Nuclear K_ATP channels trigger nuclear Ca²⁺ transients that modulate nuclear function

Abstract

Glucose, the principal regulator of endocrine pancreas, has several effects on pancreatic beta cells, including the regulation of insulin release, cell proliferation, apoptosis, differentiation, and gene expression. Although the sequence of events linking glycemia with insulin release is well described, the mechanism whereby glucose regulates nuclear function is still largely unknown. Here, we have shown that an ATP-sensitive K⁺ channel (K_ATP) with similar properties to that found on the plasma membrane is also present on the nuclear envelope of pancreatic beta cells. In isolated nuclei, blockade of the K_ATP channel with tolbutamide or diadenosine polyphosphates triggers nuclear Ca²⁺ transients and induces phosphorylation of the transcription factor cAMP response element binding protein. In whole cells, fluorescence in situ hybridization revealed that these Ca²⁺ signals may trigger c-myc expression. These results demonstrate a functional K_ATP channel in nuclei linking glucose metabolism, nuclear Ca²⁺ signals, and nuclear function.

Pancreatic beta cells play a critical role in maintaining a steady-state level of glucose in the blood and tissues. Increased levels of glucose stimulate beta cells to secrete insulin closing a well established feedback loop. Malfunction of beta cells causes the widespread pathology, diabetes mellitus. The signal transduction mechanism leading to insulin release involves the closure of plasma membrane K_ATP channels as a result of glucose metabolism by increasing both the intracellular ATP/ADP ratio and diadenosine polyphosphates (DPs; refs. 1 and 2). Channel closure leads to membrane depolarization and the opening of voltage-activated Ca²⁺ channels (3). The subsequent cytosolic Ca²⁺ signal, which is oscillatory (4), triggers a pulsatile insulin secretion. In most cells, a single second messenger as Ca²⁺ is able to provoke different responses depending on its route of entry, its localization, and a code of amplitude or frequency of Ca²⁺ oscillations (4, 5). In pancreatic beta cells, Ca²⁺ mediates not only insulin secretion but also a broad range of other processes such as gene expression (6, 7). Although it is well established that the nucleoplasmic concentration of free Ca²⁺ regulates nuclear function (5), the mechanism whereby nuclear Ca²⁺ signals are generated is still unclear. Here, we report confocal measurements of nuclear Ca²⁺ concentration ([Ca²⁺]_n) in intact beta cells exposed to glucose. Experiments in isolated nuclei revealed a K_ATP channel present on the nuclear envelope whose blockade results in a [Ca²⁺]_n rise. This increased [Ca²⁺]_n induces phosphorylation of the transcription factor cAMP response element binding protein (CREB). We further demonstrate that [Ca²⁺]_n elevation may result in c-myc expression in whole cells.

Materials and Methods

Cell Isolation, Culture, and Permeabilization.

Islets from adult (8–10 weeks old) Swiss albino male mice (OF1) killed by cervical dislocation were isolated and then dispersed into single cells after a published procedure (3). Isolated cells were cultured in RPMI medium 1640 for 24 h. Cell permeabilization was performed as described (8).

Spot Confocal Microscopy.

Cells were loaded with 2 μM Calcium Green-1/AM (Molecular Probes) for 60 min at room temperature (RT) and then were perfused in a modified Krebs–Ringer buffer [119 mM NaCl/4.7 mM KCl/1.2 mM MgSO₄/1.2 mM KH₂PO₄/25 mM NaHCO₃/2.5 mM CaCl₂/3 mM glucose bubbled constantly with a mixture of 95% O₂/5% CO₂ (pH = 7.4)]. Ca²⁺ was measured by using spot confocal microscopy, which excels in measuring minute Ca²⁺ transients because of its high signal-to-noise ratio. The spot illumination-detection configuration has been described (9, 10). Briefly, a laser-illuminated pinhole (10 μm) was focused onto a spot through the objective on a nucleus equatorial plane. Thus, the predicted detection volume is about 0.6 × 0.6 × 1.1 μm³. Changes in fluorescence in this nuclear volume were precisely detected with a photodiode (HR008; United Detector Technology) which was connected to an Axopatch-200A amplifier (50 GΩ feedback; Axon Instruments, Foster City, CA). The laser illumination time was 30 ms. Ca²⁺-induced fluorescence intensity ratio was plotted as a function of time (F_t/F₀).

Nuclei Isolation.

