Amplification of receptor signalling by Ca²⁺ entry-mediated translocation and activation of PLCγ2 in B lymphocytes

Abstract

In non-excitable cells, receptor-activated Ca²⁺ signalling comprises initial transient responses followed by a Ca²⁺ entry-dependent sustained and/or oscillatory phase. Here, we describe the molecular mechanism underlying the second phase linked to signal amplification. An in vivo inositol 1,4,5-trisphosphate (IP₃) sensor revealed that in B lymphocytes, receptor-activated and store-operated Ca²⁺ entry greatly enhanced IP₃ production, which terminated in phospholipase Cγ2 (PLCγ2)-deficient cells. Association between receptor-activated TRPC3 Ca²⁺ channels and PLCγ2, which cooperate in potentiating Ca²⁺ responses, was demonstrated by co-immunoprecipitation. PLCγ2-deficient cells displayed diminished Ca²⁺ entry-induced Ca²⁺ responses. However, this defect was canceled by suppressing IP₃-induced Ca²⁺ release, implying that IP₃ and IP₃ receptors mediate the second Ca²⁺ phase. Furthermore, confocal visualization of PLCγ2 mutants demonstrated that Ca²⁺ entry evoked a C2 domain-mediated PLCγ2 translocation towards the plasma membrane in a lipase-independent manner to activate PLCγ2. Strikingly, Ca²⁺ entry-activated PLCγ2 maintained Ca²⁺ oscillation and extracellular signal-regulated kinase activation downstream of protein kinase C. We suggest that coupling of Ca²⁺ entry with PLCγ2 translocation and activation controls the amplification and co-ordination of receptor signalling.

Introduction

In non-excitable cells, stimulation of surface membrane receptors induces Ca²⁺ signals via activation of phospholipase C (PLC), which hydrolyses phosphatidylinositol 4,5-bisphosphate (PIP₂) into inositol 1,4,5-trisphosphate (IP₃) and diacylglycerol (DG) (Berridge, 1993; Clapham, 1995). IP₃ triggers rapid Ca²⁺ release from internal Ca²⁺ stores of the endoplasmic reticulum (ER) by activating IP₃ receptors (IP₃R) and a consequent transient increase in cytosolic Ca²⁺ concentration ([Ca²⁺]_i). The initial phase of receptor-activated Ca²⁺ signalling is followed by sustained and/or oscillatory [Ca²⁺]_i changes. The molecular mechanism underlying the second phase is still elusive (Feske et al., 2001; Mori et al., 2002), despite the vital role played by this phase in various biological responses (Miyazaki et al., 1992; Thorn et al., 1993; Dolmetsch et al., 1998; Li et al., 1998).

The second phase of Ca²⁺ signalling is critically controlled by receptor-activated Ca²⁺ entry processes such as capacitative Ca²⁺ entry (CCE) mediated by store-operated Ca²⁺ channels (SOCs) coupled to IP₃-induced Ca²⁺ release or consequent depletion of Ca²⁺ stores (Putney and Bird, 1993). In electrophysiological measurement, SOCs and other Ca²⁺ channels activated by DG in a membrane-delimited and protein kinase C (PKC)-independent manner have been identified (Hofmann et al., 1999; Okada et al., 1999; Inoue et al., 2001; Gamberucci et al., 2002). In [Ca²⁺]_i measurement, it has been postulated that CCE is distinguished as an ‘off’ response elicited by readministration of extracellular Ca²⁺ to cells after pretreatment with store-depleting agents such as thapsigargin (TG) in the absence of Ca²⁺ (Parekh et al., 1997). Importantly, there is a large body of evidence that the properties of IP₃R Ca²⁺ release channels affect the temporal patterns of Ca²⁺ signals/oscillations (Tsien and Tsien, 1990; Miyazaki et al., 1992; Miyakawa et al., 1999). Thus, together with direct elevation of [Ca²⁺]_i by the Ca²⁺ entered through SOCs, the ‘capacitative’ function of CCE, which refills Ca²⁺ stores to maintain Ca²⁺ release, is a major role played by Ca²⁺ entry in eliciting the second phase.

In receptor-induced Ca²⁺ signalling, activation of PLC isoforms is controlled through distinct coupling mechanisms (Rhee, 2001). It has been commonly understood that receptor and non-receptor tyrosine kinases are responsible for activation of PLCγs. Recent evidence has added the novel view that PLCγs have multifunctional properties, and directly interact with numerous target proteins (Smith et al., 1994; Kim et al., 2000; Rebecchi and Pentyala, 2000; Rhee, 2001; Patterson et al., 2002; Putney, 2002; Runnels et al., 2002; Ye et al., 2002; Trebak et al., 2003). Here, we reveal that PLCγ2, beyond its role in initiating the Ca²⁺ response upon stimulation of tyrosine kinase-coupled receptors, mediates the second phase of Ca²⁺ signals in B lymphocytes. B cell receptor (BCR)-induced and store-operated Ca²⁺ entry elicited the translocation of PLCγ2 towards the plasma membrane in a lipase-independent manner, and enhanced PLCγ2 activation to maintain IP₃ production and IP₃-induced Ca²⁺ release. We suggest that Ca²⁺ entry is coupled with PLCγ2 translocation and activation to control physiological co-ordination and amplification of receptor-mediated signals.

Results

Activation of PLCγ2 by BCR-induced Ca²⁺ entry

In chicken DT40 B lymphocytes, ligation of BCR with anti-IgM antibody (anti-IgM) induced rapid rises in [Ca²⁺]_i followed by sustained elevation of [Ca²⁺]_i in the presence of extracellular Ca²⁺ (Figure 1A). Since the sustained phase was eliminated by omitting extracellular Ca²⁺, the sustained phase must derive from Ca²⁺ entry. Real time changes in [IP₃]_i were assessed in individual cells, using a new IP₃-selective sensor R9-PHIP₃-D106 constructed based on the pleckstrin homology (PH) domain of PLCδ1 (Morii et al., 2002). Stimulation of BCR triggered a rapid decrease in R9-PHIP₃-D106 fluorescence, indicative of an initial [IP₃]_i increase in the wild-type (WT) cells (Figure 1B). The increase in [IP₃]_i was sustained or slowly enhanced for >5 min in the presence of extracellular Ca²⁺, but was transient when Ca²⁺ was removed. Population IP₃ measurement confirmed the observation (Supplementary figure 1C, available at The EMBO Journal Online). Since PLCγ2 mediates BCR-induced Ca²⁺ signalling (Takata et al., 1995), the PLCγ2-deficient (PLCγ2^–) DT40 cell line was employed to examine whether the observed IP₃ production is in fact due to PLCγ2 activation. The BCR-induced [IP₃]_i and [Ca²⁺]_i rises terminated in the PLCγ2^– cells (Figure 1C and D). The defects were resolved by heterologous expression of rat PLCγ2 in the PLCγ2^– cells, but not by a lipase-dead PLCγ2 mutant (LD; see also Figure 4A).

