Activation of protein kinase A (PKA) pathway at presynaptic terminals plays a crucial role in the supply of synaptic vesicles (SVs) from the reserve pool, affecting the steady-state level of activity and the reconstitution of the readily releasable pool after intense stimulation. However, the identity of the stimuli activating this pathway is undefined. Using fluorescence resonance energy transfer and molecular genetic, we show that kainate, through the activation of presynaptic kainate receptors, induces PKA activation and enhances synapsin I phosphorylation at PKA-specific residues. This leads to a dispersion of synapsin I immunoreactivity, which is accompanied by a PKA-dependent increase in the rate of SV recycling at the growth cone and by an enhanced miniature excitatory postsynaptic currents frequency in mature networks. Selective activation of this pathway is induced by the native neurotransmitter glutamate, when applied in the high nanomolar range. These data identify glutamate, specifically acting on KARs, as one of the stimuli able to induce phosphorylation of synapsin at PKA sites, both at the axonal growth cone and at the mature synapse, thus increasing SV availability and contributing to plasticity phenomena.
Neurons contain 3 types of ionotropic glutamate receptors, N-methyl-d-aspartate, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), and kainate receptors (KARs), which are expressed both presynaptically and postsynaptically. While at the postsynaptic level, these receptors mediate fast synaptic transmission and provide the structural basis for the control of network excitability, and at the presynapse, they play a major role in regulation of neurotransmitter release and short-term plasticity (Liu et al. 1994; Clarke et al. 1997; Rodriguez-Moreno et al. 1997; Kamiya and Ozawa 2000; Satake et al. 2000; Schenk et al. 2003; Corlew et al. 2008; Pinheiro and Mulle 2008).
Presynaptic KARs, in particular, have been reported to control γ-aminobutyric acid (GABA) and glutamate release and to contribute to short-term plasticity in the hippocampus (Chittajallu et al. 1996; Rodriguez-Moreno et al. 1997; Contractor et al. 2011). During synaptogenesis, activation of AMPA receptors and KARs has been shown to regulate filopodia motility, stabilize the nascent synaptic contact, and favor the transition from filopodia to the mature synapse (Verderio et al. 1994; Chang and De Camilli 2001; Schenk et al. 2003; Tashiro et al. 2003).
Notwithstanding their typical topology and function as ligand-gated channels, it has been shown that some of the neuronal functions of both KARs and AMPA receptors are mediated through noncanonical metabotropic signaling pathways involving the activation of G proteins and intracellular signalling cascades linked to protein kinase C (PKC), protein kinase A (PKA), or mitogen-activated protein kinase (MAPK) (Rozas et al. 2003; Rodriguez-Moreno and Sihra 2007). In particular, kainate attenuates GABA release from interneuron terminals onto CA1 pyramidal cells through a mechanism which was suggested to involve a G protein coupled to phospholipase C, which generates diacylglycerol to activate PKC (Rodriguez-Moreno and Lerma 1998). At hippocampal mossy fiber synapses, low kainate (KA) concentrations facilitate glutamate release (Schmitz et al. 2001; Pinheiro et al. 2007; Kwon and Castillo 2008) through the involvement of AC and PKA (Rodriguez-Moreno and Sihra 2004). Whereas the existence of noncanonical metabotropic functions by KARs has received strong experimental support, many mechanistic aspects of the signaling pathway remain unresolved, including the possibility for a direct coupling between KARs and a G protein. In addition, clear evidence exists that some metabotropic actions attributed to KARs are due to the actions of neuromodulatory agents whose release or action is stimulated by KAR activation (Frerking et al. 1998; Lourenço et al. 2011)
In the past years, we have demonstrated that AMPA receptors are expressed at the presynapse (Schenk et al. 2003), where they play a crucial role in the regulation of neurotransmitter release and short-term plasticity. In particular, we have shown that stimulation of presynaptic AMPA receptors induces local activation of MAPK and phosphorylation of synapsin I at MAPK sites, followed by its dissociation from the actin cytoskeleton and increase in the rate of exocytosis (Schenk et al. 2005). In the present study, we demonstrate that functional presynaptic KARs coexist with AMPA receptors in axonal growth cones of developing neurons, and we identify a novel presynaptic target and a signaling pathway for the presynaptic action of these receptors. We show that kainate induces the local activation of PKA, but not of MAPK, leading to phosphorylation of synapsin I specifically at PKA sites. Our data point to a role of KARs as modulators of presynaptic function through a PKA-dependent tuning of the level of synapsin I phosphorylation that sets the efficiency of synaptic vesicle (SV) exocytosis.
