A role for Ubc9 in tumorigenesis - Nature

Oncogene (2005) 24, 2677–2683

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

A role for Ubc9 in tumorigenesis Yin-Yuan Mo*,1, Yanni Yu1, Elena Theodosiou2, PL Rachel Ee1 and William T Beck1,3,4 1

Department of Biopharmaceutical Sciences, University of Illinois at Chicago, Chicago, IL 60612, USA; 2Department of Medicine, Section of Hematology-Oncology, University of Illinois at Chicago, Chicago, IL 60612, USA; 3Department of Biochemistry and Molecular Genetics, University of Illinois at Chicago, Chicago, IL 60612, USA; 4Gynecologic Oncology Group Core Laboratory in Molecular Pharmacology, University of Illinois at Chicago, Chicago, IL 60612, USA

The post-translational modifications ubiquitination and sumoylation have been implicated in regulating many critical cellular pathways. Like ubiquitination, sumoylation is a multistep process involving maturation, activation, conjugation and deconjugation. Ubc9 is a sole E2-conjugating enzyme essential for sumoylation. We have previously shown that alterations of Ubc9 expression affect tumor drug responsiveness. However, it is not clear whether there is any link between sumoylation and tumorigenesis, even though alterations of the ubiquitination pathway can lead to the development of cancer. In this study, we found that Ubc9 expression levels were elevated in ovarian tumors compared to the matched normal ovarian specimens, suggesting that Ubc9 may play a role in tumorigenesis. To test this, we overexpressed a dominant-negative mutant of Ubc9 (Ubc9-DN) and wildtype Ubc9 (Ubc9-WT) in the MCF-7 human breast tumor cells. Inoculating these cells as xenografts in mice revealed that tumors expressing Ubc9-WT grew better than the vector control, while tumors expressing Ubc9-DN exhibited reduced growth. This pattern was also seen in these cells when grown in culture. To better understand the mechanism behind this observation, we profiled gene expressions in these cells by microarray analysis and found alterations in expression of the pro-oncogene bcl-2 in these Ubc9-DN- and Ubc9-WT-expressing cells. Consistent with the bcl-2 results, subsequent studies revealed a higher rate of apoptosis and poor survival for the MCF-7 cells expressing Ubc9-DN, which are associated with downregulation of bcl-2. Together, these results suggest a role for Ubc9 in tumorigenesis at least partially through regulation of bcl-2 expression. Oncogene (2005) 24, 2677–2683. doi:10.1038/sj.onc.1208210 Published online 21 February 2005 Keywords: Ubc9; sumoylation; SUMO; Bcl-2; apoptosis

*Correspondence: Current address: Y-Y Mo, Department of Medical Microbiology and Immunology, Southern Illinois University School of Medicine, 911 N. Rutledge, Room #1300A, Springfield, IL 62794, USA; E-mail: [email protected] Received 23 June 2004; revised 31 August 2004; accepted 14 September 2004; published online 21 February 2005

Introduction Post-translational modifications play an important role in in vivo functions of proteins through the regulation of protein activity, turnover and localization and/or interactions. These modifications include phosphorylation, acetylation, methylation, glycosylation and ubiquitination. Recently, another type of protein modification has been identified, that is, small ubiquitin (Ub)-related modifier (SUMO) conjugation or sumoylation (Muller et al., 2001). SUMO-1, also known as UBL1, Sentrin, GMP1 and Smt3, is a member of the Ub and Ub-like superfamily. Like ubiquitination, sumoylation modulates protein function through post-translational covalent attachment to lysine residues within targeted proteins. However, the biochemical consequences of these two pathways are different. While ubiquitination leads to protein degradation through the 26S proteosome (Hershko and Ciechanover, 1998), sumoylation does not cause protein degradation; instead, it has been implicated in other cellular processes such as regulating activity of transcription factors, mediating nuclear translocation of proteins or formation of subnuclear structures, and enhancing protein stability (Muller et al., 2001; Seeler and Dejean, 2003). The sumoylation cycle is a multistep process, involving maturation, activation, conjugation and deconjugation (Muller et al., 2001). Ubc9 is an E2 enzyme that transfers the activated SUMO to the target protein (Johnson and Blobel, 1997) and it is known to be required for conjugation of SUMOs, including SUMO-1 and -2/3 (Saitoh and Hinchey, 2000). Although E3-like ligases have been identified recently in both yeast and humans (Kahyo et al., 2001; Pichler et al., 2002), they are apparently not required for in vitro sumoylation (Desterro et al., 1999). Thus, the major function of these enzymes may be to regulate Ubc9-mediated sumoylation or increase the affinity between Ubc9 and its substrates in vivo. While several E2 enzymes have been reported in the ubiquitination pathway (Pickart, 2001), Ubc9 is the sole E2 enzyme known to be required for sumoylation. In yeast and higher eukaryotic cells, gene disruption of Ubc9 is lethal (Seufert et al., 1995; Hayashi et al., 2002). This is presumably due to its involvement in regulating several critical pathways. Ubc9 is conserved from yeast to humans and is expressed ubiquitously, but the

