Received 2 February 1999/Returned for modification 26 March
1999/Accepted 5 May 1999
The p20K gene is induced in conditions of reversible growth arrest
in chicken embryo fibroblasts (CEF). This expression is dependent on
transcriptional activation and on a region of the promoter designated
the quiescence-responsive unit (QRU). In this report, we describe the
regulatory elements of the QRU responsible for activation in resting
cells and characterize the trans-acting proteins
interacting with these elements. We show that the QRU consists of
functionally distinct domains including quiescence-specific and weak
proliferation-responsive elements. The quiescence responsiveness of the
QRU was mapped to two C/EBP binding sites, and the activity of the p20K
promoter and its QRU was inhibited by the expression of a dominant
negative mutant of C/EBP
in nondividing cells. The activation of QRU
in response to serum starvation and contact inhibition correlated with
the presence of a growth arrest-specific complex in electrophoretic
mobility shift assays. This complex was supershifted by antibody for
C/EBP. C/EBP accumulated in conditions of contact inhibition as a
result of transcriptional activation. Therefore, C/EBP was itself
regulated as a growth arrest-specific gene in CEF. Finally, we show
that the expression of p20K is regulated by linoleic acid, an essential
fatty acid binding to p20K. The addition of linoleic acid to
contact-inhibited CEF markedly repressed the synthesis of p20K without
inducing mitogenesis. The activity of the QRU was inhibited by linoleic acid or the peroxisome proliferator-activated receptor PPAR
2 in
transient expression assays. Therefore, we have identified C/EBP as
a key activator of a growth arrest-specific gene in CEF and implicated
an essential fatty acid, linoleic acid, in regulation of the QRU and
the p20K lipocalin gene.
 |
INTRODUCTION |
Mitogenic stimulation is
characterized by extensive changes in gene expression. Gene induction
occurs within minutes of the addition of a mitogen or growth factor,
often independently of de novo protein synthesis (3, 21, 38,
41). Many of the immediate-early genes expressed at the
G0/G1 transition code for a transcription
factor which has a viral or cellular oncogenic counterpart. Growth
arrest caused by serum starvation or contact inhibition is also
characterized by changes in gene expression. While the gene products
up-regulated during mitogenesis are repressed in nondividing cells, a
different set of genes, referred to as growth arrest-specific (GAS),
growth arrest- and DNA damage-inducible (GADD), or
quiescence-specific genes (8, 9, 22, 31, 77), are also
induced in conditions of reversible growth arrest or G0.
The function of their gene product is generally poorly understood, but
proteins of the extracellular matrix (18, 22, 58), proteins
with high affinity for lipids (9, 13, 41, 73), proteins
acting as negative regulators of cell proliferation (4, 13, 24,
34) or survival factors (35, 60), or proteins capable
of enhancing the response of quiescent cells to mitogens
(51) are part of the growth arrest-specific program of gene expression.
Despite the recent efforts in the cloning and characterization of the
GAS genes, little is known about the regulatory mechanisms controlling
their expression. Many of the GAS genes are regulated at the
posttranscriptional level in serum-starved cells and may depend on
transcript stabilization for expression (31, 80). Others are
regulated at multiple levels including transcriptional activation
(25, 31, 56, 58). The p53 tumor suppressor induces growth
arrest in cells irradiated with rays and activates the
transcription of the p21waf1 and
gadd45 genes (28, 43). However, these genes are
also induced in response to serum starvation even in cells devoid of functional p53 (95). Therefore, multiple mechanisms,
signaling pathways and trans-acting factors may control the
expression of p21waf1, gadd45, and
other GAS genes. Thus far, the trans-acting factors and
regulatory mechanisms controlling the expression of the GAS genes in
G0 have remained elusive, yet their characterization is
likely to be important for our understanding of growth control and
reversible growth arrest, in particular.
We have characterized the expression of a secretory protein, p20K,
expressed by quiescent chicken embryo fibroblasts (CEF) and chicken
heart mesenchymal (CHM) cells. While the function of p20K in quiescent
cells is unknown, it is a member of the lipocalin family of lipid
binding proteins (9, 30). The expression of p20K is markedly
elevated in quiescent CHM cells and is rapidly inhibited in response to
mitogenic stimulation. p20K is also synthesized by serum-starved or
density-arrested CEF. In contrast, compounds which inhibit DNA
synthesis, such as hydroxyurea, are poor inducers of p20K synthesis,
suggesting that growth arrest in G0 but not G1/S controls the expression of this gene (56).
In addition, p20K is not synthesized by senescent CEF or CEF undergoing
apoptosis (7, 52), indicating that p20K is a marker of
reversible growth arrest, i.e., of cells capable of reentering the cell
cycle following mitogenic stimulation and the establishment of
favorable growth conditions (5).
The expression of p20K is regulated at the transcriptional level in
CEF. We have previously identified a region of the p20K promoter
capable of conferring quiescence responsiveness to a heterologous,
minimal promoter. This 48-bp region, termed the quiescence-responsive
unit (QRU) of the p20K gene, is both essential and sufficient for
activation in conditions of serum starvation and contact inhibition in
CEF and rat FR3T3 fibroblasts (7, 56). In this report, we
describe the characterization of the QRU and trans-acting
factors controlling its activity in quiescent cells. We identified
C/EBP (designated NF-M [nuclear factor myeloid] for the chicken
homolog) as an essential regulator of p20K expression in normal,
density-arrested CEF and serum-starved Rous sarcoma virus
(RSV)-transformed CEF and provide evidence that the expression of
C/EBP is itself regulated during reversible growth arrest. We show
that linoleic acid, an essential fatty acid (EFA) binding to p20K
(17), inhibits the activity of the QRU and expression of
p20K in contact-inhibited cells. These results establish C/EBP as a
key regulator of a GAS gene in CEF and suggest a role for EFAs in
control of the p20K gene.
| |
MATERIALS AND METHODS |
Cells and viruses.
Early passages of CEF were cultured at
41.5°C in Richter-improved minimal essential medium containing
insulin and zinc (I+ medium; Irvine Scientific, Santa Ana,
Calif.), 5% heat-inactivated newborn bovine serum (MediaserII;
Montreal Biotech Inc., Kirkland, Quebec, Canada), 5% tryptose
phosphate broth, glutamine, penicillin, and streptomycin. CEF were also
infected with the Schmidt-Ruppin A strain of RSV (SR-A RSV) or with
recombinant viruses generated with the RCASBP(B) retroviral vector.
RSV-infected CEF were starved in the complete absence of serum in
Dulbecco's modified Eagle's medium containing 10% tryptose phosphate
broth. To maintain a constant pH, 50 mM HEPES (pH 7.4) was added
routinely to the culture medium of RSV-transformed cells. Cells,
cultured in serum-containing medium, were also treated overnight (16 h)
with various concentrations (25 to 200 µM) of fatty acids dissolved
in ethanol. Cells with lipid-containing droplets were fixed in 3.7%
formaldehyde in phosphate-buffered saline (PBS) and then stained for
1 h with a 0.25% solution of oil red O dissolved in 40% ethanol
(71).
Promoter and expression vector constructs.
