Cannabinoids are hypothesized to play an important role in modulating learning and memory formation. Here, we identified mRNAs expressed in Lymnaea stagnalis central nervous system that encode two G-protein-coupled receptors (Lymnaea CBr-like 1 and 2) that structurally resemble mammalian cannabinoid receptors (CBrs). We found that injection of a mammalian CBr agonist WIN 55,212-2 (WIN 55) into the snail before operant conditioning obstructed learning and memory formation. This effect of WIN 55 injection persisted for at least 4 days following its injection. A similar obstruction of learning and memory occurred when a severe traumatic stimulus was delivered to L. stagnalis. In contrast, injection of a mammalian CBr antagonist AM 251 enhanced long-term memory formation in snails and reduced the duration of the effects of the severe traumatic stressor on learning and memory. Neither WIN 55 nor AM 251 altered normal homeostatic aerial respiratory behaviour elicited in hypoxic conditions. Our results suggest that putative cannabinoid receptors mediate stressful stimuli that alter learning and memory formation in Lymnaea. This is also the first demonstration that putative CBrs are present in Lymnaea and play a key role in learning and memory formation.
Stress alters learning and memory formation. Roger Bacon suggested this in 1620 (Bacon, 1620) and it is codified in the Yerkes–Dodson inverted-U law (Yerkes and Dodson, 1908). Yerkes and Dodson (1908) stated that ‘an easily acquired habit may be readily formed under strong stimulation, whereas a difficult habit may be acquired only under relatively weak stimulation’ (see also Ito et al., 2015). Donald Hebb (1955) was probably the first to portray this as an inverted-U in his presidential address to the American Psychological Association in the 1950s, where he discussed stress and memory. Succinctly put: there is an optimal level of stress that results in the ‘best’ memory; too much or too little stress results in poorer memory formation and/or its recall.
In our model system, the pond snail Lymnaea stagnalis, ecologically relevant stressors such as predator detection, crowding, low environmental calcium and thermal shock all alter the ability of snails to learn associatively and form long-term memory (LTM) following operant conditioning of aerial respiratory behaviour (Lukowiak et al., 2010, 2014a,b). For example, following crowding, LTM formation is blocked (de Caigny and Lukowiak, 2008), while following exposure to a thermal shock, LTM formation is enhanced (Teskey et al., 2012). When more than one stressor is experienced by the animal, there are emergent effects arising from the combination of the stressors that result in unpredictable consequences on the snails' ability to learn and form memory (Dalesman and Lukowiak, 2012; Dalesman et al., 2013; Lukowiak et al., 2014a; Ito et al., 2015). The mechanisms by which the various stressors affect memory formation are not entirely clear. However, it has been hypothesized that cannabinoids play a key role in mediating stressor effects on adaptive behaviours (Morena and Campolongo, 2014; Wamsteeker-Cusulin et al., 2014; Goodman and Packard, 2015).
In mammals, the endocannabinoid system is composed of endogenous ligands, such as 2-arachidonoyl glycerol (2-AG) and anandamide (AEA), and mainly pre-synaptically localized receptors. These receptors, e.g. cannabinoid receptor type 1 (CB1), exert important neuromodulatory roles, such as how stress alters learning and memory (Campolongo et al., 2009; Hauer et al., 2011; Atsak et al., 2012; Tan et al., 2014). The endocannabinoid system also plays a key role in the neuronal regulation of anxiety and fear responses (Ruehle et al., 2012; Mechoulam and Parker, 2013). We hypothesize that cannabinoid receptors (CBrs) in Lymnaea are involved in a similar role and the activity of the cannabinoid system also plays a key role in how stress modifies LTM formation.
Although the endocannabinoid system is phylogenetically ancient, having been shown to occur in most phyla with the notable exception of insects (McPartland et al., 2001; Salzet and Stefano, 2002; McPartland, 2004; Elphick, 2012), it has not yet been identified in Lymnaea. Importantly, demonstrating the presence of CBrs would allow us to make use of a number of advantages offered to us by our model system for the study of the role of cannabinoids in memory formation in Lymnaea (Elphick, 2012). A chief advantage of our model system is that an identified neuron, right pedal dorsal 1 neuron (RPeD1), is a necessary site for LTM formation, its reconsolidation, extinction and forgetting following operant conditioning of aerial respiration (Scheibenstock et al., 2002; Sangha et al., 2003a,b, 2005). It would therefore be possible to begin to determine exactly how the cannabinoids alter activity in this neuron that is causal to memory formation.
