Differential regulation of FBXW7 isoforms by various stress stimuli

Abstract

Fbxw7 is a tumor suppressor mutated in a wide range of human cancers. It serves as the substrate recognition component of SCF E3 ubiquitin ligases, and intensive effort was made to identify its substrates. Some of the substrates are central regulators of the cell cycle, cell fate determination, and cellular survival. Unlike the many efforts aimed at identifying novel targets, little is known about the regulation of Fbw7 isoform expression. In this study, we examined the mRNA expression of different FBXW7 isoforms during the cell cycle and after exposure to various stress stimuli. We observed that Fbw7β is induced by all the stress stimuli tested, mostly, but not exclusively, in a p53-dependent manner. In fact, FBXW7β was found to be the most potently induced p53 target gene in HCT-116 cells. Expression of FBXWα and γ is p53-independent and their responsiveness to most stress stimuli is limited. Furthermore, their pattern of stress responsiveness is very different from that of the β isoform. Under certain conditions, the same genotoxic agent stimulates induction of β and repression of α. Analysis of FACS-sorted cells in specific phases of the cell cycle by using fluorescent ubiquitination-based cell cycle indicator (FUCCI), showed a significant repression of the γ isoform during the S phase of normal cycling HCT-116 cells. Altogether, this study suggests differential regulation of the 3 Fbw7 isoforms.

Introduction

FBXW7 acquires increasing significance during the last years as a tumor suppressor, which is mutated in many types of cancer. Novel functions of FBXW7 are constantly being discovered. Accumulated clinical and experimental data implicated it in the preservation of genomic stability,¹ prevention of transformation and metastasis,^2,3 involvement in stem cell differentiation,⁴ and cell survival in response to drug exposure.^5-7 The ability to participate in all these activities is associated with its role as a substrate recognition unit of SCF E3 ubiquitin ligases, and a huge effort was made to identify its targets during the last years. Indeed, a large variety of targets have been identified, some of which are involved in cell cycle regulation, such as cyclin E,⁸ cell cycle progression and cell size determination, such as c-Myc^9,10 and c-Jun,^11,12 regulators of cell fate determination and differentiation, such as Notch,^13,14 the anti-apoptotic Bcl-2 superfamily member Mcl-1,^5,6 and others (reviewed in ref. 3). Three isoforms of FBXW7 (α, β, and γ) direct the variety of substrates to degradation.¹⁵ All 3 isoforms are derived from the same locus on chromosome 4. They share 10 C-terminal exons, which encode for the F-box and substrate-recognition motifs. Thus, potentially every isoform can identify every substrate containing a phosphorylated consensus Cdc4 phospho-degron motif. However, additional factors determine the interactions between a substrate and a specific isoform: the subcellular localization and the abundance of the FBXW7 isoform at a specific setting. A unique N-terminal exon specific for each isoform determines their subcellular localization.¹⁶ Isoform α, the most abundant isoform in proliferating cells, is mainly located in the nucleus, whereas isoform β is cytoplasmic and γ nucleolar. Therefore, some degree of specificity is contributed by co-localization of the substrate with the specific FBXW7 isoform. Nevertheless, several levels of complexity are added to the specificity provided by localization. The ability of FBXW7 isoforms to heterodimerize provides a potential for alterations in localization via dimerization.¹⁷ In addition, a combined activity of 2 isoforms in the degradation of a substrate was reported, for instance in the degradation of cyclin E.¹⁸

