The Competence Factor βFtz-F1 Potentiates Ecdysone Receptor Activity via Recruiting a p160/SRC Coactivator

Abstract

Hormones provide generalized signals that are interpreted in a specific spatial and temporal manner by a developing or reproducing multicellular organism. The ability to respond to hormones is determined by the competence of a cell or a tissue. The βFtz-F1 orphan nuclear receptor acts as a competence factor for the steroid hormone 20-hydroxyecdysone (20E) in Drosophila melanogaster metamorphosis and mosquito reproduction. The molecular nature of the βFtz-F1 action remains unclear. We report that the protein-protein interaction between βFtz-F1 and a p160/SRC coactivator of the ecdysone receptor, FISC, is crucial for the stage-specific expression of the 20E effector genes during mosquito reproduction. This interaction dramatically increases recruitment of FISC to the functional ecdysone receptor in a 20E-dependent manner. The presence of βFtz-F1 facilitates loading of FISC and the ecdysone receptor on the target promoters, leading to enhanced local histone H4 acetylation and robust activation of the target genes. Thus, our results reveal the molecular basis of competence for the stage-specific 20E response.

In multicellular organisms, hormones orchestrate development, growth, and reproduction. Competence is the crucial mechanism for translating generalized signals, such as hormone stimulation/release, into specific spatial and temporal cellular responses (5). Therefore, an understanding of the molecular nature of competence is a fundamental issue in developmental biology (37).

In insects, the steroid hormone 20-hydroxyecdysone (20E) functions as a generalized systemic signal coordinating critical developmental events of embryogenesis, larval molting, metamorphosis, and, in some insects, reproduction (29, 33, 34). 20E mediates its biological activities through the ecdysone receptor (EcR) complex, a heterodimer consisting of two nuclear hormone receptors, EcR and the retinoid X receptor homologue Ultraspiracle (USP). In turn, the ecdysone receptor acts via binding to its cognate binding sequence, the ecdysone response element (EcRE), which is found in a number of primary-response early genes, and activates transcription of these genes in a 20E-dependent manner (35, 41). The products of these early genes, including E74, E75, and Broad, are transcriptional regulators that modulate the expression of a larger group of secondary-response effector genes, leading to the initiation of developmental events (41).

Although the 20E signaling pathway is utilized reiteratively in diverse biological processes, it remains an open question how a single hormone triggers the expression of vastly different sets of genes in a distinct spatial and temporal pattern (21, 26, 36, 50). The stage- and tissue-specific outcomes of hormone stimulation depend on cell competence. The βFtz-F1 orphan nuclear receptor has been postulated to function as a competence factor for stage-specific responses to 20E during Drosophila melanogaster metamorphosis (5, 24, 48). The elevation of 20E titers in late-third-instar larvae leads to induction of βFtz-F1 during the mid-prepupal stage. βFtz-F1 mutants pupariate normally in response to the late larval pulse of 20E but display defects in response to the subsequent 20E pulse in prepupae (48). Moreover, ectopic βFtz-F1 expression in late-third-instar larvae leads to premature induction in larval salivary glands of E93 transcription, which normally occurs in late prepupae. However, the molecular mechanism of the βFtz-F1 action as a competence factor has remained unknown.

In mosquitoes, 20E is the key hormone governing reproductive processes (32, 33). In most mosquitoes, production of each batch of eggs requires a blood meal. Female adult mosquitoes utilize amino acids and other nutrients from the blood of vertebrate hosts to produce yolk protein precursors (YPPs), which are deposited in developing oocytes (31). The ingestion of blood by a female mosquito results in a significant increase in hemolymph ecdysteroid titers. 20E controls the tissue- and stage-specific synthesis of YPPs in the fat body, an insect metabolic tissue functionally analogous to the liver in vertebrates (13). Two EcR isoforms (AaEcR-A and AaEcR-B) and two USP isoforms (AaUSP-A and AaUSP-B) have been characterized in the yellow fever mosquito, Aedes aegypti (45, 46). The transcription of AaEcR-B and AaUSP-B is increased five- to eightfold after blood intake, while elevation of 20E titers also gives rise to upregulation of E74, E75, and Broad (7, 30, 39). Transcripts of the vitellogenin (Vg) gene, encoding the main YPP, increase more than 10,000-fold and become the most abundant mRNA by 24 h post-blood meal (PBM). Functional binding sites of EcR/USP and the early gene products have been identified in the Vg promoter, and these transcriptional regulators are postulated to act in synergy to achieve the robust and precise expression of the Vg gene (19, 40).

A newly emerged female mosquito requires a 3-day preparatory stage, prior to blood feeding, to become fully responsive to 20E with respect to YPP expression. The establishment of competence appears to be under the control of juvenile hormone, which rises rapidly after eclosion (11). We have demonstrated that juvenile hormone regulates the mosquito βFtz-F1 at the posttranscriptional level (52). The presence of the βFtz-F1 protein is closely correlated with acquisition of fat body competence for 20E-activated Vg expression after a blood meal (26, 52). RNA interference (RNAi) experiments indicated that βFtz-F1 was indeed required for the fat body to attain competence for a 20E response (52).

In this study, we utilized the mosquito fat body as a system for studying the molecular nature of βFtz-F1 as a competence factor. We provide evidence that βFtz-F1 interacts directly with FISC, a p160 coactivator of the ecdysone receptor complex, and is crucial for the recruitment of FISC to the Vg gene promoter after 20E activation. Ablation of either FISC or βFtz-F1 expression by RNAi effectively impairs activation of both the 20E-inducible early genes and the 20E effector YPP genes. Thus, it appears that βFtz-F1 serves as a molecular sign to guide the recruitment of EcR/USP/FISC to 20E-inducible target promoters.

MATERIALS AND METHODS

Yeast two-hybrid screen.

