D. melanogaster flies were reared on a standard agar-cornmeal medium at 25 or 17°C under a 12:12 h light/dark cycle. w ¹¹¹⁸ served as a control strain. phm-GAL4#22 (a gift from Michael B. O’Connor, University of Minnesota, MN; McBrayer et al., 2007; Yamanaka et al., 2013) was used as the strain to drive forced gene expression in the PG. UAS-dicer2 (#24650) was obtained from the Bloomington Drosophila Stock Center. UAS-pabp-IR (#22007) and UAS-mld-IR (#17329) were obtained from the Vienna Drosophila Resource Center. Transgenic RNAi experiments were conducted by crossing these UAS RNAi lines with w, UAS-dicer2, and phm-GAL4#22/TM6 Ubi-GFP. A strain carrying the 1.45 kb spok promoter-fused GFP cassette (spok>GFP) was previously described (Komura-Kawa et al., 2015).

Rabbit polyclonal anti-Mld antibody was raised against the peptide: NH₂-MSANRRRRSASAASSIAAET-COOH, which corresponds to 1–20 amino acid residues of the mature Mld protein (Neubueser et al., 2005).

Immunostaining of prothoracic glands was performed as described previously (Imura et al., 2017). Briefly, dissected larval tissues were fixed in 3.7% formaldehyde in phosphate-buffered saline (PBS) for 30 min at room temperature. Samples were then washed with PBS + 0.3% Triton X-100 (Nacalai tesque, Kyoto Japan) and incubated overnight at 4°C with primary antibodies: rabbit anti-Mld antibody (1:200), guinea pig anti-Spok antibody (1:200; Gibbens et al., 2011), rabbit anti-Phantom (Phm) antibody (1:200; Parvy et al., 2005), rat anti-Ventral vein lacking (Vvl) antibody (1:3,000; Anderson et al., 1995), and guinea pig anti-POU domain motif 3 (Pdm3; 1:100; Chen et al., 2012). As fluorescent secondary antibodies, we used goat anti-guinea pig Alexa Fluor 488 (Life Technologies, Carlsbad, CA, USA) and goat anti-rabbit Alexa Fluor 555 (Life Technologies, Carlsbad, CA, USA). The secondary antibodies were diluted 1:200 and incubated for 2.5 h at room temperature. Nuclear stains used in this study were 4',6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich, St. Louis, MO, USA) and TOPRO3 (Thermo Fisher Scientific, Waltham, MA, USA). For DAPI staining, after the incubation with the secondary antibodies, the samples were washed and then incubated with 1 μg/ml (final concentration) of DAPI for 1 h. For TOPRO3 staining, the samples were incubated with the secondary antibodies along with 10 μg/ml (final concentration) of RNase A (Takara Bio, Kusatsu, Japan). The samples were then washed, followed by the 1-h incubation with 10 μm TOPRO3 (Thermo Fisher Scientific, Waltham, MA, USA). Confocal images were captured using the LSM 700 laser scanning confocal microscope (Carl Zeiss, Oberkochen, Germany). Images were processed using the ImageJ software (Schneider et al., 2012).

RNA was isolated from the whole bodies of the second instar larvae using the RNAiso Plus reagent (TaKaRa, Shiga, Japan). Genomic DNA digestion and cDNA synthesis were performed using the ReverTra Ace qPCR RT Kit (TOYOBO, Tokyo, Japan). Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) was performed using the THUNDERBIRD SYBR qPCR Mix (TOYOBO, Tokyo, Japan) with a Thermal Cycler Dic TP800 System (TaKaRa, Shiga, Japan). Serial dilutions of a plasmid containing the open reading frame (ORF) of each gene were used as a standard. The expression levels of the target genes were normalized to an endogenous control ribosomal protein 49 (rp49) in the same sample. Primers amplifying noppera-bo (nobo), neverland (nvd), shroud (sro), spok, phm, disembodied (dib), shadow (sad), and rp49 have been described previously (McBrayer et al., 2007; Niwa et al., 2010; Enya et al., 2014).

A spok promoter-driven GFP reporter assay was performed as described previously (Komura-Kawa et al., 2015). Briefly, spok>GFP/CyO Act-GFP; UAS-deicer2, phm22-GAL4/TM6 Ubi-GFP was established and crossed with UAS-pabp-IR. Eggs were laid on grape plates with yeast pastes at 25°C for 2 h. The GFP-negative first instar larvae were picked up and transferred into vials with standard cornmeal food. They were dissected 60 h after egg laying (AEL) and then immunostained.

