Nuclear inositide specific phospholipase C signalling  interactions and activity

Authors

  • Irene Faenza,

    1. Cell Signaling Laboratory, Department of Biomedical Science (DIBINEM), University of Bologna, Italy
    Search for more papers by this author
  • Roberta Fiume,

    1. Cell Signaling Laboratory, Department of Biomedical Science (DIBINEM), University of Bologna, Italy
    Search for more papers by this author
  • Manuela Piazzi,

    1. Cell Signaling Laboratory, Department of Biomedical Science (DIBINEM), University of Bologna, Italy
    2. Institute of Molecular Genetics, National Research Council of Italy (IGM-CNR), Bologna, Italy
    3. SC Laboratory of Musculoskeletal Cell Biology, Rizzoli Orthopedic Institute, Bologna, Italy
    Search for more papers by this author
  • Alessia Colantoni,

    1. Cell Signaling Laboratory, Department of Biomedical Science (DIBINEM), University of Bologna, Italy
    Search for more papers by this author
  • Lucio Cocco

    Corresponding author
    1. Cell Signaling Laboratory, Department of Biomedical Science (DIBINEM), University of Bologna, Italy
    • Correspondence

      L. Cocco, Cell Signaling Laboratory, Department of Biomedical Science (DIBINEM), University of Bologna, Bologna, Italy

      Fax: +39 051 251735

      Tel: +39 051 2091639

      E-mail: lucio.cocco@unibo.it

    Search for more papers by this author

Abstract

Evidence accumulated over the past 20 years has highlighted the presence of an autonomous nuclear inositol lipid metabolism, and suggests that lipid signalling molecules are important components of signalling pathways operating within the nucleus. Nuclear polyphosphoinositide (PI) signalling relies on the synthesis and metabolism of phosphatidylinositol 4,5-bisphosphate, which can modulate the activity of effector proteins and is a substrate of signalling enzymes. The regulation of the nuclear PI pool is totally independent from the plasma membrane counterpart, suggesting that the nucleus constitutes a functionally distinct compartment of inositol lipids metabolism. Among the nuclear enzymes involved in PI metabolism, inositide specific phospholipase C (PI-PLC) has been one of the most extensively studied. Several isoforms of PI-PLCs have been identified in the nucleus, namely PI-PLC-β1, γ1, δ1 and ζ; however, the β1 isozyme is the best characterized. In the present review, we focus on the signal transduction-related metabolism of nuclear PI-PLC and review the most convincing evidence for PI-PLC expression and activity being involved in differentiation and proliferation programmes in several cell systems. Moreover, nuclear PI-PLC is an intermediate effector and interactor for nuclear inositide signalling. The inositide cycle exists and shows a biological role inside the nucleus. It is an autonomous lipid-dependent signalling system, independently regulated with respect to the one at the plasma membrane counterpart, and is involved in cell cycle progression and differentiation.

Abbreviations
AML

acute myeloid leukaemia

DAG

diacylglycerol

DGK

diacylglycerol kinase

DM

myotonic dystrophy

EPO

erythropoietin

ERK

extracellular signal-regulated kinase

MDS

myelodysplastic syndrome

NLS

nuclear localization signal

PA

phosphatidic acid

PI

polyphosphoinositide(s)

