| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Reproductive Medicine and Neuroscience University of California, San Diego La Jolla, California 92093-0674
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
An excellent model system for examination of GnRH gene transcription is the GT1–7 cell line. The GT1–7 cell line was immortalized by targeting the oncogene, SV40 T antigen, to GnRH neurons using the 5'-flanking region of the rat GnRH gene in transgenic mice. The culturing of a hypothalamic tumor derived from such mice facilitated the clonal isolation of the GT1–1, GT1–3, and GT1–7 cell lines (4). These cell lines have been invaluable in studying GnRH gene expression, allowing a detailed analysis of the rat GnRH gene regulatory region. A 300-bp enhancer (-1863 to -1571) was identified by deletion analysis of a 3-kb 5'-regulatory region of the rat GnRH gene (5). Additionally, a conserved 173-bp promoter was identified by cross-species similarity (6). Fusion of the GnRH enhancer to the GnRH promoter in a reporter gene plasmid recapitulates the activity seen with the 3-kb 5'-flanking region in transient transfection assays (5) and in transgenic mice (M. A. Lawson, S. B. Nelson, and P. L. Mellon, unpublished).
Many transcription factors have been found to interact with the rat GnRH regulatory regions including Oct-1, GATA-4, SCIP/Tst-1, and Otx2 (Refs. 7, 8, 9, 10). All of these proteins were identified utilizing the GT1–7 GnRH neuronal model system, and colocalization with GnRH has been confirmed in vivo by immunohistochemistry and/or in situ hybridization. Oct-1, a POU-homeodomain transcription factor, binds to two regions within the GnRH enhancer and two regions within the GnRH promoter (7, 11). GATA-4 binds to one site in the GnRH enhancer, and its expression colocalizes with GnRH neurons during embryonic development (8, 9, 12). In vitro synthesized SCIP binds to regions within the GnRH promoter and colocalizes with GnRH expression during mouse embryonic development (10). Otx2, a homeoprotein required for anterior head development (13, 14, 15) related to the Orthodenticle gene in Drosophila, binds to a single site in the promoter and colocalizes with GnRH in the embryo during migration and in the adult hypothalamus (10A ). Although a few of the proteins that bind to the rat GnRH enhancer and promoter have now been identified, none of these proteins is restricted uniquely to the GnRH neurons, supporting the hypothesis that a unique combination of proteins control GnRH-specific expression.
Interactions between the rat GnRH enhancer and promoter are important for maintaining a high level of reporter gene transcription specifically in GT1–7 cells. The enhancer activates transcription of the heterologous Herpesvirus thymidine kinase promoter (TK) only 4 fold in GT1–7 cells, yet when placed upstream of the GnRH promoter, the enhancer activates transcription 55-fold (5). Therefore, interactions must exist between the enhancer and promoter to cause this specific increase in transcriptional activity in GT1–7 cells. We have devised a transient transfection paradigm to compare reporter gene expression between different cell lines. This approach has allowed us to determine that a 62-bp region of the GnRH promoter (footprint 2; FP2) is necessary and sufficient to confer neuronal specificity of GnRH gene expression in vitro. Additionally, a 31-bp element from within FP2 acts synergistically with the enhancer and binds two GT1–7 cell-specific protein complexes, both of which may be homeodomain proteins.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). To accomplish this, chimeric
regulatory regions were constructed using the GnRH enhancer and
promoter and the enhancer and promoter elements of the Rous sarcoma
virus (RSV) 5'-long terminal repeat (LTR). Specifically, the
heterologous RSV enhancer was fused to the GnRH promoter
(RSVe/rGnRHp-luc), and the GnRH enhancer was fused to the RSV promoter
(rGnRHe/RSVp-luc). These regulatory regions were engineered upstream of
the luciferase reporter gene in a reporter plasmid vector. As controls,
reporter genes containing the RSV enhancer on the RSV promoter
(RSVe/RSVp-luc) or the combination of the GnRH enhancer and promoter
(rGnRHe/rGnRHp-luc) were also prepared. These plasmids were transfected
into four different cell lines, NIH 3T3 (mouse fibroblasts), GT1–7
(mouse GnRH expressing neurons), JEG-3 (human placental), and CV1
(monkey kidney fibroblasts). To control for differences in transfection
efficiency and relative expression level in various cell types, the RSV
enhancer and the RSV promoter fused to the Escherichia coli
ß-galactosidase gene (RSVe/RSVp-gal) was used as an internal control
in all transfections. The ratio of RSVe/RSVp-luc values to
RSVe/RSVp-gal values was set to 1 for each cell type. The relative
expression levels of the various reporter genes in the cell types
indicated in Fig. 1 are depicted on a log scale since the differences
are dramatic. There is a significantly higher level of reporter gene
expression in GT1–7 cells than in the other three cell types with the
rGnRHe/rGnRHp-luc, rGnRHe/RSVp-luc, and RSVe/rGnRHp-luc regulatory
regions. These data show that the GnRH enhancer and the GnRH promoter
each play roles in determining GT1–7 cell-specific expression and that
the combination of the GnRH enhancer and promoter together yields a
synergistic degree of specificity.