While dispersed, islet cells were suspended in a buffer that mimicked the intracellular medium (125 mM KCl/2 mM K₂P0₄/40 mM Hepes/0.1 mM MgCl₂, pH 7.2/100 nM Ca²⁺, with 10.2 mM EGTA and 1.65 mM CaCl₂). The cell membrane and cytoskeleton were disrupted by brief sonication. Single isolated nuclei were separated by centrifugation, resuspended in the intracellular medium, and then allowed to attach onto glass chambers (11). Ethidium homodimer-1 labeling (Molecular Probes) indicated that the suspension was 90% enriched in nuclei. This probe also was applied to identify nuclei after each electrophysiological and imaging experiment. An antibody to calnexin allowed an assessment of contamination of the nuclei preparation with endoplasmic reticulum (ER) fragments (12). We further performed confocal images and three-dimensional (3D) reconstruction of nuclei labeled with rhodamine B hexyl ester to detect the existence of some debris attached to the surface of some nuclei (10%), which was associated with ER contamination (13, 14). As this ER debris was completely detectable by transmitted light or fluorescence microscopy, we only performed experiments in intact and clean nuclei (free of debris).

Glibenclamide-Dipyrrometheneboron Difluoride (BODIPY) Staining.

Cultured isolated islet cells were labeled with 40 nM of green-fluorescent glibenclamide-BODIPY-FL (30 min at 4°C; Molecular Probes) and stained with anti-insulin antibodies (15). Immunofluorescence for insulin revealed more than 80% of pancreatic beta cells in the cultures. Isolated nuclei also were stained with glibenclamide-BODIPY-FL following the same protocol. Nuclear envelope was identified with 1 μM of rhodamine B hexyl ester (5 min at RT; Molecular Probes; ref. 14); meanwhile, 1 μM of ethidium homodimer-1 (5 min at RT; Molecular Probes) stained the DNA containing nucleoplasm. Fluorescence was visualized by using a Zeiss LSM 510 confocal microscope (63× objective, 1.25 N.A.) and 1–2 μm optical slices. Intensity values were obtained by calculating the average brightness value on the corresponding ring-like staining of the nuclei and measured on an arbitrary gray scale from 0 (blackest) to 255 (whitest).

Patch-Clamp Experiments.

Single-channel currents were recorded from nuclei positively labeled with ethidium homodimer-1 by using standard patch-clamp recording procedures (16). The bath solution contained 140 mM KCl, 1 mM MgCl₂, 10 mM Hepes, 1 mM EGTA, pH 7.2, and the pipette solution contained 5 mM KCl, 135 mM NaCl, 10 mM Hepes, 1.1 mM MgCl₂, pH 7.4. Pipette potential was held at 0 mV throughout the record. Experiments were performed at RT. High resistance seals were formed (1–5 GΩ), indicating minimal contamination by the endoplasmic reticulum (ER) membrane (13). Under these circumstances, it has been proposed that nuclear pores would be occluded or nonconducting because of experimental conditions or lack of cytosolic factors (13, 14, 17).

Ca²⁺ Measurements in Isolated Nuclei.

Single isolated nuclei were loaded as described (11, 18). The membrane impermeant Ca²⁺ probe Calcium Green-1 dextran (30 μg/ml; 30 min at 4°C; Molecular Probes) loaded the nucleoplasm while the membrane permeant Ca²⁺ probe Fluo-3/AM (20 μM; 60 min at 4°C) loaded the nuclear envelope. After loading, the nuclei were washed twice with the intracellular medium and then were equilibrated in the same medium supplemented with 1 μM of ATP and 300 nM Ca²⁺ for a few minutes to load nuclei with Ca²⁺ (11, 18). After that, they were washed twice again with the intracellular buffer (without ATP and Ca²⁺). Experiments were done at RT. No probe leakage was detected during the experiment. Ca²⁺ imaging in single isolated nuclei was performed by using a Zeiss LSM 510 confocal microscope with a Zeiss 63X oil immersion lens, N.A. 1.25. Images were collected at 3 s intervals, and fluorescence was measured by using the Zeiss LSM software package. Ca²⁺-induced fluorescence intensity ratio (F_t/F₀) was plotted as a function of time.

Nuclear Transmembrane Potential (ΔΨ_n) Measurements.

Isolated nuclei were equilibrated in intracellular medium containing the ER marker DiOC₆(3) (200 nM, RT; ref. 19). After 5 min, the nuclear envelope became stained. DiOC₆(3) is a fluorescent cationic lipophyllic dye whose incorporation into lumen is proportional to ΔΨ.

