- Email: [email protected]
Significance of subnuclear localization of key players of inositol lipid cycle
ARTICLE IN PRESS
Advan. Enzyme Regul. 44 (2004) 51–60
Significance of subnuclear localization of key players of inositol lipid cycle Lucio Coccoa,*, Lucia Manzolia, Ottavio Barnabeib, Alberto M. Martellia,c,d a
Cellular Signaling Laboratory, Department of Anatomical Sciences, University of Bologna, Italy b Department of Biology, University of Bologna, Italy c ITOI-CNR, Bologna Unit, c/o IOR, Bologna, Italy d School of Pharmacy, University of Bologna, Italy
Introduction Among phospholipids, polyphosphoinositides are key players in a number of signal transducing pathways, even though they are quantitatively minor constituents of cell membranes. Indeed, phosphatidylinositol (PI) typically comprises approximately 10% of total membrane lipid, suggesting that these lipids are not likely to play a major structural role. The discovery of the second messenger role of inositol (1,4,5) trisphosphate (IP3) and diacylglycerol (DAG) has brought phosphoinositide signaling to the very forefront of biological and biomedical research. Signaling through various phosphoinositides, including 3-phosphorylated inositol lipids, has been shown to mediate cell growth and differentiation, apoptosis, intracellular vesicle trafficking, ion channel activation, insulin action, cytoskeletal changes, and motility (Toker, 2002). Therefore, at present it is difficult to think of a cell response which does not utilize some form of inositide-dependent signal transduction. Moreover, it is now clear that phosphoinositide signaling disorders are implicated in the pathogenesis of various human diseases ranging from cancer to type 2 diabetes (Pendaries et al., 2003). The generally accepted model of phosphoinositide signaling involves the generation of lipid second messengers in response to stimuli in a receptor-mediated manner at the plasma membrane. However, the existence of a nuclear PI metabolism independent from that occurring elsewhere in the cell is now widely recognized (Cocco et al., 2001a), suggesting that the nucleus constitutes both a functional and a *Corresponding author. Cellular Signaling Laboratory, Department of Anatomical Sciences, University of Bologna, via Irnerio 48, Bologna 40126, Italy. Tel.: +39-051-244467; fax: +39-051-251735. E-mail address: [email protected] (L. Cocco). 0065-2571/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.advenzreg.2003.11.009
ARTICLE IN PRESS 52
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
distinct compartment for PI metabolism (Martelli et al., 2003). Indeed, it has been demonstrated that nuclei contain many of the enzymes involved in the classical PI cycle, including kinases required for the synthesis of phosphatidylinositol (4,5)bisphosphate (PIP2), phosphoinositide-specific phospholipase C (PLC), and diacylglycerol kinase (DGK) (D’Santos et al., 1998; Martelli et al., 1999, 2000; Cocco et al., 2001b; Irvine, 2003). More importantly, specific changes in the nuclear PI metabolism have been implicated in cell growth, differentiation, and neoplastic transformation (Martelli et al., 2002a). In this review, we shall highlight the major discoveries in the field of subnuclear localization of the enzymes of the inositol lipid metabolism and pay particular attention to progress made by our group in the co-localization of both enzymes and substrates of this signaling system.
DAG in the nucleus Lipid second messengers are essential intermediates linking extracellular stimuli via receptor activation to the required cellular response (Divecha and Irvine, 1995; Martin, 1998). These messengers are generated along very complex signaling pathways. DAG is a key lipid second messenger which can derive from either phosphoinositides or phosphatidylcholine (Wakelam, 1998). DAG activates some of the members of the protein kinase C (PKC) family and thereby the signaling pathways downstream of PKC. The DAG-dependent PKC isozymes include conventional PKC-a, bI ; bII ; g; and novel PKC d; e; Z; y; m: In contrast, atypical PKC-z and i= l do not require DAG for their activation (Ron and Kazanietz, 1999). However, other molecular targets of DAG have been identified such as a- and b-chimaerins, the guanine nucleotide-exchange factor vav and guanyl nucleotide-exchange factors for Ras and Rap (van Blitterswijk and Houssa, 2000; Bregoli et al., 2001). These findings indicate that DAG might also be involved in controlling the Ras and Rho family of proteins. In addition to cell surface receptor-mediated transmembrane signaling pathways involving phospholipid metabolism, a number of reports have convincingly demonstrated that the nucleus must also be considered as a site where biologically active lipid second messengers are generated (D’Santos et al., 1998; Divecha et al., 2000; Cocco et al., 2001a, b; Martelli et al., 2001; Irvine, 2003). The control of steady-state cellular levels of DAG is crucial to cellular physiology. DAG signaling must be short-lived because persistently high levels of DAG induce malignant transformation. The transforming activity of DAG has been attributed most often to persistent activation of PKC isoforms that are clearly involved in tumorigenesis (Ron and Kazanietz, 1999). DGKs, whose activity increases upon cell stimulation by a variety of agonists, metabolize DAG by converting it to phosphatidic acid (PA). Since they can attenuate local accumulation of signaling DAG, DGKs play a pivotal role in many biological responses such as cell proliferation, differentiation, survival, and apoptosis. Moreover, PA itself can function as a second messenger. Up to now, nine mammalian DGK isoforms have been identified which differ in their tissue
ARTICLE IN PRESS L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
53
expression and structural domains (Goto and Kondo, 1999; van Blitterswijk and Houssa, 1999; Topham and Prescott, 1999; van Blitterswijk and Houssa, 2000; Kanoh et al., 2002). DAG is also produced in the nucleus. The levels of nuclear DAG fluctuate during the cell cycle progression, suggesting that it has important regulatory roles. Most likely, nuclear DAG serves as a chemoattractant for some isoforms of PKC that migrate to the nucleus in response to a variety of agonists (D’Santos et al., 1998; Divecha et al., 2000; Cocco et al., 2001a, b; Martelli et al., 2001; Irvine, 2003; Neri et al., 2002). Several independent laboratories have indicated that DGK isoforms are present in the nucleus where they may be involved in regulating the amount of DAG (Martelli et al., 2002b). In some cases, the activity of nuclear DGKs has been demonstrated to be very critical for the control of cell proliferation (Topham et al., 1998; Martelli et al., 2000; Bregoli et al., 2002). When nuclear DGK isoforms were visualized by means of immunostaining or green fluorescent protein (GFP) technology, it became apparent that they were not distributed in a diffuse manner but rather concentrated in defined domains (Bregoli et al., 2001; Topham et al., 1998; Shirai et al., 2000). It has been proposed, but never proved, that DGK-y is localized in the speckle domains of the nucleus (van Blitterswijk and Houssa, 2000). The speckle domains are nuclear subcompartments enriched in small ribonucleoprotein particles and various splicing factors (Dundr and Misteli, 2001). Previous results have highlighted that nuclear speckles contain elements of the phosphoinositide cycle, including various types of phosphatidylinositol kinase (PIPK), PIP2, and phosphoinositide 3-kinase (PI3K) (Boronenkov et al., 1998; Osborne et al., 2001; Didichenko and Thelen, 2001).
Nuclear DGK-h and co-localization at the speckle domains with PLCb1, PIP2, PIPK1a and SC-35 splicing factor Albeit the hints for the presence in discrete nuclear regions of several players, an accurate and combined cytochemical and immunochemical analysis was still lacking, mainly regarding DGKs. A recent paper from our laboratory gave new insight in this issue (Tabellini et al., 2003). Here we summarize the main findings. It has been shown that DKG-y is expressed in several cell types such as PC12, HeLa, MDA-MB453, and MCF-7 cells. A series of experiments have clarified the subnuclear localization of DKG-y. In immunostained MDA-MB-453 cells for SC-35, PIP2, and DKG-y, a confocal laser scanning microscope (CLSM) analysis showed that besides SC-35 being mainly located in discrete nonnucleolar domains, corresponding to speckles, immunostaining with monoclonal antibody to PIP2 resulted in a similar pattern. Boronenkov et al. (1998) and Osborne et al. (2001) hinted at a speckle location of PIP2. This has been confirmed in MDA-MB-453, since double immunofluorescence staining with antibodies to SC-35 (an IgG) and PIP2 (an IgM) provided evidence that the two molecules co-localized at the levels of the speckles (Tabellini et al., 2003). Moreover, the authors showed by double immunofluorescence staining with antibodies to DGK-y (an IgG) and PIP2 that the two molecules co-localized at the level of nuclear foci, corresponding to speckles.