PIP2

phosphatidylinositol 4,5-bisphosphate

PI-PLC

inositide specific phospholipase C

PKC

protein kinase C

siRNA

small interfering RNA

Introduction

The existence of a nuclear polyphosphoinositide (PI) metabolism is well documented and recognized [1-3]. The nucleus is a functional compartment for PI metabolism and is endowed with the enzymes involved in the classical PI cycle, such as kinases required for the synthesis of phosphatidylinositol 4,5-bisphosphate (PIP2), phosphoinositide-specific phospholipase C (PI-PLC) and diacylglycerol kinase (DGK). In addition, because of the role exerted in both cell growth and differentiation, there are hints that the nuclear PI metabolism could be implicated in neoplastic transformation. Most of the research on signal transduction pathways based on PI-PLC has been devoted to studying phenomena that take place at the plasma membrane. However, the work of several independent laboratories has consistently demonstrated that the phosphoinositide cycle is present in the cell nucleus, and may be important for various nuclear events such as mRNA export, DNA repair and gene transcription [4, 5]. Stemming from these achievements, it is now also clear that nuclear PI are involved in cell growth and differentiation. Although a large volume of data has accumulated, we still need to know more about the regulation, physical properties and interactions of nuclear PI. Lipid signalling molecules are essential components of the processes that allow one extracellular signal to be transduced inside the cell to the nucleus. Inside the nucleus, lipid second messengers elicit reactions capable of regulating gene transcription, DNA replication or repair and DNA cleavage, eventually resulting in cell growth, differentiation, apoptosis or many other cell functions. Inositol-containing phospholipids are the most intensively studied lipid second messengers. Although most of the research on signal transduction pathways based on PIs has been devoted to phenomena that take place at the cell periphery and plasma membrane, it has become clear that the nuclear PI cycle is regulated in a totally independent manner compared to that at the plasma membrane level. This suggests that nuclear inositol lipids themselves can modulate nuclear processes as important as transcription and pre-mRNA splicing, growth, proliferation, cell cycle regulation and differentiation. Among the enzymes of the cycle, nuclear PI-PLCβ1 appears to play a key role as a check point in the G1 phase of the cell cycle. Moreover, its activation and/or up-regulation is under the control of type 1 insulin-like growth factor receptor in both mouse fibroblast and myoblasts, suggesting that its signalling activity is essential for the normal behavior of the cell, at least in culture [6]. The recent discovery of a possible involvement of the deletion of the gene for PI-PLCβ1 in the progression of myelodysplastic syndrome (MDS) to acute myeloid leukaemia (AML) [7] in humans strengthens the contention that nuclear PI-PLC signalling is essential for physiogical processes such as cell growth and differentiation. Recent findings have highlighted the importance of understanding the nuclear protein network behind PI-PLC function and identifying the physically associated protein targets, thus providing insights into the interactors and downstream target effectors that further clarify its nuclear signalling cascade. In this review, we confine our focus to nuclear PI-PLC. We review the most updated literature on PI-PLCβ1, although it should be noted that PI-PLC γ1, δ1 and ζ are also located in the nucleus during some phases of the cell cycle.

