|Home | About | Journals | Submit | Contact Us | Français|
Peptide hormones governing many developmental processes are generated via endoproteolysis of inactive precursor molecules by a family of subtilisin-like proprotein convertases (SPCs). We previously identified mutations in the Drosophila amontillado (amon) gene, a homolog of the vertebrate neuroendocrine-specific Prohormone Convertase 2 (PC2) gene, and showed that amon is required during embryogenesis, early larval development, and larval molting. Here, we define amon requirements during later developmental stages using a conditional rescue system and find that amon is required during pupal development for head eversion, leg and wing disc extension, and abdominal differentiation. Immuno-localization experiments show that amon protein is expressed in a subset of central nervous system cells but does not co-localize with peptide hormones known to elicit molting behavior, suggesting the involvement of novel regulatory peptides in this process. The amon protein is expressed in neuronal cells that innervate the corpus allatum and corpora cardiaca of the ring gland, an endocrine organ which is the release site for many key hormonal signals. Expression of amon in a subset of these cell types using the GAL4/UAS system in an amon mutant background partially rescues larval molting and growth. Our results show that amon is required for pupal development and identify a subset of neuronal cell types in which amon function is sufficient to rescue developmental progression and growth defects shown by amon mutants. The results are consistent with a model that the amon protein acts to proteolytically process a diverse suite of peptide hormones that coordinate larval and pupal growth and development.
Many peptide hormones and neuropeptides are produced by limited proteolysis of inactive precursor peptides. The release of an active peptide hormone from its precursor molecule typically occurs by proteolytic cleavage after paired basic amino acids and is mediated by a family of proteases known as the subtilisin-like proprotein convertases (SPCs) (Bergeron, 2000). Following convertase processing, most peptides undergo removal of the C-terminal basic residues via a carboxypeptidase (Fricker and Leiter, 1999) and subsequent α-amidation of the C-terminal glycine residue (Kulathila et al., 1999). Processing of precursor proteins may liberate a single or multiple bioactive products (Sossin et al., 1989; Zhou et al., 1999), and a given precursor may be differentially processed in a cell-specific fashion depending on the SPC processing enzymes expressed (Furuta et al., 1997; Rouille et al., 1995). Thus processing of peptide hormone precursors may serve as an important regulatory step to modulate peptide and neuropeptide signaling.
Prohormone Convertase 2 (PC2) is one of seven vertebrate SPCs identified to date and displays a neuroendocrine expression pattern that implicates the enzyme in the activation of peptide hormones and neuropeptides (Muller and Lindberg, 2000). The Drosophila PC2 homolog, amontillado (amon), was identified based on sequence similarity to conserved regions between yeast Kex2, human furin, and human PC2 (Siekhaus and Fuller, 1999). amon is expressed throughout the life cycle of the fly, and amon transcripts localize to the larval central nervous system and regions of the gut (Siekhaus, 1997; Siekhaus and Fuller, 1999), suggesting that amon acts in neuroendocrine tissues. The amon protein has been shown to be an active protease on a KR containing synthetic peptide when expressed in Drosophila S2 cells with the Drosophila 7B2 protein, a homolog of the 7B2 protein that functions in maturation of vertebrate PC2 (Hwang et al., 2000). Isolation and analysis of a series of EMS induced amon mutants showed that amon is required during embryogenesis and early larval development and suggests that the amon protein may act to process peptide hormones that control hatching, larval growth, and larval molting (Rayburn et al., 2003).
The regulation of molting and metamorphosis in insects has its roots in an endocrine axis and relies on a delicate interplay between steroid and peptide hormones (Ewer, 2005; Mesce and Fahrbach, 2002; Nijhout, 1994; Riddiford, 1993; Truman, 1992; Zitnan et al., 2007). This axis includes the brain, the corpus allatum (CA), and the prothoracic gland which act as sources of peptide and steroid hormones, the effectors of molting and metamorphosis. Pulses of the steroid hormone 20-hydroxyecdysone, hereafter referred to as ecdysone, act to initiate larval molting (Riddiford, 1993). One manner in which ecdysone may exert its regulatory effect on molting and metamorphosis is by affecting the expression of genes whose products are part of the ecdysis cascade. For example, in both Manduca and Drosophila the peptide hormone gene ecdysis triggering hormone (eth) contains a putative ecdysone response element upstream of the transcription start site (Park et al., 1999; Zitnan et al., 1999). Studies in Manduca show that rising ecdysteroid levels prior to ecdysis induce eth gene expression in the Inka cells and that a drop in the ecdysteroid titer is required for ETH release (Zitnan et al., 1999).
