|Home | About | Journals | Submit | Contact Us | Français|
Following in the footsteps of traditional developmental genetics, research over the last 15 years has shown that innate immunity against bacteria and fungi is governed largely by two NF-κB signal transduction pathways, Toll and IMD. Antiviral immunity appears to stem from RNA interference, whereas resistance against parasitoids is conferred by Toll signaling. The identification of these post-transcriptional regulatory mechanisms and the annotation of most Drosophila immunity genes have derived from functional genomic studies using “model” pathogens, intact animals and cell lines. The D. melanogaster host has thus provided the core information that can be used to study responses to natural microbial and metazoan pathogens as they become identified, as well as to test ideas of selection and evolutionary change. These analyses are of general importance to understanding mechanisms of other insect host–pathogen interactions and determinants of variation in host resistance.
Metazoan animals have mechanisms that allow them to defend themselves against the main threats to their integrity as individuals. These common threats (infection, parasitism, and neoplasia) disturb the delicate balance among cells, tissues and organs. A large variety of defense responses have evolved that differ in their speed and specificity in recognizing and processing a given threat. To simplify this variety of defense responses, immunologists have classified them as either cellular or humoral and either possessing or lacking memory.
Humoral or cellular responses to infectious agents can either be innate or adaptive. The terms “innate”, “natural” or “nonspecific” immunity refer to the variety of physical, cellular or molecular features which provide the first lines of defense against infection or injury. In vertebrates, innate immune responses support and operate alongside adaptive immune responses, to generate antigen-specific mechanisms that eventually lead to the destruction and elimination of the pathogen.
Innate immunity is present in all life forms and the information for innate immune responses is inherited. In mammals, the skin and the epithelial lining of the mucosal tissues act as the primary nonspecific barriers to infection. If infectious agents gain entry into the body through damaged skin, internal innate immune responses are activated. Internal innate immune agents and responses include low pH, proteolytic enzymes and the bile, the secretion of antibacterial substances and acute phase proteins, activation of complement and phagocytosis.
Phagocytosis is a fundamental innate immune mechanism used by a variety of cell types, including macrophages and dendritic cells. Widely distributed in the human body and associated with different tissues, these cells consume micro-organisms, other foreign substances and apoptotic cells. Dendritic cells reside in an immature state in most tissues and develop from the same hematopoietic precursors that give rise to the macrophages. When they come in direct contact with the pathogens, dendritic cells mature, increase their capacity to present antigen and gain the ability to activate naïve T cells of the adaptive immune system (Janeway et al., 2004).
Cells in the mammalian innate immune system detect “microbial nonself” by recognizing pathogen-associated molecular patterns (PAMPs) (Medzhitov & Janeway, 2002). PAMPs are products of microbial metabolism conserved over evolution and distributed in a wide variety of pathogens. Lipopolysaccharide (LPS), peptidoglycan, flagellin, double-stranded RNA and hypomethylated DNA islands are examples of PAMPs. PAMP receptors, called pattern recognition receptors (PRRs) are present on surfaces of host cells and recognize specific or related PAMPs. Toll-like receptors (TLRs) are one class of PRRs expressed on cells of the innate immune system. TLRs directly sense non-self, and when activated, PRRs induce intracellular signaling, that results in the activation of genes involved in host defense. Genes typically activated include those encoding antimicrobial peptides, inflammatory cytokines or chemokines (Fig. 1).
Adaptive or acquired immunity is characterized by specificity and memory. Developed more recently with the evolution of the gnathostomata, the acquired immune response has nearly infinite adaptability. Yet, adaptive immunity depends on the innate immune responses, as in organisms where both systems are present, adaptive immunity is fashioned by innate sensing mechanisms and the ensuing responses. For a fully functional immune system, these components must act in synergy (Hoebe et al., 2004; Fig. 1).
Because of the absence of adaptive immune response (Fig. 1), Drosophila serves as a powerful model for studying aspects of the innate immune system that might otherwise be masked by the adaptive immune response. This system has been a focus of intense analysis: Genome-wide studies of immune responses are now routinely performed in Drosophila: microarrays and proteomic approaches complement RNA interference and conventional genetic methods, expanding the range of experimental possibilities.
