Vertebrate female reproduction is limited by the oocyte stockpiles acquired during embryonic development. These are gradually depleted over the organism’s lifetime through the process of apoptosis. The timer that triggers this cell death is yet to be identified. We used the Xenopus egg/oocyte system to examine the hypothesis that nutrient stores can regulate oocyte viability. We show that pentose-phosphate-pathway generation of NADPH is critical for oocyte survival and that the target of this regulation is caspase-2, previously shown to be required for oocyte death in mice. Pentose-phosphate-pathway-mediated inhibition of cell death was due to the inhibitory phosphorylation of caspase-2 by calcium/calmodulin-dependent protein kinase II (CaMKII). These data suggest that exhaustion of oocyte nutrients, resulting in an inability to generate NADPH, may contribute to ooctye apoptosis. These data also provide unexpected links between oocyte metabolism, CaMKII, and caspase-2.

 The precise role of caspase-2 in cell death has been controversial, but recent studies demonstrate that it can act upstream of the mitochondria in a number of settings to trigger cytochrome c release (Guo et al., 2002; Lassus et al., 2002; Robertson et al., 2002). Caspase-2 is activated by binding to oligomerizing adaptor proteins. In response to various stressors, caspase-2 is recruited into highmolecular-weight complexes reminiscent of, though distinct from, the Apaf-1/caspase-9 apoptosome (Read et al., 2002). The caspase-2 prodomain can recruit an adaptor protein, RAIDD, which, when overexpressed, induces caspase-2 activation (Duan and Dixit, 1997). Recently it was reported that p53 induces expression of PIDD, a protein which promotes formation of a caspase-2 activation complex containing PIDD, caspase-2, and RAIDD (Tinel and Tschopp, 2004). It is not known whether PIDD or RAIDD regulates caspase-2 in the oocyte. Although genetic analyses and microinjection studies have provided critical information for understanding oocyte apoptosis, most vertebrate oocytes are not amenable to biochemical analysis due to their small size and limited abundance. However, over a decade ago, Newmeyer et al. reported that extracts prepared from eggs of the frog, Xenopus laevis, could, upon prolonged incubation at room temperature, spontaneously recapitulate many events of apoptosis, including mitochondrial cytochrome c release, caspase activation, and nuclear fragmentation (Newmeyer et al., 1994).

 Importantly, this in vitro apoptosis could be inhibited by antiapoptotic Bcl-2 proteins and caspase inhibitors, suggesting that at least some aspects of germ-cell apoptosis are faithfully recapitulated in this system, thereby providing a biochemically manipulable setting in which to understand germ-cell apoptosis (Kluck et al., 1997). Despite the ease of manipulation of Xenopus eggs and oocytes and the manifestation of apoptotic markers, it was not clear what might be driving apoptosis in this system. Although initial reports suggested that the timing of hormone administration used to elicit egg laying might determine the susceptibility of eggs to apoptosis, research in the intervening years has failed to establish a firm correlation between the rates of apoptosis in Xenopus eggs and the hormonal regimen used to obtain the eggs. Hence, we were driven to consider other features of the oocyte/egg that might contribute to cell-death regulation. In this regard, we were interested in the fact that oocytes are uniquely endowed with a large stockpile of nutrients in the form of both yolk proteins and glycogen stores, which are used to sustain early embryonic development.  In lymphocytes and neurons, growth-factor withdrawal leads to a drop in glucose uptake and glycolytic rates prior to commitment to cell death (Deckwerth and Johnson, 1993; Deshmukh et al., 1996; Rathmell et al., 2000; Vander Heiden et al., 2001). Moreover, artificial maintenance of glycolysis confers some resistance to cell death induced by cytokine withdrawal, suggesting that active growth-factor signaling is required to maintain sufficient metabolism to prevent death (Rathmell et al., 2001, 2003). 

