Control of vitellogenin genes expression by sequences derived from transposable elements in rainbow trout.

Abstract

In most of oviparous animals, vitellogenins (VTG) are the major egg yolk precursors. They are produced in the liver under the control of estrogens. In rainbow trout (Oncorhynchus mykiss), the vtg genes cluster contains an unusually large number of almost identical gene copies. In order to identify the regulatory elements in their promoters, we used a combination of reporter plasmids containing genomic sequences including putative estrogen response elements (EREs) and we performed transient transfection assays in MCF-7 and yeast cells. We found a functional ERE corresponding to the sequence GGGGCAnnnTAACCT (rtvtgERE), which differs from the consensus ERE (ERE(cs)) by three base pairs. This non-palindromic ERE is located in the env gene of a retrotransposon relic, 180 base pairs upstream of the transcriptional start site. Fluorescence anisotropy experiments confirmed that the purified human estrogen receptor alpha (hERalpha) can specifically bind to rtvtgERE. Furthermore, we observe that the stability of hERalpha-ERE(cs) and hERalpha-rtvtgERE complexes is similar with equilibrium dissociation constants of 3.0nM and 6.2nM respectively, under our experimental conditions. Additionally, this rtvtgERE sequence displays a high E2-responsiveness through ER activation in cellulo. In the rainbow trout, the functional ERE (rtvtgERE) lies within promoter sequences which are mostly composed of sequences derived from transposable elements (TEs), which therefore may have acted as an evolutionary buffer to secure the proper expression of these genes.

Full text

Page 1

1
Control of vitellogenin genes expression
by sequences derived from transposable elements in rainbow trout

Anthony Boutera, Nicolas Buisineb, Adélaïde Le Grand, Nathalie Mouchelc, Franck
Chesneld, Catherine Le Goffd, Véronique Le Tilly*, Jacques Wolff and Olivier Sire

Laboratoire d’Ingénierie des Matériaux de Bretagne, Université de Bretagne-Sud,
CER Yves Coppens, BP573, 56017 Vannes CEDEX, France

a present address: Molecular Imaging and NanoBioTechnology, UMR 5248 CBMN, CNRS-Université
Bordeaux 1-ENITAB, IECB, 2 rue Robert Escarpit, 33607 Pessac, France.
b present address: Museum National d'Histoire Naturelle, 57 Rue Cuvier, 75231 Paris Cedex 05,
France.
c present address: Paediatric Molecular Genetics, Institute of Molecular Medicine, Oxford University,
John Radcliffe Hospital, Oxford OX3 9DS, UK.
dCNRS UMR 606, Institut Génétique et Développement de Rennes, IFR140, Campus de médecine, 2
avenue du Pr. Léon Bernard, F-35043 Rennes CEDEX, France.

* Corresponding author:
Tel. 33 297 017 135; Fax. 33 297 017 071
E-mail address: letilly@univ-ubs.fr (V. Le Tilly)

Keywords: estrogen response element, estrogen receptor, vitellogenin, rainbow trout,
transfection assays, fluorescence anisotropy.

Abbreviations: TE, transposable element; TSS, transcriptional start site; cyc, cytochrome-c
oxidase; DO, drop-out supplement; E2, 17β-estradiol; ES, estrogens; ER, estrogen receptor;
ERE(s), estrogen response element(s); EREcs, consensus estrogen response element; hER,
inserm-00511709, version 1 - 3 Sep 2010
Author manuscript, published in "Biochimica et Biophysica Acta / Biochim biophysica Acta (Amsterdam) 2010;1799(8):546-554"
DOI : 10.1016/j.bbagrm.2010.07.003

Page 2

2
human estrogen receptor; OD, optical density; rtER, rainbow trout estrogen receptor; rtVTG,
rainbow trout vitellogenin.
inserm-00511709, version 1 - 3 Sep 2010

