Calmodulin and calcium interplay in the modulation of TRPC5 channel activity. Identification of a novel C-terminal domain for calcium/calmodulin-mediated facilitation.

TRPC5 forms Ca2+-permeable nonselective cation channels important for neurite outgrowth and growth cone morphology of hippocampal neurons. Here we studied the activation of mouse TRPC5 expressed in Chinese hamster ovary and human embryonic kidney 293 cells by agonist stimulation of several receptors that couple to the phosphoinositide signaling cascade and the role of calmodulin (CaM) on the activation. We showed that exogenous application of 10 microM CaM through patch pipette accelerated the agonist-induced channel activation by 2.8-fold, with the time constant for half-activation reduced from 4.25 +/- 0.4 to 1.56 +/- 0.85 min. We identified a novel CaM-binding site located at the C terminus of TRPC5, 95 amino acids downstream from the previously determined common CaM/IP3R-binding (CIRB) domain for all TRPC proteins. Deletion of the novel CaM-binding site attenuated the acceleration in channel activation induced by CaM. However, disruption of the CIRB domain from TRPC5 rendered the channel irresponsive to agonist stimulation without affecting the cell surface expression of the channel protein. Furthermore, we showed that high (>5 microM) intracellular free Ca2+ inhibited the current density without affecting the time course of TRPC5 activation by receptor agonists. These results demonstrated that intracellular Ca2+ has dual and opposite effects on the activation of TRPC5. The novel CaM-binding site is important for the Ca2+/CaM-mediated facilitation, whereas the CIRB domain is critical for the overall response of receptor-induced TRPC5 channel activation.

TRPC5 forms Ca 2؉ -permeable nonselective cation channels important for neurite outgrowth and growth cone morphology of hippocampal neurons. Here we studied the activation of mouse TRPC5 expressed in Chinese hamster ovary and human embryonic kidney 293 cells by agonist stimulation of several receptors that couple to the phosphoinositide signaling cascade and the role of calmodulin (CaM) on the activation. We showed that exogenous application of 10 M CaM through patch pipette accelerated the agonist-induced channel activation by 2.8-fold, with the time constant for half-activation reduced from 4.25 ؎ 0.4 to 1.56 ؎ 0.85 min. We identified a novel CaM-binding site located at the C terminus of TRPC5, 95 amino acids downstream from the previously determined common CaM/IP 3 Rbinding (CIRB) domain for all TRPC proteins. Deletion of the novel CaM-binding site attenuated the acceleration in channel activation induced by CaM. However, disruption of the CIRB domain from TRPC5 rendered the channel irresponsive to agonist stimulation without affecting the cell surface expression of the channel protein. Furthermore, we showed that high (>5 M) intracellular free Ca 2؉ inhibited the current density without affecting the time course of TRPC5 activation by receptor agonists. These results demonstrated that intracellular Ca 2؉ has dual and opposite effects on the activation of TRPC5. The novel CaM-binding site is important for the Ca 2؉ /CaM-mediated facilitation, whereas the CIRB domain is critical for the overall response of receptor-induced TRPC5 channel activation.

gene and its homologue, Trp-like (Trpl), from
Drosophila melanogaster encoded calcium-permeable cationic channels activated either by store depletion or by stimulation of G q/11 -coupled receptors. These initial findings prompted the search for mammalian homologues, leading to the identification of seven TRP genes with different degrees of sequence similarity to the original insect Trp gene (3). These genes are now designated TRP-Canonical or TRPC, symbolizing their close similarity to the original Drosophila Trp. Many recently discovered cation channels are found to share some limited homology with the TRPCs. These include TRPVs (similar to the vanilloid receptor), TRPMs (named after the first identified member, melastatin), and TRPPs (named after PKD2 for polycystic kidney disease), etc. Together, there are at least 28 non-allelic TRP genes in the mammalian genome. The TRP channels serve diverse functions in many tissues from somatosensory to cardiovascular systems (4).
TRPC5 is a member of the TRPC family of Ca 2ϩ -permeable nonselective cationic channels. It has drawn attention recently because of its role in modulating hippocampal growth cone motility and neurite elongation in the mammalian brain (5). The TRPC5 channel activity is induced upon stimulation of the phosphoinositide signaling cascade by receptors that stimulate phospholipase C; however, the exact mechanism of channel activation remains controversial (6). The activation of TRPC5 is dependent on the presence of Ca 2ϩ at both the extracellular and the intracellular sides of the plasma membrane (7)(8)(9)(10). Although the extracellular effect of Ca 2ϩ has been shown to be mediated by the acidic residues, Glu 543 , Glu 595 , and Glu 598 , located at the putative pore loop of the TRPC5 protein (7), the mechanism for the intracellular Ca 2ϩ dependence of TRPC5 remains to be elucidated.