Like vertebrates, Drosophila spp. defend themselves against a range of parasites and microbes by invoking a variety of innate immune responses (Fig. 2), some of which are clearly shared with vertebrates. The immune responses include: (i) epithelial barriers (cuticle, trachea, gut) and clotting; (ii) humoral reactions of antimicrobial peptides produced by the fat body; and (iii) cellular reactions (blood cell- or hemocyte-dependent phagocytosis of microbes and encapsulation of larger pathogens). In addition, melanization accompanies coagulation, wound healing and encapsulation, adding to the armor of host defense (Hultmark, 2003; Hoffmann, 2003; Brennan & Anderson, 2004; Imler & Bulet, 2005).
Perhaps the most powerful and the best-studied of these reactions is the systemic immune response, in which antimicrobial peptides are secreted from cells of the fat body into the hemolymph. This response is induced by the invading microbial pathogens. There is ample evidence that in the adult fly, antimicrobial peptides limit bacterial or fungal growth. In general, flies with mutations in genes encoding specific components of the NF-κB (Toll and Imd) pathways are impaired in their ability to fight bacterial or fungal infections. Microarray analyses of all fly genes confirm the central importance of these post-transcriptional mechanisms in the expression of a vast majority of immunity genes. A brief overview of these physical and molecular aspects of innate immune responses in Drosophila follows.
Epithelia deter bacteria and act as the first line of defense. Epithelial cells of the epidermis and those that line the digestive (gut), respiratory (trachea), and reproductive systems are normally in contact with microbes and they serve as natural routes of microbial infection (Tzou et al., 2000; Onfelt et al., 2001). Transgenic strains carrying AMP promoter-GFP fusions reveal that AMP genes are expressed in surface epithelia locally in a cell- or tissue-specific manner. This expression may be constitutive (e.g., Drosomycin in the female spermatheca; Tzou et al., 2000) or inducible (e.g., Diptericin or Drosomycin in trachae and gut). The production of the inducible epithelial AMPs is at least partially under the control of the Imd pathway (reviewed in Lematire & Hoffmann, 2007).
Pathogenic bacteria ingested in food do not appear to be harmful to the wild-type fly host, apparently because of an antioxidant system involving the activities of dual oxidase (Duox) and an extracellular immune-regulated catalase (Ryu et al., 2006; Haet al., 2005a, 2005b). RNAinterference-mediated silencing of either of these genes disturbs the homeostatic redox balance in the gut lumen and increases susceptibility of flies to bacteria. These results indicate that immune-regulated catalase and Duox activities are important for regulating reactive oxygen species levels and protecting fly hosts against gut infection (Ha et al., 2005a, 2005b).
After the primary epithelial barrier is breached, the infected animal erects a second barrier in the form of a clot that traps microbes and also limits the loss of hemolymph (Scherfer et al., 2004). Proteomic studies have identified a dozen proteins making up the clot fibers. Chief among these is hemolectin, a multidomain protein expressed in plasmatocytes, one of the larval hemocytes. Fondue, another abundant hemolymph protein produced by plasmatocytes is also required for optimal coagulation. Animals deficient in either of these proteins remain healthy even after immune challenge; however in both cases, RNAi mutants form larger melanized scabs after wounding with a tungsten needle relative to controls (Goto et al., 2003; Scherfer et al., 2004, 2006), suggesting that these proteins play a role in wound healing. Clot formation can proceed independently of melanization, as pro-phenol oxidase-compromised mutants form normal clots (Scherfer et al., 2004). The process of hemolymph coagulation is thought to be similar to the production of extracellular fibrous traps by mammalian neutrophils after cytokine activation. These traps kill extracellular bacteria independently of phagocytosis (Brinkmann et al., 2004).
A number of microbial pathogens (viral, bacterial and fungal) and metazoan pathogens (parasitoid wasps) infect and kill D. melanogaster (Table 1). They differ in their modes of infection and effects on host physiology. A healthy immune-competent fly host coordinates a set of immune defenses; however natural pathogens have in turn evolved to circumvent some of these host defense mechanisms.
Bacteria can infect flies during ingestion or via injuries. Bacterial strains of Serratia marcescens (Db11), Erwinina carotovara (Ecc15) and Pseudomonas entomophila trigger an immune response in Drosophila after oral infection (Basset et al., 2000, 2003). However, most studies in the adult fly have been done using injection as the mode of infection. Bacterial pathogens studied after injection include Escherichia coli, Micrococcus luteus, Pseudomonas aeruginosa, Serratia marcescens, Salmonella typhimurium and Mycobacterium marinum (reviewed in Lemaitre & Hoffmann, 2007). Even though a manual injection perturbs the balance of host immune (and other) responses, each bacterial species evokes a fairly characteristic set of responses, making the fly host ideal for genetic dissection of processes and pathways governing innate immunity.
