Localization of FMRFamide-like peptides in the snail Helix aspersa

Cardioactive substances other than the classical transmitters (acetylcholine, 5-hydroxytryptamine and the catecholamines) have been detected frequently in molluscan ganglia (Frontali, Williams & Welsh, 1967; Kerkut & Laverack, 1960; Jaeger, 1966; Lloyd, 1978; Geraerts et al. 1984). Among these compounds is the family of FMRFamide-like peptides. FMRFamide itself was first isolated and sequenced from a clam, Macrocallista nimbosa, but has since been found in a number of gastropod species, including Aplysia brasiliana, Lymnaea stagnalis and Pomacea paludosa (Price & Greenberg, 1977; Lehman, Price & Greenberg, 1984; Price, 1986). Helix aspersa, a stylommatophoran pulmonate snail, contains FMRFamide, the heptapeptide pyroGlu-Asp-Pro-Phe-Leu-Arg-Phe-NH₂ (pQDPFLRF-amide) and other related peptides (Price et al. 1985).

The actions of FMRFamide-like peptides have been well studied on a variety of molluscan muscles. First, the heart of Mercenaria mercenaria and the radula protractor muscle of Busycon contrarium have been ideal assay tissues due to their sensitivity, specificity and selectivity (Greenberg, 1983). In addition, the effects of FMRFamide have been surveyed on: a vast collection of bivalve hearts (Painter & Greenberg, 1982); the Mercenaria rectum (Doble & Greenberg, 1982); the gill, anterior gizzard and various neurones of Aplysia (Weiss et al. 1984; Austin, Weiss & Lukowiak, 1983; Stone & Mayeri, 1981); the auricle, ventricle, oesophagus and penis retractor muscle of Lymnaea (Geraerts et al. 1984); the octopus heart (Wells, 1983); and the Helix tentacle muscle (Cottrell, Greenberg & Price, 1983a; Cottrell, Schot & Dockray, 1983b). Finally, the cellular mechanisms of action of FMRFamide have been investigated, primarily with clam hearts (Higgins, Price & Greenberg, 1978), the anterior byssus retractor muscle of mytilid mussels (Painter, 1982), and specific neurones of Helix aspersa (Cottrell et al. 1983 b).

In summary, FMRFamide affects all types of molluscan excitable tissue: heart, non-cardiac muscle and nerve. Notwithstanding the long list of its pharmacological actions, the physiological role of FMRFamide remains unknown. One hypothesis is that the peptide plays a cardioregulatory role in bivalves. This notion is based upon its cardioactive properties, its presence in the heart, and its presence in the blood (Greenberg & Price, 1980; Nagle, 1982). However, although FMRFamide has pronounced effects on the molluscan myocardium, other physiological sites of action have been proposed (Greenberg & Price, 1980). For example, FMRFamide may be the transmitter or modulator at particular neuromuscular junctions (Cottrell et al. 1983b; Austin et al. 1983).

As an approach to learning the physiological role of FMRFamide, this report describes the localization of FMRFamide-related immunoreactivity in the organs and tissues of Helix aspersa, while in the subsequent paper (Lehman & Greenberg, 1987) the distribution is correlated with pharmacological responsiveness and known functions. A radioimmunoassay for FMRFamide was used to determine the concentrations of FMRFamide-related activity in 23 tissues of Helix. Immunocytochemical methods were employed to examine the density of innervation, the types of cells innervated, and the locations of FMRFamidergic cell bodies; i.e. the probable sites of production of these peptides. The FMRFamide-like peptides of this snails were characterized by their elution volumes in gel chromatography and HPLC, and by their biological activity on the radula protractor muscle of Busycon contrarium.

Helix aspersa were collected near Fullerton, California and sent by express mail to Florida. The snails were aestivating upon arrival, and they remained in this condition until dissected (anatomy described by Leake, 1975; Hyman, 1967).

