Foley & O’Farrell – Genes & Development 2003, 17: 115-125 1

1

Nitric oxide contributes to induction of innate immuneresponses to gram negative bacteria in Drosophila

Edan Foley and Patrick H. O’Farrell*Dept. of Biochemistry and Biophysics, 513 Parnassus Ave, UCSF, San Francisco, CA. 94143-0448

Running title: NO signaling in innate immunityKey words: Drosophila, Nitric Oxide, Signaling, Innate Immunity,Hemocyte, Relish,

ABSTRACTStudies in mammals uncovered important

signaling roles of nitric oxide (NO), andcontributions to innate immunity.Suggestions of conservation led us toexplore the involvement of NO inDrosophila innate immunity. Inhibition ofnitric oxide synthase (NOS) increasedlarval sensitivity to gram negative bacterialinfection, and abrogated induction of theantimicrobial peptide Diptericin. NOS wasupregulated after infection. Antimicrobialpeptide reporters revealed that NOtriggered an immune response in

uninfected larvae. NO induction ofDiptericin reporters in the fat body requiredimmune deficiency (imd) and domino.These findings show that NOS activity isrequired for a robust innate immuneresponse to gram negative bacteria, NOS isinduced by infection, and NO is sufficientto trigger response in the absence ofinfection. We propose that NO mediates anearly step of the signal transductionpathway inducing the innate immuneresponse upon natural infection with gramnegative bacteria.

INTRODUCTION The innate immune system is an ancient first line

of defense against foreign organisms (Hoffmann etal. 1999). In contrast to the geneticrearrangements and clonal selection processesthat underlie adaptive immunity, innate immunityrelies on the functions of germ line encoded geneproducts. Nitric oxide (NO) is a highly reactivemolecule with innate immune functions as well asroles in responses to hypoxia, and CNSdevelopment (Dawson et al. 1991; Bredt andSnyder 1994; Gibbs and Truman 1998; Wingroveand O'Farrell 1999; Bogdan 2001). NO isproduced in mammalian macrophages by a NitricOxide Synthase isoform (iNOS/NOS2), which isstrongly upregulated following infection(MacMicking et al. 1997). Macrophages alsoproduce high quantities of superoxide by NADPHoxidase following pathogen detection (Ding et al.1988). The highly toxic molecules NO, superoxideand their derivatives are believed to play animportant role in destroying invadingmicroorganisms (Nathan and Shiloh 2000). Inseeking to develop Drosophila as a model for

studying mechanisms modulating and directingthis toxicity, we found that NO plays a signalingrole in the induction of immune responses inDrosophila larvae to gram negative bacteria.

Metazoan defenses against foreign organismsare commonly induced by pathogen detection.Components of the innate immune systemrecognize chemical structures that are hallmarks ofmicroorganisms and are not found on host cells(Akira et al. 2001). These Pathogen AssociatedMolecular Patterns (PAMPs) include β-1,3-glucanof fungi, peptidoglycan and lipopolysaccharides ofbacteria and phosphoglycan of parasites.Widespread PAMPs are recognized by PatternRecognition Receptors (PRRs). PRR activation bybinding of cognate PAMPs initiates host signalingcascades which start defense responses.

Toll (Tl) signal transduction, which was originallyidentified based on its role in dorsal/ventralpatterning of the Drosophila embryo, is nowrecognized as having an evolutionarily conservedrole in pathogen detection and induction ofimmune responses (Rosetto et al. 1995; Lemaitre

Foley & O’Farrell – Genes & Development 2003, 17: 115-125 2

2

et al. 1996; Nicolas et al. 1998; Qiu et al. 1998;Kopp and Medzhitov 1999; Aderem and Ulevitch2000; Tauszig et al. 2000; Michel et al. 2001; Ooiet al. 2002; Tauszig-Delamasure et al. 2002;Underhill and Ozinsky 2002). The Drosophila Tlpathway is particularly responsive to fungi andgram positive bacteria, which trigger processingand activation of the Tl ligand, Spaetzle (Spz)(Levashina et al. 1999). Tl receptor dependentsignaling activates two NF-κB related transcriptionfactors, Dorsal (Dl) and Dorsal-related immunityfactor (Dif), and expression of defense genes suchas those encoding antimicrobial peptides (Lemaitreet al. 1995b; Meng et al. 1999).

In Drosophila, a pathway that is independent ofTl, the Imd-pathway, is activated in response togram negative bacteria. A putative peptidoglycanrecognition protein (PGRP-LC) contributes to thispathway (Choe et al. 2002; Gottar et al. 2002;Ramet et al. 2002). Although there areuncertainties regarding its specificity, PGRP-LChas attributes of a PAMP and it acts upstream ofthe death-domain containing protein Immunedeficiency (Imd). Imd signaling proceeds throughthe MAP kinase kinase kinase homologue, dTAK1,the caspase, dredd and an I-κ-kinase to culminatein the proteolytic activation of the NF-κB familymember Relish (Rel) (Lemaitre et al. 1995a;Dushay et al. 1996; Wu and Anderson 1998;Hedengren et al. 1999; Kim et al. 2000; Silvermanet al. 2000; Stoven et al. 2000; Georgel et al.2001; Lu et al. 2001; Vidal et al. 2001). Mutantslacking Rel are highly sensitive to infection withgram negative bacteria (Hedengren et al. 1999).

Two protocols have been established forinvestigating responses of Drosophila larvae toinfection. Piercing the larval cuticle with a needlesoiled with pathogen triggers a robust induction ofthe innate immune response. While this septicinjury protocol has been used successfully todefine the outlines of the signaling cascades thatinduce host responses to infection, it has adisadvantage. The trauma of injury weaklyactivates immune responses even withoutbacteria, thus eroding the specificity of theresponse. A recently devised “natural infection”protocol activated the Imd-pathway by brieflyfeeding larvae on concentrated slurries of a gramnegative bacteria such as Erwinia carotovoracarotovora (Basset et al. 2000).

We present evidence that NO makes crucialcontributions to the Imd-pathway that activatesRel. Pharmacological inhibition of nitric oxide

synthase (NOS) increases larval and adultsusceptibility to septic and natural infection withgram negative bacteria and compromisesproduction of antimicrobial peptides. Treatment ofuninfected larvae with NO donors is sufficient toactivate antimicrobial peptide production. Geneticexperiments demonstrate that one action of NO iseither upstream of Imd or in a parallel pathwayrequired for Rel activation of innate immuneresponses in fat body. Ingestion of bacteria isfollowed by induction of NOS in the gut and inhemocytes. NO mediated induction of theantimicrobial peptide Diptericin in the fat body wasblocked in domino mutants. We propose that NO isan innate immune signal that acts early in a multi-tiered cascade in which sentinel tissues detectgram negative bacteria and recruit other tissues tohost defense.

