Effects of 5-HT and insulin on learning and memory formation in food-deprived snails

Accepted Manuscript Effects of 5-HT and insulin on learning and memory formation in food-deprived snails Hitoshi Aonuma, Yuki Totani, Mugiho Kaneda, Ryota Nakamura, Takayuki Watanabe, Dai Hatakeyama, Varvara E. Dyakonova, Ken Lukowiak, Etsuro Ito PII: DOI: Reference:

S1074-7427(17)30218-6 https://doi.org/10.1016/j.nlm.2017.12.010 YNLME 6778

To appear in:

Neurobiology of Learning and Memory

Received Date: Revised Date: Accepted Date:

22 July 2017 8 December 2017 29 December 2017

Please cite this article as: Aonuma, H., Totani, Y., Kaneda, M., Nakamura, R., Watanabe, T., Hatakeyama, D., Dyakonova, V.E., Lukowiak, K., Ito, E., Effects of 5-HT and insulin on learning and memory formation in fooddeprived snails, Neurobiology of Learning and Memory (2017), doi: https://doi.org/10.1016/j.nlm.2017.12.010

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Neurobiology of Learning and Memory (NLM-17-183-R2) Research Article

Effects of 5-HT and insulin on learning and memory formation in food-deprived snails

Hitoshi Aonuma 1,2,†, Yuki Totani 3,†, Mugiho Kaneda 4, Ryota Nakamura 3, Takayuki Watanabe 1, Dai Hatakeyama 4, Varvara E. Dyakonova 5, Ken Lukowiak 6 and Etsuro Ito 3,4,7,8,

*

1

Research Center of Mathematics for Social Creativity, Research Institute for Electronic

Science, Hokkaido University, Sapporo, Hokkaido 060-0811, Japan 2

CREST, Japan Science and Technology Agency, Kawaguchi, Saitama 332-0012, Japan

3

Department of Biology, Waseda University, Shinjuku, Tokyo 162-8480, Japan

4

Kagawa School of Pharmaceutical Sciences, Tokushima Bunri University, Sanuki,

Kagawa 769-2193, Japan 5

Institute of Developmental Biology, Russian Academy of Sciences, Moscow 119334,

Russia

1

6

Hotchkiss Brain Institute, University of Calgary, Calgary, AB T2N 4N1, Canada

7

WASEDA Bioscience Research Institute in Singapore, 138667, Singapore

8

Lipid Science and Aging Research Center and Center for Stem Cell Research,

Kaohsiung Medical University, Kaohsiung 80708, Taiwan

†

Equally contributed.

*

Author for correspondence ([email protected])

Etsuro Ito, Department of Biology, Bg. 50 (TWIns), Rm. 01C208, Waseda University, 2-2 Wakamatsucho, Shinjuku, Tokyo 162-8480, Japan

Running title: Serotonin and insulin for CTA in snails

Highlights: 1. One-day food-deprived snails that exhibit the best CTA learning had a low 5-HT content in their CNS. 2. Five-day food-deprived snails, which did not learn, had a high level of 5-HT content in their CNS.

2

3. Insulin rescues the CTA deficit and this may be due to a decrease in the 5-HT content in the CNS.

KEY WORDS: conditioned taste aversion, food-deprivation, feeding, insulin, Lymnaea, serotonin

3

ABSTRACT The pond snail Lymnaea stagnalis learns conditioned taste aversion (CTA) and consolidates it into long-term memory (LTM). How well they learn and form memory depends on the degree of food deprivation. Serotonin (5-HT) plays an important role in mediating feeding, and insulin enhances the memory consolidation process following CTA training. However, the relationship between these two signaling pathways has not been addressed. We measured the 5-HT content in the central nervous system (CNS) of snails subjected to different durations of food deprivation. One-day food-deprived snails, which exhibit the best learning and memory, had the lowest 5-HT content in the CNS, whereas 5-day food-deprived snails, which do not learn, had a high 5-HT content. Immersing 1-day food-deprived snails in 5-HT impaired learning and memory by causing an increase in 5-HT content, and that the injection of insulin into these snails reversed this impairment. We conclude that insulin rescues the CTA deficit and this may be due to a decrease in the 5-HT content in the CNS of Lymnaea.

4

1. Introduction

Serotonin (5-hydroxytryptamine: 5-HT) drives both feeding behavior and food satiety in the pond snail Lymnaea stagnalis (Kemenes et al., 1989; 1990; Hatakeyama et al., 1999; Yamanaka et al., 2000; Yeoman et al., 2008; Kawai et al., 2011; Dyakonova et al., 2015a; Yamagishi et al., 2015). The cerebral giant cells (CGC) in Lymnaea, which are 5-HTergic, play key roles in mediating various aspects of feeding (Yeoman et al., 2008). Interestingly and consistent with the proposed role played by 5-HT in feeding and food satiation in Lymnaea are the findings that the activity of 5-HT-containing neurons in the central nervous system (CNS) are sensitive to the concentration of glucose in the hemolymph (Dyakonova et al., 2015a,b). Thus, the biophysical properties (e.g., membrane potential and spontaneous firing frequency) of the 5-HT containing neurons involved with the mediation of locomotion are altered by changes in hemolymph glucose concentration in vitro. Furthermore, the exogenous application of glucose to the isolated CNS causes a decrease in excitatory input to these 5-HTergic neurons (Dyakonova & Sakharov, 2001; Chistopolsky & Sakharov, 2003).

5-HT has long been proposed to play a key role in the mediation of learning and memory

5

in molluscs. The best known example of this is dishabituation and sensitization of the gill withdrawal reflex in Aplysia is mediated by the release of 5-HT from interneurons (Kandel, 2001). In addition, 5-HT likely plays an important role in the reconsolidation of memory in Aplysia as well as in other forms of neuronal plasticity (Cai et al., 2012; Hawkins, 2013). In the terrestrial snail Helix lucolum, 5-HT signaling alters the defensive conditioning (Andrianov et al., 2015). As regards 5-HT and memory in Lymnaea, 5-HT plays a key role in the enhancement of LTM following operant conditioning of aerial respiration with exposure of the snail to a predator scent (Il-Han et al., 2010). In a taste-aversive learning in Lymnaea, administration of a 5-HT receptor antagonist before retention tests brings about the phenomenon of reversible amnesia (Nikitin & Solntseva, 2013). Acting as a neurotransmitter/neuromodulator, 5-HT in Lymnaea also plays important roles in the mediation of both locomotion and cardiovascular activity (Buckett et al., 1990; Tsyganov & Sakharov, 2000).

Previously, we have also shown that another neuromodulator/neurotransmitter, an insulin-like molecule, plays an important role in conditioned taste aversion (CTA) and its subsequent LTM formation (Murakami et al., 2013a). There is an up-regulation of molluscan insulin-related peptides (MIPs), following CTA training (Azami et al., 2006).

6

In addition, application of MIPs or mammalian insulin to the isolated CNS caused long-lasting changes in synaptic enhancement at sites thought to play key roles in CTA learning and LTM formation. This synaptic enhancement was blocked by an application of an insulin receptor antibody (Murakami et al., 2013a). An injection of the same antibody into snails before CTA training, while not blocking learning, did block CTA-LTM formation (Murakami et al., 2013a). Thus, insulin is thought to function in the long-term maintenance of synaptic enhancement in CTA learning-related circuits. The mechanisms by which insulin in snails acts in the formation of CTA-LTM have also been examined (Hatakeyama et al., 2013; Murakami et al., 2013b; Mita et al., 2014a,b; Kojima et al., 2015).

Insulin plays an important role in cognitive function across species. For example, human clinical studies suggest a link between type 2 diabetes mellitus (i.e., insulin resistance) and cognitive dysfunction (Biessels & Reagan, 2015; Heni et al., 2015; Mainardi et al., 2015). The role played by 5-HT in cognitive functioning is complicated and difficult to parse (González-Burgos & Fesia-Velasco, 2008; Olvera-Cortés et al., 2013; Olivier, 2015; Voigt & Fink, 2015). Even more complicated is the relationship between the insulin signaling pathway and the 5-HT signaling pathway in mammals (Gerozissis,

7

2008). Thus, we posit that it may be easier to determine in Lymnaea how the two systems interact to mediate cognitive processes.