Two additional peptide hormones, eclosion hormone (EH) and crustacean cardioactive peptide (CCAP) are known to interact with ETH in an ecdysis cascade to elicit the behavioral outputs that characterize molting and metamorphosis in insects. Low levels of circulating ETH trigger the release of EH from the brain. This initial EH release induces a subsequent and exhaustive release of ETH from the Inka cells which in turn cues exhaustive release of EH (McNabb et al., 1997). In Manduca, EH acts through a second messenger system to cause elevated levels of cGMP in cells that express CCAP (Ewer et al., 1997). CCAP release elicits ecdysis motor burst while suppressing the pre-ecdysis behaviors initiated by EH (Gammie and Truman, 1997). Notably, the ETH and CCAP pro-peptides of Drosophila contain putative dibasic processing sites (Park et al., 2003; Park et al., 1999) suggesting they require endoproteolytic activation. The Drosophila EH pro-peptide also contains a possible dibasic processing site (Horodyski et al., 1993), although processing at this site may not be necessary for production of the bioactive EH peptide.
Here we have used a conditional rescue system to ask whether amon is required during postembryonic developmental transitions in Drosophila. amon mutants rescued past earlier embryonic and larval requirements by heat-shock induced expression of an amon cDNA and then removed from the rescue regime at the late third instar stage display defects in head eversion, leg and wing extension, and abdominal differentiation, indicating that amon activity is required for these aspects of metamorphosis. Although amon mutants show similar phenotypes to eth mutants and to pupae resulting from CCAP cell ablation, the amon protein does not co-localize with ETH, CCAP, or EH, suggesting the existence of novel peptide hormones that regulate molting and metamorphosis. Interestingly, the amon protein localizes to cells that innervate the CA and corpora cardiaca (CC) sections of the ring gland, suggesting that amon protein may regulate the endocrine activity of this gland. Finally, expression of amon in a subset of neuronal cell types in an amon mutant background is sufficient to partially rescue developmental progression and growth.
One hundred fifty yw; hs-amon/hs-amon; Df(3R)Tl-X e/TM3 Sb Ser y+ e virgin females were crossed to 150 yw; amonC241Y or amonQ178st e/TM3 Sb Ser y+ e males in an egg collection chamber and allowed to lay eggs on 100 mm grape juice agar plates spotted with yeast paste for three days at 25°C. On the third day, a four or eight hour egg collection was taken and yellow larvae (yw; hs-amon/+; amonC241Y or amonQ178st e/Df(3R)Tl-X e) were collected at 36 hours after egg laying (ael) and placed on a fresh plate. Animals were heat shocked for 30 min at 37°C every 24 hours beginning at 36 hours ael and subsequently scored for survival until all the animals on the positive control plate eclosed. Experimental animals were heat shocked as above until 108 hours ael. After each scoring, larvae were transferred to a fresh plate unless the animal had pupariated. Dead larvae were mounted in polyvinyl lactophenol for examination. Pupae from the experimental plate were removed from the pupal case in water under a Leitz dissecting scope once a majority of the control animals had successfully eclosed. Once out of the pupal case, animals were photographed using a digital camera (Hamamatsu 3CCD) mounted to a Leitz dissecting scope. Control animals failing to eclose were also removed from the pupal case and photographed at later time points. Percent pupariated and eclosed values were calculated by dividing the number of animals that had pupariated or eclosed by the number of animals collected at 36 hours ael.
