The Drosophila innate immune system discriminates between pathogens and responds by inducing the expression of specific antimicrobial peptide‐encoding genes through distinct signaling cascades. Fungal infection activates NF‐κB‐like transcription factors via the Toll pathway, which also regulates innate immune responses in mammals. The pathways that mediate antibacterial defenses, however, are less defined. We have isolated loss‐of‐function mutations in the caspase encoding gene dredd, which block the expression of all genes that code for peptides with antibacterial activity. These mutations also render flies highly susceptible to infection by Gram‐negative bacteria. Our results demonstrate that Dredd regulates antibacterial peptide gene expression, and we propose that Dredd, Immune Deficiency and the P105‐like rel protein Relish define a pathway that is required to resist Gram‐negative bacterial infections.
Innate immune systems in insects and mammals recognize distinct classes of microbes and activate effector genes through conserved signaling pathways. In mice, the Toll‐like receptor 4 is primarily involved in the recognition of lipopolysaccharide, a component of Gram‐negative bacteria, while Toll‐like receptor 2 mediates Gram‐positive bacterial recognition (Takeuchi et al., 1999; Beutler, 2000). These two receptors utilize similar signaling cascades to activate NF‐κB, a central transactivator of many immune and inflammatory genes. In Drosophila, molecules directly involved in microbial recognition remain poorly characterized (Khush and Lemaitre, 2000; Kim et al., 2000); however, it is apparent that the Drosophila innate immune system discriminates between pathogens and responds by inducing the expression of specific antimicrobial peptide‐encoding genes (reviewed in Engstrom, 1999; Anderson, 2000; Imler and Hoffmann, 2000; Khush and Lemaitre, 2000). Genetic analysis demonstrates that the Toll signaling cascade controls an antifungal response (Lemaitre et al., 1996). This pathway is triggered by the proteolytic cleavage of the Toll ligand, Spätzle (Spz), and leads to activation of the rel proteins DIF and Dorsal (Manfruelli et al., 1999; Meng et al., 1999; Rutschmann et al., 2000). Mutations that block this cascade reduce the expression of the antifungal peptide gene drosomycin and increase susceptibility to fungal infection (Lemaitre et al., 1996). In addition, a subset of the genes encoding peptides with antibacterial activity are induced to lower levels in flies deficient for Toll signaling, indicating that this pathway also plays a role in antibacterial immune responses (Lemaitre et al., 1996). Mutations in the molecularly uncharacterized immune deficiency (imd) gene, by contrast, only affect the induction of genes with antibacterial activity and increase susceptibility to bacterial infections (Lemaitre et al., 1995). imd;Toll double mutants fail to express any antimicrobial genes, suggesting that imd and Toll define two essential pathways that regulate antimicrobial gene expression (Lemaitre et al., 1996). Recently, Hedengren et al. (1999) determined that mutations in relish, a gene encoding a P105‐like rel protein, reduce the expression of all antimicrobial genes after bacterial infection and concluded that relish may function downstream of imd. Here, we show that the Dredd caspase, which regulates apoptotic pathways in Drosophila, also mediates immune responses to infection by Gram‐negative bacteria.
The dredd caspase regulates Drosophila immune responses
To identify genes that control Drosophila antibacterial immune responses, we screened for mutations on the X chromosome that affect the expression of the antibacterial peptide gene diptericin after bacterial infection (R.S. Khush and B. Lemaitre, unpublished results). Among 2500 EMS mutagenized lines, we isolated five viable, recessive mutations (named B118, F64, L23, D55, D44) of a gene that is required for the expression of a diptericin–GFP reporter gene in larvae after bacterial infection (Figure 1A). In addition, northern blot analysis shows that adults homozygous for each of the five alleles do not express the diptericin gene after bacterial injection (Figure 1B). We mapped the B118 allele to cytological region 1B9–1B13 on the proximal tip of the X chromosome and identified a small deficiency, Df(1)R194, which does not complement B118 (data not shown). Deficiency Df(1)R194 spans four previously identified genes: rpL36, l(1)1Bi, dredd and su(s) (Chen et al., 1998) (Figure 1C). Several results demonstrate that B118 is a mutation in dredd: (i) B118 is allelic to a viable P element insertion (EP‐1412) inserted 50 bp upstream of dredd coding sequences (data not shown; Figure 1C); (ii) the two genes flanking dredd, su(s) and l(1)1Bi complement B118 (data not shown); (iii) a small deficiency, Df(1)dreddD3, which we generated by imprecise P element excision, and which removes dredd and affects the 5′ upstream sequences of su(s), blocks diptericin expression after bacterial infection (Figure 1B and C); and (iv) a P element insertion, P[dredd+], containing 7.6 kb of genomic DNA, including dredd but not su(s) and l(1)1Bi (Chen et al., 1998), fully restores diptericin expression in B118 flies (Figure 1B and C). All five dredd EMS mutations block diptericin expression after infection to the same degree as Df(1)dreddD3, indicating that they are probably null alleles (Figure 1B; see Methods). The P element insertion in line EP‐1412 generates a strong hypomorphic dredd mutation since a small amount of diptericin expression is detectable after infection (Figure 1B).
