Introduction
The molting process of arthropods is regulated by two endogenous peripheral insect hormones such as ecdysteroids and juvenile hormones. Ecdysteroids are essential to elicit molting in arthropods by binding to the ecdysone receptor (EcR) heterodimerized with the ultraspiracle (USP), an ortholog of the retinoid X receptor (RXR). These two receptor proteins [EcR and USP (or RXR)] are members of the nuclear receptor superfamily.1,2) When ecdysteroid levels in the hemolymph increase, the specific network system involved in the arthropod molting process proceeds. This molting process is triggered by the binding of 20-hydroxyecdysone (20E; Fig. 1) to EcR with the aid of USP. The chemical structure of 20E as well as its three-dimensional structure was elucidated half a century ago.3) Mostly in arthropods, 20E is used as a molting hormone, even in scorpions.4) However, other ecdysteroids such as ecdysone,5) makisterone A6) and ponasterone A (PonA; Fig. 1)7) are used as molting hormones in a few species.
Fig. 1. Chemical structures of representative ecdysone agonists.
Although the compounds that exert molting hormonal activity are candidates for insecticides,8) ecdysteroids are not good candidates due to the inappropriate physicochemical properties of steroids in general and the difficulty and expense of chemical synthesis of ecdysteroids. However, the discovery of non-steroidal ecdysone agonists such as diacylhydrazines (DAHs)9,10) allows the pursuit of compounds exhibiting potent molting activity as insecticides. To date five DAHs, tebufenozide (Fig. 1), methoxyfenozide, chromafenozide, fufenozide, and halofenozide are used as insecticides.11,12) Most of them are toxic to Lepidoptera but inactive or less active against other insect orders such as Diptera, Hemiptera, Hymenoptera, Neuroptera, and Coleoptera.13,14) This selective toxicity of DAHs is due to the differences in the amino acid sequence of the ligand binding domain (LBD) in EcR among insect species. Since the insecticidal spectrum of DAH-type compounds is narrow (specific to Lepidoptera), significant efforts have been made to identify new types of non-steroidal ecdysone agonists. To date, new compound classes such as 3,5-di-tert-butyl-4-hydroxy-N-i-butylbenzamide (DTBHIB),11,15) acylaminoketones,16) tetrahydroquinolines,17) oxadiazolines,18) and imidazoles19) have been reported to be ecdysone agonists, but they are less potent than representative DAHs. However, among them, tetrahydroquinoline-type compounds are effective against Diptera, particularly against mosquitoes.20) Previously, we found three new compounds with novel mother structures that specifically bind to EcR using in silico screening,21) but they are inactive under an in vivo assay.
Pesticides that have been developed and used recently are required to be less toxic to mammals and labile in environment. In addition, new pesticides are required to show selective toxicity between pests and beneficial insects and aquatic animals; unfortunately such selectivity is not easy to achieve for neuroactive insecticides. However, as stated above, the ecdysone receptor is a plausible target to distinguish species, leading to the discovery of selective insecticides. Since the primary sequences of EcRs are varied among species, it is possible to develop highly selective EcR-based insecticides, which are toxic to target pests but not beneficial insects and other non-target arthropods. Studies on EcR/USP cloning and ligand-receptor interaction were mostly done for pests,22–24) not for beneficial insects.
In this study, we performed cDNA cloning of EcR and USP genes of the multicolored Asian lady beetle Harmonia axyridis and the phytophagous ladybird beetle Epilachna vigintioctopunctata to investigate selective toxicity for EcR-based insecticides in pests and beneficial insects. E. vigintioctopunctata is phytophagous and may damage crops, but H. axyridis is carnivorous, eating harmful aphids. In order to gain a detailed understanding of ligand-receptor interactions, the EcR/USP proteins from these two species were prepared and the binding potency of ecdysone agonists (Fig. 1) to them was measured. In further study, docking simulations were performed to understand the differences in ligand binding between the EcR/USP proteins from H. axyridis and E. vigintioctopunctata.
