The full-length open reading frame of Drosophila Met (NM_078571) and Gce (NM_078605) open reading frame was amplified from the larval fat body cDNA library, and BirA (NC_000913.3) was amplified from E. coli. All the PCR products were cloned into the pClone007 vector (007VS, TsingKe Biotech). BirA was then inserted into the pUAST vector between the EcoRI and XbaI sites to generate a control construct expressing only the BirA protein. For the generation of Met and Gce overexpression lines, vectors were constructed with the Gibson assembly method. First, the pUAST vector was digested by EcoRI/XbaI. The P2A-BirA sequence was amplified from the pClone007-BirA vector. When a P2A peptide (ATNFSLLKQAGDVEENPGP) was added to the N-terminus of BirA, 5 amino acids (GGSGS) were added to the N-terminus, and 3 amino acids (GGS) were added to the C-terminus of the P2A peptide as a linker. The FLAG tag (DYKDDDDK) and an Avi tag (GLNDIFEAQKIEWHE) were added to the 5′ end by reamplifying the above PCR products with FLAG-Avi-P2A-F oligos and the BirA-R oligo. PCR of the coding sequence of Met or Gce from the pClone007-Met or pClone007-Gce vector was performed, and then we used a Hieff Clone Plus Multi One Step Cloning Kit (10912ES10, Yeasen) for assembly. Finally, we obtained the pUAST-Met-4 × FLAG-Avi-P2A-BirA (UAS-Met-BirA) or pUAST-Gce-3 × FLAG-Avi-P2A-BirA (UAS-Gce-BirA) vector, in which BirA was fused at the C-terminus. Meanwhile, we also constructed pUAST-BirA-P2A-Met-3 × FLAG-Avi and pUAST-BirA-P2A-Gce-3 × FLAG-Avi, with BirA fused at the N-terminus, using a similar method to assess the translational efficiency of different termini.

Moreover, the full-length cDNAs of Met, Gce, Hsp83 (NP_523899), 14-3-3ε (NM_169796) and 14-3-3ζ (NM_001273916) were amplified. Met-FLAG, Met-bHLH+(amino acids 1–97)-FLAG, Met-PAS-A+(amino acids 88–196)-FLAG, Met-PAS-B+(amino acids 187–716)-FLAG, Met-PAS-B (amino acids 403–510)-FLAG, Met-PAS-B + ^T265A-FLAG, Met-PAS-B + ^S531A-FLAG, Met-PAS-B + ^T535A-FLAG, Met-PAS-B + ^S652A-FLAG, Met-PAS-B + ^{T265A/S531A/T535A/S652A}-FLAG (4 m), Met-4m-FLAG, Met^{S256A/T265A/S531A/T535A/S651A/S652A}-FLAG (6 m), Gce-FLAG, Gce-bHLH+(amino acids 1–335)-FLAG, Gce-PAS-A+(amino acids 326–415)-FLAG, Gce-PAS-B+(amino acids 406–956)-FLAG, Hsp83-FLAG, Hsp83-V5, Hsp83^S219A-V5, 14-3-3ε-HA, and 14-3-3ζ-HA were separately generated and constructed into the pUAST vector for overexpression in Kc cells.

Western blotting was performed as previously described (He et al., 2014; Gao et al., 2021). In brief, cells were lysed on ice in NP-40 lysis buffer supplemented with protease and phosphatase inhibitor cocktail, and the concentration of the total protein was determined with a BCA reagent. An aliquot of 20 μg of the protein extracts was separated by a 10% SDS-PAGE gel and immediately transferred onto a PVDF membrane. The membranes were then blocked with 5% non-fat milk and incubated with 1: 2000 primary antibodies at 4°C overnight, followed by three washes with PBST. The protein bands were finally incubated with 1: 5,000 secondary antibodies and visualized by chemiluminescence. The primary antibodies used in this study were HRP-linked goat anti-biotin (#7075, Cell Signaling Technology, CST), mouse anti-FLAG (#F1804, Sigma-Aldrich), rabbit anti-V5 (#ab9116, Abcam), mouse anti-HA (#sc-7392, Santa Cruz), rabbit anti-phospho-PKC substrate (#2261, CST), rabbit anti-phospho-CaMKII (#12716, CST), and mouse anti-α-tubulin (#AT819, Beyotime) antibodies, and HRP-labeled secondary antibodies (#sc-2313 and sc-2318, Santa Cruz) were used.

