Rbf1 degron dysfunction enhances cellular DNA replication

Abstract

The E2F family of transcription factors contributes to oncogenesis through activation of multiple genes involved in cellular proliferation, a process that is opposed by the Retinoblastoma tumor suppressor protein (RB). RB also increases E2F1 stability by inhibiting its proteasome-mediated degradation, but the consequences of this post-translational regulation of E2F1 remain unknown. To better understand the mechanism of E2F stabilization and its physiological relevance, we examined the streamlined Rbf1-dE2F1 network in Drosophila. During embryonic development, Rbf1 is insulated from ubiquitin-mediated turnover by the COP9 signalosome, a multi-protein complex that modulates E3 ubiquitin ligase activity. Here, we report that the COP9 signalosome also protects the Cullin4-E3 ligase that is responsible for dE2F1 proteasome-mediated destruction. This dual role of the COP9 signalosome may serve to buffer E2F levels, enhancing its turnover via Cul4 protection and its stabilization through protection of Rbf1. We further show that Rbf1-mediated stabilization of dE2F1 and repression of dE2F1 cell cycle-target genes are distinct properties. Removal of an evolutionarily conserved Rbf1 C terminal degron disabled Rbf1 repression without affecting dE2F1 stabilization. This mutant form of Rbf1 also enhanced G₁-to-S phase progression when expressed in Rbf1-containing S2 embryonic cells, suggesting that such mutations may generate gain-of-function properties relevant to cellular transformation. Consistent with this idea, several studies have identified mutations in the homologous C terminal domains of RB and p130 in human cancer.

Keywords: :

• retinoblastoma
• tumor suppressor
• E2F1
• repression
• proteasome
• ubiquitin
• degron

Introduction

The RB/E2F regulatory nexus

The Retinoblastoma family of proteins consists of the RB, p107 and p130 members that control multiple processes associated with cellular proliferation, including cell cycle, differentiation, apoptosis and cellular biosynthetic potential (reviewed in ref. 1). Consistent with their regulatory governance of these processes, RB family members are frequently inactivated in human cancers.^2,3 In some diseases, such as retinoblastoma and small cell lung carcinoma, mutations in the RB1 gene itself are potentially causative for disease. In other cancer types, deregulation is accomplished through altered function of upstream regulatory factors, including the cyclin-dependent kinases (cdk) and cyclin/cdk inhibitors, with effects encompassing all RB family members.⁴ Together, these genetic changes are so pervasive as to be recognized as a hallmark of cancer.^5,6

One important target for RB family members in gene regulation is the E2F family of transcription factors that like RB are tightly linked to growth control. In humans, at least eight different E2F species (E2F1–E2F8) have been identified, and are classified as either transcriptional activators (E2F1–3) or repressors (E2F4–8), based on their sequence homology and functional properties.^7,8 In Drosophila, these pathways are streamlined with two RB family proteins, Rbf1 and Rbf2, contributing to regulation of two E2F proteins, dE2F1 and dE2F2.⁹ During G₀ and early G₁ of cell cycle progression, RB family members directly bind to different sets of E2F factors,^10,11 and at least for E2F1, RB association reverses regulatory polarity from activation to repression.^11-16 Cyclin-cdk kinase mediated phosphorylation of RB in late G₁ causes RB/E2F1 dissociation, allowing E2F to activate numerous proliferation genes that drive entry into S phase.^17-21 In human cancer, increased E2F activity is frequently observed,^22-25 and is associated with poor prognosis, particularly in melanoma and breast cancer,^26-29 highlighting the importance of imposing regulatory curbs on E2F1 expression and activity.

