Report

Spatiotemporal Dynamics of Inositol 1,4,5-Trisphosphate That Underlies Complex Ca2+ Mobilization Patterns

See allHide authors and affiliations

Science  28 May 1999:
Vol. 284, Issue 5419, pp. 1527-1530
DOI: 10.1126/science.284.5419.1527

Abstract

Inositol 1,4,5-trisphosphate (IP3) is a second messenger that elicits complex spatiotemporal patterns of calcium ion (Ca2+) mobilization and has essential roles in the regulation of many cellular functions. In Madin-Darby canine kidney epithelial cells, green fluorescent protein–tagged pleckstrin homology domain translocated from the plasma membrane to the cytoplasm in response to increased concentration of IP3. The detection of translocation enabled monitoring of IP3 concentration changes within single cells and revealed spatiotemporal dynamics in the concentration of IP3 synchronous with Ca2+ oscillations and intracellular and intercellular IP3 waves that accompanied Ca2+ waves. Such changes in IP3 concentration may be fundamental to Ca2+ signaling.

IP3production by phospholipase C (PLC)–mediated hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP2) is an early intracellular event after stimulation by hormones, autacoids, and neurotransmitters. IP3 mobilizes Ca2+from intracellular stores through the IP3 receptor, resulting in activation of Ca2+-dependent cellular events such as contraction, secretion, gene expression, and synaptic plasticity (1, 2). Ca2+ mobilization occurs in complex temporal and spatial patterns, including Ca2+ oscillations (3) and Ca2+ waves (4). However, the mechanism underlying the generation of the complex patterns has not been fully elucidated, partly because of lack of knowledge regarding IP3 dynamics in single cells.

Green fluorescent protein (GFP)–based probes have been used to analyze cellular signaling because they have the advantage that they can be DNA encoded (5). Fusion proteins consisting of GFP and a functional protein domain can function as molecular probes when their intracellular translocation pattern can be visualized (6). The GFP-tagged pleckstrin homology (PH) domain of PLC-δ1 (GFP-PHD) is one such probe because it binds to PIP2 within the plasma membrane and translocates to the cytoplasm after receptor stimulation (7). Although the translocation was thought to reflect a decrease in the PIP2concentration (7), we obtained evidence that an increase in the cytoplasmic IP3 concentration ([IP3]i) causes the translocation of GFP-PHD, and therefore we used GFP-PHD to monitor spatiotemporal changes in [IP3]i that underlie the complex Ca2+ mobilization patterns within single living cells.

We analyzed PIP2 binding of the PH domain of PLC-δ1 (8) by a surface plasmon assay (9) and obtained dissociation constants (K d) of 2.8 and 2.1 μM for PH domains with and without GFP tagging, respectively (Fig. 1A). IP3 inhibited this binding in a dose-dependent manner, and the K d for IP3 was 93 nM (Fig. 1B), indicating that IP3 binds to this PH domain with a ∼20-fold higher affinity than PIP2, consistent with previous reports (10). IP3inhibited the binding of GFP-PHD to PIP2 (Fig. 1C) with similar efficiency, indicating no impairment of IP3 binding by GFP tagging. Inositol 1,3,4-trisphosphate and inositol 1,3,4,5-tetrakisphosphate displaced little GFP-PHD from PIP2 (Fig. 1C). This result confirms the ligand recognition specificity of the PH domain (10). The off rate (k off) deduced from the dissociation curve was 5.25 s−1 (Fig. 1D), which indicates the lower limit of the actual k offbecause of the limited time resolution. Thus, IP3rapidly displaces GFP-PHD from PIP2 with high affinity and specificity. We, therefore, studied the movement of GFP-PHD within living cells in conjunction with the receptor-mediated phosphatidylinositol turnover.

Figure 1

In vitro characterization of GFP-PHD. (A) Dose dependence of the PIP2 binding of the PH domain with (•, dashed curve) and without (○, solid curve) GFP. Average ± SEM (n = 3). (B) Dose-dependent inhibition (○) of PIP2 binding of the PH domain (2 μM) and estimated IP3 binding to the PH domain (•) (n = 3). (C) Inhibition of PIP2 binding of 2 μM GFP-PHD by 2 μM IP3 (InsP3), inositol 1,3,4-trisphosphate [Ins(1,3,4)P3], and inositol 1,3,4,5-tetrakisphosphate (InsP4). (D) Dissociation of GFP-PHD from PIP2. At time zero, GFP-PHD was removed and IP3 (400 μM) was introduced. A.U., arbitrary units.

