Corresponding author L. A. Blatter: Department of Physiology, Loyola University Chicago, 2160 S. First Ave., Maywood, IL 60153, USA. Email: firstname.lastname@example.org
In cardiac myocytes the type-2 inositol 1,4,5-trisphosphate receptor (IP3R2) is the predominant isoform expressed. The IP3R2 channel is localized to the SR and to the nuclear envelope. We studied IP3-dependent nuclear Ca2+ signals ([Ca2+]Nuc) in permeabilized atrial myocytes and in isolated cardiac nuclei. In permeabilized myocytes IP3 (20 μm) and the more potent IP3R agonist adenophostin (5 μm) caused an elevation of [Ca2+]Nuc. An IP3-dependent increase of [Ca2+]Nuc was still observed after pretreatment with tetracaine to block Ca2+ release from ryanodine receptors (RyRs), and the effect of IP3 was partially reversed or prevented by the IP3R blockers heparin and 2-APB. Isolated nuclei were superfused with an internal solution containing the Ca2+ indicator fluo-4 dextran. Exposure to IP3 (10 μm) and adenophostin (0.5 μm) increased [Ca2+]Nuc by 25 and 27%, respectively. [Ca2+]Nuc increased to higher levels than [Ca2+]Cyt immediately adjacent to the outer membrane of the nuclear envelope, suggesting that a significant portion of nuclear IP3 receptors are facing the nucleoplasm. When nuclei were pretreated with heparin or 2-APB, IP3 failed to increase [Ca2+]Nuc. Isolated nuclei were also loaded with the membrane-permeant low-affinity Ca2+ probe fluo-5N AM which compartmentalized into the nuclear envelope. Exposure to IP3 and adenophostin resulted in a decrease of the fluo-5N signal that could be prevented by heparin. Stimulation of IP3R caused depletion of the nuclear Ca2+ stores by approximately 60% relative to the maximum depletion produced by the ionophores ionomycin and A23187. The fluo-5N fluorescence decrease was particularly pronounced in the nuclear periphery, suggesting that the nuclear envelope may represent the predominant nuclear Ca2+ store. The data indicate that IP3 can elicit Ca2+ release from cardiac nuclei resulting in localized nuclear Ca2+ signals.
Ca2+ plays a key role in nuclear functions, including gene transcription and DNA replication (for review see Bading et al. 1997; Hardingham et al. 1998), conformation of and transport through the nuclear pore (Stehno-Bittel et al. 1995a,b, 1996) and structuring of the nuclear envelope (Subramanian & Meyer, 1997). Whereas the importance of Ca2+ for nuclear activities is undebated (see Irvine, 2003), it remains controversial what sources of Ca2+ affect or control nuclear [Ca2+] ([Ca2+]Nuc). There is evidence that [Ca2+]Nuc is exclusively controlled by cytosolic [Ca2+] ([Ca2+]Cyt) and Ca2+ diffusion (Lipp et al. 1997). In contrast, evidence is growing that the nucleus contains the complete machinery for both inositol 1,4,5-trisphosphate receptor (IP3R)- and ryanodine receptor (RyR)-dependent Ca2+ release. There are Ca2+ pumps in the outer nuclear membrane (Irvine, 2003) and RyR-dependent Ca2+ release has been demonstrated in isolated nuclei (Gerasimenko et al. 2003). The nucleus contains the key components of the phosphoinositide–phospholipase C (PLC) signalling cascade (Chi & Crabtree, 2000) and is able, with its own PLC (presumably mainly the PLC-β1 isoform), to generate diacylglycerol (DAG) and IP3 (extensively reviewed recently by Irvine, 2003). DAG has been shown to control nuclear protein kinase C (PKC) functions (e.g. D'Santos et al. 1999). Early evidence of IP3-dependent Ca2+ release from a nuclear Ca2+ pool was first reported in isolated liver nuclei (Nicotera et al. 1990). Single channel recordings from nuclear IP3R have been described also (Stehno-Bittel et al. 1995a). There is direct evidence of IP3Rs facing the nucleoplasm influencing [Ca2+]Nuc (e.g. Hennager et al. 1995) and of IP3-mediated release of Ca2+ from a reticular membrane network within the nucleus (Echevarria et al. 2003). The data suggest that the nucleus itself contains all components required for autonomous IP3-dependent Ca2+ signalling.
