The COOH-terminal domain of Drosophila TRP channels confers thapsigargin sensitivity.

Previous studies have shown that the Drosophila cation channels designated Trp and Trpl can be functionally expressed in Sf9 insect cells using baculovirus expression vectors. The trp gene encodes a Ca-permeable channel that is activated by thapsigargin, blocked by low micromolar Gd, and is relatively selective for Caversus Na and Ba. In contrast, trpl encodes a Ca-permeable cation channel that is constitutively active, not affected by thapsigargin, blocked by high micromolar Gd, and non-selective with respect to Ca, Na, and Ba. The region of lowest sequence identity between Trp and Trpl occurs in the COOH-terminal domain. To test the hypothesis that this region is responsible for the differential sensitivity of these channels to thapsigargin, chimeric constructs of Trp and Trpl were created in which the COOH-terminal tail region of each protein was exchanged. The Trp construct with the Trpl COOH-tail was constitutively active, insensitive to thapsigargin, but retained selectivity for Ca over Na and Ba. In contrast, the Trpl construct with the Trp COOH-tail was not constitutively active, could be activated by thapsigargin, but remained non-selective with respect to Ca, Ba, and Na. These results suggest that the COOH-terminal domain of Trpl plays an important role in determining constitutive activity, whereas the COOH-terminal region of Trp contains the structural features necessary for activation by thapsigargin.

Previous studies have shown that the Drosophila cation channels designated Trp and Trpl can be functionally expressed in Sf9 insect cells using baculovirus expression vectors. The trp gene encodes a Ca 2؉permeable channel that is activated by thapsigargin, blocked by low micromolar Gd 3؉ , and is relatively selective for Ca 2؉ versus Na ؉ and Ba 2؉ . In contrast, trpl encodes a Ca 2؉ -permeable cation channel that is constitutively active, not affected by thapsigargin, blocked by high micromolar Gd 3؉ , and non-selective with respect to Ca 2؉ , Na ؉ , and Ba 2؉ . The region of lowest sequence identity between Trp and Trpl occurs in the COOH-terminal domain. To test the hypothesis that this region is responsible for the differential sensitivity of these channels to thapsigargin, chimeric constructs of Trp and Trpl were created in which the COOH-terminal tail region of each protein was exchanged. The Trp construct with the Trpl COOH-tail was constitutively active, insensitive to thapsigargin, but retained selectivity for Ca 2؉ over Na ؉ and Ba 2؉ . In contrast, the Trpl construct with the Trp COOH-tail was not constitutively active, could be activated by thapsigargin, but remained nonselective with respect to Ca 2؉ , Ba 2؉ , and Na ؉ . These results suggest that the COOH-terminal domain of Trpl plays an important role in determining constitutive activity, whereas the COOH-terminal region of Trp contains the structural features necessary for activation by thapsigargin.
The Drosophila proteins encoded by the trp and trpl genes (Trp and Trpl) are thought to form Ca 2ϩ -permeable cation channels responsible for depolarization of photoreceptor cells following stimulation by light (1,2). Since phototransduction in Drosophila requires phospholipase C activity (for review, see Refs. [2][3][4], Trp and Trpl may be insect homologues of the channels responsible for I crac , the elusive store-operated, capacitative Ca 2ϩ entry channels that are activated by receptormediated phosphoinositide hydrolysis in mammalian non-excitable cells. The first studies demonstrating that trp and trpl encode ion channels were performed in a heterologous expression system using baculovirus expression vectors (5)(6)(7)(8). Eval-uation of whole cell membrane currents in Sf9 insect cells following infection with recombinant baculovirus containing the cDNA for trp and trpl under control of the polyhedrin promoter demonstrated that trp encodes a cation channel that is selective for Ca 2ϩ over Na ϩ , can be activated by depletion of the internal Ca 2ϩ store by thapsigargin, and is a poor conductor of Ba 2ϩ (8). In contrast, trpl encodes a cation channel that is constitutively active, is relatively non-selective with respect to Ca 2ϩ , Ba 2ϩ , and Na ϩ , is unaffected by thapsigargin (7,8), but can be activated by a receptor-mediated increase in inositol-1,4,5-trisphosphate (6,9).
