Gene name - immune deficiency
Synonyms - BG5
Cytological map position - 55C9--11
Function - signal transduction
Keywords - immune response
Symbol - imd
FlyBase ID: FBgn0013983
Genetic map position -
Classification - death domain protein
Cellular location - cytoplasmic
In the mid 1990s, it became apparent that two distinct pathways control the antifungal and antibacterial responses (Lemaitre, 1995 and 1996). Indeed, it was found that the dorsoventral regulatory gene cassette spätzle/Toll/cactus directs the potent antifungal response in Drosophila adults, whereas the antibacterial defense is largely independent of these genes (Lemaitre, 1996). A mutation was discovered at that time, and referred to as imd (immune deficiency) in which the antibacterial, but not the antifungal response, was compromised (Lemaitre, 1995). Since the initial description of the imd mutation, four additional genes have been shown to participate in the antibacterial defense of Drosophila: (1) ird5, a gene encoding a homolog of mammalian IKKß (Lu, 2001; Silverman, 2000); (2) kenny, a homolog of IKKgamma/NEMO (Rutschmann, 2000; Silverman, 2000); (3) dredd, which is structurally related to vertebrate caspase-8 (Elrod-Erickson, 2000; Leulier, 2000; and (4) Relish, a member of the NF-kappaB family (Hedengren, 1999) (Georgel, 2001 and references therein).
The imd gene has now been identified. This gene encodes a 30 kDa protein with a death domain that exhibits significant similarity with the death domain of mammalian RIP (receptor interacting protein), a protein associating with the TNFalpha receptor 1 and with the Fas receptor. Epistasis experiments indicate that imd acts in the antibacterial defense upstream of the four genes mentioned above. Overexpression of the imd gene in wild-type and imd mutant flies results in forced transcription of all genes encoding antibacterial peptides, but not of that coding for the antifungal peptide Drosomycin. Similar to the dual function of RIP in vertebrates, the data also suggest a possible role for imd in apoptosis. Overexpression of imd induces transcription of the reaper gene and causes the massive appearance of TUNEL-reactive cells in the adult fat body. The effects of ectopic expression of imd on the transcription of antibacterial peptides and on the generation of TUNEL-reactive cells are blocked by coexpression of the baculovirus antiapoptotic protein P35. imd mutant flies are more resistant than control flies to UV irradiation, suggesting that imd may be involved in the apoptotic response to DNA damage (Georgel, 2001).
The induction of antimicrobial peptide genes is correlated in Drosophila to the binding of Rel proteins to kappaB-related responsive elements in their promoter sequences. This type of binding does not occur in imd1 mutant flies (Lemaitre, 1995). The ability of protein extracts from immune-challenged imd1 and imd plus transgenic flies to bind to a labeled oligonucleotide containing a kappaB-related motif from the diptericin promoter was analyzed by gel shift experiments. The binding capacity was rescued in the transgenic flies and was specific for the nucleotide sequence, as evidenced by incubation with a mutated probe (Georgel, 2001).
imd1 flies are susceptible to infection with bacteria (Lemaitre, 1995 and 1996; Leulier, 2000; Rutschmann, 2000). The survival to infection with E. coli and M. luteus (mixture) of imd1 and transgenic flies carrying a wild-type copy of BG5 was examined. As expected, imd1 flies are susceptible to infection in these experiments. However, the transgenic flies have a survival rate similar to that of wild-type flies, demonstrating that BG5 can rescue the imd1 survival phenotype. BG5 will therefore be referred to as the imd gene. The death rates of hemizygous imd1/Df (2R) 2.1 flies, which are similar to those of dredd and kenny alleles (Leulier, 2000; Rutschmann, 2000), are lower than those of imd1 homozygous flies. These data indicate that imd1 is a hypomorphic mutation. The deficiency introduced in this context has no effect by itself, as evidenced by the survival curve of the hemizygous + /Df (2R) 2.1 flies (Georgel, 2001).
