REL1, a Homologue of Drosophila Dorsal, Regulates Toll Antifungal Immune Pathway in the Female Mosquito Aedes aegypti*

Signaling by Drosophila Toll pathway activates two Rel/NF-κB transcription factors, Dorsal (Dl) and Dorsal-related immune factor (Dif). Dl plays a central role in the establishment of dorsoventral polarity during early embryogenesis, whereas Dif mediates the Toll receptor-dependent antifungal immune response in adult Drosophila. The absence of a Dif ortholog in mosquito genomes suggests that Dl may play its functional role in the mosquito Toll-mediated innate immune responses. We have cloned and molecularly characterized the gene homologous to Drosophila Dl and to Anopheles gambiae REL1 (Gambif1) from the yellow fever mosquito Aedes aegypti, named AaREL1. AaREL1 alternative transcripts encode two isoforms, AaREL1-A and AaREL1-B. Both transcripts are enriched during embryogenesis and are inducible by septic injury in larval and female mosquitoes. AaREL1 and AaREL2 (Aedes Relish) selectively bind to different κB motifs from insect immune gene promoters. Ectopic expression of AaREL1-A in both Drosophila mbn-2 cells and transgenic flies specifically activates Drosomycin and results in increased resistance against the fungus Beauveria bassiana. AaREL1-B acted cooperatively with AaREL1-A to enhance the immune gene activation in Aag-2 cells. The RNA interference knock-outs revealed that AaREL1 affected the expression of Aedes homologue of Drosophila Serpin-27A and mediated specific antifungal immune response against B. bassiana. These results indicate that the homologue of Dl, but not that of Dif, is a key regulator of the Toll antifungal immune pathway in A. aegypti female mosquitoes.

The major reasons for this tragic situation are the unavailability of effective vaccines for malaria and other mosquito-borne diseases and the lack of insecticides and drug resistance to vectors and pathogens, respectively (6,7).
Immunity plays a key role in the interaction between a pathogen and its vector host. Many genes and their products involved in mosquito innate immunity have been identified and characterized either independently or on the basis of their relationships to known immune genes of the model organism, Drosophila melanogaster. The genomic DNA sequences of Anopheles gambiae, a major disease vector involved in the transmission of Plasmodium falciparum malaria in Africa, have been revealed, and they encompass more than 278 million base pairs, which suggests the presence of ϳ14,000 proteinencoding genes (8). Of these, 242 A. gambiae genes from 18 gene families have been implicated in innate immunity by comparative genomic analysis to the Drosophila immune systems (9). The mosquito immune genes involved in recognition, signal modulation, and effector systems diverge widely from those in Drosophila. In addition, these families of immune factors have undergone significant expansion during the evolution of mosquitoes, which possibly reflects different selection pressures to a variety of pathogens encountered in the distinct lifestyles of these insects. However, the components of two principal immune transduction pathways, the Toll and IMD, have principally been conserved between these two insects. This points to the evolutionary requirement of preserving the integrity of key factors in intracellular immune signal transduction and gene activation pathways (9).
In Drosophila, three Rel/NF-B molecules, Dorsal-related immunity factor (Dif), 1 Dorsal (Dl), and Relish, affect the expression of numerous immune genes, including those encoding antimicrobial peptides (AMPs) (for review, see Refs. 10 and 11). These Rel/NF-B molecules are involved in two distinct innate immune pathways, the Toll pathway, which is mediated by Dl and Dif and responds primarily to fungal and Gram-positive bacterial infections, and the IMD pathway, which is regulated by Relish and is predominantly directed against Gram-negative bacteria. In the Toll-signaling pathway, Dif stimulates the production of the antifungal factor Drosomycin (12). In loss-offunction mutants affecting the Toll signaling pathway, the inducibility of Drosomycin in immune-challenged adults is severely compromised, whereas those of Defensin, Cecropin, and Attacin are reduced to a lesser extent. In contrast, the expres-sion of Diptericin, Drosocin, and Metchnikowin are completely unaffected (13,14). In Drosophila relish mutants, the induction of immune defense is severely reduced, and insects become extremely sensitive to bacterial and fungal infection (15). Immune effector genes under the regulation of Relish include Cecropin, Diptericin, Attacin, Defensin, and Metchnikowin (15,16). Recent microarray studies have shown that in addition to AMP genes many other immune genes are regulated cooperatively or independently by the Toll and IMD pathways (17). These genes include recognition molecules like peptidoglycan recognition proteins and Gram-negative bacteria-binding protein, which are components involved in both the IMD and Toll pathways; the protease cascade proteins; and putative components of the phenoloxidase or blood clotting pathways.
