Mode of action of a Drosophila FMRFamide in inducing muscle contraction

Neuropeptides usually elicit physiological responses via G-protein-coupled receptors (GPCRs) which, in turn, act through second messenger pathways (Mercier et al., 2007; Meeusen et al., 2003; Vauquelin and von Mentzer, 2007). Many ‘RFamide’ peptides (i.e. containing Arg-Phe-NH₂ at the carboxyl terminus) have been shown to act via GPCRs (e.g. van Tol-Steye et al., 1999; Volterra and Siegelbaum, 1988; Wang et al., 1995), although molluscs contain a sodium channel that is directly gated by FMRFamide (Cottrell, 1997; Lingueglia et al., 1995). There is considerable interest in identifying the GPCRs that mediate the effects of RFamide peptides and the intracellular messengers that mediate their physiological effects.

Early studies demonstrated the involvement of GPCRs using pharmacological evidence, such as sensitivity to non-hydrolysable GTP analogues or to pertussis and cholera toxins, but more recent studies have identified peptide-activated GPCRs using genetic approaches, comparing genomes and transcriptomes with sequences encoding known GPCRs. The Drosophila genome contains ~44 genes encoding peptide-binding GPCRs (Hewes and Taghert, 2001; Brody and Cravchik, 2000), and several of these have been functionally characterized in cell cultures, usually Chinese hamster ovary cells. Four of these genes encode GPCRs that are activated by RFamide peptides. Only one of these GPCRs, referred to as ‘FR’ (encoded in gene CG2114), was reported to bind with high affinity to Drosophila peptides containing the sequence ‘FMRFamide’, and this gene product was less sensitive to other RFamide peptides (Meeusen et al., 2002; Cazzamali and Grimmelikhuijzen, 2002). This GPCR is a good candidate for mediating the physiological effects of Drosophila FMRFamides, and its involvement in mediating physiological responses needs to be tested more thoroughly using Drosophila cells or tissues.

DPKQDFMRFamide is the most abundant of eight peptides encoded in the dFMRFamide gene (Nambu et al., 1988; Schneider and Taghert, 1988) and is localized in neurohemal sites in fly larvae (Wegener et al., 2006). This peptide enhances nerve-evoked muscle contractions and synaptic transmission at neuromuscular junctions (Hewes et al., 1998; Dunn and Mercier, 2005). DPKQDFMRFamide increases the amount of transmitter released from nerve terminals (Hewes et al., 1998; Klose et al., 2010), and it also increases the amplitude of nerve-evoked calcium signals via release of calcium from intracellular stores (Klose et al., 2010). The enhancement of synaptic transmission requires calcium/calmodulin-dependent protein kinase (CaMKII) (Dunn and Mercier, 2005) and involves the FR GPCR (Klose et al., 2010). The intracellular pathways mediating the peptide-induced contractions, however, have not been studied, and possible involvement of the FR GPCR has not been investigated.

Earlier work demonstrated that DPKQDFMRFamide also elicits contractions in muscle fibres of the Drosophila larval body wall in the absence of the CNS (Clark et al., 2008). This effect appeared to involve a direct action on muscle fibres, as the peptide induced muscle contractions even after glutamate receptors on the muscle fibres were desensitized, which would prevent activation by increasing spontaneous release of glutamate from presynaptic terminals (Clark et al., 2008). However, other neurotransmitters, such as pituitary adenylate cyclase-activating polypeptide (PACAP), octopamine, proctolin and insulin-like peptide (ILP), are present in synaptic boutons on the larval body wall muscles (Anderson et al., 1988; Gorczyca et al., 1993; Monastirioti et al., 1995; Zhong and Peña, 1995). Thus, a possible presynaptic action of the peptide, through release of these other substances, was not ruled out.

The present investigation was aimed at identifying the receptor and second messenger pathways mediating DPKQDFMRFamide-induced contractions in Drosophila larval body wall muscles and determining whether the receptor was localized presynaptically or postsynaptically. We used RNA interference (RNAi) to test the hypothesis that DPKQDFMRFamide acts via FR to elicit contractions. Our data indicate that the peptide-induced contractions require the presence of the FR in muscle fibres. We also used genetic and pharmacological methods to investigate second messengers underlying peptide-induced contractions.

As in previous experiments (Clark et al., 2008), DPKQDFMRFamide induced slow contractions lasting up to 5 min (Fig. 1). To quantify this effect (an increase in muscle tonus), the tonus during a 5 min pre-peptide period (basal tonus) was subtracted from the maximal contraction observed during 5 min of peptide application. Responses are reported as mean tonus change ± s.e.m. We also observed spontaneous phasic contractions in both the presence and absence of DPKQDFMRFamide, as reported previously (Clark et al., 2008). As our earlier studies indicated that the frequency and amplitude of such phasic contractions do not correlate with peptide concentration (Clark et al., 2008), we did not examine these contractions further.

