Oncogene (2005) 24, 8314–8325
& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc
Y-box factor YB1 controls p53 apoptotic function Craig Homer1, Deborah A Knight1, Lynne Hananeia1, Philip Sheard2, Joanna Risk1, Annette Lasham3, Janice A Royds1 and Antony W Braithwaite*,1 1
Pathology Department, Dunedin School of Medicine, University of Otago, PO Box 913, Dunedin, Otago 9001, New Zealand; Physiology Department, School of Medical Sciences, University of Otago, Box 913, Dunedin, New Zealand; 3Genesis Research and Development Corporation Limited, PO Box 50, Auckland, New Zealand 2
Nuclear localization and high levels of the Y-box-binding protein YB1 appear to be important indicators of drug resistance and tumor prognosis. YB1 also interacts with the p53 tumor suppressor protein. In this paper, we have continued to explore YB1/p53 interactions. We report that transcriptionally active p53 is required for nuclear localization of YB1. We go on to show that nuclear YB1 regulates p53 function. Our data demonstrate that YB1 inhibits the ability of p53 to cause cell death and to transactivate cell death genes, but does not interfere with the ability of p53 to transactivate the CDKN1A gene, encoding the kinase p21WAF1/CIP1 required for cell cycle arrest, nor the MDM2 gene. We also show that nuclear YB1 is associated with a failure to increase the level of the Bax protein in normal mammary epithelial cells after stress activation of p53. Together these data suggest that (nuclear) YB1 selectively alters p53 activity, which may in part provide an explanation for the correlation of nuclear YB1 with drug resistance and poor tumor prognosis. Oncogene (2005) 24, 8314–8325. doi:10.1038/sj.onc.1208998; published online 12 September 2005 Keywords: YB1; nuclear localization; p53 functions; apoptosis
Introduction YB1 is a member of the highly conserved Y-box family of proteins that regulate gene transcription by binding to either double- or single-stranded TAACC elements (the Y box) contained within many eukaryotic promoters (Wolffe, 1994; Mantovani, 1998). Transcriptional targets of YB1 include genes associated with cell death such as FAS (Lasham et al., 2000), encoding a cell deathassociated receptor, and the tumor suppressor gene, TP53 (Okamoto et al., 2000; Lasham et al., 2003). YB1 also positively regulates genes associated with cell proliferation including epidermal growth factor receptor (EGFR or c-ERBB1) (Sakura et al., 1988), matrix *Correspondence: AW Braithwaite; E-mail:
[email protected] Received 24 January 2005; revised 15 June 2005; accepted 6 July 2005; published online 12 September 2005
metalloproteinase-2 (MMP-2) (Mertens et al., 1997, 1999), and DNA topoisomerase IIa (Shibao et al., 1999). Several studies have demonstrated that YB1 is directly involved in the cellular response to genotoxic stress. In response to DNA-damaging agents, such as the chemotherapeutic drug cisplatin and UV irradiation, YB1 induces expression of the multidrug resistance 1 (MDR1) gene through increased binding to a Y-box element within the MDR1 promoter (Uchiumi et al., 1993; Asakuno et al., 1994; Ohga et al., 1998). In addition, overexpression of YB1 has been observed in cell lines that are resistant to cisplatin, and depletion of YB1 results in increased sensitivity to cisplatin (Ohga et al., 1996). These data suggest that YB1 plays an important role in conferring drug resistance on tumor cells. Furthermore, microarray analysis of cells resistant to multiple DNA-damaging agents has shown that YB1 is one of about 20 genes that are transcriptionally upregulated (Levenson et al., 2000), again suggesting a role for YB1 in DNA damage responsiveness. Y-box family members have also been shown to bind RNA in a sequence-specific manner (Chen et al., 2000), and interact with an RNA-splicing factor (Raffetseder et al., 2003). YB1 also stimulates RNA splicing from different templates (Raffetseder et al., 2003; Stickeler et al., 2001) and is a potent and general stabilizer of capped mRNA (Evdokimova et al., 2001). Thus, YB1 appears to have pleiotropic functions (Kohno et al., 2003) in transcription, RNA splicing and translation (Matsumoto and Wolffe, 1998; Sommerville, 1999). YB1 is predominantly localized to the cytoplasm in both normal and tumor cell lines, but can translocate to the nucleus in response to a variety of environmental stresses including DNA damage (Koike et al., 1997; Okamoto et al., 2000; Holm et al., 2002). However, in vitro, we found this to be uncommon occurring in less than 10% of cells examined after treatment with several DNA-damaging agents (Zhang et al., 2003). Nonetheless, several reports indicate that nuclear translocation of YB1 is necessary for it to function as a transcription factor, as might be expected (Bargou et al., 1997; Oda et al., 1998, 2003; Stenina et al., 2001; Holm et al., 2002). In addition, nuclear YB1 appears to be important clinically as it has been associated with poor prognosis in several types of human cancer (Bargou et al., 1997; Oda et al., 1998, 2003; Kamura et al., 1999;
YB1 controls p53 apoptotic function C Homer et al
8315
a
Uninduced
C/ER/2
C/ERp53/8
% nuclear YB1
Shibahara et al., 2001). Thus, the regulation of YB1 nuclear translocation is an important aspect of YB1 function. Recent studies have investigated the mechanism of YB1 nuclear translocation. One report suggests that it is associated with a transient arrest of the cell cycle at the G1/S phase boundary (Jurchott et al., 2003), another with binding the splicing factor SRp30c (Raffetseder et al., 2003) that shuttles in and out of the nucleus, and we have reported that it may require overexpression of the p53 tumor suppressor (Zhang et al., 2003). Recently, nuclear translocation of YB1 has also been shown to require phosphorylation by Akt (Sutherland et al., 2005). The relationship between these sets of observations is not clear but a role for p53 and nuclear translocation of YB1 being cell cycle associated are reconcilable, as one of the functions of p53 is to induce a cell cycle arrest (Prives and Hall, 1999). In addition, subcellular trafficking of the proteins encoded by two of p53’s target genes, MDM2 and CDKN1A, are regulated by Akt phosphorylation (Zhou et al., 2001a, b). Furthermore, it is also not clear why nuclear YB1 is a marker of poor tumor prognosis, but one possibility, is that it interferes with the normal functions of p53 in tumor suppression. In this paper, we have investigated the effect of nuclear localization of YB1 on p53 functions. We report that YB1 selectively inhibits the ability of p53 to cause cell death by preventing transactivation of proapoptotic genes.