Materials and Methods
Antibodies and R eagents
Monoclonal antibodies against synaptobrevin/vesicle-associated membrane protein (VAMP) 2, activated erk1,2, and beta-tubulin were purchased from Sigma (Milan, Italy). Polyclonal antibodies against GluK2/3 and GluK5 were purchased from Upstate Biotechnology (Lake Placid, NY). Polyclonal antibody against synapsin I were a kind gift of P. De Camilli (Yale University). Phalloidin conjugated to Texas Red was purchased from Molecular Probes (Eugene, OR). The secondary antibodies conjugated to Fluorescein Isothiocyanate, Texas Red, Cy5, Alexa Fluo, or peroxidase were obtained from Jackson ImmunoResearch (West Grove, PA) and from Molecular Probes (Invitrogen Corporation, Eugene, OR). AMPA, (S)-α-methyl-4-carboxyphenylglycine, (RS)-α-cyclopropyl-4-phosphonophenylglycine, D-(−)-2-amino-5-phosphonopentanoic acid, and (S)-1-(2-amino-2-carboxyethyl)-3-(2-carboxybenzyl)pyrimidine-2,4-dione (UBP302) were purchased from Tocris Cookson (Bristol, UK). GYKI53655 (racemic isomer, Paternain et al. 1995), a selective AMPA receptor antagonist, was a kind gift of J. Lerma (Consejo Superior de Investigaciones Científicas and Universidad Miguel Hernández). 2-(2-Amino-3-methoxyphenyl)-4H-1-benzopyran-4-one (PD98059), forskolin, kainic acid, Pluronic F-68, guanosine 5′-[β-thio]diphosphate trilithium salt (GDP-β-S), 3-isobutyl-1-methylxanthine (IBMX), and H-89 dihydrochloride hydrate were obtained from Sigma. All chemicals were diluted in Krebs’–Ringer’s–4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (KRH) (containing the following [in mM]: 125 NaCl, 5 KCl, 1.2 MgSO4, 1.2 KH2PO4, 2 CaCl2, 6 glucose, and 25 HEPES/Na, pH 7.4) at the indicated concentrations.
Hippocampal C ell C ultures
Primary neuronal cultures were prepared from the hippocampi of 18-day-old fetal Sprague–Dawley rats (Charles River Italica, Calco, Italy), as described previously (Bartlett and Banker 1984). All animal procedures were carried out according to guidelines approved by the European Communities Council Directive of 24 November 1986 (86/609/EEC). The dissociated cells were plated onto glass coverslips coated with poly-L-lysine at densities ranging from 258 to 516 cells/mm2. The cells were maintained in Neurobasal (Invitrogen, San Diego, CA) with B27 supplement and antibiotics, 2 mM glutamine, and glutamate (neuronal medium). Neurons were transfected with calcium phosphate at 2–3 days in vitro (DIV) and used for experiments at 6 DIV as described (Menna et al. 2009). In a set of experiments, 6 DIV neurons were preincubated for 1 h with a Gα inhibitor, GDP-β-S (1 mM), dissolved in 0.2% Pluronic F-68 dimethyl sulfoxide (DMSO, Meyer et al. 2000) before treating with kainate. Higher pluronic concentrations were avoided since more than 1% Pluronic F-68 DMSO-induced neuronal damage (data not shown).
The cultures were fixed and stained as previously described (Verderio et al. 2004). The images were acquired using a BioRad MRC-1024 and Zeiss LSM 510 META confocal microscopes.
RNA I nterference
GluA2 (GluR2) small interfering RNA (siRNA) experiments were performed as described previously (Liu et al. 1994; Passafaro et al. 2003). A sequence (5′-GATCCCCCACTGCAAGCTGTTCTGGATTCAAGAGATCCAGAACAGCTTGCAGTGTTTTTGGAAA-3′) ineffective in GluA2 silencing was also used as a control. GluA2 siRNA cDNA was a kind gift of Maria Passafaro (CNR, Institute of Neuroscience).
The plasmids pCDNA3-RIIRI-CFP and pCDNA3-C-YFP, coding, respectively, for the regulatory subunit of PKA fused to cyan fluorescent protein (CFP) and for the catalytic subunit of PKA fused to yellow fluorescent protein (YFP) were a gift from Zaccolo M. (University of Glasgow, UK).
Biochemical T echniques
Growth Cone Particles
Growth cones particles (GCPs) were prepared as described (Lockerbie et al. 1991; Lohse et al. 1996). Isolated GCPs were diluted in 7–8 volumes of modified Krebs’–Ringer’s solution (180 mM sucrose, 50 mM NaCl2, 5 mM KCl, 22 mM HEPES [pH 7.4], 10 mM glucose, 1.2 mM NaH2PO4, 1.2 mM MgCl) containing, if needed, an antagonist and kept on ice for 5 min. For stimulation, GCPs were transferred to 37 °C for 5–10 min, and then an agonist was added for 20 min. GCPs were processed by sodium dodecyl sulfate (SDS) gel electrophoresis followed by western blotting.
Synaptosome P reparation
The purification of synaptosomes from rat forebrain was carried out as described (Huttner et al. 1983). Samples with a protein content of 200 mg were pelleted and resuspended in KRH, containing antagonists, if needed. Samples were kept for 10 min on ice and then transferred for 10 min to 37 °C. After addition of agonist, they were left for 20 min at 37 °C, and then a fraction corresponding to 50 mg of protein was mixed with SDS containing sample buffer for phosphoprotein analysis.
