The many faces of p53: something for everyone

Forty years ago four research laboratories in London, Paris, New York/Bethesda, and Princeton uncovered the existence of the p53 protein (Deleo et al., 1979; Lane and Crawford, 1979; Linzer and Levine, 1979; Kress et al., 1979). Each laboratory came upon this protein for a different reason and with a different experimental approach that uncovered this unanticipated result. Together, the four papers permitted one to conclude the following: (i) in SV40-infected and transformed cells the SV40-encoded oncogene protein, the large T-antigen, formed a protein complex with a cellular-encoded protein of ~53000 daltons in size. (ii) This p53 protein was detected at high levels in a variety of transformed cells derived from viral, chemical, or inherited (teratocarcinomas) transformation events. (iii) Nontransformed cells expressed lower levels of the p53 protein. (iv) Animals bearing tumors produced antibodies directed against the p53 protein. A temperature-sensitive mutation in the SV40 large T-antigen gene (the oncogene of this virus) was employed to demonstrate that the p53–T-antigen complex was formed at the permissive temperature, where the cells are transformed, but not at the nonpermissive temperature, where the cells behave normally (Linzer and Levine, 1979; Linzer et al., 1979). At a later date p53 protein complexes with viral oncogene products were observed, including the adenovirus E1b-58kd protein (Sarnow et al., 1982) and the human papilloma virus E6 oncoprotein (Scheffner, et al. 1990; Werness et al., 1990; which is the cause of human cervical cancers and some head and neck cancers). In order to explore the functions of the p53 protein, several p53 cDNAs were isolated and cloned (Oren and Levine, 1983; Oren et al., 1983; Pennica et al., 1984). These clones were tested for oncogene activities and found to cooperate with the RAS oncogene in transforming embryonic cells (Eliyahu et al., 1984; Parada et al., 1984). Thus, it appeared that the p53 gene was an oncogene whose protein forms a complex with viral oncogene proteins, possibly mediating transformation. However, the cDNA clone isolated by Pennica failed to transform cells in culture and had a single amino acid change when compared with the Oren cDNA clone, which did transform cells. Was the amino acid difference between these clones significant? Was this difference a sequencing mistake? A polymorphism? Or a mutation? If it was a mutation, which clone was the wild-type and which was the mutant? To address these questions, Oren and Levine exchanged clones (and reproduced each other’s observations). By 1989 it became clear that mutations in the p53 cDNA clones resulted in cellular transformation, and wild-type p53 protein prevented transformation and functioned as a tumor suppressor (Eliyahu et al., 1989; Finlay et al., 1989). p53 mutations in both p53 alleles in colon cancers of humans resulted in the same conclusion; p53 functioned as a tumor suppressor gene that helped to prevent cancer (Baker et al., 1990a, b, Nigro et al., 1995). From 1979 to 1989 the p53 protein was alternatively referred to as a fetal antigen expressed in the teratocarcinoma stem cells, a tumor antigen that induced antibodies in animals and humans with tumors, an oncogene whose mutant forms could transform cells, and, finally, a tumor suppressor gene that prevented cancers. During this time the p53 protein was demonstrated to increase its concentration in response to DNA damage (Maltzman and Czyzyk, 1984). Over these first 10 years of research the p53 protein was shown to have many diverse faces and activities, functioning as an oncogene and a tumor suppressor gene while responding to DNA damage in a cell. The mutant forms of p53 add to the many faces of p53. There are multiple ways to inactivate p53 protein functions. There are mutations such as deletions, nonsense mutations, and frameshift mutations, which are all true loss-of-function mutations, but more commonly there are missense mutations localized in the p53 protein DNA-binding domain (Bouaoun et al., 2016). In addition, there are gene amplifications in the MDM-2 gene that overexpress this ubiquitin ligase, which is a negative regulator of the levels of the p53 protein. Thirdly, there are protein modifications that reduce or eliminate p53 transcriptional activity (Zhu et al., 2016; Levine, 2017). To date, >500 different missense mutations, located in the DNA-binding domain of the p53 protein, have been isolated and sequenced from human cancers. Missense mutations that produce high levels of mutant p53 proteins can function as dominant negative proteins, inactivating the wild-type p53 protein, in cells with normal levels of wild-type proteins. This is why mutant p53 cDNA clones transform normal cells in culture. A tetramer of the wild-type p53 protein functions as the transcription factor. Faulty or mutant monomeric proteins contributing to the tetrameric protein poison its activity and act in a dominant negative fashion. A second way in which missense mutant p53 proteins act in cells is to contribute a gain of function to the cell. This is observed when a missense mutant protein is added to a cell with no p53 protein (a deletion mutation). The cell clone expressing the missense protein is then compared to the cell clone with no p53 protein. When this is done the cells expressing a mutant missense p53 protein have new activities that enhance transformation and tumorigenesis (Dittmer et al., 1993). Thus, the mutant protein, by