pelle

The Drosophila Toll-9, acting through Pelle and Cactus activates a constitutive antimicrobial defense

The Toll family of transmembrane proteins participates in signaling infection during the innate immune response. The nine Drosophila Toll proteins were analyzed and it was found that wild-type Toll-9 behaves similar to gain-of-function Toll-1. Toll-9 activates strongly the expression of Drosomycin, and utilizes similar signaling components to Toll-1 in activating the antifungal gene. The predicted protein sequence of Toll-9 contains a tyrosine residue in place of a conserved cysteine, and this residue switch is critical for the high activity of Toll-9. The Toll-9 gene is expressed in adult and larval stages prior to microbial challenge, and the expression correlates with the high constitutive level of drosomycin mRNA in the animals. The results suggest that Toll-9 is a constitutively active protein, and implies its novel function in protecting the host by maintaining a substantial level of antimicrobial gene products to ward off the continuous challenge of microorganisms (Ooi, 2002).

In both dorsal–ventral development and antifungal response, activated Toll-1 recruits Tube and Pelle to initiate signaling. Both Tube and Pelle contain death domains, and Pelle is a kinase. Recruitment of Pelle somehow leads to degradation of the inhibitor Cactus and release of the transcription factors, Dorsal and Dif. Whether Toll-9 employs the same signaling components to activate drosomycin expression was examined. A construct for Pelle containing only the death domain (Pelle^DD), but lacking the kinase domain, was generated. This mutated Pelle protein should function as dominant negative by binding to the death domain of Tube but cannot phosphorylate downstream substrates. Transfection of wild-type Pelle activated the reporter gene efficiently, consistent with an important role of the protein in antifungal response. As expected, Pelle^DD did not activate the reporter. In contrast, the Pelle^DD construct inhibited all the Toll-1-, Toll^10b- and Toll-9-mediated drosomycin reporter activities (Ooi, 2002).

Cactus uses its ankyrin repeats to bind to the Rel homology domains of Dif and Dorsal. The Cactus protein degradation is regulated both by signal dependent and signal independent mechanisms, through the N-terminal serine residues and C-terminal PEST sequence, respectively. Therefore, a construct CactusDelta¹²⁵Delta^PEST was used that contained only the ankyrin repeats. This mutant Cactus should stably bind to and inhibit Dif and Dorsal, even when the signaling pathway is stimulated. Co-transfection of wild-type Cactus did not lead to significant changes in the activation of drosomycin by Toll-1, Toll^10b and Toll-9. In contrast, the CactusDelta¹²⁵Delta^PEST construct abolished all these Toll signaling activities. Therefore, Cactus and Pelle, and probably the binding partners Dif and Tube, are likely signaling components that mediate the activation of drosomycin by Toll-9 (Ooi, 2002).

Protein Interactions

The interaction of the serine/threonine kinase Pelle and adaptor protein Tube through their N-terminal death domains leads to the nuclear translocation of the transcription factor Dorsal and activation of zygotic patterning genes during Drosophila embryogenesis. Crystal structure of the Pelle and Tube death domain heterodimer reveals that the two death domains adopt a six-helix bundle fold and are arranged in an open-ended linear array with plastic interfaces mediating their interactions. The Tube death domain has an insertion between helices 2 and 3, and a C-terminal tail making significant and indispensable contacts in the heterodimer. In vivo assays of Pelle and Tube mutants have confirmed that the integrity of the major heterodimer interface is critical to the activity of these molecules (Xiao, 1999).

Within the Drosophila embryo, Tube and the protein kinase Pelle transduce an intracellular signal generated by the transmembrane receptor Toll. This signal directs import of the rel-related protein Dorsal into ventral and ventrolateral nuclei, thereby establishing dorsoventral polarity. Tube protein associates with the plasma membrane during interphase. Tube sequences are required for signaling interaction with pelle in a yeast two-hybrid assay. Fusion of the Pelle catalytic domain to the transmembrane receptor Torso is sufficient to induce ventral fates. This activity is independent of Toll or Tube. Fusion of the Tube protein to Torso also induces ventral fates, but only in the presence of functional Pelle. A model has been proposed wherein Tube activates Pelle by recruiting it to the plasma membrane, thereby propagating the axis-determining signal (Galindo, 1995).

A complex signal transduction pathway functions in the early Drosophila embryo to establish dorsal-ventral polarity. Activation of this pathway results in the nuclear transport of the protein dorsal (dl), a member of the rel/NF-kappaB family of transcription factors. Genetic studies have identified three intracellular components whose activities are required for activation of dl: Toll, a transmembrane receptor; pelle (pll), a serine/threonine protein kinase, and tube, a protein of unknown function. The activities of these proteins were examined when coexpressed in Drosophila Schneider cells. Coexpression of pll with dl enhances dl nuclear localization and results in a modest increase in transcriptional activity. However, when pll is coexpressed with a specific mutant derivative of Toll (TlNaeI), although not with wild-type Toll, a striking synergistic activation of dl is detected. Unexpectedly, coexpression of pll plus TlNaeI, in the absence of dl, results in a similar synergistic activation of a GAL4-tube fusion protein. Based on these and other results, a model is proposed in which pll receives a signal from activated Toll and phosphorylates tube, which then participates directly in dl activation. The C-terminal intracytoplasmic region of Toll contains a 200 residue IL-1R (type I interleukin-1 receptor) homologous domain that is essential for Toll activity, plus an extra 68 residues of unique sequence at the very C terminus. It has been shown that eletion of this unique region increases transcriptional activity of Dorsal in cotransfected Schneider cells, suggesting it plays an inhibitory function (Norris, 1996).

Tube and Pelle are required to relay the signal from Toll to the Dorsal-Cactus complex. In a yeast two-hybrid assay, Tube and Pelle interact with Dorsal. These interactions have been confirmed in an in vitro binding assay. Tube interacts with Dorsal via its C-terminal domain, whereas full-length Pelle is required for Dorsal binding. Tube and Pelle bind Dorsal in the N-terminal domain 1 of the Dorsal Rel homology region rather than at the Cactus binding site. Domain 1 has been found to be necessary for Dorsal nuclear targeting. Genetic experiments indicate that Tube-Dorsal interaction is necessary for normal signal transduction. A model is presented in which Tube, Pelle, Cactus, and Dorsal form a multimeric complex that represents an essential aspect of signal transduction (Yang, 1997).

Dorsal is an embryonic phosphoprotein, its phosphorylation state regulated by an intracellular signaling pathway initiated by the transmembrane receptor Toll. The phosphorylation state of Dorsal is altered during the time period that Toll is activated. Moreover, mutations that constitutively activate Toll stimulate Dorsal phosphorylation, while mutations that block Toll activation reduce the level of Dorsal phosphorylation. Signal-dependent Dorsal phosphorylation is modulated by three intracellular proteins: Pelle, Tube, and Cactus. Free Dorsal is a substrate for a signal-independent kinase activity. These results imply that Dorsal is a substrate for a Toll-dependent kinase. (Gillespie, 1994) However, subsequent studies (Reach, 1996) indicate that Cactus and not Dorsal is the target of Pelle phosphorylation (See above: Biological overview).

A signaling pathway active on the ventral side of the Drosophila embryo defines dorsoventral polarity. A cell surface signal relayed by Toll, Tube and Pelle releases the Rel-related protein Dorsal from its cytoplasmic inhibitor Cactus; free Dorsal translocates into nuclei and directs expression of ventral fates. Using the yeast two-hybrid system and immunoprecipitation experiments, scaffolding and anchoring interactions were defined among the pathway components. Dorsal binds specifically to Tube, Pelle and Cactus, and the protein kinase activity of Pelle differentially regulates its interactions with Dorsal and Tube. Amino acids 47-345 oof Dorsal are sufficient for interaction with both Tube and Pelle. This same region, the Rel homology domain, is also required for dimerization, for DNA binding and for interaction with Cactus. The Dorsal Rel domain is both necessary and sufficient for generation of a dorsoventral nuclear concentration gradient. Pelle and Dorsal interact with two separable domains of Tube. Pelle binds to the amino-terminal region of Tube that spans residues 25 to 173. Dorsal binds a C-terminal domain of Tube (amino acids 257 to 462). This region contains five copies of an evolutionarily conserved, 8-amino-acid repeat and is required for full Tube function. Interaction with Dorsal requires full-length Pelle. In contrast, only about 100 amino acids of Pelle (residues 26 to 129) are necessary and, most likely, sufficient for interaction with Tube. Pelle catalytic activity modulates its interaction with Dorsal and Tube. Drosophila Filamin (an Actin binding protein that localizes to the inner surface of the cell membrane) is identified as a potential adaptor linking the interaction network, via Tube, to the transmembrane receptor Toll. The Toll/IL-1 receptor homology appears to be both necessary and sufficient for the interaction of Toll with Filamin. The studies reported here have defined minimal interactions for Pelle (residues 26-129) and Tube (residues 25-173) that correspond closely to regions with similarity to a consensus death domain (see Reaper). Death domains have been identified in pathways regulating apoptosis, but their participation in the dorsoventral signaling cascade suggests a more general role in protein interactions mediating signal transduction (Edwards, 1997).

The Toll pathway recruits Tube and Pelle to the plasma membrane, a function that is required for the transmission of information from activated Toll receptors to the Dorsal-Cactus complex. An mRNA microinjection assay has demonstrated that targeting of either Tube or Pelle to the plasma membrane by myristylation is sufficient to activate the signal transduction pathway that leads to Dorsal nuclear translocation. Using confocal immunofluorescence microscopy it has also been shown that activated Toll induces a localized recruitment of Tube and Pelle to the plasma membrane. Together, these results strongly support the hypothesis that intracellular signaling requires the Toll-mediated formation of a membrane-associated complex containing both Tube and Pelle (Towb, 1998).