A peptide corresponding to the final 27 amino acids of the amon protein with three additional amino terminal residues (CKC) was synthesized by the Molecular Genetics Instrumentation Facility (MGIF) at the University of Georgia. The peptide was coupled via M-Maleimidobenzoyl N-Hydroxysuccinimide Ester (MBS, Pierce) to bovine serum albumin as described below and then injected into five mice at the Monoclonal Antibody Facility at the University of Georgia. Following an initial subcutaneous injection with Freund's complete adjuvant (Sigma Chemical), mice were boosted intraperitoneally using Freund's incomplete adjuvant every 21 days and bleeds were taken approximately seven days after each boost. Beginning with the third bleed, serum was tested via Western analysis for its ability to recognize the amon protein. Following the sixth boost one mouse produced serum specific for amon protein and was subsequently boosted a total of 15 times over the course of about one year. Bleeds taken after the seventh boost were used in the immunocytochemistry and Western analyses as described below.
Five milligrams of carrier protein (Bovine Serum Albumin, Sigma) was dissolved in 1 mL of phosphate buffered saline (PBS). 50 μL of MBS in N, N dimethylformamide (DMF) was added drop by drop to 500 μL of the carrier protein solution and the stirred at room temperature for 30 minutes. Excess MBS was removed using Micro Bio-Spin P-30 Tris Chromatography Columns (Bio-Rad) equilibrated with PBS. A 20-fold excess of synthetic peptide was added to the carrier protein-MBS solution and stirred at room temperature for three hours. To make sure that peptide was coupled to the carrier protein, the solution was run on a 10% polyacrylamide gel next to a control solution (carrier protein-MBS solution alone) and stained with Coomassie.
Wandering third instar larvae (Canton S) were dissected, fixed (2 hours in PBS and 175 mM NaCl, pH 7.4, with 4% paraformaldehyde), dehydrated, and rehydrated according to the protocol of (Cao and Brown, 2001). Tissues were washed in TBS (25 mM Tris, 137 mM NaCl, 0.27 mM KCl) containing 0.5% Triton X-100 (TBST) two times for five minutes each and then blocked in TBST+5% goat serum overnight at 4°C. All subsequent steps were performed at 4°C. Primary antiserum was added in TBST+ 2% BSA at a 1/200 dilution and left overnight. Following a wash of at least one hour in TBST+2% BSA, the secondary antibody (anti-mouse IgG conjugated to Alexa Fluor 488, 568 or 594, Molecular Probes, Inc.) was added at a dilution of 1:500-1:2000 in TBST+ 2% BSA and tissues were incubated overnight in the dark. Tissues underwent a final wash in TBST and were then mounted on slides in a 1:1 TBST/glycerol solution. Images were captured on an Olympus BX60 microscope using a JVC digital camera (model KYF70BM) and the AutoMontage (Syncroscopy) software. To obtain amon mutant third instar CNS tissues for use as a control, amon mutant larvae (yw; hs-amon/+; amonQ178st e/Df(3R)Tl-X e) were rescued past early developmental requirements for amon function by periodic expression of a hs-amon transgene for 45 min at 37°C every 24 hours, beginning at 36 hrs AEL until 84 hrs AEL. The CNS was dissected at 108 hrs AEL (24 hrs after the final 84 hr AEL heatshock) from both amon mutants and control siblings (yw; hs-amon/+; Df(3R)Tl-X e or amonQ178st e/TM3 Sb Ser y+ e).
For co-localization of amon and ETH, the same protocol was followed except that after blocking, tissues were incubated with the mouse antiserum directed against the amon peptide for three hours, washed for one to two hours, incubated in the anti-mouse secondary antibody for three hours, and washed overnight. The next day, the same procedure was followed for a rabbit antiserum directed against ETH (Park et al., 2002) used at a 1:500 dilution and a secondary antibody (anti-rabbit IgG conjugated to Alexa Fluor 568, Molecular Probes, Inc.) used at a 1:500 dilution. Alternatively, tissues were incubated in the mouse anti-amon antiserum for three hours, washed for one to two hours, incubated in the rabbit anti-ETH antiserum for three hours, washed for one to two hours, and then incubated in both mouse and rabbit secondary antibodies overnight. The following day, tissues were washed for three hours in TBST, mounted on slides, and visualized.
