Report

Requirement of the Inositol Trisphosphate Receptor for Activation of Store-Operated Ca2+ Channels

See allHide authors and affiliations

Science  03 Mar 2000:
Vol. 287, Issue 5458, pp. 1647-1651
DOI: 10.1126/science.287.5458.1647

Abstract

The coupling mechanism between endoplasmic reticulum (ER) calcium ion (Ca2+) stores and plasma membrane (PM) store-operated channels (SOCs) is crucial to Ca2+ signaling but has eluded detection. SOCs may be functionally related to the TRP family of receptor-operated channels. Direct comparison of endogenous SOCs with stably expressed TRP3 channels in human embryonic kidney (HEK293) cells revealed that TRP3 channels differ in being store independent. However, condensed cortical F-actin prevented activation of both SOC and TRP3 channels, which suggests that ER-PM interactions underlie coupling of both channels. A cell-permeant inhibitor of inositol trisphosphate receptor (InsP3R) function, 2-aminoethoxydiphenyl borate, prevented both receptor-induced TRP3 activation and store-induced SOC activation. It is concluded that InsP3Rs mediate both SOC and TRP channel opening and that the InsP3R is essential for maintaining coupling between store emptying and physiological activation of SOCs.

Receptor-induced Ca2+signals comprise two interdependent components—rapid Ca2+release from Ca2+ stores in the ER and Ca2+entry through slowly activating PM SOCs. The trigger for SOC activation is decreased Ca2+ in the ER lumen (1, 2). However, despite intense study, the ER-derived signal coupling store depletion with SOC activation remains unknown (3). Direct coupling between ER and PM has been hypothesized (4, 5), and evidence indicates that physical docking of ER with the PM is involved in SOC activation (6–8). The mammalian TRP family of receptor-operated ion channels has been suggested to share some operational parameters with SOCs (9). Kiselyov et al. (10, 11) provided evidence that human TRP3 channel activation results from interaction with InsP3Rs. However, other evidence indicates that diacylglycerol (DAG), not InsP3, is the phospholipase C (PLC) product that mediates activation of TRP3 channels (12) and that TRP3 channels operate independently of stores (12–15). We show here that physical interaction between ER and PM is necessary for activation of both TRP and SOC, and we provide new evidence that the InsP3R is an essential component for mediating and maintaining coupling between store emptying and physiological activation of SOCs.

To directly compare the function of endogenous SOCs and TRP3 channels, we used T3-65 clonal human embryonic kidney (HEK293) cells stably transfected to express the human TRP3 (hTRP3) channel (14). Both channels function in T3-65 cells, whereas only SOCs function in the control-transfected clonal C1 line. We used Sr2+ entry to identify and distinguish SOC and TRP3 channel function. After release of Ca2+ from stores with the Ca2+ pump blocker thapsigargin (TG) in the absence of external Ca2+, Sr2+ addition resulted in little Sr2+ entry in either C1 or T3-65 cells (Fig. 1, A and B). Ca2+addition resulted in Ca2+ entry in both cell lines, which is typical of SOC activation (16). The ionophore ionomycin induced more rapid and complete release of stored Ca2+ and again little Sr2+ entry in either cell type (Fig. 1, C and D). Addition of adenosine triphosphate (ATP) to stimulate PLC-coupled purinergic receptors caused entry of Sr2+ in T3-65 cells but not in C1 cells, revealing TRP3 channel function in T3-65 cells. Thus, the TRP3 channel is clearly not activated by store emptying alone, with either TG or ionomycin; instead, it appears to be activated by a product of receptor-induced PLC. Redistribution of actin (caused, for example, by phosphatase inhibitors) to form a dense cortical layer beneath the PM displaced cortical ER and prevented SOC activation by store emptying, providing evidence that ER-PM interactions are required for SOC activation (6). In T3-65 cells, the slowly developing SOC-mediated Ca2+ entry in response to store emptying (Fig. 1E) was substantially reduced after treatment with the phosphatase inhibitor calyculin A (calyA) (Fig. 1G); this action correlated with the formation of a dense cortical actin layer (Fig. 1, F and H). If TRP3 were directly activated by receptors in the PM, we reasoned that TRP3 activation might be insensitive to calyA. However, whereas calyA had no effect on receptor-induced (InsP3-mediated) Ca2+ release, it completely blocked receptor-mediated TRP3 activation in T3-65 cells (Fig. 1, I to K), suggesting that TRP activation, as for SOCs, involves coupling between the PM and another organelle.

