Identification and characterization of a tachykinin-containing neuroendocrine organ in the commissural ganglion of the crab Cancer productus

The tachykinins constitute one of the largest and most diverse groups of peptides in the animal kingdom. While the first sequenced family member was the molluscan peptide eledoisin (pEPSKDAFIGLMamide; Erspamer and Anastasi, 1962),perhaps the best known is the mammalian peptide substance P (RPKPQQFFGLMamide; Chang et al., 1971). Like all vertebrate tachykinins, and a few invertebrate isoforms sequenced from salivary tissues (e.g. eledoisin), these two peptides contain the carboxy(C)-terminal amino acid motif –FXGLMamide, where Xrepresents a variable amino acid (Nachman et al., 1999; Nässel,1999; Vanden Broeck et al.,1999; Severini et al.,2002). In invertebrates, a plethora of peptides containing the C-terminal amino acid motif–FX_IGX_IIRamide have been identified and collectively termed tachykinin-related peptides or TRPs(Nachman et al., 1999; Nässel, 1999; Vanden Broeck et al., 1999). Comparisons of sequence homology, tissue distribution, chemical/conformational requirements for receptor interaction and physiological function suggest a common evolutionary origin for both the vertebrate- and invertebrate-type peptides (Nachman et al.,1999; Nässel,1999; Vanden Broeck et al.,1999; Severini et al.,2002).

In most invertebrates, particularly insects, large numbers of species-specific TRP isoforms are common(Nachman et al., 1999; Nässel, 1999; Vanden Broeck et al., 1999; Severini et al., 2002). For example, 10 TRPs have been isolated from the cockroach Leucophea maderae (Muren and Nässel, 1996, 1997), seven from the honeybee Apis mellifera (Takeuchi et al.,2003), five from the locust Locusta migratoria (Schoofs et al., 1990a,b, 1993), five from the fruit fly Drosophila melanogaster (Siviter et al., 2000) and three from the mosquito Culex salinarius (Meola et al.,1998). In contrast to the diversity of species-specific TRPs present in insects, only a single isoform is thought to be present in decapod crustaceans (Christie et al.,1997a; Nieto et al.,1998; Li et al.,2002a; Huybrechts et al.,2003; Yasuda-Kamatani and Yasuda, 2004). This peptide, APSGFLGMRamide or Cancer borealis tachykinin-related peptide Ia (CabTRP Ia), has been isolated from or is predicted by cDNA to be present in the crab Cancer borealis, the shrimp Panaeus vannamei, the chelate marine lobster Homarus americanus, the freshwater crayfish Procambarus clarkii and the spiny lobster Panulirus interruptus(Christie et al., 1997a; Nieto et al., 1998; Li et al., 2002a; Huybrechts et al., 2003; Yasuda-Kamatani and Yasuda,2004).

While no antibody has been generated directly against CabTRP Ia, the peptide has been shown to cross-react with a rat monoclonal antibody to substance P and with several antibodies generated against insect TRPs(Cuello et al., 1979; Nässel, 1993; Christie et al., 1997a; Winther and Nässel,2001). Immunohistochemical studies using these antibodies have shown that CabTRP Ia is widely distributed within the nervous system of decapod crustaceans (Mancillas et al.,1981; Fingerman et al.,1985; Goldberg et al.,1988; Sandeman et al., 1990a,b;Schmidt and Ache, 1994, 1997; Blitz et al., 1995, 1999; Christie et al., 1995a, 1997a,b;Schmidt, 1997a,b; Langworthy et al., 1997; Fénelon et al., 1999; Glantz et al., 2000; Thirumalai and Marder, 2002; Pulver and Marder, 2002). In C. borealis, P. interruptus and H. americanus, one area of the nervous system that exhibits TRP immunoreactivity is the stomatogastric nervous system (STNS; Fig. 1),which controls the movement of the foregut musculature(Goldberg et al., 1988; Blitz et al., 1995, 1999; Christie et al., 1997a,b; Fénelon et al.,1999).

Within the STNS, one common and thus far unique feature of the CabTRP Ia labeling is a club-shaped plexus located in the anterior medial quadrant of each commissural ganglion (CoG; Goldberg et al., 1988; Fig. 1). Regardless of species, this structure originates from a fascicle of axons projecting from the circumoesophageal connective(coc) linking the CoG to the thoracic and abdominal nervous systems(Goldberg et al., 1988). The physiological role that this plexus plays in crustaceans is unknown. In the present report, we use a combination of anatomical, mass spectrometric and electrophysiological techniques to investigate the structural organization,co-transmitter complement and potential function(s) of this site in the Pacific red rock crab, Cancer productus. Collectively, our results show that this plexus is a neuroendocrine organ that provides hormonal CabTRP Ia modulation to the foregut musculature and, potentially, paracrine modulation to intrinsic CoG targets. Some of these data have appeared previously in abstract form (Messinger et al., 2004).

Cancer productus Randall were collected by hand, ring-trap or trawl at multiple locations in the greater Puget Sound area and San Juan Archipelago of Washington State (USA). Animals were maintained in either flow-through natural seawater tanks (Friday Harbor Laboratories, Friday Harbor, Washington, USA; ambient water temperature 8–12°C) or aerated natural seawater aquaria chilled to 10°C (Department of Biology,University of Washington, Seattle, WA, USA).

For tissue collection, crabs were anesthetized by packing in ice for 30–60 min, after which the dorsal carapace was removed and the foregut isolated from the animal. Following extraction of the foregut, the CoGs, and in some cases the entire STNS, were dissected free in chilled (approximately 10°C) physiological saline [composition in mmol l^–1: 440 NaCl; 11 KCl; 13 CaCl₂; 26 MgCl₂; 10 Hepes acid, pH 7.4(adjusted with NaOH)]. For some crabs, the eyestalks and dorsolateral walls of the pericardial chamber were also removed. From these structures, we isolated two well-known neuroendocrine organs: the sinus glands (SGs) and pericardial organs (POs).