100 80
C/ER/2
C/ERp53/7
60 40 20 0
b C/ERp53/8 Nuclear YB1 Cytoplasmic YB1
C/ERp53/7 Nuclear YB1
Results Induction of p53 in ERp53 cells results in translocation of YB1 to the nucleus Although nuclear YB1 appears to be commonplace in many tumors, as indicated above, we found that it is rare in tumor cells in culture even after stress with DNAdamaging agents. It is difficult therefore to determine the impact of nuclear YB1 on cellular functions. However, in our previous report we took advantage of a p53-inducible cell line we had constructed (C/ERp53/7) to confirm that p53 is important for the nuclear translocation of YB1 (Zhang et al., 2003). Quantitation showed that nearly 100% of p53-positive cells contain nuclear YB1 (Zhang et al., 2003). This result is confirmed here using the same cells as well as a different cell clone (C/ERp53/8). Nuclear YB1 is found in most cells only after p53 induction (treatment with the estrogen analogue 4-hydroxytamoxifen (4-OHT), see below) but is not found in a control cell line not expressing p53 (C/ER/2) but treated in exactly the same way (Figure 1a). This experiment was extended by inducing p53 and then fractionating cell lysates into nuclear and cytoplasmic components, followed by Western blotting. Results for both C/ERp53/7 and C/ERp53/8 cells (Figure 1b) show that YB1 is predominantly cytoplasmic until after p53 induction whereupon it translocates to the nucleus. These cells then afford useful tools to investigate the impact of YB1 on cellular functions, in particular on p53.
Cytoplasmic YB1 Figure 1 Induction of ERp53 causes nuclear translocation of YB1. C/ER2 cells and C/ERp53 cell clones were treated with increasing concentrations of 4-OHT and analysed for YB1 subcellular localization by (a) immunofluorescent staining and quantitation by direct cell counting or (b), by fractionation into nuclear and cytoplasmic components and detection of YB1 by Western blotting. Concentrations of 4-OHT used for cell counting are 0.2 and 2 mM, while in part (b) 4-OHT was added at 0.2, 0.5, 1 and 2 mM. These experiments have been carried out multiple times
ER/p53 is transcriptionally active The reported principle of the induction method for these cell lines is that the ER/p53 fusion protein (see Materials and methods) is retained in the cytoplasm until treatment with 4-OHT, which then causes the ER/p53 to translocate to the nucleus thereby becoming active as a transcription factor (Morgenstern and Land, 1990; Littlewood et al., 1995). However, in the course of analysing these cells in detail, we noticed that p53 was located in the nucleus both before and after (Figure 2a, first row) treatment with 4-OHT, but no p53 was observed in the C/ER/2 cells as expected (Figure 2a). Similar results were obtained with clone 8 and other ER/ p53-derived cell clones (data not shown). These results suggest that YB1 does not translocate to the nucleus by simply binding p53, but probably requires ER/p53 to become functionally active. Oncogene
YB1 controls p53 apoptotic function C Homer et al
8316
a
C/ERp53/7 0.2
C/ER/2 0
2
0
p53 (DO-1)
p21WAF1/CIP1
Mdm2
b
Relative Luciferase Activity
p53 (pAb1620)
CDKN1A
40
1.5
30
1
20
0.5
10
0
Induction of ER/p53 does not cause cell death but still allows cell cycle arrest
0 24 48 72 96 100
0 0
24 48 72 96 120 0 Time (hours)
24 48 72 96 120 Time (hours)
C/ERp53/7 uninduced
C/ER/2 uninduced
Figure 2 ERp53 is transcriptionally active only after treatment with 4-OHT. (a) C/ER2 cells and C/ERp53/7 cells were treated with 4-OHT and 48 h later they were analysed by immunofluorescent staining using antibodies to p53 (DO-1) and (pAb1620), p21WAF1/CIP1 and Mdm2. (b) C/ERp53/7 cells were treated with 4-OHT and transfected with a CDKN1A-luciferase reporter with and without a p53 response element (CDKN1ADp53). At indicated times after induction of p53, cells were harvested and luciferase activity determined
This idea was examined further by measuring the expression of some p53 target genes before and after treatment with 4-OHT. To do this, immunofluorescence was carried out using antibodies to Mdm2 and p21WAF1/CIP1 as well as studies using the CDKN1A promoter, the gene of which encodes p21WAF1/CIP1, and a similar promoter without the p53 response element (CDKN1ADp53). Results from the immunofluorescence experiments (Figure 2a) show that after treatment of C/ERp53/7 cells with 4-OHT, p21WAF1/CIP1 and Mdm2 are both strongly detected but are barely detectable in the uninduced cells and in the C/ER/2 cells. Similarly, treatment of cells with 4-OHT causes a transactivation of the CDKN1A promoter, but only when the p53 response element is present (Figure 2b). Thus, the ER/p53 Oncogene
becomes transcriptionally competent only after 4-OHT treatment as shown previously (Littlewood et al., 1995), but the mechanism of ER/p53 ‘activation’ is different (Morgenstern and Land, 1990; Littlewood et al., 1995). To determine whether the mechanism of p53 activation involves a conformational change, rather than a translocation, C/ERp53/7 cells were treated with 4-OHT and stained with the conformation-specific antibody pAb1620. Results show that only after 4-OHT treatment does the ER/p53 in C/ERp53/7 cells become immunoreactive with the antibody (Figure 2a, bottom panel), consistent with a conformational change occurring. However, there is no change in the total amount of p53 as determined by immunofluorescence (DO-1 result in Figure 2a) and Western blotting using different antibodies (data not shown). Collectively, the data from the above experiments allow us to conclude that nuclear translocation of YB1 does not involve direct binding to p53 but appears to require a transcriptionally competent p53.
