Activation of the B cell receptor complex in B lymphocytes causes Ca2+ release from intracellular stores, which, in turn, activates ion channels known as Icrac. We investigated the mechanisms that link Ca2+ store release to channel gating in DT40 B lymphocyte cell lines genetically manipulated to suppress the expression of several tyrosine kinases: Btk, Lyn, Syk, and the Blnk adaptor molecule. The simultaneous but not the independent suppression of Lyn and Syk expression prevents the activation of Icrac without interfering with thapsigargin-sensitive Ca2+ store release. Icrac activation by Ca2+ is reversed in mutant cells by the homologous expression of the missing kinases. Pharmacological inhibition of kinase activity by LavendustinA and PP2 cause the same functional deficit as the genetic suppression of enzyme expression. Biochemical assays demonstrate that kinase activity is required as a tonic signal: targets must be phosphorylated to link Ca2+ store release to Icrac gating. The action of kinases on Icrac activation does not arise from control of the expression level of the stromal interaction molecule 1 and Orai1 proteins.
In mature B lymphocytes, binding of antigen or antireceptor antibody to the B cell receptor (BCR) complex causes a sustained rise in intracellular free Ca2+, which, in turn, regulates proliferation and differentiation of the cells into either memory or antibody-secreting ones (for review see Kurosaki, 2000). The increase in cytoplasmic Ca2+ arises from two successive events: first, there occurs a transient Ca2+ release from intracellular stores initiated by a rise in free inositol 1,4,5-trisphosphate (IP3; for review see Berridge et al., 2003). In turn, the emptying of Ca2+ stores activates Icrac (Ca2+ release activated) ion channels. The channels are Ca2+ permeable, and Ca2+ influx via these channels results in a sustained elevation of cytoplasmic free Ca2+ (Parekh and Penner, 1997; for review see Lewis, 1999). The molecular identity of the Icrac channels has yet to be fully determined, but recent studies demonstrate that the protein Orai1 (also called CRACM1) is an integral component of the channel and is associated with its Ca2+ selectivity filter (Vig et al., 2006a; Yeromin et al., 2006). Orai1, a gene product identified in severe combined immunodeficiency patients (Feske et al., 2006), was discovered to be important in Icrac function through RNAi screens (Feske et al., 2006; Vig et al., 2006b; Zhang et al., 2005).
Tyrosine kinase activity is absolutely required for activation of the BCR–Ca2+ signaling pathway (for review see Kurosaki, 2000). In the absence of kinase function, the sequence of molecular events linking BCR activation to IP3 production fails because the normal phosphorylation and activation of PLCγ, the enzyme that produces IP3, does not occur. The BCR complex is a multimer consisting of membrane Ig associated with Ig α/β heterodimers. The BCR complex interacts with and is phosphorylated by one or more of the following members of the Src kinase family: Lyn, Fyn, and/or Blk. In addition to the Src kinase family, two other tyrosine kinases participate in the BCR–Ca2+ signaling pathway: Syk (Syk family) and Btk (Tec family). All three tyrosine kinase families participate in PLCγ activation. Furthermore, tyrosine kinase–dependent activation of PLCγ is facilitated by adaptor or linker proteins such as Blnk (for review see Kurosaki, 2000). Whether these enzymes play a role in the events that link the emptying of Ca2+ stores to Icrac activation has not been investigated directly. However, indirect studies using pharmacological blockers of tyrosine kinase function have suggested that the enzymes may play a role in linking Ca2+ store release and Ca2+ influx (Lee et al., 1993, Sargeant et al., 1993a,b). In this study, we provide direct evidence of a role for kinases in the link between Ca2+ store emptying and Icrac activation, and we identify some of the specific enzymes involved.
Two general classes of mechanisms have been proposed to link the store release of Ca2+ to Icrac activation. The first class proposes that Icrac channels are structurally linked to the Ca2+-containing stores and that their activation depends on a conformational coupling between the store and the plasma membranes (Irvine, 1990; Petersen and Berridge, 1996; Kiselyov et al., 2001). The second class of mechanisms proposes that second messenger molecules accomplish this link through Ca2+-dependent activation of target proteins. Over time, several plausible messenger proteins have been proposed: G proteins (Bird and Putney, 1993; Fasolato et al., 1993; Jaconi et al., 1993; Petersen and Berridge, 1995), PKC (Parekh and Penner, 1995), tyrosine kinases (Lee et al., 1993; Sargeant et al., 1993a,b; Rosado et al., 2000), Ca2+ influx factor (Randriamampita and Tsien, 1993; Csutora et al., 1999), inositol 1,3,4,5-tetrakisphosphate (Luckhoff and Clapham, 1992), and cytochrome P-450 (Alvarez et al., 1992). Recently, however, a compelling case has been developed that identifies stromal interaction molecule 1 (STIM1) as a messenger protein between Ca2+ release and Icrac gating (Liou et al., 2005; Roos et al., 2005; Zhang et al., 2005). Simultaneous overexpression of STIM1 and Orai1 but not either alone facilitates the activation of Icrac (Peinelt et al., 2006). STIM1 is a membrane-bound Ca2+-binding protein (its structure includes an EF hand) located in the ER. STIM1 acts as a Ca2+ sensor: store Ca2+ depletion causes STIM1 to cluster as discrete aggregates (puncta) that relocalize to ER membrane areas juxtaposed to the plasma membrane. STIM1 puncta colocalize with Orai1, and Ca2+ release–activated channels open at or near the sites where this protein nexus forms (Luik et al., 2006; Wu et al., 2006).