RESULTS

NOS contributes to pathogen destructionWe tested if inhibition of NOS influences the

ability of larvae to survive a bacterial infection.Drosophila has a single NOS homologue for whichthere are no available mutant alleles (Regulski andTully 1995; Stasiv et al. 2001). Topharmacologically inactivate NOS, we treatedlarvae with the NOS-inhibitory arginine analogueNω-Nitro-L-Arginine-Methyl-Ester (L-NAME) andused the septic injury and natural infection modelsto infect third instar larvae with the Drosophilapathogen Erwinia carotovora carotovora (Ecc)(Fig. 1A and B respectively). In the absence ofinfection, treatment with L-NAME or the inactive D-enantiomer had little impact on eclosion/survival.Furthermore, injury alone did not significantlycompromise eclosion/survival. In contrast, L-NAME but not D-NAME treatment dramaticallyreduced the ability of larvae to survive asubsequent Ecc septic infection (Fig. 1A,B).Similar effects were observed with adult filessubjected to a septic infection (data not shown).We conclude that inactivation of NOS impairs theability of Drosophila to combat subsequentinfections by two different routes.

To distinguish contributions of NOS to the Imd-dependent and Spz-dependent defense pathways,we tested the consequence of NOS inhibition inmutant strains incapable of activating one or the

Foley & O’Farrell – Genes & Development 2003, 17: 115-125 3

3

other pathway. D-NAME had no effect and L-NAME only a modest effect on the ability of spzheterozygous larvae to eliminate bacteria (Fig 1 Cand data not shown). In contrast, homozygousspz mutant larvae exhibited a remarkably differentresponse to feeding D-NAME and L-NAME: at 24h after infection, bacterial titer was 10,000 timeshigher in spz mutants fed L-NAME (Fig 1 C). Thisresult indicates that L-NAME severelycompromises Spz-independent host defenses. Themajor defense against gram negative bacteriasuch as Ecc is thought to be mediated via the Spz-independent Imd pathway, which is thus acandidate target for the action of L-NAME.

Bacterial titer remained high in imd mutantswhether or not they were fed L-NAME (data notshown). This is in accord with the possibility that L-NAME acts on the Imd-pathway, but we cannotasses from this result whether NOS regulates anyImd-independent contributions to host defense,because such pathways do not appear to conferany significant resistance to our infection protocol.

NO is a signaling molecule in the Imd/Relpathway.

The emphasized role of NO in the mammalianhost defense is its involvement in pathogen killing.Studies in insect systems, have also indicated arole for NOS in eliminating pathogens (Dimopouloset al. 1998; Luckhart et al. 1998; Han et al. 2000),but there are suggestions that its action mightinclude induction of immune responses (Nappi etal. 2000; Imamura et al. 2002). For this reason, wetested whether NO might have a role as anupstream regulator of the immune response bytesting the influence of L-NAME treatment oninduction of the antimicrobial peptide Diptericin(Dipt). Induction of Dipt in the fat body after anatural or septic infection is strictly dependent onImd and Relish (Lemaitre et al. 1995a; Hedengrenet al. 1999). We followed Dipt induction using areporter construct in which GFP expression iscontrolled by the Dipt promoter (Dipt-GFP). Thisreporter mimics the infection-induced expressionof the endogenous Dipt gene (Tzou et al. 2000).

No GFP was observed in Dipt-GFP larvae in theabsence of infection. A large fraction (79%) oflarvae naturally infected with Ecc produced GFP inthe fat body (Fig. 2B and F) within 16 h. While D-NAME did not compromise this induction (Fig. 2Cand F), the fraction of GFP expressing larvae wasdramatically reduced by L-NAME treatment (Fig.2D,F, 10%). Similar effects were seen after natural

infection with E. coli, although E. coli was not aspotent as Ecc in inducing the reporter (data notshown). Thus, inhibition of NOS greatly reducedinfection-dependent expression of the Diptreporter.

To determine if NO is sufficient to activate Diptproduction on its own, uninfected Dipt-GFP larvaewere fed the NO-donor S-Nitroso-N-acetylpenicillamine (SNAP) and examined 16 hlater. Exposure to exogenous NO induced GFP inabout 50% of the larvae despite the lack ofinfection (Fig. 2E,F). Control Dipt-GFP larvaetreated with pre-reacted SNAP that no longerreleased NO did not express the reporter (data notshown). Therefore, we conclude that NOS activityis required for Dipt expression after naturalinfection with gram negative bacteria and that NOis sufficient to activate Dipt expression in theabsence of infection. Both infection and SNAP-treatment resulted in an initial mosaic pattern ofGFP in the fat body (data not shown and Reichhartet al. 1992). Thus, exogenously provided NO notonly activates the Imd pathway in the absence ofinfection, it also recapitulates this spatial feature ofDipt expression.

To compare the efficacy of natural and septicinfections and to assess the degree of inhibition ofthe response by L-NAME, we measured β-galactosidase activity in extracts of treated larvaethat express β-galactosidase under the control ofthe Diptericin promoter (Dipt-lacZ). Septic infectionproduced a slightly faster and more persistent risein β-galactosidase than natural infection, but thisdifference was small compared to the suppressionof the response upon incubation with L-NAME(Fig. 2G). Note that there is a slow rise in β-galactosidase even in the presence of L-NAMEand that control larvae appear to be downregulating the response within 24 h (Fig. 2G). Wealso quantified the β-galactosidase activity ofuninfected, SNAP-treated Dipt-lacZ larvae 24 hafter feeding SNAP. Exposure to NO led to a clearelevation of β-galactosidase activity in uninfectedlarvae (Fig. 2H) supporting the results with Dipt-GFP (Fig 2 E, F) indicating that NO is sufficient toinduce Dipt expression in uninfected larvae.

In summary, pharmacological inhibition of NOSinhibits, or greatly retards induction of Dipt reporterconstructs in response to infection, while dietaryprovision of NO to larvae stimulates Diptproduction in the absence of infection. We proposethat NO contributes to the activation of the Imd/Relpathway. As L-NAME treated wild type larvae

Foley & O’Farrell – Genes & Development 2003, 17: 115-125 4

4

eliminate pathogens more effectively than imdmutants, we believe that pharmacological inhibitionof NOS does not completely inactivate Imd-dependent signaling. In agreement with this, wenoticed that a low percentage of L-NAME treatedinfected Dipt-GFP larvae still express GFP uponinfection and that, although greatly reduced anddelayed, low level reporter induction occurred inthe presence of L-NAME. Nonetheless, thecontribution of NOS to gram negative bacterialdestruction is substantial, as L-NAME treated spzmutants are greatly impaired in their ability to clearEcc upon natural infection when compared tountreated control spz mutant larvae.

