Elsevier

Advances in Biological Regulation

Volume 54, January 2014, Pages 242-253
Advances in Biological Regulation

DGKζ under stress conditions: “To be nuclear or cytoplasmic, that is the question”

https://doi.org/10.1016/j.jbior.2013.08.007Get rights and content

Abstract

Eukaryotic cells have evolved to possess a distinct subcellular compartment, the nucleus, separated from the cytoplasm in a manner that allows the precise operation of the chromatin, thereby permitting controlled access to the regulatory elements in the DNA for transcription and replication. In the cytoplasm, genetic information contained in the DNA sequence is translated into proteins, including enzymes that catalyze various reactions, such as metabolic processes, energy control, and responses to changing environments. One mechanism that regulates these events involves phosphoinositide turnover signaling, which generates a lipid second messenger, diacylglycerol (DG). Since DG acts as a potent activator of several signaling molecules, it should be tightly regulated to keep cellular responsiveness within a physiological range. DG kinase (DGK) metabolizes DG by phosphorylating it to generate phosphatidic acid, thus serving as a critical regulator of DG signaling. Phosphoinositide turnover is employed differentially in the nucleus and the cytoplasm. A member of the DGK family, DGKζ, localizes to the nucleus in various cell types and is considered to regulate nuclear DG signaling. Recent studies have provided evidence that DGKζ shuttles between the nucleus and the cytoplasm in neurons under pathophysiological conditions. Transport of a signal regulator between the nucleus and the cytoplasm should be a critical function for maintaining basic processes in the nucleus, such as cell cycle regulation and gene expression, and to ensure communication between nuclear processes and cytoplasmic functions. In this review, a series of studies on nucleocytoplasmic translocation of DGKζ have been summarized, and the functional implications of this phenomenon in postmitotic neurons and cancer cells under stress conditions are discussed.

Introduction

Phospholipids consist mainly of phosphatidylcholine (PC), phosphatidylserine (PS), phosphatidylethanolamine (PE), and phosphatidylinositol (PI). They are the major constituents of biological membrane, and each of these constituents plays a unique role in the functional properties of the membrane.

Importantly, the metabolism of PI, constituting approximately 10% of membrane phospholipids, is intimately involved in signal transduction, which produces a lipid second messenger diacylglycerol (DG). DG signaling is tightly regulated by phosphorylation of DG to phosphatidic acid by DG kinase (DGK) (Kanoh et al., 1990). Reportedly, DG metabolism via catalysis by DGK is widely present in various subcellular compartments. DGK is composed of a family of isozymes. Each member has a unique feature in terms of enzymatic characteristics, regulatory mechanism, binding partner, tissue distribution, and subcellular localization (Goto et al., 2007, Sakane et al., 2007, Merida et al., 2008, Topham and Epand, 2009).

Within the DGK family, DGKζ contains both a nuclear localization signal (NLS) and nuclear export signal (NES) and has been shown to localize to the nucleus (Goto and Kondo, 1996, Evangelisti et al., 2010). Thus, DGKζ is thought to be involved in nuclear DG metabolism, which is differentially regulated from that in the cytoplasm (Martelli et al., 2002, Martelli et al., 2005, Goto et al., 2006, Evangelisti et al., 2007a). Recent studies have reported that DGKζ shuttles between the nucleus and the cytoplasm in neurons under pathophysiological conditions (Hozumi et al., 2003, Ali et al., 2004, Nakano et al., 2006, Saino-Saito et al., 2011, Okada et al., 2012, Suzuki et al., 2012). In this review, we describe a series of studies on the nucleocytoplasmic translocation of DGKζ and discuss its functional implications on cellular pathophysiology under stress conditions.

DGKζ was isolated in 1996 from rat brain (the rat clone was previously also termed DGK-IV) (Goto and Kondo, 1996) and human endothelial cell libraries (Bunting et al., 1996). It closely resembles the eye-specific DGK of Drosophila that is encoded by rdgA (Masai et al., 1993) in that DGKζ contains ankyrin-like repeats, but no EF-hand motifs. Characteristically, an NLS is included in mammalian DGKζ (Goto and Kondo, 1996), but is not present in the corresponding region of Drosophila DGK (rdgA). The NLS is a bipartite type, which consists of a cluster of two adjacent basic amino acids (lysine or arginine) that are separated by 10 amino acids from a second cluster, in which three of the next five amino acids are also basic. The NLS sequence overlaps with the phosphorylation domain of myristoylated alanine-rich C-kinase substrate (MARCKS) motif (Bunting et al., 1996). As is suggested by the NLS sequence, DGKζ has been shown to localize to the nucleus in cDNA-transfected cells (Goto and Kondo, 1996, Topham et al., 1998).