Wild-type and imd1 mutant fly lines carrying a UAS-IMD transgene were generated and the effects of overexpression of imd were analyzed using different GAL4 drivers. Expression of imd directed by the hs-GAL4 driver in imd1 mutant flies restores the immune inducibility of all the antibacterial peptides to levels comparable to wild-type. Remarkably, in these flies, the transcription of all the antibacterial peptide genes occurs even in the absence of immune challenge, indicating that overexpression of imd has a dominant effect on the activation of the pathway controlling the antibacterial peptide genes. drosomycin expression is not affected in these experiments; this is consistent with previous observations that drosomycin expression is independent of imd (Lemaitre, 1996). In UAS-IMD/hs-GAL4 flies a constitutive expression of antibacterial peptide genes is observed even in the absence of heat shock. However, no spontaneous expression of these genes was detected in any of the original fly lines carrying the UAS-IMD insertion. It is therefore proposed that the basal level of GAL4 expression from the hsp promoter accounts for the UAS-IMD-dependent transcription of the antibacterial peptide genes. This assumption is supported by the observation that the hs-GAL4 driver is able to sustain a basal level of expression of a UAS-GFP and a UAS-lacZ reporter gene in the absence of heat shock. Overexpression of imd by hs-GAL4 also has a dominant effect on antibacterial genes in wild-type flies (Georgel, 2001).
The effect of overexpression of imd was further analyzed by selecting a tissue-specific driver, yolk-GAL4, which directs transcription of UAS-dependent genes in the fat body of adult females starting 3 to 5 days after hatching. diptericin is strongly expressed in these flies in the absence of immune challenge, but not in unchallenged males of the same genotype (Georgel, 2001).
Expression of most antimicrobial peptide genes is induced by LPS in the Drosophila macrophage-like S2 cell line. These cells are transfected with an imd expression vector and strong expression is observed of attacin- (and drosocin-) luciferase reporter genes in the absence of LPS. In contrast, transfection with mutated imd (imd1) is unable to induce the attacin-luciferase reporter gene. Lack of induction of the attacin reporter gene by the imd1 construct does not reflect the inability of the cells to synthesize the recombinant protein, since the wild-type and mutated IMD proteins are expressed at comparable levels in transfected cells. These results corroborate the in vivo data indicating that the Ala31 to Val31 substitution in the imd1 allele fully accounts for the immune-deficient phenotype (Georgel, 2001).
The imd gene acts upstream of DmIKKgamma and the Caspase Dredd to activate the expression of antibacterial peptide genes. The dominant effect on antibacterial peptide genes of overexpression of imd driven by the hs-GAL4 driver (even in the absence of heat shock) serves to establish the epistatic relationships of imd with other genes in the pathway. In dredd and kenny mutant flies, expression of all antibacterial peptide genes in response to bacterial challenge, as well as survival to infection, is strongly impaired (Elrod-Erickson, 2000; Leulier, 2000; Rutschmann, 2000). The overexpression of imd is unable to confer challenge-independent expression of diptericin in dreddD55 and kenny1 mutant backgrounds. Furthermore, it does not restore immune inducibility of the diptericin gene in these mutants. It is noteworthy that drosomycin expression is not affected in these flies. These results demonstrate that imd acts upstream of both dredd and kenny to activate the antibacterial peptide genes during an immune response. In keeping with these results, transcription of the attacin-luciferase reporter gene induced by overexpression of imd in S2 cells is abolished by an ird5 (IKKß homolog) dominant-negative expression construct (Georgel, 2001).
