Cyclic AMP is involved in cardioregulation by multiple neuropeptides encoded on the FMRFamide gene

The majority of studies to date have shown that neuropeptides mediate their effects via the activation of a variety of different second messenger pathways, with each signalling pathway playing an important role in producing the desired biological response (Lundberg, 1996). The second messenger-mediated effects resulting from the release of a combination of peptides are likely to be highly complex and have been little studied. We chose to investigate this problem further by examining the second messenger targets for a family of neuropeptides, with diverse physiological actions, which are encoded on a single gene in Lymnaea stagnalis, the FMRFamide gene (Benjamin and Burke, 1994). The peptides, encoded on exon II of the FMRFamide gene in L. stagnalis, were separated into three separate classes on the basis of their amino acid number and physiological effects: (1) FMRFamide and FLRFamide, (2) EFLRIamide, pQFYRIamide and pQFLRIamide (referred to as the ‘isoleucine’ peptides) and (3) SEQPDVDDYLRDVVLQSEEPLY (known as ‘SEEPLY’). pQFLRIamide is not encoded by the FMRFamide gene (Kellett et al., 1994; Santama et al., 1995), but as this peptide is present in both the motoneurones (E_he cells) that provide the FMRFamidergic innervation of the heart and the heart tissue itself (Worster et al., 1999), its effects were also examined in the present study.

The previous paper showed a clear role for the calcium-mobilizing messenger, inositol-1,4,5-trisphosphate (Ins(1,4,5)P₃), in mediating the excitatory effects of FMRFamide and FLRFamide in the heart (Willoughby et al., 1999). However, this messenger played no direct role in the mechanism of action of the two other classes of neuropeptides, which are encoded on the same exon, and are known to be present in the heart tissue. Two possible alternative second messenger targets were 3′,5′-cyclic adenosine monophosphate (cyclic AMP) and 3′,5′-cyclic guanosine monophosphate (cyclic GMP). In the present paper we examined the ability of the three classes of exon II peptide encoded on the FMRFamide gene of L. stagnalis to regulate these two messenger systems. We attempted to relate changes in the activity of these second messenger pathways to physiological actions of the neuropeptides on the heart. The cyclic nucleotides have been linked to the activity of numerous neuropeptides. Vasoactive intestinal peptide (VIP), vasopressin, secretin, small cardioactive peptide (SCP) and myomodulin have all been shown to upregulate the production of cyclic AMP (Lloyd et al., 1985; Patterson et al., 1989; Weiss et al., 1992). The opioid peptides, somatostatin and the locust F1 peptide have been shown to suppress the production of cyclic AMP (Patterson et al., 1989; Baines et al., 1995). Neuropeptides that induce increases in cyclic GMP production include bradykinin, angiotensin II, neurotensin and atrial natriuretic peptide (ANP) (Reiser et al., 1984; Gilbert et al., 1986; Schmidt et al., 1993). These neuropeptides were shown to stimulate cyclic GMP levels indirectly, via the release of calcium from intracellular stores (causing stimulation of nitric oxide synthase), or directly by interacting with the particulate guanylate cyclase (GC) enzyme. In this study we examine the effects of the peptides on the particulate GC enzyme only (not the soluble GC) using a heart membrane preparation.

Cyclic AMP has been shown to produce positive inotropic and chronotropic effects on mammalian, molluscan and insect hearts (Hartzell and Fischmeister, 1986; S.-Rózsa, 1980). In many molluscan hearts there is a correlation between the positive inotropic and chronotropic effects of 5-HT and SCP_b with increased cyclic AMP levels (Higgins, 1974; Mandelbaum et al., 1979; Sawada et al., 1984; Lloyd et al., 1985; Reich et al., 1997). In the mammalian heart acetylcholine stimulates cyclic GMP levels, or reduces cyclic AMP production, to produce negative inotropic effects (Tohse et al., 1995; Hartzell and Fischmeister, 1986). However, the effects of cyclic GMP in the invertebrate heart are poorly documented. In hearts from Locusta migratoria and Helix pomatia high doses of cyclic GMP (10 mmol l⁻¹) produced inconsistent effects on heart beat (S.-Rózsa, 1980). In L. stagnalis the effects of cyclic AMP and cyclic GMP on the heart have not been examined.

