Role of Pirh2 in Mediating the Regulation of p53 and c-Myc

Ubiquitylation is fundamental for the regulation of the stability and function of p53 and c-Myc. The E3 ligase Pirh2 has been reported to polyubiquitylate p53 and to mediate its proteasomal degradation. Here, using Pirh2 deficient mice, we report that Pirh2 is important for the in vivo regulation of p53 stability in response to DNA damage. We also demonstrate that c-Myc is a novel interacting protein for Pirh2 and that Pirh2 mediates its polyubiquitylation and proteolysis. Pirh2 mutant mice display elevated levels of c-Myc and are predisposed for plasma cell hyperplasia and tumorigenesis. Consistent with the role p53 plays in suppressing c-Myc-induced oncogenesis, its deficiency exacerbates tumorigenesis of Pirh2−/− mice. We also report that low expression of human PIRH2 in lung, ovarian, and breast cancers correlates with decreased patients' survival. Collectively, our data reveal the in vivo roles of Pirh2 in the regulation of p53 and c-Myc stability and support its role as a tumor suppressor.


Role of Pirh2 in Mediating the Regulation of p53 and c-Myc
Introduction Protein ubiquitylation is essential for a broad spectrum of cellular processes including nuclear export, endocytosis, transcriptional regulation, DNA damage repair and proteasomal degradation [1]. Impairment of the ubiquitylation process has been associated with a variety of human diseases including autoimmunity, immunodeficiency, inflammation and cancer [2,3].
One well established role for ubiquitylation in cancer is its function in regulating the stability and function of the tumor suppressor p53 and the oncogene c-MYC, two proteins with major roles in human malignancies [2][3][4]. Monoubiquitylation of p53 serves to signal its nuclear export as well as its mitochondrial translocation, while p53 polyubiquitylation targets it for proteasomal degradation. Remarkably, p53 is targeted for ubiquitylation by several ubiquitin E3 ligases including Pirh2 (Rchy1), Mdm2/ Hdm2, Cop1, E6/E6AP, ARF-BP1, Synoviolin and by atypical E3 ligases including E4F1 [5].
Similar to p53, ubiquitylation is critical for the regulation of the function and stability of c-MYC, an oncoprotein frequently overexpressed in various human cancers including breast and ovarian cancer [6,7]. The ubiquitin ligases FBW7, SKP2 and ARF-BP1/HectH9 have been shown to mediate c-MYC ubiquitylation. While c-MYC polyubiquitylation by FBW7 leads to its proteasomal degradation, its polyubiquitylation by ARF-BP1 increases its transcriptional activity. c-MYC ubiquitylation by SKP2 has been shown to mediate both its transactivation activity and proteasomal degradation [2,6]. Interestingly, some of the E3 ligases important for the regulation of c-Myc also regulate p53 function. SKP2 negatively regulates p53 by suppressing its acetylation by p300 while ARF-BP1 suppresses p53 through its polyubiquitylation and proteolysis [8,9]. Furthermore, FBW7 loss of function has been reported to attenuate p53 activity [10].
The identification of multiple E3 ligases that regulate the function and stability of p53 and c-MYC and the possible involvement of some of these E3 ligases in the regulation of both p53 and c-Myc have raised questions regarding the physiological functions of these E3 ligases. Despite the identification of Pirh2 as an E3 ligase that polyubiquitylates p53 in cell culture assays, its in vivo functions remain unknown. Here we report that mice deficient for Pirh2 are viable but display elevated levels of p53 and apoptosis in response to DNA damage. We also demonstrate that c-Myc is a novel Pirh2 interacting protein and that Pirh2 regulates c-Myc expression levels by mediating its polyubiquitylation and proteolysis. Consistent with this novel Pirh2 function, mice mutant for Pirh2 show elevated levels of c-Myc and increased risk for plasma cell hyperplasia, gammaglobulinemia, and tumorigenesis. In accordance with the role of p53 in suppressing c-Myc oncogenesis, its inactivation considerably increases spontaneous cancer susceptibility of Pirh2 deficient mice. We also report that the expression of PIRH2 is reduced in various human cancers and that lower levels of PIRH2 expression correlate with decreased survival of patients with lung, breast or ovarian cancer.
Collectively, these data support a role for Pirh2 in the in vivo regulation of p53 functions, demonstrate its negative regulation of the oncoprotein c-Myc, and support its role as a novel tumor suppressor.

