Nuclear calcium signalling has been a controversial battlefield for many years and the question of how permeable the nuclear pore complexes (NPCs) are to Ca2+ has been the subject of a particularly hot dispute. Recent data from isolated nuclei suggest that the NPCs are open even after depletion of the Ca2+ store in the nuclear envelope. Other research has suggested that a new Ca2+-releasing messenger, nicotinic acid adenine dinucleotide phosphate (NAADP), can liberate Ca2+ only from acidic organelles, probably lysosomes, rather than from the traditional Ca2+ store in the endoplasmic reticulum (ER). Recent work indicates that NAADP can release Ca2+ from the nuclear envelope (NE), which has a thapsigargin-sensitive, ER-type Ca2+ store. NAADP acts in a manner similar to inositol (1,4,5)-trisphosphate [Ins(1,4,5)P3] or cyclic ADP-ribose (cADPR): all three messengers are equally able to reduce the Ca2+ concentration inside the NE and this is associated with a transient rise in the nucleoplasmic Ca2+ concentration. The NE contains ryanodine receptors (RyRs) and Ins(1,4,5)P3 receptors [Ins(1,4,5)P3Rs], and these can be activated separately and independently: the RyRs by either NAADP or cADPR, and the Ins(1,4,5)P3Rs by Ins(1,4,5)P3.
A rise in the cytosolic Ca2+ concentration is a highly versatile signal that can regulate many cellular processes (Berridge et al., 2003), but Ca2+ signals can also arise in many other subcellular compartments including the nucleoplasm (Gerasimenko et al., 1995). Indeed, half of the articles written in the Ca2+ signalling field during the past decade are actually about nuclear calcium signalling. Here, we deal with three currently controversial areas.
The first concerns the permeability of the nuclear pore complexes (NPCs), which span the inner and outer membranes of the NE (Bootman et al., 2000). The diameter of the nuclear pore should allow free diffusion of small ions including Ca2+ (Perez-Terzic et al., 1997; Lee et al., 1998). However, there are many reports about the existence of nuclear-cytosolic Ca2+ gradients. Most are based on confocal microscopy. Although discussion continues about this topic, an overwhelming body of evidence supports the view that the NPCs are highly Ca2+ permeable (Lipp and Niggli, 1994), and most likely the reported cytoplasmic-nuclear Ca2+ gradients are due to differences in the behaviour of the fluorescent dyes in different subcellular compartments (Al-Mohanna et al., 1994; O'Malley et al., 1999).
The second controversial area concerns the nature of the nuclear Ca2+ release. So far, there is no consensus on the location of the Ins(1,4,5)P3 receptors [(Ins(1,4,5)P3Rs] and ryanodine receptor (RyR) Ca2+ channels that release Ca2+ in the nucleus. The nuclear Ca2+ store can release Ca2+ into the nucleoplasm as well as outside (Gerasimenko et al., 1995). Indirect functional evidence indicates that Ins(1,4,5)P3Rs and RyRs reside on the inner surface of the NE (Gerasimenko et al., 1995; Gerasimenko et al., 1996a; Petersen et al., 1998). This hypothesis is supported by several studies (Humbert et al., 1996; Santella and Carafoli, 1997; Adebanjo et al., 1999; Adebanjo et al., 2000), but the role of intranuclear receptors in normal intracellular responses has been disputed (Lipp et al., 1997; Bootman et al., 2000).
The third area of dispute is the role of NAADP. Genazzani et al. (Genazzani and Galione, 1996) have suggested that NAADP releases Ca2+ from stores either in lysosome-like organelles or in related granules that are separate from the ER. This would appear to exclude the possibility of NAADP-induced Ca2+ signalling in the nucleus. Nevertheless, recent data (Gerasimenko et al., 2003) indicate that NAADP can indeed release Ca2+ from the NE and thereby generate nucleoplasmic Ca2+ signals.
Nuclear-cytoplasmic Ca2+ gradients?
Nearly half of the published nuclear Ca2+ measurements in many cell types identify a Ca2+ gradient between nuclei and cytoplasm, using different fluorescent dyes. The other half do not, often using the same cell types, similar dyes and apparently similar techniques (Table 1). Mostly, the nuclear-cytosolic Ca2+ gradients described are transient. Only very few studies report persistent gradients.
