Molecular Cloning and Functional Characterization of a Novel Receptor-activated TRP Ca2+ Channel from Mouse Brain*

Characterization of mammalian homologues of Drosophila TRP proteins, which induce light-activated Ca2+ conductance in photoreceptors, has been an important clue to understand molecular mechanisms underlying receptor-activated Ca2+ influx in vertebrate cells. We have here isolated cDNA that encodes a novel TRP homologue, TRP5, predominantly expressed in the brain. Recombinant expression of the TRP5 cDNA in human embryonic kidney cells dramatically potentiated extracellular Ca2+-dependent rises of intracellular Ca2+ concentration ([Ca2+] i ) evoked by ATP. These [Ca2+] i transients were inhibited by SK&F96365, a blocker of receptor-activated Ca2+ entry, and by La3+. Expression of the TRP5 cDNA, however, did not significantly affect [Ca2+] i transients induced by thapsigargin, an inhibitor of endoplasmic reticulum Ca2+-ATPases. ATP stimulation of TRP5-transfected cells pretreated with thapsigargin to deplete internal Ca2+stores caused intact extracellular Ca2+-dependent [Ca2+] i transients, whereas ATP suppressed [Ca2+] i in thapsigargin-pretreated control cells. Furthermore, in ATP-stimulated, TRP5-expressing cells, there was no significant correlation between Ca2+ release from the internal Ca2+ store and influx of extracellular Ca2+. Whole-cell mode of patch-clamp recording from TRP5-expressing cells demonstrated that ATP application induced a large inward current in the presence of extracellular Ca2+. Omission of Ca2+ from intrapipette solution abolished the current in TRP5-expressing cells, whereas 10 nm intrapipette Ca2+ was sufficient to support TRP5 activity triggered by ATP receptor stimulation. Permeability ratios estimated from the zero-current potentials of this current wereP Ca:P Na:P Cs= 14.3:1.5:1. Our findings suggest that TRP5 directs the formation of a Ca2+-selective ion channel activated by receptor stimulation through a pathway that involves Ca2+ but not depletion of Ca2+ store in mammalian cells.

Calcium (Ca 2ϩ ) influx across the plasma membrane plays a vital role in the regulation of diverse cellular processes, ranging from ubiquitous activities like gene expression to tissuespecific functions such as neurotransmitter release and muscle contraction, by controlling the cytosolic free Ca 2ϩ concentration ([Ca 2ϩ ] i ) (1,2). Recently, in addition to the well characterized modes of Ca 2ϩ entry through voltage-dependent Ca 2ϩ channels and ligand-gated cation channels, receptor-activated Ca 2ϩ influx that occurs as a second phase of phosphatidylinositol (PI) 1dependent response, has been recognized for its physiological significance (3). Diverse ion channels activated by various triggers have been recognized to be responsible for the receptoractivated Ca 2ϩ influx (3). Among members of the group, recent attention was particularly directed to capacitative Ca 2ϩ entry (CCE; in other words, Ca 2ϩ release-activated current (I CRAC ), or store-operated channel), that is activated through Ca 2ϩ release from the intracellular Ca 2ϩ store, endoplasmic reticulum (ER), induced by inositol 1,4,5-trisphosphate (IP 3 ) and consequent depletion of Ca 2ϩ from the store (2)(3)(4)(5)(6)(7)(8)(9). Diffusible small molecules (10,11), IP 3 metabolites (12), and direct coupling of IP 3 receptors or small GTP-binding (G) proteins with the channel proteins (5,6,9) have been proposed to be involved in the activation of this Ca 2ϩ entry pathway. Other plasma membrane ion channels directly activated by second messengers such as Ca 2ϩ , IP 3 , and inositol 1,3,4,5-tetraphosphate (IP 4 ) (13)(14)(15)(16) are also categorized as receptor-activated Ca 2ϩ channels (3).
An important clue for understanding the molecular basis of receptor-activated Ca 2ϩ influx was first attained through the finding of a Drosophila visual transduction mutation transient receptor potential (trp), whose photoreceptors fail to generate the Ca 2ϩ -dependent sustained phase of receptor potential and to induce subsequent Ca 2ϩ -dependent adaptation to light (17,18). Inasmuch as the gene products of the trp and trp-like (trpl) gene (TRP and TRPL) comprise the light-activated, PI-dependent Ca 2ϩ conductance in Drosophila photoreceptors (19), the original hypothesis that the counterparts of TRP and TRPL are responsible for CCE in vertebrate cells was based upon analogy between the phototransduction mechanism in Drosophila and the PI-dependent signal transduction processes in vertebrates (18). In fact, recent molecular characterization has unveiled the existence of multiple genes encoding TRP homologues in vertebrate cells (20 -25), and cDNA expression experiments of TRP proteins present some lines of supportive evidence for the hypothesis that TRP and its homologues except TRPL are CCE channels (20,(23)(24)(25)(26)(27)(28)(29)(30)(31). However, the hypothesis is still controversial (19,32,33). Acharya et al. (33) demonstrated that photoreceptors from Drosophila with homozygous loss-of-function mutation of IP 3 receptors were indistinguishable from wild-type controls in sensitivity, kinetics, and adaptation of response to light. Furthermore, most significantly, low IP 3 concentrations can induce substantial Ca 2ϩ release from the stores without activating Ca 2ϩ entry at all in rat leukemia cells, suggesting that even the activation of CCE is not that tightly coupled to Ca 2ϩ release from the IP 3 -sensitive stores (12).
