- Email: [email protected]
ΔNp53 or p44: priming the p53 pump
The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
Molecules in focus
Np53 or p44: priming the p53 pump Heidi Scrablea,∗ , Tsutomu Sasakia,b , Bernhard Maiera a
b
Department of Neuroscience, University of Virginia, Room 6116, MR-4, Lane Road Extended, Charlottesville, VA 22908-1392, USA Neuroscience Graduate Program, University of Virginia, P.O. Box 801392, Charlottesville, VA 22908, USA Received 3 August 2004; received in revised form 15 October 2004; accepted 4 November 2004
Abstract The human protein Np53 and its murine counterpart p44 are isoforms of the tumor suppressor p53 lacking the transactivation domain present in the first 39 (40 in mouse) amino acids of the full-length protein. This makes them similar in structure to the N isoforms of the other members of the p53 superfamily of transcription factors, p63 and p73. The principle way both the human and the murine proteins are generated is by alternative translation of the p53 mRNA utilizing a start site in exon 4. Choice of start site depends on an interaction between p53 and its cognate RNA. When the balance between Np53 (p44) and full-length p53 is altered, the function of p53 as a transcription factor is disturbed. One consequence of over-expressing p44 in mice is an acceleration of the aging process and altered expression of genes in the IGF-1 signaling cascade [Maier, B., Gluba, W., Bernier, B., Turner, T., Mohammad, K., Guise, T., et al. (2004). Modulation of mammalian lifespan by the short isoform of p53. Genes & Development, 18, 306–319]. This links p53 to the single most important growth factor pathway known to regulate lifespan in lower organisms. © 2004 Elsevier Ltd. All rights reserved. Keywords: p47; N isoform; p53/47; Longevity gene; Lifespan; Aging
1. Introduction p53 is a transcription factor that regulates the cell cycle and apoptosis. Since its discovery in 1979, two more members have been added to the p53 superfamily, namely p63 and p73. Both p63 and p73 have several N isoforms, which lack the transactivator domains ∗ Corresponding author. Tel.: +1 434 982 1416; fax: +1 434 982 4380. E-mail address: [email protected] (H. Scrable).
1357-2725/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocel.2004.11.014
and vary in their carboxy termini (reviewed in Davis & Dowdy, 2001; Little & Jochemsen, 2002). The single N isoform of p53 known so far was first discovered in 1985 in isolates of spleen cells from Friend virusinfected mice as a truncated protein encoded by a mutant p53 allele missing exon 2 (Mowat, Cheng, Kimura, Bernstein, & Benchimol, 1985). It was rediscovered in 2002 as a naturally occurring isoform of p53 expressed in normal cells of human and mouse origin (Courtois et al., 2002; Yin, Stephen, Luciani, & Fahraeus, 2002).
914
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
Fig. 1. p53 gene structure, mRNA, and protein. (A) Full length p53 and the two alternative translation start sites. (B) The mutant allele encoding p44. A 3 kb deletion removed exon 2 and flanking intron sequences. Only the downstream translation start site remains. (C) Domain structure of the p53 protein (see text for description). Amino acid positions are given for murine p53. The corresponding positions in human p53 are (left to right): Ser15, Ser20, K320, K370, K372, K373, K381, K382, and K386. Exons, boxes; introns, connecting lines. Color-coding of exons in mRNAs corresponds to matching domains in the protein.
2. Structure The p53 gene comprises 11 exons with the translation start site for a 53 kDa protein located in exon 2 (Fig. 1A). Np53 (p44) is also encoded by the p53 locus, and uses an alternative translation start site located in exon 4 at codon 40 in human RNA (mouse codon 41). The resultant 44 kDa protein lacks the corresponding N-terminal amino acids (Fig. 1B). As depicted in Fig. 1C, the p53 protein has several functional domains (reviewed in Courtois, de Fromentel, & Hainaut, 2004). The N-terminal region contains the activator domains AD1 and AD2, as well as the binding site for Mdm2, which overlaps with AD1. Mdm2, the main regulator of p53 protein stability, is discussed below. The activator domains are followed by a prolinerich domain (PRD), with complex roles as a proteinbinding site and as a specific regulator of apoptosis
(Courtois et al.). The DNA-binding domain (DBD) provides sequence-specific binding capacity and the tetramerization domain (TD) is necessary for formation of the p53 tetramer, which is the functional form of the p53 transcription factor. The C-terminal amino acids comprise the basic regulatory domain (BD), a region containing multiple lysine residues, which are targets for both acetylation and ubiquitination. Accordingly, the use of the alternative translation start site in exon 4 results in loss of AD1, with retention of the remaining functional domains, in Np53.