Tube and Pelle have been shown to signal constitutively when fused to the extracellular and transmembrane domains of the Drosophila Torso receptor. It has been proposed that constitutive activation of these chimeras is the direct result of membrane localization. However, the Torso sequences could also have activated the chimeras by (for example) mediating an ectopic extracellular protein-protein interaction. To determine whether membrane association is in fact sufficient to constitutively activate Tube and Pelle, Src90, the amino-terminal 90 amino acids of the Drosophila Src tyrosine kinase, was fused to full-length Tube and to the catalytic domain of Pelle. The Src 90 domain contains a myristylation signal that directs association with the plasma membrane. The src90-pelle and src90-tube fusion constructs were transcribed in vitro and then assayed their activity by microinjecting the resulting RNAs into syncytial blastoderm embryos. The ability of microinjected RNA transcripts to stimulate dorsoventral signaling was analyzed by examining cell fate markers in the cuticle secreted by the developing embryo. In wild-type embryos, Toll-mediated signaling to Dorsal leads to the formation of filzkšrper and ventral denticle belts (cuticle structures representative of dorsolateral and ventral ectodermal fates, respectively). Mutations exist in which intracellular signaling is abolished due to a block in the extracellular pathway required for Toll ligand activation. For example, in embryos generated by females lacking function at the gastrulation defective locus (gd2 embryos), Toll is inactive and Dorsal remains exclusively cytoplasmic. Such embryos fail to form filzkšrper and ventral denticles. Injection of either SRC90-TUBE or SRC90-PELLE mRNA into gd2 embryos directs signaling at the posterior site of microinjection, as evidenced by the appearance of filzkšrper. SRC90-PELLE mRNA also induces the ectopic expression of ventral denticle belts. These results demonstrate the ability of both Src90-Tube and Src90-Pelle to direct Dorsal nuclear translocation in the absence of Toll activation. It is concluded that Pelle and Tube are activated upon targeting to the plasma membrane (Towb, 1998).

In the embryo, the majority of the Tube protein molecules associates with the cell surface and this association occurs over the entire circumference of the embryo. Since this localization is not restricted to ventral regions, it cannot represent an activated state of Tube. However, based on the Src90 fusion experiments, it is envisioned that Toll activates Tube by driving association of Tube with the plasma membrane. To detect any such signal-related membrane association, a detailed examination of Tube localization was carried out, using the nuclear concentration gradient of Dorsal as a marker for the dorsoventral axis. Surface views of wild-type embryos reveal that Tube forms a mesh-like array. Such a localization pattern is characteristic of proteins associated with the membranes and cytoskeleton that cap and surround each blastoderm nucleus. For Tube, the mesh-like pattern is most intense along the ventral midline, suggesting a concentration of Tube in this region. Confocal cross-sectional views confirm the existence of a dorsoventral asymmetry in the distribution of Tube at the embryo surface. Tube is more highly concentrated along the ventral surface than the dorsal surface; the difference in concentration is two-fold, as assayed by quantitation of staining intensity. Higher magnification ventral views reveal that Tube clusters along the embryo surface. These clusters or aggregates are approximately 2 mm in diameter and are located between adjacent ventral nuclei at sites of membrane invagination. On the dorsal side of the embryo, clusters of Tube are less readily detectable (Towb, 1998).

To determine whether the asymmetric distribution of Tube depends on Toll-mediated signaling, Tube localization was examined in genetic backgrounds that either inactivate or constitutively activate the signaling pathway. The gd2 mutation blocks Toll activation, whereas the Toll10b mutation alters the Toll extracellular domain so as to constitutively activate the receptor throughout the embryo. Both mutations eliminate the dorsoventral asymmetry in Tube staining observed in the wild type. However, Tube is more highly concentrated at the periphery of Toll10b embryos than gd2 embryos. The Toll10b and gd2 results support the model that activated Toll directs recruitment of Tube to the embryo surface. The asymmetry in Tube localization in the wild type is not, however, as striking as that observed for Cactus or Dorsal. Therefore, the distribution of Tube was also examined in P[Toll10b-bcd] embryos, in which Toll-mediated signaling is ectopically oriented along the anteroposterior axis. Females carrying the P[Toll10b-bcd] transgene express high levels of the Toll10b cDNA fused to the bicoid (bcd) 3'UTR and consequently produce embryos in which the bcd sequences localize the Toll10b mRNA to the anterior pole. Tube concentrates at the plasma membrane in and near the anterior pole of P[Toll10b-bcd] embryos. Moreover, clusters of Tube are prominent anteriorly on both the dorsal and ventral surfaces of such embryos. The Toll10b protein expressed from the P[Toll10b-bcd] construct also forms cell surface aggregates at the anterior pole. Comparison of the patterns of Tube and Toll localization in P[Toll10b-bcd] embryos indicates that the Toll aggregates colocalize with the Tube aggregates (Towb, 1998).

Since Tube and Pelle are thought to interact in embryos, it was reasoned that Pelle, like Tube, might localize to sites of Toll activation. Using a polyclonal anti-Pelle serum to stain wild-type embryos, it was found that Pelle is distributed throughout the embryo. A fraction of Pelle localizes to the surface of embryos, but there is no significant asymmetry in this distribution of Pelle across the dorsoventral axis. If Pelle undergoes a signal-dependent membrane association in wild-type embryos, this association is too short-lived or involves too few of the Pelle molecules in the embryo to be readily detectable. However, since localization studies had indicated that signal-dependent protein relocalization is enhanced in embryos expressing high levels of Toll10b protein at the anterior pole, Pelle localization was also examined in embryos generated by P[Toll10b-bcd] females. A large fraction of the Pelle protein localizes to the anterior end of P[Toll10b-bcd] embryos, indicating a significant recruitment or stabilization of Pelle protein at this pole. Confocal cross sections have revealed that the anterior-localized Pelle is predominantly at the embryonic periphery, possibly in association with the plasma membrane. High levels of activated Toll thus promote localization of both Tube and Pelle to the surface of the embryo. Since Tube localization at the anterior end of P[Toll10b-bcd] embryos mirrors that in the ventral portion of wild-type embryos, it is proposed that the localization of Pelle in P[Toll10b-bcd] embryos similarly reflects a relocalization occurring in wild-type embryos in response to Toll activation (Towb, 1998).

In the Drosophila embryo the nuclear localization of Dorsal, a member of the Rel family, is regulated by an extracellular signal, which is transmitted to the interior of the egg cell by a cascade of proteins involving the novel protein Tube and the protein kinase Pelle. The activation mechanism of Tube and Pelle and the interaction between these two components have been analyzed. Both proteins, although having different biochemical activities, are activated by the same mechanism. Membrane association alone (driven by creating Torso-Pelle or Torso-Tube fusion proteins) is not sufficient, but oligomerization (driven by creating constitutively active Torso in these fusion proteins) is required for full activation of Tube and Pelle. By deletion analysis the domains of Tube and Pelle mediating the physical interaction and the signaling to downstream components has been determined. In order to investigate the link between Pelle and the target of the signaling cascade (the Dorsal/Cactus complex) a novel but evolutionary conserved protein termed Pellino was isolated and characterised. Pellino associates with the kinase domain of Pelle (Grosshans, 1999).

Pellino is a protein with 424 amino acids. It is highly conserved to a protein of unknown function in C. elegans, with about 50% identity and 70% homology. Mouse and human homologs have been identified. Pellino is expressed during all stages of develoment, coded for by a transcript of 3kb. In addtion, there is a oocytic transcript of 1.9kb found only in females and in eggs. Pellino does not bind to a catalytically inactive form of the Pelle kinase and Pellino is not phosphorylated by Pelle. These observations indicate that Pellino, unlike Tube, is not a phosphorylation substrate of Pelle, but that it probably only binds to the autophosphorylated form of Pelle. Tube and Pellino are shown ot bind to different domains of Pelle. Pellino is probably involved in the signaling steps downstream of Pelle. Pellino could stabilize the activated state of Pelle after the dissociation from Tube oligomers or could mediate the interaction to the downstream phosphorylation substrate of Pelle. Alternatively, the signal transduction by activated Pelle oligomers may occur through the binding of Pellino and other downstream components without any phosphorylation of them by Pelle. Future studies of Pellion and the identification of mutations in the pellino gene will allow a better understanding of the molecular nature of the link between Pelle and the Dorsal/Cactus complex (Grosshans, 1999).