Dl, the first insect member of NF-B family, was identified in a screen for genes required for Drosophila embryonic development. Dl is a key regulator in determining dorsoventral polarity (18). A gradient of nuclear Dl spatially restricts the expression of zygotic genes along the dorsoventral axis and functions both as a transcriptional activator and a repressor (19). The remarkable structural and functional similarities between the mechanisms of activation of Dl during Drosophila morphogenesis and NF-B during the mammalian acute phase response have implicated Dl and Dif in the host defense of Drosophila (20,21). However, two independent studies clearly show that Dif, the possible duplicator of the Dl gene during Drosophila evolution, is the essential regulator of the Drosophila immune Toll pathway. The rescue experiments of Drosophila Dl and Dif double mutants have shown that the ectopic expression of Dif without Dl is sufficient to mediate the induction of Drosomycin (22). In addition, genetic epistasis studies with Dif mutant flies have demonstrated that Dif mediates the Toll-dependent control of the inducibility of the Drosomycin gene (23).
The Drosophila Toll pathway regulates many aspects of the immune response (for review, see Ref. 24). In addition to activating antifungal defenses, this pathway is also required for resistance against Gram-positive bacteria, for regulation of the melanization cascade and blood cell proliferation. Drosophila mutants that constitutively activate Toll pathway, such as necrotic, Toll 10b , and cactus, all show hemocyte phenotype with overreactive blood cells that form melanotic masses (25), similar to the cellular response to parasites (26).
Surprisingly, the A. gambiae genome sequence harbors only two NF-B genes, REL1 (gambif1) and REL2, which are homologues of Drosophila Dl and Relish, respectively (9,27). The A. gambiae Dl homologue Gambif1 (now called REL1) can bind to B motifs on the promoters of Drosophila, Diptericin, and Cecropin and activate transcription of a reporter gene under the control of the Diptericin B motif in cell culture studies (28). The lack of a Dif ortholog in mosquitoes raises a question about the difference between their immune responses and those of adult Drosophila and about the immune function of mosquito homologue of Dl (REL1). In particular, it has not been demonstrated in vivo whether mosquito REL1 serves as a functional Dif analog in adult mosquitoes. In this report, we have cloned and characterized a homologue to Drosophila Dl from the yellow fever mosquito Aedes aegypti, which we named AaREL1. In vitro comparative binding studies, transfection assays in both Drosophila mbn-2 cells and mosquito A. aegypti Aag-2 cells, the transgenic overexpression in flies, and in vivo studies by the RNAi knock-outs in mosquitoes provide conclusive evidence to implicate AaREL1 as the key activator of the Toll-mediated antifungal immune pathway in adult female mosquitoes.

EXPERIMENTAL PROCEDURES
Isolation of cDNA and Genomic DNA Clones and RACE-A PCR product was obtained from genomic DNA by degenerate primers based on the conserved region of the Rel homology domain (RHD) from Drosophila Dl and Anopheles REL1 (gambif1). The following primers were used, AaREL1-R2 (5Ј-CAGTG(T/C)GCCAA(A/G)AAGAAGGA-3Ј) and AaREL1-R-A (5Ј-GGCAATCTTCTCGCACAGCA-3Ј). This PCR fragment was subcloned using a TA cloning kit (Novagen) and was then utilized as a probe to screen the Lambda ZAPII cDNA library prepared from previtellogenic female A. aegypti mosquitoes. Six clones were isolated and sequenced from both the 5Ј-and 3Ј-ends. Based on sequencing and restriction mapping analyses, the six clones were subdivided into two groups. The longest representative of each cDNA group, the D4 and D6 cDNA clones, were fully sequenced from both strands. The 5Јand 3Ј-ends of D4 and D6 cDNA clones were identified by the RACE system (Invitrogen). Two overlapping PCR products spanning 13,355 bp in length were isolated using the expand long template PCR system (Roche Diagnostics). These PCR fragments were subcloned using a TOPO XL cloning kit (Invitrogen) and then sequenced. The sequences reported in this paper have been deposited in the GenBank TM data base (accession numbers AY748242 for AaREL1-A cDNA, AY748243 for AaREL1-B cDNA, and AY748244 for genomic DNA).