The abundance of each isoform at a given scenario is an additional factor determining the efficiency of substrate degradation at a given time. Expression of the 3 isoforms is controlled by different promoters and upstream regulation elements. Consequently, the mRNA expression level of the different isoforms significantly varies in 4 different human cell lines with averaged levels of FBXW7α 100-fold higher than those of the γ isoforms.¹⁹ Unlike the efforts put on identifying novel substrates, the regulation mechanisms controlling the expression of the different isoforms are still obscure. Handfuls of factors have been suggested to regulate FBXW7 expression. p53 upregulates FBXW7β,^20,21 but was also suggested to affect FBXW7α expression,²² whereas presenilin,²³ C/EBPδ,²⁴ Hes5,²⁵ and the 2 microRNAs miR223 and miR27a^26,27 might repress FBXW7 isoforms. In this paper, we further addressed the issue of isoform expression by monitoring the abundance of the mRNA of each FBXW7 isoform during the cell cycle and after exposure of cells to various stress stimuli with special emphasis on the role of p53 in the regulation of FBXW7β during stress responses.

Results

Expression of FBXW7 isoforms is altered by agents commonly used to synchronize cycling cells

The 3 isoforms of FBXW7 are frequently mutated in cancer at the substrate recognition site, which is encoded by shared exons.²⁸ Thus, mutations occurring in cancer impair the function of all isoforms. However, the abundance of each isoform at a given scenario has to be considered when assessing the normal function of FBXW7. We were particularly interested in monitoring the expression of each isoform during cell cycle progression. mRNA levels of each isoform were determined by reverse transcriptase quantitative PCR (real-time PCR, RTqPCR) using forward primers specific to the first unique exon. The low expression levels of the FBXW7 proteins and the lack of isoform specific antibodies that can distinguish between isoforms of similar size (β and γ), forced us to focus on mRNA expression. Cell cycle-dependent expression is usually measured by sampling pre-synchronized culture at different times after the release from the cell cycle arrest.²⁹ To determine the compatibility of synchronization methods for measuring FBXW7 isoforms expression, we determined the effects of 2 common synchronization protocols, aphidicolin treatment and serum starvation, on isoform expression in HCT-116 colon carcinoma cells. This cell line was chosen in this study, as they express functional p53 tumor suppressor gene. Aphidicolin is a reversible inhibitor of DNA polymerase that leads to inhibition of DNA synthesis and, therefore, arrests cells at the beginning of the S phase.³⁰ However, replication stress also activates the G₁ checkpoint to arrest the proliferation at this stage.³¹ HCT-116 cells were treated with 5 μg/ml aphidicolin for 24 h, and changes in cell cycle and FBXW7 isoform expression were monitored. As expected, the treatment changed the cell cycle distribution, enriching the G₁ population and reducing the cell fraction at the G₂/M, without any induction of apoptosis at the time-point analyzed (Fig. 1A). However, despite the desirable effect on cell cycle profile, measurement of FBXW7 isoform expression reveals that aphidicolin synchronization protocol cannot be used, as the expression of FBXW7β mRNA was elevated 11-fold by the treatment (Fig. 1B). The expression of FBXW7α and γ did not significantly change. Growing the cells at low concentrations (1%) of fetal calf serum (FCS) also arrested the cells at the G₁ phase (Fig. 1A). Under these conditions, FBXW7α expression is reduced, while FBXW7β expression is elevated (Fig. 1C). The mitotic inhibitor vinblastine elevated FBXW7β expression as well (see Fig. 5), thus excluding its use for cell cycle synchronization to study FBXW7 isoform expression pattern in a specific cell cycle phase.

Figure 1. Expression of FBXW7 isoforms is altered by aphidicolin treatment and serum starvation. HCT-116 cells were treated with 5 μg/ml aphidicolin or incubated in a low serum containing medium (1% FCS), harvested at the indicated times (hours) and analyzed for cell cycle by propidium iodide staining and flow cytometry analysis (A) or for expression of the 3 FBXW7 isoforms using RTqPCR and specific primers for each isoform (B, 5 μg/ml aphidicolin; and C, low serum containing medium). Expression of the isoforms was normalized to GAPDH, and the value of expression in control cultures was determined as 1. Continuous black line, FBXW7β; gray line, FBXW7γ; dashed lines, FBXW7α.