A yeast two-hybrid cDNA library was constructed in the pGAD10 Gal4 activation domain vector, according to the manufacturer's instructions (Clontech), using a total of 10 mg poly(A) RNA from fat bodies of adult female mosquitoes. The region of βFtz-F1 encompassing amino acids 339 to 840 (βFtz-F1^339-840) was cloned into the Gal4 DNA-binding domain vector pGBKT7 (Clontech). Saccharomyces cerevisiae strain Y190 was manipulated according to the Matchmaker library user manual (Clontech). A sequential transformation protocol was used with pGBKT7-βFtz-F1 as bait. Positive clones were identified by growth on synthetic dropout-Trp-Leu-His medium with 50 mM 3-amino-1,2,4-triazole (Sigma) and the activity of the lacZ reporter gene in filter assays. The library plasmids from positive clones that expressed both HIS3 and LacZ reporters were recovered and retransformed into yeast cells, together with the original bait, for testing the specificity of protein-protein interaction. These analyses led to a single specific clone, FISC, which was selected for further analysis in this study. Full-length FISC cDNA was cloned by means of 5′ rapid amplification of cDNA ends PCR with the adult female fat body mRNA.

RNA interference assay.

Synthesis of double-stranded RNAs (dsRNAs) and microinjection were performed as described previously (52). Mosquito females were each microinjected intrathoracically with approximately 0.5 μg dsRNA 6 h after emergence. The injected mosquitoes were allowed a period of 5 days for recovery and were then fed blood. Total RNA preparation, cDNA synthesis, and quantitative PCR using SYBR green technology were performed as described previously (52).

Antibodies and coimmunoprecipitation.

A cDNA fragment encoding amino acids 380 to 977 of FISC was cloned into pQE30 (QIAGEN) in frame with the coding region for a six-histidine tag. Purified His-FISC protein was injected into rats (Cocalico Biologicals), and antiserum was affinity purified using immobilized His-FISC. The generation of antigen-purified rabbit polyclonal antibodies against AaUSP and AaEcR, as well as against AaβFtz-F1, has been reported previously (51, 52). The monoclonal antibody against DmUSP (anti-DmUSP) was a gift from Fotis C. Kafatos (Imperial College, London, United Kingdom). Mouse monoclonal antibodies against V5 and β-actin were purchased from Invitrogen and Sigma, respectively.

The EcR-deficient Drosophila melanogaster L57-3-11 cell line, provided generously by Lucy Cherbas (Indiana University), was cotransfected by expression vectors of AaβFtz-F1, AaFISC, AaEcR, and AaUSP and cultured in the presence or absence of 1 × 10⁻⁶ M 20E. Cells were harvested 40 h after transfection in ice-cold NP-40 lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.1% NP-40) supplemented with protease inhibitors. Assays were performed using an immunoprecipitation kit (Roche) following the manufacturer's instructions. Antibodies against AaEcR, AaβFtz-F1, and Drosophila USP were used for this analysis. The resulting immune complexes were precipitated by the addition of protein G-agarose beads. After extensive washing, the complexes were separated by means of sodium dodecyl sulfate-polyacrylamide gel electrophoresis and analyzed using Western blotting with the appropriate antibodies. When the experiment was conducted with 20E, it was added to the binding and washing buffers.

Transient transfection assay.

Drosophila L57-3-11 cells were transfected according to the instructions of Hu et al. (18). The pVg1.0-Luc reporter construct has been described elsewhere (28).

The green African monkey kidney CV-1 cell line (American Type Culture Collection) was used for the mammalian two-hybrid assay and transfected as described elsewhere (51). The reporter plasmid 5× UAS-E1B-TATA-Luc containing five tandem repeats of the GAL4 upstream activating sequence was provided by Richard G. Pestell (47).

Immunohistochemistry.

Dissected fat bodies were processed and stained with affinity-purified antibodies as described previously (16). The following antibodies were used: rabbit anti-AaEcR (1:10), rat anti-FISC (1:50), rabbit anti-AaβFtz-F1 (1:50), Alexa Fluor 488- or 568-goat anti-rat antibody (1:500), and Alexa Fluor 488- or 568-goat anti-rabbit antibody (1:500) (Molecular Probes). Nuclei were detected by means of DAPI (4′,6′-diamidino-2-phenylindole) staining. Fluorescent images of mosquito fat body cells were acquired using a Leica TCS SP2 laser confocal microscope.

Chromatin immunoprecipitation assay.

The emergence of adult mosquitoes was synchronized, and female mosquitoes were collected at 6 h posteclosion (PE), 96 h PE, and 6 h PBM. Chromatin immunoprecipitation (ChIP) assays were then performed as described by Zhu et al. (51). Cross-linked extracts were precipitated with antibodies against acetylated histone H4, EcR, βFtz-F1, FISC, or preimmune serum. Following reversal of the cross-links, DNA was recovered and analyzed using PCR. PCR products were examined by means of slot blot hybridization. The PCR primers used to amplify regulatory regions of the 20E target genes were (i) for the EcRE region (positions −380 to −81) of the Vg promoter (forward, 5′-TCTGGAATCCATTGCAAGCTA-3′; and reverse, 5′-ATTCACAGCATCCTTTCGTTCG-3′) and (ii) for a region about 0.9 kb upstream from the EcRE (positions −1670 to −1282) on the Vg promoter (forward, 5′-AAGGTTCCGTGCTCACTAATGC-3′; and reverse, 5′-AAAGACCTTTCCGACGATTGTC-3′). Controls with nonspecific preimmune serum did not result in amplification of any of these sequences.

A similar protocol was utilized for the sequential ChIP assays, except that the primary immune complex obtained with the USP antibody was eluted using 10 mM dithiothreitol with agitation at 37°C for 30 min. The eluate was diluted 50 times with ChIP dilution buffer (20 mM Tris-HCl, pH 8.1, 150 mM NaCl, 2 mM EDTA, and 1% Triton X-100), followed by reimmunoprecipitation with rabbit preimmune serum or antibody against AaEcR, AaβFtz-F1, or AaFISC.