Role for Ubc9 in tumorigenesis Y-Y Mo et al

2678

a

N

T

N

T

N

T

UBC9 β2-M 1

2

3

4

5

6

7

8

9

10

UBC9 β2-M UBC9 β2-M UBC9 β2-M

b Relative Ubc9 expression

expression level differs among various tissues (Kovalenko et al., 1996). A mutation where cysteine 93 is replaced by serine (C93S) exhibits a dominant inhibitory effect on the endogenous Ubc9 (Giorgino et al., 2000). Thus, the dominant-negative form of Ubc9 (Ubc9-DN) provides a useful tool to study the cellular function(s) of Ubc9. As an essential E2-conjugating enzyme for sumoylation, Ubc9 is believed to play a central role in sumoylation-mediated cellular pathways. In support of this notion, we have recently shown that overexpression of Ubc9-DN increased tumor drug sensitivity (Mo et al., 2004). However, although much has been learned about the biology and biochemistry of Ubc9, a link between sumoylation, specifically Ubc9, and tumorigenesis is unclear. In the present study, we found that in human ovarian tumors, ubc9 is upregulated compared to matched normal ovarian specimens. This prompted us to investigate whether alterations of Ubc9 have any effect on tumor growth in the mouse xenograft carcinoma model. By using human breast tumor MCF-7 cells overexpressing Ubc9-DN or wild-type Ubc9 (Ubc9-WT), we found that Ubc9-DN reduces the tumor growth, whereas Ubc9-WT enhances the tumor growth in the mouse model. Furthermore, this Ubc9-associated alteration in tumor growth correlates with altered expression of the pro-oncogene bcl-2.

3.0 2.5

P < 0.01

2.0 1.5 1.0 0.5 0.0

Results ubc9 mRNA is upregulated in ovarian tumors We previously found that Ubc9-DN, which decreases protein sumoylation (Mo et al., 2002), increased tumor drug sensitivity (Mo et al., 2004); moreover, over dozens of proteins involved in critical cellular pathways such as tumor suppressors, oncoproteins and nuclear receptors have been identified as substrates for sumoylation (Gostissa et al., 1999; Poukka et al., 1999; Rodriguez et al., 1999; Muller et al., 2001). Therefore, we attempted to determine whether Ubc9 plays any role in tumorigenesis. As a first step toward this direction, we examined the expression level of ubc9 in ovarian tumor specimens and matched normal ovarian tissue obtained from the Gynecologic Oncology tumor bank. Semiquantitative reverse transcription–polymerase chain reaction (RT– PCR) analyses indicated that overall, the ubc9 mRNA level was higher in ovarian tumor specimens than in the matched normal ovarian epithelium specimens from the same patients (Figure 1a). After normalization to b-2microglobulin (b-2-M), the average value of the relative ubc9 mRNA level in 10 tumor specimens was significantly higher (Po0.01) than that of the matched normal ovarian tissue (Figure 1b), suggesting that aberrant ubc9 expression is associated with human malignancy. Ubc9-DN reduces the tumor growth, while Ubc9-WT enhances the tumor growth in nude mice To further determine the role for Ubc9 in tumorigenesis, we suppressed the Ubc9 function in MCF-7 cells by Oncogene