Subcellular localization in native cells has been investigated by immunohistochemical studies using a specific antibody, which confirmed nuclear localization of DGKζ in various types of cells, i.e. neurons in the brain (Hozumi et al., 2003), spinal ganglion cells (Sasaki et al., 2006), hepatocytes (Nakano et al., 2012), retinal neurons (Hozumi et al., 2013), pituitary endocrine cells (Hozumi et al., 2010), alveolar epithelial cells, and macrophages (Katagiri et al., 2005) under normal conditions. It should be mentioned that DGKζ generally localizes to the nucleus in various parts of the brain, such as the cortex and hippocampus, whereas it localizes to both the nucleus and the cytoplasm in cerebellar Purkinje cells, but mostly to the cytoplasm in these cells at an early stage of development (Hozumi et al., 2003). In spinal ganglion cells, three patterns of DGKζ staining have been discerned: nuclear, cytoplasmic, and nucleocytoplasmic types (Sasaki et al., 2006). These findings suggest that DGKζ dynamically shuttles between the nucleus and the cytoplasm in the developing brain and under certain physiological conditions.

It is generally accepted that the subcellular localization of a given molecule is never static, but is rather in dynamic equilibrium. This is particularly true for proteins containing both an NLS and an NES, which may shuttle between the nucleus and the cytoplasm, depending on the dominant effect of NLS over NES, or vice versa. Nucleocytoplasmic shuttling of such proteins is thought to be regulated by importin and Crm1/exportin systems (Yoneda, 2000). Importin α is known to serve as an adaptor for an NLS-containing cargo protein and importin β. Importin β recognizes the cargo-importin α complex and transfers it to the cytoplasmic face of the nuclear pore. Besides an NLS, a functional NES sequence (LxxxLxxLxL) is also present in DGKζ (Evangelisti et al., 2010). NES mediates signal-dependent transport of proteins from the nucleus back into the cytoplasm through the nuclear exporter Crm1/exportin.

In fact, of 7 importin isoforms, Qip1 and NP1 have been found to bind to DGKζ, and Crm1/exportin has been shown to interact with DGKζ (Okada et al., 2011). Interacting molecules may inevitably modulate or suppress the binding affinity of NLS and NES for importin and Crm1/exportin, respectively, which would affect the dominance of NLS or NES. This is exemplified by the newly identified DGKζ-interacting proteins, viz., nucleosome assembly protein 1-like 1 (NAP1L1) and NAP1-like 4 (NAP1L4) (Okada et al., 2011). NAP1 is a highly conserved histone chaperone protein that is involved in the regulation of the H2A-H2B histone heterodimer (Zlatanova et al., 2007). NAP1Ls have been shown to associate with the NLS region of DGKζ, thereby attenuating the interaction of DGKζ and importins, although this association has no effect on interaction with Crm1. At a cellular level, overexpression of NAP1Ls blocks nuclear localization of DGKζ, while RNA silencing of NAP1Ls enhances its nuclear localization. In this instance, bacterial recombinant DGKζ has been found to bind strongly to native NAP1Ls, suggesting that modification of DGKζ, such as phosphorylation, is not always necessary for the association.

In addition, the MARCKS domain that is overlapped with NLS and the C-terminal PDZ-binding domain may also be involved in the translocation. PKCα associates with DGKζ and phosphorylates its MARCKS homology domain, which may reduce its nuclear accumulation (Topham et al., 1998, Luo et al., 2003). The syntrophin family of proteins also associates with DGKζ through the PDZ-binding motif at the C-terminus of DGKζ, which anchors DGKζ to the cytoplasm in muscle cells (Hogan et al., 2001, Abramovici et al., 2003). Furthermore, it is reported that other interacting partners, such as Rac1 (Yakubchyk et al., 2005), RasGRP1 (Topham and Prescott, 2001), phosphatidylinositol 5-kinase type Iα (Luo et al., 2004), retinoblastoma protein (Los et al., 2006), and β-arrestins (Nelson et al., 2007), may also regulate the subcellular localization of DGKζ (Rincon et al., 2012), although how these proteins participate in its nucleocytoplasmic translocation remains elusive.

