| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
(Received for publication, November 9,
1994; and in revised form, January 5, 1995) From the
The tumor suppressor p53 protein possesses activities typical of
eukaryotic transcriptional activators; p53 binds to specific DNA
sequences and stimulates transcription of the target genes. By a series
of deletion and domain-swapping studies, we report here that (i) p53
has two auxiliary domains, which have little effect on the DNA binding
activity of its core domain but are capable of modulating its
transactivation activity in a target site-dependent manner; (ii) p53
contains two cell-specific transcriptional inhibitory domains, I1 and
I2, which are active in Saos-2 and HeLa cells but not in HepG2 and
Hep3B cells; and (iii) I1 inhibits the activity of several structurally
different activating regions. These results demonstrate that the
apparent transcriptional activity of p53 is determined by
collaborations among its regulatory domains, its target sites, and the
cellular environment.
The tumor suppressor p53 protein negatively regulates cell
growth (1) . Transformation of primary cells by oncogenes is
inhibited in the presence of wild type p53 protein(1) .
Reintroducing the wild type p53 gene into transformed cells blocks cell
proliferation (2) and causes these cells to accumulate in the
late G DNA tumor viruses and many human
cancers employ different mechanisms to overcome the negative effects of
p53 on cell proliferation (for a review, see (7) ). Whereas the
viruses encode oncoproteins that inactivate the braking effects of p53
on the cell cycle, the natural forms of cancers almost universally
contain mutations in the p53 gene. Mostly, they are single missense
point mutations located within the conserved regions 2-5 of p53
protein. Furthermore, many of these mutants act in a trans-dominant
fashion to promote neoplastic processes by forming hetero-oligomers
with wild type p53, therefore abrogating its function(8) . Although the molecular mechanisms responsible for the biological
activities of p53 are still the subject of intensive study, several
lines of evidence indicate that p53 functions in the regulation of
transcription. For example, p53 binds to specific DNA sequences termed
p53-responsive elements (or PREs)( Like typical
eukaryotic transcription factors, p53 is endowed with several distinct
functional domains, including those for transactivation, for DNA
binding, and for the responsiveness to regulatory signals (for a
review, see (21) ). An acidic activating region is located at
codons 20-42. The sequence-specific DNA binding domain (i.e. the so-called core domain) has been roughly mapped to the central
portion, approximately corresponding to residues 100-300. In
addition, oligomerization and nuclear localization activities have been
assigned to the carboxyl terminus. Nonetheless, a systematic study of
the structure and functions of p53 is still missing, as is a
description of the collective interactions among its functional
domains, which result in modulating the transcriptional activity of
p53. Deletion and domain-swapping approaches have been routinely and
successfully employed to define various functional domains of proteins
that have versatile biochemical activities. This testifies, in general,
to the modular nature of proteins (for a review, see (22) ).
Using these approaches to perform a systematic genetic analysis of p53,
we report here the identification and characterization of several new
transcriptional regulatory activities of p53. The functional
communications among these domains as well as superimposed effects of
the DNA target sites and the cellular environment profoundly influence
how p53 performs its transcriptional activity.
Figure 1:
The
amino terminus of p53 possesses a target site-dependent transcriptional
modulator. A, transactivation of reporter constructs
1PRE
Figure 2:
The carboxyl terminus of p53 also
possesses a target site-dependent transcriptional modulator. A, the experiments were performed as described in Fig. 1A, except that activators are the CD mutants of
p53 shown below the autoradiogram. B, protein levels
of p53 derivatives. As in Fig. 1C, except that the CD
mutants of p53 were used.
The yeast transcriptional activator HAP1 binds, as p53
does, to several dissimilar sequences ( (33) and references
therein). Deletions of HAP1 have distinct effects on its
transactivation activity via different target sites(33) . This
prompted us to look into the influence of the amino-terminal deletions
of p53 on other PREs. We chose the one found in a region upstream of
the transcription start site of the human ribosomal gene cluster
(PRE The experiments above used reporters
containing one copy of PRE. Since the amino-terminal end of p53
possesses a transactivation activity, the abrupt drop in activity
toward PRE In a
parallel study, the transactivation activity of a series of
carboxyl-terminal deletion mutants (or CD mutants) of p53 was measured.
