Allatostatin Peptides in the Crab Stomatogastric Nervous System: Inhibition of the Pyloric Motor Pattern and Distribution of Allatostatin-Like Immunoreactivity

The stomatogastric nervous system (STNS) of decapod crustaceans has served as an important model for examining the function of neural networks (Selverston and Moulins, 1987). For example, the pyloric network of the crab Cancer borealis (Harris-Warrick et al. 1992; Marder and Weimann, 1992), which produces rhythmic movements of the pyloric region of the stomach, consists of 11 neurons in the stomatogastric ganglion (STG) and is modulated by a defined set of input fibers descending from the oesophageal ganglion (OG), the paired commissural ganglia (CoGs) and the brain (Coleman et al. 1992). At least 15 neuromodulatory substances have been found in the crustacean STNS, including a number of neuropeptides (Marder, 1987; Harris-Warrick et al. 1992; Marder and Weimann, 1992), all of which have excitatory actions.

The allostatins (ASTs) have been isolated from various insects species (Diploptera punctata, D-AST-1–5, Woodhead et al. 1989; Pratt et al. 1989, 1991; Manduca sexta, M-AST, Kramer et al. 1991; Calliphora vomitoria, C-AST-1–5, Duve et al. 1993) and have been shown to inhibit the synthesis of juvenile hormone by the corpora allata (Tobe and Stay, 1985), a major endocrine organ of insects. The synthesis and secretion of juvenile hormone play an important role in larval development and in the maturation and reproductive activities of insects (Schooley and Baker, 1985). The D. punctata and C. vomitoria ASTs share a conserved C-terminal sequence (-Y-X-F-G-L-NH₂, Woodhead et al. 1989; Pratt et al. 1989, 1991). The M. sexta AST is dissimilar in sequence to the other ASTs, suggesting the existence of more than one allatostatin peptide family. The cDNA for the D-AST has been isolated and the translation of the cDNA indicates that 13 different peptides belong to this family (Donly et al. 1993).

Although, in insects, ASTs are closely connected to the synthesis of juvenile hormone, there are strong indications that they have additional neuromodulatory functions (Stay et al. 1992; Duve et al. 1993). Immunocytochemical staining shows AST-like immunoreactivity in many neurons throughout the nervous system in a variety of insects (P. americana, Agricola et al. 1992; Gryllus bimaculatus, Schildberger and Agricola, 1992; D. punctata, Stay et al. 1992; C. vomitoria, Duve et al. 1993). The only physiological action reported thus far of the AST peptides, other than juvenile hormone synthesis, is that D-AST-2 antagonizes the excitatory effect of proctolin on the antennal heart of the cockroach P. americana. However, when applied to the heart alone, D-AST-2 does not alter the rhythm (Hertel and Penzlin, 1992).

In this paper, we use immunocytochemical methods to demonstrate that AST-like peptides are found within the STNS in a crustacean species. In physiological studies we show that D-AST-1–4 are the first peptides shown to have inhibitory actions on the pyloric network of C. borealis.

Crabs, Cancer borealis Stimpson, were obtained from commercial fisheries (Neptune Seafood, James Hooke Co., and Bay State Lobster, Boston, MA). The animals were maintained in aquaria with circulating aerated, artificial sea water at 10–14°C. Data were obtained from male crabs weighing between 300 and 800 g.

Experiments were performed on the STNS, which includes the CoGs, OG and STG, and their connecting and motor nerves. Intracellular and extracellular recordings were carried out using routine methods for the STNS (Selverston and Moulins, 1987). STG motor neurons were identified as described in Hooper et al. (1986) and Weimann et al. (1991). STGs and CoGs were desheathed to facilitate penetration of the neurons for intracellular recordings and to facilitate diffusion of solutions containing peptides. Intracellular recordings were made using glass microelectrodes (15–20 MΩ) filled with a solution containing 4 mol l⁻¹ potassium acetate and 20 mmol l⁻¹ KCl. Data were collected on a chart recorder (Gould TA4000). Preparations were continuously superfused (7–15 ml min⁻¹) with physiological saline containing 440 mmol l⁻¹ NaCl, 11.3 mmol l⁻¹ KCl, 26.3 mmol l⁻¹ MgCl₂, 13.3 mmol l⁻¹ CaCl₂, 11.0 mmol l⁻¹ Tris base and 5.2 mmol l⁻¹ maleic acid; pH 7.4–7.5. The saline was cooled to 10–14°C and the bath volume was 10–20 ml. D-ASTs (Table 1) were purchased from Bachem California (Torrance, CA).

