 |
INTRODUCTION |
The p53 tumor suppressor gene is the most frequent target for
genetic alterations in human cancers (1, 2). The wild type
(wt)1 p53 protein is
apparently latent under normal conditions but, upon different types of
stress, becomes modified and activated (3) and exerts its biological
activities, including growth arrest, apoptosis, and differentiation
(4-6).
The recently discovered p53 family member, p73, shares a remarkable
homology in DNA sequence and protein structure with p53. Indeed,
p73 can be roughly divided into three domains, (a)
the N-terminal transactivation domain, which shares 29% homology with the N-terminal part of p53, (b) the sequence-specific DNA
binding domain, which shares 63% of homology with the corresponding
p53 domain, and (c) the tetramerization domain, which shares
42% of homology with the oligomerization domain of p53 (7). Unlike p53, p73 is subject to alternative splicing, giving rise to a family of
proteins whose individual function has yet to be elucidated (1, 8-11).
p73 is not inactivated by viral oncoproteins such as SV40 large T
antigen, HPV E6, and Ad E1Bp55, well known inactivators of p53
(12-15). Furthermore, although p53 is stabilized and activated by
diverse types of stress including DNA-damaging agents, radiation, oncogenes, hypoxia, and ribonucleotide depletion, to date p73 is known
to be stabilized only in response to cisplatin and 
-radiation (3, 4,
16-18). It has recently been shown that p73 can be acetylated in
response to doxorubicin and selectively directed to activate specific
target genes (19). More recent work has reported that p73 is required
for p53-dependent apoptosis in response to DNA damage
(20).
Ectopic expression of p73 in p53
/, and p53+/+ cells causes growth
arrest, apoptosis, and differentiation, as does p53 (7, 21-25). These
effects are achieved mainly through the activation of a plethora of
specific target genes. Several reports show that p73 binds to p53
binding sites in vitro and in vivo and,
consequently, activates p53 target genes (7, 22). Thus, transcriptional activation or repression of specific sets of target genes mediates the
biological effects of both p53 and p73. The importance of functional
and physical integrity of the transcriptional activation domain for p53
activity has been clearly demonstrated by the findings that mice
carrying a p53 mutated in the N-terminal transactivation domain are
prone to develop tumors, as do p53-deficient mice (26). Major
differences between p53 and p73 have been revealed by in vivo ablation of the genes. Thus, p53- and p73-deficient mice exhibit quite different phenotypes; p53-deficient mice develop normally
but undergo spontaneous tumor formation (mainly sarcoma and lymphomas),
whereas the p73 counterparts exhibit severe defects in the development
of the central nervous system and suffer from inflammations (27, 28).
Such differences are likely to depend also on activation or repression
of different sets of target genes that need to be identified. Indirect
support for such hypothesis has been provided by the recent findings
that the potent transcriptional co-activator yes-associated protein
(YAP) binds to the long forms of p73 and p63 but not to p53
(29).
The recent development of DNA microarrays has allowed global analysis
of the pattern of activated or repressed genes in response to different
types of stimuli, including p53 activation (30-32). Taking advantage
of this approach, we compared the gene expression profiles upon
ponasterone A induction of p73
or p53 in the same cellular context,
H1299. We found that p73 or wt-p53 expression modifies 141 and 320 genes, respectively. p73 up-regulates 85 genes, of which 25 are
specific and 27 are in common with p53 regulation, whereas p53 induces
153 genes, of which 63 are specific. We will focus here on the p73
activated genes in response to ponasterone induction. Functional
classification of these genes reveals that they are involved in many
aspects of cell function ranging from cell cycle and apoptosis to DNA
repair, including also several brain-specific genes involved in
synaptic regulation. Furthermore, we report that p73 or p53 are
recruited directly to some of the activated genes. Our findings
indicate that, upon ectopic expression in the same cellular context,
p73 promotes a specific transcriptional gene profile that only
partially overlaps with that of p53.
| |
EXPERIMENTAL PROCEDURES |
Cell Lines--
The H1299 cell line is derived from a human
non-small cell lung carcinoma. H1299 cells were maintained in RPMI
medium supplemented with 10% fetal calf serum. Before transfection,
the culture medium was changed to Dulbecco's modified Eagle's medium
supplemented with 10% fetal calf serum. H1299-pVgRXR cells were
maintained in the same medium containing zeocin (100 µg/ml). The
H1299-p73 and H1299-wt-p53 cell lines were maintained in medium
containing zeocin and G418 (400 µg/ml). To induce expression of
p73 or p53, ponasterone A, a synthetic analogue of ecdysone, was
added to the medium (final concentration, 2.5 µM).
