| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Splicing Variant Mediates Both Positive and Negative Effects on Gene Transcription
Michigan State University Department of Physiology East Lansing, Michigan 48824-1101
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
(ER) mRNA
is typically expressed as a mixture of transcripts. Using PCR
amplification, this heterogeneity has been shown to result largely from
an imprecise pattern of mRNA splicing that gives rise to a family of
correctly processed and exon-skipped ER transcripts. We have
reconstructed ER cDNAs representing the single exon-skipped variants
ER
E2 through ERE7 to enable their functional characterization in
a well defined cell transfection system. All six of the ER splicing
variants support the efficient expression of stable proteins in Cos7
cells, and each shows a characteristic pattern of subcellular
distribution. Each of the variants displays a dramatic reduction in
DNA-binding activity with a consensus estrogen response element (ERE)
in an in vitro gel mobility shift assay. While this
DNA-binding defect appears to be complete for ERE2, ERE3,
ERE4, and ERE6, weak DNA binding is observed for ERE5 and
ERE7. Scatchard analysis of hormone binding demonstrates that among
the variants, only ERE3 binds 17ß-estradiol
(E2) and does so with an affinity similar to
wild-type ER (wt ER). Individual variants cotransfected with the
pERE-TK-CAT reporter plasmid [a consensus ERE-driven chloramphenicol
acetyltransferase (CAT) reporter gene that is highly responsive to
E2-liganded wt ER] were ineffective at
inducing CAT expression in ER-negative HeLa cells. Only ERE5 showed
indications of positive transcriptional activity on the pERE-TK-CAT
reporter, but this activity was limited to approximately 5% of the
activity of wt ER. When variants were expressed simultaneously with
wt ER, ERE3 and ERE5 were observed to have a dominant negative
effect on wt ER transcriptional activity. Like the wild-type
receptor, both ERE3 and ERE5 interact with steroid receptor
coactivator-1e (SRC-1e) in vitro; however, only ERE3
retained the ability to dimerize with wt ER. Transcription from a
region of the ovalbumin promoter, which contains an ERE half-site and
an AP-1 motif, is positively regulated by liganded wt ER and ERE3
in phorbol ester-treated, transiently transfected HeLa cells. In both
cases, this activity was enhanced by cotransfected cJun. These
observations suggest that selected ER splicing variants are likely
to exert important transcriptional effects, especially on genes that
are regulated by nonconsensus EREs and subject to complex hormonal
control. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The transcriptional effects of estrogens are mediated by two
closely related receptor isoforms, ER and the more recently
described ERß (10, 11), each of which is encoded by a separate gene.
While ERß is also being investigated for its potential role in
various diseases, including cancer, this study focuses solely on the
ER isoform. Analysis of mRNA prepared from a variety of
estrogen-responsive cells and tissues, including breast tumors, has
established that ER mRNA is typically expressed as a mixture of
transcripts (12, 13, 14, 15). This heterogeneity results largely from a pattern
of alternative mRNA splicing that gives rise to a family of correctly
processed and exon-skipped ER mRNAs. ER mRNA comprises sequences
from 8 coding exons and is translated to yield a protein with discrete
functional domains. An N-terminal transactivation function (AF1)
encoded by exon 1 and a portion of exon 2 is thought to promote gene
transcription by interacting with nuclear receptor coactivators and
also with proteins integral to the transcription initiation complex (1, 16, 17). Derived from exons 2 and 3 is a centrally located zinc-finger
motif (commonly referred to as the DNA-binding domain or DBD) that is
essential for sequence-specific DNA binding and transcriptional
activation through canonical estrogen response elements (EREs) (18).
Within the region encoded by exon 4 are the nuclear localization
signals (NLS) and a hinge region that allows for receptor
conformational flexibility (3, 19). A ligand-binding domain (LBD)
confers regulatory function to the receptor and is encoded by the
C-terminal exons 4 through 8 (20). This region also includes
determinants for subunit dimerization and a well characterized
C-terminal transactivation function (AF2), which promotes gene
transcription by recruiting coactivators (1, 2, 3, 21). Like other nuclear
receptors, ER is a modular protein in that individual domains are
capable of demonstrating autonomous function within receptor mutants,
as well as when they are introduced into heterologous fusion proteins
(1, 18). It can reasonably be assumed that the exclusion of a
particular exon will predictably result in a protein lacking the
function ascribed to that exon. Additionally, it is probable that the
loss of a particular exon will result in unpredictable functional
deficits or perhaps even bestow a novel function on the variant
receptor. This study examines the function of ER splicing variants
from the vantage point of what is known about the functional
organization of wt ER. Concurrently, the process of examining
splicing variants, like mutational studies, improves our understanding
of wt ER function. We report results from experiments designed to
assess receptor capacity to translocate to the nucleus, bind to DNA,
bind ligand, participate in protein complexes, and promote gene
transcription.
Fuqua and colleagues (22, 23) have reported that ERE5 (which
contains the AF1 domain, but lacks AF2 and the regulatory functions
imparted by the LBD), is constitutively active in promoting
transcription from an ERE in a heterologous yeast reporter gene assay.
These authors have also described that overexpression of ERE5 in a
stably transfected breast cancer cell line (MCF-7) supported greater
proliferation compared with control cells, as well as imparting a
tamoxifen-resistant phenotype (24). In the human osteosarcoma cell line
U2-OS, it has recently been reported that coexpression of ERE5
significantly enhances ERE-directed reporter gene expression induced by
wt ER (25). The existence of a constitutively active receptor
variant (such as ERE5) able to exert a mitogenic effect in breast
tumor cells in the absence of E2 or in the
presence of tamoxifen is an appealing explanation for the acquisition
of antiestrogen resistance observed in previously responsive tumors and
cell lines (26, 27). However, this model is challenged by conflicting
observations that ERE5 and closely related, genetically engineered
ER mutants do not efficiently induce transcription from an ERE
reporter in transiently transfected ER-negative HeLa or CEF cells (2, 28), or promote proliferation in stably transfected MCF-7 cells
(28).