|
To identify the cell-specific element(s), 5' and internal deletions
were created within the GnRH promoter and transfected into GT1–7
cells, and the results were compared with those from parallel
transfections into NIH 3T3 cells. Subsections of the promoter
containing various footprinted regions were placed downstream of the
GnRH enhancer (Fig. 2A), as previously
described by Eraly et al. (6), or the heterologous RSV
enhancer (Fig. 2B). Here, we have used vectors containing the
luciferase reporter gene instead of the chloramphenicol transferase
reporter gene. When the 5' and internally deleted promoter regions are
placed downstream of the GnRH enhancer, reporter expression in GT1–7
cells was consistently higher than in NIH 3T3 cells (Fig. 2A). This
result demonstrates that the GnRH enhancer can confer specificity in a
heterologous context (5). The relative difference in reporter gene
expression between the two cell types can be calculated by dividing the
GT1–7 values by the NIH 3T3 values (Fig. 2A at right,
relative activity). It is apparent from these comparisons that a large
decrease in cell type specificity occurs when FP2 is individually
deleted and when FP2 through FP7 are deleted (note the log scale). The
role for FP2 in cell type specificity is further substantiated by
transient transfections with the deletions placed downstream of the RSV
enhancer (Fig. 2B). Here again, there is a higher level of reporter
gene expression in GT1–7 cells compared with the NIH 3T3 cells with
the full-length 173-bp promoter although the degree of difference is
diminished due to the lack of the GnRH enhancer. The individual
deletion of FP2 results in a loss of cell type specificity since both
the GT1–7 cells and the NIH 3T3 cells express the reporter gene to the
same degree. Additionally, when FP2 was reinserted, cell specificity
reappears. It remains unclear why deletion of FP2 through FP7 does not
completely abolish the cell-specific expression, but this result may
indicate that there is some specificity inherent to the TATA box
machinery.
|
). To determine whether a smaller
region of FP2 can confer cell type specificity, transcriptional
analysis of a 4-fold multimerized -63 to -33 region (4x-63/-33)
was conducted. The reporter plasmids contained the GnRH enhancer fused
to 4x-63/-33 and the RSV promoter (rGnRHe/4X63/RSVp-luc). When
transfected into GT1–7 cells, rGnRHe/4X63/RSVp-luc increased reporter
gene expression compared with the rGnrHe/RSVp-luc (Fig. 3A). This
increase was not as great as that seen with FP2, but it does suggest
that 4x -63/-33 synergizes with the GnRH enhancer. Thus, we can
conclude that both FP2 and 4x-63/-33 can act synergistically with
the GnRH enhancer to confer cell type specificity.
|
) possibly due to the additional 62 bp inserted between
the RSV enhancer and promoter. Regardless of this overall decrease in
expression, there is a higher level of reporter gene expression in
GT1–7 cells compared with NIH 3T3 cells. Additionally, transfections
into 
T3–1 (a mouse pituitary gonadotrope cell line immortalized in
the same manner as GT1–7 cells) and IMR-32 (a human neuroblastoma cell
line), showed lower relative activity than expression in NIH 3T3 cells
(data not shown), indicating that NIH 3T3 cells provide the best
comparison with GT1–7 cells. Internal block mutations within FP2 in
the context of the RSV enhancer and GnRH promoter did not result in a
change in the relative difference between reporter expression in GT1–7
vs. NIH 3T3 cells (data not shown). Thus, specific sequences
required for this GT1–7 cell-specific expression could not be further
localized. To determine whether the smaller region of FP2 is able to
confer cell type specificity, transcriptional analysis of the 4-fold
multimerized -63 to -33 region was conducted. The reporter plasmids
contained the RSV enhancer fused to 4x -63/-33 and the RSV promoter
(RSVe/4X63/RSVp-luc). Reporter gene expression levels were higher in
NIH 3T3 cells than in GT1–7 cells, demonstrating that the 4x-63/-33
is not able to act independently when placed in a heterologous context.
Thus, we have identified FP2 as the smallest region (62 bp) of the GnRH
promoter that is sufficient to autonomously confer cell-specific
expression of the GnRH gene.
Complexes Specific to GT1–7 Nuclear Extract Bind the -63/-33
Element
Interactions between the GnRH enhancer and GnRH promoter are
likely what confer the cell-specific expression of the GnRH gene based
on our evidence (
Figs. 1–3) and evidence obtained in transgenic mice
(M. A. Lawson and P. L. Mellon, unpublished observations).
Having demonstrated that both FP2 and the -63 to -33 promoter
elements interact with the enhancer to specify reporter gene expression
to GT1–7 cells, we next wanted to determine what proteins may bind to
this region to confer cell specificity. We chose to analyze the -63 to
-33 region because it contains fewer potential protein binding sites,
while still maintaining the capability of interacting with the GnRH
enhancer to confer cell specificity. Previously, we have identified
five GT1–7 nuclear protein complexes binding to -63/-33, one of
which is Oct-1 (complex 5) (11). In these experiments, we used EMSA to
compare the protein complexes bound to the -63/-33 element in GT1–7
nuclear extract with NIH 3T3 nuclear extract (Fig. 4A).
