Oncogene (2007) 26, 2191–2201
& 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00
www.nature.com/onc
REVIEW
Changing the p53 master regulatory network: ELEMENTary, my dear
Mr Watson
D Menendezl,3, A Inga2,3, JJ Jordan1,3 and MA Resnick1,3
1
Laboratory of Molecular Genetics, Chromosome Stability Section, National Institute of Environmental Health Sciences, NIH,
Research Triangle Park, NC, USA and 2Department of Translational Oncology, Molecular Mutagenesis Unit, National Institute for
Cancer Research, IST, Genoa, Italy
The p53 master regulatory network provides for the
stress-responsive direct control of a vast number of genes
in humans that can be grouped into several biological
categories including cell-cycle control, apoptosis and
DNA repair. Similar to other sequence-specific master
regulators, there is a matrix of key components, which
provide for variation within the p53 master regulatory
network that include p53 itself, target response element
sequences (REs) that provide for p53 regulation of target
genes, chromatin, accessory proteins and transcription
machinery. Changes in any of these can impact the
expression of individual genes, groups of genes and the
eventual biological responses. The many REs represent
the core of the master regulatory network. Since defects
or altered expression of p53 are associated with over 50%
of all cancers and greater than 90% of p53 mutations are
in the sequence-specific DNA-binding domain, it is
important to understand the relationship between wildtype or mutant p53 proteins and the target response
elements. In the words of the legendary detective Sherlock
Holmes, it is ‘Elementary, my dear Mr. Watson.’
Oncogene (2007) 26, 2191–2201. doi:10.1038/sj.onc.1210277
Keywords: p53; response element; mutants; transactivation; SNP; evolution
What constitutes functional p53 target response elements?
There have been two approaches to the identification of
p53-binding elements. One involves the iterative enrichment of sequences that could bind p53 in vitro, and
another relies on cloning of random human sequences
into a transcription reporter vector in yeast cells that
constitutively express high levels of p53 (Figure 1a),
followed by alignment and pattern search on the
transcriptionally active sequences (el-Deiry et al., 1992;
Tokino et al., 1994). Further experimental validation of
defined candidate sequences generally relied on competitive gel shift analysis, DNA fingerprinting or on gene
Correspondence: Dr MA Resnick, Laboratory of Molecular Genetics,
Chromosome Stability Section, National Institute of Environmental
Health Sciences, NIH, 111 Alexander Drive, PO Box 12233, Research
Traingle Park, NC, 27709, USA.
E-mail: [email protected]
3
These two authors equally contributed to this work.
reporter assays in transiently transfected mammalian
cells (Freeman et al., 1994; Thukral et al., 1995; Wang
et al., 1995; Nagaich et al., 1997, 1999; McLure and Lee,
1998; Resnick-Silverman et al., 1998). On the basis of
these approaches, a consensus sequence was developed
over 10 years ago consisting of two 10-base dimers
separated by a 0–13 base spacer (n): RRRCWWGYYY
(n) RRRCWWGYYY, where R is purine, Y is pyrimidine and W is A or T (el-Deiry et al., 1992; Funk et al.,
1992). This consensus sequence and related nucleotide
position weight matrices (PWM) have been adopted as the
standard for identifying potential p53 targets using
computational approaches (Tokino et al., 1994; Wang
et al., 2001; Hoh et al., 2002; Liu et al., 2004; Miled et al.,
2005). Although this consensus is already quite degenerate, nearly all established human response element
sequences (REs) have at least one mismatch.
Recently, the p53 consensus sequence received further
support using a very different approach, described by
Wei et al. (2006). They developed a novel, genome-wide
chromatin immunoprecipitation (ChIP) method and an
improved sequencing strategy for analysing DNA
sequences that can directly bind p53 in human cells
that are exposed to 5-fluoruracil, a base analog that
efficiently induces p53. Over 500 p53-binding loci in the
human genome were identified from the analysis of
B60 000 p53-ChIP fragments providing greater definition to the above consensus resulting in a refined matrix
of nucleotide probabilities at each position. Many
established p53 target REs were not retrieved under
the experimental conditions used, suggesting that either
the sites were not accessible to p53 or the crosslinking
agent or p53 occupancy at those sites was below the
detection threshold (Wei et al., 2006). For the individual
dimer-binding sites, the more common RRR and YYY
are AGA and (C/T)CC, respectively, and there is strong
preference for CATG as the central four bases. Those
results predicted AGACATG(C/T)CC as the strongest
consensus under the commonly held assumption that
nucleotide frequencies at any given position can be
independently related to the DNA-binding affinity of a
transcription factor to the complete RE (MacIsaac
and Fraenkel, 2006). However, complex protein–DNA
mutual allosteric modifications, particularly when a
transcription factor (TF) assembles on target sites as
a multimer, such as in the case of p53 (Nagaich et al.,
1997; McLure and Lee, 1998), can generate dependencies
Changing the p53 master regulatory network
D Menendez et al
2192
between nucleotides at different positions of the
consensus thus limiting the predictions based on PWM.
Although these various methods identify sequences
that can bind p53, they do not necessarily address the
range of sequences that could qualify as functional
binding sites in terms of the ability of p53 to both bind
and modulate transcription of an associated gene, nor
do they address the dynamics of such functional
interactions at binding sites. In the various screens that
were developed, there was a preference for sequences
that could be considered as strong binders. Each of these
screens represents a different endpoint assessment of p53
Oncogene
Figure 1 Yeast as an in vivo test tube to assess the transactivation capacity of p53 towards various response elements (REs). (a)
Constitutive expression of p53. Presented are generic, yeast-based
systems for assessing p53 functional status under conditions where the
p53 cDNA is constitutively expressed from a moderate to strong
promoter such as ADH1 on a selectable, low copy plasmid vector or
as a construct integrated into the chromosome. The p53 is considered
to form dimer molecules, which are then proposed to interact with
individual RE sequences to form tetramers. The different REs (i.e.,
REx1, REx2, REx3, etc.), which can be derived from human p53
regulated genes, are present in separate isogenic strains that differ only
by the RE sequences. The REs replace the upstream activating
sequences of a minimal promoter (i.e., cytochrome c) that regulates
the expression of a reporter that allows growth on deficient media
(i.e., HIS3 and URA3) or a color reporter (i.e., ADE2, and bgalactosidase) or a quantitative reporter (i.e., green fluorescent
protein (GFP), Ds-Red or firefly luciferase). Transactivation of the
reporter gene is p53-dependent and requires interaction with the RE
sequence. Multiple yeast strains containing different RE sequences
can be used to assess how specific deviations from the consensus
sequence affect p53 transactivation, as well as provide functional
fingerprints of the transactivation capacity of mutant p53s. Although
this system is useful in discriminating between p53 alleles with loss or
altered-function, it is limited in its ability to identify variant p53 alleles
that have subtle phenotypic alterations, owing to the high levels of p53
expression. (b) Rheostatable promoter system to quantitatively
address p53 transactivation capacity. The p53 gene is integrated and
placed under the control of a rheostatable GAL1 promoter that
provides for highly controlled expression of p53 protein. When cells
are grown on glucose, the GAL1 promoter is repressed. On raffinose
the GAL1 promoter is de-repressed but not induced, resulting in
a basal level of expression. Addition of galactose induces p53
expression. Depending on the amount of galactose added, the
expression can be varied over a 100-fold range. This feature provides
for analysis of transactivation by p53 from individual REs at various
p53 expression levels and it also enables transactivation to be assessed
for various p53 mutants under conditions of low expression. As
described in the text, several mutants differ from wild type only when
the expression is low. This system has been modified recently such that
RE::reporters are created in one set of haploid strains and various
p53 mutants are created in another set of haploids of opposite mating
type (Jordan, Inga and Resnick, unpublished). Following mating,
p53 induced transcription of the various RE::reporters can be
assessed. The diploid system allows for the rapid assessment of
the transactivation potential for many p53 variants towards
many individual RE sequences at various levels of p53 expression.