Recently, a novel mechanism for mediation of an estrogen response has
been reported to involve AP-1-directed regulation of transcription by
ER (29, 30, 31, 32). AP-1 describes the fos/jun family of transcription factors
that play a key role in transducing the effects of growth factors to
regulate cell proliferation (33, 34). A variety of estrogen-responsive
genes have been described that lack a palindromic ERE, but instead
contain one or more consensus AP-1 elements (5'-TGAG/CTCA-3'), with or
without a degenerate ERE or ERE half-site (5'-GGTCA-3' or 5'-TGACC-3').
Examples of such genes include ovalbumin, which is induced by
E2 in chicken oviduct cells (35), and the
insulin-like growth factor-I (IGF-I) gene whose expression is
stimulated by E2 in the uterus of
ovariectomized-hypophy-sectomized rats and in cultured rat
osteoblast cells (36, 37). An AP-1 enhancer motif identified in the
chicken IGF-I promoter is essential for E2 and
phorbol ester-stimulated gene transcription (31). Phorbol esters act
directly on protein kinase C to initiate a signal transduction cascade
that ultimately activates AP-1 (33). Reporter gene cotransfection
studies with expression vectors for AP-1 isoforms and ER in HeLa
cells indicate that a similar mechanism regulates the human collagenase
promoter (32). The minimal region of the collagenase promoter reported
to be responsive to tamoxifen-liganded wt ER, and to a variety of
ER mutants, harbors a critical AP-1 element and lacks a consensus
ERE. Additionally, the activity of ER on the collagenase promoter
was enhanced with AP-1 (c-jun or c-fos)
overexpression (32). Further evidence that ER regulation converges with
AP-1-directed gene transcription is provided by results from protein
binding assays indicating that c-jun is able to bind to wt
ER in vitro (32).
Although evidence for function of ER variants has been elusive,
reports that ERE5 can support weak, cell type-dependent activity
(23, 25, 28), and that, when tested on an ERE, both ERE5 and ERE3
are dominant negative receptor forms in the presence of wt ER (38, 39) indicate that it is inaccurate to label these variants as
transcriptionally inert. To investigate the capacity for ER splicing
variants to regulate gene transcription, we have expanded our
transcriptional focus to include the noncanonical ERE of the ovalbumin
promoter in addition to the consensus vitellogenin A2 ERE. Here we
present data indicating that individual variants display both
similarities and differences compared with wt ER, and that selected
splicing variants (specifically ERE3 and ERE5) have the capacity
to both positively and negatively regulate gene expression, depending
on the promoter context.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
cDNAs have now been cloned and sequenced
from breast tumors and established tumor cell lines (12, 13, 14, 15). The most
common variants harbor a precise deletion of one of the internal exons
from the eight that contribute to the structure of the mature ER
protein, suggesting that they arise as a result of imprecise splicing
of the primary ER mRNA transcript. ER cDNAs with sequence
deletions corresponding to exons 2, 3, 4, 5, and 7 have been
identified, along with a large number of more complex variants
(12, 13, 14, 15). These basic variants will be referred to as ERE2 through
ERE7, where the deleted exon is indicated numerically. Although
there is no consistent ratio of relative expression, wild-type and
variant ER transcripts are always coexpressed in ER- positive tumor
cell lines and normal and tumorous breast tissue. Quantitation of
individual variants shows that they generally represent a minority of
ER mRNA; however, as a population, splicing variants typically
constitute as much as 50% of the total ER mRNA in the tissues and cell
lines examined (Refs. 12, 13, 15 and D. P. Ankrapp and
R. J. Miksicek, unpublished observations). While there has been
extensive analysis at the RNA level of the pattern of expression and
abundance of ER splicing variants, limited information is available
on their functional activity. We have therefore constructed
cytomegalovirus (CMV) promoter-driven ER cDNA expression vectors
representing the exon-skipped variants ERE2 through ERE7 to
enable their functional characterization in a well defined cell
transfection system. Our assembled pool of ER splicing variants also
includes the hypothetical receptor ERE6, even though this variant is
not readily identified in vivo. Figure 1A
diagrams the ER mRNA splicing
variants examined, showing the positions of deleted exons and their
consequences with respect to protein structure. Deletion of exon 2, 5,
6, or 7 all cause a frame-shift mutation resulting in premature
termination of translation, thereby generating a diverse class of
C-terminally truncated receptor forms. Omission of either exon 3 or 4
does not disrupt the mRNA reading frame, but produces a receptor
protein with an internal deletion.
|
). Based on
immunoblot analysis with an N-terminal monoclonal antibody (Mab-17),
which recognizes an epitope within exon 1 common to all of the variants
(40), we observe that the mobility of the six variant proteins is
consistent with their predicted molecular weights. No immunoreactivty
is observed in mock transfected cells, confirming the specificity of
the Mab-17 antibody.
Measurement of the DNA-Binding Activity of the ER Splicing
Variants
Efficient DNA binding by ER requires the cooperation of
several functional elements within this protein, including the
centrally located DBD and a ligand-inducible subunit dimerization motif
located near the C-terminal end of the LBD (2, 18). It is also possible
that additional subunit contacts occur elsewhere in the protein.
Because all of the ER splicing variants sustained deletions within
various regions of this protein, it was of interest to systematically
assess the DNA-binding ability of each variant. For this purpose, gel
mobility shift assays were performed using extracts prepared from
E2-treated, transiently transfected Cos7 cells.
