| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, Texas 75235-8857
Address all correspondence and requests for reprints to: Michael J. McPhaul, M.D., Department of Internal Medicine, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75235-8857.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
- hydroxy-androstenedione to
estradiol, estrone, and estriol, respectively. Aromatase is expressed
in the gonadal and brain tissue(s) of all species studied to date
(1, 2, 3, 4, 5, 6, 7, 8, 9). In some species, P450arom has also been detected in placenta,
adipose, skin, adrenal, and some Leydig tumors (6, 10, 11, 12, 13, 14). It appears that aromatase expression can be regulated in at least three different ways. At one extreme, recent evidence suggests the existence of discrete, independently regulated aromatase isoforms that are expressed in porcine ovary, placenta, and blastocyst (15, 16). In contrast, tissue-specific promoters regulate the expression of a single CYP19 gene in some species (e.g. humans, cows, pigs, horses, chickens, and rats) (15, 16, 17, 18, 19, 20, 21). Finally, the differential regulation of each promoter can result in varying levels of aromatase expression. For example, we have previously shown that transcription initiation of the aromatase gene in rat ovary and in two rat Leydig tumor cells (R2C and H540) occurs immediately 5' to the initiator methionine (proximal promoter, also called promoter II) (6, 22). Although the proximal promoter regulates aromatase expression in both rat ovary and rat Leydig tumor cells, this promoter is regulated differentially by treatment with forskolin, cAMP, or glucocorticoids in these cell types (1, 6, 22, 23, 24, 25, 26, 27, 28). In the studies described here, we have focused on this latter model: the regulation of a single aromatase gene by one of its promoters (the proximal promoter) in different cell types.
By sequence analysis, a steroidogenic factor-1 (SF1)-binding site (at -90 relative to the start of transcription) and two motifs similar to cAMP response elements (5'CRE at -335 and 3'CRE at -231) have been identified in the proximal promoter (24). SF1 (also named Ad4BP), an orphan member of the nuclear receptor family, was first identified as an important regulator of adrenal steroidogenic P450s and appears to be necessary, but not sufficient, for the expression of many steroidogenic P450s (29, 30, 31, 32). The two CRE-like sequences differ from the consensus CRE by nucleotide substitutions of the two central nucleotides. A third CRE-like sequence has recently been identified at -169 by functional analysis (24, 25, 26). The latter CRE-like sequence (which we have designated the XCRE) has an extra nucleotide inserted between the second and third nucleotides of the consensus motif. Each of these CRE-like sequences is a potential candidate in the regulation of CYP19 by forskolin or cAMP analogs (6, 22, 23, 24, 25, 26, 27, 28, 32).
To define the mechanisms controlling the different regulatory patterns of the rat aromatase proximal promoter, vectors containing different sized promoter fragments fused to the luciferase reporter gene were transfected into two rat Leydig tumor cell lines (R2C and H540, which express high levels of aromatase activity), a mouse adrenocortical cell line (Y1, which expresses low, but detectable, levels of aromatase activity), and a rat embryonic fibroblast cell line (Rat2, which has no detectable aromatase activity). The results of these functional studies have been correlated using gel mobility shift assays to identify the proteins binding to functionally important promoter sequences.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The H540 cell line was obtained from Dr. J. I. Mason (University Department of Clinical Biochemistry, Royal Infirmary of Edinburgh NHS Trust, Edinburgh EH3 9YW, Scotland, UK) and was cultured in 50:50 Ham’s F-12-MEM (Mediatech) containing 5% FBS, 1 mg/ml penicillin-streptomycin, and 1% anti-PPLO (Life Technologies).
Aromatase assay
Aromatase activity was measured by the tritiated water release
assay using 0.6–0.8 mM [1ß-3H]testosterone
(New England Nuclear Research Products, Boston, MA) (7, 33, 34). The
protein content was measured by the method of Lowry et al.
(35).
Transfection assay
Promoter fragments were generated by the PCR and subcloned
into the pGL2-Basic vector (Promega Corp., Madison, WI).