Nuclear extract from T3–1 cells was included as a second cell type
that does not express GnRH. The low mobility complex formed with GT1–7
nuclear extract (complex 5) was previously identified as Oct-1 by
antibody supershift analysis (11). In EMSA with NIH 3T3 nuclear
extract, a complex comigrates with the Oct-1 band from the GT1–7
cells, but addition of an antibody against Oct-1 does not block binding
of the complex. The Oct-1 antibody blocks binding of complex 5 in
GT1–7 and T3–1 nuclear extract. These data suggest that either
Oct-1 does not bind to FP2 in NIH 3T3 cells or that another complex
binding with NIH 3T3 nuclear extract masks the binding of Oct-1.
Regardless, the major low mobility complex in NIH 3T3 extracts binding
to -63/-33 does not contain a significant amount of Oct-1, which may
account for the lower transcriptional activity observed in Figs. 1 and 2. To prove that functional Oct-1 is present in NIH 3T3 cells, we show
that Oct-1 from NIH 3T3 and GT1–7 nuclear extracts binds appropriately
to an octamer consensus sequence (Fig. 4B). Oct-1 binding to -63/-33
is not unique to GT1–7 nuclear extract since nuclear extract from
T3–1 cells forms the Oct-1 complex (Fig. 4A). Additionally, two
complexes with a high mobility (complexes 1 and 2) are present in EMSA
with GT1–7 nuclear extract but not with NIH 3T3. In T3–1 nuclear
extract, complex 1, but not complex 2, is present by EMSA. These
differences in protein complexes binding to FP2 could be responsible
for the cell type-specific interactions between the GnRH enhancer and
promoter.
|
). The
nuclear extracts were from the following cell lines: GT1–7, AtT-20
(mouse pituitary corticotrope), NLT (mouse GnRH-expressing tumor from
the nasal region outside the CNS), NIH 3T3, Y1 (mouse adrenal),
T1–1 (mouse pituitary), T3–1 (mouse pituitary gonadotrope), CV1
(monkey fibroblast), HeLa (human cervical fibroblast), and JEG-3 (human
choriocarcinoma). From the panel of nuclear extracts, it appears that
complex 2 is not present in any of the other cell types tested. In
JEG-3 cells, a complex exists that migrates slightly faster than
complex 2 from GT1–7 cells. It is possible that the protein from JEG-3
cells is a human homolog of the complex 2 protein in mice, but it is
more likely that this complex from JEG-3 cells is a different protein
from complex 2. Complex 1 is present in GT1–7 nuclear extract and
appears to comigrate with complexes seen in NLT, T3–1, and T1–1
nuclear extracts. Thus, complex 1 is partially cell specific and
complex 2 is unique to GT1–7 cells, in the tested cell types.
|
, middle site) that is
homologous to a C/EBP and a homeodomain-binding site. Previously, we
found that C/EBPß, although present in GT1–7 cells, does not bind to
this site, termed the middle site by Eraly et al. (11).
Homeodomain transcription factors recognize an ATTA DNA core motif.
Those homeodomains with a glutamine at position 50 (Q50) bind a CAATTA
or CCATTA motif while a lysine at position 50 (K50) recognizes a GGATTA
motif (18). By sequence comparison, it appears that both the middle and
downstream sites contain Q50 homeodomain binding elements (Fig. 6A). Few of the specific Q50 homeodomain
binding sites have been determined; the best characterized include
engrailed (en), antennapedia (antp), and fushi tarazu (ftz) (19, 20, 21).
Ftz is a member of the antp family of homeodomain proteins while en is
a member of a separate family (22). En and ftz are Q50 homeodomain
transcription factors that bind a core site of CAATTA, an 8 out of 10
match to the middle site in FP2 (Fig. 6A) (20, 21). The downstream site
of the -63/-33 promoter element is also homologous to a Q50
homeodomain binding site that can be bound by ftz or antp, based on the
core CAATTA motif (19). Thus, it is possible that members of the Q50
homeodomain protein family bind to one or both of these regions within
the -63/-33 sequence.
|
, we show the mutant oligonucleotide
probe sequences below the wild-type -63/-33 oligonucleotide probe
sequence. The mutant probes lack the binding of protein complexes to
either the middle site (m2c block mutation) or to the downstream site
(m2e block mutation and m2oct double point mutation) (11). Here we list
the previous results in table form with the addition of a second block
mutation in the middle site, m2d (data not shown), and four mutated
probes corresponding to mutations in the Q50 consensus sites, m2Q1,
m2Q2, m2Q3, and m2Q4 (EMSA shown in Fig. 6B). Complex 5 (Oct-1) and
complex 1 are greatly reduced by the m2e and m2oct mutation,
corroborating the binding of the corresponding proteins in the
downstream site of FP2 (diagram in Fig. 6C). The new m2d mutation, as
well as the previously documented m2c mutation, eliminates complex 4
binding, suggesting that complex 4 binds to the middle FP2 binding
site. Complex 3 is eliminated by all of the block mutations including
m2d, confirming that dramatic changes in the -63/-33 oligonucleotide
disrupt complex 3. We previously indicated that m2c does not disrupt
complex 2 formation (11), but a more detailed analysis shows a
reduction in complex 2 formation on the m2c block mutation
oligonucleotide probe (see also Fig. 7A).
|
|
for compiled data and mutation
sequences). Mutation of the middle ATTA site to GGGG (m2Q1) results in
decreases in complexes 1, 2, 3, and 5 while complex 4 is completely
eliminated (Fig. 6B). A smaller mutation in the middle site, changing
the first and fourth A’s to G’s (m2Q2), results in a decrease in
complex 2 binding and elimination of complex 4. The same mutation in
the downstream ATTA site (m2Q3) decreases binding of complexes 1 and 2
and eliminates binding of complex 5 (Oct-1), as expected (see above).
