| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
College of Pharmacy (G.H., L.P.) and Department of Anatomy and Neurobiology (J.F.H.), University of Kentucky Medical Center, Lexington, Kentucky 40536
Address all correspondence and requests for reprints to: James F. Hyde, Department of Anatomy and Neurobiology, University of Kentucky College of Medicine, 800 Rose Street (MN224), Lexington, Kentucky 40536-0084.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Only the distal sequence at -527 (gERE1) yielded an ERE-specific DNA/protein complex distinguished by mobility and cross-competition with vERE. The gel mobility pattern of the DNA/protein complex was comparable between the pituitary tissue and tumor extracts. However, DNA/protein affinity estimations demonstrated a greater affinity of pituitary proteins for gERE1 over the vERE sequence. Evidence that the human ER (hER) does recognize the gERE1 sequence in the human galanin gene was provided by electrophoretic mobility shift assays (EMSAs) with Sf9 extracts enriched in recombinant hER. In addition, antibodies specific for the hER recognized the gERE1/protein complex in supershift experiments.
Enhancer activity by gERE1 was detected in transient transfections of the rat GH3 pituitary cell line, resulting in a 4-fold induction of expression driven by the heterologous thymidine kinase promoter in the presence of estrogen. Evidence for ER regulation of the gERE1 enhancer was demonstrated by: 1) inhibition of enhancement using the specific ER antagonist ICI 164,384; and 2) enhancement in HeLa cells that was dependent upon coexpression with hER. Enhancement by gERE1 was half the magnitude as that from the vERE element and may reflect a difference in affinity or composition of the ER complex between the two sequences.
These data demonstrate the presence of a functional ERE sequence within the human galanin gene that could potentially function as a regulatory element for estrogen action in the AP.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Direct estrogen action is mediated through the ER, a member of the steroid receptor superfamily, which regulates transcription in target tissues (13, 14). Transcriptional regulation by estrogen is mediated through receptor dimers binding to the palindromic consensus element [estrogen response element (ERE)] (15, 16). Although estrogen is not necessary for high-affinity binding of ERs to the consensus sequence, bound ligand and protein/protein interactions are necessary for transcriptional activation (17, 18).
Cloning and sequencing the galanin promoter has permitted analysis of transcriptional regulators of galanin gene expression (19, 20, 21). However, a functional ERE, in either the rat or human galanin genes, has not yet been localized (21, 22). The purpose of this study was to identify any putative ERE sequences within 3 kb of the 5'-flanking region of the human galanin gene that may mediate ER action using the following criteria: 1) sequence similarity to the ERE consensus; 2) recognition and binding by the human recombinant ER; and 3) ER-mediated enhancer activity, as defined by transient expression analysis. We report the presence of an ERE-like sequence in the 5'-flanking region of the human galanin gene that can bind human ER (hER) and act as an enhancer of a heterologous promoter in pituitary cells. Thus, this sequence may potentially function to regulate galanin gene expression in the AP, in response to estrogen.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MtTW-10 pituitary tumor cell culture
Trypsin-dispersed MtTW-10 pituitary tumor cells were prepared
from one estradiol-treated rat for each experiment, as previously
described (12). Briefly, tumor cells (500,000 cells/well) were cultured
in 24-well plates (Costar, Cambridge, MA) in DMEM (GIBCO, Grand Island,
NY) containing 10% horse serum, 2.5% FBS, antibiotics, and 1
nM 17ß-estradiol. The tumor cells were treated with
either 50 nM trans-4-hydroxy-tamoxifen (TOT), 50
nM ICI 164,384, or vehicle (0.01% ethanol) for a total of
4 days. Fresh culture medium (containing the ER antagonists or vehicle)
was added after 2 days. Two days later, the culture medium was
collected and frozen at -20 C until assayed for galanin content by RIA
(10).