Figure 1

Correlation of ER-PM interactions with activation of SOC and TRP3 channels. (A to D) Comparison of Sr2+ and Ca2+ entry in fura-2-loaded clonal T3-65 HEK293 cells stably transfected with the hTRP3 gene and control-transfected clonal C1 cells (14). Cytosolic Sr2+ and Ca2+ was measured by ratiometric (F340/F380) fluorescence of groups of 5 to 10 fura-2-loaded cells as described (23). (A and B) TG (1 μM) (arrows) was added to C1 or T3-65 cells in the absence of Ca2+ followed by addition of 1 mM Sr2+. Medium was replaced with divalent cation-free medium, to which 1 mM Ca2+ was then added (bars). (C and D) Cells were treated with 5 μM ionomycin (Ion) (arrows) in the absence of Ca2+; 1 mM external Sr2+ (bars) and 100 μM ATP (arrows) were then added to the external medium. (Eto H) Effects of 50 nM calyA treatment (45 min, 22°C) on TG-induced Ca2+ responses and F-actin distribution in T3-65 cells. Cytosolic Ca2+ was measured in response to 1 μM TG (arrows) in control (E) or calyA-treated (G) T3-65 cells with or without external Ca2+ (bars). Confocal images of F-actin labeled with fluorescein isothiocyanate–conjugated phalloidin in control (F) or calyA-treated (H) T3-65 cells. Imaging and F-actin analysis were as described (6). (I toK) Cytosolic fura-2 ratiometric measurements in Ca2+-free medium in response to 100 μM ATP (arrows) in the presence (bars) or absence of 1 mM external Sr2+ in C1 cells (I), control T3-65 cells (J), or T3-65 cells treated with calyA (K) as above. TG, ionomycin, and ATP were maintained after addition.

TRP3-mediated Sr2+ entry was increased by the membrane permeable DAG analog 1-oleoyl-2-acetyl-sn-glycerol (OAG) (Fig. 2A) and increased less effectively by another permeable analog, 1,2-dioctanoyl-sn-glycerol (DOG) (Fig. 2B). The DAG lipase inhibitor RHC80267 (RHC) also induced Sr2+ entry (Fig. 2C). These results are consistent with activation of TRP channels by DAG (12). They appear to be a specific reflection of TRP3 channels because there was no action of any agent with C1 cells (Fig. 2, A to C). We obtained similar results for TRP-mediated Ca2+ entry (Fig. 2, D to F), indicating that DAG is not activating SOC. Even after agonist, OAG induced no Sr2+ entry (Fig. 2G), again suggesting no endogenous TRP activity in C1 cells. After ATP-induced Sr2+ entry through TRP3 channels in T3-65 cells, the addition of OAG induced a small further increase in Sr2+ entry (Fig. 2H). We compared the action of calyA on both agonist- and DAG-induced TRP3 activation. Whereas calyA treatment eliminated agonist-induced Sr2+ entry, the action of OAG was unaffected (Fig. 2I). This indicates that ER-PM interactions mediate only agonist-induced TRP3 activation, whereas DAG may activate TRP3 directly. The inability of agonist to stimulate TRP3 in calyA-treated cells suggests that DAG does not contribute to PLC-coupled receptor-induced TRP3 activation, which implies that receptors activate TRP3 only through InsP3.

