The ubiquitin-specific protease USP28 is required for MYC stability

A RT I C L E S

The ubiquitin-specific protease USP28 is required for MYC stability Nikita Popov1, Michael Wanzel1, Mandy Madiredjo2, Dong Zhang3, Roderick Beijersbergen2, Rene Bernards2, Roland Moll4, Stephen J. Elledge3 and Martin Eilers1,5 The MYC proto-oncogene encodes a transcription factor that has been implicated in the genesis of many human tumours. Here, we used a bar-code short hairpin RNA (shRNA) screen to identify multiple genes that are required for MYC function. One of these genes encodes USP28, an ubiquitin-specific protease. USP28 is required for MYC stability in human tumour cells. USP28 binds to MYC through an interaction with FBW7α, an F-box protein that is part of an SCF-type ubiquitin ligase. Therefore, it stabilizes MYC in the nucleus, but not in the nucleolus, where MYC is degraded by FBW7γ. High expression levels of USP28 are found in colon and breast carcinomas, and stabilization of MYC by USP28 is essential for tumour-cell proliferation. The MYC proto-oncogene encodes a transcription factor, MYC, which is a central regulator of cell growth, proliferation and apoptosis. Enhanced levels of MYC are thought to contribute to the genesis of many human tumours1. Correspondingly, expression of MYC is regulated at multiple levels; for example, the MYC protein is unstable and rapidly degraded by the ubiquitin pathway2. The stability of MYC is enhanced in several human tumours. In some lymphomas, this is because of point mutations in the MYC gene, most of which alter the sequence encoding a conserved region in the amino-terminus termed MYCBoxI3. Within MYCBoxI, Thr 58 is recognized by the F-box protein, FBW7, which targets MYC for degradation by the proteasome pathway4,5. The point mutations found in lymphomas either affect Thr 58 directly, or residues that are part of the signalling pathway that controls phosphorylation of Thr 58 by GSK3 (ref. 6). Stabilization of MYC proteins can also occur in the absence of mutations in the MYC coding sequence7,8. This may reflect alterations in regulatory pathways that control degradation by the GSK3–FBW7 pathway. Also, several additional ubiquitin ligases have been implicated in MYC turnover, but little is known about the pathways that control their activity9–11. Here, we have screened a retroviral shRNA library to identify genes that are required for MYC function. One of the genes identified in this screen encodes an ubiquitin-specific protease, USP28. Ubiquitin-specific proteases antagonize the activity of ubiquitin ligases12. We report that USP28 controls MYC stability through antagonizing the activity of the SCFFBW7 ubiquitin ligase complex, and that the stabilization of MYC by USP28 is required for proliferation of several tumour cell types and for inhibition of cell differentiation in colon carcinoma.

RESULTS Identification of USP28 To identify genes that are required for MYC function, we made use of the finding that high levels of MYC can induce apoptosis in the absence of survival factors13 and reasoned that shRNAs that inhibit or dampen MYC function might confer a survival advantage under such conditions. We used a clone of U2OS osteosarcoma cells that expresses a conditional MYC−ER protein14. On serum starvation, these cells accumulated in the G0 phase of the cell cycle; activation of MYC rapidly stimulated proliferation and subsequent apoptosis and, ultimately, all cells died within several days (Fig. 1a, d). The cells were infected with a retroviral library that comprises 23,742 shRNA vectors targeting 7,914 genes15, serum starved and then treated with 4-OHT. At several time points, DNA was harvested from the surviving cells and, as control, from cells before addition of 4-OHT. Subsequently, the relative representation of each shRNA was measured using a microarray16. Initial experiments showed a significant enrichment of individual shRNAs only when 90% of the 4-OHT treated cells had died (see Supplementary Fig. S1a, b). To minimize experimental variation, data from six individual experiments was combined. In addition to MYC itself, this approach identified 91 targeted genes, for which shRNAs were identified at least three times in these experiments (see Supplementary Information, Table S1). Several arguments suggest that a significant proportion of these genes have a role in MYC function or MYC-induced apoptosis: first, virtually all genes that are known to be involved in MYC-induced apoptosis are present on the list, including MYC itself, MAX17,

1 Institute of Molecular Biology and Tumor Research, Emil-Mannkopff-Str.2, 35033 Marburg, Germany. 2Netherlands Cancer Institute, Division of Molecular Carcinogenesis, Plesmanlaan 121, 1066 CX Amsterdam, Netherlands. 3Department of Genetics, Center for Genetics and Genomics, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, MA 02115, USA. 4Department of Pathology, University of Marburg, Baldingerstrasse, 35033 Marburg, Germany. 5 Correspondence should be addressed to M.E. ([email protected])

Received 14 December 2006; accepted 11 May 2007; published online 10 June 2007; DOI: 10.1038/ncb1601

NATURE CELL BIOLOGY VOLUME 9 | NUMBER 7 | JULY 2007

765

A RT I C L E S

40 20 0

Ethanol - 4-OHT 0

1

2 3 Day

4

5

10

0

6

0

1

2

3 Day

THOC1-1

b

4

5

6

Percentage of cells in SubG1

60

20

Percentage of cells in G2M

80

Percentage of cells in S

Percentage of cells in G1

a 30 20 10 0 0

1

2

3 Day

5

80 60 40 20 0 0

6

c

USP28-1

THOC1-2

4

100

1

2

3 Day

4

5

6

USP28 shRNA

d0 +4-OHT + 4-OHT

USP28-2

H2AZ-1

H2AZ-2

d6 +4-OHT

– 4-OHT

Mr(K) 100

60 40

72 72

20

50

0 Day

4-OHT

0

5 –

5 +

0

5 –

5 +

34

f

Anti-MYC–ER

Anti-MYC

Anti-CDK2

Scr shRNA USP28 shRNA

Percentage of BrdU positive cells

e 80

USP28 shRNA

+ 4-OHT

Scr shRNA

d

Percentage of subG1 cells

Control

Time (h):

60 Scr shRNA

USP28 shRNA

50 + 4-OHT – 4-OHT

+ 4-OHT – 4-OHT

40

30

20 0

10

20

30

0

10

20

30

Figure 1 A bar-code shRNA screen identifies USP28 as a gene required for MYC function. (a) FACS analysis of a selected U2OS cell MYC–ER clone. The panels show the percentage of G1, S, G2–M and subG1 cells after the cells were serum starved for 48 h and subsequently treated either with 200 nM 4-OHT or ethanol as control for the indicated times. (b) Validation of candidate shRNAs in a colony formation assay. Serumstarved cells infected with the indicated shRNA viruses were either treated with 4-OHT or ethanol for 6 days before the cells were fixed and subsequently stained with Giemsa. (c) Apoptosis of U2OS MYC–ER cells infected with a control vector or with USP28 shRNA. The panels show photographs of representative plates infected with either scrambled (Scr) shRNA or USP28 shRNA vectors before and 6 days after addition of 4-OHT. The scale bars represent 100 μm. (d) Percentage of subG1

cells in the absence and presence of 4-OHT after serum starvation for 48 h and addition of 4-OHT or ethanol at the indicated times. The error bars represent s.d. (n = 3). (e) Immunoblot documenting the amount of endogenous MYC and the MYC–ER chimeric protein in cells expressing USP28 shRNA. (f) Summary of FACSscan experiments of U2OS MYC–ER cells that express a control shRNA or USP28 shRNA. Cells were serum starved for 2 days and restimulated by addition of 30 nM 4-OHT or ethanol as control. Trial experiments had shown that addition of 30 nM yields approximately 80% maximal stimulation of S-phase (data not shown). For each time point, three independent cell samples were labelled with BrdU for 1 h, harvested and the percentage of S-phase cells was determined by FACS. The data plot the percentage of cells that incorporate BrdU. The error bars correspond to s.d. from three independent experiments.

E2F1 (ref. 18), BAX19, HMGA1 (ref. 20) and PAK2 (ref. 21; see Supplementary Information, Table S1). Notable exceptions are shRNAs targeting P53 and p14ARF; indeed, MYC-induced apoptosis is p53-independent, most likely because the ARF-locus is silenced in U2OS cells22 (see Supplementary Information, Fig. S1c, d). Second, sixteen shRNA vectors from this list were tested individually and thirteen partially protected from MYC-induced cell death (Fig. 1b and see

Supplementary Information, Table S2). In contrast, only one out of ten shRNA vectors that were picked from a group of genes, which had occurred only once in these screens, protected from MYC-induced apoptosis (P = 0.00168; see Supplementary Information, Table S2). Third, in addition to USP28 (see below) other genes on the list (for example, H2A.Z) are required for transcriptional regulation by MYC (data not shown).

