Molecular Cancer Research Targeting the PI3-Kinase Pathway in Cancer Bridging the Lab and the Clinic in Cancer Medicine
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zhao, L.
Right arrow Articles by Quelle, D. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zhao, L.
Right arrow Articles by Quelle, D. E.
Molecular Cancer Research 1:195-206 (2003)
© 2003 American Association for Cancer Research

Cell Cycle, Cell Death, and Senescence

Cyclin G1 Has Growth Inhibitory Activity Linked to the ARF-Mdm2-p53 and pRb Tumor Suppressor Pathways1

Lili Zhao1, Tina Samuels1, Sarah Winckler1, Chandrashekhar Korgaonkar1, Van Tompkins1, Mary C. Horne1,2 and Dawn E. Quelle1,2

1 Department of Pharmacology and 2 The Molecular Biology Graduate Program, College of Medicine, The University of Iowa, Iowa City, IA

Requests for reprints: Dawn E. Quelle, Department of Pharmacology, College of Medicine, The University of Iowa, 51 Newton Road, Iowa City, IA 52242. Phone: (319) 353-5749; Fax: (319) 335-8930. E-mail: dawn-quelle{at}uiowa.edu


   Abstract
Top
 Notes
 Abstract
Introduction
 Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 
Cyclin G1 is a p53-responsive gene that is induced in alternative reading frame (ARF)-arrested cells, yet its role in growth control is unclear. We tested its effects on growth and involvement in the ARF-Mdm2-p53 tumor suppressor pathway. We show that cyclin G1 interacts with ARF, Mdm2, and p53 in vitro and in vivo. At high levels, cyclin G1 induces a G1-phase arrest in mammalian cells that coincides with p53 activation. Conversely, lower levels of cyclin G1 lack intrinsic growth inhibitory effects yet potentiate ARF-mediated growth arrest. Notably, cyclin G1 is down-regulated by Mdm2 through proteasome-mediated degradation. These data suggest that cyclin G1 is a positive feedback regulator of p53 whose expression is restrained by Mdm2. Interestingly, growth inhibition by cyclin G1 does not require p53 but instead exhibits partial retinoblastoma protein (pRb) dependence. These findings reveal that cyclin G1 has growth inhibitory activity that is mechanistically linked to ARF-p53 and pRb tumor suppressor pathways.


   Introduction
Top
 Notes
 Abstract
 Introduction
Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 
Inactivation of the p53 tumor suppressor gene is the most frequent genetic event in human cancers (1). p53 is a checkpoint regulator that maintains genomic stability in the face of environmental and intracellular stresses, including hypoxia, DNA damage, and oncogene activation (2). Normally, p53 is a short-lived nuclear protein, but stress signals rapidly stabilize and activate p53 through post-transcriptional mechanisms, such as phosphorylation and acetylation (3). Activated p53 suppresses growth by transactivating genes that trigger growth arrest or apoptosis (4). Once cellular damage is repaired, p53 must be down-regulated to allow normal cell growth. This is primarily accomplished by Mdm2, a p53-responsive gene product that acts in a negative autoregulatory feedback loop to inactivate p53 (5). Mdm2 is an E3 ubiquitin ligase that blocks p53 function through direct binding, ubiquitination, and promotion of p53 nuclear export into cytoplasmic proteasomes (5).

ARF is an alternative reading frame product derived from the INK4a/ARF tumor suppressor locus on chromosome 9p21 (6, 7). It is the second most commonly inactivated gene in human cancer (8), and it blocks cellular transformation in response to activated oncogenes, such as Ras or Myc (9). As with mice lacking p53, specific disruption of ARF results in spontaneous tumor development (10, 11). p53 is the major effector of ARF-mediated growth inhibition, and ARF activates p53 by antagonizing Mdm2 (7, 9). Until recently, it was generally accepted that ARF neutralized Mdm2 activity by sequestering it within nucleoli, thereby allowing p53 to accumulate in the nucleoplasm and induce expression of growth inhibitory genes. However, recent studies showed that ARF can inhibit growth without relocalizing endogenous Mdm2 to nucleoli (12–14). Moreover, regions within the amino terminus of ARF were identified that were dispensable for Mdm2 binding and relocalization, but essential for its activation of p53 and inhibition of growth (14). Those findings suggested that other factors besides Mdm2 contribute to p53-dependent growth suppression by ARF. Consistent with that notion is the existence of multiple ARF signaling pathways. For instance, once p53 is activated, both p21-dependent and p21-independent pathways can contribute to the G1 and G2 arrest elicited by ARF. p21Cip1 is a p53-responsive gene and potent inhibitor of cyclin-dependent kinases (Cdks) that blocks phosphorylation of the retinoblastoma tumor suppressor protein, pRb (15). The consequent accumulation of active, hypophosphorylated pRb arrests cells in G1 phase and prevents S-phase entry. Although p21 is the primary downstream effector of ARF-mediated cell cycle arrest, a p21-independent pathway also exists that exerts a distinct biphasic growth arrest (16, 17). In addition, ARF can induce a delayed G1-phase growth arrest in cells lacking both p53 and Mdm2 (18). Importantly, regulators of the p21- and p53/Mdm2-independent pathways have yet to be identified.

We previously showed that cyclin G1 is induced by ARF in both p21-positive and p21-negative cells (14, 17). Cyclin G1 is a transcriptional target of p53 that contains two p53 binding sites within its promoter, and its up-regulation coincides with activation of p53 by various DNA-damaging agents (19–23). Cyclin G1 is also induced by transforming growth factor ß (TGF-ß), BMP-4, p63, and p73, and is often detected in cells and tissues lacking p53, indicating that it can be regulated through p53-independent pathways (21–26). Notably, there is no proven Cdk partner for cyclin G1. It has been found to interact with pRb, cyclin G1-associated kinase (GAK), Cdk5, and the regulatory B' and catalytic C subunits of protein phosphatase 2A (PP2A) (27–32). The physiological importance of those associations has not been determined, although there is suggestive new data concerning PP2A. Okamoto et al. (32) recently reported that cyclin G1 recruits PP2A to dephosphorylate Mdm2 and thereby regulate p53.

Apparently conflicting roles have been assigned to cyclin G1 in growth control. Some reports indicate that cyclin G1 promotes growth based on observations that its overexpression enhances the growth of certain cancer cell lines, whereas introduction of antisense constructs suppresses their growth (29, 33–35). Conversely, it has been suggested that cyclin G1 may have growth inhibitory activities. This is based on substantial yet largely correlative data showing that cyclin G1 expression is high in differentiated tissues and in G2-phase-arrested hepatocytes, is induced by DNA-damaging agents, and is up-regulated during TGF-ß- or BMP4-mediated growth arrest (23, 24, 36, 37). Moreover, it was recently shown that cyclin G1-/- mouse embryo fibroblasts (MEFs) are partially deficient in an irradiation-induced G2-M-phase checkpoint (38). Others found that overexpression of cyclin G1 had no effect on the growth properties of mouse fibroblasts, yet it sensitized those cells to tumor necrosis factor {alpha} (TNF{alpha})-mediated apoptosis (19, 23).

The data presented here implicate cyclin G1 as a regulator within the ARF-Mdm2-p53 and pRb tumor suppressor pathways. Moreover, our findings suggest that cyclin G1 has intrinsic growth inhibitory activity that is dependent on the magnitude of its expression. That observation may help to explain, at least in part, why there are conflicting reports concerning its role in growth control.


   Results
Top
 Notes
 Abstract
 Introduction
 Results
Discussion
 Materials and Methods
 Acknowledgements
 References
 
Cyclin G1 Has Intrinsic Growth Inhibitory Activity
We previously demonstrated that cyclin G1 protein expression is induced concomitantly with p53 activation in ARF-arrested mouse fibroblasts (14, 17). Given that ARF can inhibit growth independent of p53 and that cyclin G1 is regulated by multiple transcription factors, we tested whether up-regulation of cyclin G1 by ARF is p53 dependent. Cyclin G1 protein levels were assayed following introduction of ARF into cells expressing or lacking p53. Mouse NIH 3T3 fibroblasts (INK4a/ARF-null, wild-type p53) and triple-knockout (tko) MEFs lacking p53, Mdm2, and ARF were infected with retroviruses encoding mouse ARF or empty vector control. ARF causes a rapid p53-dependent G1- and G2-phase growth arrest in 3T3 cells versus a delayed G1-phase block in tko cells (14, 18). Immunoblots showed equivalent expression of ARF in both populations, yet cyclin G1 expression and induction was only evident in NIH 3T3 cells (Fig. 1 ). For comparison, we also examined cyclin G1 expression in human Narf cells, a derivative of U2OS osteosarcoma cells that express isopropyl-1-thio-ß-D-galactopyranoside (IPTG)-inducible ARF (39). Treatment with IPTG for 2 days resulted in complete growth arrest (data not shown) that coincided with induction of ARF, stabilization of p53, and up-regulation of cyclin G1 (Fig. 1). These data indicate that ARF-mediated induction of cyclin G1 is commonly observed and requires p53.



