p53-induced protein with a death domain (PIDD) isoforms ... - Nature

Oncogene (2008) 27, 387–396

& 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

p53-induced protein with a death domain (PIDD) isoforms differentially activate nuclear factor-kappaB and caspase-2 in response to genotoxic stress S Cuenin, A Tinel1, S Janssens2 and J Tschopp Department of Biochemistry, University of Lausanne, Epalinges, Switzerland

Cells respond to DNA damage in a complex way and the fate of damaged cells depends on the balance between proand antiapoptotic signals. This is of crucial importance in cancer as genotoxic stress is implied both in oncogenesis and in classical tumor therapies. p53-induced protein with a death domain (PIDD), initially described as a p53inducible gene, is one of the molecular switches able to activate a survival or apoptotic program. Two isoforms of PIDD, PIDD (isoform 1) and LRDD (isoform 2), have already been reported and we describe here a third isoform. These three isoforms are differentially expressed in tissues and cell lines. Genotoxic stress only affects PIDD isoform 3 mRNA levels, whereas isoforms 1 and 2 mRNA levels remain unchanged. All isoforms are capable of activating nuclear factor-kappaB in response to genotoxic stress, but only isoform 1 interacts with RIPassociated ICH-1/CED-3 homologous protein with a death domain and activates caspase-2. Isoform 2 counteracts the pro-apoptotic function of isoform 1, whereas isoform 3 enhances it. Thus, the differential splicing of PIDD mRNA leads to the formation of at least three proteins with antagonizing/agonizing functions, thereby regulating cell fate in response to DNA damage. Oncogene (2008) 27, 387–396; doi:10.1038/sj.onc.1210635; published online 16 July 2007 Keywords: PIDD; DNA repair; NF-kB; caspase-2

Introduction Maintenance of genetic integrity is a recurrent challenge for cells, both during cell replication and also when cells are exposed to DNA-damaging agents, such as ultraviolet (UV) radiation, ionizing radiation (IR), endogenous reactive oxygen species or chemotherapeutic Correspondence: Professor J Tschopp, Department of Biochemistry, University of Lausanne, Chemin des Boveresses 155, CH-1066 Epalinges, Switzerland. E-mail: [email protected] 1 Current address: Department of Pathology, Harvard Medical School, 77 Av Louis Pasteur, Boston, MA 02115, USA. 2 Current address: Department of Molecular Biomedical Research, University of Ghent-VIB, Technologiepark 927, 9052 Gent, Belgium. Received 15 February 2007; revised 24 May 2007; accepted 29 May 2007; published online 16 July 2007

agents. The accumulation of DNA damage can induce permanent changes leading to oncogenic transformation or severely impaired cellular functioning. As soon as the damage is detected, cells engage a cascade of reactions whose outcome determines cell fate. Activation of nuclear factor-kappaB (NF-kB) and cell cycle arrest in response to DNA damage are thought to be generally beneficial for the cell, allowing it some time for repair (Janssens and Tschopp, 2006). When the genotoxic insult is too severe, however, the cell will die. Although it seems obvious that the balance between pro- and antiapoptotic signals dictates cell fate, how the decision for cell survival or cell death is taken remains largely unclear. However, some proteins have been identified that can act as switches between these two cell fates. One of these proteins is p53-induced protein with a death domain (PIDD), which was initially identified by Lin et al. (2000) as a p53-inducible gene using differential display with DDP16.1 erythroleukemia cells expressing a temperature-sensitive p53ts mutant. Their experiments suggested that PIDD is an effector of p53-induced apoptosis, as it induces G1 phase cell-cycle arrest and cell death when overexpressed. Further characterization of PIDD indicated that it could act as a switch between cell survival and cell death in response to DNA damage. Our group has shown that PIDD is able to interact with receptor interacting protein 1 (RIP1) and NF-kB essential modifier (NEMO) leading to NEMO sumoylation and NF-kB activation in response to DNA damage (Janssens et al., 2005). We also demonstrated that PIDD is able to form a caspase2-activating platform by recruiting the adapter molecule RIP-associated ICH-1/CED-3 homologous protein with a death domain (RAIDD) and thus sensitize cells to genotoxic stress (Tinel and Tschopp, 2004). Different types of DNA damage exist that lead to the generation of various DNA alterations, such as cyclobutane pyrimidine dimers, oxidized bases, singlestrand breaks or double-strand breaks (DSBs) (reviewed in Sancar et al., 2004; Roos and Kaina, 2006). The latter, which can be induced by topoisomerase inhibitors as well as IR, are extremely toxic and can lead to chromosomal aberrations and genetic instability (van Gent et al., 2001). Activation of NF-kB in response to DSB depends on two parallel signaling cascades, namely those orchestrated by the PI3 kinase

