DNA binding specificity of proteins derived from alternatively spliced mouse p53 mRNAs

p53 Isoforms: An Intracellular Microprocessor?

Normal function of the p53 pathway is ubiquitously lost in cancers either through mutation or inactivating interaction with viral or cellular proteins. However, it is difficult in clinical studies to link p53 mutation status to cancer treatment and clinical outcome, suggesting that the p53 pathway is not fully understood. We have recently reported that the human p53 gene expresses not only 1 but 12 different p53 proteins (isoforms) due to alternative splicing, alternative initiation of translation, and alternative promoter usage. p53 isoform proteins thus contain distinct protein domains. They are expressed in normal human tissues but are abnormally expressed in a wide range of cancer types. We have recently reported that p53 isoform expression is associated with breast cancer prognosis, suggesting that they play a role in carcinogenesis. Indeed, the cellular response to damages can be switched from cell cycle arrest to apoptosis by only manipulating p53 isoform expression. This may provide an explanation to the hitherto inconsistent relationship between p53 mutation, treatment response, and outcome in breast cancer. However, the molecular mechanism is still unknown. Recent reports suggest that it involves modulation of gene expression in a p53-dependent and -independent manner. In this review, we summarize our current knowledge about the biological activities of p53 isoforms and propose a molecular mechanism conciliating our current knowledge on p53 and integrating p63 and p73 isoforms in the p53 pathway.

The over-expression of p53 H179Y residue mutation causes the increase of cyclin A1 and Cdk4 expression in HELF cells

Down-regulation of p53 expression has been found in a broad range of human cancers and cell proliferation disorders, indicating that p53 plays a key role in cell cycle regulation and tumor suppression. In our current study, we transfected human embryonic lung fibroblast (HELF) cells with pcDNA3-wild-type p53 (pcDNA3-wtp53) plasmid, or pcDNA3-H179Y-mutated p53 (pcDNA3-mtp53) plasmid that mimics the mutation found in some human lung tumors, and further studied the role of p53 in the regulation of cell proliferation. Over expression of wild-type p53 caused cell cycle arrest at G1 phase with reduced cell size, decreased expression of cyclin D3, cyclin E, Cdk2 and Cdk4, and increased expression of p21. In contrast, over expression of H179Y-mutant p53 promoted G1 to S phase transition with enlarged cell size and increased cyclin A1 and Cdk4 expression in HELF cells. These results indicate that mutation at the p53 H179Y residue up-regulates cyclin A1 and Cdk4 expression, and promotes HELF cell proliferation.

Recognition of RNA by the p53 tumor suppressor protein in the yeast three-hybrid system

The p53 tumor suppressor protein is a homotetrameric transcription factor whose gene is mutated in nearly half of all human cancers. In an unrelated screen of RNA/protein interactions using the yeast three-hybrid system, we inadvertently detected p53 interactions with several different RNAs. A literature review revealed previous reports of both sequence-specific and -non-specific interactions between p53 and RNA. Using yeast three-hybrid selections to identify preferred RNA partners for p53, we failed to identify primary RNA sequences or obvious secondary structures required for p53 binding. The cationic p53 C-terminus was shown to be required for RNA binding in yeast. We show that while p53 strongly discriminates between certain RNAs in the yeast three-hybrid assay, the same RNAs are bound equally by p53 in vitro. We further show that the p53 RNA-binding preferences in yeast are mirrored almost exactly by a recombinant tetrameric form of the HIV-1 nucleocapsid (NC) protein thought to be a sequence-nonspecific RNA-binding protein. However, the possibility of specific RNA binding by p53 could not be ruled out because p53 and HIV-1 NC displayed certain differences in RNA-binding preference. We conclude that (1) p53 binds RNA in vivo, (2) RNA binding by p53 is largely sequence-nonspecific in the yeast nucleus, (3) some structure-specific RNA binding by p53 cannot be ruled out, and (4) caution is required when interpreting results of RNA screens in the yeast three-hybrid system because sequence-dependent differences in RNA folding and display can masquerade as sequence-dependent differences in protein recognition.