View larger version (36K):
[in this window]
[in a new window]
  		FIGURE 1. Cyclin G1 is induced by ARF in a p53-dependent manner. NIH 3T3 cells and tko MEFs were infected with vector (V) or ARF (A) retroviruses, whereas Narf cells were treated with (+) or without (-) IPTG. Western blotting was performed to measure the expression of ARF, p53, and cyclin G1 in whole cell lysates (50 µg per lane) from the indicated cells.

 
The p53-dependent up-regulation of cyclin G1 in ARF-arrested cells suggested that it might have intrinsic growth inhibitory activity. This was supported by preliminary experiments showing that ectopically expressed cyclin G1 induced a G1-phase arrest in Chinese hamster ovary (CHO) and 293 cells.2 To further test that idea, green fluorescent protein (GFP)-tagged cyclin G1 plasmids were expressed by transient transfection in U2OS and NIH 3T3 cells. Cells expressing GFP or GFP-cyclin G1 (GFP-G1) were collected by fluorescence-activated cell sorting, and the DNA content of each population was measured by staining with Hoescht dye or propidium iodide (PI). Fig. 2 shows that cells expressing high levels of GFP remained in cycle, similar to GFP-negative cells isolated from the same populations. In contrast, cells expressing GFP-G1 were dramatically arrested in the G1 phase of the cell cycle. In experiments using live U2OS cells stained with Hoescht, cells were simultaneously stained with PI to identify dead cells within the population. No increase in cell death was observed in the GFP-G1-positive cells versus those expressing GFP (data not shown), indicating that in this system, cyclin G1 does not initiate apoptosis.



View larger version (25K):
[in this window]
[in a new window]
  		FIGURE 2. High levels of cyclin G1 induce G1-phase growth arrest. Representative histograms showing cell cycle distributions of sorted GFP-positive (+) and GFP-negative (-) U2OS and NIH 3T3 cells following transfection with GFP or GFP-cyclin G1 (GFP-G1) plasmids. The percentage of cells in S phase for each population is denoted.

 
Cyclin G1 Potentiates ARF-Mediated Growth Arrest
To examine the significance of cyclin G1 up-regulation by ARF, we tested its contribution to ARF-mediated growth arrest in NIH 3T3 cells. Mouse fibroblasts were chosen for these studies because we routinely achieve nearly complete introduction of our genes of interest into the population via retroviral-mediated infection (14, 17). As controls, retroviruses encoding cyclin G1, empty vector, or mouse ARF were individually transduced into the ARF-null cells. Western blotting confirmed expression of the exogenous hemaglutinin (HA)-tagged cyclin G1 (Fig. 3A ), although levels achieved by retroviral infection were approximately 5-fold lower than that induced by ARF. Flow cytometric analyses revealed no significant effect of cyclin G1 on 3T3 cell growth, similar to vector control cells, whereas ARF induced a complete G1- and G2-phase growth arrest (Fig. 3B and C).



View larger version (45K):
[in this window]
[in a new window]
  		FIGURE 3. Cyclin G1 enhances ARF-mediated growth arrest. A. Equivalent amounts of total cellular protein from NIH 3T3 cells infected with the indicated retroviruses were analyzed by Western blotting for expression of ARF, cyclin G1 (HA-tagged form indicated by asterisk), p53, Mdm2, and B23 (loading control). B. Cell cycle distributions of infected NIH 3T3 cells from a representative experiment, in which V/A and G1/A represent sequential infections of vector or cyclin G1 plus ARF viruses. C. The relative percentage S phase for cells expressing the indicated viruses relative to vector control was calculated from three independent experiments. Asterisk, statistically significant difference between G1/A and V/A samples, as determined by a paired, two-tailed Student's t test analysis (P = 0.008); bars, SD. D. Cyclin G1 immunoblot assessing expression levels in an equivalent number of NIH 3T3 cells infected with vector, ARF, or HA-cyclin G1 retroviruses compared to NIH 3T3 cells transfected with GFP or GFP-G1. Transfected cells were analyzed (not sorted) by fluorescence-activated cell sorting, and 33% expressed GFP-G1. Asterisks indicate endogenous cyclin G1 (*), HA-cyclin G1 (**), and GFP-G1 (***).

 
The lack of growth arrest by cyclin G1 in these cells was surprising given its ability to block U2OS and NIH 3T3 cell growth in transfection experiments (Fig. 2). However, it was consistent with earlier studies showing that cyclin G1 was unable to initiate growth arrest or apoptosis when overexpressed in mouse fibroblasts (19). Interestingly, cell counts taken 2 days after infection showed that cyclin G1 expressors divided at a slightly faster rate (1.2-fold, P < 0.01) than vector control cells (Table 1). This finding was in accordance with studies showing that cyclin G1 modestly enhanced colony formation of human diploid fibroblasts (29). It seems most likely that an inability to achieve high levels of cyclin G1 expression in the retroviral system accounts for its lack of growth inhibitory activity. Indeed, a correlation between growth arrest and high-level expression of cyclin G1 was observed (Fig. 3D). NIH 3T3 cells that were arrested, either by transfection of GFP-G1 or introduction of ARF, expressed high levels of tagged or endogenous cyclin G1, respectively (Fig. 3D). Conversely, cells infected with cyclin G1 retroviruses continued to proliferate and expressed relatively low levels of exogenous cyclin G1. Thus, the magnitude of cyclin G1 expression correlated with its effects on cell growth.


View this table:
[in this window]
[in a new window]
  		Table 1. Low Levels of Exogenous Cyclin G1 Modestly Promote Cell Growth in NIH 3T3 Fibroblasts

 
Previous studies showed that cyclin G1 was unable to induce apoptosis when overexpressed in NIH 3T3 cells, but it potentiated cell death induced by TNF{alpha} (23). To test whether exogenous cyclin G1 could enhance ARF-induced growth arrest, NIH 3T3 cells were first infected with retroviruses encoding empty vector or cyclin G1, followed by a second round of infection with ARF retroviruses. Although this method resulted in essentially complete infection with ARF (at least 96% of cells in each population expressed ARF, as determined by immunofluorescence), reduced expression of ARF was routinely achieved. Consequently, a less robust arrest and up-regulation of p53 by ARF was observed (Fig. 3A and B). Under these conditions, the addition of exogenous cyclin G1 consistently potentiated the G1-phase growth arrest exerted by ARF (Fig. 3C). Although the effect was modest, it was highly reproducible and statistically significant when subjected to a Student's t test analysis (P = 0.008). Identical results were obtained in BrdUrd-incorporation assays, in which 40% of vec/ARF-infected cells incorporated BrdUrd compared to only 30% of G1/ARF-infected cells.

The above results indicated that cyclin G1 can contribute to ARF-mediated growth arrest. It is notable that the enhanced arrest of G1/ARF-expressing cells did not coincide with increased p53 stabilization (Fig. 3A) or transcriptional activation in reporter assays (negative data not shown). The slight increase in Mdm2 expression observed in G1 + ARF- versus Vec + ARF-infected cells was not reproducible in multiple experiments (see Fig. 3A), and we have no evidence for up-regulation of another p53 target, p21, in these cells (data not shown). Therefore, the effects of exogenous cyclin G1 appeared to be independent of p53, as one would expect given that its expression is normally downstream of p53.

