In vitro conditioning induces morphological ... - Semantic Scholar

Neuroscience Research 43 (2002) 363 /372 www.elsevier.com/locate/neures

In vitro conditioning induces morphological changes in Hermissenda type B photoreceptor Ryo Kawai a, Tetsuro Horikoshi b, Takashi Yasuoka a, Manabu Sakakibara b,* a

b

Graduate School of Science, Tokai University, Kita-Kaname, Hiratsuka 259-1292, Kanagawa, Japan Laboratory of Neurobiological Engineering, Department of Biological Science and Technology, School of High-Technology for Human Welfare, Tokai University, Numazu 410-0321, Shizuoka, Japan Received 22 March 2002; accepted 23 April 2002

Abstract Short- and long-term synaptic plasticity are considered to be cellular substrates of learning and memory. The mechanisms underlying synaptic plasticity especially with respect to morphology, however, are not known. In vitro conditioning in molluscan preparations is a well established form of short-term synaptic plasticity. Five paired presentations of light and vestibular stimulation to the isolated nervous system of Hermissenda results in an increase in excitability of the identified neuron, the type B photoreceptor, indicated by 2 measures, an increase in the input resistance and a cumulative depolarization after the cessation of light stimulus recorded from the cell soma. The terminal branches of type B photoreceptors iontophoretically injected with fluorescent dye were analyzed using computer-aided 3-dimensional reconstruction of images obtained using a confocal microscope under ‘blind’ conditions. The terminal branches contracted along the centro-lateral axis within an hour after conditioning, paralleling the increase in neuronal excitability. These data suggest that in vitro conditioning in Hermissenda is a form of short-term synaptic plasticity that involves changes in macromolecular synthesis. # 2002 Published by Elsevier Science Ireland Ltd. and the Japan Neuroscience Society. Keywords: In vitro conditioning; Input resistance; Long-lasting depolarization; Terminal branch arborization; B photoreceptor; Hermissenda ; Morphology

1. Introduction Changes in synaptic efficacy is one of the cellular mechanisms underlying experience-dependent changes in behavior. Long-term synaptic plasticity depends on activation of second messenger cascades and the induction of changes in gene expression (Bartsch et al., 1995; Yin et al., 1995). While short-term synaptic plasticity does not appear to require changes in gene expression, the same second messenger cascades appear to be involved (Kandel and Schwartz, 1982). The mechanism underlying short-term synaptic plasticity, however, is not well understood. Recent data indicates that the active post-synaptic element undergoes a relatively rapid change in both

* Corresponding author. Tel.: /81-55-968-1211x4504; fax: /81-55968-1156 E-mail address: [email protected] (M. Sakakibara).

electrical characteristics and morphology (Shi et al., 1999). New synaptic spines on pyramidal cell dendrites are built within an hour after LTP induction in hippocampal slices (Engert and Bonhoeffer, 1999; Maletic-Savatic et al., 1999). Furthermore, dendritic spines at excitatory synapses undergo rapid, actindependent shape changes in primary hippocampal neuron cultures that may contribute to plasticity (Fischer et al., 2000). Similar morphological modification has been reported in molluscan neuronal circuits. The number of pre-synaptic varicosities and the extent of axonal arborization of the pre-synaptic neuron increases after long-term facilitation between synapses of the sensory neurons and the motor neurons in Aplysia (Bailey and Kandel, 1993). Pre-synaptic varicosities form after application of serotonin and this can be prevented with actin polymerization blockers in Aplysia (Hatada et al., 2000). The volume of axonal branching of the type

0168-0102/02/$ - see front matter # 2002 Published by Elsevier Science Ireland Ltd. and the Japan Neuroscience Society. PII: S 0 1 6 8 - 0 1 0 2 ( 0 2 ) 0 0 0 6 1 - 5

364

R. Kawai et al. / Neuroscience Research 43 (2002) 363 /372

B photoreceptor decreases by 50% 3 days after Pavlovian conditioning training in Hermissenda (Alkon et al., 1990). In this previous study, only the axon terminal appears to shrink. They interpreted this as the elimination of a useless synaptic contact in long-lasting memory formation; thus, they termed this modification ‘focusing’. In molluscan preparations, such as Aplysia (Nargeot et al., 1997), Lymnaea (Kemenes et al., 1997), and Hermissenda (Farley and Alkon, 1987), cellular phenomena induced by in vitro conditioning correlate well to those induced by in vivo conditioning. Among these, morphological and genetic short-term plasticity can be studied easily in Hermissenda , because it is possible to produce in vitro conditioning with natural stimulus (i.e. paired light and vestibular stimulation) and because the sensory system is intact even in the isolated preparation (Matzel et al., 1996). The cellular correlates of acquisition of classical conditioning in Hermissenda are well understood at the type B photoreceptor, which is the post-synaptic element for the vestibular hair cells and presumed to control whole-animal movement during positive phototaxis (Alkon, 1987). In the present study, we examined in vitro-conditioning-induced morphological changes in the terminal branch of type B photoreceptors, focusing on rapid changes that may underlie short-term plasticity.

