Heat Shock Protein 83 (Hsp83) Facilitates Methoprene-tolerant (Met) Nuclear Import to Modulate Juvenile Hormone Signaling*

Background: Methoprene-tolerant (Met) and Germ-cell expressed belonging to the bHLH-PAS family have been identified as juvenile hormone (JH) receptors in Drosophila. Results: Physical interaction with Hsp83 facilitates nuclear import of Met and JH action. Conclusion: Hsp83 modulates JH signaling through mediating the nuclear localization of Met. Significance: Our study helps in understanding the complicated molecular mechanisms of JH signaling. Juvenile hormone (JH) receptors, methoprene-tolerant (Met) and Germ-cell expressed (Gce), transduce JH signals to induce Kr-h1 expression in Drosophila. Dual luciferase assay identified a 120-bp JH response region (JHRR) in the Kr-h1α promoter. Both in vitro and in vivo experiments revealed that Met and Gce transduce JH signals to induce Kr-h1 expression through the JHRR. DNA affinity purification identified chaperone protein Hsp83 as one of the proteins bound to the JHRR in the presence of JH. Interestingly, Hsp83 physically interacts with PAS-B and basic helix-loop-helix domains of Met, and JH induces Met-Hsp83 interaction. As determined by immunohistochemistry, Met is mainly distributed in the cytoplasm of fat body cells of the larval when the JH titer is low and JH induces Met nuclear import. Hsp83 was accumulated in the cytoplasm area adjunct to the nucleus in the presence of JH and Met/Gce. Loss-of-function of Hsp83 attenuated JH binding and JH-induced nuclear import of Met, resulting in a decrease in the JHRR-driven reporter activity leading to reduction of Kr-h1 expression. These data show that Hsp83 facilitates the JH-induced nuclear import of Met that induces Kr-h1 expression through the JHRR.

Juvenile hormones (JH) 2 are a family of sesquiterpenoid compounds synthesized primarily by the corpus allatum (CA) in insects. JH and ecdysteroids (20-hydroxyecdysone (20E) is the most active form) coordinately control insect molting and metamorphosis. One major function of JH is to inhibit some actions of 20E during the larval stages and therefore, JH is usually referred to as the "status quo" hormone (1). In the fruit fly, Drosophila melanogaster, a dipteran insect, the status quo action of JH is subtle but functionally important. As demonstrated in a JH-deficient animal from which the CA is genetically ablated, JH prevents 20E-triggered apoptosis of the larval fat body (2) and precocious differentiation of the optic lobe of adult brain (3) during the larval-pupal metamorphosis in Drosophila.
Significant progress has been made on the identification of the JH receptor and JH signal transduction pathway during the last two decades. Based on a genetic screen to isolate mutants that are resistant to a JH agonist, methoprene, in Drosophila, methoprene-tolerant (Met), a transcription factor belonging to the bHLH-PAS family, was suggested to be involved in JH action (4,5). Met forms homodimers or heterodimers with its paralog, Germ-cell expressed (Gce), and JH reduces this dimerization (6). Although both Met and gce null mutants are viable, the Met gce double mutant dies during the larval-pupal transition (7), resembling what is seen in the JH-deficient animal (2,3). Functionally, Met and Gce mediate JH action to prevent 20E-triggered metamorphic events and thus maintaining the status quo action (2,3,7). In the red flour beetle, Tribolium castaneum, reduction of Met expression by RNAi results in precocious metamorphosis (8,9). Importantly, both Drosophila and Tribolium Met bind to JH at physiological concentrations in vitro (10,11), suggesting strongly that Met is a bona fide JH receptor.
Besides Met and Kr-h1, other proteins are also involved in JH signaling. The p160/SRC/NCoA-like molecule (known as Taiman in Drosophila, FISC in Aedes, and SRC in Tribolium and Bombyx), which belong to bHLH-PAS family transcriptional regulators, directly interact with Met and is required for JH signal transduction (15,(17)(18)(19). Cycle, which is also a bHLH-PAS transcriptional regulator, forms a heterodimer with Met to mediate circadian expression of JH-induced genes in Aedes (16). JH response elements (JHRE) in the Aedes AaET promoter and Bombyx, Aedes, and Tribolium Kr-h1 promoters have been identified as a C box sequence (CACGCG, an E-boxlike motif) or a canonical E box sequence (CACGTG), which are essential for JH-induced gene expression through Met (16 -19). A recent study in Aedes revealed that a consensus 9-mer Metbinding motif, CACG(C/T)G(A/G)(T/A)G, is conserved in the promoter regions of many JH response genes (20).