Cyclin G1 Associates with ARF, Mdm2, and p53
Previous studies showed that cyclin G1 is a nuclear protein, ARF is nucleolar, and p53 and human Mdm2 (Hdm2) reside within the nucleoplasm (22, 36, 40). Therefore, we tested whether cyclin G1 might physically bind to those regulators. In vitro binding assays were performed using glutathione S-transferase (GST)-tagged cyclin G1 mixed with recombinant ARF, p53, or Mdm2 produced in Sf9 insect cells (Fig. 4A ). GST-cyclin G2 and GST-PP2AC (the C subunit of PP2A) were included for comparison. Cyclin G2 shares significant homology with cyclin G1, and the B' and C subunits of PP2A are known to associate with both G cyclins (24, 30–32). We found that GST-cyclin G1 bound to ARF, p53, and Mdm2, although the interaction with Mdm2 was reproducibly more efficient in repeated assays (Fig. 4A and data not shown). Interestingly, p53 associated equally well with cyclin G2 and more strongly with PP2AC than it did with cyclin G1. The interaction between ARF and Mdm2 with cyclin G2 was weak in vitro. Likewise, their ability to complex with PP2AC was limited but modestly stronger given the relatively low levels of GST-PP2AC in the reactions. Thus, cyclin G1 is able to associate independently with ARF, p53, and Mdm2 in vitro, and p53 can also interact with cyclin G2 and PP2AC.



View larger version (59K):
[in this window]
[in a new window]
  		FIGURE 4. Cyclin G1 interacts with Mdm2, p53, and ARF in vitro and in vivo. A. Equal amounts of Sf9 lysates containing Mdm2, p53, and ARF were mixed with the indicated GST fusion proteins, and Western blots performed to measure binding. GST fusion proteins were detected by Ponceau S staining (bottom panel). B. GFP or GFP-G1 was co-expressed with Hdm2 or p53 in U2OS cells, and cell lysates were subjected to IP-Western blot analyses. Antibodies to GFP (G), p53 (P), and Hdm2 (H) were used in the IPs. C. U2OS cells expressing GFP or GFP-G1 with ARF, or cells transfected with ARF and Hdm2, were analyzed by IP-Western blotting, as described above. Antibodies to ARF (A) were used. D. Endogenous cyclin G1 complexes were examined in ARF-arrested NIH 3T3 cells treated with MG132 for 3 h. Direct lysates (50 µg per lane) from vector (V)- and ARF (A)-infected cells, and immunoprecipitated complexes (from 700 µg per IP) were analyzed by Western blotting. IPs were performed with the indicated antibodies, including pAb421 conjugated to Sepharose ({alpha}p53) and two different antibodies to cyclin G1 ({alpha}G1-sc, Santa Cruz Biotechnology, Santa Cruz, CA; {alpha}G1*, polyclonal 1133).

 
To define the region(s) of cyclin G1 that interact with ARF, p53, and Mdm2, similar in vitro binding studies were performed with COOH-terminal deletion mutants of cyclin G1 fused with GST. GST-G1 proteins containing NH2-terminal residues 1–24 or 1–57 of cyclin G1 were incapable of binding to either ARF, p53, or Mdm2 (Table 2). Relatively weak binding was observed between those proteins and residues 1–187 of cyclin G1, a construct that contains the conserved cyclin box (24, 41). In contrast, GST-G11–217 exhibited strong association with ARF and optimal binding to p53 and Mdm2 when compared to the binding ability of wild-type cyclin G11–294. Thus, residues 187–217 are required for efficient binding of ARF, p53, and Mdm2. That region comprises the first three {alpha} helices of the so-called "box repeat," a COOH-terminal region of five {alpha} helices containing structural similarity to the cyclin box (41).


View this table:
[in this window]
[in a new window]
  		Table 2. The COOH-Terminal "Box Repeat" Region of Cyclin G1 Is Required for Efficient in Vitro Binding to ARF, Mdm2, and p53

 
To determine whether cyclin G1 formed complexes with Mdm2, ARF, and p53 in vivo, immunoprecipitation (IP)-Western blot analyses were performed from U2OS cells expressing GFP-G1 with each of the individual proteins in U2OS cells (Fig. 4B and C). GFP control did not associate with either Hdm2, p53, or ARF. In contrast, GFP-G1 efficiently associated with ectopically expressed Hdm2 and p53 (Fig. 4B). It also coprecipitated with endogenous Hdm2 in p53 immunoprecipitations from cells overexpressing p53, suggesting that a trimeric complex between G1-Hdm2-p53 can be formed. By comparison, cyclin G1 appeared to interact weakly with ARF relative to the association between ARF and Hdm2 (Fig. 4C).

NIH 3T3 cells infected with ARF retroviruses were then used to determine if endogenous cyclin G1, p53, and Mdm2 form complexes in ARF-arrested cells. As expected from earlier studies (39, 42), protein complexes between ARF, Mdm2, and p53 were observed in the arrested cells (Fig. 4D). Two different antibodies efficiently precipitated cyclin G1 and coprecipitated a small amount of p53, yet failed to coprecipitate ARF or Mdm2. Conversely, low levels of cyclin G1 were detectably precipitated by antisera to Mdm2 and p53 compared to IgG control, suggesting that a small percentage of cyclin G1 is associated with those regulators during ARF-mediated growth arrest.

Cyclin G1 Is Relocalized by ARF and Hdm2
Given that cyclin G1 forms complexes with Hdm2, ARF, and p53 in vivo, we tested whether its localization was altered by those proteins in U2OS cells (Fig. 5 ). As expected, GFP was expressed in both the cytoplasm and nucleus, and its localization was not altered by expression of ARF, p53, or Hdm2, or vice versa. GFP-G1 was distributed throughout the entire nucleus, including the nucleoli, when expressed alone in U2OS cells. In contrast, GFP-G1 localization was altered by exogenous ARF and Hdm2. GFP-G1 became exclusively nucleolar in nearly 60% of transfected cells co-expressing ARF, whereas Hdm2 retained GFP-G1 in the nucleoplasm in 65% of the population (Fig. 5). To assess the specificity of the ARF effect, GFP-G1 was co-expressed with an ARF mutant, D1-62, that localizes to the nucleoplasm and lacks growth inhibitory activity (40, 43). GFP-G1 localization was unaffected by D1-62. Importantly, expression levels were key determinants of cyclin G1 localization because lower levels of ARF or Hdm2 were less efficient at relocalizing GFP-G1 (data not shown). Overall, colocalization and/or relocalization of cyclin G1 with p53, Hdm2, and ARF supports the notion that in vivo complexes exist between those proteins.



View larger version (47K):
[in this window]
[in a new window]
  		FIGURE 5. Cyclin G1 subnuclear distribution is distinctly altered by Hdm2 and ARF. U2OS cells were transfected with GFP or GFP-G1 constructs, with or without ARF, ARF mutant D1-62, Hdm2, or p53. Immunofluorescence was used to determine the localization of GFP and GFP-G1 proteins (green) versus ARF, Hdm2, and p53 (Texas Red). Relocalization of GFP-G1 by Hdm2 and ARF was quantified from three independent experiments in which 100 cells or more were counted per condition.

 
Cyclin G1 Stabilizes and Activates p53
The mechanism by which cyclin G1 potentiated ARF-mediated growth inhibition appeared to be p53 independent. However, the G1-phase arrest induced by high-level expression of cyclin G1 correlated with an ability to bind Mdm2 and p53. We hypothesized that an association between cyclin G1 with Mdm2 or p53 might block the ability of Mdm2 to negatively regulate p53, thereby fostering p53 activation. Consequently, we measured the stability and transcriptional activity of p53 in U2OS and NIH 3T3 cells expressing GFP-G1 (Fig. 6 ). GFP-positive cells were collected by cell sorting from populations transfected with GFP or GFP-G1 and analyzed by Western blotting for expression of p53, Mdm2, and the respective GFP proteins (Fig. 6A). In U2OS cells, GFP-ARF was included as a positive control, and as anticipated, it stabilized p53 and led to up-regulation of Mdm2 compared to GFP controls. Notably, GFP-G1 also caused a 2- to 5-fold increase in p53 and Mdm2 levels, although the magnitude of up-regulation by cyclin G1 was consistently less than that achieved by ARF. A similar result was observed in NIH 3T3 cells in which p53 and Mdm2 expression was enhanced by GFP-G1 compared to GFP-positive and GFP-negative control cells (Fig. 6A, right panel). The increased expression of p53 indicated that it was stabilized by cyclin G1, whereas the up-regulation of Mdm2 suggested that p53 was activated.