Immunohistochemistry was performed as whole mounts. Specifically, dissected tissue was pinned in a Sylgard 184 (World Precision Instruments, Inc.,Sarasota, FL, USA; catalog #SYLG184)-lined Petri dish and fixed in a solution of either 4% paraformaldehyde (EM grade; Electron Microscopy Sciences,Hatfield, PA, USA; catalog #15710) in 0.1 mol l^–1 sodium phosphate (P) buffer (pH 7.4), 4%1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC; Sigma-Aldrich, St Louis,MO, USA; catalog #E-7750) in P, 4% paraformaldehyde and 1% EDAC in P or 100%methanol (HPLC grade; Sigma-Aldrich; catalog #27047-4), depending on the primary antibody being used (see below). All solutions containing paraformaldehyde or EDAC were prepared immediately prior to use, and tissue was fixed at 4°C. Methanol fixation was done at –20°C. Regardless of solution, tissues were fixed for 12–24 h, except as noted below (see Primary antibodies). Following fixation, tissue was rinsed five times over approximately five hours in a solution of P containing 0.3% Triton X-100 (P-Triton). Incubation in primary antibody (see below) was carried out in P-Triton, with 10% normal donkey serum (NDS; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA; catalog #017-000-121) added to diminish nonspecific binding. Following incubation in primary antibody, tissue was again rinsed five times over approximately five hours in P-Triton and then incubated in a 1:300 dilution of secondary antibody (see below) for 12–24 h. As with the primary antibody, incubation with secondary antibody was performed in P-Triton containing 10% NDS. After incubation in secondary antibody, preparations were rinsed five times over approximately five hours in P and mounted between a glass microscope slide and cover slip using Vectashield mounting medium (Vector Laboratories, Inc., Burlingame, CA,USA; catalog #H-1000). Incubations in primary and secondary antibodies were done using gentle agitation at 4°C. All tissues were rinsed at room temperature (18–24°C) without agitation. Secondary antibody incubation was conducted in the dark, as was all subsequent processing. Likewise, slides were stored in the dark at 4°C until examined.

Each of the primary antibodies employed in our study has been used previously to map the distribution of its respective antigen in crustacean/insect nervous systems and has been shown to be specific for its antigen. The references provided for each antibody describe its development/specificity and/or use in arthropod neural tissue.

For the detection of TRP, a rat monoclonal antibody to substance P (clone NC1/34 HL; Abcam Incorporated, Cambridge, MA, USA; catalog # ab6338; Cuello et al., 1979; Goldberg et al., 1988; Christie et al., 1997a) was used at a final dilution of 1:300. For the detection of the small molecule transmitter γ-aminobutryic acid (GABA), a rabbit polyclonal antibody(Sigma-Aldrich; catalog #A2052; Blitz et al., 1999; Swensen et al.,2000) was used at a final dilution of 1:500. For the detection of the amine dopamine, a mouse monoclonal antibody to its biosynthetic enzyme tyrosine hydroxylase (Immunostar Inc., Hudson, WI, USA; catalog # 22941; Tierney et al., 1999; Pulver and Marder, 2002) was used at a final dilution of 1:1000. For the detection of the amine histamine,a rabbit polyclonal antibody (Immunostar; catalog # 22939; Panula et al., 1988; Christie et al., 2004b) was used at a final dilution of 1:500. For the detection of the amine serotonin, a rabbit polyclonal antibody (Immunostar; catalog # 20080; Tierney et al., 1999; Pulver and Marder, 2002) was used at a final dilution of 1:500. To assay for the gas carbon monoxide, a rabbit polyclonal antibody to its biosynthetic enzyme heme oxygenase 2(Stressgen Biotechnologies Corp, Victoria, BC, Canada; catalog # OSA-200; Christie et al., 2003) was used at a final dilution of 1:100. To assay for the gas nitric oxide, a rabbit polyclonal antibody to its biosynthetic enzyme nitric oxide synthase (Affinity Bioreagents, Golden, CO, USA; catalog # PA1-039; Christie et al., 2003) was used at a final dilution of 1:300. For the detection of allatostatin-like peptides, a mouse monoclonal antibody (Dr B. Stay, University of Iowa, Iowa City, IA, USA; Stay et al.,1992; Woodhead et al.,1992; Pulver and Marder,2002) was used at a final dilution of 1:100. For the detection of buccalin-like peptides, a rabbit polyclonal antibody (Dr K. Weiss, Mount Sinai School of Medicine, New York, NY, USA; Miller et al., 1992; Christie et al., 1994) was used at a final dilution of 1:100. For the detection of cholecystokinin(CCK)-related peptides, a rabbit polyclonal antibody (Dr G. Turrigiano,Brandeis University, Waltham, MA, USA; Turrigiano and Selverston,1991; Christie et al.,1995b) was used at a final dilution of 1:300. For the detection of corazonin-related peptides, a rabbit polyclonal antibody (Dr J. Veenstra,Université Bordeaux 1, Talence cedex, France; Veenstra, 1991; Christie and Nusbaum, 1995)was used at a final dilution of 1:500. For the detection of crustacean cardioactive peptide (CCAP), a rabbit polyclonal antibody (Dr H. Dircksen,Stockholm University, Stockholm, Sweden; Dircksen and Keller, 1988; Stangier et al., 1988; Christie et al., 1995a) was used at a final dilution of 1:500. For the detection of FLRFamide-related peptides, a rabbit polyclonal antibody (Immunostar; catalog # 20091; Christie et al., 2004a) was used at a final dilution of 1:300. For the detection of myomodulin-like peptides, a rabbit polyclonal antibody (Dr K. Weiss; Miller et al., 1991; Christie et al., 1994) was used at a final dilution of 1:300. For the detection of orcokinins, a rabbit polyclonal antibody (Dr H. Dircksen; Bungart et al., 1994; Li et al., 2002b; Skiebe et al., 2002) was used at a final dilution of 1:5000. For the detection of β-pigment dispersing hormone (β-PDH)-related peptides, a rabbit polyclonal antibody (Dr K. Rao, University of West Florida, Pensacola, FL, USA; Dircksen et al., 1987; Mortin and Marder, 1991) was used at a final dilution of 1:1000. For the detection of proctolin, a rabbit polyclonal antibody (Dr D. Nässel, Stockholm University; code K9832/13; Johnson et al., 2003) was used at a final dilution of 1:500. For the detection of red pigment concentrating hormone (RPCH), a rabbit polyclonal antibody (Dr R. Elde, University of Minnesota, Minneapolis, MN, USA; Madsen et al., 1985; Nusbaum and Marder,1988) was used at a final dilution of 1:300.

