Protein Kinase Cϵ (PKCϵ) Promotes Synaptogenesis through Membrane Accumulation of the Postsynaptic Density Protein PSD-95*

Protein kinase Cϵ (PKCϵ) promotes synaptic maturation and synaptogenesis via activation of synaptic growth factors such as BDNF, NGF, and IGF. However, many of the detailed mechanisms by which PKCϵ induces synaptogenesis are not fully understood. Accumulation of PSD-95 to the postsynaptic density (PSD) is known to lead to synaptic maturation and strengthening of excitatory synapses. Here we investigated the relationship between PKCϵ and PSD-95. We show that the PKCϵ activators dicyclopropanated linoleic acid methyl ester and bryostatin 1 induce phosphorylation of PSD-95 at the serine 295 residue, increase the levels of PSD-95, and enhance its membrane localization. Elimination of the serine 295 residue in PSD-95 abolished PKCϵ-induced membrane accumulation. Knockdown of either PKCϵ or JNK1 prevented PKCϵ activator-mediated membrane accumulation of PSD-95. PKCϵ directly phosphorylated PSD-95 and JNK1 in vitro. Inhibiting PKCϵ, JNK, or calcium/calmodulin-dependent kinase II activity prevented the effects of PKCϵ activators on PSD-95 phosphorylation. Increase in membrane accumulation of PKCϵ and phosphorylated PSD-95 (p-PSD-95S295) coincided with an increased number of synapses and increased amplitudes of excitatory post-synaptic potentials (EPSPs) in adult rat hippocampal slices. Knockdown of PKCϵ also reduced the synthesis of PSD-95 and the presynaptic protein synaptophysin by 30 and 44%, respectively. Prolonged activation of PKCϵ increased synapse number by 2-fold, increased presynaptic vesicle density, and greatly increased PSD-95 clustering. These results indicate that PKCϵ promotes synaptogenesis by activating PSD-95 phosphorylation directly through JNK1 and calcium/calmodulin-dependent kinase II and also by inducing expression of PSD-95 and synaptophysin.

Protein kinase C⑀ (PKC⑀) is one of the novel PKC isotypes and is characterized as a calcium-independent and phorbol ester/diacylglycerol-sensitive serine/threonine kinase. Among the novel PKCs, PKC⑀ is the most abundant species in the central nervous system, mediating various neuronal functions (1,2). In neuroblastoma cells overexpression of PKC⑀, but not PKC␣, -␤II, or -␦ leads to neurite outgrowth through interac-tion of actin filaments and the C1 domain of PKC⑀ (3)(4)(5). The actin binding site of PKC⑀ is also implicated in exocytosis of neurotransmitters (6). PKC⑀ is essential for many types of learning and memory (7,8) and neuroprotection (9 -13). Neuronal contact with astrocytes also promotes global synaptogenesis through PKC⑀ signaling (14). PKC⑀ activation has been shown to promote the maturation of dendritic synapses during associative learning (9). PKC⑀ activation also protects against neurodegeneration (10,15). Phosphorylation of long-tailed AMPA receptors GluA4 and GluA1 by PKC promotes their surface expression (16,17). PKC activation induces protein synthesis required for long term memory (12,18). PKC⑀ activation is also required for HuD-mediated mRNA stabilization of neurotrophic factors (19) and apoE-mediated epigenetic regulation of BDNF (20). PKC activation induces translocation of calcium/ calmodulin-dependent kinase II (CaMKII) 2 to synapses (21) where it participates in PSD-95-induced synaptic strengthening (22). PKC also promotes NMDA receptor trafficking by indirectly triggering CaMKII autophosphorylation and subsequent increased association with NMDA receptors (23).
