Intermediate and long-term memory are different at the neuronal level in Lymnaea stagnalis (L.)

Neurobiology of Learning and Memory 96 (2011) 403–416

Contents lists available at ScienceDirect

Neurobiology of Learning and Memory journal homepage: www.elsevier.com/locate/ynlme

Intermediate and long-term memory are different at the neuronal level in Lymnaea stagnalis (L.) Marvin H. Braun ⇑, Ken Lukowiak Hotchkiss Brain Institute, Department of Physiology and Pharmacology, University of Calgary, Calgary, Alberta, Canada T2N 4N1

a r t i c l e

i n f o

Article history: Received 23 September 2010 Revised 26 June 2011 Accepted 29 June 2011 Available online 3 July 2011 Keywords: Intermediate-term memory Long-term memory Lymnaea stagnalis Learning Neurons Operant conditioning

a b s t r a c t Both intermediate-term memory (ITM) and long-term memory (LTM) require novel protein synthesis; however, LTM also requires gene transcription. This suggests that the behavioural output of the two processes may be produced differently at the neuronal level. The fresh-water snail, Lymnaea stagnalis, can be operantly conditioned to decrease its rate of aerial respiration and, depending on the training procedure, the memory can last 3 h (ITM) or >24 h (LTM). RPeD1, one of the 3 interneurons that form the respiratory central pattern generator (CPG) that drives aerial respiration, is necessary for memory formation. By comparing RPeD1’s electrophysiological properties in naïve, ‘ITM-trained’, ‘LTM-trained’ and yoked control snails we discovered that while the behavioural phenotype of memory at 3 and 24 h is identical, the situation at the neuronal level is different. When examined 3 h after either the ‘ITM’ or ‘LTM’ training procedure RPeD1 activity is significantly depressed. That is, the firing rate, input resistance, excitability and the number of action potential bursts are all significantly decreased. In snails receiving the ITM-training, these changes return to normal 24 h post-training. However, in snails receiving the ‘LTM-training’, measured RPeD1 properties (firing rate, excitability, membrane resistance, and the number of action potential bursts fired) are significantly different at 24 h than they were at 3 h. Additionally, 24 h following LTM training RPeD1 appears to be functionally ‘‘uncoupled’’ from its control of the pneumostome as the link between RPeD1 excitation and pneumostome opening is weakened. These data suggest that the behavioural changes occurring during LTM are due to more widespread neuronal reorganization than similar behavioural changes occurring during ITM. Thus ITM and LTM are not just distinct in a chronological and transcriptional manner but are also distinct at the level of neuronal properties. Ó 2011 Elsevier Inc. All rights reserved.

1. Introduction Historically, the study of memory has been separated into two discrete processes, short-term memory (STM), lasting for several minutes and long-term memory (LTM), lasting days to years (Rosenzweig, Bennettt, Colombo, Lee, & Serrano, 1993). Short-term memory is dependent upon enzymes such as protein kinase C (PKC), cAMP-dependent protein kinase (PKA), and calcium/calmodulin-dependent kinase (CaM-kinase). These are activated in response to the elevation of intracellular [Ca2+] and adenylyl cyclase which occurs following synaptic excitation. These enzymes directly modulate ion channels and transmitter release, resulting in STM (reviewed in Kandel, 2001). However, with continued stimulation, the high levels of kinases migrate to the nucleus and begin ⇑ Corresponding author. Address: Hotchkiss Brain Institute, Department of Physiology and Pharmacology, University of Calgary, Room 2104 Health Sciences Centre, 3330 Hospital Drive NW, Calgary, Alberta, Canada T2N 4N1. Fax: +1 403 283 2700. E-mail addresses: [email protected] (M.H. Braun), [email protected] (K. Lukowiak). 1074-7427/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.nlm.2011.06.016

to activate additional enzymes (e.g. mitogen-activated protein kinase (MAPK)) and transcription factors (e.g. cAMP response element binding protein-1 (CREB-1)), which promote the transcription of various target genes and are crucial towards the initiation of LTM (Kandel, 2001). While this work has been critical in understanding some of the molecular underpinnings of memory, the dichotomy between STM and LTM is overly simplistic and with the last few decades has come the realization that there is an additional mechanistically distinct stage of memory, termed intermediate term memory (ITM). While first described behaviourally in chicks (Gibbs & Ng, 1979; Rosenzweig et al., 1993), similar phenomena have now been shown to occur in Aplysia (Ghirardi, Montarolo, & Kandel, 1995), Lymnaea stagnalis (Lukowiak, Adatia, Krygier, & Syed, 2000), Caenorhabditis elegans (Steidl, Rose, & Rankin, 2003), Drosophila (Margulies, Tully, & Dubnau, 2005) and Apis (Müller, 1996). Somewhat longer than STM (a few hours instead of minutes), ITM requires novel protein synthesis like LTM. However, unlike LTM, ITM is not dependent on gene transcription (Davis & Squire, 1984; Sutton, Masters, Bagnall, & Carew, 2001). This has been clearly demonstrated in both Aplysia (Sutton et al., 2001) and Lymnaea (Sangha, Scheibenstock, McComb, & Lukowiak,

404

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

2003), whereby blockade of mRNA synthesis still resulted in ITM but blockade of both protein synthesis and gene transcription resulted in neither ITM nor LTM. The study of the causal neuronal and molecular mechanisms of learning and memory in mammals is exceedingly difficult. With billions of neurons, glia and synapses to choose from, pinpointing the specific neural changes correlating with a specific memory is problematic. Therefore, using simpler model systems such as Aplysia, Lymnaea or Drosophila has often been the only method whereby many of the basic processes involved in memory formation and maintenance can be elucidated. For this reason, the pond snail, Lymnaea, is a prime model-system for the study of non-declarative associative learning and the subsequent memory formation. Bimodal breathers capable of extracting oxygen from both water and air, Lymnaea can be quickly trained (i.e. operant conditioning) to significantly decrease the amount of aerial respiration they perform (Lukowiak, Cotter, Westly, Ringseis, & Spencer, 1998; Lukowiak, Ringseis, Spencer, Wildering, & Syed, 1996) and, depending on the training procedure, these memories can last from 3 h to >7 days (Lukowiak et al., 2000, 2010; Orr & Lukowiak, 2008). Three identified interneurons (IP3, VD4, and RPeD1) make up the respiratory central pattern generator (CPG) in Lymnaea and the coordinated output of these neurons generates a pattern of action potential bursts that drives the pneumostome (the respiratory orifice) to open and close (Lukowiak, 1991; Syed, Bulloch, & Lukowiak, 1990; Syed, Ridgway, Lukowiak, & Bulloch, 1992). Not only are bursts of action potentials in RPeD1 strongly correlated with pneumostome opening (Khan & Spencer, 2009; McComb, Meems, Syed, & Lukowiak, 2003; McComb, Rosenegger, Varshney, Kwok, & Lukowiak, 2005; Spencer, Kazmi, Syed, & Lukowiak, 2002) but experimental stimulation of RPeD1 alone is usually sufficient to induce pneumostome opening as well (McComb et al., 2005; Spencer, Syed, & Lukowiak, 1999). As operant conditioning of the snails results in a decrease in the activity of these neurons, specifically RPeD1, this model has advantages over most others in that a specific behavioural memory (decrease in aerial respiration) can be linked to changes within a specific neuron (RPeD1) that has been experimentally shown to be necessary for the establishment of LTM, memory reconsolidation, extinction and forgetting (Sangha, Scheibenstock, & Lukowiak, 2003; Scheibenstock, Krygier, Haque, Syed, & Lukowiak, 2002; Sangha, Scheibenstock, McComb, et al., 2003; Sangha et al., 2005). However, despite its requirement for the establishment of LTM, following LTM training, there appears to be a reorganization of the CPG such that RPeD1 becomes functionally uncoupled from the opener motoneurons and its ability to cause pneumostome opening is diminished (Khan & Spencer, 2009; McComb et al., 2005; Spencer et al., 1999). To our knowledge, there has yet to be a rigorous comparison of the neuronal changes correlated with ITM and LTM formation. By producing either ITM or LTM in the Lymnaea model-system, we can determine if the behavioural phenotype we witness (a significant decrease in the number of attempted pneumostome openings) is the result of similar or different states of neuronal activity. Based on the molecular data we currently possess, this is an open question. In Lymnaea, appetitive classical conditioning of feeding behaviour has revealed that while CaMKII is required for consolidation of LTM, it does not appear to have a role in ITM formation (Wan, Mackay, Iqbal, Naskar, & Kemenes, 2010), evidence, along with the different transcriptional requirements, that the two processes may be different. Conversely, operant conditioning of aerial respiration in Lymnaea where the snail learns and remembers not to perform the behaviour, activation of both PKC and MAPK are required for ITM and LTM (Rosenegger & Lukowiak, 2010). This result mirrors the situation in Apis, Drosophila and Aplysia, whereby the kinases PKA, PKC and MAPK regulate both ITM and LTM induction and expression (reviewed in Stough, Shobe, &