Animals were cooled on ice for 30 min before dissections of the STNSs were carried out in ice-cold saline. STNSs were fixed overnight in 4% formaldehyde in phosphate buffer (PBS, 0.1 mol l⁻¹ phosphate buffer containing 462 mmol l⁻¹ NaCl and 16 mmol l⁻¹ KCl, pH 7.5). Following several PBS washes, the preparations were incubated in PBS containing 0.3% Triton-X 100 (PBS-X) for 1 day. Anti-D-AST-1 antibody, generously provided by Dr Hans Agricola (University of Jena, Germany), was raised in a rabbit injected with insect D-AST-1 conjugated to thyroglobulin with glutaraldehyde. The antibody was applied in a 1:5000 dilution in PBS-X for 4–6 days. Secondary anti-rabbit antibodies labeled with Cy3 (1:40, Jackson Immunochemicals, West Grove, PA) were applied for 1 day. Preparations were mounted in glycerol after washing (8 h) with frequent changes of PBS-X and were viewed using a Zeiss-Axiophot microscope equipped for confocal microscopy (BioRad system: photomultiplier MRC 600 and a krypton/argon laser). Images of the whole mounts were reconstructed from optical sections (2 μm), and photographs were taken with a Sony color video printer UP-5000 system. In addition, two preparations were stained using biotinylated secondary antibodies (1:200, Vector, Burlingame, CA) and the avidin–biotin complex (ABC) as described by Vector Laboratories.

For double labeling, a similar protocol was used except that the two primary and secondary antibodies were applied as mixtures. Preparations were incubated in monoclonal rat anti-serotonin (Accurate Chemical Scientific Corp., Westbury, NY; 1:50) and rabbit anti-D-AST-1 antibodies (1:5000) for 4–6 days. The incubations with secondary antibodies were at final concentrations of 1:50 FITC-labeled donkey anti-rat (Jackson Immunochemicals, West Grove, PA) and 1:50 Texas-Red-labeled donkey anti-rabbit (Jackson Immunochemicals, West Grove, PA) for 1 day.

Separate preabsorption controls were carried out for all four D-ASTs (Bachem California, Torrance, CA). The peptides were dissolved in distilled water, diluted with PBS-X to a final concentration of 10⁻³ mol l⁻¹ and antibody was added to a final dilution of 1:5000. After 24–48 h of incubation at 4°C, individual STGs were added and the procedures were continued as described above.

A series of experiments was performed to determine the physiological activity of the D-AST peptides 1–4. The pyloric motor pattern involves alternating bursts of activity in three different classes of motor neurons: lateral pyloric (LP), pyloric (PY) and pyloric dilator (PD) neurons. The pyloric rhythm was monitored by extracellular recordings from the lateral ventricular nerve (lvn) as shown in the top trace of Fig. 1. Bath applications of D-AST peptides 1–4 (10⁻⁶ mol l⁻¹) caused reversible decreases in the pyloric cycle frequency. In most preparations, control patterns of activity returned after 10–20 min of washing with saline. In the experiment shown in Fig. 1, the control pyloric frequency was 0.91 Hz. The pyloric frequencies in D-AST-1, D-AST-2, D-AST-3 and D-AST-4 were 0.56, 0.48, 0.37 and 0.57 Hz, respectively. The decrease in pyloric frequency produced by all four D-ASTs in this experiment was significantly different from the control (P<0.001, Student’s paired t-test). The reduction elicited by D-AST-3 differs significantly from that of D-AST-1 (P<0.01) and D-AST-4 (P<0.001), but not from that of D-AST-2 (P>0.05). An increase in LP burst duration was seen following D-AST application in most experiments. This increase is not necessarily a direct effect of D-AST, but could result from the frequency reduction. When the pyloric cycle frequency was drastically reduced to a low frequency, repeated brief disruptions of the motor pattern occurred (Fig. 1 dots below traces). The disruptions are characterized by stretches of tonic activity in LP that start when PD neurons fail to burst and are terminated by PD neuron bursts; they were only seen when the pyloric cycle frequency was drastically reduced. In each of a total of five experiments in which all four peptides were compared, the frequency reduction produced by application of D-AST-3 was always the largest. Therefore, D-AST-3 was used in the remaining physiological studies.