RESULTS

Isolation of FISC, a βFtz-F1 interacting protein.

In order to understand the role of βFtz-F1 in fat body competence for the 20E response during mosquito egg development, we employed the yeast two-hybrid system to identify proteins associated with βFtz-F1. A chimeric protein, composed of the GAL4 DNA-binding domain and a portion of βFtz-F1 (amino acids 339 to 840, encompassing the putative DNA-binding and ligand-binding domains), was used as bait. Screening of the Aedes adult fat body cDNA library using this chimeric protein resulted in isolating a 2.7-kb cDNA clone that putatively encoded a polypeptide most related to Taiman in the fly genome. First identified as a regulator of the migration of specific follicle cells (border cells), Drosophila Taiman closely resembles the steroid hormone receptor coactivator AIB1 and functions as a potent coactivator of ecdysone receptor-dependent transcription (2). The identified mosquito protein was named FISC (βFtz-F1 interacting steroid receptor coactivator). Full-length FISC cDNA was subsequently cloned; its nucleotide sequence analysis indicated that it putatively encoded a protein of 1,443 amino acid residues. Sequence alignment of the full-length FISC and Taiman revealed a striking homology at the N-terminal region, which contained a basic helix-loop-helix (bHLH) domain, two Per/ARNT/Sim domains (PAS A and PAS B), and a PAS-associated C-terminal domain (Fig. 1). Five LXXLL/LLXXL (L, leucine; X, any amino acid) motifs were localized in the middle, while two more were scattered in the carboxy-terminal glutamine-rich (Q-rich) region. These structural features found in FISC are characteristics of the p160 steroid receptor coactivator (SRC) family (6). The leucine-rich motifs have been shown to be indispensable for the interaction between p160/SRC coactivators and steroid receptors (17, 42).

FIG. 1.

FISC resembles a steroid receptor coactivator of the p160 family. (A) Deduced amino acid sequence of FISC. The bHLH domain is underlined in the deduced amino acid sequence of FISC. The PAS and PAS-associated C-terminal domains are shaded. The LXXLL motifs are shown in bold and underlined. The nucleotide sequence has been deposited in the GenBank database under accession number DQ469817. (B) Schematic alignment of mosquito FISC, Drosophila Taiman, and human AIB1. The degree of amino acid sequence identity within the bHLH and PAS domains is expressed as a percentage.

FISC is required for 20E-activated gene expression in the fat body.

Analysis of FISC expression showed that the FISC transcript was present in fourth-instar larvae and adult females but barely detectable in pupae and adult males (Fig. 2A). In the adult female, FISC mRNA was expressed in the fat body, ovary, and gut both before and after blood ingestion. Detailed analysis of FISC transcript kinetics showed that FISC gene expression was increased in the fat body shortly after blood feeding, with the second major peak at 24 h PBM (Fig. 2B).

FIG. 2.

Expression of FISC in the mosquito Aedes aegypti. (A) mRNA profile of FISC in mosquitoes of different developmental stages. Total RNAs were isolated from early-fourth-instar larvae (12 to 24 h post-third molt), early pupae (12 to 24 h postpupation), and the indicated adult mosquitoes. FISC transcripts were measured using real-time reverse transcription-PCR and normalized to β-actin expression. Arbitrary units are plotted against developmental time. Representative data (means ± standard errors of the means) from three independent experiments are shown. (B) FISC mRNA profile in the fat body of adult female mosquitoes.

To gain insight into the function of FISC during the egg developmental cycle, we utilized a reverse-genetics approach by injecting dsRNA corresponding to the FISC gene into adult female mosquitoes (Fig. 3). We compared this effect of FISC RNAi with that of knockdown of βFtz-F1. Silencing of βFtz-F1 and FISC resulted in a significant decline in their respective mRNA levels (Fig. 3, βFtz-F1 and FISC). The effectiveness of RNAi knockdowns of FISC and βFtz-F1 was additionally demonstrated by means of immunohistochemistry assays and Western blotting experiments (see Fig. 6 and 8D). Transcription of FISC was not altered in βFtz-F1 RNAi mosquitoes, suggesting that the FISC gene is not under the transcriptional control of βFtz-F1 (Fig. 3, FISC). βFtz-F1 transcription in FISC RNAi mosquitoes was not affected at the beginning of vitellogenesis, but at 12 h and 24 h PBM, its levels were reduced by about 50% (Fig. 3, βFtz-F1).

FIG. 3.

FISC and βFtz-F1 are required for a proper 20E response in the fat body after blood ingestion. dsRNAs were injected into the thoraxes of female mosquitoes as described in Materials and Methods. The transcripts of the indicated genes in female mosquitoes were measured using real-time reverse transcription-PCR at the indicated time after blood feeding and were normalized to β-actin expression. WT, uninjected A. aegypti Rockefeller/UGAL strain; Mal RNAi, injected with double-stranded RNA complementary to bacterial malE; FISC RNAi, injected with FISC dsRNA; Ftz-F1 RNAi, injected with βFtz-F1 dsRNA. Arbitrary units are plotted against developmental time expressed as hours PBM. Representative data (means ± standard errors of the means) from at least three independent experiments are shown.

FIG. 6.

EcR/USP, FISC, and βFtz-F1 are translocated into the nucleus of the fat body after blood feeding. The fat body was stained with polyclonal antibodies against EcR (green), FISC (red), and βFtz-F1 (green). Cells were also stained with DAPI to visualize nuclei (blue). Stained fat body trophocyte cells were examined using confocal microscopy. Note the altered cellular distribution of EcR and FISC in fat bodies with downregulated βFtz-F1 expression (βFtz-F1 RNAi). WT, uninjected A. aegypti Rockefeller/UGAL strain.