The functional studies on Trp and Trpl allow for specific predictions concerning the structural features of Trp that may be responsible for the differential sensitivity of these two channel proteins to thapsigargin. Trp and Trpl exhibit substantial amino acid identity in their NH 2 -terminal regions and in their proposed membrane spanning segments but differ in their COOH-terminal domains (1,10,11). We hypothesized that the COOH-terminal domain of Trp is necessary for thapsigargin sensitivity. To test this hypothesis, chimeric proteins were created in which the COOH-terminal domains of Trp and Trpl were exchanged, and the resulting channels were functionally expressed using the baculovirus-Sf9 insect cell expression system. The results demonstrate that the COOH domain of Trp confers thapsigargin sensitivity to Trpl and that the COOH domain of Trpl confers constitutive activity to Trp. Although the relative selectivity of the chimeric channels for Ca 2ϩ , Na ϩ , and Ba 2ϩ appears to be unchanged from the native channels, the sensitivity of the chimeric channels to Gd 3ϩ blockade is intermediate between that seen for Trp and Trpl. Thus, the COOH-terminal domain may influence pore characteristics. A preliminary report of this work has appeared in abstract form (12).
Cell Culture-Sf9 cells were obtained from Invitrogen and were cultured as described previously (5,6,13) using Grace's insect medium supplemented with lactalbumin hydrolysate, yeastolate, L-glutamine, 10% heat-inactivated fetal bovine serum, and 1% penicillin-streptomycin solution (Life Technologies, Inc.). Cells were grown either in spinner flasks (Bellco Glass, Vineland, NJ) or in 35-or 100-mm plastic tissue culture dishes. The cell cultures were incubated at 27°C in a humidified air atmosphere.
Construction of Chimeric cDNAs-The cDNA for trp and trpl was subcloned using standard techniques (14) into the baculovirus transfer vector pVL1393 (Invitrogen) yielding pVL-trp and pVL-trpl. The trp and trpl sequences have in common a unique cleavage site for the restriction endonuclease DraIII, 14 amino acids downstream from the final (S6) putative transmembrane domain (Fig. 1), and the pVL vector contains a cleavage site for NotI in the downstream multiple cloning site. Both pVL-trp and pVL-trpl were digested with DraIII and NotI, and the resulting fragments were purified by agarose gel electrophoresis. The COOH-tail fragment of trpl was ligated with the pVL-trp fragment to produce the trp/trpl chimera; the trpl/trp chimera was produced in an analogous fashion. The nucleotide sequence for each construct was confirmed by the dideoxynucleotide method using Sequenase version 2.0 (U. S. Biochemical Corp.).
To monitor expression at the protein level, the nucleotide sequence encoding the FLAG epitope (DYKDDDDK) was attached to the NH 2 terminus of each construct (pVL-trp, pVL-trpl, pVL-trp/trpl, and pVLtrpl/trp) using the following general procedure. Oligonucleotides (Ransom Hill Bioscience, Ramona, CA) were synthesized consisting of a methionine start codon, the FLAG sequence, and several base pairs encoding the NH 2 terminus of either trp or trpl. A second set of oligonucleotides was synthesized consisting of sequence within the coding region of trp and trpl that included a unique restriction site. These oligonucleotides were used as primer sets for amplification by polymerase chain reaction of the FLAG-attached NH 2 terminus of trp and trpl. The polymerase chain reaction products were subcloned into plasmid pCRII (Invitrogen) and subsequently transferred to the pVL-trp and -trpl constructs using convenient restriction sites. All constructs were sequenced to confirm that the nucleotides encoding the FLAG epitope were attached and that the trp and trpl coding sequence remained in frame with the new start codon.
