Receptor-activated Ca2+ Influx via Human Trp3 Stably Expressed in Human Embryonic Kidney (HEK)293 Cells

Ca2+ release from its internal stores as a result of activation of phospholipase C is accompanied by Ca2+ influx from the extracellular space. Ca2+influx channels may be formed of proteins homologous toDrosophila Trp. At least six non-allelic Trpgenes are present in the mouse genome. Full-length human, bovine, mouse, and rat cDNAs for Trp1, 3, 4, 6 have been cloned. Expression of these genes in various mammalian cells has provided evidence that Trp proteins form plasma membrane Ca2+-permeant channels that can be activated by an agonist that activates phospholipase C, by inositol 1,4,5-trisphosphate, and/or store depletion. We have stably expressed human Trp3 (hTrp3) in human embryonic kidney (HEK)293 cells. Measurement of intracellular Ca2+ concentrations in Fura2-loaded cells showed that cell lines expressing hTrp3 have significantly higher basal and agonist-stimulated influxes of Ca2+, Mn2+, Ba2+, and Sr2+ than control cells. The increase in Ca2+ entry attributable to the expression of hTrp3 obtained upon store depletion by thapsigargin was much lower than that obtained by stimulation with agonists acting via a Gq-coupled receptor. Addition of agonists to thapsigargin-treated Trp3 cells resulted in a further increase in the entry of divalent cations. The increased cation entry in Trp3 cells was blocked by high concentrations of SKF 96365, verapamil, La3+, Ni2+, and Gd3+. The Trp3-mediated Ca2+ influx activated by agonists was inhibited by a phospholipase C inhibitor, U73122. We propose that expression of hTrp3 in these cells forms a non-selective cation channel that opens after the activation of phospholipase C but not after store depletion. In addition, a subpopulation of the expressed hTrp3 may form heteromultimeric channels with endogenous proteins that are sensitive to store depletion.

The activities of a large number of enzymes are regulated through changes of intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ). 1 Under resting conditions, cells keep [Ca 2ϩ ] i at approximately 100 nM. A rise of the cytosolic Ca 2ϩ triggers a cascade of Ca 2ϩ -sensitive events that are both immediate, such as secretion, contraction, and mobilization of energy resources (e.g. glycogenolysis), and long term, such as changes in the transcription of many genes. Some of the Ca 2ϩ -responsive transcription processes are known to cause proliferation and programmed cell death (1)(2)(3). In both nonexcitable and excitable cells, Ca 2ϩ signaling pathways can be activated by a ligand binding to cell surface receptors that activate phospholipase C (PLC). These include receptors that activate heterotrimeric G proteins and receptors signaling through activation of protein tyrosine kinases. The activation of PLC leads to the production of inositol 1,4,5-trisphosphate (IP 3 ), which binds to IP 3 receptors, a class of intracellular ligand-operated Ca 2ϩ release channels. The opening of IP 3 receptors allows Ca 2ϩ to exit from its internal storage pools, causing a rapid increase in [Ca 2ϩ ] i . The increased cytosolic Ca 2ϩ level is reduced quickly by Ca 2ϩ pumps located on both the endoplasmic reticulum and the plasma membrane, causing [Ca 2ϩ ] i to decrease. The Ca 2ϩ signal is prolonged, however, by the opening of a set of plasma membrane Ca 2ϩ -permeant channels that allow Ca 2ϩ to enter cells from the extracellular space, where the concentration of Ca 2ϩ is in the millimolar range. In many, and possibly all cells, the entering Ca 2ϩ is taken up rapidly by a storage compartment from which it is re-released in the continued presence of the triggering extracellular signal. This may, although not necessarily, be accompanied by periodic oscillations of the [Ca 2ϩ ] i and allows for regulation of cytosolic as well as membrane-associated functions of the affected cells (for reviews, see Refs. 1 and 4).
Putney (5,6) coined the term capacitative Ca 2ϩ entry (CCE) for Ca 2ϩ entry that is activated upon stimulation of cells with agonists that promote Ca 2ϩ release from internal stores. After the discovery that CCE can be stimulated by mere store depletion, as occurs after inhibition of sarcoplasmic endoplasmic reticulum Ca 2ϩ -activated ATPases with thapsigargin (TG) (7,8), channels that mediate this type of CCE have been referred to as store-operated channels. In most cells, Ca 2ϩ entry after stimulation by agonists can be explained by activation of storeoperated channels, but neither the mechanism of how the store depletion signal is transmitted to the plasma membrane nor the molecular nature of the plasma membrane channels mediating Ca 2ϩ has been clearly identified. Moreover, although store-operated plasma membrane Ca 2ϩ -permeant channel activities have been found in many cell types (9 -13), channels regulated by other mechanisms, such as a G protein, IP 3 , inositol 1,3,4,5-tetrakisphosphate (IP 4 ), Ca 2ϩ , or ATP, seem to coexist with store-operated channels (14 -17). The molecular basis of this diversity and the relation of store depletion-insen-sitive Ca 2ϩ -permeant channels to store depletion-sensitive Ca 2ϩ -permeant channels are unclear. However, it is likely that multiple Ca 2ϩ influx channels are involved in Ca 2ϩ entry initiated by the activation of PLC-linked receptors.