Calcium I maging
Hippocampal cultures of 5–7 or 13 DIV were loaded with 5 μM Fura-2 pentacetoxy methylester in KRH for 45 min at 37 °C, washed in the same solution and transferred to the recording chamber of an inverted microscope (Axiovert 100; Zeiss, Oberkochen, Germany) equipped with a calcium imaging unit. Polychrome IV (TILL Photonics, Germany) as described (Verderio et al. 2004).
SV R ecycling
The 6 DIV hippocampal neurons (pre-exposed or not to Kainate and antagonist as indicated in the text) were incubated with antibodies against the intravesicular domain of synaptotagmin I (Syt-ecto SYSY) under depolarizing conditions (50 mM KCl) for a brief period (1–2 min) in order to avoid a saturation effect due to maximal recycling. Neurons were then washed, fixed, and stained for the internalized Syt-ecto with an appropriate secondary antibody and immunostained for vGLUT1 to identify excitatory terminals. Images were acquired using a Zeiss LSM 510 Meta confocal microscope and analyzed by using the Image J software. vGLUT1-positive recycling puncta were revealed by generating a binary mask of Syt-ecto/vGLUT double-positive images. The number of Syt-ecto–positive vGLUT1 puncta identified by the binary mask was calculated. Synapses were scored as positive for Syt-ecto antibody internalization when the fluorescence intensity was at least 2.5–3 times higher compared with cultures exposed only to secondary antibodies.
Photoactivation E xperiments
PAGFP:Synapsin I was a kind gift of Noam Ziv (Haifa, Israel). Hippocampal neurons were transfected at 3 DIV with calcium phosphate and imaged at 6–8 DIV. Images were recorded on a Leica TCS SP5 confocal microscope (Leica Microsystems, Wetzlar, Germany) equipped with a cell incubation chamber (OKOLAB, Naples, Italy) for temperature, humidity, and CO2 control. Photoactivation was performed by selectively scanning rectangular regions of interest (ROIs) along the axon at 405 nm using a violet diode laser. The 3D time-lapse recordings (1 Z-stack/2 min) of fluorescence images were obtained by exciting the photoactivated green fluorescent protein (GFP) at 488 nm (8000 Hz) using a ×40 objective (1.25 NA) and setting the voxel size to 151.5 × 151.5 × 250 nm. To consider all the information coming from the different focal planes of the axon, the maximal intensity projections of Z-stacks were analyzed using the ImageJ software. We identified the clusters of activated synapsin I in the ROI as the pixels that exceeded a given intensity value (threshold) at time 0 (F0) and measured the area of synapsin clusters over time (keeping the threshold constant). The area of synapsin clusters at 10 and 20 min was normalized with respect to F0.
Analysis of S ynapsin I D ispersion
To quantify the degree of synapsin I dispersion (dispersion index) in growth cones, we analyzed the profile of the mean intensity of fluorescence (mIF) of synapsin I staining measured along arbitrary lines in growth cones by using Image J software. We counted the number of peaks with respect of background level in the plot profile of mIF per micron (50 growth cones per each experimental condition) that reflects the number of synapsin I clusters in the growth cone. This value was normalized and the difference between treated and control indicated the dispersion index. For the quantification of synapsin I dispersion as cone fraction, cones with dispersion index more that 0.2 were considered as dispersed. All data are results of at least 3 independent experiments. Data were shown as means ± SEM.
Fluorescence Resonance Energy Transfer I maging of I ntracellular cAMP D ynamics
Images were obtained by using a 458- to 514-nm dichroic beamsplitter, and the META detector was set between 470 and 500 nm for CFP and between 530 and 600 nm for YFP. Under these conditions, no cross talk was observed. Fluorescence resonance energy transfer (FRET) was measured by acceptor photobleaching, where an increase in CFP signal (dequenching) during incremental photobleaching of YFP can be observed. Background fluorescence of both fluorophores was determined from a cell-free area of the image and subtracted from the overall intensity. The mean intensity was measured from regions of interest by using IMAGEJ (National Institutes of Health) and Zeiss LSM software. FRET efficiency was calculated as:
Whole-cell patch-clamp recordings were obtained from 13 DIV neurons with an Axopatch 200B amplifier and pClamp-10 software (Axon Instruments, Foster City, CA). Recordings were performed in the voltage-clamp mode. Currents were sampled at 2 kHz and filtered at 2–5 kHz. External solution (KRH) had the following composition (in mM): 125 NaCl, 5 KCl, 1.2 MgSO4, 1.2 KH2PO4, 2 CaCl2, 6 glucose, and 25 HEPES–NaOH, pH 7.4. Miniature excitatory postsynaptic currents (mEPSCs) were recorded in the presence of 1 μM tetrodotoxin (TTX). Recording pipettes were fabricated from capillary glass (World Precision Instruments) using a 2-stage puller (Narishige, Tokyo, Japan) and had tip resistances of 3–5 MΩ when filled with the intracellular solution of the following composition (in mM): 130 K-gluconate, 10 KCl, 1 ethyleneglycol-bis(2-amnioethylether)-N,N,N′N′-tetra acetic acid (EGTA), 10 HEPES, 2 MgCl2, 4 MgATP, and 0.3 Tris-GTP. Voltage-clamp recordings were performed with a holding potential of −70 mV. Recordings were performed at room temperature. Off-line analysis of mEPSCs used Clampfit-pClamp-10 software. Events had to exceed a threshold of 2 times the SD of the baseline noise. Neurons were pre-exposed to 200 nM KA in TTX-containing external solution for 20 min followed by wash. In a parallel set of experiments, neurons were exposed for 3 min to AMPA (50 μM) in KRH devoid of Ca2+ (Schenk et al. 2005), washed, and recorded in TTX-containing external solution.
