Capacitative Ca²⁺ entry (CCE) is a ubiquitous mode of Ca²⁺ influx in nonexcitable cells, leading to a sustained elevation in the cytoplasmic Ca²⁺ concentration ([Ca²⁺]_i) (Putney, 1986). This mode of Ca²⁺ influx is activated by and occurs after the intracellular Ca²⁺ stores have been depleted, either by phospholipase C-linked agonists or pharmacological agents such as thapsigargin (TG). The exact mechanism that couples store depletion to CCE remains elusive, but it has recently been postulated to involve the Ca²⁺ sensor STIM1 (Zhang et al., 2005). In addition, the molecular identity of the CCE channel is far from established.

The TRPC family of channels has been suggested as likely candidates for CCE channels (Putney, 2004). A closely related subset of this family, TRPC3, TRPC6, and TRPC7, can be directly activated by diacylglycerol (DAG) and its synthetic analog 1-oleyl-2-acetyl-sn-glycerol (OAG) (Hofmann et al., 1999). Significantly, these DAG-activated TRP subunits are present endogenously in human embryonic kidney (HEK) 293 cells (Garcia and Schilling, 1997; Zagranichnaya et al., 2005).

The Ca²⁺-stimulable adenylyl cyclase type 8 (AC8) is a focal point for cross-talk between the Ca²⁺ and cAMP signaling pathways. In nonexcitable cells, AC8 is activated by CCE and is closely associated with CCE channels. For example, BAPTA, a fast chelator of Ca²⁺, is able to prevent AC8 stimulation by CCE, whereas at equivalent concentrations, EGTA, a slower chelating agent, is ineffective (Fagan et al., 1998). Moreover, residence of AC8 in cholesterol-rich caveolae is required for its regulation by CCE (Fagan et al., 2000; Smith et al., 2002) and contributes to the apposition of the signaling molecules of the Ca²⁺ and cAMP pathways. TRPC3 localizes to caveolae and interacts directly with caveolin-1 (Lockwich et al., 2001), thereby placing it in the same microdomain as AC8, in an ideal position to mediate CCE.

Due to their close proximity to CCE channels, Ca²⁺-stimulable adenylyl cyclases, both endogenously expressed and transiently transfected, have been shown to be selective for CCE over other types of Ca²⁺ entry, such as ionophore-mediated entry (Fagan et al., 1998). This selectivity of AC8 for Ca²⁺ entry occurring via the capacitative pathway has been used previously to discriminate between the capacitative and arachidonate-activated Ca²⁺ entry pathways in HEK 293 cells (Shuttleworth and Thompson, 1999). In this study, we asked whether OAG-activated TRP subunits might form part of the CCE apparatus using the Ca²⁺ stimulation of AC8 as a discerning measure of CCE to determine whether OAG activates the same subset of Ca²⁺ channels as store depletion.

Materials and Methods

Materials. [2-³H]Adenine, [2,8-³H]cAMP, and [α-³²P]ATP and DNase were purchased from GE Healthcare (Little Chalfont, Buckinghamshire, UK). Fura-2/AM and Pluronic F-127 were purchased from Molecular Probes (Paisley, UK). Thapsigargin, forskolin, and 1-oleyl-2-acetyl-sn-glycerol were purchased from Calbiochem (Nottingham, UK). All other chemicals and culture media were purchased from Sigma (Poole, Dorset, UK), BDH (Poole, Dorset, UK), or Fisher Scientific (Loughborough, UK).

Cell Culture and Transfection of HEK 293 Cells. HEK 293 cells were cultured in minimal essential medium Eagle (MEM), supplemented with 10% fetal bovine serum, 50 μg/ml penicillin, 50 μg/ml streptomycin, 100 μg/ml neomycin, and 2 mM l-glutamine. Cells were maintained at 37°C in a humidified atmosphere of 5% CO₂/95% air. One day before transient transfection, cells were plated on 100-mm diameter dishes at ∼50% confluence. The Ca²⁺ phosphate method of transfection was adopted (Chen and Okayama, 1987) using 2 μg of AC8 cDNA, and cells were used 2 days after transfection.