PIDD isoforms display different activities S Cuenin et al

388

like ataxia-telangectasia mutated (ATM), which is activated by DSB or higher order chromatin changes (Bakkenist and Kastan, 2003) and those regulated by PIDD, which is activated by an as yet unknown signal (Janssens et al., 2005). In response to DSB, NEMO (also known as IKKg) is sumoylated on K277 and K309 by PIASg in a PIDD-dependent fashion, and SUMONEMO accumulates in the nucleus (Huang et al., 2003; Janssens et al., 2005; Mabb et al., 2006). Activated ATM phosphorylates SUMO-NEMO to allow its desumoylation and subsequent monoubiquitination on the same lysine residues (Huang et al., 2003). NEMO then shuttles back to the cytoplasm where it can associate with IKKa and IKKb to form an active IKK complex that will lead to NF-kB activation and cell survival. Alternatively, catastrophic DNA damage can also lead to cell death. Experiments to date indicate that mitochondria are the major mediators of apoptosis in response to DNA damage. Several genes upregulated by p53 target the mitochondria to induce the release of cytochrome c, Smac/DIABLO and other mediators of apoptosis. These include Bad, Bax, PUMA and Noxa (for a recent review, see Roos and Kaina, 2006). Furthermore, caspase-2 has been shown to be activated by PIDD and to act upstream of the mitochondria by promoting mitochondrial membrane permeabilization (Guo et al., 2002; Lassus et al., 2002; Robertson et al., 2002). Caspase-2 can act indirectly on the mitochondria by cleaving and activating Bid (Guo et al., 2002) or by directly targeting the mitochondria, independently of its enzymatic activity (Robertson et al., 2002). In this article, we identified several isoforms of PIDD. To understand how alternative splicing affects PIDD function, we characterized the interacting partners of three isoforms of PIDD and compared the functional outcome of these interactions. We could show that all the three isoforms are able to activate NF-kB in response to genotoxic stress, whereas only isoform 1 (PIDD) can activate caspase-2 and sensitize cells to genotoxic stress. Results PIDD isoforms description and expression pattern PIDD is a 910 amino-acid protein composed of seven N terminal leucin-rich repeats (LRRs), two ZU5 domains (present in ZO-1 and Unc5-like netrin receptors) and a C terminal death domain (DD). It is encoded by the LRDD gene on chromosome 11 at 11p15.5 and comprises 16 exons. A search of the National Center for Biotechnology Information (NCBI) database currently retrieves nine entries for PIDD complete transcripts and 148 expressed-sequence-tags (ESTs). Careful alignment and analysis of these sequences allowed the identification of five potential isoforms (Figure 1a and Supplementary Figure S1). Alignment of the isoforms reveals that isoform 2 (leucine repeat death domain containing protein (LRDD) (Telliez et al., 2000)) has a 146 aminoacid deletion at the N terminus and an additional 11 amino-acid deletion at position aa 580. Isoform 3 has Oncogene

a 17 amino-acid deletion at position aa 705 compared to isoform 1. Isoform 4 lacks the LRRs. Isoform 5 also lacks the LRRs and, in addition, it has a unique C terminus resulting from alternative splicing at the boundary of exons 9 and 10 leading to a frameshift; it thus lacks the DD in addition to the LRRs. We focused on isoforms 1, 2 and 3 for further study. To test whether these three isoforms are indeed expressed in tissues or cell lines, we developed a polymerase chain reaction (PCR) strategy to specifically amplify each isoform, taking advantage of the deletions in isoforms 2 and 3 (Supplementary Figure S2). Isoforms 1 and 3 are almost ubiquitously expressed, at various levels, with the notable exception of skeletal muscle, where no PIDD expression could be detected (Figure 1b), confirming the observations of Lin et al. (2000). In contrast, isoform 2 is only detected in liver, pancreas and leukocytes. The PIDD-expression profile in cell lines differs from that in tissues (Figure 1c), with cell lines tested expressing isoforms 1 and 2, whereas only HCT116 cells (colon carcinoma cells) express isoform 3. Interestingly, lung and kidney tumor samples exhibit the same expression profile as cell lines (Figure 1d). Regulation of PIDD expression PIDD was initially described as a p53-inducible gene (Lin et al., 2000). We observed that PIDD protein is indeed induced after DNA damage, but dependence on p53 varies with the stimulus (Figure 2). Treatment with the topoisomerase II inhibitor etoposide (VP16) or UV upregulates PIDD levels independently of the p53 status of the cells (Figures 2a and b), whereas the topoisomerase II inhibitor and DNA-intercalating agent doxorubicin requires the presence of p53 to upregulate PIDD levels (Figure 2c). As our anti-PIDD antibody directed against the DD is unable to discriminate between the three isoforms, we investigated the mRNA levels of each isoform. Upon treatment with etoposide, isoforms 1 and 2 mRNA levels remained constant, whereas isoform 3 mRNA was increased (Figures 3a and b). This induction of isoform 3 was also observed when treating the cells with tumor necrosis factor a (TNFa), but not with Brefeldin A, an endoplasmic reticulum (ER) stressinducing compound (data not shown). As observed at the protein level, the p53 status of the cell had no effect on the regulation of the isoforms mRNA level by etoposide or UV. When the cells were treated with doxorubicin, the mRNA levels of isoforms 1 and 2 remained constant as well, but isoform 3 mRNA levels decreased (Figure 3c). This effect was more pronounced in the absence of p53. PIDD isoforms 1, 2 and 3 activate NF-kB We have recently shown that PIDD isoform 1 is autoprocessed at S446 and S588 to generate a PIDD-C and a PIDD-CC fragment (Tinel et al., 2007), in agreement with previous reports (Telliez et al., 2000; Pick et al., 2006). The 11 amino-acid deletion in isoform 2 encompasses the second cleavage site used for PIDD


