Functional interplay between MDM2, p63/p73 and mutant p53.

Many cancers express mutant p53 proteins that have lost wild-type tumor suppressor activity and, in many cases, have acquired oncogenic functions that can contribute to tumor progression. These activities of mutant p53 reflect interactions with several other proteins, including the p53 family members p63 and p73. Mutations in p53 that affect protein conformation (such as R175H) show strong binding to p63 and p73, whereas p53 mutants that only mildly affect the conformation (such as R273H) bind less well. A previously described aggregation domain of mutant p53 is not required for p63 or p73 binding; indeed, mutations within this region lead to the acquisition of a mutant p53 phenotype-including a conformational shift, p63/p73 binding and the ability to promote invasion. The activity of wild-type p53 is regulated by an interaction with MDM2 and we have investigated the potential role of MDM2 in the mutant p53/p63/p73 interactions. Both mutant p53 and p73 bind MDM2 well, whereas p63 binds much more weakly. We found that MDM2 can inhibit p63 binding to p53R175H but enhances the weaker p53R273H/p73 interaction. These effects on the interactions are reflected in an ability of MDM2 to relieve the inhibition of p63 by p53R175H, but enhance the inhibition of p73 activity by p53R175H and R273H. We propose a model in which MDM2 competes with p63 for binding to p53R175H to restore p63 activity, but forms a trimeric complex with p73 and p53R273H to more strongly inhibit p73 function.

C-terminal ubiquitination of p53 contributes to nuclear export.

The growth inhibitory functions of p53 are controlled in unstressed cells by rapid degradation of the p53 protein. One of the principal regulators of p53 stability is MDM2, a RING finger protein that functions as an E3 ligase to ubiquitinate p53. MDM2 promotes p53 nuclear export, and in this study, we show that ubiquitination of the C terminus of p53 by MDM2 contributes to the efficient export of p53 from the nucleus to the cytoplasm. In contrast, MDM2 did not promote nuclear export of the p53-related protein, p73. p53 nuclear export was enhanced by overexpression of the export receptor CRM1, although no significant relocalization of MDM2 was seen in response to CRM1. However, nuclear export driven by CRM1 overexpression did not result in the degradation of p53, and nuclear export was not essential for p53 degradation. These results indicate that MDM2 mediated ubiquitination of p53 contributes to both nuclear export and degradation of p53 but that these activities are not absolutely dependent on each other.

Mdm2 is a RING finger-dependent ubiquitin protein ligase for itself and p53.

Mdm2 has been shown to regulate p53 stability by targeting the p53 protein for proteasomal degradation. We now report that Mdm2 is a ubiquitin protein ligase (E3) for p53 and that its activity is dependent on its RING finger. Furthermore, we show that Mdm2 mediates its own ubiquitination in a RING finger-dependent manner, which requires no eukaryotic proteins other than ubiquitin-activating enzyme (E1) and an ubiquitin-conjugating enzyme (E2). It is apparent, therefore, that Mdm2 manifests an intrinsic capacity to mediate ubiquitination. Mutation of putative zinc coordination residues abrogated this activity, as did chelation of divalent cations. After cation chelation, the full activity could be restored by addition of zinc. We further demonstrate that the degradation of p53 and Mdm2 in cells requires additional potential zinc-coordinating residues beyond those required for the intrinsic activity of Mdm2 in vitro. Replacement of the Mdm2 RING with that of another protein (Praja1) reconstituted ubiquitination and proteasomal degradation of Mdm2. However, this RING was ineffective in ubiquitination and proteasomal targeting of p53, suggesting that there may be specificity at the level of the RING in the recognition of heterologous substrates.

Analysis of the degradation function of Mdm2.