To examine the importance of Pabp in the PG, we observed the developmental progress of pabp-RNAi larvae, for which we used a PG-specific driver (phm22-GAL4, hereafter phm>) to knock down pabp expression by transgenic RNAi. We found that PG-specific pabp-RNAi caused a larval-arrest phenotype. Eighty eight percent of phm>pabp-RNAi animals were arrested at the second larval instar and even 132 h AEL or later, while only few animals molted into the third larval instar or pupariated at the same time point (Figures 1A–C). The arrested second instar larvae failed to pupariate and died. This result suggests that the Pabp function in the PG is essential for larval development.

Next, we examined whether pabp knockdown in the PG changes the expression of ecdysteroidogenic genes. We conducted qRT-PCR to examine the expression levels of seven ecdysteroidogenic genes (Chávez et al., 2000; Warren et al., 2002, 2004; Niwa et al., 2004, 2010; Namiki et al., 2005; Ono et al., 2006; Yoshiyama et al., 2006; Chanut-delalande et al., 2014; Enya et al., 2014) in the second instar larvae of control and pabp-RNAi animals. The expression levels of all ecdysteroidogenic genes examined were suppressed in pabp-RNAi animals. However, specifically, the suppression levels were different among these seven genes. In particular, the suppression level of spok, encoding an ecdysteroidogenic cytochrome P450 enzyme, substantially decreased (Figure 2A). Consistent with this result, the Spok protein level also substantially decreased in the PG of pabp-RNAi animals as compared to control animals, while the protein level of another ecdysteroidogenic P450 enzyme Phm was only slightly affected (Figures 2B,C).

To determine whether spok expression is transcriptionally and/or translationally disrupted in the pabp-RNAi animals, we conducted a GFP reporter assay in vivo. The spok enhancer region, which is sufficient for spok transcription in the PG, has been identified in our previous study (Komura-Kawa et al., 2015). We found that the spok enhancer-driven GFP level was almost diminished in PG cells of pabp-RNAi animals, suggesting that pabp knockdown disrupts spok transcription (Figures 2D,E).

To analyze the molecular mechanism of suppression of spok transcription, we focused on the transcription factor Mld. Mld is a zinc-finger type DNA binding protein involved in ecdysone biosynthesis in the PG (Neubueser et al., 2005). Moreover, we have previously reported that Mld is crucial for transcription of spok, acting on the upstream region of the spok gene locus (Danielsen et al., 2014; Komura-Kawa et al., 2015). In this study, we newly-generated an anti-Mld antibody to visualize Mld protein in vivo by immunostaining. Before the immunostaining experiment, we examined the specificity of the newly-generated anti-Mld antibody. We observed that the Mld signal was observed in the nucleus of PG cells in control animals. In contrast, the Mld signal in the PG nucleus disappeared in mld-RNAi animals (Figure 3), confirming that the immunostaining signal in the PG nucleus corresponds to Mld protein localization. We then conducted immunostaining with an anti-Mld antibody against pabp-RNAi PG cells. Surprisingly, we found that nuclear localization of Mld was disrupted in the pabp-RNAi PG cells, although Mld is localized in the nucleus of control PG cells (Figures 4A,B). These results suggest that Pabp regulates the transcription of spok by mediating the nuclear localization of Mld.

Next, we examined whether such mislocalization is selective for Mld, but not for other transcription factors in PG cells. To address this issue, we examined the nuclear localization of two other transcription factors such as Vvl and Pdm3, in the pabp-RNAi PG cells. Vvl is a POU-domain transcription factor involved in the regulation of the transcription of all known ecdysteroidogenic genes in PG cells (Danielsen et al., 2016). Pdm3 is also a POU-domain transcription factor that is enriched in PG cells (Ou et al., 2016), whereas, its role in PG cells has not yet been elucidated. An immunohistological analysis using anti-Vvl and anti-Pdm3 antibodies revealed that nuclear localization of Vvl and Pdm3 was maintained in the nucleus, even in pabp-RNAi PG cells. Nevertheless, the Vvl signal did slightly decrease (Figures 4C–F). Taken together, these results suggest that Pabp regulates the nuclear localization specifically of Mld.