A Chinese herbal decoction, reformulated from Kai-Xin-San, relieves the depression-like symptoms in stressed rats and induces neurogenesis in cultured neurons

Kai-Xin-San (KXS), a Chinese herbal decoction for anti-depression, is a combination of paired-herbs, i.e. Ginseng Radix et Rhizoma (GR)-Polygalae Radix (PR) and Acori Tatarinowii Rhizoma (ATR)-Poria (PO). The make-up of the paired-herbs has been commonly revised according to syndrome differentiation and treatment variation of individual. Currently, an optimized KXS (KXS2012) was prepared by functional screening different combination of GR-PR and ATR-PO. The aim of this study was to verify the effect and underlying mechanism of KXS2012 against depression in chronic mild stress (CMS)-induced depressive rats and in primary cultures of neurons and astrocytes. In rat model, the CMS-induced depressive symptoms were markedly alleviated by the treatment with KXS2012. The CMS-suppressed neurotransmitter amounts were restored in the presence of KXS2012. And the expressions of neurotropic factors and its corresponding receptors were increased under KXS2012 administration. In cultured neurons, application of KXS2012 could promote neurogenesis by inducing the expression of synaptotagmin and dendritic spine density. Moreover, application of KXS2012 in cultured astrocytes, or in H2O2-stressed astrocytes, induced the expressions of neurotrophic factors: the increase might be associated with the modification of Erk1/2 and CREB phosphorylation. Our current results fully support the therapeutic efficacy of KXS2012 against depression in cell and animal models.

KXS 2012 reverses the depression-like symptoms in CMS-induced depressive rats. Three animal behavior tests including sucrose preference, forced swimming and open field tests were employed to evaluate the function of KXS 2012 against depression in rat. After the treatment of herbal extracts for 6 weeks, KXS 2012 at three doses (low dose: 60.9 mg/kg/day, medium dose: 182.7`mg/kg/day and high dose: 548.1 mg/kg/day) alleviated sucrose preference (~20%) of CMS-induced depressive rats (Fig. 1A). In forced swimming test, the CMS-induced depressive rats doubled cumulative immobility time, while KXS 2012 at all doses restored the cumulative immobility time (Fig. 1B). In open field test, the CMS-induced depressive rats showed a decrease of over 50% movements. However, the treatment of KXS 2012 at all doses reversed the time spent in central area of CMS-induced depressive rats: this treatment also affected the horizontal and vertical movements of KXS 2012 -treated rats (Fig. 1C). Imipramine (20 mg/kg/day) was set as a positive control, and non-stressed rats were set as another control.

KXS 2012 restores the levels of neurotransmitters in CMS-induced depressive rat brain.
According to previous study, a systematic method was used to determine the amounts of neurotransmitters in rat brain by HPLC-fluorescence detector (FLD) 19,20 . The detected neurotransmitters included: norepinephrine, dopamine and serotonin. In addition, 5-hydroxyindoleacetic acid (5-HIAA), a major metabolite of serotonin (5-HT), was also determined. The ratio of 5-HIAA/5-HT represents the metabolic state of serotonin. Thus, the ratio of 5-HIAA/5-HT was calculated to interpret the metabolism of serotonin. The HPLC conditions for neurotransmitter detection were determined. The analytic parameters are listed in Table S1. The chromatograms of neurotransmitters are shown in Fig. S2. Three brain regions were selected here, i.e. hippocampus, cerebral cortex and striatum. In CMS-induced depressive rat hippocampus, the amounts of serotonin and norepinephrine, as well as the ratios of 5-HIAA/5-HT, were decreased to ~80%, ~60% and ~65%, respectively ( Fig. 2A). The treatment of KXS 2012 at all doses restored the levels of serotonin and norepinephrine. However, the KXS 2012 treatment failed to reverse the ratio of 5-HIAA/5-HT caused by CMS ( Fig. 2A). In depressive rat cerebral cortex, the reduction in amounts of serotonin (~25%), norepinephrine (~40%) and ratio of 5-HIAA/5-HT (~25%) were observed. KXS 2012 at all doses restored the levels of serotonin and norepinephrine in cortex. In parallel, KXS 2012 at medium dose (182.7 mg/kg/day) increased the ratio of 5-HIAA/5-HT to normal state (Fig. 2B). In depressive rat striatum, the amounts of serotonin, dopamine and norepinephrine, as well as the ratio of 5-HIAA/5-HT, were reduced from 20% to 50%. KXS 2012 at all doses restored the levels of serotonin, dopamine and norepinephrine, as well as the ratio of 5-HIAA/5-HT in the striatum (Fig. 2C). Imipramine (20 mg/kg/day) was used as a positive control, and non-stressed rats were set as another control.