Cardiac myocytes also express IP3Rs. The type-2 isoform of the IP3R is the main subtype expressed in cardiac muscle. It is present at one to two orders of magnitude lower density than the RyR, but atrial myocytes express more functional IP3Rs than ventricular myocytes (for references see Blatter et al. 2003; Zima & Blatter, 2004). IP3-dependent Ca2+ release in the heart appears to be important during development and in cardiac injury and pathologies (cf. Woodcock et al. 1998); however, the role of IP3R-dependent Ca2+ signalling in cardiac excitation–contraction coupling (ECC) has remained highly controversial. A number of more recent reports (Mackenzie et al. 2002; Zima & Blatter, 2004; Li et al. 2005) on atrial Ca2+ signalling and ECC indicate that in atrial myocytes IP3-dependent Ca2+ release enhances basal [Ca2+]i, and has positive inotropic effects by enhancing twitch [Ca2+]i transient amplitude. However, stimulation of IP3Rs also enhances spontaneous Ca2+ spark activity and the propensity of developing spontaneous Ca2+ release such as spontaneous Ca2+ waves and action potentials and the development of arrhythmogenic Ca2+ alternans. Most recently, IP3Rs have also been implicated in arrhythmogenesis in ventricular myoctes (Proven et al. 2006). In ventricular myocytes IP3Rs have been found to localize to the nuclear envelope (Bare et al. 2005), but the functional role of IP3Rs at this specific location has remained elusive.
Based on our previous observations that (a) in atrial cells IP3-dependent Ca2+ release from the SR makes significant contributions to Ca2+ signalling during ECC (Zima & Blatter, 2004; Li et al. 2005), (b) agonist-induced increases in [IP3] could be evidenced directly with a novel FRET-based IP3 sensor in intact cardiac myocytes (Remus et al. 2006), and (c) IP3Rs are preferentially located to the nuclear envelope (Bare et al. 2005), we set out to investigate IP3-dependent Ca2+ release from the nuclear envelope in cardiac tissue. Using permeabilized myocytes and isolated cardiac nuclei we demonstrate here that IP3 increases [Ca2+]Nuc and [Ca2+]Cyt in the close vicinity of the nuclear membrane, but with different kinetics and spatio-temporal patterns. Furthermore, the data suggest that IP3Rs are located in the inner membrane (facing the nucleoplasm) and outer membrane (facing the cytosol) of the nuclear envelope. Activation of the IP3Rs in both locations directly released Ca2+ from the nuclear envelope. A previous account of this work has been presented in abstract form (Zima et al. 2005).
Isolation of cat atrial myocytes
Single myocytes were isolated from cat atria as previously described (Wu et al. 1991; Kockskamper & Blatter, 2002; Sheehan & Blatter, 2003). Cats (n= 6) were obtained from R & R Research (Howard City, MI, USA). The vendor and the procedure for cell isolation were fully approved by the Institutional Animal Care and Use Committee of Loyola University Medical Center. Briefly, adult cats of either sex were anaesthetized with thiopental sodium (35 mg kg−1i.p.). Following thoracotomy hearts were quickly excised, mounted on a Langendorff apparatus, and retrogradely perfused with oxygenated collagenase-containing solution at 37°C. Cells were plated on coverslips for later experimentation, and bathed in Tyrode solution before permeabilization (composition of Tyrode solution in mm: NaCl 140; KCl 4; CaCl2 2; MgCl2 1; glucose 10; Hepes 10; pH 7.4 adjusted with NaOH). All experiments were carried out at room temperature (22–25°C).
