Distinct Regulation of Transmitter Release at the Drosophila NMJ by Different Isoforms of nemy

Synaptic transmission is highly plastic and subject to regulation by a wide variety of neuromodulators and neuropeptides. In the present study, we have examined the role of isoforms of the cytochrome b561 homologue called no extended memory (nemy) in regulation of synaptic strength and plasticity at the neuromuscular junction (NMJ) of third instar larvae in Drosophila. Specifically, we generated two independent excisions of nemy that differentially affect the expression of nemy isoforms. We show that the nemy 45 excision, which specifically reduces the expression of the longest splice form of nemy, leads to an increase in stimulus evoked transmitter release and altered synaptic plasticity at the NMJ. Conversely, the nemy 26.2 excision, which appears to reduce the expression of all splice forms except the longest splice isoform, shows a reduction in stimulus evoked transmitter release, and enhanced synaptic plasticity. We further show that nemy 45 mutants have reduced levels of amidated peptides similar to that observed in peptidyl-glycine hydryoxylating mono-oxygenase (PHM) mutants. In contrast, nemy 26.2 mutants show no defects in peptide amidation but rather display a decrease in Tyramine β hydroxylase activity (TβH). Taken together, these results show non-redundant roles for the different nemy isoforms and shed light on the complex regulation of neuromodulators.


Introduction
The strength of transmitter release is highly plastic and can be modulated both acutely and chronically. The modulation of synaptic strength in response to specific inputs facilitates learned behaviours and allows synapses to synchronize their outputs to specific environmental cues. Synaptic strength may be modulated by the action of various neuromodulators including neuropeptides and biogenic amines such as dopamine, octopamine and serotonin, which exert varied effects on synapses through their respective receptors [1][2][3][4][5][6][7]. The modulation of synaptic strength by such neuromodulators facilitates the integration and co-ordination of multiple inputs. In this manner, neuromodulators play an essential role in regulating synaptic plasticity, which is essential for the survival of an organism in an ever-changing environment. It is therefore not surprising that there has been considerable interest in understanding the mechanisms by which neuromodulators regulate synaptic plasticity and behaviour.
We have previously shown that nemy, a cytochrome b561 homologue in Drosophila is required for aversive olfactory memory [8]. nemy is expressed in the larval and adult CNS. In the adult CNS, nemy shows strong expression in the antennal lobe and mushroom bodies, structures known to be involved in olfactory learning and memory [8]. In the larval CNS, nemy is expressed in several cells in the midline of the ventral nerve cord. Cytochrome b561 homologues maintain a pool of reduced intra-vesicular ascorbic acid by the transfer electrons generated by the oxidation of ascorbic acid in the cytoplasm across secretory vesicle membranes to reduce intra-vesicular semi-dehydroascorbate, producing ascorbic acid (Reviewed in [9]. Ascorbic acid is a necessary co-factor for mono-oxygenases such as Peptidyl-glycine hydroxylating mono-oxygenase (PHM) and Tyramine β hydroxylase TβH [10]. Both PHM and TβH affect a wide variety of behaviours and biological processes in flies.
PHM catalyzes the first step in a two-step reaction that converts a glycine residue at the Cterminal end of neuropeptide precursors into an α-amide [11,12], thus producing an active amidated neuropeptide. It is estimated that as many as 90% of all neuropeptides in Drosophila require C-terminal amidation to be biologically active [13]. PHM activity has been directly linked to the regulation of several behaviours and biological processes including developmental transitions during embryonic and larval development [13] and circadian locomotor rhythms [14]. In addition, specific neuropeptide precursor genes such as amnesiac, and dFMRF have been shown to affect behaviours including sleep maintenance [15], learning and memory [16][17][18], synaptic function and plasticity [19][20][21]. TβH is a homologue of dopamine β-hydroxylase (DβH) in mammals, which converts dopamine into nor-epinephrine [22]. In flies, TβH converts tyramine to octopamine, which acts as both a neuromodulator and neurohormone, similar to nor-epinenphrine in mammals [22,23]. Octopamine has been shown to affect a wide array of behaviours including egg laying [22], social behaviours [24,25] learning and memory [26] and synaptic function and plasticity [27,28].
Several studies have shown that larval NMJ function is regulated by both neuropeptides and neuromodulators such as octopamine [21,[28][29][30]. It is possible therefore, that mutations in nemy may lead to changes in NMJ function. Furthermore, it is possible that nemy may exert multiple effects on NMJ function through effects on both neuropeptide amidation and octopamine synthesis. In the present study, we have addressed this question by examining synaptic strength and plasticity in third instar larval NMJs and compared the effects of mutations in nemy mutants to those observed in PHM and TβH mutants. We show that the nemy gene gives rise to multiple isoforms. We generated independent deletions within the nemy gene that differentially affect the expression of different nemy isoforms. We show that loss of the longest splice form of nemy leads to a significant increase in stimulus evoked transmitter release while loss of the shorter splice forms leads to a decrease in stimulus evoked transmitter release. We also observed significantly different effects on synaptic plasticity in the two nemy mutants. We show that the differences in synaptic function observed in the nemy mutants may be explained by differences in the effects of these mutations on PHM and TβH activity, suggesting unique roles for the different nemy isoforms with regard to PHM and TβH function.

