INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Application of neurotransmitters to Aplysia neurons activates specific lipoxygenase pathways. The application of histamine to identified neurons in the abdominal ganglion of Aplysia results in the release of arachidonic acid, through the action of a phospholipase A₂ (Shapiro et al. 1988), and the conversion of arachidonic acid to bioactive 12-lipoxygenase products (Piomelli et al. 1988, 1989). Similarly, the application of acetylcholine (ACh) to Aplysia nervous tissue leads to the generation of 8-lipoxygenase metabolites. Previously we showed that ACh induces the generation of 8(R)-hydroperoxyeicosatetraenoic acid [8(R)-HPETE], which can be reduced both enzymatically and nonenzymatically to 8(R)-hydroxyeicosatetraenoic acid [8(R)-HETE]. In addition, 8-HPETE is enzymatically converted to a ketone, 8-ketoeicosatetraenoic acid (8-KETE) (Steel et al. 1997). Other more polar enzymatic products probably related to hepoxilin have been tentatively identified (Tieman and Feinmark, unpublished data). These metabolites comprise the lipids of the 8-lipoxygenase family.

8-Lipoxygenase is a member of a large family of enzymes with members distributed widely in both plants and animals (Brash 1999; Kuhn and Thiele 1999). Since a major function of the well-characterized lipoxygenases is to generate second messengers and signaling molecules, it seems likely that the ACh-induced metabolites of 8-lipoxygenase also play a specific role in Aplysia neural function.

To further our understanding of how 8-lipoxygenase is activated by ACh, we characterized the ACh receptor (AChR) that activates 8-lipoxygenase metabolism. ACh is known to activate four pharmacologically distinct receptors. One is a G-protein-linked, metabotropic receptor that activates a potassium conductance (Kehoe 1972a; Sasaki and Sato 1987). The other three are ionotropic AChRs, one that mediates a nonspecific cationic conductance (Ascher et al. 1978) with the other two being distinct chloride conductances (Kehoe and McIntosh 1998). One of the chloride-dependent responses is rapidly desensitizing; the other is sustained during agonist application.

Both of the AChRs that control chloride conductances are activated by nicotine and suberyldicholine and both are blocked by -bungarotoxin (-BTx); only the AChR that mediates the rapidly desensitizing chloride current is blocked by -conotoxin-ImI (-CTx-ImI), however (Kehoe and McIntosh 1998; Kehoe et al. 1976).

Here we provide pharmacological evidence that 8-lipoxygenase metabolism is activated by the nicotinic AChR that mediates the sustained chloride conductance. Several of the properties of the 8-lipoxygenase activation and the sustained chloride conductance are similar. Both are activated by suberyldicholine and nicotine, blocked by -BTx, and unaffected by -CTx-ImI. Because nicotinic receptors presumably are all ionotropic, the activation of 8-lipoxygenase metabolism by a nicotinic receptor is unexpected.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Biochemical experiments

Aplysia californica weighing 20-100 g were obtained from Marinus (Long Beach, CA) and the Miami Aplysia Facility (Miami, FL) and maintained in artificial sea water (ASW) at 15°C prior to dissection. Central ganglia were removed by dissection from animals anesthetized by injecting isotonic MgCl₂ (Schwartz and Swanson 1987). The ganglia were transferred to a small volume of ASW (Eisenstadt et al. 1973), chloride-free ASW [(in mM) 460 NaOH, 55 MgSO₄, 11 Ca(OH)₂, 10 KOH, 10 Tris base, and 546 methanesulfonic acid, pH adjusted to 7.6 with NaOH], or high Mg²⁺/low Ca²⁺ solution [(in mM) 230 NaCl, 10 KCl, 1 CaCl₂, 220 MgCl₂, and 10 HEPES plus 50× MEM amino acids solution (4 ml/l; without L-glutamine; Gibco, Grand Island, NY), 100× MEM nonessential amino acids solution (2 ml/l; 10 mM), and 100× MEM vitamin solution (5 ml/l), the pH adjusted to 7.6 at room temperature]. Neural components (neurons and neuropil) were obtained by removing the connective tissue sheaths of ganglia from 60 to 100 g animals. Ganglia were pinned to a silicone plastic (Sylgard, Dow Corning, Midland, MI) and the connective tissue sheath removed under a dissecting microscope.