Wolbachia is a natural bacterial endosymbiont of Drosophila, frequently found in flies in natural and laboratory conditions, and normally transmitted vertically from a female host to her offspring. Wolbachia are able to colonize different tissues although they prefer to infect cells of reproductive tissues. Recent studies (Lazzaro et al., 2006; Cox & Gilmore, 2007) mark the beginning of the characterization of other natural microbial flora in the D. melanogaster gut and hemolymph. Both Gram-positive (Enterococcus spp., Lactococcus lactis) and Gram-negative (Acetobacter pasteurianus, Serratia marcescens, Providencia burhodogranaria) bacteria colonize Drosophila (Table 1). It is not known whether the fly host resists colonization of these species. Determining this is essential to whether flies have a stable microbial flora and if these populations play an ecological role in the bacteria–fruit fly consortia.
Beauveria bassiana, a fungus that parasitizes insects, can grow on adult flies if its spores are applied to the host. Because of this natural ability to infect flies, B. bassiana has been used as the “model” fungus for studying fungal-specific responses in Drosophila (Lemaitre et al., 1997; De Gregorio et al., 2001, 2002). Candida albicans, a common human pathogen, has also been used to examine phagocytosis in intact Drosophila and in cell culture experiments (Alarco et al., 2004; Stroschein-Stevenson et al., 2006).
DNA-binding studies of promoters of antimicrobial peptide genes (Engstrom et al., 1993; Kappler et al., 1993) and the discovery of Dif (Ip et al., 1993) and Relish (Dushay et al., 1996) gave impetus to the idea that NF-κB signaling may be important in immune-inducibility of AMPs. Genetic studies of mutant animals demonstrated that flies can discriminate between classes of surface molecules on pathogens, and further, two distinct NF-κB pathways, the Toll and IMD pathways, control AMP production by recruiting the functions of NF-κB transcription factors Dorsal/Dif and Relish respectively (Fig. 3).
The two pathway model hinged on the discovery of the Imd (for immune deficiency) mutation (Lemaitre et al., 1995). A combination of forward genetic screens and reverse genetic analyses led to the identification of eight additional Imd pathway components (Fig. 3). The Toll pathway is responsible for resistance to fungal and Gram-positive bacterial infections. The two pathway model is strengthened by evidence from additional studies including genome-wide microarray analyses (De Gregorio et al., 2001, 2002). The cytoplasmic components of both these signaling mechanisms have homologs in other insects (mosquitoes, Christophedos et al., 2002; bees, Evans et al., 2006) where they are likely to play a similar function in host defense. The mammalian homologs of these proteins also regulate innate immune responses. Evolution has thus preserved the essential circuitry of innate immunity.
The intracellular components of the Toll pathway were identified in genetic screens for maternal-effect mutations affecting dorsal-ventral polarity of the fly embryo (Govind, 1999). Unlike vertebrate Toll-like receptors, the fly trans-membrane receptor Toll1 does not bind directly to pathogens. Instead, Toll1 is activated by a processed form of Spätzle, a cysteine-knot protein with structural similarities to nerve growth factor. Toll activation leads to destabilization of the cytoplasmic complex of IκB protein Cactus and NF-κB proteins Dorsal/Dif via intracellular components MyD88, Tube and the Pelle kinase. Activation of the Toll pathway results in degradation of the IκB-family protein Cactus and localization of Dorsal and Dif to the nucleus. Dorsal/Dif activate Drosomycin expression (Fig. 3). Dorsal and Dif function redundantly and cell-autonomously in the larval fat body (Manfruelli et al., 1999). Dif is essential in the adult response (Rutschmann et al., 2000). Toll signaling remains repressed by functions of Cactus and the SUMO-conjugating enzyme dUbc9 (Chiu et al., 2005). The extracellular components of the Toll pathway necessary for Spätzle activation in embryonic development and immunity are different (see below). Mutations in most of these well-conserved intracellular signaling components render flies susceptible to fungal and Gram-positive bacterial infections (Hoffmann, 2003; Hultmark, 2003, Tanji & Ip, 2005).