ARTICLE IN PRESS 54
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
The co-localization of DGK-y with PIP2 in nuclear foci was also detected in MCF-7, PC12, and HeLa cells. Overall, DGK-y immunoreactivity was mainly found in the nucleus. HeLa and PC12 cells showed some DGK-y staining in the cytoplasm, whereas in MDA-MB-453 and MCF-7 cells cytoplasmic positivity was almost undetectable. Cytoplasmic PIP2 immunostaining was mostly seen in HeLa, MDAMB-453, and PC12 cells. An interesting feature reported in this paper is the quantitative analysis of the results provided by CLSM using both Pearson’s correlation and overlap coefficient. Pearson’s correlation provides information about the similarity of shape between images without regard to the average intensity of the signals (spatial co-localization). It is a value computed to be between 1 and 1. The overlap coefficient, simultaneously used to describe co-localization, does not perform any pixel averaging functions, so correlations are returned as values between 0 and 1. This method is not sensitive to intensity variations in the image analysis and this is especially important when considering issues typical to fluorescence imaging such as sample photo-bleaching or different setting of the detectors. Not surprisingly all the values about the co-immunostaining of SC-35/ PIP2 and DGK-y/PIP2 were higher than 74%, indicating a remarkable level of colocalization. A further confirmation of the co-localization comes from the data obtained with a amanitin. Indeed, it is known that cell treatment with the transcriptional inhibitor a amanitin at concentrations that specifically inhibit RNA polymerase II causes reorganization of nuclear speckle domains into few and/or large foci (CarmoFonseca et al., 1992). In HeLa cells a amanitin caused the expected changes in SC-35 immunofluorescence staining, in that speckles were larger and less numerous. The paper (Tabellini et al., 2003) reports that also intranuclear distribution of DGK-y changed in response to a amanitin, since the foci became less numerous, even though their size did not increase as much as for SC-35 staining. The authors have also established whether or not the level of co-localization of SC-35 with PIP2 and of DGK-y with PIP2 was maintained after treatment with a amanitin. The level of colocalization of SC-35/PIP2 was maintained in a amanitin-treated cells. In contrast, double immunofluorescence labeling for DGK-y/PIP2 revealed that, after a amanitin exposure, the extent of co-localization was not as high as in untreated samples. The data discussed above could suggest that the players of the nuclear phosphoinositide cycle locate in discrete domains which are not chromatin-related. Indeed, it has been shown that RNA, but not DNA, is essential for the association of PIP2 with nuclear speckles (Osborne et al., 2001). In MDA-MB-453 cells DNase I treatment did not affect the intranuclear distribution of DGK-y, which colocalizes with PIP2, whilst RNase A treatment completely removed intranuclear immunoreactivity for DGK-y. These results imply that RNA, but not DNA, is essential for the association of DGK-y with speckles. Therefore, the analysis of the subcellular distribution of DGK-y is a good tool to assess the whole phosphoinositide cycle and related enzymes. By Western blot analysis Tabellini et al. (2003) could demonstrate that in MDA-MB-453 cells DGK-y was enriched in the nuclear matrix fraction which was obtained by DNase I digestion and 0.6 M (NH4)2SO4 extraction performed on isolated nuclei to remove chromatin.
ARTICLE IN PRESS L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
55
The nuclear matrix fraction we prepared from these cells was enriched in 170-kDa topoisomerase IIa, an abundant component of nuclear matrix (Berrios et al., 1985) and NuMa, another structural component of the matrix (Barboro et al., 2002) but was depleted of histone H1, a well-established chromatin marker. Because of the importance in defining the fine localization of the key players of nuclear lipid signaling the same authors have chosen an immunoprecipitation approach to define the components associated with DGK-e" in interphase nuclei. MDA-MB-453 cell nuclear extracts were immunoprecipitated with either antibody to PIP2 or antibody to PLCb1. The antibody was able to PIP2 immunoprecipitate some of the phosphorylated form of RNA pol II present in the nuclear lysates. The fact that not all of the hyperphosphorylated RNA pol II was immunoprecipitated by antibody to PIP2 was consistent with the observation that hyperphosphorylated RNA pol II was only partially localized in the speckles (Mortillaro et al., 1996). However, anti-PIP2 antibody did not immunoprecipitate either DGK-y or PIPKIa, which were recovered in the supernatants of the immunoprecipitates. In contrast, antibody to PLCb1 failed to immunoprecipitate the hyperphosphorylated form of RNA pol II, but immunoprecipitated a substantial amount of both DGK-y and PIPKIa originally present in the nuclear lysates. A final touch was added by means of immunogold electron microscopy analysis. The SC-35 splicing factor was immunolocalized to structures corresponding to interchromatin granules and perichromatin fibrils, supporting previous evidence by Spector et al. (1991). Interchromatin granule clusters and perichromatin fibrils are thought to correspond to the 20–40 intensely stained nuclear speckles when immunofluorescence is performed with probes to a variety of splicing factors (Boronenkov et al., 1998). Interestingly these structures also contained PIP2, as well as DGK-y. Tabellini et al. (2003) provided a further proof that DGK-y localizes to these nuclear domains using a speckle marker different from SC-35 or PIP2. Indeed, they employed the monoclonal antibody B3 (an IgM) which recognizes a hyperphosphorylated form of RNA polymerase II associated with the speckles (Mortillaro et al., 1996). Both SC-35 and the antigen recognized by B3 were present in the same nuclear domain. A double labeling experiment with antibody to DGK-y and B3 further demonstrated that DGK-y associated with interchromatin granules and perichromatin fibrils, in which also PLCb1 could be detected.
