|Home | About | Journals | Submit | Contact Us | Français|
Expression of a dominant negative atypical protein kinase C (aPKC), PKCζ, prevents nuclear translocation of extracellular regulated kinase 2 (ERK-2), p27 nuclear reduction, and DNA synthesis induced by estradiol in human mammary cancer-derived MCF-7 cells. aPKC action upstream of these events has been analyzed. In hormone-stimulated NIH 3T3 and Cos cells ectopically expressing human estrogen receptor alpha (hERα), aPKC is activated by phosphatidylinositol 3-kinase (PI 3-kinase) and, in turn, controls the Ras/MEK-1/ERK cascade. In MCF-7 and Cos cells stimulated by hormone, PI 3-kinase activates PKCζ by Thr410 phosphorylation. Serine phosphorylation of PKCζ is simultaneously induced. PKCζ activation leads to recruitment of Ras to a multimolecular complex that also includes hERα, Src, PI 3-kinase, and aPKC. We propose that PKCζ pushes Ras and the signaling complex close together in such a way that it facilitates the Src-dependent Ras activation. This activation is crucial for the interplay between estradiol-triggered signaling and cell cycle machinery.
Atypical protein kinase C (aPKC) isoforms, such as aPKCζ and aPKCι/λ, lack the Ca2+ binding C2 domain, and their C1 domain is unresponsive to diacylglycerol and phorbol ester. Therefore, these kinases are not activated by diacylglycerol or phorbol esters, whereas they react to such lipid mediators as phosphatidylinositol 3,4,5-triphosphate, a product of phosphatidylinositol 3-kinase (PI 3-kinase) activity (39). PKCζ is a target of PI 3-kinase (22), and aPKCs trigger various biological responses, among them proliferation, apoptosis, and differentiation (39). We have now analyzed the effect of estradiol on an aPKC in MCF-7 cells and in other cells made hormone responsive by the transient expression of human estrogen receptor alpha (hERα).
PKCζ is involved in epidermal growth factor (EGF)-induced activation of p70S6K in a PI 3-kinase-dependent manner (32). Similarly, PKCζ in concert with PI 3-kinase mediates Ras-independent extracellular-regulated kinase (ERK) activation by a Gi protein-coupled receptor (38). In addition, a dominant negative mutant of PKCζ impairs serum-induced MEK-1/ERK kinase activation (2), and PKCζ has been implicated in insulin-induced ERK activation in adipocytes (33). Thus, PKCζ controls EGF-induced ERK activation in a variety of cells, although the nature of this control has not yet been clarified. Here, we report that estradiol stimulates an aPKC in MCF-7 cells, which acts downstream of PI 3-kinase and participates in the regulation of the Ras/MEK-1/ERK cascade. A central role in the novel aPKC action on Ras is played by a multimolecular complex induced by the agonist-occupied hERα. In hormone-stimulated MCF-7 and Cos cells, PKCζ interacts with hERα and recruits Ras to the complex including hERα, PI 3-kinase, and Src. We propose that Ras recruitment is required for the previously reported Src-dependent activation of Ras (26).
p27kip1 (p27) belongs to the kinase inhibitory protein (KIP) family of cdk inhibitors that regulate cyclin-cdk complexes (36). The importance of p27 in G1/S progression of breast cancer cells is corroborated by the finding that antisense-mediated inhibition of p27 expression is sufficient to induce cell cycle entry of steroid-depleted cells (3). In addition, a large body of evidence demonstrates that p27 is regulated by mitogenic signal transduction pathways, including Ras-dependent activation of the mitogen-activated protein kinase pathway (1, 7, 16). Here, we show that estradiol activation of an aPKC in MCF-7 cells is involved in cytoplasm redistribution of p27, which allows G0-arrested cells to reenter the cell cycle. Taken together, our results clarify the role of aPKC in estradiol-activated signaling and cell cycle regulation.
cDNAs encoding wild-type (HEG0) and mutant (HE241G) forms of hERα were cloned into a pSG5 expression vector as described previously (40, 44). cDNA coding for A221-MEK-1 in pEXV3 was used (8). The wild-type and dominant negative (T410A) PKCζ isoforms were subcloned in pCDNA3 as previously reported (45). They were excised as an XbaI fragment, filled with Klenow fragments, and subcloned into the blunt end of either a pSG5 or Myc-tagged pSG5 vector. The latter vector was obtained by amplifying the Myc tag and a relative polylinker from pGBKT7 vector (Clontech, Palo Alto, Calif.). After digestion with Cac8I and BglII, the amplified fragment was subcloned into blunt BglII pSG5. The dominant negative (T410A) PKCζ was also excised as an XbaI fragment, filled by Klenow fragments, and subcloned into SmaI-pEGFP (C2; Clontech). All junctions were verified by sequencing. The wild-type ERK-2 was subcloned into a 5′ Myc-tagged pCMV-expressing plasmid as previously described (17). The Myc-His-tagged dominant negative Akt (K179M) in the pUSEAmp plasmid was from UBI (Lake Placid, N.Y.).