Carew, 2006), evidence supporting the notion that the processes underlying both forms of memory are similar. Despite these similarities, however, the different molecular processes and time scales of ITM and LTM imply that they are fundamentally different states and, therefore, we hypothesize that RPeD1 properties will be different in ITM and LTM. Additionally, we hope that this comparison of ITM and LTM will shed light on whether the apparent uncoupling of RPeD1 from the control of aerial respiration is a feature of learning in either both training paradigms or specific to LTM.

2. Methods 2.1. Animals L. stagnalis (L.), originally derived from stocks obtained from Vrjie University (Amsterdam), were bred and raised in the snail facility at the University of Calgary. Adult snails were maintained at room temperature (20–23 °C) in eumoxic aquaria and were fed lettuce and spinach ad libitum. Artificial pond water (0.26 g/l Instant OceanÒ, Spectrum Brands Inc.) with additional calcium sulphate dehydrate added to make what we refer to as standard calcium 80 mg/l [Ca2+] pond water (Dalesman & Lukowiak, 2010). 2.2. Operant conditioning procedure The operant training procedure was similar to that which has been used previously in the lab (Lukowiak et al., 1996). Briefly, individually labelled snails were placed in a 1 l beaker containing 500 ml of room temperature water made hypoxic (<0.1 ml O2 l 1) by bubbling N2 through it 20 min prior to and throughout the training and testing procedure. Animals were given a 10-min acclimatization period during which they were allowed to freely respire. Following this period, during the operant conditioning training session, every time a snail attempted to open its pneumostome to perform aerial respiration, a sharpened wooden applicator was used to touch the pneumostome, inducing pneumostome closure. ‘Poking’ the snail in this manner did not elicit withdrawal behaviour. The time each poke was administered was recorded and utilized for the yoked control procedure (see below). 2.3. The single training session vs. two session training The training procedure used to produce long-term memory (LTM) consisted of two 30-min operant conditioning training sessions (TS) in hypoxic pond water separated by a 1-h rest interval. We will refer to this as the ‘LTM training’ procedure. Separate groups of animals were then tested for memory (MT) at either 3 or 24 h post-training (i.e. after the second training session) by counting the number of attempted pneumostome openings. This procedure has been demonstrated to result in a memory that lasts 24 h (Lukowiak et al., 2000). Long-term memory is operationally defined as being present if the number of attempted pneumostome openings in the memory test (MT) was significantly lower than that of TS1 and not significantly higher than that of TS2 (Lukowiak et al., 1996, 2003; Sangha, Scheibenstock, McComb, et al., 2003; Sangha, Scheibenstock, & Lukowiak, 2003). The training procedure used to produce a memory that only persists for 3 h (i.e. intermediate term memory; ITM) consisted of a single 0.5 h training session that was carried out exactly as described above for producing LTM and, as with LTM-training, separate groups of animals were then tested for memory (MT) at either 3 or 24 h post-training (i.e. after the single training session). We will refer to this as the ‘ITM training’ procedure. The only difference between the ‘LTM training’ procedure and the ‘ITM training’

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

procedure is that only one 0.5 h TS is used rather than two 0.5 h TSs separated by a 1 h interval. In the ‘ITM training’ procedure memory is operationally defined as being present if the number of attempted pneumostome openings in the memory test (MT) was significantly lower than that of TS1. Yoked control snails were used to test the effect of non-contingent pokes on the number of pneumostome openings. Randomly selected control snails were ‘‘yoked’’ to particular experimental animal and these yoked controls received the tactile stimulus not when they opened their pneumostome but rather when the experimental animals they were paired with did (Lukowiak et al., 1996, 2003). 2.4. Electrophysiological recording from RPeD1 RPeD1 activity was measured in naïve snails, operantly trained snails and yoked control snails utilizing semi-intact preparations which were dissected out following the protocol previously developed in our laboratory (Dalesman, Braun, & Lukowiak, 2010; McComb et al., 2005; Orr, El-Bekai, Lui, Watson, & Lukowiak, 2007; Orr, Hittel, & Lukowiak, 2009; Orr & Lukowiak, 2008). Thus, each snail’s head-foot musculature, buccal mass, salivary ducts and reproductive organs were removed, exposing the CNS and leaving the pneumostome and the nerves innervating the pneumostome area intact. The preparation was pinned down with the dorsal surface up, covered with saline and left for 20 min. Semi-intact preparations will spontaneously open and close the pneumostome if the saline level is below the level of the pneumostome (Inoue, Takasaki, Lukowiak, & Syed, 1996; Syed & Winlow, 1991), so prior to impaling RPeD1 with a sharp glass microelectrode, the saline level was lowered below that of the pneumostome. After impaling RPeD1 with a glass microelectrode filled with saturated K2SO4 solution (tip resistances ranged from 20 to 75 MW), the preparation was given a further 15 min stabilization period prior to collection of electrophysiological data, for a total time of 35 min between dissection and data collection. Intracellular signals were amplified using a Neurodata Instruments IR283 amplifier (New York, NY, USA) and displayed simultaneously on a Macintosh PowerLab/ 4SP (ADInstruments Inc, Colorado Springs, CO, US) and a Hitachi oscilloscope (Tokyo, Japan). Recordings were stored and analyzed via Chart 5 Software (ADInstruments Inc, Colorado Springs, CO, US) using a 600 s trace for data analysis. The trace was used for measurements of spontaneous firing, resting membrane potential, burst firing frequency, number of bursts, burst length and action potentials/burst. Previous work in our lab on isolated neurons, isolated brains and semi-intact preparations have revealed that the RPeD1 bursts of action potentials (APs) which trigger pneumostome opening have particular properties. Therefore, a burst was defined as a period of sustained depolarization during which 2 to over 20 action potentials were fired. The burst was not considered over until membrane potential had returned to the resting value. A measure of RPeD1 excitability was obtained by counting the number of APs evoked when the impaled cell was driven through a series of 10 depolarizing current steps from 0.2 to 2.0 nA. The steps were 400 ms long and the cell was allowed to recover for 300 ms between steps. The input resistance of RPeD1 was measured by driving the impaled cell through a series of 10 hyperpolarizing current steps from 0.2 to 2.0 nA (step lengths are the same as above) and calculating the resistance from the resulting current–voltage relationships according to Ohm’s law (R = V/I) (Dalesman et al., 2010). It has previously been demonstrated that injection of a depolarizing current into RPeD1 is sufficient to induce pneumostome opening (McComb et al., 2005). To test the ability of RPeD1 to in-