Thus, a number of studies have suggested that PKC activators such as bryostatin and dicyclopropanated linoleic acid methyl ester (DCPLA-ME) may be useful therapeutic candidates for the treatment of Alzheimer disease and other causes of synaptic loss such as ischemia, stroke, and fragile X syndrome (5,6,14,24). Some of these benefits have been attributed to induction of neurotrophic factors such as BDNF or the activation of anti-A␤ repair pathways and anti-apoptotic activity (10,13,20,25). However, the biochemical mechanisms by which PKC⑀ induces synaptogenesis and mediates neuroprotection are still not fully understood.
At excitatory synapses, the postsynaptic density is characterized by an electron-dense thick matrix that contains key molecules involved in the regulation of glutamate receptor targeting and trafficking (26). PSD-95 is an abundant scaffold protein in excitatory synapses, where it functions to cluster proteins such as glutamate receptors on the postsynaptic membrane and couples them to downstream signaling molecules, thereby inducing the surface expression and synaptic insertion of glutamate receptors (27)(28)(29). In addition to its role in synaptic function, PSD-95 has also been proposed to affect synapse maturation and stabilization (30 -32) and, thus, synapse number. Phosphorylation of the serine 295 residue of PSD-95 enhances the synaptic accumulation of PSD-95 and its ability to recruit surface ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors and potentiate excitatory postsynaptic currents (33).
In the present study we examined the role of PKC⑀ signaling and PKC⑀ activators in PSD-95 regulation and induction of synaptogenesis in cultured neurons and CA1 hippocampal slices. We report that PKC⑀ activation induces membrane translocation and phosphorylation of PSD-95 at the serine 295 residue, coinciding with an increased number of synapses. Our data suggest that an important mechanism by which PKC⑀ induces synaptogenesis is by increasing the phosphorylation of PSD-95 at the postsynaptic site and by regulating the expression of synaptophysin at the presynaptic site.

PKC Activation Prevents Degradation of Primary Human
Neurons-PKC⑀ is present in high concentration in central neuronal tissues and has been implicated in broad spectrum neuronal functions. To determine the effect of PKC⑀ activation on survival and maintenance, primary human neurons were treated for 40 days with two different PKC activators (bryostatin 1 and DCPLA-ME, which are relatively specific for PKC⑀) (13, 34 -36). Culture media and activators were changed every 3 days. Cells were imaged from three independent wells every 5 days, and neurite-positive cells were counted from 508-m 2 field images. Cells treated with either DCPLA-ME (100 nM) or bryostatin 1 (0.27 nM) showed an improved survival with increased neuritic branching (Fig. 1A). Untreated cells showed degeneration and 50% cell loss by 36 days, whereas the treated cells remained healthy for at least 40 days (Fig. 1B). The number of viable neurite-positive cells was also significantly higher at 40 days (F (2,6) ϭ 705.4; ANOVA, p Ͻ 0.0001) in the activatortreated cells than untreated cells (bryostatin 1, 369.7 Ϯ 12.2; DCPLA-ME, 334.7 Ϯ 1.8; untreated 109.7 Ϯ 6.4).
Bryostatin 1 and DCPLA-ME Specifically Activate PKC⑀-We then investigated whether this phenomenon is specific to PKC⑀ or whether other PKC isozymes are involved. PKC translocation to the plasma membrane generally has been considered the hallmark of activation and frequently has been used as a surrogate measure of PKC isoform activation in cells (37). Expression levels of PKC␣, PKC⑀, and PKC␦ in the soluble (cytosol) and particulate (membrane) were measured by immunoblot at 1, 4, and 24 h after either bryostatin 1 (0.27 nM) or DCPLA-ME (100 nM) treatment (Fig. 2, A and C). Both DCPLA-ME and bryostatin 1 increased membrane translocation of PKC⑀ but not PKC␣ or PKC␦ (Fig. 2, B and D), confirming that both the compounds activate PKC⑀ but not PKC␣ or PKC␦.