Figure 2

Activation of TRP3 channels by two distinct mechanisms. (A to C) Cytosolic fura-2 ratiometric measurements in T3-65 cells or C1 cells in response to 100 μM OAG (A), 100 mM DOG (B), or 50 μM RHC (C) in the presence of 1 mM external Sr2+ (bar). (D to F) Cytosolic fura-2 measurements in T3-65 cells or C1 cells in response to 100 μM OAG (D), 100 μM DOG (E), or 50 μM RHC (F) with 1 mM external Ca2+ (bar). (G to I) Cytosolic fura-2 measurements in Ca2+-free medium in response to 100 μM ATP (arrows) and 100 μM OAG (arrows) in the presence (bars) or absence of 1 mM external Sr2+ in C1 cells (G), control T3-65 cells (H), or T3-65 cells treated with calyA (I) as above. ATP and DAG analogs were maintained after addition.

Much evidence indicates that TRP3 channels operate independently of Ca2+ stores (12–15) (Fig. 1). We therefore questioned how related SOC and TRP really are and whether InsP3Rs play any role in their activation. The membrane-permeant InsP3R antagonist 2-aminoethoxydiphenyl borate (2-APB) (17) proved to be a remarkably effective probe for assessing InsP3R involvement in situ. 2-APB at 75 μM blocked receptor-induced store emptying in intact HEK293 cells and all other cells tested, regardless of PLC-coupled agonist (Fig. 3, A to E). 2-APB had no effect on basal cytosolic Ca2+ or on the size of ionophore-releasable stores in intact cells, which indicates no change in Ca2+ homeostasis. In broken cells, 2-APB directly blocks InsP3R-mediated Ca2+ release from ER (18). 2-APB has no effect on InsP3 binding, does not alter InsP3 production through agonist-sensitive PLC, and does not modify the function of ryanodine receptors or voltage-operated Ca2+ channels (17).

Figure 3

Inhibition by 2-APB of TRP3 channel activation through PLC-coupled receptors. (A to E) 2-APB (75 μM) prevented PLC-coupled receptor-induced Ca2+ release in response to 100 μM ATP and 10 μM CCh in C1 cells (A and B), 1 nM vasopressin (VP) or 1 μM serotonin (5-HT) in A7r5 cells (C and D) or 10 μM bradykinin (BK) in DDT1MF2 cells (E). (F to I) Ca2+ release and TRP3-mediated Sr2+ entry activated by ATP were blocked by 2-APB in T3-65 cells; DAG-induced TRP3 activation was unaffected. (F and G) Fura-2 responses to 100 μM ATP (arrows) and 100 μM OAG (arrows) in the presence or absence of 1 mM external Sr2+ with (F) or without (G) 75 μM 2-APB. (H and I) Effects of 100 μM ATP (arrows) and 100 μM OAG (arrows) on Sr2+ entry after store depletion with 5 μM ionomycin (Ion) (arrows). (J) Fura-2 measurements in response to 100 μM ATP (arrow) in the presence or absence of 1 mM external Sr2+ in T3-65 cells pretreated with 20 μM xestC for 20 min and maintained until removal. (Kto M) Effect of transient 2-APB (75 μM) addition on ATP-induced Sr2+ entry in T3-65 cells either without ionomycin (K and M) or after Ca2+ store release with 5 μM ionomycin (L). (N) Concentration dependence of 2-APB–induced inhibition of ATP-mediated Ca2+ release and TRP3-mediated Sr2+ entry in T3-65 cells. Both activities were measured consecutively on one coverslip after addition of 2-APB 90 s before addition of 100 μM ATP (in Ca2+-free medium) followed by addition of Sr2+ 3.5 min later. Results are single measurements made on a series of coverslips. A7r5 and DDT1MF2 smooth muscle cells were used as described (6). 2-APB synthesis was described (17) and xestC was from I. Pessah (University of California, Davis).