3. Synthesis and degradation Like the N isoforms of its family members p63 and p73, Np53 appears to be regulated at both the transcriptional and translational levels. Early work identi-
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
fied an alternative transcript for p53 (p53EII) present at approximately 5% the level of the full-length transcript in normal human foreskin fibroblasts (Matlashewski, Pim, Banks, & Crawford, 1987). p53EII has recently been found to be associated with poly-ribosomes and is capable of expressing Np53 (Ghosh, Stewart, & Matlashewski, 2004). Normal mouse cells also express a shorter p53 transcript at low level that could encode the homologous murine p44, which we have detected using a very sensitive ribonuclease protection assay (Maier et al., 2004). There is accumulating evidence, however, that the expression of Np53 is determined mainly by post-transcriptional mechanisms that utilize the N-terminus of p53 and preferentially regulate the translation and degradation of the full-length isoform. Transfection of full-length human or mouse p53 cDNA into p53-deficient cell lines results in the appearance of both the full-length and N protein isoforms (Gannon & Lane, 1991; Mosner et al., 1995; Courtois et al., 2002; Yin et al., 2002), and in-frame mutation of human p53 cDNA at codon Met40 (which is conserved in human, mouse and rat) can completely abolish the expression of Np53 isoforms. Most importantly, translation of both isoforms is attenuated by sequences in the 5 region (5 UTR) of the p53 mRNA (MokdadGargouri, Belhadj, & Gargouri, 2001). In yeast strains constructed with human p53 cDNAs without the 5 UTR, there was robust expression of full-length p53 and much lower levels of the shorter isoform (MokdadGargouri et al.). In the presence of the UTR, however, very little translation of p53 occurred at all, and what little could be detected was exclusively the short form. The authors concluded that p53 translation was inhibited by a mechanism that involved the 5 UTR and they proposed a model in which the 5 UTR interacted with the coding region of the mRNA to block translation. However, within this purely mRNA secondary structure-based model, it was difficult to explain the presence of the short isoform, which seemed to be independent of the 5 UTR. An alternate model of how p53 translation is regulated is that the effect of the 5 UTR on the translation of p53 is mediated by p53 itself. In addition to its well-known DNA binding properties, p53 can bind to RNA, and has been shown to bind to the 5 UTRs of the CDK4 and FGF-2 mRNAs, among others (reviewed in Cassiday & Maher, 2002). Mosner et al. found that translation of both a short and a full-
915
length form of p53 is attenuated by the interaction of the protein with a stem–loop structure in the 5 UTR that overlaps with 280 nucleotides of coding sequence (Mosner et al., 1995). However, because stabilization of this stem–loop structure by p53 requires the N-terminus of the protein, only full-length p53 would be capable of translational regulation via this mechanism. Similarly, both thymidylate synthase (TS) and dihydrofolate reductase (DHFR) regulate translation of their cognate mRNAs by binding to regions near the translation initiation site. In the case of DHFR, this region contains a putative stem–loop structure similar to that found in the p53 5 UTR (Chu, Takimoto, Voeller, Grem, & Allegra, 1993). In the presence of its substrates (dihydrofolate and NADPH), which bind to it more strongly than this RNA stem–loop structure, the enzyme undergoes a conformational change and releases the mRNA, which can then be translated. This brings us to the question of how the block to p53 translation imposed by the 5 UTR might be relieved. Surprisingly, this appears to involve Mdm2, a protein long associated with the ubiquitination and degradation of p53 (reviewed in Michael & Oren, 2003). When human p53 cDNA without either the 5 or 3 UTR was introduced into human cells, Mdm2 modulated the use of the alternate translation initiation sites by interacting with the N-terminus of the p53 protein (Yin et al., 2002). When introduced into yeast cells, however, which have no Mdm2 ortholog, the choice of initiation site depended only on the 5 UTR (Mokdad-Gargouri et al., 2001). p53 initiated at the upstream ATG in the absence of the 5 UTR and at the downstream ATG in its presence. These results demonstrate that, while sequences in the 5 UTR are both necessary and sufficient for significant inhibition of translation of fulllength p53, an additional factor is required to initiate translation of p53 and for high-level p53 expression. In mammalian cells, this factor appears to be Mdm2. Of course, in addition to these effects on p53 translation, Mdm2 enzymatically adds ubiquitin moieties to Lys residues in the carboxy terminus of the protein (Fig. 1C) and targets it for degradation by the proteasome (reviewed in Michael & Oren, 2003). Because the E3 ubiquitin ligase function of Mdm2 requires the presence of the N-terminus of p53, which contains the Mdm2 binding site, the N isoform is not subject to Mdm2-mediated degradation except when it is bound to full-length p53 monomers within the same tetrameric
916
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
complex (Inoue, Geyer, Howard, Yu, & Maki, 2001). Because this method is relatively inefficient, however, p44 has a prolonged half-life (9.5 h) (Lavigueur et al., 1989), compared to that of full-length p53 (0.5 h). The difference between Mdm2-mediated synthesis of p53, which affects both isoforms, and Mdm2-mediated degradation, which preferentially affects the full-length isoform, shifts the ratio of full-length to N isoforms in favor of Np53/p44. It is this shift we have associated with changes in p53 function that alter longevity.