The avian
C/EBP cDNA was inserted in the Cla12 adapter plasmid and subcloned
in the unique ClaI site of the RCASBP(B) retroviral vector
to generate plasmid RCASBP-C/EBP (16, 69). The resulting vector was transfected in CEF, and a productive infection was established by the replication-competent avian retrovirus. The same
procedure was followed to generate the RCASBP-20K virus expressing the
p20K gene. The p12E CAT (chloramphenicol acetyltransferase) reporter
plasmid containing 2.3 kb of the 5' flanking region of the p20K gene
was described previously (56). Constructs of the QRU of the
p20K promoter (corresponding to positions 
217 to 169) were
generated with plasmid pJFCAT-TATA, which includes a minimal promoter
consisting of a TATAAAA box and the initiation start site of
the human -globin gene. To generate these minimal promoter constructs, synthetic double-stranded oligonucleotides representing various regions of the QRU were multimerized and inserted in the HindIII site of plasmid pJFCAT-TATA. The intact QRU and
several mutant derivatives of this 48-bp DNA fragment were also
synthesized chemically and inserted in the SalI site of the
same plasmid. These mutations are shown in Fig.
1.
View larger version (32K):
[in this window]
[in a new window]
|
FIG. 1.
Schematic representation of the QRU. Potential binding
sites for the C/EBP and Ets families of transcription factors as well
as the position of a palindrome are indicated. The QRU was arbitrarily
divided into three overlapping regions designated regions A, B, and C
and analyzed in transient expression assays. Mutations generated in the
QRU are indicated by lowercase letters.
|
|
Transient expression assays and conditions of quiescence.
All transfections were done by calcium phosphate precipitation as
described previously (36, 56). Briefly, dense cultures of
normal or RSV-transformed CEF grown in 100-mm-diameter dishes were
transfected with a total of 30 µg of DNA consisting of 10 µg of
test reporter plasmid, 2 µg of the LacZ-containing plasmid pCH110,
and 18 µg of salmon sperm carrier DNA. In experiments with expression
plasmids, 2 to 10 µg of plasmid pCDM8 encoding avian C/EBP
or -
or the peroxisome proliferator-activated receptor (PPAR) expression
plasmid pCMV2-PPAR2 were included in the transfection mixture. The
plasmid encoding the dominant negative mutant of C/EBP (
184) was
described by Kowenz-Leutz et al. (48); 10 µg of the 184
expression plasmid was cotransfected as described above. For all
transient expression assays, the activity of the CAT reporter enzyme
was determined in lysates representing equal levels of
-galactosidase activity. All constructs were analyzed in duplicate
or triplicate in at least two separate experiments. The conversion of
chloramphenicol into its acetylated forms was quantitated in an
InstantImager (Packard/Canberra). The results of previous studies
indicated that the simian virus 40 enhancer/promoter of plasmid pCH110
is not regulated in our conditions of proliferation and quiescence and
therefore can serve as an internal standard to control for differences
in transfection efficiency.
The transfection of serum-starved or density-arrested CEF resulted in a
poor expression of any of the promoter constructs investigated thus
far, including reporter genes driven by strong, constitutive
promoter/enhancers used as internal standards (55). Therefore, two different protocols were developed to induce quiescence in actively dividing CEF transfected with p20K promoter constructs. Our
original procedure was based on the transfection of dense RSV-transformed CEF (56). In this protocol, quiescence is
obtained by resuspending the cells in serum-free medium the day after
transfection and preparing the cell lysate 48 h after
transfection. Dense cultures of RSV-transformed CEF deplete the medium
of essential nutrients and growth factors more rapidly than their
normal counterparts and do not need to be starved for more than 24 h to become quiescent and express p20K. A second protocol was developed
to study the activation of the p20K promoter in normal,
density-arrested CEF. In this protocol, cultures of normal CEF grown at
approximately 80% confluence were transfected by the procedure
described above. Calcium phosphate precipitation is essential in this
protocol because calcium acts as a strong mitogen for the cells, which must reach confluence and become contact inhibited before preparation of the lysate. In these conditions, the normal CEF reach confluence the
day after transfection, at which time the cells are refed with
serum-containing medium to maximize contact inhibition at the time of
cell lysis (48 h after transfection). Actively dividing normal or
RSV-transformed CEF were obtained by trypsinization of the culture the
day after transfection and replating of the cells at one-sixth of the
original density in fresh serum-containing medium. Both protocols
yielded the same results unless otherwise specified in the text.
Results obtained with the starvation of RSV-transformed CEF or contact
inhibition of normal CEF are presented to emphasize specific points or
differences relevant to the activation of the QRU.
Electrophoretic mobility shift assay (EMSA).
Nuclear
extracts were prepared by a modification of the methods of Briggs et
al. (14) and Dignam et al. (27). Routinely, 107 cells were washed once with cold PBS, collected for 5 min in a microcentrifuge at 1,500 × g, and resuspended
in 500 µl of buffer A (10 mM HEPES [pH 7.9], 10 mM KCl, 0.5 mM
dithiothreitol [DTT], 1.5 mM MgCl2, cocktail of protease
and phosphatase inhibitors consisting of 0.5 mM phenylmethylsulfonyl
fluoride, 0.3 µg of leupeptin per ml, 0.3 µg of antipain per ml,
0.5 µg of aprotinin per ml, 1 mM sodium fluoride, and 1 mM sodium
vanadate). Cells were lysed with a glass Dounce homogenizer (B pestle),
and the nuclei were pelleted by centrifugation at 12,000 × g for 20 min. The nuclear pellet was extracted with 1 ml of buffer
C (20 mM HEPES [pH 7.9], 25% glycerol, 1.5 mM MgCl2,
0.42 mM KCl, 0.5 mM DTT, 1 mM EDTA, protease and phosphatase inhibitors
as mentioned above) for 20 min at 4°C with continuous agitation. At
the end of the incubation, the reaction mixture was centrifuged at
12,000 × g for 20 min, and the supernatant was
dialyzed against buffer D (20 mM HEPES [pH 7.9], 20% glycerol, 0.1 M
KCl, 0.5 mM DTT, 0.2 mM EDTA, protease and phosphatase inhibitors).
Alternatively, the nuclear lysate was partially purified by ammonium
sulfate precipitation. In this case, the nuclear lysate was centrifuged at 80,000 × g for 1 h at 4°C, and 0.33 g
of ammonium sulfate was added slowly per ml of the resulting
supernatant. The mixture was then incubated for 30 min at 4°C with
gentle agitation and finally centrifuged at 21,000 × g
for 20 min. The supernatant was discarded, and the pellet was
resuspended in buffer D and dialyzed against the same buffer for 5 h at 4°C. Both methods gave the same results. The dialysate from the
purified nuclear extract was aliquoted and stored at 80°C until
needed. Protein concentration was determined by the Bradford assay
using bovine serum albumin as a standard.
Sequences of synthetic DNA oligomers used for DNA binding analysis are
shown below in Table 1; mutations of the
QRU are described in Fig. 1. The NF-M oligonucleotide is a composite
element derived from the two NF-M (C/EBP) binding sites of the
chicken myelomonocytic growth factor (cMGF) promoter (82).
All probes were generated by filling in the ends with Klenow enzyme,
using [32P]dATP, [32P]dCTP, and unlabeled
dGTP and TTP. Binding reactions were performed in a total volume of 20 µl of 1× binding buffer (20 mM Tris-HCl [pH 7.5], 1 mM
MgCl2, 60 mM NaCl, 5% glycerol) for 30 min at room temperature. For EMSAs, the 32P-end-labeled
oligonucleotides (approximately 10,000 cpm/0.1 ng) were incubated with
2 to 5 µg of nuclear extracts in the presence of 2 µg of
poly(dI-dC). Competition binding reactions were performed by
preincubating the nuclear extract with an excess of unlabeled, filled-in oligonucleotides for 15 min at room temperature. This was
followed by the addition of labeled oligonucleotides and incubation for
an additional 30 min. The DNA-protein complexes were resolved on 4.8%
nondenaturing polyacrylamide gels in 0.5× Tris-borate-EDTA buffer and
visualized by autoradiography.
