(Received for publication, July 16, 1996, and in revised form, October 18, 1996)
From the Departments of Internal Medicine and
Physiology & Biophysics, University of Iowa College of Medicine,
Iowa City, Iowa 52242 and § Molecular Biology & Hypertension Laboratory, Department of Physiology, University of
Sydney, Sydney, New South Wales 2006, Australia
We examined the DNA sequence and
transcription factor requirements for cAMP-induced transactivation of
the human renin promoter using Calu-6 cells that express human
renin mRNA endogenously. A series of constructs containing 896 base
pairs of human renin 5-flanking DNA fused to the luciferase gene and
containing either the native, a consensus, or a nonfunctional cAMP
response element (CRE) were used to assess DNA sequence requirements
mediating the cAMP response. Expression vectors encoding the CREB-1
transcription factor, a dominant negative mutant form of CREB-1, and
the catalytic subunit of protein kinase A (PKA) were used to assess
transcription factor requirements mediating the cAMP response.
Forskolin treatment alone only caused a 2-3-fold activation of the
HREN promoter in Calu-6 cells, but nearly a 10-fold
activation in JEG-3 cells, which do not express renin but are highly
responsive to cAMP. Gel shift assays revealed the binding of five
specific DNA-protein complexes consisting of the ATF-1 and CREB-1
transcription factors, one of which was an ATF-1·CREB-1 heterodimer
suggesting the potential for regulation of CREB-1 activity by ATF-1.
However, over-expression of CREB-1 did not significantly enhance
forskolin-induced human renin transcriptional activity. Transfection of
both Calu-6 and JEG-3 cells with a PKA expression vector resulted in a
10-fold induction of human renin transcriptional activity in constructs containing the native or consensus CRE and 5-fold activation in a
construct containing a nonfunctional CRE. We confirmed that the PKA
response has both a CREB-dependent and CREB-independent component by demonstrating that the PKA response was abolished by
co-transfection of a dominant negative mutant form of CREB-1 into cells
containing the native or consensus CRE construct but not in cells
containing the nonfunctional CRE construct. We therefore conclude that
the human renin promoter can be transcriptionally activated in a
renin expressing cell line through the cAMP-PKA pathway and is mediated
by both a CREB-dependent and CREB-independent mechanism.
Renin, the rate-limiting enzyme in the generation of angiotensin
II is synthesized and released from renal juxtaglomerular (JG)1 cells. In kidney, renin synthesis and
release is regulated by numerous physiological factors including sodium
chloride sensed by the macula densa, arterial pressure sensed by
baroreceptors present within the renal vasculature, sympathetic nerve
activity via -adrenergic receptor signaling mechanisms, and
angiotensin II acting on high affinity angiotensin II receptors present
on JG cells (reviewed in Ref. 1). Release of norepinephrine from sympathetic nerve terminals stimulates cell surface
-adrenergic receptors on JG cells and causes increased intracellular cyclic AMP
(cAMP) through the classical mechanism (1). cAMP has been reported to
be a major factor controlling renin synthesis and secretion (2, 3), and
numerous reports have documented that cAMP can activate the renin
promoter in both renin expressing (4, 5) and non-renin expressing (6,
7) cell lines.
Previous DNase I footprinting studies have revealed that sequences
located between 374 and +16 of the human renin (HREN) 5
-flanking region bind trans-acting factors present in
choriodecidual cell nuclear extracts (5). These include proteins that
bind to Ets, Pit-1, AGE-3, and ARP-1 like sequences. In addition, the sequence TAGCGTCA at position
225 to
218 of the HREN
5
-flanking DNA shares homology to the consensus cAMP-responsive
element (CRE) binding site previously characterized in the somatostatin
promoter (8). The CRE sequence has been reported to be required for the
cAMP stimulation of HREN promoter-reporter constructs in
primary cultures of choriodecidual cells (4, 5). In addition, a sequence present in the HREN 5
-flanking DNA which has
homology to the pituitary-specific factor Pit-1 and binds members of
the POU domain family of transcription factors has been reported to be
required for the cAMP induction mediated through the HREN
CRE (9, 10).
Although several studies have demonstrated that the CRE sequence is necessary for cAMP-mediated induction of the HREN promoter, the mechanistic details of this induction still remain largely unexplored. Most investigators have assumed that HREN transcriptional induction caused by cAMP is due to the action of the CRE binding protein-1 (CREB-1) factor. Although purified CREB-1 can specifically bind to the HREN CRE (10), other members of the CREB/activating transcription factor (ATF) family of transcription factors can also bind to CRE-like elements and activate transcription in response to cAMP (11). Moreover, CREB/ATF transcription factors, as members of the leucine-zipper family of transcription factors (bZIP), can not only activate transcription in their homodimeric form but also regulate transcriptional activation by forming heterodimers. For example, ATF-1 has been reported to antagonize the transcriptional effects of CREB-1 on the somatostatin promoter by the formation of an ATF-1·CREB-1 heterodimer (12). Additionally, the fact that mutations in the HREN CRE do not totally abolish HREN promoter activity in response to cAMP (10) suggests that alternative mechanisms may also be active in renin expressing cells. The importance and magnitude of the CRE-independent mechanisms have yet to be explored.
