(Received for publication, July 31, 1996, and in revised form, October 8, 1996)
From the Department of Molecular Genetics, Kumamoto University School of Medicine, Kuhonji, Kumamoto 862, Japan
The gene for liver-type arginase, an ornithine
cycle enzyme, is induced by glucocorticoids in a delayed secondary
manner. An enhancer element located around intron 7 of the rat arginase gene shows delayed glucocorticoid responsiveness, and it harbors two
sites binding with members of the CCAAT/enhancer binding protein (C/EBP) family. Here, we investigate the role of these C/EBP binding sites in glucocorticoid response of the arginase gene. When inserted in
front of the herpes simplex virus thymidine kinase promoter, these
C/EBP sites exhibited glucocorticoid responsiveness in reporter transfection assay using rat hepatoma H4IIE cells. In footprint analysis using nuclear extracts of H4IIE cells, profiles of the protected areas of the two C/EBP sites changed when cells were treated
with dexamethasone. In gel shift analysis, the complex formation for
the two C/EBP sites was augmented in response to dexamethasone.
Antibody supershift/inhibition analysis demonstrated that a major
portion of the binding proteins induced by dexamethasone is C/EBP.
Induction of arginase mRNA by dexamethasone was preceded by
augmentation of the C/EBP site-binding activities, which followed increase in C/EBP
mRNA. These results were consistent with the notion that the glucocorticoid response of the arginase gene is mediated by C/EBP
.
Liver-type arginase (EC 3.5.3.1) is an enzyme catalyzing the last step of the ornithine cycle (urea cycle), through which toxic ammonia is converted into less toxic urea. The mammalian arginase gene is expressed almost exclusively in the liver (1, 2) and is markedly induced in the late fetal period (3). After birth, arginase gene expression is also activated by high protein intake or starvation that imposes ammonia production (4, 5). These developmental and nutritional activations of the arginase gene are presumably mediated, at least in part, by glucocorticoids and/or glucagon. In fact, these hormones induce arginase mRNA in rat primary cultured hepatocytes (6) and a rat hepatoma cell line H4IIE (7).
Glucocorticoids activate transcription of many liver-specific genes.
The glucocorticoid response can be divided into two types. One is the
primary response, in which the glucocorticoid-receptor complex directly
activates transcription of target genes and which does not require
ongoing protein synthesis. The other is the secondary response that
follows a delayed time course, compared with the primary response, and
is blocked by protein synthesis inhibitors. Genes showing the secondary
response are exemplified by those of hepatic 2u-globulin
(8), its isoform (9, 10), and ornithine cycle enzymes including
arginase (6). A newly synthesized protein factor(s) involved in this
secondary activation process has remained to be identified.
We have studied transcriptional regulation of the arginase gene by
investigating the promoter (11-13) and the enhancer (12) of the rat
gene (14). The arginase promoter exhibits a moderate liver-selective
activity (11) but apparently no glucocorticoid responsiveness (12). On
the other hand, the enhancer region, which is located 11 kilobases
downstream from the transcription start site and spans the junction of
intron 7 and exon 8, showed delayed glucocorticoid responsiveness in
transfection analysis using rat hepatoma H4IIE cells (12). In this
enhancer region, there are four protein binding sites, two of which are
recognized by a factor(s) related to CCAAT/enhancer binding protein
(C/EBP)1 (15), and the other two bind with
unknown factors. C/EBP is characterized by the DNA-binding basic region
and the adjacent leucine zipper domain (16) and constitutes a family
with related factors (15, 17). C/EBP is one of these family members
(16) and is also known as NF-IL6 (18), IL6-DBP (19), LAP (20), AGP/EBP
(21), and CRP2 (17). While C/EBP
is enriched in the liver and plays
a role in liver-selective transcription of several genes (20, 22, 23),
this factor seems to be involved in a number of cellular processes
in various tissues (reviewed in Introductions of Refs. 24-27). In
hepatoma cells (28, 29) and primary-cultured rat hepatocytes (30),
C/EBP
mRNA is induced by glucocorticoids. This induction in
hepatocytes is caused by transcriptional activation of the C/EBP
gene in a primary manner (30).