P-CREB Immunofluorescence.

CREB phosphorylation was induced in isolated nuclei by increasing Ca²⁺ in an EGTA-buffered intracellular medium from 65 nM to 1 μM Ca²⁺ (20) or by addition of tolbutamide or AP₄A for 2 min. After 10 min, nuclei were fixed with 0.1% paraformaldehyde (wt/vol) and permeabilized with 0.1% Triton X-100 for 10 min. Nuclei were preincubated with blocking buffer; then, anti-CREB phospho-specific rabbit antibodies were applied for 16 h at 4°C (1:200, Calbiochem). P-CREB was visualized with fluorescein-conjugated secondary antibodies (1 h, RT, 1:64; Sigma). High Ca²⁺ gave the maximum response.

Fluorescence in Situ Hybridization.

Fragments (237 and 530 bp) of c-myc and β-actin cDNAs cloned in pBSSK (Stratagene) were used as templates for digoxigenin-labeled double-stranded DNA probes, synthesized by PCR using specific primers (21, 22). Isolated islet cells were cultured at least for 24 h in a culture medium containing 3 mM of glucose and then were placed in different stimulating conditions in a Krebs–Ringer buffer for 10 min. After that, cells were equilibrated in 3 mM of glucose for 45 min. Only when AP₄A was used as a stimulator, cells were permeabilized as mentioned above. Then, cells were fixed in a solution containing 4% (vol/vol) formaldehyde, 5% (vol/vol) acetic acid, and 0.9% NaCl and permeabilized with 1:1,000 Triton X-100 for 10 min. Hybridization was performed under standard conditions and was revealed by immunofluorescence detection with fluorescein-conjugated antidigoxigenin (1:500; Roche, Barcelona, Spain). Nonspecific binding was reduced with a commercial blocking reagent (Roche, Barcelona, Spain). Cells were counterstained with 1 μM ethidium homodimer for 5 min before visualization under a Zeiss LSM 510 confocal microscope (10× objective, 0.45 N.A.). The average intensity value from each stained cell was calculated and expressed on an arbitrary gray scale from 0 (blackest) to 255 (whitest). Cells whose fluorescence intensity was above the range of values of unstimulated control cells were scored as activated cells for gene expression.

Statistical Analysis.

Statistical analysis was performed by using SIGMAPLOT (Jandel, San Rafael, CA). Values are mean ± SE. Except where indicated, P < 0.05 by Student's t test.

Results

Nuclear Ca²⁺ Changes.

Ca²⁺ changes were measured in the nuclear space of isolated pancreatic beta cells by spot confocal microscopy (Fig. 1A; refs. 9 and 10). Cells equilibrated in an extracellular medium containing 16.7 mM glucose exhibited a transient increase of [Ca²⁺]_n (Fig. 1B). Exposure to the sulfonylurea tolbutamide (20 μM), which directly closes K_ATP channels, produced a similar increase (Fig. 1B). Both increases of [Ca²⁺]_n were still produced in the absence of extracellular Ca²⁺ (Fig. 1C). Increased [Ca²⁺]_n was probably due to direct Ca²⁺ release from the nuclear envelope to the nuclear space (11, 14, 23), because tolbutamide induces [Ca²⁺]_n changes not only in the absence of extracellular Ca²⁺ but in digitonin permeabilized cells (8) with a buffered Ca²⁺ concentration (Fig. 1D). The tolbutamide-induced Ca²⁺ increase was counteracted by the sulfonamide diazoxide, a K_ATP channel opener (Fig. 1D). These results strongly suggest that intracellular K_ATP channels responsible for eliciting nuclear Ca²⁺ signals might be present close to the beta cell nucleus.

Figure 1.

Ca²⁺ changes in the nuclear space in intact cells revealed by spot confocal microscopy. (A) Hybrid phase contrast image with a fluorescent spot (see arrow) focused on an intact cell nucleus loaded with Calcium-Green-1/AM. (B) Intact cells were stimulated with different agents at the time indicated by the line. Values are mean ± SE and were pooled from six cells [16.7 mM glucose (G)] and seven cells [20 μM tolbutamide (T)]. (C) Beta cells were stimulated with 16.7 mM G (n = 5) and 20 μM T (n = 6) in a Ca²⁺-buffered medium containing 5 mM EGTA. (D) Permeabilized cells were stimulated with 100 μM tolbutamide or 100 μM tolbutamide plus 100 μM diazoxide (n = 4) in a Ca²⁺-buffered intracellular medium. G, glucose; T, tolbutamide; D, diazoxide.

