Characterization of a G protein-activated phosphoinositide 3-kinase in vascular smooth muscle cell nuclei.

Recent studies highlight the existence of an autonomous nuclear polyphosphoinositide metabolism related to cellular proliferation and differentiation. However, only few data document the nuclear production of the putative second messengers, the 3-phosphorylated phosphoinositides, by the phosphoinositide 3-kinase (PI3K). In the present paper, we examine whether GTP-binding proteins can directly modulate 3-phosphorylated phosphoinositide metabolism in membrane-free nuclei isolated from pig aorta smooth muscle cells (VSMCs). In vitro PI3K assays performed without the addition of any exogenous substrates revealed that guanosine 5'-(gamma-thio)triphosphate (GTPgammaS) specifically stimulated the nuclear synthesis of phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P(3)), whereas guanosine 5'-(beta-thio)diphosphate was ineffective. PI3K inhibitors wortmannin and LY294002 prevented GTPgammaS-induced PtdIns(3,4,5)P(3) synthesis. Moreover, pertussis toxin inhibited partially PtdIns(3,4,5)P(3) accumulation, suggesting that nuclear G(i)/G(0) proteins are involved in the activation of PI3K. Immunoblot experiments showed the presence of Galpha(0) proteins in VSMC nuclei. In contrast with previous reports, immunoblots and indirect immunofluorescence failed to detect the p85alpha subunit of the heterodimeric PI3K within VSMC nuclei. By contrast, we have detected the presence of a 117-kDa protein immunologically related to the PI3Kgamma. These results indicate the existence of a G protein-activated PI3K inside VSMC nucleus that might be involved in the control of VSMC proliferation and in the pathogenesis of vascular proliferative disorders.

Vascular smooth muscle cells (VSMCs) 1 play a central role in the fibroproliferative response during the development of ath-erosclerosis and of restenosis after angioplasty (1,2). Recently, we have demonstrated that phosphoinositide 3-kinase (PI3K) was essential for the progression of VSMCs throughout the G 1 phase of the cell cycle (3), implying that a better understanding of the PI3K signaling pathway might be of pathophysiological relevance. PI3K phosphorylates the D3 position of the inositol ring in phosphoinositides (PI) to generate the putative second messengers PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 (4). In addition to proliferation, the PI3K products have also been involved in cell transformation, apoptosis, vesicle trafficking, and cytoskeleton organization (5,6).
Three distinct classes of PI3Ks have now been identified on the basis of their in vitro substrate specificity, structure, and mode of regulation (7,8). The most studied are class I PI3Ks, which in vitro phosphorylate PtdIns, PtdIns(4)P, and PtdIns(4,5)P 2 , and display a preference for PtdIns(4,5)P 2 in vivo. Class I PI3Ks form a heterodimeric complex and are subdivided according to the adaptor protein associated with the catalytic subunit. Class IA PI3Ks consist of a 110-kDa catalytic subunit (p110 ␣, ␤, ␦) (9) and a 85-kDa adaptor protein (p85 ␣, ␤) (10) containing Src homology 2 (SH2) domains that link them to tyrosine kinase signaling. In contrast, class IB PI3K or PI3K␥ defines a G protein-coupled receptor-regulated PI3K (11). It is made of a p110␥ catalytic subunit and a p101 regulatory subunit unrelated to p85 (12). The p110␥ can be activated in vitro by both the ␣ and ␤␥ subunits of heterotrimeric G proteins (11)(12)(13). This stimulation is considerably enhanced by the p101 adaptor (12).