Since complex 4 is retained by the probe containing the downstream
mutation, it is clear that this complex requires an intact middle ATTA
site although it is not a cell-specific complex (see Fig. 5). Complex 2
appears to bind to the CAATTA motif in the middle site (-55 to -50)
and an ATTA motif (-41 to -38) present in the downstream site.
Conversely, complex 2 may also be maintained by protein-protein
interactions even when its DNA recognition sequence has been disrupted.
Complex 1 is relatively cell specific (see Fig. 5) and binds the
identical site as Oct-1, as is shown by its retention on the probe with
the point mutations of the middle site and its reduction by mutations
in the downstream site. Finally, an oligonucleotide probe containing
the point mutations in both middle and downstream sites (m2Q4) reduces
binding of complex 1 and eliminates binding of the remaining complexes
(Fig. 6B). A diagram summarizing the proteins binding to -63/-33 is
shown in Fig. 6C.
-63/-33 Binds a Complex That Comigrates with a Complex Binding to
a Q50 Homeodomain Consensus Site
To determine whether proteins expressed in GT1–7 cells can bind a
consensus Q50 binding site (represented by the binding site for en and
ftz, termed Q50), EMSA was conducted with -63/-33, m2c, and Q50 as
probes (Fig. 7A). In GT1–7 nuclear extract, the Q50 probe is bound by
a protein complex that comigrates with complex 2 on the probe
-63/-33. This complex is not present in NIH 3T3 cell nuclear extract
(Fig. 7A, arrow) or other cell types (see Fig. 5). In EMSA
with nuclear extracts from hypothalamus and forebrain, but not
cerebellum, a complex is observed to comigrate with complex 2 using the
Q50 probe (data not shown). Furthermore, the Q50 oligonucleotide
competes for the binding of complex 2 bound to the -63/-33 probe
while a mutated Q50 oligonucleotide (Q50mut) does not compete for the
complex (Fig. 7B). Additionally, -63/-33 competes for binding of the
specific complex bound to Q50 (arrow) while the m2c
oligonucleotide (disruption of the middle site) does not compete (Fig. 7B). The two slower mobility complexes, which are abundant in both
GT1–7 and NIH 3T3 nuclear extract, are not competed for by the
-63/-33 oligonucleotide. These data suggest that a Q50 homeodomain
transcription member binds to -63/-33 specifically in GT1–7 nuclear
extract, perhaps modulating the cell-specific interactions between the
GnRH enhancer and promoter.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In the simplest cases, cell type-specific gene expression can be traced to the presence of transcription factors unique to the individual cell types (23, 24, 25). As more is known about tissue-specific control regions, it has been discovered that they often comprise complexes of interacting elements and regulatory proteins with or without uniquely tissue-specific factors (26, 27, 28). These more complex control regions may have evolved to integrate diverse spatial and temporal information in determining cell fate. The GnRH gene thus far falls into the latter class in that the proteins identified to date are expressed in many cell types.
The GnRH regulatory region comprises an enhancer and a proximal promoter, which combined, can confer uniquely targeted expression in transgenic animals, but the enhancer on a heterologous (RSV) promoter is inadequate (M. A. Lawson and P. L. Mellon, unpublished results). Thus, the enhancer must act coordinately with the elements in the promoter. This high degree of interdependence between the regulatory elements may be a quality adapted for specifying expression to a very rare cell type, since activation requires the simultaneous presence of multiple specific proteins, some of which bind to both the enhancer and the promoter (11, 16).
The synergistic activation of transcription by the combination of the
GnRH enhancer and promoter regulatory regions is specific to GT1–7
cells and GnRH neurons. Thus, the GT1–7 cells express the proteins
necessary to facilitate the interaction between the enhancer and
promoter, whereas other cell lines, such as NIH 3T3 cells, do not.
Using transient transfections, we previously found that the GnRH
enhancer can confer cell-type specificity without the GnRH proximal
promoter (5). In this study, we have demonstrated that the promoter can
also confer a degree of cell type specificity in the absence of the
GnRH enhancer. FP2 strongly contributes to cell type specificity, since
deletion of FP2, when the GnRH promoter is coupled with the RSV
enhancer, abolishes preferential reporter gene expression in GT1–7
cells (Fig. 2B). The specificity of FP2 action is further demonstrated
by fusion to the GnRH enhancer or the RSV enhancer, upstream of the RSV
promoter. In these instances, FP2 increased the activity ratio by only
2-fold in NIH 3T3 cells but by 15-fold in GT1–7 cells (Fig. 3, A and
B). Finally, we have identified a subregion of FP2, -63/-33, that
specifically interacts with the GnRH enhancer in GT1–7 cells.