The fixative used for all antibodies except for those generated against tyrosine hydroxylase, histamine and CCAP was 4% paraformaldehyde (tissue for GABA immunoprocessing fixed for 2–3 h only). Fixation for tyrosine hydroxylase utilized 100% methanol, while 4% EDAC was employed for histamine,and a combination of 4% paraformaldehyde and 1% EDAC was used for CCAP.

The secondary antibodies used in our experiments were donkey anti-rat immunoglobulin G (IgG) conjugated to Alexa Fluor 488 (Molecular Probes,Eugene, OR, USA; catalog #A-21208) or Alexa Fluor 594 (Molecular Probes;catalog #A-21209), donkey anti-rabbit IgG conjugated to Alexa Fluor 488(Molecular Probes; catalog #A-21206) or Alexa Fluor 594 (Molecular Probes;catalog #A-21207) and donkey anti-mouse IgG conjugated to Alexa Fluor 488(Molecular Probes; catalog #A-21202) or Alexa Fluor 594 (Molecular Probes;catalog #A-21203).

Fluorescently labeled tissue was viewed and data collected using one of two Bio-Rad MRC 600 laser scanning confocal microscopes (Bio-Rad Microscience Limited, Hemel Hempstead, UK), a Bio-Rad Radiance 2000 laser scanning confocal microscope or a Nikon Eclipse E600 epifluorescence microscope (Tokyo, Japan). The Bio-Rad MRC 600 system at Friday Harbor Laboratories is equipped with a Nikon Diaphot inverted microscope and a krypton/argon mixed gas laser. Nikon Fluor 10× 0.5NA dry, PlanApo 20× 0.75NA dry, Nikon Fluor 40×0.85NA dry and PlanApo 60× 1.4NA oil immersion objective lenses and Bio-Rad-supplied BHS, YHS and K1/K2 filter sets and Comos software were used for imaging preparations on this system (filter specifications are as described in Christie et al.,1997b). The Bio-Rad MRC 600 system located at the University of Washington (Department of Biology) is equipped with a Nikon Optiphot upright microscope and, with the exception of the Nikon Fluor 40× 0.85NA dry lens, uses the same laser, filters, software and objective lenses as the MRC 600 system located at Friday Harbor Laboratories. The Bio-Rad Radiance 2000 system is equipped with a modified Nikon Eclipse E600FN microscope and a krypton/argon mixed gas laser (568 nm excitation line used). For imaging on this system, Nikon PlanApo 10× 0.45NA DIC dry, PlanApo 20× 0.75NA DIC dry and PlanApo 60× 1.4NA DIC oil immersion objective lenses as well as a Bio-Rad-supplied E600LP emission filter and Bio-Rad LaserSharp 2000 software were used. The Nikon Eclipse E600 epifluorescence microscope is equipped with Nikon PlanFluor 10× 0.30NA, PlanFluor 20× 0.50NA and PlanFluor 40× 0.75NA dry objective lenses and ENDOW GFP HYQ (EX,450–490 nm; DM, 495 nm; BA, 500–550 nm) and G-2E/C TRITC (EX,528–553 nm; DM, 565 nm; BA, 600–660 nm) filter sets.

India ink injection into the circulatory system has long been used to map the distribution of blood vessels in neural tissue(Lane et al., 1981; Renkin et al., 1981; Renkin, 1985; Farley, 1990; Hogers et al., 1995; Murray and Wilson, 1997; Grivas et al., 2003; Sasaki et al., 2003; Cerri et al., 2004; Hutchinson and Savitzky, 2004; Marinkovic et al., 2004). Here, we developed a method using this reagent to visualize the distribution of hemolymph vessels and lacunae in the stomatogastric neuromuscular system. Specifically, Higgins Fountain Pen India ink (Eberhard Faber Inc., Lewisberg,TN, USA; catalog #723) or Koh-I-Noor Fount India drawing ink (Koh-I-Noor,Bloomsbury, NJ, USA; catalog #9150-D) was dried down and then reconstituted in a similar amount of physiological saline. The resulting saline/ink solution was injected into the pericardial chamber using a 22-gauge needle attached to a 1-ml plastic syringe via penetration through the junction of the thorax and abdomen. For small animals (<120 g), 200 μl of saline/ink solution was injected into each individual. Intermediate-sized animals(120–250 g) were injected with 500 μl and large animals (>250 g)received a 1 ml injection of the saline/ink solution. Following the injection,animals were returned to their tanks for 1–4 h, then anesthetized and dissected as described earlier. Nervous system tissue was then pinned flat in a Sylgard-lined Petri dish containing physiological saline, and ink infiltration into lacunae was visualized using either a Nikon SMZ800 or Nikon SMZ1000 stereomicroscope with incident illumination provided by a Fiber-Lite Model 190 fiber optic illuminator (Dolan-Jenner Industries, Inc., Woburn, MA,USA). Micrographs of ink infiltration were taken using either a Nikon CoolPix 4500 digital camera mounted on the SMZ800 microscope or a Nikon CoolSNAP digital camera mounted on the SMZ1000 microscope.

Following ink infiltration, some CoGs were fixed and immunoprocessed with the substance P antibody to assess the coincidence of the hemolymph lacunae and the TRP immunoreactivity. These ink-filled/substance P immunoprocessed ganglia were imaged using the Bio-Rad Radiance 2000 confocal system, with ink-filling visualized via its transmitted light detector.

For the mass spectrometric identification of TRP isoforms in the tachykininergic plexus of the CoG, ganglia were dissected and pinned in a Sylgard-lined Petri dish containing chilled physiological saline. Incident illumination was used to identify the location of the plexus in each ganglion,then the site was subsequently isolated and placed in acidified methanol [90%methanol (HPLC grade; Sigma-Aldrich), 9% glacial acetic acid (sequencing grade; Fisher Scientific, Fair Lawn, NJ, USA; catalog #BP1185-500) and 1%water (HPLC grade; Sigma-Aldrich; catalog #27073-3)]. Tissue from 20 ganglia was pooled in approximately 200 μl of acidified methanol and stored at–80°C until utilized for analysis.

