Functional visualization of the separate but interacting calcium stores sensitive to NAADP and cyclic ADP-ribose

Cyclic ADP-ribose (cADPR) and nicotinic acid adenine dinucleotide phosphate (NAADP) are Ca²⁺ releasing metabolites of NAD and NADP, respectively (Clapper et al., 1987; Lee and Aarhus, 1995; Lee et al., 1989). The cADPR-sensitive Ca²⁺ stores were first described in sea urchin eggs (Clapper et al., 1987; Lee and Aarhus, 1995; Lee et al., 1989) and have since been shown to be widely present in cells from protist and plant to human (reviewed in Lee, 1997; Lee, 2000). Both the pharmacology (Galione et al., 1991; Lee, 1993) and the characteristics of the cADPR-sensitive Ca²⁺ release channels reconstituted in lipid bilayers are similar to those of the ryanodine receptor (Lokuta et al., 1998; Perez et al., 1998; Sitsapesan and Williams, 1995; Sonnleitner et al., 1998). The cADPR-sensitive Ca²⁺ stores are thus likely to be the endoplasmic reticulum, and this is also supported by subcellular fractionation studies (Dargie et al., 1990; Lee, 1991; Lee and Aarhus, 1995). Nevertheless, the cADPR-sensitive stores have never been directly visualized.

The NAADP-sensitive Ca²⁺ stores have recently been characterized in sea urchin eggs (Lee and Aarhus, 1995). They have since been found in two other invertebrate eggs, ascidian and starfish (Albrieux et al., 1998; Santella et al., 2000), as well as in plant (Navazio et al., 2000) and mammalian cells, pancreatic acinar cells and brain microsomes (Bak et al., 1999; Cancela et al., 1999), suggesting that these stores are also generally present in cells. A unique property of the NAADP-sensitive Ca²⁺ release mechanism is its insensitivity to antagonists of all the intracellular Ca²⁺ release channels known, which include the inositol trisphosphate-receptor, the ryanodine receptor and the cADPR-sensitive Ca²⁺ channels (Chini et al., 1995; Graeff et al., 1995; Lee and Aarhus, 1995). Fractionation studies indicate that the NAADP-sensitive Ca²⁺ stores are also distinct from the mitochondria (Lee and Aarhus, 1995). In this study, we describe the direct visualization of the cADPR- and NAADP-sensitive Ca²⁺ stores in sea urchin eggs using Ca²⁺ imaging. The results show that the stores are separate but appear to be interacting during fertilization.

Lytechinus pictus eggs were stratified according to a previous procedure (Brandriff and Hinegardner, 1975; Craig and Piatigorski, 1971; Lee and Epel, 1983). The exact conditions for stratification were determined for each batch of eggs. Centrifuging for too long or at too high a speed split the eggs, resulting in loss of cell fragments containing the relevant Ca²⁺ stores. Insufficient centrifugation, on the other hand, resulted in inadequate segregation of organelles. Generally, the egg suspension (0.4 ml) was layered on top of 1 ml of 1M sucrose and centrifuged in a Beckman microfuge for 8-15 minutes at 11,000-13,000 rpm (7,000-8,000 g).

Freshly stratified eggs were used in each experiment within 15-20 minutes of stratification to avoid complications due to redistribution of organelles in the eggs. Stratified eggs were washed once with artificial sea water, attached to coverslips coated with protamine sulfate and loaded by microinjection. The injection medium contained either 350 μM caged cADPR, 18-34 μM caged NAADP or 90 μM caged IP₃, and 10 mM fluo-3, 0.5 M KCl, 0.1 mM EGTA, 20 mM Hepes, pH 6.7. The injection volume was 1-2% of the egg volume. The site of injection is usually at the center of the egg, which was elongated by stratification. The injected fluid droplet dispersed within seconds as visualized in brightfield. The diffusion of the injected substances throughout the eggs was uniform and rapid as verified by observing the distribution of fluo-3 that was included in the injection medium. Fluo-3 has a size and charge similar to the caged compounds and can serve as a good tracer. That the injected caged compounds themselves indeed diffuse uniformly throughout the eggs is shown in unstratified eggs. A uniform pattern of Ca²⁺ elevation was seen in these eggs after whole-field illumination with UV-light, photolyzing the injected caged compounds (Aarhus et al., 1996; Aarhus et al., 1995; Lee et al., 1997).

