Phosphatidylinositol 3-Kinase Couples Localised Calcium Influx to Activation of Akt in Central Nerve Terminals
- 1k Downloads
The efficient retrieval of synaptic vesicle membrane and cargo in central nerve terminals is dependent on the efficient recruitment of a series of endocytosis modes by different patterns of neuronal activity. During intense neuronal activity the dominant endocytosis mode is activity-dependent endocytosis (ADBE). Triggering of ADBE is linked to calcineurin-mediated dynamin I dephosphorylation since the same stimulation intensities trigger both. Dynamin I dephosphorylation is maximised by a simultaneous inhibition of its kinase glycogen synthase kinase 3 (GSK3) by the protein kinase Akt, however it is unknown how increased neuronal activity is transduced into Akt activation. To address this question we determined how the activity-dependent increases in intracellular free calcium ([Ca2+]i) control activation of Akt. This was achieved using either trains of high frequency action potentials to evoke localised [Ca2+]i increases at active zones, or a calcium ionophore to raise [Ca2+]i uniformly across the nerve terminal. Through the use of either non-specific calcium channel antagonists or intracellular calcium chelators we found that Akt phosphorylation (and subsequent GSK3 phosphorylation) was dependent on localised [Ca2+]i increases at the active zone. In an attempt to determine mechanism, we antagonised either phosphatidylinositol 3-kinase (PI3K) or calmodulin. Activity-dependent phosphorylation of both Akt and GSK3 was arrested on inhibition of PI3K, but not calmodulin. Thus localised calcium influx in central nerve terminals activates PI3K via an unknown calcium sensor to trigger the activity-dependent phosphorylation of Akt and GSK3.
KeywordsEndocytosis Vesicle Presynapse Calcium Exocytosis Akt GSK3
Central nerve terminals rely on the efficient recycling of a limited supply of synaptic vesicles (SVs) in order to maintain neurotransmission. This is achieved via recruitment of a number of SV endocytosis modes [1, 2]. Clathrin-mediated endocytosis is dominant during low frequency stimulation and generates SVs directly from the plasma membrane whereas ultrafast endocytosis can be triggered by single action potentials at physiological temperatures and generates SVs via an endosomal intermediate [3, 4, 5]. During higher stimulation intensities an additional endocytosis mode is triggered to increase retrieval capacity called activity-dependent bulk endocytosis (ADBE). ADBE generates large endosomes direct from the plasma membrane. These bulk endosomes then donate SVs back to the recycling pool via clathrin-dependent mechanisms [2, 6].
Triggering of ADBE is dependent on the dephosphorylation of the large GTPase dynamin I at two serine residues (Ser774 and Ser778) . This event is mediated by the calcium-dependent protein phosphatase calcineurin and only occurs during intense activity . ADBE is maintained across multiple trains of stimuli via the rephosphorylation of dynamin I by cyclin-dependent kinase 5 on Ser778 and then subsequently by glycogen synthase kinase 3 (GSK3) on Ser774 .
GSK3 has a high basal activity which is regulated via its phosphorylation status . Phosphorylation of Ser21/9 of GSK3β/α by upstream cascades inhibits this protein kinase activity. We have shown that GSK3 is phosphorylated in central nerve terminals only during high levels of neuronal activity, coincident with dynamin I dephosphorylation . We also identified that the activity-dependent GSK3 kinase was Akt (also known as protein kinase B). This activity-dependent inhibitory signalling cascade ensures that Ser774 on dynamin I is maximally dephosphorylated during periods of intense activity . Importantly, either expression of constitutively active Akt or addition of signalling molecules such as brain-derived neurotrophic factor (BDNF, which activates Akt) reduces ADBE in cultured neurons [10, 11] suggesting that this activity-dependent signalling cascade is essential for efficient presynaptic function during intense neuronal activity.
It is currently unclear how action potential stimulation is transduced into activation of Akt. To address this we monitored the phosphorylation status of both Akt and GSK3 after elevating intracellular free calcium ([Ca2+]i) via either action potentials or a calcium ionophore. We found that localised [Ca2+]i increases via voltage gated calcium channels were essential for Akt activation. This activation was dependent on phosphatidylinositol 3-kinase (PI3K) but not calmodulin activity. This reveals a key role for PI3K in the activity-dependent control of GSK3 activity, but leaves the identity of the calcium sensor still undetermined.