Extracts were incubated with a 32P-labeled
oligonucleotide containing a consensus ERE (AGGTCACAGTGACCT) from the
Xenopus vitellogenin A2 promoter. As expected, variants that
harbor a mutation within the DBD (ERE2 and ERE3) are completely
unable to recognize the ERE (Fig. 2, lanes 5–8). Less predictably, the loss of exons contributing to the
LBD also result in a strong defect in ERE recognition (Fig. 2, lanes
9–16). For ERE5 and ERE7, however, this appears to be a
quantitative defect in DNA binding. The addition of the monoclonal
antibody, MAb-17, to the binding reactions consistently results in the
recovery of weak DNA binding by ERE5 (Fig. 2, lane 12). Presumably,
the role of the bivalent antibody is to stabilize the interaction of
receptor subunits with their palindromic binding site, mimicking the
function of the missing dimerization motif present within the C
terminus of the LBD. These results suggest the possible existence of
cell-specific constituents that perform the same function in
vivo and may account for the variable activity of ERE5 and
related constructs in different cell types (2, 25, 28). We have also
observed the formation of a complex between ERE7 and labeled ERE
probe (Fig. 2, lane 16, and data not shown). Overall, however, the
relative weakness of DNA-binding observed in these studies raises
serious questions about the extent to which any of the variants,
including ERE5 and ERE7, are able to recognize and bind to a
consensus ERE, in vivo. Furthermore, that ERE7 binds an
ERE in vitro has little transcriptional relevance in light
of the observation that ERE7 is not translocated to the nucleus when
expressed in Cos 7 cells (see below).
|
E3, Like wt ER, Binds Ligand mRNA splicing variants to bind
hormone, we performed a saturation binding assay on whole-cell extracts
from Cos7 cells transiently transfected with wt ER or the ER
variants. Only wt ER and ERE3 were able to bind
3H-labeled E2, whereas all
of the remaining variants demonstrated no specific ligand binding (Fig. 3A). The individual deletion of exons 2,
4, 5, 6, and 7 effectively eliminates all, or a significant portion, of
the LBD (see Fig. 1A), consistent with their loss of hormone binding.
We confirmed these results using an in vivo ligand-binding
assay in which the binding of a fluorescent estrogen analog was
visualized in cells cultured on cover slips. Cos7 cells transiently
expressing the individual variants or wt receptor were treated with the
fluorescent ligand, nitrile tetrahydrochrysene (nitrile THC) (41). Only
those cultures transfected with wt ER or ERE3 were observed to
stain with this ligand. In both cases, staining was localized tightly
within the nucleus. This suggests that among the variants examined, wt
ER and ERE3 exclusively bind ligand and in both cases,
ligand-bound receptors are translocated normally to the nucleus of
expressing cells (Fig. 3B). With Scatchard analysis we compared the
affinity of ERE3 and wt ER for 3H-labeled
E2. The measured dissociation constants were 0.66
nM for wt ER and 0.79 nM
for ERE3 (Fig. 3C).
|
Splicing Variants splicing variants, including those that fail to bind ligand, Cos7
cells were transiently transfected with expression vectors encoding wt
ER or individual variants. These receptors were detected in
transfected cells by indirect immunofluorescence staining (using the
MAb-17 monoclonal antibody) and confocal microscopy. Similar to wt
ER, ERE3 and ERE5 localize to the nuclei of transfected cells,
although ERE5 showed perinuclear as well as nuclear staining (Fig. 4). These results are consistent with the
fact that both ERE3 and ERE5 retain a NLS immediately downstream
of the DBD (3, 19). Subcellular localization studies have also been
completed for the exon 2, 4, 6, and 7 deletion variants. Each of these
proteins can be readily detected in transfected cells, but they all
possess dramatic defects in nuclear targeting (Fig. 4). Nuclear
targeting of wt ER is governed, in large part, by a tripartite
karyophilic signal present within exon 4 (19). Loss of this signal is
therefore consistent with the cytoplasmic pattern of distribution of
mutants such as ERE2 and ERE4, both of which lack protein
corresponding to exon 4 sequences. Inappropriate presentation or
folding of this signal must account for the defects in nuclear
localization seen with ERE6 and ERE7, since the NLS is retained
in these variants. Based on their subcellular distribution, we would
predict that only ERE3 and ERE5, like wt ER, would have the
potential to exert nuclear effects, such as modulating gene
transcription. Furthermore, the inability of the cytoplasmic variants
ERE2, ERE4, ERE6, and ERE7 to dimerize with wt ER
predicts that their subcellular distribution will not be influenced by
coexpression with the nuclear isoforms (wt ER, ERE3, and
ERE5). For ERE4, this was confirmed with a cotransport assay
using a dimerization-competent ER (data not shown). It is worth
noting, moreover, that the existence of translationally stable,
cytoplasmic splicing variants such as these may provide an explanation
for cytoplasmic staining that is commonly observed during
immunohistochemical analysis of breast biopsy specimens to assess ER
status.
|
Splicing
Variants on the Vitellogenin ERE splicing
variants have failed to demonstrate any significant ability of variant
receptors to support gene activation through an ERE, with the possible
exception of the ERE5 variant, which is reported to display a low
level of constitutive transcriptional activity on an ERE-driven
reporter in some, but not all, cell types examined (23, 25, 28). In our
hands none of the variants were effective transcriptional activators of
an ERE-containing promoter. ERE5 repeatedly showed only modest
constitutive activity (
5% of wt ER induction) on an ERE-directed
reporter plasmid cotransfected into HeLa cells (see Fig. 7, inset).
|
. We have previously
reported that the ERE3 variant acts as a dominant negative mutant
when it is coexpressed with wt ER in HeLa cells treated with
E2 (39). In the human breast epithelial cell line
HMT-3522S1, ERE5 has also been reported to disrupt transactivation
by agonist-bound wt ER of an ERE reporter gene (38). To clarify
whether this is a function unique to these variants, we completed a
series of experiments to test whether the remaining exon-skipped ER
variants also support transcriptional inhibitory effects. When examined
in a HeLa cell cotransfection assay in which the expression of
pERE-TK-CAT was driven by E2-bound wt ER, a
5-fold molar excess of any of the splicing variants lacking exons 2, 4,
6, or 7 failed to inhibit the E2-dependent
induction of chloramphenicol acetyltransferase (CAT) gene expression by
intact receptor (data not shown). In agreement with previously
published results, the ERE3 and ERE5 variants both demonstrated a
dominant inhibitory activity at all molar ratios tested (Fig. 5). With the caveat that equal amounts of
plasmid DNA support similar levels of variant receptor expression (see
Fig. 1B), it appears that ERE3 and ERE5 are approximately
equivalent in their inhibitory activity in HeLa cells.
|
E3 and
ERE5E3 and ERE5 on wt ER
activity; specifically, we tested for a direct interaction of ERE3
and ERE5 with wt ER and steroid receptor coactivator-1e (SRC-1e).