Transgenic flies were examined in which imd expression was placed under the control of the ubiquitous da-GAL4 driver. It was surprising to observe 100% lethality in these flies during early larval development. The lethality was partially rescued by coexpression of the viral caspase inhibitor P35. This protein inactivates most of the executioner caspases of the death program. It was further noted that overexpression of imd by the fat body-specific driver yolk-GAL4 induced the transcription of a reaper-lacZ reporter in this tissue. Reaper is a key activator of apoptosis in Drosophila. A TUNEL analysis of transgenic flies overexpressing imd also revealed a remarkably large number of labeled nuclei in fat body cells, as compared to controls. This effect was suppressed by coexpression of the antiapoptotic protein P35. Transmission electron microscopy analysis revealed that the fat body cells exhibit the classical morphological aspects of apoptosis, that is, densification and fragmentation of the cytoplasm, membrane blebbing, and stacking of the endoplasmic reticulum (Georgel, 2001).
These observations raised the question as to whether the imd-dependent antibacterial response involves activation of an apoptotic program. However, when injecting gram-negative bacteria into flies, appearance of TUNEL-positive cells are never observed in the fat body nor any induction of the reaper reporter gene, although the antibacterial peptide genes were fully induced. It was reasoned that the imd-dependent control of an apoptotic pathway could be triggered by stimuli different from bacterial infection. One well-established experimental system that induces apoptosis in Drosophila is DNA damage by UV irradiation. Pupae were subjected to increasing doses of UV treatment and a significant mortality in emerging flies was noted at 50,000 µJ/cm2. This lethality is, at least partially, linked to induction of apoptosis by the UV treatment, since it is significantly reduced in flies overexpressing the apoptosis inhibitor P35. Remarkably, when imd mutant pupae are subjected to UV treatment, lethality is also significantly reduced. The reduction in the level of lethality is clearly related to the product of the imd gene, since in flies overexpressing imd an increased sensitivity to UV irradiation is observed. Also, reintroduction of a wild-type copy of the imd gene into imd1 mutant flies restores their sensitivity to UV irradiation to levels comparable to wild-type. Interestingly, dredd mutant flies are also resistant to UV irradiation. In contrast to the clear effect of UV damage on apoptosis-related mortality, this treatment does not induce the expression of diptericin in the absence of a microbial challenge (Georgel, 2001).
The observations that the imd-induced cell death is antagonized by P35 raises the question as to whether P35 also affects the immune induction of antibacterial peptides. Overexpressing p35 in adult fat body cells blocks the challenge-dependent transcription of diptericin. Furthermore, the induction of the diptericin gene by overexpression of imd is also abolished by coexpressing p35. Note that induction of drosomycin by an immune challenge is not affected by P35. These results indicate that P35-sensitive caspases act downstream of imd.
It is fully appreciated that wild-type imd is mandatory for the immune induction of Diptericin and the other antibacterial peptides, but not for that of the antifungal peptide Drosomycin (Lemaitre, 1995, 1996). Four distinct genes have now been characterized that, when mutated, induce a roughly similar phenotype. These genes code for three well-established components within signaling cascades: (1) Relish is a member of the NF-kappaB family of inducible transactivators. Like its mammalian counterpart, p105, it has a C-terminal I-kappaB-like domain, which accounts for its retention in the cytoplasm in the absence of challenge. Nuclear translocation of the Rel homology domain of Relish requires endoproteolytic cleavage by an as yet unidentified protease. (2) Two genes code for the components of a mammalian signalosome equivalent, that is, for homologs of IKKß (ird5) and IKKgamma (kenny). Both genes act upstream of Relish, by a mechanism which has not been established, but which may involve phosphorylation of Relish by IKKß. (3) Finally, the dredd gene encodes a close homolog of mammalian caspase-8. Since Relish is not challenge-dependently cleaved in dredd mutant flies, DREDD must act upstream of this transactivator in the signal transduction pathway (Georgel, 2001).