Caged cADPR and caged NAADP were synthesized as described previously (Aarhus et al., 1995; Lee et al., 1997). Caged IP₃ was purchased from CalBiochem (La Jolla, CA, USA) and further purified by anion exchange chromatography (Cellufine A800, Millpore, Bedford, MA, USA) to remove contaminating IP₃. A gradient of water and 1 M triethylamine bicarbonate (pH 8.8) was used for elution as previously described (Lee et al., 1997). Bodipy-thapsigargin and rhodamine 123, specific markers for the ER (Simpson and Russell, 1997) and mitochondria (Johnson et al., 1980), respectively, were from Molecular Probes (Eugene, OR, USA).

Calcium imaging and whole-field photolysis were obtained with an InCa⁺⁺ system (Intracellular Imaging Inc., Cincinnati, OH, USA) equipped with a 300-watt Xenon arc lamp. The light output was directed to the eggs via the epifluorescence attachment of a Nikon inverted microscope equipped with a dichroic mirror that reflected both 365 nm and 485 nm light. The exposure time was 0.5-2.5 seconds, which did not produce any Ca²⁺ changes in control eggs not loaded with the caged compounds. Fluorescence data were not collected during the photolysis exposure. Local photolysis used a nitrogen laser (VSL-377ND-S) obtained from LSI (Franklin, MA, USA). The laser provided 4 nanosecond pulses at 337.1 nm with a pulse energy of 120 μJ. Peak power was 30 kW and the average power was 2.4 mW at 20 Hz. Ten laser pulses were directed to a portion of the egg via a fused silica fibre (0.2 mm diameter) obtained from Ocean Optics (Dunedin, FL, USA). The fiber was heat-pulled to produce a tip of 10-20 μm?????

Eggs were gently homogenized in a glass homogenizer fitted with a Teflon pestle. The homogenates were centrifuged for about 30 seconds in a microfuge to remove egg cortices containing the cortical granules as well as intact eggs (Clapper and Lee, 1985; Clapper et al., 1987; Lee, 1991). Egg microsomes were isolated by Percoll density centrifugation. 2 ml of 15% L. pictus egg homogenates were placed on top of 10 ml of 25% Percoll and centrifuged for 40 minutes at 25,000 rpm (41,000 g) in a Beckman Ti50 rotor (10°C) as described previously (Lee and Aarhus, 1995). The supernatants were discarded and eight fractions were collected from each gradient by puncturing the bottom of the centrifuge tubes. ATPase activity was determined by diluting the microsomes twentyfold into an assay medium containing 250 mM N-methylglucamine, 250 mM potassium gluconate, 0.5 mM MgCl₂, 0.1 mM [γ-³²P]dATP, 20 mM Hepes, pH 7.2, adjusted with acetic acid. After a 3 hour incubation at 20-25°C, the reaction was stopped with 0.1 M phosphoric acid containing 25 g/l activated charcoal. After centrifugation, the released ³²P_i in the supernatant was counted.

It is generally accepted that the stimulus-sensitive Ca²⁺ stores in cells are in the ER. Indeed, our fractionation results in the sea urchin egg show that both the cADPR- and IP₃-sensitive Ca²⁺ stores copurified in Percoll gradients with ER markers, such as glucose 6-phosphatase and Ca²⁺-ATPase (Clapper et al., 1987; Lee, 1991; Lee and Aarhus, 1995). In contrast, the NAADP-sensitive Ca²⁺ stores have previously been proposed to be new and distinct from the ER (Lee and Aarhus, 1995). The main evidence is that their distribution in Percoll gradients is different from the other two Ca²⁺ stores and is also broader than the ER (Lee and Aarhus, 1995). It is possible, however, that fragmentation of Ca²⁺ stores could occur during cell homogenization, contributing to the apparently distinct fractionation patterns. To resolve this issue, it is necessary to demonstrate that the NAADP-sensitive stores can be distinguished, in intact cells, from the ER and the other two Ca²⁺ stores.

