| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Center for Neurobiology and Behavior, Columbia University, New York State Psychiatric Institute, New York, New York; and 2Department of Chemistry, University of Illinois, Urbana, Illinois
Submitted 2 June 2006; accepted in final form 26 September 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Kandel 1976). However, traditional circuit analysis techniques are inefficient when dealing with the numerous potential synaptic connections that might exist between Aplysia's 20,000 central neurons (Croll 2003). Various studies using Lymnaea and Aplysia have suggested a complementary approach—using the coordinated expression pattern of a neuropeptide-encoding gene to help delineate a neural circuit (Bogerd et al. 1991; Pulst et al. 1988; Shope et al. 1991; van Minnen et al. 1988, 1989, 1992). We have followed this strategy to identify a circuit of Aplysia neurons that express the gene for the R15 peptides and to begin to examine the role of that circuit in controlling behavior.
Since its initial description by Arvanitaki and Tchou (1942), the endogenously bursting cell R15 has been the most widely studied neuron in Aplysia (e.g., Adams and Benson 1985). R15 expresses a family of three neuropeptides—R15
1, R15
and R15
—which are encoded by a single gene (Weiss et al. 1989). The R15 neuropeptide gene is also expressed in other Aplysia neurons as a splice-variant of mRNA (Buck et al. 1987). The two resulting expression patterns differ by a single peptide substitution, R152 for R151, which differ in their pharmacological effects (Alevizos et al. 1991a). In the entire CNS, R151-like immunoreactivity (IR) was found only in R15, whereas R152-like IR was observed in 
40 neurons, including four unidentified neurons in the abdominal ganglion (Alevizos et al. 1991a).
In this study, we used more direct methods to confirm that the R15-peptide-like IR neurons express authentic R15 peptides and to examine the details of their expression. We used physiological criteria to identify the four cells in the abdominal ganglion that stain for R152 peptide and found that two of them, the L9G gill motoneurons, also co-express pedal peptide (abbreviated Pep) (Lloyd and Connolly 1989), as well as the R15 peptides.
We then tested whether R15 and the other identified neurons that express the R15 peptide gene are part of a circuit that contributes to control of a specific behavior. We identified the circuit connections and physiological firing patterns of the neurons in the abdominal ganglion that express the R152 peptide. As part of this circuit analysis, we located and identified the cholinergic cell Interneuron XIII, originally inferred to exist from indirect evidence by Kandel et al. (1967). Finally, we used in situ intra- and extracellular in vivo recording techniques to record from some of the circuit elements during escape locomotion. The results demonstrate that R15 and the other R15 peptide-expressing neurons in the abdominal ganglion are part of a larger autonomic control circuit that is activated by a rhythmic central command from the pattern generator that drives fictive escape locomotion. More generally, this study illustrates how, in some cases, neuropeptide expression patterns may be used to help delineate neural circuits.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Specimens of Aplysia californica weighing 10–300 g were supplied by Marinus, Long Beach, CA. Animals were maintained in aquaria containing artificial seawater (ASW; Tropic Marine) at 14–16°C until used, except as noted in the following text.
Chronic in vivo nerve recording from identified neurons
To record in vivo from the axons of R15, L11, or L9G neurons, the peripheral nerve of interest was placed in an extracellular cuff electrode for en passant recording using surgical procedures as described by Alevizos et al. (1991b). Nerves used included genital (R15 and L11), siphon (L9G), and a small branch of pericardial (R15). The animal was allowed 24–48 h to recover before recording. In experiments designed to test the correlation of R15 activity with egg laying, we increased the probability of egg laying by using a modification of a protocol originally developed for Lymnaea (Ter Maat et al. 1982). The subject was placed overnight with four other animals in a small (12 l) bucket of ASW that was kept without aeration or filtering. Approximately 18 h later, the subject was placed in a 4.5-l aquarium of aerated and filtered ASW at 15–16°C for the duration of the recording sessions. Spontaneous egg laying and escape locomotion were monitored live, and the timing of their occurrence was marked with a manually operated event marker. This event marker signal and the signal from the electrode were recorded with an FM tape recorder.
At the end of the recording period, which consisted of 6–10 h daily sessions for from 1 to 4 days, the animal was killed and the abdominal ganglion removed along with the extracellular electrode connected to the nerve branch. In the reduced preparation, activity in the soma of the neuron of interest was recorded with an intracellular microelectrode and fed to an audio monitor. The signal from the cuff electrode was fed to a computer-based spike recognition system (Signal Processing Systems; SPS-8701 Waveform Discriminator; Malvern, Australia). This system allows one to create a template that matches the shape and size of any arbitrary unit in the extracellular nerve record. By using the audio output as a guide, one can adjust the template settings to recognize the axon spike of the neuron from which one is recording the soma spike. When the template is optimally adjusted, the SPS system generates an event mark each time there is a soma spike. Once a template has been created, the tape of the nerve cuff recording from the egg-laying sessions is played back into the SPS device. Event markers from the SPS system that signaled matches of the template to axon spikes were recorded on chart paper together with the original behavioral event marks for analysis of a possible correlation.