Calmodulin (CaM) is a common intracellular mediator of many Ca 2ϩ -dependent regulations. All TRPC proteins possess a C-terminal CaM-binding domain that also interacts with an N-terminal sequence of the inositol 1,4,5-trisphosphate receptor (IP 3 R) (6). We have demonstrated that IP 3 R and CaM compete with each other for binding to the common CaM/IP 3 Rbinding (CIRB) site of the TRPC. In functional studies, the TRPC-binding region of the IP 3 R activated and Ca 2ϩ /CaM inhibited the activation of TRPC3 and TRPC4. Moreover, TRPC channels were activated by removing or inactivating CaM from excised inside-out membrane patches, indicating that displacement of the inhibitory CaM from the common CIRB site is sufficient for the activation of TRPC channels (11). Consistent with this, we have shown that CaM increased the delay between the release of Ca 2ϩ from internal storage compartments and the activation of Ca 2ϩ influx via endogenous TRPC1 channels in Chinese hamster ovary (CHO) cells, which were subjected to the regulation by IP 3 R and CaM in a similar fashion as the exogenously expressed TRPC3 and TRPC4 in human embryonic kidney (HEK) 293 cells (12). Additional C-terminal CaM-binding domains outside of the CIRB sites have been found on TRPC proteins (6,13). Unlike the CIRB sites, these sites are not conserved among all TRPC channels, and they do not bind to IP 3 Rs. The second CaM-binding (CBII) site of TRPC1 has been shown to be involved in the slow Ca 2ϩ -induced channel inactivation (13).
In the present study we have identified the CBII site from mouse TRPC5 (mTRPC5). We have explored the specific roles of the CIRB and CBII sites in the modulation of channel activity after activation of receptors that stimulate the phosphoinositide signaling cascade by selectively disrupting each CaMbinding site from mTRPC5 and heterologous expression of the wild type and mutant channels in CHO and HEK293 cells. We show for the first time that intracellular application of CaM accelerated the activation of mTRPC5 by receptor agonists. Although mutations in the CIRB site rendered the channel inactive, deletion of the CBII site attenuated the Ca 2ϩ /CaMinduced acceleration of receptor-evoked mTRPC5 activation.
DNA Constructs and Mutagenesis-Complementary DNA for the open reading frame of mTRPC5 was subcloned in pIRESneo (Clontech, Mountain View, CA) vector. Mutageneses were performed using the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Mutations were confirmed by DNA sequencing. Constructs for maltosebinding protein (MBP) fusion proteins that contained different fragments of mTRPC5 were made in the pAGA3 vector as described previously (14). cDNA for the guinea pig histamine type 1 receptor (H1R) in pcDNA3 was kindly provided by Dr. Michael Schaefer (Charité-Universitä tsmedizin Berlin) and that for the human type 2 bradykinin receptor (Bk2R) was a generous gift from Dr. William W. Schilling (Case Western Reserve University).
Analysis of the Expression of mTRPC5 mRNA by RT-PCR-Total RNA was extracted from wild type cells and cells transfected with mTRPC5 and the mTRPC5 mutants using the TRIzol reagent (Invitrogen). Approximately 500 ng of total RNA was used as the template for RT-PCR, which was carried out using the ONEStep system with Superscript II (Invitrogen) following the manufacturer's protocols. The reaction was performed on a TC-512 thermal cycler (TECHNE, Cambridge, UK) using 30 cycles of the following PCR protocol: 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s. The primers used for mTRPC5 were 5Ј-CTATGAGACCAGAGCTATTGATG (forward) and 5Ј-CTACCAGG-GAGATGACGTTGTATG (reverse), which amplify a 220-bp-long product. For glyceraldehyde-3-phosphate dehydrogenase, the primers used were 5Ј-GACATCAAGAAGCTGGTGAAGC (forward) and 5Ј-TACTC-CTTGGAGGCCATGTAG (reverse), which amplify a 236-bp-long product. The PCR products were sequenced to identify the products, run on a 2% agarose gel, and stained with ethidium bromide. The gel was analyzed with a Typhoon 8600 Imager (Amersham Biosciences).
