Spatial learning and memory flexibility are known to require long-term potentiation (LTP) and long-term depression (LTD), respectively, on a cellular basis. We previously showed that cyclin Y (CCNY), a synapse-remodeling cyclin, is a novel actin-binding protein and an inhibitory regulator of functional and structural LTP in vitro. In this study, we report that Ccny knockout (KO) mice exhibit enhanced LTP and weak LTD at Schaffer collateral-CA1 synapses in the hippocampus. In accordance with enhanced LTP, Ccny KO mice showed improved spatial learning and memory. However, although previous studies reported that normal LTD is necessary for memory flexibility, Ccny KO mice intriguingly showed improved memory flexibility, suggesting that weak LTD could exert memory flexibility when combined with enhanced LTP. At the molecular level, CCNY modulated spatial learning and memory flexibility by distinctively affecting the cofilin-actin signaling pathway in the hippocampus. Specifically, CCNY inhibited cofilin activation by original learning, but reversed such inhibition by reversal learning. Furthermore, viral-mediated overexpression of a phosphomimetic cofilin-S3E in hippocampal CA1 regions enhanced LTP, weakened LTD, and improved spatial learning and memory flexibility, thus mirroring the phenotype of Ccny KO mice. In contrast, the overexpression of a non-phosphorylatable cofilin-S3A in hippocampal CA1 regions of Ccny KO mice reversed the synaptic plasticity, spatial learning, and memory flexibility phenotypes observed in Ccny KO mice. Altogether, our findings demonstrate that LTP and LTD cooperatively regulate memory flexibility. Moreover, CCNY suppresses LTP while facilitating LTD in the hippocampus and negatively regulates spatial learning and memory flexibility through the control of cofilin-actin signaling, proposing CCNY as a learning regulator modulating both memorizing and forgetting processes.
Spatial learning is the process of recording the information required to navigate a location, and memory flexibility (also often called behavioral or cognitive flexibility) is the process of learning a new location that involves two courses, forming new memory and suppressing the previous memory. Synaptic plasticity, including long-term potentiation (LTP) and long-term depression (LTD), is generally considered as the cellular basis of learning and memory in the brain. Specifically, a myriad of preceding studies have reported the functional significance of LTP in mediating learning and memory with extensive molecular mechanisms [1, 2]. In addition, many studies have demonstrated that LTD is required for memory flexibility [3,4,5,6] and is also facilitated by learning per se [7,8,9,10,11,12]. Memory consolidation must be regulated within the physiological range through fine controls between memorizing and forgetting processes. Excessive memory strengthening and/or impaired forgetting processes likely contribute to a deficit in memory flexibility [13,14,15], which may lead to establishment of pathological conditions, such as post-traumatic stress disorder, autism spectrum disorder, schizophrenia, and Alzheimer’s disease [16,17,18,19]. Therefore, the maintenance of LTP, LTD, and memory strength within adequate physiological ranges is essential for the proper functioning of the brain in various cognitive and behavioral contexts.
Cyclin Y (CCNY), a cyclin family protein known to regulate the cell cycle in dividing cells [20,21,22], can regulate synaptic functions in terminally differentiated neuronal cells [23,24,25,26]. Specifically, CCNY is particularly enriched in the postsynaptic compartment in rat brains and inhibits the trafficking of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptor (AMPAR) to the synapse during glycine-induced LTP in cultured neurons . In addition, the overexpression of CCNY, biolistically transfected in organotypic hippocampal slices, inhibits LTP at the Schaffer collateral-Cornu Ammonis 1 (CA1) synapses . Furthermore, CCNY binds actin with high affinity and negatively regulates structural LTP by controlling actin dynamics . Upon observation that CCNY regulates functional and structural LTP in cultured hippocampal neurons and organotypic slice systems [25, 26], we hypothesized that CCNY also mediates hippocampus-dependent cognitive function in vivo, including learning and memory, by regulating synaptic plasticity.
A strong correlation between LTD and memory flexibility has been demonstrated by showing that mice with impaired LTD but with normal LTP and/or spatial learning exhibit impaired memory flexibility [3, 4, 6, 27]. Furthermore, LTD was suggested to play a role in the extinction of prior memory when new learning occurs during the process of memory flexibility . However, it is still unclear whether intact LTD is a particular requirement for memory flexibility, and whether LTP and LTD distinctively play a role in the memorizing and forgetting processes.
In the present study, Ccny KO mice exhibited enhanced LTP and reduced LTD at Schaffer collateral-CA1 synapses in the hippocampus, indicating a shifted spectrum in synaptic plasticity toward potentiation, and showed improved spatial learning in the Morris water maze (MWM) task as well as improved memory flexibility in the reversal learning phase of the MWM task and in a delayed nonmatch to place (DNMTP) T-maze task. Notably, transcriptome and immunoblot analyses revealed that LIM domain kinase 1 (LIMK1)-cofilin-actin signaling is differentially regulated in response to original learning and reversal learning in Ccny KO mice. Mice overexpressing a phosphomimetic cofilin-S3E in the hippocampal CA1 region, recapitulating the basal state of Ccny KO mice, also showed enhanced LTP and reduced LTD at Schaffer collateral-CA1 synapses as well as improved spatial learning and memory flexibility. Ccny KO mice overexpressing a non-phosphorylatable cofilin-S3A in the hippocampal CA1 regions showed reversion of the synaptic plasticity, spatial learning, and memory flexibility phenotypes observed in Ccny KO mice. Taken together, our findings demonstrate that reduced LTD can mediate normal and even enhanced memory flexibility when combined with normal and enhanced LTP, respectively, suggesting a crosstalk between LTP and LTD in the context of memory flexibility. In addition, this study suggests that CCNY is a learning regulator that modulates both memorizing and forgetting processes, and that spatial learning and memory flexibility are differentially modulated through the CCNY-mediated actin signaling pathway at the organismal level.
Neuronal excitability and basal excitatory synaptic properties remain unaltered in Ccny KO mice
We first investigated the role of CCNY in hippocampal synaptic plasticity and cognitive functions by characterizing electrophysiological features in Ccny KO mice . The frequency of action potential firing triggered by current injection in CA1 neurons was not altered in Ccny KO mice (Fig. 1a). The current-voltage relationship (Fig. 1b), input resistance (Fig. 1c), and sag ratio (Fig. 1d) remained unaffected, demonstrating that intrinsic neuronal excitability is normal in Ccny KO mice. Moreover, the paired-pulse ratio (PPR) was not significantly altered in Ccny KO mice (Fig. 1e), indicating that presynaptic release probability at Schaffer collateral-CA1 synapses in Ccny KO mice remains comparable to that of wild-type (WT) mice. This result is in agreement with the finding that CCNY expression is absent in the synaptic vesicle-enriched (LP2) presynaptic fraction, where a presynaptic protein synaptophysin exists in the mouse (Supplementary Fig. S1a) and rat forebrain . CCNY was enriched in postsynaptic fractions of the mouse forebrain, including the synaptic plasma membrane (SPM) and the postsynaptic density (PSD) fractions, where postsynaptic proteins, such as GluA1, a subunit of AMPARs, and PSD-95, are located (Supplementary Fig. S1a).