GFP-PHD was expressed in Madin-Darby canine kidney (MDCK) epithelial cells (11), and localization of its fluorescence was examined under a confocal microscope (12). GFP-PHD was concentrated at the plasma membrane (Fig. 2A). Cells expressing GFP alone showed homogeneous cytoplasmic and nuclear staining (13). Thus, GFP-PHD preferentially interacts with a plasma membrane component, presumably PIP2, as does full-length PLC-δ1 (14). In cells treated with adenosine triphosphate (ATP), fluorescence intensity of GFP-PHD in the cytoplasmic region increased, whereas that at the plasma membrane decreased (Fig. 2A). Removal of ATP restored the original fluorescence distribution (13). Neither binding of IP3nor binding of PIP2 in vitro changed the fluorescence intensity of GFP-PHD (13). Thus, the observed changes in the fluorescence intensity apparently reflect the translocation of GFP-PHD and were ATP concentration dependent (Fig. 2B). Similar results were observed in cells stimulated with bradykinin (13).

Figure 2

Translocation of GFP-PHD induced by IP3. (A) Cells challenged with 50 μM ATP were examined by confocal microscopy. Time courses of the fluorescence intensities in the membrane [defined by the bright peripheral region (widths, 1 μm) of the cells before stimulation] and cytoplasmic regions are shown together with the images at the time points indicated (arrows). Data shown are representative of four determinations. The membrane region was separately verified with a membrane probe, FM4-64. (B) Dependence of the extent of cytoplasmic translocation on ATP concentration. ΔF/F0, fractional changes in fluorescence intensity. (C) Effect of microinjection of IP3 (final concentration, ∼80 μM) or vehicle on cytoplasmic translocation of GFP-PHD in the absence of extracellular Ca2+. Experiments were also performed after treatment with U73122 (5 μM) and ionomycin (10 μM). Data shown are representative of three experiments. (D) IP3 dependence of GFP-PHD translocation. [IP3]i was estimated by fluorescence of rhodamine-dextran with which IP3 (100 to 500 μM) was coinjected. (E) Effect of IP3 5-phosphatase on the translocation of GFP-PHD. GFP-PHD–expressing cells transfected with pcDNA3.1-IP3 5-phosphatase (a tod) and pcDNA3.1-EBFP or pcDNA3.1 and pcDNA3.1-EBFP (e to h) were imaged with a CCD camera. The fluorescence images of GFP-PHD (a and e) and BFP (b and f) are shown. (c) and (g) are GFP-PHD images divided by the average of 10 consecutive images before stimulation. The time course of cytoplasmic translocation is shown (d and h). ATP (50 μM, open bar) was applied after ionomycin treatment and in the presence of a low extracellular Ca2+concentration (1 μM) to avoid intracellular Ca2+elevation that might secondarily augment PIP2hydrolysis. Similar results were obtained in cells without such treatment. Data shown are representative of four experiments.

Microinjection of IP3 induced translocation in a dose-dependent manner resembling that by agonist stimulation (Fig. 2, C and D). U-73122, a PLC inhibitor (15), did not block this translocation (Fig. 2C), indicating that PIP2 hydrolysis is not essential for this translocation. Nor was Ca2+ mobilization required, because depletion of the stores by ionomycin did not block the translocation (Fig. 2C). Because IP3 5-phosphatase participates in the degradation of IP3 (16), we examined the effect of its overexpression on GFP-PHD translocation. Translocation of GFP-PHD elicited by purinergic stimulation was abolished in the IP3 5-phosphatase–expressing cells (Fig. 2E), indicating that an increase in [IP3]i is necessary for the agonist-elicited translocation. Thus, an increase in [IP3]i is both necessary and sufficient for the translocation of GFP-PHD. Moreover, complete abolition of the translocation by overexpression of IP3 5-phosphatase indicates that during agonist stimulation, the concentration of free PIP2 available to GFP-PHD remains either constant or greatly in excess of the K d of PIP2binding (17).

We monitored the translocation of GFP-PHD to analyze changes in [IP3]i associated with complex Ca2+ mobilization patterns. GFP-PHD–expressing cells were incubated with the Ca2+ indicator, fura-2. The negligible overlap in the excitation spectra of GFP and fura-2 enabled us to detect both Ca2+ and IP3 signals simultaneously. ATP (1 to 3 μM) often elicited Ca2+oscillations in MDCK cells (18), and oscillatory translocation of GFP-PHD synchronous with Ca2+ oscillations was observed (Fig. 3A). IP3 oscillations have been suggested by measurement of [IP3]i in a large population of cells in which Ca2+ oscillations were synchronized by removal and restoration of extracellular Ca2+(19), although the validity of this technique has been challenged (20). Our results provide evidence for IP3 oscillations accompanying Ca2+oscillations at the single-cell level.

