Elimination of Plasmatocytes by Targeted Apoptosis Reveals Their Role in Multiple Aspects of the Drosophila Immune Response

Elimination of plasmatocytes by targeted apoptosis reveals their role in multiple aspects of the Drosophila immune response

Bernard Charroux and Julien Royet1

Institut de Biologie du De´veloppement de Marseille Luminy, Unite´Mixte de Recherche 6216, Centre National de la Recherche Scientiﬁque, Universite´dela Méditerrannée Aix-Marseille II, 13288 Marseille Cedex 9, France

Communicated by Jules A. Hoffmann, Centre National de la Recherche Scientifique, Strasbourg, France, April 17, 2009 (received for review January 14, 2009) Drosophila hemocytes have strong phagocytic capacities and pro- generating plasmatocyte-depleted individuals and by analyzing duce antimicrobial peptides (AMPs). However, the precise role of their development and immune response. blood cells during immune responses and developmental processes has only been studied using indirect means. To overcome this Results limitation, we generated plasmatocyte-depleted flies by specifi- To obtain flies devoid of plasmatocytes, we decided to trigger cally overexpressing the proapoptotic protein Hid into plasmato- cell death in blood cells by overexpressing the proapoptotic cytes. Unexpectedly, these plasmatocyte-depleted animals have a protein Hid using the plasmatocyte (and crystal cells)–specific normal larval and pupal development and do not exhibit any hml(⌬)-Gal4 driver (20, 21). A UAS-EGFP transgene was used obvious defect after birth. Remarkably, plasmatocyte-depleted to precisely characterize the spatiotemporal expression of the adults show a strong susceptibility to infections by various micro- hml(⌬)-Gal4 driver. GFP is not expressed during embryogenesis, organisms, although activation of systemic AMP gene transcription even in late embryos in which Croquemort-positive blood cells via the Toll and immune deficiency (IMD) pathways is wild-type. are detected (Fig. 1A). The expression of GFP in plasmatocytes Our data show that this susceptibility, which correlates with appears shortly after hatching in early first-instar larvae (Fig.

overproliferation of bacteria, is likely due to the absence of 1B), is maintained throughout the larval stages (Fig. 1F and data IMMUNOLOGY phagocytosis. We also demonstrate that during larval stages, not shown), and remains in the adult (Fig. 1G). GFP-positive plasmatocytes play an essential role in mediating AMP production cells are also detected in the cortical zone of the larval lymph by the fat body after oral bacterial infection. Finally, we show that gland (Fig. 1C) (12, 22). In addition to the circulating and sessile plasmatocytes are involved in immune surveillance during pupal plasmatocytes, we identified 2 yet-uncharacterized blood cell development, because they prevent bacterial infection that causes populations located in close contact to the gut cells. One group pupal lethality. is composed of approximately 50 cells, which are positioned within the larval proventriculus, where they surround the esoph- AMP ͉ blood cells ͉ innate immunity ͉ Toll agus (Fig. 1D). More posteriorly, in the larval ventriculus, few scattered GFP-positive cells are found within the gut epithelium he cellular interactions and pathways involved in the develop- (Fig. 1E). Some of these cells span the entire epithelium and are Tment and function of blood cells are, to a certain extend, therefore in contact with both the gut lumen and the hemolymph conserved between vertebrates and invertebrates (1–3). As in (supporting information Fig. S1). These cells, which are still present vertebrates, Drosophila blood cell development occurs in 2 phases. in adult guts (Fig. S1), are also labeled by the srpHemo-Gal4 driver, A first wave takes place in the embryonic procephalic mesoderm suggesting that these are indeed plasmatocytes (Fig. S1). Interest- and gives rise to both plasmatocytes and crystal cells (4, 5). ingly, both cell groups are present in gut regions competent to Plasmatocytes, the major function of which is phagocytosis, are the produce antimicrobial peptides after oral infection by the Gram- negative bacteria Erwinia carotovora (Fig. 1 D and E). Confident predominant hemocyte population present at all developmental ⌬ stages (6). In the embryo they engulf apoptotic corpses formed that the hml( )-Gal4 driver is blood cell specific, we then asked during developmental processes (4, 7, 8). In larvae and adults they whether it is indeed panhemocytic. We observed that more than 96% (ϩ3 SD) of the circulating hemocytes labeled by DAPI (n ϭ are responsible for phagocytosis of invading bacteria and fungi (9, ⌬ 10). At the pupal stage they play a fundamental role by engulfing 4585) in hml( )-Gal4, UAS-EGFP individuals are GFP positive and recycling doomed cells during metamorphosis (6). In addition (Fig. 1K). The remarkable tissue specificity of this Gal4 driver and to their phagocytic function, plasmatocytes also produce and se- its broad plasmatocyte distribution prompted us to use it as a tool to eliminate blood cells in vivo. crete a number of peptides/proteins, such as antimicrobial peptides (AMPs) (11). The crystal cells, which persist until the onset of hml(⌬)-Gal4, UAS-EGFP, UAS-hid Individuals Can Reach the Adult metamorphosis, mainly contribute to larval melanization reactions Stage. We first characterized the phenotype of Drosophila of the (5). The second wave of hematopoiesis occurs at the larval stage in hml(⌬)-Gal4, UAS-EGFP, UAS-hid/ϩ (later called hml-hid) the lymph gland. The precursors of the lymph gland derive from the genotype. As expected from the hml(⌬)-Gal4 expression pattern, dorsal thoracic mesoderm and subsequently grow during the first embryonic plasmatocytes were not ablated in these embryos and second larval instars (12–15). At the third-instar larval stage, (Fig. S2). However, third-instar larval plasmatocytes of the same the lymph gland is the main site of hemocyte production. The lymph genotype were strongly affected by Hid overexpression. First, these gland is also the production site of lamellocytes, which are cells devoted to the encapsulation of foreign bodies too large to be phagocytosed (6, 13, 16). The contribution of hemocytes to Dro- Author contributions: B.C. designed research; B.C. performed research; B.C. and J.R. ana- sophila development and immunity has so far been addressed by lyzed data; and B.C. and J.R. wrote the paper. blocking phagocytosis after bead injection or by using mutants with The authors declare no conﬂict of interest. a reduced number of hemocytes (17–19). Because these mutants 1To whom correspondence should be addressed. E-mail: [email protected]. show pleiotropic phenotypes and affect tissues other than hemo- This article contains supporting information online at www.pnas.org/cgi/content/full/ cytes, we decided to re-evaluate the function of plasmatocytes by 0903971106/DCSupplemental.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903971106 PNAS Early Edition ͉ 1of6 Downloaded by guest on September 29, 2021 A B C D E