To further our understanding of the mechanisms that couple Ca2+ store release to Icrac gating, we investigated the features of Icrac and Ca2+ store release in a chicken B lymphocytic cell line, DT40, in which the expression of specific kinases can be genetically manipulated (Takata et al., 1994; Takata and Kurosaki, 1996; Ishiai et al., 1999). Our results indicate that Lyn or Syk tyrosine kinases are absolutely required to link the Ca2+ released from intracellular stores to Icrac gating: when the simultaneous expression of the enzymes is genetically suppressed, Icrac is not activated by the release of intracellular Ca2+ stores. This effect is reversible and specific and can be mimicked by the application of drugs that suppress kinase enzymatic activity in wild-type cells. The kinase effects do not arise from changes in the expression levels of either Orai1 or STIM1. Moreover, the release of Ca2+ does not change the level of the phosphorylation of Syk, a measure of its enzymatic activity, and only small changes are detected in Lyn. Therefore, tonic signals that generate small pools of phospho-Lyn and -Syk appear to be required for Icrac activation.
To investigate the role of tyrosine kinases in the molecular mechanisms underlying Icrac activation, we studied the electrophysiological properties of wild-type and genetically modified DT40 cells engineered to suppress the expression of specific proteins of interest.
Features of Icrac in wild-type DT40 cells
We measured voltage-clamped membrane currents at room temperature and tested the response to various agents that cause the release of Ca2+ from intracellular stores: (1) thapsigargin (TG) blocks the constitutively active Ca2+ transport into ER stores mediated by the ATP-dependent Ca2+ pump (sarcoplasmic/ER Ca2+ ATPase); (2) M4, a mAb, specifically activates the B cell antigen receptor (BCR), which, in turn, activates PLCγ and elevates cytoplasmic IP3; (3) IP3 delivered via the tight-seal electrode, which causes Ca2+ release through interaction with its specific receptor on the ER membrane; and (4) ionomycin, an ionophore that renders membranes permeable to divalent cations. Each of the four reagents slowly activated a stationary inward membrane current at −90 mV (Fig. 1 and Table I). This current was unambiguously identified as Icrac because of its biophysical features (Lewis and Cahalan, 1989; Hoth and Penner, 1992; McDonald et al., 1993; Zweifach and Lewis, 1993).
The characteristics of Icrac that we report were measured at various times, starting 120 s after first adding the stimulant and as long as 10 min thereafter. We did not observe remarkable differences throughout this time. We did not measure in detail the time course of Icrac development in our experiments, but a previous study indicates that IP3 fully activates the current within 150 s (Kiselyov et al., 2001), and TG activates the current 60–120 s after adding the stimulant (see Fig. 6 and the section Effect of TG on cytoplasmic Ca2+…). We identified the current activated by the various stimulants as Icrac because (1) its I-V curve is inwardly rectifying and has a reversal potential between 40 and 50 mV; (2) it activates with a substantial delay relative to the moment of drug delivery; (3) it exhibits an anomalous mole fraction effect between Na+ and Ca2+ ions; (4) in the presence of divalent cations, it strongly selects divalent over monovalent cations but does not distinguish among some divalent cations (Ba2+ is permeant, although somewhat less than Ca2+); and (5) it is blocked by Cd2+.
The ion selectivity features of Icrac activated by IP3 are illustrated in Fig. 2. Current was activated by IP3 delivered by diffusion from the lumen of the tight-seal electrode. In the same cell, we first measured the I-V characteristics of the IP3-dependent currents in normal high Ca2+ (10 mM) bathing solution and then in solutions free of divalent cations (with 2 mM EGTA), with Ca2+ replaced by 10 mM Ba2+. As expected for authentic Icrac, Ca2+ is permeable, Ba2+ is less permeable than Ca2+, and the current is blocked by Cd2+. Moreover, there is an anomalous mole fraction effect that explains the large current amplitude in the absence of external Ca2+. We made the same observation in 10 other cells. Our results reproduce the features of Icrac in DT40 cells previously reported by others (Kiselyov et al., 2001; Mori et al., 2002).
Features of Icrac in tyrosine kinase–deficient DT40 cells
We examined membrane currents activated by intracellular Ca2+ release in DT40 cell lines in which the expression of several tyrosine kinase enzymes was suppressed by homologous recombination. We studied cells that failed to express Syk, Lyn, and Btk as well as cells lacking the adaptor molecule Blnk (B cell linker protein). Typical results in cells activated with IP3 loaded in the tight-seal electrode are illustrated in Fig. 3. Deletion of any one of the various tyrosine kinases alone did not eliminate the electrical response to IP3 or change the quantitative feature of Icrac (Table II). In contrast, when Lyn and Syk kinases were simultaneously deleted, IP3 stimulation entirely failed to elicit Icrac (Fig. 3 and Table II). That is, IP3-dependent Icrac activation appears to require the normal function of either Lyn or Syk kinases.
The effect reported in Fig. 3 (and Table II) could simply be a consequence of a change in IP3 sensitivity. After all, the concentration of IP3 required for the activation of Icrac is greater than that necessary to release Ca2+ from intracellular stores (Parekh and Penner, 1997). Furthermore, the stores that release Icrac-activating Ca2+ are but a specific subset of all intracellular Ca2+ stores in the cell (Hofer et al., 1998). We investigated the effectiveness of IP3 over a large concentration range (Fig. 4) and found that IP3 failed to activate Icrac even at the highest concentration tested (360 μM). Thus, failure of Syk- and Lyn-deficient cells to respond to IP3 is not simply caused by a shift in the sensitivity of Icrac to IP3.