A member of the tumor necrosis factor (TNF) receptor-associated factor (TRAF) family has been identified in Drosophila. DTRAF1 contains 7 zinc finger domains followed by a TRAF domain, similar to mammalian TRAFs and other members of the family identified in data bases from Caenorhabditis elegans, Arabidopsis, and Dictyostelium. Analysis of DTRAF1 binding to different members of the human TNF receptor family has shown that this protein can interact through its TRAF domain with the p75 neurotrophin receptor and weakly with the lymphotoxin-ß receptor. DTRAF1 can also self-associate and binds to human TRAF1, TRAF2, and TRAF4. Interestingly, DTRAF1 interacts with human cIAP-1 and cIAP-2 but not with Drosophila DIAP-1 and -2. By itself, DTRAF1 does not induce significant NFkappaB activation when overexpressed in mammalian cells, although it specifically increases NFkappaB induction by TRAF6. In contrast, TRAF2-mediated NFkappaB induction is partially inhibited by DTRAF1. Mutants of DTRAF1 lacking the N-terminal region inhibit NFkappaB induction by either TRAF2 or TRAF6. DTRAF1 specifically associates with the regulatory N-terminal domain of Pelle, a Drosophila homolog of the human kinase interleukin-1 receptor-associated kinase (IRAK). Interestingly, though Pelle and DTRAF1 individually are unable to induce NFkappaB in a human cell line, co-expression of Pelle and DTRAF1 result in significant NFkappaB activity. Interactions of DTRAF1 with human TRAF-, TNF receptor-, and IAP-family proteins imply strong evolutionary conservation of TRAF protein structure and function throughout Metazoan evolution (Zapata, 2000).

Altogether these results demonstrate that DTRAF1 is able to regulate NFkappaB activation in collaboration with Pelle, suggesting that DTRAF1 is a component of the Toll pathway in Drosophila. This pathway has been implicated in the regulation of the dorsal-ventral polarization of developing embryos. However, microinjection of DTRAF1 (1-124) mRNA into Drosophila embryos fails to affect the normal dorsal-ventral patterning, suggesting that DTRAF1 may not be required for this developmental process. This observation, together with the recent characterization of another member of the TRAF family in Drosophila implies that flies may contain other TRAF-family genes that create redundancy in the pathways available for NFkappaB induction, similar to the situation with mammalian TRAFs. A role for DTRAF1 in innate immune responses to pathogens in flies, however, remains to be explored (Zapata, 2000).

Signaling through the Toll receptor is required for dorsal/ventral polarity in Drosophila embryos, and also plays an evolutionarily conserved role in the immune response. Upon ligand binding, Toll appears to multimerize and activate the associated kinase, Pelle. However, the immediate downstream targets of Pelle have not been identified. Drosophila Tumor necrosis factor receptor-associated factor 2 (dTRAF2), a homologue of human TRAF6, physically and functionally interacts with Pelle, and is phosphorylated by Pelle in vitro. The N-terminal RING/zinc-finger domains of dTRAF2, but not the TRAF domain, interact strongly with both phosphorylated and unphosphorylated Pelle, and activate Dorsal in SL2 cells. Importantly, dTRAF2 and Pelle cooperate to activate Dorsal synergistically in cotransfected Schneider cells. Deletion of the C-terminal TRAF domain of dTRAF2 enhances Dorsal activation, perhaps reflecting the much stronger interaction of the mutant protein with phosphorylated, active Pelle. Taken together, these results indicate that Pelle and dTRAF2 physically and functionally interact, and that the TRAF domain acts as a regulator of this interaction. dTRAF2 thus appears to be a downstream target of Pelle (Shen, 2001).

Unlike the situation observed with mammalian TRAF6, the TRAF domain of dTRAF2 does not show a dominant-negative effect, but instead activates Dorsal to a level comparable to that observed with wild-type dTRAF2. It is possible that an interaction between the exogenously expressed TRAF domain and endogenous dTRAF2 may be responsible for this seemingly anomalous activation of signaling by the isolated TRAF domain. Mammalian TRAF domains have been shown to be involved in both homotypic and heterotypic aggregation. In TNF-induced NF-kappaB activation, the interaction between the TRAF6 N-terminal domain and zetaPKC (an atypical member of the protein kinase C family) is important for signaling. zetaPKC is unable to interact with the N-terminal domain of TRAF6 in its monomeric form, but dimerization of this domain dramatically stimulates interaction. Likewise, the interaction between the dTRAF2 N-terminal region and Pelle could be regulated by TRAF domain dimerization (Shen, 2001).

dTRAF2 is phosphorylated by activated Pelle in vitro. It is not clear how dTRAF2 functions in signaling to Dorsal, and whether phosphorylation is critical for its activity. Mammalian TRAF6 has been shown to interact with several downstream factors that appear to be required for NF-kappaB activation. Because it is unknown whether TRAF6 is phosphorylated by IRAK (or indeed phosphorylated at all), it is unclear how phosphorylation might affect these, or other, interactions. Likewise, additional studies are necessary to determine whether phosphorylated IRAK interacts with TRAF6, since Pelle is capable of binding dTRAF2. If so, TRAF6 could link active IRAK to other downstream targets. In any event, this study has provided evidence that a TRAF can be phosphorylated by an upstream kinase, which extends the possible mechanisms by which these molecules can modulate signaling pathways (Shen, 2001).

dTRAF2 is covalently modified when overexpressed in SL2 cells, and this modification requires the RING-finger domain. Such domains have been shown to be involved in protein ubiquitination, and indeed it may be that all RING-finger proteins may act as E3 ubiquitin ligases. The ladder-like modification of dTRAF2 that was detected in transfected cells is similar to the pattern expected for ubiquitination, and may target dTRAF2 and associated proteins (e.g., Pelle) for degradation. IRAK has been shown to be degraded by the proteasome after phosphorylation. Phosphorylated Pelle rapidly disappears from embryos in the 5th hour after egg laying, after completion of Toll signaling and Dorsal activation. It is conceivable that activated, phosphorylated Pelle may also be degraded by the ubiquitin-proteasome pathway by binding to dTRAF2, which would thus desensitize Pelle signaling after activation of downstream components. However, it is notable that recent studies have shown that TRAF6 is required for synthesis of an unusual polyubiquitin chain required for IKK activation, by a mechanism not involving degradation. Whether this observation is related to the apparent dTRAF2 self-ubiquitination detected here is not clear (Shen, 2001).

dTRAF1 has also been shown to interact with Pelle in vitro, and to enhance NF-kappaB-mediated expression in mammalian cells cotransfected with Pelle. Unlike dTRAF2, which interacts with the Pelle catalytic domain and can synergize strongly with Pelle to activate Dorsal, dTRAF1 interacts with Pelle in its N-terminal death domain and gives rise only to weak activation of NF-kappaB. These results suggest that the two dTRAFs play distinct roles. Supporting this possibility, dTRAF1, but not dTRAF2, was shown to activate the JNK pathway by binding to the Ste20 kinase, Msn. In addition, microinjection of mRNA encoding dTRAF1Delta, which encodes a dominant-negative protein in transfections assays in mammalian cultured cells, into Drosophila embryos failed to affect normal dorsal-ventral patterning. These results imply that the different downstream signaling events that activate Dorsal and JNK may bifurcate at the level of the Pelle-dTRAF interaction (Shen, 2001 and references therein).

The Drosophila Pelle kinase plays a key role in the evolutionarily conserved Toll signaling pathway, but the mechanism responsible for its activation has until now been unknown. In vivo and in vitro evidence is presented establishing an important role for concentration-dependent autophosphorylation in the signaling process. Pelle phosphorylation can be detected transiently in early embryos, concomitant with activation of signaling. Importantly, Pelle phosphorylation is enhanced in a gain-of-function Toll mutant (Toll^10b), but decreased by loss-of-function Toll alleles. Pelle is phosphorylated in transfected Schneider L2 cells in a concentration-dependent manner such that significant modification is observed only at high Pelle concentrations, which coincide with levels required for phosphorylation and activation of the downstream target, Dorsal. Pelle phosphorylation is also enhanced in L2 cells co-expressing Toll^10b, and is dependent on Pelle kinase activity. In vitro kinase assays reveal that recombinant, autophosphorylated Pelle is far more active than unphosphorylated Pelle. Importantly, unphosphorylated Pelle becomes autophosphorylated, and activated, by incubation at high concentrations (Shen, 2002).

What is the function of activated Pelle? Autophosphorylation allows Pelle to dissociate from both the Toll receptor and Tube. In one model, Pelle function might thus be to phosphorylate Toll and Tube to facilitate Pelle dissociation following activation. Phosphorylated Toll might then recruit other factors, analogous to cytokine receptors, that recruit STAT after being phosphorylated by JAK kinase, although there is currently no evidence for this. In another model, it seems highly likely that autophosphorylated, activated Pelle phosphorylates specific downstream targets. For example, Pellino and Dorsal were both shown to bind phosphorylated wild-type Pelle but not PelleK240R, suggesting that one or both proteins might be a Pelle target. In addition, Pelle can phosphorylate Cactus in vitro and mouse Pelle (IRAK) is able to phosphorylate IkappaBalpha, suggesting the possibility of direct physical and functional interactions between Pelle and the Cactus/Dorsal complex. Moreover, the TNF receptor-associated factor TRAF6 functions in NF-kappa B activation by mammalian TLRs, and interacts with IRAK upon IL-1 induction. Drosophila TRAF2, an apparent TRAF6 homolog, is a target of Pelle phosphorylation in vitro, and contributes strongly to Dorsal activation in transfected Schneider cells (Shen, 2002).

In summary, Pelle is transiently phosphorylated upon Toll activation in Drosophila embryos. Pelle autophosphorylation is concentration dependent and can be induced by activated Toll in co-transfected Schneider cells. More importantly, Pelle autophosphorylation is necessary for kinase activity, and dephosphorylated Pelle can be activated by autophosphorylation at high concentrations. Based on these results, a model is proposed in which Toll-bound, unphosphorylated Pelle is activated by transphosphorylation at high local concentrations, which are created by Toll multimerization induced by ligand binding. Available data suggest that this model is probably relevant to the function of TLR proteins generally. A next important question is how activated Pelle transmits its signal downstream, and ultimately to the Dorsal/Cactus complex. Analysis of Pelle-interacting factors and target proteins will help to fill the gap between the Toll receptor complex and the Dorsal/Cactus complex. But these results have provided a molecular mechanism for the first intracellular steps in the Toll pathway (Shen, 2002).