The promoter fragment was also sequenced by the dideoxynucleotide
sequencing method (36) to ensure that no unintended mutations were
introduced into the PCR product during amplification. Lipofectamine
(Life Technologies; R2C, H540, Y1: 10
µl/3.5-cm2 dish; Rat2: 5 µl/3.5-cm2 dish)
was used as directed by the manufacturer to cotransfect cells with the
pGL2-promoter constructs (1.5 µg/dish) and the cytomegalovirus
(CMV)-lacZ plasmids (1 µg/dish). Luciferase activity was
measured after 3 days of expression. ß-Galactosidase activity and
protein content (Bio-Rad Laboratories, Inc., Hercules, CA;
Bradford method) were also determined for each sample. The
luciferase data were corrected for transfection efficiency with
ß-galactosidase activity. The corrected promoter activity was then
depicted relative to the most active promoter construct (-688; Figs. 2
and 3). Samples were analyzed in triplicate.
|
|
and 5 also included 1 mM leupeptin. The
pET-
-CRE-binding protein (CREB) plasmid (39), obtained from Dr.
C. R. Mendelson (University of Texas Southwestern Medical Center,
Dallas, TX) was transcribed and translated using the T7 polymerase in
the rabbit reticulocyte lysate system as directed by the manufacturer
(Promega Corp.). Ten microliters of the reaction were used
in mobility shift assays. When the in vitro transcribed and
translated CREB protein product was labeled with 35S and
separated on a 9% SDS-polyacrylamide gel, one major protein of the
expected size (46 kDa) was detected (data not shown).
|
|
Unless specified, oligonucleotides were synthesized by a PE Applied Biosystems (Foster City, CA) 381A DNA synthesizer. The
anti-CREB monoclonal antibody (-CREB) was obtained from Dr. J.
P. Hoeffler (University of Colorado Health Science Center, Denver, CO).
The anti-SF1 antibody (-SF1) was obtained from Dr. K. L. Parker
(Duke University Medical Center, Durham, NC). The anti-Ad4BP (Ad4BP is
the bovine homolog of SF1) antibody (-Ad4BP) was obtained from Dr.
K.-I. Morohashi (Kyushu University, Fukuoka, Japan).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
contains a list of the cell lines used
in functional studies. These cell lines have been chosen to compare the
relative activities of the promoter fusions in cells with different
levels of aromatase activity. The R2C and H540 cell lines are two
independently derived Leydig tumor lines that arose in different
strains of rats (40, 41).
|
. This fragment was subcloned from
a 1.5-kb PstI fragment isolated from a rat genomic Charon
library (19). The promoter fragments, shown schematically in Fig. 2A
, were assayed by transient
transfection in each of the cell lines. The relative transcriptional
activity of this series of 5'-deletions in the rat aromatase promoter
in transient transfection assays is shown in Fig. 2B. When assayed in
the R2C cell line, an increase in transcriptional activity was seen as
the regions containing the 3'CRE and 5'CRE sequences were included in
the promoter fragments (-183 compared with -242 and -339). Inclusion
of the GT-rich sequences (located between nucleotides -339 to -475)
did not increase promoter activity. An additional increase was seen as
the promoter region from -475 to -538 was included. Addition of
sequences 5' to the -688 boundary resulted in an apparent decrease in
promoter activity.
|
As noted in Table 1, the Y1 cells express much lower aromatase activity
than either of the Leydig tumor lines. Despite this, the data obtained
from assays of the promoter fusions were very similar to the results of
assays performed in both the R2C and H540 cell lines.
The transfection experiments performed with the nonsteroidogenic Rat2 cell line (embryonic fibroblasts) yielded markedly different results. Although constitutively active plasmids (pCMV-lacZ, which was used to monitor transfection efficiency among the different aromatase promoter plasmids, and pRSV-luciferase) were transcriptionally active in the Rat2 cell line at levels similar to those observed for the other cell lines, none of the aromatase promoter fragment constructs had significant activity above that seen with the pGL2 parent reporter vector (data not shown).