dredd encodes an apical caspase and is an effector of the apoptosis activators reaper, grim and hid (Chen et al., 1998). One or more dredd transcripts are specifically enriched in cells programmed to die and dredd overexpression induces apoptosis in SL2 cells (Chen et al., 1998; Rodriguez et al., 1999). In mammals, the closest dredd homologs are caspases 8 and 10, which mediate apoptosis induced by members of the tumor necrosis factor receptor family (Reed, 1999). Caspases are produced as inactive zymogens termed pro‐caspases; when activated, mature caspases catalyze the proteolytic cleavage of death substrates that are associated with apoptosis (Reed, 1999). The isolation of mutations in dredd that block diptericin expression after infection demonstrates that Dredd also regulates immune responses. In addition, a dredd–lacZ reporter gene is constitutively expressed in all adult and larval tissues including the fat body, the major immuno‐responsive tissue in insects (Figure 1D). Infection does not, however, appear to increase dredd expression levels (data not shown).
The five dredd alleles all contain point mutations that affect different regions of the dredd protein (Figure 1E). Alleles B118, D55 and F64 generate either premature stop codons or frameshift changes in the Dredd prodomain. D44 has a missense mutation in sequences encoding the first death effector domain (DED), a region thought to mediate protein–protein interactions (Reed, 1999). In the protein encoded by allele L23, a tryptophan (W) in the caspase domain is replaced by an arginine (R) residue. The strong phenotype of alleles D44 and L23 indicates that Dredd domains affected in these alleles are essential for Dredd function in immunity.
dredd mediates antimicrobial peptide gene expression in response to Gram‐negative bacterial infections
The isolation of dredd mutations that block diptericin expression enabled us to characterize dredd‘s role in mediating Drosophila antimicrobial host defense as well as dredd’s relationship to other genes that function in this response. Pricking adult flies with a mixture of Gram‐positive and Gram‐negative bacteria activates the expression of all the genes that encode antimicrobial peptides in Drosophila (Figure 2A). In the dreddB118 mutant, however, mixed Gram‐positive/Gram‐negative infections only induce the expression of the antifungal gene drosomycin and the gene coding for Metchnikowin, which has both antifungal and antibacterial activity (Figure 2A); diptericin, cecropin A, attacin A and defensin are expressed at <5% of wild‐type levels and metchnikowin is expressed at 50% of the wild‐type level (Figure 2A; quantification data not shown). Antimicrobial gene expression is similarly affected in flies homozygous for relE20, a strong or null mutant allele of relish (Hedengren et al., 1999) and imd, although most of the antibacterial genes are expressed at slightly higher levels in imd flies (Figure 2A). By contrast, a mutation in the spz gene, which blocks Toll activation, reduces drosomycin induction by mixed Gram‐negative/Gram‐positive bacterial infection and reduces the induction of some of the antibacterial genes (defensin, attacin, cecropin A) (Figure 2A) (Lemaitre et al., 1996). These data demonstrate that mutations in dredd are phenotypically similar to mutations in imd and relish, and that these three genes regulate all Drosophila antibacterial peptide gene expression.
Hedengren et al. (1999), however, previously showed that drosomycin and metchnikowin are expressed to only 10–20% of the wild‐type level in relE20 flies after infection by the Gram‐negative bacteria Enterobacter cloacae. The difference between their results and our observation that drosomycin and metchnikowin are significantly induced in relish mutants after mixed Gram‐negative/Gram‐positive infections could be explained by the type of infection. To define further the roles of imd, dredd and relish in activating metchnikowin and drosomycin after different types of bacterial infection, we quantified metchnikowin and drosomycin expression in different mutant backgrounds 6 h after infection with either Gram‐negative Escherichia coli or Gram‐positive Micrococcus luteus bacteria. The dreddB118 and relE20 mutations strongly reduce metchnikowin and drosomycin induction by Gram‐negative bacterial infections (∼20–50% of wild‐type levels), while the imd mutation has a weak effect; by contrast, metchnikowin and drosomycin are expressed at close to wild‐type levels in the imd, dreddB118 and relE20 mutants after Gram‐positive bacterial infection (Figure 2B). We conclude, therefore, that dredd and relish play a greater role in inducing metchnikowin and drosomycin after Gram‐negative bacterial infection than after Gram‐positive bacterial infection.