Dorsoventral polarity in the Drosophila embryo is established through a signal transduction cascade triggered in ventral and ventrolateral regions. Activation of a transmembrane receptor, Toll, leads to localized recruitment of the adaptor protein Tube and protein kinase Pelle. Signaling through these components directs degradation of the IkappaB-like inhibitor Cactus and nuclear translocation of the Rel protein Dorsal. Pelle mediates feedback regulation in the Drosophila Toll signaling pathway. Pelle functions to downregulate the signal-dependent relocalization of Tube. Inactivation of the Pelle kinase domain, or elimination of the Tube-Pelle interaction, dramatically increases Tube recruitment to the ventral plasma membrane in regions of active signaling. A large collection of pelle alleles has been characterized in this study, leading to an identification of the molecular lesions in these alleles and their effects on Pelle autophosphorylation, Tube phosphorylation and Tube relocalization. These results point to a mechanism operating to modulate the domain or duration of signaling downstream from Tube and Pelle (Towb, 2001).

An activity of Pelle, possibly direct phosphorylation of Tube, is required to disaggregate membrane-associated Tube clusters. In pelle mutant embryos, Tube clustering along the ventral midline is greatly enhanced over that seen in wild-type embryos. Elimination of either of two wild-type activities of Pelle -- Tube binding or protein kinase function -- is sufficient for this enhancement (Towb, 2001).

Pelle can phosphorylate Tube in vitro. Furthermore, the pll⁰⁷⁴ mutation, which blocks Tube phosphorylation but not Pelle autophosphoryation, enhances Tube clustering. It is suggested, therefore, that in the wild-type embryo Pelle-mediated phosphorylation of Tube causes dissociation of multiprotein complexes containing Tube. The weak gradient of Tube seen in the wild type would thus represent a balance between Tube recruitment to clusters and Tube release from clusters upon Pelle-catalyzed phosphorylation (Towb, 2001).

Although the idea that Pelle directly phosphorylates Tube to disrupt Tube-containing complexes is favored, alternative mechanisms for regulating complex stability are possible. For example, Pelle might phosphorylate and activate an unidentified downstream target that would then disaggregate the Tube-containing complex. Distinguishing between these possibilities is made difficult by the fact that all known mutations that prevent the association of Tube and Pelle also block Pelle activation and hence downstream function. For instance, the tub² mutation, which prevents interaction with Pelle, completely blocks signaling to Dorsal. Similarly, mutations in Pelle that do not affect the catalytic domain, but that alter the Tube binding domain, disrupt Pelle function (Towb, 2001).

Nearly all the Pelle missense mutations sequenced in this study map to the kinase domain rather than the death domain, the site of interaction with Tube. A possible explanation for this bias became apparent in the structural analysis of the interaction between Tube and Pelle. A combination of crystallographic studies and mutational analyses has revealed that the Tube death domain can productively interact with two different sets of surface residues in the Pelle death domain. One would therefore predict that most Tube binding sites in the Pelle death domain would be redundant in function and that such sites would not mutate to a loss-of-function phenotype. Thus, mutations that block signaling would most frequently map to the death domain of Tube, but outside the death domain in Pelle, as is in fact the case (Towb, 2001).

Extensive genetic characterization of pelle mutations led to the classification of several alleles as recessive antimorphs. The term antimorphic indicates that the presence of this allele is more detrimental to development than a complete absence of the functional gene. Most of the pelle alleles classified as strong antimorphs alter residues in or near a part of the kinase structure known as the activation loop. Such activation loops often contain residues that are targets for upstream, activating kinases. In the case of the protein kinase ERK2, for example, phosphorylation of residues in the activation loop alters the conformation of the active site cleft, affecting ERK2's ability to bind ATP and changing the conformation of the substrate-recognition region. By analogy, the antimorphic effect of several pelle mutations may be due to their binding and sequestering either an upstream activating kinase for Pelle or a Pelle substrate (Towb, 2001).

Regulation of protein-protein interaction through phosphorylation has been widely documented in signal transduction pathways. In visual signal transduction, for example, modulation of the phosphorylation state of phosducin regulates the interaction of phosducin and the ßgamma subunits of transducin. In the case of Tube and Pelle, such a change in phosphorylation state could be catalyzed by Pelle itself. This could explain why a catalytically inactive form of Pelle interacts more strongly with Tube in the yeast two-hybrid system than does the wild-type Pelle (Towb, 2001).

Phosphorylation of Tube by Pelle could act negatively on the pathway, breaking apart Tube-containing complexes, as well as weakening the Tube/Pelle interaction. Since Pelle acts downstream of Tube and requires Tube-mediated targeting to the membrane for activation, both of these feedback mechanisms could act to turn Pelle activity down and decrease downstream signaling. This regulation may serve to increase the fidelity of downstream signal transduction. High, medium, and low nuclear levels of the Dorsal transcription factor control the transcription of different sets of genes in highly circumscribed ventral-to-dorsal territories in the early embryo. Feedback mechanisms in the signal transduction pathway downstream of Toll may have evolved to ensure that proper nuclear Dorsal levels are maintained in the correct spatial regions of the embryo (Towb, 2001).

In considering negative regulation of Toll signaling, it is necessary to take into account that signal transduction is taking place during the highly rapid nuclear division cycles of early Drosophila embryogenesis. The Dorsal nuclear localization gradient does not persist during mitosis, but rather reforms after every nuclear division. Furthermore, Tube becomes associated with nuclei during each mitosis. Negative feedback may therefore be required to return the signaling cascade downstream of Toll to a starting state, setting the stage for reformation of the Dorsal gradient in the next interphase. In this regard, it is noted that the activation of both Spätzle and Easter, which act upstream of Toll, appears to be subject to a distinct form of negative feedback regulation (Towb, 2001).

Although the phenomenon of Tube gradient enhancement could, as outlined above, reflect a negative feedback mechanism, it is also possible that Tube release from clusters and Tube/Pelle complex dissociation are mechanisms that act to promote or spatially regulate downstream signaling. Activated Tube molecules might need to diffuse away from a signaling complex, or other components might need to cycle through steps of activation and deactivation, in order for downstream signaling to occur. Alternatively, if the Toll receptor is only activated in a narrow ventral stripe, the diffusion of activated Tube and Pelle could serve to establish a gradient of signaling away from the ventral midline, and thus fine tune the gradient of Dorsal nuclear translocation (Towb, 2001).

The mechanism proposed for downregulation of the dorsoventral pathway signal may function in mammals. The adaptor protein, Myd88, has been reported to bind more tightly to IRAK when the IRAK kinase domain has been inactivated, similar to the situation with Tube and Pelle. Therefore, upon activation, IRAK may phosphorylate MyD88 and cause the signaling complex to dissociate. This mechanism of regulating death domain interactions through phosphorylation, would thus be another aspect of the host defense system inherited from the common ancestor of flies and humans (Towb, 2001).

MyD88 is an adaptor protein that interacts physically with Toll and with the kinase Pelle and functions upstream of Tube and Pelle

Toll-like receptors comprise a family of cell surface receptors that play a crucial role in the innate immune recognition of both Drosophila and mammals. Previous studies have shown that Drosophila Toll-1 mediates the induction of antifungal peptides during fungal infection of adult flies. Through genetic studies, Tube, Pelle, Cactus, and Dif have been identified as downstream components of the Toll-1 signaling pathway. A Drosophila homologue of human MyD88 is an adapter in the Toll signaling pathway that associates with both the Toll receptor and the downstream kinase Pelle. Expression of Drosophila Myd88 in S2 cells strongly induces activity of a Drosomycin reporter gene, whereas a dominant-negative version of Drosophila MyD88 potently inhibits Toll-mediated signaling. Drosophila MyD88 associates with the death domain-containing adapter Drosophila Fas-associated death domain-containing protein (dFADD), which in turn interacts with the apical caspase Dredd. This pathway links a cell surface receptor to an apical caspase in invertebrate cells and therefore suggests that the Toll-mediated pathway of caspase activation may be the evolutionary ancestor of the death receptor-mediated pathway for apoptosis induction in mammals (Horng, 2001).

Drosophila MyD88 is an adapter in the Toll signaling pathway that associates with both the Toll receptor and the downstream kinase Pelle. Expression of MyD88 in S2 cells strongly induces activity of a Drosomycin reporter gene, whereas a dominant-negative version of MyD88 potently inhibits Toll-mediated signaling. MyD88 associates with the death domain-containing adapter Drosophila Fas-associated death domain-containing protein (FADD), which in turn interacts with the apical caspase Dredd. This pathway links a cell surface receptor to an apical caspase in invertebrate cells and therefore suggests that the Toll-mediated pathway of caspase activation may be the evolutionary ancestor of the death receptor-mediated pathway for apoptosis induction in mammals (Horng, 2001).

A BLAST search of the Drosophila genome identified the sequence encoding MyD88, a Drosophila homolog of human MyD88. Similar to its human homolog, Drosophila MyD88 contains an N-terminal death domain, an intermediate domain, and a TIR domain. However, unlike human MyD88, Drosophila MyD88 contains an additional 81 amino acids preceding the death domain and a 162-aa-long C-terminal region following the TIR domain (Horng, 2001).