Fly Stocks
All stocks were raised on standard fly food, with a 12/12 hr light/dark cycle, at 23 ± 1°C and 45%-50% relative humidity. w 1118 flies that had previously been outcrossed for 10 generations with Canton S were used as the wild-type strain in this study. The nemy 45 excision was previously described [8]. The nemy 26.2 excision was generated via imprecise excision of P{EP} 16037 obtained from the GenExel Inc Stock centre. Both the nemy 45 and nemy 26.2 excisions were outcrossed to the cantonized w 1118 strain for at least 5 generations prior to the start of experiments. The tyramine β-hydroxlase precise (M6) and imprecise (M18) excision mutants were generated by M. Monastirioti (Institute for Molecular Biology and Biotechnology, Heraklion Greece) [22]. Both M6 and M18 lines were outcrossed to Canton S by Sarah Certel (Center for Structural and Functional Neuroscience, University of Montana). The M18 TβH mutants were maintained over a y, FM7 balancer chromosome. Experiments were performed on M18 males with a non y bristle phenotype. To eliminate any possible gender based variability, male M6 larvae were used as controls. The PHM 07623 mutant line in was a gift from P. Taghert (Department of Anatomy and Neurobiology, Washington University, St Louis). PHM 07623 flies were balanced over a 2 nd chromosome balancer with a ubiquitous GFP dominant marker to facilitate genotyping in embryonic and larval stages. To avoid genetic background issues arising from the presence of a balancer chromosome, balanced mutants were crossed to wild-type CS flies, and the non-GFP larvae were selected for experiments.

Electrophysiology
Larvae were dissected [31] in a haemolymph like solution HL3.1 [32] containing 0.2 mM calcium and pinned to the base of a Sylgard (Dow Corning) coated perfusion bath. The segmental nerves exiting the ventral ganglion were cut and drawn into a glass-stimulating pipette and stimulated with square wave pulses (0.15 ms, 5-15V) to generate action potentials in the motor neurons. Unless otherwise stated, intracellular recordings were obtained from muscle 6 of abdominal segments 3 and 4 with sharp intracellular microelectrodes filled with a 50/50 mix of 3 M potassium chloride and 3 M potassium acetate . Stimulus evoked excitatory junction potentials (EJPs) and spontaneous miniature EJPs (mEJPs) were recorded using an Axoclamp 2B amplifier and 0.1 LU head stage from Axon Instruments. Analog signals were low-pass filtered at 5 kHz, digitized at 20 kHz by a Powerlab/4SP A/D converter and saved on a computer for later analysis using Mini Analysis software (Synaptosoft). Quantal content was calculated independently for each NMJ by dividing the EJP amplitude by the mean mEJP amplitude.

RT-PCR
Total RNA was extracted from adult flies and purified using Mini RNA isolation columns from Invitrogen. Isolated RNAs were DNase treated using amplification grade DNase I (Invitrogen). DNase treated RNA was reverse transcribed using a SuperScript II reverse transcriptase kit from Invitrogen and 200-250 ng random hexamers per reaction. cDNA was amplified by PCR using the a single reverse primer that recognizes all nemy isoforms (TTCTCACAGCACAGCATTAC) and the following isoform specific forward primers: nemy RA , (CATTGAAAGGGTTTTTGAC), nemy RF , (GATTAGCAGCACATTCAACC), nemy RB, RC,RE, RG (TTCTCACAGCACAGCAT TAC). Note that due to the sequence similarity between nemy transcripts RB, RC, RE, and RG, these transcripts were all amplified in a single PCR reaction. The ribomal protein L32 (RP49) was used as a loading control and was amplified using the following primers (fwd: TTGAA GCTGGAAGGACACAA, rev: AATTCGGATCGATTCCTGTG). PCR reactions were run on a 1% agarose gel, and stained with Red-Safe nucleic acid stain.

Mass Spectrometry of amidated neuropeptides
The CNS of 50 third instar larvae were dissected and crushed in 40 ul of an ice-cold solution of methanol/water/formic acid (90:9:1) solution, filtered through Millipore spin-down filters (0.22mm, 6000 rpm for 10 min) dried and stored at -20°C. The pellets were resuspended in 5% acetonitrile and 0.1% formic acid. The peptides were separated by LCMS on a reverse phase column (Magic C18, Michrome Biosciences) using a split-free nano chromatography system (EASY-nLC, Proxeon, Odense Denmark). Samples were loaded via a C18 pre-column and analysis was carried out in duplicate. Elution was performed across a gradient from 5% to 30% acetonitrile in 0.1% Trifluoroacetic acid over 50 min. Peptide detection was carried out using a linear ion trap/FTMS hybrid mass spectrometer (LTQ-Orbitrap, Thermo-Fisher, Bremen Germany) operated in a data dependent mode. Peptide MS data was collected at a resolution of 60,000 full-width half maximum. Tandem mass spectra were obtained in parallel in the lineAr ion trap. The tandem mass spectra were compared to the D. melanogaster database (NCBI) using the Sequest search engine (Thermo-Fisher, San Jose CA). Potential modifications included in the search were amidated C-termini and oxidized methionine residues. The mass accuracy used in the search was 7 ppm for the parent ion and 0.5 Daltons for the fragment ions. Peptide quantitation was carried out using the Xcalibur 2.0 software (Thermo-Fisher, San Jose CA) generated extracted ion current for the monoisotopic isotopic peak of the doubly charged ion. The total area of the extracted peak for each ion was calculated for all samples. All measurements were done in triplicate biological replicates each of which contained two technical repetitions.

