A hallmark of the systemic antimicrobial response of Drosophila is the synthesis by the fat body of several antimicrobial peptides which are released into the hemolymph in response to a septic injury. One of these peptides, drosomycin, is active primarily against fungi. Using a drosomycin–green fluorescent protein (GFP) reporter gene, we now show that in addition to the fat body, a variety of epithelial tissues that are in direct contact with the external environment, including those of the respiratory, digestive and reproductive tracts, can express the antifungal peptide, suggesting a local response to infections affecting these barrier tissues. As is the case for vertebrate epithelia, insect epithelia appear to be more than passive physical barriers and are likely to constitute an active component of innate immunity. We also show that, in contrast to the systemic antifungal response, this local immune response is independent of the Toll pathway.
In past years, several advances have highlighted the primordial role of innate immunity in providing a quick effector response to infections in vertebrates (Janeway, 1989; Fearon, 1997). Furthermore, natural immunity appears not only to trigger the adaptive immune response but also to direct the type of effector response in clonally selected immune cells that is appropriate to fight efficiently against infections (Medzhitov and Janeway, 1997). However, it has become clear that a first line of defense of the organism consists of the local synthesis and release of antimicrobial peptides in tissues, also called barrier epithelia, which are in direct contact with microorganisms. The role of these antimicrobial peptides is illustrated dramatically in the case of cystic fibrosis patients, where β‐defensin expressed in the conducting and respiratory airway is inactivated by the high salt concentrations in the airway surface fluid due to a defect in the cystic fibrosis transmembrane conductance regulator. As a result, the airway may become colonized by microbial pathogens, and chronic inflammation ensues (Smith et al., 1996; Goldman et al., 1997).
In Drosophila, a septic wound induces the rapid appearance in the hemolymph of a battery of antibacterial peptides that includes the cecropins (Kylsten et al., 1990; Tryselius et al., 1992), diptericin (Wicker et al., 1990), drosocin (Bulet et al., 1993), insect defensin (Dimarcq et al., 1994), metchnikowin (Levashina et al., 1995), attacin (Asling et al., 1995) and one major antifungal peptide, drosomycin (Fehlbaum et al., 1995). These peptides are synthesized mostly in the fat body, a functional equivalent of the liver, and secreted into the hemolymph. We will refer to this reaction as the systemic antimicrobial response.
Since experimental wounds are restricted to a single point of entry and since all of the disseminated fat body is responding to the attack, it is thought that a signal is transmitted through the hemolymph from the entry site of microorganisms, where non‐self recognition presumably occurs, to the fat body. It is generally appreciated today that the insect host defense presents remarkable similarities to vertebrate innate immunity, pointing to a certain degree of common ancestry (reviewed in Hoffmann et al., 1996; Hoffmann and Reichhart, 1997). In this context, we have asked whether antimicrobial peptides are also expressed in barrier epithelia in Drosophila. In this study, we address this question for the antifungal peptide drosomycin in larvae and adults of this species.
To monitor the expression of drosomycin in live flies, we designed a transgenic reporter system based on the green fluorescent protein (GFP). This protein fluoresces without specific cofactors when illuminated at the right wavelengths, and can be detected in live animals provided they are reasonably transparent (Chalfie et al., 1994; Wang and Hazelrigg, 1994). Further, the fluorescence can be quantified using fluorimeters or imaging devices. Since there is no enzymatic amplification step, GFP is less sensitive as a reporter gene than the bacterial β‐galactosidase gene that has been broadly used so far. However, new generations of mutant GFPs that are brighter may compensate for these shortcomings (Heim et al., 1995; Cormack et al., 1996). Indeed, we show here that S65TGFP is an adequate reporter of inducible gene expression in Drosophila.
Using drosomycin–GFP reporter transgenes, we find in some non‐experimentally immunized animals an expression of the drosomycin–GFP reporter gene in a variety of epithelial tissues. These observations suggest the existence of a local immune response in Drosophila in addition to the classically described systemic response induced by wounding. Thus, like vertebrate epithelia, insect epithelia may be more than passive physical barriers and are likely to constitute an active component of innate immunity.