Plasmids and Transfections--
Plasmids pVgRXR and pIND were
from Invitrogen. pIND/p73 was prepared by cloning the
BamHI/EcoRV fragment of human p73, generated by PCR (the sequence of the oligonucleotides is available upon request), into pIND. pIND/wt-p53 was prepared by cloning the
HindIII/XbaI fragment of human p53 (a kind gift
of Dr. T. Unger) into pIND (33). H1299 cells were transfected with each
plasmid using the calcium phosphate method. Clone selection was done
with zeocin (100 µg/ml) for pVgRXR and with G418 (400 µg/ml) for
pIND/p73 and pIND/wt-p53. Stable clones obtained by double selection
were screened by immunoblotting using antibody Ab4 for p73 (Neomarker) and DO1 for p53.
Immunoblot Analysis--
Total cell lysates were prepared as
previously described (34), and protein content was determined with the
Bio-Rad protein assay kit (Bio-Rad). Samples containing 50 µg of
total protein were resolved by SDS-10% PAGE and transferred to
nitrocellulose membranes (Bio-Rad). For p73 detection, a p73 monoclonal
antibody was used at a 1:200 dilution; for p53 detection, a mixture of p53 monoclonal antibodies DO1 and 1801 was used at a 1:40 dilution; for
p21waf1 detection, a p21 polyclonal antibody (C19, Santa Cruz)
was used at a 1:200 dilution.
Cell Cycle Analysis--
Cells (3 × 105) were
seeded on 60-mm dishes, and 12 h later the medium was replaced by
medium containing ponasterone A (2.5 µM final
concentration). 24 h after ponasterone A addition, floating cells
and attached cells were collected, washed in phosphate-buffered saline,
resuspended in 1 ml of phosphate-buffered saline, and fixed in 5 ml of
cold methanol for 30 min at 20 °C. After centrifugation and a
further wash in phosphate-buffered saline, cells were resuspended in 1 ml of phosphate-buffered saline containing 50 µg/ml RNase and 50 µg/ml propidium iodide (Sigma) and analyzed by cytofluorimetry with
an Epics-XL analyzer (Coulter Corp.). Data were analyzed using the
Cellfit program.
RNA Extraction and Reverse Transcriptase Reaction--
Cells
from H1299-pIND clone 1, H1299-p73 clones 9 and 11, and H1299-p53
clones 23 and 16 were harvested in TRIzol reagent (Invitrogen) at
specific time points (0, 5, 9, 12, 24 h) after ponasterone A
addition, and total RNA was isolated as per the manufacturer's
instructions. Five micrograms of total RNA was reverse-transcribed at
37 °C for 45 min in the presence of random hexamers and Moloney
murine leukemia virus reverse transcriptase (Invitrogen). PCR analyses
were carried out by using oligonucleotides specific for the following
genes: CaN19 (down, 5'-CTCTGAATTCGCCACAGATCCATGATGTGC; up,
5'-CTCTGCGGCCGCCAACAGACAAAAAAAGTTTAT TGAATACAAAAC); CaN19 (down,
5'-GTAAGGGGGAAATGAAGGAAC TTCT; up, 5'-ACAAAACTCAAAGGCATCAACAGTC); 14-3-3
(down, 5'-TCTCAGTAGCCTATAAGAACGTGGTG; up, 5'-ATCTCGTAGTGGAAG ACGGAAAAGT); PIG3 (down, 5'-CCGGAAAACCTCTACGTGAA; up,
5'-CTCTGGGATAGGCATGAGGA); 1-antitrypsin (down,
5'-TTCTTCTCC CCAGTGAGCAT; up, 5'-GTGTCCCCGAAGTTGACAGT); p21waf1
(down, 5'-CCTCTTCGGCCCGGTGGAC; up, 5'-CCGTTTTCGACCCTGAGAG);
PTGF-
(down, 5'-GAGCTGGGAAGATTCGAACA; up, 5'-AGA
TTCTGCCAGCAGTTGGT); JAG2 (down, 5'-CCTTAAGGAGTACCAGGCCAA; up,
5'-AAGTGGCGCTGTAGTAGTTCTCGT). The housekeeping aldolase A
mRNA, used as an internal control, was amplified from each cDNA
reaction mixture using the following specific primers: down,
5'-CGCAGAAGGGGTCCTGGTGA; up, 5'-CAGCTCCTTCTTCTGCTGCG. Amplified PCR products were electrophoresed on a 2% agarose gel containing ethidium bromide (0.5 µg/ml) and visualized under UV light.