One problem encountered by numerous investigators examining the mechanism of HREN regulation has been the absence of suitable renin expressing cell lines. Indeed numerous attempts to establish permanent cell lines that express HREN endogenously have been largely unsuccessful (13, 14). Although the HREN gene is expressed in many cell types, its expression is particularly high in JG (1) and choriodecidual cells (15), but both cell types lose their ability to express renin mRNA and produce renin when placed in culture (14, 15). We have reported that Calu-6 cells that are derived from a human pulmonary carcinoma express HREN mRNA endogenously (16). Moreover, the steady-state level of endogenous HREN mRNA is markedly increased in response to increased intracellular cAMP as a result of both transcriptional and post-transcriptional mechanisms (17). The expression of endogenous HREN mRNA in Calu-6 cells, along with the findings that 1) renin is expressed in human fetal lung (18), 2) in human pulmonary tumor tissue (19, 20), and 3) in the lung of transgenic mice and transgenic rats containing genomic HREN transgenes expressed from their own promoters (21, 22), provided evidence to propose that lung is a bona fide site of renin expression in humans and suggested that Calu-6 cells would be a novel tool to examine HREN gene regulation. Herein we use Calu-6 cells to examine the mechanism controlling the transcriptional activation of the HREN promoter by the cAMP pathway.
Calu-6 cells were grown in Eagle's minimal essential media supplemented with 10% fetal bovine serum, sodium pyruvate, and non-essential amino acids as described previously (16). JEG-3 cells were grown in Dulbecco's modified essential media supplemented with 10% calf serum. Cells were grown to 90% confluency in T75 tissue culture flasks and were split into 60-mm tissue culture plates the day before transfection. The cells were transfected when they reached 80% confluency.
Plasmid ConstructsAll HREN promoter segments
were cloned into the pGL2-Basic vector (Promega, Madison, WI) after one
or more subcloning steps. The subclone pES3.0 (4) was digested with
XhoII which liberated a 2608-bp fragment (2595 to +13)
containing the HREN promoter. This fragment was ligated into
BglII-digested pGL2-Basic to generate the 2750Luc construct.
The plasmid 900Luc (900L) was constructed by ligating a 909-bp
HindIII fragment (
896 to +13) containing the promoter of
2750Luc into pGL2-Basic. The 900MUT and 900CRE mutants were generated
using the Chameleon site-directed mutagenesis assay system (Stratagene)
using the directions and reagents provided by the manufacturer. The
oligonucleotides used for the mutagenesis were 1) a 30-base
XmnI selection oligonucleotide
5
-GCTCATCATTGGAAAACGCTCTTCGGGGCG-3
, 2) a 34-base oligonucleotide
containing the consensus CRE sequence, 5
-CAGATAGAGGGCTGCTGACGTCACTGGACACAAG-3
, and 3) a 36-base
oligonucleotide containing the sequence that completely destroyed
protein binding to the CRE, 5
-GAGGGCTGCTGATACTGCTGGACACAAGATTGCTTT-3
.
A control plasmid, SSCRE, was generated by ligating a 240-bp
HindIII fragment containing the somatostatin promoter and
consensus CRE (a gift from by Dr. Richard Goodman, Vollum Institute,
OR) into pGL2-Basic. Two Rous sarcoma virus promoter expression vectors
which contain either the wild-type CREB (wtCREB) or a dominant negative
mutant of CREB (KCREB) were generously provided by Dr. Richard Goodman (Vollum Institute, OR) (23). KCREB contains a single amino acid change
from arginine to leucine in the DNA binding domain. We obtained the
plasmid pMtC encoding the catalytic subunit of mouse protein kinase A
(PKA) under the control of the mouse metallothionein 1 promoter from
Dr. Stephen Goodbourn (St. George's Hospital, University of London)
(24, 25). A cytomegalovirus promoter-
-galactosidase fusion construct
was used as an internal control to normalize for differences in
transfection efficiency among flasks. All plasmid DNAs were purified by
centrifugation on two cesium chloride density gradients prior to
transfection. Plasmid DNA concentration was measured by absorbance at
260 nm and was confirmed by gel electrophoresis using known
standards.
Calu-6 or JEG-3 cells,
plated at 2 × 106 at 80% confluence, were
transfected using calcium-phosphate precipitates containing 7.0 µg of
test reporter plasmid DNA, 2.5 µg of either the wtCREB or KCREB
expression vector, 2.5 µg of cPKA expression vector DNA, and 5.0 ng
of the pCMV--gal internal control. pUC19 vector DNA was used as
carrier DNA in order to maintain a constant amount of nucleic acid (12 µg) per transfection. Approximately 10-20% of the cells are
transfected using this protocol. Forskolin (10 µM) was
added 18-24 h prior to performing the luciferase activity assay.
Luciferase activity assays were performed using a commercially available kit (Promega, Madison, WI), following the directions recommended by manufacturer, and activity was measured in a Monolight 2010 automatic luminometer.
-Gal activity was measured using the
Galacto-light kit (Tropix, Bedford, MA) following the directions recommended by the manufacturer and as described previously by us (17).