Fig. 1. Extracellular Ca²⁺ elicits sustained PLCγ2 activation and receptor-evoked [Ca²⁺]_i mobilization. (A) Left, representative time courses of Ca²⁺ responses in DT40 cells upon BCR stimulation with anti-IgM (1 µg/ml) in 2 mM Ca²⁺-containing external (n = 32 cells) or EGTA-containing, Ca²⁺-free (n = 32) solution. Right, peak BCR-induced [Ca²⁺]_i rises and [Ca²⁺]_i increases sustained after 5 min stimulation. (B) Left, representative time courses of BCR-induced changes in florescence intensities of IP₃ indicator R9-PHIP3-D106 (n = 12–15). F is the fluorescence intensity and F₀ is the initial F. Right, BCR-induced [IP₃]_i changes at peak and sustained after 5 min. (C and D) BCR-induced [Ca²⁺]_i rises (C) and [IP₃]_i rises (D) at peak and sustained after 6 min stimulation in WT cells expressing GFP (GFP/WT), and in PLCγ2^– cells expressing GFP (GFP/PLCγ2^–), PLCγ2 (PLCγ2/PLCγ2^–) or LD mutant (LD/PLCγ2^–) (n = 22–60). Ca²⁺ is present in external solution. (E) PLCγ2 enhances Ca²⁺ responses induced by Ca²⁺ entry upon M1R stimulation. Ca²⁺ release was first evoked in Ca²⁺-free solution, and Ca²⁺ entry-induced Ca²⁺ responses were induced by readministration of 2 mM Ca²⁺ in WT cells expressing GFP and in PLCγ2^– cells expressing GFP, PLCγ2, LD or ΔSH3 (n = 5–13). Left, average time courses. Right, peak [Ca²⁺]_i rises in Ca²⁺-free solution and after Ca²⁺ readministration. Significance difference from control: *P < 0.05.

Fig. 4. The C2 domain mediates Ca²⁺ entry-induced activation of PLCγ2. (A) Schematic diagram for various mutant constructs of PLCγ2. (B) Peak BCR-induced [Ca²⁺]_i rises during 5 min incubation in Ca²⁺-free solution and after subsequent readmission of Ca²⁺ in PLCγ2^– cells expressing respective PLCγ2 mutants (n = 11–36). (C) Peak TG-induced [Ca²⁺]_i rises in Ca²⁺-free solution and after readmission of Ca²⁺ in PLCγ2^– cells expressing respective PLCγ2 mutants (n = 12–37). Protocol is the same as in Figure 2E. (D) Left, average time courses of BCR-induced Ca²⁺ responses in PLCγ2^– cells expressing PLCγ2, membrane-bound PLCγ2 chimera (mPLCγ2) or GFP (n = 8–14). mPLCγ2 is composed of the human CD16 extracellular domain, the human TCR ζ-chain transmembrane domain and the complete rat PLCγ2 as a cytoplasmic domain (Ishiai et al., 1999). Right, BCR- induced [Ca²⁺]_i rises at peak and sustained after 5 min. (E) Left, average time courses of BCR-induced F/F₀ changes of R9-PHIP3-D106. Right, maximal [IP₃]_i elevation and sustained [IP₃]_i rises after 5 min upon BCR stimulation.

Ca²⁺ entry induced by G protein-coupled receptor stimulation activates PLCγ2

To characterize the extracellular Ca²⁺-dependent sustained phase separately from the initial BCR-evoked phase in PLCγ2 activation, heterologously expressed M1 muscarinic acetylcholine receptor (M1R) (Sugawara et al., 1997) was stimulated with carbachol (CCh). This is because the different isoform PLCβ mediates PIP₂ hydrolysis to trigger Ca²⁺ release upon M1R and G_q protein activation (Parekh et al., 1997). In the WT and PLCγ2^– cells, stimulation of M5R triggers Ca²⁺ mobilization via G_q and PLCβ (Patterson et al., 2002). Stimulation of M1R induced smaller [Ca²⁺]_i increases in the PLCγ2^– cells compared with those in the WT cells (Supplementary figure 2). However, omission of extracellular Ca²⁺ abolished the difference between the WT and PLCγ2^– cells in M1R-activated Ca²⁺ responses (Figure 1E). The data support the view that PLCγ2 is responsible for external Ca²⁺-dependent enhancement of Ca²⁺ responses, and that Ca²⁺ entry induced by PLCβ-coupled receptor leads to late sustained PLCγ2 activation. Furthermore, the defect in Ca²⁺ entry-induced Ca²⁺ response in PLCγ2^– cells was partially restored by the LD mutant in the PLCγ2^– cells (Figure 1E), suggesting the lipase-independent (Patterson et al., 2002) and lipase-dependent roles of PLCγ2 in receptor-activated Ca²⁺ response.

Activation of PLCγ2 by store-operated Ca²⁺ influx

To analyse further the Ca²⁺ entry-induced mechanism in BCR-induced PLCγ2 activation, we used SERCA Ca²⁺-ATPase inhibitor TG, which activates SOCs via passive depletion of Ca²⁺ stores. In the presence of extracellular Ca²⁺, TG produced dramatic [Ca²⁺]_i increases in the WT cells, but only moderate [Ca²⁺]_i increases in the PLCγ2^– cells (Figure 2A and B). In the absence of Ca²⁺, TG caused indistinguishable transient Ca²⁺ rises in the WT and PLCγ2^– cells, indicating that TG-sensitive Ca²⁺ stores are not affected by the loss of PLCγ2. As shown in Figure 2C, treating the WT cells with TG induced marked [IP₃]_i rises in the presence of extracellular Ca²⁺, but elicited small [IP₃]_i rises in the absence of Ca²⁺. The [IP₃]_i increase induced by TG (1.43 ± 0.1 µM) was about half that induced by BCR (2.70 ± 0.3 µM) (Figure 2D). However, in the PLCγ2^– cells, TG produced small [IP₃]_i increases, regardless of extracellular Ca²⁺ (Figure 2C and D). These results suggest that Ca²⁺ entry activated by TG-induced store depletion leads to IP₃ production by PLCγ2.