FIG. 8.

βFtz-F1 is crucial for the recruitment of EcR/USP/FISC to the 20E-responsive promoters. (A) βFtz-F1, FISC, and EcR/USP are associated with the Vg promoter in the fat body after blood feeding. Fat bodies of female mosquitoes were collected at 6 h PE, 96 h PE, and 6 h PBM. ChIP experiments were performed using antibodies against acetylated H4 (Ac-H4), EcR, βFtz-F1, and FISC, followed by PCR analysis of the indicated regions of the Vg promoter. The amounts of input promoter DNA were also examined using PCR. (B) Binding of EcR/USP, βFtz-F1, and FISC is restricted to the proximal region harboring EcRE on the Vg promoter. ChIP was carried out as described for panel A, except that a region about 0.9 kb upstream from the EcRE (positions −1670 to −1282) on the Vg promoter was analyzed. (C) Injection of Mal dsRNA does not affect the recruitment of FISC to the Vg promoter. (D) Protein levels of EcR, βFtz-F1, and FISC in the fat bodies used for ChIP assays. Aliquots of fat body cell lysates were subjected to immunoblot analyses with antibodies as indicated. (E) Cooccupancy of EcR/USP, βFtz-F1, and FISC on the Vg promoter. Chromatin fragments from fat bodies of female adults at 6 h PBM were first incubated with USP antibody. An aliquot of total soluble chromatin was set aside without immunoprecipitation and used as the input fraction. USP-bound DNA complexes were precipitated, eluted, and reimmunoprecipitated (Re-ChIP) with antibody against EcR, βFtz-F1, or FISC. After reverse cross-linking, purified DNA was amplified with primers specific for the Vg promoter. A rabbit immunoglobulin G (IgG) was used to detect any nonspecific immunoprecipitated DNA. The gel shown of the PCR products is representative of two separate experiments. WT, uninjected A. aegypti Rockefeller/UGAL strain.

We then examined the effects of RNAi knockdowns of FISC and βFtz-F1 on the expression of the transcriptional regulators in the 20E signaling hierarchy, E74B and E75A, as well as on the fat body-specific 20E response effector genes, Vg and VCP (vitellogenic carboxypeptidase), in the fat body after the onset of vitellogenesis (Fig. 3). Injection of dsRNA corresponding to either FISC or βFtz-F1 eliminated the E74B expression peak observed at 24 h PBM (Fig. 3, E74B). Knockdown of FISC resulted in a greater decrease in E75A expression than that observed for mosquitoes injected with βFtz-F1 dsRNA; however, in both cases, the stage-specific expression of this early gene was significantly diminished (Fig. 3, E75A). RNAi-mediated gene silencing of βFtz-F1 or FISC led to a reduction of expression of Vg and VCP to background levels (Fig. 3, Vg and VCP). The fact that downregulation of both βFtz-F1 and FISC exhibited similar effects on the 20E-inducible early genes and the 20E response effector genes suggested that βFtz-F1 and FISC were functionally linked during transcriptional activation of these genes in the stage-specific 20E response (Fig. 3).

Both βFtz-F1 and FISC are essential for 20E-induced high-level expression of Vg.

Given the structural similarity of Aedes FISC, Drosophila Taiman, and other p160 coactivators, we investigated whether FISC functionally interacted with the EcR/USP and influenced its activity on the effector gene promoter. We focused our study on Vg, which is regulated cooperatively by the EcR/USP dimer and the early gene products (19, 28, 40). The pVg1.0-Luc reporter construct, harboring a 1.0-kb mosquito Vg 5′ regulatory region, was transfected into L57-3-11 cells (an EcR-deficient derivative of the Drosophila melanogaster Kc cell line), along with the expression vectors for EcR-B (referred to hereafter as EcR), USP-B (referred to as USP), βFtz-F1, and FISC. After transfection, the cells were incubated with 1 × 10⁻⁶ M 20E or a solvent vehicle. Cotransfection of EcR and USP increased activity of the Vg promoter threefold in the presence of 20E, whereas neither FISC nor βFtz-F1 alone had any marked effect on the activity of the pVg1.0-Luc reporter (Fig. 4). The addition of FISC to EcR and USP resulted in an eightfold 20E-dependent increase in the reporter activity.

FIG. 4.

βFtz-F1 potentiates 20E activation of the Vg promoter. (A) Schematic illustration of the βFtz-F1 binding sites in the region of the ecdysone response elements on the Vg promoter. (B) Schematic diagram of βFtz-F1 mutants. DBD, DNA-binding domain; LBD, ligand-binding domain; F, Ftz-F1 box. (C) Effects of βFtz-F1 and FISC on EcR/USP-mediated transactivation of the Vg promoter. L57-3-11 cells were transfected with a pVg1.0-Luc reporter construct and the indicated expression plasmids. After transfection, cells were cultured in control medium (CM) or medium with 1 × 10⁻⁶ M 20E. Data represent ratios of firefly luciferase to Renilla luciferase activity (relative luciferase activity), and the values shown are the means from three independent experiments, with error bars representing the standard deviations of the means. (D) Mutation in the DNA-binding domain or Ftz-F1 box abolishes binding of βFtz-F1 to its cognate DNA sequence. EMSA was performed using ³²P-labeled oligonucleotides containing consensus βFtz-F1 binding sites and in vitro-synthesized βFtz-F1, βFtz-F1^C357A, and βFtz-F1^414A2. The bottom panel shows a Western blot of the in vitro-synthesized βFtz-F1 and its derivatives with polyclonal βFtz-F1 antibodies. (E) Overexpression of βFtz-F1 does not affect the levels of AaFISC, AaEcR-B, and AaUSP-B in the cell transfection assay. A portion of the transfected L57-3-11 cells was analyzed with antibodies against AaβFtz-F1, AaEcR, AaUSP, and AaFISC. The expression vectors used in this experiment are indicated at the top of the panel.