NMJ morphology
To examine the synaptic morphology, we transferred 25 male and 25 female flies to fresh bottles and allowed them 24 hrs to lay eggs. The control and nemy mutant third instar larvae from these bottles were dissected to expose the musculature of the body wall in phosphate buffered saline (PBS) then fixed at room temperature in 4% formaldehyde in PBS. Neuromuscular junctions were labeled with FITC-conjugated anti-horse radish peroxidase (1: 400, ICN Biomedicals, CA). Images were captured as a Z-stack using a Leica SP8 confocal microscope. Z-stacks were compressed into a single image and the total number of boutons on muscles 6 and 7 were counted.

Fluorescent in situ hybridization
Fluorescent in situ hybridization (FISH) experiments were performed on Oregon R embryos essentially as described in Wilk et. al [33]. Two RNA probes were used to distinguish the localization of the nemy RA isoform from other nemy isoforms. The first probe was chosen to specifically anneal to nemy RA transcripts, while the second probe was designed to hybridize to all nemy transcripts. Probes were DIG-labelled for detection with tyramide signal amplification-FISH as described in Wilk et al [33].

Statistical analysis
Data were analysed statistically using an ANOVA and t-test. All statistics were performed using Excel and Sigmaplot software.

Results
nemy was originally generated in a forward genetic screen for mutants affecting courtship memory [34]. A more thorough characterization showed that in addition to defects in courtship memory, nemy mutants also have a defect in middle-term olfactory memory [8]. The mutant used in both of these studies was a P{LacW} 153 insertion in the 1 st exon of nemy which specifically affects the expression of the longest splice form of the gene. However, Northern Blot analysis showed at least 4 distinct splice forms [8] and several shorter splice forms of nemy are also predicted in flybase (www.flybase.org). To determine whether the different splice forms of nemy have distinct functions we have generated two novel mutations within the nemy locus, named nemy 45 and nemy 26.2 by excision of two independent transposable elements, P {LacW} 153 and P{EP} GE16057 respectively (Fig 1). Both excisions are homozygous viable and breakpoints were confirmed by PCR and subsequent sequencing analysis. Interestingly, the nemy 45 and nemy 26.2 excisions exert different effects on the expression of nemy isoforms. Using primers specific for each transcript (see Materials and Methods), we examined the expression of the different nemy transcripts by RT-PCR. As shown in Fig 1, expression of the nemy RA transcript is significantly reduced in nemy 45 mutants, but is not affected in nemy 26.2 mutants. Conversely, expression of the remaining nemy transcripts is significantly reduced in nemy 26.2 mutants, but not affected in nemy 45 mutants.