4. Biological function The inability of p44 to bind to p53 mRNA or to Mdm2 because it lacks the N-terminal region of p53 necessary for these interactions may provide a clue to the biological function of the short form of p53. We propose the following model. p44 might be present at a very low level at all times in order to “prime” bursts of p53 expression, allowing it to respond rapidly to situations that require p53 activity. Low level of p44 expression arises from a block of translation initiated at the ATG codon at position 1, which is imposed when the N-terminus of full-length p53 binds to p53 mRNA (Fig. 2A). In this case, translation would initiate at the ATG at position 41 and generate p44. In the presence of Mdm2, however, the association between the stem–loop structure in the p53 mRNA and the Nterminus of p53 protein would be disrupted. Freed of bound p53, the stem–loop structure would be destabilized, allowing the ribosome to access the upstream initiation codon and resulting in a burst of full-length p53 expression (Fig. 2B). The burst would be terminated when p53 is ubiquitinated by Mdm2 and degraded by the proteosome. As long as p44 is expressed at a low level, the cycle can reset itself in preparation for the next burst of p53 activity. When p44 is present in vast excess, however, it can interfere with the functions normally associated with p53, as many biochemical experiments over the past decade have shown (for example, see Unger, Mietz, Scheffner, Yee, & Howley, 1993; Chan, Siu, Lau, & Poon, 2004). As a general rule, the effect of the N form on transactivation depends on the levels of the two isoforms, with the greatest effect being seen in the complete absence of full-length p53. When the effect of increased p44 to p53 was analyzed systematically,
the lowest amount of Np53 resulted in a positive effect on transactivation, while higher doses of the short form led to a negative effect (Chan et al.). With a large increase in Np53, the ability of p53 to trans-activate target genes such as p21Cip1 (Zhu, Zhou, Jiang, & Chen, 1998; Courtois et al., 2002; Yin et al., 2002) is weakened or even lost. With only a small increase in the Np53 isoform, such as in p44 transgenic animals, however, the ability of p53 to trans-activate p21 is actually enhanced (Maier, 2004). Because p53 exhibits a stronger affinity for the short isoform than for itself (Ghosh et al., 2004), a likely explanation for these discrepancies is that the level of p44 affects its subcellular localization, and thus the activities both of itself and of p53. At low level, Np53 or p44 would be exclusively found in tetramers with full-length p53 and would exert its main effect on transcription, either by altering the DNA binding capacity of the tetramer or its ability to recruit various co-factors to the promoters of target genes activated or repressed by p53. At high level, however, excess short form would also be present as homo-tetramers or even as non-tetrameric forms, such as monomers or dimers, which could have several different effects. First of all, homotetramers of p44 would be severely compromised in their ability to activate target genes due to the complete absence of Ntermini, which carry the activation domain (AD1) that binds to the basal transcription machinery (Fig. 1C). Secondly, with higher Np53, full-length p53 relocalizes from the nucleus to the cytoplasm (Ghosh et al.), where it obviously cannot act as a transcription factor. Finally, non-tetrameric forms of p44 could replace p53 in transcription-independent (cytoplasmic) activities, such as mitochondrial cytochrome c release, altering the ability of p53 to mediate apoptosis (Mihara et al., 2003). The ability of p44 to alter the trans-activation of p21Cip1 could have profound implications in normal cells, where p21 plays a role in facilitating S-phase transition during the cell cycle (reviewed in Sherr & Roberts, 1999; Li & Blow, 2001). In this role, p21 facilitates the assembly of cyclin–CDK complexes at replication origins by stabilizing them in a transient ternary complex that can then interact with the origin of replication complex. In fact, in cells from p21deficient mice, the assembly of cyclin D1/D2–CDK4 complexes is impaired. As long as the level of p21 is normal (low), this interaction is transient because it
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
917
Fig. 2. Generation and activity of the two isoforms of p53. (A) The N-terminal domain of full-length p53 (red) is shown interacting with the stem–loop structure of the p53 mRNA. This interaction also requires the DNA binding domain of p53, which is shown wrapping around a loop of RNA. The p53 RNA–p53 protein complex prevents the ribosome at ATG(1) from being activated and translating full-length p53. Instead, the active ribosome (yellow) is positioned at ATG(41) to translate p44/Np53. (B) The active ribosome is shown at the upstream translation initiation site ready to translate full-length p53. Low-level translation of p44 in the absence of p53 allows the short isoform to build up, and as it does the level of Mdm2 (green) also increases. The stem–loop is destabilized when the concentration of Mdm2 reaches a level where it can compete successfully for binding with p53. This removes the block at ATG(1) and p53 is translated. The cycle ends when p44, together with p53, is ubiquitinylated and degraded, and begins again with the low level translation of p44 by the mechanism illustrated in (A). (C) Illustrates how increasing amounts of p44 might change the quality of p53 activity at the p21Cip1 promoter. Following phosphorylation and acetylation, p53 (red symbol) can assemble co-activators that interact to facilitate transcription of p21Cip1 (left panel). With a small increase in p44 (red stippled symbol), interactions with co-activators and with DNA are stabilized, possibly through increased acetylation of lysine residues in the basic domain of p53, and transcription of p21Cip1 is up-regulated (middle panel). However, in p44 excess, modifications that require the N-terminus of p53, such as phosphorylation, cannot occur, preventing the assembly of active transcription complexes on the p21Cip1 promoter and down-regulating p21 expression (right panel). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
leads to the phosphorylation and subsequent degradation of p21. This ephemeral nature of p21 prevents it from accumulating and arresting the cell cycle at the G1-S transition. It is possible, therefore, that in a normally cycling cell, a rise in the level of p44 might cause p21 levels to increase to the point where transit into S is facilitated. Too much p21 at the wrong time, however, would obviously disrupt the ability of ternary complexes of cyclin–CDK–p21 to be transient, thus leading to cell cycle arrest and loss of proliferative ability. In p44 transgenic mice, for example, a small increase in p44 results in an accumulation of p21, impaired ability of cells to synthesize DNA and divide, growth retardation, and premature aging (Maier et al., 2004).
5. Possible implication in aging Within a model in which p44 facilitates discrete bursts of p53 expression as cells cycle in and out of mitosis, it becomes understandable how changes in the level of p44 might adversely affect longevity. If these changes disturb the periodicity of p53 activity in normal cells, the temporal disposition of gene products whose promoters are under the control of p53 would also be disturbed. In the case of p21Cip1, this would have profound effects on proliferating cell populations, as has already been observed in mice in which the level of p44 is only slightly elevated (Maier et al., 2004). Because tissue regeneration relies on the proliferation of stem
918
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
and progenitor cells, the ability of damaged or wornout cells to be replaced would be compromised to the extent that cell cycle regulators like p21 are expressed at inappropriate times or inadequate levels during the cell cycle (see Campisi, 2003 for a discussion of this topic). Furthermore, the effects of even slightly elevated p44 levels on cell cycle progression would be compounded by the effects increased p44 would have when incorporated into the tetrameric form of p53, which is necessary for it to function as a trans-activator or trans-repressor of DNA transcription. We have demonstrated, for example, that a small increase in p44 affects the ability of p53 to function as a transcription factor, alters IGF signaling, and results in premature aging phenotypes in the mouse (Maier et al., 2004). This is consistent with a growing body of evidence from experiments in Caenorhabditis elegans and Drosophila melanogaster that loss-of-function mutation in genes that code for insulin/IGF1 signaling pathway components significantly prolong life-span (reviewed in Gems & Partridge, 2001). The Ames and Snell GH-deficient dwarf mice, as well as mice with reduced expression of IGF1 receptor or insulin receptor, also showed significantly increased lifespan (Gems & Partridge). Thus, the insulin/IGF1 axis is linked to longevity in mouse and lower organisms (discussed in Maier et al.) and would be expected to influence human lifespan as well. In conclusion, it is tempting to speculate that the level of the short isoform of p53, which is expressed in normal cycling cells in small amounts, may change over the life-time of mammals and affect the aging process. If so, it may be possible to modulate the level of p44/Np53 to decrease the rate of aging and promote mammalian health-span.