An example of a transient nuclear-cytoplasmic gradient is shown in Fig. 1. Localized Ca2+ spikes in the secretory region of acinar cells do not propagate into the nucleus or the basal area of the cell (Gerasimenko et al., 1996b). There is a reason for the transiently different Ca2+ concentrations in the nucleus and the cytosol. Perigranular and perinuclear mitochondria are able to restrict local Ca2+ spikes elicited by small doses of agonists to the secretory granule area (Johnson et al., 2003; Tinel et al., 1999). Therefore, the transient difference in Ca2+ concentration is quite understandable, since nuclei in many cell types are normally surrounded by mitochondria sequestering Ca2+ during physiological responses (Duchen, 1999; Robb-Gaspers et al., 1998). Strong stimulation can induce large Ca2+ waves that can overcome this mitochondrial barrier (Fig. 1C) and therefore quickly penetrate the nucleus. In this case, the difference between the Ca2+ concentration in the cytosol and the nucleus is marginal (Al-Mohanna et al., 1994; Lipp and Niggli, 1994).
Accurate measurements of the nuclear Ca2+ concentration are extremely difficult and require separate calibration of the nuclear dye owing to many possible artefacts, such as increased brightness of most of the fluorescent indicators in the nucleoplasm (Al-Mohanna et al., 1994; Bootman et al., 2000; Perez-Terzic et al., 1997). In addition to exhibiting differences in dissociation constant and dynamic range, the indicators can also be bound inside the nucleoplasm and sequestered in Ca2+ stores (Al-Mohanna et al., 1994; Burnier et al., 1994; Ikeda et al., 1996; Ronde and Nichols, 1997). These problems effectively invalidate the majority of the observations of nuclear-cytoplasmic Ca2+ gradients. There is thus probably no persistent nuclear-cytoplasmic gradient (Badminton et al., 1996; Badminton et al., 1998; Brown et al., 1997; Bootman et al., 2000; Bootman et al., 2002; Tovey et al., 1998). Indeed, this is convincingly shown by luminescence measurements of targeted aequorin (Brini et al., 1993).
The Ca2+ permeability of NPCs
If sustained nuclear-cytoplasmic Ca2+ gradients do not exist, then NPCs must be easily permeable to small ions, including Ca2+. Unfortunately, there are contradictory data regarding the Ca2+ permeability of NPCs. Electrophysiological recordings in isolated nuclei (Assandri et al., 1997) (reviewed by Santella, 1996) have shown that the ionic conductance of NPCs can change dramatically and that only a small proportion of NPCs are open under patch-clamp conditions. Danker and collaborators (Danker et al., 1999; Danker et al., 2001) have shown that most NPCs are in an open state and are freely permeable to ions. Atomic force microscopy experiments suggest that NPCs change their conformation when the central plug blocks the pore (Stoffler et al., 1999), but there is no proof that the plugs actually prevent Ca2+ diffusion through NPCs. Even if they do, the available evidence indicates that there are substantial numbers of NPCs without plugs at any time. We have shown (Gerasimenko et al., 1995) that Ca2+ concentration changes outside the isolated nucleus are quickly reproduced by similar Ca2+ changes in the nucleoplasm, both in isolated liver nuclei (Fig. 2) and in isolated pancreatic nuclei (Fig. 3), (Gerasimenko et al., 2003). Generally speaking, NPCs must thus be open to passive diffusion of small ions in these systems. There might of course be special conditions under which NPCs are closed, and we cannot conclude that all NPCs are open all the time for passive diffusion. But, even if half of the NPCs were closed, the NE would still be very permeable to small ions and molecules. This could resolve the apparent discrepancy between single NPC observations (reviewed by Santella, 1996) and nuclear electric conductance measurements of large surfaces of nuclei using the nuclear hourglass technique, which indicated that the NE conductance is always high (Danker et al., 1999).