The drosomycin–GFP reporter gene is expressed in the fat body and hemocytes during the systemic immune response
We have constructed several drosomycin–GFP reporter genes using either the wild‐type or the S65T versions of GFP (Materials and methods). The RNA and protein levels of the reporter constructs induced after immunization are similar to those of the endogenous drosomycin as judged by Northern and Western blot analysis (data not shown). Furthermore, the comparison of the kinetics of protein accumulation as monitored by Western blot analysis with that of its fluorescence monitored quantitatively by fluorometry revealed that S65TGFP becomes fluorescent shortly (1–2 h) after its synthesis in the fat body of the adult (M.‐C.Criqui, A.C.Jung and D.Ferrandon, unpublished data). Like the endogenous drosomycin, S65TGFP accumulates progressively during the systemic response and reaches a maximum 3 days after induction.
To monitor qualitatively the expression of the drosomycin–GFP reporter gene during the systemic immune response, we looked at adult transgenic flies using a dissecting microscope equipped with an epifluorescence illumination module and the relevant set of filters. Through the cuticle, a diffuse green fluorescence was observed all over the immunized animals, whereas control animals displayed only a weak autofluorescence (Figure 1A). Upon dissection of the animal, it appeared that most of this fluorescence was due to the expression of the GFP reporter gene in the disseminated fat body of the adult (Figure 1C). No qualitative differences were noted between different lines carrying independent insertions of the transgenes.
The S65TGFP reporter gene could first be detected 4–6 h after induction; the signal could usually be detected reliably in whole flies after 8 h with a dissecting scope. However, in keeping with the quantitative results described above, the fluorescence was much stronger after 24 h.
In the transparent third instar larva, GFP fluorescence was detected mostly uniformly in all lobes of the fat body of immunized animals starting 4–6 h after injury, whereas usually no fluorescence could be observed in control animals (Figure 1B and D). These results agree with those obtained by in situ hybridization with a drosomycin probe (Fehlbaum et al., 1995).
We also found reporter GFP expression in a subset of larval plasmatocytes and lamellocytes after a septic wound.
Expression of the drosomycin–GFP reporter gene in the larval tracheal system in the absence of experimental immune challenge
While looking at a large number of control non‐induced larvae, we noticed that a few larvae displayed an apparently spontaneous expression of the reporter gene. The proportion of drosomycin–GFP‐expressing animals was quite variable from one culture vial to the next and was probably dependent on culture conditions such as crowding, age of the culture and/or the microbial environment of the culture medium. Of these larvae, a minor percentage showed an expression in the fat body which reflected a systemic immune response due probably to natural infections. Interestingly, the reporter gene was also detected in different tissues of other larvae, especially in the tracheal epithelia. The most frequent expression pattern was observed in the anterior or posterior spiracles. This expression was often limited to one spiracle of the pair (Figure 2), suggesting a limited response to a local infection. GFP was detected in the epithelium of the tracheal trunk next to the spiracle, but not in the spiracular chamber itself. The extent of this expression was sometimes limited to a ring. In other insects, however, this spiracular expression extended to most of the tracheal dorsal trunk and transverse connectives, indicating that an aerial infection had progressed along this trachea. Rarely, some larvae expressed the reporter gene in the central but not the distal parts of the dorsal trunk. Thus, it was not always possible to correlate an expression in the tracheae with a spiracular expression. Furthermore, some insects displayed a GFP fluorescence only in the transverse connectives, lateral trunk and other smaller tracheae, but not in the dorsal trunk (Figure 2E). The expression of the reporter gene in the tracheal system was almost totally abolished in larvae reared under axenic conditions.
To check that the GFP expression we detect in the tracheal system corresponds to the actual synthesis of the endogenous drosomycin peptide in this tissue, we dissected fluorescent and non‐fluorescent parts of tracheal trunks. Strikingly, we could detect a peak corresponding to the drosomycin peptide in the fluorescent tracheae and not in the non‐fluorescent ones by MALDI‐TOF mass spectrometry on isolated tracheal trunks, thus demonstrating the relevance of the GFP reporter to epithelial tissue expression of drosomycin (Figure 3).