2. Material and methods 2.1. Animals Hermissenda crassicornis obtained from Sea Life Supply (Sand City, CA) were maintained in 60-l artificial sea water (ASW) aquaria (Aqua, Tokyo) under subdued orange light (20 mW/cm2) at 13 8C on a 12-h light:12-h dark cycle (on at 08:00) and fed goldfish pellets (Hikari). 2.2. Electrophysiology The circumesophageal ganglion was removed under dim light illumination in ASW (430 mM NaCl, 10 mM KCl, 50 mM MgCl2, 10 mM CaCl2, and 10 mM Tris / HCl, pH 7.4). All experiments were performed on the type B photoreceptor. Preparations were immobilized on a glass slide using the weight of stainless-steel pins with the ends embedded in vaseline. Prior to impaling type B photoreceptors with a microelectrode, the thin connective tissue sheath was digested by incubation in protease (type XXIV or type VIII, Sigma Chemical Co., St. Louis, MO) solution (1 mg/ml) for 8/10 min at 20 8C. Intracellular recordings were made with three fluorescent dyes, Lucifer yellow CH (5% solution in 1 M

LiCl: L-453, Molecular Probes, Eugene, OR), Alexa 488 and Alexa 594 (1% in distilled water; A-10436, A-1038, Molecular Probes, Eugene, OR) or 3 M KCl- filled glass microelectrode with input resistance ranging from 20 to 30 MV (measured with 3 M KCl filling electrode) connected by a silver chloride wire to a high input impedance amplifier (7110A, Pelagic Electronics, Falmouth, MA). For double staining experiments and for observing dynamical changes in morphology in live preparations due to conditioning, the fluorescent dyes Alexa 488 and Alexa 594 were included in electrodes. Though Alexa dyes have advantage to give much brighter fluorescence Alexa dyes are not suitable for fixed gastropoda-tissues, because fluorescence disappeared during fixation and dehydration process. In this study they were used only for live preparations. Voltage responses were recorded on a storage oscilloscope (DCS-7020, Kenwood, Tokyo) and on a chart recorder (Rectihoriz 8, NEC-Sannei, Tokyo) and analyzed using a microcomputer via an interface board (Digidata 1200, Axon Instruments, Foster City, CA) and analysis software (p-clamp, Axon Instruments). The I /V relation was obtained after correction for the electrode resistance by measuring the voltage drop due to a constant current injection from /0.2 to /0.2 nA in 0.1 nA steps. The slope input resistance was assessed from a linear regression of the I/V relation. If the electrode resistance differed by more than 10% as measured in ASW before and after the cell was impaled, it was not used. The timing of a light flash from a tungsten halogen lamp (HL-100, HOYA SCHOTT, Tokyo) was controlled using a solenoid mechanical shutter (EC-601, Copal, Tokyo) in the lamp house and the light directed underneath the preparation using a fiber optic cable. The light intensity at the preparation was always less than 5.8 mW/cm2 at 510 nm. The sensitivity of the type B photoreceptor is maximal at this wavelength. Light stimuli were presented every 2 min. To estimate the time of the long-lasting depolarization (West et al., 1982), we measured the time required for the membrane potential to repolarize to within 10% of the resting membrane potential after cessation of the flash. The resting membrane potential of typical healthy type B photoreceptors ranges from /40 to /60 mV after 10 min of dark adaptation. Thus, after dark adaptation, if the resting membrane potential was not within this range, the cell was discarded. 2.3. In vitro conditioning In the paradigm of visuo-, vestibular-associative learning, the conditioning stimulus (CS) was the light flash and the unconditioning stimulus (UCS) was the vibro-tactile stimulus to a statocyst. The CS consisted of a white light focused through a fiber optic bundle that