Unlike the Toll pathway that also regulates other processes in the fly life cycle, the Imd pathway appears to be dedicated to defending hosts against Gram-negative bacteria (Martinelli & Reichhart, 2005; Lemaitre & Hoffmann, 2007). Imd protein contains a death domain similar to the one in Receptor Interacting Protein of the TNF-receptor. The main receptor of the Imd pathway is peptidoglycan recognition protein-LC (PGRP-LC; see below). Components downstream of Imd are TAK1, TAB2, DIAP2, IKKβ/ird5, IKKγ/Kenny, the FADD adaptor, the Dredd caspase, and the transcription factor Relish (Fig. 3). The IKKβ/ird5 and IKKγ/Kenny form a complex equivalent to the mammalian IKK signalosome. The IKK complex is responsible for phosphorylation of Relish. Relish is a compound Rel-family protein with a Rel domain and ankyrin-repeat containing inhibitory domain. Phosphorylated Relish is cleaved by Dredd. Relish cleavage in the cytoplasm releases the Rel domain from its inhibitory domain, which remains in the cytoplasm. The Imd pathway remains repressed by the ubiquitin-proteasome system (Khush et al., 2002). The order of action of these proteins (see Fig. 3) was deduced from epistatic experiments in which typically one “gain-of-function” component was combined with the “loss-of-function” mutation in a second component. If the outcome of such a combination results in the suppression of the “gain-of-function” phenotype (i.e., constitutive expression of the AMP gene is reduced), then the second component is assumed to be acting “downstream” of the first.
In general, the kinetics of Drosomycin and Diptericin mRNA expression, the canonical “read-outs” of the Toll and Imd pathways, respectively, are quite different: For example, Drosomycin expression peaks a day after injury, whereas Diptericin expression peaks within the first 12 hours of septic injury. However, together, both these NF-κB pathways account for about 80% of the genes induced by septic injury (De Gregorio et al., 2001, 2002). While there is no evidence of cross-talk between the pathways, some AMP genes can be induced by both of them, whereas others show specific input from a single pathway.
In addition to the Toll and Imd pathways, other signaling mechanisms complement and contribute to the core circuitry. The expression of a subset of the immune-responsive genes such as the thiol-ester protein Tep1 and the stress response gene TotA also require input from JAK-STAT signaling (Lagueux et al., 2000; Agaisse et al., 2003). While the initial steps leading to receptor activation are less clear, events downstream of Domeless are carried out by the Janus Kinase Hopscotch and the transcription factor STAT92E. Thus, some aspects of the systemic immune response are complex and require the coordination and integration of different cell types and regulatory mechanisms (Agaisse et al., 2003). JAK-STAT-deficient flies are resistant to bacterial and fungal infections but succumb to viral infections (Dostert et al., 2005). Like the Toll pathway, the JAK-STAT pathway also regulates other physiological and developmental functions in the fly.
One of the long-standing questions in the field of insect immunity has been how, in the absence of antibodies and the major histocompatibility complex, the insect immune system distinguishes between foreign and self-antigens. In flies, members of the PRR families (PGRPs and GNBPs) sense the presence of non-self. PGRPs can bind to a specific form of peptidoglycan (PGN, a PAMP molecule) that is present in both Gram-positive and Gram-negative bacteria. There are 13 PGRPs in the fly, and all share a 160-amino acid PGRP domain (Werner et al., 2000; Royet et al., 2005). PGRPs can bind (and recognize) PGN. They may also possess key zinc-binding residues that confer a zinc-dependent N-acetylmuramoyl-L-alanine amidase activity. It has been proposed that the amidase activity of this latter class of catalytic PGRPs (PGRP-SC1A/B, LB, SB1/2 and SC2) removes peptides from the glucan chains of PGN, altering its biological activity (Mellroth et al., 2003; Mellroth & Steiner, 2006; Zaidman-Remy et al., 2006). Members of the non-catalytic class of PGRPs serve as receptors and include PGRP-LC (the major receptor of the Imd pathway) as well as PGRP-SA, SD, LA, LD, LE and LF. The three-dimensional structures of four of these proteins have been solved. These structural studies provide insights into the binding/recognition-specificity versus catalytic functions of PGRP family members (reviewed in Lematire & Hoffmann, 2007).