Materials and Methods
1. AnimalsH. axyridis was kindly provided from the Laboratory of Environmental Zoology at Meijo University, and the Laboratory of Sericulture and Entomoresources at Nagoya University. E. vigintioctopunctata was collected at an experimental farm of Meijo University.
2. ChemicalsTritiated ponasterone A ([3H]PonA, 140 Ci/mmol) was purchased from American Radiolabeled Chemicals Inc. (St. Louis, MO). RH-5849 and tebufenozide were from our stock samples.25,26)
3. cDNA cloning of EcR and USPPrepupae of H. axyridis and E. vigintioctopunctata were dissected, and their midgut and most of the fat body were removed. Total RNA was extracted using TRIzol (Life Technologies) as described previously.27) cDNAs of EcRs and USPs were synthesized from total RNA with a Ready-To-Go T-Primed First-Strand Kit (GE Healthcare). Reverse transcription polymerase chain reaction (RT-PCR) was performed with degenerate primers designed in our previous study.27,28) To amplify EcR from H. axyridis, the first PCR was performed using EcR-F1 and EcR-R2 primers with the annealing temperature of 48°C. To conduct the second and the third PCRs (nested PCR), EcR-F2/R2 and EcR-F3/R3 primers were used with the annealing temperatures of 52°C and 46°C, respectively. To amplify EcR from E. vigintioctopunctata, the first PCR was performed using EcR-F1 and EcR-R1 primers with the annealing temperature of 42°C, followed by the second PCR with EcR-F2/R2 at 44°C for annealing and the third PCR with EcR-F3/R3 at 46°C for annealing. Similarly, a cDNA fragment encoding USP from H. axyridis was amplified using USP-F1 and USP-R2 primers with the annealing temperature of 46°C, followed by a nested PCR with USP-F2/R2 primers with the annealing temperature of 48°C. To amplify a longer fragment, PCR was performed using a gene-specific primer HaUSP-RF1 and a degenerate primer USP-ER1 with the annealing temperature of 55°C, followed by a nested PCR using a gene-specific primer HaUSP-RF2 and a degenerate primer USP-ER2 with the annealing temperature of 50°C. These degenerate primers (USP-ER1 and USP-ER2) were designed based on a conserved domain in the E/F region. A cDNA fragment encoding E. vigintioctopunctata USP was amplified using USP-F1 and USP-R2 primers with the annealing temperature of 46°C, followed by a nested PCR with USP-F2/R2 primers with the annealing temperature of 44°C. To amplify a longer fragment, PCR was performed using a gene-specific primer EvUSP-RF1 and a degenerate primer USP-ER1 with the annealing temperature of 55°C, followed by a nested PCR using a gene-specific primer EvUSP-RF2 and a degenerate primer USP-ER2 with the annealing temperature of 50°C.
To obtain full-length cDNA of EcR and USP, rapid amplification of cDNA ends (RACE) was conducted as follows. Poly (A)-rich RNA was purified from the total RNA using QuickPrep Micro mRNA Purification Kit (GE Healthcare), and cDNA pools were prepared using a SMART RACE cDNA amplification kit (Clontech). Touchdown PCR and the secondary nested PCR were performed with Advantage2 Polymerase (Clontech) according to the manufacturer’s instructions. Primer sequences are listed in Table S1. These PCR products were purified, cloned into a pGEM-T Easy vector (Promega), and sequenced. The DNA sequence data have been deposited in the DDBJ/EMBL-Bank/GenBank International Nucleotide Sequence Database under accession numbers AB506665 (HaEcR-B1), AB506666 (HaEcR-A), AB506667 (HaUSP-1), AB506668 (HaUSP-2), AB506669 (EvEcR-B1), AB506670 (EvEcR-A), AB506671 (EvUSP-1) and AB506672 (EvUSP-2).