reached the preparation with an intensity less than 5.8 mW/cm2 at 510 nm. The UCS consisted of statocyst hair cell stimulation accomplished by gently vibrating the statocyst with a polished glass probe with a concave tip approximately 50 mm in diameter. This probe was connected to a piezo-driver (DPS-255, Dia Medical Co., Tokyo) that produced a 33 Hz vibration with a maximum amplitude of 19 mm. The method of the in vitro conditioning employed in this study was basically identical as previous in vitro conditioning studies (Matzel and Rogers, 1993; Matzel et al., 1996; Tomsic and Alkon, 2000). First, we recorded three light responses, which were evoked by a 1-s light flash every 2 min following 10 min dark adaptation period. In order to stabilize the membrane potential we waited an additional 10 min, then the I/V relation was obtained by measuring the voltage in response to a constant current injection of /0.2 to /0.2 nA in 0.1 nA steps. Following a second 10-min stabilization period, the in vitro paired conditioning stimulation paradigm was applied as follows. Five successive paired stimuli; a 3-s light stimulus and a 2-s vestibular stimulus with a 1-s delay, were presented every 2 min. After another 10-min stabilization period, we measured the input resistance again. The preparation was then dark adapted for 10 min and a post-conditioning light responses were measured. Altogether, the experimental procedure required more than 90 min. There were two control groups; in one group preparations received the same amount of stimulation as the paired conditioning, but received explicitly unpaired presentations of the CS and UCS (unpaired) interposed by 60 s. In another group, preparations received 3-s CS presentations alone (light). Soon after the in vitro conditioning, fluorescent dye was injected using iontophoresis with alternating current of 1 Hz, /0.5 nA in 50% duty cycle for 30 min. Most of the in vitro conditioning was conducted under double-blind conditions and the electrophysiological test score was unknown to those analyzing the morphology (except for the experiments that measured the dynamical change in morphology of living tissue caused by the conditioning protocol).

2.4. Morphology After the dye injection, all preparations were placed in the dark in ASW for 30 min to allow the dye to diffuse. The nervous tissue was fixed with 4% paraformaldehyde in 0.1 M PO4 buffer at pH 7.4 for 3 h, and then dehydrated through a 70 /100% methanol series. The whole-mount preparation was cleared using methyl salicilate on a glass slide embedded with resin (Biolet, Oukenn Co., Tokyo). Dye injection and morphological observation of the conditioning-induced dynamical change of the terminal shape in living type B cell was

365

done on a confocal microscope (TCSNT, Leica, Heerbrugg, Switzerland). 2.5. Image processing The morphology of the terminal branch in the type B photoreceptor was analyzed using a confocal microscope. Before image acquisition, we estimated the depth of a preparation by scanning along the z -axis to acquire a series of cross-sections along the axis for 3-dimensional (3-D) reconstruction. Four cross-sectional images were obtained at each depth and averaged. Each image involved 1024 /1024 pixel scanning from the top of the branch to the bottom in 1 mm steps. Each pixel was represented as the brightness index a value between 1 and 256. The maximum intensity projection was defined by the maximum brightness index along the z -axis which yielded the largest cross-sectional image of the type B photoreceptor terminal. From the overlay image of the maximum intensity projection the terminal length was measured. Since it was difficult to discriminate the border of the neuron from the background especially for living tissue, we set the threshold for each sample for border discrimination by applying a ‘Discriminant and Least Square Criteria’ algorithm combined with an ‘Average Intensity Projection’ algorithm (Kawai and Sakakibara, 2001). The algorithms yielded an objective threshold and provided a reasonable 3-D image for comparison. This procedure was carried out using an Engineering Work Station (Digital personal work station 500AU, DEC, Maynard, MA) with image analysis software (AVS EXPRESS developer, AVS Systems, Waltham, MA) and an IBM-PC compatible personal computer with Scion Image (Scion Co., Frenderick, MD) and MS-Excel (Microsoft Co., Tokyo, Japan). Finally the terminal volume was estimated from the summation of each cross-sectional area (total number of pixels) and the voxel size (depth). 2.6. Statistical tests Statistical comparisons of input resistance and terminal length ratio among three groups (paired, unpaired and light) were made with one-way ANOVA, input resistance before and after conditioning paradigms was compared by Paired t -Test, and the comparisons of morphological features of each paradigm were done by Independent t-Test (ORIGIN 5.0, Microcal, Northampton, MA).