In situ electrophysiological recording and dye filling
In situ preparations for intra- and extracellular recording and dye filling were prepared as described previously (Alevizos et al. 1989). The chamber was filled with ASW, which was cooled by a Peltier plate to 15°C. To study neuronal firing patterns during fictive locomotion, the abdominal, pleural, pedal, and cerebral ganglia were dissected with their connections to each other intact. The cut ends of one or both P9 nerves were stimulated via suction electrodes to trigger fictive escape locomotion. As described by Xin et al. (2000), nerve recording from nerve P10 was used to monitor the rhythmic pattern of fictive locomotion while recording with intracellular electrodes from the somata of various identified neurons or from the axon of one of the pedal arterial shortener (PAS) neurons in a pedal arterial nerve (PAN) (described by Skelton and Koester 1992).
Immunocytochemistry
The ganglion containing the dye-labeled cell of interest was immunostained as described by Alevizos et al. (1991a). Primary polyclonal serum, referred to here as antiserum I/II, was raised in rabbits against the synthetic R152 peptide, which shares a large overlapping region of amino acid sequence with R151 peptide (Alevizos et al. 1991a). A polyclonal serum raised against R15 peptide was generated by Babco (Richmond, CA) via the method described previously (Alevizos et al. 1991a). For whole-mount preparations, primary antibodies were used in a 1:100 dilution (antiserum I/II) or 1:50 (R15 peptide antiserum). The secondary antibody used at concentrations of 1:50 was goat anti-rabbit conjugated to rhodamine (Cappel, West Chester, PA). In some experiments, sections of 16-µm thickness were made through the ganglion and alternate sections were exposed to either antiserum I/II or the antiserum made to R15 peptide, to test whether the labeled cell bound both antisera. Controls showed that primary antibody binding is eliminated by preabsorption with the antigen used to generate the antiserum and is not mimicked by using preimmune serum (Alevizos et al. 1991a; unpublished observations).
HPLC analysis of RBHE neurons for R15-peptide synthesis
Neuropeptides in RBHE were radiolabeled with 35S-methionine as described previously (Lloyd et al. 1985). For HPLC analysis, the somata of labeled RBHE neurons were dissected from nine abdominal ganglia and transferred to glass microtubes containing 100 µl of 0.1 M acetic acid. Ten nanomoles of synthetic R151 and R152 peptides were added as separation standards. The microtubes were heated to 95°C for 5 min to inactivate proteases. The extracts from RBHE cells then were combined, filtered and subjected to a three-stage separation by reverse phase high-performance liquid chromatography (RP-HPLC) as described previously (Weiss et al. 1989). An additional peptide standard, synthetic Aplysia SCP (Mahon et al. 1985), was added to the sample prior to the second and third separations.
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS)
Ganglia containing dye-labeled R15, RBHE or L9G neurons were maintained overnight at 4°C in ASW. The following day, after 20–45 min treatment in 0.1% protease type IX (Sigma-Aldrich, St. Louis, MO), labeled cells were manually isolated using electrolytically sharpened tungsten needles and transferred onto a gold sample plate in a drop of ASW using a plastic micropipette. The ASW was removed by aspiration, and 0.5 µl of a matrix solution (2,5-dihydroxybenzoic acid, DHB, 50 mg/ml in 80% acetone) was applied to the cells.
Cell samples were analyzed by a Voyager DE STR mass spectrometer equipped with a pulsed nitrogen laser (337 nm, 4 ns) and delayed ion extraction (PE Biosystems, Framingham, MA). Several hundred positive ion mass spectra were generated in linear mode. Each representative mass spectrum is a smoothed average of 85–125 laser shots. External calibration with peptide standards, bovine insulin (m/z 5734.6), and angiotensin I (m/z 1297.5) was used for initial cell sample screening. Peptide standards were then applied to spots with previously screened cell samples, spectra were re-acquired and peak assignments made on the basis of correspondence of observed masses to the calculated masses of the predicted peptides. Alternatively, known peptides—pedal peptide (m/z 1539.7) and acidic peptide (m/z 2961.3), observed in some of the L9G neuron spectra—were used as internal calibrants. For peak assignment, multiple mass spectra were acquired from each sample and peak m/z values were averaged. The average mass assignment error for different neuron types was within 100 ppm. We used PAWS software (Genomic Solutions, Ann Arbor, MI) to aid in the interpretation of mass spectra, and SignalP (Bendtsen et al. 2004) to determine the signal sequence of the precursor polypeptide.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
In a previous immunocytochemical study, we found that five neurons in the abdominal ganglion stain strongly positive with antiserum I/II, which binds to both R151 and R152 peptides. Only one of these cells—R15 itself—also stained with antiserum I, which stains only R151peptide. Therefore the other four neurons were assumed to contain R152 peptide (Alevizos et al. 1991a). In the current study, we identified those four putative R152-containing neurons by systematically impaling neurons in the appropriate size range in the areas where R152 staining was observed. We noted for each impaled neuron its size, position, color, spontaneous firing pattern, synaptic input and output, and any potential motor effects. Each cell was stained with Lucifer yellow, and the positions of the stained cells were noted on a map of the ganglion. We then immunostained the tissue with antiserum I/II, using a rhodamine-labeled secondary antibody. With this double-labeling approach, we were able to determine the identity of all four of the R152-positive cells.