Pirh2 Is Dispensable for Development
Although p53 is a substrate for ubiquitylation by several E3ligases, to date only Mdm2 is known to be required for p53 regulation in vivo as its deficiency leads to embryonic lethality in a p53-dependent manner [11][12][13]. To characterize the physiological functions of Pirh2, an E3 ligase previously reported to polyubiquitylate p53 and regulate its stability [14], we generated Pirh2 mutant mice ( Figure S1A and S1B). Western blot analysis indicated loss of Pirh2 protein in Pirh2 2/2 cells ( Figure S1C). By contrast to Mdm2 deficiency [11][12][13], loss of Pirh2 did not affect embryonic development. Pirh2 2/2 mice were born in a Mendelian ratio, were fertile and did not present any apparent developmental defects compared to their wildtype (Wt) littermates. Analysis of total cell numbers as well as the different cell subpopulations in thymus, spleen, lymph nodes (LN) and bone marrow (BM) of 6 to 8 weekold Pirh2 2/2 mice revealed no significant differences compared to Wt littermates ( Figure S1D and S1E). In addition, [ 3 H] thymidine incorporation assay indicated that loss of Pirh2 had no significant effect on the in vitro proliferation of T-and B-cells from 6 to 8 week-old Pirh2 2/2 mice ( Figure S1F). These data indicate that Pirh2 is dispensable for embryonic and postnatal development.
Pirh2 Regulates p53 Turnover in Response to DNA Damage p53 is essential for several cellular processes including apoptosis, cell cycle, senescence and metabolism and its stability is regulated by various proteins including the E3 ligases Mdm2, Cop1, and ARF-BP1 [5,15]. To determine the role Pirh2 plays in the in vivo regulation of p53, we examined p53 levels and functions in Pirh2 deficient mice and cells. Splenocytes and thymocytes from Pirh2 2/2 and Wt mice were irradiated ex-vivo, and the levels of p53 expression and activation were determined by Western Blot analysis. Untreated Pirh2 2/2 cells displayed mildly increased expression of p53 compared to Wt controls ( Figure 1A). However, untreated Pirh2 2/2 cells, similar to Wt controls, displayed no detectable expression of the p53 transcriptional targets p21, bax, and Puma ( Figure 1A). We examined the effect of Pirh2 deficiency on irradiation (IR) induced p53 expression and have observed that irradiation of Pirh2 2/2 splenocytes resulted in higher expression levels of p53 and its targets p21, bax, and Puma compared to Wt controls ( Figure 1A and 1B). Consitent with the reported increased transcription of pirh2 in response to p53 activation in mouse embryonic fibroblasts (MEFs) [14], Pirh2 mRNA was found to increase by 1.5 fold at 1 to 4 h post irradiation of splenocytes ( Figure S2A). To examine the in vivo effects of Pirh2 loss on the expression of p53, Pirh2 2/2 mice and their Wt littermates were subjected to whole-body irradiation (6 Gc) and their tissues were collected at different time points postirradiation (0-12 h). Immunohistochemistry (IHC) examination of p53 expression indicated that while p53 was undetectable in spleen, thymus, intestinal crypts and liver from untreated Pirh2 2/2 and Wt mice, the frequency of cells expressing p53 in these organs was significantly higher post-IR of Pirh2 2/2 mice compared to Wt controls ( Figure 1C and 1D, Figure S2B-S2D). We next examined the effect of PIRH2 deficiency on P53 level using human RKO cells in which PIRH2 was knocked down by a tetracycline-inducible system [16]. Similar to murine cells, deficiency of PIRH2 in RKO cells only slightly increased P53 level under untreated conditions, but this increase was more pronounced post-IR ( Figure S2E).
Phosphorylation of p53 is critical for its stability and functions [17]. For instance, in response to DNA damage, p53 is phosphorylated on Serine 15 (p53-S15) by a number of kinases including ATM, DNA-PK and ATR and phosphorylation of this p53 site, together with others, is important for its stability and functions [17]. To further examine the effect of Pirh2 deficiency on p53 stability and function, we examined by Western blotting the level of p53-S15 in Pirh2 2/2 and Wt cells. While, the level of p53-S15 was similarly low in untreated Pirh2 2/2 and Wt splenocytes, it was significantly higher in irradiated Pirh2 2/2 cells compared to Wt controls ( Figure S3A). Similarly, IHC analysis indicated no differences in the levels of p53-S15 in spleen of untreated Pirh2 2/2 mice and Wt controls; however a marked increased expression of this phosphorylated form of p53 was observed post whole-body irradiation of Pirh2 2/2 mice compared to Wt controls ( Figure S3B).
We next examined the effect of PIRH2 expression on the level of S15-p53 using H1299 cells transiently transfected with expression vectors for MDM2 or PIRH2 and either wildtype p53, p53 S15A or p53 S15D. The presence of Alanine (A) at residue 15 of p53 prevents p53 phosphorylation at this site while the presence of Asparatic acid (D) at this site mimics its constitutive