To determine whether induction of p53 causes cell cycle arrest and apoptosis, we measured the number of viable and nonviable cells over several days after induction of p53. Results are shown in Figure 3. For uninduced C/p53ER/7, the viable cell number increases 11-fold by day 12, but after treatment with 0.2 mM 4-OHT the cell number remains essentially constant until 6 days rising fivefold by day 12. At 2.0 mM 4-OHT, the cell number is constant until 8 days after which it increases to fourfold by day 12. For C/p53ER/8, the uninduced cells increase about 10-fold over the time course of the experiment, but after treatment with 0.2 and 2.0 mM 4-OHT, the cell number remains essentially constant over the 12 days. Untreated C/ER/2 cells increase 15-fold by day 12 and 10-fold after 2.0 mM 4-OHT treatment. Thus, induction of p53 appears to reduce the rate of proliferation of clone C7/p53ER/7 but with the exception of the first 4–6 days, a complete growth arrest is not achieved. A complete arrest of cell division is achieved with clone C/p53ER/8. Thus, it appears that in both ER/p53 cell clones, induction of p53 prevents cell proliferation but fails to induce cell death to any greater degree than in control cells. Induction of p53 in the C/ERp53 cell clones does not lead to activation of cell death genes To explore why induction of p53 in our cell clones does not induce death we measured the protein levels of several p53 target genes. Cells were seeded, induced with different concentrations of 4-OHT, harvested after 2 days and cell lysates prepared for Western blotting. Results of one Western blot are shown in Figure 4 together with quantitation by laser scanning densitometry relative to actin staining on the same membrane from two separate Western blots. We find that both
YB1 controls p53 apoptotic function C Homer et al
8317 60
50
C/ER/2 Uninduced
40
Dead Cells
30
Live Cells
20 10
Cell Number (x 104)
Cell Number (x 104)
60
0
30 20 10 0
2 4 6 8 10 12 Days post induction 60
40 30 20 10 0
40 30 20 10 0
0
2 4 6 8 10 12 Days post induction
30 20 10 0 0
2 4 6 8 10 12 Days post induction
30 20 10 0
2 4 6 8 10 12 Days post induction
60
50
C/ERp53/8
40 30 20 10
Cell Number (x 104)
40
Cell Number (x 104)
C/ERp53/8 Uninduced
40
2 4 6 8 10 12 Days post induction
60
50
C/ERp53/7
50
0 0
60
2 4 6 8 10 12 Days post induction
60 C/ERp53/7
50
Cell Number (x 104)
C/ERp53/7 Uninduced
50
Cell Number (x 10 4 )
60 Cell Number (x 104)
40
0 0
Cell Number (x 104)
C/ER/2
50
50
C/ERp53/8
40 30 20 10 0
0 0
2 4 6 8 10 12 Days post induction
0
2 4 6 8 10 12 Days post induction
Figure 3 C/ERp53 cells undergo cell cycle arrest but do not die. C/ER and C/ERp53/7&8 cells were treated with 4-OHT as indicated and cell viability determined over 12 days using the method of trypan blue exclusion. The data shown is from a single experiment with each data point being an average from triplicate samples. These experiments have been carried out three times
p21WAF1/CIP1 and Mdm2 levels increase after p53 induction as was found with immunofluorecent analysis above, but Bax, Noxa, Gadd45 and 14-3-3s show no change. Although clearly this is not an exhaustive analysis it appears that these cells may be impaired for a p53 death response because the cell death genes are not induced. However, the ability of p53 to induce a G1 arrest of cell cycle progression is still maintained. Inhibition of YB1 causes p53-dependent cell death and an increase in the protein levels of p53-regulated cell death genes To determine whether the failure of p53 to induce cell death in these cells is due to the presence of nuclear YB1, as indicated from our previous studies, induced and uninduced cells were treated with YB1 antisense oligonucleotides (Lasham et al., 2003) and cell viability determined over time. These oligonucleotides have been used to treat several cell lines and we consistently find greater than 50% reduction in YB1 protein while other proteins are unaffected (Lasham et al., 2003; data not shown). In the present experiment, we find that treatment of induced cells with the antisense oligonucleotides reduces viability by greater than 60% whereas the uninduced cells do not show any change in cell viability (Figure 5a). In addition, the C/ER/2 cells that
do not express p53 show no change in viability after treatment with YB1 antisense oligonucleotides. Thus, inhibition of YB1 restores the ability of p53 to induce cell death. To determine if inhibition of YB1 also allows induction of p53-regulated cell death genes, the p53inducible cell lines were treated with 4-OHT as well as the non-p53-expressing control line C/ER/2, then treated with YB1 antisense oligonucleotides as used above. At indicated times after antisense treatment, cells were lysed and Western blotting carried out for Bax and Noxa. Again, one immunoblot is shown in Figure 5b as well as quantitation by laser scanning densitometry relative to an actin-loading control from two separate Western blots. We find that inhibition of YB1 expression increases the levels of both Bax and Noxa in cells expressing a transctriptionally active p53, but not in cells with a transcriptionally inert p53, nor in the control C/ER/2 cells expressing no p53. These data suggest that YB1 regulates the ability of p53 to induce cell death and cell death genes. YB1 regulates p53-dependent transactivation in C/ERp53 cells To test whether YB1 can regulate p53 transactivation, reporter assays were carried out. In the first part of this Oncogene
YB1 controls p53 apoptotic function C Homer et al
8318
levels. Thus, as was found by Western blotting, ER/p53 is unable to transactivate the promoters of proapoptotic genes. To determine whether this inability to transactivate proapoptotic gene promoters is due to YB1, C/ERp53/7 cells were next transfected with the NOXA reporter plasmid and subsequently with the YB1 antisense oligonucleotides used above. Results of this experiment (Figure 6b) show that treatment with YB1 antisense oligonucleotides causes a time-dependent increase in NOXA reporter activity, which does not occur with control oligonucleotides. These results suggest that YB1 prevents p53 from transactivating proapoptotic genes.
C/ERp53/7 C/ER/2
0
p21 Mdm2 Gadd45
Bax Noxa
YB1 selectively inhibits the ability of p53 to transactivate the BAX and NOXA promoters
Actin
Fold Induction
Fold Induction
3 2 1 0 4
4
p21/Actin
3 2 1 0 4
2 1
3 2 1 0 4
Fold Induction
Fold Induction
Bax/Actin 3 2 1
Mdm2/Actin
3
0 4
Gadd45/Actin Fold Induction
Fold Induction
4
Noxa/Actin
3 2
Direct binding of YB1 and p53 are required for transcriptional repression
1 0 C/ER/2 0
0 C/ER/2 0 C/ERp53/7
C/ERp53/7
Figure 4 Failure of C/ERp53 cells to induce p53-regulated cell death genes. C/ER and C/ERp53 cells were treated with either 0.2, 0.5, 1 or 2 mM 4-OHT and 48 h later cells were lysed and subjected to SDS–PAGE followed by Western blotting using antibodies to the proteins encoded by several p53 target genes. Expression levels were determined by chemiluminescence and quantitated by laser scanning densitometry. Data are plotted relative to actin staining on the same membranes and represent the average values from two independent experiments
experiment, we determined which genes ER/p53 can transactivate. C/ERp53/7 cells were transfected with luciferase reporters linked to the BAX, NOXA, MDM2 and CDKN1A genes and cells treated with 4-OHT. Results (Figure 6a) show that the MDM2 and CDKN1A promoters are markedly induced after p53 activation with 4-OHT, but BAX and NOXA remain at basal Oncogene
To test directly whether YB1 can regulate p53 transactivation, transient transfections were carried out in p53 null human fibroblasts (IIICF/c cells) using the above luciferase reporters genes along with p53 and YB1 expression constructs. The results of this experiment are shown in Figure 7. In Figure 7a, p53 is shown to transactivate all promoter constructs about 10-fold with 10 ng of transfected p53 expression plasmid DNA. When the YB1 expression plasmid was titrated into the transfection mixture, the BAX and NOXA promoters were repressed to basal levels with as little as 0.3 ng of YB1 plasmid (Figure 7b). However, no detectable repression of the CDKN1A promoter occurred until 62.5 ng of YB1 plasmid was added. A similar lack of repression by YB1 was found when the MDM2 promoter was used in the same kind of experiment (data not shown). Thus, these data suggest that YB1 selectively impairs the ability of p53 to transactivate proapoptotic genes and therefore the cell death pathway.
