Induction of Nitric Oxide Synthase in Anopheles stephensi ...thologous to those in Drosophila melanogaster (11, 48) have been described from Aedes aegypti (45, 46) and from Anopheles

INFECTION AND IMMUNITY, May 2005, p. 2778–2789 Vol. 73, No. 50019-9567/05/$08.00�0 doi:10.1128/IAI.73.5.2778–2789.2005Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Induction of Nitric Oxide Synthase in Anopheles stephensi byPlasmodium falciparum: Mechanism of Signaling and

the Role of Parasite GlycosylphosphatidylinositolsJunghwa Lim,1 D. Channe Gowda,2 Gowdahalli Krishnegowda,2

and Shirley Luckhart1*Department of Biochemistry, Virginia Tech, Blacksburg, Virginia,1 and Department of Biochemistry

and Molecular Biology, Hershey Medical Center, Pennsylvania State University,Hershey, Pennsylvania2

Received 2 November 2004/Returned for modification 15 December 2004/Accepted 7 January 2005

Malaria parasite (Plasmodium spp.) infection in the mosquito Anopheles stephensi induces significant ex-pression of A. stephensi nitric oxide synthase (AsNOS) in the midgut epithelium as early as 6 h postinfectionand intermittently thereafter. This induction results in the synthesis of inflammatory levels of nitric oxide (NO)in the blood-filled midgut that adversely impact parasite development. In mammals, P. falciparum glyco-sylphosphatidylinositols (PfGPIs) can induce NOS expression in immune and endothelial cells and aresufficient to reproduce the major effects of parasite infection. These effects are mediated in part by mimicry ofinsulin signaling by PfGPIs. In this study, we demonstrate that PfGPIs can induce AsNOS expression in A.stephensi cells in vitro and in the midgut epithelium in vivo. Signaling by P. falciparum merozoites and PfGPIsis mediated through A. stephensi Akt/protein kinase B and a pathway involving DSOR1, a mitogen-activatedprotein kinase kinase, and an extracellular signal-regulated kinase. However, despite the involvement ofkinases that are also associated with insulin signaling in A. stephensi cells, signaling by P. falciparum and byPfGPIs is distinctively different from signaling by insulin. Therefore, although mimicry of insulin by PfGPIsappears to be restricted to mammalian hosts of P. falciparum, the conservation of PfGPIs as a prominentparasite-derived signal of innate immunity can now be extended to include Anopheles mosquitoes, indicatingthat parasite signaling of innate immunity is conserved in mosquito and mammalian cells.

Anopheles stephensi, a primary vector of Plasmodium spp. inIndia and the Middle East, limits malaria parasite develop-ment with the inducible synthesis of nitric oxide (NO) (34)catalyzed by A. stephensi NO synthase (AsNOS (32, 33). Induc-tion of AsNOS expression is proportional to the intensity ofparasite infection and is detectable in the midgut by 6 h postin-fection (15, 31). Early induction is critical to inhibition ofparasite development: dietary provision of the pan-NOS inhib-itor N�-nitro-L-arginine, with a half life in blood of 3 to 6 h(13), resulted in significantly higher parasite infection intensi-ties than did the inactive enantiomer N�-nitro-D-arginine (34).

The NO-mediated defense of A. stephensi is analogous tomammalian NO-mediated inactivation of liver-invading sporo-zoites and blood-stage gametocytes (36, 42), indicating thatmosquitoes and mammals share a conserved antiparasite de-fense. The activation of mammalian immune effectors, includ-ing inflammatory cytokines, adhesion molecules and iNOS, hasbeen attributed to parasite GPIs (reviewed in (22) and tohemozoin (27, 38, 55). In general, GPIs consist of a conservedethanolamine phosphate-trimannosylglucosaminyl glycan coreattached to phosphatidylinositol. GPIs are ubiquitous in eu-karyotic cells, where their primary function is to anchor pro-teins to the cell membrane. In the case of P. falciparum GPIs

(PfGPIs), key structural features include a terminal fourthmannose, variable fatty acyl substituents with unsaturated acylresidue on sn-2 position on glycerol, and C16:0 acyl moiety onC-2 of inositol (22). Compared to animal cells, parasites ex-press GPIs at levels severalfold higher than are required forprotein-anchoring (20). A number of studies during the pastdecade have shown that GPIs of various pathogenic parasites,including Plasmodium, Trypanosoma, and Leishmania species,are biologically active. For example, parasite GPIs can inducethe production of proinflammatory cytokines and NO (49, 63).From the point of view of the host, these innate immuneresponses represent a first line of defense for recognition andelimination of parasites through responses that are toxic toinvading microorganisms.

Early studies revealed that PfGPIs could induce lipogenesisand glucose oxidation in rat adipocytes and that injection ofPfGPIs into mice could induce hypoglycemia (52). These ob-servations led to the hypothesis that PfGPIs were insulin-mi-metic. Subsequently, it was demonstrated that malaria parasiteGPIs exhibited signaling characteristics of the insulin secondmessenger phosphoinositolglycan (PIG) (7), which is releasedfrom host cell GPI by insulin stimulation of phosphatidylino-sitol-dependent phospholipase activity. Although no additionalstudies have examined the insulin-like signaling behavior ofPfGPIs in detail, studies with synthetic insulin-mimetic PIGs,developed for treatment of insulin-resistant diabetes, providerelevant insight into parasite GPI signaling. In adipocytes, theinsulin-mimetic PIGs bypass insulin receptor (INR) activationand instead interact with an unidentified cell surface protein to

* Corresponding author. Present address: Department of MedicalMicrobiology and Immunology, 3146 Tupper Hall, One Shields Ave-nue, University of California at Davis, School of Medicine, Davis, CA95616. Phone: (530) 754-6963. Fax: (530) 752-8692. E-mail:[email protected].

2778

on April 4, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

induce tyrosine phosphorylation of mammalian insulin recep-tor substrates (IRSs), including IRS-1 and IRS-3 (21, 39).PIG-dependent IRS phosphorylation is then followed by sig-naling through the two major insulin signaling pathways (Fig.1) involving phosphatidylinositol 3-kinase (PI3-K), Akt/PKB,and MEK/ERK (21).

Differential activation of mosquito immune genes by bacte-ria and Plasmodium spp. (18, 19, 34, 44). indicates a degree ofimmune recognition that may be based in part on activation ofhost pathways by parasite GPIs. Available data on well-knownGPI-linked parasite proteins suggest that GPIs derived fromboth asexual and sexual stages would be available to signalinduction of AsNOS expression in A. stephensi from bloodmealingestion through sporogonic development (16, 24, 37, 51). Inaddition, recent work suggests that relevant signaling pathwaysin mosquito cells could transduce signals from PfGPIs forAsNOS induction. Insulin signaling pathway gene products or-thologous to those in Drosophila melanogaster (11, 48) havebeen described from Aedes aegypti (45, 46) and from Anophelesgambiae (47). Insulin signal transduction can induce iNOSexpression in mammalian cells, revealing a connection betweeninsulin signaling and inflammation (3). In Caenorhabditis el-egans, the insulin signaling pathway upregulates antimicrobialresponse genes, suggesting that the link between insulin sig-naling and host defense has been conserved through evolution(40). In this study, we show that P. falciparum GPIs can induceAsNOS expression in vitro and in vivo by activating kinasesassociated with insulin signaling, indicating that both signalingby GPIs and functional relevance of GPIs to innate immunityare evolutionarily conserved.

MATERIALS AND METHODS

Materials. Chemicals, antisera, and other reagents were purchased from thefollowing companies: human serum and human red blood cells from ContinentalServices Group; RPMI 1640, Trizol reagent, and Topo TA cloning kit fromInvitrogen Life Technologies; minimal essential medium (MEM) from Cellgro;hydroxy-2-naphthalenylmethylphosphonic acid-Trisacetoxymethyl ester and geni-stein from Calbiochem; LY294002, wortmannin, PD98059, human insulin, and

monoclonal mouse anti-phospho-ERK antisera from Sigma-Aldrich; bovineserum albumin from Fisher Scientific; polyclonal rabbit antiphospho-INRantisera, polyclonal rabbit anti-phospho-PKB antisera, polyclonal rabbit anti-phospho-JNK/SAPK antisera, horseradish peroxidase (HRP)-conjugated anti-rabbit immunoglobulin G (IgG) from Biosource International; HRP-conjugatedanti-mouse IgG and SuperSignal West Pico chemiluminescent detection kit fromPierce; lactate reagent from Trinity Biotech; Moloney murine leukemia virusreverse transcriptase from Applied Biosystems; Lig’nScribe kit and MEGAscriptT7 transcription kit from Ambion; Effectene Transfection Reagent fromQIAGEN; and protease inhibitor cocktail from Roche Diagnostics.

Isolation of P. falciparum merozoites, purification of PfGPIs, and stimulationof ASE cells. For preparation of merozoites, P. falciparum (FCR-3 strain) wascultured to 30% parasitemia and incubated at 0.2% hematocrit as described (41)to prevent reinvasion of merozoites. The culture was centrifuged at 900 rpm at4°C for 5 min to remove the majority of infected and uninfected red blood cells.During subsequent centrifugation of the supernatant at 1,800 rpm, approxi-mately 50% of the total merozoites and the remaining infected and uninfectedred blood cells formed a layered pellet. The top layer of merozoites was carefullyaspirated. The supernatant was then centrifuged at 3,600 rpm to pellet theremaining merozoites. Collected merozoites were combined and washed withendotoxin-free incomplete RPMI medium and stored at �80°C. Protein-freePfGPIs were isolated as described previously and purified by high-pressure liquidchromatography (28, 41). All preparations of PfGPIs used in our work weretested for endotoxin (28).