The approach we used to demonstrate that was to take advantage of the fact that sea urchin eggs are large cells and that their organelles can be stratified into well-defined layers by centrifugation (Brandriff and Hinegardner, 1975; Craig and Piatigorski, 1971; Lee and Epel, 1983). Fig. 1 shows that Lytechinus pictus eggs were stretched to elongated shapes following centrifugation on top of a sucrose cushion. The nucleus (indicated by arrowheads) together with the oil cap (dark patch, Fig. 1B) are moved to one pole (labeled N). These stratified eggs were injected with fluo-3 together with caged NAADP (Lee et al., 1997) (Fig. 1A), caged cADPR (Aarhus et al., 1995) (Fig. 1B) or caged IP₃ (Fig. 1C) and exposed to whole-field illumination of UV-light to photolyze the intracellular caged compounds. Although the photolyzing illumination was uniform, the resulting Ca²⁺ release was highly localized. The NAADP-induced release was mainly in the pole distal to the nucleus, while cADPR-induced release was localized to the nuclear pole (N). These localized patterns of Ca²⁺ mobilization are in striking contrast to the uniform elevation of intracellular Ca²⁺ following whole-field photolysis we reported previously in normal eggs not stratified by centrifugation (Aarhus et al., 1996; Aarhus et al., 1995; Lee et al., 1997). Uniform photolysis of caged IP₃ (Fig. 1C) produced a pattern of Ca²⁺ mobilization similar to that of caged cADPR and was distinct from that of the caged NAADP.

Similar localized patterns of release were seen in eight eggs injected with caged cADPR, nine eggs injected with caged NAADP and eleven eggs injected with caged IP₃. Fig. 2 shows the averaged fluo-3 fluorescence profiles of the eggs injected with the three caged compounds. The maximal fluorescence in the eggs following photolysis of caged cADPR was seen at a distance 29±5% (± s.e.m., n=8) of the cell length starting at the nuclear pole (Fig. 2A). On the other hand, the maximal Ca²⁺ release in the caged NAADP-injected eggs determined in the same manner occurred at 74±9% (± s.e.m., n=9) of the cell length (Fig. 2A). The nuclei of these eggs were located at 16±2% (± s.e.m., n=28) of the cell length. The profile of Ca²⁺ mobilization induced by caged IP₃ peaked at 36±4% of the cell length (± s.e.m., n=11) (Fig. 2B), very similar to that induced by caged cADPR. Since it is generally accepted that the diffusion of Ca²⁺ in cells is limited, the localized release patterns observed suggest that the three Ca²⁺ stores were segregated to opposite poles of the stratified eggs.

This remarkable segregation of the three Ca²⁺ stores was further demonstrated using local photolysis. The output of a nitrogen laser was delivered to a stratified egg via a fused silica fiber with a tip diameter of about 10 μm (Parpura and Haydon, 1999) (Fig. 3). The egg was previously injected with caged NAADP and fluo-3. Local photolysis at the nuclear pole (labeled N in Fig. 3A) produced only a small and localized Ca²⁺ release that dissipated quickly. The fiber was then placed at the distal pole of the same egg (Fig. 3B) and the same numbers of photolyzing laser pulses were delivered. The resulting Ca²⁺ release spread throughout the distal half of the egg. The magnitude of the release was higher and the duration was also longer. Eight of the eleven eggs measured in this manner responded similarly. Fig. 3C shows that the averaged fluo-3 fluorescence increase (F/Fo) induced by local photolysis at the distal poles of these eggs was 6.2±1.1 (± s.e.m.) -fold as compared with an increase of only 2.5±0.6 (± s.e.m.) -fold when the photolysis was directed at the nuclear poles of the same eggs. The time courses of the fluorescence increase at the two poles (Fig. 3C) indicate that the duration of the Ca²⁺ release (time from the start of photolysis to the return of the fluo-3 fluorescence to half of the maximum) at the distal pole was also about 2.2-fold longer.