389 Isoform 3

LRR

ZU5

ZU5

Isoform 1

LRR

ZU5

ZU5

ZU5

ZU5

ZU5

ZU5

ZU5

ZU5

LRR

Isoform 2 (LRDD)

Deletion: 705-721

DD

DD

Deletion: 580-590

DD

147

Isoform 4

DD

314

Pos ctrl

H2O

Leukocytes

Colon

Small intestine

Ovary

Testis

Thymus

Spleen

584

Pancreas

Kidney

Skeletal muscle

Liver

Lung

Placenta

Brain

Heart

314

Prostate

Isoform 5

Isof 1 Isof 2

Isof 3

Pos ctrl

H2O

Lung adenocarcinoma

Kidney carcinoid tumor

Pos ctrl

H2O

HCT116 p53-/-

HCT116 p53+/+

THP-1

Jurkat

Raji

SH-EP

LAN-1

HeLa

293T

GAPDH

Isof 1

Isof 1

Isof 2

Isof 2

Isof 3

Isof 3

GAPDH

GAPDH

Figure 1 PIDD isoforms. (a) Schematic view of the five isoforms of PIDD. Positions of the deletions between isoform 1 and isoforms 2 (LRDD) and 3 are indicated. The alternative C terminus of isoform 5 is indicated as a dashed line. (b) Expression of PIDD isoforms in different tissues. The tissue cDNA library was subjected to a PCR using the primers described in Materials and methods to specifically amplify isoforms 1, 2 and 3. Amplification of GAPDH was used as a control for the quality of the cDNA. (c, d) Expression of PIDD isoforms in different cell lines. As in (b), using cDNA generated from the indicated cell lines (c) or tumor samples (d). DD, death domain; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; LRDD, leucin repeat DD; LRR, leucine-rich repeat; PCR, polymerase chain reaction.

Oncogene


390 p53 +/+

p53 -/p53 +/+

Etoposide (h) C

8 24 kDa 48

CC

33

WB: αPIDD

0

4

8 24

0

4

UV (h) WB: αPIDD

25

WB: αRAIDD

CC

33 48 33 25

25

17

17 83

WB: αp53

48

WB: αPARP

p53 +/+ Doxorubicin (h) WB: αPIDD

0 3

kDa 48

33

WB: αPARP

NT 6 24

C

WB: αCasp2

48

WB: αCasp2

NT 6 24

p53 -/-

6 12 24 0 3

83

p53 -/6 12 24

kDa

C

48

CC

33

WB: αRAIDD

25

WB: αCasp2

48

WB: αPARP 83 WB: αp53

48

Figure 2 Expression of PIDD protein in response to genotoxic stress. (a, b) PIDD protein levels increase independently of p53 in response to etoposide (a) or UV (b). HCT116 cells p53 þ / þ or p53/ were treated with 10 mM etoposide or with UV 10 mJ/cm2 and allowed to recover for up to 24 h and cell extracts were analysed by western blotting. PIDD-C and PIDD-CC fragments are indicated with an arrowhead. (c) PIDD protein level upregulation is dependent on p53 in response to doxorubicin. HCT116 cells p53 þ / þ or p53/ were treated with 10 mg/ml doxorubicin for the indicated time and cell lysates were subjected to western blot analysis. UV, ultraviolet.

autoprocessing and indeed, although all the three isoforms are present as full length and PIDD-C, only isoforms 1 and 3 generate a PIDD-CC fragment (Figure 4a, western blot anti-Flag). The two C terminal fragments generated by PIDD autoprocessing have distinct functions (Tinel et al., 2007). PIDD-C is able to interact with RIP1 and NEMO to activate NF-kB in response to DNA damage (Janssens et al., 2005). In contrast, PIDD-CC interacts with RAIDD and caspase-2 to form the so-called PIDDosome (Tinel and Tschopp, 2004), a molecular platform that induces caspase-2 activation and cleavage. We first investigated the ability of the three isoforms to activate NF-kB. Correlating with their ability to generate a PIDD-C fragment, all the three isoforms Oncogene

interacted with RIP1 and NEMO (Figure 4a). The overexpression of each isoform alone was sufficient to induce a subtle NF-kB activation (Figure 4b). Moreover, treatment of human embryonic kidney (HEK) 293T cells that stably express PIDD isoforms with etoposide resulted in increased NF-kB, as assessed by a luciferase assay (Figure 4c). In the course of a response to DNA damage, NEMO is sequentially sumoylated, phosphorylated and ubiquitinated to activate NF-kB (Huang et al., 2003; Hay, 2004). PIDD isoform 1 is essential for NEMO sumoylation and its overexpression increases SUMO-NEMO levels (Janssens et al., 2005). Figure 4d shows that all the PIDD isoforms when overexpressed in HEK293T cells enhanced NEMO sumoylation in response to etoposide treatment. The


