Functional Properties of Endogenous Receptor- and Store-operated Calcium Influx Channels in HEK293 Cells*

Activation of phospholipase C (PLC)-mediated signaling pathways in non-excitable cells causes the release of calcium (Ca2+) from inositol 1,4,5-trisphosphate (InsP3)-sensitive intracellular Ca2+ stores and activation of Ca2+ influx via plasma membrane Ca2+ channels. The properties and molecular identity of plasma membrane Ca2+ influx channels in non-excitable cells is a focus of intense investigation. In the previous studies we used patch clamp electrophysiology to describe the properties of Ca2+ influx channels in human carcinoma A431 cell lines. Now we extend our studies to human embryonic kidney HEK293 cells. By using a combination of Ca2+ imaging and whole cell and single channel patch clamp recordings we discovered that: 1) HEK293 cells contain four types of plasma membrane Ca2+ influx channels: ICRAC, Imin, Imax, and INS; 2) ICRAC channels are highly Ca2+-selective (PCa/Cs > 1000) and ICRAC single channel conductance is too small for single channel analysis; 3) Imin channels in HEK293 cells display functional properties identical to Imin channels in A431 cells, with single channel conductance of 1.2 pS for divalent cations, 10 pS for monovalent cations, and divalent cation selectivity PBa/K = 20; 4) Imin channels in HEK293 cells are activated by InsP3 and inhibited by phosphatidylinositol 4,5-bisphosphate, but store-independent; 5) when compared with Imin, Imax channels have higher conductance for divalent (17 pS) and monovalent (33 pS) cations, but less selective for divalent cations (PBa/K = 4), 6) Imax channels in HEK293 cells can be activated by InsP3 or by Ca2+ store depletion; 7) INS channels are non-selective (PBa/K = 0.4) and display a single channel conductance of 5 pS; and 8) INS channels are not gated by InsP3 but activated by depletion of intracellular Ca2+ stores. Our findings provide novel information about endogenous Ca2+ channels supporting receptor-operated and store-operated Ca2+ influx pathways in HEK293 cells.

Activation of phospholipase C (PLC) 1 -mediated signaling pathways in non-excitable cells causes the release of calcium (Ca 2ϩ ) from inositol 1,4,5-trisphosphate (InsP 3 )-sensitive intracellular Ca 2ϩ stores and activation of Ca 2ϩ influx via plasma membrane Ca 2ϩ channels. Receptor-operated (ROC) and storeoperated (SOC) Ca 2ϩ influx pathways have been described in non-excitable cells (1)(2)(3). ROC Ca 2ϩ influx pathways are activated directly as a result of PLC-coupled receptor activation. Activation of SOC Ca 2ϩ influx pathways requires depletion of intracellular Ca 2ϩ stores. Functional properties and molecular identity of channels supporting ROC and SOC Ca 2ϩ influx pathways is a focus of intense investigation. Based on their selectivity for divalent cations, SOC currents can be separated into highly Ca 2ϩ -selective channels (P D/M Ͼ 1000) named "Ca 2ϩ release-activated channels" (I CRAC ) (4 -6) and moderately Ca 2ϩ -selective (P D/M ϳ 10) channels I SOC (1)(2)(3)7). The molecular identity of I CRAC remains unclear (1)(2)(3)8). Mammalian trp channels of the TRPC family are the most likely candidates for the role of I SOC channels (3, 7, 9 -12). Properties of ROC channels are much less understood due to the difficulty in separating direct effects of PLC activation from effects resulting from depletion of intracellular InsP 3 -sensitive Ca 2ϩ stores.
Most of the functional studies of ROC and SOC Ca 2ϩ influx pathways have been performed using Ca 2ϩ imaging or whole cell current recordings. We reasoned that single channel patch clamp recordings provide unique and precise information about functional properties of individual types of channels supporting Ca 2ϩ influx in non-excitable cells. In our previous experiments we used single channel patch clamp recordings to perform detailed characterization of channels supporting Ca 2ϩ influx in A431 cells (13)(14)(15)(16)(17)(18)(19). Now we extended our studies to HEK239 cell line, which is used extensively as heterologous expression host in functional studies of TRPC family members, such as TRPC1 (20), TRPC3 (21)(22)(23), TRPC4 (24), TRPC5 (20,23,25), TRPC6 (26,27), and TRPC7 (27). HEK293 cells express endogenous TRPC1, TRPC3, TRPC4, and TRPC6 proteins (28 -30), and the functional properties of endogenous Ca 2ϩ influx channels in HEK293 cells described here will be useful for interpretation of results obtained in heterologous expression studies of TRPC channels in HEK293 cells. Obtained results also provide interesting insights into similarities and differences between Ca 2ϩ influx pathways in A431 and HEK293 cells.