Page 3

3
Abstract
In most of oviparous animals, vitellogenins (VTG) are the major egg yolk precursors. They
are produced in the liver under the control of estrogens. In rainbow trout (Oncorhynchus
mykiss), the vtg genes cluster contains an unusually large number of almost identical gene
copies. In order to identify the regulatory elements in their promoters, we used a combination
of reporter plasmids containing genomic sequences including putative estrogen response
elements (EREs) and we performed transient transfection assays in MCF-7 and yeast cells.
We found a functional ERE corresponding to the sequence GGGGCAnnnTAACCT
(rtvtgERE), which differs from the consensus ERE (EREcs) by three base pairs. This non-
palindromic ERE is located in the env gene of a retrotransposon relic, 180 base pairs upstream
of the transcriptional start site. Fluorescence anisotropy experiments confirmed that the
purified human estrogen receptor α (hERα) can specifically bind to rtvtgERE. Furthermore,
we observe that the stability of hERα-EREcs and hERα-rtvtgERE complexes is similar with
equilibrium dissociation constants of 3.0 nM and 6.2nM respectively, under our experimental
conditions. Additionally, this rtvtgERE sequence displays a high E2-responsiveness through
ER activation in cellulo.
In the rainbow trout, the functional ERE (rtvtgERE) lies within promoter sequences which
are mostly composed of sequences derived from transposable elements (TEs), which therefore
may have acted as an evolutionary buffer to secure the proper expression of these genes.

inserm-00511709, version 1 - 3 Sep 2010

Page 4

4
1. Introduction
Estrogens (ES) signaling pathways are key components of many biological processes such
as differentiation, growth and embryogenesis, and are thus critical for many life traits. ES
biological activity is mediated by the estrogen receptor (ER), a member of the superfamily of
nuclear receptors. In the presence of ES, ER binds to specific DNA sequences called Estrogen
Response Elements (EREs) [1]. These short sequences are usually located in the promoter,
although some have been identified in introns or exons [2]. Recently, genome-wide analysis
found many functional ER binding sites located at large distances from the transcriptional
start site [3-4]. It is well established that ER has the highest affinity for a 15 bp sequence
composed of two 6 bp inverted repeats separated by a 3 bp spacer [2, 5]. This sequence,
AGGTCAnnnTGACCT, is designated as the consensus ERE sequence (EREcs). The EREcs
sequence is rarely found in natural promoters of ES-regulated genes; in fact, a multitude of
imperfect palindromic-like ERE sequences has been identified as functional EREs.
Additionally, by using natural and synthetic imperfect EREs, it has been shown that single
nucleotide alteration in each half-site of the ERE palindrome affects more the ER binding and
its transcriptional activity than if two mutations occur in only one half-site of the ERE [2].
Besides, the spacer size between the two half-sites also affects the binding affinity and the
conformation of the receptor in human estrogen receptor-ERE complex: human estrogen
receptors bind strongly to EREcs exhibiting no spacer or with a spacer size of 3 bp between
half-sites. In comparison, the association is much lower with a spacer size of 1 or 2 bp [6].
Vertebrate genomes usually encode two distinct ERs, ERα and ERβ, which significantly
differ in their biological activities [7]. In rainbow trout, ERα is present as two isoforms
generated by alternative splicing (rtERS and rtERL) [8]. The rtERS expression is restricted to
the liver where it is the dominant isoform, whereas the rtERL expression pattern is more
ubiquitous, suggesting a specific role of rtERS in vitellogenesis [9]. Human ERα (hERα) and
rtERS are well-conserved (92% and 60% similarities for the DNA-binding domain and the
inserm-00511709, version 1 - 3 Sep 2010