Induction of the expression of the drosomycin–GFP reporter gene following an exposure to microbial agents
To show that the localized expression patterns of the drosomycin–GFP reporter gene in the respiratory system are indeed a local response to a microbial infection, we dipped early third instar larvae in concentrated solutions of various bacteria or fungal spores for 30 min. The larvae were then allowed to recover from the treatment for 5–7 days at 18°C in a normal fly vial. In a typical experiment, 80% of the larvae that had been treated with the Gram‐negative bacteria Erwinia carotovora carotovora developed a strong expression of the reporter gene throughout the tracheal system. In contrast, only 10% of the control animals that had been dipped in water displayed an expression of the transgene that was limited to the spiracles. In a few experiments, the proportion of control animals reacting to water treatment was somewhat higher than 10%, yet transgene expression was limited to the spiracles and did not propagate to the whole tracheal system. These experiments show that drosomycin can be synthesized in the tracheal system in response to a microbial infection.
Other expression patterns of the drosomycin–GFP reporter gene in larvae and adults in the absence of experimental immune challenge
A few larvae and pre‐pupae were found occasionally to express the reporter gene in all cells of the salivary glands (Figure 2F and G). In contrast, all transgenic adults carrying the drosomycin–GFP reporter displayed some fluorescence in the terminal coiled regions of the salivary glands (Figure 4E). The intensity of fluorescence was variable from one fly to the next. We may assume that this expression is weakly constitutive and can be induced strongly. Often, two paired fluorescent patches were seen inside the proboscis; they most likely correspond to expression of the drosomycin–GFP reporter gene in the two small labellar glands whose secretory ducts open at the beginning of the alimentary canal (Figure 4C). Very few individuals displayed expression of the reporter gene in the pseudotracheae of the proboscis (Figure 4D). The salivary and labellar gland expression was sometimes observed already in late pupal stages.
The above results raise the possibility that drosomycin is secreted into the alimentary canal by two sets of secretory organs, the salivary and labellar glands.
Some adults exhibited other expression patterns in the head. For instance, fluorescence was detected fairly often in the maxillary palps. In some cases, it was clear that expression was taking place in the epithelial cells (Figure 4F), whereas in others, the fluorescence appeared more diffuse and could possibly be due to drosomycin–GFP reporter expression in the small maxillary palp trachea. Similar observations were made in the antennal segments.
As in larvae, expression of the drosomycin–GFP reporter gene was sometimes detected in the adult respiratory system, usually of older insects (Figure 4G). Fluorescence was observed in the spiracles, abdominal tracheae, leg tracheae, the head air sacs and tracheae, and probably in the thoracic air sacs. However, the expression in most cases did not extend to the whole respiratory system, but only portions thereof, suggesting a local immune response in the trachea following an aerial infection.
Occasionally, drosomycin–GFP could be detected in the ejaculatory duct, and more rarely in the ejaculatory bulb of some males (Figure 4B). Depending on the fly, the expression in the ejaculatory duct extended more or less anteriorly. These observations suggest that infections can propagate along the genital ducts and induce a local synthesis of drosomycin to fight off this infection.
In a few adults, several rows of three cells beneath the dorsal abdominal cuticula were found to express the drosomycin–GFP reporter gene (Figure 4H). This expression pattern also included two thin continuous rows of epidermal cells that partially surround the dorsal part of the eyes, as well as two paired groups of abdominal cells that lie on the ventral side at the frontier with the thorax. To our knowledge, these groups of cells have not been identified so far.
All the expression patterns recapitulated in Table I were never induced by wounding experiments that trigger the systemic response.
The drosomycin–GFP transgene reveals a constitutive expression of drosomycin in the female sperm storage organs
All adult females carrying the drosomycin–GFP reporter gene in the different transgenic lines displayed a bright source of fluorescence that could be seen through the cuticle of the posterior part of the abdomen. Upon dissection, the fluorescence was found in the female sperm storage structures: the two spermathecae and the seminal receptacle (Figure 4A). In keeping with this expression, endogenous drosomycin RNA was also detected by in situ hybridization in these organs (L.Michaut and R.Lanot, unpublished observations). This signal is not dependent on copulation since it was also found in virgin females. Further, it was present in females grown under axenic conditions. We conclude that this expression is constitutive.
