The influence of heterodimer partner ultraspiracle/retinoid X receptor on the function of ecdysone receptor Subba R. Palli 1 , Mariana Z. Kapitskaya 2 and David W. Potter 2 1 Department of Entomology, University of Kentucky, Lexington, KY, USA 2 RheoGene Inc. Norristown, PA, USA Steroid hormones, ecdysteroids, regulate insect develop- ment, reproduction and several other physiological processes. The most active form of ecdysteroids is 20-hy- droxyecdysone (20E). The 20E transduces its signal through a heterodimeric complex of two nuclear recep- tors, the ecdysone receptor (EcR) [1] and the ultraspira- cle (USP), an ortholog of the vertebrate retinoid X receptor (RXR) [2–4]. Both EcR and USP are members of the nuclear receptor superfamily [5] and exhibit a typ- ical modular structure comprising the N-terminal A ⁄ B domain, the DNA-binding or C domain, the hinge or D domain, the ligand-binding or E domain, and the C-terminal F domain. The ligand-binding domain sup- ports ligand-dependent dimerization and transactivation functions. A ⁄ B and F domains support ligand-inde- pendent transactivation. The DNA-binding domain and the N-terminal region of the hinge region are known to support dimerization of two receptors. The EcR:USP heterodimers bind to the ecdysteroid response elements (EcRE) present in the promoter regions of ecdysteroid response genes and regulate their transcription. Most of the nuclear hormone receptors, including EcR, are modular and function as ligand-controlled transcription factors, a characteristic that renders these receptors or their key regions (e.g. the ligand-binding domain) suitable for gene switches Keywords gene switch; ligand-binding domain; nuclear receptor; steroid hormone Correspondence S. R. Palli, Department of Entomology, S225 Agricultural Science Center, College of Agriculture, University of Kentucky, Lexington, KY 40546, USA E-mail: rpalli@uky.edu (Received 14 July 2005, revised 26 August 2005, accepted 3 October 2005) doi:10.1111/j.1742-4658.2005.05003.x A pair of nuclear receptors, ecdysone receptor (EcR) and ultraspiracle (USP), heterodimerize and transduce ecdysteroid signals. The EcR and its nonsteroidal ligands are being developed for regulation of transgene expres- sion in humans, animals and plants. In mammalian cells, EcR:USP heterodimers can function in the absence of ligand, but EcR ⁄ retinoid X receptor (EcR:RXR) heterodimers require the presence of ligand for activa- tion. The heterodimer partner of EcR can influence ligand sensitivity of EcR so that the EcR ⁄ Locusta migratoria RXR (EcR:LmRXR) heterodi- mers are activated at lower concentrations of ligand when compared with the concentrations of ligand required for the activation of EcR ⁄ Homo sapiens RXR (EcR:HsRXR) heterodimers. Analysis of chimeric RXRs con- taining regions of LmRXR and HsRXR and point mutants of HsRXR showed that the amino acid residues present in helix 9 and in the two loops on either end of helix 9 are responsible for improved activity of LmRXR. The EcR:Lm-HsRXR chimera heterodimer induced reporter genes with nanomolar concentration of ligand compared with the micromolar concen- tration of ligand required for activating the EcR:HsRXR heterodimer. The EcR:Lm-HsRXR chimera heterodimer, but not the EcR:HsRXR hetero- dimer, supported ligand-dependent induction of reporter gene in a C57BL ⁄ 6 mouse model. Abbreviations CfEcR, Choristoneura fumiferana EcR; DMSO, dimethylsulfoxide; 20E, 20-hydroxyecdysone; ECD, ecdysteroid; EcR, ecdysone receptor; EcRE, ecdysone response element; G:CfEcR(DEF), GAL4:CfEcR(DEF); RLU, relative light units; RXR, retinoid X receptor; SEAP, secreted alkaline phosphatase; VP:Hs–LmRXR(EF), VP16:HsbRXR (helices 1–8) LmRXR (helices 9–12 + F); V:MmRXR(EF), VP16:MmRXR(EF); V:CfUSP(EF), VP16:CfUSP(EF); V:LmRXR(EF), VP16:LmRXR(EF); USP, ultraspiracle; WT, wild-type. FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS 5979 (ligand-dependent regulation of transgenes) in various biotechnology applications. Several nuclear receptors, including the glucocorticoid receptor (GR), the pro- gesterone receptor (PR), the estrogen receptor (ER), and the EcR are being used to develop gene switches for applications in medicine and agriculture. Because the EcR and its ligands are not found in vertebrates, they are attractive targets for the development of gene switches for applications in humans. The EcR gene switch is being developed for use in various applica- tions including gene therapy, expression of toxic pro- teins in cell lines and cell-based drug discovery assays [6–14]. EcRs function as an ecdysteroid-dependent tran- scription factor in cultured mammalian cells [15,16]. No et al. [17] used DmEcR and human RXRa to develop an EcR gene switch and demonstrated its function in mammalian cells and mice. The EcR gene switch was improved by using a nonsteroidal ecdysone agonist, tebufenozide, to induce a high level of repor- ter gene transactivation in mammalian cells through Bombyx mori EcR (BmEcR) [18,19] and endogenous RXR. Later, Hoppe et al. [20] combined DmEcR and BmEcR systems and created a chimeric Drosophila ⁄ Bombyx EcR (DBEcR) that had combined positive