Gram-positive and Gram-negative PGN show subtle but important differences in chemical structure and localization. The Gram-positive PGN (lys-type) is multilayered and exposed at the bacterial surface, whereas Gram-negative PGN (DAP-type) is single-layered and found within the periplasmic space under the outer bacterial membrane. Both these differences are thought to be important in the specificity of pathogen recognition. The Toll pathway is activated by the lys-type PGN, whereas the Imd pathway is activated by DAP-type PGN (Leulier et al., 2003; Kaneko et al., 2004; Stenbak, 2004). The second important difference in their activation is that the Toll pathway is activated by secreted members of the PGRP family (PGRP-SA and SD), whereas the Imd pathway is activated by membrane-bound or intracellular receptors (PGRP-LC and LE).
It is commonly accepted that Spätzle activation and binding triggers the Toll pathway (Weber et al., 2007). The steps intervening pathogen recognition and Spätzle processing are not completely understood. However, it is believed that a zymogen activation cascade like the one that activates Spätzle in the embryo (Govind, 1999; Hashimoto et al., 2003; Weber et al., 2003) is also responsible for processing Spätzle after immune challenge. This idea is supported by the discovery of the Spätzle Processing Enzyme (SPE) that is highly similar to Easter, the equivalent protease in the early embryo (Jang et al., 2006; Kambris et al., 2006).
Events upstream of SPE appear to be distinct, depending on the nature of the pathogen: Gram-positive bacteria are recognized by PGRP-SA, PGRP-SD and Gram-negative binding protein-1 (GNBP-1, an apparent misnomer). Immune-deficient phenotypes of PGRP-SA and GNBP-1 mutant flies are similar and there is evidence that these two proteins act in concert (Gobert et al., 2003; Pili-Floury et al., 2004). The presence of fungi is recognized by GNBP3 (Gottar et al., 2006), a circulating PRR. GNBP3 mutants are immune-compromised and cannot activate Toll signaling. GNBP3 belongs to a small family with proteins containing an N-terminal portion that binds β (1,3)-glucan and the C-terminal catalytic β-glucanase domain. Fungal infection is mediated by an extracellular cascade involving the Clip-serine protease Persephone (Ligoxygakis et al., 2002) and the serine protease inhibitor Necrotic (Levashina et al., 1999). Both these enzymes control Spätzle activation although in opposite ways: Persephone function is needed for Spätzle activation, whereas Necrotic is a negative regulator and its inhibitory function is necessary to limit Toll signaling.
The Imd pathway is activated primarily by Gram-negative PGN whose presence is detected by the membrane-bound PGRP-LC. The PGRP-LC gene is alternatively spliced to result in three isoforms: LCa, LCx and LCy. These isoforms share the intracellular domain but differ in their extracellular structures (Werner et al., 2000; Kaneko et al., 2004). The isoforms show differences in specificity to monomeric versus polymeric PGN and two PGRP-LC molecules may come together to bind polymeric PGN and activate signaling (Reiser et al., 2004; Mellroth et al., 2005; Chang et al., 2005, 2006). PGRP-LE detects DAP-type PGN and can function as a receptor both inside and outside the cell (Kaneko et al., 2006). The Imd pathway is thought to be attenuated by the amidase activities of PGRP-LB (Zaidman-Remy et al., 2006) and PGRP-SC (Bischoff et al., 2006). The amidase activity processes peptides from the glycan chain of PGN. This mechanism is thought to contain the intensity and duration of the Imd response.
Humoral defense responses in insects include localized melanization, lectin binding, and induction of antimicrobial peptides (AMPs) via activation of the IMD and Toll pathways (Fig. 3). Humoral factors such as antimicrobial proteins (AMPs) and complement-like proteins called Drosophila thiol-ester-containing proteins, (or dTEPs) are thought to be the primary immune effector molecules. Roughly 20 AMP genes in Drosophila represent seven classes: antifungal AMPs (drosomycin, metchnikowin), anti-Gram-negative AMPs (attacins, cecropins, diptericins, drosocin), and an anti-Gram-positive AMP (Defensin) (Martinelli & Reichhart, 2005). AMPs are synthesized mostly in fat body cells and released into the hemolymph. However, cells of barrier tissues such as the midgut (Diptericin, Attacin), the tracheae (Drosocin, Drosomycin), the oviduct (Drosocin, Cecropin), the seminal receptacle and the spermatheca (Cecropin, Defensin, Drosomycin) also express AMPs (Tzou et al., 2000) as part of the local defense mechanism. Antimicrobial peptides are small and cationic and are thought to affect microbial membrane properties (Imler & Bulet, 2005). The three-dimensional structures of Cecropin, Defensin, Drosocin, Drosomycin have been determined, although the structural basis of their activities is not understood.