Fig. 2. Store-operated Ca²⁺ entry induced by passive depletion of Ca²⁺ stores activates PLCγ2. (A) Average time courses of Ca²⁺-responses induced by TG (2 µM) in WT (left) and PLCγ2^– (right) cells in Ca²⁺-free or Ca²⁺-containing external solution (n = 18–33). (B) Peak [Ca²⁺]_i rises induced by TG in WT and PLCγ2^– cells. (C) Typical time courses of TG-induced changes in F/F₀ of R9-PHIP3-D106 in WT (left) and PLCγ2^– cells (right) in the presence (top) or absence (bottom) of extracellular Ca²⁺ (n = 7–15). (D) Peak TG-induced [IP₃]_i rises. (E) Dependence of TG-induced ‘off’ Ca²⁺ responses on extracellular Ca²⁺ concentrations in WT and PLCγ2^– cells. The ‘off’ responses were induced after 12.5 min exposure to TG in Ca²⁺-free solution (n = 34–53). Right, peak [Ca²⁺]_i rises plotted against extracellular Ca²⁺ concentration.

In Figure 2E, the Ca²⁺ entry-evoked ‘off’ Ca²⁺ response was induced separately from the preceding passive Ca²⁺ release/depletion by TG (Parekh et al., 1997). In the physiological range of extracellular Ca²⁺ concentrations (1–10 mM), the Ca²⁺ entry-induced ‘off’ responses were significantly higher in the WT cells than those in the PLCγ2^– cells. Replacing Ca²⁺ with Ba²⁺ failed to discriminate the ‘off’ responses between the WT and the PLCγ2^– cells (Supplementary figure 3A), suggesting a specific positive role of Ca²⁺ in PLCγ2 activation or a lack of Ba²⁺ permeability of the Ca²⁺ entry channel. We also tested the possibility that the suppressed ‘off’ response in the PLCγ2^– cells derives from a decrease in the electrical driving force for Ca²⁺ by membrane depolarization through cation channels such as TRPM7, which associates with PLCβ via the C2 domain (Runnels et al., 2002). In the solution containing K⁺ ionophore valinomycin, which maintains cell membrane potential at about –75 mV (Mori et al., 2002), and in the membrane-depolarizing high K⁺ (30 mM) solution, the ‘off’ [Ca²⁺]_i increase remained significantly smaller in the PLCγ2^– cells than in the WT cells (Supplementary figure 3B). Thus, the reduced ‘off’ response in the PLCγ2^– cells was not due to membrane depolarization.

IP₃-mediated Ca²⁺ release is involved inTG-induced ‘off’ Ca²⁺ responses

We next examined whether IP₃R-mediated Ca²⁺ release is involved in TG-induced ‘off’ Ca²⁺ responses in DT40 B cells. Importantly, larger ‘off’ Ca²⁺ responses in the WT cells compared with those in the PLCγ2^– cells were not observed after pretreatment with the IP₃R antagonists, Xestospongin C (Xest C) (Figure 3A and B) and 2-aminoethoxydiphenyl borate (2-APB) (Supplementary figure 3C). The involvement of IP₃-mediated Ca²⁺ release in the ‘off’ responses was confirmed by using ionomycin (IM), which depletes internal Ca²⁺ stores more indiscriminatingly and efficiently than TG, which selectively targets SERCA-containing stores (Supplementary figure 3D) (Parekh et al., 1997). The WT cells treated with IM showed greater Ca²⁺ release, but a smaller ‘off’ Ca²⁺ response compared with the WT cells treated with TG (Figure 3C). IM elicited nearly identical ‘off’ Ca²⁺ responses in the WT and the PLCγ2^– cells, suggesting that thorough store depletion by IM nullified the contribution of IP₃-induced Ca²⁺ release to the ‘off’ responses in the WT cells. Interestingly, IM elicited significant PLCγ2 translocation (Table I) and [IP₃]_i increase (1.55 ± 0.1 µM). In the cells deficient in three IP₃R subtypes (IP₃R-KO) (Sugawara et al., 1997), TG and IM induced nearly identical ‘off’ Ca²⁺ responses (Figure 3D). The data support the view that ‘off’ responses are promoted by IP₃-induced Ca²⁺ release from Ca²⁺ stores, when Ca²⁺ stores are partially depleted (Supplementary figure 3D).

Fig. 3. Critical involvement of IP₃-induced Ca²⁺ release in ‘off’ Ca²⁺ responses. (A and B) The IP₃R blocker Xest C significantly suppresses TG-induced ‘off’ Ca²⁺ responses in WT (A) but not in PLCγ2^– (B) cells (n = 23–48). Xest C (50 µM) was loaded with fura-2/AM using 0.1% Pluronic F-127 (Molecular Probes) for 30 min prior to [Ca²⁺]_i measurements. During measurements, Xest C (50 µM) was added to perfusion solution for 20 min, and was omitted from Ca²⁺ readministration solution. Left, average time courses. Right, peak [Ca²⁺]_i rises in Ca²⁺-free or Ca²⁺-containing external solution. (C) Left, average time courses of Ca²⁺ responses induced by TG and IM (1 µM) in WT and PLCγ2^– cells (n = 16–33). Right, comparison between peak IM- and TG-induced [Ca²⁺]_i rises in Ca²⁺-free or Ca²⁺-containing external solution. (D) Left, average time courses of Ca²⁺ responses induced by TG or IM in IP₃R knockout cells. Right, peak IM- and TG-induced [Ca²⁺]_i rises in Ca²⁺-free or Ca²⁺-containing solution.

Table I. Changes of PLCγ2 localization by BCR stimulation (% of total fluorescence).

	PLCγ2 mutant	Before BCR stimulation (membrane area)	5 min after BCR stimulation (membrane area)
WT	WT (2 mM Ca2+)	37.0 ± 3.9 (n = 5)	57.9 ± 2.3*
	LD	32.7 ± 2.7 (n = 3)	53.4 ± 2.8*
	LD + U73122	30.1 ± 2.9 (n = 3)	31.0 ± 5.0
	LD (EGTA)	35.0 ± 3.3 (n = 3)	37.6 ± 4.6
	ΔC2	26.9 ± 8.4 (n = 3)	22.4 ± 5.1
	minΔC2	33.3 ± 6.4 (n = 3)	42.1 ± 5.9
	ΔPH	25.5 ± 4.2 (n = 3)	24.5 ± 3.5
	E165A	31.8 ± 4.7 (n = 3)	56.5 ± 2.9*
	P820L	28.8 ± 3.1 (n = 3)	46.5 ± 2.2*
PLCγ2–	WT (2 mM Ca2+)	33.8 ± 4.2 (n = 4)	60.2 ± 9.3*
	WT (EGTA)	39.4 ± 5.7 (n = 5)	37.3 ± 7.5
	WT (2 mM Ba2+)	35.7 ± 6.1 (n = 3)	39.1 ± 2.3
	WT + U73122	29.3 ± 7.9 (n = 3)	35.0 ± 8.1
	WT + U73343	31.2 ± 4.0 (n = 5)	56.8 ± 3.5*
	LD	31.7 ± 2.1 (n = 4)	34.2 ± 1.9
	ΔC2	34.7 ± 8.3 (n = 3)	44.2 ± 3.8
	minΔC2	36.7 ± 5.3 (n = 3)	47.6 ± 7.7
	ΔPH	27.5 ± 4.8 (n = 3)	30.2 ± 8.8
	E165A	29.4 ± 3.6 (n = 3)	50.0 ± 5.3*
	P820L	31.3 ± 5.4 (n = 3)	60.4 ± 3.7*
	WT (ionomycin)	38.4 ± 2.0 (n = 4)	44.9 ± 2.1*