PKC⑀-mediated Membrane Localization of p-PSD-95 S295 Involves JNKI and CaMKII-Previously it has been reported that accumulation of PSD-95 in the PSD is increased by synaptic activity and by a Rac1-JNK1 signaling pathway (33). PKC⑀ is involved in JNK activation in macrophages (38,39), and CaM-KII inhibitors inhibit PKC-mediated signaling in hippocampal neurons (40). Thus we investigated the involvement of PKC⑀, JNK, and CaMKII in PSD-95 S295  PKC⑀ phosphorylated PSD-95 in vitro, incorporating 1.46 Ϯ 0.05 mol of [ 32 P]ATP/mol of PSD-95. Western blotting with p-PSD-95 S295 -specific antibody confirmed that this included the Ser-295 site (Fig. 3, A and B). PKC and JNK inhibitors fully inhibited the PKC⑀-mediated PSD-95 phosphorylation, whereas a CaMKII inhibitor partially prevented PSD-95 phosphorylation (Fig. 4E). PKC⑀ also phosphorylated JNK1 in vitro, incorporating 1.02 Ϯ 0.04 mol of [ 32 P]ATP per mol of JNK1; BisI prevented JNK1 phosphorylation (Fig. 4F). PKC is also reported to phosphorylate CaMKII in vitro (41); we also found an increase in phosphorylation of CaMKII by PKC⑀ (Fig. 4G). Because both JNK and CaMKII inhibitors prevented PSD-95 phosphorylation by PKC⑀ (Fig. 4E), we considered the possibility that the JNK inhibitor might not be specific. Therefore, we performed a siRNA knockdown of PKC⑀ and JNK in human neurons. PKC⑀ or JNK knockdown caused a 50% reduction in their respective protein expression (Fig. 4H). DCPLA-ME failed to induce the membrane accumulation of p-PSD-95 S295 in PKC⑀ and JNK knockdown human neurons (F (5,12) ϭ 24.6; ANOVA, p Ͻ 0.0001 (Fig. 4, I and J). These results confirm that PKC⑀ is required for membrane translocation of p-PSD-95 S295 and that JNK and CaMKII are intermediates in the pathway (Fig. 4K). Next we investigated the effect of bryostatin 1 on basal synaptic transmission of hippocampal CA1 pyramidal neurons to determine whether the new synapses are functional. Field potential recordings were measured from rat hippocampal slices. An input-output curve was calculated with stimulus intensity versus the slope of excitatory synaptic potentials (EPSPs) elicited in response to increasing intensity of stimulation to the Schaffer collateral. The mean EPSP slope increased with stronger intensity of stimulus. Slices preincubated with bryostatin 1 for 1 h exhibited greater EPSP slope without any change in fiber volley amplitude. This was abolished with 30-min pretreatment with the PKC inhibitor bisindolylmaleimide I (Go 6850) (BisI, 100 nM) (Fig. 5, C and D). Bryostatin 1 increased the area under the curve, which represents the overall basal synaptic transmission, and a PKC inhibitor prevented the increase (bryostatin1, 0.71 Ϯ 0.08, p ϭ 0.03; Bis1 ϩ bryostatin 1, 0.49 Ϯ 0.07; untreated (ethanol only), 0.51 Ϯ 0.06) (Fig. 5E). Treatment of slices for 4 h with bryostatin (12 slices, 3 rats) dramatically increased the EPSP slope compared with the ethanol-treated slices (6 slices, 3 rats) (Fig. 5, F and G). The smaller response in the 4-h untreated slices compared with 1-h untreated slices may be attributed to the vehicle (ethanol) added to the slices. EPSPs in hippocampal slices are reduced by a smaller percentage after ethanol treatment (42). Thus, the prolonged treatment of slices with ethanol for 4 h may have slightly reduced the EPSP slope in these groups. Bryostatin increased the area under the curve by nearly 2-fold (p Ͻ 0.0001, Fig. 5H). These results suggest that bryostatin 1 treatment facil- itates basal synaptic transmission in the Schaffer collateral commissural pathway of rat hippocampus and that the increase in EPSP slope is independent of the fiber volley.
Our results indicate that increased phosphorylation of PSD-95 by PKC⑀ leads to an increase in synapse number with increased synaptic activity. Together these data demonstrate that the new synapses are functional.
