Atrial granular cells of the snail Achatina fulica release proteins into hemolymph after stimulation of the heart nerve

Like the vertebrates, the gastropod molluscs possess a chambered myogenic heart that is controlled by neuronal and humoral factors(Jones, 1983). In these molluscs the coordinated beats of atrium and ventricle underlie both regular action potentials (Zhuravlev et al.,2002) and responsible ion channels(Kodirov et al., 2004). The peculiar feature of the heart of gastropod molluscs is numerous granular cells(GCs) attached to the myocardial surface of the atrium. The GCs have been described in the atrium of the gastropod molluscs Lymnaea stagnalis(Rozsa and Nagy, 1967), Helix pomatia (Erdelyi and Halasz, 1972), Helix aspersa, Strophocheilus oblongus(Volkmer-Ribeiro, 1970) and Achatina fulica (Martynova et al., 2007). These cells were also reported in other organs of molluscs, such as the buccal artery, optic tentacles, connective tissue of the oesophageal ring (Steinbach,1977) and in the vein reticulum of the mantle(Volkmer-Ribeiro, 1970). The GCs are large (up to 45 μm) spheroid cells containing the membrane-bound paraldehyde fuchsin (PAF)-positive secretory granules varying in diameter from 0.1 to 4 μm (Erdelyi and Halasz,1972; Rozsa and Nagy,1967; Volkmer-Ribeiro,1970). As revealed by histochemical and immunocytochemical methods, granules of the GCs contain cysteine-rich protein and/or glycoprotein(Steinbach, 1977; Sminia, 1972), atrial natriuretic peptide-immunoreactive material(Bystrova et al., 2002) and heat-shock-protein-70-immunoreactive material(Martynova et al., 2007).

In gastropod molluscs the heart is innervated by the pericardial branch of the intestinal nerve arising from the visceral ganglion. The number of fibres in this nerve was estimated to be ∼1500 in Busycon canaliculatum(Kuwasawa and Hill, 1973). In A. fulica, four groups of cardiostimulating and two cardioinhibitory neurons have been identified (Bychkov et al., 1997; Furukawa and Kobayashi, 1987a; Furukawa and Kobayashi, 1987b; Zhuravlev et al., 1997). On the inner surface of the H. pomatia atrium a typical neurohaemal area has been observed by Cottrell and Osborne(Cottrell and Osborne, 1969). Innervation of the heart by peptidergic neurons have been described in gastropods (Boyd et al., 1986; Elekes et al., 2000; Elekes and Ude, 1994; Fujimoto et al., 1991;Fujiwara-Sakata and Kobayashi, 1994; Hernadi et al., 1995). Stimulation of the heart nerve or visceral ganglia in molluscs affects the beating of the heart (Hill and Welsh,1966). Degranulation of GCs after the electrical stimulation of the heart nerve was observed almost four decades ago by both light and electron microscopy in S. oblongus(Volkmer-Ribeiro, 1970) and L. stagnalis (Rozsa and Nagy,1970). However, there are only a limited number of studies on GCs in molluscs and the role of these enigmatic cells in the physiology of molluscs remains unknown; new approaches are needed to shed light on GCs functions.

In the present study we developed a convincing approach that allowed us to obtain direct evidence of the existence of neuronal control over secretory activity of the atrial GCs. In addition, we have partially characterized proteins secreted by GCs in response to nerve stimulation. Finally, our results provide a basis for future biochemical and physiological studies of the GCs, the substances contained within them and their action on potential targets.

Experiments were carried out on the adult giant African snails, Achatina fulica Ferussac (1821; Gastropoda, Stylommatophora) bred in the laboratory. The animals were kept in a terrarium at 25°C, in a 12 h:12 h light:dark cycle, and fed a vegetable diet supplemented with dry milk and chalk.

Light microscopy was performed on freshly isolated atria. For this purpose the heart was dissected and placed in physiological solution, which was designed for Achatina and contained (in mmol l^–1):61 NaCl, 3.3 KCl, 10.7 CaCl₂, 10 MgSO₄·7H₂O, 10 Hepes (pH 7.4). Then the atrium was quickly isolated, opened by a single incision, placed in a Petri dish and observed under an inverted microscope (Axiovert 200M, Carl Zeiss, Germany). The images were acquired with a CCD camera (Leica DFC 420, Leica Microsystems,Germany).

