20-Hydroxyecdysone (20E) Primary Response Gene E75 Isoforms Mediate Steroidogenesis Autoregulation and Regulate Developmental Timing in Bombyx*

The temporal control mechanisms that precisely control animal development remain largely elusive. The timing of major developmental transitions in insects, including molting and metamorphosis, is coordinated by the steroid hormone 20-hydroxyecdysone (20E). 20E involves feedback loops to maintain pulses of ecdysteroid biosynthesis leading to its upsurge, whereas the underpinning molecular mechanisms are not well understood. Using the silkworm Bombyx mori as a model, we demonstrated that E75, the 20E primary response gene, mediates a regulatory loop between ecdysteroid biosynthesis and 20E signaling. E75 isoforms A and C directly bind to retinoic acid receptor-related response elements in Halloween gene promoter regions to induce gene expression thus promoting ecdysteroid biosynthesis and developmental transition, whereas isoform B antagonizes the transcriptional activity of isoform A/C through physical interaction. As the expression of E75 isoforms is differentially induced by 20E, the E75-mediated regulatory loop represents a fine autoregulation of steroidogenesis, which contributes to the precise control of developmental timing.

The temporal control mechanisms that precisely control animal development remain largely elusive. The timing of major developmental transitions in insects, including molting and metamorphosis, is coordinated by the steroid hormone 20-hydroxyecdysone (20E). 20E involves feedback loops to maintain pulses of ecdysteroid biosynthesis leading to its upsurge, whereas the underpinning molecular mechanisms are not well understood. Using the silkworm Bombyx mori as a model, we demonstrated that E75, the 20E primary response gene, mediates a regulatory loop between ecdysteroid biosynthesis and 20E signaling. E75 isoforms A and C directly bind to retinoic acid receptor-related response elements in Halloween gene promoter regions to induce gene expression thus promoting ecdysteroid biosynthesis and developmental transition, whereas isoform B antagonizes the transcriptional activity of isoform A/C through physical interaction. As the expression of E75 isoforms is differentially induced by 20E, the E75-mediated regulatory loop represents a fine autoregulation of steroidogenesis, which contributes to the precise control of developmental timing.
Animals undergo developmental transitions from the embryo to juvenile to adulthood, and these processes are determined by steroid hormones and their corresponding nuclear receptors (NRs). 2 In insects, 20-hydroxyecdysone (20E; ecdy-sone is the immediate precursor of 20E; 20E and ecdysone are the main ecdysteroids) is the actual steroid hormone. The ecdysone receptor (EcR) and its partner molecule, Ultraspiracle (USP), form the functional NR complex of 20E. In conjunction with EcR-USP, 20E activates a small set of early response genes encoding several transcription factors that further activate a large set of downstream late response genes. Pulses of 20E signals initiate major developmental transitions in insects, including egg hatching, larval-larval molting, and larval-pupal-adult metamorphosis (1,2).
In Drosophila, the E75 locus encodes four E75 mRNA isoforms, E75A, E75B, E75C, and E75D, which are generated by differential promoter usage and alternative splicing of 5Ј exons. The DBD of E75A/C possesses two C4 zinc fingers; E75B is incomplete and contains only one zinc finger, whereas E75D lacks a DBD. 20E-EcR-USP rapidly and abundantly induces the expression of E75A and E75B by binding to the 20E response elements present in the promoter regions. In contrast, the 20E induction of E75C expression is slow and weak (9,16). Germ line clones of E75-null mutants missing all three isoforms lead to arrest during mid-oogenesis (17). Isoform-specific E75 null mutants exhibit different phenotypes; E75A mutants show a reduced ecdysteroid titer leading to developmental retardation and molting defects; E75B mutants can survive and exhibit normal reproductive performance; and E75C mutants die within a few days after eclosion (9). E75 might regulate 20E signals through interaction with another 20E response gene HR3, which encodes another important insect NR. HR3 controls the termination of the 20E signal pulse, which triggers the larvalprepupal transition by both inhibiting 20E-EcR-USP transactivation by interacting with EcR and blocking ecdysone biosynthesis by down-regulating the Halloween family of cytochrome P450 genes (Halloween genes). HR3 also induces the expression of ␤ftz-F1, which acts as a competent factor for EcR-USP to respond to the subsequent 20E signal pulse during the prepupal-pupal transition. Importantly, E75 acts as a transcriptional repressor for HR3 in relieving HR3 inhibition on 20E signaling and HR3 induction on ␤ftz-F1 expression. E75 inhibits the transactivation ability of HR3 through physical interaction and competing for binding to the retinoic acid receptor-related receptor response elements (ROREs). Therefore, the 20E-induced transcriptional cascade, including EcR-USP, E75, HR3, and ␤ftz-F1, governs the larval-prepupal-pupal transition. In addition, because NO and CO are able to reverse the ability of E75 to interfere with HR3, the function of E75 is modulated by gas availability (10 -12, 18 -21).
