TRPV5 Is Internalized via Clathrin-dependent Endocytosis to Enter a Ca2+-controlled Recycling Pathway*

The epithelial Ca2+ channel TRPV5 plays an essential role in transcellular Ca2+ transport and is one of the most Ca2+-selective members of the transient receptor potential superfamily. Regulation of the abundance of TRPV5 at the cell surface is critical in body Ca2+ homeostasis. However, little is known about the mechanisms underlying TRPV5 endo- and exocytosis. Here, we show that TRPV5 is constitutively internalized in a dynamin- and clathrin-dependent manner. Internalized TRPV5 first appears in small vesicular structures and then localizes to perinuclear structures positive for Rab11a. TRPV5 has a half-life of more than 8 h and is stable even after internalization from the cell surface for more than 3 h. Disruption of cell surface delivery of newly synthesized TRPV5 by brefeldin A does not reduce TRPV5-mediated Ca2+ influx in cells, suggesting the presence of a stable intracellular pool of the channel capable of recycling back to the surface. Furthermore, the endocytic recycling kinetics is decreased upon treatment with Ca2+ chelator BAPTA-AM, indicating that the channel's trafficking pathways are dynamically controlled by Ca2+.

Transient receptor potential (TRP) 3 channels have a remarkably diverse set of functions ranging from roles in sensory processes like vision, osmosensation, hearing, and taste to ion homeostasis (1,2). Most investigations regarding the regulation of the TRP family of ion channels have focused on channel gating and signaling cascades controlling channel activity at the plasma membrane. Regulation of the spatial dynamics of TRP channels is a relatively unexplored area. However, recent evidence shows that translocation of TRP channels to the cell surface is a significant factor determining their overall activity (3)(4)(5)(6)(7)(8)(9). The role of regulated trafficking to and from the cell surface is especially important for TRP channels with constitutive activity. TRPV5 is one of the most Ca 2ϩ -selective members of the TRP superfamily and shows constitutive activation at physiological membrane potentials (10). A TRPV5 knock-out mouse model displays significant hypercalciuria (6-fold increase in total Ca 2ϩ excretion) and severe bone abnormalities (11,12). Together with the localization of TRPV5 in Ca 2ϩ -transporting cells in the kidney and its pronounced regulation by calciotropic hormones, these data provide strong evidence for the role of this channel as a pivotal element in epithelial Ca 2ϩ transport (13).
Recent evidence shows that factors affecting the number of TRPV5 channels at the cell surface have a major contribution to the physiological regulation of this Ca 2ϩ channel. Novel channel-associated proteins have been identified, including S100A10-annexin 2 and Rab11a, which appear to play a role in the trafficking and/or cell surface stability of TRPV5 (14,15). Furthermore, klotho, a protein with glucuronidase activity that has been linked to longevity, was postulated to slow down cell surface removal of TRPV5 via an alteration of the extracellular TRPV5 glycosylation (16). Another factor affecting turnover of TRPV5 at the plasma membrane is the serine protease kallikrein that was shown to cause accumulation of TRPV5 at the cell surface via activation of the bradykinin/diacylglycerol/protein kinase C pathway (17). Although these studies have provided new information on the regulation of TRPV5, the molecular mechanisms utilized in the endocytosis of TRPV5 are not known. Similarly, the dynamics of TRPV5 internalization and the fate of internalized TRPV5 are unclear. In this study we show that TRPV5 is constitutively internalized via a dynaminand clathrin-dependent pathway and that internalized TRPV5 enters the Ca 2ϩ -sensitive recycling pathway. These findings provide a framework to understand the physiological regulation of TRPV5 trafficking. Furthermore, our data present the first evidence for a role of clathrin-dependent endocytosis and Ca 2ϩ -sensitive protein recycling in the regulation of TRP channel cell surface expression.

EXPERIMENTAL PROCEDURES
DNA Constructs-Constructs containing FLAG TM or hemagglutinin epitope (HA)-tagged or untagged full-length TRPV5 in * This work was supported in part by grants from the German Research Association (Deutsche Forschungsgemeinschaft Grant SFB 629) and by Dutch Organization of Scientific Research Grant TOP-CW 700.55.302. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs pT7Ts, pciNeo-IRES-GFP, or pCB6 were obtained as described previously (18). TRPV5 constructs with epitope tags at predicted extracellular positions were prepared by mutagenesis according to the manufacturer's protocol (Stratagene), resulting in the replacement of amino acids in TRPV5 by an HA or FLAG TM tag or in the insertion of these tags (Fig. 1A).