Transcriptional activity of mutated aromatase promoter
fragments
The above results suggested that the segments containing the
5'CRE, 3'CRE, XCRE, and/or SF1 motif were functionally important. To
test the significance of these motifs more directly, 6 bp of the 5'CRE,
3'CRE, XCRE, and/or the SF1 motifs were replaced with a SalI
restriction endonuclease cleavage site (GTCGAC). A schematic diagram of
the single and double mutant promoter fragments that were generated and
assayed is shown in Fig. 3A. The relative
activities of these seven mutant fragments compared with that of the
wild-type promoter fragment (-688/+94) are shown in Fig. 3B. In the
R2C Leydig tumor cells, single mutations of the SF1 motif (mSF1), the
XCRE (mX), or the 3'CRE (m3) reduced promoter activity by 60–90% of
wild-type activity. In contrast, mutation of the 5'CRE (m5)
consistently resulted in a slight increase in promoter activity. The
combination of m5 with either mSF1 or m3 slightly increased the
activity of the promoter fragments compared with that of the single
mSF1 and m3 mutants. The m3 mSF1 double mutant showed a dramatic
decrease (
95%) in activity compared with the wild-type promoter
fragment (Fig. 3B).
The promoter activities measured in the H540 Leydig tumor cells and the Y1 adrenocortical cells were very similar to the results seen in R2C Leydig tumor cells, except for the mX mutant promoter fusion. The mSF1, m3, m5 mSF1, and m5 m3 promoter fusions exhibited activities decreased by 45–70% compared with that of the wild-type fragment. The m5 mutant showed an increase in promoter activity, whereas the m3 mSF1 mutant only had 10–15% of the activity of the wild-type promoter. As noted above, the only major difference in the behavior of the promoter fusions assayed in H540 and Y1 cells compared with R2C cells was the activity of the promoter fusion carrying the mutagenized XCRE (mX). In the H540 cells, the mX construct had elevated levels of activity compared with the wild-type fragment, and in the Y1 cells, the mX construct had 80% the activity of the wild-type construct, suggesting that this region is less important in the H540 and Y1 cell lines than this region is in the R2C cell line.
As described earlier, although constitutively active pRSV-luciferase and pCMV-lacZ plasmids were transcriptionally active in the nonsteroidogenic cell line Rat2, none of the promoter fragment constructs had significant activity above that seen with the pGL2 parent reporter vector (data not shown).
Nuclear proteins from Leydig tumor cells bind to the
SF1-binding site
As the SF1-binding site is functionally important in transfection
experiments, gel mobility shift assays were performed to further
characterize the nuclear proteins (putative transcription factors) that
can bind to the aromatase SF1 binding motif. Gel mobility shift assays
using the SF1 motif as probe (SF1: see Materials and
Methods) are shown in Fig. 4. The
same results were obtained using a 20-mer oligonucleotide as probe
(20SF1; data not shown).
Although the relative intensities of the shifted bands varied, similar
patterns could be demonstrated using nuclear extracts prepared from
each of the steroidogenic cell lines (Fig. 4, R2C, H540, and Y1; lanes
2, 6, and 10 in A). Three complexes (arrows) could be
specifically competed by excess unlabeled probe (lanes 3–4, 7–8, and
11–12). A 50- to 250-fold molar excess of oligomers containing a
mutated SF1 motif (mutSF1) does not compete with any of the specific
band shifts in most experiments, although a 250-fold molar excess of
mutSF1 appears to slightly compete with specific bands in Fig. 4 (data
not shown and lanes 5, 9, and 13).
The inclusion of an anti-Ad4BP antibody in the reactions eliminated the
specific band shifts of the two slower migrating bands and formed a
supershifted complex (Fig. 4, arrowhead; compare lanes 10
and 14 in A; compare lanes 19–20 and 22–23 in B), whereas incubation
with an unrelated antibody against member of the steroid hormone
receptor superfamily (-AR, which is specific for the androgen
receptor) did not alter the mobility of the specific bands visualized
(lanes 21 and 24 in B). Lanes 15, 17, and 18 are control reactions that
show that the antibody preparations do not have an intrinsic ability to
bind to the DNA probe. In data not shown, a second antibody (-SF1)
slightly retarded the mobility of the two slower migrating bands
without completely eliminating the original band containing SF1-bound
oligonucleotide.