Although nothing was known previously about the effects of the ‘isoleucine’ neuropeptides or SEEPLY on cyclic AMP or cyclic GMP levels there is evidence in molluscs that FMRFamide often acts via cyclic AMP. It has been shown that FMRFamide stimulates cyclic AMP levels and enhances contractions in the bivalve heart and in gill tissue of Aplysia californica (Higgins et al., 1978; Cawthorpe et al., 1985; Weiss et al., 1985), and FMRFamide evokes a cyclic AMP-mediated decrease in K⁺ conductance in the E11 neurone of Helix aspersa (Colombaioni et al., 1985).

Two main approaches were used in the present investigation. First, the ability of FMRFamide, EFLRIamide, pQFYRIamide, pQFLRIamide and ‘SEEPLY’ to alter the rate of cyclic AMP and cyclic GMP production in a heart membrane preparation from L. stagnalis was analysed biochemically using a radioimmunoassay (RIA) technique. Second, a series of pharmacological experiments were performed using the intact whole-heart preparation to compare the effects of the cyclic nucleotides and peptide application.

Lymnaea stagnalis (25–40 mm shell length) were supplied by Blades biological (Kent, UK), or the Biology Department of the Vrije Universiteit (Amsterdam). All snails were maintained at 20 °C and subjected to a 12 h:12 h L:D cycle for at least 14 days prior to experiments. Animals were fed ad libitum on lettuce.

FMRFamide and FLRFamide were purchased from Sigma (Dorset, UK). The peptides EFLRIamide, pQFYRIamide, pQFLRIamide and SEQPDVDDYLRDVVLQSEEPLY (‘SEEPLY’) were synthesized in our laboratory, and checked for purity by HPLC. All other pharmacological agents were purchased from Sigma unless stated otherwise.

Hepes-buffered saline (HBS) (Benjamin and Winlow, 1981) contained (in mmol l⁻¹): NaCl (24), KCl (2), CaCl₂ (4), MgCl₂ (2), NaH₂PO₄ (0.1), NaOH (35), glucose (1), Hepes (50), pH 7.9. Tris-DTT buffer (in mmol l⁻¹): Tris-HCl (7.5), DTT (1), pH 7.0. Tris-DTT-EDTA buffer (in mmol l⁻¹): Tris-HCl (7.5), DTT (1), EDTA (1), pH 7.0. Assay solution (in mmol l⁻¹): Tris-acetate buffer (375) pH 7.6, IBMX (0.5), magnesium acetate (50), guanosine 5′-triphosphate (0.25). The IBMX acts as a phosphodiesterase inhibitor and prevents the breakdown of both cyclic AMP and cyclic GMP.

Hearts were rapidly dissected from snails, washed in normal HBS and homogenized in 1 ml of ice-cold Tris-DTT-EDTA buffer. 8–10 hearts were pooled to provide sufficient material for a single experiment. The homogenate was centrifuged at 13 000 revs min⁻¹ (approximately 8000 g) for 5 min at 4 °C, after which the supernatant was discarded and the pellet resuspended in 1.5 ml of Tris-DTT-buffer. The homogenate was centrifuged and resuspended twice more as described above. After each centrifugation the supernatant was removed. The resulting membrane homogenate was stored for up to 4 months at −80 °C until required. At the time of use the membrane homogenate was resuspended in 1.5 ml of HBS, from which 200 μl samples of heart homogenate were taken. All samples were kept on ice during the addition of 80 μl of the assay solution and 100 μl of peptide or HBS (control). 20 μl of 50 mmol l⁻¹ ATP or GTP was added to each sample, at 20 °C, to begin the reaction (ATP was used to stimulate cyclic AMP production, whilst GTP stimulated the production of cyclic GMP). Samples of 50 μl were taken from each of the samples at 0, 5, 10, 20, 45, 60 or 120 s incubation time, and pipetted into Eppendorf tubes in a hot-water bath, maintained at approximately 70 °C, to stop the reaction. Cyclic nucleotides (cyclic AMP and cyclic GMP) were extracted by keeping the 50 μl samples in the hot-water bath for a further 5 min, and then probe-sonicating for 3–4 s. The samples were finally centrifuged for 5 min at 13 000 revs min⁻¹ (approximately 8000 g) and their supernatants were diluted in 200 μl of 0.5 mmol l⁻¹ sodium acetate buffer. These samples were then assayed to determine cyclic AMP or cyclic GMP production using cyclic-nucleotide-specific RIAs.

