Inositol tetrakisphosphate as a frequency regulator in calcium oscillations in HeLa cells.

Cellular signaling mediated by inositol (1,4,5)trisphosphate (Ins(1, 4,5)P(3)) results in oscillatory intracellular calcium (Ca(2+)) release. Because the amplitude of the Ca(2+) spikes is relatively invariant, the extent of the agonist-mediated effects must reside in their ability to regulate the oscillating frequency. Using electroporation techniques, we show that Ins(1,4,5)P(3), Ins(1,3,4, 5)P(4), and Ins(1,3,4,6)P(4) cause a rapid intracellular Ca(2+) release in resting HeLa cells and a transient increase in the frequency of ongoing Ca(2+) oscillations stimulated by histamine. Two poorly metabolizable analogs of Ins(1,4,5)P(3), Ins(2,4,5)P(3), and 2,3-dideoxy-Ins(1,4,5)P(3), gave a single Ca(2+) spike and failed to alter the frequency of ongoing oscillations. Complete inhibition of Ins(1,4,5)P(3) 3-kinase (IP3K) by either adriamycin or its specific antibody blocked Ca(2+) oscillations. Partial inhibition of IP3K causes a significant reduction in frequency. Taken together, our results indicate that Ins(1,3,4,5)P(4) is the frequency regulator in vivo, and IP3K, which phosphorylates Ins(1,4, 5)P(3) to Ins(1,3,4,5)P(4), plays a major regulatory role in intracellular Ca(2+) oscillations.

In response to stimulation by agonists, many non-excitable eukaryotic cells release Ca 2ϩ from intracellular stores to regulate cellular processes in an oscillatory manner (1)(2)(3). This Ca 2ϩ release is mediated by inositol 1,4,5-trisphosphate (Ins(1,4,5)P 3 ) 1 (4) or cyclic ADP-ribose (5). Many models have been proposed to describe the mechanism of intracellular Ca 2ϩ oscillations induced by various agonists based on positive and negative feedback from Ca 2ϩ (1, 2, 4, 6 -12). Nevertheless, it is generally agreed that the mechanism appears to be cell typedependent and not fully understood. We have shown that repetitive phosphorylation/dephosphorylation of the InsP 3 receptor and the Ca 2ϩ -ATPase pump catalyzed by calmodulindependent protein kinase II and a calyculin A and okadaic acid-inhibitable protein phosphatase are responsible for sustaining Ca 2ϩ oscillations in HeLa cells when stimulated with histamine (13). In agreement with others, we found that the amplitude of oscillating Ca 2ϩ spikes is relatively independent of the agonist concentration, whereas its frequency increases as a saturation function with respect to the agonist concentration. Thus, the extent of the agonist-mediated effects must reside in their ability to regulate the oscillation frequency. Recent studies using a Ca 2ϩ clamp technique (14) and a caged InsP 3 method (15) revealed that oscillations can reduce the Ca 2ϩ threshold and that variation of its frequency affects both the specificity and optimization for gene expression. Therefore, it is essential to understand by which metabolites and by which mechanism the frequency is regulated. We have previously pointed out that such regulation must involve a factor generated by a reaction related to the agonist-receptor interaction to account for the agonist concentration dependence (13). Ins(1,4,5)P 3 has been shown to undergo oscillations with Ca 2ϩ in REF52 fibroblast cells (16) and in canine kidney epithelial cells (17). To search for the frequency regulator(s), we investigated the effects of various inositol polyphosphates on both the intracellular Ca 2ϩ release in HeLa cells and on the frequency of ongoing Ca 2ϩ oscillations in histamine-stimulated cells. Our results indicate that Ins(1 3,4,5)P 4 is a frequency regulator in vivo and Ins(1,4,5)P 3 3-kinase, which catalyzes the conversion of Ins(1,4,5)P 3 to Ins(1,3,4,5)P 4 plays a major regulatory role in intracellular Ca 2ϩ oscillation.