Measurement of [Ca²⁺]_i in Cell Populations. [Ca²⁺]_i was measured in Fura-2/AM-loaded cells using a spectrofluorometer (LS50B; PerkinElmer Life and Analytical Sciences, Boston, MA). Cells were detached with phosphate-buffered saline containing EDTA (1 mM) and loaded with Fura-2/AM (2 μM) and 0.02% Pluronic F-127 in serum-free MEM containing 20 mM HEPES and 0.1% bovine serum albumin, pH 7.4, for 30 min at room temperature. Cells were washed and resuspended in MEM. Before experiments, 4 × 10⁶ cells were resuspended in 3 ml of Krebs' buffer (120 mM NaCl, 4.75 mM KCl, 1.44 mM MgSO₄, 11 mM d-glucose, and 25 mM HEPES adjusted to pH 7.4 with 2 M Tris base), supplemented with 0.1 mM EGTA, and transferred to a stirred cuvette. After a 1-min equilibrium interval, test substances were added from 100-fold stock solutions. The maximal ratio of fluorescence (R_max) was determined in the presence of Triton-X (0.1% final); R_min was determined in the presence of EGTA (5 mM) and Tris base (30 mM). Fluorescence emission ratios (F₃₄₀/F₃₈₀) were converted to Ca²⁺ concentrations by applying the Grynkiewicz equation (Grynkiewicz et al., 1985). Mn²⁺ quench experiments were carried out in the absence of EGTA, excitation wavelengths were set to 340 and 360 nm, and extracellular Ca²⁺ was substituted for Mn²⁺ (0.5 mM). Quantitation of peak [Ca²⁺]_i responses was performed by averaging four independent experiments and are shown as mean ± S.E.M. Initial rates of Ca²⁺ entry were calculated over a 10-s period, starting 2 s after the addition of extracellular Ca²⁺; a linear regression curve was fitted to the data, and the initial rate was determined from the slope of the fitted curve (n = 3). For the Gd³⁺ block experiments, MgSO₄ was substituted with MgCl₂, EGTA was omitted, and the peak [Ca²⁺]_i response was calculated as a percentage of the control response after subtraction of background (n = 3). Statistical significance was determined using Student's t test.

Measurement of [Ca²⁺]_i in Single Cells. Cells were plated onto 25-mm poly(l-lysine)-coated coverslips 24 h before loading with Fura-2/AM (2 μM) and 0.02% Pluronic F-127 for 40 min at room temperature in Krebs' buffer supplemented with 1 mM CaCl₂. After loading, cells were washed several times and then imaged using a CoolSNAP-HQ CCD camera (Photometrics, Tucson, AZ) and monochromator system (Cairn Research, Kent, UK) attached to a Nikon TMD microscope (×40 objective; Nikon, Tokyo, Japan). Emission images (D510/80M) at 340 and 380 nm excitation were collected every second using MetaFluor software (Universal Imaging, Sunnyvale, CA). Data were plotted as 340/380 nm ratio changes relative to the fluorescence ratio before the addition of extracellular Ca²⁺ (Δ340/380).

Preparation of HEK 293 Cell Membranes. Crude membranes were prepared from HEK 293 cells transiently transfected with AC8 cDNA, as described previously (Nakahashi et al., 1997). In brief, cells were sheared by passage through a 22-gauge needle 10 times in homogenization buffer [2 mM MgCl₂, 1 mM EDTA, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 1 mM benzamidine, 1 μg DNase, and 50 mM Tris, pH 7.4]. Membranes were collected by centrifugation (23,000g, 15 min, 4°C), resuspended in assay buffer [40 mM Tris-Cl, 800 μM EGTA, and 0.25% bovine serum albumin (fraction V), pH 7.4], and stored in liquid nitrogen until use.