As can be seen in Fig. 2B, mutants containing
carboxyl-terminal deletions of up to 95 amino acids were expressed to a
level about double that of wild type p53. Fig. 2A shows
that wild type p53 and these mutants exhibited similar transcriptional
activity toward PRE When the same CD
mutants were assayed with a reporter driven by PRE
Figure 3:
The transcriptional modulators have no
effect on the DNA binding activity of p53 core domain. Labeled probes
of PRE
As shown in Fig. 4A, fusion protein
GAL4-p53 barely stimulated transcription from the reporter
pG
Figure 4:
p53 protein has two cell-specific
transcriptional inhibitory domains. A, transactivation of
reporter constructs pG
Derivatives of the fusion protein with a deletion of up to
carboxyl-terminal 95 amino acids showed only a minor reduction in their
transactivation activity (compare lanes4-7 in Fig. 4A). Another transcriptional inhibitory domain,
however, was revealed when the carboxyl-terminal truncation proceeded
beyond 95 amino acids. This second inhibitory domain, or I2, presumably
resides quite close to the amino-terminal end of p53, because
GAL4-p53CD223, a fusion protein with only the amino-terminal 170 amino
acids of p53 left, still failed to enhance transcription. The
carboxyl-terminal boundary of domain I2 was localized to between amino
acids 145 and 170 of p53 since a further deletion of 25 amino acids
produced a strong activator, GAL4-p53CD248 (compare lanes11 and 12 in Fig. 4A). All of
the above experiments were done with osteosarcoma Saos-2 cells.
However, it has been demonstrated that transcriptional inhibitory
domains can function in a cell-specific manner. For instance, the To further investigate their
cell-type specificity, we tested the repression activity of I1 and I2
in two other cell lines. As shown in Fig. 4D, I1 and I2
retained their activity in HeLa but not in Hep3B cells. These results
therefore provided additional supports for the conclusion that I1 and
I2 function in a cell-specific manner. To map the boundary of I1
more precisely, we performed the experiments shown in Fig. 5.
Fusion of the p53 carboxyl-terminal 67 amino acids to activator
GAL4-p53CD267 greatly reduced its activity, while fusion of a smaller
region containing the p53 carboxyl-terminal 46 amino acids had little
effect. Immunoprecipitation analysis showed that the two proteins were
expressed to a similar level (data not shown). Thus, the p53
carboxyl-terminal 67 amino acids were required and sufficient for the
activity of I1. Moreover, since I1 could be transferred from its
original position to a completely new environment without compromising
its activity, we further concluded that I1 was a context-independent
inhibitory domain. In support of this conclusion, we found that in
addition to the p53 activation domain, I1 inhibited many activation
domains consisting of different protein motifs (see below).
Figure 5:
The
carboxyl-terminal 67 amino acids are required and sufficient for the
function of I1. The experiments were done as described in the legend to Fig. 1A, except that the reporter plasmid is
pG
Figure 6:
I1 is a general inhibitor. A, as
described in the legend to Fig. 5. B, bandshift assay.
A
Since
the most carboxyl-terminal part of p53 can inhibit its own DNA binding
activity(35) , I1, by analogy, might inhibit transcription by
blocking the DNA binding activity of the GAL4 fusion protein.
Alternatively, since I1 is overlapping with the p53 oligomerization
domain, it might inhibit transcription by forcing the fusion proteins
to form oligomers, which therefore abrogated the activity of fusion
proteins by steric hindrance. Fig. 6B shows that
neither case was responsible for the repression activity of I1. Indeed,
both GAL4VP and GAL4VPC67 had comparable affinity toward the cognate
GAL4 site, and the DNA-protein complexes exhibited similar mobility in
the bandshift assay. We therefore concluded that I1 was a general
inhibitor of various eukaryotic transactivating regions, and its
function was probably unrelated to either the proposed DNA binding
inhibition activity or the oligomerization one of the p53 carboxyl
terminus. Previous studies have shown that p53 binds to specific DNA
sequences(6, 9, 10, 11, 12, 13, 14) and
activates transcription both in vivo and in
vitro(9, 15, 16, 40) . Although
the domains responsible for the inhibition(41) ,
transactivation, and DNA binding activities of p53 have been roughly
mapped, information regarding the interactions among them in the
determination of the apparent transcriptional activity of p53 is still
missing. In this report, we identify several new functional domains in
p53 and demonstrate that the apparent transcriptional activity of p53
is determined by functional communications among its multiple domains
as well as by superimposed effects of the DNA target sites and the
cellular environment.
Conceivably, the minimal domain of p53 required for efficient DNA
binding is target site-dependent. This possibility is not without
precedent. Genetic and biochemical evidence indicates that in
transcription factor Oct-1, the region responsible for specific DNA
binding varies, depending on the DNA site(45) . Nevertheless,
this proposition apparently contradicts the result shown in Fig. 3and that of an in vitro study by Pavletich et
al.(46) , who detected specific DNA binding to PRE Alternatively, PRE Furthermore, a few
cancer cells express mutant p53 proteins that have only the
carboxyl-terminal transcriptional modulator deleted(49) . In
other words, cancers arise probably because the truncated p53 no longer
functions as a transcriptional activator. In evidence, most
p53-responsive
genes(6, 9, 13, 50, 51, 52, 53, 54, 55, 56) contain
PREs in their promoters that do not match with the consensus p53
binding sequences and therefore may become unresponsive to the
truncated p53. Such observations support that at least the
carboxyl-terminal transcriptional modulator indeed plays an important
role in determining the apparent transcriptional activity of p53.