4. Binding assay using in vitro translated proteinsFull-length DNA fragments for the ORFs of HaEcR-A, HaUSP-1, EvEcR-A and EvUSP-1 were amplified by PCR and cloned into the pBluescript II SK (+) vector (Agilent Technologies, Inc. CA, USA). HaEcR-A, HaUSP-1, EvEcR-A and EvUSP-1 proteins were prepared using these constructs and TNT T7 Coupled Reticulocyte Lysate Systems (Promega) according to the manufacturer’s instructions. The ligand binding assay was performed as described previously.27,29) Briefly, in vitro translated EcR and USP proteins from H. axyridis and E. vigintioctopunctata were incubated with [3H] PonA and a test compound for 90 min at 25°C. A 1000-fold excess of unlabeled PonA was added to measure nonspecific binding, and total binding was obtained for the treatment with a carrier. After incubation, the reaction mixtures were filtered through GF-75 glass filters (ADVANTEC, Tokyo, Japan) with the aid of vacuum pump. Filters were washed and transferred to a vial containing 3 mL of Aquasol-2 (Perkin-Elmer, MA, USA) to measure the radioactivity in an Aloka LSC-6100 liquid scintillation counter (Aloka, Tokyo, Japan).
5. Receptor modeling and ligand-receptor docking simulationThe three-dimensional (3-D) structures of LBDs (the 235 amino acids Q294-W528) of EvEcR and HaEcR were constructed from HvEcR (1R20; Fig. 2)22) using homology modeling. The full automated modeling system (FAMS, In-Silico Sciences Inc., Tokyo, Japan), an extension of FAMS originally developed by Ogata and Umeyama,30) was used. EcR-LBDs of H. axyridis and E. vigintioctopunctata were aligned to that of HvEcR as shown in Fig. 2. Although two other crystal structures of EcR-LBDs are reported as complexes with 20E (2R40) and PonA (1R1K), 1R20 bound to a DAH-type compound (BYI06830) was used as a template for FAMS. Four amino acids are mutated in 1R20 (W303Y, A361S, L456S, and C483S) compared to the wild-type EcRs (2R40 and 1R1K), but these mutations are far from the ligand binding pocket and are therefore considered unlikely to alter the binding interactions under study.
Fig. 2. Comparison of ligand binding domains of EvEcR and HaEcR with that of HvEcR (1R20) with four amino acid mutations (letters in squares). H1 to H12 means α-helices 1–12. Numbers above the sequence of HvEcR are the same as those published in Nature 2003. Yellow-colored (online version) characters mean that they are different between HaEcR and EvEcR.
These mol2 files of 3-D structures of RH-5849 and tebufenozide that were constructed from the X-ray crystal structure of RH-584931) were submitted to OMEGA (ver. 2.4.5; OpenEye, NM, USA; http://www.eyesopen.com)32) to generate 200 conformers, which were converted to sdf files and stored in a folder. The ligand binding pockets were made by the “MAKE RECEPTOR” and “FRED (Fast Rigid Exhaustive Docking)” tool of OE docking (ver. 3.0.1; Open Eye, USA; http://www.eyesopen.com). The box enclosing the binding site is automatically created around the bound ligand and then the binding site shape is defined inside the box based on the shape potential. OMEGA uses its own torsion and ring libraries to identify and enumerate rotatable bonds and flexible rings. Conformers with internal clashes or high strain are then discarded. An exhaustive search was done by rotating and translating the rigid conformer within the active site with a Chemgauss4 scoring function. The Chemgauss4 scoring function uses smooth Gaussian functions to represent the shape and chemistry of molecules and is aware of interactions such as shape, hydrogen bonds, and desolvation. The models with the highest scores in terms of Chemgauss4 scoring function were selected in this study. VIDA (ver. 4.2.1; OpenEye) was used to view the models.
Docking simulation was also executed using the GOLD Suite (ver. 5.2; CCDC, UK)33) where the tebufenozide and modeled proteins were submitted to GOLD docking. Modeled proteins are the same as those used for FRED docking. The pocket for docking was chosen by removing BYI06830 bound to EcRs. The initial 3-D structure of tebufenozide was constructed based on the X-ray structure of RH-5849.31) In the ligand flexibility setting, default setting was used for torsion angle distribution and post-process rotatable bonds. In the fitness search options, docking was performed using ChemPLP as a scoring function with the default parameter setting. ChemPLP uses the Chemscore hydrogen bonding term and multiple linear potentials to model van der Waals and repulsive terms. Early termination was chosen if the top 20 solutions were within 1.5A. In the genetic algorithm (GA) setting, “automatic” with very flexible search efficiency (200%) was selected.