As might be expected on the basis of cell lineage, nuclear proteins that bind the SF1 motif from the aromatase gene were not detected in nuclear extracts prepared from the nonsteroidogenic cell line, Rat2 (data not shown).
Nuclear proteins from steroidogenic cells bind specifically to the
3'CRE and the XCRE, but not the 5'CRE
The transfection data shown in Fig. 2 (promoter fragments -183,
-242, and -339) and Fig. 3 (m3 single and double mutants and mX)
demonstrate that the CRE-like motifs play an important role in
regulating the expression of the aromatase gene in steroidogenic cell
lines. Mobility shift assays using fragments containing these motifs
were performed using extracts from the different cell lines.
A gel mobility shift assay using the 3'CRE motif as probe is shown in
Fig. 5. In functional assays, this motif
was found to be important for promoter activity in all of the
steroidogenic cell lines that we tested (Fig. 3). Nuclear extracts from
each of these cell lines (R2C, H540, and Y1) contained proteins that
could bind to the 3'CRE motif (Fig. 5A, lanes 1, 8, and 15,
respectively). Three bands were visualized in mobility shift assays of
the 3'CRE probe using the nuclear extracts from steroidogenic cell
lines (complexes I–III). Complex II was the major protein-DNA
interaction detected, whereas the levels of complexes I and III varied
among the cell lines. No significant protein-DNA interactions were
detected in Rat2 nuclear extracts that bound to the labeled 3'CRE
oligonucleotides (data not shown).
To determine the specificity of the protein-DNA interactions and to
characterize the DNA-binding site more precisely, competition
experiments were then performed. In experiments using Leydig cell tumor
nuclear extracts, complex II could be completely competed by excess
unlabeled probe (Fig. 5A, lanes 2 and 3 and lanes 9 and 10). Complex I
was not affected by the same concentrations of unlabeled probe,
suggesting that the mobility-shifted probe is interacting with a
nonspecific DNA-binding protein (Fig. 5A, lanes 2 and 3 and lanes 9 and
10). A 100-fold molar excess of mut1 oligonucleotide, a 24-mer in which
the 3'CRE motif has been replaced by a SalI endonuclease
restriction site (Fig. 5B), slightly competed with all of the shifted
bands (Fig. 5A, lanes 4 and 11). Three additional mutant
oligonucleotides also could not eliminate protein-DNA probe
interactions. These mutants include the following: 1) mut2, in which
the first and the last nucleotide in the 3'CRE motif was altered (Fig. 5, B and A, lanes 5 and 12); 2) mut3, in which three nucleotides
preceding the 3'CRE motif were altered (Fig. 5, B and A, lanes 6 and
13); and 3) mut4, in which three nucleotides following the 3'CRE motif
were altered (Fig. 5, B and A, lanes 7 and 14). The SSCRE was not as
efficient a competitor as the unlabeled probe for the proteins that
specifically bind to the 3'CRE probe (Fig. 5C, compare lanes 5 and 6 to
lanes 7 and 8) and excess Sp1 does not compete with the labeled probe
for protein binding (Fig. 5C, compare lanes 4 and 9). Finally, an
excess of oligonucleotides containing the 5'CRE and XCRE sequences
cannot compete with the 3'CRE probe for protein binding (data not
shown).
Similar to the results using R2C and H540 nuclear extracts, complex II
could be fully competed by excess unlabeled probe in experiments using
Y1 nuclear extracts, and complex I was not affected by excess unlabeled
probe (Fig. 5A, lanes 16 and 17). In addition, complex III, which was
detected in the Leydig cell tumor extracts at lower levels, was only
slightly competed by excess unlabeled probe (lanes 16 and 17). The
mutant oligonucleotides (mut1-mut4) could not completely disrupt any of
the shifted bands (lanes 18–21). Again, the SSCRE was not as efficient
a competitor as the unlabeled probe for the proteins that specifically
bind to the 3'CRE probe and excess Sp1 did not compete with the labeled
probe for protein binding (data not shown).