Three EcREs have previously been characterized in the Vg promoter (28). Sequence analysis also revealed several βFtz-F1 binding sites within or in the vicinity of the EcREs of the Vg promoter (Fig. 4A), and in vitro binding of βFtz-F1 to these sites was confirmed by the results of electrophoretic mobility shift assay (EMSA) experiments (data not shown). In light of the requirement for both βFtz-F1 and FISC in activation of 20E-responsive genes following a blood meal, we examined whether βFtz-F1 directly participated in the EcR/USP-mediated transactivation of Vg. L57-3-11 cells were transiently transfected by reporter construct pVg1.0-Luc, together with expression vectors for EcR, USP, βFtz-F1, and FISC. Transfection of βFtz-F1 alone did not exhibit any notable effect on the reporter activity, irrespective of the presence or absence of the hormone (Fig. 4C). Cotransfection of βFtz-F1 with either EcR/USP or FISC yielded no further activation of the reporter. However, when βFtz-F1 was expressed along with the EcR/USP complex and FISC, the activity of the reporter was increased dramatically in a 20E-dependent manner (Fig. 4C). Given that FISC was isolated based on its interaction with βFtz-F1, these results raised the possibility that the competence factor βFtz-F1 modulated the 20E response by facilitating recruitment of FISC to the ecdysone receptor-dependent target genes.

To determine whether binding to the Vg promoter was required for the βFtz-F1 action, we constructed two derivatives of βFtz-F1 that failed to bind the 5′-PyCAAGGPyCPu-3′ sequence with high affinity. In βFtz-F1^C357A, a cysteine residue of the first zinc finger motif in the DNA-binding domain was replaced with an alanine residue (Fig. 4B). Members of the Ftz-F1 nuclear receptor family have a characteristic conserved basic amino acid-rich region, known as the Ftz-F1 box, which is adjacent to the C-terminal end of the zinc finger motif and is involved in stabilizing monomeric binding (44). Accordingly, in βFtz-F1^414A2, the two glycine residues at positions 414 and 415 in the Ftz-F1 box were changed to alanine residues (Fig. 4B). EMSA experiments confirmed that both mutated βFtz-F1 proteins failed to recognize their cognate binding sequences (Fig. 4D). We repeated the above-described cell transfection experiment using expression vectors of βFtz-F1^C357A and βFtz-F1^414A2 in lieu of βFtz-F1. Neither was able to substitute for the intact βFtz-F1 in the 20E-mediated activation of the Vg promoter (Fig. 4C). Instead, the βFtz-F1 point mutants reduced the activity of EcR/USP and FISC, perhaps by sequestering FISC away from the EcR/USP complex. These data imply that DNA binding of βFtz-F1 is crucial for its action on 20E-responsive genes, such as the Vg gene.

FISC forms a protein complex with both EcR/USP and βFtz-F1.

Next, we investigated interactions of FISC, βFtz-F1, and the EcR/USP complex at the protein-protein level. In the first set of experiments, L57-3-11 cells were cotransfected with the expression plasmids for EcR, USP, and βFtz-F1. The last plasmid was expressed as a fusion protein with the V5 epitope tag. Whole-cell protein extract was then incubated with the anti-EcR antibodies, precipitated using protein G-agarose, and resolved by means of sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Western blot analysis using the V5 tag antibody revealed that βFtz-F1 did not form a protein complex with the EcR/USP complex, irrespective of the hormone treatment (Fig. 5A, group 1). When the FISC-V5 construct was used in a similar experiment in place of the βFtz-F1-V5 fusion construct, the results demonstrated that FISC bound the EcR/USP complex in a 20E-dependent manner (Fig. 5A, group 2). Finally, when FISC-V5 was expressed with βFtz-F1, immunoprecipitation with anti-βFtz-F1 antibodies showed protein-protein interaction of βFtz-F1 and FISC in a hormone-independent manner (Fig. 5A, group 3). The inputs of these proteins were analyzed using Western blot analysis, which indicated that the differences we observed were not due to variation of expression of individual proteins.

FIG. 5.

FISC physically binds to both EcR/USP and βFtz-F1. (A and B) Protein complexes identified by immunoprecipitation (IP) experiments. L57-3-11 cells were transfected with the indicated expression vectors. Cell extracts were incubated with antibody (Ab) against EcR or βFtz-F1 (A) or USP (B). The precipitated proteins were detected by immunoblotting (IB) with anti-V5 antibody. In the bottom of panel A, 50% of the inputs were loaded and subjected to Western blot analysis with the indicated antibodies. An additional protein band (asterisk) reacted to FISC antibodies, presumably representing a FISC derivative generated by using an alternative start codon. (C) Protein interactions in the fat body detected by gel mobility shift assays. Nuclear proteins were extracted from fat bodies of female mosquitoes at 6 h post-blood meal. Gel shift assays were performed with a ³²P-labeled probe corresponding to the EF fragment of the Vg promoter (Fig. 4A). FBNE, fat body nuclear extracts.

We then tested whether FISC formed a multiprotein complex with both βFtz-F1 and EcR/USP. In the control experiment, in which L57-3-11 cells were cotransfected with the expression plasmids for EcR, USP, and βFtz-F1-V5 fusion construct, immunoprecipitation using anti-USP antibodies confirmed the lack of direct protein-protein interaction between EcR/USP and βFtz-F1 (Fig. 5B, group 1). However, when EcR/USP and βFtz-F1 were coexpressed together with FISC-V5, βFtz-F1 was found to be associated with EcR/USP in a 20E-dependent manner (Fig. 5B, group 2), suggesting that FISC bridges the interaction between βFtz-F1 and EcR/USP in the formation of the multiprotein complex.