Immunoprecipitation, Western blotting analysis, and production of
C/EBP antiserum.
The synthesis of p20K was investigated by
immunoprecipitation analysis of proteins metabolically labeled with
[35S]methionine followed by separation on sodium dodecyl
sulfate (SDS)-polyacrylamide gels and fluorography as described
previously (56). For Western blotting analyses, 40 µg of
total cell protein extract prepared in SDS sample buffer was subjected
to SDS-polyacrylamide gel electrophoresis and blotted on nitrocellulose
(BA85; Schleicher & Schuell). A polyclonal antibody for avian C/EBP
was kindly provided by K.-H. Klempnauer and used at a dilution of
1:2,000 in a 0.1% solution of milk dissolved in PBS (42).
This was followed by incubation with a peroxidase-conjugated secondary
antibody (Pierce, Rockford, Ill.) and detection with an enhanced
chemiluminescent substrate according to protocols provided by the
manufacturer (Amersham). Antibodies for avian C/EBP were also
generously provided by A. Leutz or generated in a rabbit by using a
recombinant avian C/EBP protein as the antigen and established
immunization protocols (8). Antibodies for ERK-1 and ERK-2
(SC-93 and SC-154, respectively) were obtained from Santa Cruz
Biotechnology (Santa Cruz, Calif.).
Northern blotting and transcription run-on analyses.
RNA was
purified by high-salt urea precipitation and analyzed on Northern blots
as described by Mao et al. (56). Run-on transcription assays
were performed as described by Dehbi et al. (23). In both
analyses, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as
an internal standard to control for loading or hybridization. The
intensity of radioactive bands on blots was quantitated with an
InstantImager (Packard/Canberra).
Screening of cDNA library.
A contact-inhibited CEF cDNA
library (9) was screened with the SacII
restriction fragment of the chicken C/EBP gene corresponding to the
DNA binding and leucine zipper domains of the protein. A similar screen
was done with a DNA fragment of the murine CHOP-10/gadd153 cDNA (also corresponding to the DNA binding domain of the protein; 75). The final wash of nitrocellulose filters
harboring phage DNAs was performed with 0.5× SSC (1× SSC is 0.15 M
NaCl plus 0.015 M sodium citrate)-0.1% SDS at 55°C (C/EBP) or
with 2× SSC-0.1% SDS at room temperature (CHOP-10). Following
secondary and tertiary screens of positive phages, excision of
phagemids was accomplished with the ExAssist helper phage according to
protocols provided by the manufacturer (Stratagene, La Jolla, Calif.).
Following excision, clones were analyzed by a combination of
restriction enzyme and sequencing analyses followed by FastA database
analysis (68).
| |
RESULTS |
Characterization of functional domains of the QRU.
The region
of the promoter required for activation of the p20K gene in quiescent
cells was termed the QRU. The 48-bp QRU confers quiescence
responsiveness to an heterologous promoter in conditions of serum
starvation or contact inhibition and is necessary and sufficient for
induction in growth-arrested CEF and rat FR3T3 cells (7,
56). EMSAs indicated that numerous DNA binding proteins interact
with the 48-bp QRU in conditions of quiescence and logarithmic
proliferation (55). To begin the characterization of the
QRU, we identified potential binding sites for known transcription factors by using the program Signal Scan (72) and performed a series of transient expression assays with individual segments of the
QRU. The analysis of potential binding sites is presented in Fig. 1 and
2. Two consensus elements for the C/EBP
and Ets families are present within the QRU. A sequence located at the 3' end of the QRU (GCGTCTG) shows similarity to the core
sequence recognized by the Maf family of transcription factors
(44, 70). A palindrome (CCTCAGG) with no
extensive similarity to known regulatory elements is also found at the
5' end of the 48-bp fragment. To identify functional domains within the
QRU, we then arbitrarily defined three partially overlapping segments
designated regions A, B, and C and analyzed their activity in transient
expression assays. Region A includes the first C/EBP binding site and
the palindrome located at the 5' end of the QRU. Region B contains the
two potential Ets response elements and the second C/EBP binding site.
Region C includes the second C/EBP binding site and the potential Maf
recognition element. Synthetic double-stranded oligonucleotides corresponding to each region were multimerized, inserted in proximity to a minimal promoter, and investigated in actively dividing or serum-starved RSV-transformed CEF. A construct containing two copies of
region A was strongly activated in quiescent cells, with a 40-fold
induction over the level observed in actively dividing cells (construct
A-2 in Fig. 3). In contrast, constructs
containing two or three copies of region B or C did not respond to
growth arrest and were in fact more active in proliferating cells. This proliferation-dependent activity was more obvious with region B but
remained modest when compared to the strong activity of region A in
nondividing cells. Therefore, region A contains a potent
quiescence-responsive element, but the QRU is composed of multiple,
functionally distinct domains.
View larger version (10K):
[in this window]
[in a new window]
|
FIG. 2.
Sequence similarity between the QRU and the binding
sites of known transcription factors. The consensus binding sites of
known transcription factors were obtained from reference
29 and other references indicated in the text.
Numbers indicate positions and strands of the elements of the QRU
(e.g., 199 to 207 refers to the minus strand, whereas 189 to
181 indicates the position of the element on the plus strand). N
represents any nucleotide.
|
|
View larger version (22K):
[in this window]
[in a new window]
|
FIG. 3.
Characterization of functional elements of the QRU.
Double-stranded oligonucleotides corresponding to various regions of
the QRU were cloned in the minimal promoter plasmid pJFCAT-TATA and
investigated in transient expression assays. For each construct, the
number of inserted copies of region A, B, or C or µB-Ets is indicated
by a numeral and the orientation is indicated by or +.
Normalized values of CAT activity expressed as the means of duplicate
samples are indicated for serum-starved CEF (0% serum) and actively
dividing RSV-transformed CEF (5% serum).
|
|
Since several members of the Ets family have been implicated in gene
activation in response to mitogens (12, 90), we synthesized a modified version of the region B oligonucleotide containing a
mutation in both potential Ets binding sites but affecting noncritical residues of the putative C/EBP binding site (oligonucleotide µB-Ets in Fig. 1 and constructs µB+2 and µB+3 in Fig. 3) (33, 66, 84). In transient expression assays, mutation of the potential Ets binding sites reduced the activity of region B in actively dividing
cells, in agreement with the notion that one or multiple proliferation-responsive elements are also present in the QRU. The
activity of regions B and C was observed in RSV-transformed CEF but not
in normal, actively dividing cells (46). Therefore, the
proliferation-specific activity of regions B and C may be restricted to
v-src-transformed cells. This activity was not characterized further.
C/EBP is required for activation of the QRU.