Figure 5. Expression of FBXW7 isoforms after exposure to stress-inducing agents. HCT-116p53^+/+ and HCT-116p53^−/− cells that were treated in parallel with those subjected to cell cycle analysis (Fig. 4), were harvested at the indicated time points, and mRNA levels of the 3 different isoforms of FBXW7 were determined by RTqPCR. GAPDH and HPRT were used for normalization, and the basal level of each isoform prior to treatment was regarded as 1. White bars, relative mRNA expression in HCT116p53^+/+ ; and black bars, relative mRNA expression in HCT116p53^−/− cells.

Cell cycle-dependent changes in the expression of FBXW7 isoforms

Since the various synchronization protocols described above induce alterations in the expression of different Fbxw7 isoforms, we tried another approach for analyzing their expression during cell cycle progression. The protocol of fluorescent ubiquitination-based cell cycle indicators (FUCCI) enables the detection and sorting of cells at various phases of the cell cycle by introduction of fluorescently labeled probes (Cdt1 and Geminin) that are specifically degraded at defined stages of the cell cycle.³² Thus, cells are stained red at the G₁ phase due to high mKO2–Cdt1 expression while stained green during the G₂/M phase, where mAG-Geminin is abundantly expressed. Only cells at the S phase are double stained. HCT-116 cells, stably expressing these 2 probes (hereafter termed HCT-FUCCI) were analyzed by flow cytometry to associate the red/green emission with different cell cycle stages, which were determined in concert with Hoechst 33342 co-staining to visualize the DNA content. Using this system, we were able to clearly define cells in G₁, S, and G₂/M phases (Fig. 2). Cells in each cell cycle phase were isolated separately by careful FACS sorting. To homogeneously isolate M-phase cells, we used the technique of mitotic shake-off. Next, using FACS, we sorted cells from asynchronous cultures which have not been pre-treated or exposed to any stress stimuli, according to their emission. RNA was extracted from the cells, and the expression of the 3 FBXW7 isoforms was determined at each cell cycle stage using RTqPCR. Cyclin E1 gene expression (CCNE1) and Stk15 were used as controls to demonstrate the validity of the method. CCNE1 expression is high at G₁ and S and declines at G₂ and M, whereas Stk15 spikes at the G₂ phase of the cell cycle, as previously reported³³ (Fig. 3A). Interestingly, all isoforms are expressed at the highest levels at the G₁ stage. However only a slight decrease in FBXW7α and FBXW7β expression were observed at the S phase relative to G₁ (P = 0.034 and 0.0328, respectively), while at other cell cycle stages, there were no statistical differences. Unlike the α and β isoforms, the expression of the FBXW7γ showed a more substantial reduction upon the transit from the G₁ to the S stage. Its expression declined more than 7-fold at the G₁/S transit and inclined later, at the M phase. These data suggest that FBXW7γ is differentially regulated during cell cycle.

Figure 2. Use of the FUCCI system for isolation of cells at different phases of the cell cycle. Asynchronous HCT-116-FUCCI cells were stained with Hoechst 33342 and subjected to flow cytometry analysis. (A) Gating of high mKO2, low mAG population (P1 = G₁); high mAG, low mKO2 population (P2 = G₂/M); and a population expressing both colors (P3 = S). (B) DNA content of each population is indicated by comparison to the DNA content measured by Hoechst 33342 staining. The histogram at the bottom represents the entire non-gated population. (C) Reconstitution of the 3 gated populations indicates that each represents the expected cell cycle specific phase.

Figure 3. Analysis of FBXW7 isoforms expression at different cell cycle stages. HCT116-FUCCI cells were sorted according to the analysis presented in Figure 2. CCNE1 and STK15 and the indicated FBXW7 isoforms were measured by RTqPCR. Mitotic cells were collected by mitotic shake off. GAPDH levels were used for normalization. Lowest expression levels of each isoform were considered as 1. *P < 0.05.