Northern hybridization
RNA from the cultured pituitary tumor cells was isolated using
guanidinium thiocyanate extraction and centrifugation over a cesium
chloride cushion. Northern blots were hybridized initially with an
oligodeoxynucleotide complementary to rat galanin messenger RNA (mRNA)
(+241 to +288) (8). After 24 h exposure to x-ray film, the blots
were stripped and subsequently hybridized to the human 28S ribosomal
RNA (rRNA), as previously described (23). Oligodeoxynucleotide probes
were end-labeled using [32P]-ATP and T4 polynucleotide
kinase. Relative densities (gray levels) of the autoradiograms were
determined using NIH Image Software. Galanin message levels in each
lane were normalized against the densitometric signal from the 28S rRNA
rehybridization.
EMSA
Oligodeoxynucleotides vERE (5' GATCCCTGGTCAGCTTGACCGGAG 3'),
gERE1 (5' GATCCGGGTACCCCATGACCTCCG 3'), and gERE2 (5'
GATCGGGGGTCAGGCTTGCTCTGG 3') and their complements were synthesized by
the Oligodeoxynucleotide Core Facility at the University of Kentucky
and annealed to form double-stranded 24-bp elements. Double-stranded
glucocorticoid response element (GRE) oligodeoxynucleotides were
purchased from Promega (Madison, WI). Oligodeoxynucleotides were
radiolabeled with [32P]-ATP with T4 polynucleotide kinase
and separated from unincorporated radionucleotides by passage over
Sephadex G-50. AP tissue or MtTW-10 pituitary tumor tissue was
dissected, frozen in liquid nitrogen, and nuclear extracts isolated
using the 1-h minipreparation technique (24). Total protein was
quantitated by Bradford assay and normalized against extraction buffer.
Gel shift assays were carried out in 20 µl of 4-mM HEPES
(pH 7.9), 50 mM KCl, 1.5 mM MgCl2,
1 mM dithiothreitol, 1 mM EGTA, 10% glycerol,
2 µg poly dI:dC, 2 µg nuclear proteins, and 0.1–0.5 ng
radiolabeled probe (ca. 50,000 cpm/rxn) for 20 min at 22 C. For
supershift assays, DNA probes were incubated under the same conditions
for EMSA simultaneously with 50 nM recombinant hER in Sf9
cell extracts and 800 ng Antibody 6, specific to recombinant hER (25).
AP nuclear extracts also were incubated with C311 or MC20 (Santa Cruz
Biotech., Santa Cruz, CA), H222 (generously provided by Geoffrey L.
Greene, The Ben May Institute, Chicago, IL) (35), or antibody 6 (25).
After incubation, DNA/protein complexes were separated by
electrophoresis on a nondenaturing 6% polyacrylamide (30:1) gel by
electrophoresis in 45 mM Tris (pH 8.0), 45 mM
borate, and 1 mM EDTA and visualized by
autoradiography.
Plasmid construction
Synthetic oligodeoxynucleotides for gERE1 (-530 to -506 of the
human galanin gene) (21) were annealed and cloned into compatible
EcoRI and BamHI sites of pBluescript II KS(+)
vector (Stratagene Cloning Systems, La Jolla, CA). The 1.8-kb
XbaI/SstI fragment from pBLCAT2 (26), containing
166 bases of the Herpes Simplex thymidine kinase (TK) promoter and the
chloramphenicol acetyltransferase (CAT) reporter gene, was
directionally cloned downstream from a single copy of the gERE1 insert
to generate the gERE1/TKCAT plasmid. Double-stranded vERE
oligodeoxynucleotides (27) were cloned into the SmaI site of
pBLCAT2. Both hybrid constructs were verified by dideoxynucleotide
sequencing through junction segments.