Cell Culture and Transfections-HeLa and HEK293 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and antibiotics. Stably transfected HeLa cells with inducible dynamin K44A expression (generously provided by Dr. S. Schmid, The Scripps Research Institute, La Jolla CA) were cultured as described (19), and dynamin was induced by culturing in the absence of tetracycline for 48 h. HEK293 cells were stably transfected with pCB6-HA-TRPV5 as described previously (20). HeLa cells were transfected with Lipofectamine 2000 (Invitrogen) or Effectene (Qiagen) following the manufacturers' instructions. TRPV5 was expressed in Xenopus laevis as described previously (18). For transient knock-down experiments HEK293 cells were seeded into 6-cm dishes and transfected the same day with annealed clathrin heavy chain siRNA duplexes (Ambion, 5Ј-CCUGCGGUCUGGAGUCAACtt-3Ј and 5Ј-GUUGACUCC-AGACCGCAGGtt) using OligofectAMINE (Invitrogen). This transfection was repeated the next day. One day after the second transfection cells were seeded onto coverslips or poly-Llysine-coated 10-cm plates. siRNA transfections in HeLa cells were performed similarly, but the second transfection was performed with both a plasmid encoding HA-TM-TRPV5-IRES-GFP and clathrin siRNA using Lipofectamine 2000 according to the manufacturer's instructions. Cells were processed for further analyses 48 h after the second transfection. All experiments were performed in parallel with non-targeting control siRNA duplexes obtained from Dharmacon (Chicago, IL) instead of the clathrin siRNA.
Fluorescence Microscopy-Cells were grown on coverslips and incubated in internalization medium (10 mM KCl, 140 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 20 mM Hepes, pH 7.4, 5.5 mM glucose, and 1% (w/v) bovine serum albumin for 30 min at 37°C. To track TRPV5 internalization in pulse-chase experiments, cells were first incubated for 1 h at 4°C in internalization medium with mouse (1:100) or rabbit (1:200) anti-HA antibodies. Cells were then washed with cold PBS and subsequently chased for various times in internalization medium at 37°C, fixed, and internalized anti-HA visualized with Texas Red-or Alexa 594-coupled secondary antibodies. Images were acquired using a LSM510 Meta confocal microscope (Carl Zeiss, Jena, Germany). TRPV5-expressing cells were identified by GFP expression due to the bicistronic vector used. Anti-clathrin heavy chain antibodies were from BD Biosciences. Antibodies directed against TfR (H68.4) and Rab11a were obtained from Zymed Laboratories Inc. (South San Francisco, CA).
Cell Surface Biotinylation and Internalization Assay-HEK293 cells stably transfected with HA-tagged TRPV5 were seeded on poly-L-lysine-coated 10-cm dishes (7 million cells/ plate). After 1-2 days proteins present at the cell surface were biotinylated at 4°C using sulfo-NHS-LC-LC-biotin (only Fig.  6A) or sulfo-NHS-SS-biotin (0.5 mg/ml, Pierce) in PBS supplemented with 0.5 mM CaCl 2 and 1.0 mM MgCl 2 (PBS-CM) for 30 min. Unreacted biotin was quenched using 0.1% (w/v) bovine serum albumin in PBS-CM. The cells were incubated at 37°C for the indicated times in culture medium or kept at 4°C. To measure TRPV5 internalization, biotin that remained at the cell surface was removed by incubation with fresh 100 mM 2-Mercaptoethanesulfonic acid sodium salt (mesna) in 100 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, pH 8.6, 0.2% (w/v) bovine serum albumin 3 times for 20 min at 4°C. Subsequently, cells were incubated for 10 min with 120 mM iodoacetic acid in PBS-CM. Finally, cells were washed with PBS-CM and PBS and then lysed in 150 mM NaCl, 5 mM EDTA, 50 mM Tris-HCl, pH 7.5, 1% (v/v) Nonidet P-40, 5 g/ml leupeptin, 1 g/ml pepstatin A at 4°C. The lysate was centrifuged for 10 min at 16,000 ϫ g, and biotinylated proteins were precipitated overnight from the supernatant using Neutravidin-coupled beads (Pierce) and analyzed by immunoblot analyses.