All luciferase and
-gal activity assays were performed in duplicate
and the values averaged to obtain n = 1. Background luciferase activity was determined by performing the assay in untransfected cells. Importantly, there was no increase in
-gal activity in response to cAMP stimulation. The pGL-promoter (SV40 promoter) and pGL2-Basic luciferase vectors were included in each experiment to monitor the quality of transfection. Assay values were
always at least 5-10 times greater than the promoterless construct.
Relative luciferase activity was calculated as a percentage of the SV40
promoter after correction for transfection efficiency in order to
eliminate day-to-day variation in transfection as described previously
(17). The data represent the mean ± S.E. of six independent
experiments performed in Calu-6 cells and four independent experiments
performed in JEG-3 cells. The data were analyzed by one-way analysis of
variance (ANOVA) (26) followed by Student's modified t test
with Bonferroni correction for multiple comparisons between means (27)
using the modified error mean square term from the ANOVA or by unpaired
t test using the SigmaStat Software package (Jandel
Scientific).
Crude nuclear
extracts from either Calu-6 or JEG-3 cells were prepared as described
previously (28) with the following modifications. Cells from 80%
confluent monolayer cultures were harvested and washed in 1 × phosphate-buffered saline. Pelleted cells were resuspended in buffer
containing 10 mM HEPES, pH 7.9, 1.5 mM
MgCl2, 10 mM KCl, and 0.5 mM
dithiothreitol, maintained on ice to swell, and lysed by being rapidly
and repeatedly drawn through a 26-gauge needle. The crude nuclear
pellet obtained by brief centrifugation was then incubated in 20 mM HEPES, pH 7.9, containing 25% (v/v) glycerol, 0.42 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride,
0.5 mM dithiothreitol, 1.0 µg/ml pepstatin, 2.0 µg/ml aprotinin, and 1.0 µg/ml leupeptin. Crude nuclear extract was dialyzed against 20 mM HEPES, pH 7.9, containing 20% (v/v)
glycerol, 0.1 M KCl, 0.2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 1.0 µg/ml pepstatin,
2.0 µg/ml aprotinin, 1.0 µg/ml leupeptin, and 0.5 mM
dithiothreitol. The dialyzed crude nuclear extract was snap-frozen in
liquid nitrogen and stored at 80 °C in aliquots. The concentration
of protein in the extracts was estimated according to the method of
Bradford (29), using a Bio-Rad protein assay kit. The amount of protein
was determined at 630 nm in a microplate reader.
Oligonucleotides corresponded to the sense and antisense strands of the
HREN 5-flanking region surrounding the putative CRE. hRen-CRE extended from
239 to
204 with respect to the transcription start site. hRenCRE* and hRenMut spanned the same sequences but were
mutated to the consensus CRE or to the sequence which abolishes CRE
function (Table I). SomCRE was homologous to the rat
somatostatin promoter between nucleotides
59 and
32 with respect to
the transcription start site and contained the consensus CRE sequence
(8). The sequences of the double-stranded gel mobility shift assay
probes and competitors are shown in Table I. Oligonucleotides were
end-labeled at the 5
-terminus using [
-32P]ATP and T4
polynucleotide kinase (30). One strand was labeled and then allowed to
anneal to a 3-fold molar excess of the complementary strand by slow
cooling from 80 °C to room temperature. Labeled double-stranded
oligonucleotides were purified from unincorporated nucleotides by
electrophoresis through a 15% nondenaturing polyacrylamide gel. The
probe was excised from the gel, eluted overnight in 1 mM
Tris-HCl, pH 8.0, 0.1 mM EDTA, 1 M NaCl and was
recovered by ethanol precipitation.
|
The presence of nuclear protein factors with the ability to
specifically bind to DNA was assessed using gel mobility shift analysis. The binding reaction consisted of 10 mM HEPES, pH
7.9, 1 mM EDTA, 5 mM MgCl2, 100 mM KCl, 6% (v/v) glycerol, 0.05% Nonidet P-40, 2 µg of
bovine serum albumin, and 0.04 µg/µl poly(dA-dT) in a final volume
of 25 µl. Crude nuclear extract was preincubated for 15 min at room
temperature in binding buffer. 0.1 pmol of end-labeled double-stranded
oligonucleotide was added, and the incubation was continued for 15 min.
Where appropriate, double-stranded competitor oligonucleotides
(100-fold molar excess or as indicated) were added at the preincubation
step. For supershift assays, monoclonal antibodies to ATF-1 (catalog
number 25C10G) or CREB-1 (catalog number X-12) (Santa Cruz
Biotechnology) was added after addition of the labeled oligonucleotide.
Nonspecific antibody controls included a monoclonal antibody directed
against Oct-1/Oct-2 octamer binding proteins (Upstate Biotechnologies
Inc., Lake Placid, NY) and polyclonal antibodies directed against the
ITF-2 and Pan-1 (E12/E47) helix-loop-helix transcription factors.
