Transcription Factor Forkhead Regulates Expression of Antimicrobial Peptides in the Tobacco Hornworm, Manduca sexta

Antimicrobial peptides (AMPs) play an important role in defense against microbial infections in insects. Expression of AMPs is regulated mainly by NF-κB factors Dorsal, Dif and Relish. Our previous study showed that both NF-κB and GATA-1 factors are required for activation of moricin promoter in the tobacco hornworm, Manduca sexta, and a 140-bp region in the moricin promoter contains binding sites for additional transcription factors. In this study, we identified three forkhead (Fkh)-binding sites in the 140-bp region of the moricin promoter and several Fkh-binding sites in the lysozyme promoter, and demonstrated that Fkh-binding sites are required for activation of both moricin and lysozyme promoters by Fkh factors. In addition, we found that Fkh mRNA was undetectable in Drosophila S2 cells, and M. sexta Fkh (MsFkh) interacted with Relish-Rel-homology domain (RHD) but not with Dorsal-RHD. Dual luciferase assays with moricin mutant promoters showed that co-expression of MsFkh with Relish-RHD did not have an additive effect on the activity of moricin promoter, suggesting that MsFkh and Relish regulate moricin activation independently. Our results suggest that insect AMPs can be activated by Fkh factors under non-infectious conditions, which may be important for protection of insects from microbial infection during molting and metamorphosis.

In insects, innate immunity is the first line of defense against pathogens and it is composed of both humoral and cellular immune responses [1][2][3][4][5] . Synthesis of small cationic antimicrobial peptides (AMPs) in the fat body of insects is an important defense mechanism of humoral immune responses against microbial infection [6][7][8][9] . Expression of insect AMPs is regulated mainly by the evolutionarily conserved Toll and immune deficiency (IMD) pathways [10][11][12][13] . The Toll pathway defends against Gram-positive bacteria, fungi and viruses 14,15 , whereas the IMD pathway acts against Gram-negative bacteria 16 . Both the Toll and IMD pathways activate the Rel/NF-κB family of transcription factors, including Dorsal, Dif (Dorsal-related immunity factor) 17 and Relish 18 . These NF-κB factors contain N-terminal Rel-homology domain (RHD) that is required for DNA binding and dimerization 19 .
In our previous study with the moricin promoter in the tobacco hornworm Manduca sexta, we found that both NF-κB and GATA-1 binding sites are crucial for activation of moricin promoter, and we also identified a 140-bp region (between −240 and −100 bp) in the moricin promoter, designated as moricin promoter activating element (MPAE), which contains binding sites for additional transcription factors required for activation of moricin promoter 20 . The purpose of this study is to identify the additional transcription factor(s) that can bind to the MPAE region, investigate whether the new factor(s) can also activate other AMP genes, and whether the new factor(s) acts independently or cooperatively with NF-κB factors. Sequence analysis of the 140-bp MPAE region showed a binding site for silk gland factor-1 (SGF-1), a transcription factor in the silk worm Bombyx mori that is homologous to Drosophila melanogaster forkhead (Fkh). SGF-1 can bind in vitro to a cis-element located in the promoter region of sericin-1, which encodes a silk protein in the middle silk gland (MSG) cells [21][22][23] . Further analysis showed that there are three Fkh-binding sites in the 140-bp MPAE region (Fig. 1A).
