Molecular and Cellular Endocrinology 246 (2006) 83–90
Understanding the human estrogen receptor-alpha
using targeted mutagenesis
Debra F. Skafar a,b,∗ , Shohei Koide c
a
Department of Physiology, Wayne State University School of Medicine, 540 E. Canfield,
Detroit, MI 48201, United States
b The Barbara Ann Karmanos Cancer Institute, Detroit, MI 48201, United States
c Department of Biochemistry and Molecular Biology, University of Chicago, Chicago, IL 60637, United States
Abstract
The estrogen receptor-alpha is a wonderfully complex protein important in normal biology, breast cancer, and as a target for anti-cancer agents.
We are using the available structures of the hER␣ as well as secondary structure predictions to guide site-directed mutagenesis in order to test the
importance of specific interactions and regions in the ligand-regulated activity of the protein.
In one area of interest, we are investigating the role of the F domain in the ligand-stimulated activity of the hER␣. Results from our laboratory
and others suggest that the F domain modulates the activity of the hER␣. In order to better understand the role of the F domain in the hER␣, we have
constructed mutants within this region. Mutations within a predicted alpha-helical region alter the response of the ER to estradiol (E2), eliminate
or impair the agonist activity of 4-hydroxytamoxifen (4-OHT), and alter the ability of E2 to overcome 4-OHT’s antagonist activity. Deleting the F
domain increases the affinity of the receptor for E2; by contrast, mutating a residue in the middle of the predicted helix to a proline does not alter
the affinity for E2, but does change the binding mechanism from a positive cooperative to a noncooperative interaction. These and other results
show the F domain exhibits substantial functional complexity, and support the idea that this domain modulates the activity of the hER␣.
In a second area of interest, we are investigating the role of hydrophobic and hydrogen-bonding interactions at the start of helix 12 in the activity
of the hER␣. Leucine-536 (L536) has been proposed to participate in hydrophobic interactions that form part of a capping motif stabilizing the start
of helix 12. When mutated, the resulting receptors exhibit a reduced response, or even an inverted response, to E2 and 4-OHT on both ERE-driven
and AP-1-driven promoters. Interestingly, these mutated receptors also exhibit altered interactions with probes that recognize the agonist-bound
and 4-OHT-bound conformations of the ER␣. Thus, L536 couples the binding of ligand with the conformation of the receptor.
Overall, these results show that combining structure-based hypotheses with functional tests of the ER’s activity can identify regions and
interactions that are important in the ligand-stimulated activity of the protein.
© 2005 Elsevier Ireland Ltd. All rights reserved.
Keywords: Estrogen receptor-alpha; Mutagenesis; 4-Hydroxytamoxifen
1. Introduction
The estrogen receptor-alpha (ER␣) is a wonderfully complex
protein important in normal biology, breast cancer, and as a target
for anti-cancer agents. We want to understand the function of two
regions of the alpha form of the estrogen receptor. The first part
of this paper considers work on the extreme C-terminal region of
the receptor, the F domain. The second part considers work on
the region at the start of helix 12 in the ligand-binding domain
(LBD) of the estrogen receptor. Although the estrogen receptor
can modify the activity of cells through a number of mechanisms,
∗
Corresponding author. Tel.: +1 313 577 1550; fax: +1 313 577 5494.
E-mail address: dskafar@med.wayne.edu (D.F. Skafar).
0303-7207/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.mce.2005.12.015
we will focus on its classical, ERE-driven, mechanism of action
(Nilsson and Gustafsson, 2000).
1.1. The F domain in the human estrogen receptor-alpha
Among the members of the nuclear receptor superfamily, the
F domain is one of the most variable regions (Evans, 1988).
Although many LBD crystal structures have been determined,
little is known about the conformation of the F domain. The F
domain is generally considered devoid of well-defined structure.
Why is the F domain important? First, consider the F domain in
another member of the nuclear hormone receptor superfamily,
the orphan receptor hepatic nuclear factor-4-alpha (HNF4-␣).
HNF4␣ possesses one of the longest F domain in the nuclear
84
D.F. Skafar, S. Koide / Molecular and Cellular Endocrinology 246 (2006) 83–90
Fig. 1. The F domain in the hER. (Top) Domain structure of the ER, with the
F domain shaded. (Below) Amino acid sequences of the F domain of the hER␣
(aa 554–595) and of the hER␤ (aa 456–485). The secondary structure of each F
domain was predicted using GOR IV (Garnier et al., 1996). In the hER␣, there is
a predicted alpha-helix (aa 559–570, bold), and two predicted extended regions
(aa 580–585 and aa 593–594, bold and underlined). The two glycines near the
predicted helix (aa556 and 557, underlined) are also of interest. In the hER␤,
there is a short predicted extended region (aa 483–484, bold and underlined).
