N-Linked Protein Glycosylation Is a Major Determinant for Basal TRPC3 and TRPC6 Channel Activity*

The TRPC family of receptor-activated cation channels (TRPC channels) can be subdivided into four subfamilies based on sequence homology as well as functional similarities. Members of the TRPC3/6/7 subfamily share common biophysical characteristics and are activated by diacylglycerol in a membrane-delimited manner. At present, it is only poorly understood whether members of the TRPC3/6/7 subfamily are functionally redundant or whether they serve distinct cellular roles. By electrophysiological and fluorescence imaging strategies we show that TRPC3 displays considerable constitutive activity, while TRPC6 is a tightly regulated channel. To identify potential molecular correlates accounting for the functional difference, we analyzed the glycosylation pattern of TRPC6 compared with TRPC3. Two NX(S/T) motifs in TRPC6 were mutated (Asn to Gln) by in vitro mutagenesis to delete one or both extracellular N-linked glycosylation sites. Immunoblotting analysis of HEK 293 cell lysates expressing TRPC6 wild type and mutants favors a model of TRPC6 that is dually glycosylated within the first (e1) and second extracelluar loop (e2) as opposed to the monoglycosylated TRPC3 channel (Vannier, B., Zhu, X., Brown, D., and Birnbaumer, L. (1998) J. Biol. Chem. 273, 8675–8679). Elimination of the e2 glycosylation site, missing in the monoglycosylated TRPC3, was sufficient to convert the tightly receptor-regulated TRPC6 into a constitutively active channel, displaying functional characteristics of TRPC3. Reciprocally, engineering of an additional second glycosylated site in TRPC3 to mimic the glycosylation status in TRPC6 markedly reduced TRPC3 basal activity. We conclude that the glycosylation pattern plays a pivotal role for the tight regulation of TRPC6 through phospholipase C-activating receptors.


From the Institut fü r Pharmakologie und Toxikologie, Philipps-Universitä t-Marburg, 35033 Marburg, Germany
The TRPC family of receptor-activated cation channels (TRPC channels) can be subdivided into four subfamilies based on sequence homology as well as functional similarities. Members of the TRPC3/6/7 subfamily share common biophysical characteristics and are activated by diacylglycerol in a membrane-delimited manner. At present, it is only poorly understood whether members of the TRPC3/6/7 subfamily are functionally redundant or whether they serve distinct cellular roles. By electrophysiological and fluorescence imaging strategies we show that TRPC3 displays considerable constitutive activity, while TRPC6 is a tightly regulated channel. To identify potential molecular correlates accounting for the functional difference, we analyzed the glycosylation pattern of TRPC6 compared with TRPC3. Two NX(S/T) motifs in TRPC6 were mutated (Asn to Gln) by in vitro mutagenesis to delete one or both extracellular N-linked glycosylation sites. Immunoblotting analysis of HEK 293 cell lysates expressing TRPC6 wild type and mutants favors a model of TRPC6 that is dually glycosylated within the first (e1) and second extracelluar loop (e2) as opposed to the monoglycosylated TRPC3 channel (Vannier, B., Zhu, X., Brown, D., and Birnbaumer, L. (1998) J. Biol. Chem. 273, 8675-8679). Elimination of the e2 glycosylation site, missing in the monoglycosylated TRPC3, was sufficient to convert the tightly receptor-regulated TRPC6 into a constitutively active channel, displaying functional characteristics of TRPC3. Reciprocally, engineering of an additional second glycosylated site in TRPC3 to mimic the glycosylation status in TRPC6 markedly reduced TRPC3 basal activity. We conclude that the glycosylation pattern plays a pivotal role for the tight regulation of TRPC6 through phospholipase C-activating receptors.
Following binding to cognate receptors on the cell membrane, many hormones, neurotransmitters, and growth factors elicit increases in the intracellular free calcium concentration ([Ca 2ϩ ] i ) subsequent to phospholipase C activation. In addition to inositol 1,4,5-trisphosphate (InsP 3 ) 1 -mediated calcium re-lease from intracellular stores, Ca 2ϩ -permeable plasma membrane ion channels are activated in a PLC-dependent manner in most cells either by second messenger-mediated pathways or by store depletion (1). To date, seven mammalian homologues of the Drosophila melanogaster transient receptor potential (TRP) visual transduction channels (2), the classical TRPC channel proteins (TRPC), have been identified. They were characterized as subunits of receptor-activated (ROCs) and or storeoperated (SOCs) cation channels (see Refs. 3-6 for recent reviews). TRPC proteins are thought to span the plasma membrane 6 times with a pore loop inserted between transmembrane segments 5 and 6 (7), and functional channel complexes are believed to be homo-and heterotetrameric. In principle, all TRPC proteins combine into multimeric complexes within the confines of TRPC subfamilies as assessed in living cells (8).
Based on structural and functional criteria, TRPC3, -6, and -7 form a distinct subgroup within the TRPC family. The primary amino acid sequences of TRPC3, -6, and -7 are ϳ75% identical, and the latter proteins represent non-selective cation channels characterized by both inward and outward rectification. A functional hallmark of the TRPC3, -6, -7 subfamily is their membrane-delimited activation by diacylglycerols (DAGs) independent of protein kinase C (9,10). Furthermore, in a heterologous expression system (8) as well as in rat brain synaptosomes (11) TRPC3, -6, and -7 interact with each other, but not with other members of the TRPC family (8). Interestingly, both TRPC3 and TRPC6 are expressed in brain (12,13). TRPC3 expression in the brain is most prominent around the time of birth, and in pontine neurons channel activation occurs through a signaling pathway initiated by binding of brainderived growth factor to its cognate receptor, the transmembrane receptor tyrosine kinase TrkB (14). The role of TRPC6 or putative TRPC3/6 heteromultimers in brain still remain elusive. TRPC3 was originally cloned from human embryonic kidney cells (HEK 293 cells) (12) and functionally contributes to calcium permeable ion channels gated by store depletion in these cells (15). Low level stable expression of TRPC3 in HEK 293 cells gave rise to ion currents only when IP 3 is bound to its receptor and internal calcium stores are depleted (16). At the molecular level, interacting domains of the IP 3 receptor and TRPC3 have been characterized in great detail, thus entertaining the notion of store-dependent activation by a conformational coupling mechanism (Ref. 17 and see Ref. 18 for a recent review) and inhibition by Ca 2ϩ /calmodulin (19). Although TRPC3 expressed in DT40 avian B-lymphocytes show activation by store depletion (20), the latter concept is not undisputed, and a predominant store-operated activation mecha-* 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.