To investigate whether the GPCR encoded by the Drosophila FR gene plays a role in mediating the effects of DPKQDFMRFamide on muscle tonus, RNAi was used to reduce expression of this gene in Drosophila larvae (Dietzl et al., 2007). We obtained homozygous transgenic flies carrying an FR inverted repeat (FR-IR) under the control of an upstream activating sequence (UAS). To express the RNAi and knockdown FR expression, these flies were crossed with flies that exhibit ubiquitous expression of GAL4 homozygously under the control of the tubulin promoter (tubP-GAL4). The progeny (UAS-FR RNAi/+; tubP-GAL4/+) exhibited a significantly lower increase in muscle tonus in response to 1 μmol l⁻¹ DPKQDFMRFamide compared with control lines (Fig. 1). Muscle tonus increased by only 9.86±3.45 μN in these larvae, which was ~30% of the response elicited in both parental control lines (tubP-GAL4/+ genotype and UAS-FR RNAi/+ genotype; one-way ANOVA, Tamhane's post hoc, P<0.01) and only 26% of the response in Canton-S larvae (one-way ANOVA, Tamhane's post hoc, P<0.05). No difference was observed between DPKQDFMRFamide-induced contractions of the two parental control lines (Tamhane's post hoc, P>0.05), or between the parental control lines and Canton-S larvae (Tamhane's post hoc, P>0.05), in which expression of the FR gene was not manipulated. Thus, combining the UAS-FR RNAi line with GAL4 to ubiquitously disrupt function of the FR was successful at inhibiting the peptide's ability to elicit contraction. Quantitative polymerase chain reaction (qPCR) showed that the ubiquitous FR knockdown line exhibited the most dramatic reduction, ~90% reduction in transcript levels compared with Canton S and outcross controls (Fig. 2). These results suggest that the FR is at least partially responsible for mediating the myostimulatory effect of the peptide on larval body wall muscles.

To reduce FR expression specifically in muscle cells, UAS-FR-RNAi flies were crossed with 24B-GAL4 flies, which express Gal4 in all embryonic and larval somatic muscles (Schuster et al., 1996). This generated a heterozygous F1 generation, UAS-FR RNAi/+; 24B-GAL4/+. To control for potential effects of P-element insertions present in all larvae, two control lines were used: (a) heterozygous UAS-FR-RNAi flies with a non-active UAS-FR-IR construct and (b) heterozygous 24B-GAL4 lines containing a Gal4 vector only. Peptide-induced contractions in larvae with RNAi-regulated expression of the FR gene (UAS-FR RNAi/+; 24B-GAL4/+) in muscle cells were only 11% and 15% of the amplitude observed in UAS-FR RNA and 24B-GAL4 controls, respectively (Fig. 1, Tamhane's post hoc, P<0.05 and P<0.01), and only 9% of the amplitude observed in Canton-S larvae (Fig. 1, Tamhane's post hoc, P<0.05). Thus, reducing FR expression in muscle cells reduced responsiveness to the peptide by ~85–90%. There was no difference between the peptide-induced body wall contraction of Canton-S, UAS-FR RNA and 24-GAL4 control larvae (Tamhane's post hoc, P>0.05). qPCR data from the muscle knockdown line demonstrated a substantial reduction in transcript levels, roughly 75% knockdown compared with controls. The results suggest that expression of FR on muscle cells is necessary for most if not all of the peptide-induced contraction. These results also corroborate previously reported evidence that DPKQDFMRFamide induces contractions via a postsynaptic action (Clark et al., 2008).

To rule out the possibility that the tonus change might be mediated by presynaptic actions of the peptide, we crossed UAS-FR RNAi flies with flies harbouring pan-neuronal GAL4 expression of the elav-Gal4 driver. The effects of 1 μmol l⁻¹ DPKQDFMRFamide on the progeny (UAS-FR RNAi /+; elav-GAL4/+) were compared with effects on two control progeny lines (UAS-FR RNAi/+ and elav-GAL4/+). The nerve-specific driver was able to reduce transcript levels by ~60% relative to controls. The peptide's effect in larvae with FR-RNAi and pan-neuronal expression of Gal4 (UAS-FR RNAi /+; elav-GAL4/+) was similar to its effect in heterozygous control elav-GAL4 (Fig. 1; Tamhane's post hoc, P>0.05) and UAS-FR-IR larvae (Fig. 1; Tamhane's post hoc, P>0.05) as well as to the effect in Canton-S larvae (Fig. 1; Tamhane's post hoc, P>0.05). Thus, pan-neuronal expression of the FR-IR construct had no effect on the ability of the peptide to increase muscle tonus.

Presynaptic effects of DPKQDFMRFamide are mediated by both FR and one of the two dromyosuppressin receptors (Dms-R2 but not Dms-R1) (Klose et al., 2010). To investigate the potential involvement of the Drosophila myosuppressin receptors, homozygous transgenic flies carrying an inverted repeat for DmsR-1 and DmsR-2 under the control of UAS (UAS-DmsR-1 RNAi and UAS-DmsR-2 RNAi) were crossed with the same three drivers as FR (tubPGal4, 24B-Gal4 and elavGal4) as well as the appropriate out-cross controls. Responses of larval progeny of these genetic crosses to 1 μmol l⁻¹ DPKQDFMRFamide did not display any significant deviations from control lines (Fig. 3). We confirmed the reduction of expression for both myosuppressin receptors using qPCR. Data for DmsR-1 are as follows: UAS-DmsR-1 RNAi/+; 24B-Gal4/+ had 46% expression relative to wild-type, UAS-DmsR-1 RNAi/+; elav-Gal4/+ had 61% expression relative to wild-type, and UAS-DmsR-1 RNAi/+; tubP-Gal4/+ had 38% expression relative to wild-type. Data for expression of DmsR-2 relative to wild-type are as follows: UAS-DmsR-2 RNAi/+; 24B-Gal4/+ 51%, UAS-DmsR-2 RNAi/+; elav-Gal4/+ 65%, and UAS-DmsR-2 RNAi/+; tubP-Gal4/+ 38%.