Results
1. cDNA cloning of EcR and USPThe 2399 bp and 2297 bp cDNA sequences were obtained from H. axyridis by executing RT-PCR, 5′-RACE and 3′-RACE. The 2032 bp and 2142 bp cDNAs were also sequenced in E. vigintioctopunctata. The longest ORFs for H. axyridis encoded 496 and 570 amino acids, which are respectively deduced analogous to EcR-A and B1 by a BLAST search (http://www.ncbi.nlm.nih.gov/BLAST/) and named HaEcR-A and HaEcR-B1. EvEcR-A (ORF; 498 a.a.) and EvEcR-B1 (ORF; 573 a.a.) were also deduced in E. vigintioctopunctata. These four EcRs are shown in Fig. 3. In a similar manner, we obtained cDNA sequences encoding HaUSP-1 (ORF; 438 a.a.) and HaUSP-2 (ORF; 407 a.a.) from H. axyridis (2131 bp and 2000 bp) and EvUSP-1 (ORF: 429 a.a.) and EvUSP-2 (ORF; 398 a.a.) from E. vigintioctopunctata (1547 bp and 1405 bp) (Fig. 4).
Fig. 3. Alignment of EcR isoforms of H. axyridis and E. vigintioctopunctata.
Fig. 4. Alignment of USP isoforms of H. axyridis and E. vigintioctopunctata.
Sequences of EvEcR-A, EvEcR-B1, HaEcR-A, and HaEcR-B1 were aligned with other EcRs (Fig. S1). A/B domains are fairly different even among Coleoptera EcRs. The identity of A/B domains is high within A isoforms of HaEcR and EvEcR, and that for their B1 isoforms is high as well. DNA binding domains are homologous in all EcRs. The ligand binding domains (LBDs) are identical between B1 and A isoforms in the same species, and the homology of LBDs is very high (89%) between EvEcR and HaEcR (Fig. 3). LBDs of EvEcR and HaEcR were also homologous to other Coleopteran EcRs such as TmEcR and TcEcR (80% identity). Since LBDs are highly conserved among taxonomic insect orders (Fig. S2), the phylogenetic tree was derived using the sequences of LBDs using CLUSTALW (http://www.genome.jp/tools/clustalw/), showing that H. axyridis and E. vigintioctopunctata were clearly grouped in Coleoptera as shown in Fig. 5. This tree is constructed using the auto-execute botton “rooted phylogenetic tree with branch length (UPGMA)” of the software CLUSTALW 2.1 (http://www.genome.jp/tools/clustalw/).
Fig. 5. Phylogenetic trees constructed using the primary sequences of the ligand binding domains of EcRs.
USPs of H. axyridis and E. vigintioctopunctata were aligned to other USPs (Fig. S3), indicating that they are homologous not only to Coleoptera insects, but early derived insect species (ancient arthropod lineage) such as Locusta migratoria (LmUSP; Orthoptera) and Blattera germanica (BgUSP; Blattodae). Even though the homology of the LBDs of EcRs is very high among the same insect orders, it is not so high for the LBDs of USPs.
2. Binding affinity of ecdysone agonists to HaEcR/HaUSP and EvEcR/EvUSPThe binding assay was performed for PonA and non-steroidal ecdysone agonists, RH-5849 and tebufenozide, against in vitro translated EcR/USP heterodimers. The 50% inhibition concentrations (IC50) of PonA against HaEcR/HaUSP and EvEcR/EvUSP were determined to be 5.6 nM and 4.0 nM, respectively by probit analysis.34) Both nonsteroidal ecdysone agonists RH-5849 and tebufenozide significantly inhibited the [3H]PonA binding against HaEcR/HaUSP heterodimer and their IC50 values were determined to be 10.5 µM and 11.0 µM, respectively, from their concentration curves (Fig. 6). However, these compounds were inactive even at high concentration (250 µM) against E. vigintioctopunctata EcR/USP (Table 1).