The C terminus of wt ER and fragments of the SRC-1e protein were
expressed as fusion proteins with glutathione S-transferase
(GST) and attached to glutathione-Sepharose beads. Binding assays with
GST fused to the C terminus of wt ER (GST-AF2) and
35S-methionine labeled in vitro
translated receptor demonstrate that ligand-bound wt ER and ERE3,
but not ERE5, dimerize with the LBD of ER in solution (Fig. 6A). In experiments with GST-SRC-1e
fragments, both ERE3 and ERE5 were observed to bind to regions of
the coactivator that also bind wt ER (Fig. 6B). ERE3 and wt ER
bind the SRC-1e fragments comprising amino acids 570–780 and 989-1240
and do so only in the presence of E2. In
contrast, binding of ERE5 to the 989-1240 amino acid fragment is
constitutive (Fig. 6B).
|
E3 Is a Positive Regulator of Gene Expression on an AP-1
Reporter splicing variants
are either inactive (ERE2, ERE4, ERE6, and ERE7) or largely
inhibitory (ERE3 and ERE5) in their effect on reporter constructs
that contain a consensus ERE. Recent reports suggest that wt ER is
also able to transactivate genes whose promoters do not contain an
obvious ERE. In particular are promoters for the human collagenase,
chicken IGF-I and ovalbumin genes that are regulated by wt ER and
contain a critical AP-1 element (29, 31, 32). Mutational analysis
revealed that the DBD was not required for ER-dependent expression
of these genes. Clearly, the mechanisms of ER transcriptional
activity and DNA targeting are complicated by these reports. We propose
that to assess the transactivating potential of ER variants, the
promoter focus must be expanded to include promoters that contain
noncanonical regulatory elements. These thoughts prompted us to test
the activity of the ER variants on the ovalbumin promoter that
contains a complex hormone response element. We performed
cotransfection experiments in HeLa cells using vectors expressing wt
ER or the exon-skipped variants ERE2 through ERE7 and a CAT
reporter gene construct, pOvalb-CAT, driven by a fragment of the
ovalbumin promoter (-1342 to +7 relative to the transcription start
site) described to encompass much of the regulatory sequence of this
gene (35, 42). Results from these experiments indicate that both wt
ER and ERE3 support inducible gene expression from the ovalbumin
promoter (Fig. 7) and that all of the
remaining single exon-skipped variants are transcriptionally inactive
on this reporter construct (data not shown). For wt ER, this
corroborates previously published reports (29, 35). Maximal activity
was measured in cultures treated with both phorbol 12-myristate,
13-acetate (PMA, a phorbol ester) and E2, where a
16-fold induction was observed. Like wt ER, ERE3 reproducibly
induced this reporter, despite its lack of an intact DBD. While the
induction shown in Fig. 7 for ERE3 (averaging 9-fold) is less than
that supported by wt ER, the activity of ERE3 equaled and
occasionally exceeded that of the intact receptor in several individual
experiments, confirming that this variant can be a potent inducer of
transcription. In both cases cotreatment with PMA and
E2 is highly synergistic as
E2 treatment alone has no significant effect, and
PMA treatment alone supports only modest induction for wt ER
(2.5-fold relative to vehicle control, P < 0.001).
Tamoxifen treatment of wt ER- or ERE3-transfected cultures,
either alone or together with phorbol ester, had no significant effect
on pOvalb-CAT expression. This contrasts with the stimulatory activity
of tamoxifen observed on other AP-1 containing estrogen-responsive
reporter genes (32). In control cells transfected with an empty CMV
expression vector, treatment with PMA yielded negligible reporter gene
activity. This suggests that, in the absence of ER, activation of
endogenous AP-1 alone is not an effective inducer of transcription from
the ovalbumin promoter in these cells. To confirm that wt ER and
ERE3 cooperate with AP-1 factors to regulate transactivation of the
ovalbumin promoter, we measured pOvalb-CAT expression in HeLa cells
cotransfected with both a receptor isoform and cJun. Transcriptional
activity of wt ER and ERE3 supported by PMA and
E2 cotreatment was enhanced by cJun
overexpression. While the presence of endogenous AP-1 tended to obscure
the synergism between cJun and wt ER in this system, the combined
effects of these transcription factors were slightly more than
additive. A greater than additive activation was also observed when
cJun and ERE3 were coexpressed (Fig. 8). Exogenous cJun alone elicited only a
modest response to PMA and E2 treatment. These
observations, combined with the dual requirement for activating both
AP-1 and ER, strongly suggest that these factors are acting
cooperatively on the ovalbumin promoter.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
mRNA
splicing variants have identified two receptor isoforms that possess
the ability to modulate estrogen signaling on genes that are targeted
by the ER. Although their protein structure is significantly altered,
the ERE3 and ERE5 splicing variants retain many of the activities
attributed to the full-length receptor. Loss of exon 3 results in a
receptor protein with an internal deletion that lacks a major portion
of the DBD and therefore prevents ERE3 from binding to a consensus
ERE, as confirmed by gel mobility shift analysis. However, ERE3
retains the LBD and NLS, thereby allowing it to bind hormone with an
affinity similar to wt ER and translocate to the nucleus. The
deletion of exon 5 causes a frame-shift mutation and results in a
C-terminally truncated form of the receptor. Loss of the LBD
predictably renders ERE5 unable to bind E2.
Nonetheless, ERE5 still retains the NLS, and immunofluorescence
analysis shows nuclear staining in Cos7 cells transfected with this
variant.