It was shown in the gastropod mollusc Aplysia californica in the 1970s that the main psychoactive ingredient of Cannabis sativa, Δ9-tetrahydrocannabinol, caused a depression in neuronal excitability (Acosta-Urquidi and Chase, 1975). Presumably, this effect was mediated by CBrs although until now the presence of such receptors in Aplysia has not been confirmed. The endocannabinoid system has also been reported to be involved in the regulation of synaptic efficacy in the terrestrial snail Helix lucorum (Lemak et al., 2007). CBr genes have been identified in some other invertebrates such as the leech Hirudo medicinalis (Stefano et al., 1997; Elphick, 1998; Matias et al., 2001; McPartland and Glass, 2003), the nematode Caenorhabditis elegans (McPartland and Glass, 2003; Oakes et al., 2017), the ascidian Ciona intestinalis (Elphick et al., 2003) and the lancelet Branchiostoma floridae (Elphick, 2007). We show here mRNA expression of two putative CBr genes in the central nervous system (CNS) of L. stagnalis. We also show pharmacologically that these putative CBrs are involved in a key role in modulating learning and memory formation. We then further demonstrate that experiencing an extreme stressor blocks learning and memory. This effect of extreme stress can be mimicked by the injection of a CBr agonist, and blockage of these receptors by a CBr antagonist mitigates the effects of the severe stressor. Moreover, the CBr antagonist itself can enhance LTM formation.
MATERIALS AND METHODS
The pond snail Lymnaea stagnalis (Linnaeus 1758) (20–30 mm shell length) was used. Snails were maintained on an 8 h light:16 h dark cycle at room temperature (20–23°C) in well-aerated pond water and fed lettuce ad libitum (Yamagishi et al., 2015; Aonuma et al., 2016, 2017; Sunada et al., 2017). Artificial pond water was made from deionized water containing 0.26 g l−1 Instant Ocean (Aquarium Systems Inc., Mentor, OH, USA) and calcium sulphate dihydrate to provide a calcium concentration of 80 mg l−1 (Dalesman and Lukowiak, 2010).
Molecular cloning of putative CBrs in L. stagnalis
We conducted a tblastn search on the Lymnaea CNS transcriptome shotgun assembly (TSA) sequence database (Sadamoto et al., 2012) to search for Lymnaea homologues of the invertebrate-type CBr. We found two TSA contigs (FX183817 and FX186161), both of which encode rhodopsin-like G-protein-coupled receptors with sequence similarity to Ciona intestinalis CBr (NP_001027653).
Total RNA was isolated from L. stagnalis CNS with TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) and treated with DNase I (TaKaRa, Shiga, Japan) at 37°C for 1 h. Total RNA was then reverse-transcribed using Superscript III reverse transcriptase (Thermo Fisher Scientific) with the 3′-RACE adapter primer from the FirstChoice RLM-RACE kit (Ambion, Austin, TX, USA). cDNAs encoding the full-length L. stagnalis CBr-like proteins were amplified using gene-specific primers designed to the 5′ and 3′ untranslated regions (UTRs) of two putative Lymnaea CBr-like genes (Lymnaea CBr-like 1, 5′-CGT TAC CAA ATT TAC CGA CCA C-3′ and 5′-CCG GAT GAA CTC ACG ATT CG-3′; Lymnaea CBr-like 2, 5′-GGA CAC GAC CAC CAC CG-3′ and 5′-CAA GGG AAG ATA CAT TCA TAG CCA G-3′). PCRs were carried out using Q5 High-Fidelity DNA Polymerase (New England Biolabs, Tokyo, Japan). The amplified DNA fragments were subcloned into the plasmid vector pGEM-T easy (Promega, Madison, WI, USA), and the nucleotide sequences were determined.
The deduced amino acid sequences of the Lymnaea CBr-like proteins were aligned with those of molluscan CBr-like proteins by the MAFFT algorithm (Katoh and Standley, 2013) in Geneious (v8.1.6; http://www.geneious.com/) and visualized with BoxShade (v3.21; http://www.ch.embnet.org/software/BOX_form.html). Transmembrane helices of Lymnaea CBr-like proteins were predicted using the GPCRHMM algorithm (Wistrand et al., 2006).
The deduced full-length protein sequences of Lymnaea CBr-like proteins and those of the G-protein-coupled receptors related to the CBr family were aligned as described above. A maximum likelihood tree was constructed from the aligned sequences using the MEGA 6 program with default settings (Tamura et al., 2011, 2013). One thousand bootstrap replications were conducted to evaluate the reliability of the reconstructed tree. Vertebrate G-protein-coupled receptors closely related to the CBr (Wolf and Grüneweld, 2015) [melanocortin receptors (Homo sapiens MC1-5), sphingosine 1-phosphate receptors (Homo sapiens S1P1-5) and lysophosphatidic acid receptors (Homo sapiens LPA1-6)] were included in the analysis as outgroups to root the tree. The obtained tree was visualized with FigTree (v1.4.2; http://tree.bio.ed.ac.uk/software/figtree/). GenBank IDs of the proteins used for phylogenetic analysis are listed in Table 1.