aspects of both systems so that the chimeric receptor was capable of binding to modified EcRE and also functioned without exogenous RXR. Saez and coworkers discovered that the RXR ligands enhance the ligand-dependent activity of EcR-based gene swit- ches [21], and Wybroski and coworkers [22] developed methods for expression of both EcR and RXR in a bicistronic vector. Although the current versions of EcR gene switch possess several fundamental features that confer great potential for enhancement, they do not satisfy all of the criteria desirable for a generally useful gene regula- tion system. To improve the EcR-based switch, we tested several combinations of GAL4 DNA-binding domain (GAL4 DBD), VP16 activation domain (VP16 AD), EcR and RXR and found that a two- hybrid format switch, in which GAL4 DBD was fused to CfEcR (DEF) and VP16 AD was fused to Mus musculus RXR (MmRXR) EF was the best combina- tion in terms of low background levels of reporter gene activity in the absence of a ligand and high levels of reporter gene activity in the presence of a ligand [23]. However, the ligand sensitivity of this two-hybrid for- mat EcR gene switch is not very high and requires a micromolar concentration of ligand for induction of genes. To improve the ligand sensitivity of EcR gene switch, we tested insect RXR and chimeras between human and insect RXRs as partners for EcR and discovered that the partner of EcR affects functioning of EcR in gene switch applications. The ligand sensi- tivity of EcR gene switch was improved by  100-fold by replacing HsRXR with a chimera between HsRXR and an insect RXR from Locusta migrotoria (LmRXR). Results Use of invertebrate RXR improves the function of EcR in mammalian cells Alignment of USP and RXR sequences showed that the RXR homologs, USPs from lepidopteran and dip- teran insects fall into one group and the RXR homo- logs identified from insects belonging to other orders (e.g. Heteroptera, Locusta migratoria and Coleoptera, Tenebrio molitor, as well as from crab and tick group with vertebrate RXRs) (Fig. 1). In other words, the RXR homologs identified in insects belonging to orders other than Lepidoptera and Diptera as well as from crab and tick are closer to vertebrate RXRs than to their counterparts in lepidopteran and dipteran insects. As shown in Fig. 2A, use of USP from the lepidop- teran insect Choristoneura fumiferana [V:CfUSP(EF)] as a partner for EcR from Choristoneura fumiferana [G:CfEcR(DEF)] resulted in the expression of a repor- ter gene in the absence of ligand and showed low levels of ligand-dependent induction. In contrast, use of Fig. 1. Phylogenetic tree of USP ⁄ RXR ligand-binding domain sequences. The phylogenetic tree was prepared using DNA STAR (DNA star Inc., Madison, WI). The sequences used are Homo sap- iens retinoid X receptor (HsRXR) [26], Xenopus laevis retinoid X receptor (XlRXR) [27], fiddler crab Uca pugilator RXR homolog (UpRXR) [28], Locusta migratoria RXR homolog (LmRXR) [29], Amblyomma americanum RXR homolog (AmaRXR) [30], Bombyx mori USP (BmUSP) [31], Manduca sexta USP (MsUSP) [32], Choris- toneura fumiferana USP (CfUSP) [33], Drosophila melanogaster USP (DmUSP) [34–36], Aedes aegypti USP (AaUSP) [37], Chirono- mus tentans USP (CtUSP) [27]. Influence of RXR on EcR function S. R. Palli et al. 5980 FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS RXR from Homo sapiens [V:HsRXR(EF)] or RXR from orthopteran insect, Locusta migratoria [V:Lm- RXR(EF)] as a partner for G:CfEcR(DEF) showed low background levels of expression of the GALRE- regulated luciferase reporter gene in the absence of lig- and and the luciferase activity increased after exposure to the ligand, RG-102240 (Fig. 2A). The increase in lu- ciferase activity occurred in cells that were transfected with the G:CfEcR(DEF) and V:HsRXR(EF) con- structs and exposed to a 5 lm or higher concentration of RG-102240 (Fig. 2A). By contrast, the increase in luciferase activity occurred at 200 nm or higher con- centrations of RG-102240 in cells that were transfected with the G:CfEcR(DEF) + V:LmRXR(EF) switch, showing that the ligand sensitivity of the G:CfEcR- (DEF) + V:LmRXR(EF) switch is higher than that of the G:CfEcR(DEF) + V:HsRXR(EF) switch. Pro- teins isolated from 3T3 cells transfected with V:LmRXR(EF), V:CfUSP(EF) or V:HsRXR(EF) were analyzed using western blots and VP16 antibodies. As shown in Fig. 2B, all three fusion proteins are expressed in similar quantities suggesting that the dif- ference observed in ligand sensitivity of LmRXR, HsRXR and CfUSP switches is due to structure of these proteins rather than due to differences in their expression levels. When the green fluorescence protein (GFP; placed under the control of GALRE) and G:CfEcR(DEF) + V:CfUSP(EF), G:CfEcR(DEF) + V:HsRXR(EF) or G:CfEcR(DEF) + V:LmRXR(EF) constructs were transfected into 3T3 cells, the cells transfected with G:CfEcR(DEF) + V:CfUSP(EF) switch constructs showed GFP fluorescence in the cells treated with dimethylsulfoxide (DMSO), 1.0 or 10 lm RG-102240 (Fig. 3). Low levels of GFP fluorescence were detected in 3T3 cells transfected with G:CfEcR(DEF) + V:LmRXR(EF) constructs and exposed