ISSN 18197124, Neurochemical Journal, 2015, Vol. 9, No. 2, pp. 101–107. © Pleiades Publishing, Ltd., 2015. Original Russian Text © E.O. Tarasova, A.E. Gaydukov, O.P. Balezina, 2015, published in Neirokhimiya, 2015, Vol. 32, No. 2, pp. 123–130.
EXPERIMENTAL ARTICLES
Methods of Activation and the Role of Calcium/CalmodulinDependent Protein Kinase II in the Regulation of Acetylcholine Secretion in the Motor Synapses of Mice E. O. Tarasova, A. E. Gaydukov1, and O. P. Balezina Moscow State University, Moscow, Russia Received December 2, 2014
Abstract—In electrophysiological studies of neuromuscular junctions of mice, it was demonstrated that evoked release of mediator triggered by calcium influx via P/Q calcium channels in the terminals remains unchanged in the presence of KN62, a blocker of calcium/calmodulindependent protein kinase II (CaMKII). Disinhibition of Ltype calcium channels by blockage of phosphatase calcineurin with cyclospo rine A (CsA) causes an increase in the quantal content of endplate potentials (EPP’s QC), which can be pre vented by the Ltype calcium channel blocker nitrendipine, as well as by preliminary inhibition of CaMKII with KN62. Thus, the acetylcholine (ACh) release is enhanced with involvement of activated CaMKII only after disinhibition of Ltype calcium channels. When disinhibition of Ltype calcium channels was induced by suppression of calciumactivated BKtype potassium channels with paxilline, an increase in EPP’s QC was also observed. The conclusion was drawn that in motor synapses of mice, calcium influx via P/Qtype calcium channels cannot provide selective activation of CaMKII to facilitate transmission. However, disin hibition of Ltype calcium channels results in the activation of CaMKII, which is accompanied by a persis tent increase in evoked ACh release. Keywords: calcium/calmodulindependent protein kinase II, calcineurin, evoked mediator release, quantum content, Ltype calcium channels DOI: 10.1134/S1819712415020099 1
INTRODUCTION Calcium/calmodulindependent protein kinase II (CaMKII) is a serine–threonine protein kinase that is composed of 12 subunits that form two tightly contact ing hexameric rings. Activation of this kinase occurs with the involvement of calmodulin and calcium with following crossautophosphorylation of its subunits at T286. After transition into a phosphorylated state, the activity of CaMKII becomes independent of calcium and calmodulin and is aimed at different intracellular targets [1]. The role of CaMKII in the regulation of pre and postsynaptic plasticity of synapses is well known. Depending on synapse type and the localiza tion of CaMKII in the pre and postsynaptic areas, calcium ions can activate the enzyme after entering the cell via both NMDA channels and voltagegated calcium channels or via ryanodyne receptors of cal cium stores [2–4]. In peripheral neuromuscular junc tions, the role of presynaptic CaMKII remains unstudied. The P/Qtype calcium channels are the main source of calcium influx into the terminals of
1 Corresponding
author; address: Leninskie gory 1, str. 12, Moscow, 119234, Russia; phone: +7(495)9392792; email:
[email protected].
motor synapses; their activity supports calcium dependent release of mediator during single or rhyth mical activity of synapses [5]. On motornerve endings Ltype calcium channels also occur [6], which are inhibited in normal conditions, but, when disinhib ited, significantly increase ACh secretion [7, 8]. The question of which type of calcium channels and cal cium signals mediated by them induce the selective activation of CaMKII and the following involvement of the enzyme in the regulation of ACh release remains unstudied. Hence, the purpose of this work was to study evoked ACh release under conditions of the functioning of voltagegated P/Qtype calcium chan nels or Ltype calcium channels after their disinhibi tion by different means. The task that we set was to detect possible interactions between the functioning of calcium channels, selective activation of CaMKII, and its participation in the regulation of mediator release. MATERIALS AND METHODS The study of evoked and spontaneous acetylcho line (ACh) secretion in mature motor synapses of male
101
102
TARASOVA et al.