In order to detect the presence of the EcR/USP/FISC/βFtz-F1 complexes in vivo, we performed EMSA experiments using nuclear proteins extracted from early vitellogenic fat bodies of female mosquitoes at 6 h PBM. Multiple bands were observed when the nuclear extract was incubated with a 40-bp oligonucleotide (the EF fragment shown in Fig. 4A) containing Vg EcRE and binding sites for βFtz-F1 (Fig. 5C). These protein-DNA interactions were insensitive to the 20E treatment (Fig. 5C, lanes 2 and 3). The band with the highest molecular weight was supershifted by antibody against USP, βFtz-F1, or FISC, suggesting that EcR/USP, βFtz-F1, and FISC bind to this specific regulatory region of the Vg promoter as components of a large protein complex. The specificity of binding was confirmed by competition with unlabeled specific and nonspecific probes (Fig. 5C, lanes 7 and 8).

To further elucidate the interaction of βFtz-F1, FISC, and EcR/USP, we investigated the subcellular localization of these factors in the previtellogenic and vitellogenic (6 h PBM) mosquitoes. Fat bodies of adult female mosquitoes were immunostained with affinity-purified polyclonal antibodies against EcR, FISC, and βFtz-F1. As shown in Fig. 6, at 120 h PE, EcR, FISC, and βFtz-F1 were present in both the cytoplasm and the nucleus in fat body cells. In contrast, at 6 h PBM, after 20E activation of vitellogenesis, all three proteins exhibited nuclear colocalization, implying that nuclear translocation occurs during the onset of vitellogenesis. Importantly, the nuclear accumulation of both EcR and FISC was severely disrupted in mosquitoes with βFtz-F1 RNAi knockdown. Thus, these data provide strong evidence that βFtz-F1 plays a key role in the nuclear accumulation of EcR and FISC.

Characterization of the interactions among FISC, βFtz-F1, and EcR/USP.

To delineate the domains that were required for these protein interactions, we resorted to mammalian two-hybrid assays. Distinct regions of FISC were fused to the activation domain of VP16, while βFtz-F1 (339 to 840 amino acids, without amino-terminal A/B domains) was fused to the DNA-binding domain of GAL4 (Fig. 7A). Full-length FISC interacted with βFtz-F1 regardless of the 20E presence, consistent with the results of coimmunoprecipitation experiments (Fig. 7B). FISC was then split into three distinct regions: FISC^1-696, the amino-terminal region containing the bHLH/PAS domain; FISC^697-972, the central region harboring the cluster of five LXXLL motifs; and FISC^973-1443, the carboxy-terminal polyglutamine region (Fig. 7A). None of these regions alone was able to interact with βFtz-F1 (Fig. 7B). After removal of the carboxy- and amino-terminal regions, the remaining FISC domains (FISC^1-972 and FISC^697-1443, respectively) were subsequently assayed in these two-hybrid interaction experiments. The results indicate that FISC^697-1443 is responsible for most of the interaction between FISC and βFtz-F1 (Fig. 7B). However, the ligand-binding domain of βFtz-F1 (βFtz-F1^646-840) was unable to recognize FISC. The protein-protein interaction with FISC was retained mostly in βFtz-F1^C357A and βFtz-F1^414A2 mutants; however, when a portion of the Ftz-F1 box was truncated (βFtz-F1^Δ405-428), βFtz-F1 lost its ability to bind FISC. These observations suggest that the Ftz-F1 box is an essential part of the interaction surface for βFtz-F1 in complex with FISC.

FIG. 7.

βFtz-F1 enhances the 20E-dependent interactions between EcR/USP and FISC. (A) Schematic diagram of EcRb, USPb, βFtz-F1, and FISC. (B to D) Protein interactions were tested in a mammalian two-hybrid system. CV-1 cells were transfected with 5× UAS-TATA-Luc reporter construct, along with expression vectors for GAL4 and VP16 fusions, and full-length USPb and βFtz-F1 proteins, as indicated. After transfection, cells were cultured in medium with or without 1 × 10⁻⁶ M 20E. Data represent ratios of firefly luciferase to Renilla luciferase activity (relative luciferase activity), and the values shown are the means from three independent experiments, with error bars representing the standard deviations of the means. For EcRΔAF2, the AF2 core was removed from GAL4-EcR^348-675. DBD, DNA-binding domain; LBD, ligand-binding domain; CM, control medium.

As for the FISC and EcR/USP interaction, the EcR/USP heterodimer, and not EcR alone, was able to bind FISC in the presence of 20E (Fig. 7C). Further analyses of FISC indicated that the central region harboring the LXXLL repeats accounts for its interaction with the ligand-binding domains of the EcR/USP heterodimer. When the AF2 activation domain was removed from EcR (pM-EcRΔAF2), the 20E-dependent association of FISC with EcR/USP was no longer detectable. This is in line with the fact that the AF2 helix in other nuclear receptors provides an essential contact surface for the binding of LXXLL motif-containing coactivators. Thus, FISC appears to interact with EcR/USP and βFtz-F1 through distinct regions, allowing the cofactor to simultaneously engage βFtz-F1 and the ecdysone receptor complex.

To unravel the molecular mechanism underlying the modulation effect of βFtz-F1 on the 20E response of Vg, we explored the possibility that βFtz-F1 reinforced interaction between FISC and EcR/USP. As shown in Fig. 7D, EcR/USP was associated with FISC only in the presence of 20E. The further addition of full-length βFtz-F1 indeed augmented the interaction between FISC and EcR/USP in a 20E-dependent manner (Fig. 7D). Importantly, this function required an intact βFtz-F1, as mutations in the DNA-binding domain and Ftz-F1 box disturbed these enhanced protein interactions. However, it is possible that, in this experiment, the βFtz-F1^C357A and βFtz-F1^414A2 mutants could bind FISC and compete it away from EcR/USP, contributing to the deteriorated protein interactions.