SYBR Green-based qRT-PCR
The procedures of primer extension analyses were performed as described previously with modifications (Hatakeyama et al., 2013). To determine the expression level of Lymnaea CBr-like 1 and 2 mRNA, we dissected the CNS, buccal mass, penis, ovotestis, gut and mantle from 12 snails. Four samples of each tissue were allocated into three groups. Total RNA samples of tissues were purified using TRI Reagent (Sigma-Aldrich, St Louis, MO, USA). Reverse transcription (RT) was performed using 20 µl of each total RNA preparation, SuperScript IV Reverse Transcriptase (Thermo Fisher Scientific), RNaseOUT (Thermo Fisher Scientific) and oligo d(T)12–18 primer (Thermo Fisher Scientific) following the product manuals. RT samples were diluted 1/20 with distilled water, then mixed with SYBR Green Realtime PCR Master Mix (ToYoBo, Osaka, Japan) and a primer set selectively amplifying Lymnaea CBr-like 1 and 2 or Lymnaea actin. We designed two primer sets for each Lymnaea CBr-like 1 and 2 using Primer Express Software v3.0 (Thermo Fisher Scientific). We used three different reference genes: actin, tubulin and Lymnaea elongation factor α (LEFα). The nucleotide sequence of each primer was as follows: Lymnaea CBr-like-1-F (1st set) 5′-TCT CCA CTG AGA TCT AAA CGA ATC TG-3′; Lymnaea CBr-like-1-R (1st set) 5′-ACA CCA GTA GGC CTT GAA TAC CTT-3′; Lymnaea CBr-like-1-F (2nd set) 5′-CAT CTA TGG GAT ACT GTT CGA GTC ATA C-3′; Lymnaea CBr-like-1-R (2nd set) 5′-ATG GTG CCC CGT GTA CTT GTA G-3′; Lymnaea CBr-like-2-F (1st set) 5′-CCG ACG GAC TCA AGG AAA AC-3′; Lymnaea CBr-like-2-R (1st set) 5′-TGA CCA GTG CGT TGC CAA T-3′; Lymnaea CBr-like-2-F (2nd set) 5′-ATG CGC AGT ATC ACC AAT ATA TTT CT-3′; Lymnaea CBr-like-2-R (2nd set) 5′-CTG CAG GTA GTG GAC AGC CTT AC-3′; Lymnaea actin-F 5′-TCC CTT GAG AAG AGC TAC GAG C-3′; Lymnaea actin-R 5′-GAG TTG TAG GTG GTT TCG TGG-3′. The nucleotide sequences of primers for tubulin and LEFα were obtained from van Nierop et al. (2006). The reaction was carried out at 95°C for 1 min followed by 50 cycles of 95°C for 15 s, 60°C for 15 s, and 72°C for 30 s each using the StepOnePlus Real-Time PCR System (Thermo Fisher Scientific). Each preparation of CBr-like 1, CBr-like 2 and actin was applied to a 96-well plate in triplicate. Relative levels of CBr-like 1 and 2 mRNA were calculated by the comparative CT (ΔΔCT) method (Thermo Fisher Scientific) (Livak and Schmittgen, 2001; Hatakeyama et al., 2013) using actin mRNA as a reference to give normalized CBr-like 1 and 2 mRNA expression levels.
Aerial respiratory operant conditioning
Lymnaea is a pulmonate bimodal breather. In the eumoxic environment (i.e. normal levels of O2), snails mainly absorb oxygen through their skin, whereas in the hypoxic environment, they move up to the water surface and open the pneumostome to inhale atmospheric air. The experimental procedures for aerial respiratory operant conditioning were described previously (Orr et al., 2009; Lukowiak et al., 2000). Briefly, a 1 litre beaker containing 500 ml pond water at room temperature was bubbled with N2 for 20 min to make it hypoxic. Snails were numbered 24 h before training, and then placed in the hypoxic training beaker for 10 min for acclimation, followed by training either with a single 0.5 h training session or with two 0.5 h training sessions separated by a 1 h interval. In both training procedures, LTM was tested 24 or 48 h later. Memory formation has been operationally defined (Orr et al., 2009; Lukowiak et al., 2000). When two 0.5 h training sessions are used, memory formation is indicated if the number of attempted pneumostome openings in the memory test session (MT) is significantly lower than the number of attempted openings in the initial training session (TS1) and not significantly greater than the number of attempted openings in the second training session (TS2). However, when only a single 0.5 h training session procedure is used, as in the data shown in Results, ‘Trauma and LTM’, the number of attempted openings in the memory test session has to be significantly lower than that in the initial training session. In the training and memory test sessions, a gentle tactile stimulus was applied to the pneumostome area using a sharpened wooden stick (approximately 0.8 mm diameter sharpened end) every time the snail began to open its pneumostome to perform areal respiration. This tactile stimulus evoked only pneumostome closure; it did not cause the snail to withdraw its foot and mantle (i.e. the whole-body withdrawal response). Snails were returned to the eumoxic water after the training sessions. Twenty-four hours after the final training session (TS1 or TS2), a memory retention test was carried out following the same procedure as in the training session, for 30 min. The time sequence of operant conditioning is shown in Fig. 4A. It should be noted that individual snails in each cohort were only used in a single experiment; a cohort was not used in multiple experiments and each snail was only tested for memory once.