Genetic analysis: a pathway other than Toll?
The systemic antifungal response in the adult has been shown to be under the control of the spätzle/Toll/tube/pelle/cactus pathway (Lemaitre et al., 1996). The inducible expression of drosomycin is abolished in spätzle (spz), recessive lack‐of‐function Toll (Tl), tube (tub) and pelle (pll) mutants. Drosomycin is constitutively expressed in cactus (cact) and Tl dominant gain‐of‐function mutants (Lemaitre et al., 1996). Most of this gene cassette, which also establishes the primary dorso‐ventral pattern (reviewed in Morisato and Anderson, 1995), has been conserved during evolution and appears to control the activation of the NF‐κB transcription factor in the vertebrate inflammatory responses (Ingham, 1994; Kopp and Ghosh, 1995; Verma et al., 1995; Baeuerle and Baltimore, 1996). When this pathway is altered, flies succumb to a fungal, but not to a bacterial infection (Lemaitre et al., 1996). We therefore investigated whether the local expression of drosomycin that is detected in several tissues of ectodermal origin is also controlled by this pathway.
We first looked at insects heterozygous for the drosomycin–GFP reporter transgene and a dominant mutation of Tl, Tl10B, that results in a constitutively activated TL receptor (Figure 5A–D). Consequently, the drosomycin gene was transcribed strongly in the absence of an immune challenge. In Tl10B larvae, we could detect, as expected, a strong GFP expression in all lobes of the fat body; furthermore, some but not all plasmatocytes and lamellocytes were also positive. The intensity of the fluorescence was considerably stronger than in wild‐type larvae 12 h after immune challenge, which is likely to result from the continuous accumulation of the stable reporter protein in these tissues since embryogenesis. Strikingly, no constitutive expression could be detected in the tracheal system nor in the salivary glands. However, as in the case of the wild‐type, expression of the reporter gene in the tracheae or spiracles could be observed in some larvae, probably in response to a local infection.
In young Tl10B adults, a strong fluorescence was at first limited to the abdomen. However, this signal disappeared within 2 days, whereas a weaker fluorescence appeared everywhere in the animal. Upon dissection, it turned out that the cells responsible for the intense fluorescence in the abdomen were larval fat body cells that had not yet lysed. The adult fat body replaces the larval one during the first 2 days of adult life. As in larvae, no constitutive expression of the drosomycin–GFP reporter was detected in tissues of Tl10B adults where a local expression is sometimes seen in wild‐type adults. These results indicate that in both the larva and the adult, the activation of the Tl pathway is not sufficient to trigger expression of drosomycin in these ectodermal tissues.
Next, we determined whether the Tl pathway is necessary for the local reporter gene expression or the constitutive expression in the female reproductive tract. To this end, we looked at the expression pattern of the reporter gene in Tl and spz lack‐of‐function mutants where the systemic antifungal response is impaired (Figure 5E–I). The spermathecae and seminal receptacles of Tl and spz females that also carry one or two copies of the GFP reporter transgene constitutively displayed a bright fluorescence. Local expressions in the trachea, salivary and labellar gland, ejaculatory duct, labial palps and pseudotracheae of the proboscis were detected in some Tl or spz adults, as in wild‐type. Likewise, expression of the transgene in a spz background was also observed in the tracheae of homozygous larvae and in the salivary glands of early pupae. Thus, the local expression of drosomycin–GFP was not eliminated in these mutants, whereas the systemic immune response in the fat body was abolished.
Here we report the expression pattern in live Drosophila of the major antifungal peptide drosomycin using several GFP reporter transgenes. The GFP reporter gene expressed under the control of the drosomycin promoter appears to reproduce faithfully the transcription pattern of the endogenous drosomycin gene, both quantitatively and qualitatively. Further, since the endogenous and the reporter protein accumulate at roughly the same rate, and since both proteins are resistant to most proteases, it appears likely that the GFP fluorescence accurately mirrors the actual synthesis of the endogenous drosomycin peptide (see Figure 3 for instance).
