| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Studies |
Department of Obstetrics and Gynecology, Jones Institute for Reproductive Medicine, Eastern Virginia Medical School, Norfolk, Virginia 23507
Address all correspondence and requests for reprints to: Ke-Wen Dong, Ph.D., Jones Institute for Reproductive Medicine, Eastern Virginia Medical School, 601 Colley Avenue, Norfolk, Virginia 23507. E-mail: DONGKW{at}EVMS.EDU
Abstract
In previous studies, we have localized four specific nuclear protein-binding elements in the human GnRH upstream promoter. To test whether these four elements are reproductive tissue specific, we placed the four elements upstream to a thymidine kinase (TK) promoter/luciferase reporter gene, and transfected the constructs into human placental choriocarcinoma (JEG-3) cells. The 272-bp fragment (-994 to -723) containing the four elements can drive heterologous TK promoter expression in JEG-3 cells about 15 times more than that of basal TK promoter activity. Deletion of element 4 (E4, -987/-968) significantly decreased (4-fold) the luciferase activity. Further deletion of the other elements (E3 individual, -960/-940 or E3 and E2 in combination, -919/-896) only slightly decreased the luciferase activity. In contrast, deletion of element 1 (E1, -876/-851) caused a 2-fold loss of luciferase activity and elimination of E2 and E3 only lost less than 2-fold of the luciferase activity. Study performed with 5' end deletion of this region confirmed these observations. Furthermore, E4 DNA-protein complex can be supershifted by Oct-1 antibody, indicating that Oct-1 binds to E4. These results clearly demonstrated that all four elements are required to confer tissue-specific expression of the hGnRH gene in JEG-3 cells. However, the E4 is the most important for the tissue-specific expression of the hGnRH gene in JEG-3 cells. Oct-1 factor binds with E4 element and may be involved in the mediation of the human GnRH upstream promoter activity.
THE HUMAN GnRH hormone (hGnRH) is the hypothalamic-releasing factor that controls the functions of pituitary gonadotropin cells and, ultimately, reproductive competence. This decapeptide is produced in hypothalamic neurons and acts as a neuroendocrine factor to regulate the synthesis and secretion of luteinizing hormone and FSH. In addition to its classical neuroendocrine functions, GnRH also exerts paracrine and autocrine actions in the reproductive tissue (1) and directly modulates reproductive function (2, 3, 4) through the GnRH receptor present in these tissues (5, 6). In human placenta, for example, expressions of GnRH gene (7) and GnRH receptor gene (8) have been well established. The studies of GnRH and GnRH receptor expression in human placenta suggest important roles of GnRH in the physiological regulation of pregnancy (9, 10, 11). This is further supported by the two-promoter system of the hGnRH gene (7). The downstream transcription start site is used mainly in the hypothalamus, and the upstream transcription start site positioned at 579 bases upstream of the downstream one is predominantly used in the reproductive tissues (7). Further studies demonstrated that the upstream promoter is used to direct tissue-specific expression of the hGnRH gene in the human placental JEG-3 cells but not in neuronal GT1–7 cells (12). Deoxyribonuclease I footprinting analysis of the upstream promoter revealed four specific DNA elements that are capable of binding to the nuclear proteins from the JEG-3 cells but not the GT1–7 cells. The interaction of nuclear proteins from the JEG-3 cells with these four elements was further confirmed by gel mobility shift assay (12). These results suggest that these four elements may act as cis-regulatory elements to direct tissue-specific expression of the hGnRH gene in human placenta. However, it remains unclear how these elements are involved in the tissue-specific expression. Whether these four elements act independently or interactively to direct tissue-specific expression of the hGnRH gene in human placenta is still unknown.