To better understand the JH signal transduction pathway in Drosophila, we identified a JHRR (which contains three E-boxlike motifs) in the Kr-h1 promoter and used it to purify nuclear proteins bound to the JHRR. The chaperone protein, heat shock protein 83 (Hsp83), was identified as one of the proteins bound to JHRR in the presence of JH. We further illustrate that, in the presence of JH, Hsp83 interacts with Met and facilitates its nuclear import. Met then participates in induction of Kr-h1 expression through the JHRR.
Synchronization of Larval Development-The synchronization was performed at egg laying for studies at 96 h AEL (equal to the end of day 1 of the third instar or the onset of day 2 of the third instar) and at the initiation of wandering for studies at 3 h after the initiation of wandering (AIW) (21). The JH titer (22,23), JH biosynthesis by the CA (24), and jhamt mRNA levels (25) are low at 96 h AEL and high at 3 h AIW, respectively.
␤-Galactosidase Activity Assay and Immunohistochemistry-The larval fat body tissues were collected from different Drosophila lines at 96 h AEL and 3 h AIW. To test the expression of JHRR-LacZ, ␤-galactosidase activity was measured using the Animal Tissue ␤-galactosidase Activity Detection Kit (GENMED, China) according to the manufacturer's instructions.
Cell Culture-Drosophila Kc cells were routinely maintained in Schneider's Drosophila medium (Lonza, Switzerland) supplemented with 5% fetal bovine serum (HyClone). When needed, 1 M methoprene (stock concentration is 100 g/l, dissolved in DMSO) (Dr. Ehrenstorfer GmbH, Germany), 10 g/ml of cycloheximide (Sigma), and/or 1 g/ml of geldanamycin (GA) (Cell Signaling) were added to Kc cells. In these experiments, DMSO was used as a control. Transient transfection assays were carried out for 48 h using Effectene (Qiagen). The final DNA concentration was 0.2 g/ml, and the DNA:Effectene ratio used was 1:25 (26 -29). Using the T7 RiboMAX Express RNAi System (Promega), dsRNAs of Met, gce, Hsp83, and EGFP (as a control) were synthesized. Reduction of gene expression by RNAi in Kc cells was performed by transfecting dsRNAs using Effectene at a final concentration of 2 g/ml of dsRNA (26 -29). The transfected cells were cultured for 48 h.
Culture of Fat Body Tissues-The larval fat body tissues were collected from different Drosophila lines at 96 h AEL and 3 h AIW and cultured in Schneider's Drosophila medium supplemented with 5% fetal bovine serum. When needed, 1 M methoprene and/or 1 g/ml of GA were added to the medium.
Dual Luciferase Assay-For identification of the Kr-h1 JHRR, a 2-kb region of the Kr-h1␣ promoter upstream of the transcription start site was cloned into SacI and EcoRI sites of pGL3 basic vector containing the hsp70 minimal promoter. The deletions and mutations of promoter regions were also constructed into pGL3. After transient transfection in Kc cells with the reporter pGL3 vector and the reference pRL vector, the dual luciferase assays were performed as previously described (26 -29).
Chromatin Immunoprecipitation Assay-Kc cells were transfected with the Met-V5 expression plasmid. At 48 h after transfection, the cells were exposed to 1 M methoprene (or DMSO as a control) for 10 min. Then the cells were fixed and subjected Hsp83 Facilitates Met Nuclear Import OCTOBER 3, 2014 • VOLUME 289 • NUMBER 40 to a chromatin immunoprecipitation assay (ChIP) using the Pierce TM agarose ChIP Kit (Thermo) and the V5 mouse monoclonal antibody (R960-25, Invitrogen). Mock immunoprecipitations with preimmune serum were used for negative controls. The precipitated DNA and input were analyzed by qPCR to detect binding between Met-V5 and JHRR using primers 5Ј-CCACACCATGATGAAGCGCA-3Ј and 5Ј-CTGGCAGA-GTTCGCGTGGC-3Ј. In addition, the binding between Met-V5 and a 5Ј-UTR of Kr-h1␣ was also performed as a control by qPCR using primers 5Ј-GAATACGACATAACAGC-C-3Ј and 5Ј-CGATTTCCGTGAATATGTTCT-3Ј.