to DMSO. However, upon exposure to 1.0 or 10 lm RG-102240, these cells showed higher GFP fluorescence (Fig. 3). In contrast, the GFP activity was not observed in 3T3 cells transfected with G:CfEcR(DEF) + V:HsRX- R(EF) constructs and exposed to DMSO. Upon expo- sure to 1.0 or 10 lm RG-102240, these cells showed GFP fluorescence (Fig. 3). The data show that G:CfEcR(DEF) + V:CfUSP(EF) switch supports the expression of the GFP gene placed under the control of GALRE even in the absence of ligand. By contrast, G:CfEcR(DEF) + V:HsRXR(EF) and G:CfEcR- (DEF) + V:LmRXR(EF) switches induce the expres- sion of GFP placed under the control of GALRE in the presence of ligand, RG-102240. In addition, the G:CfEcR(DEF) + V:LmRXR(EF) switch is more sen- sitive to ligand than the G:CfEcR(DEF) + V:HsRXR- (EF) switch. Thus, the data from these experiments confirm the results observed with the luciferase reporter. Amino acid residues present in helix 9 and in loops on either side of helix 9 of RXR are responsible for increased activity of LmRXR As shown in Fig. 2A, LmRXR performed better than HsRXR as a partner for EcR in ligand-dependent induction of reporter genes in 3T3 cells. To determine AB Fig. 2. (A) Transactivation of reporter gene by EcR + HsRXR, EcR + LmRXR and CfEcR + CfUSP gene switches. 3T3 cells were transfected with pRLUC, pFRLUC, G:CfEcR(DEF) and V:HsRXR(EF) or V:LmRXR(EF) or V:CfUSP(EF) for 4 h. The transfected cells were grown in med- ium containing DMSO, 0.04, 0.2, 1 or 5 l M RG-102240. At 48 h after addition of ligand, cells were harvested and assayed for luciferase activity. The fly luciferase activity was normalized using Renilla luciferase activity. The values presented are mean ± SD (n ¼ 3). (B) Twenty micrograms of proteins from 3T3 cells transfected with V:LmRXR(EF), V:CfUSP(EF) or V:HsRXR(EF) constructs were separated on SDS ⁄ PAGE, transferred to nitrocellulose and analyzed using VP16 antibodies. The position of 50 and 37 kDa bands form Bio-Rad Precision plus protein standards is shown on the left. Arrows point to 32, 38 and 36 kDa fusion protein bands. S. R. Palli et al. Influence of RXR on EcR function FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS 5981 which regions of LmRXR are responsible for this improved activity, we prepared five chimeras of LmRXR and HsRXR by sequentially replacing helix 6 with helix 12 of HsRXR with the corresponding regions of LmRXR. The chimeric RXRs, HsRXR and LmRXR were assayed as partners for CfEcR in ligand-dependent induction of reporter activity in 3T3 cells. The luciferase reporter gene regulated by GAL- RE (pFRLUC), G:CfEcR(DEF) and V:HsRXR(EF) or V:LmRXR(EF) or VP6 fusion of each Hs– LmRXR(EF) chimera shown in Fig. 4A were trans- fected into 3T3 cells and the transfected cells were exposed to RG-102240. Luciferase activity was measured at 48 h after addition of ligand. The G:CfEcR(DEF) + V:HsRXR(EF) switch induced lu- ciferase activity at 1 lm or higher concentration of RG-102240 and the G:CfEcR(DEF) + V:LmRXR- (EF) switch induced luciferase activity at 0.2 lm or higher concentration of RG-102240 (Fig. 4A). Repla- cing RXR with Hs–LmRXR(EF) chimera containing helices 1–7 of HsRXR and 8–12 of LmRXR or helices 1–8 of HsRXR and 9–12 of LmRXR resulted in an increase in ligand sensitivity of the EcR switch. Luci- ferase activity was induced with a 0.04 lm or higher concentration of RG-102240 (Fig. 4A) in the presence of these chimeras. The other three chimeras performed similar to HsRXR. Proteins isolated from 3T3 cells that were transfected with chimera constructs were analyzed using Western blots and VP16 antibodies. As shown in Fig. 4B, fusion proteins for all five chimeras expressed well, suggesting that the differences observed in ligand sensitivity of gene switches containing chime- ras are due to structure of these proteins rather than to differences in their expression levels. The data sug- gest that the amino acid residues present in the region of LmRXR that spans helices 8–9 and loops between helices 7–8, 8–9 and 9–10 are responsible for the increased activity of LmRXR. Comparison of amino acid sequences present in the ligand-binding domains of HsRXR and LmRXR showed that most of the differences in the amino acids between HsRXR and LmRXR are found in helix 9 and in the loops on either side of helix 9. To confirm the results observed in analysis of chimeras as well as to identify the precise region of LmRXR that is responsible for the increase in its activity when com- pared with HsRXR, we performed site-directed muta- genesis on HsRXR and changed the amino acid residues of HsRXR that are different from LmRXR to the corresponding amino acid residues present in LmRXR. The amino acids changed are shown in Fig. 5. The performance of the mutants was compared Fig. 3. Differences in the transactivation of GFP reporter gene by EcR + HsRXR (HsRXR), EcR + LmRXR (LmRXR) and CfEcR + CfUSP (CfUSP) gene switches. 