Pathological conditions, such as ischemia, are known to induce an increase in second messengers, such as DG and Ca2+ (Yoshida et al., 1986, Uematsu et al., 1988), suggesting that PI signaling is deeply involved in this pathological process. An animal model of transient forebrain ischemia is widely used to investigate the effect of ischemic stress on neurons, especially hippocampal neurons that show vulnerability compared with those in other brain areas. Under these conditions, DGKζ is translocated from the nucleus to the cytoplasm in hippocampal neurons after 20 min of ischemic stress (Ali et al., 2004). During the course of reperfusion, DGKζ is never relocated to the nucleus, but its levels are gradually attenuated. However, prolonged cessation of blood supply in an infarction model induces immediate disappearance of DGKζ in cortical neurons after 90 min, which engenders instantaneous cell death (Nakano et al., 2006). Taking the data from transient ischemic and infarction models together, it appears that the more severe the stress, the shorter the duration of cytoplasmic localization of DGKζ, which suggests that the duration of cytoplasmic localization of DGKζ correlates with the survival period after insult.

Is this phenomenon unique to ischemic stress? Cessation of blood flow, followed by deprivation of oxygen and glucose, results in insufficient ATP generation in neurons; consequently the ionic gradient is lost and glutamate is released. Furthermore, defective glutamate uptake together with glutamate release by disabled astrocytes leads to a high perisynaptic concentration of glutamate in the ischemic brain (Swanson et al., 2004). These conditions engender glutamate excitotoxic stress, which promotes overstimulation of glutamate receptors and causes a resultant massive influx of calcium that activates various catabolic processes. Glutamate excitotoxicity is known to be linked not only to ischemic injury, but also to other conditions, such as seizures and chronic neurodegenerative diseases (Lau and Tymianski, 2010).

A limbic seizure model induced by kainate injection in rats is another model available for studying excitotoxic stress (Ben-Ari, 1985). The epileptogenic actions of kainate are mediated by kainate receptors located on mossy fiber synapses that are enriched in hippocampal CA3 pyramidal neurons. Subsequently, the neuronal activity is transmitted to the CA1 region, the major output gate from the hippocampus, which engenders propagation of seizures to other limbic structures and subsequent generation of a limbic partial type of seizure. In this model, kainate-induced seizures can also induce nucleocytoplasmic translocation of DGKζ in hippocampal neurons (Saino-Saito et al., 2011). DGKζ becomes dominant in the cytoplasm by 2 h after kainate injection and remains in this location for at least 3 days. In this seizure model, neuronal damage in the hippocampal region is caused selectively by seizures per se and is not attributable to any anoxic–ischemic episode (Pinard et al., 1984). In this case, the local blood flow is in fact rather increased during seizures, which more than compensates for the increased oxygen consumption.

It has been reported that neurons damaged after hypoxia–ischemia immediately reenter the cell cycle and causes aberrant apoptosis-associated DNA synthesis, while kainate-induced excitotoxicity gradually engenders apoptotic neuronal death, without causing DNA synthesis (Kuan et al., 2004). Even though the apoptotic processes engaged in ischemia and seizures are mediated through distinct molecular pathways, the excitotoxicity is triggered by glutamate receptor activity, and subsequent massive calcium influx, in both cases. Induction of irreversible nucleocytoplasmic translocation of DGKζ is presumed to be involved in the glutamate excitotoxicity and massive calcium influx. Excitotoxicity caused by transient ischemia and kainate-seizures can result in neuronal death, which is commonly observed after several hours to days following excessive calcium influx by glutamate stimulation, a phenomenon referred to as “delayed neuronal death” (Kirino, 1982, Coyle and Puttfarcken, 1993, Dirnagl et al., 1999). Therefore, it has been suggested that nucleocytoplasmic translocation of DGKζ at an early phase of ischemic stress is “a point of no return” and is relevant to the subsequent irreversible death process in neurons.