Basal transmitter release and synaptic plasticity are impaired in nemy mutants
Both the nemy 45 and nemy 26.2 mutants were outcrossed to cantonized w 1118 flies for five generations prior to the commencement of experiments to avoid differences in the genetic background of the mutants. To assess whether either the nemy 45 or nemy 26.2 mutations affect synaptic transmission, we examined stimulus evoked and spontaneous transmitter release at the neuromuscular junctions of third instar larvae. We first examined the effect of the nemy 45 and nemy 26.2 excisions on spontaneously occurring miniature excitatory junction potentials (mEJPs) in muscle 6 of abdominal segments 3 and 4. Interestingly, we observed different effects of the nemy 45 and nemy 26.2 excisions on both the amplitude and frequency of spontaneous transmitter release compared to the w 1118 control (Fig 2). nemy 26.2 mutants showed a small but significant increase in mEJP amplitude compared to control without any significant change in the frequency of mEJPs. In contrast, nemy 45 mutants showed no significant change in mEJP amplitude, but did show a significant increase in the frequency of mEJPs. (Table 1, Fig 2B and  2C). Differences in mEJP amplitude are often equated with a change in the density of postsynaptic receptors, while changes in the frequency of mEJPs are exclusively regulated by the presynaptic neuron [35,36]. These observations of changes in mEJP amplitude and frequency may therefore indicate unique roles for the different isoforms of nemy in both the presynaptic and postsynaptic cells of the NMJ. We next examined the effect of nemy mutations on stimulus evoked transmitter release. Single pulses were delivered to cut motor neurons at a frequency of 0.2 Hz and the stimulus evoked EJPs were recorded from muscle 6 in abdominal segments 3 and 4. Compared to the control (w 1118 ), we observed a small decrease in stimulus evoked EJP amplitude in nemy 26.2 mutants (Table 1, Fig  2D). Combined with the increased amplitude of mEJPs in these mutants, this amounted to a significant decrease in quantal content of transmitter release from nemy 26.2 mutant NMJs ( Table 1, Fig 2E). By contrast, both stimulus evoked EJP amplitude and Quantal content were significantly increased in nemy 45 mutants compared to control (Table 1, Fig 2D and 2E).
Changes in transmitter release from the NMJ are often coupled with morphological changes including altered axon branching and changes in the number of synaptic boutons. To determine whether the observed changes in transmitter release observed in our nemy mutants also affected NMJ morphology, NMJs of third instar larvae were stained with FITC conjugated anti-horse radish peroxidise (FITC-HRP) that labels all post-mitotic neurons in Drosophila The nemy RA isoform is absent in nemy 45 mutants, but not affected in nemy 26.2 mutants. C. Conversely, isoforms B, E and G are present at control levels in nemy 45 mutants, but absent in nemy 26.2 mutants. Note, we were unable to detect a band that corresponded to nemy isoform C. D. The nemy isoform F is expressed at control levels in nemy 45 mutants, but significantly reduced in nemy 26.2 mutants. E. The house keeping gene RP49 was expressed at comparable levels in all genotypes and no signal was detected in 'no transcript controls' (NTC).  Basal transmitter release is altered in nemy mutants. A. Example stimulus evoked EJPs and spontaneous mEJPs. B-C. The amplitude (B) and frequency (C) of miniature EJPs in control (w) and nemy mutants. D. Stimulus evoked EJP amplitude in control (w) and nemy mutants. E. Quantal content in control (w) and nemy mutants. F. Example NMJs from control (w) and nemy mutant third instar larvae. Synaptic boutons are stained with a FITC-conjugated antibody raised against horse-radish peroxidise. G. Quantification of the number of synaptic boutons in control and nemy mutant NMJs. * indicates a significant differences of P < 0.05, ** indicates P < 0.01 and *** indicates P < 0.001. Statistical differences were calculated by ANOVA followed by multiple comparisons using the Student-Newman-Keuls method. doi:10.1371/journal.pone.0132548.g002 nemy Isoforms Differentially Affect Transmitter Release [37]. We observed a small but significant decrease in the number of synaptic boutons in nemy 26.2 mutants, but no significant change in bouton number in nemy 45 mutants (Table 1, Fig  2F and 2G). Taken together, these results indicate that nemy is required for proper function at the larval NMJ. Furthermore, these data suggest that different isoforms of nemy have distinct effects on NMJ function and morphology.
Given the observed defects in basal transmitter release, we next asked whether nemy mutants show impaired synaptic plasticity. We first looked at paired pulse plasticity by delivering pairs of stimuli with varying inter-stimulus intervals and comparing the amplitude of the second (test) EJP relative to the amplitude of the first (conditioning) EJP across genotypes ( Fig  3B). At short inter-stimulus intervals, the test pulse showed moderate facilitation in control preparations under the conditions used (0.2 mM extracellular calcium). nemy 26.2 mutant NMJs showed a greater facilitation at short inter-stimulus intervals than control NMJs (Table 1, Fig  3C). By contrast nemy 45 mutant NMJs showed a decrease in paired pulse facilitation, and even showed paired pulse depression at the shorter inter-stimulus intervals (less than 50 ms). Paired-pulse plasticity is correlated with transmitter release probability [38]. The increased paired-pulse facilitation observed in nemy 26.2 mutants suggests a decrease in the underlying probability of transmitter release, while the reduced facilitation and depression observed in nemy 45 mutants suggests an increase in transmitter release probability. These results are therefore consistent with the observations of decreased and increased Quantal content observed in nemy 26.2 and nemy 45 mutants, respectively.
We also examined a different form of synaptic plasticity known as paired pulse potentiation. Previous work has shown that the amplitude of EJPs are temporarily facilitated after a short (30s) tetanus (10 Hz stimulation) is applied [39,40]. We observed an approximately 50% increase in the amplitude of EJPs in control preparations following 30s of 10 Hz stimulation. The posttetanic-potentiation ratio (calculated as the amplitude of the post-tetanic stimulus evoked EJPs relative to the amplitude of pre-tetanic stimulus evoked EJPs) was significantly higher in nemy 26.2 mutant NMJs, but was not significantly different in nemy 45 mutant NMJs (Table 1). Differences in post-tetanic potentiation may reflect a difference in the size of the reserve pool of vesicles or in the ability to mobilize the reserve pools of vesicles, [41][42][43] suggesting that the reserve pool of nemy Isoforms Differentially Affect Transmitter Release reduced ascorbic acid in the lumen of secretory vesicles [9]. This pool of reduced ascorbic acid then serves as a co-factor for mono-oxygenase enzymes such as peptidyl-glycine hydroxylating mono-oxygenase (PHM) and tyramine β-hydroxylase (TβH) [10]. PHM performs the rate-limiting step in the C-terminal amidation of neuropeptide precursors while TβH catalyses the conversion of tyramine into octopamine. Octopamine and many neuropeptides act as neuromodulators and have been shown to affect a wide range of behaviours in flies. Previous work has shown that loss of TβH function leads to decreased transmitter release and a reduction in the number of synaptic boutons [27]. Interestingly, these larval NMJ phenotypes are strikingly similar to those observed in the present study in nemy 26.2 mutants. However, the effect of PHM mutations on NMJ function has not been characterized possibly due to the early larval lethality of homozygous PHM mutants [13]. In mice, PHM function is encoded by the bi-functional gene peptidylglycine a-mono-oxygenase (PAM) [11]. Similar to PHM mutants in flies, homozygous PAM mutants die during development [44]. However, heterozygous animals are viable but still show significant abnormalities related to reduced peptide amidation including impaired synaptic transmission [44].