Affinity Purification and Identification of Nuclear Proteins Bound to JHRR-The nuclear proteins isolated from methoprene-treated Kc cells were used for affinity purification of nuclear proteins bound to the JHRR. Cells in 10 ϫ 100-mm tissue culture dishes were treated with 1 M methoprene (or DMSO as a control) for 10 min, and the cell pellets were collected in 50-ml tubes after centrifugation for 5 min at 1,000 ϫ g. As published previously (29,30), the nuclear proteins were extracted according to the manufacturer's protocols. Total protein content in the extracts was quantified by a BCA protein assay kit (Beyotime) and the extracts were stored at Ϫ80°C until use. JHRR fragments were synthesized by PCR amplification and phosphorylated using T4 polynucleotide kinase, followed by ligation with T4 DNA ligase to produce the polymerized JHRRs (31). Polymerized JHRRs were then coupled to commercial CNBr-activated Sepharose 4B resin (GE Healthcare) according to the manufacturer's instructions. The resulting polymerized JHRR affinity column was then used to isolate nuclear proteins bound to the JHRR as described previously (30). The eluted JHRR-binding proteins were digested with trypsin and subjected to HPLC-MS/MS analysis (Shanghai Applied Protein Technology, China) (2,28).
Immunoprecipitation-Met-V5 (its mutated constructs and Flag-gce) and Myc-Hsp83 were cloned into pAc5.1 (Invitrogen) to generate the expression constructs. The constructs were transfected into Kc cells, at 48 h after transfection, the cells were harvested and lysed in ice-cold NP-40 lysis buffer (Beyotime, China). The procedure for immunoprecipitation has been previously described in detail (26,28). The immunoprecipitates were separated by SDS-PAGE and analyzed by Western blotting.
Ligand Binding Assay-Dextran-coated charcoal assay was performed as described previously (10,11). Kc cells were transfected with Met or EGFP overexpression plasmid. Prior to harvesting, cells were treated with GA for 5 h. The total proteins were extracted and incubated with 3 H-JH III (20 Ci/mmol, PerkinElmer Life Sciences; final concentration ϭ 100 nM) for 90 min at 22°C. This was followed by the addition of 5% charcoaldextran solution, gentle mixing for 2 min, and centrifugation for 1 min at 3,000 ϫ g. The supernatant was collected into a scintillation vial and incubated overnight at 4°C. Radioactivity was measured using a liquid scintillation counter (LS-6000IC, Beckman).
Statistics-Data were analyzed using the Student's t test and analysis of variance (ANOVA), t test: *, p Ͻ 0.05; **, p Ͻ 0.01. ANOVA analysis is shown as the bars labeled with different lowercase letters as significantly different (p Ͻ 0.05). The mean Ϯ S.D. from at least three independent experiments is shown.

Identification of Response Regions to Methoprene in Kr-h1
Promoter-Kr-h1 was first identified as a JH response gene in Drosophila (12). We here found that methoprene induces Kr-h1 expression in a rapid and reversible manner in Drosophila Kc cells (Fig. 1A). The mRNA level of Kr-h1 increased about 5-fold within 5 min after methoprene treatment, gradually increased up to 8-fold by half an hour, and the induction was steadily decreased after 1 h (Fig. 1A). Moreover, the rapid induction of Kr-h1 expression by methoprene was not blocked by a protein synthesis inhibitor, cycloheximide, suggesting that Kr-h1 is a JH primary-response gene (Fig. 1B).