EXPERIMENTAL PROCEDURES
Cell Culture-HEK293 cells (Cell Culture Collection, Institute of Cytology, St. Petersburg, Russia) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM glutamine. Cells were grown in an incubator at 37°C with humidified 5% CO 2 and 95% air. For patch clamp and Ca 2ϩ imaging experiments cells were seeded onto coverslips and maintained in culture for 1-3 days before use.
Generation of Recombinant Adenoviruses-Ad-InsP 3 R-N adenovirus was generated using the AdEasy system according to published procedures (31). Briefly, an amino-terminal fragment of rat InsP 3 R1 (32) (M1-K604) was amplified by PCR (with addition of 5Ј-untranslated region Kozak sequence and carboxyl-terminal His 6 tag) and subcloned into KpnI/XhoI sites of pAdTrack-CMV shuttle vector. Resulting pAdTrack-CMV-InsP 3 R-N plasmid was linearized by PmeI and cotransformed into Escherichia coli BJ5183 cells together with pAdEasy-1 adenoviral backbone plasmid. Recombinants were selected by kanamycin resistance and confirmed by PCR analysis. Recombinant adenoviral plasmid pAd-InsP 3 R-N was isolated, linearized by PacI, and transfected into HEK293 cells for packaging. Transfected cells were monitored for GFP expression, collected 10 days after transfection, and lysed by three cycles of rapid freezing and thawing in methanol/dry ice bath. Collected viral lysate (P1) was clarified by centrifugation and high titer (P3) stock of Ad-InsP 3 R-N adenovirus was obtained by repetitive rounds of HEK293 cell infection. Obtained high titer stock of Ad-InsP 3 R-N adenovirus was aliquoted, frozen at Ϫ70°C and transported on dry ice from Dallas, TX, to St. Petersburg, Russia for experiments.
Ca 2ϩ Imaging-HEK293 cells grown on glass coverslips were loaded with 5 M Fura-2AM in the presence of 0.025% Pluronic for 20 -30 min. Loaded cells were illuminated by alternating 340 and 380 nm excitation light at 2-Hz frequency. Emission fluorescence intensity was measured at 510 nm with the use of an InCyt Basic I/P dual wavelength fluorescence imaging system (Intracellular Imaging Inc., Cincinnati, OH). The change of cytosolic Ca 2ϩ concentration in cells was expressed as the ratio of emission fluorescence intensity at 340 and 380 nm excitation wavelengths (340/380 ratio). In experiments with recombinant adenovirus HEK293 cells were infected using high titer stock of Ad-InsP 3 R-N 2-3 days prior to Ca 2ϩ imaging experiments. The efficiency of infection was close to 100% based on GFP fluorescence.
Whole Cell Recordings-Whole cell recordings were performed using an Axopatch 200B patch clamp amplifier (Axon Instruments) with a conventional 10-G⍀ feedback resistance in the head stage. Resistance of Sylgard-coated, fire-polished glass microelectrodes was 3-5 M⍀. Series resistance was not compensated. The pipette solution contained (in mM) 145 N-methyl-D-glutamine aspartate, 10 Cs-EGTA (or 12 Cs-BAPTA), 10 Cs-HEPES, pH 7.3, 1.5 MgCl 2 , and either 4.5 CaCl 2 (pCa 7.0) or no CaCl 2 added (pCa Ͼ 9). Extracellular solution contained (in mM) 140 N-methyl-D-glutamine aspartate, 10 BaCl 2 , 10 Cs-HEPES, pH 7.3. During recording the currents were sampled at 5 kHz and filtered digitally at 500 Hz. pClamp6 software suite (Axon Instruments) was used for data acquisition and analysis. In all whole cell experiments the holding potential was 0 mV. Periodically (once every 4 -30 s) the membrane potential was stepped to Ϫ100 mV (for 30 ms), and a 170-ms voltage ramp to ϩ70 mV was applied. Traces recorded before current activation were used as a template for leak subtraction. The recorded currents were normalized to the cell capacitance. The mean value of cell capacitance was 40 Ϯ 5 picofarads (n ϭ 41).