2D gel electrophoresis was performed as previously described (Liu et al., 2015). Total proteins from Kc cells were extracted with NP-40 lysis buffer as described above. The protein extracts from each treatment were precipitated with ice-cold trichloroacetic acid/acetone and were then redissolved in rehydration buffer (8 M urea, 4% 3-[(3-Cholamidopropyl)-dimethyl- ammonio]-1-propane sulfonate [CHAPS], 65 mM DTT, 0.2% ampholytes, and 0.001% bromophenol blue). Equal amounts of the total proteins (150 μg) from each sample were loaded onto an immobilized pH gradient (IPG) strip with the pH ranging from 3 to 10 (Bio-Rad). Isoelectric focusing (IEF) was performed using an Ettan IPGphor 3 IEF System (GE Healthcare) following the manufacturer’s instructions. After IEF, the IPG strips were successively equilibrated in a reducing solution (0.375 M Tris-HCl, pH 8.8, 2% DTT, 6 M urea, 2% SDS, and 20% glycerin) and an alkylating solution (0.375 M Tris-HCl, pH 8.8, 2.5% iodoacetamide, 6 M urea, 2% SDS, 20% glycerin) each with 15 min. The equilibrated strips were then subjected to 2D SDS-PAGE. After electrophoresis, the proteins were transferred to polyvinylidene fluoride (PVDF) membranes for probing with an antibody against biotin.

Forty-six hours following transfection, 10 μM methoprene (stock concentration of 100 μg/μL, dissolved in DMSO) or DMSO was added to Kc cells for 2 h. The cells were then washed with PBS and harvested and lysed in ice-cold NP-40 lysis buffer containing protease and phosphatase inhibitor cocktail and protease inhibitor cocktail. The extracts were clarified by centrifugation at 12000 × g for 15 min at 4°C. The lysates were incubated with a FLAG, V5, or HA antibody for 4 h and then with Pierce protein A/G agarose overnight at 4°C. Beads were collected by slow centrifugation, washed 4 times with lysis buffer and analyzed by SDS-PAGE followed by western blotting.

The subcellular localization of Met-FLAG, Met-bHLH + -FLAG, Met-PAS-A + -FLAG, Met-PAS-B + -FLAG, Met-PAS-B-FLAG, Met-PAS-B + ^T265A-FLAG, Met-PAS-B + ^S531A-FLAG, Met-PAS-B + ^T535A-FLAG, Met-PAS-B + ^S652A-FLAG Met-PAS-B+^4m-FLAG, Gce-bHLH + -FLAG, Gce-PAS-A + -FLAG, Gce-PAS-B + -FLAG, Met-4m-FLAG, Met-6m-FLAG, 14-3-3ε-HA, and 14-3-3ζ-HA was measured by immunohistochemistry as described previously (He et al., 2017; Liu et al., 2018). Briefly, the cells were fixed with 4% formaldehyde for 30 min. After three washes with PBT (PBS with 0.3% Triton X-100 and 0.5% bovine serum albumin), the cells were incubated with primary antibodies (1: 1,000) at 4°C overnight followed by three PBT washes. Incubation of the secondary antibody (1: 1,000) was then performed for 2 h at room temperature. The primary antibodies used were rabbit anti-FLAG (#14793, CST, diluted 1: 200) and mouse anti-HA (#sc-7392, Santa Cruz) antibodies. The secondary antibodies used were Alexa Fluor 488-conjugated goat anti-rabbit IgG (#A11008, Invitrogen, 1: 200 diluted) and Alexa Fluor 594-conjugated goat anti-mouse IgG (#A11032, Invitrogen, 1: 200 diluted). The nuclei were stained with DAPI at 1: 2000 (#C1002, Beyotime). Confocal images were collected on an Olympus FluoView FV3000 confocal microscope.