mice (P30) (line 129/Sv) was performed with isolated “dissected” neuromuscular preparations of diaphragm muscle (m. diaphragma–n. phrenicus) at 20–25°C [7]. Mice were killed by rapid decapitation. Animals were kept according to the 86/609/EEC directive on the use of laboratory animals; the protocol was approved by the bioethics commission of the Department of Biol ogy of Moscow State University. The preparation was placed into a 3 mL chamber and perfused with oxy genated (95% O2, 5% CO2) Liley solution for homoio therms (pH 7.2–7.4) that contained (mM): NaCl, 135; KCl, 4; NaH2PO4, 0.9; CaCl2, 2, MgCl2, 1; NaHCO3, 16.3; and glucose, 11. Simultaneous recording of miniature endplate potentials (MEPPs) and evoked endplate potentials (EPPs) was conducted intracellularly with glass microelectrodes filled with 2.5 M KCl (at a tip resistance of 10–15 MΩ). In the study of rhythmic activity, the phrenic nerve was stim ulated by short trains of suprathreshold impulses (50 stimuli at a frequency of 50 Hz). Signals were recorded by an Axoclamp2B amplifier (Molecular Devices, United States) and recorded on a PC hard drive using an LCard E154 analogtodigital converter with a PowerGraph interface. The data were analyzed using MiniAnalysis software (Synaptosoft, United States). In each series of experiments, we used at least three neuromuscular preparations. In control, we recorded MEPPs and EPPs simultaneously in five or more dif ferent synapses. Next, we added the studied substances in a certain order into perfusion saline and recorded the synaptic activity from different synapses for 60– 90 min. Apart from evaluating mean values of the MEPP and EPP amplitudes, we estimated the quantal content of EPPs (EPP’s QC), which was calculated as the relationship of the mean corrected for nonlinear summation of EPP amplitude to the mean MEPP amplitude [9]. The reliability of the differences between samples was estimated by the Student’s ttest and Mann–Whitney test. The level of significance of the differences between two selections was 0.05 (n for the number of synapses studied). In this work, following reagents were used: calcineurin inhibitor, cyclosporine A (CsA); calcium/calmodulindependent protein kinase II inhibitor, KN62 (4[(2S)2[(5iso quinolinylsulfonyl)methylamino]3oxo3(4phenyl 1piperazinyl)propyl] phenyl isoquinolinesulfonic acid ester); a blocker of Ltype calcium channels, nitrendipine (1,4dihydro2,6dimethyl4(3nitro phenyl)3,5pyridine dicarboxylic acid ethyl methyl ester); and the inhibitor of calciumactivated BKtype sodium channels, paxilline. All reagents were from Enzo Life Sciences, United States. RESULTS It is known that under normal conditions, the exo cytosis of neurotransmitter during rhythmic activity of neuromuscular junctions is mainly provided by cal cium influx through P/Q type channels, which
induces a train of EPPs with a typical shape [5]. In these experiments, the application of the calcium/calm odulindependent protein kinase II blocker KN62 (3 μM) caused no significant changes in the mem brane potential of muscle fibers, MEPP amplitudes (0.98 ± 0.05 mV (n = 16) in the control vs 1.04 ± 0.05 mV (n = 16, p > 0.05) in the presence of KN62), and the quantal content of each EPP during a train compared to the control (for the first EPP in the train: 35.32 ± 1.44 (n = 16) in the control and 36.19 ± 1.62 (n = 16) during KN62 treatment). This means that calcium, which enters through voltagegated P/Qtype calcium channels and triggers mediator release, does not induce CaMKII activation and its following involvement into regulation of evoked neu rotransmitter release. In the next series of experiments, we used disinhi bition of Ltype calcium channels located on the ter minals in an inactive state. It was previously shown in our work that blockage of the basal activity of cal cium–calmodulindependent phosphatase cal cineurin with cyclosporine A results in the disinhibi tion of Ltype calcium channel activity and the ampli fication of the evoked mediator release [10]. Therefore, in this