To determine the dose-dependence of the D-AST-3 effect, experiments were carried out in which the frequency of the pyloric rhythm was measured at different D-AST-3 concentrations. In all experiments (N=6), D-AST-3 caused a dose-dependent decrease in the frequency of the pyloric rhythm. The results of one such experiment are shown in Fig. 2. Starting at 10⁻⁸ mol l⁻¹, increasing concentrations of D-AST-3 decreased the frequency of the pyloric rhythm (Fig. 2B). In parallel with the decrease of the pyloric cycle frequency, the rate of firing of the LP neuron within each burst was reduced (Fig. 2A).

The percentage change in pyloric cycle frequency produced by a given concentration of peptide varied from preparation to preparation. Previous work with the peptide proctolin has shown that it has strong excitatory effects on the pyloric cycle frequency when bath-applied to preparations cycling at low frequencies, but has relatively little effect when applied to rapidly cycling preparations both in the crab C. borealis and the lobster Panulirus interruptus (Marder et al. 1986; Hooper and Marder, 1987). To determine whether the variability seen with D-AST-3 was also dependent on the control frequency, the percentage reduction in frequency of the pyloric rhythm was plotted as a function of the control frequency for 17 applications of 10⁻⁷ mol l⁻¹ D-AST-3 in 13 preparations (Fig. 3). A significant correlation (r=−0.78, P<0.001) was seen between the control frequency and the percentage decrease in frequency caused by the peptide. The plot in Fig. 3 clearly shows that D-AST-3 had pronounced inhibitory effects when the preparations were slowly cycling (<0.8 Hz) under control conditions and only weak effects on preparations with robust pyloric rhythms. This indicates that the D-AST-3 effect depends on the previous physiological state of the preparation. Because of this state-dependence, it was difficult to determine accurately a threshold concentration for the D-AST-3 effect. In five of six experiments, 10⁻⁸ mol l⁻¹ D-AST-3 produced a measurable frequency reduction, and in one of four experiments in which 10⁻⁹ mol l⁻¹ D-AST-3 was tested, a small reduction (19%) in frequency was observed.

Fig. 4 shows one of the largest frequency reductions produced by D-AST-3. In this experiment, the PD, LP and PY neurons were recorded intracellularly. The control pyloric frequency was relatively slow (0.60 Hz), and 10⁻⁶ mol l⁻¹ D-AST-3 completely inhibited the pyloric rhythm. The PD neuron fired only occasionally and the amplitude of the slow depolarizing wave in the PD neuron was reduced. These recordings show tonic activity in the LP and PY neurons, with only small inhibitory synaptic potentials from the PD neurons.

Immunocytochemical methods were used to determine whether AST-like peptides are contained in neurons of the STNS of C. borealis. A polyclonal antibody against insect D-AST-1 was used, since neither crustacean ASTs nor antibodies against crustacean ASTs are available.

The anatomical staining patterns presented in this paper are based on 13 immunocytochemical whole-mount preparations of the STNS. Immunoreactive structures were found throughout the entire STNS in all preparations (Fig. 5). Immunoreactive cell bodies were located in the CoG, (12–19 cells in each CoG, N=11), in the unpaired OG (2 cells, N=13) and in the peripheral nerves, lvn and medial gastric nerve (mgn) (a single cell body in each, N=2). The two pairs of peripheral cell bodies project to the STG. A third pair of cells that project to the STG seems to originate in the CoGs, but the cell body location of the projecting neurite was not determined. This staining pattern was seen in all 13 preparations and the immunostaining was blocked by preincubation for 24–48 h of the antibody (1:5000) with all four D-ASTs (10⁻³ mol l⁻¹). No staining was observed when the primary antibody was omitted.

In the STG, fine cellular processes and varicosities are seen but no cell bodies showing immunostaining are observed (Fig. 6A). The origins of these cellular processes are two pairs of bipolar cell bodies located in the lvn and mgn (Figs 5, 6Bi) and one pair of cells that are probably located in the CoGs.