MCF-7 cells were grown and seeded onto gelatin-precoated coverslips as previously reported (5). Cells were made quiescent by steroid depletion according to the methods outlined in the same report. For ERK-2 translocation, cells were transfected with Superfect (QIAGEN GmbH) with 6 μg of purified plasmid expressing green fluorescent protein (GFP) with either sense or antisense dominant negative PKCζ. The Myc-tagged wild-type ERK-2 construct was cotransfected at 3 μg. Twelve hours later, cells were serum starved (in 0.5% charcoal-stripped serum) for 4 h and then left unstimulated or stimulated with 10 nM estradiol for 20 min. For the S-phase entry analysis, cells were transfected with 1 μg of purified plasmid expressing either Myc-tagged wild-type PKCζ or Myc-tagged dominant negative PKCζ. Eight hours later, cells were left unstimulated or stimulated with 10 nM estradiol for 18 h. For the p27 localization analysis, Myc-tagged dominant negative PKCζ, A221-MEK-1 and Myc-His-tagged dominant negative Akt (K179M) were transfected with 2 μg of purified plasmids. Eight micrograms of sheared salmon sperm DNA (Eppendorf) was included as a carrier. Twelve hours later, transfected cells were left unstimulated or stimulated with 10 nM estradiol for 8 h. For the specificity analysis of Myc-tagged dominant negative PKCζ, quiescent MCF-7 cells were transfected with 2 μg of Myc-tagged, wild-type or dominant negative PKCζ, alone or together with 2 μg of hemagglutinin (HA)-tagged Akt wild type. Twelve hours later, cells were left unstimulated or stimulated with 10 nM estradiol for 3 min, and lysates were used for Western blot analysis or immunoprecipitation experiments. Cos-7 cells were cultured and made quiescent by steroid depletion (27). They were transfected with Superfect with 3 μg of either pSG5 empty plasmid or hERα-expressing plasmid. The pCDNA3 plasmid expressing wild-type or dominant-negative PKCζ was cotransfected with 2 μg of plasmid. When indicated, 2 μg of either Myc-tagged wild-type PKCζ or Myc-tagged dominant negative PKCζ was cotransfected. After 24 h, transfected cells were either left unstimulated or stimulated with 10 nM estradiol for the indicated times. NIH 3T3 mouse fibroblasts were cultured and made quiescent as previously reported (5). The cells were transfected with 3 μg of the pSG5 empty plasmid, hERα (HEG0), or Δ250-303 ERα (HE241G)-expressing plasmid. The Δp85α-expressing plasmid was cotransfected into the cells at a concentration of 4 μg. After 24 h, transfected cells were either left unstimulated or stimulated with 10 nM estradiol for the indicated times.
Myc-His-tagged dominant negative Akt was stained as previously reported (6). The Myc-tagged wild-type ERK-2 was detected with diluted rabbit anti-Myc-tagged polyclonal antibody (UBI) (diluted 1:100 in phosphate-buffered saline [PBS]). Anti-rabbit Texas red-conjugated antibody (diluted 1:200) (Jackson Laboratories) was used as a secondary antibody. MEK-1 was visualized with rabbit anti-MEK-1 polyclonal antibody (diluted 1:200 in PBS) (C-18; Santa Cruz). Cells on coverslips were incubated with goat anti-rabbit Texas red-conjugated antibody (diluted 1:400 in PBS) (Jackson Laboratories). Myc-tagged PKCζ, either dominant negative or the wild type, was visualized with mouse anti-Myc-tagged monoclonal antibody (MAb; diluted 1:100 in PBS) (Clontech). Anti-mouse Texas red-conjugated antibody (diluted 1:200) (Jackson Laboratories) was used as a secondary antibody. Endogenous p27 was visualized as previously described (31) with the anti-p27 MAb (Transduction Laboratories). Cells on coverslips were then incubated with anti-mouse fluorescein isothiocyanate-conjugated antibodies (diluted 1:200 in PBS) (Calbiochem). In experiments with Myc-tagged PKCζ, rabbit anti-p27 polyclonal antibody (C-19; Santa Cruz) (diluted 1:100 in PBS containing 0.1% bovine serum albumin) was used. Anti-rabbit fluorescein isothiocyanate-conjugated antibody (diluted 1:200 in PBS) (Jackson Laboratories) was employed as a secondary antibody. To evaluate DNA synthesis, cells on coverslips were pulsed for 6 h with 100 μM bromodeoxyuridine (BrdU) (Boehringer). BrdU incorporation was analyzed as previously reported (5) with fluorescein-conjugated mouse MAbs to anti-BrdU (diluted 1:1 in PBS) (clone BMC 9318; Boehringer Mannheim). Nuclei were stained with Hoechst stain 33258 (Sigma), and coverslips were inverted and mounted with Mowiol (Calbiochem). Images were generated with a DMLB fluorescent microscope (Leica, Heerbrugg, Switzerland) with 40 and 100× lens objectives and processed with IM1000 (Leica) software.