N

T

Figure 1 Detection of ubc9 expression in ovarian cancer specimens by RT–PCR. (a) RT–PCR products from 10 pairs of matched normal (N) and tumor (T) tissues. (b) Relative expression of ubc9 after normalization to b-2-M

overexpression of Ubc9-DN (Giorgino et al., 2000; Mo et al., 2004). We also overexpressed Ubc9-WT in MCF7 cells. From over 20 positive clones for each construct, we chose three Ubc9-DN clones (#7, #8 and #13) and three Ubc9-WT clones (#112, #114 and #125). The expression of the exogenous Ubc9-DN or Ubc9-WT in these clones was confirmed by Western blot (Figure 2a). As a control, we pooled over 20 individual MCF-7 clones transfected with the vector alone. When inoculated into nude mice, the Ubc9-DN cell lines grew tumors much smaller in size than the vector controls. By contrast, all the Ubc9-WT cell lines formed tumors larger in size than the vector controls. Extended studies with Ubc9-DN-7 and Ubc9-WT-112 along with the vector control indicated that tumors from Ubc9-WT cells grew more rapidly than the vector controls, whereas tumors from Ubc9-DN cells grew more slowly than the vector controls (Figure 2b). After tumors were removed from the mice 25 days after injection, the tumors were individually weighed. While tumors originating from the Ubc9-DN cells weighed about 32% of those of the vector controls (or 68% reduction), tumors originating from the Ubc9-WT cells weighed 150% of those of the vector controls, indicating that alterations of Ubc9 affect tumor growth. Furthermore, when the


2679

a

DN clone # V

7

8

a

WT clone # 13

112

114

Vector

DN

WT

125

α-Ubc9 α-Myc

b Vector Ubc9-DN Ubc9-WT

1000 800

b

10000

600 400 200 0 10

c

12

14

16 18 20 22 Day after inculation

24

26

0.6

Tumor Weight (g)

0.5

Cell Number (x 1000)

Tumor volume (mm3)

1200

Vector Ubc9-DN-7 Ubc9-WT-112

1000

1

2

3

4

5

6

Day

V vs DN: P < 0.01 WT vs DN: P < 0.05

Figure 3 Effect of Ubc9 on cell morphology (a) and growth (b). (a) Cells were seeded at about 3 million cells/100 mm dish and pictures were taken after cells were grown for 16 h. Upper panel: phase-contrast images; lower panel: H&E-stained cells. (b) Cells were seeded at B300 000 cells/100 mm dish and were counted at the indicated time points. The values were averages of three separate experiments with standard errors

0.4 0.3 0.2 0.1 0.0 Vector

DN

WT

Figure 2 Effect of Ubc9 on tumor growth in nude mice. (a) Expression of Ubc9-DN and Ubc9-WT in MCF-7 cells as detected by Western blot. Total protein was extracted from the vector control, Ubc9-DN and Ubc9-WT cells; the membrane was probed with anti-Ubc9 antibody and then probed with anti-Myc antibody after stripping. The Myc antibody recognized both Ubc9-DN and Ubc9-WT. The lower bands detected by Ubc9 antibody correspond to the endogenous Ubc9, while the upper bands correspond to the exogenous Myc-tagged proteins (Ubc9-DN or Ubc9-WT). (b) Tumor growth curve. Tumors derived from the vector control, Ubc9-DN and Ubc9-WT MCF-7 cells in nude mice were measured using a caliper and the tumor volume was calculated using the formula volume ¼ D d2 p/6 (Zhang et al., 2002), where D is the longer diameter and d is the shorter diameter. (c) Average tumor weight for each group with s.d. (n ¼ 8). V, vector; DN, Ubc9-DN; WT, Ubc9-WT

tumors derived from the Ubc9-DN were allowed to grow for an extended period of time (up to 45 days), they never reached the size of those derived from either the vector control or Ubc9-WT, suggesting that Ubc9DN does not only delay, but also suppress the tumor progression.