The footprint 2 region of the GnRH gene proximal promoter is complex (reviewed in Ref. 16). It was shown previously that deletion of the FP2 element results in a 20-fold loss of transcriptional activation, the most dramatic loss in activation of any of the promoter deletions (6). Although data from DNase I footprint analysis of block mutations in the promoter suggests the binding of three protein complexes, EMSA reveals five individual complexes within the middle and downstream site (11). This region also confers responsiveness to both phorbol esters and glucocorticoids. Previous data also show the induction of a slower mobility complex by the phorbol ester TPA, a protein kinase C activator (6). Furthermore, the glucocorticoid receptor binds to the 5'-region of FP2 in the equivalent area in the mouse GnRH promoter (29). Thus, the GnRH proximal promoter element, FP2, confers neuronal specificity to GnRH gene expression, as well as responsiveness to hormones and second messengers.
Although FP2 is necessary for conferring several responses to GnRH gene expression, little is known about the proteins that bind to this region. Previously, we had identified Oct-1 binding to the 3'-portion of FP2 (11). Oct-1 may play a role in cell-specific activation of GnRH expression through unique interactions with other ubiquitous transcription factors, through interactions with a GnRH neuron-specific activator, or through interactions coupling the GnRH enhancer to the GnRH promoter. The data presented here show that FP2 of the GnRH promoter can greatly augment transcriptional activation by the GnRH enhancer, thereby demonstrating cross-talk between FP2 and the GnRH enhancer. Oct-1 may be one of the regulators required for this interaction, particularly in consideration of its critical role in transcriptional activation through binding sites in the enhancer (7).
To identify potential cell-specific proteins in GT1–7 cells, we
compared EMSA between GT1–7 and various other cell types. Three
differences exist between GT1–7 and NIH 3T3 nuclear factors binding to
the -63/-33 region of the promoter that could play a role in
cell-specific expression of GnRH. First, Oct-1 containing complexes are
not detected by EMSA that include the -63/-33 region of the GnRH
promoter and nuclear extract from NIH 3T3 cells, in contrast to
assays containing nuclear extract from GT1–7 cells. Rather, in assays
utilizing nuclear proteins derived from NIH 3T3 cells, a strong,
comigrating, but as yet unidentified, complex forms. It is possible
that this complex masks Oct-1 binding or that the formation of the
strong, low-mobility complex occludes Oct-1 binding by occupying
nucleotide sequences necessary for Oct-1 interaction with the -63/-33
site. As mentioned earlier, binding sites for Oct-1 are also present in
the GnRH enhancer, and it is possible that the synergy between the two
elements could be due to Oct-1 or an Oct-1 binding partner. Although
formation of Oct-1 complexes on the -63/-33 site with NIH 3T3 nuclear
extract is not observed, Oct-1 complexes from pituitary-derived
T3–1 nuclear extracts are detected. Furthermore, expression through
the GnRH enhancer and promoter in T3–1 cells is lower than in NIH
3T3 cells (data not shown), indicating that the ability of Oct-1 to
bind -63/-33 alone is not sufficient to confer cell-specific
expression. This observation suggests that other factors dependent on
Oct-1 interaction may be important for cell-specific activation of the
GnRH promoter rather than Oct-1 itself.
A second candidate for cell-specific activation is complex 1. This
complex is not present in NIH 3T3 nuclear extract whereas it appears to
be present in T3–1 nuclear extract by EMSA comigration experiments.
Again, since T3–1 cells express lower levels of reporter gene
driven by the GnRH enhancer and promoter than NIH 3T3 cells, the
presence of complex 1 is not sufficient to facilitate cell-specific
interactions between the GnRH promoter and enhancer. A third candidate,
that which forms complex 2, is relatively unique to GT1–7 cells by
EMSA and may account for the differences seen in transcriptional
regulation between the cell types. As described earlier, the GnRH
enhancer fused to the 4x-63/-33 region results in a higher reporter
expression in GT1–7 cells compared with NIH 3T3 cells. To further
support this idea, the m2c mutation in the -63/-33 EMSA probe
decreased binding of complex 2 (Fig. 7A). When the m2c mutation is
present in the context of the whole promoter and enhancer, reporter
gene expression is only 25% of wild type, the most significant
decrease of all of the FP2 mutations examined to date (11). Thus, it is
likely that complex 2 plays a role in maintaining cell type specificity
and in transcriptional activation of GnRH expression.
Complex 2 is our best candidate for a cell-specific protein binding to FP2. Expression of complex 2 appears to be restricted to GT1–7 cells and to bind a potential Q50 homeodomain transcription factor site, CAATTA. This site is also present in repeating elements in the 3'-region of the GnRH enhancer. EMSA experiments have confirmed that complexes that form on these elements in GT1–7 nuclear extracts comigrate with complex 2 (data not shown, C. G. Kelley and P. L. Mellon, personal communication). This similarity suggests that complex 2 may also bind to the GnRH enhancer and could subsequently foster interactions between the two regulatory regions.