Detailed sequence analysis revealed that the four elements are short AT-rich regions that are possible bound by the POU-homeodomain transcription factors. The POU-homeodomain protein Oct-1 is known to be present in a variety of tissues and cell types (13) and is believed to involved in tissue-specific gene expression by interacting with other transcription factors (14, 15) or tissue-specific coactivators (16, 17). The role of Oct-1 in tissue-specific expression has been demonstrated in neuronal GT1–7 cells for rat GnRH (18) and in placental JEG-3 cells for leukemia inhibitory factor receptor gene (19). Thus, Oct-1 may also be involved in tissue-specific expression of the human GnRH in the placenta. In the present study, we analyzed these four footprinted elements in detail by using a luciferase reporter gene to determine the contributions of each of these four elements for their ability to direct the tissue-specific expression of the hGnRH gene in JEG-3 cells.
Materials and Methods
Construction of the hGnRH upstream promoter/TK luciferase constructs
A 272 bp (-994 to -723) AflII/BamHI
fragment containing the four footprinted elements was cut out from the
full length 5'-flanking region of the hGnRH gene and ligated into a
herpes simplex thymidine kinase (TK) minimal promoter/luciferase
reporter, pT109luc. The individual or combined sequences of the four
elements (E1, E2, E3, and E4) (Fig. 1A
)
were generated by PCR using the following specific primers:
|
A2: 5'-GTCGACAATAAATGTAATAAGAGTAATGGT-3';
B1: 5'-TATCTCGAGAAGCACTCATTACAATAG-3';
C1: 5'-GTCGACTTTCAGCCTAGGGGTAAATTAG-3';
D1: 5'-CAGTCGACAAGAGATTTAAATAAG-3';
C2: 5'-GGTCGACCTACCTGGTTAAAAGCT-3';
D2: 5'-GGTCGACCTACCTGGTTAAAAGCT-3';
E2: 5'-GGTCGACCTACCTGGTTAAAAGCT-3'.
Each of the amplified DNA fragments was inserted into the pT109luc vector. All constructs were transformed into DH5 Escherichia coli competent cells, and positive clones were identified by restriction mapping, Southern blot analysis, and DNA sequencing. For serial deletion study, a set of 5' end deletion in the hGnRH upstream promoter/luciferase constructs was generated by ExonucleaseIII/Mung Bean Nuclease deletion (Stratagene, La Jolla, CA). The sites of deletion were determined by DNA sequencing. The bacteria containing the positive clone were amplified in Luria-Bertani medium and plasmid DNAs were prepared by Concert nucleic acid purification system (Life Technologies, Inc., Gaithersburg, MD).
Cell culture and transient transfection assay
Human placental choriocarcinoma cells (JEG-3) and human kidney cells (293) were purchased from American Type Culture Collection (Rockville, MD) and grown to 50–60% confluence in DMEM and MEM with 10% FBS, respectively. The cells were plated at a density of 2–5 x 10 5 cells per 60-mm dish 1 day before transfection. The medium was then replaced 2 h before transfection. Transfections were performed with the calcium-phosphate coprecipitation method (20) with 3 µg of the test plasmid for each 60-mm plate. Cells were incubated for 16–18 h with the DNA and then rinsed with ice-cold 1x PBS three times. The cells were incubated in the fresh DMEM medium for a further 24 h before harvesting for luciferase activity assay. Protein concentration of the cellular extract was determined by BCA Protein Assay (Pierce Chemical Co., Rockford, IL) to standardize the luciferase activities. To correct for different transfection efficiencies, 2 µg of pCMV ß-galactosidase construct was cotransfected into cells with each test plasmid as an internal control. A portion of the harvested cell extract was used to detect ß-galactosidase activity by a colorimetric assay at 405 nm.
Luciferase assay
To determine luciferase activity, transfected cells were washed once with 1x PBS and lysed in 200 µL of lysis buffer (1% Triton-X 100, 25 mM glycylglycine, pH 7.8, 15 mM MgSO 4, 4 mM EGTA and 1 mM DTT). After centrifugation at 4 C, 14,000 rpm for 5 min, luciferase activity was measured on a Lumat LB9501 luminometer (14) by mixing 100 µL of the cell extracts, 360 µL of luciferase assay buffer (25 mM glycylglycine, pH 7.8, 15 mM MgSO4, 4 mM EGTA, 15 mM KPB, pH 7.8, 1 mM DTT, and 2 mM ATP), and 100 µL of luciferin buffer (2 mM D-luciferin and 25 mM glycylglycine, pH7.8).