For co-localization with tissues expressing GFP, the UAS-mCD8-GFP line (Lee and Luo, 1999) was crossed to EH-GAL4 (McNabb et al., 1997), Kurs-21, (Siegmund and Korge, 2001), and CCAP-GAL4 (Park et al., 2003) to obtain progeny with GFP in the EH neurons, the CA-innervating neurons, and the CCAP neurons respectively. Wandering third instar larvae were dissected and then processed as described above for amon protein detection.
Animals were homogenized in cracking buffer (0.125 M Tris-HCl, pH 6.7, 5% β-mercaptoethanol, 2% sodium dodecyl sulfate, 4 M urea) (10 μL per adult fly, 5 μL per third instar larvae) containing a Complete Mini EDTA-free protease inhibitor cocktail tablet (Roche) using motorized hand held pellet pestles (Konte) for 1.5 mL eppendorf tubes. Protein homogenates were boiled for five minutes and centrifuged for 10 minutes at 14,000 × g using a benchtop microfuge. The supernatant was boiled for five minutes in an equal volume of sample buffer (0.125 M Tris-HCl, pH 6.8, 1% sodium dodecyl sulfate, 20% glycerol, and 0.02% bromophenol blue) and then loaded onto a 10% polyacrylamide gel.
Proteins were transferred to a nitrocellulose membrane (Schliecher and Schuell BioScience, Inc) overnight in a BioRad Transfer Cell and the membrane was blocked in blocking solution (10% non-fat dry milk, 30% goat serum in TBS/Tween (20 mM Tris-HC1, pH 7.5, 500 mM NaCl, 0.05% Tween20) overnight at -20°C. Primary antiserum was added at a 1:2000 dilution in blocking solution and incubated at room temperature for two to three hours. After three washes of at least five minutes each, a horseradish peroxidase conjugated secondary antibody (anti-mouse IgG-HRP, Promega) was added at a 1:1500 dilution and incubated for 30 minutes at room temperature. Following three more washes of at least five minutes each, peroxidase activity was detecting using Lumi-Light Western Blotting Substrate (Roche). Excess substrate solution was blotted off the membrane and then images were captured using Kodak BioMax Light film.
For the developmental Western analysis, a four hour Canton-S egg collection was taken at 25°C and animals were aged to the following time points after the end of egg laying (ael): 16 hours, 20 hours, 40 hours, 64 hours, 84 hours, and 96 hours. In our hands these time points correspond to embryogenesis (16 and 20 hour ael), first instar, second instar, early third instar, and mid-third (non wandering) instar, respectively. For the wandering third instar and white pre-pupal time points, animals were collected at the appropriate developmental stage. Animals were homogenized as described above in the following amounts of cracking buffer: 0.25 μL per first or second instar animal, 3 μL per early third instar animal, 5 μL per mid and wandering third instar animal, and 10 μL per white pre-pupa.
For antibody blocking experiments, approximately 100 μg of the C-terminal amon peptide used for injection into mice, or one of two control peptides corresponding to EcR-B1 or EH, was added to the antiserum in 1 mL of blocking solution and then allowed to incubate for 30 minutes at room temperature. The consequences of pre-absorption with amon or control peptides were then tested using a Western blot as an assay.
To determine if amon function in specific neuronal cell types is sufficient to rescue amon mutant defects, approximately 150 yw; uas-amon40L; Tl-X e/TM3 Sb Ser y+ e virgin females and 150 yw; +; Kurs21-gal4, amonC241Y e/TM3 Sb Ser y+ e males were mated in an egg collection chamber containing a grape juice agar plate with fresh yeast paste at 25°C. As a control cross, 150 yw; uas-amon40L; Tl-X e/TM3 Sb Ser y+ e virgin females and 150 yw; +; amonC241Y/TM3 Sb Ser y+ e males were mated in the same way. A 4 hr egg collection on grape juice agar plates with yeast paste was taken 24 hrs later. At 36 hrs ael amon mutant larvae, recognizable by the yellow marker, (yw; uas-amon40L/+; Tl-X e/Kurs21-gal4, amonC241Y e) and wild-type control larvae (yw; uas-amon40L/+; Tl-X e or amonC241Y e/TM3 Sb Ser y+ e) were selected and moved to fresh grape juice agar plates with yeast paste. Larvae were scored every 24 hrs based on spiracle morphology to determine whether they were first, second, or third instar larvae.