The ubiquitin-proteasome system and RB/E2F regulation

In addition to limitation through cyclin/cdk-mediated phosphorylation, the RB/E2F axis is governed by the ubiquitin-proteasome system. Indeed, inappropriate RB turnover contributes to disease as demonstrated during cellular immortalization by viral proteins leading to enhanced RB ubiquitination.³⁰ Although RB levels often appear stable in actively proliferating cells,^31,32 steady-state fluctuations have been correlated with phosphorylation changes during cellular stress,³³ suggesting that a negative correlation exists between RB levels and its activity in certain contexts. For example, in response to nocodazole blockade, U2OS osteosarcoma cells exhibit marked elevation of RB levels in the G₂/M phase of the cell cycle, and upon release into early G₁, RB destabilization reestablishes lower baseline steady-state levels (not shown). In seminal experiments linking RB to cell cycle control, microinjected RB induced cellular G₁ arrest only when introduced during the window of time immediately after nocodazole release and not when injected in asynchronously proliferating cells,¹⁰ suggesting that RB function is correlated with conditions in early G₁ amenable to its diminishing steady-state levels. An inverse relationship between steady-state levels and repressor potency was also observed for the Drosophila Retinoblastoma family member Rbf1 wherein unstable Rbf1 proteins were potent for target gene repression while stable mutant proteins were impotent.³⁴ A tight, activity-instability linkage may ensure that RB repression programs remain dynamic and sensitive to growth conditions, such as previously suggested for dynamic p53 fluctuation in response to DNA damage.^35,36 Similar to RB, both p107 and p130 exhibit differential expression during the cell cycle with p107 levels peaking in S phase and p130 levels highest in G₀,^31,33,37 and thus, multiple mechanisms likely influence turnover of these different RB family members. Interestingly, cyclin/cdk kinase activity is correlated with changes in RB family member levels,^33,37 suggesting that the cyclin/cdk and ubiquitin/proteasome regulatory arms crosstalk to govern both RB family activity and stability.

In previous studies of the Drosophila embryo, we observed that the Rbf1 and Rbf2 proteins associate with the COP9 signalosome,³⁸ a developmentally regulated complex that controls proteasome-mediated degradation of many proteins through interactions with SCF (SKP1/cullin/F-box) E3 ubiquitin ligase complexes.³⁹ CSN5, the catalytic core of COP9, contains a metalloprotease motif termed JAMM (Jab1/MPN domain-associated metalloisopeptidase) that removes Nedd8 from the cullin subunits.⁴⁰ As cullin neddylation activates E3 ligase activity, the COP9 signalosome thus serves to protect substrates from turnover. Indeed, both Rbf1 and Rbf2 are destabilized in the absence of COP9 function,³⁸ connecting the regulation of Rbf protein turnover to the ubiquitin-proteasome system. The involvement of a specific ubiquitin ligase remains unknown, although in mammals, RB and p130 turnover has been linked to MDM2^41,42 and SCF^{Skp2 33}, respectively. As with the RB family, E2F family members are degraded through ubiquitin-mediated turnover, both at defined points during the cell cycle⁴³ and in response to DNA damage^44,45 with E2F1 subjected to ubiquitination via the S-phase-specific F-box protein SCF^Skp2, and degradation in the S/G₂ phases of the cell cycle.^46,47 Other ubiquitin ligases including APC/C⁴³ and ROC-Cullin ligases⁴⁸ likely contribute to E2F1 degradation in these contexts. In contrast, a protective role is suggested for the MDM2 ubiquitin ligase,⁴⁷ and consistently, p19^ARF mediated inhibition of MDM2 encourages E2F1 turnover.^49,50 Interestingly, a key determinant of E2F1 degradation turns out to be RB itself.^51-54 RB can bind to a C terminal instability element in E2F1, and may stabilize E2F1 by occluding the cellular ubiquitination machinery.⁵⁵ In Drosophila, E2F1 destruction is mediated by the Cul4^Cdt2 E3-Ubiquitin ligase,^56,57 suggesting that both Rbf1 and COP9 may coordinately influence E2F1 stability. Herein, we show that dE2F1 levels are indeed influenced by Rbf1 and COP9, but through distinct mechanisms. While Rbf1 can stabilize dE2F1 through pocket-domain-dependent protein-protein interactions, the COP9 signalosome complex downregulates dE2F1 levels through modulation of the Cul4 E3 ligase. Rbf1 mediated dE2F1 stabilization and repression activity are separate properties as select mutant Rbf1 forms lacking dE2F1 repression capability retained their capacity to stabilize dE2F1. We further show that this class of repression-inactive Rbf1 mutants enhanced the rate of S-phase entry perhaps through their inappropriate stimulation of dE2F1 levels.