Noxious mechanical/tactile stress
We traumatized snails using repeated presentations (∼50 times) of a noxious tactile stimulus to the foot or mantle cavity with a hand-held sharp wooden stick (as for the operant conditioning procedure, described above). This stimulation procedure did not break the skin of the snail. Such stimulation always elicited the whole-body withdrawal response (Sunada et al., 2012). As Lymnaea do not possess an operculum, the foot of the snail was always visible and could thus receive continued stimulation, even though the snails withdrew as much as possible into their shell with this procedure. Snails did not have to be repositioned to enable us to present this traumatic stimulus.
Effects of CBr agonist and antagonist on behaviour changes
The mammalian CBr agonist WIN 55,212-2 (WIN 55) and its antagonist AM 251 (both from Tocris Bioscience, Bristol, UK) were dissolved in DMSO as a stock solution at 10 mmol l−1 concentration. Stock solutions were diluted with Lymnaea saline (51.3 mmol l−1 NaCl, 1.7 mmol l−1 KCl, 1.5 mmol l−1 MgCl2, 4.0 mmol l−1 CaCl2 and 10.0 mmol l−1 Hepes, pH 8.0) for both the behavioural experiments. For control experiments, 0.1% DMSO was dissolved in the saline as the vehicle solution. For behavioural experiments, snails were anaesthetized with ice-cold pond water for 30 min before injection. The CBr agonist or antagonist (10 µmol l−1) in 100 µl Lymnaea saline was injected into the abdominal cavity of a snail, giving an estimated concentration of the drug in the Lymnaea body of about 2 µmol l−1. In the experiment described in Results, ‘Trauma and LTM’, AM 251 was dissolved in 0.1% DMSO and 0.1% Tween 80 in Lymnaea saline. This stock solution was diluted in the saline to give a final concentration of 18 µmol l−1 AM 251; 100 µl of this solution was injected into the snails to give a final concentration in the body of approximately 3.6 µmol l−1 AM 251. Snails were prepared as independent groups in each behavioural experiment.
The data are represented as means±s.e.m. The mRNA expression levels of Lymnaea CBr-like genes were analysed by one-way analysis of variance (ANOVA) followed by Tukey's post hoc test. As stated above, memory was operationally defined as a significant decrease in the number of attempted pneumostome openings in MT compared with TS1, where the number was not significantly greater than that in TS2. A one-way ANOVA and Tukey's post hoc test were used to determine whether the agonists or antagonists altered homeostatic aerial respiratory behaviour. Two-way repeated measures ANOVA (rmANOVA) was used to analyse the behavioural response of the snails receiving operant conditioning of aerial respiratory behaviour, with the various endocannabinoid agonists, antagonists and trauma applications (Figs 4–7) as the between-subjects factor and the training and memory test sessions (TS1, TS2, MT) as the within-subject factors. Where significant interactions were found (P<0.05), Tukey's post hoc test was used for comparisons to determine significant differences between the responses. In the naive and vehicle conditions, a paired t-test was used to determine whether LTM was formed; in the AM 251 condition, an ANOVA followed a Tukey's test was used to determine whether memory was formed 24 and 48 h after training. In all cases, for significance to be declared, P<0.05 had to be achieved. Data were analysed using Prism 6 software for Macintosh.
Putative CBr-like genes in L. stagnalis
We found two transcriptome shotgun assembly contigs (Lymnaea CBr-like 1, GenBank: FX183817; and Lymnaea CBr-like 2, GenBank: FX186161), both of which encode rhodopsin-like G-protein-coupled receptors with sequence similarity to Ciona intestinalis CBr (NP_001027653) (Elphick et al., 2003). We next performed RT-PCR to ensure that the mRNAs of two Lymnaea CBr-like genes are expressed in the Lymnaea CNS, using gene-specific primers designed to the 5′- and 3′-UTRs of the genes, and successfully amplified their cDNA fragments: a 1622 bp cDNA fragment of Lymnaea CBr-like 1 (GenBank: LC093511) and a 1537 bp cDNA fragment of Lymnaea CBr-like 2 (GenBank: LC093512). The cDNA fragment of Lymnaea CBr-like 1 contains an open reading frame (ORF) spanning 97–1581 bp with a predicted protein product of 494 amino acids. The cDNA fragment of Lymnaea CBr-like 2 contains an ORF spanning 46–1482 bp with a predicted protein product of 478 amino acids. The amino acid sequence of Lymnaea CBr-like 1 is 37.8% identical to that of Lymnaea CBr-like 2. The transmembrane regions of Lymnaea CBr-like proteins are conserved (57.5% identity).
