N-Glycosylation Determines the Abundance of the Transient Receptor Potential Channel TRPP2*

Background: Transient receptor potential channel TRPP2 is a glycoprotein mutated in polycystic kidney disease. Results: Five N-glycosylation sites in TRPP2 and glucosidase II-mediated glycan trimming determine TRPP2 protein abundance. Conclusion: N-Glycosylation of TRPP2 in the first extracellular loop is required for efficient protein translation and stability. Significance: The mechanisms controlling the biogenesis of TRPP2 may provide insights into the pathogenesis of polycystic kidney and polycystic liver disease. Glycosylation plays a critical role in the biogenesis and function of membrane proteins. Transient receptor potential channel TRPP2 is a nonselective cation channel that is mutated in autosomal dominant polycystic kidney disease. TRPP2 has been shown to be heavily N-glycosylated, but the glycosylation sites and the biological role of N-linked glycosylation have not been investigated. Here we show, using a combination of mass spectrometry and biochemical approaches, that native TRPP2 is glycosylated at five asparagines in the first extracellular loop. Glycosylation is required for the efficient biogenesis of TRPP2 because mutations of the glycosylated asparagines result in strongly decreased protein expression of the ion channel. Wild-type and N-glycosylation-deficient TRPP2 is degraded in lysosomes, as shown by increased TRPP2 protein levels upon chemical inhibition of lysosomal degradation. In addition, using pharmacological and genetic approaches, we demonstrate that glucosidase II (GII) mediates glycan trimming of TRPP2. The non-catalytic β subunit of glucosidase II (GIIβ) is encoded by PRKCSH, one of the genes causing autosomal dominant polycystic liver disease (ADPLD). The impaired GIIβ-dependent glucose trimming of TRPP2 glycosylation in ADPLD may explain the decreased TRPP2 protein expression in Prkcsh−/− mice and the genetic interaction observed between TRPP2 and PRKCSH in ADPLD. These results highlight the biological importance of N-linked glycosylation and GII-mediated glycan trimming in the control of biogenesis and stability of TRPP2.

Glycosylation plays a critical role in the biogenesis and function of membrane proteins. Transient receptor potential channel TRPP2 is a nonselective cation channel that is mutated in autosomal dominant polycystic kidney disease. TRPP2 has been shown to be heavily N-glycosylated, but the glycosylation sites and the biological role of N-linked glycosylation have not been investigated. Here we show, using a combination of mass spectrometry and biochemical approaches, that native TRPP2 is glycosylated at five asparagines in the first extracellular loop. Glycosylation is required for the efficient biogenesis of TRPP2 because mutations of the glycosylated asparagines result in strongly decreased protein expression of the ion channel. Wildtype and N-glycosylation-deficient TRPP2 is degraded in lysosomes, as shown by increased TRPP2 protein levels upon chemical inhibition of lysosomal degradation. In addition, using pharmacological and genetic approaches, we demonstrate that glucosidase II (GII) mediates glycan trimming of TRPP2. The non-catalytic ␤ subunit of glucosidase II (GII␤) is encoded by PRKCSH, one of the genes causing autosomal dominant polycystic liver disease (ADPLD). The impaired GII␤-dependent glucose trimming of TRPP2 glycosylation in ADPLD may explain the decreased TRPP2 protein expression in Prkcsh ؊/؊ mice and the genetic interaction observed between TRPP2 and PRKCSH in ADPLD. These results highlight the biological importance of N-linked glycosylation and GII-me-diated glycan trimming in the control of biogenesis and stability of TRPP2.
Autosomal dominant polycystic kidney disease (ADPKD) 4 is one of the most common potentially fatal monogenetic diseases in man, affecting one in 400 to one in 1000 live births (1). Mutations in Polycystic Kidney Disease 2 (PKD2) cause ADPKD type 2 (OMIM no. 613,095) (2). The PKD2 gene product polycystin 2 is a member of the transient receptor potential (TRP) channel superfamily. It is classified as transient receptor potential channel polycystin 2 (TRPP2), the prototypical member of the TRPP subfamily (3). TRP channels are involved in diverse physiological processes, including sensory physiology, male fertility, regulation of vascular tone, as well as Ca 2ϩ and Mg 2ϩ homeostasis (3). They share the leitmotif of permeability to cations and rather low sensitivity to membrane voltage (3). Studies on the architecture of TRP channels yielded considerable evidence that these channels function as tetramers (4 -6). Monomeric TRPP2 channel subunits are integral membrane proteins with six transmembrane helices (S1-S6), framing a pore-forming loop between S5 and S6 (TRPP2 634 -659 ), and cytosolic amino and carboxyl termini (TRPP2 1-223 and TRPP2 680 -968 , respectively) (7). A prominent feature of TRPP2 is the large extracellular loop between S1 and S2, consisting of 223 amino acids (TRPP2 245-468 ) (Fig. 1A). Loss-of-function phenotypes in ADPKD patients and TRPP2-deficient animal models have revealed important biological roles of this ion channel, ranging from renal tubular morphogenesis to determination of embryonic left-right asymmetry (8). Yet, our knowledge about the molecular mechanisms translating ion channel activity into complex morphogenetic processes is still limited. An important step toward understanding the function of TRPP2 is a molecu-lar dissection of its posttranslational modifications and the characterization of their biological relevance. Posttranslational phosphorylation and glycosylation have been shown to significantly impact the behavior of ion channels, including channel activity, conductance, and protein stability. Several studies have demonstrated a critical role of phosphorylation in TRPP2 function (9 -14). Biochemical studies have revealed that TRPP2 is glycosylated, but a detailed mapping and characterization of glycosylated amino acid residues is lacking (15,16).
