Toll family members bind multiple Spätzle proteins and activate antimicrobial peptide gene expression in Drosophila

The Toll signaling pathway in Drosophila melanogaster regulates several immune-related functions, including the expression of antimicrobial peptide (AMP) genes. The canonical Toll receptor (Toll-1) is activated by the cytokine Spätzle (Spz-1), but Drosophila encodes eight other Toll genes and five other Spz genes whose interactions with one another and associated functions are less well-understood. Here, we conducted in vitro assays in the Drosophila S2 cell line with the Toll/interleukin-1 receptor (TIR) homology domains of each Toll family member to determine whether they can activate a known target of Toll-1, the promoter of the antifungal peptide gene drosomycin. All TIR family members activated the drosomycin promoter, with Toll-1 and Toll-7 TIRs producing the highest activation. We found that the Toll-1 and Toll-7 ectodomains bind Spz-1, -2, and -5, and also vesicular stomatitis virus (VSV) virions, and that Spz-1, -2, -5, and VSV all activated the promoters of drosomycin and several other AMP genes in S2 cells expressing full-length Toll-1 or Toll-7. In vivo experiments indicated that Toll-1 and Toll-7 mutants could be systemically infected with two bacterial species (Enterococcus faecalis and Pseudomonas aeruginosa), the opportunistic fungal pathogen Candida albicans, and VSV with different survival times in adult females and males compared with WT fly survival. Our results suggest that all Toll family members can activate several AMP genes. Our results further indicate that Toll-1 and Toll-7 bind multiple Spz proteins and also VSV, but they differentially affect adult survival after systemic infection, potentially because of sex-specific differences in Toll-1 and Toll-7 expression.

The Toll signaling pathway in Drosophila melanogaster regulates several immune-related functions, including the expression of antimicrobial peptide (AMP) genes. The canonical Toll receptor (Toll-1) is activated by the cytokine Spätzle (Spz-1), but Drosophila encodes eight other Toll genes and five other Spz genes whose interactions with one another and associated functions are less well-understood. Here, we conducted in vitro assays in the Drosophila S2 cell line with the Toll/interleukin-1 receptor (TIR) homology domains of each Toll family member to determine whether they can activate a known target of Toll-1, the promoter of the antifungal peptide gene drosomycin. All TIR family members activated the drosomycin promoter, with Toll-1 and Toll-7 TIRs producing the highest activation. We found that the Toll-1 and Toll-7 ectodomains bind Spz-1, -2, and -5, and also vesicular stomatitis virus (VSV) virions, and that Spz-1, -2, -5, and VSV all activated the promoters of drosomycin and several other AMP genes in S2 cells expressing full-length Toll-1 or Toll-7. In vivo experiments indicated that Toll-1 and Toll-7 mutants could be systemically infected with two bacterial species (Enterococcus faecalis and Pseudomonas aeruginosa), the opportunistic fungal pathogen Candida albicans, and VSV with different survival times in adult females and males compared with WT fly survival. Our results suggest that all Toll family members can activate several AMP genes. Our results further indicate that Toll-1 and Toll-7 bind multiple Spz proteins and also VSV, but they differentially affect adult survival after systemic infection, potentially because of sex-specific differences in Toll-1 and Toll-7 expression. The innate immune system of Drosophila melanogaster and other insects provides defense against infection by pathogenic viruses, bacteria, fungi, and parasites (1). One key defense response is the production of antimicrobial peptides (AMPs), 3 whose expression is primarily regulated by the Toll and IMD (immune deficiency) pathways (1)(2)(3). In Drosophila, the Toll pathway is activated when a cytokine named Spätzle (Spz-1) binds the Toll receptor (Toll-1) (4,5). Spz-1 is expressed and secreted into circulation as a zymogen (pro-Spz-1), which consists of an N-terminal pro-domain and a C-terminal cystine knot. Toll-1 structurally consists of an ectodomain containing leucine-rich repeats (LRRs), a single-pass transmembrane domain, and a cytosolic Toll/interleukin-1 receptor (TIR) homology domain (6,7). Systemic infection by bacteria or fungi stimulates the Spätzle-processing enzyme (SPE) to cleave pro-Spz-1 into Spz-1, which binds Toll-1 that is expressed on the surface of the fat body and other select cell types (5, 8 -12). Spz-1 binding promotes Toll-1 dimerization/oligomerization, which stimulates downstream signaling through the adaptor protein MyD88 (dMyD88) and kinases named Tube and Pelle. This results in phosphorylation and degradation of the IB inhibitor Cactus (13), which enables the NF-B transcription factors Dif and Dorsal to translocate to the nucleus and up-regulate AMP genes through promoter binding (1,14,15).
Vertebrates encode multiple Toll-like receptors (TLRs) that exhibit similar architecture to Drosophila Toll-1 and regulate several immune and nonimmune functions (16 -19). Similarities in downstream signaling components also support shared ancestry between the Toll and TLR pathways. However, vertebrate TLRs do not bind cytokines like Spz-1 but instead function as pattern recognition receptors (PRRs) that bind pathogen-associated ligands such as bacterial lipopolysaccharide, peptidoglycan, teichoic acid, flagella, CpG DNA (17, 20 -22), viral single-stranded RNA, and viral dsRNA (20 -21).
Comparative genomic data indicate that insects also encode multiple Toll genes. Drosophila encodes eight other Toll family members (Toll-2 to Toll-9) in addition to Toll-1 with some evidence supporting defense functions for Toll-2 (18 wheeler, 18W), Toll-5 (Tehao), Toll-8 (Tollo), and Toll-9 (23-28). Toll-6 and Toll-7 function as neurotrophin receptors (29), whereas Toll-7 is also reported to recognize vesicular stomatitis virus (VSV) and induce antiviral autophagy (30,31). In contrast, other results indicate that autophagy plays a minor role in hemocyte-mediated defense against VSV and does not depend on Toll-7 (32). Drosophila encodes five other Spz genes (Spz-2 to Spz-6) in addition to Spz-1, but it remains unknown whether any of these other family members bind to Toll-1 or other Toll proteins. It is also unclear whether AMP genes activated by Toll-1 are also activated by other Toll family members.
In this study, we assessed whether all or only some Drosophila Toll family members activate the drosomycin promoter, which is a known target for the canonical Toll-1 pathway (1,14,15). Focusing on Toll-1 and Toll-7, we also assessed binding to Spz family members and VSV and whether each similarly or differentially affects adults after infection by different microbes. Our results indicated that the TIR domains for all Toll family members activated the drosomycin promoter. We further determined that Toll-1 and Toll-7 bind multiple Spz proteins and VSV while differentially affecting adult female and male survival after systemic infection.