Page 5

5
ligand-dependent transactivation domain, respectively), except for the ligand-independent
transactivation domain, which is poorly conserved (20% similarity). hERα and rtERS also
exhibit important functional differences [10]: (i)- rtERS has a marked transcriptional activity
in the absence of estrogens; (ii)- rtERS needs a 10-fold higher estradiol (E2) concentration to
achieve maximal transactivation compared to hERα; (iii)- rtERS displays a weaker
transactivation activity compared to hERα, in yeast assays with a reporter gene containing
one, two or three copies of EREcs. It is important to note that the binding of E2 to rtERS, but
not to hERα, was shown to be temperature sensitive [11].
Vitellogenin (VTG) is the major precursor of egg yolk proteins which are essential for the
early development of non-mammalian vertebrates. VTG, produced by the liver of the mature
female, is mainly under the control of estrogens. Besides inducing vtg genes transcription, ES
also increase the stability of the corresponding messenger RNA [12-13]. VTG is then secreted
into the bloodstream and selectively incorporated into the growing oocytes [14]. Because of
their remarkable E2-mediated stimulation, vtg genes have for long been the model of choice
to decipher the molecular mechanisms of transcriptional regulation by ERs. Most of our
knowledge of the basic mechanisms of transcriptional regulation is based on Xenopus laevis
and chicken vtg genes [15-18]. Besides, the tilapia vtg gene promoter [19] was shown to
contain several regions exhibiting more than 70% similarity with the X. laevis vtgA2 gene
promoter. This, together with additional functional analysis [20], suggest that many features
of vtg gene expression have been conserved through evolution between teleosts and tetrapods.
VTG are usually encoded by small multigene families which mostly form a vtg gene
cluster in a conserved syntenic group [21]. The promoter structure of vtg genes is generally
quite simple [15-16, 19, 22-23], consisting of a consensus or imperfect EREs plus additional
enhancers located close to the transcriptional start site (TSS). In salmonids, two paralogous
vtg gene clusters arose from an ancestral tetraploïdization, at the base of salmonid radiation.
In the course of evolution, Oncorhynchus species have retained only one cluster [24]. In the
inserm-00511709, version 1 - 3 Sep 2010

Page 6

6
rainbow trout (Oncorhynchus mykiss), this cluster contains about 20 highly conserved genes,
and functional genes, plus 10 truncated pseudogenes for which the truncation breakpoint
corresponds to a putative retrotransposon located in intronic sequences [12]. Genes and
pseudogenes are arranged in a head-to-tail orientation, typical of tandemly arrayed genes
subject to concerted evolution [25-26]. Repeated units are separated from each other by a
highly conserved 4.6 kb intergenic region which is mostly composed of transposable elements
(TEs)-related sequences. Strikingly, these sequences are found very close to the TSS, raising
the question whether they participate functionally or interfere with vtg genes expression. They
show no sequence similarity with other known vtg promoter sequences, ie tilapia, Xenopus
and chicken.

In this paper, we describe the identification and the physicochemical characterization of a
functional imperfect ERE which drives rtvtg genes transcription. This rtvtgERE is located in
TE-related sequences that compose almost all the promoters of rtvtg gene, suggesting that
these sequences participate in the regulation of the expression of vtg genes and were co-opted
during the course of evolution. Therefore, the recent reshaping of the structure and
organization of rtvtg genes provides us with a unique opportunity to decode the interplay
between the evolution of gene structure and the regulation of gene expression.

inserm-00511709, version 1 - 3 Sep 2010

Page 7

7
2. Materials and methods

2.1. Vector construction
DNA fragments containing putative EREs were obtained by PCR from clone S5 of the
rtvtg gene [27] and cloned in a luciferase reporter plasmid (pGL2-b, Promega). MluI and
BglII restriction sites were included in primers to facilitate cloning. DNA was amplified
between positions -480 to +22 and -140 to +22, with respect to the transcription initiation site.
The PCR products were cloned into the MluI/BglII digested pGL2-b plasmid, giving the p-
480/+22 and p-140/+22 vectors. The resulting plasmids contain the first 21 nucleotides of the rtvtg
exon 1, in which the ATG (position +18) was mutated (ATA) to prevent interferences with
reporter gene. All constructs were controlled by sequencing. pGL2-b (promoterless plasmid)
and pGL2-p (SV40 promoter) were used as negative and positive controls respectively.