F G H I

J K LM

N O P Q

Fig. 1. Obtainment of plasmatocyte-depleted Drosophila. Although numerous Croquemort-positive cells are present in hml(⌬)-Gal4, UAS-EGFP embryos (A), these cells do not express GFP. Gal4 expression, visualized through GFP signal, is present in hml(⌬)-Gal4, UAS-EGFP first- (B) and third-instar larvae (F) and in adults (G). A magnification of a hml(⌬)-Gal4, UAS-EGFP/dome-MESO–LacZ (labeling the medullary zone) third-instar larval lymph gland (C) shows that only peripheral differentiated cells from the cortical zone express Gal4.(D and E) Two populations of GFP-positive cells are found in the larval gut. One is located in the proventriculus and surrounds the digestive tract (dashed lines) (D). The second is found in the most posterior region of the ventriculus (E). Some blood cells (arrow) span the entire gut epithelium and therefore contact both the gut lumen and the hemolymph. Ectopic expression of the proapoptotic protein Hid using the hml(⌬)-Gal4 driver gives rise to larvae (H) and adults (I) that are completely devoid of GFP-positive plasmatocytes. (J–M) When bled onto a coverslip, hemolymph from hml(⌬)-Gal4, UAS-EGFP, UAS-hid/ϩ larvae (L) contain approximately 40% of the number of DAPI-positive cells observed in hml(⌬)-Gal4, UAS-EGFP/ϩ control larvae (J). These cells are GFP negative (compare K with M), suggesting that they are dying. This is confirmed by the fact that more than 96% of these DAPI-positive cells are TUNEL positive (P and Q). This number is less than 5% in hml(⌬)-Gal4, UAS-EGFP/ϩ control larvae (N and O).