Cell transfection and CAT assay
GH3 rat pituitary cells or HeLa cells (CCL 82.1, CCL
2; ATCC, Rockville, MD) were cultured in DMEM/Ham’s F-12 containing
2.5% FBS, 10% horse serum, 100 U/ml penicillin, 0.1 mg/ml
streptomycin, and 25 U/ml nystatin. For low estrogen conditions, cells
were grown in phenol red-free DMEM/Ham’s F-12 and 10% gelded horse
serum. One day before transfection, GH3 cells (5 x
105 cells) were plated on poly-L-lysine-coated
petri dishes (35 mm). Plasmids were prepared by FPLC (Pharmacia,
Piscataway, NJ). The cells were washed twice with Opti-MEM 1 (phenol
red-free, GIBCO BRL) and then 2 µg of hybrid CAT expression vector
and cytomegalovirus ß-galactosidase internal control plasmids were
cotransfected using Lipofectin reagent (GIBCO BRL), diluted in Opti-MEM
1. HeLa cells (2.5 x 105 cells) were transfected as
above, with or without 50 ng of a hER expression vector, HEGO
[generously provided by Pierre Chambon, Strasbourg, France (28)].
Transfections were optimized according to the manufacturer’s
directions. Forty-eight hours after transfection, cell extracts were
prepared and assayed for CAT (29) and ß-galactosidase activities
(30). After normalizing against ß-galactosidase activity, CAT
activities are expressed as percent induction over pBLCAT2 values.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Furthermore,
quantitation of galanin steady-state message (
900 bp), by Northern
analysis, shows a similar decrease of 2- and 10-fold with respective
antagonist treatments (Fig. 2). These
data demonstrate that the ER is necessary to maintain elevated levels
of galanin in pituitary cells and that ER does mediate the regulation
of galanin gene expression by estrogen.
|
|
displays two galanin sequences at -527
(gERE1) and -365 (gERE2) with 77% similarity to the full ERE
palindromic consensus. Interestingly, the gERE1 sequence seems to be
composed of one perfect half-site of the ERE and one perfect half-site
of the glucocorticoid response element (35), separated by four
nucleotides. No regions of ERE similarity were found within the first
intron of the galanin gene.
|
).
However, the gERE2 sequence binds only a faster mobility protein
complex, as compared with vERE (Fig. 3b), which is not specifically
competed by an excess of unlabeled vERE competitor. Hence, the gERE2
does not demonstrate ERE-specific protein binding and is eliminated as
a potential high-affinity binding site for the ER in the AP.
|
).
|
, lanes 3 and 9). Cross-competition with
unlabeled vERE sequence resulted in partial competition for protein
binding to the gERE1 probe (Fig. 4a, lane 4). However, an excess of
unlabeled gERE1 sequence completely competed protein binding to the
vERE probe (Fig. 4a, lane 10). Not surprisingly, the gERE2 sequence did
not compete for the binding to either gERE1 or vERE probes (Fig. 4a, lanes 5 and 11). Despite the sequence similarity between gERE1 and GRE,
no competition for binding was observed with an excess of GRE
oligodeoxynucleotide (Fig. 4a, lanes 6 and 12). In addition, the
mobility pattern of protein binding to the gERE1 did not resemble that
observed for a GRE probe (data not shown). The cross-competitions
suggest that the gERE1 sequence does not primarily function as a
glucocorticoid receptor binding site and are consistent with the
presence of the ER in the AP complex.
Figure 4b repeats the same cross-competition profile as in Fig. 4a, with nuclear proteins isolated from MtTW tumor tissue. Growth of the
tumor in the presence or absence of estrogen did not influence the
protein binding to gERE1 in a manner similar to that seen with AP
tissue (36). Observations of partial competition by the vERE competitor
and lack of competition by the GRE competitor were duplicated with the
MtTW extracts (Fig. 4b, lanes 6–9), further substantiating the
similarities between AP and MtTW tumor tissues.