Total RNA samples were prepared from Kc cells using RNAiso Plus regents (Takara). All qRT-PCR was performed with at least three biological replicates, each with triplicates, by using Hieff qPCR SYBR Green Master Mix (#11202ES03, Yeasen) on a QuantStudio 6 Flex Real-Time PCR system (Applied Biosystems). The final volume of each reaction was 20 μL, and the PCR thermocycling was initiated from 95°C for 2 min, followed by 40 cycles of 95°C for 10 s, and 60°C for 30 s. After the PCR amplification, a melting program was also included to verify the specificity of the products. Relative expressions of the target genes were calculated by the 2^−ΔΔCT method, with rp49 gene as an internal reference for normalization. All the primers involved are summarized in Supplementary Table S7.

pGL3-JHRR (He et al., 2014) and the reference pRL vector were co-transfected into Kc cells with 14-3-3 protein overexpression vectors. Forty-six hours after transfection, 10 μM methoprene or DMSO was added to Kc cells, and the cells were incubated for an additional 2 h. Dual luciferase assays were conducted using a Dual Luciferase Assay System (Promega) as previously reported (Tian et al., 2013).

Unless otherwise specified, all the data subjected to statistics are mean ± standard deviations from independent replicates, and the statistics was analyzed with IBM SPSS Statistics 19.0. Significant differences between two groups and among multi-treatments were determined with the Student’s t-test and one-way analysis of variance (ANOVA) followed by Duncan multiple comparisons test, respectively, at the level of p < 0.05.

The phosphorylation and transcriptional activity of Met is regulated by JH membrane signaling pathway which involves RTK-mediated phosphorylation of CaMKII and PKC (Liu et al., 2015; Ojani et al., 2016; Gao et al., 2021). We thus first confirmed whether methoprene (JH analog, JHA) induced the phosphorylation of PKC and CaMKII in Drosophila Kc cells. In response to JHA, the phosphorylation of both proteins in was elevated at 2–4 h, and the elevation was sustained for 6 h (Figure 1A). By performing 2D gel electrophoresis of the protein extracts followed by western blotting, we detected multiple molecular forms of Met and Gce in a descending order of isoelectric point (pI), and some of these forms migrated to a lower-pH region (Figure 1B). Incubation of JHA-treated protein extracts with λ phosphatase (λpp) before electrophoresis prevented this migration, resulting in a distribution of Met or Gce being similar to that in the control group (Figure 1B). Together with the previous reports (Liu et al., 2015; Ojani et al., 2016; Gao et al., 2021), the above results suggest functional conservation of the PKC and CaMKII in the regulation of phosphorylation of Met and Gce in Drosophila.

To isolate and purify Met/Gce, we used an in vitro biotinylation strategy and then examined the phosphorylation profile of Met/Gce by liquid chromatography-tandem mass spectrometry (LC-MS/MS). We generated vectors that express Met-BirA or Gce-BirA, a synthetic linear construct with an added biotin ligase (BirA) molecule at the C-terminus (Figure 2A). Each target protein in the construct had a FLAG and biotin acceptor peptide (Avi) tag that could be effectively biotinylated by the Escherichia coli BirA enzyme (Beckett et al., 1999; de Boer et al., 2003). We also included a porcine teschovirus-1 2A (P2A) self-cleaving peptide that shows high efficiency in facilitating the expression of multiple transgenes in Drosophila (Daniels et al., 2014; Liu et al., 2017). These constructs as well as a control construct with only the BirA were expressed in Kc cells using the Act5c-Gal4. The protein-biotin strategy was first used to purify JH intracellular receptors and their conjugates in either JHA- or dimethyl sulfoxide (DMSO)-treated cells. To assess the translational efficiency of different termini, we initially tested Met or Gce protein levels in Kc cells. We found that Met and Gce protein levels were higher when BirA was fused at the C-terminus than that at the N-terminus (Figure 2B). For consistency, UAS-Met-BirA and UAS-Gce-BirA were used in all subsequent studies. Both types of cells expressing Met and Gce, which all included biotin and FLAG, showed bands of the same size (Figure 2C), indicating that this approach was viable for our purpose.

To identify the phosphorylation sites, Met and Gce proteins were separated by 10% SDS-PAGE, and the target bands were subjected to protease digestion and MS analysis (Figure 2D). A total of 19 phosphorylation sites of Met were identified from the JHA treatment, two of which were also present in the control DMSO treatment. We identified 6 and 13 phosphorylation sites of Gce in the control and JHA treatment groups, respectively; three sites were shared by both groups (Figure 2E; Supplementary Tables S1–S3), indicating that these sites may be basal phosphorylation sites. Moreover, the identified phosphorylation sites of Met were rich in its N-terminus, whereas most of the phosphorylation sites of Gce were distributed in the C-terminus (Figure 2F). This finding suggests that phosphorylation sites in Met and Gce may differ in their molecular functions.