Genotoxic stress stimuli alter the expression of specific FBXW7 isoforms

As of today, the responsiveness of specific isoforms of FBXW7 to various genotoxic agents has been poorly studied. Therefore, we examined the expression of FBXW7 isoforms after exposure of HCT-116 cells to genotoxic agents that arrest cell cycle progression at different stages or to agents promoting cell death. HCT-116 cells were exposed to the chemotherapeutic agents cisplatinum (CDDP), etoposide or vinblastine, or to the specific p53 activator Nutlin 3, which prevents the interaction between Mdm2 and p53. The cells were also exposed to UV radiation, the most common environmental carcinogen. Cell cycle was analyzed at different time points and apoptotic death determined 48 h post-treatment. Vinblastine, CDDP, and etoposide treatments transiently arrested proliferation after 24 h as depicted in Figure 4. After 48 h, treatments with these agents led to apoptosis, killing 24 ± 2, 30 ± 3, and 35 ± 4% of the cells, respectively (Fig. 4). Nutlin 3 treatment resulted in prolonged G₁ and G₂ growth arrest. HCT-116 showed relatively high resistance to UV irradiation, with relatively low appearance (11 ± 1%) of apoptotic cells 48 h post-treatment. These conditions were used to assess the responsiveness of FBXW7 isoforms expression to the different stress stimuli. mRNAs were extracted at different time points in parallel to flow cytometry analyses, and the expression of each isoform was determined by RTqPCR. The expression analyses revealed a very high responsiveness of FBXW7β to all the stress conditions examined, spiking to more than 30-fold 48 h after exposure to vinblastine and 24-fold after exposure to CDDP (Fig. 5). Unlike the β isoform, except for UV exposure, the 2 other isoforms showed significantly fewer alterations (Fig. 5). FBXW7α is transiently repressed by CDDP, and repressed in a time-dependent manner following etoposide and Nutlin 3 treatment. In all these cases, the extent of repression was small and did not exceed a 2-fold change. Similar to FBXW7α, the changes in FBXW7γ expression were only minor and transient after exposure to CDDP and vinblastine. In contrast, UV radiation resulted in unique consequences with a strong repression of FBXW7α (up to 10-fold) and weaker repression of FBXW7γ in addition to 6-fold induction of FBXW7β, summing in a 60-fold difference between FBXW7α and FBXW7β 24 h post-radiation, a phenomenon that has been previously described.³⁴ Altogether, these data demonstrate that the 3 FBXW7 isoforms are differentially regulated under stress conditions.

Figure 4. Cell cycle analysis of HCT-116 cells before and after treatment with various stress-inducing agents. HCT-116 cells were treated for the indicated times with 20 μM CDDP, 20 μM etoposide, 100 nM Vinblastine, 15 μM Nutlin-3, or harvested at the indicated time after exposure to 30 J/m² UVC. Harvested cells were fixed in methanol, stained with propidium iodide, and subjected to cell cycle analysis by flow cytometry. The numbers at 48 h represent the percentage of sub-G₁ cells with less than 2N DNA content, which are considered apoptotic cells.

As p53 was suggested to affect FBXW7 isoforms expression,^20,22 it was intriguing to detect its contribution to the responsiveness of each isoform to stress stimuli. To that end we analyzed the expression of the isoforms in the isogenic HCT-116p53^−/− cells³⁵ to find out whether the above-described changes are dependent on p53. Nutlin-3, which specifically activates p53, did not induce FBXW7β in HCT116p53^−/− cells (Fig. 5), indicating a p53-dependent induction. The induction of FBXW7β by UV radiation and vinblastine is also p53-dependent (Fig. 5), while its induction by CDDP or etoposide is only partially dependent on p53, since this isoform is induced in HCT-116p53^−/− cells (Fig. 5). In fact, 24 h after etoposide treatment, the levels of FBXW7β were elevated higher in the p53-deficient cells than in the p53-proficient ones (21-fold and 13-fold, respectively) (Fig. 5). These data suggest that FBXW7β is the only isoform regulated by p53, and its induction under certain conditions may also occur in a p53-independent manner.