TRPV5 Is Internalized via Clathrin-dependent Endocytosis
Ca 2ϩ Concentration Measurements-PciNeo TRPV5-IRES-GFP-transfected HEK293 cells were grown on coverslips and incubated with 3 M Fura2-AM for 20 min at room temperature, washed twice, and allowed to equilibrate for another 10 min in HEPES-Tris buffer (132 mM NaCl, 4.2 mM KCl, 1.2 mM CaCl 2 , 1 mM MgCl 2 , 5.5 mM D-glucose, and 10 mM HEPES, pH 7.4) Coverslips were placed in an incubation chamber mounted on an inverted microscope (Axiovert 200M, Carl Zeiss) equipped with a 40ϫ, 1.3 NA Plan NeoFluar objective. Changes in cytosolic [Ca 2ϩ ] were measured with Fura2 alternatively excited at 340 and 380 nm using a monochromator (Polychrome IV, TILL Photonics, Gräfelfing, Germany). Fluorescence emission light was directed by a 415DCLP dichroic mirror (Omega Optical Inc., Brattleboro, VT) through a 510WB40 emission filter (Omega Optical Inc.). After background correction for each wavelength, the fluorescence emission ratio after excitation at 340 and 380 nm was calculated. In each experiment the value at t ϭ 0 for non-transfected cells was set at 100%, to which all other values were related.
Statistical Analysis-In all experiments the data are expressed as the mean Ϯ S.E. Overall statistical significance was determined by analysis of variance. p values below 0.05 were considered significant.

RESULTS
Internalization of Cell Surface TRPV5-Studies into the mechanisms underlying the regulated endocytosis of TRPV5 had been hampered by the difficulty of detecting TRPV5 at the cell surface. Various attempts to create antibodies against regions of TRPV5 that face the extracellular milieu or to create functional TRPV5 channels with epitope tags in these regions were unsuccessful ( Fig. 1A and data not shown). Finally, we successfully employed a different strategy where we added an additional transmembrane domain to the amino terminus to create a topology with an extracellular facing amino terminus tagged with HA (HA-TM-TRPV5, Fig. 1A). This construct yielded a functional TRPV5 channel, as small but distinct TRPV5-specific currents were detected in transiently transfected HEK293 cells (data not shown), and significant 45 Ca 2ϩ uptake (63 Ϯ 3% of untagged TRPV5) was measured in X. laevis oocytes ectopically expressing the channel construct. Furthermore, co-immunoprecipitation experiments demonstrated that HA-TM-fused TRPV5 is also able to form multimers with GFP-TRPV5, as expected from a channel which functions in a tetrameric configuration (18) (Fig. 1B). As a negative control, TRPV2-yellow fluorescent protein (YFP) could not be co-immunoprecipitated with HA-TM-TRPV5, in line with the previously described inability of TRPV2 and TRPV5 to form heteromultimers (21).
To visualize TRPV5 internalization, we incubated live HA-TM-TRPV5-expressing HeLa cells with anti-HA antibodies at 37°C. Fig. 2A shows that only the transfected cells accumulated significant amounts of anti-HA antibody, indicative of internalized TRPV5. No antibody uptake was observed with anti-VSV antibodies, used as isotype control antibody (data not shown). Incubation with anti-HA antibodies at 4°C only yielded a signal at the cell surface (Fig. 2B). Subsequent warming of the cells to 37°C resulted in almost complete internalization of the antibody already within 5 min. The internalized TRPV5 first (5-10 min) appeared in numerous small vesicles in the periphery of the cell and then was found in a number of larger vesicles. Continuous labeling resulted in the signal in larger vesicles with some small peripheral vesicles remaining. Multiple z-sections of cells of this experiment are given in supplemental Fig. 1 showing similar results in all focal planes. This system is applicable to other cell types as anti-HA uptake was also observed in HA-TM-TRPV5-expressing HEK293 cells, commonly used for heterologous expression of TRPV5 (Fig.  2C) and in HA-TM-TRPV5-expressing Chinese hamster ovary cells (data not shown).