work, we applied CsA in order to determine whether the observed facilitation of ACh secretion may be related not only to the appearance of an extra calcium source in the terminal but also to CaMKII activation. CsA (1 μM) significantly increased the quantal content of the first EPP from 30.62 ± 2.53 (n = 11) in the control to 39.24 ± 2.30 (n = 15, p < 0.05); a similar increase was present in the entire train (Fig. 1). During further action of the blocker of Ltype calcium chan nels, nitrendipine (1 μM), the quantal content of every EPP in train returned to the control level, and the first EPP returned to 30.48 ± 2.29 (n = 14, p > 0.05 as com pared with control). During the entire experiment, the membrane potential of muscle fibers (43.55 ± 1.44 mV (n = 11) in the control, 44.67 ± 1.71 mV (n = 15) in the presence of CsA, and 44.36 ± 1.68 mV (n = 14, p > 0.05) in the presence of CsA and nitrendipine) and the MEPP amplitude did not change (Fig. 1). Thus, in case of calcineurin blockage, the significant facilita tion of evoked ACh release occurs due to the addition of calcium influx via Ltype channels to the influx via the functioning pool of P/Q type calcium channels. To detect whether CaMKII is activated during dis inhibition of Ltype calcium channels due to cal cineurin inhibition and if it participates in the observed facilitation of evoked ACh release under these conditions, we studied the rhytmic activity of motor synapses in the presence of CsA in combination with the following or preceding application of CaMKII blocker KN62. It turned out that after preliminary action of KN62 (3 μM), which itself did not influence evoked secretion in neuromuscular junctions, CsA (1 μM) lost the ability to facilitate the neurotransmitter release NEUROCHEMICAL JOURNAL
Vol. 9
No. 2
2015
METHODS OF ACTIVATION AND THE ROLE OF … PROTEIN KINASE II mV 1.5
50 45
1.0
40 Quantal content of EPP
103
0.5 0
35 30 25
*
20 15 1
15
8
29 22 EPP number in a train
36
43
50
Fig. 1. Changes in the quantal content of EPPs during a short rhythmic train (50 Hz): in the control (䊉); 1 µM CsA (䊉); nitren dipine (1 µM) in the presence of CsA (䉫). In inset, the MEPP amplitudes are shown. *, p < 0.05 compared with the control.
(Fig. 2a). The quantal content of the first EPP in a short rhythmic train was 35.32 ± 1.44 (n = 16) in the control, 37.64 ± 2.63 (n = 16) in the presence of KN62 (3 μM), and 35.99 ± 2.27 (n = 18, p > 0.05) in the presence of CsA (1 μM) and KN62 (3 μM). Thus, we found that after inhibition of calcineurin activity, potentiation of ACh release mediated by disinhibition of Ltype calcium channels strictly requires CaMKII activation. However, KN62 application during action of CsA and an already developed increase in ACh secretion caused no effect on the evoked activity of synapses (Fig. 2b). Thus, in the control, the quantal content of the first EPP in the train was 24.49 ± 1.58 (n = 14); it increased to 38.26 ± 3.43 in the presence of CsA (n = 12); and remained increased following KN62 application in the presence of CsA, 39.64 ± 3.85 (n = 22, p < 0.05). Since the blockage of calcineurin can induce not only disinhibition of Ltype calcium channels but also CaMKII activation due to removal of the direct inhib itory action of calcineurin on this kinase, the situation remained unclear: which factor exactly triggers CaMKII activity: calcium influx by Ltype channels or the inhibition of phosphatase activity of calcineurin itself? To examine this, in the next series of experi ments, we used another known method of the disinhi bition of Ltype calcium channels, inhibition of cal ciumactivated BKtype potassium channels by pax illine while leaving the activity of calcineurin intact [11,12]. NEUROCHEMICAL JOURNAL
Vol. 9
No. 2
2015