The lvn neurons are located just posterior to the mgn. In a few preparations, the distal axons of the mgn cells were traced to the gm9a muscles, where they branch extensively. The proximal axons of mgn and lvn cells project through the lvn and the dorsal ventricular nerve (dvn) into the STG (Fig. 5). At the anterior end of the STG, the weakly labeled axons of the mgn and lvn cells leave the STG through the stomatogastric nerve (stn). They bifurcate into the right and left superior oesophageal nerves (son) and end in the CoGs. These morphological characteristics are similar to those of previously described serotonin (5-HT)/acetylcholine (ACh)-containing cells, the gastropyloric receptors (GPR, Katz et al. 1989). Accordingly, a double-labeling experiment was performed to determine whether the AST-immunoreactive cells are the GPR cells. Secondary antibodies for rabbit and rat IgG labeled with Texas Red and FITC were used to detect AST and 5-HT (see Materials and methods). As illustrated in Fig. 6B, the mgn and lvn AST-labeled cells also showed immunostaining for 5-HT. Cross-reactivity between these antibodies is unlikely, because different labeling patterns were seen for the antibodies in the STG, CoG and OG in the same preparations.

The set of AST-labeled cells innervates the STG from the anterior ganglia via the stn. Axons of these cells are labeled more strongly than those of the GPR neurons. These axons were traced back to the CoG, but following their path through the intensely labeled neuropil was not possible. Therefore, the cell bodies of origin of these neurites were not identified.

In each CoG, about 12–19 cell bodies (N=11) stain for AST (Fig. 6C). Two large cell bodies (approximately 50 μm) are stained reliably in the anterior part of the ganglion. They are surrounded by about eight medium-sized cell bodies (15–25 μm) and up to nine small cell bodies (5–10 μm). Neurites of all the CoG neurons project into the neuropil, where they disappear in a network of intensely labeled fibers (Fig. 6C). Immunostained fibers were observed in the inferior oesophageal nerve (ion) and the son. Stained cell bodies were also found in the son, close to where it exits the CoG, and in the dorsal posterior oesophageal nerve (d-pon, Fig. 5).

Two pairs of immunoreactive fibers were found in each ion (Fig. 6D). Their polarity and origin could not be determined. A pair of these fibers passes through the OG and connects the two CoGs. The other neurite pair joins the inferior ventricular nerve (ivn).

They are accompanied by the neurites of the two large, monopolar cell bodies stained in the OG. These six fibers reach the supraoesophageal ganglion (brain) via the ivn. The neuropil in the OG was not labeled.

The results presented in this paper demonstrate direct physiological actions of D-AST peptides on the crustacean STNS. D-AST-1–4 reduce the frequency of the pyloric rhythm in a dose-dependent manner. The strength of the D-AST-3 effect is correlated with the control frequency of the pyloric rhythm, demonstrating that the effect of D-AST-3 depends on the preceding physiological state of the preparation. The existence of AST-like immunoreactive neurons projecting to the stomatogastric ganglion in crabs, where they elaborate in a rich array of varicosities, and the reproducible dose-dependent physiological effects of these peptides on the pyloric network suggest that crustacean AST-like peptides probably exist in the stomatogastric system, where they are likely to act as neuromodulators. Most neuromodulators of the STG have pronounced excitatory effects on the networks of the STNS. These include the peptides proctolin (Hooper and Marder, 1987; Heinzel and Selverston, 1988), red pigment-concentrating hormone (RPCH, Nusbaum and Marder, 1988; Dickinson and Marder, 1989), the endogenous FMRFamide-like peptides SDRNFLRFamide and TNRNFLRFamide (Weimann et al. 1993) and other classic transmitters, such as ACh, dopamine and 5-HT (Harris-Warrick et al. 1992; Marder and Weimann, 1992). Only two non-peptidergic substances, γ-aminobutyric acid (GABA, Cazalets et al. 1987) and histamine (Claiborne and Selverston, 1984), are known to inhibit significantly the rhythmic motor patterns of the STG. Cazelets et al. (1987) showed that GABA inhibits pyloric activity in the lobster Homarus gammarus, and Claiborne and Selverston (1984) showed that histamine inhibits many neurons of the pyloric network of the lobster Panulirus interruptus, thereby inhibiting the rhythm. D-AST is the first peptide that has been shown to inhibit the STG. The inhibitory effects of D-AST-3 are not limited to the pyloric rhythm. It also inhibits the gastric mill rhythms and reduces the EJPs of some gastric mill muscles (Skiebe et al. 1993), suggesting that the ASTs have coordinated actions centrally and peripherally to inhibit STG rhythms, including turning ‘off’ the STNS at appropriate times in behavior. Another role for these peptides may be in sensory feedback, since AST-like substances are co-localized with ACh and 5-HT in the sensory GPR cells. In concert with other modulators, the ASTs could induce a different rhythmic pattern from that of either one of the modulators alone. This would increase the number of possible rhythmic patterns.