766


RN A

hR NA

sh

Mr(K) 130

USP28 mRNA

MYC mRNA Mr(K) 130

Anti-USP28

+

+ MYC

c –

+

++

HA–USP28

72 50

Anti-MYC

34

Anti-CDK2

– –

Mr(K) 130 100

Anti-HA

MYC mRNA 72

C171A

P2 US

b

WT

8-4 s P2

US

8- 5

hR NA

hR NA

8-3 s

8-1 s US

P2

P2

– +

Flag–USP28

Anti-USP28

72 50

Anti-Myc

34

Anti-CDK2

Anti-MYC

50 34

r

–M

d M (K)

yc

Anti-CDK2

0 90 180 270 360

0

– – + +

e 90 180 270 360 CHX (min)

130 100

Mr(K)

Anti-HA

C171A

GAPDH mRNA

WT

US

a

Sc rs

hR NA

A RT I C L E S

– Flag–USP28 + + + His–UB + + + FBW7α + + + MYC

72

Anti-MYC

50

72

Input

Anti-Flag

50

Anti-Flag

100

34

Anti-CDK2 130 + HA–USP28

g Anti-Myc

50

Mr(K)

130 100 72 50

Control

Anti-USP28

Anti-USP28

72 50

+ FBW7α + USP28

Control

72

– +

Anti-MYC

Mr(K)

130

– –

Input

f

UB– MYC

100

UB–MYC

130 100

Anti-USP28

Mono MYC IgG

Figure 2 Stabilization and deubiquitination of MYC by USP28. (a) Depletion of USP28 decreases MYC protein levels. HeLa cells were transfected with the indicated pRetroSuper vectors. Two days after transfection, extracts were analysed by RT–PCR and immunoblotting, respectively. (b) Ectopic expression of USP28 increases steady-state levels of endogenous MYC protein. The upper panels show immunoblots of HeLa cells transiently transfected with a CMV-driven expression vector encoding HA-tagged USP28. + and ++ denote increasing amounts of the HA–USP28 expression vector. The lower panels show RT–PCR assays from a parallel experiment documenting levels of MYC and GAPDH mRNAs. (c) The catalytic activity of USP28 is required to regulate MYC levels. The panels show immunoblots of HeLa cells transiently transfected with CMV-driven expression vectors encoding either Flag-tagged USP28 or Flag-tagged USP28C171A, in which the catalytic cysteine is replaced by alanine. (d) Expression of USP28 stabilizes MYC. HeLa cells were transfected with an expression vector for MYC and a CMV-driven HA-USP28 expression vector or a control. The panels


72

Anti-MYC

50

show immunoblots of the indicated proteins before and at the indicated times after addition of cycloheximide. Quantification of the data is shown in Fig. 3e. (e) USP28 reduces ubiquitination of MYC in vivo. Cells were cotransfected with the indicated plasmids and ubiquitinated MYC was recovered on Ni-NTA resin. (f) USP28 deubiquitinates MYC in vitro. Cells were transfected with expression plasmids encoding MYC and His-tagged ubiquitin and ubiquitinated MYC recovered on Ni-NTA resin. The beads were mixed with USP28 that was immunoprecipitated from cells either expressing USP28 or USP28 together with FBW7α, and incubated in vitro. After the incubation, the Ni-agarose beads were re-isolated, boiled in SDS-sample buffer and an immunoblot of the samples probed with anti-MYC antibodies. (g) Interaction of endogenous USP28 with MYC. Lysates of HeLa cells were precipitated with either anti-MYC, anti-USP28 or control (anti-CDK2) antibodies as indicated. Precipitates were probed with either anti-MYC or anti-USP28 antibodies. The input lane corresponds to 10% of the amount of protein used in the immunoprecipitates.

767

A RT I C L E S Two secondary assays were performed: using FACS, whether individual shRNAs reduced the extent of MYC-induced apoptosis was measured (data not shown and Fig. 1c, d); and immunoblots were performed to determine the level of MYC–ER and endogenous MYC proteins (Fig. 1e). From these secondary screens, an shRNA directed against USP28 (USP28 shRNA) was selected for further characterization, as it showed a robust reduction of apoptosis and was the only shRNA that affected steady-state levels of MYC proteins. USP28 encodes an ubiquitin-specific protease; members of this class of enzymes can enhance the stability of individual proteins by antagonizing the activity of ubiquitin E3 ligases12. Recently, USP28 has been identified as a 53BP1-associated protein that controls the levels of 53BP1 and claspin23. In that study, the deubiquitinating activity of USP28 was demonstrated using the tetra-ubiquitin cleavage assay. Consistent with such a role, expression of USP28 shRNA decreased the levels of both the ectopically expressed MYC–ER and the endogenous MYC protein, demonstrating that it regulates MYC levels at a post-transcriptional level (Fig. 1e). Furthermore, USP28 shRNA reduced both basal and MYC-induced proliferation of U2OS MYC–ER cells, demonstrating that it does not selectively regulate the pro-apoptotic function of MYC (Fig. 1f). Three additional shRNA vectors directed against USP28 inhibited apoptosis in U2OS-MYC–ER cells (data not shown) and reduced the level of endogenous MYC protein in HeLa cells without affecting the levels of MYC mRNA (Fig. 2a); therefore, the regulation of MYC protein levels by USP28 shRNA does not reflect an off-target effect of a single shRNA. Identical results were obtained in several other tumour cell lines (see below). A screen of shRNA vectors directed against 30 additional ubiquitin-specific proteases did not identify another vector with a similar effect, suggesting that the effect of USP28 shRNA is specific (see below). USP stablizes MYC Expression of a HA-tagged cDNA encoding USP28 enhanced steadystate levels of cotransfected (data not shown) and endogenous MYC proteins in HeLa cells without altering MYC mRNA levels (Fig. 2b). Expression of a mutant allele of USP28, in which the catalytic cysteine has been replaced by alanine (C171A), had no effect on MYC levels (Fig. 2c). To determine whether USP28 regulates MYC stability, both MYC and USP28 were expressed by transient transfection in HeLa cells and the synthesis of new proteins was blocked by cycloheximide. Timecourse experiments showed that the half-life of transfected MYC in the absence of USP28 was 1.2 h, approximately twofold higher than what has been reported for endogenous MYC7. Transfection of USP28 increased the half-life of cotransfected MYC to 2.6 h (Fig. 2d; a quantification of the results is shown in Fig. 3e). Expression of wild-type USP28, but not of USP28C171A, in the presence of His-tagged ubiquitin decreased the amount of ubiquitinated MYC, demonstrating that USP28 can deubiquitinate MYC in vivo (Fig. 2e). To determine whether USP28 can deubiquitinate MYC in vitro, ubiquitinated MYC was purified from cells that were cotransfected with expression vectors encoding MYC and His-tagged ubiquitin by precipitation with Ni-agarose. The predominant species recovered in these experiments corresponded to mono-ubiquitinated MYC, as judged by its mobility (Fig. 2f). Mixing these preparations with USP28 that had been immunoprecipitated from transfected cells led to a reduction of ubiquitinated MYC. Experiments described below led us to conclude that USP28 binds to MYC indirectly through the F-box protein 768