A local immune response in the respiratory system of Drosophila
In Drosophila, a septic injury induces the rapid synthesis in the fat body of a cocktail of antimicrobial peptides that are secreted in the hemolymph where they neutralize invading microorganisms (Hoffmann and Reichhart, 1997). This reaction constitutes the systemic response to infection. Our data suggest strongly that Drosophila is also able to fight off infections at a local level in epithelia that are in contact with the external environment without mobilizing the systemic response. The expression of the drosomycin–GFP reporter gene in the tracheal system clearly illustrates this phenomenon. First, in most cases, the expression pattern is asymmetrical, both in adults and in larvae, being for instance limited to one of the tracheal trunks of the larva. Furthermore, the GFP signal is detected in the transverse connectives that are connected to the GFP‐expressing parts of the dorsal trunk, and not in those connected to non‐expressing parts (Figure 2). Second, the expression is often confined to the entry points of the tracheal system, i.e. the anterior or posterior spiracle, suggesting that drosomycin is synthesized there as a response to a local infection of the aerial pathway. In this view, expression patterns that extend to the whole tracheal trunk may be the results of infections that have not been stopped by the spiracular synthesis of drosomycin. Third, local expression of the reporter gene in the tracheal system of larvae grown under axenic conditions is almost totally abolished. Finally, it is possible to induce the expression of the drosomycin reporter gene systematically throughout the tracheal system by dipping larvae in concentrated microbial solutions. Thus, drosomycin is expressed in the tracheal system as a local immune response to an infection of the respiratory system. However, we cannot exclude that other stimuli, e.g. wounding, can also trigger a local expression in epithelia.
A local immune response in ‘barrier epithelia’
Strikingly, we detect reporter gene expression in most tissues that are in contact with the external environment, suggesting that drosomycin is expressed there as a local response to infections, as is the case for the respiratory system.
The digestive tract may be a route of entry for infectious microorganisms as they are likely to be abundant in food. We did not detect any expression of drosomycin–GFP in the gut, in contrast to what has been found in the mosquito Anopheles gambiae and in the blood‐sucking fly Stomoxys calcitrans, which express defensins in the anterior midgut (Dimopoulos et al., 1997; Lehane et al., 1997). However, the reporter gene is frequently strongly expressed in the distal part of the salivary gland of the adult, and less frequently in the labellar glands. These observations suggest that drosomycin may be secreted into the food before reabsorption by the fly to prevent infections of the digestive tract.
The great reproductive abilities of insects are central to their evolutionary success. The strong constitutive expression of drosomycin in the female sperm storage structures may reflect an evolutionary constraint to prevent infection of these structures by sperm infected during copulation. Furthermore, other antibiotic peptides have been found in the Drosophila reproductive tract. Drosocin has been observed in the calix and oviducts of fertilized females (Charlet et al., 1996). Andropin is constitutively synthesized by the male ejaculatory duct and may be transferred to the female genital tract (Samakovlis et al., 1991). Indeed, an antibacterial activity has been found in the ejaculatory duct that is transferred to females in the sperm (J.Postlethwait, personal communication). In the medfly Ceratitis capitata, ceratotoxins are constitutively expressed in the accessory gland under the control of juvenile hormone, and their expression is enhanced by mating (Rosetto et al., 1996; Manetti et al., 1997). It is not clear whether the increased expression is due to mating per se or to microbial contamination during copulation. In the Drosophila male, drosomycin expression in the ejaculatory duct is more likely to constitute a local immune response to protect the male genital tract from an ongoing infection, since this expression is not constitutive, is not induced by mating and its extent is variable from one individual to another, most displaying no response however.