50 μl samples of cyclic nucleotide standards or sample supernatants were pipetted in duplicate and acetylated with 5 μl (for cyclic AMP) or 2.5 μl (for cyclic GMP) of a mixture of triethylamine:acetic anhydride (2:1) to improve assay sensitivity. 50 μl of antibody and 50 μl of ¹²⁵I-labeled cyclic nucleotide were then added to each standard or sample duplicate. The antibodies (provided by the Biochemistry Department, University of Sussex, UK) were prepared in rabbits against succinyl cyclic AMP/cyclic GMP-conjugated albumin. Cross-reactivity of the antisera with other cyclic nucleotides was less than 0.1 %. In addition, duplicate tubes were set up containing labeled cyclic nucleotide only (total count), and labeled cyclic nucleotide plus 0.5 mol l⁻¹ sodium acetate buffer only (non-specific binding). The tubes were incubated for a minimum of 18 h at 4 °C. Bound and free label were then separated in each tube (with the exception of the total count duplicate) following the addition of 4 ml polyethylene glycol/gamma globulin mix and a 20 min centrifugation at 30 000 revs min⁻¹ (approx. 20 000 g) at 4 °C. After centrifugation the tubes were inverted to pour off the supernatant and left to drain on absorbent paper for a minimum of 1 min. The tops of the tubes were blotted dry by gentle tapping of the inverted tubes against the tissue. The amount of bound cyclic nucleotide label in the resulting pellet, and the amount of radioactivity in the ‘total counts’, was determined using an RIA CALC program and an LKB-Pharmacia multi-gamma counter. Each sample was counted for 1 min and its cyclic nucleotide content was then determined by comparison with a standard curve.

The protein content (mg ml⁻¹) of each homogenate preparation used for the cyclic nucleotide radioimmunoassay was determined using the Bio-Rad protein assay (Bradford method). Bovine plasma gamma globulin was used as standard. 50 μl of each protein standard or sample of unknown protein content was pipetted into a cuvette with 450 μl distilled water and then 100 μl of the dye-reagent. The contents were then mixed by inversion. After approximately 15 min of incubation the absorbance at 595 nm was measured against the ‘blank’ sample (i.e. zero protein content) using a spectrophotometer. Unknown protein concentrations were determined using a standard curve. It was common for the samples to require dilution prior to the protein assay so that their protein content would fall within the working region of the standard curve.

Hearts were removed from the snails and mounted in an experimental perfusion chamber for isotonic recording as described in Willoughby et al. (1999). The time of drug solution perfusion varied between 30 s and 2 min, depending on the type of experiment. Normal HBS was perfused between each application of drug for a minimum of 8 min to allow the heart to return to a steady beat rate. In addition, the outside of the heart was constantly perfused with HBS via a peristaltic pump. Contractions were permanently recorded on a brush chart recorder (Gould Instruments, Hainault, UK). A wide range of peptides and drugs were applied to the heart. All agents tested were dissolved in HBS and adjusted to pH 7.9.

Data are reported as mean ± S.E.M. Means were compared using the Student’s t-test. P<0.05 was considered significant and N was the number of experiments performed in the RIAs, or the number of hearts in the pharmacological experiments.

The ‘isoleucine’ peptides (EFLRIamide, pQFYRIamide and pQFLRIamide) and FMRFamide, applied separately, all significantly increased the rate of cyclic AMP production in homogenized heart membrane preparations from L. stagnalis (Fig. 1). Concentrations of 1 μmol l⁻¹ FMRFamide (N=21), 1 μmol l⁻¹ EFLRIamide (N=21) and 1 μmol l⁻¹ pQFYRIamide (N=15) produced similar maximum increases in the rate of cyclic AMP production (AC activity) of 18.9±2.9, 19.7±2.5 and 19.4±2.9 pmoles cyclic AMP min⁻¹ mg⁻¹ protein, respectively, at 5 s of peptide incubation. These values were significantly greater than control levels of AC activity (P<0.001), which remained at approximately 5–8 pmoles cyclic AMP min⁻¹ mg⁻¹ protein throughout the 2 min incubation period. During peptide incubation AC activity steadily declined from the peak value at 5 s and slowly returned towards control activity levels over the 120 s incubation period. The third ‘isoleucine’ neuropeptide, pQFLRIamide, also stimulated AC activity significantly at 5 s of peptide incubation (P<0.01, N=6) but was less potent than the other peptides. The peptide pQFLRIamide produced a maximum increase in the rate of cyclic AMP production to 13.8±2.4 pmoles cyclic AMP min⁻¹ mg⁻¹ protein (Fig. 1). 10 μmol l⁻¹ SEEPLY (N=11) did not significantly affect the rate of production of cyclic AMP in L. stagnalis heart membrane homogenate (Fig. 1). The adenylate cyclase (AC) activator, forskolin (FSK) (1 μmol l⁻¹) was used as a positive control for cyclic AMP production and caused a fivefold increase in AC activity. At 5 s of forskolin incubation the rate of cyclic AMP production was increased to 33.4±3.3 pmoles min⁻¹ mg⁻¹ protein (P<0.001, N=6) (Fig. 1). A similar rate of cyclic AMP production was maintained throughout the remainder of the 2 min incubation period with 1 μmol l⁻¹ forskolin.