The following identified neurons were isolated from mature animals weighing 20-50 g (Camardo et al. 1983): cells B3, 6, 8, 9, and 10 and certain neighboring cells from the buccal ganglia (Gardner and Kandel 1977); cells of the medial group in the pleural ganglia (Kehoe 1972a); and cells of the RB group of the abdominal ganglion (Kandel et al. 1967). Pleural, buccal, and abdominal ganglia were incubated with protease (10 mg/ml; Sigma Type IX; Sigma Chemical, St. Louis, MO) in modified Leibovitz 15 (L15; Sigma) medium. To make this medium, additional salts [(in mM) 385 NaCl, 10 KCl, 28 MgCl₂, 27 MgSO₄, 2.3 NaHCO₃, 35 glucose, and 11 CaCl₂] and penicillin-streptomycin (1% vol/vol; Gibco; 10,000 units of penicillin, 10 mg streptomycin/ml) were added to L15 medium, and the final solution was filtered through a 0.22-µm filter (Millipore Products Division, Bedford, MA). Ganglia were incubated with protease at 35°C for 90-150 min depending on the age of the animals. The treated ganglia were washed three times for at least 5 min each in modified L15 medium before the connective tissue sheath was removed with fine forceps. Cells were isolated from ganglia with a glass needle and transferred to glass test tubes containing ASW.

Labeling and incubation of neural tissue

In most experiments, membrane lipids were labeled by incubating the neural components with [³H]arachidonic acid (10 µCi/200 µl ASW; 100 µCi/mmol, DuPont/NEN, Boston, MA) in microcentrifuge tubes for 2 h at 15°C. Neural components were washed by gently replacing the incubation fluid with ASW containing the test agonist. Isolated cells were labeled immediately after the dissection for 2 h in glass tubes containing [³H]arachidonic acid in ASW (5 µCi/1 ml); in most experiments, these cells were washed to remove excess radioactivity, then rested overnight in ASW with Aplysia hemolymph (50:50, vol/vol). These methods result in the incorporation of [³H]arachidonic acid into membrane phospholipids. Piomelli et al. (1987a) have described the uptake and distribution of the label in Aplysia lipids.

To test the activity of agonists, radioactively labeled neural components or cells were incubated for 10 min at 15°C in ASW or in ASW containing an agonist [histamine, serotonin (5-HT), -aminobutyric acid (GABA), glutamate, octopamine, the neuropeptide Phe-Met-Arg-Phe-amide (FMRFamide), dopamine, ACh, myomodulin, arecoline, carbachol, or suberyldicholine, which were purchased from Sigma, Research Biochemicals International, Natick, MA, or Peninsula Laboratories, Belmont, CA]. Incubations were stopped by adding acetone (2 vol). For studies of antagonists, labeled isolated neurons or neural components were exposed to an antagonist before adding an agonist. The incubation was then continued for 10 min and stopped with acetone. The cholinergic antagonists tested were added 2 min before adding agonists and included tubocurarine, atropine, hexamethonium, tetraethylammonium (TEA), and -BTx, which were from Research Biochemicals International or Calbiochem (La Jolla, CA). -CTx-ImI (a gift of J. Michael McIntosh, University of Utah) was applied to neural components for 10 min and isolated cells for 15 min before testing an agonist. For experiments with pertussis toxin (PTx; List Biological Laboratories, Inc., Campbell, CA), neural components were incubated with [³H]arachidonic acid for 2 h, washed twice with ASW (500 µl) containing bovine serum albumin (0.5% wt/vol), and then incubated for 6 h with the holotoxin (0.1 µg/ml) or toxin that had been heat inactivated (100°C for 5 min).