To identify response regions to methoprene in the Kr-h1 promoter, we employed the dual luciferase assay system and Kc cells. A 2-kb Kr-h1␣ promoter region (Ϫ2000 to ϩ1 bp upstream of the transcriptional start site) was cloned into pGL3 vector. This promoter region supported ϳ10and 140-fold increase in the luciferase activity within 2 and 12 h after methoprene treatment, respectively. Three other regions (E1, E2, and E3, about 120 bp in each region) containing canonical E-boxes of possible JHREs (17), which are present in the promoter region beyond 2 kb from the transcriptional start site and the first intron of Kr-h1␣ (Fig. 1C), were also cloned into pGL3. Among these three regions, E2 showed response to methoprene treatment similar to the 2-kb promoter region, whereas E1 and E3 showed lower response (Fig. 1D). Although the 2-kb pro- moter region and E2 showed the highest response to methoprene, it is likely that all four regions are involved in JH-induced Kr-h1 expression.

Identification of a JHRR in 2-kb Promoter Region of Kr-h1-
The 2-kb promoter region of Kr-h1␣ was chosen for further studies because it showed high response to methoprene treatment and is located right before the transcriptional start site. Subsequent reporter assays with a series of deletion constructs (Ϫ2000 to Ϫ863, Ϫ998 to ϩ1, Ϫ863 to Ϫ649, and Ϫ440 to Ϫ224) showed that the region crucial for response to methoprene is located between Ϫ440 and Ϫ224 ( Fig. 2A). Further reporter analyses with shorter deletion constructs (Ϫ440 to Ϫ336, Ϫ336 to Ϫ224, Ϫ348 to Ϫ224, Ϫ440 to Ϫ313, Ϫ440 to Ϫ321, and Ϫ386 to Ϫ321) identified a 120-bp region (Ϫ440 to Ϫ321) that is responsible for the majority of response to methoprene and is referred to as the JHRR (Fig. 2B).
We noticed that the Ϫ386 to Ϫ321 region, which showed a much lower response to methoprene than the JHRR, contains a single C box (CACGCG), whereas the JHRR contains two additional B boxes (CATGTG) (Fig. 2, B and C). bHLH-PAS transcription factors are known to bind E-box-like motifs (CANNTG), including both C box and B box (32). To elucidate the individual contribution of each E-box-like motif in the JHRR, point mutations were introduced. Mutations to both B boxes resulted in significant reduction in the JH-induced luciferase activity, whereas a single mutation to the first B box caused no obvious reduction. However, mutation to the C box or all three E-box-like motifs completely abolished the luciferase activity (Fig. 2D). Moreover, the pGL3 reporter constructs carrying two copies of the two B boxes or four copies of the C box showed no response to methoprene within 2 h. The former construct still showed no induction of luciferase activity after 12 h of methoprene treatment, whereas the later construct showed 4.5-fold induction (Fig. 2E). In conclusion, all three E-box-like motifs in the JHRR are required for full response, but their relative contributions require further studies. Met and Gce Transduce JH Signals to Induce Kr-h1 Expression through JHRR-It has been documented that Met and Gce transduce JH signals to induce Kr-h1 expression in Drosophila (7). In Kc cells, the JH-induced Kr-h1 expression (Fig. 3A) and JHRR-driven luciferase activity (Fig. 3B) were significantly decreased when the mRNA levels of Met and gce were simultaneously reduced by RNAi (Fig. 3, C and CЈ). To avoid possible off-targeting effects of RNAi (33), multiple sets of Met and gce dsRNAs were generated and similar results were obtained (data not shown).
Similar to the previous report that JH fails to induce Kr-h1 expression in the Met gce double mutant, Met 27 gce 2.5k , during larval molts (7), Kr-h1 expression was significantly reduced in the fat body at 3 h AIW (Fig. 3D). In comparison with 3-h AIW when the JH titer is high, Kr-h1 expression was much lower in the wild type w 1118 larvae at 96 h AEL when the JH titer is low (Fig. 3DЈ). Moreover, at 96 h AEL, Kr-h1 expression was significantly enhanced by the addition of methoprene to the cultured fat body tissues isolated from the w 1118 larvae (Fig. 3DЉ, left panel) but not from the Met 27 gce 2.5k larvae (Fig. 3DЉ, right  panel).