A meaning for the discrete subnuclear localization of the inositide cycle players Data obtained prior to those reported recently (Tabellini et al., 2003) suggest that nuclear speckles contain some elements of phosphoinositide metabolism (Boronenkov et al., 1998; Osborne et al., 2001; Didichenko and Thelen, 2001). The novelty comes from the demonstration for the first time that also DGK-y localizes to the nuclear speckle domains in different cell lines. Such a conclusion derives from several lines of evidence. Immunofluorescence staining coupled to CLSM analysis showed nuclear DGK-y to be concentrated in discrete structures that, for their number and
ARTICLE IN PRESS 56
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
size, reminded us of nuclear speckles. Double immunofluorescence staining experiments showed that nuclear DGK-y and PIP2 almost completely co-localized. Previous results have demonstrated that PIP2 is concentrated in nuclear speckles (Boronenkov et al., 1998; Osborne et al., 2001). In MDA-MB-453 cells the immunofluorescent staining pattern given by a monoclonal antibody highly specific for PIP2 is largely superimposable to that given by an antibody that recognizes the splicing component SC-35, a well-recognized marker of nuclear speckles. Moreover, the nuclear DGK-y fluorescent pattern was sensitive to a amanitin, a fungal toxin known to affect the number and/or the size of nuclear speckles. Taken together, these data strongly suggested that DGK-y is located in the nuclear speckles. An interesting observation is that the level of co-localization of PIP2 with DGK-y was not completely maintained after treatment with a amanitin. On the other hand, an almost complete co-localization of SC-35 with PIP2 was detected also in samples treated with the fungal toxin. These findings seem to indicate that most of DGK-y does not co-localize with SC-35 under conditions of impaired mRNA synthesis. This might depend on a different sensitivity and/or behavior of the speckle components to a amanitin. Also, immuno-electron microscopy unequivocally demonstrated that, in MDA-MB-453 cells, DGK-y was present in the interchromatin granules and perichromatin fibrils that, at the ultrastructural level, correspond to nuclear speckles (Spector et al., 1991), showing co-localization of DGK-y with a hyperphosphorylated form of RNA polymerase II, which is known to be present in the speckles as well as with both PIP2 and PLCb1. Cell fractionation experiments showed DGK-y to be strongly enriched in nuclei prepared from MDA-MB-453, HeLa, and PC12 cells. These results were consistent with the findings provided by immunofluorescence staining, which demonstrated that DGK-y was preferentially present in the nucleus. DGK-y which can be recovered associated to the nuclear matrix might represent the active fraction of the enzyme. Indeed, the nuclear matrix is viewed by several investigators as the fundamental organizing principle of the nucleus where many functions take place, including DNA replication, gene expression, and protein phosphorylation (Nickerson, 2001; Berezney, 2002; Martelli et al., 2002c). Since several enzymes of the inositol lipid metabolism have been found associated with the nuclear matrix (Maraldi et al., 1999) it is conceivable that the matrix may also be involved in intranuclear signal transduction pathways. The presence of DGK in the nucleus was first shown by immunocytochemical staining of rat retina and brain (Goto and Kondo, 1996). Following this report, other groups have demonstrated that DGK-z (Topham et al., 1998), a or g (Shirai et al., 2000), and DGK-y (Bregoli et al., 2001) without any clue for localization to specific domains are nuclear. DAG kinase-z possesses a bipartite nuclear targeting motif located close to the second zinc finger-like sequence in its regulatory domain (Topham et al., 1998). However, the nuclear localization sequences of other DGK isoforms remain to be identified. In any case, it will also be very important to analyze whether or not DGK-y displays a sequence element directing it to the speckles, such as the motif recently identified in the essential spliceosomal protein SF3b155 (Eilbracht and Schmidt-Zachmann, 2001). Evidence
ARTICLE IN PRESS L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
57
indicates that DGK-y was responsible for the increased nuclear DGK activity which followed stimulation of quiescent IIC9 cells with a-thrombin (Bregoli et al., 2002). It should be emphasized that resting IIC9 cells displayed immunoreactivity for DGK-y both in the cytoplasm and in the nucleus. The nuclear staining pattern given by the same monoclonal antibody to DGK-y employed in this study was reminiscent of nuclear speckles. However, after stimulation with athrombin there was a translocation of some DGK-y from the cytoplasm to the nucleus, as shown by Western blot, but the overall nuclear fluorescent pattern did not change. The association of DGK-y with nuclear speckles appears extremely interesting, because of other results showing that several elements of the phosphoinositide cycle are present within this subnuclear compartment. These elements include PIPKIa, PIPKIIa, PIPKIIb, PIP2, and PI3K C2a (Boronenkov et al., 1998; Osborne et al., 2001; Didichenko and Thelen, 2001). The issue of a domain-located multienzymatic complex, endowed with enzymes of phosphoinositide cycle, at the nucleus seems to be elucidated by the recent data we have reviewed. Indeed, PIP2 is associated with a hyperphosphorylated form of RNA pol II, DGK-y is not recovered in the immunoprecipitates obtained with anti-PIP2 antibody, but rather in those collected after immunoprecipitation with an antibody to PLCb1, that also contain PIPKIa. This actually indicates the existence in the nuclear speckles of a multienzymatic complex composed of proteins involved in phosphoinositide metabolism. PLCb1 could hydrolyze PIP2 yielding DAG and inositol 1,4,5-trisphosphate, while DGK-y may convert DAG into PA. The presence of PIPKIa (which phosphorylates phosphatidylinositol 4-phosphate to PIP2) suggests that the speckles are also the site for intranuclear PIP2 synthesis. The detergent-resistant PIP2 pool in the speckles has also been related to a role played by this inositol lipid in pre-mRNA splicing (Osborne et al., 2001). Even though DGK-y apparently did not associate with hyperphosphorylated RNA pol II, RNase A digestion experiments revealed that RNA was important for determining the association of DGK-e" with the speckles. Such a requirement for the presence of RNA has been reported for other components of the speckles, such as PI3K C2a (Didichenko and Thelen, 2001) or protein 4.1N (Lallena and Correas, 1997). However, it may be that this simply reflects the fact that DGK-y is nuclear matrix-bound, because RNA is very important for the structural integrity of the matrix (Nickerson, 2001; Barboro et al., 2003). The localization of DGK-y to a specific subnuclear site is an important step toward a better understanding of compartmentalization and function of phosphoinositide nuclear metabolism mainly because of the co-localization with other key players of this signaling system. Therefore, taking into account previous reports, which hinted at both transcription factors such as E2F and NF-E2 in mammals (Faenza et al., 2000, 2002) and mRNA synthesis/transport in yeast (York et al., 2001) as downstream targets of nuclear inositide signaling, the observations we have discussed in this review strengthen the contention that inositol lipid signaling and its related pathways in the nucleus are key steps in transcriptional events.
ARTICLE IN PRESS 58
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
Summary The cell nucleus contains enzymes that are involved in lipid-dependent signal transduction. DAG is an important lipid second messenger that is produced in the nucleus. Several reports have shown that the nucleus contains DGKs, i.e. the enzymes that, by converting DAG into PA, terminate DAG-dependent events. Here, we have discussed how by immunofluorescence staining and confocal analysis, DGK-y can be found at high extent in the nucleus of various cell lines, such as MDA-MB-453, MCF-7, PC12, and HeLa. DGK-y co-localizes with PIP2 in nuclear domains corresponding to nuclear speckles. The spatial distribution of nuclear DGK-y is dynamic and correlates to the transcription of mRNA. DGK-y, PIP2, and PLCb1 associate with electron-dense particles within the nucleus that correspond to interchromatin granule clusters. Cell fractionation experiments indicate a preferential association of DGK-y with the nucleus and the immunochemical analysis shows that DGK-y is mainly located in the nuclear matrix. Moreover, immunoprecipitation experiments reveal an association between PLCb1 and both DGK-y and PIPKIa. It appears evident that speckles represent a crucial site for the nuclear-based inositol lipid cycle, a site were DGK-y, upon cell stimulation with an agonist, converts to PA the DAG derived from PLCb1-dependent PIP2 hydrolysis.
Acknowledgements This work was supported by Associazione Italiana per la Ricerca sul Cancro (A.I.R.C.), Italian ‘‘MIURCofin 1999 and 2001’’, Selected Topics Research Fund from Bologna University, and CARISBO Foundation, Bologna. The authors are grateful to A.M. Billi for the preparation of this manuscript.
References Barboro P, D’Arrigo C, Diaspro A, Formino M, Alberti I, Parodi S, Patrone E, Balbi C. Unraveling the organization of the internal nuclear matrix. RNA-dependent anchoring of NuMa to a lamin scaffold. Exp Cell Res 2002;279:2002–218. Barboro P, D’Arrigo C, Mormino M, Coradeghini R, Parodi S, Patrone E, Balbi C. An intranuclear frame for chromatin compartmentalization and higher-order folding. J Cell Biochem 2003;88:113–20. Berezney R. Regulating the mammalian genome: the role of nuclear architecture. Adv Enzyme Regul 2002;42:39–52. Berrios M, Osheroff N, Fisher PA. In situ localization of DNA topoisomerase II, a major polypeptide component of the Drosophila nuclear matrix. Proc Natl Acad Sci USA 1985;82:4142–6. Boronenkov IV, Loijens JC, Umeda M, Anderson RA. Phosphoinositide pathways in nuclei are associated with nuclear speckles containing pre-mRNA processing factors. Mol Biol Cell 1998;9:3547–60. Bregoli L, Baldassare JJ, Raben DM. Nuclear diacylglycerol kinase-y is activated in response to athrombin. J Biol Chem 2001;276:23288–95. Bregoli L, Tu-Sekine B, Raben D. DGK and nuclear signaling. Nuclear diacylglycerol kinases in IIC9 cells. Adv Enzyme Regul 2002;42:213–26.