Extraction of lipids

Before we extracted the lipids, we added 8-HETE or 12-HETE to each sample as internal standard. The samples were then acidified to pH 3.5 with HCl and extracted with diethyl ether (Steel et al. 1997). The lipid extract was evaporated to dryness under reduced pressure and resuspended in mobile phase for chromatographic analysis or stored in ethanol.

Reverse phase-high performance liquid chromatography (RP-HPLC)

Lipid extracts were fractionated on a Novapak C₁₈ column (Waters Chromatography, Milford, MA) eluted isocratically at 0.7 ml/min with acetonitrile/water (50:50, vol/vol; pH adjusted to 4.5 with acetic acid). Between injections, the column was washed with acetonitrile to elute nonpolar lipids. HPLC analyses were performed on a Hewlett-Packard 1090M (Hewlett-Packard Instruments, Paramus, NJ) with a photodiode array UV detector in series with a flow-through radioactivity monitor (-Ram, IN/US Systems, Tampa, FL). The internal standard was quantified by its absorbance at 235 nm and used to correct for losses during extraction. Radioactivity was measured with Scintflow software and normalized to the recovery of the internal standard. Samples were mixed with UniverSol scintillation fluid (ICN Biochemical, Costa Mesa, CA) in a ratio of 1:3.

Electrophysiological experiments

Electrophysiology was done on the neurons described in the preceding text for the isolated cell experiments. Ganglia were prepared as described by (Kehoe 1985). The ASW used in these studies contained (in mM) 480 NaCl, 10 KCl, 10 CaCl₂, 50 MgCl₂, and 10 Na-HEPES, pH 7.8. The composition of the chloride-free ASW was the same as given in the preceding text for the biochemical experiments. All drugs used were obtained from Sigma.

Recordings were usually made in the two-electrode voltage-clamp mode as described (Kehoe 1985). For the experiments designed to assess the role of a G protein in the chloride-dependent responses, whole cell patch-clamp methods were used (Hamill et al. 1981; Kehoe 1994). The internal solution in the patch pipettes consisted of (in mM) 411 K₂SO₄, 8.3 Na₂SO₄, 1 CaCl₂, 2 MgCl₂, and 5 EGTA (buffered with 20 mM K-HEPES to pH 7.4). For control recordings, GTP (1 mM) and ATP (10 mM) were included in the internal solution. In some experiments, GTP--S (10 mM) or GDP--S (10 mM) was also included with or without ATP and GTP. Continuous recordings were made on a Servogor 340 paper recorder and on a digital audio tape recorder. Records of agonist-elicited currents were digitized and sampled on-line using a Cambridge Electronic Design 1401 interface and the whole cell electrophysiology program from Strathclyde Electrophysiological Software. Agonists and antagonists were applied using the fast perfusion system described by Kehoe and McIntosh (1998).

Statistics

Lipid products were quantified and normalized to an internal standard. Data are reported as the means ± SE. Differences between means were assessed with a repeated measures ANOVA; P < 0.05 was taken as significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As shown in the representative HPLC trace of Fig. 1, application of ACh to Aplysia neuronal tissue induced the formation of 8-HETE and 8-KETE. The effect of ACh was dose dependent: lipid metabolites were produced at concentrations of the transmitter as low as 1 µM and increased up to 1 mM, the highest dose tested. The response to ACh was specific. We tested histamine [previously shown to activate 12-lipoxygenase in Aplysia neurons (Piomelli et al. 1987a)], 5-HT, GABA, glutamate, octopamine, dopamine, FMRFamide, and myomodulin. No neurotransmitter other than ACh led to a significant production of either 8-HETE or 8-KETE (Fig. 2). In addition, depolarization of neurons with KCl (60 mM) did not cause the generation of the metabolites. In two trials, the mean 8-HETE production was: ASW, 375 cpm; ACh (100 µM), 5,635 cpm; and KCl (60 mM), 690 cpm.