Loading of Fluo-3 and Monitoring of Ca 2ϩ Signal-HeLa cells, grown to 70 -80% confluency, were detached from dishes by 0.25% trypsin with 1 mM EDTA for 2 min and then washed at least twice with and resuspended in Buffer A (10 mM Hepes, pH 7.4, 121 mM NaCl, 5 mM KCl, 1.8 mM CaCl 2 0.8 mM MgCl 2 , and 5 mM D-glucose). After 0.5 h of incubation at 37°C with 2 M Fluo-3/AM (a fluorescent Ca 2ϩ indicator), the cells were washed three times and resuspended in fresh buffer. They were either used immediately or kept at 4°C for use within 4 h. A fluoromicroscope equipped with an image analysis setup to record at wavelengths longer than 510 nm was used to monitor the Ca 2ϩ signal in HeLa cells excited at 480 nm. The images were recorded on videotape at 30 frames/s and transferred to a computer at 1 frame/s. The light intensity in each frame was calculated by the computer and included the correction for dye bleaching. Experiments were carried out at room temperature, 23 Ϯ 2°C.
Electroporation-The home-built electroporator used was described previously (16). Briefly, electric pulses of desired amplitude and pulse width were applied across cells placed in a home-built electroporation chamber. The chamber was constructed with two polished stainless steel electrodes fixed on a 75 ϫ 25-mm microscope glass slide. The inter-electrode separation was 1 mm. A suspension of cells was placed between the electrodes, and the chamber was mounted on an inverted fluorescence microscope (Ziess, Axiovert 100TV). The cells were allowed to settle in the glass chamber, and a field of view with about 5-10 cells was chosen prior to field application and initiation of experiments. Fluorescence images were acquired with an intensified camera and recorded on a VHS tape for analysis. The loading of anti-mouse IgG or * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
anti-IP3K antibody into a population of cells was performed using a commercial electroporation cuvette (Bio-Rad) with an inter-electrode separation of 0.2 cm.
In the present study, a constant electric field of 1.2 kV/cm with a pulse width of 200 s was used for electroporating inositol polyphosphates, and 500 s was used for electroporating antibodies. The efficiency of electroporation was determined to be about 10% of the bath concentration for inositol polyphosphate using 3 H-labeled Ins(1,4,5)P 3 as a probe and about 0.3% for antibody using fluorescein-labeled IgG as a probe.

RESULTS AND DISCUSSION
For this study, HeLa cells were chosen because they contain mainly the InsP 3 -sensitive Ca 2ϩ stores, and histamine consistently induces Ca 2ϩ oscillations in more than 80% of these cells (13). Although ryanodine receptors have been found in HeLa cells, their population is so low that an acute opening of these receptors has little effect on the concentration of intracellular Ca 2ϩ (18). The electroporation technique (19) was used to introduce membrane impermeable inositol polyphosphates and the monoclonal anti-IP3K antibody into the cells. To ensure that observed Ca 2ϩ signals originated from the intracellular Ca 2ϩ stores, all experiments involving the use of electroporation were carried out in the absence of extracellular Ca 2ϩ . Fig.  1 shows the Ca 2ϩ oscillation pattern in HeLa cells induced by 200 M histamine observed in a Ca 2ϩ -free buffer that also contained 10 M EGTA. This addition of 10 M EGTA was to ensure the absence of free extracellular Ca 2ϩ as indicated by the fact that no Ca 2ϩ signal was observed when the cells were electroporated in the absence or presence of a low Ins(1,4,5)P 3 concentration (see Fig. 2). We found that, unlike the oscillation pattern in the presence of extracellular Ca 2ϩ that exhibits constant amplitude and frequency, the absence of external Ca 2ϩ causes a small reduction in the amplitude and frequency. Furthermore, under our electric pulse conditions (1.2 kV/cm, 200 -500 s), which induced rapid transient membrane pore formation (20), we observed no visible effects on the intracellular Ca 