Membrane regions of fluorescent areas in DT40 cells were as described in Figure 5A.

n, number of experiment. *P < 0.05, significantly different from cells before BCR stimulation.

Contrary to the above notion, the IP₃R-KO cells showed TG-induced ‘off’ Ca²⁺ responses indistinguishable from those in the WT cells (Supplementary figure 4A). This discrepancy seems partly due to a compensatory enhancement in DG-sensitive Ca²⁺ channel expression; notable La³⁺-sensitive ‘off’ [Ca²⁺]_i increases due to the DG analogue 1-oleoyl-2-acetyl-sn-glycerol (OAG, 30 µM) were observed in about half of the IP₃R-KO cells (18/41) (Supplementary figure 4B), while fewer WT cells (15/50) showed small, transient OAG-induced ‘off’ responses. As reported previously (Venkatachalam et al., 2001), extracellular application of Ba²⁺ for Ca²⁺ elicited a tiny increase in the fluorescence ratio (data not shown), suggesting the importance of extracellular Ca²⁺ in differentiating WT cells from IP₃R-KO cells in the OAG-induced ‘off’ response. Thus, TG-induced ‘off’ Ca²⁺ responses in IP₃R-KO cells may result from store-operated Ca²⁺ entry and Ca²⁺ entry activated by DG via PLCγ2 activation, while the contribution of the latter pathway may be minor in WT cells.

C2 domain is responsible for Ca²⁺ influx-dependent activation of PLCγ2

The domains critical for Ca²⁺ entry-dependent activation were determined in PLCγ2. Three potential Ca²⁺-binding domains (EF-hand, catalytic domain and C2 domain), and the PH and SH3 domains are distinguished in PLCγ2 (Bairoch et al., 1990; Essen et al., 1996; Rhee, 2001; Ananthanarayanan et al., 2002). In PLCγ1, PH mediates membrane-targeting through phosphatidylinositol 3,4,5-trisphosphate (PIP₃) binding, while SH3 acts as a physiological guanine nucleotide exchange factor (GEF) for the nuclear GTPase PIKE, an activator of nuclear phosphatidylinositol 3-kinase (PI3K) (Ye et al., 2002) that produces the PIP₃ required for BCR-induced [Ca²⁺]_i elevation (Falasca et al., 1998). Mutations were introduced in the PLCγ2 domains tagged with yellow fluorescent protein (YFP) (Figure 4A). As expected, expression of the LD mutant in PLCγ2^– cells failed to restore either the BCR-induced [Ca²⁺]_i increase or to potentiate a TG-induced ‘off’ Ca²⁺ response (Figure 4B and C). Strikingly, deletion of C2 domain (ΔC2), which is involved in the formation of the Ca²⁺-phospholipid ternary complex in PLCδs (Ananthanarayanan et al., 2002), ablated the PLCγ2 function to restore the BCR- and TG-induced Ca²⁺ responses. Deleting one of the expected Ca²⁺ binding regions (Δ1138E-1140D) (Essen et al., 1996) in C2 (minΔC2) similarly blocked the restoration of the ‘off’ responses but not that of the BCR-induced Ca²⁺ release, strikingly differentiating Ca²⁺ entry-mediated responses from Ca²⁺ release. The PH deletion mutant (ΔPH) also failed to restore the PLCγ2^– deficiency, as reported in PLCγ1 activation (Falasca et al., 1998). In contrast, the SH3 deletion (ΔSH3) mutant elicited partially impaired the Ca²⁺ responses, and the point SH3 mutant with leucine at proline 820 (P820L) that corresponds to P842 essential for GEF activity in PLCγ1 (Ye et al., 2002), or the EF-hand mutant (E165A), behaved like WT. Thus, the C2 domain is critical for Ca²⁺ entry-dependent activation of PLCγ2.

C2 domain mediates Ca²⁺ influx-dependent translocation of PLCγ2

The above data and the requirement for C2 when various proteins target the plasma membrane (Sutton et al., 1995; Ananthanarayanan et al., 2002; Delmas et al., 2002) motivated us to visualize the localization of YFP-tagged PLCγ2 (Figure 5). Transfected PLCγ2 was diffusively distributed in the cytosol from the perinuclear area to the plasma membrane of the WT and the PLCγ2^– cells. Within 4 min, but not within 2 min, stimulation of BCR efficiently concentrated PLCγ2 near the plasma membrane in the presence of extracellular Ca²⁺ (Figure 5A). This PLCγ2 translocation was abolished by removing Ca²⁺, using Ba²⁺ for Ca²⁺, using ΔC2 and minΔC2 in the WT and the PLCγ2^– cells (Figure 5A and B; Table I). ΔPH also suppressed translocation. The EF and SH3 mutants showed significant translocation (Table I). Importantly, even in the absence of extracellular Ca²⁺, expression of membrane-bound PLCγ2 (mPLCγ2) (Ishiai et al., 1999) in the PLCγ2^– cells promoted sustained [IP₃]_i and [Ca²⁺]_i increases (Figure 4D and E). Thus, C2 mediates the Ca²⁺ entry-induced transportation of PLCγ2 for subsequent activation. Furthermore, in the WT cells, but not in the PLCγ2^– cells, the LD mutant of PLCγ2 showed significant BCR-induced plasma membrane localization inhibited by the PLC blocker, U73122 (Figure 5C). This suggests that transfected PLCγ2 does not require its own lipase activity for Ca²⁺ entry-induced translocation to the plasma membrane when endogenous lipase-active PLCγ2 is present.