(c) Qualitative color assay on petri plates to assess p53 transactivation.
The ADE2 phenotypic color assay (RE::ADE2) determines the ability
of p53 variants to transactivate from specific REs in stationary cells
on plates. In the absence of the ADE2 gene product, cells growing on
minimal medium with just a small amount of adenine accumulate a
red pigment. With increased p53 transactivation from the RE, cells
are able to complete the adenine biosynthesis and the red pigment is
no longer accumulated. Hence, colony pigmentation is dependent on
the ability of p53 to transactivate from a specific RE sequence. If p53
(wild type or mutant) is capable of strongly transactivating from a
RE, colonies are large in size and white, whereas if p53 is not able to
transactivate from a RE, colonies are smaller and red. Inducible
expression of p53 under the rheostatable GAL1 promoter allows for
p53-transactivation capacity to be assessed at variable levels of protein
expression. Therefore, the range of pigmentation (pink to white)
correlates with the extent of p53 transactivation from a particular
RE. To compare the responsiveness of various REs to p53 induction,
levels of galactose (and, therefore, protein) are identified at which
comparable levels of reporter are produced (Resnick and Inga, 2003).
(d) Quantitative assessment of p53 transactivation capacity in vivo. A
luciferase reporter assay is used to assess the ability of wild-type or
mutant p53 to transactivate from specific REs in yeast. The
quantitative luciferase assay provides ‘kinetics’ of transactivation for
wild-type and mutant p53 over a range of p53 expression levels, from
basal to high levels, in logarithmically growing cells. As presented,
transactivation by wild-type p53 can differ considerably between REs.
Furthermore, the effect of specific mutations on the ability of p53 to
transactivate from different REs can be readily assessed.
Changing the p53 master regulatory network
D Menendez et al
2193
binding. For in vitro detection of REs, there must be
sufficiently strong binding to enable detection in a gel
shift. In the human cell ChIP analysis, the p53-binding
to a sequence would need to be sufficient to allow
the formation of crosslinked protein–DNA that
could be recovered by immunoprecipitation in quantities that would be distinguishable from background
by polymerase chain reaction-based analyses. In the
yeast-based screen using expressed p53, there was a
requirement for sufficient interaction with the RE to
enable detectable transcription of the promoter. Where
the yeast system depends on a combination of binding
and transactivation, the human cell ChIP and the
in vitro electrophoretic mobility shift assay (EMSA)
measurements strictly address p53-binding to DNA
in the context of the cellular milieu or in pure solution.
Hence, the role of chromatin versus naked DNA and/or
the processes that lead to the transcriptional modulation
upon p53-binding to a RE can greatly impact the
results obtained with those various approaches. In
fact, there is a dramatic difference between the p53
consensus sequence determined in vitro versus p53
target sequence within yeast and human cells in that the
spacer of an efficient, functional RE in vivo is limited to a
few bases (el-Deiry et al., 1992; Tokino et al., 1994; Wang
et al., 1995; Inga et al., 2002b; Tomso et al., 2005; Wei
et al., 2006). However, we have found that a consensus
sequence can retain limited function with larger separations between the 10 bp dimers in mammalian cells
(Menendez et al. in preparation). Moreover, even a halfsite can provide low level p53-mediated transactivation of
a gene (Menendez et al., 2006b) (discussed below).
Importantly, once a consensus sequence was established, the sequence could be used to determine, if the
transcriptional responsiveness to the stress of various
genes was mediated through direct p53 regulation,
indicated by p53 sequences associated with the genes
in question, or through downstream targets (Wang et al.,
2001; Hoh et al., 2002; Weinberg et al., 2005). Although
these approaches have been useful for identifying
potential sequences that may interact with p53, they do
not provide for the in vivo assessment of relative
transactivation capability from different REs. Sequence
differences and strength of binding have been invoked to
explain possible differences in transactivation of groups
of genes subject to direct control by p53, as for the case
of cell-cycle control versus apoptotic genes (ResnickSilverman et al., 1998; Vousden, 2000; Zhao et al., 2000;
Kannan et al., 2001; Qian et al., 2002; Lim et al., 2006).
Relationship between target sequence and transactivation
for wild-type p53
Although p53 must bind a RE to effect transactivation,
the relationship between binding and the actual level
of transcription of a target gene can be determined
by many more factors including other co-activators, as
noted above, as well as post-translational modifications
(Espinosa and Emerson, 2001; Szak et al., 2001; Kaeser
and Iggo, 2002; Espinosa et al., 2003). Even if p53 is
bound to a RE in the absence of stress, additional
proteins may determine the extent of transactivation
(Espinosa et al., 2003). Also, as the levels of p53 can
vary in response to stress and between cell types, the
extent of transactivation in a simple protein–DNAbinding model might be expected to be determined by
binding (Kd) and the amount of activated p53 that is
available, after accounting for other factors. There are
alternative possibilities to how p53 might act including,
for instance, a simple off-on switch that is dependent
only on the presence of some level of p53 (Vousden,
2000; Kaeser and Iggo, 2002).
To address strictly the relationship between sequence
and p53 transactivation, in vivo, isogenic systems are
required in which the only variables are the level of p53
and the target sequence, a situation difficult to attain
in mammalian cells. For example, experiments relying
on transient transfection do not provide the control of
p53 or the number of copies of reporter plasmids
with upstream target sequence. This led to the development of a surrogate isogenic system based in yeast
(which has no p53 gene) that provides strict control
of p53 levels along with a single integrated copy of a
target RE upstream of a reporter, as diagrammed in
Figure 1b (Inga et al., 2002b; Resnick and Inga, 2003).
The system extends previous yeast-based approaches
(Figure 1a) that were developed to address the
functionality of p53 mutants (Ishioka et al., 1993;
Flaman et al., 1995; Brachmann et al., 1996; Campomenosi et al., 2001). Briefly, the p53 coding sequence is
placed under the control of a GAL1 promoter, which
provides for rheostatable control of p53 expression
through variation in the amount of galactose in the
medium. Transcription of the p53 complementary DNA
is dependent on the activation of GAL1 promoter by
GAL4 (Johnston, 1987; Lohr et al., 1995). Gal4 binds to
the upstream activation sequence (UAS) of GAL1. In
the absence of galactose, Gal80 interacts with Gal4
inhibiting its ability to function as an activator,
although it may bind the UAS. Addition of galactose
leads to Gal3 inhibition of the Gal8–Gal4 interaction,
allowing Gal4 to stimulate the activation of its target
genes. Glucose prevents GAL1 expression through
catabolite repression mechanism. There is a basal level
of p53 expression on raffinose, and p53 expression can
be increased several 100-fold in proportion to the
amount of galactose added. p53 dependent transactivation of a reporter gene from a RE can be detected with
color reporters using colony plate assays (Figure 1c).
Efficiencies of transactivation can thus be compared
across different levels of p53. A luciferase-based assay
has provided the opportunity to address transactivation
in cultures of logarithmically growing cells at very low
levels of p53 (Figure 1b and d). An additional distinct
advantage of yeast is the opportunity to rapidly
modify in vivo either the target REs or to create mutant
p53 coding sequences utilizing a highly efficient recombination-based system, known as delitto-perfetto (Italian
for ‘perfect murder’) that targets desired changes
with oligonucleotides (Storici et al., 2001, 2003; Storici
and Resnick, 2006).