References Campisi, J. (2003). Cancer and ageing: Rival demons? Nature Reviews. Cancer, 3, 339–349. Cassiday, L. A., & Maher, L. J., 3rd. (2002). Having it both ways: Transcription factors that bind DNA and RNA. Nucleic Acids Research, 30, 4118–4126. Chan, W. M., Siu, W. Y., Lau, A., & Poon, R. Y. (2004). How many mutant p53 molecules are needed to inactivate a tetramer? Molecular and Cellular Biology, 24, 3536–3551. Chu, E., Takimoto, C. H., Voeller, D., Grem, J. L., & Allegra, C. J. (1993). Specific binding of human dihydrofolate reductase pro-
tein to dihydrofolate reductase messenger RNA in vitro. Biochemistry, 32, 4756–4760. Courtois, S., de Fromentel, C. C., & Hainaut, P. (2004). p53 protein variants: Structural and functional similarities with p63 and p73 isoforms. Oncogene, 23, 631–638. Courtois, S., Verhaegh, G., North, S., Luciani, M. G., Lassus, P., Hibner, U., et al. (2002). DeltaN-p53, a natural isoform of p53 lacking the first transactivation domain, counteracts growth suppression by wild-type p53. Oncogene, 21, 6722–6728. Davis, P. K., & Dowdy, S. F. (2001). p73. The International Journal of Biochemistry & Cell Biology, 33, 935–939. Gannon, J. V., & Lane, D. P. (1991). Protein synthesis required to anchor a mutant p53 protein which is temperature-sensitive for nuclear transport. Nature, 349, 802–806. Gems, D., & Partridge, L. (2001). Insulin/IGF signalling and ageing: Seeing the bigger picture. Current Opinion in Genetics & Development, 11, 287–292. Ghosh, A., Stewart, D., & Matlashewski, G. (2004). Regulation of human p53 activity and cell localization by alternative splicing. Molecular and Cellular Biology, 24, 7987–7997. Inoue, T., Geyer, R. K., Howard, D., Yu, Z. K., & Maki, C. G. (2001). MDM2 can promote the ubiquitination, nuclear export, and degradation of p53 in the absence of direct binding. The Journal of Biological Chemistry, 276, 45255–45260. Lavigueur, A., Maltby, V., Mock, D., Rossant, J., Pawson, T., & Bernstein, A. (1989). High incidence of lung, bone, and lymphoid tumors in transgenic mice overexpressing mutant alleles of the p53 oncogene. Molecular and Cellular Biology, 9, 3982–3991. Li, A., & Blow, J. J. (2001). The origin of CDK regulation. Nature Cell Biology, 3, E182–E184. Little, N. A., & Jochemsen, A. G. (2002). p63. The International Journal of Biochemistry and Cell Biology, 34, 6–9. Maier, B., Gluba, W., Bernier, B., Turner, T., Mohammad, K., Guise, T., et al. (2004). Modulation of mammalian life span by the short isoform of p53. Genes & Development, 18, 306–319. Matlashewski, G., Pim, D., Banks, L., & Crawford, L. (1987). Alternative splicing of human p53 transcripts. Oncogene Research, 1, 77–85. Michael, D., & Oren, M. (2003). The p53-Mdm2 module and the ubiquitin system. Seminars in Cancer Biology, 13, 49–58. Mihara, M., Erster, S., Zaika, A., Petrenko, O., Chittenden, T., Pancoska, P., et al. (2003). p53 has a direct apoptogenic role at the mitochondria. Molecules and Cells, 11, 577–590. Mokdad-Gargouri, R., Belhadj, K., & Gargouri, A. (2001). Translational control of human p53 expression in yeast mediated by 5 -UTR-ORF structural interaction. Nucleic Acids Research, 29, 1222–1227. Mosner, J., Mummenbrauer, T., Bauer, C., Sczakiel, G., Grosse, F., & Deppert, W. (1995). Negative feedback regulation of wild-type p53 biosynthesis. The EMBO Journal, 14, 4442–4449. Mowat, M., Cheng, A., Kimura, N., Bernstein, A., & Benchimol, S. (1985). Rearrangements of the cellular p53 gene in erythroleukaemic cells transformed by Friend virus. Nature, 314, 633–636. Sherr, C. J., & Roberts, J. M. (1999). CDK inhibitors: Positive and negative regulators of G1-phase progression. Genes & Development, 13, 1501–1512.
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919 Unger, T., Mietz, J. A., Scheffner, M., Yee, C. L., & Howley, P. M. (1993). Functional domains of wild-type and mutant p53 proteins involved in transcriptional regulation, transdominant inhibition, and transformation suppression. Molecular and Cellular Biology, 13, 5186–5194. Yin, Y., Stephen, C. W., Luciani, M. G., & Fahraeus, R. (2002). p53 Stability and activity is regulated by Mdm2-mediated induction
919
of alternative p53 translation products. Nature Cell Biology, 4, 462–467. Zhu, J., Zhou, W., Jiang, J., & Chen, X. (1998). Identification of a novel p53 functional domain that is necessary for mediating apoptosis. The Journal of Biological Chemistry, 273, 13030–13036.