Fig. 5. Ca²⁺ influx induces translocation of PLCγ2 via the C2 domain. (A) Confocal fluorescence images indicating BCR-induced changes of subcellular localization of YFP-tagged PLCγ2 in PLCγ2^– cells in Ca²⁺-containing or Ca²⁺-free solution (n = 5). Right, representative fluorescent changes of PLCγ2–EYFP distributed in 1 µm widths peripheral regions (Mem.) and in the rest the cytosolic areas (Cyt.). (B) Confocal images indicating changes of localization of PLCγ2 constructs (WT, LD, ΔC2 or ΔPH) in WT and PLCγ2^– cells by 5 min BCR stimulation (n = 5). (C) PLC inhibitor U73122 blocks BCR-induced localization changes of LD in WT cells. Treatment with U73122 (3 µM) or its inactive analogue U73343 was started 20 min prior to BCR stimulation. Lower left, average time courses of BCR-induced F/F₀ changes of R9-PHIP₃-D106 in WT cells expressing GFP or LD, or pretreated with U73122 or U73343. Lower right, maximal [IP₃]_i elevation upon BCR stimulation.

Functional and physical association of PLCγ2 with TRPC3

Among the TRPC channels responsible for receptor-activated Ca²⁺ entry (Montell et al., 2002), expression of TRPC1, C3, C4, C5 and C7 was identified in DT40 cells (Figure 6A). The respective TRPC channels were tested for functional coupling with PLCγ2 by co-expressing PLCγ2 in HEK cells. Upon stimulation of the endogenous muscarinic acetylcholine receptors (MRs) with CCh, TRPC3, C5, C6 and C7 alone promoted increased [Ca²⁺]_i (Figure 6C), while co-expression of PLCγ2 exerted further potentiation only in the TRPC3-expressing cells (Figure 6B and C). LD elicited the intact potentiation of peak Ca²⁺ responses but failed to sustain Ca²⁺ response. In the patch-clamp recording, PLCγ2 significantly sustained CCh-activated TRPC3 currents (Figure 6D). Upon MR stimulation, PLCγ2–YFP fluorescence near the plasma membrane showed increases that indicate translocation in TRPC3-expressing cells (47.8 ± 0.7 to 52.3 ± 0.3%, P < 0.05). Furthermore, immumoprecipitation analysis revealed a physical association between PLCγ2 and TRPC3 (Figure 6E), which was impaired by ΔSH3, but not by LD or minΔC2 (Figure 6F), in the HEK cells. In the DT40 cells, co-immunoprecipitation of TRPC3 with PLCγ2 enhanced by BCR stimulation was observed (Figure 6G). These results suggest that receptor-activated TRPC3 channels functionally and physically interact with PLCγ2 via SH3 to enhance Ca²⁺ response and to maintain their own channel activity.

Fig. 6. Functional and physical association of PLCγ2 with the TRPC3 Ca²⁺ channel promotes receptor-evoked TRPC3 currents and Ca²⁺ responses. (A) RT–PCR analysis of RNA expression of TRPC homologues in DT40 cells. (B) Average time courses of Ca²⁺ responses upon stimulation of MRs in HEK cells expressing TRPC3 and/or PLCγ2 (WT or LD) (n = 32–38). (C) Peak MR-activated [Ca²⁺]_i rises in HEK cells expressing respective TRPC1–C7 channels alone or along with PLCγ2. (D) Left, representative traces of ionic currents recorded from cells expressing TRPC3 alone or plus PLCγ2 at a holding potential of –60 mV. Right, peak and sustained (30 s after reaching peak) amplitudes of the MR-activated TRPC3 currents (n = 9–10). (E) Co-immunoprecipitation of TRPC3 and PLCγ2 via SH3. Immunoprecipitates (IP) with anti-GFP polyclonal antibody from TRPC3/GFP- and TRPC3/PLCγ2–YFP-expressing HEK cells (upper) or GFP/PLCγ2- and TRPC3–GFP/PLCγ2-expressing HEK cells (lower) were separated on 8% SDS-PAGE, and blotted with anti-PLCγ2 or anti-TRPC3 antibody, respectively, in western blot analysis (WB). (F) IP with the anti-GFP antibody from LD/TRPC3-, minΔC2/TRPC3- and ΔSH3/TRPC3-expressing HEK cells. Anti-TRPC3 antibody was used to assess co-immunoprecipitation in WB. All the PLCγ2 constructs were tagged with YFP. (G) IP with the anti-GFP antibody from TRPC3-GFP-expressing WT and TRPC3–GFP-expressing PLCγ2^– cells with (+) or without (–) 5 min BCR stimulation (anti-IgM). Co-immunoprecipitation was assessed using anti-PLCγ2 antibody in WB.

Critical involvement of Ca²⁺ entry-dependent PLCγ2 activation in receptor-evoked Ca²⁺ oscillation

Periodic increases in [Ca²⁺]_i (Ca²⁺ oscillation) are important for the efficiency and specificity of gene expression in lymphocytes (Dolmetsch et al., 1998) and other biological responses (Miyazaki et al., 1992; Thorn et al., 1993). Figure 7A depicts a typical Ca²⁺ oscillation induced by stimulation of BCR in DT40 cells. LD and ΔC2 expression almost eliminated the late oscillatory phase of the BCR-induced Ca²⁺ response in the WT cells, while they only elicited partial suppression of the initial transient Ca²⁺ response (Figure 7B and C). Considering the translocation of LD to the plasma membrane in the WT cells (Figure 5B), this evidence supports the view that lipase activity is essential for the translocated PLCγ2 in maintaining downstream Ca²⁺ oscillation. The expression of PLCγ2 restored Ca²⁺ oscillation in the PLCγ2^– cells (Figure 7F). However, ΔSH3 incompletely restored Ca²⁺ oscillation, and ΔC2 failed to elicit a detectable Ca²⁺ response in the PLCγ2^– cells (Figure 7G and H). The results demonstrate the physiological significance of Ca²⁺ influx-dependent PLCγ2 activation in BCR-induced Ca²⁺ oscillation.

Fig. 7. PLCγ2 requires its lipase activity and C2 domain to elicit BCR-evoked Ca²⁺ oscillation. (A–D) Representative time courses of BCR-induced Ca²⁺ responses in WT and PLCγ2^– cells in Ca²⁺-containing (A–C and E–H) (n = 16–72) or Ca²⁺-free (D) (n = 50) solution (long time course, 60 min). PLCγ2 constructs were the same as those in Figure 4.