Cell lysates were prepared as previously described (26), and protein concentration was measured with a protein assay kit (Bio-Rad, Richmond, Calif.). Equal amounts of cell lysates (concentration, 2 mg of protein/ml) were used for the immunoprecipitation and ERK-2 kinase assay (26). Src and p85-associated PI 3-kinase were immunoprecipitated from cell lysates as previously described (6, 27). Cell lysates were used at a concentration of 4 mg of protein/ml for the Raf/Ras binding domain (Raf-RBD) assay, and the Raf-RBD pullout assay was done with the Ras activation assay kit (UBI). PKCζ was immunoprecipitated from cell lysates (concentration, 2 mg of protein/ml) with 5 μg of rabbit anti-PKCζ polyclonal antibody (Gibco-BRL). When indicated, PKCζ was immunoprecipitated with 2 μg of rabbit polyclonal anti-PKCζ antibody (clone C-20; Santa Cruz). Myc-tagged PKCζ (wild type or dominant negative) was immunoprecipitated with 2 μg of mouse anti-Myc-tagged MAb (clone 9E10; Clontech). The HA-tagged Akt wild type was immunoprecipitated with 2 μg of mouse anti-HA MAb (Covance). PKCζ activity in the immunocomplexes was assayed as previously described (38) with myelin basic protein (MBP) as a substrate.
The glutathione S-transferase (GST)-HEG14 fusion protein was produced, extracted, and purified on glutathione-agarose beads (Fluka) as previously reported (24). The matrix with the adsorbed fusion protein was used for the protein-protein interaction assay. Coupled in vitro transcription-translation reactions were used to produce 35S-labeled K-Ras and PKCζ in rabbit reticulocyte lysate (Promega). The protein-protein interaction assay was performed as previously described (24). Eluted proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and revealed by fluorography.
All electrophoresis and immunoblotting procedures have been described previously (27). The anti-phospho-PKCζ/λ (Thr410/Thr403), phospho-PKCδ/θ (Ser643/Ser66), phospho-PKCα/βII (Thr638/Thr641), phospho-PKC (pan), and phospho-Akt (Ser473) antibodies were obtained from Cell Signaling. Src was revealed with the mouse anti-Src MAb (clone 327; Calbiochem); hERα and Δ250-303 ERα were detected with the H222 rat anti-ER MAb (11); p85-associated PI 3-kinase was immunoblotted as previously described (6). Ras was detected with anti-pan Ras antibody (Calbiochem), PKCζ was revealed with rabbit anti-PKCζ polyclonal antibody (C-20; Santa Cruz), PKCδ was immunoblotted with the rabbit anti-PKCδ polyclonal antibody (C-20; Santa Cruz), and PKCα was revealed with the rabbit anti-PKCα polyclonal antibody (H-300; Santa Cruz). Phosphorylated ERK (phospho-ERK) and ERK were revealed using the mouse anti-phospho-ERK MAb (E4; Santa Cruz) and the rabbit anti-ERK polyclonal antibody (C-20; Santa Cruz). The Myc-tagged PKCζ as well as the Myc-tagged ERK-2 were detected with the mouse anti-Myc-tagged MAb (Clontech), the HA-tagged Akt was detected with the mouse anti-HA-tagged MAb (Covance), and the phosphoserine Western blotting kit (Zymed) was used to detect phospho-Ser proteins. Immunoreactive proteins were detected with the ECL detection system (Amersham, Little Chalfont, United Kingdom) according to the manufacturer's instructions.