3T3 cells were transfected with pFRGFP, G:CfEcR(DEF) and V:HsRXR(EF) or V:LmRXR(EF) or V:CfUSP(EF) for 4 h. The transfected cells were treated with DMSO, 1 l M or 10 lM RG-102240, the cells were photographed 48 h after addition of ligand. Influence of RXR on EcR function S. R. Palli et al. 5982 FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS with the parent RXRs in supporting EcR gene switch activity in 3T3 cells. As shown in Fig. 5, the mutants in which HsRXR amino acid residues were replaced with LmRXR residues in helix 9 as well as in loops on either side of helix 9 performed better than wild-type HsRXR as partners of EcR in supporting ligand- dependent induction of reporter activity. One partic- ular mutant, in which three amino acids present in the loop between helix 8 and 9 of HsRXR were replaced with three amino acids present in the same region of LmRXR (D450E ⁄ A451V ⁄ K452R), performed even better than LmRXR as a partner for EcR in ligand- CH6 CH8 CH9 CH10 CH11 A B Fig. 4. (A) 3T3 cells were transfected with pRLUC, pFRLUC, G:CfEcR(DEF) and V:HsRXR(EF) or V:LmRXR(EF) or VP6 fusion of one the Hs–LmRXR(EF) chimeras. Trans- fected cells were exposed to DMSO, 0.04, 0.2, 1 or 5 l M RG-102240 for 48 h. The cells were harvested and assayed for luciferase activity. The fly luciferase activity was nor- malized using Renilla luciferase activity. The values presented are mean ± SD (n ¼ 3). (B) Twenty micrograms of proteins isolated from 3T3 cells transfected with V:Hs– LmRXR(EF) chimera constructs were separ- ated on SDS ⁄ PAGE, transferred to nitrocel- lose and analyzed using VP16 antibodies. Arrow points to fusion protein bands. Fig. 5. Sequence of chimeras between HsRXR and LmRXR. The amino acids that are from HsRXR are shown with a pink background. The amino acids from LmRXR are shown with a green background. The amino acids that were mutated are shown with a yellow background. S. R. Palli et al. Influence of RXR on EcR function FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS 5983 dependent induction of reporter activity (Fig. 6A). However, the performance of this mutant is not as good as Hs–LmRXR (EF) chimera 9 (Fig. 4A) sug- gesting that not only these three amino acids but also other amino acids that are different between HsRXR and LmRXR in helix 9 as well as in the loops on either side of helix 9 contribute to the improved activ- ity of LmRXR. Western blot analysis of proteins iso- lated from 3T3 cells that were transfected with RXR mutant constructs showed that the fusion proteins for all nine mutants expressed well suggesting that the dif- ferences observed in ligand sensitivity of gene switches containing chimeras are due to structure of these pro- teins rather than due to differences in their expression levels (Fig. 6B). To confirm that the amino acid residues present in helix 9, as well as in the loops on either side of helix 9, are responsible for improved performance of LmRXR, we produced a chimera in which the region of LmRXR containing helix 9 and the two loops on either side of helix 9 were replaced with the corres- ponding region present in HsRXR. The performance of this chimera and two parent RXRs, LmRXR and HsRXR was evaluated as partners of EcR in ligand- dependent induction of reporter activity in 3T3 cells. As shown in Fig. 7, The G:CfEcR(DEF) + V:Lm- HsRXR(EF) chimera switch induced the luciferase activity with 1 lm or higher concentration of RG- 102240. This is similar to the ligand sensitivity of the G:CfEcR(DEF) + V:HsRXR(EF) switch, but lower than that of the G:CfEcR(DEF) + V:LmRXR(EF) switch in which the luciferase activity was induced with 0.2 lm or higher concentration of RG-102240 (Fig. 7). These data confirmed the results that the region of LmRXR containing helix 9 and the two loops on either side of helix 9 is responsible for improved per- formance of LmRXR as a partner for EcR in ligand- dependent induction of reporter activity. 123456789 A B Fig. 6. 3T3 cells were transfected with pRLUC, pFRLUC, G:CfEcR(DEF) and V:HsRXR(EF) or V:LmRXR(EF) or mutants of HsRXR. (A) Trans- fected cells were exposed to DMSO, 0.04, 0.2, 1 or 5 l M RG-102240 for 48 h. The cells were harvested and assayed for luciferase activity. The fly luciferase activity was normalized using Renilla luciferase activity. The values presented are mean ± SD (n ¼ 3). (B) Twenty micro- grams of proteins isolated from 3T3 cells transfected with V:HsRXR(EF) mutant constructs were separated on SDS ⁄ PAGE, transferred to nitrocellulose and analyzed using VP16 antibodies. The arrow points to fusion protein bands. Mutant 1, D450E ⁄ A451V ⁄ K452R; mutant 2, S455K ⁄ N456S ⁄ P457A ⁄ S458Q; mutant 3, V462L; mutant 4, S470A; mutant 5, T473E; mutant 6, C475T ⁄ K476R ⁄ Q477T ⁄ K478T ⁄ Y475H; mutant 7, E481D ⁄ Q482E ⁄ 483P; mutant 8, A495S; mutant 9, A528S. Influence of RXR on EcR function S. R. Palli et al. 5984 FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS To determine whether the region of LmRXR (helix 9 and two loops either side of it) that improved EcR performance would also affect RXR perform- ance mediated through 9-cis-retinoic acid, we com- pared the performance of HsRXR, LmRXR and the two chimeras, Hs–LmRXR (helix 1–8 of HsRXR and 9–12 of LmRXR) and Lm-HsRXR (helix 9 and two loops on either side of helix 9 of LmRXR were replaced with the corresponding regions form HsRXR) in transactivation assays. As