View larger version (18K): [in this window] [in a new window]   Fig. 1. Formation of metabolites of the 8-lipoxygenase pathway is stimulated by acetylcholine (ACh). A: standards of 8-HETE and 8-KETE were separated by reverse phase-high performance liquid chromatography (RP-HPLC) on a Novapak C18 column eluted with acetonitrile/water (50:50, vol/vol; pH adjusted to 4.5 with acetic acid) at a flow rate of 0.7 ml/min. Metabolites were detected by UV absorbance monitored at 235 and 270 nm. Steel et al. (1997) described the identification and characterization of these products. B and C: neural components were labeled with [3H]arachidonic acid (10 µCi) for 2 h, washed, and exposed to ACh [100 µM in artificial seawater (ASW)] or vehicle at 15°C for 10 min. Acetone was added to stop the reaction, and the incubation mix was acidified and lipids were extracted with diethyl ether and fractionated by RP-HPLC. Radioactive metabolites were detected with a flow-through radioactivity monitor. These traces are typical of those obtained in at least 25 independent experiments.

View larger version (27K): [in this window] [in a new window]   Fig. 2. The 8-lipoxygenase pathway is specifically activated through an ACh receptor (AChR). Neurotransmitters dissolved in ASW were tested for their ability to initiate the synthesis of 8-lipoxygenase metabolites. Aplysia neural components were labeled with [3H]arachidonic acid (10 µCi) for 2 h as described in the legend to Fig. 1 and exposed to histamine (50 µM), serotonin (5-HT, 50 µM), GABA (1 mM), glutamate (50 µM), octopamine (50 µM), Phe-Met-Arg-Phe-amide (FMRFamide, 50 µM), dopamine (10 µM), ACh (100 µM), myomodulin (10 µM), or ASW for 10 min. We chose doses of these transmitters known to elicit physiological responses in Aplysia. Lipids were extracted and analyzed as described in the legend to Fig. 1. Baseline release of radioactive products (in the presence of ASW alone) is indicated by the horizontal line. Baseline release of 8-HETE in a recent set of experiments was 1,716 ± 511 cpm (n = 9). Data are expressed as mean values ± SE from 3 to 29 independent experiments; *P < 0.05.

Pharmacology of 8-lipoxygenase activation

To identify the receptor that mediates activation of 8-lipoxygenase, we tested several cholinergic agonists and antagonists previously shown to be effective at Aplysia ACh receptors. Carbachol, a nonspecific cholinergic agonist, was as effective as ACh in activating this pathway (data not shown). Nicotine and suberyldicholine, which activate only the receptors that mediate chloride conductances (Ger and Zeimal 1977; Kehoe 1979; Kehoe and McIntosh 1998), also were effective (Fig. 3).

View larger version (17K): [in this window] [in a new window]   Fig. 3. Nicotine and suberyldicholine activate the 8-lipoxygenase pathway in Aplysia neural components. A: nicotine (1 µM) and ACh (100 µM) were applied for 10 min to neural components that had been labeled with [3H]arachidonic acid for 2 h as described in Fig. 1. Lipid products were then extracted and analyzed as described in Fig. 1. B: suberyldicholine (100 µM) also activated the 8-lipoxygenase pathway in the same assay. Data are expressed as mean values ± SE from 9 (A) or 8 (B) independent experiments. *P < 0.05; **P < 0.01.

Of the cholinergic antagonists tested, only -BTx blocked the activation of 8-lipoxygenase (Fig. 4). The concentration of toxin used in this experiment has been shown to block the ACh-activated chloride conductances in Aplysia (Kehoe et al. 1976) but does not interfere with the known metabotropic AChR. Tubocurarine, like atropine, hexamethonium, and TEA failed to inhibit the response (Fig. 4). Although tubocurarine has been shown to be a weak antagonist of the two ACh-activated chloride conductances (Kehoe and McIntosh 1998), concentrations higher than those used here would have been needed to block the sustained chloride-dependent response.