Materials and Methods
Foetal bovine serum was from Biosera. Penicillin/streptomycin and Minimum Essential Medium were from Invitrogen. The primary antibodies phospho-Akt Ser473 and GSK3α/β Ser21/9 were from Cell Signalling. Bovine serum albumin was from Roche. BAPTA-AM, CNQX and AP-5 were from Tocris. EGTA-AM was from Cambridge Bioscience and ionomycin from Merck Chemicals. Secondary antibodies, β-Actin antibody, LY294002, wortmannin and all other chemicals were bought from Sigma.
Cerebellar Granule Neuron Cultures
Cerebellar granule neuron (CGN) cultures were prepared from the cerebella from P7 Sprague–Dawley rat pups of either sex . Cultures were used between 8 and 14 days in vitro (DIV).
CGN Lysis Experiments
Cerebellar granule neurons (CGNs) were repolarised in incubation medium (170 mM NaCl, 3.5 mM KCl, 400 µM KH2PO4, 20 mM TES [N-tris(hydroxy-methyl)-methyl-2-aminoethane-sulfonic acid], 5 mM NaHCO3, 5 mM glucose, 1.2 mM Na2SO4, 1.2 mM MgCl2, 1.3 mM CaCl2 at pH 7.4) for 10 min. This medium was then replaced by media supplemented with (2R)-amino-5-phosphonovaleric acid (AP-5, 50 µM), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 10 µM) and various antagonists as described in the respective figure legends. For electrical field stimulation experiments coverslips were placed in a Warner imaging chamber (RC-21 BRFS) and either challenged with a train of 800 action potentials delivered at 80 Hz or no stimulus (basal). In experiments using ionomycin, cells were treated in an identical manner, with the exception that the challenge with 2.5 µM ionomycin or a vehicle control (DMSO) was for 1 min. In all experiments antagonists were present during stimulation with the exception of those using either BAPTA-AM or EGTA-AM. Low calcium incubation medium had the following alterations, CaCl2 was reduced to 100 µM, MgCl2 was increased to 10 mM and EGTA was added at 50 µM. After stimulation with either action potentials or ionomycin, CGNs were immediately lysed in 25 µl of SDS sample buffer (67 mM Tris, 9.3 % glycerol, 2 mM EGTA, 67 mM SDS, bromophenol blue, 12 % β-mercaptoethanol) and boiled at 95 °C for at least 5 min.
Cerebellar granule neuron (CGN) lysates were subjected to Western blotting as described before . Primary antibodies were all used at 1:1000, apart from HRP-coupled Anti-β-Actin (1:50,000). Secondary antibodies were used at 1:5000 except for Anti-β-Actin where no secondary was required. The intensity of the detected chemiluminescence signal was determined using Image J (NIH). Raw densitometry values were normalised firstly to the highest value within an individual blot and then to individual corresponding loading controls (β-Actin, which was also normalised to the highest value). All Western blot data is displayed after normalisation to the basal signal from each blot. All data was analysed using GraphPad Prism and with students t tests performed between basal and stimulated states as well as between stimulated states across conditions.
An Intracellular Free Calcium Increase is Essential for Phosphorylation of Akt/GSK3
A Localised [Ca2+]i Increase is Essential for Akt/GSK3 Phosphorylation
PI3K Activity is Required for Activity-Dependent Akt and GSK3 Phosphorylation
The maintenance of neurotransmission during intense neuronal activity is reliant on the integration of activity-dependent calcium influx, multiple signalling cascades and several SV endocytosis modes [1, 2]. One such cascade is the activity-dependent inhibition of GSK3 via Akt, which is essential for the efficient dephosphorylation of dynamin I, a key event in the triggering of ADBE [10, 11]. We show here that localised calcium influx via voltage-gated channels is essential for this event to occur. Furthermore we have demonstrated a key role for PI3K, but not calmodulin, in transduction of activity-dependent calcium influx into activation of Akt.