2ϩ signal (data not shown). Fig. 2 shows that electroporation in the presence of 2.5 M Ins(1,4,5)P 3 causes the release of a Ca 2ϩ spike from intracellular stores. The Ca 2ϩ signal was not observed when the concentration was reduced to 1.2 M or in the absence of Ins(1,4,5)P 3 . Because the efficiency of the electroporation is about 10% under our experimental conditions (1.2 kV/cm, 200 s), the resulting cytosolic concentration of Ins(1,4,5)P 3 was about 0.25 M when Ca 2ϩ was mobilized from the intracellular stores. Using this method, we probed a series of inositol polyphosphates, at bath concentrations ranging from 2 to 150 M, for their capacity to mobilize the release of intracellular Ca 2ϩ and to alter the frequency of histamine-initiated Ca 2ϩ oscillations. It should be pointed out that, similar to Ins(1,4,5)P 3 , the Ca 2ϩ releases induced by Ins(1,3,4,5)P 4 and Ins(1,3,4,6)P 4 were observed with a bath concentration of 3.5 and 2.0 M, respectively. Therefore, the observed Ca 2ϩ release cannot be because of the potential 5% contamination of Ins(1,4,5)P 3 in Ins(1,3,4,5)P 4 or Ins(1,3,6)P 3 in Ins(1,3,4,6)P 4 samples, because, under our conditions, a minimum bath concentration of 1.2 M is required for Ins(1,4,5)P 3 to cause a Ca 2ϩ release, and Ins(1,4,5)P 3 is a much more potent Ca 2ϩ mobilizer than Ins(1,3,6)P 3 (21). The Ca 2ϩ release data shown in Table I are similar to those reported for SH-SY5Y neuroblastoma cells (22) and bovine adrenal microsomes (23), except that in HeLa cells Ins(1,4,5)P 3 and Ins(1,3,4,5)P 4 exhibit a similar potency for inducing Ca 2ϩ release based on the concentration-dependent study (data similar to those shown in Fig. 2). In addition, Ins(1,3,4,6)P 4 , 2,3-didoeoxy-Ins(1,4,5)P 3 , and a high concentration of Ins(2,4,5)P 3 can also induce Ca 2ϩ release in resting cells.
To probe the effect of inositol polyphosphates on the frequency of Ca 2ϩ oscillations, we electroporated the histaminetreated cells with 20 -35 M of inositol polyphosphate (150 M for Ins(2,4,5)P 3 ), which gave a 2-3 M elevation inside the cells. This elevated concentration is within the range found in ago-

TABLE I Effects of inositol polyphosphates on the intracellular Ca 2ϩ release in resting cells and on the frequency of ongoing oscillations in HeLa cells
The data were obtained from three independent experiments. n is the number of cells observed, and -fold increase represents the increase in frequency, which lasted for 1-2 min, relative to that of ongoing oscillations prior to the electroporation of inositol phosphates. nist-stimulated cells (24). Although, as indicated in Fig. 2, the concentrations in the bath required to induce Ca 2ϩ release are relatively low, i.e. about one-tenth of those shown in Table I, a higher concentration range was used to observe an increase in the frequency for a duration of 1 to 2 min (Table I). This finding suggests that the frequency regulator(s) may be generated due to the added inositol polyphosphates, and it undergoes intracellular clearance. Fig. 3A shows that no effect was observed on the frequency of histamine-stimulated Ca 2ϩ oscillation when the cells were electroporated in the presence of 150 M Ins(2,4,5)P 3 . However, in similar cells, electroporation in the presence of 150 M Ins(1,3,4,5)P 4 causes a transient increase in the frequency (Fig. 3B). The observed Ins(1,3,4,5)P 4 -induced frequency increase and the duration of the transient high frequency phase is concentration-dependent. Among the inositol polyphosphates tested (see Table I), only Ins(1,4,5)P 3 and Ins(1,3,4,5)P 4 enhance the frequency of the ongoing Ca 2ϩ oscillations effectively. Ins(1,3,4,6)P 4 also causes an increase in frequency but with much reduced efficiency. The two poorly metabolizable analogs of Ins(1,4,5)P 3 , Ins(2,4,5)P 3 and 2,3-dideoxy-Ins(1,4,5)P 3 , which can induce a Ca 2ϩ spike, fail to enhance the oscillation frequency. Thus, the data suggest that the frequency regulator is not Ins(1,4,5)P 3 per se but its metabolite(s).