Oncogene
Changing the p53 master regulatory network
D Menendez et al
2194
The combination of a rheostatable system with over 50
REs, most of which are found in human cells, has revealed
a several 100-fold variation in the ability of putative REs
to support transactivation by p53 (Inga et al., 2002b).
For a few of these REs, available data from competitive
gel shift assays supported our results (Freeman et al.,
1994; Wang et al., 1995; Resnick-Silverman et al., 1998;
Thornborrow et al., 2002). These findings contrast with
results obtained with PWM algorithms that would predict
comparable activity between the REs. More recently, a
sophisticated in vitro quantitative analysis of sequencespecific p53-binding to DNA revealed large differences in
affinities for various REs (Weinberg et al., 2005).
When all other factors are kept equal, the yeast-based
results suggest a wide range potential for transactivation
that is determined by the RE sequence and level of
protein. Thus, for any level of p53 protein, there is an
intrinsic capability for the transactivation responsiveness of each RE. Although other factors are clearly
important, these results have been interpreted in terms
of a model based on a piano (Figure 2) (Resnick and
Inga, 2003). The ‘hand’ is p53 that can press the various
‘keys’ (i.e., genes) with different intensities. As a result, a
‘chord’ is created from the various genes being turned
on to various extents that results in a ‘sound,’ the
biological response or phenotype.
Relationship between sequence and transactivation for
altered-function p53 mutants
Most cancer-associated p53 alterations are due to
missense mutations located in the DNA-binding domain
(Olivier et al., 2002; Hamroun et al., 2006) (http://
p53.bii.a-star.edu.sg/index.php). This suggests that the
DNA-binding and consequent transactivation capability
of p53 are especially important in preventing cancer,
which is consistent with the various categories of genes
controlled by p53 (Vogelstein et al., 2000; Levine et al.,
2006). However, many p53 mutations do not eliminate
transactivation capability (Kato et al., 2003; Hamroun
et al., 2006). In an exhaustive yeast-based study that
investigated nearly all possible single base substitution
mutations, approximately 30% of the mutant proteins
retained function towards at least one RE, consistent
with other yeast-based studies (Kato et al., 2003;
Shiraishi et al., 2004). These altered-function mutants
differ from the ‘gain-of-function’ mutants described in
human cells, which have lost sequence-specific transactivation capability. The gain-of-function mutants are
able to influence the transactivation of other genes
through RE-independent mechanisms, such as protein–
protein interactions with large protein complexes that
are recruited to specific promoters by other sequencespecific transcription factors (Blandino et al., 1999; Di
Agostino et al., 2006), or alternatively the mutant p53s
are able to impact on chromatin metabolism (Deppert
et al., 2000; Rubbi and Milner, 2003).
Altered-function mutants have been examined extensively with the rheostatable yeast-based system (Resnick
and Inga, 2003). Even though many of the mutants
Oncogene
Figure 2 A piano model to describe the master regulator p53 and
mutants transactivation of target genes. As described in Resnick
and Inga (2003), a master regulatory gene such as p53 directly
controls the expression of many genes, similar to hands playing
keys on a piano (bottom of figure). The p53 piano has many ‘keys’
corresponding to the directly targeted genes, and it is played by the
p53 ‘hand’ where the fingers interact directly with promoter REs.
As a result, a ‘chord’ is created such that each note or key may be
played with a different intensity (i.e., level of transactivation). For p53
mutants that retain partial function or have altered DNA-binding
specificity, the p53 ‘hand’ can have some fingers that are stronger
(bold type) or weaker (smaller and lighter type) and some inactive
fingers that do not strike keys. In addition it is possible that new keys
will be played. The resulting ‘sounds’ played by wild-type and mutant
p53 hands correspond to the cellular phenotype. In the case of p53
mutants, some of these changes can result in new chords and sounds
with novel biological consequences, for example, different sensitivities
to DNA-damaging-agents resulting from expression of wild-type
versus the T125R mutant (top panel) (Menendez et al., 2006a). (This
figure is adapted from (Resnick and Inga, 2003)).
simply result in an overall dampening of transactivation
(i.e., reduction in the ‘volume’ of the chord in Figure 2),
surprisingly, several were change-in-spectrum such that
new-sounding chords were produced, where intensities
of individual keys were changed and some keys were
even lost from the chord. The nature of the chord would
be determined in part by the amount of p53 available,
that is, the strength of the hand. Among the change-inspectrum mutants were those that resulted in much
stronger transactivation than wild-type at low p53 levels
(Inga et al., 2001). These super-trans mutants might be
expected to dramatically affect the relevant biological
pathways. An interesting category of mutants was
revealed that resulted in cell killing when expressed at
other than low levels (Inga and Resnick, 2001). Not only
were these mutants toxic, they were also shown to
exhibit greatly increased transactivation towards
some REs when expressed at low p53 levels. A UVinduced skin cancer hot spot mutation T122L (corresponds to the T125 position in humans) was identified
in repair-deficient mice that were also heterozygous
for p53 (Reis et al., 2000). This mutation resulted in a
change-in-spectrum phenotype towards many REs in yeast
Changing the p53 master regulatory network
D Menendez et al
2195
(Inga et al., 2002a), suggesting that the altered p53 chord
could be highly selected under certain circumstances.
The piano analogy for transactivation was found to
apply to the expression of endogenous target genes when
certain p53 mutants were expressed at biologically
meaningful levels in human cells (see example in
Figure 2) (Menendez et al., 2006a). To test directly the
piano interpretation of p53 change-in-spectrum mutants, several were examined in human osteosarcoma
SaOS2 (p53-/-) cells in a tetracycline (Tet) off-on
regulatable system. Wild-type and various p53 mutants
were placed under Tet control and integrated into
SaOS2 cells. Individual clones were selected such that
the level of expression attainable was comparable with
that measured following UV-irradiation of other cell
lines containing a functional p53 gene. Following
induction, each of the mutants exhibited unique induction patterns for the target genes examined. Interestingly, under these conditions the level of p53-induced
transcription of the endogenous target genes reflected
the extent of binding, as measured by ChIP at promoter
regions containing REs, which suggests that under the
off-on conditions for p53 induction used, there is a
direct relationship between promoter occupancy and
transactivation. In this case, differences in transactivation capability towards the various REs appear to be
due to the ability of the proteins to bind rather than in
the capability for transcription once bound.
When a p53-inducing genotoxic stress (UVC-irradiation) was applied to the cells, p53 protein
levels increased. This increase was associated with the
appearance of post-translational modifications that
resulted in higher p53 protein stability and proportionally higher occupancy levels. Overall, those results
suggest that allosteric modification of p53 protein does
not provide a major level of regulation of p53 activities
except as expected when more p53 is available, consistent with previous reports (Espinosa and Emerson,
2001; Kaeser and Iggo, 2002).
The changes in the transactivation of p53 target genes
induced by some change-in-spectrum mutants resulted
in a substantial alteration of cellular responses to
ionizing and UV-irradiation in terms of clonal survival
and apoptosis (Menendez et al., 2006a). For example,
the T125R mutation reversed the sensitivities to these
two agents (Figure 2) as compared with wild-type p53.
Thus, the ‘chord’ of genes played and the intensities
were changed so that the resulting biological response –
the ‘sound’ – was altered. The finding that a single p53
mutation can dramatically alter the pattern of genes
induced as well as their levels has several implications
for cancer. For example, the biological consequences
could depend on the specific altered-function mutation,
and this might influence opportunities for treatment.
Evolution of the p53 network
In a broader sense, the results with change-in-spectrum
single mutants as well as variation in the responsiveness
of REs to p53 transactivation suggest mechanisms of
selection in the evolution of master regulatory networks.