Degradation of the p53 tumor suppressor protein has been shown to be regulated by Mdm2. In this study, we identify regions of Mdm2 that are not required for p53 binding but are essential for degradation. Mdm2 mutants lacking these regions function in a dominant negative fashion, stabilizing endogenous p53 in cells by interfering with the degradative function of the endogenous Mdm2. p53 protein stabilized in this way does not strongly enhance the expression of p21(Waf1/Cip1), the product of a p53-responsive gene, supporting the model in which binding of Mdm2 to the NH2-terminal domain of p53 inhibits interaction with other components of the basal transcriptional machinery. Interestingly, COOH-terminal truncations of Mdm2 that retain p53 binding but fail to mediate its degradation are also stabilized themselves. Because Mdm2, like p53, is normally an unstable protein that is degraded through the proteasome, this result suggests a direct link between the regulation of Mdm2 and p53 stability.

Regulation of Mdm2-directed degradation by the C terminus of p53.

The stability of the p53 tumor suppressor protein is regulated by interaction with Mdm2, the product of a p53-inducible gene. Mdm2-targeted degradation of p53 depends on the interaction between the two proteins and is mediated by the proteasome. We show here that in addition to the N-terminal Mdm2 binding domain, the C terminus of p53 participates in the ability of p53 to be degraded by Mdm2. In contrast, alterations in the central DNA binding domain of p53, which change the conformation of the p53 protein, do not abrogate the sensitivity of the protein to Mdm2-mediated degradation. The importance of the C-terminal oligomerization domain to Mdm2-targeted degradation of p53 is likely to reflect the importance of oligomerization of the full-length p53 protein for interaction with Mdm2, as previously shown in vitro. Interestingly, the extreme C-terminal region of p53, outside the oligomerization domain, was also shown to be necessary for efficient degradation, and deletion of this region stabilized the protein without abrogating its ability to bind to Mdm2. Mdm2-resistant p53 mutants were not further stabilized following DNA damage, supporting a role for Mdm2 as the principal regulator of p53 stability in cells. The extreme C terminus of the p53 protein has previously been shown to contain several regulatory elements, raising the possibility that either allosteric regulation of p53 by this domain or interaction between this region and a third protein plays a role in determining the sensitivity of p53 to Mdm2-directed degradation.

Activation of p53 DNA binding activity by point mutation.

The p53 tumor suppressor protein can adopt both latent, non-DNA binding and active, DNA binding forms, and p53 activity is thought to be regulated in cells, at least in part, through a conformational shift which leads to sequence specific DNA binding. In vitro, this allosteric regulation of DNA binding by p53 has been shown to be mediated through the C-terminus of the protein. We show here that although deletion of the C-terminal 16 amino acids of p53 did not activate DNA binding, deletion of a further eight amino acids resulted in constitutive activation of DNA binding activity. Simultaneous mutation of the three lysine residues within these eight amino acids also resulted in constitutive DNA binding activity, although this was reduced when only two of these lysines were altered. The deletion or point mutants of p53 showing constitutive DNA binding activity did not display clear evidence of DNA binding site specificity, although some binding site preference was seen with the point mutants. Each of the constitutively active p53 mutants retained transcriptional activity and induced both cell cycle arrest and apoptosis in transiently transfected cells at rates comparable with the wild type protein.

Differential activation of target cellular promoters by p53 mutants with impaired apoptotic function.

The p53 tumor suppressor protein is a sequence-specific transcriptional activator, a function which contributes to cell cycle arrest and apoptosis induced by p53 in appropriate cell types. Analysis of a series of p53 point mutants has revealed the potential for selective loss of the ability to transactivate some, but not all, cellular p53-responsive promoters. p53 175P and p53 181L are tumor-derived p53 point mutants which were previously characterized as transcriptionally active. Both mutants retained the ability to activate expression of the cyclin-dependent kinase inhibitor p2lcip1/waf1, and this activity correlated with the ability to induce a G1 cell cycle arrest. However, an extension of this survey to include other p53 targets showed that p53 175P was defective in the activation of p53-responsive sequences derived from the bax promoter and the insulin-like growth factor-binding protein 3 gene (IGF-BP3) promoter, while p53 181L showed loss of the ability to activate a promoter containing IGF-BP3 box B sequences. Failure to activate transcription was also reflected in the reduced ability of the mutants to bind the p53-responsive DNA sequences present in these promoters. These specific defects in transcriptional activation correlated with the impaired apoptotic function displayed by these mutants, and the results suggest that activation of cell cycle arrest genes by p53 can be separated from activation of genes with a role in mediating the p53 apoptotic response. The cellular response to p53 activation may therefore depend, at least in part, on which group of p53-responsive genes become transcriptionally activated.