The data analysed in Fig. 1, which show a clear stimulation of the rate of cyclic AMP production by FMRFamide and the ‘isoleucine’ peptides, were obtained from a heart membrane homogenate with control rates of cyclic AMP production of below 10 pmoles min⁻¹ mg⁻¹ protein. Other assays were performed on heart membrane with higher control AC activity (22.5±4.1 pmoles cyclic AMP min⁻¹ mg⁻¹ protein, N=14). However, there was no increase in AC activity in response to the peptides in these batches of tissue homogenate, suggesting that the stimulatory effects of the peptides were limited to tissue in which control AC activity levels were below 10 pmoles min⁻¹ mg⁻¹ protein. This is supported by the data in Fig. 2, which shows scatter plots of the percentage increase in total cyclic AMP production evoked by a 2 min incubation with each of the four peptides, plotted against control AC activity. Analysis of the data in this manner showed an apparent maximum rate of basal cyclic AMP production in the range of 10–15 pmoles min⁻¹ mg⁻¹ protein; above this level the peptides were unable to stimulate cyclic AMP production. Importantly, in this heart membrane homogenate with high AC activity, 1 μmol l⁻¹ forskolin was still able to increase the rate of cyclic AMP production by almost threefold to 62.3±7.4 pmoles min⁻¹ mg⁻¹ protein (P<0.01, N=3). Hence, although the tissue was apparently at ‘saturation’ for the rate of cyclic AMP production to be increased by the peptides, the AC activity could still be directly stimulated by forskolin.

Fig. 3 illustrates the ability of FMRFamide (N=6) (Fig. 3A), EFLRIamide (N=6) and pQFYRIamide (N=6) (Fig. 3B), to stimulate the rate of cyclic AMP production in a concentration-dependent manner, over a concentration range known to produce physiological effects in the heart. Tissue membrane homogenates were incubated with the peptide for 1 min and total cyclic AMP production over the 1 min period was measured. The threshold peptide concentration necessary to produce an increase in the rate of cyclic AMP production was approximately 10 nmol l⁻¹. Peptide concentrations of 100 μmol l⁻¹ (the highest concentration tested) produced a significant increase in the rate of cyclic AMP production for all three peptides when compared to controls (P<0.05, N=6). The increases in cyclic AMP production rate seen in this set of experiments, particularly for EFLRIamide and pQFYRIamide, were less than those predicted from earlier experiments in which heart membrane homogenate was incubated for various times with 1 μmol l⁻¹ concentrations of the peptides (Fig. 1). This might be accounted for by the control rate of cyclic AMP production lying close to ‘saturation’ point. Dose–response data for SEEPLY (N=6) (Fig. 3A) showed no significant changes in the rate of cyclic AMP production compared to controls.

In summary, it appears that the ‘isoleucines’ and FMRFamide might mediate some of their physiological effects via cyclic AMP. In the case of FMRFamide, this would be an additional mechanism to the stimulation by this peptide of the inositol phosphate signalling pathway (Willoughby et al., 1999).