β-galactosidase reporter vectors are derivatives of the YRPE2 vector, in which 2 EREcs are
located upstream of a minimal cyc promoter [28]. A unique XhoI site upstream of the 2 EREcs
and a unique BamHI in the lacZ gene permitted the replacement of this region by appropriate
sequences prepared in pBK plasmid (Stratagene). The cyc-lacZ sequence was amplified from
YRPE2 with primers containing XhoI-BglII sites (forward) and BamHI site (reverse) and
subcloned into pBK at XhoI and BamHI sites (pBKcyc). Genomic fragments encompassing
the putative EREs were amplified using primers containing XhoI (forward) and BglII (reverse)
sites. These sequences were subcloned into the XhoI-BglII sites of pBKcyc and XhoI-BamHI
fragment was used to replace the corresponding fragment in YRPE2. The resulting plasmids
were designated Y-450/-130, Y-260/-130, Y-210/-130, Y-450/-190 and Y-450/-310. The same strategy was
used to construct vectors with a single or no consensus ERE (YEREcs and Ycyc, respectively).
Site mutations were introduced in Y-450/-130 and the oligonucleotides used for mutation of the -
180 bp fragment were GCTAAATGGCAGTGCCGAAggtAAACCTAACCTTTAT and
inserm-00511709, version 1 - 3 Sep 2010

Page 8

8
ATAAAGGTTAGGTTTaccTTCGGCACTGCCATTTAGC (bold represents mutated
nucleotides).
Yeast rtERS and hERα expression plasmids (pY60rtERS and pY60hERα, respectively) are
derivatives of pYeDP60-AQPcic [29] from which the cDNA encoding AQPcic was excised
by EcoRI and SacI and replaced by the complete coding sequence of the rtERS cDNA
amplified from pCMV5/rtERS [30], or the human ERα cDNA amplified from YEPE15 [28].
The plasmid has the yeast ura marker and the bacterial ampR selection markers. rtERS and
hERα expression is under the control of a GAL10-cyc promoter and can be induced by 2%
galactose.

2.2. Cell culture and transient expression assays
MCF-7 human breast cancer cells were seeded at 106 cells per 60 mm dish in phenol red
free Dubelcco's modified medium and Ham's nutrient mix F12 (DMEM Ham F12)
supplemented with 10% Fetal Calf Serum (FCS), 10 U/ml penicillin, 10 µg/ml streptomycin,
0.0025 µg/ml amphotericin, 25 mM Hepes and 4.8 mM of bicarbonate. After 24 h of
incubation at 37 °C in this medium (DMEM Ham F12 + 10 % FCS), cells were kept for 12 h
in DMEM Ham F12 plus 5 % dextran coated charcoal stripped FCS (DCC-FCS). MCF-7 cells
were then transfected by the calcium phosphate procedure using 5 µg of reporter plasmids
(pGL2-b, pGL2-p, p-140/+22 or p-480/+22). The plasmid pCH110 (Pharmacia) containing the β-
galactosidase reporter gene was used as an internal control. To improve transfection
efficiency, a glycerol shock was performed 6 h after transfection. Cells were then incubated
with DMEM Ham F12 supplemented with 1 % DCC-FCS with or without E2 (10-8 M). Cells
were harvested 36 h after transfection. Luciferase activity was measured by scintillation
counting with the luciferase assay system (Promega), and counts were normalized to levels of
β-galactosidase activity. Displayed values are the means of 3 independent experiments.