Nuclear PI-PLCβ

It is known that PIs are present in the inner part of the cell nucleus and that their metabolism changes during cell growth and differentiation [8]. PI-PLCβ1 is highly expressed in some brain areas, even though PI-PLCβ1 and other mammalian PI-PLCβs are distributed in almost all tissues [9]. Inside the nucleus of several cell lines, PI-PLCβ1 is the main isoform [10, 11] because of its nuclear localization signal at the C-terminal tail [12]. Moreover, PI-PLCβ1 exists as alternatively spliced variants β1a and β1b, which differ in their carboxy-terminal residues [13]. The subcellular localization of the two PI-PLCβ1 proteins was examined in rat C6Bu-1 glioma cells using immunological techniques. Immunoblot analysis revealed that the two forms of PI-PLCβ1 were detectable in both cytosolic and nuclear fractions (Fig. 1). PI-PLCβ1a appeared to be located preferentially in the cytosol, whereas PI-PLCβ1b was found predominantly in the nuclei of C6Bu-1 cells. Immunocytochemical experiments confirmed the differential localization of the two PI-PLCβ1 species in C6Bu-1 cells. These results suggest that the two PI-PLCβ1 proteins may have different physiological roles in the cell [14]. Another study out a detailed analysis of the expression of these two proteins with respect to the regulation of the cell cycle in Friend erythroleukaemia cells. It was already known that nuclear PI-PLCβ1 is down-regulated when Friend erythroleukaemia cells treated with dimethylsulfoxide differentiate and synthesize γ-globin [15]. Matteucci et al. [16] decided to overexpress PI-PLCβ1a and 1b using both wild-type cDNA and a mutant in the COOH-terminal region lacking the ability to localize at the nucleus, aiming to discriminate between the nuclear and cytoplasmatic signalling activity of this PI-PLC during erythroid differentiation. Initially, it was shown that nuclear PI-PLCβ1 is involved in maintaining the undifferentiated state of Friend erythroleukaemia cells, possibly by opposing the inhibition of the cell cycle progression necessary for erythroid differentiation [16]. Subsequently, PI-PLCβ1 1a and 1b was overexpressed in the nuclear compartment and a mutant-PI-PLCβ1, which lacks the nuclear localization sequence, was overexpressed in the cytoplasmatic compartment to investigate how nuclear localization of PI-PLCβ1 acts on cell cycle progression and on the key signalling events that regulate passage through the G1 phase of Friend erythroleukaemia cells. The clones overexpressing the PI-PLCβ1 1a, 1b and the cytoplasmatic mutant-PI-PLCβ1 for the nuclear localization sequence, respectively, show that, although the cytoplasmatic mutant-PI-PLCβ1 has lost the capacity for nuclear localization, the clones overexpressing the two subtypes PI-PLCβ1 1a and 1b have the PLC in the nucleus, with the b form being entirely nuclear and the a form being distributed in both nucleus and cytoplasm, even though, in the latter compartment, it is expressed less [17]. The correlation of nuclear PI-PLCβ1 expression and activity with that of cyclin D3 was demonstrated for the first time. The evidence suggests that the overexpression of PI-PLCβ1 in the nucleus is directly responsible for the overexpression and activation of the cyclin D3-cdk4 complex, which is known to stimulate progression through G1 rather than promote the G1-S transition. Moreover, it has been reported that mitogen-activated protein kinasess, in particular Jun N-terminal kinase and extracellular signal-regulated kinase (ERK)1/2, play a critical role in transducing the mitogenic stimulus, and also that nuclear PI-PLCβ1 is activated during the G2/M phase concomitantly with the recruitment of protein kinase C (PKC)-α/PKCβI to the nuclear compartment[3]. By means of specific inhibitors of PKC-α or PKCβI and by small interfering RNA (siRNA) silencing, this research group [6] provided evidence indicating that, in the nucleus, PKC-α phosphorylates lamin B1 and thus enables cell cycle progression. The evidence for colocalization rested on two lines of evidence. First, immunocytochemical analysis by transmission electron microscopy showed that PI-PLCβ1 and lamin B1 were in close juxtaposition. In particular, in proliferating cells, both proteins predominantly decorate regions of euchromatin and, to a lesser extent, heterochromatin. In G2/M cells, both proteins, again localized in close juxtaposition, predominantly decorated chromosomal structures and, at a low frequency, spaces among chromosomes. Second, coimmunoprecipitation studies provided evidence indicating that lamin B1 was coimmunoprecipitated with PI-PLCβ1 and vice versa, arguing in favour of a physical interaction. Overall, the current data provide evidence that the PI-PLCβ1-mediated nuclear inositide signalling plays a functional role in the G2/M transition [18]. The nuclear localization of the PI-PLCβ1 forms is connected to specific targets during the differentiation of Friend erythroleukaemia cells. The expression of transcription factor p45/NF-E2, a prerequisite in the erythroid differentiation of Friend cells, is modulated by nuclear PI-PLCβ1 because overexpression of the nuclear subtypes, instead of mutant PI-PLCβ1 for the nuclear localization sequence, almost completely abolishes dimethylsulfoxide-induced p45/NF-E2 expression [19]. Other work has strengthened the contention that nuclear PI-PLCβ1 constitutes a key step in erythroid differentiation in vitro [20]. The novel finding that emerged from the microarray experiments was an up-modulation of the plasma membrane marker CD24 induced by nuclear PI-PLCβ1 but not by the cytoplasmatic mutant-PI-PLCβ1. CD24 is an antigen that plays a role in differentiation and haematopoiesis in the early stages and is also involved in the metastasizing ability of solid tumours and the increases in several leukaemias. These findings, obtained by combining microarrays, phenotypic analysis and siRNA technology, identify CD24 as an effector of the nuclear PI-PLCβ1 signalling pathway in murine erythroleukaemia cells and reinforce the notion that nuclear PI-PLCβ1 is a key player in erythroid differentiation as well as in other differentiation systems. During the differentiation of C2C12 mouse myoblasts stimulated with insulin, there is a dramatic