Affinity of AP nuclear proteins is greater for gERE1 than vERE
sequences
To compare affinities of the two sequences for the AP protein
complex, nuclear protein binding to the gERE1 probe was titrated with
varying amounts of unlabeled gERE1 or vERE oligodeoxynucleotides (Fig. 5). Approximately a 10-fold greater
amount of the vERE element was required to yield the same degree of
competition by the gERE1 element, thus, demonstrating a greater
affinity of AP proteins for the gERE1 sequence over the vERE sequence.
The ability of gERE1 to completely compete for protein bound to the
vERE probe in Fig. 4 suggests that the gERE1 sequence can bind ER.
However, the partial competition by vERE suggests that different
components of the ER complex seem to be uniquely bound to the gERE1
sequence for which the vERE cannot abolish binding (Fig. 4a, lane 4;
Fig. 4b, lanes 6 and 7; Fig. 5, lane 9).
|
). Figure 6b demonstrates strong binding
of the hER to the vERE sequence (lane 2), which is supershifted by the
anti-hER antibody 6 (lane 3). As shown in Fig. 6a, the gERE1 sequence
is also bound by proteins recognized by antibody 6 (lanes 2 and 3). The
bound complex is of similar mobility to that seen with the vERE probe
in Fig. 6b. Lastly, the gERE2 sequence did not yield a supershifted hER
complex when incubated with Sf9 extracts and antibody 6 (data not
shown), consistent with the previous EMSA data with pituitary extracts.
Competitions similar to those demonstrated in Fig. 4 were repeated with
the supershifted complex. Self-competition with excess unlabeled probe
was able to diminish the supershifted complex for both gERE1 and vERE
sequences (Fig. 6, a and b; lane 4). Cross-competition by vERE
completely competes for the gERE1 supershifted complex (Fig. 6a, lane
5). Although binding of the hER seems to favor the vERE more than the
gERE1 sequence, antibody 6 recognition of the protein/gERE1 complex and
partial competition of the vERE supershifted complex by gERE1 provides
evidence that the hER, indeed, can recognize the gERE1 sequence.
|
GH3 pituitary cells, grown under low estrogen conditions
(see Materials and Methods), were transfected with each of
the hybrid vectors in the absence or presence of 10 nM 17
ß-estradiol (Fig. 7a). The gERE1/TKCAT
hybrid construct resulted in an approximately 4-fold induction of CAT
expression in the presence of estrogen over the basal level of
gERE1/TKCAT without estrogen. Transfection of the vERE/TKCAT hybrid
vector yielded a 9-fold induction, whereas pBLCAT2 gave no estrogen
induction over basal expression levels. Enhancement from gERE1/TKCAT
was approximately half the magnitude as that measured for the
vERE/TKCAT positive control. To test whether the observed enhancement
from gERE1/TKCAT was dependent upon ER, the pure estrogen antagonist
ICI 164,384 (33) was used to block ER activity (Fig. 7b). Treatment
with 50 nM of the antagonist ICI led to a significant
3-fold inhibition of enhanced expression from both the gERE1 and vERE
hybrid vectors, indicating involvement of the ER in enhancer activity
from the gERE1 element.
|
). The
positive control vector, vERE/TKCAT, also demonstrated a 9-fold
enhancement of expression after cotransfection with the HEGO expression
vector. Taken together, the findings from the various functional
studies demonstrate estrogen-responsiveness for the gERE1 sequence and
are consistent with our binding data that suggest that enhancement from
the gERE1 sequence, in part, is caused by ER interaction.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The conclusion of our study is in contrast to that recently reported by
Kofler et al. (21), who did not detect a functional ERE
within the 5'-flanking region of the human galanin gene. However, the
two studies differ in several aspects. Kofler et al.
identified three ERE-like sequences in the human galanin 5'-flanking
region located at -1162, -361, and -122, based on similarity to the
5' GGTCA 3' half-site sequence on the sense strand. Our study, however,
analyzed galanin gene sequences for similarity to the full-length
bipartite palindromic sequence (Table 1), which identified two
sequences located at -527 and -365. Consequently, the gERE2 sequence
at -365 corresponds to one of the ERE-like sequences identified by
Kofler et al., which we report here to be nonfunctional in
EMSA. Analyzing sequences on the sense strand, the gERE1 at -527 does
not contain a perfect match to the sequence used by Kofler et
al. (21). However, it does contain a perfect match to the more
proximal ERE half-site, 5' TGACC 3', which previously has been shown to
be necessary for binding of the ER in vitro (17). Comparison
between gERE1 and the 5'-flanking sequences of galanin genes from other
species may reveal similar sequences.