Preparation of Labeled cRNA and Hybridization of
Microarrays--
Total RNA was isolated from H1299-PIND clone 1, H1299-p73 clone 9, and H1299-p53 clone 23 cells after the addition
of ponasterone A for 0, 5, 9, 12, and 24 h. Biotin-labeled cRNA
was synthesized and hybridized as described (35) to the Genechip HuGene
FL array (Affymetrix, Santa Clara, CA), which contains probes for
~11,000 mRNA species, and one chip was hybridized to cRNA from
each time point. Scanned output files were visually inspected for
hybridization artifacts. Arrays lacking significant artifacts were
analyzed using Genechip 3.3 software (Affymetrix). Arrays were scaled
to an average intensity of 1200/gene and analyzed independently. The
expression value for each gene was determined by calculating the
average of differences (perfect match intensity minus mismatch intensity) of the probe pairs in use for this gene. Ratios were determined by dividing the average difference of H1299-p73 or H1299-p53 for each time point with those of the 0-h time point.
Clustering Analysis--
Clustering analysis was performed by
the super-paramagnetic clustering method (36) on the 211 genes that
were up-regulated more than 2-fold in at least 3 time points by the
induction of p53 (153 genes) or p73 (85 genes), where 27 genes were
common to both p73 and p53. The gene had to be "Present" in the
Present/Absent call provided by Affymetrix software at least at one
time point.
Each gene was represented by eight components representing the ratio of
the average difference value provided by the Affymetrix software at
each time point divided by that of 0-h time point (before the addition
of ponaterone). The 0-h time point was the average of the 0 h of
the two cell lines and the eight components containing four from the
p73 and four from the p53 induced cells, each at the indicated time
point (5, 9, 12, and 24 h). Before clustering we normalized each
row such that its mean vanishes, and its norm is one (35) as follows.
Aij represents the ratios of the expression of gene
i (where i = 1-211), measured at
experiment (and time point), j = 1-8. We subtract from
Aij its average <Ai> (see Equation 1) and divide
the difference by i, its S.D. for gene i (see
Equation 2); the resulting 8-component vector represents gene
i, and the resulting normalized matrix is denoted by
Bij (see Equation 3).
|
(Eq. 1)
|
|
(Eq. 2)
|
|
(Eq. 3)
|
The clustering algorithm measured the distance between the genes
using the regular Euclidean distance between their normalized values.
Genes with similar expression profiles (over the eight time points) are
represented by two nearby vectors and are placed in the same cluster.
Formaldehyde Cross-linking and Chromatin
Immunoprecipitation--
H1299-p73 clone 9 and H1299-wt-p53 clone
23 cells were treated with ponasterone A to induce the expression of
p73 and wt-p53 for 24 h. DNA and proteins were cross-linked by
the addition of formaldehyde (1% final concentration) 10 min before
harvesting, and cross-linking was stopped by the addition of glycine pH
2.5 (125 µM final concentration) for 5 min at room
temperature. Cells were scraped off the plates, resuspended in
hypotonic buffer, and passed through a 26-gauge needle. Nuclei were
spun down, resuspended in 300 µl of SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8, and a protease
inhibitor mixture), and sonicated to generate 500-2000-bp fragments.
After centrifugation, the cleared supernatant was diluted 10-fold with
immunoprecipitation buffer (50 mM Tris-HCl, pH 8, 150 mM NaCl, 5 mM EDTA, 0.5% Nonidet P-40). The
cell lysate was precleared by incubation at 4 °C with 15 µl of
protein G beads preadsorbed with sonicated single-stranded DNA and
bovine serum albumin. The cleared lysates were incubated overnight with
a mixture of anti-p73 polyclonal antibodies (C17 and C19) (Santa Cruz), anti-green fluorescent protein polyclonal antibody (Invitrogen), or
anti-p53 (Ab7) or without any antibody. Immune complexes were precipitated with 30 µl of protein G beads preadsorbed with sonicated single-stranded DNA and bovine serum albumin. After centrifugation the
beads were washed, and the antigen was eluted with 1% SDS, 100 mM sodium carbonate. DNA-protein cross-links were reversed by heating at 65 °C for 4-5 h, and DNA was phenol-extracted and ethanol-precipitated. Levels of CaN19, p21waf1, 14-13-3,
PIG3, 1-antitrypsin, JAG2, and PTGF- DNAs were
determined by PCR using oligonucleotides spanning the p53/p73 binding
sites. The following specific oligonucleotides were used: CaN19-up site (down, 5'-GTGTTCAAAGCCTGACACCTAACTT; up,
5'-TGGATCATAGCTCACTGTAATCTCG); CaN19-down site (down,
5'-AAGTAGCTGGGACTACAAGCGTATG; up, 5'-GGGATAGAAAAGCCCAGCTAAGATA); p21waf1-up site (down, 5'-CTATTTGGGACTCCCCAGTCTCTT; up,
5'-GGTTTACCTGGGGTCTTTAGAGGTC); p21waf1-down site (down,
5'-ATGTATAGGAGCGAAGGTGCA GAC; up, 5'-CCTCCTTTCTGTGCCTGAAACA); 14-3-3
(down, 5'-CTGTACTTCAGCCTGGACATCAGAG; up, 5'-CCGACCTAATAGTTGAGCCAG GAT);
PIG3 (down, 5'-CAGGACTGTCAGGAGGAGGCGAGTGATAAGG; up,
5'-GTGCGATTCTAGCTCTCACTTCAAGCAGAGG); JAG2 (down,
5'-ACTGCTGCCTTCTGGAAACTC; up, 5'-CAAGTGGTGAACAAGGGAGACT). Oligonucleotides specific for thymidine kinase promoter (down, 5'-GTGAACTTCCCGGAGGCGCAA; up, 5'-GCCCCTTTAAACTTGGTGGGC)
were used as negative control.