Next, we searched for molluscan genes encoding a protein that structurally resembles the Lymnaea CBr-like proteins. We found predicted genes encoding CBr-like proteins from three molluscan species (Aplysia californica, Biomphalaria glabrata and the owl limpet Lottia gigantea). We also searched for genes related to the Lymnaea CBr-like proteins in the gene database of Brachiopoda, which is a closely related phylum to Mollusca in the superphylum Lophotrochozoa, and found a gene encoding a putative CBr-like protein from the brachiopod Lingula anatina. We compared the amino acid sequences of Lymnaea CBr-like proteins and other molluscan/brachiopodan CBr-like proteins (Fig. 1, Table 1), and found that these CBr-like proteins can be classified into two subgroups according to the sequence and structure of the third cytoplasmic loop, the region between the fifth and sixth transmembrane helices. Lymnaea CBr-like 1 and Biomphalaria CBr-like protein have a long third cytoplasmic loop (104 amino acids and 112 amino acids, respectively) with high sequence homology (65.8% identity; Fig. 1A). Lymnaea CBr-like 2 and CBr-like proteins of A. californica, L. gigantea and L. anatina have a short third cytoplasmic loop (59–66 amino acids) with low sequence homology (14–32% identity; Fig. 1A). A molecular phylogenetic analysis of the CBrs with other G-protein-coupled receptors revealed that molluscan/brachiopodan CBr-like proteins are closely related to the vertebrate CBrs (CB1 and CB2) and other CBr-like proteins (Fig. 1B). Our phylogenetic analysis also revealed that gene duplication events of the CBr-like gene have independently occurred in molluscan and vertebrate lineages.
Comparison of expression levels of CBr-like genes among several tissues of Lymnaea
We next examined tissue-specific mRNA expression of the two Lymnaea CBr-like genes by SYBR Green-based qRT-PCR. The levels of Lymnaea CBr-like 1 and 2 mRNA among snail CNS, buccal mass, penis, ovotestis, gut and mantle were compared using three different reference genes (actin, tubulin and LEFα) according to the ΔΔCT method for qRT-PCR (Fig. 2). Lymnaea CBr-like 1 mRNA was detected in all tested tissues containing the CNS (Fig. 2A). Expression in the ovotestis and gut was significantly higher than that in the CNS (Fig. 2A, one-way ANOVA followed by Tukey's post hoc test). In contrast, Lymnaea CBr-like 2 mRNA expression in the CNS was significantly higher than that in other tissues (Fig. 2B, one-way ANOVA followed by Tukey's post hoc test). Lymnaea CBr-like 2 mRNA levels in the other tissues were too low (around 10−8 to 10−4) to quantify. These results suggest that the Lymnaea CBr-like 1 gene was ubiquitously transcribed in tissues and the Lymnaea CBr-like 2 gene was specifically expressed in the CNS of Lymnaea.
Effects of injection of a mammalian CBr agonist or antagonist on aerial respiratory behaviour and aerial respiratory operant conditioning
To test whether the putative CBrs found in the CNS play a role in modulating adaptive behaviours in Lymnaea, we used a mammalian CBr agonist and antagonist to examine whether adaptive behaviour would be altered. We first determined whether the agonist or antagonist altered homeostatic aerial respiratory behaviour elicited in response to a hypoxic environment (Fig. 3A). We found that aerial respiratory behaviour of snails in hypoxic pond water was unaffected by injection of the mammalian CBr agonist WIN 55. Both the total number of breaths and the total breathing time during the 30 min session were unaltered. Likewise, we found that when we injected the mammalian CBr antagonist AM 251, there was no significant change in aerial respiratory behaviour (one-way ANOVA followed by Scheffé's post hoc test, P>0.05; Fig. 3B).
We found that injecting WIN 55 into a cohort of snails 1 h before operant conditioning resulted in the snails exhibiting neither learning nor LTM (Fig. 4A,B). To show that WIN 55 obstructed learning and memory, we injected snails with only the vehicle and found that learning and memory were not obstructed (Fig. 4D; n=12; one-way ANOVA followed by Tukey's post hoc test, P<0.05). We then injected WIN 55 into three cohorts and tested the snails' ability to learn and form LTM 2, 4 and 7 days later to allow us to determine the duration of WIN 55's effect (Fig. 4C,D). Learning and memory formation were obstructed when WIN 55 was injected either 2 days (n=11) or 4 days (n=15) before operant conditioning (Fig. 4D; one-way ANOVA followed by Tukey's post hoc test, P>0.05 in both cohorts). Learning and memory, however, were seen in the cohort of snails that received WIN 55 injection 7 days before training (Fig. 4D; n=16, one-way ANOVA followed by Tukey's post hoc test, P<0.05 in TS1, P<0.01 in MT).