In Drosophila eyes, the light-induced current is carried by channels formed from at least two related photoreceptor-specific proteins, Trp and Trpl (Trp-like). Trp was identified by molecular cloning as a protein missing from the transient receptor potential (trp) mutant (18). Trpl was identified biochemically as a calmodulin-binding protein and was cloned by standard techniques revealing a structure that shares a high degree of homology with Trp (19). Because the phototransduction pathway in insects resembles the PLC/IP 3 signaling pathway in mammalian cells, it was speculated that mammalian homologs of Trp may exist and form Ca 2ϩ influx channels. In support of this, expression of Trpl in Sf9 cells led to development of an agonist-stimulated non-selective Ca 2ϩ -permeable ion conductance, and that of Trp led to the development of a TG-stimulated ion conductance (20 -22). However, neither Trpformed channels nor Trpl-formed channels displayed the ion permeation properties of endogenous insect CCE channels and, while forming channels, Trpl was not activated by TG treatment. More recently, work from the laboratories of Minke (23) and Montell (24) showed that Trp and Trpl can form heteromultimeric ion channels with properties that differ from those of their parental molecules, suggesting that native voltageindependent Ca 2ϩ influx channels in insects are likely to be formed of more than one type of subunit, i.e. to be heteromultimers.
Several mammalian sequences homologous to the Drosophila Trps 2 have been identified from the data base of expressed sequence tags and by reverse transcriptase-polymerase chain reaction using degenerate oligonucleotide primers based on the Drosophila sequences (25)(26)(27)(28)(29)(30). In the mouse we found the existence of at least six non-allelic Trp genes that can be divided into four major types based on primary amino acid sequence similarity (26,31). Full-length cDNAs for Trp1, 3, 4, and 6 from human, murine, rat and bovine sources have been reported (25-27, 29 -32). Functional expression of human Trp1, Trp3, bovine Trp4, or mouse Trp6 in COS, Chinese hamster ovary, or human embryonic kidney (HEK)293 cells results in enhancement of either agonist-stimulated Ca 2ϩ entry or of IP 3 -or TG-stimulated inward currents that are at least in part carried by Ca 2ϩ (26,29,31,32). More importantly, agonist-stimulated Ca 2ϩ influx in murine L cells was blocked by transfection of murine Trp cDNA fragments in the antisense direction (26,31), showing that mammalian Trp proteins are involved in agoniststimulated Ca 2ϩ influx.
For the present work we developed stable HEK293 cell lines expressing human Trp3 (hTrp3) and studied their Ca 2ϩ influx properties. We report that the expression of hTrp3 caused an increase of [Ca 2ϩ ] i under basal and agonist-stimulated conditions. The influx pathway formed by hTrp3 permeates Ca 2ϩ , Sr 2ϩ , and Ba 2ϩ equally well, whereas the pathways intrinsic to HEK293 cells seem to be more Ca 2ϩ selective. The influx due to hTrp3 can be blocked by SKF 96365, a known CCE blocker, or by verapamil, a nonspecific blocker for L-type voltage-sensitive Ca 2ϩ channels. Although treatment with TG caused a small increase of Ca 2ϩ influx over the basal in cell lines expressing hTrp3, a large portion of the influx pathway due to hTrp3 is not sensitive to store depletion and seems to depend primarily on the activation of PLC.

EXPERIMENTAL PROCEDURES
Materials-SKF 96365 hydrochloric acid was purchased from Calbiochem. TG, U73122, (Ϯ) verapamil were from Research Biochemicals International. Carbachol, 3-amino-9-ethyl-carbozole tablets, deoxycholic acid, phenylmethanesulfonyl fluoride, soybean trypsin inhibitor, leupeptin, and Nonidet P-40 were purchased from Sigma. 35  cDNAs and Expression Vectors-The wild type hTrp3 was subcloned into the mammalian expression vector pcDNA3 (26). The insert, located in between the EcoRI and the XbaI site of the polylinker of pcDNA3, contains nucleotide Ϫ15 (A of the first ATG is nucleotide 1) to 3285 of hTrp3 cDNA which includes a poly(A) tail (n ϭ 30) and is flanked by a 36-nucleotide sequence from the Marathon cDNA adaptor (CLON-TECH) on each side. To introduce an epitope for hemagglutinin (HA) at the C terminus of hTrp3, oligonucleotide A, 5Ј-CCCAGCATGCTGTAC-CCGTACGATGTTCCTGATTACGCGAGATGTGAATGATGCAGCA-3Ј, which contains a sequence (underlined) encoding the HA epitope (YPY-DVPDYA) in between codons for Leu-845 and Arg-846 at the C terminus of hTrp3 cDNA, was synthesized. A partial hTrp3 sequence was amplified by polymerase chain reaction using hTrp3 in pcDNA3 as a template and oligonucleotide A and Sp6 as primers. The polymerase chain reaction product was subcloned back into the original plasmid by ligation of an EcoRI/SphI fragment generated from hTrp3/pcDNA3, the SphI/XbaI fragment of the polymerase chain reaction product, and the EcoRI/XbaI fragment of pcDNA3. The presence of the HA epitope sequence in the hTrp3-HA/pcDNA3 plasmid was confirmed by sequencing. The rat V1aR/pcDNA3 expression vector was a gift from Dr. Mariel Birnbaumer.