View larger version (48K):
[in this window]
[in a new window]
  		FIGURE 6. Cyclin G1 stabilizes and activates p53. A. U2OS and NIH 3T3 cells were transfected with GFP, GFP-G1, or GFP-ARF plasmids, and GFP-positive cells (as well as GFP-negative cells from NIH 3T3 experiments) were collected by sorting. Cell lysates were electrophoresed on separate gels and immunoblots performed to determine expression of p53 and GFP (upper blots) or Mdm2 (lower blots), using Stat5 as the loading control for each membrane. B. U2OS and NIH 3T3 cells stably expressing a p53 luciferase reporter construct were similarly transfected, and the relative luciferase activity within GFP-G1- or GFP-ARF-expressing cells was calculated compared to GFP controls. Data are representative of three independent experiments.

 
To directly measure the effects of cyclin G1 on p53 activity, U2OS and NIH 3T3 stable cell lines expressing a p53 luciferase reporter construct were generated and transiently transfected with GFP, GFP-ARF, or GFP-G1. Luciferase assays revealed that GFP-ARF and GFP-G1 activated p53 transcription, with an average 5- to 8-fold increase above background levels observed in GFP-expressing cells (Fig. 6B). Thus, cyclin G1 is a positive regulator of p53. Because U2OS and NIH 3T3 cells lack the INK4a/ARF gene, these results also demonstrate that cyclin G1 is able to activate p53 in the absence of ARF.

Cyclin G1 Function Does Not Require p53 Yet Shows Partial Dependence on pRb
To test whether growth arrest mediated by cyclin G1 required p53 activity, GFP or GFP-G1 was expressed in mouse and human cells lacking p53 (Fig. 7 ). As shown previously, GFP expression alone had minimal effects on the cell cycle distributions of each cell line tested, whereas p53-positive NIH 3T3 and U2OS were efficiently arrested by cyclin G1 (Fig. 7A). Cyclin G1 also exhibited growth suppressive activity in murine 10-1 cells, an immortalized derivative of Balb 3T3 fibroblasts that lacks p53 (44). In contrast, cyclin G1 failed to block growth in p53-null Saos-2 osteosarcoma cells. Besides lacking p53, Saos-2 cells carry a homozygous deletion of RB genes (45), whereas 10-1 cells are pRb-positive.3



View larger version (26K):
[in this window]
[in a new window]
  		FIGURE 7. Cyclin G1-mediated growth arrest does not require p53 but exhibits partial pRb dependence. A. The relative percentage S phase for sorted cells expressing GFP (black bar) or GFP-G1 (gray bar) compared to GFP-negative cells within the same populations. Two independent experiments were performed for each cell type. B. U2OS cell lines expressing the human papilloma viral proteins, E6 or E7, or empty vector (Vec) were transfected with GFP (black bar) or GFP-G1 (gray bar). GFP-positive cells were collected by sorting and their DNA content measured by PI staining and flow cytometry. The mean percentage S phase for each sample is shown from two independent experiments.

 
The above data suggested that growth inhibition by cyclin G1 was p53 independent but might require pRb. To test that idea more directly, we established isogenic derivatives of U2OS cells stably expressing the human papilloma viral proteins, E6 or E7, or empty vector. The E6 protein targets p53 for degradation, whereas E7 binds to and inactivates pRb (46–48). Cyclin G1 effectively arrested U2OS-vector and U2OS-E6 cells, supporting the notion that p53 is not required for cyclin G1-mediated growth suppression (Fig. 7B). By comparison, cells expressing E7 were only partially arrested by cyclin G1, consistent with the data obtained in Saos-2 cells that suggested a role for pRb in cyclin G1-induced growth arrest. Identical results were obtained in BrdUrd incorporation assays using the same cell types transfected with either GFP or GFP-G1 (data not shown).

Mdm2 Targets Cyclin G1 for Degradation in the 26S Proteasome
During the course of our studies, we found it difficult to achieve and maintain high levels of ectopically expressed cyclin G1. This was evident in both transfection and infection experiments in which exogenous cyclin G1 levels dramatically decreased from 30 to 72 h post-expression (data not shown). As such, we tested whether cyclin G1 was degraded by the 26S proteasome. NIH 3T3 cells were infected with retroviruses encoding vector or cyclin G1, and 2 days later, the populations were treated with the proteasome inhibitor, MG132, for increasing amounts of time. As shown in Fig. 8A , treatment of vector control cells with MG132 resulted in the stabilization of endogenous cyclin G1, p53, and Mdm2.



View larger version (40K):
[in this window]
[in a new window]
  		FIGURE 8. Cyclin G1 degradation by the 26S proteasome is promoted by Hdm2. A. NIH 3T3 cells were infected with retroviruses encoding vector (V) or HA-cyclin G1 (marked with an asterisk), and treated with DMSO at time 0 or 50 µM MG132 for the indicated times. ARF-infected cells (A) were left untreated. Cyclin G1, p53, Mdm2, and B23 (loading control) expression was examined by immunoblotting. B. GFP-G1-transfected NIH 3T3 cells were treated for 6 h with (+) or without (-) 20 µM MG132. Matched images using the same confocal settings are represented. Although not shown, GFP expression was not altered by MG132 treatment. C. U2OS cells were transfected with the indicated plasmids, and the expression of p53, Hdm2, GFP, GFP-G1, and Stat5 (loading control) was determined by Western blotting.

 
As a target of p53, the up-regulation of endogenous cyclin G1 could be attributed to induction by stabilized p53 rather than inhibition of proteasome-mediated degradation. Therefore, we analyzed expression of HA- or GFP-tagged cyclin G1 constructs that lack p53 promoter sites. Both HA-cyclin G1 and GFP-G1 were markedly stabilized by MG132, indicating that cyclin G1 is a target of the 26S proteasome (Fig. 8A and B). Interestingly, the stabilization of exogenous HA-cyclin G1 was maximal after 1 h treatment with MG132, and this coincided with accelerated stabilization and up-regulation of endogenous p53 and Mdm2 compared to vector control cells (Fig. 8A). Such data are consistent with the ability of cyclin G1 to stabilize and activate p53. However, the high levels of HA-cyclin G1 obtained in MG132-treated cells, which are comparable to the levels of endogenous cyclin G1 induced by ARF, did not stabilize p53 as well as ARF. This indicates that additional events besides up-regulation of cyclin G1 contribute to ARF-mediated stabilization of p53.

Given that cyclin G1 interacts with Hdm2, an ubiquitin ligase that targets p53 for degradation via the proteasome, we assayed whether degradation of cyclin G1 was mediated by Hdm2. Equivalent amounts of either p53, GFP-G1, or GFP were co-expressed with increasing amounts of Hdm2 in U2OS cells, and the expression of each protein was assessed by Western blotting (Fig. 8C). As a control for Hdm2 activity and to establish the validity of our assay, we showed that p53 expression was progressively reduced as Hdm2 expression was increased. Moreover, p53 was not destabilized by an Hdm2 mutant (Hdm2.Ala466–473) which is disrupted in the RING domain required for ubiquitin ligase activity (49). Identical results were obtained with GFP-G1, whereas GFP stability was not affected by Hdm2 expression. These results strongly suggested that cyclin G1, like p53, is a target of Hdm2-mediated degradation.


   Discussion
Top
 Notes
 Abstract
 Introduction
 Results
 Discussion
Materials and Methods
 Acknowledgements
 References
 
A key finding of this work was that cyclin G1 overexpression caused a G1-phase growth arrest. Since its discovery in 1994 as a transcriptional target of p53 (19), cyclin G1 has been described as either a positive or negative regulator of cell growth. For instance, stable overexpression of cyclin G1 was found to accelerate clonal expansion in primary fibroblasts (29), whereas transient overexpression of cyclin G1 in cell lines and primary hepatocytes induced apoptosis (36). Our finding that low levels of cyclin G1 lack growth inhibitory activity (and may even promote growth) while high levels suppress it may help to reconcile data suggesting opposite roles for cyclin G1 in growth control. One likely reason the growth inhibitory effects of cyclin G1 have not been widely appreciated is that its expression is markedly down-regulated by the proteasome in a potentially Mdm2-dependent manner. As such, it is difficult to achieve high levels of cyclin G1 except in transient assays. Also, studies that relied on stable overexpression would naturally exclude cells arrested by cyclin G1 (23, 29, 35).