Internalization of Cell Surface TRPV5 Is Dependent on Clathrin and Dynamin-Anti-HA antibody uptake was employed to determine the role of dynamin in the endocytosis of TRPV5 using HeLa cells expressing the dominant negative dynamin K44A mutant in a tetracycline-inducible manner (HeLa K44A) (19). The anti-HA signal remained at the cell surface in most of  FEBRUARY 15, 2008 • VOLUME 283 • NUMBER 7 JOURNAL OF BIOLOGICAL CHEMISTRY 4079 the cells when dynamin K44A expression was induced by cell culture in the absence of tetracycline, whereas HA-TM-TRPV5 internalized normally when the cells were cultured in the presence of tetracycline, i.e. under non-inducing conditions ( Fig. 3A and supplemental Fig. 2). Similar results were obtained in Tf uptake experiments, with poor Tf internalization when dynamin K44A was overexpressed but significant Tf uptake into non-induced cells (Fig. 3B).

TRPV5 Is Internalized via Clathrin-dependent Endocytosis
To exclude that the dynamin-dependent internalization of TRPV5 was only a consequence of the anti-HA antibody addition, we employed as another model system HEK293 cells stably expressing TRPV5 (without the additional transmembrane domain of HA-TM-TRPV5). These cells show a low expression level of the channel that could still be detected using immunoblot analysis, but the cell surface expression of TRPV5 was too low to be detected by immunocytochemistry. Therefore, to probe for endocytosis of plasma membrane-resident TRPV5, cell surface molecules were labeled with sulfo-NHS-ss-biotin, and cells were then placed at 37°C to permit constitutive internalization of biotinylated proteins. Biotin remaining at the cell surface was removed using cell-impermeable mesna. Maximum internalization was achieved within 15-30 min (supplemental Fig. 3), although the internalization signal, i.e. mesna-resist-ant TRPV5, always remained significantly lower than 100% (referred to as the "no mesna" control) even at incubation times up to 2 h. To elucidate the involvement of dynamin in constitutive TRPV5 internalization analyzed by this surface biotinylation assay, we used dynasore, a cell-permeable inhibitor of dynamin (22). The addition of 80 M dynasore blocked TRPV5 uptake into a mesna-resistant structure, further confirming a role for dynamin in the endocytosis of TRPV5 (Fig. 3C).
We next analyzed the contribution of clathrin-dependent mechanisms to TRPV5 uptake. First, the addition of 450 mM sucrose to the medium (added after the biotinylation) was cho-

TRPV5 Is Internalized via Clathrin-dependent Endocytosis
sen as an unspecific tool to disrupt the formation of clathrincoated vesicles, thereby blocking clathrin-dependent endocytosis (23). This led to a complete block of TRPV5 uptake (Fig.  3C). To test the role of clathrin more directly, we depleted cells of clathrin heavy chain by siRNA transfection and first tested the effect on the internalization of TRPV5 by analyzing the uptake of cell surface-biotinylated channels. Although only partial depletion of clathrin was obtained in this transient approach (clathrin heavy chain expression was decreased by ϳ60%), TRPV5 internalization was reduced in the clathrin siRNA-transfected cells as compared with control cells (Fig. 4,  A and B). TfR internalization in the clathrin siRNA-transfected cells was also partially reduced. That TfR internalization is clathrin-dependent (24) suggests the remaining internalized signal of both TRPV5 and TfR mainly originates from cells in the population showing poor or no depletion of clathrin. Therefore, we also tested the role of clathrin at the single cell level, employing the antibody internalization assay and HA-TM-TRPV5-expressing cells. To this end HeLa cells were transfected with clathrin siRNA and the next day again co-transfected with clathrin siRNA and HA-TM-TRPV5. Immunocytochemistry using anticlathrin antibodies showed a significant reduction of clathrin heavy chain protein in 40 -60% of these cells (Fig. 4D). Furthermore, Tf uptake (20 min, 37°C) was clearly inhibited in cells showing the clathrin heavy chain depletion (Fig. 4E). Likewise, HA-TM-TRPV5 internalization was inhibited in the clathrin knock-down cells as revealed by the anti-HA signal remaining at the cell surface as opposed to the control cells showing HA-TM-TRPV5 internalization and accumulation of the signal in the perinuclear region (Fig. 4, F and G).