Figure Figure1A1A shows that BrdU incorporation of estradiol-stimulated MCF-7 cells were unaffected by the indolocarbazole GO6976, a compound that inhibits conventional PKCs but has no effect on the Ca2+-independent PKCs, including aPKCs (23). In an attempt to establish the role of aPKC in estradiol-stimulated DNA synthesis, we verified the effect of a dominant negative form of PKCζ (T410A-PKCζ) (45) on the estradiol-stimulated S-phase entry of MCF-7 cells. The construct was subcloned into a Myc-tagged pSG5 plasmid to facilitate the identification of the overexpressed mutant, and then it was transiently transfected into quiescent cells. The cells were then left unstimulated or stimulated with 10 nM estradiol, labeled in vivo with BrdU, fixed, and stained. Multiple coverslips from different experiments were analyzed. The number of BrdU-positive cells expressing Myc-tagged T410A PKCζ was compared to the number of BrdU-positive nontransfected cells from the same coverslips. Data were pooled and statistically analyzed. It was found that overexpression of the dominant negative PKCζ strongly inhibited estradiol-induced S-phase entry of MCF-7 cells (Fig. (Fig.1A).1A). As a control, quiescent MCF-7 cells were transfected with a construct expressing Myc-tagged wild-type PKCζ. Overexpression of the Myc-tagged wild-type PKCζ did not affect entry into S phase (Fig. (Fig.1A).1A). Similarly, overexpression of the Myc-tagged pSG5 empty plasmid did not modify BrdU incorporation of MCF-7 cells, irrespective of the presence of estradiol (Fig. (Fig.1A1A legend).
In quiescent MCF-7 cells, nuclear accumulation of the transfected Myc-tagged ERK-2 wild-type reached a peak after 20 min of estradiol treatment (results not shown). Since PKCζ controls EGF-induced ERK activation in various cells, we analyzed the effect of the dominant negative PKCζ on estradiol-induced ERK-2 translocation in MCF-7 cells. The construct was subcloned in sense and antisense directions into the pEGFP plasmid and then transiently transfected into quiescent cells with the Myc-tagged wild-type ERK-2. After 20 min of estradiol treatment, cells were fixed and stained. Images of representative fields are shown in Fig. Fig.1B.1B. Upon estradiol stimulation, Myc-tagged wild-type ERK-2 was localized mainly in the nuclear compartment of cells expressing the GFP-antisense dominant negative PKCζ (Fig. (Fig.1B,1B, top) and mostly in the cytoplasm of cells expressing the GFP-sense dominant negative PKCζ (Fig. (Fig.1B,1B, bottom).
The specificity of the dominant negative PKCζ in MCF-7 cells was then analyzed. Overexpression of Myc-tagged dominant negative PKCζ (Fig. (Fig.1C,1C, top) did not interfere with either phosphorylation or expression of other PKC isotypes. Nevertheless, overexpression strongly interferes with ERK activation by estradiol. We tested whether dominant negative PKCζ would interfere with Akt activation in another set of experiments. Following an approach described previously (32), we found that overexpression of dominant negative PKCζ did not modify Akt activation in a cotransfection experiment with HA-tagged Akt (Fig. (Fig.1C,1C, bottom). Unlike ERK activation, phosphorylation of either PKC isotypes or Akt was unaffected by overexpression of the dominant negative we used.
p27 nuclear localization inhibits cell cycle progression (30) and estradiol causes redistribution of a p27 fraction to cytoplasm of MCF-7 cells (31). Immunofluorescence analysis of p27 in MCF-7 cells shows that p27 was localized in the nuclear compartment of unstimulated cells, whereas 8 h of estradiol stimulation induced a diffuse cytoplasmic staining of p27 (Fig. (Fig.2A).2A). Therefore, we analyzed the endogenous p27 localization of unstimulated or estradiol-stimulated MCF-7 cells in response to the expression of either the kinase-dead MEK-1 (Ser221 changed to alanine; A221-MEK-1) (8), the Myc-tagged-PKCζ dominant negative, or the kinase-dead Akt (Myc-His-tagged Akt K−; Lys 179 changed to methionine) (18). Quiescent cells were transiently transfected with the indicated construct and left unstimulated or stimulated with estradiol. The endogenous p27 localization in cells scored for overexpression of either A221-MEK-1, Myc-tagged dominant negative PKCζ, or Myc-His-tagged dominant negative Akt was analyzed. The data were compared with results obtained with nontransfected cells from the same coverslips (Fig. (Fig.2B).2B). p27 was localized in about 50% of the nuclei of quiescent, untransfected MCF-7 cells, and similar values were tabulated for quiescent, transfected cells. Treatment with 10 nM estradiol, which allows S-phase entry of MCF-7 cells, reduced the level of p27 nuclear labeling. Interestingly, overexpression of the Myc-tagged dominant negative PKCζ or the kinase-dead A221-MEK-1, which inhibits the estradiol-stimulated S-phase entry of MCF-7 cells (Fig. (Fig.1A1A and Fig. Fig.22 legend), restored the nuclear localization of p27. Irrespective of the presence or absence of hormone, overexpression of the kinase-dead Akt did not affect p27 nuclear localization. A previous study showed that kinase-dead Akt efficiently inhibits the S-phase entry of hormone-stimulated MCF-7 cells through inhibition of cyclin D1 transcription (6).