We further produced JHRR-LacZ transgenic flies, and put JHRR-LacZ in the Met gce double mutant background to generate Met 27 gce 2.5k /y; JHRR-LacZ/ϩ. At 3 h AIW, the JHRRdriven ␤-galactosidase activity was very high in the fat body of JHRR-LacZ larvae, but became nearly undetectable in Met 27 gce 2.5k /y; JHRR-LacZ/ϩ (Fig. 3E). In comparison to the 3-h AIW, the JHRR-driven ␤-galactosidase activity was lower in the JHRR-LacZ larvae at 96 h AEL (Fig. 3EЈ). Moreover, at 96 h AEL, the JHRR-driven ␤-galactosidase activity was significantly increased by addition of methoprene to fat body tissues isolated from JHRR-LacZ larvae (Fig. 3EЉ, left panel) but not from the Met 27 gce 2.5k /y; JHRR-LacZ/ϩ larvae (Fig. 3EЉ, right panel). Taken together, these in vitro and in vivo data show that Met and Gce transduce JH signals to induce Kr-h1 expression through the JHRR.
Furthermore, we performed ChIP to examine whether Met directly binds to the JHRR, which contains three E-box-like motifs, for inducing Kr-h1 expression. The binding of Met-V5 to DNA was detected using the V5 antibody and cross-linked chromatin isolated from Kc cells, which were transfected with the Met-V5 expression plasmid. As measured by qPCR, V5 antibody weakly precipitated JHRR DNA, and the addition of methoprene increased the amount of precipitated JHRR DNA by 5-fold (Fig. 3F, left panel). By contrast, V5 antibody did not precipitate DNA from the 5Ј-UTR region of Kr-h1␣, which lies in the first exon of Kr-h1␣ and contains no E-box-like motifs, and methoprene did not increase the precipitation (Fig. 3F,  right panel). The ChIP-qPCR experimental data conclusively showed that Met directly binds to the JHRR for inducing Kr-h1 expression.
Identification of Hsp83 as One of the Proteins Bound to JHRR-To identify the proteins binding to the JHRR, nuclear proteins isolated from methoprene-or DMSO-treated Kc cells were purified using the polymerized JHRR affinity column and subjected to HPLC-MS/MS analysis. The chaperone protein Hsp83 (homolog of the mammalian Hsp90) was identified as a JHRR-bound nuclear protein present only in purified nuclear proteins isolated from methoprene-treated Kc cells but not from the DMSO-treated cells (Fig. 4). The other JHRR-bound nuclear proteins identified in either control or methoprenetreated Kc cells are listed in Tables 1 and 2.
Hsp83 Is Required for JH-induced Kr-h1 Expression-The mammalian Hsp90 is critical for regulating transcriptional activity by interacting with ligand-activated transcription factors, including some nuclear receptors such as glucocorticoid receptor (GR) (34,35) and a bHLH-PAS transcription factor, the aryl hydrocarbon receptor (AhR) (36,37). In Drosophila, Hsp83 is necessary for activating the 20E nuclear receptor complex, EcR-USP, that mediates 20E signal transduction (38). To this end, we investigated whether Hsp83 is involved in the JHinduced Kr-h1 expression.
First, we tested this hypothesis in Kc cells. The JH-induced Kr-h1 expression (Fig. 5A) and JHRR-driven luciferase activity (Fig. 5B) were significantly decreased when the cells were pretreated with GA, a potent and highly specific inhibitor of Hsp90/Hsp83, which binds to the ADP/ATP binding pocket located in the N terminus of Hsp90/Hsp83 for inactivating its ATPase activity (39), or after the mRNA level of Hsp83 was reduced by RNAi.
Next, we tested this hypothesis in the larval fat body at 3 h AIW. In comparison with the w 1118 larvae, the Kr-h1 mRNA level of an Hsp83 weak mutant, Hsp83 08445 , was reduced to about 50%. To avoid any non-tissue autonomous effects, the binary GAL4/UAS system was used to reduce Hsp83 expression specifically in the larval fat body by RNAi. Kr-h1 expression was reduced to ϳ25% in two lines of the Lsp2-GAL4ϾUAS-Hsp83-RNAi larvae (32996 and 33947) (Fig. 5C). Moreover, the JHRRdriven ␤-galactosidase activity in JHRR-LacZ::Lsp2-GAL4/ UAS-Hsp83-RNAi larvae was decreased to about 15% of that in the control larvae, JHRR-LacZ::Lsp2-GAL4/ϩ (Fig. 5D). These data demonstrate that Hsp83 is required for JH-induced Kr-h1 expression through the JHRR and suggest that Hsp83 might influence the action of Met in JH signal transduction.