Three of the R152-positive cells were found to be previously identified motoneurons. One is RBHE, a serotonergic heart exciter motoneuron, which increases the frequency, and in some cases the strength, of the myogenic heartbeat (Mayeri et al. 1974). Two of the cells are the twin L9G gill motoneurons, the transmitter of which is unknown (Kupfermann and Kandel 1969).
The fourth immunopositive cell was found to be a previously unidentified neuron, which we have named L40. It is located at the deepest layer of the cortex of the left dorsal rostral quadrant of the abdominal ganglion—just superficial to the neuropil (Fig. 1, A and B). The cell body of L40 is 80 µm in diameter. L40 is a spontaneously active cell (Fig. 5A) that receives primarily background fast excitatory postsynaptic potential (EPSP) inputs. When fired by current injection it increases the rate of firing of a large, spontaneously active EPSP input to R15 (Fig. 1C). This input was found to come from Int XIII (see following text). L40 was shown, by filling with Lucifer yellow, to have a single axon leaving the ganglion via the right pleural-abdominal connective (Fig. 1B).
|
|
2-positive neurons described in the preceding text could have been the result of nonspecific antibody binding. To rule out this possibility for RBHE, we used two separate neurochemical tests to determine whether it synthesizes R152 peptide. In the first set of experiments, we sequentially fractionated the pooled cell extracts of nine radiolabeled RBHE neurons in the presence of R151 and R152 peptide standards. In the first stage of separation, with triethylamine as the counter ion, the major radioactive peak precisely co-migrated with synthetic R152 peptide. This fraction was collected and re-chromatographed in the presence of a different counter-ion, heptafluorobutyric acid (HFBA) and an SCPA peptide standard, to resolve other peaks that may have been superimposed on the peak obtained with TEA. Again, the most intense peak of radioactivity co-eluting with synthetic R152 peptide at the second stage was collected and subjected to a third stage of chromatography in the presence of trifluoroacetic acid (TFA) as the counter-ion. No significant radioactive peaks co-eluted with R151 and SCP. Thus the peak of radioactivity had the same migration time as the synthetic R152 peptide in three different steps of purification (Fig. 2). These results strongly support the hypothesis that R152 peptide is synthesized by neuron RBHE. To provide more definitive evidence for this notion, as well as to examine whether RBHE also synthesizes the R15 and peptides, we directly analyzed the contents of the cell using a second neurochemical approach—MALDI-TOF MS.
|
; Weiss et al. 1989), we expected to find masses matching R151, R15, R15, and pGluR15. They were indeed all observed (Table 1 and Fig. 3 A). We also detected a peak at m/z 1834, which by mass corresponds to R15f, a C-terminal fragment of R15 peptide that has also been purified from R15 cell extracts (Weiss et al. 1989). The observed mass of the R151 peptide is two daltons smaller than its theoretical mass, which is indicative of possible disulfide bond formation between two cysteine residues. This finding directly confirms a similar conclusion based on biochemical and pharmacological properties of R151 peptide (Weiss et al. 1989). The peak corresponding to the pGlu form of R15 was always observed at higher intensity than the peak corresponding to the unmodified peptide.
|
|
, R15f, and R15, with both Gln and pGlu forms, were detected. The 38-amino-acid-long peptide R151 was absent from the RBHE spectra. Instead, the 24-amino-acid-long alternate, R152, was observed (Table 1 and Fig. 3B). The sequence of the R152 peptide suggests a disulfide bond between two cysteine residues located in the stretch of sequence unaffected by the nucleotide substitution in the mRNA. The observed mass of R152 differs by two daltons from its theoretical mass, thereby suggesting that cysteines are also bound in this peptide.
In the L9G neurons, we observed masses characteristic of R152, R15, R15f, and pGluR15 peptides. In 20% of our mass spectra, the pGluR15 peak was accompanied by a weak and poorly resolved peak that may indicate the presence of the Gln form of R15 peptide. However, unlike in the other neurons examined, we did not detect the mass corresponding to the unmodified R15 peptide (Fig. 3C).