This concept is supported by findings with the Spalax blind
mole rat (Ashur-Fabian et al., 2004; Avivi et al., 2005),
which lives underground and is subjected to hypoxic
conditions. It appears to have uniquely adapted to the
p53-inducing conditions of life with less oxygen by two
lysine changes at the conserved arginine codons corresponding to R174 and R209 in human p53. The p53 protein
alignment from 44 species and functional analysis of the
mutated proteins suggested that the R174K is an alteredfunction p53 mutation, also found in human tumors. This
mutant p53 protein apparently results in less induction of
the apoptosis set of genes and this could contribute to the
adaptation of the species to a hypoxic environment.
A proline to arginine polymorphic variant at p53
codon 72 is another example of a p53 variant, which
may represent a genetic source of adaptive phenotypic
diversification. The two alleles differ in the potential for
protein–protein interactions and induction of apoptosis
and the allelic status could impact the outcome of viral
infections, the risk of developing cancer as well as
impact the progression of the disease (Storey et al., 1998;
Thomas et al., 1999; Bergamaschi et al., 2003; Bergamaschi
et al., 2006) (for a summary of cancer risk-association studies
see http://p53.bii.a-star.edu.sg/aboutp53/snps/snpdetail.
php?geneid ¼ X54156&snppos ¼ 12139). Interestingly,
the allele frequency broadly varies in the human population
and follows a geographical gradient.
Evolution can also occur through changes in the
target REs of the p53 master regulatory network. Along
this line, the DDB2 gene involved in global repair of
UV-induced damage is not responsive to direct p53
transactivation in mice unlike the situation in humans
(Tan and Chu, 2002). A closer examination of the DDB2
gene indicates that unlike for humans, there is not a
functional p53 target RE, consistent with lack of p53binding at the mouse promoter. As a result, this could
result in mice being much more sensitive to UV-induced
skin cancers in areas of exposure.
There are now several examples of human diversity in
the p53 network that are due to changes in REs
(Contente et al., 2002; Tomso et al., 2005). In one case,
a single nucleotide polymorphism (SNP) in the promoter region of the MDM2 enhances the opportunity
for binding by the transcription factor SP1 and results
in higher basal levels of the MDM2 protein with a
consequent reduction in p53 level/functionality (Bond
et al., 2004; Arva et al., 2005). Several studies have
examined the impact of this SNP (known as SNP309) on
the risk of developing cancers. Although an initial report
showed a clear effect on the age of onset for familial as
well as sporadic sarcomas (Bond et al., 2004), other
studies produced conflicting results on the association
between SNP status and cancer risk. Recently, it was
recognized that the impact of this regulatory SNP is
influenced by estrogen receptors (ER) (via a nearby target
RE). When ER status and hormonal levels were taken
into account, SNP309 resulted in higher risk or aggressiveness of colon and breast cancer (Bond et al., 2006a, b).
Human diversity in target REs has been explored
directly using an approach that combines bioinformatics
Oncogene
Changing the p53 master regulatory network
D Menendez et al
2196
and functional analysis (Tomso et al., 2005). Reasoning
that single base changes in putative p53 responsive REs
might be expected to impact significantly the potential
for transactivation, an extensive SNP database was
perused for pairs of alleles, where one might be expected
to be a well-functioning RE for p53-transactivation and
the other would be a poor p53 target. For example,
alleles that differed by C or G in the central CATG
sequence would be expected to have a marked effect on
transactivation function. The search focused on a region
5 kb from the start site of a gene, the region of the vast
majority of p53 target REs. Among the many candidates, over 40 pairs of such SNP alleles were identified
(Tomso et al., 2005). There were no SNP pairs identified
for any of the established human p53 REs. Most of the
newly discovered putative p53 REs had not been
reported earlier. Several of the pairs of REs were then
functionally compared through transactivation analysis
in yeast and in human cells (Figure 3). For the latter, the
RE alleles were placed in plasmids upstream of reporters
and transfected along with p53-expressing plasmids into
SaOS2 cells. In all cases, the predicted stronger allele
always yielded greater transactivation than the predicted
weaker allele for the eight pairs of alleles tested in yeast
or human cells (Figure 3). For example, the SNP alleles in
the promoter of the ADAR2 adenine deaminase gene
(GGACAAGTTg-AAACTT(G/A)CaC) and Toll-like receptor gene (TLR8) (AGGCAAGaTg-AAACAT(G/A)
TCa) exhibited dramatic differences in p53 transactivation.
These findings of SNP variations in REs demonstrate
that there may be considerable diversity between
humans in terms of responsiveness to wild type and
even altered-function, cancer-associated p53 mutants
(Menendez and Resnick, in preparation). In terms of the
piano model (Figure 2), this would correspond to
variations in the keys that are available for playing on
different individual pianos.
The rules of engagement are not so ELEMENTary, my
dear Mr Watson
As discussed above, for many years the metric for
identifying p53 target REs has been a consensus
sequence consisting of two RRRCWWGYYY decamers
separated by a small number of bases.
Figure 3 Modulation of the p53-transactivation network by SNPs in p53–REs that alter p53-transactivation capacity. The ability
of pairs of SNP–RE alleles to support p53 transactivation was determined under isogenic conditions in yeast and in human SaOS2
(p53/) cells (adapted from (Tomso et al., 2005)). The REs are located upstream from minimal promoters and luciferase reporter genes.
(a) Transactivation capacity in yeast. Luciferase activity for the predicted strong and predicted weak SNP RE allele is presented as relative
induction between p53-expressing cells and control cells lacking p53. (b) Transactivation capacity in mammalian cells. p53-null SaOS2
cells were transiently co-transfected with CMV vectors containing p53 and the luciferase reporter constructs that contains one of the pair
of RE alleles. Luciferase activity was measured relative to the empty pGL3-promoter vector in the absence or presence of p53. The
nucleotide beneath each bar corresponds to the specific SNP allele examined. The AIP1 and P21-5’ p53 REs served as positive controls.
Oncogene
Changing the p53 master regulatory network
D Menendez et al
2197
Evaluation of the ability of p53 to transactivate from
a RE with yeast-based assays has demonstrated that the
specific sequence of the p53 RE can contribute greatly to
the level of transactivation obtained by wild-type p53.
Additionally, there is an increasing need to examine
further if the organization of the RE, in terms of spacer
distance and the requirement for the presence of four
quarter-sites alters the ability of p53 to function from a
specific sequence. Examination of many REs through
functional assays to empirically obtain the rules of
transactivation has indicated that p53 transactivation
could also be accomplished with a 3/4 site (i.e., a halfsite with an adjacent quarter site) based on results with
the proliferating-cell nuclear antigen promoter gene
which contains the sequence GAACAAGTCCGGG
CATaTgT. The p53 protein can strongly interact with
this sequence in vitro (Weinberg et al., 2005) and lead to
modest induction in the yeast-based system (Jordan and
Resnick, in preparation). In addition, functional studies
have demonstrated that spacer size between decamer
half-sites can impact the ability of p53 to transactivate
from a RE in yeast and human cells (unpublished).
It was recently reported that even a putative half-site
can provide good responsiveness to p53 (Menendez
et al., 2006b). The promoter of the human Flt1 gene,
which codes for the VEGF receptor 1, a component of
the VEGF pathway, was examined for possible SNPs
that might affect its regulation. A SNP was identified
(C/T) that resulted in one allele having a consensus p53
half-site: GGACA(c/T)GCTC. Immediately beyond this
half-site was the sequence (CCCCTG)GGACcTGagC
that was poorly related to a p53 consensus. Unlike the
C-allele, the T-allele was able to bind p53 and
functioned in p53-induced transactivation. Subsequent
investigations have shown that the p53 T-allele half-site
binds p53 and supports a small amount of p53
transactivation on its own, although it binds p53
(Menendez et al., in press). Surprisingly, the presence
of a nearby ER target RE enabled p53 transactivation.