In the present study we address the question of whether TRPC3 and TRPC6 are functionally redundant or whether they display distinct functional characteristics. Thus, we monitored the basal activity of TRPC3 and compared it to that of TRPC6 by analysis of TRPC3-and TRPC6-expressing HEK 293 cells in Mn 2ϩ -quench experiments and electrophysiological recordings. In contrast to TRPC6, which turns out to be a tightly receptorand phospholipase C-regulated channel, TRPC3 shows pronounced basal channel activity despite a high degree of overall amino acid identity between the two proteins (5). To identify the underlying cause for the different functional characteristics, we examined the glycosylation pattern of TRPC6 and identified two glycosylated asparagine residues as opposed to the monoglycosylated TRPC3 channel (7). Elimination of the distal glycosylation site in TRPC6 (N561Q) by in vitro mutagenesis was sufficient to convert the tightly regulated TRPC6 protein into a constitutively active channel, resembling TRPC3 in this regard. Conversely, engineering of a second distal glycosylation site into TRPC3 (E512N) reduced basal activity. Thus, post-translational modifications of TRPC3 and TRPC6 determine distinct functional characteristics of genetically closely related cation channels.

EXPERIMENTAL PROCEDURES
Materials-All cell culture media and supplements were purchased from PAA laboratories (Linz, Austria) Cell Culture and Transient Transfections-HEK 293 cells were grown in Earle's medium supplemented with 10% fetal calf serum, 50 units/ml penicillin, and 50 units/ml streptomycin. For transient transfections, cells were seeded into 6-well dishes on glass coverslips and transfected the following day (80% confluency) with Metafectene reagent (Biontex, Munich, Germany). The concentration of transfected plasmid DNA was kept constant by combining 0.2 g/dish green fluorescent protein (GFP) plasmid (Invitrogen, San Diego), 0.2 g/dish of the guinea pig histamine H 1 receptor in pcDNA3 and 1.6 g/dish of the indicated TRPC subunit in pcDNA3 together with 6 l of Metafectene reagent in 100 l of serum-free medium. The mixture was added to the medium of the cells, which were analyzed 48 h later.
Generation of Fluorescent Fusion Proteins and TRPC6GM1/2/1.2-The C-terminal fusion of hTRPC6 and hTRPC3 with green (GFP) and yellow fluorescent protein (YFP), respectively, is described elsewhere (8). Glycosylation mutants of TRPC6 were produced by using the Quick-Change XL mutagenesis kit (Stratagene Europe, Amsterdam, The Netherlands) and the following oligonucleotides (mutated nucleotides are underlined): GM1 (N473Q) forward primer, 5Ј-GGC ACA AAA CTC CTT CCT CAA GAA ACC AGC ACA GAT, reverse primer, 5Ј-ATC TGT GCT GGT TTC TTC AGG AAG GAG TTT TGT GCC; GM2 (N561Q) forward primer, 5Ј-GC ATC:ATT GAC GCA CAA GAT ACT TTG AAG GAC TTG ACG; reverse primer, 5Ј-CGT CAA GTC CTT CAA AGT ATC TTG TGC GTC AAT GAT GC. The double mutant GM1.2 (N473Q/561Q) was constructed by amplifying the GM1 mutant with primers specific for GM2. The TRPC3Gly2 mutant (E512N) was constructed by a twostep PCR using the following primers: product A forward primer, 5Ј-GGA AGG ACC TAG GGA ATA CA, reverse primer, 5Ј-CTC TGG TGG GAG TGT CAC GTT ACT GAG GTC ACT CTC TTG; product B forward primer, 5Ј-CAA GAG AGT GAC CTC AGT AAC GTG ACA VCTC CCA GAG, reverse primer, 5Ј-TGG GAA AGT TAA CAA TTC GCA TGA TG. Products A and B were mixed and amplified by using the product A forward primer and the product B reverse primer. The resulting PCR product was digested with the restriction enzymes AvrII and HpaI and used to replace the corresponding fragment in the TRPC3 cDNA. All mutations were verified by sequencing. YFP-or GFP-tagged TRPC proteins were used for detection in immunoblots and in confocal microscopy, whereas all functional data were acquired with untagged counterparts.
Tunicamycin Treatment and Immunoblotting-HEK 293 cells were transiently transfected with cDNAs coding for the indicated TRPC subunit in pcDNA3 (3 g) using 9 l of the FuGENE 6 reagent (Roche Applied Science, Mannheim, Germany) according to the manufacturer's recommendations. 14 h after transfection one-fifth of the medium was replaced with medium containing 0.1% Me 2 SO or tunicamycin (10 M, Calbiochem, San Diego) in 0.1% Me 2 SO giving rise to a final tunicamycin concentration of 2 M in 0.02% Me 2 SO. Cells were washed in phosphate-buffered saline 36 h after transfection and resuspended in a buffer containing 20 mM Tris/HCl (pH 7.5) and protease inhibitors provided through MiniProtean tablets (Roche Applied Science). After repetitive aspiration through a 26-gauge needle, one volume of 2ϫ Laemmli buffer was added. The cell lysate was heated (65°C for 5 min), and proteins were resolved by 7% SDS-PAGE. After electroblotting to nitrocellulose (Hybond N, Amersham Biosciences, Freiburg, Germany), membranes were blocked in phosphate-buffered saline supplemented with 5% nonfat dry milk for 1 h and incubated with an anti-TRPC6 (24) or an anti-TRPC3 antibody (14) (1:1000 in phosphate-buffered saline supplemented with 0.05% Tween 20). After washing with phosphatebuffered saline/0.05% Tween 20 (3 ϫ 15 min), membranes were incubated with a secondary peroxidase-coupled goat anti-rabbit Ig-G (1: 2000 in blocking medium; Sigma, Taufkirchen, Germany) for 2 h, washed as before, and proteins were visualized by chemiluminescence (ECL; Amersham Biosciences, Freiburg, Germany).