Verification of K_ATP Channels on Nuclei.

To verify the existence of a K_ATP channel on the nuclear envelope from mouse beta cells, we used a high-specificity, high-affinity (K_is = 4 nM) binding assay of glibenclamide-BODIPY-FL to the sulfonylurea receptor (SUR1; ref. 14). SUR1 is the molecular complement of the potassium inward rectifier (K_ir; refs. 24 and 25). The association of SUR1 and K_ir6.2 forms the K_ATP channel in beta cells. Fig. 2 A–C shows a confocal optical section of pancreatic beta cells with an intracellular ring-like labeling, suggesting the binding of glibenclamide-BODIPY to the nuclear envelope. Similar results were observed in isolated nuclei (Fig. 2 D–F). The glibenclamide-BODIPY binding site colocalized with rhodamine B hexyl ester, a marker of the nuclear membrane (ref. 14; Fig. 2E). The binding was specific because it was displaced by the nonfluorescent sulfonylurea tolbutamide in a dose-dependent manner (Fig. 2G). This evidence agrees with previous observations reporting an intracellular location of K_ATP channels including perinuclear sites (26–31) in addition to plasma membrane K_ATP channels.

Figure 2.

K_ATP channels on the nuclear envelope. (A) Intact islet cells incubated with glibenclamide-BODIPY-FL showed a strong ring-like labeling around the nucleus. (B) Cell in A identified as beta cell by immunofluorescence against insulin. (C) Colocalization of images A and B. Optical sections range between 1–2 μm. (Bar = 10 μm.) Similar results were obtained in 19 cells from four coverslips from three different cultures. (D) Glibenclamide-BODIPY-FL ring-like staining of an isolated nucleus. (E) Nucleus in D with rhodamine B hexyl ester staining the nuclear envelope. (F) Colocalization of images D and E (n = 21). (Bar = 5 μm.) Confocal images were taken with 1 μm optical slices. (G) Glibenclamide-BODIPY competition with nonfluorescent unlabeled tolbutamide reflected a displacement of the binding site at the nuclear envelope. Glibenclamide labeling is expressed as the percentage of fluorescence intensity with respect to the control condition (0 μM tolbutamide). Results (mean ± SE) were pooled from 94 nuclei for 0 μM tolbutamide, 29 for 1 μM, 20 for 10 μM, and 24 for 1.000 μM from 12 coverslips of 6 different nuclei preparations. (H) Image showing a patch pipette on an isolated nucleus previously identified using ethidium homodimer-1. (I) Single-channel records (filtered at 1 KHz) were obtained from a nucleus membrane excised patch (n = 5). The pipette potential was held at 0 mV. Upward currents represent currents going into the pipette. Single-channel current gives an estimate of 25 pS. A dashed line represents changes in the external solution. (i) ADP (200 μM) increased the burst length. The change from 200 μM ADP to 2 mM ATP rapidly and completely closed K⁺ channels. (ii) The K⁺ channel activity absent in 2 mM ATP was rapidly restored when ATP was removed. (iii) The change from 40 μM to 20 μM ATP increased channel activity. (iv) Diazoxide (200 μM) activates the channel in the presence of ATP, whereas tolbutamide (100 μM) blocks it.

Excised patch-clamp recordings (16) from the nuclear envelope of isolated nuclei in a 140 mM K⁺/5 mM K⁺ solution exhibited K⁺-channel activity with conductance of approximately 25 pS (ref. 25; Fig. 2I). This channel was activated by ADP and inhibited by ATP in a concentration-dependent manner (Fig. 2I). Nuclear K_ATP (nK_ATP) channels display kinetic properties similar to channels found on the beta cell plasma membrane, with openings grouped in bursts separated by long closing periods (data not shown). The pharmacological profile of the nK_ATP channel also resembles that of the plasma membrane K_ATP channel (25). At concentrations similar to those acting on the plasma membrane K_ATP channel, tolbutamide (100 μM) blocks the nK_ATP channel whereas diazoxide (200 μM) opens it (Fig. 2Iiv) in the presence of ATP. Therefore, this nK_ATP channel seems remarkably similar to the plasma membrane K_ATP channel (25).

Blockade of nK_ATP Channels Elicits Ca²⁺ Signals in Isolated Nuclei.