Fig. 6. Concentration–response curves for the ligand–receptor binding of tebufenozide (A) and RH-5849 (B).
| Compounds | IC50 (µM) |
|---|
| EvEcR | HaEcR | LdEcRa) |
|---|
| Ponasterone A | 0.00398 | 0.00562 | 0.0074 |
| RH-5849 | >250 (30.6%) | 10.5 | 10.7 |
| Tebufenozide | >250 (5%) | 11.0 | 6.6 |
In order to find the reason for the selective binding of DAHs (RH-5849 and tebufenozide) to EcRs between H. axyridis and E. vigintioctopunctata, we examined the ligand-receptor interactions using an in silico docking technique. In the docking study using FRED, RH-5849 and tebufenozide are nicely accommodated in the ligand binding pocket of both EvEcR and HaEcR, which are docked to the receptor in a manner similar to BYI06830. The hydrogen bonds (HBs) between TYR113 (identical to Y408 for HvEcR: Fig. 2) and the carbonyl group far from N(t-Bu) moiety of both RH-5849 and tebufenozide were detected in the ligand binding pocket of both EcRs (Fig. 7), although Y408 formulated an HB toward NH of BYI06830 in HvEcR. The docking simulation of tebufenozide to HaEcR is shown in Fig. 7.
Fig. 7. Docking simulation of tebufenozide to HaEcR using FRED. The structure shown in gray (online version) is BYI06830 that was co-crystalized with HvEcR.
Docking results using GOLD are shown in Fig. 8. In the GOLD docking, only 10 results were obtained for tebufenozide against both HaEcR and EvEcR, even though the maximum number of results has been set at 20. The docked conformations are rather varied in EvEcR as shown in Fig. 8. Figure 8a is for HaEcR and Fig. 8b is for EvEcR. Docking scores for 10 results for HaEcR are from 71.89 to 72.86, and those for EvEcR are from 72.84 to 75.86. Variations of docking scores of 10 structures to HaEcR are smaller than those for EvEcR, suggesting that tebufenozide is more closely associated with the ligand binding pocket of HaEcR than that of EvEcR.
Fig. 8. Docking simulations of tebufenozide to HaEcR (a) and EvEcR (b) using GOLD. Ten conformers are overlapped.
Discussion
1. Primary sequences of lady beetle EcRsThe A/B regions of HaEcR and EvEcR are very different between them, even in the same family, Coccinellidae (Fig. 3 and Fig. S1). The difference in EcR or USP isoforms is generally created by the link of different exons in the A/B regions. In the case of H. axyridis and E. vigintioctopunctata, the A/B regions of EcR-B1s are longer than those of EcR-As, as reported in Bombyx mori,35,36) Manduca sexta,37,38) and Drosophila melanogaster.39,40) A/B regions are homologous between B1 isoforms (89% identity for HaEcR-B1/EvEcR-B1) and between A isoforms (80% identity for HaEcR-A/EvEcR-A) irrespective of the Coccinellidae species. The difference between EcR-B1 and EcR-A is fairly large in terms of the length of primary sequence (75 amino acids for HaEcR; 74 amino acids for EvEcR), being similar to that (77 amino acids) of Leptinotarsa decemlineata (A: 488 a.a. vs. B1: 565 a.a.)27) and that (61 amino acids) of TcEcR (A: 484 a.a., B1: 549 a.a.),41) but the difference between B1 and A isoforms is small for Lepidoptera Chilo suppressalis (29 amino acids; A: 518 a.a. vs. B1: 547 a.a.)42) and B. mori (31 amino acids; A: 512 a.a. vs. B1: 543 a.a. ).35,36) Even though the transactivation potency and gene expression are not examined in this study, EcR-B1 probably has different transactivation potency from that of EcR-A, and the expression of EcR isoform genes is thought to be tissue and stage specific.
