TAp73 knockout shows genomic instability with infertility and tumor suppressor functions

Characterization of TAp73 deletion and its effects on the expression and activities of Trp53 family members

To determine the specific role of the TAp73 isoforms of p73 (Fig. 1A), we generated TAp73-deficient mice in which deletion of exons 2 and 3 of the Trp73 gene (Fig. 1B) selectively abolished expression of all TAp73 isoforms (Supplemental Fig. S1A–C). Firstly, we confirmed the specific deletion of TAp73, and we analyzed the expression patterns of ΔNp73, p53, and p63 in various tissues of TAp73^−/− mice. Although p53 and p63 expression were not modified (Supplemental Fig. S2A,D; data not shown), ΔNp73 mRNA was up-regulated in most tissues of TAp73^−/− mice (Fig. 1C; Supplemental Fig. S2B). The exception was the lung, where ΔNp73 mRNA was not significantly different from that of wild-type mice (Fig. 1C). We also observed enhanced ΔNp73 mRNA expression in TAp73^−/− mouse embryonic fibroblasts (MEFs), and this increased expression was persistent even after the excision of the PGK-neo cassette (Supplemental Fig. S1D,E), indicating that the observed up-regulation of ΔNp73 mRNA is due to the absence of TAp73 proteins rather than to the insertion of the PGK-neo cassette close to the ΔNp73 promoter. However, the enhanced ΔNp73 mRNA level is not reflected in a significant increase in ΔNp73 protein level. Indeed, we analyzed ΔNp73 protein level in lung and ovaries from TAp73^−/− tissues versus wild-type tissues by using two antibodies—one specific of the ΔN region (IMG-313) and one specific of the a-C-terminal region from all p73 isoforms (C-17)—and we compared the bands corresponding to the ΔNp73 isoforms (Fig. 1D). Therefore, the antibody data would suggest that the increases in ΔNp73 transcript levels are not accompanied by parallel increases in expression of ΔNp73 protein. The elevated levels of ΔNp73 mRNA in TAp73^−/− mice were unexpected because TAp73 is a known transcriptional inducer of ΔNp73 (Grob et al. 2001). However, further examination showed that, in wild-type mice, ΔNp73 levels decreased more rapidly than levels of TAp73 following DNA damage (Maisse et al. 2004; Toh et al. 2004) due to the ubiquitination and proteasomal degradation of ΔNp73 by an as-yet-uncharacterized ubiquitin ligase (B.S. Sayan, C. Piro, A.L. Yang, and G. Melino, unpubl.). These data suggest that there are multiple mechanisms that operate post-transcriptionally to control the levels of various p73 isoforms, resulting in a complex inteplay between TAp73 and ΔNp73. Nonetheless, our TAp73^−/− knockout seems to be a pure TAp73-deficient mouse at the protein level, reminiscent of the TAp73/ΔNp73 imbalance observed in human tumors.

Since ΔNp73 is a functional inhibitor of p53 and TAp73 isoforms (Grob et al. 2001; Deyoung and Ellisen 2007), and as we observed an increase in ΔNp73 mRNA level, we wanted to be sure that no ΔNp73 activity was induced, potentially interfering with p53 functions. We showed that p53 activation was comparable in wild-type and TAp73^−/− MEFs (Supplemental Fig. S2C,D). Furthermore, we detected no differences between wild-type and TAp73^−/− thymocytes (Fig. 1E), primary MEFs (Supplemental Fig. S2E), or E1A-transformed MEFs (Supplemental Fig. S2F) in p53-dependent cell death induced by γ-irradiation, etoposide, cisplatin, or doxorubicin. The normal sensitivity to death of TAp73^−/− thymocytes is particularly interesting in view of the finding that E2F binding to the TAp73 promoter is required for the TCR-induced death of human peripheral T cells from wild-type mice as well as for the death of cancer cells (Irwin et al. 2000; Lissy et al. 2000; Stiewe and Pützer 2000). Our data suggest that the ΔNp73 mRNA increase observed in most tissues from TAp73-deficient mice is not relevant since ΔNp73 protein levels appear normal. A compensatory mechanism may have been elicited in the TAp73-deficient mice to guarantee normal ΔNp73 protein expression under normal or physiological conditions.