Hsp83 Interacts with Met through Its PAS-B and bHLH Domains-It is well known that mammalian Hsp90 physically interacts with AhR to facilitate its transcriptional activity (36,37). Thus, we investigated whether Hsp83 physically interacts with Met after Myc-Hsp83 and Met-V5 were co-overexpressed in Kc cells. Co-immunoprecipitation experiments revealed that Met associated with Hsp83 and JH was able to induce physical interaction between Met and Hsp83 (Fig. 6, A and B). Importantly, GA significantly decreased the basal and JH-induced Met-Hsp83 interaction (Fig. 6B). We then individually mutated the bHLH, PAS-A, and PAS-B domains of Met. Mutation to the PAS-B domain of Met abolished the Met-Hsp83 interaction, mutation to the bHLH domain significantly decreased the interaction, but mutation to the PAS-A domain had no obvious effect (Fig. 6C), indicating that the PAS-B and bHLH domains of Met are essential for the Met-Hsp83 interaction. In addition, Gce also associated with Hsp83, and GA decreased the basal and JH-induced Gce-Hsp83 interaction (Fig. 6D).
Hsp83 Is Required for JH Binding to Met-The JH binding activity of total proteins extracted from the Met-overexpressed Kc cells was signfiicatnly higher than proteins isolated from control Kc cells overexpressing EGFP. Importantly, GA  OCTOBER 3, 2014 • VOLUME 289 • NUMBER 40

JOURNAL OF BIOLOGICAL CHEMISTRY 27879
treatment had no apparent effect on the JH binding activity of total proteins extracted from EGFP-overexpressed cells, but decreased that of the Met-overexpressed cells to the control level (Fig. 6E), suggesting that Hsp83 facilitates JH binding to Met.
Hsp83 Facilitates JH-induced Met Nuclear Import-The mammalian Hsp90 regulates subcellular localization of the ligand-activated GR (40,41) and AhR (36). It has been reported that Met is predominantly localized in the nuclei of Drosophilacultured cells (10,42) and tissues (43). As detected by immunohistochemistry using a V5 antibody, Met-V5 was predominantly localized in the nuclei of fat body cells isolated from Lsp2-GAL4ϾUAS-Met-V5 (Fig. 7A, first image in the upper  panel) and Lsp2-GAL4::UAS-Met-V5 (Fig. 7A, first image in the lower panel) larvae at 3 h AIW. Interestingly, the nuclear localization of Met-V5 was inhibited when fat body tissues were exposed to GA (Fig. 7A, second image in the upper panel). To confirm the GA effects, we examined whether mutation to or RNAi of Hsp83 affects subcellular localization of Met-V5 at 3 h AIW. Consistent with the GA effects, Met-V5 became mainly distributed in the cytoplasm of the fat body cells isolated from the Hsp83 08445 ::UAS-Met-V5/Hsp83 08445 ::Lsp2-GAL4 larvae (Fig. 7A, third image in the upper panel) and two lines of the Lsp2-GAL4::UAS-Met-V5/UAS-Hsp83-RNAi larvae (Fig. 7A,  second and third images in the lower panel). To avoid any noncell autonomous effects of Hsp83, we further used the Flp-out technique to detect the cytoplasm-nucleus shuttling of Met-V5 in clones of Met-V5-overexpressed fat body cells, which were marked with GFP, at 3 h AIW. Again, Met-V5 was predominantly localized in the nuclei of cloned Met-V5-overexpressed fat body cells isolated from the hsFlpase; ActϾCD2ϾGAL4; UAS-GFP/UAS-Met-V5 larvae (Fig. 7B, first images in both panels), and GA significantly inhibited the nuclear import of Met-V5 resulting in predominant localization of Met-V5 in the cytoplasm (Fig. 7B, second images in both panels). Moreover, Met-V5 was mainly distributed in the cytoplasm of the cloned  fat body cells isolated from the hsFlpase; ActϾCD2ϾGAL4; UAS-GFP/UAS-Met-V5::UAS-Hsp83-RNAi larvae (Fig. 7B, third images in both panels). These data demonstrate that Hsp83 is required for Met nuclear import. We also detected the subcellular localization of Met-V5 in