To determine whether the defects in transmitter release observed in nemy mutants may reflect changes in either PHM or TβH activity, we first examined transmitter release and synaptic plasticity in TβH mutants and PHM P07623 hypomorphic mutants. Male TβH M18 mutants [22] were compared to male M6 precise excision controls (see Materials and Methods for details). Due to the early lethality of homozygous PHM 07623 hypomorphic mutants, we compared transmitter release in heterozygous PHM mutants to CantonS controls (see Materials and Methods for details). Consistent with previously published results, we observed a decrease in stimulus evoked transmitter release from the NMJs of TβH M18 mutants relative to the M6 control ( Table 2, Fig 4C). In our hands, we also observed an increase in the amplitude of spontaneous mEJPs in TβH M18 mutants relative to M6 control ( Table 2, Fig 4B). By contrast, PHM heterozygous mutants showed a marked increase in stimulus evoked transmitter release relative to control ( Table 2, Fig 4C), but did not show any significant difference in the amplitude of the spontaneous mEJPs. These defects in NMJ function observed in TβH and PHM mutants are strikingly similar to those observed in nemy 26.2 mutants and nemy 45 mutants respectively. To further compare the similarities between TβH or PHM mutants and our nemy mutants, we next examined the effect of mutations in TβH and PHM on paired pulse plasticity. We observed a significant decrease in paired pulse plasticity in heterozygous PHM mutants relative to the control ( Table 2, Fig 4E), but did not see any significant differences in paired pulse plasticity in TβH mutants ( Table 2, Fig 4D). The observation of a decrease in paired pulse plasticity in PHM mutants is consistent with our observations in nemy 45 mutants, and supports the hypothesis that nemy 45 mutants exert an effect on NMJ function via reduced PHM activity. The lack of any difference in paired pulse plasticity in the TβH mutants however was not consistent with nemy 26.2 mutants, suggesting that the defects in NMJ function observed in nemy 26.2 mutants may be only partially due to changes in TβH function.
Studies of TβH mutants have shown that defects in egg laying and locomotion can be rescued by feeding octopamine to the animals [22,45,46]. We hypothesized that if the defects observed in nemy 26.2 mutants are due, at least in part, to a reduction in TβH activity, we should be able to rescue this by feeding octopamine to larvae. Furthermore, the increased transmitter release observed in nemy 45 mutants may be partially masked by a concomitant reduction in TβH activity in these mutants. To test these possibilities, control and mutant larvae were transferred to food containing 10 mg/ml octopamine 24-36 hours prior to experiments. We then compared NMJ function in these animals to larvae that were maintained on regular food. Consistent with our earlier results, nemy 26.2 mutant larvae maintained on regular food showed reduced stimulus evoked transmitter release and increased altered paired pulse plasticity while nemy 45 mutants maintained on regular food showed increased stimulus evoked transmitter release and altered paired pulse plasticity (Table 3, Fig 4F and 4G). Control larvae that were fed octopamine showed no significant differences in stimulus evoked transmitter release or paired pulse plasticity relative to larvae maintained on regular food (Table 3, Fig 4F  and 4G). Similarly, nemy 45 mutants did not show any change in stimulus evoked transmitter release or synaptic plasticity following octopamine feeding. These results suggest that dietary derived octopamine is not able to exert a lasting effect in control and nemy 45 mutant larvae. However, nemy 26.2 mutants fed octopamine showed a significant increase in stimulus evoked transmitter release and a concomitant decrease in paired pulse plasticity compared to larvae maintained on regular food (Table 3, Fig 4F and 4G). Interestingly, stimulus evoked transmitter release and synaptic plasticity in octopamine fed nemy 26.2 mutant larvae was not significantly different to that observed in control larvae maintained on regular food, suggesting that supplying a source of octopamine to nemy 26.2 mutant larvae is sufficient to rescue the defects in NMJ function. Furthermore, since the effect of dietary octopamine was specific to nemy 26.2 mutants, these results suggest that the rescue of synaptic function in octopamine fed nemy 26.2 mutants must result from regulated release of octopamine from secretory vesicles.
Finally, to confirm an effect of the nemy mutations on PHM activity, we measured the level of amidated neuropeptides in control and nemy mutants. Previous work showed that peptide amidation levels were reduced in nemy 45 mutants [8], however peptide amidation in nemy 26.2 mutants was not examined. As in the previous study, we used the amidation of the FMRF family of neuropeptides derived from the dFMRF precursor [8] as a measure of peptide amidation. The ANOVA analysis did not reveal any differences (F 2,12 = 3.44, p = .066) between genotypes in the level of a nonamidated peptide (WFGDVNQKPI) that served as an internal control. Thus the intensity of amidated neuropeptides was normalized against the intensity of non-amidated peptides in the same sample. We observed significant differences in neuropeptide amidation in nemy 45 and nemy 26.2 mutants (Fig 5). Consistent with our previous work, we saw reduced amidation of the FMRFamide neuropeptides in nemy 45 mutants. We observed a significant reduction in the levels of both TPAEDFMRFamide (P < 0.05) and PDNFMRFamide (P < 0.005) in nemy 45 mutants. We also compared the level of amidation between nemy mutants and another wild-type Canton S control line (Fig 1 in S1 File). Although this line is less appropriate as a genetic background control than cantonised w 1118 , it is nonetheless relevant since all our lines were converted to a Canton S background. Moreover, the cantonised w 1118 control line does not differ from Canton S in the levels of amidation for all four FMRFalike peptides (S1 File). The ANOVA did not find any differences between Canton S, nemy 26  nemy isoforms show distinct localization The above results suggest the different nemy isoforms differentially affect PHM and TβH activity. We hypothesized that this likely arises from differences in the cellular or sub-cellular localization of the different nemy isoforms. Using an antibody that detects all isoforms of Nemy, we have previously shown that Nemy is expressed in both the larval and adult brain and can be detected both pre-and post-synaptically at NMJs of third instar larvae (Figs 1-3 in S1 File).
Lacking an antibody capable of distinguishing between individual isoforms of Nemy, we utilized fluorescent in situ hybridization (FISH) using two probes against nemy, one that specifically detects the longest splice form and a second probe that recognizes all nemy isoforms. Since nemy expression is highest during embryonic and pupal stages [47], we focussed on the expression of nemy transctipts in embryos. As seen in Fig 6, using a probe that recognizes all nemy transcripts, we detected nemy expression in the embryonic CNS and ventral nerve cord, Quantal content, while PHM/+ mutants showed an increase in Quantal content, but no change in mEJP amplitude. D-E. Paired pulse plasticity in TβH (D) and PHM (E) mutants and the appropriate controls. Paired pulse plasticity in TβH M18 mutants was not significantly different to that in TβH M6 controls. Paired pulse plasticity in PHM/+ mutants was reduced compared to CS controls. F. Quantal content in control (w) and nemy mutants maintained on regular food (Black bars), or on food supplemented with 10 mg/ml Octopamine (grey bars). On regular food, nemy 26.2 mutants showed a decrease in Quantal content while nemy 45 mutants showed an increase in Quantal content. Transferring larvae to octopamine supplemented food had no effect on the Quantal content of either control or nemy 45 animals, but significantly increased the Quantal content of nemy 26.2 mutants. G. Paired pulse plasticity in control and nemy mutants maintained on regular food (no drug) or food supplemented with 10 mg/ml octopamine (octopamine). The paired pulse ratio was not affected in either control (Black symbols) or nemy 45 mutants (red symbols) on an octopamine supplemented diet (compare open symbols to closed symbols). However, the paired pulse ratio was reduced to control levels in nemy 26.2 mutants (blue symbols) on an octopamine supplemented diet (compare solid inverted triangles in nemy 26.2 mutants on regular food with open triangles of nemy 26.2 mutants on octopamine supplemented food). * indicates P < 0.05 relative to the genetic control. ** indicates P < 0.01 relative to the genetic control. *** indicates P < 0.001 relative to the genetic control. # indicates P < 0.05 relative to the pharmacological control (regular food). Statistical differences were determined by ANOVA followed by multiple comparisons using the Student-Newman-Keuls method.
doi:10.1371/journal.pone.0132548.g004 Table 3. Effect of octopamine diet on basal transmitter release and paired pulse plasticity. we also observed fluorescence in processes exiting the ventral nerve cord that may be peripheral axons. In contrast, we were unable to detect expression of the nemy A transcript in the ventral nerve cord indicating that the different transcripts of nemy are differentially localized.