Loss of TAp73 leads to infertility due to defects in early embryonic development

A significant percentage of TAp73^−/− mice die during embryogenesis, resulting in an abnormal Mendelian ratio at birth (+/+, 29%; +/−, 52%; −/−, 19%) either in mixed or pure background mice (Supplemental Fig. S3A). This partial embryonic lethality could theoretically be caused by the potential enhanced ΔNp73 protein not balanced by the TAp73 at specific stages of embryonic development. Indeed, strongly deregulated ΔNp73 expression in ΔNp73 transgenic mice leads to 100% embryonic lethality partially due to inhibition of p53 activities (Erster et al. 2006; Huttinger-Kirchhof et al. 2006).

We were unable to obtain pups from mating trials involving mating a TAp73^−/− male or female with wild-type mice, suggesting that both sexes are infertile (Supplemental Fig. S3B). In contrast to Trp73^−/− mice, in which defects in sensory and hormonal pathways contribute to reproductive and behavioral phenotypes that lead to male and female infertility, TAp73^−/− mice mate normally and the females display normal cyclicity (Supplemental Fig. S3B; data not shown). To investigate the nature of TAp73^−/− female infertility, we induced superovulation in 3- to 4-wk-old TAp73^−/− females and their wild-type littermate sisters using exogenous gonadotropins. In contrast to wild-type females, no oocytes were present in the fallopian tubes of TAp73^−/− females 16 h after administration of hCG (human choriogonadotropin). When the ovaries and oviducts of TAp73^−/− females were dissected, we found that the ovulated oocytes were trapped under the bursa and did not progress toward the fallopian tubes (Fig. 2A, right). Upon collection of these oocytes, it became clear that TAp73^−/− females also ovulated fewer gametes (Fig. 2B).

To establish if altered follicular development was responsible for the oocyte deficit in TAp73^−/− female mice, we performed morphometric analyses (based on the counting of nonatretic follicles) of ovaries obtained from adolescent TAp73^+/+ and TAp73^−/− females. Although significant differences were observed in the number of quiescent (primordial) or early growing (primary) follicles, numbers of early secondary follicles were comparable (Fig. 2C). To examine the later “Graffian follicle” stages, which are dependent on gonadotropins, we treated prepubertal TAp73^+/+ and TAp73^−/− females with a single dose of pregnant mare serum gonadotropin (PMSG) to synchronize follicular development. Ovaries collected 46 h later revealed no significant differences in the competence of TAp73^−/− gonads to form large antral ovulatory follicles (Fig. 2D, left). However, when we assessed the ovulatory capacity of these follicles, we found that a greater number of oocytes remained trapped within the luteininzing granulosa cells of TAp73^−/− ovaries compared with wild-type females (Fig. 2D, right). To confirm that ΔNp73 was not responsible for that phenotype, we determined that ΔNp73 mRNA expression level was even reduced in MII oocytes from TAp73^−/− mice compared with wild-type MII oocytes (Supplemental Fig. S4A). These results suggest that the decreased number of ovulated oocytes in TAp73^−/− mice is likely due to the retention of some germ cells within the ovary.