405

duce pneumostome opening in the various groups of snails, 2.3 nA was injected for 4 s. If the pneumostome opened within 5 s of this current injection, the link between RPeD1 and the opener motoneuron was considered functional. 2.5. Data analysis As much of the electrophysiological data were nonparametric, they were analyzed for significant differences using a Kruskal Wallis test followed by a post hoc Dunn’s multiple comparisons test to find between group differences. Membrane resistance was parametric and differences were calculated using an ANOVA with a TUKEY post hoc test to find group differences. The percentage of snails that breathed following depolarization of RPeD1 was examined using a contingency table and a Chi-square test for independence. All statistics were calculated using InStat (GraphPad InStat version 3.0a for Macintosh, GraphPad Software, San Diego California USA). 3. Results 3.1. The ITM training procedure To our knowledge there has been relatively little work attempting to understand the causal neuronal changes that constitute intermediate-term memory (ITM). We therefore trained a naive cohort of snails with the ITM-training procedure and tested whether snails formed memory. There was a significant decrease in the number of attempted pneumostome openings compared to TS1 and the yoked animals 3 h after training (MT (3 h)) (Fig. 1). However, when testing a second group of ITM-trained snails 24 h after TS1 (MT (24 h)), the number of attempted pneumostome openings in the memory test session was not significantly different than the number of attempted openings in TS. Therefore, memory was present 3 h but not 24 h after training. We then determined if there were concomitant changes in the electrophysiological properties of RPeD1 following this training procedure. As with previous studies (e.g. Orr & Lukowiak, 2008), measurements of burst firing frequency and the resting membrane potential of RPeD1 in the semi-intact preparation showed no significant differences between the trained, naive and yoked control snails (Table 1). However, there was a significant decrease in the number of spontaneous APs fired by RPeD1 following this training procedure both 3 h and 24 h after TS1 (Figs. 2 and 3A). In addition, the number of spontaneous bursts of APs that we recorded in RPeD1 was significantly decreased compared to the number in naive and yoked control preparations at both 3 h and 24 h after training (Figs. 2 and 3B). While the number of bursts significantly decreases following training, there was no difference in the pattern of the bursts themselves as both the length of the bursts (Fig. 4A) and number of spikes/burst (Fig. 4B) were no different from either the control or yoked snails. Thus, even though ‘behavioural’ memory was not apparent 24 h after the single 0.5 h TS there were significant changes in the electrophysiological properties of RPeD1. These data may explain a previous finding of a residual memory trace (Parvez, Stewart, Sangha, & Lukowiak, 2005) that is present 24 h after a single TS even though the behavioural phenotype of memory was not apparent. However, we did find that following the ITM-training procedure there were significant differences in the electrophysiological properties of RPeD1 that paralleled the behavioural differences shown in Fig. 1. First, there was a significant change in the number of APs elicited when the impaled cell was driven through a series of 10 depolarizing current steps from 0.2 to 2.0 nA 3 h after training compared to the number elicited by the series of current steps in naive and yoked control preparations (Fig. 5). Notice that 24 h after

406

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

Fig. 1. Operant conditioning using the ITM-training procedure (1–0.5 h training session (TS)). Memory tests (MT) at 3 h post-training showed a significant fall in the number of attempted pneumostome opening. Data are given as mean ± S.E.M. p < 0.05 (MT (3 h), n = 12; MT (24 h), n = 12; yoked (3 h), n = 12; yoked (24 h), n = 12).

A

B

decreased excitability at 3 h appears to be due to both a change in firing threshold (p < 0.05) and in the relationship between injected current and action potential generation (p < 0.01). There were no significant differences in these relationships at 24 h after TS. This alteration in excitability of RPeD1 appears to be due in part to intrinsic changes in the membrane properties of RPeD1 as there was a concomitant significant decrease in the input resistance 3 h after the ITM-training procedure (Fig. 6). To summarise, following the ITM-training procedure behavioural memory is only observed 3 h after training. However, the electrophysiological findings are more complicated. Some of the measured parameters (e.g. number of spontaneous APs/10 min and number of bursts) are still apparent 24 h after the single TS while others (number of APs elicited by a series of current steps and the input resistance) are only observed 3 h after training. Finally, resting membrane potential and spike frequency/burst are not altered at all (Table 1). This will be discussed below. 3.2. The LTM-training procedure

C

Fig. 2. Representative recordings of the bursting behaviour of RPeD1 from semiintact preparation following the ITM training procedure. The number of bursts decreases at both 3 h and 24 h following training. (A) Naïve (B) 3 h after TS and (C) 24 h after TS.

the ITM-training procedure the number of APs elicited by the depolarizing steps was not significantly different from naive or yoked control preparations. However, the number of APs elicited by the depolarizing steps at 24 h after TS was significantly greater than the number elicited 3 h after the single TS. Comparisons of the trend lines indicating the relationship between # of action potentials fired and the depolarizing current step show that the

We were curious about whether there would be similar or different effects on RPeD1‘s electrophysiological parameters following a training procedure which results in behavioural long-term memory (LTM). When subjected to two 0.5 h training session separated by 1 h, snails demonstrated memory both 3 h and 24 h after TS2 (Fig. 7). Similar to what we found with the ITM-training procedure, snails subjected to the LTM-training procedure exhibited a significant decrease in spontaneous AP firing compared to naïve and yoked animals (Figs. 8 and 9A,). In fact, there was no significant difference in the firing rate recorded 3 h after training between snails receiving the ITM-training procedure compared to those receiving the LTM-training procedure (p = 0.41). However, unlike the snails that received just the ITM-training procedure (Figs. 2 and 3A), snails receiving the LTM-training procedure showed a further significant decrease in the firing rates 24 h post-training. In these preparations, there was a significantly lower number of APs recorded at 24 h post-training compared to the number recorded 3 h post-training (Figs. 8B and C and 9A). As was the case with snails receiving the ITM-training procedure the number of bursts recorded in RPeD1 from preparations receiving the LTM-training procedure dropped significantly at both 3 h and 24 h post training (Figs. 8 and 9B). However, in these