The Drosophila genome encodes six members of the dTEP gene family (Lagueux et al., 2000). dTEPs are acute-phase glycoproteins that belong to the C3/α2-macroglobulin superfamily. Through their thioester motifs, TEPs form covalent links to bacterial surfaces. In mammals, similar modification of bacterial surfaces by the complement protein C3 marks microbes for opsonization. Because these molecules are expressed in the fat body and presumably secreted into the hemolymph, dTEPs are considered to be part of the humoral response. The expression of TEPs is under the control of the JAK-STAT pathway (Lagueux et al., 2000).
More than 25 viruses have been documented to infect Drosophila (Brun & Plus, 1980; Ashburner et al., 2005), and the effects of few of these have been studied recently (Lemaitre & Hoffmann, 2007; Cherry & Silverman, 2006). Viruses can spread by vertical infection from the egg to adult as is the case with the Sigma virus. Alternatively, they are transmitted horizontally (feeding or contact) as in the case of Drosophila C virus (DCV) or Drosophila X virus (DXV). DCV is a picorna-like RNA virus (Johnson & Christian, 1998). The DCV genome is single-stranded positive genomic RNA. DXV is a Birnavirus, with two linear, double-stranded RNA molecules (Zambon et al., 2005). In general, even though virus-infected flies may appear otherwise normal, injections of adults with virus make them sensitive to carbon dioxide exposure and cause early death of the animal (Seecoff, 1965; Cherry & Perrimon, 2004; Zambon et al., 2005).
The Toll and JAK-STAT pathways are involved in resistance against DXV (Zambon et al., 2005) and DCV (Dostert et al., 2005). However, neither of these mechanisms appears to be central to the antiviral defense in the fly. Instead, flies employ the RNA interference pathway (Fig. 4) to resist virus infection. (Wang et al., 2006; Galiana et al., 2006; Zambon et al., 2006). Several core members of the RNAi machinery such as Dicer-2, r2d2, Argonaute-2 and piwi are important in executing the antiviral response to DXV, DCV and other unrelated infectious viruses (cricket paralysis virus, flock house virus): mutations in these loci lead to increased sensitivity to viral infection and enhanced mortality with higher viral load in mutant adults. Thus, like the role of NF-κB signaling in innate immunity, RNAi shows structural and functional conservation in host defense from plants to mammals.
Recent studies on the Drosophila C virus provide interesting insights into Drosophila-DCV interaction. DCV injections into the fly abdomen result in dose-dependent lethality (Cherry & Perrimon, 2004). As mentioned above, the RNA-induced silencing compex RISC component Argonaute 2, which has endonculease activity is essential for protecting the host from virus-induced mortality (Galiana-Arnoux et al., 2006; Wang et al., 2006; van Rij et al., 2006). In ago-2 mutant flies, an increased accumulation of viral RNA levels is observed and this increase is accompanied by higher RNA titers. Interestingly, DCV encodes a suppressor of the RNAi machinery (van Rij et al., 2006). This suppressor specifically inhibits only RNAi silencing (that is initiated by long dsRNA) but apparently not miRNA biogenesis or function. Tentatively named DCV-1A, the suppressor gene encodes an RNA-binding protein with a canonical dsRNA-binding domain (dsRBD). This protein (and its N-terminal 99 amino acids containing the dsRBD) can suppress RNAi. Suppression of RNAi is dependent on the highly conserved residues of the RNA-binding domain. Transgenic flies expressing DCV-1A can suppress the dsRNA-induced silencing of the white gene in adult flies. Finally, DCV-1A has dsRNA-binding activity and it can inhibit Dicer processing of dsRNA. The authors (van Rij et al., 2006) propose that by binding to the long double-stranded RNA, DCV-1A inhibits Dicer-2-mediated dsRNA processing into siRNAs. Thus DCV-1A protein allows DCV to overcome the RNAi defense mechanism of the host.
Microbes in the hemolymph are consumed by phagocytic cells called plasmatocytes. Plasmatocytes are the predominant hemocyte type in healthy animals and they carry receptors for recognizing bacteria, yeast, viruses and apoptotic cells. Pathogen recognition is followed by internalization. Recent studies have identified a number of phagocytic receptors. In addition, cultured cells of Drosophila called S2 are highly phagocytic. This makes it easy to perform genome-wide RNAi screens for genes involved in the phagocytic process. These studies show that there are many similarities in pathogen recognition and engulfment between flies and mammals.