Here, we asked whether the C/EBP sites in the arginase enhancer mediate
the glucocorticoid response. We also examined changes in profiles of
protein binding to these sites in response to glucocorticoids. Our
findings suggest that C/EBP is involved in the secondary glucocorticoid response of the arginase gene.
The basal chloramphenicol acetyltransferase (CAT)
plasmid pBLCAT5-ABA was constructed starting from the plasmid pBLCAT5
(31) that harbors the CAT gene under the control of herpes simplex virus thymidine kinase gene (tk) promoter. pBLCAT5 was cut
with BglII, blunt-ended with Klenow fragment, and
self-ligated. The resulting plasmid was cut with SacI, and
the linker 5-CGGGCCCAGATCTGGGCCCGAGCT-3
bearing the
ApaI-BglII-ApaI sites was inserted
into the SacI site, yielding the plasmid pBLCAT5-ABA.
Oligonucleotides each containing C/EBP binding sites A and B
(previously designated protein-binding sites I and III (12),
respectively) of the arginase enhancer were inserted into the
BamHI site located just upstream of the tk
promoter of pBLCAT5-ABA. The direction and sequence integrity of the
inserts were verified by nucleotide sequencing.
Rat hepatoma H4IIE cells were grown in Eagle's
minimal essential medium supplemented with 10% fetal calf serum.
Transfection was carried out by the calcium phosphate precipitation
method (32), with a total of 15 µg of DNA mixture containing 10 µg of the CAT gene recombinant plasmid and 5 µg of an internal standard plasmid pAc-lacZ (33) bearing the -galactosidase gene. Cells were
cultured with or without 1 µM dexamethasone for 48 h
prior to harvesting. During this period, 1 mM dibutyryl
cAMP was also added into the medium to increase the basal CAT activity
(12). 72 h after transfection, CAT activity of cell extracts was
measured as described (34), quantified using a bio-image analyzer
BAS2000 (Fuji Photo Film, Tokyo), and normalized for
-galactosidase
activity. Relative CAT activities were shown by mean values plus
standard errors of at least three independent experiments.
Nuclear extracts from rat liver (35) and H4IIE cells (36) were prepared as described. The probe used was the 32P-labeled DNA fragment corresponding to the lower strand of the XbaI-HincII enhancer segment (12). The binding reaction, DNase I digestion, and electrophoresis were done as described previously (12).
Gel Shift AssaysExpression and purification of the
recombinant C/EBP protein fused to maltose-binding protein (MBP) was
as described (13). Double-stranded oligonucleotide probes were 5
-end
labeled with [
-32P]ATP and T4 polynucleotide kinase.
The binding reaction for 30 min on ice was carried out as described
previously (12). Electrophoresis was done in a 5% polyacrylamide gel
made in 22 mM Tris, 22 mM boric acid, and 0.6 mM EDTA. The antibodies used in supershift/inhibition analysis (1 µg of IgG in 1 µl) against C/EBP
and C/EBP
were purchased from Santa Cruz Biotechnology, Inc.
Isolation of total RNA (37),
electrophoresis, blotting, and hybridization (30) were performed as
described. The following 32P-labeled DNA probes were used:
the mouse C/EBP gene, about a 1.6-kilobase BstXI fragment
of pEF-C/EBP
(22); rat arginase cDNA, about a 850-base pair
EcoRI-EcoRV fragment of pARGr-2 (38); and rat
actin, a polymerase chain reaction product, nucleotide positions
417-1223 (39). Relative mRNA levels quantified with a bio-image
analyzer were shown by mean values with standard errors of at least
three independent experiments.