We next investigated whether oocyte quality contributed to TAp73^−/− female sterility. Germinal vesicle oocytes as well as ovulated oocytes from 3- to 4-wk-old TAp73^+/+ and TAp73^−/− mice were induced to mature in vitro. In both cases, oocytes of both genotypes showed the same rate of arrest in metaphase II (Fig. 3A,B, left). However, TAp73^−/− oocytes exhibited a striking increase in spindle abnormalities (Fig. 3A,B, right), which included multipolar spindles, spindle relaxation, and spindle scattering accompanied by varying degrees of misalignment (Fig. 3C). We next compared the developmental competence of ovulated TAp73^+/+ and TAp73^−/− oocytes by assessing their ability to undergo in vitro fertilization (IVF) and preimplantation development in vitro. While the fertilization rate was the same, only 28% of zygotes originating from TAp73^−/− oocytes became blastocysts, compared with 75% of zygotes originating from TAp73^+/+ oocytes (Fig. 3D). The majority of embryos obtained from TAp73^−/− oocytes arrested during early cleavage, resulting in embryos with multinucleated blastomeres, and blastocysts of inferior quality with an abnormal cell number (Fig. 3E,F). Thus, Trp73 is a maternal-lethal effect gene, because a lack of TAp73 in developing oocytes leads to a failure of preimplantation embryonic development.

Poor oocyte quality leading to abnormal embryonic development is also associated with maternal aging, and is one of the most important prognostic factors of human infertility. Because oocytes from young TAp73^−/− mice displayed a similar spectrum of spindle abnormalities to that observed in oocytes from aged wild-type mice (Tarin et al. 2001), we evaluated TAp73 expression in the ovaries of 8- and 55-wk-old wild-type mice. We found that natural aging abolishes TAp73 expression in oocytes (Supplemental Fig. S4B,C). This observation suggests that loss of TAp73 may be responsible for the compromised developmental capacity of aged normal oocytes.

To identify TAp73 transcriptional targets potentially involved in the regulation of developmental competence, we decided to perform a gene microarray analysis of ovarian tissue. We first used immunocytochemistry to determine what stages of oogenesis were most likely to be affected by a lack of TAp73. Growing wild-type oocytes from the primary follicle stage onward displayed only cytoplasmic expression of TAp73, but the oocytes of wild-type primordial follicles often exhibited nuclear staining (Fig. 3G; Supplemental Fig. S4B). Thus, we used neonatal ovaries, which are enriched in oocytes at this developmental stage, as the target tissue for our microarray analysis. Our results showed that the expression of 79 annotated genes in neonatal ovary was altered by at least 1.8-fold by an absence of TAp73. We confirmed for some genes their modulation of expression in TAp73^−/− ovaries compared with wild-type ovaries (Supplemental Fig. S5A,B; data not shown). Among these 79 genes, many were involved in the regulation of transcription and translation via control of RNA processing (see Tables 1, 2). We confirmed that most of these genes are potential or already known p53 targets (Supplemental Fig. S5C). To determine functional links among the differentially regulated targets, we mapped them to the I²D protein interaction database (Brown and Jurisica 2007). Extensive analyses identified several proteins that interacted strongly with TAp73, and also showed that all of these proteins were at points of network disconnect, which suggests that they are essential for signaling within this network (Fig. 3H). Indeed, several of these genes (as example: PEG10 and Ctr9) have been disrupted in animal models with resulting embryonic lethality or infertility (Ono et al. 2006; Akanuma et al. 2007). Taken together, these data suggest that these target genes may normally be regulated by TAp73 or are physical interactors of TAp73, as, for example, in the case of PIAS1 that regulates cell cycle checkpoint progression by sumoylation of TAp73 (Munarriz et al. 2004), and thus their dysregulation may contribute to preimplantation lethality in the absence of TAp73.

TAp73 is necessary for hippocampal development but not for neural cell maintenance