ARTICLE IN PRESS L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
59
Carmo-Fonseca M, Pepperkok R, Carvalho MT, Lamond AI. Transcription-dependent colocalization of the U1, U2, U4/U6, and U5 snRNPs in coiled bodies. J Cell Biol 1992;117:1–14. Cocco L, Martelli AM, Barnabei O, Manzoli FA. Nuclear inositol lipid signaling. Adv Enzyme Regul 2001a;41:361–84. Cocco L, Martelli AM, Gilmour RS, Rhee SG, Manzoli FA. Nuclear phospholipase C and signaling. Biochim Biophys Acta 2001b;1530:1–14. Didichenko SA, Thelen M. Phosphatidylinositol 3-kinase C2a contains a nuclear localization sequence and associates with nuclear speckles. J Biol Chem 2001;276:48135–42. Divecha N, Clarke JH, Roefs M, Halstead JR, D’Santos C. Nuclear inositides: inconsistent consistencies. Cell Mol Life Sci 2000;57:379–93. Divecha N, Irvine RF. Phospholipid signalling. Cell 1995;80:269–78. D’Santos CS, Clarke JH, Divecha N. Phospholipid signalling in the nucleus. Een DAG uit het leven van de inositide signalering in de nucleus. Biochim Biophys Acta 1998;1436:201–32. Dundr M, Misteli T. Functional architecture in the cell nucleus. Biochem J 2001;356:297–310. Eilbracht J, Schmidt-Zachmann MS. Identification of a sequence element directing a protein to nuclear speckles. Proc Natl Acad Sci USA 2001;98:3849–54. Faenza I, Matteucci A, Bavelloni A, Marmiroli S, Martelli AM, Gilmour RS, Suh PG, Manzoli L, Cocco L. Nuclear PLCb1 acts as a negative regulator of p45/NF-E2 expression levels in Friend erythroleukemia cells. Biochim Biophys Acta 2002;1589:305–10. Faenza I, Matteucci A, Manzoli L, Billi AM, Aluigi M, Peruzzi D, Vitale M, Castorina S, Suh PG, Cocco L. A role for nuclear phospholipase Cb1 in cell cycle control. J Biol Chem 2000;275:30520–4. Goto K, Kondo H. A 104-kDa diacylglycerol kinase containing ankyrin-like repeats localizes in the cell nucleus. Proc Natl Acad Sci USA 1996;93:11196–201. Goto K, Kondo H. Diacylglycerol kinase in the central nervous system-molecular heterogeneity and gene expression. Chem Phys Lipids 1999;98:109–17. Irvine RF. Nuclear lipid signalling. Nat Rev Mol Cell Biol 2003;4:1–12. Kanoh H, Yamada K, Sakane F. Diacylglycerol kinases: emerging downstream regulators in cell signaling systems. J Biochem 2002;131:629–33. Lallena M-J, Correas I. Transcription-dependent redistribution of nuclear protein 4.1 to SC35-enriched nuclear domains. J Cell Sci 1997;110:239–47. Maraldi NM, Zini N, Santi S, Manzoli FA. Topology of inositol lipid signal transduction in the nucleus. J Cell Physiol 1999;181:203–17. Martelli AM, Capitani S, Neri LM. The generation of lipid signaling molecules in the nucleus. Prog Lipid Res 1999;38:273–308. Martelli AM, Bortul R, Tabellini G, Aluigi M, Peruzzi D, Bareggi R, Narducci P, Cocco L. Re-examination of the mechanisms regulating nuclear inositol lipid metabolism. FEBS Lett 2001;505:1–6. Martelli AM, Bortul R, Tabellini G, Bareggi R, Manzoli L, Narducci P, Cocco L. Diacylglycerol kinases in nuclear lipid-dependent signal transduction pathways. Cell Mol Life Sci 2002b;59:1129–37. Martelli AM, Falcieri E, Zweyer M, Bortul R, Tabellini G, Cappellini A, Cocco L, Manzoli L. The controversial nuclear matrix: a balanced point of view. Histol Histopathol 2002c;17:1193–205. Martelli AM, Manzoli L, Faenza I, Bortul R, Billi AM, Cocco L. Nuclear inositol lipid signaling and its potential involvement in neoplastic transformation. Biochim Biophys Acta 2002a;1603:11–7. Martelli AM, Tabellini G, Borgatti R, Bortul R, Capitani S, Neri LM. Nuclear lipids: new functions for old molecules? J Cell Biochem 2003;88:455–61. Martelli AM, Tabellini G, Bortul R, Manzoli L, Bareggi R, Baldini G, Grill V, Zweyer M, Narducci P, Cocco L. Enhanced nuclear diacylglycerol kinase activity in response to a mitogenic stimulation of quiescent Swiss 3T3 cells with insulin-like growth factor-I. Cancer Res 2000;60:815–21. Martin TE. Phosphoinositide lipids as signaling molecules: common themes for signal transduction, cytoskeletal regulation, and membrane trafficking. Annu Rev Cell Dev Biol 1998;14:231–64. Mortillaro MJ, Blencowe BJ, Wei X, Nakayasu H, Du L, Warren SL, Sharp PA, Berezney R. A hyperphosphorylated form of the large subunit of RNA polymerase II is associated with splicing complexes and the nuclear matrix. Proc Natl Acad Sci USA 1996;93:8253–7.
ARTICLE IN PRESS 60
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
Neri LM, Borgatti P, Capitani S, Martelli AM. Protein kinase C isoforms and lipid second messengers: a critical nuclear partnership? Histol Histopathol 2002;17:1311–6. Nickerson JA. Experimental observations of a nuclear matrix. J Cell Sci 2001;114:463–74. Osborne SL, Thomas CL, Gschmeissner S, Schiavo G. Nuclear PtdIns(4,5)P2 assembles in a mitotically regulated particle involved in pre-mRNA splicing. J Cell Sci 2001;114:2501–11. Pendaries C, Tronch"ere H, Plantavid M, Payrastre B. Phosphoinositide signaling disorders in human diseases. FEBS Lett 2003;546:25–31. Ron D, Kazanietz MG. New insights into the regulation of protein kinase C and novel phorbol ester receptors. FASEB J 1999;13:1658–76. Shirai Y, Segawa S, Kuriyama M, Goto K, Sakai N, Saito N. Subtype-specific translocation of diacylglycerol kinase a and g and its correlation with protein kinase C. J Biol Chem 2000;275:24760–6. Spector DL, Fu XD, Maniatis T. Associations between distinct pre-mRNA splicing components and the cell nucleus. EMBO J 1991;10:4367–81. Tabellini G, Bortul R, Santi S, Riccio M, Baldini G, Cappellini A, Billi AM, Berezney R, Ruggeri A, Cocco L, Martelli AM. Diacylglycerol kinase-y is localized in the speckle domains of the nucleus. Exp Cell Res 2003;287:143–54. Toker A. Phosphoinositides and signal transduction. Cell Mol Life Sci 2002;59:761–79. Topham MK, Bunting M, Zimmerman GA, McIntyre TM, Blackshear PJ, Prescott SM. Protein kinase C regulates the nuclear localization of diacylglycerol kinase-z. Nature 1998;394:697–700. Topham MK, Prescott SM. Mammalian diacylglycerol kinases, a family of lipid kinases with signaling functions. J Biol Chem 1999;274:11447–50. van Blitterswijk WJ, Houssa B. Diacylglycerol kinases in signal transduction. Chem Phys Lipids 1999;98:95–108. van Blitterswijk WJ, Houssa B. Properties and functions of diacylglycerol kinases. Cell Signal 2000;12:595–605. Wakelam MJO. Diacylglycerol—when is it an intracellular messenger? Biochim Biophys Acta 1998;1436:117–26. York JD, Guo S, Odom AR, Spiegelberg BD, Stolz LE. An expanded view of inositol signaling. Adv Enzyme Regul 2001;41:57–71.
