Cyclin Y inhibits plasticity-induced AMPA receptor exocytosis and LTP

Cyclin Y (CCNY) is a member of the cyclin protein family, known to regulate cell division in proliferating cells. Interestingly, CCNY is expressed in neurons that do not undergo cell division. Here, we report that CCNY negatively regulates long-term potentiation (LTP) of synaptic strength through inhibition of AMPA receptor trafficking. CCNY is enriched in postsynaptic fractions from rat forebrain and is localized adjacent to postsynaptic sites in dendritic spines in rat hippocampal neurons. Using live-cell imaging of a pH-sensitive AMPA receptor, we found that during LTP-inducing stimulation, CCNY inhibits AMPA receptor exocytosis in dendritic spines. Furthermore, CCNY abolishes LTP in hippocampal slices. Taken together, our findings demonstrate that CCNY inhibits plasticity-induced AMPA receptor delivery to synapses and thereby blocks LTP, identifying a novel function for CCNY in post-mitotic cells.

CCNY localizes in dendritic spines at perisynapstic sites, negatively regulates plasticity-induced AMPA receptor delivery to synapses and thereby blocks LTP. Our findings reveal CCNY as an inhibitory regulator of synaptic plasticity of hippocampal LTP in the vertebrate central nervous system.

Results
CCNY is enriched in postsynaptic fractions and is localized adjacent to postsynaptic sites. We first investigated whether CCNY is expressed in the mammalian brain. There has been a previous report on the CCNY mRNA level in various human tissues 7 . In addition, in situ hybridization shows CCNY expression in brain regions, including hippocampus, cortex, striatum, olfactory bulb, and cerebellum (Supplementary Fig. 1; the Allen Brain Atlas). However, protein expression of CCNY in brain has not been examined. Using immunoblot analysis with several brain region homogenates, we found that CCNY is expressed throughout the brain with relatively higher levels in the striatum and hippocampus (Fig. 1c). In addition, CCNY is expressed in the dentate gyrus (DG), Cornu Ammonis 3 (CA3), and CA1 region of the hippocampus (Fig. 1d). CCNY protein expression in the hippocampus increases over development in vivo (Fig. 1e) and in vitro (Fig. 1f). We next asked whether CCNY is located at synapses. For this purpose, we performed subcellular fractionation from rat forebrains and found that CCNY is enriched in postsynaptic fractions (Fig. 1g). To examine the subcellular localization of CCNY relative to postsynaptic density (PSD) in dendritic spines, we co-expressed CCNY wild-type (CCNY-WT) and PSD-95, a postsynaptic scaffolding protein in cultured hippocampal neurons. Confocal imaging (Fig. 1h) and 3D rendering ( Fig. 1hi-hv) revealed that CCNY is localized in dendritic spines where it concentrates adjacent to the PSD as labeled by PSD-95.
To further examine CCNY function in AMPA receptor-mediated synaptic transmission, we performed surface immunostaining of the AMPA receptor subunit GluA1. Consistent with the results in EPSC AMPA amplitudes (Fig. 2a,b), knockdown of CCNY increased endogenous surface level of GluA1 in dendritic protrusions compared to control cells (Fig. 2c,d) whereas co-expression of shRNA-resistant CCNY with the CCNY shRNA rescued the increase in surface levels of GluA1 caused by CCNY knockdown (Fig. 2c,d). Labeling of NMDA receptors in dendritic protrusions was unaffected by CCNY knockdown (Fig. 2e). Moreover, CCNY knockdown had no effect on the total levels of GluA1 or the NMDA receptor subunit GluN1 (Fig. 2f-h). The reduction of synaptic, but not total AMPA receptor levels upon CCNY knockdown suggests regulation of receptor trafficking.
The inhibition of SEP-GluA1 surface expression by CCNY overexpression could be explained by either (1) a lack of an intracellular pool of SEP-GluA1 available to be exocytosed or (2) blockade of the SEP-GluA1 exocytic pathway per se by CCNY. We reasoned that if the former is the case, the relative increase in the SEP-GluA1 signal after glycine stimulation should be comparable to that of control neurons after CCNY knockdown. Conversely, if the latter is the case, it should be significantly higher than in control neurons after CCNY knockdown.
Whereas  Fig. 4c). This suggests that CCNY regulates surface GluA1 level by inhibiting their exocytosis during LTP. In further support of these findings, overexpression of CCNY-WT decreased the number of SEP-GluA1 insertion events in spines following glycine stimulation (Fig. 4d), while CCNY knockdown significantly increased the number of these events (Fig. 4d). This augmentation of exocytic events was rescued by co-expression of the shRNA-resistant CCNY-WT plasmid in CCNY knockdown cells (Fig. 4d), confirming that these events are attributed specifically to CCNY. Importantly, glycine stimulation did not affect the overall expression level of CCNY (Fig. 4e,f).