Previous analyses of Trp73^−/− mice have demonstrated that p73 is critical for the development and maintenance of the nervous system (Yang et al. 2000). Indeed, Trp73^−/− mice exhibit hippocampal dysgenesis with a gradual but persistent postnatal loss of neurons with greatly enlarged ventricles (hydrocephalus) and greatly reduced cortical tissue (Yang et al. 2000; Pozniak et al. 2002). To investigate whether either of these phenotypes might be due to the loss of TAp73 isoforms, we carried out coronal sectioning of the brains of 10-wk-old wild-type and TAp73^−/− mice. Nissl staining of the forebrain revealed that TAp73^−/− brains were similar to their wild-type counterparts with regard to the size of the lateral ventricles and the thickness of the cortex (Fig. 4A,B) but that hippocampal anatomy was abnormal. In particular, the lower blade of the dentate gyrus was either completely missing or greatly truncated, depending on the level of the section (Fig. 4C–F). This hippocampal dysgenesis was strikingly similar to that seen in Trp73^−/− mice at postnatal day 14, before the occurrence of the ventricular enlargement and loss of neural tissue that further perturb hippocampal morphology in adult Trp73^−/− mice. The similarity between TAp73^−/− mice and p73^−/− mice on this particular phenotype suggests that it is due to TAp73 loss and not to ΔNp73 loss. Thus, our data support previous work (Pozniak et al. 2000, 2002; Walsh et al. 2004) showing that TAp73 is essential for normal hippocampal development, while ΔNp73 seems to prevent neural tissue loss and ventricular enlargement in adult life.

TAp73^−/− mice are tumor-prone and sensitive to carcinogens

While infertility and neural abnormalities detected in the TAp73^−/− mice were not identical but reminiscent of those described for Trp73^−/− mice (Yang et al. 2000), other aspects of the phenotype were markedly different. Indeed, although the life span of TAp73^−/− mice was significantly shorter than that of wild-type mice (Fig. 5A), it still exceeded that of Trp73^−/− mice (Yang et al. 2000). Furthermore, whereas untreated Trp73^−/− mice do not develop tumors, 30% of TAp73^+/− and 73% of TAp73^−/− mice of a mixed genetic background spontaneously developed malignancies (Fig. 5B,C). Lung adenocarcinoma was the most frequent cancer that developed in TAp73^−/− mice, being observed in 32% of all TAp73^−/− mice analyzed and representing 44% of all tumors excised from TAp73^−/− mice. In the lung of the TAp73^−/− knockout, no changes in either ΔNp73 mRNA or protein compared with wild-type were observed (see above). Importantly, no correlation was found between tumor onset and p53, p63, or ΔNp73 protein expression in any tumor analyzed (data not shown), suggesting that the appearance of tumors in TAp73^+/− and TAp73^−/− mice was specifically due to TAp73 deletion rather than to ΔNp73, p63, or p53 deregulation. Consistent with a previous report (Roman-Gomez et al. 2006), loss-of-heterozygosity (LOH) of p73 had occurred in 66% of tumors developing in TAp73^+/− mice (Fig. 5D).

To confirm the tumor-prone phenotype of TAp73^−/− mice, we induced carcinogenesis by intraperitoneal injection of DMBA into TAp73^+/+ and TAp73^−/− female littermates. The mean latency time for tumor development in TAp73^−/− mice was 18.7 ± 7.1 wk, significantly shorter than the 32 ± 3.4-wk mean latency observed in DMBA-treated TAp73^+/+ mice (Fig. 5E). However, there was no difference between DMBA-treated TAp73^+/+ and TAp73^−/− mice in sites of tumor development, which included the colon, intestine, liver, and stomach (data not shown). These results indicate that TAp73 isoforms mediate the tumor-suppressive function of Trp73, and that loss of TAp73 alone is sufficient to promote oncogenic transformation.

Taking into account the observation that TAp73^−/− mice develop predominantly lung adenocarcinoma, we decided to investigate the expression of p73 isoforms in human lung cancer. Examination of the expression of TAp73 and ΔNp73 in matched normal and tumoral lung tissue samples from 18 lung cancer patients revealed that most tumoral lung samples exhibited a down-regulation of TAp73 and an up-regulation of ΔNp73, with a consequent decrease in the TAp73/ΔNp73 ratio (Supplemental Fig. S6). Notably, in the two samples (patients 133 and 177) in which TAp73 expression was up-regulated, greater amounts of ΔNp73 were also expressed. These results are consistent with the postulated inactivation of p53 in human lung cancer (Haruki et al. 2001).