Samples from neurons RBHE and L9G contain other neuropeptides in addition to the R15 peptides. In samples from both cell types, we observed strong signals at m/z 5014 and 5022, representing the recently characterized Aplysia beta thymosin proteins (Romanova et al. 2006) as well as a number of unknown peptides (Table 2). Some of these unassigned peaks, namely m/z 6053 and m/z 7616, are often seen together with the beta thymosin proteins in Aplysia neuronal and connective tissue samples (Vanmali et al. 2003). Masses matching pedal peptide (Pep) (Lloyd and Connolly 1989) were detected in all L9G cell samples. Egg-laying hormone (ELH) and related peptides (Newcomb and Scheller 1987) occasionally were seen in the L9G mass spectra (data not shown). Given the sporadic nature of their appearance, we suspect they result from processes from the bag cell neurons that infiltrate the ganglion (Chiu and Strumwasser 1981).
|
). In R15, RBHE and L9G neurons, R15 peptide is further cleaved to yield a 17-amino-acid-long C-terminal fragment, R15f.
The small size and location of L40 makes it a challenge to use for single-cell MALDI-TOF analysis. We therefore used a simpler, less direct method to test whether the R152-staining either represents real R152 peptide, or alternatively is due to cross-reaction with another peptide. We reasoned that if L40 shows immunoreactivity for both R152 and R15 peptides, the probability of such a pattern of double reactivity being due to nonspecific binding would be low, especially because the two peptides have completely unrelated amino acid sequences. After labeling the cell with Lucifer yellow, we subjected alternate sections of the ganglion to immunocytochemical staining using antiserum I/II (for R152) and R15 antisera. Application of either antiserum resulted in staining of L40, strongly supporting the conclusion that the cell expresses at least two of the R15 peptides (Fig. 1D). The antiserum to R15 peptide also labeled neurons R15, RBHE and L9G, as expected (data not shown).
Synaptic inputs and outputs of L40 and identification of Int XIII
When L40 is fired by current injection it elicits a barrage of giant EPSPs in R15 (Fig. 1C). The EPSPs occur only if the right pleural-abdominal connective remains connected to the right pleural-pedal ganglion complex. Previous investigators had attributed a similar EPSP that could be triggered by stimulation of the right connective to an unidentified cell named Int XIII [sometimes called Input I (Parnas et al. 1974) or RC1 (Schlapfer et al. 1976)]. Although the location of Int XIII's cell body was unknown, the EPSP it elicits in R15 could be monitored either by observing spontaneous activity of Int XIII or by stimulating the right pleural-abdominal connective. In this way, the Int XIII-to-R15 EPSP has been used for numerous studies of synaptic plasticity and connectivity patterns in Aplysia since its first description by Kandel et al. (1967) (see Adams and Benson 1985 for review).
To test directly whether the EPSPs that L40 elicits in R15 are from Int XIII, we searched for the soma of Int XIII. We successfully identified the cell body of Int XIII based on the criterion that it elicits a large EPSP in R15 when fired by current injection (Fig. 4 B).
|
. Its cell body, which is located on the lateral edge of the ganglion, midway between dorsal and ventral surfaces, is 70–80 µm in diameter. Int XIII is a spontaneously active neuron, receiving primarily sporadic background fast EPSP inputs. It was shown, by filling with Lucifer yellow, to have a single axon that exits the ganglion via the right pleural-abdominal connective (Fig. 4A). According to standard Aplysia nomenclature, once the cell body of an unidentified interneuron is found, it is given a name identifying it with the ganglion in which it is located. We gave Int XIII the identified cell name "RP5," which we will use in the following text.
We conclude that the cell we identified as RP5 is identical to Int XIII (also known as Input I or RC1) that we and others have studied over the years, for the following reasons. 1) The spontaneous giant EPSPs in R15 attributed to Int XIII (Woodson and Schlapfer 1979) are completely eliminated by hyperpolarizing RP5 by current injection (Fig. 5C). 2) Firing RP5 by injecting brief current pulses into its soma produces EPSPs in R15 with amplitudes and kinetics of facilitation and posttetanic potentiation (data not shown) similar to those observed by others who stimulated the axon of Int XIII in the cut connective (Schlapfer et al. 1976). 3) The RP5-R15 EPSP is reversibly blocked by 500 µM hexamethonium (data not shown) as previously has been found for the Int XIII-R15 EPSP elicited by connective stimulation (Segal and Koester 1982). 4) RP5 spikes generate 1:1 synaptic potentials in the RB, L9G, and L11 abdominal ganglion neurons, similar in sign and amplitude to those previously attributed to Int XIII by Segal and Koester (1982) (data not shown). 5) The RP5-R15 connection appears to be monosynaptic, as it follows RP5 soma spikes1:1 with constant latency (Fig. 4B).
We also investigated additional synaptic connections to and from L40, beginning with RP5. By hyperpolarizing RP5, we were able to demonstrate that it mediates all of the synaptic input that is generated in R15 by firing L40 (Fig. 4C). However, the EPSPs in RP5 do not follow L40 spikes 1:1 with constant latency, suggesting that there may be one or more interposed interneurons between them. The spontaneous firing of L40 provides a major component of the drive for RP5's baseline firing rate in situ, which in turn modulates spontaneous R15 bursting activity (Fig. 5).