Statistical analysis was performed using SigmaStat 3.5 software for Windows (Systat Software Inc., Germany). After testing whether data were normally distributed or not, the appropriate statistical test has been used: Student's t-test or Mann–Whitney rank’s test, one-way ANOVA or Kruskal–Wallis test, followed by post hoc analysis. Detailed information is reported in the figure legends. Statistical significance is indicated in graphs as follows: *P < 0.05; **P < 0.01, and ***P < 0.001.
Activation of KA R eceptors by KA I nduces D ispersion of S ynapsin I mmunoreactivity in the G rowth C ones of H ippocampal N eurons
The AMPA GluA2/3 (GluR2/3) receptor subunits and the kainate GluK5 (KA2) and GluK2/3 (GluK2/7) receptor subunits are expressed in GCPs isolated from fetal rat cortices. Both GluK2/3 and GluK5 subunits were heavily enriched in GCPs with respect to the total homogenate, along with the SV protein synaptotagmin I and the AMPA receptor subunit GluA2/3 (Fig. 1A). Calcium imaging recordings of 6 DIV hippocampal cultures, loaded with the Ca2+-sensitive dye Fura-2 and imaged by single-cell recording, revealed [Ca2+]i increases that originated at the axonal growth cone and propagated retrogradely toward the soma (Fig. 1B), upon stimulation with 10–50 μM KA in the presence of TTX. No calcium elevations were detected upon neuron exposure to KA concentrations lower than 10 μM (Supplementary Fig. S1A). KA-induced, but not AMPA-induced, [Ca2+]i increases were largely prevented by the rather selective KAR inhibitor UBP302 (Supplementary Fig. S1B,B′,C,C′). Thus, KARs expressed at the growth cones of cultured hippocampal neurons are functional, in line with previous results (Tashiro et al. 2003; Ibarretxe et al. 2007).
We have previously shown that synapsin I, a major presynaptic regulatory phosphoprotein involved in the assembly and maintenance of the reserve pool of SVs (Cesca et al. 2010; Fornasiero et al. 2010), is the substrate of presynaptic AMPA receptors activation (Schenk et al. 2005). We therefore aimed at investigating whether synapsin I may also be the substrate of KAR activation at the growth cone. We exposed hippocampal cultures to low KA concentrations (200 nM) for 20 min in the presence of TTX, washed, and stained for synapsin I. Synapsin I immunoreactivity, which is normally clustered in small puncta along the axon and in the core of the axonal growth cone, became fully dispersed and entered filopodia upon KA stimulation (Fig. 1C), as shown by quantification of either the immunoreactivity dispersion index (Fig. 1D) or the cone fraction (Fig. 1F) (see Materials and Methods for details). A similar protein dispersion was detected in neurons transfected with a photoactivatable variant of GFP-tagged synapsin I (Patterson and Lippincott-Schwartz 2002; Tsuriel et al. 2006), activated with high-intensity illumination at 405 nm and imaged upon KA exposure (Fig. 1E). Preincubation of cultured hippocampal neurons with the KAR blocker UBP302 also blocked the synapsin I relocation consequent to KA (Fig. 1C,D) in a dose-dependent manner (Supplementary Fig. S1D). Conversely, the AMPA receptor blocker GYKI53655 was completely ineffective in preventing synapsin dispersion consequent to 200 nM KA exposure (Fig. 1F). Synapsin I dispersion was detected even when cultures were exposed to 200 nM KA in the presence of EGTA or 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) (Supplementary Fig. S2A,B) or in a Ca2+-free Krebs’–Ringer’s solution, in which choline was substituted for sodium (Fig. 1F). Furthermore, kainate application did not induce significant dispersion of synapsin I immunoreactivity in cells treated with the general inhibitor of all Gα subunits GDP-β-S (1 mM) (Supplementary Fig. S2C). These data indicate that a KAR-dependent, calcium-independent, most likely G protein–dependent, pathway leading to synapsin I dispersion is activated by low doses of KA at the growth cone. Conversely, the dispersion of synapsin I immunoreactivity induced by high doses of KA (100 μM) was partially dependent on the presence of extracellular calcium and was partially inhibited by GYKI53655 (Supplementary Fig. 1G), besides being blocked by UBP302, indicating that high doses of KA also activate different pathways, possibly partially involving AMPA receptors (Patneau and Mayer 1991). On the other hand, reduction of GluA2 expression by siRNA did not prevent synapsin I dispersion and calcium responses induced by high doses of kainate, while completely inhibiting synapsin I dispersion and calcium transients induced by AMPA (Fig. 1H,I). These data indicate that KAR activation at the growth cone induces synapsin I relocation and provide evidence that distinct glutamate receptor types converge on the same molecular target at the presynaptic level. They also show that dispersion of synapsin I upon 200 nM KA results from selective activation of KAR; for this reason, most of the experiments in the present study were carried out using this agonist concentration.