KXS 2012 up-regulates neurotrophic factor expressions in CMS-induced depressive rat brain.
The deficiency of neurotrophic factors is one of the indicative pathogens during depression. Here, the protein levels of nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF) and glia cell-derived neurotrophic factor (GDNF) in depressive rat cerebral cortex were detected by ELISA. The standard curves for calibration were: Y = 0.002X + 0.026 (NGF), R 2 = 0.998; Y = 0.001X− 0.044 (BDNF), R 2 = 0.990; Y = 0.000516X + 0.00648 (GDNF), R 2 = 1.000 (Fig. S3). The protein levels of NGF, BDNF and GDNF in cortex were markedly decreased in CMS-induced depressive rat cortex; however, this reduction was alleviated after KXS 2012 treatment (Fig. 3A). In the same rat samples, the CMS-reduced mRNA expressions of neurotrophic factor receptors, including tropomyosin receptor kinase (Trk) A, Trk B and Trk C, were restored by the treatment of KXS 2012 (Fig. 3B). In particularly, KXS 2012 at medium and high doses largely increased the mRNA expression levels of Trk B and Trk C, indicating that KXS 2012 might exert anti-depression through up-regulating neurotrophic factor signaling pathway (Fig. 3B). The induction of Trk receptors by KXS 2012 was robust from 2 to 6 folds as compared to the control, which was much better than the positive control imipramine at 20 mg/kg/day. KXS 2012 promotes neurogenesis in cultured neurons. The impairment of the brain's ability to promote neurogenesis underlies depression. As expected, the cellular study of KXS 2012 on primary cultured rat cortical neurons and hippocampal neurons were employed to investigate the function of KXS 2012 on neurogenesis. Cultured cortical neurons could undergo the progress of differentiation from day in vitro (DIV) 1 to DIV 30. Thus, DIV 5 and DIV 15 were chosen as early and middle stage of neuron differentiation. Synapse formation involves pairing of the pre-and post-synaptic partners at a specific neuro-spatial coordinate [21][22][23] . The optimized dose of KXS 2012 in cultured cortical neurons was determined by MTT assays, i.e. the dose did not affect the cell number, as shown in Fig. S4A. In DIV 5 of cortical neurons, KXS 2012 increased the expression of synaptic vesicle protein, synaptotagmin: the maximal induction could reach over 4 folds (Fig. 4A). Synaptotagmin is a Ca 2+  The CMS-induced depressive rats were randomly divided into six groups as the protocol in Fig. 1. After drug treatment, the hippocampus was collected. The amounts of norepinephrine (NE), serotonin (5-HT) and ratio of 5-HIAA/5-HT in the extracts of hippocampus were calculated, as described in method session. Similar procedures were applied onto cerebral cortex (B) and striatum (C) of CMS-induced depressive rats. Additionally, the level of dopamine in striatum (C) was determined (bottom right). Dopamine level was too low to be measured in hippocampus and cortex. Values are expressed in μ g/g or the number of ratio, Mean ± SEM, n = 8, * p < 0.05, * * p < 0.01 and * * * p < 0.001 compared to the corresponding control.