With CTA training, Lymnaea learn to associate their innate response to a substance such as sucrose (i.e., an increase in the number of feeding bites) with an aversive stimulus (e.g., electric shock) that suppresses biting behavior such that following CTA conditioning sucrose now no longer signals an increased feeding response. This CTA learning then is consolidated into CTA-LTM (Kojima et al., 1996; 1998; Ito et al., 1999; 2013; 2015b; Sunada et al., 2017c). To produce CTA in snails, an appetitive stimulus (e.g., sucrose) is used as the conditioned stimulus (CS), and an aversive stimulus (e.g., electric shock) is used as the unconditioned stimulus (US). The US causes the snails immediately inhibit feeding (Kojima et al., 1997; 2001). In the CTA training procedure, the CS is paired with the US. After repeated CS-US temporal pairings, the CS no longer elicits feeding, and this learned CTA response persists for at least a month (Kojima et al., 1996).

Furthermore, competency to learn and form CTA-LTM is dependent on the level of food satiation of the snail. Thus, 1-day food-deprived snails learn and form CTA-LTM better than food-satiated snails or 5-day food-deprived snails (Mita et al., 2014a,b). Correlated

8

with these aforementioned behavioral data are measurements of dopamine and octopamine in the CNS in snails subjected to the different lengths of food deprivation (Aonuma et al., 2016, 2017).

In the present study, we examined the relationship between the 5-HT content and insulin signaling in the snail CNS following CTA training. We found that there was a small range of 5-HT concentrations where feeding occurred. On either end of this range feeding behavior was compromised. In typically competent snails (i.e., 1-day food deprived), poorer CTA-LTM was coincident with altered 5-HT levels. Injection of insulin in these snails recused the snails’ ability to learn and form CTA-LTM.

2. Materials and methods

2.1. Snails with various food-deprivation states Specimens of Lymnaea stagnalis (Linnaeus, 1758) with a 15-25 mm shell obtained from our snail-rearing facility (original stocks from Vrije Universiteit Amsterdam, The Netherlands) were used. The rearing methods have previously been described

9

(Wagatsuma et al., 2004; Sunada et al., 2017b). To examine the effects of food-deprivation on CTA learning and subsequent LTM formation, we employed four cohorts of snails: (1) Snails that were in the act of eating just before the commencement of the CTA training procedure (i.e., Eating snails); (2) Snails that had been given ad libitum access to food but were not eating just prior to the CTA training procedure (Day -1 snails); (3) Snails that had been food-deprived for 1 day (Day 1 snails); and (4) Snails that had been food-deprived for 5 days (Day 5 snails) (Fig. 1A). We have employed similar cohorts in previous studies (Mita et al., 2014a,b; Aonuma et al., 2016; 2017). The day when snails began food deprivation is called Day 0 (Fig. 1A).

2.2. Measurement of 5-HT and its metabolites in the Lymnaea CNS The measurement method of 5-HT and its precursor and catabolites was as described previously (Aonuma & Watanabe, 2012; Aonuma et al., 2016, 2017). Briefly, snails were quickly frozen using liquid N2, and the CNS was dissected out in ice-cold Lymnaea saline. Ten CNSs were collected for each group. Each CNS was homogenized in 50 µl of ice-cold 0.1 M perchloric acid containing 5 ng of N-ω-methyl-5-hydroxytryptamine oxalate (NMET; Sigma-Aldrich, St. Louis, MO, USA) as an internal standard. After centrifugation of the homogenate [0°C, 21500 g (15000 rpm), 30 min], 40 µl of

10

supernatant was collected. We measured the following compounds in each of the four groups: 1) Serotonin (5-HT); 2) a precursor of 5-HT, (5-hydroxytryptophan, 5-HTP); and two catabolites of 5-HT 3) (N-acetylserotonin, Nac-5-HT; and 4) 5-hydroxyindole acetoaldehyde, 5-HIAA) using high-performance liquid chromatography with electrochemical detection (HPLC-ECD; EICOM, Kyoto, Japan). The mobile phase containing 0.18 M chloroacetic acid and 16 µM disodium EDTA was adjusted to pH 3.6 with NaOH. Sodium-1-octanesulfonate at 1.85 mM as an ion-pair reagent and CH3CN at 8.40% (v/v) as an organic modifier were added to the mobile phase solution. The chromatographs were acquired using the computer program PowerChrom (eDAQ Pty, Denistone East, NSW, Australia). The supernatants of samples were injected directly onto the HPLC column. The snails used for the measurement of 5-HT and its related molecules were independent from those used for behavioral experiments but were in a similar food state. This was done to avoid the effects of sucrose application on the 5-HT contents.

2.3. CTA training procedure

The CTA training procedure has been previously described (Hatakeyama et al., 2006; Aonuma et al., 2016; 2017) and is shown in Fig. 1B. All snails were first given a pretest,

11

in which we counted by visual inspection the number of feeding responses (i.e., bites) in a 1-min period in distilled water after 15-s application of 10 mM sucrose (CS). In the CTA training procedure, we paired the CS (10 mM sucrose solution) with the unconditioned stimulus (US, 3-s electric shock). A 12-s interval followed the presentation of the US, because following the 3-s electric shock a 12-s recovery time was needed before snails re-emerge from their shell. The inter-trial interval was 10 min. Snails received 20 paired presentations of the CS-US (see, Ito et al., 2015a). Controls included a backward-conditioned (US-CS) group and a naive group to validate associative learning. For the naive control group, only distilled water was applied to the lips instead of the CS and US. In the post-test sessions, snails were again challenged with the CS, and the number of bites was recorded in the 1-min interval in distilled water after a 15-s application of the CS. The post-tests were performed 10 min, 1 h, 1 day and 1 week after training. In all experiments, after the 10-min post-test, snails were allowed ad libitum access to food. All tests were performed blindly. The behavioral experiments were performed in the morning, because it has been shown that the CTA scores are better in the morning than at other times (Wagatsuma et al., 2004).

2.4. Modulation of 5-HT signaling

12

Serotonin (5-HT) and a precursor of 5-HT (5-HTP) were purchased from Tokyo Chemical Industry (Tokyo, Japan). To increase the 5-HT content of the CNS, we used 5-HTP. This strategy has previously been used in crickets (Ureshi et al., 2002) and molluscs (Dyakonova et al., 2009). 5-HTP is thought to be almost immediately decarboxylated to yield 5-HT (Cooper et al., 2003). In the experiments reported below, we immersed snails in 200 μM 5-HTP. A 5-HT2 receptor antagonist, cinanserin was purchased from Tocris Bioscience, Bristol, UK, and a 5-HT3 receptor antagonist, 3-tropanyl-3,5-dichlorobenzoate (MDL72222) was purchased from Sigma-Aldrich. These two antagonists have been confirmed to function as a 5-HT receptor blocker in the Lymnaea CNS (Kawai et al., 2011). The effective concentrations of these two antagonists were 10 μM and 100 μM, respectively, in isolated CNS preparations (Kawai et al., 2011). Here we applied these chemicals by transdermal application, that is, we dissolved them in DW and immersed the snails in these solutions before training (see Shields-Johnson et al., 2013 for the studies using Aplysia). Previous studies have shown that a 100 μM 5-HTP solution had a sufficient effect on an isolated neuron within tens of seconds after application (Dyakonova et al., 2009). Thus, we used 100 μM (cinanserin) and 500 μM (MDL72222) for the snail immersion experiments. MDL72222 was dissolved into DMSO as a stock solution. Because the final concentration of DMSO was under 0.1%, a

13

DMSO control injection was not deemed necessary.

2.5. Modulation of insulin signaling Bovine insulin was purchased from Sigma-Aldrich (St. Louis, MO, USA; Murakami et al., 2013a; Mita et al., 2014b), and 50 mg of insulin was dissolved in 10 μl of 1 M HCl as a stock solution. This stock insulin solution was further diluted with Lymnaea saline, and the diluted solution of 40 μl was injected into the abdominal body cavity, whose estimated volume was 400 μl. The final concentration of insulin was estimated as 100 nM (Murakami et al., 2013; Mita et al., 2014b). The control experiments were performed by injection of the vehicle (i.e., HCl), whose estimated final concentration in Lymnaea saline was 1.1 μM. Lymnaea saline consisted of NaCl 50 mM, KCl 1.6 mM, MgCl2 2.0 mM, CaCl2 3.5 mM, and HEPES 10 mM (pH 7.9). To block the action of insulin, we used the mouse monoclonal [47-9] antibody to human insulin receptor alpha subunit (ab982, Abcam, Cambridge, UK), whose final concentration in the body was estimated as 17.5 nM (Murakami et al., 2013a; Mita et al., 2014b). The validity of a use of bovine insulin and human insulin receptor antibody in Lymnaea was described in the previous studies (Murakami et al., 2013a; Mita et al., 2014b). Briefly, the binding site of Lymnaea MIP receptors is well conserved across phyla. For example, in comparison with mammalian

14

insulin receptors, the homology is 34% for the whole amino acid sequences; 56% for the ligand-binding domain 1 (L1 domain); and 33% for the ligand-binding domain 2 (L2 domain) between Lymnaea, humans and mice. In addition, previous studies using another mollusc, Aplysia, have also demonstrated that application of bovine insulin activated the bag cell neuron insulin receptor by stimulating its autophosphorylation on tyrosine residues (Jonas et al., 1996) and evoked egg-laying hormone secretion (Jonas et al., 1997). The control experiments were performed by injection of mouse normal IgG (whole molecule; Jackson ImmunoResearch Laboratories, PA, USA; 17.5 nM as its final concentration in Lymnaea saline; Murakami et al., 2013a; Mita et al., 2014b). The injection procedures of the insulin and insulin receptor antibody were the same as described earlier (Mita et al., 2014b). That is, the snails that were injected with these drugs were kept in an upside-down position without immersion of pond water for 1 h.