Absence of Lyn and Syk specifically prevents Icrac channel gating
Failure of IP3-dependent Icrac activation could arise from one or all of the following mechanisms: failure of the IP3 receptor, failure of Ca2+ release from intracellular stores, or failure of the link between Ca2+ store release and Icrac activation. To distinguish among these possibilities, we investigated the effects of TG on the membrane current. TG inhibits the sarcoplasmic/ER Ca2+ ATPase pump and, thus, causes the release of Ca2+ from intracellular stores.
Fig. 5 illustrates voltage-clamped currents measured in the presence of TG in wild-type and Lyn- and Syk-deficient cells. TG activates Icrac in wild-type cells but failed to do so in the Lyn- and Syk-deficient ones (Fig. 5 and Table II). The concentration of TG we tested, 1 μM, is ∼10-fold higher than the concentration typically sufficient to saturate the effect (Premack et al., 1994). The functional defect in Lyn- and Syk-deficient cells is not in the release of Ca2+.
Effect of TG on cytoplasmic Ca2+ in Lyn- and Syk-deficient DT40 mutant cells
In wild-type DT40 cells bathed in normal Ringer's solution, TG stimulation causes a sustained elevation of cytoplasmic free Ca2+ that slowly reaches a peak in 100–150 s (Fig. 6 A; Broad et al., 2001; Ma et al., 2001; Mori et al., 2002). The sustained elevation arises from the sum of two sequential events: (1) the release of Ca2+ from intracellular stores as a result of the TG inhibition of Ca2+ pumps and (2) the influx of Ca2+ from the extracellular space through Icrac channels activated by the initial rise in free Ca2+. To resolve Ca2+ store release from membrane Ca2+ influx, it is common to measure the TG-dependent Ca2+ response in the absence of extracellular Ca2+ (4 mM EGTA added to the DT40 Ringer's solution; ∼100 nM of free Ca2+). Under this condition, TG causes an increase in cytoplasmic Ca2+ (released from ER) that is smaller in amplitude and much shorter in duration than the changes observed while in the continuous presence of extracellular Ca2+ (Fig. 6 B).
We explored the functional role of Lyn and Syk on TG-dependent changes in cytoplasmic free Ca2+. In Lyn- and Syk- deficient cells maintained in normal Ringer's solution, TG stimulation caused a small and transient increase in cytoplasmic free Ca2+ (Fig. 6 C) rather than the large and sustained change observed in normal cells (Fig. 6 A). On the other hand, the response to TG in mutant cells suspended in a Ca2+-free medium (Fig. 6 D) was not different from that of normal cells (Fig. 6 B): TG caused the expected small and transient increase in free Ca2+. We made similar observations in six additional experiments. That is, in Lyn- and Syk-deficient cells, the TG-dependent release of Ca2+ from intracellular stores is not different from that in normal cells, but the absence of tyrosine kinases causes a failure of Icrac channel activation.
The functional defect in Lyn- and Syk-deficient cells is rescued by heterologous expression of the enzymes
The failure of Icrac activation in Lyn- and Syk-deficient cells could arise specifically from the absence of tyrosine kinase activity but also from an unintended consequence of the genetic manipulation. If the failure of Icrac gating depends specifically on the expression of active kinases, restoring enzyme expression in the mutant cells should restore the wild-type–like Ca2+ response.
We tested this thesis by transfecting lyn/syk−/− cells with DNA constructs that command the homologous expression of the Lyn and Syk proteins alone or in combination. Furthermore, the DNA constructs commanded the expression of a reporter GFP in tandem with the kinase. This tandem expression allowed us to identify successfully transfected cells, which are expected to express the kinases, because they also fluoresced under chromatically appropriate illumination. We used an imaging instrument to measure cytoplasmic Ca2+, an alternative to the mass measurement reported in the previous section and in Fig. 6, to permit us to analyze the cytoplasmic Ca2+ in a population consisting of more than a single cell type. We could distinguish the response of reconstituted cells (fluorescent cells expressing GFP) from that of Lyn- and Syk-deficient ones that had been identically treated but did not express the kinases.
Cells were loaded with Fura2, and the emission intensity of individual images, which were excited at 340 and 380 nm, was measured. The ratio of these intensities reports free Ca2+ concentration. The ratio in individual cells was averaged. The time course of the mean change in free cytoplasmic Ca2+ in a cell batch is illustrated in Fig. 7 (each cell averaged). Wild-type cells were initially bathed in normal Ringer's solution. They were then locally superfused with a Ca2+-free Ringer's solution (4 mM EGTA) also containing 1 μM TG. A rise in the emission intensity ratio (F340/F380) revealed the expected rise in free Ca2+ as a result of release from intracellular stores (Fig. 7 A). The superfusing solution was then switched to one containing the same 1 μM TG but with added Ba2+ (10 mM total and 8 mM of free Ca2+). The addition of Ba2+ caused a second transient in the F340/F380 emission ratio (Fig. 7 A). The second transient arises from binding to Fura2 of Ba2+ that flows into the cells through the Icrac channels activated by the TG-dependent Ca2+ transient (Vazquez et al., 2002). Ba2+ was used to ensure that changes in Fura2 fluorescence arose specifically from the activation of Icrac (and associated Ba2+ influx), not from Ca2+ flux originating in unrecognized cellular sources or influx through other Ca2+-permeable ion channels (see Fig. 9; compare this data with Vazquez et al., 2002).