Direct tissue analysis of individual tissue pieces was performed as previously described (Kutz et al.,2004). Briefly, a tissue fragment was desalted in 10 mg ml^–1 2,5-dihydroxybenzoic acid (DHB; ICN Biomedicals Incorporated, Costa Mesa, CA, USA; catalog #190209). Next, 0.2 μl of saturated DHB solution (in 50:50 methanol:water, v/v) was added to one facet of a matrix-assisted laser desorption/ionization (MALDI) target plate, and the desalted tissue fragment quickly placed into the matrix. The tissue-containing matrix was then crystallized at room temperature.

Mass spectrometric analysis of peptides was carried out using an IonSpec HighRES MALDI Fourier transform mass spectrometer (FTMS; IonSpec Corporation,Lake Forest, CA, USA). In order to increase the signal-to-noise ratio, in-cell accumulation was performed to allow multiple packets of ions resulting from the tissue sample to accumulate in the analyzer cell prior to detection. For the detection of low-mass peptides (i.e. approximately 1000 m/z), a pulse sequence was used to more efficiently transport ions into the analyzer cell(Kutz et al., 2004).

To determine if CabTRP Ia is a circulating hormone, we used MALDI-FTMS to assay the hemolymph. Hemolymph was collected by inserting a 22-gauge needle attached to a 3-ml plastic syringe through the junction of the thorax and abdomen into the pericardial chamber. A fresh needle and syringe were used for removal of hemolymph from each animal. Approximately 2 ml of hemolymph was withdrawn from each animal and immediately placed in twice its volume of acidified methanol and vortexed for 3 min at 10°C using a Thermolyne Maxi Mix II tabletop vortexer (Barnstead/Thermolyne, Dubuque, IA, USA). Following vortexing, the hemolymph/acidified methanol mixture was centrifuged for 5 min at 15 800 g using an Eppendorf 5415C tabletop centrifuge(Eppendorf AG, Hamburg, Germany), also at 10°C. After centrifugation, the resulting supernatant was removed and stored at –80°C until utilized for analysis.

Prior to mass spectrometric analysis, large proteins and salts were removed from the extracted hemolymph. Large proteins were removed by placing 500 μl of crude extract in a 10 000 Da molecular mass cutoff tube (Argos Technologies, catalog #VS0101) and centrifuging it at 16 100 gfor 10 min at room temperature. The resulting low-molecular-mass filtrate was concentrated using a Savant SC 110 SpeedVac concentrator (Thermo Electron Corporation, West Palm Beach, FL, USA) and then resuspended in 10 μl of 0.1% formic acid (puriss grade; Sigma-Aldrich; catalog #94318). The acidified sample was desalted by passing it through a ZipTip_C18 pipette tip(Millipore, Billerica, MA, USA; catalog #ZTC18S096) and eluting the bound peptides with 4 μl of 50% acetonitrile. Desalted extract was mixed 1:1 with DHB matrix on a MALDI plate and allowed to crystallize at room temperature,after which MALDI-FTMS analysis was performed as per the CoG fragments.

Neuromuscular preparations were dissected from the C. productusforegut and pinned flat in 5-ml Sylgard-lined Petri dishes. Nerves and muscles were identified according to the nomenclature of Maynard and Dando(1974). During recording sessions, the bath volume was maintained at approximately 3 ml and preparations were continuously superfused (4–5 ml min^–1) with physiological saline. (It should be noted that the saline used in the physiological experiments was buffered using 11.2 mmol l^–1 Trizma base and 5.1 mmol l^–1 maleic acid rather than the 10 mmol l^–1 Hepes acid used in the saline employed for our anatomical and mass spectrometric experiments.) CabTRP Ia was bath-applied by means of a switching port at the inflow of the superfusion system. This peptide was synthesized and purified using standard techniques(Christie et al., 1997a) by the Cancer Research Center of the University of Pennsylvania School of Medicine (Philadelphia, PA, USA) and was a gift from Dr Michael P. Nusbaum(Department of Neuroscience, University of Pennsylvania School of Medicine). Synthetic CabTRP Ia was dissolved in distilled water at a concentration of 10^–3 mol l^–1 and stored at –20°C. Immediately before use, samples of dissolved peptide were thawed and diluted to final bath concentrations(10^–9–10^–7 mol l^–1)in physiological saline. During all experiments, the saline temperature was cooled with an ice bath and regulated to within a few tenths of a degree at a temperature of approximately 10°C.

Measurements of excitatory junctional potentials (EJPs) were made from the gastric mill 4 (gm4), gastric mill 5a (gm5a), gastric mill 6a (gm6a), gastric mill 8a (gm8a), pyloric 1 (p1) and pyloric 2 (p2) muscles using conventional 2 mol l^–1 potassium acetate-filled microelectrodes with resistances of 7–10 MΩ. The gm4 muscle is innervated by the dorsal gastric (DG) neuron via the dorsal gastric nerve (dgn). The gm5a muscle is innervated by the inferior cardiac (IC) neuron via the medial ventricular nerve (mvn). The other four muscles are innervated via the lateral ventricular nerve (lvn); gm6a and gm8a by the lateral gastric (LG) neuron, p1 by the lateral pyloric (LP) neuron, and p2 by the pyloric (PY) neurons (Maynard and Dando, 1974; Selverston and Moulins, 1987; Weimann et al., 1991; Harris-Warrick et al., 1992). Innervating nerves were stimulated extracellularly via stainless steel pin electrodes driven by an A-M systems Model 2100 isolated pulse stimulator (A-M Systems, Carlsborg, WA, USA). EJPs were measured with an Axoclamp 2-B intracellular amplifier (Axon Instruments, Union City, CA, USA), amplified 10-fold using a Model 440 instrumentation amplifier(Brownlee Precision, San Jose, CA, USA) and recorded using a Digidata 1322A acquisition system (Axon Instruments). EJP amplitude was defined as the peak membrane potential of the EJP relative to the baseline potential prior to stimulation. EJP amplitudes were analyzed using routines written in Matlab(The MathWorks, Natick, MA, USA).