| |
RESULTS |
Generation of Stable Cell Lines Expressing Inducible p73 or
p53--
Inducible cell lines overexpressing p73 or p53 were
generated in two steps. First, H1299 cells were transfected with pVgRXR followed by zeocin selection (33, 34). The resulting clones were
transiently transfected with pIND/p73 or with pIND/wt-p53 plasmid
and maintained in the presence of ponasterone A. The highest expressor
of either p73 or wt-p53 was chosen and stably transfected with the
above-mentioned plasmids or the pIND control vector followed by G418
selection. Western blot and immunostaining with an anti-p73 monoclonal
antibody or with a mixture of anti-p53 DO1 and 1801 monoclonal
antibodies were performed to screen for p73- and p53-inducible expression and intracellular localization. As seen in Fig.
1, A and B,
upper panels), for 2 representative clones (H1299-p73 clone 9 and H1299-wt-p53 clone 23) expression of both p73 and p53
was induced by ponasterone A (2.5 µM). To verify whether
the induced proteins were transcriptionally active we checked the expression of p21waf1 and found it induced by both p73 and
p53 (Fig. 1, A and B, middle and
lower panels). In addition, both of these proteins were
predominantly nuclear, as expected (data not shown). Identical analyses
performed in H1299-pIND cells revealed no p53 protein, since H1299
cells are p53 null, and no p21waf1 induction was shown
upon ponasterone A (data not shown). With regard to endogenous
p73, H1299 cells express very low protein levels, undetectable by
direct immunoblot (data not shown) and detectable only in p73
immunoprecipitates followed by Western blot with specific anti-p73
antibodies (34).
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 1.
Ponasterone A induced expression of
p73, p53, and p21waf1 in H1299
cells. Total cell lysates (100 µg/lane) derived from
H1299-p73 clone 9 and H1299-p53 clone 23 cells, respectively, were
lysed at the indicated time points after the addition of 2.5 µM ponasterone A. Levels of p73, p53, and
p21waf1 were detected by probing the nitrocellulose filter with
a monoclonal anti-p73 antibody (Ab4) with a mixture of anti-p53
monoclonal antibodies DO1 and 1801 and with an anti-p21waf1
polyclonal serum, respectively. Equal loading of protein amount for
each line was determined by probing with anti-tubulin or anti-Hsp70
antibodies.
|
|
To further verify whether transcriptionally active p73 and p53 were
capable of altering the growth of H1299 cells, we analyzed the cell
cycle profile of each clone upon ponasterone A induction. We found that
p73 or p53 expression causes an increase in the G1
population accompanied by a decrease in the S phase population (Fig.
2, B and C). No
alterations in the cell cycle profile of H1299-pIND clone 1 cells were
found (Fig. 2A).
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 2.
Ectopic expression of either
p73 or p53 promotes cell cycle modification in
H1299 cells. Shown is cell cycle analysis of H1299-pIND clone 1, H1299-p73 clone 9, and H1299-p53 clone 23 cells by propidium iodide
staining. To induce p73 and p53, cells were stimulated for 24 h
with 2.5 µM ponasterone A (Pon-A). At that
time floating and adherent cells were pooled, fixed, and stained as
described under "Experimental Procedures." The cytofluorimetric
analysis was performed with the aid of an Epics-XL analyzer (Coulter).
Data were analyzed using the Cellfit program.
|
|
Identification of p73 or p53 Target Genes by DNA Microarray
Analysis--
H1299-pIND clone 1, H1299-p73 clone 9, and
H1299-wt-p53 clone 23 cells were stimulated with ponasterone A and
harvested in TRIzol reagent after 0, 5, 9, 12, and 24 h (Fig.
1). To eliminate background noise in the analysis of the
hybridization experiments, we chose a very stringent filter and
considered only those genes that showed more than 2-fold induction or
repression at 3 or more time points in the p73 or p53 clones over
the average of the 0-h time points of the cell lines.