itself, has gained new functions. The 500 plus different missense mutations in the p53 gene that have been observed in cancers occur at different frequencies in all types of human cancers from 7% for some alleles to 0.005% for other mutant alleles. The top 10 mutant alleles account for 33% of human cancer mutations, whereas the top 50 mutant alleles account for ~50% of human cancer mutations. Thus, the frequencies with which different mutant p53 alleles are observed in many human cancers occur over a four log difference in their levels (Balachandran et al., 2017; Levine, 2019). Why? These 500 mutant proteins have a wide variety of structures. There are both mutations that alter amino acids that make a contact with the template DNA sequence, and there are mutations that alter the structure of the p53 DNA-binding domain. Different missense mutants transcribe seven different p53-responsive genes in a yeast cell assay at a wide variety of different efficiencies (Kato et al., 2003; Mathe et al., 2006; Petitjean et al., 2007). The top 10 most frequent mutant p53 alleles largely fail to transcribe p53-responsive genes. Some p53 mutant alleles produce proteins that can act to induce a CD-8 T-cell response in long-term survivors of cancers (Łuksza et al., 2017; Levine, 2018, 2019). The immunological activity of a p53 mutation depends upon the binding efficiency of HLA-A, HLA-B, or HLA-C class 1 receptor alleles and the presence of a T-cell receptor that recognizes the mutant peptide, giving rise to an extraordinary diversity of the immune response to p53 mutant proteins in human populations. The extraordinary diversity of mutant missense p53 alleles, occurring at different frequencies with different phenotypes, and the diversity of the immune system in response to p53 mutant alleles have led Sabapathy and Lane (2018) to term this ‘a rainbow of p53 mutations’ where ‘some mutations are more equal than others’. Does this mean that it will take a rainbow of drugs to treat cancers with p53 mutations (Levine, 2019)? The p53 protein is a transcription factor. It binds to a degenerate DNA sequence (Pu,Pu,Pu,C,A/T,A/T,G, Py,Py,Py) repeated twice with a variable spacer, where Pu is a purine, Py a pyrimidine, and A, T, G, and C are the four usual bases. About 200–300 genes are thought to be regulated at the transcription level by the p53 protein, but this is at best an estimate. Indeed, a mega-analysis of several publications exploring p53-regulated genes has identified only ~60 genes found in common in all the publications. The p53 gene enhances the transcription of the MDM-2 gene, which produces the ubiquitin ligase that promotes the degradation of the p53 protein. This produces an autoregulatory loop so that p53 and MDM-2 proteins oscillate out of phase (Wu et al., 1993). Thus, p53 protein levels are regulated at the posttranslational level (Oren et al., 1982). The half-life of the p53 protein is quite short, between 6 and 20 min. When a cell is exposed to one or many stresses, termed the input signals, posttranslational modifications of the MDM-2 and p53 proteins increase the half-life of the p53 protein and activate it for transcription. Table 1 lists some of the p53 input signals. It is striking that a wide variety of diverse stresses activate the transcriptional activity of the p53 protein. The activated p53 protein responds to different stresses by expressing a number of transcriptional programs, which are output signals of p53-mediated transcription that are listed in Table 2. These outputs may result in quite different outcomes, resulting in cell death or the repair of the consequences of stressful inputs. There are two p53 transactivation domains, and they regulate different sets of genes as outputs (Lin et al., 1994; Raj and Attardi, 2016). In addition to the posttranslational modifications of p53, the oscillations, the alterations in these oscillations, or the lack of oscillations all can influence the output transcriptional programs (Stewart-Ornstein et al., 2017). The output transcriptional program mediated by an activated p53 protein can be influenced by the cell type, a stem or progenitor cell, or a differentiated cell, whether or not the cell is transformed or normal, or even by the nature of the microbiome (bacterial or viral). p53 input signals. XTPs, nucleoside triphosphates; ARF, alternative reading frame; ROS, reactive oxygen species. p53 input signals. XTPs, nucleoside triphosphates; ARF, alternative reading frame; ROS, reactive oxygen species. p53 output signals. p53 output signals. The input and output signal transduction pathways place p53 and MDM-2 at a central hub that extends out to a very large number of diverse functions and other signal transduction pathways (Vogelstein et al., 2000). It has become possible to identify connections among many different signal transduction pathways that carry out a large number of cellular functions creating a network of cellular interactions. In this network the central p53 hub communicates with a large number of inputs (Table 1) and responds with a large number of outputs (Table 2). The p53 hub both receives and transmits information about the intracellular and extracellular environment. It has become possible to calculate the amount of information (entropy) that each hub in the network receives and transmits. It turns out that the p53 hub has the highest information content when compared to all other nodes in the network of a cell (Pouryahya et al., 2018). It is not surprising that the p53 node