Scientific RepoRts | 6:30014 | DOI: 10.1038/srep30014 sensor regulating neurotransmitter release and hormone secretion. The expression of PSD95, a post-synaptic marker, however was not altered. In DIV 15 of hippocampal neurons, KXS 2012 increased the dendritic spine density (number of spine per 10 μ m segment) at post-synaptic sites by ~4 folds, which was higher than the positive control BDNF application (Fig. 4B,C). The KXS 2012 -induced dendritic spine density could serve as an indicator of increased possible contacts between neurons. In addition, a remarkable increase of synaptotagmin expression was revealed indicating the potential function of KXS 2012 in synaptogenesis. Thus, the function of KXS 2012 against depression might be mediated by promoting neurogenesis in the brain.
KXS 2012 up-regulates the expressions of neurotrophic factors in normal and H 2 O 2 -stressed astrocytes. Neurotrophic factors secreted from astrocytes play key roles in neurogenesis, and which are down-regulated in brain during depression. The optimized dose of KXS 2012 in cultured astrocytes was determined by MTT assays, i.e. the dose did not affect the cell number, as shown in Fig. S4B. In cultured astrocytes, application of KXS 2012 greatly increased the mRNA and protein levels of NGF, BDNF and GDNF: the maximal induction was at ~2 folds ( Fig. 5A). For the underlying mechanism, the expressions of neurotrophic factors are mediated by cAMP-protein kinase A (PKA) and mitogen-activated protein kinases (MAPK) signaling pathways in astrocytes [24][25][26] . The phosphorylation levels of cAMP response element-binding protein (CREB) and extracellular regulated protein kinases (Erk) 1/2 were determined. In cultured astrocytes, KXS 2012 increased the phosphorylation of Erk1/2 (~1.5 folds) and CREB (~2 folds) in a time-dependent manner (Fig. 5B,C). Then, H89 and   U0126, the inhibitors of PKA and MEK1/2, were selected for further study. In all scenarios, application of both inhibitors in cultured astrocytes blocked the KXS 2012 -induced mRNA levels of NGF, BDNF and GDNF (Fig. 5D). The KXS 2012 -mediated phosphorylation of CREB was fully blocked by H89, and reduced by ~60% with U0126. On the other hand, U0126 fully blocked KXS 2012 -mediated phosphorylation of Erk1/2, while H89 did not show any effect (Fig. S5). These results of inhibitors therefore suggested the activation of KXS 2012 should be firstly at PKA, then to Erk1/2. Erk1/2, which was reported to be located downstream of PKA, participated in the PKA-mediated CREB phosphorylation 27 . In addition, the condition medium derived from KXS 2012 -treated astrocytes did not significantly increase the expressions of synaptotagmin and PSD-95 in cultured cortical neurons (Fig. S6). These results suggested that the potential functional relationship of KXS 2012 between neurons and astrocytes might be complicated and need further study.
Oxidative stress is considered as a common mechanism associated with most neurodegenerative diseases, e.g. depression. Thus, the protective role of KXS 2012 against oxidative stress was further demonstrated in cultured astrocytes. In H 2 O 2 -stressed astrocytes, the H 2 O 2 -induced toxicity was alleviated by the pre-treatment with KXS 2012 (Fig. 6A). The H 2 O 2 -induced loss of NGF, BDNF and GDNF proteins were reversed in the presence of KXS 2012 (Fig. 6B). N-acetyl-L-cysteine (NAC) served as a positive control here. Moreover, the application of KXS 2012 reduced H 2 O 2 -induced phosphorylations of CREB and Erk1/2 in cultured astrocytes (Fig. 6C,D), which might be associated with the function of neuronal adaptation based on changes in gene expression; these changes of gene expression were shown to be linked to stress 28,29 . The differential gene expression related to stress, e.g. induction of apoptotic protein, subsequently down-regulates the expressions of neurotrophic factors. These findings demonstrated that in stressed astrocytes, KXS 2012 could effectively prevent the loss of neurotrophic factors so as to support neurogenesis, which showed similarity to that of KXS 2012 in depressive rats.