Early studies found that E75B interferes with HR3 induction of ␤ftz-F1 expression (18), and later studies revealed that at least E75A has the same function (11), indicating that E75 isoforms play similar roles in HR3 regulation. However, in female adults, E75A induces apoptosis in the egg chamber at stages 8 and 9, whereas E75B prevents E75A function and thus allows egg development, indicative of opposite roles in regulating female reproduction (22). Similarly, E75 isoforms also play distinct roles in regulating female reproduction in the mosquito, Aedes aegypti (23). Given that both E75A and E75B have similar effects on HR3, HR3 clearly cannot account for the opposite functions of the E75 isoforms, suggesting that E75 isoforms may employ novel mechanisms to differentially regulate insect development.
Bombyx E75 processes at least three isoforms, E75A, E75B, and E75C, showing similar gene organization and 20E response to Drosophila E75 (24,25). Likewise, Bombyx E75A/C interacts with HR3 and represses its transactivation activity by physical interaction and competing for ROREs (26). We reasoned that Bombyx could be a good model to solve the E75 isoform-specific mechanism, because this insect species has a comparatively longer life cycle for phenotypic observations and can be genetically modified for functional analyses (27). A molecular dissection of E75 isoforms in Bombyx found that, in addition to acting as transcriptional repressors of HR3, E75 isoforms also regulate ecdysteroid biosynthesis by directly controlling Halloween gene expression. Mechanistically, E75A/C functions as a transcriptional factor to directly induce Halloween gene expression, whereas E75B antagonizes the transactivation ability of E75A/C. Given that the expression of E75 isoforms is differentially induced by 20E, our study revealed an E75-medi-ated regulatory loop that contributes to steroidogenesis autoregulation and thus developmental timing. Regarding the ultimate regulation of ecdysteroid biosynthesis, E75 first functions directly and then acts through inhibition of HR3.
To determine the function of E75 during larval-pupal metamorphosis, expression of all three E75 isoforms was reduced by RNAi (E75 RNAi) at the initiation of the wandering stage (IW). E75 RNAi caused lethal phenotypes, with ϳ60 and 10% lethality during the prepupal and pupal stages, respectively. Some E75 RNAi larvae died during the wandering stage, and others failed to form normal pupae and died as larval-pupal intermediates, whereas others were arrested during the pupal stage or immediately after adult emergence (Fig. 1, A-Aٞ).
Importantly, E75 RNAi inhibited fat body remodeling, which is controlled mainly by the 20E-triggered transcriptional cascade during larval-pupal metamorphosis (28 -31). Twenty four hours after injection with E75 double-stranded RNA (dsRNA) (supplemental Fig. S1, A and BЈ), LysoTracker Red staining, the number and size of autophagosomes and the ATG8 protein levels decreased significantly, suggesting that the 20E-induced fat body autophagy is affected by E75 RNAi (Fig. 1, B-BЉ). Meanwhile, labeling with Hoechst 33342 and propidium iodide, TUNEL staining, and measurement of caspase 3 activity revealed significant reductions in 20E-induced fat body apoptosis by E75 RNAi (Fig. 1, C-CЉ). In addition, the 20E-induced fat body cell dissociation that occurred 24 h after pupation in the EGFP RNAi control pupae was significantly prevented in the E75 RNAi pupae (Fig. 1D).
The effects of E75 RNAi on fat body remodeling suggest that E75 is required for maintaining 20E signaling to promote larvalpupal metamorphosis. The expression levels of several key genes in the 20E-triggered transcriptional cascade were determined by quantitative real time PCR (qPCR) using the total RNA isolated from the fat body collected 24 h after E75 dsRNA injection. The mRNA levels of all the 20E-response genes decreased by 60 -90% compared with their levels in the control larvae (Fig. 1E). Moreover, Western blottings using EcR-B1, USP, and Br-C antibodies revealed a decrease in their protein levels in the E75 RNAi larvae (Fig. 1EЈ), indicating that E75 RNAi disrupts the 20E-triggered transcriptional cascade in the fat body during larval-pupal metamorphosis. Overall, E75 RNAi disrupted 20E signaling, prevented fat body remodeling, and caused lethality during metamorphosis.
The prothoracic glands produce and secrete ecdysone; once released into the hemolymph, ecdysone is converted to 20E in the peripheral tissues, such as the fat body and midgut (34). Ecdysone and 20E in the mixture of hemolymph ecdysteroids were separated and individually collected using reverse-phase high performance lipid chromatography (rpHPLC) and then measured by EIA. As expected, both titers of ecdysone and 20E decreased in E75 RNAi larvae; moreover, the ratio between 20E and ecdysone was further decreased in E75 RNAi larvae (Fig. 2, B-BЉ).