All homeodomain transcription factors bind to the core motif ATTA, which makes it difficult to identify the protein binding by simple binding site homology. Additionally, only a few specific homeodomain DNA-binding sites have been well characterized. To date, antennapedia, engrailed, and fushi tarazu are the only homeodomain proteins known to bind to a CAATTA site (21), and the crystal structure of engrailed bound to TAATTA has been shown (30). In EMSA experiments we have identified a complex that binds specifically with GT1–7 nuclear extract to both the -63/-33 probe and the CAATTA site. We have not yet identified the protein(s) in this complex since supershift antibodies are not available for Q50 homeodomain candidates, antennapedia, engrailed, or fushi tarazu. We have found engrailed 2 RNA and protein in GT1–7 neurons (data not shown) but engrailed 2 has not been shown to be expressed in regions known to contain GnRH neurons (31, 32). Furthermore, Hox proteins, mammalian homologs of antennapedia, are not expressed in the forebrain, and fushi tarazu has no known mammalian homolog. Thus, complex 2 binds to a core homeodomain site in the GnRH promoter and may be a novel or previously identified member of the Q50 homeodomain family. Future experiments will focus on identification of this cell-specific regulator of GnRH gene expression.
In conclusion, we have identified a 31-bp region (-63 to -33) of the rat GnRH promoter that interacts with the GnRH enhancer to increase cell-specific transcription. This region binds Oct-1, which may interact with other ubiquitous or specific transcription factors to control the cell-specific expression. Binding of Oct-1 to the GnRH enhancer and to FP2 of the promoter also may be crucial for the interaction between these two DNA regulatory elements. The data presented here substantiate the assertion that the interactions occurring between the GnRH enhancer and GnRH promoter are necessary for cell-specific reporter gene expression. We have identified cell-specific complexes binding the GnRH promoter that may play a role in interactions between the GnRH enhancer and promoter to control cell-specific expression of GnRH. Here, we suggest that a Q50 homeodomain transcription factor binds to a region of the GnRH promoter that is crucial for neuron-specific interactions with the GnRH enhancer. This protein may interact with other proteins, such as Oct-1 bound to the GnRH promoter and/or enhancer to confer cell-specific expression. Further investigations of cognate binding proteins and their protein-protein interactions will help to clarify the role of the FP2 element in GnRH transcription, elucidating the molecular mechanisms underlying the interactions between the enhancer and promoter of the GnRH gene.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
To create rGnRHe/FP2/RSVp-luc, the FP2 region (-82 to -21) was inserted between the rat GnRH enhancer and RSV promoter in the rGnRHe/RSVp-luc vector. RSVe/FP2/RSVp-luc contains RSVe/RSVp-luc with FP2 (-82 to -21) inserted in the polylinker between the RSV enhancer and RSV promoter. The -63/-33 region was multimerized by using a synthesized oligonucleotide with the following sequence: 5'-CTAGAAGGTGTTCCAATTACATTCCTCATTAAATGG3-' and 5'-CTAGTCCATTTAATGAGGAATGTAATTGGAACACCTT-3'. The oligonucleotide was annealed and inserted into pBSK+. To create the four multimer site, multiple rounds of digestion and ligation were conducted. The 4x -63/-33 multimer was then inserted between the enhancer and promoter of rGnRHe/RSVp-luc and RSVe/RSVp-luc to create rGnRHe/4x63/RSVp-luc and RSVe/4x63/RSVp-luc, respectively.
Cell Culture and Transfections
GT1–7 and NIH 3T3 cells were cultured in DMEM containing 10%
FCS (Omega Scientific, Tarzana, CA), Penn/Strep, glucose, and sodium
bicarbonate. These cells were incubated in 5%
CO2 at 37 C. For transient transfections, GT1–7
cells were split 1:3 and NIH 3T3 were split 1:25 from 100% confluent
plates in 6-cm plates. Cells were incubated overnight and were
transfected with calcium phosphate (33). Briefly, 7.2 µg of reporter
DNA and 2.2 µg of internal control DNA were added to a 15-ml conical
tube; 2x HBS (0.5 ml) was added to the DNA and briefly vortexed, and
0.25 M calcium chloride (0.5) was added drop wise into the
tube while vortexing at low speed. The precipitate was incubated 5 min
at room temperature before addition of 0.5 ml to one 6-cm plate of
GT1–7 cells and one 6-cm plate of NIH 3T3 cells. Sixteen hours later
the cells were washed two times with PBS, and DMEM 10% FCS was
replaced. Cells were incubated 24 h longer and harvested.