Depending upon their lipophylicity, the Ca²⁺-sensitive fluorescence dyes Fluo-3/AM or Calcium Green-1 dextran localized preferably in the nuclear envelope or the nucleoplasm (11, 14, 18), respectively (Fig. 3 A and B). In isolated nuclei, confocal microscopy revealed that tolbutamide induced opposite Ca²⁺ changes in the nuclear envelope and the nucleoplasm (Fig. 3 A and B). Either 100 μM or 500 μM of tolbutamide generated Ca²⁺ increases in the nucleoplasm, which paralleled Ca²⁺ decreases observed in the nuclear envelope (Fig. 3 A and B). Tolbutamide did not produce any Ca²⁺ transient in the presence of the K_ATP channel opener diazoxide (data not shown). Moreover, the diadenosine polyphosphate AP₄A (100 μM), which blocks the K_ATP channel (2), induced a similar Ca²⁺ release (Fig. 3C). Because AP₄A is membrane-impermeant, this experiment also suggests that the regulatory site of this K⁺ channel does not face the perinuclear space. Thus, the nuclear envelope may act as a Ca²⁺ reservoir that is mobilized as a result of glucose metabolism.

Figure 3.

Nuclear Ca²⁺ signals induced by K_ATP channel blockade. (A) Nuclei were loaded with the membrane permeant probe Fluo-3/AM, which was preferentially accumulated in the nuclear envelope (ref. 11; see image). [Ca²⁺]_n was measured by using confocal microscopy. Tolbutamide applied to Fluo-3-loaded nuclei produced a Ca²⁺ decrease (n = 3). (B) Calcium Green-1 dextran, a nonpermeant probe, was distributed uniformly in the nucleoplasm (see image). Tolbutamide provoked a transient Ca²⁺ increase in the nucleoplasm (n = 6). Spot confocal methods also were used to observe this tolbutamide-induced nuclear Ca²⁺ release (n = 3; data not shown). (Bar = 5 μm.) (C) Same experiment described in B showing a Ca²⁺ release induced by 100 μM AP₄A (n = 3). (D) Replacement of 100 mM K⁺ by NMG⁺ (N-methyl-d-glucamine) led to a similar Ca²⁺ transient (n = 4) in the nucleoplasm. (E) Nuclear envelope was loaded with DiOC₆(3), a voltage-sensitive probe (see Materials and Methods). Fluorescence was increased upon addition of tolbutamide (n = 4). (F) Ruthenium red (RR) blocked almost 90% of the tolbutamide-induced Ca²⁺ release (n = 4). Two consecutive Ca²⁺ transients were induced in this experiment. Before the second discharge, nuclei were loaded with Ca²⁺ (see Materials and Methods) and then treated with ruthenium red. When the blocker was not present, the two transients had the same amplitude. Thus, the first transient was used as a control (C).

All of these results suggested that this Ca²⁺ pool was sensitive to changes in the nuclear membrane K⁺ permeability but also raised the possibility that Ca²⁺ release may be elicited by means of a voltage-dependent mechanism. To explore the hypothesis that blockade of nK_ATP channels provokes a Ca²⁺ release from the nuclear envelope by voltage variations, we changed the concentrations of extraluminal K⁺. Reduction of the K⁺ concentration (by N-methyl-d-glucamine replacement) and the ensuing change in the K⁺ electrochemical gradient led to a corresponding Ca²⁺ discharge (Fig. 3D). Tolbutamide failed to produce a Ca²⁺ transient in nuclei pretreated with 10 μM of the K⁺ ionophore valinomycin, further suggesting that the collapse of K⁺ electrochemical gradient can suppress Ca²⁺ release (data not shown). We monitored changes of nuclear transmembrane potential (ΔΨ_n) by using DiOC₆(3), which has been validated as a potentiometric probe in mitochondria (ref. 19; see Materials and Methods). DiOC₆(3) accumulated in the nuclear envelope of beta cells. Tolbutamide elicited a consistent increase in DiOC₆(3) fluorescence, indicating that the perinuclear lumen became more negative (Fig. 3E). These results pointed to the existence of a ΔΨ_n, which may change as a result of a decrease in K⁺ permeability. In beta cells, a [Ca²⁺]_n pathway sensitive to this ΔΨ_n may exist.