3. Results Paired stimulation of visual and vestibular sensory receptors results in the suppression of positive phototactic behavior (Farley et al., 1983; Crow and Forrester,

366


1986), as well as cumulative depolarization and an increase in input resistance (West et al., 1982). Thus, input resistance and long-lasting depolarization after light stimulation, a reliable indication of acquisition of learning, were used to assess in vitro conditioning. 3.1. Input resistance Fig. 1 compares the input resistance measured preand post-treatment from type B photoreceptors of both

the paired in vitro-conditioned group and the control groups. There was no significant difference between the pre- and post-treatment input resistance for the control groups (light-pre: 17.09/5.9 MV, n /5, light-post: 17.19/8.9 MV, n/5; unpaired-pre: 14.99/4.4 MV, n/ 5, unpaired-post: 16.09/3.7 MV, n/5) while the posttraining input resistance of the in vitro-conditioned group increased significantly (paired-pre: 18.99/2.4 MV, n /5, paired-post: 38.9 9/16.6 MV, n /5) (t// 2.90, P /0.044). The statistical analysis of the input resistance for each group is summarized in Table 1. In this comparison no significant difference was observed between the pre- or post-treatment input resistance of the control group and the pre-training input resistance of the in vitro-conditioned group. There was, however, a significant increase in the post-training input resistance of the in vitro-conditioned group when compared with the pre- or post-treatment control group or the pretraining input resistance of the in vitro-conditioned group. These data suggest that five paired presentations of light and vestibular stimulation are sufficient for a significant increase in input resistance in vitro. In contrast, to achieve stable conditioning in vivo, it is necessary to present at least 150 paired stimuli for 3 days (Farley and Alkon, 1987). 3.2. Light response

Fig. 1. Membrane input resistance of control (light and unpaired) and in vitro-paired-conditioned type B photoreceptors. Membrane input resistance was assessed by measuring the voltage change in response to a constant current injection of /0.2 to /0.2 nA in 0.1 nA step from light, unpaired and paired preparation. Each slope resistance was the average value from five preparations and the bars of each point was standard deviation. The input resistance neither pre- and posttreatment of the control group nor pre-treatment of the control and pre-training of the in vitro-conditioned groups were significantly different (a). The input resistance of post-training of in in vitroconditioned group was significantly (*P B/0.05) increased when compared with that pre- or post-treatment of the control or pretraining of the in vitro-conditioned groups (b).

One of the other characteristics of the conditioned type B photoreceptor is a prominent elongation of the time course of repolarization of the membrane potential after a light stimulus (West et al., 1982). This feature*/ called long-lasting depolarization */has also been reported in the in vitro-conditioned type B photoreceptor (Farley et al., 1983; Farley and Alkon, 1987). Thus, in vitro conditioning was confirmed in the present study by the presence of a significant long-lasting depolarization (Fig. 2). The time required for repolarization of the membrane potential was 3.7 times longer in the in vitroconditioned type B photoreceptor (Fig. 2) compared with the control (light-pre: 129/5 s, light-post: 149/4 s; paired-pre: 119/3 s, paired-post: 539/36 s), and this cumulative depolarization was statistically significant (P B/0.05). This conditioning specific cumulative depolarization confirmed that five paired presentations of light and vestibular stimulation increased neuronal excitability in the conditioned type B photoreceptor. 3.3. Morphology Overlay images of the maximum intensity projections obtained from control (unpaired and light) and the in vitro-paired /conditioned type B photoreceptor axon terminals are shown in Fig. 3. Fig. 3a was obtained from the same sample in each group before and after the in vitro conditioning. The length of the terminal branch


367

Table 1 Statistical analysis comparing input resistance and the terminal length ratio from pre- and post-treatment for each group Measurements

Statistics

Input resistance Pre (light vs. unpaired vs. paired) Post (light vs. unpaired vs. paired) Light pre vs. post Unpaired pre vs. post Paired pre vs. post

F(2,12) 1.010 F(2,12) 6.786 t 0.051 t 0.762 t 2.897

P 0.393 P 0.011 P 0.961 P 0.489 P 0.044

N.S. Significantly different N.S. N.S. Significantly different

Terminal length ratio Light vs. unpaired vs. paired Light vs. unpaired Light vs. paired Unpaired vs. paired

F(2,14) 8.215 t 1.928 t 4.574 t 2.275

P 0.004 P 0.083 P 0.002 P 0.046

Significantly different N.S. Significantly different Significantly different

NS, not significant.