The NPC is a supramolecular assembly that has a molecular weight of ∼125,000 kDa (Greber and Gerace, 1995) and a large aqueous channel with a diameter of ∼9 nm. Normally, molecules <40 kDa pass through such a channel without the need for a nuclear localization signal. Some studies suggest that free diffusion of such molecules could be inhibited by the depletion of the nuclear Ca2+ store (Greber and Gerace, 1995; Lee et al., 1998; Lyman and Gerace, 2001; Perez-Terzic et al., 1999). The mechanism might involve binding of intra-ER Ca2+ to the NPC protein gp210 (Greber and Gerace, 1995). However, we and others have not observed this (Strubing and Clapham, 1999; Stehno-Bittel et al., 1995; Perez-Terzic et al., 1996; Wei et al., 2003; Gerasimenko et al., 2003). Very convincing data in favour of free diffusion through NPCs have been obtained recently using fluorescence recovery after photobleaching (FRAP) of 27 kDa enhanced green fluorescence protein (EGFP) in intact cells (Wei et al., 2003). These recent studies employed a transfection technique that does not affect nuclear transport; by contrast, microinjection, which was used previously, can disrupt the cytoskeleton (Swaminathan et al., 1997).
Our own results indicate that, after depletion of the NE Ca2+ stores with Ins(1,4,5)P3 or NAADP, subsequent external application of a high Ca2+ concentration followed by the Ca2+ chelator EGTA induces a large rise and thereafter a fast decrease in the nucleoplasmic Ca2+ concentration (Gerasimenko et al., 2003). These measurements demonstrate rapid movement of Ca2+ across the NE, most likely through the NPCs. On the basis of all these data, we favour the argument that the NPCs are permeable to Ca2+ even after depletion of the NE Ca2+ stores.
Ins(1,4,5)P3Rs in the nuclear membranes
The presence of Ins(1,4,5)P3Rs in the NE is well documented (reviewed by Vermassen et al., 2004), and their localization in the outer membrane of the NE, which is a continuation of the ER membrane, is beyond dispute (Malviya et al., 1990; Nicotera et al., 1990; Gerasimenko et al., 1996a; Humbert et al., 1996; Meldolesi and Pozzan, 1998). More intriguing evidence has emerged about the localization of Ins(1,4,5)P3Rs on the inner NE membrane, which would enable release of Ca2+ into the nucleoplasm (Gerasimenko et al., 1995; Hennager et al., 1995; Humbert et al., 1996; Santella and Kyozuka, 1997b), since nuclei can produce inositol polyphosphates (Cocco et al., 1994; Divecha et al., 1991). The inner nuclear membrane appears to lack the sarco-endoplasmic reticulum Ca2+-activated ATPase (SERCA)-type Ca2+-activated ATPases present on the outer membrane (Humbert et al., 1996; Gensburger et al., 2003). These authors also suggested a Ca2+ uptake mechanism, which is active at elevated Ca2+ concentrations induced by inositol (1,3,4,5)-tetrakisphosphate [Ins(1,3,4,5)P4] through Ins(1,3,4,5)P4Rs located exclusively on the outer nuclear membrane. The Ca2+ concentration in the nucleoplasm nevertheless returns to normal following messenger-mediated release, probably by diffusion of Ca2+ through NPCs and subsequent action of Ca2+-ATPases/Ins(1,3,4,5)P4Rs in the outer nuclear membrane. Such an intranuclear Ca2+ signalling system might be able to work locally and independently, although at the moment there is no convincing demonstration of its physiological importance.
Interesting data in favour of such a system (Clubb et al., 1998; Fricker et al., 1997) show that the NE has invaginations into the nucleoplasm in many cell types and could explain Ca2+ release sites inside the nucleoplasm. Recent data by Echevarria et al. confirm these suggestions and show that local Ca2+ release from the nucleoplasmic reticulum occurs through Ins(1,4,5)P3Rs and induces translocation of protein kinase C (PKC) to the NE membrane (Echevarria et al., 2003). These new data argue for an independent intranuclear Ca2+ signalling system.
Unfortunately, because accurate measurement of the Ca2+ concentration in the nucleoplasm of intact cells is difficult (as discussed above), there are no reliable Ca2+ measurements that unequivocally show independent activation of an intranuclear Ca2+ signalling system with a physiological agonist.