Generation of Recombinant Baculovirus-Recombinant baculoviruses were produced using the BaculoGold TM transfection kit (Phar Mingen, San Diego, CA) by cotransfecting Sf9 cells with pVL-trp, -trpl, and -chimeric constructs and linearized BaculoGold viral DNA as described in the instructions provided with the kit. Single viral plaques were isolated and amplified two to four times to obtain a high titer viral stock solution, which was stored at 4°C until use.
Infection of Sf9 Insect Cells with Recombinant Baculovirus-For routine infection, Sf9 cells in Grace's medium were allowed to attach to the bottom of plastic culture dishes (ϳ10 5 cell/cm 2 ). Following incubation for 15 min, an aliquot of viral stock was added (multiplicity of infection was 10), and the cells were maintained at 27°C in a humidified air atmosphere. Unless otherwise indicated, cells were used at 48 h postinfection.
Isolation of Membrane-associated FLAG Proteins-Cells were harvested at 48 h postinfection time, subjected to centrifugation at 500 ϫ g for 5 min, and resuspended at a density of 5 ϫ 10 6 cells/ml in lysis buffer containing 20 mM Tris-Cl, 5 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 0.5 M aprotinin, and 2 M leupeptin. The cell suspension was sonicated on ice using a Sonic Dismembranator (Fisher) on a power setting of 2.5. The cell suspension was sonicated three times for 10 s with a 10-s rest between pulses. Lysis was monitored by light microscopy. The cell lysate was subjected to centrifugation at 6000 ϫ g for 10 min at 4°C. The resulting pellets were discarded, and the supernatants were centrifuged at 42,000 ϫ g for 30 min. The microsomal pellets were resuspended in lysis buffer at a protein concentration of 5-10 mg/ml and stored at Ϫ80°C until use. Protein was determined by the method of Lowry (15) using bovine serum albumin as the standard. An excess volume of sample buffer (60 mM Tris-Cl, pH 6.8, 2% SDS, 10% glycerol, 100 mM dithiothreitol, and 0.025% bromphenol blue) was added to an aliquot of membrane preparation, and the reaction was incubated at 100°C for 1.5 min. Samples were subjected to electrophoresis on 8% polyacrylamide gels (16) along with biotinylated molecular weight standards (Bio-Rad).
Immunoblotting of the FLAG Proteins-Proteins were transferred electrophoretically from the polyacrylamide gels to Millipore polyvinylidene difluoride membranes (100 -150 mA for 12 h on ice). The membranes were blocked with dry milk for 1 h and probed with mouse M2 anti-FLAG antibody (VWR, Bridgeport, NJ) for 2 h at room temperature. After washing, the membranes were probed for 1 h with goat anti-mouse IgG conjugated with horseradish peroxidase and with avidin-conjugated horseradish peroxidase for reaction with biotinylated molecular weight standards. Protein bands were visualized by addition of horseradish peroxidase substrates in peroxidase buffer (Pierce).
Measurement of Free Cytosolic Ca 2ϩ Concentration-[Ca 2ϩ ] i was measured in Sf9 cells using the fluorescent indicator, fura-2, as described previously (5, 6). Briefly, cells were dispersed, washed, and resuspended at a concentration of 1.5-2 ϫ 10 6 cell/ml in MBS contain-ing 2 M fura-2/AM. Following a 30-min incubation at room temperature (22°C), the cell suspension was subjected to centrifugation, resuspended in an equal volume of MBS, and incubated for an additional 30 min. The cells were washed, and fluorescence was measured using an SLM 8100 spectrophotofluorimeter. Excitation wavelength alternated between 340 and 380 nm, and fluorescence intensity was monitored at an emission wavelength of 510 nm. All measurements on Sf9 cells were performed at room temperature (22°C). Calibration of the fura-2 associated with the cells was accomplished using Triton lysis in the presence of a saturating concentration of Ca 2ϩ followed by addition of EGTA (pH 8.5). [Ca 2ϩ ] i was calculated by the equation of Grynkiewicz et al. (17) using a K d value for Ca 2ϩ binding to fura-2 of 278 nM for 22°C (18). Statistical differences were determined by Student's t-test using Bonferroni's correction for multiple comparisons (19).