Transfection of MyD88 into Drosophila S2 cells potently induces a Drosomycin reporter gene but not an Attacin reporter gene. This preferential ability to induce an antifungal gene is similar to that of Toll 10b, a constitutively active form of Toll, and suggests that MyD88 may be a component of the Toll-Tube-Pelle-Cactus-Dif signaling pathway. Previous studies have demonstrated that Toll-mediated Drosomycin induction requires the nuclear translocation of Dif. Dif is normally retained in the cytoplasm by the IkappaB inhibitor Cactus and is released only in response to signal-dependent degradation of Cactus. To test whether MyD88-mediated Drosomycin induction also depends on Cactus degradation, a Cactus mutant was constructed that contains mutations of the conserved serine residues that, in mammalian IkappaB, are the targets of signal-dependent phosphorylation. A Cactus mutant inhibits Drosomycin induction by MyD88 and, as expected, by Toll. This result indicates that, similar to Toll, MyD88 regulates Drosomycin induction through the Cactus-dependent pathway (Horng, 2001).

For further analyses, various deletion mutants of MyD88 were generated. Two of the deletion mutants, one containing the TIR domain and the C-terminal domain (amino acids 237-537) and another containing the intermediate, TIR, and C-terminal domains (amino acids 176-537), activate the Drosomycin reporter weakly (10-fold) in comparison to full length MyD88, indicating that the intact protein is required for optimal activity. However, the fact that these truncation mutants can still induce signaling is surprising, since they lack the death domain that mediates interactions with downstream signaling components. Moreover, similar analyses of human MyD88 have shown that a combination of the death domain and the intermediate domain is sufficient to induce signaling activity comparable to that of the wild-type protein. An equivalent truncation of dMyD88 (amino acids 1-237) retains no residual activity despite being well expressed, suggesting that there are some differences in domain function between human and Drosophila MyD88 proteins (Horng, 2001).

To determine whether MyD88 is a component of the Toll signaling pathway, attempts were made to identify a deletion mutant that would have dominant-negative activity. Therefore, three MyD88 deletion mutants that do not activate the Drosomycin reporter were tested for their ability to inhibit Toll-mediated Drosomycin induction. The strongest inhibitor was the death domain- and middle domain-containing construct (amino acids 1-237), which at low concentrations potently inhibits Toll-mediated Drosomycin induction in a dose-dependent manner (Horng, 2001).

To order MyD88 in the pathway with respect to Pelle, MyD88 was tested for its ability to be inhibited by PelleN, a dominant-negative form of Pelle that consists of the N-terminal death domain-containing region of Pelle. MyD88, like Toll, is strongly inhibited by PelleN. MyD88, however, does not inhibit Pelle, demonstrating that, similar to the mammalian pathway, MyD88 functions upstream of Pelle (Horng, 2001).

To further establish MyD88 as a component of the Toll pathway, whether MyD88 interacts with Toll was tested by coimmunoprecipitation assays. The TIR domain-containing MyD88 construct is detected in anti-Toll immunoprecipitates. Interestingly, when cotransfected with Toll 10b, MyD88 reproducibly appears as two distinct bands -- a slower migrating upper band that may correspond to phosphorylated MyD88 construct and a faster migrating lower band. The predominant form of MyD88 detected in immunoprecipitates is the faster migrating species. MyD88 therefore associates with Toll, presumably through TIR domains, and is a component of the active receptor complex (Horng, 2001).

Because human MyD88 associates with IRAK through death domains, a likely immediate downstream target of MyD88 is the IRAK homolog Pelle. Interaction between the death domain-containing dMyD88 construct (amino acids 1-237) and Pelle was examined. MyD88 is detected in Pelle immunoprecipitates, indicating that MyD88 interacts with Pelle, presumably through their death domains (Horng, 2001).

These results therefore demonstrate that MyD88 is an adaptor in the Toll signaling pathway downstream of the receptor and upstream of Pelle. From genetic analyses, the adaptor protein Tube has also been implicated to be downstream of Toll and upstream of Pelle in the Toll signaling pathway. The death domain of Tube also interacts with Pelle. Because Tube and MyD88 also contain death domains that could potentially mediate their interaction, tests were performed for association between these two proteins in immunoprecipitation assays; Tube and MyD88 do indeed interact. Therefore, MyD88 and Tube both function as adaptors downstream of Toll, exist in the same active complex along with Pelle, and are probably both involved in the recruitment and/or activation of Pelle. Understanding functional differences between these two adapters will require further analysis (Horng, 2001).

To identify other potential downstream targets of MyD88, a search of the Drosophila genome was performed for other sequences that encode death domain-containing proteins that may interact with MyD88. One such sequence encodes a protein with a death domain as well as a death effector domain and appears to be a homolog of mammalian FADD. This cDNA has been identified and named FADD (Hu, 2000). Whether FADD can interact with MyD88 was tested. Lysates from S2 cells transfected with MyD88 were incubated with anti-Flag beads to immunoprecipitate FADD, and immunoprecipitates were blotted with anti-V5 antibody to look for associated MyD88. A strong band corresponding to MyD88 was observed, indicating that MyD88 can interact with FADD through death domains. Overexpression of FADD in S2 cells, however, does not lead to activation of either the Drosomycin or Attacin reporters (Horng, 2001).

Mammalian FADD is recruited to the tumor necrosis factor receptor complex through homophilic death domain interactions with the adapter TNFR-associated death domain-containing protein (TRADD). In turn, FADD recruits procaspase-8 through homophilic death effector domain associations. It is speculated that Drosophila FADD may likewise recruit a Drosophila caspase to the Toll receptor complex. A potential candidate caspase is Dredd, an apical caspase with a long prodomain shown to be essential for induction of antibacterial genes. Indeed, analysis of immunoprecipitated lysates from cells cotransfected with Drosophila FADD, and either full length Dredd or the death effector domain of Dredd showed strong association of Dredd with FADD. A second study (Hu, 2000) has also shown interaction of dFADD with Dredd (Horng, 2001).

Thus Drosophila MyD88 is an adapter in the Toll signaling pathway. MyD88 associates with both Toll and Pelle and functions upstream of Pelle. Tube is known from genetic studies to be an adapter in the Toll pathway that functions upstream of Pelle. Why Toll should signal through MyD88 and Tube, two receptor-proximal adapters with seemingly similar functions, is not yet clear. MyD88 associates with the receptor Toll as well as the downstream adapter FADD, which in turn interacts with the apical caspase Dredd. Because caspases are essential executioners of the apoptotic machinery in organisms from nematodes to mammals, and because Dredd has been shown to be involved in apoptosis during Drosophila development, it is possible that Toll-1 or some of the other eight Tolls that exist in Drosophila may induce apoptosis (or another Dredd-dependent pathway) through the MyD88/dFADD/Dredd pathway in a cell-type specific and/or developmental stage-specific manner. The pathway comprised of Toll, MyD88, dFADD, and Dredd would be the first description of a pathway in invertebrates that links a cell surface receptor to an apical caspase. Such a pathway, if it exists, would enable extracellular stimuli, perhaps ligands secreted by other cells during development or pathogen-derived products during infection, to instruct invertebrate cells to undergo cell death. In addition, the Toll/MyD88/dFADD/Dredd pathway is remarkably similar to that activated by the receptors of the tumor necrosis factor receptor (TNFR) superfamily in mammals, in which FADD-mediated recruitment of caspase-8 leads to induction of apoptosis. Since the Drosophila genome does not encode any cell surface receptors homologous to TNFRs, it appears that the Toll/MyD88/dFADD/Dredd pathway is the evolutionary ancestor of the mammalian death receptor pathways. This possibility is further supported by the recent finding that human TLR2 can induce apoptosis through the MyD88/FADD/Caspase-8 pathway (Horng, 2001).

MyD88 encodes the Drosophila homolog of mammalian MyD88. DmMyD88 combines a Toll-IL-1R homology (TIR) domain and a death domain. Overexpression of DmMyD88 is sufficient to induce expression of the antifungal peptide Drosomycin, and induction of Drosomycin is markedly reduced in DmMyD88-mutant flies. DmMyD88 interacts with Toll through its TIR domain and requires the death domain proteins Tube and Pelle to activate expression of Drs, which encodes Drosomycin. DmMyD88-mutant flies are highly susceptible to infection by fungi and Gram-positive bacteria, but resist Gram-negative bacterial infection much as do wild-type flies. Phenotypic comparison of DmMyD88-mutant flies and MyD88-deficient mice shows essential differences in the control of Gram-negative infection in insects and mammals (Tauszig-Delamasure, 2002).

A heterotrimeric death domain complex involving Tube, Pelle and Myd88

Tube is a scaffolding protein containing an N-terminal interaction motif belonging to the death domain family, as well as C-terminal Tube repeats that mediate binding to Dorsal. Pelle is a serine/threonine-specific protein kinase with a death domain N-terminal to its catalytic domain (Sun, 2002).

Although no Tube homolog has been found in mammals, four Pelle homologs, named IL-1 receptor-associated kinases (IRAKs), have been identified: IRAK1, -2, -M, and -4 . IRAKs function in signaling by a family of Toll-like receptors, as well as the IL-1 receptor (IL-1R), each of which contains a TIR domain, a conserved cytoplasmic signaling motif. An adaptor molecule, Myd88, associates with the C-terminal TIR domain of Toll-like receptors and the IL-1 receptor and with the N-terminal death domain (DD) of IRAKs (Sun, 2002).