The small and brief Ca²⁺ release in the nuclear end (blue squares, Fig. 3C) suggests that some NAADP-sensitive stores remained there. This is consistent with the results shown in Fig. 2A. After whole-field illumination, the averaged Ca²⁺-release profile of the eggs injected with caged NAADP showed fluorescence in the nuclear pole that was about 2-3 times lower than the peak in the distal end. Fractionation rarely separates organelles completely, even when done in vitro by density centrifugation. It is thus reasonable that some NAADP-sensitive stores remained in the nuclear pole even after stratification.

The two types of Ca²⁺ stores have different sensitivity to thapsigargin, which effectively discharges the cADPR-sensitive stores by inhibiting the Ca²⁺-ATPase, while the NAADP-sensitive stores are resistant (Genazzani and Galione, 1996). Bodipy-thapsigargin, a fluorescent analog, was used to visualize the stores containing the Ca²⁺-ATPase (Simpson and Russell, 1997). Eggs were incubated with the fluorescent probes and subsequently stratified. Fig. 4A shows that the eggs exhibited prominent staining, mainly in regions close to nuclear pole (N), similar to the distribution of the cADPR- and IP₃-sensitive Ca²⁺ stores. The absence of fluorescence around the distal pole is consistent with the NAADP-stores possessing a thapsigargin-insensitive Ca²⁺-ATPase. That the eggs do have this type of Ca²⁺-ATPase is shown directly by measuring ATPase activity in the egg microsomes. The presence of 100 μM Ca²⁺ stimulated the activity about twofold (Fig. 4B). A portion of 80-90% of the Ca²⁺-ATPase was completely blocked by 10 nM of thapsigargin while the remaining 10-20% of the Ca²⁺-ATPase was not sensitive to as high as 10 μM thapsigargin (Fig. 4B). Fig. 4C shows the distribution of the thapsigargin-insensitive Ca²⁺-ATPase in egg homogenates fractionated using Percoll density centrifugation. 40-60% of the Ca²⁺-ATPase activity measured in the denser fractions (1 and 2) was insensitive to thapsigargin-inhibition. We have previously shown that the NAADP-sensitive Ca²⁺ stores, unlike the cADPR-sensitive stores, have a broad distribution in the Percoll gradient (Lee and Aarhus, 1995). In fractions 3 and 4, which contained the cADPR-sensitive stores (Lee and Aarhus, 1995), the majority of the Ca²⁺-ATPase was inhibited by thapsigargin and only about 2-5% was insensitive.

Fig. 5 summarizes the distributions of various organelles in the stratified eggs. The nucleus, the ER containing the thapsigargin-sensitive Ca²⁺-ATPase, the cADPR-sensitive and the IP₃-sensitive Ca²⁺ stores were all concentrated in about a third of the egg (36±4%) closest to the nuclear pole. In contrast, the NAADP-sensitive Ca²⁺ stores were mainly localized in the distal half (74±3% of the cell length distal to the nuclear pole). Included in Fig. 5 is another organelle marker, the mitochondria, which were visualized by Rhodamine-123 staining. The inset shows that they were also concentrated in the distal part of the stratified eggs.