Internalized TRPV5 Is Not Degraded but Targeted to a Recycling Compartment-We next addressed the fate of internalized TRPV5. To avoid effects due to unphysiologically high overexpression, we used our model characterized by a relatively low and stable expression of TRPV5. To determine the stability of surface-expressed TRPV5, biotinylated cells were placed back at 37°C for 0 or 3 h, and the amount of biotinylated channel protein was determined by immunoblot analysis using anti-TRPV5 antibodies. To test for possible targeting of internalized TRPV5 to the degradation machinery, inhibitors of the proteasomal (MG132) or lysosomal (chloroquine) degradation pathway were used. Remarkably, the amount of biotinylated TRPV5 in these cells remained constant even 3 h after the biotinylation and was not affected by 3 h of preincubation of the cell MG132 or chloroquine degradation pathway (Fig. 5A). As controls we also immunoblotted for the receptors for the epidermal growth factor or Tf (TfR), which are either degraded or recycled upon internalization, respectively. We next treated cells with brefeldin A (BFA) to block the insertion of newly synthesized channels into the plasma membrane. In the presence of BFA a slow decrease in channel activity at the cell surface is indicative of internalized channels being degraded and not recycled, as was shown for the epithelial Na ϩ channel (25). In contrast, TRPV5 activity, as measured by ruthenium red-sensitive 45 Ca 2ϩ uptake, was not affected by prolonged incubation with BFA even up to 4h (Fig. 5B). Based on these results, we postulated that TRPV5 enters a recycling compartment after internalization. To test  FEBRUARY 15, 2008 • VOLUME 283 • NUMBER 7 this hypothesis, we performed colocalization studies of internalized HA-TM-TRPV5 with Rab11 or CD63, markers of a recycling or degradative compartment, respectively. Early after internalization, when HA-TM-TRPV5 localized to numerous small peripheral vesicles, little colocalization was observed between Rab11 and internalized anti-HA (Fig.  5C). However, after incubation of the HA-TM-TRPV5-expressing cells with anti-HA antibodies at 37°C for 1 h, a significant co-localization of internalized TRPV5 with endogenous Rab11 was observed. Co-localization was most prominent in larger perinuclear structures and less in the small peripheral TRPV5-positive punctae (Fig. 5D). In contrast, no obvious co-localization with CD63 was observed, even after prolonged incubation with anti-HA antibodies, again suggesting a limited targeting of the channel to late endosomes and then to the degradative lysosomal pathway (Fig. 5E).

TRPV5 Is Internalized via Clathrin-dependent Endocytosis
The stability of (internalized) TRPV5 was also determined on a longer time scale by pulse-chase analysis. Immediately after metabolic labeling using [ 35 S]methionine/cysteine (0 h), the TRPV5 signal was detected in SDS-PAGE mainly at ϳ75 kDa, most likely reflecting the core-glycosylated state of the channel. A minor background signal at ϳ70 kDa was also visible that was already seen in the non-transfected cells (Fig. 6A). After 1 h, most of the signal had shifted into the high molecular mass TRPV5 band, which runs at 85-100 kDa (Fig. 6, A and B). This shift in molecular weight likely reflects processing of the glycosylation moiety of TRPV5 in the Golgi compartment. Analysis of the total TRPV5 and only the upper, fully glycosylated-TRPV5 band revealed that TRPV5 was stable for 8 h but showed a large reduction in the signal at 16 h (Fig. 6, B and C). In comparison to the fully glycosylated form, it seems that the core-glycosylated TRPV5 disappears slightly more rapidly. FIGURE 5. TRPV5 is recycled rather than degraded upon internalization. A, cell surface proteins of HEK293 cells stably expressing TRPV5 were biotinylated after preincubation in control medium (CT) or medium containing inhibitors of proteasomal (MG132) or lysosomal (chloroquine) degradation. Subsequently, cells were placed at 37°C for 0 or 3 h in medium containing 10% fetal calf serum supplemented with 50 nM epidermal growth factor. Cells were lysed and biotinylated proteins precipitated using Neutravidin beads and analyzed by immunoblot. The TRPV5 signal did not decrease in a significant extent in this 3-h chase at 37°C, similar to TfR. In contrast, the epidermal growth factor receptor (EGFR), used as an example of a receptor routed to the degradative pathway, was completely