In the presence of paxilline (5 μM), the quantal content of EPPs increased throughout the train (Fig. 3a). The quantal content of the first EPP in train was 25.52 ± 1.45 in the control (n = 17) and 33.74 ± 1.81 with paxilline (n = 19, p < 0.05). The preliminary blockage of CaMKII with KN62 (3 μM) completely prevented the increase in the quan tal content of EPPs that was induced by paxilline (5 μM) and related to disinhibition of Ltype calcium channels (Fig. 3b). The shape of the train does not dif fer from the control, either after KN62 alone or after combined application of KN62 and paxilline. In this case, the quantal content of the first EPP in the train was 35.09 ± 1.25 (n = 23) in the control; 33.33 ± 0.93 (n = 23) in the presence of KN62, and 34.29 ± 1.53 (n = 32, p > 0.05) during paxilline action in the pres ence of KN62. The results of this series suggest that it is the calcium influx via Ltype channels that is required for the activation of CaMKII but not the blockage of calcineurin activity. DISCUSSION In our studies, the specific inhibitor of CaMKII, KN62, did not influence evoked mediator secretion during a train that is triggered by influx of calcium ions into the terminal via P/Qtype channels. This calcium influx determines not only single ACh release but also typical changes in secretion during the train of EPPs [13]. KN62 did not cause any changes in the shape of the train of EPPs. This allows us to conclude that this influx of calcium via P/Qchannels and its accumula
104
TARASOVA et al.
Quantal content of EPP
50
(a)
45
mV 1.5
40
1.0 0.5
35
0 30 25 20 15 1
8
15
Quantal content of EPP
50
29 22 EPP number in a train
(b)
36
43
50
mV 1.5
45
1.0
40
0.5
35
0
30 25
*
20 15 1
8
15
22 29 EPP number in a train
36
43
50
Fig. 2. The effects of CaMKII blockage by KN62 on neuromuscular transmission under conditions of short rhythmic (50 Hz) activity of synapses: (a) Changes in the quantal content of EPPs during a short rhythmic train: in the control (䊉); 3 µM KN62 (䊊); 1 µM CsA in the presence of KN62 (䊏). (b) Changes in the quantal content of EPPs during a train: in the control (䊉); 1 µM CsA (䊉); 3 µM KN62 in the presence of CsA (䊐). The insets show the MEPP amplitudes. *, p < 0.05 as compared with control.
tion as socalled “residual calcium” during shortterm rhythmic activity of synapses are not sufficient for activation of CaMKII and its involvement in the regu lation of evoked ACh release in the motor synapses of mice. This feature distinguishes murine motor syn apses from literature examples of close interaction of CaMKII and P/Qchannels in hippocampal neurons [3], as well as from motor synapses in frogs, where the functioning of the major pool of voltagegated Ntype calcium channels is influenced by CaMKII, which plays a role in the regulation of shortterm plasticity and potentiation of mediator release during highfre quency stimulation [14]. In glutamatergic neuromus cular junctions of drosophila, it was also shown that
during tetanic stimulation of nerve endings CaMKII inhibitors suppress an increase in the motility of vesi cles [15]. These differences in the sources of CaMKII activation may be related to characteristics of the experimental objects, the mode of synaptic activity, the specifics of calcium inputs, and the character of their colocalization with enzyme. In this work, it was shown for the first time that CaMKII activation in murine motor synapses and manifestation of its potentiating action on mediator release requires the disinhibition of the additional cal cium input in terminals, Ltype channels. We used two different methods for the disinhibition of latent Ltype calcium channels: first, inhibition of NEUROCHEMICAL JOURNAL
Vol. 9
No. 2
2015
METHODS OF ACTIVATION AND THE ROLE OF … PROTEIN KINASE II
105
mV
45
(a) mV 1.5
Quantal content of EPP
40
1.0
35
0.5 0
30 25
*
20 15 1
8
15
22 29 EPP number in a train
36
43
50
36
43
50
Quantal content of EPP
45 (b)
40
mV 1.5 1.0
35
0.5 30 0 25 20 15 1
8
15
22 29 EPP number in a train
Fig. 3. Changes in the quantum content of an EPP during a rhythmic burst: (a) in the control (䊉) and in the presence of 5 µM paxilline (䊊); (b) in the control (䊉), in the presence of 3 µM KN62 (䉭), in the presence of 3 µM KN62 and 5 µM paxilline (䊏). The insets show the MEPP amplitudes. *, p < 0.05 as compared with the control.