Molecules in focus
Np53 or p44: priming the p53 pump Heidi Scrablea,∗ , Tsutomu Sasakia,b , Bernhard Maiera a
b
Department of Neuroscience, University of Virginia, Room 6116, MR-4, Lane Road Extended, Charlottesville, VA 22908-1392, USA Neuroscience Graduate Program, University of Virginia, P.O. Box 801392, Charlottesville, VA 22908, USA Received 3 August 2004; received in revised form 15 October 2004; accepted 4 November 2004
Abstract The human protein Np53 and its murine counterpart p44 are isoforms of the tumor suppressor p53 lacking the transactivation domain present in the first 39 (40 in mouse) amino acids of the full-length protein. This makes them similar in structure to the N isoforms of the other members of the p53 superfamily of transcription factors, p63 and p73. The principle way both the human and the murine proteins are generated is by alternative translation of the p53 mRNA utilizing a start site in exon 4. Choice of start site depends on an interaction between p53 and its cognate RNA. When the balance between Np53 (p44) and full-length p53 is altered, the function of p53 as a transcription factor is disturbed. One consequence of over-expressing p44 in mice is an acceleration of the aging process and altered expression of genes in the IGF-1 signaling cascade [Maier, B., Gluba, W., Bernier, B., Turner, T., Mohammad, K., Guise, T., et al. (2004). Modulation of mammalian lifespan by the short isoform of p53. Genes & Development, 18, 306–319]. This links p53 to the single most important growth factor pathway known to regulate lifespan in lower organisms. © 2004 Elsevier Ltd. All rights reserved. Keywords: p47; N isoform; p53/47; Longevity gene; Lifespan; Aging
1. Introduction p53 is a transcription factor that regulates the cell cycle and apoptosis. Since its discovery in 1979, two more members have been added to the p53 superfamily, namely p63 and p73. Both p63 and p73 have several N isoforms, which lack the transactivator domains ∗ Corresponding author. Tel.: +1 434 982 1416; fax: +1 434 982 4380. E-mail address: [email protected] (H. Scrable).
1357-2725/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocel.2004.11.014
and vary in their carboxy termini (reviewed in Davis & Dowdy, 2001; Little & Jochemsen, 2002). The single N isoform of p53 known so far was first discovered in 1985 in isolates of spleen cells from Friend virusinfected mice as a truncated protein encoded by a mutant p53 allele missing exon 2 (Mowat, Cheng, Kimura, Bernstein, & Benchimol, 1985). It was rediscovered in 2002 as a naturally occurring isoform of p53 expressed in normal cells of human and mouse origin (Courtois et al., 2002; Yin, Stephen, Luciani, & Fahraeus, 2002).
914
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
Fig. 1. p53 gene structure, mRNA, and protein. (A) Full length p53 and the two alternative translation start sites. (B) The mutant allele encoding p44. A 3 kb deletion removed exon 2 and flanking intron sequences. Only the downstream translation start site remains. (C) Domain structure of the p53 protein (see text for description). Amino acid positions are given for murine p53. The corresponding positions in human p53 are (left to right): Ser15, Ser20, K320, K370, K372, K373, K381, K382, and K386. Exons, boxes; introns, connecting lines. Color-coding of exons in mRNAs corresponds to matching domains in the protein.
2. Structure The p53 gene comprises 11 exons with the translation start site for a 53 kDa protein located in exon 2 (Fig. 1A). Np53 (p44) is also encoded by the p53 locus, and uses an alternative translation start site located in exon 4 at codon 40 in human RNA (mouse codon 41). The resultant 44 kDa protein lacks the corresponding N-terminal amino acids (Fig. 1B). As depicted in Fig. 1C, the p53 protein has several functional domains (reviewed in Courtois, de Fromentel, & Hainaut, 2004). The N-terminal region contains the activator domains AD1 and AD2, as well as the binding site for Mdm2, which overlaps with AD1. Mdm2, the main regulator of p53 protein stability, is discussed below. The activator domains are followed by a prolinerich domain (PRD), with complex roles as a proteinbinding site and as a specific regulator of apoptosis
(Courtois et al.). The DNA-binding domain (DBD) provides sequence-specific binding capacity and the tetramerization domain (TD) is necessary for formation of the p53 tetramer, which is the functional form of the p53 transcription factor. The C-terminal amino acids comprise the basic regulatory domain (BD), a region containing multiple lysine residues, which are targets for both acetylation and ubiquitination. Accordingly, the use of the alternative translation start site in exon 4 results in loss of AD1, with retention of the remaining functional domains, in Np53.