Electrophysiological Techniques-The patch clamp technique for whole cell recording was utilized (8,20). Unless otherwise indicated, the bath (extracellular) and pipette (intracellular) solution contained 100 mM sodium gluconate, 10 mM HEPES, 2 mM EGTA, pH 6.5 (free Ca 2ϩ concentration (determined using fura-2 fluorescence) was ϳ150 nM). In some experiments (where indicated) the bath solution was changed to one containing either 50 mM calcium gluconate or barium gluconate, 10 mM HEPES, pH 6.5. Extracellular solutions were rapidly changed using a perfusion system adapted from one previously described (21). The osmolarity of all solutions was adjusted to 340 mosM with mannitol. The patch pipette resistance was ϳ5 megaohm, and the average whole cell capacitance was ϳ10 picofarads. Data were obtained using an Axopatch 200A amplifier (Pacer Scientific, Los Angeles, CA) and sampled on line using pClamp 5.5 software. All recordings were made at room temperature (22°C). To generate current-voltage (I-V) relations, voltage pulses were applied from Ϫ100 to ϩ60 mV in 20-mV increments with a 400-ms duration during each voltage step and a 2-s interval between steps. The holding potential between steps was 0 mV. Where indicated, n equals the number of cells examined under each condition.

RESULTS
Previous studies have suggested that Trp and Trpl differ in their sensitivity to thapsigargin (8). Comparison of the primary structure of the two proteins demonstrates that the region of greatest difference between Trp and Trpl occurs in the COOHterminal domain (Fig. 1A). To test the hypothesis that this region is responsible for the differential sensitivity of these channels to thapsigargin, chimeric constructs were created in which the COOH-terminal domains of the two proteins were exchanged, and the resulting proteins were expressed using the baculovirus expression system. The predicted molecular mass for Trp and Trpl is 142.1 and 127.6 kDa, respectively (1, 10, 11). Recombinant Trp and Trpl expressed in Sf9 insect cells migrate on SDS-polyacrylamide gel electrophoresis as 152-and 127-kDa proteins, respectively (Fig. 1B, lanes d and e). Since a major difference between Trp and Trpl is the length of the COOH-terminal domain (Fig. 1A), the chimeras should have a molecular weight determined by the COOH-domain present in each construct. As seen in Fig. 1B, the Trpl/Trp and the Trp/ Trpl chimeras have apparent molecular masses of 150 and 134 kDa, respectively (lanes f and g). The predicted molecular masses based on the cDNAs are 142.3 and 128.0 kDa for the Trpl/Trp and Trp/Trpl chimeras, respectively. Thus, all constructs that contain Trp sequences migrate with an apparent molecular weight that is greater than that predicted by the DNA sequence. The reason for this discrepancy is unknown. It seems unlikely to be related to expression in Sf9 cells since Trp expressed in Xenopus oocytes also migrates as a 150-kDa protein (22). Specific protein bands were not observed in either non-infected Sf9 cells or in cells infected with recombinant baculovirus containing the cDNA for the human bradykinin receptor (Fig. 1B, lanes b and c). Likewise, no specific bands were observed on Western blots of membrane proteins isolated from Sf9 cells infected with baculovirus containing cDNA for native (non-FLAGed) proteins (data not shown). Estimates of relative protein expression level determined from band intensities on a number of gels were Trpl Ͼ Trp Х Trp/Trpl Ͼ Trpl/Trp.

Effect of Protein Expression on Basal [Ca 2ϩ
] i -Previous studies have shown that Trpl is constitutively active when expressed in Sf9 insect cells. This activity gives rise to an increase in basal [Ca 2ϩ ] i as a function of postinfection time (7). As seen in Fig. 1C, basal [Ca 2ϩ ] i was significantly increased 225% at 48 h postinfection in Trp/Trpl-expressing cells relative to that in cells expressing native Trp (p Ͻ 0.00001). Likewise, basal [Ca 2ϩ ] i was significantly decreased 54% in cells expressing Trpl/Trp chimeric protein relative to that observed in cells expressing native Trpl (p Ͻ 0.00001). These results suggest a correlation between the presence of the Trpl COOH-terminal domain and constitutive activity.