shown in Fig. 8, 9-cis-retinoic acid induced reporter genes through G:CfEcR(DEF) + V:HsRXR(EF) switch at 0.2 lm or higher concentration of ligand. However, 25 lm concentration of 9-cis-retinoic acid was needed to induced reporter gene via the G:CfEcR(DEF) + V:LmRXR(EF) switch. The two chimeras per- formed similar to parents in this assay. The G:CfEcR(DEF) + V:Hs–LmRXR(EF) chimera switch supported reporter gene induction at 0.2 lm or higher concentration of 9-cis-retinoic acid and the G:CfEcR- (DEF) + V:Lm-HsRXR(EF) chimera switch suppor- ted reporter induction at 25 lm concentration of 9-cis-retinoic acid. The data suggest that the chimeras prepared by swapping helix 9 and the two loops on either side of helix 9 do not affect 9-cis-retioic acid activity through RXR. To determine whether the influence of USP ⁄ RXR on the EcR function is mediated at the level of heterodimerization between EcR and USP ⁄ RXR, we performed pull-down assays. Bacterially expressed fusion protein of GST and CfEcR(DEF) was used to pull down in vitro translated HsRXR(EF), LmRXR(EF), CfUSP(EF) and Hs–LmRXR(EF) chi- mera in the absence and presence of 1 lm RG- 102240. There was no difference in the amount of CfUSP(EF) pulled down by EcR in the presence of DMSO or 1 lm RG-102240, suggesting that EcR and USP can heterodimerize in the absence of ligand (Fig. 9). In contrast, the amount of HsRXR, Fig. 7. Comparison of two parent RXRs and Lm-HsRXR(EF) chi- mera in transactivation assays. 3T3 cells were transfected with pRLUC, pFRLUC, G:CfEcR(DEF) and V:HsRXR(EF) or V:LmRXR(EF) or VP6 fusion of Lm–HsRXR(EF) chimera (LmRXR helix 9 and loops on either side of it were replaced with the corresponding region of HsRXR). The transfected cells were exposed to DMSO, 0.04, 0.2, 1or5l M RG-10240 for 48 h. The cells were harvested and assayed for the luciferase activity. The fly luciferase activity was normalized using Renilla luciferase activity. The values presented are mean ± SD (n ¼ 3). Fig. 8. Comparison of two parent RXRs, Lm-HsRXR(EF) and Hs–LmRXR(EF) chimeras in 9-cis-retinoic acid induced transactiva- tion assays. 3T3 cells were transfected with pRLUC, pFRLUC, G:CfEcR(DEF) and V:HsRXR(EF) or V:LmRXR(EF) or VP6 fusion of Lm-HsRXR(EF) chimera (LmRXR helix 9 and loops on either side of it were replaced with the corresponding region of HsRXR) or Hs–LmRXR(EF) chimera 9. The transfected cells were exposed to DMSO, 0.04, 0.2, 1, 5 or 25 l M 9-cis-retinoic acid for 48 h. The cells were harvested and assayed for the luciferase activity. The fly luciferase activity was normalized using Renilla luciferase activity. The values presented are mean ± SD (n ¼ 3). Asterisks on top of the bars indicate significant difference from DMSO-treated cells at P < 0.5 determined by t-test. Fig. 9. GST:CfEcRDEF and [ 35 S]-methionine labeled Hs–LmRXR chi- mera (C) or HsRXREF (HsR) or LmRXREF (LmR) or CfUSP (CfU) were incubated in binding buffer containing DMSO or one lM RG-102240 and the complexes were precipitated with glutathione agarose beads. The pellet was washed and resolved on SDS ⁄ PAGE and the gel was dried and exposed to X-ray film. S. R. Palli et al. Influence of RXR on EcR function FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS 5985 LmRXR and Hs–LmRXR chimera pulled down by EcR increased after addition of 1 lm RG-102240. The increase in amount of RXR pulled down by EcR in the presence of ligand was maximum in the case of HsRXR and minimum in the case of the Hs–LmRXR chimera, and LmRXR was between these two (Fig. 9). These data suggest that the EcR and USP can heterodimerize in the absence of lig- and. In contrast, EcR:RXR heterodimer stability is increased by the presence of ligand. The EcR:Hs–LmRXR chimera switch is ligand sensitive and functions in mice Based on the data, we selected a RXR chimera that contains helices 1–8 from HsRXR and 9–12 from LmRXR and evaluated its performance as a partner of EcR in gene switch applications. The EcR:Hs– LmRXR chimera switch initiated the induction of the luciferase reporter activity beginning at 0.04 lm RG-102240 and the luciferase activity reached peak levels in the presence of 1 lm RG-102240 (Fig. 10A). This is a significant improvement in ligand sensitivity when compared with the EcR:HsRXR switch that requires 1 lm RG-102240 to initiate induction of the luciferase reporter gene and the reporter activity rea- Fig. 10. Dose-dependent induction of reporter gene by gene switch receptors. (A) 3T3 cells were transfected with G:Cf(DEF), V:Hs–LmRXR(EF), pFRLUC and pRLUC. The transfected cells were grown in the medium containing 0, 0.04, 0.2, 1 or 5 l M concentra- tion of RG-102240. The cells were collected at 48 h after adding ligand and reporter activity was quantified. The fly luciferase activity was normalized using Renilla luciferase activity. The values presen- ted are mean ± SD (n ¼ 3). (B) 3T3 cells were transfected with G:CfEcR(DEF) + V:HsRXR(EF) gene switch. 