TAp73 participates in maintenance of genomic stability

Since we observed an increased incidence of spontaneous tumors and also an abnormal spindle formation in the oocytes of TAp73^−/− mice, we next investigated the possible impact of TAp73 deficiency on the maintenance of genomic stability. Accurate chromosome segregation during meiosis and mitosis is critical to the preservation of euploidy in eukaryotic cells (Yuen et al. 2005), and errors in the molecular mechanisms regulating segregation result in aneuploidy (Hassold and Hunt 2001; Thomas et al. 2001). We analyzed the ability of TAp73-deficient cells to undergo mitotic arrest and observed that TAp73^−/− MEFs showed a significant decrease in nocodozole-induced mitotic arrest. Indeed, at 25 h post-nocodozole, the percentages of cells in G1 and >G2/M were increased in TAp73^−/− MEFs compared with TAp73^+/+ MEFs (Fig. 6A; Supplemental Fig. S7A). Immunoblotting of mitotic phase marker proteins such as cyclin B1 and securin revealed that mitotic “slippage” had occurred in TAp73^−/− MEFs (data not shown). Premature mitotic exit in TAp73^−/− MEFs was confirmed by immunostaining to detect phospho-histone H3-positive cells (Fig. 6B) and by counting mitotic figures (Fig. 6C, right). Thus, in the presence of nocodazole, TAp73 deficiency results in the abnormal mitotic progression of cells that otherwise would not be permitted to pass into the G1 phase. We observed a similar premature mitotic exit in p73^−/− MEF cells treated with nocodazole, strongly suggesting that TAp73 isoforms are the functional isoforms in mitotic regulation rather than ΔNp73 isoforms (Supplemental Fig. S7B,C).

After 25 h of nocodazole treatment, an 8N population of polyploid cells appeared frequently among TAp73^−/− MEFs when compared with TAp73^+/+ MEFs (Fig. 6A). Indeed, ∼14% of TAp73^−/− MEFs showed aneuploidy compared with 5% of cells in TAp73^+/+ MEF cultures. The percentage of multinuclear cells was also increased in TAp73^−/− MEFs treated with taxol or nocodozole (Fig. 6C, left). When H1299 cells were engineered to overexpress the TAp73β isoform, they were protected from taxol-induced formation of multinuclear cells (Fig. 6D). In contrast, overexpression of ΔNp73α or ΔNp73β in Trp73^−/− MEFs did not reduce the enhanced formation of multinuclear cells (data not shown).

We next examined lung and thymic cell populations from TAp73^−/−, Trp53^−/−, and wild-type mice that were treated or not with nocodazole for 12 h. TAp73^−/− lung fibroblasts showed a higher frequency of >G2/M cells, demonstrating a loss of euploidy, which are likely aneuploid, than did Trp53^−/− or wild-type lung fibroblasts, whereas no differences were observed among TAp73^−/−, Trp53^−/−, and wild-type thymic cells (Fig. 6E,F). Thus, the effect of TAp73 on aneuploidy appears to be tissue-specific, perhaps explaining why TAp73^−/− mice preferentially develop lung adenocarcinomas. Our results indicate that, at least in lung tissue, TAp73 loss is associated with decreased genomic stability.

To confirm more directly that TAp73 deficiency is associated with an altered DNA content, we karyotyped TAp73^−/− and TAp73^+/+ 3T3 cells and observed that the mutant cultures contained a significantly higher percentage of cells with an abnormal karyotype compared with wild-type controls (Fig. 6G). The affected cells showed a loss or gain of a single chromosome, indicating aneuploidy. These data further support a role for TAp73 in maintaining genomic stability.