Northern Hybridization and RT-PCR-Adult 2-or 3-day-old A. aegypti females were injected with a stationary phase culture of Enterobacter cloacae and/or Micrococcus cloacae. For the stage-specific study, adult males, females, and fourth instar larvae were collected with or without bacterial challenge. The vitellogenic mosquitoes were collected 1 and 2 days after blood feeding. The unfertilized eggs were obtained by dissecting ovaries from mosquitoes 3 days after blood feeding, whereas fertilized eggs were collected 1 day after egg laying. Total RNA was prepared by the TRIzol technique (Invitrogen). Samples of 5 g of total RNA were separated on a formaldehyde gel, blotted, and hybridized with a corresponding DNA probe. RT-PCR was performed by using the Titan one-step RT-PCR kit (Roche Diagnostics) with samples of 0.2 g of total RNA as templates. Tubes containing RNA and RNase inhibitor (1 u/l, Roche Applied Science) were incubated for 30 min at 50°C for RT-reaction. Amplification conditions included rapid heating to 94°C for 2 min followed by 25-30 cycles of 55°C for 1 min, 72°C for 3 min, and 94°C for 45 s.
Electrophoretic Gel Mobility Shift Assay (EMSA)-Each protein was synthesized by a coupled in vitro transcription-translation (TNT) system (Promega). The corresponding cDNA clones were subcloned into pcDNA3.1/Zeo (ϩ) (Invitrogen). The in vitro transcription-translation reactions programmed by the circular plasmid DNA utilized the T7 promoter. To confirm the synthesis of proteins with expected size, the control TNT reactions of each protein were performed in the presence of [ 35 S]methionine, and the resulting reactions were analyzed by SDS-PAGE and autoradiography. The annealed deoxyoligonucleotide of B motifs were purified from 15% TBE (90 mM Tris borate, 2 mM EDTA, pH 8.0) criterion precast gel (Bio-Rad) and labeling of double-stranded oligonucleotides, and EMSA was performed with a gel-shift assay system (Promega). The protein-DNA complex was separated on 5% TBE criterion precast gel and visualized by autoradiography.
Transfection Assay in Drosophila mbn-2 Cell Line-Coding region sequences of A. aegypti, ⌬REL2 (Relish C8 cDNA, which encodes truncated Rel-type protein without N-terminal transactivation domain, Ref. Drosophila mbn-2 cells were seeded at 2 ϫ 10 6 cells/ml in 35-mm plates of Drosophila Schneider's medium (Invitrogen) supplemented with 5% fetal bovine serum (Invitrogen) and 1ϫ Gluta Max-1 (Invitrogen). Twenty-four hours later, the cells were co-transfected with 2 g of expression plasmid. The cells were incubated for 4 h in serum-free Schneider's medium, which was then removed and replaced with complete medium. Overexpression was induced by 48 h of incubation with the addition of 500 M CuSO 4 . The immune response was activated by the addition of heat-inactivated Escherichia coli for 4 h.
Transfection Assay in Aag-2 Cell Line-Cell line Aag-2 from A. aegypti was maintained in Schneider's medium supplemented with 10% fetal bovine serum (Hyclone). Aedes REL2ϩ, AaREL1-A, and AaREL1-B were inserted into the pAc5.1/V5/HisA (Invitrogen) vector. Cells were incubated at 26°C for 36 h prior to transfection. After transfection, the cells were incubated at 26°C until they reached at least 70% confluency (ϳ24 h). Following transfection, for the bacterial challenge, heat-inactivated E. coli was added to the well and incubated for an additional 4 h, then cells were harvested for RNA extraction.
Generation of Transgenic Flies-AaREL1-A cDNA (D4) was subcloned downstream of the upstream activating sequence in piggyBac 3xP3-EGFP-based vector. Transformants were established by microinjection of wild-type Canton-S strain homozygous for white mutation. Two independent Drosophila lines were examined, and both produced similar results in overexpression of AaREL1-A transgene driven by Hsp70-GAL4 control element-using driver line P{GAL4-Hsp70.PB}2, which was kindly provided by the Indiana University Drosophila stock center (Bloomington, IN).