To test whether YB1 must bind p53 to cause repression of the proapoptotic promoters, the experiment in Figure 7b was repeated with the NOXA reporter along with a C-terminal p53 deletion mutant (DCT30) that no longer contains the YB1-binding site (Okamoto et al., 2000). However, this mutant is still transcriptionally competent (see legend to Figure 7c). Results of this experiment show again that overexpression of YB1 causes a dramatic reduction in the ability of p53 to transactivate the NOXA promoter, but DCT30 only causes repression when large amounts of YB1 expression plasmid are transfected. These data are consistent with YB1 repression of p53-dependent transactivation requiring direct binding to p53. BAX and CDKN1A promoters have different affinities for p53 To determine how YB1 selectively impairs p53 transactivation ability, we first examined the phosphorylation
YB1 controls p53 apoptotic function C Homer et al
8319
a
C/ERp53/7
70 60 50 40 30
Cell Number (x 104)
Cell Number (x 104)
C/ER/2
20 10 0
b C/ERp53/7
70 60 50 40
Dead Cells
30
Live Cells
20 10 0
Un C AS Un C AS
Un C AS Un C AS
18 42 Hours post Transfection
18 42 Hours post Transfection 3
AS YB1 N0
N10
0
2
4
6
8
Bax
10
Fold Activation
Time after oligo transfection (h)
Bax/Actin
Noxa/Actin
2 1
N0 N10
0 2 4 6 8 10 Time (h)
N0 N10
0
Noxa
0 2 4 6 8 10 Time (h)
Actin C/ER/2 AS YB1
3
N
0
4
8
Bax
Fold Induction
Time after transfection (h)
Noxa Bax Noxa Actin
Fold Induction
Control Oligo
Noxa/Actin
Bax/Actin -
Noxa /Actin
2 1 0 3
C/ERp53/7
Bax/Actin
2 1 0 N
0 4 8 Time (h)
N
0 4 Time (h)
8
Figure 5 Inhibition of YB1 causes p53-dependent cell death and an increase in the protein levels of p53-regulated cell death genes. C/ER2 and C/ERp53/7 cells were induced with 2 mM 4-OHT and treated with 200 pmol of either antisense oligonucleotides (AS) to YB1 or control oligonucleotides (C). (a) At indicated times after transfection the number of viable cells and nonviable cells was determined by the method of trypan blue exclusion. (b) At indicated times over the 10 h following oligonucleotide transfection, cells were harvested and Western blotting was carried out using antibodies to Bax and to Noxa. Quantitation relative to actin was carried out by laser scanning densitometry as above and again, the average values from two independent experiments are shown. Un ¼ uninduced cells (no 4-OHT); N0, N10 ¼ no oligonucleotide at 0 and 10 h after (mock) transfection
status of p53 before and after treatment of ER/p53 cells with 4-OHT, that is, under conditions when YB1 is in the cytoplasm or in the nucleus. The hypothesis was that the inability of ER/p53 to induce cell death might be due to a failure to phosphorylate on serine 46, which has been implicated in the apoptotic response of p53 (Oda et al., 2000). However, we failed to detect phosphorylation on serine 46, nor indeed on serine 15, associated with cell cycle arrest, although treatment of human fibroblasts with cisplatin did induce such phosphorylations (data not shown).
We next examined the possibility that YB1 selectively prevents the transactivation of proapoptotic promoters because the p53 response elements in these promoters have a lower affinity for p53 than the CDKN1A and MDM2 promoters (Kaeser and Iggo, 2002). If this were true, small amounts of YB1 would preferentially repress the promoters with lower p53-binding affinities, as YB1 would complex with p53 and reduce the amount of p53 available for promoter binding. To address this issue, luciferase reporters were transfected into IIICF/c fibroblasts along with increasing amounts of a plasmid Oncogene
YB1 controls p53 apoptotic function C Homer et al
8320
Fold Induction
a 50
C/ERp53/7
200
IIICF/c
160
CDKN1A MDM2 BAX NOXA
120 80 40
40 Fold induction
a
0
5
NOXA NOXA + Control Oligo NOXA + AS YB1
3 2 1 0
expressing human p53. Results (Figure 7d) show that the CDKN1A promoter is transactivated with about 10 times less p53 plasmid than for the BAX promoter. These data are therefore consistent with the suggestion that the p53 response element in the BAX promoter has a lower binding affinity for p53 than the response element in the CDKN1A promoter and is therefore more likely to be inhibited by YB1. Nuclear YB1 is associated with a failure to induce Bax in mammary epithelial cells Most of our studies suggesting a requirement for p53 in YB1 nuclear translocation and that YB1 may selectively alter p53 function, have been carried out using overexpression approaches. To determine whether our conclusions extend to more physiological conditions we used the early passage mammary epithelial cell line Bre-80 (Huschtscha et al., 1998). These cells are essentially normal with only a defect in expression of the gene encoding the p16INK4A tumor suppressor having been reported, and they undergo cellular senescence at about 33 population doublings (Huschtscha et al., 1998; data not shown). Preliminary analysis of Bre-80 cells showed that a significant proportion contained low levels of nuclear YB1 (see Table 1). Bre-80 cells were treated with UV to activate p53 and about 4 h later cell
CDKN1A
IIICF/c 1 CDKN1A BAX NOXA
0.75 0.5 0.25
p53 + YB1 plasmid
c 1.5 IIICF/c Relative p53 activity
Figure 6 Selective control of p53 transactivation by YB1. (a) C/ERp53/7 cells were transfected with indicated luciferase reporter constructs and after induction of p53 with differing concentrations of 4-OHT (0.2, 0.5, 1 and 2 mM), cells were harvested and luciferase activity determined. (b) C/ERp53/7 cells were transfected with the NOXA reporter and 24 h later they were also transfected with YB1 antisense and control oligonucleotides. Cells were then treated with 1.0 mM 4-OHT to induce a transcriptionally competent p53 and luciferase activity determined at indicated times
NOXA Promoter
0
24 2 6 4 Hours after transfection
Hwtp53
1
0.5
0 p53
d 120 Fold induction
0
BAX
b
C/ERp53/7 4 Fold Induction
20
4-OHT
Relative p53 activity
Un
Oncogene
Mock +10ng p53
10
0
b
30
+ YB1 plasmid
IIICF/c
80
CDKN1A BAX
40