Thus, criteria of activation trigger, other than depletion of Ca 2ϩ store, should be considered in functionally establishing cloned TRP channels to be correlated with native Ca 2ϩ channels responsible for receptor-activated Ca 2ϩ influx, including CCE.
We have here isolated cDNA that encodes a novel TRP homologue, TRP5, predominantly expressed in the brain. The recombinant expression of the TRP5 cDNA in human embryonic kidney (HEK) cells potentiated an extracellular Ca 2ϩ -dependent increase of [Ca 2ϩ ] i evoked by ATP, but not by an inhibitor of ER Ca 2ϩ -ATPases, thapsigargin. Whole-cell mode of patch-clamp recordings from TRP5-expressing cells demonstrated that ATP application induced a large inward current in the presence of extracellular Ca 2ϩ , which reversed at a positive potential. Our findings suggest that TRP5 directs the formation of a highly Ca 2ϩ -permeable ion channel that can be activated through receptor-operative pathways other than depletion of Ca 2ϩ from Ca 2ϩ stores in brain neurons.

EXPERIMENTAL PROCEDURES
cDNA Cloning and Sequence Determination-A mixture of oligo(dT)primed cDNAs synthesized from the mouse (BALB/c or 129/SvJ) brain poly(A) ϩ RNA was subjected to PCR amplification using a Marathon cDNA amplification kit (CLONTECH). Degenerate oligonucleotide primers used were 5Ј-TGGGGCC(T/C/A)(T/C)TGCAGAT(A/C)TC(T/A)-CTGGGA-3Ј and 5Ј-(G/T)G(A/T)TCG(A/G)GCAAA(C/T)TTCCA(C/T)TC-3Ј. Obtained PCR products were subsequently subcloned into the T/A cloning plasmid, pCRII (Invitrogen, Carlsbad, CA), to yield pTRP15. Sequence comparison with a PCR product amplified using a pair of specific oligonucleotide primers T5-1 (5Ј-TATCTACTGCCTAGTACTA-CTGG-3Ј) and T5-2 (5Ј-GCAATGAGCTGGTAGGAGTTATTC-3Ј) according to the partial genomic nucleotide sequence given in Zhu et al. (24) confirmed that the cDNA insert carried by pTRP15 encodes mouse TRP5. Further screening was performed to obtain entire coding cDNAs for mouse TRP5. Oligo(dT)-primed, size-selected (Ͼ1 kilobase pairs (kb)) cDNA libraries constructed in the Uni-ZAP XR vector (Stratagene, La Jolla, CA) using poly(A) ϩ RNA from the brain of adult BALB/c or postnatal 14-day-old (P14) C57BL/6J mice were screened to yield mouse TRP5 clones m37 (1020 -3166; nucleotide numbers from the first residue of the initiation ATG triplet) and O4 (1134 -3287 followed by a poly(dA) tract). Additional clones harboring cDNAs for the further upstream regions of mouse TRP5 were isolated by screening a specific oligonucleotide-primed cDNA transcripts of poly(A) ϩ RNA from the brain of P14 C57BL/6J mice constructed in the Uni-ZAP XR vector. The specific oligonucleotide primer (5Ј-GAGAGAGAGAGAGAGAGAG-AACTAGTCTCGAGTCAAGCAGCATTCGTCCC-3Ј) was according to the sequence in the clone m37 and was designed to contain an additional sequence of a XhoI site protected by GA repeats for subcloning into the Uni-ZAP XR vector. O15 (Ϫ232 to 1547) and the other 13 hybridization-positive clones were isolated through hybridization with the 657-base pair EcoRI(on vector)/BamHI(1665) fragment from m37. cDNA clones were sequenced on both strands using an automated sequencer (model 373S; Perkin Elmer).