The observation that drosomycin and metchnikowin expression is almost completely abolished in imd;Toll double mutants (Lemaitre et al., 1996; Levashina et al., 1998) suggests that Gram‐positive bacterial infection triggers the expression of metchnikowin and drosomycin via the Toll pathway. In agreement, our analysis shows that mutations in spz affect drosomycin gene expression more strongly after Gram‐positive than after Gram‐negative bacterial infection, and that the constitutive activation of the Toll pathway in the Tl10b mutant (Lemaitre et al., 1996) leads to drosomycin expression in the absence of dredd activity (Figure 2C). metchnikowin, however, is still expressed to a high level in spz mutants after Gram‐positive bacterial infection, indicating that metchnikowin induction by Gram‐positive bacterial infection may also be mediated in part by the Imd pathway.
dredd mediates resistance to Gram‐negative bacterial infections
The susceptibility to microbial infection observed in dredd, imd, relish, spz and imd;spz mutants is correlated with the expression pattern of antimicrobial genes in these mutants. dreddB118, relE20 and imd;spzrm7 adults are highly susceptible to bacterial infection by Gram‐negative bacteria, and imd adults are slightly less susceptible (Figure 3A for E. coli; data not shown for Erwinia carotovora carotovora). These survival results confirm that the activation of defense responses to Gram‐negative bacterial infection require imd, dredd and relish. Only the imd;spzrm7 double mutants, however, are highly susceptible to bacterial infection by Gram‐positive bacteria (Figure 3B for M. luteus; data not shown for Aerococcus viridans), indicating that resistance to Gram‐positive bacteria is regulated by both the Toll and Imd pathways. Finally, only spzrm7 and imd;spzrm7 mutants are highly sensitive to natural infection by the entomopathogenic fungus Beauveria bassiana (Figure 3C) or injection of Aspergillus fumigatus spores (data not shown), confirming that responses to fungi are largely activated by the Toll pathway (Lemaitre et al., 1996, 1997).
In this study we have identified Dredd, a Drosophila apical caspase previously shown to function in apoptosis (Chen et al., 1998), as a regulator of the antibacterial response. dredd mutants display very low inducibility of all the genes that encode peptides with antibacterial activity and a high susceptibility to Gram‐negative bacterial infection. This dredd immune phenotype is similar to the relish and imd phenotypes; we predict that the Imd, Dredd and Relish proteins function in a common signaling pathway that regulates antibacterial peptide gene expression. Based on the respective activites of Dredd as a caspase and Relish as a transcriptional transactivator (Dushay et al., 1996), we also hypothesize that Dredd functions upstream of Relish in the control of antimicrobial gene expression. Our hypothesis is supported by the observation that Dredd is required for Relish activation via endoproteolytic cleavage (Stöven et al., 2000). We believe that the weaker effects of the imd mutation on antibacterial gene expression place the imd gene product at an early stage of the antibacterial cascade where multiple responses, some of which bypass imd, trigger the activation of the pathway. Alternatively, the imd mutation may represent a hypomorphic allele.
Caspases were originally identified as effectors of apoptosis, but there is increasing evidence that caspases also function in other physiological processes. Recent studies suggest that the recruitment of the caspase‐8 precursor to the TNF‐R1 signaling complex either activates NF‐κB through a Traf2‐, RIP‐, NIK‐ and IKK‐dependent pathway or, after proteolytic processing of caspase‐8, induces apoptosis (Hu et al., 2000). Our data indicate that Dredd, a close homolog of caspase‐8, may also have dual functions in NF‐κB signaling and apoptosis in Drosophila. Further biochemical analysis is necessary to determine whether Dredd participates directly in Relish activation or functions further upstream.