Harvesting cells entailed washing the cells three times with PBS and adding harvesting buffer (0.5 ml; 0.15 M NaCl, 1 mM EDTA, and 40 mM Tris-HCl, pH 7.4). Cells were scraped from the plate and placed in 1.5-ml tubes and spun for 30 sec at 14,000 rpm. Buffer was removed, lysis buffer (50 µl; 100 mM potassium phosphate, pH 7.8, 0.2% Triton X-100) was added, and cells were resuspended by vortexing. Cells were spun for 5 min, and supernatant was placed into a new tube and assayed for luciferase and ß-galactosidase activity (Galacto-Light Plus Kit, Tropix, Inc., Bedford, MA). For the luciferase assay, cell lysate (10 µl) was assayed in a 96-well plate, which was read in a luminometer (Microlumat Plus; Microplate Luminometer LB96V; EG&G Berthold, Gaithersburg, MD) using luciferin assay buffer (100 µl; 100 mM Tris, pH 7.8, 15 mM MgSO4, 10 mM ATP and 65 µM luciferin). For the ß-galactosidase assay, cell lysate (10 µl) and 0.25 M Tris, pH 7.8 (10 µ l), were combined and incubated at 48 C for 50 min. This was transferred to a 96-well plate, and diluted Galacton-Plus Substrate (70 µl; 1:100 in Galacto-Light Reaction Buffer Diluent) was added to each well. The plate was incubated for 15 min and assayed using the luminometer. Light Emission Accelerator (100 µl) was injected into each well and read for 10 sec after a 2-sec delay. The luciferase and ß-galactosidase values for a nontransfected plate of cells were subtracted from each transfected plate value. Then luciferase values were divided by the ß-galactosidase values to control for transfection efficiency.
Normalizing Transfection Data and Statistics
To control for differences in expression between the different
cell types, each experiment was normalized. The RSV enhancer fused to
the RSV promoter driving luciferase (RSVe/RSVp-luc) was transfected in
duplicate in each experiment. The internal control, RSVe/RSVp fused to
ß-galactosidase (RSVe/RSVp-gal), was used as an internal control for
each transfected plate of cells. The RSVe/RSVp-luc luciferase values
were divided by the RSVe/RSVp-gal ß-galactosidase values and
averaged. The average was set to 1, and the values for the other plates
were normalized to this value in the individual cell types. Thus, the
values from the individual cell types can be directly compared. The
mean of at least three experiments is depicted. The error bars
represent SEM. In Fig. 1, a single-factor ANOVA and
Dunnett’s Least Significant Difference (LSD) Test were used to
determine significant difference. In Figs. 2 and 3, significance was
measured by paired t test analysis, P 
0.05 as indicated by asterisks.
Oligonucleotides
The -63/-33 oligonucleotide corresponds to the sequences
-63 to -38 and -59 to -33. The m2c, m2d, and m2e oligonucleotides
are identical to the -63/-33 oligonucleotide except for the
substitution of 5'-GCGGCCGC-3' at -58 to -51, -53 to -46, and -45
to -38, respectively. The m2oct oligonucleotide is identical to
-63/-33 with a substitution of a G for the T at positions -47 and
-40. The m2Q oligonucleotides are identical to -63/-33 with the
substitution of G’s for the ATTA from -50 to -53 (m2Q1), a G for the
A at positions -50 and -53 (m2Q2), a G for the A at positions -38
and -41 (m2Q3), and a G for the A at positions -38, -41, -50, and
-53 (m2Q4). The fushi tarazu/engrailed (ftz/en) consensus binding
oligonucleotide (Q50) corresponds to the sequences
CTAGGAAATGTCAATTAAATATCAAG (top strand), and GATCGCTTGATATTTAATTGACATTC
(bottom strand). The ftz/en consensus mutant (Q50mut) oligonucleotide
corresponds to the sequences CTAGGAAATGTCAGGGAATATCAAG (top strand) and
GATCGCTTGATATTCCCCTGACATTTC (bottom strand). Oligonucleotides were
synthesized by Operon Technologies (Alameda, CA) and were
annealed in 50 mM NaCl by heating to 95 C for 5 min and
slowly cooling to room temperature.
EMSA
Nuclear extracts were prepared according to the method
described by Schreiber et al. (34). Annealed wild-type and
mutant oligonucleotides (1 pmol) containing sequences of the GnRH
promoter and consensus sequences were filled in with
[32P]dATP (3000 Ci/mmol Dupont NEN Life Science Products, Boston, MA) and Klenow using standard
procedures (35). Probes were phenol/chloroform extracted and passed
over G-50 micro columns (Amersham Pharmacia Biotech,
Piscataway NJ). Probes were counted in a scintillation counter and
diluted in 50 mM NaCl. The competitor
oligonucleotide was end-filled with Klenow. Binding reactions were
carried out in 10 mM HEPES-KOH, pH 7.8, 50
mM KCl, 1 mM EDTA, 5
mM spermidine, 5 mM DTT,
0.2 mg/ml BSA, 0.5 mM
phenylmethylsulfonylfluoride, 12.5–25 µg/ml
polydeoxyinosinic-deoxycytidylic acid, 10% (vol/vol) glycerol, and 20
mg/ml Ficoll. One femtomole of each probe was incubated with 2 µg of
GT1–7 crude nuclear extract in 20 µl reactions. Reactions were
incubated at room temperature for 5 min, loaded, with current on, into
a 5% polyacrylamide gel [30:1 acrylamide/bisacrylamide, 0.25 x
TBE (130 mM Tris, 45 mM
boric acid, 2.5 mM EDTA), 5% glycerol], and
electrophoresed for 2 to 3 h at 175 V. Gels were prerun for 30 min
in 0.25 x TBE. After electrophoresis, gels were dried and
subjected to autoradiography. Competition reactions were performed by
preincubating the reactions with the specified amount of excess
unlabeled oligonucleotide for 20 min before the addition of probe.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
This work was supported by NIH Grant R01 DK-44838 (to P.L.M.). S.B.N. was partially supported by NIH Training Grant T32 AG00216. C.G.K. was supported by a predoctoral fellowship from the Howard Hughes Medical Institute.