Several intracellular channels found in the endo-sarcoplasmic reticulum (ER/SR) and nucleus, including inositol-1,4,5-trisphosphate receptors (InsP₃-R) and ryanodine receptors (RyR; refs. 11, 14, and 32) among others, have been shown to be voltage-sensitive. Ca²⁺ release induced by tolbutamide decreased 81.3 ± 9.3% in isolated nuclei exposed to 10 μM ruthenium red, a RyR blocker (Fig. 3F). Higher ruthenium red concentrations (100 μM) completely blocked the Ca²⁺ release. Similar results were observed when isolated nuclei were preincubated with anti-RyR antibodies (1:100, 30 min; data not shown). Conversely, heparin, a blocker of the InsP₃-R (15) did not produce a significant effect (data not shown).

Nuclear Function of nK_ATP Channels.

Several roles have been suggested for nuclear channels, yet very few studies relate these channels to nuclear functions. As has recently been proposed, nuclear Ca²⁺ signals may target the CREB protein, whose phosphorylation may activate transcription (20, 33). By using immunofluorescence in isolated functional nuclei (20), we have observed CREB phosphorylation (Fig. 4) induced by tolbutamide and AP₄A, both K_ATP blockers that provoked an increase of [Ca²⁺]_n (Fig. 3). As shown in Fig. 1D, the K_ATP channel opener diazoxide counteracted the tolbutamide effects (Fig. 4D). These observations further suggested that the closure of nK_ATP channels initiate the transduction of nuclear signals.

Figure 4.

CREB phosphorylation in isolated nuclei. Immunofluorescence detection of P-CREB in nuclei treated with different stimuli. (A) Ca²⁺ (1 μM). (B) Tolbutamide (100 μM). (C) Ca²⁺ (65 nM). Immunodetection is shown in green, and ethidium homodimer-1 staining is shown in red. (D) Percentage of labeled nuclei relative to the maximal response 1 μM Ca²⁺ (HC; n = 145) measured in the following conditions: 100 μM tolbutamide (T; n = 112); 100 μM tolbutamide plus 200 μM diazoxide (T+D; n = 92); 100 μM AP₄A (n = 56); and 65 nM Ca²⁺ (LC; n = 134).

In addition to transcription factors, other specific intranuclear Ca²⁺ targets may exist that regulate specific nuclear events, such as gene induction. These targets might include nuclear kinases, polymerases and chromatin remodeling, among others (34). To evaluate the relationship between these nuclear Ca²⁺ signals and gene expression, we conducted fluorescence in situ hybridization experiments in intact beta cells. C-myc expression was chosen as a model because this immediate early gene is activated by glucose in a Ca²⁺-dependent manner in both cultured cell lines (21; E.R., unpublished work) and in islet cells (35). Fig. 5 shows c-myc expression in islet cells under different stimuli. Remarkably, tolbutamide and AP₄A activated gene expression (30 and 20%, respectively) even in the absence of extracellular Ca²⁺. In this condition, both K_ATP blockers induced Ca²⁺ changes within the nuclear space (Figs. 1C and 3 A–C). Diazoxide abolished these effects when present (data not shown).

Figure 5.

Gene expression induced by K_ATP channel blockers in the absence of extracellular Ca²⁺. Fluorescence in situ hybridization and transmitted light images of islet cells stimulated in different conditions. (A) Glucose (16 mM). (B) Tolbutamide (100 μM) in the absence of extracellular Ca²⁺ (buffered with EGTA). (C) Glucose (3 mM) as a control condition. The green color shows c-myc expression of beta cells counterstained with ethidium homodimer (red). (Bar = 10 μm.) (D) Percentage of activated cells relative to unstimulated control cells (3 mM glucose; n = 360) in different conditions: 40 mM potassium (K⁺; n = 290); 16 mM glucose (G; n = 369); 100 μM tolbutamide (T; n = 329); 100 μM tolbutamide in the absence of extracellular Ca²⁺ [T₍₀₎; n = 466]; 100 μM AP₄A in the absence of extracellular Ca²⁺ (n = 105); and absence of extracellular Ca²⁺ [Ca²⁺₍₀₎; n = 211]. Results were pooled from three independent experiments. β-actin constitutive expression was used as an invariable control under the same conditions.