In T3-65 cells, ATP-induced Ca2+ release and subsequent TRP3-mediated Sr2+ entry were almost completely inhibited by 75 μM 2-APB (Fig. 3, F and G). However, the action of OAG was unaffected. To test whether the block of ATP-induced store emptying might have prevented TRP3 activation, we conducted the same experiment after complete store emptying with ionomycin and obtained similar results (Fig. 3, H and I). Thus, the InsP3R antagonist blocked receptor-induced TRP3 activation. The lack of effect of 2-APB on DAG-induced TRP3 activation confirms the distinction between TRP3 activation by DAG and agonist and indicates that 2-APB does not block the TRP3 channel per se.

Another permeant InsP3R antagonist, xestospongin C (xestC) (19), induced similar effects. However, the action of xestC was very slow, requiring 20 min at 20 μM to prevent TRP activation, and only partly prevented agonist-induced Ca2+ release (Fig. 3J). This likely reflects a latency of blockade of InsP3Rs deeper within cells (20). Upon removal of xestC, there was no reversal of inhibition (Fig. 3J). In contrast, 2-APB rapidly prevented Ca2+ release and TRP-mediated Sr2+ entry and was fully reversible. Thus, removal of 2-APB resulted in immediate return of entry (Fig. 3K).

We also conducted this experiment after store depletion with ionomycin (to eliminate Ca2+ release through reactivated InsP3Rs), revealing the return of TRP-mediated Sr2+ entry alone after 2-APB removal (Fig. 3L). Prior application and removal of 2-APB resulted in the return of full agonist-induced store release and Sr2+ entry (Fig. 3M). The 2-APB dose-response curve for TRP-mediated Sr2+ entry was close to that for receptor-induced Ca2+ store release (Fig. 3N), with median inhibitory concentration (IC50) values of 10 and 25 μM, respectively. The small difference in IC50 again may reflect a latency of action of 2-APB on more remote InsP3Rs mediating Ca2+ release as opposed to those coupled to PM TRP3 channels.

The results with calyA and 2-APB provide good evidence that ER InsP3Rs are required for coupling PLC activation to TRP3 channel opening. Considering the important differences between TRP and SOC function described here and by others (12–15), a crucial question was whether SOC activation has any similar InsP3R requirement. SOC was measured after TG-induced store emptying in C1 cells (Fig. 4A); removal and readdition of external Ca2+ revealed the familiar overshoot response due to reactivation of SOCs (6,16). The store-release component was observed in the absence of external Ca2+, and later addition of Ca2+ again resulted in full SOC-mediated Ca2+entry (Fig. 4B). Addition of 2-APB just before addition of TG resulted in complete inhibition of SOC activation; even the response to Ca2+ removal and readdition was eliminated (Fig. 4C). 2-APB did not prevent Ca2+ release with TG (21). Prolonged incubation with xestC reduced but did not eliminate TG-induced SOC activation (Fig. 4D). The action of both 2-APB and xestC on SOC indicates that InsP3Rs are required for SOC activation. The dose-response relationship for 2-APB on SOC blockade was close to that for TRP blockade (IC50 10 to 15 μM). After complete ionomycin-induced store emptying, 2-APB eliminated Ca2+ entry (Fig. 4E).