All other residues are predicted to be random coil.
receptor superfamily, having a length of 60–80 amino acids
(Sladek et al., 1999). Secondary structure predictions suggest it
may contain an alpha helix and two extended (beta-strand-like)
regions (Sladek et al., 1999; Bogan et al., 2000). A modulatory
role in the activity of the HNF4␣ is suggested by the observation that a splice variant having an additional 10 residues in the F
domain, HNF4␣2 increases the activity of the protein in transient
co-transfection assays by four-fold, and is also more responsive
to stimulation by the coactivators GRIP1 and CBP (Sladek et
al., 1999). Additional support for this is provided by the observation that the F domain in HNF4␣ modulates recruitment of
coactivators and corepressors (Ruse et al., 2002; Sladek et al.,
1999). Of perhaps even greater biological interest, a mutation
in a predicted extended region of the HNF4␣ F domain, V393I,
not only reduces the activity of the receptor by 50% in transient
transfection assays, but is also associated with the development
of maturity-onset diabetes of the young (Hani et al., 1998).
Now consider the F domain in the estrogen receptor. The F
domains of the alpha and beta forms of the estrogen receptor are
quite different in both length and in sequence (Fig. 1). In the beta
form, the F domain is approximately 30 amino acids long; in the
alpha form, the F domain is approximately 42 amino acids long
(Kumar et al., 1987; Kuiper et al., 1996; Mosselman et al., 1996).
There is less than 25% identity between the F domains of the
two isoforms, and they differ in predicted secondary structure
as well (Kumar et al., 1987; Kuiper et al., 1996; Mosselman et
al., 1996; Schwartz et al., 2002). The F domain of the ER␤ is
predicted to be primarily random coil, with two residues near
the end predicted to be extended (Schwartz et al., 2002). By
contrast, the F domain of the ER␣ is predicted to contain an
Fig. 2. Energy-minimized model of residues 551–595 of the hER␣. Minimization was carried out using biopolymer and discover as described in Schwartz et
al. (2002). The backbone ribbon and side chains are shown in gray. The residues
mutated in our published studies (Schwartz et al., 2002) are in black. The side
chain of R555 has been omitted for clarity.
alpha helix, shown in bold, and two extended regions (Schwartz
et al., 2002). Of note, the predicted secondary structure of the
F domain of the ER␣ is reminiscent of the predicted secondary
structure of F domain of the HNF4␣. Between the end of the
ER␣ LBD and the beginning of the predicted helical region are
two vicinal glycines, which is also a region of interest (Schwartz
et al., 2002).
Functionally, the F domain of the human estrogen receptoralpha is reported to inhibit dimerization, as well as the interaction
of the ER␣ with the coregulator RIP140 (Peters and Khan, 1999).
Although the F domain is not required for estrogen-stimulated
activity on ERE-driven promoters (Kumar et al., 1987; Montano
et al., 1995; Schwartz et al., 2002), Stephan Safe’s laboratory
has recently shown that it is essential for the estrogen-stimulated
activity of the ER␣ via interaction with Sp1 (Kim et al., 2003).
Perhaps most interestingly, deletion of the F domain eliminates the ability of tamoxifen to act as an agonist (Montano
et al., 1995; Schwartz et al., 2002) and alters its antagonist
activity as well (Nichols et al., 1999). Overall, these studies
of the F domain in the HNF4␣ and the estrogen receptor-␣ suggest that it is a region of substantial biological, and clinical,
import.
We used secondary structure predictions to efficiently identify regions of interest within the F domain and constructed
mutants of the estrogen receptor to test in functional assays
(Table 1, Fig. 2). Note that the secondary structure predictions
serve primarily as guides to efficiently identify regions of interest
Table 1
Sequence and secondary structure prediction for the F domain of the hER␣, and the F domain mutants used herein
The secondary structure, h = helix, e = extended, c = coil, was predicted using GOR IV (Garnier et al., 1996). Mutated residues are in bold.
D.F. Skafar, S. Koide / Molecular and Cellular Endocrinology 246 (2006) 83–90
in the ER: the results of these functional studies do not depend
on whether the predicted structure corresponds with the actual
structure.
We focused on the region of the ER␣ F domain near and
within the predicted alpha helix. When testing a prediction for
a helix, the classic experiment is to disrupt it by inserting a proline residue into the middle of the predicted helical region, as we
did by mutating glutamine-565 to proline (Q565P). Next, since
hydrogen bond formation between the side chains of residues at
the start of an alpha helix and the peptide backbone can stabilize
the helix (Aurora and Rose, 1998), we wanted to disrupt this
potential interaction, and so mutated serine-559 (S559) and glutamic acid-562 (E562) to alanines (S559A/E562A mutant). The
two vicinal glycines, G556 and G557, are also of interest. Since
glycine has the smallest side chain, comprising only a hydrogen
atom, we mutated these to bulky, hydrophobic, leucines in the
G556L/G557L mutant. Finally, mutating serine-554 to a stop
codon (S554stop) truncated the receptor and deleted the entire
F domain.