Glycosylation of proteins plays crucial roles in biology, from quality control of protein folding to involvement in a large number of biological recognition events. The addition and processing of asparagine-linked (N-linked) oligosaccharides in the endoplasmic reticulum (ER) is the most common form of glycosylation. N-Linked glycosylation promotes proper folding and assembly of glycoproteins and increases their stability. Glycosylation begins with the cotranslational addition of a preformed, 14-residue glycan to asparagine residues at the NX[ST] consensus sequences, where N is asparagine, X can be any amino acid except proline, followed by either serine or threonine ([ST]), respectively. For all N-glycans, processing of the core glycan continues with trimming of terminal glucose residues by ER glucosidases I and II (17,18). Glucosidase II-mediated trimming of carbohydrates eventually facilitates glycoprotein trafficking to the Golgi apparatus. Persistent glucosylation of unfolded or misfolded proteins inhibits ER exit and triggers degradation (19). Glucosidase II (GII) is of particular interest to glycan processing of TRPP2 because in vivo studies have placed TRPP2 and the non-catalytic glucosidase II ␤ (GII␤) subunit of this enzyme in a common biogenetic pathway (20). Although the kidney-specific elimination of GII␤ causes mild cystic kidney disease in mice, a severe PKD phenotype manifests on a Pkd2 ϩ/Ϫ -sensitized background (20). Loss of GII␤ is accompanied by a reduction of TRPP2 protein expression (21,20). In humans, mutations in Protein Kinase C Substrate 80K-H (PRKCSH), the gene encoding for GII␤, cause autosomal dominant polycystic liver disease (ADPLD), which manifests with bile duct cysts that are indistinguishable from the liver phenotype in ADPKD (evident in 94% of ADPKD patients who are older than 35 years) (OMIM no.174,050) (22)(23)(24)(25). GII is an ER enzyme with a key role in glycoprotein folding and quality control. It is comprised of two subunits: catalytic GII␣ and accessory GII␤ (26). GII␤ has been shown to physically interact with TRPP2, and it has been suggested that GII may be involved in the processing of TRPP2 glycosylation (21,20). However, direct experimental data supporting this model are missing.
We are only beginning to understand the importance of glycosylation of TRP channels. It has been shown for some TRP channels that N-glycosylation plays a critical role for cellular localization and function (27)(28)(29)(30)(31)(32). Biochemical studies have revealed that TRPP2 is glycosylated but the glycosylation sites and relevance of glycosylation for TRPP2 biogenesis and stability have not been investigated (15,16). Here we present a comprehensive mapping of N-glycosylated amino acid residues in TRPP2. Furthermore, we show that GII␤ is required for glycan trimming of TRPP2 and that proper N-linked glycosylation is essential for adequate TRPP2 protein expression.
Mice-C57BL/6 mice were used as the wild type in Fig. 1B. PRKCSH experiments were performed on a C57BL/6 -129 mixed background (Fig. 8, B and C). The conditional Prkcsh flox mice have been described previously (20). Deletion of exons 6 and 7 by Cre recombinase results in a functional null allele (20). Ksp-Cre mice with constitutive Cre recombinase expression in the thick ascending limb of the loop of Henle, distal convoluted tubule, and collecting duct starting at 9.5 days after fertilization have been described previously (35).
Protein Isolation-Native TRPP2 was isolated from confluent M1 cells or from the kidneys of 8-day-old mice. Transiently transfected cells were harvested 24 h after transfection if not stated otherwise. Samples were lysed in ice-cold IP buffer (1% Triton X-100, 20 mM Tris-HCl (pH 7.5), 50 mM NaCl, 50 mM NaF, 15 mM Na 4 P 2 O 7 , and 0.1 mM EDTA (pH 8)) supplemented with 2 mM Na 3 VO 4 and complete protease inhibitor mixture (Roche). Lysates were first centrifuged at 4°C for 15 min at 20,000 ϫ g and subsequently subjected to ultracentrifugation at 4°C for 30 min at 100,000 ϫ g. Supernatants were either directly denatured 30 min at 42°C with 2ϫ Laemmli sample buffer and 100 mM DTT or used for immunoprecipitation and glycosylation studies.
TRPP2 Immunoprecipitation-For TRPP2 immunoprecipitation, protein G-Sepharose beads (GE Healthcare) were incubated with 2 g of anti-TRPP2 (G-20, Santa Cruz Biotechnology) antibody in PBS (Biochrom) for 90 min at 4°C. Antibodycoupled beads were washed three times for 10 min at 4°C with PBS and then combined with the respective lysate (Ͼ5 mg of total protein). Immunoprecipitation took place overnight on a rotating wheel at 4°C. Subsequently, beads were washed three times for 10 min at 4°C with IP buffer. Proteins were eluted for 30 min at 42°C with 2ϫ Laemmli sample buffer and 100 mM DTT or, for glycosylation assays, with 0.5% SDS and 40 mM DTT.
Glycosylation Assays of TRPP2 Immunoprecipitates-Endoglycosidase H (Endo H, New England Biolabs) sample digestion was performed for 24 h at 37°C with 5000 units of Endo H in G5 reaction buffer (New England Biolabs). For peptide-Nglycosidase F (PNGase F, New England Biolabs) cleavage, samples were incubated in G7 reaction buffer (New England Biolabs) with 2.5% Nonidet P-40 (New England Biolabs) and 500 units of PNGaseF for 24 h at 37°C.
SDS-PAGE, Coomassie Staining, Western Blot Analysis, and Charge-coupled Device Camera-based ECL Detection-Protein samples (ϳ20 g of total protein for cell lysates) were separated by SDS-PAGE using Mini-PROTEAN TGX 4 -15% precast gels (Bio-Rad). For MS, gels were stained with colloidal Coomassie (36). Otherwise, proteins were Western-blotted semidry on Immobilon-P PVDF (Millipore) membranes. Membranes were blocked with 5% BSA (Serva) and probed with anti-TRPP2 698 -799 diluted 1/400 in PBS (Biochrom). Immobilized antigens conjugated to horseradish peroxidase-labeled antibodies were detected using SuperSignal West Pico chemiluminescent substrate (Thermo Scientific). A 16-bit ChemoCam (Intas) was used for light detection. Images were taken within the dynamic range of the charge-coupled device sensor. None of the depicted bands were saturated.