Measurement of Changes in [Ca 2ϩ ] i -Cells on coverslips were loaded with fura-2 by incubation in Dulbecco's modified Eagle's medium containing 5 M fura-2/AM (Dojindo Laboratories, Kumamoto, Japan) and 10% fetal bovine serum at 37°C for 30 min, and washed with HEPESbuffered saline (HBS) containing (in mM): 107 NaCl, 6 KCl, 1.2 MgSO 4 , 2 CaCl 2 , 1.2 KH 2 PO 4 , 11.5 glucose, 20 HEPES, adjusted to pH 7.4 with NaOH. The coverslips were then placed in a perfusion chamber mounted on the stage of the microscope. Fluorescence images of the cells were recorded and analyzed with a video image analysis system (ARGUS-50/CA, Hamamatsu Photonics, Hamamatsu, Japan) according to the method of Hazama and Okada (38). The fura-2 fluorescence at an emission wavelength of 510 nm (bandwidth, 20 nm) was observed at room temperature by exciting fura-2 alternately at 340 and 380 nm (bandwidth, 11 nm). The 340/380 nm ratio images were obtained on a pixel by pixel basis, and were converted to Ca 2ϩ concentrations by in vitro calibration. The calibration procedure was performed according to Ueda and Okada (39). ATP (100 mM in water), SK&F96365 (10 mM in water), and thapsigargin (2 mM in dimethyl sulfoxide) were diluted to their final concentrations in HBS or Ca 2ϩ -free HBS containing (in mM): 107 NaCl, 6 KCl, 1.2 MgSO 4 , 1.2 KH 2 PO 4 , 0.5 EGTA, 11.5 glucose, 20 HEPES, adjusted to pH 7.4 with NaOH, and applied to the cells by perfusion. LaCl 3 and GdCl 3 (100 mM in water) were diluted in HBS or Ca 2ϩ -free HBS from which KH 2 PO 4 was omitted. The number of CD8positive cells ranged from 2 to 8 in the field of view during an experiment. Data were accumulated under each condition from two to four experiments using cells prepared through two to three transfections.
Electrophysiology-For electrophysiological measurements, coverslips with cells were placed in dishes containing the solutions. Cells prepared in this manner had membrane capacitance of 21.3 Ϯ 2.3 picofarads (n ϭ 25). Currents from cells were recorded at room temperature using patch-clamp techniques of whole-cell mode (40) with an EPC-7 patch-clamp amplifier (List-Medical, Darmstadt, Germany).
Patch pipettes were made from borosilicate glass capillaries (1.5 mm, outer diameter; Hilgenberg, Malsfeld, Germany) using a model P-87 Flaming-Brown micropipette puller (Sutter Instrument, San Rafael, CA). Pipette resistance ranged from 2 to 4 megohms when filled with the pipette solution described below. Currents were sampled at 200 Hz after low-pass filtered at 1 kHz (Ϫ3 dB) using an 8-pole Bessle filter (900, Frequency Devices, Haverhill, MA) for Rapid exchange of the external solutions was made by a modified "Y-tube" method (41). Fig. 1A shows the amino acid sequence of the mouse TRP5 deduced from the open reading frame of the corresponding cDNA sequence. The translation initiation codon is assigned to the first in-frame methionine codon downstream of a stop codon. TRP5 is composed of 975 amino acid residues with a hydropathy profile revealing eight hydrophobic segments and hydrophilic N and C termini (Fig.  1B), similar to those of other TRP subtypes (21,22,24,25,42,43). Sufficient length of hydrophobic regions to span the membrane, together with the lack of a hydrophobic N-terminal sequence indicative of the signal sequence, suggests that TRP5 is a membrane protein with a core of transmembrane segments and the flanking N-and C-terminal regions disposed on the FIG. 1. Primary structure and hydropathy analysis of TRP5, and phylogenetic tree of the TRP family. In A, the amino acid sequence (in single-letter code) of the mouse brain TRP5 deduced from the cDNA sequence is shown. The hydrophobic regions H1-H8 are enclosed with solid lines. The domains predicted to form coiled-coil structure are underlined with dashed lines. In B, the Kyte-Doolittle hydrophobicity profile of TRP5 was generated with a window size of 10 amino acids (66). In C, the phylogenetic tree for the TRP family was generated using the Clustalw program (46). Members of the TRP family are as follows: dTRP (42), dTRPL (43), hTRP3 (24), bTRP4 (25), ceTRP (67), and mTRP1 (68). cytoplasmic side, like other TRP subtypes (21-25, 42, 43). Domains that form coiled-coil structure were predicted on each side of the hydrophobic core ( Fig. 1A) (44). Potential cAMP-and cGMP-dependent protein kinase phosphorylation sites Ser 122 and Thr 167 are assigned to the putative cytoplasmic regions. Fig. 1C depicts the phylogenetic tree of the TRP family constructed by the neighbor-joining method (45), based on the sequence alignment carried out by the Clustalw program (46). Sequence identity/similarity between TRP5 and bCCE (25), a bovine counterpart of TRP4 (24), was relatively high (67/79%), compared with identities/similarities between TRP5 and other TRP homologues (36 -46/57-66%). Homology of TRP5 with TRP homologues is localized in the N-terminal region and the hydrophobic core (data not shown).