The
overexpression of several members of the Maf (c-MAFI, c-MAFII, MAF B,
MAF F, MAF G, and MAF K) and Ets (PEA3 and c-Ets-1) families did not
affect the activity of the QRU in CEF (46). In contrast,
members of the C/EBP family were potent activators of the QRU. Results
shown in Fig. 4
demonstrate the action of C/EBP and - on the p20K promoter
(p12E), on the 48-bp QRU, and on individual regions of the QRU. The QRU
constructs studied in this analysis are depicted in Fig. 4A. C/EBP
and - were strong activators of p12E, the QRU, and a multimer of
region A (construct 3XA in Fig. 4). Likewise, regions B and C
(constructs 3XB and 3XC) were also activated by C/EBP, but to a lesser
extent (35- to 60-fold activation for region A, versus 3- to 12-fold
for regions B and C). pJFCAT-TATA, the parental promoter construct, did
not respond to the overexpression of C/EBP or -. The replacement of two nucleotides known to be critical for DNA binding by C/EBP markedly impaired the activation of a construct of region A by C/EBP
or C/EBP (construct 2XµA in Fig. 4). Constructs µB+2 and µB+3,
harboring mutations in the two potential Ets binding sites of region B,
were still activated by the overexpression of C/EBP or -
(7). We also studied the effect of a mutation in one or both
of the potential C/EBP binding sites of the QRU. As expected, the
mutation of both sites nearly abolished the activation of the QRU by
C/EBP (QRU-µ2C/EBP in Fig. 4). The single mutants were still well
activated, suggesting that both sites are effective targets of C/EBP
or -, at least in conditions where these factors are overexpressed
(QRU-µA and QRU-µB). The mutation of the palindrome of region A was
also activated efficiently (QRU-µpalA). This is not surprising,
considering that the palindrome does not overlap with the core sequence
of the C/EBP binding site of region A (Fig. 1). Therefore, we conclude
that the QRU contains two C/EBP-responsive elements and is strongly
transactivated by the and members of this family. In contrast,
none of the QRU constructs were affected by CHOP-10/Gadd153 when this
member of the C/EBP family was expressed alone or in combination with
C/EBP or - (46).
View larger version (28K):
[in this window]
[in a new window]
|
FIG. 4.
Activation of the p20K promoter and QRU by C/EBP. (A)
Schematic representation of promoter constructs studied in this
analysis. p12E is a construct of the p20K promoter comprising the
2289 to +42 region of p20K genomic DNA. All other constructs were
generated in the minimal promoter plasmid pJFCAT-TATA. (B) Effects of
C/EBP on the p20K promoter (p12E), the QRU, and various regions of the
QRU, determined by transient expression assays in actively dividing
CEF. Constructs of the QRU consist of one copy of the wt QRU or a
mutant derivative containing a mutation in one (QRU-µA or QRU-µB)
or both (QRU-µ2C/EBP) C/EBP binding sites. QRU-µpalA is a
derivative of the QRU containing nucleotide changes in the palindrome
identified in region A. Constructs 3XA, 3XB, and 3XC contain three
copies of regions A, B and C, respectively, inserted in the minimal
promoter plasmid pJFCAT-TATA. A similar construct was generated with a
derivative of region A containing a substitution of two critical
residues of the C/EBP binding site (2XµA). pJFCAT-TATA is the
parental CAT plasmid devoid of any element of the QRU. Cotransfection
was done with 10 µg of test plasmid, 10 µg of effector plasmid, and
2 µg of the LacZ-containing plasmid pCH110 in a 100-mm-diameter dish.
The activity of CAT was determined in duplicate samples for CEF
transfected with an expression vector for C/EBP or C/EBP or with
the parental expression vector, pCDM8.
|
|
To confirm the role of C/EBP in activation of the QRU, we studied the
effect of a mutation in a single or both C/EBP binding sites of the QRU
in actively dividing and contact-inhibited CEF. As shown in Fig.
5A, the quiescence-dependent activation
of the QRU was completely eliminated by the mutation of either one of the two C/EBP binding sites. Interestingly, the site located in region
B or C was essential when studied in the context of the QRU since a
mutation in this element abolished the activation by contact
inhibition. In contrast, the mutation of the palindrome of region A had
no effect on the activation by quiescence (construct QRU-µpalA in
Fig. 5A).
View larger version (16K):
[in this window]
[in a new window]
|
FIG. 5.
(A) Role of the C/EBP binding sites in the activity of
the QRU. Activities of QRU constructs containing a mutation in one
(QRU-µA or QRU-µB) or both (QRU-µ2C/EBP) C/EBP binding sites or
in the palindrome of region A (QRU-µpalA) were analyzed by transient
expression assays in actively dividing and contact- inhibited CEF. The
activation is indicated as the ratio of the activity in quiescent over
actively dividing cells (defined as 1). (B) Effect of a dominant
negative mutant of C/EBP on the activity of the p20K promoter.
Plasmids p12E, QRU, and AP1+3 were transfected alone, with a plasmid
encoding a dominant negative mutant of C/EBP (D184 for 184), or
with the parental vector (pCDM8) and analyzed by transient expression
assays in contact-inhibited CEF. Plasmid p12E is a construct of the
p20K promoter. Plasmid QRU contains a single copy of the QRU in the
minimal promoter vector pJFCAT-TATA; plasmid AP1+3 contains three
copies of the tetradecanoyl phorbol acetate-responsive element inserted
in the same vector. The results represent the average of duplicate
samples.
|
|
The activity of the QRU was then analyzed in the presence of a dominant
negative mutant of C/EBP (184). This mutant, constructed by
Kowenz-Leutz and coworkers, encodes a C/EBP protein lacking a
transactivating region (48). The expression of the deleted form of C/EBP markedly reduced the activity of a construct of the
QRU or p20K promoter (p12E) in contact-inhibited cells (Fig. 5B). In
contrast, plasmid 184 did not affect the activity of a construct
controlled by three copies of the tetradecanoyl phorbol acetate-responsive element (AP1+3). Likewise, the overexpression of a
dominant negative mutant of c-Jun (Tam-67 [2]) had no effect on the activity of the QRU in contact-inhibited cells
(46). Taken together, these results indicate that the
activation of the QRU is dependent on a member of the C/EBP family in
growth-arrested CEF.
C/EBP (NF-M) binds to the QRU in quiescent CEF.
To
determine if a member of the C/EBP family interacts with the QRU, we
performed a series of EMSAs with regions of the QRU used as probes.
Nuclear extracts were prepared from normal actively dividing and
density-arrested CEF and were incubated with double-stranded radiolabeled oligonucleotides. A quiescence-specific complex, designated C1 in Fig. 6A (lanes 1 and 2),
was observed with region A. Complexes migrating approximately at the
same position were also apparent when regions B and C were incubated
with the quiescent cell nuclear extract (lanes 3 to 6). A series of
fast-migrating complexes, referred to collectively as C2, were also
obtained with all three regions of the QRU. However, complexes C2 were not observed with region A containing a mutation in the palindrome, indicating that their formation required sequences adjacent to the
C/EBP binding site (Fig. 6B, lane 8). Since complexes C2 were obtained
with all three regions of the QRU, it is probable that their formation
depends on the putative Ets binding sites (TCCT or TTCC) disrupted in
the µpalA mutant. Complex C1 comigrated with a complex obtained with
an unrelated C/EBP binding site, namely, the NF-M oligonucleotide,
while complexes C2 were not formed with this probe (lane 9). Complex C1
was also observed when a nuclear extract was prepared from
serum-starved but not actively dividing SR-A RSV-transformed CEF (lanes
10 to 12).
View larger version (59K):
[in this window]
[in a new window]
|
FIG. 6.
Analysis of nuclear proteins binding to the QRU by EMSA.
