Requirement of Inositol Pyrophosphates for Full Exocytotic Capacity in Pancreatic β Cells

See allHide authors and affiliations

Science  23 Nov 2007:
Vol. 318, Issue 5854, pp. 1299-1302
DOI: 10.1126/science.1146824


Inositol pyrophosphates are recognized components of cellular processes that regulate vesicle trafficking, telomere length, and apoptosis. We observed that pancreatic β cells maintain high basal concentrations of the pyrophosphate diphosphoinositol pentakisphosphate (InsP7 or IP7). Inositol hexakisphosphate kinases (IP6Ks) that can generate IP7 were overexpressed. This overexpression stimulated exocytosis of insulin-containing granules from the readily releasable pool. Exogenously applied IP7 dose-dependently enhanced exocytosis at physiological concentrations. We determined that IP6K1 and IP6K2 were present in β cells. RNA silencing of IP6K1, but not IP6K2, inhibited exocytosis, which suggests that IP6K1 is the critical endogenous kinase. Maintenance of high concentrations of IP7 in the pancreatic β cell may enhance the immediate exocytotic capacity and consequently allow rapid adjustment of insulin secretion in response to increased demand.

Phosphoinositides have a prominent role in cellular signal-transduction events (13). Highly phosphorylated inositol polyphosphates, distant derivatives of the inositol 1,4,5-trisphosphate (IP3) second messenger, function in signal-transduction and cellular regulation (46). The pyrophosphate derivatives of IP6 diphosphoinositol pentakisphosphate, and bis-(diphospho)inositol tetrakisphosphate are commonly referred to as IP7 and IP8 (also, InsP7 and InsP8, respectively). These inositol pyrophosphate derivatives rapidly turnover and are estimated to have similar free energy of hydrolysis to that of adenosine 5′-triphosphate (ATP) (4). A striking consequence of this high-energy phosphate group is the ability of IP7 to phosphorylate a subset of proteins directly in an ATP- and enzyme-independent manner (7). The variety of cellular responses that are apparently controlled by inositol pyrophosphates (4, 8) may be facilitated by the differential intracellular distribution of the kinases that make them (9). The concentrations of inositol pyrophosphates can be dynamically regulated during key cellular events. For example, IP7 concentrations change during cell cycle progression (10), and IP7 regulates cyclin–cyclin-dependent kinase complexes (11), whereas IP8 increases acutely in response to cellular stress (8). IP6 also functions as an enzymatic cofactor (4), and so, by analogy, it is possible that even at concentrations found in unstimulated cells, IP7 could be an important regulatory molecule.

Phosphoinositides are key regulators of the insulin-secreting pancreatic β cell (12). These cells influence blood glucose homeostasis by coupling increases in the concentration of glucose and other circulatory or neuronal-derived regulators to the exocytosis of insulin. IP6 activates voltage-dependent L-type Ca2+ channels (13), exocytosis (14, 15), and dynamin-mediated endocytosis (16), all key processes in insulin secretion. A role for IP7 in the β cell has not yet been determined. However, given the suggested involvement of inositol pyrophosphates in vesicle trafficking (4), the critical nature of such trafficking events for the process of insulin exocytosis and the high β cell concentration of IP6 (13), the immediate precursor of IP7, we explored a possible role for IP7 in the regulation of insulin exocytosis.

We used [3H]myo-inositol labeling to examine insulin-secreting cells and pancreatic islets for the presence of various inositol pyrophosphate species. IP7 was identified by its coelution with a bona fide IP7 standard (fig. S1) (17). Very little IP8 was detectable. In normal mouse pancreatic islets (60% β cells), the relative concentration of IP7 was 5.21 ± 0.12% (± SEM, n = 3) of IP6. In contrast, the percentage of IP7 in islets from ob/ob mice, which have more than 90% β cells, was significantly greater, i.e., 7.71 ± 0.93% of IP6 (± SEM, n = 4; P< 0.05; unpaired, two-tailed t test). This suggests that the increased IP7 concentrations might be restricted primarily to the β cells. When we normalized the primary mouse data, on the assumption that only β cells make IP7 (Fig. 1B), they suggested that the β cells maintain IP7 levels at about 9% of those of IP6. Of the insulin-secreting cell lines, only hamster HIT-T15 cells have a considerable amount of IP7 (10% of that of IP6) (Fig. 1A). We used equilibrium labeling techniques (13) to estimate that the basal concentration of IP7 in HIT-T15 cells was 5.8 ± 0.14 μM (± SEM, n = 3), which reflects a concentration at the top end of the range that has been estimated in other mammalian cells or yeast (1 to 5μM) (4). Because IP7 is in a state of rapid exchange with the cellular IP6 pool in mammalian cells (4) and because the cellular concentration of IP6 in β cells is also high (13), it is perhaps not surprising that IP7 in these cells is at high concentrations.