391

0

4

8

p53 -/24

0

4

8

H2O

p53 +/+

24

+ Isof 1

Isof 2

Isof 3

GAPDH

NT

6h

p53 -/-

24h

NT

6h

24h

H2O

p53 +/+

+ Isof 1

Isof 2

Isof 3

GAPDH

HCT116 p53 +/+ 3

6

12

24

0

3

6

12

24

H2O

0

HCT116 p53 -/+ Isof 1 Isof 2 Isof 3 GAPDH

Figure 3 Regulation of expression of PIDD isoforms. (a, b) PIDD isoform 3 mRNA is increased in response to etoposide (a) or UV (b), independently of p53. RNA was extracted from the cells used in Figures 2a or b, respectively. After reverse transcription, cDNA was amplified by PCR to detect PIDD isoforms 1, 2 and 3. GAPDH was used as a control for cDNA quality. (c) PIDD isoform 3 is downregulated in response to doxorubicin. Same as in (a, b), but with cells from Figure 2c. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; PCR, polymerase chain reaction; UV, ultraviolet.

effect of isoform 1 on NEMO sumoylation was weaker than that of the other two isoforms, correlating with the NF-kB luciferase assay. PIDD isoforms 2 and 3 are unable to activate caspase-2 We next tested the ability of the three PIDD isoforms to form an active PIDDosome (Figures 5a and b). The PIDDosome spontaneously forms in hypotonic lysates activated at 301C (Tinel and Tschopp, 2004). We therefore activated in vitro hypotonic lysates of HEK293T cells stably overexpressing each isoform and monitored PIDDosome formation by caspase-2 cleavage

and RAIDD recruitment to PIDD. As expected, isoform 2, which does not generate a PIDD-CC fragment, is unable to interact with RAIDD and does not induce more caspase-2 activation than that observed in mock-transfected cells. Surprisingly, isoform 3 expression was also unable to activate caspase-2, although it gave rise to a PIDD-CC fragment. This incapacity to activate caspase-2 correlates, however, with the inability to recruit RAIDD. As overexpression of PIDD sensitizes cells to genotoxic stress-induced cell death (Tinel and Tschopp, 2004), we assessed the sensitization to doxorubicininduced cell death conferred by each isoform. In good Oncogene


392 IP: αFlag mock 0 10 30

Isof1

lysates

Isof2

Isof3

mock

0 10 30 0 10 30 0 10 30

Isof1

Isof2

Isof3

0 10 30 0 10 30 0 10 30 0 10 30

C

48

CC

33

WB: αFlag

WB: αRIP1

48

WB: αNEMO

48

30 no treatment

6

eto

4

Fold induction

Fold induction

25

2

20 15 10

0 mock

Isof 1

Isof 2

Isof 3

5 0 Isof 1

mock mock

WB : αsumo

T

Isof 2 E -

T

Isof 1 E

-

T

Isof 2

Isof 3

Isof 3 E

-

T

E NEMO-sumo

WB : αNEMO

WB : αFlag

PIDD-FL PIDD-C PIDD-CC

Figure 4 All the three isoforms are able to activate NF-kB. (a) PIDD isoforms 1, 2 and 3 interact with RIP1 and NEMO. Cells stably expressing the different isoforms were lysed in a hypotonic buffer and the lysates were activated at 301C for the indicated times. PIDD was then immunoprecipitated through its Flag-tag and co-immunoprecipitating proteins were analysed by western blotting. PIDD-C and PIDD-CC fragments, as well as NEMO (lower band), are indicated with an arrowhead. (b) Overexpression of PIDD isoforms 1, 2 and 3 activates NF-kB. HEK293T cells were transfected with either a mock or one of the PIDD isoform expression plasmids along with luciferase reporter plasmids and an NF-kB luciferase assay was performed. (c) Stable expression of all the three PIDD isoforms increases NF-kB activation in response to genotoxic stress. An NF-kB luciferase assay was performed on cells stably expressing PIDD isoforms following treatment with or without 40 mM etoposide for 16 h. (d) Increased DNA damage-induced NEMO sumoylation in cells stably expressing one of the PIDD isoforms. HEK293T cells stably expressing the different PIDD isoforms were treated with TNFa (T) or etoposide (E) for 16 h and cell lysates were analysed by western blotting. HEK, human embryonic kidney; IP, immunoprecipitation; NEMO, NF-kB essential modifier; NF-kB, nuclear factor-kappaB; RIP1, receptor interacting protein 1; TNFa, tumor necrosis factor a.