Trauma and LTM
We next asked whether snails experiencing a severe traumatic stimulus exhibit a similar obstruction of learning and memory to that seen in the cohorts injected with WIN 55 2 and 4 days previously (Fig. 5A). We first trained a cohort of snails that did not receive the traumatic stimulus. These snails (Fig. 5B; n=31) showed learning and LTM formation (one-way ANOVA, P<0.001). A second cohort of snails experienced the traumatic stress (i.e. 50 pokes; see Materials and methods) followed 2 days later by training using the same two 0.5 h training sessions with a memory test 24 h later (Fig. 5B). This cohort of snails (n=20) exhibited neither learning nor LTM (one-way ANOVA followed by Tukey's post hoc test, P>0.05). In a third cohort of snails (n=32) trained 4 days after receiving the traumatic stress, learning and memory formation were also not observed (Fig. 5B; one-way ANOVA followed by Tukey's post hoc test, P>0.05). Finally, a fourth cohort of snails (n=34) that had received the traumatic stimulus 1 week previously showed learning and memory (Fig. 5B; one-way ANOVA followed by Tukey's post hoc test, P<0.001).
We next injected snails with a CBr antagonist (AM 251) before we applied the traumatic stimulus and then trained snails 2 days later (Fig. 6). In this cohort (Fig. 6B), learning and memory were still obstructed (one-way ANOVA followed by Tukey's post hoc test, P>0.05). In a second cohort, we injected the antagonist 4 h before the snails (n=14) received the traumatic stress. Again, when we trained these snails 2 days later, both learning and memory formation were obstructed (Fig. 6B; one-way ANOVA followed by Tukey's post hoc test, P>0.05). Finally, snails in a third cohort (n=20) were injected with the antagonist 1 h before receiving the severe traumatic stimulus, and were trained 4 days later. In this cohort of snails, learning and memory formation occurred (Fig. 6B; one-way ANOVA followed by Tukey's post hoc test, P<0.01). We performed control experiments with vehicle injection rather than AM 251, and found that in these control snails, learning and memory were still obstructed. Thus, AM 251 reduced the duration of the obstruction of memory formation by the traumatic stimulus.
We then determined the effect of the CBr antagonist on learning and memory formation in snails that were not subjected to the traumatic stress (Fig. 7). We first trained a cohort of snails (n=20) with the single 0.5 h training procedure, and LTM was not present when memory was tested 1 day later (two-tailed paired t-test, P>0.05; Fig. 7B, naive snails). In a second cohort of snails (n=20), we injected vehicle 1 h before training. In these snails, there was no enhancing effect on LTM formation, and these snails did not exhibit LTM (two-tailed paired t-test, P>0.05; Fig. 7B). Finally, in a third cohort of snails (n=24) injected with AM 251 1 h before training, LTM formation occurred 24 h (one-way ANOVA followed a Tukey's post hoc test, P<0.01) but not 48 h later (one-way ANOVA followed by Tukey's post hoc test, P>0.05; Fig. 7B).
Finally, we reanalysed all the data obtained in the first training procedures (i.e. TS1) in the three cohorts of snails shown in Fig. 7. There were no significant differences in the number of attempted pneumostome openings in TS1 of the three different cohorts, which each received a different treatment (one-way ANOVA followed by a Tukey's post hoc test, P>0.05). This shows that while AM 251 enhanced memory formation, it did not alter the sensitivity of the snails to the tactile stimulus used in the training procedure or the response to the hypoxic environment. Three further points need to be made here. The first is that this specific procedure and the operant conditioning procedures were performed by two different individuals, and this accounts in part for the difference in the number of attempted pneumostome openings. The second is that we also used a different batch of snails, as these experiments were performed some months after the other experiments. Finally, a different concentration of AM 251 was used in this particular set of experiments (18 µmol l−1 rather than 10 µmol l−1 used in the other experiments). Again, the concentration used here did not alter the responsiveness of snails to the tactile stimulus used.
Here, we showed that Lymnaea possess two G-protein-coupled receptor genes, which encode proteins closely related to well-characterized vertebrate CBrs, and these mRNAs are expressed in the Lymnaea CNS. The results obtained with the use of mammalian CBr agonists and antagonists suggest that cannabinoids play key roles in modulating learning and memory formation in Lymnaea. Further, we showed that subjecting snails to a severe traumatic tactile stressor blocks the ability of the snails to learn and form memory for at least 4 days. This obstruction of learning and memory was mimicked by application of the CBr agonist WIN 55. However, pre-treatment of snails with the CBr antagonist AM 251 decreased the effect of the severe traumatic stressor on the snails' ability to learn and remember. That is, the duration of the impact of the applied stressor was shortened. Finally, injection of AM 251 into snails led to enhancement of LTM formation, suggesting that there is an ongoing tonic suppression on the neuronal circuits that mediate LTM formation.