All histochemical experiments were conducted on hearts that were fixed overnight in 4% paraformaldehyde buffered with PBS at 4°C. The hearts were then dehydrated in ethanol, embedded in paraffin in the usual manner and 5μm thick sections were cut. These sections were examined using an Axiovert microscope and images were acquired with a digital camera. Paraldehyde fuchsin(PAF) staining was performed according to Gomori(Gomori, 1950) in order to analyze the sulphur-containing proteins and sulphated proteoglycans. Briefly,deparaffinized sections were oxidized in a 1:1 mixture of 0.3% aqueous solution of KMnO₄ and 0.3% H₂SO₄ for 2min. Following destaining in 2% K₂S₂O₅ and washing in distilled water, sections were stained with PAF reagent for 5min. The carbohydrates containing 1-2 glycols, amino or alkylamino grouping were revealed with the periodic acid–Schiff (PAS) reaction. Deparaffinized sections were treated with 0.5% aqueous periodic acid for 10min, washed, and stained with Schiff reagent for 30min. The acidic (at pH2.5) and sulphated proteoglycans (at pH1.0) were analyzed with Alcian Blue staining. Deparaffinized sections were stained with a 1% solution of Alcian Blue 8GX in 3% aqueous acetic acid (pH2.5) or in an aqueous solution of 0.1moll^–1 HCl (pH1.0). The total proteins were revealed as follows: sections embedded in paraffin were stained with mercuric Bromophenol Blue for 2 h (an aqueous solution contained 0.05% Bromophenol Blue, 1%HgCl₂ and 2% acetic acid).

The hearts were prefixed by perfusion of ice-cold 2.5% glutaraldehyde buffered with sodium cacodylate (pH7.4). Then atria were quickly isolated and fixed in the same fixative for 2 h at 4°C. Following the fixation, atria were washed with sodium cacodylate buffer. The samples were postfixed with 1%OsO₄ in sodium cacodylate buffer for 1 h at 4°C. After washing with sodium cacodylate buffer, samples were dehydrated in a graded ethanol series, followed by a graded propylene oxide series. Next samples were infiltrated with a mixture of propylene oxide and Epon–Araldite, and then embedded in Epon–Araldite resin. Polymerized blocks were sectioned for light microscopy at 1 μm using glass knives and for transmission electron microscopy at 50–70 nm with a diamond knife. Sections were cut using a LKB Ultratome III. Sections for light microscopy were mounted on glass microscope slides and subsequently stained with an aqueous solution of 1%Toluidine Blue, 1% borax for 60 s at 60°C. For transmission electron microscopy, ultrathin sections were mounted on copper grids and stained with a saturated solution of uranyl acetate in 70% ethanol and with Reynold's lead citrate (Reynolds, 1963) for 10 min each. Material was examined using a JEM 7A transmission electron microscope operated at 80 kV.

All experiments were performed at room temperature. The shell was removed and snail was fixed in the dissecting chamber. The pulmonary vein and aorta were cut in order to stop the influx of hemolymph into the heart and possible leakage of secreted substances due to partial degranulation of GCs during preparation. The first cannula was inserted into the main pallial vein and tied at the entrance to the atrium and the second one at the ventricular apex. Accessory veins leading to the heart were ligated. The pericardial branch of the intestinal nerve (which will from now on be referred to as the heart nerve) was surgically exposed and hooked on bipolar electrodes connected to an electrical stimulator (Accupulser Signal Generator, WPI, Germany). The schematic view of the experimental set-up is shown in Fig. 1. After the preparation,the heart was perfused with physiological solution in situ, using a peristaltic pump (MityFlex Peristaltic Pump, WPI, Germany) with a constant flow rate of 0.5 ml min^–1. After hemolymph proteins were washed out of the heart two consecutive 2 ml fractions of perfusate were collected before the stimulation. During square-wave electrical stimulation(10 V, 2 ms, 10 Hz) of the heart nerve 4 ml of perfusate were collected. The stimulation parameters were adjusted to induce beating of the heart. In each 12 independent experimental sets, perfusates from two animals were taken(12×2 animals). For morphological examinations, hearts were perfused for 10 min with or without stimulation (control group). Then atria were isolated and prepared for transmission electron and whole-mount immunofluorescence microscopy.