in the network has maximal entropy in that it deals with the reproduction of the cell and the efficient extraction of energy from nutrients, two important evolutionary properties. Changes in entropy are positively correlated with a notion of robustness and create an architecture where information is processed through multiple cellular hubs (Tannenbaum et al., 2015). Cellular stresses during replication result in an enhanced rate of mutations or mistakes, and under these conditions p53 ensures fidelity by repair or by death. Surely one of the reasons why the p53 hub has control over an extraordinary level of cellular information is that there are so many diverse stresses that input to the p53 hub. Why is this? It would appear that this configuration, where many different types of stress act through a single gene and protein (p53), to ensure a response, would create a vulnerable node, liable to failure if that gene is nonfunctional via a mutation. Why not build a network with 10 different stress responders for 10 stresses and then connect them together? The most common explanation for the existing configuration (one stress responder) is that two or more different stress signals at the same time must be integrated or communicated to result in a proper response. Networks are better at integrating this information at a single node (protein) than by requiring additional connections among nodes that must coordinate the response. If this is correct, then p53 takes on the informational function of coordinating situations where multiple things go wrong by receiving and then integrating the information before programing the right kind of outputs. All of this is consistent with p53 having a position in the network that ensures a higher order of information transfer integrating the signals. This may well be the reason why the p53 gene is the single most common gene to be mutated in human cancer. The higher order function of the p53 gene and protein is to ensure fidelity and homeostasis by integrating responses to stresses. Furthermore, this explanation would suggest that no one or two or even three downstream p53-regulated genes are responsible for tumor suppression by the p53 gene. So, deletion of selected downstream genes without loss of tumor suppression is not a surprise. The p53 circuitry is designed to compensate for that possibility. Rather, tumor suppression is a regulated, integrated, and informed coordinated set of responses by the p53 protein to environmental perturbations resulting in the maintenance of cellular homeostasis. It seems likely that the communication of stress signals to p53 is mediated by protein modifications from the epigenetic activities that also communicate with chromatin, histones, and other transcription factors helping to integrate the information about environmental changes. That confers maximum information content upon the p53 node. It may be of some interest that the set of nodes in the cellular network with the second highest information content is the set of 14-3-3 genes that position various proteins at cellular locations, permitting functions or preventing them. With p53 involved in enforcing homeostasis under conditions of varying environments and 14-3-3 proteins creating a regulated topological framework for cellular proteins, these two sets of nodes act through time and to cellular The gene and protein are regulated in by the p53 protein et al., these The protein induced by p53 in a cell and the activity that an important in and cell in some cancers et al., The amino acid sequence of the human p53 protein DNA-binding domain acids out of is over a of about one years of and Levine, The of derived from the evolutionary of an amino acid sequence with a DNA-binding domain et al., These are among the or most the evolutionary and have been with a DNA-binding a of p53 in both the DNA-binding domain protein and the DNA sequence it binds with are in the from to humans et al., In all of these there is a single gene, which is expressed in the but not the If the DNA of these is by then p53 is activated and the or cells. is by and even the genes in the are from to In of the p53 and cells are This is to in human where cells to et al., Thus, the p53 gene is first observed in fidelity of the DNA under stress and by cells with by diversity in a to to act and to create new that are better to in the environment. The gene to prevent the of diversity through that are not or It is for an to and with many of its At the this is the and some such as A and human and at this that are at this can but function and Thus, the of diversity and ensure A gene with functions p53 has its in the of through and the is largely the cells and in With the of a new The stem cells that cells and many over the of the With the of stem cells and of and the stem cells of the mutations. a stem cell mutations, a for stem cell clones that other stem cells in each At this is functioning both stem cell and cell of an With additional mutations, cancers from these clones and alter and functions of the It is at about this in and when the single gene three p53, and each of which takes on diverse functions. upon the rate and of amino acid in the DNA-binding domain of these three the p53 gene the most and while and genes change only p53 activities from the to the fidelity and homeostasis in stem and progenitor cells. It a tumor suppressor gene and its as the central stress in cells. the p53 gene from to it from the to stem cells. functions largely the same to ensure fidelity by