We next examined the two L9G cells and RBHE as potential targets of L40, based on the fact that they all express the R15-peptide gene. Both cell types were driven strongly by firing L40 (Fig. 6 A, 1 and 2). Given that L40 excites RP5 and that the L9G and RBHE neurons are excited relatively weakly by RP5, their strong excitation by L40 is surprising. High-speed recording showed that both the L9G cells and RBHE receive EPSPs not only from RP5 but also from at least one other unidentified larger input. In the case of RBHE, the second L40-elicited EPSP resembles in size and shape an input described previously as coming from an unidentified cell named Interneuron XII (Int XII) (Koester et al. 1974).
|
). Because spontaneous bursting of L40 causes some of the spontaneous increases in RP5 firing rate, we examined whether L40 also excites the PAS neurons. Firing L40 indeed excited the PAS motoneurons (Fig. 6A3). It is not known whether this is a direct or indirect effect.
The only inhibitory effect noted for L40 was its weak inhibitory action on identified neuron L11 (data not shown). This effect is mediated at least in part by RP5, which previously had been shown to produce small inhibitory postsynaptic potentials (IPSPs) in L11 (Segal and Koester 1982).
Two synaptic inputs to L40 were identified: a small EPSP from RP5 via a positive feedback loop (Fig. 6B1) and a small EPSP from cell L10 (Fig. 6B2). L10 is a hybrid interneuron-motoneuron that helps to coordinate cardiovascular function with renal pore opening (Koester and Alevizos 1989).
R15 does not burst during egg-laying behavior
To test whether the cells that express R15 peptides are part of a functional circuit that contributes to behavior, we recorded from R15 in vivo during spontaneous egg-laying. This behavior is triggered by the release of ELH and other peptides by the neuroendocrine bag cells when they fire in a population burst, which lasts 20 min (Dudek et al. 1979). Mayeri et al. (1979) had shown that when the bag cells burst in situ they enhance the spontaneous bursting tendency of R15. In vivo, the start of a bag cell burst precedes the first appearance of the egg string by 35 ± 5 min at 16–20°C (Begnoche et al. 1996; Ferguson et al. 1989). We therefore hypothesized that strong bursting of R15 might begin during a period starting 30–45 min before egg laying, perhaps extending into the initial period of the behavior. Chronic recording from the axon of R15 in the intact, freely behaving animal had earlier revealed no sign of R15 bursting with the animal in a resting state (Alevizos et al. 1991b). We found a similar lack of activity in the resting animal in this study, as expected. However, contrary to expectations from in situ studies, R15 did not burst preceding spontaneous egg laying or during the egg laying itself (n = 5; data not shown). We cannot rule out the possibility that this negative result could be an artifact, caused for example, by unknown effects of the stress of the electrode-implantation procedure on R15 function.
Elements of the R15 peptide-circuit are driven by a central command during locomotion
Results from an earlier study led us to test whether the R15 peptide-expressing neurons might be active during escape locomotion. This study showed that the PAS motoneurons, which are excited by L40, are active during escape locomotion, causing aortal shortening in phase with contraction of the rostral body wall (Skelton and Koester 1992).
We found that eliciting fictive locomotion in the isolated nervous system does indeed generate rhythmic activity in a network that includes all of the known R15-peptide-expressing neurons as well as several neurons to which they are known to be connected. We used the bursting pattern of nerve P10, which is known to fire during the contraction phase of locomotion in the intact animal (Xin et al. 2000), to determine the relative phases of fictive locomotion in situ. During fictive locomotion, L40 and most of the neurons that it excites directly or indirectly—the L9Gs, R15, RBHE, and RP5—all fire in phase with one another during the extension phase of locomotion (Figs. 7 and 8). The one exception is the pair of PAS neurons, which fire during the contraction phase of both fictive and real locomotion (Skelton and Koester 1992; Xin et al. 1996; unpublished observations).
|
|
). Cell L11 sends an axon to the gill but its function is unknown. Because it has no obvious excitatory effect, L11 is assumed to be either a modulatory or an inhibitory neuron. The LDHI cells are heart inhibitory modulator cells. We found that during fictive locomotion L11 and the LDHI neurons fire in phase with the PAS neurons during the contraction phase, and out of phase with the L9Gs, R15, RBHE, L40, Int XII and RP5 (Figs. 7 and 8). We used chronic nerve recording to test whether this same pattern of circuit activity occurs during locomotion in vivo. It was not feasible to recognize the spikes of RBHE or the LDHI cells because they are not reliably distinguishable from other spikes in the pericardial nerve. Although the R15 spike was clearly distinguishable in nerve recordings, tests for R15 bursting during locomotion were negative (n = 7). We cannot say whether R15's silence during locomotion is normal or whether it is a side effect of the implantation operation. Alternatively, perhaps R15 firing during locomotion in the intact animal is gated on by specific environmental or physiological conditions that are not replicated by our experimental conditions. In contrast, we did observe L9G and L11 bursting rhythmically during locomotion, out of phase with each other and with the same phasing relative to locomotion as that recorded during fictive locomotion (Fig. 9).