View larger version (34K): [in this window] [in a new window]   Fig. 4. -Bungarotoxin (-BTx) inhibits 8-lipoxygenase metabolism. AChR antagonists were tested for their ability to block receptor-mediated 8-lipoxygenase activation. Neural components were labeled with [3H]arachidonic acid (10 µCi) for 2 h as described in Fig. 1 and incubated with tubocurarine (100 µM), hexamethonium (100 µM), TEA (1 mM), atropine (100 µM), or -BTx (10 µM) for 2 min. ACh (100 µM) was then applied for 10 min, and the lipids were extracted and analyzed as in Fig. 1. Baseline release of radioactive products (in the presence of ASW with no addition) is indicated by the horizontal line. Data are expressed as mean values ± SE for 4 independent experiments. *P < 0.05.

Two pharmacologically distinct receptors activate ACh-mediated increases in chloride conductance: one mediates a rapidly desensitizing chloride conductance that is blocked by -CTx-ImI; the other, a sustained chloride conductance that is not affected by the toxin (Kehoe and McIntosh 1998). We found that -CTx-ImI did not inhibit the production of 8-lipoxygenase metabolites induced by ACh in intact neural components (Fig. 5) nor did the toxin block suberyldicholine-activated 8-lipoxygenase metabolism either in intact neural components or in isolated cells (Table 1).

View larger version (15K): [in this window] [in a new window]   Fig. 5. 8-lipoxygenase metabolism is not blocked by -conotoxin-ImI in Aplysia neurons. Neural components were dissected and labeled with [3H]arachidonic acid (10 µCi) for 2 h as described in Fig. 1 and treated with -conotoxin (-CTx, 10 µM) for 10 min before ACh (100 µM) was applied for 10 min more. Lipids were extracted and analyzed as described in Fig. 1. Data are expressed as mean values ± SE from 5 independent experiments. *P < 0.05.

Arecoline, which activates a G-protein-linked receptor mediating a potassium conductance but does not activate a chloride conductance in Aplysia neurons (Kehoe 1972b), also stimulated the production of 8-lipoxygenase products (Fig. 6). Arecoline failed to activate 8-lipoxygenase metabolism when the experiments were repeated in high Mg²⁺/low Ca²⁺ sea water (220 mM Mg²⁺/1 mM Ca²⁺), a condition under which chemical synaptic transmission is blocked (Gardner 1977). In contrast, under the same conditions, ACh-induced activation of 8-lipoxygenase metabolism persisted (Fig. 6). Thus arecoline does not appear to activate the 8-lipoxygenase-linked AChR directly, but rather indirectly through a polysynaptic, ACh-dependent mechanism. This idea was supported by the finding that -BTx blocks arecoline-induced 8-lipoxygenase activation (data not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 6. Arecoline stimulates 8-lipoxygenase metabolism. Aplysia ganglia were dissected in normal ASW or in a high Mg2+/low Ca2+ seawater. Neural components were then labeled with [3H]arachidonic acid for 2 h as described in the legend to Fig. 1 and incubated with ACh (100 µM; ) or arecoline (100 µM; ) for 10 min. The lipids were extracted and analyzed as in the legend to Fig. 1. Arecoline is without effect in the high Mg2+/low Ca2+ seawater. A plausible explanation is that arecoline stimulates 8-lipoxygenase metabolism polysynaptically. Data are expressed as percent of control release (corrected for losses during extraction) and are mean values ± SE for 9 independent experiments. *P < 0.01; **P < 0.05.

Co-localization of an ACh-induced chloride conductance with 8-lipoxygenase metabolism

The AChR that activates the -CTx-ImI-insensitive, sustained chloride conductance is differentially distributed in Aplysia neurons (Kehoe and McIntosh 1998). If this receptor is also linked to 8-lipoxygenase metabolism, 8-lipoxygenase activity would be expected to be similarly distributed. We therefore used isolated identified cell bodies to see if the receptors that mediate the chloride-dependent responses occur in the same cells as the ACh-induced production of 8-lipoxygenase metabolites. We first tested selected cells from the buccal ganglia and medial cells from pleural ganglia in which ACh elicits both of the chloride-dependent responses (Fig. 7, A and B, insets). These neurons generated metabolites of the 8-lipoxygenase pathway when exposed to nicotine (Fig. 7, A and B). ACh elicits only an inward, cationic current in RB cells from the abdominal ganglia (Fig. 7C, inset, bottom trace). Neither suberyldicholine (see inset) nor nicotine (not shown) elicits a change in membrane conductance. Thus AChRs that mediate chloride conductances are absent (Fig. 7C, inset, top trace). Therefore as expected, exposure of RB cells to several cholinergic agonists, including nicotine and suberyldicholine (Fig. 7C), failed to elicit 8-HETE production.