Discussion

Exposure of beta cells to glucose results in a Ca²⁺-mediated activation of a broad range of functions (3–6). However, the link between glucose-induced Ca²⁺ signals and specific organelles remains obscure. Here, we focused on the mechanism of glucose-regulated [Ca²⁺]_n fluctuations and the effect of these Ca²⁺ signals on nuclear function. Exposure of intact cells to glucose and tolbutamide led to a [Ca²⁺]_n rise both in cells equilibrated in normal Krebs–Ringer containing 2.5 mM Ca²⁺ and in Ca²⁺-free Krebs–Ringer solution (Fig. 1 B and C). These results suggest that both secretagogues must be acting on intracellular Ca²⁺ compartments and are in agreement with previous observations in beta cells (36) showing that tolbutamide can increase cytosolic [Ca²⁺] in the absence of extracellular Ca²⁺. Mitochondria have been suggested as an intracellular target of sulfonylureas in beta cells (36). However, the intracellular source of this sulfonylurea-induced Ca²⁺ release was not univocally identified. Although Ca²⁺ discharge from an organelle could diffuse to the nuclear space, our observations that tolbutamide can increase [Ca²⁺]_n as well in permeabilized cells equilibrated in Ca²⁺-buffered intracellular medium (Fig. 1D), indicate that in this case the source of Ca²⁺ may be located very close to the nucleus. This tolbutamide effect was readily blocked by diazoxide, a K_ATP channel opener (Fig. 1D). This observation indicates that K_ATP channels may be present near the nucleus and their closure may control [Ca²⁺]_n in beta cells. Excised patches of nuclear membrane confirmed the presence of a nuclear nK_ATP channel that exhibits kinetics and pharmacological properties very similar to the K_ATP channels found in the plasma membrane (ref. 25; Fig. 2I). Patch scission can provoke membrane remodeling, making it difficult to know the channel orientation by using the patch-clamp approach. However, the effect of the membrane-impermeant K_ATP channel blocker AP₄A (2) on Ca²⁺ release in isolated nuclei (Fig. 3D) suggests that the regulatory site of this channel is not facing the perinuclear space.

In addition to this functional characterization we have demonstrated here a specific binding site for sulfonylureas at the beta cell nuclear envelope (Fig. 2) that was revealed by glibenclamide-BODIPY-FL binding. This method has been successfully used to detect the presence of K_ATP channels in different kinds of cells including beta cells, neurons, and monocytes (15). Our results are in agreement with previous work in cells transfected with green fluorescent protein-tagged constructs linked to SUR1 or Kir6.2 that showed fluorescence patterns not only in plasma but in perinuclear membranes, as well as in ER and Golgi (30, 31). Thus, three independent lines of evidence point to the presence of a K_ATP channel in the nuclear envelope, broadening the idea that together with the well characterized plasma membrane K_ATP channels (24, 25), sulfonylureas might have multiple sites of action (26), including mitochondria (27), secretory granules (28, 29), and, as shown here, the nuclear envelope.

Our results show that nK_ATP channels are involved in beta cell nuclear Ca²⁺ signaling because specific K_ATP blockers such as tolbutamide (25) and AP₄A (2) induced Ca²⁺ release from the nuclear envelope to the nucleoplasm in isolated nuclei (Fig. 3 A–C). Several lines of evidence suggest that the nuclear Ca²⁺ pathway revealed here depends on a change in K⁺ permeability and ΔΨ_n in beta cell nuclei. First, valinomycin abolished tolbutamide-induced Ca²⁺ transients. Second, a change in K⁺ electrochemical gradient by NMG⁺ replacement in the extraluminal medium produced a similar Ca²⁺ discharge (Fig. 3D). Third, tolbutamide provoked a transient increase in ΔΨ_n (Fig. 3E) as revealed the voltage-sensor DiOC₆(3) (19, 37). It is plausible that a potential difference between perinuclear space and cytosol exists because the two membranes of the nuclear envelope surround a lumen that separates different subcellular compartments and contains a diversity of channels with different permeabilities (23, 34). The nuclear envelope is a structural extension of the ER/SR network. Thus, similar properties and molecular composition are shared among these organelles. The existence of a low resting transmembrane potential described in the SR (38) is compatible with transient voltage increases during stimulation similar to those reported here.

Although the understanding of the molecular components of the nuclear envelope and their regulation have undergone important progress, their involvement in nuclear Ca²⁺ signaling still remains a challenge for exploration (23, 34). A variety of ER/SR and nuclear channels that may contribute to regulate [Ca²⁺]_n and other ion species have been reported (13, 32, 39–44). An important feature is that several ER/SR and nuclear channels are voltage-sensitive. Those include Cl⁻ (13), K⁺ (42), RyR (32), InsP₃-R (14), and InsP₃-insensitive Ca²⁺ channels (43, 44). Our results are consistent with RyR channels at the nuclear envelope—as observed in exocrine cells (11) and oocytes (45)—that may control the nuclear Ca²⁺ release reported here, because their blockade by ruthenium red or anti-RyR antibodies suppressed the tolbutamide-induced nuclear Ca²⁺ transients (Fig. 3F). It is noteworthy that RyR channels undergo an active state of elevated open probability upon changes of voltage (32). Thus, it is very likely that RyR channels may be affected by tolbutamide-induced nuclear voltage changes (Fig. 3E). Their activation may lead to Ca²⁺-release, which may be amplified by a Ca²⁺-induced Ca²⁺-release process, as has been observed in ER of pancreatic beta cells (46).