PKC⑀ Knockdown Reduces the Expression of PSD-95 and Synaptophysin-PKC⑀ is known to perform important functions both in presynaptic (14) and postsynaptic sites. To investigate whether PKC⑀ is essential for the expression of synaptic proteins, we measured the effect of PKC⑀ knockdown (PKC⑀ KD) and PKC⑀ overexpression (PKC⑀ OE) on the expression of postsynaptic PSD-95 and presynaptic synaptophysin. Knockdown of PKC⑀ was achieved by transfecting the neurons with a mixture of siRNA containing a pool of three to five siRNAs.
Knockdown of PKC⑀ Reduces Synaptogenesis-To further establish the role of PKC⑀ in synaptogenesis and its underlying role in expression of PSD-95 and synaptophysin we used confocal microscopy to measure the effect of PKC⑀ knockdown on the localization of PSD-95 and colocalization of PSD-95 and synaptophysin. Punctate colocalization (clusters of proximal pre-and post-synaptic markers on neurites) of PSD-95 and synaptophysin is widely accepted as an indicator of synapses (43,44). Primary human neurons were treated with bryostatin 1or DCPLA-ME alone or after PKC⑀ KD for 10 days. PSD-95 clusters and colocalized PSD-95 and synaptophysin (as recognized by staining grains along a 40-m length of neurite, n ϭ 10) were counted in 4 independent experiments (Fig. 7A). In normal cells, PKC⑀ activation by bryostatin 1 and DCPLA-ME significantly induced PSD-95 clustering in the neurites compared with untreated controls (p Ͻ 0.05) (Fig. 7B). The number of synapses was also significantly higher in cells treated with bryo-statin1 and DCPLA-ME than in untreated neurons at 10 days (Fig. 7C). The increase in the number of synapses was independent of the neuron density. We found no change in neuron density (measured by NeuN staining) after 10 days of PKC⑀ activator treatment (supplemental Fig. 1). In PKC⑀ KD cells, immunofluorescence staining of human neurons showed a loss of synaptic networks, and bryostatin 1 and DCPLA-ME had no effect. PKC⑀ KD prevented the effect of PKC⑀ activators, and more importantly, reduced the basal level of PSD-95 clusters and synapses by 50%. (Fig. 7, A-C). We also quantified the expression levels of PKC⑀, p-PSD-95 S295 , PSD-95, and synaptophysin by immunoblot after 10 days of PKC⑀-siRNA transfection (Fig. 7D). PKC⑀ KD cells expressed significantly lower amounts of PSD-95 (F (5,12) ϭ 19.24, ANOVA p Ͻ 0.0001) (Fig. 7, E and F) and synaptophysin (F (5,12) ϭ 12.79, ANOVA p ϭ 0.0002) (Fig. 7G). Bryostatin 1 and DCPLA-ME failed to induce PSD-95 and synaptophysin expression in PKC⑀ KD neurons. Bryostatin 1, but not DCPLA-ME, produced a 40% decrease in PKC⑀ protein staining (Fig. 7D). No loss in PKC⑀ mRNA was found in bryostatin 1-treated neurons (data not shown). Down-regulation of PKC after activation by bryostatin 1 is a well documented phenomenon (45, 46). We further confirmed the effect of long term PKC⑀ activation on synaptogenesis using rat hippocampal brain slices. Slices were treated with bryostatin 1 and DCPLA-ME for 10 days. The serum-free culture medium was changed every 3 days with fresh additions of activators. Synapse number in each case was quantified using electron microscopy (Fig. 8A). Bryostatin 1 (7.97 Ϯ 0.68, p ϭ 0.013, n ϭ 29 CA1 areas) and DCPLA-ME (8.71 Ϯ 0.78, p ϭ 0.001, n ϭ 24 CA1 areas) treatment increased the number of synapses in hippocampal slices compared with vehicle-only treated slices (4.5 Ϯ 0.45; n ϭ 24 CA1 area) (Fig.  8B). Presynaptic vesicle density was also significantly higher in the bryostatin 1 (59.6 Ϯ 6.4, p Ͻ 0.05, n ϭ 19 presynaptic boutons)-and DCPLA-ME (60.4 Ϯ 5.1, p ϭ 0.04, n ϭ 19 presynaptic boutons)-treated slices than vehicle-treated controls (48.4 Ϯ 4.3, n ϭ 20 presynaptic boutons) (Fig. 8, C and D). Together, these findings confirm that PKC⑀ is essential for bryostatin 1and DCPLA-ME-mediated increase in PSD-95 and synaptophysin expression leading to increased synaptogenesis at 10 days.