Primary Structure of TRP5-
Tissue Distribution of TRP5-RNA preparations from different mouse tissues were subjected to Northern blot analysis using TRP5 cDNA as a specific probe ( Fig. 2A). TRP5 mRNA was exclusively detected in the mouse central nervous system. An intense TRP5 signal with size of ϳ8.5 kb was present in the forebrain region (olfactory bulb, cerebrum, and midbrain) and a relatively weak signal was detected in the hindbrain region (cerebellum and medulla-pons). This contrasts with the predominant localization of TRP3 and TRP4 in the forebrain region and hindbrain region, respectively (47). To detect trace levels of TRP5 RNA expression in the tissues other than the brain, a pair of primers was designed to amplify cDNA sequences within the TRP5 cDNA probe used in the Northern analysis (Fig. 2B). RT-PCR amplification for 40 cycles, which is beyond the exponential phase of amplification, and subsequent Southern blot hybridization using a TRP5-specific oligonucleotide probe, disclosed TRP5 expression not only in the brain regions, but also in liver, kidney, testis, and uterus (48). In addition to the main hybridizable PCR product of ϳ330 base pairs, which corresponds to the expected size, a second hybridizable product of ϳ250 base pairs was detected in the hindbrain region, liver, kidney, testis, and uterus, but not in the forebrain.
Functional Characterization of TRP5: Cytosolic Ca 2ϩ Measurements-HEK293 cells are capable of serving as an excellent expression system for studying functional properties of TRP5 as a receptor-activated Ca 2ϩ channel, inasmuch as they have been known to endogenously express the P 2 purinoceptor coupled to activation of G q protein and phospholipase C (49). HEK cells were also reported for the absence of endogenous TRP5 expression (48). TRP5, together with a marker protein CD8, was transiently expressed in HEK cells, and intracellular Ca 2ϩ concentration was monitored in transfectants and nontransfected control HEK cells using fura-2 as an indicator. In the presence of 2 mM extracellular Ca 2ϩ , application of 100 M ATP to control cells induced a rapid rise in [Ca 2ϩ ] i that peaked within 30 s and gradually decreased to the resting level within 300 s (Fig. 3A). This transient rise in [Ca 2ϩ ] i was presumed to be mainly due to release from the intracellular Ca 2ϩ store, because omission of extracellular Ca 2ϩ little affected on the peak level (Fig. 4A), whereas the decay phase was accelerated in the absence of extracellular Ca 2ϩ . When 100 M ATP was applied to TRP5-transfected, CD8-positive cells in the presence of extracellular Ca 2ϩ , the peak [Ca 2ϩ ] i level greatly increased (Fig. 3B). In TRP5-expressing cells, Ca 2ϩ influx across the plasma membrane was likely to be a major cause of the [Ca 2ϩ ] i rise, because the amplitude of [Ca 2ϩ ] i rise was much smaller in the absence of extracellular Ca 2ϩ than that in the presence of extracellular Ca 2ϩ at ATP concentrations above 1 M (Fig. 3C), and remained almost the same as that of [Ca 2ϩ ] i transient evoked in control cells in the absence of extracellular Ca 2ϩ (Fig.  4, A and B). [Ca 2ϩ ] i rise evoked by ATP in TRP5-expressing cells in the presence and absence of extracellular Ca 2ϩ , and in control cells in the presence of extracellular Ca 2ϩ increased in a similar dose-dependent manner (Fig. 3C).
To separate contribution to [Ca 2ϩ ] i rise of Ca 2ϩ influx from that of Ca 2ϩ release, ATP was first applied in the absence of extracellular Ca 2ϩ , and 2 mM Ca 2ϩ was then added to the extracellular solution when [Ca 2ϩ ] i returned to the resting level (3 min after addition of ATP). Addition of Ca 2ϩ to the extracellular solution only slightly raised [Ca 2ϩ ] i above the resting level in control cells (Fig. 4, A and C), whereas in TRP5-transfected cells, it elicited dramatic [Ca 2ϩ ] i transients (Fig. 4, B and C), which reached maximum in the presence of ATP Ն 10 M (Fig. 4C). The second [Ca 2ϩ ] i rise evoked by extracellular Ca 2ϩ did not seem to correlate with the preceding first [Ca 2ϩ ] i rise caused by ATP-dependent Ca 2ϩ release from the intracellular Ca 2ϩ store. The second [Ca 2ϩ ] i rise was 613 Ϯ 77 (mean Ϯ S.E.) nM for the TRP5-expressing cells that showed the first rise staying below 10 nM (n ϭ 16), and was not significantly different from that (452 Ϯ 40 nM) observed in the cells where the first rise was above 10 nM (n ϭ 48). The time lag between start of ATP stimulation and addition of extracellular Ca 2ϩ did not significantly affect amplitude of [Ca 2ϩ ] i rise up to 5 min (Fig. 4D). Interestingly, after 3 min of stimulation by ATP in Ca 2ϩ -free solution, thapsigargin induced intact [Ca 2ϩ ] i rises in untransfected cells (102 Ϯ 4 nM, n ϭ 53), as compared with control cells without ATP stimulation (113 Ϯ 5 nM, n ϭ 51) (see below). Furthermore, after initial application of ATP for 3 min and subsequent omission of ATP up to 5 min in Ca 2ϩ -free solution, untransfected cells did not show significant [Ca 2ϩ ] i rise induced by the second application of ATP (n ϭ 38). These results suggest that ATP receptors are rapidly desensitized by incubating with 100 M ATP, and thereby internal stores are replenished with Ca 2ϩ within 3 min. Without ATP, [Ca 2ϩ ] i rise was not observed in TRP5-transfected cells that were immersed in the Ca 2ϩ -containing solution after preincubation in the Ca 2ϩ -free solution for up to 7 min (data not shown).