These issues raised in animal models have been addressed using acute hippocampal slice culture (Suzuki et al., 2012) and primary neuron culture systems (Okada et al., 2012). First, acute hippocampal slice culture is a common experimental tool for the analysis of hippocampal neurons by histochemical, physiological, and pharmacological methods. It confers an advantage in that it maintains neuronal connectivity and interaction with glial cells. In addition, this system makes it possible to precisely manipulate the environmental conditions (Gahwiler et al., 1997). Hippocampal slices are exposed to oxygen–glucose deprivation (OGD) in an artificial cerebrospinal fluid at 30 °C to simulate an ischemic model of the brain, in which the triggering factors and detailed time course can be evaluated. The DGKζ nucleocytoplasmic translocation phenomenon is well recapitulated in hippocampal CA1 neurons in slice culture (Suzuki et al., 2012). Under continuous OGD conditions, DGKζ gradually translocates from the nucleus to the cytoplasm after 20 min. However, in transient OGD, followed by reoxygenation, DGKζ remains in the nucleus at 10 min OGD, although it becomes dominant in the cytoplasm after 20 min of reoxygenation and is detected almost exclusively in the cytoplasm after 30 min of reoxygenation (Fig. 1). Under these OGD–reoxygenation conditions, a minimum of 8–10 min OGD is sufficient to trigger the cascade to induce irreversible DGKζ cytoplasmic translocation, which appears to be more extensive than that under continuous OGD. These findings suggest that cytoplasmic translocation of DGKζ may be controlled through an active, energy-requiring process. Furthermore, this cytoplasmic translocation is well correlated with nuclear shrinkage, an early indicator of neuronal injury in hippocampal slices (Bonde et al., 2002), indicating a functional link between DGKζ translocation and neuronal degeneration.

Pharmacological approaches have revealed that the N-methyl-d-aspartate (NMDA)-type glutamate receptor antagonist AP5 inhibits DGKζ cytoplasmic translocation under OGD–reperfusion conditions, whereas its activation by means of an NMDA agonist induces this translocation under normal conditions. In this case, elimination of extracellular Ca2+ abolishes the translocation under OGD–reperfusion conditions. Furthermore, Ca2+ influx via other routes, such as the voltage-dependent Ca2+ channel and Ca2+ ionophore, fails to induce this translocation. These data clearly indicate that DGKζ cytoplasmic translocation is triggered by an influx of extracellular Ca2+ via the NMDA receptor. This feature seems to be closely related to a previous finding that Ca2+ loading via NMDA receptor channels is toxic, whereas identical Ca2+ loads incurred through voltage-sensitive Ca2+ channels are completely innocuous (Sattler et al., 1998). Therefore, the DGKζ cytoplasmic translocation phenomenon is compatible with the ‘source-specificity’ hypothesis that Ca2+ toxicity is linked to the route of Ca2+ entry and the distinct second messenger pathways (Tymianski et al., 1993, Arundine and Tymianski, 2004).

A downstream cascade of Ca2+ signaling in neurons leads to the activation of calmodulin, calcium–calmodulin-dependent protein kinase II, PKC, and calcineurin (Orrenius et al., 2003). Of the inhibitors for these molecules, the selective PKC inhibitor GF109203X suppresses DGKζ cytoplasmic translocation under OGD–reperfusion conditions, whereas other inhibitors have no effect, suggesting that PKC activity partially, if not entirely, regulates DGKζ cytoplasmic translocation. Whether PKC directly phosphorylates the MARCKS domain of DGKζ or indirectly regulates DGKζ remains elusive.

What are the functional implications of cytoplasmic translocation of DGKζ? Is it proapoptotic or antiapoptotic? A second experimental system, i.e., primary neuron culture, has been used to address these issues in neurons at the cellular level (Okada et al., 2012). After 30 min of stimulation with a toxic concentration of glutamate (300 μM) followed by 24 h of incubation in the regular medium, DGKζ translocates from the nucleus to the cytoplasm and is then greatly attenuated at the protein level. The DGKζ mRNA transcript levels show no apparent change during the course of such an experiment, suggesting that regulation occurs at the protein level. Moreover, after exposure to glutamate excitotoxicity or hypoxic stress, DGKζ protein undergoes polyubiquitination followed by degradation via the proteasome system. In this regard, it should be noted that proteolytic degradation of DGKζ is executed in the cytoplasm, as evidenced by the fact that leptomycin B, a Crm1/exportin nuclear export carrier protein inhibitor, suppresses DGKζ polyubiquitination.