(A) Analysis of nucleoprotein complexes obtained with regions A, B, and
C incubated with a nuclear extract from actively proliferating CEF (P)
or quiescent, density-arrested CEF (Q). The positions of complexes C1
and C2 are indicated. (B) Patterns of nucleoprotein complexes obtained
with region A (lane 7) or a derivative of region A containing mutations
disrupting a palindromic sequence identified in this region (lane 8;
µpalA). A radiolabeled, double-stranded oligonucleotide corresponding
to the C/EBP element designated NF-M was used as a probe in lane 9. (C)
Nucleoprotein complexes obtained when region A is incubated with a
nuclear extract from SR-A RSV-transformed CEF cultured in the presence
(5%) and absence (0%) of serum (lanes 10 and 11, respectively). The
pattern obtained with a nuclear extract prepared from normal,
density-arrested CEF (P) is shown in lane 12. (D) Supershift of complex
C1 with (+) and without () a C/EBP antibody (ab; arrowhead in
lanes 15 and 16). As a control, the preimmune serum was used in lanes
17 and 18.
|
|
The specificity of the quiescence-specific complex C1 was determined by
competition EMSA (Fig. 7). A 50-fold
excess of the region A oligonucleotide competed for the formation of
complex C1 (lane 12 in Fig. 7). In contrast, similar oligonucleotides differing by two nucleotides critical for C/EBP binding did not compete
complex C1 even when present at a 50-fold molar excess (oligonucleotide
µA [lanes 10 and 11]). A 50-fold excess of region B and NF-M
oligonucleotides also could compete for formation of complex C1 (lanes
15 and 19, respectively). However, region C was a poor competitor of
complex C1 even at a 50-fold molar excess (lane 17). Similar
competition assays were performed with oligonucleotides corresponding
to the QRU or derivatives containing a mutation in one or both C/EBP
binding sites. Complex C1 was competed by a 50-fold excess of the
wild-type (wt) QRU and the QRU containing a mutation in either one of
the two C/EBP binding sites (lanes 3, 5, and 7). In contrast, a QRU
fragment containing a mutation in both C/EBP binding sites did not
eliminate complex C1 (lane 9). Taken together, these results indicate
that the formation of the quiescence-specific complex C1 is dependent
on nucleotides that are also critical for binding by C/EBP.
View larger version (90K):
[in this window]
[in a new window]
|
FIG. 7.
Complex C1 is competed by oligonucleotides containing a
C/EBP binding site. Region A was used as a probe in EMSA and incubated
with a nuclear extract from density-arrested CEF. The specificity of
the nucleoprotein complexes was assessed by preincubating the extract
with a 10- or 50-fold molar excess of unlabeled oligonucleotides
corresponding to arbitrarily defined regions of the QRU (A, B, and C),
a derivative of region A containing a mutation in the C/EBP binding
site (µA), or the C/EBP element designated NF-M. In lanes 2 to 9, competition is performed with the wt QRU or a derivative of the QRU
containing a mutation of the C/EBP element in region A (QRU-µA), a
mutation of the C/EBP element in region B (QRU-µB), or a mutation in
both C/EBP binding sites (QRU-µ2C/EBP).
|
|
To identify a relevant activator(s) of the QRU, we screened a
contact-inhibited CEF cDNA library with a DNA fragment corresponding to
the conserved DNA binding domain of avian C/EBP. More than 15 independent clones were isolated and then characterized by restriction
digest and sequencing analyses. The member most frequently isolated was
NF-M, i.e., the avian C/EBP (82). A minority of the
clones encoded C/EBP, which was the only other member of the family
isolated in this screen. A similar screen was performed with the DNA
binding region of murine CHOP-10 but no C/EBP-related clones, including
one encoding CHOP-10, were isolated, in agreement with the notion that
the expression of this factor is induced predominantly in conditions of
glucose depletion and stress (89).
To determine if C/EBP binds to the QRU in quiescent cells, a nuclear
extract prepared from contact-inhibited CEF was preincubated with a
C/EBP antibody or with the preimmune serum and analyzed by EMSA. As
shown in Fig. 6D, the majority of complex C1 was supershifted by the
C/EBP antibody but not by the preimmune serum, indicating that
C/EBP is a component of C1. The results of other assays not shown in
this report confirmed that the entire C1 complex can be supershifted or
eliminated by antibodies for C/EBP. A less abundant
C/EBP-containing complex was also supershifted by the antibody in
the sample corresponding to actively dividing CEF (arrowhead in Fig.
6D). Complex C1 was also supershifted or eliminated by C/EBP
antibodies when a nuclear extract of serum-starved RSV-transformed CEF
was incubated with region A (7). Therefore, C/EBP is a
component of the quiescence-specific complex C1 of normal
density-arrested and serum-starved RSV-transformed CEF.
p20K and the QRU are regulated by linoleic acid in CEF.
To
determine if C/EBP can regulate the expression of the resident p20K
gene and not only that of promoter constructs activated transiently, we
constructed a replication-competent retrovirus capable of infecting CEF
and overexpressing C/EBP ectopically. CEF infected with the parental
virus, RCASBP, were used as a control in this experiment. The
overexpression of C/EBP was verified on a Western blot (Fig.
8A), while its nuclear localization was confirmed by indirect immunofluorescence analysis (46).
Total RNA was purified from CEF infected with RCASBP-C/EBP or RCASBP or from uninfected, contact-inhibited CEF. Subconfluent cells infected
with a retrovirus were investigated in serum-containing medium, i.e.,
in conditions where p20K is normally not expressed. As shown in Fig.
8B, CEF infected with RCASBP-C/EBP but not RCASBP expressed elevated
levels of the p20K mRNA. This level was comparable to that found in
contact-inhibited CEF. Likewise, the synthesis of p20K was markedly
stimulated in CEF overexpressing C/EBP but not in CEF infected with
the parental virus, RCASBP (Fig. 8C). CEF infected with a virus
encoding p20K (RCASBP-20K) express high levels of the protein and thus
provided a positive control in this experiment.
View larger version (39K):
[in this window]
[in a new window]
|
FIG. 8.
Analysis of p20K expression in CEF infected with a
retrovirus encoding C/EBP (RCASBP-C/EBP). (A) Western blotting
analysis of C/EBP in CEF infected with RCASBP-C/EBP or the
parental vector RCASBP. (B) Northern blotting analysis of p20K mRNA in
CEF infected with RCASBP or RCASBP-C/EBP or in uninfected but
density-arrested CEF (quiescent CEF). RNA loading was controlled by
probing the blot with a GAPDH cDNA. (C) Synthesis of the p20K protein
examined by metabolic labeling with [35S]methionine
followed by immunoprecipitation, separation on a polyacrylamide gel,
and fluorography. As a positive control, p20K was immunoprecipitated
from CEF infected with a retrovirus expressing the p20K cDNA
(RCASBP-20K).
|
|
CEF infected with RCASBP-C/EBP rapidly developed vesicles
reminiscent of the lipid droplets found in adipocytes (37).
Staining of the cells with oil red O confirmed that these vesicles
contained lipids (Fig. 9B).
Lipid-containing vesicles were not observed in serum-starved or
density-arrested CEF (Fig. 9C or D, respectively). The same was true
for actively dividing or serum-starved RSV-transformed CEF
(7). However, since p20K is a lipid binding protein of the
lipocalin family (9, 30), it is possible that its expression reflects the differentiation of CEF into adipocytes and not reversible growth arrest. To resolve this issue, we treated CEF with increasing concentrations of different lipids and examined the synthesis of p20K
in these conditions. The addition of a high concentration (200 µM) of
the polyunsaturated fatty acid linoleic acid induced the formation of
lipid droplets as efficiently as overexpressed C/EBP (Fig. 9H).
Smaller vesicles appeared with concentrations of linoleic acid as low
as 25 µM in cells treated for 16 h. In contrast, the addition of
palmitic or oleic acid induced the formation of only very small
vesicles even at a concentration of 200 µM (Fig. 9F or G,
respectively). CEF treated with lipids were metabolically labeled with
[35S]methionine, and the synthesis of p20K in actively
dividing or density-arrested CEF was analyzed by immunoprecipitation.
No synthesis of p20K was detected when lipid-treated CEF were actively
dividing (Fig. 10A, lanes 1 to 4).