Forkhead transcription factors belong to the Fox (Forkhead box) superfamily proteins 24 . Members of the Fox family proteins consist of a conserved long DNA binding domain connected to a pair of loops or "wings" via a small β-sheet 25 . The long DNA binding domain, also known as winged-helix or forkhead domain, is about 90 residues and composed of three α-helices with a helix-turn-helix core, thus, the Fox family proteins are also known as winged helix transcription factors 26 . Fox proteins have been identified in eukaryotic organisms from yeast to human, and they play important roles in various biological processes, including cellular differentiation, development, metabolism, insulin signaling, immune regulation, cancer development, and aging [27][28][29][30][31][32][33] . Forkhead gene A (FoxA) was first discovered in D. melanogaster since mutation in this gene resulted in fork-headed structure in the embryos with defects in the anterior and posterior gut formation 34 . Later, cDNA encoding hepatocyte nuclear factor 3α (HNF3α), also known as FoxA1, was cloned in rat 35 . The central 90-residue DNA-binding domain (forkhead domain) of D. melanogaster FoxA and human FoxA1 shows shockingly high (>85%) identity 36 . Since the discovery of Drosophila FoxA, hundreds of Fox genes have been identified in various species. There are at least 4 Fox genes in yeast, 16 in D. melanogaster, 15 in Caenorhabditis elegans, 44 in mice, and 50 in humans [37][38][39] . The genome of Aedes aegypti contains eighteen loci that encode putative Fox factors and six of them are involved in reproduction 40 . Based on phylogenetic analysis, metazoan Fox family is divided into 19 subfamilies (FoxA-S) 24 . Human Fox subfamilies are classified into two classes based on the sequence homology within and beyond the forkhead domain: Fox A-G, I-L and Q are class 1, Fox H and M-P are class 2 Fox proteins 24 .
In D. melanogaster, several Fox transcription factors have been described. Crocodile (FoxC) is involved in early patterning in embryo 41 , Biniou (FoxF) plays a role in visceral mesoderm 42 , sloppy paired 1 & 2 (FoxG) are involved in early embryo segmentation 43,44 , domina (FoxN) is involved in vitality and fertility 45 , and FoxO regulates the insulin signaling in the brain and fat body 46 . Drosophila FoxO is also involved in innate immunity and expression of AMPs 47,48 . Forkhead (Fkh) is the founding member of the FoxO family and it is activated upon TOR inhibition by rapamycin 48 . However, functions of Fox transcription factors in regulation of AMPs in other insect species have yet to be reported.
In this study, we identified Fkh-binding sites in the promoters of M. sexta moricin and lysozyme genes, cloned Fkh genes from both M. sexta and D. melanogaster, and demonstrated activation of AMP gene promoters by Fkh factors. We also found that Fkh mRNA was undetectable in Drosophila S2 cells, which may account for low activity of some AMP promoters in S2 cells. In addition, co-expression study showed that M. sexta Fkh (MsFkh) interacted with Relish (Rel2)-RHD, but did not show any interaction with Dorsal-RHD. However, co-expression of MsFkh and MsRelish-RHD did not have an additive effect on the activity of moricin promoter compared to expression of MsFkh or MsRelish-RHD alone, suggesting that MsFkh and MsRelish regulate AMP gene Electrophoretic mobility shift assay (EMSA). Cytosolic (Cyto) and nuclear (Nuc) proteins were prepared from Sf9 and S2 cells and incubated with biotinylated MPAE-1, −2 or −3 fragments for EMSA assays as described in the Materials and Methods. Lanes 1, 6 and 15 contained only biotinylated MPAE-1, −2 and −3; lanes 2-5, lanes 7-10 and lanes 11-14 contained biotinylated MPAE-1, MPAE-2 and MPAE-3, respectively, with cytosolic or nuclear proteins. Only nuclear proteins from Sf9 cells (Sf9-Nuc) bound to all three biotinylated MPAE fragments and caused mobility shift of the DNA fragments (lanes 5, 10 and 14, arrow). expression independently. Activation of AMPs by Fkh factors under non-infectious conditions is particularly important for insects during molting and metamorphosis to protect them from microbial infections.

