diffusion of essential components [28–30]. PTH acutely induces hypophosphatemia through regulated endocytosis of NPT2a, thus providing a model process to study non-genomic actions of this hormone. Importantly, both the PTH1R and NPT2a appear to be components of the ezrin-NHERF-based scaffold located on the brush-border of the PCT, thus identifying a likely signaling complex directed at regulating the reabsorption of phosphate ions in vivo. To date, most reports examining PTH1R/scaffold interactions have relied on the use of in vitro models. With the aid of the PTH-TMR probe, the current research analyzes the in vivo assembly of these complexes in the proximal tubule of the mouse kidney.

Due to the general poor quality of PTH1R antibodies, effective analysis of receptor localization in native tissues has been limited. We believe that the florescent PTH-TMR probe is an effective, novel tool to investigate PTH1R dynamics for several reasons. First of all, PTH-TMR binding displays an impressive degree of specificity when using cell culture models in the absence or presence of the PTH1R [21, 22]. Second, PTH-TMR luminal staining of the PCT is consistent with several reports that demonstrate apical localization of the PTH1R [4–7, 23]. Most of these reports have used the Covance PTH1R antibody. Although the background is high for this antibody in our hands, the unique patterns generated by the PTH-TMR-labeled vesicles clearly overlap the immunostaining pattern displayed by this antibody, indicating that this ligand probe likely marks the receptor (Figure 1). Furthermore, the clathrin adapter, AP1B, binds to the PTH1R and directs basolateral expression in polarized cell models; however, the PCT does not express this adapter, resulting in apical expression of the receptor, as demonstrated in models of MDCK cells [31]. This finding and apical localization of many components associated with GPCR signaling, shown herein and by others [32–34], lend credence to the assertion that PTH within the primary filtrate activates apical receptors. Third, megalin, a scavenger receptor that binds PTH with micromolar affinity [23], fails to readily co-localize with PTH-TMR-labeled, cytoplasmic vesicles. Consistent with this finding, Bacic et al [35] failed to detect any PTH-induced megalin endocytosis in mouse PCT upon injection of unlabeled ligand. Lastly, early PCT segments labeled with PTH-TMR display functional effects that are consistent with PTH1R activation, such as marked increases in staining of PKA-substrate antibodies and robust, acute internalization of NPT2a. Notably, Picard et al. [36] reported that PTH-elicited NPT2a internalization is strongest in the S1 segments of the PCT, directly overlapping the primary location of PTH-TMR binding. Combined, these findings reveal that acute effects of a bolus PTH injection in the kidney occur immediately after filtration through the glomeruli, accessing apical PTH1R sub-populations.

Despite the obvious utility, limitations for the use of the PTH-TMR probe exist. Upon cloning of the PTH1R gene, probing of standard multi-tissue Northern blots revealed that the receptor is markedly expressed in the kidney when compared to the low level expression found in most tissues [37]. This finding, combined with the PCT as a primary target, supports the contention that the PTH1R is expressed to a level in this tissue that allows for detection using the PTH-TMR ligand. Conversely, the absence of staining is likely due to the fact that PTH1R expression levels in other regions of the kidney or other tissues, such as bone (unpublished observations), are probably below the detection limit of the PTH-TMR ligand, thus limiting the probe’s usefulness to the PCT. In other words, the lack of PTH-TMR staining does not necessarily indicate the absence of receptor expression. Lastly, it is conceivable that the PTH-TMR ligand is not directly marking the PTH1R, but to some unknown receptor. If true, then it is possible that this probe is still an indicator of those segments expressing the PTH1R due to the functional correlation demonstrated with the PKA-substrate antibodies, NPT2a internalization and apparent conformational changes in ezrin.

The existence of two PTH receptors, the PTH1R and megalin, within the PCT puts forth a compelling paradigm, one high affinity-low capacity and one low affinity-high capacity, respectively. Stressing mice with a low calcium diet reveals markedly reduced PTH-TMR labeling, an effect that correlates with a loss of PTH1R protein in brush-border membranes and no apparent effect on overall megalin expression. However, it is also possible that chronic PTH stimulation may induce changes in the PTH1R that lower ligand-binding affinities; thus reducing PTH-TMR staining. Notably, PTH-TMR labeling that does exist on the low calcium diet readily co-localizes with megalin in the parietal cell layer and very early PCT segments. Upon receptor internalization in the hyperparathyroid hormone state, PTH-TMR binds to megalin, but only immediately after entering the tubules upon filtration through the glomeruli, which suggests that the initial concentration of ligand is high enough to interact with this scavenger. Under normal conditions, however, dilution of PTH-TMR as it transits down the tubule maintains a high enough concentration capable of binding to the high affinity PTH1R but not the low affinity megalin receptor. These findings strongly suggest that PTH-TMR is labeling the PTH1R primarily with a lessor amount binding to megalin.