Assessment of Mn 2ϩ Entry-Mn 2ϩ quenching experiments were conducted 48 h after transfection as described previously (9). Briefly, cells were loaded with fura-2-acetoxymethyl ester (5 M) in HEPES-buffered saline (HBS: 140 mM NaCl, 6 mM KCl, 1.25 mM MgCl 2 , 2 mM CaCl 2 , 10 mM HEPES, 5 mM glucose and 0.1% bovine serum albumin, pH 7.4) at room temperature for 45 min. Cells of interest were selected and analyzed using a Polychrome II monochromator and an IMAGO CCD camera (Till Photonics, Martinsried, Germany) coupled to an inverted Olympus IX70 microscope. Cells expressing GFP were selected at 470 nm and analyzed in Mn 2ϩ quench experiments at the isosbestic wavelength of fura-2 (360 nm in our system) and at 340 and 380 nm for measuring [Ca 2ϩ ] i . Cells were superfused with 2 or 0.2 mM MnCl 2 in HBS and 100 M histamine in HBS/Mn 2ϩ at the time points indicated. Fluorescence signals of cells expressing GFP and the histamine receptor were recorded at 330-ms intervals at 360 nm and normalized to 100% of the initial value. Basal decreases of fluorescence after adding Mn 2ϩ were calculated in percent/s, and initial decreases due to bleaching were subtracted. Each experiment was performed at least three times with similar results.
Electrophysiological Techniques-HEK 293 cells were transiently transfected with cDNAs coding for the indicated TRPC subunits in pcDNA3 (2 g), guinea pig H 1 histamine receptor in pcDNA3 (0.5 g) and GFP reporter plasmid (0.1 g) using FuGENE 6 reagent (7.8 l). For TRPC3 the clone published by Zitt et al. (21) was used and compared with the one published by Zhu et al. (12). No functional differences were observed. The TRPC6GM1.2 and TRPC6GM2 mutants were subcloned into the pIRES2-EGFP vector (Invitrogen, San Diego) containing the coding sequence for the enhanced green fluorescent protein (eGFP) downstream of the internal ribosome entry site (IRES) and cotransfected with the guinea pig H 1 receptor (0.5 g) as described above. Whole cell patch clamp recordings were carried out at room temperature (23°C) 12-48 h after transfection, except for the TRPC6GM1.2 and TRPC6GM2 mutants, which were analyzed 48 -120 h after transfection. Cells were superfused with bath solution (in mM): NaCl 140, CsCl 5, CaCl 2 2, MgCl 2 1, glucose 10, HEPES 10, buffered to pH 7.4 with NaOH. Data of whole cell recordings were collected with an EPC9 patch clamp amplifier (HEKA, Lambrecht, Germany) using the Pulse software. Patch pipettes made of borosilicate glass (Science Products, Hofheim, Germany) had resistances of 2.5 to 3.5 M⍀ when filled with intracellular solution (in mM): CsCl 120, NaCl 9.4, MgCl 2 1, GTP 0.2, 100 nM free Ca 2ϩ buffered with BAPTA 10, HEPES 10, titrated to pH 7.2 with CsOH. The liquid junction potential was ϩ4.3 mV and corrected for by the Pulse software. Current-voltage relations were obtained from triangular voltage ramps from Ϫ100 to ϩ60 mV with a slope of 0.4 V s Ϫ1 applied at a frequency of 1 Hz. Data were acquired at a frequency of 5 kHz after filtering at 1.67 kHz. Inside-out singlechannel measurements were carried out 48 -96 h after transfection with TRPC6 or TRPC6GM2 cDNAs as described above with a pipette solution (in mM): CsCl 120, NaCl 10, MgCl 2 1, buffered at 100 nM free Ca 2ϩ with BAPTA 10, HEPES 10, titrated to pH 7.2 with CsOH and a bath solution containing (in mM): CsCl 120, NaCl 10, MgCl 2 1, glucose 10, buffered at 200 nM free Ca 2ϩ with EGTA 10, HEPES 10, titrated to pH 7.2 with CsOH. The liquid junction potential was ϩ0.5 mV. Data acquisition was performed with an EPC9 and EPC10 patch clamp amplifier using the Pulse software. For the evalution of channel activity, defined as NPo (the product of the number of channels and open probability) PC DAC 1.1.4 of Marburg University Software Team was FIG. 1. Basal and receptor-mediated activity of TRPC3 and TRPC6. A, electrophysiological recordings. HEK 293 cells were transiently cotransfected with plasmids encoding eGFP, the H 1 histamine receptor and human TRPC3 or human TRPC6. Current-voltage relationships were analyzed by patch clamp recordings in the whole cell mode in non-transfected unstimulated cells and in transfected cells before (basal) and after application of receptor agonist (histamine). Insets, continuous recordings of TRPC3-and TRPC6-expressing cells before and after application of histamine. Duration of histamine applications is indicated by bars. B, current-voltage relationships (n ϭ 11) of TRPC3-expressing HEK 293 cells were analyzed by patch clamp recordings in the whole cell mode after application of receptor agonist (100 M histamine). Reversal potentials were used. NPo values were calculated for voltage ladders from Ϫ60 to ϩ60 mV, recorded in 1-s steps. 1 M 1-stearoly-2-arachidonoyl-sn-glycerol (SAG), a potent non membrane-permeable diacylglycerol derivative, was added to stimulate the channels in inside-out patches. The sampling rate was 15 kHz and the data were filtered with a Ϫ3dB cut-off frequency of 5 kHz. Only patches with baseline noise of Ͻ560 fA r.m.s. were used.