Knockdown of CCNY increases phosphorylation of GluA1 at S845 during LTP-inducing stimulation.
GluA1 has two well-characterized phosphorylation sites on the C-terminus such as serine (S) 831 and S845, regulatory phosphorylation of which has been known to play a crucial role in synaptic plasticity [29][30][31][32][33][34][35] . Phosphorylation of S845 by protein kinase A (PKA) controls synaptic trafficking of GluA1 during LTP [36][37][38][39][40] . To further support the finding that CCNY inhibits plasticity-induced AMPA receptor trafficking, we tested whether the glycine-induced increase in phosphorylation of GluA1 at S845 is affected under conditions of altered CCNY levels. Glycine stimulation increases phosphorylation of GluA1 at S845 (Fig. 4g,h) as it has been known to be observed in LTP. This increase was even further enhanced by CCNY knockdown (Fig. 4g,h). These results suggest that CCNY negatively controls GluA1 phosphorylation at S845 during glycine-induced LTP.

Discussion
In the present study, we showed that the cyclin protein CCNY is expressed in the hippocampus, and is located in perisynaptic domains of dendritic spines. In addition, CCNY inhibits plasticity-induced AMPA receptor trafficking to the synapse. Given that knockdown of CCNY enhances LTP in hippocampal slices, we postulate that CCNY inhibits functional plasticity by restricting the synaptic delivery of AMPA receptors during LTP (Fig. 6).
Recent studies have begun to define novel roles for cyclin proteins in non-proliferating neuronal cells 42 . Our findings reveal that CCNY regulates synapse function, while the canonical role of the cyclin proteins is to regulate cell proliferation. This unique function of CCNY in the nervous system could contribute to  the increased complexity and diversity of brain function. Like other cyclin proteins, CCNY forms a complex with CDKs, such as PFTK1/CDK14 and PCTK1/CDK16, to control several biological processes 11,43,44 . Interestingly, our data support that CCNY exerts its inhibitory roles on the AMPA receptor exocytosis during LTP. It will be important for future studies to determine whether CCNY performs other neuronal functions independent from, or in concert with, its known CDK partners PFTK1 and/or PCTK1.
Our study provides the first demonstration of CCNY function in the vertebrate nervous system. Our biochemical subcellular fractionation and high-resolution confocal imaging results indicate that a significant amount of CCNY is located in the immediate vicinity of the plasma membrane in spines. This localization of CCNY could provide for rapid regulation of CCNY during the activity-dependent AMPA receptor trafficking at synapses. Further investigation is necessary to determine whether CCNY functions away from the plasma membrane in other parts of neurons and to determine what mechanisms are involved for this function. AMPA receptors recruited to the synapse during LTP are originated from recycling endosomes 26 . Given the role of CCNY as a negative regulator of the AMPA receptor insertion during LTP, CCNY might exert its function by inhibiting the exit of AMPA receptors from the intracellular compartments, such as recycling endosomes or by inhibiting the AMPA receptor-containing vesicle fusion process to the plasma membrane 24 .
The actin cytoskeleton is abundant in spines and plays a critical role in dynamic changes in the structure of dendritic spines. During LTP, spine enlargement and synaptic recruitment of AMPA receptors occur together 13,17,45 . In addition, AMPA receptors recruited to synapses utilize the actin-based motor myosin Vb to arrive at the synapse 27,46 . Our findings show that the activation of synaptic NMDA receptors causes CCNY to play an inhibitory role in the AMPA receptor insertion, and LTP. Therefore, it would be interesting to investigate if CCNY-mediated inhibition of the AMPA receptor insertion during LTP involves actin remodeling. If so, CCNY could be proposed to be a factor to link structural and functional changes during LTP through the regulation of actin dynamics, leading to the control of both AMPA receptor delivery and spine enlargement. It will be important to delineate the cellular and molecular mechanisms required for CCNY signaling during neuronal structural and functional plasticity. The cyclins were first identified by their oscillating and cell cycle-dependent expression patterns and were reported to regulate cell division. Neurons in the central nervous system are postmitotic, terminally differentiated cells that are no longer capable of undergoing cell division 1,2,47 . Thus, studying the function of cyclin proteins in postmitotic neuronal cells may, at first glance, appear rather contradictory. Yet, non-mitotic roles of cell cycle proteins have been reported in the nervous system 42 . For instance, ablation of cyclin E using conditional cyclin E knockout mice reduces the number of synapses and spines and causes impairments in synaptic plasticity and memory formation 42 .
Given that the CCNY interacting partners responsible for the neuronal function of CCNY are unknown at the moment, identifying binding partners or regulatory mechanisms for CCNY at the synapse will be important to define the precise mechanism by which CCNY regulates synaptic strength.