Nuclear extract preparation
Nuclear extracts were prepared as described previously (12). Briefly, JEG-3 cell pellets were resuspended in ice-cold buffer A [10 mM HEPES-NaOH, pH 7.9; 10 mM KCl; 1.5 mM MgCl 2; 1 mM DTT; 50 µg/ml phenymethylsulfonylfluoride (PMSF)] and incubated on ice for 10 min. The cell nuclei were separated from cytoplasmic supernatant by centrifugation at 4 C, 12,500 rpm for 25 min, and resuspended in ice-cold buffer C (20 mm HEPES-NaOH, pH 7.9; 10 mM KCl; 1.5 mM MgCl 2; 1 mM DTT, 50 µg/ml PMSF, 0.25 mM EDTA, and 25% glycerol). Then the suspension was homogenized in Dounce homogenizer. After rocking at 4 C for 30 min, the nuclei extract was spun down and the supernatant was dialyzed against 250 mL buffer D (20 mM HEPES-NaOH, pH 7.9; 100 mM KCl; 1 mM DTT, 50 µg/ml PMSF, 0.2 mM EDTA, and 20% glycerol). Then the nuclei extract solution was collected and stored at -80 C for subsequent gel-shift assay. Protein concentration was measure by BCA protein assay kit (Pierce Chemical Co., Rockford, IL).
Electrophoretic mobility shift assay (EMSA) and supershift assay
For the EMSA, annealed wild-type oligonucleotides
containing sequences of the four footprinted elements (E1: -876 to
-851, 5'-AATAAATGTAATAAGAGTAATGGTTG-3'; E2: -919 to -896,
5'-TTTCAGCCTAGGGGTAAATTAG-3'; E3: -960 to -940,
5'-GAGAGATTTAAATAAGATG-3'; E4: -987 to -968,
5'-CTCACTAAATGCCGGGGG-3') were labeled with [
-32P]ATP (6,000 Ci/mmol; NEN Life Science Products, Boston, MA) by using T4 polynucleotide kinase. Binding
reactions were carried out in 10 mM Tris-HCl, pH 8.0, 50
mM KCl, 1 mM EDTA, 1 mM DTT, 4 µg
of poly (dI:dC), 5% glycerol, 3–4 µg of nuclear extract, and 20,000
cpm of labeled probe in total volume of 25 µL. Reaction mixes were
incubated at room temperature for 10 min, loaded into a 10%
polyacrylamide gel (acrylamide-bisacrylamide [37.5:1], 0.25x TBE),
and electrophoresed for 2 h at 200 V in a cold room. Gels were
prerun for 30 min in 0.25x TBE. Then the gels were dried and subjected
to autoradiography. For competition reactions, the reactions were
preincubated with the specified amount of excess unlabeled
oligonucleotide at room temperature for 20 min before the addition of
probe. Supershift assays were performed by incubating the reactions
with 1 µL of Oct-1 antibody (sc-232x, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or rabbit IgG control at
room temperature for 1 h before the addition of probe.
Data analysis
All transient transfection assays were carried out in triplicate and repeated on at least three separate occasions. The results were standardized by protein concentration and ß-galactosidase activity, and expressed as relative light unit (x 1000) and mean ± SEM. Statistical analysis was evaluated by Student’s t test and one-way ANOVA followed by Dunnett’s test.
Results
Four cis-regulatory elements (E4–E1) confer the tissue-specific expression of the full hGnRH upstream promoter activity in JEG-3 cells
To determine the role of the sequence (-994 to -723) that
contains the four cis-regulatory elements in regulating the
hGnRH upstream promoter activity in human placenta, transient
transfection with the four-element TK/luciferase reporter plasmid were
performed in JEG-3 cells, which were used as in vitro human
placental model. The results (Fig. 1B
) showed that the promoter
activity conferred by these four elements alone (TKU) was about 15
times more than that of basal activity (P < 0.01),
which was similar to that of the full-length hGnRH upstream promoter
activity. In contrast, the TK promoter alone had very low activity in
JEG-3 cells. All tested constructs were not active (P
> 0.01, compared with control TKU) in the human kidney 293 cells.