larvae were almost totally devoid of GFP-positive cells (Fig. 1 H and role of blood cells during immune response using our model of M). Second, when bled onto a coverslip, we observed that hml-hid plasmatocyte-depleted Drosophila. larval hemolymph contains 40% (Ϯ4 SD; n ϭ 1,702) DAPI-positive cells when compared with controls (n ϭ 4,263; Fig. 1 J and L). Plasmatocytes Are Dispensable for Immune Deficiency (IMD) and Toll Third, more than 96% (Ϯ2 SD) of these remaining DAPI-positive Pathways Activation in Adults. To test the role of plasmatocytes cells were undergoing apoptosis, as indicated by TUNEL staining during the immune response, we compared the ability of control (Fig. 1 P and Q). The number of TUNEL-positive cells in controls and hml-hid adults to survive infection. Flies devoid of plas- was much lower (5%, Ϯ2 SD; Fig. 1 N and O). Thus, we estimated matocytes showed a clear increased susceptibility to infection by that hml-hid larvae possess less than 2% of nonapoptotic plas- both Gram-positive and Gram-negative bacteria (Fig. 2). These matocytes. Because plasmatocytes have been described as par- observed susceptibilities were weaker in plasmatocyte-depleted ticipating in tissue remodeling by phagocytosing larval tissues adults than in mutant for genes in either the Toll (spz)orthe rm7 during pupariation, we were surprised to note that a significant IMD (PGRP-LC) pathways. Whereas all spz mutant flies percentage (45%, Ϯ12 SD; n ϭ 432) of the hml-hid larvae could succumbed to infection by Enterococcus faecalis within 28 h, only reach the pharate adult stage and emerge as phenotypically 67% of the plasmatocyte-depleted adults were dead 68 h after normal adults (Fig. 1I and Fig. S3) (6). Such adults contain very infection. Similar susceptibility was observed when flies were infected with fungi (Fig. 2C). As expected, this lethality was few remaining plasmatocytes (5%, Ϯ2.1 SD; n ϭ 9) when correlated with a faster and higher bacterial growth in compared with controls (set up to 100%; n ϭ 176). This plasmatocyte-depleted vs. wild-type flies (Fig. 3A). In addition, observation also emphasizes the fact that other mutants used a possible explanation for the observed phenotypes could be that previously to study hemocyte immune functions, such as domino, plasmatocytes are required to mount normal antimicrobial pep- l-(3)hem,orpsidin, which are all larval lethal, are very likely tide production during the systemic immune response. To test affected in tissues other than hemocytes (17–19, 23–25). It this hypothesis, we compared Toll and IMD pathway activation should also be noted that unlike these larval-lethal mutants that in wild-type and plasmatocyte-depleted flies. Toll pathway ac- have delayed development, plasmatocyte-depleted individuals tivation quantified by Drosomycin transcription was clearly not reach the pupal stage with the same timing as wild-type siblings reduced in the plasmatocyte-depleted background but was (data not shown). The temporal delay of domino, l-(3)hem, and rather upregulated after infection by various Gram-positive psidin mutants could account at least partially for the immune bacteria (Fig. 3B and data not shown). No obvious effects of the phenotypes observed earlier, given that the intensity of the absence of plasmatocytes on the level of both Diptericin and immune response has clearly been correlated with developmen- Attacin-A, 2 IMD pathway targets, were noted after Gram- tal timing and ecdysone titers (26). We also noted that l-(3)hem negative infection (Fig. 3B). Defensin transcription, which relies and domino larvae contain residual hemocytes of abnormal size on Toll and IMD pathway activation, was clearly upregulated and shape (Fig. S4). Finally, unlike domino mutants for instance, after Gram-negative infection. Kinetics of Toll and IMD- our plasmatocyte-depleted animals do not contain melanotic dependent AMP gene transcription were similar in control and tumors or any other obvious phenotypes (data not shown). plasmatocyte-depleted flies during the first 8 h after infection Altogether, these observations prompted us to re-evaluate the (Fig. 3C and data not shown). Note that the production of

2of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903971106 Charroux and Royet Downloaded by guest on September 29, 2021 IMMUNOLOGY Fig. 2. Plasmatocyte-depleted adults are susceptible to various microorganisms. Survival curves after infection by pricking with various microorganisms. E shows that spa¨tzle mutant ﬂies simultaneously plasmatocyte-depleted have a stronger susceptibility to infection by E. faecalis than spa¨tzle mutants. Error bars are shown as vertical bars. Survival curves with an asterisk (*) are statistically signiﬁcantly different (P Ͻ 0.05) from controls.