The snails were processed in groups of 30—100. The first step was to take blood through a syringe needle inserted directly into the ventricle through a hole in the shell. Thereafter each of the 22 parts of interest was dissected out of the animal and placed immediately into ice-cold acetone, a solvent successfully used to extract FMRFamide and its congeners (Lehman et al. 1984; Price et al. 1985; Price, 1986).

The tissues were homogenized for 5 min (Polytron Model PT 10), the homogenate was centrifuged (10000 g, 15 min), and the pellet was re-extracted with 80% acetone. The two acetone supernatants were pooled and evaporated to dryness. The dried material was taken up in 1 ml of distilled water. Ten microlitres was removed, diluted, and these solutions were assayed with a radioimmunoassay (RIA) for FMRFamide.

The immunoreactive peaks recovered after gel filtration were pooled, lyophilized, taken up in buffer and subjected to high-pressure liquid chromatography (HPLC) (Waters Associates Model 720 System, with a μ-Bondapak C-18 reverse-phase column). Two solvent systems were used to characterize the peptides: an isocratic system consisting of 80% butanol-saturated buffer (0·5 mol l^-1 ammonium acetate, 0·1 moll^-1 acetic acid, pH5·5); and a gradient system comprising methanol and 1·0% trifluoroacetic acid (from 30% to 50% methanol in 20min) at pH 2·0. The buffers were pumped at a rate of 2 ml min^-1, and the u.v. absorbance was monitored at 254 nm.

The antiserum was prepared in 1980 as follows. Tyr-Gly-Gly-Phe-Met-Arg-Phe-NH₂ (YGGFMRFamide) was coupled to thyroglobulin with a water-soluble carbodiimide (using the method of Skowsky & Fisher, 1972). The reaction mixture was diluted to approximately lmgml^-1 and emulsified in an equal volume of Freund’s complete adjuvant. Approximately 1 ml of the emulsion was injected into many intradermal sites on the back of each of six rabbits. A booster injection of the same concentration, emulsified in Freund’s incomplete adjuvant, was given 4 weeks later. Further booster injections were given 2 weeks prior to bleeding. An antiserum (S-253) obtained from a single bleeding of one rabbit has been used in all of the experiments described in this report. The structure-activity relationships of the RIA using this antibody have been well studied (Price, 1983). The most general requirement for binding is the minimal sequence Arg-Phe-NH₂; other known peptides, including SCPB and Met-enkephalin, do not significantly cross-react in the RIA (Fig. 1).

YGGFMRFamide was iodinated monthly by the chloramine T method (Greenwood, Hunter & Glover, 1963). Chloramine T (10μg) was added to ¹²⁵I (1·5 mCi) and YGGFMRFamide (1·0 μg) in 10·0 μl of sodium phosphate buffer, pH 7·6. After 15 s, 40μg of sodium metabisulphate was added, and the iodinated YGGFFMRF-amide was separated from free iodide with a Sep-Pak C-18 cartridge.

Each assay tube contained 10μl of diluted antiserum (1:30000), 50μl of diluted trace (7000 counts min^-1) and 50μl of either RIA buffer [1 % bovine serum albumin in phosphate-buffered saline (PBS), 0·01 % merthiolate, 0·0·25 mol l^-1 sodium-EDTA] or sample diluted in the RIA buffer. The tubes were incubated at 4°C overnight, and 1 ml of a charcoal dextran solution (0·01 moll^-1 sodium phosphate, pH7·6; 0·01% merthiolate; 2·5gl^-1 charcoal; and 0·25gl^—I dextran) was added, and centrifuged (4000 rev. min^-1, 10 min, 4°C) to separate the bound tracer from free tracer. Each sample was counted for 1 min and the percentage bound for unknowns was compared to an FMRFamide standard curve.

The various tissues of Helix aspersa were fixed for 4 h with 4 % paraformaldehyde in Helix saline (pH 7·6; Kerkut & Meech, 1965) and washed overnight with 30% sucrose in distilled water.