FBXW7β is the most potently induced p53 target gene in HCT-116 cells

The previous experiment demonstrated the dependence of FBXW7β induction on p53. Therefore, it was of interest to compare the extent of stress responsiveness as well as the p53-dependence of FBXW7β to other known p53 target genes. To determine the basal contribution of p53 to FBXW7 expression, we compared the expression of known p53 target genes such as p21/Cip1, NOXA, HDM2, and GADD45α to FBXW7β in p53-proficient vs. p53-deficient HCT-116 cells. FBXW7α and γ were used as controls for p53-independent expression. As shown in Figure 6A, the basal expression of FBXW7β in p53-deficient cells is only 1/10 of that in p53-proficient HCT116 cells. This reduced expression was the most significant among the group of p53 target genes examined. For comparison, expression of p21, the next highly repressed target gene, was reduced 5.9-fold. HDM2 expression was repressed 2-fold and the others only 1.4-fold. The expression of FBXW7α and γ, which are not regulated by p53 under stress conditions, was not affected at all by the absence of p53, indicating the specificity of the effect.

Figure 6. Comparison of expression of FBXW7 isoforms to other known p53 regulated genes. (A) mRNA levels of the indicated genes were measured by RTqPCR, and the ratio between mRNA expression in HCT116p53^−/− and HCT116p53^+/+ cells is presented. HCT116 cells were treated with 15 μM Nutlin-3 (B) or 20 μM etoposide (C) and the expression of the indicated genes was tested at different time points (hours) after treatment. GAPDH and HPRT were used as an internal control.

Next we compared the p53-dependent expression of FBXW7β and other known p53 target genes in HCT-116 following treatment with the p53 activator Nutlin-3 (Fig. 6B). Again, both the kinetics and extent of FBXW7β induction was the fastest and highest among the examined genes. p21 was induced to a little lower extent, followed by HDM2 and PUMA, which reached at their maximal levels to 75% and 56%, respectively, of the induction observed in FBXW7β. Exposing the cells to etoposide, which induces FBXW7β in both p53-dependent and p53-independent manners, resulted in even a bigger difference between the different p53-target genes (Fig. 6C). Here, the maximal induction of FBXW7β was 2-fold higher than that of p21. These results demonstrate that in HCT-116 cells, FBXW7β is in general the most responsive p53 target gene among the genes examined in this study.