|
previously had shown that respiratory pumping, a coordinated stereotyped contraction of gill and siphon, is recruited during locomotion. We would therefore expect that most of the identified motoneurons that fire during respiratory pumping also would be recruited in phase with locomotion. However, R15, the L9G motoneurons and RBHE are not excited during spontaneous respiratory pumping in situ (Adams and Benson 1985; Kupfermann and Kandel 1969; Koester et al. 1974), so it is particularly significant that they are recruited in phase with fictive locomotion.
|
). These mismatches presumably reflect the fact that only a small fraction of all the interneurons that are active during locomotion have been identified, and during actual locomotion unidentified interneurons may override some of the synaptic actions we recorded during fictive locomotion. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Alternative splicing of R15-peptide mRNA and processing of peptide precursor
We confirmed that expression of two alternative R15 mRNA transcripts in the CNS results in the differential distribution of distinct but overlapping sets of R15 peptides to different identified neurons in the abdominal ganglion. With good mass accuracy and high-quality spectra, we detected the presence of all known R15 peptides in the cells we studied that express the R15 peptide gene. We observed directly a selective expression of R152 peptide in RBHE and the L9G neurons and R151 peptide in neuron R15, indicating that RBHE and L9G neurons make a different mRNA splice choice than that of neuron R15. Our findings are in agreement with and extend earlier in situ hybridization and immunocytochemistry reports based on R15 and on unidentified neurons (Alevizos et al. 1991a; Buck et al. 1987). We thus provide the most direct evidence to date that mutually exclusive expression of alternative splice forms of R15 mRNA is a feature of single identified neurons in the abdominal ganglion. Our immunocytochemical data suggest that neuron L40 uses the same splice choice as do the RBHE and L9G cells.
We confirmed the earlier proposed structures for R151 and R152 peptides by direct measurement of peptide molecular weights. As expected, the disulfide bonds between two cysteine residues located in the amino acid segments were found not to be affected by alternative splicing (Weiss et al. 1989).
We determined that in addition to R15 gene peptides, the RBHE and the L9G cells express other known and unidentified neuropeptides. For example, pedal peptide (Pep) is observed in mass spectra from all L9G cells. This is consistent with the results of Pearson and Lloyd (1990), who found Pep-like immunoreactivity in two large unidentified neurons in the abdominal ganglion in the area where the L9G somata are found. Moreover, the L9G neurons send their axons to the gill via the siphon nerve (Lukowiak 1979), and Pep-like IR was found in these two structures as well (Pearson and Lloyd 1989, 1990). Our MALDI-TOF MS data demonstrate that in addition to Pep, the 1, 2, , and pGlu forms of R15 peptides are also present in the siphon nerve (data not shown).
What do the R15-peptides and pedal peptide do?
We presume that R15 peptides and Pep are synaptically released. Hall and Lloyd have demonstrated synaptic release of Pep at junctions within pedal muscle. Pharmacological application of Pep modulates the contractility of that muscle, and it is hypothesized that neuronal release of Pep has a similar effect during locomotion (Hall and Lloyd 1990). Perfusion of the gill with R152 peptide causes contractions (Alevizos et al. 1991a), so it seems plausible that Pep and the R152 peptide in the L9G neurons also may be released and have excitatory or modulatory effects on gill muscle. R152 peptide-immunoreactivity has been found in the heart and the R152 peptide has been shown to have excitatory effects on heart muscle that resemble the actions of serotonin (Skelton and Koester 1992). Conceivably, serotonin and the R152 peptide may be released together as co-transmitters by RBHE. Alternatively, Pep and/or R152 peptide and the other R15 peptides may have trophic effects on synaptic connections, similar to that recently shown for the neuropeptide sensorin at Aplysia sensory neuron-to-motoneuron synapses (Hu et al. 2004). More work is required to determine if these peptides are released, and if so, what physiological effects they have.
Central command coordinates autonomic and locomotor function in Aplysia
In both vertebrates and invertebrates, autonomic adjustments to exercise are in some cases driven by a "central command"—a neural signal generated in parallel with the signal from the higher center that controls somatic musculature. Reflex action may then fine-tune the autonomic response (Arshavsky et al. 1990; Eldridge et al. 1985; Mateika and Duffin 1995). Our results, combined with earlier results in the literature, demonstrate that such a central command coordinates autonomic activity with escape locomotion in Aplysia and that the R15 peptide-containing cells we have identified are part of a circuit that relays the central command to visceral organs (Fig. 10).