We tested the activity of the wild-type or mutated estrogen
receptors using a transient transfection assay in HeLa cells. We
use this system because although there is no standard promoter
and cell model for evaluating the effects of mutations of the estrogen receptor, when reports in the literature describe a mutated
estrogen receptor, its activity on an ERE-driven reporter in HeLa
cells is generally included.
We first examined the effect of mutating the F domain on the
ability of the estrogen receptor to stimulate transcription of an
ERE-driven promoter in response to estradiol (fold-stimulation)
(Fig. 3). As you would expect, the addition of estradiol increased
the activity of the wild-type receptor. Next, note that each mutant
was stimulated by estradiol, showing that mutation or deletion
of the F domain did not eliminate the ability of the recep-
Fig. 3. Transcription activation by F domain mutants of the ER␣. The pEREBLCAT reporter was transiently cotransfected into HeLa cells along with the
pSG5–HEGO expression vector encoding the wt or mutated hER␣, exposed to
ethanol vehicle (nh), or increasing concentrations of E2, and the resulting CAT
activity was measured as described in Schwartz et al. (2002). The results are
expressed as “fold-stimulation” and are the mean ± S.E.M. of three to seven
independent experiments, each carried out in duplicate (Schwartz et al., 2002).
85
Fig. 4. Transcription activation by F domain mutants of the ER␣. The pEREBLCAT reporter was transiently cotransfected into HeLa cells along with the
pSG5–HEGO expression vector encoding the wt or mutated hER␣, exposed to
ethanol vehicle (nh), or increasing concentrations of 4-OHT, and the resulting
CAT activity was measured as described in Schwartz et al. (2002). The results
are expressed as “fold-stimulation” and are the mean ± S.E.M. of three to seven
independent experiments, each carried out in duplicate (Schwartz et al., 2002).
tor to respond to estradiol (Schwartz et al., 2002). Of these
mutated proteins, the S559A/E562A mutant exhibited the greatest effect—its activity at the highest ligand concentrations was
substantially greater than that of the wild-type protein (Schwartz
et al., 2002). In addition, substituting a proline in the middle of
the predicted helical region may have attenuated the response
to estradiol (Schwartz et al., 2002). Thus, mutations in the F
domain altered the response of the human estrogen receptoralpha to estradiol.
We next examined the effects of mutations within the F
domain on the weak agonist activity of 4-hydroxytamoxifen,
using the same ERE-driven promoter (Fig. 4). Tamoxifen is not
solely an antagonist—it is a selective estrogen receptor modulator (SERM), and can exert weak agonist activity in some
circumstances. Tamoxifen approximately doubled the activity
of the wild-type receptor on an ERE-driven reporter (Schwartz
et al., 2002). Also, as originally shown by Montano et al., deleting the F domain (S554stop) eliminated the ability of tamoxifen
to stimulate the activity of the receptor. Mutating the vicinal
glycines to leucines (G556L/G557L) did not eliminate the agonist activity of tamoxifen (Schwartz et al., 2002). However, the
mutations directed to the predicted helical region of the protein did alter the agonist activity of tamoxifen. The Q565P
mutant lost the ability to respond to tamoxifen as an agonist, while the S559A/E562A mutant responded to 4-OHT as
an agonist only at the very highest concentration of ligand
used, if at all (Schwartz et al., 2002). Thus, not only deletion
of the entire F domain, but mutation of the predicted helical
region, greatly impaired or eliminated the agonist activity of
tamoxifen.
We next examined the effects of mutations in the F domain on
tamoxifen’s antagonist activity. In this series of experiments, we
maintained the concentration of 4-hydroxytamoxifen constant
at 100 nM, and added increasing concentrations of estradiol.
By comparing the activity of the receptor in the presence of
D.F. Skafar, S. Koide / Molecular and Cellular Endocrinology 246 (2006) 83–90
86
Table 2
The concentration of E2 calculated to overcome 50% of the inhibition by 100 nM
4-OHTa
Receptor
E2 concentration (␮M)
wt hER␣
Q565P
S559A/E562A
G556L/G557L
S554stop
0.16
320
0.32
0.16
0.008
a
Schwartz et al. (2002).
both tamoxifen and estradiol with its activity in the presence
of estradiol alone, we calculated the concentration of estradiol
that would be needed to overcome tamoxifen’s inhibition of the
activity of the receptor (Table 2) (Schwartz et al., 2002). For the
wild-type protein, 160 nM estradiol was needed to overcome
50% of the inhibitory effect of 100 nM 4-OHT (Schwartz et al.,
2002). The G556L/G557L and the S559A/E562A mutations had
little or no effect on the concentration of estradiol needed to overcome tamoxifen inhibition (Schwartz et al., 2002). However,
when the F domain was deleted, only 8 nM estradiol was required
to overcome inhibition by tamoxifen—1/20 of the amount of
estradiol needed by the wild-type protein to overcome tamoxifen’s inhibitory activity (Schwartz et al., 2002). By contrast, we
calculated it would take substantially more estradiol—320 ␮M
to overcome the inhibition by tamoxifen in the Q565P mutant
(Schwartz et al., 2002). (Note that this value is obtained by
extrapolation, and would be unlikely to be achieved in aqueous
solution.) Thus, mutations in the F domain increased (S554stop)
or decreased (Q565P) the ability of estradiol to overcome inhibition by tamoxifen.