Synthesis and Microinjection of dsRNA-The synthesis of dsRNA was accomplished by simultaneous transcription of both strands of template DNA with a HiScribe RNAi Transcription Kit (New England Biolabs). The plasmid LITMUS 28iMal containing a nonfunctional portion of the E. coli malE gene that encodes a maltose-binding protein was used to generate control dsRNA. After RNA synthesis, the samples were treated by phenol/chloroform extraction and ethanol precipitation. The dsRNA was then suspended in diethyl pyrocarbonate-treated distilled water with a final concentration of 5 g/l. The formation of dsRNA was confirmed by running 0.2 l of these reactions in a 1.0% agarose gel in TBE. A Picospritzer II (General Valve, Fairfield, NJ) was used to introduce 200 nl of dsRNA into the thorax of CO 2 -anesthetized mosquito females, at 2-3 days post-eclosion.
Septic Injury and Survival Experiments-Septic injuries were performed by pricking the flies or mosquitoes in the rear part of the abdomen with a Hamilton 31S needle dipped into bacterial culture or a fungal spore suspension. Survival experiments were carried out under the same conditions. Groups of 20 female Drosophila or A. aegypti female mosquitoes after 2 days of recovery following AaREL1 dsRNA injection were challenged by bacteria culture or a spore suspension (ϳ5 ϫ 10 7 viable spores/ml) of B. bassiana strain GHA. The viable spore number was calculated by spreading the suspension onto Sabouraud dextrose agar plates.

Cloning of Two REL1 Isoform cDNAs from A. aegypti-A
376-bp DNA fragment was obtained using a set of sense and antisense primers (AaREL1-R2 and AaREL1-R-A) and was used as a probe to screen the cDNA library prepared from previtellogenic A. aegypti female mosquitoes. The nucleotide sequence of the longest 2,250-bp D4 clone out of a total of 6 clones contained a full open reading frame that was conceptually translated into a 579-amino acid-long polypeptide. Additional 5Ј-untranslated region (UTR, 41-bp) and 3Ј-UTR (11-bp) sequences that were followed by poly(A) sequence were identified by RACE PCR. The sequence and overall structure of the conceptual protein from D4 was highly similar to Drosophila Dl and A. gambiae REL1 (AgREL1), and thus the A. aegypti homologue was named AaREL1-A (Fig. 1A). A phylogenetic analysis using Rel-homology domains demonstrated that AaREL1 and AgREL1 clustered with Drosophila Dl. This Dl/REL1 subgroup, together with Drosophila Dif, formed a separate cluster that was different from the group of mammalian Rel proteins (Fig. 1B).
Sequencing of another cDNA clone, D6, indicated differences in the 5Ј-UTR sequence that contained an alternative start codon. Both D4 and D6 cDNA clones shared identical sequences only in the RHD region, but the C-terminal domains (CTDs), including the nucleus localization signal (NLS), were completely divergent. An additional 5Ј-UTR sequence of 263 bp of D6 cDNA was obtained by 5Ј-RACE PCR. Two major PCR products, short and long form, obtained by 3Ј-RACE PCR, added the 3Ј-truncated region of the D6 cDNA. Both short and long RT-PCR products were subcloned and additional transcript sequences of 1,290 and 1,967 bp were obtained. The complete short and long form transcripts were 3,433 and 4,110 bp in size, which encoded the same protein. The additional 3Ј-UTR of 4,110-bp transcript was identical to the 3Ј-region of D4 cDNA clone, indicating that long form transcript might be the precursor mRNA of the D4-type transcript. The 844-residue protein encoded by these transcripts lacked N-terminal 39 residues that were encoded by the AaREL1-A transcript, which indicated the presence of an alternative transcription start site. We were able to isolate 2,894-bp (844 residues) and 2,968-bp (883 residues) cDNA fragments by RT-PCR that yielded polypeptides with or without N-terminal 39 residues. Northern blot analysis clearly showed that all A. aegypti REL1 transcripts used both start sites, which resulted in doublet bands for each alternate transcript (Fig. 2A).
The CTD of the second A. aegypti REL1 form (named AaREL1-B) was highly conserved with the rear portion of Drosophila Dl-B CTD, which possessed trans-activating properties (Fig. 1A). A crucial region for the transactivation properties of Drosophila Dl-B, located between amino acids 576 and 683, contains two putative acidic activation modules (30). These two modules and a bipartite NLS were highly conserved in the CTDs of Drosophila and AaREL1-B. A Q-rich region, present in the CTD of Drosophila Dl-B, was absent in AaREL1-B. The Rel domains of Dl-B and AaREL1-B lacked an NLS at their ends. Instead, both Dl-B and AaREL1-B had a bipartite NLS in their CTDs. This was a common variant for NLS in which a small cluster of basic residues positioned 10 -12 residues at the Nterminal of a monopartite-like NLS sequence. The additional binding energy contributed by the upstream cluster of basic residues relaxes the requirements for the downstream monopartite-like sequences (31). As shown in Fig. 1A, the monopartite NLS sequence of AaREL1-A consisted of all basic residues, KRKKRK, whereas the monopartite-like sequence of AaREL1-B was SRKKSK and included an upstream cluster of basic residues, KNKK.