Statistical Analysis-Data are presented as means Ϯ S.E. of the mean (S.E.). In the case of Mn 2ϩ quench measurements, one representative experiment out of three is shown. Unless stated otherwise, data were compared by an unpaired Students' t test, if a Gaussian distribution was confirmed by applying a Kolmogorov-Smirnov (normality) test, and significance was accepted at p Ͻ 0.05. For multiple comparisons a one-way ANOVA was applied.

RESULTS
Different Basal Activities of TRPC3 and TRPC6 -To address the issue as to whether TRPC3 and TRPC6 are functionally redundant or whether they have discernible characteristics, we set out to compare basal channel activities. To this end, human TRPC3 or TRPC6 were transiently coexpressed with the H 1 histamine receptor in HEK 293 cells. Current-voltage relationships before and after application of histamine to the bath solution were recorded in whole cell patch clamp experiments (Fig. 1A). In both TRPC3-and TRPC6-expressing cells H 1 receptor activation gave rise to rapidly developing transient outward and inward currents at holding potentials of ϩ60 and Ϫ60 mV (Fig. 1A, insets), respectively. The current-voltage relations revealed rectification in the inward and outward direction, a property which is consistently encountered with members of the TRPC3/6 subfamily (25). Most notably, basal TRPC3 activity in the absence of histamine was higher than that of TRPC6 when recorded both at negative and at positive potentials (Fig. 1A). Reversal potentials for TRPC3-expressing cells were Ϫ2.1 Ϯ 0.8 and Ϫ1.7 Ϯ 1.1 mV before and after histamine application, respectively, and Ϫ1.5 Ϯ 1.0 and Ϫ2.6 Ϯ 1.0 mV for TRPC6-expressing cells and were not significantly different (p Ͼ 0.05 One-way ANOVA). By plotting reversal potentials of TRPC3-expressing cells against current densities at ϩ60 mV and at Ϫ60 mV no correlation between both values was detected (Fig. 1B). When basal TRPC3 and TRPC6 activities at Ϫ60 and ϩ60 mV were normalized to maximal currents in response to histamine challenge monitored at the same holding potentials, TRPC3-expressing cells displayed significantly (p Ͻ 0.002) increased constitutive channel activities (-46.0 Ϯ 5.0% and 48.7 Ϯ 4.3% at Ϫ60 and ϩ60 mV, respectively) as compared with HEK 293 cells studied after transfection of TRPC6 cDNA (-10.6 Ϯ 2.3% and 20.9 Ϯ 4.3% at Ϫ60 and ϩ60 mV, respectively) (Fig. 1C). The functional differences between the two TRPC proteins did not depend on the type of G q/11 -coupled receptor coexpressed, because TRPC3 and TRPC6 could likewise be distinguished by means of their discrepant basal activities in HEK 293 cells expressing m5 muscarinic instead of H 1 histamine receptors (data not shown).
By adding Mn 2ϩ to the bath solution, the influx of extracellular cations into single TRPC protein-expressing cells can be monitored by the quenching of fura-2 fluorescence by Mn 2ϩ ions recorded at the isosbestic wavelength. Subsequent to the addition of 2 mM Mn 2ϩ to the extracellular medium, a profound acceleration of the basal decrease of fura-2 fluorescence (Ϫ0.13 Ϯ 0.01%/s) was observed only in TRPC3-expressing cells (Fig. 1D). On the contrary, the presence of TRPC6 did not give rise to increased fura-2 quenching by Mn 2ϩ ions (Ϫ0.034 Ϯ 0.008%/s, similar to untransfected cells, data not shown), thus revealing a significantly reduced (p Ͻ 0.001) basal activity. These data lend further credence to the concept of high basal TRPC3 activity as opposed to the quiescence of TRPC6 in the absence of phospholipase C activation.
Glycosylation Status of TRPC3 and TRPC6 -TRPC3 and TRPC6 are closely related members of the same subfamily of TRPC proteins. We hypothesized that in addition to differences in the primary structure post-translational modifications like protein glycosylation may also contribute to the distinctive functional properties of the two cation channels. Therefore, we searched for potential glycosylation consensus sites (NX(S/T)) in the TRPC6 primary amino acid sequence. We detected seven potential glycosylation sites, only two of which are predicted to face the extracellular space as deduced from the TRPC3 transmembrane topology proposed by Vannier et al. (7). Thus, the N-terminal site, Asn-473, appears to be located in the first extracellular loop (e1) and corresponds to the single glycosylated asparagine residue in TRPC3 (7), while Asn-561 in the second extracellular loop (e2) has no counterpart in the TRPC3 primary amino acid sequence. To test whether these putative glycosylation sites were indeed post-translationally modified in TRPC6 and whether such modifications would entail any functional consequences, we initially created three glycosylation mutants (TRPC6GM1[N473Q], TRPC6GM2[N561Q], TRPC6GM1.2[N473/561Q]) as illustrated in Fig. 2A. In order to minimize the possibility of conformational disruption and to preserve the local charge distribution within the ion channel proteins, we exchanged Asn for structurally very similar Gln residues differing only by one methylene group in the amino acid side chains. TRPC6 and its three mutated versions Cterminally fused to GFP were expressed in HEK 293 cells and analyzed by immunoblotting (Fig. 2B) and confocal laser-scanning microscopy (Fig. 2C).