Lanthanides La 3ϩ and Gd 3ϩ were reported to block currents induced by recombinant expression of Drosophila TRP, TRPL, human TRP1, TRP3 (23,24,50), and native Ca 2ϩ channels (51)(52)(53) and I CRAC (54). The imidazol derivative, SK&F96365, inhibits various types of ion channels including receptor-activated channels (55,56). In Fig. 5, 100 M ATP was added alone (Fig. 5A) or together with one of the agents (25 M SK&F96365 (Fig. 5B) or 100 M La 3ϩ (Fig. 5C)) to the Ca 2ϩ -free extracellular solution, and 2 mM Ca 2ϩ was added 3 min later. As shown in Fig. 5 (B-D), 25 M SK&F96365 and 100 M La 3ϩ signifi- cantly suppressed the second [Ca 2ϩ ] i increase due to Ca 2ϩ influx. However, compared with the two agents, the effect of 100 M Gd 3ϩ on the amplitude of the second [Ca 2ϩ ] i rise was not as significant (Fig. 5D). These results indicate that ATPinduced Ca 2ϩ influx in TRP5-transfected cells is sensitive to blockade by SK&F96365 and La 3ϩ (Fig. 5D).

FIG. 4. Separation of ATP-induced [Ca 2؉ ] i transients due to Ca 2؉ release from internal stores and Ca 2؉ influx in TRP5-transfected HEK cells. Cytosolic Ca 2ϩ was measured in fura-2-loaded control HEK293 cells (A) or HEK293 cells transfected with TRP5 plus CD8 (B). In
To examine whether the [Ca 2ϩ ] i transient due to TRP5mediated Ca 2ϩ influx is activated by depletion of the intracellular Ca 2ϩ store induced by IP 3 -dependent Ca 2ϩ release via phospholipase C stimulation, we used, instead of ATP, the specific inhibitor of sarcoplasmic and endoplasmic reticulum ATPases, thapsigargin (57). As the cells were perfused with Ca 2ϩ -free solution containing 2 M thapsigargin, [Ca 2ϩ ] i was transiently increased and thereafter reduced to the basal level (Fig. 6, A and B). Subsequent addition of Ca 2ϩ (2 mM) to the extracellular solution transiently increased [Ca 2ϩ ] i in TRP5transfected cells (Fig. 6B) to levels similar to those in control cells (Fig. 6A), indicating that CCE is not potentiated by expression of TRP5. This observation suggests that TRP5 is not activated by store depletion.
We further tested whether ATP is capable of activating Ca 2ϩ influx in the TRP5-expressing cells where thapsigargin-induced Ca 2ϩ influx is already activated. When cells were perfused with the solution containing 2 mM Ca 2ϩ and 2 M thapsigargin, transient [Ca 2ϩ ] i increase developed similarly in the control and transfected cells, decreasing within 1000 s after addition of thapsigargin to stationary levels that stayed slightly above the initial basal levels (Fig. 6, C and D). Replacement of thapsigargin with 100 M ATP transiently increased [Ca 2ϩ ] i without any latency in TRP5-transfected cells (Fig. 6D), whereas it slightly decreased [Ca 2ϩ ] i in control cells (Fig. 6C). The results overall indicate that ATP activates Ca 2ϩ influx by a trigger different from depletion of the Ca 2ϩ store.