Atomic force microscopy reveals kinks in the p53 response element DNA

p53 is a 53 kDa nuclear phosphoprotein. Its function as a tumor suppressor critically lies in its ability to recognize its target DNA response elements as a tetramer. Here, we report the structural theme intrinsic to the response element DNA that governs this recognition phenomenon. The intrinsic flexibility or dynamic bending between two distinctly different, but naturally occurring p53 response elements has been compared by ring closure. Results show that DNA binding sites containing helically phased d(CATG.CATG) tetra-nucleotide sequences at the centers of quasi-dyad symmetry in each half-response site are more intrinsically flexible (i.e. preferentially bent under axial stress) than their d(CTTG.CTTG) counterparts. Intriguingly, p53 binding sites containing these more flexible d(CATG.CATG) sequence elements also exhibit a stronger tendency for tetrameric binding of the p53 DNA binding domain peptide. Examination of the shapes of DNA microcircles obtained by circularization of oligomers constructed from such flexible p53 target DNA sequences in tandem using MacMode atomic force microscopy directly revealed sequence-specific kinks in solution. The tetra-nucleotide sequence d(CATG.CATG) is highly conserved in most functional p53 response elements. Consequently, we propose that the sequence-specific kinks originating from d(CATG.CATG) sequences could be a common structural theme in p53 response elements and as evident from the results reported here, could be a determinant of binding site recognition by the p53 protein and the subsequent stability of the p53-DNA complex.

p53-dependent apoptosis is regulated by a C-terminally alternatively spliced form of murine p53

It is now well accepted that the p53 C-terminus plays a central role in controlling the activity of the wild-type molecule. In our previous studies, we observed that a C-terminally altered p53 protein (p53AS), generated by an alternative spliced p53 mRNA, induces an attenuated p53-dependent apoptosis, compared to that induced by the regularly spliced form (p53RS). In the present study we analysed the interrelationships between these two physiological variants of wild-type p53, and found that in cells co-expressing both forms, in contrast to the expected additive effect on the induction of apoptosis, p53AS inhibits apoptosis induced by p53RS. This inhibitory effect is specific for p53-dependent apoptosis and was not evident in a p53-independent apoptotic pathway induced by growth factor deprivation. Furthermore, the expression of p53AS in transiently transfected cells caused both inhibition of apoptosis and inhibition of the p53RS-dependent transactivation of a number of p53 target genes. These results suggest that expression of an alternatively spliced p53 form may serve as an additional level in controlling the complexity of p53 function by the C-terminal domain.

Expression in vitro of alternatively spliced variants of the messenger RNA for human apolipoprotein E receptor-2 identified in human tissues by ribonuclease protection assays

The apolipoprotein E receptor-2 (apoER2), also called LR7/8B, is a member of the low-density lipoprotein (LDL)-receptor family that is expressed in brain. We have identified mRNA splicing variants in human tissues by ribonuclease protection assays and found that some variants are preferentially amplified by reverse transcription-polymerase chain reaction (RT-PCR). Transcripts were found that lacked sequences encoding three repeats in the putative ligand-binding domain, the O-linked sugar domain or a novel region in the cytoplasmic domain. When mammalian expression vectors for eight potential protein isoforms were transfected into LDL-receptor-deficient Chinese hamster ovary cells, the proteins were all expressed on the cell surface, as detected by immunoblotting of cell extracts with a specific antipeptide antiserum to apoER2 before and after treatment of intact cells with pronase. Although cells expressing all the variants bound very low-density lipoprotein of beta mobility (beta-VLDL), it was with lower affinity and capacity than binding by the LDL-receptor and none was able to degrade beta-VLDL. Ligand blotting of cell extracts showed that all variants bound recombinant histidine6-tagged receptor-associated protein (His6-RAP) with high affinity, although variants lacking exon 5 bound less strongly. The presence of vestiges of the novel insert in the cytoplasmic domain of apoER2 in the LDL- or VLDL-receptor genes was investigated, but nucleotide sequencing showed that no sequences homologous to it could be detected in the final intron of these genes.

Activities and response to DNA damage of latent and active sequence-specific DNA binding forms of mouse p53

The mouse p53 protein generated by alternative splicing (p53as) has amino acid substitutions at its C terminus that result in constitutively active sequence-specific DNA binding (active form), whereas p53 protein itself binds inefficiently (latent form) unless activated by C-terminal modification. Exogenous p53as expression activated transcription of reporter plasmids containing p53 binding sequences and inhibited growth of mouse and human cells lacking functional endogenous p53. Inducible p53as in stably transfected p53 null fibroblasts increased p21(WAF1/Cip-1/Sdi) and decreased bcl-2 protein steady-state levels. Endogenous p53as and p53 proteins differed in response to cellular DNA damage. p53 protein was induced transiently in normal keratinocytes and fibroblasts whereas p53as protein accumulation was sustained in parallel with induction of p21(WAF1/Cip-1/Sdi) protein and mRNA, in support of p53as transcriptional activity. Endogenous p53 and p53as proteins in epidermal tumor cells responded to DNA damage with different kinetics of nuclear accumulation and efficiencies of binding to a p53 consensus DNA sequence. A model is proposed in which C-terminally distinct p53 protein forms specialize in functions, with latent p53 forms primarily for rapid non-sequence-specific binding to sites of DNA damage and active p53 forms for sustained regulation of transcription and growth.