The arginase enhancer, which shows
glucocorticoid responsiveness in a delayed secondary manner, contains
four protein binding sites (12). Two sites (sites A and B) bind C/EBP
family members (Fig. 1). To examine the contribution of
the two C/EBP binding sites to glucocorticoid responsiveness, we
inserted each C/EBP binding site just in front of the herpes simplex
virus tk promoter that drives the Escherichia
coli CAT gene, and transient transfection assay was done using rat
hepatoma H4IIE cells. The cells were treated with 1 µM
dexamethasone for 48 h, and cell extracts were subjected to CAT
enzyme assay (Fig. 1). CAT activity derived from the native
tk promoter was increased slightly (1.8-fold) by the dexamethasone treatment. When site A was inserted in the same direction
of the tk promoter, CAT activity was enhanced 13.6-fold in
the absence of dexamethasone and was not further augmented by
dexamethasone. On the other hand, when inserted in the opposite direction, site A enhanced basal CAT activity 4.4-fold, and this enhanced activity was further stimulated 7.8-fold by dexamethasone. Therefore, site A can respond to dexamethasone if appropriately arranged with the tk promoter. The reason why site A does
not respond to dexamethasone in the forward direction is not known. A
possible explanation is that in this promoter construct an unidentified transcriptional activator(s) occupies site A and is insensitive to
dexamethasone. Another C/EBP binding site B had little effect on basal
CAT activity when a single copy of this site was inserted in forward or
reverse direction. On the other hand, CAT activities derived from these
constructs increased 5.3- and 3.9-fold, respectively, in response to
dexamethasone. Two copies of site B inserted in the forward direction
led to a 5.3-fold increase in the basal tk promoter
activity, which was further augmented 4.2-fold by dexamethasone.
Therefore, site B exhibits glucocorticoid responsiveness in the context
of the tk promoter. Thus, both C/EBP binding sites A and B
have potency to mediate the glucocorticoid response.
Changes in Footprint Profiles of the C/EBP Binding Sites by Dexamethasone
To examine whether dexamethasone affects
protein-binding profiles of the C/EBP sites, DNase I footprint analysis
was carried out (Fig. 2). Using nuclear extracts from
H4IIE cells, protection against DNase I digestion was observed at C/EBP
sites A and B corresponding to footprint areas previously detected with
rat liver nuclear extracts (12). Treatment of H4IIE cells with 1 µM dexamethasone for 12 h changed the footprint
profiles of both sites A and B. In both sites, the footprint profiles
obtained with extracts from dexamethasone-treated cells resembled those seen with liver extracts. Therefore, binding activity (or activities) to C/EBP sites A and B similar to that in the liver seems to be induced
in H4IIE cells by dexamethasone.
Induction of C/EBP
We further characterized dexamethasone-induced
changes in a factor(s) binding to sites A and B with gel mobility shift
assay (Fig. 3A). We previously showed that
C/EBP and C/EBP
bind with these sites and that C/EBP
dominates
in the rat liver, whereas C/EBP
dominates in the
dexamethasone-treated H4IIE cells (12). When H4IIE cells were treated
with 1 µM dexamethasone for 12 h, binding of nuclear
extracts to sites A and B was dramatically enhanced (compare
lanes 5 to lanes 1). To examine whether C/EBP
is involved in dexamethasone-induced binding activities, antibody supershift/inhibition analysis was performed. The antibody specific to
C/EBP
diminished the complexes induced by dexamethasone (lanes 7), while the C/EBP
-specific antibody and the control serum
caused little change (lanes 6 and 8). Therefore,
C/EBP
accounts for a major portion of site A- and B-binding
activities induced by dexamethasone in H4IIE cells. Similar results
were obtained by using as a probe the authentic C/EBP site (40) of the
Rous sarcoma virus long terminal repeat (data not shown). To confirm
recognition of sites A and B by C/EBP
, we examined binding of the
recombinant C/EBP
protein fused to MBP (Fig. 3B). When
sites A and B, as well as the authentic Rous sarcoma virus C/EBP site,
were used as probes, the MBP-C/EBP
fusion protein but not MBP gave a
shifted band. On the other hand, no shifted band was detected when an unrelated probe (site IV (12) of the arginase enhancer) was used. Thus,
sites A and B specifically bind with the recombinant C/EBP
protein.
All these results taken together lead to the proposal that binding
activity for C/EBP sites A and B of the arginase enhancer is induced in
H4IIE cells by dexamethasone administration and that this binding can
be attributable to C/EBP
.