Functional Characterization of TRP5: Electrophysiological Measurements-To directly demonstrate that TRP5 is responsible for Ca 2ϩ influx activated by ATP in HEK cells, ionic currents triggered by ATP stimulation in TRP5-transfected cells were characterized in comparison with those in control nontransfected HEK cells, using whole-cell mode of patchclamp. When 200 nM free Ca 2ϩ was present in the patch pipette, and ATP was added to the 0Ca 2ϩ external solution, 16 out of 18 CD8-positive, TRP5-transfected HEK cells showed inward currents accompanied with an increase in the amplitude of current fluctuation (Fig. 7B). Rapid exchange of the 0Ca 2ϩ external solution with the 10Ca 2ϩ external solution induced a rapid development of large inward currents, followed by a gradual increase of inward currents in 7 out of the 16 cells (Fig. 7B). The remaining nine cells did not show further increase of currents by the solution change. The augmentation of FIG. 7. Electrophysiological characterization of the TRP5 channel. In A, shown is a time course of ionic current recorded from a control HEK293 cell at a holding potential of Ϫ50 mV. During application of ATP (indicated by the hatched bar above the current) to the control HEK293 cell, the external solutions were changed from the 0Ca 2ϩ solution (open bar) to the 10Ca 2ϩ solution (filled bar). Finally, ATP was washed out with the 0Ca 2ϩ solution. In B, a time course of ionic current recorded from a HEK293 cell transfected with TRP5 plus CD8 is shown. In C, current-voltage relationships of the TRP5 channel are shown. Currents were evoked by 1.5-s negative voltage ramps from 40 to Ϫ70 mV. Five consecutive ramps were applied every 5 s in the 0Ca 2ϩ solution or the 10Ca 2ϩ solution with 100 M ATP. The averaged currents were drawn. The currents shown in B and C were recorded from different TRP5-transfected cells. A and B, the perfusion solution was changed to Ca 2ϩfree HBS containing 0.5 mM EGTA, and 2 M thapsigargin (TG) was applied to the cells in the absence of extracellular Ca 2ϩ , which was followed by the addition of 2 mM extracellular Ca 2ϩ . In C and D, the cells were treated with 2 M thapsigargin in the presence of extracellular Ca 2ϩ , then thapsigargin was replaced with 100 M ATP. The duration of exposure to Ca 2ϩ -containing HBS, Ca 2ϩ -free HBS, 100 M ATP, and 2 M thapsigargin is indicated by the filled, open, hatched, and shaded bars, respectively, above the graphs. Data points are the means Ϯ S.E.

Thapsigargin-induced [Ca 2؉ ] i transients and ATP-induced [Ca 2؉ ] i changes after store depletion in control and TRP5-transfected HEK cells. Cytosolic Ca 2ϩ was measured in fura-2-loaded control HEK293 cells (A and C) or HEK293 cells transfected with TRP5 plus CD8 (B and D). In
[Ca 2ϩ ] i in the indicated number of cells. the inward current was abolished quickly upon the removal of Ca 2ϩ , although the fluctuation of currents remained until ATP was washed out. In the control cells (n ϭ 7), 100 M ATP did not induce significant ionic currents regardless of the presence of 10 mM Ca 2ϩ in extracellular solution (Fig. 7A). When Ca 2ϩ concentration in pipette solution was reduced to 50 nM, similar proportion of the CD8-positive, TRP5-transfected cells (11 out of 12 cells) showed responsiveness to ATP. However, usage of 10 nM free Ca 2ϩ intrapipette solution resulted in a slightly reduced number of the CD8-positive, TRP5-transfected HEK cells responsive to ATP, inducing inward currents in six out of nine cells. The CD8-positive, TRP5-transfected HEK cells measured using the EGTA-containing, Ca 2ϩ -free pipette solution were not responsive to ATP (n ϭ 5).
Current-voltage relationship of ionic current triggered by ATP in TRP5-expressing cells was examined using negative voltage ramps from 40 to Ϫ70 mV for 1.5 s. Five consecutive voltage ramps were applied. The averages of current traces generated by five consecutive ramps every 5 s in the 0Ca 2ϩ external solution and the 10Ca 2ϩ external solution with 100 M ATP were drawn in Fig. 7C. The current-voltage relationship recorded in 10Ca 2ϩ was nonlinear, showing a significant inward current at physiological potentials. Permeability ratios among Na ϩ , Cs ϩ , and Ca 2ϩ were estimated on the basis of the Goldman-Hodgkin-Katz equation using the reversal potentials in Fig. 7C. On the assumption that activity coefficients are 0.3 for Ca 2ϩ and 0.75 for both Na ϩ and Cs ϩ , the reversal potentials of 8 mV in the 0Ca 2ϩ external solution and 17 mV in the 10Ca 2ϩ solution lead to permeability ratios P Ca :P Na :P Cs ϭ 14.3:1.5:1.