various AMPs was not affected in noninfected phagocyte- and posterior ends of the larva (Fig. 4 E and F), others in the depleted animals (Fig. 3), showing that plasmatocyte killing does entire fat body (Fig. 4 A–D), and some larvae did not show any not, by itself, trigger the production of AMPs. These results sign of IMD pathway activation in the fat body (data not shown). indicate that plasmatocytes are not required per se for systemic The percentage of larvae with obvious IMD pathway activation immune response in adults. They also indicate that plasmatocyte- was significantly reduced in the plasmatocyte-depleted larvae depleted flies in which AMPs are produced at wild-type levels (Fig. 4 D and M). Interestingly, in infected domino larvae, nevertheless exhibit increased susceptibility to a number of micro- Dpt-cherry activation in the fat body was always almost com- organisms, thus demonstrating the essential role of plasmatocytes pletely restricted to the anterior and posterior lobes (Fig. 4 G and in Drosophila immune response. One mechanism by which plas- H). Furthermore, Dpt-cherry activation was not uniform as in matocytes could participate in resistance to infection is phagocy- controls but patchy, with only few activated cells in a given fat tosis. To test this hypothesis, we challenged either eater mutant flies body lobe (Fig. 4 I and J). Because domino encodes a ubiquitous or flies in which the phagocytic receptors Nimrod or Eater were SWI2/SNF2 DNA-dependent ATPase, we propose that the specifically inactivated by RNAi in the plasmatocytes with various effects previously reported in domino mutants are rather due to microorganisms. Both mutants reportedly exhibit reduced phago- modifications in transcriptional competence of fat body cells cytic activities (9, 27). Survival curves obtained with eater or Nimrod than to a reduction of blood cell numbers (17). All these results flies are very similar to those observed with plasmatocyte-depleted were confirmed by quantitative RT-PCR measurement (Fig. flies (Fig. 2D), suggesting that plasmatocytes act mainly via phago- 4N). It should be noted that we did not observe any reduction of cytosis. Our data also indicate that AMP production by itself is not Diptericin, Attacin-A,orDefensin activation in l-(3)hem mutants sufficient to mount a full immune response. The fact that when we allowed larvae to reach late L3 stage (Fig. 4N). This plasmatocyte-depleted flies mutant for the spz gene show a higher suggests that the remaining hemocytes in l-(3)hem larvae are susceptibility to infection than spz mutant flies shows, as previously sufficient to convey the signal from the digestive tract to the fat proposed, that AMP production acts together with but indepen- body. In addition, we showed that larvae with reduced phago- dently of blood cells to fight infection (Fig. 2E) (15). cytosis capacity (i.e., eater larvae) are as competent as wild-type larvae to induce AMP genes in the fat body after oral infection Larval Plasmatocytes Are Required to Mount an Efficient Immune (Fig. 4N). This suggests that phagocytosis is not a major con- Response After Natural Infection. When larvae are orally infected tributor in this plasmatocytes-dependent mechanism. However, with the bacteria E. carotovora, a systemic IMD-dependent the role of phagocytosis in this process cannot be completely immune response is triggered in fat body cells (17). The mech- ruled out, given that eater mutants do not totally abolish but only anisms by which these bacteria, which are unable to cross the gut reduce plasmatocyte-dependent phagocytosis. epithelium, signal from the gut to the fat body remain unknown. We assessed the role of plasmatocytes in this particular aspect of Gut-Associated Blood Cells Are Not Required for Local AMP Produc- the immune response. As usual with systemic immune reactions tion. As mentioned earlier, blood cells positive for both hemolec- triggered by orally transmitted bacteria, the response in wild- tin and serpent expression are located in 2 domains of the gut, type larvae was heterogeneous. Some showed activation of the around the esophagus in the proventriculus and at the most Dpt-cherry reporter only in fat body lobes close to the anterior posterior part of the ventriculus. To test whether these plas-

Charroux and Royet PNAS Early Edition ͉ 3of6 Downloaded by guest on September 29, 2021 essential for local Dpt activation (Fig. S6). Targeted removal of this larval plasmatocyte subpopulation will be required to specifically investigate their function(s).