View larger version (20K): [in this window] [in a new window]   Fig. 7. Activation of 8-lipoxygenase occurs in Aplysia neurons with the ACh-activated chloride conductance. Identified neurons were transferred to ASW (1 ml) and labeled with [3H]arachidonic acid (5 µCi) for 2 h, washed, and rested overnight as described in METHODS. These cells were then incubated with nicotine (200 µM; buccal cells, A and medial cells, B) or suberyldicholine (100 µM; RB cells, C) for 10 min. Lipids were extracted and analyzed as described in the legend to Fig. 1. The retention time of the internal standard is indicated by the arrow. Differences in retention time occur due to variations in the pH of the mobile phase but do not affect the detection or quantification of the compounds of interest. In parallel experiments, cholinergic conductances activated in the various cell-types were recorded in conventional voltage clamp at a holding potential of 30 mV. A and B (insets): current was recorded from cells in the buccal ganglion and medial cells from the pleural ganglion during a 7-min application of nicotine. C (inset): suberyldicholine (top trace) and ACh (bottom trace) both were applied to RB cells for 10 s. Typical currents recorded under these conditions are shown. No conductance change is elicited by suberyldicholine (flat trace) and ACh activates only an inward current (i.e., a pure cationic response). These traces are typical of at least 6 independent experiments.

Chloride influx and ACh-induced 8-lipoxygenase metabolism

GABA and glutamate both induce increases in chloride conductance in Aplysia neurons (King and Carpenter 1989). We examined the action of these two transmitters using identified neurons known to have the AChRs that mediate the chloride conductances. GABA (1 mM) and glutamate (200 µM) were applied to these cells at concentrations sufficient to activate a chloride conductance similar in amplitude to that elicited by ACh in the same cell types. Neither transmitter activated 8-lipoxygenase metabolism, suggesting that chloride influx alone is not sufficient for activating lipid metabolism. To determine whether the influx of chloride is necessary for the activation of 8-lipoxygenase, we replaced chloride with an impermeant anion, methanesulfonic acid, in a modified, chloride-free ASW. Under these conditions, ACh failed to elicit influx of chloride ion and the generation of 8-lipoxygenase products was unaffected (Fig. 8).

View larger version (17K): [in this window] [in a new window]   Fig. 8. Chloride influx is not required for the production of 8-lipoxygenase metabolites. Neural components were bathed in ASW or chloride-free ASW (chloride substituted by methanesulfonic acid) and labeled with [3H]arachidonic acid (10 µCi) for 2 h as described in Fig. 1. Nicotine (100 µM) was applied and after 10 min, the lipids were extracted and analyzed as described in the legend to Fig. 1. Data are expressed as means ± SE from 11 independent experiments. *P < 0.05; **P < 0.01 vs. control.

Activation of 8-lipoxygenase is blocked by pertussis toxin

ACh-activated chloride conductances do not depend on a G protein since they persist during whole cell recordings made in the absence of ATP and GTP in the pipette solution. Under the same conditions, the potassium conductance that is known to be G-protein dependent is eliminated (Kehoe 1994). Furthermore, chloride-dependent responses persist in cells in which G proteins are blocked by GDP--S or are "desensitized" by GTP--S (data not shown). Is a G protein involved in ACh-induced 8-HETE synthesis? Shapiro et al. (1988) showed that agonist-induced release of arachidonic acid from Aplysia neural membrane lipids is mediated by G proteins and blocked by PTx. Suberyldicholine-activated production of 8-HETE was reduced when neural components are treated with PTx (Fig. 9A; 89% inhibition) but was not affected when inactivated PTx was used (data not shown). The release of free arachidonic acid in these experiments also was reduced when the neural components were treated with PTx (Fig. 9B; 65% inhibition), as would be expected if the activation of a phospholipase is blocked.