Determination of Free Ca²⁺Concentrations. Free Ca²⁺ concentrations were established using an EGTA buffering system as described previously (Ahlijanian and Cooper, 1987). The BAD4 computer program (Brooks and Storey, 1992) was used to estimate free Ca²⁺ concentrations by solving equations, which described the complexes formed within a mixture composed of the assay buffer components (see Measurement of Adenylyl Cyclase Activity). Final assay mixture concentrations of free Ca²⁺ (in the presence of 200 μM EGTA) are shown.

Measurement of Adenylyl Cyclase Activity. AC8 activity was measured as described previously (Boyajian et al., 1991) with some modifications. In brief, HEK 293 membranes (see Preparation of HEK 293 Cell Membranes) were incubated (15 min, 30°C) in the presence of 12 mM phosphocreatine, 1.4 mM MgCl₂, 40 μM GTP, 100 μM cAMP, 100 μM ATP, 25 units/ml creatine kinase, 70 mM Tris-Cl, 500 μM isobutylmethylxanthine (IBMX), 0.5 μCi [α-³²P]ATP, and Ca²⁺ (see Determination of Free Ca²⁺ Concentrations) in the presence and absence of 30 μM OAG. Reactions were terminated with the addition of 100 μl of stopper solution (0.5% w/v SDS, 22 mM ATP, and 1.5 mM cAMP), and the [³²P]cAMP formed was quantified using standard Dowex/alumina methods and using [³H]cAMP as a recovery marker (Salomon et al., 1974).

Measurement of cAMP Accumulation. cAMP accumulation in intact cells was measured as described previously (Evans et al., 1984) with some modifications (Fagan et al., 1996). HEK 293 cells were transiently transfected with AC8 and grown on 24-well plates. Cells were incubated (2 h, 37°C) with [³H]adenine (1.5 μCi/well) and then washed and incubated (10 min at 30°C) in nominally Ca²⁺-free Krebs' buffer (900 μl/well) supplemented with 0.1% bovine serum albumin. Experiments were carried out at 30°C in the presence of EGTA (100 μM) and IBMX (100 μM), which were preincubated with the cells for 10 min before a 1-min assay. Cells were treated with TG (0.1 μM) for 4 min to empty intracellular Ca²⁺ stores before the addition of 100 μl of assay solution (Krebs', 10 μM forskolin, and CaCl₂ as shown) and terminated by the addition of trichloroacetic acid (5% w/v final). Cells were incubated on ice for at least 30 min followed by the addition of unlabeled cAMP (1 μM), unlabeled ATP (0.65 μM), and [α-³²P]ATP (∼5000 cpm) used as a recovery marker. After pelleting (4000g, 6 min), the [³H]ATP and [³H]cAMP content of the supernatant was quantified using the standard Dowex/alumina methodology (Salomon et al., 1974). Accumulation of cAMP is expressed as the percentage of conversion of [³H]ATP into [³H]cAMP and shown as means ± S.D. of triplicate determinations.

Results

Characterization of OAG-Activated Ca²⁺ Entry. TRPC3, TRPC6, and TRPC7 are expressed endogenously in HEK 293 cells (Garcia and Schilling, 1997; Zagranichnaya et al., 2005). These TRP subunits are activated by DAG and its membrane-permeant synthetic analog, OAG (Hofmann et al., 1999). It was advantageous to use OAG to activate TRP subunits in our experiments because, unlike DAG, OAG is not metabolized to arachidonate, thereby allowing the DAG-mediated pathway to be distinguished from the arachidonate-activated pathway. Cells were suspended in nominally Ca²⁺-free Krebs' solution and treated for 1 min with a range of OAG concentrations (0-100 μM). When extracellular Ca²⁺ (2 mM) was restored, a sharp increase in [Ca²⁺]_i was evoked. The amplitude of the Ca²⁺ signal was dependent on the OAG concentration (Fig. 1A). The increase in [Ca²⁺]_i induced by OAG was not due to CCE because during the 5-min period in which the cells were in Ca²⁺-free medium, no depletion of the stores was observed (data not shown).