FBW7; correspondingly, USP28 immunopurified from cells expressing both USP28 and FBW7 had an enhanced capacity to deubiquitinate MYC (Fig. 2f). Furthermore, endogenous USP28 coimmunoprecipitated endogenous MYC from cell extracts, demonstrating that the two proteins form a complex in vivo (Fig. 2g). We were unable to coprecipitate a significant amount of USP28 with anti-MYC antibodies; most likely, this is due to the very low number of MYC molecules present in most cells. Our data suggest that USP28 directly deubiquitinates MYC in vivo. Immunoblots of HeLa cells revealed that USP28 shRNA not only reduced protein levels of MYC, but also those of cyclin E (Fig. 3a), both of which are degraded by the same ubiquitin ligase, SCFFBW7 (also called hCDC4 or archipelago)4,5,24,25. In contrast, USP28 shRNA slightly increased protein levels of p27 and p130, two substrates of the SCF–SKP2 complex26,27, and did not affect levels of SKP2 and cyclin A, which are ubiquitinated by the APC complex28,29. Ectopic expression of USP28 abrogated FBW7-mediated degradation of MYC (Fig. 3b). In contrast, USP25, which is the deubiquitinating enzyme most closely related to USP28, did not antagonize FBW7-mediated degradation of MYC (Fig. 3c). Furthermore, neither ectopic expression nor depletion of USP28 had any effect on the stability of MYCT58A, which is not recognized by FBW7 (Fig. 3d, e and data not shown)4,5. Also, ectopic expression of USP28 did not further stabilize MYC in cells depleted of FBW7 (data not shown). Similarly, ectopic expression of USP28 stabilized wild-type cyclin E, but had no effect on a point mutant (cyclin ET380A) that does not bind FBW7 (Fig. 3f)30. Finally, depletion of FBW7 using an shRNA vector abolished the requirement for USP28 in maintaining the stability of endogenous MYC protein (Fig. 3g). We concluded that USP28 specifically antagonizes the action of FBW7 in MYC and cyclin E degradation. There are three isoforms of FBW7 (α, β and γ) that differ in their amino-terminal sequences and in their subcellular localization. Both the nuclear (FBW7α) and the nucleolar (FBW7γ) isoforms bind to and degrade MYC when overexpressed; in contrast, FBW7β is localized in the cytosol and does not bind MYC31. Surprisingly, MYC is selectively degraded in the nucleolus by FBW7γ. The reason for this specificity is unknown, as the WD40 domains that interact with phosphorylated Thr 58 of MYC are localized in the shared carboxyl-terminus of FBW7 (ref. 31). Immunofluorescence microscopy experiments showed that HA–USP28 was predominantly localized in the nucleus excluding the nucleolus, similar to FBW7α and MYC, but distinct from FBW7γ (Fig. 4a). Consistently, transfection of HA–USP28 enhanced steady levels of MYC in the nucleus, but not in the nucleolus, whereas inhibition of the proteasome led to a nucleolar accumulation of MYC, suggesting that USP28 might specifically antagonize the function of FBW7α (see Supplementary Information, Fig. S2a). USP28 binds to MYC via FBW7α Several deubiquitinating enzymes are recruited to their substrate indirectly through binding to the E3 ligase32,33. Consistent with this notion, a binary complex between USP28 and FBW7 was detectable in lysates from transfected cells (see Supplementary Information, Fig. S2b) — endogenous FBW7 was not detected using several different antibodies. Both FBW7α and FBW7γ formed binary complexes with USP28. In contrast, USP28 did not bind to ROC1, SKP1 or CUL1. Furthermore, USP28 did not interact with SKP2, another F-box protein that interacts with MYC9 (see Supplementary Information, Fig. S3a and data not shown). Neither deletion of the amino-terminus nor of the carboxy-terminal WD40 repeats NATURE CELL BIOLOGY VOLUME 9 | NUMBER 7 | JULY 2007

Mr(K) Mr(K) 72 34

b

Anti-p27

50 50

Anti-cyclin E

43

Anti-actin

50

Mr(K) 130 100

+ – –

+ + –

+ – +

+ + +

MYC Flag–FBW7α HA–USP28 Anti-HA

c

50

34

d

WT +

+

–

+

+

HA–USP28 –

–

+

–

–

+

Anti–USP28

Anti-Flag

Mr(K) 130 100 72

Anti-MYC

50 34

Anti-CDK2

+ + 28

+ MYC + Flag–FBW7α 25 HA–USP Anti-HA

Anti-MYC

Anti-Flag

100

Anti-CDK2

e

T58A

–

+ + –

50

Anti-SKP2 100

+ – –

72 MYC

43

– – –

Mr(K) 130 100

Anti-cyclin A 72

34

MYC

– – –

Percentage of remaing MYC protein

Anti-MYC

hR NA Us p2 8s hR NA

Sc rs

a

Sc rs

hR NA Us p2 8s hR NA

A RT I C L E S

34

Anti-CDK2

100

MYCT58A MYCT58A/USP28 MYC MYC/USP28

10

100

0

200

300

WT +

–

T380A + +

–

–

+

–

–

+

FBW7α +

+

+

+

+

+

Anti-HA

g Mr(K) 72 Mr(K) 130 100

50 34

–

+

FBW7 shRNA + USP28 shRNA – Anti-MYC Anti-CDK2

72 Anti-cyclin E

50

Anti-CDK2

34

MYC mRNA

FBW7 mRNA

USP28 mRNA

1

0.5

0

FBW7 shRNA USP28 shRNA

+

FBW7 shRNA

–

USP28 shRNA

Cyclin E HA–USP28

Control

f

MYC protein levels (relative to CDK2 and control)

Time CHX (min)

Figure 3 USP28 antagonizes the function of FBW7. (a) Depletion of USP28 decreases abundance of cyclin E and of MYC, but not of other unstable proteins. Whole cell extracts from HeLa cells transfected with the indicated plasmids were harvested two days after transfection and immunoblotted using the antibodies shown. (b) FBW7α-mediated degradation of MYC is blocked by expression of USP28. HeLa cells were transfected with the expression plasmids shown and protein levels were monitored by immunoblotting. (c) FBW7α-mediated degradation of MYC is inhibited by USP28, but not by USP25. The experiment was performed as in b. (d) Immunoblots documenting steady-state levels of wild-type MYC or MYCT58A after expression of HA–USP28. To show an exposure in the linear range, a threefold shorter exposure of the MYCT58A blot is shown.

(e) Quantification of immunoblots documenting the change in stability of wild-type MYC, but not of MYCT58A on coexpression of HA–USP28. Protein stability was measured as described for Fig. 2c. (f) Immunoblots showing steady state levels of wild-type cyclin E or of cyclin ET380A in the absence or presence of coexpressed USP28. FBW7α was expressed in all lanes. (g) Depletion of FBW7 abolishes the effect of USP28 shRNA on endogenous MYC levels. HeLa cells were cotransfected with USP28 shRNA and with a pRetroSuper vector targeting FBW7, as indicated. The upper panels show protein levels of MYC and CDK2, the lower panels RT–PCR assays documenting levels of MYC, FBW7 and USP28 mRNAs. A summary of four independent experiments documenting the average levels of MYC protein (normalized to CDK2 and control) in cells expressing the indicated shRNAs is also shown.

abrogated interaction of FBW7 with USP28, suggesting that more than one domain of FBW7 can mediate the formation of a binary complex with USP28 (see Supplementary Information, Fig. S2b and Discussion). We then examined whether FBW7 mediates the interaction between MYC and USP28 in vivo. On cotransfection of HA–USP28 and MYC, only a small amount of complex formed (not visible in the exposure

of Fig. 4b, but visible in Fig. 4c). Coexpression of FBW7α strongly stimulated complex formation between HA–USP28 and MYC. In contrast, expression of FBW7γ did not stimulate complex formation between MYC and HA–USP28 (Fig. 4b). Furthermore, expression of FBW7αΔΝ, in which the amino-terminus is deleted, abrogated coprecipitation of HA–USP28 with MYC (Fig. 4c), strongly suggesting that the


769

A RT I C L E S

a USP28

FBW7α

FBW7γ

b Mr(K)

– + –

α + +

– + +

γ + +

Flag–FBW7 HA–USP28 MYC

72 50

c Mr(K)

Anti-MYC

FBW7α (anti-Flag)

Input

– + –

– + +

ΔN Flag–FBW7 + HA–USP28 + MYC

130 100

Anti-HA

72

130

Anti-HA FBW7α (anti-Flag)

100 72

72

ΔN FBW7 + HA–USP28 + MYC Anti-HA

34

+

+

+

e – +

72

Anti-MYC

Mr(K) 130

Mr(K) 130

IP: anti-HA

+

+

HA–USP28 Anti-HA

Anti-HA Input

72 50

72 Anti-CDK2

Anti-Flag

f MYC HA–USP28

Anti-MYC

50

72 50

Scr shRNA

γ + +

Anti-MYC

FBW7 shRNA

α + +

T58

– + +

T58/S62A

– – +

72

Anti-HA

WT

– – Mr(K) – 130 100

IP: anti-MYC

FBW7γ

130 100

d

Input

FBW7

50

Anti-MYC

130

50

Anti-MYC

Anti-HA IP: anti-MYC

100 130

Anti-HA

IP: anti-MYC

100

Figure 4 USP28 binds to MYC via FBW7α. (a) USP28 and FBW7α localize to the nucleoplasm, whereas FBW7γ localizes to the nucleolus. Immunostaining was performed on transiently transfected HeLa cells with anti-HA (USP28) and anti-Flag (FBW7) antibodies. The scale bars represent 20 μm. (b) FBW7α enhances binding of USP28 to MYC. Cells were cotransfected with HA-tagged USP28, MYC and Flag-tagged FBW7α or FBW7γ. Two days after transfections, cells were treated with the proteasome inhibitor MG132 for 3 h, lysed, and protein complexes were recovered with anti-MYC N262 antibody. (c) The N-terminal region of FBW7 mediates recruitment of MYC to USP28. HeLa cells were cotransfected as indicated, treated with MG132, lysed, and immunoprecipitation was performed with anti-HA antibody. (d) USP28

inhibits MYC-induced degradation by FBW7α, but not by FBW7γ and FBW7ΔN. The panels show immunoblots of HeLa cells after transfection with the indicated expression plasmids. (e) MYCT58A and MYCT58A/S62A show a reduced interaction with USP28. HeLa cells were cotransfected with HA-tagged USP28 and expression plasmids encoding either wild-type MYC, MYCT58A or MYCT58A/S62A. Cell lysates were immunoprecipitated with antiMYC (N262) antibody and probed with antibodies against MYC (9E10) or USP28. Input blots show a 5% aliquot. (f) Depletion of FBW7 decreases binding of USP28 to MYC. HeLa cells were cotransfected with HA–USP28 and either Scr shRNA or USP28 shRNA. After transfection, cells were lysed and aliquots immunoprecipitated with an anti-MYC antibody (N262). Precipitates were probed with an anti-HA antibody.