Discussion
The modulation of synaptic strength and plasticity by neuromodulators facilitates learned behaviours and allows synapses to synchronize their outputs to specific environmental cues.
The Drosophila NMJ has served as a useful model for examining genetic and pharmacological factors that regulate synaptic strength and plasticity for over 30 years. In the present study, we have examined the role of the cytochrome b561 homologue nemy on synaptic function and plasticity. We found that nemy isoforms differentially affect PHM and TβH function and may therefore indirectly influence some, or all, of the synapses that are modulated by amidated neuropeptides or octopamine. Since the nemy 45 excision specifically affects the expression of a single splice form (nemy A ) while the nemy 26.2 excisions affects the expression of several splice forms (nemy B , nemy E and nemy G ), we considered the possibility that differences in synaptic function of these observed mutants may reflect differences in the total level of nemy expression. However, if nemy transcripts were functionally redundant, and the phenotype was a reflection of the number of transcripts being expressed, we would expect both mutants to show similar defects in synaptic function, distinguished only by the relative severity of the defect. However, the two nemy excisions were not merely quantitatively different to each other, they exhibited unique phenotypes. Transmitter release probability was reduced and synaptic plasticity was enhanced in nemy 26.2 mutants, while the opposite was true for nemy 45 mutants. PHM function and thus neuropeptide amidation was impaired in nemy 45 mutants, but not affected in nemy 26.2 mutants, and finally, defects in nemy 26.2 mutants could be rescued by feeding octopamine to larvae, while the same feeding protocol had no effect on nemy 45 mutants. These results suggest that different isoforms of nemy have unique functions.