TIR domains of several Drosophila Toll family members activate the drosomycin promoter in S2 cells
Prior studies indicate that binding of Spz-1 to Toll-1 activates the drosomycin promoter as well as the promoters for other select AMP genes (1,14,15). To determine whether other Toll family members can also activate the Drosophila drosomycin promoter, we conducted dual-luciferase assays in S2 cells that were co-transfected with a pGL3B-drosomycin reporter plus pMT/BiP/V5-His that inducibly expressed the TIR domain for each Drosophila Toll family member as well as Toll-1 from the moth Manduca sexta (33). We also assessed whether any of these TIRs activated the Drosophila diptericin promoter, because this AMP is not activated by Toll-1 signaling but is activated by the IMD pathway (1). We first confirmed by immunoblotting that each TIR was expressed (Fig. 1A). We then conducted dual-luciferase assays, which indicated that each TIR significantly activated the drosomycin reporter (7-54-fold) above an empty plasmid/pGL3B control (Fig. 1B). However, the strongest responses were elicited by the TIRs from Toll-1 (54-fold), Toll-7 (39-fold), and M. sexta Toll-1 (48fold) (Fig. 1B). In contrast, no TIR significantly activated the diptericin promoter when compared with the empty vector control (Fig. 1B). We thus concluded that the TIR domains for each Drosophila Toll family member can activate the drosomycin promoter not the diptericin promoter.