Isolation of cardiac nuclei
Nuclei from cardiac myocytes were recovered following their selective sedimentation through 2.4 m sucrose buffer with centrifugation (50 000 gmax) utilizing the low-speed pellet (3800 gmax) fractionated from homogenized whole rat hearts following procedures previously described (Bare et al. 2005). Frozen rat hearts were obtained from Pel-Freez Biologicals (Rogers, AR, USA). The obtained nuclei were resuspended in buffer with 0.32 m sucrose. A step-down exchange and equilibration was made from this buffer into an internal solution of the same composition used for permeabilized cells (see below).
Intracellular calcium measurements
[Ca2+]Nuc and [Ca2+]Cyt were measured in permeabilized atrial myocytes with fluorescence laser scanning confocal microscopy (Radiance 2000 MP, Bio-Rad, UK) using the fluorescent Ca2+ indicator fluo-3. Cells were permeabilized by exposure to 0.005% (w/v) saponin (Zima et al. 2003). After 30 s the bath solution was exchanged to a saponin-free internal solution composed of (mm): K+ aspartate 100; KCl 15; KH2PO4 5; MgATP 5; EGTA 0.4; CaCl2 0.12; MgCl2 0.75; phosphocreatine 10; creatine phosphokinase 5 U ml−1; dextran (MW: 40 000) 8%; Hepes 10; fluo-3 potassium salt 0.03; pH 7.2 (KOH). Free [Ca2+] and [Mg2+] of this solution were 100 nm and 1 mm, respectively (calculated using WinMAXC 2.05, Stanford University, CA, USA). Fluo-3 was excited with the 488 nm line of an argon ion laser and fluorescence was measured at wavelengths > 515 nm. Ca2+ spark frequencies are expressed as number of observed sparks per second and per 100 μm of scanned distance (sparks s−1 (100 μm)−1).
[Ca2+] measurements from isolated nuclei
Two fluorescent probes were used to study nuclear Ca2+ dynamics. Nucleoplasmic [Ca2+] ([Ca2+]Nuc) was measured with fluo-4 dextran (potassium salt; MW 10 000; 20 μm). Isolated nuclei were loaded with fluo-4 dextran by incubation in an internal solution for 1 h at 4°C (Gerasimenko et al. 2003). After loading extranuclear indicator was washed (except in the experiment shown in Fig. 5), whereas nucleoplasmic dye remained entrapped. Fluo-4 was calibrated by exposure of isolated nuclei to an internal solution with free [Ca2+] ranging from 5 nm to 1 mm. The measured Kd of fluo-4–Ca binding was 350 nm inside the nucleus, compared to 327 nm in the extranuclear environment. Changes in [Ca2+] inside the nuclear envelope were assessed by fluo-5N (MW 958). The isolated nuclei were loaded with 20 μm of the membrane-permeant fluo-5N AM for 3–4 h at room temperature. Fluo-4 dextran and fluo-5N were excited with the 488 nm line of an argon ion laser and fluorescence was measured at wavelengths > 515 nm.
Fluo-3 pentapotassium salt, fluo-4 dextran and fluo-5N AM were obtained from Molecular Probes/Invitrogen (Carlsbad, CA, USA). Inositol-1,4,5-trisphosphate (IP3) and adenophostin were obtained from EMD Biosciences/Calbiochem (San Diego, CA, USA). 2-Aminoethoxydiphenyl borate (2-APB), A23187, heparin (MW: 6000), ionomycin and tetracaine were from Sigma (St Louis, MO, USA).
Results are reported as means ± standard error of the mean (s.e.m.) for the indicated number (n) of cells or isolated nuclei. Statistical significance was evaluated using Student's t test.