NO effects on Drs expression.Whereas Dipt is induced strictly via the Imd

pathway, Drosomycin (Drs) is also induced by theToll pathway (Lemaitre et al. 1996). To testwhether NOS is required for this response, weexamined induction of a reporter expressing GFPunder the control of the Drosomycin promoter(Drs-GFP). A few uninfected larvae (7%) exhibitedfoci of Drs-GFP expression, but natural infectionwith Ecc dramatically increased the incidence andextent of GFP expression 16 h post infection (Fig.3B). The number of larvae exhibiting this infection-induced Drs expression was not appreciablyaffected by D-NAME or L-NAME treatment (88%and 80% respectively, Fig. 3F). Thus, it appearsthat inhibition of NOS does not prevent infection-dependent Drs induction. Nonetheless, the NOdonor SNAP induced GFP in about 50% ofuninfected Drs-GFP larvae (Fig. 3E,F).

To quantify contributions of NO to Drsexpression, we prepared extracts from naturallyinfected larvae that express β-galactosidase undercontrol of the Drosomycin promoter (Drs-lacZ).Following infection, L-NAME fed larvae hadsignificantly less (by nearly half) β-galactosidaseactivity than D-NAME fed larvae (Fig. 3G). Thus, itappears that although NOS is not required for allDrs reporter activity, NOS activity makes aquantitative contribution. Consistent with aquantitative contribution of NO to the induction, weobserved that SNAP mediated a substantialinduction of Drs-lacZ in uninfected larvae (Fig. 3H).While the most straightforward interpretation ofthese results is that infection-mediated induction ofDrs occurs via two pathways (see discussion), aNOS dependent and independent one, we cannotexclude the possibility that L-NAME only partiallyblocks a single NO-dependent pathway.

Epistatic relations of NOS and Imd in theImd/Rel pathway.

To test whether NO induction of a Dipt-reporterinvolves the Imd pathway, we tested the responsein the fat bodies of imd mutant larvae. As shownby activity stain for β-galactosidase in dissected fatbodies from Dipt-lacZ larvae (Fig. 4), NO inductionof the reporter requires Imd function. Reciprocally,overexpression of Imd, which activates immuneresponses (Georgel et al. 2001) and reporterexpression (Fig. 4 H) in the absence of infection,retained its ability to activate Dipt-lacZ expressionin the presence of the NOS inhibitor, L-NAME (Fig.4 I). These findings are most simply consistentwith action of NO upstream of Imd in the signaltransduction pathway inducing Dipt in the fat body,although it remains possible that NO acts in aparallel pathway.

NOS involvement in hemocyte responses toinfection.

In addition to the fat body, other cells such ashemocytes produce antimicrobial peptides afterinfection. To test whether NO impinges uponhemocyte expression of antimicrobial peptides, wenaturally infected or SNAP treated Dipt-lacZlarvae, collected hemocytes 16 h later and stainedfor β-galactosidase activity. We did not observeDipt-lacZ activity in hemocytes from uninfectedlarvae (Fig. 5A), while a subset of hemocytes fromlarvae infected with Ecc had deposits of X-galderived stain (Fig. 5B). This induction was alsoobserved following treatment of uninfected larvaewith SNAP (Fig. 5C). Infection dependent inductionof Dipt was blocked by L-NAME (Fig. 5D). Weconclude that, NO is necessary for efficientinduction of Dipt in larval hemocytes followingnatural infection.

In mammalian systems, one of three NOSisoforms, inducible NOS (iNOS/NOS2), isupregulated in macrophages in response toseveral types of infection or challenge withbacterial lipopolysaccharides. We asked if infectionupregulates NOS in Drosophila hemocytes, themajority of which are phagocytic cells that functionlike mammalian macrophages. Larvae wereinfected by feeding with Ecc and hemocytes werecollected after 5 h and examined for NOS proteinlevels by immunofluorescence. We used twodifferent anti-NOS antibodies; a universal anti-NOS and an antibody we prepared against a C-terminal peptide of Drosophila NOS. Similar results

Foley & O’Farrell – Genes & Development 2003, 17: 115-125 5

5

were observed for both antibodies. Hemocytes ofinfected larvae showed elevated NOS proteinlevels as early as 5 h after infection (Fig. 5H).

Since host pathogen interaction following naturalinfection is likely to first occur in the gut, weconsidered whether there might also be aresponse in this tissue. We monitored NOSactivity in the guts of infected larvae using ahistochemical stain (diaphorase staining). Weobserved a striking increase in NOS activity levelsin the guts of naturally infected larvae 5 h afterinfection with Ecc (Fig. 5L).

NO-dependent activation of Dipt production infat bodies requires the function of domino.

Mutations in the chromatin remodeling factor,domino, affect proliferating tissues, includingimaginal tissues and larval hemocytes and result inpupal lethality (Ruhf et al. 2001). It has beendemonstrated previously that domino mutants areincapable of producing Dipt in their fat body afternatural infection with Ecc, while fat bodyproduction of Drs was not impaired (Basset et al.2000). This result has implicated hemocytes,which are missing in domino mutants, in theresponse to bacteria; however, as the mutation isrelatively pleiotropic, other explanations for thedomino dependence are possible. Nonetheless,we tested whether the domino mutation similarlyimpairs SNAP-mediated induction of Drs-GFP orDipt-GFP.

Uninfected Drs-GFP larvae occasionally expressGFP (Fig. 3F) and a low level of backgroundexpression persisted in domino mutants (Fig. 6A).As described previously, Drs-GFP; dominomutants expressed GFP after natural infection withEcc (Fig. 6C). Interestingly, uninfected dominomutant larvae also induced Drs-GFP after feedingSNAP (Fig. 6D). Thus, Drs reporter induction inresponse to exposure to an NO donor or tobacteria can occur in the absence of hemocytes.Thus, cells other than, or in addition to, hemocytesmust respond to NO and have inputs into theimmune response.

We observed different effects when weexamined domino; Dipt-GFP larvae. In this case,we observed no GFP fluorescence after eithernatural infection with Ecc or treatment with SNAP(Fig. 6F and G respectively). These data show aparallel between NO signaling and the Imd

pathway, in that both are blocked in the dominomutant. Furthermore, this finding suggests that thefat body does not have an autonomous ability torespond to NO, and that the NO role in theinduction of Dipt-GFP occurs at stages of signalingupstream of the fat body. Finally the distinctionbetween the effect of the domino mutation on Drsand Dipt reporters demonstrates that NOstimulates Drs expression via a pathway that is, atleast partially, distinct from the pathway by whichNO induces Dipt. We speculate that both pathwaysinvolve distinct sentinel tissues that act upstreamof the fat body (Fig. 7).