Ectodomains of Toll-1 and Toll-7 interact with multiple Spz proteins
We next considered whether Toll family members interact with only one or multiple Spz proteins. For these and subsequent experiments, we focused on comparing Drosophila Toll-1 to Toll-7 because in the preceding assays these two family members most strongly activated the drosomycin promoter. Previous co-immunoprecipitation (co-IP) assays indicated that M. sexta Toll-1 binds the cystine knot domain of M. sexta Spz-1 but not the full-length pro-Spz-1 (33). We therefore used previously developed approaches (29) to first predict the cystine knot domains for the six Drosophila Spz family members (Fig.  S1A). Alignments further indicated that Spz-1, Spz-2, and Spz-5 are more closely related to one another than to other family members, with Spz-1 sharing 63 and 71% similarity with Spz-2 and Spz-5, respectively, and Spz-2 sharing 62% similarity with Spz-5 (Fig. S1B). We then cloned each cystine knot into pMT/BiP and expressed them in S2 cells as recombinant proteins with FLAG epitope tags, while cloning either the ectodomains (Toll-1 ecto and Toll-7 ecto ) or full-length Drosophila Toll-1 and Toll-7 into pMT/BiP/V5-His for expression as recombinant proteins with V5 epitope tags. Immunoblotting assays showed that each Spz cystine knot was present in both S2 cell lysates and the medium that indicated each was secreted (Fig. S2). Toll-1 ecto and Toll-7 ecto were also detected in both lysates and medium, whereas full-length Toll-1 and Toll-7 were, as expected, readily detectable in S2 cells but were not secreted due to the presence of transmembrane domains (Fig. S2). in extracts prepared from S2 cells co-transfected with pGL3B-drosomycin, pGL3B-diptericin, or pGL3B (empty vector) plus plasmids expressing each TIR domain. Three biological replicates were generated for each treatment. For the drosomycin promoter, bars with different letters indicate treatments that significantly differed from one another (p Ͻ 0.05; one-way ANOVA followed by a post hoc Tukey HSD test). No significant differences were detected between treatments for the diptericin reporter or empty vector.

Multiple Toll-1-Spz and Toll-7-Spz pairs activate the drosomycin promoter in S2 cells
Given evidence that Toll-1 ecto and Toll-7 ecto bound multiple Spz family members, we conducted dual-luciferase assays to determine whether the same Spz family members activate the drosomycin promoter in S2 cells expressing full-length Toll-1 or Toll-7 plus our drosomycin-luciferase reporter. Results indicated that the cystine knots for Spz-1, -2, and -5 activated the drosomycin promoter in S2 cells expressing full-length Toll-1 492-, 188-, and 122-fold, whereas other Spz family members had no significant effect on promoter activation (Fig. 3A). Spz-1, -2, and -5 also activated the drosomycin promoter in S2 cells expressing full-length Toll-7 98-, 87-, and 83-fold (Fig. 3B). In contrast, adding the cystine knot for Spz-6 and other family members either only weakly activated or had no effect on the drosomycin promoter in cells expressing full-length Toll-7 (Fig.  3B). Control assays where 1) no Spz was added to cells expressing full-length Toll-1, Toll-2, and Toll-7, and 2) cystine knot proteins for Spz-1, Spz-2, or Spz-5 were added to cells lacking full-length recombinant Toll-1 or Toll-7 or cells co-expressing Toll-2 and the cystine knot for Spz-1 also resulted in either little (3.5-14-fold) or no activation of the drosomycin promoter (Fig.  3C). Thus, with the exception of Spz-6, Spz family members that bound Toll-1 ecto and Toll-7 ecto also strongly activated the drosomycin promoter. In contrast, weak activation of the drosomycin promoter in select control assays could reflect interactions with endogenous ligands or receptors.

Ectodomains of Toll-1 and Toll-7 interact with VSV, which also activates AMP gene promoters
As noted previously, Toll-7 has previously been reported to interact with VSV, which suggested that Toll-7 could function as a PRR (30). Given that Toll-7 and Toll-1 bound largely the same Spz family members in our co-IP assays, we asked whether Toll-1 also binds VSV. The co-IP assays where we added VSV-GFP virions to culture medium containing Toll-1 ecto or Toll-7 ecto resulted in an anti-V5 antibody plus protein G beads co-immunoprecipitating VSV virions, whereas control assays in which VSV virions, culture medium containing Toll-1 ecto or Toll-7 ecto plus protein G beads alone resulted in no pulldown of VSV virions, Toll-1 ecto or Toll-7 ecto (Fig. 4A). Dualluciferase assays further showed that adding VSV-GFP virions to S2 cells co-expressing full-length Toll-1 or Toll-7 and reporter plasmids significantly activated the drosomycin promoter as well as the promoters for two other AMP genes (attacin and metchnikowin) that are known to be activated by Toll signaling (Fig. 4, B and C).