Activation Mechanism of TRP5 Essential for Receptor-activated Ca 2ϩ
Influx-In the present investigation, we have cloned and functionally characterized the mouse TRP homologue, designated as TRP5, predominantly expressed in the brain. Recombinant expression of the TRP5 cDNA in HEK cells potentiated transient increases in [Ca 2ϩ ] i evoked by ATP in the presence of extracellular Ca 2ϩ (Fig. 3, A and B). When Ca 2ϩ was added to the extracellular solution after preincubating the cells in the Ca 2ϩ -free solution under constant ATP stimulation, potentiation of transient [Ca 2ϩ ] i rise induced by TRP5 expression became more prominent (Fig. 4, A and B). In this experiment, the second [Ca 2ϩ ] i rise due to Ca 2ϩ influx showed no significant correlation with the first [Ca 2ϩ ] i rise due to Ca 2ϩ release from IP 3 -sensitive internal stores in the absence of extracellular Ca 2ϩ . In an extreme case, [Ca 2ϩ ] i rise through Ca 2ϩ influx was induced in a TRP5-expressing cell where Ca 2ϩ release was hardly detectable. When thapsigargin was substituted for ATP to deplete the internal Ca 2ϩ store by inhibiting ER Ca 2ϩ -ATPases, the second [Ca 2ϩ ] i rise due to CCE was not potentiated by TRP5 expression (Fig. 6, A and B). These results indicate that TRP5 is responsible for Ca 2ϩ influx activated by ATP via mechanisms other than Ca 2ϩ depletion from the internal Ca 2ϩ store.
The independence of TRP5-activating cascades from depletion of the internal Ca 2ϩ stores is confirmed by additional lines of experimental evidence. In the presence of extracellular Ca 2ϩ , after the thapsigargin-induced [Ca 2ϩ ] i transient decayed to a plateau level, another [Ca 2ϩ ] i rise was induced in TRP5-expressing cells by ATP (Fig. 6D), which did not induce [Ca 2ϩ ] i transients and even elicited slight decreases of [Ca 2ϩ ] i in control cells (Fig. 6C). The lack of ATP-induced Ca 2ϩ release from ER in control cells after thapsigargin treatment (Fig. 6C) indicates that the ATP-sensitive stores are included in the thapsigargin-sensitive stores, excluding the possibility that TRP5 is activated via depletion of the ATP-sensitive stores independent of the thapsigargin-sensitive stores. It is also unlikely from this finding that TRP5 activation is directly coupled with the Ca 2ϩreleasing process involving the IP 3 receptors. Furthermore, our results indicate that after 3 min of stimulation by ATP in Ca 2ϩ -free solution, TRP5 channels are still activable (Fig. 4B), but endogenous receptor-activated channels including CCE channels, whose activation by thapsigargin is clearly seen in Fig. 6A, are dormant in HEK cells (Fig. 4A). This is presumably due to differences between the TRP5 channel and endogenous channels in susceptibility to effects of ATP receptor desensitization. Desensitization of ATP receptors within 3 min is suggested from our experimental observation that in Ca 2ϩ -free solution, [Ca 2ϩ ] i transient was not any more induced by ATP after initial application of ATP for 3 min and subsequent washing out of ATP for 5 min. Rapid desensitization of ATP receptors, compared with other types of receptors, was also reported by other groups (49). After a 3-min application of ATP in Ca 2ϩ -free solution, [Ca 2ϩ ] i rise induced by subsequent application of thapsigargin was intact, suggesting that internal stores were rapidly replenished with Ca 2ϩ . Thus, after desensitization of ATP receptors and replenishment of Ca 2ϩ stores, activation signals for TRP5 channel still persist, whereas the activation trigger for CCE is already abolished.
From our experiments, some insights can be gained into the activation mechanism for TRP5. Present data imply an important role of Ca 2ϩ in activation of TRP5. Whole-cell inward currents in ATP-stimulated, TRP5-transfected cells measured using the pipette solution containing free Ca 2ϩ exhibited rapid and dramatic increases upon addition of Ca 2ϩ to the extracellular solution (Fig. 7B), whereas those that were measured using the EGTA-containing, Ca 2ϩ -free pipette solution were not responsive to ATP. [Ca 2ϩ ] i could be lowered to 10 nM, which is considerably lower than physiological [Ca 2ϩ ] i , to elicit TRP5 current in HEK cells, whereas higher percentage of the TRP5 current-positive cells was obtained when [Ca 2ϩ ] i was elevated to 50 nM and 200 nM. This observation suggests that TRP5 is activable in the physiological range of [Ca 2ϩ ] i even at basal levels. Ca 2ϩ may act through Ca 2ϩ -binding proteins such as calmodulin and Ca 2ϩ -dependent enzymes, although it is yet preliminary to make any conclusion with regard to the Ca 2ϩ effect.
In thapsigargin-treated, TRP5-expressing cells, ATP primed the activity of TRP5 presumably not through [Ca 2ϩ ] i elevation (Fig. 6D), given that the action of ATP on [Ca 2ϩ ] i was toward decrease from slightly elevated levels in thapsigargin-treated control cells (Fig. 6C). This excludes the possibility that Ca 2ϩ is a sole activation trigger for TRP5, strongly suggesting involvement of other factors in TRP5 activation. It is possible that slight decrease from the elevated level (in the presence of thapsigargin) optimizes [Ca 2ϩ ] i in the range that activates but does not inactivate TRP5. Activation of G q protein, phospholipase C-␤, and protein kinase C, and production of phosphoinositide metabolites such as IP 3 and IP 4 , which are all triggered by stimulation of ATP receptors, should be considered as candidate activators of TRP5.