Lethality of Plasmatocyte-Depleted Pupae Is Suppressed in Axenic Conditions. Although plasmatocyte-depleted larvae are able to develop into adults, we noticed that 55% of hml-hid individuals died during postembryonic development, more specifically at the pupal stage (Fig. 5 and data not shown). As expected, we found that this lethality was completely rescued by coexpressing the antiapoptotic baculovirus P35 protein, which also restores plasmatocyte GFP- positive cells both in larvae and adults (Fig. 5 and Fig. S7). This specific phenotype could either be explained by a requirement of plasmatocytes during the metamorphosis process itself or could be interpreted as a need for plasmatocytes to ensure a function of immune surveillance against pathogens during normal larval and pupal stages. To distinguish between these hypotheses, we raised wild-type and plasmatocyte-depleted stocks in the presence of antibiotics. Results presented in Fig. 5 unambiguously show that the pupal lethality associated with depletion of plasmatocytes was almost completely suppressed by antibiotic treatment. This indicates that plasmatocytes play an immune function during developmental stages to prevent bacteria-induced lethality. The fact that such pupal lethality is not observed in eater mutants suggests that phagocytosis may not be involved in this plasmatocyte function. Further experiments are now needed to dissect the mechanisms by which plasmatocytes protect pupae against bacteria-induced lethality. Discussion By overexpressing the Hid protein, we generated Drosophila containing only few remaining plasmatocytes. This suggests that plasmatocytes are largely dispensable from third-instar larval and pupal development. We cannot, however, completely rule out that the few remaining plasmatocytes can successfully perform some essential function, such as removal of histolyzed tissue during metamorphosis. Our results show that plasmatocytes play various roles in setting up an efficient immune response. For larval immune response, plasmatocytes are critical for an efficient activation of AMP production by the fat body cells. They might be required to convey a signal from the digestive tract to the fat body, after oral infection by the Gram-negative bacteria E. carotovora. Bacterial peptidoglycan (PGN), which is the main bacterial immune elicitor in flies, could Fig. 3. AMP production in plasmatocyte-depleted flies. (A) Bacterial growth be such a signal. Alternatively, hemocytes could produce after infection by E. faecalis, presented as the relative number of E. faecalis factor(s) that facilitate activation of the immune signaling pathways present per fly 20 h after infection. Gray bars, independent experiments; black by circulating PGN. The fact that Dipt activation is not totally bars, mean ϩ SD. *P Ͼ 0.05. Other differences are statistically significant (**P Ͻ abolished in hml-hid individuals indicates that plasmatocytes con- 0.05). Similar results were obtained with other Gram-positive bacteria, such as tribute only partially to this process and that other plasmatocyte- S. aureus.(B) AMP mRNA inductions are shown in control and plasmatocyte- independent mechanisms are involved. Our results also indicate depleted adults. mRNAs inducibility measured in control flies (UAS-hid/ϩ) was that, in contrast to what has been published earlier, l (3) hem mutant set to 100, and values obtained with other genotypes are expressed as a larvae behave as wild-type controls when allowed to reach the same percentage of this value. Each histogram corresponds to the mean value of 6 experiments in which the SD is shown. Values indicated by identical symbols developmental stage. In contrast to what happens after larval oral (*, ***, or ***) are not significantly different (P Ͼ 0.05) from one another. All infection, we show here that blood cells are not required in adults other differences are statistically significant (P Ͻ 0.05). Similar results were to mount normal Toll and IMD pathway activation after septic obtained with other Gram-positive bacteria, such as S. aureus and E. faecalis injuries. It is therefore unlikely that, as proposed before, plasmato- (not shown). (C) Quantitative RT-PCR analysis of Drosomycin induction after cytes produce a signal (such as the Toll ligand Spa¨tzle) necessary to M. luteus septic injury in adults. Relative ⌬CtDrosomycin/⌬Ctrp49 ratios of unchal- activate the Toll pathway in the fat body during systemic immune lenged UAS-hid/ϩ controls were anchored in 1 to indicate fold induction. response (28). The overactivation of the Toll pathway observed in Graphs represent the mean and SD of relative ratios detected in 3 biologic plasmatocyte-depleted flies rather suggests that plasmatocytes play repetitions of a pool of 5 females. Histograms with a single asterisk (*) are a role to dampen Toll pathway activation. By phagocytosing some significantly (P Ͻ 0.05) different from control genotypes. All other differences are not statistically significant (P Ͼ 0.05). of the invading bacteria, plasmatocytes could reduce the amount of available Toll pathway elicitor. Alternatively, plasmatocytes could dampen Toll pathway activation by producing negative regulators of matocytes are required for AMP production by the gut epithe- immune signaling, such as Serpins, which have been shown to lium, we compared Dpt-cherry expression patterns in guts from inhibit the cleavage of Spz (29). Consistently, serpin genes are control and plasmatocyte-depleted larvae after oral infection by expressed in plasmatocytes, and the transcription levels of some of E. carotovora. No difference could be detected between the 2 them are downregulated upon infection (28). The normal produc- genotypes, indicating that gut-associated plasmatocytes are not tion of AMP in adult-depleted flies also has implications for our