Dispersion of S ynapsin I mmunoreactivity by KA I nvolves PKA A ctivation
It is well known that synapsins are regulated by distinct protein kinases acting on multiple phosphorylation sites of the protein (De Camilli et al. 1983; Chi et al. 2001, 2003; Bonanomi et al. 2005; Menegon et al. 2006; Cesca et al. 2010). In order to gain insights into the molecular pathways coupling KAR activation to synapsin I relocation, we first tested the possible involvement of the Erk/MAPK pathway in this process. Opposite to presynaptic AMPA receptors signaling, the KA-induced dispersion of synapsin I was not accompanied by MAPK activation, as indicated by immunofluorescence (Fig. 2A,B) or western blotting (Fig. 2C) with antibodies recognizing the phosphorylated form of MAPK (p-MAPK). Also, AMPA-induced, but not KA-induced, synapsin I relocation was inhibited by the specific MAPK inhibitor PD98059 (Fig. 2D), thus indicating that the MAPK signaling cascade is not involved in the KAR pathway.
We next checked for the possible involvement of PKA in the KA-induced synapse dispersion. Thirty minutes preincubation of cultured hippocampal neurons with the PKA inhibitor H89 (10 μM) largely reduced synapsin I dispersion induced by 200 nM KA, thus suggesting the potential involvement of PKA in this effect (Fig. 3A). H89 only partially inhibited synapsin I dispersion induced by 100 μM KA or by the potent PKA activator, forskolin (Fig. 3A), possibly due to the fact that additional signaling pathways may be activated downstream massive cAMP production, that is, EPAC and CaMKII (Pereira et al. 2007; Willoughby and Cooper 2007). The direct demonstration that synapsin I dispersion upon KA stimulation is in fact produced by phosphorylation at PKA phosphorylation sites, was obtained by 2 independent approaches. First, 6 DIV hippocampal neurons were treated with 200 nM KA for 20 min in Ca2+-free KRH, fixed, and labeled with an antibody directed against synapsin I phosphorylated by PKA at Ser9 (site 1), in combination with anti synaptobrevin/VAMP2 antibodies to visualize axonal growth cones; 200 nM KA induced a significant synapsin phosphorylation and increased the percentage of growth cones immunopositive for PKA–phosphosynapsin in a dose-dependent manner (Fig. 3B), an effect which was accompanied by synapsin dispersion (Fig. 3C). Second, hippocampal neurons were transfected with a cDNA encoding a synapsin I point mutant (Ser9Ala; syn Imut) that cannot be phosphorylated by PKA. Figure 3D,E shows that the PKA nonphosphorylatable synapsin I mutant is not relocalized by KA treatment, although it is efficiently dispersed by AMPA stimulation, which acts on the MAPK/Erk pathway (Lourenço et al. 2011).
To directly investigate whether the PKA signaling pathway is activated downstream of KAR, we took advantage of the genetically encoded biosensor for cAMP (Zaccolo et al. 2000; Evellin et al. 2004). The exogenous expression of the regulatory and catalytic subunits of the PKA, fused with CFP and YFP, respectively, are used for measuring changes in FRET that result from changes in intracellular cAMP levels. Indeed, at low intracellular cAMP concentrations, PKA is in the inactive tetrameric form, CFP and YFP are in close proximity, and FRET is maximal. When cAMP levels rise, the catalytic and regulatory subunits dissociate, thus leading to PKA activation, and FRET decreases. FRET was monitored by following the YFP/CFP fluorescence emission ratio (Zaccolo and Pozzan 2002; Evellin et al. 2004; Lissandron et al. 2005; Zaccolo et al. 2005). Cultured hippocampal neurons were transiently transfected with the PKA-based sensor. Upon exposure to 25 μM forskolin and 100 μM IBMX, a phosphodiesterase inhibitor, a decrease in YFP/CFP fluorescence emission ratio was detected as a consequence of the maximal generation of cAMP (Dal Molin et al. 2008; Dinant et al. 2008; Fig. 4A). To quantify changes of FRET efficiency consequent to KA treatments, the acceptor photobleaching method was applied that allows measuring the increase of CFP fluorescence emission following YFP photobleaching (Amiri et al. 2003; Wallrabe et al. 2006; Dinant et al. 2008; Gupta and Srivastava 2009). Transfected neurons were stimulated with 25 μM forskolin or KA in the presence of IBMX for 3 min, fixed, and processed for acceptor photobleaching. FRET experiments were performed in neurons stimulated with 100 μM KA in the presence of TTX, to increase the probability that transfected growth cones also responded to KA stimulation. The results showed an FRET efficiency (E) in the soma of 12.09% ± 0.55 on average under resting conditions (Fig. 4B,C), which decreased to 3.40% ± 0.36 upon forskolin and to 6.81% ± 0.42 upon KA treatments; in growth cones, E was 10.60% ± 0.64 under resting conditions and decreased to 5.75% ± 0.89 upon KA treatment (Fig. 4C′). These data indicate that activation of KA receptors leads to generation of cAMP and PKA activation.