Immunofluorescence and Live Cell Imaging-HeLa cells were cultivated in 6 wells (Greiner Bio-One) on cover glasses (Carl Roth) until ϳ80% confluence for transient transfection. 24 h after transfection, HeLa cells were either fixed or used for live cell imaging. For immunofluorescence staining, cells were rinsed twice with PBS (Biochrom) and fixed for 15 min at 21°C using 3% paraformaldehyde (Electron Microscopy Sciences) and 5% sucrose (Sigma-Aldrich) in PBS. After washing twice for 3 min with PBS, cells were blocked for 30 min at 21°C with 5% horse serum (Invitrogen) and 0.1% Triton X-100 (Merck) in PBS. Subsequently, samples were incubated for 3 min with PBS and for 1 h with primary antibodies (1/500 in PBS) at 21°C. Cells were then washed for 3 min with PBS and incubated with secondary antibodies (1/1000 in PBS) for 30 min at 21°C. After an additional 3 min at 21°C with PBS, Prolong Gold and DAPI (Invitrogen) antifade reagent were added. Finally, samples were transferred to microscope slides (R. Langenbrinck) and sealed with nail varnish (L'Oréal). Acidic organelles in transfected HeLa cells were visualized in live cells with Lysotracker Red DND-99 (Invitrogen) according to the instructions of the manufacturer. Microscopy images were recorded using an Axio Observer microscope (Zeiss).
Metabolic Labeling-Cells were cultured until ϳ80% confluent in DMEM minus-Met/Cys (Invitrogen) with 10% heat-inactivated FBS (Gemini Bioproducts). Subsequently, cells were incubated in medium plus 100 -200 Ci/ml [ 35 S]Met/[ 35 S]Cys (PerkinElmer Life Sciences), washed with PBS (Invitrogen), and then maintained in chase medium (DMEM (Invitrogen) with 10% FBS (Gemini Bioproducts)). Cells were then lysed, and the protein of interest was immunoprecipitated, followed by SDS-PAGE and Western blot analysis. Depending on the experiment, the beads were incubated with jack bean mannosidase (20 units/mg of protein, Sigma-Aldrich) prior to SDS-PAGE. Wherever specified, cells were preincubated with 2 mM N-butyldeoxynojirimycin (NB-DNJ, Toronto Research Chemicals) for 96 h or 5 g/ml tunicamycin (Sigma-Aldrich) for 2 h before and during labeling.
Small Molecule Inhibitors-Drug incubation of HeLa cells was started 24 h after transfection. Small molecule inhibitors were added to DMEM (Lonza) supplemented with 10% heatinactivated FBS (Biochrom). Cells were maintained in a humid-ified 10% CO 2 incubator at 36.5°C. Drug concentrations and incubation times were as follows: 200 M chloroquine (Sigma-Aldrich) or 6 mM ammonium chloride (Sigma-Aldrich) for 24 h, 210 M MG-132 (Enzo Life Sciences) for 6 h, and 18 M cycloheximide (Sigma-Aldrich) for 2-10 h. Cells were counted after harvesting using a TC10 automated cell counter (Bio-Rad). Sample volumes were normalized for cell count.
Glucosidase II Inhibition Assay-NB-DNJ (Toronto Research Chemicals) was added either to the cell lysate (8 -80 M) or to the cell medium (2 mM) for 24 -96 h prior to experiments. Glucosidase II activity was measured in the lysate of LLCPK cells. The substrate used was 4-methyl-umbelliferyl ␣-D-glucopyranoside (4MUG, Toronto Research Chemicals). Cell lysates were incubated with 1 mM 4MUG in a total reaction volume of 100 l for 30 min at 37°C. The reactions were stopped with 100 l of 2 M Tris, and the fluorescence was recorded immediately (excitation 365 nm and emission 455 nm).
Statistical Analysis-For data quantitation, at least three independent experiments were evaluated. Samples were normalized for ␤-actin or ␣-tubulin Western blot signals. Relative expression was calculated for samples on the same gel. Western blot signal intensities were measured using ImageJ. Student's t test was performed to assess statistical significance.

RESULTS
Native TRPP2 Is N-glycosylated-TRPP2 is a six-transmembrane (S1-S6) protein with a large extracellular loop between S1 and S2 (TRPP2 245-468 ), a pore-forming loop between S5 and S6 (TRPP2 634 -659 ), and cytosolic amino and carboxyl termini (TRPP2 1-223 and TRPP2 680 -968 , respectively) ( Fig. 1A) (7). Endogenous human TRPP2 has been shown to be heavily N-glycosylated (15,16). TRPP2 contains seven potential NX[ST] glycosylation sites: 3, 299, 305, 328, 362, 375, and 580 (NetNGlyc 1.0 Server) (2). To further study the N-glycosylation of TRPP2 in a native context, we immunoprecipitated TRPP2 from wild-type mouse kidneys and analyzed the immunoreactive profile to TRPP2 in a Western blot analysis. Invariably, two distinct bands were observed, one at ϳ120 kDa and one below 100 kDa. Enzymatic deglycosylation of TRPP2 immunoprecipitates with Endo H, which cleaves within the chitobiose core of high mannose glycans, led to an electrophoretic mobility shift of both bands to ϳ110 and ϳ80 kDa, respectively (Fig. 1B). N-Linked glycosylation contributes ϳ15 kDa to the molecular mass of mature TRPP2. Similar results were obtained for endogenous TRPP2 from murine cortical collecting duct M1 cells (Fig. 1C). Interestingly, the size of both TRPP2-reactive bands was decreased after deglycosylation, and the faster migrating band was significantly smaller than the predicted size of 109 kDa for unmodified murine TRPP2 (966 amino acids). The high similarity of the Endo H-mediated glycosylation pattern of TRPP2 from mouse kidney and M1 renal epithelial cells make the latter an attractive model system to investigate N-linked glycosylation in detail.