Binding reactions were stopped by adding glycerol loading dye. Reaction
products were then electrophoresed through a 4% polyacrylamide gel
(30:1 acrylamide/ bis-acrylamide) in 0.5 × TBE buffer at room
temperature on a BRL sequencing apparatus. After electrophoresis, the
gel was dried for 30 min in a vacuum dryer. Gels were exposed to x-ray film at 80 °C with intensifying screens.
To confirm that transcription of the
luciferase gene was initiated from the HREN transcription
start site, mRNA from Calu-6 cells co-transfected with 900Luc plus
cPKA or JEG-3 cells transfected with 900Luc and treated with forskolin
were isolated, and the start site was determined using an RNase
protection assay. A probe which spans nucleotides 97 to +13 of the
HREN 5
-flanking DNA together with 137 bp of the 5
-end of
the luciferase gene was generated by polymerase chain reaction
amplification of 900Luc using oligonucleotides derived from
HREN 5
-flanking DNA and the luciferase gene. The polymerase
chain reaction fragment was cloned into the pCR-II vector (Invitrogen)
using the AT cloning kit and the directions recommended by the
manufacturer. A partial human
-actin cDNA was obtained from the
manufacturer of the RPAII kit (Ambion, Austin, TX). Total RNA was
isolated from cells by homogenization in 2.5 ml of guanidine
isothiocyanate followed by phenol emulsion extraction at pH 4.0 as
described previously (31). Twenty-microgram samples of total cell RNA
were hybridized to single stranded labeled ([
-32P]UTP)
antisense RNA probes generated by SP6-RNA polymerase using the
Maxiscript Kit (Ambion). RNase protection assay was performed as
described by the manufacturer. Protection products were visualized by
electrophoresis through a 6% polyacrylamide, 7 M urea
sequencing gel that was then dried for autoradiography.
Duplicate flasks of confluent Calu-6 or JEG-3 cells were lysed in sodium dodecyl sulfate sample buffer. Cells were either treated with forskolin for 3 h to minimize the dephosphorylation of CREB or co-transfected with cPKA for 24 h. Whole-cell lysates from each sample (50 µg of protein) were split and resolved on a 10% SDS-polyacrylamide gel, then transferred to polyvinylidene difluoride membrane (Millipore Co., Bedford, MA). A single membrane containing the duplicated samples was cut in half, and the blots were separately probed with antibodies that recognize CREB or specifically the phosphorylated form of CREB using the PhosphoPlus CREB (Ser 133) Antibody Kit (New England Biolabs) according to the instructions provided by the manufacturer. Blots were developed using enhanced chemiluminescence and exposed to x-ray film.
The HREN CRE contains a 6/8-bp match to the consensus
somatostatin CRE. In order to examine the importance of this sequence in the transcriptional response to cAMP in Calu-6 cells, we performed transient transfections with luciferase (L) fusion constructs containing 896 bp of HREN 5-flanking DNA (
896 to +13) and
containing either the native HREN CRE (900L), a mutation to
the somatostatin consensus CRE (900CRE), or a nonfunctional CRE mutant
(900MUT, Fig. 1). Transfections were also repeated in
JEG-3 cells that do not express renin but are highly responsive to
cAMP. The highly cAMP-responsive somatostatin promoter (SSCRE) was also
fused to luciferase and used as a positive control. Forskolin caused a 70-fold increase in transcriptional activity of the somatostatin promoter in JEG-3 cells (Fig. 2B). The
transcriptional activity of each HREN promoter construct
increased significantly in proportion to the strength of the CRE in
JEG-3 cells (900CRE > 900L > 900MUT, Fig. 2B).
In Calu-6 cells, forskolin caused a 2-3-fold induction of luciferase
activity that was greater in cells transfected with 900CRE than either
900L or 900MUT (Fig. 2A). The observation that forskolin
could still stimulate transcriptional activity of the HREN
promoter without a functional CRE (900MUT) suggests the presence of a
non-CRE-dependent mechanism. This aspect of the response will be addressed below. The observation that the SSCRE control construct was induced 70-fold in JEG-3 cells but only 2-fold in Calu-6
cells prompted us to ask 1) whether Calu-6 cells are deficient in
transcription factors binding to the CRE, 2) whether the
HREN promoter could be transcriptionally transactivated in
Calu-6 cells via the cAMP pathway, and 3) whether this transactivation
occurred via a CREB-1-dependent mechanism requiring the
HREN CRE.
CREs are known to form DNA-protein complexes with the CREB/ATF family
of transcription factors. In order to identify the nuclear factor(s)
that bind to the HREN CRE, we performed gel mobility shift
assays using nuclear extracts from Calu-6 cells and labeled double-stranded oligonucleotides hRenCRE (Fig.
3A), hRenCRE* (Fig. 3C), and
somatostatin CRE (SomCRE, Fig. 3B). In preliminary
experiments, the shift products (labeled A-E in
Figs. 3, 4, 5) migrated as a diffuse set of DNA-protein complexes close to
a major nonspecific DNA-protein complex (labeled NS in Figs.