Previous studies on the pharmacological actions of the ‘isoleucine’ peptide EFLRIamide on isolated whole-heart preparations indicated that it produced an initial slowing of heart beat followed by a longer lasting increase in beat frequency when perfused through the lumen of the L. stagnalis heart (Santama et al., 1995). More detailed results on the actions of the ‘isoleucine’ peptides (EFLRIamide and pQFYRIamide) were obtained in this study and are presented in Figs 4, 5. The effects of the ‘isoleucine’ peptides proved to be complex and variable, but could be summarized as dual action, causing both inhibitory and excitatory effects on heart beat. The biphasic responses to both EFLRIamide (N=10) and pQFYRIamide (N=5) (Fig. 4A) were comparable to those described by Santama and colleagues (1995). During the inhibitory period the force of contraction was reduced and the heart beat slowed for a few seconds (Fig. 4Ai, I). The excitatory phase was more delayed and prolonged by comparison (Fig. 4Ai, E). During this phase beat rate was increased and the tonus and force of contraction were also increased in some preparations. In other hearts EFLRIamide produced purely inhibitory (N=13) (Fig. 4Bi) or purely excitatory responses (N=8) (Fig. 4Ci), suggesting that the other type of response was absent or too weak to influence heartbeat. Similar inhibitory (N=3) (Fig. 4Bii) or excitatory responses (N=10) (Fig. 4Cii) in the heart were produced by pQFYRIamide. The production of an inhibitory response by pQFYRIamide was less likely and was always weak compared with that evoked by EFLRIamide (Fig. 4Bi). The type of response seen for any particular heart was consistent for repeated applications of the ‘isoleucine’ peptides. Detailed dose–response data for the inhibitory and excitatory components of the response seen during 1 min applications of EFLRIamide (N=31) and pQFYRIamide (N=18) for all of the whole-heart preparations tested are shown in Fig. 5.

Statistical analysis of the data revealed that threshold concentrations of 100 nmol l⁻¹ EFLRIamide and 1 μmol l⁻¹ pQFYRIamide were required to cause a slowing of heart beat (Fig. 5A). The maximal decrease in beat frequency was seen during a 1 min application of 100 μmol l⁻¹ EFLRIamide (highest concentration tested) (P<0.01), or 10 μmol l⁻¹ pQFYRIamide (P<0.05). Higher concentrations of pQFYRIamide (100 μmol l⁻¹) produced purely excitatory responses. Increases in beat rate during the excitatory phase (Fig. 5D) were seen at concentrations ⩾100 nmol l⁻¹ for both peptides, and were maximal during the application of 100 μmol l⁻¹ EFLRIamide and pQFYRIamide (P<0.01), the highest concentrations tested. The effects of the peptides on beat amplitude were only seen during the inhibitory phase of the response (Fig. 5B). Maximal decreases in the beat amplitude were produced by 10 μmol l⁻¹ EFLRIamide or pQFYRIamide (P<0.01). Statistical analysis showed that beat amplitude was not significantly changed during the excitatory phase of the ‘isoleucine’ response (Fig. 5E), despite apparent increases in some preparations (in 5 out of 31 preparations for EFLRIamide; Fig. 4Ai, or in 4 out of 18 preparations for pQFYRIamide). During the inhibitory phase, EFLRIamide (10 μmol l⁻¹ maximal response), but not pQFYRIamide, significantly reduced the underlying tonus of the isolated heart muscle (P<0.01) (Fig. 5C). In contrast, both peptides were capable of increasing the underlying tonus of the heart during the delayed excitatory phase when EFLRIamide or pQFYRIamide were applied at concentrations >1 μmol l⁻¹ (Fig. 5F). This ability of pQFYRIamide and EFLRIamide to increase tonus was significantly different when compared to control values (P<0.05 and P<0.01, respectively).