The results obtained are also indicative of a role of Ca 2ϩ as a negative regulator for TRP5. In the presence of extracellular Ca 2ϩ , [Ca 2ϩ ] i transients induced by ATP stimulation decreased almost to the basal level (Fig. 3B) at the time when the second [Ca 2ϩ ] i rise induced by Ca 2ϩ addition with time lag of 3 min after ATP stimulation reached peak (Fig. 4B). Negative regulatory action of Ca 2ϩ has been reported for Drosophila TRP, TRPL (30), and CCE in Xenopus oocytes (58).
Human TRP3 has been reported to form a nonselective cation channel that is not sensitive to Ca 2ϩ store depletion (59,60). Specifically, Zitt et al. (59) have shown that Ca 2ϩ neither act alone or act together with calmodulin directly on the TRP3 protein to activate the channel. Since the submission of the first version of this manuscript, we have learned that mouse TRP6 encodes a nonselective cation channel stimulated by the muscarinic M5 receptor, but not by intracellular store depletion (61). It is therefore possible that the Ca 2ϩ -selective TRP5 channels are activated via common activation mechanisms that operate in opening the TRP3 channel and/or the TRP6 channel.
Functional Correlation of TRP Homologues and Native Receptor-activated Ca 2ϩ Channels-Functional correspondence between cloned TRP homologues and Ca 2ϩ channels responsible for receptor-activated Ca 2ϩ influx, including CCE, in native preparations is still very controversial. Ca 2ϩ selectivity in permeation has been one of the important criteria in correlating recombinant TRP homologues with native Ca 2ϩ channels (32). Our whole-cell current measurements using patch pipettes filled with the solution containing 200 nM free Ca 2ϩ demonstrated that rapid exchange of the external 0Ca 2ϩ solution with the 10Ca 2ϩ solution elicited instantaneous and dramatic increases of inward TRP5 currents (Fig. 7B), which reversed at positive potentials (Fig. 7C). This, together with the permeability ratios (P Ca :P Na :P Cs ϭ 14.3:1.5:1) calculated from the reversal potentials, indicates that TRP5 is selective for Ca 2ϩ over monovalent cations. Of the other recombinantly expressed TRP homologues, Drosophila TRP and mammalian TRP4 were demonstrated for Ca 2ϩ selectivity (25,26,30), whereas Drosophila TRPL and mammalian TRP1 and TRP3 were rather classified as nonselective cation channels (23,27,30,59,62). In the native systems, the TRP and TRPL components of light-activated current isolated through usage of trpl and trp mutant photoreceptors showed ion selectivity comparable with those of the recombinant TRP and TRPL (19,63). Among the receptoractivated Ca 2ϩ channels in vertebrate cells, known for diversity in ion permeation properties (3), some display ion selectivity that may correspond well to TRP homologues. However, establishing functional correlation of TRP with native receptoractivated Ca 2ϩ channels becomes considerably unsuccessful by introduction of activation trigger as a second distinguishing criterion. Although I CRAC is similar to TRP5 in selectivity for Ca 2ϩ over Na ϩ , Ba 2ϩ , and Mn 2ϩ (64), 2 depletion of the intracellular Ca 2ϩ store activates I CRAC and the nonselective cation channels TRP1 (23) and TRP3 (62), but not Ca 2ϩ -selective TRP5. In contrast to TRP5, IP 4 and Ca 2ϩ -sensitive channels in endothelial cells are highly permeable not only to Ca 2ϩ , but also to other divalent cations such as Ba 2ϩ and Mn 2ϩ (16). It has been also reported that Ba 2ϩ or monovalent cations are as permeant as Ca 2ϩ in other receptor-activated Ca 2ϩ channels triggered by second messengers such as IP 3 or by activation of G-proteins (3). Thus, each vertebrate TRP homologue expressed in heterologous systems does not really correspond to the native receptor-activated Ca 2ϩ channels in both the two functional criteria: activation trigger and Ca 2ϩ selectivity.
Heteromultimer formation by multiple TRP isoforms (30) may be necessary to elicit native type receptor-activated Ca 2ϩ entry. In our expression studies of TRP5, there was no clear functional indication for presence of heterogeneous populations of heteromultimer and homomultimer, although Garcia and Schilling (48) have shown expression of TRP1, TRP3, TRP4, and TRP6 mRNAs in HEK 293 cells. This may derive from the usage of ATP receptor stimulation in activating Ca 2ϩ entry, or low mRNA expression levels of endogenous TRP isoforms compared with the level of TRP5 overexpression. It would be necessary to characterize functional properties of neuronal recep-tor-activated Ca 2ϩ channels at exact expression sites of individual TRP homologues determined by in situ hybridization (47,65) and immunohistochemistry in native tissues, and to compare them with those of recombinant receptor-activated channels composed of appropriate TRP combinations.