A particularly novel observation was that growth inhibition by cyclin G1 coincided with stabilization and activation of p53. As depicted in the model in Fig. 9 , this shows that cyclin G1 can act in a positive regulatory feedback loop to activate p53. The molecular mechanisms underlying p53 activation by cyclin G1 are presently undefined, but we speculate that Mdm2 function is somehow blocked. This could result from cyclin G1 stoichiometrically limiting the binding between p53 and Mdm2, in keeping with the finding that both proteins associate with the same COOH-terminal region of cyclin G1. However, cyclin G1 overexpression actually enhanced the detection of Mdm2-p53 complexes in vivo (Fig. 4 and Ref. 32), and binding studies revealed that p53 and cyclin G1 interact with distinct regions of Mdm2.4 Thus, Mdm2 might bridge the interaction between cyclin G1 and p53 in vivo. As such, introduction of cyclin G1 into Mdm2-p53 complexes may partially disrupt Mdm2 conformation and cause reduced p53 ubiquitination.



View larger version (15K):
[in this window]
[in a new window]
  		FIGURE 9. Model depicting the functional relationship between cyclin G1 with ARF, Mdm2, p53, and pRb. Cyclin G1 and Mdm2 are transcriptional targets of p53; ARF activates p53 and induces their expression by antagonizing Mdm2, an ubiquitin ligase and negative regulator of p53. Mdm2 may also down-regulate cyclin G1 activity because it enhances proteasomal degradation of cyclin G1. At high levels of expression, cyclin G1 inhibits cell growth and stimulates p53 activity. We hypothesize that cyclin G1 activates p53 by disrupting Mdm2 function (dashed line), possibly by altering its phosphorylation status via PP2A. Once activated by cyclin G1, p53 would be expected to induce expression of p21, an inhibitor of Cdks that blocks growth by suppressing pRb phosphorylation. Consistent with that idea is our finding that cyclin G1 exhibits partial dependence on pRb to suppress growth. However, our observation that growth inhibition by cyclin G1 does not require p53 is inconsistent with that model. Consequently, a more direct functional link between cyclin G1 and pRb is likely, perhaps mediated through PP2A. Arrows, activating events; perpendicular bars, inhibitory processes.

 
It is also conceivable that cyclin G1 promotes p53 activation via its association with PP2A. That interaction causes the dephosphorylation of Mdm2 at S166 (32), an event associated with the cytoplasmic relocalization of Mdm2 and consequent stabilization of p53 in the nucleus (50). A notable complication with that notion is that cyclin G1-PP2A complexes also promote dephosphorylation of Mdm2 at T216 (32). Dephosphorylation at that residue is thought to result in p53 degradation, not stabilization, because cells lacking cyclin G1 express elevated p53 and hyperphosphorylated Mdm2 at T216. Those results led Okamoto et al. (32) to suggest that cyclin G1 can negatively regulate p53, although given the opposing effects of cyclin G1-PP2A on Mdm2 phosphorylation, they also speculated that cyclin G1 could activate p53. Unfortunately, the biological effects of cyclin G1 on p53 activity or cell growth were not tested in that study.

Because of its ability to activate p53, we were somewhat surprised that growth inhibition by cyclin G1 did not require p53. Several cell types lacking functional p53 were efficiently arrested by cyclin G1. Although those results suggested that p53 is not important for cyclin G1-mediated growth arrest, they do not rule out the alternative possibility that cyclin G1 serves a role in p53 signaling. Indeed, a transient surge in cyclin G1 expression following p53 stimulation might mimic the high levels achieved in our transfection assays and amplify p53 activity. Subsequent down-regulation of cyclin G1 by the proteasome, presumably mediated by Mdm2, would then reduce cyclin G1 expression and remove its contribution to p53 signaling. Consistent with such a temporal response, others showed that cyclin G1 expression precedes Mdm2 up-regulation in response to DNA damage, and over time, cyclin G1 levels decrease while Mdm2 expression increases (51, 52). The circumstance or cellular context in which cyclin G1 expression is induced may also determine its impact on p53-dependent events. There is evidence that cyclin G1 contributes to DNA damage-induced checkpoints and G2-M-phase arrest in primary fibroblasts and hepatocytes (36–38).

On the other hand, several lines of evidence support our observation that cyclin G1 can function independent of p53. First, cyclin G1 is expressed in cells and tissues lacking p53 (21–25). Second, it can sensitize cells to undergo apoptosis irrespective of p53 status (23), and its ability to potentiate ARF-induced growth arrest did not correspond with greater stabilization or activation of p53. Third, if the sole function of cyclin G1 was to regulate p53, cyclin G1-null animals would be expected to either lack p53 function and be predisposed to cancer or possibly die during embryogenesis due to unchecked p53 activity. Neither outcome was observed in mice lacking cyclin G1 (38). Indeed, the lack of tumorigenesis or overt developmental defects in cyclin G1-null mice suggests that it does not function as a tumor suppressor or essential regulator of growth. Rather, it may contribute to growth control in response to genotoxic stresses or at particular times in development, and interactions with other regulators, such as pRb and PP2A, may dictate its role. An important point when considering cyclin G1 function, however, is the possibility that cyclin G2 may have redundant or compensatory functions. That idea is bolstered by findings that cyclin G2 also inhibits cell growth (31), and it associates with many of the same proteins, including p53, PP2A, Mdm2, and ARF (data herein and Refs. 31, 32, and 52).

Our data showed that cyclin G1-mediated growth suppression was partially dependent on pRb. Cyclin G1 had no growth inhibitory activity in RB-null Saos-2 cells and only partial activity in U2OS cells expressing E7, an oncoprotein known to cause the degradation and inactivation of pRb (46, 48, 53, 54). Earlier work showed that cyclin G1 can associate with pRb, and that the effects of cyclin G1 on growth in RKO colon cancer cells were lost on inactivation of pRb (29). While the authors of that study proposed that cyclin G1 promoted growth in a pRb-dependent manner, we suggest that pRb may be required for growth inhibition by cyclin G1. In fact, both ideas may be correct. As mentioned earlier, the level of cyclin G1 expression may determine its effects on growth. The noteworthy point of agreement is that pRb may be essential for cyclin G1 action.

The pRb tumor suppressor protein and its closely related family members, p107 and p130, are phosphoproteins that play an essential role in controlling the G1-to-S phase transition (55). When hypophosphorylated, the pRb proteins sequester E2F transcription factors and block S-phase entry (56). Protein phosphatase 1 (PP1) and PP2A represent two different classes of serine/threonine phosphatases that have been implicated in a number of biological processes, including phosphorylation of the pRb proteins (57). Considerable evidence suggests that PP1 dephosphorylates pRb and activates pRb-dependent growth arrest (58–60), whereas PP2A preferentially targets p107 (61, 62). However, there is a possibility that PP2A can also act on pRb (63). Given its ability to associate with both PP2A and pRb, it is conceivable that cyclin G1 induces G1-phase arrest by promoting PP2A-mediated dephosphorylation of either pRb or p107. A potential role for PP1 still cannot be excluded, nor can the possibility that cyclin G1 binding to pRb directly blocks its phosphorylation by Cdks. At this point, additional studies are warranted to define the molecular basis of the functional relationship between pRb and cyclin G1.

An important discovery during the course of these studies was that cyclin G1 expression is limited by proteasomal degradation. Moreover, Mdm2 overexpression accelerated that process, whereas a RING finger mutant of Mdm2 that lacks ubiquitin ligase activity failed to promote cyclin G1 degradation. Those findings are exciting because they represent the first demonstration of cyclin G1 regulation by the proteasome, and they suggest that Mdm2 normally controls cyclin G1 expression via ubiquitination. In fact, preliminary data from a variety of in vivo ubiquitination assays support that notion.5 The ability of Mdm2 to regulate both p53 and cyclin G1 is striking, particularly because cyclin G1 can activate p53. As such, this work implies that Mdm2 can negatively regulate p53 function by promoting cyclin G1 degradation, in addition to its more direct effects on p53.

The significance of cyclin G1 relocalization to nucleoli in cells overexpressing ARF is presently unclear. Binding studies revealed relatively weak association between cyclin G1 and ARF in vivo; therefore, the relocalization of cyclin G1 to nucleoli likely requires other factors. Although Hdm2 and p53 showed more efficient interaction with cyclin G1 in vivo, both remained largely or completely nucleoplasmic in cells expressing cyclin G1 and ARF (data not shown), suggesting that they were not responsible for directing cyclin G1 to nucleoli. It remains to be determined whether the localization of cyclin G1 correlates with its effects on growth and ability to activate p53. Various types of tumors, including osteosarcomas, and breast and prostrate cancers, express high levels of cyclin G1 (22). Because we found that high expression of cyclin G1 is growth inhibitory, it is presumed that those tumor cells lack essential regulators of G1 function, such as pRb. It is also possible that mislocalization of cyclin G1 in cancer cells cancels its growth inhibitory effects. Others showed that cyclin G1 failed to cluster in discrete nuclear foci in response to DNA damage in transformed cells, but did so in normal breast epithelial cells, prompting them to postulate that "clustering" enabled cyclin G1 to act as a p53 effector (22).