Generation of TAp73^−/− mice and LOH analysis

Mutant mice deficient for Trp73 exons 2 and 3, which are specific for the TAp73 isoforms, were generated by conventional gene targeting procedures in Sv129Ola embryonic stem cells. Blastocysts were transferred to pseudopregnant C57BL/6J female mice. TAp73 wild-type (wild type, +/+; TAp73^+/+), heterozygous (+/−; TAp73^+/−), and homozygous (−/−; TAp73^−/−) littermate mice (Sv129Ola × C57BL/6J) were created by intercrossing TAp73^+/− mice. Offspring were genotyped by PCR analysis using specific primer pairs to detect the wild-type Trp73 and Trp73^ΔTA alleles. Sense and antisense primers for the wild-type Trp73 allele were 5′-CTGGTCCAGGAGGTGAGACTGAG GC-3′ and 5′-CTGGCCCTCTCAGCTTGTGCCACTTC-3′, respectively. Sense and antisense primers for the gene-targeted Trp73^ΔTA allele were 5′-GTGGGGGTGGGATTAGATAAAT GCCTG-3′ and 5′-CTGGCCCTCTCAGCTTGTGCCACTTC3′, respectively. Predicted PCR product sizes were 1.0 kb and 1.2 kb for the wild-type Trp73 and Trp73^ΔTA alleles, respectively. F4 mixed background mice were used for all experiments except the analysis of the Mendelian ratio, for which we also used animals of the fifth-generation backcross in a C57B6 background. All animals were treated in accordance with the NIH Guide for Care and Use of Laboratory Animals as approved by the Ontario Cancer Institute Animal Care Committee (Toronto, ON, Canada). LOH analysis was performed using genotyping primers and PCR amplification of Trp73 exons 2 and 3 in genomic DNA from tumor and normal (tail) tissues of tumor-bearing TAp73^+/− mice.

Analyses of spontaneous and induced tumors

Life span and spontaneous tumor incidence were determined in untreated TAp73^+/+ and TAp73^−/− mice. Mice with visible tumors and moribund mice showing weight loss or difficulties in moving were sacrificed upon detection. The remaining surviving mice were sacrificed at 26 mo of age. For induced tumorigenesis, TAp73^+/+ and TAp73^−/− mice (23 mice/group) were injected i.p. with 1 mg/kg dimethylbenz[a]anthracene (DMBA; Sigma-Aldrich) starting at 63 d post-partum and continuing once a week for 6 wk. This dose was sufficient to induce 100% tumor incidence in controls. Tumors and tissues of control and mutant mice were collected and processed for histopathology and RNA/DNA isolation by standard procedures.

Oocyte collection, in vitro maturation, and in vitro fertilization

Female mice (3–7 wk old) were superovulated, and mature oocytes were collected as previously described (Perez et al. 2005). For in vitro maturation, germinal vesicle stage oocytes from 3-wk-old females were induced to undergo spontaneous in vitro maturation (Yao et al. 2004), and in vitro fertilization was carried out as described previously (Acton et al. 2004). Embryonic developmental competence was assessed daily. On day 4, embryos were fixed and cell number, cell death, and mitotic rates were determined as described previously (Jurisicova et al. 1998). Statistics were obtained using the χ² or Student’s t-test.

Microarray studies

Microarray analysis was performed on day 4 ovarian tissue from two sibling pairs of TAp73^+/+ and TAp73^−/− females by the Center for Applied Genomics (Hospital for Sick Children, Toronto) using the Affymetrix platform (Mouse MOE 430 2.0) with two cycle amplification steps. Data were preprocessed using RMAExpress 1.0 beta 3 (Bolstad et al. 2003), log₂ transformed, and then analyzed using SAM version 3.0 (Tusher et al. 2001). Using FDR (False Discovery Rate) = 0.603%, we identified 2538 genes that were differentially expressed (1178 up-regulated and 1360 down-regulated in TAp73^+/+ ovaries). To further diminish false positives, we prioritized the genes by considering only fold change >1.8, which resulted in 16 genes up-regulated and 60 genes down-regulated in wild-type ovaries.