calcineurin, calcium and calmodulindependent phosphatase; and, second, removal of hyperpolariza tion caused by calciumactivated BKtype potassium channels. Earlier we proved that in murine motor ter minals, calcineurin possesses tonic inhibitory activity, which suppresses Ltype calcium channels [10]. How ever, it is known that this phosphatase is able to indi rectly influence the dephosphorylation of activated CaMKII through another phosphatase, PP1 [16], and thereby, indirectly interact with CaMKII. Therefore, we used another method for the disinhibition of Ltype calcium channels that does not affect cal cineurin activity. It is known that there is a reciprocal interaction between the activity of calciumdependent BKtype potassium channels and the condition of Ltype calcium channels. Accordingly, the blockage of calciumdependent potassium channels by selective NEUROCHEMICAL JOURNAL
Vol. 9
No. 2
2015
inhibitors (paxilline or iberiotoxin) results in disinhi bition of activity of Ltype calcium channels with intact calcineurin activity [11, 12]. We used paxilline for disinhibition of Ltype calcium channels and com bined its action with preliminary application of the CaMKII inhibitor. After both inhibition of BKchannels with paxilline [11] and blockage of calcineurin, the increment of the quantal content of EPPs, which was uniform inside the entire train, was observed. This increment was pre vented by selective blockers of Ltype calcium chan nels. As well, regardless of the method of the disinhi bition of Ltype calcium channels, the preliminary blockage of CaMKII with KN62 prevented the incre ment of ACh release, which is related to the activity of Ltype calcium channels. However, KN62 applica tion after CsA was inefficient and did not influence the
106
TARASOVA et al.
CsAinduced increment of ACh release. These differ ences in the effects of KN62 before and after CsA correspond well to the data on the nature of the activa tion and action of CaMKII. Indeed, it is well known that an increase in calcium concentration is necessary only at the first stage of CaMKII activation for its autophosphorylation, which results in sustained calciumindependent enzyme activity [1]. It is also known that KN62 inter acts with calciumcalmodulinbinding CaMKII site and thereby prevents only the initial stage of its activa tion [17]. This is the reason why KN62 application in the presence of CsA, i.e., when CaMKII is already converted into the active phosphorylated state due to an increase in the calcium level in the terminal, was inefficient because the inhibitor is unable to block CaMKII action in this case. These data suggest that it is calcium influx via Ltype calcium channels (regard less of the method of their disinhibition) that is the cause of CaMKII activation and the following increase in mediator secretion. Until now, it seemed that in motor terminals the involvement of Lcalcium channels creates an addi tional calcium signal in the terminal, which sums with calcium entering via P/Qtype channels. As a result, the intracellular calcium level increases, which leads to enhancement of calciumdependent exocytosis and release of ACh quanta. However, this work revealed for the first time the necessity of CaMKII activation in disinhibition of Ltype channels as well. Potentiation of mediator release occurs only in this case. It is note worthy that CaMKII is not the only calciumdepen dent enzyme that is activated under these conditions. Earlier, it was shown that protein kinase C is also acti vated. The possibility of activation of several calcium dependent enzymes at once, including PKC and CaMKII, during calcium influx via Ltype channels was described as well in other excitable cells with Ltype calcium channels (in cardiomyocytes and smoothmuscle cells). These enzymes often “service” Ltype channels and form a positive feedback [18, 19]. For example, along with exocytosis proteins (Munc13), Ltype calcium channels are considered as possible targets of presynaptic PKC, which, when acti vated, facilitate