The anticoagulant solution (2 ml; in mmol l^–1: 61 NaCl,3.3 KCl, 45EDTA, 10 Hepes, pH7.4) and 10μl of protease inhibitors cocktail(PIC; Sigma-Aldrich, St Louis, MO, USA) were added to each fraction. The samples were centrifuged at 400 g for 15 min and supernatants were centrifuged at 1700 g for 5 min. The total protein content was measured as described by Bradford(Bradford, 1976). Then proteins in the samples were precipitated with trichloracetic acid and sodium deoxycholate, centrifuged at 7500 g for 10 min, washed twice with cold acetone and once with cold diethyl ester, and dried. The pellets were dissolved in the sample buffer [2% SDS, 10% glycerol, 0.01% Bromophenol Blue, 100 mmol l^–1 dithiothreitol (DTT), 60 mmol l^–1 Tris-HCl, pH 6.8] and heated at 95°C for 5 min.

The hemolymph was collected with a syringe via the pulmonary vein of intact animals. The sample was mixed with anticoagulant solution; PIC was added, the sample was centrifuged, and total protein was measured as described above. Then the supernatant was mixed with fourfold sample buffer (3/1) and heated at 95°C for 5 min.

The atria were isolated and homogenized in cold lysis buffer (in mmol l^–1: 20 Tris-HCl, 20 NaCl, 2 EDTA, 1 DTT, PIC, 1% Triton X-100; pH 7.4). The homogenate was freeze-thawed twice and centrifuged at 7500 g for 10 min. Total protein was measured as described above.

Proteins in the perfusate obtained upon stimulation were precipitated as described above and separated by SDS-PAGE in 12.5% gels. Gels were stained with Coomassie Brilliant Blue G250. Bands of the major 16 kDa protein were excised from gels and used for immunization.

Before the immunization, normal serum was collected as a control for immunochemistry. A male rabbit was immunized first by injections into the digital pulp of hind legs and intracutaneously into flanks with 10 bands of protein homogenized in Hank's solution with Freund's complete adjuvant. After a 3-week interval, the second injection of Freund's incomplete adjuvant was given into the hind leg lymph nodes. Two booster injection of Freund's incomplete adjuvant, were given at one month intervals, intramuscularly into the thigh. The serum was collected 8 days after the second booster injection.

Fragments of nitrocellulose membrane with immobilized 16 kDa protein were used as a solid-phase affinity sorbent to purify anti-16 kDa immunoglobulins. After semidry transfer from SDS-PAGE gels to nitrocellulose membrane, the 16 kDa protein bands were visualized with Ponceau S, excised (100 pieces, total active surface ∼20 cm²) and blocked with BSA. The affinity sorbent was incubated overnight with 1 ml of serum diluted fivefold with PBS in a 15 ml tube at 4°C. In order to remove non-specifically bound proteins, sorbent was extensively washed with the PBS containing 0.5% Tween 20(TPBS) and with double strength PBS. Immunoglobulins were eluted with 1 ml of 0.1 mol l^–1 glycine-HCl buffer (pH 3.0). After elution, the pH was immediately adjusted to physiological range. Then the purified antibody was dialyzed overnight at 4°C against PBS and stabilized with 5 mg ml^–1 BSA and 0.002% sodium azide. Specificity of the anti-protein 16 (AP16) antibody was tested by western blotting.

SDS-PAGE was performed on 12.5% gels according to Laemmli(Laemmli, 1970) with a Mini-Protean II electrophoretic cell (Bio-Rad). To visualize proteins, the gel slabs were silver-stained according to Schagger(Schagger, 2006). For the histochemical visualization of proteins with PAF, the gel slabs was stained according to Gomori (Gomori,1950). For the molecular mass estimation of unknown proteins a set of marker proteins was used (Weber et al.,1972). For western blotting, after electrophoresis, proteins were transferred to the nitrocellulose membrane using the semi-dry method(Towbin et al., 1979). Then the membrane was blocked in 5% non-fat dry milk diluted in TPBS and incubated overnight with the AP16 antibody (dilution 1:200) at 4°C. Following washing in TPBS three times for 30 min each, the blot was incubated in goat anti-rabbit IgG horseradish-peroxidase-conjugated antibody and revealed using the enhanced chemiluminescence (ECL) reaction with ECL Plus reagent (GE Healthcare, Little Chalfont, Buckinghamshire, UK).