For constructing lentiviral vectors expressing CCNY shRNAs, the insert containing H1 promoter and CCNY shRNAs was isolated from pSuper-CCNY shRNAs, and subcloned between the HIV-flap and ubiquitin promoter of FUGW lentiviral vector (a gift from Michael Ehlers, Pfizer Neuroscience). For constructing lentiviral vectors expressing CCNY-WT, the insert containing the CCNY-WT PCR fragment was subcloned into EcoRI/BstBI sites of FUGW lentiviral vector. FUGW lentiviral vector contains the EGFP gene under a ubiquitin promoter to indicate viral production and infection.
Production of lentivirus. Lentiviral vector FUGW harboring CCNY-WT or CCNY shRNA, the packaging vector Δ 8.9, and VSVG envelope glycoprotein vector were cotransfected into HEK 293T cells using Fugene HD (Promega). Supernatants containing the lentivirus were harvested 36− 48 hours after transfection, and ultracentrifuged at 25,000 rpm to concentrate the lentivirus. The pellet was resuspended in phosphate-buffered saline (PBS), aliquoted, and stored at − 80 °C.
Preparation of brain homogenates and neuronal cell lysates. Hippocampi were rapidly removed from adult rat brain and homogenized with a Dounce glass tissue grinder homogenizer (Wheaton Industries) in ice− cold homogenization buffer (mM: 320 sucrose, 10 HEPES, 2 EDTA, protease inhibitor cocktail, 1 PMSF, pH 7.4). The neuronal cells were collected in lysis buffer (mM: 50 Tris− HCl, 150 NaCl, 5 EDTA, 1% Triton X-100, protease inhibitor cocktail, 1 PMSF, pH 7.4) on ice, and lysed by incubating for 1 hr at 4 °C. After centrifugation at 1,000 g for 10 min at 4 °C, supernatants were collected, and protein concentrations were measured by Bradford assays (Bio-Rad Protein Assay kit, Bio-Rad Laboratories). Subcellular fractionation. Subcellular fractionation was performed from P30 Sprague-Dawley (SD) rat forebrain as described previously [48][49][50] . In brief, the cerebellum and the brain stem were removed from thirty-day-old (P30) SD rat brain. The three rat forebrains were homogenized in buffer A (0.32 M sucrose, 20 mM HEPES, 5 mM EDTA, protease inhibitor cocktail, 1 mM PMSF pH 7.4) using a glass-teflon homogenizer with 30 strokes. Homogenate was centrifuged for 10 min at 1,000 g to produce a nuclear fraction (P1). The supernatant (S1) was centrifuged at 9,200 g for 10 min. The resulting pellet was washed by resuspending in buffer A and then centrifuged at 10,000 g for 20 min to produce crude synaptosomal fraction (P2). The supernatant was further centrifuged at 12,000 g for 30 min to collect the supernatant (S2). S2 was centrifuged at 165,000 g for 2 hours at 4 °C using NVT90 rotor to produce the cytosolic supernatant (S3) and the microsomal pellet (P3). P2 was resuspended in buffer A and lysed by hypo-osmotic shock using 9 volumes of H 2 O and 3 strokes with a glass-teflon homogenizer, and rapidly adjusted to 4 mM HEPES/5 mM EDTA (pH 7.4) and kept on ice for 30 min. The lysate was centrifuged at 25,000 g for 20 min at 4 °C to produce the synaptosomal membrane pellet (LP1) and the synaptic vesicle and cytosolic supernatant (LS1). LS1 was further centrifuged at 165,000 g for 2 hours at 4 °C using NVT90 rotor to produce the synaptic cytosolic supernatant (LS2) and the synaptic vesicle-enriched pellet (LP2). LP1 was resuspended and loaded on top of a discontinuous sucrose gradient solution containing 0.8 M, 1 M and 1.2 M sucrose. The gradient was centrifuged at 150,000 g for 2 hours at 4 °C using SW41Ti rotor. The cloudy band between 1.0 M and 1.2 M sucrose was collected and then diluted to buffer A. The diluted suspension was further centrifuged at 150,000 g for 30 min using SW41Ti rotor to produce the synaptic plasma membrane fraction (SPM). SPM was resuspended with 0.5% Triton X-100 in buffer A and kept on ice for 15 min and then centrifuged at 32,000 g for 20 min to divide into soluble and insoluble fractions (Triton X-100 soluble fraction and Postsynaptic density fraction). Triton X-100 insoluble PSD fraction was resuspended in buffer A. Five μ g of proteins of each fraction was analyzed by immunoblotting.