Cell Lines and Cell Culture Conditions-HEK293 cells were cultured in Dulbecco's modified Eagle's medium containing 4.5 mg/ml glucose, 10% heat-inactivated fetal bovine serum, 50 units/ml penicillin, and 50 mg/ml streptomycin. The hTrp3-HA/pcDNA3 (100 ng) was transfected into the HEK cells, plated at 2.8 ϫ 10 6 cells/10-cm tissue culture dish 20 h before transfection, by the calcium phosphate/glycerol shock method (33). After 24 h, cells were harvested, diluted in medium supplemented with 400 g/ml G418, and transferred to wells of 96-well plates at different dilutions. G418-resistant transformants were expanded into 12-well plates. To identify the clones expressing hTrp3-HA, cells were seeded in 96-well plates precoated with poly-D-lysine (100 g/ml). Immunocytochemical staining was performed as described by Vannier et al. (34) using a monoclonal HA antibody (12CA5) as the primary antibody, anti-mouse IgG conjugated with peroxidase (Amersham) as the secondary antibody, and 3-amino-9-ethyl-carbozole as the colorant. Positive cells were stained red and were visualized through a light microscope. Eleven cell lines expressing the HA epitope were further expanded and analyzed for the expression of HA-tagged hTrp3 by immunoprecipitation with 12CA5 as described below. All cell lines synthesized an immunoprecipitated protein of approximately 100 kDa when analyzed by SDS-polyacrylamide gel electrophoresis, which corresponds to the predicted size of hTrp3. Two cell lines, HEKTrp3-9 (T3-9) and HEKTrp3-65 (T3-65) were used for further analysis. Control cells were transfected with V1aR/pcDNA3 under the same conditions, two G418-resistant clones were randomly selected and designated as control-1 (C1) and control-2 (C2). The stable cell lines were diluted twice weekly and maintained in medium supplemented with 400 g/ml G418.
Immunoprecipitation-Cells (2 ϫ 10 6 ) were plated in 6-cm tissue culture dishes at least 16 h before the experiments. Cells were washed once with Hanks' balanced salt solution and then incubated in 1 ml of methionine/cysteine-free Dulbecco's modified Eagle's medium (ICN) containing 5% fetal bovine serum at 37°C for 1 h. The medium was replaced by the same medium containing 50 Ci/ml 35  from the culture dish in 1 ml of ice-cold Dulbecco's phosphate-buffered saline containing 1 mM EDTA and 0.1 mM phenylmethanesulfonyl fluoride, and pelleted by centrifugation at 5,000 rpm in a microcentrifuge at 4°C for 5 min. Cells were then homogenized in 0.5 ml of RIPA buffer (150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS, 50 mM Tris-HCl, pH 8.0) containing 0.1 mM phenylmethanesulfonyl fluoride, 1 g/ml soybean trypsin inhibitor, and 0.5 g/ml leupeptin. Immunoprecipitation was performed using the anti-HA monoclonal antibody (12CA5) as described by Innamorati et al. (35). For glycosidase treatment, cell lysate was incubated with peptide-N-glycosidase F (2 units/ml) or endoglycosidase H (10 milliunits/ml) at room temperature for 1 h before antibody (12CA5) was added. The immunoprecipitated proteins were eluted in 80 l of 2 ϫ Laemmli buffer (1 ϫ ϭ 62.5 mM Tris-HCl, 1% SDS, 10% glycerol, 10% ␤-mercaptoethanol, pH 6.8) and separated by SDS-polyacrylamide gel electrophoresis in 9% acrylamide gels at 40 mA for 2 h. The gels were stained with Coomassie Blue to visualize the molecular markers, destained, and then dried for autoradiography.
Measurement of [Ca 2ϩ ] i -Changes of [Ca 2ϩ ] i in individual cells were monitored after loading cells with Fura2 by fluorescence videoimaging microscopy using an Attofluor Digital Imaging and Photometry attachment of a Carl Zeiss Axiovert inverted microscope as described before (26).
Changes of [Ca 2ϩ ] i in cell populations were measured using an Aminco-Bowman Series 2 luminescence spectrofluorometer (SLM Instruments, Inc.). Briefly, cells grown to confluence in 15-cm dishes were trypsinized and collected in a 50-ml capped polypropylene tube. After a centrifugation for 5 min at 400 ϫ g and removal of the supernatant, the cell pellets were resuspended at room temperature in an extracellular solution (ECS) composed of 140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 10 mM glucose, 0.1% bovine serum albumin, 15 mM Hepes, pH 7.4. Cells were incubated in ECS supplemented with 5 M Fura2/AM and 0.05% Pluronic F-127 at 37°C for 40 min. Cells were then washed once and resuspended in ECS at 2 ϫ 10 6 cells/ml. Aliquots of 2 ml were kept in the dark at room temperature until use. Before each measurement, cells were washed twice in ECS or twice in ECS to which no CaCl 2 was added if a Ca 2ϩ -free/Ca 2ϩ readdition protocol was used. 2 ml of cell suspension was transferred to a quartz cuvette and maintained at 32°C under continuous stirring. Changes in intracellular Fura2 fluorescence intensity were measured by alternating excitation at 340 and 380 nm at 3-s intervals and detecting emission at 510 nm. Autofluorescence of the cells at 340 and 380 nm was determined from unloaded cells of an equivalent cell density and subtracted from values obtained for the Fura2-loaded cells. Drugs were added by a small aliquot (Ͻ20 l) of concentrated stocks (dissolved either in water or dimethyl sulfoxyide) to achieve the final concentrations. At the end of each recording, 0.1% Triton X-100 was added to the cells to determine the maximal fluorescence ratio (R max ). This was followed by an addition of 20 mM EGTA to determine the minimal fluorescence ratio (R min ).
In most experiments, Ca 2ϩ entry under the basal or stimulated conditions was measured by a Ca 2ϩ -free/Ca 2ϩ readdition protocol, in which cells were incubated in a nominally Ca 2ϩ -free ECS stimulated or not by the addition of an agonist or a store depletion drug. The stimulation normally caused a transient [Ca 2ϩ ] i increase due to the release of Ca 2ϩ from intracellular Ca 2ϩ stores. After allowing the first [Ca 2ϩ ] i peak to decrease to a steady-state level (normally 3-7 min), CaCl 2 was added to give a final concentration of 1.8 mM. Typically, this led to a second [Ca 2ϩ ] i increase. This extracellular Ca 2ϩ -dependent [Ca 2ϩ ] i increase is assumed to be caused by Ca 2ϩ influx.