Results
Binding of nuclear proteins in Sf9 cells to MPAE of M. sexta moricin promoter. In our previous study, we identified a 140-bp MPAE region in the M. sexta moricin, which was responsible for activation of moricin and D. melanogaster drosomycin promoters in S. frugiperda Sf9 cells 20 . To identify nuclear factor(s) in Sf9 cells that can bind to the MPAE region, we divided the 140-bp MPAE region into three fragments (MPAE-1 to 3) (Fig. 1A), and performed electrophoretic mobility shift assay (EMSA). EMSA results showed that proteins from nuclear extracts of Sf9 cells, but not S2 cells, bound to all three MPAE fragments (Fig. 1B, lanes 5, 10 and 14, arrow). We then analyzed transcription factor binding sites (AliBaba 2.1 program, http://www.gene-regulation. com) in the three MPAE fragments, and found that there is an SGF-1 binding site in MPAE-3 (Fig. 1A). SGF-1 is a transcription factor in the silkworm B. mori that regulates expression of silk protein Sericin-1 in the middle silk gland cells 22,23 , and it is a homolog of D. melanogaster forkhead (Fkh). Fkh is a member of the FoxO family proteins and FoxO binds to the consensus sequence of (T/C)(G/A)AAACAA 49 . We found an ATAAACA sequence in both MPAE-2 and MPAE-3, and an ATAAAGA sequence in MPAE-1 (Fig. 1A). Thus, we speculate that Fkh transcription factor in Sf9 nuclear extracts bound to all three MPAE fragments.
To further determine DNA binding motif in the MPAE region, we selected MPAE-3 fragment and divided it into three smaller fragments (22-24 bp) with 10-bp overlapping region (MPAE-3a to 3c) ( Fig. 2A) for competitive EMSA assays. The results showed that only MPAE-3b fragment, which contains the ATAAACA sequence, competed with binding of the labeled MPAE-3 to nuclear proteins of Sf9 cells and the competition by MPAE-3b was dose-dependent ( Since MPAE did not bind to nuclear proteins in S2 cells (Fig. 1B), and Drosophila Fkh (DmFkh, CG10002) is not expressed in S2 cells (Flybase), we performed real-time PCR to determined transcripts of several Drosophila Fox genes in S2 cells, including DmFkh, FoxK long and short isoforms, jumeau (jumu) and dFoxO. The results showed that mRNAs of FoxK, jumu and dFoxO, but not DmFkh, were detected in S2 cells, with higher transcript level of FoxK and jumu than dFoxO (Fig. 4A), confirming that DmFkh was not expressed (or was expressed at an undetectable level) in S2 cells. To test whether overexpression of DmFkh in S2 cells can activate expression of AMP genes, DmFkh was also cloned.
To investigate activation of M. sexta moricin and lysozyme promoters by forkhead factors, MsFkh and DmFkh were over-expressed in S2 cells, and dual luciferase assays were performed with several truncated moricin and lysozyme promoters. MsFkh is 355 residues long with a theoretical molecular weight of 39.5 kDa and pI of 9, while DmFkh (Forkhead isoform A) is 510 residues with molecular weight of 54.3 kDa and pI of 8.7. The full length MsFkh shows 96% identity in amino acid sequence to B. mori SGF-1 (Genbank accession no. NP_001037329.1), 95% identity to Helicoverpa armigera (Genbank accession no. AAW56613.1) and Spodoptera exigua (Genbank accession no. ACA30303.1) forkhead domain transcription factors, but only 42% identity to the full length DmFkh. However, the forkhead domains of MsFkh and DmFkh proteins share 98% identity. Dual luciferase results showed that both MsFkh and DmFkh activated M. sexta moricin, lysozyme, defensin-1, defensin-3 and attacin-2 promoters, but did not activate M. sexta cecropin, attacin-1, or defensin-2 promoter (Fig. 4B-D). DmFkh stimulated the activity of AMP promoters to a similarly high level as or to a higher level than MsFkh. Comparing different truncated promoters of M. sexta moricin and lysozyme promoters, the 1.4-kb moricin promoter (Mor-1400) was activated by Fkh factors to a similarly high level as the 242-bp truncated moricin promoter

Identification of Fkh-binding sites in M. sexta moricin and lysozyme promoters. To identify