These findings with the Flt1 promoter demonstrate that
a much smaller RE sequence is needed for p53 to function
in the genome provided that other co-transactivating
sequences are available. Furthermore, the Flt1-T allele
merges three important biological pathways: stress (p53),
hormonal responses (ER), and angiogenesis (VEGF
receptor Flt1). Importantly, these findings suggest that
master regulatory partial target REs might provide a
genetically wired means for crosstalk between pathways.
Implications
The ELEMENTary components of master regulatory
networks are the REs along with the sequence-specific
TFs. Differences in RE sequences, changes in p53 levels
and post-translational modifications, along with amino
acid alterations in p53 provide a large matrix, in which
the expression network of target and downstream
genes can be varied in the context of disease states,
inter-individual genetic variability (e.g., SNPs) or
species-specific evolutionary separation. Considering
the central role that p53 plays in preventing tumorigenesis, understanding p53 biology and defining new
components and interactions within its network will be
valuable in cancer prevention. Along these lines, an
emerging issue relates to the health impact of RE
sequence variation within the human population as well
as the consequences of somatically acquired or inherited
altered-function p53 mutants. There are limited studies
where a functional classification that distinguishes
altered versus loss-of-function p53 mutants was adopted
to query clinical data on cancer likelihood and outcomes. For breast cancer, functional classification of
p53 does not appear to add prognostic value to the
knowledge about p53 mutant status (Olivier et al.,
2006). It is interesting that several p53 mutations
associated with breast cancer, particularly in BRCA1related cases, appear wild-type for transactivation under
normal expression conditions, but some are clearly
different from wild type when expressed at low levels
(Resnick and Inga, 2003). For germinal p53 mutations
resulting in familial cancer proneness, complete loss of
function, as compared with altered function, appeared to
be associated with earlier cancer onset and more severe
clinical manifestations (Fronza et al., submitted).
The importance of sequence-specific binding in p53
transactivation versus the role of other factors continues
to be a fundamental issue in understanding the action of
TFs. Beyond TFs there is a complex interplay between
stress-signal transducers, their regulated TFs and
chromatin states at target promoters. In addition, there
is a plethora of induced epigenetic changes that can
impact the outcome of protein–protein interactions
between TFs, cofactors, mediators and basal transcriptional components (Ito et al., 1999; Espinosa et al., 2003;
Wallberg et al., 2003; An et al., 2004; Minsky and Oren,
2004; Zhang et al., 2005). Although the recruitment of
p53 at the target RE(s) is a necessary step, it may not be
sufficient for changing transcription rates, unless specific
cofactors are recruited and/or other cis-acting factors are
concertedly activated (Espinosa et al., 2003; Gomes et al.,
2006). Formally, the outcome of transcriptional modulation elicited by a sequence-specific TF could be largely
dictated by the cofactor(s) tethered by the TF at the site.
Along this line, an important question that needs to
be further examined is to what extent RE sequence can
act as an allosteric modifier of promoter bound TFs,
impacting the assembly of higher order transcriptional
complexes (Lefstin and Yamamoto, 1998). Based on
biochemical and biophysical characterization, several
models for p53 action have been suggested that focused
on the ability of p53 to bind specific genomic sequences
resulting in the transactivation of adjacent genes (Prives
and Hall, 1999; Vogelstein et al., 2000). Results have
supported p53-induced, RE-dependent conformational
changes that could be related to differences in binding
affinities and provided an initial framework for explaining transactivation specificity. However, a critical issue
that could not be properly addressed experimentally was
the impact of RE sequence on the quaternary structure
of p53–DNA complexes, namely the possibility of
conformational variation of DNA-bound p53 tetramers,
Oncogene
Changing the p53 master regulatory network
D Menendez et al
2198
which could be an underlying source of diversity in
transcriptional outcomes. For nearly 12 years, there has
been only one crystal structure of human p53 bound to
DNA (Cho et al., 1994). However, because the p53 used
was restricted to monomers, the more general nature of
p53–RE interactions could not be addressed. Recently,
Kitayner et al. (2006) reported the structure of three
p53:DNA crystals obtained with the p53–DNA binding
domain and different half-site RE sequences (GGA
CATGTCC, AGGCATGCCT; GGGCATGCCC plus
one flanking base). In every case, four p53 molecules
self-assembled on two DNA molecules resulting in a p53
tetramer consisting in a dimer of dimers. Differential
p53–DNA binding affinity could be related to sequencespecific variations in the protein–DNA contact geometry. The structures also predicted that the four
transactivation domains in the p53 amino terminus
would have a common orientation with respect to the
tetramer bound to DNA, which may facilitate cooperative recruitment of co-factors. These findings provide a
new structural understanding of p53 binding to REs and
they are likely to form the basis for addressing the large
differences in transactivation from various REs, the
consequences of mutations in p53, and the mechanisms
for searching REs among nonspecific genomic DNA
(Hupp and Lane, 1994; Ahn and Prives, 2001; Espinosa
and Emerson, 2001; McKinney et al., 2004).
Our finding that half-site p53 REs can provide
responsiveness to p53 has many implications and
demonstrates that the p53 master regulatory network is
even more complex and larger than envisioned based on
the standard 20 base consensus sequence. In light of the
recent structural characterizations discussed above, it will
be interesting to understand how p53 can transactivate
from a half-site or a 3/4 site. In addition, the size and
complexity of the p53 master regulatory network may be
much greater than envisaged based on the standard
20 base consensus sequence. The results with a half-site
raises the possibility of a multifactorial synergistic
interconnection of master regulators involving partial or
even nonfunctional REs. Along this line, although wholegenome ChIP analysis (Wei et al., 2006), has provided
opportunities to identify strong p53–RE interactions,
there are likely to be many more that will require even
more sensitive methods for detection. In addition to roles
in gene regulation, the sequence-specific interactions with
partial sites might have other functions, including the
suggested possibility of ‘chromatin relaxation’ in order to
facilitate DNA repair (Rubbi and Milner, 2003).
In its role as a master transcriptional regulator, p53
also appears able to trans-repress genes (Ho and
Benchimol, 2003) including those affecting signaling
pathways such as cell proliferation (cyclin B, RNA
polymerase I, c-myc), apoptosis (bcl-2, survivin) and
cytoskeleton organization (stathmin and map4)
(Ahn et al., 1999; Xu and el-Gewely, 2001; Ho and
Benchimol, 2003). In general, the mechanism(s) of
repression is often unclear and may not be dependent
on site-specific DNA-binding activity (Zhai and Comai,
2000; Koumenis et al., 2001; Ho et al., 2005). For
example, the responsiveness of some genes to p53 can be
Oncogene
indirect in that target genes (transactivated or transrepressed) might be transcription factors, which then
affect downstream responses to p53 signaling (Xu and
El-Gewely, 2003). Clear mechanisms of p53-dependent
repression have been identified in the association
between p53 and histone deacetylases and other
chromatin remodeling co-factors, as well as co-repressors (Murphy et al., 1999; Ho and Benchimol, 2003).
Although sites are suggested, there is no clearly
identified consensus p53-binding site within repressed
promoters. How these sites differ from those in
transactivated promoters has only begun to be elucidated (May and May, 1999; Johnson et al., 2001;
Hoffman et al., 2002). Alternatively, bound p53 may
impede the binding of other transcriptional activators,
as is the case of the DNA polymerase delta gene (Li and
Lee, 2001) and the antiapoptotic gene survivin (Hoffman
et al., 2002). Future studies are likely to further elucidate
the consensus sites at which p53 could have a repressive
effect as well as structure-function relationships, based on
recent crystal structures (Kitayner et al., 2006).