Differential enhancement of [Ca2+]Cyt and [Ca2+]Nuc signals by IP3
Permeabilized atrial myocytes were exposed to 20 μm of the physiological IP3R agonist inositol-1,4,5-trisphosphate (IP3). As shown in Fig. 1A, IP3 had differential effects on [Ca2+]Cyt and [Ca2+]Nuc. In the cytosol, IP3 caused an overall increase of basal [Ca2+] and a transient enhancement of Ca2+ spark activity (lower trace in Fig. 1A). This is consistent with our previous observations (Zima & Blatter, 2004) that IP3-dependent Ca2+ release facilitates SR Ca2+ release via RyR, and thus the enhanced Ca2+ spark frequency. On average (Fig. 1B) IP3 increased Ca2+ spark frequency from 6.6 ± 0.7 (control; n= 7 cells) to 9.8 ± 1.2 sparks s−1 (100 μm)−1 after 30 s exposure to IP3 (P < 0.05). After 120 s exposure to IP3 Ca2+ spark frequency had returned to control levels (7.1 ± 1.0 sparks s−1 (100 μm)−1). The increase of basal [Ca2+]Cyt was also transient (Fig. 1C). In contrast, in the nucleus IP3, after a delay of 10–15 s, caused a steady rise of [Ca2+]Nuc (Fig. 1D), while the enhanced Ca2+ spark activity and the elevation of basal [Ca2+]Cyt were transient.
IP3-dependent nuclear Ca2+ signals during RyR inhibition
While the data shown in Fig. 1 indicate some degree of independent regulation of [Ca2+]Cyt and [Ca2+]Nuc, they did not allow us to determine whether this was due solely to Ca2+ release from IP3R controlled stores or whether IP3R-dependent Ca2+ release facilitated Ca2+ release from RyRs in the nuclear compartment. To distinguish between these possibilities we inhibited RyRs with tetracaine (0.7 mm) prior to exposure of permeabilized atrial cells to IP3. Cells were bathed in an internal solution containing the Ca2+ indicator fluo-3. We have shown previously that with the tetracaine protocol RyR activity is inhibited and allows for visualization of IP3R-dependent Ca2+ release (including elementary IP3R-dependent Ca2+ release events, termed Ca2+ puffs) in cardiac tissue (Zima & Blatter, 2004). As shown in Fig. 2A in the presence of tetracaine IP3 caused a gradual increase of [Ca2+]Nuc which was partially reversed by the subsequent application of the IP3R blocker heparin. On average IP3 increased [Ca2+]Nuc by 15 ± 1% (n= 7; P < 0.01). The subsequent addition of heparin decreased [Ca2+]Nuc to 9 ± 2% above control, i.e. ∼40% of the IP3-dependent increase of [Ca2+]Nuc was sensitive to subsequent exposure to heparin. As shown in Fig. 2B a similar rise of [Ca2+]Nuc was observed with adenophostin. Adenophostin is one of the most potent IP3R agonists, binds to the receptor with much higher affinity than IP3 itself and is not subject to cellular enzymatic degradation (Takahashi et al. 1994). Adenophostin increased [Ca2+]Nuc on average by 12 ± 2% (n= 8; P < 0.01). The data suggest the presence of IP3-dependent Ca2+ release from the nucleus in cardiac myocytes.