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

407

Fig. 3. ITM-trained snails show a decrease in the number of spikes and bursts fired by RPeD1. (A) Number of spikes/10 min fired by RPeD1 in naïve, trained (3 h and 24 h) and yoked controls (3 h and 24 h) semi-intact preparations. (B) The number of bursts fired in 10 min by RPeD1 in naïve, yoked controls and trained snails at 3 and 24 h. Data are given as mean ± S.E.M. p < 0.05, Where p values are not given, there were no significant differences (naïve, n = 8; trained (3 h), n = 11; trained (24 h), n = 12; yoked (3 h), n = 7; yoked (24 h), n = 7).

trained snails the bursts themselves were different from those observed in the naïve preparations (Fig. 10). As in preparations receiving the ITM-training procedure session, there were no changes to either burst duration or the number of spikes fired/ burst at 3 h, but after 24 h, the duration of the bursts from snails receiving the LTM-training procedure was significantly longer than those of the naïve and yoked snails (Fig. 10A). While the frequency of APs occurring during the bursts did not change with training (data not shown), due to the increased burst length, the overall numbers of action potentials fired per burst also significantly increased (Fig. 10B). The excitability of RPeD1 in snails following the LTM-training procedure was significantly lower than either the naïve or yoked control snails at both 3 and 24 h (Fig. 11). While the excitability at 3 h was no different from the results obtained from snails receiving the ITM-training procedure (Fig. 5) (p = 0.15); at 24 h following the second training session, the excitability of RPeD1 continued to decrease and was significantly lower than that at 3 h post-training. The slopes of the trend lines indicating the relationship between # of action potentials fired and the depolarizing current step show that the decreased excitability at 3 h appears to be primarily due

to a change in firing threshold. However, at 24 h, the decreased excitability is due to both an altered threshold and a change in the relationship between injected current and action potential generation. The input resistance of RPeD1 also decreased following the second training session in a manner which mirrored the excitability changes (Fig. 12). Thus, 3 h post-training the input resistance was significantly lower than that of the naïve or the yoked controls and was identical to the observed 3 h after the ITM-training procedure preparations (Fig. 6) (p = 0.36). As with excitability, by 24 h following the second training session, the input resistance was also significantly lower than that at 3 h. While bursting occurred at 3 h and 24 h in preparations made from snails receiving both the ITM and LTM-training procedures, in the preparations made from snails that received the LTM-training procedure a significant finding was the fact that the bursts at 24 h no longer correlated with pneumostome opening and initiation of aerial respiration. In the example shown in Fig. 13, while the burst in the naïve preparation resulted in pneumostome opening, the much longer burst in the preparation receiving the LTMtraining procedure did not. The naïve burst appears to be the result

408

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

A

B

Fig. 4. ITM-trained snails show no differences in RPeD1 burst properties. Comparison of RPeD1 burst properties in naïve, trained (3 h and 24 h) and yoked control (3 h and 24 h) semi-intact preparations. (A) Burst duration (s) (B) # of spikes/burst. Data are given as mean ± S.E.M. p < 0.05 where p values are not given, there were no significant differences.

of summating EPSPs while the 24 h burst is the product of large amplitude EPSPs suggesting the possibility that the bursts in the different groups of animals are a result of different synaptic inputs. However, there were no significant differences between the spike frequencies within bursts (Table 1) and while the mean burst length in the naïve group was significantly smaller than the mean length in the 24 h group (Fig. 10), there were occasional long bursts in the naïve preparations, bursts which did result in pneumostome opening (M. Braun, personal observations), Therefore, despite the differences, the bursts in both groups are likely the result of interactions of RPeD1 with the other cells of the CPG. It has previously been demonstrated that injection of a depolarizing current into RPeD1 is sufficient to induce pneumostome opening (McComb et al., 2005). In our study passing 2.30 nA for 4 s was sufficient to induce pneumostome opening in 68% of the naïve preparations, 67% and 76% of the preparations at 3 h posttraining (in preparations made from snails receiving either the ITM or LTM-training procedure respectively) and 81% of the 24 h preparations receiving the ITM-training procedure (Fig. 14). However, 24 h after the second training session only 15% of the preparations opened their pneumostomes in response to a depolarizing current. Calculation of the Chi-square value for these data indicated that there was a significant difference in the ability of RPeD1 depolarization to cause a pneumostome opening (Chi-square 35.26, df 5, p < 0.001). If the data for the trained snails (2TS-24 h)

is removed and the Chi-square recalculated, there is no longer a significant difference (Chi-square 1.962, df 4, p = 0.74). 4. Discussion Despite the apparent similarities between the behavioural output and molecular machinery of ITM and LTM (Lukowiak et al., 2000; Rosenegger & Lukowiak, 2010; Stough et al., 2006), there are differences, the most significant of which is the transcriptional requirement of LTM. While mRNA production is a requirement of LTM induction, ITM is transcriptionally independent (Sangha, Scheibenstock, McComb, et al., 2003; Sutton et al., 2001; Yin et al., 1994). In fact, ITM can be produced in neurons that no longer possess a cell body, thus demonstrating that the protein synthesis needed for ITM occurs in the neurite, possibly at the synapse (Sangha, Scheibenstock, & Lukowiak, 2003; Scheibenstock et al., 2002). We hypothesized that this distinction in the ‘production’ of the two processes at the molecular level would result in obvious electrophysiological changes in RPeD1, a neuron that is a necessary site for LTM formation (Sangha et al., 2004). The data obtained in our present study are consistent with this hypothesis, as many of the significant changes in the electrophysiological properties of RPeD1 3 h post-training are significantly different than those 24 h post-training. As such, this study both supports and builds on the previous work showing that ITM and LTM are distinct at

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

409

Fig. 5. Neuronal excitability is significantly decreased 3 h following ITM training. Excitability of RPeD1 as measured by the number of action potentials fired in response to 10 depolarizing current steps. Comparisons between naïve and trained (3 h and 24 h post-training) snails. Yoked control (3 h and 24 h) snails were not significantly different from naïve snails and are omitted for clarity. Trend lines and their slopes are added to the histograms. Inset: total # of action potentials fired over the ten current steps. Data are given as mean ± S.E.M. p < 0.05.

Fig. 6. ITM-trained snails have a significantly lower neuronal resistance 3 h following training. Input resistance of RPeD1 calculated from hyperpolarizing current steps. Comparison between naïve, trained (3 h and 24 h post-training) and yoked control (3 h and 24 h) snails. Data are given as mean ± S.E.M. p < 0.05 where p values are not given, there were no significant differences.

both the molecular and electrophysiological levels (Li, Tully, & Kalderon, 1996; Sangha, Scheibenstock, McComb, et al., 2003; Müller, 1996; Sangha, Scheibenstock, & Lukowiak, 2003; Sutton et al., 2001; Tully & Quinn, 1985). Despite exhibiting a similar

behavioural phenotype 3 h and 24 h after the second training session (i.e. a significant decrease in the number of attempted pneumostome openings), RPeD1 exhibits a different electrophysiological profile 3 and 24 h post-training.

410

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

Table 1 Resting membrane potential and spike frequency/burst measured in RPeD1 neurons from naïve, yoked, and trained snails at 3 h and 24 h.