Rather than serving to stimulate transcription on a consensus ERE,
results from transient transfection experiments in HeLa cells that
combine either ERE3 or ERE5 with wt ER and an ERE-driven
reporter gene indicate that these isoforms actually function to inhibit
transcriptional activation by wt ER. These observations agree with
our previous results and with those reported by others from similar
experiments using HMT-3522S1 cells (38, 39). A 70% inhibition of
transcriptional activation by E2-liganded wt
ER on an ERE-driven CAT reporter gene was observed in HeLa cells
when ERE3 and wt ER expression vectors were cotransfected at a
ratio of 5:1 (39). In the ER-negative cell line HMT-3522S1,
coexpression of an equal amount of ERE5 significantly inhibited
stimulation of an ERE reporter construct by wt ER (38). Increasing
the ratio of transfected variant to wt ER demonstrates that the
repression of wt ER by ERE3 and ERE5 is dose-related and
becomes nearly complete when the variants are present in sufficient
excess (38, 39). This observation has physiological significance in the
case of breast tumor cells that predominantly express one of these
splicing variants (12, 22). Castles et al. (22) report that
ERE5 is the major ER transcript in BT-20 and MDA MB 330 breast tumor
cell lines. In BT-20 cells the ERE5 variant comprises 68% of the ER
mRNA population while wt ER measures 8%. Studies by Erenburg
et al. (43) indicate that, while ERE3 tends to be
underrepresented in breast tumors and tumor cell lines, it typically
constitutes 50% or more of ER mRNA in both stromal fibroblasts and
epithelial cells isolated from reduction mammoplasty specimens. These
authors further demonstrated that stable overexpression of ERE3 in
MCF-7 cells to levels seen in normal mammary epithelial cells
dramatically reduced the expression and estrogen inducibility of
endogenous pS2 mRNA, as well as reducing their anchorage-independent
growth and in vivo invasiveness (43).
The dominant negative character of ERE3 and ERE5 suggests that,
like wt ER, these variants are able to interact with at least one
component of the ERE-directed transcription complex in a manner that
disrupts positive gene regulation by wt ER. Based on gel mobility
shift assay analysis, it is unlikely that transcriptional interference
by these variants involves binding to an ERE to the exclusion of wt
ER. Our DNA binding analysis indicates that ERE5 can bind only
weakly to DNA, and only when the formation of this complex is
stabilized by the addition of a bivalent antibody. The role of the
antibody in this case is presumably to substitute for the missing
dimerization interface and to tether receptor subunits together in a
form more able to interact with DNA. DNA binding by ERE7 similarly
requires the addition of antibody, but this binding is even less
efficient than binding by ERE5. Interestingly, a correlation exists
among the ER variants between their ability to translocate to the
nucleus and their transcriptional inhibitory effect on wt ER
activity in mammalian cells. As of yet, no clear function has been
established for the ERE7 variant in mammalian cells, despite an
earlier report that ERE7 is a dominant inhibitor of wt ER
function in yeast (44). This is noteworthy since a number of
quantitative studies have indicated that, as a rule, ERE7 represents
the most abundant of the ER splicing variants in breast tumors
(summarized in Ref. 15).
We have previously reported that although ERE3 is unable to bind to
an ERE itself, it can prevent wt ER from binding to DNA (39). That
ERE3 inhibits both DNA complex formation and transactivation by wt
ER suggests that the potential targets of interaction by ERE3 may
include protein-protein contacts with wt ER itself or interactions
with nuclear receptor coactivators or other receptor-associated
factors. ER function may be disrupted when ERE3, which lacks the
DBD but retains the hormone-inducible dimerization domain, forms mixed
dimers with wt ER that are inefficient at binding stably to DNA. We
are able to show that, in the presence of E2,
ERE3 (but not ERE5) can form a stable complex with the LBD of
ER fused to GST attached to glutathione-Sepharose beads. This is
consistent with a model for direct inhibition of the DNA binding
activity of full-length receptor by ERE3. Experiments using
fragments from SRC-1e fused to GST indicate that both ERE3 and
ERE5 can bind a nuclear receptor coactivator. Similar to the pattern
of wt ER interaction with SRC-1e, in vitro translated
ERE3 is able to associate in an E2-dependent
manner with two regions of the steroid coactivator SRC-1e (amino acids
570–780 and 989-1240). This agrees with previous reports that also
describe three conserved nuclear receptor-binding motifs (LXXLL) within
the 570–780 amino acid region and a distinct site for AF-1 interaction
within the 989-1240 amino acid fragment (17, 45). A site for SRC-1
interaction within ER corresponds with the AF2 domain (46), a region
that is retained in the ERE3 variant. Isoforms of SRC-1 are potent
enhancers of agonist-bound ER and are required for its full
transcriptional activity (47, 48). Transfection experiments in
E2-treated HeLa cell cultures demonstrate that
coexpression of mutants containing the C terminus of ER can
attenuate ER-dependent gene expression and that this decreased
activity can be overcome with simultaneous overexpression of the
SRC-1-related coactivator transcriptional intermediary factor 2
(TIF2) (49). These results suggest that coactivators are
limiting factors for which the receptors are competing and that
ERE3, like wt ER, is a target for SRC-1 binding.
In a surprising result from cotransfection studies using engineered
mutants of ER, maximal expression of an ERE-containing reporter gene
could be observed when SRC-1 was transfected simultaneously with
separate N- and C-terminal fragments of ER, containing the AF-1/DBD
and the LBD/AF2 regions, respectively (50). These results suggest that
separate AF1- and AF2-containing ER polypeptides can interact in a
transcriptionally productive manner, provided they are brought together
by SRC-1. Furthermore, they provide an initial indication that SRC-1
interacts separately and perhaps directly with both the AF1 and AF2
domains. More support for this notion is provided by our observations
that ERE5 binds to the SRC-1e amino acid fragment 989-1240 in
solution. These results suggests the possibility that the inhibitory
function of ERE5, which itself is relatively inefficient at binding
DNA or activating transcription through an ERE, most likely results
from competition with wt ER for interaction with SRC-1 or other
cellular factors.