When lysates of wild-type TRPC6GFP-expressing cells were resolved by SDS-PAGE and probed with an anti-TRPC6 antibody, two specific bands of ϳ120 and 160 kDa were detected (Fig. 2B). The diffuse band migrating with a higher molecular mass disappeared upon preincubation of cells with the glycosylation inhibitor tunicamycin, commensurate with the assumption that it represented a glycosylated form of TRPC6, while the 120-kDa band conforms to the non-glycosylated ion channel protein. Bands with apparent molecular masses around 80 kDa were also detected in non transfected control cells (left panel of Fig. 2B). Elimination of the N-terminal glycosylation site in TRPC6GM1 did not prevent the appearance of a diffuse slowly migrating, tunicamycin-sensitive band of ϳ140 kDa indicating that protein glycosylation still occurred at the C-terminal site in e2. Reciprocally, removal of the Cterminal Asn-561 in TRPC6GM2 did not abolish protein glycosylation, because the relevant Asn residue in e1 could still be post-translationally modified. Only lysates of cells expressing TRPC6GM1.2 in which both possible glycosylation sites were eliminated, did no longer show slowly migrating tunicamycinplotted against current densities at ϩ60 mV (black squares) and at Ϫ60 mV (gray circles). C, summary of basal activities of TRPC3-(n ϭ 11) and TRPC6-(n ϭ 12) expressing cells analyzed at Ϫ60 and ϩ60mV. Values were normalized to histamine-induced currents at the aforementioned holding potentials. D, Mn 2ϩ quench experiments. HEK 293 cells expressing TRPC3 (n ϭ 16) or TRPC6 (n ϭ 12) were loaded with fura-2 and mounted onto a microscope connected to a fluorescence imaging system. Mn 2ϩ (2 mM) was added to the Ca 2ϩ -containing bath solution as indicated by bars. Time course of total fura-2 fluorescence at the isosbestic wavelength is shown. The black lines represent calculated means, whereas the gray areas indicate standard errors of the mean (Ϯ S.E.). Total fura-2 fluorescence corrected for background signals were recorded by exciting fura-2 at the isosbestic wavelength (360 nm) and normalized to the initial values of each single cell. Stippled lines indicate the initial slopes, i.e. the rates of fura-2 quenching before addition of Mn 2ϩ . responsive protein bands. Thus, in contrast to TRPC3 modified only at one site in e1, our results strengthen the notion of a dually glycosylated TRPC6 channel. As protein glycosylation is known to have a profound impact on proper folding and plasma membrane targeting, we asked whether the glycosylation mutants would be correctly inserted into the cell membrane and assessed the cellular localization of TRPC6 proteins C-terminally fused to GFP by confocal laser microscopy. As shown in Fig. 2C, the cellular distribution of the three TRPC6 mutants was indistinguishable from the wild-type channel, and the overlay of fluorescence and differential interference contrast images demonstrated that all TRPC6 proteins were correctly targeted to the plasma membrane.
Increased TRPC6 Basal Activity of TRPC6 Glycosylation Mutants-To evaluate the effect of protein glycosylation on TRPC6 basal activity, TRPC6GM2, devoid of N-linked glycosylation in e2, thereby mimicking the glycosylation status of TRPC3, and TRPC6GM1.2 lacking any oligosaccharide modifications (Fig.  2, A and B) were transiently expressed in HEK 293 cells in conjunction with the G q/11 -coupled H 1 histamine receptor. Ionic currents before and after addition of histamine were examined with the whole cell patch clamp technique. When compared with wild-type TRPC6 (Fig. 1A, lower panel), the current-voltage relationships of HEK 293 cells expressing either glycosylation mutant revealed increased basal channel activity, which was most conspicuous at negative membrane potentials (Fig.  3A). Histamine stimulation further increased inward and outward whole cell currents.
When comparing the current-voltage relationships characteristic for TRPC6 ( Fig. 1A) with those obtained with the two glycosylation mutants (Fig. 3A), we noted that apart from differences in basal channel activity outward rectification upon histamine stimulation was reduced in the case of the mutant TRPC6 proteins. Therefore, it appeared deemed legitimate to ask whether our normalization procedure might systematically overestimate basal channel activity at positive potentials. To address this issue, current densities were determined under basal conditions immediately after establishing the whole cell configuration in cells expressing the wild-type or mutant TRPC6 proteins. The synopsis of cells expressing wild-type TRPC6 or one of the glycosylation mutants (Fig.  3C) illustrates that progressive ablation of protein glycosylation (wild-type TRPC6 3 TRPC6GM2 3 TRPC6GM1.2) resulted in an increase in current densities at negative and Me 2 SO (0.02% (Ϫ)) or tunicamycin (2 M) as a glycosylation inhibitor dissolved in 0.02% Me 2 SO(ϩ) was added to the cells as described under "Experimental Procedures." Proteins were detected by the TRPC6-specific antibody. Bands corresponding to the glycosylated form of TRPC6GFP are indicated by black triangles. Numbers indicate the molecular masses of the marker proteins in kDa. C, subcellular localization TRPC6 and TRPC6GM1, 2, 1.2. Human TRPC6 and TRPC6GM1, 2, 1.2 were C-terminally fused to GFP, transiently transfected into HEK 293 cells and analyzed by confocal laser scanning microscopy. GFP fluorescence, differential interference contrast (DIC) images, and overlays of GFP fluorescence and corresponding DIC images are depicted. Expression of TRPC6 mutants significantly amplified basal current densities when compared with wild-type TRPC6 (p Ͻ 0.05). At the same time, peak current densities measured after histamine challenge did not significantly differ between wild-type TRPC proteins and any mutant tested at Ϫ60 and ϩ60 mV (data not shown).
Single-channel recordings of TRPC6-and TRPC6GM2-expressing HEK 293 cells revealed conductances of 54 Ϯ 3 pS (n ϭ 4) for TRPC6 and 186 Ϯ 6 pS (n ϭ 3) for the TRPC6GM2 glycosylation mutant. NPo values derived from inside-out single channel recordings of TRPC6-or TRPC6GM2expressing HEK 293 cells at holding potentials of Ϫ60, Ϫ40, 40, and 60 mV before and after application of SAG are shown in Fig. 3D. NPo values in unstimulated patches were higher in the case of TRPC6GM2 when compared with TRPC6 at all voltages tested. When NPo values obtained in unstimulated patches were normalized to those elicited by SAG challenge, basal channel activity amounted to 8.2% at Ϫ60 mV and 6.4% at ϩ60 mV in the case of TRPC6, while 42.3% at Ϫ60 mV and 43.3% at ϩ60 mV were observed in TRPC6GM2-containing isolated patches. Thus, basal channel activities determined by two different electrophysiological approaches were very similar in isolated inside-out patches and in the whole cell configuration of TRPC6 and TRPC6GM2-expressing cells (see Figs. 1C and 3B) and strengthen the concept that the protein glycosylation pattern has a profound impact on basal channel activity.