IP3-dependent Ca2+ signals in isolated cardiac nuclei
While the data shown in Figs 1 and 2 are strong indications that the source of Ca2+ responsible for the observed IP3-dependent rise of [Ca2+]Nuc resides in the nucleus, we had to exclude the possibility that the elevation was due solely to Ca2+ diffusion from the cytosolic compartment (although it seemed that this would have to occur against a concentration gradient, cf. Fig. 1). To address the issue of nuclear IP3R-dependent Ca2+ release further, we exposed individual nuclei isolated from cardiac tissue to IP3R agonists. Cardiac nuclei were bathed in an internal solution and loaded with the Ca2+ indicator fluo-4 dextran. With a molecular weight of ∼10 000, the Ca2+ dye fluo-4 dextran can be considered a rather immobile Ca2+ buffer that remained entrapped in the nucleoplasm. As shown in Fig. 3 IP3 and adenophostin caused a transient increase of [Ca2+]Nuc that averaged 25 ± 1% (n= 36; P < 0.001 compared to control levels) and 27 ± 1% (n= 12; P < 0.001), respectively. In contrast, caffeine caused only a small and brief elevation of [Ca2+]Nuc (10 ± 2%; n= 23; P < 0.001). While all three agonists caused a statistically significant elevation of [Ca2+]Nuc, in comparison the effect of caffeine was significantly (P < 0.001) smaller than the effect of IP3 and adenophostin, respectively. Figure 4 shows that application of IP3 after pretreatment with the IP3R blockers heparin and 2-APB failed to increase [Ca2+]Nuc significantly. Figure 5 illustrates the differential changes [Ca2+]Nuc and [Ca2+]Cyt. [Ca2+]Cyt was measured immediately adjacent to the nuclear envelope from a ∼1 μm wide region surrounding the nucleus (Fig. 5A). For this experiment fluo-4 dextran was kept in the bathing solution. Stimulation with IP3 caused a significantly larger and prolonged increase in [Ca2+]Nuc compared to [Ca2+]Cyt immediately adjacent to the nuclear border; however, [Ca2+]Cyt reached its peak faster than [Ca2+]Nuc (Fig. 5B), i.e. [Ca2+]Nuc continued to rise while [Ca2+]Cyt already began to decline. The observation of a prolonged rise of [Ca2+]Nuc that outlasted the rise of [Ca2+]Cyt, and [Ca2+]Nuc exceeding [Ca2+]Cyt in amplitude supports the hypothesis that a significant fraction of IP3Rs embedded in the nuclear envelope membrane face towards the nucleoplasm and are capable of directly releasing Ca2+ into this compartment. The elevation of [Ca2+]Cyt immediately adjacent to the nucleus could be due to release of Ca2+ from IP3Rs facing the cytosol (i.e. anchored in the outer membrane of the nuclear envelope) and/or diffusion of Ca2+ out of the nucleus. The presence of nuclear IP3Rs facing the cytosol would be supported by the observation that [Ca2+]Cyt began to rise shortly before [Ca2+]Nuc and had a faster time to peak (TTP).
IP3-dependent depletion signals from the nuclear envelope
The low affinity Ca2+ indicator fluo-5N has been used successfully to measure intra-SR [Ca2+] and Ca2+ release via RyR directly in cardiac myocytes (Shannon et al. 2003; Picht et al. 2006; Wu et al. 2006). Here we applied this technique to isolated nuclei for the direct measurement of Ca2+ depletion of the nuclear envelope. The top panel of Fig. 6A shows a series of images from an isolated nucleus loaded with fluo-5N. Fluo-5N was loaded into the nuclear envelope using the membrane-permeant ester form of the indicator. The three images show fluo-5N in the nuclear envelope under control conditions, after exposure to IP3 followed by treatment with the Ca2+ ionophore A23187. IP3 and A23187 led to a progressive decrease of fluo-5N fluorescence that was particularly pronounced in the nuclear periphery (see profile top right). This observation was consistent with the idea that the nuclear envelope is the nuclear Ca2+ store and that the decrease in fluorescence was the result of Ca2+ depletion of the nuclear envelope. Adenophostin caused a similar Ca2+ depletion of the nuclear envelope (Fig. 6A, middle), whereas caffeine had little effect on the fluo-5N signal (Fig. 6A, bottom), suggesting that the membrane of the cardiac nuclear envelope is poorly equipped with RyRs compared to IP3Rs. In support of these results, Bare et al. (2005) showed that an isolated cardiac nuclei fraction contained only sparse levels of RyRs compared to the enrichment of IP3R2, and immunostaining of isolated ventricular myocytes with an IP3R2 specific antibody demonstrated receptor localization primarily on the nuclear envelop whereas RyRs were not visualized on the envelope or in the perinuclear area in contrast to their SR localization. The Ca2+ ionophores ionomycin and A23187 were used to deplete the nuclear Ca2+ stores completely. As summarized in Fig. 6B, IP3 and adenophostin caused an approximately 20% decrease of the nuclear fluo-5N signal which was further decreased by the ionophores to ∼65–70% of control. Thus, IP3 and adenophostin released on average ∼55–70% of the total releasable Ca2+ from the nuclear Ca2+ stores. Figure 7 shows that preincubation of fluo-5N loaded nuclei with heparin largely prevented the IP3-induced depletion of the nuclear envelope.