Anopheles stephensi rearing, infection with P. falciparum, and provision ofPfGPIs by artificial bloodmeal. Anopheles stephensi Liston were reared at 27°Cand 75% relative humidity. Use of mice and hamsters as bloodmeal sources inthe rearing of A. stephensi is in compliance with all federal guidelines andinstitutional policies. For infection with P. falciparum, 4- to 5-day-old mosquitoeswere allowed to feed on an artificial bloodmeal containing cultured P. falciparum(NF54 strain), filtered human serum, washed human red blood cells and RPMI1640 with HEPES. Mosquitoes fed on the above bloodmeal without parasiteswere used as controls. The use of anonymously collected human blood compo-nents for these procedures is in compliance with all federal guidelines andinstitutional policies. Bloodmeals were provided through 37°C water-jacketedbaudruche membranes. Mosquitoes were allowed to feed for approximately 30min to ensure that the majority of insects were engorged. Midguts of blood-fedmosquitoes were dissected immediately after the 30-min feeding period (0 h) andat various times post-bloodmeal. Total midgut RNA was isolated with Trizolreagent.

AsNOS expression was analyzed by quantitative reverse transcription (RT)-PCR using an ABI Prism 7700 Sequence Detection System (PE Applied Biosys-tems). The amplification efficiencies of AsNOS and S7 ribosomal protein genewere optimized so that AsNOS expression level could be normalized against S7ribosomal protein gene expression by the comparative Ct method as described(14). PCR of AsNOS was performed with 700 nM each primer and 200 nMprobe: AsNOS forward 5�GACCAAACCGGTCATCCTGAT3�; AsNOS reverse5�GGAATCTTGCAGTCAACCATTTC3�; probe 5�CACCGTTCCGTTCGTTCTGGCA3�. For all samples, the reaction was duplicated.

For provision of PfGPIs, 5 �g of PfGPIs were dissolved in 10 �l 80% ethanoland this solution was added to 1 ml of bloodmeal mixture to yield a finalconcentration of 2.5 �M PfGPIs. As a control, 10 �l of 80% ethanol was addedto a separate aliquot of the bloodmeal mixture for feeding to matched mosqui-toes from the same cohort. Midguts were dissected from each group at varioustimes post-blood meal and total RNA was isolated and used for quantitativeRT-PCR of AsNOS expression as described above. For analyses of signalingprotein activation, a third control group of mosquitoes fed only the bloodmealmixture was added. At 0 h and 0.5 h post-blood meal, 30 midguts were dissectedfrom each group for Western blots. Blood was removed by puncturing themidguts with minuten probes and washing twice with phosphate-buffered salinecontaining a protease inhibitor cocktail on ice. Midgut tissues were triturated in80 �l of lysis buffer containing 10 mM Tris-HCl (pH 7.4), 100 mM NaCl, 1 mMEDTA, 1 mM EGTA, 1 mM NaF, 20 mM Na4P2O7, 2 mM Na3VO4, 0.1%sodium dodecyl sulfate (SDS), 0.5% sodium deoxycholate, 1% Triton X-100,10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 60 �g/ml aprotinin, 10 �g/mlleupeptin, and 1 �g/ml pepstatin. Cell debris was removed by centrifugation for10 min at 4°C. Protein concentration was measured by Bradford assay (Bio-Rad)and proteins were mixed with 2� Sample buffer containing 125 mM Tris-HCl(pH 6.8), 10% glycerol, 10% SDS, 0.006% bromophenol blue, and 130 mMdithiothreitol. Equivalent amounts of proteins per lane were electrophoreticallyseparated by SDS-polyacrylamide gel electrophoresis (PAGE). Western blotswere performed as described below.

FIG. 1. Generalized insulin signal transduction pathways. Selectedorthologous receptor and pathway elements from human (Hs), D.melanogaster (Dm) (11, 48), A. aegypti (Aa) (45, 46) and A. gambiae(Ag) (47) are indicated.

VOL. 73, 2005 P. FALCIPARUM SIGNALING OF ANOPHELES NOS INDUCTION 2779


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

Stimulation of A. stephensi cells. Immortalized, embryo-derived A. stephensicell lines, ASE (29, 31) and MSQ43 (generously provided by the Department ofEntomology, Water Reed Army Institute of Research (31), were cultured inmodified MEM containing 5% heat-inactivated fetal bovine serum (E5 medium)at 28°C under 5% CO2. For stimulation, 1 � 106 cells were seeded in 96-well

plates and allowed to grow overnight. Cells were stimulated with merozoites,incomplete medium, human insulin, or HEPES buffer. For kinase inhibitionstudies, cells were pretreated with inhibitors or diluents of inhibitors as controlsfor 30 to 60 min and then stimulated. For Western blots, cells were harvested at5 to 30 min after stimulation, lysed in buffer described above and prepared foranalysis. At 48 h after stimulation, RNA isolation and AsNOS expression anal-yses were performed as described above.

Western blot analyses. Lysates prepared from stimulated cells and midguttissues were centrifuged at 10,000 rpm at 4°C for 10 min to remove insolublematerial. Supernatant proteins were electrophoretically separated by SDS-PAGE and then transferred to nitrocellulose membrane using a semidry blotter(Bio-Rad). The membranes were blocked with Tris-buffered saline (pH 7.4;TBS) containing 3% bovine serum albumin. After washing with TBS containing0.1% Tween 20, the membranes were incubated with a 1:1,000 dilution ofpolyclonal rabbit anti-phospho-INR antisera, a 1:1,000 dilution of polyclonalrabbit anti-phospho JNK/SAPK antisera, a 1:1,000 dilution of polyclonal rabbitanti-phospho-PKB antisera, or a 1:10,000 dilution of monoclonal mouse anti-phospho-ERK antisera for 2 h at room temperature. The INR antiserum recog-nizes three phosphotyrosine residues within the activation loop of the receptor,while the PKB antiserum recognizes a threonine phosphorylated by PDK1 andthe ERK and JNK/SAPK antisera recognize bisphosphorylated (pT/pY) ERKand bisphosphorylated (pT/pY) JNK/SAPK, respectively. The sequences of pep-tides used to generate these antisera are 100% conserved with predicted aminoacid sequences among orthologous proteins from mammals, D. melanogasterand/or other mosquito species, and, as such, were expected to recognize relevantA. stephensi proteins. Membranes were then washed and incubated with a1:250,000 dilution of horseradish peroxidase (HRP)-conjugated anti-rabbit IgGor a 1:50,000 dilution of HRP-conjugated anti-mouse IgG. Peroxidase activitywas detected with the SuperSignal West Pico chemiluminescent detection kit.For JNK/SAPK Western blot, signal intensities were measured using a GS-800calibrated densitometer (Bio-Rad) and normalized against untreated, controlcells.

Measurement of lactate release. ASE cells (1 � 106 per well in 96-well plates)were stimulated with human insulin, PfGPIs, or P. falciparum merozoites for 4 h.Cells stimulated with HEPES buffer, 80% ethanol, or incomplete medium wereused as controls. Culture media were collected after 4 h and lactate level wasmeasured as described (8) using lactate reagent.

Expression of INR, DSOR1, and Akt/PKB gene orthologs in the A. stephensimidgut. Total RNA was isolated from 10 to 15 P. falciparum infected anduninfected A. stephensi midguts at 24 h post-blood meal and from ASE cells asdescribed. First strand cDNA was synthesized using MuLV reverse transcriptase.Fragments of A. stephensi INR, DSOR1, and Akt/PKB genes were amplified withdegenerate primers: INR forward 5� GGNTCGTTNGGTATGGTTTA 3�; INRreverse 5� CGTTCCATTACNCCACCGTC 3�; DSOR1 forward 5� CGGANACGCCGAAATCAC 3�; DSOR1 reverse 5� TTCTANGGCGCNTTCTACAG 3�;Akt/PKB forward 5� TTCACCTTCATCATCCGCGG 3�; Akt/PKB reverse 5�ATCATCTCGTACATGACNACGCC 3�. PCR products were cloned into TopoTA plasmid. Double-stranded, partial sequences of A. stephensi INR, DSOR1,and Akt/PKB gene orthologs were deposited in GenBank and these sequenceswere used to design gene-specific primers for RT-PCR.

Conditions for RT-PCR analyses of the A. stephensi genes were as follows.INR: forward primer 5� GGGTCGTTGGGTATGGTTTA 3� and reverse primer5� CGTTCCATTACCCCACCGTC 3�, 1 cycle of 95°C for 10 min, 35 cycles of94°C 30 s, 53°C 30 s, 72°C 1 min, and 1 cycle of 72°C 10 min. DSOR1: forwardprimer 5� CGGAGACGCCGAAATCAC 3� and reverse primer 5� TTCTATGGCGCGTTCTACAG 3�, 1 cycle of 95°C 10 min, 35 cycles of 94°C 30 s, 56°C 30 s,72°C 30 s, and 1 cycle of 72°C 10 min. Akt/PKB: forward primer 5� TTCACCTTCATCATCCGCGG 3� and reverse primer 5� ATCATCTCGTACATGACGACGCC 3�, 1 cycle of 95°C 10 min, 35 cycles of 94°C 30 s, 59°C 30 s, 72°C 30 s, and1 cycle of 72°C 10 min. Control reactions for each target gene were performed inthe absence of RT to confirm a lack of contaminating genomic DNA.