It was hypothesized that the delayed excitatory actions of the ‘isoleucine’ peptides were related to the production of cyclic AMP, a common mediator of excitatory effects in the heart (Hartzell and Fischmeister, 1986; Lloyd et al., 1985). As predicted, hearts that produced a delayed excitatory response to 10 μmol l⁻¹ EFLRIamide (N=9) showed an increase in the rate of cyclic AMP production in a membrane preparation of the hearts. AC activity peaked at 5 s of EFLRIamide incubation, and remained higher than that seen in control tissue throughout the 2 min incubation period (Fig. 6A). Hearts that showed no delayed excitatory response to EFLRIamide perfusion (N=11) showed no increase in AC activity compared to control (Fig. 6B). The rate of cyclic AMP production was determined, as described earlier, following examination of the pharmacological response of the isolated hearts to 10 μmol l⁻¹ EFLRIamide. Hearts producing an excitatory or no excitatory pharmacological response were divided into the two groups and pooled. The hearts within each group were combined to provide sufficient tissue homogenate for a single RIA sample at 0, 5, 10, 20, 45 and 120 s of EFLRIamide incubation. The data from these two groups suggested that the potentiation of heart beat was dependent on control AC activity in the heart membrane homogenate. The membrane preparation from hearts showing excitatory responses to EFLRIamide exhibited, on average, a control rate of cyclic AMP production of approximately 5 pmoles min⁻¹ mg⁻¹ protein. In contrast, a membrane preparation of those hearts showing no excitatory response to EFLRIamide had a greater control rate of cyclic AMP production, around 15 pmoles min⁻¹ mg⁻¹ protein. This data is consistent with that shown in Fig. 2 and supports the hypothesis that the increased rate of cyclic AMP production mediates the excitatory effects of the ‘isoleucine’ peptides. In contrast, the inhibitory effect of the isoleucines, shown for EFLRIamide in Fig. 6B, did not appear to be mediated by cyclic AMP. An interesting observation from these experiments was that hearts in which AC activity was higher (15 pmoles min⁻¹ mg⁻¹ protein) had a higher resting beat rate (23±2 beats min⁻¹) than those in which AC activity was lower (11±2 beats min⁻¹ on average). This suggests that cyclic AMP might also have some role in the controlling the heart of L. stagnalis at rest.

To provide further evidence that cyclic AMP was responsible for the cardioexcitatory effects of the ‘isoleucine’ peptides and FMRFamide, the pharmacological effects of the peptides in the isolated heart were compared with those mediated by a membrane-permeable analogue of cyclic AMP, 8-bromo-cyclic AMP (8-Br-cyclic AMP). As predicted 8-Br-cyclic AMP produced excitatory effects on the heart of L. stagnalis that were comparable to the delayed excitatory effects seen following application of the peptides (Fig. 7). Following a 10 s delay, 1 mmol l⁻¹ 8-Br-cyclic AMP mediated increases in beat frequency and force of contraction of the isolated heart (Fig. 7D). These effects were highly comparable to the excitatory phase of the response of the heart seen following a 30 s perfusion of 10 μmol l⁻¹ EFLRIamide (Fig. 7B) and 10 μmol l⁻¹ pQFYRIamide (Fig. 7C) on the heart. The cyclic AMP analogue also mimicked some of the customary excitatory effects seen during a 30 s perfusion of the isolated heart with 1 μmol l⁻¹ FMRFamide (Fig. 7A), except that it was slower to act than FMRFamide. An inhibitory response to 8-Br-cyclic AMP was never observed. Similarly, perfusion of the AC activator, forskolin, stimulated the heart and produced increases in both the frequency and the amplitude of heart beat (Fig. 7E). The effects of 8-Br-cyclic AMP in the heart of L. stagnalis were concentration-dependent (Fig. 8, N=11). The graphs illustrate the ability of 8-Br-cyclic AMP to produce cardioexcitatory effects during 2 min applications of a range of concentrations (1 μmol l⁻¹ to 1 mmol l⁻¹). This cardioexcitation by cyclic AMP, or its analogues, is consistent with studies in other molluscan hearts (Higgins, 1974; Mandelbaum et al., 1979; Sawada et al., 1984).

Activity of the particulate (membrane bound) GC enzyme was investigated using the heart membrane preparation. SEEPLY, FMRFamide and the ‘isoleucine’ peptides had no significant effect on the rate of cyclic GMP production in heart membrane preparations from L. stagnalis.

This study of peptidergic regulation of the heart in L. stagnalis provides an important insight into the diversity of peptide action in a well-characterized system. The results presented in this paper suggest that cyclic AMP is an important second messenger mediating the effects of the ‘isoleucine’ peptides and FMRFamide in the heart. In the case of FMRFamide, this mechanism would be additional to the activation of inositol phosphate second messengers (Willoughby et al., 1999). The peptides EFLRIamide, pQFYRIamide and FMRFamide all significantly increased the rate of cyclic AMP production over control levels in responsive tissue. The increases in AC activity were most prominent within the first 5 s of peptide incubation. The third ‘isoleucine’ peptide, pQFLRIamide, which is not encoded by exon II of the FMRFamide gene but is present in the heart (Santama et al., 1995; Worster, 1999), also stimulated the rate of cyclic AMP production in heart membrane homogenate, but was less potent than the other peptides.