Expression of Channel Activity-Trpl-expressing cells exhibit substantial membrane current under basal non-stimulated conditions (7), which is unaffected by thapsigargin (8). In contrast, membrane currents observed in Trp-expressing cells are small under basal conditions but increase dramatically following application of thapsigargin (8). Whole cell currents were recorded in symmetrical sodium gluconate solutions in cells expressing either Trp/Trpl or Trpl/Trp chimeric proteins (Fig. 2). Inward and outward membrane currents were observed in Trp/Trpl-expressing cells under basal, non-stimulated conditions (n ϭ 16). The current-voltage relationship was linear with a reversal potential near 0 mV. In 13 out of 16 cells, addition of thapsigargin (200 nM) to Trp/Trpl expressing cells had no effect on inward or outward currents. In three cells, current increased ϳ5% after thapsigargin addition. In sharp contrast, currents recorded in Trpl/Trp cells were small under basal conditions but increased dramatically following the addition of thapsigargin to the bath (14 out of 17 cells). Thapsigargin-induced currents were linear and reversed close to 0 mV when recorded in symmetrical sodium gluconate solutions. In 3 of 17 cells expressing Trpl/Trp, current was seen within 1 min after establishment of the whole cell recording mode and did  (lanes a and h). With the exception of Trpl, which was isolated at 30 h, all membrane preparations were isolated from infected cells at 48 h postinfection. The protein bands seen in lanes b and c arise from nonspecific binding of the avidin-conjugated horseradish peroxidase used to visualize the biotinylated molecular weight standards. Panel C, effect of protein expression on basal, non-stimulated [Ca 2ϩ ] i . Basal [Ca 2ϩ ] i in Sf9 cells expressing the indicated protein was estimated using fura-2 fluorescence as described under "Experimental Procedures." All cells were examined at 48 h postinfection, and the values represent the mean Ϯ S.E. (n ϭ 6 -10). Note that the FLAG-Trpl/Trp construct was used for these experiments; all others were native non-FLAGed proteins. not increase after thapsigargin.
Chimeric Channel Ion Selectivity-Trp channels are relatively selective for Ca 2ϩ versus Na ϩ and Ba 2ϩ . In contrast, Trpl channels are non-selective with respect to Ca 2ϩ , Na ϩ , and Ba 2ϩ (8). To determine the ionic selectivity for the chimeric channels, currents were first recorded in symmetrical sodium gluconate solutions and the bath solution was subsequently changed to one containing either 50 mM calcium gluconate or barium gluconate (Fig. 3). The I-V relationships for basal currents in the Trp/Trpl chimera were linear in symmetrical Na ϩ solutions with reversal potentials near 0 mV. Changing the bath to Ca 2ϩ decreased inward current and produced a positive shift in reversal potential to approximately 35 mV; the current was inwardly rectifying with sodium gluconate in the pipette and calcium gluconate in the bath. Changing the bath solution to one containing Ba 2ϩ decreased both inward and outward current without affecting the reversal potential. The thapsigargininduced membrane currents in the Trpl/Trp chimera were also linear over a wide range of membrane potentials when recorded in symmetrical Na ϩ solutions. Changing the bath to one containing Ca 2ϩ had little effect on the currents in the Trpl/Trp chimera-expressing cells. Inward current was reduced with Ba 2ϩ bath solution, but the reversal potential remained near 0 mV. These results suggest that the relative permeability of the chimeras for Na ϩ , Ca 2ϩ , and Ba 2ϩ is not affected by switching the COOH-terminal domains.