Figure 3

Temporal dynamics of IP3. (A) IP3oscillations (bottom traces) monitored by GFP-PHD translocation during Ca2+ oscillations (top traces) elicited by the application of 3 μM ATP (open bar). Representative data from at least 11 oscillating cells are shown. (B) Time courses of GFP-PHD translocation (solid traces) and [Ca2+]i(dashed traces) during ATP (3 μM) application (open bar) and subsequent introduction of extracellular Ca2+ (gray bar) in the cells pretreated with ionomycin (10 μM) in the absence of Ca2+. Black, blue, and red traces denote those time courses with the introduction of 0, 100, and 300 μM Ca2+, respectively. (C) Additional increase in GFP-PHD translocation 60 s (red) or 120 s (blue) after the introduction of Ca2+ (0, 10, 100, and 300 μM), plotted against [Ca2+]i. Mean ± SEM (n = 7 to 14).

Generation of oscillations in [IP3]i is thought to require Ca2+-dependent activation of PLC (19,21). We therefore examined the effect of intracellular Ca2+ concentration ([Ca2+]i) on ATP-induced increase in [IP3]i. When cells were incubated with ionomycin or thapsigargin to deplete the Ca2+ stores and then stimulated with ATP, translocation of GFP-PHD was observed without any change in [Ca2+]i. A subsequent increase in the extracellular Ca2+ concentration induced Ca2+influx, which then elicited further translocation (Fig. 3, B and C). However, the increase in [IP3]i was transient and began to decrease, whereas the [Ca2+]i continued to increase (Fig. 3B). The relation between [Ca2+]i and translocation of GFP-PHD changed with time, and at higher [Ca2+]i, a pronounced time-dependent inhibition was observed (Fig. 3C). These results indicate that Ca2+ has both enhancing and inhibitory effects on [IP3]i increase.

We analyzed spatial changes in [IP3]iin detail during the early phase of the increase in [Ca2+]i accompanying intracellular Ca2+ waves in MDCK cells after purinergic stimulation (Fig. 4). Translocation of GFP-PHD occurred concomitantly with Ca2+ wave propagation, indicating the presence of IP3 waves. Taking into consideration the inherent kinetic and diffusional delay in the GFP-PHD signal, [Ca2+]i and [IP3]i waves appeared almost simultaneously, supporting the models in which regenerative Ca2+-mediated IP3 production accompanies Ca2+ waves or oscillations (19, 21).

Figure 4

(left).Intracellular dynamics of IP3. Representative data from 16 cells that showed intracellular Ca2+ waves upon application of ATP (3 μM). The images normalized by the average of 10 images before stimulation are shown. The time courses of the signals from the regions, indicated by the numbered boxes, are plotted. Both Ca2+ and IP3 waves start from the lower region of the cells. The upper region of the cell was another focus of the Ca2+ wave, which propagated downward until it reached the perinuclear region. Although the very early elevation of the IP3 signal is difficult to see, the IP3 wave is suggested by the delayed elevation in the perinuclear region.

Mechanical stimulation of MDCK cells initiated intercellular Ca2+ waves that spread toward peripheral cells (Fig. 5) (22). Simultaneous imaging of GFP-PHD showed that the increase in [IP3]i also spread in a wave pattern similar to that of the Ca2+ wave.

Figure 5

(right).Mechanical stimulation of the cell numbered “1” with a fine glass capillary elicited an intercellular Ca2+ wave propagating to the neighboring cells. Ca2+ and IP3waves are monitored with fura-2 and GFP-PHD, respectively. The time courses of the signals are plotted for the cells numbered as indicated. Data shown are representative of four experiments.

Our results provide insight into the mechanism of generation of complex Ca2+ signals. Two alternative mechanisms underlying the complex Ca2+ mobilization patterns have been proposed: Ca2+-mediated positive and negative feedback mechanisms may control either the Ca2+ release process itself (mechanism 1) or IP3 production (mechanism 2) (1,3, 4). Our results are consistent with mechanism 2, which proposes the occurrence of oscillations in the [IP3]i and IP3 waves. We also observed Ca2+-mediated enhancement and suppression of [IP3]i increase, both of which are postulated in mechanism 2. However, our results do not exclude mechanism 1, and Ca2+-mediated regenerative mechanisms of both Ca2+ release and IP3 production may participate cooperatively in the generation of complex Ca2+signaling patterns. Regarding the intercellular Ca2+ wave, intercellular diffusion of IP3 may be also involved (23). The relative contribution of these mechanisms remains to be clarified.

  • * To whom correspondence to be addressed. E-mail: hirose{at}calcium.cmp.m.u-tokyo.ac.jp

REFERENCES AND NOTES

View Abstract

Navigate This Article