Moreover, there is now considerable evidence that a nuclear PI cycle, apart from that occurring in the plasma membrane, is involved in the regulation of nuclear functions (14). Indeed, it has been demonstrated that nuclei contain almost all the enzymes involved in the classical PI cycle, including kinases required for the synthesis of PtdIns(4,5)P 2 , phospholipases C, and diacylglycerol kinase (15,16). Furthermore, specific changes in the nuclear levels of PI have been implicated in both cell growth and differentiation (17)(18)(19). To date, information concerning the role and the regulation of nuclear PI3K are still very limited. Immunocytochemical and biochemical analyses demonstrate the presence of the p85␣ regulatory subunit in the nuclei of rat and human cells (20 -22) and the growth factordependent nuclear translocation of the p110␤ catalytic subunit in osteoblast-like cells (23), suggesting that class IA PI3Ks exist in the nucleus. Recently, a study based on the immunolocalization of epitope-tagged p110␥ in HepG2 cells reported that PI3K␥ translocates to the nucleus after serum stimulation (24). Since a nuclear G protein-regulated PI3K activity has not yet been demonstrated, we investigated whether GTP-binding proteins directly modulate 3-phosphorylated phosphoinositide (3-PI) metabolism in membrane-free nuclei isolated from pig * This work was supported by grants from the Association pour la Recherche contre le Cancer, the Ligue Nationale contre le Cancer, and the Fondation pour la Recherche Médicale. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Chemicals and Antibodies-All culture reagents were obtained from Life Technologies Inc. U73122, wortmannin, and LY294002 were obtained from Biomol (Plymouth Meeting, PA). GTP␥S, GDP␤S, and RNase-free DNase I were from Roche Molecular Biochemicals.
[␥-32 P]ATP (3000 Ci/mmol) was purchased from PerkinElmer Life Sciences. Monoclonal anti-p110␥, fluorescein isothiocyanate-conjugated anti-rabbit antibodies, and the enhanced chemiluminescence (ECL) system were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Polyclonal anti-p110␥ was kindly provided by S. Roche. Specific polyclonal anti-G␣ 0 antibodies directed against the last 10 amino acids of the common carboxyl-terminal sequence of the G␣ 01 and G ␣ 02 (anti-G␣ 0 (C-ter) or against the amino acid 291-302 of the ␣ subunit of G 01 (anti-G␣ 01 ) were obtained and characterized as previously described (25). Horseradish peroxidase-conjugated anti-rabbit/ mouse antibodies and polyclonal anti-p85␣ antibody were, respectively, supplied from New England Biolabs Inc. (Saint Quentin, France) and Upstate Biotechnology Inc. (Lake Placid, NY). Monoclonal anti-nucleoporin p62 antibody was from BD Transduction Laboratories (San Diego, CA). Lactate dehydrogenase and 5Ј-nucleotidase kits, polyclonal antitubulin antibodies, and all other reagents were obtained from Sigma. COS-7 cells overexpressing human p110␥ or p110␣ were a generous gift of A. Yart and P. Raynal.
Cell Culture and Isolation of VSMC Nuclei-VSMCs were prepared from 6-week-old pig thoracic aorta using the explant technique (26) and cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, as previously described (27). For all the experiments, VSMCs were used from the third to the sixth passage.
Growing VSMCs were washed twice with ice-cold calcium-and magnesium-free PBS and once with a hypotonic buffer containing 5 mM Tris-HCl, 1.5 mM KCl, 2.5 mM MgCl 2 , pH 7.4. All subsequent procedures were carried out at 4°C. Medium was then switched to hypotonic buffer supplemented with 200 M Na 3 VO 4 , 1 mM NaF, 1 mM EDTA, 1 mM EGTA, 100 M phenylmethylsulfonylfluoride, 10 g/ml each aprotinin, benzamidin, and leupeptin for 1 min, and VSMCs were lyzed by the addition of 1% Nonidet-P40 and 1% deoxycholic acid. Cells were allowed to swell for 1 min and were sheared by three passages through a 25-gauge needle. The cell lysate was layered over a 0.3 M/2 M sucrose discontinuous gradient and was centrifuged at 500 ϫ g for 15 min in polypropylene tubes pretreated with a siliconizing agent (Sigmacote) in a further attempt to reduce nuclei adsorption. Then, nuclei were recovered at the interface of 0.3 M/2 M sucrose and washed once with the assay buffer containing 1 mg/ml fatty-acid free bovine serum albumin, 40 mM Hepes, pH 7.5, 2 mM EGTA, 2 mM EDTA, 1 mM dithiothreitol, 50 mM NaCl, 4 mM MgCl 2 , 200 M Na 3 VO 4 , 1 mM NaF, and proteases inhibitors, as above. As to the yield of the nuclear isolation, an average of 0.5 ϫ 10 6 nuclei were obtained from 1 ϫ 10 6 cells. 1 ϫ 10 6 cells and 1 ϫ 10 6 nuclei contained 300 g and 30 g of proteins, respectively.