DISCUSSION.Drosophila has emerged as a potent model for

elucidation of mechanisms underlying innateimmunity. Our data show that NO performs asignaling function in Drosophila in the induction ofhost defenses to gram negative bacteria.Pharmacological inactivation of Drosophila NOSmakes larvae more susceptible to infection andblocks Rel-dependent activation of theantimicrobial peptide Diptericin. NOS isupregulated in the gut and hemocytes followingnatural infection, consistent with a role for NOS incombating invading microorganisms. Furthermore,dietary provision of an NO donor is sufficient toactivate promoters of antimicrobial genes in thehemocytes and fat body of uninfected hosts. Testsof the ability of NO to induce fat body expressionof Diptericin in mutant backgrounds showed that itcannot bypass the requirement for Imd, acomponent of the signal transduction pathwayinducing Dipt.

Inhibition of NOS impairs the host defenseresponse.

Our initial experiments showed thatpharmacological inactivation of NOS compromisedthe ability of Drosophila larvae or adults to surviveinfection by septic injury or natural infection.Furthermore, NOS inhibition by L-NAME limitedthe ability of naturally infected larvae to eliminategram negative bacteria (E. carotovora carotovoraand E. coli). L-NAME attenuation of pathogenelimination was particularly dramatic in larvaeincapable of mounting Spz-dependent responsesto infection. In contrast, an imd mutant was notfurther compromised by NOS inhibition. These

Foley & O’Farrell – Genes & Development 2003, 17: 115-125 6

6

results suggest that NOS activity is of particularimportance in the Imd-dependent immuneresponse pathway.

Production of Dipt in the fat body followinginfection relies entirely on the Imd signalingcascade, as mutations in imd or the Imd-responsive NF-κB homologue rel completelyabrogate Dipt production following infection(Lemaitre et al. 1995a; Hedengren et al. 1999;Georgel et al. 2001). Overproduction of Imd in theabsence of infection activates Dipt production(Georgel et al. 2001). Thus, the Imd pathway isrequired for Dipt activation and is sufficient toinduce Dipt expression. Similarly, we show thatpharmacological inhibition of NOS reduces Diptexpression while introduction of NO donorsactivates Dipt expression. Thus, NO is required foroptimal Dipt induction and NO is sufficient totrigger Dipt expression.

We propose that NO has a signaling role in theinnate immune system of Drosophila andcontributes to pathogen resistance in this capacity(Fig. 7). We note that this role in host defense isdistinct from roles that NO might play as acytotoxic agent used in the destruction ofpathogens. Previous studies have indicatedsignaling functions for NO in the immune responseof mammals and plants (Delledonne et al. 1998;Durner et al. 1998; Diefenbach et al. 1999)although the exact manner in which NO functionshas not been elucidated. Our data demonstratesthat NO is essential for activation of a specificbranch of the innate immune response pathway;the Imd pathway in Drosophila which is analogousto the TNF pathway of mammals.

Interactions between NO and the Tl-dependentresponse pathway.

Our data imply a crucial role for NO in Imd-dependent signaling. However, NO also canactivate production of Imd-independentantimicrobial peptides, as dietary provision of NOis sufficient to activate Drs production. While thisfinding does not necessarily imply that NOnormally acts in this pathway, it is of interest toconsider the relationship to the results we haveobtained for NO induction of the Imd pathway.

There is some cross-talk between the Toll andImd pathways such that activation of the Imdpathway by NO might contribute to the induction ofDrs. However the Imd pathway is not necessaryfor induction of Drs by exogenous NO, as a dietaryNO source induces Drs to an equal extent in wild

type, imd and relE20 mutants (data not shown).Hence, there must be another route of activation ofDrs by exogenous NO. However, since inhibition ofNOS does not effectively inhibit the Tl-pathway asit does the Imd-pathway, further analyses will berequired to determine whether NO normally makescontributions to the Tl-dependent innate immuneresponses. Here, we focus on NO contributions tothe Imd-pathway.

The domino mutation and NO-dependentactivation of antimicrobial peptide production.

The Domino gene product is a transcriptionfactor involved in chromatin remodeling (Ruhf et al.2001). Mutations in domino affect proliferatingtissues such as the imaginal discs and lymphglands, resulting in larvae with several defects,including elimination or severe reduction in thenumber of hemocytes (Braun et al. 1998). Whiledomino mutants produce Dipt in the proventriculusafter natural infection, they no longer produce it inthe fat body. These observations led to theproposal that hemocytes may produce a cytokine-like molecule following natural infection whichactivates antimicrobial peptide production in the fatbody (Basset et al. 2000). While the proposal thatthe fat body response is secondary to a hemocytesignal remains tentative because of the pleiotropyof the domino mutation, it is clear that a gut(proventriculus) response can occur independentof hemocytes. We find that NO also fails to induceDipt in the fat body of domino mutant larvae,suggesting that NO does not act directly on the fatbody. Perhaps it mediates signaling at earliersteps in the response.

We propose that different host organs playdistinct roles in the response to foreign organisms.Tissues that contact a pathogen early duringinfection are well positioned to perform a “sentinel”function by detecting the pathogen and signalingother tissues to recruit their contributions to hostdefense (Fig. 7). Our results are consistent withthe suggested involvement of hemocytes in theresponse of the fat body (Basset et al. 2000). Thefinding that exogenous NO is unable to induce Diptin the fat body of domino mutants, which lackhemocytes, suggests that the fat body cannotrespond autonomously to NO. Although thepleiotropy of the mutation allows otherinterpretations, the result encourages us to thinkthat NO acts upstream of the fat body. In supportof this, the initial fat body response to NO orinfection appears as a mosaic of Dipt expressing

Foley & O’Farrell – Genes & Development 2003, 17: 115-125 7

7

and nonexpressing cells. This feature is not easilyconsistent with a humoral signal, but is compatiblewith a induction by cell contact with activatedhemocytes. Furthermore, hemocytes respond tonatural infection by inducing Dipt. Since thisresponse is mimicked by NO treatment andinhibited by pharmacological inhibition of NOS, wesuggest that NO acts upstream of this hemocyteresponse. It was shown previously that the gut(proventriculus) can respond to infectionindependent of, and perhaps upstream of,hemocytes (Basset et al. 2000). We show thatNOS activity is induced in the gut following naturalinfection. These findings suggest the tentativemodel shown in Fig. 7 in which cells in the gutdetect the pathogen, activate hemocytes via anNO dependent signal, and hemocytes act in turn toinduce the fat body by an as yet unknown signal.Since pathogens have numerous ports of access,we expect that, like the diverse dendritic system inhumans, there will be a number of alternativesentinel tissues and multiple pathways of triggeringactivation of innate immunity.