Recordings of contractions from the gm8 muscle were obtained using a FT03 force displacement transducer (Astro-Med, West Warwick, RI, USA). The neuromuscular preparation consisted of fibers of both the gm8a and gm8b muscles, which were not separated in order to minimize damage to the muscle fibers and innervating nerves. One of the muscle insertions was pinned down to the Sylgard in the recording dish, while the other insertion was tied to the transducer with a short piece (∼3 cm) of size 6/0 silk suture thread (Fine Science Tools, Foster City, CA, USA). The transducer was positioned so that the muscle was stretched just past its relaxed length. The innervating nerve was stimulated with a train of pulses applied through the pin electrode. This resulted in muscle shortening, and the force transducer measured the increased tension. The transducer signal was amplified by a factor of 10 000 using the Brownlee Precision Model 440 amplifier and recorded using the Digidata interface board. Contraction amplitudes were analyzed using the Clampfit program (Axon Instruments).

In the experiments in which the effect of a single concentration(10^–7 mol l^–1) of CabTRP Ia on the EJP amplitude or contraction of a muscle was tested, a paired t-test was used to test for statistical significance. Error bars on plots correspond to standard errors.

Anatomy figures were produced using a combination of Photoshop (version 7.0; Adobe Systems Inc., San Jose, CA, USA) and Canvas (version 8.0; Deneba Systems Inc., Miami, FL, USA) software. Contrast and brightness were adjusted as needed to optimize the clarity of the printed images. Mass spectra were collected using IonSpec99 version 7.0 software (IonSpec Corp.). Boston University Data Analysis (BUDA) version 1.4 was used to export the spectra as a bitmap into Macromedia Fireworks MX 2004 Version 7.0 (Macromedia Incorporated, San Francisco, CA, USA). Resolution was increased and peaks were labeled with mass and peptide identity in Fireworks. Physiology figures were produced using Sigma Plot (version 8.0; Systat Software, Point Richmond, CA,USA).

To determine the distribution of TRPs in the STNS of C. productus,we immunoprocessed this portion of the nervous system with a rat monoclonal antibody generated against the vertebrate tachykinin substance P(Cuello et al., 1979). This antibody was used previously for mapping the distribution of TRPs in the STNS of the crab C. borealis, the spiny lobster P. interruptusand the chelate lobster H. americanus(Goldberg et al., 1988; Blitz et al., 1995; Christie et al., 1997a) and is known to cross-react with the peptide CabTRP Ia, the only TRP thus far isolated from crustaceans and probably the sole isoform of this peptide family present in this group of arthropods(Christie et al., 1997a; Nieto et al., 1998; Li et al., 2002a; Huybrechts et al., 2003; Yasuda-Kamatani and Yasuda,2004).

Within the STNS of C. productus, numerous TRP-immunopositive profiles were identified (Fig. 1). As with the previously mapped crab and lobster species(Goldberg et al., 1988), the most striking of these structures is a club-shaped plexus located in the anterior medial portion of each CoG (Figs 1, 2; N=148 ganglia). Labeling in the plexus consisted of a dense aggregation of nerve terminals that originate from a fascicle of small (<1 μm)-diameter axons projecting from the coc connecting the CoG to the thoracic nervous system (Fig. 2A,B). On average,the individual terminals that comprise the plexus were rather elongate, with major cross-sectional diameters of <1 to >15 μm. In the posterior portion of the plexus, the terminal aggregates were often extensively fenestrated by unstained tubular areas(Fig. 2C).

In addition to the plexus just described, a number of other TRP-immunopositive structures were also seen in the STNS (N=41 intact nervous systems examined). The distribution of this immunoreactivity is summarized schematically in Fig. 1. Seven immunopositive somata were present in each CoG(Fig. 2A; three somata shown,with a single labeled soma present in the oesophageal ganglion). Immunopositive axons were present in the inferior oesophageal (ion;one axon per ion), oesophageal (on; two axons) and stomatogastric (stn; two axons) nerves, and neuropil was present in each CoG (Fig. 2A), at the junction of the superior oesophageal nerves (sons) with the on and the stn and also in the stomatogastric ganglion(STG). No TRP labeling was seen in any of the motor nerves posterior to the STG nor was any immunoreactivity seen in the anterior cardiac plexus (ACP), a neuroendocrine structure located on each of the paired anterior cardiac nerves(Christie et al., 2004a).

In addition to examining the STNS for the presence of TRP-immunopositive profiles, we also examined two well-known neuroendocrine organs located outside this system, namely the SG and PO, for TRP-like labeling. In brachyuran crabs, the SG is recognizable as a superficial, intensely iridescent structure located between the medulla interna and medulla externa of the eyestalk ganglia (Cooke and Sullivan, 1982; Beltz,1988; Fingerman,1992). The PO is located on the dorsolateral wall of the pericardial chamber and, in brachyurans, consists of two or more longitudinal nerve trunks that are interconnected by vertical nerve bars(Cooke and Sullivan, 1982; Beltz, 1988; Fingerman, 1992). In neither the SG (N=6) nor the PO (N=6) was any TRP immunoreactivity evident (data not shown).

The existence of hemolymph sinuses and lacunae penetrating deep into the ganglia of crustaceans is well documented (Abbott, 1971, 1972; Skinner, 1985; McGaw and Reiber, 2002). In the blue crab, Callinectes sapidus, mapping by radiography and corrosion casting showed that the CoGs are among the ganglia fenestrated by hemolymph lacunae (McGaw and Reiber,2002). As just described, the TRP-like labeling in the posterior portion of the C. productus CoG plexus is fenestrated by tubular unlabeled areas. One explanation for these regions of immunolabel avoidance is that they are hemolymph lacunae. To investigate this hypothesis, we injected physiological saline containing India ink into the pericardial chamber surrounding the heart and allowed it to be pumped throughout the vascular system, thereby filling hemolymph vessels with an easily visualized substrate.

As can be seen in the ganglion shown in Fig. 3, numerous ink-filled lacunae are present in the CoG. While the location, size, shape and depth of invagination of the individual lacunae varied between ganglia, each CoG taken from an ink-injected animal (N=14 ganglia) possessed some deeply invaginated channels in its anterior medial quadrant, particularly in the posterior portion of this region.