Suppression of Akt/PKB and DSOR1 mRNA levels in MSQ43 cells by RNAinterference (RNAi) and analyses of AsNOS Induction. To produce templates fordouble-stranded RNA synthesis, 858bp and 550bp of A. stephensi Akt/PKB andDSOR1 were amplified from plasmid clones by PCR. Sense and antisense ssRNAwere synthesized using the Lig’nScribe kit and the MEGAscript T7 transcriptionkit. Double-stranded RNAs of A. stephensi Akt/PKB and DSOR1 were producedas described (30). For each transfection, 2.5 � 106 MSQ43 cells in 10 ml E5medium were transfected with 2 �g of double-stranded RNA using EffecteneTransfection Reagent; control cells were treated in an identical manner butwithout double-stranded RNA (e.g., mock transfection). To examine the reduc-tion of mRNA levels of Akt/PKB and DSOR1 in transfected MSQ43 cells, totalRNA was isolated from 6 h to 5 d posttransfection and RT-PCR was performed

FIG. 2. AsNOS expression is induced by P. falciparum merozoitesand PfGPIs. (A) ASE cells were stimulated with 15.6 merozoites percell (equivalent to 0.25 �M GPIs; n � 3), with 1.56 merozoites per cell(equivalent to 0.025 �M GPIs; n � 2) or incomplete medium as acontrol for 48 h. (B) ASE cells were stimulated with 2.5 �M (n � 3),0.25 �M (n � 5), or 0.025 �M PfGPIs (n � 1) in 80% ethanol or withan equivalent volume of 80% ethanol as a control for 48 h. (C) Midgutswere dissected from A. stephensi immediately after (0 h) and from 1–48h post-blood meal after PfGPI or control feeding for analysis ofAsNOS expression. Data were derived from two replicates of experi-mental and control feeds with two separate cohorts of A. stephensi. Inpanels A to C, AsNOS expression levels were divided by control ex-pression levels to show relative induction. Values represent means �SEs. Data within each treatment or time point versus the control wereanalyzed using the Student t test; P values are shown.

2780 LIM ET AL. INFECT. IMMUN.


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

as described above. For stimulation experiments, culture medium was removedfrom MSQ43 cells at 12 h posttransfection and 1 � 106 MSQ43 cells werereseeded in 96-well plates. Cells were allowed to stabilize for 6 h and thenstimulated with P. falciparum merozoites, incomplete medium, human insulin, orHEPES buffer for 48 h. Following stimulation, AsNOS expression was measuredby quantitative RT-PCR as described above.

DNA sequencing and analysis. DNA sequencing was carried out by using adideoxy dye termination method on an ABI sequencer (Perkin-Elmer Cetus) byUC Davis Sequencing. Nucleotide sequences were compared against standarddatabases and deposited in GenBank for AsINR (AY697415), AsDSOR1(AY697414) and AsAkt/PKB (AY697413).

Statistical analyses. Data from replicated analyses are represented as means� standard errors (SEs) and were analyzed using the Student t test. P values areshown in graphical representations of the data.

RESULTS

AsNOS expression in A. stephensi cells is induced by P. fal-ciparum merozoites and PfGPIs. To identify the P. falciparumfactor(s) responsible for AsNOS induction, we established anin vitro system using A. stephensi embryo-derived ASE andMSQ43 cell lines, tractable and proven models for AsNOSinduction (31). Anopheles stephensi cells were stimulated withP. falciparum merozoites at a ratio of 15.6 or 1.56 merozoitesper cell. These parasite to host cell ratios are comparable tonatural infection in vivo. Specifically, ingestion of 1 to 2 �l ofblood by A. stephensi from a host with a 10% parasitemia,typical in peripheral blood samples from life-threatening P.falciparum infections (5), would lead to midgut infections of250 to 550 parasites per midgut epithelial cell. Our in vitroassays of 15.6 or 1.56 merozoites per mosquito cell would betypical of mosquito feeding on hosts with lower (0.25 to 0.5%)parasitemias (35). At 48 h after treatment, 15.6 and 1.56 mero-zoites per ASE cell induced AsNOS expression 3-fold (P �0.08) and 1.5-fold (P � 0.05), respectively, compared tomedium-treated controls (Fig. 2A). Merozoites induced simi-lar levels of AsNOS expression in MSQ43 cells (see below; Fig.8D and 9D). These results are consistent with levels of AsNOSexpression induced by natural parasite infection in vivo (31)and demonstrated that whole parasites contained AsNOS-in-ducing factors that were recognized by A. stephensi cells.

To determine whether PfGPIs could induce AsNOS expres-sion, ASE cells were stimulated with PfGPIs at concentrationsequivalent to ingestion by A. stephensi of 40,000 or 400,000parasites (1% or 10% parasitemia [5]). Given that each ma-laria parasite contributes 107 GPIs (20), meals of 40,000 or400,000 parasites would contain 0.32 �M or 3.2 �M GPIs.Based on these calculations, we selected 0.025 to 2.5 �MPfGPIs for stimulation. Treatment with 0.25 �M and 2.5 �MPfGPIs induced AsNOS expression in ASE cells 1.7-fold (P �0.01; Fig. 2B) and 5.2-fold (P � 0.03; Fig. 2B), respectively,compared to controls, indicating that biologically relevant lev-els of PfGPIs could induce AsNOS expression in our in vitromodel system. Our results also suggested that P. falciparumcomponents other than GPIs could induce AsNOS expressionsince 15.6 and 1.56 merozoites per ASE cells (equivalent to0.25 �M and 0.025 �M PfGPIs, respectively) induced AsNOSexpression 3-fold and 1.5-fold, respectively, (Fig. 2A), whereas0.25 �M PfGPIs induced AsNOS expression only 1.7-fold and0.025 �M PfGPIs did not induce AsNOS expression relative tocontrols (Fig. 2B).

AsNOS Expression in the A. stephensi midgut epithelium isinduced by PfGPIs. To determine whether PfGPIs could func-tion as an AsNOS-inducing ligand in vivo, we provided 2.5 �MPfGPIs in artificial bloodmeals to two separate cohorts of A.stephensi. Age-matched mosquitoes fed equivalent bloodmealswith only 80% ethanol added were used as controls. At varioustimes after feeding, samples of total RNA from dissected mid-guts were analyzed for AsNOS expression by quantitative RT-PCR. AsNOS expression in midguts from PfGPI-treated A.stephensi was induced 1.4-fold (P � 0.05), 2.5-fold (P � 0.1),and 2-fold (P � 0.03, P � 0.01) relative to controls at 0 h(immediately after feeding) and at 1 h, at 24 h and 48 hpost-blood meal (Fig. 2C), indicating that PfGPIs signalAsNOS induction in vivo.

Putative INR-, DSOR1-, and Akt/PKB-encoding genes areexpressed in A. stephensi cell lines and in the midgut. Based onprevious observations that PfGPIs are insulin-mimetic (7, 52)and that insulin signaling can induce NOS expression in mam-malian cells (3), we hypothesized that parasite GPIs signalAsNOS induction through pathways involving PI3-K, Akt/PKBand MEK/ERK. Initially, to establish the presence of thesesignaling molecules in mosquitoes, we characterized expres-sion of A. stephensi genes encoding a predicted INR, the MEKhomolog DSOR1, and Akt/PKB (Fig. 3). In nonquantitativeRT-PCR assays, we determined that all three genes are ex-pressed in the A. stephensi midgut, in the presence and absenceof P. falciparum infection, and in the ASE cell line (Fig. 4). Wealso determined that ASE cells and the midgut epithelium areresponsive to human insulin. Specifically, 1.7 �M insulin in-duced AsNOS expression in ASE cells maximally to 2.2-foldrelative to controls at 48 h after treatment (Fig. 5A and B),while the same insulin concentration induced AsNOS expres-sion in the midgut to 1.9-fold at 6 h and to 3.6-fold at 36 hpostfeeding relative to controls (not shown).

Plasmodium falciparum merozoites signal A. stephensi cellsthrough non-INR protein tyrosine kinase. To determinewhether P. falciparum signaled AsNOS induction through ac-tivation of an A. stephensi INR, ASE cells were pretreated withHNMPA-(AM)3, an INR tyrosine kinase inhibitor, prior tostimulation. In these and subsequent inhibitor assays, humaninsulin was used as a standard to verify involvement of insulinsignaling pathway components. Plasmodium falciparum mero-zoites were used to stimulate ASE cells to account for potentialinvolvement of signaling factor(s) other than PfGPIs. Pretreat-ment of ASE cells with 0.1 �M or 1 �M HNMPA-(AM)3 ledto, respectively, 45% (P � 0.02) and 32% (P � 0.02) decreasesin insulin-induced AsNOS expression and 92% (P � 0.0002)and 84% (P � 0.0001) decreases in merozoite-induced AsNOSexpression relative to controls (Fig. 6A). These results sug-gested that INR activation is necessary for both insulin andparasite induction of AsNOS. In A. aegypti, HNMPA-(AM)3inhibited bovine insulin signaling in ovaries, although a 90%reduction in signaling required a much higher concentration (1mM) of HNMPA-(AM)3 than was used in our assays (46).