CaM Binding Assay-Fragments of mTRPC5 fused to the C terminus of MBP were prepared by in vitro synthesis using the transcription-and translation-coupled rabbit reticulocyte lysates in the presence of [ 35 S]Met and [ 35 S]Cys as described previously (11,14). The 35 S-labeled proteins were incubated with CaM-Sepharose at room temperature for 30 min in a binding solution that contained 120 mM KCl, 1 mM CaCl 2 , 0.5% Lubrol, 20 mM Tris-HCl, pH 7.5. After several washes, bound proteins were separated by SDS-PAGE and then revealed by x-ray autoradiography as described (11). For studying Ca 2ϩ dependence of mTRPC5-CaM binding, 10 mM EGTA or HEDTA and the desired concentrations of CaCl 2 (calculated using the MaxChelator program (www.stanford.edu/ϳcpatton/maxc.html)) were included in the binding solution, and the apparent affinity (K 1/2 ) for Ca 2ϩ was determined as described (14).
Cell Surface Biotinylation Assay-Transfection, biotinylation, and streptavidin precipitation were performed as described previously (15). Immunoblotting was performed using anti-mTRPC5 antibodies (Alomone Labs, Jerusalem, Israel). Sample loading for the crude cell lysate and streptavidin precipitated portion represents 12 and 355 g, respectively, of total proteins in cell lysates. Identical exposure time was used to reveal the chemiluminescent signals for the mTRPC5 proteins in crude lysates and streptavidin-precipitated samples.
Fluorescence-based Membrane Potential Measurements-HEK293 cells were cotransfected with mTRPC5 and H1R in wells of a 96-well plate as described previously (16). One day after the transfection, cells were washed once with Hanks' balanced salt solution and then incubated for 30 min with 80 l of FLIPR membrane potential dye (Molecular Devices, Sunnyvale, CA) diluted in the Hanks' solution. Changes in membrane potential were measured at 32°C using a fluid handling integrated fluorescence plate reader, FlexStation (Molecular Devices). Histamine was diluted in Hanks' solution at 300 M, and 40 l were delivered to the sample plate by the integrated robotic 8-channel pipettor at 20 s after readings began. Samples were excited at 530 nm, and emission of 565 nm was collected from the bottom of the plate at 0.67 Hz.
Stable Cell Lines-HEK293 and CHO cells were transfected with bicistronic plasmids containing the wild type mTRPC5 or the mutants described in this study followed by an internal ribosome entry site and the gene conferring resistance to neomycin. Cells were maintained in 1 mg/ml G418 for 1 month and later assayed by RT-PCR to confirm the presence of the mRNA for mTRPC5 and aminoglycoside phosphotransferase. Stably transfected cell clones were isolated by progressive dilution and tested for mTRPC5 expression. Cells (both CHO and HEK293) were transfected also with a similar bicistronic plasmid containing the cDNA for Bk2R and the gene conferring resistance to puromycin. Clonal cell lines expressing both mTRPC5 and Bk2R were isolated using the two selection markers, neomycin and puromycin.
Whole-cell Measurements of mTRPC5 Currents-The whole-cell configuration of the patch clamp technique was utilized to study mTRPC5mediated currents as described previously (12). Briefly, cells were plated on glass coverslips and mounted on the stage of an inverted microscope (Nikon Instruments, Japan). The amplifier used was the Axopatch 200A (Axon Instruments, Union City, CA). Pipette resistance was 10 -12 megohms when tested with pipette and bath solutions. Whole-cell resistances were in the range of 1-2 gigohms, and cell capacitance ranged from 10 to 12 picofarads for CHO and 12 to 16 picofarads for HEK293 cells. Cells were held at the holding potential of Ϫ80 mV and repetitively stepped to Ϫ120 through ϩ60 mV for 500 ms each in a 20-mV increment once every second. Receptor agonists were applied through bath perfusion, and CaM was introduced intracellularly by dialysis through the patch pipette.
Data Analysis and Curve Fitting-All data were plotted and fitted using Sigmaplot 8 (SYSTAT, Point Richmond, CA). To obtain activation time constants, the mean outward current was plotted over time for the duration of the experiments under the different experimental conditions indicated in each figure legend. Data were fitted to sigmoidal Equation 1 as follows: where a indicates the maximum outward current value (top asymptote); b indicates the minimum outward current value (bottom asymptote); X 0 indicates time constant for half-activation, and x indicates the time explored for each data point.

