Activation of transcription through the ligand-binding pocket of the orphan nuclear receptor ultraspiracle

Authors


G. Jones, Molecular and Cellular Biology Section, Department of Biology, University of Kentucky Lexington, KY 40506, USA. Fax: + 1 859 257 7505, Tel.: + 1 859 257 2105, E-mail: gjones@pop.uky.edu

Abstract

The invertebrate nuclear receptor, ultraspiracle (USP), an ortholog of the vertebrate RXR, is typically modelled as an orphan receptor that functions without a ligand-binding activity. The identification of a ligand that can transcriptionally activate USP would provide heuristic leads to the structure of potentially high affinity activating compounds, with which to detect unknown regulatory pathways in which this nuclear receptor participates. We show here that the application of the sesquiterpenoid methyl epoxyfarnesoate (juvenile hormone III) to Sf9 cells induces transcription from a transfected heterologous core promoter, through a 5′-placed DR12 enhancer to which the receptor ultraspiracle (USP) binds. Isolated, recombinant USP from Drosophila melanogaster specifically binds methyl epoxyfarnesoate, whereupon the receptor homodimerizes and changes tertiary conformation, including the movement of the ligand-binding domain α-helix 12. Ligand-binding pocket point mutants of USP that do not bind methyl epoxyfarnesoate act as dominant negative suppressors of methyl epoxyfarnesoate-activation of the reporter promoter, and addition of wild-type USP rescues this activation. These data establish a paradigm in which the USP ligand-binding pocket can productively bind ligand with a functional outcome of enhanced promoter activity, the first such demonstration for an invertebrate orphan nuclear receptor. USP thus establishes the precedent that invertebrate orphan receptors are viable targets for development of agonists and antagonists with which to discern and manipulate transcriptional pathways dependent on USP or other orphan receptors. The demonstration here of these functional capacities of USP in a transcriptonal activation pathway has significant implications for current paradigms of USP action that do not include for USP a ligand-binding activity.

Abbreviations
core

a reporter core promoter from the JHE gene

DR

direct repeat

RA

retinoic acid

hRAR

human retinoic acid receptor

hRXR

human retinoid X receptor

USP

ultraspiracle

Nuclear hormone receptors are a primary tranduction mechanism through which extracellular hormonal signals are transduced into genetic regulation of metabolic pathways and developmental programs. The past two decades have seen the steady identification of mammalian receptors of well-known ligands (steroids, thyroid hormone, all-trans retinoic acid (RA) [1,2]), as well as the identification of endogenous ligands for initially orphaned receptors [3–5]. Similarly, steroid nuclear receptors in invertebrate models of transcriptional regulation, such as the Drosophila melanogaster ecdysteroid receptor (dEcR), were isolated a decade ago and used to develop important concepts in hormone action [6–12].

In parallel to the search for receptors that can be activated by known ligands, has been the search for ligands of orphan receptors, which are members of the steroid nuclear receptor superfamily whose natural ligands are unknown [13]. The biological relevance of identification of agonistic or antagonistic ligands for orphan receptors is several fold. First, the ability of a chemical structure to fit into the ligand-binding pocket of an orphan receptor and thereby transcriptionally activate the orphan receptor would raise the possibility that the orphan receptor ligand-binding pocket has a conformation enabling it to bind with and be activated by a natural ligand of similar structure. Second, the identification of ligands that transcriptionally activate or antagonize an orphan receptor would aid the discovery of regulatory pathways in which the receptor participates. Finally, transcriptional agonists and antagonists of orphan receptors provide leads to pharmacologically significant structures that, through the orphan receptor, can selectively intercede in disease pathways or that can disrupt disease-causing or disease-transmitting organisms, and not affect related receptors in humans or other nontarget organisms [14].

Identification of chemical compounds that bind to the ligand-binding pocket of ultraspiracle, the Drosophila RXR ortholog [15–17], has been stymied in part by difficulty in demonstrating specific binding of a test compound to the purified receptor and that such binding then induces conformational changes in the receptor. Indeed, the current paradigm expressed in most published models for USP function is that USP does not bind to any ligand in exerting its regulatory functions [18, Fig. 1; 19, Fig. 8; 20, Fig. 8; 21, Fig. 3B; 22, Fig. 4; 6, 23, Fig. 8; 24, 25, 26, Fig. 8]. A demonstration that endogenous USP can become transcriptionally activated upon binding to an agonist would have major implications for the current paradigms of hormone action in invertebrates.

The orthology between invertebrate USP and vertebrate RXR offers the possibility that the USP ligand-binding pocket may be conformed so as to be susceptible to binding and transcriptional activation by a terpenoid-related ligand [27]. In a previous report, we observed that methyl epoxyfarnesoate (juvenile hormone III) appeared to bind to USP in biochemical assay, and application of methyl epoxyfarnesoate to cells activated a transfected reporter construct containing direct repeat elements to which recombinant USP bound in gel shift assay [28]. However, these indirect experiments did not address whether methyl epoxyfarnesoate actually binds to the ligand-binding pocket of the receptor, nor whether endogenous USP in the transfected cells actually binds to the direct repeat elements, nor do they address whether methyl epoxyfarnesoate-activation of the reporter is dependent upon liganded USP, all of which are crucial underpinnings to the concept of the USP ligand-binding pocket as a viable target for experimental or practical agonistic or antagonistic ligands. In the present report we demonstrate a functional transcriptional outcome of occupancy of the ligand-binding pocket of the nuclear receptor ultraspiracle.