3T3 cells were trans- fected with G:Cf(DEF), V:HsRXR(EF), pFRLUC and pRLUC. The transfected cells were grown in the medium containing 0, 0.2, 1, 5 and 25 l M concentration of RG-102240. The cells were collected at 48 h after adding ligand and reporter activity was quantified. The fly luciferase activity was normalized using Renilla luciferase activity. The values presented are mean ± SD (n ¼ 3). Fig. 11. (A) Time course of induction of reporter gene by gene switch plasmids. (A) 3T3 cells were transfected with G:Cf(DEF), V:Hs–LmRXR(EF), pFRLUC and pRLUC. The transfected cells were grown in the medium containing 1 l M concentration of RG-102240. The cells were collected at 0, 1, 3, 6, 12, 24, 48 and 72 h after add- ing ligand and reporter activity was quantified. The fly luciferase activity was normalized using Renilla luciferase activity. The values presented are mean ± SD (n ¼ 3). (B) 3T3 cells were transfected with G:Cf(DEF), V:Hs–LmRXR(EF), pFRLUC and pRLUC. The trans- fected cells were grown in the medium containing 1 l M concentra- tion of RG-102240. At 48 h after addition of ligand, the cells were washed with fresh medium and maintained in the fresh medium. The cells were collected at 0, 1, 3, 6, 12, 24, 48 and 72 h after transfer to the fresh medium and the luciferase activity was quanti- fied. The fly luciferase activity was normalized using Renilla luci- ferase activity. The values presented are mean ± SD (n ¼ 3). Influence of RXR on EcR function S. R. Palli et al. 5986 FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS ches peak levels in the presence of 25 lm RG-102240 (Fig. 10B). The reporter gene regulated by the EcR:Hs–LmRXR chimera switch was induced begin- ning at 1 h after addition of ligand and reached peak levels of 19 000-fold induction by 48 h after addition of ligand (Fig. 11A). The turn off of reporter activity after withdrawal of ligand is also fast. More than 50% of reporter activity was reduced by 12 h after with- drawal of ligand and by 24 h after withdrawal of lig- and, most of the reporter activity disappeared (Fig. 11B). To evaluate the performance of the G:CfEcR(DEF): V:Hs–LmRXR(EF) switch in vivo in mice, reporter (secreted alkaline phosphatase regulated by GALRE), G:CfEcR(DEF) and V:Hs–LmRXR(EF) chimera or V:HsRXR(EF) plasmids were electroporated into the quadriceps of C57BL ⁄ 6 mice. The animals were treated with 5 mg RG-102240 ⁄ 50 lL DMSO ⁄ mouse by intraperitoneal injection at three days after electroporation of plasmids. Secreted alkaline phos- phate (SEAP) in mouse sera was evaluated at various time points after ligand administration. The G:CfEcR- (DEF) + V:Hs–LmRXR(EF) chimera switch induced SEAP activity that reached peak levels at five days after the administration of ligand (Fig. 12). By con- trast, the G:CfEcR(DEF) + V:HsRXR(EF) switch did not cause induction of SEAP activity up to 25 days after the administration of ligand (Fig. 12). Thus, the G:CfEcR(DEF) + V:Hs–LmRXR(EF) chimera switch but not G:CfEcR(DEF) + V:HsRXR(EF) switch is sensitive enough to support ligand-dependent induction of reporter gene expression in vivo. Discussion The EcR heterodimerizes with the nuclear receptor USP, binds to ecdysteroids and ecdysone response ele- ments and regulates the expression of ecdysteroid responsive genes. Because ecdysteroids and their lig- ands are absent in vertebrates, including humans, they are being developed to regulate transgenes in various applications including gene therapy, functional genom- ics, drug discovery, and biopharmaceutical production [24]. In mammalian cells, the CfEcR and CfUSP het- erodimer induces the expression of reporter genes regu- lated by EcRE even in the absence of ligand; therefore, they are not useful for gene switch applications. Ver- tebrate RXR such as human or mouse RXR have been used in the place of USP in all EcR gene switches developed to date. One of the major limitations of cur- rent versions of EcR gene switches is the requirement of micromolar concentrations of ligand for induction of gene expression. In phylogenetic analysis, Locusta migratoria RXR (LmRXR) falls into vertebrate RXR but not insect USP group (Fig. 1). We hypothesized that unlike the CfEcR:CfUSP heterodimer, the CfEcR:LmRXR heterodimer does not induce reporter genes regulated by EcRE in the absence of ligand. Furthermore, compared with the CfEcR:HsRXR het- erodimer, CfEcR:LmRXR heterodimers may induce reporter genes regulated by response elements at lower concentrations of ligand. We tested these hypotheses by comparing the performance of CfUSP, LmRXR and HsRXR as partners for CfEcR in induction of reporter genes regulated by GALRE in 3T3 cells and found that both hypotheses are true. Pull-down experi- ments showed that the CfUSP heterodimerizes with the CfEcR in the absence of ligand (Fig. 9), there- fore, CfEcR:CfUSP heterodimers can induce gene expression in the absence of ligand. However, CfEcR:HsRXR heterodimers are at very low levels in the absence of ligand and upon addition of ligand, increased quantities of HsRXRs were pulled down by CfEcR suggesting that CfEcR:HsRXR heterodimers increase in the presence of ligand and induce gene expression. We created chimeric receptors comprised of regions from HsRXR and LmRXR and mutants of HsRXR, and evaluated their performance as partners of CfEcR in ligand-dependent induction of gene expression. These analyses