Effect of Gd 3ϩ on Chimeric Channel Currents-The effect of Gd 3ϩ was determined on basal, non-stimulated currents in Trpl and Trp/Trpl-expressing cells and on thapsigargin-induced currents in Trp and Trpl/Trp-expressing cells (Fig. 4). Trp and Trpl inward currents exhibit profound differences in sensitivity to blockade by Gd 3ϩ ; inward Na ϩ currents in Trpexpressing cells are almost completely blocked by 10 M Gd 3ϩ added to the extracellular bath, whereas currents in Trpl cells required 1 mM Gd 3ϩ to achieve the same level of inhibition. The sensitivity of the Trp/Trpl and Trpl/Trp chimera cells is intermediate, with 10 M Gd 3ϩ producing 50% inhibition of inward Na ϩ currents. Thus, exchanging the COOH-terminal domains of Trp and Trpl affects sensitivity of both channels to Gd 3ϩ blockade.

DISCUSSION
Previous studies on the functional expression of Trp in Sf9 insect cells suggest that although Trp appears to be a Ca 2ϩ store-operated channel that can be activated by thapsigargin, the ionic selectivity is different from both the endogenous I crac recorded in the Sf9 insect cells (8) and I crac observed in several mammalian cell types (23)(24)(25)(26); I crac appears to be highly selective for Ca 2ϩ over Na ϩ and to be inwardly rectifying even in the presence of symmetrical Na ϩ solutions (25). Thus, Trp may not be the insect homologue of the I crac channel. Trp does, however, appear to be a member of a large protein family found in Drosophila, Calliphora, Xenopus, mouse (22), and human (27,28), although there is no functional information on the Trp homologues from sources other than insect. Petersen et al. (22) recently reported that expression of Drosophila Trp in Xenopus oocytes following injection of cRNA gives rise to an enhanced thapsigargin-induced increase in Ca 2ϩ influx estimated by the magnitude of the Ca 2ϩ -activated Cl Ϫ current, consistent with the activation of Trp by depletion of the internal Ca 2ϩ store. Thus, it seems likely that although Trp may not be identical to I crac channels, it may be regulated in a fashion similar to I crac . Understanding the mechanism by which Trp is regulated by thapsigargin and identification of the the structural domain of Trp necessary for this regulation could provide important clues to structure and function of mammalian I crac channels. Toward this end, the purpose of the present study was to determine the general region of Trp that is necessary or sufficient for regulation of channel activity by thapsigargin. To accomplish this goal, we exploited the differences in sensitivity of Trp and Trpl to thapsigargin. With exception of amino acid residues 330 -500, Trp and Trpl are very similar over the first two-thirds of the predicted amino acid sequence. We therefore focused our attention on the COOH-terminal domain as the region that gives rise to the differential sensitivity of these two channel proteins to thapsigargin.
Exchanging the COOH-terminal domains of Trp and Trpl produced several important functional changes. First, and most importantly, thapsigargin sensitivity was conveyed to Trpl by the COOH domain of Trp, and the presence of the Trpl COOH-terminal domain on Trp eliminated thapsigargin sensi- tivity and increased constitutive activity. These results provide strong evidence that the region of Trp necessary for thapsigargin sensitivity resides in the COOH-terminal domain. This altered sensitivity of the chimeras to thapsigargin occurred without a change in the relative permeability of the channels for Ca 2ϩ , Na ϩ , or Ba 2ϩ , suggesting that pore characteristics are predominately determined by the first two-thirds of the protein structure and probably by the central hydrophobic core, which contains the proposed transmembrane segments S1-S6. However, the Trp/Trpl channel had a lower sensitivity to Gd 3ϩ compared to native Trp channels, and the Trpl/Trp chimera had a higher sensitivity to Gd 3ϩ compared to native Trpl channels. Thus, it appears that the COOH-terminal domains of Trp and Trpl may have subtle influences on pore characteristics. One possibility is that the COOH-terminal domain of Trp forms an extension of the pore structure into the cytoplasm beyond the inner leaflet of the phospholipid membrane in a fashion analogous to the nicotinic acetylcholine receptor, where a large extracellular domain forms the vestibule of the channel, which extends out from the membrane structure. Presumably this region would contain the sequence responsible for binding Gd 3ϩ . Alternatively, the COOH-terminal domain may influence the conformation of the pore region producing subtle alterations in the Gd 3ϩ binding site. A similar phenomenon has been observed in the inwardly rectifying K ϩ channel, where the COOH-terminal domain appears to have a major role in specifying pore properties (29).