In our assays with HNMPA-(AM)3 and other inhibitors (seebelow), we noted an unexpected pattern of higher AsNOSinduction levels at higher inhibitor concentrations. We deter-mined that treatment of ASE cells with high concentration ofinhibitors induced phosphorylation of a putative JNK/SAPK(Fig. 7), a signaling protein whose activation has been associ-



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

FIG. 3. Alignment of predicted amino acid sequences of AsINR, AsDSOR1, and AsAkt/PKB with predicted orthologous sequences from A.gambiae, A. aegypti, and D. melanogaster. Conserved amino acids are represented in bold type. (A) The predicted amino acid sequence of AsINRgene fragment (AY697415) was aligned with A. gambiae INR (XP_320130), A. aegypti INR (AAB17094), and D. melanogaster INR (NP_524436).(B) The predicted amino acid sequence of AsDSOR1 gene fragment (AY697414) was aligned with A. gambiae DSOR1 (XP_322064) and D.melanogaster DSOR1 (NP_511098). Nucleotide sequences encoding overlined amino acid sequences were used to design gene specific primers forRNAi. (C) The predicted amino acid sequence of AsAkt/PKB gene fragment (AY697413) was aligned with A. gambiae Akt/PKB (EAA03708), A.aegypti Akt/PKB (AAP37655) and D. melanogaster Akt/PKB (CAA81204). Nucleotide sequences encoding overlined amino acid sequences wereused to design gene specific primers for RNAi.

2782


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

ated with stress-related induction of NOS expression and NOsynthesis in mammalian cells (10, 23). Therefore, it appearsthat JNK/SAPK activation in A. stephensi cells may account forslightly higher AsNOS induction levels at higher inhibitor con-centrations.

To verify INR activation by insulin and P. falciparum mero-zoites, anti-phospho-INR Western blots were performed onlysates prepared from ASE cells stimulated with 1.7 �M hu-man insulin or 15.6 merozoites per cell for 5, 15, and 30 min.A cross-reacting band of 98 kDa, a molecular mass consistentwith that reported for the A. aegypti INR � subunit (45), wasdetected in insulin-stimulated cells, but not in merozoite-stim-ulated cells (Fig. 6B). Plasmodium falciparum signaling was notassociated with INR activation, a result inconsistent with inhi-bition by HNMPA-(AM)3. However, HNMPA-(AM)3 can in-hibit PTKs other than those associated with INRs, includingnonreceptor PTKs (50) and p59 hck, a src family PTK, whichhas been implicated in PfGPI signaling in mammalian cells (58,59).

To determine whether PTKs other than the INR tyrosinekinase were involved in parasite signaling of AsNOS induction,we pretreated ASE cells with genistein, a PTK inhibitor that isinactive against INR (1), prior to parasite stimulation. Pre-treatment with 10�4 or 10�3 �M genistein reduced AsNOSexpression by 47% (P � 0.04) and 21% (P � 0.01), respec-tively, in merozoite-stimulated cells relative to controls (Fig.6C), suggesting that A. stephensi non-INR PTKs are involved inparasite signaling.

Plasmodium falciparum merozoites activate ERK and re-quire DSOR1 for AsNOS induction. To determine whether P.falciparum signals AsNOS induction through MEK activation,ASE cells were pretreated with PD98059, a MEK inhibitor,prior to stimulation. Pretreatment of ASE cells with 0.04 �Mor 4 �M PD98059, respectively, reduced AsNOS expression by17% (P � 0.03) and 22% (P � 0.09) in insulin-stimulated cellsand by 75% (P � 0.002) and 60% (P � 0.001) in merozoite-

stimulated cells relative to controls (Fig. 8A), suggesting thatMEK activation is critical to parasite and insulin induction ofAsNOS. To verify this conclusion, we first examined activationof ERK by Western blotting in cells stimulated by merozoitesand human insulin, then used RNAi-dependent gene silencingto determine whether A. stephensi DSOR1, the putative acti-vator of ERK, was necessary for parasite- and insulin-inducedAsNOS expression.

Anti-phospho-ERK Western blots were performed on pro-tein samples prepared from ASE cells stimulated with 1.7 �Mhuman insulin or 15.6 merozoites per cell for 5 or 15 min. Asingle cross-reacting band of 50 kDa, somewhat larger than the44-kDa Drosophila melanogaster ortholog (4), was detected insamples collected from insulin- and parasite-stimulated cells(Fig. 8B). As expected, insulin induced ERK phosphorylationwithin 5 min in ASE cells relative to controls (Fig. 8B, lane 2versus 3) and PD98059 pretreatment reduced activation ofERK (Fig. 8B, lane 2 versus 5) suggesting the specificity of thisinhibitor. Unexpectedly, merozoite stimulation reduced ERKphosphorylation relative to controls after 5 min (Fig. 8B, lane

Fig. 4. Anopheles stephensi INR, DSOR1, and Akt/PKB genes areexpressed in ASE cells and in the midgut epithelium. Fragments of A.stephensi INR, DSOR1, and Akt/PKB genes were amplified from cDNAsamples using gene-specific primers and cycling conditions described inthe text. Control reactions were performed on RNA in the absence ofreverse transcriptase to confirm a lack of contaminating genomicDNA. Results show amplifications from cDNA and RNA from A.stephensi midgut tissue 24 h postinfection with P. falciparum (Inf mid-gut), amplifications from cDNA and RNA prepared from A. stephensimidgut tissue 24 h after an uninfected bloodmeal (Uninf midgut),amplifications from cDNA and RNA prepared from ASE cells, andcontrol amplifications without added cDNA or RNA (no-templatecontrol, NTC). A representative figure of RT-PCR results is shown.

FIG. 5. AsNOS expression is induced maximally in ASE cells by 1.7�M human insulin at 48 h after stimulation. (A) ASE cells werestimulated with various concentrations of human insulin in HEPESbuffer (n � 3 for 0.17 �M, n � 7 for 0.85 �M, n � 8 for 1.7 �M, n �3 for 3.4 �M, n � 4 for 17 �M) or HEPES buffer only as a control for48 h. (B) ASE cells were stimulated with 1.7 �M human insulin orHEPES buffer for 2- 48 h (n � 4 for 2 h, n � 4 for 6 h, n � 4 for 12 h,n � 4 for 24 h, n � 4 for 36 h, n � 9 for 48 h). For panels A and B,AsNOS expression levels were measured by quantitative RT-PCR andwere divided by control expression levels to show relative induction.Values represent means � SEs.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

1 versus 3); identical results were obtained at 15 min afterstimulation (not shown). Further, PD98059 pretreatment re-duced ERK phosphorylation in response to merozoites to anundetectable level (Fig. 8B, lane 1 versus 4). Because PD98059also inhibited parasite induction of AsNOS (Fig. 8A), we inferthat some level of ERK phosphorylation is required for para-site induction of AsNOS.

To determine whether DSOR1, the predicted upstream ac-tivator of A. stephensi ERK, was necessary for P. falciparumand insulin induction of AsNOS, we silenced DSOR1 withRNAi in MSQ43 cells prior to stimulation. In DSOR1 double-stranded RNA-transfected cells, DSOR1 mRNA levels wereundetectable from 6 h to 5 d posttransfection (Fig. 8C). ForAsNOS induction assays, MSQ43 cells were stimulated at 18 hposttransfection with 15.6 merozoites per cell, 1.7 �M humaninsulin, medium or HEPES buffer for 48 h. DSOR1 RNAireduced merozoite induction of AsNOS by 87% relative tomock transfected cells (P � 0.04), while insulin induction wasreduced by only 41% relative to mock transfected cells (P �0.007; Fig. 8D). These data mirrored levels of AsNOS induc-tion observed following pretreatment with PD98059 (Fig. 8A)and confirmed that DSOR1 and its likely impact on ERKactivation are necessary for P. falciparum induction of AsNOS.