increase in nuclear PI-PLCβ1, hinting at the fact that nuclear PI-PLCβ1 is an important step in several differentiation systems [21]. Skeletal muscle differentiation is characterized by terminal withdrawal from the cell cycle, the activation of muscle-specific gene expression and morphological changes, including myoblast alignment, elongation and fusion of mononucleated myotubes. These events are coordinated by a family of four muscle-specific basic helix-loop-helix transcription factors: MyoD1, Myf5, myogenin and Mrf4, termed the muscle regulatory factors [22]. When the expression of PI-PLCβ1 is forced in the cytoplasm by means of a cytoplasmatic PI-PLCβ1 mutant, the differentiation of C2C12 myoblasts is inhibited [21]. The C2C12 cell line contains also PI-PLCβ3, and β4. Inside the nucleus, PI-PLCβ1a and 1b splicing variants are present, with the 1b form being totally nuclear even though PI-PLCβ3 is detectable to a low extent. During myotube differentiation, both PI-PLCβ1 splicing variants increase; at the same time, PI-PLCβ4 decreases in both the cytoplasmatic and perinuclear compartments [23]. More detailed investigations showed that nuclear PI-PLCβ1 activates cyclin D3 promoter during the differentiation process. Cyclin D3 expression plays a critical role in the Myo-D-mediated arrest of the cell cycle, which precedes myoblast differentiation. Therefore, PI-PLCβ1 appears to be a key player in the skeletal muscle differentiation programme with respect to the regulation of cyclin D3 gene activity. Further studies showed that PI-PLCβ1 catalytic activity is essential for the transduction of the differentiating signals elicited by insulin, targeting cyclin D3 promoter through the activation of c-jun/AP1 transcription factor [24]. Recently, it was shown that forced expression of cyclin D3 is able to promote and correct the differentiation process, leading to a recovery of myogenin and desmin levels in myotonic dystrophy (DM). The lack of elevation of cyclin D3 in DM1 differentiating cells appears to be a critical event leading to impaired myoblast fusion [25, 26]. DM is the most prevalent form of muscular dystrophy in adults. DM type 1 (DM1) and type 2 (DM2) are dominantly inherited multisystem disorders. The role of cyclin D3 is thus of particular interest as a result of the close correlation between the levels of PI-PLCβ1 expression and that of cyclin D3 during the process of muscle differentiation. Given the data obtained in C2C12 cells, as well as the available data concerning the decrease in expression of cyclin D3 in DM1, Faenza et al. [27] determined whether a modulation of the expression of PI-PLCβ1 could enhance differentiation in DM1 and DM2. Becauase cyclin D3 expression plays a critical role in the Myo-D-mediated arrest of the cell cycle preceding myoblast differentiation, PI-PLCβ1 appears to affect the skeletal muscle differentiation programme. Our research group demonstrated that a modulation of PI-PLCβ1 expression increases cyclin D3 and myogenin expression, rescuing the early steps of the differentiation programme in DM1 and DM2 cells [27]. The myogenesis of DM cells was characterized by a strong decrease in the expression of PI-PLCβ1. This decrease in PI-PLCβ1 protein expression did not correspond with a decrease in gene expression as evaluated by real-time PCR, thus leading us to hypothesize that the lack of increased expression was the result of a problem at the translational level or an alteration in one of the pathways that leads to the processing of the protein, which is typical of DMs. The accumulation of aberrant RNA in the nucleus can lead to a blockage of the normal processes involved in the translation [26]. Cyclin D3 levels are increased in differentiated cells from control participants but not in DM1 and DM2 cells. By contrast, myogenin was demonstrated to increase expression during differentiation, although at very low levels. Overexpression of PI-PLCβ1 1a and 1b in differentiated cells from control participants and patients with DM1 or DM2 showed that both isoforms were required for the regulation of cyclin D3 and myogenin expression [27]. It was also shown that, even in human myoblasts in the process of differentiation induced by insulin, the expression of cyclin D3 is regulated by a PI-PLCβ1-dependent signal transduction pathway. Activation of cyclin D3 by PI-PLCβ1 links the differentiation programme of MEL cells and C2C12 cells. Nuclear PI-PLCβ1 activates cyclin D3 in both systems. Indeed, cyclin D3 has opposite effects in these cell types, promoting the differentiation of myoblasts to myotubes in the case of C2C12 cells, and stimulating the progression through the G1 phase of the cell cycle in the case of MEL cells [28, 29]. Nuclear PI-PLCβ1 is also required for 3T3-L1 adipocyte differentiation and regulates expression of cyclin D3 in this differentiating system. PI-PLCβ1 acts in two phases during 3T3-L1 adipocyte differentiation. The first takes place within 5 min of treatment with differentiation media and does not require new PI-PLCβ1 to enter the nucleus, being regulated by phosphorylated ERK and PKCα. The second phase occurs from day 2 of differentiation, requires new PI-PLCβ1 to enter the nucleus and is not regulated by phosphorylated ERK and PKCα. In sum, nuclear PI-PLCβ1 activity is up-regulated during 3T3-L1 differentiation. Over-expression of the cytoplasmatic PI-PLC mutants, which lack the ERK phosphorylation site or nuclear localization sequence, showed that both phases of PI-PLCβ1 activity are required for terminal differentiation to occur. Inhibition of PI-PLCβ1 activity prevents the up-regulation of cyclin D3 and cdk4 protein, suggesting that PI-PLCβ1 plays a role in the control of the cell cycle during differentiation [30]. New data have identified a nuclear role for PLC in insulin secretion. In MIN6 β cells, PI-PLCβ1 was localized in both nucleus and cytoplasm: PLCδ4 in the nucleus and PLCγ1 in the cytoplasm. By silencing each isoform, it was observed that they all affected glucose-induced insulin release both at basal and high glucose concentrations [31]. These findings highlight a novel pathway by which nuclear PI-PLCs affect insulin secretion and identify PPARγ as a novel molecular target of nuclear PI-PLCs [31].