In conclusion, the results of our experiments (Fig. (Fig.11 and and2)2) show that aPKC is involved in S-phase entry, ERK-2 nuclear translocation, and p27 nuclear localization regulated by estradiol in MCF-7 cells.
Since PI-3 kinase controls ERK activation through aPKC in a variety of cells (2, 32, 33, 38), we analyzed the effect of estradiol on aPKC activity; we also analyzed the effect of the PI 3-kinase inhibitor, LY294002, on estradiol activation of aPKC and ERK-2 in MCF-7 cells. The results shown in Fig. Fig.33 indicate that hormonal treatment rapidly stimulates MBP-phosphorylating activity, specifically immunoprecipitated by antibody raised against either aPKC (Fig. (Fig.3A)3A) or ERK-2 (Fig. (Fig.3B).3B). LY294002 prevented hormonal stimulation of aPKC and ERK-2, whereas GO6976 had no effect (Fig. (Fig.3).3). These data show that estradiol stimulates aPKC activity in MCF-7 cells and that PI 3-kinase controls aPKC and ERK-2 activities in the same cells. The lack of effect shown with GO6976 indicates that conventional kinases do not interfere with this regulatory process.
To further address the role of PI 3-kinase in the estradiol activation of aPKC and its dependent signaling, we used NIH 3T3 fibroblasts. They represent a well-established model to analyze signaling pathway activation as well as the signaling-dependent DNA synthesis triggered by steroid hormones (5, 6). Cells were transiently transfected with hERα-expressing plasmid (40) in the presence or absence of a construct expressing a dominant negative p85 regulatory-adapter subunit of PI 3-kinase (Δp85α) (9). Quiescent cells were then challenged with estradiol, and overexpressed proteins were revealed by Western blotting with the appropriate antibodies (Fig. (Fig.3C).3C). In fibroblasts transfected with hERα, coexpression of Δp85α inhibited the hormone stimulation of ERK-2 activity (Fig. (Fig.3D).3D). Interestingly, such a dominant negative also abolished the estradiol-induced Ras activation evaluated by the Ras pullout assay (Fig. (Fig.3E).3E). As was observed with MCF-7 cells, estradiol stimulated aPKC activity of hERα-expressing fibroblasts, and such an activation was abolished by Δp85α coexpression (Fig. (Fig.3F3F).
NIH 3T3 fibroblasts, which do not express ERα or β (4, 5), were then transiently transfected with either wild-type hERα or its transcriptionally inactive mutant (Δ 250-303 ERα; HE241G) (40, 44). ERα transcriptional activity is not a prerequisite for estradiol stimulation of aPKC activity (Fig. (Fig.3G).3G). Kinase activation by estradiol was equally strong in NIH 3T3 fibroblasts, whether they expressed hERα or its mutant. It is noteworthy that although this mutant does not bind DNA (44), it stimulates signaling pathway-dependent S-phase entry of NIH 3T3 fibroblasts (5).
In conclusion, p85α-associated PI 3-kinase controls estradiol-induced activation of aPKC, Ras, and ERK-2.
After the transient expression of steroid receptors, Cos cells respond to steroids by activating signaling pathways (4, 24, 26, 27). Therefore, we used these cells to analyze the role of aPKC in estradiol-stimulated Ras and ERK-2 activation. Cos cells expressing hERα were cotransfected with a plasmid expressing either the wild-type or dominant negative PKCζ. Ectopically expressed proteins were revealed by immunoblot analysis of lysates (Fig. (Fig.4A).4A). As was observed with NIH 3T3 fibroblasts (Fig. (Fig.3)3) and MCF-7 cells (26), estradiol strongly activates Ras (and ERK-2 (Fig. (Fig.4B).4B). Coexpression of the dominant negative PKCζ abolished this effect, thereby indicating that Ras and ERK-2 activities are regulated by the estradiol-activated aPKC.