The most compelling evidence that some of the ER mRNA splicing
variants may indeed be transcriptionally active is seen in transfection
experiments involving ERE3 and reporter gene constructs containing a
nonconsensus hormone-regulatory element. Recently, a novel mechanism
for mediation of an estrogen response has been reported to involve
AP-1-directed regulation of transcription by ER (29, 30, 31, 32). AP-1 and
its isoforms represent a family of nearly ubiquitous transcription
factors whose activity is crucial for the efficient expression of a
wide variety of genes. As an important downstream target for the
mitogen-activated protein kinase (MAPK)- and Jun kinase
signaling cascades, AP-1 is a central player in mediating the effects
of serum and growth factors on cellular proliferation (33, 34). A
variety of estrogen-responsive genes have been described that lack a
palindromic ERE, but instead contain one or more consensus AP-1
elements that often occur near a degenerate ERE or ERE half-site (29, 31, 32). It should be noted that these imperfect EREs may in some cases
serve dual function as cryptic AP-1 response elements whose consensus
sequence (5'-TGAG/CTCA-3') bears superficial similarity to an ERE
half-site (5'-GGTCA-3' or 5'-TGACC-3').
An important observation from the analysis of genes regulated by
noncanonical EREs is that the structure-activity requirements for
activation by ER (both for the ligand and for the receptor) are
different than those for transcriptional activation through a
palindromic ERE. Using a region of the collagenase gene promoter
(-73/+63) that lacks an ERE but harbors an essential AP-1 element,
Kushner and co-workers (32) demonstrated that a DBD-deleted mutant of
ER was significantly more effective at supporting
E2-induced reporter gene expression than wt ER
in transfected HeLa cells. Similar to collagenase gene expression,
ER-dependent activation of the chicken ovalbumin promoter, which
lacks a palindromic ERE, does not require an intact DBD (29).
Furthermore, an ERE half-site was determined to be the site for
synergistic regulation of ovalbumin gene expression by AP-1 and ER
(29). Our studies involving cotransfection of an ER splice variant
with the ovalbumin promoter construct, pOvalb-CAT, agree with these
findings. Results from Fig. 7 demonstrate that, compared with
mock-transfected HeLa control cultures, pOvalb-CAT is strongly
activated by either wt ER or ERE3. A breakdown of treatments
indicates that maximal activity of both wt ER and ERE3 clearly
requires E2 in addition to an AP-1 activator.
In vitro assays demonstrate an interaction between cJun and
the N terminus of ER fused to GST (32). Additional evidence
suggesting that ERE3 and wt ER cooperate with activated AP-1 to
maximally transactivate the ovalbumin promoter is provided by our
observation that receptor activity is enhanced by simultaneous cJun
overexpression. In our studies tamoxifen treatment had little or no
effect on the activity of wt ER or ERE3, either with or without
PMA cotreatment. This contrasts with results observed when wt ER was
cotransfected with a collagenase reporter construct in HeLa cells,
where tamoxifen supported a significant induction of reporter gene
expression (32).
Our transfection results show that, while several of the ER splicing
variants are functionally incapacitated by their deletions, two of the
variants clearly retain significant transcriptional activity. For the
ERE3 and ERE5 variants, this activity is quite complex. Both of
these variants represent stable receptor isoforms that, like the
full-length receptor, localize efficiently to the nucleus where
they can interact with the transcription apparatus. However, when
acting through a consensus ERE, these variants completely lack
(ERE3) or show only weak (ERE5) transcriptional stimulatory
activity, consistent with their poor DNA binding ability. On the
contrary, both variants serve to blunt the ability of coexpressed wt
ER to promote transcription of ERE-containing genes. At the same
time, the ability of ERE3 (and presumably also ERE5) to interact
productively with nuclear receptor coactivators or other transcription
factors gives these ER splicing variants the potential to stimulate
or otherwise modulate gene expression through nonconsensus hormone
response elements that are targeted by AP-1 motifs or other DNA-binding
sites. We have clearly shown this to be true for ERE3 and the
chicken ovalbumin promoter and believe that this is also likely to be
true for many other genes, such as those encoding collagenase,
cathepsin D, IGF-I, transforming growth factor-ß, c-fos, heat shock
protein-27, and retinoic acid receptor-, all of which lack an
obvious ERE and yet still respond to estrogen. In this respect, ER
splicing variants may actually serve to redirect transcription away
from ERE-containing genes to genes such as these that appear to be
regulated nonclassically by estrogens.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
mRNA splicing variant cDNAs were generated as
derivatives of pCMV4 (51) and pcDNA3.1 (Invitrogen, San
Diego, CA), which support high levels of receptor expression in HeLa
and Cos7 cell lines (41). Plasmids expressing ERE4, ERE5, and
ERE6 were generated using synthetic oligonucleotides to construct
the variant splice junctions within an otherwise wt ER cDNA
expression plasmid. The remaining plasmids were constructed with the
use of flanking restriction sites to shuttle cloned cDNAs (39) into the
appropriate expression vectors. Mouse cJun cDNA cloned into the pCMV2
expression vector was provided by L. McCabe (Michigan State University,
East Lansing, MI).
Cell Culture, Transfection, and CAT Assays
Cos7 and Hela cells were grown in phenol red-free DMEM
supplemented with 10% calf-serum, 5 mM HEPES (pH 7.4), 2
mM glutamine, penicillin (50 U/ml), and streptomycin (50
µg/ml). Cells were transfected by the CaPO4
method, as previously described (52). HeLa cells (2 x
10-6 cells per 100-mm dish) were transfected
with 1 µg of the indicated ER expression plasmid, 2 µg of the
cJun expression plasmid (where indicated) and 16 µg of the
estrogen-responsive reporter plasmid, pERE-TK-CAT (53) or pOvalb-CAT (a
reporter gene construct containing -1342 to +7 bp of the chicken
ovalbumin promoter relative to its transcription start site) (42). Calf
thymus DNA (10 µg) was added as carrier. After overnight incubation
with DNA, culture medium was replaced with 5% charcoal-treated
serum-supplemented DMEM containing the indicated hormones. After a 24-h
incubation, cells were harvested and CAT assays were performed as
previously described (54) using 100 µg protein. Quantification of CAT
activities was performed by phosphorimage analysis of thin layer
chromatographs (ImageQuaNT, Molecular Dynamics, Inc.,
Sunnyvale, CA). For experiments involving biochemical or cytochemical
analysis of ER variants, Cos7 cells were similarly transfected with
10 µg of the indicated expression plasmid and 10 µg of calf thymus
carrier DNA. After overnight exposure to DNA, cells were cultured for
48 h in 10% calf serum-supplemented DMEM. All experiments
involving extracts from transfected cells were normalized with respect
to protein, as measured using the method of Lowry et al.