The repression activity of domain
I2 can be demonstrated only in the context of GAL4-p53 fusion protein,
for the overlapping of I2 with the p53 DNA binding domain makes the
assay on p53 impossible. The activity of I2 seems to be regulated by
the region with a carboxyl-terminal boundary around amino acids
279-298 of p53 because the repression activity of I2 appeared
only when this region was removed (see Fig. 4A). The
underlying mechanism of the regulation is unknown. Notably,
carboxyl-terminal deletions of GAL4-p53 that terminate within the
conformationally sensitive DNA binding domain of p53 mostly fail to
transactivate. It is thus arguable that I2 might simply reflect some
structural requirement for an intact p53 core domain. The cell-specific
activity of I2, however, is incompatible with this argument concerning
the conformational state of the p53 fragments. Rather, a cell-specific
factor(s) may be involved in the modulation of I2 activity. Both
domains I1 and I2 function in a cell-specific manner. Although we do
not know the molecular mechanism responsible for this, it is not simply
a consequence of differential protein degradation in one cell line versus in another one (for an example, see Fig. 4C). More possibly, a cell-specific cofactor(s)
may be responsible for the observed difference. The finding of specific
proteins coimmunoprecipitated with GAL4-p53 from Saos-2 but not from
HepG2 cells should provide a clue for investigating the mechanisms
responsible for the cell-specific negative effect of I1 on
transcription.
Volume 270,
Number 12,
Issue of March 24, 1995 pp. 6966-6974
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
FUNCTIONAL INTERACTIONS AMONG MULTIPLE REGULATORY DOMAINS (*)
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
phase of the cell cycle(3) . Loss of p53
function results in genome instability(4, 5) and
eliminates growth arrest at the G phase in response to DNA
damage(6) , indicating that p53 functions as a checkpoint that
is important in arresting the cell cycle progression under inadequate
or detrimental growth conditions.)(6, 9, 10, 11, 12, 13, 14) and
stimulates transcription of the target genes in vivo and in vitro(9, 15, 16) . p53 mutants
that fail to suppress cell growth also cannot perform specific DNA
binding and transactivation(15) . Paradoxically, p53 also
represses transcription of many viral and cellular genes that
apparently do not have PREs (17, 18, 19) .
How p53 down-regulates gene transcription is not well understood,
although a number of studies have shown that p53 interacts with the
TATA binding protein, and, depending on the promoter context (20) , may somehow repress transcription.
Plasmid Constructions
Plasmids pE1BCAT,
pG
E1BCAT, pSG424, pSGVP, and pSGE1A have been previously
described(23) . Plasmids p1PRE
CAT and
p3PRECAT were cloned by inserting one and three copies,
respectively, of a consensus p53 binding site oligomer
(PRE) 5`-AGCTAGGCATGTCTAGACATGCCT-3` (12) into the HindIII site of pE1BCAT. p1PRE
CAT and
p3PRECAT were made similarly by replacing the above
oligomer with the one (PRE) that is found in a region
upstream of the transcription start site for the human ribosomal gene
cluster (10) and has the sequence
5`-AGCTTGCCTGGACTTGCCTGGCCTTGCCTTTTC-3`. For expression of p53
derivatives and GAL4-p53 fusion proteins in mammalian cells, polymerase
chain reaction-generated p53 DNA fragments were cloned between the HindIII and BamHI sites of pCEP4 (Invitrogen) and
pSG424, respectively. Plasmid pSGC67 was constructed by ligating a DNA
fragment encoding the carboxyl-terminal 67 amino acids of p53 between
the KpnI and XbaI sites of pSG424, resulting in an
in-frame fusion between the GAL4 and p53 modules. Plasmids pSGVPC67,
pSGE1AC67, pSGProC67, and pSGGlnC67 were cloned by inserting DNA
fragments encoding the activating regions of HSV VP16 (residues
413-490)(24) , adenovirus E1A (residues
121-222)(23) , NF1/CTF (residues
401-480)(25) , and Sp1 (residues
339-500)(26) , respectively, in-frame between the EcoRI and BamHI sites of pSGC67. Plasmids pSGPro and
pSGGln were constructed by deleting the 67 amino acids of p53 from
plasmids pSGProC67 and pSGGlnC67. Finally, for the expression of p53
protein in bacteria, pHisp53 was constructed by inserting the p53