In Bombyx, the Halloween genes, spook (spo), phantom (phm), disembodied (dib), and shadow (sad), mediate the sequential steps of ecdysone biosynthesis in the prothoracic glands, whereas shade (shd) catalyzes the conversion from ecdysone to 20E in the fat body and other peripheral tissues (35,36). The mRNA levels of spo, phm, dib, and sad decreased by more than 90% in the prothoracic glands isolated from E75 RNAi larvae (Fig. 2C). The prothoracic glands were dissected out from the E75 RNAi larvae and cultured in vitro, and the ecdysone released into the medium was measured by EIA. Importantly, the ratio of ecdysone release by the cultured prothoracic glands decreased by about half in E75 RNAi larvae (Fig.  2CЈ). Overall, the prothoracic glands of E75 RNAi larvae exhibited normal morphology without apparent autophagy, apoptosis, and cell dissociation (supplemental Fig. S2, A and C). Nevertheless, a large number of mitochondria, which are essential for hormone production in endocrine organ cells, were misshaped in the prothoracic gland cells from the E75 RNAi larvae, supporting the reduced ecdysone production (Fig. 2CЉ) (Fig. 2D). The fat body tissues were dissected and cultured in vitro with the addition of ecdysone in the medium, and ecdysone and the newly converted 20E in the medium were separated by rpHPLC and measured by EIA. The conversion from ecdysone to 20E also decreased by 80% in the fat body dissected from the E75 RNAi larvae (Fig.  2DЈ). Taken together, these data demonstrated that E75 RNAi down-regulates Halloween genes that are responsible for ecdysone biosynthesis in the prothoracic glands and the conversion from ecdysone to 20E in the fat body, resulting in the disruption of ecdysteroid biosynthesis and 20E-induced metamorphosis.
Overexpression of E75A/C Up-regulates Halloween Genes, Promotes 20E Signaling, and Accelerates Metamorphosis-Initial experiments using RNAi to reduce the expression of each E75 isoform showed variable results, mostly because their AF-1 domains are too short to generate reliable isoform-specific dsRNAs. We generated an ecdysteroid UDP-glucosyltransferase (egt) mutant of B. mori nucleopolyhedrosis baculovirus (BmNPV) to overexpress each E75 isoform on day 2 of the fifth instar (L5D2). Five and a half days after BmNPV infection, only 30% of the EGFP-overexpressed larvae began wandering, whereas 80 and 70% of the E75A-and E75C-overexpressed larvae entered the wandering stage with reduced body sizes, respectively, although the wandering behavior and the body size of the E75B-overexpressed larvae were slightly prevented compared with the control larvae (Fig. 3, A and AЈ and supplemental Fig. S1, C and Dٞ). Moreover, 72 h after BmNPV infection, both titers of ecdysone and 20E increased in the E75A/C-overexpressed larvae, but they slightly decreased in the E75B-overexpressed larvae (Fig. 3, B-BЉ).
Seventy two hours after BmNPV infection, we further examined the effects of each E75 isoform on the prothoracic glands and fat body (supplemental Fig. S1, C and Dٞ). The mRNA levels of spo, phm, dib, and sad increased by 4 -7-and 3-5-fold in the prothoracic glands of the E75A-and E75C-overexpressed larvae, respectively; however, they decreased by 20 -40% in the E75B-overexpressed larvae (Fig. 3, C-CЈ). The amount of ecdysone released by the cultured prothoracic glands increased by 200 and 150% in the E75A-and E75C-overexpressed larvae, respectively, but they slightly decreased in the E75B-overexpressed larvae (Fig. 3D). Meanwhile, the mRNA levels of shd increased by 3-fold in the fat body from the E75A/C-overexpressed larvae, but they slightly decreased in the E75B-overexpressed larvae (Fig. 3E). Similarly, the conversion from ecdysone to 20E increased in the fat body from the E75A/Coverexpressed larvae, but they slightly decreased in the E75Boverexpressed larvae (Fig. 3F). In conclusion, overexpression of E75A/C up-regulates Halloween genes, promotes ecdysteroid biosynthesis, and accelerates metamorphosis, whereas E75B overexpression might have opposing effects. E75A/C, but Not E75B, Binds to ROREs and Directly Induces Halloween Gene Expression-Because E75 binds to ROREs to antagonize the transactivation ability of HR3, we hypothesized that E75 might also bind to ROREs and thus directly induce Halloween gene expression. Using a dual-luciferase assay system established in heterologous human HEK 293 cells, we investigated whether the three E75 isoforms can directly bind the promoter of the five Halloween genes, including spo, phm, dib, sad, and shd. The ϳ2.5-kb promoter region of each Halloween gene was cloned into the pGL3 vector. Upon E75A/C overexpression, all five ϳ2.5-kb promoter regions supported a 2.5-4-fold increase in luciferase activity, whereas E75B overexpression had no effect (Fig. 4, A-E). In BmN cells, the luciferase activities of all five ϳ2.5-kb promoter regions increased 1.5-3fold upon E75A/C overexpression. Interestingly, E75B overexpression slightly reduced the luciferase activities (Fig. 4, F-J), resembling the effects of E75B overexpression in vivo (Fig. 3, C-CЉ and E).