Measurements of Influxes for Sr 2ϩ , Ba 2ϩ , and Mn 2ϩ -A Ca 2ϩ free/ cation addition protocol, analogous to the Ca 2ϩ -free/Ca 2ϩ readdition protocol in which Sr 2ϩ or Ba 2ϩ was added during the cation readdition phase, was used to study the entry of Sr 2ϩ and Ba 2ϩ into cells. The extent of influx of these divalent cations is expressed in the figures as changes in the ratio of 340 to 380 nm fluorescence without the estimation of their intracellular concentrations.
In a similar manner, MnCl 2 was added to a final concentration of 50 M to cells incubated in a nominally Ca 2ϩ -free ECS under either unstimulated or stimulated conditions. Fluorescence quenching was studied using the Fura2 isosbestic excitation wavelength at 360 nm and recording emitted fluorescence at 510 nm. At the end of each recording, 0.1% Triton X-100 was added to cells to determine the maximal quenching of Fura2 by Mn 2ϩ . Results of fluorescence quenching are expressed (in percent) as the amount of fluorescence decreased from the initial value (before the addition of Mn 2ϩ ) divided by the maximal loss after the addition of Triton X-100.

RESULTS
Activity of HA Epitope-tagged hTrp3-We showed previously (26) that transient expression of hTrp3 in COS-M6 cells results in a 2-fold increase of Ca 2ϩ influx in response to the activation of a coexpressed M5 muscarinic receptor. To study the function and biochemical properties of hTrp3 in more detail, we decided to develop stable cell lines expressing hTrp3 at high levels. To aid the identification of cell lines expressing hTrp3, we added an HA epitope at the C terminus of hTrp3 (hTrp3-HA). The resulting cDNA construct was first transfected to COS-M6 cells, and its function on Ca 2ϩ influx was compared with the wild type hTrp3, which was transfected in parallel in the same experiment. Ca 2ϩ influx induced by 20 M carbachol (CCh) was identical in cells transfected with the wild type hTrp3 or hTrp3-HA ( Fig. 1). We transfected hTrp3-HA/pcDNA3 into HEK293 cells and selected 11 cell lines that expressed the HA-tagged hTrp3. Two cell lines, HEKTrp3-9 (T3-9) and HEK-Trp3-65 (T3-65), expressed relatively high levels of hTrp3 and were used in studies presented in this report. Two control cell lines (C1 and C2) were obtained by transfecting cDNA for the rat V1aR into the HEK293 cells. Thus the hTrp3 cells and the control cells underwent the same type of treatment and selection processes and were kept under the same culturing conditions.
Glycosylation of hTrp3-Immunoprecipitation of HA-tagged hTrp3 from metabolically labeled T3-9 and T3-65 cells showed multiple bands at around 100 kDa (bands 1, 2, 3, 4) and a haze of radioactivity above these bands ( Fig. 2A). Treatment with endoglycosidase H selectively eliminated only one of the middle bands (band 2) and increased the intensity of the lower band (band 1), suggesting that band 2 is a high mannose-containing immature form of the glycosylated hTrp3. Treatment with peptide-N-glycosidase F removed bands 2-4 and the haze while greatly enhancing the intensity of band 1. Thus, bands 2, 3, 4 and the haze of radioactivity seen on the SDS-polyacrylamide gel electrophoresis represent hTrp3 at different stages of glycosylation during remodeling of the protein-attached sugar moieties, whereas band 1 represents the non-glycosylated hTrp3. In COS cells transiently transfected with hTrp3-HA, the predominant protein was the high mannose-containing im-FIG. 1. Introduction of an HA epitope at the C terminus of hTrp3 does not alter its ability to enhance Ca 2؉ influx in response to stimulation of a G q -coupled receptor. COS-M6 cells were transiently cotransfected with cDNAs encoding wild type hTrp3 (a), hTrp3-HA (b), or murine luteinizing hormone receptor (c, control) plus the rat M5 muscarinic receptor as described previously (26). Changes of [Ca 2ϩ ] i in individual cells loaded with Fura2 were measured by videoimaging microscopy 40 h post-transfection as described in Zhu et al. (26). CCh at 20 M was added at 1 min and was present throughout the experiment as indicated by the horizontal bar. A solution containing 0.5 mM EGTA was replaced by a solution containing 1. mature glycosylated form (band 2), as shown by its sensitivity to the endoglycosidase H treatment (see Fig. 7 in Ref. 31). This is not surprising since transiently transfected COS cells are known to synthesize a large quantity of the exogenous proteins of which only a fraction is processed to mature cell surface form (37).
The maturation process of hTrp3 can be followed in a pulsechase experiment, as shown in Fig. 2B. After 15 min of labeling with a mixture of [ 35 S]Met and [ 35 S]Cys, the high mannose glycosylated hTrp3 was the major form. At 30 min, the more mature forms of glycosylated hTrp3 started to appear. At 60 and 120 min, the intensities of all of the bands increased proportionally compared with that at 30 min. The haze of radioactivity became more evident. When cells were labeled for 15 min and then the labeled 35 S was chased with culture medium containing unlabeled amino acids, the intensities of both the non-glycosylated and the high mannose-containing form of hTrp3 decreased as the chase time increased (Fig. 2C). . Ca 2ϩ influx was studied using the Ca 2ϩ -free/Ca 2ϩ readdition protocol described under "Experimental Procedures." CCh was used to activate the PLC pathway via an endogenous muscarinic receptor found in the HEK293 cells. Fig. 4A shows the abundance of the muscarinic receptor sites in each cell line determined according to Liao et al. (38). Similar results were obtained by stimulating cells with 100 M ATP, which activates an endogenous purinergic receptor. To study Ca 2ϩ influx induced by store depletion, we used 0.2 M TG to block the endoplasmic reticulum Ca 2ϩ -ATPase, which passively depletes the internal store without increasing the production of IP 3 (4). Fig. 4 summarizes results obtained for two Trp3 cells (T3-9 and T3-65) and two control cells (C1 and C2).