(55). Two-way ANOVA and comparison with Student’s t test
were used to assess statistical differences between groups. Statistical
significance was set at the P < 0.05 or
P < 0.001 level as indicated in Figs. 7 and 8.
E2 Binding Analysis
Ligand-binding assays were performed as previously described
(40). Whole-cell extracts were prepared from transfected Cos7 cells
that were resuspended and sonicated in extraction buffer (20
mM HEPES, pH 7.4, 20% glycerol, 0.4 M KCl, 1
mM MgCl2) supplemented immediately
before use with protease inhibitors (0.05 mg/ml each of chymostatin,
trypsin inhibitor, antipain, leupeptin, aprotinin, and pepstatin).
Aliquots containing 200 µg of protein were incubated overnight at 4 C
with various concentrations (0.1 nM–10 nM) of
3H-labeled E2 (NEN Life Science Products, Boston, MA) in the presence or absence of
a 200-fold molar excess of unlabeled E2. Free
ligand was separated from bound ligand by treatment with dextran-coated
charcoal. For determination of equilibrium binding constants, these
data were plotted according to the method of Scatchard (56).
DNA Binding Assays
DNA binding assays were performed as previously described (40).
Aliquots containing 30 µg of protein from extracts prepared as above
from transfected Cos7 cells were preincubated for 15 min at room
temperature in 10 µl binding buffer [10 mM HEPES (pH
7.4), 1 mM MgCl2, 1 mM
dithiothreitol, and 20% glycerol] containing 1 µg poly (dI-dC),
with or without 1 µl of added human ER-specific monoclonal antibody
(Mab-17), generated as described by Neff et al. (40).
Approximately 6 fmol (40,000 cpm) of a 32P-labeled
double-stranded ERE oligonucleotide (39) were added to the samples and
incubated for 30 min at room temperature, followed by an additional
5-min incubation at 4 C. Samples were then loaded on a
preelectrophoresed nondenaturing 5% polyacrylamide gel that was run in
0.5 x Tris-Borate-EDTA at 275 V for 2 h. The gel was dried
and exposed for autoradiography.
Immunoblot Analysis
Discontinuous 12% SDS-PAGE was carried out as previously
described (57). After electrophoresis of 30 µg of whole-cell protein
from extracts of transfected Cos7 cells, proteins were
electrophoretically transferred to nitrocellulose filters with a Trans
Blot apparatus (Bio-Rad Laboratories, Inc. Richmond, CA)
using the procedure of Erickson et al. (58). Immunoblots
were probed with the ER-specific monoclonal antibody, Mab-17, obtained
from a hybridoma culture supernate that was diluted with an equal
volume of PBS (40). Immunoreactive protein was visualized by enhanced
chemiluminescence using a horseradish peroxidase-conjugated goat
antimouse IgG, following manufacturer’s instructions (Amersham Pharmacia Biotech, Arlington Heights, IL).
In Vitro Protein-Protein Interaction Assays
Variant and wt ER receptor protein was translated in the
presence of [35S]methionine using the TNT
Coupled Reticulocyte System (Promega Corp., Madison, WI).
GST-fusion proteins were expressed in the pGEX system (Pharmacia Biotech, Uppsala, Sweden) (45, 59). Overnight cultures of
transformed bacteria were diluted 1:20 and cultured for 2 h before
protein expression was induced with the addition of isopropyl
ß-D-thiogalactoside (IPTG, 0.2 mM final
concentration). Bacteria were collected by centrifugation 2 h
following IPTG induction, and pellets were resuspended in 400 µl of
extraction buffer supplemented with protease inhibitors. Cells were
sonicated briefly, and the resulting lysates were centrifuged for 20
min at 20,000 rpm, 4 C. Protein concentrations were determined (55) and
extracts were diluted to 2 µg/µl extraction buffer and stored at
-70 C until binding assays were performed.
Before use in protein interaction assays, 25 µl of glutathione-Sepharose 4B beads (Pharmacia Biotech) were washed three times in 100 µl NETN [0.5% Nonidet P-40, 1 mM EDTA, 20 mM Tris (pH 8.0) 100 mM NaCl] and suspended in 100 µl NETN, 0.5% powdered milk. Washed beads were incubated with 40 µg of GST-fusion protein for 2 h, rotating at room temperature. Beads complexed with GST-fusion proteins were washed three times with 100 µl NETN, 4 C. For protein-protein interaction assays, 5 µl of in vitro translated receptor were added to washed complexed beads resuspended in 100 µl NETN supplemented with protease inhibitors (as above) with and without 2.5 µM E2. After a 2-h incubation during which the samples were rotated at room temperature, the beads were pelleted and washed four times with 100 µl NETN, 4 C. Bound proteins were separated on a discontinuous 10% polyacrylamide SDS-PAGE gel (57). The gels were dried and exposed for autoradiography.
Immunohistochemical and Cytochemical Analysis
Indirect immunofluorescence analysis was performed as previously
described (40) using Cos7 cells that were plated and transfected on
glass cover slips. On the second day after transfection, cells
were washed three times with Tris-buffered saline (TBS), fixed for 3
min in cold 95% methanol, rehydrated by three washes with TBS, and
incubated 30 min at 37 C with primary antibody (Mab-17 hybridoma
supernate used at a 1:10 dilution in TBS). Bound antibody was detected
by staining with a rhodamine-conjugated affinity-purified goat
antimouse IgG (Roche Molecular Biochemicals, Indianapolis,
IN) diluted 1:2000 in TBS, and incubating for 30 min at 37 C in the
presence of 0.02 µg/ml of 4',6-diamidine-2-phenylindole
dihydrochloride. Confocal images were recorded using the Odyssey system
(Noran Instruments, Middleton, WI) on an Optiphot 2 Nikon
(Melville, NY) microscope. Fluorescent ligand staining of transfected
Cos7 cells was performed as described by Miksicek et al.