Rat embryo fibroblasts immortalized with simian virus 40 large T antigen undergo senescence upon its inactivation.

Introduction of simian virus 40 T antigen into rodent fibroblasts gives rise to cells that can proliferate indefinitely but are dependent upon it for maintenance of their growth once the normal mitotic life span has elapsed. Inactivation of T antigen in these immortalized cells causes rapid and irreversible cessation of growth. To determine whether this growth arrest is associated with entry into senescence, we have undertaken a genetic and biological analysis of conditionally immortal (tsa) cell lines derived by immortalizing rat embryo fibroblasts with the thermolabile tsA58 T antigen. This analysis has identified the following parallels between the tsa cells after inactivation of T antigen and senescent rat embryo fibroblasts: (i) growth arrest is irreversible; (ii) it occurs in G1 as well as G2; (iii) the G1 block can be partially overcome by stimulation with 20% fetal calf serum, but the G2 block cannot be overcome; (iv) 20% fetal calf serum induces c-fos, but c-myc is unaltered; and (v) fibronectin and p21(Waf1/Cip1/Sdi1) are upregulated upon growth arrest. These results suggest that T-antigen-immortalized fibroblasts are committed to undergo senescence but are prevented from undergoing this process by T antigen. Inactivation of T antigen removes this block and results in senescence of the cells. Thus, these cell lines may represent a powerful system for study of the molecular basis of entry into senescence.

Sensitivity of p53 lysine mutants to ubiquitin-directed degradation targeted by human papillomavirus E6.

The activity of the p53 tumor suppressor protein is regulated, at least in part, through the stability of the protein. p53 degradation in normal cells is controlled by ubiquitin-dependent proteolysis, and activation of p53 following DNA damage is associated with an increase in the stability of the protein. The human papillomavirus-encoded E6 protein abrogates p53 function by targeting it for rapid degradation, also through the ubiquitin pathway. Although the p53 protein is ubiquitinated following interaction with E6, we show here that none of the lysine residues within p53 are specifically required for E6-targeted degradation. Mutation of lysine residues within the C-terminus of p53 resulted in resistance to E6-mediated degradation in vitro, although the ability of the two proteins to form a complex was not affected. The same mutant was efficiently targeted for degradation in cells, however, illustrating a lack of correlation between the in vitro and the in vivo assays.

Specific loss of apoptotic but not cell-cycle arrest function in a human tumor derived p53 mutant.

The p53 tumor-suppressor gene product is frequently inactivated in malignancies by point mutation. Although most tumor-derived p53 mutants show loss of sequence specific transcriptional activation, some mutants have been identified which retain this activity. One such mutant, p53175P, is defective for the suppression of transformation in rodent cells, despite retaining the ability to suppress the growth of p53-null human cells. We now demonstrate that p53175P can induce a cell-cycle arrest in appropriate cell types but shows loss of apoptotic function. Our results therefore support a direct role of p53 transcriptional activation in mediating a cell-cycle arrest and demonstrate that such activity is not sufficient for the full apoptotic response. These data suggest that either p53 can induce apoptosis through a transcriptionally independent mechanism, a function lost by p53175P, or that this mutant has specifically lost the ability to activate genes which contribute to cell death, despite activation of genes responsible for the G1 arrest. This dissociation of the cell-cycle arrest and apoptotic activities of p53 indicates that inactivation of p53 apoptotic function without concomitant loss of growth inhibition can suffice to relieve p53-dependent tumor-suppression in vivo and thereby contribute to tumor development.

Retinoblastoma gene deletions in human glioblastomas.