Discussion
The outgrowth of neurites and formation of synapses depends on interactions among a number of regulatory proteins. These interactions are required for synaptic structure rearrangement, spinogenesis, and synaptogenesis. PKC⑀ is one of the key regulators of synaptogenesis (3,24), and PKC⑀ acti-vators promote the maturation of dendritic spines (9,47). PSD-95 is a scaffold protein that also plays an important role in formation of excitatory synapses (48,49).
Here we showed that PKC⑀ activation induces translocation and phosphorylation of PSD-95 at the serine 295 residue leading to PSD-95 accumulation at the postsynaptic density. Our findings showed that PKC⑀ activation not only increased the survival of neurons but also preserved the neuronal structure. Untreated cells showed gradual degeneration over 25 days, suggesting that PKC⑀ activation is beneficial for both maturation and survival of neurons, confirming a previous report by Hama et al. (14). We have shown that short term acute changes in PKC⑀ activity induce structural and biochemical changes in post-synaptic density scaffolding protein PSD-95 as well as increased synaptic activity. Synaptic activity is important for neuronal survival. Synaptic activity induces expression of survival genes and suppresses pro-death genes (50). Therefore, the increased survival of neurons treated with PKC⑀ activators may be due to the increased connectivity induced in the early stages; however, other factors such as elevated neurotrophins may also play a role. PKC⑀ induces BDNF (10,19), and elevated expression and release of BDNF is associated with elevated synaptic activity, which contributes to neuroprotection (51,52).
PKC⑀ activation and membrane translocation occur both presynaptically (14,53) and postsynaptically (8) where it phos- phorylates important substrate proteins required for synaptic facilitation and synaptogenesis. We found that p-PSD-95 S295 accumulation increased in the membrane of PKC⑀-activated neurons and followed the same time course as PKC⑀ activation at 1 h and 4 h. The serine 295 residue was essential for the PKC⑀-mediated membrane accumulation of PSD-95. In vitro, PKC⑀ phosphorylated both PSD-95 and JNK1. The JNK1 inhibitor also prevented PKC⑀ activation-mediated increase in p-PSD-95 S295 , confirming previous findings showing that serine 295 phosphorylation of PSD-95 is regulated by Rac1-JNK1 and PP1/PP2A signaling (33,54). PKC⑀ is involved in JNK activation; PKD, a downstream effector of PKC, also regulates JNK (38,39,55). Knockdown of either PKC⑀ or JNK inhibited the PKC⑀ activator-mediated p-PSD-95 S295 accumulation in the membrane, thus confirming that PKC⑀ and JNK act collectively in regulating PSD-95. Although it has been reported that synaptic localization of PSD-95 is regulated by JNK signaling and not by CaMKII (33,56), our data demonstrate a role of both JNKI and CaMKII. This is possible as PKC activation induces a simultaneous translocation of CaMKII to synapses (21), and CaMKII activation is needed for PSD-95-induced synaptic strengthening (22). CaMKII is a downstream target of PKC⑀ in many pathways, including the events responsible for the induction of neuroplastic changes associated with hyperalgesic priming (57). In this study we found that both JNK1 and CaMKII inhibitors prevented the PKC⑀-mediated membrane association of p-PSD-95 S295 . These results suggest that JNK1 and CaMKII are downstream to PKC⑀ in events responsible for phosphorylation and membrane accumulation of PSD-95.