There are 2, 1, 3, 2, and 4 potential ROREs in the ϳ2.5-kb promoter regions of spo, phm, dib, sad, and shd, respectively (supplemental Fig. S3). We then performed chromatin immunoprecipitation (ChIP) in Bm-N cells to examine how the three E75 isoforms bind to ROREs. The binding of E75 isoforms to DNA was detected using the V5 antibody and cross-linked chromatin isolated from Bm-N cells that were transfected with the E75A/B/C-V5 expression plasmids. As measured by qPCR, the V5 antibody increased the precipitation of 13 ROREs (except one in shd) when E75A was overexpressed, 12 ROREs (except one in sad and the other in shd) when E75C was overexpressed, but no ROREs when E75B was overexpressed (Fig. 4, All the responsive ROREs in the ϳ2.5-kb promoter regions of each Halloween gene were deleted, and the mutated ϳ2.5-kb promoter regions of all the five Halloween genes were cloned into the pGL3 vector. E75A/C overexpression did not increase luciferase activity for any of the mutated constructs ( Fig. 4, P-T). Together, the dual-luciferase assays and ChIP-qPCR data revealed that E75A/C, but not E75B, binds to ROREs in the promoter regions of all five Halloween genes and directly induces gene expression.    AUGUST 26, 2016 • VOLUME 291 • NUMBER 35

JOURNAL OF BIOLOGICAL CHEMISTRY 18167
E75B Antagonizes the Transactivation Ability of E75A/C-The above overexpression results raise the possibility that E75B antagonizes the transactivation ability of E75A/C. To test this hypothesis, E75A or E75C and E75B were co-transfected into HEK 293 cells. The effect of the expressed proteins on spo and shd promoter activities was determined. As shown above (Fig.  4, A-E), E75A/C overexpression, but not E75B overexpression, showed significant increases in the luciferase activity. Importantly, co-transfection of E75B antagonized the transactivation ability of E75A/C in a dose-dependent manner, whereas E75A and E75C did not affect each other (Fig. 5, A-BЈ). Similar results were obtained in BmN cells (Fig. 5, C-DЈ).
To further verify the hypothesis in vivo, equal amounts of two BmNPVs of EGFP, E75A, E75B, or E75C were co-infected to L5D2 larvae. Five days after BmNPV infection, E75A/C-, but not E75B-, overexpressed larvae showed precocious wandering behavior and reduced body size compared with the EGFP-overexpressed control larvae (Fig. 5E). Importantly, co-infection with E75B nearly blocked the ability of E75A/C to reduce body size (Fig. 5EЈ). Seventy two h after BmNPV infection, ecdysteroid titers significantly increased in the E75A/C-overexpressed larvae, and this increase was blocked by co-infection with E75B (Fig. 5F), suggesting that E75B antagonizes the transactivation ability of E75A/C.    CRISPR/Cas9-mediated genome editing is becoming a powerful tool for functional studies in Bombyx (37, 38). Because RNAi was not able to sufficiently and specifically reduce E75B expression, we performed CRISPR/Cas9-mediated knock-out of E75B. Interestingly, all of the E75B-knock-out larvae successfully survived to adults but showed accelerated wandering   Ecdysteriod titer (ng/mL) behavior and elevated ecdysteroid titers (Fig. 5, G and H, and  supplemental Fig. S4). Overall, the E75B-knock-out larvae underwent phenotypic changes similar to those of the E75A/Coverexpressed larvae. Both in vitro and in vivo experimental data revealed that E75B antagonizes the transactivation ability of E75A/C for regulating Halloween gene expression, ecdysteroid biosynthesis, and metamorphosis. Incomplete DBD in E75B Mediates Physical Interactions and Thus the Opposing Actions between E75B and E75A/C-Finally, we investigated whether E75B antagonizes the transactivation ability of E75A/C through protein-protein interactions. Two constructs of E75A, E75B, or E75C, the C termini of which were fused to different tags, were co-transfected in HEK 293 cells. Immunocytochemistry was performed to examine their possible protein-protein interactions. When E75A and/or E75C were co-transfected, they evenly localized in the nuclei (Fig.  6A). By contrast, when E75B was co-transfected with E75A, E75B, or E75C, the two proteins frequently co-localized at some aggregating chromatin spots (Fig. 6B). Similar results were obtained in BmN cells (Fig. 6, C and D). Both data in HEK 293 and BmN cell lines suggested that E75B might associate with all three E75 isoforms.