The possibility exists that OAG may exert its effects by stimulating protein kinase C (PKC) (Rozengurt et al., 1984). To address this issue, the Ca²⁺ entry experiment in response to OAG was repeated in the presence of two PKC inhibitors, staurosporine (Fig. 1B), a broad spectrum inhibitor of protein kinases (Tamaoki et al., 1986), and bisindolylmaleimide I (Fig. 1C), which shows high selectivity for PKC isoforms α, β_I, β_II, γ, δ, and ϵ (Toullec et al., 1991). Only in the presence of staurosporine and at the lowest concentration of OAG (30 μM) was the peak [Ca²⁺]_i response after treatment with OAG (305 ± 51 nM; n = 4) significantly different relative to the control response in the absence of PKC inhibitor (196 ± 29 nM; n = 4). High OAG concentrations produced saturating responses in HEK 293 cells; therefore, a concentration of OAG of 30 μM was used in subsequent experiments.

OAG Causes Ca²⁺ Entry through the Plasma Membrane. To investigate the possibility that, rather than causing Ca²⁺ entry, OAG may be elevating [Ca²⁺]_i by inhibiting Ca²⁺ extrusion or causing release from intracellular stores, a Mn²⁺ quench experiment was carried out. Addition of extracellular Mn²⁺ caused a leak of Mn²⁺ into the cell, quenching the fluorescence of Fura-2 at both 340 and 360 nm. Once the fluorescence signal was stabilized, cells were treated with TG (Fig. 2A). Release of Ca²⁺ from intracellular stores, caused by TG, evoked an increase in fluorescence at 340 nm. On the other hand, at 360 nm, the isobestic point of Fura-2, Mn²⁺ entry through the capacitative pathway caused a decrease in fluorescence. In contrast, treatment with OAG decreased the fluorescence of Fura-2 at both wavelengths, indicating that OAG is only causing Mn²⁺ entry through the plasma membrane (Fig. 2B).

OAG-Activated Ca²⁺Entry Evokes a Larger and Faster Increase in [Ca²⁺]_i than CCE in Populations of HEK 293 Cells. To determine whether OAG activates the same subset of Ca²⁺ channels as those mediating CCE, the profiles of the [Ca²⁺]_i responses arising after treatment with OAG and store depletion were compared in populations of cells. TG, a sarcoplasmic-endoplasmic reticulum calcium pump inhibitor, is routinely used to empty intracellular stores independently of receptor activation, thereby avoiding the production of DAG through the phospholipase C pathway. CCE was triggered by depleting the stores with 0.1 μM TG after 1-min equilibration in Ca²⁺-free Krebs (Fig. 3A). Depletion of the Ca²⁺ stores gave rise to a modest increase in [Ca²⁺]_i, which returned to basal level within a few minutes. At 5 min, a range of Ca²⁺ concentrations (0-4 mM) was added to the extracellular medium, and CCE was observed as a sharp increase in [Ca²⁺]_i, the magnitude of which was a function of the extracellular [Ca²⁺]. Although a longer incubation of TG would give rise to a larger CCE response (Fagan et al., 1998), the chosen conditions enabled OAG-activated Ca²⁺ entry to be triggered at the same time point as CCE while minimizing passive depletion of intracellular Ca²⁺ stores.

For comparative purposes, OAG-activated Ca²⁺ entry was triggered in an analogous fashion; instead of depleting stores with TG, cells were treated for 1 min with OAG (30 μM) (Fig. 3B). At 5 min, the same range of Ca²⁺ concentrations (0-4 mM) was added. Entry of Ca²⁺ by the OAG-activated pathway led to a dramatic increase in [Ca²⁺]_i. As in the case of CCE, the amplitude of the increase in [Ca²⁺]_i was dependent on the extracellular [Ca²⁺]. Under these experimental conditions, the overall amplitude and the initial rate of the Ca²⁺ entry through the OAG-activated pathway were significantly greater than for CCE (Fig. 3C).