Ins(1,4,5)P 3 is easily metabolized to Ins(1,3,4,5)P 4 and Ins(1,4)P 2 catalyzed by IP3K and 5-phosphatase, respectively (25). Table I shows that Ins(1,4)P 2 failed to induce an intracellular Ca 2ϩ release or an increase in the frequency of Ca 2ϩ oscillations. Thus, Ins(1,3,4,5)P 4 is likely to be involved, directly or indirectly, in regulating the frequency in the cells. To verify this possibility, we investigated the effect of IP3K inhibitors on Ca 2ϩ oscillations. Adriamycin is a potent inhibitor of IP3K (26). Fig. 4 shows that, when HeLa cells were incubated with 10 M adriamycin for 2 h before histamine treatment, the histamine-induced Ca 2ϩ oscillations were abolished. In its place, only a single Ca 2ϩ spike was observed, suggesting that Ins(1,4,5)P 3 was generated to cause the release of Ca 2ϩ from the intracellular stores. However, without active IP3K to catalyze the formation of Ins(1,3,4,5)P 4 , the cells fail to produce Ca 2ϩ oscillations. Other evidence supporting Ins(1,3,4,5)P 4 playing an important role in regulating the frequency was obtained with monoclonal anti-IP3K antibody (Table II). The antibody was introduced into the cells by electroporation in a Ca 2ϩ -free buffer containing 1 mg/ml antibody. Five min after electroporation, the cells were transferred into Buffer A containing 1.8 mM Ca 2ϩ , and the histamine was added to a final concentration of 200 M. As a control, a similar procedure was applied to another population of cells, except that the anti-IP3K antibody was replaced with a nonspecific anti-mouse IgG. The results showed that only 35% of the anti-IP3K antibodytreated cells exhibited Ca 2ϩ oscillations in response to histamine, relative to 73 and 75% observed with anti-IgG-treated and untreated cells, respectively. In addition, statistical analysis revealed that the oscillation frequency exhibited by the anti-IP3K-treated cells, 0.9 Ϯ 0.6 spike/min, was significantly lower than the control, which gave 1.6 Ϯ 0.5 spike/min. The heterogeneity of the anti-IP3K antibody-loaded cells indicates that the quantity of the antibody loaded in each cell was not the same. The population of cells that received a sufficient quantity of antibody to fully inactivate the IP3K exhibited one single Ca 2ϩ spike. Those cells that received a lower quantity of antibody, such that their IP3K was partially inhibited, could undergo Ca 2ϩ oscillations but with reduced frequency.
The data shown in Fig. 4 and Table II clearly demonstrate that the metabolism of Ins(1,4,5)P 3 via phosphorylation at the 3-position is essential for sustaining Ca 2ϩ oscillations and for controlling the oscillation frequency in HeLa cells. Because it  can be metabolized to other inositol phosphates, we cannot directly confirm that Ins(1,3,4,5)P 4 itself is the frequency regulator. The most accepted and well characterized route for metabolizing Ins(1,3,4,5)P 4 is the 5-phosphatase-catalyzed hydrolysis that yields Ins(1,3,4)P 3 , which can then be metabolized further to Ins(1,3)P 2 , Ins(3,4)P 2 , and Ins(1,3,4,6)P 4 , with the latter as a minor product (25). The data in Table I show that Ins(1,3,4)P 3 in the ϳ2 M intracellular concentration range failed to induce both intracellular Ca 2ϩ release and an increase in the frequency of ongoing Ca 2ϩ oscillations. The latter observation also implies that inositol phosphate 6-kinase-catalyzed phosphorylation of Ins(1,3,4)P 3 is not a major pathway in HeLa cells. This is in accordance with the finding that phosphorylation of Ins(1,3,4)P 3 by 6-kinase is a minor route relative to total Ins(1,3,4)P 3 metabolism during receptor activation (27). The concentration of Ins(1,3,4,5)P 4 in the agonist-stimulated cells has been reported to be similar to (25) or higher than (28,29) that of Ins(1,4,5)P 3 . Furthermore, IP3K is a Ca 2ϩ /CaM-activated enzyme, in which activity can be enhanced 4-fold by micromolar concentrations of the Ca 2ϩ /CaM complex (30). This finding suggests that the concentration of Ins(1,3,4,5)P 4 will transiently increase following each Ca 2ϩ spike. Based on our data and the properties of IP3K reported elsewhere (30), it is feasible that the frequency of Ca 2ϩ oscillation is achieved by altering the concentration of Ins(1,3,4,5)P 4 .