protein coding sequence between the NdeI and BamHI
sites of plasmid 8His-pET11d(27) .Transfection and CAT Assay
Saos-2, HepG2,
Hep3B, and HeLa cells were maintained in Dulbecco's modified
Eagle's medium with 10% fetal bovine serum. Calcium
phosphate-mediated DNA transfection was performed as previously
described(28) , except that 5 µg each of the CAT reporter
and activator plasmids were used and 2 µg of a LacZ reporter
plasmid were included to monitor the transfection efficiency. CAT
activity was measured and quantified according to Carey et
al.(29) .Nuclear Extracts and DNA Binding
Assays
Nuclear extracts were prepared from Saos-2 and HepG2
cells transiently transfected with the mammalian expression plasmids of
interest. Cells from a 100-mm diameter plate were lysed by briefly
mixing them in 500 µl of ice-cold buffer A (10 mM Hepes
(pH 8.0), 1.5 mM MgCl
, 10 mM KCl, 0.5
mM dithiothreitol, and 0.5% Nonidet P-40). Nuclei were
pelleted at 3,000 
g for 3 min and then resuspended and
extracted on ice in 100 µl of buffer C (20 mM Hepes (pH
8.0), 25% glycerol, 0.42 mM NaCl, 1.5 mM MgCl, 0.2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 0.5 mM dithiothreitol)
for 30 min. Nuclear extracts were then collected by centrifugation at
15,000 g for 10 min. For bandshift assays, the probe
was the 
P end-labeled HindIII-EcoRI
fragment of pMH100(23) . 1 µg of nuclear proteins was used
per reaction. All other methods concerning bandshift assays have been
previously described(30) . Assays for the DNA binding activity
of p53 derivatives were performed as described(10) , except
that probes were the P end-labeled XhoI-SalI fragment of p3PRECAT and
p3PRECAT and that 400 µg of proteins of nuclear
extracts from transiently transfected cells and 10 µl of the
polyclonal anti-p53 anti-serum described below were used.Nuclear-Cytoplasmic
Fractionation
Nuclear/cytoplasmic fractionation was done as
follows. Saos-2 cells from a transiently transfected 100-mm diameter
plate were lysed on ice in 100 µl of hypotonic buffer (25 mM Tris-HCl (pH 7.4), 1 mM MgCl, 5 mM KCl, 1 mM phenylmethylsulfonyl fluoride, and a mixture of
protease inhibitors containing 1 µg/ml each of leupeptin,
pepstatin, antipain, and chymostatin) plus 0.5% Nonidet P-40 by
pipetting up and down a few times. 2 M sucrose was immediately
added to a final concentration of 0.25 M, and nuclei were
pelleted at 1,000 g in a microcentrifuge for 2 min at
4 °C. The supernatant (i.e. the cytosol fraction) was
further spun at 12,000 g for 2 min to remove
contaminating nuclei. Nuclei were washed three times in hypotonic
buffer plus 0.5% Nonidet P-40 and then lysed in 100 µl of hypotonic
buffer plus 1% SDS. After addition of an equal volume of 2
protein sample buffer, the nuclear fraction was sonicated to reduce its
viscosity. These nuclear fractions were used for the immunoblotting
studies shown in Fig. 1C and Fig. 2B.
CAT (lanes1-6) and
1PRECAT (lanes7-12) by the ND
mutants of p53 in Saos-2 cells. Experiments were repeated three times.
Plasmid pCH110 (Pharmacia Biotech Inc.), which contains a functional
LacZ gene, was used as an internal control to monitor transfection
efficiency. An autoradiogram of a typical experiment is shown. Diagrams
of the structure of the activators are shown below the
autoradiogram. The activator and relative CAT activity (RCA)
are indicated above each track of the autoradiogram. B, experiments were performed as in A, except that
reporter plasmids are 3PRECAT (lanes1-6) and 3PRECAT (lanes7-12). The activators are the same as in A, and therefore their diagrams are omitted. C,
protein levels of p53 derivatives. Proteins of the nuclear fraction of
Saos-2 cells transiently transfected with the vector (lane1) or p53 derivatives (lanes2-5)
were fractionated on a 10% SDS-polyacrylamide gel electrophoresis gel.
p53 derivatives were detected by immunoblotting as described under
``Materials and Methods.'' The p53 derivative is indicated above each track of the immunoblot. The positions of
molecular mass markers in kilodaltons are indicated on the left.
Antibody Preparation and Western Immunoblotting
Analysis
Anti-GAL4 antibody was purchased from UBI (New
York). The His-tagged p53 protein was expressed in and purified from
bacteria JM109DE3 (Promega) by the published procedure(27) .