A major concern in this study was to clarify the relationship between imd and these four other known genes in the pathway. The results clearly show that imd acts upstream of both kenny and dredd. Taken together with other studies they indicate that IMD functions upstream of the DmIKK signaling complex, which in turn activates Relish directly or indirectly. These data establish an imd/kenny/ird5/Relish gene cassette in the control of the antibacterial peptide genes. The presence in IMD of a death domain with similarity to that of the adaptor protein RIP suggests that the IMD protein may be part of a receptor-adaptor complex. By analogy with the role of RIP in TNFR1 signaling, IMD could function by recruiting the signalosome (DmIKKgamma/DmIKKß) to the upstream receptor-adaptor complex (Devin, 2000; Poyet, 2000; Zhang, 2000) (Georgel, 2001 and references therein).
Both the receptor and the ligand(s) activating the IMD pathway remain unknown. This is in sharp contrast with the Toll pathway, which controls the antifungal response of Drosophila and namely the synthesis of the antifungal peptide Drosomycin. In the latter case, genetic analysis indicates that microbial ligands activate an extracellular proteolytic cascade culminating in the cleavage of the cytokine-like polypeptide Spätzle. Cleaved Spätzle in turn interacts with Toll and initiates a signal transduction pathway that leads to the transcription of drosomycin. This does not, however, activate the synthesis of the antibacterial peptides (Lemaitre, 1996). Conversely, overexpression of imd leads to the expression of the antibacterial peptide genes but not that of drosomycin. These data indicate that within the predominant immune-responsive tissue of adult flies, that is, the fat body, the Toll and the IMD pathways are functionally separated and that each controls, via specific Rel transcription factors, a set of given antimicrobial peptide genes. This functional separation of the two signaling pathways within the cells is corroborated by the results that in flies subjected to infection by a bacterial mix (gram-positives and gram-negatives), expression of the antiapoptotic protein P35 abolishes inducibility of the antibacterial peptides, but not that of drosomycin (Georgel, 2001).
Overexpressing imd results in larval lethality, expression of the reaper gene in fat body cells of adult flies, and apoptosis of these cells. Most of these effects could be, at least partially, rescued by coexpressing the antiapoptotic protein P35. The presence of a death domain in imd, however, might bring these results in line with recent studies showing that overexpression of death domain proteins induces apoptosis in mammalian cells. Overexpressing the death domain proteins RIP, FADD, and TRADD has been shown to lead to cell death by recruitment of apical caspases, namely caspase-8 and -10, which in turn activate downstream effector caspases to execute the apoptotic program. RIP, whose death domain is similar to that of IMD, has been proposed to induce cell death by recruiting FADD, which subsequently may activate caspase-8. Drosophila has a FADD homolog (DmFADD) and a caspase-8 homolog (DREDD) (Georgel, 2001). Furthermore, IMD, DmFADD, and DREDD can associate into a multimeric complex in Drosophila S2 cells (R. Medzhitov, personal communication to Georgel, 2001).
At this stage, it cannot be determine whether imd plays a role in developmentally regulated apoptosis. imd is expressed at significant levels during stages of embryogenesis and pupariation, when massive apoptosis is known to occur. However, imd mutants are viable and show no obvious developmental defect, indicating that either imd is not involved in these processes or that redundant control mechanisms could substitute for its role. Interestingly, it has been reported that reaper loss-of-function mutants are also viable, despite the accepted role of reaper as a key player in Drosophila apoptosis through caspase activation and inactivation of inhibitor of apoptosis proteins. In the context of the studies of immune defenses, the crucial question pertains to the potential links between the control of expression of the antibacterial peptides and the induction of apoptosis. Here, two lines of evidence draw a clear distinction: for one, bacterial challenge induces antibacterial peptide genes but does not result in detectable apoptosis; further, UV irradiation leads to P35-sensitive lethality, but does not induce antibacterial peptides. In other words, the data available today do not point to a role for apoptosis in the host response to bacterial infection (Georgel, 2001).