The stimulatory effects of EFLRIamide, pQFYRIamide, pQFLRIamide and FMRFamide on AC activity were only apparent when control rates of cyclic AMP production were, on average, below 10 pmoles min⁻¹ mg⁻¹ protein. In batches of heart membrane homogenate with higher rates of cyclic AMP production the peptides did not enhance further cyclic AMP production. In homogenate where the peptides had no effect on the rate of cyclic AMP production the ‘high’ basal rates of AC activity were comparable to the maximal rates of cyclic AMP production produced by the peptides in tissue with ‘low’ basal AC activity. Forskolin significantly increased the rate of cyclic AMP production in heart membrane homogenates with both ‘low’ and ‘high’ levels of basal AC activity. The actions of forskolin in this tissue were presumably by direct stimulation of the AC enzyme. This suggested that ‘saturation’ of the peptide-stimulated cyclic AMP production might occur upstream of the AC enzyme, for example at the G-protein or at the peptide receptor itself. Alternatively, it is possible that there are AC enzymes not associated with the ‘isoleucine’ receptor(s) that will still be stimulated by forskolin. In support of the concept of saturation of second messenger production upstream from the AC enzyme are recent findings by Freedman et al. (1996), suggesting that high levels of intracellular cyclic AMP are capable of uncoupling the β-receptor from its stimulatory G-protein (G_s). The variability in control rates of cyclic AMP production between different batches of snails cannot be explained at present.

The pharmacological effects of 8-Br-cyclic AMP (a membrane-permeant cyclic AMP analogue) and forskolin (an AC activator) were comparable to the delayed excitatory effects of the ‘isoleucine’ peptides. The excitatory effects of 8-Br-cyclic AMP and forskolin were also comparable to the excitatory effects seen during perfusion of the isolated heart with FMRFamide. It seemed likely, therefore, that an increase in cyclic AMP production might be responsible for the excitatory effects on heartbeat seen during application of both types of these peptides. The most striking similarity between the cyclic AMP analogues and FMRFamide, EFLRIamide and pQFYRIamide was their ability to increase beat frequency. The maximal increase in beat frequency, of approximately 60–80 % on average compared to the control beat rate, could similarly be achieved by applying cyclic AMP analogues or any of the individual peptides.

Detailed pharmacological effects of the ‘isoleucine’ peptides, EFLRIamide and pQFYRIamide, on the heart of L. stagnalis were described. The ‘isoleucine’ peptides appeared to have dual activity, producing complex biphasic responses in the heart consisting of both an inhibitory and an excitatory component. Similar data for the effects of EFLRIamide on the heart of L. stagnalis have been reported previously by Santama et al. (1995), and in the related snail, Helix aspersa by Lesser and Greenberg (1993). The inhibitory effects of the ‘isoleucines’ dominated the early phase of the response to peptide application, reducing frequency and amplitude of heart beat in particular. This was followed by a more prolonged increase in beat frequency and tonus that appears to be mediated by increased cyclic AMP production. Our RIA data show a peak increase in the rate of cyclic AMP production after 5 s of peptide incubation with heart membrane homogenate. However, during such experiments the peptides have direct access to membrane-associated receptors. It is likely that the peak increase in cyclic AMP production occurs later, when the peptides are perfused through the intact heart. Indeed, this would be more consistent with our pharmacological data, where the peptides often take longer to produce a clear cardioexcitatory response.

In several of the hearts only inhibitory or excitatory effects were observed. Our data show a clear relationship between an increased rate of cyclic AMP production and the excitatory phase in hearts with relatively ‘low’ control AC activity levels. The rate of cyclic AMP production in membrane homogenate from hearts showing no excitatory response to the ‘isoleucine’ peptides was not stimulated further by the application of peptides.

At present the inhibitory phase of the ‘isoleucine’ response cannot be accounted for by any of the second messengers analysed in this or the previous paper, and might be due to direct effects of the peptides at an ion channel. Green et al. (1994) have identified two types of amiloride-sensitive Na⁺ channel in the C2 neuron of Helix that are directly gated by FMRFamide. A homologue of this channel has recently been cloned in our laboratory in L. stagnalis. Although activity of the FMRFamide-gated Na⁺ channel could not account for the inhibitory effects of the isoleucines, it is possible that a novel isoleucine-gated channel might also exist, which could account for the inhibitory response in the heart.