Identification of Five N-glycosylation Sites in the First Extracellular Loop of TRPP2-To determine the individual asparagine residues occupied with N-glycans, endogenous TRPP2 was immunoprecipitated and deglycosylated with Endo H. This generated truncated carbohydrate molecules with one N-acetylglucosamine residue remaining on the respective asparagine, i.e. a definite mass for N-linked glycosylation (asparagine ϩ N-acetylglucosamine). After SDS-PAGE and colloidal Coomassie staining, samples were subjected to MS analyses. TRPP2 peptides were identified in gel fragments corresponding to a molecular weight from 35-150 kDa (Fig. 1D). The obtained data represent 63% of TRPP2 amino acids, covering almost all of the protein except for hydrophobic parts (Fig.  1E). The computational analysis of MS/MS spectra identified four glycosylated asparagine residues in murine TRPP2: 297, 303, 326, and 360 (Fig. 2, A-F). The sites are evolutionarily conserved in vertebrates and correspond in human TRPP2 to amino acid positions 299, 305, 328, and 362. To validate N-linked glycosylation in TRPP2, two constructs were generated by site-directed mutagenesis, replacing the four asparagines with either glycine or glutamine: N299G/N305G/ N328G/N362G (TRPP2 ⌬4-Glyc ) and N299Q/N305Q/N328Q/ N362Q. However, upon heterologous expression in HEK293T and HeLa cells, these were still sensitive to Endo H-mediated deglycosylation, as seen by an electrophoretic mobility shift of TRPP2 bands in SDS-PAGE (ϳ3 kDa) (Fig. 3). A fifth N-glycosylation site was identified by taking advantage of the in silico predictions for TRPP2 (2). The additional mutation of asparagine 375 in TRPP2 (TRPP2 ⌬5-Glyc ), which is partially conserved in vertebrates, abrogates any size shift after enzyme-mediated deglycosylation of the protein (Fig. 4A).
The comprehensive analysis of N-linked glycosylation identified five widely conserved glycosylated asparagines in human TRPP2: 299, 305, 328, 362, and 375. Notably, all glycosylation sites are located in the first extracellular loop (Fig. 4B). These asparagines appear to be the only N-glycosylation sites of the protein because neither mass spectrometry nor enzymatic digestion with Endo H or PNGase F detected evidence for additional glycosidic bonds. Further in vitro analysis was facilitated by the recapitulation of native glycosylation patterns with highmannose glycans by heterologously expressed TRPP2 (Figs. 1B and 4A).
N-Glycosylation of TRPP2 Is Required for Adequate TRPP2 Protein Levels-It has been shown that N-linked glycans determine the homeostasis of secretory proteins and membrane proteins in eukaryotic cells (19). Therefore, we hypothesized that aberrant N-glycosylation of TRPP2 might impair its protein biogenesis. To test this hypothesis, TRPP2 ⌬5-Glyc was expressed heterologously in HeLa and HEK293T cells. Indeed, TRPP2 ⌬5-Glyc shows a significant reduction in protein levels by 86.2% compared with wild-type TRPP2 in HeLa cells (n ϭ 3, p ϭ 0.003) (Fig. 4C). This effect is not cell type-specific because a similar reduction is observed in HEK293T cells (mean, Ϫ74.4%; n ϭ 3; p ϭ 0.016) (Fig. 4D). To ensure that this decrease is not solely caused by a change in kinetics of TRPP2 protein processing, we measured protein levels after 96 h in HEK293T cells. TRPP2 ⌬5-Glyc protein levels compared with wild-type TRPP2 remain significantly reduced after 96 h (mean, Ϫ79%; n ϭ 4; p ϭ 0.00002).
Lower protein levels may be caused by either transcriptional down-regulation, impaired translation, or decreased protein stability. To evaluate a possible impact on mRNA transcription or stability, RNA of transiently transfected HeLa cells was iso-lated. TRPP2 wild-type and TRPP2 ⌬5-Glyc mRNA abundance was similar, as assessed by qPCR (Fig. 5A). Protein stability of TRPP2 was investigated using the Streptomyces griseus metabolite cycloheximide, which inhibits protein translation by blocking elongation factor 2-mediated ribosomal translocation (38). Cells were incubated with 18 M cycloheximide for 2-10 h, and TRPP2 expression was evaluated by Western blot analysis (Fig. 5, B and C). These experiments show a reduced stability of TRPP2 ⌬5-Glyc compared with wild-type TRPP2 after 10 h (mean, Ϫ46.8%; n ϭ 3; p ϭ 0.006) (Fig. 5D).
TRPP2 Is Subject to Lysosomal Degradation-Nascent polypeptides transit between a significant number of intermediate folding states before attaining their native state. It has been demonstrated that the fidelity of this process may be compromised by the absence of posttranslational glycosylation. Missing glycosylation may cause newly synthesized proteins to be degraded or to attain their native states more slowly (39). To investigate the cause for decreased levels of glycosylation-deficient TRPP2, small molecule inhibitors were employed.  TRPP2 is an N-glycosylated protein. A, TRPP2 is a six-transmembrane (S1-S6) protein with a large extracellular loop between S1 and S2 (TRPP2 245-468 ), a pore-forming loop between S5 and S6 (TRPP2 634 -659 ), as well as cytosolic amino and carboxyl termini (TRPP2 1-223 and TRPP2 680 -968 ). B, immunoprecipitation (IP) of native TRPP2 from wild-type C57BL/6 mouse kidneys were analyzed by anti-TRPP2 Western blot (WB) staining. Two distinct bands were observed, one at ϳ120 kDa and one below 100 kDa. Endo H-mediated deglycosylation led to an electrophoretic mobility shift of both bands to ϳ110 kDa and ϳ80 kDa, respectively. C, enzymatic deglycosylation of endogenous TRPP2 immunoprecipitates with Endo H from M1 renal epithelial cells. Similar to mouse kidneys, two distinct bands were observed. D, immunoprecipitation of native TRPP2 from M1 cells. After SDS-PAGE, samples were stained with colloidal Coomassie and cut into eight pieces. TRPP2 peptides were identified in gel fragments corresponding to a molecular mass from 35-150 kDa. IHC, immunoglobulin heavy chain; ILC, immunoglobulin light chain. E, total nano-LC-MS/MS coverage of murine TRPP2 was 63%. aa, amino acids.