Although in a nominally Ca 2ϩ -free buffer, basal [Ca 2ϩ ] i in Trp3 cells is not significantly different from that of control cells; when the cells are maintained in a normal physiological solution containing 1.8 mM Ca 2ϩ , the resting [Ca 2ϩ ] i in Trp3 cells is 40 -70% higher than that in control cells (Fig. 4B). located on plasma membrane and endoplasmic reticulum, as well as other Ca 2ϩ -buffering systems inside cells, which actively extrude Ca 2ϩ out of the cell or into its intracellular stores. Opening of the Ca 2ϩ influx channels results in a rapid rise of [Ca 2ϩ ] i , which in turn causes an increase in extrusion of Ca 2ϩ mediated by the Ca 2ϩ pumps. Therefore, at any given time, [Ca 2ϩ ] i is a result of the dynamic interplay of Ca 2ϩ pumps, intracellular Ca 2ϩ channels (e.g. IP 3 receptors), and plasma membrane Ca 2ϩ influx channels. Upon readdition of Ca 2ϩ to the extracellular medium, three influx activities could contribute positively to the increase of [Ca 2ϩ ] i in CCh-or TG-stimulated Trp3 cells. The first is the basal activity mediated by Trp3. The second is influx intrinsic to the HEK cells stimulated by CCh or TG. The third is a Trp3-mediated influx stimulated by the same drugs. Because a significant fraction of Ca 2ϩ entering from external space is removed by the Ca 2ϩ pumps, the net increase of [Ca 2ϩ ] i will not be the sum of influx arising from the three activities. Therefore, from the experiments shown in Figs. 3 and 4, it is difficult to conclude whether stimulation by CCh or TG increases a Trp3-mediated Ca 2ϩ influx pathway or if the stimulated increase in [Ca 2ϩ ] i observed in the Trp3 cells results merely from the opening of Ca 2ϩ influx channels endogenous to the HEK cells. More difficult is to see whether store depletion causes more Ca 2ϩ influx in Trp3 cells than in control cells because the increase in [Ca 2ϩ ] i in the two Trp3 cell lines is not significantly higher than that in C2 cells (Fig. 4C). To test whether any stimulated Ca 2ϩ influx is caused by the expression of hTrp3, we selectively blocked the endogenous Ca 2ϩ influx with 10 M Gd 3ϩ . As shown in Fig. 5, A, B, and E), in the presence of 10 M Gd 3ϩ , the endogenous Ca 2ϩ influx activated by CCh and TG decreased Ͼ85% in control cells. In contrast, addition of 10 M Gd 3ϩ to the Trp3 cells reduced Ͻ10% of the basal Ca 2ϩ influx (Fig. 5E), about 15% of Ca 2ϩ influx stimulated by CCh (Fig. 5, C and E), and 55-60% stimulated by TG (Fig. 5, D and E). Assuming that the Trp3mediated influx pathway is insensitive to 10 M Gd 3ϩ , we conclude that for Trp3 cells treated with CCh, a significant proportion of Ca 2ϩ influx is mediated by Trp3, whereas in cells treated with TG, most of the Ca 2ϩ influx is carried by the endogenous CCE channels.
Although in the presence of 10 M Gd 3ϩ , the maximal increase in [Ca 2ϩ ] i in Trp3 cells in response to the addition of 1.8 mM Ca 2ϩ under basal and TG-stimulated conditions is very similar, the rate of [Ca 2ϩ ] i increase in TG-treated cells appears to be faster than in the non-stimulated cells (Fig. 6A). In fact, in both cell lines expressing hTrp3, maximal [Ca 2ϩ ] i was reached within 40 s after the addition of external Ca 2ϩ , and  Fig. 6B shows TG-stimulated Ca 2ϩ influx resistant to 10 M Gd 3ϩ in control and Trp3 cells. Date are from subtraction of [Ca 2ϩ ] i changes of TG-stimulated cells from that of non-stimulated cells. Control cells have very low CCE activity when 10 M Gd 3ϩ is present, whereas Trp3 cells have a small and transient CCE activity that lasts for less than 1 min. Therefore, although very small, there is an increase of storeoperated Ca 2ϩ influx due to expression of hTrp3 in the HEK cells. Furthermore, TG also appears to block the basal Ca 2ϩ influx seen in Trp3 cells. This could be due to the closing of the constitutively active hTrp3 channel or the enhancement of intracellular Ca 2ϩ removal. Neither function has been described for TG.