Yellow. The averaged intensity profile in Fig. 4b demonstrated that the paired B cell shortened in length by 10% when comparing the extent of pixels with an

Fig. 2. Response to a light stimulus of in vitro-conditioned type B photoreceptor. The time required for the membrane potential to repolarize to the resting membrane level (indicated with a dashed line) was significantly longer in in vitro-conditioned type B photoreceptors. Stim, light stimulus. The waveforms shown were the light response of pre- and post- of the in vitro conditioning.

in the paired animal clearly shortened in comparison with the controls, light and unpaired. This was also demonstrated in Fig. 3b. These images of pre- and posttreatment were made from living tissue using the fluorescent dyes Alexa 488 and Alexa 594. The image of the maximum intensity projection revealed the maximum cross-sectional area of a terminal along z axis. The length of the axon terminal in paired / conditioned type B photoreceptors shortened compared with control type B photoreceptors. The ratio of pretreatment terminal length to post-treatment terminal length along the centro-lateral axis was significantly different between paired and control (F(2,14) /8.215; P /0.004) while no significant difference was observed among controls as shown in Fig. 3b. Fig. 4 shows the overlay images of the maximum intensity projection and the averaged intensity along the axis from the type B photoreceptor in a light and a paired preparation. These images were made from fixed tissue using Lucifer

Fig. 3. Dynamical changes in morphology of the axon terminal of type B photoreceptor due to the paired conditioning. Overlay images of the maximum intensity projection from the live photoreceptor in a light, unpaired and paired preparation were obtained using a confocal microscope (a). Cells were intracellularly stained with 1% Alexa 488 by iontophoresis. After the conditioning protocol preparations were stabilized for 30 min before image acquisition. Axon running from the right terminated at the left shown by interposed line. Note that the terminal of light and unpaired did not show remarkable change both in pre- and post-treatment, however, the terminal of the paired preparation shortened in length by 6%. This shortening in paired preparation was significant on the terminal length ratio defined by length (post)/ length (pre) (b). *P B/0.05 error bars were standard deviations.

368


estimated from the product of cross-sectional area and depth of preparation. Both the total amount of intensity and the terminal volume obtained from control and paired animals were significantly different between control and paired animals (Table 2). Fig. 5 shows reconstructed 3-D images obtained from the paired and the control type B photoreceptor axon terminals recorded at two different angles; one along the centrolateral axis and the other along the dorso-ventral axis. The length and depth of the in vitro-conditioned terminals were significantly reduced compared to controls, while the width did not differ between the in vitroconditioned and control terminals (Fig. 6, Table 2). The statocyst hair cell is assumed to be the presynaptic element of the type B photoreceptor in associative learning in Hermissenda. In this last experiment we observed the morphological relation between the pre-synaptic and post-synaptic elements of the type B photoreceptors. First, simultaneous recordings were made from a caudal hair cell and a lateral type B cell connected with a functional synapse. Before dye injection, the synaptic connection was confirmed by observing that excitation of the type B cell, caused by light stimulus or current injection, inhibited hair cell activity (data not shown) as previously mentioned (Tabata and Alkon, 1982). Then fluorescent dyes were injected using iontophoresis. Fig. 7 shows the 3-D images from two functional synaptic pairs. The caudal hair cell terminal is in red, and the lateral type B cell terminal is in green; the right images were obtained from a preparation that received in vitro conditioning treatment and the left images were obtained from a preparation that received control (light alone) treatment. Fig. 7 demonstrates that the area of shrinkage at the terminal induced by in vitro condition-

Table 2 Statistical comparison of control and in vitro-conditioned amount of intensity, terminal length, width, depth, and volume

Fig. 4. Overlay images of the maximum intensity projection from the type B photoreceptor along the centro-lateral axis in a control (light) and in vitro-paired-conditioned preparation (a). Note that the length along the axis was shorter in the in vitro-conditioned terminal. Scale bars: 20 mm. The averaged intensity profile along the contra-lateral axis was displayed (b).

intensity greater than 15. The total amount of intensity obtained from the spatial summation along the x position was proportional to the terminal volume

Measurements

Statistics

Amount of intensity control vs. in vitro conditioned Terminal length control vs. in vitro conditioned Terminal width control vs. in vitro conditioned Terminal depth control vs. in vitro conditioned Terminal volume control vs. in vitro conditioned

Significantly different t 2.459; P 0.018 Significantly different t 4.456; P 6.587 10 5 N.S. t 1.565; P 0.126 Significantly different t 2.94; P 0.005 Significantly different t 2.838; P 0.007

Note that two (length and depth) out of three indices were statistically significantly different and thus the product of these indices, volume, was also significantly reduced in the in vitro-conditioned terminals.