Advan. Enzyme Regul. 44 (2004) 51–60
Significance of subnuclear localization of key players of inositol lipid cycle Lucio Coccoa,*, Lucia Manzolia, Ottavio Barnabeib, Alberto M. Martellia,c,d a
Cellular Signaling Laboratory, Department of Anatomical Sciences, University of Bologna, Italy b Department of Biology, University of Bologna, Italy c ITOI-CNR, Bologna Unit, c/o IOR, Bologna, Italy d School of Pharmacy, University of Bologna, Italy
Introduction Among phospholipids, polyphosphoinositides are key players in a number of signal transducing pathways, even though they are quantitatively minor constituents of cell membranes. Indeed, phosphatidylinositol (PI) typically comprises approximately 10% of total membrane lipid, suggesting that these lipids are not likely to play a major structural role. The discovery of the second messenger role of inositol (1,4,5) trisphosphate (IP3) and diacylglycerol (DAG) has brought phosphoinositide signaling to the very forefront of biological and biomedical research. Signaling through various phosphoinositides, including 3-phosphorylated inositol lipids, has been shown to mediate cell growth and differentiation, apoptosis, intracellular vesicle trafficking, ion channel activation, insulin action, cytoskeletal changes, and motility (Toker, 2002). Therefore, at present it is difficult to think of a cell response which does not utilize some form of inositide-dependent signal transduction. Moreover, it is now clear that phosphoinositide signaling disorders are implicated in the pathogenesis of various human diseases ranging from cancer to type 2 diabetes (Pendaries et al., 2003). The generally accepted model of phosphoinositide signaling involves the generation of lipid second messengers in response to stimuli in a receptor-mediated manner at the plasma membrane. However, the existence of a nuclear PI metabolism independent from that occurring elsewhere in the cell is now widely recognized (Cocco et al., 2001a), suggesting that the nucleus constitutes both a functional and a *Corresponding author. Cellular Signaling Laboratory, Department of Anatomical Sciences, University of Bologna, via Irnerio 48, Bologna 40126, Italy. Tel.: +39-051-244467; fax: +39-051-251735. E-mail address: [email protected] (L. Cocco). 0065-2571/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.advenzreg.2003.11.009
ARTICLE IN PRESS 52
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
distinct compartment for PI metabolism (Martelli et al., 2003). Indeed, it has been demonstrated that nuclei contain many of the enzymes involved in the classical PI cycle, including kinases required for the synthesis of phosphatidylinositol (4,5)bisphosphate (PIP2), phosphoinositide-specific phospholipase C (PLC), and diacylglycerol kinase (DGK) (D’Santos et al., 1998; Martelli et al., 1999, 2000; Cocco et al., 2001b; Irvine, 2003). More importantly, specific changes in the nuclear PI metabolism have been implicated in cell growth, differentiation, and neoplastic transformation (Martelli et al., 2002a). In this review, we shall highlight the major discoveries in the field of subnuclear localization of the enzymes of the inositol lipid metabolism and pay particular attention to progress made by our group in the co-localization of both enzymes and substrates of this signaling system.
DAG in the nucleus Lipid second messengers are essential intermediates linking extracellular stimuli via receptor activation to the required cellular response (Divecha and Irvine, 1995; Martin, 1998). These messengers are generated along very complex signaling pathways. DAG is a key lipid second messenger which can derive from either phosphoinositides or phosphatidylcholine (Wakelam, 1998). DAG activates some of the members of the protein kinase C (PKC) family and thereby the signaling pathways downstream of PKC. The DAG-dependent PKC isozymes include conventional PKC-a, bI ; bII ; g; and novel PKC d; e; Z; y; m: In contrast, atypical PKC-z and i= l do not require DAG for their activation (Ron and Kazanietz, 1999). However, other molecular targets of DAG have been identified such as a- and b-chimaerins, the guanine nucleotide-exchange factor vav and guanyl nucleotide-exchange factors for Ras and Rap (van Blitterswijk and Houssa, 2000; Bregoli et al., 2001). These findings indicate that DAG might also be involved in controlling the Ras and Rho family of proteins. In addition to cell surface receptor-mediated transmembrane signaling pathways involving phospholipid metabolism, a number of reports have convincingly demonstrated that the nucleus must also be considered as a site where biologically active lipid second messengers are generated (D’Santos et al., 1998; Divecha et al., 2000; Cocco et al., 2001a, b; Martelli et al., 2001; Irvine, 2003). The control of steady-state cellular levels of DAG is crucial to cellular physiology. DAG signaling must be short-lived because persistently high levels of DAG induce malignant transformation. The transforming activity of DAG has been attributed most often to persistent activation of PKC isoforms that are clearly involved in tumorigenesis (Ron and Kazanietz, 1999). DGKs, whose activity increases upon cell stimulation by a variety of agonists, metabolize DAG by converting it to phosphatidic acid (PA). Since they can attenuate local accumulation of signaling DAG, DGKs play a pivotal role in many biological responses such as cell proliferation, differentiation, survival, and apoptosis. Moreover, PA itself can function as a second messenger. Up to now, nine mammalian DGK isoforms have been identified which differ in their tissue
ARTICLE IN PRESS L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
53
expression and structural domains (Goto and Kondo, 1999; van Blitterswijk and Houssa, 1999; Topham and Prescott, 1999; van Blitterswijk and Houssa, 2000; Kanoh et al., 2002). DAG is also produced in the nucleus. The levels of nuclear DAG fluctuate during the cell cycle progression, suggesting that it has important regulatory roles. Most likely, nuclear DAG serves as a chemoattractant for some isoforms of PKC that migrate to the nucleus in response to a variety of agonists (D’Santos et al., 1998; Divecha et al., 2000; Cocco et al., 2001a, b; Martelli et al., 2001; Irvine, 2003; Neri et al., 2002). Several independent laboratories have indicated that DGK isoforms are present in the nucleus where they may be involved in regulating the amount of DAG (Martelli et al., 2002b). In some cases, the activity of nuclear DGKs has been demonstrated to be very critical for the control of cell proliferation (Topham et al., 1998; Martelli et al., 2000; Bregoli et al., 2002). When nuclear DGK isoforms were visualized by means of immunostaining or green fluorescent protein (GFP) technology, it became apparent that they were not distributed in a diffuse manner but rather concentrated in defined domains (Bregoli et al., 2001; Topham et al., 1998; Shirai et al., 2000). It has been proposed, but never proved, that DGK-y is localized in the speckle domains of the nucleus (van Blitterswijk and Houssa, 2000). The speckle domains are nuclear subcompartments enriched in small ribonucleoprotein particles and various splicing factors (Dundr and Misteli, 2001). Previous results have highlighted that nuclear speckles contain elements of the phosphoinositide cycle, including various types of phosphatidylinositol kinase (PIPK), PIP2, and phosphoinositide 3-kinase (PI3K) (Boronenkov et al., 1998; Osborne et al., 2001; Didichenko and Thelen, 2001).
Nuclear DGK-h and co-localization at the speckle domains with PLCb1, PIP2, PIPK1a and SC-35 splicing factor Albeit the hints for the presence in discrete nuclear regions of several players, an accurate and combined cytochemical and immunochemical analysis was still lacking, mainly regarding DGKs. A recent paper from our laboratory gave new insight in this issue (Tabellini et al., 2003). Here we summarize the main findings. It has been shown that DKG-y is expressed in several cell types such as PC12, HeLa, MDA-MB453, and MCF-7 cells. A series of experiments have clarified the subnuclear localization of DKG-y. In immunostained MDA-MB-453 cells for SC-35, PIP2, and DKG-y, a confocal laser scanning microscope (CLSM) analysis showed that besides SC-35 being mainly located in discrete nonnucleolar domains, corresponding to speckles, immunostaining with monoclonal antibody to PIP2 resulted in a similar pattern. Boronenkov et al. (1998) and Osborne et al. (2001) hinted at a speckle location of PIP2. This has been confirmed in MDA-MB-453, since double immunofluorescence staining with antibodies to SC-35 (an IgG) and PIP2 (an IgM) provided evidence that the two molecules co-localized at the levels of the speckles (Tabellini et al., 2003). Moreover, the authors showed by double immunofluorescence staining with antibodies to DGK-y (an IgG) and PIP2 that the two molecules co-localized at the level of nuclear foci, corresponding to speckles.