Two overlapping PCR products spanning 13,355 bp in length were isolated and sequenced. This genomic region did not contain a 5Ј-UTR, N-terminal region (M1-C150), and short 3Ј-UTR as shown in the D4-type transcript. The sequence result, however, gave us sufficient information about how the RNA was processed in generating the different types of AaREL1 transcripts (Fig. 1C). These exon-intron structures and the alternative splicing of the AaREL1 gene were very similar to the Drosophila Dl gene (30), showing that both insect genes originated from a common ancestor gene.
The Inducible AaREL1 Isoform Transcripts Are Expressed in Adult A. aegypti Females-Northern blot analyses were performed to examine the expression of AaREL1 transcripts. Utilization of the total RNA revealed three transcript bands of ϳ2.5, ϳ3.5, and ϳ4.3 kb in size that were constitutively expressed during the normal unchallenged state of the naïve larvae in adult males and females ( Fig. 2A). The probe from the AaREL1-B-specific CTD region (F400-L568 of AaREL1-B) did not hybridize to the ϳ2.5-kb band, suggesting that it represented the AaREL1-A transcript ( Fig. 2A). According to the sizes of the transcripts, the ϳ3.5and ϳ4.3-kb bands corresponded to the short and long AaREL1-B transcripts, respectively. The expression level of all transcripts in larvae and adult female mosquitoes were more elevated 5 h after septic injury with a mixture of stationary-phase culture of E. cloacae and Micrococcus luteus. AaREL1 transcripts remained unchanged upon bacterial challenge in adult males.
Without the immune challenge, the expression level of AaREL1 transcripts was very low in the fat body and mainly occurred in the ovary of naïve previtellogenic female mosquitoes. The AaREL1 gene expression was elevated after blood feeding as well as after bacterial challenge, and it increased further during egg development. All REL1 transcripts showed the highest level in unfertilized eggs that were dissected from the ovaries of female mosquitoes 3 days post-blood feeding and in 1-day post-oviposition eggs (Fig. 2A). These AaREL1 expression data in the ovary and during vitellogenesis, and in eggs of A. aegypti, indicated REL1 participation in the Toll-mediated FIG. 1. Molecular characterization of two Aedes REL1 isoforms, AaREL1-A, and AaREL1-B. A, amino acid sequence comparison of AaREL1-A and AaREL1-B to Drosophila Dl, Dl-B, and Anopheles REL1 (AgREL1). Alignments were done by ClustalW and were manually adjusted. CTDs with a Gln (Q)-rich or Gln/His (Q/H)-rich region that was shown to be important for transactivation in Drosophila Dl and Dif are early embryogenesis similar to that of Drosophila Dl. In addition, AaREL1 expression was greatly induced after septic injury in the fat body of previtellogenic female mosquitoes, which implicated AaREL1 involvement in innate immunity at this stage of the mosquito life cycle (Fig. 2A).
Comparative Analysis of A. aegypti REL1 Isoforms and REL2 (Relish) Binding to B Motifs-We expressed AaREL1-A and AaREL1-B by using an in vitro coupled transcription-translation assay. We used D4 cDNA for AaREL1-A and a 2,894-bp RT-PCR product containing the alternative open reading frame of 844 residues for AaREL1-B. In vitro translation of D4 cDNA resulted in two REL1-A protein bands, indicating that an alternative start site could be used as the translation initiation codon (data not shown). Only a single major band was observed after in vitro translation of the REL1-B RT-PCR product. The sizes of translated proteins matched well with those expected from deduced proteins (data not shown). For in vitro translation of AaREL2, we used our previous construct (29). Seven B motifs, including those from Drosophila Cecropin A1 and A. aegypti Defensin promoters (29), were designed based on the comparison of promoter sequences of various insect immune genes (Fig. 2B).