Materials and methods

Cell culture and transfections

Spodoptera frugiperda cell line, Sf9, was maintained and transfected as described previously [29,30]. As an internal control to compare activities of different constructs, 0.3 µg of a constituitive heat-shock promoter-driven β-galactosidase gene was cotransfected. To study the role of USP in activation of the reporter promoter in methyl epoxyfarnesoate-treated cells, cloned D. melangaster USP (dUSP) cDNA and its derivatives containing mutations in the ligand-binding pocket were cotransfected with the reporter and internal control plasmids. At 36 h after the transfection, the cells were treated with 75 µm methyl epoxyfarnesoate (Sigma) in ethanol carrier (1% final ethanol concentration) or just ethanol carrier only (previous studies demonstrated methyl epoxyfarnesoate effects were dose dependent, with maximum near 75 µm[28]). 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). β-Galactosidase activity was measured using chlorophenol red-α-d-galactopyranoside monosodium (CPRG; Roche Molecular Biochemicals) as a colorimetric substrate.

Plasmid constructs

The sequences and characteristics of the core promoter (−61 to +28) of the JHE gene were described by Jones et al. [28,29], and it was previously verified to respond to methyl epoxyfarnesoate through a heterologous 5′ flanking direct repeat motif in cell transfection assay [28]. This Core promoter reporter was cloned into KpnI/BglII sites of pGL3. An NheI site was then placed immediately 5′ to the KpnI site, and multiple direct repeat (DR) sequences were cloned into the NheI site by the following method. Complementary oligonucleotides encoding the particular DR motif were synthesized, with each oligonucleotide possessing at its 5′ end a 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. Specific DR sequences for the oligonucleotides were (upper strand) for DR1: 5′-CAAGGTCAAAGGTCAG-3′, for DR4: 5′-CAAGGTCAAGAAAGGTCAG-3′, for DR12: 5′-CAAGGTCAAGAAGGCCAAAGAGGTCAG-3′ (repeat motif underlined; CTAG on 5′ ends not shown). The recovered YDRXCore constructs (X representing 1, 4 or 12 intervening bases; Y representing the number of tandem pairs of direct repeats) were verified by sequencing. The intervening sequences in the DR1 and DR4 motifs were randomly chosen, while the DR12 sequence used is found in the ecdysteroid-sensitive ng-1 and ng-2 genes that are expressed during metamorphosis of D. melanogaster, and can serve in vitro as a binding site for the various receptor dimers involving USP (ecdysteroid receptor (EcR)/USP heterodimer, USP/DHR38 heterodimer and USP/USP homodimer [28,31,32]).

Point mutations in the ligand-binding domain of dUSP were made with a ChameleonTM double-stranded site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. The selection primer used to change the unique NdeI (underlined) site in the pIE1-4 vector was CGGTATTTCACACCGCAcATGGTGCACTCTCAGTACAATC. The primer to mutate Q288 to alanine in the ligand-binding pocket was: GTGCCAAGTGGTCAACAAAgcGCTCTTCCAGATGGTCGAATAC. A primer that targeted two amino acids was used to make the double mutation in C473A and H476L because of their adjacent locations, with the sequence: GCGATCGATCAGCCTGAAGgcCCAGGATCtCCTGTTCCTCTTCCGCATTAC. A primer that replaced two proline residues (P498, P499) at the end of α-helix 12 with tryptophan residues was: 5′-CTTTCTCGAGCAGCTGGAGGCGtgGtgGCCACCCGGCCTGGCGATGAAACT-3′. All mutant constructs were confirmed by DNA sequencing.

For expressing dUSP in Sf9 cells, PCR-generated full-length wild-type and point-mutated dUSP coding sequences were cloned into PmeI and NotI sites of the pIE1-4 vector (Novagen) and confirmed by sequencing, and for bacterial overexpression were cloned into pET32EK (Novagen).