Toll-1 and Toll-7 loss-of-function mutants differentially affect male and female survival after infection
Real-time PCR assays indicated that transcript abundance for Toll-1 was significantly higher in 5-day-old WT (w 1118 ) adult females than adult males, whereas transcript abundance for Toll-7 was higher in adult males than adult females (Fig.  S3A). This finding suggested to us that Toll-1 and Toll-7 could with a V5 epitope tag were combined with lysates from cells expressing the cystine knot domain for each Spz with a FLAG epitope tag. A, adding an anti-V5 antibody plus protein G beads to the combined lysates immunoprecipitated Toll-1 ecto in all treatments that were detected on immunoblots using anti-V5 (upper blot), while only co-immunoprecipitating the cystine knot domains for Spz-1, -2, and -5, which were detected on immunoblots using anti-FLAG (lower blot). B, reciprocal experiment of adding anti-FLAG plus protein G beads immunoprecipitated the cystine knot domains for each Spz (upper blot), while only co-immunoprecipitating Toll-1 ecto in the Spz-1, -2, and -5 treatments (lower blot). C and D, outcomes of co-IP experiments where lysates from cells expressing Toll-7 ecto were combined with lysates from cells expressing the cystine knot domain of each Spz. C, anti-V5 plus protein G beads immunoprecipitated Toll-7 ecto in all treatments (upper blot) while co-immunoprecipitating the cystine knot domains for Spz-1, -2, -5, and -6 (lower blot). D, anti-FLAG plus protein G beads immunoprecipitated the cystine knot domain for each Spz (upper blot) while co-immunoprecipitating Toll-7 ecto in the Spz-1, -2, -5, and -6 treatments (lower blot).

Toll family member functions
differentially affect defense in females versus males after systemic infection by different microorganisms. To assess this possibility, we compared the survival of adult female and male loss-of-function Toll-1 (Tlr 632 /Tl I-RXA and Tl I-RXA ) and Toll-7 mutants (Toll-7 g1-5 /Cyo and Df/Toll-7 g1-5 ) to WT (w 1118 ) flies after infection by two species of bacteria (Enterococcus faecalis or Pseudomonas aeruginosa), the fungus Candida albicans, or VSV-GFP. For females, both Toll-1 mutants exhibited lower survival than WT flies after infection by E. faecalis, C. albicans, and VSV (Fig. 5, A-D). Both Toll-7 mutants also exhibited lower survival when compared with WT flies after infection by VSV but did not differ from WT flies after infection by E. faecalis, P. aeruginosa, and C. albicans (Fig. 5, A-D). For males, we observed the reverse trend with one or both Toll-7 mutants exhibiting significantly lower survival rates than WT flies after infection by each microbe, whereas Toll-1 mutants exhibited lower survival rates in response to infection by E. faecalis and C. albicans but did not differ from WT flies in response to infection by P. aeruginosa (Fig. 5, A-D). Overall, loss of Toll-1 adversely affected adult females more than adult males after . Spz-1, -2, and -5 partner with Toll-1 and Toll-7 to activate the drosomycin promoter. A, mean relative luciferase activity Ϯ S.E. in extracts prepared from S2 cells that were co-transfected with pGL3B-drosomycin, pMT/BiP/V5-His A-Toll-1, and pMT/BiP-Spz-1, -2, -3, -4, -5 or -6. B, mean relative luciferase activity Ϯ S.E. in extracts prepared from S2 cells that were co-transfected with pGL3B-drosomycin, pMT/BiP/V5-His A-Toll-7, and pMT/BiP-Spz-1, -2, -3, -4, -5, or -6. C, mean relative luciferase activity Ϯ S.E. in extracts prepared from S2 cells that were transfected with pGL3B-drosomycin, and pMT/BiP/V5-His-Toll-1, -2, or -7, pMT/BiP-Spz-1, -2, or -5, or pMT/BiP/V5-His A-Toll-2 plus pMT/BiP-Spz-1. Four biological replicates were generated for each treatment. For each graph, different letters above bars indicate treatments significantly differed from one another (p Ͻ 0.05; one-way ANOVA followed by a post hoc HSD test).