These results indicate that the four cis-regulatory
elements, but not the TK promoter, confer tissue-specific expression of
the TKU construct in JEG-3 cells.
Serial deletion analysis supports the role of the four cis-regulatory elements in the tissue-specific expression of the hGnRH gene in the JEG-3 cells
To investigate the relative roles of the four elements in the
tissue-specific expression of the hGnRH upstream promoter in JEG-3
cells, a detailed deletion analysis was carried out. As shown in Figure 2, deletion of the hGnRH upstream
promoter up to -991 bp retained the majority of upstream promoter
activity. However, mutation of E4 by replacing with an unrelated
sequence significantly decreased the upstream promoter activity
(P < 0.01, compared with hU5). Further deletion up to
-935 bp (E3 and E4 were deleted away) resulted in only a slightly
decrease of the full upstream promoter activity (P >
0.01). Deletion of three elements (E2 to E4) up to -880 bp did not
result in a significant additional decrease in the upstream promoter
activity (P > 0.01). However, deletion of all four
elements up to -827 bp, human GnRH upstream promoter activity was
dramatically decreased to a level almost 10-fold below that of the
construct (hU5) that contains all four elements (P <
0.01). In contrast, the vector itself (pxp2) produced only negligible
luciferase activity. Thus, the results of detailed deletion analysis
indicated that the four cis-regulatory elements are
essential for tissue-specific expression of the hGnRH gene in the human
placental JEG-3 cells. Among these cis-regulatory elements,
the E4 element may be the most important for the human GnRH upstream
promoter activity.
|
To determine how these four cis-regulatory elements
contribute to the tissue-specific expression of the hGnRH upstream
promoter in JEG-3 cells, different combinations of these four elements
-TK/luciferase reporter plasmids were constructed (as described
in Materials and Methods) and transfected into JEG-3 cells.
As showed in Fig. 3, deletion of E4
dramatically decreased the luciferase activity by nearly 4-fold
(P < 0.01), but deletion of E1 caused only a 2-fold
decrease of luciferase activity (P < 0.01).
Elimination of E2 or E3 by PCR knock-out studies caused only less than
a 2-fold decrease of the luciferase activity (Fig. 3, TKU431 and
TKU421) (P < 0.01). Additional deletion of E3 or E2
from either side only slightly decreased the luciferase activity (Fig. 4, TKU21 and TKU 43). Further
deletion of three of the four elements from either side (E4/E3/E2 or
E3/E2/E1) resulted in no significant additional decrease (Fig. 4, TKU1
and TKU4) (P > 0.01).
|
|
, deletion of E4 significantly decreased
the luciferase activity (nearly 4-fold, P < 0.01).
Further deletion of the other elements (individual E3 or E3 and E2 in
combination) only slightly decreased the luciferase activity
(P > 0.01). In contrast, deletion of E1 caused a
2-fold loss of luciferase activity (P < 0.01). The E2
or E3 knock-out decreased only less than 2-fold of luciferase activity,
respectively (P < 0.01). These results together with
5' serial deletion and E4 mutation studies clearly demonstrated that
all four cis-regulatory elements are required to confer
tissue-specific expression of the hGnRH gene in the JEG-3 cells.
However, the E4 is the most important cis-regulatory
sequence for the tissue-specific expression of the hGnRH gene in the
JEG-3 cells.
|
Gel mobility shift assay was conducted to determine the
transcriptional factors binding to the four footprinted elements. As
expected, all of the four cis-regulatory elements bind with
proteins isolated from the JEG-3 nucleus (12). All the
bound protein can be competed out with 500-fold excess of unlabeled
corresponding oligonucleotides, indicating the specific DNA-protein
binding (12). Supershift assay using Oct-1 antibody was
carried out to determine whether Oct-1 protein is directly involved in
the DNA-protein interaction. In Fig. 6A, the EMSA using E4 as probe showed that E4 DNA-protein complex was
supershifted by Oct-1 antibody but not by rabbit IgG (data not shown).