The fixed tissues were embedded in a gelatin-albumin mixture (35 ml of 6% gelatin, 15 ml of egg albumin) and frozen in liquid nitrogen. Sections (8–10μm thick) were cut on a cryostat, mounted on slides coated with Histostik (Accurate Biochemicals), and frozen overnight. On the following day, the sections were incubated for 30min with FMRFamide antiserum (S-253, diluted 1/200 in PBS), washed for 15 min with PBS, incubated for 30 min with fluorescein-labelled, goat anti-rabbit IgG (Boehringer Mannheim, diluted 1/50 in PBS and washed for 30 min with PBS). Two controls for specificity were applied: the antiserum was preincubated with FMRFamide, pQDPFLRFamide or SCP_B; and selected sections, prepared as described, were stained with non-immune serum.

The distribution of immunofluorescence was examined with a fluorescence microscope (Leitz Dialux 20) with incident light passed through a 436 nm excitation filter and a 490 nm suppression filter. All photographs were taken on Kodak Ektachrome 400 ASA film with a photoautomatic system (Wild MPS 50).

Radula protractor muscles, isolated from the whelk Busycon contrarium and prepared by the method of Hill (1958), were suspended in an organ bath containing aerated, natural sea water. Tension was recorded with a force-displacement transducer connected to an inkwriting oscillograph (Grass Instrument Co.). All drugs were added directly to the bath, and doses of FMRFamide are expressed in terms of their final concentrations in the bath (in moll^-1).

The concentrations of immunoreactive(ir-)FMRFamide in acetone extracts of various tissues were determined (Table 1).

In the central nervous system (CNS), ir-FMRFamide was concentrated in the suboesophageal and cerebral ganglia, and lesser quantities were found in the buccal ganglia. In addition to the material in the CNS, very large amounts of activity were found in the viscera.

In the reproductive system, especially high levels of the peptides occurred in the male reproductive tract. The muscular epiphallus/flagellum complex contained levels exceeding even those in the ganglia, but the penis also possessed large amounts. The common duct, with both male and female components, had modest amounts of ir-FMRFamide, whereas other, particularly female, regions of the reproductive system had relatively low amounts (Table 1).

The distribution of immunocytochemically stained FMRFamidergic fibres in the reproductive system was also not uniform; i.e. the fibres were very dense in the male reproductive tract, but were rarely found in the female system. In the male portions, immunofluorescent fibres and varicosities were particularly localized in both the circular and longitudinal muscle layers in the thick walls of the penis, epiphallus, flagellum and dart sac (Fig. 2A). Only thin, sparsely distributed immunoreactive fibres were found in the vas deferens, and they were limited to the thin musculae sheath. Densely staining aggregates resembling cell bodies occurred between the muscle fibres in the outermost regions of the male tract.

Within the digestive system, the rectum had the highest concentration of ir-FMRFamide, whereas more anterior regions of the tract contained less of the immunoreactive material. Significant levels were also found in the salivary glands (Table 1). The highest density of ir-FMRFamidergic nerves in the digestive system was found in the muscular layers of the rectum. Immunoreactive varicosities occurred in the circular muscle layer, and fibres extended into the muscular villi. No specific immunofluorescence was associated with the cells of the secretory mucosal epithelium of the rectum.

In all other regions of the digestive tract, including the intestine, crop and oesophagus, ir-FMRFamidergic fibres and varicosities were seen in the thin muscular sheath along the entire length of the digestive tract (Fig. 2D). As for the salivary glands, the muscular sheath surrounding the exocrine cells was densely stained.

The auricle contained most of the ir-FMRFamide in the heart; less was detected in the ventricle (Table 1). Immunoreactive FMRFamidergic fibres were scattered throughout the auricle in the outermost longitudinal muscle cell layer and the internal circular muscular layers. The ventricle contained relatively few immunoreactive fibres, but some were occasionally found at the auriculoventricular junction (Fig. 2B).

The concentration of ir-FMRFamide in the blood was determined from three pooled samples from 40 individuals and ranged from 10⁻⁹ to 10⁻⁸moll^-1. These values are in good agreement with those of Price et al. (1985).