Fig. 1.

Concentrations of IP7 in pancreatic β cells. (A) Comparison of 3H-labeled IP7 as a percentage of 3H-labeled IP6 in pancreatic islets or insulin-secreting HIT-T15 cells. Data are from three or four separate experiments. *P < 0.05, two-tailed t test. (B) The islet data from (A) were transformed to take into account the different β cell composition of lean mouse (60%) versus ob/ob mouse (90%) islets.

To investigate whether high IP7 concentrations keep β cells in a responsive state, we overexpressed all three reported mammalian IP6Ks in primary β cells and examined whether stimulated exocytosis was enhanced. We used cell capacitance as a measure of exocytosis, detecting the increase in β cell surface area that occurs when the insulin-containing granules fuse with the plasma membrane (18). The perforated patch whole-cell technique was used to allow measurements in metabolically intact cells, and exocytosis was elicited by groups of four 500-ms depolarizing pulses from –70 to 0 mV. Inmock-transfected cells, capacitance increased by 79 ± 11 f F (n = 8) (Fig. 2, A and B). In cells overexpressing IP6K1, the amplitude of the capacitance increase averaged 198 ± 12 f F (n = 10; P < 0.05), whereas no effect on exocytosis was observed in cells overexpressing a catalytically inactive version of IP6K1 (Fig. 2, A and B). The capacitance increase evoked by the first depolarization was augmented by 293% in cells overexpressing wild-type IP6K1. Exocytosis during the first depolarization is thought to largely represent the content of the readily releasable pool (RRP) of exocytotic granules (19). The size of the RRP (in femtofarads) can be estimated by using the equation: RRP = S/(1 – R2), where S is the sum of the responses to the first (ΔC1) pulse and the second (ΔC2) pulse, and R is the ratio ΔC2C1 (19). We estimate that the RRP averaged 96 ± 9 fF (n = 8) and 225 ± 21 fF (n = 10) in mock-transfected and wild-type IP6K1 transfected cells, respectively. Thus, IP6K1 increased the size of the RRP by 134%. Using a conversion factor of 3 fF per granule (20), we estimated that the RRP contained 30 granules in mock-transfected and 75 granules in wild-type IP6K1 transfected cells. The stimulatory action of IP6K1 was restricted to the first depolarization, and little enhancement was seen over that observed in mock-transfected cells during the final three pulses (Fig. 2A). This became obvious when the net increase per pulse was plotted (Fig. 2B). The exhaustion of the exocytotic response during the train appeared not to reflect inactivation of the depolarization-induced Ca2+ current, because the change in the integrated Ca2+ current (QCa) was much smaller (20%) than the decrease in exocytosis (fig. S2), but also see (21).

Fig. 2.

Increased exocytosis in pancreatic β cells after expression of IP6Ks. (A) Individual mouse β cells were transfected with enhanced green fluorescent protein (EGFP) (mock) or a combination of EGFP and either wild-type (IP6K1) or a catalytically inactive (IP6K1-K/A) variant of IP6K. Increases in cell capacitance (ΔCm) were elicited by a train of four 500-ms depolarizations (1 Hz) by using the perforated patch configuration. The extracellular medium contained 3 mM glucose. Recordings are representative of 8 to 12 different experiments. (B) Histogram summarizing the average increment in cell capacitance per pulse (ΔCm,n – ΔCm,n – 1) during the train in cells mock-transfected or overexpressing either IP6K1 or IP6K1-K/A. Values ± SEM are from 8 to 12 experiments. *P < 0.05 from Dunnett's test for multiple comparisons. (C) Average total increase in cell capacitance at the end of the train in mock-transfected cells or cells overexpressing either IP6Kn or IP6Kn-K/A type 1, 2, and 3 kinase, respectively. Values ± SEM are from 7 to 12 experiments. *P < 0.05 from Dunnett's test for multiple comparisons. (D) INS-1E cells cotransfected with pCMV5-hGH and empty vector (pcDNA3) (mock-transfected) or with pCMV5-hGH and IP6Kn or IP6Kn-K/A types 1, 2, and 3 kinase, respectively. hGH secretion was measured in Krebs-Ringer bicarbonate Hepes buffer with 3 mM glucose. The amount of secreted hGH is expressed as a percentage of total hGH in the cells. Values ± SEM are from three experiments (each in triplicate). *P < 0.05 from Dunnett's test for multiple comparisons.