Through an atypical, C-terminal PDZ interaction motif, the PTH1R readily interacts with both NHERF1 and 2 in vitro, in cells and in kidney extracts, as determined by co-immunoprecipitation [11, 12]. Within in the PCT, NHERF1 is abundantly expressed within the microvilli, whereas NHERF2 is predominantly localized to the base of the microvilli, patterns that are consistent with previously published findings [18]. In the early time points, PTH-TMR binding localizes to the base of the microvilli primarily overlapping NHERF2; however, one cannot rule out the possibility that ligand initiates a rapid relocalization from the microvilli to the base. In general, the current findings reveal that as endocytosis proceeds, the PTH1R likely dissociates from the apical scaffold complex, which is consistent with data published by Friedman and co-workers using transformed cell lines [38]. Furthermore, our data are consistent with those of Weinman and co-workers [39], revealing that NHERF1 dissociates from both NPT2a and possibly the PTH1R upon endocytosis via a process that likely involves phosphorylation of NHERF1.

From our earlier studies [10, 21], It was fully anticipated that the PTH-TMR-labeled luminal surface of the PCT would readily localize with ezrin, a scaffold that exists in the brush-border of this segment. It is well established that ERM proteins, such as ezrin, possess intrinsic, auto-regulatory activation mechanisms mediated through conformational changes induced by intramolecular interactions between the amino-terminal FERM domain and a carboxy-terminal domain [24, 40]. Our findings reveal that the CS3145 antibody, which possesses an amino-terminal epitope, is unable to recognize the inactive, closed form of ezrin, a mechanism that is likely due to epitope masking by the carboxy-terminal domain upon fixation. Therefore, the inability of this antibody to detect the bulk of ezrin in PTH-TMR stained segments strongly suggests that this scaffold is a target for downstream signaling processes mediated by the PTH1R. In contrast, the carboxy-terminal epitope of the monoclonal, 3C12 antibody is not masked and thus capable of detecting both active and inactive forms of ezrin. Consequently, the 3C12 antibody readily detects the bulk of ezrin at the 5-minute time point, as well as ezrin sub-population associated with the PTH-TMR probe. Notably, these findings strongly suggest that the PTH1R, marked with the ligand probe, interacts with ezrin and that PTH signaling induces acute conformational changes in the bulk of ezrin.

With respect to the apparent PTH-induced conformational changes in ezrin, two key findings exist. First, the abundant levels of ezrin within the brush-border certainly far outweigh expression levels of the PTH1R. Therefore, one would assume that only a small fraction of ezrin would co-localize with the PTH1R. This contention is supported by the fact that low levels of ezrin readily co-localize with the PTH-TMR probe during the initial stages of endocytosis. Considering the high background of the receptor-specific antibodies, co-immunolocalization of these low levels of ezrin, using the monoclonal antibodies, and the PTH1R would be difficult. This distinct co-localization also reveals that PTH-TMR does not interfere with ezrin immunolocalization. Furthermore, the ability to detect PTH-TMR co-localization with both ezrin antibodies within the nascent vesicles suggests that ezrin is in the active, open conformation when bound to the PTH1R. This assertion is consistent with our previous findings demonstrating that the PTH1R C-tail binding to the FERM domain is inhibited by the carboxy-terminal half of ezrin [10], revealing preferential binding to the open, active form. Other established ezrin binding partners also preferentially interact with the active form [41]. Second, the apparent PTH-elicited changes in the conformation of the bulk of ezrin likely arise from PTH1R activation and the inherent signal amplification associated with second messenger production and stimulation of downstream effector enzymes, such as kinases, that possess high levels of processivity and substrate turnover.