degraded in this chase in a lysosomal fashion. Biotinylated TRPV5 was equally stable in the absence of epidermal growth factor (not shown). B, radioactive Ca 2ϩ uptake experiments were performed in HEK293 cells stably expressing TRPV5, preincubated for various times with 5 g/ml BFA to block insertion of newly synthesized TRPV5 into the plasma membrane. Clear ruthenium red (RR)-sensitive TRPV5 activity was observed. No reduction in the level of TRPV5 activity at the cell surface was detected even at prolonged BFA treatment. The asterisk indicates a significant difference (p Ͻ 0.05) from a situation without ruthenium red. The activity of BFA was verified by the disruption of the Golgi, as visualized using the Golgi marker 58K. C, HA-TM-TRPV5-transfected HeLa cells were incubated with anti-HA at 4°C to label TRPV5 and then warmed to 37°C for 5 min, fixed, and permeabilized. Endogenous Rab11 was visualized using rabbit anti-Rab11 Secondary antibodies used were anti-mouse Texas Red and anti-rabbit Cy5. Note that internalized TRPV5 signal is present in a large number of small peripherally localized vesicles that do not co-label with anti-Rab11. D, HA-TM-TRPV5-expressing cells were incubated for 1 h at 37°C with mouse anti-HA. After this prolonged incubation time, TRPV5 is found in large perinuclear structures that show colocalization with Rab11. E, labeling was carried out as in C. Only a low percentage of vesicles containing internalized TRPV5 shows colocalization with CD63. HA-TM-TRPV5-transfected cells were identified by the GFP signal (left inset in C, D, and E). The bar indicates 10 m. NT indicates non-transfected cells.
Recycling of Internalized TRPV5 Is Ca 2ϩ -dependent-Using the Ca 2ϩ -dye Fura2, we demonstrated that expression of TRPV5 in HEK293 cells resulted in a ϳ2-fold increase in the resting intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) (Fig. 7C). This difference was due to TRPV5-mediated Ca 2ϩ influx as [Ca 2ϩ ] i decreased to control levels when extracellular Ca 2ϩ was omitted, and EDTA was added (Fig. 7C). Subsequently, we tested whether the cytosolic Ca 2ϩ concentration affects the endocytic trafficking of TRPV5. However, [Ca 2ϩ ] i was not significantly altered when we lowered the extracellular [Ca 2ϩ ] to values as low as 10 M (not shown). Further reduction of the extracellular Ca 2ϩ concentration using 2 mM EGTA resulted in massive cell detachment during the mesna-based internalization assay. Therefore, we tested whether lowering the cytosolic Ca 2ϩ concentration using the Ca 2ϩ chelator BAPTA would affect the endocytic trafficking of TRPV5. The addition of BAPTA (ϪAM) to cells has been performed in studies on TRPV5 function, as this treatment decreases the Ca 2ϩ -dependent channel inactivation. In vivo this Ca 2ϩ -buffering function is mediated by calbindins (26,27). Interestingly, the amount of mesna-protected TRPV5 increased ϳ4-fold (408 Ϯ 98%) upon treatment of the cells with BAPTA-AM (Fig. 7, A and B). The amount at the cell surface (no mesna condition) at the start of the chase was not significantly affected (100 Ϯ 26 versus 113 Ϯ 29% of control, p ϭ 0.7), and the expression of TRPV5 was similar in all conditions (Fig. 7A, lower panel). In control experiments the effect of Ca 2ϩ chelation on TfR trafficking was ana-lyzed. Preincubation with BAPTA-AM for 30 min before cell surface biotinylation resulted in a similar increase in the amount of internalized biotinylated TfR compared with the effect of BAPTA-AM on the amount of internalized TRPV5 FIGURE 6. Determination of the stability of TRPV5 by pulse-chase analysis after metabolic labeling. A, HEK293 cells stably expressing TRPV5 were incubated with [ 35 S]methionine/cysteine for 1 h at 37°C to metabolically label all newly synthesized proteins. Cells were then washed and incubated in normal culture medium (containing only unlabeled amino acids) for the times indicated. Subsequently, cells were lysed and incubated with guinea pig anti-TRPV5 immobilized on protein A-agarose beads. Immunoprecipitates were subjected to SDS-PAGE and autoradiography. Specific TRPV5 signal was detected at ϳ75 kDa, the size of core-glycosylated TRPV5, immediately after labeling. NT indicates non-transfected cells. B, within 1 h, an extra band appeared at the size of complex-glycosylated TRPV5, indicating that a significant fraction of the newly produced TRPV5 has matured beyond the endoplasmic reticulum. The fully glycosylated TRPV5 is stable for 8 h, but a large fraction is degraded in the following 8 h, as the signal is greatly reduced at 16h. C, quantification of the stability of the high molecular weight (complex glycosylated) TRPV5 band (left) or the total TRPV5 signal (right).  