Discussion

In this study we monitored the expression of FBXW7 isoforms during cell cycle progression and after exposure to various stress stimuli. We have demonstrated that common methods routinely used for synchronization of cell cultures prior to analysis, such as chemical inhibitor of the S phase, inhibitors of the M phase, or even growth at low serum containing medium, may affect the expression of genes subjected to analysis. In our case, S-phase inhibitor aphidicolin generates a replicative stress resulting in FBXW7β induction, probably through the activation of p53. Mitotic inhibitors also activate p53 to prevent endoreduplication, a process also controlled by FBXW7.⁷ Consequently, FBXW7β is potently induced, frequently, but not exclusively, in a p53-dependent manner. Using Nutlin-3 treatments, we have demonstrated that FBXW7β is the only p53-responsive isoform, confirming previously reports.^20,21 Moreover, in this study we show that the responsiveness of FBXW7β in HCT-116 cells is exceptionally high. It exceeded the induction of any other p53-regulated gene examined in this study, even the highly responsive p21/CIP1 gene. The most striking example is the p53-dependent 32-fold induction following vinblastine treatment. Surprisingly, currently the biological functions of FBXW7β within the p53 pathway are unknown. In vivo, tumors derived from irradiated p53-heterozygous mice acquire mutations in the Fbw7 gene, whereas tumors derived from p53-null mice do not.³⁶ However, these mutations usually affect all isoforms and cannot exclude the possible activity of another isoform upstream of p53. More important is the observation that inactivation of FBXW7β expression by promoter hypermethylation is associated with favorable prognosis in primary breast cancer,³⁷ suggesting that in some cases, this isoform may contribute to the progression of tumorigenesis. Biochemically, mice deficient in FBXW7β develop normally, but primary cultures of neurons prepared from the mutant mice were more vulnerable to oxidative stress than were those prepared from wild-type mice.³⁸ Although suggested to reside in the ER, FBXW7β does not affect apoptosis driven by ER stress.³⁸ Aside from the high p53 responsiveness, FBXW7β expression is also regulated in a p53-independent manner, especially after exposure to etoposide, presumably relying on other members of the p53 family p63 and p73, which are activated by the drug.^39,40 No other upregulators of FBXW7β are currently known. In contrast, the transcription of the β isoform was recently shown to be repressed by the Notch-regulated HES5, thus forming a Notch/Hes5/Fbw7β positive feedback loop required for proper intestinal progenitor cell and neural stem cell differentiation.²⁵ The α isoform is indirectly repressed by Presenilin²³ and directly by C/EBPδ ²⁴. Fbxw7 expression is also repressed by microRNAs such as miR-27²⁷ and miR-223.²⁶ As the recognition sequences of the microRNAs reside at the shared 3′UTR of the FBXW7 gene, their presence is expected to down-modulate all 3 isoforms to a similar extent. Therefore, they are unlikely to contribute to the differential expression of the isoforms observed under the stress conditions described in this paper.

During cell cycle progression, FBXW7α expression is moderately higher at the G₁ phase. In one report, the protein level of FBXW7α in chemically synchronized cells did not differ significantly through the cell cycle,¹⁹ suggesting the existence of post-transcriptional events affecting the stability of this isoform. However, another report²⁷ suggests that FBXW7α is indeed regulated through the cell cycle. A cooperative effect of the α and γ isoforms in the degradation of cyclin E was reported.¹⁸ A later report suggested that the cooperation is cell type-dependent, and FBXW7α is the major regulator of cyclin E in HCT-116 cells.¹⁹ Nevertheless, some degree of redundancy/cooperative degradation is implied by the higher levels of cyclin E in cells deficient in all isoforms in comparison to cells deficient in specific ones.¹⁹ Furthermore, cell cycle analysis using chemical inhibitors revealed high levels of cyclin E during the S phase even in FBXW7α-deficient cells. We suggest that the decline in FBXW7γ expression at this stage may contribute to the elevation in cyclin E levels. The basal expression of FBXW7γ is the lowest among the 3 isoforms in HCT-116 as well as several other cell lines, suggesting that low levels of the protein are sufficient to carry out its activity.¹⁹ Further reduction may result in the temporal abrogation of FBXW7γ activity. Keeping c-Myc activity in balance was suggested to be controlled by FBXW7γ.^9,10 However, recently it was reported that the isoform specific-dependence of c-Myc degradation is also cell type-specific and differs between U2OS and HCT-116 cells.¹⁹ Fbxw7γ regulates the nucleolar pool of c-Myc in U2OS, while FBXW7α does it in HCT-116 cells. The higher expression of FBXW7β at the G₁ and M phases is expected, as the activity of its regulator p53 increases when approaching the G₁/S and G₂/M checkpoints.

Altogether, the data presented in this paper show that the 3 isoforms of FBXW7 are differentially regulated and therefore cannot be categorized as one protein. Seemingly, due to different abundance and intracellular compartmentalization, their cellular functions are expected to differ. Emerging data suggest that this is the case, and further studies are required to identify their separate physiological roles.