Author Summary
Tumor suppressors and oncogenes play critical roles in cancer development. The tumor suppressor p53 and the oncogene c-MYC are among the most frequently deregulated genes in human cancer, and their ubiquitylation mediated by several E3 ligases is critical for their turnover and their functions. P53 has been shown to be ubiquitylated by Pirh2; however, the physiological significance of this modification of p53 remains unknown. In this study we have generated mice deficient for Pirh2 and have observed that loss of Pirh2 results in a higher level of p53 and cell death, especially in response to radiation. Remarkably, we also identified that Pirh2 interacts with c-Myc and mediates its polyubiquitylation and degradation. c-Myc accumulates in the absence of Pirh2 and this accumulation is accompanied by increased tumorigenesis of Pirh2-deficient mice. We also report that dual deficiency of Pirh2 and p53 synergizes cancer development. Examination of the expression level of PIRH2 in human cancers indicated that its lower expression level associates with poor survival of patients with lung, ovarian, or breast cancers. Collectively, these data indentify Pirh2 as a novel tumor suppressor involved in the regulation of both p53 and c-Myc.
phosphorylation. Western blot analysis indicated that overexpression of MDM2 or PIRH2 decreased the level of wildtype p53 ( Figure S3C). Consistent with previous data, overexpression of MDM2 was able to decrease the stability of p53 S15A and p53 S15D [17][18][19]. Interestingly, while overexpression of PIRH2 resulted in decreased expression of both forms of p53, its effect was more pronounced on p53 S15D. Collectively, these data support a role for Pirh2 in the regulation of p53.
Pirh2 Deficiency Leads to Elevated Irradiation-Induced Apoptosis p53 plays central roles in IR-induced apoptosis; therefore we have examined the effect of Pirh2 deficiency on the levels of apoptosis under untreated and DNA damage conditions. Splenocytes and thymocytes from Pirh2 2/2 and Wt mice were either untreated or irradiated (6 Gc) and their apoptotic levels examined using the AnnexinV/PI assay (Figure 2A). We also examined the level of apoptosis in thymus, spleen and intestinal crypts from whole-body irradiated (6 Gc) mice using TUNEL assay and cleaved caspase 3 IHC ( Figure 2B and 2C, Figure S3D and S3E). Our data indicated no increased level of spontaneous apoptosis in the absence of Pirh2; however consistent with the elevated IRinduced expression of p53 and its proapototic target genes Bax and Puma in the absence of Pirh2, IR-induced cell death was significantly elevated in Pirh2 2/2 cells and mice compared to Wt controls. These data demonstrate that Pirh2 deficiency results in elevated IR-induced apoptosis.

Pirh2 Interacts with the Oncoprotein c-Myc
In our search for novel Pirh2 interacting proteins we identified interaction of Pirh2 and c-Myc. We observed that human PIRH2 and c-MYC reciprocally co-immunoprecipitate upon overexpression in HEK293 cells ( Figure 3A). Endogenous interaction of murine Pirh2 and c-Myc was confirmed in NIH3T3 cells where immunoprecipitation of Pirh2 pulled down c-Myc and immunoprecipitation of c-Myc pulled down Pirh2 ( Figure 3B).
We next examined whether PIRH2 and c-MYC interaction was direct and characterized the domains involved in this interaction. We performed Glutathione S-transferase (GST) pull-down assays using purified GST or His-fusion proteins of the full length or fragments of PIRH2 and c-MYC. These pull-down assays indicated interaction of PIRH2 amino acids (aa) 138-261 with the N-terminus of c-MYC (aa 120-160) that contains MYCboxII (MBII) but not with N-terminus of c-MYC (aa 1-69) that contains MYCboxI (MBI) ( Figure 3C and 3E, Figure S4A). Notably, MBII (aa 128-143) is important for c-MYC proteasomal degradation, transcriptional repression and activation and for controlling c-MYC transformation ability [20]. Pull-down experiments using His-C-terminus of c-MYC (aa 353-434) demonstrated the interaction of this region of c-MYC with the N-terminus of PIRH2 (aa 1-137) ( Figure 3D and 3E). These data reveal c-MYC as a novel interacting protein for PIRH2 and indicate the importance of different domains of c-MYC and PIRH2 for this interaction.

Pirh2 Mediates c-Myc Polyubiquitylation
Polyubiquitylation of c-Myc is critical for the regulation of its stability and functions [2,6]. Our observation that c-MYC interacted with the E3 ligase PIRH2 raised the possibility that it might be a target for PIRH2 mediated ubiquitylation. We examined this hypothesis in HEK293T cells that overexpress HA-Ubiquitin (Ub), human c-MYC, and/or either human full length PIRH2 or a PIRH2 lacking the ring finger domain (PIRH2 DR ). 48 h post-transfection, immunoprecipitation followed by Western Blot analysis revealed high-molecular-weight c-MYC species where full length PIRH2, but not PIRH2 DR , were overexpressed in the presence of c-MYC and HA-Ub ( Figure 4A). Anti-HA immunoblot of c-MYC immunoprecipitates confirmed that the observed high-molecular-weight c-MYC species in the presence of the full length PIRH2 were indeed due to its polyubiquitylation. Our data also demonstrated the requirement for the RING finger domain of PIRH2 for its polyubiquitylation of c-MYC ( Figure 4B). Similarly, we examined the ability of PIRH2 to polyubiquitylate T58A c-MYC as phosphorylation of c-MYC at this site is critical for its polyubiquitylation by FBW7. Consitent with our finding that MBI that contains T58 is dispensible for the interaction c-MYC-PIRH2, PIRH2 was able to ubioquitylate T58A mutant c-MYC ( Figure S4B). We also performed in vitro ubiquitylation assays and observed that PIRH2 directly polyubiquitylates c-MYC ( Figure 4C and 4D).
We next examined the role of Pirh2 in the ubiquitylation of the endogenous c-Myc using Pirh2 2/2 and Wt B-cells. Anti-IgM activated B-cells were pre-treated for 3 h with the proteasome inhibitor MG132, and c-Myc immunoprecipitates were examined for their levels of ubiquitylation using anti-c-Myc and antiubiquitin antibodies. Pirh2 deficiency resulted in reduced level of both high-molecular-weight and polyubiquitylated forms of c-Myc ( Figure 4E and 4F). Taken together, these data identified c-Myc as a novel substrate for Pirh2-mediated polyubiquitylation.