The mechanism by which the COOH-terminal domain affects constitutive activity is unknown. This does not reflect variation in protein expression, since a difference in basal [Ca 2ϩ ] i is seen between Trp and the Trp/Trpl chimera, yet both are expressed to approximately the same levels in the Sf9 cell. Furthermore, the constitutive activity (or lack thereof) of the chimeras is clear from the whole cell current recordings; the COOH-terminal domain of Trp maintains the channel in a non-conducting state. In a fashion analogous to other channel types, part of COOH-terminal domain of Trp may act as the "gate," which is closed in the absence of thapsigargin but opens in response to depletion of the internal Ca 2ϩ store.
The present structure-function study may also provide insight into the mechanism by which store depletion activates surface membrane channels like I crac . In this regard, there are basically two hypotheses. The conformational coupling hypothesis suggests that close association between the endoplasmic reticulum and the plasmalemma allows for direct physical coupling between the inositol 1,4,5-trisphosphate receptor and Ca 2ϩ entry channels and that information concerning the repletion status of the internal store is related to surface membrane channels via a conformational change (30). Alternatively, a soluble, diffusible messenger may be generated and/or released upon depletion of the Ca 2ϩ store, which then either  Fig. 2. Currents were first determined in each cell with 100 mM sodium gluconate in both the bath and pipette solution. EGTA was omitted from the bath solution for this set of experiments. Currents were recorded under basal, non-stimulated conditions for Trpl (Ⅺ) and Trp/Trpl (E), and in the presence of 200 nM thapsigargin for Trp (F) and Trpl/Trp (f) expressing cells. The bath solution was then changed for one containing 100 mM sodium gluconate plus the indicated concentration of Gd 3ϩ . All values represent the mean Ϯ S.E. (n ϭ 4 -5) current amplitudes taken at 200 ms after a step change in membrane potential from 0 to Ϫ100 mV. Currents in the presence of Gd 3ϩ were normalized to the control current amplitudes obtained in each cell before application of Gd 3ϩ . In all cells reported, currents returned to control levels following washout of Gd 3ϩ . directly or indirectly activates the surface membrane channels (31,32). The COOH-terminal domain of Trp contains a unique proline-rich region in which the dipeptide KP is repeated 27 times at relatively even intervals and includes a highly charged segment where the sequence DKDKKP(G/A)D is repeated 9 times. Interestingly, the bacterial protein TonB also has a proline-rich region in which a string of EP and KP repeats is thought to form a mechanical linkage between the inner and outer bacterial membranes (33,34). By analogy, the highly charged proline-rich segment of Trp may perform the same function linking proteins of the endoplasmic reticulum to the Trp channel. Alternatively, another region of the COOH-terminal domain of Trp may be responsive to depletion of the Ca 2ϩ store. In this regard, the region may contain the binding site for CIF, the putative Ca 2ϩ influx factor (31). Another possibility derives from studies suggesting the involvement of tyrosine kinase activity in Ca 2ϩ entry activated by depletion of the stores (35)(36)(37)(38). In this regard, Trp has 5 tyrosine residues in the COOH-terminal domain at positions 665, 687, 745, 756, and 922. The first three tyrosine residues are conserved in Trpl. It is possible that phosphorylation of Tyr-756 or Tyr-922 may play a role in activation or regulation of Trp by thapsigargin. Additional chimeric constructs in which different regions of the COOH-terminal domains of Trp and Trpl are exchanged may help determine the specific regions necessary for thapsigargin sensitivity, for constitutive activity, and for determination of pore characteristics.