Journal ClubImmunology

Drosophila Toll Pathway: The New Model

See allHide authors and affiliations

Science Signaling  06 Jan 2009:
Vol. 2, Issue 52, pp. jc1
DOI: 10.1126/scisignal.252jc1


In Drosophila, recognition of microbe-specific molecules (such as bacterial peptidoglycans) activates serine protease cascades that converge to activate the Toll pathway. Recent data show that the serine protease Grass, which is activated downstream of pattern recognition receptors and was initially thought to be a component only of the Gram-positive bacteria–induced signaling cascade, is also required for the induction of the Toll pathway after fungal infection. Persephone, a serine protease known to be specifically activated by fungal proteases, was also found to be required for sensing Gram-positive bacterial proteases. Thus, Persephone serves as a sensor for microbial activities from both fungi and Gram-positive bacteria. With these new discoveries, a new model has been proposed for activation of the Drosophila Toll pathway.

In Drosophila, the Toll and immunodeficiency pathways are the major innate immune signaling pathways that sense microbes in hemolymph, the insect equivalent of blood. The Toll pathway in Drosophila is activated by fungi and also detects Gram-positive bacteria with lysine-type peptidoglycans (Lys-type PGNs) in their cell walls (i.e., most Gram-positive bacteria). Some Gram-positive bacteria, such as Bacillus, have meso-diaminopimelic acid–type peptidoglycans (DAP-type PGNs) similar to those of Gram-negative bacteria, and these do not activate the Toll pathway [reviewed in (1)]. These microbe-specific molecules, known as pathogen-associated molecular patterns (PAMPs), are recognized by a set of germline–encoded receptors called pathogen recognition receptors (PRRs). Although vertebrate Toll-like receptors (TLRs) function as PRRs, the Drosophila Toll receptor does not function as a PRR but instead serves as a cytokine receptor activated by its ligand Spätzle (Spz), a cytokine secreted in the hemolymph (2). In response to Toll activation, the fat body (the equivalent of liver in insects) produces antimicrobial peptides (AMPs) and secretes them into the hemolymph to kill microbes.

The Drosophila Toll pathway is activated by proteolytic cascades initiated through both PAMP-dependent and PAMP-independent mechanisms. Lys-type PGNs are recognized by the peptidoglycan recognition protein SA (PGRP-SA), PGRP-SD, and glucan-binding protein 1 (GNBP1) [reviewed in (1)]. Glucans (polymers of d-glucose) present in the fungal cell wall are detected by GNBP3, the fungal cell wall PRR (3). In contrast, a proteolytic cascade thought to be independent of PRRs relies on a serine protease called Persephone (Psh) that is activated by fungal virulence factors (substances that enhance the infectivity of the microbe) (4). Psh activation is independent of GNBP3 (3). Instead, Psh detects virulence factors such as proteases and chitinases secreted by spores of fungi that infect insects (entamopathogenic fungi) (3). These virulence factors degrade the cuticle to enable the fungi to gain entry into the host.

According to the classical model, these were thought to represent three independent proteolytic cascades that converged on Spätzle processing enzyme (SPE), which cleaves the cytokine Spz into a ligand that can activate the Toll receptor (5). The serine proteases Grass (Gram-positive–specific serine protease) and Psh were respectively thought to be components of the Gram-positive bacteria signaling cascade and of the fungal virulence factor pathway (Fig. 1A). However, research by El Chamy et al. has demonstrated a role for Grass in fungal detection and for Psh as a sensor for Gram-positive bacterial virulence factor activity (6). This research has had a profound impact on how we view Drosophila Toll receptor signaling that has made us reevaluate the model proposed earlier.

Fig. 1

(A) The Drosophila Toll pathway (old model). Three independent proteolytic cascades lead to activation of the Toll pathway. The first cascade, involving the protease Grass, is activated downstream of the peptidoglycan recognition protein SA (PGRP-SA), PGRP-SD, and glucan-binding protein 1 (GNBP1) detecting Lys-type PGNs from Gram-positive bacteria. The second cascade is activated by glucans from fungi, which are detected by GNBP3. The third cascade relies on a serine protease called Persephone (Psh) that is activated by fungal proteases. These cascades converge on SPE, which cleaves Spz, the Toll receptor ligand, into its active form. (B) The proposed model of the Drosophila Toll pathway. Two independent proteolytic cascades lead to the activation of the Toll pathway: the PRR-dependent cascade and the danger signal–derived cascade. The PRR-dependent cascade, comprising the protease Grass, is activated upon recognition of conserved pathogen-associated molecular patterns (PAMPS) by cognate PRRs. The danger signal extracellular cascade, comprising the protease Psh, is activated by microbial secreted proteases or abnormal proteolytic activities in the hemolymph. Glucans and Lys-type PGNs from fungi and Gram-positive bacteria activate their respective PRRs, which culminate on Grass, which in turn activates other proteases that activate SPE to cleave Spz. Cleaved Spz acts as a ligand for the Toll receptor.