The purified protein was used to raise polyclonal antibodies against
human p53 protein in rabbits according to standard protocols. For
immunoblotting, approximately 50 µg of nuclear proteins were
separated on a 10% SDS-polyacrylamide gel. All other methods concerning
Immunoprecipitation and Western immunoblotting assays were as
previously described(23, 31) .
p53 Contains Two Auxiliary Regions of Target
Site-dependent Transcriptional Modulation
Partly because
p53 binds to several dissimilar DNA sequences(6, 9, 10, 11, 12, 13, 14) and
partly because its DNA binding domain, located approximately between
amino acids 100 and 300, is structurally distinct from other known DNA
binding domains(32) , we asked whether any regions outside the
central DNA binding domain influenced the transactivation activity of
p53. Since the p53 activation domain has been mapped to its
amino-terminal end, to facilitate studies on amino-terminal deletion
mutants (or ND mutants) we decided to fuse the VP16 activation domain
of herpes simplex virus to the carboxyl-terminal end of p53. The VP16
fusion apparently caused no detriment to the transactivation activity
of p53 (Fig. 1A, compare lanes1 and 3). Mutants p53ND50-VP and p53ND100-VP, which have a deletion
of 50 and 100 amino acids, respectively, from the amino-terminal end of
the fusion protein, showed a 3-4-fold reduction in activity
toward a reporter gene containing one copy of the consensus PRE (or
PRE) in the promoter. The reduction in activity was
probably due to a difference in the level of protein being made (Fig. 1C, lanes2-5) and in the
number of activation domains (see below). Interestingly, although the
protein level of p53ND100-VP was about one-third of that of p53ND50-VP,
both stimulated transcription almost equally well. It suggested that
even the low level of p53ND100-VP being expressed was probably
sufficient to saturate all of the PRE sites. A further
deletion of 25 amino acids completely eliminated the transactivation
activity of the fusion protein, although the protein was made to a
level about half of that of p53-VP (Fig. 1C, compare lanes2 and 5). Thus, our data demonstrate
that the amino-terminal boundary of p53 required for specific DNA
binding toward the PRE is located between amino acids 100
and 125.) because it is well characterized. Fig. 1A shows that p53 and p53-VP did not distinguish PRE from
PRE (compare lanes1 and 3 to lanes7 and 9). Surprisingly, both deletion
mutants, p53ND50-VP and p53ND100-VP, failed to activate transcription
via PRE (Fig. 1A, lanes10 and 11). observed from p53-VP to p53ND50-VP or to
p53ND100-VP might merely reflect the difference in the number of
activation domains (two in p53-VP but one in each of mutants p53ND50-VP
and p53ND100-VP). If this were correct, then an increase in the number
of p53 binding sites would compensate the loss via the mechanism of
transcriptional synergism(29, 34) . As shown in Fig. 1B, whereas the activity of p53ND50-VP and
p53ND100-VP approached that of the parental p53-VP16 when the number of
PRE increased from one to three, only a marginal increase
in their activity was observed when the same assays were repeated with
three copies of PRE. In other words, the data collected
from the above study of transcription synergism suggest that the low
transcriptional activity of the two ND mutants toward PRE is unlikely due to a difference in the number or strength of
their activation domains. These data clearly demonstrated that the p53
amino-terminal region containing the first 100 residues is not required
for the core domain to activate transcription via PRE but
is indispensable for its transactivation via PRE. Thus, our
experiments uncovered a novel function of the amino-terminal end of
p53: a target site-dependent transcriptional modulator.. We noted a discrepancy between the
levels of proteins being made and their transcriptional activity. For
instance, p53 and p53CD55 stimulated transcription approximately
equally, though the latter was about twice as high in protein
concentration as the former. We have no reasonable explanation for
this, except for mutant p53CD30, whose high activity was probably due
to the deletion of a repression domain ( (35) and see below).
Note that p53CD55, p53CD75, and p53CD95 do not contain the
oligomerization domain but still stimulated transcription quite well,
indicating that the oligomerization domain is dispensable for the
transactivation activity of p53 via PRE. p53CD115 was
completely inactive, although the protein was expressed and localized
to the nucleus (Fig. 2B and data not shown). By
densitometer, the level of p53CD115 expressed was about 73% that of
wild type p53, a reduction too small to account for the inaction of
p53CD115. The experiment done with p53CD115 thus suggested that the
carboxyl-terminal boundary of p53 required for specific DNA binding
toward PRE was located between amino acids 278 and 298. In
addition, since activator p53CD95 lacks all three previously identified
nuclear localization signals (NLSI, residues 316-322; NLSII,
residues 369-375; and NLSIII, residues 379-384) ( (21) and references therein) but was still able to stimulate
transcription, this implied that all three NLSs are not required for
p53 derivatives to be localized to the nucleus., a
different result was obtained. Activators p53 and p53CD30 behaved
similarly, regardless of which PRE they were assayed with (Fig. 2A, compare lanes2 and 3 with lanes9 and 10). Interestingly,
p53CD55, p53CD75, and p53CD95 exhibited hardly any significant
transactivation activity toward PRE. Thus, in analogy to
its amino terminus, the p53 carboxyl-terminal region between residues
298 and 363 also contained a target site-dependent transcriptional
modulator.The Transcriptional Modulators Have No Effect on the
DNA Binding Activity of p53 Core Domain
We imagined that
the transcriptional modulators may influence the transactivation
activity of p53 derivatives by modulating the DNA binding activity of
the core domain. To test this idea, we compared the PRE binding
activity of mutants p53CD30 and p53-VP with that of mutants p53CD95 and
p53ND50-VP, respectively, because of their obvious difference in
transactivation activity toward PRE. To quantitatively
reflect the affinity of each protein to the PRE sites, the amount of
nuclear extract was titrated so that the incubation was carried out
within the linear range of the protein-DNA interaction. As shown in Fig. 3, all four p53 derivatives seemed to bind PRE equally well (lanes4, 6, 8,
and 10). Unexpectedly, they also exhibited similar activity in
binding to PRE (lanes3, 5, 7, and 9). The binding of p53 derivatives to PREs was
specific as demonstrated by the result that a nuclear extract of Saos-2
cells transfected with vector alone showed only residual binding (lanes1 and 2). Thus, our results did not
support the idea that the modulator regions may augment transcription
by increasing the DNA binding activity of p53 core domain toward
PRE.