By this criteria, 153 genes were found up-regulated by p53 and 167 genes were found down-regulated. In the p73-expressing cells, 85 and
56 genes fulfilled the threshold criteria for up and down regulation,
respectively. To characterize the global gene expression due to p53 or
p73 activation we employed a scatter plot analysis of these genes at
5 and 24 h compared with the 0-h time point (before ponasterone A
addition). We found that the expression level of this collection of
genes changed very little in the control cell line (H1299-pIND clone 1)
but showed extensive changes in the p53- and p73-expressing cell lines
(data not shown). It also appears that in the p53 cell line there may
be leaky expression of p53 since some of the genes showed increased
expression even at 0 h (data not shown). Indeed, low level
expression of p53 at 0 h, before induction, is detectable in Fig.
1.
Analysis of the genes whose expression was altered in both cell lines
identified a group of common genes that were activated by both p53 and
p73 and a group of genes that responded only to p53 or p73 (Fig.
3). For example, 27 and 21 genes were up-
and down-regulated, respectively, by both proteins using the filter of
2-fold change at three or more time points. These were defined as
"common genes." Fig. 3 shows that 65 genes (shaded circles) passed
this filter for p53 but were up-regulated by p73 over 2-fold only once
or twice (but not three times). Similarly 33 genes (shaded circles, Fig. 3A) passed the filter for up-regulation
by p73 and were also up-regulated more than 2-fold by p53 once or
twice. These groups of genes therefore showed preference in their
response to activation by either p73 or p53. Last, two groups of genes were up-regulated more than 2-fold at least at three time points by
only one of the transcription factors and not even once by the other.
p73 induced 25 such genes (denoted "p73", see Table I
and Fig. 3A), and p53 induced
63 such genes (denoted "p53"; see Fig.
3A).
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 3.
Venn diagram of genes regulated by p53 and
p73. Only genes that showed at least 2-fold up
(A) or down (B) regulation at a minimum of 3 time
points as compared with the control were included in the analysis. The
diagram shows that 27 and 21 genes were up- and down-regulated,
respectively, by both p53 and p73. These genes are denoted
"common" genes. The number of genes regulated, respectively, by
either p53 or p73 are shown in the upper part of the
circle, whereas genes that passed the cutoff (2-fold
overexpression at 1 or 2 time points only) for one of the transcription
factors (p73 or p53) are shown in the shaded small
circles.
|
|
View this table:
[in this window]
[in a new window]
|
Table I
Target genes modulated by p73 and p53
The data were derived from affymetrix chip analysis and show fold
change of expression at the indicated time points for 85 genes (see
Fig. 3). The letter C indicates genes that were clustered together with
the p73-only genes (10 genes, see "Results"). The asterisks
(p73*) indicates genes that were regulated by p73 and
partially regulated by p53 (33 genes, Fig. 3A) but only once
or twice above 2-fold. HIV, human immunodeficiency virus; MDR,
multidrug resistance; PML, promyelocytic leukemia.
|
|
Identifying the common, p53-only, and p73-only Groups of Genes by
Clustering Analysis--
The grouping of induced genes (Fig. 3) was
based on an arbitrary cutoff, and it seemed conceivable that clustering
the genes according to their correlated expression profiles may yield a better understanding of their functional role. Cluster analysis (super-paramagnetic clustering) was done on the 211 genes up-regulated by p53 or p73 or both (see Fig. 3). Each of the 211 genes is
represented by 8 components, which represent the expression ratio at
each time point over that of 0 h (before the addition of the inducer).
The results are summarized in the dendrogram of Fig.
4A. The parameter T
controls the resolution at which the data are viewed. At
T = 0 all the 211 genes are in a single cluster; as
T increases, large groups split into smaller ones. When we
ordered the genes according to their position in the dendrogram,
i.e. rearrange the rows of the expression data matrix
according to the order imposed by the clustering process, the
color-code matrix of Fig. 4B is obtained. In Fig.
4A, the boxes indicate clusters that contain at
least five genes, and each box is colored according to its "purity," the percentage of the members of a given group
(e.g. the 25 genes of the p73-only group, see Fig.
3A) among the genes contained in the corresponding cluster.
The cluster of the p73-only genes (cluster a) contains 23 (92%) of the
25 genes identified in Fig. 3. The position of the members of this
group is marked by red ×s, and their change of expression
is shown in the matrix of Fig. 4B in the area marked p73.
Similar results were also obtained for the p53-only genes. Of the 63 genes identified in Fig. 3 as p53-only, 52 (83%) are contained in
cluster b (Fig. 4A).
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 4.