To determine functional links among identified target genes, we mapped them to the Interologous Interaction Database I²D ver. 1.71 (http://ophid.utoronto.ca/i2d). A network was formed containing 531 proteins and 977 interactions on seven targets up-regulated in TAp73^+/+ ovaries and 26 targets up-regulated in TAp73^−/− ovaries. Further visualization and analysis of the resulting protein interaction network were performed using NAViGaTOR ver 2.0 (http://ophid.utoronto.ca/navigator). Shortest path calculations were used to determine critical proteins and interactions. Briefly, the shortest path among all nodes within the network was computed, and the frequency of inclusion of individual proteins and interactions was counted. The most frequent nodes and lines form the critical subgraph within the network. We applied blending to reduce network complexity (NAViGaTOR version 2.0).

Immunohistochemical analysis

For histological analysis of mouse brains, sacrificed animals were perfused with 4% paraformaldehyde, and their brains were cryoprotected, sectioned, and Nissl-stained as described previously (Pozniak et al. 2002). For histological analysis of oocytes, oocyte fixation and immunocytochemistry were performed as described previously (Yao et al. 2004). Samples were viewed on a deconvolution fluorescence microscope. Follicular histomorphology was performed on fixed ovarian tissue obtained from 3-wk-old females (Canning et al. 2003), and immunocytochemistry on ovarian sections was performed as described previously (Matikainen et al. 2001). The antibodies used for analyses of oocytes and/or ovarian sections were rat monoclonal anti-tubulin (YL1/2; Abcam), rabbit polyclonal anti-p73 (H79; Santa Cruz Biotechnologies), and goat anti-p73α (C17; Santa Cruz Biotechnologies). For spindle analyses, secondary antibody was conjugated to Oregon green (Molecular Probes). For ovarian sections, visualization was performed using a biotinylated anti-rabbit antibody detection kit (DAKO).

Cell lines and overexpression experiments

MEFs obtained from embryonic day 13.5 (E13.5) TAp73^+/+, TAp73^−/−, and p73^−/− embryos, or 3T3 cells, were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FCS without antibiotics. For overexpression experiments, the human lung carcinoma cell line H1299 was plated at 60% density the day before transfections. Transient transfection of plasmids was carried out using the FuGENE reagent (Roche Applied Science) according to the manufacturer’s protocol. Control vector pcDNA3-HA was from Clontech. DNA sequences encoding the TAp73α and TAp73β isoforms were all cloned in pcDNA3-HA.

Flow cytometric analyses

TAp73^−/− MEFs were seeded at 3 × 10⁵ cells per 100-mm plate. Cultures were then either left untreated or treated with nocodazole (200 ng/mL; Sigma-Aldrich). Immediately (control) or at 16, 21, 25, or 48 h after nocodazole treatment, cells were harvested and fixed for at least 1 h at 4°C in PBS/70% ethanol. Fixed cells were resuspended in PBS containing 100 μg/mL RNase A and 50 μg/mL propidium iodide (Sigma-Aldrich) and immediately analyzed using a FACSCalibur (Becton Dickinson Biosciences). Data were collected and analyzed with CellQuest Pro (BD Biosciences) and FlowJo 7.2.2 Data Analysis Software. Phospho-histone H3 staining was performed using an anti-phospho-histone H3 antibody (#9708; Cell Signaling) according to the manufacturer’s protocol. Cell cycle analysis of total cell populations from fresh TAp73^+/+, TAp73^−/−, and p53^−/− thymus or lung tissues was performed as previously described (Yamamoto et al. 2003).

Cell death analysis was performed by flow cytometry using propidium iodide staining together with AnnexinV-FITC as described previously.

Western blotting

Protein extracts (100 μg) were fractionated by SDS-PAGE and subjected to Western blotting by standard procedures. Western blots were quantified by band densitometry using an Odyssey Infrared Imaging System (LI-COR Biosciences). For immunoprecipitations, protein extracts from MEFs were analyzed according to standard protocols. The antibodies used for Western blot were mouse anti-ΔNp73 (IMG-313; Imgenex), goat anti-p73 (C-20; Santa Cruz), rabbit anti-phospho-p53 ser15 (#9284L; Cell Signaling), and rabbit anti-p53 total (CM5; Vector Laboratories).