(i) The IgSF-domain containing protein isoforms Dscam (E. coli, Watson et al., 2005); (ii) members of the class C scavenger receptor (dSR-CI, dsRNA; Ramet et al., 2001); (iii) PGRP family (PGRP-SC1a; Staphylococcus aureus; Garver et al., 2006); (iv) epidermal growth factor-domain containing protein Eater (dsRNA, E. coli, S. aureus; Kocks et al., 2005); and (v) CD36-related protein Peste (Mycobacterium, Listeria; Philips et al., 2005). These receptors show selectivity in pathogen binding/internalization (reviewed in Cherry & Silverman, 2006; Lemaitre & Hoffmann, 2007).
In addition to these receptor molecules, TEPVI (also called Mcr, for macroglobulin complement-related and a TEP family member without the active thioester motif) is important in the uptake of the fungus Candida albicans, but not that of S. cerevisae. Other TEP family members also modestly contribute to phagocytosis of E. coli (TepII) or S. aureus (TepIII), suggesting that these complement-like proteins facilitate microbe recognition or uptake (Stroschein-Stevenson et al., 2006). Significantly, Anoph-eles mosquito has 19 TEPs, and TEP1 of A. gambiae has been shown to mediate parasite killing (Blandin et al., 2004).
Cell culture studies (Cherry & Perrimon, 2004; Agaisse et al., 2005; Philips et al., 2005; Benghezal et al., 2006; Stroschein-Stevenson et al., 2006) have identified proteins required after pathogen internalization. These include proteins important for actin cyctoskeletal reorganization (such as actin capping proteins, Arp2/3 complex, cdc42, and cofilin) and those involved in clathrin-mediated endocytosis pathway or in vesicle trafficking. In vivo, the requirement of WASp homolog D-Scar and Chickadee has been shown (Pearson et al., 2003). These recent findings set the stage for a fine-grained analysis of steps involved in pathogen recognition, internalization, and the subsequent intracellular steps that lead to microbe destruction.
In addition to microbial pathogens, parasitoid wasps infect Drosophila larvae and pupae. In some natural populations, more than half of Drosophila larvae collected are infected, indicating that wasps can be a potent cause of mortality. Parasitoids develop within the larval body cavity. Most parasitoid species studied in the lab are highly virulent on the D. melanogaster. With their sharp ovipositors, female wasps inject one or more eggs into the host hemocoel. Leptopilina, Asobara, Ganaspis and Trichopria species infect Drosophila (Table 1), although, of these, host interactions with the Leptopilina spp. are the best studied. L. boulardi, L. heterotoma and L. victoriae are highly virulent but. L. boulardi differs from the other two in its host range restricted to the melanogaster group of flies (Schlenke et al., 2007).
Being much larger than the 10-μ plasmatocyte, the wasp egg cannot be phagocytosed. Instead, hemocytes (plasmatocytes, crystal cells and lamellocytes) cooperate and physically surround and encase the egg. This encapsulation reaction involves hemocyte division and differentiation. This reaction is relatively slow and capsules are not observed until 1–2 days after infection. Whether (and how quickly) the egg becomes encapsulated also depends on where in the larval hemocoel the egg is localized after oviposition, the physicochemical features of the egg itself and the presence of immune-suppressive factors or virus-like particles.
Encapsulation of the wasp egg involves at least three phases: (i) recognition of wasp egg; (ii) the differentiation of highly flattened and adhesive blood cells called lamellocytes; and (iii) construction of the capsule. Lamellocyte precursors reside in the larval lymph gland (Lanot et al., 2001; Sorrentino et al., 2002) although their existence in circulation has not been ruled out. Components of the Toll pathway are also expressed in the lymph gland (Qiu et al., 1998) and Toll signaling is required for wasp egg encapsulation. Many other genes and pathways, especially JAK-STAT signaling also contribute to differentiation of hemocytes after immune challenge (Crozatier & Meister, 2007). Mutations in components of Toll and JAK-STAT pathways affect hemocyte counts and compromise the host’s ability to mount an effective encapsulation reaction in response to L. boulardi infection (Sorrentino et al., 2004).