amon mutants arrest during embryogenesis and early larval development (Rayburn et al., 2003). To determine the requirements for amon during later development, we have rescued amon mutants by periodic heat induced expression of an amon cDNA. In Table 1, we show that continuous daily expression of amon (see Materials and Methods) starting at 36 hours after egg laying (ael) rescues two amon mutant hemizygotes, amonC241Y and amonQ178st, to adulthood. In contrast, removal of heat-shock treatment at the late third instar stage (108 hours ael) results in a failure to eclose.
As shown in Figure 1C-F, the predominant phenotype of the arrested amonC241Y animals was a failure to evert the head coupled with a failure to extend the legs and wings (91%, Table 2). This phenotype also predominated in arrested amonQ178st animals (33%, Table 2). Interestingly, the failures in head eversion and leg and wing extension did not preclude further development of these structures as evidenced by tanning and bristle formation (Fig. 1C,D), nor did they prevent the differentiation and pigmentation of the eye (data not shown). Because events such as bristle formation on the legs and eye pigmentation typically occur after a successful head eversion (Bainbridge and Bownes, 1981), the phenotypes of the amon mutant pupae indicate an uncoupling of the early event of head eversion and the events characteristic of late stages of pupal development.
A smaller proportion of amon mutants displayed failures in abdominal differentiation (Fig. 1C, D, Table 2). For example, 58% of the amonC241Y animals and 27% of the amonQ178st did not complete abdominal differentiation. The undifferentiated abdomens resembled those of stage P5 (Bainbridge and Bownes, 1981) which are transparent and lack segmentation and bristles. The failure in abdominal differentiation did not, however, prevent the thoracic segment from acquiring pigment and bristles, indicating that these two body segments are governed by distinct developmental controls. An additional phenotype displayed by 15% of the amonC241Y mutants was a failure to displace the pre-pupal gas bubble, an event that typically marks the transition from the P2 to the P4 stage of development (Bainbridge and Bownes, 1981) (Fig. 1E,F, Table 2). Interestingly, all of the pupae failing to displace the mid-body gas bubble developed bristled abdomens, a characteristic of a much later developmental stage (P12i) (Bainbridge and Bownes, 1981). Collectively, our results show that amon is required for successful coordination and completion of pre-pupal and pupal development.
To determine the localization pattern of the amon protein, we developed a mouse polyclonal antibody against the amon protein and used it in Western and immunocytochemistry experiments to evaluate the presence and localization of the protein. On Western blots, the polyclonal antiserum recognizes a 97 kDa band that is present in all developmental stages examined (Fig. 2). The levels of the 97kDa protein increase in abundance over developmental time in accordance with RNA expression data reported earlier for amon (Siekhaus and Fuller, 1999). This pattern of expression is consistent with the requirements we have uncovered for amon during larval and pupal life (Rayburn et al., 2003 and this paper). Recognition of the 97kDa band on Western blots is reproducibly abolished in blocking experiments (see Materials and Methods) in which the amon polyclonal antiserum is pre-incubated with the amon peptide antigen but not abolished when control peptides corresponding to EcR-B1 or EH are used (data not shown). Recognition of a second 66kDa band seen at some developmental stages (Fig. 2) is not blocked by pre-incubation with the amon peptide. The 66kDa protein thus appears to be antigenically unrelated to amon.