The results presented above strongly suggest that the effect of DPKQDFMRFamide on muscle tonus is mediated via FR GPCRs localized in muscle fibres. As the ability of DPKQDFMRFamide to increase neurotransmitter release from synaptic terminals involves FR GPCRs in neurons (Klose et al., 2010) and requires CaMKII activity (Dunn and Mercier, 2005), we sought to determine whether peptide-induced contractions are also mediated by CaMKII. This question was addressed using pharmacological agents known to reduce CaMKII activity and a transgenic Drosophila line (ala1) that increases expression of the inhibitory peptide of CaMKII in response to heat shock (Griffith et al., 1993). We also used an additional transgenic line (UAS-ala), which enabled us to drive the expression of the inhibitory peptide of CaMKII using a muscle-specific driver (24B), independently of heat shock.

Contractions of body wall muscles were compared between ala1 larvae and a control line (UAS-ala, which contains the inserted gene for the inhibitory peptide but under the control of Gal4 rather than heat shock). In the absence of heat shock, ala1 and control larvae showed similar increases in muscle tonus in response to 1 μmol l⁻¹ DPKQDFMRFamide (Fig. 4; t-test, P>0.05). Hence, constitutive expression of the CaMKII inhibitory protein in ala1 larvae had no effect on the DPKQDFMRFamide-induced tonus change. Griffith et al. (Griffith et al., 1993) reported some expression of the alanine inhibitory peptide in ala1 flies at 25°C (1.9–3.9 μmol l⁻¹) and there is probably some expression at room temperature (22°C), but well below IC₅₀ (13 μmol l⁻¹ in Drosophila). A 1 h heat shock (37°C) increases expression of the inhibitory peptide and decreases CaMKII activity by 90–100%. Exposing ala1 flies to an hour-long heat shock (37°C, which has been shown to decrease CaMKII activity by 70–100%) (Griffith et al., 1993) did not reduce peptide-induced contractions in ala1 larvae but enhanced them by ~150% compared with non-heat-shocked ala1 larvae (t-test, P<0.05). Heat shock also increased peptide-induced contractions in control larvae (UAS-ala) by ~100% (t-test, P<0.05). There was no significant difference between the responses of heat-shocked ala1 and heat-shocked UAS-ala larvae (t-test, P>0.05). Driving the expression of the inhibitory peptide of CaMKII in muscle cells did not alter muscle responsiveness to exogenously applied DPKQDFMRFamide (Fig. 4; UAS-ala/24B-Gal4: 23.9±2.3, UAS-ala/+: 24.5±5.4, P>0.05). These results suggest that enhanced responsiveness to the peptide is attributable to the heat-shock treatment rather than to CaMKII inhibition, and that CaMKII inhibition does not antagonize the peptide's ability to induce muscle contraction.

A cell-permeable inhibitor of CaMKII activity, KN-93, was applied at a concentration of 1 μmol l⁻¹ for 20 min, and immediately thereafter the preparation was exposed to a solution containing both 1 μmol l⁻¹ DPKQDFMRFamide and 1 μmol l⁻¹ KN-93. There was no significant difference between contractions elicited by DPKQDFMRFamide alone (29.1±9.6 μN) or in the presence of KN-93 (25.8±3.3 μN; t-test, P>0.05).

Excitatory effects of RFamide peptides have been linked to phospholipase C (PLC) and the generation of IP₃ in Lymnea stagnalis (Willoughby et al., 1999) and Helix aspersa (Falconer et al., 1993). If DPKQDFMRFamide induces contractions via the PLC-IP₃ pathway, it would be predicted that such contractions would be attenuated or abolished by disrupting either PLC or IP₃ receptor activity. To test this, we took advantage of Drosophila fly lines in which either PLC or IP₃ receptor functionality are disturbed with mutations in genes (norpA and Plc21C) that encode PLCβ or in the single gene known to encode the IP₃ receptor, itpr, respectively.

Heterozygous IP₃ receptor mutant larvae (Itp-r83A⁰⁵⁶¹⁶) exhibit ~50% reduction in IP₃ receptor transcript levels (Klose et al., 2010). Fig. 5A shows the averaged change in tonus of heterozygous IP₃ receptor mutants (Itp-r83A⁰⁵⁶¹⁶) and wild-type larvae in response to DPKQDFMRFamide. The mutant larvae responded to the peptide with an average muscle tonus increase of 27.74±2.87 μN, which was not significantly lower than the response elicited in wild-type larvae (t-test, P>0.05).

Several mutant lines with deficits in PLC activity were used. Each of the mutants (w* norpA³³, w* norpA³⁶ and norpA⁷) carries a point mutation in the norpA gene induced by ethyl methanesulphonate mutagenesis on the Oregon-R line. PLC activity is reduced to ~1–1.5% of wild-type levels in norpA³³ and norpA³⁶ mutants (Pearn et al., 1996), and PLC activity in norpA⁷ mutants is reduced to about 2–3% of normal levels (Inoue et al., 1988). We also used the Plc21C gene mutant y¹ w¹¹¹⁸; Plc21C^A246, which was produced using P element insertion mutagenesis. It has been reported that the plc21C gene encodes two transcripts, of which one is expressed in the adult head only and the other in adult head and body tissue throughout development (Shortridge et al., 1991), but exact levels of their expression in y¹ w¹¹¹⁸; Plc21C^A246 mutant flies have not been reported. All the PLC mutants we examined showed an increase in muscle tonus in response to 1 μmol l⁻¹ DPKQDFMRFamide (Fig. 5B). Peptide-induce contractions in w* norpA³⁶, w* norpA³³ and norpA⁷ larvae were not significantly different from contractions in control (Oregon R) larvae and were not significantly different from each other (one-way ANOVA, P>0.05). Peptide-induced contractions were not significantly different between y¹ w¹¹¹⁸; Plc21C^A246 mutant larvae and y,w control larvae (t-test, P>0.05). These data do not support the hypothesis that either PLC or IP₃ is involved in mediating myostimulatory effects of DPKQDFMRFamide.