RESULTS
Time Course of the Activation of mTRPC5 by Agonists-Agonist stimulation of CHO cells stably expressing the wild Novel CaM-binding Domain Facilitates TRPC5 Channel Activation type mTRPC5 channel results in a time-dependent activation of outwardly rectifying currents (Fig. 1). We followed current activation in the presence of agonists for 15 min (only the first 8 min are shown). The current reached a steady-state level after ϳ7 min in the continuous presence of the agonists. After this period of time, 30% of the cells showed a small (about 5%) reduction of current amplitude. This phenomenon was not further studied in the present work. Fig. 1A shows a family of currents evoked by voltage pulses from a holding potential of Ϫ80 to Ϫ120 mV through ϩ60 mV in 20-mV steps. Thrombin (1 unit) induced activation of an outwardly rectifying current with the reversal potential near 0 mV (Ϫ6 Ϯ 4 mV, n ϭ 25). The time course of the current-voltage (I-V) relationships is illustrated in Fig. 1B. The gray rectangle in Fig.  1B shows the limit of current induced by the agonist in cells not expressing mTRPC5. Notice that basal current level (before agonist stimulation) in mTRPC5-expressing cells was slightly higher when compared with wild type cells (ϳ10% more current). Although this current may reflect mTRPC5 channel activity, this was not further explored in the present study.
To confirm the expression of mTRPC5, we performed RT-PCR experiments with total RNA obtained from control cells (wild type CHO cells) and cells expressing mTRPC5. The inset in Fig. 1B shows that that only mTRPC5-transfected cells contained the mTRPC5 transcript.
The time course of mTRPC5 current activation was independent of the agonist and receptor utilized. Fig. 1 only illustrates the response to thrombin. We have compared the responses to thrombin and bradykinin (Bk) at concentrations that gave the maximal response (1 unit and 100 nM, respectively) in CHO cells stably expressing mTRPC5, and we ob-  (Table I). Because inward currents were relatively small compared with the amplitude of outward currents, and because they were difficult to separate from endogenous inward currents (this was especially difficult with HEK293 cells that showed larger inward currents than CHO cells), all the analyses in this study were conducted only on the outward currents.
Identification of Two CaM-binding Sites at the C Terminus of mTRPC5-The above results suggest that CaM plays an important role in facilitating the receptor-induced activation of mTRPC5 channels. We have reported previously (14) the presence of a conserved common CIRB site at the near Cterminal regions of all TRPC proteins. For mTRPC5, the CIRB site lies in between Glu 701 and Lys 733 and requires higher Ca 2ϩ concentrations (apparent K 1/2 ϭ 44.2 M) for binding to CaM than the CIRB sites of other TRPC proteins. We also reported that the rest of the mTRPC5 C terminus downstream from the CIRB site (Gly 762 -Leu 975 , clone J1170 in Fig. 3A) bound to CaM as well. To narrow down the CBII site of mTRPC5, we have tested a large number of smaller fragments generated from J1170 by using an in vitro pulldown assay. As shown in Fig. 3, the minimal binding site for CaM is confined to a stretch of 27 residues, Pro 828 -Asn 854 , 95 amino acids downstream from the CIRB site. The isolated mTRPC5 CBII fragment bound to CaM in a Ca 2ϩ -dependent manner (Fig. 3D) with an apparent K 1/2 for Ca 2ϩ of 3.1 Ϯ 0.2 M (n ϭ 3). This value is lower than the previously determined K 1/2 value (44.2 M) of Ca 2ϩ for CaM binding to the mTRPC5 CIRB site but within the range of those values (1.6 -12.9 M) for the CIRB sites of other TRPC isoforms (14).
Mutations on CaM-binding Sites Alter mTRPC5 Channel Activation-To investigate the function for each of the CaMbinding sites of mTRPC5, we destroyed the CIRB and CBII sites by substituting Arg 718 , Lys 722 , and Arg 723 with alanines (CIRBm1) and deleting Pro 828 -Asn 854 (⌬CBII), respectively (Fig. 4A). The loss of CaM binding in these mutants was confirmed by in vitro binding studies (Fig. 4B). Receptor-mediated activation of mTRPC5 currents was studied after coexpression of the full-length clones of the mutated mTRPC5 with Bk2R or by stimulation of the endogenous thrombin receptor in CHO. As shown in Fig. 4C, without CaM in the pipette, the activation of mTRPC5⌬CBII was slightly delayed as compared with the wild type mTRPC5. Inclusion of 10 M CaM in the pipette only weakly accelerated the activation of the mutant channel, indicating that CBII is important for the Ca 2ϩ /CaM-mediated facilitation of mTRPC5 activation. After fitting the outward current, the time constants for mTRPC5⌬CBII of 6.04 Ϯ 0.5 (control) and 5.48 Ϯ 0.7 min (with 10 M CaM) were obtained.