Fkh-binding sites in the MPAE region of M. sexta moricin promoter and the 345-bp region of lysozyme promoter that are responsible for activation by Fkh factors, we first analyzed the Fkh-binding sites in the 140-bp MPAE and the 345-bp region of lysozyme promoter. Three and four Fkh-binding sites with the core sequence of AAACA were predicted within the 140 bp MPAE of moricin and 345 bp of lysozyme promoters, respectively (Figs. 1A, 5A,B and S1). To determine the active Fkh-binding sites in moricin and lysozyme promoters, the truncated Mor-242 and Lyz-345 promoters were selected for mutation of each Fkh-binding site for dual luciferase assays (Fig. 5A,B). Mutation of Fkh-binding site 2 or 3 alone in the Mor-242 promoter significantly decreased the activity of the mutant promoter by more than 60% compared to Mor-242 promoter, and mutation of both sites 2 and 3 together completely abolished activation of the mutant Mor-242 promoter by DmFkh (Fig. 5C). Mutation of Fkh-binding site 1 decreased the activity of the mutant promoter by ~20% compared to Mor-242 promoter, while mutation of both sites 1 and 2 or sites 1 and 3 together did not significantly decrease the activity further compared to mutation of site 2 or site 3 alone (Fig. 5C). These results suggest that Fkh-binding sites 2 and 3 in the MPAE region of moricin promoter play an equally important role in activation of moricin, while Fkh-binding site 1 may also contribute to activation of moricin. Similarly, mutation of the Fkh-binding site 1 or 2 alone in the Lyz-345 promoter significantly decreased the activity by more than 50%, and mutation of the Fkh-binding site 3 or 4 Dual-Luciferase ® Reporter Assay System as described in the Materials and Methods. Bars represent the mean of three independent measurements ± SEM. For transcription levels of Fox genes in S2 cells (A), or relative luciferase activity (B-D) among different promoters activated by one transcription factor (comparing striped bars or solid bars), identical letters are not significant difference (p > 0.05) while different letters indicate significant difference (p < 0.05). For the activity of the same promoter stimulated by different transcription factors (between MsFkh and DmFkh), the significance of difference was also determined by an unpaired t-test (*p < 0.05; **p < 0.01). MsCec, MsAtt-1, MsAtt-2, MsDef-1, MsDef-2 and MsDef-3 are M. sexta cecropin, attacin-1, attacin-2, defensin-1, defensin-2 and defensin-3 promoters (See Fig. S1 for the promoter sequences).
alone completely abolished DmFkh-activated Lyz-345 promoter activity (Fig. 5D), suggesting that Fkh-binding sites 3 and 4 play an important role in activation of lysozyme by Fkh factor whereas Fkh-binding sites 1 and 2 also contribute to regulation of lysozyme.

Interaction of M. sexta Fkh with Relish-RHD.
In the Mor-242 promoter, an NF-κB and a GATA-1 binding sites, which were both required for activation of moricin by immune signaling pathway 20 , were still present near the transcription initiation site (Fig. 5A), while in the Lyz-345 promoter, the NF-κB site was absent (the only NF-κB site in the 1.2-kb lysozyme promoter was between −1191 and −1182 bp, Fig. S1). In order to test whether NF-κB and Fkh factors regulate AMP genes independently or cooperatively, we first determined interaction of Fkh factor with NF-κB factors Dorsal and/or Relish (Rel2). Co-immunoprecipitation (Co-IP) assays showed that V5-tagged MsFkh co-precipitated with Flag-tagged M. sexta Rel2-RHD (Fig. 6B, D, lane 4), but did not co-precipitate with Flag-tagged Dorsal-RHD (Dl-RHD) (Fig. 6F, H, lane 4), suggesting that MsFkh interacts with MsRelish but did not interact with MsDorsal.  used Mor-242 and its seven Fkh mutant promoters (Fig. 5A) for the dual luciferase assays. For MsFkh-stimulated activity (Fig. 7, solid bars), mutations of individual Fkh-binding sites alone (site 1, 2 or 3) or combination of Fkh-binding sites (sites 1 and 2, 1 and 3, 2 and 3, or all three sites) significantly decreased the activity of Fkh mutant promoters, a result similar to that in Fig. 5C, indicating that Fkh-binding sites indeed play a role in activation of moricin promoter by Fkh factor. Co-expression of MsFkh and MsRel2-RHD had similar effect as overexpression of MsRel2-RHD alone (Fig. 7, comparing stripe bars and dotted bars across different promoters, as well as between the stripe and dotted bars in each promoter), suggesting that presence/absence of MsFkh and Fkh-binding sites does not have an impact on activation of moricin promoter by MsRel2-RHD, and the overall activity of these moricin promoters is due to MsRel2-RHD binding to NF-κB site in the moricin promoter.  Comparing the activity of the same promoter stimulated by different transcription factors (between solid and stripe bars, solid and dotted bars, as well as stripe and dotted bars for each promoter), the significance of difference was also determined by an unpaired t-test (*p < 0.05; **p < 0.01), and "n" indicates not significant.