In other arthropods cAMP mediates peptide-induced muscle contractions by modulating K⁺ channels (Erxleben et al., 1995) and Ca²⁺ channels (Bishop et al., 1991; Bishop et al., 1987). As the ability of DPKQDFMRFamide to induce contractions requires L-type Ca²⁺ channels (Clark et al., 2008), it seemed plausible that the peptide might activate Ca²⁺ channels via the cAMP pathway. To investigate whether cyclic nucleotide monophosphates are involved in mediating peptide-induced contractions, cAMP and cGMP levels in body wall muscles were determined with enzyme immunoassay after incubation in various peptide concentrations in the presence of 3-isobutyl-1-methylxanthine (IBMX), a phosphodiesterase inhibitor that slows the breakdown of cAMP and cGMP (Beavo et al., 1970; Beavo and Reifsnyder 1990; Goy, 1990).

A 10 min exposure of body wall muscles to DPKQDFMRFamide at concentrations ranging from 10 nmol l⁻¹ to 1 μmol l⁻¹ in the presence of 0.5 mmol l⁻¹ IBMX did not significantly increase cAMP levels above the control level measured in IBMX alone (Fig. 6A, one-way ANOVA, Tamhane's post hoc, P>0.05). There was also no significant difference between cAMP levels following the exposure of larval muscles to 0.5 mmol l⁻¹ IBMX alone and saline alone (Fig. 6A, one-way ANOVA, Tamhane's post hoc, P>0.05), which raised concern that this concentration of IBMX might not be sufficient to slow the breakdown of cAMP in Drosophila larval muscles. This same concentration of IBMX, however, potentiated the increases in cAMP induced by forskolin, an adenylate cyclase activator (Fig. 6B). Forskolin increased cAMP levels by ~230% compared with saline controls (one-way ANOVA, Tamhane's posthoc, P<0.01), but it increased cAMP by 650% in the presence of 0.5 mmol l⁻¹ IBMX, representing a 2.3-fold increase in forskolin's effectiveness (one-way ANOVA, Tamhane's post hoc, P<0.01). Thus, the concentration of IBMX should have been sufficient to act synergistically with other cAMP-elevating compounds to increase cAMP levels.

The effect of the peptide on cGMP levels was also determined (Fig. 7). One-way ANOVA revealed a significant difference in cGMP levels between treatment groups (P<0.01). IBMX (0.5 mmol l⁻¹) increased cGMP levels by 260% compared with saline (one-way ANOVA, Tamhane's post hoc, P<0.01). Treatment with 0.01, 0.1 and 1 μmol l⁻¹ DPKQDFMRFamide in the presence of IBMX, however, did not change cGMP from the level observed in IBMX alone (Tamhane's post hoc, P>0.05). These data suggest that cGMP does not play a role in mediating the peptide's effects on muscle tonus.

To further investigate a possible role of cyclic nucleotide signal transduction pathways, we determined whether selective pharmacological agents would mimic the ability of DPKQDFMRFamide to induce contractions. Treatment with IBMX for 10 min had no significant effect on muscle tonus (1.1±7.3 μN, N=7). Treatment with 50 μmol l⁻¹ forskolin for 10 min appeared to cause a small change in tonus (11.6±4.7 μN, n=7), but this change was just at the threshold level for discernible effects reported previously (~10 μN) (Clark et al., 2008). This apparent effect of forskolin was substantially weaker than the effect of 1 μmol l⁻¹ DPKQDFMRFamide (Fig. 8A), despite the fact that forskolin increased cAMP levels at this concentration and DPKQDFMRFamide did not (Fig. 6).

If the peptide were to act by increasing cAMP or cGMP levels, its effectiveness should be enhanced by slowing hydrolysis of the cyclic nucleotide monophosphates with IBMX, particularly at threshold concentrations of the peptide, where cAMP and cGMP have not reached maximal levels. When DPKQDFMRFamide was applied at the threshold concentration for eliciting contractions (~0.01 μmol l⁻¹) (Clark et al., 2008), its effectiveness was not significantly increased by the presence of 0.5 mmol l⁻¹ IBMX (Fig. 6A, t-test P>0.05). Inhibitors of the cAMP-dependent protein kinase (Rp-cAMPs) and the cGMP-dependent protein kinase (Rp-8-pCPT-cGMPS) did not significantly alter the amplitude of contractions elicited by 1 μmol l⁻¹ DPKQDFMRFamide (Fig. 8B, t-test, P>0.05). Thus, the results do not support a role for cAMP, cGMP or their target kinase enzymes in mediating peptide-induced contractions.

To determine whether the myogenic effects of DPKQDFMRFamide might result from the involvement of signalling pathways mediated by arachidonic acid, we perfused third-instar larvae with arachidonic acid at concentrations ranging from 10⁻¹⁰ to 10⁻⁴ mol l⁻¹ and looked for changes in muscle tonus that might mimic the effect of the peptide. No significant change in tonus was observed at any of the concentrations tested (N=7–10 preparations for each concentration; 9.3±11.8 μN at 10⁻⁴ mol l⁻¹).

After determining that the various second messengers above are not involved in mediating the effects of DPKQDFMRFamide on muscle fibres, one possible explanation is a direct G-protein interaction. To examine this possibility, we incubated dissected third-instar tissue with an inhibitor of GPCR subunit uncoupling, pertussis toxin (PTX) (Fig. 9). Incubation with 500 ng ml⁻¹ PTX for 20 min resulted in a significant reduction in muscle responsiveness to exogenously applied DPKQDFMRFamide (PTX: 4.6±0.5, heat-inactivated PTX: 19.3±3.3; Mann–Whitney rank sum, P=0.002).