As did IP6K1, IP6K2 and IP6K3 also stimulated exocytosis (Fig. 2C). Overexpression of a catalytically inactive version of IP6K2 and IP6K3 did not affect the exocytotic capacity (Fig. 2C). We also overexpressed IP6K1-3 in rat insulinoma INS-1E cells in a human growth hormone (hGH) transient cotransfection assay, in which hGH acted as a reporter of exocytosis from transfected cells. Total increases in cell capacitance in INS-1E cells overexpressing IP6K1 (212 ± 28 fF; n = 5) were comparable to those observed in primary mouse β cells (201 ± 16 fF; n = 12). Overexpression of each IP6K-stimulated hGH secretion was 150% above basal (n = 9 to 12; P < 0.05), an effect that was not shared by their catalytically inactive mutants (Fig. 2D).

IP6Ks can also use IP5 as a substrate, generating a different subset of inositol pyrophosphates (4). Therefore, we verified that 5-IP7 (mammalian IP7 with pyrophosphate in the 5 position) directly promoted exocytosis (Fig. 3, A to C). We also assessed other isomers of IP7 (Fig. 3D). To measure the effects of 5-IP7 on exocytosis, we applied trains of depolarizations in standard whole-cell patch-clamp experiments in which the β cell was dialyzed for 2 min with a solution containing 3 μM 5-IP7. A group consisting of four 500-ms depolarizations was then applied to evoke exocytosis. In a series of six experiments, the total increase in cell capacitance amounted to 231 ± 12 fF (P < 0.01) in the presence of 3 μM 5-IP7 in the pipette-filling solution and 77 ± 11 fF under control conditions, respectively (Fig. 3A). As was the case for cells overexpressing the IP6Ks, the capacitance was primarily increased by the first depolarization in the presence of 5-IP7, with little effect of the subsequent three depolarizations (Fig. 3B). The ability of 5-IP7 to stimulate exocytosis was not associated with a change in the whole-cell Ca2+ current (fig. S3). The stimulatory action of 5-IP7 on exocytosis was concentration-dependent (Fig. 3C), with a half-maximal stimulatory effect at 1.02 μM, and a cooperativity factor of 1.5. Thus, 5-IP7 enhanced exocytosis within the physiological range of IP7 concentrations (1 to 10 μM). Other isomers of IP7 also stimulated exocytosis at 10 μM. CH-PP (monocyclohexyl trisodium diphosphate), a simple pyrophosphate based on cyclohexane, was ineffective (Fig. 4E). Under the conditions used to examine 5-IP7's effect on exocytosis, the net effect of IP6 appeared to promote endocytosis not exocytosis (fig. S4), because the effect of IP6 on exocytosis could only be discerned under conditions in which endocytosis was inhibited (15). This was not the case for 5-IP7. Furthermore, the effect of IP6 on exocytosis, when endocytosis was inhibited, did not selectively promote secretion from the RRP (fig. S4). Our experiments do notprecludearolefor amorephosphorylated pyrophosphate.

Fig. 3.

Promotion of Ca2+-dependent exocytosis by 5-IP7. Individual mouse β cells were subjected to a train of four 500-ms depolarizations (1 Hz) by using the standard whole-cell patch-clamp configuration. (A) Exocytosis was observed under control conditions and in the presence of 3 μM5-IP7 in the pipette-filling solution. 5-IP7 was allowed to diffuse into the cell for 2 min before initiation of the experiment. Recordings are representative of six experiments for both conditions. (B) Histogram summarizing the average increment in cell capacitance per pulse (ΔCm,n – ΔCm,n – 1) during the train in the absence or presence of 3 μM5-IP7 in the pipette-filling solution. Values ± SEM are from six experiments. *P < 0.05 from Dunnett's test for multiple comparisons. (C) Concentration dependence of stimulatory action of 5-IP7 on exocytosis. The curve represents aleast-squares fit of the mean data points to the Hill equation. Values ± SEM are from five to seven experiments. *P <0.05 from Dunnett's test for multiple comparisons. (D) A comparison of several isomers of IP7 at a 10 μM concentration on total increase in exocytosis with the same protocols as in (A). Values ± SEM are from five to eight experiments. *P < 0.05 from Dunnett's test for multiple comparisons.