We next examined the function of AMPAR and N-methyl-D-aspartate (NMDA) receptor (NMDAR). The ratio of NMDAR- to AMPAR-mediated synaptic transmission (NMDA/AMPA ratio) was normal in Ccny KO mice (Fig. 1f), and the amplitude and frequency of miniature excitatory postsynaptic currents (mEPSCs) at Schaffer collateral-CA1 synapses were not altered (Fig. 1g). In addition, the input-output relationship between presynaptic stimulus intensity and the slope of the resulting field excitatory postsynaptic potentials (fEPSPs) against the presynaptic stimulus intensity (Fig. 1h) and the ratio of fEPSP slope to the fiber volley amplitude (Fig. 1i) were also not affected in Ccny KO mice. These results together indicate that basal excitatory synaptic transmission remained unchanged in Ccny KO mice. In support of these findings, the postsynaptic subcellular localizations of GluA1 and PSD-95 were not altered in the forebrain of Ccny KO mice (Supplementary Fig. S1b, c). The unaltered basal excitatory synaptic transmission in Ccny KO mice is rather contradictory to that reported in a previous study showing that a short hairpin RNA (shRNA)-mediated knockdown of CCNY increases the amplitude but not the frequency of mEPSCs in cultured rat hippocampal neurons . This difference could be due to the use of different systems; the chronic depletion of CCNY in Ccny KO mice might be systemically complemented in vivo, whereas 3- to 4-day knockdown of CCNY in cultured neuronal system rather exhibits an acute response to reduced CCNY levels. Together, these results suggest that the complete depletion of CCNY does not affect intrinsic neuronal excitability and basal excitatory synaptic transmission in the hippocampus.
CCNY regulates hippocampal LTP and LTD at Schaffer collateral-CA1 synapses
Hippocampal synaptic plasticity, including LTP and LTD, is widely considered as a fundamental mechanism underlying cognitive brain functions of learning and memory [5, 9, 10, 30,31,32,33]. We found that the magnitude of LTP induced by high-frequency stimulation (HFS) at Schaffer collateral-CA1 synapses was slightly but significantly enhanced in acute hippocampal slices prepared from Ccny KO mice compared to WT littermates (Fig. 1j, k). Furthermore, LTD induced by low-frequency stimulation (900 pulses, 1 Hz, 15 min) at Schaffer collateral-CA1 synapses was expressed with reduced magnitude in Ccny KO mice compared to WT littermates (Fig. 1l, m). In contrast, both voltage-gated calcium channel (VGCC)-dependent LTP (Supplementary Fig. S2a, b) and metabotropic glutamate receptor (mGluR)-dependent LTD (Supplementary Fig. S2c, d) were normal in Ccny KO mice. Collectively, these results indicate that Ccny KO mice possess a shifted spectrum of synaptic plasticity toward potentiation in the hippocampus and specifically affect the NMDAR-dependent form of synaptic plasticity, thereby potentially affecting hippocampus-dependent cognitive functions, such as spatial learning and memory flexibility.
Hippocampus-dependent spatial learning is enhanced in Ccny KO mice
We next investigated whether CCNY regulates hippocampus-dependent spatial learning and memory using the MWM task equipped with surrounding spatial cues (Fig. 2a). The latency to platform was significantly shorter on days 3, 4, and 5 in Ccny KO mice compared to WT littermates during the five consecutive training days (Fig. 2b), indicating an enhanced learning ability in Ccny KO mice. In addition, during the probe test on day 6, the latency (Fig. 2b) and travel distance (Fig. 2c) to the platform zone where the platform was originally located was significantly shorter in Ccny KO mice. Even though the number of entries to the target quadrant was not significantly different between WT and Ccny KO mice (Fig. 2d), the time spent in the target quadrant was increased in Ccny KO mice compared to WT littermates (Fig. 2e). Furthermore, the frequency of passing the platform zone was increased in Ccny KO mice compared to WT littermates (Fig. 2f). These results together indicate an enhancement of spatial memory ability in Ccny KO mice. In the visible platform test, which is a hippocampus-independent non-spatial learning task, the performance of Ccny KO mice was comparable to that of WT mice in terms of the latency to platform during the training sessions (Fig. 2g) and to the platform zone during the probe test (Fig. 2h), suggesting that the improved spatial learning and memory of Ccny KO mice in the MWM task (Fig. 2b−f) were not due to the enhanced motivation for escape or improved vision and/or motor skills.
In addition, the total distance moved, the time spent in the center, and the number of center entries in an open field task were not different between WT and Ccny KO mice (Supplementary Fig. S3a−h), further supporting that basal locomotion, anxiety, and willingness to explore were not affected in Ccny KO mice compared to WT littermates. Moreover, no significant difference was observed between WT and Ccny KO mice in a novel object recognition task (Supplementary Fig. S3i−k) used to test non-spatial memory [34, 35] or in a passive avoidance task (Supplementary Fig. S3l, m) used to assess aversive associative memory . These results together demonstrate that hippocampus-dependent spatial learning and memory is specifically improved in Ccny KO mice and suggest future studies exploring the roles of CCNY in other spectrums of learning behaviors and/or other behavioral/non-behavioral functions in other brain regions.
Memory flexibility is improved in Ccny KO mice and normal in Grin2b +/C456Y mice
LTD is associated with behavioral and memory flexibility [3,4,5,6, 9, 10, 27, 28]. Because hippocampal LTD was significantly altered in Ccny KO mice (Fig. 1l, m), we further investigated whether CCNY is involved in the regulation of memory flexibility. To this end, a spatial reversal learning task was conducted by moving the hidden platform to a new location (i.e., the quadrant opposite to the previously memorized quadrant) following the original learning task in the MWM. The latency to platform was significantly shorter on days 2 and 3 during the training sessions in Ccny KO mice (Fig. 2i). In addition, on day 4 of the reversal learning probe test, the latency (Fig. 2i) and travel distance (Fig. 2j) to the platform zone were also significantly reduced in Ccny KO mice, and the time spent in the target quadrant was significantly increased in Ccny KO mice compared to WT littermates (Fig. 2k). Furthermore, both the time spent in the previous target quadrant (Fig. 2k) and the frequency of entries to the previous target quadrant during the reversal learning probe test (Fig. 2l) were significantly decreased in Ccny KO mice compared to WT littermates. Both WT littermates and Ccny KO mice more frequently entered to the target quadrant (Fig. 2l) and target platform zone (Fig. 2m) than to the previous quadrant (Fig. 2l) and previous platform zone (Fig. 2m), respectively, but the degrees of increase in the frequency of entries to the target quadrant and target platform zone were much higher in Ccny KO mice than in WT littermates (Fig. 2l, m). Altogether, these results indicate that memory flexibility (or new learning and memory ability) is improved in Ccny KO mice.