KA I nduces I ncreases in SV E xocytosis before and a fter S ynaptogenesis
To evaluate whether KA-induced PKA activation and synapsin dispersion is accompanied by an increase in the rate of SV recycling, 6 DIV neuronal cultures were exposed to 200 nM KA in the presence of TTX, washed, and then incubated with antibodies to the luminal domain of synaptotagmin (Syt-ecto), in the presence of 50 mM KCl. Figure 5A,B indicate that preincubation with 200 nM KA enhances the rate of SV recycling in isolated axons of developing neurons, in a UBP302- and H89-sensitive manner. The KA-induced enhancement of SV fusion was also maintained after the establishment of synaptic contacts, as indicated by the analysis of mEPSCs in 13 DIV hippocampal cultures. Culture incubation with 200 nM KA in TTX-containing external solution for 20 min, followed by wash and subsequent recording, revealed a statistically significant increase in mEPSC frequency relative to untreated neurons (Fig. 5C,D, mEPSCs frequency, Ctr, normalized value = 1.02 ± 0.07, n = 25; mEPSCs frequency, KA, normalized value = 1.39 ± 0.09, n = 26); this effect was completely prevented by pretreatment of neuronal cultures with the PKA inhibitor H89 (10 μM; mEPSCs frequency, KA plus H89, normalized value = 0.90 ± 0.11 n = 13; Fig. 5D). No significant change in the amplitude of miniature events was detected, indicating a predominant involvement of the presynaptic compartment in the KA effects (mEPSC amplitude Ctr = 32.03 pA ± 2.17; amplitude KA = 36.94 pA ± 2.45; Mann–Whitney rank sum test, P = 0.087). Differently from kainate, pre-exposure of cultures to 50 μM AMPA (Schenk et al. 2005) did not affect mEPSC frequency (mEPSCs frequency, Ctr, normalized value = 1 ± 0.17419, n = 12; AMPA = 1.11 ± 0.15495, n = 9; t-test, P = 0.971). These data indicate that transient stimulation of KARs, but not AMPA receptors, can facilitate spontaneous release through PKA activation by increasing SV recycling, possibly through a priming effect that persists in the absence of KA.
Glutamate A ctivates KARs and S timulates the PKA P athway
We next investigated whether exogenous application of the natural agonist, glutamate, might be able to activate PKA, through the activation of KARs; 6 DIV hippocampal neurons were incubated with glutamate concentrations ranging from 200 to 300 μM (data not shown), and the dispersion of synapsin immunoreactivity, together with its sensitivity to different blockers, was analyzed. Among the tested doses, we found that 500 nM glutamate, applied in the absence of extracellular calcium for 20 min, induced a significant dispersion of synapsin I immunoreactivity which was prevented by the KARs inhibitor UBP302 or the PKA inhibitor H89, but not by the MAPK inhibitor PD98059 or the AMPA inhibitor GYKI53655 (Fig. 6A). These data indicate that glutamate may induce synapsin I dispersion by selectively recruiting presynaptic KAR, through the PKA pathway. Accordingly, preincubation of 13 DIV neurons with 500 nM glutamate for 20 min, followed by extensive washes, resulted in increase in mEPSC frequency, which was inhibited by the KARs inhibitor UBP302 (Fig. 6B).