The anti-CREB antibody (-CREB), which recognizes CREB 341, CREM,
and ATF-1, seemed to affect the binding of proteins in complex II from
the Leydig tumor cells (R2C and H540) by slightly increasing the
mobility of complex II. In contrast, -CREB did not affect the
binding of proteins detected as complex I or III in any of the cell
lines (Fig. 5C, compare lanes 3 and 4). As shown in lane 2 (Fig. 5C),
-CREB did not interact with the 3'CRE probe. Anti-ATF2 (-ATF2)
and anti-ATF3 (-ATF3) antibodies had no effect of the mobility of
any shifted band (data not shown).
Gel mobility shift assays were also performed using oligonucleotides
containing the XCRE motif (Fig. 6A). In
functional assays this motif was required for promoter activity only in
the R2C cell line. Specific nuclear protein binding was detected in the
Leydig tumor cells (R2C and H540; lanes 2 and 9), but not in Y1 or Rat2
cells (lanes 16–23). The mobility-shifted bands detected with Y1 and
Rat2 nuclear extracts increased as the amount of nonspecific DNA
competitor, poly(dI-dC) decreased, and these bands could not be
competed by an excess of unlabeled probe. Three complexes were detected
when using nuclear extracts from R2C cells, but only the two slower
migrating bands were formed with H540 extracts. In the Leydig cell
experiments, the band shifts could be competed by excess unlabeled
probe (lanes 3, 4, and 10), but not by excess, unlabeled
oligonucleotides containing the rat aromatase 3'CRE or 5'CRE (lanes 5,
6, 11, and 12) or by excess unlabeled Sp1 oligonucleotide (lanes 7 and
13). The inclusion of the anti-CREB antibody altered the pattern of
band shifts, suggesting that CREB, CREM, and/or ATF-1 are involved in
the formation of the complexes visualized (lanes 8 and 14). As shown in
Fig. 6B, in vitro transcribed and translated CREB protein
was able to form complexes with our XCRE probe in mobility shift
assays.
|
Additional mobility shift experiments were performed to examine the
interaction of nuclear extracts from R2C cells with an oligonucleotide
probe containing a consensus CRE from the somatostatin promoter (SSCRE;
Fig. 7). As shown in lane 3, multiple
protein binding could be detected. An excess of unlabeled probe was
able to compete with visualized bands (compare lanes 3 and 5). It is
also of interest that the XCRE motif from the aromatase promoter was
able to completely compete with the probe, whereas the 5'CRE and 3'CRE
motifs were not efficient competitors (lanes 6–8). In supershift
experiments, -CREB clearly decreased the mobility of some of the
shifted bands (lane 4). The results of experiments using nuclear
extracts from H540 cells and the SSCRE were essentially the same as the
results shown in Fig. 7.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The orphan nuclear receptor, SF1, has been shown to be an important
regulator of steroidogenic P450s (29, 30, 31, 32, 42). The bovine homolog has
been named Ad4BP (30, 32). Our work builds on the data supporting the
paradigm that the SF1-binding site plays a crucial role in controlling
transcription of cytochrome P450 genes in cell lines of both gonadal
and adrenal origins. Mutagenesis of the SF1 binding site (Fig. 3) and
the results of our deletion analysis (Fig. 2, promoter fragments -85
and -183) indicate that the SF1-binding site is necessary, but not
sufficient, for regulating the aromatase gene in the R2C cell line [in
agreement with the studies of Lynch et al. (31) and Carlone
and Richards (26)] and in two additional steroidogenic cell lines
(H540 and Y1). These results are paralleled by the results of mobility
shift assays, which indicate that nuclear extracts from these three
cell lines contain factors that bind to radiolabeled DNA fragments
containing the SF1 motif. Competition experiments indicate that the
protein binding is specific for the SF1 motif.
