A Direct Role for Dual Oxidase in Drosophila Gut Immunity

See allHide authors and affiliations

Science  04 Nov 2005:
Vol. 310, Issue 5749, pp. 847-850
DOI: 10.1126/science.1117311


Because the mucosal epithelia are in constant contact with large numbers of microorganisms, these surfaces must be armed with efficient microbial control systems. Here, we show that the Drosophila nicotinamide adenine dinucleotide phosphate (NADPH) oxidase enzyme, dual oxidase (dDuox), is indispensable for gut antimicrobial activities. Adult flies in which dDuox expression is silenced showed a marked increase in mortality rate even after a minor infection through ingestion of microbe-contaminated food. This could be restored by the specific reintroduction of dDuox, demonstrating that this oxidase generates a unique epithelial oxidative burst that limits microbial proliferation in the gut. Thus, oxidant-mediated antimicrobial responses are not restricted to the phagocytes, but rather are used more broadly, including in mucosal barrier epithelia.

The innate immune system provides an essential means of host defense in eukaryotes against a broad spectrum of microorgansims (1), and the production of microbicidal reactive oxygen species (ROS) is a key feature of this protective response (26). To date, most studies have focused on the molecular mechanism of respiratory burst in the professional phagocytes in response to microbial infection (26). In contrast, the oxidant-dependent antimicrobial properties in mucosal epithelia, which are in permanent contact with the microbial realm, remain largely unknown. In Drosophila, the nuclear factor κB (NF-κB) pathways are critical during systemic infection (711) but appear to be less than crucial for host survival after epithelial infection (12). Natural gut infection has been associated with the rapid synthesis of ROS (12), and the dynamic cycle of ROS generation and elimination appears to be vital in Drosophila, because flies that lack ROS-removal capacity have an increased mortality (12). Such observations suggest an imortant role for ROS generation in controlling epithelial infection.

To directly examine if the epithelial oxidative burst system is required for host survival, we tested the potential superoxide-producing activity of intestinal epithelia in vitro (13). A basal level of superoxide generation was maintained in the membrane fraction of dissected intestines, and this increased markedly in the presence of calcium in a dose-dependent manner (Fig. 1A). Treatment with EGTA, or diphenylene iodonium (DPI), which is a flavo-protein inhibitor that also inhibits the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase–dependent oxidative burst, completely blocked calcium-activated intestinal superoxide-producing activity (Fig. 1A). In humans, phagocytic cells generate the ROS precursor, superoxide, via the phagocytic oxidase (phox) complex (2, 5). Recently, the human genome has been shown to contain several NADPH oxidase family members [currently designated the Nox 1-5 and dual oxidase (Duox) 1-2], each of which is homologous to the phox catalytic subunit, gp91phox/Nox2 (14, 15). The Duox family can be distinguished from the Nox family based on the presence of an N-terminal extracellular peroxidase-homology domain (PHD) in addition to the gp91phox-like oxidase domain (14, 15). The Nox/Duox family of enzymes are expressed in a variety of nonphagocytic cells, suggesting that they require oxidase functions similar to those of gp91phox/Nox2 (1618). Recently, Duox has been shown to be expressed in the barrier epithelia, including epithelial cells of mucosal surfaces of colon, rectum, salivary gland ducts, and bronchi (1820), and it has been suggested that Duox may provide an epithelial ROS source in host defense (1820). To determine the in vivo role of Drosophila Nox and Duox homologs (dNox and dDuox, respectively) (fig. S1) with regard to epithelial immunity, we generated a set of loss-of-function transgenic flies carrying specific RNAi (RNA interference) constructs (fig. S2) (13). Natural infection experiments revealed that ubiquitous expression of dDuox-RNAi resulted in a consistently increased mortality rate for a number of microorganisms (Fig. 1B and fig. S3). A similar result was obtained when dDuox-RNAi was restricted to the intestine (by using cad-GAL4 driver), but not when it was introduced in the fat body/hemocytes—the main immune tissues in systemic immunity—by using c564-GAL4 driver (Fig. 1B) (13). In contrast, the transgenic flies carrying the UAS-dNox-RNAi, as well as NF-κB pathway mutant flies, were completely unaffected by gut infection (Fig. 1B and fig. S4). The dDuox-dependent ROS did not affect the NF-κB–dependent antimicrobial peptide gene expression in the gut (fig. S5), and conversely, the NF-κB pathway was not involved in the basal and infection-induced expression of intestinal dDuox (fig. S6). The infection-inducible nature of the intestinal dDuox also suggests that the transcriptional control of dDuox may play an important role in the regulation of the intestinal ROS generation. Taken together, these observations demonstrate that intestinal dDuox plays a major role in host resistance during natural infection. In the intestinal membrane fraction of the dDuox-RNAi flies, measurable superoxide-producing activity was 30% of that observed in the control flies (Fig. 1C), indicating that dDuox provides the main source of ROS within the intestines. In addition, the basal and infection-induced levels of in vivo ROS measured from the dissected intestines of the dDuox-RNAi flies were also significantly lower than those observed in the control intestines (Fig. 1D) (13). In a control experiment, immune-regulated catalase (IRC)RNAi flies showed significantly increased basal and infection-induced ROS levels (Fig. 1D), which can be explained by a diminution of infection-induced ROS-removal capacity (12). Double knock-down flies carrying both dDuox-RNAi and IRC-RNAi exhibited normal levels of ROS (both basal and infection-induced ROS) (Fig. 1D) and displayed normal survival rates during natural infection (Fig. 1E). This phenomenon most likely reflects the fact that dDuox-RNAi used in the study resulted in a partial loss of function (fig. S2). The residual ROS level of dDuox-RNAi flies (Fig. 1, C and D) was apparently augmented to the normal control level by IRC-RNAi (Fig. 1D). Thus, the suppression of the infection-induced mortality of dDuox-RNAi flies by IRC-RNAi can be attributed to a reciprocal compensatory effect of the two opposing knock-down phenotypes. In theory, if we used adult flies carrying the dDuox null allele that show a total absence of a basal level of ROS, reducing the expression of IRC with IRC-RNAi may not protect the dDuox null mutant flies from infection. However, we could not reduce dDuox expression further using more effective dDuox-RNAi lines to confirm this possibility because of their larval/pupal lethality even in the presence of IRC-RNAi. At present, we cannot rule out the possible existence of other minor ROS-generating enzyme(s) in the adult gut. However, because dDuox is the main enzyme system involved in de novo synthesis of infection-induced epithelial ROS during gut infection, the redox balance between “dDuox-dependent ROS generation” and “IRC-dependent ROS removal” appears to represent a principal determinant of host survival during natural infection.

