Interactions of ultraspiracle with ecdysone receptor in the transduction of ecdysone- and juvenile hormone-signaling


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


Analyses of integration of two-hormone signaling through the vertebrate nuclear hormone receptors, for which the retinoid X receptor is one partner, have generated a number of mechanistic models, including those described as ‘subordination’ models wherein ligand-activation of one partner is subordinate to the liganded state of the other partner. However, mechanisms 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 invertebrate 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 DR1JHECoreLuciferase (or IR1JHECoreLuciferase) reporter promoter each induced expression of the reporter. Cotreatment of transfected cells with both hormones yielded a greater than additive effect on transcription, for especially the IR1JHECoreLuciferase reporter. Overexpression in Sf9 cells of recombinant Drosophila melanogaster ultraspiracle (dUSP) fostered formation of dUSP oligomer (potentially homodimer), as measured by coimmunoprecipitation 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 recombinant D. melanogaster ecdysone receptor (dEcR) in the transfected cells generated 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 integration of JH III and 20OHE signaling is through the USP/EcR heterodimer, but that when the EcR partner is unliganded, the USP partner in this system is unable to transduce the JH III-activation.


20-OH ecdysone

DR1 and IR1

direct repeat and inverted repeat with one intervening base between repeats, respectively


ecdysone receptor


electrophoretic mobility shift assay


juvenile hormone


core promoter from juvenile hormone esterase gene


retinoic acid





Nuclear hormone receptors play key roles in metazoan development, metabolic homeostasis, and response to xenobiotics. Some of these nuclear receptors bind ligands that modulate the regulatory effect of these receptors on gene transcription, while other receptors 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 regulation 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 receptor complexes, and their corresponding multiple ligands, 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 complex 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 dissociate 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 function 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 ligand-dependent activity of the heterodimer partner FXR [11].

Thus far, essentially all of such studies on integration of hormone signaling using cloned nuclear receptors have involved vertebrate receptors. 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 (20OHE) 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 binding receptor have focussed EcR binding 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 transcriptional 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 activation.


JH III and 20OHE synergism of promoter activity

In the Sf9 cell transfection system, the JHECore promoter does not significantly respond to treatment of transfected cells with either 20OHE alone or JH III (Fig. 1A). Placement of DR1 hormone response elements 5′ to the JHECore promoter confers a twofold induction of promoter activity by 1 µm 20OHE and a twofold induction by 100 µm JH III. However, more significantly, the combination of the two hormones yielded a sixfold induction (Fig. 1B). This transcriptional interaction of JH III and 20OHE occurred over the 10 nm to 1 µm range of dose-dependent response of the DR1JHECore to the 20OHE (Fig. 1C).

Figure 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 µm the DR1JHECore responds more strongly to cotreatment with 100 µm JH III than to 20OHE alone.

We 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 difference 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 construct was exposed to either a dose range of JH III (0.1–30 µm) in the presence of 1 µm 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 DR1bJHE construct shows that the IR1JHECore construct transduced a greater effect of JH III on 20OHE than occurred with the DR1JHECore construct. For the IR1JHECore, JH III alone at 30 µm yielded an approximately twofold induction, 20OHE alone at 1 µm 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 hormone 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 promote this interaction between JH III and 20-OHE.

Figure 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 µm 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.

USP Interaction with DNA and EcR

We have demonstrated previously that under our conditions, purified dUSP can bind to the DR12 motif [19]. Using similar conditions, we confirmed by electrophoretic 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 complex is observed (Fig. 3B). The specificity of the binding is again confirmed by the competition of unlabelled DR1 probe. The presence of USP in the complex was confirmed by a supershift formed upon the addition of a monoclonal antibody (AB11) specific for USP (arrow). The specificity of the supershift was confirmed by the absence of a supershift when a monoclonal 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 DR1JHECore transfection construct.

Figure 3.

Binding of ultraspiracle (USP) to DR1 motif and to ecdysone receptor (EcR). (A) Electrophoretic mobility shift assay (EMSA) of recombinant 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 complex 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 supershift (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.

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 coimmunoprecipitation assay. When we transfected the Sf9 cells with a dEcR-expressing plasmid (to again use the dEcR-specific antibody), in order to also increase the amount of intracellular USP to a level detectable in the coimmunoprecipitation procedure, we also contransfected the same cells with a plasmid overexpressing dUSP. The cells were harvested and nuclear extracts were prepared. After incubation of the extracts with monoclonal 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 monoclonal antibody against dUSP for both the immunoprecipitation 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 dependence of JH III action to increase the transcriptional induction of 1 µm 20OHE. As shown in Fig. 4A, the region of the dose-dependent JH III action to increase the transcriptional activation by 20OHE closely parallels 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 transcription of the DR1JHECore is the same as, or has a titration curve indistinguishable in this transfection assay from, the site of JH III action for its transcriptional interaction with 20OHE.

Figure 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 µm 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 concentrations 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 (•), does not result in enhancement of 20OHE-activation by JH III (○). 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.