The autonomic adjustments mediated by this central command circuit are only partially understood. 1) Cyclical firing of the LPAS neurons in phase with locomotion helps to adjust the length of the anterior aorta to the length of the rostral half of the body as it alternately shortens and lengthens (Fig. 10) (Skelton and Koester 1992; Xin et al. 1996). 2) Activation of the LDHI neurons and RBHE in a push-pull manner during locomotion may facilitate the redistribution of blood required to support the changes in body wall shape that occur during stepping (Fig. 10). Hening had previously shown that in vivo respiratory pumping occurs either with each cycle of stepping or, in some cases, on alternate cycles, and that it occurs just before initiation of the extension phase of stepping. He postulated that the gill contraction that is a component of respiratory pumping forces a bolus of hemolymph through the inhibited (relaxed) heart into the arterial system, in support of the initial phase of anterior foot extension (Hening 1982). Because the LDHIs are known to fire during respiratory pumping in situ (Koester et al. 1974), the activity we see in them just before the start of the extension phase of fictive locomotion is presumed to be driven by the pattern generator for respiratory pumping. In contrast to the LDHIs, the heart exciter, RBHE, fires during the extension phase of stepping. This may result in an increase in blood flow to the hydroskeleton, thereby helping to extend the body wall throughout the extension phase of stepping. 3) The L9G cells are excited in phase with foot extension. Their firing presumably causes a modest increase in vascular resistance of the afferent and efferent veins to the gill (Swann et al. 1982) although the functional consequences of such a resistance increase are not clear. 4) L11 fires in phase with foot contraction with the same phase relationship as the LDHIs. The function of L11 is unknown. 5) As mentioned in the preceding text, despite the modulation of R15 in phase with fictive locomotion, such modulation was not observed in vivo. It remains to be determined whether during actual locomotion there may be other experimental conditions in which it fires, and whether L40, Int XII, Int XIII, RBHE, and the LDHI cells fire in vivo as they do during fictive locomotion.
Use of gene-expression patterns to identify functionally related neurons
Our results, together with two earlier studies from Aplysia, suggest that using neuropeptide expression to search for functionally related neurons may be useful in two ways. 1) Identification of neurons with similar functional roles in behavior. A recent study of neuropeptide expression patterns in different neurons throughout the nervous system showed that neuronal expression of the neuropeptide sensorin is strongly correlated with high-threshold mechanoreceptor function (Walters et al. 2004). 2) Identification of neural circuits. Based on the shared expression patterns of the genes that encode ELH, Brown et al. (1989) identified neurons in the right pleural ganglion that excite the neuroendocrine bag cells in the abdominal ganglion. Likewise, we have found such an approach to be effective on a somewhat broader scale in the case of the R15 peptides. Because L40 has a small soma and is located deep in the ganglion, it would have been particularly difficult to identify without the clue provided by the immunohistochemical data.
It will be interesting to extend our examination of neuropeptide-expression patterns to other neurons in Aplysia. For example, Pep expression previously had been observed in various unidentified neurons in Aplysia as well as in a cluster of neurons in each pedal ganglion. These Pep-containing pedal neurons are excited in phase with the locomotor rhythm and appear to be involved in modulation of pedal contractions (Hall and Lloyd 1990). It is striking that the L9G neurons, which are also active in phase with locomotion, likewise express Pep. It would be of interest to identify other neurons that express Pep and/or R152 peptide and to determine whether they also are driven by the central command that accompanies locomotion.
Various investigators studying neuropeptide and neurohormone distributions and actions in Lymnaea have hypothesized that expression patterns of caudal dorsal cell hormone, molluscan insulin-like peptide, APGWamide, and VD1/RPD2 peptide may be localized primarily to circuits of neurons that control oviposition, blood glucose levels, male reproductive behavior, and ventilatory function, respectively (Bogerd et al. 1991; Croll and Van Minnen 1992; van Minnen et al. 1989). Likewise, Brown et al. (1989) and Painter et al. (1989) have suggested that unidentified ELH gene-expressing neurons in the cerebral ganglion of Aplysia may be connected to the members of the network of cells in the pleural ganglion and the bag cell clusters that express the same gene. In light of our results, it will be interesting to test these notions experimentally.
The use of neuropeptide distributions to map circuits may be considered as a special case of a more general approach based on expression patterns of various types of molecules. For example, Marinesco et al. (2004) have found that 80% of the serotonin-IR neurons in Aplysia are activated during defensive withdrawal. In vertebrates, there is an emerging hypothesis that some neural circuits may be defined by patterns of transcription factor expression (Brunet and Pattyn 2002; Dasen et al. 2005; Qian et al. 2001).
The argument that neuropeptide-expression patterns can be used to help define functional circuits is not based on the naïve premise that there is a one peptide-one behavior (or 1 gene-1 peptide) principle. This obviously is not correct as there are numerous counter-examples of individual neuropeptides that are found in different neural circuits with quite different behavioral functions. Moreover, it has been well documented that under different conditions individual premotor neurons can take part in more than one circuit controlling more than one behavior (reviewed by Marder and Calabrese (1996). With these caveats in mind, the results of this and other studies suggest that when attempting to define new neural circuits, a strategy of looking for correlated neuropeptide expression can be valuable as an initial screening approach.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Present addresses: K. R. Weiss, Dept. of Physiology and Biophysics, Mount Sinai School of Medicine, New York 10029–6574; N. McKay, 300 Carew St., Suite 1, Springfield, MA 01104.