Lipid Kinase Assay-Nuclei were disrupted by sonication (10 kHz for 3 ϫ 1 s) using an ultrasonic cell disrupter (Branson Sonifier 250) and treated for 1 h at 4°C with 10 units/1.5 ϫ 10 6 nuclei RNase-free DNase I. All assays were conducted in a final volume of 100 l of assay buffer containing 5 ϫ 10 6 nuclei, 1 M thapsigargin (ATPase inhibitor), and 5 M U73122 (phospholipase C inhibitor). For experiments in the presence of PI3K inhibitors and a G i /G 0 inhibitor, nuclei were preincubated with 20 nM wortmannin or 10 M LY294002 and 5 ng/ml pertussis toxin, respectively, for 15 min on ice. When indicated, 100 M GTP␥S or 100 M GDP␤S was added for another 15 min. The assays were then started by the addition of 1 mM ATP (10 l) containing 65 Ci of [␥-32 P]ATP, and the 32 P incorporation was allowed for 15 min at 30°C under shaking. For exogenous lipid phosphorylation, 30 l of lipid vesicles containing 100 M PtdIns and 200 M phosphatidylserine were added 5 min before starting the assay. Reactions were stopped by the addition of 1360 l of chloroform/methanol (1:1, v/v), 300 l of 2 N HCl, and 280 l of 200 mM EDTA. Lipids were immediately extracted after the modified procedure of Bligh and Dyer (28). Lipids were then analyzed either on oxalate-coated thin-layer chromatography plates (Silica Gel 60, Merck) developed in isopropanolol:acetic acid:H 2 O (65:1:34) or by HPLC on a Partisphere SAX column (Whatman International Ltd, U. K.) after deacylation, as previously described (29). The synthesis of radioactive standard PtdIns(3)P, PtdIns(3,4)P 2 , and PtdIns(3,4,5)P 3 was performed from specific anti-p85␣ immunoprecipitates, essentially as described in Payrastre et al. (30).
Immunofluorescence-VSMCs were plated on glass coverslips, stimulated for 3 days with 10% fetal calf serum, and fixed in 70% ethanol for 30 min at 4°C. Cells were then permeabilized, and the DNA was denaturated for 30 min in a solution of 2 N HCl, 0.5% Triton X-100, 0.5% Tween 20 in PBS. Cells were washed extensively with PBS and were blocked for 1 h in PBS containing 0.5% bovine serum albumin (w/v) and 0.5% Tween 20. Coverslips were incubated with the anti-p85␣ antibody (1/20, overnight at 4°C) then with fluorescein isothiocyanate-conjugated anti-rabbit antibody (1/100, 1 h), mounted in Mowiol, and examined under epifluorescent illumination using a Zeiss microscope coupled to a Micromax camera (Princeton Instruments Inc.).
Electron Microscopy-The purified nuclei were immediately fixed in 3% glutaraldehyde in PBS, postfixed in osmium tetroxide, dehydrated in graded ethanol series, and embedded in Epon 812. Ultrathin sections were then cut (Reichert ultratome), placed on 300 mesh copper grids, counterstained with uranyl acetate and lead citrate, and examined in a Hitachi 300 transmission electron microscope.

Purity of VSMC Nuclear Preparations-Membrane-depleted
nuclei from pig aorta VSMCs were isolated by hypotonic shock combined with detergents. The purity of nuclear preparations was evaluated by biochemical and immunochemical analyses. Lactate dehydrogenase and 5Ј-nucleotidase activities, recognized as markers for cytoplasm and plasma membrane were, respectively, found to be 0.29 Ϯ 0.12% (n ϭ 3) and 0.18 Ϯ 0.08% (n ϭ 3) of the activity in the total homogenate. Furthermore, Western blot analysis using anti-tubulin antibody showed the absence of immunoreactivity to the cytoskeletal proteins in the purified nuclei (Fig. 1A). In addition, the nuclear fraction was highly enriched with nucleoporin p62, a protein of nuclear pore complex (Fig. 1B). Finally, electron microscopy analysis confirmed that the isolation procedure yielded nuclei of high purity (Fig. 1C). No appreciable morphological change of the nuclei was noted during the isolation procedure, the nucleolar structure was maintained, and the lysis procedure completely removed the nuclear envelope to leave the naked laminar layer and nuclear pore remnants (Fig. 1C, right panel).