We further investigated whether the lack of CCNY also influences the performance in the DNMTP T-maze task as an independent test to examine hippocampus-dependent memory flexibility [3, 27, 28, 37]. Similar to the reversal learning of MWM task which requires memory flexibility to learn and memorize new spatial information, the location of rewards for the choice runs in the DNMTP T-maze task alternates between the right and left arms in each trial, thus requiring memory flexibility to make a correct choice. Consistent with the enhanced reversal learning (Fig. 2i−m), Ccny KO mice also showed improved learning flexibility in the DNMTP T-maze task (Fig. 2n). In agreement with this finding, both travel distance (Supplementary Fig. S4a) and travel time (Supplementary Fig. S4b) taken to make a correct choice in the DNMTP T-maze task were shorter in Ccny KO mice compared to WT littermate. The body weights of the mice subjected to this task were maintained at approximately 80% of the initial body weights during the handling period throughout the experiments (Supplementary Fig. S4c). Because the animal behaviors in the open field task were not significantly different between WT and Ccny KO mice (Supplementary Fig. S4a−h), the improved memory flexibility in Ccny KO mice was not due to a change in basal locomotion, anxiety, or willingness to explore. Collectively, our findings indicate that even reduced LTD in Ccny KO mice (Fig. 1l, m) supports memory flexibility of both reversal learning in MWM (Fig. 2i−m) and the DNMTP T-maze task (Fig. 2n).
To further examine the notion that reduced LTD is sufficient to support memory flexibility, we performed a DNMTP T-maze task using GluN2B-C456Y-mutant mice (Grin2b+/C456Y) that exhibited reduced LTD induced by low-frequency stimulation but normal LTP at Schaffer collateral-CA1 synapses (i.e., presenting an altered spectrum in synaptic plasticity) along with normal original and reversal learning in MWM . Despite the reduced hippocampal LTD , the GluN2B-C456Y-mutant mice showed memory flexibility comparable to WT littermates in the DNMTP T-maze task (Fig. 2o), consistent with the normal reversal learning of MWM observed in GluN2B-C456Y-mutant mice , confirming that reduced LTD is sufficient to support memory flexibility. In agreement with this finding, both travel distance (Supplementary Fig. S4d) and travel time (Supplementary Fig. S4e) taken to make the correct choice in GluN2B-C456Y-mutant mice were comparable to WT littermates. The body weights of the mice were maintained at approximately 80% of their initial body weight during the handling period throughout the experiments (Supplementary Fig. S4f). Considering that the GluN2B-C456Y-mutant mice showed a difference in terms of normal LTP and normal original learning when compared to Ccny KO mice, the improved memory flexibility in Ccny KO mice seems to be accounted for by the enhanced LTP and improved original learning ability.
To further support the requirement of residual LTD in Ccny KO mice for the improved memory flexibility, the residual LTD in Ccny KO mice was completely blocked using a selective GluN2B antagonist and a specific LTD inhibitor, Ro25-6981 [4, 39]. Ro25-6981 treatment did not affect LTP (Fig. 3a, b), whereas it completely removed residual LTD in both WT and Ccny KO mice (Fig. 3c, d). In agreement with LTP results, Ro25-6981 administration to Ccny KO mice did not affect spatial learning in the MWM task, as it resulted in no significant changes in the latency to the platform (zone) (Fig. 3e), travel distance (Fig. 3f), number of entries to the target quadrant (Fig. 3g), time spent in the target quadrant (Fig. 3h), and frequency of passing the platform zone (Fig. 3i). In agreement with the notion that the requirement of the residual LTD in Ccny KO mice is required for the improved memory flexibility, latency to the platform (zone) was significantly delayed (Fig. 3j) and travel distance to the platform zone tended to increase with Ro25-6981 treatment (Fig. 3k). Both Ro25-6981-treated and untreated Ccny KO mice spent more time in the target quadrant than in the previous quadrant (Fig. 3l) and more frequently passed the target platform zone than the previous target platform zone (Fig. 3m). However, the degrees to which both the time spent in the target quadrant and the frequency of passing to the target platform zone increased were significantly lower in Ro25-6981-treated Ccny KO mice than in vehicle-treated Ccny KO mice (Fig. 3l, m). These data indicate that the enhanced memory flexibility in Ccny KO mice is dependent on the residual LTD. Taken together, these findings suggest that to improve memory flexibility, it is critical to facilitate not only the extinction or weakening of the previously encoded spatial memory, but also the formation of new spatial memory. The facilitation of the weakening of previously acquired memory information and the formation of new spatial memory are presumably mediated by weakened LTD and enhanced LTP, respectively, in Ccny KO mice.
Original learning upregulates genes related to the actin cytoskeleton, synaptic plasticity, or learning in WT mice with similar changes in untrained Ccny KO mice
Spatial learning and memory are correlated with spine morphology and plasticity [2, 40, 41], which requires actin remodeling [42,43,44,45,46,47,48]. CCNY is an actin-binding protein and regulates spine plasticity induced by LTP through the cofilin-actin signaling pathway , implying the functional significance of CCNY in modulating spatial learning and memory through the actin signaling pathway.