2.6. Statistics The data are expressed as the mean ± SEM. Both one-way ANOVA and where appropriate two-way ANOVA were performed, and significant differences at P < 0.05 were examined. Where significance is shown the appropriate post-hoc test was performed; either a Tukey’s test or Scheffé’s test.

15

3. Results

3.1. Contents of 5-HT and related molecules in the CNS Four separate cohorts of snails in four different satiated/hungry states were used (Fig. 1A): (1) Snails that were in the act of eating just before the commencement of the 5-HT measurement (i.e., Eating snails); (2) Snails that had been given ad libitum access to food but were not eating just prior to the measurement (Day -1 snails); (3) Snails that had been food-deprived for 1 day (Day 1 snails); and (4) Snails that had been food-deprived for 5 days (Day 5 snails). We measured the contents of 5-HTP, 5-HT, Nac-5-HT and 5-HIAA in the CNS of these snails (Fig. 2). 5-HTP is the precursor of 5-HT, and both Nac-5-HT and 5-HIAA are the catabolites of 5-HT (Aonuma & Watanabe, 2012).

The 5-HTP content in the CNS of Day -1 snails was higher than those of Eating, Day 1 and Day 5 snails (One-way ANOVA, F(3, 25) = 6.57, followed by a Scheffé’s post-hoc test, P < 0.01; Fig. 2A). No significant difference was found between the 5-HTP content in the CNS of Day 1 snails and that of Day 5 snails (P > 0.05; Fig. 2A). Next, the 5-HT

16

content in the CNS of Day 1 snails, which were food-deprived for 1 day, was lower than those of Eating, Day -1 and Day 5 snails (One-way ANOVA, F(3, 35) = 12.40, followed by a Scheffé’s post-hoc test, P < 0.01; Fig. 2B). The 5-HT content of Day 5 snails returned to almost the same level as the Eating and Day -1 snails (Fig. 2B). The Nac-5-HT content in the CNS of Eating snails was higher than those of Day 1 and Day 5 snails (One-way ANOVA, F(3, 34) = 4.01, followed by a Scheffé’s post-hoc test, P < 0.05; Fig. 2C). No significant difference was found between the Nac-5-HT content in the CNS of Day 1 snails and that of Day 5 snails (P > 0.05; Fig. 2C). About the 5-HIAA content, although there was a tendency for changes similar to that seen for 5-HT content (Fig. 2B), no significant differences were observed among the 4 cohorts (One-way ANOVA, F(3, 35) = 2.33, followed by a Scheffé’s post-hoc test, P > 0.05; Fig. 2D).

To more easily compare the changes in the CNS contents of 5-HTP, 5-HT, Nac-5-HT and 5-HIAA, the mean values of the data were normalized to that of Eating snails (Fig. 2E). These data show that the CNS of Day 1 snails have the lowest relative contents of the measured molecules; whereas the contents in the CNS of Day 5 snails returned to the levels seen in food-satiated snails (i.e., Day -1). The only exception was the 5-HTP content.

17

3.2. CTA learning and memory We examined the ability to successfully learn and form CTA-LTM in the different cohorts of food-deprived snails. The results of the CTA learning and LTM formation for the 4 cohorts of snails (Eating, Day -1, Day 1 and Day 5 snails), were almost the same as those previously reported (Aonuma et al., 2016; 2017). That is, the best CTA learning and LTM were seen in Day 1 snails (i.e., snails food-deprived for 1 day).

In all 4 cohorts of snails, we performed CTA training, as well as two control experiments (i.e., backward conditioning and naive training) (Fig. 3). In all these groups, no significant differences were found at the pretests (one-way ANOVA, F(11, 228) = 1.32, P > 0.05). The two control groups showed neither learning nor LTM formation in all 4 cohorts. Snails trained that were in the act of eating (i.e., Eating snails) exhibited poor learning and memory (Fig. 3A). That is, there was only a significant difference in these snails to their pre-test response in the 1-week post-test session (one-way ANOVA, F(14, 285) = 5.17, followed by a Tukey’s post-hoc test, P < 0.01). Moreover, their performance did not differ from the two control groups with the exception of the 1 h post-test session, and then only with the naive group. Their overall performance was poorer than Day -1

18

and Day 1 snails and about equal to Day 5 snails.

The next cohort of snails tested was Day -1 cohort. In these snails following training, the response to the CS was significantly suppressed compared to the pretest response in all post-test sessions (one-way ANOVA, F(14, 285) = 17.26, followed by a Tukey’s post-hoc test, P < 0.01, Fig. 3B). Thus, these snails learned and formed CTA memory. We then examined the Day 1 cohort. As expected from previous results (Mita et al., 2014a,b; Aonuma et al., 2016; 2017), the snails exhibited excellent learning and memory formation in all the post-test sessions (one-way ANOVA, F(14, 285) = 63.58, followed by a Tukey’s post-hoc test, P < 0.01, Fig. 3C). As can be seen, the suppression of the feeding response to the sucrose CS in the post-tests in the Day 1 snails is greater than that in Day -1 snails.

Finally, we examined the Day 5 cohort. We found that there was no significant difference in the post-test sessions compared to the pretest sessions indicating that neither learning occurred nor memory formed. In these snails, the feeding response elicited by the CS in the CTA-trained snails was suppressed (10 min and 1 week post-tests) in comparison with that in the backward-conditioned snails (one-way ANOVA, F(14, 285) = 3.62, P < 0.01,

19

followed by a Tukey’s post-hoc test, P < 0.05, Fig. 3D). However, because no significant difference was found between the feeding response in the CTA-trained snails and that in the naive control snails (Tukey’s post-hoc test, P > 0.05, Fig. 3D), we conclude that learning and memory formation did not occur in these snails. Our conclusion, therefore, is that Day 1 snails exhibit the best CTA-LTM, whereas Day 5 snails, which are more severely food-deprived, show poorer learning and memory.

3.3. Changes in feeding responses to CS after changes in 5-HT levels in Day 1 and Day 5 snails Our 5-HT measurements showed that Day 1 snails had the lowest 5-HT content, whereas Day 5 snails had a higher level of 5-HT similar to the levels in the satiated snails (Eating and Day -1 snails) (Fig. 2B). As was just seen, Day 1 snails exhibited the best learning and memory, whereas Day 5 snails did not (Fig. 3C,D). Here, we attempted to determine if the 5-HT content of the CNS altered the ability of snails to learn and form CTA memory. We did this by manipulating 5-HT signaling in the CNS by using both 5-HT receptor agonists and antagonists in Day 1 and Day 5 snails. Specifically, we tested whether bathing Day 1 snails in a solution that increases the 5-HT CNS content caused them to perform poorer, whereas bathing Day 5 snails in a 5-HT receptor antagonist medium

20

resulted in better memory formation.

Day 1 snails were immersed in 200 μM 5-HTP for 1 h (Fig. 4A). We did not inject 5-HTP, because it takes some time for snails to sufficiently recover for experiments involving feeding behavior (Kemenes et al., 1989). In Figure 4A, we have plotted the percent change in feeding in Day 1 snails following immersion in the 5-HTP bath. As can be seen, there was a significant depression in the number of bites in response to an application of 10 mM sucrose solution. The suppression of the feeding response continued for 3 h after immersion (two-way ANOVA; snails: F(17, 85) = 2.96, P < 0.01; days: F(5, 85) = 4.58, P < 0.01; followed by a Tukey’s post-hoc test, P < 0.01, Fig. 4A). Five hours later, the feeding response to sucrose returned to the basal before 5-HTP application (P > 0.05). Thus, in Day 1 snails, increasing the 5-HT content of the CNS depressed the feeding response elicited by sucrose.