Our work and other studies indicate that NO hasa global involvement in the immune response.Work in mammalian cells predicted involvement ofNO in pathogen elimination during an oxidativeburst in phagocytic cells, while our workdocuments a role for NO as an inducer of theinnate immune response. It remains to be testedfully, but several observations suggest that both ofthese roles will be conserved.

MATERIALS AND METHODS

Fly strains and culturing

We used Sevelen flies as a wild type strain. The spzrm7,domino, and imd mutant strains used have been describedelsewhere (Lemaitre et al. 1995a; Lemaitre et al. 1996; Braunet al. 1998; Georgel et al. 2001). The reporter strains Dipt-lacZ, Drs-GFP and Drs-lacZ have been described in(Reichhart et al. 1992; Ferrandon et al. 1998; Manfruelli et al.1999) respectively. Dipt-GFP flies were a gift of BrunoLemaitre. It is a homozygous viable strain with two insertionson each third chromosome. In this line GFP is expressedunder the control of the Dipt 2.2kb promoter, whichrecapitulates in vivo Dipt expression. Flies were cultured onstandard fly food at 25ºC prior to infection. After infection flieswere transferred to a 29ºC incubator.

Infection experimentsNatural infection experiments with Erwinia carotovora

carotovora or E. coli were done as described previously(Basset et al. 2000). Briefly, for 30min third instar larvae werefed a concentrated bacterial pellet mixed into crushed banana

with 500mM L-NAME or D-NAME. Larvae were thentransferred to 29ºC for the remainder of the experiment. Todetermine the consequences of NOS inhibition on larvalsurvival after infection larvae were transferred to vials thatcontained instant fly food with 100mM L-NAME or D-NAMErespectively. Bacterial titer was determined by plating serialdilutions of homogenate from infected larvae at various timesafter infection onto LB-Amp plates. The strain of Erwiniacarotovora carotovora used (015) is ampicillin resistant and E.coli was transfected with a β-lactamase encoding plasmid. Forseptic injury experiments on larvae late second instar larvaewere transferred from standard fly food to instant fly foodcontaining 100mM D-NAME or L-NAME. 120 third instarlarvae were picked 16 h later and infected by puncturing thecuticle with a dissecting needle that had been previouslysoaked in a concentrated bacterial pellet. In parallel controlexperiments, 120 larvae were treated in an identical mannerand injured with a sterile needle. For septic injury of adultslate second instar larvae were transferred from standardcornmeal food to instant fly food with 100mM L-NAME or D-NAME. 200 1-2 day old adult flies that eclosed were infectedby puncturing their thorax with a sharp needle that had beendipped in a concentrated bacterial pellet.

MicroscopyAll images were taken on a Leica DMRD. To visualize

expression of GFP in larvae darkfield and green fluorescentimages (FITC) were taken of representative larvae andmerged using Adobe Photoshop 5.5. Figures were assembledusing Adobe Illustrator. For immunohistochemistry of larvalhemocytes, hemocytes were deposited on Superfrost PlusGold microscope slides (Fisher) and fixed for 5min in 4%glutaraldehyde in PBS. Cells were incubated with a 1:200dilution of Rabbit-anti-NOS and Rhodamine-labeled anti-rabbitwas used a secondary antibody. NOS antibody staining wasdone using a universal anti-NOS antibody (Oncogene) orusing an antibody raised against a peptide(TAEIHTKSRATARIRMASQ) that corresponds to a C-terminalsection of the mature NOS protein.

-galactosidase assaysTo monitor induction of Dipt-lacZ in fat bodies, fat bodies

were dissected from larvae in PBS and fixed for 10min in PBSwith 0.5% glutaraldehyde on ice. Fat bodies were thenincubated at room temperature in staining buffer (30µl X-gal(5% 5-bromo-4-chloro-3-indolyl b-D-galactosidase in DMFO)per ml staining solution (10m sodium phosphate buffer pH7.2,150mM NaCl, 1mM MgCl2, 3.5mM K3FeCN6 and 3.5mMK4FeCN6)). β-galactosidase activity in hemocytes from Dipt-lacZ larvae was monitored by depositing hemolymph onSuperfrost Plus Gold microscope slide (Fisher), fixing for 30sin PBS with 0.5% glutaraldehyde and staining overnight at37ºC with staining buffer. β-galactosidase titration was doneas described previously (Basset et al. 2000).

Diaphorase stainingThird instar larvae were washed and dissected in PBS andthen transferred to fixing buffer (4% paraformaldehyde,100mM PIPES pH 7.4, 2mM MgSO4, 1mM EGTA) for 30minat room temperature. Larvae were then washed in PBS andincubated in staining solution (1mM NADPH, 0.2mMnitrobluetetrazolium, 100mM Tris pH 7.2, 0.2% Triton X-100)for 20min at RT. Larvae were washed in PBS and mounted inglycerol.

Foley & O’Farrell – Genes & Development 2003, 17: 115-125 8

8

PharmacologyThe NO donor SNAP was fed to third instar larvae for 15min

at RT by mixing 15mM SNAP in crushed banana and thenfeeding to larvae.

AcknowledgementsFly strains were kindly provided by Kathryn Anderson, Jules

Hoffmann, Bruno Lemaitre and Marie Meister. We are gratefulto Bruno Lemaitre for advice and discussions. We also wish tothank to Pascale Dijkers, Arnaud Echard, Renny Feldman,Gilles Hickson, Devin Parry and Anita Sil for critical reading ofthe manuscript. This work was supported by a basic ScienceAward from the Sandler Family Supporting Foundation andNIH grant GM60988. Supported in part by Fellowship DRG-1713-02 from the Damon Runyon Cancer ResearchFoundation.

ReferencesAderem, A. and R.J. Ulevitch. 2000. Toll-like receptors in the

induction of the innate immune response. Nature 406: 782-7.

Akira, S., K. Takeda, and T. Kaisho. 2001. Toll-like receptors:critical proteins linking innate and acquired immunity. NatImmunol 2: 675-80.

Basset, A., R.S. Khush, A. Braun, L. Gardan, F. Boccard, J.A.Hoffmann, and B. Lemaitre. 2000. The phytopathogenicbacteria Erwinia carotovora infects Drosophila and activatesan immune response. Proc Natl Acad Sci U S A 97: 3376-81.

Bogdan, C. 2001. Nitric oxide and the immune response. NatImmunol 2: 907-16.

Braun, A., J.A. Hoffmann, and M. Meister. 1998. Analysis of theDrosophila host defense in domino mutant larvae, whichare devoid of hemocytes. Proc Natl Acad Sci U S A 95:14337-42.