Single Channel Recordings-Single channel recordings were performed using an Axopatch 200B patch clamp amplifier and glass pi- The single channel recordings were digitized at 5 kHz and filtered at 100 Hz for analysis and presentation. pClamp6 software suite (Axon Instruments) was used for data acquisition and analysis. Unitary current amplitude was determined from current records and all-point amplitude histograms. The experiments were carried out at room temperature (22-24°C (30) ) as a standard way to compare open channel probability among different experiments.

Store-operated and Receptor-operated Ca 2ϩ
Influx Pathways in HEK293 Cells-Application of 100 M UTP to Fura-2-loaded HEK293 cells maintained in Ca 2ϩ -free media resulted in transient cytosolic Ca 2ϩ elevation due to release of Ca 2ϩ from intracellular InsP 3 -sensitive Ca 2ϩ stores (Fig. 1A). Subsequent re-addition of 2 mM extracellular Ca 2ϩ resulted in second elevation of cytosolic Ca 2ϩ due to Ca 2ϩ entry via plasma membrane Ca 2ϩ channels (Fig. 1A). Application of membrane-permeable low affinity Ca 2ϩ chelator TPEN leads to rapid and reversible reduction of intraluminal Ca 2ϩ concentration without disturbing cytosolic Ca 2ϩ levels. We found that depletion of Ca 2ϩ stores in HEK293 cells by 1 mM TPEN in Ca 2ϩ -free solution induces activation of Ca 2ϩ influx following Ca 2ϩ readdition ( Fig. 1B). Ca 2ϩ influx induced by depletion of intraluminal Ca 2ϩ (Fig. 1B) is a signature of store-operated Ca 2ϩ (SOC) channels, which have been previously described in many non-excitable cells (1-3, 7). Thus, Ca 2ϩ influx in response to UTP (Fig. 1A) can be mediated by plasma membrane receptoroperated Ca 2ϩ (ROC) channels as a direct result of P2Y2 receptor activation or as result of SOC activation following depletion of InsP 3 -sensitive intracellular Ca 2ϩ stores. To determine if depletion of intracellular Ca 2ϩ stores is required for UTP-induced Ca 2ϩ influx, we infected HEK293 cells with an adenovirus encoding the InsP 3 -binding amino-terminal region of rat InsP 3 R1 (Ad-InsP 3 R-N). Consistent with published observations (34), UTP-induced Ca 2ϩ release was abolished in Ad-InsP 3 R-N-infected cells (Fig. 1C), presumably due to ability of InsP 3 R-N fragment to function as the "InsP 3 sponge" (34). Despite absence of Ca 2ϩ release from InsP 3 -sensitive stores, addition of extracellular Ca 2ϩ resulted in Ca 2ϩ influx in Ad-InsP 3 R-N-infected cells exposed to UTP (Fig. 1C). Thus, we concluded that depletion of intracellular stores is not required for UTP-induced Ca 2ϩ influx in HEK293 cells, which appear to be at least in part mediated by ROC channels.
Whole Cell Analysis of Ca 2ϩ Influx Pathways in HEK293 Cells-In the next series of experiments we measured whole cell Ca 2ϩ currents activated in HEK293 cells by extracellular UTP. We found that in some HEK293 cells (2 out of 8) application of UTP induced highly Ca 2ϩ -selective current with strong inward rectification ( Fig. 2A). In other HEK293 cells (6 out of 8) the currents induced by UTP displayed linear currentvoltage relationship with significant outward current at test potentials more positive than ϩ30 mV (Fig. 2B). To characterize Ca 2ϩ currents activated by depletion of intracellular Ca 2ϩ stores, we repeated whole cell recordings with addition of 12 mM BAPTA into pipette. We found that in some cells (2 out of 12) depletion of Ca 2ϩ stores induced highly Ca 2ϩ -selective current with strong inward rectification (Fig. 2C). In other cells (5 out of 12) less Ca 2ϩ selective linear current was induced by store depletion (Fig. 2D). In our previous whole cell experiments with A431 cells we observed activation of highly Ca 2ϩselective (I CRAC ) and moderately Ca 2ϩ -selective (I SOC ) currents by extracellular UTP and intracellular BAPTA (19). It appears that similar currents are also present in HEK293 cells (Fig. 2, A-D). In addition, some HEK293 cells (5 out of 12) express non-selective outwardly rectifying current activated by store depletion (Fig. 2E). This type of current was also rarely observed in whole cell experiments with A431 cells (data not shown).