Naïve ITM – 3 h ITM – 24 h Yoked – ITM – 3 h Yoked – ITM – 24 h LTM – 3 h LTM – 24 h Yoked – LTM – 3 h Yoked – LTM – 24 h

Spike frequency/ Resting membrane n burst (Hz) potential (mV)

Significance

1.09 ± 0.11 1.64 ± 0.45 0.89 ± 0.23 1.11 ± 0.24 1.09 ± 0.18 1.08 ± 0.11 1.51 ± 0.29 1.47 ± 0.26 1.28 ± 0.30

NSD NSD NSD NSD NSD NSD NSD NSD NSD

59.1 ± 1.8 59.0 ± 2.3 60.8 ± 1.9 61.7 ± 1.4 60.8 ± 1.2 62.0 ± 2.7 58.4 ± 1.2 60.1 ± 1.9 62.5 ± 2.5

25 11 12 7 7 9 9 8 6

A

B

4.1. The ITM-training procedure vs. The LTM-training procedure The most obvious difference between the two training procedures is that memory is evident both 3 h and 24 h (i.e. LTM) after the LTM-training procedure whereas only a 3 h memory (i.e. ITM) is apparent after the ITM-training procedure. However, the behavioural phenotype of the memory at 3 h produced by the two different training procedures (i.e. number of attempted pneumostome openings) is indistinguishable. Most interestingly, at the neuronal level there are also no major differences between the electrophysiological phenotype (i.e. firing rate, bursting rate, burst length, input resistance and excitability) possessed by RPeD1 3 h after training irrespective of whether the snail received ‘ITM’ or LTM’ training. However, in the preparations that had received the LTM-training procedure electrophysiological changes in RPeD1 continue to occur so that by 24 h after the second training session there are significant differences from what was seen at 3 h. These data suggest that both training procedures result in similar behavioural and electrophysiological phenotypes at the 3 h mark following the cessation of training. 4.2. Changes at 3 h RPeD1 is the neuron that initiates rhythmogenesis in the CPG that drives aerial respiratory behaviour (Lukowiak, 1991; Syed et al., 1990) so it is unsurprising that the excitability of RPeD1 decreases following training which results in fewer attempted pneumostome openings. Our primary measure of neuronal excitability

C

Fig. 8. Representative recordings of the bursting behaviour of RPeD1 from a semiintact preparation following the LTM-training procedure. The number of bursts decreases at both 3 h and 24 h following training. (A) Naïve (B) 3 h after TS2 and (C) 24 h after TS2.

is to count the number of APs elicited by a series of depolarizing current steps applied to the neurons. The significantly diminished ability of the LTM trained neuron to fire APs over a range of depolarizing currents compared to naive and yoked control preparations occurs concomitantly with a significant decrease in membrane input resistance. This suggests that an alteration in the membrane ion channels has occurred, increasing basal ionic conductance, and in turn, decreasing input resistance. In Aplysia (Kandel, 2001), modulation of ion channels is the first target of increases in cAMP and PKA and in Lymnaea, altering the balance of

Fig. 7. Operant conditioning using the LTM-training procedure (2–0.5 h training sessions (TS) separated by 1 h). Memory tests (MT) at 3 and 24 h post-training showed a significant fall in the number of attempted pneumostome opening. Data are given as mean ± S.E.M. Where p values are not given, there were no significant differences. p < 0.05. (MT (3 h), n = 37; yoked (3 h), n = 15; MT (24 h), n = 37; yoked (24 h), n = 27).

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

411

A

B

Fig. 9. LTM-trained snails show a significant fall in both spikes and bursts following training. (A) Number of spikes/10 min fired by RPeD1 in naïve, trained (3 h and 24 h) and yoked control (3 h and 24 h) semi-intact preparations. (B) The number of bursts fired in 10 min by RPeD1 in naïve, yoked controls and trained snails at 3 and 24 h. Data are given as mean ± S.E.M. Where p values are not given, there were no significant differences (naïve, n = 17; trained (3 h), n = 9; trained (24 h), n = 9; yoked (3 h), n = 8; yoked (24 h), n = 6).

phosphatases and kinases can significantly enhance the ability of the snails to form LTM (Rosenegger & Lukowiak, 2010; Rosenegger, Parvez, & Lukowiak, 2008). However, while modulation of ion channels by both cAMP and PKA occurs during the process of short-term sensitization (Kandel, 2001) the resulting memory only lasts a few minutes. It may be that the longer lasting changes of ITM (3 h) require either the production of more channels or simply different isoforms of a specific channel to change the membrane properties and result in a less excitable neuron. That protein synthesis is required for ITM means the properties of the ion channels could have been modulated via enzymes such as kinases or phosphatases. However, it is also possible that there are newly made additional or different ion transporters which have been inserted into the membrane. Lymnaea synapses have been documented to change the population of receptors they contain (Magoski & Bulloch, 2000), resulting in a switch in the synaptic response. Therefore, the change at ITM (and LTM) may be in the makeup of the channel population as training may initiate a process whereby the dominant response to a particular transmitter is altered. Classical appetitive conditioning of feeding behaviour in Lymnaea appears to involve the paired cerebral giant cells (CGCs). Memory is accompanied by a change in a persistent sodium current in the CGCs, an increase in which results in a depolarized resting membrane potential and an increased excitability (Nikitin

et al., 2008). While increased neuronal excitability is the opposite effect to that which we observed in RPeD1 following operant conditioning, it does show that ion channels are feasible targets in the induction of long-lasting memory in Lymnaea. Thus, this so-called non-synaptic substrate for long-term associative memory (Nikitin et al., 2008) may also occur in the respiratory CPG. While increases in a potassium or decrease in a sodium or calcium conductance should all result in a generalized decrease in neuronal excitability, the fact that membrane resistance simultaneously decreases implies that there is now greater ionic flux across the membrane. Therefore, an increase in the potassium current is the most likely cause. Such an increase in a potassium conductance could create both a decreased membrane input resistance and a decrease in cell excitability (i.e. fewer APs elicited by an injected depolarizing current). In rat and mice CNS neurons, modulation of the small conductance Ca2+-activated K+ (SK) channels are involved in synaptic plasticity and, as a result, memory formation (Brosh, Rosenblum, & Barkai, 2007; Faber, 2009; Faber & Sah, 2007; Stackman et al., 2002). SK channels are thought to affect learning and memory in the mammalian brain via adjustments of intrinsic neuronal properties such as repetitive spike firing frequency as well as overall synaptic transmission. Following learning, the effects of the SK channels are amplified, resulting in an increase in the after-hyperpolarization rate of the post-synaptic potential. Both a delayed rectifying and a calcium-dependent potassium current have been

412

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

A

B

Fig. 10. LTM-trained snails have a significant increase in burst length and # spikes/burst following training.. Comparison of RPeD1 burst properties in naïve, trained (3 h and 24 h) and yoked control (3 h and 24 h) semi-intact preparations. (A) Burst duration (s) (B) # of spikes/burst. Data are given as mean ± S.E.M. p < 0.05 where p values are not given, there were no significant differences.