E 7 5 B O E
E75A, E75B, and E75C contain different AF-1 domains and DBDs (25). To identify the actual E75B domain(s) that are responsible for its association with all three E75 isoforms, we generated three E75 mutant constructs: coE75A/C that shares the complete DBD of E75A/C and the common C terminus, coE75A/B/C that shares the incomplete DBD of E75A/B/C and the common C terminus, and E75noN that only retains the common C terminus (Fig. 6E). When coE75A/C was co-transfected with E75A, E75B, or E75C, only coE75A/C and E75B co-localized at the aggregating chromatin spots in HEK 293 cells (Fig. 6F). When coE75A/B/C was co-transfected with E75A, E75B, or E75C, the two proteins always co-localized at the aggregating chromatin spots (Fig. 6G). Nevertheless, E75noN had no co-localization with E75A/C but co-localized with E75B (Fig. 6H). The immunocytochemistry experiments demonstrated that the incomplete DBD in E75B is indispensable for the association between E75B and E75A/C. Furthermore, in HEK 293 cells, co-transfection with coE75A/B/C antagonized the transactivation ability of E75A/C in a dose-dependent manner (Fig. 6, I and IЈ), indicating that the incomplete DBD in E75B mediates physical interactions and thus the opposing actions between E75B and E75A/C.

E75A/C Is a Bona Fide Transcription Factor That Induces
Halloween Gene Expression-The majority of research on Drosophila focused on showing that E75 is a transcriptional repressor of HR3 through physical interaction and competing for ROREs. Nevertheless, HR3 inhibition is not able to explain the isoform-specific phenotypes of E75 mutants. Here, we demonstrate for the first time that, in addition to HR3 inhibition, E75A/C is a bona fide transcription factor that directly drives Halloween gene expression and thus induces ecdysteroid biosynthesis. First, E75 RNAi resulted in a decrease in expression of all five Halloween genes responsible for ecdysteroid biosynthesis, low ecdysteroid titers, impaired 20E signaling, repressed fat body remodeling, and lethality during metamorphosis (Figs.  1 and 2A). These E75 RNAi silkworms exhibit phenotypic defects similar to Drosophila E75A and E75C null mutants (9). Second, E75A/C overexpression up-regulates Halloween genes, promotes ecdysteroid biosynthesis, and accelerates metamorphosis (Fig. 3). Consistently, these phenotypes are similar to those observed after overexpression of E75A specifically in the Drosophila prothoracic glands (21). Third, dual-luciferase assays and ChIP-qPCR experiments together showed that E75A/C binds to ROREs in the Halloween gene promoter regions and thus induces expression of these genes ( Fig. 4 and supplemental Fig. S3), providing strong evidence that E75A/C is a bona fide transcription activator. By binding to ROREs, E75A/C might act as a transcriptional repressor for competing with HR3, but more importantly, E75A/C functions as a transcriptional activator that induces Halloween gene expression. Fourth, E75B processes an incomplete DBD, which is not able to bind to ROREs to induce Halloween gene expression but still can inhibit HR3, confirming that Halloween gene induction and HR3 inhibition are indeed two separated functions of E75 (Figs. 3 and 4). Finally, in both the prothoracic glands and fat body of Bombyx, the E75 mRNA levels peak nearly 1 day earlier than that of HR3 (supplemental Fig. S5, A, B, D, and E) (25), implying that E75A/C functions as a transcriptional activator in inducing Halloween gene expression during the wandering stage and then as transcriptional repressor of HR3 during the larval-pupal transition. Regarding the ultimate regulation of ecdysteroid biosynthesis, E75 should first function directly and then act through inhibition of HR3 (supplemental Fig. S5H). Conclusively, E75A/C is a bona fide transcription activator that drives Halloween gene expression and thus induces ecdysteroid biosynthesis (Fig. 7).
E75B Antagonizes E75A/C to Regulate Halloween Gene Expression-The second important discovery of this study is on E75B antagonism of the transactivation ability of E75A/C to regulate Halloween gene expression. First, in contrast to E75A/C overexpression, E75B overexpression down-regulates Halloween genes, reduces ecdysteroid biosynthesis, and delays metamorphosis (Fig. 3). Second, co-transfection of E75B antagonizes the transactivation ability of E75A/C both in vitro (Fig. 5, A-DЈ) and in vivo (Fig. 5, E and F). Third, knock-out of E75B exhibits phenotypes (Fig. 5, G and H) similar to those seen after E75A/C overexpression (Fig. 3, A and AЈ). Finally and most importantly, the incomplete DBD of E75B mediates physical interactions and thus opposing actions between E75B and E75A/C (Fig. 6). Different from that, all E75 isoforms utilize their common C terminus to interact with and to antagonize transcriptional activity of HR3 (19,26).