In EMSA, in vitro expressed Dl-A specifically bound to four B motifs among seven tested motifs (Fig. 2B). The highest affinity of AaREL1-A binding was to the B motif (I) from Drosophila CG16978 promoter. This B motif, 5Ј-GGGAAAT-TCC-3Ј, from the promoter of Drosophila unknown immune gene CG16978, matched well to B motif consensus, GGG-(W)nCCM, of Drosophila zen ventral repression element (32). This type of B motif was present in the promoter regions of Drosophila immune genes whose expression was activated only by the Toll pathway (17). Similar B motifs were found from Drosophila GNBP-like (CG13422) and Nec gene promoters (Fig. 2B). The other six sites were representatives of various B motifs found in upstream promoter regions of Drosophila and mosquito immune genes (Fig. 2B). The addition of 25-fold excess of the unlabeled specific oligonucleotide (I) effectively completed binding to the labeled probe, whereas the addition of a nonspecific competitor, AP2, affected binding very weakly (Fig.  2C). The competition of AaREL1-A binding to I motif by three B motifs (V, VI, and VII) was similar to that of the nonspecific competitor, indicating that binding of AaREL1-A to these motifs was nonspecific (Fig. 2D). The binding specificities of AaREL1-A and AaREL2 to various B motifs from insect immune gene promoters were clearly different (Fig. 2B). To test whether AaREL1-A and AaREL2 could bind to B motifs as heterodimers, equal amounts of both these Rel proteins, cotranslated in vitro (data not shown), were subjected to EMSA. Formation of additional complexes was not detected, indicating that AaREL1-A and AaREL2 bound to tested B motifs only as homodimers ( Fig. 2A). Binding activity of in vitro translated AaREL1-B was not detected with any B motif (data not shown). In addition, its co-translation with either AaREL1-A or AaREL2 did not affect B binding of these two factors.

AaREL1-A Activates Drosomycin Expression in Drosophila mbn-2 Cell Line and Transgenic Flies-Drosophila tumorous blood cells (mbn-2 line) can be induced to express Diptericin
and Cecropin genes by the addition of lipopolysaccharide to the culture medium (33,34). Moreover, expression of the Drosomycin gene is highly activated by immune challenge in this cell line (35). We examined the expression of Drosophila AMP genes in the presence of exogenous mosquito NF-B proteins. The overexpression of mosquito AaREL1-A in mbn-2 cells activated strong expression of the Drosophila Drosomycin gene in the absence of bacterial challenge (Fig. 3A). The expression pattern of Diptericin, Cecropin, and Attacin genes remained constant during overexpression of any mosquito Rel protein (Fig. 3A, data not shown).
Next, the mutant flies, which overexpressed AaREL1-A by the heat shock (hs)-GAL4/UAS system, were constructed to show in vivo activity of AaREL1-A. Because of the leaky promoter, AaREL1-A had a background expression which could activate the Drosomycin gene independent of septic injury with fungal infection (Fig. 3B). The overexpression of AaREL1-A by heat shock resulted in an increased expression level of the Drosomycin gene but not other Drosophila AMP genes such as Diptericin, Cecropin, and Attacin (data not shown). The level of Drosomycin expression is significantly higher in heat-shocked flies overexpressing AaREL1-A than that fully induced by immune challenge in the presence or absence of bacterial challenge. (Fig. 3B). These transgenic flies exhibited increased resistance to the infection with B. bassiana spores (Fig. 3C).

REL1-A and REL1-B Synergistically Activate the Immune Genes in Aag-2 Mosquito Cells-Aag-2 cell line originated from
A. aegypti embryonic tissues is responsive to the immune challenges (36). Immune factors such as cecropin, defensin, lysozyme, and transferrin were inducibly expressed in this cell line. Drosophila Serpin-27A, which regulates the melanization cascade through the specific inhibition of the prophenoloxidaseactivating enzyme, is an acute immune-responsive gene mainly regulated by the Toll pathway (37). We have isolated the mosquito homologues of Drosophila Serpin-27A from A. aegypti. 2 In addition to Aedes Cecropin A and Defensin A, the expression of mosquito Serpin-27A was induced by heat-killed bacteria in Aag-2 cells (Fig. 4A).
Transfection of AaREL1-A alone in Aag-2 cells could weakly activate mosquito immune genes. However, when both AaREL1-A and AaREL1-B were co-transfected together, the same immune genes were strongly activated (Fig. 4A). This experiment showed that AaREL1-B acted cooperatively with AaREL1-A in enhancing the immune gene activation. This response was similar to that observed for Drosophila Dl-B, which increased in vitro activation of B-like promoters when it was co-expressed with Dl-A (30). Because AaREL1-B did not bind to any B motifs directly, it might act as a cofactor to AaREL1-A.