showed that the amino acids present in helix 9 and in the two loops present on either side of helix 9 are responsible for improved performance of Fig. 12. In vivo comparison of human RXRb and Hs-RXRb –Lm- RXRb fusion protein in a C57BL ⁄ 6 mouse model. The gene swit- ches, composed of plasmids containing pCMV ⁄ GAL4-pCfEcR(DEF), pCMV ⁄ VP16-Hs-RXR(EF) (s) or pCMV ⁄ VP16-HsRXRb(H1-8)- LmRXRb(H9-12) fusion (m), and 6xGAL4RE-TTR-SEAP, were elec- troporated into the quadriceps of C57BL ⁄ 6 mice. Animals were treated with 5 mg RG-102240 ⁄ 50 lL DMSO ⁄ mouse by IP injection three days after electroporation of plasmid. SEAP in mouse sera was evaluated for up to 17 days after ligand administration. Values are the average from seven animals ± SD. S. R. Palli et al. Influence of RXR on EcR function FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS 5987 LmRXR when compared with the performance of HsRXR as a partner for CfEcR. Structural studies on EcR:USP and RAR:RXR heterodimers showed that helices 7 and 10, present in two nuclear receptors, form heterodimerization interfaces and play critical roles in heterodimerization [25]. The data presented here showed that besides helices 7 and 10, amino acid residues present in helix 9 and in the two loops present on either side of helix 9 play critical roles in hetero- dimerization of EcR:USP ⁄ RXR. In fact, there is only one amino acid each in helix 7 and helix 10 that is dif- ferent between HsRXR and LmRXR. Replacing these two amino acids in HsRXR with the corresponding amino acids present in LmRXR did not increase the performance of HsRXR as a partner for CfEcR. In contrast, replacing HsRXR amino acids present in helix 9 and in the loops on either side of helix 9 with the amino acids present in corresponding positions in LmRXR resulted in an increase in the performance of HsRXR as a partner for CfEcR. Particularly, repla- cing three amino acids (DAK) located in the loop between helices 8 and 9 of HsRXR with the amino acids (EVR) present in the corresponding positions of LmRXR resulted in a HsRXR mutant that performed even better than LmRXR as CfEcR partner. Examina- tion of structures of EcR and RXR indicated that, when compared with the aspartic acid (D) residue pre- sent in the loop between helices 8 and 9 of HsRXR, the glutamic acid (E) residue present in the corres- ponding region of LmRXR is located at a more favo- rable distance to interact with the arginine (R) residue present in helix 7 of EcR [25]. Taken together, these studies conclusively show that the amino acid residues present in helix 9 and in the two loops on either side of helix 9 contribute to the heterodimerization of EcR and RXR. The diacylhydrazine nonsteroidal ligands of EcR are not highly polar, therefore; higher concentrations of these ligands or highly sensitive gene switches are required for in vivo applications. Because the CfEcR:HsRXR switch requires micromolar concen- trations of ligand for the transactivation of genes, it does not function very well for in vivo applications. However, the CfEcR:Hs–LmRXR chimera switch requires only nanomolar concentrations of ligands for the transactivation of genes and functions well in vivo. The RXR chimeras containing most of HsRXR and helix 9 and the loops on either side of helix 9 from HsRXR or mutants of HsRXR with a change in just three amino acids (D450E ⁄ A451V ⁄ K452R) will definitely help in the develop- ment of gene switches, especially those that require in vivo applications. Experimental procedures Constructs The construction of GAL4:CfEcR(DEF) [G:CfEcR(DEF)], VP16:MmRXR(EF) [V:MmRXR(EF)] and VP16:CfUS- P(EF) [V:CfUSP(EF) has been described previously [23] VP16:LmRXR(EF) [V:LmRXR(EF)] was constructed by amplifying EF domains of LmRXR using primers contain- ing EcoRI and BamHI sites in the forward and reverse pri- mer, respectively, followed by cloning of the PCR product into EcoRI and Bam HI digested pVP16 vector (Clontech Inc. Palo Alto, CA). pFRLUC reporter plasmid was pur- chased from Stratagene (La Jolla, CA). pRLUC is reporter plasmid expressing Renilla luciferase under the control of thymidine kinase promoter (Promega, Madison, WI). The GST fusion construct of MmR(EF) was made by cloning MmR(EF) domain into pGEX-5X-1 vector (Amersham Pharmacia Biotech, Piscataway, NJ) forward and reverse primers, respectively. Ligands RG-102240 [N-(1,1-dimethylethyl)-N¢-(2-ethyl-3-methoxy- benzoyl)-3,5-dimethylbenzohydrazide] also known as GS TM -E and RheoSwitchÒ ligand 1 (RSL1) is a synthetic stable diacylhydrazine ecdysone agonist synthesized by RheoGene Inc.; 9-cis-retinoic acid was purchased from Sig- ma Chemical Co. (St Louis, MO, USA). The ligands were applied in DMSO and the final concentration of DMSO was maintained at 0.1% in both controls and treatments. Cells and transfections and reporter assays 3T3 cells were grown to 60% confluency. Fifty thousand cells were plated per well of 12-well plates. The next day, cells were transfected with 0.25 lg of receptor(s) and 1.0 lg of reporter constructs using 4 lL of SuperFect (Qiagen Inc., Valencia, CA). A second reporter, Renilla luciferase, expressed under a thymidine kinase