It is necessary to note that the E75B mutants in both Drosophila (9) and Bombyx (Fig. 5, G and H) are viable and fertile. Taking advantage of the comparatively long life cycle and the newly developed CRISPR/Cas9-mediated genome editing method in Bombyx, we have observed accelerated wandering behavior and elevated ecdysteroid titers in the Bombyx E75B mutants. Because E75B antagonizes transcriptional activity of both HR3 and E75A/C, the outcome of E75B in tuning ecdysteroid biosynthesis and developmental timing should be context-specific. We assume that this role of E75B should be conserved in Bombyx and Drosophila, because the protein structure and 20E induction of expression are the same in both animals. Previous studies also showed that E75A and E75B have opposing effects on the apoptosis/development choice of the egg chamber in Drosophila (22). Similarly, E75 isoforms play distinct roles in regulating female reproduction in the mosquito A. aegypti (23). These studies indicate that E75 isoforms have an isoform-specific function in regulating insect reproduction, in line with our findings that E75A/C and E75B oppositely regulate Halloween gene expression, ecdysteroid biosynthesis, and developmental timing. In conclusion, lacking a complete DBD, E75B does not act as an independent transcription activator, but antagonizes the transactivation ability of E75A/C by binding to and changing the conformation of E75A/C (Fig. 7).
Correlations among E75, HR3, and NO-Phylogenetic analysis reveals that E75 and HR3 belong to NR subfamily 1 and are closely related (3). Multiple lines of evidence support that, by physical interaction and by competing for ROREs, E75 isoforms indiscriminately act as transcriptional repressors for HR3. By being either transcriptional repressors for HR3 in relieving HR3 inhibition on Halloween gene expression (21) or transcriptional activators in inducing Halloween gene expression (Figs. 1-4), E75A/C ultimately promotes ecdysteroid biosynthesis and developmental transition (supplemental Fig. S5H). Both gainof-function and loss-of-function results clearly show that E75B inhibits Halloween gene expression and thus ecdys-teroid biosynthesis in vivo (Figs. 3 and 5), indicating that its inhibition of ecdysteroid biosynthesis via antagonizing the transactivation ability of E75A/C (Figs. 3-6) is more crucial than its possible promotion via relieving HR3 inhibition (supplemental Fig. S5 H).
NO and CO are able to reverse the ability of E75 to interfere with HR3; thus, the function of E75 in counteracting HR3 might vary depending on the availability of these gases. We investigated the developmental profiles of NO synthetase (NOS1 and NOS2) in the prothoracic glands and the fat body in Bombyx. Interestingly, the expression peaks of NOS1 and NOS2 (supplemental Fig. S5, C, CЈ, F, and FЈ) never match that of E75 (supplemental Fig. S5, A, B, D, and E). Moreover, the transcriptional activity of E75A/C in inducing Halloween gene expression was able to be reversed by NO (supplemental Fig. S5G). Thus, the ability of E75A/C to promote ecdysteroid biosynthesis and developmental transition could be reversed by NO (supplemental Fig. S5 H). We suppose that the binding of E75B with E75A/C, the binding of all E75 isoforms with HR3, and the binding of NO with all E75 isoforms will result in changes of conformation and transactivation ability of the latter ones.
E75-mediated Steroidogenesis Autoregulation Contributes to the Precise Control of Developmental Timing-Steroidogenesis autoregulation in insects involves a fine regulatory loop between ecdysteroid biosynthesis and 20E signaling. A number of genes in the 20E-triggered transcriptional cascade regulate ecdysone biosynthesis and thus ecdysteroid titers in both Drosophila and Bombyx (9,21,(31)(32)(33). ␤ftz-F1, Br-C, HR3, and E75 regulate Halloween gene expression in the prothoracic glands (21,32,33). Here, we found that E75 binds to ROREs and induces Halloween gene expression ( Fig. 4 and supplemental Fig. S3). Importantly, E75A/C induces the Halloween gene expression that is responsible for not only ecdysone biosynthesis in the prothoracic glands but also the conversion from ecdysone to 20E in the fat body, whereas E75B has opposing roles. The composite data support the central role of E75 in the regulatory loop of ecdysteroid biosynthesis (Fig. 7).
In summary, 20E induces the expression of E75 isoforms differently, and E75A/C and E75B oppositely regulate ecdysteroid biosynthesis, forming a fine regulatory loop between ecdysteroid biosynthesis and 20E signaling (Fig. 7). Acting independently or through HR3 inhibition in a context-specific manner, E75 isoforms are involved in the fine regulation of ecdysteroid biosynthesis, which contributes to the precise control of developmental timing (supplemental Fig. S5H). This study provides a paradigm for how NR isoforms accurately mediate steroidogenesis autoregulation and thus developmental timing in animals.