Two major proteolytic pathways have been implicated in protein turnover in intact cells: lysosome-and proteasome-mediated degradation. Chloroquine is a weak base that accumulates in acidic cellular organelles, raises the pH, and, thereby, blocks lysosomal function. The peptidyl aldehyde MG-132 is a specific and potent inhibitor of the proteolytic activity of the 26 S proteasome complex (40). Wild-type TRPP2 is significantly more abundant in cells treated with lysosomotropic chloroquine (mean, 575.36%; n ϭ 4; p ϭ 0.0003). MG-132 shows no effect (Fig. 6A). Similarly, MG-132 has no effect on TRPP2 ⌬5-Glyc , but chloroquine also significantly increases TRPP2 ⌬5-Glyc protein levels (mean, 329.64%; n ϭ 4; p ϭ 0.048) (Fig. 6B). The observation that chloroquine increases TRPP2  protein levels suggests lysosome-mediated degradation as a pathway for TRPP2 degradation. However, chloroquine is known to have pleiotropic effects on cellular metabolism. To test the hypothesis that lysosomes degrade TRPP2, we inhibited lysosomal function using another structurally different compound, ammonium chloride, which neutralizes the acidic pH value of lysosomes. Both wild-type TRPP2 (mean, 203.07%; n ϭ 3; p ϭ 0.04) and TRPP2 ⌬5-Glyc (mean, 218.69%; n ϭ 3; p ϭ 0.02) protein levels rise under ammonium chloride treatment (Fig.  6C). Indirect immunofluorescence suggests a partial colocalization of TRPP2 and lysosomal LAMP1 (Fig. 7, A and B). However, the majority of TRPP2 is detected in a reticular pattern, consistent with the published ER localization of the overexpressed channel (41). The gross distribution, number, and appearance of lysosomes are not affected by transfection of TRPP2 or TRPP2 ⌬5-Glyc (Fig. 7, A-C).
Protein turnover is determined by the dynamic balance between protein synthesis and degradation. Our data support a critical role for lysosomes in the degradation of TRPP2. The relatively modest increase of TRPP2 ⌬5-Glyc protein levels upon lysosomal inhibition together with the insensitivity of TRPP2 ⌬5-Glyc expression to proteasomal inhibition suggests reduced translation efficiency as an important additional mechanism for reduced expression of TRPP2 ⌬5-Glyc .
Inactivation of Glucosidase II Results in Defects in N-Glycan Trimming of TRPP2-N-linked glycans are processed and monitored in the ER by enzymes and molecular chaperones (19). Because addition of N-linked oligosaccharides is required for normal protein levels and mutations in at least one protein involved in processing N-linked glycans, GII␤, result in reduced TRPP2 levels and polycystic kidney and liver disease, we next studied the effects of GII␤ on processing of the precursor gly- ) of wild-type TRPP2 causes a size decrease, that was slightly more pronounced after PNGase F treatment (24 h at 37°C). No electrophoretic mobility shift is observed after Endo H or PNGase F treatment of TRPP2 ⌬5-Glyc , suggesting that TRPP2 ⌬5-Glyc was not N-glycosylated. After immunoprecipitation (IP), normalized amounts of protein were used to visualize SDS-PAGE electrophoresis by anti-TRPP2 Western blot (WB) staining. B, illustrative schematic of TRPP2 N-glycosylation. All five N-linked glycosylation sites are located in the first extracellular loop (amino acids 244 -469) of human TRPP2: 299, 305, 328, 362, and 375 (corresponding to the conserved asparagines 297, 303, 326, 360, and 373 in murine TRPP2). C, heterologous expression of wild-type TRPP2 and TRPP2 ⌬5-Glyc in HeLa cells. Cells were counted to normalize for cell number. After lysis and SDS-PAGE, anti-TRPP2 and anti-␤-actin Western blot signal intensities were measured using ImageJ. Sample intensities were normalized for ␤-actin. Wild-type TRPP2 protein expression was set to 100%, and relative TRPP2 ⌬5-Glyc amounts were calculated. TRPP2 ⌬5-Glyc shows a significant reduction in protein levels by 86.2% compared with wild-type TRPP2 (n ϭ 3, p ϭ 0.003). D, overexpression of TRPP2 ⌬5-Glyc in HEK293T cells. Similar to C, TRPP2 ⌬5-Glyc shows a significant reduction in protein levels compared with wild-type TRPP2 (Ϫ74.4%, n ϭ 3, p ϭ 0.016).
can of TRPP2. TRPP2 and GII␤ interact physically in cells expressing tagged versions of both proteins (Fig. 8A) (21,20). Kidney-specific conditional deletion of GII␤ in mice using Ksp-Cre results in significantly reduced tissue levels of the TRPP2 protein in 2-month-old mice (mean, 66.75%; n ϭ 4; p ϭ 0.04) (Fig. 8, B and C). Because the Ksp-Cre is active in a portion of kidney tubule cells primarily in the distal nephron, the actual reduction in TRPP2 in cells lacking GII␤ is likely more pronounced than what can be detected at the whole organ tissue level.