We have detected a local expression of drosomycin–GFP in epithelial tissues of the trachea and of the salivary glands, but not in the cuticular epithelium, except for the maxillary palps. Due to the size of the larvae, it appears difficult to reproduce abrasion experiments that have been performed on the silkworm Bombyx mori which have demonstrated a localized synthesis of cecropins in the cuticular epithelium (Brey et al., 1993). Except for one single case, we never detected a local epidermal response at the wounding site. In naturally occurring and post‐challenge systemic responses, we cannot exclude that fluorescence in the fat body would have masked such a response. The results obtained with the silkworm abrasion experiments indicate that a local immune response is not limited to drosomycin or to Drosophila. Preliminary results obtained with the GFP reporter gene system suggest that other antimicrobial peptides also display a local immune response in Drosophila (S.Ohresser, personal communication).
Characteristics of the local immune response
One striking feature of the localized drosomycin response is that it does not involve the systemic response. This implies that epithelial cells carry receptors that are able to recognize non‐self determinants, presumably conserved structural patterns found on the surface of microorganisms (Janeway, 1989). These activated receptors subsequently induce a transduction cascade that results in the local synthesis and secretion of drosomycin by these very same cells. It is interesting in this respect that this pathway does not involve the spz–Tl signaling pathway that is used in the systemic response, a process that may be mediated by putative cytokines in the hemolymph. Alternatively, proteases secreted by microorganisms could trigger the localized response by altering a receptor on the cell surface of these tracheal cells. The case of the up‐regulated adult salivary gland drosomycin–GFP expression may be an exception in that it is unlikely that the distal salivary gland cells directly detect the infection. One may wonder whether the infection might be detected at the level of the maxillary palps cells that would in turn signal to the salivary glands.
The Drosophila systemic antifungal response in adults has been shown to be controlled by the spätzle/Toll/cactus gene cassette (Lemaitre et al., 1996). The constitutive activation of the Tl pathway is not sufficient to trigger expression in the tracheal system or in the salivary glands; the absence of the putative ligand SPZ, or of its receptor TL, does not prevent the localized expression in both the adult and the larva. Therefore, the local host response is regulated by another, as yet undetermined, genetic pathway. Furthermore, the dissection of the drosomycin promoter has revealed that the systemic and local immune responses are controlled by distinct regulatory elements (L.Michaut and E.Levashima, unpublished data). In this respect, it is worth noting that the systemic response takes place in tissues of mesodermal origin (fat body and hemocytes), whereas the local expression which we detect is found in epithelia of ectodermal origin.
The local host response: a common strategy in the animal kingdom?
Similarities between the insect systemic immune response and the mammalian acute phase response of inflammation have been underscored. Here, we report a localized immune response in barrier tissues that displays striking parallels with vertebrate mucosal immunity. Indeed, several antimicrobial peptides, α‐ and β‐defensins, have been found in different mammalian epithelia, including the lingual epithelium (LAP), the aerial tract mucosa (TAP, β‐defensin 1), the submaxillary salivary glands, kidney, prostate, testis, female genitourinary tract (β‐defensin 1) and Paneth cells of the intestine (α‐defensins) (Diamond et al., 1993, 1996; Schonwetter et al., 1995; Ouelette and Selsted, 1996; Russell et al., 1996; Zhao et al., 1996). A recent study reported widespread expression of bovine β‐defensins in a variety of epithelial tissues. Interestingly, cows suffering from a chronic gut infection with Mycobacterium paratuberculosis displayed an intense expression in the ileal mucosal region. Similarly, β‐defensins were markedly up‐regulated in the lung after an experimental respiratory infection with Pasteurella haemolytica (Stolzenberg et al., 1997). Thus, vertebrate defensins are thought to be important components of the innate immune response in mucosal barriers; the role of β‐defensins in the immune defense of the respiratory airway is illustrated dramatically in cystic fibrosis patients (Smith et al., 1996; Goldman et al., 1997). Strikingly, drosomycin is expressed in analogous tissues, among them the tracheal epithelia, salivary glands and genital tract. For lack of adequate mutants, we cannot yet demonstrate that the local drosomycin expression in these tissues is necessary for a local host defense. However, this is very likely since mutants in which the systemic antifungal response is abolished are more susceptible to fungal infections (Lemaitre et al., 1996).