Whole-mount immunofluorescence was performed according to Swidbert(Swidbert, 2008). Hearts were prefixed by perfusion with ice-cold zinc-formaldehyde fixative (ZnFA) in situ. The ZnFA fixative contained 0.25% ZnCl₂, 4%paraformaldehyde, 100 mmol l^–1 NaCl, 15 mmol l^–1 Tris-HCl, pH 6.8 (pH above 7.0 causes precipitation of ZnCl₂). Then atria were quickly isolated, opened by a single incision and fixed overnight with ZnFA containing 1% paraformaldehyde at 4°C. Fixation lasted for 24 h and was terminated by washing in 100 mmol l^–1 NaCl, 15 mmol l^–1 Tris-HCl, pH 6.8 for 2 h at room temperature. The atria were then transferred into a 1:4 mixture of dimethyl sulphoxide (DMSO) and methanol. After 2 h incubation at room temperature, atria were transferred into absolute methanol and incubated overnight at –25°C. Then atria were transferred directly into PBS. All following incubations and washes were carried out in PBS containing 1%DMSO. Atria were first incubated in 0.5% BSA for 1 h. The same solution was used to dilute the primary and secondary antibodies. The AP16 primary antibody was applied at 1:50 dilution for 60 h at 4°C with constant agitation. Following five washes of 30 min each, a secondary FITC-labelled anti-rabbit antibody (1:300; Sigma) was applied for 24 h at 4°C. Following three washes of 30 min each, the atria were mounted in a 1:2 mixture of PBS and glycerol. Immunofluorescence was visualized using a confocal laser scanning microscope (Leica TCS SL, Leica Microsystems, Germany). The control immunocytochemical reaction was performed by substituting preimmune rabbit serum for the primary antibody.

To strengthen our confidence that the 16 kDa protein imunoreactivity seen in the atrium was due to the presence of 16 kDa protein, we conducted a preabsorption specificity test. In this experiment, 16 kDa protein purified from atria of A. fulica by preparative electrophoresis and immobilized on nitrocellulose membrane was used as blocking agent. The AP16 antibody (100 μl) was diluted tenfold with PBS containing 1% DMSO and 0.5%BSA, and then incubated with blocking agent (active surface ∼15 cm²) for 17 h at 4°C, before applying to the tissue. Immunoprocessing on whole mounts was performed as described above.

Statistical differences in protein concentration between fractions of the perfusate were determined by nonparametric test for dependent variables(Wilcoxon matched pairs test). Differences were considered significant at P<0.01. Results are expressed as the mean ± s.d.

The GCs are easy to recognize under the microscope. Fig. 2A shows the GCs on the surface of muscle bundles in a freshly isolated atrium. They are large spheroid cells packed with secretory granules. The granules are intensely stained (Fig. 2B) either dark or light blue in semithin sections. The atrial GCs are strongly PAF positive(Fig. 3A) and their secretory granules stained with mercuric Bromophenol Blue, revealing the presence of proteins (Fig. 3B). The GCs were negative to the PAS reaction and Alcian Blue staining (data not shown).