Key aspects of the ability of the estrogen receptor to respond
to ligands are the affinity of binding, as well as the mechanism of interaction between the ligand and the receptor. A
non-cooperative interaction is characterized by a Hill coefficient
near 1, while a positive cooperative interaction is characterized
by a Hill coefficient greater than 1 (Hill, 1910). We measured
the binding of [3 H] estradiol to baculovirus-expressed wt ER,
the Q565P mutant, and the S554stop mutant in an equilibrium
binding assay (Fig. 5, Table 3). As we and others have reported,
the wild-type receptor bound estradiol with high affinity, having a Kd less than 1 nM, and a positive cooperative binding
mechanism, as shown by a Hill coefficient of approximately
1.6 (Notides et al., 1981; Obourn et al., 1993; Schwartz et
al., 2002; Yudt et al., 1999). A positive cooperative binding
mechanism means that binding of the first molecule of ligand
facilitates binding of the next molecule, and so the Hill coefficient is a measure of site–site interactions within the receptor
(Hill, 1910). The S554stop mutant did not exhibit an altered
Hill coefficient of binding, which indicates that the site–site
interactions of the receptor and the positive cooperative binding
mechanism between the ligand and the receptor was unaltered
by deletion of the F domain (Schwartz et al., 2002). However,
the affinity of the F-domain-deleted receptor for estradiol was
increased, to a Kd of 0.05 nM (Schwartz et al., 2002). Substitution of a proline in the middle of the predicted helical region
did not affect the afffinity for estradiol – the Kd of the Q565P
mutant was ∼0.2 nM – but reduced the Hill coefficient to a
value near 1, which indicates that the binding mechanism had
been converted from a positive cooperative to a non-cooperative
interaction (Schwartz et al., 2002). These results show that
mutations in the F domain altered the affinity of the receptor for estradiol, as well as the site–site interactions of the
receptor.
Taken together, our results show that mutations in the F
domain altered the responses of the ER to estradiol and 4hydroxytamoxifen, as well as the affinity for estradiol and the
site-site interactions of the receptor. This raises an important
question: how might this be? In other words, how could a region
of only 42 amino acids modulate so many different activities of
the receptor?
To address this question, consider a model of the estrogen
receptor alpha F domain based on secondary structure predictions (Fig. 2) (Schwartz et al., 2002). In this model, there is
a predicted alpha-helical region, a predicted extended region,
and two additional predicted extended residues almost at the
extreme C-terminus of the protein. Of note, the larger predicted extended region contains residues – isoleucine and threonine – that are characteristic of surface beta-strands (Palliser
et al., 2000; Palliser and Parry, 2001). Most importantly, this
model shows that although the F domain contains only about 42
residues, it is potentially 100 Å or more of peptide.
Table 3
The binding of [3 H]estradiol to the wild-type, S554stop, and Q565P mutant
hER␣
Receptor
Hill coefficient, nH
Affinity, Kd (nM)
wt hER␣
S554stop
Q565P
1.58 ± 0.18
1.65 ± 0.36
0.94 ± 0.1
0.64 ± 0.41
0.05 ± 0.007
0.23 ± 0.01
Estradiol-binding was measured using baculovirus-expressed S554stop, Q565P,
and wt ER receptors (Schwartz et al., 2002). The values of the affinity (Kd ) and
Hill coefficient (nH ) were calculated by fitting the untransformed binding data
to the Hill equation by nonlinear regression using GraphPad Prism. These data
are the mean ± S.E.M. of two to three independent experiments (Schwartz et al.,
2002).
Fig. 5. Binding of the wt and mutant hER␣ to [3 H] estradiol. The binding of
[3 H] estradiol to wt hER␣ (filled squares), the S554stop mutant (open circles)
or the Q565P mutant (filled triangles) was measured in Sf9 insect cell extracts
containing baculovirus-expressed receptor (Schwartz et al., 2002). (Left) Nontransformed saturation binding data; the lines shown are the best fit by nonlinear
regression to the Hill equation using GraphPad Prism. (Right) Scatchard plot of
the same data (Scatchard, 1949). Similar data are in Schwartz et al. (2002).