In the mature nervous system, presynaptic KARs play crucial roles in the control of transmitter release and in plasticity phenomena (Lerma 2003; Pinheiro and Mulle 2008; reviewed in Contractor et al. 2011). Early in development, KA regulates the motility of axonal filopodia (Chang and De Camilli 2001), with low KA concentrations increasing filopodia motility and high KA concentrations having the opposite effect (Tashiro et al. 2003), most likely by increasing spiking activity (Ibarretxe et al. 2007). Thus, an increased local concentration of glutamate at sites of synapse formation may play a role, via an autocrine mechanism, in stabilizing the nascent contact, promoting a transition from filopodia to a mature synapse (Chang and De Camilli 2001; Tashiro et al. 2003). The inappropriate activation of KARs may be involved in excitotoxic processes and epilepsy (Sander et al. 1995; Rodríguez-Moreno et al. 1997; Smolders et al. 2002; Epsztein et al. 2005; Sloviter et al. 2006; Vincent and Mulle 2009). While some actions of KARs are classically ionotropic, others seem to involve the activation of second-messenger cascades, involving PKC, PKA, and MAPK (Rodriguez-Moreno and Lerma 1998; reviewed in Contractor et al. 2011 and Rodriguez-Moreno and Sihra 2007), a process that could involve the action of neuromodulatory agents whose release is stimulated by KAR activation (reviewed in Contractor et al. 2011). Although the notions of noncanonical metabotropic signaling of KARs and of their action in the modulation of transmitter release are established, the identity of the stimuli activating this pathway is still undefined. The present study identifies a novel presynaptic target and a signaling pathway for the presynaptic action of KAR activation by demonstrating that, in developing axons and axonal growth cones of cultured hippocampal neurons, activation of KAR induces synapsin I dispersion via activation of PKA, as directly indicated by transfection of a genetically encoded fluorescent sensor, which allows the monitoring of changes in intracellular cAMP concentrations by FRET as well as by transfection of a PKA nonphosphorylatable synapsin I mutant. KA-induced relocation of synapsin I was inhibited by the KAR blocker UBP302, a known antagonist of GluK1 (More et al. 2004) as well as of GluK3 (Perrais et al. 2009) KARs. However, this drug has been recently reported to be also an antagonist of recombinant GluK2/GluK5 receptors (P Pinheiro, C Mulle, unpublished observation). This suggests that not only GluK1 or GluK3 but also GluK2/GluK5-containing KAR may be involved in this process.
Activation of PKA/Ca2+/calmodulin–sensitive PKC upon pharmacological activation of KARs has been shown at mossy fibers–CA3 synapses and linked to facilitation of synaptic transmission (Rodriguez-Moreno and Sihra 2004). It has been proposed that facilitation might depend on Ca2+ influx, occurring through Ca2+-permeant KARs or intracellular Ca2+ release, which might activate the Ca2+/calmodulin–sensitive AC isotypes AC1 and AC8 (reviewed in Rodríguez-Moreno and Sihra 2007). This facilitatory action is well in line with the role of PKA in presynaptic facilitation at mossy fiber–CA3 synapses (Weisskopf et al. 1994). It remains to be demonstrated that physiological activation of presynaptic KARs by endogenous release of glutamate also relies on direct PKA-signalling mechanisms. Interestingly, we show that PKA activation by KAR does not necessarily require intracellular calcium increases. Although the specific molecular mechanism allowing PKA activation by presynaptic KAR stimulation remains to be identified, our results further support the notion that KARs can be engaged in metabotropic functions, possibly through the involvement of a G protein. Coupling to transduction pathways, independently of channel receptor activity, may be a common feature of ionotropic glutamate receptors. Indeed, AMPA receptors are also endowed with metabotropic properties (Wang et al. 1997) as their activation stimulates MAPK through a nonreceptor tyrosine kinase-dependent mechanism (Schenk et al. 2005), possibly involving the nonreceptor tyrosine kinase lyn (Hayashi et al. 1999).
One of the major accomplishments of the present study is the demonstration that KAR-induced PKA activation results in phosphorylation of synapsin I. Synapsins are SV-associated phosphoproteins that represent the major presynaptic targets of PKA during development and in mature neurons (reviewed in Cesca et al. 2010; Fornasiero et al. 2010). A plethora of data point to an important role of PKA-mediated phosphorylation of synapsin I in synapse maturation and function. Thus, at early stages, synapsin phosphorylation at site 1 by PKA and CaMK I/IV modulates axon elongation and synapse formation (Kao et al. 2002; Perlini et al. 2011). Previous work has also shown that the state of association of synapsin I with SVs in the growth cone compartment is also entirely dependent on site 1 phosphorylation (Bonanomi et al. 2005). As a consequence of synapsin I detachment, SVs are mobilized, moving from the central compartment of the growth cone to the periphery. cAMP-triggered PKA activation seems therefore to be one of the most important players in the modulation of SV distribution in the growth cone, before the onset of synaptic specialization.
A causal relationship has been shown to occur between synapsin I phosphorylation and enhancement of the neurosecretory response, as indicated by the findings that in synapses expressing the PKA nonphosphorylatable mutant of synapsin I, the size of the recycling SV pool and the rate of evoked SV exocytosis were reduced, with a more intense depression and a much slower recovery after sustained high frequency stimulation (Menegon et al. 2006). All together, these observations suggest a role for PKA phosphorylation in the supply of SVs from the reserve pool, thus affecting both the steady-state level of activity and the reconstitution of the readily releasable pool after intense stimulation. However, thus far nothing was known concerning the possible physiological stimuli that initiate this sequence of events. The demonstration that glutamate, via KAR activation, induces PKA-mediated synapsin phosphorylation, SV dispersion, and increase in recycling adds an important piece of information to our knowledge on the physiology of the growth cone.