This study was initiated to determine if cyclin G1 is a regulator within the ARF signaling pathways. The data suggest that it is, because it can contribute to ARF-mediated cell cycle arrest and has intrinsic growth inhibitory activity. Notably, the ability of cyclin G1 to potentiate growth arrest by ARF did not involve activation of p53, suggesting that cyclin G1 functions as a downstream effector of p53 in the p21-independent pathway induced by ARF. We can conclude that cyclin G1 is not a participant in the p53/Mdm2-independent pathway, because it is not expressed in p53-null tko cells arrested by ARF. At present, we have not addressed whether cyclin G1 is required for ARF-induced arrest. That seems highly unlikely, however, because ARF can inhibit growth in the absence of p21 or p53, demonstrating that it has multiple mechanisms of action (16–18). Rather, cyclin G1 may affect the magnitude or kinetics of growth suppression by ARF and p53.

In conclusion, these studies provide some explanation for how cyclin G1 can exhibit differential effects on cell growth. We propose that the levels and timing at which it is expressed largely dictate its function, possibly by modulating the regulators with which it associates. In agreement with that idea, cyclin G1 affected both the accumulation and degradation of p53 depending on whether it was in complexes with Mdm2/ARF or Mdm2/PP2A, respectively (52). In that regard, interesting parallels can be drawn with another p53 target, p21, because low levels of p21 facilitate the assembly of growth-promoting cyclin D/Cdk4 complexes while high levels are redistributed among the Cdks and consequently inhibit growth (15). Our current understanding of cyclin G1 suggests that its different roles in growth control correlate, at least in part, with differential regulation of the ARF-Mdm2-p53 pathway.


   Materials and Methods
Top
 Notes
 Abstract
 Introduction
 Results
 Discussion
 Materials and Methods
Acknowledgements
 References
 
Cell Culture and Protein Expression
NIH 3T3 fibroblasts and U2OS osteosarcoma cells (both ARF-null, p53 and Mdm2 wild type) were grown in DMEM containing 10% fetal bovine serum, 2 mM glutamine, and 100 µg/ml of penicillin and streptomycin. Two p53-null cell lines, Saos-2 and 10-1 (44), and primary MEFs lacking p53, Mdm2, and ARF (kindly provided by Gerry Zambetti, St. Jude Children's Research Hospital) (18), were similarly maintained. Narf6 cells (kindly provided by Gordon Peters, ICRF) were treated with 1 µg/ml IPTG for 2 days to induce ARF expression. Cells were treated with 20–50 µM MG132 (Calbiochem, San Diego, CA) to inhibit the proteasome.

Retroviruses containing HA-tagged wild-type ARF or murine cyclin G1 were produced and infected into NIH 3T3 cells, as described (6). For sequential infections, cells were first infected with vector or cyclin G1 viruses for 6 h, followed by 4 ml of ARF retrovirus overnight. The next day, fresh medium was added and cells were collected 24 h later. Plasmid DNAs were transfected by a modified calcium phosphate precipitation method (64). To select stable lines expressing human papilloma virus E6 or E7 proteins, amphotropic viruses were first collected from PA317 cells expressing either the empty pLXSN retroviral vector, pLXSN-E6 or pLXSN-E7 (kindly provided by Denise Galloway, Fred Hutchison Cancer Center). U20S cells were then infected with 10 ml of each virus containing 8 µg/ml polybrene for 18 h, and selected in 0.6 mg/ml neomycin (G418) for 2–3 weeks. Expression of E7 was confirmed by Western blotting (Zymed, San Francisco, CA), whereas E6 expression was confirmed by reduced p53 expression (measured by Western blotting) and diminished ability of ARF to induce growth arrest (data not shown).

DNA Constructs
Expression constructs containing HA-tagged ARF and its mutant, D1-62, in pcDNA3.1, pSR{alpha}-MSV-tk-neo, or pEGFP, have been described elsewhere (6, 65). DNA constructs for GFP-tagged cyclin G1 were prepared by polymerase chain reaction amplification of murine cyclin G1 cDNA (forward primer: 5'-GCGAAGCTTGGATCCACCATGGTAGAAGTACTGACAACTGACTCTC-3' and reverse primer: 5'-GAGCCCGGGAATTCTTACAAATGGTCTCAGGAATCGTTGG-3'). A 950-bp BamHI-SmaI cyclin G1 cDNA was subcloned into pEGFP-N1 (Clontech Laboratories, Inc., Palo Alto, CA) at BglII-SmaI sites. Cyclin G1 cDNA was further amplified from pEGFP-cyclinG1 (forward primer: 5'-GCGAAGCTTGGATCCACCATGGTAGAAGTACTGACAACTGACTCTC-3' and reverse primer: 5'-GCGGAATTCTCAACTCGAGGTCGACTGACAAATGGTCTCAGGAATCGT-3'). Products were subcloned into pcDNA3.1, a pBlueScript vector containing an HA epitope, and HA-cyclin G1 was ligated into pSR{alpha}-MSV-tk-neo. GST expression constructs for cyclin G1, cyclin G2, and PP2A/C have been described (31). Mutants of cyclin G1 in the pGex vector were generated by deletion of internal fragments from full-length cyclin G1 and vector religation, including XbaI/XhoI (GST-G11–24), BglII/XhoI (GST-G11–57), StuI/XhoI (GST-G11–187), and SnaBI/XhoI (GST-G11–217).

Analyses for Growth Arrest
The DNA content of GFP-positive and GFP-negative cells was determined by staining live cells with Hoescht dye 33342 (Sigma Chemical Co., St. Louis, MO), exactly as described (31). Dead cells within the populations were identified by staining with 5 µg/ml PI for 5 min at room temperature. Cells were sorted by an Epics 753 dual laser cytometer (Beckman Coulter Corporation, Miami, FL). PI-positive cells and doublets were excluded to ensure that only single viable cells were used for analysis of DNA content and GFP expression. Alternatively, 1.5 x 105 GFP-positive and GFP-negative cells were sorted from unstained populations, stained with PI, and analyzed for DNA content using a FACScan (Becton Dickinson, San JOse, CA) (6). For infected cells, DNA content was determined 48 h post-infection by PI staining and FACScan analysis. Final cell cycle distributions were determined using ModFit (Verity Software House, Topsham, ME) or Watson Pragmatic (FlowJo, Tree Star Inc., San Carlos, CA) software. Cell cycle progression into S phase was also monitored by BrdUrd incorporation (14).

Protein Interaction Analyses
Cells were lysed (1 x 107 cells/ml) for 1 h on ice in NP40 buffer [50 mM Tris (pH 7.5), 120 mM NaCl, 1 mM EDTA, 0.5% NP40] supplemented with 0.1 mM sodium vanadate, 1 mM sodium fluoride, 5 µg/ml leupeptin, and 30 µM phenylmethylsulfonyl flouride. Lysates were sonicated (1 x 5 s pulse) and clarified by centrifugation at 12,000 rpm for 10 min at 4°C. Equivalent amounts of protein were immunoprecipitated with protein A- or G-Sepharose at 4°C using antibodies to mouse ARF (6), GFP (Abcam, Cambridge, UK), Mdm2 [2A10 or Ab-1 (Oncogene Research Products, Cambridge, MA)], p53 [DO-1 (Santa Cruz Biotechnology) or pAb421 conjugated to Sepharose (generously provided by Ettore Appella from NIH)], and cyclin G1 [SCH-46 (Santa Cruz Biotechnology) or polyclonal 1133 (generated against human residues 11–25)6]. For in vitro interactions, insect cell lysates containing recombinant ARF, p53, or Mdm2 were incubated with GST fusion proteins on glutathione S-Sepharose (Amersham Biosciences, Piscataway, NJ), as described (31). Protein complexes, as well as lysates (50–100 µg protein per lane), were separated on denaturing gels and transferred onto polyvinylidene difluoride membranes (Millipore, Bedford, MA). Proteins were detected by enhanced chemiluminescence (ECL, Amersham) according to the manufacturer's specifications using the antibodies listed above, as well as antisera to Stat5a (kindly provided by Fred Quelle, University of Iowa) and nucleophosmin/B23 (Zymed).

p53 Reporter Assays
G418-resistant clones of U2OS and NIH 3T3 cells were established that stably expressed the p53 luciferase reporter construct, p53-luc (Stratagene, La Jolla, CA). Resulting reporter lines were transfected with pEGFP, pEGFP-G1, or pEGFP-ARF. Cells were collected 36–48 h later, lysed, and samples measured in triplicate for luciferase activity (Promega Luciferase Assay System). The same populations were simultaneously examined for GFP (transfected cells) or ARF-BrdUrd (infected cells) immunofluorescence to determine the efficiency of expression and to assess the growth arrest by ARF, respectively. Relative luciferase activities were calculated by normalizing luciferase readings to the percentage of cells expressing GFP proteins or ARF.