To confirm that the ink-filled lacunae just described are coincident with the areas of immunolabel avoidance seen in the CoG plexus, we fixed and immunoprocessed a number of ink-filled ganglia (N=6) with the substance P antibody. Upon fixation, ink-filling was often lost in some of the smaller lacunae, particularly the superficially located ones. Irrespectively,in each of the immunoprocessed ganglia, many of the unstained regions fenestrating the label in the posterior portion of the plexus were found to contain ink (Fig. 4),thereby confirming that these areas of immunolabel avoidance are hemolymph lacunae. This organization (i.e. nerve terminals in direct apposition to hemolymph lacunae/sinuses) is similar to that of the SG(Cooke and Sullivan, 1982; Beltz, 1988; Fingerman, 1992). Thus, this finding suggests strongly that the TRP-immunopositive CoG plexus, like the SG,is a neuroendocrine release site. Based on its location within the nervous system, we have named this structure the anterior commissural organ or ACO.

Previous work has shown that most crustacean neuroendocrine sites contain large and often complex complements of signaling molecules(Christie et al., 1995a; Li et al., 2002a, 2003; Fu et al., 2005a,b). In an attempt to determine what other substances might be co-localized with TRP in the ACO, we immunolabeled CoGs with antibodies generated against a number of known crustacean hormones and/or neuromodulators and examined them for labeling of the plexus. In some preparations, double-labeling with substance P antibody was undertaken to unambiguously delimit the location of the ACO. The assayed substances were the small molecule transmitter GABA, the amines dopamine, histamine and serotonin, the gases carbon monoxide and nitric oxide as well as the peptides allatostatin, buccalin, CCK, corazonin, CCAP,FLRFamide, myomodulin, orcokinin, β-PDH, proctolin and RPCH. While each of the antibodies labeled neuronal profiles in the CoG and/or other regions of the C. productus nervous system (data not shown), none gave rise to labeling within the ACO (N≥6 ganglia for each antibody; Fig. 5).

Interestingly, while no co-transmitters were found in the ACO, our double-label experiments did show that this structure is surrounded in three dimensions by neuropil containing many other neuroactive compounds(Fig. 5). In many areas, these immunopositive neuropils directly abut ACO terminals. Thus, in addition to neuroendocrine release, the ACO may also function to modulate the CoG neuropil in a paracrine fashion.

As shown earlier, the ACO is immunolabeled by an antibody generated against the vertebrate tachykinin substance P. In crustaceans, the only TRP thus far identified is APSGFLGMRamide, commonly referred to as CabTRP Ia(Christie et al., 1997a; Nieto et al., 1998; Li et al., 2002a; Huybrechts et al., 2003; Yasuda-Kamatani and Yasuda,2004). Previous studies have shown that the substance P antibody cross-reacts with this peptide (Christie et al., 1997a); thus, it is likely that the substance P immunoreactivity seen in the ACO results from the presence of authentic CabTRP Ia in this structure. To confirm that this is indeed the case, we isolated the quadrant of the CoG containing the ACO and subjected it to direct tissue MALDI-FTMS analysis.

As the mass spectrum in Fig. 6A shows, many peptides are present in the portion of the CoG containing the ACO. Among the detected peaks are ones corresponding to a number of previously identified peptides including several isoforms of orcokinins, corazonin, three FLRFamide-related peptides,Gly¹-SIFamide and RPCH. Of particular interest is the mass/charge(m/z) peak at 934.501, which is essentially identical to the theoretical m/z of authentic CabTRP Ia (i.e. 934.493). This result is strongly suggestive of CabTRP Ia being present in authentic form in the anterior medial quadrant of the CoG and hence in the ACO.

Since the ACO possesses the organization of a neuroendocrine organ, it is logical to expect that the CabTRP Ia contained within it would be released into the circulatory system. As none of the other known neuroendocrine organs in C. productus (i.e. the SG, the PO and the ACP) exhibits any evidence of TRP immunoreactivity, the ACO may well be the sole source of this peptide in the hemolymph. To determine if CabTRP Ia does circulate, we collected hemolymph samples from three crabs and subjected the extract to MALDI-FTMS analysis. In two of three animals, the mass spectra of the hemolymph extract showed an m/z peak corresponding to that of authentic CabTRP Ia (i.e. m/z 934.491 in Fig. 6B). This finding strongly supports the notion that this peptide is a circulating hormone potentially released from the ACO.

In the crab C. borealis, Jorge-Rivera and colleagues showed that many peptides that do not directly innervate the muscles of the gastric mill and pyloric regions of the foregut are nonetheless potent modulators of the musculature (Jorge-Rivera and Marder, 1996, 1997; Jorge-Rivera, 1997; Jorge-Rivera et al., 1998). As shown by our schematic diagram of substance P immunolabeling in the STNS, no direct TRP innervation of the gastric mill or pyloric muscles is present in C. productus (Fig. 1). To determine if circulating CabTRP Ia can similarly modulate stomatogastric muscles in C. productus, we studied its actions on EJP and contraction amplitude for a subset of the intrinsic muscles of the foregut.

We tested the effect of 10^–7 mol l^–1CabTRP Ia on the amplitude of nerve-evoked EJPs in six stomatogastric muscles(Fig. 7). Three of the muscles(gm4, gm6a and gm8a) are innervated by motor neurons (i.e. DG or LG) involved in generating the gastric mill motor pattern, while the other three muscles(gm5a, p1 and p2) are innervated by neurons (IC, LP or PY, respectively) that participate in the pyloric circuit(Maynard and Dando, 1974; Selverston and Moulins, 1987; Weimann et al., 1991; Harris-Warrick et al., 1992). For all muscles, a statistically significant and reversible increase in EJP amplitude was noted (gm4, N=5; gm5a, N=6; gm6a, N=8; gm8a, N=6; p1, N=8; p2, N=6). For all muscles, the maximum effect of CabTRP Ia occurred approximately 5–10 min after bath application was initiated and required at least 30 min of rinsing in physiological saline to return to baseline. No effect of CabTRP Ia on membrane potential was observed for any of the muscles tested.