We next wanted to directly examine the role of GII on the trimming of N-linked glycans in TRPP2. Because this required the sensitive determination of differences in the migration of TRPP2 on the basis of variable presence or absence of two glucose residues per N-linked glycan, we used a carboxyl terminaltruncated form of TRPP2, truncated at codon 703 (TRPP2 L703X-HA ), to allow us to better resolve migration differences on the basis of loss of GII activity. We first used a potent inhibitor of ␣-glucosidases, NB-DNJ, to establish our ability to reliably detect the lack of N-glycan deglycosylation in TRPP2 L703X-HA . NB-DNJ provides potent inhibitory activity when applied in vitro to cell lysates or in vivo to intact cells (Fig.  8, D and E). GII catalytic activity was measured by the fluorescence units released through the cleavage of the specific substrate 4MUG. The studies with the synthetic substrate establish 2 mM NB-DNJ as sufficient to provide at least 70% GII inhibi-tion in intact LLCPK1 cells. To examine the GII-dependent glycosylation status of TRPP2 L703X-HA , LLCPK1 cells stably expressing TRPP2 L703X-HA were incubated with 2 mM NB-DNJ for 96 h. Cells were pulsed for 30 min with medium containing [ 35 S]Met/[ 35 S]Cys and chased with non-radioactive medium for 30 min. After lysis, TRPP2 L703X-HA was immunoprecipitated using anti-HA antibodies and digested with jack bean mannosidase, which cleaves mannose residues from deglycosylated N-glycans but not from N-glycan moieties that were not deglycosylated because of inhibition of GII activity. Thus, the jack bean mannosidase treatment improves the resolution of the migrating protein species to differentiate deglycosylated forms from non-deglycosylated forms. Inhibition of GII by NB-DNJ resulted in a distinct electrophoretic mobility pattern for TRPP2 L703X-HA , migrating at a higher molecular weight compared with TRPP2 L703X-HA from untreated cells, in which GII was active to permit deglycosylation of the protein (Fig. 8F). This showed the ability to detect a defect in glucose trimming by this method.
GII function in vivo is dependent on GII␤ (42). We therefore tested the hypothesis whether GII␤ deficiency as seen in ADPLD results in defective glucose trimming of TRPP2. TRPP2 L703X-HA was transiently expressed in mouse kidney epithelial cells either wild-type (Prkcsh flox/flox ) or null (Prkcsh Ϫ/Ϫ ) for GII␤. Cells were labeled metabolically, lysed, immunoprecipitated with anti-HA, digested with jack bean mannosidase, FIGURE 5. TRPP2 ⌬5-Glyc protein stability is decreased significantly compared with wild-type TRPP2. A, in transiently transfected HeLa cells, mRNA expression of TRPP2 ⌬5-Glyc is similar to wild-type TRPP2, as assessed by qPCR. B, wild-type TRPP2 protein stability was probed using cycloheximide, which blocks protein translation. Cells were incubated with 18 M cycloheximide for 2-10 h, samples were normalized for cell count, and TRPP2 protein persistence was evaluated by Western blot (WB) analysis. ␤-Actin was used as loading control. C, TRPP2 ⌬5-Glyc protein stability was probed as described in B. D, group data from B and C. Western blot signal intensities at 0 and 10 h were measured using ImageJ. After 10 h of 18 M cycloheximide treatment, 95.5% of wild-type TRPP2 was still detected (n ϭ 3). TRPP2 ⌬5-Glyc , in contrast, shows a significant reduction of protein levels to 44.72% (n ϭ 3, p ϭ 0.006). and resolved by SDS-PAGE prior to autoradiography. TRPP2 L703X-HA expressed in Prkcsh Ϫ/Ϫ cells showed slower electrophoretic mobility compared with TRPP2 L703X-HA expressed in wild-type cells, indicating impaired glucose trimming in the absence of the non-catalytic GII␤ subunit (Fig. 8G). These data show that the enzymatic activity of GII is indeed impaired in the absence of GII␤ in ADPLD cells and leads to the accumulation of di-or monoglycosylated TRPP2 L703X-HA intermediates. Because all five N-linked glycosylation sites are located in the first extracellular loop (TRPP2 245-468 ) of TRPP2 and should not be affected by the TRPP2 L703X-HA truncation, it is likely that the observed impaired glycan trimming is analogous in wild-type TRPP2.
We identified five N-glycosylation sites in TRPP2 and showed that their glycosylation is required for adequate TRPP2 protein expression. Processing of these N-linked glycans is dependent on GII, highlighting the molecular association of the ADPKD protein TRPP2 and its ADPLD partner GII␤.

DISCUSSION
The addition and processing of N-linked oligosaccharides in the ER play a pivotal role in the biogenesis and quality control of many membrane proteins (43,18). It has been shown that TRPP2 is a glycoprotein, but the glycosylation sites and the importance of glycosylation for TRPP2 biogenesis and stability have not been investigated (15,41). In this study, we identified five N-glycosylation sites in TRPP2. All glycosylated asparagines in TRPP2 are found in the first extracellular loop (TRPP2 245-468 ) between the first two transmembrane domains (Asn-299, Asn-305, Asn-328, Asn-362, and Asn-375 in human TRPP2). These asparagines are probably the only N-linked glycosylation sites in TRPP2 because enzyme-mediated deglycosylation of TRPP2 with mutations in all five residues does not cause any additional electrophoretic mobility shift of the protein. Our data indicate that glycosylation of TRPP2 regulates the abundance of the channel complex and that deglycosylation results in inefficient translation and decreased stability. TRPP2 As a negative control, untreated (ø) cells were processed in parallel. Cells were counted after harvesting, and sample volume was normalized for cell count. ␤-Actin was used as the Western blot (WB) loading control. Wild-type TRPP2 protein expression in cells without drug treatment was set to 100%, and relative TRPP2 amounts after drug treatment were calculated. Wild-type TRPP2 was significantly more abundant in cells treated with chloroquine compared with control cells (mean ϭ 575.36%, n ϭ 4, p ϭ 0.0003). B, likewise, TRPP2 ⌬5-Glyc protein levels increased in the presence of chloroquine (mean, 329.64%; n ϭ 4; p ϭ 0.048). C, incubation of wild-type TRPP2 (mean, 203.07%; n ϭ 3; p ϭ 0.04) and TRPP2 ⌬5-Glyc (mean, 218.69%; n ϭ 3; p ϭ 0.02) with ammonium chloride significantly increased protein levels compared with untreated controls. appears to be degraded through lysosomes. Incubation of cells with the lysosomal inhibitor chloroquine leads to a significant increase of both wild-type and glycosylation-deficient TRPP2, whereas proteasomal inhibition with MG-132 shows no effect. In addition, we show that glucosidase II mediates trimming of TRPP2 glycosylation. This is required for efficient maintenance of steady-state levels of the protein. These combined results provide a mechanism for the decrease of TRPP2 levels in PRKCSH-deficient cells, which may contribute to the pathogenesis of ADPLD.