Results and Discussion

The COP9 signalosome regulates the Rbf1/E2F1 pathway

To test whether loss of COP9 function is associated with destabilized E2F1, endogenous dE2F1 steady-state levels were examined in S2 embryonic cells that were depleted of the largest subunit of the COP9 complex, CSN1, using dsRNA. This treatment strongly reduced levels of transfected flag-tagged RBF1 and endogenous RBF2, as expected, while E2F1 levels were increased significantly (Fig. 1A). This result shows that the COP9 signalosome stabilizes Rbf proteins, as previously noted,³⁸ but instead of protecting E2F1, the COP9 complex contributes to its turnover. Similar results were obtained during CSN4 and CSN5 knockdowns (not shown), suggesting that the COP9 complex, rather than individual COP9 subunits, contribute to E2F modulation. Previous studies showed that the COP9 signalosome can stabilize cullin E3 ligase-containing SCF complexes,⁵⁸ and a Cul4-containing SCF complex contributes to dE2F1 ubiquitination in S2 cells.⁵⁶ Therefore, Cul4 levels were examined in Csn1 knockdown cells, ascertaining that Cul4 was indeed diminished during COP9 knockdown. Direct knockdown of Cul4 but not Cul5 also led to increased levels of E2F1, consistent with previous reports.⁵⁶ Thus, a pathway emerges wherein the Cul4 E3 ligase responsible for E2F1 turnover is stabilized by the COP9 signalosome (Fig. 1B). Direct modulation of Cul4 and Cul5 levels had no discernable effect on Rbf levels, indicating that Rbf1 and E2F1 are ubiquitinated through distinct pathways. We conclude that COP9 signalosome is associated with opposing roles for Rbf1 and E2F1, contributing to Rbf1 stabilization but E2F1 destabilization.

Figure 1. Dual roles of the COP9 signalosome in regulation of the Rbf1-dE2F1 network. (A) The COP9 signalosome complex governs both Rbf and E2F1 stability. S2 cells were treated with indicated dsRNA and proteins were measured by western blot analyses. Endogenous dE2F1 levels were dramatically increased due to reduced Cul4 levels when Cul4 (lane 4 and 6) or its upstream regulator COP9 (CSN1 subunit, lane 3) were depleted. Flag-Rbf1 and endogenous Rbf2 levels were substantially decreased by the CSN1 knockdown, but were not affected by Cullin knockdowns. (B) COP9 is a dual-functional regulator of dE2F1 stability. First, COP9 plays a protective role on Rbf1, which in turn stabilizes dE2F1. Second, COP9 restrains dE2F1 level by stabilizing a Cul4-based E3 ligase, which targets dE2F1 for degradation.

Drosophila RBF1 enhances dE2F1 levels

The data presented above highlights that a complicated network governs E2F stability with the COP9 signalosome contributing to low E2F1 levels during normal function. Previous studies have shown that in humans, RB can stabilize E2F1.^51,52,55 We therefore examined dE2F1 levels using Drosophila S2 cells that harbor wild-type COP9 function in the absence or presence of increased Rbf1 expression (Fig. 2A). Three days post-transfection, steady-state protein levels of Myc-tagged dE2F1 were measured in the presence or absence of the MG132 proteasome inhibitor. Consistent with previous studies on mammalian E2F1, increased levels of Drosophila dE2F1 were observed during Rbf1 expression and at levels comparable to those observed with MG132 proteasome inhibition (Fig. 2A). Under the conditions selected for this experiment, Rbf1 was relatively stable and its levels largely unaffected by MG132 treatment, although under different growth conditions, Rbf1 is proteasome sensitive.³⁴ These data indicate that dE2F1 is targeted by the ubiquitin-proteasome pathway and is responsive to the steady-state levels of Rbf1.