Ultrastructural examination has shown that GCs contain numerous spheroid membrane-bound granules (Figs 4and 5). The rough endoplasmic reticulum, ribosomes and mitochondria are minor cytoplasmic component. The nucleus is eccentrically located and granules are frequently found in the nuclear infoldings. Granules can be subdivided into three types according to ultrastructural matrix appearance (Fig. 5). Granules filled with finely granulated material of low electron density were classified as type 1. The type 2 granules appeared to be filled with electron-dense finely granulated material. The type 3 granules were filled with smooth, disorganized material of reticulated structure. Closer examination of the granule matrix revealed the existence of intermediate states between type 2 and 3 granules(Fig. 6). We have observed granule fusions that formed a chain of consequently emptying granules(degranulation channel) permeating the cytoplasm of the GC and filled with loosened matrix (Fig. 4). The fusions were observed only for type 3 granules. The degranulation channels are eventually opened to the extracellular space through the pores and emptied altered granule material. This process appears to be a common way of releasing granular content by GCs into the extracellular space. The degranulation channels and different types of granules occasionally occur within the same cell. In addition, we noted accumulation of numerous tubular and vesicular membrane structures under basal lamina surrounding the GC(Fig. 4).

A further common characteristic of the GCs is that they are located in close proximity to neurosecretory nerve fibres densely innervating atrial tissue. The nerve fibres form membrane-to-membrane contacts with GCs without any pre- or postsynaptic membrane specialization(Fig. 7A,B). All these contacts are characterized by tightly apposed sites of axolemma and GC plasmalemma running parallel and forming a 15–20 nm wide cleft. Two types of contacts were revealed: (1) axon profiles deeply embedded into the GCs; (2)axon profiles on the surface of the GCs. A number of axon profiles did not form close contact with, but remained relatively distant (several hundred nanometres) from the GCs or were in contact with variable widths of the cleft. Occasionally, several axon profiles were found to contact one GC, suggesting multiterminal innervations. The axon profiles that establish contacts with GCs were found to contain a large ovoid (90–160 nm in diameter) dense-core granule. Interestingly, GCs were frequently seen in close proximity to nerves. The latter consisted of a neurosecretory fibres enwrapped by glial processes that were also present within the nerve.

The morphology of the GCs was examined by transmission electron microscopy following heart nerve stimulation. The ultrastructural study showed that GCs in control atria were mostly well-preserved and contained all types of granules (Fig. 8A). Following the nerve stimulation a massive granule fusion and degranulation channel formation were observed. The GCs contain mostly type 3 granules and showed varying states of granules emptying (Fig. 8B). Some of the stimulated GCs were entirely permeated by an interconnected channel system filled with amorphous disorganized matrix(Fig. 8C). Identification of actively releasing GCs was often possible by the presence of completely emptied degranulation channels.

In this set of experiments, after two consecutive washes of 2 ml each, the heart nerve was stimulated, and 4 ml of perfusate was collected. Results of analysis of total protein concentration in perfusate are presented in Fig. 9. Total protein concentration decreased significantly from 58±44.3μgml^–1 in the first fraction collected immediately after heart preparation to 4.7±3.5μgml^–1 in the second fraction of the washes. The total protein content in the third fraction,obtained upon stimulation of the nerve, increased significantly(11.5±7.4 μgml^–1) compared with that in the second fraction.

SDS-PAGE and silver staining of proteins precipitated from the third fraction revealed three major bands with molecular masses of ∼16, ∼22 and ∼57 kDa (Fig. 10, lane 2); the two first were undetectable in hemolymph of the normal snails(Fig. 10, lane 3). The proteins with similar mobility were weakly represented in the first fraction(Fig. 10, lane 1), but in this fraction the high-molecular-mass proteins similar to those in hemolymph were present. The PAF staining of proteins from the third fraction resolved with SDS-PAGE showed that only the 16 kDa band was PAF positive(Fig. 10, lane 4).

Western blot analysis revealed that antiserum raised against the 16 kDa protein recognized a single band with apparent molecular mass of 16 kDa in the atrium lysate (data not shown). We have purified immunoglobulins specific to the 16 kDa protein using immobilized denaturated antigen as affinity sorbent. The AP16 antibody recognizes a single band (16 kDa) in the fraction obtained after stimulation and in lysate of the intact atrium(Fig. 10, lanes 5 and 6). The optimal titres for western blot and immunofluorescence were of 1:200 and 1:50 dilutions, respectively. We conducted experiment to confirm that the immunofluorescence we report is suppressed specifically by preabsorption of the antibody with 16 kDa protein. Incubation of the AP16 antibody with the 16 kDa protein completely abolished immunolabeling in the atrium (data not shown).