Chromatin immunoprecipitation assays identify EcR, βFtz-F1, and FISC on the native 20E-responsive promoter after blood feeding.

To elucidate whether EcR/USP, βFtz-F1, and FISC assemble into relevant transcriptional complexes on the 20E-responsive promoter, we analyzed occupancies of EcR, FISC, and βFtz-F1 in the vicinity of EcREs in the Vg gene. Fat bodies were collected from female mosquitoes at 6 h PE, 96 h PE, and 6 h PBM, and ChIP assays were carried out. As shown in Fig. 8A, acetylation of histone H4 associated with the Vg promoter increased after blood ingestion, concomitant with activation of Vg. EcR, βFtz-F1, and FISC were barely detectable on the Vg promoter in unfed female mosquitoes at 6 h and 96 h PE. Conversely, ChIP analysis clearly reflected a dramatic increase in the association of EcR, βFtz-F1, and FISC with the Vg EcRE at 6 h PBM. The protein-DNA association was specific for the region containing the Vg EcRE, as evidenced by the lack of detectable products using primer pairs complementary to a distal upstream regulatory region of the Vg gene lacking EcRE (Fig. 8B).

Having established that EcR, βFtz-F1, and FISC selectively associated with the EcRE region of the Vg promoter after the activation of its expression, we further explored the role of βFtz-F1 in the formation of this multiprotein complex. In this case, female mosquitoes were injected individually with dsRNA complementary to EcR, FISC, βFtz-F1, or MalE (negative control) at 6 h PE. Fat body samples were then collected at 96 h PE and 6 h PBM and subjected to ChIP assays. As a control, injection of dsRNA corresponding to MalE did not incur any results markedly different from those with the uninjected female mosquitoes (Fig. 8C). Knockdown of any one of EcR, FISC, or βFtz-F1 led to a decrease in H4 acetylation at 6 h PBM (Fig. 8A). Knockdown of EcR resulted in significant reduction of EcR and FISC signals but not βFtz-F1, suggesting that EcR binding and βFtz-F1 binding to chromatin are not linked. Likewise, knockdown of βFtz-F1 resulted in a significant decline of βFtz-F1 and FISC signals as well as diminished chromatin binding by EcR. The reduction of FISC did not affect chromatin binding of EcR and βFtz-F1 on the Vg promoter (Fig. 8A). Western blot analysis demonstrated that, at 6 h PBM, the amounts of EcR and FISC in fat bodies of βFtz-F1 RNAi mosquitoes were comparable to those of the control mosquitoes (Fig. 8D), suggesting that βFtz-F1 plays an essential role in recruiting FISC on the Vg promoter.

To determine unequivocally whether EcR/USP, FISC, and βFtz-F1 occupied the same portion of chromatin at the same time, we performed sequential ChIP experiments. Cross-linked chromatin from fat bodies of mosquitoes at 6 h PBM was first treated with the USP antibody. After extensive washes, the resulting immune complex was then eluted and subjected to immunoprecipitation separately with EcR, βFtz-F1, and FISC antibodies. As shown in Fig. 8E, the Vg EcRE region present in the first immune complex was pulled down by any of the three antibodies. Taken together, these data strongly indicate that in the mosquito fat body, βFtz-F1 defined the 20E response after a blood meal by enhancing the recruitment of FISC to the EcR/USP complex at the regulatory sites of their target genes and that this is achieved through protein-protein interaction with FISC.

DISCUSSION

In Drosophila, during the late prepupal stage, βFtz-F1 mutations abolish expression of the intermediate transcriptional regulators in the 20E signaling hierarchy, including BR-C, E74A, E75A, and E93 (5, 23). The βFtz-F1 protein appears to be directly involved in the regulation of these genes, as antibodies against the Ftz-F1 protein detect staining of some 20E-regulated puffs in the late prepupal salivary gland polytene chromosomes, including the 74EF and 75B early puff loci that contain the E74 and E75 genes (24). Broadus et al. (5) postulated that the possible mechanism of βFtz-F1 action is enhancing the activity of the ecdysone receptor via direct interaction with it at target promoters. However, the molecular mechanism by which βFtz-F1 exerts its function as a competence factor has not been clearly understood. Our present study has revealed that βFtz-F1 regulates the expression of the stage-specific 20E effector genes in mosquitoes by playing a crucial role in the recruitment of a p160/SRC coactivator, FISC, to the ecdysone receptor complex.

In vitro tissue culture experiments indicated that expression of FISC in the female fat body is not regulated by 20E ligand (J. Zhu and A. S. Raikhel, unpublished data). In vitellogenic mosquitoes, injection of FISC dsRNA substantially reduced expression of E74B, E75A, and the fat body-specific 20E response effector genes, Vg and VCP, and impeded egg development after blood feeding (data not shown), suggesting that recruitment of this cofactor is essential for proper 20E responses in both vitellogenic tissues, the fat body and the ovary. Further studies showed that FISC binds EcR/USP only in the presence of 20E ligand and enhances transactivation properties of the heterodimer. FISC and EcR/USP are colocalized in the fat body and translocated into the nucleus after a blood meal. Moreover, ChIP assays suggested that FISC is required for acetylation of histone H4 associated with the Vg promoter. All of these data substantiate FISC as a coactivator of the ecdysone receptor complex in the vitellogenic 20E response.