I min Channels Act as ROC Channels in HEK293 Cells-In the previous studies we utilized single channel patch clamp recordings to characterize Ca 2ϩ influx channels in A431 cells (13)(14)(15)(16)(17)(18)(19). What are the channels supporting Ca 2ϩ influx in HEK293 cells? To answer this question, we performed patch clamp recordings with HEK293 cells using 105 mM Ba 2ϩ in the pipette as a current carrier. We found that application of 100 M UTP to external surface of plasma membrane in cell-attached (c/a) patch evoked the activity of small conductance channels in 56 out of 118 attempts (Fig. 3A). Channels with similar properties were activated by application of 2.5 M InsP 3 to cytosolic surface of inside-out (i/o) patches (Fig. 3B) in 18 out of 80 experiments. The unitary current-voltage relationship of UTP-activated (in c/a experiments) and InsP 3 -activated (in i/o experiments) channels is shown on Fig. 3C. The linear fit to the current-voltage relationship resulted in a slope conductance of 1.2 pS for both UTP-activated (Fig. 3C, filled symbols) and InsP 3 -activated (Fig. 3C, open symbols) channels. The extrapolated reversal potential of these channels was equal to ϩ50 mV (Fig. 3C), which corresponded to P Ba/K ϭ 20 (Equation 1) for ionic conditions used in our patch clamp experiments. Single exponential fit to the open dwell time distribution of UTPactivated channels yielded a mean open time of 9 ms (Fig. 3D). In our previous studies with A431 cells (13-15, 18, 19) we described UTP-and InsP 3 -activated channels with similar conductance and gating properties, which we called "I min ." Based on similarity in single channel properties (Table I) we concluded that HEK293 cells also express I min channels that can be activated by UTP (in c/a) and InsP 3 (in i/o) (Table II).
I min channels in A431 cells can also be activated by depletion of intracellular Ca 2ϩ stores (18,19). Does depletion of intracellular Ca 2ϩ stores activate I min channels in HEK293 cells? To answer this question we performed a series of c/a patch clamp recording experiments with HEK293 cells. In these experiments passive depletion of intracellular Ca 2ϩ stores was induced by application of 1 mM TPEN in the bath (n ϭ 84), loading cells with 100 M BAPTA-AM in the presence of 1 M thapsigargin in the bath (n ϭ 20), or addition of 1 M thapsigargin to the pipette solution (n ϭ 30). Although these manipulations induced I min activity in A431 cells (18,19), none of these experiments yielded I min activity in the c/a patches from HEK293 cells (Table II). Thus, we concluded that I min channels in HEK293 cells are store-independent (Table I).
PIP 2 Inhibits I min Channels in HEK293 Cells-Experiments described in the previous section led us to conclude that I min channels in HEK293 cells function as InsP 3 -gated ROC channels. What is a mechanism that couples PLC-coupled P2Y2 receptors with activation of I min channels in intact cells? In the previous studies we proposed that I min channels in A431 cells are functionally coupled to PLC⅐PIP 2 ⅐InsP 3 R signaling complex (17,18). Are I min channels in HEK293 cells also coupled to PIP 2 ? To answer this question, we performed a series of insideout (i/o) patch clamp experiments with anti-PIP 2 monoclonal antibodies (PIP 2 Ab). We found that addition of PIP 2 Ab facilitated the activity of InsP 3 -gated I min channels (Fig. 4A). By reversing the order of InsP 3 and PIP 2 Ab additions to the patch, we established that PIP 2 Ab alone was not sufficient for I min activation but subsequent application of InsP 3 was effective in 48 of 60 experiments (Fig. 4B). Thus, InsP 3 was able to activate I min channels in i/o patches pretreated with PIP 2 Ab in 80% of experiments (48 out of 60), much higher than the 22% (18 out of 80) success rate in untreated patches (Table II). Thus, we reasoned that a large fraction of I min channels in unstimulated HEK239 cells are under tonic inhibition by PIP 2 . To test this hypothesis further, we repeated i/o experiments with addition of 100 M UTP into the pipette solution. We reasoned that addition of UTP to the pipette will activate P2Y2 receptors in the patch leading to local hydrolysis of PIP 2 by PLC and relieve of I min inhibition. Consistent with our expectations, addition of InsP 3 activated I min channels in 7 out of 18 experiments (40%) when UTP was present in the pipette. Thus, addition of UTP in the pipette increased the success rate of I min channels activation by InsP 3 from 22% to 40%. These results are in agreement with our previous studies of I min activation mechanisms in A431 cells (17,18), indicating that I min channels in both A431 in Ca 2ϩ -free medium is followed by a Ca 2ϩ elevation upon re-addition of Ca 2ϩ to the extracellular medium. B, Ca 2ϩ store depletion was induced by 1 mM TPEN added to the bath in the Ca 2ϩ -free medium. Subsequent re-addition of 2 mM extracellular Ca 2ϩ resulted in Ca 2ϩ entry. C, application of 100 M UTP to HEK293 cells infected with Ad-InsP 3 R-N adenovirus fails to induce Ca 2ϩ release in Ca 2ϩ -free medium, but subsequent re-addition of 2 mM extracellular Ca 2ϩ resulted in Ca 2ϩ elevation due to Ca 2ϩ entry. The data shown in panels A-C are representative of three to five experiments. and HEK293 cells are inhibited by PIP 2 (Table I).