Figure 1.

Schematic diagram depicting the targets of nuclear PI-PLCβ1.

The modulation of PI-PLCβ1 at a nuclear level is implicated in the progression of MDS to AML. MDS, a heterogeneous group of bone marrow disorders displaying impaired stem cell differentiation and cytopenia, has a variable risk of evolution to AML. Recently, it was shown that azacitidine specifically targets PI-PLCβ1 [32]. PI-PLCβ1 promoter methylation and gene expression have been analyzed in high-risk MDS patients during azacitidine administration and compared with the expression observed in patients treated with only best supportive care, as well as that observed in healthy subjects [32]. It is of clinical interest that promoter methylation and gene expression had an opposite trend, in that reduced promoter methylation and increased PI-PLCβ1 expression anticipate a positive clinical outcome. These changes were detectable before clinical improvement. By contrast, unresponsive patients did not show these early changes. One possible explanation is that the altered expression of nuclear PI-PLCβ1 and activated (phosphorylated) Akt could lead to a deregulation of cell cycle processes, therefore negatively influencing the apoptotic processes and affecting the survival of primary MDS cells [33]. After azacitidine treatment in high-risk MDS patients, an increase in PI-PLCβ1 levels is followed by a reduction in activated Akt levels, thus indicating that PI-PLCβ1 and Akt could play opposite roles [34]. PI-PLCβ1 promoter hypermethylation has been associated with the progression of high-risk MDS into AML and, in addition, the effect of erythropoietin (EPO) treatment on Akt activation and PI-PLCβ1 expression strengthens the contention that correct nuclear lipid signalling is essential for physiological processes such as cell growth and differentiation in MDS [35]. EPO responder patients showed an activation of Akt, as expected, whereas the same cases displayed a decrease of PI-PLCβ1. Interestingly, the decrease of PI-PLCβ1 was statistically significant after 4–6 months of therapy, which is consistent with previous findings showing that, after an early transient increase, PI-PLCβ1 is down-regulated in primary human erythroblasts treated with EPO for up to 96 h [36], therefore suggesting that PI-PLCβ1 could be required at the beginning of erythroid differentiation but is dispensable, if not inhibitory, at later stages. At the same time, the Akt phosphorylation that we detected in EPO responder cases is in agreement with the findings obtained other previous in vitro studies showing that EPO can induce a nuclear translocation of active Akt, which is required for erythroid differentiation [36]. Taken together, these results not only confirm the inverse correlation between PI-PLCβ1 and Akt, but also hint at a role for PI-PLCβ1 as a negative regulator of erythroid differentiation, as previously hypothesized as a result of in vitro studies conducted in erythroleukaemia cells [19].