0 log p53 plasmid Figure 7 YB1 represses p53-dependent transactivation of proapoptotic gene promoters. (a) IIICF/c fibroblasts were transfected with indicated promoter-reporter constructs and with an expression plasmid for human p53 to show that each reporter is transactivated by p53. (b) Using the amount of p53 plasmid in (a), differing amounts of YB1 plasmid (from 0.3 to 250 ng) were titrated into the transfection mixture and luciferase activity measured 48 h later. (c). As for (b) but only the NOXA reporter was used and also the p53 mutant DCT30. DCT30 transactivated the NOXA promoter 10-fold in this experiment and wild-type p53 15-fold. (d) Activity of the CDKN1A and BAX promoters was determined 48 h after transfection in the presence of different amounts of the p53 expression plasmid (from 20 pg to 40 ng). All experiments have been carried out at least three times, each in duplicate
YB1 controls p53 apoptotic function C Homer et al
8321
clots were prepared. The clots were fixed, sectioned and stained for p53, YB1, Bax and p21WAF1/CIP1 using specific antibodies and HRP detection. The frequency and intensity of staining was determined qualitatively using light microscopy. The results of this experiment are summarized in Table 1 and examples are shown in Figure 8. A low level of p53 staining (grade 1–2) was observed in untreated cells, which was significantly increased by UV irradiation (grade 3). Some YB1 nuclear staining was observed in the untreated cells (mostly grade 1 with some cells higher). This was again markedly increased after exposure to UV (grade 3 þ ). Modest p21WAF1/CIP1 staining was observed in the untreated cells but staining was significant after stress
(grade 3). By contrast, no Bax staining of Bre-80 cells was observed under any conditions however, it was observed after infection of A549 cells with adenovirus (Ad5) as previously reported (Lomonosova et al., 2002) demonstrating that the conditions used in the experiment do allow Bax detection. Thus, under conditions when YB1 localizes to the nucleus, a selective p53 response is observed similar to that observed in the inducible p53 cell lines and with plasmid transfection experiments. These data suggest that YB1 selectively regulates the apoptotic function of p53 under physiological conditions.
Discussion Table 1
Quantitation of gene expression as determined by immunohistochemistry
Cell type
Treatment
p53
p21
Bax
YB1 nu
YB1 cyto
A549 A549 C/ERp53/7 C/ERp53/7 Bre-80 Bre-80
Control Wild-type Ad5 Control 4-OHT Control UV
1 1–3 3 3 1–2 3
ND ND 1 4 1 3
1 3 0 0 0 0
ND ND 0–1 4 1–4 3+
ND ND 2 1 1 1
ND ¼ not determined
Control
UV
p53
YB1
WAF1/CIP1
p21
Bax
Figure 8 Selective activation of p53 responsiveness in early mammary epithelial cells. The mammary epithelial cell line Bre-80 were treated with UV irradiation (20 J/m2) and about 4 h later cell clots were prepared, fixed and immunostained for p53, YB1, p21WAF1/CIP1 and Bax
In this paper, we have confirmed our previous report that p53 is necessary for efficient nuclear translocation of the Y-box factor YB1. This was carried out using p53 inducible cell lines in which functional activation of p53 occurs after treatment with 4-OHT. Only after p53 induction are substantial amounts of YB1 found in the nucleus (Figures 1 and 2). This result suggests that rather than YB1 being transported to the nucleus by complexing with p53 (Okamoto et al., 2000), p53 must transactivate one of its downstream target genes, the product of which is the ‘effector’ of YB1 nuclear translocation. Further support for this conclusion comes from the observations in Figure 2, where under uninduced conditions p53 is in the nucleus, but YB1 is predominantly in the cytoplasm. Only when p53 becomes transcriptionally competent does YB1 translocate to the nucleus. What the target gene of p53 responsible for nuclear translocation is currently unknown, but a good candidate would be Mdm2 as this is known to undergo nucleocytoplasmic shuttling in a p53dependent manner (Tao and Levine, 1999). Using the p53 inducible cell lines, we next asked what the impact of nuclear YB1 is on p53 functions. We found that induction of p53 resulted in a cell cycle arrest but no increase in cell death (Figure 3), but that inhibiting YB1 expression with antisense oligonucleotides allowed p53 to cause cells to die (Figure 5a). We also observed that the proteins encoded by p53-dependent cell death genes, Bax and Noxa, were not increased after p53 activation with 4-OHT (Figure 4), but both were increased after treatment with antisense YB1 oligonucleotides and this occurred in a p53-dependent manner (Figure 5b). This selective effect of YB1 on p53 activity was also confirmed using promoter-reporter assays. The NOXA and BAX promoters were not induced after p53 activation in the inducible cells (Figure 6a), but they were induced after treatment with antisense YB1 oligonucleotides (Figure 6b). In addition, YB1 was found to inhibit induction of proapoptotic gene promoters by p53 in cotransfection experiments (Figure 7a, b), but had little effect on the ability of p53 to induce the CDKN1A promoter. The mechanism of YB1 selectivity on p53 functions was also investigated. No evidence of differential p53 Oncogene
YB1 controls p53 apoptotic function C Homer et al
8322 Cytoplasm
Nucleus
Cytoplasm
Nucleus
Death Genes (low affinity) Death Genes (low affinity) High affinity genes
Genotoxic stress
High affinity genes (eg CDKN1A, MDM2) YB1
p53
p53
X
Gene X (High affinity)
p53
? Gene X (High affinity) X
X
YB1 Figure 9
X
X
Model of how YB1 shuttles in and out of the nucleus and affects p53 function