I max Channels Act as Both ROC and SOC Channels in HEK293 Cells-If I min channels in HEK293 cells are storeindependent, then what channels support store depletion activated Ca 2ϩ influx (Fig. 1B) in HEK293 cells? One possibility is that store-dependent Ca 2ϩ influx in HEK293 cells is supported by highly Ca 2ϩ -selective I CRAC channels (Fig. 2C). However, whole cell recordings revealed that in some cells depletion of intracellular Ca 2ϩ stores causes activation of moderately Ca 2ϩselective I SOC currents (Fig. 2D). Which channels mediate I SOC currents in HEK293 cells? Examination of current records obtained in c/a experiments with HEK293 cells revealed that in 47 out of 118 experiments UTP activated channels with higher amplitude and different gating kinetics than I min (Fig. 5A). To distinguish these channels from I min we called them "I max ." Frequently (n ϭ 39) I min and I max channels were observed in the same patch. For single channel analysis we selected c/a patches where only I max channels were observed in response to UTP (n ϭ 8). Analysis of these experiments revealed that the unitary current-voltage relationship of I max channels was nonlinear ( Fig. 5B) with the slope conductance of 17 pS between Ϫ100 and Ϫ80 mV and 10 pS between Ϫ40 and Ϫ10mV. The extrapolated reversal potential of I max channels was equal to ϩ30 mV (Fig. 5B), which corresponded to P Ba/K ϭ 4 (Equation 1) for ionic conditions used in our patch clamp experiments. The open lifetime distribution of I max channels could be fitted by sum of two exponential functions with time constants 1 ϭ 2.1 ms and 2 ϭ 32 ms (Fig. 5C). What is a mechanism of I max channels activation? Similar to I min channels, I max channels could be activated by application of InsP 3 to cytosolic surface of inside-out patches from HEK239 cells in 15 out of 80 experiments (Fig. 5D) (Table II). PIP 2sensitivity of I max channels has not been investigated in this study. These results indicated that I max channels in HEK293 cells can function as receptor-activated (ROC) channels. In our previous i/o patch clamp experiments with A431 cells we described InsP 3 -gated channels with properties similar to I max (13). Application of 1 mM TPEN resulted in activation of I max channels in 18 out of 84 c/a patches from HEK293 cells (Fig. 5E) (Table II), indicating that I max channels are storeoperated. Thus, in contrast to "ROC-only" I min channels, I max channels contribute to both ROC and SOC currents in HEK293 cells.
I min and I max Channels Display Different Conductance for Monovalent Cations-Experiments described above suggested an existence of at least two distinct Ca 2ϩ influx channel types (I min and I max ) in plasma membrane of HEK293 cells. To investigate this possibility further, we performed a series of i/o patch clamp recordings in divalent-free media containing 140 mM Na ϩ . In the absence of divalent cations, Na ϩ and other monovalent cations carry a substantial current through voltage-gated Ca 2ϩ channels (35,36). In our previous studies, we characterized monovalent currents mediated by I min channels in A431 cells (19). We found that addition of 10 M of InsP 3 to cytosolic surface of i/o patches from HEK293 cells resulted in channel activity in 12 out of 24 attempts. Two distinct channel types were recorded in these experiments (Fig. 6). In six experiments we observed 10 pS conductance channels (Fig. 6, A and  B) with the properties similar to I min channels recorded in divalent-free media in A431 cells (19) (Table I). In another six experiments we observed 33 pS channels (Fig. 6, C and D). Because both I min and I max can be activated by application of InsP 3 to i/o patches taken from HEK239 cells (Figs. 3B and 5D), we reasoned that 10 pS channels correspond to monovalent currents via I min channels and 33 pS channels correspond to monovalent currents via I max channels.