In contrast, in Lyn- and Syk-deficient cells tested with the same protocol, TG/EGTA Ringer's solution caused the same initial transient release of Ca2+ from the ER observed in wild-type cells, but adding Ba2+ to the extracellular medium did not change Fura2 cytoplasmic fluorescence (Fig. 7 B). That is, in Lyn- and Syk-deficient cells, Icrac channels are not active and, thus, extracellular Ba2+ cannot flow into the cell to interact with Fura2. The same observation was made in 10 repeat experiments. Just like the findings in the measurements of cell suspensions, these results verify that lack of tyrosine kinase activity causes a loss of the linkage between Ca2+ store release and Icrac activation.
We measured TG-dependent Ca2+ changes in transiently reconstituted mutant cells that expressed tyrosine kinases 24 h after DNA transfection. In image analysis, we distinguished cells by their GFP fluorescence: GFP-negative cells were mutant lyn/syk−/− cells, whereas GFP-positive ones expressed the specific kinase of interest. In general, ∼20% of the cells expressed GFP under our experimental protocol, but the level of expression varied among cells. We measured Ca2+ only in the cells expressing the highest level of GFP, ∼5–10% of the cells. Typical results are illustrated in Fig. 7. The same observations were made in three different experiments for each transfection condition.
Kinase-deficient cells that did not reconstitute the expression of tyrosine kinases but were nonetheless subjected to the electroporation protocol of transfection behaved indistinguishably from untreated Lyn- and Syk-deficient cells. Stimulation with TG/EGTA Ringer's solution caused a release of Ca2+ from intracellular stores, but Icrac was not activated, and extracellular Ba2+ failed to flow into the cells (Fig. 7, C–E). In contrast, the expression of kinases Lyn (Fig. 7 C) or Syk (Fig. 7 D) alone or together (Fig. 7 E) restored wild-type–like behavior in these cells. TG caused intracellular Ca2+ release, which activated Icrac channels that, in turn, sustained Ba2+ influx.
Examination of data in Fig. 7 reveals that the expression of tyrosine kinases restored Icrac gating, but it should be noted that the Ba2+-dependent fluorescence changes are not quantitatively similar to the response in wild-type cells and, in fact, differed among the various transfected cell batches tested. This is not surprising because we could not control the level of expression of the kinases in the transfected cells, and, moreover, there is no reason to expect that these levels would be identical to those in wild-type cells or amid different cell batches. Nonetheless, restoring kinases in Lyn- and Syk-deficient cells always rescued the cell phenotype, indicating that functional defects in the deficient cells arise specifically from the lack of these kinases. As we observed in electrical studies of mutant cell lines (Fig. 3), the presence of either Lyn or Syk alone is sufficient to obtain normal function.
Mechanisms of Lyn and Syk action
Although the detailed molecular mechanisms whereby Lyn and Syk function to link Ca2+ release to Icrac gating remain to be determined, we tested two possibilities: (1) that these mechanisms act on the events linking Ca2+ release and channel gating and (2) that the enzymatic activity of the kinases change in the course of Ca2+ store release and/or channel gating.
A parsimonious thesis suggests that kinases act on the molecular events that link Ca2+ release and channel gating. However, it can be argued that the action of kinases is not direct but arises from long-term effects of enzyme activity on events such as gene expression or cellular protein trafficking. To assess whether the kinase action is immediate and direct or not, we tested the effectiveness of drugs known to act as specific tyrosine kinase inhibitors. If the role of kinase activity is direct, drug treatment should have the same effect on TG-dependent Icrac activation as on the genetic suppression of enzyme expression.
We tested the effectiveness of two well-established kinase inhibitors: LavendustinA, a specific tyrosine kinase inhibitor that does not distinguish among enzyme types (Onoda et al., 1989; Ping et al., 1999), and PP2, a selective inhibitor of the Src family tyrosine kinase. We first used an electrophysiological assay to test the effect of LavendustinA on TG-dependent Icrac activation. Cells were treated with 10 μM LavendustinA starting 10 min before the functional assay and throughout the assay. Electrophysiological data in Fig. 8 are typical of experiments that were repeated five times. In normal cells, TG caused activation of an inward current at −90 mV that developed slowly and reached its maximum steady amplitude ∼60–100 s after initiating TG stimulation (Fig. 8 A, inset). The I-V curve of this TG-dependent current was measured by taking the difference in I-V curves before TG stimulation (time 0 data point) and 85 s after initiating TG stimulation (time 165 s data point). The difference in the I-V curve thus computed is typical of Icrac (Fig. 8 A). In contrast, in LavendustinA-treated cells, TG failed to activate an inward current at −90 mV (Fig. 8 B, inset). In three different LavendustinA-treated cells, we followed the current for up to 20 min after initiating TG stimulation and did not observe any substantial activation. Also, in the presence of LavendustinA, there was no difference in the I-V curves measured at time 0 and at 165 s (Fig. 8 B).
Having demonstrated that LavendustinA specifically prevents TG-dependent Icrac activation, we also tested the drug effect in a single-cell Ca2+ imaging assay of Fura2-loaded cells. As illustrated in Fig. 7 A, in wild-type cells and in the absence of extracellular Ca2+, TG causes a transient increase in cytoplasmic Ca2+ because of the cation release from intracellular stores. If Ba2+ is then added to the extracellular medium, the cation flows into the cell only if Icrac channels have been activated and causes a second, prominent increase in the Fura2 emission intensity ratio. In contrast, in Lyn- and Syk-deficient cells, Icrac is not activated, and, therefore, extracellularly added Ba2+ does cause any changes in Fura2 fluorescence (Fig. 8 C). The same phenotype is observed in LavendutinA-treated cells: TG causes an initial rise in cytoplasmic Ca2+, but the addition of extracellular Ba2+ does not cause any change in the Fura2 emission intensity (Fig. 8 D).