Figure 2. The Rbf1 pocket domain contributes to dE2F1 protection from proteasome-mediated degradation. (A) dE2F1 is sensitive to proteasome inhibition and is robustly stabilized by Rbf1 WT protein but not by forms of Rbf1 lacking the central pocket domain. Under these experimental conditions, Rbf1 WT and ΔPocket forms were expressed at equivalent levels and both are insensitive to proteasome inhibition. Endogenous tubulin levels are shown as loading controls. The experiment shown is representative of three biological replicates. (B) The Rbf1 pocket domain is insufficient for robust dE2F1 stabilization. The Rbf1 WT protein stabilizes dE2F1 protein, whereas at equivalent levels of expression, the Rbf1 pocket deletion mutant is incapable of stabilizing dE2F1 protein, while the Rbf1 pocket-only mutant provides only partial stabilization. Endogenous tubulin levels are shown as loading controls.

We next tested whether the direct binding of Rbf1 is required for dE2F1 stabilization. The conserved RB-family pocket domain is the primary site for E2F1 interaction,^59,60 therefore, a deletion mutant of Rbf1 lacking this domain was tested. Unlike with wild-type Rbf1, dE2F1 levels were unaffected by this mutant form of RBF1, while retaining responsiveness to proteasome inhibition. The pocket domain alone was sufficient to confer at least partial stabilization on dE2F1, suggesting that this domain is necessary but not sufficient for complete stabilization (Fig. 2B). These data are consistent with a model wherein dE2F1 is stabilized by direct contacts with Rbf1. Combined with our previous analysis of COP9 function, we conclude that the COP9 signalosome complex influences dE2F1 levels through two separate pathways, positively by stabilization of Rbf1, which binds dE2F1 to enhance cellular levels, and negatively by stabilizing the E3 ligase Cul4.

Our previous studies of Rbf1 indicated that an evolutionarily conserved C terminal instability element (IE) functions as an autonomous degron that stimulates both Rbf1 ubiquitination and repression potency^34,61. RB family proteins require multiple domains to mediate gene repression, including the pocket domain that facilitates E2F interaction and co-factor recruitment,^62-65 and the C terminal region that harbors the IE,³⁴ and may provide additional dE2F1 contacts, as was shown for human RB and p107.⁶⁰ Interestingly, Rbf1 degron deletion mutants retained their capacity to physically associate with dE2F1 at target gene promoters.³⁴ To determine whether Rbf1 repression activity and dE2F1 stabilization are biochemically separable, we generated a series of Rbf1 deletion constructs that were tested for both properties (Fig. 3A). Consistent with previous studies, transcription from the PCNA promoter was activated by dE2F1, but was repressed upon co-expression with wild-type Rbf1 (Fig. 3B). Also consistent, wild-type Rbf1 (1–845) robustly stabilized dE2F1. Mutant forms of Rbf1 lacking the pocket domain (1–375 and Δ376–727) were inactive for both repression and dE2F1 stabilization, attesting to the importance of this domain for both these properties. Significantly, three different mutant forms of Rbf1 that lacked the IE entirely or had mutations in four key lysines were defective for repression, but continued to stabilize E2F1. Together, these data identify one class of Rbf1 mutations that disable repression without affecting dE2F1 stabilization, and another class that disables both repression and stabilization.

Figure 3. Mutant Rbf1 lacking IE function stabilizes dE2F1 but cannot fully repress its transcriptional activity. (A) Schematic representation of Rbf1 proteins used for functional testing show the relative positions of the pocket domain and the IE region. (B) Functional characterization of Rbf1. Mutations in the IE (Δ728–786 and 4K–A) compromise transcriptional repression activities of Rbf1 proteins measured on the PCNA-luciferase reporter gene (bar graph), but do not affect dE2F1 stabilization property (anti-Myc western blot). Under these transfection conditions, Rbf proteins were expressed at similar levels (anti-Flag western blot). Data represents at least three biological replicates.