Pirh2 Mediates c-Myc Proteolysis and Its Absence Leads to Accumulation of c-Myc Protein
Since c-Myc polyubiquitylation can control its turnover, we examined the steady state levels of c-Myc in Wt and Pirh2 2/2 MEFs in the presence of the protein biosynthesis inhibitor cycloheximide (CHX). These studies indicated that while less than 40% of c-Myc remained in Wt MEFs 1 h post-CHX treatment, near 80% of c-Myc remained in Pirh2 2/2 MEFs ( Figure 5A). Similarly, knockdown of human PIRH2 in RKO cells resulted in increased c-MYC stability ( Figure 5B).
Because Pirh2 interacted with c-Myc and mediated its polyubiquitylation and proteolysis, we examined the effect of its deficiency on the level of c-Myc protein. We first assessed by Western blotting the expression of c-Myc in B-cells from young (6-8 weeks old) Pirh2 2/2 mice and their Wt littermates. c-Myc protein level was significantly increased in naïve and in anti-IgM activated Pirh2 2/2 B-cells compared to Wt controls ( Figure 5C). Immunohistochemistry analysis of the expression level of c-Myc protein in spleen from young Pirh2 2/2 and Wt mice also showed increased c-Myc level in the absence of Pirh2 ( Figure 5D). Quantitative real time PCR analysis of cDNA from Pirh2 2/2 and Wt splenocytes indicated that despite the increased protein expression of c-Myc in Pirh2 2/2 cells, c-Myc mRNA abundance was not significantly affected compared to Wt controls ( Figure 5E). However, examination of c-Myc transcriptional targets indicated that the RNA expression level of E2f1, Ccna2, Ccnd2 and Brca1 was significantly increased while Mcl-1 RNA expression was repressed in accordance with the elevated levels of c-Myc in Pirh2 2/2 cells ( Figure 5E). Increased level of Ccnd2 protein was also observed in splenocytes from young Pirh2 2/2 mice compared to Wt littermates ( Figure 5F). Western blot examination of c-Myc expression in various tissues from young Pirh2 2/2 mice and Wt littermates indicated increased c-Myc level in the absence of Pirh2, with the highest increase observed in spleen, testis and mammary glands ( Figure 5F and 5G).
The tumor suppressor p19 ARF has been shown to be upregulated in response to high, but not to low, levels of deregulated c-Myc [21]. Using Western blotting we evaluated the level of p19 ARF in cells and tissues from young Pirh2 2/2 mice and their Wt littermates. The level of p19 ARF protein was found elevated in Pirh2 2/2 activated B-cells and in tissues that express high levels of c-Myc (e.g. spleen, testis, mammary glands) ( Figure 5C, 5G and 5J).
We next examined whether the loss of human PIRH2 would also lead to increased level of c-MYC protein. Western blot analysis indicated that similar to Pirh2 deficiency in mice, knock down of the human PIRH2 in RKO cells significantly increased the protein expression level of c-MYC ( Figure 5H). Conversely, overexpression of human PIRH2 in RKO cells downregulated their protein level of c-MYC ( Figure 5I).
Since Skp2 and Fbw7 mediate c-Myc polyubiquitylation and proteasomal degradation [2,6,22,23], we examined their expression levels in Pirh2 2/2 cells and Wt controls. Western blot analyses indicated no difference in the expression levels of Skp2 in spleen, liver and kidney from Pirh2 2/2 and Wt mice, while Fbw7 expression was barely detectable in these organs independently of the genotypes ( Figure 5J). We next sought to examine whether PIRH2-c-MYC complex may also contains SKP2 or FBW7. While Western blotting analysis of whole cell lysate of RKO cells showed that these cells expressed PIRH2, SKP2 and FBW7;  PIHR2 immunoprecipitation from these cells pulled down c-MYC, but failed to pull down SKP2 or FBW7 ( Figure S4C). Therefore SKP2 and FBW7 are not part of the PIRH2-c-MYC complex.
Collectively, these data demonstrate that Pirh2 deficiency leads to increased level of c-Myc protein and that Pirh2 regulates c-Myc stability.