Plasmodium falciparum merozoites require Akt/PKB forAsNOS induction. To determine whether P. falciparum signalsAsNOS induction through activation of A. stephensi PI3-K andAkt/PKB, ASE cells were pretreated with LY294002 and wort-mannin, PI3-K inhibitors, prior to stimulation with humaninsulin or P. falciparum merozoites. Pretreatment of ASE cellswith 1 �M or 20 �M LY294002 had no effect on the insulin-mediated induction of AsNOS compared to controls (Fig. 9A).Similar results were obtained when cells were pretreated with1 � 10�4 to 10 �M wortmannin prior to insulin stimulation(not shown). In contrast to our results, 10�2 to 100 �MLY294002 reduced insulin-stimulated steroidogenesis in A. ae-gypti ovary cells by 60% (46), suggesting that insulin signalingvaries among mosquito cell types. Treatment with 1 �M or 20�M LY294002 reduced AsNOS expression by 65% (P �0.0001) and 51% (P � 0.0002), respectively, in merozoite-stimulated cells relative to controls (Fig. 9A), suggesting thatPI3-K activity is critical to parasite induction of AsNOS. Toverify this conclusion, we examined activation of Akt/PKB byWestern blotting in ASE cells stimulated by human insulin andmerozoites, then used RNAi-dependent gene silencing to de-

FIG. 6. Plasmodium falciparum merozoites signal A. stephensi cellsand AsNOS induction through non-INR PTK. (A) ASE cells werepretreated for 1 h with HNMPA-(AM)3, an inhibitor of INR activa-tion, dissolved in dimethyl sulfoxide or with an equivalent volume ofdimethyl sulfoxide as a mock pretreatment (0 �M). After pretreat-ment, cells were stimulated with 1.7 �M human insulin in HEPESbuffer (n � 7 for 0.1 �M inhibitor, n � 4 for 1 �M inhibitor) or with15.6 P. falciparum merozoites per cell in incomplete RPMI 1640 me-dium (n � 2 for both inhibitor concentrations) for 48 h prior to analysisof AsNOS expression. (B) Anti-phospho-INR Western blot of ASEcells stimulated with 1.7 �M human insulin or with 15.6 P. falciparummerozoites per cell for 5–30 min. Unstimulated ASE cells (NS) wereused as controls (incomplete medium- and HEPES buffer-stimulatedcells yielded identical results, not shown). Anti-phospho-INR antiseradid not cross-react with P. falciparum merozoites (not shown). Thefigure shown is representative of replicated Western blots from inde-pendent experiments. (C) ASE cells were pretreated for 1 h withgenistein, a PTK inhibitor, dissolved in dimethyl sulfoxide or with anequivalent volume of dimethyl sulfoxide as a mock pretreatment (0�M). After pretreatment, cells were stimulated with P. falciparummerozoites as in A (n � 3 for 10�3 �M inhibitor, n � 2 for 10�4 �Minhibitor). For panels A and C, AsNOS expression levels were dividedby mock pretreatment expression levels to show relative effects of theinhibitor on parasite induction of AsNOS expression. Values representmeans � SEs. Data (inhibitor versus mock pretreatment) were ana-lyzed using the Student t test; P values are shown.

FIG. 7. Treatment with kinase inhibitors induces JNK/SAPK phos-phorylation in ASE cells. ASE cells were treated with 100 �MLY294002 for 30 min, 100 �M genistein for 1 h, 100 �M HNMPA-(AM)3 for 1 h or 40 �M PD98059 for 30 min. After stimulation, cellswere collected and lysed as described in the text and applied to anti-phospho-JNK/SAPK Western blots. Signal intensity of each band isshown. The figure shown is representative of replicated Western blotsfrom independent experiments.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

termine whether A. stephensi Akt/PKB was necessary for par-asite-induced AsNOS expression.

Anti-phospho-Akt Western blots were performed on lysatesprepared from ASE cells stimulated with 1.7 �M human insu-lin or 15.6 merozoites per cell for 10 or 30 min. A 56 kDa bandand a slightly slower migrating band of 58 kDa cross-reactedwith the anti-phospho-Akt antisera (Fig. 9B). We suggest thatthis 58 kDa band is the result of a mass shift from multiplephosphorylation events as has been described for mouse Akt/PKB (17). As predicted from results with LY294002 and wort-mannin (see above), insulin did not induce phosphorylation ofa putative Akt/PKB (Fig. 9B, lane 1 versus 2 and lane 4 versus5) while merozoites induced phosphorylation of Akt/PKB rel-ative to controls at 10 and 30 min (Fig. 9B, arrow, lane 1 versus3 and lane 4 versus 6).

To determine whether Akt/PKB was necessary for P. falci-parum induction of AsNOS, we silenced Akt/PKB with RNAi inMSQ43 cells prior to stimulation. In Akt/PKB double-strandedRNA-transfected cells, Akt/PKB mRNA levels were undetect-able from 6 h to 5 d posttransfection (Fig. 9C). For AsNOSinduction assays, MSQ43 cells were stimulated at 18 h post-transfection with 15.6 merozoites per cell or medium for 48 h.As expected, Akt/PKB RNAi reduced merozoite induction ofAsNOS by 65% relative to mock transfected cells (P � 0.00002;Fig. 9D), indicating a prominent role for Akt/PKB activation inparasite induction of AsNOS.

Plasmodium falciparum merozoites and PfGPIs are not in-sulin-mimetic to A. stephensi cells. Based on observations thatPfGPIs are insulin-mimetic to mammalian cells (7, 52), wehypothesized that merozoites and PfGPIs may be perceived asinsulin-mimetic to A. stephensi cells. Because insulin induceslactate release rather than glucose uptake in D. melanogasterKc cells (8), we assayed lactate release in ASE cells stimulatedwith 0.17, 1.7, or 17 �M human insulin or HEPES buffer from1 to 8 h after treatment. At 4 h after treatment with 1.7 �Minsulin, lactate release relative to control cells was maximal andnearly identical to the 1.3-fold induction reported for Kc cells(not shown) (8). Although these assays confirmed that ASEcells were similar to Kc cells in response to insulin, stimulationwith PfGPIs and merozoites failed to induce significant lactaterelease relative to control treatments (Fig. 10). Our data indi-cate that, while merozoites and PfGPIs activate mosquito cellsthrough signaling proteins that are associated with insulin sig-nal transduction pathways, neither merozoites nor PfGPIs areinsulin-mimetic to A. stephensi cells.

PfGPIs signal the A. stephensi midgut epithelium throughactivation of Akt/PKB and ERK. Based on our results that P.

FIG. 8. Plasmodium falciparum merozoites activate ERK and re-quire DSOR1 for AsNOS induction. (A) ASE cells were pretreated for30 min with 0.04 �M or 4 �M PD98059, a MEK inhibitor, dissolved indimethyl sulfoxide or with an equivalent volume of dimethyl sulfoxideas a mock pretreatment (0 �M). After pretreatment, cells were stim-ulated with 1.7 �M human insulin in HEPES buffer (n � 8 for 0.04 �Minhibitor, n � 5 for 4 �M inhibitor) or with 15.6 P. falciparum mero-zoites per cell in incomplete RPMI 1640 medium (n � 2 for 0.04 �Minhibitor, n � 3 for 4 �M inhibitor) for 48 h. AsNOS expression levelswere divided by mock pretreatment expression levels to show relativeeffects of the inhibitor on insulin (line) or parasite (bars) induction ofAsNOS. Values represent means � SEs. Data (inhibitor versus mockpretreatment) were analyzed using the Student t test; P values areshown. (B) Anti-phospho-ERK Western blot of ASE cells pretreatedwith 0.4 �M PD98059 or dimethyl sulfoxide then stimulated with 1.7�M human insulin or with 15.6 P. falciparum merozoites per cell for 5min. Unstimulated ASE cells were used as controls (incomplete me-dium- and HEPES buffer-stimulated cells yielded identical results, notshown). Anti-phospho-ERK antisera did not cross-react with P. falci-parum merozoites (not shown). The figure shown is representative ofreplicated Western blots from independent experiments. (C) RT-PCR

of DSOR1 expression levels following RNAi of A. stephensi DSOR1 inmock control (C) and double-stranded RNA-transfected (T) MSQ43cells from 6 h to 5 d posttransfection. Ribosomal S7 protein geneRT-PCR amplification was used as a loading control and to confirmsample integrity. (D) AsNOS expression levels in mock- and DSOR1double-stranded RNA-transfected cells were divided by expressionlevels in matched transfected cells stimulated with HEPES buffer orincomplete RPMI 1640 medium to show relative AsNOS induction bytreatment with 1.7 �M insulin (n � 3) or with 15.6 P. falciparummerozoites per cell (n � 2), respectively. Values represent means �SEs. Data within each treatment were analyzed using the Student ttest; P values are shown.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

falciparum signals AsNOS induction in vitro through DSOR1/ERK- and PI3-K/Akt-dependent pathways, we reasoned thatPfGPI signaling in vivo would recapitulate these pathway as-sociations. To address this question, two separate cohorts offemale A. stephensi were provided with a bloodmeal supple-mented with 2.5 �M PfGPIs in 80% ethanol, a bloodmealsupplemented with an equivalent volume of 80% ethanol, oran unmodified bloodmeal. Midguts were dissected at 0 h (im-mediately after) and at 0.5 h after feeding and prepared forWestern blots of ERK and Akt/PKB phosphorylation. At bothtime points, PfGPIs induced phosphorylation of ERK and Akt/PKB in the A. stephensi midgut epithelium relative to bothcontrols (Fig. 11) indicating that merozoites (Fig. 8B and 9B)and PfGPIs signal through activation of A. stephensi Akt/PKBand ERK.