D.F. Skafar, S. Koide / Molecular and Cellular Endocrinology 246 (2006) 83–90
Fig. 6. Plausible pathways for the start of the F domain in the hER␣ LBD. The
peptide backbone of the DES-bound ER LBD dimer (3ERD, Shiau et al., 1998)
in complex with a peptide derived from the coactivator GRIP1 is represented as
a ribbon, with different subunits in gray and cyan. The van der Waals surface
of DES is in yellow; the GRIP1 peptide is in blue, and the residues comprising
helix 12 are in orange. The arrows denote plausible directions for the start of the
F domain. (For interpretation of the references to colour in this figure legend,
the reader is referred to the web version of the article.)
Now consider the size of the ligand-binding domain of
the estrogen receptor-alpha. The LBD monomer is approximately 50 Å × 40 Å × 30 Å, and has different surfaces involved
in dimerization and coactivator binding, as well as surfaces that
are potentially involved in ligand association and dissociation
Fig. 7. L536 in the DES-bound (left) and 4OHT-bound (right) hER␣. (Left) An
energy-minimized model of built on the crystallographic structure of the DESbound, wt hER␣ LBD (3ERD, Shiau et al., 1998). (Right) The crystallographic
structure of the wt hER␣ LBD bound with the SERM 4-OHT (3ERT, Shiau et al.,
1998). The peptide backbone is shown as a ribbon in gray, with the exception that
the residues in helix 12 (left, 537–548; right, 537–551) are in cyan. The ligands
(left, DES; right, 4-OHT) are depicted as Connelly surfaces in yellow. The van
der Waals surface of L536 is in blue; the van der Waals surface of L541 is in
magenta. The Connelly surface of a crystallographic water is shown on the left in
orange. Republished with permission of the American Society for Biochemistry
and Molecular Biology, from Zhao et al. (2003); permission conveyed through
Copyright Clearance Center. (For interpretation of the references to colour in
this figure legend, the reader is referred to the web version of the article.)
87
Fig. 8. Response of the wt and mutant hER␣ to E2 on an ERE-driven promoter.
Transcription activation was measured using a transient cotransfection assay
with wt or mutant ER␣ (HEGO or mutated ER in pSG5), an ERE-driven reporter
(p2ERE-luciferase), and a Renilla luciferase transfection control (pRL-SV40)
in HeLa cells (Zhao et al., 2003). The activity of the ER is measured by the ratio
of firefly luciferase activity to Renilla luciferase activity (RLU). The activity
was measured in the absence of hormone (vehicle control) and in response
to increasing concentrations of E2 (10−10 , 10−9 , 10−8 , and 10−7 M E2). The
values are the mean ± S.E.M. of three to four independent experiments, each
carried out in triplicate (Zhao et al., 2003). Republished with permission of the
American Society for Biochemistry and Molecular Biology, from Zhao et al.
(2003); permission conveyed through Copyright Clearance Center.
(Brzozowski et al., 1997; Shiau et al., 1998). Because the peptide chain is continuous and the F domain immediately follows
the ligand-binding domain, the F domain has to start immediately after helix 12. We suggest that the F domain is highly
dynamic and extended, so it can easily wrap around the surface of
the estrogen receptor ligand-binding domain. There are at least
three plausible ways of starting to wrap the F domain around the
ER LBD—under helix 12, over to the other monomer, or up the
dimer interface (Fig. 6). Because of the length of the F domain,
Fig. 9. Response of the wt and mutant hER to 4OHT and ICI-182,780 on an EREdriven promoter. The activity of the ER in the absence of ligand and in response
to 4-OHT (10−7 M) and ICI-182,780 (10−6 M) was measured using transiently
transfected HeLa cells as described in Fig. 6. The values are the mean ± S.E.M.
of three independent experiments, each carried out in triplicate (Zhao et al.,
2003). Republished with permission of the American Society for Biochemistry
and Molecular Biology, from Zhao et al. (2003); permission conveyed through
Copyright Clearance Center.
88
D.F. Skafar, S. Koide / Molecular and Cellular Endocrinology 246 (2006) 83–90
different parts of the F domain could interfere with or help to
form different surfaces on the ligand-binding domain. In this
way, the F domain could modulate multiple activities of the
receptor.
1.2. What interactions determine the conformation of the
human estrogen receptor-alpha?
One of the fundamental properties of the estrogen receptor
is to change shape, or conformation, in response to the binding of ligand. Most notably, the position of helix 12 changes,
depending on the ligand that is bound—it covers the ligandbinding pocket in the presence of a strong agonist, DES, and it
blocks the coactivator binding site in the presence of a SERM, 4hydroxytamoxifen (Brzozowski et al., 1997; Shiau et al., 1998).
We want to understand what determines the conformation of the
estrogen receptor. Consider the region at the start of helix 12: the
residue in blue, leucine-536 (L536), occupies different positions,
and makes totally different contacts in the two different structures of the receptor (Fig. 7). In the agonist-bound receptor, it
appears to be involved in a hydrophobic interaction with leucine541 (L541) in magenta; in the tamoxifen-bound receptor, the
interaction between the two residues is completely disrupted
(Shiau et al., 1998). We therefore wanted to understand the role
of this residue in the conformation changes of the protein.