constitutive promoter was cotransfected into cells and used for normal- ization. After transfection, cells were grown in a medium containing ligands for 24–48 h. The cells were harvested, lyzed and the reporter activity was measured in an aliquot of lysate. All transfection experiments were performed in triplicate and the experiments were repeated at least three times. Luciferase and Renilla luciferase activities were measured using the Dual-luciferase TM reporter assay sys- tem (Promega). Construction of chimeras and mutants Site-directed mutagenesis was carried out using the QuikchangeÒ site directed mutagenesis kit (Stratagene). Mutations were verified by sequencing. RXR chimeras Influence of RXR on EcR function S. R. Palli et al. 5988 FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS [...]... encodes an ecdysone receptor, a new member of the steroid receptor superfamily Cell 67, 59–77 2 Thomas HE, Stunnenberg HG & Stewart AF (1993) Heterodimerization of the Drosophila ecdysone receptor with retinoid X receptor and ultraspiracle Nature 362, 471–475 3 Yao TP, Forman BM, Jiang Z, Cherbas L, Chen JD, McKeown M, Cherbas P & Evans RM (1993) 15 16 Functional ecdysone receptor is the product of EcR... Gage FH (1998) High level transactivation by a modified Bombyx ecdysone receptor in mammalian cells without exogenous retinoid X receptor Proc Natl Acad Sci USA 95, 7999–8004 19 Swevers L, Drevet JR, Lunke MD & Iatrou K (1995) The silkmoth homolog of the Drosophila ecdysone receptor (B1 isoform): cloning and analysis of expression during follicular cell differentiation Insect Biochem Mol Biol 25, 857–866... Adenovirusmediated inducible gene expression in vivo by a hybrid ecdysone receptor Mol Ther 1, 159–164 21 Saez E, Nelson MC, Eshelman B, Banayo E, Koder A, Cho GJ & Evans RM (2000) Identification of ligands and coligands for the ecdysone- regulated gene switch Proc Natl Acad Sci USA 97, 14512–14517 22 Wyborski DL, Bauer JC & Vaillancourt P (2001) Bicistronic expression of ecdysone- inducible receptors in mammalian... T & Hill R (1995) Functional transfer of an elementary ecdysone gene regulatory system to mammalian cells-transient transfections and stable cell-lines Eur J Entomol 92, 379–389 FEBS Journal 272 (2005) 5979–5990 ª 2005 The Authors Journal Compilation ª 2005 FEBS 5989 Influence of RXR on EcR function S R Palli et al 17 No D, Yao TP & Evans RM (1996) Ecdysone- inducible gene expression in mammalian cells... al Influence of RXR on EcR function were constructed using the Seamless cloning kit (Stratagene) 4 Western blots 3T3 cells were transfected with V:LmRXR(EF) or V:CfUSP(EF) or V:HsRXR(EF) Twenty micrograms of proteins isolated form these cells at 48 h after transfection were separated by SDS ⁄ PAGE and transferred to nitrocellulose membranes After blocking with 1% dry milk powder, the membranes were exposed... modulates ecdysone receptor function via heterodimer formation Cell 71, 63–72 Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M & Chambon P (1995) The nuclear receptor superfamily: the second decade Cell 83, 835–839 Albanese C, Reutens AT, Bouzahzah B, Fu M, D’Amico M, Link T, Nicholson R, Depinho RA & Pestell RG (2000) Sustained mammary gland-directed, ponasterone... ponasterone A-inducible expression in transgenic mice FASEB J 14, 877–884 Zhou Y & Ratner L (2001) A novel inducible expression system to study transdominant mutants of HIV-1 Vpr Virology 287, 133–142 Fussenegger M (2001) The impact of mammalian gene regulation concepts on functional genomic research, metabolic engineering, and advanced gene therapies Biotechnol Prog 17, 1–51 Yarovoi SV & Pederson T (2001)... cell lines expressing hormone regulated T7 RNA polymerase localized at distinct intranuclear sites Gene 275, 73–81 Yam JW, Chan KW & Hsiao WL (2001) Suppression of the tumorigenicity of mutant p53-transformed rat embryo fibroblasts through expression of a newly cloned rat nonmuscle myosin heavy chain-B Oncogene 20, 58– 68 Sparacio S, Pfeiffer T, Schaal H & Bosch V (2001) Generation of a flexible cell... Improved ecdysone receptor- based inducible gene regulation system Eur J Biochem 270, 1308–1315 24 Palli SR, Hormann RE, Schlattner U & Lezzi M (2005) Ecdysteroid receptors and their applications in agriculture and medicine Vitamins Hormones in press 25 Billas IM, Iwema T, Garnier JM, Mitschler A, Rochel N & Moras D (2003) Structural adaptability in the ligand-binding pocket of the ecdysone hormone receptor. .. F, Asahina M & Riddiford LM (1997) Identification and mRNA developmental profiles of two ultraspiracle isoforms in the epidermis and wings of Manduca sexta Insect Mol Biol 6, 41–53 Perera SC, Palli SR, Ladd TR, Krell PJ & Retnakaran A (1998) The ultraspiracle gene of the spruce budworm, Choristoneura fumiferana: cloning of cDNA and developmental expression of mRNA Dev Genet 22, 169–179 Henrich VC, Sliter . The influence of heterodimer partner ultraspiracle/retinoid X receptor on the function of ecdysone receptor Subba R. Palli 1 ,. results that the region of LmRXR containing helix 9 and the two loops on either side of helix 9 is responsible for improved per- formance of LmRXR as a partner