Aedes REL1 Regulates Antifungal Immune Response in Vivo-We tested the expression profiles of several immune genes in A. aegypti mosquitoes after infection with Gram-negative bacteria (E. cloacae), Gram-positive bacteria (M. luteus), and fungal spores (B. bassiana). The expression levels of all tested immune genes were highly elevated after any type of infection. However, the specific response of A. aegypti REL1 and Serpin-27A was elicited only by the fungal challenge (Fig. 4B).  Right panel, a DNA fragment encoding AaREL1-B-specific CTD was used as a probe. Aedes actin gene was used to reprobe the same blot as an RNA loading control. B, mobility shift assay for the binding of AaREL1-A and AaREL2 to various insect B motifs. Each B motif used in this experiment was listed in this figure. The putative B binding sites from various insect AMP and other immune genes were collected. Motifs similar in their sequences and patterns were grouped, and the representative B motif from each group was used for EMSA assay. D4 cDNA (AaREL1-A) and AaREL2 C8 cDNA (29) were used for in vitro transcription-coupled translation. The binding affinity of seven representative B binding motifs to AaREL1-A and AaREL2 were compared. Equal quantities of D4 and C8 cDNAs were transcribed in one reaction to test any interaction between two NF-B proteins. The unbound primers were shown as the labeling and loading control of each B motif (Input). C, the binding of AaREL1-A to B motif (I) from the Drosophila CG16978 gene promoter is almost abolished by the addition of 25-fold (25ϫ) of unlabeled specific competitor but not by the nonspecific probe AP2. D, relative binding affinity of each B motif (I-VII) to AaREL1-A was shown by competition assay. The binding of AaREL1-A to B motif (I) from the Drosophila CG16978 gene promoter was competed with 25-fold unlabeled B motifs (I-VII).
Next, dsRNA complementary to the RHD of AaREL1 was synthesized in vitro and injected into the thorax of newly emerged female mosquitoes. When AaREL1 dsRNA was introduced into the mosquitoes, the mRNA level of both AaREL1 transcripts greatly decreased (Fig. 4C). The mRNA quantity of A. aegypti Serpin-27A in AaREL1 dsRNA-treated mosquitoes also significantly declined, suggesting that AaREL1 was involved in the regulation of A. aegypti Serpin-27A gene expression. The mRNA level of AaREL2 was not affected by the treatment of AaREL1 dsRNA.
Genetic analyses have shown that Drosophila Toll pathway mutants are sensitive to fungal infection (13). To address the role of Aedes Dorsal homologue in mosquito immune response, the susceptibility of either AaREL1 or AaREL2 dsRNA-treated mosquitoes was compared after bacterial or fungal challenge. After 2-4-day recovery following injection with either dsRNA, the treated mosquitoes were challenged by either the spores of entomopathogenic fungus, B. bassiana, or the mixture of E. cloacae and M. luteus. As illustrated in Fig. 4D, the AaREL1 dsRNA-treated mosquitoes were significantly more sensitive to the fungal infection than AaREL2 or MalE dsRNA-treated mosquitoes. In contrast, only AaREL2 dsRNA-treated mosquitoes were susceptible to the bacterial challenge (Fig. 4E). We did not observe any increased susceptibility to Gram-positive bacteria in either AaREL1 or AaREL2 dsRNA-treated mosquitoes (data not shown). DISCUSSION We have investigated the contribution of REL1 in the antifungal immune defense in adult females of A. aegypti mosquitoes. This investigation has been prompted by the lack of a homologue of Drosophila Dif in mosquitoes (9). Dif is the im-mune factor responsible for the antifungal immune defense in adult Drosophila (12,13). In this report, we have cloned and characterized a homologue to Drosophila Dl from the yellow fever mosquito A. aegypti, which we named AaREL1. Although A. gambiae REL1 (Gambif1) has previously been cloned, the conclusive in vivo data on its role in adult A. gambiae had been lacking (28). Our study has convincingly demonstrated that mosquito REL1 serves as a functional Dif analog in adult A. aegypti females.
The AaREL1 expression profile is different from Drosophila Dl. In Drosophila, Dl transcripts were previously shown to be markedly enhanced in adult males upon bacterial challenge (20). In a later study by Gross et al. (30), Drosophila Dl-B transcript was increased upon bacterial challenge in larvae and adult males. AaREL1 transcripts, however, were elevated in larvae and adult females after bacterial challenge. Moreover, the tissue-specific expression in female mosquitoes, the enriched ovary expression in the naïve state, and the elevated fat body expression after septic injury suggest AaREL1 as an immune factor of female mosquitoes and a morphogen during egg development.