VSMC Nuclei Contain a GTP-dependent PI3K Activity-Previous studies suggest the existence of a nuclear PI3K pathway (20 -24), but the intranuclear location and regulation of a PI3K activity has not clearly been demonstrated. Therefore, we first investigated whether membrane-free nuclei were able to produce 3-PI from endogenous precursors. We assessed nuclear PI3K activity in vitro by phosphate incorporation from [␥-32 P]ATP into inositol lipids. All assays were conducted in the presence of 2 mM EGTA and 5 M U73122, a phospholipase C (PLC) inhibitor, so that nuclear PLC activity would not interfere with PI3K activity.
As shown in Fig. 2A (upper panel), HPLC analyses revealed a peak of [ 32 P]phosphatidic acid, suggesting a residual PLC activity, but we never detected radiolabeled 3-PI. To look for the possible presence of a G protein-regulated PI3K (11-13), we next tested whether 100 M GTP␥S, a nonhydrolyzable GTP analogue, could trigger the production of 3-PI in isolated nuclei ( Fig. 2A, middle panel). Under this condition, we detected the synthesis of only one PI3K product, the PtdIns(3,4,5)P 3 , and the incorporation into [ 32 P]PtdIns(3,4,5)P 3 was 49073 Ϯ 2000 cpm/10 7 nuclei (n ϭ 3) (Fig. 3A). This peak coincided with the HPLC profile of pure [ 32 P]PtdIns(3,4,5)P 3 used as a control (Fig. 2B). In addition, the [ 32 P]phosphatidic acid peak was always present, and we never observed the production of any other inositol lipids, especially PtdIns(3)P and PtdIns(3,4)P 2 . No further stimulation of PtdIns(3,4,5)P 3 was observed with 200 M GTP␥S (data not shown). Moreover, 100 M GDP␤S was unable to activate PtdIns(3,4,5)P 3 synthesis, as presented in Fig. 2A (lower panel). These results were confirmed by TLC showing the specific activation of only PtdIns(3,4,5)P 3 synthesis by GTP␥S in isolated nuclei (Fig. 3B, lanes 1 and 3). A weak wortmannin-insensitive production of PtdIns(3,4,5)P 3 was observed in the absence of GTP␥S (Fig. 3B, lanes 1 and 2), but in general, the PI3K activity was too small to be detectable. These data strongly suggest that a nuclear GTP-dependent PI3K phosphorylates a pre-existing intranuclear pool of PtdIns(4,5)P 2 to produce PtdIns(3,4,5)P 3 . We next investigated the effects of two specific PI3K inhibitors on GTP␥S-induced nuclear PI3K activation. Both HPLC (Fig. 3A) and TLC studies (Fig. 3B) showed that accumulation of PtdIns(3,4,5)P 3 was reduced by 60 -80% after preincubation of nuclei with 20 nM wortmannin or 10 M LY294002, demonstrating that the nuclear GTP-dependent PI3K is sensitive to PI3K inhibitors in a classical range of concentrations.
Substrate Specificity of the Nuclear GTP-dependent PI3K-The above data suggest either that the GTP-dependent PI3K is selective for nuclear PtdIns(4,5)P 2 or that PtdIns and PtdIns(4)P present in the nuclear membrane were removed during nuclei isolation. To address this question, vesicles containing PtdIns/phosphatidylserine (100 M/200 M, final concentrations) were added to nuclear PI3K assays, and the formation of 32 P-labeled lipids derived from PtdIns was analyzed by HPLC (Fig. 4). Incubation of nuclei with exogenous PtdIns in the absence of GTP␥S resulted in the synthesis of [ 32 P]PtdIns(4)P, as already described (15), and also of [ 32 P]PtdIns(3)P (36,963 Ϯ 4105 cpm/10 7 nuclei, n ϭ 3) but not of any other 3-PI (Fig. 4A, upper panel, and Fig. 4B). This result suggests that membrane-free VSMC nuclei also contain a GTPindependent PI3K phosphorylating in vitro exogenous PtdIns to PtdIns(3)P but unable to phosphorylate the intranuclear pool of PtdIns(4,5)P 2 . In the presence of GTP␥S, [ 32 P]PtdIns(3)P synthesis was increased by 40% (57114 Ϯ 572 cpm/10 7 nuclei, n ϭ 3, p Ͻ 0.01), and under that condition, we also measured PtdIns(3,4,5)P 3 synthesis (Fig. 4A, lower panel). These results indicate that the nuclear GTP-dependent PI3K can phosphorylate both endogenous nuclear PtdIns(4,5)P 2 and exogenous PtdIns.