RyRs in nuclear membranes
The intracellular Ca2+-releasing messenger cADPR (Lee et al., 1989) has been shown to be an endogenous activator of RyR Ca2+ channels (Meszaros et al., 1993) and plays a role similar to Ins(1,4,5)P3 in Ca2+ signalling (Ehrlich et al., 1994; Fitzsimons et al., 2000; Solovyova et al., 2002; Wakui et al., 1990). The demonstration that the NE releases Ca2+ in response to cADPR and ryanodine indicates that RyRs are present on the inner nuclear membrane (Gerasimenko et al., 1995). Indeed, the presence of RyRs has been confirmed both by RyR-specific antibodies (Adebanjo et al., 1999; Adebanjo et al., 2000; Santella and Kyozuka, 1997) and by specific staining with fluorescent BODIPY FL ryanodine (Fig. 3D) (Gerasimenko et al., 2003). Release of Ca2+ through cADPR induces a reduction of the free Ca2+ concentration in the NE Ca2+ store, similar to that evoked by Ins(1,4,5)P3. These data indicate that the cADPR-induced Ca2+ release in the nucleus can also be directed into the nucleoplasm (similar to what has been shown for Ins(1,4,5)P3-induced Ca2+ release). However, we cannot exclude the possibility that Ca2+ released outside the outer membrane of the NE can diffuse into the nucleoplasm through open NPCs (Gerasimenko et al., 1995; Lipp et al., 1997; Gerasimenko et al., 2003).
NAADP-induced Ca2+ release
NAADP is a recently identified Ca2+-releasing messenger, capable of inducing Ca2+ signals in different cell types (Cancela et al., 1999; Cancela, 2001; Cancela et al., 2003; Chini et al., 1995; Chini and Dousa, 1996; Chini and de Toledo, 2002; Genazzani and Galione, 1996; Genazzani and Galione, 1997; Lee and Aarhus, 1995; Johnson and Misler, 2002; Mitchell et al., 2003; Masgrau et al., 2003; Petersen and Cancela, 1999). NAADP can enhance neurosecretion in response to Ins(1,4,5)P3, cADPR and sphingosine 1-phosphate at the frog neuromuscular junction (Brailoiu et al., 2003), and ER-type Ca2+ stores might be coupled to separate NAADP-sensitive Ca2+ stores (Brailoiu et al., 2003). Churchill and co-authors (Churchill et al., 2002; Churchill et al., 2003; Yamasaki et al., 2004) have suggested that the NAADP-sensitive Ca2+ stores are reserve granules, the functional equivalent of lysosomes in sea urchin eggs (Armant et al., 1986; Jadot et al., 1984). However, the mechanistic details of NAADP-induced Ca2+ release are still unclear, since the NAADP receptor (NAADPR) remains to be identified and NAADP might even act through RyRs (Hoheneger et al., 2002; Mojzisova et al., 2001).
We found recently that NAADP can release Ca2+ into the nucleoplasm in isolated pancreatic nuclei, a preparation that only contains an ER-type Ca2+ store and does not contain lysosomes or other acidic organelles (Gerasimenko et al., 2003). The NAADP-sensitive nuclear store is ATP dependent and thapsigargin sensitive, and is distinct from any acidic, endocytic or Golgi-type Ca2+ store (Gerasimenko et al., 1996c; Gerasimenko et al., 1998; Pinton et al., 1998). Remarkably, the Ca2+ release normally induced by NAADP [but not by Ins(1,4,5)P3] is completely blocked by inhibitors of RyRs (e.g. high ryanodine concentration or Ruthenium Red). These findings indicate that, in pancreatic nuclei, NAADP releases Ca2+ from the same thapsigargin-sensitive store as Ins(1,4,5)P3 and cADPR. These findings also cast doubt on the idea that NAADP releases Ca2+ only from a distinct pool.