Nuclear extracts and electrophoretic mobility shift assay

Nuclear extracts were isolated from Sf9 cells as previously described [28,29]. For the DR12 probe, the double stranded DR12 oligonucleotide (sequence as shown above) was 5′ end-labelled with 32P by T4 polynucleotide kinase (New England Biolabs Inc.), and then purified from a 20% native polyacrylamide gel. The same double stranded DR12 oligonucleotide was used in 100-fold excess as a self competitor. For the 4DR12Core probe, the 4DR12Core sequence was liberated from the vector as a 148-bp ClaI/HindIII fragment, and was 5′ end-labelled with 32P and purified. The same, unlabelled fragment was used at 100-fold excess as a self competitor. As a negative control for specificity in gel shifts, the 36 bp BglII/KpnI polylinker region fragment of the pGL3 vector was liberated and recovered from low melting point agarose gels and used as a 100× nonself competitor (sequence: GGTAC CGAGCTCTTACGCGTGCTAGCCCGGGCTCGA). Either a final concentration of 500 nm of His-tagged wild-type USP or His-tagged mutant Cys472Ala/His475Leu (= C472A/H475L), or five micrograms of nuclear proteins, were incubated with the given probe on ice for 30 min in binding buffer (10 mm Tris/HCl, pH 7.5; 50 mm NaCl, 0.5 mm EDTA, 5% glycerol, 1 mm MgCl2, and 1 mm dithiothreitol). In some experiments, nuclear proteins were preincubated with the probe for 30 min followed by incubation with anti-USP mAb (a gift from F. Kafatos, EMBL, Heidelberg), or monoclonal Elav antibody (Developmental Studies Hybridoma Bank, University of Iowa), for an additional 1 h on ice. Samples were then subjected to 4% (w/v) polyacrylamide gel electrophoresis in 0.5 × Tris/borate/EDTA buffer. After electrophoresis, the gels were dried and exposed to Kodak film at −70 °C for 12–48 h.

Extraction of total proteins and immunoblotting analysis

Total Sf9 cell protein extracts from transfected Sf9 cells were fractionated by SDS/PAGE, 8% (w/v) polyacrylamide gel, and then transferred onto a nitrocellulose membrane. USP was detected using a primary USP AB11 monoclonal antibody and with an anti-mouse IgG-AP secondary Ig (Bio-Rad) by a BCIP/NBT color development solution (Bio-Rad). The USP signals were normalized by an internal control, α-actin, which was detected by a primary polyclonal α-actin antibody (Sigma) and with an anti-rabbit IgG-AP secondary Ig (Southern Biotechnology Associates, Inc.).

Purification of the His–USP fusion protein and ligand-binding assay

The homodimer-enriched fraction of bacterial recombinant His–dUSP fusion protein was purified by nickel resin selection, elution with imidazole, centrifugal concentration, and then gel permeation chromatography (Superdex 200) with procedures and chemical sources exactly as already described previously [27]. The homodimer-enriched fraction of the purified His–USP fusion protein was raised to 2 mL of NaCl/Pi and a final concentration of 0.5 µm. For a fluorescence-based ligand-binding assay based on intrinsic tryptophan fluorescence [28,33], ligand or ethanol carrier was added and the receptor preparation excited at 290 nm and monitored for emission at 340 nm, until the signal from the receptor had stabilized. Fluorescence was measured three times for each sample, with standard deviation typically smaller than the graphical plotted datum point. Each fluorescence experiment was replicated on three or more independent occasions, each time with similar results.

Modelling of hRXRα and D. melanogaster USP

Tertiary conformation of human RXRα and D. melanogaster USP was analyzed by rasmol software, using the coordinates reported into the Protein Data Bank by Bourguet and Moras (deposition number 1LB) and by Schwabe and Clayton (deposition number 1HG4), respectively. Using a minimum energy conformation of farnesol as a scaffold, a conformation of epoxyfarnesoic acid was prepared and placed by hand into the ligand-binding pocket of USP along a generally similar trace as was reported for the (more bent) 9-cis retinoic acid ligand when the latter was cocrystalized with hRXRα (Egea et al. [52]).

Results

Placement of four tandem copies of a DR12 motif (CAAGGTCA(N)12AGGTCAG, Fig. 1A) at 5′ to the Core promoter reporter (4DR12Core construct, Fig. 1B) yielded a 10-fold induction in promoter activity in response to treatment of the transfected Sf9 cells with methyl epoxyfarnesoate (Fig. 1B). In contrast, insertion of a cassette containing four tandem copies of either a DR1 or DR4 motif yielded only a 2.5- and 3.5-fold induction, respectively (Fig. 1B). This differential result confirms that the 10-fold activation observed with the 4DR12Core construct was caused by the sequence of the inserted DR12 cassette itself, and was not due to either insertional disruption or creation of a putative cryptic regulatory element at the vector multiple cloning site. Due to the highest reporter activity being obtained with the DR12 motif, we focussed on the DR12 repeat construct, towards the goal of the study of ligand activation of USP.

Figure 1.

Activation of transfected Core promoter reporter through DR12 enhancer. (A) The sequences of single copies of DR1, DR4, DR12 and mutant DR12 enhancer motifs used in the promoter constructs. Each half site is dashed underlined. Mutated residues are shown in lower case letters. (B) On the left are the designs of the vector construct encoding the luciferase reporter enzyme, of the vector construct containing the Core promoter reporter, and of the three vector constructs in which the Core promoter is preceded by four tandem copies either a DR1-, a DR4- or a DR12-based enhancer, with the orientation of each motif shown by the small arrows. On the right are the activations of the indicated promoter reporter construct in response to treatment of transfected cells with 75 µm methyl epoxyfarnesoate.