To determine whether DPKQDFMRFamide application leads to changes in membrane potential, we assessed over 300 intracellular recordings and found that the peptide does not produce any significant change in membrane voltage (control −41.3±1.3 mV, peptide −43.6±1.9 mV, P>0.05).

The present investigation provides evidence that the Drosophila FMRFamide, DPKQDFMRFamide, elicits contractions in larval muscle fibres via the GPCR FR. Peptide-induced contractions were reduced by 70–90% in heterozygous larvae in which RNAi reduces expression of the FR gene either in all cells or specifically in muscle cells. No reduction in peptide-induced contractions, however, was observed in larvae with reduced FR expression only in neurons. These observations confirm earlier work (Clark et al., 2008) indicating that DPKQDFMRFamide induces contractions via a direct action on larval muscle fibres.

Although peptide-induced contractions were reduced dramatically in larvae with reduced FR expression ubiquitously or in muscle, such contractions were not completely abolished. At least two possibilities could account for this result. First, the larvae used for physiological recordings were heterozygous for the GAL4 and UAS-RNAi element, which would make it plausible that larval cells retained sufficient expression of the dFMRFamide receptor to mediate a weak response to the peptide. We assessed this possibility through qPCR. In ubiquitous FR knock-down lines, a 70% reduction in responsiveness to DPKQDFMRFamide corresponded to a 90% reduction in expression, and in muscle-specific FR knock-down lines an 85% reduction in peptide responsiveness corresponded to a 70% reduction in expression. However, qPCR results are not a direct measure of protein expression levels, which we did not assess. Thus, in both cases the lack of a complete reduction in responsiveness to exogenous application of DPKQDFMRFamide correlates with incomplete inhibition of FR expression. An alternative possibility is that DPKQDFMRFamide may act via targets other than the FMRFamide receptor. Johnson et al. (Johnson et al., 2003) showed that 100 nmol l⁻¹ DPKQDFMRFamide elicits effects that are mediated by both the FMRFamide receptor and a myosuppressin receptor when expressed in Human embryonic kidney (HEK) cells. The FMRFamide receptor (FR) and the myosuppressin receptor (DmsR-2) are also necessary for DPKQDFMRFamide to enhance transmitter release from synaptic terminals (Klose et al., 2010). None of the DmsR knock-down lines, however, showed any change in responsiveness to DPKQDFMRFa compared with controls (Fig. 3) despite the reduction in gene expression, which was as high as a 62% reduction in lines where the RNAi was expressed ubiquitously. Thus, it is unlikely that the myogenic effects of DPKQDFMRFamide are mediated through the myosuppressin receptors.

Although earlier work demonstrated a role for CaMKII in modulating transmitter release in Drosophila larvae by DPKQDFMRFamide (Dunn and Mercier, 2005; Klose et al., 2010), data reported here indicate that CaMKII activity is not necessary for the peptide to induce muscle contraction. The requirement of different intracellular pathways for these two physiological responses indicates that they are distinct and strengthens the interpretation that the peptide's ability to induce contractions represents a direct effect on muscle fibres rather than some presynaptic effect, such as an increase in spontaneous release of neurotransmitters. In the context of earlier work, the current findings indicate that postsynaptic and presynaptic effects of one modulatory substance may be mediated by different intracellular signalling pathways and by different receptor combinations.

Heat-shock treatment potentiated contractions elicited by DPKQDFMRFamide in ala1 and UAS-ala larvae by ~150% and 100%, respectively. The reason for such potentiation is not clear but probably involves an increase in intracellular calcium levels in the muscle fibres. Warming Drosophila larval salivary gland cells to 35°C increases intracellular Ca²⁺ concentration 10-fold, with a slow recovery of [Ca²⁺]_i, starting ~45 min after the temperature had cooled to 25°C (Drummond et al., 1986). If heat shock has a similar effect in muscle cells, an increase in intracellular Ca²⁺ might last long enough to augment the response to the peptide. Alternatively, a substantial rise in [Ca²⁺]_i might activate intracellular processes that potentiate peptide-induced contractions even after Ca²⁺ levels have subsided.

Our observation that DPKQDFMRFamide-induced contractions were not impaired in the IP₃ receptor mutant (Itp-r83A⁰⁵⁶¹⁶/+) suggests that this receptor may not be required for the increase in muscle tonus. There is only one IP₃ receptor-encoding gene in Drosophila, and larvae of this particular mutant were shown to exhibit a 50% reduction in itpr-83 mRNA levels compared with wild-type larvae (Klose et al., 2010). Such a reduction in expression was sufficient to abolish the ability of DPKQDFMRFamide to enhance peak Ca²⁺ levels in presynaptic terminals and excitatory junction potential amplitude. The present observations indicate that postsynaptic induction of muscle contraction does not exhibit the same dependence on the IP₃ receptor as the presynaptic effects for this neuropeptide. As the IP₃ receptor mutant still expresses 50% of the normal level of itpr-83 mRNA, we cannot completely rule out the involvement of this receptor.

DPKQDFMRFamide-induced contractions were neither impaired nor potentiated in four PLC mutant fly lines, each containing a different mutation. Although there are several PLC types, only PLC-β is activated by a GPCR (Vauquelin and von Mentzer, 2007). In D. melanogaster two genes, norpA and Plc-21c, encode PLC-βγ. The norpA mutants exhibit a 97–99% reduction in PLC activity compared with control flies, as determined by in vitro PLC assays of eye and whole-head homogenates (Inoue et al., 1988; Pearn et al., 1996). Unfortunately, the level of PLC activity in tissues of Plc-21C mutants is unknown. The lack of impairment in peptide-induced contractions in any of the norpA mutants, however, suggests that PLC is probably not required for the postsynaptic response. This would suggest that DPKQDFMRFamide does not initiate muscle contractions by PLC-dependent mechanisms, such as generation of diacylglycerol and IP₃.