Received for publication October 28, 1999. Revision received May 24, 2000. Accepted for publication May 30, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 globin gene using a new SV40 host-vector
system. Cell 27:279–288[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  H. Otani, F. Otsuka, M. Takeda, T. Mukai, T. Terasaka, T. Miyoshi, K. Inagaki, J. Suzuki, T. Ogura, M. A Lawson, et al. Regulation of GNRH production by estrogen and bone morphogenetic proteins in GT1-7 hypothalamic cells J. Endocrinol., October 1, 2009; 203(1): 87 - 97. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Li, D. Mitchell, J. Luo, Z. Yi, S.-G. Cho, J. Guo, X. Li, G. Ning, X. Wu, and M. Liu Estrogen Regulates KiSS1 Gene Expression through Estrogen Receptor {alpha} and SP Protein Complexes Endocrinology, October 1, 2007; 148(10): 4821 - 4828. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. H. Kim, A. Wolfe, R. N. Cohen, S. C. Eames, A. L. Johnson, C. N. Wieland, and S. Radovick In Vivo Identification of a 107-Base Pair Promoter Element Mediating Neuron-Specific Expression of Mouse Gonadotropin-Releasing Hormone Mol. Endocrinol., February 1, 2007; 21(2): 457 - 471. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Q. Tang, M. Mazur, and P. L. Mellon The Protein Kinase C Pathway Acts through Multiple Transcription Factors to Repress Gonadotropin-Releasing Hormone Gene Expression in Hypothalamic GT1-7 Neuronal Cells Mol. Endocrinol., November 1, 2005; 19(11): 2769 - 2779. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Rave-Harel, N. L. G. Miller, M. L. Givens, and P. L. Mellon The Groucho-related Gene Family Regulates the Gonadotropin-releasing Hormone Gene through Interaction with the Homeodomain Proteins MSX1 and OCT1 J. Biol. Chem., September 2, 2005; 280(35): 30975 - 30983. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. L. Givens, N. Rave-Harel, V. D. Goonewardena, R. Kurotani, S. E. Berdy, C. H. Swan, J. L. R. Rubenstein, B. Robert, and P. L. Mellon Developmental Regulation of Gonadotropin-releasing Hormone Gene Expression by the MSX and DLX Homeodomain Protein Families J. Biol. Chem., May 13, 2005; 280(19): 19156 - 19165. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. L. Givens, R. Kurotani, N. Rave-Harel, N. L. G. Miller, and P. L. Mellon Phylogenetic Footprinting Reveals Evolutionarily Conserved Regions of the Gonadotropin-Releasing Hormone Gene that Enhance Cell-Specific Expression Mol. Endocrinol., December 1, 2004; 18(12): 2950 - 2966. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. R. Thanky, R. Slater, and A. E. Herbison Sex Differences in Estrogen-Dependent Transcription of Gonadotropin-Releasing Hormone (GnRH) Gene Revealed in GnRH Transgenic Mice Endocrinology, August 1, 2003; 144(8): 3351 - 3358. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Heger, M. Seney, E. Bless, G. A. Schwarting, M. Bilger, A. Mungenast, S. R. Ojeda, and S. A. Tobet Overexpression of Glutamic Acid Decarboxylase-67 (GAD-67) in Gonadotropin-Releasing Hormone Neurons Disrupts Migratory Fate and Female Reproductive Function in Mice Endocrinology, June 1, 2003; 144(6): 2566 - 2579. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. G. Kelley, M. L. Givens, N. Rave-Harel, S. B. Nelson, S. Anderson, and P. L. Mellon Neuron-Restricted Expression of the Rat Gonadotropin-Releasing Hormone Gene Is Conferred by a Cell-Specific Protein Complex that Binds Repeated CAATT Elements Mol. Endocrinol., November 1, 2002; 16(11): 2413 - 2425. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Lawson, L. A. Macconell, J. Kim, B. T. Powl, S. B. Nelson, and P. L. Mellon Neuron-Specific Expression in Vivo by Defined Transcription Regulatory Elements of the GnRH Gene Endocrinology, April 1, 2002; 143(4): 1404 - 1412. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. H. Kim, A. Wolfe, G. R. Smith, S. A. Tobet, and S. Radovick Promoter Sequences Targeting Tissue-specific Gene Expression of Hypothalamic and Ovarian Gonadotropin-releasing Hormone in Vivo J. Biol. Chem., February 8, 2002; 277(7): 5194 - 5202. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Roy, N. L. Angelini, H. Fujieda, G. M. Brown, and D. D. Belsham Cyclical Regulation of GnRH Gene Expression in GT1-7 GnRH-Secreting Neurons by Melatonin Endocrinology, November 1, 2001; 142(11): 4711 - 4720. [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 |