The tentacle and pharyngeal retractor muscles are both part of the columnellar muscle group (Jones, 1975), but the concentrations of ir-FMRFamide in these two muscles were very different (Table 1). The tentacle retractor muscle contained the highest concentration of FMRFamide, and the largest number of ir-FMRFamidergic nerve fibres of all the muscles examined in Helix. These fibres were arranged in parallel rows, perpendicular to the longitudinal muscle cells (Fig. 2C). No staining was observed in the optic nerve or retina of the tentacle. Fluorescent fibres were also detected in the buccal musculature, and in the muscles of the foot and body wall. The pharyngeal retractor muscle and penis retractor muscle lacked ir-FMRFamide fibres.

Immunoreactive FMRFamidergic fibres were observed in the muscular epithelium surrounding the kidney, collar, ovo-testis and digestive gland.

When an extract of any tissue of Helix was subjected to gel filtration on a Sephadex G-15 column, the FMRFamide-like immunoreactivity migrated as two distinct peaks (Fig. 3). The first peak (designated ‘peak C’), eluted in 2·5 void volumes and corresponds to the peak C originally described by Frontali et al. (1967). The second immunoreactive peak was retained longer and eluted in approximately 5–6 void volumes, but much earlier than the peak D of Frontali et al. (1967).

Peak C was separated into two major peaks by passage through a C-18 HPLC column in either of the two solvent systems (Fig. 4). The earliest peak eluted quickly and coincidentally with FMRFamide. This early peak migrated as a doublet in the butanol solvent system (Fig. 4A), the minor component eluting slightly before the FMRFamide standard. The second component of peak C appeared much later; it coeluted with pQDPFLRFamide and was indistinguishable from a pQDPFLRFamide standard added to the extract before chromatography (Fig. 5).

The two HPLC peaks recovered from peak C were both active on the radula protractor muscle of Busycon (Fig. 6). Moreover, when doses equiactive on the muscle were quantified by RIA, the peak eluting with FMRFamide contained 7–10 times less peptide (i.e. was more potent) than the peak eluting with pQDPFLRFamide; the potencies of the synthetic peptides are of the same order on this assay system (Price et al. 1985).

FMRFamide and pQDPFLRFamide were not equally distributed in the tissues of Helix (Table 2). Whereas FMRFamide predominated in the reproductive system, retractor muscles and ventricle, pQDPFLRFamide was more abundant in the ganglia and digestive system.

The second peak eluting from the gel column (Fig. 3) migrated on HPLC as a single peak in both the isocratic and the gradient solvent systems; it eluted between FMRFamide and pQDPFLRFamide (Fig. 7). The material recovered from HPLC is immunoreactive, yet its immunoreactivity is not affected by treatment with carboxypeptidase Y, and its binding curve is steeper than that of either FMRFamide or peak C (Fig. 8). Moreover, 500 pmol of this HPLC peak, as determined by RIA, had no effect on the radula protractor muscle of Busycon which is sensitive to 4 pmol of authentic FMRFamide.

The ganglia, male reproductive organs and the posterior end of the digestive tract of Helix contained the highest levels of immunoreactive FMRFamide as determined by RIA, and these organs also had the highest density of immunocytochemically positive neural structures. In the non-neural organs, FMRFamide-like immunoreactivity was confined to muscle tissue and was observed mostly as nerve fibres, though a few nerve cell bodies were seen in the digestive and male reproductive tracts. These results are very similar to those of two previous immunocytochemical studies on Helix aspersa, one on the circumoesophageal ganglia (Marchand, Wijdenes & Schot, 1982) and the other on the peripheral nervous system (Cardot & Fellman, 1983). In particular, Cardot & Fellman found, as we did, that most of the immunoreactivity is contained in muscle tissue, with the digestive and male reproductive systems having especially dense innervations and even some reactive perikarya.