We next examined the food response to a 10 mM sucrose solution (i.e., the CS) after immersion of Day 1 snails in a 5-HT bath for 24 h (Fig. 4B). The feeding response elicited by the CS did not change when snails were immersed in either a 5 μM solution or a 10 μM solution of 5-HT. However, as the concentration of 5-HT was increased, the feeding

21

response to the CS decreased in a dose dependent manner (one-way ANOVA, F(4, 65) = 75.13, followed by a Scheffé’s post-hoc test, P < 0.01, Fig. 4B). Thus, as the 5-HT content of the CNS increased, there was a suppression of the feeding response to the sucrose CS similar to the decrease feeding observed with the 5-HTP bath.

We then employed the 5-HT receptor antagonists, cinanserin (5-HT2) and MDL 72222 (5-HT3) in Day 1 snails (Fig. 4C). Previous studies showed that both metabotropic serotonin (5-HT2) and ionotropic (5-HT3) receptors are present in the synapses between the serotoninergic cerebral giant cell (CGC) and a follower motor neuron (the B1 cell) (Kawai et al., 2011). Moreover, the CGC is a key neuron for CTA-LTM (Otsuka et al., 2013; Sunada et al., 2017a). Simultaneous application of cinanserin and MDL 72222 (100 μM and 500 μM, respectively) to these snails suppressed the feeding response elicited by the sucrose CS (one-way ANOVA, F(2, 57) = 191.90, followed by a Tukey’s post-hoc test, P < 0.01, Fig. 4C), whereas the results by 100 μM and 500 μM application were almost same (P > 0.05).

3.4. Effects of 5-HT on CTA We hypothesized that Day 1 snails, which exhibit the best learning and memory, when

22

immersed in 5-HT would show impaired CTA learning and memory (Fig. 3C). We arrived at this hypothesis because the snails with a high 5-HT CNS content, such as Eating, Day -1 and Day 5 snails, demonstrated poorer or no CTA-LTM (Fig. 3A,B,D). To test this hypothesis, we immersed the Day 1 snails in a 300 μM 5-HT solution for 24 h, during which time snails did not have access to food (Fig. 5). That is, these snails are referred to as ‘Day 1 snails with a 300 μM 5-HT solution’. In these snails neither learning nor CTA memory formation was observed (one-way ANOVA, F(4, 45) = 0.12, followed by a Tukey’s post-hoc test, P > 0.05, Fig. 5).

3.5. Effects of insulin on 5-HT contents in the CNS The Ito laboratory previously has shown a key role of insulin in CTA-LTM formation (Hatakeyama et al., 2013; Murakami et al., 2013a,b; Mita et al., 2014a; Kojima et al., 2015). In the course of forming CTA-LTM, some molluscan insulin-related peptides (MIPs) are up-regulated (Azami et al., 2006). When MIPs or mammalian insulin was superfused over an isolated CNS, a long-term change in synaptic efficacy (i.e., enhancement) was observed (Murakami et al., 2013a). When mammalian insulin was injected directly into Day 5 snails, learning and CTA-LTM formation now occurred (Mita et al., 2014b). That is, the injected insulin allowed these snails to learn and form

23

CTA-LTM. On the other hand, injection of an insulin receptor antibody into the Day 1 snails obstructed the memory consolidation process (Murakami et al., 2013a).

We thus examined the effects of the injection of insulin and an insulin receptor antibody on the CNS 5-HT contents of Day 1 and Day 5 snails (Fig. 6). In Day 1 snails, application of the 4 drugs used (i.e., 100 nM insulin, its vehicle of 1.1 μM HCl, 17.5 nM insulin receptor antibody, and its control of 17.5 nM mouse normal IgG) did not alter the 5-HT content (one-way ANOVA, F(3, 36) = 2.14, followed by a Tukey’s post-hoc test, P > 0.05; Fig. 6A). In Day 5 snails, the application of 100 nM insulin decreased the 5-HT content of the CNS (P < 0.01, see Fig. 6B and Fig. 2B). This seemed convincing because our previous study showed that the insulin injection into Day 5 snails rescued the poor CTA learning and memory (Mita et al., 2014b). However, we did not find any differences in 5-HT contents in the cases of the HCl vehicle-, the insulin receptor antibody- or the IgG-injected Day 5 snails from that in the insulin-injected Day 5 snails (one-way ANOVA, F(3, 36) = 1.21, followed by a Tukey’s post-hoc test, P > 0.05, Fig. 6B).

3.6. Rescue of poor CTA scores brought by 5-HT by insulin injection

24

The data presented in Figure 5 showed that Day 1 snails immersed in 5-HT for 24 h did not exhibit learning nor CTA-LTM. This result lead us to ask whether an injection of insulin into snails treated in this manner would confer on them the ability to learn and form CTA-LTM; just as previously it was shown that insulin conferred the ability to learn and form memory in Day 5 snails (Mita et al., 2014b). We prepared the following 4 groups of Day 1 snails: 1) Snails immersed in a 300 μM 5-HT solution for 24 h, during which time snails did not have access to food, and then injected the snails with a 100 nM insulin 1 h before CTA training; 2) Snails immersed in the same concentration of 5-HT solution for 24 h, but injected with a vehicle, HCl, 1 h before CTA training; 3) Snails immersed in the same concentration of 5-HT solution for 24 h without any insulin injection and then trained in the backward US-CS pairings; and 4) Snails immersed in the same 5-HT solution for 24 h without any insulin injection, but then trained in the naive control protocol.

As shown in Figure 7, no significant differences were found in the pretest among all the 4 groups. Only Day 1 snails immersed in the 300 μM 5-HT solution and injected with the 100 nM insulin solution exhibited both CTA learning and CTA-LTM formation (one-way ANOVA, F(19, 176) = 3.43, followed by a Tukey’s post-hoc test, P < 0.01, Fig. 7). None

25

of the 3 other groups of snails exhibited CTA learning or memory formation. Thus, we conclude the following two points: 1) increasing 5-HT levels in Day 1 snails obstructs CTA learning and memory; and 2) insulin rescues the ability of these snails to learn and form CTA-LTM in those snails.

4. Discussion

Day 1 snails that exhibit the best CTA learning and memory formation were found to have the lowest CNS contents of 5-HTP, 5-HT, Nac-5-HT and 5-HIAA (Figs. 2 and 3). Moreover, when the CNS 5-HT content was increased in the Day 1 snails, their response to the CS was decreased and they exhibited poorer learning and memory formation (Figs. 4B and 7). Further, when the 5-HT signaling cascade was blocked by the two 5-HT receptor antagonists, the feeding response was also decreased. These results indicate that the feeding response elicited by the sucrose CS is controlled by a narrow window concentration of 5-HT in the CNS and that a high 5-HT content results in poorer learning and memory formation following CTA training. Finally, insulin injection was able to reverse the learning and memory deficit produced by immersing Day 1 snails in 5-HT that

26

increases the 5-HT content in these snails.

Previously, it was found that the 5-HT content in both the buccal and pedal ganglia were higher after 12-h or 48-h food deprivation than those in the control food satiated snails (Dyakonova et al., 2015a). Here we used the whole CNS for the measurement of the 5-HT content. We did this as we also measured both the 5-HT precursor and two 5-HT metabolites that occur in tiny amounts in the CNS (Fig. 2). Changes in CNS 5-HT content have previously been correlated with acute stress responses occurring with termination of feeding as a result of avoidance behavior in another species of mollusc (Gillette, 2006). In the future, we will attempt to determine if there are similar changes in the CNS 5-HT content as a result of differing stressors that alter learning and memory of aerial respiratory behavior in Lymnaea (Hughes et al., 2016; 2017). It should be noted that while we showed that a 1-h bath application of 5-HTP to snails decreased the feeding response to sucrose (Fig. 4A), Dyakonova and Sakharov had previously demonstrated an increase in the amount of consumed food after 12 h of 5-HTP incubation (Dyakonova and Sakharov, 1994). This contradiction may occur by the difference of incubation time. Recent studies have shown that different periods of 5-HTP incubation time induced different transcriptional effects in Lymnaea (Benatti et al., 2017). We intend in the future

27

to examine the effects of 5-HTP on the feeding response with differing 5-HTP incubation times.