phosphorylation was observed in the presence and absence of nuclear YB1, but p53 was found to preferentially transactivate the CDKN1A promoter compared to the BAX promoter (Figure 7d). Thus, given that repression of the proapoptotic promoters requires a direct interaction with p53 (Figure 7c), the simplest explanation of the ability of YB1 to selectively regulate p53 transactivation is that YB1 binds to p53 thereby reducing it’s availability for promoter binding, and promoters with a low-binding affinity for p53 are therefore disabled first. Such an interpretation is consistent with published work comparing p53 response element affinities for different p53-regulated genes (Kaeser and Iggo, 2002) and with all our data. Selective regulation of p53 activity was also observed after UV treatment of early passage human mammary epithelial cells (Bre-80, Figure 8 and Table 1) and this is associated with nuclear YB1. These data suggest that the phenomenon established above using overexpression approaches has physiological relevance, at least in breast tissue. Taking all the data in this paper together a model of how YB1 might be involved in tumorigenesis has been developed. Cells are exposed to stresses of various kinds that lead to p53 being activated/stabilized to become transcriptionally competent. p53 then transactivates several genes, including gene X, the product of which binds YB1 and transports it to the nucleus (Figure 9). YB1 then selectively impairs p53 activity. However, Mdm2 levels can continue to rise, as YB1 does not affect transactivation of MDM2 (Figures 2, 4 and 6). Increased YB1 then causes p53 levels to decline by ubiquitin-dependent proteolysis (Haupt et al., 1997) or p53 activity impaired by direct blocking of the transactivator domain (Lin et al., 1994), which in turn reduces the levels of X, thereby preventing more YB1 entering the nucleus. If this process took place commonly in some normal tissues, they would be chronically impaired for p53 tumor suppressor function and therefore unable to delete (by death) cells harboring Oncogene
YB1
(genetic) abnormalities. An impaired cell death response, combined with a growth advantage conferred by YB1, as recently shown (Berquin et al., 2005; Sutherland et al., 2005), would clearly predispose cells to cancer formation. Preliminary data (not shown) from an analysis of normal breast tissue has shown nuclear YB1 in many ductal cells. This provides some support for the above model, which might help explain why such tissues are cancer prone. Why these tissues have evolved processes that predispose them to cancer is not clear. However, it could be that because the tissues are chronically stressed, there is a trade-off whereby cell death is minimized thereby maintaining tissue integrity, but a partial protective stress response (cell cycle arrest and DNA repair) is still allowed.
Materials and methods Plasmids and construction of p53-inducible cell lines The plasmid pBabepuro-p53ER4OHT (p53ER) was a gift from Dr T Littlewood. The p53ER construct contains human p53 that is fused at its 30 end to the mutated ligand-binding domain of the murine estrogen receptor (a-isoform) described previously (Littlewood et al., 1995). The p53ER insert was cloned into the BamH1 and EcoR1 sites of the multiple cloning site within the retroviral vector pBabe-puro. The pBabe vectors are defective retroviral vectors that maintain the ability to integrate into the host cell genome. Inducible cell lines were prepared using the two vectors p53ER (Morgenstern and Land, 1990) and pBabepuroER4OHT (pER). pER was constructed by removing the p53 insert from p53ER and religating the plasmid. IIICF/C fibroblasts (see below) were transfected with the above constructs and selected using 2 mg/ml puromycin. After 2–3 weeks between 20 and 50 puromycin-resistant colonies were obtained giving a stable transfection frequency of around 1/105. Colonies were expanded and individual clones derived after limiting dilution. Only a small proportion of the original colonies grew sufficiently well to isolate clones. Clones were then screened by PCR for the presence of the appropriate plasmid and by
YB1 controls p53 apoptotic function C Homer et al
8323 immunoblotting to detect p53 protein. Four ER/p53 clones designated C/ERp53/7-10 were finally isolated and two ER only clones designated C/ER/1-2. Cells and cell culture IIICF/c LiFraumeni-derived skin fibroblasts (Rogan et al., 1995) were used for some of the experiments in this paper and were the parent line from which the ER and ER/p53 cell clones were derived. These cells were grown in Dulbecco’s Modified Eagle Medium (DMEM), supplemented with 10% fetal bovine serum (FBS), penicillin/streptomycin (500 U/ml) and L-glutamine (2 mM) but without phenol red (for the ER cell lines). Dominant selection of the ER and ER/p53 constructs were maintained with the addition of 1 mg/ml puromycin. Cells were incubated at 371C in a humidified atmosphere with 10% CO2. A549 lung cancer cells were similarly maintained. Infection with wild-type Ad5 was carried out as previously described (O’Carroll et al., 2000) using one infectious unit of virus per cell for 48 h. Bre-80 mammary epithelial cells were maintained in MCDB.170 (Invitrogen Life Technologies) with 5 mg/ml G418 sulfate and 5 ml of bovine pituitary extract (MCDB100x). Bre-80 cells were incubated at 371C in a humidified atmosphere with 5% CO2. Other plasmids The expression plasmids CMV human wild-type p53 and PCDNA3 rat YB1 used in this paper, along with the control plasmid PCDNA3, have been described previously (Zhang et al., 2003). In addition, CDKN1A-luciferase and a mutant lacking the p53-response element (CDKN1ADp53), MDM2luciferase, NOXA-luciferase and BAX-luciferase were used as reporter plasmids in the transcription assays. Oligonucleotides For transfection, phosphorothioated oligonucleotides were synthesized as follows: antisense YB1 oligonucleotides were used as a combination of 2; 50 -CTGCACAGGAGGGTTG GAATAC-30 , and 50 -GGAATCGTGGTCTATATCCCCG-30 , negative control phosphorothioated oligonucleotides were synthesized as follows; 50 -GCGGATAACAATTTCACACA GG-30 . Details of preparation and the specificity of the antisense oligonucleotides have been previously reported (Lasham et al., 2003). Cell growth analysis Cells were seeded into six-well plates at a density of 2 or 3 104 cells per well. At indicated times after treatment, cells were harvested by trypsin treatment and cell number determined by direct cell counting on a hemocytometer using the method of trypan blue exclusion. Both viable and nonviable cells were counted. All data points were determined from the average of three independent wells and repeated at least twice. Immunocytochemistry Cells were seeded at 2 104 cells per well (24-well plate). They were then induced and grown on coverslips for 48 h. Coverslips were fixed in ice-cold methanol for 10 min and either kept at 201C until required or washed in PBS for 10 min for further processing. Coverslips were blocked (2% normal goat serum, 5% BSA and 0.2% Triton in PBS), incubated with primary antibodies, washed five times in PBS and incubated with secondary antibody (1 : 400). All incubations were on an