3. Synthesis and degradation Like the N isoforms of its family members p63 and p73, Np53 appears to be regulated at both the transcriptional and translational levels. Early work identi-
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
fied an alternative transcript for p53 (p53EII) present at approximately 5% the level of the full-length transcript in normal human foreskin fibroblasts (Matlashewski, Pim, Banks, & Crawford, 1987). p53EII has recently been found to be associated with poly-ribosomes and is capable of expressing Np53 (Ghosh, Stewart, & Matlashewski, 2004). Normal mouse cells also express a shorter p53 transcript at low level that could encode the homologous murine p44, which we have detected using a very sensitive ribonuclease protection assay (Maier et al., 2004). There is accumulating evidence, however, that the expression of Np53 is determined mainly by post-transcriptional mechanisms that utilize the N-terminus of p53 and preferentially regulate the translation and degradation of the full-length isoform. Transfection of full-length human or mouse p53 cDNA into p53-deficient cell lines results in the appearance of both the full-length and N protein isoforms (Gannon & Lane, 1991; Mosner et al., 1995; Courtois et al., 2002; Yin et al., 2002), and in-frame mutation of human p53 cDNA at codon Met40 (which is conserved in human, mouse and rat) can completely abolish the expression of Np53 isoforms. Most importantly, translation of both isoforms is attenuated by sequences in the 5 region (5 UTR) of the p53 mRNA (MokdadGargouri, Belhadj, & Gargouri, 2001). In yeast strains constructed with human p53 cDNAs without the 5 UTR, there was robust expression of full-length p53 and much lower levels of the shorter isoform (MokdadGargouri et al.). In the presence of the UTR, however, very little translation of p53 occurred at all, and what little could be detected was exclusively the short form. The authors concluded that p53 translation was inhibited by a mechanism that involved the 5 UTR and they proposed a model in which the 5 UTR interacted with the coding region of the mRNA to block translation. However, within this purely mRNA secondary structure-based model, it was difficult to explain the presence of the short isoform, which seemed to be independent of the 5 UTR. An alternate model of how p53 translation is regulated is that the effect of the 5 UTR on the translation of p53 is mediated by p53 itself. In addition to its well-known DNA binding properties, p53 can bind to RNA, and has been shown to bind to the 5 UTRs of the CDK4 and FGF-2 mRNAs, among others (reviewed in Cassiday & Maher, 2002). Mosner et al. found that translation of both a short and a full-
915
length form of p53 is attenuated by the interaction of the protein with a stem–loop structure in the 5 UTR that overlaps with 280 nucleotides of coding sequence (Mosner et al., 1995). However, because stabilization of this stem–loop structure by p53 requires the N-terminus of the protein, only full-length p53 would be capable of translational regulation via this mechanism. Similarly, both thymidylate synthase (TS) and dihydrofolate reductase (DHFR) regulate translation of their cognate mRNAs by binding to regions near the translation initiation site. In the case of DHFR, this region contains a putative stem–loop structure similar to that found in the p53 5 UTR (Chu, Takimoto, Voeller, Grem, & Allegra, 1993). In the presence of its substrates (dihydrofolate and NADPH), which bind to it more strongly than this RNA stem–loop structure, the enzyme undergoes a conformational change and releases the mRNA, which can then be translated. This brings us to the question of how the block to p53 translation imposed by the 5 UTR might be relieved. Surprisingly, this appears to involve Mdm2, a protein long associated with the ubiquitination and degradation of p53 (reviewed in Michael & Oren, 2003). When human p53 cDNA without either the 5 or 3 UTR was introduced into human cells, Mdm2 modulated the use of the alternate translation initiation sites by interacting with the N-terminus of the p53 protein (Yin et al., 2002). When introduced into yeast cells, however, which have no Mdm2 ortholog, the choice of initiation site depended only on the 5 UTR (Mokdad-Gargouri et al., 2001). p53 initiated at the upstream ATG in the absence of the 5 UTR and at the downstream ATG in its presence. These results demonstrate that, while sequences in the 5 UTR are both necessary and sufficient for significant inhibition of translation of fulllength p53, an additional factor is required to initiate translation of p53 and for high-level p53 expression. In mammalian cells, this factor appears to be Mdm2. Of course, in addition to these effects on p53 translation, Mdm2 enzymatically adds ubiquitin moieties to Lys residues in the carboxy terminus of the protein (Fig. 1C) and targets it for degradation by the proteasome (reviewed in Michael & Oren, 2003). Because the E3 ubiquitin ligase function of Mdm2 requires the presence of the N-terminus of p53, which contains the Mdm2 binding site, the N isoform is not subject to Mdm2-mediated degradation except when it is bound to full-length p53 monomers within the same tetrameric
916
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
complex (Inoue, Geyer, Howard, Yu, & Maki, 2001). Because this method is relatively inefficient, however, p44 has a prolonged half-life (9.5 h) (Lavigueur et al., 1989), compared to that of full-length p53 (0.5 h). The difference between Mdm2-mediated synthesis of p53, which affects both isoforms, and Mdm2-mediated degradation, which preferentially affects the full-length isoform, shifts the ratio of full-length to N isoforms in favor of Np53/p44. It is this shift we have associated with changes in p53 function that alter longevity.
4. Biological function The inability of p44 to bind to p53 mRNA or to Mdm2 because it lacks the N-terminal region of p53 necessary for these interactions may provide a clue to the biological function of the short form of p53. We propose the following model. p44 might be present at a very low level at all times in order to “prime” bursts of p53 expression, allowing it to respond rapidly to situations that require p53 activity. Low level of p44 expression arises from a block of translation initiated at the ATG codon at position 1, which is imposed when the N-terminus of full-length p53 binds to p53 mRNA (Fig. 2A). In this case, translation would initiate at the ATG at position 41 and generate p44. In the presence of Mdm2, however, the association between the stem–loop structure in the p53 mRNA and the Nterminus of p53 protein would be disrupted. Freed of bound p53, the stem–loop structure would be destabilized, allowing the ribosome to access the upstream initiation codon and resulting in a burst of full-length p53 expression (Fig. 2B). The burst would be terminated when p53 is ubiquitinated by Mdm2 and degraded by the proteosome. As long as p44 is expressed at a low level, the cycle can reset itself in preparation for the next burst of p53 activity. When p44 is present in vast excess, however, it can interfere with the functions normally associated with p53, as many biochemical experiments over the past decade have shown (for example, see Unger, Mietz, Scheffner, Yee, & Howley, 1993; Chan, Siu, Lau, & Poon, 2004). As a general rule, the effect of the N form on transactivation depends on the levels of the two isoforms, with the greatest effect being seen in the complete absence of full-length p53. When the effect of increased p44 to p53 was analyzed systematically,