ARTICLE IN PRESS 54
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
The co-localization of DGK-y with PIP2 in nuclear foci was also detected in MCF-7, PC12, and HeLa cells. Overall, DGK-y immunoreactivity was mainly found in the nucleus. HeLa and PC12 cells showed some DGK-y staining in the cytoplasm, whereas in MDA-MB-453 and MCF-7 cells cytoplasmic positivity was almost undetectable. Cytoplasmic PIP2 immunostaining was mostly seen in HeLa, MDAMB-453, and PC12 cells. An interesting feature reported in this paper is the quantitative analysis of the results provided by CLSM using both Pearson’s correlation and overlap coefficient. Pearson’s correlation provides information about the similarity of shape between images without regard to the average intensity of the signals (spatial co-localization). It is a value computed to be between 1 and 1. The overlap coefficient, simultaneously used to describe co-localization, does not perform any pixel averaging functions, so correlations are returned as values between 0 and 1. This method is not sensitive to intensity variations in the image analysis and this is especially important when considering issues typical to fluorescence imaging such as sample photo-bleaching or different setting of the detectors. Not surprisingly all the values about the co-immunostaining of SC-35/ PIP2 and DGK-y/PIP2 were higher than 74%, indicating a remarkable level of colocalization. A further confirmation of the co-localization comes from the data obtained with a amanitin. Indeed, it is known that cell treatment with the transcriptional inhibitor a amanitin at concentrations that specifically inhibit RNA polymerase II causes reorganization of nuclear speckle domains into few and/or large foci (CarmoFonseca et al., 1992). In HeLa cells a amanitin caused the expected changes in SC-35 immunofluorescence staining, in that speckles were larger and less numerous. The paper (Tabellini et al., 2003) reports that also intranuclear distribution of DGK-y changed in response to a amanitin, since the foci became less numerous, even though their size did not increase as much as for SC-35 staining. The authors have also established whether or not the level of co-localization of SC-35 with PIP2 and of DGK-y with PIP2 was maintained after treatment with a amanitin. The level of colocalization of SC-35/PIP2 was maintained in a amanitin-treated cells. In contrast, double immunofluorescence labeling for DGK-y/PIP2 revealed that, after a amanitin exposure, the extent of co-localization was not as high as in untreated samples. The data discussed above could suggest that the players of the nuclear phosphoinositide cycle locate in discrete domains which are not chromatin-related. Indeed, it has been shown that RNA, but not DNA, is essential for the association of PIP2 with nuclear speckles (Osborne et al., 2001). In MDA-MB-453 cells DNase I treatment did not affect the intranuclear distribution of DGK-y, which colocalizes with PIP2, whilst RNase A treatment completely removed intranuclear immunoreactivity for DGK-y. These results imply that RNA, but not DNA, is essential for the association of DGK-y with speckles. Therefore, the analysis of the subcellular distribution of DGK-y is a good tool to assess the whole phosphoinositide cycle and related enzymes. By Western blot analysis Tabellini et al. (2003) could demonstrate that in MDA-MB-453 cells DGK-y was enriched in the nuclear matrix fraction which was obtained by DNase I digestion and 0.6 M (NH4)2SO4 extraction performed on isolated nuclei to remove chromatin.
ARTICLE IN PRESS L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
55
The nuclear matrix fraction we prepared from these cells was enriched in 170-kDa topoisomerase IIa, an abundant component of nuclear matrix (Berrios et al., 1985) and NuMa, another structural component of the matrix (Barboro et al., 2002) but was depleted of histone H1, a well-established chromatin marker. Because of the importance in defining the fine localization of the key players of nuclear lipid signaling the same authors have chosen an immunoprecipitation approach to define the components associated with DGK-e" in interphase nuclei. MDA-MB-453 cell nuclear extracts were immunoprecipitated with either antibody to PIP2 or antibody to PLCb1. The antibody was able to PIP2 immunoprecipitate some of the phosphorylated form of RNA pol II present in the nuclear lysates. The fact that not all of the hyperphosphorylated RNA pol II was immunoprecipitated by antibody to PIP2 was consistent with the observation that hyperphosphorylated RNA pol II was only partially localized in the speckles (Mortillaro et al., 1996). However, anti-PIP2 antibody did not immunoprecipitate either DGK-y or PIPKIa, which were recovered in the supernatants of the immunoprecipitates. In contrast, antibody to PLCb1 failed to immunoprecipitate the hyperphosphorylated form of RNA pol II, but immunoprecipitated a substantial amount of both DGK-y and PIPKIa originally present in the nuclear lysates. A final touch was added by means of immunogold electron microscopy analysis. The SC-35 splicing factor was immunolocalized to structures corresponding to interchromatin granules and perichromatin fibrils, supporting previous evidence by Spector et al. (1991). Interchromatin granule clusters and perichromatin fibrils are thought to correspond to the 20–40 intensely stained nuclear speckles when immunofluorescence is performed with probes to a variety of splicing factors (Boronenkov et al., 1998). Interestingly these structures also contained PIP2, as well as DGK-y. Tabellini et al. (2003) provided a further proof that DGK-y localizes to these nuclear domains using a speckle marker different from SC-35 or PIP2. Indeed, they employed the monoclonal antibody B3 (an IgM) which recognizes a hyperphosphorylated form of RNA polymerase II associated with the speckles (Mortillaro et al., 1996). Both SC-35 and the antigen recognized by B3 were present in the same nuclear domain. A double labeling experiment with antibody to DGK-y and B3 further demonstrated that DGK-y associated with interchromatin granules and perichromatin fibrils, in which also PLCb1 could be detected.