The notion of increased basal activity of TRPC6 glycosylation mutants was additionally supported by measuring basal Mn 2ϩ entry into fura-2-loaded cells expressing TRPC6 mutants (Fig.  3D). In the case of all three glycosylation mutants tested, the addition of 2 mM Mn 2ϩ to the extracellular medium was followed by a pronounced decline of fura-2 fluorescence (TRPC6GM1: Ϫ0.16 Ϯ 0.02%/s; TRPC6GM2: Ϫ0.15 Ϯ 0.02%/s; TRPC6GM1.2: Ϫ0.17 Ϯ 0.02%/s), reminiscent of genuine TRPC3 properties, but in contrast to the quiescence of wildtype TRPC6 (p Ͻ 0.001 for all three glycosylation mutants compared with TRPC6) under these experimental conditions (cf. Fig. 1D, lower panel). FIG. 4. Glycosylation status of TRPC3 and TRPC3Gly2. A, schematic representation of the glycosylated site in TRPC3 and a mutant (TRPC3Gly2) with an additional potential glycosylation site (E512XS 3 N512XS). B, immunoblot analysis of TRPC3 and TRPC3Gly2 expressed in HEK 293 cells and analysis of their glycosylation status. Lysates (50 g) of HEK 293 cells transfected with pcDNA3 or pcDNA3 with an inserted TRPC6GFP-or TRPC3YFP-cDNA (left panel) were analyzed with an anti-TRPC6 antibody (TRPC6-Ab). Cell lysates (10 g) expressing human TRPC3 or TRPC3Gly2 were analyzed by the same antibody (right panel). Me 2 SO (0.02% (Ϫ)) or tunicamycin (2 M) as a glycosylation inhibitor dissolved in 0.02% Me 2 SO(ϩ)) was added to the cells as described under "Experimental Procedures." Proteins were detected by the TRPC3-specific antibody. Bands corresponding to the glycosylated forms of TRPC3 are indicated by black triangles. C, subcellular localization TRPC3 and TRPC3Gly2. Human TRPC3 and TRPC3Gly2 were C-terminally fused to YFP, transiently transfected into HEK 293 cells and analyzed by confocal laser scanning microscopy. YFP fluorescence, differential interference contrast (DIC) images, and overlays of YFP fluorescence and corresponding DIC images are depicted.
in the whole cell mode at Ϫ60 and ϩ60 mV. D, determination of NPo values using inside-out single channel recordings of TRPC6-or TRPC6GM2-expressing HEK 293 cells at holding potentials of Ϫ60, Ϫ40, 40, and 60 mV (n values are given on top of each bar) before (gray bars) and after (black bars) application of SAG (stars indicate significance at p Ͻ 0.05 using one-way ANOVA). E, Mn 2ϩ quench experiments. HEK 293 cells expressing TRPC6 glycosylation mutants (TRPC6GM1 (n ϭ 13), TRPC6GM2 (n ϭ 10), TRPC6GM1.2 (n ϭ 11)) were loaded with fura-2 and mounted onto a microscope connected to a digital fluorescence imaging system. Mn 2ϩ (2 mM) was added to the Ca 2ϩ -containing bath solution as indicated by bars. Time course of total fura-2 fluorescence is depicted. The black lines represent the calculated means, whereas the gray areas indicate standard errors of the mean (Ϯ S.E.). Background-corrected total fura-2 fluorescence was recorded by exciting fura-2 at the isosbestic wavelength (360 nm) and normalized to the initial values of each single cell.
Engineering of a Dually Glycosylated TRPC3 Mutant (TRPC3GM2)-After having learned that a reduction of the TRPC6 glycosylation pattern resulted in increased basal channel activity, we embarked on a reciprocal approach and introduced an additional glycosylation consensus site into e2 of TRPC3 to generate TRPC3Gly2 [E512N], hypothesizing that such a measure should reduce receptor-independent activity. TRPC3Gly2 was hence assumed to be subject to protein glycosylation at two sites corresponding to those in the dually mod-ified TRPC6 (Fig. 4A). TRPC3 and its glycosylation mutant were C-terminally fused with the yellow fluorescent protein (YFP), expressed in HEK 293 cells and examined by immunoblotting (Fig. 4B) and confocal laser scanning microscopy (Fig. 4C).
Lysates of TRPC3-expressing cells were subjected to SDS-PAGE and probed with an anti-TRPC3 antibody. An incompletely resolved diffuse TRPC3 doublet was detected, the higher molecular mass component of which disappeared after FIG. 5. Basal and receptor-mediated activity of TRPC3 and TRPC3Gly2. A, electrophysiological recordings. HEK 293 cells were transiently cotransfected with a pcDNA3 plasmid encoding the H 1 histamine receptor and a pIRES plasmid encoding the TRPC3Gly2 and the cDNA of eGFP upstream of the internal ribosomal entry site (IRES). Current-voltage relationships were analyzed by patch clamp recordings in the whole cell mode before (basal) and after application of the receptor agonist (histamine) to the pipette solution. B, summary of basal activities of TRPC3Gly2-expressing cells (n ϭ 6) analyzed at Ϫ60 and ϩ60 mV. Values were normalized to the histamine-induced increases at the respective holding potentials. C, analysis of current densities of TRPC3-(n ϭ 11) and TRPC3Gly2-expressing cells (n ϭ 21). Transfected cells were analyzed by patch clamp recordings in the whole cell mode at Ϫ60 and ϩ60 mV, D, Mn 2ϩ quench experiments. HEK 293 cells were transiently cotransfected with plasmids encoding the green fluorescent protein (GFP), the H 1 histamine receptor and human TRPC6 (n ϭ 13), TRPC3 (n ϭ 30), or TRPC3Gly2 (n ϭ 30). After adding Mn 2ϩ (0.2 mM) to the Ca 2ϩ -containing bath solution, cells were stimulated by histamine (His; 100 M as indicated by bars). Time course of total fura-2 fluorescence is depicted. The black lines represent calculated means, whereas the gray areas indicate standard errors of the mean (Ϯ S.E.). Background-corrected total fura-2 fluorescence was recorded by exciting fura-2 at its isosbestic wavelength (360 nm) and normalized to the initial values of each single cell. Stippled lines indicate the initial slopes, i.e. the rates of fura-2 quenching by Mn 2ϩ . pretreatment of cells with tunicamycin indicating that it represented the mono-glycosylated version of wild-type TRPC3 (Fig. 4B), while the faster migrating band of the doublet indicated the non-glycosylated protein. Two to three faint bands migrating with an apparent molecular mass smaller than 85 kDa were only detected in TRPC3-expressing HEK 293 cells (Fig. 4B, left panel) and were most likely proteolytic cleavage products of TRPC3. The antibody showed some minor crossreactivity toward overexpressed recombinant TRPC6GFP (50 g of lysate protein loaded; see: Fig. 4B, left panel). Introduction of a second glycosylation site in e2 allowed for an improved separation of two bands, most likely because of an increase in the molecular mass of the upper, tunicamycin-sensitive band due to dual glycosylation (Fig. 4B, right panel). To assess cell membrane targeting of TRPC3 and its glycosylation mutant TRPC3Gly2, we performed laser scanning microscopy with GFP-tagged proteins. Both wild-type as well as mutant TRPC3 were inserted into the plasma membrane as illustrated by the overlay of fluorescence and differential interference contrast images (Fig. 4C). In line with a previous observation (26) both C-terminally GFP-tagged TRPC3 proteins yielded a punctated fluorescence staining of the cell membrane, clearly set apart from the uniform distribution of the fluorescence signal in the case of TRPC6 (Fig. 2C).