We also demonstrated that PKC⑀ activation increases the levels of PSD-95 and the number of synapses. In adult hippocampal slices, bryostatin 1 increased basal synaptic activity.
Our results indicate an important link between PKC⑀ activation and the membrane localization of PSD-95, specifically enriching the membrane with the p-PSD-95 S295 form, which is known to strengthen the excitatory synapses (33). PSD-95 also regulates membrane insertion of AMPA receptor and dendritic spine morphology during synaptic plasticity (22, 30 -32).
Overexpression of PSD-95 converts silent synapses to functional synapses (58), whereas synaptophysin may be required for increased presynaptic vesicle density, thereby facilitating neurotransmitter release (59). We found that overexpressing PKC⑀ in primary human neurons induces the mRNA and protein levels of PSD-95 and synaptophysin, whereas knockdown of PKC⑀ reduces PSD-95 and synaptophysin mRNA and protein levels. Our results indicate that PKC⑀ regulates the gene expression of PSD-95 and synaptophysin. PKC⑀ may play a critical role in synapse maintenance by regulating the synthesis of PSD-95 and synaptophysin (18). PKC⑀ is known to drive the mitogenic response and DNA synthesis (60) via the Raf-MEK-ERK cascade and regulates transcription of essential genes through JNK/AP1, NF-B, and JAK/STAT cascades (61,62). PSD-95 is a critical transcriptional target of NF-B, which is known to induce excitatory synapse formation and regulate dendritic spine formation and morphology in murine hippocampal neurons (63). Synaptophysin mRNA expression is induced by the BDNF-cFos pathway (64). NF-B and synapto-physin have a common regulator in BDNF (65). PKC⑀ up-regulates BDNF expression (19 -20, 66).
In conclusion, PKC⑀ has two specific roles in synaptogenesis; at the postsynaptic site it regulates PSD-95, either directly or through JNKI and CaMKII, and at the presynaptic site it induces the expression of synaptophysin. Repeated treatment with PKC⑀ activators induces synthesis of PKC⑀, PSD-95, and synaptophysin, resulting in an increased number of synapses. PKC⑀ knockdown inhibits the synthesis of PSD-95 and synaptophysin leading to a reduced number of synapses. Besides the PKC-JNK1/CaMKII-PSD-95 pathway, PKC⑀ can also induce synaptogenesis through the HuD-BDNF pathway. PKC⑀ stabilizes HuD, which increases the stability and rate of translocation of target mRNAs. HuD increases as a result of PKC⑀ activation after learning (67) and stabilizes the mRNA for BDNF, nerve growth factor (NGF), and neurotrophin-3 (NT-3) (19). PKC⑀ activation induces the synthesis of BDNF (10,20,47), and BDNF induces transport of PSD-95 to the dendrites (68), which is required for maintenance of mature spines (69). Deficits of PKC⑀ function could also contribute to the synapse loss in Alzheimer disease (15), whereas the therapeutic elimination of such deficits may offer a strategy for the treatment of synaptic loss in Alzheimer disease and other synaptic disorders.
Cell Culture-Human primary neurons (hippocampal neurons, catalogue #1540, ScienCell Research Laboratories, Carlsbad, CA) were plated on poly-L-lysine-coated plates and were maintained in neuronal medium (ScienCell) supplemented with the neuronal growth supplement (NGS, ScienCell). For maintenance of neurons half of the media was changed every 3 days. Fresh activators were added with every media change. Human HEK-293 cells were obtained from ATCC, Manassas, VA. Cells were maintained in Eagle's minimum essential medium and 10% fetal bovine serum.