Experimental Procedures
Silkworms and Cells-Bombyx larvae (p50 strain) were provided by the Sericultural Research Institute, Chinese Academy of Agricultural Sciences (Zhenjiang, China), and fed fresh mulberry leaves at 25°C under 14-h light/10-h dark cycles (25,29). Bm-N cells were maintained in TC-100 medium (PAN-BIOTECH, Germany) supplemented with 10% heat-inactivated fetal bovine serum (Gibco). HEK 293 cells were maintained in Dulbecco's modified Eagle's medium (HyClone) supplemented with 10% fetal bovine serum (25,31). E75 RNAi in Bombyx Larvae-The E75 dsRNA (25) was synthesized using a T7 RiboMAX TM Express RNAi kit (Promega, P1700). The EGFP dsRNA was used as a control. Thirty g of dsRNA per larva was injected at IW. The prothoracic glands, peripheral fat body tissues from the 5th abdominal segment, and hemolymph samples were collected at the indicated times for further analysis (25).
Baculovirus-mediated Overexpression of E75 Isoforms in Bombyx Larvae-Using the homologous recombination technique (39), we generated the BmNPV egt mutant that allows silkworms, which survive until pupation, to produce sufficient E75 protein. The BmNPVs expressing E75A, E75B, and E75C were obtained in the same manner as the Autographa californica nucleopolyhedrovirus (25). Five l of P2 BmNPV (ϳ10 5 pfu) was injected into each Bombyx larva on L5D2, and then the prothoracic glands, fat body, and hemolymph were collected at the indicated times for further analysis.
CRISPR/Cas9-mediated Knock-out of E75B in Bombyx-Our colleagues previously developed efficient approaches for CRISPR/Cas9-mediated genome editing in Bombyx (37,38) and helped us to perform the E75B knock-out experiment in this study. Cas9 mRNA (mMESSAGE mMACHINE kit, Ambion, Austin, TX) and E75B sgRNA (MAXIscript T7 Kit, Ambion, Austin, TX), TAATACGACTCACTATAGGTGCT-AGTGAGCATGCTGGAGGGTTTTAGAGCTAGAAATAG-CAAGTTAAAATAAGGCTAGTCCGTTATCAACTTGAA-AAAGTGGCACCGAGTCGGTGCTTTT, was synthesized and purified separately. A mixture of Cas9 mRNA (300 ng/l) and E75B sgRNA (300 ng/l; with EGFP sgRNA as a control) was injected into the non-diapause preblastoderm p50 embryos prepared within 6 h after oviposition using a micro-injector (Narishige, Tokyo, Japan), and then the embryos were incubated at 25°C in a humidified chamber for 10 -12 days until larval hatching. Approximately 24 h after IW, genomic DNA was extracted for mutagenesis analysis. The prothoracic glands, fat body, and hemolymph were collected for further analysis.
Conventional Molecular, Biochemical, and Cellular Methods-Details of caspase-3 activity measurement, qPCR, and Western blotting have been previously described (25,31,40). Production of the EcR, Met1, Br-C, and E75 antibodies have been reported in our publications (25,30). The AB11 USP antibody was a kind gift from Dr. Fotis Kafatos. The Western blotting images were obtained with a Tanon-5500 Chemiluminescent Imaging System (Tanon, China).
Fluorescence Microscopy and Transmission Electron Microscopy-The prothoracic glands and fat body were dissected and processed for fluorescence microscopy and transmission electron microscopy analyses as described previously (25,29,31,40). TUNEL (Beyotime, China) labeling and LysoTracker Red (Invitrogen) staining were used to estimate caspase activity and autophagy, respectively. Cell death was also detected by propidium iodide staining (red nuclei) and nuclei with Hoechst 33342 (blue) (Beyotime). Pictures were taken under an FV10-ASW confocal microscope (Olympus, Japan) at ϫ40 magnification, and each type of observation was performed under the same conditions. A H7650 transmission electron micro-scope (Hitachi, Japan) was used to observe autophagic components, mitochondria, and other cell structures.
Tissue Culture and rpHPLC-EIA Measurements of Ecdysteroids-For measurements of ecdysteroid titers in the hemolymph, we used EIA (Cayman Chemical) (38). In some cases, ecdysone and 20E in the hemolymph were separated using a modified rpHPLC procedure (41) followed by quantification using EIA. In brief, total ecdysteroids in the hemolymph samples were extracted with methanol, dried, and re-dissolved in 20% acetonitrile (ACN) containing 0.1% trifluoroacetic acid (TFA). An Agilent 1100 Series HPLC system (Agilent Technologies) equipped with a variable UV wavelength detector (set at 240 nm) was employed. All samples were separated by an Eclipse Plus C18 (4.6 ϫ 250 mm) column (Agilent Technologies) using a variable mobile phase consisting of 20% ACN containing 0.1% TFA for 5 min and a linear gradient of 20 -80% ACN containing 0.1% TFA for 20 min. The flow rate was 1 ml/min. 20E and ecdysone standards (Sigma) were eluted after 7 and 14 min, respectively. All sample fractions were collected at 6.7-8.7 and 13-15 min for 20E and ecdysone, respectively, dried, re-dissolved in EIA buffer, and measured by EIA. The ratio of ecdysone and 20E was calculated.