described in RPeD1 (Sakakibara et al., 2005), either of which could be potential targets of similar modulation. With the decrease in the excitability of the neuron, both the number of action potential bursts and the overall number of APs fired decreased. However, it is the decrease in the number of bursts that is most germane since it is the RPeD1 bursts that result in pneumostome opening (Syed et al., 1990; Syed & Winlow, 1991). Therefore, these changes in the membrane properties of RPeD1 alone could be sufficient to effectively decrease aerial respiratory behaviour. The fact that RPeD1 from snails exhibiting ITM is still competent to initiate breathing when depolarized shows that the synaptic connections between RPeD1 and the motor circuits are intact and functional. 4.3. Changes at 24 h Compared to the changes in electrophysiological properties at 3 h following training, the only significant changes in RPeD1 24 h after the cessation of training are those in preparations that received the LTM-training procedure. Following the LTM-training procedure, the changes in RPeD1 at 24 h have been accentuated such that the resistance, excitability and total number of APs fired are significantly lower than they were at 3 h. This may indicate

that there is nothing more than a time effect occurring: the changes at 3 h were merely snapshots from an ongoing process to decrease the tendency of RPeD1 to initiate pneumostome opening. The continued decrease in a number of neuronal properties indicates that the accentuation of effects that were seen at 3 h must be important for continued maintenance of the memory. The bursting behaviour of RPeD1 is an emergent property of its connection with the other two neurons of the CPG. (Lukowiak, 1991; Syed et al., 1990, 1992). That RPeD1 continues to burst, albeit in an altered manner, strongly suggests that the CPG connections are still there. There are undoubtedly other connections with RPeD1 in the nervous system but it is the interaction of the three CPG neurons which results in the majority of the bursting behaviour in RPeD1. However, the significantly increased length of the burst seen 24 h after the second training session implies that a fundamental change has occurred somewhere in the CPG neural circuit, altering the normal timing. The increased burst length in the semi-intact preparation mirrors the situation noted by Spencer et al. (2002), whereby the length of pneumostome opening in semi-intact preparations significantly increased following LTM training. Along with the changes in burst length, evidence that the similar behavioural phenotypes seen in ITM and LTM are due

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

413

Fig. 11. Neuronal excitability is significantly decreased 3 h and 24 h following LTM training. Excitability of RPeD1 as measured by the number of action potentials fired in response to 10 depolarizing current steps. Comparisons between naïve and trained (3 h and 24 h post-training) snails. Yoked control (3 h and 24 h) snails were not significantly different from naïve snails and are omitted for clarity. Trend lines and their slopes are added to the histograms. Inset: total # of action potentials fired over the ten current steps. Data are given as mean ± S.E.M. p < 0.05.

Fig. 12. LTM-trained snails have a significant decrease in neuronal resistance at 3 h and 24 h following training. Input resistance of RPeD1 calculated from hyperpolarizing current steps. Comparison between naïve, trained (3 h and 24 h post-training) and yoked (3 h and 24 h) controls. Data are given as mean ± S.E.M. p < 0.05. where p values are not given, there were no significant differences.

to fundamentally different neuronal states (i.e. changes in excitability, etc.) comes from the discovery that RPeD1’s ability to drive aerial respiratory behaviour has been lost or severely degraded in

LTM. While RPeD1 24 h after the second training session is still capable of spontaneously firing bursts of AP, these bursts no longer drive (i.e. they are not correlated with) pneumostome opening.

414

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

A

A

B

B

Fig. 13. 24 h following LTM training, bursts of action potentials are no longer correlated with pneuomstome opening. Representative recordings of a single burst of action potentials from (A) a naïve semi-intact preparation and (B) a semi-intact preparation 24 h after the LTM-training procedure.

Spencer et al. (1999), McComb et al. (2005) and Khan and Spencer (2009) observed similar effects and it appears that, despite the importance and necessity of RPeD1 to encoding LTM (Scheibenstock et al., 2002; Sangha, Scheibenstock, McComb, et al., 2003; Sangha, Scheibenstock, & Lukowiak, 2003; Sangha et al., 2005), following LTM training, a reorganization of the nervous system occurs, resulting in the uncoupling of RPeD1 from direct control of the pneumostome. The research to elucidate the causal neuronal basis of LTM has typically been centred on synaptic reorganization (Hotulainen & Hoogengraad, 2010; Kandel, 2001). Synaptic plasticity is probably the main underlying process causing LTM; however as shown here and previously in the literature (e.g. Nikitin et al., 2008) there are also non-synaptic mechanisms that play a role in the mediation of LTM. Here we have concentrated our remarks on the non-synaptic changes that occur in RPeD1 because previously we have shown (e.g. McComb et al., 2005; Spencer et al., 1999, 2002) that there are synaptic changes within the CPG circuit that concomitantly occur with LTM. While translation of already made mRNA in the neurites of a soma-less RPeD1 is sufficient for ITM (Scheibenstock et al., 2002), all models of LTM also require altered gene activity. Increases in kinase activity (e.g. PKC, PKA, MAPK) leading to changes in transcription factors (e.g. CREB) are essential to altering the changes in neurons that constitute the neural basis of memory formation. These changes in transcription factors ultimately lead to changes in protein levels such as have been detected following operant conditioning in Lymnaea (Rosenegger, Wright, & Lukowiak, 2010). Twenty-four hours following ITM training, LTM can be induced in Lymnaea with a single reinforcing stimulus (a poke to the pneumostome) (Parvez et al., 2005). Those results suggested that Lymnaea maintained a ‘‘memory trace’’ following ITM training which

Fig. 14. The ability of RPeD1 depolarization to initiate pneumostome opening is limited following LTM training. (A) Representative trace demonstrating opening of pneumostome following injection of 2.30 nA of current for four seconds (black bar). Closed and open refer to the pneumostome: an upward deflection of the lower trace indicates pneumostome opening and downward deflection indicated pneumostome closing. (B) The percentage of total semi-intact preparations which opened the pneumostome following injection of 2.3 nA of depolarizing current into RPeD1 for 4 s. Comparing naïve, the ITM-training procedure), the LTM-training procedure and yoked snails at 3 h and 24 h. (naïve, n = 17; trained (1TS-3 h) n = 11, trained (2TS3 h) n = 9; trained (1TS-24 h) n = 12, trained (2TS-24 h) n = 9; yoked (2TS-24 h), n = 6). Chi-square results indicate that the data are significantly different.

could be consolidated into a longer, more stable memory when appropriately triggered. This built on the previously expressed hypothesis that ITM and LTM are interconnected at the behavioural level (Smyth, Sangha, & Lukowiak, 2002). The intriguing finding that 24 h following ITM training RPeD1 activity is significantly depressed suggests that the interconnectedness of ITM and LTM extends to the neuronal level as well. Despite no behavioural phenotype of memory, the rates of AP firing and bursting in RPeD1 were significantly decreased when compared to both the naïve and yoked control animals (although not as low as in the LTM-trained animals). The decreased activity in RPeD1 24 h following ITM training may be the electrophysiological evidence of the previously hypothesized ‘‘memory trace’’. It appears that this trace exists in RPeD1 for some time after the behavioural memory, keeping the neuron in a ‘‘primed’’ state whereby an appropriate stimulus can trigger reconsolidation or reintroduction of the previously ‘‘forgotten’’ memory. 4.4. Neuronal vs behavioural: peripheral control Changes in the neuronal properties of the CNS from naïve to 3 h post-training to 24 h post-training presumably contribute to the observed behavioural changes. However, this study indicates that the behavioural memory witnessed at 3 and 24 h, while outwardly identical, have different underlying neuronal substrates. Our data