Localization Assays
U20S cells (2.5 x 105) were seeded onto glass coverslips in six-well dishes, fixed 36–48 h after transfection with 4% paraformaldehyde, and permeabilized with 0.2% Triton X-100 for 15 min. Wild-type ARF and D1-62 were detected with ARF antisera (6). Hdm2 was detected with SMP-14 (Santa Cruz Biotechnology, 1:80 dilution), and p53 was detected with 1 µg/ml DO-1 antibody (Santa Cruz Biotechnology). Secondary antibodies were used as follows: biotinylated antirabbit at 1:500, biotinylated antimouse at 1:100, and Streptavidin Texas Red (Amersham) at 1:200. Protein localization was analyzed using Zeiss or Bio-Rad confocal microscopes. Z-sections were performed to verify protein co-localization.


   Acknowledgements
Top
 Notes
 Abstract
 Introduction
 Results
 Discussion
 Materials and Methods
 Acknowledgements
References
 
The authors thank Al Klingelhutz, Gerry Zambetti, Chuck Sherr, Fred Quelle, Karen Vousden, Ettore Appella, Denise Galloway, and Gordon Peters for reagents. We also thank Jussara Hagen, Aruni S. Arachichige Don, and Brian Haugen for technical assistance. These studies were performed with assistance from the University of Iowa Flow Cytometry Facility, the Holden Comprehensive Cancer Center, and core facilities of the Diabetes and Endocrinology Research Center at the University of Iowa.


   Notes
Top
 Notes
Abstract
 Introduction
 Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 
1 D.E.Q. from the American Cancer Society (RSG-98-254-04-MGO) and NIH (RO1 CA90367), and by a grant from the NIH to M.C.H. (RO1 GM56900). Back

2 S. Winckler and M.C. Horne, unpublished observations. Back

3 C. Korgaonkar and D.E. Quelle, unpublished observations. Back

4 L. Zhao and D.E. Quelle, unpublished observations. Back

5 T. Samuels, S. Winckler, and M.C. Horne, unpublished observations. Back

6 M.C. Horne, unpublished data. Back

Received July 5, 2002; revised November 1, 2002; accepted December 13, 2002.


   References
Top
 Notes
 Abstract
 Introduction
 Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 