Similar to our results, a complex pattern of mobility shifts was also
detected in experiments using probes containing the SF1 motifs from the
side-chain cleavage cytochrome P450, steroid 21-hydroxylase, and
aldosterone synthase promoters and nuclear extracts from rat granulosa
cells at various stages of follicular development, Y1 cells, and bovine
adrenal cortex (29, 43). Under the conditions of our antibody
supershift experiments, we consistently observed a slight supershifting
of the proteins bound to the SF1 motif with -SF1. In contrast, the
-SF1, which recognizes the DNA-binding domain of the SF1 protein,
interferes with protein-DNA interactions in experiments performed in
other laboratories (42, 44). The supershifts we detected with
-Ad4BP, which was raised against the full-length protein, is similar
to other reports using the same antibody (45). Taken together, the
oligonucleotide competition experiments and the two antibody studies
support our conclusion that the SF1 protein is involved in forming at
least the two slower migrating complexes. Further experiments analyzing
the complex mobility shift pattern of SF1 will be described
elsewhere.
Our inability to detect SF1 by mobility shift assays in the Rat2 cell line is consistent with the derivation of these cells from a nonsteroidogenic cell lineage. The absence of a transcription factor that can bind to the SF1 motif is in keeping with the lack of detectable steroidogenic P450 expression in this cell line.
FSH can regulate aromatase expression in ovarian granulosa cells and in testicular Sertoli cells (3, 23), and LH can regulate aromatase expression in Leydig cells (1, 27). All of these effects are believed to be modulated at least in part via alterations in the intracellular concentrations of cAMP. In contrast, aromatase activity in the R2C and H540 Leydig tumor cells is not responsive to cAMP, but instead, aromatase activity decreases in R2C cells and appears to be constitutively active in H540 cells (6, 22, 24, 29). These differences suggest the possibility that transcription factors expressed in the Leydig tumor cells can bind to these CRE-like motifs in the aromatase promoter and constitutively activate transcription.
Inspection of the nucleotide sequence surrounding the site of transcription initiation of aromatase messenger RNA identified three potential CRE-like elements. One cis element that was found to regulate aromatase expression was the CRE-like sequence at -231 (3'CRE: TGAAATCA), which bears substantial similarity to the consensus CRE (TGACGTCA) (46, 47, 48). The functional importance of the 3'CRE and the proteins that bind to this regulatory element has not been previously characterized. Deletion analysis and mutagenesis of promoter fragments in transient transfection assays show that the 3'CRE contributes to the activity of the aromatase promoter in each of the steroidogenic cell lines examined in this paper. Using mobility shift assays, we demonstrate that nuclear extracts from steroidogenic cell lines contain factors that bind to the radiolabeled 3'CRE probes, whereas Rat2 nuclear proteins do not interact with this motif. The pattern of band shifts detected in nuclear extracts prepared from R2C, H540, and Y1 cells were similar, but the ratios of the three complexes visualized (I, II, and III) in each cell line differed. Competition experiments suggest that complexes I and III resulted from nonspecific protein binding, because these bands were not competed by an excess of unlabeled probe. These observations suggest that the proteins contained in complexes I and III are not functionally important.
Additional competition experiments were performed for the following
reasons: 1) nuclear proteins from the steroidogenic cell lines could
also bind to an oligonucleotide containing the consensus CRE from the
somatostatin promoter; 2) some of these mobility-shifted bands could
clearly be supershifted by -CREB; 3) in mobility shift assays, an
excess of unlabeled SSCRE was not an efficient competitor for the
proteins bound to the aromatase 3'CRE; and 4) the -CREB supershift
assays were not conclusive for CREB binding to the 3'CRE (the mobility
of complex II slightly increased, instead of decreased).
By mutagenesis and comparisons of natural variants of the consensus
CRE, Deutsch et al. (47) showed that nucleotides at
positions 1 and 4, and to a lesser extent 8
(TGACGTCA; indicated
residues underlined), are required for transcriptional
activity of a CRE. In the aromatase 3'CRE, nucleotide 4 is already
different from the consensus motif
(TGAAATCA),
which begins to suggest that CREB or CREB family members are not
involved in binding to this regulatory element. In mobility shift
assays, competitor oligonucleotides containing substitutions of the
first six nucleotides (mut1) slightly affected nuclear protein
interactions with the 3'CRE, which suggests that these nucleotides may
participate in protein binding or stabilizing protein binding. A longer
promoter fragment that contains the same mutation as mut1 had low
functional activity in transfection assays (Fig. 3). Mut2, containing
substitutions of bases 1 and 8 of the CRE-like motif, was designed to
decrease binding of CREB family members, but in competition
experiments, mut2 still slightly affected nuclear protein interactions
with a 3'CRE probe. Deutsch et al. (47) also found that
flanking sequences can affect both transcriptional activity of a
consensus CRE and protein binding to a consensus CRE. In mut3 and mut4,
three nucleotides flanking the aromatase 3'CRE were changed to the
flanking sequences of -CG CRE, which is very responsive to cAMP.