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: J. Koester, Center for Neurobiology and Behavior, Columbia University, New York State Psychiatric Institute, 1051 Riverside Dr., New York, NY
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Alevizos A, Karagogeos D, Weiss KR, Buck L, and Koester J. R15 alpha-1 and R15 alpha-2 peptides from Aplysia: comparison of bioactivity, distribution, and function of two peptides generated by alternative splicing. J Neurobiol 22: 405–417, 1991a.[CrossRef][ISI][Medline]
Alevizos A, Weiss KR, and Koester J. SCP-containing R20 neurons modulate respiratory pumping in Aplysia. J Neurosci 9: 3058–3071, 1989.
Alevizos A, Weiss KR, and Koester J. Synaptic actions of identified peptidergic neuron R15 in Aplysia. I. Activation of respiratory pumping. J Neurosci 11: 1263–1274, 1991b.[Abstract]
Arshavsky YI, Deliagina TG, Gelfand IM, Orlovsky GN, Panchin YV, Pavlova GA, and Popova LB. Neural control of heart beat in the pteropod mollusk Clione limacina: coordination of circulatory and locomotor systems. J Exp Biol 148: 461–475, 1990.
Arvanitaki A and Tchou S. Les lois de la croissance relative individuelle des cellules neveuses chez l'Aplysie. Bull Histol Appl Tech Microscop 244–256, 1942.
Begnoche VL, Moore SK, Blum N, Van Gils C, and Mayeri E. Sign stimulus activates a peptidergic neural system controlling reproductive behavior in Aplysia. J Neurophysiol 75: 2161–2166, 1996.
Bendtsen JD, Nielsen H, von Heijne G, and Brunak S. Improved prediction of signal peptides: SignalP 3.0. J Mol Biol 340: 783–795, 2004.[CrossRef][ISI][Medline]
Bogerd J, Geraerts WP, Van Heerikhuizen H, Kerkhoven RM, and Joosse J. Characterization and evolutionary aspects of a transcript encoding a neuropeptide precursor of Lymnaea neurons, VD1 and RPD2. Brain Res Mol Brain Res 11: 47–54, 1991.[Medline]
Brown RO, Pulst SM, and Mayeri E. Neuroendocrine bag cells of Aplysia are activated by bag cell peptide-containing neurons in the pleural ganglion. J Neurophysiol 61: 1142–1152, 1989.
Brunet JF and Pattyn A. Phox2 genes - from patterning to connectivity. Curr Opin Genet Dev 12: 435–440, 2002.[CrossRef][ISI][Medline]
Buck LB, Bigelow JM, and Axel R. Alternative splicing in individual Aplysia neurons generates neuropeptide diversity. Cell 51: 127–133, 1987.[CrossRef][ISI][Medline]
Chase R. Behavior and Its Neural Control in Gastropod Molluscs. NY: Oxford, 2002.
Chiu AY and Strumwasser F. An immunohistochemical study of the neuropeptidergic bag cells of Aplysia. J Neurosci 1: 812–826, 1981.
Croll RP. Complexities of a simple system: new lessons, old challenges and peripheral questions for the gill withdrawal reflex of Aplysia. Brain Res Brain Res Rev 43: 266–274, 2003.
Croll RP and Van Minnen J. Distribution of the peptide Ala-Pro-Gly-Trp-NH2 (APGWamide) in the nervous system and periphery of the snail Lymnaea stagnalis as revealed by immunocytochemistry and in situ hybridization. J Comp Neurol 324: 567–574, 1992.[CrossRef][ISI][Medline]
Dasen JS, Tice BC, Brenner-Morton S, and Jessell TM. A Hox regulatory network establishes motor neuron pool identity and target-muscle connectivity. Cell 123: 477–491, 2005.[CrossRef][ISI][Medline]
Dudek FE, Cobbs JS, and Pinsker HM. Bag cell electrical activity underlying spontaneous egg laying in freely behaving Aplysia brasiliana. J Neurophysiol 42: 804–817, 1979.
Eldridge FL, Millhorn DE, Kiley JP, and Waldrop TG. Stimulation by central command of locomotion, respiration and circulation during exercise. Respir Physiol 59: 313–337, 1985.[CrossRef][ISI][Medline]
Ferguson GP, Ter Maat A, Parsons DW, and Pinsker HM. Egg laying in Aplysia. I. Behavioral patterns and muscle activity of freely behaving animals after selectively elicited bag cell discharges. J Comp Physiol [A] 164: 835–847, 1989.[CrossRef][Medline]
Hall JD and Lloyd PE. Involvement of pedal peptide in locomotion in Aplysia: modulation of foot muscle contractions. J Neurobiol 21: 858–868, 1990.[CrossRef][ISI][Medline]
Hening WA. Central Generation and Coordination of a Complex Behavior R