In contrast, the mTRPC5-CIRBm1 mutant failed to respond to the stimulation by thrombin and Bk in both CHO and HEK293 cells. In a fluorescence-based membrane potential assay using the FLIPR membrane potential dye, coexpression of mTRPC5 with H1R in HEK293 cells resulted in a histamineevoked membrane depolarization (Fig. 4D). Although cells expressing mTRPC5⌬CBII showed a slightly reduced depolarization as compared with those expressing the wild type mTRPC5, cells expressing mTRPC5-CIRBm1 failed to show any difference from those expressing H1R alone. This result confirms that mTRPC5-CIRBm1 is a loss-of-function mutant.
In order to determine whether mTRPC5-CIRBm1 is sufficiently expressed on the plasma membrane, we performed surface biotinylation assays for mTRPC5 and its mutants transiently transfected in the HEK293 cells. Two days after  the transfection, although a lower amount was detected for mTRPC5⌬CBII than for the wild type, the expression of mTRPC5-CIRBm1 in the total cell lysate as well as in the biotinylated fraction was similar to that of wild type mTRPC5 (Fig. 4E). The fact that mTRPC5-CIRBm1 protein was labeled by biotin in the nonpermeabilized cells at cold temperature indicates that the mutant protein is efficiently delivered to the plasma membrane.
To confirm that the CIRB site is critical for the receptorinduced mTRPC5 activation, we made another mutant, I717D/ L720E/V721A, at the CIRB site. Again the mutant was irresponsive to agonist stimulation but showed normal cell surface expression as the wild type channel (not shown). Thus, the lack of function for these CIRB mutants is most likely because of a defect in channel gating and/or channel activation by agonists, rather than a dysfunction in the trafficking and translocation of the channel protein.
Inhibitory Effect of Constantly High Intracellular Ca 2ϩ on the Activation of mTRPC5-To investigate further the role of intracellular Ca 2ϩ and CaM on the activation rate and amplitude of mTRPC5 currents, studies were performed by utilizing two concentrations (0.4 and 5 M) of intracellular Ca 2ϩ in the presence and absence of CaM in the patch pipette with cells expressing wild type mTRPC5 and mTRPC5⌬CBII mutant (Fig. 5). As compared with 0.4 M, the high intracellular Ca 2ϩ (5 M) produced a significant inhibition of the maximal currents activated by thrombin in CHO cells that expressed wild type mTRPC5 from 995 Ϯ 110 (Fig. 5A, n ϭ 22) to 302 Ϯ 28 pA (Fig. 5B, n ϭ 19), which accounts for 30.3% of the current obtained with 0.4 M free Ca 2ϩ in the pipette. The maximum current inhibition obtained with 10 M free Ca 2ϩ (75%) was not different from that obtained with 5 M free Ca 2ϩ .
For mTRPC5⌬CBII, 5 M free Ca 2ϩ also reduced the maximal amplitude of receptor-induced currents (at ϩ60 mV) to 28.2% of that obtained with 0.4 M free Ca 2ϩ (Fig. 5C). This level of inhibition is similar to that of the wild type mTRPC5, suggesting that the CBII site is not involved in the inhibition of channel activation by high concentrations of intracellular Ca 2ϩ .
As illustrated in previous figures, introduction of CaM in the patch pipette in the presence of 0.4 M free Ca 2ϩ did not alter the maximal current amplitude but accelerated the rate of current activation in response to agonist stimulation of the wild type mTRPC5. Similarly, introduction of 10 M CaM in the pipette in the presence of 5 M free Ca 2ϩ did not overcome the inhibitory effect of the high intracellular Ca 2ϩ concentration on the maximal current amplitude (Fig. 5C) but reduced the time constant for half-activation of the outward current at ϩ60 mV from 4.85 Ϯ 0.8 (n ϭ 12) to 2.13 Ϯ 1.02 min (n ϭ 16) (Fig. 5D). On the other hand, there is no significant difference in the time constants between 0.4 and 5 M free Ca 2ϩ for mTRPC5 activation in the absence or presence of exogenously applied CaM, indicating that either the endogenous CaM is more than enough to mediate the inhibitory effect of high intracellular Ca 2ϩ or the inhibition is independent of CaM.
To explore further the possible role of CaM on the Ca 2ϩinduced inhibition of mTRPC5 and mTRPC5⌬CBII, we used two potent CaM inhibitors in the patch pipette. As illustrated in Fig. 5C, neither 500 nM of W7 nor 5 M of TFP showed any effect on the inhibition of current amplitude induced by 5 M Ca 2ϩ . We had shown previously (12) that these concentrations of the inhibitors were sufficient to eliminate the participation of CaM in the modulation of store-operated calcium entry in CHO cells.