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 participation 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 overexpressing 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 DR12JHECore promoter 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 favoring USP dimerizing 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 overexpression of more exogenous USP.

We 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 homodimer 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 conditions 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.

Figure 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 transfected with GFP-dUSP-expressing plasmid and not in cells transfected 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 subjected 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.

Is EcR part of the JH III-activation pathway?

The indication that overexpressed USP does not aid the 20OHE-activation pathway prompted us to examine a reciprocal question: whether EcR is participating in the JH III alone-activation pathway. As one approach to this question, we tested the effect of overexpression 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. However, 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 pathway 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 α-helix 3, α-helix 4, and the N-terminus of α-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, functions of this region include: (a) binding the α-helix 12 of the receptor into its own hydrophobic groove [22]; (b) stabilization of the α-helix 12 near the hydrophobic groove in a position that the C-terminus of the α-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 negative 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, addition of JH III to Sf9 cells transfected with the DR1JHECore plasmid resulted in an induction of promoter activity. However, cotransfection with progressively greater amounts of the mutant L314R dUSP-expressing plasmid caused a progressive decrease in the JH III-inducibility of the DR1JHECore promoter (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 increasing dose of wtUSP progressively rescued the JH III-activation of the DR1bJHECore promoter (Fig. 6, lower panel). These results further confirmed the participation of wtUSP in the JH III-activation pathway, and in particular establish that the wild-type conformation of the surface near L314 is necessary for USP transduction of JH III-activation.

Figure 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 suppression of JH III-activation of the reporter promoter. (B) Restoration of JH III-activation from the suppressive effects of dominant negative L314R dUSP, by cotransfection with the indicated increasing concentrations of plasmid expressing wild-type dUSP.


Juvenile hormone transcriptional activation through USP

An important area of investigation in the mechanisms of invertebrate hormone action is the identification of specific 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 dominant negatives in the USP-dependent, model JH-activation pathway. In this study, we have used a DR1JHECore reporter promoter to study JH III signaling through USP. Thus, the transduction of signaling for activation of the JHECore promoter here is effected through the DR1 element, which we also demonstrated by EMSA is a binding site for both purified, recombinant dUSP and for USP endogenous to nuclear 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 invertebrate 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 identified 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 complex was not investigated. Recently, Kethidi et al. [35] have reported a regulatory element through which JH signaling suppresses ecdysone activation, but the components 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 anticipated 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 homodimer, 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 heterodimer 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 overexpression of EcR increases the activation of DR1JHECore by 20OHE. We infer from these results that the integration of JH III and 20OHE in the activation of the DR1JHECore promoter reporter in Sf9 cells is mediated through the EcR/USP complex binding 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 heterodimer. 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 integration of action of JH III and 20OHE to activate DR1JHECore transcription in Sf9 cells, the site of JH 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 binding, 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 vivo ligand 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 experiments, 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 reporter; F. Fang, Y. Xu, D. Jones and G. Jones, unpublished observation.) In the vertebrate system, there is also recent evidence for the existence of RXR homodimer-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 responded 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 heterodimer, did not yield increased cellular response to treatment with JH III alone, even though USP is present as a potential JH III target in the EcR/USP heterodimer. 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 activation 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 ‘subordination’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 system, 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 direction, Willy and Mangelsdorf [44] showed that binding of agonist by RXR manifests as activated transcription through the coactivator binding site of the LXR partner. 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 differently with respect to coactivator/corepressor interaction [45].

Under the conditions of our Sf9 cultured cell system, exogenous JH III acts through the ligand binding pocket of USP to activate transcription of the DR1JHECore. Overexpression of EcR in the transfected cells results in the loading of the DR1 enhancer of the DR1JHECore reporter with the USP/EcR heterodimer, 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 liganded 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.

Juvenile hormone transcriptional activation requires a specific USP surface feature

Our present study found that the transcriptional activation by USP in response to JH III requires the presence of wild-type amino acid sequence 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 α-helix 12 to become repositioned to one edge of this hydrophobic groove. When USP is concentrated to 10 mg·mL−1 (orders of magnitude above physiological levels) and crystallized with a stabilizing fortuitous pseudoligand (phospholipid), its α-helix 12 is observed in an antagonist position covering this groove [46]. Our studies with USP prepared at 200× lower concentration (much closer to physiological levels) have shown that binding of JH III causes the α-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 α-helix 12 to move in such a way that the surface involving L314 is accessible to participate in the JH III-dependent transcriptional activation mechanism.

Experimental procedures


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 (elution with imidazole), and then passage of the eluted USP fraction over Superdex 200 resin in 50 mm sodium phosphate 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 immunoprecipitation 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 strand, 5′-AGCTACCCATACGACGTGCCAGACTACGCATCTCTG-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 was: 5′-CGACCAGGTGATTCTGagTGAAAGCCGCTTGGATCG-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 confirmed by sequencing. Expression of each receptor in transfected Sf9 cells was confirmed by immunoblotting with AB11 monoclonal antibody against dUSP (a gift from F. Kafatos, European Molecular Biology Laboratory, Heidelberg, 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 International, 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 JHECore 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. Complementary oligonucleotides encoding a direct repeat motif (underlined) separated by one base (DR1, 5′-AGGTCAAAGGTCA-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-stranded 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 recovered 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 ‘→’ indicating a single motif of the above, upper strand sequence reading toward the downstream JHECore promoter, is ←→→←→. The IR1JHECore promoter reporter [containing an inverted repeat (underlined) with the half sites separated by one base, IR1] was prepared by the annealing of single stranded oligomers encoding the sequence (upper strand) 5′-CTAGGAGGTCAATGACCTC-3′, which is 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 a construct containing a single motif in the → direction was used in this study.