A 117-kDa PI3K␥ but Not p85␣ Is Expressed Inside VSMC Nuclei-We next performed immunoblot experiments to iden- tify the two PI3K isoforms expressed in isolated nuclei (Figs. 5 and 6). The existence of a nuclear PI3K that phosphorylates both PtdIns and PtdIns(4,5)P 2 in a GTP-dependent manner suggests that it could be a class IB PI3K. Indeed, we detected the presence of the catalytic subunit of the G protein-regulated PI3K by using two different anti-PI3K␥ antibodies (Fig. 5). The analysis with a monoclonal antibody against the amino-terminal region of human p110␥ revealed a 117-kDa protein in both the total cell homogenate (Fig. 5A, lane 1) and the nuclear fraction (Fig. 5A, lane 2), assayed here in similar amounts in terms of nuclei number. A positive control with human p110␥ overexpressed in COS-7 cells also showed a single band at 117 kDa (Fig. 5A, lane 3). Moreover, the p110␥ detection was specific, as the monoclonal antibody did not cross-react with human p110␣ overexpressed in COS-7 and as no signal was detected in VSMC nuclei when the primary antibody was omitted (Fig. 5A, lanes 4 and 5). The polyclonal anti-p110␥ antibody also confirmed the presence of a 117-kDa protein in VSMC nuclei (Fig. 5B). Unfortunately, as previously noted in HepG2 cells (24), all available antibodies were unable to detect endogenous PI3K␥ by immunofluorescence in VSMCs.
Having shown that a nuclear PI3K phosphorylates exogenous PtdIns in the absence of GTP␥S and considering that p85␣ has been reported in rat liver nuclei (20), we checked whether class IA PI3Ks would also be present in VSMC nuclei (Fig. 6). We used an anti-p85␣ antibody to detect this adaptor protein and loaded the same amount of proteins (15 g) from cell homogenate or from purified nuclei (Fig. 6A). In contrast to previous reports, p85␣ was undetectable in VSMC nuclei when compared with total cell homogenates, although 10 times as much nuclei were assayed. The absence of nuclear p85␣ expression was further confirmed by indirect immunofluorescence microscopy. As shown in Fig. 6B, the p85␣ labeling was evident as a ring at the perinuclear region and as fluorescent dots in the cytoplasm and the plasma membrane, whereas the nuclear interior remained unstained. These results strongly suggest that the heterodimeric PI3K p85␣/p110 is absent from our nuclear preparations but could be present in the nuclear envelope, whereas a 117-kDa PI3K ␥-like kinase is, significantly, located inside VSMC nuclei.
The Nuclear GTP-dependent PI3K Could Be Regulated by G Proteins-We next sought to determine whether nuclear heterotrimeric G i /G 0 proteins could be responsible for the nuclear PI3K activation by using PTX. Pretreatment of isolated nuclei with 5 ng/ml PTX inhibited about 50% of the GTP␥S-induced PI3K activity (Fig. 7A). Moreover, to identify PTX-sensitive G proteins inside VSMC nuclei, we performed immunoblots using specific antibodies directed against both G␣ 01 /␣ 02 subunits (anti-G␣ 0 (C-ter)) or against the ␣ subunit of G 01 (anti-G␣ 01 ). As shown in Fig. 7B, a protein of 42-43 kDa is recognized by both antibodies in our nuclear preparations. These results suggest that PTX-sensitive G proteins are present in VSMC nuclei and could be involved in nuclear PI3K activation. DISCUSSION In the present study, we provide the first evidence that a GTP-dependent PI3K generates the second messenger PtdIns(3,4,5)P 3 from a pre-existing nuclear pool of PtdIns (4,5)P 2 , directly within VSMC nucleus. Furthermore, we showed that this PI3K activity, which could be related to a PI3K␥, is coupled to nuclear G i /G 0 heterotrimeric G proteins.