We mutated L536 to a number of different residues: to alanine
(A), which has a small side chain, to glutamic acid (E), which
has a negatively charged side chain, to glycine (G), which has
only a hydrogen atom for a side chain, to isoleucine (I), which
is similar to leucine, to lysine (K), which is large and positively
charged, and to asparagine (N), which is small and polar. We
evaluated the activity of the mutants in response to estradiol on
an ERE-driven luciferase reporter (Fig. 8) (Zhao et al., 2003).
As expected, estradiol stimulated the activity of the wildtype receptor. All of the mutated receptors except the isoleucine
substitution exhibited increased basal activity of the receptor
(Zhao et al., 2003). The glycine and asparagine mutants exhibited the highest basal activity (Zhao et al., 2003). However, of the
mutated receptors, only the substitution with isoleucine retained
the ability to be stimulated by estradiol (Zhao et al., 2003). Thus,
mutations of L536 eliminate the ability of estradiol to stimulate
the activity of the receptor on an ERE-driven promoter.
We next investigated the effect of these mutations on the
weak agonist activity of 4-hydroxytamoxifen (Fig. 9). Tamoxifen stimulates the activity of the wild-type receptor; the antagonist ICI-182,780 has no effect by itself (Zhao et al., 2003). Again,
the mutated receptors exhibited an increased basal activity (Zhao
et al., 2003). Most interestingly, tamoxifen either had no effect
on the activity of a mutated receptor, or decreased its basal
activity—that is, tamoxifen exhibited an “inverse agonist” effect.
Fig. 10. Interactions of hER␣ mutants with hER/agonist complex-specific probes as measured using a yeast two-hybrid system. ␤-Galactosidase activity (Miller
Units, Y-axis) was measured in the presence of 1 ␮M of the indicated ligand, from left to right: E2 (E), ICI-182,780 (I), 4-hydroxytamoxifen (T), raloxifene (R),
vehicle control (V) or progesterone (P). Column A, results obtained using the receptor-interacting domain of SRC-1 (aa190–400); column B, results obtained using
monobody “E3#6”. The data are from triplicate measurements. Reactivity with the wt (top panels) and L536E mutant (lower panels) is shown. Different panels
represent experiments using different yeast cells and thus the absolute activity cannot be normalized (or quantitatively compared) across panels (Zhao et al., 2003)
Republished with permission of the American Society for Biochemistry and Molecular Biology, from Zhao et al. (2003); permission conveyed through Copyright
Clearance Center.
D.F. Skafar, S. Koide / Molecular and Cellular Endocrinology 246 (2006) 83–90
The antagonist ICI-182,780 decreased the activity of all of the
mutated receptors (Zhao et al., 2003). Thus, mutations at L536
also eliminate the ability of 4-hydroxytamoxifen to act as an
agonist on an ERE-driven promoter.
In order to understand how these mutations affected the ability of the estrogen receptor to interact with coactivator proteins,
we used a two-hybrid system. In this system, interaction between
the two proteins of interest drives expression of a reporter gene.
We use a two-hybrid system in yeast to investigate the interaction
between estrogen receptor and the coactivator protein SRC-1
(Koide et al., 2002). In addition, probes based on a fibronectin
backbone, called “monobodies”, are used that specifically recognize the agonist-bound conformation of the receptor; these are
designated E3#6 and E2#23 (Koide et al., 2002). For the sake
of simplicity, focus on the activity of the wild-type protein and
one mutant, the L536E mutant (Fig. 10).
Each probe, SRC-1, E3#6 and E2#23, reacts with the
wild-type receptor in the presence of estradiol, but not in
the absence of ligand, or in the presence of ICI-182,780, 4-
89
hydroxytamoxifen, raloxifene, or progesterone (Koide et al.,
2002; Zhao et al., 2003). When leucine-536 is mutated, two
of the probes, SRC-1 and E3#6, show increased reactivity
with the receptor in the absence of ligand that is blocked by
4-hydroxytamoxifen (Zhao et al., 2003). Similar results are
observed with the other mutants, with the exception of the
isoleucine substitution (Zhao et al., 2003). Thus, mutations at
leucine-536 increase the basal reactivity with two probes that
specifically recognize the agonist-bound conformation of the
receptor.