369

4. Discussion

Fig. 5. Images of type B photoreceptor terminals reconstructed in three dimensions. Terminal length, width, depth, and volume were determined from these images. Terminal length was determined from the length of the trunk to the farthest terminal obtained from the maximum intensity projection. The depth was determined from the level of the first appearance image and the last one along the z -axis (dorso-ventral axis) from a series of confocal images recorded in 1 mm steps. Images were also recorded along the dorso-ventral axis to determine the width. The terminal volume was estimated from the summation of each cross-sectional area (total number of pixels) and the voxel size. Scale bars: 20 mm.

ing was not the area where the B cell made contact with the hair cell. The area of contact is the same both preand post-treatment, in both conditioned and control preparations.

The present study is the first in which morphological changes have been examined after in vitro conditioning of type B photoreceptors of Hermissenda . The in vitro conditioning produced by 5 paired light and rotation stimuli has been previously shown to result in an increase in the duration of evoked action potentials and a decrease in the amplitude of the spike afterhyperpolarization (Gandhi and Matzel, 2000) together with enhanced excitability (i.e. input resistance, evoked spike rate, long-lasting depolarization) in type B photoreceptors (Farley and Alkon, 1987; Matzel et al., 1996). The in vitro conditioning produced by our pairing protocol was confirmed using physiological measures in one set of animals. In a separate set of animals, morphological changes in terminal branches were examined after in vitro-conditioning, using the same measurement procedure used to examine morphological changes after in vivo conditioning (Alkon et al., 1990). The main result of this study is that, using this procedure, changes in terminal branch morphology are observed within an hour after the paring protocol. Contraction of the terminal branch of type B photoreceptors is one of the prominent features in Pavlovianconditioned Hermissenda observed at least 3 days after acquisition of learning (Alkon et al., 1990). The results of the present study are consistent with contraction of the terminal branches of type B photoreceptors observed in Hermissenda at least 3 days after conditioning. The striking feature of the present study is, however, based on the dynamical change in morphology due to the

Fig. 6. Summarization of the terminal length (a), width (b), depth (c), and volume (d). The in vitro-conditioned terminals were significantly reduced by all measures except terminal width. **P B/0.01. Error bars were standard deviations.

370


Fig. 7. The 3-D morphology of the pre-synaptic element, the caudal hair cell (red) and the post-synaptic element, the type B photoreceptor terminal (green) were shown. Cells were stained with fluorescent dye (Alexa 488 and Alexa 594). Both pair showed the mutual inhibition (data not shown).

conditioning in living tissue as shown in Fig. 3. The terminals changed shape even faster than we predicted, indicating that the morphological changes parallel the learning process and that even short-term synaptic plasticity may require changes in macromolecular synthesis. We have to mention that the degree of long-term morphological modifications at the synapse revealed by Alkon et al. correlated well with phototactic behavior of animals and the effective number of pairings for the associative training to show the marked phototaxis for whole animals is at least 45 pairings per day for 4 days (Alkon et al., 1990). Though we have no means to show the behavioral correlation after the in vitro conditioning, Farley and Alkon showed that even five pairings of light and rotation resulted in short-term suppression of phototactic behavior together with the increase of input resistance of the type B cell. For a whole animal conditioning this short-term suppression, however, does not last long because it takes time to modify every synapse resulting in phototaxis assuming synaptic interactions of type B and A photoreceptors, inter-neurons, and motor-neurons (Farley and Alkon, 1987). It has been suggested previously that short-term synaptic plasticity does not require macromolecular synthesis, but only requires changes in second messenger system activity (Kandel and Schwartz, 1982). Recent studies, however, indicate that there are rapid actin-

dependent morphological changes in mammalian (Fischer et al., 2000) and invertebrate preparations (Hatada et al., 2000). Morphological modification requires not only newly synthesized synaptic element but also a decrease or elimination of some synaptic elements. Dendritic spines in cultural hippocampal neurons change their length in response to glutamate application depending on the stimulation intensity and duration. Stimuli causing a massive increase in cytosolic calcium level cause a decrease in spine length; stimuli that give rise to a moderate increase in calcium cause an increase in spine length (Korkotian and Segal, 1999; Segal et al., 2000). The results of the present study suggest that the morphological changes in the terminal branch may underlie the in vitro conditioning. The mechanism underlying the morphological changes of the terminal remains unknown, but several previously identified molecules are candidates. For classical conditioning in Hermissenda , the GTP binding protein calexcitin has a key role in associative learning in type B photoreceptors (Nelson et al., 1994, 1996). This protein is the substrate of protein kinase C and inactivates K channels when PKC is activated by calcium release from ryanodine receptors (Blackwell and Alkon, 1999; Nelson et al., 1999). Furthermore, calexcitin reduces retrograde but not anterograde transport of organelles (Moshiach et