4of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903971106 Charroux and Royet Downloaded by guest on September 29, 2021 A B G H

C D I J

E F K L

M IMMUNOLOGY

Fig. 4. Larval plasmatocytes are required for full induction of an immune systemic response after oral infection. Larvae were orally infected with E. carotovora and analyzed for Dpt-cherry expression. Three classes of larvae were observed: (i) larvae showing IMD pathway activation in the entire fat body (A and B), (ii) larvae in which the reporter is only expressed in the anterior and posterior fat body lobes (E and F), and (iii) larvae in which no reporter activation is noticeable (data not shown). When plasmatocyte-depleted larvae were orally infected, the number of nonresponsive larvae increases at the expense of larvae with global fat body IMD pathway activation (M). Each histogram corresponds to the mean value of 5 independent experiments (200 larvae analyzed per experiment) in which the SD is shown. Empty bars, independent experiments; dark gray bars, mean Ϯ SD. Histograms with a single asterisk (*) are not significantly different (P Ͼ 0.05) from controls. Histograms with two asterisks (**) are statistically significant (P Ͻ 0.05) from controls. Whereas fat body of control or plasmatocyte- depleted larvae exhibited a uniform expression of Dpt-cherry in the fat body lobes (K and L), this expression was patchy in domino fat bodies with fewer cells expressing the transgene (G–J). (N) AMP mRNA inductions are shown in control, plasmatocyte-depleted, eater, l (3)hem, and domino mutant larvae orally infected with E. carotovora. AMP mRNA inducibility measured in control flies (UAS-hid/ϩ) was set to 100, and values obtained with other genotypes are expressed as a percentage of this value. Histograms correspond to the mean Ϯ SD of 5 experiments. *P Ͼ 0.05. Other differences are statistically significant (**P Ͻ 0.05).

understanding of the mode of activation of peptidoglycan recog- senting’’ PGN for detection by PGRP-LC. Our results demonstrate nition protein LC (PGRP-LC). The IMD pathway is mainly acti- that this is not the case, given that IMD pathway activation by vated by Gram-negative bacteria (30–32). In these bacteria PGN is PGRP-LC is normal in the absence of plasmatocytes. Finally, we buried underneath the LPS layer and therefore is not directly found that plasmatocytes are involved in immune surveillance accessible for recognition by the membrane receptor PGRP-LC. It during pupal development: they prevent bacterial infection that has been proposed that plasmatocytes could play a role by ‘‘pre- causes pupal lethality. One possible explanation would be that

Charroux and Royet PNAS Early Edition ͉ 5of6 Downloaded by guest on September 29, 2021 Septic Injuries, Fly Survival Experiments, and Bacterial Growth. For each set of experiments, triplicate groups of 25 6-day-old ﬂies were infected by pricking the thorax with a thin tungsten needle dipped into bacteria. Results are expressed as percentage of living ﬂies at different time points after infection. Experiments shown are representative of at least 5 independent experiments. The following OD was used: E. faecalis OD ϭ 100 (except for Fig. 2 E,ODϭ 5), Pseudomonas entomophila OD ϭ 0.5, and Candida albicans OD ϭ 100.

TUNEL Assay. TUNEL assay was performed on hemocytes using the Apotag Red In Situ Kit (Chemicon). The value obtained with hml(⌬)-Gal4, UAS-EGFP/ϩ larvae was set up to 100, and values obtained with hml(⌬)-Gal4, UAS-EGFP, UAS-hid/ϩ larvae are expressed as a percentage of this value. The data Fig. 5. Lethality of plasmatocyte-depleted pupae can be rescued by antibiotics presented in Fig. 1 are representative of 3 independent experiments. treatment. More than 50% of the hml(⌬)-Gal4, UAS-GFP, UAS-hid/ϩ individuals die during postembryonic development, during the pupal stage. This lethality is Immunostaining on Embryos and Larvae. Standard procedures were used for suppressed by coexpression of the antiapoptotic protein P35, demonstrating the embryo and larval staining, except that signals were amplified using the specificity of the effect observed. When plasmatocyte-depleted flies are raised on Tyramide Signal Amplification kit (NEN Life Science). Primary antibodies were food supplemented by antibiotics, most of the pupal lethality is suppressed. The rabbit anti-Crq (1:500; ref. 7) and rabbit anti-GFP (1:500; Molecular Probes). number of adult flies of the control genotype UAS-hid/ϩ was set up to 100, and values obtained with other genotypes are expressed as a percentage of this value. Natural Infection of Larvae. Three hundred fifty microliters of an overnight Histograms correspond to the mean Ϯ SD of 5 experiments. *P Ͼ 0.05. Other bacterial culture of Ecc-15 (OD ϭ 200) or Luria-Bertani (LB) were directly added differences are statistically significant (**P Ͻ 0.05). on top of feeding third-instar larvae into medium at 25 °C. Larvae were monitored for Diptericin transcription by fluorescence analysis using the Dpt-cherry reporter and by quantitative RT-PCR.