Monobody probes that recognize the tamoxifen-bound conformation of the receptor have also been developed (Koide et
al., 2002). Again, focus on the wild-type receptor and the glutamic acid mutant (Fig. 11). Each probe – OHT#1 and OHT#33
– reacts with the wild-type receptor in the presence of tamoxifen, but exhibits little reactivity in the absence of ligand or in the
presence of estradiol, ICI-182,780, raloxifene, and progesterone
(Zhao et al., 2003). Mutation of leucine-536 to glutamic acid
substantially increases the reactivity of the receptor with OHT#1
Fig. 11. Interactions of hER␣ mutants with hER/OHT complex-specific probes as measured using a yeast two-hybrid system. ␤-Galactosidase activity (Miller Units,
Y-axis) was measured in the presence of 1 ␮M of the indicated ligand, from left to right: E2 (E), ICI-182,780 (I), 4-hydroxytamoxifen (T), raloxifene (R), vehicle
control (V), or progesterone (P). Column A, results obtained using monobody “OHT#1”, and column B, results obtained with monobody “OHT#33”. Data are from
triplicate measurements. Reactivity with the wt ER (top panel) and L536E (lower panels) is shown. Different panels represent experiments using different yeast cells
and thus the absolute activity cannot be normalized (or quantitatively compared) across panels (Zhao et al., 2003). Republished with permission of the American
Society for Biochemistry and Molecular Biology, from Zhao et al. (2003); permission conveyed through Copyright Clearance Center.
90
D.F. Skafar, S. Koide / Molecular and Cellular Endocrinology 246 (2006) 83–90
in the presence of raloxifene, but not in the absence of ligand. The
mutated receptor also interacts with OHT#1 in the presence of
tamoxifen. All mutants, except the isoleucine substitution, show
increased reactivity with this probe in the presence of raloxifene
(Zhao et al., 2003). This means the raloxifene-bound, mutated
receptor, resembles the tamoxifen-bound, wild-type receptor, in
the part of the protein that is recognized by this probe.
When reactivity with the other probe, OHT#33 is examined, the glutamic acid mutant shows increased reactivity in
the presence of ICI-182,780, but not in the absence of ligand
(Fig. 11) (Zhao et al., 2003). The mutated receptor also interacts
with OHT#33 in the presence of tamoxifen (Zhao et al., 2003).
All mutants, with the exception of the isoleucine substitution,
show increased reactivity with this probe in the presence of ICI182,780. This means the ICI-182,780-bound, mutated receptor,
resembles the tamoxifen-bound, wild-type receptor, in the part
of the protein that is recognized by this probe (Zhao et al., 2003).
Thus, mutations of leucine-536 increase the interaction of the
raloxifene-bound and ICI-182,780-bound receptor with probes
that recognize the tamoxifen-bound conformation of the protein.
What do all these results mean? They show that leucine-536
is critical for coupling the binding of agonist and antagonist
ligands to the conformation of the receptor. In other words, this
residue is a tie rod (Zhao et al., 2003). (A tie rod is part of a
car’s steering linkage, which connects the steering wheel with
the driving wheels.) These results also show that leucine-536
essentially “reads” the side chain of tamoxifen, raloxifene, and
ICI-182,780, and so distinguishes the conformations of these
SERM-bound or antiestrogen-bound ligand-binding domains.
Acknowledgements
We thank the members of our laboratories who have contributed so much over the years. We also thank the agencies who
have provided support for our research over the years, including
the National Institutes of Health, the National Science Foundation, and the Department of Defense Breast Cancer Research
Program, and internal funds from the Karmanos Cancer Institute, the Environmental Health Sciences Center, and the Wayne
State University School of Medicine.
References
Aurora, R., Rose, G.D., 1998. Helix capping. Protein Sci. 7, 21–38.
Bogan, A.A., Dallas-Yang, Q., Ruse, M.D., Maeda, Y., Jiang, G., Nepomuceno, L., Scanlan, T.S., Cohen, F.E., Sladek, F.M., 2000. Analysis of
protein dimerization and ligand binding of orphan receptor HNF4␣. J.
Mol. Biol. 302, 831–851.
Brzozowski, A.M., Pike, A.C.W., Dauter, Z., Hubbard, R.E., Bonn, T.,
Engstrom, O., Ohman, L., Greene, G.L., Gustafsson, J.-A., Carlquist,
M., 1997. Molecular basis of agonism and antagonism in the oestrogen
receptor. Nature (London) 390, 753–758.
Evans, R.M., 1988. The steroid and thyroid hormone receptor superfamily.
Science 240, 889–895.
Garnier, J., Gibrat, J.-F., Robson, B., 1996. Meth. Enzymol. 266, 540–
553.
Hani, E.H., Suaud, L., Boutin, P., Chevre, J.-C., Durand, E., Philippi, A.,
Demenais, F., Vionnet, N., Furuta, H., Velhi, G., Gell, G.I., Laine, B.,
Froguel, A., 1998. A missense mutation in hepatocyte nuclear factor-4,
resulting in a reduced transactivation activity, in human late-onset noninsulin-dependent diabetes mellitus. J. Clin. Invest. 101, 521–526.
Hill, A.V., 1910. J. Physiol. (London) 40, iv–vii.