amino-terminally deleted protein competes with endogenous FBW7 for binding to MYC. Consistent with this model, both FBW7γ and FBW7αΔN were able to promote degradation of MYC in the presence of USP28, in contrast with FBW7α (Fig. 4d). If FBW7α mediates complex formation between USP28 and MYC, mutations in MYC that disrupt binding of MYC to FBW7 should also affect complex formation between MYC and USP28. Indeed, both MYCT58A and the double mutant MYCT58A/S62A (refs 4, 5), showed a strongly diminished interaction with USP28 (Fig. 4e). Furthermore, depletion of FBW7 reduced the amount of endogenous MYC that was bound to HA-tagged USP28 (Fig. 4f). Our data strongly suggest that in growing cells MYC is stabilized in the nucleus, but not in the nucleolus, as only FBW7α is associated with USP28. These data might also explain why FBW7α, in contrast to FBW7γ, is unable to polyubiquitinate and promote degradation of cyclin E34.

USP28 is required for tumour-cell proliferation To determine whether inhibition of USP28 has an impact on proliferation of human tumour cells, a panel of cell lines derived from breast, lung and colon carcinomas, and from glioblastoma, was used (Fig. 5a,b). Expression of USP28 shRNA led to a strong decrease in MYC levels in all cell lines we tested (Figs 2a and 5a, b). Consistent with a requirement for high MYC levels for proliferation of some human tumour cells, depletion of USP28 strongly inhibited growth of the human tumour cell lines that were analysed in a colony formation assay. Parallel experiments showed that MYC shRNA inhibited growth of these lines to a similar extent (Fig. 5a, b and data not shown). To determine whether USP28 regulates expression of downstream targets of MYC in vivo, MYC shRNA or USP28 shRNA was expressed in Ls174T colon carcinoma cells and microarray experiments were performed. The majority of genes that were regulated in response to

770


hR NA Mr(K) 72 50

d

r =0.62

2.0

–2.0

Scr shRNA

Tumour type

MYC levels

HeLa

Cervix

+

+

A549

Lung

+

n.d.

Lung

+

n.d.

A172

Glioblastoma

+

+

Ls174T

Colon

+

+

Anti-CDK2

HCT116

Colon

+

n.d.

MCF7

Breast

+

+

USP28-1 shRNA e 15

MYC

2.0 MYCT58A –2.0

Log fold regulation by USP28 shRNA

Proliferation

H157

Δ(USP28 shRNA-Scr shRNA)

Log fold regulation by MYC shRNA

c

Cell line

Anti-MYC

Ls174T 34

Effect of USP28 shRNA on:

b

USP28-4 shRNA

HeLa

Scr shRNA USP28-1 shRNA USP28-2 shRNA USP28-3 shRNA

P2 US

MY Cs

8-4 s

sh 8-1 US

Sc rs

a

P2

hR NA

RN A

hR NA

A RT I C L E S

f

G1 10 5

S

G2 MYC

Myc: WT USP28 shRNA: – +

T58A – +

USP28

MYCT58A 45SrRNA

0 –5

tRNALEU

–10

5SRNA

–15

Retroviral MYC

Figure 5 Depletion of USP28 inhibits growth and proliferation via regulation of MYC protein levels. (a) A colony assay documenting inhibition of cell growth of HeLa and Ls174T colon carcinoma cells after stable transfection with two different USP28 shRNA vectors or a MYC shRNA vector is shown. Plates were stained 10 days after transfection. Immunoblots documenting the reduction in MYC levels in Ls174T cells after stable transfection with four different USP28 shRNA vectors are slos shown. (b) Summary of the effects of USP28 shRNA in several human tumour cell lines. In each case, cells were transfected with four different USP28 shRNA vectors. + indicates that two or more of these vectors led to a significant downregulation of levels of MYC protein or of colony formation, respectively. n.d., not determined. (c) Summary of the effects of USP28 shRNA on gene expression. The result of a microarray experiment of Ls174T colon carcinoma cells expressing USP28 shRNA, MYC shRNA or Scr shRNA as control, is shown. Each dot represents the expression change of an

individual gene. All genes that were regulated by more than 1.5-fold in response to expression of either MYC shRNA or USP28 shRNA relative to control cells are shown. For each gene, the value on the x-axis shows the fold change in expression that is induced in response to USP28 shRNA, the y-axis the expression change for the same gene that is induced by MYC shRNA. (d) Depletion of USP28 suppresses colony formation of HeLa cells expressing wild-type MYC, but not cells expressing MYCT58A. HeLa cells expressing either protein were stably transfected with USP28 shRNA or with control vectors as indicated. Plates were stained 10 days after transfection. (e) FACS scan documenting the changes in cell-cycle distribution induced by depletion of USP28 in HeLa cells expressing either wild-type MYC or MYCT58A. (f) RT–PCR assays documenting expression of USP28 mRNA, FBW7 mRNA and of the 45S rRNA precursor, tRNALEU and 5SRNA in HeLa cells expressing either wild-type MYC or MYCT58A after stable transfection with either Scr shRNA (–) or USP28 shRNA (+).

USP28 shRNA were also regulated in the same direction by MYC shRNA (P 90%) of the tumour cell nuclei, but excluding nucleoli. Among normal cells, there was weak nuclear staining of some endothelial, smooth muscle and mononuclear cells. High levels of USP28 expression were also found in metastases derived from colon carcinoma (Fig. 6c). Western blotting of normal colon and of tumour samples confirmed that the staining did reflect differential expression of USP28 (see Supplementary Information, Fig. S3d). Similarly, nine of the ten invasive ductal breast carcinomas revealed strong staining of most (> 90%) tumour cell nuclei (Fig. 6d). To demonstrate that levels of USP28 can be rate limiting for proliferation, wild-type USP28 and catalytically inactive USP28C171A were stably expressed in Ls174T colon carcinoma cells. Consistent with previous observations, western blotting revealed that cells expressing wild-type USP28, but not cells expressing USP28C171A expressed higher levels of MYC protein (Fig. 6e). FACS analysis showed that these cells had an increased percentage of cells incorporating BrdU relative to control cells and to cells expressing USP28C171A (Fig. 6e). Similar data were obtained in NIH3T3 cells (data not shown). Ectopic expression of USP28 was not sufficient to transform primary mouse embryonic fibroblasts in conjunction with Ras, potentially due to low levels of c-myc mRNA in these cells (M.W. and M.E., unpublished observations). In colon carcinoma cells, high levels of MYC are required to maintain a crypt and/or progenitor phenotype and downregulation of MYC induces terminal differentiation37,38. Consistent with the requirement for USP28 in stabilizing MYC in these cells, depletion of MYC or USP28 led to an increase in expression of P21CIP1 and markers of terminal differentiation including GAL4, EFNB1 and MUC2 (Fig. 6e and data not shown)38. DISCUSSION MYC proteins activate transcription by both RNA polymerase I in the nucleolus and by RNA polymerase II in the nucleus, suggesting that they may coordinate nuclear and nucleolar transcription39. One mechanism that links the function of MYC in the nucleus to nucleolar events is proteasomal degradation, as MYC is selectively degraded by the nucleolar isozyme of FBW7, FBW7γ31,40. The observation is surprising as the WD40 domains of FBW7, which recognize phosphorylated Thr 58 in MYC, are present in both FBW7α and FBW7γ. We now show that the action of FBW7α is antagonized by the ubiquitin-specific protease USP28, which forms a complex with MYC and FBW7α, but not with FBW7γ. The formation of the ternary complex depends on the aminoterminus of FBW7α, which is not present in FBW7γ. USP28 does not bind to MYC directly, but binds MYC through interaction with FBW7. There are previous examples in which an ubiquitin-specific protease interacts directly with an ubiquitin ligase — USP7 interacts with Mdm2, the E3 ligase that degrades p53, and USP33 interacts with VHL, which targets Hif-1α33,41. These interactions may regulate the stability of the E3 ligase and potentially protect it from auto-ubiquitination41; conversely, the bound USP may be a substrate for the E3 ligase. NATURE CELL BIOLOGY VOLUME 9 | NUMBER 7 | JULY 2007