Figure 5 PIDD isoforms 2 and 3 do not activate caspase-2. (a) Activation of caspase-2 by PIDD isoforms. Cells stably expressing the different isoforms were lysed in a hypotonic buffer and the lysates were incubated at 301C for the indicated times. Lysates were then analysed by western blotting. PIDD-C and PIDD-CC fragments are indicated with an arrowhead. (b) Interaction with RAIDD. PIDD was immunoprecipitated from lysates of HEK293T cells stably expressing the different isoforms through its Flag-tag and coimmunoprecipitating RAIDD was analysed by western blotting. (c) Sensitization to genotoxic stress. HeLa cells stably expressing the three isoforms were treated with 10mg/ml doxorubicin for the indicated time. Cell death was assayed by Trypan blue staining and cell lysates were analysed by western blotting for caspase-3-dependent PARP cleavage, caspase-2, RAIDD and Flag-PIDD expression. (d) Effect of co-expression of isoforms 2 and 3 on isoform 1. HeLa cells stably expressing PIDD isoform 1 were transfected with isoforms 2, 3 or Bcl-2 and their sensitivity to genotoxic stress was assessed by Trypan blue staining after doxorubicin treatment as in (c). HEK, human embryonic kidney; PARP, poly(ADP-ribose) polymerase. Oncogene


393

correlation with the immunoprecipitation data, only isoform 1 sensitized cells to genotoxic stress-induced cell death, as measured by Trypan blue staining (Figure 5c). We then tested the ability of isoforms 2 and 3 to protect cells from isoform 1-induced sensitiIsof 3

IP Flag

10 30 0 10 30 0 10 30 0 10 30

kDa

FL

83

C

62 48

CC

WB: αRAIDD

WB: αRAIDD

25

WB: αCasp2

pro-Casp2

kDa 25

WB: αFlag

33

Isof 3

Isof 2

Isof 2

0

Isof 1

Isof 1

WB: αFlag

mock

mock

Time (min) at 30°C

zation to genotoxic stress (Figure 5d). Isoform 2 was able to protect cells, albeit to a lesser extent than Bcl-2. Isoform 3, on the contrary, enhanced the sensitization to genotoxic stress when co-expressed with isoform 1.

C

83 62 48

CC

33

FL

33 25 p19

80 0 3 5 5 + zVAD

70

% cell death

60 50 40 30 20 10 0 mock Doxo (h)

0

3

5

Isof 1 5z 0

3

5

Isof 2 5z 0

3

Isof 3

5 5z 0

3 5 5z

kDa

WB: αPARP 83 WB: αCasp2

pro-Casp2

p19 WB: αRAIDD

17 83

WB: αPIDD

62

C

48

CC

% cell death

33 90 80 70 60 50 40 30 20 10 0

0

3

isof1/mock

4

0

3

4

isof1/isof2

0

3 isof1/isof3

4

0

3

4

isof1/Bcl2

Time doxo treatment (h)

Oncogene


394

Discussion The decision that will determine cell fate after DNA damage is a crucial one, especially in the context of an organism. Indeed, cells dying too easily or, on the other hand, cells too resistant to potentially mutagenic DNA damage would be detrimental to the organism. There is thus a need for strict regulation of survival or death signals. Some proteins have the potential to act as switches between these two antagonistic cell fates. PIDD is one such protein, as it is both able to mediate NF-kB induction and survival or to activate caspase-2 and cell death in response to DNA damage. Autoprocessing and interacting partners of PIDD are key determinants in the outcome of the signaling cascade engaged in response to DNA damage. PIDD isoforms can influence cell survival in different ways. Although all the three isoforms can activate NF-kB, only isoform 1 is able to induce caspase-2 cleavage and to sensitize cells to genotoxic stress. The fact that PIDD isoform 2 is expressed in all cell lines tested, is absent from most normal tissues and protects from the deadly action of PIDD isoform 1 indicates that it may confer a survival advantage. The opposite might apply to isoform 3. Whether isoforms 2 and 3 could be used as transformation markers remain an interesting open question. Cells treated with different DNA-damaging agents, such as UV and the topoisomerase inhibitor etoposide, upregulate PIDD at the protein level in a p53-independent manner. Analysis of the mRNA levels to discriminate between the three isoforms reveals that only isoform 3 mRNA levels increase in response to these genotoxic stresses, which would sensitize cells to PIDD isoform 1 pro-apoptotic function. Treatment with doxorubicin, which is not only an inhibitor of topoisomerase II but also a DNA-intercalating agent, leads to a p53-dependent upregulation of PIDD at the protein level. The mRNA analysis shows a downregulation of isoform 3 mRNA levels, without affecting isoform 1 or 2. Thus, the upregulation at the protein level could be due to the stabilization of the protein by p53. However, arguing against this hypothesis, treating the cells with the protein synthesis inhibitor cycloheximide (CHX) resulted in the abrogation of doxorubicin-induced PIDD protein upregulation (data not shown). The fact that doxorubicin treatment decreases PIDD isoform 3 mRNA levels, whereas TNFa has the opposite effect indicates that PIDD expression could be, in part, NF-kB dependent. Indeed, doxorubicin has been shown to repress some NF-kB-driven genes by assembling a histone deacetylase-dependent transcriptional repressor complex (Campbell et al., 2004, 2006). It has also been observed that doxorubicin leads to a deficiency in phosphorylation and acetylation of p65, resulting in unstable binding of NF-kB to endogenous loci (Ho et al., 2005). The loss of p53 accentuates the decrease in isoform 3 mRNA levels, indicating that both p53 and (at least) another transcription factor, maybe NF-kB, control PIDD expression. PIDD isoform 1 is the only isoform able to interact with RAIDD and to activate caspase-2, whereas all the Oncogene