Rbf degron mutations enhance cellular S-phase entry

Previous studies have shown that elevated E2F levels are associated with increased cellular proliferation, and therefore, we hypothesized that enhanced dE2F1 stabilization enabled by repression-incompetent Rbf1 would facilitate ectopic S phase and contribute to deregulated cell growth. To test this possibility, GFP-Rbf1 WT and the Rbf1 ΔIE mutant were expressed in S2 cells and the effect on cell cycle progression was examined by FACS analysis (Fig. 4). Consistent with the established role of Rbf1 in G₁-to-S phase transition, GFP-Rbf1 WT induced a strong G₁ arrest in the transfected cells (GFP positive) that was not observed in untransfected cells (GFP negative) from the same culture (not shown), or in cells expressing the GFP-Rbf1 ΔIE mutant. The lack of G₁ arrest by GFP-Rbf1 ΔIE is consistent with a parallel lack of dE2F1 repression potency associated with IE loss. Interestingly, GFP-Rbf1ΔIE-expressing cells also displayed a modest increase in their S-phase percentage, as estimated by Modfit analysis (Fig. 4, inset). Therefore, as a direct measure of the ability of this mutant form of Rbf1 to stimulate S-phase entry, we performed BrdU incorporation assays. In this assay, cells expressing wild-type or mutant Rbf1 were visualized by anti-Flag epitope immunofluorescence, and cells undergoing de novo DNA synthesis were identified by BrdU staining (Fig. 5A). To assess the effect of transfected proteins on cell cycle, we calculated a proliferation index comparing the percentage of BrdU-positive cells in the transfected Rbf1-expressing population to the total population of cells. If a transfected protein exhibits no effect on cell cycle entry, the index should be equal to one, whereas the index should be less than one should the transfected protein cause cell cycle arrest. A protein that induces ectopic S phase entry should result in an index greater than one. As shown in Figure 5B, expression of Rbf1 lacking the pocket domain had no effect on DNA synthesis (P.I. = 1.03), consistent with observations that this protein is defective for repression and is unable to interact with and stabilize dE2F1. Therefore, in the context of S2 cells that express endogenous Rbf1, this defective Rbf1 protein had no effect. Cells expressing wild-type Rbf1 experienced substantial cell cycle arrest (P.I. = 0.63), suggesting that repression-competent Rbf1 downregulated key genes required for S phase entry. Importantly, expression of the repression-defective Rbf1ΔIE protein resulted in a substantial increase in the percentage of cells undergoing DNA replication (P.I. = 1.24), presumably through effects on dE2F1 stabilization. We conclude that expression of mutant Rbf1 harboring alterations to IE function confers a distinct growth advantage due to increased rate of S phase entry. In the model proposed in Figure 5C, dE2F1 activity increases during early G₁ concomitant with increases in its steady-state levels mediated, in part, by either wild-type or mutant Rbf1. However, loss of Rbf1-mediated repression associated with IE mutation permits premature S phase entry, a process that would normally be delayed in the presence of wild-type Rbf1 until licensing by cyclin/cdk phosphorylation. Based on these observations, we hypothesized that mutations in the IE domain of human RB family members may be selected for in human cancers. Indeed, in one study of non-small lung carcinoma, a substantial percentage of patients (84%) were found to harbor mutations in the p130 IE region.⁶⁶ We previously showed that mutant Rbf1 harboring alanine substitutions of lysine residues within the IE diminish repressor potency but do not eliminate Rbf1-dE2F1 protein-protein interactions.³⁴ It is interesting to note that many of the p130 IE cancer-associated mutations were observed in these conserved basic residues. Together, these studies suggest an unexpected role for RB family turnover in cellular proliferation and cancer.