We found that the expression levels of Ccny mRNA detected by quantitative real-time polymerase chain reaction (qRT-PCR) (Fig. 4a) and CCNY protein detected by immunoblot analysis (Fig. 4b) were significantly reduced in the hippocampus after original learning in WT mice. Notably, improved spatial learning and memory ability in Ccny KO mice (Fig. 2b−f) and the decreased CCNY levels in WT mice following the original learning and memory task (Fig. 4a, b) together suggest a possible correlation between CCNY expression levels and cognitive function. Therefore, we aimed to determine whether Ccny KO mice in the naive state share molecular changes similar to those in WT mice subjected to original learning and memory. The potential candidates of molecular processes associated with original learning were examined by analyzing the transcriptome using high-throughput RNA sequencing. A total of 560 genes were differentially expressed when comparing WT mice under basal state with WT mice after original learning, in which 265 genes were upregulated and 295 genes were downregulated in the WT mice subjected to original learning (Fig. 4c; WT_Basal vs WT_OL). In untrained naive Ccny KO mice (KO_Basal), the expression levels of 314 genes were differentially regulated compared to untrained naive WT mice (WT_Basal), with 95 upregulated and 219 downregulated genes (Fig. 4c). Importantly, 29 genes that were commonly upregulated in both WT mice subjected to original learning and in untrained Ccny KO mice compared to WT mice under basal state (Fig. 4d, e) were significantly associated with the regulation of the actin cytoskeleton and the regulation of neuronal synaptic plasticity and learning, as revealed by the Kyoto Encyclopedia of Genes and Genomes enrichment analysis and the Gene Ontology analysis, respectively (Fig. 4f). These results strongly suggest that actin signaling plays a crucial role in mediating both the learning in WT mice and more improved learning ability in Ccny KO mice.
Distinct expression of actin-related proteins induced by original and reversal learning in WT and Ccny KO mice
Next, we profiled dynamic changes in the expression levels of genes closely associated with actin reorganization and spine morphology, including molecular components in the pathways of Rac, Cdc42, and Rho GTPases, in the hippocampus of WT and Ccny KO mice before and after original and reversal learning and memory tasks (Fig. 4g). Without learning (Basal), the level of Rac3 transcript was significantly higher in Ccny KO mice compared to WT mice, whereas the levels of Acta2 and Actb transcripts were significantly lower in Ccny KO mice (Fig. 4hi). Original learning reduced the levels of Rac3 and Ssh3 transcripts in Ccny KO mice without significant changes in WT mice. In contrast, original learning decreased the levels of Pak3 and Wasl transcripts in WT mice without significant changes in Ccny KO mice (Fig. 4hii). Reversal learning kept the reduction level of Rac3 in Ccny KO mice but significantly decreased the Rac3 level in WT mice when compared to original learning (Fig. 4hiii). Interestingly, changes in the Ssh3 transcript levels were reversed by reversal learning in both WT and KO mice when compared to original learning (Fig. 4hiii). Original learning increased the levels of Pak6, Limk1, Wasf3, and Acta1 transcripts in WT mice with no significant changes in Ccny KO mice (Fig. 4hii). The Limk1 and Wasf3 transcript levels remained unaltered after reversal learning versus original learning in both WT and Ccny KO mice, but the increase in the levels of Pak6 and Acta1 transcripts were greatly reduced after reversal learning in WT mice without significant changes in Ccny KO mice (Fig. 4hii−iv). These results together indicate that original and reversal learning processes are distinct at the molecular level and that such distinct molecular signaling is differentially modulated depending on the presence or absence of CCNY. Importantly, the expression of Actb transcripts increased after both original and reversal learning in WT mice, and Ccny KO mice exhibited more substantial increases in the Actb transcript level after original and reversal learning (Fig. 4hii−iv, i), highlighting the significance of the actin-mediated pathway underlying the improved original and reversal learning ability in Ccny KO mice, as observed in MWM behavioral tests (Figs. 2, 3).
CCNY affects the LIMK1-cofilin-actin pathway distinctively in original and reversal learning
Given that CCNY is an actin-binding protein and inhibits structural LTP by hindering the cofilin-actin pathway , we further investigated whether the improved original and reversal learning in Ccny KO (Figs. 2, 3) is associated with a change in the activity of cofilin, an actin-depolymerizing factor essential for actin reorganization [26, 48,49,50,51,52]. The levels of the phosphorylated inactive form of cofilin (p-cofilin) tended to increase after original learning in WT mice, but the p-cofilin level was significantly decreased in Ccny KO mice after original learning (Fig. 5a, b). Interestingly, Ccny KO mice without learning (basal) exhibited significantly higher level of the phosphorylated active form of LIMK1 (p-LIMK1), a kinase that phosphorylates cofilin, compared to the littermate WT mice, and also showed a tendency to increase the p-cofilin levels, albeit not significantly (Fig. 5b, c). Therefore, considering that Ccny KO mice under the basal state and WT mice subjected to original learning show comparable levels of p-LIMK1 and p-cofilin, it is likely that certain components of learning-induced molecular changes in the cofilin-actin pathway are already present in Ccny KO mice. This was supported by the finding that the actin cytoskeleton pathway was enriched by the set of genes that were commonly upregulated in WT mice after original learning and naive Ccny KO mice (Fig. 4f). In addition, reversal learning significantly increased p-cofilin levels in WT mice and rescued p-cofilin levels in Ccny KO mice to the level comparable to that observed in the basal condition (Fig. 5a, b). The level of p-LIMK1 was also altered similarly to the p-cofilin levels after original and reversal learning in both WT and Ccny KO mice (Fig. 5a, c), which was expected as LIMK1 is an upstream kinase for cofilin. In addition, the Limk1 transcript level was increased after original learning in WT mice; however, this was not observed in Ccny KO mice (Fig. 4hii). Taken together, these results suggest that a decrease in the phosphorylation levels of cofilin (thus leading to elevated cofilin activity that facilitates actin depolymerization and enhances actin dynamics), and dynamic reversal of cofilin activity likely serve as the molecular basis for the improved original and reversal learning in Ccny KO mice, respectively (Fig. 5).
The phosphorylation level of cofilin affects synaptic plasticity, spatial learning, and memory flexibility
Based on the RNA-seq data, we further determined whether the cofilin activity in Ccny KO mice is a critical factor that regulates synaptic plasticity, spatial learning and memory flexibility. We therefore investigated whether the increased p-cofilin level observed in Ccny KO mice at the basal state (Fig. 5a, b) is sufficient to modulate hippocampal synaptic plasticity, spatial learning, and memory flexibility and whether decreasing p-cofilin levels in Ccny KO mice could revert the observed altered synaptic plasticity and hippocampus-dependent learning. To test these possibilities, we generated another mouse model whose spectrum in synaptic response might be affected by increasing the phosphorylation levels of cofilin in WT mice as examined in Ccny KO mice through the stereotaxic injection of AAV expressing cofilin-S3E, a phosphomimetic mutant of cofilin, or by decreasing the phosphorylation levels of cofilin through the stereotaxic injection of AAV expressing cofilin-S3A, a non-phosphorylated active form of cofilin, into the CA1 region of Ccny KO mice (Fig. 5d). Intriguingly, viral-mediated overexpression of cofilin-S3E-EGFP reduced the expression levels of endogenous CCNY in the primary hippocampal neuron culture (Supplementary Fig. S5a, b) as well as in the CA1 region of hippocampal tissue (Supplementary Fig. S5d, e) compared to the controls. In addition, overexpression of phosphomimetic cofilin-S3E decreased the ratio of p-cofilin to the total levels of endogenous cofilin in the cultured hippocampal neurons (Supplementary Fig. S5a, c), as well as in the CA1 region of the hippocampus (Supplementary Fig. S5d, f). The AAV-mediated overexpression of cofilin-S3A-EGFP in the CA1 region increased the ratio of p-cofilin to the total levels of endogenous cofilin in WT (Supplementary Fig. S5g, i) and Ccny KO mice (Supplementary Fig. S5g, j) without affecting endogenous CCNY protein levels in WT mice (Supplementary Fig. S5g, h). The reduction and increase in the ratio of p-cofilin to the total endogenous cofilin levels observed upon the overexpression of cofilin-S3E and cofilin-S3A, respectively, might be caused by an intrinsic feedback mechanism required to maintain the total levels of p-cofilin, including endogenous p-cofilin and phosphomimetic cofilin-S3E or non-phosphorylatable cofilin-S3A, within appropriate ranges.