The functional consequence of DGKζ downregulation is suggested by previous studies that show that nuclear DGKζ negatively regulates cell cycle progression (Topham et al., 1998, Evangelisti et al., 2007b). On the other hand, DGKζ overexpression causes accumulation of cells in the G0/G1 phase of the cell cycle; both DGK activity and a functional NLS are required for this cell cycle block (Topham et al., 1998), suggesting that this process involves nuclear DG metabolism catalyzed by DGKζ. Furthermore, DGKζ interacts with the retinoblastoma protein (pRB) (Los et al., 2006), and DGKζ overexpression downregulates cyclin D1 expression in C2C12 myoblast cells through upregulation of BTG, a transcriptional regulator of cyclin D1 with a strong antiproliferative function, coincidentally with decreased levels of Ser807/811 phosphorylated pRB (Evangelisti et al., 2009). In contrast, downregulation of DGKζ by siRNA results in increased phosphorylation of pRB at these sites, thereby liberating E2F from its complex with pRB. Accordingly, E2F can upregulate the transcription of genes required for the G1/S phase transition (Dyson, 1998, Cobrinik, 2005). De novo expression of cell cycle proteins, such as cyclin D1, induced by seizures and other neurodegenerative disorders, leads to cell cycle reentry in postmitotic neurons, culminating in apoptotic cell death (Freeman et al., 1994, Herrup and Yang, 2007, Koeller et al., 2008). Knockdown of DGKζ results in enhanced expression of cyclins and phosphorylated pRB, together with increased apoptotic marker expression and DNA fragmentation in hippocampal neurons after kainate-induced seizures (Okada et al., 2012). With regard to the time scales of this phenomenon, it should be mentioned that, in dissociated cultured neurons, cytoplasmic translocation of DGKζ is observed after several hours of glutamate stimulation (Okada et al., 2012), whereas it occurs after 20 min of hypoxia in the animal model (Ali et al., 2004), or within 1 h under OGD–reperfusion conditions in hippocampal slice culture (Suzuki et al., 2012). This time scale difference may be ascribed to the distinct environmental conditions surrounding neurons: in primary neuron culture systems, neurons are principally dissociated from each other and have no intimate contact with glial cells, whereas neurons retain physiological interactions with glial cells in the brain in vivo and in slice culture. Presumably, cytoplasmic translocation of DGKζ may be facilitated by glial cells, which surround perisynaptic space and are involved in the regulation of the local glutamate concentration. Further study is needed to elucidate this point.

Collectively, the experiments on primary cultured neurons have revealed a series of events that can be observed upon DGKζ in neurons suffering from hypoxic or excitotoxic stress. Under stress conditions, DGKζ translocates from the nucleus to the cytoplasm. Once translocated, cytoplasmic DGKζ is susceptible to degradation through the ubiquitin–proteasome system. Prolonged absence of DGKζ from the nucleus makes neurons susceptible to aberrant cell cycle reentry, followed by apoptotic cell death. These findings are supported by an experiment showing that DGKζ-deficient neurons are more vulnerable to excitotoxicity, because of aberrant cell cycle reentry, although those neurons do not succumb directly to apoptosis under normal conditions (Okada et al., 2012).

Previous reports have shown that cardiac overexpression of DGKζ exerts beneficial effects on hearts suffering from pressure-overload or infarction and increases the survival rate thereafter (Niizeki et al., 2007, Niizeki et al., 2008). These results implicate DGKζ in a protective role in cellular pathophysiology in an as yet unspecified manner; moreover, how DGKζ cytoplasmic translocation is engaged in the stress response machinery has remained elusive. A recent report has revealed a pathophysiological link between DGKζ cytoplasmic translocation and p53-mediated cytotoxicity after DNA damage (Tanaka et al., 2013). The tumor suppressor gene product p53 plays a central role in the coordination of cellular responses to various stresses, which include DNA-damage, hypoxia, and oncogene activation (Polager and Ginsberg, 2009, Vousden and Prives, 2009). Under stress conditions, p53 triggers various cellular reactions that engender DNA repair, cell cycle arrest, apoptosis, and differentiation. In the brain, p53 activation is implicated in the apoptotic cell death of postmitotic neurons after ischemia, traumatic injury, and neurodegenerative disorders (Culmsee and Mattson, 2005). At the cellular level, primary cultured neurons of the hippocampus or cortex derived from p53-deficient mice are resistant to kainate and glutamate excitotoxicity (Xiang et al., 1996). This is also true for the hippocampal neurons of p53-null mice in the kainate-induced seizure model (Morrison et al., 1996).