PIRH2 Expression Is Downregulated in Human Cancer
PIRH2 was reported to be overexpressed in lung and prostate cancer [24,25]; however, to date no human pathologies have been associated with its decreased expression. To assess whether reduced Pirh2 expression associates with cancer, we examined PIRH2 expression in primary human cancers using microarray studies that associated mRNA levels to patient outcome. Median  dichotomization was used to identify patients with high-and low-PIRH2 expression levels. We found that lower levels of PIRH2 mRNA were associated with reduced patient survival in breast cancer [26,27] (Figure 6A), ovarian cancer [28,29] (Figure 6B) and squamous cell carcinomas of the lung [30,31] (Figure 6C).
The expression level of PIRH2 mRNA was also surveyed in microarray studies of human cancer using the Oncomine database [32]. Expression of PIRH2 mRNA was significantly reduced in ovarian clear cell adenocarcinoma [33], adult germ cell tumors [34] and invasive bladder cancer [35] compared to controls ( Figure S5). These data indicate the high prognostic significance of PIRH2 expression for patients with these tumor types.
Pirh2 Mutation Leads to Plasma Cell Hyperplasia, Gammaglobulinemia, Kidney Failure, and Premature Death Deregulated c-Myc has been associated with a number of diseases including plasma cell hyperplasia, gammaglobulinemia, kidney failure and cancer development [7,[36][37][38]. Monitoring of cohorts of Pirh2 mutant and Wt mice indicated that both Pirh2 2/2 and Pirh2 +/2 mice had reduced lifespan compared to Wt littermates ( Figure 7A). Necropsy of Pirh2 mutant mice indicated that these mutants developed a lymphoproliferative disorder characterized by plasma cell hyperplasia. 100% of sick Pirh2 mutants presented splenomegaly while lymphadenopathy was observed in 40% and 20% of sick Pirh2 2/2 and Pirh2 +/2 mice respectively ( Figure 7B and 7C). Hematoxylin and Eosin (H&E) staining of spleen and lymph nodes from sick Pirh2 mutant mice, together with in-touch blood smears, illustrated the presence of plasma cells with large basophilic cytoplasm, eccentric nuclei with clock face appearance and occasional multiple Russell bodies ( Figure 7D-7G), all typical features of well differentiated plasma cells. FACS analysis and IHC staining for CD138, a marker for plasma cells, indicated increased proportion of CD138 + B220 2 cells in the spleen of sick Pirh2 2/2 mice ( Figure S6A and S6B).
Infiltration of these plasma cells was evident in regional lymph nodes and kidneys ( Figure 7E and 7G). Pirh2 2/2 plasma cell perivascular infiltrates to non lymphoid organs displayed elevated levels of c-Myc that was accompanied by increased proliferation as indicated by their level of Ki67 ( Figure S7A). Consistent with these data and the increased expression of c-Myc in Pirh2 2/2 splenocytes, activated T-cells or B-cells from Pirh2 2/2 mice displayed higher proliferation rates compared to Wt controls ( Figure S7B). In addition these Pirh2 2/2 activated cells also displayed increased apoptosis compared to Wt controls ( Figure  S7C).
As the cytokine interleukin-6 (IL-6) is important for the differentiation and survival of plasma cells [39], and Pirh2 2/2 mice develop a plasma cell disorder, we next examined by ELISA the serum level of IL-6 in Pirh2 2/2 mice and their controls. We observed elevated levels of IL-6 in the serum of sick Pirh2 mutant mice compared to Wt littermates ( Figure S6D).
Macroscopic analysis of sick Pirh2 2/2 and Pirh2 +/2 mutant mice indicated the presence of pale kidneys in about 40% of the cases. In order to examine the cause for this kidney defect, H&E staining was performed on kidneys from Pirh2 mutants and Wt littermates. Since infiltration of plasma cells was observed in kidneys from Pirh2 mutant mice ( Figure 7G), we investigated whether the kidney defects of these mice would be associated with immunoglobulin (Ig) deposits. Staining of kidney sections with anti-Ig antibodies indicated the presence of glomerular Ig deposition in Pirh2 2/2 and Pirh2 +/2 mice but not in age matched Wt littermates ( Figure S6C).
In accordance with the plasma cell hyperplasia and the elevated level of glomerular Ig deposition observed in Pirh2 mutant mice, ELISA analysis of serum Ig levels of 10 to 12 month-old Pirh2 2/2 mice indicated that these mice suffered gammaglobulinemia. Significantly elevated levels of IgG1, IgG2b and IgA were observed in the serum of Pirh2 2/2 mice compared to their Wt littermates ( Figure S6E). While plasma cells in sick Pirh2 mutant mice showed several histological and morphological features of malignancy, their lack of clonality as assessed by rearrangement of Ig loci, and the polyclonal gammaglobulinemia of Pirh2 mutant mice prevented us from characterizing the disease as a plasma cell neoplasm. These data indicate that Pirh2 deficiency results in plasma cell hyperplasia, gammaglobulinemia, glomerular Ig deposition and kidney failure that likely contribute to the premature death of Pirh2 mutant mice.
P53 is the most frequently inactivated tumor suppressor in human cancer [45], and its loss facilitates c-Myc induced oncogenesis [46]. To examine the effect of p53 deficiency on development and tumorigenesis of Pirh2 mutant mice, we crossed Pirh2 2/2 and p53 2/2 mice and generated Pirh2 2/2 p53 2/2 mice. Double mutant mice were born in a Mendelian ratio and showed no developmental defects. Remarkably, Pirh2 2/2 p53 2/2 mice had a dramatically reduced lifespan compared to p53 2/2 and Pirh2 2/2 mice with all double mutants dying before 126 days of age compared to 230 days for p53 2/2 littermates ( Figure 8A). Necropsy examination of Pirh2 2/2 p53 2/2 mice indicated that 60% of them died due to their tumor burden, while the cause of death remained unknown for the remaining mutants. The tumor spectrum of Pirh2 2/2 p53 2/2 mice was significantly altered compared to Pirh2 2/2 and p53 2/2 mutants and consisted of adenocarcinomas, sarcomas, thymomas and T-cell lymphomas ( Figure 8B, 8E and 8F; Table S2).
These data support Pirh2 function as a tumor suppressor and demonstrate that tumorigenesis of Pirh2 deficient mice is exacerbated by the loss of p53.