The A/B domain is related to the ligand-independent transactivation (activation function 1; AF1).43) On the other hand the ligand-dependent transactivation (AF2) is triggered by the LBD that is highly conserved among insect taxonomic orders. Therefore, LBDs related to the transactivation are critical for the ligand structure. LBDs are highly conserved among Coleoptera insects (EvEcR, TmEcR, LdEcR, and TcEcR), as shown in Fig. S2. The LBDs of both HaEcR and EvEcR were homologous to those of Coleopteran EcRs (86–92%), Lepidopteran EcR (61–64%) and Dipteran EcR (63–69%). The LBDs of EcRs are important for the ligand binding as well as heterodimerization to USP and the transactivation that is regulated by the binding of the cofactors, co-repressor (CoR) and co-activator (CoA).39,44) This domain is varied among species, in particular the sequence difference is remarkable among insect taxonomic orders, but DNA binding domains are highly conserved (about 90% or more) among all species.
In the alignment of EcR-LBDs to HvEcR-LBD (Fig. 2, Fig. S2), the numbering of amino acid residues of HvEcR22) was used to make the discussion easy. Crystal structure analysis indicated that EcR-LBDs are formed by 12 α-helices numbered from helix-1 (H1) to helix-12 (H12) and β-sheets for PonA-bound EcRs. Twenty-five amino acids of LBDs are different between HaEcR and EvEcR: 306 (S, N), 308 (H, Y), 312 (S, P), 313 (E, P), 316 (V, I), 317 (K, E), 320 (I, S), 321 (N, D), 322 (A, T), 328 (D, E), 332 (C, S), 364 (L, M), 391 (N, A), 404 (S, T), 410 (L, I), 429 (N, Y), 439 (L, I), 453 (M, V), 455 (G, A), 475 (Y, N), 480 (A, K), 481 (H, P), 482 (K, R), 491 (L, M), and 492 (A, S). Interestingly, the basic amino acid residues (K317) of HaEcR are replaced with acidic amino acid residues (E317) in EvEcR, and the neutral amino acid (A480) of HaEcR is replaced with the basic amino acid (K480) in EvEcR. The acidic amino acid (D321) of EvEcR is replaced with the neutral amino acid N321 in HaEcR.
It was demonstrated that E309, T343, R383 and A398 are involved in hydrogen-bond formation in the PonA binding to BtEcR,45) which is also conserved in Lepidoptera, Orthoptera and Hemiptera EcRs as shown in Fig. S2. In the case of alanine (A398), the backbone amide is implicated in the hydrogen-bond formation. However, this alanine is replaced with valine (V398) in Coleoptera EcRs including EvEcR and HaEcR and glycine (G398) in one of crustacean EcRs, UpEcR. Three other HB forming amino acid residues (E309, T343, and R383) are all conserved irrespective of species except for Daphnia magna. In addition to HB formation, hydrophobicity is important for the ligand–receptor interaction. According to Carmichael, 9 amino acid residues of BtEcR (I342, L345, L349, R383, M384, R387, I395, L396, and F397) are involved in the hydrophobic interaction between the ligand binding pocket and the ecdysone core (steroidal skeleton).23) These hydrophobic residues are conserved in EvEcR and HaEcR. Amino acid residues constructing the binding pocket accommodating an alkyl side chain [I339, T340, M380, M381, Y408, and L420] are conserved in most EcRs including HaEcR and EvEcR, while I339 is replaced with T339 only in TmEcR. Residues M413, N504, C508, and L511 surrounding the terminal region of the alkyl chain are also conserved in HaEcR and EvEcR. In a few species methionine (M413) is changed to leucine or valine (Fig. S2). Taken together, these similarities of LBDs between EvEcR and HaEcR may have given the same IC50 values against PonA.