How does the increased cytoplasmic pool of DGKζ and attenuated nuclear DGKζ modulate p53 function under stress conditions? This question has been addressed using two experimental models, viz., a HeLa cell model of doxorubicin (DOX)-induced DNA damage and an animal model of kainate-induced seizures (Tanaka et al., 2013). These experiments have revealed that DGKζ interacts with p53 under normal and stress conditions and that overexpression of wild-type DGKζ suppresses p53 induction and reduces apoptosis after DOX-induced DNA damage. Notably, these effects on p53 is more pronounced for an NLS-deleted mutant, DGKζΔNLS (cytoplasmic), than for a wild-type DGKζ (nuclear-dominant). In this instance, cytoplasmic DGKζ anchors p53 and facilitates its degradation through the ubiquitin–proteasome system in the cytoplasm in a kinase-independent manner. Furthermore, all events mediated by the cytoplasmic mutant DGKζΔNLS are recapitulated by coexpression of wild-type DGKζ and its binding partners NAP1L1s, which are shown to translocate DGKζ to the cytoplasm by inhibiting its association with importins (Okada et al., 2011). Knockdown of DGKζ abolishes p53 suppression and the cytoprotective effects of NAP1Ls, showing that these effects are ascribed to DGKζ cytoplasmic translocation.

Levels of p53 protein are maintained at low levels in homeostasis by an autoregulatory negative feedback loop (Wu et al., 1993). An E3 ligase, Mdm2, comprises a major component of this autoregulatory feedback system, in which p53 regulates its own levels by inducing Mdm2 expression (Kubbutat et al., 1997, Herrup and Yang, 2007). It has been shown that cytoplasmic DGKζ may be involved in the regulatory mechanism of Mdm2 stabilization. Furthermore, DGKζ has been identified as a p53-inducible gene product in human cells overexpressing wild-type p53 (Vrba et al., 2008). In this condition, p53 induces transcription of the DGKζ gene by binding to its promoter and by subsequent acetylation of promoter-associated histones H3 and H4. Thus, the combined data suggest that DGKζ cytoplasmic translocation may provide a novel negative feedback loop for p53 under stress conditions.

What about the effects of attenuated nuclear localization of DGKζ on p53 transcriptional activity? Interestingly, under DGKζ-knockdown conditions, p53 transcriptional activity is substantially downregulated, albeit p53 protein levels are significantly upregulated (Tanaka et al., 2013). Moreover, under these conditions, both the proarrest target p21 and the proapoptotic target Bax are downregulated at the protein level. These findings suggest that DGKζ is engaged in the basic mechanism of p53 transcriptional activity. At the organismal level, p53 protein is strongly upregulated in the brain of DGKζ-deficient mice under normal conditions, as well as after kainate-induced seizures. It should be pointed out that DGKζ-deficient mice are viable and show no apparent phenotype despite upregulated p53 levels, which is contrary to Mdm2-null mice, which exhibit embryonic lethality, due to the activity of increased levels of p53 protein (Jones et al., 1995, Montes de Oca Luna et al., 1995). This apparent paradox may be explained by the fact that p53 transcriptional activity is greatly compromised in the absence of DGKζ; i.e., p53-inducible gene expression is suppressed to a great extent in DGKζ-deficient mice.

Taken together, nucleocytoplasmic translocation of DGKζ has dual functional implications on cellular responses to stress conditions. The cytoplasmic pool of DGKζ functions as a scaffolding protein that facilitates p53 protein degradation, whereas attenuation of nuclear DGKζ downregulates p53 transcriptional activity. Both of these effects in the cytoplasm and nucleus synergistically repress p53-mediated cytotoxicity. Therefore, it has been suggested that DGKζ serves as a sentinel to control p53 function both during normal homeostasis and during the stress response (Tanaka et al., 2013). In addition, DGKζ cytoplasmic translocation may provide an additional layer to the autoregulatory negative feedback loop of p53, which is finely tuned in stress conditions.

Finally, what is the significance of these findings? Models depicting the roles of DGKζ are summarized in Fig. 2. Nuclear DGKζ negatively regulates cell cycle progression and activates p53 transactivation activity. In postmitotic neurons, NMDA receptor-mediated excessive Ca2+ entry triggers nucleocytoplasmic translocation of DGKζ, and cytoplasmic DGKζ may serve as a scaffolding protein to facilitate p53 degradation in a pathway mediated through the Mdm2-dependent ubiquitin–proteasome pathway. This translocation reduces nuclear DGKζ, which results in suppression of p53 transcriptional activity. Both of the effects on p53 attenuate p53-mediated cytotoxicity. However, prolonged absence of nuclear DGKζ leaves its original role as a negative regulator of cell cycle progression unattended, which can render postmitotic neurons susceptible to the initiation of aberrant cell cycle reentry under stress conditions. This situation is reminiscent of the fact that cells have a host of protective stress responses, most of which switch into execution mode during prolonged activation (Herrup and Yang, 2007). These responses are a double-edged sword that could be either protective or destructive, depending on the cell type, characteristics of the stress, and its severity and duration. This may be the case with DGKζ cytoplasmic translocation in postmitotic neurons.