DISCUSSION

Current efforts to control malaria parasite transmission in-clude the development of genetically modified Anopheles mos-quitoes that exhibit enhanced refractoriness to Plasmodiumspp. The search for antiparasite effectors has identified somepromising targets in the immune-responsive mosquito, includ-ing melanotic encapsulation and the synthesis of toxic reactiveoxygen species, antimicrobial peptides, and NOS, as well asputative nonself recognition factors and signaling cascades im-plicated in transducing responses to bacterial challenge (re-

FIG. 9. Plasmodium falciparum merozoites activate Akt and re-quire Akt for AsNOS induction. (A) ASE cells were pretreated for 1 hwith 1 �M or 20 �M LY294002, a PI3-K inhibitor, dissolved in 100%ethanol or with an equivalent volume of ethanol as a mock pretreat-ment (0 �M). After pretreatment, cells were stimulated with 1.7 �Mhuman insulin in HEPES buffer (n � 2 for both concentrations ofinhibitors) or with 15.6 P. falciparum merozoites per cell in incompleteRPMI 1640 medium (n � 3 for both concentrations of inhibitors) for48 h. AsNOS expression levels were divided by mock pretreatmentexpression levels to show relative effects of the inhibitor on insulin(line) or parasite (bars) induction of AsNOS. Values represent means� SEs. Data (inhibitor versus mock pretreatment) were analyzed usingthe Student t test; P values are shown. (B) Anti-phospho-Akt/PKBWestern blot of ASE cells stimulated with 1.7 �M human insulin orwith 15.6 P. falciparum merozoites per cell for 10 min or 30 min.Unstimulated ASE cells were used as controls (incomplete medium-and HEPES buffer-stimulated cells yielded identical results, notshown). Anti-phospho-Akt/PKB antisera did not cross-react with P.falciparum merozoites (not shown). Arrow indicates slower migrating,fully phosphorylated putative A. stephensi Akt/PKB. The figure shown

FIG. 10. Plasmodium falciparum merozoites and PfGPIs do notmimic the effects of insulin on lactate release by A. stephensi cells. ASEcells were stimulated with 2.5 �M PfGPI in 80% ethanol (n � 2), with15.6 P. falciparum merozoites per cell in incomplete medium (n � 2),or with 1.7 �M human insulin in HEPES buffer (n � 2) for 4 h.Matched control cells for each replicate were stimulated with equiva-lent volumes of HEPES buffer, incomplete medium or 80% ethanol.Lactate levels are represented as means � SEs. Data within eachtreatment were analyzed using the Student t test.

is representative of replicated Western blots from independent exper-iments. C, RT-PCR of Akt expression levels following RNAi of A.stephensi Akt in mock control (C) and double-stranded RNA-trans-fected (T) MSQ43 cells from 6 h to 5 d posttransfection. Ribosomal S7protein gene RT-PCR amplification was used as a loading control andto confirm sample integrity. (D) AsNOS expression levels in mock- andAkt double-stranded RNA-transfected cells were divided by expressionlevels in matched transfected cells stimulated with incomplete RPMI1640 medium to show relative AsNOS induction by treatment with 15.6P. falciparum merozoites per cell (n � 2), respectively. Values repre-sent means � SEs. Data within each treatment were analyzed using theStudent t test; P values are shown.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

viewed in reference 17). Although the repertoire of reportedresponses to malaria parasite infection and invasion may beconfounded by resident microbial flora in Anopheles (56), sig-nificant expression of immune genes in antibiotic-treated, par-asite-infected mosquitoes has consistently predicted the exis-tence of Plasmodium-specific mechanisms of gene induction(34, 44). However, the identity of parasite-derived signalingfactors capable of inducing specific responses in Anophelescells has remained unknown until completion of the work de-scribed here.

In mammalian cells, PfGPIs are sufficient to account for themost notable effects of P. falciparum (53, 54, 57). We havedemonstrated that 2.5 �M PfGPIs can induce AsNOS expres-sion 5-fold in A. stephensi cells, results that are consistentwith inductions of NO synthesis of 1.5-fold and 4-fold by 1 �Mand 10 �M PfGPIs, respectively, in mouse macrophages (57).During parasite infection, induction of AsNOS expression inthe midgut is biphasic, with 2-fold inductions at 6 h, 36 h, and48 h after feeding (31). Provision of PfGPIs in the bloodmealalso induced a biphasic pattern of AsNOS expression in themosquito midgut (Fig. 2C), with the earlier initial inductioncompared to natural infection (0 to 1 h versus 6 h) likely dueto the more immediate availability of a larger concentration ofPfGPIs in the midgut after feeding. The similarity of AsNOSinduction patterns following feeding on PfGPIs and naturalinfection suggests that parasite GPIs are an important signalfor AsNOS induction prior to (�24 h) and during parasiteinvasion (24 to 48 h) of the midgut.

Plasmodium falciparum and PfGPIs signal A. stephensi cellsthrough insulin-responsive PI3-K/Akt and DSOR1/ERK. Inmouse macrophages, PfGPIs induced rapid phosphorylation ofERK2, although PD98059 pretreatment suggested that ERKsignaling was not involved in induction of macrophage NOsynthesis by PfGPIs (63). In our studies with A. stephensi,stimulation of different target cells (ASE cells in vitro andmidgut cells in vivo) with both whole parasites and PfGPIsrevealed important information about ERK signaling. Al-though some level of ERK phosphorylation is necessary forAsNOS induction by P. falciparum in vitro (Fig. 8A and B),ERK phosphorylation by PfGPIs in vivo (Fig. 11) was moreprominent.

In mammalian cells, nonphosphorylated, monophosphory-lated, and fully bisphosphorylated forms of ERK2 are detect-

able (9, 60), with the balance of these forms in different celltypes maintained by the opposing action of MEK1 and multi-ple protein phosphatases. Based on these observations, it wasproposed that diverse signals may be integrated at the phos-phatase level, rather than the kinase level, to dictate the cel-lular responses to external stimuli (61). Indeed, the activity ofmonophosphorylated ERK2 was determined to be intermedi-ate to that of unphosphorylated and fully active bisphosphory-lated ERK2 (62). The reduction in ERK phosphorylationwithin 5 min of parasite stimulation, together with our knock-out results, suggests that an ERK pool with diminished levelsof bisphosphorylated ERK and perhaps higher levels of func-tional, monophosphorylated ERK drives the MEK-dependentcellular response to P. falciparum in ASE cells. As with ERKactivation, we noted subtle differences in Akt activation in A.stephensi cells in vitro and in vivo. Specifically, PfGPI-inducedAkt/PKB phosphorylation in midgut cells (Fig. 11) did notresult in the protein mass shift observed following P. falcipa-rum stimulation of ASE cells (Fig. 9B). While some of thesedifferences may be attributable to physiological differences be-tween ASE and midgut cells, they also suggest that multiplesignals from whole parasites, including at least PfGPIs andperhaps hemozoin (27, 38, 55) and others, contribute toAsNOS induction.

In addition to mimicry of insulin, PfGPIs signal mammaliancells through protein kinase Cε, PTK p59 hck, and nuclearfactor-B/c-rel (57, 59). Additional data indicate that Toll-likereceptors are activated by malaria parasites (28), a signalingmechanism that is well established for GPIs of the parasiticprotozoan Trypanosoma cruzi (2, 6). Although our data indi-cate involvement of kinases associated with insulin signaling inAsNOS induction by P. falciparum and PfGPIs, these agentsare not insulin-mimetic to A. stephensi cells. We conclude thatactivation of A. stephensi Akt/PKB and DSOR1/ERK by P.falciparum merozoites and PfGPIs is likely due to activation ofpathways that share signaling components with insulin signaltransduction pathways. Akt/PKB, for example, phosphorylatesmore than 50 known mammalian substrates associated withcell growth, defense, survival, and metabolism (25), suggestingthat malaria parasite activation of A. stephensi Akt/PKB per-turbs multiple pathways and physiological processes in A. ste-phensi cells. Further, inhibition of parasite signaling of AsNOSinduction by genistein, a PTK inhibitor that is inactive againstthe INR (1), indicates that other PTKs, perhaps includingrepresentatives of the src family, are involved in parasite sig-naling of AsNOS induction. Therefore, mimicry of insulin byPfGPIs appears to be restricted to mammalian hosts of P.falciparum, but the conservation of PfGPIs as a prominentparasite-derived signal of innate immunity can now be ex-tended to include Anopheles mosquitoes. This novel findingsignificantly increases the likelihood of identifying additionalsignaling pathways and downstream effectors associated withmosquito resistance to parasite development.

In general, the context of parasite signaling of AsNOS in-duction in the mosquito midgut is likely to be complicated bysimultaneous exposure of midgut cells to dynamic concentra-tions of parasite-derived factors, mosquito-derived factors andexogenous factors in mammalian blood ingested during feed-ing. The latter factors include human insulin, which can induceAsNOS expression after feeding, and human transforming

FIG. 11. Plasmodium falciparum GPIs rapidly activate putative A.stephensi Akt/PKB and ERK in the mosquito midgut epithelium. Anti-phospho-Akt and anti-phospho-ERK Western blots of A. stephensimidgut tissue proteins immediately after (0 h) and 0.5 h after feedingon artificial bloodmeals with PfGPIs, 80% ethanol, or nothing added(blood only). The figure shown is representative of replicated Westernblots from independent experiments.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

growth factor -�1, which alters AsNOS induction and parasiteinfection in A. stephensi (31). We are challenged, therefore, todetermine whether these factor(s) synergize or interfere withother signals, endogenous and exogenous in the blood-filledmidgut, to mediate anti-parasite immunity in the mosquito.Cross talk among pathways of interest to us is well known frommammalian systems. For example, transforming growth factor�1-mediated growth inhibition is dependent on activation ofIRS proteins (26), while the effects of transforming growthfactor �1 are regulated by PI3-K/Akt at the level of directinteraction between Akt and Smad3 (12, 43). Hence, an un-derstanding of the complexity of the signaling milieu in themosquito, which exhibits remarkable conservation with that inthe mammalian host, is necessary for the success of efforts tomanipulate signaling factors, pathways or effector genes toenhance mosquito refractoriness.