Amplification of DG signalling promotes ERK activation

The PLCγ2-mediated signal amplification should have a significant impact on the responses downstream of DG, since DG production is also enhanced by Ca²⁺ entry-induced PLCγ2 activation and PIP₂ hydrolysis. Activation of PKC by DG promotes activation of extracellular signal-regulated kinase (ERK) through phosphorylation in DT40 cells (Kurosaki, 2002). Figure 8A depicts sustained BCR-induced ERK activation in an extracellular Ca²⁺-dependent manner. The ERK activation was inhibited by a PKC inhibitor, bisindolylmaleimide (BIM, 10 µM), and more prominently by PLCγ2 deficiency, which was partially reversed by a PKC activator phorbol-12-myristrate-13-acetate (PMA) (Figure 8B). Hence, signal cascades other than PKC activation may be evoked by PLCγ2 to activate ERK. Furthermore, TG gradually increased the phosphorylated levels of ERK in an extracellular Ca²⁺- and PLCγ2-dependent manner (Figure 8C and D). Thus, Ca²⁺ entry-dependent PLCγ2 activation is important in DG-induced activation via PKC of ERK in mitogen-activated protein kinase (MAPK) pathways.

Fig. 8. Requirement of Ca²⁺ influx-dependent PLCγ2 activation for sustained BCR-induced activation of ERK in DT40. (A) WT or (B) PLCγ2^– cells were stimulated with anti-IgM (1 µg/ml) for indicated time in the presence or absence of extracellular Ca²⁺. Treatment with BIM (10 µM) was started 10 min before BCR stimulation. PMA (50 nM) was applied to anti-IgM in PLCγ2^– cells (B). (C and D) Average time courses of ERK activation evoked by TG (2 µM) in WT (C) and PLCγ2^– (D) cells. The ERK activities were assessed by the quantification of its phosphorylation using phospho-specific antibody.

Discussion

Ca²⁺ entry amplifies BCR-induced Ca²⁺ signalling via PLCγ2 translocation and activation

This study reveals a novel molecular mechanism that amplifies cellular signals in the second phase of the Ca²⁺ response in B lymphocytes. The amplification mechanism is a positive feedback cycle, in which a key event is the Ca²⁺ entry-evoked translocation towards the plasma membrane and the subsequent PLCγ2 activation/IP₃ production then leads to Ca²⁺ release and entry (Figure 9). The C2 domain mediates the translocation of PLCγ2 (Figure 5B). Importantly, C2 is also found in PLCδ (Ananthanarayanan et al., 2002), phospholipase A₂ (PLA₂) (Clark et al., 1991), synaptotagmin I (Sutton et al., 1995) and PKC isoforms (Schaefer et al., 2001; Delmas et al., 2002), among which PLA₂ and PKCs are translocated by Ca²⁺ entry (Clark et al., 1991; Schaefer et al., 2001). For targeting membrane microdomains and activating PLCγ2, its PH domain is essential (Figure 5B). PIP₃ by PI3K is required for receptor-stimulated Ca²⁺ mobilization (Bae et al., 1998), and PIP₃ binds to the PH with the highest affinity among the phosphoinositides to activate PLCγ1 through membrane targeting (Falasca et al., 1998). In fact, our data demonstrate that the ΔPH mutant failed to restore the Ca²⁺ responses (Figure 4B and C), and that wortmannin (1 µM), which suppresses PIP₂ and PIP₃ synthesis by inhibiting PI3K and PI4K, significantly suppressed BCR-stimulated Ca²⁺ oscillation, and attenuated the TG-induced [Ca²⁺]_i increase in the WT cells, but not in the PLCγ2^– cells (Supplementary figure 5).

Fig. 9. Model for amplification of Ca²⁺ signalling by PLCγ2 in B lymphocytes. BCR stimulation induces activation of PLCγ2 via phosphorylation by Syk tyrosine kinase, and the activated PLCγ2 hydrolyses PIP₂ to produce IP₃. IP₃ leads to Ca²⁺ release from ER and Ca²⁺ influx via TRPC3-containing receptor-activated Ca²⁺ channels (RACCs), which may correspond to SOCs activated through interaction with IP₃R or to DG-activated Ca²⁺ channels. Ca²⁺ influx via RACC promotes lipase-independent PLCγ2 translocation from the cytosol to the plasma membrane (PM) by acting on the C2 domain. The targeted PLCγ2 is activated to hydrolyse PIP₂ and produce IP₃ and DG, which in turn induces Ca²⁺ release from ER and Ca²⁺ influx via Ca²⁺ entry channels such as TRPC3. PLCγ2 may also positively regulate TRPC3 channels via direct interaction.

The second key process in the amplification mechanism is successive Ca²⁺ release. We observed a striking difference in the ‘off’ response between the WT and the PLCγ2^– cells (Figure 3C) when TG was used to selectively deplete SERCA-containing stores, but not when IM was employed for more indiscriminate and thorough store depletion. This suggests that IP₃-induced Ca²⁺ release from partially depleted stores significantly contributes to the TG-induced ‘off’ response, being consistent with the previous report in which TG frequently failed to induce Ca²⁺ release-activated Ca²⁺ currents in patch-clamp recording despite its efficiency in inducing ‘off’ responses in [Ca²⁺]_i measurement (Parekh et al., 1997). Efficient activation of Ca²⁺ entry channels via IP₃-induced store depletion (Zhu et al., 1996) following PLCγ2 activation is the third key process. This process may be elicited from the physical association of TRPC3 with PLCγ2 and IP₃Rs (Boulay et al., 1999; Kiselyov et al., 1999) (Figure 6E). Activation of type 3 IP₃R in the plasma membrane (Vazquez et al., 2002) and direct DG action on Ca²⁺ channels (Hofmann et al., 1999; Trebak et al., 2003) are also important processes to be considered. Ultimately, determinating the composition of the signalplex that contains PLCγ2, TRPC3 and other proteins such as BLNK (Ishiai et al., 1999) is essential for fundamental understanding of the amplification mechanism.