  1. Hainaut, P., Soussi, T., Shomer, B., Hollstein, M., Greenblatt, M., Hovig, E., Harris, C. C., and Montesano, R. Database of p53 gene somatic mutations in human tumors and cell lines: updated compilation and future prospects. Nucleic Acids Res., 25: 151–157, 1997.[Abstract/Free Full Text]
  2. Vogelstein, B., Lane, D., and Levine, A. J. Surfing the p53 network. Nature, 408: 307–310, 2000.[Medline]
  3. Giaccia, A. J. and Kastan, M. B. The complexity of p53 modulation: emerging patterns from divergent signals. Genes Dev., 12: 2973–2983, 1998.[Free Full Text]
  4. Levine, A. J. p53, the cellular gatekeeper for growth and division. Cell, 88: 323–331, 1997.[Medline]
  5. Momand, J., Wu, H. H., and Dasgupta, G. MDM2—master regulator of the p53 tumor suppressor protein. Gene, 242: 15–29, 2000.[Medline]
  6. Quelle, D. E., Zindy, F., Ashmun, R. A., and Sherr, C. J. Alternative reading frames of the INK4a tumor suppressor gene encode two unrelated proteins capable of inducing cell cycle arrest. Cell, 83: 993–1000, 1995.[Medline]
  7. Sherr, C. J. Tumor surveillance via the ARF-p53 pathway. Genes Dev., 12: 2984–2991, 1998.[Free Full Text]
  8. Ruas, M. and Peters, G. The p16INK4a/CDKN2A tumor suppressor and its relatives. Biochim. Biophys. Acta, 1378: F115–F177, 1998.[Medline]
  9. Sherr, C. J. The INK4a/ARF network in tumour suppression. Nat. Rev. Mol. Cell Biol., 2: 731–737, 2001.[Medline]
  10. Kamijo, T., Zindy, F., Roussel, M. F., Quelle, D. E., Downing, J. R., Ashmun, R. A., Grosveld, G., and Sherr, C. J. Tumor suppression at the mouse INK4a locus mediated by the alternative reading frame product p19(ARF). Cell, 91: 649–659, 1997.[Medline]
  11. Donehower, L. A., Harvey, M., Slagle, B. L., McArthur, M. J., Montgomery, C. A., Jr., Butel, J. S., and Bradley, A. Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature, 356: 215–221, 1992.[Medline]
  12. Llanos, S., Clark, P. A., Rowe, J., and Peters, G. Stabilisation of p53 by p14ARF without relocalization of Mdm2 to the nucleolus. Nat. Cell Biol., 3: 445–452, 2001.[Medline]
  13. Lin, A. W. and Lowe, S. W. Oncogenic ras activates the ARF-p53 pathway to suppress epithelial cell transformation. Proc. Natl. Acad. Sci. USA, 98: 5025–5030, 2001.[Abstract/Free Full Text]
  14. Korgaonkar, C., Zhao, L., Modestou, M., and Quelle, D. E. ARF function does not require p53 stabilization or Mdm2 relocalization. Mol. Cell. Biol., 22: 196–206, 2002.[Abstract/Free Full Text]
  15. Sherr, C. J. and Roberts, J. M. CDK inhibitors: positive and negative regulators of G(1)-phase progression. Genes Dev., 13: 1501–1512, 1999.[Free Full Text]
  16. Groth, A., Weber, J. D., Willumsen, B. M., Sherr, C. J., and Roussel, M. F. Oncogenic ras induces p19ARF and growth arrest in mouse embryo fibroblasts lacking p21Cip1 and p27Kip1 without activating cyclin D-dependent kinases. J. Biol. Chem., 275: 27473–27480, 2000.[Abstract/Free Full Text]
  17. Modestou, M., Puig-Antich, V., Korgaonkar, C., Eapen, A., and Quelle, D. E. The ARF tumor suppressor inhibits growth through p21-dependent and p21-independent pathways. Cancer Res., 61: 3145–3150, 2001.[Abstract/Free Full Text]
  18. Weber, J. D., Jeffers, J. R., Rehg, J. E., Randle, D. H., Lozano, G., Roussel, M. F., Sherr, C. J., and Zambetti, G. P. p53-independent functions of the p19ARF tumor suppressor. Genes Dev., 14: 2358–2365, 2000.[Abstract/Free Full Text]
  19. Okamoto, K. and Beach, D. Cyclin G is a transcriptional target of the p53 tumor suppressor protein. EMBO J., 13: 4816–4822, 1994.[Medline]
  20. Zauberman, A., Lupo, A., and Oren, M. Identification of p53 target genes through immune selection of genomic DNA—The cyclin-G gene contains 2 distinct p53 binding-sites. Oncogene, 10: 2361–2366, 1995.[Medline]
  21. Bates, S., Rowan, S., and Vousden, K. H. Characterisation of human cyclin G(1) and G(2): DNA damage inducible genes. Oncogene, 13: 1103–1109, 1996.[Medline]
  22. Reimer, C. L., Borras, A. M., Kurdistani, S. K., Garreau, J. R., Chung, M., Aaronson, S. A., and Lee, S. W. Altered regulation of cyclin G in human breast cancer and its specific localization at replication foci in response to DNA damage in p53+/+ cells. J. Biol. Chem., 274: 11022–11029, 1999.[Abstract/Free Full Text]
  23. Okamoto, K. and Prives, C. A role of cyclin G in the process of apoptosis. Oncogene, 18: 4606–4615, 1999.[Medline]
  24. Horne, M. C., Goolsby, G. L., Donaldson, K. L., Tran, D., Neubauer, M., and Wahls, A. F. Cyclin G1 and cyclin G2 comprise a new family of cyclins with contrasting tissue-specific and cell cycle-regulated expression. J. Biol. Chem., 271: 6050–6061, 1996.[Abstract/Free Full Text]
  25. Morita, N., Kiryu, S., and Kiyama, H. p53-independent cyclin G expression in a group of mature neurons and its enhanced expression during nerve regeneration. J. Neurosci., 16: 5961–5966, 1996.[Abstract/Free Full Text]
  26. Irwin, M. S. and Kaelin, W. G. p53 family update: p73 and p63 develop their own identities. Cell Growth & Differ., 12: 337–349, 2001.[Free Full Text]
  27. Kanaoka, Y., Kimura, S. H., Okazaki, I., Ikeda, M., and Nojima, H. GAK: a cyclin G associated kinase contains a tensin/auxilin-like domain. FEBS Lett., 402: 73–80, 1997.[Medline]
  28. Kimura, S. H., Kataoka, T. R., Endo, Y., and Nojima, H. Genomic structure and chromosomal localization of mouse cyclin G1 gene. Genomics, 46: 483–486, 1997.[Medline]
  29. Smith, M. L., Kontny, H. U., Bortnick, R., and Fornace, A. J. The p53-regulated cyclin G gene promotes cell growth: p53 downstream effectors cyclin G and GADD45 exert different effects on cisplatin chemosensitivity. Exp. Cell Res., 230: 61–68, 1997.[Medline]
  30. Okamoto, K., Kamibayashi, C., Serrano, M., Prives, C., Mumby, M. C., and Beach, D. p53-dependent association between cyclin G and the B' subunit of protein phosphatase 2A. Mol. Cell. Biol., 16: 6593–6602, 1996.[Abstract]
  31. Bennin, D. A., Arachchige Don, A. S., Brake, T., McKenzie, J. L., Rosenbaum, H., Ortiz, L., DePaoli-Roach, A. A., and Horne, M. C. Cyclin G2 associates with protein phosphatase 2A catalytic and regulatory B' subunits in active complexes and induces nuclear aberrations and a G1/S phase cell cycle arrest. J. Biol. Chem., 277: 27449–27467, 2002.[Abstract/Free Full Text]
  32. Okamoto, K., Li, H. Y., Jensen, M. R., Zhang, T. T., Taya, Y., Thorgeirsson, S. S., and Prives, C. Cyclin G recruits PP2A to dephosphorylate Mdm2. Mol. Cell, 9: 761–771, 2002.[Medline]
  33. Kampmeier, J., Behrens, A., Wang, Y., Yee, A., Anderson, W. F., Hall, F. L., Gordon, E. M., and McDonnell, P. J. Inhibition of rabbit keratocyte and human fetal lens epithelial cell proliferation by retrovirus-mediated transfer of antisense cyclin G1 and antisense MAT1 constructs. Hum. Gene Ther., 11: 1–8, 2000.[Medline]
  34. Gordon, E. M., Liu, P. X., Chen, Z. H., Liu, L., Whitley, M. D., Gee, C., Groshen, S., Hinton, D. R., Beart, R. W., and Hall, F. L. Inhibition of metastatic tumor growth in nude mice by portal vein infusions of matrix-targeted retroviral vectors bearing a cytocidal cyclin G1 construct. Cancer Res., 60: 3343–3347, 2000.[Abstract/Free Full Text]
  35. Skotzko, M., Wu, L. T., Anderson, W. F., Gordon, E. M., and Hall, F. L. Retroviral vector-mediated gene-transfer of antisense cyclin G1 (CYCG1) inhibits proliferation of human osteogenic-sarcoma cells. Cancer Res., 55: 5493–5498, 1995.[Abstract/Free Full Text]
  36. Jensen, M. R., Factor, V. M., and Thorgeirsson, S. S. Regulation of cyclin G1 during murine hepatic regeneration following dipin-induced DNA damage. Hepatology, 28: 537–546, 1998.[Medline]
  37. Shimizu, A., Nishida, J., Ueoka, Y., Kato, K., Hachiya, T., Kuriaki, Y., and Wake, N. CyclinG contributes to G2/M arrest of cells in response to DNA damage. Biochem. Biophys. Res. Commun., 242: 529–533, 1998.[Medline]
  38. Kimura, S. H., Ikawa, M., Ito, A., Okabe, M., and Nojima, H. Cyclin G1 is involved in G2/M arrest in response to DNA damage and in growth control after damage recovery. Oncogene, 20: 3290–3300, 2001.[Medline]
  39. Stott, F. J., Bates, S., James, M. C., McConnell, B. B., Starborg, M., Brookes, S., Palmero, I., Ryan, K., Hara, E., Vousden, K. H., and Peters, G. The alternative product from the human CDKN2A locus, p14(ARF), participates in a regulatory feedback loop with p53 and MDM2. EMBO J., 17: 5001–5014, 1998.[Medline]
  40. Weber, J. D., Taylor, L. J., Roussel, M. F., Sherr, C. J., and Bar-Sagi, D. Nucleolar Arf sequesters Mdm2 and activates p53. Nat. Cell Biol., 1: 20–26, 1999.[Medline]
  41. Noble, M. E. M., Endicott, J. A., Brown, N. R., and Johnson, L. N. The cyclin box fold: protein recognition in cell-cycle and transcription control. Trends Biochem. Sci., 22: 482–487, 1997.[Medline]
  42. Kamijo, T., Weber, J. D., Zambetti, G., Zindy, F., Roussel, M. F., and Sherr, C. J. Functional and physical interactions of the ARF tumor suppressor with p53 and mdm2. Proc. Natl. Acad. Sci. USA, 95: 8292–8297, 1998.[Abstract/Free Full Text]
  43. Quelle, D. E., Cheng, M. G., Ashmun, R. A., and Sherr, C. J. Cancer-associated mutations at the INK4a locus cancel cell cycle arrest by p16(INK4A) but not by the alternative reading frame protein p19(ARF). Proc. Natl. Acad. Sci. USA, 94: 669–673, 1997.[Abstract/Free Full Text]
  44. Harvey, D. M. and Levine, A. J. p53 alteration is a common event in the spontaneous immortalization of primary BALB/c murine embryo fibroblasts. Genes Dev., 5: 2375–2385, 1991.[Abstract/Free Full Text]
  45. Isfort, R. J., Cody, D. B., Lovell, G., and Doersen, C. J. Analysis of oncogenes, tumor-suppressor genes, autocrine growth-factor production, and differentiation state of human osteosarcoma cell-lines. Mol. Carcinog., 14: 170–178, 1995.[Medline]
  46. Chellappan, S., Kraus, V. B., Kroger, B., Munger, K., Howley, P. M., Phelps, W. C., and Nevins, J. R. Adenovirus-E1A, simian virus-40 tumor-antigen, and human papillomavirus-E7 protein share the capacity to disrupt the interaction between transcription factor-E2F and the retinoblastoma gene-product. Proc. Natl. Acad. Sci. USA, 89: 4549–4553, 1992.[Abstract/Free Full Text]
  47. Scheffner, M., Werness, B. A., Huibregtse, J. M., Levine, A. J., and Howley, P. M. The E6 oncoprotein encoded by human papillomavirus type-16 and type-18 promotes the degradation of p53. Cell, 63: 1129–1136, 1990.[Medline]
  48. Dyson, N., Howley, P. M., Munger, K., and Harlow, E. The human papilloma virus-16 E7-oncoprotein is able to bind to the retinoblastoma gene-product. Science, 243: 934–937, 1989.[Abstract/Free Full Text]
  49. Lohrum, M. A. E., Ashcroft, M., Kubbutat, M. H. G., and Vousden, K. H. Identification of a cryptic nucleolar-localization signal in MDM2. Nat. Cell Biol., 2: 179–181, 2000.[Medline]
  50. Zhou, B. H. P., Liao, Y., Xia, W. Y., Zou, Y. Y., Spohn, B., and Hung, M. C. HER-2/neu induces p53 ubiquitination via Akt-mediated MDM2 phosphorylation. Nat. Cell Biol., 3: 973–982, 2001.[Medline]