During the past few years, genomic sequencing has allowed the identification of Drosophila genes with mammalian homologs functioning in Toll/IL-1 receptor-signaling pathways. These genes include IKK (a homolog of mammalian IKKalpha/ß), Kenny (a homolog of mammalian IKKgamma), IK2 (a homolog of mammalian TBK1/IKKepsilon), Myd88, TAK1, three TRAF loci, and ECSIT. These genes have been studied systematically by using RNA interference (RNAi). RNAi provides a ready means to inactivate a given gene or genes and has facilitated the dissection of Drosophila signaling pathways in cultured S2 cells. To search for essential components of the Toll pathway, an RNAi-based screen was performed among these potential Drosophila NF-kappaB regulators. This approach, coupled with genetic and biochemical analyses, has allowed the dissection of the molecular interactions among death domain-containing proteins in the Drosophila Toll pathway (Sun, 2002).

To investigate the mechanism of Toll signaling, a reporter assay was used in conjunction with RNAi in cultured Drosophila cells. A constitutively active form of Toll, TollDeltaLRR, was stably expressed in S2 cells under the control of a metallothionein promoter, such that the addition of CuSO₄ to the cell culture medium initiates Toll signal transduction. To assay signal transduction downstream of Toll, these S2 cells were transiently transfected with a Drosomycin-luciferase construct. Expression of TollDeltaLRR consistently induces a significant activation (~100 fold) of the Drosomycin reporter (Sun, 2002).

To confirm the efficacy of RNAi in these cells, dsRNA was generated for several genes known to function in the Drosophila Toll pathway. RNAi against Pelle, Tube, or Dorsal significantly inhibits the activation of the Drosomycin reporter, with the effect of Dorsal RNAi relatively stronger than that of Pelle or Tube RNAi. In contrast, RNAi against Cactus dramatically enhances the activation of the Toll pathway. These observations are consistent with the fact that Pelle, Tube, and Dorsal promote Toll signaling, whereas Cactus plays an inhibitory role in the pathway. In this and all subsequent experiments, Easter RNAi serves as a negative control for any nonspecific effect of dsRNA, because Easter acts upstream, and not downstream, of Toll (Sun, 2002).

Next, RNAi-based screening was performed against fly counterparts of mammalian Toll and tumor necrosis factor pathway components, specifically Drosophila IKK, IKKgamma (Kenny), IK2, Myd88, TAK1, ECSIT, TRAF1, TRAF2, and TRAF3. Expression of each of these genes was interrupted individually by RNAi in S2 cells and the effect on Toll signaling was assayed. RNAi was also conducted against combinations of genes, in particular IKK and Pelle;IKK and IKKgamma;IKK and IK2;TRAF1, 2, and 3. To determine whether requirements are specific to the Toll pathway, the same panel of RNAi analysis was conducted in S2 cells treated with LPS. An Attacin-luciferase reporter was used to indicate LPS-mediated activation of the response pathway for Gram-negative bacteria (Sun, 2002).

When the effects of RNAi on the Toll and LPS pathways were compared, it was found that Drosophila Myd88, like Tube and Pelle, is required for activation of the Drosomycin, but not the Attacin, reporter; Drosophila Myd88 is thus essential for Toll signaling. In contrast, a requirement for TAK1 was found only in the LPS pathway and no essential role was found for fly IK2, ECSIT, or TRAF 1, 2, and 3 in either Toll or LPS signaling. Inactivating IKK and IKKgamma affected both types of signaling, with the LPS pathway being more severely inhibited than the Toll pathway. These results are consistent with the fact that inactivating IKK in flies disrupts Toll-dependent axis formation in a small fraction of embryos, although neither IKK nor IKKgamma is strictly required for Toll signal transduction (Sun, 2002).

It is known that Tube acts downstream of Toll and upstream of Pelle in signal transduction. To place Drosophila Myd88 in this pathway precisely, the epistatic relationship was examined among Myd88, Tube, and Pelle. Expression of wild-type Myd88, which has been shown to activate the Drosomycin reporter, was induced. RNAi against Pelle, Tube, or Myd88 blocks this Myd88-induced activation. These results, as well as similar findings in adult flies (Tauszig-Delamasure, 2002), indicate that Myd88 acts either upstream of or in parallel to Tube (Sun, 2002).

To dissect the signaling hierarchy further, a constitutively active form of Tube was used. This Tube-initiated activation of the Drosomycin reporter does not require Myd88, but does require Pelle. Furthermore, Pelle-induced activation of the Drosomycin reporter is diminished only by RNAi against Pelle, but not Tube or Myd88. Thus, epistasis analysis defines a linear order of action, with Tube downstream of Myd88 and upstream of Pelle (Sun, 2002).

Myd88, Tube, and Pelle each contain a death domain, a motif known to form homotypic interactions. Tube and Pelle interact directly by means of their death domains. Furthermore, Myd88 has been found to coimmunoprecipitate with Pelle in S2 cells (Tauszig-Delamasure, 2002). It was therefore interesting to discover the role of binding interactions mediated by death domains in the hierarchy defined by epistasis analysis (Sun, 2002).

To assay the interaction of Myd88 with either Pelle or Tube, full-length Myd88, as well as the death domain of Pelle (PelleDD) and a slightly larger Tube death domain peptide (TubeDD*) were epitope tagged. Also, an antiserum was generated against Drosophila Myd88. Immunoprecipitation experiments, using the alpha-Myd88 for the precipitation step and alpha-V5 to detect the tagged peptides was carried out. In pair-wise experiments, substantial interaction was detected between Myd88 and the Tube death domain. (In addition, a reduction occurred in the abundance of a fast migrating Myd88 species, perhaps reflecting a TubeDD-mediated protection from proteolysis). In contrast, only a trace amount of PelleDD coprecipitated with Myd88 (Sun, 2002).

PelleDD, TubeDD, and Myd88 were co-expressed to assay for higher-order complexes. Under such conditions, the amount of Myd88-associated PelleDD was dramatically increased. Indeed, the relative amount of TubeDD and PelleDD coimmunoprecipitated with Myd88 was indistinguishable. It is concluded that Tube forms a stable complex with Myd88 and is also strictly required for the recruitment of Pelle into a complex with Myd88 (Sun, 2002).

Two alternative models were envisioned for the role of Tube in complex formation. In one, the interaction of Pelle and Tube is essential for Pelle to join the Myd88 complex. In the alternative model, Pelle can stably associate with Myd88, provided Myd88 is bound by Tube. To discriminate between these two models, interaction surface mutations were used in characterizing a dimer between the Tube and Pelle death domains (Sun, 2002).

The crystal structure of the complex formed by the death domains of Tube and Pelle suggests that residue E50 in Tube and R35 in Pelle form a salt bridge that is critical for dimer formation. By using an RNA injection bioassays, it has been demonstrated that mutation of residue 50 in Tube to lysine (E50K mutation) abolish Tube function in Toll signaling. It was therefore surprising to find that the E50K mutation has no discernible effect on the binding of the Tube death domain to Myd88 (Sun, 2002).

Although Tube E50K has an apparently wild-type interaction with Myd88, this mutation blocks the binding of Tube to Pelle in the coimmunoprecipitation assay. Furthermore, a mutational change in Pelle (Pelle R35E) that is predicted to reconstitute the salt bridge, fully restores the Tube-Pelle interaction, just as these compensatory mutations in Tube and Pelle together allow signaling in embryos. Thus, at least two types of death domain contacts are in the Toll signaling complex: one between Tube and Pelle that involves Tube E50 and a second between Tube and Myd88 that is E50-independent (Sun, 2002).

To determine whether binding to Tube is essential for Pelle recruitment into the Myd88 complex, advantage was taken of the compensatory mutations in Tube and Pelle. In cells coexpressing TubeDD and PelleDD, the association of PelleDD with Myd88 is greatly inhibited by either individual mutation, Pelle R35E or Tube E50K, that blocks the Tube-Pelle interaction. Remarkably, the simultaneous presence of these compensatory mutations restores the recruitment of PelleDD to the Myd88 complex. It is therefore concluded that Pelle must bind directly to Tube to join the Myd88 complex (Sun, 2002).

A model is proposed that describes the three way interaction between Pelle, Tube and Myd88. The Tube-mediated complex formation involves two distinct binding surfaces on Tube death domain, which allow simultaneous association of Myd88 and Pelle. Direct binding of Myd88 to Toll and of both Tube and Pelle to Cactus-bound Dorsal would result in a complex facilitating efficient signal transduction from Toll to Dorsal. On the basis of this model of the heterotrimeric death domain complex, it is predicted that expression of the wild-type death domain of either Tube or Pelle might disrupt formation of an endogenous trimeric complex and thereby interfere with the normal function of the Toll pathway. Moreover, distinct outcomes are expected for expression of mutant forms of the Tube and Pelle death domains. The E50K mutant of TubeDD, although incapable of interacting with Pelle, nevertheless binds to Myd88 and hence might interfere with the formation of the complex of Myd88, Tube, and Pelle. By the same logic, expressing the R35E mutant of Pelle, which cannot stably interact with Tube, and hence the trimeric complex, might not interfere with signaling (Sun, 2002).

To test these hypotheses, the effect of expressing Tube and Pelle death domains was assayed in the context of an active Toll pathway. Wild-type and E50K TubeDD each significantly block TollDeltaLRR-induced activation of the Drosomycin reporter, as does wild-type PelleDD. However, the R35E mutant of PelleDD, expressed at the same level as its wild-type counterpart, has no discernible effect on Drosomycin activation. These results thus confirm the predictions of the model for heterotrimer formation and demonstrate that formation of the trimeric Myd88, Tube, and Pelle complex is a critical step in Toll signaling (Sun, 2002).