The finding that Src and p85 coimmunoprecipitate with ERα in estradiol-treated MCF-7 cells after direct interaction with the receptor suggested that this multimolecular complex is responsible for the simultaneous activation of the two main signaling pathways induced by estradiol, as well as for the cross talk between Src and PI 3-kinase (6). Our experiments showed that aPKC and Ras are novel components of the putative single complex. In the results shown in Fig. Fig.5,5, lysates from MCF-7 cells treated with estradiol in the presence or absence of the antiestrogen ICI 182,780 were immunoprecipitated with an antibody against p85 or Src. The immunoprecipitates were examined by Western blot analysis with the appropriate antibodies. The results were similar, irrespective of the antibody used. Estradiol triggered association of ERα, p85, Src, and aPKC. The antagonist ICI 182,780 prevented this association, which was undetectable in control immunoprecipitates.
To further investigate the nature of the ERα/aPKC/Ras complex assembly, we used anti-aPKC antibodies in immunoprecipitation of MCF-7 cell lysates. In addition to demonstrating the association of aPKC with ERα and Ras, Western blot analysis of immunoprecipitates with anti-phospho-PKCζ/λ (Thr410/Thr403) antibody revealed hormone-stimulated aPKC activation (Fig. (Fig.6A).6A). At the same time aPKC activation occurred, estradiol induced serine phosphorylation of aPKC. The antagonist ICI 182,780 abolished the hormone-induced aPKC activation as well as its phosphorylation on serine. At the same time, ICI 182,780 inhibited the association of ERα with aPKC and Ras (Fig. (Fig.6A),6A), implying that aPKC activation and serine phosphorylation are required for association. Similar results were obtained by Western blot analysis of complexes immunoprecipitated with anti-ERα antibody (results not shown).
The role of aPKC in ERα-Ras association was further investigated in Cos cells cotransfected with ERα and either the Myc-tagged wild-type PKCζ or the Myc-tagged dominant negative PKCζ. The results shown in Fig. Fig.6B6B demonstrate that, like in MCF-7 cells, Ras was coimmunoprecipitated with aPKC by anti-ERα antibodies in Cos cells expressing wild-type PKCζ and challenged with estradiol. An antagonist, ICI 182,780 or OH-tamoxifen, abolished the association of Myc-tagged wild-type PKCζ and Ras with ERα (Fig. (Fig.6B).6B). No coimmunoprecipitation of either aPKC or Ras in lysates from cells expressing the Myc-tagged dominant negative PKCζ was detected.
In another set of experiments, Cos cells were cotransfected with ERα and either the Myc-tagged wild-type PKCζ or the Myc-tagged dominant negative PKCζ. Analysis of immunoprecipitates with the anti-Myc-tagged antibody revealed that, as in MCF-7 cells, estradiol significantly stimulates Thr410 PKCζ phosphorylation, as well as serine phosphorylation of Myc-tagged PKCζ (Fig. (Fig.6C).6C). Expectedly, estradiol failed to activate the phospho-Thr410 PKCζ of Cos cells transfected with the Myc-tagged dominant negative PKCζ, because of its replacement of Thr410 by alanine (T410A-PKCζ) (45). In fact, such a mutant cannot be phosphorylated by phosphoinositide-dependent kinase 1 (PDK1) in a PI 3-kinase-dependent fashion (22). Interestingly, expression of such a dominant negative abolished the estradiol-stimulated serine phosphorylation of the Myc-tagged dominant negative PKCζ, suggesting that the kinase undergoes autophosphorylation on serine. It also abolished the hormone-induced ERα/aPKCζ/Ras complex assembly, which, in contrast, was detectable after estradiol treatment of Myc-tagged wild-type PKCζ-expressing cells (Fig. (Fig.6C).6C). Noteworthily, in the same experiment, neither serine, threonine, tyrosine phosphorylation of Ras, nor tyrosine phosphorylation of wild-type PKCζ was detected (results not shown), although the tyrosine phosphorylation of wild-type PKCζ in nerve growth factor-stimulated PC12 cells has been previously described (reference 43 and references therein).
Finally, pulldown experiments with GST-HEG14 fusion protein (HEG14 is the C-terminal domain of ERα, which includes the hormone BD) showed hormone-dependent interaction between 35S-labeled PKCζ synthesized in reticulocyte lysate and purified GST-HEG14 (Fig. (Fig.6D,6D, lane 4). Furthermore, the lack of interaction between 35S-labeled Ras and GST-HEG14 observed in the absence of 35S-labeled PKCζ (Fig. (Fig.6D,6D, lane 2), together with the hormone-dependent interaction of 35S-labeled Ras and GST-HEG14 in the presence of PKCζ (Fig. (Fig.6D,6D, lane 6), indicates that this aPKC is required for Ras interaction with GST-HEG14. Under the same conditions as those of the pulldown assay, phosphorylation of PKCζ in Thr410 and serine in reticulocyte lysates was detected (data not shown).