To determine the location of the 16 kDa protein in the atrial tissue, the immunoreactivity was investigated by confocal microscopy and whole-mount immunofluorescence technique. The immunofluorescence experiments on whole-mount atria preparations proved to be ideal to trace the degranulation of GCs following nerve stimulation because of the specific reaction of the AP16 antibody with granular contents. Results revealed the presence of AP16-immunoreactive material in the atrial granules of A. fulica(Fig. 11). In control snails,most of atrial GCs contained fluorescent granules(Fig. 11A–D), whereas in cells that underwent nerve stimulation there were only a few(Fig. 11E–H).

In this study, we have used several microscopy techniques, and physiological and biochemical methods to investigate the effect of electrical stimulation of the heart nerve on atrial GCs in A. fulica; the heart was perfused in situ.

Our ultrastructural data clearly demonstrated an intimate relationship between GCs and nerve fibres in the A. fulica atrium. In most cases,axon profiles form a close contact with or are deeply embedded into the GCs;however, they never form a true specialized synapse. According to previous ultrastructural studies on peripheral tissues of gastropod molluscs(Elekes, 2000; Elekes and Ude, 1994; Plesch, 1977) the neuromuscular and neuroglandular contacts are close but unspecialized contacts; in vertebrates, similar contacts are observed between nerve terminals and mast cells (Keith and Saban,1995). Although true synapses are lacking between nerve fibres and GCs, a transmission via paracrine peptides and neurotransmitters is still possible. The axon profiles in contact with GCs contain dense-core vesicles of typical appearance and size for peptide storage and, therefore,can be classified as peptidergic. According to Elekes and Ude(Elekes and Ude, 1994), five types of ultrastructurally distinct FMRFamide immunoreactive axon profiles are present in the Helix atrium. A non-synaptic modulatory role of neuropetides in the peripheral nervous system in Helix has also been proposed (Elekes, 2000). The axon profiles at a distance from GCs or forming contact with variable widths of clefts may exert another form of modulatory input. Thus, our results suggest that GCs are directly innervated.

Upon electrical stimulation of the heart nerve, nearly all secretory granules undergo a series of morphological changes: (1) granule matrix alteration; (2) granule to granule membrane fusions; (3) degranulation channel formation and (4) granule matrix extrusion through the pores. This is consistent with previous observations on other secretory systems, e.g. barnacle cyprid cement secretion following the chemical stimulation(Odling et al., 2006) and mast cells degranulation following chemical or nerve stimulation(Dvorak et al., 1991; Monteforte et al., 2001). Interestingly, the exact mechanism of exocytosis may vary in different secretory systems depending upon the physiological time course required for exocytotic events. In vertebrates, a similar pattern of exocytotic event can be observed in mast cells undergoing the rapid process of anaphylactic degranulation (Dvorak et al.,1991). Similar observations of extracellular vesicular structures in concert with granule release were made in mast cells in conjunction with anaphylactic degranulation (Dvorak et al.,1985; Dvorak et al.,1991). According to Dvorak et al., shedding of membrane structures and focal pieces of cytoplasm after massive additions of granule containers(membranes) to plasmalemma serves as one of the mechanisms of recovery of the mast cells following rapid degranulation. The granules with partial loss of contents and those completely depleted in GCs of control snails' atria suggest that some background secretion in these cells occurs under physiological conditions. Another possibility is that the dissection procedure may lead to a moderate secretion in atrial GCs of control snails.

Our in situ perfusion experiments demonstrated that nerve stimulation causes degranulation of the atrial GCs and release of three proteins into the circulation, thus providing physiological evidence for interaction between nerve fibres and GCs. The molecular masses of three newly isolated proteins were estimated to be 16, 22 and 57 kDa. Based on the electrophoretic band intensity, the 16 kDa protein appears to be the major protein secreted after the stimulation. Interestingly, as shown by silver staining, 16- and 22 kDa proteins are not abundant in hemolymph of intact snails. The presence of small amounts of the 16-, 22- and 57 kDa proteins in the perfusate taken before the nerve stimulation may have resulted from the partial degranulation of the GCs during the preparation procedure.