A meaning for the discrete subnuclear localization of the inositide cycle players Data obtained prior to those reported recently (Tabellini et al., 2003) suggest that nuclear speckles contain some elements of phosphoinositide metabolism (Boronenkov et al., 1998; Osborne et al., 2001; Didichenko and Thelen, 2001). The novelty comes from the demonstration for the first time that also DGK-y localizes to the nuclear speckle domains in different cell lines. Such a conclusion derives from several lines of evidence. Immunofluorescence staining coupled to CLSM analysis showed nuclear DGK-y to be concentrated in discrete structures that, for their number and
ARTICLE IN PRESS 56
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
size, reminded us of nuclear speckles. Double immunofluorescence staining experiments showed that nuclear DGK-y and PIP2 almost completely co-localized. Previous results have demonstrated that PIP2 is concentrated in nuclear speckles (Boronenkov et al., 1998; Osborne et al., 2001). In MDA-MB-453 cells the immunofluorescent staining pattern given by a monoclonal antibody highly specific for PIP2 is largely superimposable to that given by an antibody that recognizes the splicing component SC-35, a well-recognized marker of nuclear speckles. Moreover, the nuclear DGK-y fluorescent pattern was sensitive to a amanitin, a fungal toxin known to affect the number and/or the size of nuclear speckles. Taken together, these data strongly suggested that DGK-y is located in the nuclear speckles. An interesting observation is that the level of co-localization of PIP2 with DGK-y was not completely maintained after treatment with a amanitin. On the other hand, an almost complete co-localization of SC-35 with PIP2 was detected also in samples treated with the fungal toxin. These findings seem to indicate that most of DGK-y does not co-localize with SC-35 under conditions of impaired mRNA synthesis. This might depend on a different sensitivity and/or behavior of the speckle components to a amanitin. Also, immuno-electron microscopy unequivocally demonstrated that, in MDA-MB-453 cells, DGK-y was present in the interchromatin granules and perichromatin fibrils that, at the ultrastructural level, correspond to nuclear speckles (Spector et al., 1991), showing co-localization of DGK-y with a hyperphosphorylated form of RNA polymerase II, which is known to be present in the speckles as well as with both PIP2 and PLCb1. Cell fractionation experiments showed DGK-y to be strongly enriched in nuclei prepared from MDA-MB-453, HeLa, and PC12 cells. These results were consistent with the findings provided by immunofluorescence staining, which demonstrated that DGK-y was preferentially present in the nucleus. DGK-y which can be recovered associated to the nuclear matrix might represent the active fraction of the enzyme. Indeed, the nuclear matrix is viewed by several investigators as the fundamental organizing principle of the nucleus where many functions take place, including DNA replication, gene expression, and protein phosphorylation (Nickerson, 2001; Berezney, 2002; Martelli et al., 2002c). Since several enzymes of the inositol lipid metabolism have been found associated with the nuclear matrix (Maraldi et al., 1999) it is conceivable that the matrix may also be involved in intranuclear signal transduction pathways. The presence of DGK in the nucleus was first shown by immunocytochemical staining of rat retina and brain (Goto and Kondo, 1996). Following this report, other groups have demonstrated that DGK-z (Topham et al., 1998), a or g (Shirai et al., 2000), and DGK-y (Bregoli et al., 2001) without any clue for localization to specific domains are nuclear. DAG kinase-z possesses a bipartite nuclear targeting motif located close to the second zinc finger-like sequence in its regulatory domain (Topham et al., 1998). However, the nuclear localization sequences of other DGK isoforms remain to be identified. In any case, it will also be very important to analyze whether or not DGK-y displays a sequence element directing it to the speckles, such as the motif recently identified in the essential spliceosomal protein SF3b155 (Eilbracht and Schmidt-Zachmann, 2001). Evidence
ARTICLE IN PRESS L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
57
indicates that DGK-y was responsible for the increased nuclear DGK activity which followed stimulation of quiescent IIC9 cells with a-thrombin (Bregoli et al., 2002). It should be emphasized that resting IIC9 cells displayed immunoreactivity for DGK-y both in the cytoplasm and in the nucleus. The nuclear staining pattern given by the same monoclonal antibody to DGK-y employed in this study was reminiscent of nuclear speckles. However, after stimulation with athrombin there was a translocation of some DGK-y from the cytoplasm to the nucleus, as shown by Western blot, but the overall nuclear fluorescent pattern did not change. The association of DGK-y with nuclear speckles appears extremely interesting, because of other results showing that several elements of the phosphoinositide cycle are present within this subnuclear compartment. These elements include PIPKIa, PIPKIIa, PIPKIIb, PIP2, and PI3K C2a (Boronenkov et al., 1998; Osborne et al., 2001; Didichenko and Thelen, 2001). The issue of a domain-located multienzymatic complex, endowed with enzymes of phosphoinositide cycle, at the nucleus seems to be elucidated by the recent data we have reviewed. Indeed, PIP2 is associated with a hyperphosphorylated form of RNA pol II, DGK-y is not recovered in the immunoprecipitates obtained with anti-PIP2 antibody, but rather in those collected after immunoprecipitation with an antibody to PLCb1, that also contain PIPKIa. This actually indicates the existence in the nuclear speckles of a multienzymatic complex composed of proteins involved in phosphoinositide metabolism. PLCb1 could hydrolyze PIP2 yielding DAG and inositol 1,4,5-trisphosphate, while DGK-y may convert DAG into PA. The presence of PIPKIa (which phosphorylates phosphatidylinositol 4-phosphate to PIP2) suggests that the speckles are also the site for intranuclear PIP2 synthesis. The detergent-resistant PIP2 pool in the speckles has also been related to a role played by this inositol lipid in pre-mRNA splicing (Osborne et al., 2001). Even though DGK-y apparently did not associate with hyperphosphorylated RNA pol II, RNase A digestion experiments revealed that RNA was important for determining the association of DGK-e" with the speckles. Such a requirement for the presence of RNA has been reported for other components of the speckles, such as PI3K C2a (Didichenko and Thelen, 2001) or protein 4.1N (Lallena and Correas, 1997). However, it may be that this simply reflects the fact that DGK-y is nuclear matrix-bound, because RNA is very important for the structural integrity of the matrix (Nickerson, 2001; Barboro et al., 2003). The localization of DGK-y to a specific subnuclear site is an important step toward a better understanding of compartmentalization and function of phosphoinositide nuclear metabolism mainly because of the co-localization with other key players of this signaling system. Therefore, taking into account previous reports, which hinted at both transcription factors such as E2F and NF-E2 in mammals (Faenza et al., 2000, 2002) and mRNA synthesis/transport in yeast (York et al., 2001) as downstream targets of nuclear inositide signaling, the observations we have discussed in this review strengthen the contention that inositol lipid signaling and its related pathways in the nucleus are key steps in transcriptional events.
ARTICLE IN PRESS 58
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
Summary The cell nucleus contains enzymes that are involved in lipid-dependent signal transduction. DAG is an important lipid second messenger that is produced in the nucleus. Several reports have shown that the nucleus contains DGKs, i.e. the enzymes that, by converting DAG into PA, terminate DAG-dependent events. Here, we have discussed how by immunofluorescence staining and confocal analysis, DGK-y can be found at high extent in the nucleus of various cell lines, such as MDA-MB-453, MCF-7, PC12, and HeLa. DGK-y co-localizes with PIP2 in nuclear domains corresponding to nuclear speckles. The spatial distribution of nuclear DGK-y is dynamic and correlates to the transcription of mRNA. DGK-y, PIP2, and PLCb1 associate with electron-dense particles within the nucleus that correspond to interchromatin granule clusters. Cell fractionation experiments indicate a preferential association of DGK-y with the nucleus and the immunochemical analysis shows that DGK-y is mainly located in the nuclear matrix. Moreover, immunoprecipitation experiments reveal an association between PLCb1 and both DGK-y and PIPKIa. It appears evident that speckles represent a crucial site for the nuclear-based inositol lipid cycle, a site were DGK-y, upon cell stimulation with an agonist, converts to PA the DAG derived from PLCb1-dependent PIP2 hydrolysis.
Acknowledgements This work was supported by Associazione Italiana per la Ricerca sul Cancro (A.I.R.C.), Italian ‘‘MIURCofin 1999 and 2001’’, Selected Topics Research Fund from Bologna University, and CARISBO Foundation, Bologna. The authors are grateful to A.M. Billi for the preparation of this manuscript.
References Barboro P, D’Arrigo C, Diaspro A, Formino M, Alberti I, Parodi S, Patrone E, Balbi C. Unraveling the organization of the internal nuclear matrix. RNA-dependent anchoring of NuMa to a lamin scaffold. Exp Cell Res 2002;279:2002–218. Barboro P, D’Arrigo C, Mormino M, Coradeghini R, Parodi S, Patrone E, Balbi C. An intranuclear frame for chromatin compartmentalization and higher-order folding. J Cell Biochem 2003;88:113–20. Berezney R. Regulating the mammalian genome: the role of nuclear architecture. Adv Enzyme Regul 2002;42:39–52. Berrios M, Osheroff N, Fisher PA. In situ localization of DNA topoisomerase II, a major polypeptide component of the Drosophila nuclear matrix. Proc Natl Acad Sci USA 1985;82:4142–6. Boronenkov IV, Loijens JC, Umeda M, Anderson RA. Phosphoinositide pathways in nuclei are associated with nuclear speckles containing pre-mRNA processing factors. Mol Biol Cell 1998;9:3547–60. Bregoli L, Baldassare JJ, Raben DM. Nuclear diacylglycerol kinase-y is activated in response to athrombin. J Biol Chem 2001;276:23288–95. Bregoli L, Tu-Sekine B, Raben D. DGK and nuclear signaling. Nuclear diacylglycerol kinases in IIC9 cells. Adv Enzyme Regul 2002;42:213–26.