three isoforms activate NF-kB. This indicates that an intact intermediate domain (ID) is required for caspase2 activation but not for NF-kB activation. The functional properties of isoform 2 can easily be explained by the deletion of the second cleavage site: it retains the ability to generate a NF-kB-inducing PIDDC fragment, but loses that of generating a caspase-2activating PIDD-CC fragment. However, the inability of isoform 3 to activate caspase-2 was unexpected and highlights a crucial role of the ID for PIDD-induced caspase-2 activation. The processing of isoform 3 is slightly less efficient than that of isoform 1 leading to less PIDD-CC generation, but this cannot account for the loss of caspase-2 activation. Indeed, smaller deletions, even as small as two amino acids in the region deleted in isoform 3, are enough to completely abrogate the caspase-2-activating ability of PIDD, although these constructs are efficiently processed (data not shown). Given the ability of PIDD to act as a potential switch between cell survival and cell death in response to DNA damage, its regulation is of great importance. Two levels of regulation had already been reported: at the transcriptional level, as PIDD is a known p53 target gene (Lin et al., 2000), and at the post-translational level, as autoprocessing of PIDD determines its partners of interaction and, therefore, the downstream signaling events (Tinel et al., 2007). The alternative splicing of PIDD that we report here could represent a third level of regulation, acting at the post-transcriptional level. What controls the preferential splicing into one or the other isoform remains to be elucidated. Interestingly, a recent article reported alternative splicing mediated by ATM and ataxia telangiectasia and Rad3-related protein (ATR) on TATA-binding protein (TBP)-associated factor I (TAF-1), a component of the general transcription factor transcription-factor IID (TFIID), in response to DNA damage in Drosophila (Katzenberger et al., 2006). Understanding how the splicing machinery is affected by DNA damage will probably help to elucidate the alternative splicing of PIDD in response to a genotoxic stimulus.

Materials and methods Cell culture and biological reagents HEK293T cells and HeLa cells were grown in Dulbecco’s Modified Eagle’s medium and Glutamax (Life Technologies, Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal calf serum (FCS). HCT116 p53 þ / þ and p53/ cells (gift of Dr B Vogelstein, John Hopkins University, Baltimore, MD, USA) were grown in Eagle’s minimum essential medium (EMEM) (Cambrex Bio Science, Walkersville, MD, USA) supplemented with 10% FCS. All cells were grown at 371C with 5% CO2. Etoposide and human recombinant TNFa were purchased from Alexis (Lausen, Switzerland). Doxorubicin and BrefeldinA were obtained from Sigma (St Louis, MO, USA). Expression vectors pCR3-PIDDisof1-Flag and pMSCV-PIDD-Flag have been described previously (Tinel and Tschopp, 2004). PIDDisof2