The nuclear localization of Met depends on the PAS-B domain in the presence of JH (Greb-Markiewicz et al., 2011). We therefore divided the full-length Met into three distinct fragments corresponding to conserved domains predicted by the Conserved Domain Database, including bHLH+ (amino acids 1–97), PAS-A+ (amino acids 88–196) and PAS-B+ (amino acids 187–716) (Figure 3A). Compared to the control, JHA dramatically induced subcellular trafficking of the PAS-B+ part (Figures 3B–I’). By contrast, PAS-B lacked surrounding amino acids was exclusively localized to the cytoplasm (Figures 3J–K’). Likewise, each of the three Gce fragments remained a stable subcellular localization after JH treatment (Supplementary Figure S1). Given the known function of phosphorylation modification in controlling the subcellular localization of transcription factors, these observations suggest that the region surrounding the PAS-B domain of Met that contains phosphorylation sites might affect the subcellular localization of Met.

For Met, there were 6 phosphorylation sites, including 2 basal phosphorylation sites. To determine whether these phosphorylation sites in the PAS-B+ region of Met are essential for JH-induced subcellular trafficking, we mutated either phosphorylated serine (S) or threonine (T) to alanine (A) and generated 5 phosphor-inactivated PAS-B+ mutant constructs. All these mutant forms failed to prevent their subcellular trafficking in the presence of JHA (Figure 4). Likewise, even if all 4 phosphorylation sites were mutated in full-length Met (Met-4m), nuclear localization was maintained (Figures 5A–C’). These findings indicate that the phosphorylation sites surrounding the PAS-B domain of Met are dispensable for nuclear import of Met. Moreover, the Kr-h1 transcriptional level was similar to that in intact Met-overexpressing cells after JHA exposure (Figure 5D).

Since the bHLH domain is responsible for DNA binding (Li et al., 2006), two JH-induced phosphorylation sites (S55 and T76) in the bHLH domain were mutated (Met-6m) on the basis of Met-4m (Figure 5A). Consistent with the findings for Met-4m, mutation of 6 sites caused the protein to be present primarily in the nucleus, regardless of the presence or absence of JHA (Figures 5E–F’). The Kr-h1 transcriptional level was significantly higher in Met-overexpressing cells treated with JH than in Met-6m-overexpressing cells (Figure 5G). These results suggest that phosphorylation sites in the bHLH domain of Met are necessary for JH intracellular signaling activity.

In addition to posttranslational modifications, partner proteins of the Met/Gce complex, such as Tai, ftz transcription factor 1 (Ftz-f1), and Hsp83, are indispensable for modulating JH action or Met nuclear import (Charles et al., 2011; Bernardo and Dubrovsky, 2012; He et al., 2014). To determine potential Met/Gce binding proteins involved in regulating the subcellular localization of JH intracellular receptors, we expressed Avi-tagged Met, Gce, or both constructs within Kc cells and performed pulldown assays with streptavidin beads and LC-MS/MS analysis (Figure 6A). Datasets from competitive elution and boiling elution were pooled for bioinformatics analysis. After deducting the peptides identified from the BirA control, we obtained 135, 149, and 143 peptides for Met expression, Gce expression and Met/Gce co-expression in Kc cells (Figure 6B; Supplementary Table S4); 69 of these peptides were common to all treatments (Figure 6B; Supplementary Table S5). Notably, two 14-3-3 protein isoforms, 14-3-3ε and 14-3-3ζ, were identified in all these examined samples. 14-3-3 proteins often act redundantly in many cellular processes involved in the regulation of subcellular localization (Benton and St Johnston, 2003; Acevedo et al., 2007). They form homo- and heterodimers and participate in many cellular processes, including signal transduction, by binding to specific phosphorylated sites on their target partners (Mackintosh, 2004; Morrison, 2009). These findings suggest that 14-3-3 proteins might affect the subcellular distribution of Met or Gce.