The transfected LD PLCγ2 mutant was translocated to the plasma membrane area in the WT cells, but not in the PLCγ2^– cells (Figure 5B). This result suggests that Ca²⁺ influx is mediated by endogenous PLCγ2 to elicit subsequent translocation of LD in WT cells, and that PLCγ2 does not require its own lipase activity in the translocation process. Independent of the hydrolysis of PIP₂, the translocated PLCγ2 may exert positive regulatory action on TRPC3 Ca²⁺ channel activity at the plasma membrane through physical interaction via SH3 (Figure 6) (Patterson et al., 2002). It is also possible that PLCγ2 translocation is accompanied by associated TRPC3 targeting the plasma membrane (Figure 9). In fact, on M1R stimulation, we observed partial but significant restoration of Ca²⁺ entry-induced [Ca²⁺]_i elevation by the ΔSH3 mutant in the PLCγ2^– cells (Figure 1E). Thus, the enzyme activity-independent translocation and consequent interaction of PLCγ2 with TRPC3 at the plasma membrane may underlie a non-enzymatic role of PLCγ2 (Patterson et al., 2002; Putney, 2002), cooperating with the second lipase activation in enhancing Ca²⁺ entry to fully restore the Ca²⁺ deficiency in mutant cells. This view is consistent with a recent report by Patterson et al. (2002) in which LD restored Ca²⁺ entry-induced [Ca²⁺]_i increases in M5R-stimulated PLCγ2^– cells, although the restoration was comparable to that by WT. The difference in peak responses between our data using M1R and the data of Patterson et al. (2002) using M5R is attributable to different activation and/or desensitization speeds of receptor-activated Ca²⁺ entry and subsequent lipase activity-dependent PLCγ2 signalling, considering smaller Ca²⁺ release but larger Ca²⁺ entry-induced Ca²⁺ response upon M1R stimulation (Figure 1E) compared with that upon M5R stimulation. The slower signal transduction through PLCγ2 may result in the Ca²⁺ entry and lipase activity-dependent response that persisted throughout the protocol (Figure 1E), in contrast to the almost complete suppression of Ca²⁺ entry and absence of lipase activity-dependent response in the M5R-expressing PLCγ2-deficient cells (Figure 6 of Patterson et al., 2002). Since deleting SH3 also terminated the enhancing effect of PLCγ1 on the Ca²⁺ entry-induced Ca²⁺ responses in the PC12 cells and the TRPC3-expressing HEK cells (Patterson et al., 2002), PLCγ1 and PLCγ2 may share an enzyme-independent function in receptor-induced Ca²⁺ responses.

Physiological impact of Ca²⁺ entry-induced PLCγ2 activation in BCR signalling

Our finding is contrary to the prevailing view that receptor-activated Ca²⁺ entry channels amplify Ca²⁺ responses predominantly through the direct effect of Ca²⁺ permeation and replenishing Ca²⁺ stores. The PLCγ2-mediated amplification of the Ca²⁺ signal would therefore dramatically change the view of the physiological roles played by receptor-activated Ca²⁺ entry. The PLCγ2-mediated signal amplification should have a significant eventual impact on downstream Ca²⁺-dependent responses, including regulation of gene expression via transcription factors such as NF-AT (Dolmetsch et al., 1998), cell proliferation and apoptosis (Verkhratsky and Petersen, 2002; Irvine, 2002). Importantly, PIP₂ hydrolysis and DG production are also enhanced in the amplification mechanism. Activation of PKC by DG promotes ERK activity in the MAPK pathway to induce phosphorylation of the transcription factor CREB, which interacts with an enhancer cAMP-response element (Kurosaki, 2002). Our results provide direct evidence that ERK activation is promoted by Ca²⁺ entry via PLCγ2 and PKC. DG is also known to activate the TRPC3, C6 and C7 channels in a PKC-independent manner (Hofmann et al., 1999; Okada et al., 1999; Trebak et al., 2003), while PIP₂ is linked to the pathway composed of PI3K, Bruton’s tyrosine kinase and the serine threonine kinase Akt/PKB (Kurosaki, 2002). Hence, the model predicts extensive regulatory roles of Ca²⁺ entry-induced PLCγ2 activation in the amplification and/or co-ordination of multiple BCR-induced signalling pathways.

PLCγ2 is also expressed in the anterior pituitary and cerebellar Purkinje and granule cells (Rebecchi and Pentyala, 2000). Previously, regarding Ca²⁺ oscillation in REF52 fibroblasts, Harootunian et al. (1991) proposed that a dominant feedback mechanism is the stimulation of PLC by the [Ca²⁺]_i increase through population IP₃ measurement, while pathways that evoke the [Ca²⁺]_i increase remained unidentified. Recent research has also demonstrated production of IP₃ induced by voltage-dependent Ca²⁺ influx (Okubo et al., 2001) and regulation of metabotropic glutamate receptor-mediated neuronal signalling by extracellular Ca²⁺ in Purkinje neurons (Tabata et al., 2002). When eggs are fertilized by sperm, the Ca²⁺ oscillations essential for embryo development are mediated by PLCγ (Sato et al., 2000) or homologous PLCζ (Saunders et al., 2002). Furthermore, the activation of recombinantly overexpressed PLCδ1 by Ca²⁺ entry that follows the PLCβ activation upon bradykinin stimulation has been reported in PC12 cells (Kim et al., 1999). Since the ubiquitous PLCγ1 and PLCδ also contain C2 domains (Rebecchi and Pentyala, 2000), the signal amplification mediated by Ca²⁺ entry-induced PLC translocation and activation may be universal among various cell types.

Materials and methods

Cell cultures and cDNA expression

The PLCγ2 mutants were constructed using PCR techniques. The WT PLCγ2 (Takata et al., 1995), LD (H327F) (Patterson et al., 2002) and other mutated constructs were fused at the C-terminus with YFP using an EYFP-N1 vector (Clontech). In the ΔC2, ΔPH and ΔSH3 mutants, residues 1043M-1237V, 20R-131Q and 810D-829A were deleted, respectively. Cell cultures and cDNA expression in the DT40 cells using the vesicular somatitis virus and in the HEK293 cells were performed as described previously (Hara et al., 2002; Mori et al., 2002).

Measurement of changes in [Ca²⁺]_i

The cells on coverslips were loaded with 1 µM fura-2/AM in a physiological salt solution (PSS) containing (in mM): 150 NaCl, 8 KCl, 2 CaCl₂, 1 MgCl₂, 5 HEPES, and 5.6 glucose, adjusted to pH 7.4 with NaOH. Fluorescence images of the cells were recorded and analysed with a video image analysis system (ARGUS CA-20; Hamamatsu Photonics) as described previously (Mori et al., 2002). All ratio data were calculated to [Ca²⁺]_i using an in vivo calibration method.

Fluorometric measurement of changes in [IP₃]_i

The basic strategy in designing the optical direct IP₃ sensors is described in Morii et al. (2002). Conjugation with fluorophore 6-bromoacetyl-2-dimetyl-aminonaphtalene (DAN) at Cys106 introduced near the mouth of the IP₃-binding site prevents semi-specific ligand binding (Morii et al., 2002), conferring an IP₃-selective sensitivity to the PH protein. The dissociation constant was only slightly affected by Ca²⁺ and ranged from 2.2 to 3.3 µM in 0–1 mM Ca²⁺. The protein was further fused to arginine (Arg) nonapeptide for improved cellular incorporation. The modified sensor, R9-PHIP3-D106 [Arg (R)₉-PH IP₃ sensor conjugated with DAN at residue 106), showed an efficient in vivo IP₃-sensing ability. In in vivo measurements, cells were plated onto poly-l-lysine-coated glass coverslips. After a 2 h incubation for cell adhesion, the cells were transferred to Hank’s balanced salt (HBS) solution containing 1 µM of IP₃ sensor and 0.1% F-127 (Molecular Probes) for 10 min at 37°C. After the cells were washed, changes in the fluorescence images of the cells were observed using a video image analysis system at an emission wavelength of 510 nm (bandwidth 20 nm) at room temperature by excitation at 380 nm (bandwidth 11 nm) in PSS. [IP₃]_i increases were detected as fluorescence quenching of R9-PHIP₃-D106 due to IP₃ binding. All traces were displayed after subtracting the averaged time courses of fluorescence reduction due to photobleaching or photodegradation.