PI3K is implicated in the regulation of secretion and membrane trafficking in a variety of systems [18, 19, 20, 21]. One member of this family, PI3K-Class 2 α (PI3K-C2α), regulates exocytosis either indirectly via modulation of signalling pathways  or directly via control of the molecular fusion machinery [23, 24]. PI3K-C2α enzymatic activity is stimulated by calcium  and the N-terminus contains a clathrin binding site required for recruitment of the vesicle uncoating machinery . This suggests that PI3K-C2α may be a good candidate involved in the transduction of activity-dependent calcium influx to the Akt signalling cascade in this work. However, PI3K-C2α is insensitive to LY294002 and wortmannin at the concentrations used in this study , making it unlikely that PI3K-C2α is involved in the activity-dependent activation of Akt. LY294002 and wortmannin sensitive Class I PI3Ks have been implicated at several stages of SV recycling in central nerve terminals, including replenishment of the readily releasable pool (RRP) from the reserve pool , generation of bulk endosomes  and SV generation from bulk endosomes . PI3K is also required for the negative regulation of ADBE via BDNF . Therefore it would be of interest to determine the activity-dependent role of PI3K in both dynamin I dephosphorylation and ADBE.
Activity-Dependent Akt Phosphorylation at Nerve Terminals
We assessed the inhibition of GSK3 and activation of Akt respectively by monitoring their phosphorylation status. Inhibition of GSK3 occurs via phosphorylation of either Ser9 of GSK3α or Ser21 of GSK3β . In this study we found that GSK3α and GSK3β behaved in an identical manner across all stimulation conditions and pharmacological manipulations. We assessed activation of Akt by determining the phosphorylation status of Ser473. Phosphorylation of both Ser473 and Thr308 are required for maximal Akt activity [31, 32] and we have previously shown that the phosphorylation status of both sites respond in a similar manner to action potential stimulation in CGNs . These sites are proposed to be phosphorylated by different protein kinases however. Thr308 is a PDK1 substrate, which is activated by PI3K [31, 33], whereas Ser473 can be phosphorylated by either the mTORC2 complex  or DNA-PK . However we show that the PI3K antagonists LY294002 and wortmannin inhibit the activity-dependent phosphorylation of Ser473. This suggests that Akt phosphorylation is mediated via PI3K at central nerve terminals and studies that attempt to model this process should take this into account.
Akt and GSK3 have a wide cellular distribution, suggesting that the increased phosphorylation we observed during action potential stimulation may not be occurring at nerve terminals. We contend that these events are presynaptic, since identical stimulation protocols and modulation of both Akt and GSK3 activity-dependent phosphorylation directly impact on the phosphorylation status of dynamin I and ADBE [10, 11]. Since both dynamin I expression and ADBE is restricted to nerve terminals [2, 35] we argue that the activity-dependent changes in Akt and GSK3 phosphorylation are presynaptic. This statement is almost certainly not true for ionomycin-evoked changes, since this ionophore will raise [Ca2+]i uniformly throughout the neuron. Ionomycin is used throughout this study as a positive control to confirm the specificity of both flunarizine and EGTA-AM in inhibiting calcium channels and global calcium increases respectively. The inhibition of ionomycin-evoked Akt and GSK3 phosphorylation by calmidazolum is intriguing however and may reflect altered activity of these enzymes by postsynaptic calcium, where Akt and GSK3 are known to have postsynaptic functions [36, 37].
A number of studies have investigated the relationship between both neuronal activity and calcium influx in the activation of Akt. Previous work in CGNs showed that chronic hyperkalemic stimulation resulted in increased Akt phosphorylation, however this was insensitive to PI3K antagonists . This most likely reflects postsynaptic activation of Akt, since permanent depolarisation will reverse electrogenic glutamate transporters on the plasma membrane  raising extracellular glutamate concentrations and thus activating postsynaptic glutamate receptors. In agreement, shorter periods of stimulation with elevated KCl resulted in increased Akt phosphorylation which was reversed by PI3K antagonists .
Transduction of Calcium Influx into Akt Phosphorylation
We have shown that intense neuronal activity causes Akt phosphorylation via localised calcium microdomains. This suggests that a sensor must be located within the vicinity of this microdomain to couple calcium influx to PI3K/Akt activation. We have shown that calmodulin is unlikely to be the sensor since the antagonist calmidazolium was unable to inhibit action potential-evoked Akt phosphorylation. Furthermore the ability of calmidazolium to arrest ionomycin-evoked Akt phosphorylation is in agreement with the proposed location of calmodulin in the wider cytosol .
How could activity-dependent calcium influx be coupled to Akt phosphorylation? One potential candidate for the activity-dependent calcium sensor is the family of Ras-GRFs . Ras-GRF activates the small G-protein Ras in a calcium-dependent manner  and overexpression of dominant negative Ras inhibits Akt phosphorylation evoked by KCl depolarisation in peripheral neurons . Importantly Ras-GRF2 knockout mice display altered paired pulse facilitation, suggesting it is localised in the correct location to be activated by presynaptic calcium influx . However the mechanism via which Ras-GRF is activated by calcium influx is thought to be via calcium-bound calmodulin binding to its IQ domain [42, 44]. Since calmodulin inhibition had no effect on activity-dependent Akt phosphorylation, it may be that Ras-GRF is not the sensor for this event.