inserm-00511709, version 1 - 3 Sep 2010

Page 9

9
Yeast Saccharomyces cerevisiae cells W303.1B (α, leu2, his3, trp1, ura3, ade2-1, canR,
and cyr+) were grown in a rich YPRE medium (2% tryptone, 1% yeast extract, 0.5%
raffinose, pH 7 and 3% ethanol) or selective SD medium (0.67% nitrogen base without amino
acids, 1% raffinose or 2% glucose, pH 5.8, plus drop-out supplements: DO-Ura or DO-Ura-
Leu). Yeast cells were transformed with YRPE2-derivative reporter plasmids and/or
expression plasmids by the lithium acetate chemical method (Yeast protocols handbook,
Clontech). Transformed yeast cells were grown in the selective medium at 170 rpm and 30°C
up to
1cm=l 600nm,
OD
= 1 and diluted 10 times into the YPRE medium. They were grown again to
1cm=l600nm,
OD
= 0.6 before induction with 2% galactose, and/or hormonal stimulation with 17β-
estradiol (E2), for 16 h. Activity determined by β-galactosidase assays (Yeast protocols
handbook, Clontech) was expressed in Miller units according to:
cm
1
=
lnm
cm
1
=
l nm
ODVt
OD
Activity
, 600
,420
1000
××
×
=
(1)
with t, the time of reaction (min) at 30°C and V the culture volume used for the assay (0.05
mL). Displayed values are the means of at least 3 independent experiments. The data of the
dose-response curves have been fitted according to the Hill equation from which the
recovered value of
50
EC was extracted:
n
n
L
L EC
L
+
a
ActivityActivity
][
][
50
0][
×
+=
→
(2)

2.3. In vitro binding assays
Double stranded oligonucleotide solutions were prepared from complementary 21 bp-long
oligonucleotides synthesized (with labeling or not), purified and adjusted to 100 µM by the
Proligo company (Paris, France). The sense strand of each 21 bp oligonucleotide sequence
was labeled with fluorescein at the 5’ end. These oligonucleotide sequences are
GTCAGGTCACAGTGACCTGAT (EREcs), AGTGGGGCAGGTTAACCTAAC (pERE2),
inserm-00511709, version 1 - 3 Sep 2010

Page 10

10
AGTGCCGAAGGTAAACCTAAC (mutated pERE2), CCCACGTAAAACTGACCATCC
(pERE1) and CCATTAGACCGTTAGG (mim, a negative control corresponding to a Myb
response element). Double stranded oligonucleotides were used at a working concentration of
1 nM. Purified human ERα expressed from baculovirus was provided by PanVera
Corporation at a concentration of 3 µM. One hour before experiments, protein samples were
diluted four times in buffer B (10 mM Tris-HCl, 140 mM KCl, pH 7.5, 1 mM DTT, 0.1 mM
EDTA and 10% glycerol) and kept in ice for the duration of the experiments. The experiment
could not be carried out with rtERs because we could not produce a functional protein. We
note that functional estrogen receptors are notoriously difficult to over-express and purify.
Instead of using electro-mobility gel shift assay (EMSA), KD values have been inferred from
anisotropy fluorescence experiments. As opposed to the limited window of experimental
conditions available with EMSA (especially pH), fluorescence anisotropy has the
considerable advantage of being able probe the physical properties of protein-DNA
interaction in a wide range of experimental conditions, and thus providing us with
biologically relevant parameters.
Fluorescein fluorescence anisotropy values were monitored with a 480 nm linear polarized
excitation light (vertical or horizontal) using an SLM 8100 spectrofluorometer. Fluorescein
emission anisotropy ( A) was obtained from parallel,
// F , and perpendicular,
⊥
F , emission
components:
⊥
⊥
F
×+
−
=
F
FF
A
2
//
//
(3)
Parallel and perpendicular fluorescence intensities were monitored through a 520 nm cut-off
Oriel filter. For each displayed anisotropy value, the solvent contribution was subtracted.
Fluorescence anisotropy measurements were carried out with an integration time of 5 s on
each emission component. Each binding curve was performed at 10°C and repeated at least 3
times. The values of the dissociation equilibrium constant (KD) were inferred by nonlinear
inserm-00511709, version 1 - 3 Sep 2010

Page 11

11
least-squares curve fitting using commercial Peakfit software and by assuming that only one
equilibrium is observed within the range of protein concentration being used.