The whole-mount immunofluorescence technique and confocal microscopy revealed that the 16 kDa protein is localized in granules of the atrial GCs and can be released into the hemolymph after nerve stimulation. The cellular location of the 22- and 57 kDa proteins remains to be elucidated. We suggest that secretory granules of atrial GCs is a most probable source of these proteins; however, the granules of gliointerstitial cells associated with nerve fibres and neurosecretory granules of nerve endings could also be involved.

The granular content of GCs in A. fulica, were found in the present study to be strongly PAF positive. This data is in line with previous observations on other gastropod species(Erdelyi and Halasz, 1972; Rozsa and Nagy, 1967; Volkmer-Ribeiro, 1970). PAF staining was first introduced by Gomori(Gomori, 1950) in order to stain elastic fibres, mast cells, gastric chief cells, certain pituitary basophils, and pancreatic β-cells. It is well established that the substances stained with PAF are of two classes: (1) sulfur-containing peptides and proteins, and (2) sulphated proteoglycans(Mowry, 1978). PAF has also been widely used to stain neurosecretory material in both vertebrates(Cameron and Steele, 1959) and invertebrates (Fritsch et al.,1976; Duve and Thorpe,1979; Duve and Thorpe,1983). As revealed by mercuric Bromophenol Blue staining, the granular content of atrial GCs in A. fulica is proteinaceous in nature, whereas negative PAS reaction and Alcian Blue staining suggest absence of carbohydrates containing 1,2-glycols, amino or alkylamino groups, sulphated and acidic proteoglycans. Thus, the PAF reaction reveals sulfur-containing protein(s) in granules of the GCs. In the present study, attention was focused on the 16 kDa protein, as this is the only protein from the total bulk of electrophoretically resolved proteins of perfusate, which is PAF positive. Thus, 16 kDa protein appears to be one of the sulfur-containing proteins contained in GCs. It is reasonable to suggest that the PAF-positive granular cells in H. pomatia (Erdelyi and Halasz, 1972) and L. stagnalis(Rozsa and Nagy, 1967) may also contain 16 kDa protein in their granules. Possible candidates for the 16 kDa protein may be found among proteins of A. fulica hemolymph that contains three multimeric lipopolysaccharide-binding lectins composed of identical 15 kDa subunits (Basu et al.,1986; Biswas et al.,2000; Mitra and Sarcar,1988). Alternatively, another possible candidate for the 16 kDa protein is an atrial natriuretic peptide precursor with a molecular mass of exactly 16 kDa. In our earlier report, we showed that granules of H. pomatia atrial GCs reacted with polyclonal antiserum against rat atrial natriuretic peptide (Bystrova et al.,2002). However, subsequent proteomic analysis is needed to characterize newly isolated proteins and specific roles of these proteins remain to be elucidated.

The snail atrial GCs show interesting morphological and physiological parallels with vertebrate mast cells – peculiar large tissues granulocytes. The mast cells produce and secrete a wide range of bioactive agents participating in various physiological, immune and regeneration processes and actually are one-cell glands. There is convincing evidence that in a variety of tissues the mast cells are in close anatomical and functional proximity to the peripheral nervous system(Newson et al., 1983; Johnson and Krenger, 1992; Keith and Saban, 1995; Laine et al., 2000). Electrical nerve stimulation affects mast cell number and degranulation(Kiernan, 1990; Keller et al., 1991; Monteforte et al., 2001). The mast cells similarly to GCs do not form any specialized contacts with the nerve endings. It is reasonable to suggest that the atrial GCs in snails, like vertebrate mast cells, could be an important component of the endocrine system. The GCs in the A. fulica atrium are ideally situated for endocrine secretion and released substances could easily spread over the entire circulatory system and reach peripheral targets within a short time. The circulation time of the hemolymph in pulmonates ranges between 4–6 min (Jones, 1983).

In summary, our results reveal that the atrial GCs are a fast-acting secretory component, and the central nervous system can modulate the secretion from these cells via non-synaptic pathways. In addition, the level of 16 kDa protein in hemolymph can be used as a `marker' for the secretory activity in atrial GCs in response to either internal or external(environmental) stimuli. A better understanding of these stimuli and specific features of newly isolated proteins could clarify the role of GCs in the physiology of molluscs.