Both cAMP/PKA and PLC/PKC are activated by the PTH1R, pathways that likely target the ezrin scaffold complex. With the use of kinase-specific substrate antibodies, Nashiki et al [42] demonstrated that PTH invokes both PKA- and PKC-dependent phosphorylation of ezrin in opossum kidney cells. Yao and coworkers [43] reported that PKA-mediated phosphorylation of ezrin on Ser66 mediates apical cytoskeleton remodeling associated with gastric parietal cell activation. Furthermore, ezrin has been shown to be an AKAP (A-kinase anchoring protein) [44] and AKAP79 directly interacts with NTP2a in OK cells [45], thus providing additional levels of regulation that is possibly mediated by the scaffold complex. Transient conformational changes in ezrin could also arise from local depletion of PIP2 via activation of PLCβ. It is well established that PIP2 is required to maintain the active form of ezrin [26, 46, 47], as confirmed by the effect of the PIPase shown in Figure 7. Certainly, PTH-elicited depletion of PIP2 is expected to be far less extensive than the membrane-anchored PIPase, an effect that is consistent with the transient nature of the conformational change in ezrin demonstrated by the PTH-TMR ligand. Considering that ezrin is an intrinsic component of brush-border membranes, the ability of PTH to modify cytoskeletal structures in primary proximal tubule cells [48], LLC-PK1 cells [49] and in osteoblasts [50] establishes a possible functional link to the dynamic regulation of this scaffold protein.

The ability to readily identify PTH1R localization using the PTH-TMR probe is a novel technique that will allow for the better understanding of receptor function within the PCT target tissue, especially signals emitted from the apical sub-population. With this novel tool in hand, it is readily apparent that following an injection of a bolus of hormone, PTH is readily filtered through the glomeruli and accesses receptors located on the luminal surface of the PCT. Furthermore, the brush-border membrane compartment not only provides the ezrin-based scaffold complex, but also contains a full complement of G proteins and related effector molecules that establish a fully functional PTH/PTH1R signalsome. Future directions include analyzing the nature of the PTH-induced conformational change in ezrin and possible mechanisms associated with the down regulation of NPT2a.

5. Conclusions

1. The florescent PTH-TMR probe, upon subcutaneous injection, labels the luminal surface of early (S1) segments of the kidney proximal convoluted tubule that co-localize with PTH1R-specific antibodies but not readily with megalin antibodies, thus marking the PTH1R.

2. Activation of PKA and internalization of NPT2a are evident functional effects that correspond to PTH-TMR labeled segments, demonstrating a fully active PTH1R.

3. Heterotrimeric G proteins and effector molecules associated with GPCRs distinctly localize to the brush-border surface of the proximal tubule.

4. During the initial stages of ligand binding, PTH-TMR/PTH1R co-localizes with a sub-population of ezrin and mediates a conformational change in the bulk of uncomplexed ezrin consistent with inactivation of the scaffold complex.

5. As endocytosis progresses, the PTH-TMR/PTH1R complex dissociates from the ezrin-NHERF scaffold.

Highlights.

• Parathyroid hormone-TMR (a florescent dye) marks the location of the PTH1R in kidney

• PTH1R is prominently expressed on the luminal side of proximal convoluted tubules

• Apical PTH1R is functional, including activation of PKA and NPT2a regulation

• PTH-TMR/PTH1R directly co-localizes to a sub-population of ezrin

• PTH induces a conformational change in bulk of ezrin consistent with inactivation

Abbreviations

PTH

parathyroid hormone

PTHrP

parathyroid hormone-related protein

PTH1R

parathyroid hormone 1 receptor

PTH-TMR

parathyroid hormone labeled with the florescent tetramethyl-rhodamine dye

NPT2a

type IIa sodium-phosphate co-transporter

NHERF1 and 2

sodium-hydrogen exchanger regulatory factors

PIP2

phosphatidyinositol-4,5-bisphosphate

ERM

ezrin-radixin-moesin family of proteins

PDZ

psd95, discs large, ZO-1 interaction domain

PIPase

membrane-targeted catalytic domain of the type IV phosphoinositide-5-phosphatase

PCT

proximal convoluted tubule of the kidney

Aqp1

aquaporin 1

PKA-sub

protein kinase A substrate antibodies

CS3145

polyclonal ezrin antibodies

3C12

monoclonal ezrin antibody