bcl-2 is downregulated in the Ubc9-DN cells To determine the underlying mechanism of this Ubc9associated alteration in tumor growth, we examined the effect of alterations of Ubc9 on cell morphology, cell growth and cell cycle progression. Morphologically, MCF-7 cells expressing vector alone behaved no differently from the parental controls, that is, they appeared as typical epithelial cells and tended to form cohesive patches of cells (Figure 3a). By contrast, the Ubc9-DN cells tended to form loose cell–cell contacts, while the Ubc9-WT cells appeared to form larger clusters than the vector controls. In terms of cell growth, the Ubc9-WT cells grew faster than the Ubc9-DN cells or the vector controls, their doubling time being about 19 h. By contrast, the Ubc9-DN cells grew slightly more slowly than the vector controls, although the error bars overlapped with those of the vector controls at most time points (Figure 3b). The doubling times for the vector controls and the Ubc9-DN cells were about 23 and 24 h, respectively. Flow cytometric analysis showed similar cell cycle profiles for all these cells (not shown). To better understand the molecular mechanism of Ubc9-associated alterations in tumor growth, we Oncogene


2680

profiled gene expression of the Ubc9-DN and Ubc9-WT cells by microarray analysis. Among about 22 000 gene sequences on the chips, over 300 genes exhibited over a twofold up- or downregulation when comparing MCF-7 cells expressing Ubc9-DN and the vector control or comparing Ubc9-WT and the vector control. While experiments are still under way to characterize systematically the differences in gene expression in these cells, we found, intriguingly, that expression of the antiapoptotic gene bcl-2 was decreased 4.7-fold in the Ubc9-DN cells compared to the vector controls and decreased 5.3fold compared to the Ubc9-WT cells. Semiquantitative RT–PCR confirmed the microarray data (Figure 4a). Furthermore, we found that the level of Bcl-2 protein was also reduced in the Ubc9-DN cells, as detected by Western blot (Figure 4b). Densitometric analysis of three separate experiments indicated that the level of Bcl-2 was remarkably lower in the Ubc9-DN cells (4.4-fold) than in the Ubc9-WT cells, and also lower than in the vector controls (3.8-fold), suggesting a good correlation between Ubc9-DN and downregulation of Bcl-2. We also analysed bcl-2 expression in tumors derived from Ubc9-DN, Ubc9-WT and vector control MCF-7 cells in nude mice and found a similar downregulation of bcl-2

Relative expression of bcl-2

a 6 5 4 3 2 1 0

3 2 1 0 V

DN

WT

DN

Cell culture

Tumor

Relative level of Bcl-2

b 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 DN

WT

V

DN

WT Bcl-2 β-actin

Cell culture

Tumor

Figure 4 Expression of the bcl-2 gene in the vector control (V), Ubc9-DN-7 (DN) and Ubc9-WT-112 (WT) cell cultures or tumors derived from these cells as indicated. (a) mRNA levels of bcl-2 as detected by RT–PCR. (b) Protein levels of Bcl-2 as detected by Western blot. Values in (a) and (b) are averages of three separate experiments7s.d. Oncogene

We reasoned that because bcl-2 is a well-known antiapoptotic gene that has been shown to play a key role in regulating cell apoptosis and tumorigenesis, alteration of bcl-2 expression should have a direct impact on cell apoptosis. Indeed, we found that more apoptotic cells were present in the Ubc9-DN-expressing MCF-7 cell population than either the vector control or Ubc9-WT cell population (Figure 5a and b), which may in part explain why these cells are more sensitive to anticancer drugs (Mo et al., 2004). In addition, we found a more obvious effect of Ubc9 on cell survival (Figure 5c). Although the Ubc9-DN cells grew only slightly more slowly (Figure 3b), the survival rate of the Ubc9-DN cells was significantly lower than that of the vector control cells (Figure 5c). By contrast, the Ubc9WT cells exhibited a higher survival rate than the vector controls or Ubc9-DN cells. These results suggest that downregulation of bcl-2 could be a major contributing factor for the high rate of cell death and poor survival in the Ubc9-DN cells.

WT

-actin

V

Reduced expression of bcl-2 in the Ubc9-DN cells is associated with a high rate of cell death and poor survival

Discussion V

bcl-2

2.5 2.0 1.5 1.0 0.5 0.0

mRNA and Bcl-2 protein (Figure 4a and b) by Ubc9DN. In contrast to bcl-2, other members of the bcl-2 gene family, mcl-1, bcl-x and bax, revealed only slight or no change in their mRNA and protein levels (not shown). Thus, it appears that bcl-2 is affected by Ubc9DN to the greatest extent among members of the bcl-2 gene family.