Fig. 6A shows that fertilization of these stratified eggs produced a striking Ca²⁺ wave, starting at the nuclear pole (N, nucleus; arrow) (Eisen and Reynolds, 1985) and propagating throughout the entire egg. The Ca²⁺ concentration in the distal end that contained the NAADP-sensitive Ca²⁺ stores elevated also, suggesting that the NAADP-stores were activated during fertilization. Even though sperm were added uniformly as a suspension to the eggs, in all seven eggs observed, the Ca²⁺ waves were initiated near the nuclear pole, close to the cADPR-sensitive stores. This is further documented in Fig. 6C by averaging the fluo-3 fluorescence profiles along the long axis of the eggs shortly after fertilization. On average, the initial Ca²⁺ release was most prominent in regions at distances about 10% of the cell length from the nuclear pole. 1-2 minutes afterwards, the release pattern changed to be essentially uniform throughout the cells. These results suggest that a signal is sent from the initiation site in the nuclear half of the egg, which contains the majority of the cADPR-sensitive Ca²⁺ stores, and propagates to the distal half, resulting in the activation of the NAADP-sensitive Ca²⁺ stores.

The remarkable selectivity of the initiation of the Ca²⁺ wave in the nuclear half suggests that the fertilizing sperm may have a preference for the nuclear half. The fertilizing sperm can be visualized by the nuclear stain, Hoechst 33342. The nucleus of the fertilizing sperm, once incorporated into the egg, accumulated the dye preincorporated into the eggs and appeared as a bright fluorescent spot (Lee et al., 1993; Twigg et al., 1988). In all seven monospermic eggs observed in this manner, the fertilizing sperm nuclei were located in the nuclear half. An example is shown in Fig. 6B. In three polyspermic eggs, one showed three sperm nuclei in the nuclear half and two showed one sperm nucleus in each half of the egg (data not shown). It thus appears that under the normal monospermic condition, the fertilizing sperm do have a remarkable preference for the nuclear half, where the cADPR-sensitive Ca²⁺ stores are present.

Direct and visual evidence is provided in this study to show that the cADPR- and the NAADP-sensitive Ca²⁺ stores are distinct and separable, firmly establishing the reality of these two novel Ca²⁺ stores. That centrifugation can segregate these stores indicates they are free cytoplasmic organelles not attached to the cortex of the egg. The cADPR-stores are likely to be the ER since they segregate near the nuclear pole similar to the thapsigargin-sensitive Ca²⁺-ATPase, a marker for the ER. This is consistent with subcellular fractionation studies showing that the cADPR-stores copurify with the Ca²⁺ sequestration activity and a marker of the ER (Lee, 1991; Lee and Aarhus, 1995).

It is generally believed that the nuclear envelope is closely associated, or even continuous, with the ER. Indeed, recent work shows that not only the cADPR-sensitive Ca²⁺ release mechanism but also CD38, a homolog of ADP-ribosyl cyclase, are both present in the nucleus (Adebanjo et al., 1999; Gerasimenko et al., 1995; Khoo et al., 2000; Santella, 1996). That the nucleus is found to colocalize with the cADPR-sensitive stores in the stratified eggs further indicates the close association of the two organelles. This strategic localization may allow cADPR to play a direct and important role in regulating gene expression, which has, in fact, been demonstrated in plants. Thus, cADPR is shown to be the Ca²⁺ messenger mediating gene expression induced by the plant hormone, abscisic acid (Wu et al., 1997), as well as that induced by nitric oxide and viral infection (Durner et al., 1998). The clear separation of the cADPR and IP₃ stores from the NAADP-stores indicates the latter are organelles distinct from the ER. In the stratified eggs, they appear to colocalize with mitochondria to the distal pole as shown by rhodamine staining, which suggests that they may be the same. Mitochondria have also been shown to contain an NADase capable of synthesizing cADPR (Ziegler et al., 1997). However, neither cADPR nor NAADP have been reported to be involved in mobilizing the mitochondrial Ca²⁺. Evidence against the possibility that the NAADP-stores are mitochondria comes from previous fractionation studies, which shows that the distribution of the NAADP stores is distinct from cytochrome c oxidase, a mitochondrial marker (Lee and Aarhus, 1995). More importantly, mitochondria in unfertilized sea urchin eggs are essentially devoid of Ca²⁺. Thus, treatment with mitochondrial uncouplers produces Ca²⁺ release only in fertilized but not unfertilized eggs (Eisen and Reynolds, 1985). This is in striking contrast to the dramatic Ca²⁺ release induced by NAADP in the unfertilized eggs. The NAADP stores are also unlikely to be the cortical granules since we have previously shown by electron microscopy that they remain attached to the egg cortex after stratification (Lee and Epel, 1983). The NAADP stores may thus be a new Ca²⁺ storing organelle that has not been previously described.