We tested the effect of PP2, a potent and specific inhibitor of the Src kinase family (that is, a specific inhibitor of Lyn), on TG-dependent Ca2+ fluxes in suspensions of cells loaded with the Ca2+ indicator Indo1 (Fig. 9). Again, in wild-type cells, 1 μM TG delivered in the presence of EGTA caused a transient increase in cytoplasmic free Ca2+, which was followed by a much larger and sustained change in concentration when Ca2+ was added to the extracellular medium (Fig. 9 A). In Lyn- and Syk-deficient cells, TG caused the initial rise in cytoplasmic Ca2+, but, because Icrac channels are not activated, the added external Ca2+ caused only a modest change in cytoplasmic free Ca2+ when compared with the normal cells (Fig. 9 B). Treatment of wild-type cells with 20 μM PP2 caused a functional phenotype that was indistinguishable from that of the Lyn- and Syk-deficient cells. In PP2-treated cells, TG added in a medium free of divalent cations caused a transient increase in cytoplasmic Ca2+, but the addition of Ca2+ to the extracellular medium did not cause any further changes in cytoplasmic free Ca2+ (Fig. 9 C). These pharmacological findings together with the results of genetic manipulations demonstrate that Lyn activity is required for Icrac gating even if Syk activity is unchanged. That is, the tyrosine kinases are required, and Lyn (and other Src kinases) is likely upstream of Syk in this pathway.
Biochemical assay of Lyn and Syk phosphorylation activity
We measured the level of phosphorylation of Lyn and Syk kinase proteins. This autophosphorylation is a parameter that can reflect kinase activity. To increase sensitivity, all tyrosine-phosphorylated proteins in cell lysates were first immunoprecipitated with panspecific phosphotyrosine mAbs. These immunoprecipitates were then probed in a Western blot using antibodies that bound to specific phosphotyrosine residues in Lyn and Syk (Fig. 10 A).
TG stimulation had different effects on the two enzymes. Syk exhibited a low level of phosphorylation before stimulation, which was not detectably changed by TG. In contrast, Lyn had a low level of phosphorylation before stimulation, and TG increased this level ∼2.5-fold. As a control, we used immunoprecipitates isolated from Lyn- and Syk-deficient cells, and, as expected, there was no detectable signal for either of these kinases (Fig. 10 A). Also, we measured the phosphorylation of these kinases in BCR-stimulated cells, and, consistent with previous observations (Mizuguchi et al., 1992; Ma et al., 2001), we observed a stimulated phosphorylation of approximately sevenfold for each kinase.
Expression levels of STIM1 protein and of STIM1 and Orai1 mRNA in Lyn- and Syk-deficient cells
Recent experiments have demonstrated that two membrane-associated proteins, STIM1 (Liou et al., 2005; Roos et al., 2005; Zhang et al., 2005) and Orai1 (Feske et al., 2006, Vig et al., 2006b), participate in the events linking Ca2+ store release to Icrac channel activation in lymphocytes. It is possible, then, that the effects of Lyn and Syk arise from their action on these proteins. Future work must explore whether these proteins are phosphorylated and what role tyrosine kinases might play. In this study, we tested the possibility that Syk and Lyn act as regulators of the expression of either or both of these proteins. We tested this thesis by assaying the protein expression level of STIM1 and the mRNA expression level of both STIM1 and Orai1. We find that none of these is detectably different in wild-type and mutant cells.
STIM1 protein levels were measured in Western blots. Individual lanes in SDS-PAGE were loaded with the same total protein amount. The relative amount of protein identified with an STIM1 antibody was assessed by chemiluminescence. Samples from wild type and the following mutants were run alongside each other in the same gel: lyn−/−, syk−/−, lyn/syk−/−, blnk−/−, and IP3−/− receptor (Fig. 10 B). The expression levels of tubulin in each sampled lane were also measured to confirm that the amount of protein loaded in each lane was essentially the same (Fig. 10 B). There was no detectable difference in the level of STIM1 protein between wild type and the various mutant cells tested.
mRNA levels of STIM1 and Orai1 were assayed with semiquantitative RT-PCR. Total cell RNA was used to synthesize cDNA, which, in turn, was amplified using PCR. The amount of gene-specific DNA produced under identical reaction conditions was compared for STIM1 and Orai1-specific primers in wild type and in the following mutants: lyn−/−, syk−/−, lyn/syk−/−, blnk−/−, and IP3−/− receptor. PCR products were loaded at various dilutions into the lanes of an agarose gel and separated by size. The level of β-actin mRNA was also assessed in each of the cell samples. The amount of DNA in each well was quantified by the intensity of fluorescence in images of ethidium bromide–stained gels (Fig. 10 C). There was no detectable difference in the STIM1 and Orai1 mRNA levels between wild type and the various mutant cells tested.