Our results show that this tolbutamide-induced [Ca²⁺]_n conductance seems to be associated with a change in ΔΨ_n as a result of a decrease in K⁺ permeability (Fig. 3 D and E). This process is not exclusive to beta cells. In fact, there is a body of evidence that establishes a link between K⁺ and Ca²⁺ conductances in the ER/SR. For instance, Ca²⁺ release from SR has been induced not only by K⁺ channel general blockers (47) but also by the specific K_ATP channel blocker glibenclamide (48), suggesting the presence of this K⁺ channel type and its role in intracellular Ca²⁺ release. Moreover, lowering the extraluminal K⁺ produced a Ca²⁺ discharge from SR microsomes of myocytes, suggesting a voltage change that is not explained well by current ideas about E–C coupling in muscle cells (49). Furthermore, a functional K_ATP channel has been proposed to control the K⁺ permeability and the granular membrane potential in glucagon-containing secretory granules (29). The presence of nuclear Ca²⁺ channels sensitive to changes in transmembrane potential explains tolbutamide-induced Ca²⁺ release in beta cell nuclei. Nonetheless, we do not exclude the involvement of other voltage-sensitive pathways that may direct or indirectly trigger the nuclear Ca²⁺ release reported here. These possible pathways may include ion exchange fluxes coupled to Ca²⁺ (50, 51) or other mechanisms such as the ER Ca²⁺ leak pathway (52). In fact, a remarkable characteristic is that the ER leak pathway is increased with high concentrations of ATP, a blocker of the K_ATP channel (52). In summary, we suggest that blockade of K_ATP channels at the nuclear envelope may elicit a transient transmembrane potential rise as a result of K⁺ permeability decrease, which may trigger a voltage-sensitive Ca²⁺ discharge, mainly through RyR channels.

By using immunofluorescence, we have proved that the nuclear Ca²⁺ transients reported here increased the phosphorylation rate of the transcription factor CREB (Fig. 4) in isolated nuclei, in agreement with similar results described in hippocampal neuron nuclei (20). Our working model only allows Ca²⁺ mobilization from the nucleus and unambiguously involves nK_ATP channels in pancreatic beta cell nuclear function.

In a whole-cell model, we evaluated the possibility of other nuclear processes such as gene expression being affected. Tolbutamide and AP₄A, which induced Ca²⁺ release in isolated nuclei (Fig. 3), triggered c-myc expression in the absence of extracellular Ca²⁺ (Fig. 5). In the whole cell, we cannot rule out the possibility that other Ca²⁺ release in the cytosol may drive gene expression in our conditions. However, as shown here, the nucleus probably accounts for an important part of these intracellular Ca²⁺ signals (Figs. 1 and 3). Be that as it may, we have demonstrated that these nuclear Ca²⁺ signals are sufficient to drive CREB phosphorylation in isolated nuclei in these cells. Hence, Ca²⁺ signals generated at the nuclear envelope have an important role for nuclear function in beta cells.

We propose that signals generated by glucose metabolism not only inhibit the plasma membrane K_ATP channel, initiating insulin release, but also interact with nK_ATP channels, triggering nuclear Ca²⁺ signals that modulate nuclear functions such as phosphorylation of the transcription factor CREB and likely gene expression. Our data shows a new signal-transduction pathway linking nuclear K⁺ ion channels to fluctuations of [Ca²⁺]_n and nuclear function and also contributes to an understanding of the mechanism of action of sulfonylureas, drugs that are broadly used in the clinic for the treatment of the diabetic patient.

Abbreviations

K_ATP

ATP-sensitive K⁺ channel

[Ca²⁺]_n

nuclear Ca²⁺ concentration

CREB

cAMP response element binding protein

RT

room temperature

BODIPY

dipyrrometheneboron difluoride

nK_ATP

nuclear K_ATP channel

ER

endoplasmic reticulum

ER/SR

endo-sarcoplasmic reticulum

RyR

ryanodine receptors