al., 1993). This raised the possibility that such alteration of organelle movement induced by calexcitin may contribute to memory specific changes in morphology in type B photoreceptor terminals shortly after in vitro conditioning. Recent findings by Crow and Xue-Bian (2000) indicate that the phosphoprotein homologous to the b-thymosin family may also be involved in the modification of the terminal shape. In some preparations it was possible to label both the terminal branches of pre-synaptic hair cells and the terminal branches of post-synaptic type B photoreceptors. The terminal contraction was not observed at the functional synapse between hair cell and type B photoreceptor synapse; rather the morphological change occurred at other parts of type B photoreceptor terminal branch which makes synaptic contact with the type A photoreceptor or post-synaptic neurons in the optic ganglion. Several lines of evidence suggest that in vitro conditioning induces synaptic facilitation at the postsynaptic target, the type A photoreceptor (Farley and Alkon, 1987; Matzel et al., 1996; Talk and Matzel, 1996; Gandhi and Matzel, 2000). This double labeling technique can be to assess whether synaptic facilitation is accompanied by an increase in synaptic contact between type B and type A photoreceptors.

Acknowledgements We thank Dr Blackwell for critical reading of the manuscript and giving valuable suggestion. This study was supported by Grant-in-Aids (07279105, 11168231, 12680783) for Scientific Research, the Ministry of Education, Science, Sports, and Culture of Japan to MS, and by the Proposed-Based New Industry Creative Type Technology R&D Promotion Program from the New Energy and Industrial Technology Development Organization (NEDO) of Japan in the field of Biocybernetics (98S18-001-2) to MS. ASW, prepared with ‘Sea Life’ was the cordial gift from Marine-Tech Co., Tokyo.

References Alkon, D.L. 1987. Memory Traces in the Brain. Cambridge University Press, New York. Alkon, D.L., Ikeno, H., Dworkin, J., McPhie, D.L., Olds, J.L., Lederhendler, I., Matzel, L., Schreurs, B.G., Kuzirian, A., Collin, C., Yamoah, E. 1990. Contraction of neuronal branching volume: an anatomic correlate of Pavlovian conditioning. Proc. Natl. Acad. Sci. USA 87, 1611 /1614. Bailey, C.H., Kandel, E.R. 1993. Structural changes accompanying memory storage. Annu. Rev. Physiol. 55, 397 /426. Bartsch, D., Ghirardi, M., Skehel, P.A., Karl, K.A., Herder, S.P., Chen, M., Bailey, C.H., Kandel, E.R. 1995. Aplysia CREB2 represses long-term facilitation: relief of repression converts