inserm-00511709, version 1 - 3 Sep 2010

Page 12

12
3. Results

3.1. The promoter of rtvtg genes is composed almost entirely of sequences derived from TEs
In the rainbow trout genome, vtg genes are arranged in tandem arrays. They are separated
by a 4.6 kb highly conserved intergenic region (Fig. 1). Sequence analysis between -1090 and
+410 of a vtg gene did not reveal blocks of regulatory elements such as those described for
the Xenopus and chicken genes [15, 18]. No obvious similarity in the organization and
sequence of the promoter region was observed between the rtvtg and other vertebrate vtg
genes. This is not surprising considering that most of the 5' region of rtvtg genes corresponds
to sequences derived from retroelements [31]: the region between positions -1380 and -290
encodes for a truncated reverse transcriptase of a LINE-type retrotransposon (50% similarity
with a jockey element) whereas the sequence from positions -1495 to -1380 displays 70%
similarities with an env coding sequence (position 41 to 69, EMBL accession number:
U56288) of a HIV-type retrovirus [31]. Additional sequence comparisons carried out with
BLASTX against NCBI's databases and REPEATMASKER against a collection of salmonid
specific TE sequences (http://lucy.ceh.uvic.ca/repeatmasker/cbr_repeatmasker.py) revealed
that the sequence between positions -191 and -107 also shares significant similarities (64%)
with the HIV env coding sequence (position 176 to 196, EMBL accession number: U56288),
strongly suggesting that the jockey-like element has been inserted into a retrovirus-like
element already present in the rtvtg promoter region. In addition, we found that the sequence
between -3987 and -2851 is related to a Salmo salar LINE-like element (SsaL2.1).
Altogether, these data show that almost all the entire intergenic region up to position -107
(which contains the promoter) unambiguously originates from TE-like sequences, leaving a
particularly short proximal sequence, putatively corresponding to the ancestral promoter. We
note that the TE dynamics and the extent of genomic rearrangement in the intergenic region
were much greater than initially reported [24, 31].
inserm-00511709, version 1 - 3 Sep 2010

Page 13

13
3.2. Putative imperfect ERE located in TE sequences within intergenic region
Sequence analysis of the rtvtg promoter revealed no EREcs but two motifs corresponding to
putative imperfect EREs (pERE1 and pERE2) within the first ~ 450 bp proximal sequences
(Fig. 1). An additional motif located, at -220, is composed of two half EREs separated by a
single nucleotide spacer and almost certainly does not correspond to a structural and
functional ERE [5-6, 32]. Compared to the consensus sequence, the two putative identified
EREs, denoted as pERE1 (ACGTAAaacTGACCA) and pERE2 (GGGGCAggtTAACCT),
exhibit 3 nucleotide differences distributed in the two half-sites. Importantly, both putative
EREs are located in sequences derived from TEs. Nonetheless, given that half EREs may act
in synergism, we cannot rule-out that they functionally participate to the regulation of rtvtg
genes expression. Therefore, we undertook in cellulo and in vitro experiments to probe the
biological activity of these putative EREs, acting alone or synergistically. To this end, we first
tested their ability to induce reporter gene expression in an E2-dependent manner. We then
characterized the physical properties of the interactions between putative EREs and ER, with
in vitro binding experiments.

3.3. A putative ERE exhibits an E2-dependent activity in MCF-7 cells
In order to assess the ability of the region containing the putative EREs to mediate E2-
dependent gene induction, the fragments, -480 to +22 and -140 to +22, were inserted in the
promoterless plasmid pGL2-b, upstream of the luciferase reporter gene. The resulting
plasmids, p-480/+22 and p-140/+22, were used to transfect breast cancer MCF-7 cells, which
naturally express hER. Luciferase activity was measured with or without E2 treatment (10-8
M). Control experiments were carried out in similar conditions with MCF-7 cells transfected
with pGL2-p vector, containing the SV40 promoter, and showed no E2-dependent activity
(Fig. 2). Cells transfected with p-140/+22, devoid of putative ERE, exhibit no significant
luciferase activity with and without E2 treatment. Luciferase activities measured with this
inserm-00511709, version 1 - 3 Sep 2010

Page 14

14
construct were similar to those of the promoterless pGL2-b vector and thus represent the
background level of the assay. Cells transformed with the construct p-480/+22, in the absence of
E2, also showed a background level of luciferase activity, whereas an increased luciferase
activity (X ~12) is observed upon E2 stimulation. This result shows that the sequence between
-480 and -140 contains a transcriptional enhancer whose activity is dependent on E2.