Moreover, the synthesis of p20K was markedly reduced in
density-arrested cells treated with linoleic but not palmitic acid
(lanes 8 and 9). Linoleic acid did not induce DNA synthesis in
growth-arrested cells, indicating that the repression of p20K synthesis
was not the result of mitogenic stimulation (Table
2). Likewise, the addition of linoleic
acid, in the experimental conditions described for Fig. 10, did not
induce apoptosis (46). Therefore, we conclude that linoleic
acid, a known ligand of p20K (17), is a negative regulator
of p20K expression. These results imply that the expression of p20K is
not the result of growth-arrested CEF being committed to adipogenesis.
View larger version (137K):
[in this window]
[in a new window]
|
FIG. 9.
The overexpression of C/EBP or the addition of
linoleic acid induces the formation of lipid droplets in CEF. CEF
infected with RCASBP (A) or RCASBP-C/EBP (B) were stained for
lipid-containing vesicles with the lipophilic dye oil red O. (C) Oil
red O staining was also performed on CEF transferred to serum-free
medium and starved for 2 days. The formation of lipid-containing
vesicles was also investigated in density-arrested CEF (D); in this
sample, confluent CEF were treated with fresh serum-containing medium
and stained with oil red O the day after. Staining was also performed
on uninfected CEF treated with 200 µM palmitic, oleic, or linoleic
acid for 16 h (F, G, or H, respectively) or treated with the
diluent for the same period (0.1% ethanol) (E).
|
|
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 10.
Regulation of p20K expression by linoleic acid and
PPAR2. (A) The synthesis of p20K was examined by metabolic labeling
with [35S]methionine and immunoprecipitation analysis in
CEF treated for 16 h with 200 µM palmitic acid or linoleic acid
in conditions of logarithmic proliferation (lanes 2 and 3) or contact
inhibition (lanes 8 and 9). As a control, the synthesis of p20K was
also examined in actively dividing CEF infected with RCASBP or
RCASBP-C/EBP (lanes 5 and 6). CEF treated with the diluent, 0.1%
ethanol, were also analyzed (lanes 1 and 7) and compared with untreated
CEF (lanes 4 and 10). (B) The activity of a QRU construct was examined
by transient expression assays in growth-arrested (quiescent [Q]) CEF
treated with a low concentration of linoleic acid (LA; 25 µM) for
16 h and/or cotransfected with 2 µg of an expression vector for
PPAR2. (C) The activation of the QRU was examined upon ectopic
expression of C/EBP or - in cells cotransfected or not with an
equal amount (2 µg) of an expression vector for PPAR2.
|
|
To investigate the mechanism of action of linoleic acid, we performed a
series of transient expression assays with the QRU construct
transfected in CEF treated with this lipid or the diluent alone (Fig.
10B). No stimulation of the QRU was observed at any of the added
concentrations of linoleic acid (7). A suboptimal concentration of linoleic acid (25 µM) reduced the activity of the
QRU in density-arrested CEF, suggesting that linoleic acid regulates
the expression of p20K at least in part at the transcriptional level.
Two transcription factors, C/EBP and PPAR2, promote adipogenesis in model cell systems of adipogenesis (53). Since the former is a potent activator of the QRU (Fig. 4B), we examined the effect of
the overexpression of PPAR2 in density-arrested CEF. As shown in
Fig. 10B, PPAR2 considerably reduced the activity of the QRU in
quiescent cells. This potent activity of PPAR2 did not require the
addition of linoleic acid to the culture medium. However, PPAR2 did
not significantly decrease the activation of the QRU when C/EBP or
- was overexpressed in transient expression assays (Fig. 10C). Taken
together, these results indicate that p20K and the QRU are regulated
negatively by factors promoting adipogenesis, namely, linoleic acid and
PPAR2.
The expression of C/EBP is induced by contact inhibition.
CEF express nearly undetectable levels of C/EBP (NF-M) compared to
cells of the myelomonocytic lineage (45). Therefore, we
examined the expression of C/EBP in actively dividing and density-arrested CEF. As determined by Western blotting analysis (Fig.
11B), a significant accumulation of the
44-kDa C/EBP protein was observed in contact-inhibited cells. No
other forms of C/EBP were detected in this analysis using different
antisera, including one recognizing the conserved C terminus of the
protein (corresponding to the dimerization and DNA binding domains
[32]). Thus, a low-molecular-weight protein
corresponding to the liver-enriched inhibitory protein form of C/EBP
was not detected (26). As expected, p20K was also more
abundant in conditions of growth arrest whereas the level of ERK-1 and
-2 was not regulated by the proliferative state of the cell. The
transcription of the C/EBP gene was also more active in
contact-inhibited CEF, indicating that the accumulation of the
protein depends at least in part on transcriptional activation of the
gene (Fig. 11A). Following quantitation in an InstantImager, we
determined that contact inhibition caused a sixfold activation of
transcription of the C/EBP gene. Hence, C/EBP was itself regulated as a GAS gene in CEF. The accumulation of C/EBP is thus
one of the mechanisms controlling the induction of p20K in quiescent
CEF.
View larger version (27K):
[in this window]
[in a new window]
|
FIG. 11.
Activation of C/EBP expression by contact
inhibition. (A) Transcription of the p20K, C/EBP, and GAPDH genes
was examined by run-on transcription assays performed with nuclei
isolated from contact-inhibited (Q [quiescent]) or actively
proliferating (P) CEF. (B) The abundance of p20K and C/EBP was
determined by Western blotting analysis of total cell protein extract
of density-arrested (Q) or actively proliferating (P) CEF. Protein
loading was monitored by incubating the blot with antibodies for ERK-1
and -2.
|
|
| |
DISCUSSION |
Activation of the p20K gene is dependent on C/EBP.
Gene
activation has been intensively investigated in cells submitted to
ionizing radiations. In these conditions, the p53 tumor suppressor is
essential for the induction of p21waf1,
gadd45, and gadd153 (28, 43, 95).
However, in response to other stresses and conditions causing growth
arrest, these genes remain at least partly inducible in the absence of
functional p53, suggesting that they are regulated by multiple
transcription factors (43, 95). Little is known about the
trans-acting proteins and regulatory mechanisms of the GAS
genes in conditions of reversible growth arrest. Promoter fragments
activated by confluence or serum starvation have been described for the
gas1 and decorin genes, but a full characterization of
relevant transcription factors and their cognate elements has not been
reported (25, 58). A member of the Maf family has been
implicated in the activation of the retina-specific QR1 gene in the
chicken and appears to play an essential role for expression of this
gene in resting neuroretinal cells (70). Interestingly,
these authors outlined the sequence similarity of the QR1 and p20K QRUs
and provided evidence that the Maf-like factor responsible for the
activation of QR1 is competed by the p20K QRU in EMSA (70).
However, despite these observations, we failed to show a role for Maf
in activation of the p20K promoter.
We now report that C/EBP is responsible for the activation of p20K in
quiescent CEF. The overexpression of a dominant negative mutant of
C/EBP reduced markedly the activity of the QRU and p20K promoter in
growth-arrested CEF (Fig. 5B). The mutation of two C/EBP binding sites
of the QRU abolished its activation by quiescence (Fig. 5A). It is not
absolutely clear why the C/EBP element of region A was the only one
capable of conferring quiescence responsiveness to a heterologous,
minimal promoter (Fig. 3). It is possible that the two C/EBP binding
sites are qualitatively different. By EMSA, we identified a
quiescence-specific complex designated C1 when region A was used as a
probe (Fig. 6). A similar but more diffuse complex was also observed
with region B. While these complexes were both supershifted or
eliminated by a C/EBP antibody (Fig. 6D and reference
7), they may represent a different C/EBP dimer.