Discussion
Ubiquitylation plays critical roles in the regulation of stability and function of p53 and c-Myc, two of the most important tumor suppressors and oncogenes, respectively. Several E3 ligases have been shown to regulate p53 stability [5]; however, the physiological functions of most of these E3 ligases have not yet been addressed. Mdm2 is well accepted as the master regulator for p53 stability and has been demonstrated for its requirement for embryonic and postnatal development [11][12][13]47]. Previous studies also indicated that while Mdm2 is required for p53 stability due to its central role in the polyubiquitylation and degradation of p53 [13,17], Mdm2 independent regulation of p53 degradation also exists. In this study we report that in contrast to Mdm2 deficiency, absence of Pirh2 did not affect embryonic development and only mildly affected p53 steady state levels. However, in response to DNA damage and despite the presence of other E3 ligases for p53 (e.g. Mdm2), Pirh2 deficiency resulted in higher p53 levels in several tissues. Although, we cannot exclude a tissue specificity for the increased IR-induced p53 expression in the absence of Pirh2, collectively our current data indicate a role for Pirh2 in the regulation of p53 stability in response to DNA damage and support a model in which different E3 ligases regulate p53 turnover in a developmental, tissue, stress or time specific manners.
The oncogene c-MYC is frequently deregulated in human cancer [7] and similar to p53, its stability and function are regulated by a number of posttranslational modifications including ubiquitylation. SKP2, FBW7 and ARF-BP1 have been demonstrated to mediate c-MYC ubiquitylation [2,6]. In this study we identified c-MYC as a novel interacting protein for PIRH2 and demonstrated that this interaction requires both the N-and Cterminal domains of PIRH2 as well as the MBII and the C-terminal domain of c-MYC. Similar to its interaction with PIRH2, c-MYC interaction with SKP2 also involves both its MBII and Cterminal domain [22,23].
In this study, we also demonstrate that PIRH2 polyubiquitylates c-MYC and that this c-Myc polyubiquitylation controls c-MYC stability and proteolysis. Pirh2 deficient murine cells and tumors, and human RKO cells knocked down for PIRH2, displayed significantly elevated levels of c-Myc protein, supporting that ubiquitylation control of c-Myc turnover is not only mediated by Skp2 and Fbw7, but also involves Pirh2. Our in vitro data showing that PIRH2 interacts with c-MYC and mediates its polyubiquitylation strongly support a direct role for this E3 ligase in c-MYC polyubiquitylation and degradation.
The embryonic lethality of Fbw7 mutant mice [48] and the postnatal developmental defects of Skp2 mutants [49,50] contrast with the normal embryonic and postnatal development observed with Pirh2 mutants. These developmental differences, together with the differential requirement for these E3 ligases for cancer suppression [2,43], highlight the complex in vivo functions of these E3 ligases.
Although c-Myc protein level was elevated in all examined tissues of Pirh2 2/2 mice, spleen, testis and mammary glands displayed the highest levels of c-Myc. In agreement with the upregulation of p19 ARF in response to high levels of deregulated c-Myc [21], spleen, testis and mammary glands of Pirh2 2/2 mice displayed increased p19 ARF expression. However, p19 ARF expression was not increased in Pirh2 2/2 tissues (e.g. lung, liver) that only displayed a moderate elevation of c-Myc expression. As c-Myc induced p19 ARF expression results in the sequestration of mdm2 in the nucleolus and the subsequent activation of p53 [51], it is possible that increased levels of c-Myc and p19 ARF in Pirh2 2/2 mice might contribute to their higher IR-induced p53 responses compared to control littermates. However, our observation that liver of Pirh2 2/2 mice display elevated IR-induced p53 expression compared to Wt controls whithout any detectable level of p19 ARF , supports that increased IR-induced p53 expression in Pirh2 mutants can take place independently of p19 ARF . Further biochemical and genetic studies are required to determine whether the increased expression of c-Myc and p19 ARF in Pirh2 deficient cells affects their p53 responses.
Our finding that Pirh2 polyubiquitylates c-Myc and mediates its proteolysis, and the elevated c-Myc level associated with Pirh2 deficiency support the possibility that deregulated c-Myc in Pirh2 2/2 mice may increase their cancer susceptibility. Indeed, Pirh2 mutant mice displayed increased predisposition to develop tumors including sarcomas, testis, mammary and lung tumors. In accordance with the role p53 plays in suppressing c-Myc induced oncogenesis [46], Pirh2 2/2 p53 2/2 mice showed a markedly increased cancer predisposition compared to single mutants. The accumulation of c-Myc in Pirh2 mutant mice is also consistent with their elevated risk for developping plasma cell hyperplasia, gammaglobulinemia and kidney failure [36][37][38].
Human PIRH2 has been reported to be overexpressed in lung and prostate cancer [24,25]. However, the human PIRH2 locus at 4q21 is lost in a diverse subset of solid tumors including epithelial tumors of the breast and lung [52,53]. We report in this study that PIRH2 expression is downregulated in various human cancers and that lower PIRH2 expression correlates with decreased survival of patients with lung squamous cell carcinomas, breast or ovarian cancer. Interestingly, Pirh2 mutant mice also displayed increased risk for lung and breast cancer.
The ability of the E3 ligase Pirh2 to negatively regulate IRinduced p53 and c-Myc steady state level, together with the increased risk for Pirh2 mutant mice to develop various pathologies, including cancer, all highlight the importance of this novel tumor suppressor and demonstrate the requirement for its tight regulation.