death. its functions of homeostasis and fidelity of the in et al., while a stem cell for The gene in several during and functions in many including cell the central and the immune system et al., 2018). in the p53 gene occur in ~50% of human it the most commonly gene in human cancers. the plus publications about the p53 gene and protein, the of them on cancer. p53 are all about cancers. It did not have to this The reason for a cancer in the p53 was that all four of the papers of the p53 protein were about cancers and that and all the this were cancer the p53 gene and protein function as a stress in Table and these stresses have an upon many and functions that can to many diverse or can even regulate normal functions (Table 2). It is clear that p53, and an important in reproduction et al., The p53 gene a in the of the and the of the It so by the transcription of the gene. is for et al., et al., The sets a between the immune system of the and the is one of the that that to T-cell is by a number of tumors when p53 is and antibodies directed against in tumor by the immune the of to the immune system would be in some of the that the immune of the and of tumors in in the In addition, tumors a number of normal to and and can a from the of p53 in the of cells during immune and 1 and make it clear that the p53 gene is a of the immune It is involved with both the microbiome and the DNA and only the tumor induce p53 activity in cells. have to inactivate the p53 activity that the viral functions. A number of intracellular and that with the cell can activate a p53 response, which can the cell. These activated p53 functions signal to and which the of the microbiome for responses by the immune system et al., 2015). the p53 protein been by the of the p53 would have been quite p53-mediated cell by oncogene or DNA can result in a p53 that cells that the cells and that eliminate the cells. Indeed, one of the of why and with time is that mutations with cells that are then by and cells. the efficiency of the system the cells are not and a of result in and With in The p53 response to this is cell It is not surprising then that high levels of p53 activity in cells can about and 2017). There is an between and alterations in the DNA in cells of the and 2018). At selected in the alterations occur in the levels of at as a function of However, there are between when is compared with which is by the rate of alterations with time and not the number of years The et al., et al., Levine, that p53 a in enforcing the and homeostasis of epigenetic in cells. This is between programs and p53 gene and protein functions. is to a clear of the of p53 in contributing to the program in each and diverse species. Each has a of the number of years a of the will there a p53 function that between and with other functions that contribute to this of the between p53 functions with and time will the of the p53 network with The of most are with the of and that have the in the evolutionary In many there is a time in when and can and the creating and even such as new transcription binding to genes to and This seems to evolutionary change and this the are at and transcription and are for the of In some conditions cancers p53 is and these are activated and create a et al., 2016). p53 epigenetic homeostasis resulting in a homeostasis and diversity of the et al., 2016). These a new for of important that is DNA The many different DNA repair pathways to different types of DNA has and DNA repair pathways the central this in and In humans there are a large number of such pathways and genes to these functions. It has become clear that a number of human cancers have inherited in these DNA repair genes. It is also clear that p53 the transcription of a number of genes in different DNA repair pathways (Table and this is of its normal tumor suppressor functions. It is possible that p53 would be efficient in DNA repair if it the gene with a of that repair (and this could be the and the between DNA repair pathways and they may could be in new not only to cancer but also to many other of the a high for upon reproduction and efficient energy from A of the p53-regulated genes that have an upon functioning (Table that p53 and acid of reactive and These have an upon the between with of high levels for cell in and and energy from and in and The p53-regulated genes upon or and pathways for replication maintenance of the There is more to The that the and for possible of the p53 gene and protein have several things in all are found in from to humans and have evolutionary to them. From to they all have common p53 functions that each of (i) (ii) (iii) and epigenetic (iv) DNA a program and and and the immune Each of these of research can be by responds with p53 and downstream genes that the of stress or eliminate the cell so that the will not be to have first identified the p53 gene and selected it for and possibly in In stress was to the and so and were the of p53 stress In and control was added to p53 functions in to prevent by and By the and in the higher all seven of the functions were and one of these that p53 to integrate responses to stress may well a with its set of and and it is likely that not on this will a for the p53 gene and protein. The many faces and functions of the p53 gene and protein are to be as upon the most important for and The p53 gene and protein is one of a of genes selected and by over one This to be of the of and at the from the of This was by a from the of The is a of the of and of which for p53 mutant