These findings show that muscle is the primary peripheral target of FMRFamidergic innervation. Other studies had indicated that these peptides, serving as transmitters or as modulators, are able to regulate the contractility of Helix muscles (Cottrell et al. 1983a, b; Lehman & Greenberg, 1987). Since we also found relatively high levels of FMRFamide-like immunoreactivity in the blood, some or all of these peptides might also be hormones. We presume that their actions are, again, on muscle, but we cannot rule out other targets.

The FMRFamide-like immunoreactivity extracted from Helix is composed of several chromatographically distinct substances. On Sephadex G-15, two large, broad peaks of immunoreactivity are separated. The earlier, eluting at 2·5–3·0 void volumes, can be identified with the peak C defined by Frontali et al. (1967). Peak C is found in all molluscan species and several FMRFamide-like peptides have been isolated from it, including FMRFamide, FLRFamide and pQDPFLRFamide (Price et al. 1985; Price, 1986). Elution at this position is characteristic of these peptides because of the interaction of their aromatic residues with the gel (Gelotte, 1960).

The second peak on Sephadex G-15 is strongly retarded by the gel; its elution position is somewhat variable (Fig. 3) — about two column volumes (5–6 void volumes) - and it has virtually no mechanical activity on the radula protractor muscle. On HPLC, it elutes as a single peak of immunoreactivity between FMRFamide and pQDPFLRFamide, but its dilution curve in the RIA is steeper than that of FMRFamide or pQDPFLRFamide. Furthermore, it is insensitive to carboxypeptidase Y digestion. The exact nature of this non-specific immunoreactivity is still unknown, but it appears to have none of the biological or chemical features of FMRFamide. PseudoFMRFamide has been found in other molluscan species and has been detected with antisera other than our own (Price, 1986).

Although crude extracts contain pseudoFMRFamide, the data in Table 1 are still useful because the levels of pseudoFMRFamide present in the tissues are roughly proportional to the level of authentic peptide. This relationship is evident in Fig. 3 which shows that where the authentic peptide peak is large (tentacle and suboesophageal ganglion), the pseudoFMRFamide peak is also large, and where the peptide peak is small (crop and heart), the pseudoFMRFamide peak is also small. It is this proportionality which allowed us to use the immunoreactivity of the acetone extracts as a crude measure of the amount of FMRFamide-like peptide contained in the tissues of the snail.

When peak C (i.e. authentic peptide) is chromatographed on a reverse-phase column, two major peaks of immunoreactivity can be identified with known peptides. The first of these is FMRFamide and the second is pQDPFLRFamide. FMRFamide had previously been reported to be absent from Helix, but the work presented here clearly shows that this conclusion was wrong; it also provides an explanation for the erroneous conclusion, as follows.

The absence of FMRFamide was first suggested by Cottrell, Price & Greenberg (1981). Using an extensive extraction and purification protocol, and a bioassay to detect activity, these authors reported that, unlike authentic FMRFamide, the majority of FMRFamide-like activity was not retained by a cation exchange column. Later Price (1982), equipped with a more efficient extraction technique, better purification methods (HPLC) and a more sensitive assay (radioimmunoassay), detected several peaks of FMRFamide-like activity but focused on the sequence of the major peptide, which was finally shown to be pQDPFLRFamide (Price et al. 1985). In the earlier study the extracts were handled extensively, increasing the likelihood that FMRFamide would be oxidized. In this investigation, we handled the material minimally, but still observed a peak on HPLC which preceded that of FMRFamide and co-eluted with oxidized, synthetic FMRFamide. Moreover, this peak had very little biological activity compared with its immunoreactivity (Fig. 6). Therefore, if the majority of FMRFamide in an extract were to be oxidized, and only a bioassay were to be available, the pQDPFLRFamide would be the only peptide detected.

In addition to the major FMRFamide and pQDPFLRFamide peaks we also see some minor peaks on HPLC. One is probably an analogue of pQDPFLRFamide with a different N-terminal amino acid, and another may be FLRFamide (Price et al. 1985 ; Price, 1986). But these minor peaks occur at such low levels that we have not yet isolated enough material to make an unequivocal identification. FLRFamide not only occurs at a very low level, but it also has a low immunoreactivity with our antibody, which compounds the problem of detecting it.