Recently, it was shown in Day 1 snails that with soma ablation of the 5-HTergic CGCs CTA learning occurs but CTA-LTM formation does not occur (Sunada et al., 2017a). Presumably, this occurs because LTM formation requires altered gene activity and removal of the soma obviously prevents that from occurring. This has previously been shown in Lymnaea (Scheibenstock et al., 2002). Thus, it was assumed that 5-HT played a critical role in CTA-LTM formation. However, as shown here the best CTA learning and memory formation occurs in Day1 snails that have the lowest CNS 5-HT content (Fig. 2). Moreover, as we also showed here increasing 5-HT levels in Day 1 snails caused these snails that typically have the best learning and memory to not be able to learn or form LTM (Fig. 5). The assumption that could be drawn for those experiments was that 5-HT levels were required to be low in order for learning and LTM to occur. However, we further found that learning and memory formation could be rescued in Day 1 snails that had been immersed in 5-HT by the injection of insulin. In those snails, learning and memory were again seen (Fig. 7).

28

However, the situation is more complicated than just the content of 5-HT and memory formation. When we injected insulin into Day 5 snails, the 5-HT content was significantly decreased (Figs. 2B and 6B). This is convincing because we have previously shown that an insulin injection rescued the ability of the Day 5 snails to learn CTA and form LTM (Mita et al., 2014b, Ito et al., 2017). However, the injection of other insulin-related drugs also caused the decrease in 5-HT content in Day 5 snails (Fig. 6B). If only insulin injection in the Day 5 snails had caused the 5-HT content to decrease to the level seen in Day 1 snails, the explanation as to why insulin injection allowed these snails to learn and remember would have been relatively straightforward. As we showed with Day 1 snails that were immersed in 5-HT to cause an increase in 5-HT and a diminution of their ability to learn and remember that was subsequently rescued by insulin injection, the interaction between insulin and 5-HT in mediating cognitive ability is complicated and at this time unknown.

We also do not fully understand how the CNS 5-HT content is regulated, especially the role and function of the 5-HT transporters (Sadamoto et al., 2008). It is feasible that an alteration of this transporter occurs during memory formation following CTA training in Lymnaea. This possibility has been studied in mammalian systems (Meneses, 2013). For

29

example, the amygdala, which is a key player in the processing of fear, is prominently modulated by 5-HT. The genetic and pharmacological alterations of the 5-HT transporter alter amygdala activation in response to emotional stimuli (Boccio et al., 2016).

The results obtained here raise a number of questions that must be answered. Firstly, there were differences seen in the response of Day 1 snails to the sucrose CS. That is, the initial response to the CS before CTA training (i.e., the pre-test response) was different (e.g., the pretest sessions of Figs. 3C, 5 and 7). An explanation may be that the experiments were performed at different times during the day. We have previously shown a circadian effect on learning and memory in Lymnaea with CTA training (Wagatsuma et al., 2004). In addition, experiments were performed in differing months of the year and this has been noted both in peer-reviewed papers (Ivashkin et al., 2015) and in anecdotal reports from other Lymnaea researchers. Thus, as we did here it is necessary to directly compare the feeding behavior between CTA trained and control snails directly.

Secondly, there are important differences in the training procedures used here and in the study of Murakami et al., (2013a). In Murakami’s study, the human insulin receptor antibody was injected into Day 1 snails before CTA training, but here we injected insulin

30

into Day1 snails that had been immersed in 5-HT. As shown, immersing Day 1 snails in 5-HT diminished their ability for CTA learning and memory. However, the injection of insulin into these snails confers upon them the ability to exhibit learning and memory with CTA training. Here we also did not inject the human insulin antibody. Further, in the earlier study (Murakami et al., 2013a), the injection of the insulin antibody had no effect on the 10-min post-test but did obstruct the response in the 1h post-test. As shown (Murakami et al., 2013a), the injection of the insulin did not obstruct learning but did block LTM formation. Here we showed that insulin rescues a learning and memory deficit produced by increased levels of 5-HT in the Day 1 snails.

Thirdly, Lymnaea possess an open circulatory system. Thus, the injection of a substance not only can affect the CNS but can also affect the peripheral nervous system as well as other organs. For example, the injection of 5-HT directly affects the cardiovascular system (i.e., rate and force of the heart) (Yamagishi et al., 2015). Insulin injection also affects the hemolymph glucose in molluscs (Horn et al., 1998). Thus, we believe it is necessary to always measure and compare responses with backward conditioning and naive controls groups.

31

Finally, we are cognizant that we must develop methods to measure the insulin content of the CNS in Lymnaea in order to be able to correlate the various roles this molecule plays in maintaining homeostatic functions in Lymnaea as well as playing a key role in how and when memories are formed. To this end we have developed an ultrasensitive measurement for a trace amount of human insulin (Watabe et al., 2014, 2016; Ito et al., 2015c). However, we need to develop suitable antibodies for Lymnaea insulin (i.e., MIPs) for this purpose. In the near future, we hope to be able to measure MIP levels after application of 5-HT to Lymnaea and determine their downstream targets in the Lymnaea CNS.

Competing interests The authors declare no competing or financial interests.

Author contributions H.A. and E.I. planned the experiments. V.E.D. and K.L. helped to design the experiments. H.A., Y.T., M.K., R.N. and E.I. performed the experiments. H.A., Y.T., M.K., W.T., D.H. and E.I. analyzed the data. Y.T. prepared the figures. H.A., W.T., D.H., V.E.D., K.L. and

32

E.I. wrote the paper.

Funding This work was supported by a grant from Network Joint Research Center for Materials and Devices in Japan [2015001 to H.A. and E.I.], KAKENHI grants from the Japanese Society for the Promotion of Science [24657055 and 25291074 to E.I.], Waseda University grants for Specific Research Projects in Japan [2016B-068, 2016B-069 and 2016S-037 to E.I.], and a grant from the Natural Sciences and Engineering Research Council of Canada [227993-2013 to K.L.]

33

References

Andrianov, V. V., Bogodvid, T. K., Deryabina, I. B., Golovchenko, A. N., Muranova, L. N, Tagirova, R. R., Vinarskaya, A. K., & Gainutdinov, K. L. (2015). Modulation of defensive reflex conditioning in snails by serotonin. Frontiers in Behavioral Neuroscience, 9, 279. http://dx.doi.org/10.3389/fnbeh.2015.00279. Aonuma, H., & Watanabe, T. (2012). Changes in the content of brain biogenic amine associated with early colony establishment in the Queen of the ant, Formica japonica. PLoS One, 7, e43377. http://dx.doi.org/10.1371/journal.pone.0043377. Aonuma, H., Kaneda, M., Hatakeyama, D., Watanabe, Y., Lukowiak, K., & Ito, E. (2016). Relationship between the grades of a learned aversive-feeding response and the dopamine contents in Lymnaea. Biology Open, 5, 1869-1873. http://dx.doi.org/10.1242/bio.021634. Aonuma, H., Kaneda, M., Hatakeyama, D., Watanabe, Y., Lukowiak, K., & Ito, E. (2017). Weak involvement of octopamine in aversive taste learning in a snail. Neurobiology of Learning and Memory, 141, 189-198. http://dx.doi.org/10.1016/j.nlm.2017.04.010. Azami, S., Wagatsuma, A., Sadamoto, H., Hatakeyama, D., Usami, T., Fujie, M.,

34

Koyanagi, R., Azumi, K., Fujito, Y., Lukowiak, K., & Ito E. (2006). Altered gene activity correlated with long-term memory formation of conditioned taste aversion in Lymnaea. Journal of Neuroscience Research, 84, 1610-1620. http://dx.doi.org/10.1002/jnr.21045. Benatti, C., Colliva, C., Blom, JMC, Ottaviani, E., & Tascedda, F. Transcriptional effect of serotonin in the ganglia of Lymnaea stagnalis. (2017). Invertebrate Survival Journal, 14, 251-258. Biessels, G. J., & Reagan, L. P. Hippocampal insulin resistance and cognitive dysfunction. (2015). Nature Reviewers Neuroscience, 16, 660-671. http://dx.doi.org/10.1038/nrn4019. Bocchio, M., McHugh, S. B., Bannerman, D. M., Sharp, T., & Capogna, M. (2016). Serotonin, amygdala and fear: assembling the puzzle. Frontiers in Neural Circuits, 10, 24. http://dx.doi.org/10.3389/fncir.2016.00024. Buckett, K. J., Dockray, G. J., Osborne, N. N., & Benjamin, P. R. (1990). Pharmacology of the myogenic heart of the pond snail Lymnaea stagnalis. Journal of Neurophysiology, 63, 1413-1425. Cai, D., Pearce, K., Chenm S., & Glanzman, D. L. (2012). Reconsolidation of long-term memory in Aplysia. Current Biology, 22, 1783-1788.