A RT I C L E S It is therefore possible that binding of USP28 to FBW7 regulates either USP28 or FBW7 function. Indeed, USP28 can form a binary complex with both FBW7α and FBWγ and depletion of FBW7 stabilizes USP28, suggesting that USP28 may be able to interact with the WD40 repeats and be a substrate of FBW7 (N.P., unpublished observations). Alternatively, however, USP28 forms a ternary complex with FBW7α and MYC and antagonizes the degradation of MYC by FBW7α. In our view, this may serve two functions: first, USP28 might couple degradation of MYC to external stimuli. For example, USP28 has been implicated in the cellular response to DNA damage23 and steady-state levels of MYC proteins decline rapidly after exposure of cells to etoposide42 or UV irradiation, even when expressed from a constitutive promoter (Herold, S. and M.E., unpublished observations), suggesting that MYC is degraded in response to DNA damage. It will be interesting to determine whether USP28 is involved in this response. Second, in analogy to similar cycles in metabolism, this form of regulation may allow both rapid and sharp alterations of MYC levels in response to small or gradual changes in external stimuli. Mutations in MYC at Thr 58 and Ser 62, which control association with FBW7 and USP28, do not only affect protein stability, but also alter the apoptotic and gene regulatory functions of MYC43,44. It is possible, therefore, that FBW7 and USP28, via selective degradation or stabilization of MYC that is phosphorylated at Thr 58, alter the functional properties of the pool of MYC that is present in a cell, in addition to controlling total MYC levels. Several previous studies have illustrated the potential of shRNA screens to identify novel targets for tumour therapy45. Also, targeting MYC may be of significant therapeutic value for different tumour types46. USP28 is required for stabilization of MYC in human tumour cells and depletion of USP28 mimics the effect of MYC depletion in tumour cells. The stabilizing effect of USP28 depends on its catalytic activity and its ability to reverse FBW7-mediated ubiquitination. USP28 is a cysteine protease and small molecule inhibitors can selectively inhibit this class of enzymes47. We propose, therefore, that inhibition of USP28 may be a pharmacologically feasible approach to interfere with MYC function in human tumours. METHODS Cell culture. Cells were cultured in DMEM (for HeLa and U2OS cells), McCoy’s (HCT116), or RPMI (Ls174T) media, all supplemented with 10% FBS. Transfections were performed by calcium phosphate precipitation. After transfection (24 h), cells were washed with PBS, and supplemented with fresh medium. For transfections with shRNA vectors, cells were selected with 2 μg ml–1 puromycin. Cells were harvested 48–72 h after transfection. To generate U2OS MYC–ER cells, U2OS cells were stably transfected with plasmids expressing the ecotropic retroviral receptor and subsequently infected with viruses expressing MYC–ER. Individual clones that expressed high levels of MYC–ER proteins were plated at about 50% confluency, serum starved with 0.1% serum for 48 h and then treated either with 4-OHT or ethanol. Medium (supplemented with 0.1% FCS) containing either freshly prepared 4-OHT or ethanol was changed every second day. Colony formation was assayed six days later. For FACS analysis, 50,000 cells were counted and their cell-cycle profile was analysed. Apoptosis was analysed by measuring cells in subG1-phase or by staining live cells with propidium iodide. Bar-code screens. Six independent screens were performed. For each screen, 1.5 × 107 U2OS MYC–ER cells were infected with the retroviral NKI shRNA library, selected and subjected to 4-OHT treatment for 14 days. Reference samples were harvested before adding 4-OHT. Genomic DNA was prepared and the shRNA cassettes were recovered by PCR. PCR products were amplified by linear RNA amplification and labelled with Cy3 (before treatment) or Cy5 (4-OHT


treated) using ULS system (Universal Linkage system; Kreatech Biotechnology, Amsterdam, The Netherlands). The labelled RNA probes were purified and used for hybridization to oligonucleotide arrays containing all specific 19mer sequences. shRNA bar code-hits were recloned in pRetroSuper and individually tested for their ability to inhibit MYC-induced apoptosis. Plasmids. To construct shRNA vectors for USP28, MYC and FBW7, hairpin-encoding oligonucleotides were annealed and ligated into pRetroSuper vector48. The following targeting sequences were used: USP28-1, GTATGGACAAGAGCGTTGG; USP28-2, CAAGAGCGTTGGTTTACAA; U SP 2 8 - 3 , G G AG TG AG AT TG A AC A AG A ; U SP 2 8 - 4 , GTGGCATGAAGATTATAGT; MYC, GATGAGGAAGAAATCGATG; FBW71, CAACAACGACGCCGAATTA; FBW7-2, ACAGGACAGTGTTTACAAA. A full-length mouse cDNA encoding USP28 was obtained from German Resource Center for Genome Research (RZPD), and subcloned into pCMVHA vector. Full-length human USP28 and USP25 cDNAs were Flag- and HA-tagged and subcloned into pDZ vector23. pUHD–MYC and pUb–His6–MYC–Ub vectors were described previously49. The Flag–FBW7 expression vectors were obtained from M. Welcker and B. Clurman (Fred Hutchinson Cancer Research Center, Seattle, WA) and have been published previously31. Immunoblotting, immunoprecipitation and immunostaining. Immunoblots and ubiquitination assays were performed as described49. The following reagents and antibodies were used: anti-MYC (9E10, N262), anti-CDK2 (M2), anti-cyclin A (H432), anti-SKP2 (H435) and anti-cyclin E (HE111) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Flag (M2) and anti-β-actin (AC-15) were from Sigma (St Louis, MO) ; anti-HA (12CA5) was from Roche (Basel, Switzerland). A polyclonal antiserum raised against human USP28 was used where indicated23. For immunoprecipitations, cells were lysed in buffer containing 20 mM Tris at pH 8.0, 1% Triton X-100, 250 mM NaCl and protease inhibitor cocktail (Roche). Lysates containing 500 μg total protein were precipitated with the indicated antibodies and protein G–sepharose beads for 2 h, and immunoprecipitates were analysed by immunoblotting. For immunostaining, transfected HeLa cells grown on coverslips were fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100, and blocked with 10% BSA in PBS for 30 min at 37%. Cells were incubated with the indicated antibodies in 5% BSA–PBS and slides were mounted with Moviol. Stability and ubiquitination assays. Cycloheximide and MG132 were purchased from Sigma and Calbiochem (Nottingham, UK), respectively. For MYC turnover analysis, cycloheximide was added to cell culture medium to 20 μM and cells were harvested at the indicated times. In vivo ubiquitination assays were performed as previously described49. For in vitro ubiquitination assays, HeLa cells were cotransfected with MYC and His–ubiquitin, and ubiquitinated MYC was recovered on Ni-NTA agarose beads. Flag–USP28 was immunoprecipitated from cells transfected with Flag–USP28 alone or with FBW7α. The beads were incubated with purified Flag–USP28 as previously described23. RT–PCR. Total RNA was isolated using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instruction. cDNA was synthesized with random hexanucleotides and MMLV reverse transcriptase. 25 cycles of PCR were carried out with the following primers: MYC, 5'-CCTACCCTCTCAACGACAGC and 5'CTCTGACCTTTTGCCAGGAG; USP28, 5'- GGAACAGCAGCAAGATGTGA and 5'GGCCGAAGGTCTCATTGTTA; FBW7, 5'CAGCAGTCACAGGCAAATGT and 5'- GCATCTCGAGAACCGCTAAC. Human tissues. In this study, tumour tissue of 10 adenocarcinomas of the colon (five moderately differentiated and five poorly differentiated cases), one liver metastasis of a colon carcinoma and ten cases of invasive ductal carcinomas of the breast (one case well differentiated, seven cases moderately differentiated, two cases poorly differentiated) was analysed. Normal colon tissue distant from the tumours was also included. Paraffin blocks previously used for diagnostic purpose were taken from the files of the Institute of Pathology of the University of Marburg (Marburg, Germany). The tissues, after removal, were fixed with 10% formalin, and specimens were embedded in paraffin. Immunohistochemistry. For immunohistochemistry, 3–4 μm thick paraffin sections were mounted on poly-l-lysine-coated slides, incubated at 58 °C and deparaffinized. For heat-induced antigen retrieval, sections were incubated in