Clustering results using the
super-paramagnetic clustering algorithm for the 211 genes that were
up-regulated at 3 or more time points upon ponasterone A activation of
p73 or p53. A, dendrogram of
the genes that includes clusters of size 5 and larger. Each cluster is
represented by a box colored according to the percent of
p73-only target genes (25 genes) contained in the cluster. The
distribution of these 25 genes is marked by red ×'s at the
right. The distribution of the common genes (27 genes up-regulated by
both p73 and p53) is shown by the black ×'s. The
clusters marked by a and b were used to plot the
expression profile of the genes in that cluster (see Fig. 5).
B, the expression matrix of the genes according to the
dendrogram on the left. The color represents induction
(red) or repression (blue). T is a
parameter of the super-paramagnetic clustering algorithm that controls
the resolution at which the cluster is found (36). The genes activated
by either p73 or p53 as defined by the clustering procedure are
shown on the right-hand side of the expression matrix.
|
|
An illustration of the advantage of clustering in pointing out genes
that show correlated expression in contrast to the grouping by
arbitrary filtering (Fig. 3A) is shown in Table I and Fig. 4. The cluster that contains most of the p73-specific genes (23 of 25)
also includes an additional 10 genes that show a similar expression
profile (marked by "C" in column 2 of Table I). Only 2 of these 10 genes are common, whereas the other 8 can be considered p73-specific,
although they also show once or twice expression levels that are
above 2-fold over the 0-h time point in the p53 induced cell
line. Two of the p73-specific genes defined in Fig. 3 diverge from
cluster a (Fig. 4B) since their profile is different (these
are COL5A2 and GPPK5, see Table I). Hence the clustering may group
together co-regulated genes that may escape the grouping by the use of
the filter threshold. This clustering procedure expanded the group of
p73-only by recruiting genes that are co-regulated and may be
p73-specific.
We selected representative clusters from Fig. 4A to analyze
the profile of their expression patterns. Fig.
5 shows the expression profile of
clusters a (p73-only) and b (p53-only). The average expression profiles
of these clusters correlate with their definitions as shown in Fig.
5.
View larger version (9K):
[in this window]
[in a new window]
|
Fig. 5.
Average expression profile of genes in
clusters a and b of Fig.
4A. The expression profile of each
gene in the cluster was normalized as described under "Experimental
Procedures." The profiles provide representative examples of the
expression profiles of the groups of genes defined as p73-only
(a) and p53-only (b) obtained by the clustering
shown in Fig. 4A.
|
|
The group of genes designated common (indicated by black
×'s in Fig. 4A) is clearly not homogeneous and segregates
into several clusters (Fig. 4B). It is possible that such
subdivision spreading represents different affinities of the target
sites of these genes for either p53 or p73. This indicates that the
filtering of genes according to an arbitrary threshold is not always
sufficiently informative with regard to the profile and level of their
expression, whereas the clustering analysis allows a better
classification of the response of the target genes.
Classification of the Genes Up-regulated by p53 and p73--
The
functional classification and expression data for the 85 genes
up-regulated in the ponasterone-induced cell lines (see Fig.
3A) is presented in Table I as genes activated by p73 or both by p53 and p73. The p73-induced genes, which also show partial response to p53 (i.e. only once or twice more than 2-fold
over control) are indicated by an asterisk (p73*) in Table I. The list
does not contain the genes activated by p53 except the 27 common genes
(Fig. 3). The full list of the 211 genes (Fig. 3) used for the
clustering analysis in the numerical order shown in Fig. 4B,
with their fold expression, is available upon request.
It has been previously shown that the cell line used in this study,
H1299, is resistant to p53-induced apoptosis (Ref. 29 and Fig.
1C). Cell cycle analysis shows that p73 also does not lead
to programmed cell death in H1299 cells (Fig. 2) and, indeed, there is
no induction of a significant number of genes known to be related to
apoptosis. On the other hand, p21waf1 is highly expressed in
response to both p53 and p73 induction and may be responsible for the
growth arrest observed after ponasterone A addition (Fig. 2).
We compared the results of Table I to those previously reported on
p73-induced genes by Vikhanskaya et al. (37). Of 16 genes
that were found to be up-regulated by p73, only PIG 3 was found to be
common to our work. Such discrepancy may be related to the different
cell line (ovarian) and p73-overexpressing system (stably transfected
clones) used in that work (37).
From the list in Table I and from Figs. 4 and 5 it is clear that there
are common and distinctive genes activated by p73 and p53. The
function of these distinct genes may be one of the reasons for the
different phenotypes of p73 and p53 knock-outs in mice.
To confirm the microarray data, reverse transcriptase analyses (RT/PCR)
were performed on a pool of activated genes with an aliquot of the RNA
used in the DNA chip analysis. As shown in Fig.