Recruitment of FISC to the stage-specific 20E target genes is greatly enhanced by the competence factor βFtz-F1. In female βFtz-F1 RNAi mosquitoes, recruitment of FISC by the activated mosquito EcR/USP complex is relatively weak and unable to achieve high-level expression of Vg. Although βFtz-F1 directly binds to the coactivator FISC in vitro in a 20E-independent manner, this interaction is incapable of activating the Vg promoter without participation of the ecdysone receptor complex, as shown in cell transfection (Fig. 4C) and ChIP assays (Fig. 8A). Instead, through its direct binding to FISC, βFtz-F1 enhances the 20E-dependent interaction between FISC and EcR/USP on the target promoters, presumably by bringing high levels of FISC into close proximity to the ecdysone receptor complex.

There are a few precedents indicating that some transcriptional regulators extend the range of coactivator and corepressor action and pass on these cofactors to other transcriptional regulators through protein-protein interactions. For instance, the retinoblastoma tumor suppressor protein (Rb) has been shown to enhance the activity of a subset of nuclear receptors (NGFI-B, HNF-4, SF-1, and ER) through direct interactions with those receptors and their coactivator, SRC-2, at the same time (3). The interactions seem to depend on the levels of Rb in the cell. Rb does not possess a DNA-binding domain and is believed to adjust hormone responsiveness by playing a key role in the stabilization of the nuclear receptor/SRC-2 complexes but not in the recruitment of SRC-2 to the nuclear receptors. As another example, activity of the murine ortholog of Ftz-F1 (SF-1) is regulated by interaction with the nuclear receptor Dax-1, which recruits nuclear receptor corepressor N-CoR to SF-1. SF-1 becomes susceptible to repression by N-CoR only in the presence of DAX-1, as N-CoR does not interact directly with SF-1 (10). Our present work reveals a novel mechanism of interaction among transcriptional regulators and a coactivator, which allows the functional integration of multiple transcriptional factors and enables the outputs of particular signaling pathways to be activated in a stage-specific manner. In recent years, several cofactors, including Bonus, NURF, Rig, SMRTER, Taiman, and TRR, have been implicated in ecdysone signaling during Drosophila development (1, 2, 4, 12, 38, 43). In this study, we have shown that the coactivator FISC acts as a bridge between βFtz-F1 and EcR/USP in the formation of the multiprotein complex, while there is no detectable direct physical attachment between βFtz-F1 and EcR/USP. It appears that the binding of βFtz-F1 at the Vg promoter does not depend on that of EcR/USP and vice versa, although their simultaneous binding seems to contribute to the stability of a functional multiple protein complex on the gene. In this case, the dual binding of βFtz-F1 and EcR/USP serves as a selector of stage-specific target sites in the genome. Simultaneous binding of βFtz-F1 and EcR/USP may also force FISC to adopt a conformation that effectively recruits auxiliary coactivators. Thus, the DNA binding and protein interaction provide a combinatorial code required for specific gene activation by 20E.

Cell transfection experiments in conjunction with mutagenesis have clearly shown that DNA binding of βFtz-F1 is critical for its action on the Vg gene (Fig. 4C). One complexity of βFtz-F1 is the DNA sequences that it recognizes. The consensus PyCAAGGPyCPu sequence encompasses a six-nucleotide nuclear receptor core half-site, and βFtz-F1 binds some imperfect consensus sequences in vitro (Zhu and Raikhel, unpublished). The multiple potential βFtz-F1 binding sites on the Vg promoter, especially those overlapped with or within the EcREs, make it very challenging to explore the function of βFtz-F1 binding by site mutagenesis. Although βFtz-F1 forms a complex with EcR-USP and FISC without involving DNA in the immunoprecipitation experiment, βFtz-F1 appears to require an intact DNA-binding domain and Ftz-F1 box to boost 20E activation of the Vg promoter. Moreover, in the EMSA experiments, the EcR, USP, FISC, and βFtz-F1 proteins from nuclear extracts of the early vitellogenic fat body did not form the multiprotein complex on probes containing either the IR-1 or DR-1 EcRE sequence (data not shown), providing another line of evidence to support the DNA-binding requirement of βFtz-F1. Certain mechanistic details, such as the basis for promoter recruitment of βFtz-F1, remain to be established. In mosquitoes, although the βFtz-F1 protein is present in the fat body of female adults at 3 to 5 days PE, nuclear accumulation and loading of βFtz-F1 on the Vg promoter take place only after activation by blood ingestion. Conversely, in vitro-synthesized proteins fail to form the EcR/USP/FISC/Ftz-F1 complex on the Vg EF fragment in EMSA experiments (Zhu and Raikhel, unpublished). Therefore, these data leave open the possibility that βFtz-F1 itself is modulated by other signals associated with blood feeding and that only the “activated” βFtz-F1 can target the 20E-responsive promoters as needed and facilitate the recruitment of FISC.

The action of βFtz-F1 family members has been believed to be ligand independent because they are constitutively active in cell-based assays. However, recent studies demonstrated that activity of SF-1, the vertebrate ortholog of insect Ftz-F1, is modulated by phospholipids, although the true endogenous ligand(s) has not been established with certainty (22, 27). Moreover, SF-1 activity, in terms of its transactivation, DNA binding, subcellular translocation, and interactions with transcriptional cofactors, is regulated by covalent modifications, such as phosphorylation mediated by mitogen-activated protein kinase, sumoylation, and acetylation by p300 in response to the cyclic AMP signaling (8, 9, 14, 20, 25). In the adult female mosquito, an increase in extracellular amino acid levels after a blood meal is critical for 20E stimulation of YPP gene expression. This amino acid signal is conveyed through the nutrient-sensitive target of rapamycin pathway (15, 16). Target of rapamycin signaling is responsible for a wide range of cellular responses (49). The absolute requirement of this pathway for the regulatory circuitry of postvitellogenic events in the female mosquito raises the possibility that βFtz-F1 is the target of posttranslational modifications. Future studies should provide new insights into how the activity of βFtz-F1 is regulated and whether βFtz-F1 is a convergence point of nutritional inputs and hormonal cues, ensuring precise genetic and biological responses.