395 and PIDDisof3 were generated by double PCR on PIDDisof1. PIDDisof2 corresponds to EST IMAGE 3908102. PIDDisof3 corresponds to NM_145887. All PIDD constructs are C terminally Flag-tagged. NFconluc, containing the luciferase reporter gene driven by a minimal NF-kB responsive promoter, was a gift of Dr A Israel (Institut Pasteur, Paris, France). pactbgal, containing the b-galactosidase gene under the control of the b-actin promotor was obtained from Dr Inoue (Institute of Medical Sciences, Tokyo, Japan). PCR isoforms PCR to amplify PIDD isoforms were performed using the following primer pairs: GCA CGC TTC CAG GTC ACA CAC and GGT CCA GAG TGG TGG TCA CGT for isoform 1; GAG CTC ACC CAC CTG TAC TGG and GGT CCA GAG TGG TGG TCA CGT for isoform 2; GAA GAA TGT GAA GGA GGT GTC and CAA GAC TGC GCG CAC CTC TTC AG for isoform 3. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a control and amplified with primers GAA ATC CCA TCA CCA TCT TCC AGG and CGC GGC CAT CAC GCC ACA GTT TCC. PCR for isoforms 1 and 2 and GAPDH had 25 rounds of amplification, versus 30 rounds for isoform 3. Betaine (1 M) (Sigma) was added to the PCR mix for isoform 3. Taq DNA polymerase was from New England Biolabs (Ipswich, MA, USA) and dNTP from Invitrogen. The Human MTC panel I and II tissue cDNA libraries and human kidney and lung tumors total RNA were purchased from Clontech/Takara Bio (Palo Alto, CA, USA). Database searches Online BLAST searches, EST mining and protein sequence retrieval were performed using the NCBI (http://www. ncbi.nlm.nih.gov/) and the ExPASy proteomics tools server (www.expasy.org). Alignments were performed using MacVector v.8.1.1. Exon–intron architecture of the gene product was found in Ensembl (www.ensembl.org), except for Ensembl-described exon 6 that was absent from EST sequences. PIDD Swissprot accession number is Q9HB75. The following NCBI nucleotide sequence accession numbers were obtained: AB208949 (contains part of intron 6–7 and full intron 9–10), AL833849 (contains part of intron 2–3, full intron 10–11 and full intron 14–15), AB208832 (contains part of intron 2–3, full intron 10–11 and full intron 14–15), AB209529 (contains part of intron 2–3 and full intron 14–15), AK074893 (isoform 5), AF465246 (isoform 4), AF229178 (isoform 2), BC014904 (isoform 3) and AF274972 (isoform 1). PIDD induction HCT116 p53 þ / þ or p53/ cells were treated with etoposide, UV (Stratalinker, Stratagene, La Jolla, CA, USA), TNFa, BrefeldinA or doxorubicin. A proportion of each sample was lysed in E1A lysis buffer (1% NP-40, 20 mM N-2hydroxyethylpiperazine-N0 -2-ethanesulfonic acid (HEPES) (pH 7.9), 250 mM NaCl, 20 mM b-glycerophosphate, 10 mM NaF, 1 mM sodium orthovanadate, 2 mM dithiothreitol (DTT), 1 mM ethylene diamine tetraacetic acid (EDTA) and a protease-inhibitor cocktail) and subjected to western blot analysis. The remainder of each sample was treated with Trizol to extract RNA for reverse transcription (RT)–PCR and PCR analysis. RT–PCR RNA was extracted with Trizol (Invitrogen) according to the manufacturer’s protocol, and 10 mg was subjected to a T-primed RT–PCR reaction (Ready-to-go, Amersham,

Piscataway, NJ, USA). Quality of the cDNA was assessed by a PCR to amplify GAPDH. Generation of stable cell lines HEK293T cells were transfected with vectors encoding retroviral structural genes along with PIDD isoforms in the pMSCV plasmid (Clontech, Palo Alto, CA, USA) to produce VSV-G pseudotyped retroviruses. HeLa or HEK293T cells were then infected with the retroviruses and selected with puromycin (0.5 mg/ml; Sigma). In vitro activation, immunoprecipitation and western blotting Preparation of the lysates in hypotonic buffer and in vitro activation and immunoprecipitation were performed, as described previously (Tinel and Tschopp, 2004). Briefly, washed cells were resuspended in hypotonic buffer (20 mM HEPES–KOH, 10 mM KCl, 1 mM MgCl2, 1 mM ethylene glycol tetraacetic acid (EDTA), 1 mM EGTA, 1 mM DTT (pH 7.5)) supplemented with protease inhibitors (Complete, Roche, Basel, Switzerland) and mechanically lysed through a 22 G needle. Lysates were incubated at 301C for the indicated times or left untreated on ice. After incubation, an equivalent volume of hypotonic buffer containing 0.1% NP-40 and 300 mM NaCl was added to perform overnight immunoprecipitation with M2-Flag beads (Sigma). Total lysates and immunoprecipitates were then separated by sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS–PAGE), transferred to nitrocellulose and analysed by immunoblotting. Antibodies used were anti-Flag M2 (Sigma), rabbit antiRAIDD (Apotech, Epalinges, Switzerland), anti-caspase-2 (clone 11B4, Apotech), anti-RIP1 (BD Transduction Laboratories, San Jose, CA, USA), rabbit anti-NEMO (Santa Cruz), mouse anti-NEMO (BD Transduction Laboratories), antiSUMO-1 (Zymed, Invitrogen), anti-poly(ADP-ribose) polymerase PARP (Cell Signaling, Danvers, MA, USA), anti-tubulin (Sigma), anti-phospho-IkB (Cell Signaling), anti-PIDD Anto-1 (Apotech) and anti-p53 DO-1 (gift from Dr R Iggo). Transient transfection and NF-kB reporter gene assay NF-kB luciferase assays were performed, as previously described (Janssens et al., 2005). Luciferase values were normalized for differences in transfection efficiency on the basis of b-galactosidase activity in the same extracts and expressed as fold-induction values relative to the unstimulated empty-vector control. Where indicated, cells were stimulated 48 h post-transfection for 16 h with 40 mM etoposide or left untreated. Cell death assay HeLa cells stably expressing the different PIDD isoforms were treated with 10 mg/ml doxorubicin for the indicated time, and cell death was assayed using Trypan blue. To study the effect of co-expressing different isoforms, stable HeLa cells expressing isoform 1 were transiently transfected using lipofectamine 2000 (Invitrogen) following the manufacturer’s instructions. Cells were treated with doxorubicin 48 h post-transfection and cell death was assessed by Trypan blue exclusion. Acknowledgements We thank Saskia Lippens, Marie-Cećile Michallet, Virginie Pe´trilli and Helen Everett for comments and critical reading of the manuscript and all members of the laboratory for helpful discussions and comments. SJ is the recipient of an EMBO long-term fellowship. Oncogene