773

A RT I C L E S DakoCytomation Target Retrieval Solution at pH 6.1 (DakoCytomation, Hamburg, Germany) for 30 min in a household steamer. The following incubations were performed on an automated immunohistochemistry apparatus (Autostainer plus; DakoCytomation), using a labelled streptavidin–biotin immunostaining (Dako REAL Detection System Peroxidase/DAB+, Rabbit/Mouse; DakoCytomation), which includes blocking of endogenous peroxidase and finally the staining reaction based on 3,3'-diaminobenzidine (DAB). For mild counterstaining, Mayer’s haematoxylin solution was used. For negative controls, the primary antibody was replaced by buffer or an irrelevant monoclonal antibody. Accession number. The microarray data summarized in the Supplementary Information, Table S3 can be accessed at http://www.ebi.ac.uk/arrayexpress/. The accession number is E-MEXP-1070. Note: Supplementary Information is available on the Nature Cell Biology website. ACKNOWLEDGEMENTS This study was supported by grants from the European Union (through the FP6 Integrated Project INTACT) to M.E. and R.B., the Deutsche Forschungsgemeinschaft (Forschergruppe Chromatin and Transregio17) to M.E. and the Netherlands Genomics Initiative/Netherlands Organisation for Scientific Research (NWO) to R.B. This work was supported by grants from the National Institutes of Health (NIH) and Cooperative Center for Medical Countermeasures Against Radiation (CMCR) to S.J.E. S.J.E. is a Howard Hughes Medical Institute Investigator. We thank D. Dobrin, B. Jebavy and R. Baumann for expert technical assistance, and M. Welcker and B. Clurman for FBW7 expression plasmids. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. Published online at http://www.nature.com/naturecellbiology/ Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions/ 1. Oster, S. K., Ho, C. S., Soucie, E. L. & Penn, L. Z. The myc oncogene: MarvelouslY Complex. Adv. Cancer Res. 84, 81–154 (2002). 2. Salghetti, S. E., Kim, S. Y. & Tansey, W. P. Destruction of MYC by ubiquitin-mediated proteolysis: cancer-associated and transforming mutations stabilize MYC. EMBO J. 18, 717–726 (1999). 3. Bahram, F., von der Lehr, N., Cetinkaya, C. & Larsson, L. G. c-MYC hot spot mutations in lymphomas result in inefficient ubiquitination and decreased proteasome-mediated turnover. Blood 95, 2104–2110 (2000). 4. Welcker, M. et al. The FBW7 tumor suppressor regulates glycogen synthase kinase 3 phosphorylation-dependent c-MYC protein degradation. Proc. Natl Acad. Sci. USA 101, 9085–9090 (2004). 5. Yada, M. et al. Phosphorylation-dependent degradation of c-MYC is mediated by the F-box protein FBW7. EMBO J. 23, 2116–2125 (2004). 6. Yeh, E. et al. A signalling pathway controlling c-MYC degradation that impacts oncogenic transformation of human cells. Nature Cell Biol. 6, 308–318 (2004). 7. Gregory, M. A. & Hann, S. R. c-MYC proteolysis by the ubiquitin-proteasome pathway: stabilization of c-MYC in Burkitt’s lymphoma cells. Mol. Cell Biol. 20, 2423–2435 (2000). 8. Malempati, S. et al. Aberrant stabilization of c-MYC protein in some lymphoblastic leukemias. Leukemia 20, 1572–1581 (2006). 9. Kim, S. Y., Herbst, A., Tworkowski, K. A., Salghetti, S. E. & Tansey, W. P. Skp2 regulates myc protein stability and activity. Mol. Cell 11, 1177–1188 (2003). 10. von der Lehr, N. et al. The F-box protein Skp2 participates in c-MYC proteosomal degradation and acts as a cofactor for c-MYC-regulated transcription. Mol. Cell 11, 1189–1200 (2003). 11. Gross-Mesilaty, S. et al. Basal and human papillomavirus E6 oncoprotein-induced degradation of MYC proteins by the ubiquitin pathway. Proc. Natl Acad. Sci. USA 95, 8058–8063 (1998). 12. Nijman, S. M. et al. A genomic and functional inventory of deubiquitinating enzymes. Cell 123, 773–786 (2005). 13. Evan, G. I. et al. Induction of apoptosis in fibroblasts by c-myc protein. Cell 69, 119–128 (1992). 14. Littlewood, T. D., Hancock, D. C., Danielian, P. S., Parker, M. G. & Evan, G. I. A modified oestrogen receptor ligand binding domain as an improved switch for the regulation of heterologous proteins. Nucleic Acids Res. 23, 1686–1690 (1995). 15. Berns, K. et al. A large-scale RNAi screen in human cells identifies new components of the p53 pathway. Nature 428, 431–437 (2004).

774

16. Brummelkamp, T. R. et al. An shRNA barcode screen provides insight into cancer cell vulnerability to MDM2 inhibitors. Nature Chem. Biol. 2, 202–206 (2006). 17. Amati, B., Littlewood, T. D., Evan, G. I. & Land, H. The c-MYC protein induces cell cycle progression and apoptosis through dimerization with Max. EMBO J. 13, 5083–5087 (1993). 18. Leone, G. et al. MYC requires distinct E2F activities to induce S phase and apoptosis. Mol. Cell 8, 105–113 (2001). 19. Dansen, T. B., Whitfield, J., Rostker, F., Brown-Swigart, L. & Evan, G. I. Specific requirement for Bax, not Bak, in MYC-induced apoptosis and tumor suppression in vivo. J. Biol. Chem. 281, 10890–10895 (2006). 20. Rothermund, K. et al. c-MYC-independent restoration of multiple phenotypes by two cMYC target genes with overlapping functions. Cancer Res. 65, 2097–2107 (2005). 21. Benitah, S. A., Frye, M., Glogauer, M. & Watt, F. M. Stem cell depletion through epidermal deletion of Rac1. Science 309, 933–935 (2005). 22. Park, Y. B. et al. Alterations in the INK4a/ARF locus and their effects on the growth of human osteosarcoma cell lines. Cancer Genet. Cytogenet. 133, 105–111 (2002). 23. Zhang, D., Zaugg, K., Mak, T. W. & Elledge, S. J. A role for the deubiquitinating enzyme USP28 in control of the DNA-damage response. Cell 126, 529–542 (2006). 24. Koepp, D. M. et al. Phosphorylation-dependent ubiquitination of cyclin E by the SCFFBW7 ubiquitin ligase. Science 294, 173–177 (2001). 25. Strohmaier, H. et al. Human F-box protein hCdc4 targets cyclin E for proteolysis and is mutated in a breast cancer cell line. Nature 413, 316–322 (2001). 26. Bhattacharya, S. et al. SKP2 associates with p130 and accelerates p130 ubiquitylation and degradation in human cells. Oncogene 22, 2443–2451 (2003). 27. Sutterlüty, H. et al. p45skp2 promotes p27kip1 degradation and induces S phase in quiescent cells. Nature Cell Biol. 1, 207–214 (1999). 28. Sorensen, C. S. et al. A conserved cyclin-binding domain determines functional interplay between anaphase-promoting complex–Cdh1 and cyclin A–Cdk2 during cell cycle progression. Mol. Cell Biol. 21, 3692–3703 (2001). 29. Bashir, T., Dorrello, N. V., Amador, V., Guardavaccaro, D. & Pagano, M. Control of the SCF(Skp2–Cks1) ubiquitin ligase by the APC/C(Cdh1) ubiquitin ligase. Nature 428, 190–193 (2004). 30. Welcker, M. et al. Multisite phosphorylation by Cdk2 and GSK3 controls cyclin E degradation. Mol. Cell 12, 381–392 (2003). 31. Welcker, M., Orian, A., Grim, J. A., Eisenman, R. N. & Clurman, B. E. A Nucleolar isoform of the FBW7 ubiquitin ligase regulates c-MYC and cell size. Curr. Biol. 14, 1852–1857 (2004). 32. Kee, Y., Lyon, N. & Huibregtse, J. M. The Rsp5 ubiquitin ligase is coupled to and antagonized by the Ubp2 deubiquitinating enzyme. EMBO J. 24, 2414–2424 (2005). 33. Li, M., Brooks, C. L., Kon, N. & Gu, W. A dynamic role of HAUSP in the p53–Mdm2 pathway. Mol. Cell 13, 879–886 (2004). 34. van Drogen, F. et al. Ubiquitylation of cyclin E requires the sequential function of SCF complexes containing distinct hCdc4 isoforms. Mol. Cell 23, 37–48 (2006). 35. Gomez-Roman, N., Grandori, C., Eisenman, R. N. & White, R. J. Direct activation of RNA polymerase III transcription by c-MYC. Nature 421, 290–294 (2003). 36. Grandori, C. et al. c-MYC binds to human ribosomal DNA and stimulates transcription of rRNA genes by RNA polymerase I. Nature Cell Biol. 7, 311–318 (2005). 37. Sansom, O. J. et al. MYC deletion rescues Apc deficiency in the small intestine. Nature 446, 676–679 (2007). 38. van de Wetering, M. et al. The β-catenin/TCF-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 111, 241–250 (2002). 39. Oskarsson, T. & Trumpp, A. The MYC trilogy: lord of RNA polymerases. Nature Cell Biol. 7, 215–217 (2005). 40. Arabi, A., Rustum, C., Hallberg, E. & Wright, A. P. Accumulation of c-MYC and proteasomes at the nucleoli of cells containing elevated c-MYC protein levels. J. Cell Sci. 116, 1707–1717 (2003). 41. Li, Z., Wang, D., Messing, E. M. & Wu, G. VHL protein-interacting deubiquitinating enzyme 2 deubiquitinates and stabilizes HIF-1α. EMBO Rep. 6, 373–378 (2005). 42. Herbst, A. et al. A conserved element in MYC that negatively regulates its proapoptotic activity. EMBO Rep. 6, 177–183 (2005). 43. Hemann, M. T. et al. Evasion of the p53 tumour surveillance network by tumour-derived MYC mutants. Nature 436, 807–811 (2005). 44. Benassi, B. et al. c-MYC phosphorylation is required for cellular response to oxidative stress. Mol. Cell 21, 509–519 (2006). 45. Ngo, V. N. et al. A loss-of-function RNA interference screen for molecular targets in cancer. Nature 441, 106–110 (2006). 46. Shachaf, C. M. et al. MYC inactivation uncovers pluripotent differentiation and tumour dormancy in hepatocellular cancer. Nature 431, 1112–1117 (2004). 47. Leung-Toung, R. et al. Thiol proteases: inhibitors and potential therapeutic targets. Curr. Med. Chem. 13, 547–581 (2006). 48. Brummelkamp, T. R., Bernards, R. & Agami, R. Stable suppression of tumorigenicity by virus-mediated RNA interference. Cancer Cell 2, 243–247 (2002). 49. Adhikary, S. et al. The ubiquitin ligase HectH9 regulates transcriptional activation by MYC and is essential for tumor cell proliferation. Cell 123, 409–421 (2005).