ARTICLE IN PRESS L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
59
Carmo-Fonseca M, Pepperkok R, Carvalho MT, Lamond AI. Transcription-dependent colocalization of the U1, U2, U4/U6, and U5 snRNPs in coiled bodies. J Cell Biol 1992;117:1–14. Cocco L, Martelli AM, Barnabei O, Manzoli FA. Nuclear inositol lipid signaling. Adv Enzyme Regul 2001a;41:361–84. Cocco L, Martelli AM, Gilmour RS, Rhee SG, Manzoli FA. Nuclear phospholipase C and signaling. Biochim Biophys Acta 2001b;1530:1–14. Didichenko SA, Thelen M. Phosphatidylinositol 3-kinase C2a contains a nuclear localization sequence and associates with nuclear speckles. J Biol Chem 2001;276:48135–42. Divecha N, Clarke JH, Roefs M, Halstead JR, D’Santos C. Nuclear inositides: inconsistent consistencies. Cell Mol Life Sci 2000;57:379–93. Divecha N, Irvine RF. Phospholipid signalling. Cell 1995;80:269–78. D’Santos CS, Clarke JH, Divecha N. Phospholipid signalling in the nucleus. Een DAG uit het leven van de inositide signalering in de nucleus. Biochim Biophys Acta 1998;1436:201–32. Dundr M, Misteli T. Functional architecture in the cell nucleus. Biochem J 2001;356:297–310. Eilbracht J, Schmidt-Zachmann MS. Identification of a sequence element directing a protein to nuclear speckles. Proc Natl Acad Sci USA 2001;98:3849–54. Faenza I, Matteucci A, Bavelloni A, Marmiroli S, Martelli AM, Gilmour RS, Suh PG, Manzoli L, Cocco L. Nuclear PLCb1 acts as a negative regulator of p45/NF-E2 expression levels in Friend erythroleukemia cells. Biochim Biophys Acta 2002;1589:305–10. Faenza I, Matteucci A, Manzoli L, Billi AM, Aluigi M, Peruzzi D, Vitale M, Castorina S, Suh PG, Cocco L. A role for nuclear phospholipase Cb1 in cell cycle control. J Biol Chem 2000;275:30520–4. Goto K, Kondo H. A 104-kDa diacylglycerol kinase containing ankyrin-like repeats localizes in the cell nucleus. Proc Natl Acad Sci USA 1996;93:11196–201. Goto K, Kondo H. Diacylglycerol kinase in the central nervous system-molecular heterogeneity and gene expression. Chem Phys Lipids 1999;98:109–17. Irvine RF. Nuclear lipid signalling. Nat Rev Mol Cell Biol 2003;4:1–12. Kanoh H, Yamada K, Sakane F. Diacylglycerol kinases: emerging downstream regulators in cell signaling systems. J Biochem 2002;131:629–33. Lallena M-J, Correas I. Transcription-dependent redistribution of nuclear protein 4.1 to SC35-enriched nuclear domains. J Cell Sci 1997;110:239–47. Maraldi NM, Zini N, Santi S, Manzoli FA. Topology of inositol lipid signal transduction in the nucleus. J Cell Physiol 1999;181:203–17. Martelli AM, Capitani S, Neri LM. The generation of lipid signaling molecules in the nucleus. Prog Lipid Res 1999;38:273–308. Martelli AM, Bortul R, Tabellini G, Aluigi M, Peruzzi D, Bareggi R, Narducci P, Cocco L. Re-examination of the mechanisms regulating nuclear inositol lipid metabolism. FEBS Lett 2001;505:1–6. Martelli AM, Bortul R, Tabellini G, Bareggi R, Manzoli L, Narducci P, Cocco L. Diacylglycerol kinases in nuclear lipid-dependent signal transduction pathways. Cell Mol Life Sci 2002b;59:1129–37. Martelli AM, Falcieri E, Zweyer M, Bortul R, Tabellini G, Cappellini A, Cocco L, Manzoli L. The controversial nuclear matrix: a balanced point of view. Histol Histopathol 2002c;17:1193–205. Martelli AM, Manzoli L, Faenza I, Bortul R, Billi AM, Cocco L. Nuclear inositol lipid signaling and its potential involvement in neoplastic transformation. Biochim Biophys Acta 2002a;1603:11–7. Martelli AM, Tabellini G, Borgatti R, Bortul R, Capitani S, Neri LM. Nuclear lipids: new functions for old molecules? J Cell Biochem 2003;88:455–61. Martelli AM, Tabellini G, Bortul R, Manzoli L, Bareggi R, Baldini G, Grill V, Zweyer M, Narducci P, Cocco L. Enhanced nuclear diacylglycerol kinase activity in response to a mitogenic stimulation of quiescent Swiss 3T3 cells with insulin-like growth factor-I. Cancer Res 2000;60:815–21. Martin TE. Phosphoinositide lipids as signaling molecules: common themes for signal transduction, cytoskeletal regulation, and membrane trafficking. Annu Rev Cell Dev Biol 1998;14:231–64. Mortillaro MJ, Blencowe BJ, Wei X, Nakayasu H, Du L, Warren SL, Sharp PA, Berezney R. A hyperphosphorylated form of the large subunit of RNA polymerase II is associated with splicing complexes and the nuclear matrix. Proc Natl Acad Sci USA 1996;93:8253–7.
ARTICLE IN PRESS 60
L. Cocco et al. / Advan. Enzyme Regul. 44 (2004) 51–60
Neri LM, Borgatti P, Capitani S, Martelli AM. Protein kinase C isoforms and lipid second messengers: a critical nuclear partnership? Histol Histopathol 2002;17:1311–6. Nickerson JA. Experimental observations of a nuclear matrix. J Cell Sci 2001;114:463–74. Osborne SL, Thomas CL, Gschmeissner S, Schiavo G. Nuclear PtdIns(4,5)P2 assembles in a mitotically regulated particle involved in pre-mRNA splicing. J Cell Sci 2001;114:2501–11. Pendaries C, Tronch"ere H, Plantavid M, Payrastre B. Phosphoinositide signaling disorders in human diseases. FEBS Lett 2003;546:25–31. Ron D, Kazanietz MG. New insights into the regulation of protein kinase C and novel phorbol ester receptors. FASEB J 1999;13:1658–76. Shirai Y, Segawa S, Kuriyama M, Goto K, Sakai N, Saito N. Subtype-specific translocation of diacylglycerol kinase a and g and its correlation with protein kinase C. J Biol Chem 2000;275:24760–6. Spector DL, Fu XD, Maniatis T. Associations between distinct pre-mRNA splicing components and the cell nucleus. EMBO J 1991;10:4367–81. Tabellini G, Bortul R, Santi S, Riccio M, Baldini G, Cappellini A, Billi AM, Berezney R, Ruggeri A, Cocco L, Martelli AM. Diacylglycerol kinase-y is localized in the speckle domains of the nucleus. Exp Cell Res 2003;287:143–54. Toker A. Phosphoinositides and signal transduction. Cell Mol Life Sci 2002;59:761–79. Topham MK, Bunting M, Zimmerman GA, McIntyre TM, Blackshear PJ, Prescott SM. Protein kinase C regulates the nuclear localization of diacylglycerol kinase-z. Nature 1998;394:697–700. Topham MK, Prescott SM. Mammalian diacylglycerol kinases, a family of lipid kinases with signaling functions. J Biol Chem 1999;274:11447–50. van Blitterswijk WJ, Houssa B. Diacylglycerol kinases in signal transduction. Chem Phys Lipids 1999;98:95–108. van Blitterswijk WJ, Houssa B. Properties and functions of diacylglycerol kinases. Cell Signal 2000;12:595–605. Wakelam MJO. Diacylglycerol—when is it an intracellular messenger? Biochim Biophys Acta 1998;1436:117–26. York JD, Guo S, Odom AR, Spiegelberg BD, Stolz LE. An expanded view of inositol signaling. Adv Enzyme Regul 2001;41:57–71.