In the present study, we first identified two Lymnaea cDNAs encoding CBr-like G-protein-coupled receptors with structurally distinct third cytoplasmic loops. To our knowledge, this is the first report of G-protein-coupled receptor genes closely related to the mammalian CBrs in molluscs. Interestingly, there were tissue-specific differences of the expression patterns of Lymnaea CBr-like 1 and 2 mRNA. Lymnaea CBr-like 1 mRNA was ubiquitously transcribed, whereas Lymnaea CBr-like 2 mRNA was specifically expressed in the CNS (Fig. 2). As shown in Fig. 1B, two subtypes of CBrs, termed CB1 and CB2, have been identified in mammals, and the tissue specificities of these receptors have already been reported. CB1 is expressed mainly in the brain, but also in the lung, liver and kidney (Gérard et al., 1991); and CB2 is expressed mainly in the immune system and haematopoietic cells (Pacher and Mechoulam, 2011). The CBr-like genes of Lymnaea were widely expressed in the whole body, suggesting that the cannabinoid system functions in the various tissues in Lymnaea as it does in mammals. Although we did not examine the pharmacological properties of the gene products of Lymnaea CBr-like genes and their fine distribution in the Lymnaea CNS, we believe that they are strong candidates for targets of WIN 55 and AM 251 in the Lymnaea CNS. In addition to Lymnaea, our in silico analyses also indicated that similar CBr-like genes exist in three other molluscan species (A. californica, B. glabrata and L. gigantea) and a brachiopod species (L. anatine). A previous report in Aplysia showing that Δ9-tetrahydrocannabinol caused a depression in nerve cell excitability is consistent with the CBr-like genes being present in Aplysia (Acosta-Urquidi and Chase, 1975). The existence of the CBrs in another lophotrochozoan species, the leech (H. medicinalis), has previously been both physiologically and pharmacologically demonstrated (Li and Burrell, 2009).
Application of both the CBr antagonist AM 251 and 2-AG, the inhibitor of diacylglycerol lipase (which is necessary for the synthesis of the cannabinoid transmitter) blocked a change in synaptic plasticity in the leech (Li and Burrell, 2009). However, bath application of 2-AG and the CBr agonist CP55 940 induced a change in synaptic plasticity (Li and Burrell, 2009). It is unclear how the behaviour of the leech was altered by the induced changes in synaptic plasticity via CBr. In addition, Stefano et al. (1997) reported a partial cDNA sequence of the leech CBr (Stefano et al., 1997) but its deduced protein sequence partially resembles the mammalian CBr and adrenocorticotropic hormone receptor, and it was thought to be an artifact of PCR contamination (Elphick, 1998). Therefore, it is still unclear whether the leech has a complete set of genes involved in endocannabinoid signalling. CBr-like genes were also found in genomes of the tunicate chordates and lancelets (C. intestinalis and B. floridae; Elphick et al., 2003; Elphick, 2007), but their pharmacological properties have not yet been investigated. Further investigations involving identification and pharmacological characterization of invertebrate CBr-like proteins are required to understand molecular evolution and diversification of the endocannabinoid system in invertebrates.
In the mammalian nervous system, CBrs are mainly found on the presynaptic terminals of central and peripheral neurons, where they modulate the release of different excitatory and inhibitory neurotransmitters, which include glutamate and gamma-aminobutyric acid (GABA), respectively, as well as many other transmitter types (Slanina and Schweitzer, 2005; Roberto et al., 2010; Katona and Freund, 2012; Wamsteeker-Cusulin et al., 2014). In Lymnaea, whether the putative CBrs are also primarily localized to pre-synaptic sites remains to be determined, and is the subject of ongoing experimentation. However, the data to date from numerous studies show that CBrs are indispensable for various forms of synaptic plasticity that are thought to play key roles in learning and memory formation (Castillo et al., 2012; Lee et al., 2015).
In mammals, the activation of the endocannabinoid system in many cases suppresses the formation of both working memory and LTM, whereas both forms of the memory are enhanced when the endocannabinoid system is inhibited (Puighermanal et al., 2012). It must be kept in mind, however, that often contradictory results are obtained when CBr agonists or antagonists are used. This may be dependent on the emotional state of the animal when the stressor is applied at the time of training. The arousal level (think of the Yerkes–Dodson/Hebb curve) of the animal is a key factor in determining the effects of the endocannabinoid system on memory. For example, in rodents, WIN 55 administration enhanced long-term object recognition memory when they were trained under a high arousal condition, but WIN 55 was ineffective with low-arousal training (Campolongo et al., 2013; Basavarajappa et al., 2014). Thus, WIN 55 enhanced long-term object recognition memory only when trained under high arousal conditions (O'Shea et al., 2004; Morena and Campolongo, 2014; Wamsteeker-Cusulin et al., 2014). Whether a similar arousal-dependent situation exists in Lymnaea will be tested in the future.
We know in Lymnaea that different stressors tweak the ability to learn and form LTM (Lukowiak et al., 2014a). Some stressors (e.g. predator detection) lead to enhanced memory formation (Orr and Lukowiak, 2008), whereas other stressors (e.g. crowding) lead to suppression of memory formation (de Caigny and Lukowiak, 2008). In addition, a combination of stressors (e.g. crowding and low environmental calcium) leads to complete obstruction of learning and all forms of memory (i.e. short, intermediate and long-term memory) (Dalesman et al., 2011). It is unclear at present how the stressors actually cause memory formation to be enhanced or suppressed. However, at least for enhanced memory formation following exposure to certain stressors (e.g. predator detection) and bioactive agents (e.g. methamphetamine), epigenetic changes (i.e. DNA methylation) have been shown to play a necessary role (Lukowiak et al., 2014b). From the data obtained here, we are confident that CBrs play a key role in both modulating learning and the subsequent formation of memory as the result of snails experiencing different stressors.