orbital shaker for 1 h and were at room temperature. Coverslips were then washed five times in PBS and mounted with 80% glycerol þ 20% PBS. The primary antibodies used were DO-1 (1 : 1000, Santa Cruz Biotechnology) and pAb1620 (1 : 500, Oncogene Science) for p53; SMP-14 for Mdm2 (1 : 1000, Santa Cruz Biotechnology); F-5 for p21WAF1/CIP1 (1 : 1000, Santa Cruz Biotechnology); and a rabbit polyclonal antibody against YB1 prepared in-house (1 : 1000). Immunofluorescence was then carried out using Alexa Fluors 488 (green) conjugated goat anti-mouse or rabbit secondary antibodies (Molecular Probes) using an Olympus Model BX50 fluorescence microscope and a Spot RT digital camera and software (Diagnostic Instruments). All antibodies were diluted in 1% normal goat serum in PBS. Immunohistochemistry Bre-80 cells with and without UV treatment (20 J/m2) were harvested, centrifuged and the cell pellet washed with PBS multiple times. An equal volume of plasma and thromboplastin (Innovin, Dade Behring) was added, and incubated for 20 min. The clot was fixed in 10% normal-buffered formalin for 4 h, then mounted in paraffin blocks. For immunostaining 7 mm sections were cut onto APES slides, then deparaffinized and endogenous peroxidase activity quenched in methanol and H2O2. Antigen retrieval wash was performed by microwave with citrate buffer (pH 6.0) for 5 min. This was repeated twice with sequentially less power. Blocking and all subsequent steps were performed using the VECTORSTAIN ABC kit (Vector Labortories Inc.) according to manufacturer’s instructions. Immunostaining was developed using the DAB chromagen and counterstained with Gills #1 hematoxylin. Sections were coverslipped with DPX. The primary antibodies used were DO-1 (1 : 500, Santa Cruz Biotechnology) for p53; F-5 for p21WAF1/CIP1 (1 : 100, Santa Cruz Biotechnology); Bax antibody (1 : 200, BD Pharmigen) and a rabbit polyclonal antibody against YB1 (1 : 500). Immunoblotting In general, about 106 cells were seeded and left overnight. Cells were induced with 4-OHT and incubated for indicated times. Protein lysates were prepared (lysis buffer; 10 mM Tris-HCl pH 8.0; 140 mM NaCl; 5 mM DTT; 1 mM EDTA; 1% SDS; 1% NP40). The protein content of each sample was determined by spectrophotometry (280 nm) and standardized accordingly. Samples were then mixed with an equal volume of DSLloading buffer (50 mM Tris-HCl pH 6.8; 200 mM DTT; 4% SDS, 40% glycerol; 0.08% bromophenol blue), boiled for 5 min then loaded on a 10–12% SDS–PAGE. Following electrophoresis, protein extracts were transferred to PVDF membrane (Amersham). Detection was carried out according to standard procedures and bands visualized using the WesternBreeze chemiluminescent system (Invitrogen). Immunoblotting experiments were carried out with the following antibodies: p53 ¼ DO-1 (1 : 500, Santa Cruz Biotechnology); YB1 ¼ A16 (1 : 500, Santa Cruz Biotechnology); Mdm2 ¼ SMP-14 (1 : 500, Santa Cruz Biotechnology); p21WAF1/CIP1 ¼ F-5 (1 : 500, Santa Cruz Biotechnology); Gadd45 ¼ C-4 (1 : 500, Santa Cruz Biotechnology); 14-33s ¼ N-14 (1 : 500, Santa Cruz Biotechnology); Bax ¼ (1 : 200, BD Pharmagen); and Noxa ¼ N-15 (1 : 500, Santa Cruz Biotechnology); Actin ¼ C11 (1 : 500 Santa Cruz Biotechnology) was used as a loading control, except for Figure 1 where equal loading was confirmed using Ponceau S staining of the membrane. Oncogene
YB1 controls p53 apoptotic function C Homer et al
8324 Promoter-reporter assays Six-well plates were seeded with 2 10 cells/well and incubated overnight. Cells were transfected according to the manufacturer’s instructions (FuGENE 6 Transfection Reagent, Roche) with indicated amounts of plasmid DNA. Luciferase promoter reporter constructs were transfected (1 mg per well) and incubated for 3–4 h before induction of p53 with 4-OHT for experiments involving ER cells. Cells were harvested at indicated times, cell number established and cell pellets frozen. Cells were lysed and luciferase activity determined according to the manufacturer’s instructions (Luciferase Assay System, Promega). Results were standardized according to cell number and all experiments were carried out in duplicate. 5
Nuclear extracts Nuclear extracts were prepared according to previously described methods (Koike et al., 1997). Briefly, cells were
resuspended in 200 ml of buffer A (10 mM HEPES pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM PMSF, 0.5 mM DTT) for 40 min, then lysed by passing 10 times through a 25-gauge needle attached to a 1-ml syringe. The nuclear pellets were centrifuged for 6 min at 4300 g, lysed by high salt, centrifuged and dialysed against buffer D (20 mM HEPES pH 7.9, 20% (v/v) glycerol, 100 mM KCl, 0.2 mM EDTA, 0.5 mM PMSF, 0.5 mM DTT) overnight. Extracts were centrifuged for 15 min (10 000 g) and stored at 801C. The cytoplasmic fraction was adjusted with buffer D.
Acknowledgements This work was funded by the Cancer Society of New Zealand (Grant 99/11), the Health Research Council of New Zealand (Grant 04/234), and the New Zealand Lottery Board (Grant AP76795). We also thank Moshe Oren (Israel) and Karen Vousden (Scotland) for promoter-reporter constructs and Lily Huschtscha (Australia) for Bre-80 cells.