35

http://dx.doi.org/10.1016/j.cub.2012.07.038. Chistopolsky, I. A., & Sakharov, D. A. (2003). Non-synaptic integration of the cell bodies of neurons into the central nervous system of the snail. Neuroscience and Behavioral Physiology, 33, 295-300. http://dx.doi.org/10.1023/A:1022163701311. Cooper, J. R., Bloom, F. E., & Roth, R. H. (2003). The Biochemical Basis of Neuropharmacology, 8th Ed, Oxford University Press, Oxford. Dyakonova, V. E., & Sakharov, D. A. (1994). Involvement of endogenous opioid system in control of feeding and defensive behaviour in the mollusc Lymnaea stagnalis. Zhurnal Vysshei Nervnoi Deyatelnosti Imeni I. P. Pavlova, 44, 316-322. [in Russian]. Dyakonova, V. E., & Sakharov, D. A. (2001). An isolated serotonergic neuron: the mechanism of excitation induced by enhanced synthesis of neurotransmitter. Doklady Biological Sciences, 378, 230-232. http://dx.doi.org/10.1023/A:1019258523049. Dyakonova, V. E., Chistopolsky, I. A., Dyakonova, T. L., Vorontsov, D. D., Sakharov, D. A. (2009). Direct and decarboxylation-dependent effects of neurotransmitter precursors on firing of isolated monoaminergic neurons. Journal of Comparative Physiology A, 195, 515-527. http://dx.doi.org/10.1007/s00359-009-0428-5.

36

Dyakonova, V. E., Hernádi, L., Ito, E., Dyakonova, T. L., Chistopolsky, I. A., Zakharov, I. S., & Sakharov, D. A. (2015a). The activity of isolated neurons and the modulatory state of an isolated nervous system represent a recent behavioural state. Journal of Experimental Biology, 218, 1151-1158. http://dx.doi.org/10.1242/jeb.111930. Dyakonova, V., Hernádi, L., Ito, E., Dyakonova, T., Zakharov, I., & Sakharov, D. (2015b). The activity of isolated snail neurons controlling locomotion is affected by glucose. BIOPHYSICS, 11, 55-60. http://dx.doi.org/10.2142/biophysics.11.55. Gerozissis, K. (2008). Brain insulin, energy and glucose homeostasis; genes, environment and metabolic pathologies. European Journal of Pharmacology, 585, 38-49. http://dx.doi.org/10.1016/j.ejphar.2008.01.050. Gillette, R. (2006). Evolution and function in serotonergic systems. Integrative and Comparative Biology, 46, 838-846. http://dx.doi.org/10.1093/icb/icl024. González-Burgos, I., & Feria-Velasco, A. (2008). Serotonin/dopamine interaction in memory formation. Progress in Brain Research, 172, 603-623. http://dx.doi.org/10.1016/S0079-6123(08)00928-X. Hatakeyama, D., & Ito, E. (1999). Three-dimensional reconstruction and mapping of serotonin-like immunoreactive neurons in the central nervous system of the pond

37

snail, Lymnaea stagnalis, with the confocal laser scanning microscope. Bioimages, 7, 1-12. Hatakeyama, D., Sadamoto, H., Watanabe, T., Wagatsuma, A., Kobayashi, S., Fujito, Y., Yamashita, M., Sakakibara, M., Kemenes, G., & Ito, E. (2006). Requirement of new protein synthesis of a transcription factor for memory consolidation: paradoxical changes in mRNA and protein levels of C/EBP. Journal of Molecular Biology, 356, 569-577. http://dx.doi.org/10.1016/j.jmb.2005.12.009. Hatakeyama, D., Okuta, A., Otsuka, E., Lukowiak, K., & Ito, E. (2013). Consolidation of long-term memory by insulin in Lymnaea is not brought about by changing the number of insulin receptors. Communicative & Integrative Biology, 6, e23955. http://dx.doi.org/10.4161/cib.23955. Hawkins, R. D. (2013). Possible contributions of a novel form of synaptic plasticity in Aplysia to reward, memory, and their dysfunctions in mammalian brain. Learning & Memory, 20, 580-591. http://dx.doi.org/10.1101/lm.031237.113. Heni, M., Kullmann, S., Preissl, H., Fritsche, A., & Häring, H. U. (2015). Impaired insulin action in the human brain: causes and metabolic consequences. Nature Reviews Endocrinology, 11, 701-711. http://dx.doi.org/10.1038/nrendo.2015.173. Horn, C. C., Koester, J., & Kupfermann, I. (1988). Evidence that hemolymph glucose in

38

Aplysia californica is regulated but does not affect feeding behavior. Behavioral Neuroscience, 112, 1258-1265. Hughes, E., Shymansky, T., Sunada, H., & Lukowiak, K. (2016). Qualitatively different memory states in Lymnaea as shown by differential responses to propranolol. Neurobiology of Learning and Memory, 136, 63-73. http://dx.doi.org/10.1016/j.nlm.2016.09.013. Hughes, E., Shymansky, T., Swinton, E., Lukowiak, K. S., Swinton, C., Sunada, H., Protheroe, A., Phillips, I., & Lukowiak, K. (2017). Strain-specific differences of the effects of stress on memory in Lymnaea. Journal of Experimental Biology, 220, 891-899. http://dx.doi.org/10.1242/jeb.149161. Il-Han, J., Janes, T., & Lukowiak, K. (2010). The role of serotonin in the enhancement of long-term memory resulting from predator detection in Lymnaea. Journal of Experimental Biology, 213, 3603-3614. http://dx.doi.org/10.1242/jeb.048256. Ito, E., Kobayashi, S., Kojima, S., Sadamoto, H., & Hatakeyama, D. (1999). Associative learning in the pond snail, Lymnaea stagnalis. Zoological Science, 16, 711-723. http://dx.doi.org/10.2108/zsj.16.711. Ito, E., Otsuka, E., Hama, N., Aonuma, H., Okada, R., Hatakeyama, D., Fujito, Y., & Kobayashi, S. (2012). Memory trace in feeding neural circuitry underlying

39

conditioned taste aversion in Lymnaea. PLoS ONE, 7, e43151. http://dx.doi.org/10.1371/journal.pone.0043151. Ito, E., Kojima, S., Lukowiak, K., & Sakakibara, M. (2013). From likes to dislikes: conditioned taste aversion in the great pond snail Lymnaea stagnalis. Canadian Journal of Zoology, 91, 405-412. http://dx.doi.org/10.1139/cjz-2012-0292. Ito, E., Yamagishi, M., Takigami, S., Sakakibara, M., Fujito, Y., & Lukowiak, K. (2015a). The Yerkes-Dodson law and appropriate stimuli for conditioned taste aversion in Lymnaea. Journal of Experimental Biology, 218, 336-339. http://dx.doi.org/10.1242/jeb.113266. Ito, E., Yamagishi, M., Hatakeyama, D., Watanabe, T., Fujito, Y., Dyakonova, V., & Lukowiak, K. (2015b). Memory block: A consequence of conflict resolution. Journal of Experimental Biology, 218, 1699-1704. http://dx.doi.org/10.1242/jeb.120329. Ito, E., Kaneda, M., Kodama, H., Morikawa, M., Tai, M., Aoki, K., Watabe, S., Nakaishi, K., Hashida, S., Tada, S., Kuroda, N., Imachi, H., Murao, K., Yamashita, M., Yoshimura, T., & Miura T. (2015c). Immunoreactive insulin in diabetes mellitus patient sera detected by ultrasensitive ELISA with thio-NAD cycling. Biotechniques, 59, 359, 361-367. http://dx.doi.org/10.2144/000114355.