There is still rather limited information on the role of N-linked glycosylation of TRP channels. For a subset of TRP channels it has been shown that N-glycosylation plays an important role in cellular localization and ion channel function (31). In TRPV5, a channel critical for Ca 2ϩ reabsorption in the distal tubule of the kidney, cleavage of a single N-linked oligosaccharide by Klotho, a glucuronidase enzyme, stabilizes the channel at the apical plasma membrane (29). In TRPC6 and TRPC3, two closely related cation channels, differences in the N-glycosylation patterns are responsible for their distinct functional properties, switching the channel from a tightly receptorregulated mode to a constitutively active mode (27). In TRPV4, an osmosensitive TRP channel, mutation of a single N-glycosylation site within the pore-forming loop increases the expression of plasma membrane-bound channels compared with the wild type (30). The temperature-sensitive TRP channels TRPV1 and TRPM8 are also N-glycosylated at a position near the pore loop. Glycosylation of these asparagines affects the biophysical properties and response to canonical activators of both channels (44,45,32). All these TRP channels display glycosylation at one or two asparagines. TRPP2, however, is more heavily glycosylated akin to TRPML1, the channel mutated in mucolipidosis type IV (46,47). One hallmark of the TRPP and the TRPML subfamilies of TRP channels is the prominent first extracellular loop, which harbors the glycosylation sites ( Fig.  4B) (48). Interestingly, TRPML1 is cleaved within this first extracellular loop, and there is evidence that TRPP2 undergoes proteolytic cleavage as well (46,47,49). In addition to fulllength TRPP2, we invariably detect TRPP2 antibody-reactive bands on Western blot analyses that are considerably smaller than the calculated size of the full-length protein. This applies to native and overexpressed protein (Figs. 1, B and C, and 4, C and D). The biological function of cleavage and of the particularly big first extracellular loop of these TRP channels is not known. However, on the basis of data from other ion channels and membrane proteins, it is tempting to speculate that cleavage may be a regulatory mechanism that extends beyond mere inactivation of TRPP2 channels (50,51). Whether the biological roles of cleavage and glycosylation in the first extracellular loop intersect remains to be determined. Unfortunately, the impact of glycosylation on channel function and cellular localization of TRPP2 cannot be studied easily using heterologous expression systems because overexpressed wild-type TRPP2 is trapped in the ER and does not traffic to the plasma membrane (15,41). Although this precludes the straightforward investigation of the effect of glycosylation on the biophysical properties of TRPP2 using electrophysiological approaches, the biogenesis and degradation of the channel can be studied using heterologously expressed TRPP2.
The proteasome and lysosome are the main sites of cellular proteolysis (52). The majority of regulated degradation occurs through the ubiquitin-proteasome system in the cytoplasm. Proteins are marked for degradation by covalent attachment of multiple copies of ubiquitin molecules. The polyubiquitinated protein is recognized by the proteasome, a multisubunit protease that degrades the substrate into short polypeptides (53,54). In contrast, selective lysosomal protein degradation is achieved by controlling the trafficking of proteins into the acidic lysosome, where they are exposed to resident proteases. Trans-membrane proteins require special mechanisms for their degradation because different parts of such proteins are on opposite sides of a membrane. Eukaryotic cells have devised several solutions to this problem. Membrane proteins in the ER and in mitochondria are removed from the lipid bilayer by ret- FIGURE 8. Inactivation of glucosidase II results in defects in TRPP2 N-glycan trimming. A, TRPP2 and GII␤ interact physically. A HA epitope-tagged TRPP2 fusion protein (TRPP2 HA ) was coexpressed with either FLAG-tagged GII␤ (GII␤ FLAG ) or with the control protein GFP FLAG in HEK293T cells. After affinity purification with immobilized anti-FLAG antibodies, TRPP2 HA was detected by Western blotting (WB) in the immunoprecipitates (IP) of GII␤ FLAG . Expression of TRPP2 HA in the cell lysates was confirmed with anti-HA antibody. B, Ksp-Cre-mediated kidney-specific deletion of GII␤ results in significantly reduced TRPP2 protein levels in comparison with wild-type TRPP2 in 2-month-old mice. Western blot signal intensities were measured using ImageJ. Samples were normalized for ␣-tubulin. C, the mean reduction of relative TRPP2 protein levels was 66.75% (n ϭ 4, p ϭ 0.04). D, NB-DNJ-mediated inhibition of endoplasmic reticulum GII in vitro. LLCPK cell lysates were incubated with 8 -80 M for 24 -96 h prior to experiments. Cell lysates were then incubated with the GII substrate 4MUG (1 mM), and fluorescence (excitation 365 nm, emission 455 nm) was recorded immediately after stopping the reaction at the indicated time points. NB-DNJ (80 M) provided potent in vitro inhibition with a ϳ95% reduction in GII enzyme activity. DMSO, dimethyl sulfoxide. E, GII inhibition was assayed in vivo in live cells by application of 2 mM NB-DNJ to cell culture medium for 24 -96 h prior to experiments. Cells were lysed and assayed for GII activity using 4MUG (1 mM) at the indicated time points. GII activity was reduced by ϳ70%. F, autoradiogram of metabolically labeled anti-HA immunoprecipitates from cells following inhibition of GII by NB-DNJ treatment (NBDN, 2 mM for 96 h). The electrophoretic mobility pattern of TRPP2 L703X-HA at a higher molecular weight is the result of resistance to jack bean mannosidase digestion compared with TRPP2 L703X-HA of untreated cells. Tunicamycin (TUN, 5 g/ml), an inhibitor of N-linked glycosylation, served as a control, shifting TRPP2 L703X-HA toward a lower molecular weight product, indicative of the unglycosylated native protein. G, TRPP2 L703X-HA was transiently expressed in mouse kidney epithelial Prkcsh flox/flox (wt) and Prkcsh Ϫ/Ϫ (GII␤ Ϫ/Ϫ ) cells. Cells were labeled metabolically, lysed, immunoprecipitated with anti-HA, treated with jack bean mannosidase, and resolved by SDS-PAGE, followed by autoradiography. The absence of the non-catalytic GII␤ subunit resulted in an electrophoretic mobility shift to a larger molecular weight similar to the inhibition of the GII by NB-DNJ, indicating that loss of GII␤, as occurs in ADPLD, results in a glucose trimming defect in N-glycans. rotranslocation into the cytoplasm, where they are ubiquitinated and degraded by the proteasome (43). In contrast, integral membrane proteins from the plasma membrane and the Golgi apparatus are collected in vesicles and sent to multivesicular bodies via endosomes. Fusion of multivesicular bodies with lysosomes leads to degradation of proteins and lipids by lysosomal proteases and lipases (55).