Similar to Ccny KO mice with increased p-cofilin levels (Fig. 5a, b), overexpression of cofilin-S3E-EGFP in the CA1 region significantly enhanced LTP (Fig. 5e, f) and reduced LTD (Fig. 5g, h) at Schaffer collateral-CA1 synapses in acute hippocampal slices compared to the control slices overexpressing EGFP in the CA1 region. In the original learning in the MWM task, the latency to the platform was significantly shorter on training days 4 and 5 in mice whose CA1 regions were infected with AAV-Cofilin-S3E-EGFP than in mice infected with AAV-EGFP (Fig. 5i). In addition, during the probe test on day 6, the latency (Fig. 5i) and travel distance (Fig. 5j) to the platform zone were significantly shorter in mice overexpressing cofilin-S3E-EGFP in the CA1 region. Even though the number of entries to the target quadrant was not significantly different between cofilin-S3E-EGFP-overexpressing and EGFP-overexpressing mice (Supplementary Fig. S6a), the time spent in the target quadrant was increased in mice overexpressing cofilin-S3E-EGFP compared to the mice overexpressing EGFP (Fig. 5k). The frequency of passing the platform zone tended to increase and was significantly increased in the third trial in mice overexpressing cofilin-S3E-EGFP compared to the control mice (Fig. 5l).
In a spatial reversal learning task, the latency to the platform was significantly shorter during the training sessions in mice overexpressing cofilin-S3E-EGFP compared to the control mice overexpressing EGFP (Fig. 5m). In addition, on day 4 of the reversal learning probe test, the latency (Fig. 5m) and travel distance (Fig. 5n) to the platform zone were also significantly reduced in mice overexpressing cofilin-S3E-EGFP. During the reversal learning probe test, mice overexpressing cofilin-S3E-EGFP spent more time in the target quadrant than in the previous target quadrant compared to the control mice overexpressing EGFP (Fig. 5o). The frequency of entries to the previous target quadrant relative to the current target quadrant (Fig. 5p) and the frequency of passing the previous target platform zone relative to the current target platform zone (Fig. 5q) were decreased in mice overexpressing cofilin-S3E-EGFP compared to the control mice. These results indicate that memory flexibility is improved in mice with enhanced phosphorylation levels of cofilin.
In agreement with the enhanced reversal learning (Fig. 5m−q), an improved learning flexibility in the DNMTP T-maze task was observed in mice overexpressing cofilin-S3E-EGFP (Fig. 5r). Both travel distance (Supplementary Fig. S6b) and travel time (Supplementary Fig. S6c) taken to make the correct choice in mice overexpressing cofilin-S3E-EGFP were comparable to those of the control mice. The body weights of the mice subjected to this task were maintained at approximately 80−87% of the initial body weight during the handling period throughout the experiments (Supplementary Fig. S6d).
We next examined whether overexpression of the non-phosphorylatable form of cofilin would reduce the synaptic and learning behavioral phenotypes of Ccny KO mice. Overexpression of cofilin-S3A-EGFP in the CA1 region did not affect LTP (Fig. 6a, b)  and LTD in WT littermates (Fig. 6c, d), whereas the overexpression of cofilin-S3A-EGFP in Ccny KO mice reverted the enhanced LTP (Fig. 6a, b) and reduced LTD (Fig. 6c, d) that were observed in Ccny KO mice to levels similar to those observed in WT littermates overexpressing EGFP in the CA1 region. In agreement with LTP results, the overexpression of cofilin-S3A-EGFP in Ccny KO mice slightly but significantly increased latency time to arrive at the platform (zone) during the original learning period in the MWM task (Fig. 6e) compared to that observed with EGFP-overexpressing control Ccny KO mice. Even though time spent in the quadrant was not altered in Ccny KO mice overexpressing cofilin-S3A-EGFP (Fig. 6h), travel distance to the platform zone tended to increase in the first trial (T1) and significantly increased in the second trial (T2) compared to that observed with EGFP-overexpressing control Ccny KO mice (Fig. 6f). The number of entries to the target quadrant (Fig. 6g) and the frequency of passing the platform zone in T1 and T2 (Fig. 6i) also tended to decrease with cofilin-S3A-EGFP-overexpressing Ccny KO mice compared to EGFP-overexpressing Ccny KO mice. Furthermore, the shortened latency to the platform (zone) during the reversal learning in the MWM task observed with Ccny KO mice reverted with the overexpression of cofilin-S3A-EGFP (Fig. 6j). Even though Ccny KO mice overexpressing cofilin-S3A-EGFP spent less time in the previous quadrant and more time in the target quadrant (Fig. 6l), those mice tended to show a longer travel distance (Fig. 6k) and further showed a significant reduction in the frequency of passing the target quadrant (Fig. 6m).
In summary, these results indicate that the phosphorylation levels of cofilin are sufficient to modulate synaptic plasticity, spatial learning, and memory flexibility in Ccny KO mice. In addition, the data from Ccny KO, cofilin-S3E-EGFP-overexpressing, or cofilin-S3A-EGFP-overexpressing mice indicate that CCNY expression level and cofilin phosphorylation signaling reciprocally affect each other, which ultimately regulates synaptic plasticity, spatial learning, and memory flexibility. Furthermore, all the data from Ccny KO, Ccny KO treated with Ro25-6981, GluN2B-C456Y-mutant, or cofilin-S3E-EGFP-overexpressing mice indicate that memory flexibility does not necessarily require intact LTD, but reduced LTD (examined in Ccny KO, GluN2B-C456Y-mutant, or cofilin-S3E-EGFP-overexpressing mouse models) supports memory flexibility when combined with normal LTP/spatial learning (as in GluN2B-C456Y-mutant mice) or enhanced LTP/improved spatial learning (as in Ccny KO or cofilin-S3E-EGFP-overexpressing mice). Therefore, CCNY is proposed as a learning regulator that modulates both memorizing (in the processes of both spatial learning and memory flexibility) and forgetting (in the process of memory flexibility) processes.