In addition, glucose6-phosphate dehydrogenase activity, which promotes the metabolism of glucose through the pentose phosphate pathway, can protect CHO cells from death induced by ionizing radiation (Tuttle et al., 2000). Conversely, inhibition of both glycolytic and pentose phosphate pathways can promote the apoptotic death of some cultured cells (Comin-Anduix et al., 2002; Tian et al., 1999; Tuttle et al., 2000). Thus, it was attractive to hypothesize that nutrient stores in the oocyte might support cell survival through modulation of metabolic pathways. According to such a scenario, depletion of energy stores within oocytes over time would result either in loss of a critical survival pathway or the de novo engagement of a cell-death pathway. Here, using the facile biochemistry provided by the Xenopus system, we report the identification and characterization of a novel caspase-2 regulatory pathway responsive to the metabolic state of the oocyte/egg. We demonstrate that glucose-6-phosphate (G6P) sufficient to drive continual operation of the pentose phosphate pathway can greatly prolong germ-cell survival. Moreover, our data indicate that NADPH generation by this pathway is critical for promoting survival and that a surfeit of NADPH promotes a calcium/calmodulindependent protein kinase II (CaMKII) dependent inhibitory phosphorylation of caspase-2. Mutant caspase-2 resistant to CaMKII phosphorylation can circumvent this metabolism-dependent suppression of apoptosis, both in egg extracts and intact oocytes. 

Accordingly, we added G6P directly to the in vitro egg extract system and monitored cleavage of the model substrate, DEVDpNA, as a measure of caspase-3 activity. As shown in Figure 1A, the caspase-3 activation observed following incubation of the extract at room temperature was entirely suppressed by G6P addition. Note that G6P, which in oocytes is derived largely from gluconeogenesis or glycogenolysis, was used in these experiments rather than glucose due to the presence of very low hexokinase activity in these extracts (C.E.H. and L.K.N., unpublished data). We also found that G6P could inhibit apoptotic fragmentation of sperm nuclei added to egg extracts (Figure 1B). Consistent with the critical role for mitochondria in the induction of Xenopus egg apoptosis, the suppressive effects of G6P could be seen at the level of cytochrome c release (Figure 1A, lower panel). Moreover, the inhibitory effect of G6P on caspase activation and cytochrome c release could be overridden by addition of truncated Bid (tBid) to forcibly promote cytochrome c release (Figure 1C). These data indicate that G6P-containing extracts were still capable of activating caspases and that the observed block to apoptosis following G6P treatment occurred upstream of or at the level of the mitochondria. Pentose-Phosphate-Pathway Intermediates Suppress Cell Death In Xenopus eggs/oocytes, G6P is used for glycogen deposition or is metabolized through the pentose phosphate pathway. Indeed, the fate of radiolabeled G6P has been followed in amphibian oocytes, and little is metabolized by glycolysis (Dworkin and Dworkin-Rastl, 1989). Consistent with this, we found that glycolytic intermediates, including glyceraldehyde-3-phosphate and pyruvate, had no effect on caspase-3 activation in egg extracts (Figure 1D).

 In contrast, 6-phosphogluconate, a pentose-phosphate-pathway intermediate, could mirror the ability of G6P to inhibit apoptosis (Figure 1E). These data suggested that some feature of pentosephosphate-pathway operation could suppress apoptosis. A key metabolic by-product of the pentose phosphate pathway is NADPH, produced when G6P is metabolized to 6-phosphogluconolactone and 6-phosphogluconate is converted to ribulose-5-phosphate. Hypothesizing that production of NADPH might be the relevant consequence of G6P addition for apoptotic inhibition, we sought to generate NADPH using an alternative route and to examine the effects of direct NADPH addition to extracts. Addition of malate, which promotes NADPH production in concert with malic enzyme, potently inhibited mitochondrial cytochrome c release and caspase activation in the egg extract (Figure 1F). Moreover, despite its relative instability in solution, NADPH could markedly suppress caspase activation (Figure 1E). These data suggest that G6P flux through the pentose phosphate pathway, leading to NADPH production, can potently suppress egg extract apoptosis. Although NADPH can act as a reducing agent, we do not believe that it is this feature of NADPH function that is responsible for apoptotic suppression as neither the SOD mimetic MnTBAP, the electron scavenger NAC, nor other reducing agents (TEMPO or reduced glutathione) could suppress extract apoptosis (Figure 1G and data not shown).