during tissue remodeling, which occurs throughout metamorphosis, Quantitative Real-Time PCR. Real-time quantitative PCR was performed as bacteria normally constrained to the gut lumen can accidentally described previously (33). Rp49, Diptericin, Attacin-A, Defensin, and Droso- gain access to the pupal tissue. If not kept in check by plasmatocytes, mycin primers have been described previously (33). these bacteria can proliferate and kill the pupae. Interestingly, this control seems independent of AMP production because double Cloning of Dpt-cherry Construct. A fragment containing the Diptericin gene upstream sequences from position Ϫ2266 to ϩ42 (34) was cloned upstream of mutants for the IMD and Toll pathways (Dif, key mutant flies; Fig. the cherry coding sequence, into a pUAST plasmid. Molecular detail can be 5) do not show such pupal lethality. obtained upon request.

Materials and Methods ACKNOWLEDGMENTS. We thank Floriane Bertaina and Ahmed Fatmi, who participated in some of the experiments; and Marie Meister and Thomas Rival Drosophila melanogaster. Stocks were raised on standard cornmeal-agar me- for comments on the manuscript. This work was supported by Centre National dium at 25 °C. For antibiotic treatment, 150 ␮L of a solution containing 5 de la Recherche Scientiﬁque, Agence Nationale de la Recherche, Fondation mg/mL of ampicillin and kanamycin were added to each vial. pour la Recherche Médicale, and Institut Universitaire de France.