Several lines of evidence presented in this study suggest that cAMP, cGMP and their respective protein kinase enzymes are not required for DPKQDFMRFamide to increase muscle tonus in Drosophila larvae. First, the peptide did not increase cAMP or cGMP levels in Drosophila larval muscles in the presence of the phosphodiesterase inhibitor IBMX. Second, IBMX failed to potentiate the peptide's effect on muscle contraction. Third, neither IBMX nor forskolin mimicked the peptide's effect on muscle tonus very well. IBMX had no effect on muscle contraction, despite its ability to increase cGMP levels by 260%. The contractions elicited by 50 μmol l⁻¹ forskolin were much smaller than those elicited by DPKQDFMRFamide, despite the fact that forskolin increased cAMP levels by 230% and the peptide did not alter cAMP significantly. Fourth, inhibitors of cAMP-dependent and cGMP-dependent protein kinases failed to reduce peptide-induced contractions. These observations suggest very strongly that changes in the concentration of cAMP or cGMP are not necessary to mediate the peptide's ability to increase muscle tonus. This would differentiate the peptide-induced contractions in Drosophila larval muscles from the effects of several neuropeptides (Bishop et al., 1991; Bishop et al., 1987; Erxleben et al., 1995; Nykamp and Lange, 2000; Trim et al., 1998) and biogenic amines (Clark and Lange, 2003; Knotz and Mercier, 1995; Nykamp and Lange, 2000) that are reported to modulate contraction of invertebrate muscles via cAMP.

In our efforts to illuminate the putative second messenger cascade initiated by the activation of the Drosophila FR, we also examined the effects of arachidonic acid, which failed to elicit changes in tonus at concentrations ranging from 10⁻¹⁰ to 10⁻⁴ mol l⁻¹. This is not surprising as two previous reports provide evidence that arachidonic acid may not be present or utilized in Drosophila (Chyb et al., 1999; Yoshioka et al., 1985). Thus, it is unlikely that DPKQDFMRFamide mediates its effects through pathways utilizing arachidonic acid.

Overall, the current results confirm that DPKQDFMRFamide increases muscle tonus in Drosophila larval muscles by acting directly on the muscle fibres via at least one GPCR. The mechanisms underlying this modulatory effect are still not known but do not appear to involve CaMKII, PLC, cAMP, cGMP or arachidonic acid. We have not completely ruled out involvement of the IP₃ receptor, but the data presented here suggest this is unlikely. Other second messengers, such as linoleic acid, have not been examined. The ability of PTX to reduce peptide-induced contractions (Fig. 9) suggests that the FR acts on a G-protein of G_i/o type (Li et al., 1995). G-protein sub-units have been reported to act directly without the involvement of second messengers in other systems (e.g. Dascal, 2001; Jiang and Bajpayee, 2009; Soejima and Noma, 1984; Wickman et al., 1994). As DPKQDFMRFamide-induced contractions are completely abolished by nifedipine and nicardipine and by lowering extracellular Ca²⁺ concentration (Clark et al., 2008), they appear to require Ca²⁺ influx through L-type channels associated with the sarcolemma, which have been shown to be present in body wall muscles of Drosophila larvae (Gielow et al., 1995). A similar dependence on extracellular Ca²⁺ and L-type channels has been reported for peptide-induced contractions in other arthropod muscles (Donini and Lange, 2002; Kravitz et al., 1980; Lange et al., 1987; Quigley and Mercier, 1997; Wilcox and Lange, 1995). One possible model is that DPKQDFMRFamide activates the PTX-sensitive G-protein, whose subunits act directly on the L-type channels to elicit Ca²⁺ influx, possibly shifting their activation voltage. Peptide-induced Ca²⁺ influx would induce contraction via calcium-induced release of calcium from the sarcoplasmic reticulum (Sullivan et al., 2000; Peron et al., 2009; Ushio et al., 1993). G_βγ subunits have been shown to act directly on N-, P/Q-, R- and T-type calcium channels (Herlitze et al., 1996; Ikeda, 1996; Wolfe et al., 2003) but they inhibit these channels rather than excite them. Furthermore, there is no evidence that L-type channels are modulated directly by G-protein subunits. Alternatively, the peptide might act downstream of calcium, possibly by reducing its re-uptake into the sarcoplasmic reticulum or by changing the sensitivity of the contractile proteins to calcium.

Canton S (CS) flies, obtained from Bloomington Drosophila Stock Center (BDSC) were used for the experiments unless otherwise indicated.

The IP₃ receptor mutant Itp-r83A⁰⁵⁶¹⁶ {P{PZ}Itp-r83A⁰⁵⁶¹⁶ Nmdar1⁰⁵⁶¹⁶ ry⁵⁰⁶/TM3, ry^RKSb¹Ser¹} (Spradling et al., 1999), was obtained from BDSC. As this mutation was balanced over a Sb¹Ser¹ dominantly marked balancer chromosome, a genetic cross scheme was performed to remove the balancer. First, a cross between the IP₃ mutant receptor line and a Ser¹ balancer line (BDSC) was performed. The progeny line was further crossed with w¹¹¹⁸ flies (BDSC) to remove the marked balancer chromosome.