Generation Pirh2 Mutant Cells and Mice
E14K embryonic stem cells were electroporated with the linearized targeting construct for Pirh2 ( Figure S1) and Pirh2 fl2-3-neo ES clones identified by PCR and Southern blotting. Pirh2 D2-3 ES clones lacking exon 2 and 3 and the Neomycin resistance cassette were obtained following transient transfection of two targeted Pirh2 fl2-3-neo ES clones with CMV-Cre recombinase. Pirh2 D2-3 ES clones were used to derive two independent lines of mutant mice (referred here as Pirh2 mutants). All mice were in 129/C57BL/6 genetic background and were maintained in the animal facility of the Ontario Cancer Institute in accordance with the established ethical care regulations of the Canadian Council on Animal Care. P53 mutant mice [54] were obtained from Taconic.
MEFs were derived according to standard procedures. HEK293T cells, RKO cells and PIRH2-KD RKO clone #11 in which PIRH2 can be knocked down by a tetracycline-inducible system [16], have been also used.

Flow Cytometry
Single cell suspensions from thymi, spleens, lymph nodes and bone marrow of Wt and Pirh2 mutant mice were stained with the following antibodies against the cell surface markers: CD4, CD8, CD3, CD43, CD95, CD138, IgM, IgD, B220, and Thy1.2 (ebioscience and Pharmingen). Stained cells were analyzed by flow cytometry using the FACScan and Cellquest software.

Cell Activation Assay
Activation of T and B-cells was performed as previously described [55].

GST Pull-Down Assays
Recombinant purified full length PIRH2 and its fragments were incubated with His-c-MYC fusion proteins in an assay buffer containing PBS (pH 7.4), 100 mM NaCl, 1 mM Benzamidine, 0.5 mM PMSF and 5 mM bME in a 1:2 molar ratio at 4uC overnight. Then, proteins were incubated with the 100 ml of glutathione-Sepharose beads (GE Healthcare) for an additional 2 h. The mixture was transferred to a microcolumn and was extensively washed with the assay buffer. Bound proteins were eluted with 30 mM reduced glutathione and detected by SDS-PAGE and Coomassie staining
HEK293T cells were transfected with expression plasmids encoding human PIRH2 (pcDNA3-hPIRH2) and c-MYC (pcDNA3-c-MYC) using the calcium phosphate method. After 48 h, the cells were lysed in a solution containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Triton-X, 1 mM phenylmethylsulfonyl fluoride, 400 mM Na 3 VO 4 and protease inhibitor cocktail tablet (Roche). The cell lysates from transfected HEK293T and non transfected NIH3T3 were centrifuged at 14,0006g for 10 min at 4uC, and the resulting supernatant was incubated with anti-c-MYC antibody (N-262) for 2 h at 4uC. Protein G-Sepharose (Amersham) was added to the mixture and the sample rotated overnight at 4uC. The resin was separated by centrifugation, washed four times with ice-cold lysis buffer, and then boiled in SDS sample buffer. Immunoblot analysis was performed with the anti-hPIRH2 (BETHYL laboratories) or antic-MYC antibody (mouse monoclonal antibody; 9E10), horseradish peroxidase-conjugated antibodies to mouse or rabbit immunoglobulin G (1:10,000 dilutions, Cell Signaling) and an enhanced chemiluminescence system (ECL, Amersham). Immunoprecipitation of endogenous PIRH2 from RKO cells was performed using anti-PIRH2 antibody from Santa Cruz (FL261). pcDNA3-hPIRH2 was transfected in RKO cells using the calcium phosphate method.