Immunocytochemistry. For staining surface AMPA receptors, hippocampal neurons were fixed with 4% paraformaldehyde/4% sucrose in PBS. Then, surface GluA1 was labeled with rabbit anti-GluA1-N (1816, a gift from Michael Ehlers, Pfizer Neuroscience or Millipore) for 1 hr at room temperature. Neurons were washed and incubated with Cy3-conjugated anti-rabbit secondary antibody for 50 min at room temperature to visualize surface GluA1s. For staining HA-tagged CCNY or the total level of GluA1, hippocampal neurons were fixed with 4% paraformaldehyde/4% sucrose in PBS and permeated with 0.1% Triton X-100 in PBS. Then, HA-tagged CCNY or the total level of GluA1 was labeled with mouse anti-HA (Convance) or rabbit anti-GluA1-C (Abcam), respectively for 1 hr at room temperature. Neurons were washed and incubated with Cy3-conjugated secondary antibody for 50 min at room temperature to visualize HA-tagged CCNY or the total level of GluA1. Cell culture and DNA transfection. HEK 293T cells were grown in DMEM (HyClone) supplemented with 10% fetal bovine serum. Hippocampal neuron cultures were prepared from E18 SD rat embryos and maintained for 10−21 days in vitro (DIV) (Park et al., 2006). Neurons were transfected between 10−14 DIV using the Lipofectamine 2000 (Invitrogen) for 1−2 or 4−7 days for overexpression or shRNA knockdown experiments, respectively.
Live−cell imaging. Live neurons grown on the coverslip that were transfected appropriately were transferred to the imaging chamber equipped with heating plate base (Live Cell Instrument, Seoul, Korea); filled with imaging solution (mM: 120 NaCl, 3 KCl, 2 CaCl 2 , 2 MgCl 2 , 15 glucose, 15 HEPES, pH 7.35), and imaged at 32 °C. Confocal images were acquired using the Revolution XD System (Andor Technology) equipped with Yokogawa CSU-X1 spinning disk confocal unit, 488 nm solid state laser, 561 nm solid state laser, 640 nm diode laser, and Andor 6-line laser combiner. Images were taken using a 60x (NA 1.4) or 100x Plan Apochromat objective (NA 1.4) and a 14-bit iXON3 DU-885 EMCCD camera (Andor Technology) using the Metamorph software program (Molecular Device Inc.). We acquired a complete confocal z-sectioning of the region of interest, followed by maximal intensity projection to produce a two-dimensional image using Metamorph. Image analysis and quantification. To analyze the surface AMPA receptor intensity, integrated intensity of individual puncta of endogenous surface GluA1 on the dendritic protrusions was measured. For NMDA receptor analysis, integrated intensity of an NMDA receptor subunit GluN1 from the dendritic protrusions was measured. To evaluate the changes of SEP-GluA1 intensity in the spine, the change in fluorescence intensity, ΔF was normalized to F 0 as ΔF/F 0 . Δ F was calculated by F t − F 0 where F t indicates the intensity at each time point, and F 0 indicates the average intensity of all time points prior to glycine treatment. For 3D volume rendering (Fig. 1h), 4D viewer for Metamorph NX software was used. Image XY calibration was 0.02-0.12 μ m per pixel, and distance between planes was 0.15-0.22 μ m.
Excitatory postsynaptic currents (EPSCs) were recorded using an Axopatch 700B amplifier (Axon Instruments). Pipette solution was comprised of 130 mM CsMeSO 4 , 8 mM NaCl, 4 mM Mg-ATP, 0.3 mM Na-GTP, 0.5 mM EGTA, 10 mM HEPES and 6 mM QX-314. The pH was adjusted to 7.2-7.3 using CsOH and osmolality was adjusted to 270-290 mOsm with sucrose as necessary. Electrodes were pulled using a horizontal Flaming Brown puller (P-97, Sutter Instruments). Electrode resistance was in the range of 4-6 MΩ. CA1 pyramidal neurons were voltage clamped at −70 mV. Only cells with series resistance <20 MΩ with a change in series resistance <10% from the baseline were included in this study. The amplitude of EPSCs was measured and these measurements were expressed relative to the normalized preconditioning baseline. LTP was induced by pairing 2 Hz stimulation with depolarization of the postsynaptic cell to 0 mV for 100 s. AMPA receptor-mediated EPSC amplitude (EPSC AMPA ) was measured as the peak EPSC amplitude at a holding potential of −70 mV, and NMDA receptor-mediated EPSC amplitude (EPSC NMDA ) was measured at + 40 mV at 80-200 ms after the peak of EPSC AMPA . Data pooled across slices are expressed as the mean ± SEM, and effects of conditioning stimulation were measured 30-35 min after induction of LTP. Data are expressed relative to baseline (100% = no change). Significance (p < 0.05) from baseline was tested using two-tailed t tests.