Figure 4. VSV binds Toll-1 and Toll-7 and activates multiple AMP gene promoters.
A, outcomes of co-IP experiments where VSV virions were added to conditioned medium from S2 cells that were transfected with pMT/BiP/V5-His-Toll-1 ecto or -Toll-7 ecto . Adding anti-V5 antibody plus protein G beads immunoprecipitated Toll-1 ecto or Toll-7 ecto that was detected on immunoblots using anti-V5 (upper left blot) and VSV virions that were detected using anti-VSV-G (lower left blot). Adding protein G beads alone (negative control) resulted in no immunoprecipitation of Toll-1 ecto , Toll-7 ecto , or VSV (upper and lower right blots). B, mean relative luciferase activity Ϯ S.E. in extracts prepared from VSV-infected and -noninfected S2 cells expressing full-length Toll-1 plus pGL3B-attacin (att), -drosomycin (drs), or -metchnikowin (mtk). C, mean relative luciferase activity Ϯ S.E. in extracts prepared from VSV-infected and -noninfected S2 cell line expressing the full-length Toll-7 plus pGL3B-attacin, -drosomycin, or -metchnikowin. Three biological replicates were generated for each treatment. An asterisk indicates that relative luciferase activity significantly differed between VSV-infected and -noninfected cells (t test; p Ͻ 0.001).

Toll family member functions
Toll family member functions infection by different microbes, whereas the loss of Toll-7 adversely affected adult males more than adult females.

Discussion
Toll-1 and Spz-1 were originally identified in D. melanogaster as components of the Toll signaling pathway that regulates dorso-ventral patterning during embryogenesis (34). Thereafter, Toll-1, Spz-1, and Toll signaling were further implicated in regulating AMP genes and other immune-related functions (1-3). Select other Toll family members (Toll-2, -5, -8, and -9) have also been reported to affect immune defenses (24 -28, 35, 36). However, no standardized platform has previously been developed that could be used to assess AMP promoter activation and ligand binding by different Toll family members.
Our approach of co-expressing the TIR domain for each Toll family member with a drosomycin promoter construct suggests all Drosophila Toll family members can activate this AMP above background levels, although activation is most strongly elicited by Toll-1 and Toll-7. Because no Toll family member activated the diptericin promoter further affirms that diptericin is primarily if not exclusively regulated by the IMD pathway. Tauzig et al. (23) in contrast took a different approach by coexpressing the TIR domain from Toll-2, -5, -6, -7, and -8 with a truncated form of Toll-1 (Toll ⌬LRR ) to produce chimeric receptors in S2 cells. However, the only chimeric construct that activated the drosomycin promoter was Toll-5 (23). Given evidence that receptor dimerization/oligomerization is required for the downstream signaling events that precede AMP gene promoter activation (1, 2), we speculate that expressing only the TIR for each Toll family member potentially enabled dimers/ oligomers to more readily form, whereas the approach of producing chimeric Tolls may have resulted in only Toll-5 forming dimers/oligomers. However, our results also indicate the Toll-5 TIR more weakly activated the drosomycin promoter than the chimeric Toll-5 construct produced by Tauzig et al. (23), which could suggest the latter forms a more stable dimers/oligomer.
Focusing on Toll-1 and Toll-7, our co-IP assays indicate the ectodomains of each bind Spz-1, -2, and -5, but not other family members, while also suggesting that binding of Spz-1, -2, and -5 to full-length Toll-1 or Toll-7 results in activation of the drosomycin promoter. Our results also indicate that Toll-7 binds Spz-6, but surprisingly, adding the cystine knot for Spz-6 to S2 cells expressing full-length Toll-7 did not activate the drosomycin promoter. Thus, three Spz family members (Spz-1, -2, and -5) exhibit features consistent with functioning as cytokines that bind to Toll-1 and Toll-7, which activates downstream Toll signaling and the drosomycin promoter. In contrast, our results do not reveal why Spz-6 does not exhibit similar activity or what the function of this family member might be.
As noted previously, Toll-6 and Toll-7 function as neurotrophin receptors in the central nervous system with Spz-2 and Spz-5 being candidate ligands for both (29). However, Toll-7 is also reported to recognize VSV, and function as a PRR that stimulates antiviral autophagy (30,31). Other results raise questions about both the importance of autophagy as an antiviral immune defense and the requirement for Toll-7 (32). Nonetheless, our results corroborate that Toll-7 binds VSV while showing that Toll-1 also binds this virus. Furthermore, infection of S2 cells expressing full-length Toll-1 and Toll-7 with VSV activated the expression of several AMP promoter constructs. Our results thus overall suggest Toll-1 and Toll-7 bind multiple Spz proteins and VSV and that both types of ligands can activate downstream signaling that leads to AMP gene expression. Finally, our finding that adult males and females differentially express Toll-1 and Toll-7 combined with differential sensitivity to systemic infection by VSV and other microbes suggests gender may affect the relative importance of different Toll family members in immune defense.
That Toll family members bind multiple Spz proteins raises important questions for future study regarding the inputs that regulate processing of different Spz zymogens and whether SPE or other unidentified serine proteases are involved. The function of Spz-6 is a second question of interest as is the relative importance of Toll family members binding different Spz family members versus pathogen-associated molecular pattern molecules on microbes like VSV in regulating different immune defense responses.