It is now well established that deregulation of ubiquitination pathways can lead to the development of cancer through the mechanisms that control the stability of important proteins such as those involved in regulating transcription and growth factors (Lipkowitz, 2003). In addition, some E3 ubiquitination ligases themselves, such as MDM2, are oncoproteins (Honda et al., 1997). By contrast, there is little information available on the role of protein sumoylation, specifically Ubc9, in cancer development. However, given that SUMO/Ubc9 affects the activity and the subcellular localization of target proteins, including tumor suppressors, oncoproteins and cell cycle-related proteins, one would expect that alterations of SUMO/Ubc9 will directly or indirectly affect the pathways linked to tumorigenesis. Our initial findings of the elevated expression of ubc9 mRNA in ovarian tumors prompted us to carry out this study. Since Ubc9 is the sole known E2 enzyme for conjugating SUMOs, including SUMO-1, -2 and -3, alterations of ubc9 expression will directly affect sumoylation. For instance, expression of Ubc9-DN was shown to have a dominant inhibitory effect on endogenous Ubc9, thus suppressing protein sumoylation (Giorgino et al., 2000; Mo et al., 2002). To determine the effect of Ubc9 on tumor growth, we established stable transfectants of Ubc9-DN and


2681

a

b

Apoptotic Cells

20000

15000

10000

5000

0 DN

V

Relative Cell Number

c

WT

10 MCF7-vector MCF7-Ubc9-WT-112 MCF7-Ubc9-DN-7

1

0.1

0.01 *

* *

0.001 5

10

15 Day

20

25

30

Figure 5 Effect of Ubc-9 on cell apoptosis and survival. (a) Detection of apoptosis. Cells were seeded at 40% confluence in a 100 mm culture dish and grown for 20 h. The cells in the suspension were harvested by centrifugation, fixed and stained with Hoechst dye. The attached cells were also harvested and counted. Apoptotic cells showing typical nuclear bladders were counted, as shown in the inset, and were calculated per 1 million total cells (both suspension cells and attached cells). (b) DNA laddering assays also suggest that the Ubc9-DN cells collected from the suspension are under apoptosis. (c) Cell survival curve. Cells were seeded and counted as described in Materials and methods. *No surviving cells were recovered after that point. The values are averages of three independent experiments with s.d. for both (a) and (c)

Ubc9-WT in human breast cancer MCF-7 cells. The finding in our study supports the notion that alterations of SUMO/Ubc9 could ultimately impact tumorigenesis and thus, proteins involved in the sumoylation pathway could be potential targets for anticancer therapy. Although the mechanisms of Ubc9-associated alteration in tumor growth remain to be determined, the reduction of bcl-2 expression in the Ubc9-DN cells observed in this study offers a plausible explanation. Programmed cell death or apoptosis is critical for maintaining tissue homeostasis, and impaired apoptosis is believed to be a key step in tumorigenesis. It is well known that apoptosis is a consequence of the activation of caspases, a family of proteases that cause a rapid selfdisassembly of cells. The initiation of activation of caspases is tightly regulated by a variety of factors, among which Bcl-2 family proteins play a pivotal role.