Here, we showed that snails lose their ability to learn and form memory following exposure to the severe traumatic stressor. The inability to learn and form memory persists for at least 4 days and we found that snails regained their ability to learn and form LTM 1 week after the traumatic event. We mimicked this traumatic effect by injecting the CBr agonist WIN 55 into the snails, and we saw a similar time course for the snails' inability to learn and form memory. Consistent with our hypothesis that putative CBrs play a key role in the inability to learn and form memory after a traumatic event, we found that if we injected a CBr antagonist (i.e. AM 251) before the snails received the traumatic stimulus, the effect of the trauma on learning and memory was reduced. Finally, again consistent with our hypothesis that putative CBrs play a key role in memory formation, we found that injection of AM 251 into snails (i.e. untrained and not traumatized) enhanced their ability to form LTM.
In our previous investigations on the effect of different stressors on LTM formation, we have not seen such long-lasting effects as we saw here following either presentation of the severe traumatic stressor or injection of WIN 55. Typically, the induced modification persists for only a few hours. For example, experiencing a low calcium environment blocks LTM formation (Dalesman et al., 2011). However, even after snails experienced the low calcium environment for 1 week, it only took 1 h in the normal calcium environment to relieve the suppressive effect on LTM formation. We also showed that predator detection enhances LTM formation (Orr and Lukowiak, 2008). Again, however, snails had to be trained immediately after predator detection; a 1 h interval between detection of the predator and training in a predator-free environment was sufficient to negate the enhancing effects of predator detection on LTM formation. In contrast, here the effects of the traumatic experience were long lasting (4 days) as were the effects of WIN 55 injection. A working hypothesis to explain these findings is that the traumatic experience and/or the injection of WIN 55 causes an extreme fear state that is incompatible with learning and memory, as Ruehle et al. (2012) pointed out the endocannabinoid system regulates the establishment and maintenance of the fear state.
We conclude that the endocannabinoid system normally plays a role in learning and memory formation when some level of stress is involved in the snails. Neither WIN 55 nor AM 251 injection had a significant effect on homeostatic aerial respiratory behaviour. This suggests to us that the central terminations of the peripheral sensory neurons that drive this behaviour (Lukowiak et al., 1996; Karnik et al., 2012) via the 3-neuron central pattern generator (Syed et al., 1990, 1992) are not affected by the endocannabinoid agonists or antagonists. Our data also suggest that the tactile sensory neurons (Steffensen et al., 1995; Inoue et al., 1996) that convey mechanosensory information are also not affected by the agonists or antagonists. If these sensory neurons were affected, we would expect to see differences in the number of pokes delivered to the pneumostome during the operant conditioning procedure, when snails received either the agonist or antagonist.
We did not expect that injection of AM 251 into snails would cause enhancement of memory formation. This finding suggests that there may be a tonic inhibitory effect of the putative CBrs on the ability of snails to form LTM. Suppression of memory formation may occur to ‘keep down the cost’ of making memories that are of too little consequence for the snail. Data complementary to this idea have been obtained in rodent work previously (Takahashi et al., 2005; Terranova et al., 1996; Robinson et al., 2008). In those studies, blockade of the CBrs produces memory-enhancing effects on spatial and associative memory tasks. In addition, in birds, similar data were obtained with a slight twist (Shiflett et al., 2004), in that there was a ‘cost’ to the improved memory as a result of blocking the CBrs. This cost is that newly formed memories may be blocked from forming as a result of proactive interference. It may be that this notion will be more easily elucidated at the neuronal level in the Lymnaea model system. Further experimentation will be necessary to establish whether this is the correct hypothesis. It could be that stressors or bioactive agents that cause enhancement of LTM formation in Lymnaea do so via the putative CBrs. The present results will enable us to better understand how the endocannabinoid system is involved in learning acquisition and memory formation, and these results will also reveal how to control the traumatic stress effect.
Investigation: H.S., T.W., D.H., S.L., J.F.; Data curation: H.S., M.S., E.I., K.L.; Writing - original draft: H.S., T.W., D.H.; Writing - review & editing: M.S., E.I., K.L.; Funding acquisition: E.I., K.L.
This work was supported by a grant from Canadian Institutes of Health Research (MOP 64339) and Natural Sciences and Engineering Research Council of Canada (227993-2013) to K.L. and KAKENHI from Japan Society for the Promotion of Science (24657055 and 25291074) and Waseda University grants for Specific Research Projects in Japan (2016B-068, 2016B-069 and 2016S-037) to E.I.
The authors declare no competing or financial interests.