Since gm8a showed the largest increase in EJP amplitude in the presence of CabTRP Ia, we expected that nerve-evoked contractions of this muscle might also be potentiated due to application of the peptide. For these contraction experiments, the preparation consisted of fibers of both the gm8a and gm8b muscles, which were not separated so as not to damage the muscle fibers or innervating nerve. As Fig. 8shows, bath application of 10^–7 mol l^–1CabTRP Ia produced a statistically significant increase in the tension exerted by the gm8 muscle in response to a 3 s duration, 10 Hz stimulation of the dgn (a frequency chosen to be comparable with that of measured LG activity during a gastric mill rhythm; Beenhakker et al., 2004).

We had hoped to construct a dose–response curve characterizing the concentration dependence of the effect of CabTRP Ia on gm8 contraction but this was hindered by dramatic and consistent desensitization to the applied peptide. While the effects of a short application (5 min) of 10^–7 mol l^–1 CabTRP Ia could be reversed with a 10–15 min rinse of control saline, a second application of 10^–7 mol l^–1 CabTRP Ia an hour later resulted in a much smaller (∼50%) increase in contraction. However, in measurements on individual preparations, we found that application of 10^–8 mol l^–1 CabTRP Ia increased contraction appreciably in each of three experiments (mean increase=35%), while the contraction increased in only one out of five experiments when 10^–9 mol l^–1 CabTRP Ia was applied. Hence,we conclude that the threshold concentration for a reproducible effect of CabTRP Ia on gm8 contraction is between 10^–9 and 10^–8 mol l^–1.

To date, a number of well-delimited neuroendocrine organs have been identified in the nervous systems of decapod crustaceans(Cooke and Sullivan, 1982; Beltz, 1988; Fingerman, 1992; Christie et al., 2004a). Included among these structures are the SG, the PO, the ACP and the post-commissural organ (PCO). In terms of gross organization, each of these structures is composed of aggregations of nerve terminals that are in direct apposition to the circulatory system. For some sites (i.e. the PO, the ACP and the PCO), proximity to the circulatory system is achieved by possessing a superficial location just below the sheath of the nerves containing/comprising them (Cooke and Sullivan,1982; Beltz, 1988; Fingerman, 1992; Christie et al., 2004a). For neuroendocrine sites whose terminals are not superficially located (i.e. the SG), invaginating hemolymph lacunae/sinuses provide released hormones with access to the hemolymph (Cooke and Sullivan, 1982; Beltz,1988; Fingerman,1992).

In the present study, we have shown that the ACO of C. productusconsists of aggregations of tachykinin-immunopositive nerve terminals that directly abut/surround hemolymph lacunae. Thus, the general organization of this structure is homologous to those of other crustacean neuroendocrine sites, and to the SG in particular. Based on this homology, we believe that,like the SG, the ACO is a neuroendocrine release organ. Our identification of the ACO as a neuroendocrine center now brings the number of characterized neuroendocrine organs in C. productus to four (i.e. the SG, PO, ACP and ACO; Christie et al.,2004a; Christie and Messinger,2005; Fu et al., 2005a,b; Messinger et al., 2005).

Given that structures apparently homologous to the C. productusACO have been identified in species from six infraorders of decapods [i.e. Brachyura (true crabs), Anomura (hermit crabs, porcelain crabs and squat lobsters), Palinura (spiny lobsters), Thalassinidea (ghost shrimp), Astacidea(chelate lobsters and freshwater crayfish) and Caridea (caridean shrimp); Goldberg et al., 1988; Messinger et al., 2004], the ACO appears to be a highly conserved structure and we hypothesize that its function as a neuroendocrine center is conserved in other decapod species.

Recently, much work has focused on the identification of hormone complements in the neuroendocrine organs (i.e. the SG, PO and ACP) of C. productus (Hsu et al., 2004, 2005a,b; de la Iglesia et al., 2005; Fu et al., 2005a,b; Messinger et al., 2005). Like those of other crustacean species(Christie et al., 1995a; Li et al., 2002a, 2003), each of these sites has been shown to contain multiple neuroactive compounds. For example, using a combination of anatomical, mass spectrometric and molecular techniques, over 50 peptide hormones have been identified from the C. productus SG(Hsu et al., 2004, 2005a,b; de la Iglesia et al., 2005; Fu et al., 2005a,b). Similarly, over 40 putative hormones (peptides, amines and small molecules)have been identified in the C. productus PO(Hsu et al., 2004; Fu et al., 2005b). Even in the ACP, which is innervated by just four neurons, multiple hormones have been identified (Christie et al.,2004a; Hsu et al.,2004; Christie and Messinger,2005; Messinger et al.,2005).

In contrast to the chemical complexity seen in the hormone complements of the other neuroendocrine organs of C. productus, only CabTRP Ia has been identified as a putative hormone in the ACO. Immunohistochemistry using antibodies to other known crustacean hormones/neuromodulators failed to identify any co-transmitters in this site. However, in our mass spectrometric characterization of the anterior medial quadrant of the CoG that contains the ACO, we detected a number of peptides for which no antibodies are currently available (e.g. Gly¹-SIFamide and HLGSLYRamide) as well as a number of currently unidentified ones, and it is certainly possible that some of these peptides are contained within and utilized as signaling agents by the ACO.

Thus far, the only neuroactive compound identified in the C. productus ACO is CabTRP Ia. In our study, we show via mass spectrometry that this peptide is also present in the hemolymph of this crab. Using immunohistochemistry, we screened the other known neuroendocrine organs of C. productus (i.e. the SG, PO and ACP) for the presence of CabTRP Ia and found that each tissue lacked labeling. Moreover, mass spectrometric analysis also failed to detect this peptide in these sites(Christie and Messinger, 2005; Fu et al., 2005b; Messinger et al., 2005). Collectively, these findings suggest that if the CabTRP Ia seen in the hemolymph is derived from a neuroendocrine organ, the ACO is a likely source of the peptide.