Nuclear PI-PLC interactors

The complexity of nuclear inositide signalling suggests that it is likely to be involved in the regulation of most if not all nuclear processes. At this point, there was a need to identify the specific nuclear proteins that interact with and are regulated by phosphoinositides. These are not simple tasks, although the fact that nuclear phosphoinositides control gene expression, mRNA stability and export and chromatin remodelling, and also that specific regulatory factors are required that are only utilized within the nuclear compartment, means that the rewards are likely to be great. Co-localization studies show that, in C2C12 cells, nuclear PI-PLCβ1 is bound to DGK-ζ, in nuclear speckles [38]. Similar to PI-PLCβ1, nuclear DGK-ζ also increases during myoblast differentiation, and impairment of DGK-ζ up-regulation markedly inhibits differentiation. If the function of nuclear diacylglycerol (DAG) is to attract DAG-dependent PKC isoforms within the nucleus, then a mechanism should exist to terminate the signal. This role could be fulfilled by DGK, the enzyme that phosphorylates DAG, yielding phosphatidic acid (PA). The fact that both isolated nuclear envelopes and nuclei produced in vitro radiolabelled PA, suggested the presence of DGK at the nuclear level. Several independent groups have demonstrated the existence of DGK isoforms within the nucleus and have shown that these enzymes are indeed involved in controlling nuclear DAG mass upon agonist stimulation [3-39]. The physical binding of these two enzymes could be functionally significant. PI-PLCβ1 produces DAG, which is the substrate for DGK. DGK produces PA, which, in turn, has been shown to activate PI-PLCβ1 by binding to its C-terminal domain. It should be noted that DGK-ζ interacts with, and is activated by the Rb family proteins, pRb, p107 and p130 [40-42]. Another target of nuclear PI-PLCβ1 in Friend erythroleukaemia cells has been identified using a proteomic approach. In this study [43], the aim was to systematically analyze proteins whose expression levels are constitutively activated and to gain a global view of the molecular regulatory network that operates in Felc overexpressing nuclear PI-PLCβ1. The proteomic approach allowed the display of several proteins expressed in the cellular line of interest, as well as the analysis and comparison of maps of protein expression of wild-type cells and cells in which PI-PLCβ1 was amplified. To identify novel downstream effectors of nuclear PI-PLCβ1-dependent signalling in Friend erythroleukaemia cells, high-resolution 2-DE-based proteomic analysis was performed. Using this approach, SRp20, a member of the highly conserved SR family of splicing regulators, was found to be down-regulated in cells overexpressing nuclear PI-PLCβ1 compared to wild-type cells. Moreover, nuclear PI-PLCβ1 was shown to be bound to the SRp20 splicing factor. Indeed, by immunoprecipitation and subcellular fractioning, it was demonstrated that endogenous PI-PLCβ1 and SRp20 physically interact in the nucleus. The existence of a PI-PLCβ1-specific target was demonstrated, namely the splicing factor SRp20, whose expression is specifically down-regulated by the nuclear signalling evoked by PI-PLCβ1 [43]. All together, these data show that it is even more urgent to understand the effectors and interactors of PI-PLCβ1 to enable the precise targeting of the mechanisms implicated in these diseases. Another interesting interactor of PI-PLCβ1 is α-synuclein, which is a small protein that is highly expressed in brain tissues and is also the major component of neurodegenerative plaques and mutations associated with sporadic forms of familial Parkinson's disease [44]. α-Synuclein binds strongly to PI-PLCβ1 and promotes the release of Ca2+ in cells. It has been shown that the expression of α-synuclein increases the cellular level of PI-PLCβ1 in two neuronal cell lines. α-Synuclein can increase the cellular level of PI-PLCβ by protecting against degradation from enzymes such as μ-calpain [44]. Very recently, a systematic analysis of the PI-PLCβ1 protein interactome was undertaken [45], which aimed to characterize both the mechanism of nuclear localization and the molecular function of nuclear PI-PLCβ1 by identifying its interactome in Friend's erythroleukaemia isolated nuclei, utilizing a procedure that coupled immuno-affinity purification with tandem MS analysis. Accordingly, 160 proteins were demonstrated to be in association with PI-PLCβ1b. The study focused on PI-PLCβ1b interactome because, as noted above, two isoforms of inositide-dependent phospholipase Cβ1 are present within the nucleus but, in contrast to PI-PLCβ1a, the vast majority in the nuclear compartment is PI-PLCβ1b. PI-PLCβ1b was shown to interact with Phb2, a protein that shuttles between the nucleus and mitochondria, and with 12 importins, Kpna2, Kpna4 and Kpnb1, which are required for nucleus-cytoplasmatic transport of a certain set of proteins. Similar to MEL/PI-PLCβ1b cells, PI-PLCβ1b was also found in complex with Srsf3, Lmnb1, Phb2, Kpnb1 and Kpna2 in nuclei of the murine pro-B lymphoid Ba/F3 cells overexpressing PI-PLCβ1b. PI-PLCβ1b associated with proteins involved in cellular metabolic processes (31%), gene expression (19%), transport (12%), developmental processes (8%), translation (8%), RNA splicing and processing (7%), response to oxidative stress (5%), and regulation of apoptosis (4%). A minor percentage of proteins was related to