In this regard, the response of cardiomyocytes, which, like neurons, are postmitotic, terminally differentiated cells, to ischemia, should be mentioned. It seems that cardiomyocytes exhibit reversible changes in the subcellular localization of DGKζ under ischemia/reperfusion conditions (Akiyama et al., 2011): it translocates to the cytoplasm in response to transient ischemic stress, but relocates to the nucleus after reoxygenation in cardiomyocytes. This phenomenon may be interpreted as indicating that an emergency response system is relieved speedily in cardiomyocytes under stress conditions. In this respect, importin α has been shown to accumulate reversibly in the nucleus in response to cellular stresses, and the classical nuclear import pathway is downregulated via depletion of importin α from the cytoplasm (Miyamoto et al., 2004). This feature seems compatible with the reversible translocation of DGKζ in cardiomyocytes under transient ischemic stress. Therefore, mechanisms underlying DGKζ translocation in neurons and cardiomyocytes may differ. These properties may distinguish cardiomyocytes from some other cell types, such as neurons (Bossenmeyer et al., 1998, Chihab et al., 1998), and confer prolonged viability in the face of severe hypoxia to cardiomyocytes, suggesting the existence of an intrinsic mechanism that protects these cells from hypoxia-related apoptosis (Malhotra and Brosius, 1999).

What does the cytoplasmic localization of DGKζ in cancer cells, in which aberrant cell cycle reentry is not of concern, imply? In addition to its role as a transcription factor, p53 elicits a transcription-independent pro-apoptotic response in the cytoplasm, where p53 localizes to the mitochondria and activates a direct mitochondrial death program (Marchenko et al., 2007, Green and Kroemer, 2009). As shown in the HeLa cell model, cytoplasmic DGKζ anchors p53 to the cytoplasm, facilitates its degradation, and exerts a protective effect against p53-mediated cytotoxicity under conditions of DNA damage. This suggests that cytoplasmic DGKζ can confer refractoriness to chemotherapy to cancer cells and may contribute to the pathogenesis of malignant transformation, although it still remains undetermined under what conditions DGKζ is localized to the cytoplasm in cancer cells. However, it has been shown that Ras, an important factor contributing to the maintenance and progression of many types of cancer, is regulated by DGKζ (Topham and Prescott, 2001). DGKζ, but not other DGKs, completely abrogates the activation of Ras that is induced by RasGRP in a DGK activity-dependent manner, suggesting that DGKζ acts as a Ras suppressor. The unclear behaviors of cytoplasmic DGKζ during cancer progression may be dynamically and temporally regulated under different conditions.

Section snippets

Concluding remarks

Careful observation of animal models led to the discovery of that DGKζ translocates between the nucleus and cytoplasm in neurons under stress conditions. This consequently raises several questions and speculations. Experimental investigations addressing these questions at the cellular and molecular levels have yielded some answers, but have also raised more questions. A series of studies has certainly elucidated a new aspect of the p53 regulatory mechanism, although much remains to be

Acknowledgments

This work was supported by Grants-in-Aid from The Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (K.G.).

References (80)

  • K. Goto et al.

    Diacylglycerol, phosphatidic acid, and the converting enzyme, diacylglycerol kinase, in the nucleus

    Biochim Biophys Acta

    (2006)
  • K. Goto et al.

    Cell biology and pathophysiology of the diacylglycerol kinase family: morphological aspects in tissues and organs

    Int Rev Cytol

    (2007)
  • A. Hogan et al.

    Interaction of γ1-syntrophin with diacylglycerol kinase-ζ. Regulation of nuclear localization by PDZ interactions

    J Biol Chem

    (2001)
  • H. Kanoh et al.

    Diacylglycerol kinase: a key modulator of signal transduction?

    Trends Biochem Sci

    (1990)
  • T. Kirino

    Delayed neuronal death in the gerbil hippocampus following ischemia

    Brain Res

    (1982)
  • H.B. Koeller et al.