Lsp2-GAL4ϾUAS-Met-V5 larvae at 96 h AEL. To our great surprise, Met-V5 was mainly distributed in the cytoplasm of fat body cells at this stage. Moreover, when fat body tissues were exposed to methoprene, Met-V5 shifted from the cytoplasm to the nucleus, and the methoprene-induced nuclear import of Met-V5 was dramatically inhibited by GA (Fig. 7C). Consistently, Met-V5 was mainly distributed in the cytoplasm of cloned Met-V5-overexpressed fat body cells in the hsFlpase; ActϾCD2ϾGAL4; UAS-GFP/UAS-Met-V5 larvae at 96 h AEL, methoprene induced the nuclear import of Met-V5, and GA blocked the methoprene-induced nuclear import of Met-V5 (Fig. 7D). The combined genetic data demonstrate that Hsp83 facilitates the JH-induced nuclear import of Met in the larval fat body.
Hsp83 Is Accumulated in the Cytoplasm Area Adjunct to the Nucleus in the Presence of JH and Met/Gce-Because JH induces Met nuclear import, we investigated whether JH also induces Hsp83 nuclear import at 96 h AEL. As detected by immunohistochemistry, Hsp83 was almost evenly distributed in the fat body cells isolated from w 1118 larvae (Fig. 8A). When the fat body tissues were exposed to methoprene, Hsp83 was accumulated in the cytoplasm area adjunct to the nucleus (Fig.  8AЈ). Importantly, Hsp83 was slightly accumulated in the cytoplasm area adjunct to the nucleus in fat body cells isolated from the Lsp2-GAL4ϾUAS-Met-V5 larvae (Fig. 8B), whereas methoprene further increased this accumulation (Fig. 8BЈ). Moreover, it appears that localization of Hsp83 was evenly distributed in the fat body cells isolated from the Met 27 gce 2.5k larvae (Fig. 8C), and methoprene had no effect on its subcellular localization (Fig. 8CЈ). These data show that Hsp83 is accumulated in the cytoplasm area adjunct to the nucleus in the presence of JH and Met/Gce.

DISCUSSION
Mammalian Hsp90 forms a molecular chaperone complex with other co-chaperone proteins, including Hsp70, Hop, p23, and a tetratricopeptide repeat-domain protein to regulate the correct maturation and activation of the "client proteins" (44). Ligand-activated transcription factors (e.g. GR and AhR) usually undergo a constant and dynamic cytoplasm-nucleus shuttling, which is regulated by the Hsp90 chaperone complex (35,36,40,41,44,45). In this study Hsp83 was identified as a nuclear protein bound to the JHRR in the presence of JH (Fig. 4). Importantly, we have shown that Hsp83 facilitates JH binding and JH-induced nuclear import of Met, which then binds to the JHRR and induces Kr-h1 expression (Figs. 3-7). Surprisingly, Met, Gce, Taiman, and Cycle were not detected in either nuclear protein isolated from control and methoprene-treated  Kc cells in affinity purification experiments. This is probably due to their relatively low abundance in nuclear proteins. As shown by cross-linking studies for several nuclear receptors, the stoichiometry of the receptor to Hsp90 is 1:2 (34), so the amount of Hsp83 bound to JHRR should be relatively more abundant than Met/Gce and/or Taiman/Cycle in the nuclear proteins.
Previous studies have shown that the PAS-B domain of Met accounts for Met-Met and Met-Gce dimerization (6) as well as JH binding (11), and that its PAS-A and PAS-B domains are indispensable for Met-FISC dimerization (18). Apparently, bHLH, PAS-A, and PAS-B domains of Met have overlappng binding sites, which also occur in AhR (37). In the absence of JH, Met/Gce forms dimers, and JH reduces Met/Gce dimerization (6) and induces the Met-Hsp83 interaction (Fig. 6, A-C). Moreover, GA attenuated the JH-binding activity of total proteins isolated from Met-overexpressed Kc cells (Fig. 6E), suggesting that the JH-induced Met-Hsp83 interaction is important for Met to form a proper conformation for binding to JH.