Here, we demonstrate that reduced LTD, but not necessarily intact LTD, is able to mediate memory flexibility when LTP is intact or enhanced, suggesting a possible intercommunication between LTP and LTD in the context of new learning and existing memory weakening during memory flexibility (Supplementary Fig. S7①). In addition, the expression levels of genes associated with actin reorganization were differentially regulated by original and reversal learning in a CCNY-dependent manner (Fig. 4g−i). Both original (spatial learning, Supplementary Fig. S7, red line) and reversal learning (memory flexibility, Supplementary Fig. S7, blue line) increased p-LIMK1 and p-cofilin levels (Fig. 5). Interestingly, CCNY facilitated phosphorylation of LIMK1 (active) and cofilin (inactive) following original learning (Fig. 5; Supplementary Fig. S7②), whereas it inhibited the phosphorylation of LIMK1 and cofilin following reversal learning (Fig. 5; Supplementary Fig. S7③), indicating that spatial learning and memory flexibility are distinctively modulated by the LIMK1-cofilin-actin pathway in a CCNY-dependent manner. Furthermore, increasing the phosphorylation level of cofilin, as observed in the Ccny KO mice, in the CA1 regions of the hippocampi by the transduction of AAV overexpressing cofilin-S3E plays a causative role in the changes observed in synaptic plasticity, including enhancement of LTP (Fig. 5e, f; Supplementary Fig. S7④, black thin arrow) and reduction of LTD (Fig. 5g, h; Supplementary Fig. S7⑤, black dotted line), as well as for the enhancement of spatial learning (Fig. 5i−l; Supplementary Fig. S7④, black thick arrow) and memory flexibility (Fig. 5m−r; Supplementary Fig. S7⑤, black thick arrow). The present study also suggests that CCNY is a learning regulator that modulates both the memorizing and forgetting processes. Further studies will be needed to determine how cofilin phosphorylation levels are regulated differently in the presence or absence of CCNY in the context of synaptic plasticity and cognitive functions such as spatial learning and memory flexibility.
LTP serves as a cellular mechanism underlying learning and memory formation in the brain and is bi-directionally modulated by enhancers and suppressors. LTD is also required for spatial reversal learning and memory (i.e., memory flexibility) [3,4,5,6, 54], and learning per se facilitates LTD [7,8,9,10,11,12]. Memory extinction is a process that involves new learning that suppresses the expression of prior memory , and importantly, both the time spent in the previous target quadrant (Fig. 2k) and the frequency of entries to the previous target quadrant (Fig. 2l) during the probe test after reversal learning were significantly deceased in Ccny KO mice, thereby implying the facilitated extinction of the initial spatial memory in Ccny KO mice. This facilitated memory extinction, a weakening of the previously acquired memory information required for new learning or memory flexibility, could be mediated by LTD expressed in Ccny KO mice (Fig. 1l, m). Importantly, however, the magnitude of LTD expression was reduced in Ccny KO mice compared to WT littermates (Fig. 1l, m), suggesting that modest LTD, but not necessarily strongly expressed LTD, could be sufficient for weakening the existing memory. Moreover, considering that memory flexibility also involves new learning in addition to the extinction of prior memory, and that learning is improved in Ccny KO mice, it is possible that both LTD and LTP play a critical role in facilitating memory flexibility.
Even though a complete blockade of LTD (but with normal spatial learning and memory ability) was reported to be associated with impaired memory flexibility [3, 27], our observations indicate that even weak LTD, together with enhanced spatial learning and memory ability, can exhibit memory flexibility. These findings add more mechanistic insights into the current simple view of the relationship between LTD and memory flexibility. In memory flexibility, which comprises new learning and a weakening of previous memory, new learning may be mediated by LTP, and a weakening of previous memory may be mediated by LTD. Our results suggest that these two processes of new learning and weakening of previous memory are mechanistically interconnected and that the precise tuning of these two processes can determine the degree of memory flexibility. Along with the findings of the present study, several previous studies support this notion. Double KO mice lacking both type 1 and 8 adenylyl cyclases show defective LTP and LTD, along with a deficit in both original and reversal learning . In addition, mice overexpressing type 1 adenylyl cyclase exhibit enhanced LTP and impaired LTD at the Schaffer collateral-CA1 synapses and show superior original as well as reversal learning . Therefore, in some instances, memory flexibility can be exhibited by the combination of enhanced LTP and reduced LTD. Considering that Ccny KO mice with enhanced LTP and moderate LTD show superior original and reversal learning, it is likely that enhanced LTP improves certain aspects of memory flexibility, presumably by facilitating new memory formation, which may compensate for the weakened LTD-derived ineffective extinction of the previously established memory.
Memory formation should be adequately regulated within physiological ranges, which requires maintaining the balance between memorizing (acquisition and consolidation) and forgetting (extinction) processes. Learning and memory have been extensively investigated in terms of acquisition and consolidation [13, 58,59,60,61], and the extinction and reconsolidation processes have also received attention in the field of learning and memory [14, 62,63,64]. Failure to maintain physiological memory strength may cause pathological malfunctioning of learning and memory, such as exhibition of weak or strong memory beyond the physiological range, resulting in deficits in learning and memory flexibility. In the present study, we demonstrate that CCNY plays a pivotal role in both spatial learning and memory flexibility, and Ccny KO mice will be useful as learning and memory research models to investigate the brain circuitry involved in spatial learning and memory flexibility in physiology and pathology.
Materials and methods
All materials and methods, except animals and data availability, are included in the Supplementary information. A summary of the statistical analysis in each figure is provided in Supplementary Table S3.
All experiments involving animals were performed in accordance with the guidelines and regulations of the Korea Institute of Science and Technology (KIST) or Korea Advanced Institute of Science and Technology (KAIST). All experimental protocols were approved by the KIST or KAIST Institutional Animal Care and Use Committee (approval number 2019-054 at KIST; KA2012-19 at KAIST).
RNA-seq data were deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) and the Korean Nucleotide Archive (KoNA) under the accession numbers GSE129324 and PRJKA220513, respectively.
Bliss TV, Collingridge GL. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 1993;361:31–9.