Divalent Cation Selectivity of Influx Channels in Control and the Trp3 HEK Cells-We compared the influx of Ba 2ϩ and Sr 2ϩ through the Trp3-mediated pathway in Trp3 cells and the endogenous pathway in control cells. Increasing concentrations of both Sr 2ϩ and Ba 2ϩ have been shown to produce a shift in Fura2 excitation wavelength spectrum similar to that produced by Ca 2ϩ , but with somewhat higher dissociation constants (39,40). Therefore, increases in the fluorescence ratio of Fura2 obtained at excitation wavelengths of 340 and 380 nm reflect increases of intracellular concentrations of Ba 2ϩ or Sr 2ϩ . Control or Trp3 cells were incubated in nominally Ca 2ϩ -free solutions either unstimulated or stimulated by CCh before 1.8 mM Sr 2ϩ or Ba 2ϩ was added into the ECS substituting for Ca 2ϩ . Fig. 7, A-D, shows that the rates of Ba 2ϩ and Sr 2ϩ influx into Trp3 cells are much higher than into control cells. Because Ba 2ϩ is a poor substrate for Ca 2ϩ pumps responsible for extrusion from the cytosol (39), a declining phase of intracellular [Ba 2ϩ ] is normally not seen, as is the case with Ca 2ϩ and Sr 2ϩ (Fig. 7, C and D). In CCh-stimulated control cells, the initial rate of fluorescence ratio increase with Ca 2ϩ (2.0 units/min) is at least two times faster than the rates with Sr 2ϩ (0.6 unit/min) and Ba 2ϩ (0.6 unit/min). In contrast, the rates of fluorescence ratio increase in Trp3 cells are very similar regardless of which divalent cation was added (5.5 for Ca 2ϩ , 6.8 for Sr 2ϩ , and 6.2 for Ba 2ϩ , in units/min). Therefore, in HEK293 cells, the endogenous influx pathway activated by CCh seems to be more selective for Ca 2ϩ than Ba 2ϩ and Sr 2ϩ , whereas the Trp3-mediated influx is not selective for these cations.
Mn 2ϩ can enter cells through certain types of Ca 2ϩ influx channels and quenches Fura2 fluorescence (41,42).  (Fig. 7E). The rate of Mn 2ϩ entry into Trp3 cells was faster than into control cells. In CCh-stimulated cells, Mn 2ϩ entry was increased in both the control and the Trp3 cells (Fig. 7F).
Effect of Ca 2ϩ Channel Blockers-Although not specific, SKF 96365 has been shown to block agonist-stimulated Ca 2ϩ influx in many cells (43). Fig. 8A shows that at 25 M, SKF 96365 inhibits completely the basal Ca 2ϩ influx of Trp3 cells. In CCh-treated cells, the drug inhibits Ca 2ϩ influx in both Trp3 and control cells. However, SKF 96365 seems to inhibit the Trp3-mediated influx more effectively than the endogenous influx. The residual Ca 2ϩ influx left in the Trp3 cells is comparable to that seen in control cells. The IC 50 for SKF 96365 was about 5 M as estimated from measuring the CCh-stimulated Ba 2ϩ influx in the Trp3 cells (Fig. 8B). Verapamil, a blocker of L-type voltage-gated Ca 2ϩ channels, not known to affect CCE or agonist-stimulated Ca 2ϩ entry in general, also prevents divalent cation entry via the Trp3-mediated pathway with an IC 50 of about 4 M (Fig. 8C). Both the basal and the stimulated cation influx in Trp3 cells can be blocked by high concentrations of Ni 2ϩ (6 mM), La 3ϩ (150 M), and Gd 3ϩ (200 M).
PLC-dependent and Store Depletion-insensitive Cation Influx in Trp3 Cells-To explore further the activation mechanism of Trp3-mediated Ca 2ϩ influx, we tested whether depleting internal stores with TG would prevent agonist-stimulated [Ca 2ϩ ] i increase. In control cells maintained in a nominally Ca 2ϩ -free solution, the addition of 200 M CCh after incubating cells with 200 nM TG for 6 min caused no significant [Ca 2ϩ ] i increase (not shown). In the Trp3 cells, a small rise in [Ca 2ϩ ] i (27 Ϯ 6 nM net increase at the peak, n ϭ 4) was induced by CCh after cells had been incubated with TG for 6 min. However, when the cells were treated with TG in a medium containing 1.8 mM Ca 2ϩ and the endogenous influx pathway was blocked by 10 M Gd 3ϩ , the addition of CCh caused a transient increase of [Ca 2ϩ ] i of 105 Ϯ 8 nM (n ϭ 11) in Trp3 cells (Fig. 9A). A similar increase of [Ca 2ϩ ] i was not seen in control cells. In the experiment shown in Fig. 9B, the endogenous Mn 2ϩ influx stimulated by either CCh or TG was blocked by 10 M Gd 3ϩ . Addition of CCh to Trp3 cells pretreated with TG caused an increase in the rate of Mn 2ϩ entry, which was not seen in control cells (right panel). The rate of Mn 2ϩ entry under these conditions (TG followed by CCh) was not significantly different from the rate of entry observed in Trp3 cells treated only with CCh (left panel). In similar exper-iments, Ba 2ϩ or Sr 2ϩ was added to the TG-treated cells. Stimulation by CCh in the presence of extracellular Ba 2ϩ or Sr 2ϩ caused additional influx of these cations in Trp3 cells (Fig. 9, C and D) but not in control cells. These results suggest that in HEK cells expressing hTrp3, not only does receptor stimulation activate the influx of divalent cations better than store depletion, but signals from store depletion do not prevent or occlude the Trp3-mediated influx activated via receptor stimulation. Most of the influx activity that resulted from overexpression of hTrp3 alone in the HEK293 cells is thus activated by a mechanism other than store depletion.
The signaling cascade initiated by an agonist includes the activation of its receptor, of a G protein, of PLC, and thus the production of IP 3 and the release of Ca 2ϩ from its internal stores. Although store depletion does not appear to be the cause for agonist-activated Ca 2ϩ influx through the Trp3-mediated pathway, activation of PLC seems to be necessary. When 15 M U73122, an inhibitor of PLC (44), was used, not only Ca 2ϩ mobilization (not shown) but also Ca 2ϩ and Ba 2ϩ influx stimulated by CCh was blocked (Fig. 10). Basal Ca 2ϩ influx in the Trp3 cells was little affected by the treatment of U73122. Therefore, either the production of IP 3 or the active form of the PLC is required for the activation of the agonist-stimulated Trp3-mediated influx pathway.