In summary, these experiments indicate that in cardiac myocytes the nuclear envelope is a functional Ca2+ store from which release is predominantly controlled by IP3Rs, with a significantly smaller contribution from RyRs. Furthermore, it appears that a significant fraction of Ca2+ release from the nuclear envelope occurred through IP3Rs facing the nucleoplasm.
In this study we demonstrated directly, using permeabilized myocytes and isolated cardiac nuclei, that Ca2+ can be released from the nuclear envelope by stimulation with IP3. Thus, the nuclear envelope of cardiac myocytes is a functional Ca2+ store from which release is predominantly controlled by IP3 receptors and to a significantly lesser extent by release of Ca2+ via ryanodine receptors.
It is well established that many nuclear functions are regulated by [Ca2+]Nuc; however, the source of Ca2+ for observed changes of [Ca2+]Nuc has and still is a matter of debate. Experimental evidence indicated that the nuclear envelope with its nuclear pores is freely permeable to Ca2+ ions, and therefore nuclear Ca2+ signals are determined by cytosolic Ca2+ signals and diffusion of Ca2+ into the nucleus (e.g. Lipp et al. 1997). On the other hand, there is growing evidence that the nuclear envelope is a functional Ca2+ store, which is contiguous with the ER/SR Ca2+ store, and outer and inner membranes of the nuclear envelope express both types of Ca2+ release channels, RyRs and IP3Rs (for recent review see, e.g. Gerasimenko & Gerasimenko, 2004; Vermassen et al. 2004). Nuclear Ca2+ release via these receptor types has been demonstrated for a number of cell types. For example, in pancreatic acinar cells Ca2+ can be released from a thapsigargin-sensitive Ca2+ store in the nuclear envelope through RyRs by stimulation with nicotinic acid adenine dinucleotide phosphate (NAADP) and cyclic ADP ribose (Gerasimenko et al. 2006). Prolonged nuclear Ca2+ release events that originated from the nuclear envelope (or a closely associated structure) and were sensitive to RyR inhibition were described for adult rat ventricular myocytes (Yang & Steele, 2005).
IP3R-dependent Ca2+ release involving the nucleus has also been demonstrated for a number of cell types and tissues, but not for the adult heart. First evidence for a functional IP3-controlled nuclear Ca2+ store was obtained in liver (Nicotera et al. 1990; Divecha et al. 1993). Subsequently IP3R-dependent nuclear Ca2+ signals were also demonstrated, for example in pancreatic acinar cells (Gerasimenko et al. 2003), neurons (Marchenko et al. 2005), skeletal muscle (Cardenas et al. 2005), SKHep1 (Echevarria et al. 2003) and HepG2 (Leite et al. 2003) liver cell lines, Xenopus oocytes (Hennager et al. 1995; Stehno-Bittel et al. 1995a), and neonatal cardiomyocytes (Garcia et al. 2004; Luo et al. 2007). In contrast, direct evidence for IP3R-mediated release of Ca2+ from the nuclear envelope in adult cardiac myocytes has been lacking until now. In adult ventricular myocytes IP3Rs are preferentially located to the nuclear envelope (Bare et al. 2005). Wu et al. (2006) demonstrated Ca2+ depletion of the nuclear envelope in adult ventricular myocytes after stimulation with endothelin-1 in intact cells and upon application of adenophostin in permeabilized myocytes. However, in the study by Wu et al. depletion of the nuclear envelope was observed in situ, i.e. the nuclear envelope still forming a contiguous network with the SR. The study did not distinguish whether Ca2+ was released via IP3R located in the membrane of the nuclear envelope or through passive depletion following IP3R-dependent release of Ca2+ from the SR. Thus, the data presented in the study here are, to our best knowledge, the first direct demonstration of IP3R-dependent Ca2+ release from the nuclear envelope of adult cardiomyocytes.