The identification of the complete repertoire of p53regulated genes is still far from complete (Miled et al.,
2005; Wei et al., 2006). In recent years, a variety of
genomic techniques (microarrays, SAGE, ChIP, EMSA)
have been utilized to identify p53-regulated target genes.
Gene expression has been examined in various cell types,
with several DNA-damaging-agents that lead to p53
activation, or simply by restoring the p53 expression in
parental p53 null cells (el-Deiry, 1998; Yu et al., 1999;
Zhao et al., 2000; Wang et al., 2001; Xu and El-Gewely,
2003; O’Farrell et al., 2004). However, in most studies it
has been difficult to distinguish between direct and
indirect transactivation effects of p53 or mutants.
Approaches utilizing isogenic analysis and variation in
levels of p53 expression can resolve some of these issues.
Finally, the approaches for characterizing wild-type
p53 and mutant proteins using variable expression and
interactions with REs under isogenic conditions can be
applied to other sequence-specific TFs. For example, the
NKX2-5 protein is required for human heart development. Using the yeast-based system, single and multiple
somatic mutations identified in congenital heart disease
have been shown to result in altered abilities to
transactivate from NKX2-5 target REs (Inga et al.,
2005). The findings with p53 and NKX2-5, as well as
other genes currently under study (Inga and Resnick, in
preparation), also support the view that sequencespecific master regulators could be master genes of
diversity, where single amino acid changes can greatly
alter the target and downstream genes included in
master regulatory networks as found for some cancerassociated p53 mutations (Resnick and Inga, 2003).
Acknowledgements
We appreciate the critical comments on the paper by Chris
Halweg and Tom Darden. This work was partially supported by
a grant from the Italian Association for Cancer Research, AIRC
(to AI) and intramural research funds from NIEHS. Jennifer
Jordan is supported by a Department of Defense Breast Cancer
Research Program Predoctoral Traineeship Award (BC051212).
Changing the p53 master regulatory network
D Menendez et al
2199
References
Ahn J, Murphy M, Kratowicz S, Wang A, Levine AJ,
George DL. (1999). Down-regulation of the stathmin/
Op18 and FKBP25 genes following p53 induction. Oncogene
18: 5954–5958.
Ahn J, Prives C. (2001). The C-terminus of p53: the more you
learn the less you know. Nat Struct Biol 8: 730–732.
An W, Kim J, Roeder RG. (2004). Ordered cooperative
functions of PRMT1, p300, and CARM1 in transcriptional
activation by p53. Cell 117: 735–748.
Arva NC, Gopen TR, Talbott KE, Campbell LE, Chicas A,
White DE et al. (2005). A chromatin-associated and
transcriptionally inactive p53-Mdm2 complex occurs in
mdm2 SNP309 homozygous cells. J Biol Chem 280:
26776–26787.
Ashur-Fabian O, Avivi A, Trakhtenbrot L, Adamsky K,
Cohen M, Kajakaro G et al. (2004). Evolution of p53 in
hypoxia-stressed Spalax mimics human tumor mutation.
Proc Natl Acad Sci USA 101: 12236–12241.
Avivi A, Ashur-Fabian O, Amariglio N, Nevo E, Rechavi G.
(2005). p53-a key player in tumoral and evolutionary
adaptation: a lesson from the Israeli blind subterranean
mole rat. Cell Cycle 4: 368–372.
Bergamaschi D, Gasco M, Hiller L, Sullivan A, Syed N,
Trigiante G et al. (2003). p53 polymorphism influences
response in cancer chemotherapy via modulation of p73dependent apoptosis. Cancer Cell 3: 387–402.
Bergamaschi D, Samuels Y, Sullivan A, Zvelebil M, Breyssens
H, Bisso A et al. (2006). iASPP preferentially binds p53
proline-rich region and modulates apoptotic function of
codon 72-polymorphic p53. Nat Genet 38: 1133–1141.
Blandino G, Levine AJ, Oren M. (1999). Mutant p53 gain of
function: differential effects of different p53 mutants on
resistance of cultured cells to chemotherapy. Oncogene 18:
477–485.
Bond GL, Hirshfield KM, Kirchhoff T, Alexe G, Bond EE,
Robins H et al. (2006a). MDM2 SNP309 accelerates tumor
formation in a gender-specific and hormone-dependent
manner. Cancer Res 66: 5104–5110.
Bond GL, Hu W, Bond EE, Robins H, Lutzker SG, Arva NC
et al. (2004). A single nucleotide polymorphism in the
MDM2 promoter attenuates the p53 tumor suppressor
pathway and accelerates tumor formation in humans. Cell
119: 591–602.
Bond GL, Menin C, Bertorelle R, Alhorpuro P, Aaltonen LA,
Levine AJ. (2006b). MDM2 SNP309 Accelerates colorectal
tumour formation in women. J Med Genet 43: 950–952.
Brachmann RK, Vidal M, Boeke JD. (1996). Dominantnegative p53 mutations selected in yeast hit cancer hot spots.
Proc Natl Acad Sci USA 93: 4091–4095.
Campomenosi P, Monti P, Aprile A, Abbondandolo A,
Frebourg T, Gold B et al. (2001). p53 mutants can often
transactivate promoters containing a p21 but not Bax or
PIG3 responsive elements. Oncogene 20: 3573–3579.
Cho Y, Gorina S, Jeffrey PD, Pavletich NP. (1994). Crystal
structure of a p53 tumor suppressor-DNA complex: understanding tumorigenic mutations [see comments]. Science
265: 346–355.
Contente A, Dittmer A, Koch MC, Roth J, Dobbelstein M.
(2002). A polymorphic microsatellite that mediates induction of PIG3 by p53. Nat Genet 30: 315–320.
Deppert W, Gohler T, Koga H, Kim E. (2000). Mutant p53:
‘gain of function’ through perturbation of nuclear structure
and function? J Cell Biochem Suppl 35: 115–122.
Di Agostino S, Strano S, Emiliozzi V, Zerbini V, Mottolese M,
Sacchi A et al. (2006). Gain of function of mutant p53: the
mutant p53/NF-Y protein complex reveals an aberrant
transcriptional mechanism of cell cycle regulation. Cancer
Cell 10: 191–202.
el-Deiry WS. (1998). Regulation of p53 downstream genes.
Semin Cancer Biol 8: 345–357.
el-Deiry WS, Kern SE, Pietenpol JA, Kinzler KW, Vogelstein B.
(1992). Definition of a consensus binding site for p53. Nat
Genet 1: 45–49.
Espinosa JM, Emerson BM. (2001). Transcriptional Regulation by p53 through intrinsic DNA/chromatin binding and
site-directed cofactor recruitment. Mol Cell 8: 57–69.
Espinosa JM, Verdun RE, Emerson BM. (2003). p53 functions
through stress- and promoter-specific recruitment of transcription initiation components before and after DNA
damage. Mol Cell 12: 1015–1027.
Flaman JM, Frebourg T, Moreau V, Charbonnier F, Martin C,
Chappuis P et al. (1995). A simple p53 functional assay
for screening cell lines, blood and tumors. Proc Natl Acad
Sci USA 92: 3963–3967.
Freeman J, Schmidt S, Scharer E, Iggo R. (1994). Mutation of
conserved domain II alters the sequence specificity of DNA
binding by the p53 protein. EMBO J 13: 5393–5400.
Funk WD, Pak DT, Karas RH, Wright WE, Shay JW. (1992).
A transcriptionally active DNA-binding site for human p53
protein complexes. Mol Cell Biol 12: 2866–2871.