Bredt, D.S. and S.H. Snyder. 1994. Nitric oxide: a physiologicmessenger molecule. Annu Rev Biochem 63: 175-95.

Choe, K.M., T. Werner, S. Stoven, D. Hultmark, and K.V.Anderson. 2002. Requirement for a peptidoglycanrecognition protein (PGRP) in Relish activation andantibacterial immune responses in Drosophila. Science296: 359-62.

Dawson, T.M., D.S. Bredt, M. Fotuhi, P.M. Hwang, and S.H.Snyder. 1991. Nitric oxide synthase and neuronal NADPHdiaphorase are identical in brain and peripheral tissues.Proc Natl Acad Sci U S A 88: 7797-801.

Delledonne, M., Y. Xia, R.A. Dixon, and C. Lamb. 1998. Nitric oxidefunctions as a signal in plant disease resistance. Nature394: 585-8.

Diefenbach, A., H. Schindler, M. Rollinghoff, W.M. Yokoyama, andC. Bogdan. 1999. Requirement for type 2 NO synthase forIL-12 signaling in innate immunity. Science 284: 951-5.

Dimopoulos, G., D. Seeley, A. Wolf, and F.C. Kafatos. 1998.Malaria infection of the mosquito Anopheles gambiaeactivates immune- responsive genes during criticaltransition stages of the parasite life cycle. Embo J 17: 6115-23.

Ding, A.H., C.F. Nathan, and D.J. Stuehr. 1988. Release of reactivenitrogen intermediates and reactive oxygen intermediatesfrom mouse peritoneal macrophages. Comparison ofactivating cytokines and evidence for independentproduction. J Immunol 141: 2407-12.

Durner, J., D. Wendehenne, and D.F. Klessig. 1998. Defense geneinduction in tobacco by nitric oxide, cyclic GMP, and cyclicADP-ribose. Proc Natl Acad Sci U S A 95: 10328-33.

Dushay, M.S., B. Asling, and D. Hultmark. 1996. Origins ofimmunity: Relish, a compound Rel-like gene in the

antibacterial defense of Drosophila. Proc Natl Acad Sci U SA 93: 10343-7.

Ferrandon, D., A.C. Jung, M. Criqui, B. Lemaitre, S. Uttenweiler-Joseph, L. Michaut, J. Reichhart, and J.A. Hoffmann. 1998.A drosomycin-GFP reporter transgene reveals a localimmune response in Drosophila that is not dependent onthe Toll pathway. Embo J 17: 1217-27.

Georgel, P., S. Naitza, C. Kappler, D. Ferrandon, D. Zachary, C.Swimmer, C. Kopczynski, G. Duyk, J.M. Reichhart, and J.A.Hoffmann. 2001. Drosophila immune deficiency (IMD) is adeath domain protein that activates antibacterial defenseand can promote apoptosis. Dev Cell 1: 503-14.

Gibbs, S.M. and J.W. Truman. 1998. Nitric oxide and cyclic GMPregulate retinal patterning in the optic lobe of Drosophila.Neuron 20: 83-93.

Gottar, M., V. Gobert, T. Michel, M. Belvin, G. Duyk, J.A. Hoffmann,D. Ferrandon, and J. Royet. 2002. The Drosophila immuneresponse against Gram-negative bacteria is mediated by apeptidoglycan recognition protein. Nature 416: 640-4.

Han, Y.S., J. Thompson, F.C. Kafatos, and C. Barillas-Mury. 2000.Molecular interactions between Anopheles stephensimidgut cells and Plasmodium berghei: the time bombtheory of ookinete invasion of mosquitoes. Embo J 19:6030-40.

Hedengren, M., B. Asling, M.S. Dushay, I. Ando, S. Ekengren, M.Wihlborg, and D. Hultmark. 1999. Relish, a central factor inthe control of humoral but not cellular immunity inDrosophila. Mol Cell 4: 827-37.

Hoffmann, J.A., F.C. Kafatos, C.A. Janeway, and R.A. Ezekowitz.1999. Phylogenetic perspectives in innate immunity.Science 284: 1313-8.

Imamura, M., J. Yang, and M. Yamakawa. 2002. cDNA cloning,characterization and gene expression of nitric oxidesynthase from the silkworm, Bombyx mori . Insect Mol Biol11: 257-65.

Kim, Y.S., S.J. Han, J.H. Ryu, K.H. Choi, Y.S. Hong, Y.H. Chung,S. Perrot, A. Raibaud, P.T. Brey, and W.J. Lee. 2000.Lipopolysaccharide-activated kinase, an essentialcomponent for the induction of the antimicrobial peptidegenes in Drosophila melanogaster cells. J Biol Chem 275:2071-9.

Kopp, E.B. and R. Medzhitov. 1999. The Toll-receptor family andcontrol of innate immunity. Curr Opin Immunol 11: 13-8.

Lemaitre, B., E. Kromer-Metzger, L. Michaut, E. Nicolas, M.Meister, P. Georgel, J.M. Reichhart, and J.A. Hoffmann.1995a. A recessive mutation, immune deficiency (imd),defines two distinct control pathways in the Drosophila hostdefense. Proc Natl Acad Sci U S A 92: 9465-9.

Lemaitre, B., M. Meister, S. Govind, P. Georgel, R. Steward, J.M.Reichhart, and J.A. Hoffmann. 1995b. Functional analysisand regulation of nuclear import of dorsal during theimmune response in Drosophila. Embo J 14: 536-45.

Lemaitre, B., E. Nicolas, L. Michaut, J.M. Reichhart, and J.A.Hoffmann. 1996. The dorsoventral regulatory gene cassettespatzle/Toll/cactus controls the potent antifungal responsein Drosophila adults. Cell 86: 973-83.

Levashina, E.A., E. Langley, C. Green, D. Gubb, M. Ashburner,J.A. Hoffmann, and J.M. Reichhart. 1999. Constitutiveactivation of toll-mediated antifungal defense in serpin-deficient Drosophila. Science 285: 1917-9.

Lu, Y., L.P. Wu, and K.V. Anderson. 2001. The antibacterial arm ofthe drosophila innate immune response requires anIkappaB kinase. Genes Dev 15: 104-10.

Luckhart, S., Y. Vodovotz, L. Cui, and R. Rosenberg. 1998. Themosquito Anopheles stephensi limits malaria parasitedevelopment with inducible synthesis of nitric oxide. ProcNatl Acad Sci U S A 95: 5700-5.

MacMicking, J., Q.W. Xie, and C. Nathan. 1997. Nitric oxide andmacrophage function. Annu Rev Immunol 15: 323-50.

Manfruelli, P., J.M. Reichhart, R. Steward, J.A. Hoffmann, and B.Lemaitre. 1999. A mosaic analysis in Drosophila fat bodycells of the control of antimicrobial peptide genes by the Relproteins Dorsal and DIF. Embo J 18: 3380-91.