Similarly, addition of CaM to the pipette of mTRPC5⌬CBIIexpressing cells did not alter the rate of channel activation in response to agonists, as indicated in Fig. 5D, even though the current amplitude was significantly reduced (Fig. 5C). Activation time constants obtained for mTRPC5⌬CBII with 5 M Ca 2ϩ were 5.96 Ϯ 0.7 in the absence of CaM and 5.21 Ϯ 0.8 min with 10 M CaM (n ϭ 21).
The Slow Activation of mTRPC5 in Response to Agonist Stimulation Does Not Involve Translocation to the Plasma Membrane-It has been shown recently (17) that epidermal growth factor induced rapid vesicular translocation of mTRPC5 channels to the plasma membrane in HEK293 cells. To determine whether the effect of CBII domain on facilitation of channel activation in response to agonists may be the result of mTRPC5 translocation to the plasma membrane, we performed two types of experiments. In the first set of experiments we compared the time course of channel current activation in response to agonists at 4 and 27°C. As illustrated in Fig. 6, A and B, no significant differences in the time courses of mTRPC5 and mTRPC5⌬CBII were observed between the two temperatures At 4°C all vesicle transport and plasma membrane translocation of vesicles were inhibited, and therefore if the slow activation induced by agonists was the result of more mTRPC5 reaching the plasma membrane, it was expected to observe differences in the time courses of current activation between 4 and 27°C. Furthermore, if CaM via the CBII domain was affecting the translocation of mTRPC5 to the plasma membrane, it would be expected to observe differences in time constants between 4 and 27°C. As illustrated in Fig. 6B, this was not the case.
To explore further any possible role of channel translocation to the plasma membrane induced by agonist stimulation, we performed biotinylation assays with mTRPC5-expressing cells under resting conditions and after 1 M Bk was applied to the cells for 1 or 10 min. As illustrated in Fig. 6C, Bk induced a significant increment in biotinylated mTRPC5 after only 1 min of Bk stimulation. No significant differences were detected between 1 and 10 min of Bk exposure (Fig. 6C). These results demonstrate that Bk induced a rapid (less than 1 min) translocation of mTRPC5 to the plasma membrane. However, the rapid translocation cannot explain the slow activation of mTRPC5 currents induced by agonists. Therefore, it appears that the CaM modulation of mTRPC5 facilitation via the CBII domain does not involve channel translocation to the plasma membrane.

Ca 2ϩ /CaM-mediated Facilitation of mTRPC5 Activation by
Receptor Agonists-Stimulation of many cell surface receptors triggers intracellular Ca 2ϩ signaling through activation of phospholipase C and breakdown of phosphoinositides. Members of the TRPC family of Ca 2ϩ -permeable nonselective cation channels have emerged as important players for the receptorinduced Ca 2ϩ signaling because they are activated downstream from phospholipase C activation and mediate Ca 2ϩ entry into the cells. Exactly which step(s) or component(s) of the phosphoinositide signaling cascade is involved in gating the TRPC channels remains to be elucidated. In the present study, we have expressed mTRPC5 in two cell lines (CHO and HEK293), and we studied kinetics and the magnitudes of its activation after agonist stimulation of two different receptors (thrombin receptor and Bk2R). For both receptors, the activation of the mTRPC5 channel was slow, requiring ϳ7 min to reach steadystate values in the continuous presence of the agonists. The two different receptor agonists induced channel activity with similar time constants (Table I), indicating that the slow activation kinetics is a property of the mTRPC5 channels independent of receptor and cell types.
Most interestingly, rat TRPC5 expressed in HEK293 cells showed a slow initial activation phase that was clearly distinguishable from the later fast activation in response to the simulation of coexpressed H1R (18). In our study, there was no clear transition from the slow to fast activation, and the rate of activation for mTRPC5 was much slower than that for the rat  Fig. 3A. Equivalent regions from the mutants were tested under the same condition as described in Fig. 3. wt, wild type. C, time course of the current activation of wild type mTRPC5 (squares) and the mutants mTRPC5⌬CBII (circles) and mTRPC5-CIRBm1 (triangles) in the presence or absence of 10 M CaM. Currents were induced by the application of 100 nM Bk. The outward current was measured by a voltage pulse to ϩ60 mV as described under "Experimental Procedures." D, histamine-induced membrane depolarization in HEK293 cells that coexpressed H1R and mTRPC5 or its mutant. 100 M histamine was added as indicated. Fluorescence values shown are 1/100,000 of those displayed by the instrument. E, biotinylation experiments illustrating that all mutants reached the plasma membrane similarly to wild type mTRPC5. Actin was used as loading control on each lane. kD illustrates the molecular mass markers (in kilodaltons).
clone. These differences may be due to the different concentrations of EGTA (0.5 versus 10 mM) used between the two studies.