A second possibility is that there may not be calcium a sensor at all. In this model the requirement for localised calcium increases is an indirect result of the presynaptic release of growth factors such as BDNF. This event only occurs in neurons during intense neuronal activity, very similar to that observed in this study . This is potentially important for Akt phosphorylation, since PI3K can activate either PDK1 or the mTORC2 complex via growth factor receptors . In support, we have shown that exogenous addition of BDNF triggers phosphorylation of Ser473 in CGNs via a potential PI3K-dependent route . Thus our manipulations to either block calcium entry or buffer the calcium microdomain should also arrest BDNF release during intense activity, stopping it feeding back to increase Akt phosphorylation via PI3K. This hypothesis is testable using agents that sequester extracellular BDNF and we are currently investigating the effect of these agents on activity-dependent Akt phosphorylation .
In summary, we have shown that during intense neuronal activity phosphorylation of Akt (and subsequent inhibition of GSK3) is dependent on localised calcium microdomains formed by voltage-gated calcium channels. The identity of the sensor or release mechanism which transduce increased [Ca2+]i remains unidentified. This is an important question to address, since we predict that modulation of these events should impact on the extent of ADBE in central nerve terminals.
This work was supported by the Medical Research Council (Grant Number: G1002117) and a Ph.D. studentship from the Biotechnology and Biological Sciences Research Council.
- 8.Clayton EL, Sue N, Smillie KJ, O’Leary T, Bache N, Cheung G, Cole AR, Wyllie DJ, Sutherland C, Robinson PJ, Cousin MA (2010) Dynamin I phosphorylation by GSK3 controls activity-dependent bulk endocytosis of synaptic vesicles. Nat Neurosci 13(7):845–851. doi: 10.1038/nn.2571 CrossRefPubMedPubMedCentralGoogle Scholar
- 23.Meunier FA, Osborne SL, Hammond GR, Cooke FT, Parker PJ, Domin J, Schiavo G (2005) Phosphatidylinositol 3-kinase C2alpha is essential for ATP-dependent priming of neurosecretory granule exocytosis. Mol Biol Cell 16(10):4841–4851. doi: 10.1091/mbc.E05-02-0171 CrossRefPubMedPubMedCentralGoogle Scholar
- 24.Dominguez V, Raimondi C, Somanath S, Bugliani M, Loder MK, Edling CE, Divecha N, da Silva-Xavier G, Marselli L, Persaud SJ, Turner MD, Rutter GA, Marchetti P, Falasca M, Maffucci T (2011) Class II phosphoinositide 3-kinase regulates exocytosis of insulin granules in pancreatic beta cells. J Biol Chem 286(6):4216–4225. doi: 10.1074/jbc.M110.200295 CrossRefPubMedPubMedCentralGoogle Scholar
- 25.Wen PJ, Osborne SL, Morrow IC, Parton RG, Domin J, Meunier FA (2008) Ca2+-regulated pool of phosphatidylinositol-3-phosphate produced by phosphatidylinositol 3-kinase C2alpha on neurosecretory vesicles. Mol Biol Cell 19(12):5593–5603. doi: 10.1091/mbc.E08-06-0595 CrossRefPubMedPubMedCentralGoogle Scholar
- 37.Peineau S, Taghibiglou C, Bradley C, Wong TP, Liu L, Lu J, Lo E, Wu D, Saule E, Bouschet T, Matthews P, Isaac JT, Bortolotto ZA, Wang YT, Collingridge GL (2007) LTP inhibits LTD in the hippocampus via regulation of GSK3beta. Neuron 53(5):703–717. doi: 10.1016/j.neuron.2007.01.029 CrossRefPubMedGoogle Scholar
- 45.Matsuda N, Lu H, Fukata Y, Noritake J, Gao H, Mukherjee S, Nemoto T, Fukata M, Poo MM (2009) Differential activity-dependent secretion of brain-derived neurotrophic factor from axon and dendrite. J Neurosci 29(45):14185–14198. doi: 10.1523/JNEUROSCI.1863-09.2009 CrossRefPubMedPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.