An interesting question concerns the location of nuclear IP3Rs. The localization of IP3Rs in the outer membrane of the nuclear envelope, which is a continuation of the ER or SR) is well accepted (e.g. Gerasimenko et al. 1996; Gerasimenko & Gerasimenko, 2004). Much more intriguing is the growing evidence for a location of IP3R on the inner membrane, which would allow for a direct release of Ca2+ into the nucleoplasm (reviewed by Gerasimenko & Gerasimenko, 2004). Early evidence suggested the possibility of Ca2+ release from the nuclear envelope being directed into the nucleoplasm (Gerasimenko et al. 1995). IP3R-dependent Ca2+ release was even demonstrated from a reticular network expressing IP3Rs that extends within the nucleus in SKHep1 epithelial cells (Echevarria et al. 2003). Ca2+ release from the nucleoplasmic reticulum caused nuclear protein kinase C translocation. The study suggested that the nucleoplasmic reticulum can regulate nuclear Ca2+ in localized subnuclear regions, which would provide a potential mechanism by which Ca2+ could regulate various nuclear processes in parallel and independently. Our own study generated some indirect evidence that a significant portion of nuclear IP3Rs are located on the inner membrane and face the nucleoplasm. As shown in Fig. 5, upon stimulation of isolated nuclei with IP3, the increase of [Ca2+]Nuc was more than twice as large as the increase of [Ca2+]Cyt at the nuclear envelope–cytosol interface, and the rise of [Ca2+]Nuc outlasted the increase of [Ca2+]Cyt by more than a minute. Calibration of the fluo-4 signal in the nuclear and the surrounding extranuclear environment yielded virtually identical binding constants for the fluo-4–Ca complex (see Methods). Thus, the observed differences in the F/F0 signal inside and outside the nucleus (Fig. 5) reflect indeed differences in absolute [Ca2+] and were not the result of changes in Ca2+ binding affinity of fluo-4 in the nuclear environment, a property of fluorescent Ca2+ indicators that has been recognized as a potential problem for the accurate quantification of nuclear Ca2+ signals (see discussion below). The [Ca2+] gradient maintained between nucleoplasm and extranuclear space during exposure to IP3 cannot be explained solely by IP3R located exclusively in the outer membrane of the nuclear envelope and facing the cytoplasm. If the rise of [Ca2+]Nuc would occur through Ca2+ release from such receptors, followed by diffusion of Ca2+ through the nuclear pores, it would have to occur against a concentration gradient. On the other hand the relative magnitude of [Ca2+]Nuc and [Ca2+]Cyt does not allow any quantitative conclusions regarding the relative proportions of IP3Rs located on each side of the nuclear envelope. Dissipative diffusion of Ca2+ away from the IP3Rs facing the cytosol would tend to reduce [Ca2+]Cyt and underestimate the amount of Ca2+ released from these nuclear envelope IP3Rs, even though the use of the rather immobile Ca2+ indicator fluo-4 dextran reduces this problem. Despite these uncertainties, it is safe to say that IP3Rs on both inner and outer membranes of the nuclear envelope can release Ca2+ from this store.
In light of the evidence that the nucleus contains the key components of the phosphoinositide–PLC signalling cascade (Chi & Crabtree, 2000) and is capable of generating IP3 (e.g. Irvine, 2003), the possibility arises that the nucleus represents an autonomous compartment regarding the regulation and modulation of Ca2+-dependent nuclear processes such as the regulation of Ca2+-dependent transcription factors and gene regulation (Berridge et al. 2003; Dolmetsch, 2003).