Our results allow us to identify Ins(1,3,4,5)P 4 as a second messenger in sustaining, directly or indirectly, Ca 2ϩ oscillation and regulating its frequency in HeLa cells. Hirose et al. (17) reported that in Madin-Darby canine kidney epithelial cells, the concentration of Ins(1,4,5)P 3 oscillates synchronously with Ca 2ϩ oscillation. Strictly speaking, because Ca 2ϩ release follows Ins(1,4,5)P 3 binding to Ins(1,4,5)P 3 receptor and most of the Ins(1,4,5)P 3 is likely to be converted to Ins(1,3,4,5)P 4 by the cooperative interaction of Ca 2ϩ , calmodulin, and IP3K, the decrease in Ins(1,4,5)P 3 concentration should at least lag somewhat behind Ca 2ϩ concentration decreases even if IP3K instantaneously phosphorylates free Ins(1,4,5)P 3 . In fact, such "synchrony," barring the presence of other major Ca 2ϩ -activated Ins(1,4,5)P 3 -binding proteins, can be achieved by a sufficient amount of fully activated IP3K. In this sense, the finding of Hirose et al. (17) supports our observation that IP3K plays a pivotal role in Ca 2ϩ oscillations. It is also likely that the product, Ins(1,3,4,5)P 4 , in turn regulates the frequency of oscillation. It should be pointed out that Ins(1,3,4,5)P 4 has been implicated in modulating Ca 2ϩ entry across the plasma membrane from the extracellular Ca 2ϩ pool (31), which is known to alter the frequency and amplitude of intracellular Ca 2ϩ oscillations in HeLa cells (13,32) by an unknown mechanism. However, because our experiments were conducted in a Ca 2ϩfree medium, it indicates that this mechanism is probably not responsible for the observed effects. Two Ins(1,3,4,5)P 4 -specificbinding proteins, GAP1 IP4BP and GAP1 m , have been purified and found ubiquitously expressed in human tissues (33). Both are members of the GAP1 family of Ras GTPase-activating proteins, and they exhibit relatively low affinity for Ins(1,3,4,6)P 4 . Whether and how the frequency-regulating properties of Ins(1,3,4,5)P 4 may be mediated through one of these proteins is unclear. On the other hand, we have shown that a phosphorylation/dephosphorylation cycle, which in-volves the Ca 2ϩ /CaM-dependent protein kinase II and a calyculin A-inhibitable protein phosphatase, plays a pivotal role in sustaining Ca 2ϩ oscillations in HeLa cells. The addition of calyculin A to an ongoing Ca 2ϩ oscillation causes a decrease in its frequency and, at high concentrations, it reduces the oscillation to a single Ca 2ϩ spike (13). Several inositol trisphosphates and tetrakisphosphates have been shown to activate a phosphoprotein phosphatase (34). Thus, Ins(1,3,4,5)P 4 may exert its function via direct or indirect activation of a specific phosphoprotein phosphatase in which the activity determines the duration between two adjacent Ca 2ϩ spikes, hence controlling the frequency of Ca 2ϩ oscillations.