Fig. 1.

Duox is responsible for infection-induced ROS generation, which is indispensable for host survival during natural gut infection. The genotypes of the flies used are indicated in the supporting text. All results are expressed as the mean ± SD of three different experiments. (A) Superoxide-generating activity in the intestinal membrane fraction. The superoxide production level in the intestine in the absence of calcium was taken arbitrarily to be 100. (B) Survival rates of dDuox-RNAi and dNox-RNAi flies were assessed after natural Ecc15 infection. (C) Reduced in vitro intestinal superoxide-generating activity of dDuox-RNAi flies. (D) Reduced in vivo ROS of dDuox-RNAi flies. The total intestinal ROS levels were quantified with flies after natural Ecc15 infection (13). The ROS level in the uninfected control intestine was taken arbitrarily to be 100. (E) The survival rates of the flies carrying both UAS-dDuox-RNAi and UAS-IRC-RNAi after natural Ecc15 infection.

To investigate the in vivo relation between Duox activity and microbial persistence, we next examined the intestines of dDuox-RNAi flies during natural infection, using a green fluorescence protein (GFP)–tagged pathogen (13). Notably, unlike normal flies, ingested bacteria were shown to persist and proliferate throughout the intestinal tracts of the dDuox-RNAi flies [∼300 times as many colony-forming units (CFUs) as those of the control intestines at 60 hours after infection] (Fig. 2, A and B). These results demonstrate that dDuox plays an important role in limiting the extent of microbial proliferation within the gut.

Fig. 2.

Duox is responsible for limiting the onset of microbial proliferation in the gut. The genotypes of the flies used are indicated in the supporting text. All results are expressed as the mean ± SD of three different experiments. (A) Bacterial persistence in the guts of dDuox-RNAi flies. Bacterial persistence was measured using spectinomycin-resistant Ecc15-GFP (13). Time-course analysis (left) and representative plates of Ecc15-GFP recovered (60 hours after infection) from the intestines (right). (B) Incomplete Ecc15-GFP clearance in the dDuox-RNAi flies. Representative image of naturally infected flies and the dissected guts (60 hours after infection). (C) The peroxidase activity of the recombinant PHD of dDuox. SDS–polyacrylamide gel electrophoresis and Western blot analyses of the His/V5-tagged recombinant protein purified from the culture medium of S2 cells stably expressing the recombinant PHD (lane 1) or from the culture medium of control S2 cells (lane 2). The purified recombinant PHD proteins were then subjected to in vitro peroxidase activity (13). Human MPO was used as a positive control. The peroxidase inhibitors, aminotriazol (ATZ) and azide, were used (10 mM). The values were expressed as relative peroxidase activity, with the activity of the PHD (4 μg) arbitrarily set to 100. (D) Microbicidal activity of PHD. The chloride-dependent microbicidal activity was performed in the presence of H2O2 (13). The values were expressed as relative CFUs, with the number of CFU in the untreated bacteria arbitrarily set to 100%.