Figure 4

Inhibition and reversal of SOC opening by InsP3R inhibitors. (A and B) Ca2+ release and entry in control-transfected C1 cells activated by 1 μM TG in the presence or absence of external Ca2+. (C) Ca2+ entry after 1 μM TG in the presence of 1 mM Ca2+ and after subsequent removal and addition of Ca2+ was blocked by 75 μM 2-APB. (D) TG-induced Ca2+ entry as for (C) was partly blocked after xestC treatment (20 min, 20 μM). (E) SOC-mediated Ca2+ entry after Ca2+ release with 5 μM ionomycin was blocked by 75 μM 2-APB; addition of 100 μM CCh with continued 2-APB was without effect. (F) Sustained SOC-mediated Ca2+ entry after 5 μM ionomycin-induced store release was rapidly terminated by addition of 75 μM 2-APB. (G and H) Comparison of 2-APB–induced termination of sustained SOC-mediated Ca2+ entry in C1 cells (G) and sustained TRP3-mediated Sr2+ entry in T3-65 cells (H). Results in (G) were from an experiment identical to that shown in (F) and mean lag time was 12.4 ± 0.6 s (n = 4). Results in (H) were for Sr2+ entry in T3-65 cells activated by 100 μM ATP with addition of 75 μM 2-APB; lag-time was <3 s. (I) Rapidity of action of 2-APB to inhibit ATP-induced Ca2+store release in C1 cells: 100 μM ATP was added either alone (left), simultaneously with 75 μM 2-APB (middle), or 20 s after addition of 75 μM 2-APB (right). (J and K) Receptor-induced InsP3 production enhanced reversal of 2-APB–induced blockade of SOC activated by ionomycin in C1 cells. (J) 2-APB (75 μM) was added (bar) after store release with 5 μM ionomycin, and Ca2+ was added and 2-APB was removed. (K) Identical to (J) except 100 μM CCh was added and 2-APB was removed.

Addition of 2-APB while SOC-mediated entry was maximally active resulted in rapid termination of SOC (Fig. 4F). This indicates that the InsP3R is required for maintenance as well as activation of SOCs. Analysis of the 2-APB–induced reversal of SOC revealed a 12-s delay (Fig. 4G), whereas 2-APB–induced reversal of TRP3-mediated Sr2+ entry had no detectable lag (Fig. 4H). The rapid action of 2-APB caused us again to question whether the action of 2-APB on Ca2+ entry was InsP3R related. 2-APB blocked intracellular InsP3Rs extremely rapidly (Fig. 4I); added with agonist, InsP3R-mediated Ca2+release from stores was reduced 50%; added 20 s before agonist release, it was reduced 80%. Thus, 2-APB can reach and block function of intracellular InsP3Rs with sufficient rapidity to account for its blockade of SOC activation. After 2-APB–induced SOC blockade, removal of 2-APB resulted in a slow return of SOC activity over several minutes (Fig. 4J). PLC activation with carbachol (CCh) to produce a high level of cytosolic InsP3 resulted in larger and faster reappearance of Ca2+ entry (Fig. 4K), whereas CCh induced no return of Ca2+ entry if 2-APB remained present (Fig. 4E). This supports the conclusion that the activation and return of SOC function is mediated through InsP3R activation.

Our results support a central role of the InsP3R in mediating Ca2+ entry. The results also highlight differences in the operation of the two entry channels. TRP3 is activated rapidly in response to PLC-coupled receptors or DAG but is not directly activated by store emptying. SOC activity develops relatively slowly in response to store emptying (6, 22) but is not activated by DAG. Despite these differences, both channels can be uncoupled by calyA-induced cortical actin rearrangement, which suggests that physiological activation in each case requires ER-PM interactions. Receptor-induced activation of the TRP3 channel appears to be mediated by the InsP3R. Furthermore, the elusive coupling of the physiologically important endogenous SOC itself appears to be InsP3R-mediated. Thus, SOC and TRP can both be considered InsP3R-mediated channels, a conclusion highly compatible with the evidence here and previously (6) for the role of ER-PM interactions in SOC activation. We suggested that store emptying promotes reversible docking of ER with PM to activate SOC (6, 7). Yao et al. (8) considered that a similar docking process might activate Ca2+ entry through insertion of SOCs into the PM. Our results reveal that InsP3R blockade rapidly deactivates already activated SOCs, suggesting that channel activation is maintained by continued close contact with InsP3Rs and militating against an insertion model. Instead, it provides strong evidence in favor of a conformational coupling model for SOC (4,5). The latency in SOC activation, together with our recent structural and functional evidence (6), suggests that SOC activation requires that InsP3Rs on the ER membrane are moved to the vicinity of SOCs and that this trafficking is the basis of activation by store emptying.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: dgill{at}umaryland.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article