In B lymphocytes, binding of antigen or antireceptor antibody to the BCR complex recruits and activates tyrosine kinases as part of the events that lead to the production of IP3 (Rawlings, 1999; Wollscheid et al., 1999; for reviews see Kurosaki, 2000; Wienands, 2000). Results presented in the current study indicate that these kinases also play a role in the link between the IP3-dependent release of Ca2+ from ER intracellular stores and activation of the Icrac ion channels in the plasma membrane. We compared the electrophysiological features of Icrac and changes in free cytoplasmic Ca2+ between wild-type and Lyn- and Syk-deficient DT40 cells and found that cells simultaneously lacking Lyn and Syk tyrosine kinases fail to activate Icrac after the release of Ca2+ stores. Interestingly, the genetic deletion of either enzyme alone did not prevent Icrac activation, revealing redundancy in the biological function of these kinases in this response. The physiological effects on Icrac and cytoplasmic Ca2+ flux of pharmacological blockers of these enzymes leads us to suggest that the enzymes may operate in a sequential cascade with the Src family enzyme (Lyn) upstream from Syk. The failure to activate Icrac in the Lyn- and Syk-deficient cells reflects that protein phosphorylation is a required event in the mechanism that couples the emptying of Ca2+ stores to channel activation or in the function of the channels themselves. These alternatives cannot now be unequivocally discerned.
There are no previous studies on the electrical properties of lymphocytes with the genetically suppressed expression of tyrosine kinases. However, the hypothesis that tyrosine kinases may play a role in the link between Ca2+ store release and Icrac activation has been proposed previously (Lee et al., 1993; Sargeant et al., 1993a,b; Rosado et al., 2000). Lymphocytes of mice in which the Btk-related Itk tyrosine kinase is knocked out respond to T cell receptor activation with a normal release of intracellular Ca2+ stores, but Icrac activation by TG is less than in wild-type animals (Liu et al., 1998). In addition to observations in animals subjected to genetic manipulation, a possible role for tyrosine kinases in lymphocyte Ca2+ transduction paths has been suggested by experimental observations of some effects of the pharmacological inhibition of tyrosine kinase activity (Whisler et al., 1991; Mizuguchi et al., 1992; Schieven et al., 1993; Nagao et al., 1996).
Currently, the mechanisms that couple Ca2+ store release to Icrac gating fall into two main categories: (1) physical coupling mechanisms propose that activation of the IP3 receptors are directly linked through conformational changes to the opening of Icrac channels. An alternative form of physical coupling, the vesicle fusion hypothesis proposes a fusionlike event between intracellular vesicles and the plasma membranes that delivers Icrac channels or their activators to the plasma membrane. Vesicle fusion is mediated by SNARE proteins. (2) Messenger mechanisms propose that the released Ca2+ is detected by particular cytoplasmic Ca2+ sensor molecules that, in turn, gate channel activity.
The direct coupling hypothesis was developed to explain findings that Icrac fails to occur after pharmacological treatments expected to disrupt cytoskeletal elements (for example, cytochalasin D [Patterson et al., 1999], calyculin A [Patterson et al., 1999], 2APB [Patterson et al., 1999], or clostridium C3 transferase injection [Yao et al., 1999]). Direct coupling supposes that molecular receptors in the ER membranes, IP3 receptors, or ryanodine receptors are linked to Icrac activation through conformational changes. In lymphocytes in particular, this hypothesis is unlikely because mutant DT40 cells lacking all IP3 receptor isoforms exhibit normal TG-activated Icrac (Prakriya and Lewis, 2001) and ryanodine receptor knockout animals have normal T cell activation (Takeshima et al., 1996).
The vesicle fusion hypothesis (for review see Rosado et al., 2005) was developed to explain the fact that interference with the function of all or even some of the various proteins that constitute the SNARE complex can block TG-dependent Ca2+ elevation. Although it is evident that SNARE protein can alter the cell's Ca2+ response, it remains questionable whether their specific activation by IP3-linked events is required in the course of normal events linking Ca2+ store release and Icrac gating (Scott et al., 2003).
Recent studies have identified two novel proteins, STIM1 and Orai1, that participate in the signaling between store Ca2+ release and Icrac activation (Zhang et al., 2005; Feske et al., 2006; Peinelt et al., 2006). They are both integral membrane proteins located either in the ER membrane (STIM1) or in the plasma membrane (Orai1). Their simultaneous coexpression enhances Ca2+-dependent Icrac gating. These may be the first identified constituents of a larger family of proteins involved in the linkage between Ca2+ store release and channel gating. The results we present here firmly indicate that the link between the release of Ca2+ stores and Icrac activation in B cells also involves protein phosphorylation catalyzed by cytoplasmic tyrosine kinases. The target of phosphorylation remains to be identified in future work; it could be STIM1, Orai1, or other proteins yet to be discovered. However, we have demonstrated that kinase activity does not affect the expression levels of STIM1 or Orai1.
We emphasize that our results suggest that the state of phosphorylation of a target protein is important, but we have no evidence, nor do we suggest that Ca2+ sensing or channel gating is mediated by phosphorylation. We assayed kinase autophosphorylation as an index of enzymatic activity. Our biochemical results reveal that the activity of one of the critical kinases (Syk) does not change with Ca2+ release, whereas the other (Lyn) changes only by a small extent and with a slow time course. We propose that phosphorylation is required as a tonic signal that sustains the normal function of its target protein rather than a phasic signal activated in the course of Ca2+ release, Ca2+ sensing, or channel gating.
Materials And Methods
IP3, TG, ionomycin, fibronectin, antiphosphotyrosine antibody 4G10, LavendustinA, and PP2 (4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d] pyrimidine) were obtained from Calbiochem. The STIM1 antibody was purchased from BD Biosciences. M4, an anti-DT40 BCR chain mAb, was prepared as described previously (Zhao and Weiss, 1996). Indo1–acetoxy-methyl and Fura2–acetoxy-methyl were obtained from Invitrogen. Phospho-Src family (P-Y416) and phospho–Zap-70 (P-Y319)/Syk (P-Y352) antibodies were purchased from Cell Signaling Technology. pLynGFP and pSykEGPF, which were gifts from A. Defranco (University of California, San Fransisco, San Fransisco, CA), are vectors that promote the expression of Lyn/GFP or Syk/EGFP proteins.