References Asakuno K, Kohno K, Uchiumi T, Kubo T, Sato S, Isono M and Kuwano M. (1994). Biochem. Biophys. Res. Commun., 199, 1428–1435. Bargou RC, Jurchott K, Wagener C, Bergmann S, Metzner S, Bommert K, Mapara MY, Winzer KJ, Dietel M, Dorken B and Royer HD. (1997). Nat. Med., 3, 447–450. Berquin IM, Pang B, Dziubinski ML, Scott LM, Chen YQ, Nolan GP and Ethier SP. (2005). Oncogene, 24, 3177–3186. Chen CY, Gherzi R, Andersen JS, Gaietta G, Jurchott K, Royer HD, Mann M and Karin M. (2000). Genes Dev., 14, 1236–1248. Evdokimova V, Ruzanov P, Imataka H, Raught B, Svitkin Y, Ovchinnikov LP and Sonenberg N. (2001). EMBO J., 20, 5491–5502. Haupt Y, Maya R, Kazaz A and Oren M. (1997). Nature, 387, 296–299. Holm PS, Bergmann S, Jurchott K, Lage H, Brand K, Ladhoff A, Mantwill K, Curiel DT, Dobbelstein M, Dietel M, Gansbacher B and Royer HD. (2002). J. Biol. Chem., 277, 10427–10434. Huschtscha LI, Noble JR, Neumann AA, Moy EL, Barry P, Melki JR, Clark SJ and Reddel RR. (1998). Cancer Res., 58, 3508–3512. Jurchott K, Bergmann S, Stein U, Walther W, Janz M, Manni I, Piaggio G, Fietze E, Dietel M and Royer HD. (2003). J. Biol. Chem., 278, 27988–27996. Kaeser MD and Iggo RD. (2002). Proc. Natl. Acad. Sci. USA, 99, 95–100. Kamura T, Yahata H, Amada S, Ogawa S, Sonoda T, Kobayashi H, Mitsumoto M, Kohno K, Kuwano M and Nakano H. (1999). Cancer, 85, 2450–2454. Kohno K, Izumi H, Uchiumi T, Ashizuka M and Kuwano M. (2003). Bioessays, 25, 691–698. Koike K, Uchiumi T, Ohga T, Toh S, Wada M, Kohno K and Kuwano M. (1997). FEBS Lett., 417, 390–394. Lasham A, Lindridge E, Rudert F, Onrust R and Watson J. (2000). Gene, 252, 1–13. Lasham A, Moloney S, Hale T, Homer C, Zhang YF, Murison JG, Braithwaite AW and Watson J. (2003). J. Biol. Chem., 278, 35516–35523. Levenson V, Davidovich I and Roninson I. (2000). Cancer Res., 60, 5027–5030. Lin J, Chen J, Elenbaas B and Levine AJ. (1994). Genes Dev., 8, 1235–1246. Oncogene
Littlewood TD, Hancock DC, Danielian PS, Parker MG and Evan GI. (1995). Nucleic Acids Res., 23, 1686–1690. Lomonosova E, Subramanian T and Chinnadurai G. (2002). J. Virol., 76, 11283–11290. Mantovani R. (1998). Nucleic Acids Res., 26, 1135–1143. Matsumoto K and Wolffe AP. (1998). Trends Cell. Biol., 8, 318–323. Mertens PR, Alfonso-Jaume MA, Steinmann K and Lovett DH. (1999). J. Am. Soc. Nephrol., 10, 2480–2487. Mertens PR, Harendza S, Pollock AS and Lovett DH. (1997). J. Biol. Chem., 272, 22905–22912. Morgenstern JP and Land H. (1990). Nucleic Acids Res., 18, 3587–3596. O’Carroll SJ, Hall AR, Myers CJ, Braithwaite AW and Dix BR. (2000). Biotechniques, 28, 408–410. Oda K, Arakawa H, Tanaka T, Matsuda K, Tanikawa C, Mori T, Nishimori H, Tamai K, Tokino T, Nakamura Y and Taya Y. (2000). Cell, 102, 849–862. Oda Y, Ohishi Y, Saito T, Hinoshita E, Uchiumi T, Kinukawa N, Iwamoto Y, Kohno K, Kuwano M and Tsuneyoshi M. (2003). J. Pathol., 199, 251–258. Oda Y, Sakamoto A, Shinohara N, Ohga T, Uchiumi T, Kohno K, Tsuneyoshi M, Kuwano M and Iwamoto Y. (1998). Clin. Cancer Res., 4, 2273–2277. Ohga T, Koike K, Ono M, Makino Y, Itagaki Y, Tanimoto M, Kuwano M and Kohno K. (1996). Cancer Res., 56, 4224–4228. Ohga T, Uchiumi T, Makino Y, Koike K, Wada M, Kuwano M and Kohno K. (1998). J. Biol. Chem., 273, 5997–6000. Okamoto T, Izumi H, Imamura T, Takano H, Ise T, Uchiumi T, Kuwano M and Kohno K. (2000). Oncogene, 19, 6194–6202. Prives C and Hall PA. (1999). J. Pathol., 187, 112–126. Raffetseder U, Frye B, Rauen T, Jurchott K, Royer HD, Jansen PL and Mertens PR. (2003). J. Biol. Chem., 278, 18241–18248. Rogan EM, Bryan TM, Hukku B, Maclean K, Chang AC, Moy EL, Englezou A, Warneford SG, Dalla-Pozza L and Reddel RR. (1995). Mol. Cell. Biol., 15, 4745–4753. Sakura H, Maekawa T, Imamoto F, Yasuda K and Ishii S. (1988). Gene, 73, 499–507. Shibahara K, Sugio K, Osaki T, Uchiumi T, Maehara Y, Kohno K, Yasumoto K, Sugimachi K and Kuwano M. (2001). Clin. Cancer Res., 7, 3151–3155.
YB1 controls p53 apoptotic function C Homer et al
8325 Shibao K, Takano H, Nakayama Y, Okazaki K, Nagata N, Izumi H, Uchiumi T, Kuwano M, Kohno K and Itoh H. (1999). Int. J. Cancer, 83, 732–737. Sommerville J. (1999). Bioessays, 21, 319–325. Stenina OI, Shaneyfelt KM and DiCorleto PE. (2001). Proc. Natl. Acad. Sci. USA, 98, 7277–7282. Stickeler E, Fraser SD, Honig A, Chen AL, Berget SM and Cooper TA. (2001). EMBO J., 20, 3821–3830. Sutherland BW, Kucab J, Wu J, Lee C, Cheang MC, Yorida E, Turbin D, Dedhar S, Nelson C, Pollak M, Leighton Grimes H, Miller K, Badve S, Huntsman D, Blake-Gilks C, Chen M, Pallen CJ and Dunn SE. (2005) advance online publication, 4 April 2005 (doi:10.1038/sj.onc.1208590).
Tao W and Levine AJ. (1999). Proc. Natl. Acad. Sci. USA, 96, 3077–3080. Uchiumi T, Kohno K, Tanimura H, Matsuo K, Sato S, Uchida Y and Kuwano M. (1993). Cell Growth Differ., 4, 147–157. Wolffe AP. (1994). Bioessays, 16, 245–251. Zhang YF, Homer C, Edwards SJ, Hananeia L, Lasham A, Royds J, Sheard P and Braithwaite AW. (2003). Oncogene, 22, 2782–2794. Zhou BP, Liao Y, Xia W, Spohn B, Lee MH and Hung MC. (2001a). Nat. Cell. Biol., 3, 245–252. Zhou BP, Liao Y, Xia W, Zou Y, Spohn B and Hung MC. (2001b). Nat. Cell. Biol., 3, 973–982.
Oncogene