Although many complications have been reported concerning specificity of response and gene regulation by IMD or Toll pathways (24), the fungal-specific late induction of Drosomycin expression and antifungal activity against some fungi including B. bassiana are still two representative characteristics of the Drosophila Toll immune pathway. We overexpressed mosquito AaREL1-A in both in vivo and in vitro Drosophila systems and examined the expression of Drosomycin in their native chromosomal environment. Our results demonstrated that AaREL1-A could fully up-regulate Drosomycin expression prior to bacterial challenge. In addition, the ectopic expression of AaREL1-A increased the resistance against B. bassiana in transgenic flies. In Drosophila S2 cells, the Drosomycin expression was more affected by Dif, than by Dl (39). Moreover, Dif, not Dl, was sufficient to mediate the induction of Drosomycin and antifungal immunity against B. bassiana in mutant adult flies (22,23). The characteristics of AaREL1 in both a Drosophila cell line and transgenic flies clearly showed that AaREL1-A, a mosquito Dl homologue, could function as Drosophila Dif in antifungal immune response. However, there were differences between AaREL1 and Drosophila Dif and/or Dl. The expression of Dif or Dl prior to lipopolysaccharide incubation caused a modest up-regulation of Drosomycin expression in S2 cells, and, after lipopolysaccharide treatment, further up-regulated the expression (39). With immune challenge, the overexpression of Dif or Dl under a heat shock promoter rescued the lack of Drosomycin inducibility in larval fat body cells of TW119 mutant flies with a deficiency uncovering both Dif and Dl genes (40). In contrast, the overexpression of AaREL1-A fully upregulated the Drosomycin expression without the immune challenge, and no further up-regulation was found after the challenge. These results suggest that AaREL1 might not interact with some components of the Drosophila Toll pathway. In Drosophila loss-of-function mutant of cactus, a Drosophila I-B inhibitor specific to the Toll pathway, the Drosomycin genes are constitutively transcribed (14,41), and the overexpression of Drosophila Dl could fully activate the Drosomycin expression without immune challenge (40).
Because of the absence of Drosomycin homologue in mosquitoes, we surveyed other mosquito immune genes possibly regulated by Toll pathway. In general, the Drosophila immune genes dependent upon Toll pathway show a specific activation profile by fungal challenge. Drosomycin expression was shown to be partially affected by the IMD pathway after bacterial infection but to be regulated predominantly by the Toll path- way during fungal infections (17,23). In addition, the microarray analysis shows that the Toll pathway controls most of the late genes induced by fungal infection (17). To find the mosquito immune genes regulated by Toll pathway, A. aegypti females were challenged by different types of microorganisms (E. cloacae, M. luteus, and B. bassiana spore). The results showed that the expression of Aedes Serpin-27A was elicited specifically by the fungal challenge, suggesting regulation by the Toll pathway similar to Drosophila Serpin-27A (37).
The increased fungal susceptibility of AaREL1 dsRNAtreated mosquitoes more clearly indicated that AaREL1 is an essential factor of the Toll antifungal immune response. Our dsRNA knock-down results have demonstrated that mosquito antifungal immune response by AaREL1 can be distinguished from the immune response against Gram-negative bacteria mediated by AaREL2, previously reported as a key regulator of mosquito IMD pathway (38). Transgenic mosquitoes with a stable dominant negative immune-deficient phenotype for AaREL2 showed a marked susceptibility to Gram-negative bacteria infection, which severely compromised induction in the expression levels of both Defensin A and Cecropin A. This indicated that the IMD pathway is generally conserved between Drosophila and mosquitoes (38). Our RNAi experiments demonstrated the AaREL1 regulated antifungal immune response, distinctly separated from the IMD pathway immune response regulated by REL2.
In vitro comparative binding studies, transfection assays in both Drosophila mbn-2 cells and mosquito A. aegypti Aag-2 cells, and the ectopic overexpression in transgenic flies provide conclusive evidence to implicate AaREL1 as a Dif analog of the Toll-mediated antifungal immune pathway in adult female mosquitoes. In addition, in vivo studies by the RNAi knock-outs in mosquitoes strongly suggest that AaREL1 serves as a key regulator of mosquito Toll immune pathway but not as a redundant factor of the immune response.