    Cyclin D1 in excitatory neurons of the adult brain enhances kainate-induced neurotoxicity

    Neurobiol Dis

    (2008)
  • A.P. Los et al.

    The retinoblastoma family proteins bind to and activate diacylglycerol kinase ζ

    J Biol Chem

    (2006)
  • B. Luo et al.

    Diacylglycerol kinase ζ regulates phosphatidylinositol 4-phosphate 5-kinase Iα by a novel mechanism

    Cell Signal

    (2004)
  • R. Malhotra et al.

    Glucose uptake and glycolysis reduce hypoxia-induced apoptosis in cultured neonatal rat cardiac myocytes

    J Biol Chem

    (1999)
  • M. Okada et al.

    Interaction of nucleosome assembly proteins abolishes nuclear localization of DGKζ by attenuating its association with importins

    Exp Cell Res

    (2011)
  • M. Okada et al.

    DGKζ is degraded through the cytoplasmic ubiquitin-proteasome system under excitotoxic conditions, which causes neuronal apoptosis because of aberrant cell cycle reentry

    Cell Signal

    (2012)
  • E. Pinard et al.

    Blood flow compensates oxygen demand in the vulnerable CA3 region of the hippocampus during kainate-induced seizures

    Neuroscience

    (1984)
  • E. Rincon et al.

    Diacylglycerol kinase ζ: at the crossroads of lipid signaling and protein complex organization

    Prog Lipid Res

    (2012)
  • S. Saino-Saito et al.

    Excitotoxicity by kainate-induced seizure causes diacylglycerol kinase ζ to shuttle from the nucleus to the cytoplasm in hippocampal neurons

    Neurosci Lett

    (2011)
  • F. Sakane et al.

    Diacylglycerol kinases: why so many of them?

    Biochim Biophys Acta

    (2007)
  • M.K. Topham et al.

    Mammalian diacylglycerol kinases: molecular interactions and biological functions of selected isoforms

    Biochim Biophys Acta

    (2009)
  • K.H. Vousden et al.

    Blinded by the light: the Growing Complexity of p53

    Cell

    (2009)
  • H. Abramovici et al.

    Diacylglycerol kinase-ζ localization in skeletal muscle is regulated by phosphorylation and interaction with syntrophins

    Mol Biol Cell

    (2003)
  • H. Akiyama et al.

    Nuclear relocation of DGKζ in cardiomyocytes under conditions of ischemia/reperfusion

    Histol Histopathol

    (2011)
  • M. Arundine et al.

    Molecular mechanisms of glutamate-dependent neurodegeneration in ischemia and traumatic brain injury

    Cell Mol Life Sci

    (2004)
  • R. Chihab et al.

    Lack of correlation between the effects of transient exposure to glutamate and those of hypoxia/reoxygenation in immature neurons in vitro

    J Neurochem

    (1998)
  • D. Cobrinik

    Pocket proteins and cell cycle control

    Oncogene

    (2005)
  • J.T. Coyle et al.

    Oxidative stress, glutamate, and neurodegenerative disorders

    Science

    (1993)
  • N. Dyson

    The regulation of E2F by pRB-family proteins

    Genes Dev

    (1998)
  • C. Evangelisti et al.

    Nuclear diacylglycerol kinases: emerging downstream regulators in cell signaling networks

    Histol Histopathol

    (2007)
  • C. Evangelisti et al.

    Identification of a functional nuclear export sequence in diacylglycerol kinase-ζ

    Cell Cycle

    (2010)
  • C. Evangelisti et al.

    Nuclear diacylglycerol kinase-ζ is a negative regulator of cell cycle progression in C2C12 mouse myoblasts

    FASEB J

    (2007)
  • K. Goto et al.

    A 104-kDa diacylglycerol kinase containing ankyrin-like repeats localizes in the cell nucleus

    Proc Natl Acad Sci U S A

    (1996)
  • D.R. Green et al.

    Cytoplasmic functions of the tumour suppressor p53

    Nature

    (2009)
  • K. Herrup et al.

    Cell cycle regulation in the postmitotic neuron: oxymoron or new biology?

    Nat Rev Neurosci

    (2007)
  • Cited by (0)

    1

    Present address: Department of Molecular Cancer Science, Yamagata University School of Medicine, Yamagata 990-9585, Japan.

    View full text