FEBRUARY 15, 2008 • VOLUME 283 • NUMBER 7 (Fig. 7D). In these experiments it was also tested whether BAPTA-AM affects either the internalization or endosomal recycling kinetics. To this end the effect of BAPTA-AM treatment was determined 5 min after warming up the cells, when TfR had not reached the recycling compartment, and at 30 min, when TfR had accumulated in recycling endosomes. mesnaprotected (internalized) TfR was clearly visible at 5 min of endocytosis, but BAPTA-AM treatment did not result in an increased signal at this time point, indicating similar internalization rates under both conditions. However, at 30 min of chase, significantly more TfR was mesna-protected in the BAPTA-AM-treated cells as compared with mock-treated cells. The validity of the assay was also shown in control experiments using 50 M monensin, a known inhibitor of endosomal recycling (28). This drug, when added during the chase, also increased the amount of mesna-protected TfR. Together, this shows that lowering the cytosolic Ca 2ϩ concentration using BAPTA-AM reduces the endosomal recycling of TfR and TRPV5.

DISCUSSION
The present study analyzed the endocytic trafficking of the epithelial Ca 2ϩ channel TRPV5 and obtained the following key findings; (i) TRPV5 is constitutively internalized via a dynamin and clathrin-dependent mechanism, (ii) internalized TRPV5 enters the recycling pathway, where it is stable for at least 3 h, (iii) the recycling rate of TRPV5 is modulated by the intracellular Ca 2ϩ concentration. These findings provide novel insight into TRPV5 trafficking and also introduce new tools to study and rationalize the function of calciotropic factors that affect the abundance of TRPV5 at the cell surface.
TRPV5 is constitutively active at physiological membrane potentials (10). Therefore, TRPV5-mediated Ca 2ϩ influx is largely dependent on the amount of TRPV5 channels at the cell surface. Recent studies describe two proteins, i.e. klotho and kallikrein, that affect the body Ca 2ϩ balance via alterations in the trafficking of TRPV5 (16,17). However, due to a lack of appropriate tools, these studies measured a possible decrease in TRPV5 endocytosis only indirectly. Because only a minute fraction of TRPV5 is present at the cell surface at any given moment, direct visualization of internalization is obscured by the dominant intracellular signal. Therefore, two approaches were established in the present study to address this issue. First, we generated a TRPV5 protein with an epitope tag facing the extracellular environment to label cell surface-expressed channels with exogenously added antibodies. This principle has been used to study endocytosis of a number of ion channels and transporters, including the glucose transporter 4 (29). Using this construct, we could visualize the internalization of TRPV5 and even track the localization of internalized channels in the cell. Second, as an independent method we used cell surface biotinylation to obtain more quantitative data on channel internalization. Both methods, which had never been employed together before to study TRP channel trafficking, yielded novel and complementary insights into the endocytosis and subsequent trafficking of TRPV5.

TRPV5 Shows Constitutive Dynamin-and Clathrin-dependent Internalization and Enters a Recycling Pathway-Several
lines of evidence show that TRPV5 is internalized by clathrinmediated endocytosis. First, using the HA-TM-TRPV5 construct we observed a rapid and constitutive internalization of anti-HA antibodies. The internalization was specific for TRPV5-expressing cells and did not occur upon incubation with nonspecific antibodies, showing that general fluid phase uptake, e.g. by macropinocytosis, could be ruled out as internalization mode. Second, the role of dynamin in the internalization of TRPV5 was demonstrated by blocking TRPV5 endocytosis by overexpression of the dominant-negative dynamin K44A mutant and by chemical inhibition of dynamin using the cell permeable inhibitor dynasore (22). Third, hypertonic medium (450 mM sucrose) and, more specifically, depleting cells of clathrin using clathrin heavy chain siRNA also blocked TRPV5 internalization.