Non-selective I NS Channels Contribute to SOC Pathway in HEK293 Cells-Divalent ion selectivity of I max channels (Fig.  5B) is consistent with reversal potential of SOC currents observed in some HEK293 cells (Fig. 2D). However, other HEK293 cells displayed non-selective SOC current (Fig. 2E) with reversal potential incompatible with I max permeability properties. Are there additional store-operated channels in HEK239 cells? Careful examination of our data revealed that in 10 out of 118 attempts application of 100 M UTP induced activity of yet another channel type in c/a patches (Fig. 7A) (Table II). In contrast to I min and I max channels, the current via these channels had an outward direction when recorded at membrane potentials more positive than Ϫ10mV. The unitary current-voltage relationship of these channels is shown on Fig.  7B. The current-voltage relationship of these channels was linear between Ϫ90 and Ϫ50 mV with the slope conductance of 5 pS. The current reversal potential was Ϫ10 mV, consistent with non-selective channels (P Ba/K ϭ 0.4, Equation 1). Thus, we called these channels I NS (non-selective channels).  (Table II). However, in 12 out of 84 attempts application of 10 mM TPEN induced activity of I NS channels in c/a patches (Fig. 7D and Table II). The I NS channels can also be activated by loading HEK293 cells with 100 M BAPTA-AM (Fig. 7E) in 8 out of 20 experiments. Thus, we concluded that I NS channels are store-operated and contribute to SOC currents, but not to ROC currents, in HEK293 cells. Based on similarity in ionic selectivity, it is likely that non-selective SOC current observed in some of our whole cell experiments with TPEN (Fig. 2E) is primarily carried by I NS channels. DISCUSSION Ca 2ϩ Influx Channels in HEK239 and A431 Cells-By using a combination of whole cell and single channel patch clamp recordings we identified four distinct types of endogenous   Ca 2ϩ influx channels in HEK293 cells: I CRAC , I min , I max , and I NS . Each of the channels has unique permeability properties and activation profiles. I CRAC channels are highly selective for divalent cations (P D/M Ͼ 1000) and display strong inward rectification. In HEK293 cells I CRAC channels can be activated by application of PLC-linked receptor agonist such as UTP ( Fig. 2A) or by depletion of intracellular Ca 2ϩ stores (Fig. 2C). In the previous studies we observed I CRAC currents with similar properties in A431 cells (19). As determined in previous studies by us (19) and others (6, 37) unitary conductance of I CRAC channels is too small (estimated at 24 f S) to allow single channel analysis. Thus, we have not been able to investigate the properties of I CRAC channels in HEK293 cells at the single channel level. I min channels display 1.2 pS conductance for divalent cations (Fig. 3C), 10 pS conductance for monovalent cations (Fig. 6B) and moderate selectivity for divalent cations (P Ba/K ϭ 20) (Fig.  3C). The conductance, permeability, and gating properties of I min channels in HEK293 cells are identical to the properties of I min channels that we described in A431 cells (Table I). In HEK293 cells I min channels are activated by UTP (c/a) (Fig. 3A) or InsP 3 (i/o) (Fig. 3B), but not by store depletion (Table II). Thus, in HEK293 cells I min channels function as ROC channels, but not as SOC channels. Importantly, we confirmed an existence of ROC Ca 2ϩ influx pathway in HEK293 cells in Ca 2ϩ imaging experiments with Ad-InsP 3 R-N adenovirus (Fig. 1C). Similar to A431 cells (17), I min channels are functionally coupled to PIP 2 in HEK239 cells (Fig. 4). Quantitative comparison of results obtained in experiments with PIP 2 Ab in inside-out patches taken from HEK293 cells (Fig. 4) and A431 cells (17,18) shows that I min channels are under stronger tonic inhibition by PIP 2 in HEK293 cells than in A431 cells. Most likely the explanation of this phenomenon is higher levels of global or local PIP 2 levels in HEK293 cells.