Previous work with the peptide proctolin has shown that it activates the pyloric rhythm depending on the control frequency of the preparation both in the crab C. borealis and in the lobster P. interruptus (Marder et al. 1986; Hooper and Marder, 1987). This state-dependence does not only apply for bath-applied proctolin. The effects of stimulating the proctolin-containing neurons in C. borealis are also dependent on the frequency of the pyloric rhythm before stimulation (Nusbaum and Marder, 1989). The effects of both proctolin and D-AST-3 are similar with respect to their state-dependent modulation of the pyloric rhythm. Both peptides elicit pronounced effects when the preparations are slowly cycling under control conditions and only weak effects on preparations with robust pyloric rhythms. Nusbaum and Marder (1989) suggest that different hormonal or physiological states in the animal might cause the variability of the pyloric rhythms and the state-dependent excitation of the peptide proctolin. It is likely that the state-dependent effects of the ASTs have their origin in changing physiological and hormonal states as well.

Since the D-ASTs were bath-applied simultaneously onto all ganglia (CoGs, OG and STG) our results may be due to (i) inhibiting excitatory inputs from the CoGs and the OG, (ii) inhibiting the pyloric network itself, or (iii) both of these. When D-AST-3 was applied only to the STG, the inhibitory effect on the pyloric rhythm was still present (data not shown), demonstrating that D-AST-3 must have at least some direct action on STG neurones or on presynaptic terminals of excitatory inputs. To examine targets of the D-AST action, additional experiments will need to be performed.

D-AST peptides 1–4 from insects have similar physiological actions on the crustacean STNS, although the effectiveness of the peptides varied. While this suggests that the specificity of action of these peptides for the receptor may be dependent on the conserved C terminus, isolation of the crustacean peptides will be required before further characterization of putative peptide receptors is possible.

This paper demonstrates the presence of immunoreactive AST-like substances within neurons and fibers of the STNS. Preabsorption of the anti-D-AST-1 antibody with all four D-ASTs completely eliminated any staining in the STNS. Preabsorption with proctolin, locustatachykinin II, FMRFamide and neurohormone D did not block the staining in insects (Agricola et al. 1992). It is likely that the D-AST antibody detects a crustacean homologue of the insect peptides, although purification and sequencing of the crustacean peptides will be required to demonstrate this.

One interesting feature shown in the immunocytochemical studies is the staining for D-AST in the GPR neurons that also contain 5-HT and ACh (Katz and Harris-Warrick 1989). In previous studies, the physiological actions of GPR neurons have been attributed to their contents of either ACh or 5-HT (Katz et al. 1989; Kiehn and Harris-Warrick, 1992). It would be interesting to determine additional properties conferred upon the GPR neurons by their AST content.

In conclusion, this paper illustrates that insect ASTs have inhibitory modulatory actions on the stomatogastric ganglion in a crustacean nervous system, in addition to their already known neurohemal functions in insects. AST-like peptides, demonstrated by immunostaining, are found in neuropil processes from which they are likely to be released by neurons that project into the STG. In further investigations we hope to reveal the functions that the inhibitory AST peptides have in shaping the rhythms generated by the STNS both centrally and peripherally.

We thank Dr Hans Agricola for the generous gift of the anti-AST-1 antibody, Andrew Christie for his assistance with the confocal microscopy, Drs Eve Marder and Ed Kravitz for their insight and support and Joan McCarthy-Griffin and Joe Gagliardi for manuscript preparation. Supported by NS17813 (to Eve Marder), DFG SK 38/1-1 (to P.S.) and DFG SCHN 368/1-1 (to H.S.).