We show that introduction of CaM in the patch pipette significantly accelerated mTRPC5 channel activation after agonist stimulation. Half-time constants were reduced almost 3-fold with thrombin and Bk in CHO cells and Bk in HEK293 cells (Table I). These results indicate that the effects of CaM on mTRPC5 activation are also independent of the receptor stimulated and the cell line used.
The Second C-terminal CaM-binding Site Is Critical for the Ca 2ϩ /CaM-mediated Facilitation of mTRPC5 Activation-We have now identified a novel CaM-binding domain (CBII) on the mTRPC5 C terminus, and we show that it is critical for the Ca 2ϩ /CaM-mediated facilitation of receptor-induced activation of mTRPC5. Disruption of the CBII site prevented the acceleration of channel activation in response to agonist stimulation mediated by the exogenously applied CaM. More interestingly, even though it is important for the Ca 2ϩ -mediated facilitation, this CaM-binding site is not absolutely required for the channel activity. The mTRPC5⌬CBII mutant showed about two-thirds of the maximal currents and membrane depolarization as compared with the wild type mTRPC5 but maintained other properties of the mTRPC5 channel, including the shape of the I/V curve, reversal potential, and inhibition by high concentrations of the intracellular Ca 2ϩ . Differences in the amount of current may be related to the amount of protein expressed/transported to the membrane, due to clonal differences.
These data suggest that the CBII site has a modulatory role in controlling the rate of channel recruitment following the activation of G-protein-coupled receptors. This positive feedback modulation by Ca 2ϩ through CaM is a novel mechanism for TRPC channels and may have important physiological implications. The rate of mTRPC5 activation is thus controlled by how fast and how high the Ca 2ϩ concentration near the cytoplasmic side of the channel can increase. This explains why the activation of TRPC5 is heavily influenced by the Ca 2ϩ -buffering capacity of intracellular solutions (7,9).
Biotinylation and electrophysiological experiments carried out at 4°C strongly suggest that the effects of CaM via CBII do not involve significant changes in mTRPC5 channel translocation to the plasma membrane. Although we observed a rapid increment in biotinylated mTRPC5 in response to Bk, no significant differences were observed at the time periods when we observed the acceleration of mTRPC5 current activation in response to agonists.
The activation of TRPC5 causes Na ϩ and Ca 2ϩ influxes, which in turn lead to membrane depolarization and intracellular Ca 2ϩ increase, respectively. In unclamped cells, the positive feedback mechanism should be more important in the initial phase of channel activation before membrane depolarization becomes predominant, in which case the reduced driving force for Ca 2ϩ influx would prevent further intracellular Ca 2ϩ elevation through TRPC5. However, membrane depolarization in excitable cells would lead to the activation of voltage-gated Ca 2ϩ channels, causing further increases in intracellular Ca 2ϩ . Because TRPC5 activation is also inhibited by high concentrations of intracellular Ca 2ϩ , the Ca 2ϩ increases mediated by the voltage-gated channels could cause either further potentiation or inhibition of the TRPC5 channel depending on the relative density of each channel type and the level of receptor activation and CaM present. The relative activity of TRPC5 would again affect the voltage-gated channels through membrane potential regulation. Such an intricate interplay among receptors, TRPC5, and voltage-gated Ca 2ϩ channels provides a mechanism to fine tune Ca 2ϩ signal in the growth cones of hippocampal neurons, in which both TRPC5 and voltage-gated Ca 2ϩ channels are abundantly expressed (5).