Serial Inductions of C/EBP
Based on the observations
described above and the primary glucocorticoid response of the C/EBP
gene (30) and the secondary response of the arginase gene (6), we
hypothesized that C/EBP
induced primarily by dexamethasone is
responsible for the secondary induction of the arginase gene. We then
examined whether time courses for increases in C/EBP
mRNA level,
C/EBP site-binding activity, and arginase mRNA level by
dexamethasone are concordant with this hypothesis. In Fig.
4A, RNA blot analysis was carried out after
exposure of H4IIE cells to dexamethasone for various periods. Arginase
mRNA began to increase with a lag of 6 h following the hormone
addition and increased markedly at 24 h. On the other hand,
C/EBP
mRNA began to increase as early as 0.5 h and reached a plateau at 2 h.
actin mRNA, measured as a standard, was
practically unchanged. Time courses of changes in binding activities
for C/EBP sites A and B were also monitored (Fig. 4B). For
both sites A and B, binding activities were increased markedly 6 h
after the addition of dexamethasone and were sustained by 24 h. No
obvious change was observed for NF-I site-binding activity monitored as a standard. These results were quantified in Fig. 4C.
Concordant with our hypothesis, the increase in arginase mRNA was
preceded by the increases in C/EBP site-binding activities, which were further preceded by the increase in C/EBP
mRNA.
In this report, we presented evidence supporting the notion that
C/EBP mediates the secondary glucocorticoid response of the arginase
gene. First, C/EBP sites of the arginase enhancer, when linked to the
heterologous tk promoter, exhibited glucocorticoid responsiveness in reporter transfection assay. Second, protein binding
profiles of these C/EBP sites were changed by dexamethasone, and
C/EBP
occupied the major portion of newly induced binding activities. Third, induction of arginase mRNA by dexamethasone followed an increase in C/EBP site-binding activity that was preceded by C/EBP
mRNA induction. We previously showed that C/EBP
mRNA is induced by dexamethasone through transcriptional activation without requiring de novo protein synthesis (30). On the
other hand, the induction of arginase mRNA follows a delayed time
course and requires ongoing protein synthesis (6, 30). Based on these
observations, we propose the following: glucocorticoids primarily
activate the C/EBP
gene and the induced C/EBP
secondarily activates the arginase gene by binding to the enhancer region.
In cotransfection assay using non-hepatic Chinese hamster ovary cells,
CAT activity derived from the reporter construct containing one copy of
site B in the forward direction was activated weakly (1.7-fold) by a
C/EBP expression plasmid.2 It remains to
be clarified whether an additional factor(s) is required for the
maximal action of dexamethasone.
In general, gene cascades controlled by steroid hormones underlie
various biological processes. A well documented example is effects of
ecdysone in metamorphosis of Drosophila. Ecdysone triggers
induction of a relatively small number of primary response genes, some
of which encode transcription factors. These induced factors in turn
activate more than 100 secondary response genes (reviewed in Ref. 41).
In mammals, several transcription factors, as well as C/EBP, were
shown to be primarily induced by glucocorticoids in various tissues:
C/EBP
in adipocytes (42), peroxisome proliferator-activated receptor
(PPAR)
in hepatocytes (43), and I
B in immune cells and Hela
cells (44, 45). On the other hand, especially in the liver,
glucocorticoid induction of a number of genes exhibits more or less
secondary aspects, i.e. delayed time course and/or sensitivity to protein synthesis inhibitors, as exemplified by genes
for
2u-globulin (8-10),
1-acid
glycoprotein (28, 46), albumin (47), tryptophan oxygenase (48),
phosphoenolpyruvate carboxykinase (49), and ornithine cycle enzymes
(6). In mammals, little is known about the transcription factors
responsible for secondary activation of target genes. Involvement of
C/EBP
in glucocorticoid response of the arginase gene provides a
typical example for hierarchical gene regulation by
glucocorticoids.
We thank F. Matsuno and other colleagues for suggestions and discussions; we also thank M. Ohara for helpful comments.