We then confirmed that sequences in the AGGTCA half sites themselves of the DR12 motif were necessary for transducing the methyl epoxyfarnesoate signalling. We took advantage of the previous report that mutation of each half site abrogated the ability of DR12 motif to enhance ecdysteroid transcriptional activation [31]. When we mutated here each half site of the DR12 motif (in a construct containing a single DR12 in order to simplify mutational analysis; 1DR12mutCore), the responsiveness of the 1DR12mutCore to methyl epoxyfarnesoate was no greater than the background of a Core promoter with no enhancer (Fig. 2B); in contrast to the responsiveness of the Core promoter in the presence of a wild-type DR12 (1DR12Core, Fig. 2B). As an independent confirmation of the important role of the two direct repeat half sites in the DR12 motif, we demonstrated that in a gel mobility shift assay with Sf9 nuclear extracts, the DR12 motif probe yielded a shifted probe band that could be competed with excess, unlabelled wild-type DR12. However, the same DR12 mutated in its two half sites that had failed to support methyl epoxyfarnesoate-enhanced transcription in the cell transfection assay also correspondingly failed to compete with the wild-type DR12 probe in the gel shift assay (Fig. 2A), confirming the functional necessity of the two half sites for interaction with a nuclear component(s). Thus, the lack of binding to the mutant DR12 combined with the lack of a transcriptional effect of that same mutant DR12 suggests that the specific binding to the wild-type DR12 observed here relates to its positive action to transduce the methyl epoxyfarnesoate signalling observed in the transfection assay. The gel mobility shift assay using Sf9 nuclear extracts detected a single major complex binding to the DR12 probe (Fig. 2C). An anti-dUSP mAb (AB11, epitope on DNA binding domain) displaced the endogenous USP in the major complex binding to the DR12 probe (Fig. 2C). The specificity of the AB11 monoclonal antibody effect on USP binding was further confirmed in that no such effect was produced by a negative control monoclonal antibody against the transcription factor Elav.

Figure 2.

Functional analysis of the DR12 motif. (A) Gel mobility shift assay, using Sf9 nuclear extracts, of the same single DR12 motif that was used as an enhancer in the cell transfection assay in (B). The shifted probe band was competitively displaced by 100× of the unlabelled DR12 motif (self), but was not competed with either by the same mutant DR12 motif as failed to act as an enhancer in cell transfection assay in B (mutDR12) or by the negative control polylinker sequence (nonself). (B) Activations of the indicated promoter reporter constructs in response to treatment of transfected cells with 75 µm methyl epoxyfarnesoate. (C) Intracellular USP binds to DR12 hormone response element. Gel mobility shift assay using Sf9 nuclear extracts (N.E.) and a 32P-labelled probe that is the four tandem DR12 motifs (‘4DR12’) shown in Fig. 1, performed as described in [28]. The USP in the Sf9 nuclear extract that is the major binding complex (small arrow) is displaced by the AB11 monoclonal antibody, just as we have previously shown is the effect of this antibody on recombinant dUSP binding to a DR12 probe [28]. The lack of similar effect by monoclonal antibody against the negative control nerve transcription factor (Elav) shows the specificity of the AB11 result.

In many vertebrate nuclear receptors, residues in a narrow region of α-helix 11 make contact with the endogenous ligand of the given receptor [27]. In hRXRα, Cys432 and His435 on α-helix 11 make contact with the distal end of the 9-cis RA ligand at two methyl branches (C16, C17) and also at the terpene backbone (Fig. 3A,B). The homologous two residues on α-helix 11 of dUSP (Cys472 and His475) are highly conserved in other USPs [27], and point into the ligand-binding pocket of USP crystal structures (Fig. 3C[34,35]). Other residues that contact 9-cis RA in hRXRα are also conserved in identity and similar location in the ligand-binding pocket of dUSP, such as Gln288, Trp318 and Leu367 in dUSP corresponding to Gln275, Try305 and Leu326 of hRXRα (Fig. 3B–D). If an epoxy farnesoid-like ligand were to reside in the dUSP ligand-binding pocket along a similar trace as does 9-cis RA in hRXRα, then the terpene backbone and the methyl branches C12 and C15 at the distal end of the epoxy farnesoid ligand might be similarly placed to interact with His475 and Cys472 in dUSP, as does 9-cis RA interact with Cys432 and His435 in hRXRα (Fig. 3B–D).

Figure 3.

Comparision of dUSP and hRXR ligand binding domains. (A) Selected contacts made between 9-cis RA and residues in the hRXRα ligand-binding pocket as determined from cocrystals (4.2 Å or less, from Doyle et al. [55] and Egea et al. [53]). On the left is also shown a conformation of epoxyfarnesoic acid, exhibiting similarities between its structure and that of the terpenoid backbone and carboxyl group of 9-cis retinoic acid. (B and C) rasmol-generated ribbon diagrams for the ligand-binding domains of the hRXRα and dUSP, respectively. (B) This shows in the hRXRα ligand-binding pocket the structure of the ligand 9-cis retinoic acid (carbon backbone in light blue, terminal carboxylate oxygens in dark blue, adapted from Egea et al. [52]). (C) This shows methyl epoxyfarnesoic acid (yellow carbon backbone and blue terminal carboxylate oxygens) lain manually in the dUSP ligand-binding pocket with the carboxy and distal (epoxy) ends, respectively, situated in similar regions of the pocket as the carboxyl end and distal end of 9-cis RA in hRXRα. (D) An overlay of the dUSP and hRXRα ribbon diagrams of B and C, with emphasis (white arrows) on the similar placement of Gln275,Trp305, Leu326, Cys432 and His435 in hRXRα as compared to Gln288, Trp318, Leu367, Cys472 and His475 in dUSP.