View larger version (14K): [in this window] [in a new window]   Fig. 9. Pertussis toxin blocks the suberyldicholine-activated release of arachidonic acid from neuronal membranes and its subsequent metabolism to 8-HETE. Neural components were labeled with [3H]arachidonic acid (10 µCi) for 2 h as described in Fig. 1, washed twice with ASW containing bovine serum albumin (0.5%), and then incubated with or without PTx (0.1 µg/ml) for 6 h. The neural components were then washed and exposed to suberyldicholine (100 µM) or ASW for 10 min. The lipids were extracted and analyzed as described in Fig. 1. Data are expressed as mean values ± SE from 7 independent experiments. A, 8-HETE; B, arachidonic acid. *P < 0.05;**P < 0.01 compared with all other treatments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The known lipoxygenases generate metabolites from arachidonic acid that can act as signaling molecules, both between cells and as second messengers. Thus 5-lipoxygenase produces the leukotrienes that play an important role in neutrophil function during inflammation and in asthma (Samuelsson 1983), while 8-lipoxygenase generates molecules that regulate oocyte maturation in some invertebrates (Holland and East 1985; Meijer et al. 1986).

In the nervous system, the actions produced by lipoxygenase metabolites are cell- and neurotransmitter-specific. In neurons of Aplysia, arachidonic acid is metabolized by 12-lipoxygenase to produce substances that mimic the actions of the neurotransmitter that activated their production. For example, application of 12-lipoxygenase metabolites produces a dual-action response in cell L14 (rapid depolarization followed by a slow hyperpolarization), mimicking the responses evoked by the neurotransmitter, histamine (Piomelli et al. 1989), and a lipid from this pathway causes a hyperpolarization in sensory neurons like that induced by FMRFamide, which results in desensitization (Piomelli et al. 1987b). We now show that 8-lipoxygenase products are generated in specific Aplysia neurons in response to ACh, and we have characterized the receptor pharmacologically. While we have not yet identified their physiological roles, we presume that these metabolites also act within neurons as second messengers or as intercellular or retrograde signaling molecules.

There are four AChR known to exist in Aplysia neurons. One metabotropic receptor mediates a potassium conductance, and three ionotropic receptors mediate a nonspecific cation conductance and two distinct chloride conductances (one that is rapidly desensitizing and the other that is sustained). The AChR described here that activates 8-lipoxygenase responds to the selective cholinergic agonists, suberyldicholine and nicotine, as well as to ACh but to no other neurotransmitter that we have tested. The ACh-induced 8-lipoxygenase metabolism is inhibited by -BTx but is unaffected by -CTx-ImI, thus corresponding pharmacologically to the receptor mediating the sustained chloride conductance. One apparent discrepancy is that 8-lipoxygenase metabolism was activated by arecoline in normal sea water. Since in Aplysia arecoline has been shown to activate a potassium conductance selectively (Kehoe 1972b) through a G-protein-linked receptor (Sasaki and Sato 1987), we suggest that the observed activation of 8-lipoxygenase by this agonist is produced polysynaptically. In support of this idea, we found that arecoline failed to activate the lipoxygenase in neural components bathed in high Mg²⁺/low Ca²⁺ sea water. This condition blocks synaptic transmission but not the 8-lipoxygenase metabolism induced by ACh. In addition, arecoline-induced 8-HETE production is blocked by -BTx, an antagonist that does not inhibit the metabotropic receptor known to be activated by arecoline. Rather the response to arecoline is the result of at least two AChRs linked polysynaptically, one a metabotropic arecoline receptor and the second the nicotinic AChR that mediates the sustained chloride response.