Discussion
Depression is a serious mental illness with high incidence in clinics. It causes mood disorder, cognitive and behavioral disturbances, and even increases suicidal tendency 1 . Current treatment of depression is hindered due to the lack of effective agents. Chinese medicine has been widely used in cases of shock, low mood, forbid forgetfulness, anhedonia and dizziness that share similar syndromes to depression. In TCM, depression is manifested as mental disease caused by insufficiency of "xin-qi" and exuberance of "dampness". KXS is one of the common Chinese herbal formulae for the treatment of depression. In KXS, the two paired-herbs (GR-PR and ATR-PO) with the function of enhancing "xin-qi" and eliminating "dampness" could promise superior efficacy against depression. However, the make-up of these two paired-herbs has been commonly revised according to syndrome differentiation and treatment variation. In addition, traditional herbal formulae do not have intellectual property that hinders the possible development of herbal extract in developing as a treatment for anti-depression. Thus, KXS 2012 was developed here by optimizing the pairing of GR-PR and ATR-PO on cell study 14 , and we investigated the functions of KXS 2012 against depression in vivo and in vitro.
The CMS rat model is probably the most valid animal model of depression, which aims to model a chronic depression-like state that develops gradually over time in response to stress with more natural induction 15,16 . Based on our findings, KXS 2012 could significantly alleviate depression-like symptoms in CMS-induced depressive rat model. Depression is closely related to decreased levels of neurotransmitters in the brain 30,31 . The treatment of KXS 2012 could restore the low levels of neurotransmitters, i.e. norepinephrine, serotonin, dopamine and 5-HIAA/5-HT, in cerebral cortex, hippocampus and striatum of CMS-induced depressive rats. Abundant evidence indicates that neurotrophic factors play important roles against depression 32,33 . According to the results, the most striking feature was that the expressions of neurotrophic factors as well as their receptors could be up-regulated in KXS 2012 -adminstrated group, as compared to CMS depressive group. Our findings implied that KXS 2012 attenuated depression through up-regulating the levels of neurotransmitters and neurotrophic factors.
The neurogenesis hypothesis suggests that depression is largely caused by an impairment of nerve growth in the brain, which also helps to explain that anti-depressants may take several weeks for restoring neurogenesis to exert an effect 17,18 . Here, the potential drug targets and global cellular mechanism on neurogenesis could be addressed through cell study on primary cultured neurons and astrocytes. Our results indicated that KXS 2012 increased the expression of synaptic vesicle protein, synaptotagmin, at pre-synaptic neurons, and induced the dendritic spine density at post-synaptic neurons. In this regard, KXS 2012 may serve to promote the possible synapse development between neurons. The development of synapse is also dependent on neurotrophic factors, e.g. BDNF 34,35 . In line to this notion, KXS 2012 increased the expressions of neurotrophic factors in astrocytes through up-regulating cAMP-PKA signaling pathway. Erk1/2, a downstream effector of PKA, was also shown to be participated in the KXS 2012 -induced gene expressions. In addition, the condition medium derived from KXS 2012 -treated astrocytes did not significantly increase the expressions of synaptotagmin and PSD-95 in cultured neurons, suggesting that the potential functional relationship of KXS 2012 between neurons and astrocytes might be complicated and need further study.
Oxidative stress is considered as a common mechanism associated with most neurodegenerative diseases, e.g. depression. H 2 O 2 is a potent reactive oxygen species 28   and Erk1/2, suggesting that KXS 2012 might restore the levels of neurotrophic factors in stressed astrocytes, as to support neurophysiological functions, e.g. neurogenesis. These protective effects were illustrated both in vitro and in vivo here. Interestingly, Erk1/2 and CREB can be activated not only in response to the pro-growth and pro-survival stimuli, but in response to oxidative stress in neurons. The sustained phosphorylations of Erk1/2 and CREB correlate well with cell death and apoptotic protein expression, which subsequently down-regulate the expressions of neurotrophic factors. Our results implied that KXS 2012 could down-regulate the phosphorylations of Erk1/2 and CREB in stressed astrocytes, as to improve neuron adaptation 29 .