S U P P L E M E N TA R Y I N F O R M AT I O N

a.

b. Experimental design

100

50

25

50

10

2

10

% cells remaining (+ 4-OHT)

10

% cells remaining (+ 4-OHT)

5

EtOH

0

5

6

10

4-OHT

-5 -10 00

Isolate RNA, PCR, Array Hybridization

c.

EtOH

5

10

15

0

d.

EtOH

10

15

A=2log( Cy5xCy3)

A=2log( Cy5xCy3)

4-OHT

5

4-OHT

1 2 3 4 5 6 1 2 3 4 5 6 days 64

p53

Vector

50

16

dnp53

6 36

64

Cdk2

p53

50

P21CIP1

26

p21

36

Cdk2

dnp53

S16

Figure S1 Characterization of the Bar-code Screen. (a) Design of the barcode screen. U2OS-MycER cells were infected with the NKI shRNA library, serum-starved and treated with 4-OHT until less than 10% of the infected cells were still alive. Reference samples were harvested before adding 4OHT. shRNA cassettes were recovered by PCR on genomic DNA and the relative abundance was analyzed on bar-code microarrays. (b) Representative results of hybridization experiments performed at 50 and at 90% cell death. (c,d) Myc-induced apoptosis is independent of p53 in U20S cells. (c) U2OS cells expressing retroviral MycER were either treated with 4-OHT or EtOH.

The upper panel shows immunoblots documenting expression of p53 and as a loading control Cdk2 at the indicated time points after treatment. The lower panel shows RT-PCR assays documenting expression of p21CIP1 and as a control S16 at the indicated time points. (d) U2OS cells expressing retroviral MycER were infected with viruses expressing a dominant-negative allele of p53 or a control vector and subsequently selected. The left panels show colony formation assays of these cells six days following treatment with either 4-OHT or EtOH. The right panels show immunoblots documenting the expression of dnp53, endogenous p53, p21, and Cdk2.

WWW.NATURE.COM/NATURECELLBIOLOGY

1

© 2007 Nature Publishing Group


a

Control

HA-Usp28

Control + MG132

-Myc

DAPI

-HA

b

HA-FBW7 FLAG-USP28 + + +

+ + + +

130 100

-HA

72 50

-Flag

130 100 130 100

-HA

Input

72

-Flag IP

50

Figure S2 Characterization of Usp28 function and interaction with Fbw7. (a) Expression of Usp28 leads to accumulation of Myc in the nucleus. The individual panels show immune-fluorescence pictures documenting staining with α-Myc antibodies (upper row), DAPI (lower row) and α-HA antibodies (bottom panel) of representative cells. The left panels show controltransfected cells, middle panels show cells transfected with HA-Usp28 and the right panels show control-transfected cells treated with MG132 for 4 hours. Equal exposures are shown for each panel. Scale bars correspond to

20 μM. (b) Binary complexes of Usp28 with different alleles of Fbw7. HeLa cells were transfected with HA-tagged Usp28 and the indicated alleles of Flag-tagged Fbw7. Lysates were immunoprecipitated with α-HA antibodies and probed with α-Flag antibodies. The FBW7 alleles used in these experiments have been described previously31; BD1 refers to a basic domain in Fbw7 involved in nuclear localization. FBW7Nα-γ carries the aminoterminus of Fbw7α grafted on Fbw7γ.

2




b.

a.

FLAG-FBW7 FLAG-SKP2 HA-USP28 -HA

-FLAG

+ - + + +

MYC - + + + HA-USP28 - - - + FLAG-FBW7 - - + +

+ - + + +

-Usp28

130

FLAG-Fbw7

100 72 50

-Myc -FLAG

FLAG-Skp2

43

FLAG-SKP2

130

130

50

50

100 34

-Cdk2

Input -FLAG IP

- + + + - - - + - - + +

43 34

d.

c.

N N T T T T T

RPS16

CSRP2

-Usp28

PIM1

PLCB4

-Myc

72 50

POLA2

SNTB2

-tubulin

50

CDC2

H1F0

PPAP2B

130

COP1

ENPP1

Figure S3 Characterization of Usp28 function and expression. (a) Usp28 does not bind to Skp2. The experiment was performed as described in panel (a), except that Flag-tagged Fbw7 and Flag-tagged Skp2 were used in parallel. (b) Usp28 specifically antagonizes Fbw7-mediated degradation. The experiment was performed as described in Figure 3, panel (b), using expression plasmids for either Fbw7α (left panel) or Skp2 (right panel) in parallel. (c) Examples of genes targeted by both shUsp28 and shMyc. The

experiment was performed as described in Figure 5c; RT-PCR was performed with specific primers for the indicated genes. RPS16 was used as control. (d) Expression of Usp28 in normal colon and in colon carcinomas. The upper panel shows an immunoblot of independent samples of normal colon (N) and of colon carcinomas (T) documenting enhanced expression of Usp28 in colon carcinomas relative to normal colon. The lower panel shows an αtubulin immunoblot that was used as control for equal loading.


3



Figure S4 Full-length Scans of all western blots depicted in the individual figures. The individual panels represent the full scans of the western blots shown in the individual panels.

4




Supplementary Table 1 List of genes targeted by shRNAs that were recovered three times or more in the six independent experiments and for which more than one shRNA species was found. „N hits“ refers to the number of times that vectors targeting the gene were found in the six screens; „Number of shRNAs“ refers to the number of shRNA species found for this gene (a maximum of 3). Supplementary Table 2 List of shRNAs validated by our own experiments and those linked to Myc function and/or apoptosis by previously published work. The column “Validation” refers to experiments such as the ones shown in Figure 1b and 1c. The indicated shRNA vectors were individually re-cloned and tested for inhibition of Myc-induced apoptosis as measured by colony formation assays and by the measurement of DNA fragmentation. Supplementary Table 3 Summary of Microarray Data obtained from Ls174T cells. Shown is the list of genes that are regulated by more than 1.5 fold either in response to shUsp28 or to shMyc relative to shScr cells. The file also shows the Chi2-square test mentioned in the text.