We found that activation of the sustained chloride conductance and activation of 8-lipoxygenase metabolism occur together in the same cells: both are observed in identified neurons from the buccal and pleural ganglia and both are absent from RB cells of the abdominal ganglia. This suggests that both the chloride conductance and the 8-lipoxygenase metabolism are activated by the same receptor. But are the chloride conductance and the lipid metabolism mediated by the same molecular entity? The pharmacological similarities between the biochemical and electrophysiological events triggered by ACh and their correlated expression in selected cells indicate that the two responses are mediated by the same AChR, even though the influx of chloride ion is neither necessary nor sufficient for activating 8-lipoxygenase.

A possible way in which an ionotropic receptor might trigger metabotropic activity is through a receptor-mediated increase in intracellular Ca²⁺. Some nicotinic AChRs are highly permeable to Ca²⁺ (McGehee and Role 1995), and Ca²⁺-dependent chloride conductances are present in many neurons. Moreover, release of arachidonic acid from membrane phospholipids is usually a prerequisite for its metabolism and the release often depends on the activation of a Ca²⁺-dependent phospholipase. Nevertheless, a role for Ca²⁺ in the ACh-induced activation of 8-lipoxygenase studied here can be excluded since there is no evidence for an ACh-dependent Ca²⁺ flux through the receptors that activate the sustained chloride conductance. First, there is no inward current in the selected cells used for these experiments, even when the agonists are applied by fast perfusion. Second, the rapid kinetics of the ACh-induced increase in chloride conductance strongly suggest that the change in conductance results from direct activation of a ligand-gated chloride channel. Finally, the ACh-induced chloride responses persist in external solutions that are Ca²⁺-free, as well as in cells that have been loaded with bis-(o-aminophenoxy)-N,N,N',N'-tetraacetic acid (BAPTA; 10 mM) (Kehoe, unpublished data).

The sustained chloride conductance is mediated by an ionotropic AChR, but the nicotinic AChR that activates 8-lipoxygenase appears to have metabotropic characteristics. Thus activation of 8-lipoxygenase depends on a PTx-sensitive G protein (Fig. 9) that presumably links the receptor to a phospholipase by a mechanism that is similar to the way that histamine causes arachidonic acid to be released in other identified Aplysia neurons (Shapiro et al. 1988; Vogel et al. 1989). In contrast, the ACh-induced chloride conductance does not depend on the activation of a G protein. Although no AChRs have yet been cloned from Aplysia, ionotropic and metabotropic receptors known to date have been found to be similar across phylogeny (Bertrand and Changeux 1995; Le Novere and Changeux 1999; Sargent 1993).

Ionotropic receptors are not usually linked to metabotropic activities by G proteins, but some exceptions have been reported. Both a postsynaptic neuronal AMPA receptor (Wang et al. 1997) and a presumptive presynaptic kainate receptor (Rodriguez-Moreno and Lerma 1998) appear to initiate metabotropic events through PTx-sensitive G proteins. In fact, Wang et al. (1997) immunoprecipitated the AMPA receptor together with a G protein, suggesting an unexpected physical interaction between this ionotropic receptor and a metabotropic effector molecule. It has further been shown that activation of the G protein is independent of the ion flux through the AMPA-activated channel. Further, it is interesting that the G-protein-linked kainate receptor activates a phospholipase C since the AChR that we have studied also must activate a phospholipase, although most likely a phospholipase A₂. Rather than a direct interaction between the ionotropic receptor and the G protein, it is possible that the AChR initiates the synthesis or release of an as yet unknown factor that subsequently activates the G protein causing lipid metabolism through the 8-lipoxygenase pathway. This mechanism has recently been proposed for the nicotine-induced activation of MAP kinase in small-cell lung carcinoma cells (Cattaneo et al. 1997).

Another possible explanation for the paradoxical nature of a nicotinic AChR having metabotropic properties would be a physical interaction between two receptor types as has recently been described. Liu et al. (2000) reported functional and biochemical evidence for a direct molecular interaction between the G-protein-linked D5 dopamine receptor and an ionotropic GABA_a receptor. In addition, there is growing evidence the receptors from different families may form oligomers, thus permitting unanticipated levels of cross-talk (Milligan 2000).