Low Concentrations of Gd³⁺ Differentially Inhibit CCE and OAG-Activated Ca²⁺Entry. Because Gd³⁺ has been reported to inhibit CCE relatively specifically at low concentrations (Broad et al., 1999; Luo et al., 2001b), we investigated whether the cation could distinguish between CCE and OAG-activated Ca²⁺ entry. Although Gd³⁺ blocked both CCE (Fig. 4A) and OAG-activated Ca²⁺ entry (Fig. 4B), there was a difference in the sensitivities of the two pathways for this inhibitor. Whereas high concentrations of Gd³⁺ (5 and 10 μM) failed to discriminate between the two types of Ca²⁺ entry, at 1 μM, Gd²⁺ was more effective against CCE than OAG-activated Ca²⁺ entry (Fig. 4C). Thus, Gd³⁺ reveals subtle distinctions between the two pathways.

OAG-Activated Ca²⁺Entry Has No Significant Effect on AC8 Activity. Using the selective activation of AC8 by CCE in nonexcitable cells as a sensor for CCE, we sought to determine whether CCE and OAG-activated Ca²⁺ entry are mediated by the same channels or two distinct sets of channels. When transiently transfected in HEK 293 cells, AC8 retains its selective activation by CCE (Fagan et al., 1996) and does not affect the profiles of Ca²⁺ entry in response to TG and OAG observed in the wild-type cells (data not shown). The two pathways of Ca²⁺ entry were triggered using a protocol analogous to that described for the [Ca²⁺]_i measurement experiments. cAMP accumulation was used as a measure of AC8 activity, which was quantified during the period when [Ca²⁺]_i increased to its maximum, in the presence of forskolin and the phosphodiesterase inhibitor IBMX (Fig. 5A). CCE, triggered by store depletion after TG pretreatment (0.1 μM), caused a robust stimulation of AC8 activity (▪, Fig. 5A). In contrast, treatment with OAG (30 μM) had little effect (□, Fig. 5A), even though OAG caused a greater increase in [Ca²⁺]_i than CCE under the same conditions (Fig. 3).

Because PKC is activated by OAG (Rozengurt et al., 1984) and inhibits another Ca²⁺-sensitive adenylyl cyclase, AC6 (Lin et al., 2002), we sought to determine whether PKC might also inhibit AC8 and thus obscure possible stimulatory effects as a result of OAG-activated Ca²⁺ entry. Thus, we measured cAMP accumulation after treatment with OAG (30 μM) in the presence of the PKC inhibitors staurosporine (SSP, 1 μM; Fig. 5, ) or bisindolylmaleimide I (BIM I, 0.5 μM; ), which were preincubated for 20 min and remained throughout the experiment. However, even in the presence of the PKC inhibitors, OAG-activated Ca²⁺ entry had little effect on AC8 activity in intact cells (Fig. 5A).

We also tested whether OAG might directly affect AC8 activity and obscure any effect of OAG-activated Ca²⁺ entry on AC8. Indeed, because OAG is membrane-permeant, it could conceivably have a nonspecific effect on AC8 activity. AC8 activity was measured in vitro, on crude membranes isolated from HEK 293 cells stably expressing AC8, in the presence of increasing concentrations of Ca²⁺ (0 to 4.5 μM), and in the absence and presence of OAG (30 μM) (Fig. 5B). The profiles of stimulation observed in each case were very similar. Therefore, OAG did not affect the ability of AC8 to be stimulated by Ca²⁺, despite being present for longer than in the cAMP measurements.