Fig. 4.

(A) RNA silencing of IP6K1, but not IP6K2, inhibits release of granules from the RRP. Total RNA was extracted from islets and MIN6m9 cells and reverse-transcribed. Relative expression of mRNA was measured by quantitative real-time PCR with the use of appropriate primers and probes. Primers and probe for 18S ribosomal RNA were used as endogenous control. (B) Mouse β cells were transfected with fluorescently tagged siRNA to IP6K1 or a negative control at 25 nM and were subjected to a train of four 500-ms depolarizations (1 Hz) with the perforated patch configuration. Increases in cell capacitance (ΔCm) were measured in fluorescent cells at 3 mM glucose in the extracellular medium. Recordings are representative of six (control) and eight (siRNA-treated cells) experiments. (C) Histogram summarizing the average increment in cell capacitance per pulse (ΔCm,n – ΔCm,n – 1) during the trainincells mock-transfected or overexpressing either siRNA to IP6K1 or negative control. Values ± SEM are from six and eight experiments. *P < 0.05 from Dunnett's test for multiple comparisons. (D) Effect of siRNA to IP6K1 and IP6K2 on total capacitance increase. Values ± SEM are from five to eight experiments. *P < 0.05 from Dunnett's test for multiple comparisons. (E) Effect of 5-IP7 on exocytosis under control conditions and in cells with reduced expression levels of IP6K1 (whole-cell patch, see Fig. 3). Values ± SEM are from five or six experiments. *P < 0.05 from Dunnett's test for multiple comparisons.

To test whether endogenous IP7 contributed to the exocytotic capacity in a physiologically relevant manner, we first examined the mRNA encoding IP6Ks from β cells or islet lysates by quantitative polymerase chain reaction (PCR). Message encoding IP6K1 and IP6K2 but not IP6K3 was detected (Fig. 4A). We then decreased expression of IP6K1 and IP6K2 in β cells with small interfering RNA (siRNA). Mouse-specific siRNAs were screened by using the mouse β cell line MIN6m9 and TaqMan realtime PCR gene expression assays (fig. S5). Elimination of either IP6K1 or IP6K2 reduced cellular IP7 levels (fig. S6). In order to be able to select primary β cells transfected with siRNA for further electrophysiological studies, suitable siRNA candidates were fluorescently tagged at the 5′ end of sense and antisense strands and transfected. Cell-capacitance measurements on mouse β cells transfected with fluorescently tagged siRNA were recorded by using the perforated patch-clamp technique. Inhibition of IP6K1, but not of IP6K2, reduced the exocytotic capacity (Fig. 4D), and the effect of silencing was again most pronounced on the first pulse, reflecting depletion of the RRP of granules (Fig. 4, B and C). Furthermore, addition of 5-IP7 in the presence of siRNA to IP6K1 restored a normal exocytotic response (Fig. 4E). Thus, endogenous IP7 generated by IP6K1, but not IP6K2, appears to account for the enhanced exocytotic capacity in β cells. The discrepancy between our exogenous versus endogenous systems may reflect a differential distribution or cellular associations of the two kinases in vivo. IP6K1 associates with proteins involved in exocytosis but IP6K2 does not (22). Studies on IP6K2 have also revealed a discrepancy between overexpression studies and gene silencing (23).

IP6K1 siRNA did not alter number of L-type Ca2+ channels per patch or channel open probability, mean closed time, or mean open time (fig. S7). Hence, in contrast to IP6, IP7 appears not to affect L-type Ca2+ channel activity (13).

We find that the pancreatic β cell maintains high concentrations of IP7. This apparently functions in the insulin secretory process by regulating the RRP of insulin-containing granules, thereby maintaining the immediate exocytotic capacity of the β cell. It is noteworthy that a putative disruption of the IP6K1 gene in a family with type 2 diabetes (24) and reduced plasma insulin levels in mice in which the IP6K1 gene has been deleted (25). This may be of interest in the context of understanding the molecular mechanisms underlying the development of diabetes.

Supporting Online Material

Materials and Methods

Figs. S1 to S7


References and Notes

Stay Connected to Science

Navigate This Article