Materials and Methods

Cell lines

Human HCT-116 and HCT-116 p53^−/− colon carcinoma cells were grown in McCoy medium supplemented with 10% FCS at 37 °C in 5% CO₂ humidified atmosphere. For generation of HCT-FUCCI cells, the cells were infected with viruses prepared from the plasmids encoding mKO2-hCdt1 (Δ30–120) and mAG-Geminin (Δ1–110), which were previously described.³² Infected cells were selected with zeocin, and the nascent cultures were subjected to FACS sorting without prior treatment at least one week after the end of the selection with antibiotics.

FACS sorting of HCT-FUCCI

HCT-FUCCI cells were seeded 36–40 h prior to FCS sorting to reach 50–70% confluence. The cells were trypsinized, re-suspended in ice cold PBS, and subjected to FACS sorting according to color using gates determined by a preliminary Hoechst 33342 staining (Fig. 2). The untreated sorted cells were collected directly into TRI-Reagent (Sigma-Aldrich) for isolation of RNA. The efficiency of FACS sorting was analyzed in parallel by analyzing the collected cells, showing more than 95% purity.

mRNA analysis

Reverse transcriptase quantitative PCR (RTqPCR) was used for analysis of the mRNA levels. mRNA isolated from cells extracted in TRI-Reagent (Sigma) according to the manifacturer's instructions, was converted to cDNA using the AB high-capacity cDNA kit (Applied Biosystems). Twenty ng of converted cDNA was used for each reaction, using Kapa Sybr-Green Master Mix (Kapa Biosystems) and 300 nM of forward/reverse primer set. Internal standards were GAPDH, β-actin, and HPRT. The real-time PCR was performed in an AB real-time PCR apparatus (7300 Real-Time PCR System). The primers used in this study are presented in Table 1.

Table 1. Primers used for real-time PCR.

Gene	Forward primer	Reverse primer
GAPDH	TCGACAGTCA GCCGCATCTT CTTT	ACCAAATCCG TTGACTCCGA CCTT
β-Actin	ATGGATGATG ATATCGCCGC GCTC	ATAGGAATCC TTCTGACCCA TGCC
HPRT	ACTGGCAAAA CAATGCAGAC TTT	GGTCCTTTTC ACCAGCAAGC T
FBXW7α	AGTAGTATTG TGGACCTGCC CGTT	GACCTCAGAA CCATGGTCCA ACTT
FBXW7β	TATTGTCAGA GACTGCCAAG CAGC	GACCTCAGAA CCATGGTCCA ACTT
FBXW7γ	CCATGGCTTG GTTCCTGTTG ATCT	GCCTTGGGCA ATGATGCTAA TGCT
p21	GGCAGACCAG CATGACAGAT T	GCGGATTAGG GCTTCCTCTT
HDM2	GAGCTTCAGG AAGAGAAACC TTCA	GGCGTTTTCT TTGTCGTTCA C
NOXA	CGGAGATGCC TGGGAAGAA	ACACTCGACT TCCAGCTCTG C
GADD45a	TGGTGACGAA TCCACATTCA TC	ATTGATCCAT GTAGCGACTT TCC
CCNE1	GCAGGATCCA GATGAAGAAA TGGC	CAGACTGCAT TATTGTCCCA AGGC
STK15	TTGGGTGGTC AGTACATGCT CCAT	ACCTTCTCAT CATGCATCCG ACCT

Cell cycle analysis

Trypsinized cells together in the medium were collected and fixed in ice-cold methanol, and kept at −20 °C for at least 1 d. The fixed cells were rehydrated in PBS, and incubated in PBS containing 50 μg/ml RNase, followed by incubation with 25μg/ml propidium iodide. Ten thousand events were read on a FACScan analyzer (Becton Dickinson) using low speed flow. Sub-G₁ cells were considered apoptotic cells.

Glossary

Abbreviations:

FUCCI

fluorescent ubiquitination-based cell cycle indicator

RTqPCR

reverse transcriptase quantitative PCR

CDDP

cisplatin

FCS

fetal calf serum

10.4161/cc.26591

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.