5


Supplementary Table 1: Summary of Barcode Screens

N hits Number shRNAs 13 3 9 3 6 3 6 2 5 3 5 3 5 2 5 2 4 3 4 3 4 3 4 3 4 3 4 3 4 3 4 4 4 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

3 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

Acc. No. AF176039 NM_01506 NM_002467 NM_004324 X59869 NM_005343 NM_005186 NM_024426 AB014599 NM_004350 NM_003743 NM_000926 NM_000293 NM_005045 NM_001754 X52078 NM_012111 AF116508 NM_002908 NM_014570 NM_001799 X95325 U43188 NM_000880 NM_005693 NM_002541 NM_001708 NM_000369 NM_152891 NM_005359 NM_005375 NM005060 NM_005225 NM_134260 NM_145259 NM_007371 NM_000562 BM806537 NM_004449 NM_012304 NM_005551 NM_138980 NM_022443 NM_002577

Gene ID HMGA1 TRIM33 MYC BAX TCF7 HRAS CAPN1 WT1 BICD2 RUNX3 NCOA1 PGR PHKB RELN RUNX1

Description high mobility group AT-hook 1 tripartite motif-containing 33 Transcript v-myc myelocytomatosis viral oncogene homolog BCL2-associated X protein transcription factor 7 (T-cell specific, HMG-box) v-Ha-ras Harvey rat sarcoma viral oncogene homolog calpain 1, (mu/I) large subunit Wilms tumor 1 coiled-coil protein BICD2 runt-related transcription factor 3 nuclear receptor coactivator 1 progesterone receptor phosphorylase kinase, beta reelin runt-related transcription factor 1 (aml1 oncogene) transcription factor 3 (E2A immunoglobulin enhancer binding TCF3 factors E12/E47) AHSA1 AHA1, activator of heat shock 90kDa protein ATPase homolog 1 ETV7 ets variant gene 7 (TEL2 oncogene ) REL v-rel reticuloendotheliosis viral oncogene homolog ARFGAP3 ADP-ribosylation factor GTPase activating protein 3 CDK7 cyclin-dependent kinase 7 (, cdk-activating kinase) CSDA cold shock domain protein A ELF2 E74-like factor 2 (ets domain transcription factor) IL7 interleukin 7 NR1H3 nuclear receptor subfamily 1, group H, member 3 OGDH oxoglutarate (alpha-ketoglutarate) dehydrogenase (lipoamide) OPN1SW opsin 1 (cone pigments), short-wave-sensitive TSHR thyroid stimulating hormone receptor EOS serine protease EOS MADH4 mothers against decapentaplegic homolog 4 MYB v-myb myeloblastosis viral oncogene homolog RORC RAR-related orphan receptor C E2F1 E2F transcription factor 1 RORA RAR-related orphan receptor A ALK anaplastic lymphoma kinase (Ki-1) BRD3 bromodomain containing 3 C8A complement component 8, alpha polypeptide EPIM epimorphin ERG v-ets erythroblastosis virus E26 oncogene like FBXL7 F-box and leucine-rich repeat protein 7 KLK2 kallikrein 2 MAPK10 mitogen-activated protein kinase 10 MLF1 myeloid leukemia factor 1 PAK2 p21 (CDKN1A)-activated kinase 2


N hits Number shRNAs Acc. No. 3 3 NM_000335 3 3 NM_006142

Gene ID SCN5A SFN

Description sodium channel, voltage-gated, type V, alpha stratifin

3

3

NM_000372

TYR

3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

NM_005709 NM_004330 NM_147780 NM_013421 NM_006003 NM_000025 NM_001182 NM_004281 NM_014299 NM_003775 NM_000125 NM_012300 NM_005479 NM_005316 NM_002097 NM_002106 NM_003883 NM_000523 NM_012218 NM_002229 NM_002275 NM_005356 NM_002343 M64240 NM_006983 D50692 NM_014397 NM_002545 NM_030905 NM_002604 NM_002657 NM_000936 NM_002860 NM_004794

USH1C BNIP2 CTSB GGT1 UQCRFS1 ADRB3 ALDH7A1 BAG3 BRD4 EDG6 ESR1 FBXW1B FRAT1 GTF2H1 GTF3A H2AFZ HDAC3 HOXD13 ILF3 JUNB KRT15 LCK LTF MAX MMP23B MYCBP NEK6 OPCML OR2J2 PDE7A PLAGL2 PNLIP PYCS RAB33A

tyrosinase (oculocutaneous albinism IA)

3 3 3 3 3 3 3

2 2 2 2 2 2 2

M64749 NM_000538 NM_000326 NM_000330 NM_006931 NM_005131 NM_016936

RDC1 RFXAP RLBP1 RS1 SLC2A3 THOC1 UBN1

G protein-coupled receptor regulatory factor X-associated protein retinaldehyde binding protein 1 retinoschisis (X-linked, juvenile) 1 solute carrier family 2 (facilitated glucose transporter), member 3 THO complex 1 ubinuclein 1

3 3 3

2 2 2

NM_017481 NM_020886 NM_004724

UBQLN3 USP28 ZW10

ubiquilin 3 ubiquitin specific protease 28 ZW10 homolog, centromere/kinetochore protein

Usher syndrome 1C (autosomal recessive, severe) BCL2/adenovirus E1B 19kDa interacting protein 2 cathepsin B gamma-glutamyltransferase 1 ubiquinol-cytochrome c reductase adrenergic, beta-3-, receptor aldehyde dehydrogenase 7 family, member A1 BCL2-associated athanogene 3 bromodomain containing 4 endothelial differentiation, G-protein-coupled receptor 6 estrogen receptor 1 F-box and WD-40 domain protein 1B frequently rearranged in advanced T-cell lymphomas general transcription factor IIH, polypeptide 1 general transcription factor IIIA H2A histone family, member Z histone deacetylase 3 homeo box D13 interleukin enhancer binding factor 3 jun B proto-oncogene keratin 15 lymphocyte-specific protein tyrosine kinase lactotransferrin MAX protein matrix metalloproteinase 23B c-myc binding protein NIMA (never in mitosis gene a)-related kinase 6 opioid binding protein/cell adhesion molecule-like olfactory receptor, family 2, subfamily J, member 2 phosphodiesterase 7A pleiomorphic adenoma gene-like 2 pancreatic lipase pyrroline-5-carboxylate synthetase RAB33A, member RAS oncogene family


Supplementary Table 2: Genes validated in secondary screens and/or previously implicated in Myc function

(a) Genes that were identified 3 times or more and represented by two or more independent shRNAs Gene ID BAX

N 2

Val. 2/2

E2F1

3

1/1

ELF2 HMGA1

2 3

2/2 1/1

H2AFZ HRAS MAX MYC MYCBP PAK2 USP28 THOC1 TRIM33 TSHR PYKS ADRB3 ALDH7 BRD3 CDK7

2 3 2 3 2 3 2 2 3 2 2 2 2 2 2

2/2 3/3 1/1 2/2 2/2 1/1 2/2 2/0 2/0 2/0 2/2 1/2

Proposed function Required for Myc Induced apoptosis Required for Myc Induced apoptosis

Reference Dansen et al., J Biol Chem. 2006; 281(16):10890-5 Leone et al.2001. Mol. Cell 8: 105-113.

Target of Myc

Rothermund et al., Cancer Res. 2005; 65(6):2097-107

Co-factor of Myc

Amati et al., EMBO J. 1993; 12(13):5083-7

Co-activator of Myc Phosphorylates Myc Antagonist of Fbw7

Taira et al., Genes Cells 1998; 3(8):549-65 Benitah et al., Science 2005: 309(5736):933-5 This communication

(b) Genes that were identified once by a single shRNA Gene ID HIST1c PML1 WRN SEI-1 PMS2 P53INP1 RAD9 PRMT1 SUMO SOX STKB17B

N 1 1 1 1 1 1 1 1 1 1 1

Val. 1/1 1/0 1/0 1/0 1/0 1/0 1/0 1/0 1/0 1/0 1/0

“N” denotes the number of different shRNAs against each gene that were found in the screen. „Val“ refers to the validation experiments (colony assays or apoptosis assays): the first number refers to the number of different shRNAs that were tested and the second number to the number of shRNAs that were positive in these assays.


Supplementary Table 3: Summary of Microarray Data ChiSquare Test

shMyc down shMyc up

shUsp28 down shUsp28 up 166 24 15 51

X2= 98,7801 p