Figure 4. The IE region contributes to Rbf1-mediated G₁ arrest. Wild-type or mutant GFP-Rbf1 ΔIE proteins were expressed in Drosophila S2 cells and the effect on cell cycle was determined by propidium iodide staining and FACS analyses. An overlay of the DNA content histograms for wild-type GFP-Rbf1 (gray) and mutant GFP-Rbf1 ΔIE-expressing cells (solid unfilled) shows that the loss of IE function is correlated with a diminished proportion of cells in the G₁ phase. (Inset) Bar graph shows the ratios of total S-phase percentages for GFP positive vs. GFP negative populations for GFP-Rbf1 WT and GFP-Rbf1 ΔIE transfected samples. In two separate experiments, GFP-Rbf1 ΔIE expressing cells exhibited an increased proportion of cells in S phase, as estimated using Modfit analysis.

Figure 5. Mutation of the RB-family degron positively influences DNA replication frequency. (A) Rbf1 ∆IE enhances S-phase entry. S2 cells were transfected with the indicated Rbf1 proteins and the effect on DNA synthesis was monitored by BrdU incorporation. Transfected cells and BrdU-positive cells were visualized by immunofluorescent staining with anti-Flag and anti-BrdU antibodies. Arrows indicate representative cells that are transfected/BrdU negative (green) or transfected/BrdU positive (yellow). Rbf1∆IE expression was associated with increased BrdU positive staining compared with Rbf1 WT expressing cells. (B) The effect of Rbf1 proteins on DNA replication was indicated as a proliferation index calculated as a ratio of the percentage of BrdU-positive cells in transfected cells to that in the total population. The Rbf1∆Pocket mutant that is unable to both stabilize and repress dE2F1 also showed no effect on BrdU incorporation. Rbf1 WT expressing cells exhibited diminished BrdU incorporation consistent with increased G₁ arrest, whereas Rbf1∆IE expressing cells exhibited enhanced S phase. Data from three biological replicates were analyzed. Error bars indicate standard deviation, and asterisks indicate p < 0.05. (C) A model for the regulation of E2F function and stability during cell cycle progression. In this model, E2F levels are stabilized during early G₁ by both wild-type and mutant Rbf1 proteins. The COP9 signalosome contributes to enhanced dE2F1 steady-state levels during this stage by stabilization of Rbf1. However, dE2F1 activity is not restrained by mutant Rbf1 leading to premature S phase entry. After Rbf1-dE2F1 estrangement mediated by cyclin cdk phosphorylation, the COP9 signalosome contributes to dE2F1 destruction via protection of Cul4.

Materials and Methods

Expression constructs

Generation of Rbf1 WT and mutant expression constructs was described previously.³⁴ To generate GFP-Rbf1 fusion proteins, eGFP cDNA was amplified from phs-eGFP and inserted into KpnI site of pAX-Rbf1 WT and pAX-Rbf1 Δ728–786 vectors. The Myc-E2F1 expression construct was a gift from Dr. Maxim Frolov).

E2F1 stabilization assay

1.5 million Drosophila S2 cells were transfected with 0.2 µg of pAXRbf1 WT or mutants and 0.2 µg of pIE-E2F1 constructs using the Effectene transfection reagent (Qiagen) according to the manufacturer’s protocol. The cells were grown for 3 d after which protein levels were analyzed through western blotting. For proteasome inhibition in Figure 2, the cells were treated with 50 µg/ml MG132 (Sigma-Aldrich) or the vehicle DMSO for 2 h.

Western blot analysis

To measure protein levels in Drosophila S2 cell culture, cells were harvested 3 d post-transfection and lysed by three freeze-thaw cycles in lysis buffer (50 mM TRIS-HCl pH8.0, 150 mM NaCl, 1% Triton X-100, protease inhibitors). Fifty µg of S2 cell lysates were run on 12.5% SDS-PAGE gels, transferred to a PVDF membrane, and probed with M2 anti-Flag (mouse monoclonal, 1:10,000, Sigma; F3165), anti-Myc (mouse monoclonal, 1:3000, Roche; 9E10), anti-tubulin (mouse monoclonal, 1:20,000, Iowa Hybridoma Bank), anti-Cul4 (1:1,000, a gift from Dr. Robert Duronio), anti-E2f1 (1:1,000, gift from Dr. Maki Asano), anti-Rbf2 (rabbit polyclonal, 1:5,000, R2C1) and anti-CtBP (rabbit polyclonal, 1:5,000, DNA208).