Bannerman DM, Sprengel R, Sanderson DJ, McHugh SB, Rawlins JN, Monyer H, et al. Hippocampal synaptic plasticity, spatial memory and anxiety. Nat Rev Neurosci. 2014;15:181–92.
Kim JI, Lee HR, Sim SE, Baek J, Yu NK, Choi JH, et al. PI3Kgamma is required for NMDA receptor-dependent long-term depression and behavioral flexibility. Nat Neurosci. 2011;14:1447–54.
Dong Z, Bai Y, Wu X, Li H, Gong B, Howland JG, et al. Hippocampal long-term depression mediates spatial reversal learning in the Morris water maze. Neuropharmacology 2013;64:65–73.
Connor SA, Wang YT. A place at the table: LTD as a mediator of memory genesis. Neuroscientist 2016;22:359–71.
Kim DH, Kang M, Kim CH, Huh YH, Cho IH, Ryu HH, et al. SPIN90 modulates long-term depression and behavioral flexibility in the hippocampus. Front Mol Neurosci. 2017;10:295.
Lemon N, Manahan-Vaughan D. Dopamine D1/D5 receptors gate the acquisition of novel information through hippocampal long-term potentiation and long-term depression. J Neurosci. 2006;26:7723–9.
Wong TP, Howland JG, Robillard JM, Ge Y, Yu W, Titterness AK, et al. Hippocampal long-term depression mediates acute stress-induced spatial memory retrieval impairment. Proc Natl Acad Sci USA. 2007;104:11471–6.
Ge Y, Dong Z, Bagot RC, Howland JG, Phillips AG, Wong TP, et al. Hippocampal long-term depression is required for the consolidation of spatial memory. Proc Natl Acad Sci USA. 2010;107:16697–702.
Dong Z, Gong B, Li H, Bai Y, Wu X, Huang Y, et al. Mechanisms of hippocampal long-term depression are required for memory enhancement by novelty exploration. J Neurosci. 2012;32:11980–90.
Kakegawa W, Katoh A, Narumi S, Miura E, Motohashi J, Takahashi A, et al. Optogenetic control of synaptic AMPA receptor endocytosis reveals roles of LTD in motor learning. Neuron 2018;99:985–98.e6
Kono M, Kakegawa W, Yoshida K, Yuzaki M. Interneuronal NMDA receptors regulate long-term depression and motor learning in the cerebellum. J Physiol. 2018;597:903–20.
McGaugh JL. Memory−a century of consolidation. Science 2000;287:248–51.
Davis RL, Zhong Y. The biology of forgetting-a perspective. Neuron 2017;95:490–503.
Mariottini C, Munari L, Gunzel E, Seco JM, Tzavaras N, Hansen J, et al. Wilm’s tumor 1 promotes memory flexibility. Nat Commun. 2019;10(Aug):3756.
Albert MS. Cognitive and neurobiologic markers of early Alzheimer disease. Proc Natl Acad Sci USA. 1996;93:13547–51.
Yerys BE, Wallace GL, Harrison B, Celano MJ, Giedd JN, Kenworthy LE. Set-shifting in children with autism spectrum disorders: reversal shifting deficits on the Intradimensional/Extradimensional Shift Test correlate with repetitive behaviors. Autism 2009;13:523–38.
Orellana G, Slachevsky A. Executive functioning in schizophrenia. Front Psychiatry. 2013;4:35.
Keith J, Velezmoro R, O’Brien C. Correlates of cognitive flexibility in veterans seeking treatment for posttraumatic stress disorder. J Nerv Ment Dis. 2015;203:287–93.
Evans T, Rosenthal ET, Youngblom J, Distel D, Hunt T. Cyclin: a protein specified by maternal mRNA in sea urchin eggs that is destroyed at each cleavage division. Cell 1983;33:389–96.
Lew DJ, Dulic V, Reed SI. Isolation of three novel human cyclins by rescue of G1 cyclin (Cln) function in yeast. Cell 1991;66:1197–206.
Noble ME, Endicott JA, Brown NR, Johnson LN. The cyclin box fold: protein recognition in cell-cycle and transcription control. Trends Biochem Sci. 1997;22:482–7.
Ou CY, Poon VY, Maeder CI, Watanabe S, Lehrman EK, Fu AK, et al. Two cyclin-dependent kinase pathways are essential for polarized trafficking of presynaptic components. Cell 2010;141:846–58.
Park M, Watanabe S, Poon VY, Ou CY, Jorgensen EM, Shen K. CYY-1/cyclin Y and CDK-5 differentially regulate synapse elimination and formation for rewiring neural circuits. Neuron 2011;70:742–57.
Cho E, Kim D-H, Hur Y-N, Whitcomb DJ, Regan P, Hong J-H, et al. Cyclin Y inhibits plasticity-induced AMPA receptor exocytosis and LTP. Sci Rep. 2015;5:12624.
Hwang H, Hur YN, Sohn H, Seo J, Hong JH, Cho E, et al. Cyclin Y, a novel actin-binding protein, regulates spine plasticity through the cofilin-actin pathway. Prog Neurobiol. 2021;198:101915.
Mills F, Bartlett TE, Dissing-Olesen L, Wisniewska MB, Kuznicki J, Macvicar BA, et al. Cognitive flexibility and long-term depression (LTD) are impaired following beta-catenin stabilization in vivo. Proc Natl Acad Sci USA. 2014;111:8631–6.
Nicholls RE, Alarcon JM, Malleret G, Carroll RC, Grody M, Vronskaya S, et al. Transgenic mice lacking NMDAR-dependent LTD exhibit deficits in behavioral flexibility. Neuron 2008;58:104–17.
Hwang H, Seo J, Choi Y, Cho E, Sohn H, Jang J, et al. Ccny knockout mice display an enhanced susceptibility to kainic acid-induced epilepsy. Pharm Res. 2020;160:105100.
Bear MF. A synaptic basis for memory storage in the cerebral cortex. Proc Natl Acad Sci USA. 1996;93:13453–9.
Kemp A, Manahan-Vaughan D. Hippocampal long-term depression and long-term potentiation encode different aspects of novelty acquisition. Proc Natl Acad Sci USA. 2004;101:8192–7.
Herring BE, Nicoll RA. Long-term potentiation: from CaMKII to AMPA receptor trafficking. Annu Rev Physiol. 2016;78:351–65.
Park M. AMPA receptor trafficking for postsynaptic potentiation. Front Cell Neurosci. 2018;12:361.
Lopez-Atalaya JP, Valor LM, Barco A. Epigenetic factors in intellectual disability: the Rubinstein-Taybi syndrome as a paradigm of neurodevelopmental disorder with epigenetic origin. Prog Mol Biol Transl Sci. 2014;128:139–76.