3.4. An imperfect ERE is located in TE-like sequences within 480 bp upstream of the
transcription start site of rtvtg
The functional characterization of the putative EREs was performed by transfection assay
in yeast expressing either hERα or rtERS. We used this simple cell system over human breast
cancer cells because it allows to study, in the same cellular context, the properties of two ERs
originating from different species (namely human -hERα- and rainbow trout -rtERS-), towards
consensus and imperfect EREs. This system also has the advantage of being appropriate to
monitor ligand-dependent transactivation mediated by both ERs [28, 33-34], without the
confounding effects of endogenous steroid factors.
Yeast cells were co-transformed with a plasmid expressing ER (pY60ER) under the control
of a galactose-inducible promoter and with a β-galactosidase reporter plasmid under the
control of one EREcs (YEREcs) or various genomic sequences located between -450 and -130
positions from rtvtg gene. Five constructs were tested: Y-450/-130 containing both pEREs, Y-450/-
190 and Y-450/-310 containing only pERE1, Y-260/-130 and Y-210/-130 containing only pERE2.
Cells containing only the reporter construct (i.e. without ER expression vector or non-induced
expression of ER) exhibited only a basal level of β-galactosidase activity (< 1 Miller unit)
with or without E2 treatment (data not shown).
In order to compare the ability of hERα or rtERS to induce reporter gene expression, cells
were co-transformed with pY60hERα or pY60rtERS together with the YEREcs plasmid. ER
expression was induced with addition of galactose and activities were measured in the
inserm-00511709, version 1 - 3 Sep 2010

Page 15

15
absence or in the presence of 10-7 M of E2 (Fig. 3). Preliminary experiments showed that this
concentration induced maximal β-galactosidase activity. In the absence of E2, only cells
expressing rtERS exhibited a significant activity. Following E2-stimulation, β-galactosidase
activity was strongly increased in hERα-expressing cells whereas rtERS-expressing cells
exhibited an activity increased by twice in comparison to without E2. This control experiment
clearly indicates that hERα transcriptional activity is strictly E2-dependent, whereas rtERS
exhibits a marked E2-independent activity when acting towards EREcs, as previously
described [8, 10].
The functional analysis of the two putative ERE-like sequences was performed in yeast
cells expressing hERα, or rtERS, and transfected with reporter plasmids (Fig. 4A and 4B),
which contain a single or a combination of the putative ERE sequences identified in silico.
Yeast cells co-transformed with the expression plasmid and a reporter plasmid containing
either no ERE (Ycyc; lane 1) or one EREcs (YEREcs; lane 7) were used as negative and positive
controls, respectively.
Cells expressing ER (hERα or rtERS) transformed with Y-450/-190 (lane 5) or Y-450/-310 (lane
6) exhibited only background level of β-galactosidase activity. This result suggests that there
is no E2-dependent regulatory sequence within positions -450 and -190 and that pERE1 plays
no direct role in ERs E2-dependent transactivation. In contrast, cells transformed with any of
the three constructs containing pERE2 (Y-450/-130, Y-260/-130 or Y-210/-130, lanes 2-4), displayed a
strong E2-induced β-galactosidase activity, which can be mediated by both hERα and rtERS.
This result indicates that the region between positions -190 and -130 is responsible for E2-
stimulation, strongly suggesting that pERE2 sequence is the active ERE.

3.5. E2 responsiveness of pERE2-containing sequence
We examined the E2 responsiveness of pERE2-containing sequence with dose-response
curves carried out with yeast cells transformed with the reporter plasmids, YEREcs (positive
inserm-00511709, version 1 - 3 Sep 2010