Fly stocks and S2 cells
WT w 1118 flies were obtained from the laboratory of Dr. Leonard Dobens (School of Biological Sciences, University of Missouri-Kansas City, MO). The Toll-7 g1-5 /Cyo mutant line was a gift from Dr. Yashimasa Yagi (Division of Biological Science, Nagoya University, Nagoya, Japan), and the Toll-7 g1-5 mutant line was created by homologous recombination of an ends-in knockout system followed by hs-ICreI treatment to generate a Toll-7 knockout line with a point mutation (37). The Toll-7 g1-5 line was balanced over Cyo to obtain the Toll-7 g1-5 / Cyo mutant line, and heterozygotes were screened based on the existence of curled wings. The Tlr 632 /Tl I-RXA and Tl I-RXA / TM6B (Tl I-RXA ) Toll (also called Toll-1) mutant lines were obtained from the laboratory of Dr. Kontoyiannis (Department of Infectious Diseases, University of Texas M. D. Anderson Cancer Center, Houston) (38). Tlr 632 /Tl I-RXA flies were generated by crossing Tlr 632 /TM6B and Tl I-RXA /TM6B Toll-deficient flies. Tlr 632 is a thermosensitive loss-of-function allele with a strong phenotype at 29°C; thus, these flies were maintained at 29°C during infection. Both the Tlr 632 and Tl I-RXA mutant lines were balanced over TM6B and were recognized by multiple hair-type bristles in the upper lateral thorax/torso. The Df(2R)BSC22/SM6a line (stock no. 6647) was purchased from the Bloomington Stock Center (Indiana); in this line, the

Toll family member functions
56D7-56F12 chromosome segment was deleted by exploiting hybrid element insertion and resolution, and this line was later balanced over SM6a to obtain flies that can be recognized by curly wings. We generated Df(2R)BSC22/Toll-7 g1-5 (Df/Toll-7 g1-5 ) flies by crossing Toll-7 g1-5 /CyO and Df(2R)BSC22/SM6a flies, which uncover the Toll-7 locus to obtain Toll-7 mutants that can be screened by the presence of curly wings. All the flies were cultured on a cornmeal diet (31) and transferred to fresh food at least 24 h prior to injection/infection. The Drosophila S2 cell line was obtained from the American Type Culture Collection (ATCC) and maintained in serum-free SFX medium (Hyclone) that was supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin, and 1% Gibco L-glutamine (25030081, Thermo Fisher Scientific) (the complete growth medium) by passaging weekly in 15-cm culture flasks (Corning).

Gene cloning
For this study, we cloned full-length Toll-1 from cDNA using total RNA from Drosophila adult females as the template. Fulllength Toll-7 without introns was previously cloned into pAC5.1-A (37). All nine Drosophila Toll TIR domains and the TIR from M. sexta Toll-1 (33), the ectodomains of Toll-1 and Toll-7, and full-length Toll-1 and Toll-7 were amplified by PCR using the forward and reverse primers listed in Table S1 and cloned into the pMT/BiP/V5-His A vector (V413020, Invitrogen) for expression of the recombinant proteins with a V5 epitope tag at the C terminus. Active Spz-1 to Spz-6 were generated by PCR amplifying the domains underlined in Fig. S1A using sequence-specific primers (Table S1) and RNA from Drosophila adult females as template followed by cloning into a modified pMT/BiP A vector (33) for expression of the recombinant Spz proteins with a FLAG-tag at the N terminus. The PCRs were performed with the following conditions: 94°C for 3 min, 35 cycles of 94°C for 30 s, Tm-5°C for 30 s, 72°C for 30 s to 4 min, and final extension at 72°C for 10 min. The PCR products were recovered using an Agarose Gel Electrophoresis-Wizard SV gel and PCR clean-up system (A9285, Promega) and then subcloned into the T-Easy vector (A1360, Promega). Recombinant plasmid DNAs were purified using a PureYield TM plasmid miniprep system (A1222, Promega) according to the manufacturer's instructions and digested with respective restriction enzymes, and DNA fragments were recovered and inserted into the pMT/BiP/V5-His A or modified pMT/BiP A vector using T4 DNA ligase (M0202L, New England Biolabs). The recombinant expression plasmids were then purified and sequenced in the sequencing facility at University of Missouri-Columbia for further experiments.