Collectively, the data shown in Fig. Fig.66 indicate that hormone treatment of MCF-7 cells or transfected Cos cells rapidly stimulates the PI 3-kinase-dependent PKCζ phosphorylation in Thr410 and serine. This event triggers the association of PKCζ/Ras with ERα.
We recently described a complex, nontranscriptional effect of sex steroid hormones that, like growth factors, stimulate a network of signal-transducing pathways. In cell lines derived from human breast and prostate cancers or in cell lines transiently expressing steroid receptors, exposure to steroids rapidly induces association of steroid receptors with Src and the p85α-adapter subunit of PI 3-kinase (4, 6, 24, 26, 27). This association is direct and simultaneous with activation of these two pathways, which drive into S-phase resting cells (4, 5, 6, 11, 24). The aim of this study was to further define the effect of estradiol on signaling network activation. Our data demonstrate that aPKC is involved in estradiol-activated signaling. Its action depends on PI 3-kinase, as shown by experiments with the PI 3-kinase inhibitor, LY294002 and the dominant negative Δp85α. Analysis of phospho-threonine 410 shows that the estradiol-stimulated PI 3-kinase/PDK1 pathway (22) activates PKCζ in MCF-7 cells and transfected Cos cells. In such cell types, estradiol stimulation of aPKC controls the ERK activation. The rapid activation of PKCζ does not require ERα transcriptional activity, as shown by experiments with NIH 3T3 fibroblasts that ectopically expressed a transcriptionally inactive form of ERα. This is in line with our view that classic steroid receptors acting outside the nuclear compartment are responsible for the hormone-triggered signaling pathway activation and related biological effects (4, 5, 6, 24, 26, 27). In apparent contrast with our findings, it has been reported that receptor signaling activity has no role in the proliferation of MCF-7 cells (21, 42). It has been observed that estren, a synthetic compound which stimulates signaling pathways in osteoblasts through either ER or androgen receptor (AR) (21), is unable to increase the MCF-7 cell number. No evidence, however, of signaling activation in MCF-7 cells by estren was presented in this report. In addition, it has been reported that a cis Decoy against estrogen responsive element inhibits DNA synthesis of estradiol-stimulated MCF-7 cells without affecting the hormonal ERK activation (42). However, reduced ERK activation was clearly detectable in cis Decoy-treated cells compared to those treated under the same conditions with the corresponding Scramble. It is thus difficult to draw conclusions from these reports. The results of experiments with human cancer-derived cell lines stimulated with sex steroid hormones and treated with signaling effector inhibitors, together with the results of experiments with NIH 3T3 cells ectopically expressing receptor mutants, prove that nongenomic action of steroid receptors is needed for the S-phase entry of these cells (4, 5, 6, 11, 24). That signaling activation is required for hormone-triggered DNA synthesis is highlighted by the recent identification of the classic mouse androgen receptor in NIH 3T3 fibroblasts. Because of its low expression level, the mouse androgen receptor is unable to enter nuclei and activate transcriptional machinery. Nevertheless, it activates signaling pathways upon androgen stimulation of cells, thus promoting S-phase entry and cytoskeleton changes of cells (4). We also found that inhibition of the rapid ERK-2 activation reduces the estradiol-induced S-phase entry of resting MCF-7 cells (25). On this basis, it is possible that while rapid signaling activation is required for G1/S transition, slow receptor transcriptional activity is involved in subsequent steps of cell proliferation. There is compelling evidence that in different cell types and under different conditions, hormone activation of signaling pathways triggers such diverse effects as neuroprotection, nitric oxide production, and survival (15, 20, 37) in addition to DNA synthesis. These findings indicate that activation of signaling pathways plays a key role in various aspects of steroid hormone action.