A unique feature of these novel stores is the presence of a thapsigargin-insensitive Ca²⁺-ATPase (Genazzani and Galione, 1996), which is consistent with the NAADP stores segregating to a pole opposite to the one stained by Bodipy-thapsigargin. We have also directly verified that the eggs do possess a thapsigagin-insensitive Ca²⁺-ATPase. An intracellular Ca²⁺-ATPase called PMR1, showing similar insensitivity to thapsigargin, has been cloned from yeast (Sorin et al., 1997). A homologous protein has also been identified in rat, containing 919 amino acids and having the same apparent transmembrane organization and all of the conserved domains present in other P-type ATPases (Gunteski-Hamblin et al., 1992). The approach described in this study provides the first visualization of this new Ca²⁺ store and should be useful for characterizing it further.

A remarkable observation reported in this study is that the Ca²⁺ wave associated with fertilization always begins at the nuclear pole of the stratified eggs, which is consistent with that reported previously (Eisen and Reynolds, 1985). This long-standing puzzle can now be accounted for by the equally remarkable preference for sperm-egg fusion to occur at the nuclear pole of the stratified eggs. These results suggest an intriguing possibility that the sperm-egg fusion sites on the surface of the eggs may be physically connected to internal organelles, such as the cADPR-sensitive Ca²⁺ stores. Stratification of the eggs could thus result in the sperm-egg fusion sites concentrating in the nuclear halves, together with the stores. Alternatively, the egg surface may contain hot zones for sperm-egg fusion. Unfertilized sea urchin eggs are perfectly symmetrical with no visible marks for distinguishing one point on the egg surface from another. Stratification of the eggs provides the necessary asymmetry that allows these hot zones to be visualized.

The presence of multiple, separate and interacting Ca²⁺ stores provides cells with a versatile Ca²⁺ signaling mechanism. Diverse stimuli could be specifically linked to separate Ca²⁺ stores. Thus, activation of surface receptors by impermeant molecules is generally linked to the inositol trisphosphate pathway (Berridge, 1993), while nitric oxide, a permeant molecule, has been shown, in some cases, to signal via the cADPR-sensitive stores (Clementi et al., 1996; Reyes-Harde et al., 1999; Willmott et al., 1996). In plants, the cADPR- and not the inositol trisphosphate-sensitive Ca²⁺-stores are involved in mediating the response to abscisic acid (Wu et al., 1997). Similar exclusive involvement of the cADPR-sensitive stores has been reported for insulin secretion by pancreatic β-cells (Takasawa et al., 1998; Takasawa et al., 1993). The distribution of different types of Ca²⁺ stores need not be uniform. Segregated distribution of the stores in cells allows them to serve localized signaling functions (Albrieux et al., 1998; Zhang et al., 1999). This is the case in Ascidian oocytes, in which the NAADP- and cADPR-sensitive stores appear to be restricted to the cortex and their mobilization effectively modulates membrane events, while the inositol trisphosphate-sensitive stores are distributed in the cytoplasm mediating cytosolic Ca²⁺ oscillations (Albrieux et al., 1998). Likewise, in pancreatic acinar cells, the NAADP-induced Ca²⁺ signal has been proposed to function as a local trigger, which is then amplified by Ca²⁺-induced Ca²⁺ release involving the cADPR- and inositol trisphosphate-sensitive stores to produce a global and propagative Ca²⁺ wave (Cancela et al., 1999). Clearly, our understanding of Ca²⁺ mobilization as a signaling mechanism would not be complete without elucidating the complexity of regulation and interaction among these diverse and novel Ca²⁺ stores.

The study was supported by NIH grants HD17484 and GM61568 to H.C.L.