396 References Bakkenist CJ, Kastan MB. (2003). DNA damage activates ATM through intermolecular autophosphorylation and dimer dissociation. Nature 421: 499–506. Campbell KJ, O’Shea JM, Perkins ND. (2006). Differential regulation of NF-kappaB activation and function by topoisomerase II inhibitors. BMC Cancer 6: 101. Campbell KJ, Rocha S, Perkins ND. (2004). Active repression of antiapoptotic gene expression by RelA(p65) NF-kappa B. Mol Cell 13: 853–865. Guo Y, Srinivasula SM, Druilhe A, Fernandes-Alnemri T, Alnemri ES. (2002). Caspase-2 induces apoptosis by releasing proapoptotic proteins from mitochondria. J Biol Chem 277: 13430–13437. Hay RT. (2004). Modifying NEMO. Nat Cell Biol 6: 89–91. Ho WC, Dickson KM, Barker PA. (2005). Nuclear factor-kappaB induced by doxorubicin is deficient in phosphorylation and acetylation and represses nuclear factor-kappaB-dependent transcription in cancer cells. Cancer Res 65: 4273–4281. Huang TT, Wuerzberger-Davis SM, Wu ZH, Miyamoto S. (2003). Sequential modification of NEMO/IKKgamma by SUMO-1 and ubiquitin mediates NF-kappaB activation by genotoxic stress. Cell 115: 565–576. Janssens S, Tinel A, Lippens S, Tschopp J. (2005). PIDD mediates NF-kappaB activation in response to DNA damage. Cell 123: 1079–1092. Janssens S, Tschopp J. (2006). Signals from within: the DNA-damageinduced NF-kappaB response. Cell Death Differ 13: 773–784. Katzenberger RJ, Marengo MS, Wassarman DA. (2006). ATM and ATR pathways signal alternative splicing of Drosophila TAF1 premRNA in response to DNA damage. Mol Cell Biol 26: 9256–9267. Lassus P, Opitz-Araya X, Lazebnik Y. (2002). Requirement for caspase-2 in stress-induced apoptosis before mitochondrial permeabilization. Science 297: 1352–1354.

Lin Y, Ma W, Benchimol S. (2000). Pidd, a new death-domaincontaining protein, is induced by p53 and promotes apoptosis. Nat Genet 26: 122–127. Mabb AM, Wuerzberger-Davis SM, Miyamoto S. (2006). PIASy mediates NEMO sumoylation and NF-kappaB activation in response to genotoxic stress. Nat Cell Biol 8: 986–993. Pick R, Badura S, Bosser S, Zornig M. (2006). Upon intracellular processing, the C-terminal death domain-containing fragment of the p53-inducible PIDD/LRDD protein translocates to the nucleoli and interacts with nucleolin. Biochem Biophys Res Commun 349: 1329–1338. Robertson JD, Enoksson M, Suomela M, Zhivotovsky B, Orrenius S. (2002). Caspase-2 acts upstream of mitochondria to promote cytochrome c release during etoposide-induced apoptosis. J Biol Chem 277: 29803–29809. Roos WP, Kaina B. (2006). DNA damage-induced cell death by apoptosis. Trends Mol Med 12: 440–450. Sancar A, Lindsey-Boltz LA, Unsal-Kacmaz K, Linn S. (2004). Molecular mechanisms of mammalian DNA repair and the DNA damage checkpoints. Annu Rev Biochem 73: 39–85. Telliez JB, Bean KM, Lin LL. (2000). LRDD, a novel leucine rich repeat and death domain containing protein. Biochim Biophys Acta 1478: 280–288. Tinel A, Janssens S, Lippens S, Cuenin S, Logette E, Jaccard B et al. (2007). Autoproteolysis of PIDD marks the bifurcation between pro-death caspase-2 and pro-survival NF-kappaB pathway. EMBO J 26: 197–208. Tinel A, Tschopp J. (2004). The PIDDosome, a protein complex implicated in activation of caspase-2 in response to genotoxic stress. Science 304: 843–846. van Gent DC, Hoeijmakers JH, Kanaar R. (2001). Chromosomal stability and the DNA double-stranded break connection. Nat Rev Genet 2: 196–206.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene