This report investigates the integration of signaling of inverteb-rate juvenile hormone JH and 20-OH ecdysone 20OHE at the level of identified nuclear receptors ultraspiracle and ecdysone
Trang 1transduction of ecdysone- and juvenile hormone-signaling Fang Fang2, Yong Xu1, Davy Jones2 and Grace Jones1
1 Department of Biology, University of Kentucky, Lexington, KY, USA
2 The Graduate Center for Toxicology, University of Kentucky, Lexington, KY, USA
Keywords
ecdysone receptor; juvenile hormone;
methyl epoxyfarnesoate; retinoid-x-receptor;
ultraspiracle
Correspondence
G Jones, Department of Biology, University
of Kentuckey, 304 Morgan Building,
Lexington, KY 40506, USA
Fax: +1 859 257 7505
Tel: +1 859 257 2105
E-mail: gjones@pop.uky.edu
(Received 14 September 2004, revised 13
January 2005, accepted 21 January 2005)
doi:10.1111/j.1742-4658.2005.04578.x
Analyses of integration of two-hormone signaling through the vertebrate nuclear hormone receptors, for which the retinoid X receptor is one part-ner, have generated a number of mechanistic models, including those des-cribed as ‘subordination’ models wherein ligand-activation of one partner
is subordinate to the liganded state of the other partner However, mecha-nisms by which two-hormone signaling is integrated through invertebrate nuclear hormone-binding receptors has not been heretofore experimentally elucidated This report investigates the integration of signaling of inverteb-rate juvenile hormone (JH) and 20-OH ecdysone (20OHE) at the level of identified nuclear receptors (ultraspiracle and ecdysone receptor), which transcriptionally activate a defined model core promoter (JH esterase gene), through specified hormone response elements (DR1 and IR1) Application
of JH III, or 20OHE, to cultured Sf9 cells transfected with a DR1JHE-CoreLuciferase(or IR1JHECoreLuciferase) reporter promoter each induced expression of the reporter Cotreatment of transfected cells with both hor-mones yielded a greater than additive effect on transcription, for especially the IR1JHECoreLuciferase reporter Overexpression in Sf9 cells of recom-binant Drosophila melanogaster ultraspiracle (dUSP) fostered formation of dUSP oligomer (potentially homodimer), as measured by coimmunopreci-pitation assay and electrophoretic mobility assay (EMSA) on a DR1 probe, and also increased the level of transcription in response to JH III, but did not increase the transcriptional response to either 20OHE treatment alone
or to the two hormones together Inapposite, overexpression of recombin-ant D melanogaster ecdysone receptor (dEcR) in the transfected cells gen-erated dUSP⁄ dEcR heterodimer [as measured by EMSA (supershift) on a DR1 probe] and increased the transcriptional response to 20OHE-alone treatment, but did not increase the transcriptional response to the JH III-alone treatment Our studies provide evidence that in this model system,
JH III-activation of the reporter promoter is through USP oligomer (homodimer) that does not contain EcR, while the 20OHE-activation is through the USP⁄ EcR heterodimer These results also show that the integ-ration of JH III and 20OHE signaling is through the USP⁄ EcR hetero-dimer, but that when the EcR partner is unliganded, the USP partner in this system is unable to transduce the JH III-activation
Abbreviations
20OHE, 20-OH ecdysone; DR1 and IR1, direct repeat and inverted repeat with one intervening base between repeats, respectively; EcR, ecdysone receptor; EMSA, electrophoretic mobility shift assay; JH, juvenile hormone; JHECore, core promoter from juvenile hormone esterase gene; RA, retinoic acid; RXR, retinoid-X-receptor; USP, ultraspiracle.
Trang 2Nuclear hormone receptors play key roles in metazoan
development, metabolic homeostasis, and response to
xenobiotics Some of these nuclear receptors bind
lig-ands that modulate the regulatory effect of these
recep-tors on gene transcription, while other receprecep-tors are
apparently constituitively active and unregulated by
dynamic equilibrium with ligand [1] For example, the
retinoic acid receptor (RAR) responds to signaling by
all-trans retinoic acid (at-RA) and 9-cis RA, in
regula-tion of fetal limb development [2] The receptor FXR
binds to catabolites in regulation of sterol pathways,
the receptor PXR is activated by binding to certain
xenobiotic compounds, while transcription-enhancing
activity of the receptor CAR is suppressed by binding
to xenobiotic compounds [3]
Each of the above ligand-binding receptors and their
close relatives functions as a heterodimer with the
(also homodimer-forming) retinoid-X-receptor (RXR),
which itself can be activated by the RAR ligand 9-cis
RA [4] The occurrence of such multiple partner
recep-tor complexes, and their corresponding multiple
lig-ands, raises questions about the integration of multiple
ligand signaling through the same receptor complex
There is some controversy in the nuclear receptor field
as to whether RXR, the vertebrate ortholog of USP,
can independently bind ligand when RXR is in
com-plex with certain nuclear receptors For example, Kersten
et al [5] detected independent binding of 9-cis RA by
RXR when in complex with RAR However, data
of Thompson et al [6] suggested that ligand-induced
transcriptional activation by RXR is ‘subordinate’ to
whether its RAR partner has bound ligand There has
also been a question as to whether the allosteric effect
of RAR ligand onto RXR occurs by permitting RXR
to bind ligand or by allowing liganded RXR to
disso-ciate corepressors and recruit coactivators [7,8] For
other receptor partners of RXR, such as ligandless
NGFI-B, it is argued that the dynamic equilibrium is
more in favor of RXR binding coactivator in response
to ligand even if its heterodimer partner (e.g NGFI-B)
is not liganded [9] Finally, Germain et al [10] propose
that the ability of liganded RXR to bind p160-family
coactivators when in complex with apoRAR is a
func-tion of the endogenous titer of the coactivator
There also appears to be variation among
RXR-partners as to whether ligand binding by the partner is
additive to the effect of RXR binding of 9-cis RA [11],
or is synergistic [8,12], or even otherwise allosteric
in affecting RXR ligand-activated function [13] For
example, a synthetic ligand of RXR exerts an allosteric
effect on the RAR partner such that the unliganded
RAR partner adopts an activated conformation [14]
In the opposite direction, there is evidence that agonist
ligand for RXR can allosterically antagonize the lig-and-dependent activity of the heterodimer partner FXR [11]
Thus far, essentially all of such studies on integra-tion of hormone signaling using cloned nuclear recep-tors have involved vertebrate receprecep-tors However, one
of the most dramatic examples of integrated signaling
by lipoidal regulators is insect metamorphosis, which
is regulated by the interplay between the steroid
20-OH ecdysone (2020-OHE) and the terpenoid methyl epoxyfarnesoate [juvenile hormone III, (JHIII)] The 20OHE receptor (EcR) has been isolated for over a decade [15], and was determined to function in vivo in heterodimer with ultraspiracle (USP), an ortholog of RXR [16] Most studies on the EcR as a ligand bind-ing receptor have focussed EcR bindbind-ing of its ligand while USP is premised or utilized as a ligandless dimer partner More recently, our studies in a model Sf9 cell transfection system utilizing a natural core promoter
of the JH-activated juvenile hormone esterase gene [17] have established that USP can transduce transcrip-tional activation by JH III by way of binding of JH III
or closely related structures to the ligand binding pocket of USP [18] This binding of JH III by USP induces change in USP tertiary conformation [18] and induces or stabilizes USP homodimerization [19] There has been renewed interest in how JH and 20-OH ecdysone signaling are integrated at a molecular level [20,21] Yet, thus far, there have been no reports
on the existence and nature of integration of JH III and 20OHE signaling through the molecules of the EcR⁄ USP heterodimer complex In this study, we used cloned EcR and USP to assess the relationship of the
JH III-USP axis in JH III-transcriptional activation to the 20OHE-EcR axis in ecdysone-transcriptional acti-vation
Results
JH III and 20OHE synergism of promoter activity
In the Sf9 cell transfection system, the JHECore pro-moter does not significantly respond to treatment of transfected cells with either 20OHE alone or JH III (Fig 1A) Placement of DR1 hormone response ele-ments 5¢ to the JHECore promoter confers a twofold induction of promoter activity by 1 lm 20OHE and a twofold induction by 100 lm JH III However, more significantly, the combination of the two hormones yielded a sixfold induction (Fig 1B) This transcrip-tional interaction of JH III and 20OHE occurred over the 10 nm to 1 lm range of dose-dependent response
of the DR1JHECore to the 20OHE (Fig 1C)
Trang 3We tested whether this transcriptional interaction
between JH III and 20-OHE is influenced by the
nature of the hormone response element For this
purpose, we compared the DR1JHECore promoter
reporter with an IR1JHECore promoter reporter
These two motifs (DR1 and IR1) contain identical half
sites and the same intervening single base, the
differ-ence being the orientation of the second half site as in
the same direction (DR1) or opposite direction (IR1)
as the first half site Each transfected promoter
con-struct was exposed to either a dose range of JH III
(0.1–30 lm) in the presence of 1 lm 20-OHE or the
dose range of JH III alone, and the activation ratio in
relation to the solvent (EtOH) treatment calculated As
shown in Fig 2, when the data for activation by
JH III alone vs JH III plus 20-OHE were plotted
for each promoter construct, the resultant slopes of
the plots are not the same The shallower slope for the
IR1JHECore promoter construct compared to the
DR1bJHEconstruct shows that the IR1JHECore
con-struct transduced a greater effect of JH III on 20OHE
than occurred with the DR1JHECore construct For
the IR1JHECore, JH III alone at 30 lm yielded an
approximately twofold induction, 20OHE alone at
1 lm yielded an approximately 35-fold induction, but
together the two hormones yielded a 45-fold induction,
which is distinctly greater than an additive effect
These results show that the transcriptional interaction
of these two hormones is transduced through the
hor-mone response element, and not through some other
region of the transfected plasmids Functionally
important is that this result shows that the opposite
orientation of the second half site of what is otherwise
an identical hormone response element causes the two
response elements to differ in their effectiveness to
pro-mote this interaction between JH III and 20-OHE
USP Interaction with DNA and EcR
We have demonstrated previously that under our con-ditions, purified dUSP can bind to the DR12 motif [19] Using similar conditions, we confirmed by elec-trophoretic mobility shift assay (EMSA) and supershift with anti-USP Ig that purified dUSP can bind to the same DR1 motif as is contained in the DR1JHECore reporter used in the cell transfection experiments (Fig 3A, arrow) We next assessed whether USP, as it exists in Sf9 nuclear extracts, can bind to this same DR1 motif In EMSA assay, a single protein–DR1
Fig 1 Induction of DR1JHECore promoter
by juvenile hormone III (JH III) and 20-OH
ecdysone (20OHE) (A) JHECore promoter is
unresponsive to JH III and 20-OHE (B)
Placement of five DR1 motifs immediately
5¢- to the JHECore renders it mildly
inducible by JH III or 20-OHE, and strongly
inducible by treatment with the two
hormones combined (C) Over a range of
0.01–1.0 l M the DR1JHECore responds
more strongly to cotreatment with 100 l M
JH III than to 20OHE alone.
Fig 2 Differential interaction of JH III with 20OHE as mediated by different hormone response elements The JHECore promoter was placed under the enhancement of either DR1 or IR1 hormone response elements, and subjected to Sf9 cell transfection assay and subsequent treatment with ethanol, a dose-range of JH III, and ⁄ or 1 l M 20OHE Plotted here for the two constructs is the relationship between the level of activation obtained for a given construct due to JH III only vs the level of activation obtained when treated with both JH III and 20OHE The IR1 motif, by its flatter slope, yielded a much stronger effect of adding JHIII to 20OHE than was exhibited through the DR1 motif.
Trang 4complex is observed (Fig 3B) The specificity of the
binding is again confirmed by the competition of
unla-belled DR1 probe The presence of USP in the
com-plex was confirmed by a supershift formed upon the
addition of a monoclonal antibody (AB11) specific for
USP (arrow) The specificity of the supershift was
con-firmed by the absence of a supershift when a
monoclo-nal antibody to an unrelated antigen (ELAV) was
used These results confirm that USP binds to the
same DR1 hormone response motif that is used in the
DR1JHECoretransfection construct
We also used this system to test for the binding of
EcR to the same DR1 hormone response element As
the monoclonal antibody to the Drosophila ecdysone
receptor (dEcR) does not cross-react well with the
endogenous Sf9 ecdysone receptor, we transfected Sf9
cells with a plasmid expressing dEcR, and harvested
the cells for binding of the extract to the DR1 probe
As shown in Fig 3C, a specific complex was observed
binding to the probe, which could be competed by
unlabelled probe (self) DNA, but not by an unrelated
(nonself) DNA The complex could be supershifted by
the monoclonal antibody to dEcR (arrow), establishing
that EcR expressed in Sf9 cells can bind to the DR1
motif
The direct interaction of USP with EcR in the Sf9
nuclear extracts was assessed by
coimmunoprecipita-tion assay When we transfected the Sf9 cells with a
dEcR-expressing plasmid (to again use the
dEcR-speci-fic antibody), in order to also increase the amount of
intracellular USP to a level detectable in the
coimmuno-precipitation procedure, we also contransfected the same cells with a plasmid overexpressing dUSP The cells were harvested and nuclear extracts were pre-pared After incubation of the extracts with monoclo-nal antibody to the dUSP, and immunoprecipitation of the complex, the protein complex was subjected to SDS⁄ PAGE and immunoblotting with monoclonal antibody to dEcR As shown in Fig 3D, a positive immunoblot signal of the correct molecular size for dEcR was obtained As a negative control, no signal was obtained when only dUSP or only dEcR was overexpressed (Preliminary experiments using a mono-clonal antibody against dUSP for both the immuno-precipitation and for the immunoblot of the precipitate confirmed that the dUSP was immunoprecipitated in the control receiving a dUSP-expressing plasmid only) Thus, a heterodimer complex of USP and EcR exists
in the extracts of Sf9 cells
JH III activation pathway is not identical
to 20OHE activation pathway The dose dependence of the action of JH III to induce transcription of the DR1JHECore construct was then assessed Induction of this promoter construct by treatment of the transfected cells with JH III was detectable in the mid-micromolar range (Fig 4A) This was compared at the same time to the dose depend-ence of JH III action to increase the transcriptional induction of 1 lm 20OHE As shown in Fig 4A, the region of the dose-dependent JH III action to increase
Fig 3 Binding of ultraspiracle (USP) to DR1 motif and to ecdysone receptor (EcR) (A) Electrophoretic mobility shift assay (EMSA) of recom-binant dUSP binding to DR1 probe (left lane), and supershift (arrow) with AB11anti-dUSP monoclonal Ig (middle lane) The negative control shows no supershift with monoclonal antibody against irrelevant ELAV protein (right lane) (B) EMSA with DR1 probe using nuclear extract from Sf9 cells The major shift-band is specific on account of its competition with self but not with nonself unlabelled competitor The com-plex on the DR1 probe contains USP, as seen by the supershift (arrow) with the AB11 anti-USP monoclonal Ig (C) EMSA with DR1 probe using nuclear extract from Sf9 cells transfected with expression plasmid for dEcR The major shift-band (leftmost lane) is specific on account
of its competition with self but not with nonself unlabelled competitor The complex on the DR1 probe contains dEcR, as seen by the super-shift (arrow) with the antidEcR monoclonal antibody (D) USP binds with EcR in Sf9 nuclear extracts Lysates from Sf9 cells cotransfected with plasmids expressing dUSP, or dEcR, or both, or empty expression vector, were first immunoprecipitated with AB11 anti-dUSP mAb, and the pellet subjected to immunoblotting (following SDS ⁄ PAGE), using anti-dEcR mAb.
Trang 5the transcriptional activation by 20OHE closely
paral-lels the mid-micromolar region of the action of JH III
alone to induce transcription These results suggest
that the mechanism of synergism is not one in which
the presence of 20OHE lowers the concentration of
JH III at which JH III exerts its transcriptional action
These results also are also evidence that the site of
JH III action for its effect to alone induce
transcrip-tion of the DR1JHECore is the same as, or has a
titra-tion curve indistinguishable in this transfectitra-tion assay
from, the site of JH III action for its transcriptional
interaction with 20OHE
Our previous studies have demonstrated that JH III
induction of DR12JHECore promoter activity in the
Sf9 cells system operates through JH III binding to
USP [18,19] We have presently also shown that both
recombinant dUSP and Sf9 cell USP can bind to the
DR1 motif (Fig 3, above) Thus, we tested the
partici-pation of USP in JH III activation of DR1JHECore
by observing the effect of increasing the concentration
of exogenous dUSP on the induction caused by JH III
As shown in Fig 4B, as the amount of plasmid
over-expressing USP is progressively increased, the
fold-induction caused by treatment with JH III alone increased Yet, under the same conditions, transfection
of increasing amounts of USP-expressing plasmid does not cause an increase in response to 20OHE alone, but instead causes a decline in the level of induction caused
by treatment with 20OHE alone All published reports
on USP function thus far indicate that USP acts as a dimer and not as a monomer, so it is unlikely that the overexpressed dUSP here is transducing JH signaling
as a monomer, and we have shown previously that both half-sites of the DR12 motif are necessary for DR12JHECorepromoter to transduce JH III signaling [18] Thus, in this experiment, the exogenous amount
of USP that enhances JH III signaling but not 20OHE signaling could be (a) overexpressed in sufficient excess over the endogenous EcR that by mass-action is favor-ing USP dimerizfavor-ing with its (abundant) self or with a partner that is not EcR, and by competitive binding thereby prevents EcR⁄ dUSP complex from binding to the hormone response element, or (b) that some other factor that is needed for 20OHE activation, but not
JH III-activation, is already limiting before the over-expression of more exogenous USP
Fig 4 Effect of treatment with hormones, and of receptor expression, on activation of the DR1JHECore reporter promoter (A) Activation of the DR1JHECore promoter by JH III alone (black histogram bars) or together with 1 l M 20-OHE (white histogram bars) The values for JH III alone are plotted at 20· their actual value, to enable more direct visualization that the range of JH III action alone is over a similar range of its action in the presence of 20OHE These data are from a single experiment (B) Differential effect of transfection of increasing concentra-tions of dUSP-expressing plasmid on action of JH III alone (black histogram bars) or 20OHE alone (white histogram bars) to activate DR1JHECore reporter promoter Data are average (±SE) of three independent experiments (C) Transfection of increasing concentrations of dUSP-expressing plasmid, while increasing the transcriptional activation arising from treatment with JH III alone (d), does not result in enhancement of 20OHE-activation by JH III (s) Data points are the average of two independent experiments (D) Effect of transfection of increasing concentrations of dEcR-expressing plasmid on action of JH III alone (black histogram bars) or 20OHE alone (white histogram bars)
to activate DR1JHECore reporter promoter Data are average (± SE) of two independent experiments.
Trang 6We then assessed the effect of overexpression of
dUSP on the transcriptional activation pathways that
are induced by 20OHE alone and on the interaction of
JH III and 20OHE In order to visualize more clearly
the effects, the data were analyzed for the fold change
in the hormonal activation that was caused by the
transfection of a particular amount of USP-expressing
plasmid So, for example, the value of 2.0· for
1000 ng of wtdUSP-expressing plasmid means that the
JH III induction was yet another 2· higher than the
induction already caused by JH III in the absence of
transfected dUSP As shown in Fig 4C, as more
USP-expressing plasmid was transfected, there was an
increasingly greater transcriptional activation by
JH III, above and beyond that being transduced in the
absence of exogenous USP However, under the same
conditions, increasing the amount of dUSP-expressing
plasmid did not increase the transcriptional activity
caused by either 20OHE alone or by 20OJE together
with JH III These results raise the possibility that the
nature of the dUSP requirement for transcriptional
activation by JH III alone is not the same activation
pathway (i.e not the same molecular complex) as that
involved in the action of JH III and 20OHE together
We tested the hypothesis that USP forms a
homo-dimer under the above condition where overexpressed
dUSP aids JH III activity, but which does not aid transcription induced by either the 20OHE activity or the JH III⁄ 20OHE We cotransfected Sf9 cells with dUSP possessing two different tags (as a GFP-dUSP fusion and as HA-tagged dUSP) As shown in Fig 5A, the control of direct immunoblotting of total cell lysate proteins showed that the higher molecular size (101 kDa) GFP-dUSP fusion and the HA-USP (55 kDa) are indeed expressed in cells transfected with their expression plasmids, but not in cells transfected with empty pIE1-4 vector In Fig 5B, we show that when anti-HA Ig was used to precipitate HA-USP, and then anti-GFP was used to probe the pellet, GFP-dUSP was found in the pellet only in the treatment in which cells were transfected with both GFP-dUSP and HA-USP These results indicate that under the condi-tions of overexpression of dUSP that further aids
JH III in activation of the DR1JHECore promoter (but which does not aid 20OHE activation), dUSP oligomer (we interpret this to include homodimer) exists in Sf9 extracts
Is EcR part of the JH III-activation pathway? The indication that overexpressed USP does not aid the 20OHE-activation pathway prompted us to
exam-Fig 5 Immunoprecipitation of USP homodimer Sf9 cells were transfected with the indicated expression plasmids (A) Aliquots from total cell lysates were loaded directly to SDS ⁄ PAGE for immunoblotting with the indicated antibody, which confirmed that the GFP-dUSP and HA-dUSP were expressed in the cells transfected with the respective expression plasmid, and as a negative control neither was detected in cells transfected with the empty expression plasmid The upper blot in panel A shows the reactive band for GFP-dUSP present in cells trans-fected with GFP-dUSP-expressing plasmid and not in cells transtrans-fected with empty vector The lower blot in (A) shows the reactive band for HA-dUSP present in cells transfected with HA-dUSP-expressing plasmid and not in cells transfected with empty vector (B) Lysates from cells transfected with the indicated plasmid constructs were first immunoprecipitated with anti-HA Ig, and the immunoprecipitate then sub-jected to immunoblotting and probing with anti-GFP Ig The only treatment to yield a 101 kDa corresponding to GFP-dUSP was that for cells cotransfected with both the plasmids encoding GFP-dUSP and HA-dUSP This result shows the presence of USP homodimer in the cell lysates.
Trang 7ine a reciprocal question: whether EcR is
participa-ting in the JH III alone-activation pathway As one
approach to this question, we tested the effect of
over-expression of dEcR on the JH III activation pathway
The pattern observed when dEcR was overexpressed
was the opposite of that observed for the case when
dUSP was overexpressed As shown in Fig 4D, as the
amount of plasmid overexpressing EcR was increased
progressively, the fold induction caused by treatment
with 20OHE alone also increased progressively
How-ever, at the same time, there was no increase in the
level of induction caused by treatment with JH III
alone This result suggests that the overexpressed level
of EcR does not participate in (does not aid) the
path-way that transduces the USP-dependent transcriptional
activation by JH III alone This result is particularly
relevant, as it has been well-established that the
EcR-containing receptor complex that transduces
20OHE transcriptional activation in insect cells is the
EcR⁄ USP complex
Is the putative USP coactivator surface a part
of the JH activation pathway?
Extensive mutational and cocrystallographic studies on
vertebrate receptors, including human RXR (which is
the ortholog of USP) have shown that an area on the
surface of the ligand binding domain involving the
C-terminus of a-helix 3, a-helix 4, and the N-terminus
of a-helix 5 comprise a surface that is important for
the function of the receptor Conserved in this area
across a wide range of nuclear hormone receptors is a
hydrophobic groove Depending on the receptor,
func-tions of this region include: (a) binding the a-helix 12
of the receptor into its own hydrophobic groove [22];
(b) stabilization of the a-helix 12 near the hydrophobic
groove in a position that the C-terminus of the a-helix
12 will interact with the receptor dimer partner [23] or
(c) to recruit coactivator⁄ corepressor proteins to bind
with the receptor at the hydrophobic groove [24] In
human RXR, mutation of the residue corresponding
to dUSP L314 converts hRXR into a dominant
negat-ive receptor [25] We therefore mutated this residue
in the dUSP (L314R), and tested its effect to act as
a dominant negative in the Sf9 cell transfection,
JH III-activation pathway As shown in Fig 6,
addi-tion of JH III to Sf9 cells transfected with the
DR1JHECore plasmid resulted in an induction of
promoter activity However, cotransfection with
pro-gressively greater amounts of the mutant L314R
dUSP-expressing plasmid caused a progressive decrease
in the JH III-inducibility of the DR1JHECore
promo-ter (Fig 6, upper panel) To further test whether the
JH III activation pathway requires the presence of the wild-type L314 USP, cells were transfected with a dominant-negative acting dose of L314R-expressing plasmid (that suppressed JH III-activation), but were also cotransfected with increasing amounts of plasmid expressing wtUSP The outcome was that the increas-ing dose of wtUSP progressively rescued the JH III-activation of the DR1bJHECore promoter (Fig 6, lower panel) These results further confirmed the parti-cipation of wtUSP in the JH III-activation pathway, and in particular establish that the wild-type confor-mation of the surface near L314 is necessary for USP transduction of JH III-activation
Discussion
Juvenile hormone transcriptional activation through USP
An important area of investigation in the mechanisms
of invertebrate hormone action is the identification of
Fig 6 Role of JH III-activated transcription by hydrophobic residue (L314) in putative coactivator-binding hydrophobic groove of USP (A) Activity of DR1JHECore promoter in Sf9 cells cotransfected with the indicated increasing concentrations of plasmid expressing dominant negative L314R mutant dUSP, resulting in increasing sup-pression of JH III-activation of the reporter promoter (B) Restor-ation of JH III-activRestor-ation from the suppressive effects of dominant negative L314R dUSP, by cotransfection with the indicated increas-ing concentrations of plasmid expressincreas-ing wild-type dUSP.
Trang 8specific receptors that can bind and transduce juvenile
hormone signaling for transcriptional activation Until
recently, identification of a nuclear receptor site of JH
action has been frustratingly difficult The results of
this study extend our previous findings that indicated
in Sf9 cells JH III(-like) molecules can bind to the
ligand binding pocket of USP, with the effect of
transcription being activated at the transfected,
DR12JHECore promoter [18,19] In those previous
studies, mutations to the ligand binding pocket that
weakened JH III binding to USP also acted as
domin-ant negatives in the USP-dependent, model
JH-acti-vation pathway In this study, we have used a
DR1JHECorereporter promoter to study JH III
signa-ling through USP Thus, the transduction of signasigna-ling
for activation of the JHECore promoter here is
effec-ted through the DR1 element, which we also
demon-strated by EMSA is a binding site for both purified,
recombinant dUSP and for USP endogenous to
nuc-lear extracts from Sf9 cells
Receptor complex involved in JH III)20OHE
signaling
The physiological integration of juvenile hormone and
20OHE signaling has for several decades been an
underpinning of models for regulation of the complex
developmental transition of insect metamorphosis
[26,27], but the molecular mechanisms by which that
integration of signaling may be accomplished has
been frustratingly elusive [28,29] USP does not bind
20OHE [30–32], and in fact EcR is the only
inverteb-rate nuclear hormone receptor shown to undergo
direct transcriptional activation by 20OHE [33] It is
thus unlikely that the integration of nuclear JH III
and 20OHE signaling is mediated directly by a USP
homodimer, or by a USP heterodimer with another
partner other than EcR Przibilla et al [34] have
iden-tified mutations to USP residues that exert allosteric
effects on the activation of the EcR⁄ USP complex by
20-OHE alone, but the operation of JH on that
com-plex was not investigated Recently, Kethidi et al [35]
have reported a regulatory element through which JH
signaling suppresses ecdysone activation, but the
com-ponents of the complex binding at the element were
not ascertained Also recently, Dubrovsky et al [21]
have identified a specific target gene (E75A) for which
ecdysone activation is synergized by JH, but the
direct site of JH action in that synergism was not
reported
In this study, we have demonstrated the enhanced
activation of the JHECore reporter promoter by
cotreatment with JH III (or its metabolite in cultured
Sf9 cells) and 20OHE This action is mediated through the DR1 enhancer that we placed 5¢- to the JHECore promoter As the EcR⁄ USP is the direct receptor target site for 20OHE, then if there is a single DR1-binding complex that is a target through which JH III and 20-OHE signaling is integrated, it would be antici-pated that the complex could be EcR⁄ USP, because that complex contains both EcR, the known target of 20OHE, and USP, which we have shown can bind
JH III and transduce JH III-signaling Overexpressing dUSP alone causes an increase in activation of the DR1JHECore (concomitant with the formation of dUSP oligomers, such as homodimers) by treatment with JH III alone However, the USP⁄ USP homo-dimer, by missing the EcR component, does not enhance the transcriptional activation that is otherwise observed following cotreatment with 20OHE and
JH III Reciprocally, intracellular EcR will bind to direct repeat hormone response elements only as a het-erodimer with USP, not as an EcR⁄ EcR homodimer Concordantly, when EcR is overexpressed, both the dEcR and USP in nuclear extracts are in the complex that binds to the DR1 motif in EMSA, and this over-expression of EcR increases the activation of DR1JHECore by 20OHE We infer from these results that the integration of JH III and 20OHE in the acti-vation of the DR1JHECore promoter reporter in Sf9 cells is mediated through the EcR⁄ USP complex bind-ing to the DR1 enhancer
The transduction of JH III-signaling through USP, which is increased by overexpression of USP, and the transduction of 20OHE-signaling through the EcR component of the EcR⁄ USP heterodimer, prompts a hypothesis that the enhancement conferred by JH III, when cells are cotreated with JH III and 20OHE, is through the USP component of the EcR⁄ USP het-erodimer Consistent with that hypothesis is our observation that the dose-dependence of JH III action (alone) to activate the DR1JHECore was in the same dose range as its effect on transcriptional activation together with 20OHE This result indicates that the site of JH III action for activation of DR1JHECore (i.e, USP) could be the same site as is involved in the
JH III synergism of 20OHE (i.e the USP component
of EcR⁄ USP) Also consistent with that model, we observed that under conditions of EcR overexpression, increased 20OHE activation, there exists an EcR⁄ USP complex in nuclear extracts that can bind to the DR1 enhancer, rendering the USP partner available
at the DR1 motif to integrate the JH III-signaling
We therefore postulate that during the integra-tion of acintegra-tion of JH III and 20OHE to activate DR1JHECore transcription in Sf9 cells, the site of