371

transient facilitation into long-term functional and structural change. Cell 83, 979 /992. Blackwell, K.T., Alkon, D.L. 1999. Ryanodine receptor modulation of in vitro associative learning in Hermissenda crassicornis . Brain Res. 822, 114 /125. Crow, T., Forrester, J. 1986. Light paired with serotonin mimics the effect of conditioning on phototactic behavior of Hermissenda . Proc. Natl. Acad. Sci. USA 83, 7975 /7978. Crow, T., Xue-Bian, J.J. 2000. Identification of a 24 kDa phosphoprotein associated with an intermediate stage of memory in Hermissenda . J. Neurosci. 20, RC74. Engert, F., Bonhoeffer, T. 1999. Dendritic spine changes associated with hippocampal long-term synaptic plasticity. Nature 399, 66 / 70. Farley, J., Alkon, D.L. 1987. In vitro associative conditioning of Hermissenda : cumulative depolarization of type B photoreceptors and short-term associative behavioral changes. J. Neurophysiol. 57, 1639 /1668. Farley, J., Richards, W.G., Ling, L.J., Liman, E., Alkon, D.L. 1983. Membrane changes in a single photoreceptor cause associative learning in Hermissenda . Science 221, 1201 /1203. Fischer, M., Kaech, S., Wagner, U., Brinkhaus, H., Matus, A. 2000. Glutamate receptors regulate actin-based plasticity in dendritic spines. Nat. Neurosci. 3, 887 /894. Gandhi, C.C., Matzel, L.D. 2000. Modulation of presynaptic action potential kinetics underlies synaptic facilitation of type B photoreceptors after associative conditioning in Hermissenda . J. Neurosci. 20, 2022 /2035. Hatada, Y., Wu, F., Sun, Z.Y., Schacher, S., Goldberg, D.J. 2000. Presynaptic morphological changes associated with long-term synaptic facilitation are triggered by actin polymerization at preexisting varicositis. J. Neurosci. (Online) 20, RC82. Kandel, E.R., Schwartz, J.H. 1982. Molecular biology of learning: modulation of transmitter release. Science 218, 433 /443. Kawai, R., Sakakibara, M. 2001. The segmentation method for confocal microscopic images. J. Adv. Sci. 12, 397 /402. Kemenes, G., Staras, K., Benjamin, P.R. 1997. In vitro appetitive classical conditioning of the feeding response in the pond snail Lymnaea stagnalis . J. Neurophysiol. 78, 2351 /2362. Korkotian, E., Segal, M. 1999. Bidirectional regulation of dendritic spine dimensions by glutamate receptors. Neuroreport 10, 2875 / 2877. Maletic-Savatic, M., Malinow, R., Svoboda, K. 1999. Rapid dendritic morphogenesis in CA1 hippocampal dendrites induced by synaptic activity. Science 283, 1923 /1927. Matzel, L.D., Muzzio, I.A., Talk, A.C. 1996. Variations in learning reflect individual differences in sensory function and synaptic integration. Behav. Neurosci. 110, 1084 /1095. Matzel, L.D., Rogers, R.F. 1993. Postsynaptic calcium, but not cumulative depolarization, is necessary for the induction of associative plasticity in Hermissenda . J. Neurosci. 13, 5029 /5040. Moshiach, S., Nelson, T.J., Sanchez-Andres, J.V., Sakakibara, M., Alkon, D.L. 1993. G-protein effects on retrograde axonal transport. Brain Res. 605, 298 /304. Nargeot, R., Baxter, D.A., Byrne, J.H. 1997. Contingent-dependent enhancement of rhythmic motor patterns: an in vitro analog of operant conditioning. J. Neurosci. 17, 8093 /8105. Nelson, T.J., Cavallaro, S., Yi, C.L., McPhie, D., Schreurs, B.G., Gusev, P.A., Favit, A., Zohar, O., Kim, J., Beushausen, S., Ascoli, G., Olds, J., Neve, R., Alkon, D.L. 1996. Calexcitin: a signaling protein that binds calcium and GTP, inhibits potassium channels, and enhances membrane excitability. Proc. Natl. Acad. Sci. USA 93, 13808 /13813. Nelson, T.J., Yoshioka, T., Toyoshima, S., Han, Y.F., Alkon, D.L. 1994. Characterization of a GTP-binding protein implicated in both memory storage and interorganelle vesicle transport. Proc. Natl. Acad. Sci. USA 91, 9287 /9291.

372


Nelson, T.J., Zhao, W.Q., Yuan, S., Favit, A., Pozzo-Miller, L., Alkon, D.L. 1999. Calexcitin interaction with neuronal ryanodine receptors. Biochem. J. 341, 423 /433. Segal, I., Korkotian, I., Murphy, D.D. 2000. Dendritic spine formation and pruning: common cellular mechanisms? Trends Neurosci. 23, 53 /57. Shi, S.H., Hayashi, Y., Petralia, R.S., Zaman, S.H., Wenthold, R.J., Svoboda, K., Malinow, R. 1999. Rapid spine delivery and redistribution of AMPA receptors after synaptic NMDA receptor activation. Science 284, 1811 /1816. Tabata, M., Alkon, D.L. 1982. Positive synaptic feedback in visual system of nudibranch mollusk Hermissenda crassicornis . J. Neurophysiol. 48, 174 /191.

Talk, A.C., Matzel, L.D. 1996. Calcium influx and release from intracellular stores contribute differentially to activity-dependent neuronal facilitation in Hermissenda photoreceptors. Neurobiol. Learn Mem. 66, 183 /197. Tomsic, D., Alkon, D.L. 2000. Background illumination effects upon in vitro conditioning in Hermissenda . Neurobiol. Learn Mem. 74, 56 /64. West, A., Barnes, E., Alkon, D.L. 1982. Primary changes of voltage responses during retention of associative learning. J. Neurophysiol. 48, 1243 /1255. Yin, J.C., Del Vecchio, M., Zhou, H., Tully, T. 1995. CREB as a memory modulator: induced expression of a dCREB2 activator isoform enhances long-term memory in Drosophila . Cell 81, 107 / 115.