6A, the
transcripts of the genes CaN19, 14-3-3, PIG-3,
1-antitrypsin, and Jag2 are specifically induced by
p73 but not by p53, whereas the p21waf1 and PTGF-
(indicated as "TGF- superfamily protein" in Table I) transcripts
were activated by both proteins. These results overlap significantly
with those obtained by microarray analysis (Table I). The expression
profile of the above-mentioned genes was also analyzed by RT/PCR using
RNA from additional p73 (H1299-p73 clone 11) or p53 (H1299-p53
clone 16) clones and from the control H1299-pIND clone 1. We found
(Fig. 6B) that the CaN19, 14-3-3, PIG-3,
1-antitrypsin, and JAG2 transcripts as well as those of p21waf1 and PTGF- were induced in H1299-p73 clone 11 and
H1299-p53 clone 16 with a similar kinetic and specificity as those
reported in Fig. 6A. No induction of any of these
transcripts was detected in H1299-pIND clone 1 cells (Fig.
6C). These results together with those of the microarray
analysis indicate that overexpression of p73 or p53 in the same
cellular context promotes distinct and partially overlapping gene
expression profiles.
View larger version (39K):
[in this window]
[in a new window]
|
Fig. 6.
Analysis of induced genes upon ponasterone
A-induced expression of p73 or p53. RNA
was extracted from the indicated cell lines at the specific time points
after the addition of ponasterone A and subjected to a RT/PCR reaction
as indicated under "Experimental Procedures." A, RT/PCR
analysis of an aliquot of the RNA from H1299- p73 clone 9 and
H1299-p53 clone 23 used in the microarray analysis. B,
RT/PCR analysis of RNA from additional p73- and p53-expressing cells,
H1299-p73 clone 11, and H1299-p53 clone 16. C, RT/PCR
analysis of RNA from H1299-pIND clone 1. The number of the PCR cycles
employed for each gene is indicated on the left side. The
length of the amplified fragments is indicated on the right
side.
|
|
p73 Is Recruited Directly onto Its Target Genes--
Using
MatInspector Professional software (genomatrix.gsf.de) to analyze
TRANSFAC 5.0 data base (transfac.gbf.de/TRANSFAC) we examined whether
p53/p73 consensus sites were contained within the promoter region or
the first intron of the pool of activated genes analyzed in RT-PCR
assays. As shown in Table II, CaN19, 14-3-3, PIG-3, 1-antitrypsin, p21waf1,
PTGF-, and JAG2 genes contain some putative or already characterized p53/p73 binding sites within their promoter or first intron regions. To
verify whether p73 and p53 are able to directly bind their consensus
sequences in vivo, we performed chromatin
immunoprecipitations. Cross-linked chromatin from H1299-p73 clone 9 or H1299-p53 clone 23 was immunoprecipitated with the indicated
antibodies (Fig. 7). We found that p73
specifically binds the regulatory regions of CaN19 (up and down),
14-3-3, p21waf1 (up and down), 1-antitrypsin,
PTGF-, PIG-3, and JAG2 (Fig. 7). No specific binding to any of the
analyzed regulatory regions was revealed in the chromatin
immunoprecipitates with anti-green fluorescent protein serum (Fig. 7).
Unlike p73, p53 binds only the two consensus motifs on the
p21waf1 promoter (Fig. 7). The thymidine kinase (TK)
promoter, which does not contain any p53/p73 consensus site, was used
as negative control, and indeed, p73 and p53 were not recruited to
such a promoter (Fig. 7, lower panels). Of note, PTGF-
whose transcript was induced by p73 or p53, does not directly
recruit p53, at least in our cell system and experimental conditions.
However, it is not excluded that the transcriptional activation could
occur through some other p53/p73 binding sites present along PTGF- gene, since our search for consensus sequences was confined to the
promoter and first intron regions. These results partially overlap with those derived from the microarray analysis and strongly contribute to the identification of genes that are direct targets of
p73.
View this table:
[in this window]
[in a new window]
|
Table II
Location and sequence of p73/p53 binding sites in the regulatory
regions of the indicated genes are reported
|
|
View larger version (68K):
[in this window]
[in a new window]
|
Fig. 7.
In vivo recruitment of
p73 and p53 to their target genes.