There are over a dozen Bcl-2-related proteins in mammalian cells, consisting of proapoptotic and antiapoptotic members (Cory et al., 2003). As an antiapoptotic protein, Bcl-2 has been shown to contribute to neoplastic cell expansion, as well as to resistance to radiation therapy and chemotherapy, by blocking apoptosis (Cory et al., 2003); moreover, aberrant expression of Bcl-2 has been found in a variety of tumors (Alderson et al., 1995; Catz and Johnson, 2003; Cory et al., 2003). An important feature of Bcl-2 family proteins is the ability to form homo- or heterodimers. For instance, the Bcl-2 heterodimer with the proapoptotic protein Bax (Bcl-2/Bax) blocks cell apoptosis (Sato et al., 1994). Thus, the ratio of antiapoptotic proteins and proapoptotic proteins (e.g. Bcl-2/Bax) will largely determine whether a cell should live or die. As the Ubc9DN cells express a reduced level of Bcl-2 with no change in Bax levels, one would expect that the Ubc9-DN cells will have a high cell death rate, and consequently, they will probably be less tumorigenic, as we have shown here. Expression of the bcl-2 gene can be regulated at both transcriptional and translational levels. Since our microarray data, along with RT–PCR, indicate that the bcl-2 mRNA is downregulated in the Ubc9-DN cells, it is likely that Ubc9 affects the transcription of bcl-2. Many factors have been implicated in controlling expression of bcl-2. Among them, nuclear receptors such as estrogen receptor (ER) are the most prominent candidates, because mounting evidence from clinical specimens as well as cell lines has revealed a good correlation between ER and Bcl-2 (Leek et al., 1994; Teixeira et al., 1995; Zhang et al., 1998; MartinezArribas et al., 2003). Moreover, several putative ERresponsive elements (ERE) have been found in the bcl-2 promoter (Dong et al., 1999; Perillo et al., 2000). Although there is no evidence that ER is sumoylated, recent studies indicate that steroid receptor coactivator family (SRC) such as SRC-1 can be sumoylated. Sumoylation of SRC stimulates expression of the downstream genes. For instance, sumoylation of SRC1 increases the interactions of progesterone receptor (PR) with SRC-1 and thus, prolongs SRC-1 retention in the nucleus (Chauchereau et al., 2003). As SRC interacts with a variety of nuclear receptors and regulates signaling pathways mediated by the nuclear receptors (Leo and Chen, 2000), induction of SRC activity by sumoylation is believed to have a profound effect on expression of nuclear receptor-mediated genes, including those ER-mediated genes such as bcl-2. Since MCF-7 cells express ER protein, suppression of sumoylation by Ubc9-DN should affect ER-mediated gene expression. However, downregulation of bcl-2 alone cannot account for the observed reduction of tumor growth associated with Ubc9-DN because more than 60 proteins are subjected to sumoylation (Seeler and Dejean, 2003). Overexpression of Ubc9-DN or Ubc9WT will also affect these proteins in terms of their activity, subcellular localization or expression of their downstream targets. Therefore, Bcl-2 may be just one component in this complex interconnected system. A great challenge is to identify which protein(s) plays a Oncogene


2682

central role in this system. We expect that our ongoing microarray analysis of cells and tumors established in this study will help us address this issue. Moreover, future study will include a transgenic model overexpressing Ubc9 in breast epithelium to better understand the role of Ubc9 in tumorigenesis.

Materials and methods Ovarian tissue In all, 10 pairs of snap-frozen ovarian tumor and matched normal tissue were obtained from the GOG Tumor Bank (Columbus, OH, USA). The specimens were obtained from patients who provided written consent to participate in the GOG Banking Protocol, GOG 136. The use of these specimens in this study was approved by the Institutional Review Board of the University of Illinois at Chicago. Total RNA was isolated using Trizol reagent (Invitrogen). Single-stranded cDNAs were synthesized using SuperScriptase III (Invitrogen) along with random hexamer primers. To amplify ubc9 mRNA, PCR primers were as follows: Ubc9-5.1 (sense) GGATCCTCGGGGATCGCCCTCAGCA and Ubc9-3.1 (antisense) CTTATGAGGGCGCAAACTTCTTG. Cell culture MCF-7 cells were purchased from American Type Cell Collection (ATCC, Manassas, VA, USA) and grown in RPMI medium (BioWhittaker, Walkersville, MD, USA) at 371C in a humidified chamber in the presence of 5% CO2. All media were supplemented with 10% fetal bovine serum (Sigma, St Louis, MO, USA), 2 mM L-glutamine (BioWhittaker), 100 U of penicillin (BioWhittaker)/ml and 100 mg streptomycin (BioWhittaker)/ml. To determine cell growth, cells were seeded at the density of B300 000 cells/100 mm dish; cell numbers for each cell line were determined by Coulter Counter (BeckmanCoulter) at 24 h intervals for duplicated dishes at each time point. To determine cell survival, cells were seeded at the density of 0.6 million cells/well in six-well plates. Cells were harvested at different time points and stained with trypan blue. Surviving cells were counted under a microscope. Generation of stable transfectants overexpressing Ubc9-DN and Ubc9-WT in MCF-7 cells Construction of Myc-tagged Ubc9-DN and Ubc9-WT has been described previously (Mo et al., 2002). Transfection of MCF-7 cells was performed using a published method with a slight modification (Agami and Bernards, 2000). Briefly, a plasmid carrying Ubc9-DN or Ubc9-WT was introduced into MCF-7 cells along with pTK-hg (BD Biosciences, San Jose, CA, USA) by electroporation. The cells (0.6 million cells/ 0.4 ml) were subjected to a pulse at 220 V and 950 mF. Stable transfectants were selected in the presence of 200 mg hygromycin (Invitrogen, Carlsbad, CA, USA)/ml. Expression of the exogenous genes was confirmed by Western blot using antiUbc9 (BD Biosciences) or anti-Myc antibody (Zymed, South San Francisco, CA, USA). Ubc9 antibody was purchased from BD Biosciences (San Diego, CA, USA); Bcl-2, Bcl-X, Mcl-1, and Bax antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA); and antibody against b-actin was purchased from Sigma-Aldrich (St Louis, MO, USA). Oncogene