Four fly lines with mutations in two genes encoding PLCβ were used. Three were mutants of the norpA (no receptor potential) gene: w*norpA³³, w*norpA³⁶ and norpA⁷ (BDSC). As the norpA mutants originated from Oregon R fly stocks, wild-type Oregon R was used as a control. The Plc21C gene mutation line and its control line were from BDSC.

To investigate the role of CaMKII, we used an ala1 transgenic line containing a synthetically generated alanine inhibitory peptide gene on the first chromosome under the control of a heat-shock promoter (Griffith et al., 1993). The UAS-ala line contains an alanine inhibitory peptide gene inserted downstream of the UAS and was used as a control for the ala1 line. Both transgenic lines were gifts from Dr Leslie Griffith (Brandeis University, Waltham, MA, USA). Additionally, we used a commercially available ala line containing the inhibitory peptide for CaMKII under UAS control which we crossed with the muscle-specific driver 24B.

A transgenic line containing the FMRFamide receptor inverted repeat (FR-IR) downstream UAS was obtained from the Vienna Drosophila RNAi Center (VDRC no. 9594). The generation of this line (UAS-FR RNAi) was described previously (Dietzl et al., 2007). Briefly, an FMRFamide receptor gene fragment was cloned as a 301 bp long inverted repeat (IR) in antisense–sense orientation into a modified pUAST vector, pMF3, with multiple UAS sites. This cloned construct was then inserted on the second chromosome of an isogenic w¹¹¹⁸ host, generating a homozygous viable UAS-FR RNAi line. Two transgene lines containing inverted repeats downstream of UAS for the Drosophila myosuppressin receptors 1 and 2 (DmsR-1 and DmsR-2) were obtained from the Viennna Drosophila RNAi Center (VDRC nos 9369 and 49952).

The following driver lines were used to express the RNAi for FR, DmsR-1 and DmsR-2: elav-GAL4 (BDSC), 24B-GAL4 (BDSC) and tubP-GAL4 (BDSC). We used elav-GAL4 for pan-neuronal expression of the UAS-FR-IR transgene (Luo et al., 1994; Sink et al., 2001). 24B-GAL4 (Luo et al., 1994; Brand and Perrimon, 1993) was used to express UAS-FR-IR in all larval somatic muscles (Schuster et al., 1996). tubP-GAL4 is an insert on the third chromosome that is balanced over TM3, Sb and allows for ubiquitous expression of Gal4 (Lee and Luo, 1999).

All flies were raised on a cornmeal-based medium (Boreal Laboratories Ltd, St Catharines, ON, Canada), supplemented with dry yeast, at 21°C on a 12 h:12 h light:dark cycle.

Tissue-specific expression of the UAS-FR IR construct was driven using the UAS/GAL4 system as described elsewhere (Brand and Perrimon, 1993). To express the UAS-FR IR construct ubiquitously, homozygous UAS-FR RNAi virgin females were crossed to tubP-GAL4 males that were previously balanced. In addition, balanced tubP-GAL4 males were crossed with w¹¹¹⁸ virgin females to remove the balancer chromosome and generate a heterozygous control line (+/+; tubP-GAL4/+). Expression of the UAS-FR IR construct in muscle and neuronal tissues was accomplished by crossing homozygous UAS-FR RNAi virgin females to homozygous 24B-GAL4 and elav-GAL4 males, respectively. Heterozygous larvae of the F1 generation were used for the experiments. To generate appropriate heterozygous Gal4 controls, homozygous elav-GAL4 and 24B-GAL4 males were crossed to w¹¹¹⁸ virgin female flies. To generate appropriate control larvae with the non-activated UAS-FR construct, homozygous UAS-FR RNAi virgin females were mated to w¹¹¹⁸ virgin female flies. The same approach was used for the expression of UAS-DmsR-1 IR and UAS DmsR-2 IR. All progeny and parent lines were kept at 27°C on a 12 h:12 h light:dark cycle.

Knockdown of RNAi-lines was assessed by RT-qPCR. Total RNA was isolated using Norgen's Total RNA Purification Kit (St Catharines, ON, Canada) and 500 ng of total RNA were reverse transcribed with iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA). For real-time qPCR, SYBR Green qPCR Supermix (Invitrogen) was added to cDNA and primers. Each sample was amplified for 40 cycles in a thermocycler (Bio-Rad): 5 min at 95°C, 15 s at 95°C, 90 s at 58°C and 30 s at 72°C. The delta delta Ct (2^−ΔΔCt) method was used for data analysis, with rp49 as a house-keeping gene for data normalization. Primers used in RT-qPCR reactions are listed in Table 1.

The ala1 and UAS-ala flies were placed in empty glass vials and transferred to a dry incubator for 1 h at 37°C. After heat shock, flies were allowed to recover at room temperature (21°C) for approximately 2 h before muscle contractions were recorded.

Wandering stage third-instar larvae were used for all experiments. They were collected from the sides of their culture vials immediately before dissections were made. All dissections were made at room temperature (~21°C) in modified haemolymph-like (HL6) Drosophila physiological saline (Macleod et al., 2002). The saline contained (in mmol l⁻¹): 23.7 NaCl, 24.8 KCl, 0.5 CaCl₂, 15 MgCl₂, 10 NaHCO₃, 80 trehalose, 20 isethionic acid, 5.7 l-alanine, 2.0 l-arginine, 14.5 glycine, 11.0 l-histidine, 1.7 l-methionine, 13.0 l-proline, 2.3 l-serine, 2.5 l-threonine, 1.4 l-tyrosine, 1.0 l-valine and 5 BES, pH 7.2. The saline was stored at −20°C in 200 ml aliquots.