Intracellular Ubiquitylation Assay
HEK293T cells were transfected with expression plasmids encoding the full length or a mutant lacking the ring finger domain (DR) of human PRIH2 together with Wt or T58A c-MYC and HA-tagged ubiquitin by the calcium phosphate method. After 48 h, the cells were lysed in a modified RIPA buffer as described earlier. The cell lysates were centrifuged at 14,0006g for 10 min at 4uC, and the resulting supernatant was incubated with anti-c-MYC antibody (N262) for 2 h at 4uC. Protein G-Sepharose was added to the mixture, which was then rotated overnight at 4uC. The resin was separated by centrifugation, washed four times with ice-cold lysis buffer, and then boiled in SDS sample buffer. Immunoblot analysis was performed with the anti-ubiquitin (FK2; BIOMOL international) or anti-c-MYC antibody (9E10), horseradish peroxidase-conjugated antibodies to mouse or rabbit immunoglobulin G and ECL.

In Vivo Ubiquitylation Assay
Splenocytes from Wt and Pirh2 2/2 8 week-old mice were cultured in the presence or absence of anti-IgM (20 mg/ml) and/or MG132 (20 mM; calbiochem) for 3 h. Cells were then lysed in a modified RIPA buffer as described earlier and cell lysates immunoprecipitated with anti-c-Myc antibody (N262). The immunocomplexes were denatured in Laemmli's sample buffer (25 mM Tris pH 8.3, 192 mM glycine, 0.1% SDS) to dissociate contaminant proteins associated with c-MYC. c-MYC proteins were reimmunoprecipitated (2 nd IP) with anti-c-Myc antibody. The resin was separated by centrifugation, washed four times with ice-cold lysis buffer, and then boiled in SDS sample buffer. Immunoblot analyses were performed with the anti-ubiquitin (FK2) or anti-c-Myc antibody (9E10), horseradish peroxidaseconjugated antibodies to mouse immunoglobulin G and ECL.

In Vitro Ubiquitylation Assay
Human ubiquitin activating enzyme E1 and 6xHis tagged ubiquitin were purchased from Boston Biochem. The ubiquitylation reaction was performed in a volume of 20 mL in a buffer of 50 mM Tris pH 7.6, 5 mM MgCl 2 , 2 mM ATP, 2 mM DTT. The reaction mixture typically contained E1 (50 ng, Calbiochem), UBE2D2/UbcH5b (100 ng), Ubiquitin (5 mg, Sigma), Pirh2 (1 mg) and full-length c-Myc (0.5 to 1 mg). After incubation at 30uC for 1.5 h, the reactions were stopped by addition of SDS-PAGE sample buffer and resolved on 7.5-10% SDS-PAGE gels. Ubiquitylated proteins were visualized and evaluated by Western Blot using anti-c-MYC (N262) and anti-Ub antibody (U0508, Sigma).

Immunohistochemistry
Tissues and tumors fixed in buffered formalin were processed for paraffin-embedded sectioning at 5 mm, and stained with H&E (Fisher). IHC was performed using anti-c-Myc (C19, Santa Cruz), anti-cleaved caspase 3 (Cell Signaling) or Ki67 (Dako). TUNEL staining was performed using an in situ cell death detection kit (Boehringer Mannheim).

Microarray Data Analysis
Microarray data from public studies were downloaded from GEO [26][27][28][29][30]. All data was log 2 transformed and normalized according to the original authors, with the exception of the lung adenocarcinoma dataset, whose pre-processing has been described elsewhere [56]. Unadjusted Cox proportional hazards models were fit in the R statistical environment (v2.6.2) using the survival package (v2.34). P-values were calculated using the Wald test. Correlation analyses used Pearson's correlation as implemented in R (v2.6.2). Scatter-and box-plots were generated using the lattice package (v0. . Long-term (.7 years) vs. short-term survival groups in ovarian cancer dataset [29] were analyzed using a t-test with Welch's adjustment for heteroscedascity. We have used all 25 long-term and all 30 short-term survival samples available from the raw data. Analyzing only strictly long-and strictly short-term survival samples (n = 23 and n = 27 respectively) does not change the average expression and standard error; it only reduces the significance to p = 0.002162. Additional data were considered from Oncomine Research Edition v.3.6, with statistical analysis as described previously [32].  Figure S3 Absence of Pirh2 leads to higher levels of Serine 15 phosphorylated p53 and increased apoptosis in response to irradiation. (A) Time course immunoblot analysis of the expression level of Serine 15 phosphorylated p53 (S15-p53), total p53, Bax and Pirh2 in response to irradiation (6 Gc) of Wt and Pirh2 2/2 splenocytes. *: non specific. (B) 6 to 8 week-old Wt and Pirh2 2/2 mice either untreated or 1 h post whole-body irradiation (6 Gc) were sacrificed and IHC was performed to assess the level of S15-p53 in their spleen. Bar = 50 mm. (C) H1299 cells were cotransfected with either pcDNA3.1, pcDNA3.1-Mdm2 or pcDNA3.1-PIRH2 and a p53 expression vector (Wt, S15A, S15D). Lysates prepared 40 h post-transfection were examined by Western blotting using the indicated antibodies. (D) 6 to 8 weekold Wt (n = 3) and Pirh2 2/2 mice (n = 3) were subjected to wholebody irradiation (6 Gc) and the levels of apoptosis in thymus at different time points post-IR were examined using active caspase 3 (casp3) and IHC. Bar = 100 mm. (E) Active casp3 positive cells in intestinal crypts of untreated (0 h) and irradiated Wt (n = 3) and