ACKNOWLEDGMENTS

This work was supported by Public Health Service grants AI41027and AI60664 from the National Institute of Allergy and InfectiousDiseases at the National Institutes of Health.

We thank Jackie Williams and Megan Dowler of the Walter ReedArmy Institute of Research for their assistance with and provision ofmaterials for P. falciparum infections.

REFERENCES

1. Abler, A., J. A. Smith, P. A. Randazzo, P. L. Rothenberg, and L. Jarett. 1992.Genistein differentially inhibits postreceptor effects of insulin in rat adipo-cytes without inhibiting the insulin receptor kinase. J. Biol. Chem. 267:3946–3951.

2. Almeida, I. C., M. M. Camargo, D. O. Procopio, L. S. Silva, A. Mehlert, L. R.Travassos, R. T. Gazzinelli, and M. A. Ferguson. 2000. Highly purifiedglycosylphosphatidylinositols from Trypanosoma cruzi are potent proinflam-matory agents. EMBO J. 19:1476–1485.

3. Begum, N., L. Ragolia, J. Rienzie, M. McCarthy, and N. Duddy. 1998.Regulation of mitogen-activated protein kinase phosphatase-1 induction byinsulin in vascular smooth muscle cells. Evaluation of the role of the nitricoxide signaling pathway and potential defects in hypertension. J. Biol. Chem.273:25164–25170.

4. Biggs, W. H., 3rd, and S. L. Zipursky. 1992. Primary structure, expression,and signal-dependent tyrosine phosphorylation of a Drosophila homolog ofextracellular signal-regulated kinase. Proc. Natl. Acad. Sci. USA 89:6295–6299.

5. Bruce-Chwatt, L. J. 1971. Essential Malariology. John Wiley and Sons, NewYork, N.Y.

6. Campos, M. A., I. C. Almeida, O. Takeuchi, S. Akira, E. P. Valente, D. O.Procopio, L. R. Travassos, J. A. Smith, D. T. Golenbock, and R. T.Gazzinelli. 2001. Activation of toll-like receptor-2 by glycosylphosphatidy-linositol anchors from a protozoan parasite. J. Immunol. 167:416–423.

7. Caro, H. N., N. A. Sheikh, J. Taverne, J. H. Playfair, and T. W. Rademacher.1996. Structural similarities among malaria toxins insulin second messengers,and bacterial endotoxin. Infect Immun. 64:3438–3441.

8. Ceddia, R. B., G. J. Bikopoulos, A. J. Hilliker, and G. Sweeney. 2003. Insulinstimulates glucose metabolism via the pentose phosphate pathway in Dro-sophila Kc cells. FEBS Lett. 555:307–310.

9. Cha, H., and P. Shapiro. 2001. Tyrosine-phosphorylated extracellular signal-regulated kinase associates with the Golgi complex during G2/M phase of thecell cycle: evidence for regulation of Golgi structure. J. Cell Biol. 153:1355–1367.

10. Chan, E. D., and D. W. Riches. 2001. IFN-gamma � LPS induction of iNOSis modulated by ERK, JNK/SAPK, and p38(mapk) in a mouse macrophagecell line. Am. J. Physiol. Cell Physiol. 280:C441–450.

11. Clemens, J. C., C. A. Worby, N. Simonson-Leff, M. Muda, T. Maehama, B. A.Hemmings, and J. E. Dixon. 2000. Use of double-stranded RNA interferencein Drosophila cell lines to dissect signal transduction pathways. Proc. Natl.Acad. Sci. USA 97:6499–6503.

12. Conery, A. R., Y. Cao, E. A. Thompson, C. M. Townsend, Jr., T. C. Ko, andK. Luo. 2004. Akt interacts directly with Smad3 to regulate the sensitivity toTGF-beta induced apoptosis. Nat. Cell Biol. 6:366–372.

13. Conner, E. M., S. Aiko, M. Fernandez, H. D. Battarbee, L. Gray, and M. B.Grisham. 2000. Duration of the hemodynamic effects of N(G)-nitro-L-argi-nine methyl ester in vivo. Nitric Oxide. 4:85–93.

14. Crampton, A., and S. Luckhart. 2001. The role of As60A, a TGF-beta

homolog, in Anopheles stephensi innate immunity and defense against Plas-modium infection. Infection, Genetics Evolution. 1:131–141.

15. Crampton, A. L., and S. Luckhart. 2001. Isolation and characterization ofAs60A, a transforming growth factor-beta gene, from the malaria vectorAnopheles stephensi. Cytokine 13:65–74.

16. del Carmen Rodriguez, M., P. Gerold, J. Dessens, K. Kurtenbach, R. T.Schwartz, R. E. Sinden, and G. Margos. 2000. Characterisation and expres-sion of pbs25, a sexual and sporogonic stage specific protein of Plasmodiumberghei. Mol. Biochem. Parasitol. 110:147–159.

17. Dimopoulos, G. 2003. Insect immunity and its implication in mosquito-malaria interactions. Cell Microbiol. 5:3–14.

18. Dimopoulos, G., A. Richman, H. M. Muller, and F. C. Kafatos. 1997. Mo-lecular immune responses of the mosquito Anopheles gambiae to bacteriaand malaria parasites. Proc. Natl. Acad. Sci. USA 94:11508–11513.

19. Dimopoulos, G., D. Seeley, A. Wolf, and F. C. Kafatos. 1998. Malaria infec-tion of the mosquito Anopheles gambiae activates immune-responsive genesduring critical transition stages of the parasite life cycle. EMBO J. 17:6115–6123.

20. Ferguson, M. A., J. S. Brimacombe, S. Cottaz, R. A. Field, L. S. Guther, S. W.Homans, M. J. McConville, A. Mehlert, K. G. Milne, J. E. Ralton, and et al.1994. Glycosyl-phosphatidylinositol molecules of the parasite and the host.Parasitology 108 Suppl.:S45–54.

21. Frick, W., A. Bauer, J. Bauer, S. Wied, and G. Muller. 1998. Insulin-mimeticsignalling of synthetic phosphoinositolglycans in isolated rat adipocytes. Bio-chem J. 336:163–181.

22. Gowda, D. C. 2002. Structure and activity of glycosylphosphatidylinositolanchors of Plasmodium falciparum. Microbes Infect. 4:983–990.

23. Guan, Z., S. Y. Buckman, L. D. Springer, and A. R. Morrison. 1999. Bothp38alpha(MAPK) and JNK/SAPK pathways are important for induction ofnitric-oxide synthase by interleukin-1beta in rat glomerular mesangial cells.J. Biol. Chem. 274:36200–36206.

24. Han, Y. S., and C. Barillas-Mury. 2002. Implications of Time Bomb modelof ookinete invasion of midgut cells. Insect Biochem. Mol. Biol. 32:1311–1316.

25. Hanada, M., J. Feng, and B. A. Hemmings. 2004. Structure, regulation andfunction of PKB/AKT-a major therapeutic target. Biochim. Biophys. Acta.1697:3–16.

26. Huang, S. S., S. M. Leal, C. L. Chen, I. H. Liu, and J. S. Huang. 2004.Cellular growth inhibition by TGF-beta1 involves IRS proteins. FEBS Lett.565:117–121.

27. Jaramillo, M., D. C. Gowda, D. Radzioch, and M. Olivier. 2003. Hemozoinincreases IFN-gamma-inducible macrophage nitric oxide generation throughextracellular signal-regulated kinase- and NF-kappa B-dependent pathways.J. Immunol. 171:4243–4253.

28. Krishnegowda, G., A. M. Hajjar, J. Zhu, E. J. Douglass, S. Uematsu, S.Akira, A. S. Woods, and D. C. Gowda. Induction of proinflammatory re-sponses in macrophages by the glycosylphosphatidylinositols (GPIs) of Plas-modium falciparum: Cell signaling receptors, GPI structural requirement,and regulation of GPI activity. J. Biol. Chem., in press.

29. Kurtti, T. J., and O. G. Munderloh. 1989. Advances in the definition ofculture media for mosquito cells, vol. 1. CRC Press, Inc, Boca Raton, FL.

30. Levashina, E. A., L. F. Moita, S. Blandin, G. Vriend, M. Lagueux, and F. C.Kafatos. 2001. Conserved role of a complement-like protein in phagocytosisrevealed by double-stranded RNA knockout in cultured cells of the mos-quito, Anopheles gambiae. Cell. 104:709–718.

31. Luckhart, S., A. L. Crampton, R. Zamora, M. J. Lieber, P. C. Dos Santos,T. M. Peterson, N. Emmith, J. Lim, D. A. Wink, and Y. Vodovotz. 2003.Mammalian transforming growth factor beta1 activated after ingestion byAnopheles stephensi modulates mosquito immunity. Infect. Immun. 71:3000–3009.

32. Luckhart, S., and K. Li. 2001. Transcriptional complexity of the Anophelesstephensi nitric oxide synthase gene. Insect Biochem. Mol. Biol. 31:249–256.

33. Luckhart, S., and R. Rosenberg. 1999. Gene structure and polymorphism ofan invertebrate nitric oxide synthase gene. Gene 232:25–34.