Epithelia that are in contact with microorganisms constitute both a physical and an immune barrier. The secretion of antimicrobial peptides locally helps to prevent the spreading of infections and thus provides a first line of defense, the local immune response, which does not require the triggering of a systemic immune response. This strategy is energetically inexpensive for dealing with frequent infections propagated by contact or by minor wounds. Indeed, during the systemic response, drosomycin is found in concentrations as high as 100 μM that are likely to exert an important metabolic toll. Similarly, during the mammalian hepatic acute phase response, the major acute phase reactants such as C‐reactive protein or serum amyloid A protein can be induced up to 1000‐fold over normal levels to reach a 1 mg/ml concentration (reviewed in Steel and Whitehead, 1994). Indeed, it has been shown recently that the artificial selection in Drosophila melanogaster of improved resistance against the parasitoid Asobora tabida through the cellular arm of immunity is correlated with a reduced larval competitive ability under a high competition environment (Kraaijeveld and Godfray, 1997). These observations indicate that an improved immune resistance comes with a selective cost. Selection is likely to have favored the constitutive expression of antimicrobial peptides in tissues that are continually exposed to infections or that are especially important, e.g. those of the reproductive tract. Future studies will reveal whether local immune responses have originated early in evolution and have been thereafter conserved, or whether insect and mammals have developed a similar strategy independently.
Materials and methods
We constructed two sets of reporter genes carrying GFP: one expressing only GFP, and the other a GFP fusion to the drosomycin pre‐propeptide so that it would be secreted. Using PCR‐directed mutagenesis, we created a BamHI site after the GFP stop codon contained in plasmid pTU#58 (a kind gift of M.Chalfie) (Chalfie et al., 1994). Next, we added at this site 800 bp of drosomycin downstream sequences (drosomycin terminator) contained in a BamHI–XhoI fragment. Using site‐directed mutagenesis, we created an NheI site after the starting ATG (non‐secreted form) or after position 59 of the drosomycin coding sequences (secreted fusion). SalI–NheI fragments containing 2450 bp of the drosomycin promoter (up to the starting ATG or amino acid 59) were fused to the drosomycin–GFP 3′ end sequences. Finally, both cassettes were inserted as SalI–XhoI fragments at the XhoI site of P‐element‐mediated transformation vector pCasPer NNSXS, a modified version of pCasPer (Ferrandon, 1994), yielding plasmids pJM608 [drosomycin promoter–GFP–drosomycin downstream sequences (non‐secreted)] and pJM612 [drosomycin promoter–59 amino acids of pre‐prodrosomycin–GFP–drosomycin downstream sequences (the secreted fusion protein contains 33 amino acids of the drosomycin peptide)]. It turned out that the product of the wild‐type GFP transgene could be detected by its fluorescence only 2 days after induction. This is most likely due to a long delay in the formation of the fluorophore in the fat body since the protein itself is synthesized much earlier (M.‐C.Criqui, A.C.Jung and D.Ferrandon, unpublished observations). We therefore tested the S65T mutant of GFP.
To reconstitute the GFP S65T version, we mutated the GFP gene contained in plasmid pJM703 using PCR‐directed mutagenesis (Higuchi et al., 1988), with the following primers: IMU 161 TGTCACTACTTTC ACT TATGGTGTTCA and IMU 162 TGAACACCATAAG TGA AA GTAGTGACA. This mutated PCR fragment was cut with NcoI and NdeI and inserted in an NcoI–NdeI‐cut pJM704 plasmid, yielding plasmid pJM705 in which the GFPS65T gene is included in a pBluescript KS+ vector. A partial GFP S65T NheI–HpaI fragment was excised from pJM705 and cloned either into an NheI–HpaI‐cut pJM608 or an NheI–HpaI‐cut pJM612 transformation vector, yielding respectively pJM802 (non‐secreted version) and pJM804 (N‐terminal end drosomycin–GFP fusion).