Nuclear translocation of ERK-2 is a crucial step in signal transduction that leads to gene expression and promotes cell cycle reentry upon mitogenic stimulation (17). The data reported herein show that dominant negative PKCζ prevents the nuclear translocation of ERK-2, which is rapidly induced by estradiol in MCF-7 cells. Such an event could inhibit the estradiol-elicited S-phase entry of the same cells. aPKC has been reported to control the ERK-2 activity induced by various stimuli (28, 33, 38), but how this control occurs is obscure. Our results with hormone-treated cells add novel pieces of information to this point. Our data indicate that ERK-2 activation by aPKC is Ras dependent. The mechanism of PKCζ action on Ras in different experiments has been analyzed. In MCF-7 and Cos cells, estradiol simultaneously stimulates kinase activation by Thr410 phosphorylation and its phosphorylation on serine. Activation and serine phosphorylation are required for the ER/aPKC/Ras complex assembly. Indeed, inhibition of estradiol-induced Thr410 and serine phosphorylation of aPKC by antiestrogen or expression of the dominant negative form of PKC prevents this association. The interaction of aPKCs with several proteins, including Src and Ras, has previously been described (10, 35). It has also been reported that phosphorylation of PKCζ/ιλ regulates protein-protein interactions as well as their intracellular movement (43). We previously observed that Src phosphorylates the p46 Shc and the p190 GAP-associated protein and upregulates Ras (26). We also reported that binding of signaling effectors such as Src and PI 3-kinase to ERα triggers their activation (6, 24). We now observe that PKCζ is activated by the ER-associated PI 3-kinase. Such an activation is responsible for PKCζ/Ras recruitment to the ER/Src/PI 3-kinase signaling cassette. Phosphorylation of PKCζ might modify the kinase surface, thereby allowing association of PKCζ with Ras and ERα. Thus, PKCζ might act as an adapter molecule that, once phosphorylated, couples Ras to the ER-mediated signal complex and makes Src-dependent activation of Ras possible. This view is supported by pulldown experiments showing that PKCζ interacts with ERα, whereas Ras interacts only with a PKCζ-associated receptor. The possibility that a single complex is involved in the hormone action described here is suggested by coimmunoprecipitation of the same proteins by antibodies directed against different members of the complex as well as by the common hormone-regulated mechanism for the complex assembly. Indeed, estradiol induces direct association of various signaling effectors (p85, Src, and PKCζ) with ERα, and steroid antagonists prevent the complex assembly.
Signaling pathways regulate, albeit by different mechanisms, the inhibitory function of p27 (12, 13, 19). Modulation of cell cycle inhibitors by aPKC is a novel, though not unexpected, finding. Indeed, PDK1-dependent activation of aPKC during insulin stimulation leads to p21Cip degradation (34). Here, we show that expression of dominant negative PKCζ induces nuclear accumulation of p27 in estradiol-treated MCF-7 cells, thereby inhibiting cell cycle progression. Interestingly, inhibition of MEK-1 also prevents the effect of estradiol on the p27 nuclear localization in MCF-7 cells. This confirms the findings of a previous report (14) and further shows that MEK-1 and PKCζ act on the same pathway that regulates p27 localization. Thus, such a pathway controls a different nuclear target from that of the hormone-activated Akt (6). Although Akt activation has been implicated in p27 localization of asynchronous MCF-7 cells (41), we found that estradiol-activated Akt does not affect p27 localization, whereas it regulates estradiol-induced cyclin D1 transcription in MCF-7 cells (6). It has also been reported that the presence of cyclin D1 or c-Myc is adequate for inducing the estradiol-induced S-phase entry of MCF-7 cells and that cyclin D1 and c-Myc both converge to shift the balance toward more active Cdk-2/cyclin E complexes (29). Our data (reference 6 and the present study) indicate that different cell cycle regulators, targeted by multiple effectors, are needed for estradiol-induced G1/S progression of responsive cells. Experimental conditions, such as overexpression of nuclear targets and use of estradiol antagonists to synchronize MCF-7 cells (29), could explain these differences. Estradiol regulation of p27 through aPKC could have important implications in hormone-dependent breast cancer therapy, since the rescue of p27 protein levels in the nuclear compartment with its consequent inhibitory function by aPKC targeting could restrain the growth of hormone-dependent human mammary cancers.
The model in Fig. Fig.77 is a schematic representation of the signaling cascade activated by the multimolecular complex assembled in the presence of estradiol and the role of PKCζ in this cascade.
We thank P. Chambon, M. H. Cobb, J. Downward, and P. Parker for generously providing plasmids. We also thank E. V. Avvedimento for kindly providing the Ras wild type in pGEX and I. Gout for the HA-tagged Akt wild type in pSG5. The H222 anti-ERα MAb was a gift of Abbott Laboratories (Abbott Park, Ill.). Zeneca (Milan, Italy) provided the antiestrogens, ICI 182,780 and OH-tamoxifen. The technical assistance of Flavia Vitale is acknowledged, and we are grateful to Jean Ann Gilder for editing the text.
This work was supported by grants from the Associazione Italiana per la Ricerca sul Cancro, the Ministero dell'Università e della Ricerca Scientifica (Cofinanziamenti 2002 and 2003; FIRB 2001), and the Ministero della Salute (Programmi Speciali; Lgs 229/99). M.L. is the recipient of a Fondazione Italiana per la Ricerca sul Cancro fellowship.
We declare that we do not have competing financial interests.