Trang 9JH III action is the USP component of the USP⁄ EcR
heterodimer
Our recent experiments show that transgenic dUSP,
mutated in the ligand pocket for reduced JH III
bind-ing, is unable to rescue Drosophila melanogaster (null
for USP) from a lethal period from pupariation to
adult emergence, further evidencing that dUSP has an
in vivoligand binding function (R Thomas, D Jones &
G Jones, unpublished data)
Site of JH action in DR1JHECore activation
by JH III alone
As discussed above, our data indicate that the
increased JH III activation of the JHECore by way of
the DR1 motifs (and probably the DR12 motif [18,19])
is through the USP homodimer, or at least not
through USP⁄ EcR heterodimer (In preliminary
experi-ments, under similar conditions of cotransfection of
plasmids expressing dUSP and dDRH38, we have not
detected dUSP⁄ DHR38 heterodimer, and transfection
of dDHR38-expressing plasmid does not increase
JH III-activation of the DR1JHECore promoter
repor-ter; F Fang, Y Xu, D Jones and G Jones,
unpub-lished observation.) In the vertebrate system, there is
also recent evidence for the existence of RXR
homo-dimer-dependent pathways that are activated by RXR
ligand [36], including DR1 binding sites [37] This
model explains the cotransfection results of Baker
et al [38], who observed that under conditions of EcR
overexpression (which our data suggest would foster
sequestering of USP into an EcR⁄ USP heterodimer
complex), their cotransfected model promoter
respon-ded to treatment with 20OHE but did not respond to
treatment with JH alone
Subordination of JH III signaling through USP
to status of EcR activation
There remains the question of why overexpression of
EcR, leading to increased formation of EcR⁄ USP
het-erodimer, did not yield increased cellular response to
treatment with JH III alone, even though USP is
pre-sent as a potential JH III target in the EcR⁄ USP
het-erodimer There has been considerable controversy in
the vertebrate nuclear receptor field concerning the
mechanistic context of ‘subordination’ of 9-cis RA
signaling through RXR in relation to the particular
RXR heterodimer partner and the liganded status of
that partner Some reports have indicated that
activa-tion of RXR by ligand is not permitted by the
heterodimer partner thyroid hormone receptor (VDR),
and thus is a ‘subordinate’ partner to VDR [6] Other
studies find that RXR in the RXR⁄ TR heterodimer can bind ligand but just not dissociate corepressor bound to RXR [39], while yet other investigators adduce the RXR subunit binds ligand but with the effect to dislodge corepressor from the TR subunit [40] The inability of the RXR partner to respond
in vivo to RXR ligand when partnered to RAR has also been taken as evidence that the ligand-dependent activity of RXR is ’subordinated’ to that of RAR [41,42] Evidence has been presented showing the ‘sub-ordination’ in vivo of RXR to the liganded state of RAR arises from the inability of RXR ligand to effect dislodging of corepressor [10] In the RXR⁄ VDR3 sys-tem, RXR has been modeled as a nonligand-binding and therefore silent partner, but a recent report finds that liganded VDR allosterically modifies the apo-RXR from an unliganded conformation to a liganded-like receptor conformation, thus enabling the apo-RXR to recruit coactivators to itself [43] In the opposite direc-tion, Willy and Mangelsdorf [44] showed that binding
of agonist by RXR manifests as activated transcription through the coactivator binding site of the LXR part-ner In yet another twist, a recent report indicates that activation of FXR by FXR ligand is suppressed when the FXR heterodimer partner, RXR, is bound to agonist [11] Further complexity exists in the vertebrate systems in that different vertebrate isoforms of the same hormone receptor may respond to ligand differ-ently with respect to coactivator⁄ corepressor inter-action [45]
Under the conditions of our Sf9 cultured cell sys-tem, exogenous JH III acts through the ligand binding pocket of USP to activate transcription of the DR1JHECore Overexpression of EcR in the trans-fected cells results in the loading of the DR1 enhancer
of the DR1JHECore reporter with the USP⁄ EcR het-erodimer, at least in EMSA assay with nuclear extracts from those cells Yet, despite the presence of USP in the heterodimer complex, the DR1JHECore promoter activity does not further increase in response
to treatment alone with the USP-agonist JH III, though it responds quite well to treatment alone with EcR-agonist 20OHE The response to JH III under conditions of EcR⁄ USP loading on to the DR1 appears to only occur when the EcR partner is ligan-ded with its cognate hormone, 20OHE This result provides evidence that a mechanism of subordination
of the USP response to JH III is operating in the presence of an unliganded EcR heterodimer partner Thus, our study has offered the first invertebrate model system in which the subordination relationships can be tested for two identified nuclear receptors for which an activating ligand is available for each
Trang 10Juvenile hormone transcriptional activation
requires a specific USP surface feature
Our present study found that the transcriptional