The importance of the antiserum is illustrated by the results of Dockray et al. (1981), who also used an RIA to measure the level of FMRFamide-like immunoreactivity in the circumoesophageal ganglia of Helix. The concentration that they measured in the suboesophageal ganglia (18·2 ± 6·6 nmol g^-1, ±s.e.) was very similar to our determinations: 33·69 ± 27·5 nmolg^-1 (±s.d.) (Table 1) and 15·4-nmolg^-1 (the sum of FMRFamide and pQDPFLRFamide in Table 2). The agreement is particularly surprising because the extraction procedures and the antisera were different, and the antisera even have radically different specificities (Dockray, 1985). In the cerebral ganglia, however, Dockray et al. (1981) found only 4·2 nmol g^-1 of FMRFamide immunoreactivity (about one-quarter of the level in the suboesophageal ganglia), whereas we found that the total immunoreactivity was about equal in the two tissues (Tables 1,2). But we did find that the cerebral ganglia have only one-quarter as much FMRFamide as the suboesophageal ganglia, the balance in the cerebral ganglia being achieved by a higher level of pQDPFLRFamide (Table 2). The two sets of results are reconcilable in terms of known differences in the antisera. Our RIA reacts about equally with pQDPFLRFamide and FMRFamide (Fig. 1; Price et al. 1985; Price, 1986), whereas Dockray’s antiserum should only cross-react 10 % as well with pQDPFLRFamide as with FMRFamide, based on its reactivity with other peptides having a leucine residue in place of the methionine (Dockray, 1985). Therefore, the measurements of Dockray et al. (1981) support the data reported here, and they are also consistent with our finding that the ratio of the levels of FMRFamide and pQDPFLRFamide vary from tissue to tissue (Figs 4A, 5A; Table 2).

The variation in the ratio of FMRFamide and pQDPFLRFamide might reflect differential processing of a single gene product (at either the RNA or the protein level) or the expression of distinct FMRFamide-encoding genes in different anatomical regions. Both mechanisms are known to account for neuropeptide diversity in Aplysia (Scheller et al. 1982; Schaefer et al. 1985). However, the variable ratio might only reflect differential recovery or degradation in the different regions. It is important to clone the genes involved to answer this question more definitively.

This variable ratio suggests that FMRFamide and pQDPFLRFamide might serve different functions, a notion supported by the differential actions of the two peptides on Helix muscle (Lehman & Greenberg, 1987). Another possibility is that one of these peptides is delivered preferentially to the tissues via the blood. Since pQDPFLRFamide has a blocked N-terminal, it is partially protected from enzymatic degradation and, therefore, might serve as a hormone. Moreover, pQDPFLRFamide does occur at pharmacologically active levels in the blood of Helix (Price et al. 1985).

Several workers have examined the actions of FMRFamide on molluscan neurones (Cottrell, Davies & Green, 1984; Stone & Mayeri, 1981), gut (Austin et al. 1983; Doble & Greenberg, 1982) and retractor muscles (Cottrell et al. 1983a; Painter, 1982; Nagle & Greenberg, 1982) and all have suggested that FMRFamide is a neurotransmitter or modulator in these tissues. This study suggests that in addition to their action in those tissues, FMRFamide and pQDPFLRFamide may act as a transmitter or modulator in the mediation of muscular movements of the male reproductive tract. In other molluscs, Aplysia and Lymnaea, immunoreactive FMRFamide is highly concentrated in the male reproductive organs (H. K. Lehman, unpublished observations). Thus, in pulmonates and opisthobranchs, the FMRFamide family of peptides may be important regulators of reproduction, as well as of other central and visceral functions.

This work was supported by NIH grant HL28440 to MJG and is contribution no. 260 from the Tallahassee, Sopchoppy and Gulf Coast Marine Biological Association. The assistance of Lynn Milstead in preparing the manuscript is gratefully acknowledged.