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

clearly demonstrated that in Lymnaea, while RPeD1 exists in different ‘states’ of excitability, the behavioural phenotypes are quite similar. As decreased aerial respiration is entrained by the operant conditioning procedure, RPeD1 is first modulated by a down regulation of its excitability 3 h following the cessation of training (i.e. when ITM would be expressed with the ITM-training procedure) before a more global suppression of excitability (24 h following the cessation of the LTM-training procedure) maintains the memory in the long-term. With pneumostome opening now independent of RPeD1 activity, there is the possibility that there is an alternative central control of breathing that can initiate pneumostome opening. This implies to us that aerial respiratory behaviour may ultimately be mediated by an interaction of the centrally located CPG and peripherally located neurons in the pneumostome area (e.g. Osphradial ganglion); similar to what occurs in the mediation of siphon and gill withdrawal reflexes in Aplysia (Lukowiak & Colebrook, 1988; Lukowiak & Jacklet, 1972). Furthermore, as RPeD1 possesses both central and peripheral axonal connections (Allison & Benjamin, 1985) it may be mediating central and peripheral signals. In support of this notion of a central-(CNS) peripheral (PNS) nervous system interaction are the observations that pneumostome openings in naïve animals without associated RPeD1 bursts have been reported in the literature (Spencer et al., 2002) and were also seen in these experiments (M. Braun, personal observations). Following LTM training, it may be that the peripheral nervous system is playing a larger role in the control of respiration. In higher vertebrates, the reflexes and sensors within the lungs allow respiration to continue without any central input (Bamford & Jones, 1976) and in mollusks, the peripheral nervous system (PNS) has been shown to possess abilities thought to be confined to the CNS (e.g. learning and memory formation (Lukowiak & Colebrook, 1988; Lukowiak & Jacklet, 1972). Under normal conditions in Lymnaea, peripheral input significantly slows the bursting behaviour of the Lymnaea respiratory CPG (Inoue, Hague, Lukowiak, & Syed, 2001; McComb et al., 2005) and it may be that normal respiratory behaviour is due to the integration of CNS–PNS activity. Following operant conditioning training there is diminished CNS activity (i.e. CPG activity is less) and this may allow the PNS to be more independent and to have a greater role in the control of pneumostome activity. Acknowledgments This study was funded by a Canadian Institute for Health Research grant to K.L. References Allison, P., & Benjamin, P. R. (1985). Anatomical studies of central regeneration of an identified molluscan interneuron. Proceedings of the Royal Society of London Series B, 226, 135–157. Bamford, O. S., & Jones, D. R. (1976). Respiratory and cardiovascular interactions in ducks: The effect of lung denervation on the initiation of and recovery from some cardiovascular responses to submergence. Journal of Physiology, 259(3), 575–596. Brosh, I., Rosenblum, K., & Barkai, E. (2007). Learning-induced modulation of SK channels-mediated effect on synaptic transmission. European Journal of Neuroscience, 26, 3253–3260. Dalesman, S., Braun, M. H., & Lukowiak, K. (2010). Low environmental calcium blocks long term memory formation in a freshwater pulmonate snail. Neurobiology of Learning and Memory, 95(4), 393–403. Dalesman, S., & Lukowiak, K. (2010). Effect of acute exposure to low environmental calcium on respiration and locomotion in Lymnaea stagnalis (L.). Journal of Experimental Biology, 213(9), 1471–1476. Davis, H. P., & Squire, L. R. (1984). Protein synthesis and memory: A review. Psychological Bulletin, 96, 518–559. Faber, E. S. (2009). Functions and modulation of neuronal SK channels. Cell Biochemistry and Biophysics, 55(3), 127–139. Faber, E. S., & Sah, P. (2007). Functions of SK channels in central neurons. Clinical and Experimental Pharmacology and Physiology, 34(10), 1077–1083.

415

Ghirardi, M., Montarolo, P. G., & Kandel, E. R. (1995). A novel intermediate stage in the transition between short- and long-term facilitation in the sensory to motor synapse of Aplysia. Neuron, 9, 479–489. Gibbs, M. E., & Ng, K. T. (1979). Behavioural stages in memory formation. Neuroscience Letters, 13, 279–283. Hotulainen, P., & Hoogengraad, C. C. (2010). Actin in dendritic spines: Connecting dynamics to function. Journal of Cell Biology, 189(4), 619–629. Inoue, T., Hague, Z., Lukowiak, K., & Syed, N. I. (2001). Hypoxia-induced respiratory patterned activity in Lymnaea originates at the periphery. Journal of Neurophysiology, 86(1), 156–163. Inoue, T., Takasaki, M., Lukowiak, K., & Syed, N. (1996). Inhibition of the respiratory pattern-generating neurons by an identified whole-body withdrawal interneuron of Lymnaea stagnalis. Journal of Experimental Biology, 199(9), 1887–1898. Kandel, E. R. (2001). The molecular biology of memory storage: A dialogue between genes and synapses. Science, 294, 1030–1038. Khan, M. A., & Spencer, G. E. (2009). Novel neural correlates of operant conditioning in normal and differentially reared Lymnaea. Journal of Experimental Biology, 212, 922–933. Li, W., Tully, T., & Kalderon, D. (1996). Effects of a conditional Drosophila PKA mutant on olfactory learning and memory. Learning & Memory, 2, 320–333. Lukowiak, K. (1991). Experimental reconstruction of neuronal pattern generators. Current Opinion in Neurobiology, 1(4), 577–582. Lukowiak, K., Adatia, N., Krygier, D., & Syed, N. (2000). Operant conditioning in Lymnaea: Evidence for intermediate- and long-term memory. Learning & Memory, 7, 140–150. Lukowiak, K., & Colebrook, E. (1988). Neuronal mechanisms of learning in an in vitro Aplysia preparation: Sites other than the sensory-motor neuron synapse are involved. Journal of Physiology (Paris), 83(3), 198–206. Lukowiak, K., Cotter, R., Westly, J., Ringseis, E., & Spencer, G. (1998). Long-term memory of an operantly conditioned respiratory behaviour pattern in Lymnaea stagnalis. Journal of Experimental Biology, 201(6), 877–882. Lukowiak, K., & Jacklet, J. W. (1972). Habituation and dishabituation: Interactions between peripheral and central nervous systems in Aplysia. Science, 178, 1306–1308. Lukowiak, K., Orr, M., de Caigny, P., Lukowiak, K. S., Rosenegger, D., Han, J. I., et al. (2010). Ecologically relevant stressors modify long-term memory formation in a model system. Behavioural Brain Research, 214(1), 18–24. Lukowiak, K., Ringseis, E., Spencer, G., Wildering, W., & Syed, N. (1996). Operant conditioning of aerial respiratory behaviour in Lymnaea stagnalis. Journal of Experimental Biology, 199, 683–691. Lukowiak, K., Sangha, S., Scheibenstock, A., Parvez, K., McComb, C., Rosenegger, D., et al. (2003). A molluscan model in the search for the engram. Journal of Physiology (Paris), 97(1), 69–76. Magoski, N. S., & Bulloch, A. G. M. (2000). Stability and variability of synapses in the adult molluskan CNS. Journal of Neurobiology, 42, 410–423. Margulies, C., Tully, T., & Dubnau, J. (2005). Deconstructing memory in Drosophila. Current Biology, 15, R700–R713. McComb, C., Meems, R., Syed, N., & Lukowiak, K. (2003). Eletrophysiological differences in the CPG aerial respiratory behavior between juvenile and adult Lymnaea. Journal of Neurophysiology, 90, 983–992. McComb, C., Rosenegger, D., Varshney, N., Kwok, H. Y., & Lukowiak, K. (2005). Operant conditioning of an in vitro CNS-pneumostome preparation of Lymnaea. Neurobiology of Learning and Memory, 84(1), 9–24. Müller, U. (1996). Inhibition of nitric oxide synthesis impairs a distinct form of longterm memory in the honeybee, Apis mellifera. Neuron, 16, 541–549. Nikitin, E. S., Vavoulis, D. V., Kemenes, I., Marra, V., Pirger, Z., Michel, M., et al. (2008). Electrophysiological and behavioral evidence demonstrating that predator detection alters adaptive behaviors in the snail Lymnaea. Journal of Neuroscience, 28, 2726–2734. Orr, M. V., El-Bekai, M., Lui, M., Watson, K., & Lukowiak, K. (2007). Predator detection in Lymnaea stagnalis. Journal of Experimental Biology, 210, 4150– 4158. Orr, M., Hittel, K., & Lukowiak, K. (2009). Different strokes for different folks: Geographically isolated strains of Lymnaea stagnalis only respond to sympatric predators and have different memory forming capabilities. Journal of Experimental Biology, 212, 2237–2247. Orr, M. V., & Lukowiak, K. (2008). Electrophysiological and behavioural evidence demonstrating the predator detection alters adaptive behaviours in the snail Lymnaea. Journal of Neuroscience, 28(11), 2726–2734. Parvez, K., Stewart, O., Sangha, S., & Lukowiak, K. (2005). Boosting intermediateterm into long-term memory. Journal of Experimental Biology, 208, 1525–1536. Rosenegger, D., & Lukowiak, K. (2010). The participation of NMDA receptors, PKC, and MAPK in the formation of memory following operant conditioning in Lymnaea. Molecular Brain, 3, 24. Rosenegger, D., Parvez, K., & Lukowiak, K. (2008). Enhancing memory formation by altering protein phosphorylation balance. Neurobiology of Learning and Memory, 90(3), 544–552. Rosenegger, D., Wright, C., & Lukowiak, K. (2010). A quantitative proteomic analysis of long-term memory. Molecular brain, 3(1), 9. Rosenzweig, M. R., Bennettt, E. L., Colombo, P. J., Lee, D. W., & Serrano, P. A. (1993). Short-term, intermediate-term, and long-term memories. Behavioural Brain Research, 57, 193–198. Sakakibara, M., Okuda, F., Nomura, K., Watanabe, K., Meng, H., Horikoshi, T., et al. (2005). Potassium currents in isolated statocyst neurons and RPeD1 in the pond snail, Lymnaea stagnalis. Journal of Neurophysiology, 94, 3884–3892.