Considering the major ingredients in KXS 2012 herbal extract should be crucial for anti-depression, different chemicals within the herbs are known for specific functions. Ginsenosides derived from GR were the main chemicals with strong neurotrophic and neuroprotective effects in vitro and vivo [36][37][38][39] . 3,6′ -Disinapoyl sucrose derived from PR was reported to show obvious anti-depressive effect on depressive animal models, and the efficiency was mediated by increasing BDNF levels and CREB phosphorylation via the CaMKII and ERK1/2 pathway [40][41][42] . In parallel, ATR-derived α -and β -asarones were shown to reverse learning and memory abilities in dementia mice 43 . Pachymic acid of PO regulated the expression of 5-HT3A receptors in Xenopus oocytes 44 . Among all the active ingredients, asarones were considered as toxic. β -Asarone shows genotoxic, hepatocarcinogens and carcinogenicity in rodents. In our previous study, the making up of KXS 2012 should decrease the toxicity by conversion of β -asarone to a less toxic form α -asarone, as to enhance therapeutic efficacy 14 . Meanwhile, due to compatibility of paired-herbs in TCM that the combination of PO and PR with ATR might help reduce side or adverse effects. Therefore, it is reasonable that KXS 2012 could be used as a new regimen for anti-depression for the new make-up could exert robust effect against depression through multiple targets, whereas the combination could eliminate toxicity and side/adverse effects.

Materials and Methods
Preparation of herbal extracts. The  The procedures of CMS were conducted with some adjustments 11,45 . Briefly, a series of stressors were applied onto the animals: (1) water deprivation for 24 h, (2) stroboscopic illumination for 10 h, (3) cage tilt (45°) for 7 h, (4) noise for 10 h, (5) soiled cage (200 mL water in 100 g sawdust bedding) for 12 h, (6) exposure to an empty bottle for 1 h, (7) forced swimming at 8 °C for 6 min, (8) tail-clipping restraint for 6 min and (9) food deprivation for 21 h. These stressors were randomly arranged in one-week and repeated for 6 weeks. The rats in control group were left undisturbed in their home cages except for house maintaining such as cage cleaning. At the end of CMS procedures, sucrose preference test was determined to evaluate the CMS model.
The SD rats were randomly divided into six groups (n = 12). The control animals and CMS model were given with saline. For the other four groups, KXS 2012 at low dose (60.9 mg/kg/day), medium dose (182.7 mg/kg/day), high dose (548.1 mg/kg/day) and imipramine (20 mg/kg/day) were intra-gastrically given 30 min before stress exposure for 6 weeks. Sucrose preference test was carried out at the end of CMS procedures 11 . In Brief, rats in each group were learned to adapt to 2 bottles of 1% sucrose solution (w/v) 72 h before the test, and 24 h later, one bottle of 1% sucrose solution (w/v) was replaced with tap water for 24 h. Then, rats were deprived of water and food for 24 h. Sucrose preference test was conducted at 17:00 p.m., where rats were kept in individual cages with two bottles, one with 100 mL of 1% sucrose solution (w/v) and the other with 100 mL of water. After 3 h, the volumes of consumed sucrose solution and water were recorded and the sucrose preference was calculated by the following formula: sucrose preference = sucrose consumption/(water consumption + sucrose consumption)× 100% 45 .
Open field test was carried out at the end of herbal treatment. Briefly, the open field apparatus consisted of a square arena (100 cm × 100 cm × 50 cm). The floor was divided equally into 25 squares. In the test, a single rat was placed in the center of the arena and allowed to explore freely. The time that rats spent into the central area was recorded during a test period of 5 min. This apparatus was cleaned with a detergent and dried after occupancy by each rat. The open field test was started at 9: 00 a.m. in a quiet room 45 .
Forced swimming test was carried out at the end of herbal treatment. Rats in each group were placed in large glass cylinders (50 cm height and 20 cm diameter) with 30 cm height water at 22 ± 2 °C, so that rats were not able to support themselves by hind limbs. The test consisted of two parts: the first 15 min was for pre-swimming and then 24 h later, the swimming behavior was observed in 5 min, and the latency to float was measured and analyzed 46 .