Organotypic Slice Culture-Organotypic hippocampal slices were prepared mainly according to the method described by Stoppini et al. (71) with slight modifications (72). Rats were sacrificed and immediately decapitated under sterile conditions. Brains were rapidly removed and placed into a chilled dissection medium composed of Hibernate A (BrainBits, Springfield, IL), 2% B27 supplement, 2 mM L-glutamine by Glu-taMax and antibiotic-antimycotics (all from Invitrogen). The hippocampi were dissected out in fresh chilled dissection medium. Isolated hippocampi were washed in new chilled dissection medium and placed on a wet 3-mm paper on the Teflon stage of a manual tissue slice chopper (Vibratome Co., Saint Louis, MO) for coronal sectioning at 300 m. Each slice with intact pyramidal and granular layers was transferred to one membrane insert (Millipore, Bedford, MA) in 12-well plates containing Neurobasal A, 20% horse serum, 2 mM L-glutamine, and antibiotics-antimycotics for 4 days. For long term maintenance slices were cultured in serum-free medium consisting of Neurobasal A with 2% B27, 2 mM l-glutamine, and antibioticantimycotics. Slices were incubated in a humidified 5% CO 2 atmosphere at 37°C. The entire medium was replaced with fresh medium at day 1. After that, half the medium was removed and replaced with fresh medium twice a week.
Cell Lysis and Western Blotting Analysis-Cells were harvested in homogenizing buffer containing 10 mM Tris-Cl (pH 7.4), 1 mM phenylmethylsulfonyl fluoride), 1 mM EGTA, 1 mM EDTA, 50 mM NaF, and 20 M leupeptin and lysed by sonication. The homogenate was centrifuged at 100,000 ϫ g for 15 min at 4°C to obtain the cytosolic fraction (soluble) and membrane (particulate). The pellet was resuspended in the homogenizing buffer by sonication. For whole cell protein isolation from primary neurons the homogenizing buffer contained 1% Triton X-100. Protein concentration was measured using the Coomassie Plus (Bradford) Protein Assay kit (Pierce). After quantification, 20 g of protein from each sample was subjected to SDS-PAGE analysis in a 4 -20% gradient Tris-glycine polyacrylamide gel (Invitrogen). The separated protein was then transferred to a nitrocellulose membrane. The membrane was blocked with BSA and incubated with primary antibody overnight at 4°C. All the primary antibodies were used at a 1:1000 dilution except rabbit polyclonal anti-p-PSD-95 S295 (1:10000) and rabbit polyclonal anti-synaptophysin (1:10000). After incubation, it was washed 3ϫ with Tris-buffered saline-Tween 20 and further incubated with alkaline phosphataseconjugated secondary antibody at 1:10,000 dilution for 45 min. The membrane was finally washed 3ϫ with Tris-buffered saline-Tween 20 and developed using the 1-step NBT-BCIP (nitro blue tetrazolium-5-bromo-4-chloro-3-indolyl phosphate) substrate (Pierce). The blot was imaged in an Image-Quant RT-ECL (GE Healthcare), and densitometric quantification was performed using IMAL software. For quantifying expression of a protein, the densitometric value for the protein of interest was normalized against ␤-actin (loading control).
Electrophysiology-Rats (1 month old) were euthanized, and hippocampus was isolated and sliced into 300-m slices on a Leica VT1200S Vibratome. Slices were incubated in ACSF at room temperature for 1 h until recording (ACSF: 124 mM NaCl, 3 mM KCl, 1.2 mM MgSO 4 , 2.1 mM CaCl 2 , 1.4 mM NaH 2 PO 4 , 26 mM NaHCO 2 , and mM 20 dextrose, saturated with 95% O 2 and 5% CO 2 , which maintains the pH at 7.4). Slices were treated with ethanol or bryostatin 1 for 1 h or 4 h. All recordings were made at room temperature. For synaptic stimulation and field EPSP recordings, pyramidal neurons in the CA1 field were identified with an Olympus BX50WI microscope. Field potential recordings were measured to determine synaptic function. A bipolar stimulating electrode (100-m separation, FHC, Bowdoinham, ME) was placed in the hippocampal Schaffer collateral pathway to elicit EPSPs in CA1 stratum radiatum, EPSPs were recorded through patch pipettes (2-5 megaohms, 1.5 mm outer diameter, 0.86 mm inner diameter, P87 Brown-Flaming Puller, Sutter Instruments) filled with ACSF. All parameters including pulse duration, width, and frequency were computercontrolled. Constant-current pulse intensity was controlled by a stimulus isolation unit. Basal synaptic transmission, represented by input-output responses, was determined by the slopes of stabilized EPSP to different stimulus intensities. The strength of EPSPs was assessed by measuring the slopes (initial 20 -80%) of the EPSPs rising phase.