Decreased Basal Activity of the TRPC3Gly2 Mutant-In order to test the hypothesis that the high basal activity of TRPC3 could be reduced by introducing an additional glycosylation site in e2, thus imitating the situation in TRPC6, we functionally analyzed the TRPC3Gly2 mutant after transient expression in HEK 293 cells. The current-voltage relationship of TRPC3Gly2 after application of histamine to the bath solution (Fig. 5A) resembles that obtained with wild-type TRPC3 (Fig. 1A, upper panel) and shows identical rectifying properties. However, Fig.  5A illustrates that additional glycosylation in e2 diminishes basal activity, a phenomenon which was evident when comparing basal outward currents between TRPC3 and TRPC3Gly2 (cf. Fig. 1A, upper panel, and Fig. 5A). Constitutive activities normalized to maximal histamine-induced currents at Ϫ60 or ϩ60 mV (Ϫ29.4 Ϯ 6.7% and 43.3 Ϯ 8.9%, respectively) ( Fig. 5B) in TRPC3Gly2-expressing cells showing outwardly rectifying currents were found to be reduced when compared with TRPC3 (Fig. 1C, upper panel) with differences attaining statistical significance at Ϫ60 mV (p ϭ 0.046). A comparison of current densities in TRPC3-and TRPC3Gly2-expressing cells recorded immediately after patch rupture at holding potentials of Ϫ60 or ϩ60 mV revealed a significantly (p ϭ 0.027) lower current density for TRPC3Gly2 at ϩ60 mV (Fig. 5C). As emphasized earlier, maximal current densities measured in response to histamine stimulation were not significantly different between TRPC3 and TRPC3Gly2 (data not shown), thus ruling out discrepant expression levels of the channel proteins in the cell membrane as the underlying cause of the divergent basal activities noted. Finally, we compared TRPC6, TRPC3, and TRPC3Gly2 by measuring Mn 2ϩ entry in fura-2-loaded cells. (Fig. 5D). In this set of experiments, the extracellular Mn 2ϩ concentration was reduced from 2 mM to 200 M, thus diminishing fura-2 quenching due to basal channel activity, yet allowing for detection of receptorstimulated Mn 2ϩ influx. In contrast to TRPC3, TRPC6 expression did not give rise to fura-2 quenching upon Mn 2ϩ addition to the bath. Of note, basal Mn 2ϩ entry characteristic for wild-type TRPC3 (Ϫ0.061 Ϯ 0.005%/s) was nearly abolished in the case of TRPC3Gly2 (Ϫ0.026 Ϯ 0.008%/s; p Ͻ 0.001; Fig. 5D, lower panel). Histamine challenge resulted in rapid and massive Mn 2ϩ influx in TRPC6-expressing cells. TRPC3-expressing cells consistently reacted to agonist stim-ulation with a transient de-quenching effect followed by an increased rate of Mn 2ϩ entry. The transient de-quenching effect is due to a higher transient rise in the cytoplasmatic Ca 2ϩ concentration after emptying the internal stores of TRPC3-expressing cells and is completely abolished by preincubation of cells with thapsigargin (data not shown). Even though basal activity was no longer observed with TRPC3Gly2, H 1 receptor stimulation initiated cation influx with a rate comparable to that seen with agonist-activated TRPC3.

DISCUSSION
In this study, we characterized the impact of different Nglycosylation patterns on functional differences between TRPC3 and TRPC6. Together with TRPC7 these channels constitute a TRPC subfamily, whose members are activated by diacylglycerol (DAG) in a membrane-delimited fashion (9). Recently, it was demonstrated that proteins of the TRPC3/6/7 subfamily can form heterotetramers by combining with each other, but not with other members of the TRPC family (8), giving rise to a select array of distinct channel tetramers, which may differ in terms of biophysical properties and regulation. To clarify the issue as to whether members of the TRPC3/6/7 subfamily are functionally redundant or whether they serve distinct physiological roles, a thorough characterization of their functional properties and regulation is an indispensable first step.