Previously the phylogenetic trees were constructed using sequences of EcR and USP.28) To understand the evolutional difference based on feeding habits, the phylogenetic tree was constructed using the primary sequences of the LBDs of EcRs as shown in Fig. 5, resulting in both those of H. axyridis and E. vigintioctopunctata being similar to those of other Coleoptera insects such as Tenebrio moritor, L. decemlineata, and Tribolium castaneum. Sequences of LBDs seem to be better for constructing phylogenetic trees because the discrimination of isoforms is not necessary.
2. Primary sequences of lady beetle USPsThe A/B regions of HaUSP and EvUSP were somewhat longer than those of other Coleoptera species such as L. decemlineata, T. molitor, and T. castaneum. Interestingly, this region is fairly different between HaUSPs and EvUSPs (65.8–76.1%). The identities of the A/B domains between USP isoforms of H. axyridis and E. vigintioctopunctata were very high (HaUSP-1/HaUSP-2=92.3%; EvUSP-1/EvUSP-2=92.4%), although the identity between different species was reduced [HaUSP-1:EvUSP-1 (76.1%); HaUSP-1:EvUSP-2 (65.8%); HaUSP-2:EvUSP-1 (66.7%); HaUSP-2:EvUSP-2 (70.5%)]
Primary sequences of USPs of Coleoptera are aligned as shown in Fig. S4. DNA binding domains were 100% identical between HaUSPs and EvUSPs, but two amino acids (T160 and L190 in Fig. S4) of these USPs were different from those (S and V/N) of other Coleoptera (LdUSP, TmUSP, and TcUSP). In particular leucine (L190) in these lady beetles is varied (V, N, A, S). The identity of the ligand binding domains of USPs was not very high among four USPs (HaUSP-1, HaUSP-2, EvUSP-1, and EvUSP-2) (77.7%) compared to that (89.4%) of EcR-LBDs which are highly conserved in the same taxonomic orders.
The USP is the ortholog of RXR for mammals that belongs to the nuclear receptor super family and regulates molting and metamorphosis as a key transcription factor after heterodimerization with EcR. However, the ligand for USP is not identified, even though 9-cis-retinoic acid (9cRA) is known as the ligand for RXR. The specific ligand for each USP might be a selective insect growth regulator like EcR ligands. LBDs of USPs are not highly conserved between HaUSP and EvUSP. It may therefore be possible to develop USP ligand-derived insecticides that are selective for harmful lady beetles.
3. Ligand–receptor bindingAlthough DAH-type compounds did not inhibit [3H]PonA binding against EvEcR/EvUSP at the concentration of 250 µM, the binding activity of RH-5849 and tebufenozide against HaEcR/HaUSP was determined to be 10.5 µM and 11.0 µM, respectively, in terms of IC50. These IC50 values were close to those (10.7 µM for RH-5849 and 6.6 µM for tebufenozide) against another Coleoptera insect species L. decemlineata.27,46) The binding affinity of RH-5849 to H. axyridis is close to that against D. melanogaster but 1/30 that of C. suppressalis. The binding affinity of tebufenozide to H. axyridis is 1/10 that of D. melanogaster but 1/5000 that against C. suppressalis, which is consistent with our previous results for selective activity among the three insect orders.47) The insecticidal activity of DAH-type compounds is measured against H. axyridis, which is 1/50 that against L. decemlineata.48) The potency of RH-5849 and tebufenozide is also measured against another Coleoptera Anthonomus grandis to be 44.6 µM and 7.09 µM, respectively.49)
As described above, ligand binding sites are different between steroid and non-steroid bound forms. According to the PDB database, 17 amino acids (E309,48) Q310, I339, M342, T343, T346, R383, V384, R387, L396, F397, A398, Y408, M413, V416, L420, and N504) are participating in PonA-receptor docking, and 12 amino acids (T343, S377, M380, V384, Y408, M413, V416, L420, L500, Q503, N504, and M507) are in BYI06830-HvEcR (PDB: 1R20) binding. Among residues surrounding ligand molecules, the underlined 7 amino acids (T343, V384, Y408, M413, V416, L420, and N504) are common for the binding of both steroidal and non-steroidal ligands in HvEcR.22) Among them two amino acids (T343 and Y408) are implicated in the hydrogen-bond interaction network of the ligand–receptor complex for both non-steroid and steroid compounds. T343 formulates HB toward C=O near the t-Bu group of BYI06830 and 14-OH of PonA, and Y408 does for NH and 20-OH of PonA. N504 is only implicated in the hydrogen-bond interaction toward the other carbonyl group of BYI06830. Although these three residues (T343, Y408, and N504) are conserved in both HaEcR and EvEcR, the binding affinity of DAH-type compounds was significantly different between H. axyridis and E. vigintioctopunctata. In this ligand–receptor docking model, only one HB between C=O and Y408 (Y113 in Fig. 7) was observed, although Y408 of Lepidoptera HvEcR is formulating an HB to NH group. Hydrogen-bond interactions of the BYI06830 molecule to EcRs observed in the docking simulations are different between HaEcR and EvEcR. Even though it is still not easy to predict the ligand–receptor binding by in silico simulation, the models will give us ideas for drug discovery and optimization of structure.