However, the EMSA experiments using E1, E2, and E3 as probe did not
demonstrate any supershift by Oct-1 antibody (data not shown). These
results indicated that Oct-1 binds to E4 and may function as a
transcriptional regulatory protein. In addition, the gel shift of E4 in
Figure 6A also exhibited a specific doublet DNA-protein complex, which
was competed out with cold E4. This doublet may represent other POU
homeodomain protein(s) and/or other transcription factor(s) with E4
complex. To determine whether the Oct-1 protein is expressed in our
JEG-3 cell, Western blot assay was carried out using Oct-1 antibody. As
show in Figure 6B, the Oct-1 protein is predominantly presented in the
nuclei of JEG cells.
|
Complex mechanisms have been evolved to ensure proper spatial and temporal expression of genes in differentiated tissues. In part, this control is achieved at the transcriptional level so that regulatory regions containing multiple cis-acting sequence elements control not only the level of gene expression but also its restriction to appropriate cell types. In previous studies, we demonstrated that the hGnRH gene utilizes the downstream transcription start site in the hypothalamus and the upstream transcription start site in the reproductive tissues to produce different GnRH mRNAs (7). The upstream promoter is used to direct tissue-specific expression of the human GnRH gene in the human placental JEG-3 cells but not in the neuronal GT1–7 cells (12). This differential promoter activity has also been confirmed (21). To further characterize the upstream promoter in tissue-specific expression of human GnRH gene, our previous studies (12) have localized a 325 bp region between -1048 and -723 involved in mediating the activity of the upstream promoter of the hGnRH gene in human placental JEG-3 cells. Deoxyribonuclease I footprint study and gel mobility shift analysis indicated four specific elements in this region that bind to nuclear extracts only from human placental cells (JEG-3) but not from human hypothalamic neuronal cells (GT1–7; 12). These results strongly suggest that these four cis-regulatory elements may be involved in tissue-specific expression of the hGnRH gene in the human placental JEG-3 cells.
In this study, we have further characterized the contribution of these
four cis-regulator elements to the human GnRH upstream
promoter activity. To determine the reproductive tissue specificity of
these elements, we placed the four elements upstream to a TK
promoter/luciferase reporter gene and transfected the constructs into
human placental JEG-3 cells. The 272-bp fragment (-994 to -723)
containing the four elements can drive heterologous TK promoter
expression in JEG-3 cells about 15 times more than that of basal TK
promoter activity. In contrast, analysis of the individual elements for
their ability to mediate the upstream promoter activity demonstrated
that deletion of the E4 element resulted in a significant decrease
(nearly 4-fold) of luciferase activity. Further deletion of the E3
element individually, or the E3 and E2 elements in combination produced
only a slightly additional decrease. In contrast, deletion of the E1
element caused a 2-fold decrease of luciferase activity. Elimination of
the E2 or E3 element only lost less than 2-fold luciferase activity.
Moreover, mutational analysis (Fig. 2) demonstrated that
replacement of the E4 element with an unrelated sequence caused a
significant decrease of upstream promoter activity. These data strongly
suggest that the E4 element is very important for the human GnRH
upstream promoter activity. However, our study also demonstrated that
the E4 element alone produced only 40% of full TKU activity (Fig. 5).
Furthermore, serial deletion analysis revealed that removal all of the
four elements resulted in almost complete loss of the upstream promoter
activity. Thus although the E4 element is the major
cis-regulatory element, all four elements are required to
confer tissue-specific expression of the hGnRH gene in the JEG-3
cells.