Luciferase reporter assay

Drosophila S2 cells were transfected using the Effectene transfection reagent (Qiagen) according to the manufacturer’s protocol. Typically, 1.5 million cells were transfected with 1 µg of PCNA-Luciferase reporter, 0.25 µg of pRL-CMV Renilla luciferase reporter (Promega), 200 ng of pIE-E2F1 and 200 ng of one of pAXRbf1 constructs. Cells were harvested 72 h after transfection and luciferase activity was measured using the Dual-Glo luciferase assay system (Promega) and quantified using the Veritas microplate luminometer (Turner Biosystems). Firefly luciferase activity was normalized to renilla luciferase activity.

Fluorescence-activated cell sorting and cell cycle analysis

To analyze the effects of Rbf1 proteins on Drosophila S2 cell cycle, FACS analyses were performed using cells expressing GFP-tagged Rbf1 proteins. One and a half million cells were transfected with 1 µg of pAXGFP-Rbf1 WT or pAXGFP-Rbf1 ΔIE constructs using the Effectene transfection reagent (Qiagen). Cells were harvested 4 d post-transfection and analyzed by flow cytometry to separate the GFP positive and GFP negative populations. Sorted cells were fixed with 70% ethanol and stained with propidium iodide (PI) for DNA content measurements using a BD Bioscience Vantage SE flow cytometer. The cell cycle data was analyzed through ModFit LT v3.3 (Verity Software House).

BrdU incorporation assay

Cell proliferation was assessed by examining bromodeoxyuridine (BrdU) incorporation 20 h after the addition of BrdU to the S2 cell cultures that were transiently transfected with the indicated pAXRbf1 expression constructs. Two million S2 cells were plated on polylysine coated glass coverslips and transfected with 400 ng pAX vector expressing Flag-tagged Rbf1 WT or mutants. Cells were incubated at 25°C for 1 d and then incubated in medium with 100 uM BrdU for 20 h. Cells were washed once with PBS and fixed in 4% paraformaldehyde for 15 min at room temperature. Cells were permeabilized with 0.4% Triton X-100 for 10 min and blocked with 1% BSA for 1 h. Cells were then incubated with rabbit polyclonal anti-Flag (1:250, Sigma-Aldrich; F7425) for 1 h. After two washes with TBST (20 mM Tris pH 7.5, 120 mM NaCl, 0.1% Tween 20), cells were incubated with Alexa488-cojugated chicken anti-rabbit (1:500). For detection of BrdU, cells were fixed again in 4% paraformaldehyde for 30 min, treated with 2 M HCl for 30 min and blocked with 1% BSA for 30 min. cells were then incubated with mouse monoclonal anti-BrdU (1:250, BD PharMingen) for 1 h. After two washes with TBST, cells were incubated with Alexa555-cojugated goat anti-mouse (1:500). Coverslips were mounted in Vectashield mounting medium containing 1.5 ug/ml DAPI.

RNA interference

Double-stranded RNAs were transcribed with MEGAscript T7 High Yield Transcription Kit (Ambion). S2 cells were maintained in Sf-900 II serum-free medium (GIBCO) supplemented with 0.5% Penicillin-Streptomycin. One and a half million cells were incubated with fresh medium containing 15 µg dsRNA for 30 min and then transfected with 200 ng pAX-Rbf1. Cells were grown for 5 d at 25°C.

Acknowledgments

We thank Nick Dyson for sharing the PCNA-luciferase reporter gene, Maxim Frolov for the pIE-E2F1 construct, Robert Duronio for the anti-Cul4 antibody, Maki Asano for the anti-E2F1 antibody, Louis King for expert assistance with flow cytometry and Satyaki Sengupta for his thoughtful comments. This work was supported by the MSU Gene Expression in Development and Disease Group, and a grant from the NIH to D.N.A and R.W.H (1R01GM079098).