The death domain was originally identified as a protein module transducing apoptotic signals. It has been found, for example, that death domain mediated interactions between Fas and FADD or between tumor necrosis factor receptor and TRADD provide the basis for assembling the death-inducing signaling complex. The death effector domain and caspase recruitment domain also form homotypic interactions involved in apoptotic signaling and are structurally similar to a death domain. These motifs, together with the death domain, comprise the death domain superfamily (Sun, 2002 and references therein).

Experimental data demonstrate that PelleDD and Myd88 are found in the same complex when each is physically associated with TubeDD. The association of three different death domains has also been implied by studies on the tumor necrosis factor receptor complex, in which TRADD was found to facilitate the recruitment of FADD or RIP to tumor necrosis factor receptor. This study has probed the nature of such a complex and it was found that Myd88, Pelle, and Tube form a heterotrimer, with the TubeDD interacting with Myd88 and PelleDD by distinct binding surfaces (Sun, 2002).

Recently, molecular modeling based on available structural data has suggested that the homotypic interaction among death domain superfamily modules could be multivalent. Higher-order multimers, such as a heterohexamer, can be modeled by docking the death domains of Fas and FADD together. Furthermore, the structural plasticity observed in the PelleDD:TubeDD dimerhypothetically allows it to accept a third death domain into a three-fold symmetric structure. Whether the death domains of Myd88, Tube, and Pelle can form such a structure, as opposed to a linear array, awaits biophysical characterization of this trimeric complex. It is noted, however, that no evidence has been provided for any physical interaction between Pelle and Myd88. In addition, the fact that the PelleDD R35E mutant fails to dominantly interfere with Toll signaling in a functional assay argues against the possibility of such a direct contact between Pelle and Myd88 (Sun, 2002).

Because Myd88 binds to Toll through interaction between TIR domains on both proteins, it is envisioned that Myd88 connects both Tube and Pelle to Toll. Toll-initiated aggregation of these signaling molecules could trigger Pelle activation. Such a model is consistent with epistasis analyses indicating a linear order of action for Toll, Myd88, Tube, and Pelle in primary signaling (Tauszig-Delamasure, 2002). Furthermore, because Dorsal binds directly to Pelle, Tube, and Cactus, it is conceivable that the entire signaling cassette exists, at least transiently, in a single complex. As suggested by both biochemical and biological assays, Pelle-catalyzed phosphorylation may then lead to both Dorsal nuclear transport and complex dissociation (Sun, 2002).

In mammalian signaling pathways initiated by either Toll-like receptors or IL-1 receptors, Myd88 associates with IRAK. Because this study shows that a third death domain is required to mediate the interaction between Myd88 and Pelle in Drosophila, does a parallel exist in mammals? Although no known Tube ortholog exists in mammals, multiple IRAKs are present. It is speculated, therefore, that two or more IRAKs may participate in one protein complex, with the death domain of one IRAK bridging the interaction of another with Myd88. In this way, a particular IRAK isoform might act together with Myd88 to regulate the activity of a second IRAK through the oligomerization of death domains, resulting in isoform-specific biological functions (Sun, 2002).

Myd88 interacts with Pelle and Tube in the establishment of dorsoventral pattern in the Drosophila embryo

In Drosophila, the dorsoventral axis is set up by the action of the dorsal group of genes and cactus, which have been ordered genetically in a linear pathway. krapfen (kra) has been identified as a new member of the dorsal-group genes. kra (accepted FlyBase name: Myd88) encodes for the Drosophila homolog of MyD88, an adapter protein operating in the mammalian IL-1 pathway. Epistasis experiments reveal that kra acts between the receptor Toll and the cytoplasmic factor Tube. There is a direct interaction between Kra and Tube presumably mediated by the death domains present in both proteins. Tube in turn interacts with its downstream effector Pelle through death domain association. It is therefore suggested that upon Toll activation, Kra associates with Pelle and Tube, in an heterotrimeric complex (Charatsi, 2003).

Whether Kra participates in the activation of the Toll receptor or whether Kra cooperates in the signal transduction downstream of Toll was investigated. The effect of loss of maternal kra was tested in a dominant Toll background, Toll^9Q, which is a ligand-independent gain-of-function allele of Toll. Embryos laid by Toll^9Q heterozygous flies show a strongly ventraliz ed phenotype due to the constitutively active Toll that signals throughout the embryo circumference. Embryos laid by females with the genotype kra¹/kra¹; Toll^9Q/TM3 show a complete dorsalized phenotype. Thus, kra suppresses the constitutive Toll signal, indicating that Kra acts downstream of the Toll receptor. This finding suggests that Kra operates in the cytoplasmic compartment of the Drosophila early embryo. In order to place Kra upstream or downstream of the cytoplasmic protein Tube, the phenotype of kra embryos was analyzed after microinjection of the gain-of-function construct of Tube, pBtor⁴⁰²¹Tube. In this construct, the intracellular kinase domain of a gain-of-function allele of the receptor tyrosine kinase torso and is replaced by the tube coding sequence. pBtor⁴⁰²¹ fusions have been used previously to show that Tube operates upstream of Pelle. Whereas uninjected kra embryos develop only dorsal epidermis, kra embryos injected with pBtor⁴⁰²¹Tube RNA can specify ventrolateral fates and restore ventrolateral pattern elements, such as ventral denticle belts and Filzkörper, that are never observed in kra mutant embryos. Through these methods kra can be placed downstream of Toll and upstream of tube (Charatsi, 2003).

The genetic positioning of kra between Toll and tube, however, is not informative as to the physical interactions that take place during signal transduction. In order to investigate the molecular role of kra, a yeast two hybrid assay was performed. kra encodes the Drosophila homolog of MyD88, a predicted cytoplasmic adapter protein, operating in the mammalian IL-1 (Interleukin-1) pathway. It shows a modular organization, a death domain [which is generally found in many members of the tumour necrosis factor receptor (TNF-R) superfamily], an intermediate domain, and a Toll cytoplasmic domain similar to that found in the Toll/Interleukin-1-like receptors family, the TIR domain. In the yeast two hybrid assay, wild type Kra as well as Kra⁵⁶, which is an EMS allele carrying a missense mutation within the TIR domain, were both able to interact strongly with Tube. This indicates that the TIR domain is not required for a Kra-Tube interaction (Charatsi, 2003).

Pelle interacts with Tube through death domain association. In the same yeast two hybrid experiment no direct interaction was found between Pelle and Kra. This suggests that Tube could mediate the formation of a complex by association with both Pelle and Kra (Charatsi, 2003).

The epistasis experiment placed Kra downstream of the receptor Toll. Both Toll and Kra contain TIR domains, which could potentially mediate their interaction. Additionally, the kra⁵⁶ allele which shows a dorsalized phenotype presumably caused by a defective TIR domain, strongly suggests that this cytoplasmic domain plays an essential role in signal transduction. An interaction between Kra and Toll, which is not necessarily direct, is supported by immunoprecipitation experiments in which Kra/dMyD88 coimmunoprecipitates with Toll. However, in the yeast two hybrid assay, no direct interaction was found between the TIR cytoplasmic domain of Toll and Kra (Charatsi, 2003).

In order to understand how the signal is transduced to Kra through Toll, the possibility that Kra homodimerises was investigated. In mice, MyD88 was shown to form homodimers in vivo through death domain-death domain and TIR-TIR interactions. Kra does not homodimerize in yeast two hybrid assay nor in immunoprecipitation experiments (Charatsi, 2003).

Therefore Drosophila Myd88 plays a crucial role in the dorsoventral patterning of the embryo through the Toll pathway. Embryos mutant for three new kra alleles display a dorsalized phenotype. Two of these alleles, kra¹ and kra², carry large deletions removing the ATG, and represent null alleles. Embryos from homozygous females carrying these alleles are completely dorsalized, whereas the EMS missense allele, kra⁵⁶, causes a partial dorsalization. Since the EP insertions do not cause the dorsalization of the embryo, it is evident that the Toll pathway is more sensitive to the amount of dMyd88 in the immune response than in the dorsoventral system (Charatsi, 2003).

RNA injection as well as two hybrid data reveal that Myd88/Kra associates with and activates Tube, but not Pelle. Thus, Tube interacts both with Kra and Pelle. A model is proposed in which Tube as a central molecule binds to Kra and Pelle to form a heterotrimeric complex which is essential for the signal transduction to downstream components. This view is supported by immunoprecipitation experiments in which Kra/MyD88 was shown to coimmunoprecipitate with Pelle. These associations are presumably mediated by the death domains present in all these proteins. This model is supported by studies on other death domain containing proteins in the apoptotic signalling pathways that form multimeric complexes (Charatsi, 2003).

Immunoprecipitation experiments with the Drosophila and the murine Myd88 suggest that Kra associates with the receptor Toll to transduce the extracellular signal. However, Kra does not interact with Toll in the two hybrid assay. Moreover, in contrast to the murine Myd88, Kra does not homodimerize. These findings together with the comparison of the sequences point out that despite the homology shared by MyD88 and Kra, functional differences may exist. Unlike MyD88, Kra contains an additional 81 amino acids preceding the death domain and a 162 amino-acid long C-terminal region, both regions are as yet of unknown function. It is therefore concluded that homodimerization plays no role in the mechanism of signal transduction (Charatsi, 2003).