Foley & O’Farrell – Genes & Development 2003, 17: 115-125 9

9

Meng, X., B.S. Khanuja, and Y.T. Ip. 1999. Toll receptor-mediatedDrosophila immune response requires Dif, an NF- kappaBfactor. Genes Dev 13: 792-7.

Michel, T., J.M. Reichhart, J.A. Hoffmann, and J. Royet. 2001.Drosophila Toll is activated by Gram-positive bacteriathrough a circulating peptidoglycan recognition protein.Nature 414: 756-759.

Nappi, A.J., E. Vass, F. Frey, and Y. Carton. 2000. Nitric oxideinvolvement in Drosophila immunity. Nitric Oxide 4: 423-30.

Nathan, C. and M.U. Shiloh. 2000. Reactive oxygen and nitrogenintermediates in the relationship between mammalian hostsand microbial pathogens. Proc Natl Acad Sci U S A 97:8841-8.

Nicolas, E., J.M. Reichhart, J.A. Hoffmann, and B. Lemaitre. 1998.In vivo regulation of the IkappaB homologue cactus duringthe immune response of Drosophila. J Biol Chem 273:10463-9.

Ooi, J.Y., Y. Yagi, X. Hu, and Y.T. Ip. 2002. The Drosophila Toll-9activates a constitutive antimicrobial defense. EMBO Rep 3:82-7.

Qiu, P., P.C. Pan, and S. Govind. 1998. A role for the DrosophilaToll/Cactus pathway in larval hematopoiesis. Development125: 1909-20.

Ramet, M., P. Manfruelli, A. Pearson, B. Mathey-Prevot, and R.A.Ezekowitz. 2002. Functional genomic analysis ofphagocytosis and identification of a Drosophila receptor forE. coli. Nature 416: 644-8.

Regulski, M. and T. Tully. 1995. Molecular and biochemicalcharacterization of dNOS: a Drosophila Ca2+/calmodulin-dependent nitric oxide synthase. Proc Natl Acad Sci U S A92: 9072-6.

Reichhart, J.M., M. Meister, J.L. Dimarcq, D. Zachary, D.Hoffmann, C. Ruiz, G. Richards, and J.A. Hoffmann. 1992.Insect immunity: developmental and inducible activity of theDrosophila diptericin promoter. Embo J 11: 1469-77.

Rosetto, M., Y. Engstrom, C.T. Baldari, J.L. Telford, and D.Hultmark. 1995. Signals from the IL-1 receptor homolog,Toll, can activate an immune response in a Drosophilahemocyte cell line. Biochem Biophys Res Commun 209:111-6.

Ruhf, M.L., A. Braun, O. Papoulas, J.W. Tamkun, N. Randsholt,and M. Meister. 2001. The domino gene of Drosophila

encodes novel members of the SWI2/SNF2 family of DNA-dependent ATPases, which contribute to the silencing ofhomeotic genes. Development 128: 1429-41.

Silverman, N., R. Zhou, S. Stoven, N. Pandey, D. Hultmark, and T.Maniatis. 2000. A Drosophila IkappaB kinase complexrequired for Relish cleavage and antibacterial immunity.Genes Dev 14: 2461-71.

Stasiv, Y., M. Regulski, B. Kuzin, T. Tully, and G. Enikolopov. 2001.The Drosophila nitric oxide synthase (dNOS) gene encodesa family of proteins which can modulate NOS activity byacting as dominant negative regulators. J Biol Chem 28: 28.

Stoven, S., I. Ando, L. Kadalayil, Y. Engstrom, and D. Hultmark.2000. Activation of the Drosophila NF-kappaB factor Relishby rapid endoproteolytic cleavage. EMBO Rep 1: 347-52.

Tauszig, S., E. Jouanguy, J.A. Hoffmann, and J.L. Imler. 2000. Toll-related receptors and the control of antimicrobial peptideexpression in Drosophila. Proc Natl Acad Sci U S A 97:10520-5.

Tauszig-Delamasure, S., H. Bilak, M. Capovilla, J.A. Hoffmann, andJ.L. Imler. 2002. Drosophila MyD88 is required for theresponse to fungal and Gram- positive bacterial infections.Nat Immunol 3: 91-7.

Tzou, P., S. Ohresser, D. Ferrandon, M. Capovilla, J.M. Reichhart,B. Lemaitre, J.A. Hoffmann, and J.L. Imler. 2000. Tissue-specific inducible expression of antimicrobial peptide genesin Drosophila surface epithelia. Immunity 13: 737-48.

Underhill, D.M. and A. Ozinsky. 2002. Toll-like receptors: keymediators of microbe detection. Curr Opin Immunol 14:103-10.

Vidal, S., R.S. Khush, F. Leulier, P. Tzou, M. Nakamura, and B.Lemaitre. 2001. Mutations in the Drosophila dTAK1 genereveal a conserved function for MAPKKKs in the control ofrel/NF-kappaB-dependent innate immune responses.Genes Dev 15: 1900-12.

Wingrove, J.A. and P.H. O'Farrell. 1999. Nitric oxide contributes tobehavioral, cellular, and developmental responses to lowoxygen in Drosophila. Cell 98: 105-14.

Wu, L.P. and K.V. Anderson. 1998. Regulated nuclear import of Relproteins in the Drosophila immune response. Nature 392:93-7.

FIGURE 1 Foley and O'Farrell

0

10

100

1000

10000

100000

1000000

241 8 48

cfu

/larv

a

spz/spz (L-NAME)

spz/spz (D-NAME)

spz/+ (L-NAME)

0

10

20

30

40

50

60

70

80

3 4 5 6 7

WT D-NAME L-NAME D-NAME (inj)

L-NAME (inj) D-NAME (Ecc) L-NAME(Ecc)

BA

C

0

10

20

30

40

50

5 6 7 8 9

% o

f la

rvae

ecl

osi

ng

Time after infection (d)

% o

f la

rvae

ecl

osi

ng

D-NAME L-NAME(Ecc015)

D-NAME(Ecc015)

L-NAME

spz/+ (D-NAME)

time (h)

Time after infection (d)

Figure 1. NOS activity contributes to pathogen destruction. (A and B) Depict the ability of third instar wild-type larvae to survive septic (A) and natural infections (B) and eclose. In both cases, inhibiting NOS by feeding L-NAME compromised the ability of the larvae to survive infection with Erwinia carotovora (L-NAME Ecc). In contrast, larvae survived infection when treated with a control isomer of the inhibitor (D-NAME Ecc015) are well as enduring control treatments such as treatment with the drugs with, or without injury (inj). (C) The bacterial titer at different times after Ecc015 natural infection of third instar larvae shows that host killing of bacteria is promoted by Spz-dependent and NOS-dependent processes. While spz heterozygous larvae treated with L-NAME or D-NAME and spz mutants treated with D-NAME successfully destroyed invading bacteria, inactivation of Spz by mutation and NOS by L-NAME prevented host elimination of bacteria.