In tissue stains (Fig. 3AB), we find that the amon antiserum recognizes cells within the central nervous system (CNS) of wild-type third instar larvae but that no signal is seen in this tissue in amon mutant third instar larvae (see Materials and Methods). The localization pattern in the larval CNS consistently includes paired ventral nerve cord cells as well as a pair of brain cells known as the medial neurosecretory cells (MNCs) (Cao and Brown, 2001; Rulifson et al., 2002) or the insulin producing cells (IPCs) (Rulifson et al., 2002). We report elsewhere that the amon protein co-localizes with DILP2 in these cells (Rhea, Rayburn and Bender, in preparation). The amon antiserum also recognizes cells within the midgut of wild-type third instar larvae. These cells include endocrine cells within thicker portions of the gut, a pattern consistent with in situ hybridization experiments using amon nucleic acid probes (Siekhaus, 1997). Midgut endocrine cells have been identified in a variety of insect species and are hypothesized to act as monitors of the presence or absence of food in the gut, releasing their peptide hormone contents to direct food movement and digestion (Brown, 2003; Brown and Lea, 1989).
The homology of amon to proprotein convertase genes and the phenotypes of amon mutant larvae and pupae suggest that peptide hormones governing larval molting, larval growth, head eversion, leg and wing extension and abdominal differentiation are potential substrates of the amon protein. Using immunocytochemistry, we looked for co-localization of amon protein and candidate peptide hormone targets in the CNS.
The timing and behavioral events of larval molting are controlled by the peptide hormones eclosion hormone (EH), ecdysis triggering hormone (ETH), and crustacean cardioactive peptide (CCAP) (Mesce and Fahrbach, 2002). The localization of these three peptide hormones in Drosophila have been established (Horodyski et al., 1993; Park et al., 2003; Park et al., 2002), and both ETH and CCAP contain potential dibasic processing sites (Park et al., 2003; Park et al., 2002). In other insects, mature bioactive EH appears to be produced directly following cleavage of the signal sequence by signal peptidase (Horodyski et al., 1989; Wei et al., 2008; Zhang and Xu, 2006). The Drosophila EH pro-peptide includes a dibasic KR motif located 9 amino acids after the putative signal peptidase cleavage site (Horodyski et al., 1993). However, it has not been demonstrated that this potential processing site is used or is necessary for production of the mature bioactive Drosophila EH peptide. We used an antibody against ETH as well as EH and CCAP-cell GAL4 drivers and GFP reporters to identify ETH, EH, or CCAP expressing cells in the brains of wild-type larvae. We then asked whether the amon protein was co-expressed in these cells (Fig. 4). As seen in Fig. 4A-C, the amon protein and ETH do not co-localize to the epitracheal Inka cell. Similarly, amon protein does not co-localize with EH in the two ventromedial EH neurons (Fig. 4D-F), nor does it localize to the CCAP-expressing cells (data not shown). Our results with ETH and CCAP confirm the previous study of Park et al. (2004) and together, these results suggest that ETH, EH, and CCAP are not direct substrates of the amon protein.
The lack of abdominal differentiation in both amonC241Y and amonQ178st mutant pupae (Fig. 1C, D, Table 2) resembles that seen in late third instar larvae and pre-pupae treated with excess amounts of the terpenoid juvenile hormone (JH) and its mimics (Ashburner, 1970; Madhavan, 1973; Zhou and Riddiford, 2002). Because the release of JH from the corpus allatum (CA) is inhibited by the peptide hormone allatostatin (Nijhout, 1994), animals treated with excess JH and animals lacking an active allatostatin peptide hormone could be predicted to display similar phenotypes. Recently Siegmund and Korge (2001) identified a GAL4 driver that strongly identifies three neurons innervating the CA, and is weakly expressed the VM and CCMS 1. Because these three neurons account for all observable synapses on the CA using synapse-specific antibodies, Siegmund and Korge argue that these are the only neurons in Drosophila that form synapses on the CA (Siegmund and Korge, 2001) making them candidates for allatostatin expression. We used the Kurs-21 driver (Fig. 5A) and a GFP reporter to identify these neurons and then examined them for amon protein localization.
Figure 5B-D shows that the amon protein localizes to the three CA-innervating cells, CA-LP1 and CA-LP2. In addition, the amon protein localizes to two cells also identified by the Kurs-21 GAL4 driver known as CC-MS2 (Fig. 5E-G) which innervate the corpora cardiaca (CC) portion of the ring gland. It is therefore tempting to speculate that amon activity in cell types that innervate the CA and CC may influence the regulatory activities of these two important endocrine signaling centers.