nemy and TβH function
Octopamine signalling is the invertebrate equivalent of adrenergic signalling in vertebrates. Octopamine receptors are structurally and functionally similar to adrenergic receptors [48,49] and the action of octopamine on these receptors affects a wide array of behaviours [22,[24][25][26]. Even the biosynthesis of octopamine in invertebrates shares similarities with nor-adrenaline biosynthesis in vertebrates. TβH shares 39% identity and 55% similarity with vertebrate Dopamine β hydroxylase [DβH,22] and converts tyramine to octopamine in the presence of ascorbic acid [10]. Octopamine has been shown to enhance transmitter release and muscle contraction in invertebrates including at the Drosophila NMJ [27,50] and has also been shown to affect synaptic outgrowth during starvation [27,28]. The results of the present study suggest that the shorter splice forms of nemy may indirectly regulate TβH function.
Previous analysis of nemy expression has shown strong localisation of nemy in several cells along the midline of the ventral nerve cord [8]. This staining pattern is consistent with reports nemy Isoforms Differentially Affect Transmitter Release of TβH expression [22]. In the present study, we showed reductions in transmitter release in nemy 26.2 mutants similar to those observed in TβH mutants [27]. Furthermore, we were able to rescue the defects in synaptic function by feeding larvae on a diet containing 10 mg/ml octopamine for 24-36 hr prior to experiments. These results suggest that the defects in synaptic function observed in nemy 26.2 mutants are due to a decrease in TβH function. Interestingly, we did not see any change in transmitter release in nemy 45 mutants following octopamine feeding suggesting that TβH function, and thus octopamine levels, may not be significantly altered in the nemy 45 mutants [27,28].
It is worth noting that rescue of synaptic function in nemy 26.2 mutants on an octopamine supplemented diet cannot be attributed to an acute action of octopamine on its receptors. Muscles 6 and 7 are not directly innervated by octopaminergic neurons [27,51,52], such that octopaminergic modulation of these synapses must occur by paracrine or endocrine release of octopamine from neighbouring muscles or the CNS. Given the relatively large volume of a physiology chamber (1-2 ml), such distant sources of octopamine are unlikely to be effective in an in situ assay of synaptic function. As such the rescue of nemy 26.2 mutants on octopamine supplemented food is indicative of a lasting effect of octopamine released into the haemolymph while the larvae are still intact and may indicate a persistent requirement for octopamine to maintain normal synaptic function.
Another interesting observation of the present study is the role of octopamine in short-term synaptic plasticity. We observed an enhancement of paired pulse plasticity in nemy 26.2 mutants that was rescued by addition of octopamine to the diet suggesting a role for octopamine in the regulation of paired pulse plasticity. Interestingly however, paired pulse plasticity did not show any significant impairment in TβH mutants, which lack octopamine. One possible explanation for these results may be the presence of multiple octopamine receptors at the NMJ. Recent work examining the role of octopamine in starvation induced synaptic growth has shown that octopamine induces synaptic outgrowth by activation of Octβ2R receptors [27], but also impairs synaptic outgrowth via an action on Octβ1R receptors [28]. It is possible that octopamine may exert similarly antagonistic effects on synaptic function through these receptors. In the absence of any octopaminergic signalling, as seen in TβH mutants, the net effect results in no significant change in paired pulse plasticity. In nemy 26.2 mutants however, TβH function may be impaired but not entirely absent, since nemy 26.2 mutants do not show defects in egg laying such as those seen in TβH mutants [22]. In these mutants, the loss or reduction of octopamine from selected sources leads to an enhancement of paired pulse plasticity that can be rescued by feeding octopamine to the larvae.