Tumor growth in nude mice Female nude (nu/nu) mice (5 weeks old) were purchased from Charles River Laboratory (Wilmington, MA, USA). Mice were maintained in the UIC’s AAALAC accredited animal facility. All experiments were performed in compliance with NIH animal use guidelines and a protocol approved by UIC Animal Care Committee. MCF-7 cells expressing Ubc9-WT, Ubc9-DN or vector alone were injected (B1 million cells/ injection) into mammary pads of female nude mice; all of these cells were viable based on nuclear staining. To facilitate tumor growth, a 0.72 mg 17b-estradiol pellet (Innovative Research of America, Sarasota, FL, USA) was implanted beneath the back skin. To reduce variations among inoculations, each mouse received two injections, one with vector control cell and the other with Ubc9-WT or Ubc9-DN cells. Primers and RT–PCR Total RNA was extracted from tumor cell lines or tumors derived from these cells using Trizol reagent (Invitrogen) as per the manufacturer’s instructions. Primers used in this study were as follows: Bcl-2-5.1 (sense, ACCTGGATCCAGGATAA CGGAG), Bcl-2-3.1 (antisense, CCAACAACATGGAAAG CGAATC); Bcl-x-5.1 (sense, ACGAGTTTGAACTGCGGTA CC); Bcl-x-3.1 (antisense, CCTGCATCTCCTTGTCTACGC); Mcl-1-5.1 (sense, AAAGAAACGCGGTAATCGGAC); Mcl1-3.1 (antisense, AGGGTAGTGACCCGTCCGTAC); Bax-5.1 (sense, AGCTCTGAGCAGATCATGAAG); and Bax-3.2 (antisense, AAGTTGCCGTCAGAAAACATG). Single-stranded cDNA was reverse transcribed using Invitrogen’s RT-kit as per the manufacturer’s protocol. PCR reactions: 941C for 3 min followed by 25–30 cycles of 941C for 30 s, 581C for 1 min and 721C for 30 s and then extended by 721C for 10 min. Microarray analysis Cells were harvested at a log phase, and total RNA was extracted and isolated using RNeasy isolation kit (Qiagen, Valencia, CA, USA) as per the manufacturer’s protocol. Three separate RNA extractions for each cell line were carried out on different days and pooled for each cell line. The pooled RNA samples were used as templates for cDNA and cRNA synthesis. Probe labeling and hybridization were all carried out at UIC Genomic Core Facility, using the Affymetrix Human Genome U133A GeneChip containing 22 283 gene sequences (with two chips per sample). Affymetrix Microarray Suite 5.0 was used to scan and analyse the arrays.

Abbreviations DN, dominant negative; RT, reverse transcription; PCR, polymerase chain reaction; SUMO, small ubiquitin-related modifier; Ub, ubiquitin; WT, wild type. Acknowledgements This work was supported by Grants CA40570, CA30103 (both to WTB) and CA102630 (to YM) from the National Institutes of Health, DAMD17-99-19220 (to WTB and YM) from DOD, and National Cancer Institute Grants CA27469 to the Gynecologic Oncology Group (GOG) with subcontracts for the GOG Molecular Pharmacology Core Laboratory and GOG Tissue Bank. We thank Martina Vaskova for her critical review of the manuscript.