Cell culture and transfections

Spodoptera frugiperda cell line, Sf9, was maintained and transfected as described previously [19]. To study the role of USP in activation of the reporter promoter in ligand-treated cells, dUSP cDNA in pIE1-4 and its mutant derivative (L314R) were cotransfected with the reporter construct. At 36 h after transfection, the cells were treated with the respective compound in ethanol solvent (0.1% final ethanol concentration) or just ethanol solvent only, or left as a no-treatment control. After 48 h of the treatment, the cells were harvested and the activity of the luciferase reporter was measured using a luciferase assay kit (Promega) in a multipurpose scintillation counter (Beckman, Fullerton, CA, USA). β-galactosidase activity was measured using chlorophenol red-β-d-galactopyranoside monosodium (CPRG; Roche Molecular Biochemicals, Indianapolis, IN, USA) as a colorimetric substrate. On each occasion that the given cell transfection assay condition was tested, three separate replications (wells) were included, and the given cell transfection assay condition was performed on at least two or three independent occasions (days). We observed, as is commonly observed with such cell line experiments, that the pattern of result for the given ligand was consistent, although the absolute level of the corresponding reporter enzyme activity varied from one occasion to the next in relation to the vigor of the cell culture at the time. Thus, unless otherwise indicated, for each experiment we show a typical result of one of the multiple occasions that the experiment was performed. Sf9 cells express endogenous EcR and endogenous USP. Thus, transfection of plasmid expressing dEcR and dUSP is assessing the effect of the exogenous receptor on the reporter promoter, above and beyond the effect of the endogenous receptor.

Immunoprecipitation assay

Sf9 cells were cotransfected with either pIE1-4-dUSP, pIE1-4-HA-USP, pIE1-4-GFP-USP and/or pIE1-4-dEcR, as described above. Upon harvest of cells, the concentration of JH III in the lysate was maintained at 100 µm during subsequent processing. For immunoprecipitation, the lysate was incubated overnight at 4 °C with rabbit polyclonal antibody against HA (Abcam Ltd). The incubate was then precipitated at 4 °C with protein A-Sepharose (Sigma) after another 2 h incubation, the precipitate washed with a washing buffer (50 mm Tris/acetate, pH 7.5, 0.3 m NaCl, 0.5% IGEPAL® CA-630, 0.1% SDS, 0.02% NaN3) and the pellet then subjected to SDS/PAGE (8% acrylamide). The fractionated proteins were electroblotted to Immobilon membrane and then probed with either mouse monoclonal Ig against GFP (Chemicon International) or with AB11 mouse monoclonal Ig against dUSP. After washing with a TBS buffer (25 mm Tris/HCl, pH 8.0, 0.138 m NaCl, and 2.7 mm KCl), the blot was incubated with horseradish peroxidase-labelled anti-mouse (Bio-Rad, Hercules, CA, USA) secondary Ig, and then developed, as described by [18].

Electrophoretic mobility shift assay

The double-stranded probe containing a single DR1 motif (5′-AGGTCAAAGGTCA- 3′) was end-labeled with 32P, and used in gel-shift assay to test for specific binding by either purified recombinant dUSP, or by components of Sf9 nuclear extracts. Binding conditions for recombinant dUSP were 1.0 µg of dUSP, 1 µg dIdC, 3 µL of 5× binding buffer (1×: 10 mm of Tris/HCl, pH 7.5, 33 mm KCl, 1 mm MgCl2, 0.5 mm EDTA, 0.5 mm dithiothreitol, 4% glycerol), 10 fmoles of labeled probe, self or nonself competitor at 100× (nonself competitor = pGL3 vector polylinker sequence), and in some cases also mAb AB11 anti-dUSP, all in a binding reaction volume of 20 µL. Similar conditions were used for assays performed with Sf9 nuclear extracts, except that 10 µg of nuclear extract were used and the dIdC was titrated to an optimum, depending on the particular nuclear extract preparation, and that in some assays there was included a monoclonal antibody against dEcR, or a monoclonal antibody against dUSP or a monoclonal antibody against the unrelated protein ELAV [47]. The antibody was added 30 min before addition of the probe. After incubation at 4 °C, the binding reaction contents were subjected to native polyacrylamide gel electrophoresis (6% acrylamide). After electrophoresis the gel was dried, exposed to X-ray film (Kodak, Rochester, NY, USA), and the resulting image scanned into Adobe photoshop.


This research was supported, in part, by NIH DK39197 and GM463713.