(41) on live, whole-cell mounts treated in DMEM with
10-7 M nitrile THC. For
these studies, cells were visualized using a Nikon UFX
microscope equipped with a 100 watt mercury lamp for fluorescence
excitation, and a 40 x 0.7 numerical aperture Plan objective.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
This work was supported by Awards DAMD1794-J-4372 and DAMD1799–1-9293 to R.J.M. and A.B. from the US Army Breast Cancer Research Program.
Received for publication February 15, 1999. Revision received January 31, 2000. Accepted for publication February 3, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
E3) in breast cancer and
the consequences of its reexpression: interference with
estrogen-stimulated properties of malignant transformation. Mol
Endocrinol 11:2004–2015This article has been cited by other articles:
 |
|
 |
|
  J. Wong and C. S. Weickert Transcriptional Interaction of an Estrogen Receptor Splice Variant and ErbB4 Suggests Convergence in Gene Susceptibility Pathways in Schizophrenia J. Biol. Chem., July 10, 2009; 284(28): 18824 - 18832. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. S. Weickert, A. L. Miranda-Angulo, J. Wong, W. R. Perlman, S. E. Ward, V. Radhakrishna, R. E. Straub, D. R. Weinberger, and J. E. Kleinman Variants in the estrogen receptor alpha gene and its mRNA contribute to risk for schizophrenia Hum. Mol. Genet., August 1, 2008; 17(15): 2293 - 2309. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. Gallagher, C. D. Langefeld, C. J. Gordon, J. K. Campbell, J. C. Mychalecky, M. Bryer-Ash, S. S. Rich, D. W. Bowden, and M. M. Sale Association of the Estrogen Receptor-{alpha} Gene With the Metabolic Syndrome and Its Component Traits in African-American Families: The Insulin Resistance Atherosclerosis Family Study Diabetes, August 1, 2007; 56(8): 2135 - 2141. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Gallagher, K. L. Keene, J. C. Mychaleckyj, C. D. Langefeld, J. N. Hirschhorn, B. E. Henderson, C. J. Gordon, B. I. Freedman, S. S. Rich, D. W. Bowden, et al. Investigation of the Estrogen Receptor-{alpha} Gene With Type 2 Diabetes and/or Nephropathy in African-American and European-American Populations Diabetes, March 1, 2007; 56(3): 675 - 684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Longo, B. Peruzzi, D. Fortunati, V. De Luca, S. Denger, G. Caselli, S. Migliaccio, and A. Teti Modulation of human estrogen receptor {alpha} F promoter by a protein kinase C/c-Src-dependent mechanism in osteoblast-like cells J. Mol. Endocrinol., December 1, 2006; 37(3): 489 - 502. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. A. Ishunina, D. F. Swaab, and D. F. Fischer Estrogen Receptor-{alpha} Splice Variants in the Medial Mamillary Nucleus of Alzheimer's Disease Patients: Identification of a Novel MB1 Isoform J. Clin. Endocrinol. Metab., June 1, 2005; 90(6): 3757 - 3765. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Veldhuis, J. N. Roemmich, E. J. Richmond, A. D. Rogol, J. C. Lovejoy, M. Sheffield-Moore, N. Mauras, and C. Y. Bowers Endocrine Control of Body Composition in Infancy, Childhood, and Puberty Endocr. Rev., February 1, 2005; 26(1): 114 - 146. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Bryant, A. E. Snowhite, L. W. Rice, and M. A. Shupnik The Estrogen Receptor (ER){alpha} Variant {Delta}5 Exhibits Dominant Positive Activity on ER-Regulated Promoters in Endometrial Carcinoma Cells Endocrinology, February 1, 2005; 146(2): 751 - 759. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. del Rio, J. M. G. Pedrero, C. Martinez-Campa, P. Zuazua, P. S. Lazo, and S. Ramos Melatonin, an Endogenous-specific Inhibitor of Estrogen Receptor {alpha} via Calmodulin J. Biol. Chem., September 10, 2004; 279(37): 38294 - 38302. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Bardin, N Boulle, G Lazennec, F Vignon, and P Pujol Loss of ER{beta} expression as a common step in estrogen-dependent tumor progression Endocr. Relat. Cancer, September 1, 2004; 11(3): 537 - 551. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Garcia Pedrero, P. Zuazua, C. Martinez-Campa, P. S. Lazo, and S. Ramos The Naturally Occurring Variant of Estrogen Receptor (ER) ER{Delta}E7 Suppresses Estrogen-Dependent Transcriptional Activation by Both Wild-Type ER{alpha} and ER{beta} Endocrinology, July 1, 2003; 144(7): 2967 - 2976. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Garcia Pedrero, B. del Rio, C. Martinez-Campa, M. Muramatsu, P. S. Lazo, and S. Ramos Calmodulin Is a Selective Modulator of Estrogen Receptors Mol. Endocrinol., May 1, 2002; 16(5): 947 - 960. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Shughrue, G. R. Askew, T. L. Dellovade, and I. Merchenthaler Estrogen-Binding Sites and Their Functional Capacity in Estrogen Receptor Double Knockout Mouse Brain Endocrinology, May 1, 2002; 143(5): 1643 - 1650. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. A. LaVoie, D. C. DeSimone, C. Gillio-Meina, and Y. Y. Hui Cloning and Characterization of Porcine Ovarian Estrogen Receptor {beta} Isoforms Biol Reprod, March 1, 2002; 66(3): 616 - 623. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Pasqualini, D. Guivarc'h, J.-V. Barnier, B. Guibert, J.-D. Vincent, and P. Vernier Differential Subcellular Distribution and Transcriptional Activity of {{Sigma}}E3, {{Sigma}}E4, and {{Sigma}}E3-4 Isoforms of the Rat Estrogen Receptor-{{alpha}} Mol. Endocrinol., June 1, 2001; 15(6): 894 - 908. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. H. Price Jr., C. A. Butler, P. Webb, R. Uht, P. Kushner, and R. J. Handa A Splice Variant of Estrogen Receptor {beta} Missing Exon 3 Displays Altered Subnuclear Localization and Capacity for Transcriptional Activation Endocrinology, May 1, 2001; 142(5): 2039 - 2049. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