Kim, K., Thu, N., Saville, B., Safe, S., 2003. Domains of estrogen receptor
(ER) required for ER/Sp1-mediated activation of GC-rich promoters by
estrogens and antiestrogens in breast cancer cells. Mol. Endocrinol. 17,
804–817.
Koide, A., Abbatiello, S., Rothgery, L., Koide, S., 2002. Probing protein
conformational changes in living cells by using designer binding proteins:
application to the estrogen receptor. Proc. Natl. Acad. Sci. U.S.A. 99,
1253–1258.
Kuiper, G.G., Enmark, E., Pelto-Huikko, M., Nilsson, S., Gustafsson, J.-A.,
1996. Cloning of a novel receptor expressed in rat prostate and ovary.
Proc. Natl. Acad. Sci. U.S.A. 93, 5925–5930.
Kumar, V., Green, S., Stack, G., Berry, M., Jin, J.R., Chambon, P., 1987.
Functional domains of the human estrogen receptor. Cell 51, 941–951.
Montano, M.M., Muller, V., Trobaugh, A., Katzenellenbogen, B.S., 1995.
The carboxy-terminal F domain of the human estrogen receptor: role
in the transcriptional activity of the receptor and the effectiveness of
antiestrogens as estrogen agonists. Mol. Endocrinol. 9, 814–825.
Mosselman, S., Polman, J., Dijkema, R., 1996. ER␤ identification and characterization of a novel human estrogen receptor. FEBS Lett. 392, 49–53.
Nichols, M., Rientjes, J.M., Stewart, A.F., 1999. Different positioning of the
ligand-binding domain helix 12 and the F domain of the estrogen receptor accounts for functional differences between agonists and antagonists.
EMBO J. 17, 765–773.
Nilsson, S., Gustafsson, J.-A., 2000. Estrogen receptor transcription and
transactivation. Basic aspects of estrogen action. Breast Cancer Res. 2,
335–344.
Notides, A.C., Lerner, N., Hamilton, D.E., 1981. Positive cooperativity of the
estrogen receptor. Proc. Natl. Acad. Sci. U.S.A. 78, 4926–4930.
Obourn, J.D., Koszewski, N.J., Notides, A.C., 1993. Hormone- and DNAbinding mechanisms of the recombinant human estrogen receptor. Biochemistry 32, 6229–6236.
Palliser, C.C., MacArthur, M.W., Parry, D.A., 2000. Surface beta-strands in
proteins: identification using an hydropathy technique. J. Struct. Biol.
132, 63–71.
Palliser, C.C., Parry, D.A., 2001. Quantitative comparison of the ability of
hydropathy scales to recognize surface beta-strands in proteins. Proteins
42, 243–255.
Peters, G.A., Khan, S.A., 1999. Estrogen receptor domains E and F: role in
dimerization and interaction with coactivator RIP-140. Mol. Endocrinol.
13, 286–296.
Ruse, M.D., Privalsky, M.L., Sladek, F.M., 2002. Competitive cofactor
recruitment by orphan receptor hepatocyte nuclear factor 4␣1: modulation
by the F domain. Mol. Cell. Biol. 22, 1626–1638.
Scatchard, G., 1949. Ann. NY Acad. Sci. 51, 660–672.
Schwartz, J.A., Zhong, L., Deighton-Collins, S., Zhao, C., Skafar, D.F., 2002.
Mutations targeted to a predicted helix in the extreme carboxy-terminal
region of the human estrogen receptor-alpha alter its response to estradiol
and 4-hydroxytamoxifen. J. Biol. Chem. 277, 13202–13209.
Shiau, A.K., Barstad, D., Loria, P.M., Cheng, L., Kushner, P.J., Agard, D.A.,
Greene, G.L., 1998. The structural basis of estrogen receptor/coactivator
recognition and the antagonism of this interaction by tamoxifen. Cell 95,
927–937.
Sladek, F.M., Ruse, M.D., Nepomuceno, L., Huang, S.-H., Stallcup, M.R.,
1999. Modulation of transcriptional activation and coactivator interaction
by a splicing variation in the F domain of nuclear receptor hepatocyte
nuclear factor 4␣1. Mol. Cell. Biol. 19, 6509–6522.
Yudt, M.R., Vorojeikina, D., Zhong, L., Skafar, D.F., Sasson, S., Gasiewicz,
T.A., Notides, A.C., 1999. The function of estrogen receptor tyrosine 537
in hormone-binding. DNA-binding and transactivation. Biochemistry 38,
14146–14157.
Zhao, C., Koide, A., Abrams, J., Deighton-Collins, S., Martinez, A.,
Schwartz, J.A., Koide, S., Skafar, D.F., 2003. Mutation of Leu-536 in
human estrogen receptor-␣ alters the coupling between ligand binding,
transcription activation, and receptor conformation. J. Biol. Chem. 278,
27278–27286.