Taken together, these results indicate that, whereas OAG gives rise to a Ca²⁺ entry that is greater than CCE, it has little effect on AC8 activity. Therefore, AC8 clearly discriminates between CCE and OAG-activated Ca²⁺ entry. The differential ability of the two types of Ca²⁺ entry to activate AC8 indicates that different subsets of Ca²⁺ channels are involved.

In Single Cells, OAG-Activated Ca²⁺Entry Occurs as a More Heterogeneous [Ca²⁺]_i Response Than Does CCE. The ability of AC8 to discriminate between CCE and OAG-activated Ca²⁺ entry indicates that these occur through distinct pathways, such that the [Ca²⁺]_i profiles of the two responses could be expected to differ. The averaged response measured in cell populations may mask more profound heterogeneity in individual cells. Thus, Ca²⁺ measurements were performed in individual HEK 293 cells, using protocols of stimulation analogous to those used for cell population experiments. After store depletion by TG, CCE occurred immediately upon the addition of 2 mM extracellular Ca²⁺ in all cells (Fig. 6A). Although cells varied in terms of the amplitude of the response, all responded with a smooth and sustained increase in [Ca²⁺]_i. In contrast, after treatment with OAG, the addition of 2 mM extracellular Ca²⁺ caused a heterogeneous response (Fig. 6B). Strikingly, only a few cells responded immediately or simultaneously to OAG. The delay before the response ranged from 0 to over 7 min. In addition, the amplitude and duration of the responses varied greatly between cells. Therefore, in contrast to the sustained increase in [Ca²⁺]_i through CCE, the response to OAG was more oscillatory in nature, revealing profound differences between the two pathways in individual cells.

Because different methodologies were used, the data obtained in single cells are not directly comparable with the data obtained from cell populations. However, in view of the delayed and complex response to OAG observed in many individual cells, we considered the possibility that these delayed transient increases in [Ca²⁺]_i triggered by OAG were capable of activating AC8, but that the effect might not be detected in the first minute after the addition of extracellular Ca²⁺, when cAMP accumulation is routinely measured. Therefore, cAMP accumulation was measured after treatment with TG or OAG as described previously over a 5-min period (Fig. 6C). Whereas CCE robustly activated AC8 at all concentrations tested, OAG-activated Ca²⁺ entry only caused a slight increase in AC8 activity at the highest concentration tested. Compared with CCE, OAG-activated Ca²⁺ entry was again ineffective.

The [Ca²⁺]_i Increases Resulting from OAG-Activated Ca²⁺Entry and CCE Are Additive. The fact that AC8 can distinguish between OAG-activated Ca²⁺ entry and CCE suggests that these two pathways are distinct. This argument would be strengthened if the increase in [Ca²⁺]_i resulting from treating the same batch of cells with both TG (0.1 μM, 4 min) and OAG (30 μM, 1 min) was equivalent to the sum of the increases triggered by treating one batch of cells with TG and another equivalent batch with OAG (Fig. 7A). To determine whether this is the case, leakage of Ca²⁺ into the cell, determined by adding 2 mM Ca²⁺ to untreated cells after 5-min incubation in Ca²⁺-free Krebs' buffer, was subtracted from each of the traces corresponding to the conditions described above. The traces corresponding to the single pathways of Ca²⁺ entry, triggered in two separate populations of cells, were added (Fig. 7A, inset). The resulting trace superimposed closely with the trace corresponding to the double treatment with TG and OAG of one cell population. Therefore, the [Ca²⁺]_i increases resulting from OAG-activated Ca²⁺ entry and CCE are additive.

The effects of CCE and OAG-activated Ca²⁺ entry on AC8 activity were examined separately and in combination (Fig. 7B). As observed previously (Fig. 5A), OAG-activated Ca²⁺ entry had little effect, whereas CCE caused a robust increase in AC8 activity. Treatment with both TG (0.1 μM, 4 min) and OAG (30 μM, 1 min) to trigger both pathways at once upon the addition of extracellular Ca²⁺, stimulated AC8 activity to the same degree as CCE alone. These data show that OAG-activated Ca²⁺ entry does not interfere with, or add to, the activation of AC8 by CCE.