We tested this model by overexpressing the His-tagged dUSP double mutant Cys472Ala/His475Leu (C472A/H475L) in methyl epoxyfarnesoate-treated Sf9 cells that were cotransfected with the 4DR12Core reporter plasmid. Cells transfected with either empty pIE1-4 vector, or that vector expressing wild-type dUSP, responded to methyl epoxyfarnesoate application with a similar induction of the 4DR12Core promoter (Fig. 4A). However, cells transfected with the plasmid expressing the C472A/H475L mutant exhibited a distinct suppression in the level of methyl epoxyfarnesoate-induced activation, as compared with the activation observed for cells transfected with either the empty plasmid or plasmid expressing wild-type dUSP (Fig. 4A). In addition, cotransfection of the empty vector, or vector expressing either wild-type dUSP or the C472A/H475L mutant, did not affect the basal activation exhibited when the Core promoter without DR12 motifs was used. Together, these data demonstrate that the suppression in methyl epoxyfarnesoate-induced activation caused by overexpression of the C472A/H475L double mutant was not due to nonspecific titration of coactivators required by a receptor other than USP and not due to disruption of Core-binding basal transcription components independent of action through the DR12 enhancer. In addition, overexpression of either the C472A/H475L double mutant or the wild-type dUSP did not change the level of endogenous USP (Fig. 4A), confirming that overexpression of exogenous dUSP did not indirectly affect the methyl epoxyfarnesoate-activation pathway by disruption of endogenous USP expression.

Figure 4.

Dominant negative activity of USP ligand-binding pocket mutants. (A) Histogram (shaded boxes) shows the dominant negative effect of transfected dUSP mutant and the double mutant (C472A/H475L) on methyl epoxyfarnesoate-activation of 4DR12Core reporter promoters, whereas transfected wild-type dUSP shows no such suppression of methyl epoxyfarnesoate activation, in comparison with transfection of Core reporter vector (reporter and expression plasmids transfected at 1 : 1 ratio). Transfection of neither the wild-type USP nor either mutant had any effect on the minimal basal activation of the Core promoter in the absence of the DR12 motif (clear boxes). Immunoblot of transfected cellular extracts with anti-(α-actin) and anti-dUSP (AB11) mAbs verified that the overexpression of mutant and wild-type dUSP did not affect the level of expression of endogenous USP, and that the transfected mutant and transfected wild-type dUSP were expressed at similar levels to each other. The molecular weights of the transfected and endogenous USPs detected by immunoblotting were ≈ 50 and 52 kDa, respectively, as estimated by molecular size standards run in parallel lanes (not shown). (B) Progressive increase in ratio of transfected dominant negative plasmid DNA relative to 4DR12Core reporter plasmid DNA yielded an increasing dominant negative suppression of methyl epoxyfarnesoate activation of reporter plasmid. Immunoblot verifies that the progressively higher overexpression of the mutant dUSP (C472A/H475L) did not affect the level of expression of endogenous USP. Inset above shows calculation of transcriptional activation ratio of reporter promoter activity in methyl epoxyfarnesoate- treated cells relative to EtOH treated cells, as a function of the ratio of the amount of transfected mutant dUSP plasmid relative to amount of transfected reporter plasmid. (C) Transfection of plasmid expressing wild-type USP rescues the dominant negative-suppression of methyl epoxyfarnesoate-activation of the reporter promoter. Open circle, methyl epoxyfarnesoate activation of 4DR12Core in the absence of USP expressing plasmid. Hashed circle, methyl epoxyfarnesoate activation is suppressed by transfection with the C472A/H475L dominant negative mutant. Filled circles, methyl epoxyfarnesoate activation is progressively restored by increasing doses of plasmid expressing wild-type dUSP. In A–C, hormone-treated cells received 75 µm of methyl epoxyfarnesoate.

Concerning the proximal end of the hRXRα ligand, cocrystals of 9-cis RA and hRXRα have also established that a glutamine residue on α-helix 3 (Gln275) makes contact with both the carbonyl carbon and a carboxylate oxygen (Figs 3A,B and 5). This glutamine residue is conserved in all reported USPs (Fig. 5C[27]). Therefore, we mutated this Gln288 in dUSP to alanine (Gln288A), and found that this mutant dUSP also acted as a dominant negative suppressor of activation of the DR12Core reporter promoter in methyl epoxyfarnesoate-treated Sf9 cells (Fig. 4A).

Figure 5.