Cell lines and solutions
DT40 cells were cultured at 37°C in 5% CO2 in RPMI 1640 with 10% FCS, 1% chicken serum, and 2 mM glutamine. Generation of every cell mutant DT40 cell line has been reported previously: lyn/syk−/− and btk−/− (Takata and Kurosaki, 1996), lyn− and syk− (Takata et al., 1994), and bln− (Ishiai et al., 1999). Mutant cells were transfected with pLynGFP or pSykEGPF by electroporation and grown for 24 h after transformation before Ca2+ imaging experiments.
DT40 Ringer's solution consisted of 145 mM NaCl, 2.8 mM KCl, 10 mM CsCl, 10 mM CaCl2, 2 mM MgCl2, 10 mM glucose, and 10 mM Hepes-NaOH, pH 7.4. Tight-seal electrodes were filled with a quasi-intracellular solution of the composition 145 mM Cs glutamate, 8 mM NaCl, 1 mM MgCl2, 2 mM MgATP, 10 mM EGTA, and added Ca2+ to yield 90 nM of free Ca2+. Free Ca2+ in the presence of the various Ca2+-binding molecules was calculated using WinMaxC.
In some experiments, we tested the ion selectivity of Icrac by measuring the current in an extracellular solution consisting of 150 mM Na methanesulfonate, 10 mM Hepes, pH 7.4, and 2 mM EGTA with varying amounts of Ca2+, Ba2+, and Cd2+. Osmotic pressure was 300 mosM (adjusted with glucose). In only these experiments, we filled the tight-seal electrode with 128 mM Cs glutamate, 3.16 mM MgCl2, and 10 mM EGTA and added Ca2+ to yield 90 nM of free ion.
Electrophysiological studies in single cells
Cells were washed free of incubation medium and resuspended at ∼105/ml in Ringer's solution. 200 μl of the cell suspension was applied onto a fibronectin-coated glass coverslip that formed the bottom of a recording chamber held on an upright microscope (Diaphot; Nikon) equipped with differential interference contrast optics. The cells were maintained at room temperature and continuously superfused with oxygenated recording solution. Drugs of interest were tested by adding them to the superfusing solution.
We measured membrane currents under voltage clamp at room temperature in the whole cell mode configuration using tight-seal electrodes produced from aluminosilicate glass (1.5 × 1.0 mm [outer diameter × inner diameter]; 1724; Corning). We used a patch-clamp amplifier (Axopatch 1B; Axon Instruments, Inc.), with which we compensated (and measured) membrane capacitance but not series resistance. Analogue signals were low-pass filtered below 1 KHz with an 8-pole Bessel filter (Frequency Devices) and digitized online at 3 KHz with FastLab (Indec Systems) or pClamp (Axon Instruments, Inc.). Bath was grounded through an Ag/AgCl electrode connected to the solution via a 1-M KCl agar bridge to avoid shifts in electrode potential as solutions changed.
Two voltage command protocols were used, steps or ramp, depending on the purpose of the experiment. In the voltage step protocol, voltage was held at 0 mV and was stepped for 160 ms to values between −100 and 90 mV in 10-mV increments. This protocol was applied immediately after attaining whole cell mode and before adding any stimulant. The I-V curve thus generated was defined as the cells leak. Cells were chemically stimulated, and the voltage step protocol was executed again 2 min thereafter. The I-V curves are corrected for nonspecific electrical leakage. This correction was carried out by measuring the I-V curve before chemical subtraction and subtracting these data from the I-V curve measured after stimulation. Continuous I-V functions were fit to discrete data points with a second degree polynomial using nonlinear least-square minimization algorithms (Origin; Microcal).
To investigate the time course of Icrac development and the effects of LavendustinA, we used a voltage ramp protocol. Voltage was held at 0 mV, and acquisition epochs were repeated continuously at 3-s intervals starting immediately after attaining whole cell mode and for as long as 25 min. In each acquisition epoch, voltage was initially stepped to −100 mV for 100 ms, and then a linear voltage ramp was applied between −100 and 100 mV over 100 ms before returning to −100 mV for another 100 ms. To determine the time course of Icrac development, we measured in each of the successive epochs the difference between the mean current amplitude at 0 mV and the amplitude 90 ms after stepping to −90 mV. The cell's I-V curve was measured directly from the ramp data.
Measurement of cytoplasmic free Ca2+ in cell suspensions
To load, Indo1 was loaded into cells by incubating ∼3 × 106 cells/ml in RPMI 1640 with 10% FBS and 3 μM Indo1–acetoxy-methyl (Invitrogen) for 1 h at 37°C. The cells were then washed three times at room temperature with a DT40 Ringer's solution consisting of 125 mM NaCl, 5 mM KCl, 1 mM CaCl2, 0.5 mM MgCl2, 1 mM Na2HPO4, 11 mM glucose, and 25 mM Hepes, pH 7.4, with 0.1% wt/vol BSA added. The last cell pellet was maintained on ice until just before measurements. To measure Ca2+, ∼2 × 106 Indo-loaded cells/ml were suspended in DT40 Ringer's solution with BSA and maintained at 37°C in a stirred cuvette in a fluorimeter (F-4500; Hitachi). Fluorescence emission intensity ratio at 400 and 500 nm (F400/F500) was measured continuously and assessed in response to stimulants injected into the stirred cuvette.