Regulation of AMP genes independently by
Co-expression of MsFkh and MsRel2-RHD activated the activity of moricin promoters to either significantly higher or significantly lower level than that activated by overexpression of MsFkh alone, depending on the presence/absence of Fkh-binding sites (Fig. 7, comparing the solid bars with dotted bars in each promoter), further confirming that when MsFkh and MsRel2-RHD are expressed together, MsRel2-RHD tends to bind NF-κB sites regardless of MsFkh and Fkh-binding sites. Together, these results suggest that even though MsFkh can interact with MsRelish, formation of MsRelish homodimers may be predominant, and MsFkh and MsRelish regulate moricin promoter activation independently.

Discussion
Synthesis of AMPs is a major defense mechanism against infection in insects [50][51][52][53][54] , and expression of AMPs is regulated by the Toll and IMD pathways via activation of NF-κB transcription factors Dorsal, Dif and Relish 55,56 . Other proteins and factors that can modulate the Toll and/or IMD pathways have been identified. For example, a Zn finger homeodomain 1 (zfh1) transcription factor has been reported as a negative regulator of Drosophila IMD pathway downstream of, or parallel to Relish 57 ; Dorsal interacting protein 3 (Dip3) can bind to the RHD of Dorsal and Relish via its BESS domain, and it functions in both dorsoventral patterning and immune response 58 . In Drosophila, it has been reported that activation of AMPs can be achieved by the transcription factor FoxO independent of the immune signaling pathways 47 , and induction of two AMPs, diptericin and metchnikowin, after downregulation of TOR by rapamycin is regulated by the transcription factor forkhead (Fkh) 48 . FoxO is an important regulator of stress, metabolism and aging [59][60][61][62] , and a key transcription factor in the insulin signaling pathway 63,64 ; whereas Fkh is the founding member of the FoxO family. However, very little is known about regulation of AMPs in other insect species by the FoxO family transcription factors.