Ca 2ϩ -mediated Inhibition of TRPC5 Activation-The regulation of TRPC5 by Ca 2ϩ is further complicated by the finding that at constantly high intracellular Ca 2ϩ concentrations, the receptor-induced activation of TRPC5 is inhibited. Different from the Ca 2ϩ /CaM-mediated facilitation discussed above, the Ca 2ϩ -mediated inhibition did not affect the rate of channel activation but rather the maximal current amplitudes. Introduction of exogenous CaM did not overcome the inhibitory effect of high Ca 2ϩ on the current amplitudes but still facilitated the rate of channel activation by receptor agonist. Furthermore, the mTRPC5⌬CBII mutant was inhibited to a similar degree by high Ca 2ϩ as the wild type channel, indi-cating that the CBII site is not involved in the Ca 2ϩ -mediated inhibition. Therefore, different sites are involved in the Ca 2ϩ / CaM-mediated facilitation and Ca 2ϩ -mediated inhibition of the receptor-induced activation of TRPC5 channels. Given that the CIRB site is important for the Ca 2ϩ -dependent inhibition of TRPC3 activation by IP 3 Rs (11), it is possible that the TRPC5 CIRB site is responsible for channel inhibition by high concentrations of intracellular Ca 2ϩ . This is supported by the finding that the Ca 2ϩ affinity for CaM binding to the CIRB site is more than 10 times lower than to the CBII site of mTRPC5. Thus, much higher Ca 2ϩ concentrations are needed for CaM to affect the CIRB than the CBII site. Unfortunately, the role of the CIRB site on this effect could not be determined because mutations introduced at this domain rendered the channel inactive.
Dual regulations by Ca 2ϩ have been shown for the Drosophila TRPL channel (19,20), which was originally identified in a CaM binding assay (21). Low intracellular Ca 2ϩ enhanced the activation of the TRPL channel with an EC 50 value of 0.45 M (20), whereas higher (micromolar) intracellular Ca 2ϩ inhibited the channel activity (20,22). Whether or not CaM is involved in the Ca 2ϩ -mediated regulation of TRPL has not been completely resolved, although at least two CaM-binding sites (23), in addition to the CIRB domain (14), are present at its C terminus. Thus, for TRPC5 whether or not the Ca 2ϩ -mediated inhibition on current density is mediated by CaM and the critical site(s) involved in this regulation warrant further investigation.
The Critical Role of the CIRB Site in the Receptor-induced Activation of TRPC5-We have also found that the common (see under "Experimental Procedures") to obtain the half-activation time constants. C, HEK293 cells were cotransfected with Bk2R and wild type TRPC5. One day after transfection, 1 M Bk was applied to the cells. 1 or 10 min later, the stimulation was terminated by washing cells with cold phosphate-buffered saline three times. Endocytosis was avoided by placing the culture dishes on ice. Total proteins were evaluated by the total lysate, and the surface parts were evaluated by biotinylated proteins. Actin (anti-actin) was probed as the control for loading. CIRB site is essential for the function of TRPC5 channels. Mutations at the CIRB site rendered the mTRPC5 channel inactive. However, the biotinylation experiments showed that a similar amount of the mTRPC5-CIRB mutants reached the plasma membrane as the wild type mTRPC5, ruling out the possibility that the lack of function was because of a defect in channel protein synthesis, stability, or trafficking and translocation. Thus it appears that an intact CIRB site is required for the activation of mTRPC5 channels via receptor agonists. This is in contrast to the finding that deletion of the CIRB site impaired the plasma membrane translocation of TRPC3 (24), arguing against the claim that trafficking-related defect is the cause for the lack of TRPC5 channel activity resulting from the disruption of CaM and/or IP 3 R binding at the CIRB domain.
It has been shown previously (25) that activation of IP 3 R is required for the opening of the mTRPC5 channels. The results presented here are in agreement with this early finding, while highlighting the key role of the CIRB domain during agonistinduced channel activation. Because this domain is conserved in all TRPC members, it is important to determine to what extent a functional CIRB domain, with perhaps a separable binding capability to IP 3 R and CaM, is required for channel activation in the TRPC family.
In summary, the current study demonstrates that CaM binding to a C-terminal site 95 residues downstream from the previously identified CIRB domain is critical for the Ca 2ϩ /CaMmediated facilitation of receptor-induced activation of TRPC5 channels. It shows that in addition to facilitating the rate of channel opening, intracellular Ca 2ϩ also blocks the receptorinduced channel activation at high concentrations. This effect appears to be independent of CaM, as indicated by the use of two potent CaM antagonists. Moreover, it provides a further support for a critical role of the CIRB site on the activation of TRPC5 channels by receptor agonists. Together, these results reveal a complex modulation of TRPC5 channel function, which provides cells with sophisticated means to exquisitely control the rate of cell membrane depolarization and Ca 2ϩ influx in response to agonist stimulation of cell surface receptors. The role of TRPC5 in modulating hippocampal growth cone motility and neurite elongation has been established recently (5). Most interestingly, it is well known that CaM is also important during neuron growth and regeneration (26). Future experiments using mTRPC5⌬CBII or similar mutant channels could provide insightful information about the role of CaM and TRPC5 in controlling neurite elongation in the mammalian brain.