Given the evidence for intranuclear Ins(1,4,5)P3 (Cocco et al., 1994; Divecha et al., 1991) and cADPR production systems (Adebanjo et al., 1999), the existence of a completely independent intranuclear Ca2+ signalling system is proposed, which would not require any cytoplasmic release although subsequent cytoplasmic Ca2+ uptake would be needed (Bootman et al., 2000). In spite of the evidence for the existence of such a complete Ca2+ release system in the nucleus, there is little evidence (Lui et al., 1998; Santella and Kyozuka, 1997; Santella et al., 1998) of its independent involvement in nuclear Ca2+ signal generation. In fact, there are more reports indicating that Ca2+ released primarily into the cytoplasm (Ca2+ puffs) can diffuse into the nucleoplasm (Bootman et al., 1997; Bootman et al., 2000; Hennager et al., 1995; Lipp et al., 1997; Shirakawa and Miyazaki, 1996). The existence of a complete, independent, intranuclear Ca2+ signalling system is potentially very interesting, but its physiological significance is currently questionable and requires convincing evidence of activation by physiological agonists. Future work in this area will undoubtedly bring more clarity and understanding of mechanisms involved.
In the intact pancreatic acinar cells, the Ca2+ store in the NE is part of one lumenally continuous ER store (Petersen et al., 2001); however, it is highly specialized in different parts of the cell (Gerasimenko et al., 2002). The data on isolated nuclei (Gerasimenko et al., 2003), as compared with the studies on local Ca2+ spiking in the secretory pole in intact cells (Cancela et al., 2000; Cancela et al., 2002; Osipchuk et al., 1990; Park et al., 2001; Petersen et al., 1991; Petersen et al., 1994), reveal that local control of Ca2+ release can operate differently in different parts of the cell. All three messengers can induce Ca2+ release at both sites. The local Ca2+ spiking in the secretory region of pancreatic acinar cells [as well as the global Ca2+-induced Ca2+ waves (Ashby and Tepikin, 2002; Ashby et al., 2003; Thorn et al., 1993; Thorn et al., 1994)] are highly dependent on cooperative interaction of Ins(1,4,5)P3Rs and RyRs (Ashby et al., 2002; Boittin et al., 1998; Cancela et al., 2000; Kidd et al., 1999; Koizumi et al., 1999; Lipp et al., 2000). By contrast, there is little cooperativity of Ins(1,4,5)P3Rs and RyRs in the nucleus; in fact, the receptors can function independently, inducing similar release of Ca2+ into the nucleoplasm (Gerasimenko et al., 2003).
The difference between the NAADP effects in the nucleus and the secretory granule area are even more striking: whereas there is a mutually potentiating interaction of Ins(1,4,5)P3, cADPR and NAADP in the secretory granule area (Cancela et al., 2002), there is no such potentiation in the nucleus (Gerasimenko et al., 2003). These differences could be explained in two different ways. Three separate channels (receptors) could interact. Such a model would depend on the presence of so-far-uncharacterized NAADPR Ca2+ channels. A much simpler hypothesis (Fig. 4) is that there are only two types of Ca2+ release channels – Ins(1,4,5)P3Rs and RyRs – in a single store. Ca2+ can enter the nucleoplasm through Ins(1,4,5)P3Rs or RyRs in the inner nuclear membrane, although we cannot exclude Ca2+ release from the Ins(1,4,5)P3Rs or RyRs in the outer nuclear membrane and subsequent entry through NPCs. All three Ca2+-releasing messengers, Ins(1,4,5)P3, cADPR and NAADP, can be produced inside the NE, although they can also enter the nucleoplasm through the NPCs. Ca2+ is pumped into the NE by SERCA on the outer nuclear membrane. Potentiating or non-potentiating interactions of Ins(1,4,5)P3Rs and RyRs (see discussion above) in this simpler model would depend on the degree of closeness of the two types of Ca2+ release channel and of the cADPR and NAADP receptors. This exciting area of research will definitely attract more attention in the future.
We thank O. H. Petersen for critical reading and helpful suggestions. We thank R. Sever for many valuable discussions and suggestions, and the reviewers for helpful comments on the manuscript. This work is supported by an MRC Programme grant (G8801575).
- © The Company of Biologists Limited 2004