1. Evans CJ, Banerjee U (2003) Transcriptional regulation of hematopoiesis in Drosophila. 19. Braun A, Hoffmann JA, Meister M (1998) Analysis of the Drosophila host defense in Blood Cells Mol Dis 30:223–228. domino mutant larvae, which are devoid of hemocytes. Proc Natl Acad Sci USA 2. Evans CJ, Hartenstein V, Banerjee U (2003) Thicker than blood: Conserved mechanisms 95:14337–14342. in Drosophila and vertebrate hematopoiesis. Dev Cell 5:673–690. 20. Goto A, Kadowaki T, Kitagawa Y (2003) Drosophila hemolectin gene is expressed in 3. Crozatier M, Meister M (2007) Drosophila haematopoiesis. Cell Microbiol 9:1117–1126. embryonic and larval hemocytes and its knock down causes bleeding defects. Dev Biol 4. Tepass U, Fessler LI, Aziz A, Hartenstein V (1994) Embryonic origin of hemocytes and 264:582–591. their relationship to cell death in Drosophila. Development 120:1829–1837. 21. Sinenko SA, Mathey-Prevot B (2004) Increased expression of Drosophila tetraspanin, 5. Lebestky T, Chang T, Hartenstein V, Banerjee U (2000) Specification of Drosophila Tsp68C, suppresses the abnormal proliferation of ytr-deficient and Ras/Raf-activated hematopoietic lineage by conserved transcription factors. Science 288:146–149. hemocytes. Oncogene 23:9120–9128. 6. Lanot R, Zachary D, Holder F, Meister M (2001) Postembryonic hematopoiesis in 22. Jung SH, Evans CJ, Uemura C, Banerjee U (2005) The Drosophila lymph gland as a Drosophila. Dev Biol 230:243–257. developmental model of hematopoiesis. Development 132:2521–2533. 7. Franc NC, Dimarcq JL, Lagueux M, Hoffmann J, Ezekowitz RA (1996) Croquemort, a 23. Braun A, Lemaitre B, Lanot R, Zachary D, Meister M (1997) Drosophila immunity: Analysis novel Drosophila hemocyte/macrophage receptor that recognizes apoptotic cells. of larval hemocytes by P-element-mediated enhancer trap. Genetics 147:623–634. Immunity 4:431–443. 24. Foley E, O’Farrell PH (2003) Nitric oxide contributes to induction of innate immune 8. Franc NC, Heitzler P, Ezekowitz RA, White K (1999) Requirement for croquemort in responses to gram-negative bacteria in Drosophila. Genes Dev 17:115–125. phagocytosis of apoptotic cells in Drosophila. Science 284:1991–1994. 25. Brennan CA, Delaney JR, Schneider DS, Anderson KV (2007) Psidin is required in 9. Kocks C, et al. (2005) Eater, a transmembrane protein mediating phagocytosis of Drosophila blood cells for both phagocytic degradation and immune activation of the bacterial pathogens in Drosophila. Cell 123:335–346. fat body. Curr Biol 17:67–72. 10. Ramet M, et al. (2001) Drosophila scavenger receptor CI is a pattern recognition 26. Ligoxygakis P, Bulet P, Reichhart JM (2002) Critical evaluation of the role of the Toll-like receptor for bacteria. Immunity 15:1027–1038. receptor 18-Wheeler in the host defense of Drosophila. EMBO Rep 3:666–673. 11. Dimarcq JL, et al. (1997) Treatment of l(2)mbn Drosophila tumorous blood cells with 27. Kurucz E, et al. (2007) Nimrod, a putative phagocytosis receptor with EGF repeats in the steroid hormone ecdysone amplifies the inducibility of antimicrobial peptide gene Drosophila plasmatocytes. Curr Biol 17:649–654. expression. Insect Biochem Mol Biol 27:877–886. 28. Irving P, et al. (2005) New insights into Drosophila larval haemocyte functions through 12. Krzemien J, et al. (2007) Control of blood cell homeostasis in Drosophila larvae by the genome-wide analysis. Cell Microbiol 7:335–350. posterior signalling centre. Nature 446:325–328. 29. Levashina EA, et al. (1999) Constitutive activation of toll-mediated antifungal defense 13. Crozatier M, Ubeda JM, Vincent A, Meister M (2004) Cellular immune response to in serpin-deficient Drosophila. Science 285:1917–1919. parasitization in Drosophila requires the EBF orthologue collier. PLoS Biol 2:E196. 30. Gottar M, et al. (2002) The Drosophila immune response against Gram-negative 14. Mandal L, Martinez-Agosto JA, Evans CJ, Hartenstein V, Banerjee U (2007) A Hedge- bacteria is mediated by a peptidoglycan recognition protein. Nature 416:640–644. hog- and Antennapedia-dependent niche maintains Drosophila haematopoietic pre- 31. Ramet M, Manfruelli P, Pearson A, Mathey-Prevot B, Ezekowitz RA (2002) Functional cursors. Nature 446:320–324. genomic analysis of phagocytosis and identification of a Drosophila receptor for E. coli. 15. Elrod-Erickson M, Mishra S, Schneider D (2000) Interactions between the cellular and Nature 416:644–648. humoral immune responses in Drosophila. Curr Biol 10:781–784. 32. Choe KM, Lee H, Anderson KV (2005) Drosophila peptidoglycan recognition protein LC 16. Rizki TM, Rizki RM (1992) Lamellocyte differentiation in Drosophila larvae parasitized (PGRP-LC) acts as a signal-transducing innate immune receptor. Proc Natl Acad Sci USA by Leptopilina. Dev Comp Immunol 16:103–110. 102:1122–1126. 17. Basset A, et al. (2000) The phytopathogenic bacteria Erwinia carotovora infects Dro- 33. Maillet F, Bischoff V, Vignal C, Hoffmann J, Royet J (2008) The Drosophila peptidogly- sophila and activates an immune response. Proc Natl Acad Sci USA 97:3376–3381. can recognition protein PGRP-LF blocks PGRP-LC and IMD/JNK pathway activation. Cell 18. Ruhf ML, et al. (2001) The domino gene of Drosophila encodes novel members of the Host Microbe 3:293–303. SWI2/SNF2 family of DNA-dependent ATPases, which contribute to the silencing of 34. Reichhart JM, et al. (1992). Insect immunity: Developmental and inducible activity of homeotic genes. Development 128:1429–1441. the Drosophila diptericin promoter. EMBO J 11:1469–1477.

6of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903971106 Charroux and Royet Downloaded by guest on September 29, 2021