Construction and melanization of a capsule involve hundreds of cells and require the coordinated action of many genes. The capsule is constructed largely of lamellocytes and these highly mobile cells organize to construct a multi-layered capsule with septate junctions and integrin-dependent cell interactions (Irving et al., 2005). Melanization is the blackening of the capsule surrounding the parasite egg. Melanin formation involves multiple steps that include the oxidation of monophenols and diphenols. These steps are catalyzed by the enzyme pro-phenol oxidase or proPO. The importance of the phenoloxidase cascade in the encapsulation reaction is controversial; some evidence against its involvement comes from the analysis of phenoloxidase-deficient mutants which encapsulate parasite eggs but fail to melanize the capsule (Rizki et al., 1980). However, in a genome-wide microarray study, several genes encoding enzymes of the melanization cascade were found to be upregulated by L. boulardi infection (Schlenke et al., 2007 see below).
Surprisingly, larvae infected by Asobara tabida or Leptopilina boulardi, but not L. heterotoma result in the upregulation of the Toll/Imd pathways and expression of antimicrobial peptides (Wertheim et al., 2005; Schlenke et al., 2007). Whole genome microarrays of hosts infected by each of these wasps reveals that hosts infected by A. tabida or L. boulardi induce numerous genes encoding proteolytic enzymes, components of the Toll and JAK/STAT pathways, and the melanization cascade as part of a combined cellular and humoral immune response. However, flies attacked by L. heterotoma do not initiate much of an immune transcriptional response. The difference could be attributable to the rapid VLP-mediated lysis of host hemocytes by L. heterotoma (see below). These studies suggest that the systemic immune responses initiated by L. boulardi or A. tabida play a role in host defense against parasitoid infections. The molecules involved in parasitoid egg recognition are not known, although microarray studies reveal that parasitoid infection activates transcription of specific PGRP genes, suggesting that their products may serve to recognize the presence of the foreign invaders (Werhteim et al., 2005; Schlenke et al., 2007).
The cynipid sister species L. heterotoma and L. victoriae produce virus-like particles (VLPs) that are implicated in suppression of egg encapsulation. VLPs are produced in the long gland-reservoir complex, a tripartite structure connected to the ovipositor and associated with the ovary (Rizki & Rizki, 1990; Morales et al., 2005). VLPs are deposited in the host hemocoel during oviposition. VLPs of L. heterotoma and L. victoriae have the ability to lyse lamellocytes in vitro and in vivo (Rizki & Rizki, 1990; Morales et al., 2005). Infection by these parasitoids also results in depletion of the hematopoietic precursors in the lymph gland (Chiu et al., 2002). L. heterotoma infection does not activate AMP gene expression (Schlenke et al., 2007) but it is not known whether the active immune suppressive factors that target hemocytes also suppress or inactivate gene expression in the fat body. Future research in this system will reveal how the activation of cellular and humoral immune pathways are coordinated by some wasps (A. tabida and L. boulardi) and how immune suppressive or evasive mechanisms of other species (L. heterotoma and L. victoriae) contribute to parasitoid virulence and determine host range.
Many human conditions such as allergy, autoimmune disorders, rheumatoid arthritis-related inflammation and septic shock are linked to dysfunctional innate or innate/adaptive immune responses. For many of these conditions, the molecular genetic lesions have not been pinpointed. For example, it is not known if these problems arise due to a failure of self/non-self discrimination, if the downstream signaling mechanisms are anomalous, or if the connections at the innate/adaptive interface are aberrant. The crucial role of the innate immune system in limiting parasitic or microbial diseases in humans is also not well understood and its analysis is complicated by the presence of adaptive mechanisms that are closely tied to innate mechanisms. The Drosophila immune system shares many features of the mammalian innate immune system at tissue, cell and molecular levels, is multifaceted and is now relatively well characterized. It is therefore a powerful model for study of the more subtle and complex aspects of the infection process, including disease progression after infection by (one or more) natural pathogens, coordination and intersection of different “arms” of the immune system, and many aspects of host physiology and immune-competence. Given that the basic scaffold of innate immune regulation is highly conserved, it will be interesting to uncover the nature and evolution of virulence in diverse organisms. These studies will provide a clearer picture of how complex biological processes and systems, such as adaptive immunity, were assembled, maintained, and altered in evolution.
Primary references can be found in the recent and extensive reviews cited here. Financial support from USDA (NRI/USDA CSREES 2006-03817), NIH (G12-RR03060 and GM08-618) and PSC-CUNY is gratefully acknowledged.