Bacterially overexpressed double mutant dUSP (C472A/H475L) and wild-type dUSP analyzed for binding to DNA or to ligand. (A) The wild-type dUSP and the C472AH475L mutant both similarly bound in part as a homodimer (upper band) and in part as a monomer (lower band) to a 4DR12 motif probe (identification of monomer and homodimer bands was made by comparative analysis of binding by other dimer-enriched vs. monomer-enriched fractions obtained from Superdex 200 chromatography, not shown). Control competitions with self and nonself unlabelled excess probes confirmed the specificity of binding. The similar formation of the homodimer form by the wild-type USP and mutant USP, along with the similar binding to DNA of the wild-type USP and mutant USP, confirm that the mutation to the ligand-binding pocket in C475A/H475L did not generally disrupt the structure of the receptor. (B) The homodimer-enriched fraction of each receptor preparation was then analyzed for binding to 75 µm methyl epoxyfarnesoate, using an intrinsic fluorescence assay method that tracks ligand binding (by suppression in receptor fluorescence) [27,28]. The wild-type dUSP exhibited binding to methyl epoxyfarnesoate in this assay. However, the double mutant dUSP exhibited no binding activity. Arrows show time of addition of methyl epoxyfarnesoate or EtOH carrier.

Under the model that overexpression of the C472A/H475L double mutant competed with endogenous USP in the pathways for transduction of the exogenous methyl epoxyfarnesoate signal, the level of effect of the double mutant ought to be dependent on its dose. Indeed, we determined that a progressive increase in the intracellular concentration of this double mutant (with endogenous USP level remaining unchanged) caused progressive suppression in the methyl epoxyfarnesoate-activation of the DR12Core promoter, down to the transcriptional level observed for the Core promoter without DR12 enhancers (Fig. 4B). Over the range of the progressive suppression of the methyl epoxyfarnesoate-activated transcription there was no effect of the double mutant on the basal level of transcription in EtOH-treated controls. We then used this background of the blocked activation pathway to test whether activation by methyl epoxyfarnesoate treatment was actually dependent on the presence of wild-type USP. As shown in Fig. 4C, the activation of the 4DR12Core promoter in methyl epoxyfarnesoate-treated cells was monotonically restored in a manner dependent on the increasing dose of the added wild-type dUSP. Again, over the range of the monotonic restoration of methyl epoxyfarnesoate-activated transcription, there was no effect of the transfected wild-type dUSP on the basal level of transcription in EtOH-treated controls.

We examined the ability of the C472A/H475L mutant to bind DNA and to homodimerize to confirm that the mutations to the ligand-binding pocket did not generally deform receptor structure. As shown in Fig. 5A, under electrophoretic mobility shift assay conditions, both the wild-type dUSP and the C472A/H475L mutant dUSP similarly bound to a DR12 motif. In addition, both receptor preparations bound to the probe similarly in part as monomer and in part as homodimer. The homodimerization of RXR and other steroid receptor superfamily members is primarily due to contacts in the ligand-binding domain that are outside of the ligand-binding pocket (in addition to some contacts also in the DNA-binding domain). The similar DNA binding and homodimerization capacities of the wild-type dUSP and mutant C472A/H475L dUSP is strongly indicative that the DNA-binding domain, and the parts of the ligand-binding domain that are outside of the ligand-binding pocket, are in a functionally similar conformation for both the wild-type and mutant receptors. Thus, any difference detected in ligand binding of the two receptors is most reasonably inferred as arising from differences in the architecture inside the cavity of the ligand-binding pocket due to the C472A/H475L point mutations.

We then tested the ability of the wild-type dUSP and dominant negative, ligand-binding pocket mutant dUSP to bind methyl epoxyfarnesoate. In a ligand-binding assay that detects methyl epoxyfarnesoate binding through its effects to suppress intrinsic fluorescence of dUSP [28,33], the bacterially overexpressed His-tagged wild-type dUSP indeed exhibited suppressed the fluorescence due to the binding of methyl epoxyfarnesoate (Fig. 5B). However, the C472A/H475L mutant dUSP did not exhibit a significant response to epoxymethyl farnsoate (Fig. 5B). This result was reproduced with independent preparations of the wild-type dUSP and C472/H475L dUSP. These results strongly support the inference that the behavior of C472A/H475L as a dominant negative mutant in the pathway for methyl epoxyfarnesoate activation of the 4DR12Core promoter is due to the effect of the C472A/H475L mutations on the ligand-binding activity of USP.