3, 4, 5). In order to resolve these complexes we ran our GMSA on
sequencing length gels. In each experiment the double-stranded
oligonucleotide was present in high excess but migrated four times
further than the lowest point shown in each figure. Five specific shift
products (and one nonspecific product) were observed with all three
probes (Fig. 3). These band shift products were shown to be
specifically competed with excess unlabeled hRenCRE, SomCRE, and
hRenCRE* but were not competed with the same CRE mutant employed in the
900MUT construct (hRenMut oligo) or a nonspecific competitor encoding
the binding site for nuclear factor Y. Since AP-2 and AP-1 can interact
with CREs to mediate a cAMP response (32) and there is an AP-2
consensus homology residing adjacent to the HREN CRE on the
opposite strand (33), two different double-stranded competitor
oligonucleotides were generated that contain the AP2 sequence. The
first contains the AP-2 consensus binding site from the human
metallothionein IIA promoter (34) and the second contains the
HREN AP-2 site and overlapping CRE that incorporated the
hRenMut mutation (AP2Mut). No competition was observed when either
oligonucleotide was added. However, we observed some competition
for hRenCRE when an AP-1 site from the human collagenase gene (34)
was used as a competitor (Fig. 3A).
To identify specifically which of the transcription factors belonging to the ATF/CREB family contribute to the DNA-protein complexes observed in Calu-6 nuclear extracts, we performed a gel mobility supershift assay using ATF-1 and CREB-1-specific monoclonal antiserum (Fig. 4). The ATF-1 antiserum was able to supershift products B, C, and D (and possibly E which forms a weak complex) to a position comigrating with product A. Only band B was supershifted with CREB-1 antiserum. These results suggest that products B-D contain ATF-1 and that product B is a CREB-1·ATF-1 heterodimer. Identical results were obtained with the SomCRE (Fig. 4A) and hRenCRE (Fig. 4B) oligonucleotides. No supershifted products were formed with three different nonspecific antibodies, demonstrating the specificity of the binding (Fig. 4C).
We also examined the relative affinities of CREB-1/ATF-1 proteins in
Calu-6 nuclear extracts for binding to hRenCRE, hRenCRE*, and SomCRE.
Radiolabeled oligonucleotides hRenCRE, hRenCRE*, and SomCRE were
incubated with Calu-6 extracts and increasing amounts of cold hRenCRE*
(Fig. 5A) and hRenCRE (Fig. 5B)
competitor oligonucleotides. Unlabeled hRenCRE was very effective in
competing for its own probe (lanes 2-5 in Fig.
5B), was much less effective as competition for SomCRE
(lanes 12-15 in Fig. 5B), and failed to compete
for binding to hRenCRE* even at the highest concentration tested
(lanes 7-10 in Fig. 5B). In contrast, unlabeled
hRenCRE* was able to compete for binding to all three probes, and the
binding for hRenCRE was competed even at the lowest concentration used
(Fig. 5A). These observations suggest that hRenCRE, SomCRE,
and hRenCRE* all have the potential to interact with the CREB/ATF
proteins in Calu-6 cells but with markedly different affinities
(hRenCRE* > SomCRE hRenCRE).
Previous studies have demonstrated that ATF-1 can inhibit
CREB-1-mediated activity of the somatostatin promoter when present as a
CREB-1·ATF-1 heterodimer (12), and the GMSA results described above
indicate that Calu-6 cells contain the CREB-1 transcription factor in a
heterodimeric complex with ATF-1. We therefore asked whether
HREN promoter activity could be transactivated by
co-transfection of the luciferase constructs with an expression vector
encoding CREB-1. Calu-6 and JEG-3 cells cotransfected with the
luciferase and CREB-1 expression vectors were either left untreated or
were treated with forskolin. In Calu-6 cells, forskolin caused a
significant induction of reporter gene activity in cells containing
900CRE plus CREB-1 but only a small induction in cells containing 900L or 900MUT plus CREB-1 (Fig. 6A).
Interestingly, the increase in 900L and 900CRE transcriptional activity
induced by forskolin in cells containing the CREB-1 expression vector
was only slightly greater than in cells without CREB-1 (compare the
filled bars to the diagonal cross-hatched bars,
Fig. 6A). Similarly, only a 2-fold induction in 900L and
900CRE activity was obtained after forskolin treatment of Calu-6/WT-11
cells, a subline of Calu-6 stably transfected with the CREB-1
expression vector and containing abundant CREB-1 mRNA (data not
shown). In JEG-3 cells (Fig. 6B), there was a significant
induction in 900CRE and 900L promoter activity in forskolin-treated
cells co-transfected with CREB-1. The effects of forskolin treatment
was greatest in cells containing 900CRE, intermediate in cells
containing 900L, but not significant in cells containing 900MUT.
The transactivation of genes through a CRE is proposed to occur by
binding of phosphorylated ATF/CREB transcription factor(s) to the CRE
with recruitment of CREB binding protein (CREBP), and phosphorylation
of CREB/ATF family members is essential for their trans-activation activity. In order to determine if the
HREN promoter could be transactivated by phosphorylation of
endogenous CREB/ATF proteins present in Calu-6 cells, cotransfection
experiments were performed using the test luciferase plasmids and an
expression vector encoding the catalytic subunit of PKA (cPKA).