Knockdown and Overexpression-Human PSD-95 was cloned into pCDNA3.1 plasmid (GenScript, Piscataway, NJ). Mutant PSD-95 mutated at serine 295 residue was also cloned into pCDNA3.1 plasmid and was obtained from GenScript. PKC⑀ knockdown was done using PKC⑀-siRNA constructs purchased from Santa Cruz Biotechnology. JNK knockdown was done using SAPK/JNK-siRNA from Cell Signaling Technology. Overexpression of PKC⑀ was obtained by transfecting pCMV6-ENTRY vector containing human PKC⑀ cDNA (Origene). Transfection was done using Lipofectamine 3000 (Invitrogen). Medium was changed after 6 h of Lipofectamine treatment. Protein expression was measured after 72 h of transfection.
Electron Microscopy-Electron microscopy of slices were done following methods described earlier (9). Hippocampi were sectioned with a vibratome at 100 m. Hippocampi were fixed in 1% OsO 4 . Electron micrographs (100 m 2 CA1 area at ϫ 5000) were made of Epon-embedded hippocampal sections with a JEOL 200CX electron microscope. These sections were 90 nm thick and had been previously stained with uranyl acetate and lead citrate. During quantification, electron micrographs were digitally zoomed up to ϫ20,000 magnification. Spines were defined as structures that formed synapses with axon boutons and did not contain mitochondria. Presynaptic vesicle density was measured from within the presynaptic axonal boutons that were seen to form synapses with dendritic spines of diameter Ͼ600 nm. Increased numbers of presynaptic vesicles in axon boutons were measured as an increase in the frequency of axon boutons with presynaptic vesicles that occupied Ͼ50% of the cross-section space not occupied by other organelles.
Immunofluorescence and Confocal Microscopy-Cells were grown in four-chambered slides (Nunc) at low density. For immunofluorescence staining the cells were washed with PBS (pH 7.4) and fixed with 4% paraformaldehyde for 4 min. After fixation, cells were blocked and permeabilized with 5% horse serum and 0.3% Triton X-100 in 1ϫ PBS for 30 min. Cells were washed 3ϫ with 1 ϫ PBS and incubated with primary antibodies (rabbit polyclonal anti-PSD-95, mouse monoclonal anti-synaptophysin, and chicken polyclonal anti-NeuN) for 1 h at 1:100 dilution. After the incubation slides were again washed 3ϫ in 1 ϫ PBS and incubated with the FITC anti-rabbit IgG, rhodamine antimouse IgG, and Cy5 anti-chicken IgY for 1 h at 1:400 dilution. Cells were further washed and mounted in Pro Long Gold antifade mounting solution (Invitrogen). Stained cells were viewed under the LSM 710 Meta confocal microscope (Zeiss) at 350-, 490-, 540-, and 650-nm excitation and 470-, 525-, 625-, and 667-nm emission for DAPI, FITC, rhodamine, and Cy5, respectively. Six individual fields at 40ϫ or 63ϫ oil lens magnification were analyzed for the mean fluorescence intensity in each channel. Punctate colocalization was done following the methods described earlier (43,44).
Statistical Analysis-All experiments were performed at least three times. Data are represented as the mean Ϯ S.E. All data were analyzed by one-way ANOVA and Newman-Keuls multiple comparison post test. Significantly different paired groups were further analyzed by two-tailed Student's t test using GraphPad Prism 6.1 software (La Jolla, CA). p values Ͻ 0.05 were considered statistically significant.