DISCUSSION
Calcium mobilization from internal stores and subsequent Ca 2ϩ entry from extracellular space are the two major components of Ca 2ϩ signaling following activation of PLC by cell surface receptors. It has long been established that Ca 2ϩ is released from its internal stores via IP 3 receptors in response to the increase of IP 3 production (1). Only recently has evidence accumulated showing that Ca 2ϩ influx can occur via plasma membrane channels formed of Trp homologs (26,29,32). The two Drosophila photoreceptor Trp proteins have different functional features. DTrp was found to form a Ca 2ϩ influx channel activated either by IP 3 or by store depletion with TG, when expressed in insect Sf9 cells (22), Xenopus oocytes (23, 28), or 293T cells (24). On the other hand, DTrpl forms a non-selective cation channel that is activated by stimulation of G q -coupled receptors but is not sensitive to store depletion (20,21,24). In addition, DTrpl also displayed significant basal activity (20). By comparing the amino acid sequences of the mammalian Trp proteins with that of Drosophila Trp or Trpl, it is difficult to predict whether any of the mammalian Trps would functionally resemble DTrp or DTrpl. A recent report by Philipp et al. (29) showed that bovine Trp4 transiently expressed in HEK293 cells forms a channel that is relatively more selective for Ca 2ϩ than for monovalent cations and is activated to the same extent by either IP 3 or TG, suggesting that Trp4 may functionally resemble the DTrp. Zitt et al. (32) also reported that hTrp1 expressed in Chinese hamster ovary cells can be activated by store depletion treatment, although the channel formed in this case is non-selective. Thus, hTrp1 seems to resemble partly both DTrp and DTrpl. Experiments to be reported elsewhere 3 indicate that the stable expression of hTrp3 in the HEK293 cells gives rise to a novel Ca 2ϩ -permeant cation influx current that is spontaneously active, non-selective, and can be further stimulated by activation of G q -coupled receptors. In the present experiments, the majority of the increased Ca 2ϩ influx due to hTrp3 is neither activated by store depletion alone nor occluded by store depletion with TG. Thus, hTrp3 functionally resembles Drosophila Trpl.
Ca 2ϩ influx mediated by channels insensitive to store depletion have been shown to coexist with the store depletion-activated pathway in many systems. Orrenius and colleagues (42,46,47) reported that in cells such as hepatocytes, T lymphocytes, adrenal glomerulosa cells, fibroblasts, platelets, and anterior pituitary cells, Ca 2ϩ influx activated by emptying the agonist-sensitive stores could not completely mimic the influx activated by receptor agonists. More importantly, the storeinsensitive cation influx in hepatocytes was less selective for Ca 2ϩ than influx activated by TG. Clementi et al. (48) also reported that stimulation of PC12 cells with carbachol activated two Ca 2ϩ influx responses, of which one was store-dependent and the other was directly dependent on receptor activation. Work by Montero et al. (49) showed that in differ-  Fig. 9, A and C. Note that the store depletion-insensitive influx activated by CCh as seen in Fig. 9, A and C, is blocked by the PLC inhibitor. entiated HL60 cells, N-formyl-leucyl-phenylalanine activated an additional Ca 2ϩ influx after the internal Ca 2ϩ store had been completely emptied by TG. In electrophysiological studies, multiple types of Ca 2ϩ -conducting currents bearing distinct characteristics are often observed by investigators using carefully designed protocols. In mast cells, Ca 2ϩ enters through both a Ca 2ϩ release-activated current (I CRAC ) and a 50-picosiemen channel that is activated more directly by receptor/G protein coupling (16,50). In A431 epithelial cells, up to four different types of Ca 2ϩ -permeant channels could be involved in Ca 2ϩ influx following the activation of G protein (shown by the action of GTP␥S) or perfusion of IP 3 (14). In addition, IP 3 was found to activate plasma membrane channels in B lymphocytes (51) and substance P receptor-transfected Chinese hamster ovary cells (52). Moreover, IP 3 may activate the same channel that is sensitive to store depletion. In inside-out patches of vascular endothelial cells, IP 3 was shown to modulate the activity of a Ca 2ϩ influx channel which is indistinguishable from that activated by treatment with 2Ј,5Ј-di(tert-butyl)1,4-benzohydroquinone, i.e. by store depletion (17). Thus, under physiological conditions, it is likely that after receptor activation, Ca 2ϩ enters through multiple Ca 2ϩ -permeant channels. The relative contribution of each influx pathway is likely to differ depending on the cell type and probably the type and the degree of the stimulation. An estimate has been made for rat mast cells in which, under physiological conditions, the amount of Ca 2ϩ conducted by I CRAC is three times of that carried by the 50-picosiemen channel (16,53).
The activation mechanism of agonist-stimulated Ca 2ϩ influx via hTrp3 is unclear. Controversy exists with respect to the mechanism of activation of the Drosophila Trpl expressed in the Sf9 cells. Dong et al. (54) reported that DTrpl is activated by intracellular perfusion with IP 3 , whereas Obukhov et al. (55) found that in excised patches, DTrpl is activated by the active forms of the ␣ subunits of the G protein G 11 and G q , but not by IP 3 . The fact that agonist-stimulated store-insensitive Ca 2ϩ influx via hTrp3 can be prevented by an inhibitor of PLC suggests that the channel formed by hTrp3 is activated by a factor downstream of PLC activation, and upstream of store depletion. Since our experiments were carried out under conditions in which store depletion occurred either before or upon the addition of the agonist, we cannot rule out the possibility that Ca 2ϩ mobilization is also a necessary step for agonistinduced Ca 2ϩ influx through Trp3. If that is the case, stimulation of hTrp3 would require both store depletion and the generation of this other factor downstream of PLC activation. The likely candidates for this factor are IP 3 and its derivatives, for instance IP 4 . According to a so-called conformational coupling model (56), it is also possible that the channel is activated through direct interaction between the channel and the activated IP 3 receptor. Less likely, although not impossible, is that the channels could be stimulated by the activated PLC itself. A more recent study by Zitt et al. (45) showed that hTrp3 may be activated by Ca 2ϩ . However, treatment with TG also induces an increase in [Ca 2ϩ ] i . This increase does not seem to activate Ca 2ϩ influx mediated by Trp3 as well as that stimulated by receptor activation, suggesting that factors more than just intracellular Ca 2ϩ are involved in activating Trp3. The 50-picosiemen channel found in mast cells is not activated by IP 3 but instead requires the activation of receptor/G protein system (16). Thus, whether hTrp3 expressed in the HEK293 cells forms a channel that resembles any of the Ca 2ϩ influx channels found in native tissues or isolated primary cells remains to be elucidated.