the lowest amount of Np53 resulted in a positive effect on transactivation, while higher doses of the short form led to a negative effect (Chan et al.). With a large increase in Np53, the ability of p53 to trans-activate target genes such as p21Cip1 (Zhu, Zhou, Jiang, & Chen, 1998; Courtois et al., 2002; Yin et al., 2002) is weakened or even lost. With only a small increase in the Np53 isoform, such as in p44 transgenic animals, however, the ability of p53 to trans-activate p21 is actually enhanced (Maier, 2004). Because p53 exhibits a stronger affinity for the short isoform than for itself (Ghosh et al., 2004), a likely explanation for these discrepancies is that the level of p44 affects its subcellular localization, and thus the activities both of itself and of p53. At low level, Np53 or p44 would be exclusively found in tetramers with full-length p53 and would exert its main effect on transcription, either by altering the DNA binding capacity of the tetramer or its ability to recruit various co-factors to the promoters of target genes activated or repressed by p53. At high level, however, excess short form would also be present as homo-tetramers or even as non-tetrameric forms, such as monomers or dimers, which could have several different effects. First of all, homotetramers of p44 would be severely compromised in their ability to activate target genes due to the complete absence of Ntermini, which carry the activation domain (AD1) that binds to the basal transcription machinery (Fig. 1C). Secondly, with higher Np53, full-length p53 relocalizes from the nucleus to the cytoplasm (Ghosh et al.), where it obviously cannot act as a transcription factor. Finally, non-tetrameric forms of p44 could replace p53 in transcription-independent (cytoplasmic) activities, such as mitochondrial cytochrome c release, altering the ability of p53 to mediate apoptosis (Mihara et al., 2003). The ability of p44 to alter the trans-activation of p21Cip1 could have profound implications in normal cells, where p21 plays a role in facilitating S-phase transition during the cell cycle (reviewed in Sherr & Roberts, 1999; Li & Blow, 2001). In this role, p21 facilitates the assembly of cyclin–CDK complexes at replication origins by stabilizing them in a transient ternary complex that can then interact with the origin of replication complex. In fact, in cells from p21deficient mice, the assembly of cyclin D1/D2–CDK4 complexes is impaired. As long as the level of p21 is normal (low), this interaction is transient because it
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
917
Fig. 2. Generation and activity of the two isoforms of p53. (A) The N-terminal domain of full-length p53 (red) is shown interacting with the stem–loop structure of the p53 mRNA. This interaction also requires the DNA binding domain of p53, which is shown wrapping around a loop of RNA. The p53 RNA–p53 protein complex prevents the ribosome at ATG(1) from being activated and translating full-length p53. Instead, the active ribosome (yellow) is positioned at ATG(41) to translate p44/Np53. (B) The active ribosome is shown at the upstream translation initiation site ready to translate full-length p53. Low-level translation of p44 in the absence of p53 allows the short isoform to build up, and as it does the level of Mdm2 (green) also increases. The stem–loop is destabilized when the concentration of Mdm2 reaches a level where it can compete successfully for binding with p53. This removes the block at ATG(1) and p53 is translated. The cycle ends when p44, together with p53, is ubiquitinylated and degraded, and begins again with the low level translation of p44 by the mechanism illustrated in (A). (C) Illustrates how increasing amounts of p44 might change the quality of p53 activity at the p21Cip1 promoter. Following phosphorylation and acetylation, p53 (red symbol) can assemble co-activators that interact to facilitate transcription of p21Cip1 (left panel). With a small increase in p44 (red stippled symbol), interactions with co-activators and with DNA are stabilized, possibly through increased acetylation of lysine residues in the basic domain of p53, and transcription of p21Cip1 is up-regulated (middle panel). However, in p44 excess, modifications that require the N-terminus of p53, such as phosphorylation, cannot occur, preventing the assembly of active transcription complexes on the p21Cip1 promoter and down-regulating p21 expression (right panel). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
leads to the phosphorylation and subsequent degradation of p21. This ephemeral nature of p21 prevents it from accumulating and arresting the cell cycle at the G1-S transition. It is possible, therefore, that in a normally cycling cell, a rise in the level of p44 might cause p21 levels to increase to the point where transit into S is facilitated. Too much p21 at the wrong time, however, would obviously disrupt the ability of ternary complexes of cyclin–CDK–p21 to be transient, thus leading to cell cycle arrest and loss of proliferative ability. In p44 transgenic mice, for example, a small increase in p44 results in an accumulation of p21, impaired ability of cells to synthesize DNA and divide, growth retardation, and premature aging (Maier et al., 2004).
5. Possible implication in aging Within a model in which p44 facilitates discrete bursts of p53 expression as cells cycle in and out of mitosis, it becomes understandable how changes in the level of p44 might adversely affect longevity. If these changes disturb the periodicity of p53 activity in normal cells, the temporal disposition of gene products whose promoters are under the control of p53 would also be disturbed. In the case of p21Cip1, this would have profound effects on proliferating cell populations, as has already been observed in mice in which the level of p44 is only slightly elevated (Maier et al., 2004). Because tissue regeneration relies on the proliferation of stem
918
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919
and progenitor cells, the ability of damaged or wornout cells to be replaced would be compromised to the extent that cell cycle regulators like p21 are expressed at inappropriate times or inadequate levels during the cell cycle (see Campisi, 2003 for a discussion of this topic). Furthermore, the effects of even slightly elevated p44 levels on cell cycle progression would be compounded by the effects increased p44 would have when incorporated into the tetrameric form of p53, which is necessary for it to function as a trans-activator or trans-repressor of DNA transcription. We have demonstrated, for example, that a small increase in p44 affects the ability of p53 to function as a transcription factor, alters IGF signaling, and results in premature aging phenotypes in the mouse (Maier et al., 2004). This is consistent with a growing body of evidence from experiments in Caenorhabditis elegans and Drosophila melanogaster that loss-of-function mutation in genes that code for insulin/IGF1 signaling pathway components significantly prolong life-span (reviewed in Gems & Partridge, 2001). The Ames and Snell GH-deficient dwarf mice, as well as mice with reduced expression of IGF1 receptor or insulin receptor, also showed significantly increased lifespan (Gems & Partridge). Thus, the insulin/IGF1 axis is linked to longevity in mouse and lower organisms (discussed in Maier et al.) and would be expected to influence human lifespan as well. In conclusion, it is tempting to speculate that the level of the short isoform of p53, which is expressed in normal cycling cells in small amounts, may change over the life-time of mammals and affect the aging process. If so, it may be possible to modulate the level of p44/Np53 to decrease the rate of aging and promote mammalian health-span.