Insects can synthesize a variety of AMPs, some AMPs are common to most insects, whereas some other AMPs are found only in certain insect species 52 . For example, moricin and gloverin have been identified only in the lepidopteran insects. We have previously shown that M. sexta moricin is regulated by NF-κB factor Relish and GATA-1 factor 20, 65 , and it can also be activated by unidentified nuclear factor(s) that bind to the MPAE region of moricin promoter 20 . In this study, we identified three Fkh-binding sites in the MPAE region and demonstrated that Fkh-binding sites 2 and 3 played an equally important role in activation of moricin promoter by Fkh factor. We also identified four Fkh-binding sites in M. sexta lysozyme promoter and all four Fkh-binding sites are important for activation of lysozyme promoter by Fkh factor. Since the consensus sequence for FoxO is about 49 with a core sequence of AAACA, there are many predicted/potential Fkh-binding sites in promoters. Thus, we used truncated promoters first to narrow the length of promoters and then focused on potential active Fkh-binding sites in the moricin and lysozyme promoters with the core sequence of AAACA, as the core sequence may be crucial for binding of Fkh factor. Among the three and four Fkh-binding sites in the Mor-242 and Lyz-345 truncated promoters, all six Fkh-binding sites but the binding site 1 in the Mor-242 contains the AAACA core sequence (Figs. 1A and 5B). Indeed, mutation of Fkh-binding site 1 (with the core sequence of AAAGA) did not decrease the activity of Mor-242 promoter as significantly as did mutation of Fkh-binding site 2 or 3, suggesting that the core sequence of AAACA is crucial for activation of genes by Fkh factor. In addition to moricin and lysozyme promoters, Fkh factor also activated M. sexta defensin-1, defensin-3 and attacin-2 promoters. To our knowledge, this is the first report about activation of AMPs by Fkh factors in a lepidopteran insect. Activation of AMPs by Fkh factor under non-infectious conditions may be important for insects during molting and metamorphosis, since insects at these particular developmental stages are vulnerable to infection, and induced expression of AMPs by Fkh factor may protect insects from microbial infection. We observed that DmFkh stimulated the activity of some AMP promoters to a higher level than MsFkh did in S2 cells. This may be because MsFkh is not completely compatible in Drosophila S2 cells or the longer DmFkh (510 residues) may contain some other domains/motifs that can help Fkh to activate AMP gene promoters.
Different transcription factors may regulate gene expression independently or cooperatively. For M. sexta moricin, we previously found that both NF-κB and GATA-1 factors are required for activation of moricin promoter 20 . We also found that M. sexta Dorsal can interact with Relish (Rel2), and Dorsal/Relish heterodimers serve as negative regulators to prevent over-activation of M. sexta AMP genes 65 . We found that MsFkh interacted with MsRel2-RHD. In the Mor-242 promoter, which contains both NF-κB and Fkh binding sites, overexpression of MsFkh factor alone activated the promoter activity to a significantly higher level compared to overexpression of MsRel2-RHD alone, suggesting that Fkh factor plays an important role in activation of moricin under non-infectious conditions. In the mutant Mor-242 promoters in which the Fkh-binding sites were mutated, co-expression of MsFkh and MsRel2-RHD activated the activity of Fkh-binding site mutant promoters to a similar high level as that activated by MsRel2-RHD alone, indicating that MsFkh and MsRel2 regulate moricin activation independently. This result also suggests that although MsFkh can interact with MsRelish to form heterodimers, homodimers of MsRelish may be predominant. This may be because the distance between Fkh and NF-κB binding sites in the moricin promoter (~58 bp between NF-κB site and Fkh-binding site 3, Fig. 5A) is too far away for the two factors to form heterodimers. For NF-κB and GATA-1 sites in the moricin promoter, the two sites are separated by only 2 bp (Fig. 5A), and we showed that both NF-κB and GATA-1 sites are required for activation of moricin promoter 20 .
We also confirmed that Fkh mRNA was undetectable in Drosophila S2 cells, but the transcripts for FoxK (long and short isoforms), jumu and dFoxO were detected. Nuclear proteins from S2 cells did not bind to MPAE region, further supporting that DmFkh is not expressed (or was expressed at an undetectable level) in S2 cells. This information is important when performing promoter reporter assays for Fkh factor in S2 cells, as stress conditions may not activate the promoters in S2 cells due to lack of endogenous Fkh factor. In future, we will investigate activation of AMPs under non-infectious conditions via Fkh factor and maybe other members of the FoxO family and whether enhanced expression of AMPs by FoxO family members can protect insects from microbial infection during molting and/or metamorphosis.

Methods
Manduca sexta and insect cell lines. We purchased M. sexta eggs from Carolina Biological Supplies (Burlington, NC, USA) and reared larvae to fifth-instar on an artificial diet at 25 °C 66 for all the experiments. D. melanogaster Schneider S2 cells were obtained from American Type Culture Collection (ATCC), and Spodoptera frugiperda Sf9 cells were from Invitrogen (12552-014, Invitrogen). Cells were maintained at 27 °C in Insect Cell Culture Media (SH30610.02, Hyclone) supplemented with 1% penicillin-streptomycin solution (G6784, Sigma-Aldrich) and 10% heat-inactivated fetal bovine serum (#10082063, Invitrogen).