It is possible that essential nucleotides located outside the known
consensus binding site of C/EBP were missing in the arbitrarily defined
regions B and C, rendering these elements nonfunctional in transient
expression assays. We also investigated the role of nucleotides
diverging between the C/EBP binding sites of the QRU or found in other
promoters such as that of the cMGF cytokine gene (82). We
observed that differences at the 5' end did not significantly affect
the activity of these elements in transient expression assays. In
contrast, the presence of a C residue at the extreme 3' end of the
element (thus diverging from the T or G residue of the established
C/EBP consensus binding site) generates an element more responsive to
C/EBP but not C/EBP (7). Thus, while region A harbors
a classical C/EBP binding site on the minus strand, it also contains
the more C/EBP-responsive sequence on the plus strand. We showed
that the expression of C/EBP is induced at contact inhibition (Fig.
11), but even in these conditions, this factor is rare in CEF. Thus,
given the limiting amounts of C/EBP, differences in the sequence of
C/EBP binding sites are likely to be translated into differences in the
activity of the elements. That being said, the mutation analysis of the
QRU indicated that both C/EBP elements are required for activation by
quiescence, implying that factors binding to these elements cooperate
in the activation of the QRU. In this respect, the C/EBP binding site
of region B behaved as a typical type "B enhanson," as defined
originally by Fromental et al. (31a).
The results of EMSAs indicated that C/EBP is a component of the
quiescence-specific complex C1. However, it is not known if complex C1
corresponds to the homodimer of C/EBP or includes a second member of
the C/EBP family. A potential partner of C/EBP would be the
structurally related CHOP-10 protein, itself the product of a GAS gene
known as gadd153 (31). The C/EBP binding site of
region A fits approximately the extended consensus sequence defined for
the CHOP-10/C/EBP heterodimer (PuPuPuTGCAAT[A/C]CCC [87]). However, the overexpression of CHOP-10 alone or
in combination with C/EBP or C/EBP did not alter the activity of
the QRU in transient expression assays. Moreover, our cDNA library
screen for C/EBP-related factors did not yield any clone for CHOP-10 even when a DNA fragment corresponding to the conserved DNA binding domain of the murine CHOP-10 protein was used as a probe. Since CHOP-10
is thought to be induced principally by stress to the endoplasmic
reticulum, it may not be expressed in our conditions of quiescence
(89). Likewise, a clone for C/EBP
, a factor induced by
serum starvation in mammary epithelial cells (67), was not isolated in our screen. Using a probe for avian C/EBP and Northern blotting analysis, we were also unable to detect any transcript for
this C/EBP family member expressed in quiescent CEF (46). Since contact inhibition induced the expression of C/EBP and this
protein is a component of the quiescence-specific C1 complex (Fig. 6
and 11), our results point to this member of the C/EBP family as the
key activator of p20K expression in quiescent CEF.
Multiple pathways of p20K regulation.
During development,
C/EBP is involved in the commitment of the hematopoietic cell
lineage and is a critical regulator of macrophage and helper T-cell
functions (20, 62, 78, 83). It is also essential for the
acute-phase response of the liver, and it participates with C/EBP in
the control of energy metabolism and lipid homeostasis (6, 53,
88). It is thus perhaps not surprising that p20K, a lipid binding
protein, is regulated by C/EBP in growth-arrested CEF and by
linoleic acid, an EFA. Thus, our results support and expand the role of
C/EBP as a switch factor for the expression of
differentiation-specific and now growth arrest-regulated genes.
Using a retroviral construct (RCASBP-20K), we observed that the
formation of lipid droplets by CEF treated with increasing amounts of
linoleic acid is enhanced by the ectopic overexpression of p20K
(7). Cancedda and collaborators showed that p20K has high
affinity for polyunsaturated fatty acids such as linoleic acid and
therefore is thought to act as a transporter for this class of
molecules (17). Our preliminary results thus suggest that
p20K promotes the uptake of linoleic acid and therefore functions as a
modulator of EFA metabolism in CEF. An intriguing possibility is that
EFAs become limiting in serum-starved or density-arrested CEF and that
the induction of p20K represents an adaptive response to deleterious
growth conditions. Interestingly, Iyer et al. reported recently that
the genes for several enzymes involved in cholesterol biosynthesis are
most dramatically regulated by serum starvation in human fibroblasts
(41). These investigators speculated that the repression of
these genes by serum stimulation was the consequence of the addition of
lipoproteins provided by the serum. In agreement with this hypothesis,
we observed that the addition of linoleic acid to quiescent CEF
inhibited the expression of p20K (Fig. 10), suggesting the existence of
a negative feedback mechanism involving linoleic acid. Several
predictions can be made from this model and are presently being tested
in our laboratory.
Lu et al. showed previously that several nonsteroidal anti-inflammatory
drugs induce apoptosis in RSV-transformed CEF and markedly inhibit the
expression of p20K (52). Since fatty acid methyl esters and
several nonsteroidal anti-inflammatory drugs activate members of the
PPAR family (50, 76), we sought to determine if PPAR2, a
nuclear receptor induced by the expression of C/EBP and a key
regulator of adipogenesis in preadipocyte cell systems (53, 85,
93), controls the activity of the QRU. As shown in Fig. 10B,
ectopically expressed PPAR2 was indeed a potent inhibitor of a QRU
construct in density-arrested CEF. Thus, it is possible that a member
of the PPAR family is part of a regulatory network of proteins
controlling the expression of p20K. The QRU does not harbor any
consensus binding site for the nuclear receptor family. Moreover, we
found that PPAR2 did not diminish the activation of the QRU by
ectopically expressed C/EBP, an observation which may explain why
CEF infected with RCASBP-C/EBP express elevated levels of p20K and
form lipid vesicles. Whether PPAR2 can regulate the activity or
expression of C/EBP under physiological conditions remains to be
investigated. Finally, it is not known if the reciprocal is true, i.e.,
that the ectopic expression of C/EBP stimulates the expression of a
member of the PPAR family and PPAR2 in particular in CEF, as
reported for other fibroblasts (53). These questions are
important for our understanding of p20K regulation in growth-arrested
cells and are presently under investigation.
The regulation of p20K by EFAs may also explain the high level of p20K
expression in quiescent CHM cells since these cells were made quiescent
by culture in lipid-poor plasma (8). However, the study of
CHM cells also suggests that p20K is not regulated solely by the
availability of lipids in the medium. Indeed, the addition of a wide
variety of growth factors and mitogens was sufficient to induce
mitogenesis of CHM cells in plasma and invariably cause the repression
of p20K synthesis. In these conditions, CHM cells underwent several
rounds of cell division and were clearly not starved for lipids or
other nutrients. Second, we showed that CHM cells transformed by a
temperature-sensitive mutant of RSV stop dividing in plasma when
transferred to the nonpermissive temperature (9). Following
temperature shift, the p20K mRNA accumulated in quiescent CHM cells
within a few hours but not in their transformed counterparts, which
continued to proliferate at the permissive temperature. Third, the
synthesis of p20K was repressed in serum-starved CEF by the addition of
fresh serum-free medium (56); fourth, our conditions
promoting contact inhibition in CEF include the addition of
serum-containing medium, which would ensure a fresh supply of nutrients
to the cell (Materials and Methods). Thus, p20K is likely to be
regulated by the products of immediate-early genes and/or components of
the cell cycle machinery as cells exit G0. c-Myc is a good
candidate for this role since it inhibited the expression of the QRU in
transient expression assays (54). A role for c-Myc in the
inhibition of C/EBP-dependent gene expression has been described by
Mink et al. (61). In contrast, Rb and the structurally
related p107 protein interact with and antagonize the action of c-Myc
and promote gene activation by C/EBP (10, 19, 40).