Gomes NP, Bjerke G, Llorente B, Szostek SA, Emerson BM,
Espinosa JM. (2006). Gene-specific requirement for P-TEFb
activity and RNA polymerase II phosphorylation within the
p53 transcriptional program. Genes Dev 20: 601–612.
Hamroun D, Kato S, Ishioka C, Claustres M, Beroud C,
Soussi T. (2006). The UMD TP53 database and website:
update and revisions. Hum Mutat 27: 14–20.
Ho J, Benchimol S. (2003). Transcriptional repression
mediated by the p53 tumour suppressor. Cell Death Differ
10: 404–408.
Ho JS, Ma W, Mao DY, Benchimol S. (2005). p53-Dependent
transcriptional repression of c-myc is required for G1 cell
cycle arrest. Mol Cell Biol 25: 7423–7431.
Hoffman WH, Biade S, Zilfou JT, Chen J, Murphy M. (2002).
Transcriptional repression of the anti-apoptotic survivin
gene by wild type p53. J Biol Chem 277: 3247–3257.
Hoh J, Jin S, Parrado T, Edington J, Levine AJ, Ott J. (2002).
The p53MH algorithm and its application in detecting p53responsive genes. Proc Natl Acad Sci USA 99: 8467–8472.
Hupp TR, Lane DP. (1994). Allosteric activation of latent p53
tetramers. Curr Biol 4: 865–875.
Inga A, Monti P, Fronza G, Darden T, Resnick MA. (2001).
p53 mutants exhibiting enhanced transcriptional activation
and altered promoter selectivity are revealed using a sensitive,
yeast-based functional assay. Oncogene 20: 501–513.
Inga A, Nahari D, Velasco-Miguel S, Friedberg EC,
Resnick MA. (2002a). A novel p53 mutational hotspot in
skin tumors from UV-irradiated Xpc mutant mice alters
transactivation functions. Oncogene 21: 5704–5715.
Inga A, Reamon-Buettner SM, Borlak J, Resnick MA. (2005).
Functional dissection of sequence-specific NKX2-5 DNA
binding domain mutations associated with human heart
septation defects using a yeast-based system. Hum Mol
Genet 14: 1965–1975.
Inga A, Resnick MA. (2001). Novel human p53 mutations that
are toxic to yeast can enhance transactivation of specific
promoters and reactivate tumor p53 mutants. Oncogene 20:
3409–3419.
Inga A, Storici F, Darden TA, Resnick MA. (2002b).
Differential transactivation by the p53 transcription factor
Oncogene
Changing the p53 master regulatory network
D Menendez et al
2200
is highly dependent on p53 level and promoter target
sequence. Mol Cell Biol 22: 8612–8625.
Ishioka C, Frebourg T, Yan YX, Vidal M, Friend SH,
Schmidt S et al. (1993). Screening patients for heterozygous
p53 mutations using a functional assay in yeast. Nat Genet 5:
124–129.
Ito M, Yuan CX, Malik S, Gu W, Fondell JD, Yamamura S
et al. (1999). Identity between TRAP and SMCC complexes
indicates novel pathways for the function of nuclear
receptors and diverse mammalian activators. Mol Cell 3:
361–370.
Johnson RA, Ince TA, Scotto KW. (2001). Transcriptional
repression by p53 through direct binding to a novel DNA
element. J Biol Chem 276: 27716–27720.
Johnston M. (1987). A model fungal gene regulatory mechanism: the GAL genes of Saccharomyces cerevisiae. Microbiol
Rev 51: 458–476.
Kaeser MD, Iggo RD. (2002). Chromatin immunoprecipitation analysis fails to support the latency model for
regulation of p53 DNA binding activity in vivo. Proc Natl
Acad Sci USA 99: 95–100.
Kannan K, Kaminski N, Rechavi G, Jakob-Hirsch J,
Amariglio N, Givol D. (2001). DNA microarray analysis
of genes involved in p53 mediated apoptosis: activation of
Apaf-1. Oncogene 20: 3449–3455.
Kato S, Han SY, Liu W, Otsuka K, Shibata H, Kanamaru R
et al. (2003). Understanding the function-structure and
function-mutation relationships of p53 tumor suppressor
protein by high-resolution missense mutation analysis. Proc
Natl Acad Sci USA 100: 8424–8429.
Kitayner M, Rozenberg H, Kessler N, Rabinovich D,
Shaulov L, Haran TE et al. (2006). Structural basis of
DNA recognition by p53 tetramers. Mol Cell 22: 741–753.
Koumenis C, Alarcon R, Hammond E, Sutphin P, Hoffman W,
Murphy M et al. (2001). Regulation of p53 by hypoxia:
dissociation of transcriptional repression and apoptosis
from p53-dependent transactivation. Mol Cell Biol 21:
1297–1310.
Lefstin JA, Yamamoto KR. (1998). Allosteric effects of DNA
on transcriptional regulators. Nature 392: 885–888.
Levine AJ, Hu W, Feng Z. (2006). The P53 pathway:
what questions remain to be explored? Cell Death Differ
13: 1027–1036.
Li B, Lee MY. (2001). Transcriptional regulation of the
human DNA polymerase delta catalytic subunit gene
POLD1 by p53 tumor suppressor and Sp1. J Biol Chem 276:
29729–29739.
Lim YP, Lim TT, Chan YL, Song AC, Yeo BH, Vojtesek B
et al. (2006). The p53 knowledgebase: an integrated
information resource for p53 research. Oncogene [Epub
ahead of print].
Liu S, Mirza A, Wang L. (2004). Generation of p53 target
database via integration of microarray and global p53
DNA-binding site analysis. Methods Mol Biol 281: 33–54.
Lohr D, Venkov P, Zlatanova J. (1995). Transcriptional
regulation in the yeast GAL gene family: a complex genetic
network. FASEB J 9: 777–787.
MacIsaac KD, Fraenkel E. (2006). Practical strategies for
discovering regulatory DNA sequence motifs. PLoS Comput
Biol 2: e36.
May P, May E. (1999). Twenty years of p53 research:
structural and functional aspects of the p53 protein.
Oncogene 18: 7621–7636.
McKinney K, Mattia M, Gottifredi V, Prives C. (2004). p53
linear diffusion along DNA requires its C terminus. Mol
Cell 16: 413–424.
Oncogene
McLure KG, Lee PW. (1998). How p53 binds DNA as a
tetramer. EMBO J 17: 3342–3350.
Menendez D, Inga A, Resnick MA. (2006a). The biological
impact of the human master regulator p53 can be altered by
mutations that change the spectrum and expression of its
target genes. Mol Cell Biol 26: 2297–2308.
Menendez D, Krysiak O, Inga A, Krysiak B, Resnick MA,
Schonfelder G. (2006b). A SNP in the flt-1 promoter
integrates the VEGF system into the p53 transcriptional
network. Proc Natl Acad Sci USA 103: 1406–1411.
Menendez D, Inga A, Snipe J, Krysiak O, Schönfelder G,
Resnick MA. (2007). A SNP in a half-binding site creates
p53 and estrogen receptor control of VEGFR1. Mol Cell
Biol in press.
Miled C, Pontoglio M, Garbay S, Yaniv M, Weitzman JB.
(2005). A genomic map of p53 binding sites identifies novel
p53 targets involved in an apoptotic network. Cancer Res
65: 5096–5104.
Minsky N, Oren M. (2004). The RING domain of Mdm2
mediates histone ubiquitylation and transcriptional repression. Mol Cell 16: 631–639.
Murphy M, Ahn J, Walker KK, Hoffman WH, Evans RM,
Levine AJ et al. (1999). Transcriptional repression by wildtype p53 utilizes histone deacetylases, mediated by interaction with mSin3a. Genes Dev 13: 2490–2501.