(lanes2, 4, 6, 8, 10) and PRE (lanes1, 3, 5, 7, 9) were
incubated with a nuclear extract of Saos-2 cells transfected with the
activator plasmid indicated above each track of the
autoradiogram. The DNA-protein complexes were precipitated with an
anti-p53 antibody. The binding reaction was performed under an
equilibrium condition (that is, an increase in the amount of nuclear
extract proportionally brought down more probes). DNAs were purified
and separated by electrophoresis on a denaturing 5% polyacrylamide gel.
The positions of probes are indicated on the left. LaneM, P-labeled DNA size markers of MspI-digested pBR322 DNA.
p53 Has Two Cell-specific Transcriptional Inhibitory
Domains
The existence of an inhibition domain in p53
protein has been proposed(35) . Our results (Fig. 2A) demonstrated that removal of 30 amino acids
from the carboxyl terminus of p53 caused an increase in its
transactivation activity. Partly because of the rapid degradation of
p53 (Fig. 1C and Fig. 2B) and partly to
search for other p53 inhibition domain(s), which may overlap with the
DNA binding domain and therefore cannot be characterized in the p53
background, a new assay system is needed. Fortunately, the GAL4-p53
hybrid seems to be a useful system because it has been shown to be
relatively stable(36) , and the GAL4 module can provide a DNA
binding activity. Furthermore, two lines of evidence indicate that
there is probably little change in the configuration and activity of
p53 when fused to the GAL4 DNA binding domain. First, it is known that
various p53 mutants and their corresponding GAL4-p53 hybrids behave
similarly in terms of transcriptional activation(36) . Second,
the temperature-sensitive feature of the Val-135 mutant of p53 is
maintained in the background of GAL4-p53 hybrid, regardless of which
site (either a PRE or GAL4 binding site) it is assayed
with(36, 37) . Thus, it should be appropriate and
relevant to transfer conclusions drawn from studies of the fusion
protein to p53.E1BCAT. A deletion of the carboxyl-terminal 30 amino
acids increased the transactivation activity about 40-fold (compare lanes3 and 4). Since the fusion protein
binds DNA via the GAL4 module, the low transcriptional activity of
GAL4-p53 was incompatible with the argument that the region containing
the carboxyl-terminal 30 amino acids inhibits transcription by reducing
the DNA binding activity of p53(35) . Rather, our results
suggested that a repression domain, I1, exists in the carboxyl terminus
of p53. More support for this conclusion is provided below.
E1BCAT by derivatives of the GAL4-p53
fusion protein in Saos-2 cells (otherwise as in Fig. 1A). B, as described in the legend to
part A, except that experiments were performed in HepG2 cells, and only
activators B, C, J, and K were
assayed. C, immunoprecipitation of GAL4-p53 derivatives
expressed in Saos-2 (lanes2-5) and HepG2 (lanes7-10). Precipitated proteins were
separated by 10% SDS-polyacrylamide gel electrophoresis. The positions
of GAL4-p53 derivatives are indicated on the right of the
autoradiogram. The positions of proteins coprecipitated with GAL4-p53
from Saos-2 but not HepG2 cells are indicated on the left. The numbers on the left of the autoradiogram indicate
molecular sizes in kilodaltons. D, as described in the legend
to B, except that Hep3B (leftpanel) and
HeLa (rightpanel) cells were used
instead.