Single-cell Ca2+ imaging
In single-cell studies, we used the fluorescent dye Fura2 in a modified epifluorescent microscope. Cells were loaded by incubation in Fura2–acetoxy-methyl (Invitrogen) as described above for Indo1–acetoxy-methyl. Immediately after the last wash, cells were suspended in DT40 Ringer's solution at ∼105 cells/ml, and 200 μl was deposited on an imaging chamber held on the stage of the inverted microscope. The bottom of this chamber was a glass coverslip coated with 0.1 mg/ml poly-l-lysine, to which cells adhered. Cell fluorescence was excited at either 340 or 380 nm (narrow band interference filters selected with a rotating wheel; Sutter Instrument Co.). Cells were observed using a 20× 0.75 NA Fluor objective (Nikon), and images were captured through a 510 ± 40-nm bandpass optical window using a cooled high resolution CCD camera (MicroMAX 1300; Roper Scientific), an ST133 controller, and WinView32 software (Roper Scientific). Image pairs at 340 or 380 nm were acquired for 5 s each, separated by 0.5 s from one another. Image pairs were repeated at 30-s intervals over 20 min. To minimize fluorophore bleaching, an electronic shutter (LPS25; Vincent Associates) was used to restrict cell illumination to the time of image acquisition.
In off-line analysis (Simple PCI software; Compix, Inc.), we selected images of individual cells that met three criteria: diameter within 10% of the population mean, low resting cytoplasmic Ca2+, and no detectable movement or loss of focus over the time course of the measurements. The mean gray level (at 10 bit resolution) of the emission intensity at 340 and 380 nm was measured for each selected cell at each time point and averaged over the cell population. Absolute cell emission intensity was calculated by subtracting the mean intensity of the background.
The cell-bathing Ringer's solution in the microscope imaging chamber was exchanged throughout an experimental measurement at a rate of ∼0.5 ml/min. In addition, the medium specifically bathing cells in a selected field of view was controlled by local superfusion with solutions flowing from a 300-μm–inner diameter polyimide capillary tubing (PT Technologies) placed with its tip in the same plane as the cells and ∼300 μm away. This superfusing medium rapidly (≤200 ms) changed in the course of cytoplasmic Ca2+ level measurements. The polyimide tubing was the exit port of micromanifold (MM-6; Warner Instruments) that allowed selection among six possible solutions. The identity and duration of flow of the superfusing solution were controlled with electronic switch valves (MPS-2; WPI Instruments).
Tyrosine phosphorylation assay
1.0 × 108 DT40 cells were suspended in RPMI, stimulated by the addition of TG or M4, and sampled at different time points after stimulation. Cells were spun, lysis buffer was added, cells were spun again, and the supernatant was collected. 4G10 antibody was added to the supernatant and incubated for 1 h at 4°C on a rocker. Protein G beads were then added, and the suspension was incubated for an additional 2 h at 4°C. The beads with bound tyrosine-phosphorylated proteins were washed four times with lysis buffer and dissolved in standard SDS sample buffer with DTT. Proteins were separated in gradient 5–15% SDS-PAGE, blotted onto nitrocellulose, and reacted with antibodies. The reaction product was located on the blot with secondary antibodies linked to HRP and reacted with the chemiluminescent substrate Luminol. Within 1 min, luminescent images were acquired with a cooled high resolution CCD camera (MicroMAX 1300; Roper Scientific). In off-line image analysis (Simple PCI software; Compix), we measured the absolute gray level of each luminescent band.
Expression levels of STIM1 and Orai1
Expression levels of STIM1 protein were assessed in a Western blot using anti-STIM1 primary antibody and a secondary antibody linked to HRP. SDS-PAGE lanes were loaded with 200 μg of protein and transferred onto a nitrocellulose membrane, which was then incubated for 2 h at room temperature with the primary antibody dissolved in 3% (wt/vol) BSA in Tween-PBS. After exhaustive washing, the membrane was further incubated with the secondary antibody for 1 h at room temperature. The extent of antibody binding was assessed using chemiluminescence (PerkinElmer).
The expression levels of STIM1 and Orai1 mRNA were assessed with semiquantitative RT-PCR. Total RNA from wild-type and mutant DT40 cells was purified using the RNeasy kit (QIAGEN). First-strand cDNA was produced using the Superscript III first-strand synthesis system (Invitrogen). Serial 1:5 dilutions of the cDNAs were then used in PCR reactions using the primers of the following sequences: STIM1, forward (GGGACTGTGCTGAAGATGACAGACCG) and reverse (TCCTTGGCCACCAGGAGCTGCTTCTC); Orai1, forward (GGCTTCGCCATGGTGGCCATGGTAGA) and reverse (TCACCCAACACAGCAACACCACCTC); and β-actin, forward (CCCTCCCCCATGCCATCCTC) and reverse (CCGGTTTAGAAGCATTTGCGG).
We thank T. Brdicka and M. Mollenauer for their invaluable advice and technical assistance as well as Z. Xu, A. Picones, M.-P. Faillace, C. Paillart, and T. Rebrik for helpful discussions. We also thank Drs. David Julius and Barbara Ehrlich for their comments on the manuscript.
This research was supported by National Institutes of Health grant 5F32EY06823 to S.C. Chung, grant CA72531 to A. Weiss, and grant EY11349 to J.I. Korenbrot.
Abbreviations used in this paper: BCR, B cell receptor; IP3, inositol 1,4,5-trisphosphate; STIM1, stromal interaction molecule 1; TG, thapsigargin.