Measurement of neurotransmitters. The brain tissues were dissected from the manipulated rats. In brief, the rats were sacrificed by decapitation, and the cortex, hippocampus and striatum were separated, respectively, rapidly frozen in liquid nitrogen and kept in − 80 °C for storage. The procedure was described: the tissue samples (1 g in 5 mL) were treated by tissue lysate (0.6 mol/L perchloric acid, 0.5 mM Na 2 EDTA and 0.1 g/L L-cysteine), and centrifuged (14,000 g, 15 min, 4 °C) twice. The supernatant was collected and treated with perchloric acid precipitation agent (0.6 mol/L perchloric acid, 1.2 M K 2 HPO 4 and 2 mM Na 2 EDTA). After centrifugation (14,000 g, 15 min, 4 °C) and filtration, the samples were challenged by HPLC-fluorescence detector (HPLC-FLD, Shimadzu, Japan) 47 .
The chromatographic separation for monoamine neurotransmitters was performed on a Shimadzu HPLC-RF series which was equipped with a degasser, a binary pump, an auto-sampler and a thermo-stated column compartment. The tissue samples were separated on an Agilent zorbax SB-C 18  Cell culture. Cortical and hippocampal neurons, from SD rat embryos at day 18, were isolated and cultured as described previously 50 with modifications. Briefly, the whole brain was dissected in Hank's Balanced Salt Solution (Sigma, St. Louis, MO) supplemented with 1 mM sodium pyruvate (Life Technologies) and 10 mM HEPES (Sigma, pH 7.4) without Ca 2+ and Mg 2+ to obtain cortex. The tissues were treated with trypsin (Life Technologies, 2.5%) for 20 min, followed by centrifugation at 1,500x g for 4 min or washing with plating medium (DMEM with 10% HS, supplemented with 0.5 mM GlutaMAX, 25 μ M monosodium glutamic acid, 100 U/mL penicillin and 100 μ g/mL streptomycin) to remove trypsin. Then, the tissues were triturated for several times to get single cells. The number of cells were counted using trypan blue (Life Technologies), and the cells were seeded on a Scientific RepoRts | 6:30014 | DOI: 10.1038/srep30014 poly-L-lysine (Sigma, 100 μ g/mL) coated 60-mm culture plates (10 6 cells/plate) and incubated at 37 °C in 5% CO 2 . Plating medium was changed to culturing medium (Neurobasal supplemented with 1 × B27, Life Technologies) in 24 h after plating. Cytosine arabinoside (Ara-C, Sigma, 2.5 μ M), a mitotic inhibitor was applied onto the cultures on the third day in vitro (DIV) to eliminate glial cells. One third of the medium was replaced by fresh culture medium every 5 day afterwards. At DIV 5 or 15, the medium was changed by neurobasal supplemented with 0.1× B27 for 3 h before herbal treatment. The time of herbal treatment was 96 h. Forskolin (Sigma, 50 nM) and BDNF (Alomone lab, Israel, 15 ng/mL) were used as positive controls, respectively 51 .
Astrocytes, from postnatal SD rat at day 1, were isolated and cultured as described previously 52 . The cortex was dissected in Hank's Balanced Salt Solution without Ca 2+ and Mg 2+ . After being trypsinized for 20 min, the cortex was washed with culture medium and triturated several times. The culture medium was DMEM supplemented with 10% FBS, 100 units/mL penicillin and 100 μ g/mL streptomycin. The cells were centrifuged at 1,000× g for 4 min. The cell pellet was re-suspended in the culture medium. The cells were seeded on plastic culture plates with the density of 2 × 10 4 cell/cm 2 and incubated at 37 °C in 5% CO 2 for 48 h before medium change. The culture medium was changed twice a week. Every time, the astrocytes were pipetted up and down several times to get rid of any loosely adherent oligodendrocytes, microglia and neurons 53 . In herbal treatment, drugs were diluted by DMEM with 0.5% FBS and 100 units/mL penicillin and 100 μ g/mL streptomycin, and the medium was changed by fresh DMEM with 0.5% FBS and 100 units/mL penicillin and 100 μ g/mL streptomycin 3 h before herbal treatment. The time of herbal treatment was 48 h. H89 (Sigma, 5 μ M) and U0126 (Sigma, 20 μ M) were applied onto cultures 3 h before herbal treatment, respectively. Forskolin (Sigma, 50 nM) and tetra-decanoylphorbol acetate (TPA, Sigma, 50 nM) were used as positive controls, respectively. In H 2 O 2 -stressed astrocytes, the cells were pre-treated with herbal extracts for 24 h before exposure to H 2 O 2 (400 μ M) for another 24 h. N-acetyl-L-cysteine (NAC, Sigma, 1 mM) was used as a positive control.