For the in vivo calibration of [IP₃]_i, synthesized IP₃ (Dojindo) was applied to R9-PHIP3-D106-preloaded DT40 cells after treatment with β-escin (1 µM; Sigma) (Supplementary figure 1). The initial decrease in fluorescence was fitted to a binding isotherm:

ΔF = ΔF_max {[IP₃]/(K_d + [IP₃])} (1)

where ΔF is the change in fluorescence intensity. ΔF_max is ΔF at saturation, K_d is the dissociation constant, and [IP₃] is the concentration of IP₃. [IP₃]_i was calculated based on Equation 1, where the K_d value is 1.24 µM.

Electrophysiology

The whole-cell patch–clamp technique was performed as described previously (Mori et al., 2002). The cells on coverslips were perfused in an external solution containing in (mM): 140 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 10 glucose, adjusted to pH 7.4 with NaOH (300 mOsm). The pipette solution contained in (mM): 75 CsOH, 75 asparate, 40 CsCl, 4 MgCl₂, 10 BAPTA, 2 ATP Na₂, 5 HEPES, and 8 phosphocreatine, adjusted to pH 7.4 with CsOH. mTRPC3-pCIneo (Okada et al., 1999) was co-transfected with pIRES2-EGFP (Clontech) or pIRES2-EGFP that contained PLCγ2 cDNA in the HEK cells, and GFP-positive cells were selected for recording.

Confocal visualization of PLCγ2–YFP fusion constructs

PLCγ2–YFP-expressing cells were plated onto poly-l-lysine coated glass coverslips after 72 h of retroviral infection. Fluorescence images were acquired with a confocal laser-scanning microscope (FV500; Olympus) using the 488 nm line of an argon laser for excitation and a 505 nm long-pass filter for emission. The specimens were viewed at high magnification using plan oil objectives (60×, 1, 40 NA; Olympus).

Expression analysis

The PCR protocol used for the expression analysis is as described previously (Inoue et al., 2001). The PCR primers used are listed in Table II.

Table II. Sequences of antisense and sense oligonucleotides used.

Gene	Orientation	Pair of primers for first PCR (5′ to 3′)	Pair of primers for nested PCR (5′ to 3′)
TRPC1	Sense	TGGATTATTGGGATGATTTGGTCAGAC
	Antisense	TCGGGCGAATTTCCATTCCTTATCTTC
TRPC3	Sense	GTCTGAAGTAACCTCTGTGTCCTG	GAGAACATGGATATGTTCTCTATGGC
	Antisense	CACATCTCAACATATTGGGATTGAG	ATGGCTAACTCTTCTGTCGCCTGGGAC
TRPC4	Sense	GAGCTGGGACATGTGGCATCCAACGC	GAGGCACTGTTTGCAATTGCAAACATC
	Antisense	GCACAAGGTGTCTCCATTATCCACC	GCTTGGTATGACATTGAAAGGAGTGGGC
TRPC5	Sense	TTGAGGTCTTCTGAAGGACCACGCAG	CTGCAAGTGGATTCACAATCAGCTGTGC
	Antisense	GGCTTCTCTTCCGCCATGTCAATCCCAG	GTTTTAATTAGCGCAGACACATTGAAGTC
TRPC7	Sense	ACGCCTGATTAAGCGGTATGTGCTGAAG	CTCAGGTGGACCGAGAAAATGATGAA
	Antisense	GCCAAAAGAGAGGCTCGGTGTGTTG	GCAAAGAGGAAGTTTTTGGCTTGGAATC

Immunoprecipitation and western blot analysis

TRPC3–GFP was first established in the EGFP-N1 vector (Clontech), and transferred into pMXΔ. HEK or DT40 Cells (∼3 × 10⁶) were lysed in 1 ml of RIPA buffer (pH 8.0) containing 150 mM NaCl, 0.5% NP-40, 0.1% sodium deoxycholate, 0.1% SDS, 50 mM Tris, 2 mM EGTA, 1 mM PMSF and 10 µg/ml leupeptin. GFP-tagged mTRPC3 or YFP-tagged PLCγ2 protein was immunoprecipitated in an RIPA buffer with 3 µl of anti-GFP polyclonal antibody (Clontech) in the presence of a 25 µl bed volume of Sepharose G–agarose beads (Amersham Pharmacia) rocking overnight at 4°C (Patterson et al., 2002). The immune complexes were washed three times with the RIPA buffer for 5 min at room temperature, and resuspended in 50 µl of the SDS sample buffer. The protein samples (5 µl) were fractionated by 8% SDS–PAGE and electrotransferred onto a nitrocellulose membrane (Inoue et al., 2001). The blots were incubated with an anti-PLCγ2 (Ishiai et al., 1999) or an anti-mouse TRPC3 antiserum, and stained using the ECL system (Amersham Pharmacia). The anti-mouse TRPC3 rabbit antiserum was directed against the C-terminus (EELAILHKLSEKLNPSVLRCE).

Analysis of ERK activity

Activation of ERK was determined by western blot with anti-phosphospecific ERK antibody (Cell Signaling). A total of ∼5 × 10⁵ cells were stimulated in serum-free PSS, and then lysed in 50 µl of lysis buffer (50 mM Tris–HCl pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 100 µM Na₃VO₄, 2 mM dithiothreitol, 2 mM phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin and 5 µg/ml aprotinin) for 10 min on ice. After centrifugation at 22 000 g for 10 min, 10 µl of supernatants were subjected to 10% SDS–PAGE and electrotransferred onto a nitrocellulose membrane. The total amount of ERK was detected by using anti-ERK polyclonal antibody (Santa Cruz Biotech). The bands were scanned and the density of each band was determined by NIH image software.

Statistical analysis

All data are expressed as means ± SEM The data were accumulated under each condition from at least three independent experiments. The P values are the results of Student’s t-tests.

Supplementary data

Supplementary data are available at The EMBO Journal Online.