References Campisi, J. (2003). Cancer and ageing: Rival demons? Nature Reviews. Cancer, 3, 339–349. Cassiday, L. A., & Maher, L. J., 3rd. (2002). Having it both ways: Transcription factors that bind DNA and RNA. Nucleic Acids Research, 30, 4118–4126. Chan, W. M., Siu, W. Y., Lau, A., & Poon, R. Y. (2004). How many mutant p53 molecules are needed to inactivate a tetramer? Molecular and Cellular Biology, 24, 3536–3551. Chu, E., Takimoto, C. H., Voeller, D., Grem, J. L., & Allegra, C. J. (1993). Specific binding of human dihydrofolate reductase pro-
tein to dihydrofolate reductase messenger RNA in vitro. Biochemistry, 32, 4756–4760. Courtois, S., de Fromentel, C. C., & Hainaut, P. (2004). p53 protein variants: Structural and functional similarities with p63 and p73 isoforms. Oncogene, 23, 631–638. Courtois, S., Verhaegh, G., North, S., Luciani, M. G., Lassus, P., Hibner, U., et al. (2002). DeltaN-p53, a natural isoform of p53 lacking the first transactivation domain, counteracts growth suppression by wild-type p53. Oncogene, 21, 6722–6728. Davis, P. K., & Dowdy, S. F. (2001). p73. The International Journal of Biochemistry & Cell Biology, 33, 935–939. Gannon, J. V., & Lane, D. P. (1991). Protein synthesis required to anchor a mutant p53 protein which is temperature-sensitive for nuclear transport. Nature, 349, 802–806. Gems, D., & Partridge, L. (2001). Insulin/IGF signalling and ageing: Seeing the bigger picture. Current Opinion in Genetics & Development, 11, 287–292. Ghosh, A., Stewart, D., & Matlashewski, G. (2004). Regulation of human p53 activity and cell localization by alternative splicing. Molecular and Cellular Biology, 24, 7987–7997. Inoue, T., Geyer, R. K., Howard, D., Yu, Z. K., & Maki, C. G. (2001). MDM2 can promote the ubiquitination, nuclear export, and degradation of p53 in the absence of direct binding. The Journal of Biological Chemistry, 276, 45255–45260. Lavigueur, A., Maltby, V., Mock, D., Rossant, J., Pawson, T., & Bernstein, A. (1989). High incidence of lung, bone, and lymphoid tumors in transgenic mice overexpressing mutant alleles of the p53 oncogene. Molecular and Cellular Biology, 9, 3982–3991. Li, A., & Blow, J. J. (2001). The origin of CDK regulation. Nature Cell Biology, 3, E182–E184. Little, N. A., & Jochemsen, A. G. (2002). p63. The International Journal of Biochemistry and Cell Biology, 34, 6–9. Maier, B., Gluba, W., Bernier, B., Turner, T., Mohammad, K., Guise, T., et al. (2004). Modulation of mammalian life span by the short isoform of p53. Genes & Development, 18, 306–319. Matlashewski, G., Pim, D., Banks, L., & Crawford, L. (1987). Alternative splicing of human p53 transcripts. Oncogene Research, 1, 77–85. Michael, D., & Oren, M. (2003). The p53-Mdm2 module and the ubiquitin system. Seminars in Cancer Biology, 13, 49–58. Mihara, M., Erster, S., Zaika, A., Petrenko, O., Chittenden, T., Pancoska, P., et al. (2003). p53 has a direct apoptogenic role at the mitochondria. Molecules and Cells, 11, 577–590. Mokdad-Gargouri, R., Belhadj, K., & Gargouri, A. (2001). Translational control of human p53 expression in yeast mediated by 5 -UTR-ORF structural interaction. Nucleic Acids Research, 29, 1222–1227. Mosner, J., Mummenbrauer, T., Bauer, C., Sczakiel, G., Grosse, F., & Deppert, W. (1995). Negative feedback regulation of wild-type p53 biosynthesis. The EMBO Journal, 14, 4442–4449. Mowat, M., Cheng, A., Kimura, N., Bernstein, A., & Benchimol, S. (1985). Rearrangements of the cellular p53 gene in erythroleukaemic cells transformed by Friend virus. Nature, 314, 633–636. Sherr, C. J., & Roberts, J. M. (1999). CDK inhibitors: Positive and negative regulators of G1-phase progression. Genes & Development, 13, 1501–1512.
H. Scrable et al. / The International Journal of Biochemistry & Cell Biology 37 (2005) 913–919 Unger, T., Mietz, J. A., Scheffner, M., Yee, C. L., & Howley, P. M. (1993). Functional domains of wild-type and mutant p53 proteins involved in transcriptional regulation, transdominant inhibition, and transformation suppression. Molecular and Cellular Biology, 13, 5186–5194. Yin, Y., Stephen, C. W., Luciani, M. G., & Fahraeus, R. (2002). p53 Stability and activity is regulated by Mdm2-mediated induction
919
of alternative p53 translation products. Nature Cell Biology, 4, 462–467. Zhu, J., Zhou, W., Jiang, J., & Chen, X. (1998). Identification of a novel p53 functional domain that is necessary for mediating apoptosis. The Journal of Biological Chemistry, 273, 13030–13036.