Therefore, the activation of the p20K promoter may depend on multiple
mechanisms and trans-acting factors controlling the
expression or activity of C/EBP. In this respect, the relationship
between progression through the cell cycle, lipid metabolism, and
expression of p20K remains to be defined.
Regulation of C/EBP in quiescent CEF.
Unlike cells of the
myelomonocytic lineage, CEF express very low levels of C/EBP.
However, results presented in Fig. 11 indicated that transcription of
the C/EBP gene is stimulated by contact inhibition, a process
providing a mechanism of induction of the p20K gene. Hence, C/EBP is
regulated as a GAS gene in CEF. Therefore, it is of interest to
characterize the transcriptional activation of C/EBP in
growth-arrested CEF and to identify the trans-acting proteins responsible for this process.
The activity of C/EBP is known to be controlled by conserved
negative regulatory domains adjacent to and interacting with the
trans-activating region of the protein. According to models proposed by two independent groups, derepression of the concealed transactivating region involves the phosphorylation of critical residues of the regulatory domains or association with unrelated trans-acting factors resulting in the unfolding of the
C/EBP protein (48, 92). Experimental evidence supports
the existence of these regulatory processes as C/EBP is
phosphorylated and activated by several kinases (63, 86, 91)
and cooperates with an increasing number of transcription factors,
including Myb AP-1, NF-
B, Pu.1, serum response factor, the
glucocorticoid receptor, and SP1, in gene activation (11, 15, 47,
49, 57, 64, 65, 74, 79, 81). It is therefore probable that the
activity of C/EBP is also regulated at growth arrest in CEF. The
characterization of these regulatory mechanisms is likely to be
important for our understanding of GAS gene expression and
G0.
This work was made possible by grants from the Natural Sciences
and Engineering Research Council and Medical Research Council of Canada
to P.-A.B.
We thank A. Leutz, C. F. Calkhoven, K.-H. Klempnauer, D. Ron, J. Hassell, J. Capone, S. Hughes, and M. Nishizawa for providing reagents
used in these studies. We also thank Yves Villeneuve and Gordon Temple
for preparation of the figures. The Signal Scan program was kindly
provided by D. S. Prestidge.
| 1.
|
Akira, S., and T. Kishimoto.
1992.
IL-6 and NF-IL6 in acute-phase response and viral infection.
Immunol. Rev.
127:25-50[Medline].
|
| 2.
|
Alani, R.,
P. Brown,
B. Binetruy,
H. Dosaka,
R. Rosenberg,
P. Angel,
M. Karin, and M. J. Birrer.
1991.
The transactivating domain of the c-Jun proto-oncoprotein is required for cotransformation of rat embryo cells.
Mol. Cell. Biol.
11:6286-6295[Abstract/Free Full Text].
|
| 3.
|
Almendral, J. M.,
D. Sommer,
H. Macdonald-Bravo,
J. Burckhardt,
J. Perera, and R. Bravo.
1988.
Complexity of the early genetic response to growth factors in mouse fibroblasts.
Mol. Cell. Biol.
8:2140-2148[Abstract/Free Full Text].
|
| 4.
|
Barone, M. V.,
A. Crozat,
A. Tabaee,
L. Philipson, and D. Ron.
1994.
CHOP (Gadd153) and its oncogenic variant, TLS-CHOP, have opposing effects on the induction of G1/S arrest.
Genes Dev.
8:453-464[Abstract/Free Full Text].
|
| 5.
|
Baserga, R.
1976.
Multiplication and division in mammalian cells, vol. 6. The biochemistry of disease. , p. 53-77.
Marcel Dekker Inc., New York, N.Y.
|
| 6.
|
Baumann, H.,
K. K. Morella,
S. P. Campos,
Z. Cao, and G. P. Jahreis.
1992.
Role of CAAT-enhancer binding protein isoforms in the cytokine regulation of acute-phase plasma protein genes.
J. Biol. Chem.
267:19744-19751[Abstract/Free Full Text].
|
| 7.
| Bedard, P.-A. Unpublished results.
|
| 8.
|
Bedard, P.-A.,
S. D. Balk,
H. S. Gunther,
A. Morisi, and R. L. Erikson.
1987.
Repression of quiescence-specific polypeptides in chicken heart mesenchymal cells transformed by Rous sarcoma virus.
Mol. Cell. Biol.
7:1450-1459[Abstract/Free Full Text].
|
| 9.
|
Bedard, P.-A.,
Y. Yannoni,
D. S. Simmons, and R. L. Erikson.
1989.
Rapid repression of quiescence-specific gene expression by epidermal growth factor, insulin, and pp60v-src.
Mol. Cell. Biol.
9:1371-1375[Abstract/Free Full Text].
|
| 10.
|
Beijersbergen, R.,
E. M. Hijmans,
L. Zhu, and R. Bernards.
1994.
Interaction of c-Myc with the pRb-related protein p107 results in inhibition of c-Myc-mediated transactivation.
EMBO J.
13:4080-4086[Medline].
|
| 11.
|
Betts, J. C.,
J. K. Cheshire,
S. Akira,
T. Kishimoto, and P. Woo.
1993.
The role of NF-B and NF-IL6 transactivating factors in the synergistic activation of human serum amyloid A gene expression.
J. Biol. Chem.
268:25624-25631[Abstract/Free Full Text].
|
| 12.
|
Bhat, N. K.,
C. B. Thompson,
T. Lindsten,
C. H. June,
S. Fujiwara,
S. Kooizumi,
R. J. Fischer, and T. S. Papas.
1990.
Reciprocal expression of human ets1 and ets2 genes during T-cell activation: regulatory role for the proto-oncogene ets1.
Proc. Natl. Acad. Sci. USA
87:3723-3727[Abstract/Free Full Text].
|
| 13.
|
Bohmer, F.-D.,
R. Kraft,
A. Otto,
C. Wernstedt,
U. Hellman,
A. Kurtz,
T. Muller,
K. Rohde,
G. Etzold,
W. Lehman,
P. Langen,
C.-H. Heldin, and R. Grosse.
1987.
Identification of a polypeptide growth inhibitor from bovine mammary gland.
J. Biol. Chem.
262:15137-15143[Abstract/Free Full Text].
|
| 14.
|
Briggs, M. R.,
J. T. Kadonaga,
S. P. Bell, and R. Tjian.
1986.
Purification and biochemical characterization of the promoter-specific transcription factor Sp1.
Science
234:47-52[Abstract/Free Full Text].
|
| 15.
|
Burk, O.,
S. Mink,
M. Ringwald, and K. H. Klempnauer.
1993.
Synergistic activation of the chicken mim-1 gene by v-myb and C/EBP transcription factors.
EMBO J.
12:2027-2038[Medline].
|
| 16.
|
Cabannes, E.,
M.-F. Vives, and P.-A. Bedard.
1997.
Transcriptional and post-transcriptional regulation of B-controlled genes by pp60v-src.
Oncogene
15:29-43[Medline].
|
| 17.
|
Cancedda, F. D.,
M. Malpeli,
C. Gentili,
V. Di Marzo,
P. Bet,
M. Carlevaro,
S. Cermelli, and R. Cancedda.
1996.
The developmentally regulated avian ch21 lipocalin is an extracellular |
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
(NF-M) Is Essential for Activation of the
p20K Lipocalin Gene in Growth-Arrested Chicken Embryo
Fibroblasts
Top
Abstract
Introduction