Because of some homology between the last four putative transmembrane segments of Trp and those of the subdomains of voltage-gated Ca 2ϩ and Na ϩ channels, it has been proposed that a Trp-based channel may be a tetramer (19). Based on our finding that the mammalian genome contains at least six Trp genes, we speculated that a channel assembled by Trps can be either homotetrameric or heterotetrameric and that this could be a mechanism to create functionally diverse Ca 2ϩ -permeant channels, including store depletion-activated, store depletioninsensitive, and channels activated by both store depletion and independently by agonists (31). In the stable HEK293 cell lines, a homotetrameric hTrp3 may have become the predominant influx channel formed because of overexpression of this protein. However, it may not represent any of the native Ca 2ϩ influx channels present in this or other cell types. Because the cDNA for hTrp3 was isolated from HEK293 cells (26), we believe that there is endogenous hTrp3 protein in these cells. However, we do not know whether channels formed by hTrp3 alone are present in the native HEK293 cells because a component of Ca 2ϩ influx resembling that expressed in the Trp3 cell lines is either missing or too small to be detected in the control cells. One possibility is that the endogenous hTrp3 in the HEK293 cells plays a very minor role in Ca 2ϩ influx in response to agonist stimulation. The other possibility is that hTrp3 and other endogenous Trp proteins form heterotetrameric channels that are responsible for agonist and store depletion-activated Ca 2ϩ entry in these cells. Evidence for the formation of a heteromultimeric Trp-based Ca 2ϩ influx channel was recently obtained by Gillo et al. (23) who observed that coexpression of Drosophila Trp and Trpl in Xenopus oocytes leads to the appearance of a channel with an ion selectivity and La 3ϩ sensitivity different from those seen in oocytes expressing either protein alone. Interestingly, the new channel is activated to the same extent either by IP 3 or by TG, even though one of its components, DTrpl, is not sensitive to store depletion. More recently, Montell and colleagues (24) demonstrated that the N termini of DTrp and DTrpl interact with each other both in vitro and in vivo. Coexpression of the two proteins in 293T cells gave rise to a store depletion-sensitive cation influx channel that had features from both DTrp and DTrpl. Interestingly, cells expressing both DTrp and DTrpl did not have an increased basal inward current as found in cells expressing DTrpl alone. The authors proposed that DTrpl in Drosophila photoreceptors does not form homomultimers by itself, and its spontaneous activity is prevented by forming heteromultimeric channels with DTrp or other Trp-related proteins. A similar conclusion may be drawn for hTrp3 since it behaves very similarly to DTrpl when expressed alone. Because Drosophila Trp proteins can form store-operated heteromultimeric channels even if only one subunit is capable of detecting the signal from store depletion, it is possible that the exogenously expressed hTrp3 in the stably transfected cell lines also forms heteromultimers in combination with the endogenous Trp proteins, of which some can be activated by store depletion. This would explain the small but significant Gd 3ϩ -resistant TG-stimulated Ca 2ϩ influx in Trp3 cells (Fig. 6C). The smaller TG-stimulated increase of Ca 2ϩ influx than the CCh-stimulated increase found in transiently transfected COS cells (26) can also be explained this way. On the other hand, the contribution of hTrp3-containing heteromultimeric channels to TG-stimulated Ca 2ϩ influx could be more significant than that being revealed by the protocol shown in Fig. 6 if these channels are more sensitive to Gd 3ϩ than the homomultimeric channel formed by hTrp3. This is even more so when the new channels formed cannot be distinguished from the native Ca 2ϩ influx channels present in the HEK293 cells by Ca 2ϩ channel blockers. Therefore, although our results show that overexpression of hTrp3 in HEK293 cells gives rise to a cation entry pathway insensitive to regulation by a store-operated manner, we cannot at this point rule out the possibility that Trp3 is involved in forming native store-operated channels in HEK293 and other cells. Other Trp-unrelated proteins may also participate in the formation of the Ca 2ϩ influx channels. It is yet to be firmly established that Trps are the true or the only subunits of storeor agonist-stimulated Ca 2ϩ entry channels, as this will require purification of the channel complexes followed by reconstitution into either vesicles or planar lipid bilayers and analysis of channel activity. Further, even if they were channel subunits, other auxiliary proteins are expected to be required for the formation the native channel(s).
In conclusion, Ca 2ϩ influx following the activation of the PLC/IP 3 signaling pathway is an important phenomenon. Channels with different properties have been described in various cell types (for review, see Refs. 3 and 53). These include the store-operated channels that differ both in conductance and ion selectivity. These also include channels that do not appear to be store-operated. The presence of six Trp homologs in mammals and the possibility that they form heteromultimeric channels provide a plausible explanation for the tissue-and cell-specific behavior of Ca 2ϩ influx pathways (and channels) found in different systems.