2683 References Agami R and Bernards R. (2000). Cell, 102, 55–66. Alderson LM, Castleberg RL, Harsh GRt, Louis DN and Henson JW. (1995). Cancer Res., 55, 999–1001. Catz SD and Johnson JL. (2003). Apoptosis, 8, 29–37. Chauchereau A, Amazit L, Quesne M, Guiochon-Mantel A and Milgrom E. (2003). J. Biol. Chem., 278, 12335–12343. Cory S, Huang DC and Adams JM. (2003). Oncogene, 22, 8590–8607. Desterro JM, Rodriguez MS, Kemp GD and Hay RT. (1999). J. Biol. Chem., 274, 10618–10624. Dong L, Wang W, Wang F, Stoner M, Reed JC, Harigai M, Samudio I, Kladde MP, Vyhlidal C and Safe S. (1999). J. Biol. Chem., 274, 32099–32107. Giorgino F, de Robertis O, Laviola L, Montrone C, Perrini S, McCowen KC and Smith RJ. (2000). Proc. Natl. Acad. Sci. USA, 97, 1125–1130. Gostissa M, Hengstermann A, Fogal V, Sandy P, Schwarz SE, Scheffner M and Del Sal G. (1999). EMBO J., 18, 6462–6471. Hayashi T, Seki M, Maeda D, Wang W, Kawabe Y, Seki T, Saitoh H, Fukagawa T, Yagi H and Enomoto T. (2002). Exp. Cell Res., 280, 212–221. Hershko A and Ciechanover A. (1998). Annu. Rev. Biochem., 67, 425–479. Honda R, Tanaka H and Yasuda H. (1997). FEBS Lett., 420, 25–27. Johnson ES and Blobel G. (1997). J. Biol. Chem., 272, 26799–26802. Kahyo T, Nishida T and Yasuda H. (2001). Mol. Cell, 8, 713–718. Kovalenko OV, Plug AW, Haaf T, Gonda DK, Ashley T, Ward DC, Radding CM and Golub EI. (1996). Proc. Natl. Acad. Sci. USA, 93, 2958–2963. Leek RD, Kaklamanis L, Pezzella F, Gatter KC and Harris AL. (1994). Br. J. Cancer, 69, 135–139.

Leo C and Chen JD. (2000). Gene, 245, 1–11. Lipkowitz S. (2003). Breast Cancer Res., 5, 8–15. Martinez-Arribas F, Nunez-Villar MJ, Lucas AR, Sanchez J, Tejerina A and Schneider J. (2003). Anticancer Res., 23, 565–568. Mo YY, Yu Y, Ee PL and Beck WT. (2004). Cancer Res., 64, 2793–2798. Mo YY, Yu Y, Shen Z and Beck WT. (2002). J. Biol. Chem., 277, 2958–2964. Muller S, Hoege C, Pyrowolakis G and Jentsch S. (2001). Nat. Rev. Mol. Cell. Biol., 2, 202–210. Perillo B, Sasso A, Abbondanza C and Palumbo G. (2000). Mol. Cell. Biol., 20, 2890–2901. Pichler A, Gast A, Seeler JS, Dejean A and Melchior F. (2002). Cell, 108, 109–120. Pickart CM. (2001). Annu. Rev. Biochem., 70, 503–533. Poukka H, Aarnisalo P, Karvonen U, Palvimo JJ and Janne OA. (1999). J. Biol. Chem., 274, 19441–19446. Rodriguez MS, Desterro JM, Lain S, Midgley CA, Lane DP and Hay RT. (1999). EMBO J., 18, 6455–6461. Saitoh H and Hinchey J. (2000). J. Biol. Chem., 275, 6252–6258. Sato T, Hanada M, Bodrug S, Irie S, Iwama N, Boise LH, Thompson CB, Golemis E, Fong L and Wang HG et al. (1994). Proc. Natl. Acad. Sci. USA, 91, 9238–9242. Seeler JS and Dejean A. (2003). Nat. Rev. Mol. Cell. Biol., 4, 690–699. Seufert W, Futcher B and Jentsch S. (1995). Nature, 373, 78–81. Teixeira C, Reed JC and Pratt MA. (1995). Cancer Res., 55, 3902–3907. Zhang GJ, Kimijima I, Tsuchiya A and Abe R. (1998). Oncol. Rep., 5, 1211–1216. Zhang W, Ran S, Sambade M, Huang X and Thorpe PE. (2002). Angiogenesis, 5, 35–44.

Oncogene