Cell viability assay. Cell  Protein phosphorylation. Primary cultures of astrocytes were seeded onto 12-well plate to reach the confluence of above 90%, the culture medium was changed to DMEM without serum for 5 h. The inhibitors were added 3 h before herbal treatment. The cells were treated with herbal extracts at different time points (0 min, 5 min, 10 min and 30 min). In stressed astrocyte culture, herbal extracts were applied onto the cells for 24 h, followed by H 2 O 2 (400 μ M) at different time points (0 min, 5 min, 10 min and 30 min). Then, the medium was expelled, and the cells were digested with 2x direct lysis buffer. The degrees of phosphorylation were determined by their specific anti-phospho-kinase and total kinase antibodies. Results are presented as intensities of phospho-bands relative to total bands and expressed as x Basal 54 .

SDS-PAGE and immunoblotting.
The cell lysate was collected, and protein content was determined by Bradford method. Proteins (~20 μ g) were separated on 8% SDS-polyacrylamide gels and transferred to a nitrocellulose. Transfer and equal loading of the samples was confirmed by staining the ponceau-S. The nitrocellulose was blocked with 5% fat-free milk in Tris-buffer saline/0.1% Tween 20 (TBS-T), and then incubated in the primary antibodies diluted in 2.5% fat-free milk in TBS-T over night at 4 °C. The primary antibodies were: anti-synaptotagmin (Gift from Prof. H. B. Peng, HKUST), anti-PSD-95 antibody (BD Biosciences), anti-phospho-Erk1/2 (Cell Signaling, Banvers, MA), anti-total Erk1/2 (Cell Signaling), anti-phospho-CREB (Cell Signaling) and anti-total CREB (Cell Signaling). After that, the nitrocellulose was rinsed with TBS-T and incubated for 2 h at room temperature in peroxidase (HRP)-conjugated anti-mouse secondary antibody, or peroxidase (HRP)-conjugated anti-rabbit secondary antibody (Life Technologies), diluted in 2.5% fat-free milk in TBS-T. After intensive washing with TBS-T, the immune complexes were visualized using the enhanced chemiluminescence (ECL) method (GE Healthcare, Piscataway, NJ). The intensities of bands in control and samples, run on the same gel and under strictly standardized ECL conditions, were compared on an image analyzer, using a calibration plot constructed from a parallel gel with serial dilutions of one of the sample 49,54 . Spine density analysis. Rat hippocampal neurons were seeded on 18-mm coverslips (1.5 × 10 5 cells per coverslip) coated with poly-D-lysine (1 mg/mL). At DIV 19, hippocampal neurons were changed by neurobasal medium supplemented with 1× B27 1 h before transfecting with pEGFP-N1 (Clontech, Mountain View, CA) using calcium phosphate precipitation, as previously described 55 . For imaging of dendritic spines in EGFP-transfected hippocampal neurons, the images were acquired by Zeiss LSM710 confocal microscope with a 40X oil-immersion objective using z serial scanning mode, and image analyses were performed with Imaris software. For quantification, two to three dendrite segments from each neuron were analyzed. The basal threshold values for the background of all images in each experiment were measured. The average of these threshold values was then applied to all images in the same experiment 56 .
Statistical analysis. All data were analyzed using one-way ANOVA followed by the Students t-test.