Vesicular stomatitis virus stock culture
VSV-expressing green fluorescent protein (GFP) inserted between the 3Ј leader and the N gene (39) was obtained from the laboratory of Dr. Sean Whelan (Harvard Medical School, Boston) (40). VSV-GFP was cultured and maintained in HEK293 cells in DMEM supplemented with 10% heat-inactivated fetal bovine serum (10082063, Invitrogen) and 1% penicillin/streptomycin solution (G6784, Sigma). The viral titer was measured by plaque assay using HEK293 cells (41). For infec-tion assays with Drosophila S2 cells, 10,000 pfu/ml VSV-GFP was used, and for the infection assays with adult flies, 10,000 pfu of VSV-GFP (in 50 nl of PBS) were injected into each fly.

Infection assays
Drosophila adult males and females (5-7 days of age) in a batch of 20 -30 flies were infected with Gram-positive E. faecalis V583 (a gift from Dr. Michael Gilmore, Harvard Medical School, Boston), Gram-negative P. aeruginosa PA-14 (a gift from Dr. Kalai Mathee, Florida International University), C. albicans (a gift from Dr. Theodore White, School of Biological Sciences at the University of Missouri-Kansas City, MO), or VSV-GFP. Briefly, overnight bacterial and fungal cultures were diluted to A 600 ϭ 0.2 and A 600 ϭ 0.5, respectively, washed with phosphate-buffered saline (PBS, pH 7.4), and resuspended in PBS for injection. Flies were anesthetized with CO 2 (for no longer than 15 min at a time), and 50 nl of diluted E. faecalis V583, P. aeruginosa PA-14, C. albicans, or VSV-GFP (10,000 pfu/50 nl) was injected into each fly at the left intrathoracic region using a Drummond nanoinjector and pulled glass capillary needles. These flies were maintained in clean bottles with fresh cornmeal diet, and the diet was changed every day throughout the course of the experiment. The flies that died within 3 h of injection were excluded from the study on the presumption of injury. Flies injected with bacteria or C. albicans were monitored hourly, while flies injected with VSV-GFP were monitored daily by recording the number of dead males and females.

Transient transfection and establishment of stable S2 cell lines
Transient transfection experiments and establishment of S2 cell lines stably expressing full-length Toll-1 and Toll-7 followed previously established protocols (33). Briefly, cells were seeded overnight in complete growth medium (see above), washed with serum-free medium, and transiently transfected using GenCarrier-2 TM transfection reagent according to the manufacturer's instructions. The DES-inducible/secreted kit with pCoBlast (K5130-01, Invitrogen) was used to generate stable S2 cell lines. Lines stably expressing recombinant Toll-1 and Toll-7 were selected by co-transfecting pCoBlast (Invitrogen) with the pMT/BiP/V5-His A vector, which uses the metallothionein promoter for inducible expression of the gene of interest following the addition of copper sulfate. Forty eight hours after transfection, S2 cells were centrifuged and resuspended in the complete growth medium containing 25 g/ml blasticidin S hydrochloride (no. 15205, Sigma). Resistant colonies appeared 1 week later.

Dual-luciferase assays
Dual-luciferase assays were performed as described previously (42). S2 cells were plated in 24-well culture plates (3 ϫ 10 5 Toll family member functions cells/well) overnight in the complete growth medium, washed with serum-free medium, and then transiently co-transfected with recombinant pMT/BiP/V5-His A expression plasmid (500 ng), pGL3B, pGL3B-drosomycin or pGL3B-diptericin firefly luciferase reporter plasmid (250 ng), and Renilla luciferase reporter plasmid (25 ng) (as an internal standard) (pRL-TK, Promega) with GenCarrier-2 TM . After overnight transfection, serum-free medium was replaced with the complete growth medium containing copper sulfate (to a final concentration of 500 M) for protein expression, and 36 h after protein expression, the firefly luciferase and Renilla luciferase activities were measured using the Dual-Luciferase Reporter Assay System (E1980, Promega) with a GloMax multimicroplate luminometer (Promega). The relative luciferase activity (RLA) was obtained as the ratio of firefly luciferase activity to Renilla luciferase activity. The RLA obtained from S2 cells co-transfected with empty pMT/BiP/V5-His A and pGL3B (empty reporter vector) plasmids was used as the calibrator. These experiments were repeated at least three times (three independent biological samples or three independent cell cultures), and a representative set of data was used to prepare the figures.