A technical issue encountered when studying nuclear Ca2+ signalling relates to altered properties of fluorescent Ca2+ indicator dyes in the nuclear environment. Various fluorescent Ca2+ indicators have been shown to alter their properties in different subcellular environments (e.g. Thomas et al. 2000). For quantitative [Ca2+] studies the dependence of the apparent Ca2+ binding affinities in a specific subcellular compartment is of particular concern. Indeed, it has been suggested that reported [Ca2+]Cyt–[Ca2+]Nuc concentration gradients might be artifactual and the result of erroneous dye calibrations (for discussion see, e.g. Thomas et al. 2000; Gerasimenko & Gerasimenko, 2004). In our study we also found that Ca2+ indicator dyes belonging to the fluo family showed a higher basal fluorescence in the nucleus which was unlikely to be the result of a maintained higher [Ca2+]Nuc. More likely, altered dye properties and/or larger accumulation of the dye in the nucleus are responsible for this observation (Perez-Terzic et al. 1997; Thomas et al. 2000). We have shown previously using the ratiometric fluorescent indicator fura-2 that in resting smooth muscle and cardiac myocytes [Ca2+] is evenly distributed throughout the cell, including the nucleus, even though the raw fluorescence images showed remarkably higher fluorescence in the nucleus and nucleoli (Wier & Blatter, 1991; Blatter & Wier, 1992; see also Lipp & Niggli, 1993). Therefore, in our study all comparisons between [Ca2+]Cyt and [Ca2+]Nuc (e.g. Figs 1, 2 and 6) were done by only comparing the relative (%) changes in the respective compartment, i.e. by normalizing (F/F0) the fluorescence signal (F) to its own basal reference value (F0). This normalization would correct for differences in dye accumulation between cellular compartments. Furthermore, we established empirically that the Ca2+ binding affinity (Kd) of fluo-4 in the nucleus and the extranuclear environment differed by less than 10%.
In summary, we have shown that in cardiac myocytes the nuclear envelope is a functional Ca2+ store from which Ca2+ can be mobilized by IP3. This raises the question of the functional consequences of nuclear Ca2+ release for ECC and ECC-independent Ca2+ signalling in cardiac myocytes. IP3-dependent Ca2+ release from the SR has been shown to have positive inotropic, but also arrhythmogenic effects in atrial (Mackenzie et al. 2002; Zima & Blatter, 2004; Li et al. 2005) and ventricular (Proven et al. 2006) myocytes. It has been proposed that IP3-dependent Ca2+ release affects ECC through Ca2+-dependent sensitization of the RyR to Ca2+-induced Ca2+ release, and not through a significant direct contribution to the amount of Ca2+ required to initiate contraction. Since the nuclear envelope represents only a small fraction of the SR and the rise of [Ca2+]Cyt measured in the immediate vicinity of the nucleus upon stimulation with IP3 (Fig. 5A) is very small (maximum increase of F/F0 was ∼10%; cf. Fig. 5B), it is assumed that Ca2+ release through nuclear IP3Rs does not contribute to ECC. However, the finding that IP3 can trigger release of Ca2+ directly into the nucleoplasm may have important ramifications for ECC-independent Ca2+ signalling in cardiac myocytes. It is well established that a number of transcription factors (e.g. HDAC, NFAT) that become activated in cardiac hypertrophy and hypertrophy-related remodelling processes in the heart are shuttled between the nucleus and the cytoplasm in a Ca2+-dependent fashion (for reviews see Bueno et al. 2002; Crabtree & Olson, 2002; Wilkins & Molkentin, 2002, 2004). While the Ca2+ dependence of several key steps in this process is undebated, it has remained a conundrum how a cardiac cell can decode defined spatio-temporal Ca2+ signal(s) for hypertrophic signalling that are distinct from the ‘background calcium noise’ of ECC, i.e. the continuous complex and rapid (beat-to-beat) changes of global [Ca2+]i during ECC. The observation that the nucleus is surrounded by its own Ca2+ store, the nuclear envelope, raises the possibility that nuclear IP3-dependent Ca2+ release plays a crucial role for excitation–transcription processes, providing a mechanism of regulation that acts locally and autonomously from global cytosolic Ca2+ signals underlying ECC.
This work was supported by National Institutes of Health Grants HL80101 (to L.A.B. and G.A.M.), MH53367 (to G.A.M.), HL62231 (to L.A.B.) and American Heart Association Grants AHA0550170Z (to L.A.B.), and AHA0530309Z (to A.V.Z.).