Overexpression of cPKA in Calu-6 cells significantly increased
HREN promoter activity of 900L by 10.3-fold, 900CRE by
9.1-fold, and 900MUT by 5.7-fold (Fig. 7A).
Similar results were obtained in JEG-3 cells (Fig. 7B).
Interestingly, the cPKA induction in JEG-3 cells mirrored the response
to forskolin (compare filled bars to cross-hatched bars, Fig. 7B). These data demonstrate that the
HREN promoter can be transactivated in Calu-6 cells by
PKA-mediated phosphorylation of transcription factors including,
presumably, CREB-1. Moreover, these results suggest the possibility
that either the basal activity of PKA is low in Calu-6 cells, resulting
in a low level of phosphorylated CREB-1 in response to forskolin, or
that CREB-1 is rapidly dephosphorylated after an initial burst of
activity. We also cannot rule out the possibility that the cPKA
response is due to the activation of other transcription factors
besides CREB-1, since a cPKA-mediated increase in HREN
transcriptional activity was observed in the 900MUT construct (see
below).
In order to confirm that the luciferase activity correlated with
transcription initiating from the HREN start site, total RNA
was isolated from Calu-6 or JEG-3 cells transfected with the 900L
construct and analyzed by RNase protection assay (see strategy in Fig.
1). In order to boost expression to detectable levels by RNase
protection assay, we treated transfected JEG-3 cells with forskolin and
co-transfected Calu-6 cells with the cPKA expression vector. A 150-base
fragment indicating correct transcriptional initiation was clearly
evident in the transfected cells, whereas no protected fragment was
detected in nontransfected control cells (Fig. 8). A
127-base protected fragment encoding human -actin was detected in
all samples indicating equal loading of RNAs in each lane. The results
conclusively indicate that the luciferase activity measured in our
experiments was derived from transcripts initiating faithfully from the
HREN promoter in both Calu-6 and JEG-3 cells.
In order to assess whether the HREN transcriptional response
to PKA was mediated via a CREB-1-dependent mechanism, we
introduced a dominant negative mutant form of CREB-1 (KCREB) into
Calu-6 and JEG-3 cells and repeated the transient expression analyses. In these experiments, luciferase test constructs were co-transfected along with the cPKA expression vector and either a wild-type or dominant negative mutant CREB-1 expression vector. In Calu-6 cells, overexpression of cPKA led to a 9.1-fold increase in 900CRE promoter activity, which was further increased to 19.6-fold with the addition of
wild-type CREB (Fig. 9A). Similar results
were obtained with 900L. Importantly, significant attenuation of the
PKA response was observed in KCREB-transfected 900L and 900CRE but not
in 900MUT cells (compare horizontal hatched bar to
filled and diagonal cross-hatched bars, Fig.
9A). That the PKA response in 900MUT was not attenuated by
KCREB strongly suggests that the induction of HREN promoter activity caused by phosphorylation of transcription factors by PKA has
both a CREB-dependent and CREB-independent component. Moreover, the CREB-independent component is illustrated by the observation that KCREB did not return promoter activity of 900L and
900CRE to base line (compare horizontal cross-hatched bar to
open bar, Fig. 9A), but to a level similar to the
promoter activity of 900MUT in response to cPKA. Similar results were
obtained in JEG-3 cells (Fig. 9B).
Since KCREB has been reported to be dominant repressor of both CREB-1
and ATF-1 (23), we examined whether there was a mechanistic link
between PKA-mediated transactivation of the HREN promoter and phosphorylation of CREB-1. We therefore performed Western blot
analysis to determine whether CREB-1 and ATF-1 were phosphorylated after forskolin treatment in JEG-3 and Calu-6 cells and after cotransfection of Calu-6 cells with cPKA using antiserum specific for
the phosphorylated form of CREB (phospho-CREB). Phospho-CREB antiserum
detected a 47-kDa CREB-1 polypeptide in cells either treated with
forskolin or co-transfected with cPKA (Fig.
10B). The phospho-CREB antiserum does not
detect purified unphosphorylated CREB (Fig. 10B, lane 2)
demonstrating the specificity of the antibody. There was an increase in
the amount of phospho-CREB detected in cPKA co-transfected cells when
compared with forskolin-treated cells (Fig. 10B). This
finding was replicated in three separate transfection experiments. It
is likely that the amount of phospho-CREB in the PKA-induced Calu-6
cells (Fig. 10B, lane 7) when compared to forskolin-treated
Calu-6 cells (lane 6) is underestimated since only a
fraction of the cells in the dish (approximately 20%) were transfected
with the cPKA construct, whereas all of the cells in the dish were
exposed to the forskolin. Equal loading of extracts on the blots is
demonstrated by the equivalent signals observed when antiserum that
recognize both the phosphorylated or unphosphorylated form of CREB was
used (Fig. 10A). The phospho-CREB antiserum also detects the
phosphorylated form of ATF-1 (according to the specifications provided
by the manufacturer of the antiserum). The level of phosphorylated ATF-1 appeared similar in untreated, forskolin-treated, or in cPKA-transfected Calu-6 cells (Fig. 10B).