The retinoblastoma susceptibility gene, RB, is the best characterised of the tumour suppressor genes, or 'anti-oncogenes'. Abnormal function of the RB protein is thought to result in loss of an inhibitory effect on cell growth, and thus contribute towards the development of certain human cancers. One group of human cancers of particular interest in relationship to retinoblastoma gene function are the gliomas, which are central nervous system tumours thought to originate from the neuroectoderm, the embryological tissue which also gives rise to retinoblastomas. We have therefore examined a group of benign and malignant gliomas for evidence of structural alterations of the RB gene. Four out of nine (44%) glioblastomas, the most malignant gliomas, showed loss of heterozygosity of a locus within this gene. In addition, one of these hemizygous tumours showed deletion of part of the RB protein-coding region, and this abnormality was also present in cells cultured from the tumour. These findings suggest that RB gene abnormalities may contribute to the development of glioblastomas.

cDNA cloning of porcine transforming growth factor-beta 1 mRNAs. Evidence for alternate splicing and polyadenylation.

Most eukaryotic cells encode principally a 2.5-kilobase (kb) transforming growth factor (TGF)-beta 1 mRNA. However, we have found two major TGF-beta 1 RNA species, 3.5 and 2.5 kb long, in porcine tissues. The 3.5-kb species has a longer 3'-untranslated sequence generated by the selection of an alternate polyadenylation site. There is a 117-nucleotide sequence within this unique 3' region, which is similar to the PRE-1 repetitive sequence of unknown function, reported earlier in the porcine genome. We have also cloned and characterized an alternately spliced mRNA species specific for the TGF-beta 1 gene, in which exons IV and V of the corresponding human TGF-beta 1 gene are deleted. The nucleotide sequence of this cDNA clone predicts a putative precursor protein of 256 amino acids; the N-terminal 211 amino acids of this putative protein are identical to the TGF-beta 1 precursor protein (exons I, II, and III of the human TGF-beta 1 gene), but the C-terminal 45 amino acids are distinct, due to a frameshift in the translation of exons VI and VII. In addition we provide data for the existence of other mRNA species generated in a tissue-specific manner either by alternate splicing or by heterogeneous 5' leader sequences.

Complementary deoxyribonucleic acid cloning of bovine transforming growth factor-beta 1.

Transforming growth factor-beta 1 (TGF beta 1) has been purified from a number of different sources and has a broad species specificity. To deduce the complete amino acid sequence of bovine TGF beta 1 we have isolated cDNA clones encoding the protein from a bovine fibropapilloma library using a human cDNA probe. Sequence analysis of two independent cDNA clones revealed that the 112 amino acids corresponding to bovine TGF beta 1 are identical to those of the human and porcine proteins. This unusually high degree of conservation in the primary structure of the human and bovine proteins reflects the strong evolutionary constraints for maintenance of structure and function of the molecule. As in the human, murine, and porcine systems, the mature form of TGF beta 1 is derived by proteolytic cleavage of a larger precursor. Small differences in amino acid sequence were observed in the portion of the precursor that does not include mature TGF beta 1, although 92% of the residues are still conserved. A 2.25 kilobase (kb) mRNA was identified in total bovine wart and bone RNA, whereas no message was detected in polyadenylated spleen or brain RNA. In addition to the major 2.25 kb message, we observed a 1.9 kb transcript in poly(A+) RNA from wart tissue.

Sacrococcygeal chordoma. A clinicoradiological study of 60 patients.

Sixty patients with sacrococcygeal chordoma, who were seen at this center between 1946 and 1985, were studied with particular attention to the radiographic findings. This study was undertaken because of the large number of these cases and comparison was made between the plain films available in 39 patients and the computed tomography CT studies in 22. Bone destruction was found in 78% on plain films but in 90% on CT. A soft tissue mass was identified in plain films in 60% but in 90% on CT. Calcific debris was found in plain films in 44% but in 87% on CT. Mostly the debris consisted of coarse irregular fragments and probably represented sequestrated necrotic bone. Myelography was performed in only 15 patients. Angiography was studied in 10 cases. Of the 60 patients 88% underwent surgical resection. The tumor recurred in 80% and in only 20% was there no evidence of recurrence. Distant metastases occurred in 24% of patients. Fifty percent survived 5 years; 28% survived 10 years; mean survival 7.5 years.