acti-vation by USP in response to JH III requires the
pres-ence of wild-type amino acid sequpres-ence at the receptor
surface corresponding to the coactivator binding site in
the ortholog RXR In the model vertebrate receptors,
this hydrophobic groove that serves as the binding site
of coactivators when the binding of ligand causes the
a-helix 12 to become repositioned to one edge of this
hydrophobic groove When USP is concentrated to
10 mgÆmL)1 (orders of magnitude above physiological
levels) and crystallized with a stabilizing fortuitous
pseudoligand (phospholipid), its a-helix 12 is observed
in an antagonist position covering this groove [46]
Our studies with USP prepared at 200· lower
concen-tration (much closer to physiological levels) have
shown that binding of JH III causes the a-helix 12 to
move in relative position [18] In this study, we have
observed that the USP mutation L314R converts USP
into a dominant negative mutant of the JH-activation
pathway in cultured Sf9 cells This result is consistent
with a model in which binding of a JH-like ligand to
wild-type USP can cause a-helix 12 to move in such a
way that the surface involving L314 is accessible to
participate in the JH III-dependent transcriptional
acti-vation mechanism
Experimental procedures
Chemicals
Juvenile hormone III (75% enantiomeric mixture) was from
Sigma and 20-hydroxy ecdysone was from Sigma (St
Louis, MO, USA) Each were dissolved as stock in ethanol
Expression and reporter constructs
The full length coding sequence of Drosophila melanogaster
wild-type USP (dUSP) cloned into the pET32EK vector
(Novagen, Madison, WI, USA), providing for a
trx-His-s-USP fusion protein, and its purification by nickel resin
(elu-tion with imidazole), and then passage of the eluted USP
fraction over Superdex 200 resin in 50 mm sodium
phos-phate buffer, have been detailed in Jones et al [19] The full
length wild-type dUSP, except for the first 9 amino acids,
and wild-type D melanogaster EcR (dEcR, isoform A) were
cloned into the PmeI⁄ NotI sites of the pIE1-4 expression
vector The green fluorescent protein (GFP)-dUSP fusion
protein was prepared in pIE1-4 by subcloning the GFP
coding sequence for into the EcoRI⁄ SmaI sites of the
pIE1-4 vector, upstream of and in the same reading frame as
USP The GFP ‘tag’ provides a total fusion protein size of
101 kDa, which separates it well from the migration on SDS⁄ PAGE of wild-type USP or HA-tagged USP For im-munoprecipitation experiments, a hemagluttanin (HA)-tag was placed at the N-terminus of the dUSP by cloning a double-stranded oligomer of the following sequence (upper
CATCTCTG-3¢) into the BamHI site of the above pIE1-4 vector already containing the dUSP coding sequence in the PmeI⁄ NotI sites The dUSP mutant L314R was made from the above wild-type dUSP in pIE1-4, by a QuikChange XL Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA) where the sequence of the mutation-target primer
GGATCG-3¢ Antisense dEcR in pIE1-4 was prepared by cloning its PCR product with a reverse orientation in the PmeI⁄ NotI sites of the vector All constructs were con-firmed by sequencing Expression of each receptor in trans-fected Sf9 cells was confirmed by immunoblotting with AB11 monoclonal antibody against dUSP (a gift from
F Kafatos, European Molecular Biology Laboratory, Hei-delberg, Germany), with monoclonal antibody against the ecdysone receptor (a gift from R Evans, Salk Institute,
La Jolla, CA, USA), with rabbit polyclonal antibody against HA-tag (Abcam Ltd, Cambridge, UK), or with monoclonal antibody against GFP (Chemicon Interna-tional, Temecula, CA, USA)
The DR1JHECore promoter reporter construct in pGL3 luciferase reporter vector (Promega, Madison, WI, USA) was prepared as described in Xu et al [18] First, the JHE-Core promoter ()61 to +28 [17]); was subcloned into the KpnI⁄ BglII sites of this reporter vector An NheI site was then manufactured immediately 5¢- to the KpnI site Com-plementary oligonucleotides encoding a direct repeat motif (underlined) separated by one base (DR1, 5¢-AGGTCAA AGGTCA-3¢) were synthesized with each oligonucleotide possessing at its 5¢-end the four base overhang of an NheI restriction site (CTAG) Upon annealing, the double-stran-ded oligonucleotides would then have a CTAG overhang at each 5¢-end The annealed oligonucleotides were then ligated into concatamers, fractionated by native PAGE and the gel fractions corresponding to higher concatamer forms recov-ered and ligated into the NheI site In this study we used a recovered construct containing five tandem DR1 motifs The orientation of the five DR1 motifs, with ‘fi’ indicating
a single motif of the above, upper strand sequence read-ing toward the downstream JHECore promoter, is
‹fifi‹fi The IR1JHECore promoter reporter [contain-ing an inverted repeat (underlined) with the half sites separ-ated by one base, IR1] was prepared by the annealing of single stranded oligomers encoding the sequence (upper
flanked by an overhang of NheI restriction site, to make an annealed double-stranded fragment containing an NheI overhang at each end The fragment was cloned into the NheI site of the pGL3 luciferase vector described above and