Some models of nuclear hormone receptor action include the component that binding of ligand to the ligand-binding pocket induces a tertiary conformational change involving the movement of α-helix 12 to a new position [36]. However, the two published crystal structures of USP in complex with a phospholipid located at the opening of the ligand-binding pocket show α-helix 12 in a position that the investigators described as so firmly ‘locked’ against other residues of the ligand-binding domain that α-helix 12 would not be able to move even if the phospholipid were not present [33,34]. We therefore tested the hypothesis that α-helix 12 is so firmly locked in position that it does not move, by replacing two of the four continuous proline residues at the end of α-helix 12 with tryptophan residues. Under the model that USP α-helix 12 does not move upon binding of methyl epoxyfarnesoate in the ligand-binding pocket, these two tryptophan residues would only raise the constant background intrinsic fluorescence of the receptor, but, on account of the fact that they (as part of the fixed α-helix 12) do not move in position, their level of fluorescence would not change upon binding of methyl epoxyfarnesoate into the pocket. Therefore, their constant background fluorescence would not enhance or disguise the suppression in fluorescence exhibited by the two other natural tryptophan residues (on α-helix 5) upon binding of methyl epoxyfarnesoate. Alternatively, if α-helix 12 does move in position upon binding of methyl epoxyfarnesoate, then the change in the local environment of the two added tryptophan residues on α-helix 12 may change their fluorescence in a way that yields a markedly different overall fluorescence pattern for the receptor. Indeed, as Fig. 6B shows, in this test the wild-type USP with only two natural tryptophan resides on α-helix 5 exhibits a distinct suppression in fluorescence upon binding of methyl epoxyfarnesoate. In contrast, the mutant USP containing two additional tryptophan residues at the end of α-helix 12 showed a much different profile, instead sharply increasing in fluorescence before then decreasing (Panel C). Collectively, these markedly different patterns of fluorescent response are most easily explained by a model in which α-helix 12 does move in relative position, upon the binding of methyl epoxyfarnesoate into the ligand-binding pocket of USP.

Figure 6.

Fluorescence response of wild-type and P498W/P499W mutant USP to farnesoid ligands. (A) The location of the mutational placement of the two tryptophan residues at the end of (red colored) α-helix 12. USP also possesses two natural tryptophan residues on helix 5 (W318, shown in green extending into pocket; W328, not shown, extending out of pocket). (B) Methyl epoxyfarnesoate binding to wild-type USP results in suppression of receptor fluorescence, while farnesol and ethanol carrier do not have that effect. (C) Methyl epoxyfarnesoate binding to P498W/P499W mutant results in a very different pattern of fluorescence response than wild-type USP in B, evidencing that α-helix 12 moves in its relative location upon USP binding of methyl epoxyfarnesoate. The wild-type USP and P498W/P499W similarly bound in part as monomer and in part as dimer to a DR12 probe in gel shift assay, evidencing that the P498W/P499W mutations did not affect receptor structure globally (not shown).

Discussion

With the inception of the original model by Ashburner on hierarchical, steroid-driven genetic programs for invertebrate development [37], the sophistication of the models has progressively increased as more transcription factors have been discovered to participate in these complex developmental programs [38]. However, despite the inclusion of ultraspiracle in these conceptual models since its discovery over 10 years ago, there has been much angst over whether this receptor possesses a ligand-binding activity. In the absence of an experimental demonstration that ultraspiracle can bind ligand and transduce that binding into transcriptional modulation, models of genetic programs that include ultraspiracle have not overtly included a ligand-binding role for ultraspiracle [18–26,39,40]. While there is genetic evidence that the ligand-binding domain of USP globally contributes to function of the EcR/USP heterodimer [41], other models expressly envision that ultraspiracle does not have any ligand-binding role in certain pathways [23,42].

We have previously demonstrated [28,33] that dUSP can specifically bind to small terpenoid-derived compounds such as epoxy methyl farnesoate and bisepoxy methylfarnesoate, in a saturable, dose-dependent manner, causing a conformational change to the receptor that suppresses its intrinsic fluorescence, while compounds such as farnesol and epoxyfarnesoic acid, and the steroid 20-OH ecdysone do not have this effect. We have also shown elsewhere [28] that the marked increase in transcription of the model DR12Core reporter promoter, with methyl epoxyfarnesoate (Fig. 1), is dose-dependent, but that neither retinoic acid nor T3 yield this effect. However, these previous results do not demonstrate whether methyl epoxyfarnesoate binds to the receptor in its ligand-binding pocket, nor whether such binding induces movement in α-helix 12, nor whether endogenous USP in the transfected cells can bind to the direct repeat motifs that 5′ flank the reporter promoter, nor do they address whether methyl epoxyfarnesoate-activation of the reporter is dependent upon liganded USP, all of which are crucial underpinnings to the concept that the USP ligand-binding pocket is a viable target for experimental or practical agonistic or antagonistic ligands.

In the present report, we have demonstrated that methyl epoxyfarnesoate does indeed bind to the ligand-binding pocket, and that point mutations to the dUSP ligand-binding pocket that disrupt methyl epoxyfarnesoate binding cause the mutant receptor to act as a dominant negative in a model transcription pathway that is activated by methyl epoxyfarnesoate treatment. These data suggest further inquiry is warranted into farnesoid-derived ligands as agonists for USP. Our demonstration here that the USP ligand-binding pocket is conformed such that it can bind methyl epoxyfarnesoate-like compounds, with a resultant change in USP conformation, including the movement of α-helix 12, and with an effect to activate transcription in methyl epoxyfarnesoate-treated cells is the first such identification of the activating binding of any compound, natural or synthetic, to the ligand-binding pocket of an invertebrate orphan receptor. This precedent establishes that invertebrate orphan receptors are not qualitatively different from the situation for vertebrate orphan receptors for which a number have now been shown to have ligand-binding pockets with the functional capacity to bind and be transcriptionally activated by appropriately structured compounds.