Transcriptional activation by CREB serves as the final step in the
signal transduction cascade initiated at cell surface receptors to
activate adenylyl cyclase, increase intracellular cAMP, and activate
cAMP-dependent protein kinase (PKA). Phosphorylation of
CREB by PKA is essential for its trans-activating functions (35). In the present study, we have shown that transcriptional activity
of the HREN promoter in Calu-6 cells can be transactivated by the PKA pathway and is mediated by an interaction of phosphorylated CREB-1 with the HREN CRE located at position 225 to
218,
but also by a CREB-1-independent mechanism.
The essential
cis-elements controlling transcriptional regulation by cAMP
have been well characterized (36). The consensus CRE contains an 8-bp
palindrome with the sequence 5-TGACGTCA-3
and was shown to mediate
cAMP induction of the somatostatin promoter. It is also highly
conserved in the promoters of other genes regulated by cAMP (8). CREs
are known to form DNA-protein complexes with members of the ATF/CREB
family of transcription factors. The minimum sequence required for a
functional CRE is the downstream half-site, 5
-CGTCA-3
, although
binding to this site is generally weaker (37). Although our
transfection and affinity results show that the HREN CRE has
a much lower affinity for binding of CREB in comparison to the
consensus CRE, it fits the description of a relatively weak or
asymmetric binding site, containing one copy of the pentamer (CGTCA),
and has the potential to modulate HREN promoter activity to
a small degree as previously reported (4, 5). In some cases, the level
of induction that can be directed by the asymmetric site exceeds that
of the symmetrical site (38). Moreover, CREB-1 binds DNA as a dimer at
the CRE, and while phosphorylation by PKA is not necessary for CREB-1
dimerization (39), phosphorylation of CREB-1 by PKA is important for
the recruitment of CREBP to the transcription complex and may be
required for the binding of CREB-1 to lower affinity asymmetric CREs
(37). Taken together, this may provide a mechanism for tightly
regulating the range of cAMP-induced activation of genes, such as
HREN, which contain an asymmetric CRE. Indeed, this is what
we observed in our studies. Despite our observation that the binding
affinity of ATF-CREB factors in vitro was much higher with
hRenCRE* (consensus) than with the native hRenCRE (Fig. 5), the level
of transcriptional activation observed with 900CRE (consensus) and 900L
(native) in vivo was similar once CREB-1 was phosphorylated
by PKA (Fig. 7).
The similar level of activity of 900L and 900CRE in fully induced (cPKA-transfected) cells may be partially explained by the sequences flanking the HREN CRE, as it has been shown that in addition to the CRE core sequence the composition of the flanking sequences is also a strong determinant of the affinity of transcription factor binding and can influence the strength of the transcriptional response observed (40). In this regard, it is interesting to note that the affinity of ATF-1 and CREB-1 for binding to the hRenCRE* sequence exceeded that of the somatostatin CRE and that the hRenCRE sequence was more effective in competing with SomCRE than hRenCRE* for binding to these factors (Fig. 5).
Mechanism of HREN Promoter TransactivationThe mutant form of
CREB (KCREB) we used in this study has a single nucleotide change from
arginine to leucine in the DNA binding domain and acts as a dominant
repressor of wild-type CREB. KCREB was previously shown to block the
ability of CREB to mediate induction of a somatostatin-promoter
reporter construct in differentiated F9 cells (23). Our results
demonstrate that KCREB attenuated the PKA-mediated induction of
transcriptional activity of constructs containing the native
HREN CRE or consensus CRE, but not of constructs containing
a nonfunctional CRE mutant. The observation that PKA induction of
900MUT transcriptional activity is not attenuated by KCREB suggests
that the induction in HREN promoter activity has a
CREB-dependent and CREB-independent component and that the CREB-dependent component requires the HREN CRE.
It remains unclear what other factors may mediate the CREB-independent
portion of the PKA response. In this regard, an AP-2 homology lies
adjacent to the HREN CRE (on the opposite strand) (4).
Although our GMSA competition results show that the AP-2 sequence could
not compete for ATF-1 and CREB-1 binding to the CRE, we cannot rule out
the possibility that AP-2 plays some role in regulation of the
HREN promoter by cAMP. An AP-2 recognition site mediates the cAMP induction of the metallothionein IIA, growth hormone, prolactin, and proenkephalin genes (reviewed in Ref. 41). In addition, a Pit-1
sequence between coordinates 82 to
58, which binds the pituitary-specific transcription factor Pit-1 and can bind members of
the POU domain family of transcription factors, has also been reported
to be an essential component of the HREN cAMP response (10,
42, 43). HREN promoter constructs containing the Pit-1 site
in the absence of the CRE are still induced 1.7-fold by cAMP (10).
Therefore, both the AP-2 and Pit-1 sites, along with their cognate
factors, may play a role in the CREB-1-independent portion of the
HREN transcriptional response to PKA.