Cohen SJ, Stackman RW Jr. Assessing rodent hippocampal involvement in the novel object recognition task. A review. Behav Brain Res. 2015;285:105–17.
LeDoux JE, Moscarello J, Sears R, Campese V. The birth, death and resurrection of avoidance: a reconceptualization of a troubled paradigm. Mol Psychiatry. 2017;22:24–36.
McHugh SB, Niewoehner B, Rawlins JN, Bannerman DM. Dorsal hippocampal N-methyl-D-aspartate receptors underlie spatial working memory performance during non-matching to place testing on the T-maze. Behav Brain Res. 2008;186:41–7.
Shin W, Kim K, Serraz B, Cho YS, Kim D, Kang M, et al. Early correction of synaptic long-term depression improves abnormal anxiety-like behavior in adult GluN2B-C456Y-mutant mice. PLoS Biol. 2020;18:e3000717.
Liu L, Wong TP, Pozza MF, Lingenhoehl K, Wang Y, Sheng M, et al. Role of NMDA receptor subtypes in governing the direction of hippocampal synaptic plasticity. Science 2004;304:1021–4.
Martin SJ, Grimwood PD, Morris RG. Synaptic plasticity and memory: an evaluation of the hypothesis. Annu Rev Neurosci. 2000;23:649–711.
Hayashi ML, Choi SY, Rao BS, Jung HY, Lee HK, Zhang D, et al. Altered cortical synaptic morphology and impaired memory consolidation in forebrain- specific dominant-negative PAK transgenic mice. Neuron 2004;42:773–87.
Threadgill R, Bobb K, Ghosh A. Regulation of dendritic growth and remodeling by Rho, Rac, and Cdc42. Neuron 1997;19:625–34.
Matus A. Actin-based plasticity in dendritic spines. Science 2000;290:754–8.
Parnas D, Haghighi AP, Fetter RD, Kim SW, Goodman CS. Regulation of postsynaptic structure and protein localization by the Rho-type guanine nucleotide exchange factor dPix. Neuron 2001;32:415–24.
Penzes P, Beeser A, Chernoff J, Schiller MR, Eipper BA, Mains RE, et al. Rapid induction of dendritic spine morphogenesis by trans-synaptic ephrinB-EphB receptor activation of the Rho-GEF kalirin. Neuron 2003;37:263–74.
Kim Y, Sung JY, Ceglia I, Lee KW, Ahn JH, Halford JM, et al. Phosphorylation of WAVE1 regulates actin polymerization and dendritic spine morphology. Nature 2006;442:814–7.
Nikolic M. The Pak1 kinase: an important regulator of neuronal morphology and function in the developing forebrain. Mol Neurobiol. 2008;37:187–202.
Hotulainen P, Hoogenraad CC. Actin in dendritic spines: connecting dynamics to function. J Cell Biol. 2010;189:619–29.
Fukazawa Y, Saitoh Y, Ozawa F, Ohta Y, Mizuno K, Inokuchi K. Hippocampal LTP is accompanied by enhanced F-actin content within the dendritic spine that is essential for late LTP maintenance in vivo. Neuron 2003;38:447–60.
Gu J, Lee CW, Fan Y, Komlos D, Tang X, Sun C, et al. ADF/cofilin-mediated actin dynamics regulate AMPA receptor trafficking during synaptic plasticity. Nat Neurosci. 2010;13:1208–15.
Bosch M, Castro J, Saneyoshi T, Matsuno H, Sur M, Hayashi Y. Structural and molecular remodeling of dendritic spine substructures during long-term potentiation. Neuron 2014;82:444–59.
Borovac J, Bosch M, Okamoto K. Regulation of actin dynamics during structural plasticity of dendritic spines: Signaling messengers and actin-binding proteins. Mol Cell Neurosci. 2018;91:122–30.
Havekes R, Park AJ, Tudor JC, Luczak VG, Hansen RT, Ferri SL, et al. Sleep deprivation causes memory deficits by negatively impacting neuronal connectivity in hippocampal area CA1. Elife. 2016;5:e13424.
Xiong CH, Liu MG, Zhao LX, Chen MW, Tang L, Yan YH, et al. M1 muscarinic receptors facilitate hippocampus-dependent cognitive flexibility via modulating GluA2 subunit of AMPA receptors. Neuropharmacology 2019;146:242–51.
Haubrich J, Machado A, Boos FZ, Crestani AP, Sierra RO, Alvares LO, et al. Enhancement of extinction memory by pharmacological and behavioral interventions targeted to its reactivation. Sci Rep. 2017;7:10960.
Zhang M, Storm DR, Wang H. Bidirectional synaptic plasticity and spatial memory flexibility require Ca2+-stimulated adenylyl cyclases. J Neurosci. 2011;31:10174–83.
Zhang M, Wang H. Mice overexpressing type 1 adenylyl cyclase show enhanced spatial memory flexibility in the absence of intact synaptic long-term depression. Learn Mem. 2013;20:352–7.
Menzel R, Muller U. Learning and memory in honeybees: from behavior to neural substrates. Annu Rev Neurosci. 1996;19:379–404.
Jonides J, Lewis RL, Nee DE, Lustig CA, Berman MG, Moore KS. The mind and brain of short-term memory. Annu Rev Psychol. 2008;59:193–224.
Kandel ER, Dudai Y, Mayford MR. The molecular and systems biology of memory. Cell 2014;157:163–86.
Dudai Y, Karni A, Born J. The consolidation and transformation of memory. Neuron 2015;88:20–32.
Dunsmoor JE, Niv Y, Daw N, Phelps EA. Rethinking extinction. Neuron 2015;88:47–63.
Giustino TF, Maren S. The role of the medial prefrontal cortex in the conditioning and extinction of fear. Front Behav Neurosci. 2015;9:298.
Hemstedt TJ, Lattal KM, Wood MA. Reconsolidation and extinction: Using epigenetic signatures to challenge conventional wisdom. Neurobiol Learn Mem. 2017;142:55–65.
We thank S. Kim for providing excellent technical support with the maintenance of the mouse line, including genotyping. This work was supported by the National Research Foundation of Korea (NRF) grants funded by the Korean government (MSIT) (NRF-2018M3C7A1021848 and NRF-2021R1A2C2004606) and by the Korea Institute of Science and Technology Institutional Program (No. 2E31512).
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Seo, J., Hwang, H., Sohn, H. et al. Cyclin Y regulates spatial learning and memory flexibility through distinct control of the actin pathway. Mol Psychiatry (2022). https://doi.org/10.1038/s41380-022-01877-0