Valine (V384) was replaced with methionine (M) in most of other insect species and other arthropods (T384 for scorpion, G384 for mite, A384 for Crustacea) (Fig. S2). These changes may be due to the reduction or disappearance of the activity of DAHs. Tebufenozide did not bind to the scorpion EcR.4) In HaEcR and EvEcR, valines at 384 and 416 were also changed to methionine (M384) and threonine (T416), respectively. These differences may be the reason for the reduction of the binding affinity of DAHs to HaEcR and EvEcR. In the FRED and GOLD docking simulations, we could not find the significant difference in the best docking results between HaEcR and EvEcR. The selective binding activity of DAH-type compounds to these ladybird beetles may be due to the small difference of ligand–receptor interaction between HaEcR and EvEcR. Although the IC50 values of RH-5849 and tebufenozide were determined for HaEcR, they were very weak compared to other ligand–receptor binding activity.47) The physicochemical reasoning for why DAH-type compounds weakly bind to HaEcR but not to EvEcR is not disclosed. It might be due to the trivial difference of the overall shapes of ligand binding pockets between HaEcR and EvEcR. In HaEcR all the docked structures are located in similar positions, but they are fairly varied in EvEcR as shown in Fig. 8.
Billas and co-workers compared the transcriptional activity between wild and mutated EcRs, demonstrating that the activity was markedly reduced in N504A and W526A mutants.22) This was explained by the change of ligand binding and ligand-mediated stabilization of the agonist H12 conformation. This agreed with the experiment for the D. melanogaster EcR mutant (W650A: corresponding to W526 in HvEcR), where neither ligand binding nor transactivation activity is observed.50) It is also reported that A398 is critical to discriminate between DAHs and steroidal compounds (20E, PonA, and MurA) using Lepidoptera CfEcR mutants.51) In Coleoptera EcRs, this residue is changed to valine (V398). Even though the compounds tested against HaEcR and EvEcR are only RH-5849 and tebufenozide, other DAH analogs or other ecdysone agonists12) may be potent against them.
In conclusion, two EcR isoforms (A and B1) and two USP isoforms (USP-1 and USP-2) were identified from H. axyridis and E. vigintioctopunctata by cDNA cloning. Open reading frames (ORFs) of EcR-A and USP-1 were inserted to the protein expression plasmids to prepare EcR and USP proteins. The binding assay was performed toward the EcR/USP heterodimer that is prepared using a rabbit reticulocyte lysate in an in vitro translation system. The IC50 values of PonA were determined for EvEcR and HaEcR and are consistent with the activity previously determined against other EcRs. RH-5849 and tebufenozide weakly bound to the EcR of the beneficial insect H. axyridis but were not active against E. vigintioctopunctata. Although we could not determine the reason for the difference in the binding affinity of DAHs between HaEcR and EvEcR, the obtained evidence provides new insight into the advanced knowledge of selective toxicity for the biorational drug design of an EcR agonist. It becomes possible to execute the chemical control without disturbing biological control, which fits with the concept of the IPM program.