Cross-linked chromatin was extracted either from H1299-p73 clone 9 or H1299-p53 clone 23 upon induction with ponasterone A (2.5 µM) and subjected to immunoprecipitation with the
indicated antibodies. The length of the amplified fragment is indicated
on the left side. TK, thymidine kinase.
|
|
| |
DISCUSSION |
The recent discovery of two p53 homologues, p73 and p63, has
established a new family of transcription factors. Furthermore, p73 and
p63 are subject to alternative splicing, giving rise to a complex
family of proteins that might exert distinct as well as overlapping
functions (38, 39). This obviously adds a level of complexity to p53
signaling in normal and cancer cells. In the present study, we have
compared the gene expression profiles promoted by ectopic expression of
p73 or p53 in the same cellular context, H1299 cells. We found that
overexpression of p73 and p53 modifies the expression of 141 and 320 genes, respectively. A further analysis of these genes revealed that
p73 up-regulates 85 genes, of which 25 are specific and 27 are in
common with p53 regulation. Our results, together with the already
described p53 target genes, contribute to the identification of a large
repertoire of genes that serve as transcriptional mediators of the p53
family activities. We further demonstrate that in p53-null cells,
p53 target genes such as 14-3-3, p21waf1, PIG3, and PTGF-
(40-43) can directly recruit and be activated by p73.
Although a certain level of knowledge regarding protein-protein
interactions among the p53 family members in cancer cells has been
provided by in vitro and in vivo studies, there
is still a lack of basic information on the transcriptional cross-talk among p53 and its close relatives (44-48). A recent report has clearly
shown that DNA damage resulting in p53-dependent apoptosis requires p73 and p63 activities (20, 49) and that in p73/ and
p63/ cells, p53 is not recruited to apoptotic target genes such as
Bax, PERP, and NOXA. It has
also been reported that DNA damage-induced acetylation of p73
selectively alters the choice of target genes of this protein (19).
Here we report the identification of novel p73 target genes such as
CaN19, 1-antitrypsin, and JAG2. These genes are not
up-regulated by p53, at least under our experimental conditions, and
indeed, unlike p73, p53 was not found to occupy their regulatory
regions. The gene product of CaN19 is a member of the S100 family
proteins and was originally isolated from primary human keratinocytes
by subtractive hybridization (50); it seems to be involved in skin and
regenerative differentiation and may also play a role in suppressing tumor cell growth (51, 52). To date no sufficient information is
available for a complete understanding of the molecular mechanisms underlying its biological activities. We are currently investigating the functional link between p73, p63, and CaN19 during keratinocytic differentiation.
1-Antitrypsin is the major serine proteinase inhibitor
(serpin A1) in human plasma. Its target proteinase is neuthrophil elastase, and its main physiological function is the protection of the
lower respiratory tract from the destructive effects of neuthrophil
elastase during an inflammatory response (53-55). p73-deficient mice
suffer from inflammation, but very little is known on the pathogenesis
of such process. A rather speculative hypothesis would suggest that
p73-mediated anti-inflammatory effects might include the induction of
1-antitrypsin.
The finding that JAG2 gene is a target of p73 correlates with a
recent study showing that JAG1 and JAG2 are specific p63- and
p73-responsive genes (56). JAG1 and JAG2 genes encode transmembrane proteins that serve as ligands for Notch receptors (57). Our findings
together with those of Sasaki et al. (56) indicate that
p73, p73, and p63 but not wt-p53 up-regulate JAG-1 and JAG-2
genes. Mutations in the Notch ligands cause developmental defects.
Indeed, JAG2-deficient mice exhibit defects of limb and craniofacial
development that closely resemble the abnormalities of ectrodactyly
ectodermal dysplasia patients carrying p63 mutations (58, 59).
Thus, JAG1 and JAG2 are direct transcriptional targets of either p73 or
p63, a finding that suggests a direct involvement of these activators
in Notch signaling pathways.
Among the genes induced by both p73 and p53, PTGF- is quite
peculiar. Although the PTGF- transcript is induced by both p73
and p53 (Fig. 6, A and B), at least under our
experimental conditions only p73 was recruited directly to the gene
(Fig. 7). Up-regulation of the PTGF- transcript has been previously reported and linked to p53-mediated growth suppression through an
autocrine as well as paracrine mechanism (40, 42). It has also been
reported that induction of PTGF- can occur through a p53-independent
mechanism that, in agreement with our findings, might be induced by
p73. TGF- is a family of secreted factors that play pivotal
functions during embryonic development and adult tissue homeostasis
(60). Despite the heterogeneity of TGF- mediated cellular responses,
these cytokines signal to the nucleus through a quite simple mechanism.
Ligand activation of TGF- receptors results in the nuclear
translocation of SMAD family proteins that control target gene
expression (61). A certain level of specificity of the effects mediated
by TGF- cytokines might be dictated by different transcriptional
activators. Thus, p53, p73, and probably p63 induction of PTGF-
might result in quite distinct effects ranging from growth arrest to
the regulation of development and homeostasis.
§,
,
TOP
ABSTRACT
INTRODUCTION
Partially supported by the German-Israel Science Foundation
and the Israel Science Foundation.