Electrophoretic mobility shift assay (EMSA). Cytosolic and nuclear proteins were isolated from S2
and Sf9 cells using the Nuclear Extraction Kit (2900, EMD Millipore) following the manufacturer's instructions. Briefly, S2 and Sf9 cells with 70-80% confluency were collected (2 × 10 8 cells) and homogenized in 500 μl of 1 × Cytoplasmic Lysis Buffer with 27-gauge needle and centrifuged at 8,000 × g for 20 min at 4 °C. The supernatants containing the cytosolic proteins were transferred to fresh tubes and stored at −80 °C for later use, whereas the pellets were resuspended in 100 μl of ice-cold Nuclear Extraction Buffer containing 0.5 mM DTT and 1/1000 Protease Inhibitor Cocktail. The nuclei were disrupted using 27-gauge needle, the mixture was incubated at 4 °C for 60 min with gentle agitation, and then centrifuged at 16,000 × g for 5 min at 4 °C. These supernatants containing nuclear proteins were removed to fresh tubes and stored at −80 °C for later use.

Analyses of transcripts of M. sexta forkhead (MsFkh) in larvae and D. melanogaster
To determine induced expression of MsFkh transcript in M. sexta larvae, day 2 fifth-instar naïve larvae were injected with heat-killed Staphylococcus aureus, Bacillus subtilis, Escherichia coli strain XL1-blue, Serratia marcescens (each at 5 × 10 7 cells/larva), or Saccharomyces cerevisiae (10 7 cells/larva), or with water as a control. Hemocytes, fat body and midgut were collected separately at 24 h post-injection for total RNA extraction and cDNA preparation as described previously 67 . Total RNA and cDNA were also prepared from Drosophila S2 cells.
Briefly, S2 cells (5 × 10 6 cells) were collected in 1 ml of TRIzol ® Reagent (T9424, Sigma-Aldrich) and homogenized using hand held pestle and mixer (Argos Technologies, Elgin, IL). Then, 200 μl of chloroform were added, and the mixture was vortexed and centrifuged at 12000 × g at 4 °C for 15 min. The top aqueous phase (200 μl) was transferred to fresh Eppendorf tubes and isopropanol (500 μl) was added. RNA was precipitated by centrifugation at 12,000 × g at 4 °C for 10 min. The RNA pellets were washed with chilled 70% ethanol, air dried and re-suspended in 50 μl of nuclease free water and stored at −80 °C for later use. used as an internal standard to normalize the amount of RNA template. Real-time PCR program was 50 °C for 2 min, 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s, 60 °C for 1 min and the dissociation curve analysis. Data from three replicas of each sample were analyzed by the ABI 7500 SDS software (Applied Biosystems) using a comparative method (2 −∆∆CT ) 68,69 . These experiments were repeated with three different biological samples.