The negative results of the two hybrid experiments could be explained by the interaction of Kra and Toll mediated by an unknown component. Since no apparent signalling domains can be identified either in Kra or in Toll, it is likely that additional proteins are recruited. Alternatively, Kra could interact only with an activated form of Toll. It has been suggested that transduction of the signal from Toll to downstream components could be dependent on a conformational change of the Toll receptor upon ligand binding, an idea that is supported by structural studies of the Toll receptor. However, this alternative can not be tested by a yeast two hybrid assay (Charatsi, 2003).

The Toll pathway shows striking similarities to the IL-1 pathway. Toll in Drosophila and the IL-1R in mammals are related transmembrane receptors. As is the case during Toll signalling, binding of IL-1 to the receptor complex IL-1R/IL-1AcP (Interleukin-1 Receptor/Interleukin-1 Accesory Protein) leads to activation of transcription factors of the Rel family. Upon ligand binding MyD88 associates with IL-1AcP and mediates the formation of the receptor complex. In the Drosophila Toll pathway and the homologous mammalian pathway most components are conserved. However, since Tube has been identified within only the Drosophila genus, it had previously been proposed that Tube is the functional homolog of the vertebrate MyD88, operating between Toll and Pelle. The identification of a functional Myd88 in Drosophila shows that Tube is an additional factor. Further studies on the interaction between Tube and Kra/dMyd88 should help to understand the enigmatic role of Tube in the establishment of the dorsal gradient in the early Drosophila embryo (Charatsi, 2003).

NF-kappaB, IkappaB, and IRAK control glutamate receptor density at the Drosophila NMJ

NF-κB signaling has been implicated in neurodegenerative disease, epilepsy, and neuronal plasticity. However, the cellular and molecular activity of NF-κB signaling within the nervous system remains to be clearly defined. This study shows that the NF-κB and IκB homologs Dorsal and Cactus surround postsynaptic glutamate receptor (GluR) clusters at the Drosophila NMJ. Mutations in dorsal, cactus, and IRAK/pelle kinase specifically impair GluR levels, assayed immunohistochemically and electrophysiologically, without affecting NMJ growth, the size of the postsynaptic density, or homeostatic plasticity. Additional genetic experiments support the conclusion that cactus functions in concert with, rather than in opposition to, dorsal and pelle in this process. Finally, evidence is provided that Dorsal and Cactus act posttranscriptionally, outside the nucleus, to control GluR density. Based upon these data it is speculated that Dorsal, Cactus, and Pelle function together, locally at the postsynaptic density, to specify GluR levels (Heckscher, 2007).

NF-κB signaling has been implicated in the mechanisms of neural plasticity, learning, epilepsy, neurodegeneration, and the adaptive response to neuronal injury. The data presented in this study advance the understanding of neuronal NF-κB signaling in two ways. First, multiple lines of evidence are presented that NF-κB/Dorsal signaling is required for the control of GluR density at the NMJ. These data provide a synaptic function for NF-κB signaling that may be directly relevant to the diverse activities ascribed to NF-κB in the nervous system. Second, molecular and genetic evidence is provided that Dorsal, Cactus, and Pelle may function posttranscriptionally, at the postsynaptic side of the NMJ, to specify GluR density during postembryonic development (Heckscher, 2007).

Several independent lines of experimentation suggest that Cactus, Dorsal, and Pelle function together at the PSD to specify GluR density. Evidence is provided that Cactus and Dorsal localize to a similar postsynaptic domain. In addition, overexpression of a GFP-tagged Pelle protein that is sufficient to rescue a pelle mutation, can traffic to the PSD where Cactus and Dorsal reside. Next, genetic evidence is presented that cactus, dorsal, and pelle function together, in the same genetic pathway, to control GluR density. It is particularly surprising that mutations in cactus behave similarly to dorsal and pelle. In other systems (embryonic patterning and immunity), Cactus inhibits Dorsal-mediated transcription by binding and sequestering cytoplasmic Dorsal protein. As a result, in these other systems, cactus mutations cause phenotypes that are opposite to those observed in dorsal mutations. This study used the same cactus and dorsal mutations that previously have been observed to generate the predicted opposing phenotypes during embryonic patterning, and yet it was observed that cactus phenocopies the dorsal mutations. In addition, genetic epistasis experiments indicate that these genes function together to facilitate GluR density. Thus, at the NMJ, Cactus functions in concert with, rather than in opposition to, Dorsal (Heckscher, 2007).

One explanation for this observation could be that Dorsal does not function as a nuclear transcription factor during the control of GluR levels. In support of this idea it has been demonstrated that (1) Dorsal protein is not detected in the nucleus, (2) reporters of Dorsal-dependent transcription fail to show activity in muscle nuclei, and (3) mutation of the Dorsal transactivation domain, dl^U5 does not impair GluR abundance even though this same mutation has been shown to impair transcription-dependent patterning during embryogenesis. An alternative explanation for the observation that dorsal and cactus have similar phenotypes at the NMJ could be that Cactus and Dorsal act synergistically to control the transcription of GluRs at the NMJ. Indeed, there is evidence in other systems that IκB can shuttle with NF-κB to the nucleus. A previous study shows Cactus accumulation in Drosophila larval muscle nuclei in a dorsal mutant background (Cantera, 1999). However, this result could not be repeated despite examination of Cactus localization in five allelic combinations of dorsal. Furthermore, the data from vertebrate systems suggest that IκB should shuttle into the nucleus with NF-κB, not in its absence. Thus, a model is favored in which Dorsal and Cactus function together at the postsynaptic membrane to facilitate GluR abundance during development (Heckscher, 2007).

If this model is correct, then it is predicted that NF-κB does not control GluR density through transcriptional regulation. This prediction is supported by two experimental observations: (1) GluR transcript levels (assessed by QT PCR) are not statistically different from wild-type in dorsal and cactus mutations that cause an ~50% decrease in GluR abundance; (2) it was demonstrated that overexpression of a myc-tagged GluRIIA cDNA using a heterologous, muscle-specific promoter is not able to restore synaptic GluRIIA levels in either the cactus or dorsal mutant backgrounds. These data are consistent with Dorsal and Cactus acting posttranscriptionally to control GluR density at the NMJ. There are two general mechanisms by which GluR levels could be controlled posttranscriptionally: (1) altered receptor delivery to the NMJ or (2) altered receptor internalization/degradation. If receptor internalization/degradation were enhanced in the cactus, dorsal, or pelle mutant backgrounds, one might expect GluRIIA-myc overexpression to overcome this change and restore normal receptor levels. In addition, less myc-tagged protein might be seen in the mutants in comparison to wild-type. This is not what was observed. Therefore, the hypothesis is favored that Cactus, Dorsal, and Pelle function together to promote the delivery of glutamate receptors to the NMJ during development (Heckscher, 2007).

The possibility that Cactus, Dorsal, and Pelle act posttranscriptionally to control GluR density raises many questions. For example, do Dorsal and Cactus exist as a protein complex at the PSD? If so, is this complex regulated and how might such a complex influence GluR density? Since pelle kinase-dead mutants impair GluR density, it is possible that Dorsal and Cactus recruit Pelle to the PSD. If so, what are the targets of Pelle kinase that are relevant to establishing or maintaining the proper density of glutamate receptors at the PSD? Finally, the demonstration that cytoplasmic NF-κB/Dorsal can influence GluR density does not rule out the possibility that NF-κB/Dorsal may also translocate to the muscle nucleus at the Drosophila NMJ under certain stimulus conditions. Indeed, in both the vertebrate central and peripheral nervous systems NF-κB is found within neuronal and muscle nuclei, and nuclear translocation can be stimulated by neuronal activity, glutamate, injury, and disease. For nuclear entry of Dorsal, two events must occur: (1) Cactus must be degraded and (2) Dorsal must be phosphorylated. It remains possible that one or both of these criteria are not met during the normal development of the Drosophila NMJ but could be met under as-yet-to-be-identified stimulus conditions. The possibility that NF-κB acts both locally at the synapse and globally via the nucleus is not unique to this signaling pathway. A similar organization has been documented for wingless/wnt signaling where noncanonical cytoplasmic signaling can impact cytoskeletal organization while canonical signaling involves the nuclear translocation of downstream beta-catenin and TCF-dependent gene transcription (Heckscher, 2007).

It remains unknown how NF-κB signaling is activated at the Drosophila NMJ. In Drosophila embryonic patterning and innate immunity, NF-κB signaling is initiated through activation of Toll or Toll-like receptors. There are nine Toll and Toll-like receptors encoded in the Drosophila genome. However, none of these receptors appear to be present in Drosophila larval muscle. The Toll receptor is expressed in a subset of embryonic muscle fibers, but is absent from larval muscle. None of the Toll-like receptors are expressed in Drosophila embryonic muscle and none appear to be expressed in larval muscle. An alternative possibility is that TNF-α receptors activate NF-κB in Drosophila muscle as has been observed in vertebrate skeletal muscle. Indeed, a TNF-α receptor homolog (Wengen) has been identified, and it is expressed in Drosophila skeletal muscle. The possibility that TNF-α signaling is mediated via NF-κB is intriguing given the recent demonstration that TNF-α regulates GluR abundance in the vertebrate central nervous system. In both cultured neurons and hippocampal slices glial-derived TNF-α signaling is required for the increase in postsynaptic AMPA receptor abundance observed following chronic activity blockade. Thus, the current data in combination with work in the vertebrate CNS raise the possibility that a conserved TNFα/NF-κB signaling system controls GluR abundance at both neuromuscular and central synapses during development and in response to chronic activity blockade (Heckscher, 2007).

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.