While 14-3-3 proteins have mostly been reported to specifically interact with phosphoserine-containing motifs such as RSXpSXP (mode I) and RXXXpSXP (mode II) in their ligands (Dougherty and Morrison, 2004), some non-consensus phosphorylation and non-phosphorylation motifs in a set of target proteins of 14-3-3 have been reported (Rajan et al., 2002; Ottmann et al., 2007; Ji and Ostap, 2020). However, upon analyzing the amino acid sequences of Met and Gce, we did not find any consensus or putative 14-3-3 binding phosphorylation motifs, implying that 14-3-3 proteins may interact with Met or Gce in a phosphorylation-independent manner. Moreover, neither Met nor Gce was able to coimmunoprecipitate with 14-3-3ε or 14-3-3ζ after DMSO or JHA treatment (Figures 6C, D). These data suggest that 14-3-3 proteins are unable to directly bind to JH intracellular receptors.

We next asked whether there exists a protein(s) acting as molecular bridge between Met/Gce and 14-3-3 proteins. As an important partner protein of Met, Hsp83 is required for Met nuclear import (He et al., 2014). We therefore analyzed the Hsp83 amino acid sequence and identified a putative 14-3-3 binding motif, RADNSEP, which was located at amino acids 161–167 (Figure 6E). Indeed, Hsp83 coimmunoprecipitated with 14-3-3ε or 14-3-3ζ in the absence or presence of JHA, respectively (Figures 6E, F), and the binding affinity of 14-3-3ζ for Hsp83 was dramatically increased by JHA treatment (Figures 6F, F’). Consistent with previous reports that JH promotes the binding affinity of Hsp83 to Met/Gce (Figures 6G–H’) (He et al., 2014), we further showed that in the presence of JHA, Hsp83 interacted with Met/Gce and 14-3-3 to form a tripolymer complex (Figure 6I).

14-3-3 binding regulates the function of its targets through several mechanisms, such as by affecting protein complex stability and their subcellular localization (Tzivion and Avruch, 2002). We next investigated whether 14-3-3 proteins affect Met nuclear import and consequently JH action. In Kc cells, 14-3-3 proteins prevented Met nuclear import regardless of JHA treatment (Figures 7C–D’’, compared to Figures 7A–B’’). Meanwhile, JHA treatment significantly increased Kr-h1 expression, whereas 14-3-3 transcript levels remained unchanged (Figure 7E), suggesting that JH evokes an antagonistic 14-3-3 response at the protein level. Moreover, overexpression of 14-3-3 proteins decreased Kr-h1 expression or JH response region (JHRR) binding activity in Kc cells exposed to JHA (Figures 7F, G). Taken together, these results demonstrate that 14-3-3 proteins play a conserved role in the negative regulation of JH signal transduction by antagonizing Met nuclear import in an Hsp83-dependent manner.

Normally, 14-3-3 proteins are characterized by their ability to bind target proteins through recognition of phosphorylated consensus motifs. Therefore, we speculated that Hsp83 is phosphorylated under JH regulation. As revealed by phosphoproteomics analysis, Hsp83 S219 was identified as a single phosphorylation site in the presence of JHA (Figure 8A), suggesting that JH induces Hsp83 phosphorylation and that 14-3-3 proteins might interact with Hsp83 in a phosphorylation-independent manner. Interestingly, as determined by aligning the Hsp83 protein of D. melanogaster Hsp83: AAF47734, Homo sapiens Hsp90AA1: NP_001017963.2, and H. sapiens Hsp90AB1: NP_001258898.1, it appears that phosphorylation of S219 appears to be highly conserved across Drosophila and other taxa (Figure 8B). To study whether S219 phosphorylation in Hsp83 is crucial for Met nuclear import, site-directed mutagenesis was performed to replace serine with alanine. S219A mutation significantly reduced the binding affinity of Met for Hsp83 but not for 14-3-3ζ under JHA treatment (Figures 8C–D’). This finding demonstrates that phosphorylation of Hsp83 at S219 promotes the interaction of Hsp83 with Met and further confirms that 14-3-3 proteins bind to Hsp83 in a phosphorylation-independent manner. Consequently, the S219A mutation attenuated Kr-h1 transcription in response to JHA (Figure 8E). These results suggest that JH-induced Hsp83 phosphorylation at S219 facilitates the interaction between Hsp83 and Met to enhance JH intracellular signaling.