Construction of recombinant MsFkh and DmFkh pAC5.1/V5-His A expression vectors. Recombinant MsRelish-RHD (MsRel2-RHD) and
MsDorsal-RHD (MsDl-RHD) with a Flag-tag in pAC5.1/V5-His A expression vectors were already constructed as described previously 65 . To construct V5-tagged Fkh into pAC5.1/V5-His A expression vector, cDNA fragments encoding MsFkh (residues 1-355) and DmFkh (residues 1-510) were amplified by PCR using forward and reverse primers (Table S1) Construction of luciferase reporter plasmids. The moricin and lysozyme truncated promoters were constructed as described previously 20,65 . To construct Fkh-binding site mutation promoters, site-directed mutagenesis was performed using the truncated M. sexta moricin-242 (242 bp) and lysozyme-345 (345 bp) promoters as templates. Primers with specific mutation sites were designed for each mutated promoter and listed in Table S1. There are three and four predicted Fkh-binding sites in moricin-242 and lysozyme-345 promoters, respectively. Primers Lyz-D3-Fkh-1, Lyz-D3-Fkh-2, Lyz-D3-Fkh-3 and Lyz-D3-Fkh-4 (Table S1) were used to generate mutations of Fkh-binding site 1, 2, 3 and 4, respectively, in lysozyme-345 promoter. Whereas primers MPAE-Fkh-1, MPAE-Fkh-2 and MPAE-Fkh-3 were used to generate mutations of Fkh-binding site 1, 2 and 3, respectively, in moricin-242 promoter. To pre-screen positive colonies prior to DNA sequencing, restriction enzyme cleavage sites of EcoR I, Nde I and Bam HI were engineered in the mutant Fkh-binding site 1, 2 and 3 of the moricin-242 promoter, respectively. To generate Mor mut-1&2 , Mor mut-1&3 , Mor mut-2&3 and Mor mut-1,2&3 mutant promoters, site-directed mutagenesis was performed using the mutant promoter as the template with the second pairs of primers. For example, to generate Mor mut-1&2 promoter, Mor mut-1 was used as the template with MPAE-Fkh-2 primers, and to obtain Mor mut-1,2&3 promoter, Mor mut-1&2 was used as the template with MPAE-Fkh-3 primers. PCR program was 95 °C for 3 min, and then 17 cycles of 95 °C for 1 min, 55 °C for 2 min, 68 °C for 15 min, followed by a final extension of 68 °C for 30 min. The PCR products were recovered, digested with Dpn I, and then transformed into competent E. coli XL1 Blue cells. The mutant reporter plasmids were then purified and sequenced by an Applied Biosystems 3730 DNA Analyzer in the DNA Sequencing and Genotyping Facility at University of Missouri -Kansas City, and used for transient transfection in S2 cells.
Dual-luciferase reporter assays. For DNA transfection, S2 cells were placed overnight to 70% confluence prior to transfection in serum-free medium (SH30278.01, Hyclone). GenCarrier-1 TM transfection reagent (#31-00110, Epoch Biolabs) was used for transient transfection according to the manufacturer's instructions. After overnight transfection, S2 cells were centrifuged and resuspended in complete growth medium to induce protein expression for 48 h. Protein expression in cell culture media and cell extracts were analyzed by immunoblotting.
Co-immunoprecipitation (Co-IP) assays were performed using cell extracts from S2 cells overexpressing MsFkh, MsRel2-RHD and MsDorsal-RHD proteins. Cell extracts were mixed and Co-IP was performed as described previously 12 . Proteins immunoprecipitated with anti-Flag M2 or anti-V5 primary antibody were captured by protein G Sepharose pre-swollen beads (#17-0618-01, GE Healthcare). Captured proteins were eluted with 30 μl of sample buffer mixed with 0.1% bromphenol blue, heated to 95 °C for 3 minutes, centrifuged at 12,000 × g for 1 min, and the supernatants were loaded onto SDS-PAGE for immunoblotting analysis as described previously 12 . Data Analysis. All the experiments were performed in 3-4 replicates and repeated with three independent biological samples. The means of a typical set of data were used to prepare the figures by GraphPad Prism (GraphPad, San Diego, CA). Statistical significance was calculated by one way ANOVA followed by a Tukey's multiple comparison tests using GraphPad Prism for comparisons of MsFkh mRNA in different tissues, or MsFkh mRNA in hemocytes, fat body or midgut by different treatments (Fig. 3), expression levels of D. melanogaster Fox genes in S2 cells (Fig. 4A), or relative luciferase activity across different promoters by overexpression of DmFkh or MsFkh (Figs. 4, 5 and 7), and identical letters are not significant difference (p > 0.05) while different letters indicate significant difference (p < 0.05). The significance of difference was also determined by an unpaired t-test with the GraphpadInStat software (*p < 0.05; **p < 0.01) to compare the activity of a promoter stimulated by different transcription factors.