nemy and PHM function
Neuropeptides regulate a diverse array of biological processes through their actions as neuromodulators and endocrine hormones. All neuropeptides are synthesized as part of a larger, inactive neuropeptide precursor protein, which is post-translationally modified to generate one or more active neuropeptides [7]. The most common form of post-translational modification of neuropeptide precursors in Drosophila is C-terminal amidation of glycine-extended peptides by the combined actions of PHM and PAL (peptidyl-glycine α-lyase) [12]. It has been suggested that as many as 90% of all neuropeptides in Drosophila undergo C-terminal amidation [13]. PHM catalyzes the first and rate-limiting step of this C-terminal amidation, and is thought to be a key regulatory point in the processing of neuropeptides [11]. PHM activity has been directly linked to the regulation of several behaviours and biological processes including developmental transitions during embryonic and larval development [13] and circadian locomotor rhythms [14]. In addition, specific amidated neuropeptides such as the pituitary adenylate cyclise peptide (PACAP) and the FMRF family of neuropeptides have also been shown to affect behaviours including sleep maintenance [15], learning and memory [16][17][18] and synaptic function and plasticity [19][20][21].
The results of the present study confirm our previous observation of reduced neuropeptide amidation in nemy 45 mutants [8]. This decrease in peptide amidation is likely due to a decrease in PHM function. It should be noted however that nemy 45 mutants do not completely abolish PHM function. Homozygous PHM mutants die as late embryos or early larvae [13] while nemy 45 mutants are completely viable. It is possible that the remaining nemy isoforms also affect PHM function. While we did not observe a difference in the amidation of FMRF peptides in nemy 26.2 mutants, we cannot rule out a role for shorter nemy isoforms in the amidation of other neuropeptides. We have previously compared the expression of nemy and PHM in the larval CNS, and while there is considerable overlap in the expression patterns, not all PHM expressing cells showed expression of nemy [8]. Recent work identified a second cytochrome b561 homologue in Drosophila that is regulated by the transcription factor DIMMED [53]. DIMMED has been shown to be the primary transcriptional regulator of peptidergic cell fate in Drosophila [54]. Interestingly, while the other cytochrome b561 homologue (current designation CG1275) was upregulated by DIMMED expression along with PHM, the study did not find any increase in nemy expression [53]. Combined with the results of the present study, these results suggest that the cytochrome b 561 homologues in Drosophila play unique roles in regulating the function of PHM and possibly other mono-oxygenases. It should be noted however that our results may indicate the presence of a feedback loop between the expression/function of the cytochrome b561 homologues. The increased amidation of TPAEDFMRF neuropeptides observed in nemy 26.2 mutants may be the result of a compensatory increase in the expression of CG1275. While nemy is not the only cytochrome 561 homologue associated with peptidergic cells in Drosophila, the results of the present study suggest that the longest splice form of nemy affects PHM function and is required for the normal amidation of at least a subset of neuropeptides.
An interesting observation from the present study is that both nemy 45 and heterozygous PHM mutants result in elevated neurotransmitter release and decreased plasticity at the NMJ. However, the molecular identity of the neuropeptide(s) responsible for this change in NMJ function remains to be determined. Despite a wealth of knowledge surrounding the roles of specific neuropeptides in specific functions [7], relatively little information exists as to the effects of these peptides on NMJ function. One family of neuropeptides that has been well characterized at the NMJ are FMRFamides. This family of neuropeptides are characterized by the signature C-terminal tetra-peptide sequence Phe-Met-Arg-Phe-NH 2 [55] and have been shown to synaptic function in both vertebrates and invertebrates [20,21,56,57].
In the present study, we showed a decrease in the levels of amidated FMRF neuropeptides, including the most abundant species (DPKQDFMRFa) in nemy 45 mutants. Given that acute application of FMRFamide peptides enhances presynaptic calcium influx and transmitter release [21], the reduction of FMRFamide peptides might be expected to reduce transmitter release. However, we observed a significant increase in transmitter release in the nemy 45 mutants. While we did observe a decrease in transmitter release in the nemy 26.2 mutants, this is unlikely to be due to changes in the level of FMRFamides. First, MS data showed no significant changes in the levels of amidated neuropeptides in the nemy 26.2 mutants, and second, the reduction in transmitter release observed in the nemy 26.2 mutants was completely rescued by feeding octopamine to the larvae, suggesting that this reduction was due to an effect of TβH function and octopamine synthesis. Interestingly, the increase in transmitter release observed in nemy 45 mutants was strikingly similar to that observed in heterozygous PHM mutants, strengthening the suggestion that these defects are due to a decrease in PHM function. While the identity of the neuropeptide(s) affected in nemy 45 and PHM heterozygous mutants remains to be determined, the results suggest that a predominant regulatory feature of NMJ function includes an as yet unidentified inhibitory neuropeptide. Further work will be required to uncover the identity of this peptide.

Conclusion
The results of the present study show isoform specific roles for nemy in regulating NMJ function. Taken together, these results suggest that the longest nemy splice form, which is reduced in nemy 45 mutants, affects NMJ function predominantly by a reduction in PHM function. Alternatively, the shorter splice forms of nemy, which are reduced in the nemy 26.2 mutant, appear to affect NMJ function predominantly by a reduction in TβH function. In situ localization data suggests that the specific roles for the nemy isoforms may arise from different cellular localizations of the transcripts. While these effects of nemy on synaptic function are indirect, the different phenotypes observed in independent nemy mutants suggest cytochrome b561 may be an important regulator of neuromodulation. Regulation at this level may be particularly important for the integration and/or synchronization of different neuromodulatory signals. While neuropeptide release normally depends on extracellular calcium entry through voltage-gated channels, a recent study showed that octopamine could stimulate neuropeptide release independently of extracellular calcium at the NMJ [58]. Other recent studies have also observed interesting and novel interactions between the neuromodulatory effects of different biogenic amines [5,59]. While neuromodulation of synaptic strength has been viewed as an important factor in the regulation of behavioural plasticity for some time, these studies, combined with the present study suggest that neuromodulation of behavioural plasticity may be more complex than previously understood.