FIGURE 2 Foley and O'Farrell

A

D E

CB

G

β-g

alac

tosi

das

e ac

tivi

ty (

a.u

.)

0

200

400

600

800

0 3 8 24time (h)

Natural (L-NAME)

Natural (D-NAME)

Septic

0

50

100

150

200

250

300H

β-g

alac

tosi

das

e ac

tivi

ty (

a.u

.)

UninfectedSNAPControl Infected

0

25

50

75

100

%o

f la

rvae

exp

ress

ing

GF

P

F

InfectedUninfected

SNAP L-NAMED-NAMEControl Control

Figure 2. NO signaling activates expression of a Dipt reporter. (A-E) GFP fluorescence in the anterior half of larvae that express GFP under the control of the Dipt promoter (Dipt-GFP). Uninfected larvae do not express the reporter (A) (the weak green signal is background autofluorescence), while infection with Erwinia carotovora carotovora induces reporter activity (B). GFP expression persists in D-NAME treated naturally infected larvae (C), while it is absent in L-NAME treated naturally infected larvae 16 h post infection (D). The NO donor SNAP induces GFP in uninfected larvae (E). (F) Bar graph showing the percentage of larvae expressing GFP 16 h following the different treatments. (G) Graph showing the b-galactosidase activity measured in larval extracts prepared from third instar larvae that express b-galactosidase under the control of the Dipt promoter (Dipt-lacZ) as function of time after infection. While D-NAME fed larvae display a robust b-galactosidase activity 8 h after natural infection, L-NAME fed larvae exhibit a lower b-galactosidase. The value for each time point is the average of three independent measurements The high levels of b-galactosidase activity in extracts from dipt-lacZ larvae exposed to a septic injury with Ecc015 persist somewhat longer than the levels in naturally infected larvae, but the initial responses differ only slightly. (H) Graph showing b-galactosidase activity in Dipt-lacZ larvae prior to treatment with SNAP and 24 h after treatment. Provision of SNAP to uninfected larvae significantly increases b-galactosidase activity. Data are the average of three independent measurements.

uninfected uninfected (SNAP)infected

FIGURE 4 Foley and O'Farrell

A B C

WT

imd

ED

Dipt-lacZ

F

Dipt-lacZ; hsGAL4/UASimd

HG I

Figure 4. NO action in the fat body requires Imd. (A-I) Show dissected fat bodies stained for b-galactosidase activity derived from a Dipt-lacZ reporter transgene. In wild-type larvae carrying the dipt-lacZ transgene (A-C), natural infection or exposure of uninfected larvae to the NO donor SNAP induces Dipt-lacZ expression within 24 h (B and C, respectively: see Fig. 2 H for measurement of induction levels). In contrast, no staining was observed in the fat bodies of imd mutant larvae 24 h after exposure to bacteria or to SNAP (E and F, respectively). hsGAL4 and UAS-imd transgenes allowed indirect induction of Imd expression in response to heat shock. The fat bodies of dipt-lacZ; hsGAL4/UAS-imd larvae showed no detectable lacZ staining in the absence of a heat shock (G), but staining was readily evident 24 h after administration of a heat shock irrespective of whether the larvae had been larvae raised on D-NAME (H) or L-NAME (I) containing fly food.

UninfectedUninfected (SNAP)Infected

A B C

Figure 5 Foley and O'Farrell

Infected (L-NAME)

D

DNA NOS

G

LK

DNA NOS

G

NOSDNA

J

DNA

IH

NOS

E

NOS DNA

F

Figure 5. Expression of Dipt-lacZ in hemocytes requires NOS. (A-D) b-galactosidase activity in hemocytes from Dipt-lacZ larvae. Hemocytes from uninfected larvae have no b-galactosidase activity (A), while a subset of hemocytes from naturally infected larvae show staining (B). Provision of SNAP to uninfected larvae induces b-galactosidase (C), while L-NAME treatment of larvae inhibits induction upon natural infection (D). Bar: 20µm. (E-J) Immunohistochemical staining showing NOS protein levels and Hoechst staining showing DNA in hemocytes from uninfected (E-G) and infected (H-J) larvae. Panels E and F were merged in G and panels H and I were merged in J with NOS in blue and DNA in red. The images in E and H were taken at the same magnification with identical exposure times. (K-L) shows the diaphorase staining of guts dissected from uninfected and naturally infected larvae. The diaphorase staining is significantly lower in the guts of uninfected larvae compared to naturally infected larvae (K and L respectively)

Uninfected Infected Uninfected-SNAP

domino;Dipt-GFP

Drs-GFP; domino

FIGURE 6 Foley and O'Farrell

E

BA C

F G

D

Figure 6. The domino mutant interferes with NO induction of Dipt, but not induction of Drs. (A-D) Expression of a Drs-GFP reporter transgene in the anterior of domino mutant third instar larvae. Localized GFP fluorescence is seen in a subset of uninfected larvae (A, arrow), while the majority of uninfected larvae exhibit detectable fluorescence (B). Natural infection or provision of SNAP to uninfected larvae induces the Drs reporter (C and D, respectively). (E-G) Fluorescence images of domino mutant third instar larvae carrying the Dipt-GFP transgene show that neither bacterial feeding (F) nor exposure to SNAP (G) induce expression of the Dipt reporter.

Hemocyte X

Rel

Fat body

Responses

NOS

NO

Gut

bacteria

Figure 7 Foley and O'Farrell

Figure 7: A model for NO signaling in immunity after natural infection of Drosophila larvae. Receptors, the PRRs, on the surface of sentinel tissues recognize molecules peculiar to pathogens. We suggest that the gut serves such a sentinel function in the case of natural infection, but note that other tissues such as hemocytes might adopt this role upon systemic infection. We propose that, host-pathogen interactions upregulate NOS in sentinel tissues (Figure 5 and Han et al. 2000) and activate a signaling cascade that leads to the release of a cytokine-like activity that recruits additional tissues to defend against the pathogen. We propose that hemocytes function as an intermediary in the Imd pathway that culminates in fat body production of antimicrobial peptides. It should be noted that the classical genes of the Imd pathway might act in different tissues or might act repeatedly at more than one tier of the cascade. As the septic injury model circumvents the gut as a point of entry for pathogens, it is likely that other tissues will act as sentinels mediating the response or the fat body might respond directly under these circumstances.