co-IP assays
S2 cell lysates (300 l, approximately equivalent to 10 6 cells) or equivalent cell culture media containing recombinant proteins were precleared with protein G-Sepharose (50% slurry, 17-0618-01, GE Healthcare) prior to co-IP assays (43). Cell lysates were mixed with an anti-FLAG M2 or anti-V5 antibody (final concentration of 1 g/l) followed by incubation at 4°C for 10 h with gentle rocking. Protein G-Sepharose (30 l of 50% slurry) in lysis buffer was added to the protein/antibody mixture, and the resulting mixture was incubated overnight at 4°C with gentle rocking. The Sepharose beads containing immunoprecipitated proteins were collected after centrifugation, washed three times with lysis buffer, resuspended in 50 l of 1ϫ SDS sample buffer, boiled at 95°C for 5 min, and used for immunoblot analysis (see above).
The co-IP assays were also performed by collecting culture medium from S2 cells expressing Toll-1 ecto or Toll-7 ecto proteins 48 h post-transfection and mixing it with DMEM from VSV-GFP-infected HEK293 cells that contained VSV-GFP virions as described above. Anti-V5 antibody plus protein G-Sepharose were then added as described above. Pulldown of Toll-1 ecto or Toll-7 ecto was detected on immunoblots using anti-V5, whereas VSV-GFP was detected using an antibody that binds VSV glycoprotein (anti-VSV-G) ([P5D4], ab50549, Abcam, 1:5000 dilution). As a control, cell culture media containing Toll-1 ecto , Toll-7 ecto , or VSV-GFP virions were incubated with protein G-Sepharose beads alone followed by immunoblotting and incubation with anti-V5 or anti-VSV glycoprotein as described above.

Real-time PCR analysis
Total RNA from adult flies or S2 cells was extracted, and the expression of target genes was determined by real-time PCR as described previously (33). The flies were anesthetized on a CO 2 bed, placed in 1.5-ml tubes, and homogenized with disposable pestles in 1 ml of TRIzol reagent (T9424, Sigma), and the total RNA from flies and S2 cells was extracted according to the manufacturer's instructions. The RNA pellets were air-dried and resuspended in 50 l of nuclease-free water, and the concentration of RNA was determined using a Nanodrop UV-visible spectrophotometer (ND-1000, Thermo Fisher Scientific).
Total RNA (2 g from each sample) was treated with RQ1 RNase-free DNase (M6101, Promega) to remove contaminated genomic DNA and then used for the synthesis of cDNAs in 25-l reactions using Moloney murine leukemia virus reverse transcriptase (M1701, Promega) and an anchor-oligo(dT) 18 primer following the manufacturer's instructions. The cDNA sample (diluted 1:50) was used as the template for quantitative real-time PCR analysis. The Drosophila ribosomal protein 49 (rp49) gene was used as an internal standard to normalize the expression of target mRNA. Real-time PCR was performed in 20-l reactions containing 10 l of 2ϫ SYBR GreenER TM quantitative PCR SuperMix Universal (no. 204141, Qiagen), 4 l of H 2 O, 4 l of diluted cDNA template, and 1 l (10 pmol) of each of the forward and reverse primers. The real-time PCR program was 2 min at 50°C, 10 min at 95°C, 40 cycles of 95°C for 15 s, and 60°C for 1 min, and the dissociation curve analysis. The data from four replicates of each sample were analyzed with a comparative method (2 Ϫ⌬⌬CT ) using ABI 7500 SDS software (Applied Biosystems). The baseline was automatically set by the software to maintain consistency. The cDNA sample from S2 cells transfected with empty pMT/BiP/V5-His A plasmid or WT flies (w 1118 ) was used as the calibrator. The expression level of target genes was calculated by the 2 Ϫ⌬⌬CT method (44), which provides the n-fold difference in relative expression compared with the calibrator. All the data are presented as relative mRNA expression levels, and all the experiments were repeated at least three times.

Toll family member functions Data analysis
Three or four replicates were performed for each assay using independently prepared and collected biological samples. Data were analyzed and figures generated using the GraphPad Prism platform (GraphPad, San Diego). Luciferase assays were analyzed by one-way ANOVA followed by a post hoc Tukey's honest significant difference test. Survival of Toll-1 and Toll-7 loss-of-function mutants after systemic infection by different microbes was compared with WT flies by a pairwise log-rank test.