 Pentose-Phosphate-Pathway Inhibition Induces Apoptosis in Egg Extracts and Oocytes Since hyperstimulation of the pentose phosphate pathway or NADPH addition inhibited caspase activation, we speculated that inhibition of the endogenous pentose phosphate pathway in egg extracts should accelerate apoptosis, particularly if the spontaneous extract apoptosis relied upon nutrient depletion during incubation. To test this, we treated extracts with the G6P dehydrogenase inhibitor, dehydroisoandrosterone (DHEA), to prevent entry of endogenous G6P into the pentose phosphate pathway (Schwartz and Pashko, 2004). This treatment accelerated apoptosis, largely eliminating the lag normally seen before caspase activation, as would be expected if pentose-phosphate-pathway function were important for extract “survival” (Figure 2A). We also examined extracts to determine whether endogenous G6P was depleted over time upon incubation. For this purpose, extracts were incubated at room temperature and assayed for caspase activation (Figure 2B, top panel). Aliquots of extract withdrawn at the start of the incubation (0 hr) or just prior to caspase activation (2 hr) were incubated with a vast excess of G6P dehydrogenase and NADP and assayed over time for generation of NADPH, which, under these circumstances, depended solely on the available G6P. As shown in Figure 2B (middle panel), G6P levels (as measured by the rate of NADPH generation) dropped substantially within the 2 hr of incubation. Note also that NADPH generation remained high in extracts supplemented with excess G6P (Figure 2B, lower panel). We wished to extend these observations to whole oocytes to determine if the pentose phosphate pathway was important for suppressing death in these intact cells. 

For this, we developed a visual assay for oocyte death. Normally, oocytes contain a light-colored vegetal half and a dark, uniformly pigmented animal half. When induced to undergo maturation with progesterone, oocytes exhibit a white spot resulting from nuclear-envelope breakdown and clearance of surrounding pigment granules. We noted that a similar but considerably larger white spot appeared at the animal poles of oocytes injected with cytochrome c to induce caspase activation (C. Holley, personal communication). These oocytes did not exhibit elevated Cdc2/ cyclin kinase activity characteristic of maturing oocytes but rather showed high levels of caspase-3 activity (data not shown). Morphological events of apoptosis have been observed in oocytes by others, both in Xenopus and in other species (Bagowski et al., 2001; Bhuyan et al., 2001; Braun et al., 2003; Demirci et al., 2001, 2002, 2003; Voronina and Wessel, 2001). Interestingly, we had observed that oocytes taken freshly out of a virgin female frog are remarkably refractory to apoptosis. With the idea that this resistance might stem from continual functioning of the pentose phosphate pathway, we soaked fresh oocytes in a DHEA-containing solution and monitored the appearance of large white spots at the animal poles.

 As shown in the micrographs in Figure 2C and the graph in Figure 2D, 200
M DHEA rapidly killed oocytes, while control treated oocytes remained healthy. These data were corroborated in experiments where cytochrome c was released from mitochondria in DHEA-treated but not control oocytes (Figure 2E). Importantly, this acceleration of oocyte death occurred at even lower doses of DHEA (100
M) and could be reversed by simultaneous treatment with the cell permeant D-methylmalate, which would be expected to promote NADPH generation within oocytes (Figure 2F). These data demonstrate that pentosephosphate-pathway inhibition promotes rapid oocyte death, suggesting that this pathway is normally critical for oocyte viability. Caspase-2 Is a Target of NADPH-Mediated Apoptotic Inhibition Since G6P (or NADPH) could inhibit apoptosis, we wished to identify the target of this inhibition. As mentioned above, genetic analyses in mice had implicated caspase-2 as an important constituent of oocyte apoptotic pathways (Bergeron et al., 1998). To determine if caspase-2 was affected by G6P, we first examined processing of radiolabeled caspase-2 during spontaneous apoptosis. As shown in Figure 3A, addition of G6P, pentose-phosphate-pathway intermediates, malate, or NADPH all prevented caspase-2 processing, while glycolytic intermediates had no effect. Moreover, G6P inhibited endogenous caspase-2 processing, as detected by immunoblotting (Figure 3H). Caspase-2 processing in untreated extracts was inhibitable by the baculoviral caspase inhibitor p35 and the peptide inhibitor of caspase-2, VDVAD-fmk, but was not inhibited by BclxL, which prevented cytochrome c release and caspase-3 activation (Figures 3B, 3C, and 3E).