The potential for FMRFamide to stimulate the production of cyclic AMP and inositol-1,4,5-trisphosphate (Ins(1,4,5)P₃) (Willoughby et al., 1999) suggested that there might be two different receptors for FMRFamide linked to second messenger pathways in the heart of L. stagnalis. As yet the FMRFamide receptor(s) of the heart of L. stagnalis have not been characterized, although a Gs-coupled FMRFamide receptor has been isolated in the squid (Chin et al., 1994). The identification of both a Gs-and a Gq-coupled FMRFamide receptor in the pond snail would account for the ability of the peptides to increase the production of both cyclic AMP and Ins(1,4,5)P₃ in the heart. Alternatively, there might be a single class of FMRFamide receptor in the heart of L. stagnalis that is linked to two different G proteins, to give the subsequent activation of two separate signalling pathways. Wang et al. (1995) suggested that both inhibitory and excitatory FMRFamide-related peptides (FaRPs) in the locust oviduct muscle might act via a single receptor linked to two different G proteins. However, definitive evidence for a single receptor in the intact muscle, rather than FaRPs actions via two pharmacologically similar receptor subtypes, was not presented. In vertebrates, β-adrenergic receptors have been shown to couple with both G_s and G_i (Abramson et al., 1988; Xiao et al., 1995), whilst the human thyrotropin receptor couples to members of four G-protein families (Laugwitz et al., 1996). Finally, it is also possible that the activation of one second messenger pathway by a single transmitter substance stimulates the production of another second messenger. Studies in a range of neuronal cell types have identified three isoforms of AC that are stimulated by elevated levels of intracellular calcium (Cooper et al., 1995), such as the increase that might be seen following Ins(1,4,5)P₃ production. Alternatively, studies by Prier et al. (1994) on heart of Manduca sexta showed that the effects of the cardioacceleratory peptides (CAP₂s), which act via the production of Ins(1,4,5)P₃, may be upregulated via an octopamine-sensitive cyclic AMP-dependent mechanism. The authors suggested a multiplicative interaction of the two pathways at the site of intracellular calcium release.

In L. stagnalis, we believe that the activation of two different signalling pathways, probably acting in parallel, can account for FMRFamide-induced cardioexcitation. Noradrenaline similarly activates these two signalling pathways in the vertebrate heart to mediate cardioexcitation (Irisawa et al., 1993; Hartzell and Fischmeister, 1986; Difrancesco and Tortora, 1991).

The rate of production of cyclic GMP via the particulate GC enzyme in the L. stagnalis heart was not significantly effected by any of the peptides encoded by exon II of the FMRFamide gene of L. stagnalis. We have not yet investigated whether the peptides modulate activity of a soluble form of GC linked to the production of nitric oxide, or indeed whether such a form of GC enzyme is present in the heart of L. stagnalis. Perfusion of isolated whole heart preparations with a membrane-permeable cyclic GMP analogue (8-bromo-cyclic GMP) (N=6, data not shown) produced a relaxation of the heart. In particular, 8-bromo-cyclic GMP mediated a reduction in the underlying tonus of the muscle. It is possible that the inhibitory effects of the ‘isoleucine’ peptides might be mediated by stimulation of the soluble GC and cyclic GMP production, but this has not yet been investigated.

In summary, the results obtained from this study provided sufficient evidence to indicate activation of at least two second messenger pathways in the molluscan heart by peptides encoded on exon II of the FMRFamide gene of L. stagnalis. The excitatory effects of the ‘isoleucines’ are likely to be mediated by enhanced cyclic AMP production, while FMRFamide produces more potent excitatory effects through the apparent activation of both cyclic AMP-and InsP₃-producing pathways. The modulatory peptide, SEEPLY, has no significant direct effect on cyclic nucleotide and inositol phosphate production but could modify the production of these second messengers by other exon II peptides. Other work in our laboratory has suggested that the exon II-encoded peptides of the FMRFamide gene may be coreleased from the heart. The functional consequences of the release of such a ‘cocktail’ of all three exon II-encoded peptide classes are at present undetermined, but are likely to be complex given the potential for interaction at the second messenger level.

We are grateful to the BBSRC for funding, and would like to thank Dr Irene Green (Biochemistry Department, University of Sussex) for help with the cyclic nucleotide radioimmunoassay.