Larvae were pinned dorsal side up at the anterior and posterior ends to a dish containing saline. They were stretched slightly longitudinally, and a mid-dorsal cut was made along the length of the animal (Fig. 1, top). The viscera were removed, and the segmental nerves were cut near the ventral ganglion, which was subsequently removed along with the brain (Jan and Jan, 1976). These preparations were used for muscle contraction recording or enzyme immunoassay.

After dissection, the pin in the anterior end of the larva was removed, and the anterior tip of the larva was hooked to a Grass FT03 tension transducer (Grass Instruments, Quincy, MA, USA) as described previously (Clark et al., 2008). Contractions were amplified using a MOD CP 122A amplifier (Grass Telefactor, West Warwick, RI, USA) and were recorded on a chart recorder. The recording dish had a volume of ~0.2–0.4 ml and was perfused continuously at a rate of 0.7 ml min⁻¹. Excess fluid was removed by continuous suction.

Dissected preparations of larval body wall muscles were placed in cold saline, and pools of two to nine larval preparations were transferred into 100 μl of desired incubation solution. Individual pools of tissues to be tested for cAMP content were incubated separately in (a) saline, (b) a non-selective phosphodiesterase inhibitor, 0.5 mmol l⁻¹ IBMX, (c) 0.5 mmol l⁻¹ IBMX containing either DPKQDFMRFamide at concentrations ranging from 0.01 to 1 μmol l⁻¹ or 0.05 mmol l⁻¹ forskolin (an adenylate cyclase activator), (d) 0.5 mmol l⁻¹ IBMX containing 0.05 mmol l⁻¹ forskolin or (e) 0.05 mmol l⁻¹ forskolin alone. Preparations to be assayed for cGMP were incubated in either (a) saline alone, (b) saline containing 0.5 mmol l⁻¹ IBMX or (c) saline containing 0.5 mmol l⁻¹ IBMX with concentrations of DPKQDFMRFamide ranging from 0.01 to 1 μmol l⁻¹. All incubation solutions contained 0.25% DMSO with the exception of solutions containing forskolin, which contained 0.65% DMSO. After 10 min of incubation at room temperature, each larval pool was placed into 100 μl of 0.1 mol l⁻¹ hydrochloric acid (HCl) for 5 min to inhibit endogenous phosphodiesterase activity. To halt the chemical reactions, larvae were frozen on a steel plate cooled over dry ice, and tissue was homogenized in 500 μl of 0.1 mol l⁻¹ HCl and centrifuged for 12 min at 11,000 g. The supernatant was stored at −80°C until ready to assay. The pellet was further submitted to a protein determination assay.

Levels of cAMP or cGMP were determined in duplicate from supernatant (100 μl) following acetylation protocols for direct cAMP and cGMP enzyme immunoassay kits (Assay Designs, Ann Arbor, MI, USA). Optical densities were read at 405 nm with a BioTek Synergy HT microplate reader (BioTek, Winooski, VT, USA), and the concentrations of cAMP or cGMP (pmol ml⁻¹) were estimated using a 4-parameter logistic curve-fitting program provided in the BioTek KC4 Software. The sensitivities of the acetylated version of the cAMP and cGMP assays were 0.037 and 25 pmol ml⁻¹, respectively.

Each pellet was dissolved in 100 μl of 1 mol l⁻¹ NaOH and placed for 2 h into a hot waterbath (40°C). Protein content was measured from 10–15 μl of solubilized pellet with a Bio-Rad Protein Assay kit (Bradford, 1976), using BSA as a protein standard. To account for variations in the number and size of larval body wall preparations, cAMP and cGMP levels were expressed as pmol mg⁻¹ protein.

PTX was obtained from Cedarlane (Burlington, ON, Canada). Dissected third-instar larval tissue was incubated in 500 ng ml⁻¹ PTX for 20 min prior to examination of tonus. For control experiments, we heat-inactivated PTX by placing the solution in a waterbath just below boiling for 30 min.

The Drosophila peptide DPKQDFMRFamide was synthesized by Cell Essentials (Boston, MA, USA) and was 98% pure as determined by reverse-phase HPLC. Peptide was stored at −20°C and was dissolved in saline to yield a 10 mmol l⁻¹ stock solution. IBMX was purchased from Sigma-Aldrich (Oakville, ON, Canada). IBMX stock solution (5 mmol l⁻¹) was made in 0.5% DMSO containing saline and kept at −20°C. The DMSO concentration in the final IBMX solution used for physiological recordings did not exceed 0.05%. Forskolin, Rp-cAMPS (adenosine 3′,5′-cyclic monophosphorothioate, Rp-isomer, triethylammonium salt), Rp-8-pCPT-cGMPS [guanosine 3′,5′-cyclic monophosphorothioate, 8-(4-chlorophenylthio)-, Rp-isomer, triethylammonium salt] and KN-93 were obtained from Calbiochem. Forskolin and KN-93 were dissolved in 100% DMSO and stored at 4°C as stock solutions (12 and 10 mmol l⁻¹, respectively), which were subsequently diluted in saline to yield the desired drug concentration with final DMSO concentrations of 0.4% and 0.1%, respectively. Rp-cAMPS (5 μmol l⁻¹) and Rp-8-pCPT-cGMPS (1 μmol l⁻¹) were stored at −20°C until ready to dilute in saline to yield the desired final concentration. Arachidonic acid sodium salt was purchased from Sigma-Aldrich (Oakville, ON, Canada) readily dissolved in saline. All experimental solutions were made fresh on the day of testing.

Both one-way ANOVA and t-test for independent samples, unless otherwise stated, were used where appropriate to determine statistical significance, and P<0.05 was used for acceptance of statistical significance. All data are expressed as means ± s.e.m.

We thank Dr Satpal Singh for advice on fly line crosses.