domain of c-jun is a cell-specific transcriptional inhibitor (38) . We therefore measured the activity of I1 and I2 in
another cell line. Fig. 4B shows that, in contrast to
the results obtained from Saos-2 cells, fusion proteins GAL4-p53 and
GAL4-p53CD223 were able to activate transcription to a high level in
HepG2 hepatoma cells (see lanes2 and 4). We
therefore concluded that domains I1 and I2 did not act as
transcriptional inhibitors in HepG2 cells. As a control, we performed
an immunoprecipitation study to show that GAL4-p53, GAL4-p53CD30,
GAL4-p53CD223, and GAL4-p53CD248 were all expressed and, as expected,
quite stable in both Saos-2 and HepG2 cells (Fig. 4C),
indicating that the observed cell-specific difference in the ability of
GAL4-p53 and GAL4-p53CD223 to stimulate transcription was not simply a
trivial result of differential protein degradation between the two cell
lines. We noted that GAL4-p53 reproducibly brought down two proteins
with molecular masses of about 56 and 80 kDa from Saos-2 but not from
HepG2 cells. They probably bind to the carboxyl end of p53 because a
30-amino acid deletion from the carboxyl terminus of p53 completely
eliminated the interaction (Fig. 4C, compare lanes2 and 3).
E1BCAT.
I1 Is a General Inhibitor
The finding
that I1 functioned as a movable inhibitory domain toward the p53
activation region compelled us to ask whether it could repress other
transactivation domains of different protein motifs. For this purpose,
we focused on those of VP16, E1A, Sp1, and NF-1/CTF, which represent
four well characterized prototypes of eukaryotic transactivators:
acidic, zinc-containing, glutamine-rich, and
proline-rich(23, 39) , respectively. In the absence of
I1, fusion proteins of GAL4 and various activation domains all
stimulated transcription to certain extents (Fig. 6A).
VP16 and E1a were strong activators, while Sp1 and NF-1/CTF were weak
ones. However, a large reduction in activity was observed for all four
activating regions when I1 was linked in cis to them (Fig. 6A), though all the proteins were expressed to a
similar level (for an example, see Fig. 6B).
P end-labeled HindIII-EcoRI fragment
of pMH100 (23) was incubated in a buffer alone (lane1) or a nuclear extract from Saos-2 cells transiently
transfected with the vector (lane2), with activator
GAL4VP (lane3) or with activator GAL4VPC67 (lane4), followed by electrophoresis on a nondenaturing 5%
polyacrylamide gel. Nuclear extracts containing approximately equal
amounts of GAL4VP and GAL4VPC67 as quantified by immunoprecipitation
were used for the binding reaction. The positions of the free probe and
DNA-protein complex are indicated on the right.
The Transcriptional Modulators
In conjunction
with the VP16 activating region, the p53 core domain containing
residues 100-298 was sufficient to activate transcription via a
consensus binding site, PRE. However, this core domain
requires two additional regions, which we refer to as transcriptional
modulators, to stimulate transcription efficiently via PRE.
We do not know how the two modulators help the core domain in the
process of transactivation of a reporter driven by PRE. The
oligomerization state of p53 is not a viable explanation. First, all
the ND mutants we tested have an intact oligomerization domain, which
is composed of the most carboxyl-terminal 50-60 amino
acids(42, 43, 44) , but they failed to
activate transcription via PRE. Second, although the CD
mutant, p53CD30, contains an oligomerization domain that is partially
truncated, it functions better than the wild type protein. with the purified core domain. may function as a composite regulatory element (47) and
interacts with transcription factors other than p53. Accordingly, p53
would be required but not sufficient by itself to stimulate
transcription from PRE. Transcriptional stimulation by
PRE would be an outcome of the concerted interaction
between p53 and other factors. This model predicts that the p53
transcriptional modulators must be somehow involved in the interaction
and that p53 mutants without the modulators either cannot interact or
interact in a nonproductive way with those factors and, therefore, fail
to activate transcription through PRE. The finding that
PRE binds, in addition to p53, several nuclear proteins
with high affinity and that the p53 amino terminus is involved in
interaction with transcription factor Sp1 (48) is in agreement
with this hypothesis. In light of this, it is worthwhile to note that
most of the known p53 binding sites could be classified into two
groups. One group, represented by PRE, mainly exhibits a
discernible palindromic motif of 20 base pairs(12) . The other
group, typified by PRE, usually consists of multiple copies
of TGCCT repeat, frequently with an Sp1-like GC-rich region adjacent to
the TGCCT repeats(9, 10) .The Inhibitory Domains
Either as part of
p53 (Fig. 2A, compare lanes9 and 10) or as part of GAL4-p53 fusion (Fig. 4A,
compare lanes3 and 4), domain I1 represses
the activity of the p53-activating region. Thus, the most
carboxyl-terminal part of p53 not only represses the DNA binding
activity of p53 (35) but is itself also a true transcriptional
inhibitor. The activity of domain I1 varies; it represses much better
in the context of GAL4-p53 than in that of p53 (compare Fig. 2A and Fig. 4A), which probably
reflects a difference in the degree of protein degradation between
them. Wild type p53 is degraded very quickly (Fig. 1C and Fig. 2B), probably resulting in a large
fraction of protein without an intact domain I1. Thus, the relatively
high transactivation activity detected with the wild type p53 protein
(see Fig. 1, A and B, and 2A) may be
largely contributed by some carboxyl-terminal truncated products. In
contrast, the GAL4-p53 fusion protein is quite stable in the cell lines
we worked with (for an example, see Fig. 4C),
indicating that domain I1 is left intact to fully perform its
inhibition function. The basis for us to propose that heterogeneous
pattern of p53 derivatives probably due to a difference in the degree
of protein degradation, but not of phosphorylation, is from the
observation that all of the p53 carboxyl-terminal deletion mutants end
with a degraded product of about 32 kDa in molecular mass (Fig. 2B). Moreover, the degradation products probably
reflect the state of p53 derivatives within the cell, because neither
the addition of protease inhibitors to the extraction buffers (Fig. 2B) nor boiling transfected cells directly in the
Laemmli SDS-containing sample buffer (data not shown) had much effect
on the profile of p53 derivatives.