All fly stocks are described in Lindsley and Zimm (1992). w− flies were used as wild‐type flies and were also used as recipients for transformation. Transgenic lines were established as described (Driever et al., 1990). We obtained at least nine independent lines for construct pJM802, and 16 for pJM804, with transposons located on all three major chromosomes. Northern blot analysis showed that both the endogenous and the reporter genes are expressed at similar levels with the same kinetics of accumulation (data not shown). To obtain Tl mutant flies, we crossed Tl632 (thermosensitive mutation) and Tl9QRE, both of them in a homozygous transgenic reporter (802 or 804) background at 20°C, and we analyzed the phenotypes of the transheterozygous flies at 29°C. spz alleles were spzrm7 and spz097. The systemic antifungal response is abolished in these mutant backgrounds (Lemaitre et al., 1996).
Live flies and larvae were anesthetized (adults with ether; larvae on ice) and viewed under epifluorescent illumination (excitation filter 480/40 nm; dichroic filter 505 nm LP; emission filter 510 nmLP) with a Leica MZ12 dissecting scope. Dissections were performed in phosphate‐buffered saline (PBS) under the dissecting scope and were viewed using either the dissecting microscope or a Leitz Diaplan epifluorescence microscope using a fluorescein isothiocyanate (FITC) filter set. Photographs were taken on a 400 ASA Fujicolor film. They were numerized, enhanced if necessary and assembled using Photoshop 3.0 (Adobe). We are confident that the observed expression patterns reflect the actual synthesis patterns of the endogenous drosomycin since similar expression patterns have been observed by in situ hybridization with a drosomycin probe (systemic response, larval salivary glands, sperm storage structures). See also Figure 3.
Matrix‐assisted laser desorption/ionization time‐of‐flight mass spectrometry (MALDI‐TOF MS)
Instrumentation. MALDI mass spectra were acquired on a Bruker (Bremen, Germany) BIFLEX™ matrix‐assisted laser desorption time‐of‐flight mass spectrometer equipped with SCOUT™ High Resolution Optics, an X–Y multisample probe, a gridless reflector and the HIMAS™ linear detector. Ionization was accomplished with the 337 nm beam from a nitrogen laser. All spectra were obtained in the linear positive ion mode and externally calibrated with a mixture of three standard peptides (synthetic drosocin without sugars, synthetic metchnikowin and recombinant drosomycin with MH+ at m/z 2199.5, 3046.4 and 4890.5, respectively). Once the conditions to detect drosomycin in the trachea were found (higher laser energy than for external calibration), a recombinant drosomycin sample was analyzed again under these conditions and found to display an m/z between 4900 and 4910.
Sample preparation. From a 1/1 mixture of α‐cyano‐4‐hydroxycinnamic acid (4HCCA, 40 mg/ml in acetone, Sigma) and nitrocellulose (NC, 40 mg/ml in acetone, Millipore) half diluted with isopropanol, 0.5 μl was deposited on the probe tip and allowed to air dry. The drosomycin–GFP− or –GFP+ tracheae were dissected under the dissecting microscope with epifluorescence illumination, rinsed in PBS and then directly loaded onto this NC/4HCCA bed and covered by 0.5 μl of a second matrix solution, which is 4HCCA at 7 mg/ml in H2O, 0.1% trifluoroacetic acid (TFA)/acetonitrile (1/1, v/v). After air drying, this preparation was washed with 1 μl of 0.1% TFA.
We thank John Postlethwaith for sharing unpublished data, M.Chalfie for the GFP plasmid, A.Van Dorsselaer for assistance with MALDI‐TOF spectrometry, C.Haas for expert technical assistance, René Lanot for help with dissections and in situ hybridization, and colleagues for stimulating discussions. This work has been supported by institutional funds from the CNRS, the Université Louis Pasteur, grants from the Association Française de Lutte contre la Mucoviscidose, the Human Frontiers in Science program (Phylogenetic Perspective of the Innate Immune Response) and Rhone Poulenc Rohrer. D.F. wishes to acknowledge the support of the Fondation pour la Recherche Médicale (aide à l'implantation de nouvelles équipes).
- Copyright © 1998 European Molecular Biology Organization