For the measurement of ecdysone release, the prothoracic glands were dissected out and cultured in Grace's medium (Sigma) at 25°C. After pre-incubation for 1 h, the medium was replaced with fresh medium. Four hours after incubation, the medium was collected, dried, and re-dissolved in EIA buffer, and the ecdysteroid concentration was determined by EIA. To measure the conversion from ecdysone to 20E, the fat body was cultured in Grace's medium at 25°C. After pre-incubation for 1 h, the medium was replaced with fresh medium containing 5 M ecdysone. Four hours after incubation, the medium was collected and concentrated. Ecdysone and 20E in the medium were separated by rpHPLC, and the fractions were dried, re-dissolved in EIA buffer, and measured by EIA. The conversion from ecdysone to 20E was calculated.
ChIP Assay in Bm-N Cells-The modified pIEx-4 vector containing the BmNPV ie1 promoter (42) was used to overexpress E75A/B/C-V5 in Bm-N cells. Bm-N cells were grown in 10-cm dishes (70% confluent) and transfected with the E75A/B/C-V5 expression plasmid for 48 h using the Effectene transfection reagent (Qiagen, Germany). Then, the cells were fixed and subjected to ChIP assay (31,42,43) using the agarose ChIP kit (Pierce) and the V5 antibody (Sigma). Mock immunoprecipitations with pre-immune serum were used for negative controls. The precipitated DNA and input were analyzed by qPCR to detect the binding between E75A/B/C-V5 and ROREs in promoter regions and CDS regions (as negative control) of the five Halloween genes.
Dual-Luciferase Assay in HEK 293 Cells and BmN Cells-To examine whether the promoter regions of the five Halloween genes are responsive to E75, the 2.5-kb regions of each Halloween gene promoter upstream of the transcription start site (or the RORE-deleted mutant constructs) were cloned into the pGL3 basic vector containing the hsp70 minimal promoter (Promega). The pRL vector (Promega) carrying Renilla luciferase driven by the Actin3 promoter was used for normalization. E75A/B/C (or coE75A/B/C) was cloned into the pcDNA 3.1(ϩ) vector (Invitrogen) to create the expression constructs. After co-transfection of E75A/B/C expression construct, a reporter pGL3 vector, and the reference pRL vector into HEK 293 cells for 48 h using the Effectene transfection reagent (Qiagen), the cells were collected. The relative luciferase activity was calculated by normalizing the reporter firefly luciferase level to the reference Renilla luciferase level. Dual-luciferase assays were conducted using the dual-luciferase assay system (Promega) and a Modulus luminometer (Turner BioSystems) (29,31,42,43). For some experiments, two constructs of EGFP, E75A, E75B, E75C, or coE75A/B/C were co-transfected into HEK 293 cells equally or in a dose-dependent manner. When necessary, the NO donor, 2,2Ј-(hydroxynitrosohydrazino)bis-ethanimine (Sigma; 200 M) was added to the medium (10). Dual-luciferase assays in BmN cells were performed the same as in HEK 293 cells except the expression vector was pIEx-4 containing the BmNPV ie1 promoter as above described.
Cytohistochemistry in HEK 293 Cells and BmN Cells-Microscope coverslips (Fisher, 12-542A) were sterilized before use and placed into 6-well plates during HEK 293 cell plating. After 1 day of pre-incubation, the cells were co-transfected with two pcDNA 3.1(ϩ) vectors expressing E75A, E75B (or its mutants coE75A/C, coE75A/B/C, and E75noN), or E75C, the C termini of which were fused to different tags (V5, HA, FLAG, and Myc), for 48 h. After extensive washing, the coverslips were fixed in 4% paraformaldehyde for 45 min at room temperature, blocked in phosphate-buffered saline containing 5% BSA and 1% Triton X-100 (PBSBT) for 1 h, and incubated with two different primary tag antibodies (V5, HA, FLAG, and Myc, Sigma) (diluted 1:200) at 4°C overnight. The coverslips were washed for 1 h in PBSBT and incubated with two counterpart FITC green/redconjugated secondary antibodies from mouse/rabbit (diluted 1:200) for 2 h at room temperature (43). Images were captured using the Olympus FV10-ASW confocal microscope at ϫ40 magnification. Cytohistochemistry in BmN cells were performed the same as in HEK 293 cells except the expression vector was pIEx-4 containing the BmNPV ie1 promoter as described above.
Statistics-The experimental data were analyzed using Student's t test and analysis of variance. For the t test, *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. For analysis of variance, bars labeled with different lowercase letters are significantly different (p Ͻ 0.05). Throughout the study, values are represented as the mean Ϯ S.D. of five independent experiments.