)
We acknowledge Dr. C. Shih for the human p53 clone. We
thank Drs. J. Y. Chen, C. Fletcher, J. Yen, and K. King for comments.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S.-L. Tzeng, Y.-W. Cheng, C.-H. Li, Y.-S. Lin, H.-C. Hsu, and J.-J. Kang Physiological and Functional Interactions between Tcf4 and Daxx in Colon Cancer Cells J. Biol. Chem., June 2, 2006; 281(22): 15405 - 15411. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Shenk, C. J. Fisher, S.-Y. Chen, X.-F. Zhou, K. Tillman, and L. Shemshedini p53 Represses Androgen-induced Transactivation of Prostate-specific Antigen by Disrupting hAR Amino- to Carboxyl-terminal Interaction J. Biol. Chem., October 12, 2001; 276(42): 38472 - 38479. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.-F. Huang, Y.-C. Wang, D.-A. Tsao, S.-F. Tung, Y.-S. Lin, and C.-W. Wu Antagonism between Members of the CNC-bZIP Family and the Immediate-Early Protein IE2 of Human Cytomegalovirus J. Biol. Chem., April 14, 2000; 275(16): 12313 - 12320. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hong, M.-D. Lai, Y.-S. Lin, and M.-Z. Lai Antagonism of p53-dependent Apoptosis by Mitogen Signals Cancer Res., June 1, 1999; 59(12): 2847 - 2852. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-F. Tung, J.-Y. Chuang, C.-T. Lin, M.-Y. Lai, C.-W. Wu, and Y.-S. Lin Inhibition of hTAFII32-binding Implicated in the Transcriptional Repression by Central Regions of Mutant p53 Proteins J. Biol. Chem., March 19, 1999; 274(12): 7748 - 7755. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-K. Hwang and C.-T. Lin Co-localization of Endogenous and Exogenous p53 Proteins in Nasopharyngeal Carcinoma Cells J. Histochem. Cytochem., July 1, 1997; 45(7): 991 - 1004. [Abstract] [Full Text]
|
|
|
|
|
|
S. Corroyer, B. Maitre, V. Cazals, and A. Clement Altered Regulation of G1 Cyclins in Oxidant-induced Growth Arrest of Lung Alveolar Epithelial Cells. ACCUMULATION OF INACTIVE CYCLIN E-CDK2 COMPLEXES J. Biol. Chem., October 11, 1996; 271(41): 25117 - 25125. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.-L. Tsai, G.-H. Kou, S.-C. Chen, C.-W. Wu, and Y.-S. Lin Human Cytomegalovirus Immediate-Early Protein IE2 Tethers a Transcriptional Repression Domain to p53 J. Biol. Chem., February 16, 1996; 271(7): 3534 - 3540. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-Y. Chuang, C.-T. Lin, C.-W. Wu, and Y.-S. Lin A Movable and Regulable Inactivation Function within the Central Region of a Temperature-sensitive p53 Mutant J. Biol. Chem., October 13, 1995; 270(41): 23899 - 23902. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L.-J. Juan, W.-J. Shia, M.-H. Chen, W.-M. Yang, E. Seto, Y.-S. Lin, and C.-W. Wu Histone Deacetylases Specifically Down-regulate p53-dependent Gene Activation J. Biol. Chem., June 30, 2000; 275(27): 20436 - 20443. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T.-M. Hong, J. J. W. Chen, K. Peck, P.-C. Yang, and C.-W. Wu p53 Amino Acids 339-346 Represent the Minimal p53 Repression Domain J. Biol. Chem., January 5, 2001; 276(2): 1510 - 1515. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||