In mammals, the gp91phox/Nox2-like oxidase domain of Duox has been reported to generate superoxide that is spontaneously dismutated to H2O2 after being secreted to the extracellular compartment (14, 18). However, it has also been shown that H2O2 is only microbicidal at high concentrations and that exogenously generated superoxide does not directly kill microbes (3, 21). Therefore, a variety of secondary oxidants have been proposed to account for the destructive capacity of phagocytes in mammals (3, 6). One such secondary oxidant is the highly microbicidal HOCl, which is generated from H2O2 by neutrophil-derived myeloperoxidase (MPO) in the presence of chloride (3, 4, 6). Given that the PHD of dDuox exhibited significant in vitro peroxidase activity (Fig. 2C) (13), we tested whether H2O2 can also be used by the PHD to generate secondary oxidant of higher antimicrobial activity (2, 6). The chloride-dependent microbicidal activity assay (13) showed that the recombinant PHD displays a significantly amplified antimicrobial effect, but only in the presence of both chloride and H2O2 (Fig. 2D). These data demonstrate that the PHD exerts an MPO-like activity, which induces bacterial death in a chloride-dependent manner. Furthermore, we found that the levels of oxidative damage of ingested bacteria as assessed on the basis of protein carbonylation and lipid peroxidation (13) were significantly reduced in the intestines of the dDuox-RNAi flies as compared with those seen in the bacteria from control intestines (fig. S7). These results demonstrate that the infection-induced ROS generation by dDuox is responsible for the direct oxidative damage inflicted on ingested microbes.

To demonstrate that the immune susceptibility of dDuox-RNAi flies could indeed be attributed to the insufficient enzymatic activity of Duox, we attempted to protect dDuox-RNAi flies by reintroducing a variety of Duox enzymes [human Duox (hDuox) 1-2, dDuox, and dDuox-ΔPHD] (Fig. 3A). The reintroduction of both hDuox and dDuox, but not that of dDuox-DPHD, markedly augmented the survival rates of the dDuox-RNAi flies after natural infection (Fig. 3B). These results are consistent with the previous observations (Fig. 2D), indicating that PHD is required for the microbicidal effects of Duox. The reduced levels of in vitro superoxide-producing activities and of in vivo intestinal ROS in the dDuox-RNAi flies were almost completely restored to normal levels by reintroducing the dDuox (Fig. 3, C and D). Consistent with this, we detected that microbial persistence within the intestines of the dDuox-RNAi flies was reduced to control levels upon reintroduction of dDuox, but not upon dDuox-DPHD expression (Fig. 3E). Taken together, our results demonstrated that intestinal dDuox is responsible for the generation of infection-induced microbicidal ROS and that ROS thus generated are required for limiting the proliferation of local pathogens during gut-microbe interactions.

Fig. 3.

The immune susceptibility of dDuox-RNAi flies can be markedly ameliorated by the reintroduction of either dDuox or hDuox. The genotypes of the flies used are indicated in the supporting text. All results are expressed as the mean ± SD of three different experiments. (A) Schematic presentation of various Duox constructs used for the generation of transgenic flies. (B) Rescue experiment. The dDuox-RNAi flies were crossed with flies carrying a variety of Duox constructs to determine the survival rates after natural Ecc15 infection. (C and D) Reintroduction of dDuox into the dDuox-RNAi flies resulted in the complete restoration of in vitro intestinal superoxide-generating activity (C) and total in vivo intestinal ROS levels (D). The ROS levels in the uninfected control intestines were arbitrarily set at 100. (E) Persistence of Ecc15-GFP in the guts of dDuox-RNAi flies is completely abolished by reintroduction of dDuox, but not upon dDuox-DPHD expression. Representative images of naturally infected flies (top) and dissected intestines (middle), and representative plates of Ecc15-GFP recovered from the intestines (bottom), at 60 hours after infection.

ROS perform a variety of functions in many biological events, including host defense, development, hormone biosynthesis, fertilization, and diverse intracellular signaling (25, 14, 15, 2228). In the present study, we have demonstrated the in vivo role of Duox in innate immunity via mediating epithelial oxidative burst in Drosophila gut. Our study broadens the concept of ROS-based immunity by demonstrating that the oxidant-dependent defense system is not restricted to the phagocytes but rather is found in barrier epithelia. In addition to the NF-κB pathway–mediated defense system (711), the Duox-mediated ROS-dependent defense system involving both gp91phox-like activity and MPO-like activity constitutes another microbicidal arm of Drosophila innate immunity. Further delineation of dDuox/IRC-mediated gastrointestinal redox homeostasis will provide important insight into innate immunity and the host-pathogen interaction.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S7


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article