40

Ito, E., Totani, Y., & Oike, A. (2017). Necessity knows no law in a snail. European Zoological Journal, 84, 457-464. http://dx.doi.org/10.1080/24750263.2017.1363303. Ivashkin, E., Khabarova, M. Y., Melnikova, V., Nezlin, L. P., Kharchenko, O., Voronezhskaya, E. E., Adameyko, I. (2015). Serotonin mediates maternal effects and directs developmental and behavioral changes in the progeny of snails. Cell Reports, 12, 1144-1158. http://dx.doi.org/ 10.1016/j.celrep.2015.07.022. Jonas, E. A., Knox, R. J., Kaczmarek, L. K., Schwartz, J. H., & Solomon, D. H. (1996). Insulin receptor in Aplysia neurons: Characterization, molecular cloning, and modulation of ion currents. Journal of Neuroscience, 16, 1645-1658. Jonas, E. A., Knox, R. J., Smith, T. C., Wayne, N. L., Connor, J. A., & Kaczmarek, L. K. (1997). Regulation by insulin of a unique neuronal Ca2+ pool and of neuropeptide secretion. Nature, 385, 343-346. http://dx.doi.org/10.1038/385343a0. Kandel, E.R. (2001). The molecular biology of memory storage: a dialog between genes and synapses. Bioscience Reports, 21, 565-611. http://dx.doi.org/10.1023/A:1014775008533. Kawai, R., Kobayashi, S., Fujito, Y., & Ito, E. (2011). Multiple subtypes of serotonin receptors in the feeding circuit of a pond snail. Zoological Science, 28, 517-525.

41

http://dx.doi.org/10.2108/zsj.28.517. Kemenes, G., Elekes, K., Hiripi, L., & Benjamin, P. R. (1989). A comparison of four techniques for mapping the distribution of serotonin and serotonin-containing neurons in fixed and living ganglia of the snail, Lymnaea. Journal of Neurocytology, 18, 193-208. http://dx.doi.org/10.1007/BF01206662. Erratum in: 18, 565. http://dx.doi.org/10.1007/BF01474551. Kemenes, G., Hiripi, L., & Benjamin, P. R. (1990). Behavioural and biochemical changes in the feeding system of Lymnaea induced by the dopamine and serotonin neurotoxins 6-hydroxydopamine and 5,6-dihydroxytryptamine. Philosophical Transactions of the Royal Society London B, 329, 243-255. http://dx.doi.org/10.1098/rstb.1990.0168. Kojima, S., Yamanaka, M., Fujito, Y., & Ito, E. (1996). Differential neuroethological effects of aversive and appetitive reinforcing stimuli on associative learning in Lymnaea stagnalis. Zoological Science, 13, 803-812. http://dx.doi.org/10.2108/zsj.13.803. Kojima, S., Nakamura, H., Nagayama, S., Fujito, Y., & Ito, E. (1997). Enhancement of an inhibitory input to the feeding central pattern generator in Lymnaea stagnalis during conditioned taste-aversion learning. Neuroscience Letters, 230, 179-182.

42

http://doi.org/10.1016/S0304-3940(97)00507-7 Kojima, S., Kobayashi, S., Yamanaka, M., Sadamoto, H., Nakamura, H., Fujito, Y., Kawai, R., Sakakibara, M., & Ito, E. (1998). Sensory preconditioning for feeding response in the pond snail, Lymnaea stagnalis. Brain Research, 808, 113-115. http://dx.doi.org/10.1016/S0006-8993(98)00823-3. Kojima, S., Hosono, T., Fujito, Y., & Ito, E. (2001). Optical detection of neuromodulatory effects of conditioned taste aversion in the pond snail Lymnaea stagnalis. Journal of Neurobiology, 49, 118-128. http://dx.doi.org/10.1002/neu.1069. Kojima, S., Sunada, H., Mita, K., Sakakibara, M., Lukowiak, K., & Ito, E. (2015). Function of insulin in snail brain in associative learning. Journal of Comparative Physiology A, 201, 969-981. http://dx.doi.org/10.1007/s00359-015-1032-5. Mainardi, M., Fusco, S., & Grassi, C. (2015). Modulation of hippocampal neural plasticity by glucose-related signaling. Neural Plasticity, 2015, 657928. http://dx.doi.org/10.1155/2015/657928. Meneses, A. (2013). 5-HT systems: emergent targets for memory formation and memory alterations. Reviews in the Neurosciences. 24, 629-664. http://dx.doi.org/10.1515/revneuro-2013-0026.

43

Mita, K., Okuta, A., Okada, R., Hatakeyama, D., Otsuka, E., Yamagishi, M., Morikawa, M., Naganuma, Y., Fujito, Y., Dyakonova, V., Lukowiak, K., & Ito, E. (2014a). What are the elements of motivation for acquisition of conditioned taste aversion? Neurobiology of Learning and Memory, 107, 1-12. http://dx.doi.org/10.1016/j.nlm.2013.10.013. Mita, K., Yamagishi, M., Fujito, Y., Lukowiak, K., & Ito, E. (2014b). An increase in insulin is important for the acquisition conditioned taste aversion in Lymnaea. Neurobiology of Learning and Memory, 116, 132-138. http://dx.doi.org/10.1016/j.nlm.2014.10.006. Murakami, J., Okada, R., Sadamoto, H., Kobayashi, S., Mita, K., Sakamoto, Y., Yamagishi, M., Hatakeyama, D., Otsuka, E., Okuta, A., Sunada, H., Takigami, S., Sakakibara, M., Fujito, Y., Awaji, M., Moriyama, S., Lukowiak, K., & Ito, E. (2013a). Involvement of insulin-like peptide in long-term synaptic plasticity and long-term memory of the pond snail Lymnaea stagnalis. Journal of Neuroscience, 33, 371-383. http://dx.doi.org/10.1523/JNEUROSCI.0679-12.2013. Murakami, J., Okada, R., Fujito, Y., Sakakibara, M., Lukowiak, K., & Ito, E. (2013b). Paired pulse ratio analysis of insulin-induced synaptic plasticity in the snail brain. Journal of Experimental Biology, 216, 1771-1773.

44

http://dx.doi.org/10.1242/jeb.083469. Nikitin, V. P., & Solntseva, S. V. (2013). Peculiarities of amnesia development during memory reconsolidation impairment induced by isolated or combined treatment with neurotransmitter receptor antagonists. Bulletin of Experimental Biology and Medicine, 155, 6-10. http://dx.doi.org/10.1007/s10517-013-2066-8. Olvera-Cortés, M. E., Gutiérrez-Guzmán, B. E., López-Loeza, E., Hernández-Pérez, J. J., & López-Vázquez MA. (2013). Serotonergic modulation of hippocampal theta activity in relation to hippocampal information processing. Experimental Brain Research, 230, 407-426. http://dx.doi.org/10.1007/s00221-013-3679-x. Olivier, B. (2015). Serotonin: a never-ending story. European Journal of Pharmacology, 753, 2-18. http://dx.doi.org/10.1016/j.ejphar.2014.10.031. Otsuka, E., Matsunaga, M., Okada, R., Yamagishi, M., Okuta, A., Lukowiak, K., & Ito E. (2013). Increase in cyclic AMP concentration in a cerebral giant interneuron mimics part of a memory trace for conditioned taste aversion of the pond snail. BIOPHYSICS, 9, 161-166. http://dx.doi.org/10.2142/biophysics.9.161. Sadamoto, H., Serfőző, Z., & Ito, E. (2008). Localization of serotonin transporter mRNA in the CNS of Lymnaea stagnalis. Acta Biologica Hungarica, 59, Suppl: 61-64. http://dx.doi.org/10.1556/ABiol.59.2008.Suppl.9.

45

Scheibenstock, A., Krygier, D., Haque, Z., Syed, S., & Lukowiak, K. (2002). The soma of RPeD1 must be present for long-term memory formation of associative learning in Lymnaea. Journal of Neurophysiology, 88, 1584- 1591. Shields-Johnson, M. E., Hernandez, J. S., Torno, C., Adams, K. M., Wainwright, M. L., & Mozzachiodi, R. (2012). Effects of aversive stimuli beyond defensive neural circuits: reduced excitability in an identified neuron critical for feeding in Aplysia. Learning & Memory, 20, 1-5. http://dx.doi.org/doi: 10.1101/lm.028084.112. Sunada, H., Lukowiak, K., & Ito, E. (2017a). Cerebral giant cells are necessary for both the formation and recall of memory of conditioned taste aversion in Lymnaea. Zoological Science, 34, 72-80. http://dx.doi.org/10.2108/zs160152. Sunada, H., Watanabe, T., Hatakeyama, D., Lee, S., Forest, J., Sakakibara, M., Ito, E., & Lukowiak, K. (2017b). Pharmacological effects of cannabinoids on learning and memory in Lymnaea. Journal of Experimental Biology, 220, 3026-3038. http://dx.doi.org/10.1242/jeb.159038. Sunada, H., Totani, Y., Nakamura, R., Sakakibara, M., Lukowiak, K., & Ito, E. (2017c). Two strains of Lymnaea stagnalis and the progeny from their mating display differential memory-forming ability on associative learning tasks. Frontiers in Behavioral Neuroscience, 11, 161. https://doi.org/10.3389/fnbeh.2017.00161.