Mut3 and mut4 were designed to have a stronger affinity for CREB family
members, but when mut3 and mut4 were used in mobility shift competition
assays, we found that these mutant competitors only partially
eliminated protein-DNA interactions, suggesting that mut3 and mut4
retained low levels of protein binding. Comparison of the elimination
of complex II by an excess of unlabeled probe to the remaining protein
binding after competition with an excess of mutated oligonucleotides
suggests that the nucleotides in the wild-type 3'CRE sequence
(TGATGAAATCACAT; indicated residues
underlined) contribute to protein binding. Collectively, our
data also suggest that the protein that binds to the functionally
important 3'CRE is not a member of the CREB family of transcription
factors.
The CRE-like sequence at -169 (XCRE: TGCACGTCA) appears to be
important for the expression of aromatase only in R2C Leydig tumor
cells, but not in H540 Leydig tumor cells, Y1 adrenocortical cells, or
Rat2 embryonic fibroblastic cells. Previous work by Fitzpatrick and
Richards (24, 25) and Carlone and Richards (26) demonstrated that this
motif was required for aromatase promoter activity in R2C cells and in
forskolin-stimulated primary rat granulosa cells. As shown in Fig. 2, a
promoter fragment containing the XCRE and the SF1 motifs (fragment
-183) is inactive when assayed in all of the steroidogenic cells
studied here. Although Carlone and Richards (26) reported that a
similarly sized promoter fragment had transcriptional activity in R2C
cells and in forskolin-treated primary granulosa cells, our results
suggests that the XCRE and SF1 motifs together are not sufficient for
promoter activity. This discrepancy might be explained by the way the
data from the respective laboratories are presented. Carlone and
Richards expressed promoter activity as the percent conversion of
substrate to acetylated substrate (chloramphenicol acetyltransferase
activity), whereas our data are expressed as luciferase activity of the
-183 fragment relative to the activity of a longer more active
promoter fragment (-688). The results of our analysis of the
5'-deletions showed concordant behavior in all of the steroidogenic
cell lines. Surprisingly, however, mutagenesis of the XCRE motif only
had a marked effect on promoter activity in the R2C cell line (Fig. 3),
indicating that the XCRE is much less important for aromatase
expression in the H540 and Y1 cell lines. Examination by mobility shift
assays demonstrates that although R2C and H540 nuclear extracts contain
proteins that can bind to the XCRE motif in mobility shift assays, no
protein-DNA interactions were detected with Y1 and Rat2 nuclear
extracts. In the -CREB supershift experiments, the appearance of
faster migrating bands, with mobilities similar to those of the
nonspecific bands seen with the Y1 and Rat2 nuclear extracts, suggests
that the antibody interferes with specific protein binding and allows
nonspecific protein binding to occur. This interpretation of the
mobility shift assay in which the -CREB antibody was included
suggests that CREB341, CREM, and/or ATF1 is a component of the
shifted bands visualized. In support of this, in vitro
CREB is able to form complexes with DNA probes containing the XCRE
(25), and nuclear proteins from our steroidogenic cell lines
specifically bound to a consensus CRE (SSCRE) were competed by an
excess of unlabeled XCRE oligonucleotides. Of interest, experiments
using R2C nuclear extracts produce a shifted band that is not detected
in the assays using H540 nuclear extracts. This unique complex may
include the transcription factor that is required for aromatase
promoter activity in transient transfection assays in R2C cells. In
addition, this unique complex may be responsible for the elevated
levels of aromatase activity that are characteristic of the R2C tumor
cells relative to the aromatase activity measured in H540 and Y1 cell
lines.