The role of Grass in Toll pathway activation was identified in a study that used RNA interference (RNAi) in vivo (7). In this study, four proteases other than Grass were shown to be involved in detecting both fungi and Gram-positive bacteria. In experiments in which flies were infected with fungi and grass was knocked down with hairpin RNA constructs, Toll activation was not impaired, which suggests that Grass is not involved in fungal detection (7). RNA-mediated interference was used against grass, which led to a decrease in grass mRNA of only 60%—a reduction that is inadequate to eliminate a role for grass in fungal detection (7). Thus, El Chamy et al. generated homozygous grass null mutant flies by mobilizing a P-element inserted 500 base pairs upstream of the grass gene; this led to an imprecise excision of about 1200 base pairs, resulting in deletion of the first 101 amino acids of the protein. This mutant allele was named grassHerrade. As expected, these flies showed decreased activation of the Toll pathway when challenged with Gram-positive bacteria and reduction of life span in survival assays when challenged with Gram-positive bacteria. Unexpectedly, however, these mutant flies also showed reduced Toll activation after entamopathogenic fungal infection, implicating Grass in the sensing of fungal infections (6).

Because Grass and Psh were both shown to be implicated in responding to fungal infections, El Chamy et al. reconsidered the involvement of Psh in bacterial infection. After infection with the Gram-positive bacterium Enterococcus faecalis, Toll pathway activation was not affected in grassHerrade or Psh mutant flies but was impaired in flies mutant for both proteases. Thus, psh and grassHerrade double-mutant flies were more vulnerable to infection by E. faecalis than were grassHerrade mutants (6). These experiments suggested that Psh also detects Gram-positive bacteria. To clarify whether Psh detects PAMPs or virulence factors from Gram-positive bacteria, the authors injected heat-killed E. faecalis into grassHerrade mutant flies and found that Toll activation was nearly totally impaired. There was no additional decrease in Toll activation in psh and grassHerrade double-mutant flies. Heat-killed E. faecalis showed lower activation of the Toll pathway after infection with live E. faecalis; this result suggests that, when alive, these bacteria express some factors that, along with PAMPs, activate the Toll pathway. Furthermore, when purified proteases from Gram-positive bacteria such as Bacillus subtilis were injected into wild-type and grassHerrade flies, a strong activation of Toll was apparent, whereas psh mutants showed lower activation. Therefore, it was concluded that Gram-positive bacteria are also detected by the Psh pathway (6).

With the discovery of new roles for Grass as a component of the fungal cell wall–detecting cascade and Psh as a sensor for proteolytic activities from Gram-positive bacteria, a new model has been proposed. The earlier model suggested that three proteolytic cascades converge to activate SPE, which in turn activates the Toll receptor (see above). The redefined model for the Drosophila Toll pathway suggests that only two proteolytic pathways activate Toll. The first pathway, which could be called the PRR-dependent pathway, recognizes glucans from fungi and Lys-type PGNs from Gram-positive bacteria. These PAMPs are detected by their respective PRRs, which activate a proteolytic cascade involving Grass, leading to the activation of SPE. The second pathway, called the “danger signal” pathway, is activated by proteases secreted by fungi and Gram-positive bacteria, which are sensed by Psh as abnormal proteolytic activity in the hemolymph (Fig. 1B) (6).

The involvement of the Drosophila Toll receptor in innate immunity was uncovered more than a decade ago (8). Since then, numerous components both upstream and downstream of Toll have been identified [reviewed in (1)]. However, our current knowledge of how PRRs transduce signals to proteases and the mechanism of action of AMPs on microbes remains hazy. Further research on organisms from other phyla—such as mollusks, some of which are hosts for Schistosoma (blood flukes)—may provide us with novel insights into the evolution of TLR signaling and innate immunity. The factor(s) that activate Psh and in turn are activated by it still await discovery. Understanding immune signaling pathways may provide us with novel insights about insect-pathogen interactions. Diseases such as malaria and plague are caused by arthropod vectors, and such research may even reveal a way to prevent infection in vectors, thereby preventing transmission of pathogens to humans.


View Abstract

Stay Connected to Science Signaling

Navigate This Article