Discussion

In nonexcitable cells, it is well-established that AC8 is preferentially activated by CCE and is unaffected by alternative forms of increase in [Ca²⁺]_i, such as release from intracellular stores or ionophore-mediated Ca²⁺ influx (Fagan et al., 1996, 1998). This property of AC8 has been successfully used to discriminate between capacitative and arachidonate-activated Ca²⁺ entry pathways (Shuttleworth and Thompson, 1999). Although the molecular identity of CCE channels remains unknown, likely candidates include the TRPC family of TRP channels, a subset of which can be activated by OAG. In this study, we explored the possibility that these OAG-activated TRP subunits may form part of the CCE apparatus by using AC8 as a sensor of CCE to determine whether OAG activates the same subset of Ca²⁺ channels as store depletion.

OAG promotes the influx of Ca²⁺ across the plasma membrane (Fig. 2), which is consistent with the presence of endogenous TRPC3, TRPC6 (Garcia and Schilling, 1997), and TRPC7 (Zagranichnaya et al., 2005) in HEK 293 cells. Furthermore, suppression of these TRPs by small interfering RNA reduces OAG-activated Ca²⁺ entry (Zagranichnaya et al., 2005). Thus, it is likely that these OAG-activated TRP subunits are responsible for mediating the observed OAG-activated Ca²⁺ response. This conclusion is reinforced by the fact that OAG-activated Ca²⁺ entry is enhanced when PKC is inhibited by staurosporine (Fig. 1), which is consistent with the reported role of PKC in inhibiting TRPC3 activity in HEK 293 cells (Venkatachalam et al., 2003). Therefore, PKC is not responsible for mediating the response and, on the contrary, has an inhibitory effect, such that the activation of Ca²⁺ entry in HEK 293 cells by OAG is likely to result from the direct activation of TRP subunits by OAG (Hofmann et al., 1999).

Under the chosen experimental conditions, not only was the overall magnitude of OAG-activated Ca²⁺ entry greater than that of CCE, but the rate of Ca²⁺ entry via the OAG-activated pathway was faster than via the capacitative pathway (Fig. 3). The two pathways showed different sensitivities to low concentrations of Gd³⁺, which inhibited CCE more potently than OAG-activated Ca²⁺ entry (Fig. 4). In other cell lines, such as T-lymphocytes (Gamberucci et al., 2002) and salivary gland HSY cells (Liu et al., 2005), Gd³⁺ did not discriminate between the two pathways. Conceivably, the relative magnitude of the two pathways may affect the potency of Gd³⁺.

The preferential activation of AC8 by CCE was used to determine whether OAG activates the same subset of Ca²⁺ channels as store depletion. CCE robustly activated AC8, whereas OAG-activated Ca²⁺ entry had no statistically significant effect (Fig. 3). We excluded the possibilities that OAG-mediated activation of PKC might inhibit AC8 and mask a possible activation of the enzyme by OAG-activated Ca²⁺ entry (Fig. 3A) and that OAG, which is membranepermeant, might directly disrupt AC8 activity (Fig. 3B). Thus, AC8 can discriminate between CCE and OAG-activated Ca²⁺ entry, which indicates that the two types of Ca²⁺ entry are mediated by distinct Ca²⁺ channels.

The contrasting effects of CCE and OAG-activated Ca²⁺ entry on AC8 activity led us to explore the profiles of the two pathways in single HEK 293 cells. Whereas CCE caused a sustained increase in [Ca²⁺]_i, which occurred immediately in all cells, OAG caused a more heterogeneous response in terms of delay, amplitude, and duration. This profile is consistent with that reported in single astrocyte cells, in which OAG evoked low-frequency, high-amplitude [Ca²⁺]_i oscillations after a brief delay (Grimaldi et al., 2003). In the individual cells, the large increases in [Ca²⁺]_i evoked by OAG may conceivably transiently activate AC8. This may be significant in a single cell but not in a population of cells over the short duration of the experiment. However, even extended periods after pretreatment with OAG failed to reveal similar effects of the two modes of Ca²⁺ entry.