In contrast to our studies of the 3'CRE and XCRE, the CRE-like sequence at -335 (5'CRE: TGAACTCA) does not appear to be important for expression of the aromatase gene. In transient transfection assays in steroidogenic cell lines (R2C, H540, and Y1), mutagenesis of the 5'CRE did not significantly affect promoter activity. In mobility shift assays, no specific protein binding to the 5'CRE was detected in experiments using nuclear extracts from any cell line studied. This finding suggests that the increase in promoter activity observed in functional assays when the region between -242 to -339 is included in the promoter fusions is not due to the proteins that bind the 5'CRE motif at -335. This finding implies that sequences other than the 5'CRE within the -242 to -339 segment mediate this observed increase in promoter activity.
In summary, the goal of these experiments was to examine how the proximal promoter of the rat aromatase gene regulates different patterns of gene expression in different cell types. Functional analysis has identified the following three elements in the proximal promoter of the rat aromatase gene that are important for basal activity in at least one of three steroidogenic cell lines (R2C, H540, and Y1) used in this study: 1) in R2C, H540, and Y1, the SF1-binding site at -90 relative to the start of transcription; 2) in R2C, H540, and Y1, the CRE-like sequence at -231 relative to the start of transcription (3'CRE); and 3) in the R2C cell line, the CRE-like sequence at -169 relative to the start of transcription (XCRE). Further, in these cell lines, not only does there appear to be differential use of multiple cis regulatory elements, but mobility shift assays suggest that in the different cell lines different binding activities are present that recognize specific elements. These observations suggest that the varied levels of aromatase activity in the cell lines that we have studied are probably a consequence of the combinations of proteins in each that bind to these three regulatory elements.
|
Acknowledgments |
|---|
CREB plasmid. We also thank Drs. D. J. Mangelsdorf and
C. R. Mendelson (both at University of Texas Southwestern Medical
Center) for helpful advice during the preparation of this manuscript,
and Michele A. Herbst for isolating the genomic clone containing the
rat aromatase promoter. |
Footnotes |
|---|
2 Current address: The Jackson Laboratory, Bar Harbor, Maine
04609.
Received July 30, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  H. K. Choi, S. H. Roh, H. G. Kim, E. H. Han, H. G. Jeong, and K. W. Kang Enhanced expression of aromatase in p53-inactivated mammary epithelial cells Endocr. Relat. Cancer, March 1, 2008; 15(1): 139 - 147. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Sirianni, A. Chimento, R. Malivindi, I. Mazzitelli, S. Ando, and V. Pezzi Insulin-Like Growth Factor-I, Regulating Aromatase Expression through Steroidogenic Factor 1, Supports Estrogen-Dependent Tumor Leydig Cell Proliferation Cancer Res., September 1, 2007; 67(17): 8368 - 8377. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Silandre, C. Delalande, P. Durand, and S. Carreau Three promoters PII, PI.f, and PI.tr direct the expression of aromatase (cyp19) gene in male rat germ cells J. Mol. Endocrinol., August 1, 2007; 39(2): 169 - 181. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D.-S. Wang, T. Kobayashi, L.-Y. Zhou, B. Paul-Prasanth, S. Ijiri, F. Sakai, K. Okubo, K.-i. Morohashi, and Y. Nagahama Foxl2 Up-Regulates Aromatase Gene Transcription in a Female-Specific Manner by Binding to the Promoter as Well as Interacting with Ad4 Binding Protein/Steroidogenic Factor 1 Mol. Endocrinol., March 1, 2007; 21(3): 712 - 725. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Galmiche, N. Richard, S. Corvaisier, and M.-L. Kottler The Expression of Aromatase in Gonadotropes Is Regulated by Estradiol and Gonadotropin-Releasing Hormone in a Manner that Differs from the Regulation of Luteinizing Hormone Endocrinology, September 1, 2006; 147(9): 4234 - 4244. [Abstract] [Full Text] [PDF]
|
|
|
< |