The initial reports of the imd mutation (imd1; Corbo, 1996; Lemaitre, 1995) included complementation analysis based on available deficiencies, which mapped the mutation to the 55C-55E interval on the right arm of the second chromosome. To refine the mapping of the imd gene, additional deficiencies were generated by transposase-induced male recombination. The newly generated deficiencies were analyzed for complementation of the imd1 mutation by monitoring the immune inducibility of the gene encoding the antibacterial peptide diptericin. Out of 60 deficiencies, two were of particular interest; one, Df (2R) 2.2, was the smallest deficiency which did not complement imd1, and the other, Df (2R) 2.3, was the largest deficiency complementing imd1. These two deficiencies define a 1,935 bp region that contains at least part of the imd transcription unit. The genome annotation database of Drosophila (GadFly) predicts the presence of one gene in this region, BG5 (CG5576). Df (2R) 2.2 uncovers the upstream promoter region of this gene, since the breakpoint of the deficiency corresponds to the second nucleotide of the corresponding cDNA (EST GH20785). The nucleotide sequence of BG5 codes for a 273-residue protein with an apparent 80-residue death domain (DD) in the C-terminal part. This domain shows marked sequence similarity (33%) with the death domain of the mouse and human 75 kDa RIPs. In contrast to mammalian RIP, the N-terminal region of the protein deduced from the BG5 sequence has no apparent kinase domain. Significant similarity was also observed between the DD encoded by BG5 and those of the death receptors TNFR1, DR3, and DR5, and the TNFR1- and Fas-associated DD proteins TRADD and FADD (Georgel, 2001).
The genomic region surrounding BG5 was amplified in imd1 homozygous and wild-type flies by PCR. The comparison of the two DNA sequences revealed a single nucleotide substitution, changing amino acid Ala31 in the wild-type to Val31 in the imd1 mutant flies. The transcription profile of BG5 was also investigated and high expression levels were observed in 6- to 24-hr-old embryos and at the time of pupariation. Note that the imd gene is also upregulated by immune challenge (Georgel, 2001).
The imd gene encodes a 30 kDa protein with a C-terminal death domain (DD). This 80-residue conserved domain is shared by a growing number of proteins. It is characterized by the presence of six sequential alpha helices interrupted by loop regions and is usually located at the C terminus. Death domain-containing proteins include death receptors, adaptor proteins, transcription factors, and structural proteins, which serve diverse cellular functions essentially by mediating homo- and hetero-typic protein-protein interactions. In addition to imd, four death domain-containing proteins have been described in Drosophila: (1) the serine/threonine kinase Pelle, a homolog of the mammalian kinases IRAKs, which are associated with the IL-1 and the Toll-like receptors; (2) the adaptor protein Tube; (3) DmFADD, the Drosophila homolog of FADD. Pelle and Tube play a role in the Toll signaling pathway both during embryogenesis and in the antifungal response. The role of DmFADD is not fully elucidated. It has been reported to interact with the caspase DREDD. (4) The protein MyD88 in the group associates an N-terminal death domain with a C-terminal TIR (Toll-Interleukin receptor homolog) domain and appears homologous to mammalian MyD88 (Georgel, 2001).
The sequence of the DD encoded by the imd gene is closest to that of mammalian RIP, which reportedly interacts with the death receptors TNFR1 and Fas. It has been proposed that recruitment of RIP to the receptor complex plays an important role in mediating NF-kappaB activation (TNFR1) and/or apoptosis (Fas, TNFR1) by these receptors, and RIP has been shown to be mandatory for TNFR1-dependent NF-kappaB activation in response to TNFalpha. It is of relevance to note that the DD of Imd also shows significant homology to that present in TNFR1, TRADD, and FADD. In the mammalian system, TNFR1, TRADD, RIP, and FADD interact through their DDs in the context of inflammatory and apoptotic responses to TNFalpha. Outside of the death domain, however, the IMD and RIP proteins are largely divergent. In fact, in contrast to RIP, the much shorter Imd lacks an N-terminal serine/threonine kinase domain (which, however, is dispensable for NF-kappaB activation). Thus, structurally, Imd does not qualify as a sensu stricto homolog of RIP (Georgel, 2001).
date revised: 30 October 2001
Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.