cell cycle (2%), cell proliferation (2%), regulation of (myeloid) differentiation (1%) and cell fate determination (1%). Subdividing the gene expression categories revealed PI-PLCβ1b to be specifically associated with chromatin organization proteins, chromatin silencing by methylation and epigenetic regulation. One of the more interesting findings is the association of PI-PLCβ1b with the classical import proteins, Kpna2, Kpna4, Kpnab1, Ran and Rangap1. Both PI-PLCβ1a and PI-PLCβ1b present with a bipartite consensus sequence for nuclear localization (K1055, K1056 separated from K1069, K1071, by a linker of 12 amino acids); however, a cluster of three lysines (K1056, K1063 and K1070) was demonstrated to be essential for nuclear translocation. It was therefore suggested that these three lysines were critical sites for establishing interactions that retain PI-PLCβ1 within the nucleus. A recent study [46] reported that PKC phosphorylation on S887 (in the C-terminal region of both PI-PLCβ1a and 1b) also regulated the subcellular distribution of PI-PLCβ1 because the lack of phosphorylation kept the enzyme located within the nucleus. The sequence of 75 amino acids exclusive to PI-PLC β1a also contains a putative nuclear export sequence, which likely explains the predominant presence of PI-PLC β1a in the cytoplasm [14]. A study by Montana et al. [9] describes a complete and detailed neuroanatomical distribution map of the PI-PLCβ1 isoform along the adult rat neuraxis, and defines the phenotype of cells expressing PI-PLCβ1, along with its subcellular localization in cortical neurones as assessed by double-immunofluorescence staining and confocal laser scanning. Montana et al. [9] also describe the relationships of PI-PLCβ1-positive cells with GABAergic neurones and glial fibrillary acidic protein-immunopositive radial glia-like processes of the spinal cord white matter. Moreover, histological and biochemical evidence is provided for the presence of PI-PLCβ1 in the nuclear compartment of rat cortical neurones. The expression levels of the splice variants PI-PLCβ1a and PI-PLCβ1b were measured in both nuclei and cytoplasmatic fractions, using an antibody raised against a peptide mapping at the N-terminus of PI-PLCβ1 (common to both PI-PLCβ1a and PI-PLCβ1b variants). The comparative study between PI-PLCβ1a/PI-PLCβ1b ratios revealed that it was significantly higher in cytoplasmatic than in nuclei and plasmamembrane fractions. In intact nuclei isolated from homogenates of the cerebral cortex, a high overlap was observed between PI-PLCβ1-immunostaining and the signals provided by the markers of the nuclear speckles NeuN/Fox3 and SC-35. A recent study [47] described a proteomics approach for identifying potential effectors of nuclear PI-PLCβ1-dependent signalling during insulin-stimulated myogenic differentiation. A search was made of phosphosubstrates of cPKC in nuclei of insulin-stimulated C2C12 cells. Immunoprecipitation experiments on nuclei of C2C12 myoblasts were performed with a phospho(Ser)-PKC substrate antibody. The immunoprecipitated samples were electrophoretically resolved by SDS/PAGE and Coomassie-stained. Bands that appeared to be visibly present in the control and insulin treated lanes were excised from the gel, digested with trypsin and subjected to LC-MS/MS. Using a co-immunoprecipitation and MS based methodology, a new protein was identified (i.e. eEF1A), which interacted with and was phosphorylated by a cPKC, indicating that PKCβI can directly phosphorylate eEF1A on Ser53 [47]. Taken together, these data point to a diverse set of nuclear activities regulated, in part, by the presence of PIs within nuclear speckles. The global significance of nuclear PI–protein interactions, however, is still poorly understood, largely as a result of the small number of known nuclear effector proteins. However, in cases where this has been investigated in detail, it appears that such interactions have profound physiological effects. The identification of other nuclear PI effectors is therefore likely to shed more light on these and, importantly, other nuclear PI functions. To address this, Lewis et al. [48] developed a proteomic approach to enrich for (and identify) potential nuclear PI-binding proteins by nano LC-MS/MS. Neomycin extraction was used as a means of reducing sample complexity and also to enrich for a pool of potential PI-binding proteins devoid of bound PI. Several previous observations suggested that incubating isolated nuclei in the presence of excess neomycin would produce an extract that is relatively less complex than a total nuclear lysate and, importantly, also is enriched in PI-binding proteins whose PI-binding domain(s) are devoid of any endogenous nuclear derived PI. In total, 349 nuclear proteins were identified, 48% of which harbour PI binding domains. Clustering analysis of these proteins revealed over-represented functions related to RNA splicing, chromatin assembly and DNA topological change. Furthermore, this method was validated first by identifying a subset of proteins displaced by neomycin as PIP2 interacting proteins using quantitative lipid pull downs and, second, by biochemically characterizing the interaction of PIP2 with DNA topoisomerase IIα, an enzyme with hitherto indirect links with nuclear PI metabolism. This proteomics approach provides compelling evidence suggesting that nuclear PIs interact with a wide range of nuclear proteins regulating numerous nuclear functions.