By applying whole cell patch clamp recordings as well as fluorescence imaging techniques we characterized TRPC6 as a tightly receptor-activated channel with very low basal activity. On the contrary, pronounced channel activation even in the absence of receptor agonist was invariably noted upon TRPC3 expression. To uncover potential molecular determinants for the different functional behavior of two closely related cation channels, we analyzed the glycosylation pattern of TRPC6 as compared with TRPC3. Our findings demonstrate that the TRPC6 protein is dually glycosylated as opposed to the monoglycosylated TRPC3 channel (7). TRPC6 glycosylation not only occurs in heterologous expression systems, but has recently also been demonstrated in pulmonary vascular smooth muscle cells (PASMC) where TRPC6 is involved in PASMC proliferation (27). By deleting glycosylated site(s) in TRPC6 we were able to create mono-and non-glycosylated TRPC6 variants, which were found to be correctly inserted into the plasma membrane as assessed by confocal microscopy. N-linked glycosylation is implicated in correct protein folding and the assembly of functional ion channel complexes as demonstrated for Shaker potassium channels structurally related to the TRPC family (28). Thus, it is worthwhile mentioning that TRPC glycosylation mutants could be expressed as functional ion channels at a cell membrane density comparable to that of wild-type proteins as inferred from peak current densities after agonist challenge. Consistent with the notion that the N-glycosylation pattern is responsible for functional differences TRPC6 glycosylation mutants showed a similarly high basal activity when compared with TRPC3 in Mn 2ϩ quench experiments. Electrophysiological recordings in the whole cell mode confirmed these observations, but revealed differences in the outwardly rectifying properties compared with wild-type TRPC6 indicating an additional as yet unappreciated contribution of e2 glycosylation to TRPC6 function distinct from the control of basal channel activity. The analysis of NPo values derived from isolated inside-out patches support our conclusion based on whole cell experiments and demonstrate that also under cell-free conditions basal channel activity is regulated by the protein glycosylation status.
Conversely we asked if insertion of an additional glycosy-lated site in TRPC3 would lead to reduced basal activity and engineered a glycosylation site into the second extracellular loop (e2) of the TRPC3 protein. Immunoblotting and confocal microscopy experiments showed that the new second consensus site was in fact glycosylated and that the mutant channel was inserted into the plasma membrane. In accordance with our initial hypothesis we noted that the basal activity of TRPC3Gly2 was significantly reduced as opposed to wild-type TRPC3. While TRPC6 was shown to be a receptor-and phospholipase C-controlled cation channel activated by intracellular DAG independent of store depletion (9), the situation with TRPC3 is considerably more complex. Heterologous expression of TRPC3 in HEK 293 and CHO cells revealed TRPC3 channel activity under resting conditions (21,22,23). A HEK 293 cell line stably expressing human TRPC3 also showed high basal Sr 2ϩ , Ba 2ϩ , and Mn 2ϩ -influx when compared with control cells (22). Using patch clamp recordings from inside-out patches, TRPC3 exhibited high intrinsic activity under resting conditions, which could be silenced by Ca 2ϩ /CaM (19). Furthermore, a tight physical and functional interaction of TRPC3 channels and InsP 3 receptors was reported (16,17). In a recent detailed analysis, however, Trebak et al. (29) provided compelling evidence to support the concept that in close analogy to TRPC6 (9) TRPC3 is activated in a receptor-and phospholipase C-dependent fashion via DAG, independently of InsP 3 production and store depletion.
In light of similar biophysical pore properties (25) and channel activation mechanisms, the question arises as to whether TRPC3 and TRPC6 are functionally redundant and thus freely interchangeable in any given cellular setting. Systematic studies to describe principal rules of TRPC subunit combination in functional ion channel complexes revealed that in fact TRPC3 and TRPC6 can be incorporated into the same tetrameric complexes (8,11). To assess the functional contribution of either channel subunit in such heterotetramers, the delineation of distinctive properties allowing us to differentiate between TRPC3 and TRPC6 is a necessary prerequisite. Apart from being activated by DAG all members of the TRPC3/6/7 subfamily display a potentiating effect of decreasing [Ca 2ϩ ] o (30,10). A first distinctive feature is the enhancement of TRPC6 currents by the frequently used cation channel blocker flufenamate which inhibits TRPC3 and TRPC7 (30,31). In addition, we show here by direct comparison that TRPC3 exhibits significant basal activity under resting conditions while TRPC6 is quiescent and under tight regulation by phospholipase C-activating receptors. Furthermore, we present evidence for a cardinal contribution of differences between the N-linked glycosylation pattern to the discrepant intrinsic activities observed. As yet, cell physiological implications of the constitutive activity of TRPC3 are unknown.
Usually N-glycosylation has a great impact on protein folding, intracellular trafficking or membrane targeting. We can exclude such effects for TRPC3 and TRPC6 by showing the insertion of all glycosylation mutants into the plasma membrane. The impact of protein glycosylation on functional characteristics of ion channels is considerably less understood. Our results show that in the case of TRPC3 and TRPC6 a dual glycosylation status is required for low basal and tight receptor-regulated channel activity. The molecular mechanism, however, by which TRPC protein glycosylation accomplishes low resting channel activity still remains elusive. Potential cues to a mechanistic understanding might be derived from studies on voltage-gated ion channels, some of which are heavily glycosylated with terminal sialic acid residues generating a cloud of negative charges at the external surface (see Ref. 32 for a recent review) thereby affecting the S4 voltage sensors of the channels. Obviously, such a mechanism cannot simply be extended to TRPC channels because they are devoid of S4 voltage sensors.
Recently, TRPC6 was reported as a receptor-activated cation channel in platelets (33), and in the aortic smooth muscle cell line A7r5 (31). TRPC6 may further be a likely molecular candidate for the vascular ␣ 1 -adrenoreceptor-activated non-selective cation channel (30). Coexpression of TRPC3 and TRPC6 in smooth muscle cells of cerebral arteries (34), portal vein myocytes (30), but also in mouse spermatozoa (35) and in other tissues was demonstrated (36). At present it is unknown, however, whether TRPC3 and TRPC6 have distinct physiological roles in these cells and tissues or whether they are freely interchangeable without functional consequences. Here, we show in a heterologous cellular expression system that TRPC3 and TRPC6 are not functionally redundant, but display distinct properties: high basal activity of TRPC3 as opposed to tight regulation of TRPC6. Future studies with appropriate in vivo models like for other TRPC channels (37) will ultimately clarify whether these distinctions made at a cellular level also have a bearing on the living organism.