I max channels display 17-pS conductance for divalent cations (Fig. 5B), 33-pS conductance for monovalent cations (Fig. 6D), and less selective for divalent cations than I min channels (P Ba/K ϭ 4) (Fig. 5B). In HEK293 cells I max channels can be activated by UTP (c/a) (Fig. 5A), InsP 3 (i/o) (Fig. 5D), or by store depletion (Fig. 5E) (Table II). Thus, in HEK293 cells I max channels function as both ROC and SOC channels. In addition to I min and I max , we also identified non-selective I NS channels with 5-pS conductance (Fig. 7B) and P Ba/K ϭ 0.4 (Fig. 7B). I NS channels are not InsP 3 -gated but can be activated by depletion of Ca 2ϩ stores (Fig. 7, D and E) (Table II). Thus, in HEK293 cells I NS channels function as SOC channels but not as ROC channels. In A431 cells we also observed I max and I NS channels (Ref. 13 and data not shown), but in our recent studies we focused primarily on I min (14,15,18,19).
Obtained results lead us to conclude that in HEK239 cells ROC currents are supported by a combination of I min and I max channels (with the possible contribution of I CRAC ), and SOC currents are supported by I CRAC , I max , and I NS channels. Single channel recordings reveal that each of these channels displays unique conductance and selectivity properties. Although used widely, Ca 2ϩ imaging (Fig. 1) and whole cell currents (Fig. 2) report combined Ca 2ϩ influx via multiple channel types. Thus, conclusions regarding ion selectivity and activation mechanisms of ROC and SOC currents based on Ca 2ϩ imaging and whole cell experiments can be misleading without knowledge of single channel properties of channels supporting these currents.
The molecular identity of channels encoding I CRAC Ca 2ϩ influx channels in HEK293 and A431 cells remains unknown, but the members of TRP family are the most likely candidates for the role of I min , I max , and I NS channels. HEK293 cells express endogenous TRPC1, TRPC3, TRPC4, and TRPC6 proteins (28 -30). Our preliminary molecular analysis (in collaboration with William Schilling) confirmed an expression of multiple TRPC isoforms in A431 cells. The conductance and selectivity properties of I max and I NS channels are consistent with functional properties displayed by TRPC proteins when expressed as homo-or heteromers (3, 7, 9 -12). In particular, conductance and selectivity of I max channels resembles properties of the channels formed by TRPC1 when expressed in the human submandibular gland cell line (38). Previous Ca 2ϩ imaging studies demonstrated that suppression of TRPC4 expression in HEK293 cells by RNA interference lead to reduction of ROC, but not SOC Ca 2ϩ influx pathway (30). In contrast, suppression of TRPC3 expression in the same study affected both ROC and SOC Ca 2ϩ influx pathways in HEK293 cells (30). Thus, TRPC3 and TRPC4 subunits are likely candidates to encode I max and I NS channels in HEK293 cells. Single channel conductance of I min channels is lower and selectivity for divalent cations is higher than the range of values reported for the channels formed by TRPC subunits. Thus, I min channels may be encoded by a more selective member of TRP family such as TRPV6 (39 -42). Application of molecular techniques, such as RNA interference (43), should help to clarify molecular composition of I min , I max , and I NS Ca 2ϩ influx channels in HEK293 and A431 cells (30,44). In addition, recently discovered specific SOC inhibitors tetrapandin peptides (45) will be useful in identification of SOC-supporting channels in HEK293 and A431 cells.
From functional analysis we concluded that I min channels display identical functional properties in A431 and HEK293 cells (Table I) and most likely encoded by the same member of TRP family in both types of cells. However, I min channels are store-operated in A431 cells (18,19) and store-independent in HEK293 cells (Table II). Conflicting results regarding store dependence of TRP-supported channels have been obtained by a variety of groups (3, 7, 9 -12). Usually these discrepancies are interpreted as an artifact of overexpression system. In our studies we find that endogenous I min channels are store-dependent in A431 cells and store-independent in HEK293 cells. Thus, most likely coupling of I min channels to store depletion depends on an accessory protein that is present in A431 cells but is absent in HEK293 cells. In addition, different InsP 3 R isoforms expressed in A431 and HEK293 cells may affect store dependence of I min activation. Definitive molecular identification of I min -encoding proteins is required to address this issue.