It is generally thought that ER can regulate gene transcription by binding to a perfect ERE palindromic sequence, 5' GGTCANNNTGACC 3', as a homodimer (15). However, ER can also bind with lower affinity to imperfect ERE sequences by forming heterodimers with nonhormone proteins or through other protein-protein interactions (17, 18, 37, 38, 39). Through such protein interactions, imperfect EREs can render tissue-specificity to hormonal regulation that is not observed with perfect palindromic sequences (38, 39). Chambon and co-workers (37) have documented that several half-palindromic ERE sequences can act synergistically to achieve optimal and cell-specific estrogen induction. Such may be the case with the gERE1 element.
When compared with binding of ER homodimers to a complete palindromic sequence, palindromic half-sites only bind ER monomers weakly in vitro (16, 17, 37). Our EMSA and competition assays suggest an affinity difference between bound proteins from AP extracts and the Sf9 extracts enriched in hER. This difference could be indicative of a species difference between the affinity of rat (AP) and human (Sf9) ER for the human gERE1 sequence. Alternatively, this difference may reflect the presence of tissue-specific accessory proteins, present only in pituitary tissue, which may enhance the affinity of the receptor for the imperfect palindrome (39). Lastly, a recent report by Friend et al. (40) describes the existence of two estrogen-inducible forms of the ER, TERP-1 and TERP-2, present in pituitary tissue. TERP-1 displays both a pituitary-specific and female-specific pattern of expression. However, the physiological significance of these ER isoforms in mediating estrogen response is not yet known. Using a panel of antibodies against the ER in supershift assays (see Materials and Methods), we have been unable to achieve antibody recognition of the pituitary complex bound to the vERE element (G. Howard, L. Peng, and J. F. Hyde, unpublished observation). This may be indicative of a pituitary-specific form of the ER present in the vERE-bound complex or blockade of antibody recognition by tissue-specific accessory proteins. Whether our EMSA data, demonstrating strong binding by pituitary nuclear proteins to the gERE1 sequence, represents binding of the specific TERP-1 isoform remains to be determined.
The conclusions of Kofler et al. (21) are based primarily
upon transient transfections performed in neuroblastoma/glioma cells
and not pituitary cells. However, estrogen-dependent regulation from an
imperfect palindrome may not be seen in all cell types (37). In
neuroblastoma cells, expression from 1.9 kb of the human galanin
promoter increased 30-fold in the presence of 17ß-estradiol but only
at nonphysiological levels of cotransfected ER (21). Using
GH3 pituitary cells, we were able to detect
estrogen-dependent induction of expression from various reporter
constructs (Fig. 7). In addition, cotransfection of the gERE1 hybrid
construct with a hER expression vector resulted in enhanced CAT
expression in HeLa cells (Fig. 8). Considering both the detected
enhancer function and ER binding data, the gERE1 sequence could be
responsible, in part, for the 10-fold estrogen induction observed in
Fig. 2. Based on our results, the estrogen-dependent increase of the
1.9 kb of promoter reported by Kofler et al. may indeed
occur at physiological levels of ER when tested in pituitary cells.
The magnitude of endogenous galanin gene expression in MtTW cells (Fig. 2) is in agreement with the preliminary data from pituitary tissue
reported by Kaplan et al. (22) for the rat promoter.