PtdIns (3,4,5)P 3 Synthesis in VSMC Nuclei-We prepared highly purified VSMC nuclei stripped of their nuclear envelope. Our data demonstrated that the GTP␥S-responsive PI3K and the PtdIns(4,5)P 2 resist treatment with non-ionic detergents, suggesting that this enzyme and its substrate are tightly associated with non-membrane nuclear structures. In agreement with these results, PtdIns(4,5)P 2 was localized in nucleolusassociated heterochromatin in rat pancreas (31), and picomolesensitive mass assays have revealed that 35% of nuclear PtdIns(4,5)P 2 (about 30 pmol/mg of protein) remained in rat liver nuclei treated with Triton X-100 (32). In addition, recent data from Boronenkov et al. (33) provide new insights about the localization of PtdIns(4,5)P 2 within the nucleus. They demonstrated that PtdIns(4,5)P 2 is spatially organized to "nuclear speckles" in mammalian cells. Interestingly, nuclear speckles are separated from known membrane structures and contain pre-mRNA processing factors, and their morphology is tightly linked to the status of mRNA transcription. These observations suggest that speckles might function as centers for PI-signaling pathways in nuclei. However, PI3K activity was not addressed in these studies. In this respect, our finding of a nuclear G protein-regulated PI3K is of special interest, and we can speculate that phosphorylation of PtdIns(4,5)P 2 into PtdIns (3,4,5)P 3 might modulate speckle functions.

FIG. 7. Effects of PTX on GTP␥S-induced nuclear PtdIns(3,4,5)P 3 synthesis and expression of G␣ 0 proteins in VSMCs.
Membrane-depleted nuclei were isolated, and nuclear lipid kinase activity was assayed as described in Fig. 2. A, nuclei were pretreated for 15 min with 5 ng/ml pertussis toxin before incubation with 100 M GTP␥S. Lipids were extracted and analyzed by HPLC. Quantification of [ 32 P]PtdIns(3,4,5)P 3 synthesis is expressed as cpm/10 7 nuclei. Data for GTP␥S alone are the mean Ϯ S.E. from three independent experiments. Data for PTX represent the mean of two independent experiments. B, Western blotting analysis of heterotrimeric G␣ 0 protein. Protein from 1 ϫ 10 5 whole cells or 5 ϫ 10 5 purified nuclei were fractionated on 10% SDS-PAGE, transferred, and probed with polyclonal anti-G␣ 0 (C-ter) or anti-G␣ 01 antibodies. Experiments illustrated are representative of at least three distinct experiments. PIP 3 , PtdIns(3,4,5)P 3 .
The PtdIns(3,4,5)P 3 -dependent signaling pathways in the nucleus are still unknown. However, PtdIns(3,4,5)P 3 synthesis was associated with activation of the serine/threonine kinases, protein kinase C, and Akt/protein kinase B (34,35), and the nuclear translocation of these two PtdIns(3,4,5)P 3 effectors was demonstrated upon mitogenic stimulation (36,37). More recently, Neri et al. (38) show that nuclear PtdIns(3,4,5)P 3 production correlates both with p85␣/p110 PI3K and protein kinase C translocations to the nucleus of nerve growth factortreated PC12 cells. Thus, the G protein-activated PI3K could produce PtdIns(3,4,5)P 3 inside VSMC nuclei to recruit and/or activate downstream effectors. In this respect, we also observed the presence of protein kinase C and active Akt/protein kinase B in VSMC nuclei (data not shown). The nuclear targets of protein kinase C and Akt/protein kinase B only begin to be described. Indeed, a major component of the nucleolus, nucleolin, has been shown to be phosphorylated by protein kinase C (36), and Akt/protein kinase B has been reported to promote phosphorylation of the nuclear transcription factors CREB (cAMP-response element-binding protein) (39) and FKHR1 (forkhead in rhabdomyosarcoma 1) (40). Finally, evidence has been provided that PIP3BP, a PtdIns(3,4,5)P 3 -binding protein, is exported out of the nucleus by the expression of constitutively activated PI3K (41).