Deciphering the mechanisms that enable Drosophila to differentiate between pathogens and mount specific immune responses is essential for understanding innate immunity. Recent studies indicate that the Toll pathway is mainly activated in response to fungal and Gram‐positive bacterial infection (Lemaitre et al., 1997; Rutschmann et al., 2000). We now present several observations that suggest that imd, dredd and relish mediate most of the responses to Gram‐negative bacterial infection: (i) these genes regulate the antimicrobial peptide genes that are most highly induced by Gram‐negative bacterial infection (Lemaitre et al., 1997); (ii) dredd and relish control the induction of metchnikowin and drosomycin after Gram‐negative bacterial infection; and (iii) these three genes are required for resistance to Gram‐negative bacterial infection. We propose a model whereby antimicrobial gene expression in Drosophila adults is regulated by a balance of inputs from the Toll pathway and the Imd pathway, which includes Imd, Dredd and Relish, and that these two pathways are differentially activated by different classes of microorganisms. Identifying the receptors that discriminate between invading microbes and stimulate these pathways presents an exciting challenge in the study of innate immunity.
CantonS flies were used as a wild‐type standard. imd, spz, Tl alleles are described elsewhere (Lemaitre et al., 1996). relE20 is a strong or null allele of relish (Hedengren et al., 1999). B118, F64, D44, D55 and L23 are five EMS mutations of dredd that were generated in a y,w chromosome. Flies homozygous or hemizygous (in trans to Df(1)dreddD3) for these five mutations are phenotypically identical to flies homozygous for Df(1)dreddD3 with respect to both diptericin and attacin induction after challenge and susceptibility to E. carotovora carotovora infection, thereby indicating that they are genetically null mutations of dredd (data not shown). Df(1)dreddD3 was generated by imprecise P element excision of EP‐1412. Other mutant lines are described in the text. diptericin–GFP is P transgene containing a fusion between 2.2 kb of upstream sequence of diptericin and the Green Fluorescent Protein gene. The precise fly genotypes are: imd: br, pr, imd; spzrm7: spzrm7/spzrm7; rel: relE20, e; dreddB118: y,w, dreddB118. For survival studies, we have used a relE20 stock from which the ebony mutation was removed since ebony affects survival levels (Lemaitre et al., 1996). Similar survival rates were observed with y,w, dreddB118 and y+ Df(1)dreddD3/y,w, dreddB118 flies, indicating that our results do not reflect deleterious effects due to the y marker. Drosophila stocks were maintained at 25°C. Infected animals were incubated either at 25°C (northern blot) or at 29°C (survival).
Bacterial infections were performed by pricking third instar larva or adults with a thin needle previously dipped into a concentrated culture (OD ∼200) of E. coli, M. luteus or a mixture of the two bacteria. Natural infections with B. bassiana were performed by shaking anesthetized flies for 30 s in a Petri dish containing a sporulating fungal culture (Lemaitre et al., 1997). Bacterial and fungal strains were previously described (Lemaitre et al., 1997).
Northern blot analysis.
Total RNA extraction and northern blotting experiments were performed as described in Lemaitre et al. (1997).
Sequencing of dredd alleles.
DNA was extracted from adult flies and the dredd ORF was amplified by PCR using the Expand Long Template PCR kit (Boehringer Mannheim) for each of the mutant alleles and then sequenced using an ABI 373 automated sequencer.
Cloning and transformation of P[dredd–lacZ].
To determine if 5′ regulatory sequences regulate dredd expression after bacterial infection in the fat body, transgenic lines carrying an ∼3.3 kb promoter fragment fused to lacZ were tested for β‐galactosidase activity in third instar larvae. The dredd 3.3 kb fragment was generated by PCR from genomic DNA obtained from yw flies as described above. The following modified oligonucleotides containing an EcoRI and a BamHI site, respectively, were used: 5′‐GGCACTCGCGAATTCAATGCTGCTAGCATGGTC‐3′ and 5′‐ATTGGATCCGGCCATGGCGATCGGGAGTATACA‐3′. The PCR‐amplified fragment was then digested with EcoRI and BamHI, gel purified and ligated into EcoRI–BamHI‐cut pCaSpeR‐AUG‐β‐gal vector. Flies bearing this construct were generated by P element‐mediated germline transformation. X‐gal staining was described previously (Lemaitre et al., 1995).
We thank Dan Hultmark for providing relish mutants and our colleagues Isaline Rowe, Stéphanie Gobin, Thomas Ranson, Sheila Vidal, Ennio de Gregorio and Pedro Santamaria for assistance and stimulating discussions.
- Copyright © 2000 European Molecular Biology Organization