It has been long considered that in Drosophila, Met localizes in the nuclei no matter whether JH is present or not (10,42). Inconsistent with those reports, we observed that Met-V5 is predominantly localized in the nuclei of fat body cells at 3 h AIW when the JH titer is high. In contrast, Met-V5 is mainly distributed in the cytoplasm of fat body cells at 96 h AEL when the JH titer is low, and moreover, JH induces Met-V5 nuclear import (Fig. 7). It was reported that the JH-binding activity of proteins in the cytoplasm isolated from fat body cells of the Met mutants was much lower than that of wild type larvae (46), suggesting that Met may also present in the cytoplasm, at least in some circumstances. A constant and dynamic cytoplasmnucleus shuttling of Met and the JH-induced Met nuclear import are supported by the identification of nuclear import and export signals of Met, which are predominantly localized in PAS-A and PAS-B domains, respectively (42). In conclusion, similar to the other ligand-activated transcription factors, nuclear import of the JH receptor Met is aided by Hsp83 in Drosophila (Fig. 7). Moreover, Hsp83 is accumulated in the cytoplasm area adjunct to the nucleus in the presence of JH and Met/Gce (Fig. 8). This increase is likely due to an increase in import of Met/Gce, JH, and Hsp83 complex. A recent report on Helicoverpa armigera (Lepidoptera) also suggests that both 20E and JH stimulate nuclear import of Hsp90/Hsp83 in cultured cells (47). Nup358, a nuclear pore protein containing a tetratri- copeptide repeat domain, was also identified as a JHRR-bound nuclear protein from methoprene-treated Kc cells (Table 1). Our preliminary data imply Hsp83 physically interacts with the tetratricopeptide repeat domain of Nup358, suggesting that this protein may be involved in Hsp83-facilicated JH-induced Met nuclear import and thus JH signaling (data not shown). These observations suggest that whereas Hsp83 is involved in JH-induced Met nuclear localization, it is not sufficient to complete the process and other components are necessary.
Hsp90/Hsp83 does not contain a nuclear localization signal. In mammalian cells, nuclear import of the GR-Hsp90 complex through the nuclear pore complex is mediated by its interaction with Nup62, the integral nuclear pore glycoprotein, and Importin, the classical pathway of nuclear import. By interacting with the Hsp90 chaperone complex, motor proteins are involved in passing the GR-Hsp90 complex through the nuclear pore (40,41,44,45). Whether the above molecular mechanism applies to the JH-induced Met nuclear import remains unknown and requires further studies.
Based on the data presented in this paper and previous studies, we propose a model that Hsp83 facilitates Met nuclear import to modulate JH signaling in Drosophila (Fig. 9). Prior to  binding to JH in the cytoplasm, Met remains in an inactive conformation and only weakly associates with Hsp83, and JH induces the physical interaction between the PAS-B and bHLH domains of Met and Hsp83. In turn, the JH-induced Met-Hsp83 interaction permits a proper conformational change in Met and an increase in binding to JH. Thus, JH induces nuclear import of the Met-Hsp83 complex, possibly along with other transcription co-factors (e.g. Taiman or Cycle). Upon entering the nucleus, the JH-bound Met directly binds to the JHRR and induces the expression of JH response genes such as Kr-h1. It is well documented that the mammalian Hsp90 chaperone complex is critical for regulating transcriptional activity of several ligand-activated transcription factors (i.e. GR and AhR). However, this is the first time to show that Hsp90/Hsp83 plays an essential role in the regulation of a bHLH-PAS transcription factor in invertebrates. Interestingly, nuclear import of both AhR and Met is regulated by the Hsp90 Hsp83 molecular chaperone complex, suggesting this mechanism is evolutionarily conserved for the ligand-activated bHLH-PAS transcription factors. This study adds to our knowledge on the comprehensive molecular mechanism of JH signaling and the general roles of bHLH-PAS transcription factors.