To examine the functional consequences of restoring amon expression in the CA-LP1 and CA-LP2 cells in an amon mutant background, we expressed a uas-amon construct in these cells using the Kurs-21 GAL4 driver. Fig. 6A shows that amon function in these cells is sufficient to partially rescue developmental progression. Only 14% of first instar amon mutant larvae successfully complete the first to second instar larval molt and emerge as second instar larvae. In contrast, when amon function is restored in the CA- and CC-innervating neurons in an amon mutant background, 64% of amon mutant larvae complete this molt and emerge as second instar larvae (Fig. 6A). Restoring amon expression to the CA- and CC-innervating cells is also sufficient to partially rescue growth in amon mutants (Fig. 6B-D). Rescued amon mutant larvae (Fig. 6D) are intermediate in size compared to amon mutant (Fig. 6C) and control sibling larvae (Fig. 6B). Thus, restoring amon function in the cells defined by the Kurs-21 GAL4 enhancer trap line is sufficient to partially rescue the molting and growth defects exhibited by amon mutants.
Analysis of loss of function amon mutations has shown that amon is required during early larval development and for the completion of the first to second instar larval molt (Rayburn, et al. 2003). Here we have used a conditional rescue system to show that amon is also required for successful completion of pupal development and eclosion of the adult fly (Table 1). The predominant phenotypes of two amon mutant alleles in which amon expression was removed at the late third instar larval stage indicate a requirement for amon for behavioral events characterizing early stages of pupal development (Fig. 1, Table 2) and suggest that the developmental events of metamorphosis can be uncoupled.
The arrested amon mutant pupae fail to undergo head eversion (Fig. 1), an event that occurs in wild-type animals at 12-14 hours after pupariation (Bainbridge and Bownes, 1981), yet they complete processes such as eye pigmentation and head bristle formation (data not shown) that typically take place only after successful head eversion. Interestingly, despite the failure to complete head eversion and leg and wing extension, two of the earliest events of pupal development, the thoracic segment of these animals appears to develop normally as evidenced by its pigmentation and bristle formation (Fig. 1). Similarly, in a majority of the amonC241Y mutant pupae (58%) and a significant percentage of the amonQ178st mutant pupae (27%), the abdominal segment fails to differentiate despite the developmental progression of the anterior segments (Table 2, Fig. 1). In addition, a small subset of arrested amonC241Y pupae (15%) fail to anteriorly displace the mid-body gas bubble, an event that typically happens prior to head eversion, yet they display abdominal differentiation. These observations suggest that the head, thoracic, and abdominal segments of the fly respond to endocrine cues independently of one another and that interruption of these signals can cause an uncoordinated pattern of metamorphosis.
The amon protein localizes to cells in the central nervous system and gut in third instar animals (Fig. 3). To identify potential amon protein substrates, we compared the localization of three neuropeptides involved in molting and metamorphosis, CCAP, ETH, and EH (Park et al. 2003, Ewer, et al. 1997, McNabb et al. 1997, Park et al. 2002) to that of the amon protein. Phenotypic similarities between amon and ETH mutants (Park et al. 2002) and amon mutant pupae and pupae lacking the CCAP neurons (Park et al. 2003) suggested a relationship between amon and these peptide hormones. However, the amon protein does not co-localize with CCAP (data not shown), ETH, or EH (Fig. 4) arguing against a direct enzyme-substrate relationship between the amon protein and these neuropeptides. It is interesting to note, however, that amon transcripts were detected in the Inka cells by Siekhaus et al. (1999), but that another group (Park et al., 2004) using a different antiserum against the amon protein (Hwang et al., 2000) have also determined that the amon protein does not localize to the Inka cells, suggesting possible post-transcriptional regulation of the amon mRNA. This same group also examined amon protein's localization in relation to the CCAP-expressing cells and observed the same lack of co-localization we report here (Park et al., 2004). Together, these results suggest that additional, as yet unidentified peptide hormones are involved in the regulation of larval molting and metamorphosis. Such a conclusion is supported by a study of the relationship among EH, ETH, and CCAP which suggests that exogenous ETH's ability to elicit pre-ecdysis requires the EH cells and that EH may act independently of CCAP to elicit ecdysis (Clark et al., 2004).