As just described, CabTRP Ia is detectable in the hemolymph of C. productus, and the ACO is one potential source of this circulating peptide. Given a hormonal mode of delivery, multiple target tissues are likely to exist for this peptide in the crab. As we have shown here, one target of hemolymph-borne CabTRP Ia is the musculature of the foregut. In this species,none of the muscles of the gastric mill or pyloric regions of the foregut is directly innervated by TRP-immunopositive axons. Nevertheless, a hormonally relevant concentration (10^–7 mol l^–1) of CabTRP Ia increased EJPs in three of three gastric mill and three of three pyloric muscles tested. Moreover, concentrations as low as 10^–8 mol l^–1 were observed to consistently increase the contraction amplitude of at least one muscle, with the enhancement presumably due in part to the CabTRP Ia-evoked increase in EJP amplitude.

The most thoroughly studied peptide modulators of stomatogastric musculature are TNRNFLRFamide and SDRNFLRFamide in the crab C. borealis (Jorge-Rivera and Marder,1996; Jorge-Rivera et al.,1998). Our investigation of the effects of CabTRP Ia in C. productus, while not nearly as extensive as the earlier work, has produced some similar observations. Application of either FLRFamide peptide increased the nerve-evoked contraction in 15 out of 17 muscles tested, with a threshold concentration of 10^–10 mol l^–1(TNRNFLRFamide) or 10^–9–10^–8 mol l^–1 (SDRNFLRFamide)(Jorge-Rivera and Marder,1996). We measured the threshold concentration of CabTRP Ia for an increase on gm8 contraction to be between 10^–9 mol l^–1 and 10^–8 mol l^–1 and observed effects on EJP amplitude in all six muscles tested.

Although we have not explored the biophysical mechanisms responsible for the increase in EJP amplitude, we note that no shift in muscle membrane potential was observed that would clearly indicate a change in muscle input resistance. In the crab C. borealis, CabTRP Ia and four other peptides (including TNRNFLRFamide) have been observed to activate an inward current (Swensen and Marder,2000) in subsets of STG motor neurons(Swensen and Marder, 2001). Interestingly, the other four peptides, all presumably acting as hormones,also increase muscle contractions in C. borealis(Jorge-Rivera et al., 1998). Given the low-frequency motor discharge that drives many of the foregut muscles, we hypothesize that circulating CabTRP Ia may be crucial for maintaining appreciable muscle contractions in the gastric mill and pylorus,as has been proposed for other peptide hormones, including the extended FLRFamides (Jorge-Rivera and Marder,1996).

Our results show that many of the nerve terminals constituting the ACO envelop and/or directly abut hemolymph lacunae. This organization should allow released hormone direct access to the circulatory system and is widely recognized as the defining characteristic of a crustacean neuroendocrine release site (Cooke and Sullivan,1982; Beltz, 1988; Fingerman, 1992; Christie et al., 2004a). Intriguingly, there are some areas of the ACO that do not directly abut the circulatory system, but instead are in apposition to putative synaptic neuropil. Here, peptide released from the ACO may also function as a paracrine modulator of intrinsic CoG targets.

One possible role for paracrine signaling by the ACO is the coordination of multiple neuroendocrine systems (Fig. 9). The soma of the large (L)-cell, which provides both peptidergic and aminergic innervation to the PO(Cooke and Sullivan, 1982; Beltz, 1988; Fingerman, 1992), is located within the CoG and, in C. productus, extends processes to the general vicinity of the ACO (A.E.C., unpublished observations). Likewise, the somata innervating the ACP [i.e. anterior commissural neurons 1 and 2 (ACN1/2); Christie et al., 2004a; Christie and Messinger, 2005; Messinger et al., 2005] are contained within the CoG and arborize in the vicinity of the plexus. If the L-cell and/or ACN1/2 are modulated by CabTRP Ia (or some other unidentified co-transmitter within the ACO), then elements of at least two other neuroendocrine centers (i.e. the PO and ACP) could be simultaneously modulated/synchronized locally within the CoG, with concurrent release of hormones to the circulatory system.

Non-endocrine neurons within the CoG may also be targets of paracrine signals emanating from the ACO (Fig. 9). InCancer species, the area of the CoG containing this plexus is known to be the location of the intrinsic arbors of many projection neurons that innervate and modulate the STG(Coleman and Nusbaum, 1994;Norris et al., 1994, 1996; Coleman et al., 1995; Bartos and Nusbaum, 1997; Blitz et al., 1999; Nusbaum, 2002; Wood et al., 2004). If these neurons are influenced by CabTRP Ia (or another compound) released from the ACO in paracrine fashion, then a profound reorganization of the motor patterns produced by the stomatogastric circuit could occur, in addition to the effects that hemolymph-borne hormones have on the foregut musculature.

We thank all members of the Friday Harbor Laboratories' Spring 2004 Research Apprenticeship `From Neurons to Behavior: Comparative Neurobiology of Marine Invertebrates' for their support. Likewise, we thank Jackson ImmunoResearch Inc., Leica Microsystems, Molecular Probes and Vector Laboratories for their contributions of equipment and reagents to the above class. A. Countryman, T. Larson and T. Halatchev are thanked for their help with immunohistochemistry. Drs H. Dircksen, R. Elde, D. Nässel, K. Rao,B. Stay and J. Veenstra are thanked for their generous contributions of antibodies and Dr M. Nusbaum is thanked for his contribution of synthetic CabTRP Ia. Drs B. Swalla and D. Tauck are thanked for allowing us use of their epifluorescence microscope system and microelectrode puller, respectively. A.E.C., D.I.M. and C.T.N. acknowledge financial support from the University of Washington Department of Biology's Casey Fund for Undergraduate Research, the Mary Gates Endowment for Students, the Howard Hughes Medical Institute (HHMI),the Ellis B. Ridgeway Fellowship Fund and the University of Washington Office of the Provost. L.L. acknowledges support from the University of Wisconsin,Wisconsin Alumni Research Foundation, a research award from the American Society for Mass Spectrometry and a National Science Foundation CAREER Award(CHE-0449991). K.K.K. acknowledges support from NIH Chemistry–Biology Interface Training Grant (T32 GM008505). J.T.B. and D.R.V. acknowledge support from the Research Corporation, the Grass Foundation, Santa Clara University and an award to SCU under the Undergraduate Biological Sciences Education Program of the HHMI.