Nuclear GTP-dependent and -independent PI3Ks-In response to GTP␥S, we never detected nuclear PtdIns(3)P or PtdIns(3,4)P 2 synthesis from endogenous substrates, suggesting that PtdIns and PtdIns(4)P, which have been reported located in the nuclear membrane (14,32), were totally removed during nuclei isolation. On the other hand, the GTP-dependent PI3K might be PtdIns(4,5)P 2 -selective. To address this question, we added exogenous PtdIns in assays. We found that GTP␥S stimulated a PtdIns(3)P synthesis, demonstrating that the nuclear GTP-dependent PI3K uses PtdIns as a substrate in vitro. This result suggests that GTP-binding proteins activate a class IB PI3K in VSMC nucleus. Accordingly, immunoblot analysis revealed the presence of a 117-kDa catalytic subunit of G protein-regulated PI3K␥ inside VSMC nuclei, suggesting that the nuclear GTP-dependent PI3K could be PI3K␥ or a PI3K␥-like kinase. In this respect, Stephens et al. (12) purified two G protein-activated PI3Ks from pig neutrophils. Both were heterodimers composed of the 101-kDa regulatory protein and either a 120-kDa or a 117 kDa catalytic subunit. Only, the p120 cDNA was cloned and was shown to be highly related to PI3K␥. Furthermore, two recent reports demonstrated the presence of a PI3K␥-mediating ion channel stimulation in smooth muscle cells (42,43). However, the mechanism governing the nuclear targeting of the GTP-dependent PI3K is unclear. Nevertheless, Metjian et al. (24) showed that serum induces translocation of tagged p110␥ into HepG2 nuclei and suggested that p101 could regulate this relocation. In contrast with this observation, preliminary data from our laboratory seem to indicate that serum does not modify nuclear PI3K␥-like kinase expression. 2 We further showed that VSMC nuclei contain a PI3K activity able to generate PtdIns(3)P from exogenous PtdIns in the absence of GTP␥S. Immunoblots and immunofluorescence analyses failed to detect the adaptor protein p85␣ inside the VSMC nuclei, suggesting that the GTP-independent PI3K is different from p85␣/p110 PI3Ks (class IA). This conflicts with the data from Lu et al. (20) and Marchisio et al. (44), who found p85␣ in the nuclear matrix of rat liver and of differentiated HL60 cells, respectively. Such a discrepancy might be due to differences in cell type or in nuclei isolation. However, p85␣/p110 could be expressed in the nuclear envelope of VSMCs. We also cannot exclude the presence of a class IA PI3K containing the p85␤ adaptor protein or class II/III PI3K inside VSMC nuclei. These hypotheses are currently under investigation. However, it is noteworthy that this GTP-independent PI3K was unable to phosphorylate the pool of PtdIns(4,5)P 2 within the nucleus.
Our experiments with bacterial toxin and the nuclear identification of G␣ 01 allowed us to propose that PTX-dependent G proteins are present in the nucleus and participate in the activation of the GTP-dependent PI3K. In support of this hypothesis, the association between G␣ 0 and the mitotic spindle was observed in certain cancer cell lines (45). It has also been shown that growth factor-induced cell division is paralleled by the translocation of ␣ i to the nucleus (46). Moreover, G␤␥ subunits might also play an important role in GTP-dependent PI3K activation, since G␤␥ directly stimulates PI3K␥ activity in vitro (47), and a ␤␥ subunit-like activity has been reported in rat liver nuclei (48). Interestingly, PI3K␥ was implicated in both G␣ i -and ␤␥-mediated survival pathways elicited by G protein-coupled receptors in COS-7 cells (49).
Clearly, further investigations will be required to answer important points, in particular to clarify the role of the nuclear heterotrimeric G protein-activated PI3K in VSMC pathophysiology. It will be of special interest to study the regulation of this PI3K throughout the cell cycle. In this regard, the recent demonstration that serum-induced VSMC proliferation is mediated primarily via G␤␥ in vitro and that targeted inhibition of G␤␥ reduces intimal hyperplasia and limits restenosis in vivo (50) appears essential.