416

M.H. Braun, K. Lukowiak / Neurobiology of Learning and Memory 96 (2011) 403–416

Sangha, S., Scheibenstock, A., & Lukowiak, K. (2003). Reconsolidation of a long-term memory in Lymnaea requires new protein and RNA synthesis and the soma of right pedal dorsal 1. Journal of Neuroscience, 23, 8034–8040. Sangha, S., Scheibenstock, A., McComb, R., & Lukowiak, K. (2003). Intermediate and long-term memories of associative learning are differentially affected by transcription versus translation blockers in Lymnaea. Journal of Experimental Biology, 206(10), 1605–1613. Sangha, S., Varshney, N., Fras, M., Smyth, K., Rosenegger, D., Parvez, K., et al. (2004). Memory, reconsolidation and extinction in Lymnaea require the soma of RPeD1. Advances in Experimental Medicine and Biology, 551, 311–318. Sangha, S., Scheibenstock, A., Martens, K., Varshney, N., Cooke, R., & Lukowiak, K. (2005). Impairing forgetting by preventing new learning and memory. Behavioral Neuroscience, 119(3), 787–796. Scheibenstock, A., Krygier, D., Haque, Z., Syed, N., & Lukowiak, K. (2002). The Soma of RPeD1 must be present for long-term memory formation of associative learning in Lymnaea. Journal of Neurophysiology, 88(4), 1584–1591. Smyth, K., Sangha, S., & Lukowiak, K. (2002). Gone but not forgotten: the lingering effects of intermediate-term memory on the persistence of long-term memory. Journal of Experimental Biology, 205(1), 131–140. Spencer, G. E., Kazmi, M. H., Syed, N. I., & Lukowiak, K. (2002). Changes in the activity of a CPG neuron after the reinforcement of an operantly conditioned behaviour in Lymnaea. Journal of Neurophysiology, 88(4), 1915–1923. Spencer, G. E., Syed, N. I., & Lukowiak, K. (1999). Neural changes after operant conditioning of the aerial respiratory behavior in Lymnaea stagnalis. Journal of Neuroscience, 19(5), 1836–1843. Stackman, R. W., Hammond, R. S., Linardatos, E., Gerlach, A., Maylie, J., Adelman, J. P., et al. (2002). Small conductance Ca2+-activated K+ channels modulate synaptic plasticity and memory encoding. Journal of Neuroscience, 22(23), 10163–10171.

Steidl, S., Rose, J. K., & Rankin, C. H. (2003). Stages of memory in the nematode Caenorhabditis elegans. Behavioral and Cognitive Neuroscience Reviews, 2(1), 3–14. Stough, S., Shobe, J. L., & Carew, T. J. (2006). Intermediate-term processes in memory formation. Current Opinion in Neurobiology, 16, 672–678. Sutton, M. A., Masters, S. E., Bagnall, M. W., & Carew, T. J. (2001). Molecular mechanisms underlying a unique intermediate phase of memory in Aplysia. Neuron, 31, 143–154. Syed, N. I., Bulloch, A. G., & Lukowiak, K. (1990). In vitro reconstruction of the respiratory central pattern generator of the mollusk Lymnaea. Science, 250, 282–285. Syed, N. I., Ridgway, R. L., Lukowiak, K., & Bulloch, A. G. (1992). Transplantation and functional integration of an identified respiratory interneuron in Lymnaea stagnalis. Neuron, 8, 767–774. Syed, N. I., & Winlow, L. (1991). Coordination of locomotor and cardiorespiratory networks of Lymnaea stagnalis by a pair of identified interneurons. Journal of Experimental Biology, 158, 37–62. Tully, T., & Quinn, W. G. (1985). Classical conditioning and retention in normal and mutant Drosophila melanogaster. Journal of Comparative Physiology (A), 157, 263–277. Wan, H., Mackay, B., Iqbal, H., Naskar, S., & Kemenes, G. (2010). Delayed intrinsic activation of an NMDA-independent CaM-kinase II in a critical time window is necessary for late consolidation of an associative memory. Journal of Neuroscience, 30(1), 56–63. Yin, J. C., Wallach, J. S., Del Vecchio, M., Wilder, E. L., Zhou, H., Quinn, W. G., et al. (1994). Induction of a dominant negative CREB transgene specifically blocks long-term memory in Drosophila. Cell, 79, 49–58.