USPs, which compared to RXR are unusual for their stretch of additional amino acids inserted after α-helix 5, have recently been cocrystalized with fortuitous phospholipid pseudoligands [34,35]. These cocrystals had a relatively large total van der Waals volume of the USP ligand-binding pocket (≈ 1300 Å3), compared to the volume of JH III (259 Å3[43]). However, the volume of the PPARγ ligand-binding pocket (similar to that of USP, ≈ 1300 Å3[44]) is also much larger than that of its natural ligand 15-deoxy-Δ12,14-prostaglandin J2 (which has a volume similar to that of JH III, at 301 Å3[43]), yet this prostaglandin ligand is able to bind and transcriptionally activate the PPARγ[45]. In addition, the volume of β-estradiol (which at 245–251 Å3 is smaller than methyl epoxyfarnesoate [43]), is approximately half the volume of the ligand-binding pocket of the estrogen receptor (450–500 Å3[46,47]), Yet, β-estradiol is nonetheless able to bind to and activate the estrogen receptor. Thus, PPARγ and the estrogen receptor demonstrate that endogenous compounds much smaller than the total ligand-binding pocket volume of a nuclear hormone receptor can and do serve as natural activating ligands. The recently crystallized PXR, which binds with, and is activated by, a variety of small and large ligands, also possesses a large 1300 Å3 ligand-binding pocket [48], and possesses an unusual additional stretch of amino acids that the authors postulated enables what would otherwise be a smaller PXR ligand-binding pocket to enlarge to accommodate a large ligand. Important in these considerations is whether there is a subregion in the ligand-binding pocket in which the local conformation corresponds well to the conformation of a particular small ligand. Although the overall volume of the ligand-binding pocket observed in the cocrystals of USP (≈ 1300 Å3) is much larger than that of hRXRa (≈ 500 Å3), the proximal subregion of the ligand-binding pocket of hRXRα and USP are much more similar in volume and shape [34]. The proximal subregion of each of the two receptors also has a similar placement of conserved amino acids that in hRXRα interact with the terpenoid backbone of 9-cis RA (Fig. 4A–D). In addition, 9-cis RA and methyl epoxyfarnesoate have similar van der Waals volumes of 291 and 258 Å3, respectively [43]. These considerations suggest that methyl epoxyfarnesoate-like metabolites cannot be dismissed a priori as potential USP agonists, merely on the basis of comparison of the volume of methyl epoxyfarnesoate vs. the reported total volume of the USP ligand-binding pocket.

Our combined use of an equilibrium, fluorescence binding assay and a transfection transcriptional assay that is activated by treatment with methyl epoxyfarnesoate will be very useful in identifying new, higher-affinity ligands for USP. The molecular interactions between a receptor and a synthetic activating ligand have previously provided insight to the molecular basis by which agonist ligand(s) activates the receptor. Crystal structures of the vitamin D receptor in complex with natural activating ligand vs. with synthetic agonists revealed that both induced the same intramolecular conformational changes in the receptor [49]. Cocrystal structure analysis showed that human RARα was induced to undergo similar intramolecular conformational changes by either natural 9-cis RA or a synthetic agonist [50]. We have shown that binding of methyl epoxyfarnesoate by dUSP promotes not only an intramolecular conformational change of movement of α-helix 12, but also homodimerization [28], which together appears reminiscent of the way in which 9-cis RA induces an activating intramolecular conformational change in human RXRα (e.g. movement of α-helix 12) as well as that receptor's homodimerization [51–53]. These results have considerable significance for current popular models of USP function as a heterodimeric partner with EcR, because most of these models do not envision the binding effect of an agonist by USP.

It is becoming increasingly appreciated that not all core promoters are alike in their ability to respond to the same transcriptional enhancer, as additional DNA sequence in and around the TATA box and initiator motifs confer selectivity in the nature of the components that nucleate to form the basal transcription apparatus at the core promoter. Indeed, a number of different parameters have been identified under which different EcR/USP heterodimer DNA binding sites exert very different levels of effect in transducing ecdysteroid signalling [54–57]. Therefore, we do not anticipate that the DR12 motif used here will necessarily function to enhance the activity of all model core promoters in response to methyl epoxyfarnesoate-like molecules. However, it is clear that this model system of the DR12Core promoter in Sf9 cells will be appropriate and useful as a tool in exploring the functional structure of the ligand-binding pocket of USP with respect to USP activation upon binding of methyl epoxyfarnesoate and other agonistic compounds.

Acknowledgements

The research reported herein was supported, in part, by NIH grants 462795 and 463713. We express our appreciation to Drs David Mangelsdorf, Carl Thummel and Mietek Wozniak for helpful discussions on the framing of hypotheses on the functional structure of ligand-activated nuclear receptors.

Ancillary