46

Tsyganov, V. V. & Sakharov, D. A. (2000). Locomotor rhythms in the pond snail Lymnaea stagnalis: site of origin and neurotransmitter requirements. Acta Biologica Hungarica, 51, 189-195. Ureshi, M., Dainobu, M., & Sakai M. (2002). Serotonin precursor (5-hydroxytryptophan) has a profound effect on the post-copulatory time-fixed sexually refractory stage in the male cricket, Gryllus bimaculatus DeGeer. Journal of Comparative Physiology A, 188, 767-779. http://dx.doi.org/10.1007/s00359-002-0364-0. Voigt, J. P., & Fink, H. (2015). Serotonin controlling feeding and satiety. Behavioral Brain Research, 277, 14-31. http://dx.doi.org/10.1016/j.bbr.2014.08.065. Wagatsuma, A., Sugai, R., Chono, K., Azami, S., Hatakeyama, D., Sadamoto, H., & Ito, E. (2004). The early snail acquires the learning. Comparison of scores for conditioned taste aversion between morning and afternoon. Acta Biologica Hungarica, 55, 149-155. http://dx.doi.org/10.1556/ABiol.55.2004.1-4.18. Watabe, S., Kodama, H., Kaneda, M., Morikawa, M., Nakaishi, K., Yoshimura, T., Iwai, A., Miura, T., & Ito, E. (2014). Ultrasensitive enzyme-linked immunosorbent assay (ELISA) of proteins by combination with the thio-NAD cycling method. BIOPHYSICS, 10, 49-54. http://dx.doi.org/10.2142/biophysics.10.49. Watabe, S., Morikawa, M., Kaneda, M., Nakaishi, K., Nakatsuma, A., Ninomiya, M.,

47

Yoshimura, T., Miura, T., & Ito, E. (2016). Ultrasensitive detection of proteins and sugars at single-cell level. Communicative & Integrative Biology, 9, e1124201. http://dx.doi.org/10.1080/19420889.2015.1124201. Yamagishi, M., Watanabe, T., Hatakeyama, D., & Ito, E. (2015). Effects of serotonin on the heartbeat of pond snails in a hunger state. BIOPHYSICS, 11, 1-5. http://dx.doi.org/10.2142/biophysics.11.1. Yamanaka, M., Hatakeyama, D., Sadamoto, H., Kimura, T., & Ito, E. (2000). Development of key neurons for learning stimulates learning ability in Lymnaea stagnalis. Neuroscience Letters, 278, 113-116. https://doi.org/10.1016/S0304-3940(99)00916-7.

48

Figure legends

Fig. 1. The training procedure for CTA training in snails in different food-deprivation states. (A) Food deprivation status. The day when snails began food deprivation was designated Day 0. Day -1 snails: snails were able to feed ad libitum. Day 1 snails: snails were food-deprived for 1 day. Day 5 snails: snails were food-deprived for 5 days. Further, snails that were in the act of eating just before the commencement of the application of drugs or CTA training were referred to as Eating snails (not indicated in this figure). (B) CTA training procedures. A 10 mM sucrose solution was used as the CS, whereas a 3-s high voltage electric shock was used as the US. DW was distilled water. Controls included a backward-conditioning (US-CS) group and a naive group to validate associative learning.

Fig. 2. Changes in CNS content of 5-HTP, 5-HT, Nac-5-HT and 5-HIAA in the CNS isolated from snails in the different food-deprived states. We utilized 4 cohorts: (1) Eating snails; (2) Day -1 snails; (3) Day 1 snails; and (4) Day 5 snails as explained in Fig. 1 legend. The y-axis shows the content per a single CNS. (A) 5-HTP (the precursor of 5-HT) content. (B) 5-HT content. (C) Nac-5-HT (one of the catabolites of 5-HT) content.

49

(D) 5-HIAA (one of the catabolites of 5-HT) content. (E) The data were normalized to that of Eating snails. *P < 0.05, **P < 0.01.

Fig. 3. CTA learning and memory formation in the different food states. We prepared CTA-trained snails (black bars), backward-conditioned control snails (hatched bars) and naive control snails (white bars). N = 20 each. The feeding response to 10 mM sucrose was indicated as the number of bites per minute. (A) Eating snails. After CTA training, the feeding response was weakly suppressed in comparison with those of control snails. However, the learning ability was considered extremely unstable. (B) Day -1 snails. The feeding responses were suppressed, and CTA memory persisted from 10 min to at least 1 week. (C) Day 1 snails. The feeding response was strongly suppressed, and CTA memory persisted for at least one week. As can be seen, the learning ability is better in Day 1 snails than Day -1 snails. (D) Day 5 snails. The feeding response in the CTA-trained snails was weakly suppressed in comparison with those of control snails, but we concluded that CTA memory was not formed because a significant difference was not observed between the CTA-trained snails and the naive control snails. That is, Day 1 snails form CTA-LTM, whereas Day 5 snails do not. *P < 0.05, **P < 0.01.

50

Fig. 4. Effects of different 5-HT CNS levels on the elicited feeding response to sucrose in food-deprived snails. (A) Application of 5-HTP to Day 1 snails. The feeding response in Day 1 snails (N = 18) was measured to 10 mM sucrose after 1-h application of 200 μM of 5-HTP. In this cohort of snails, there was a decrease in the feeding response compared to the control (i.e., 0 μM of 5-HTP as indicated at the time of -1 h) that persisted for 3 h. (B) Application of 5-HT to Day 1 snails. The feeding response of snails elicited by 10 mM sucrose was examined after 24-h application of different concentrations of 5-HT. During the time, the snails were food-deprived. The numbers of snails used were 10 - 20. A high concentration (> 300 μM) of 5-HT decreased the feeding response in a dose-dependent manner. (C) Application of cinanserin and MDL72222 to Day 5 snails. The feeding response of sails elicited by 10 mM sucrose was examined after 24-h application of different concentrations of the two 5-HT receptor antagonists (cinanserin and MDL72222). These two antagonists were applied simultaneously. The snails were food-deprived for 5 days, and on the last one day the snails were given these antagonists. The x-axis shows each concentration of cinanserin and MDL72222. The number of snails used in each experiment was 20. The antagonist application suppressed the feeding response at the concentration of 100 μM each and 500 μM each, whereas the results by 100 μM and 500 μM application were almost same. *P < 0.05, **P < 0.01.

51

Fig. 5. Alteration of learning and memory formation by immersion of Day 1 snails in a 5-HT bath. A 300 μM bath of 5-HT alters 5-HT level in the CNS. CTA scores in Day 1 snails with a 300 μM 5-HT solution for 24 h. The number of snails used was 10 each. We tested CTA-trained snails, and found that the feeding response of CTA-trained snails to the CS (i.e., 10 mM sucrose) was not suppressed in comparison with the pretest. That is, CTA-LTM did not occur.

Fig. 6. Effects of injection of insulin or the insulin receptor antibody into Day 1snails on the 5-HT CNS content. (A) Day 1 snails. There are no significant differences. (B) Day 5 snails. There are no significant differences, either. That is, we conclude that the insulin signaling pathway does not affect the 5-HT signaling pathway. The number of snails used was 10 each.

Fig. 7. Insulin rescues the learning and CTA-LTM impairment caused by 5-HT in Day 1 snails. We trained the following 4 groups: (1) CTA training after 300 μM 5-HT immersion for 24 h and 100 nM (estimated final concentration) insulin injection for 1 h; (2) CTA training after 300 μM 5-HT immersion for 24 h and 1.2 μM HCl (estimated final

52

concentration) injection for 1 h; (3) Backward conditioning (US-CS) after 300 μM 5-HT immersion for 24 h; and (4) Naive control after 300 μM 5-HT immersion for 24 h. The concentrations of injected insulin and vehicle (i.e., HCl) were estimated by the same methods described earlier (Mita et al., 2014b). The number of snails used was 10 each. No significant differences were found in the pretest among all the 4 groups. The feeding response only in the snails following CTA training after 300 μM 5-HT immersion for 24 h and 100 nM insulin injection for 1 h was suppressed at the 10-min, 1-h, 1-day and 1-week post-tests. **P < 0.01.

53

Highlights: 1. One-day food-deprived snails that exhibit the best CTA learning had a low 5-HT content in their CNS. 2. Five-day food-deprived snails, which did not learn, had a high level of 5-HT content in their CNS. 3. Insulin rescues the CTA deficit and this may be due to a decrease in the 5-HT content in the CNS.

54