Like CREB-1, ATF-1 exists as both a homodimer and ATF-1·CREB-1 heterodimer in a variety of cell types (44). ATF-1 was reported to be capable of antagonizing CREB-1-dependent activation of the somatostatin promoter by limiting the amount of CREB-1 that can form homodimers (12). Our band shift assays indicate that there were several DNA-protein complexes consisting of ATF-1, one of which was an ATF-1·CREB-1 heterodimer, but no detectable CREB-1 homodimers. In addition to CREB-1 and ATF-1, other Calu-6 factors may have the ability to form DNA-protein complexes with the CRE, and it remains unclear what other transcription factors make up the ATF-1·CRE complexes observed in our band shift experiments. These results are consistent with previous observations that the majority of CREB-1 in HeLa cells is associated with ATF-1 (44-46) and suggests that ATF-1 may suppress cAMP responsiveness by displacing or preventing binding of CREB-1 to the CRE or by preventing the interaction between CREB-1 and other members of the transcription complex. Overexpression of wild-type CREB-1 in Calu-6 cells was not sufficient, however, to induce HREN promoter activity, even when treated with forskolin. Although we are not ruling out a role for ATF-1 in regulating CREB-1 activity in Calu-6 cells, the data suggest that the inactivity of CREB-1 in response to forskolin may be due to another mechanism, such as low basal levels of PKA activity in Calu-6 cells. It is interesting to note in this regard that the levels of phosphorylated ATF-1 are approximately equal under basal or forskolin-stimulated conditions (Fig. 10). This suggests the possibility that phosphorylation of CREB and ATF-1 is differentially regulated in these cells. This is likely since both CREB and ATF-1 are targets of multiple kinases (see below).
Alternatively, it is possible that high level activity of phosphatases may dephosphorylate CREB after forskolin stimulation in Calu-6 cells. Indeed, as phosphorylation of CREB-1 by PKA on Ser-133 is associated with increased transcriptional activity (35, 47, 48), the rate of CREB-dependent transcription has been shown to be diminished by phosphatase treatment (49, 50). Consistent with this possibility is the observation that phosphorylated CREB-1 is detectable after 3 h of forskolin treatment (Fig. 10), but not after 24 h of treatment (data not shown) unless the cells contain the cPKA expression vector. Since our luciferase measurements were performed 24 h after forskolin treatment, it is possible that the absence of significant induction may be due to dephosphorylation of CREB-1. Rapid phosphorylation and dephosphorylation of CREB-1 in renin expressing cells may provide a mechanism to tightly regulate the transcriptional activity of the HREN promoter in response to a changing environment. Clearly, it is necessary to establish whether similar mechanisms are operating in renin expressing cells in vivo.
cAMP is thought to be the major stimulatory second messenger
responsible for causing increased renin secretion in response to
sympathetic nerve stimulation via the -adrenergic receptor mechanism, whereas the suppression of renin secretion is thought to
involve increases in intracellular calcium (51). Angiotensin II causes
increased calcium mobilization and decreased renin release as a result
of angiotensin II type 1 receptor activation (52-54). Of the members
of the CREB/ATF family, both CREB-1 and ATF-1 have been shown to be
responsive to both the cAMP and calcium pathways. Both cAMP, by acting
through cAMP-dependent protein kinase A, and calcium, by
acting through Ca2+/calmodulin-dependent
protein kinases, are capable of phosphorylating CREB-1 on Ser-133 and
thus each can regulate CREB-1-mediated transcription of target genes
(47, 55). Transient expression studies have also shown that ATF-1 can
be activated by increases in either cAMP or calcium (11, 56). Different
members of the Ca2+/calmodulin-dependent
protein kinase family can either activate or inhibit the activation of
CREB-1, suggesting the calcium pathway can either antagonize or act
jointly with the cAMP pathway for activation of CREB-1. Such convergent
regulation of CREB activity by cAMP and calcium may be used to
integrate multiple extracellular signals in renin-producing cells
in vivo. Such a mechanism could be fundamental to the
control of HREN promoter activity by intracellular signal
transduction pathways that either increase transcription by
PKA-mediated phosphorylation of CREB-1 or suppress transcription by
Ca2+/calmodulin-mediated inhibition of CREB-1 activity,
perhaps via ATF-1.
In summary, we have demonstrated that the HREN promoter can be transactivated by the cAMP pathway by both CREB-dependent and CREB-independent mechanisms in both renin expressing and non-renin expressing cell lines. Although the transcriptional response to forskolin alone did not require a functional CRE, the CREB-dependent portion of the response to elevated PKA required the CRE. Our results suggest that either the basal levels of PKA in Calu-6 cells may be low or, alternatively, that CREB is rapidly dephosphorylated after forskolin treatment, and experiments to differentiate between these possibilities are in progress. It is now critical to determine whether these mechanisms can be extended to those cells in vivo which express and release renin in response to physiological inputs. Accordingly, it is interesting to note that HREN promoter activity in transfected primary cultures of renin expressing choriodecidual cells was recently reported to be induced by forskolin only 2.4-fold in a construct containing an intact CRE and only 1.7-fold in a construct without a functional CRE (10) suggesting that similar mechanisms may be operating in this cell type as well.
We thank Julie Lang for her superb technical assistance, Drs. Richard Goodman and Stephen Goodbourn for the gift of plasmids used in this study, and Dr. Andrew Russo for the gift of antibodies.