Other nuclear phospholipases C

PI-PLCγ1 was also described in the nucleus. It has been reported that the SH3 domain of PI-PLCγ1 can act as a guanine nucleotide exchange factor for the brain-specific nuclear GTPase phosphoinositide 3-kinase enhancer and for dynamin-1 [49, 50]. The ability of PI-PLCγ1 of aiding phosphoinositide 3-kinase enhancer does not depend on its phospholipase activity and underlies several features of PI-PLCγ1 cellular function, including the regulation of proliferation and survival [51]. Although PI-PLCδ1 is generally found in the cytoplasm of quiescent cells, it localizes in nuclear structures in the G1/S boundary of the cell cycle. PI-PLCδ1 has both nuclear export and import sequences that contribute to its shuttling between the cytoplasm and nucleus [52]. Stallings et al. [53] demonstrated that siRNA silencing of PI-PLCδ1 increases the level of cyclin E, a key regulator of the G1/S boundary, alters S-phase progression, and inhibits cell proliferation. Moreover, transient expression of PI-PLCδ1 suppresses the expression of cyclin E at the G1/S boundary, hinting at PI-PLCδ1 being an important player in cell cycle progression. Okada et al. [54] reported that PI-PLCδ1 shuttles between the cytoplasm and nucleus and also that a massive increase in the intracellular Ca2+ concentration causes the nuclear import of PI-PLCδ1 through a Ca2+-dependent direct interaction with importin β1. Moreover, they examined the distribution of PI-PLCδ1 using primary cultures of rat hippocampal neurones. Treatment of 7DIV neurones with ionomycin or thapsigargin caused the nuclear localization of PI-PLCδ1. Similar results were obtained with neurones treated with glutamate, suggesting that the nuclear localization of PI-PLCδ1 plays a role in the excitotoxicity associated with ischaemic stress. Generally, cells undergoing ischaemic or hypoxic cell death show nuclear shrinkage. They demonstrate that a massive influx of Ca2+ gave similar results [54]. Furthermore, overexpression of GFP-PLCδ1 facilitated ionomycin-induced nuclear shrinkage in embryonic fibroblasts derived from PI-PLCδ1 gene-knockout mice. Therefore, nuclear translocation and the PI-PLC activity of PI-PLCδ1 may regulate the nuclear shape by controlling the nuclear scaffold during stress-induced cell death as a result of high levels of Ca2+. It was concluded that overstimulation of the glutamate signalling pathway and treatment with reagents that cause massive increases in the intracellular Ca2+ concentration results in the nuclear import of PI-PLCδ1 in primary cultured rat hippocampal neurones similar to the responses of other cell types to Ca2+ flux, such as a rat adrenal pheochromocytoma cell line (PC12) and Madin–Darby canine kidney cells,. Moreover, overexpression of PI-PLCδ1 with enzymatic activity (but not that of a PI-PLC activity-null mutant) stimulated ionomycin-induced nuclear shrinkage. These findings suggest that PI-PLCδ1 enhances the nuclear shrinkage-associated cell death in a Ca2+-dependent manner. Nuclear localization of sperm-specific PI-PLC-ζ isozyme has also been demonstrated. Originally, PI-PLC-ζ was discovered in soluble sperm extracts from mouse, hamster and pig and, subsequently, PI-PLC-ζ was detected in soluble cytosolic fractions of sperm extracts, as well as in extracts of sperm heads [55]. PI-PLC-ζ triggers Ca2+-oscillations in the egg that are essential for egg activation, fertilization and embryo development. It has been demonstrated that these oscillations are cell cycle-dependent and are associated with PI-PLC-ζ distribution. When the oscillations stop, PI-PLC-ζ is found to be distributed in the pronuclei, whereas during the first mitosis, when the oscillations restart, PI-PLC-ζ returns to the cytosol [56, 57]. Moreover, if exogenous PI-PLC-ζ is injected into an embryo at a later stage of development, it continues to undergo nuclear sequestration. PI-PLC-ζ nuclear sequestration appears to be a result of the presence of a putative nuclear localization signal, which promotes its accumulation in the pronuclei [58]. Moreover, Cooney et al. [59] demonstrated that bovine PI-PLC-ζ failed to accumulate in the pronucleus of bovine or murine zygotes, despite possessing a putative nuclear localization signal. Conversely, murine PI-PLC-ζ accumulated in the pronucleus of both murine and bovine zygotes. Cooney et al. [59] demonstrated that murine PI-PLC-ζ and bovine PI-PLC-ζ possess species-specific differences in activity suggesting potential differences in the mode of action of the protein between the two species [56-58]. Ca2+ signalling is important for controlling gene transcription. Changes in the cytosolic Ca2+ concentration may promote the migration of transcription factors or transcriptional regulators to the nucleus. Changes of the nucleoplasmic Ca2+ concentration can also directly regulate gene expression. The idea of an independent Ca2+ signalling system in the nucleus is most intriguing, favoured by the existence of intranuclear ryanodine receptors, inositol trisphosphate receptors and the enzymatic machinery necessary for the synthesis of the messengers involved in these transduction systems. The increase of Ca2+ concentration also induces shuttling of PI-PLCδ1 to the nucleus, where it can induce the production of inositol trisphosphate and diacylglycerol, thus affecting several nuclear functions [60]. Vanzela et al. [61] disrupted one of three putative phosphatidylinositol phospholipase C genes of Aspergillus nidulans and studied its effect on carbon source sensing linked to vegetative mitotic nuclear division. It was demonstrated that glucose does not affect nuclear division rates during early vegetative conidial germination in either the wild-type or the plcA-deficient mutant. Only after 8 h of cultivation on glucose did the mutant strain show any decrease in nuclear duplication. However, decreased nuclear division rates were observed in the wild-type when cultivated in media amended with polypectate, whereas the plcA-deficient mutant did not show slow nuclear duplication rates when grown on this carbon source, even though it requires the induction and secretion of multiple pectinolytic enzymes to be metabolized. Thus, plcA appears to be directly linked to high-molecular-weight carbon source sensing [61].

Conclusions

We have focused our attention on PI-PLCβ1 because of its predominant presence in the nucleus compared to other PI-PLC isoforms, as well as the extensive investigations carried out on this signalling molecule. In addition to the role of PI-PLCβ1 in the regulation of nuclear inositol lipid signalling, it could also be involved in the organization of nuclear domains. In the nucleus, PI-PLCβ1 interacts with a number of proteins involved in nuclear import, differentiation, cell cycle progression, mRNA processing and apoptosis. Such data give significant insights into the molecular environment that surrounds PI-PLCβ1 and provide evidence for the interaction of nuclear PI-PLCβ1 with a number of proteins involved in nuclear processing. These discoveries also give new insights into the possible mechanisms of nuclear trafficking and the functioning of this critical signalling molecule. In sum, the data reviewed here show that nuclear inositide signalling has a functional significance that is still not fully understood. Therefore, further investigations, conducted both under physiological conditions and in diseases, mainly in humans, are necessary and could give rise to new and unexpected findings that are important for our understanding of the pathophysiology of a number of diseases.

Acknowledgements

This work was supported by Italian MIUR-FIRB Human Proteomic Network and MIUR-FIRB 2010.

Ancillary