Western analysis shows that the amon protein is present in all stages of development (Fig. 2). We note that the band recognized by our antiserum on Western blots does not migrate with a size similar to that reported for the amon protein in a previous study in which two bands migrating at 80 and 75 kDa were detected from extracts of third instar larvae and the medium of S2 cells expressing amon and d7B2 (Hwang et al., 2000). We cannot currently explain the apparent size differences between these studies but note that the 97kDa band detected here is competed by addition of amon peptides and not by a non-specific peptide competitor and that heat-shock induced expression of an amon cDNA leads to leads to increases in abundance of the 97kDa band (data not shown). It is possible that differences in the apparent mobility of the amon protein result from differing polyacrylamide gel electrophoresis conditions in the two studies.
Although in vivo mutational studies are critical to understanding the physiological roles played by SPCs (Furuta et al., 1997; Kass et al., 2001; Rayburn et al., 2003), the fact that these proteases have multiple processing targets creates difficulties in interpretation of phenotypes seen in mutants completely lacking function for a given SPC. Here we have begun to dissect the signaling contribution of subsets of neuronal cells to normal Drosophila development by restoring amon expression to these cells in an amon mutant background. We find that expression of amon using the Kurs-21 GAL4 driver partially rescues larval growth and molting defects exhibited by amon mutants (Fig. 6). This result ties amon function in cells defined by the Kurs-21 GAL4 driver to the regulation of larval molting and growth and suggests that the Kurs-21 cells may produce a processed peptide hormone signal that functions in these pathways. Because subsets of Kurs-21 cells make synaptic connections to either the CA or the CC, in the future it will be interesting to explore whether the rescuing activity demonstrated here acts via the CA or CC, or both, and whether this effect intersects the known signaling pathways controlling ecdysis and growth.
Previous genetic analysis of the peptide hormone processing enzyme PHM, which acts to α-amidate C-terminal residues of secretory peptides, has shown the importance of secreted peptide hormones in the modulation of developmental transitions and highlighted the potential of genetic methods in dissecting the contribution of secreted peptides to normal developmental progression (Jiang et al., 2000). The rescue approach that we describe here is a flexible genetic tool that can be used to correlate amon function in specific neuronal cells with the control of normal aspects of development and physiology as initially inferred from amon loss of function genetic studies and to ask whether and how amon contributes to the regulation of peptide hormone signaling. A powerful complementary approach to determine cell types in which amon function is required for normal growth and development will be the reduction of amon function in specific neuronal cell types using RNA inactivation. Numerous peptide hormones possessing potential dibasic cleavage sites have been identified in Drosophila (Hewes and Taghert, 2001). Ultimately, the genetic techniques described above may identify cells in which these known peptide signals are used in new ways and have the potential to lead to the identification of novel peptide signals.
The phenotypic and localization data presented here suggest that the amon protein functions in a wide variety of neuroendocrine cell types to process and activate a diverse suite of peptide hormones that coordinate larval and pupal development. In the future, we hope to identify cell types in which amon function is required for control of specific physiological, behavioral, or developmental events through cell-type specific rescue and inactivation experiments. A second important challenge for the future is biochemical identification of direct proteolytic substrates of the amon protein.
We thank Mark Brown for advice on immunocytochemistry and for use of imaging equipment. We also thank Richard Meagher and Anne Marie Zimeri for help with the antibody generation. This work was funded by grants to M.B. from the National Institutes of Health (GM53681) and the National Science Foundation (IOS-0823472). L.Y.M.R. was supported by an ARCS Foundation Fellowship, an American Association of University Women Educational Foundation Dissertation Fellowship and a National Institutes of Health Training Grant (GM07103). J. R. was supported by a National Institutes of Health Training Grant (GM07103). S.R.J. was funded by a Center for Undergraduate Research Opportunities (CURO) summer fellowship from the University of Georgia. The comments of reviewers improved the manuscript and are gratefully acknowledged.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.