Once internalized, TRPV5 is rather stable and shows virtually no degradation within the first 3 h. To test whether this was due to internalized channels entering the recycling pathway, we subjected cells to BFA treatment for different time points. BFA has been used before to demonstrate internalization-dependent degradation of the epithelial Na ϩ channel and of ROMK, a K ϩ channel expressed in the kidney (25,30). For both epithelial Na ϩ channel and ROMK, prolonged incubation (3-4 h) with BFA strongly reduced their activity, indicative of a reduced number of channels at the cell surface, and this process was attenuated by inhibition of endocytosis. In contrast, BFA treatment did not affect TRPV5 activity even after 4 h, suggesting that transport of newly synthesized TRPV5 from the Golgi to the plasma membrane does not play a significant role within this time frame. Therefore, an endosomal intracellular pool of TRPV5, which replenishes the constitutively internalized TRPV5 at the cell surface, is likely to exist. In addition, the pulse-chase results indicate an almost complete processing of TRPV5 to the fully-glycosylated form with considerable stability, suggesting that a significant pool of TRPV5 is present in a post-endoplasmic reticulum compartment for a significant amount of time. We postulate that this pool is located in a recycling compartment continuously exchanging TRPV5 with the plasma membrane. This is in line with the prominent colocalization of Rab11a with internalized TRPV5 (Fig. 5) and corroborated by previous data on endogenously expressed TRPV5 in primary renal epithelial cells (15).
Modulation of TRPV5 Recycling by the Intracellular Ca 2ϩ Concentration-Previous studies indicated that regulatory pathways controlling the abundance at the cell surface of membrane proteins can target distinct steps in their trafficking. Two recent examples include the cargo-regulated rate of clathrincoated vesicle formation (31) and cellular signaling-regulated sorting of the internalized material (32,33). Our findings show constitutive clathrin/dynamin-dependent internalization of TRPV5, rendering it likely that control of this step could enable the cell to accumulate the channel at the cell surface, as observed with klotho or kallikrein treatment. Another mechanism to regulate protein trafficking in the endosomal pathway is modulation of endocytic recycling. Until recently, recycling was considered to be a constitutive, non-regulated process that would not play a significant role in the regulation of protein trafficking (34). However, it was demonstrated that protein kinase C regulates the recycling rate of proteins, i.e. E-cadherin (35,36). Here we show that treatment of cells with BAPTA-AM reduces the endocytic recycling rate without affecting early steps in internalization. These findings are in line with earlier results on a stimulatory role of Ca 2ϩ on TfR recycling (37,38). Using electropermeabilized cells, Knight (37) previously showed that TfR recycling is half-maximal at 1 M, compatible with values observed in TRPV5-expressing cells (Fig. 7C). Modulation of protein recycling could be relevant in physiological terms for the regulation of TRPV5, which is one of the most Ca 2ϩ -selective members of the TRP family. It is known that the intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) has evident effects on channel properties of TRPV5, showing rapid inactivation upon increases in [Ca 2ϩ ] i , as determined by patch-clamp analysis (26). Furthermore, recovery of TRPV5 from a Ca 2ϩ -induced inactivated state is a relatively slow process, with halfmaximal recovery reached only after ϳ96 s (26). This timescale could allow a role of endocytic recycling, where inactivated TRPV5 at the cell surface is replaced with active TRPV5 from the intracellular TRPV5 pool in this recovery process. The effect of [Ca 2ϩ ] i on TRPV5 recycling described in this study suggests a physiological mechanism where a Ca 2ϩ -dependent increase in TRPV5 recycling would enable the cell to rapidly exchange inactivated cell surface TRPV5 with the active channel from a recycling compartment to restore Ca 2ϩ influx.
The field of TRP channel trafficking is still relatively unexplored and has focused on exocytic events (39,40). Until recently, exploration of TRP channel internalization has remained very sparse (41)(42)(43). This study presents the elucidation of the internalization pathway utilized to extract TRPV5 from the cell surface and of the subsequent targeting to the recycling machinery. Clathrin-dependent internalization and endocytic recycling might represent a phenomenon shared among several TRP channels, although the regulatory mechanisms underlying the kinetics of these processes could be tailored to individual TRP channels. Furthermore, this knowledge provides a framework to understand and further study the regulation of TRPV5 abundance at the plasma membrane.