34. Luckhart, S., Y. Vodovotz, L. Cui, and R. Rosenberg. 1998. The mosquitoAnopheles stephensi limits malaria parasite development with inducible syn-thesis of nitric oxide. Proc. Natl. Acad. Sci. USA 95:5700–5705.

35. Markell, E. K., D. T. John, W. A. Krotoski. 1999. Medical Parasitology, 8thed. W.B. Saunders Company, Philadelphia, PA.

36. Mellouk, S., S. L. Hoffman, Z. Z. Liu, P. de la Vega, T. R. Billiar, and A. K.Nussler. 1994. Nitric oxide-mediated antiplasmodial activity in human andmurine hepatocytes induced by gamma interferon and the parasite itself:enhancement by exogenous tetrahydrobiopterin. Infect. Immun. 62:4043–4046.

37. Moran, P., and I. W. Caras. 1994. Requirements for glycosylphosphatidyli-nositol attachment are similar but not identical in mammalian cells andparasitic protozoa. J. Cell Biol. 125:333–343.

38. Mordmuller, B., F. Turrini, H. Long, P. G. Kremsner, and P. Arese. 1998.Neutrophils and monocytes from subjects with the Mediterranean G6PDvariant: effect of Plasmodium falciparum hemozoin on G6PD activity, oxi-dative burst and cytokine production. Eur. Cytokine Netw. 9:239–245.

39. Muller, G., N. Hanekop, W. Kramer, W. Bandlow, and W. Frick. 2002.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

Interaction of phosphoinositolglycan(-peptides) with plasma membrane lipidrafts of rat adipocytes. Arch. Biochem. Biophys. 408:17–32.

40. Murphy, C. T., S. A. McCarroll, C. I. Bargmann, A. Fraser, R. S. Kamath,J. Ahringer, H. Li, and C. Kenyon. 2003. Genes that act downstream ofDAF-16 to influence the lifespan of Caenorhabditis elegans. Nature 424:277–283.

41. Naik, R. S., E. A. Davidson, and D. C. Gowda. 2000. Developmental stage-specific biosynthesis of glycosylphosphatidylinositol anchors in intraerythro-cytic Plasmodium falciparum and its inhibition in a novel manner by man-nosamine. J. Biol. Chem. 275:24506–24511.

42. Naotunne, T. S., N. D. Karunaweera, K. N. Mendis, and R. Carter. 1993.Cytokine-mediated inactivation of malarial gametocytes is dependent on thepresence of white blood cells and involves reactive nitrogen intermediates.Immunology 78:555–562.

43. Remy, I., A. Montmarquette, and S. W. Michnick. 2004. PKB/Akt modulatesTGF-beta signalling through a direct interaction with Smad3. Nat. Cell Biol.6:358–365.

44. Richman, A. M., G. Dimopoulos, D. Seeley, and F. C. Kafatos. 1997. Plas-modium activates the innate immune response of Anopheles gambiae mos-quitoes. EMBO J. 16:6114–6119.

45. Riehle, M. A., and M. R. Brown. 2002. Insulin receptor expression duringdevelopment and a reproductive cycle in the ovary of the mosquito Aedesaegypti. Cell Tissue Res. 308:409–420.

46. Riehle, M. A., and M. R. Brown. 1999. Insulin stimulates ecdysteroid pro-duction through a conserved signaling cascade in the mosquito Aedes aegypti.Insect Biochem. Mol. Biol. 29:855–860.

47. Riehle, M. A., S. F. Garczynski, J. W. Crim, C. A. Hill, and M. R. Brown.2002. Neuropeptides and peptide hormones in Anopheles gambiae. Science.298:172–175.

48. Rintelen, F., H. Stocker, G. Thomas, and E. Hafen. 2001. PDK1 regulatesgrowth through Akt and S6K in Drosophila. Proc. Natl. Acad. Sci. USA98:15020–15025.

49. Ropert, C., and R. T. Gazzinelli. 2000. Signaling of immune system cells byglycosylphosphatidylinositol (GPI) anchor and related structures derivedfrom parasitic protozoa. Curr. Opin. Microbiol. 3:395–403.

50. Saperstein, R., P. P. Vicario, H. V. Strout, E. Brady, E. E. Slater, W. J.Greenlee, D. L. Ondeyka, A. A. Patchett, and D. G. Hangauer. 1989. Designof a selective insulin receptor tyrosine kinase inhibitor and its effect onglucose uptake and metabolism in intact cells. Biochemistry 28:5694–5701.

51. Schmidt, A., R. T. Schwarz, and P. Gerold. 1998. Plasmodium falciparum:asexual erythrocytic stages synthesize two structurally distinct free and pro-tein-bound glycosylphosphatidylinositols in a maturation-dependent man-ner. Exp. Parasitol. 88:95–102.

52. Schofield, L., and F. Hackett. 1993. Signal transduction in host cells by aglycosylphosphatidylinositol toxin of malaria parasites. J. Exp. Med. 177:145–153.

53. Schofield, L., S. Novakovic, P. Gerold, R. T. Schwarz, M. J. McConville, andS. D. Tachado. 1996. Glycosylphosphatidylinositol toxin of Plasmodium up-

regulates intercellular adhesion molecule-1, vascular cell adhesion mole-cule-1, and E-selectin expression in vascular endothelial cells and increasesleukocyte and parasite cytoadherence via tyrosine kinase-dependent signaltransduction. J. Immunol. 156:1886–1896.

54. Schofield, L., L. Vivas, F. Hackett, P. Gerold, R. T. Schwarz, and S. Tachado.1993. Neutralizing monoclonal antibodies to glycosylphosphatidylinositol,the dominant TNF-alpha-inducing toxin of Plasmodium falciparum: pros-pects for the immunotherapy of severe malaria. Ann. Trop. Med. Parasitol.87:617–626.

55. Sherry, B. A., G. Alava, K. J. Tracey, J. Martiney, A. Cerami, and A. F.Slater. 1995. Malaria-specific metabolite hemozoin mediates the release ofseveral potent endogenous pyrogens (TNF, MIP-1 alpha, and MIP-1 beta) invitro, and altered thermoregulation in vivo. J. Inflamm. 45:85–96.

56. Straif, S. C., C. N. Mbogo, A. M. Toure, E. D. Walker, M. Kaufman, Y. T.Toure, and J. C. Beier. 1998. Midgut bacteria in Anopheles gambiae and An.funestus (Diptera: Culicidae) from Kenya and Mali. J. Med. Entomol. 35:222–226.

57. Tachado, S. D., P. Gerold, M. J. McConville, T. Baldwin, D. Quilici, R. T.Schwarz, and L. Schofield. 1996. Glycosylphosphatidylinositol toxin of Plas-modium induces nitric oxide synthase expression in macrophages and vas-cular endothelial cells by a protein tyrosine kinase-dependent and proteinkinase C-dependent signaling pathway. J. Immunol. 156:1897–1907.

58. Tachado, S. D., P. Gerold, M. J. McConville, T. Baldwin, D. Quilici, R. T.Schwarz, and L. Schofield. 1996. Glycosylphosphatidylinositol toxin of Plas-modium induces nitric oxide synthase expression in macrophages and vas-cular endothelial cells by a protein tyrosine kinase-dependent and proteinkinase C-dependent signaling pathway. J. Immunol. 156:1897–1907.

59. Tachado, S. D., P. Gerold, R. Schwarz, S. Novakovic, M. McConville, and L.Schofield. 1997. Signal transduction in macrophages by glycosylphosphati-dylinositols of Plasmodium, Trypanosoma, and Leishmania: activation ofprotein tyrosine kinases and protein kinase C by inositolglycan and diacyl-glycerol moieties. Proc. Natl. Acad. Sci. USA. 94:4022–4027.

60. Yao, Z., Y. Dolginov, T. Hanoch, Y. Yung, G. Ridner, Z. Lando, D. Zharhary,and R. Seger. 2000. Detection of partially phosphorylated forms of ERK bymonoclonal antibodies reveals spatial regulation of ERK activity by phos-phatases. FEBS Lett. 468:37–42.

61. Zhou, B., Z. X. Wang, Y. Zhao, D. L. Brautigan, and Z. Y. Zhang. 2002. Thespecificity of extracellular signal-regulated kinase 2 dephosphorylation byprotein phosphatases. J. Biol. Chem. 277:31818–31825.

62. Zhou, B., and Z. Y. Zhang. 2002. The activity of the extracellular signal-regulated kinase 2 is regulated by differential phosphorylation in the activa-tion loop. J. Biol. Chem. 277:13889–13899.

63. Zhu, J., G. Krishnegowda, and D. C. Gowda. Induction of proinflammatoryresponses in macrophages by the glycosylphosphatidylinositols (GPIs) ofPlasmodium falciparum: The requirement of ERK, p38, JNK and NF-kappaB pathways for the expression of proinflammatory cytokines and nitric oxide.J. Biol. Chem., in press.

Editor: W. A. Petri, Jr.



http://iai.asm.org/

Dow

nloaded from
http://iai.asm.org/

Documents

Induction of Nitric Oxide Synthase in Anopheles stephensi ...thologous to those in Drosophila melanogaster (11, 48) have been described from Aedes aegypti (45, 46) and from Anopheles