This pathway appears to be active not only at the axonal growth cone but also after the establishment of synaptic contacts, when it increases SVs availability, possibly contributing to plasticity phenomena. In this respect, it has previously been shown that PKA-mediated synapsin I phosphorylation could be a fundamental step in the mechanism by which mature synapses resist to fatigue and recover from depression (Menegon et al. 2006). The present work for the first time shows that this important sequence of events can be initiated by glutamate via KARs. Further work is required to identify which KAR subunits are involved in the present process and characterize their functional properties that vary according to subunit composition (Contractor et al. 2011).
The finding that stimulation of both KAR and AMPA receptors, via activation of PKA and MAPK, respectively, impacts synapsin I phosphorylation indicates that distinct glutamate receptor types converge on the same molecular target at the presynaptic level. Since AMPA and KA receptors are coexpressed in the same growth cone, it is tempting to speculate that a preferential activation of glutamate receptor subtypes is linked to the agonist concentration; indeed, we provide evidence that low glutamate concentrations mainly act on presynaptic KARs and induce a PKA-dependent synapsin phosphorylation.
Being excellent substrates for multiple protein kinases, synapsins are at the convergence of several signalling pathways impinging on a nerve terminal and activated by distinct extracellular messengers. The observation that glutamate, acting on different presynaptic receptors, leads to synapsin phosphorylation on distinct sites by different kinases may have important functional consequences. Although its phosphorylation on serine residues by either PKA or MAPK triggers apparently similar effects on the relocation of the protein and on the availability of SVs for release, it is known from in vitro studies that PKA-mediated synapsin phosphorylation affects the interaction of the protein with both actin and SVs, whereas MAPK-mediated phosphorylation of synapsin I on sites 4, 5, and 6 moderately decreases its interaction with actin, while having no effect on its interaction with SVs. Moreover, the levels of activation of PKA and MAPK are modulated by other extracellular signals impinging on the terminal, which may activate them in a differential manner, and the time courses of synapsin phosphorylation by the 2 kinases in mature neurons are completely different, given the complex regulation by electrical activity and [Ca2+]i of the balance between kinase and phosphatase activation (reviewed in Cesca et al. 2010). In fact, it is known that depolarization-induced Ca2+ entry into the nerve terminal increases phosphorylation of synapsin I on the CaMKI/PKA site, whereas the MAPK sites are first dephosphorylated by calcineurin and subsequently phosphorylated by activation of MAPK (Jovanovic et al. 2001; Menegon et al. 2006). Thus, synapsins might act as central integrators of distinct signal transduction pathways impinging on a nerve terminal, and the interplay between stimulation of presynaptic KAR and/or AMPA receptors may have a key role in the fine activity-dependent tuning of nerve terminal function.
Altogether these data raise the possibility that at early developmental stages, when neurons are still immature and glutamate is present in the environment at low concentrations, activation of KARs first occurs at the growth cones (Tashiro et al. 2003). KARs activation may increase filopodia motility (Tashiro et al. 2003) and induce PKA-mediated synapsin I phosphorylation and SV dispersion, leading to increased vesicle recycling (see also Menegon et al. 2006). Limitation of the extracellular space around the nascent synapse during neuronal development may then favor the rise of extracellular glutamate to the levels required to activate AMPA receptors and trigger MAPK activation at immature presynaptic sites (Schenk et al. 2005). High glutamate concentrations may then reduce filopodia motility, promoting stabilization of the contact (Tashiro et al. 2003; Ibarretxe et al. 2007). Ca2+entry, occurring mainly through KARs themselves upon high agonist concentrations, may cooperate with activation of the MAPK and PKA pathways in the regulation of motility and secretion. Thus, glutamate secreted either by neuronal activity or neighboring astrocytes may act as an autocrine/paracrine factor, regulating the growth cone and filopodia and favoring the transition from filopodia to presynaptic structures during the process of synaptogenesis (Chang and De Camilli 2001; Goda and Davis 2003; Tashiro et al. 2003) or modulating network activity (Liu et al. 2004).
European Union 7th Framework Program (EUSynapse Integrated Project to M.M. and C.M.) under grant agreement no. HEALTH F2-2009-241498 (EUROSPIN Project to M.M.); Compagnia di San Paolo, Torino (to M.M., F.B., and F.V.); PRIN 2008 2008T4ZCNL (to M.M., F.B., and F.V.); and Telethon-Italy (grant GGP09134 to F.B. and F.V.).
Supplementary Figures S1 and S2 can be found at: http://www.cercor.oxfordjournals.org/
We wish to thank Dr J. Lerma (Consejo Superior de Investigaciones Científicas and Universidad Miguel Hernández) for discussion and for kindly sharing the GYKI53655 inhibitor, and Dr N. Ziv (The Rappaport Family Institute for Research in the Medical Sciences, Technion Faculty of Medicine) for the PAGFP:Synapsin I construct. Conflict of Interest: None declared.