The pathways responsible for the degradation of proteins can be studied by incubating cells with specific and selective inhibitors of proteasomes or lysosomes, such as MG-132 or chloroquine, respectively. We used these inhibitors to investigate where wild-type and non-glycosylated TRPP2 are degraded. Because overexpressed TRPP2 is an ER-resident protein and because glycosylation has been shown to be important for folding and quality control of membrane proteins, we speculated that TRPP2 is likely degraded by the ubiquitin-proteasome system via ER-associated degradation (43). Indeed, previous studies have shown that TRPP2 is ubiquitinated (56,21). Surprisingly, inhibition of the proteasome with MG-132 did not show any effect on wild-type or glycosylation-deficient TRPP2 levels. Inhibition of lysosomal degradation with chloroquine or ammonium chloride, however, resulted in a significant increase of wild-type and glycosylation-deficient TRPP2 levels. Because overexpressed TRPP2 has been shown to be fully Endo H-sensitive and because immunofluorescence shows exclusive ER localization of the channel, lysosomal degradation of the protein appears to be rather counterintuitive (15,41). It is conceivable, though, that transient trafficking of TRPP2 to the Golgi apparatus and plasma membrane as well as to the lysosomal compartment cannot be detected under steady-state conditions because the retrieval of the channel to the ER profoundly exceeds its forward transport. It has been reported that wildtype TRPP2 can be found at the plasma membrane under certain circumstances (57,58). Furthermore, truncation mutations of the carboxyl terminus that delete an ER retention/ retrieval motif in TRPP2 can be detected at the plasma membrane (15,10,59). A recent study has shown that TRPP2 takes different routes to the plasma membrane and the membrane of the primary cilium (59). Notably, this study showed that TRPP2 does not traverse the Golgi apparatus on its way to the primary cilium, raising the possibility that similar mechanisms apply to lysosomal targeting. The stability of TRPP2 ⌬5-Glyc is decreased significantly, as shown in the experiments with cycloheximide-treated cells (Fig. 5D). However, it should be noted that lysosomal degradation may not explain the strong decrease of glycosylation-deficient TRPP2 because the increase in protein levels upon lysosomal inhibition is not increased for TRPP2 ⌬5-Glyc compared with wild-type TRPP2. The relatively modest increase of TRPP2 ⌬5-Glyc protein levels upon inhibition of lysosomes suggests reduced efficiency of translation and folding of TRPP2 as an important additional mechanism for low-level expression of TRPP2 ⌬5-Glyc .
The general importance of N-glycosylation in protein function is highlighted by severe and pleiotropic clinical phenotypes of patients with impaired glycosylation. These disorders are caused by several known heritable defects in N-glycan synthesis and processing (60,18). Glycosylation begins with the cotrans-lational addition of a preformed, 14-residue precursor glycan to asparagine residues that are part of the consensus sequence NX [ST]. Processing of the core glycan commences with trimming of two of the three terminal glucose residues by ER glucosidases I and II in the calnexin/calreticulin cycle (17,18). Mutations in these glucosidases result in human disease. The non-catalytic ␤ subunit of glucosidase II, encoded by the PRKCSH gene, is of particular interest to glycan processing of TRPP2 because mutations in this gene cause ADPLD (22)(23)(24). Mutations in PRKCSH manifest with bile duct cysts indistinguishable from the liver phenotype in ADPKD. It has been shown that PRKCSH is part of a genetic interaction network with several PKD genes, including PKD1 and PKD2 (20). This study established that polycystin 1 is the rate-limiting component in cyst formation in human polycystic kidney and liver disease. Furthermore, it has been shown that GII␤ is required for adequate expression of polycystin 1 and TRPP2 (21,20). Whether TRPP2 is a substrate for glucosidase II has not been addressed previously. Here we show that GII is required for glycan trimming of TRPP2 using pharmacological and genetic approaches. This observation provides a mechanism for the reduction of TRPP2 expression in cells lacking GII␤. It should be noted that the reduction of TRPP2 protein levels in GII␤ mutant kidneys was performed with Ksp-Cre-mediated deletion of GII␤. This approach leads to deletion of the gene in less than 30% of the TRPP2-expressing cells in the kidney. 5 Therefore, the reduction in TRPP2 expression depicted in Fig. 8, B and C, would probably be more profound in Prkcsh null kidneys. This is supported by the observation that shRNA-mediated knockdown of PRKCSH in renal epithelial cells leads to a considerably stronger reduction of TRPP2 expression (21). Although it has been shown convincingly that polycystin 1 is the central component for cyst formation, the reduction of TRPP2 expression in cells lacking GII␤ also contributes to cyst formation or modulates the severity of the disease.
In summary, we demonstrate that TRPP2 is glycosylated at five asparagines in the first extracellular loop. N-glycosylation and glycan trimming through glucosidase II regulate the abundance of TRPP2, highlighting the importance of these posttranslational modifications for the biogenesis and stability of the protein. These results provide a basis for further studies addressing additional roles of TRPP2 glycosylation in health and disease.