In the rat GnRH gene promoter, the results of studies from several groups (22, 23, 24, 25, 26) demonstrated the presence of potent neural-specific enhancer as well as cis-acting regulatory elements. This neural-specific enhancer is located in the distal promoter region between -1863 and -1571 in the rat GnRH promoter, while the cis-acting regulatory elements are crowded in the proximal region of 173 bp 5' of the transcriptional start site (22, 26). Results of studies using GT1–7 neural cell line have demonstrated that the hGnRH gene also possesses a neural-specific enhancer (21) but its location in the proximal promoter region (between -535 and -47) is different from that of the rat neural-specific enhancer. Further studies have demonstrated that the sequences of the proximal region of GnRH genes are conserved across species and consist of multiple cis-acting elements that are necessary for basal and regulated GnRH gene transcription (26). A nucleotide sequence comparison of the GnRH promoter between rats and humans revealed that there is high homology in the proximal promoter region between -1 to -551 of human GnRH and -1 to -424 of rat sequences (21). In contrast, the human GnRH promoter has little similarity in the distal promoter region -1131 to -551) when compared with that of rat -1031 to -424). It is interesting to find that human GnRH gene upstream promoter is located in this region, but the rat does not have second promoter (27). The results of our studies suggest a species-specific difference in the structural organization of cis-acting promoter elements used to confer tissue-specific expression in human placenta.
The presence of putative DNA-binding sites for transcription factors on the 5' flanking region of the hGnRH gene has been reported (21). Since the POU homeodomain protein Oct-1 is expressed in a variety of tissues and cell types (13) including placental JEG-3 cells and neuronal GT1–7 cells, it is believed that Oct-1 may participate in tissue-specific gene expression by interaction with either other transcription factors (14, 15) or tissue-specific coactivators (16, 17). In the rat GnRH promoter, it has been reported that binding of Oct-1 to promoter elements is required for tissue-specific expression in the hypothalamic cell line, GT1–7 (18). Close examination of the four footprinted elements from the human GnRH upstream promoter revealed short AT-rich sequences, all containing the octamer-like sequence TAAAT, which is most likely the target of POU homeodomain proteins. Although the octamer-like motifs within these four elements deviated somewhat from the octamer consensus sequence ATGCAAAT, Oct family proteins are known to interact with AT-rich sequences with relaxed specificity (28). In the present study, we demonstrated that Oct-1 binds only to the E4 element (5'-CTCACTAAATGCCGGGGG-3'), which shows a four of eight match to the octamer consensus sequence but not the E3 element (5'-GAGATTTAAATAAG-3'), which represents a six of eight match, as well as other elements. Furthermore, Western blot assay has demonstrated the expression of Oct-1 protein in our JEG-3 cells. Therefore, our study suggests that Oct-1 may participate in the placental cell-specific expression by directly binding to tissue-specific E4 cis-regulatory element and interacting with the other tissue-specific proteins. Because Oct-1 is not known to be a strong transcriptional activator by itself but in conjunction with other coactivators, the Oct-1 complex can promote potent transactivation of target genes. Thus, it will be of interest to identify the coactivators that interact with Oct-1 to mediate hGnRH gene expression in human placental cells by using yeast two-hybrid system.
In conclusion, the results of the present study demonstrate that all four cis-regulatory elements, E4-E1, are required for full human GnRH upstream promoter activity in human placental JEG-3 cells. However, the E4 element is the most important cis-regulatory element for the tissue-specific expression of the hGnRH gene in the JEG-3 cells. Furthermore, our studies demonstrate that Oct-1 binds to E4 and mediates the human GnRH upstream promoter activity providing more evidence to support the involvement of POU family in GnRH gene expression.
Acknowledgments
We would like to thank Chen Ya-di and Liu Chi-ting for technical assistance, and Dr. Ki-Lie Yu for critical reading of the manuscript.
Footnotes
1 This work was supported by NIH Grant [IR29HD/CA30244 (to
K.W.D.)]. 
2 Current address: Molecular Therapeutics Section, NEI, NIH, 9000
Rockville Pike, Bethesda, Maryland 20892.
Received October 11, 2000.
Revised December 14, 2000.
Accepted February 22, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  B. A. Adams, J. A. Tello, J. Erchegyi, C. Warby, D. J. Hong, K. O. Akinsanya, G. O. Mackie, W. Vale, J. E. Rivier, and N. M. Sherwood Six Novel Gonadotropin-Releasing Hormones Are Encoded as Triplets on Each of Two Genes in the Protochordate, Ciona intestinalis Endocrinology, May 1, 2003; 144(5): 1907 - 1919. [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 |