Nagaich AK, Appella E, Harrington RE. (1997). DNA
bending is essential for the site-specific recognition of
DNA response elements by the DNA binding domain
of the tumor suppressor protein p53. J Biol Chem 272:
14842–14849.
Nagaich AK, Zhurkin VB, Durell SR, Jernigan RL, Appella E,
Harrington RE. (1999). p53-induced DNA bending and
twisting: p53 tetramer binds on the outer side of a DNA
loop and increases DNA twisting. Proc Natl Acad Sci USA
96: 1875–1880.
O’Farrell TJ, Ghosh P, Dobashi N, Sasaki CY, Longo DL.
(2004). Comparison of the effect of mutant and
wild-type p53 on global gene expression. Cancer Res 64:
8199–8207.
Olivier M, Eeles R, Hollstein M, Khan MA, Harris CC,
Hainaut P. (2002). The IARC TP53 database: new online
mutation analysis and recommendations to users. Hum
Mutat 19: 607–614.
Olivier M, Langerod A, Carrieri P, Bergh J, Klaar S, Eyfjord J
et al. (2006). The clinical value of somatic TP53 gene
mutations in 1,794 patients with breast cancer. Clin Cancer
Res 12: 1157–1167.
Prives C, Hall PA. (1999). The p53 pathway. J Pathol 187:
112–126.
Qian H, Wang T, Brachmann RK. (2002). Not all p53 DNA
binding sites are created equal. Proceeding of the American
Association for Cancer Research 43: 1141.
Reis AM, Cheo DL, Meira LB, Greenblatt MS, Bond JP,
Nahari D et al. (2000). Genotype-specific Trp53 mutational
analysis in ultraviolet B radiation- induced skin cancers
in Xpc and Xpc Trp53 mutant mice. Cancer Res 60:
1571–1579.
Resnick-Silverman L, St Clair S, Maurer M, Zhao K,
Manfredi JJ. (1998). Identification of a novel class of
genomic DNA-binding sites suggests a mechanism for
selectivity in target gene activation by the tumor suppressor
protein p53. Genes Dev 12: 2102–2107.
Resnick MA, Inga A. (2003). Functional mutants of the
sequence-specific transcription factor p53 and implications
for master genes of diversity. Proc Natl Acad Sci USA 100:
9934–9939 . Epub 2003 Aug 8.
Changing the p53 master regulatory network
D Menendez et al
2201
Rubbi CP, Milner J. (2003). p53 is a chromatin accessibility
factor for nucleotide excision repair of DNA damage.
EMBO J 22: 975–986.
Shiraishi K, Kato S, Han SY, Liu W, Otsuka K, Sakayori M
et al. (2004). Isolation of temperature-sensitive p53 mutations from a comprehensive missense mutation library.
J Biol Chem 279: 348–355.
Storey A, Thomas M, Kalita A, Harwood C, Gardiol D,
Mantovani F et al. (1998). Role of a p53 polymorphism
in the development of human papillomavirus-associated
cancer. Nature 393: 229–234.
Storici F, Durham CL, Gordenin DA, Resnick MA. (2003).
Chromosomal site-specific double-strand breaks are efficiently targeted for repair by oligonucleotides in yeast. Proc
Natl Acad Sci USA 100: 14994–14999.
Storici F, Lewis LK, Resnick MA. (2001). In vivo site-directed
mutagenesis using oligonucleotides. Nat Biotechnol 19:
773–776.
Storici F, Resnick MA. (2006). The delitto perfetto approach
to in vivo site-directed mutagenesis and chromosome
rearrangements with synthetic oligonucleotides in yeast.
Methods Enzymol 409: 329–345.
Szak ST, Mays D, Pietenpol JA. (2001). Kinetics of p53 binding
to promoter sites in vivo. Mol Cell Biol 21: 3375–3386.
Tan T, Chu G. (2002). p53 Binds and activates the xeroderma
pigmentosum DDB2 gene in humans but not mice. Mol Cell
Biol 22: 3247–3254.
Thomas M, Kalita A, Labrecque S, Pim D, Banks L,
Matlashewski G. (1999). Two polymorphic variants of
wild-type p53 differ biochemically and biologically. Mol
Cell Biol 19: 1092–1100.
Thornborrow EC, Patel S, Mastropietro AE, Schwartzfarb EM,
Manfredi JJ. (2002). A conserved intronic response element
mediates direct p53-dependent transcriptional activation
of both the human and murine bax genes. Oncogene 21:
990–999.
Thukral SK, Lu Y, Blain GC, Harvey TS, Jacobsen VL.
(1995). Discrimination of DNA binding sites by mutant p53
proteins. Mol Cell Biol 15: 5196–5202.
Tokino T, Thiagalingam S, el-Deiry WS, Waldman T,
Kinzler KW, Vogelstein B. (1994). p53 tagged sites from
human genomic DNA. Hum Mol Genet 3: 1537–1542.
Tomso DJ, Inga A, Menendez D, Pittman GS, Campbell MR,
Storici F et al. (2005). Functionally distinct polymorphic
sequences in the human genome that are targets for
p53 transactivation. Proc Natl Acad Sci USA 102:
6431–6436.
Vogelstein B, Lane D, Levine AJ. (2000). Surfing the p53
network. Nature 408: 307–310.
Vousden KH. (2000). p53: death star. Cell 103: 691–694.
Wallberg AE, Yamamura S, Malik S, Spiegelman BM,
Roeder RG. (2003). Coordination of p300-mediated
chromatin remodeling and TRAP/mediator function
through coactivator PGC-1alpha. Mol Cell 12: 1137–1149.
Wang L, Wu Q, Qiu P, Mirza A, McGuirk M, Kirschmeier P
et al. (2001). Analyses of p53 target genes in the human
genome by bioinformatic and microarray approaches. J Biol
Chem 276: 43604–43610.
Wang Y, Schwedes JF, Parks D, Mann K, Tegtmeyer P.
(1995). Interaction of p53 with its consensus DNA-binding
site. Mol Cell Biol 15: 2157–2165.
Wei CL, Wu Q, Vega VB, Chiu KP, Ng P, Zhang T et al.
(2006). A global map of p53 transcription-factor binding
sites in the human genome. Cell 124: 207–219.
Weinberg RL, Veprintsev DB, Bycroft M, Fersht AR. (2005).
Comparative binding of p53 to its promoter and DNA
recognition elements. J Mol Biol 348: 589–596.
Xu H, el-Gewely MR. (2001). P53-responsive genes and the
potential for cancer diagnostics and therapeutics development. Biotechnol Annu Rev 7: 131–164.
Xu H, El-Gewely MR. (2003). Differentially expressed downstream genes in cells with normal or mutated p53. Oncol Res
13: 429–436.
Yu J, Zhang L, Hwang PM, Rago C, Kinzler KW, Vogelstein B.
(1999). Identification and classification of p53-regulated
genes. Proc Natl Acad Sci USA 96: 14517–14522.
Zhai W, Comai L. (2000). Repression of RNA polymerase I
transcription by the tumor suppressor p53. Mol Cell Biol 20:
5930–5938.
Zhang X, Krutchinsky A, Fukuda A, Chen W, Yamamura S,
Chait BT et al. (2005). MED1/TRAP220 exists predominantly in a TRAP/ Mediator subpopulation enriched in
RNA polymerase II and is required for ER-mediated
transcription. Mol Cell 19: 89–100.
Zhao R, Gish K, Murphy M, Yin Y, Notterman D, Hoffman WH
et al. (2000). Analysis of p53-regulated gene expression
patterns using oligonucleotide arrays. Genes Dev 1 4:
981–993.
Oncogene