The selective activation of AC8 by CCE is postulated to arise from a very close apposition of the enzyme and the CCE channel (Fagan et al., 1996, 1998). In addition, the localization of AC8 in cholesterol-rich caveolae is necessary for CCE regulation of AC8 (Fagan et al., 2000; Smith et al., 2002). It is possible that the channels mediating OAG-activated Ca²⁺ entry may reside in different compartments, such that the increase in [Ca²⁺]_i caused by OAG does not occur in the vicinity of AC8. Even though both AC8 and TRPC3 have been reported to reside in lipid rafts (Fagan et al., 2000; Lockwich et al., 2001), these proteins may well segregate to separate rafts.

In HEK 293 cells, a study using small interfering RNA techniques showed that the same endogenously expressed OAG-activated TRP subunits assemble with different partners to form the complexes mediating OAG-activated Ca²⁺ entry and CCE. Thus, whereas TRPC1, TRPC3, and TRPC7 were found to coassemble to mediate CCE, a complex including TRPC3, TRPC4, TRPC6, and TRPC7 was responsible for the response to OAG (Zagranichnaya et al., 2005). Because our results indicate that the complexes mediating CCE and OAG-activated Ca²⁺ entry are distinct, they suggest that the TRP subunits, when part of the complex mediating CCE, are not responsive to OAG. Hence, the subunits per se are not activated by OAG; however, in combination with specific partners, they assemble into an OAG-activated channel. If TRP subunits are recruited by a factor such as STIM1 (Zhang et al., 2005) upon store-depletion to form CCE channels, this may account for the smooth, homogeneous response observed in all cells. In contrast, if OAG does not cause channel recruitment, this may account for the dramatic cell-to-cell variation. Moreover, the preferential activation of AC8 by CCE could be explained if the combination of subunits which assembles to form the CCE apparatus but not the one that mediates OAG-activated Ca²⁺ entry includes AC8 as an integral part of the complex.

The pathways mediating CCE and OAG-activated Ca²⁺ entry function independently of one another, because they can be triggered simultaneously in the same population of cells to give rise to an influx of Ca ²⁺, which is equivalent to the sum of the influxes generated through each individual pathway (Fig. 7). In this respect, the OAG-activated pathway differs from the arachidonic-activated pathway, which was found to potently inhibit CCE in HEK 293 cells and vice versa (Luo et al., 2001a). This observation reinforces the fact that, although diacylglycerol can be metabolized to arachidonic acid, the pathways activated by each of these messengers are distinct. However, the relationship between the OAG-activated and CCE pathways seems to be cell type-specific. For example, the properties of these two pathways in T-lymphocytes are similar to those in HEK 293 cells; the effects of TG and OAG are additive, and TRPC6 has been implicated in mediating OAG-activated Ca²⁺ entry (Gamberucci et al., 2002). In contrast, a recent study in human parotid gland ductal (HSY) cells found that the effects of TG and OAG were nonadditive (Liu et al., 2005). Moreover, antisense and coimmunoprecipitation experiments implicated TRPC1 and TRPC3 in mediating OAG-activated Ca²⁺ entry in these cells. Hence, the relationship between CCE and OAG-mediated Ca²⁺ entry may well depend on the types of TRP channels expressed, which in turn dictates the possible combinations of TRP subunits.

In conclusion, our data show that OAG-activated Ca²⁺ entry is an independent pathway of Ca²⁺ entry in HEK 293 cells, which is mediated by channels that are distinct from the elusive CCE channels, even though they may share common TRP subunits. The inability of OAG-activated Ca²⁺ entry to regulate AC8 further reinforces the absolute dependence of this enzyme on CCE.