However, whereas estrogen induces endogenous galanin gene expression
10-fold in pituitary tissue, a single copy of the gERE1 sequence was
able to enhance expression of the heterologous TK promoter only 4-fold.
The cellular and promoter context of the bound receptor influences the
magnitude of transcriptional regulation of the promoter (37). Such
quantitative differences between gERE1/TKCAT and the endogenous galanin
gene may reflect the absence of additional enhancers that impact the
magnitude of the estrogen induction. In addition, pituitary-specific
enhancer elements also may be required for optimal induction from
estrogen. Whether the other three ERE-like sequences reported by Kofler
et al. (21) or the full promoter context influences the
magnitude of induction from the gERE1 palindromic half-site at -527
remains to be determined.
The physiological roles of galanin in the pituitary are still to be elucidated. The correlation of increased galanin gene expression with estrogen-induced pituitary tumors, and decreased galanin gene expression when tumor formation is inhibited (8, 9), has led to the hypothesis that galanin may mediate the hyperplastic events associated with estrogen-induced pituitary tumor formation (9). Indeed, galanin increases the proliferation of human small-cell lung carcinoma cells and 235–1 pituitary cells in vitro (41, 42). Moreover, we recently reported that estrogen-treated lactotrophs that contain galanin mRNA secrete significantly greater amounts of PRL than those lactotrophs not containing galanin message (43), suggesting a possible autocrine regulation of PRL secretion by galanin in estrogen-stimulated, hyperplastic pituitary cells. The role of galanin in the human pituitary gland awaits further study.
In conclusion, identification of a functional ERE element within the gERE1 sequence supports the possible mechanism of direct regulation of galanin transcription by the ER in AP tissue. Whether the tissue-specific estrogen induction of galanin gene expression occurs through the binding of pituitary-specific forms of the ER to gERE1 or whether it involves interactions with additional tissue-specific proteins warrants further investigations.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received May 5, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Y.-L. Yu, C.-H. Pon, H.-C. Ku, C.-T. Wang, and Y.-H. Kao A preprogalanin cDNA from the turtle pituitary and regulation of its gene expression Am J Physiol Regulatory Integrative Comp Physiol, April 1, 2007; 292(4): R1649 - R1656. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Merchenthaler, G. E. Hoffman, and M. V. Lane Estrogen and Estrogen Receptor-{beta} (ER{beta})-Selective Ligands Induce Galanin Expression within Gonadotropin Hormone-Releasing Hormone-Immunoreactive Neurons in the Female Rat Brain Endocrinology, June 1, 2005; 146(6): 2760 - 2765. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Counts, S. E. Perez, S. D. Ginsberg, S. de Lacalle, and E. J. Mufson Galanin in Alzheimer Disease Mol. Interv., May 1, 2003; 3(3): 137 - 156. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Peng and A. H. Payne AP-2gamma and the Homeodomain Protein Distal-less 3 Are Required for Placental-specific Expression of the Murine 3beta -Hydroxysteroid Dehydrogenase VI Gene, Hsd3b6 J. Biol. Chem., March 1, 2002; 277(10): 7945 - 7954. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Freeman, B. Kanyicska, A. Lerant, and G. Nagy Prolactin: Structure, Function, and Regulation of Secretion Physiol Rev, October 1, 2000; 80(4): 1523 - 1631. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Giustina and J. D. Veldhuis Pathophysiology of the Neuroregulation of Growth Hormone Secretion in Experimental Animals and the Human Endocr. Rev., December 1, 1998; 19(6): 717 - 797. [Abstract] [Full Text]
|
|
|
|
|
|
C. Faure-Virelizier, D. Croix, S. Bouret, V. Prevot, S. Reig, J.-C. Beauvillain, and V. Mitchell Effects of Estrous Cyclicity on the Expression of the Galanin Receptor Gal-R1 in the Rat Preoptic Area: A Comparison with the Male Endocrinology, October 1, 1998; 139(10): 4127 - 4139. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