their work [12, 20]. For CaMKII, there are examples of submembrane CaMKII localization in close proximity to voltage gated calcium channels [21, 22]. In these cases, the calcium current via calcium channels turns out to be an immediate cause of the activation of the enzyme [3, 23, 24]. At the same time, in nerve terminals, the release of stored calcium may be a source of CaMKII activation [4, 15, 25, 26]. In turn, Ltype calcium channels quite often trigger release in neurons [27]. Earlier, we showed that in motor synapses with disin hibited Ltype calcium channels, RyRs becomes acti vated and stored calcium is released; this release is necessary for the potentiation of ACh secretion [11]. A decrease in the activity of at least one of the kinases
(PKC or CaMKII) probably leads to a partial decrease in calcium influx via Ltype channels, partially atten uates the Ca signal, and makes it insufficient for RyR activation, which disrupts the functioning of the fol lowing cascade: Ltype calcium channels → calcium influx → calciumdependent activation of ryanodine sensitive calcium stores → release of stored calcium → potentiation of mediator secretion. It is possible that CaMKII activation results not only from calcium current entering via Ltype channels but also from RyRmediated release of stored calcium, which is coupled to the previous calcium entry. It is known, for example, that in motornerve terminals of drosophila the release of stored calcium results in CaMKII activation, which is followed by translocation of the enzyme to the locations of target proteins [4]. Currently, the bestknown targets of CaMKII in nerve terminals of different types of synapses are the following: protein synapsin 1 [28], BKtype potassium channels [29], as well as voltagegated Ntype, P/Qtype, and Ltype calcium channels [3, 30, 31]. From this list, in our view, Ltype channels themselves are the most probable target of CaMKII action in the situation of the disinhibition of Ltype calcium chan nels in motor terminals. At present, it is universally recognized that regard less of the neuron or terminal type, Ltype calcium channels are the target of various phosphatase–kinase and voltagedependent regulations [27, 32–34], although, in motor terminals, inhibitory phosphatase actions dominate resulting in the inhibition of Ltype calcium channels [10, 20]. However, in cases where Ltype calcium channels are disinhibited, CaMKII activity is generally present in the cell; quite often, it is a form of positive feedback for the maintenance of the functioning of a channel itself [21, 30, 35]. This sce nario is well known for cardiomyocytes, where cal cium channels that possess several described sites for phosphorylation by calmodulin kinase are a target of CaMKII activated by calcium influx via Ltype cal cium channels [30, 32]. It is possible that the studied motor synapses also have a similar positive feedback between partially disinhibited Ltype calcium chan nels and CaMKII activity, which shifts the balance of regulatory actions on Ltype channels towards ampli fication and maintenance of their activity and thus potentiates calcium influx and the following calcium dependent ACh release. Our considerations, however, are hypothetical: the final solution of the problem of the mechanisms of the facilitation of mediator release that are related to the activation of Ltype calcium channels and CaMKII in motor terminals requires further study. CONCLUSIONS In this work, it was found for the first time that in motor terminals calcium influx via P/Q type channels after the shortterm activity of synapses is incapable of NEUROCHEMICAL JOURNAL
Vol. 9
No. 2
2015
METHODS OF ACTIVATION AND THE ROLE OF … PROTEIN KINASE II
activation of CaMKII to potentiate transmission. However, calcium influx via Ltype voltagegated channels, regardless of the methods of the disinhibi tion of these channels (either by suppression of phos phatase calcineurin or by blockage of calciumacti vated BKtype potassium channels), results in selec tive CaMKII activation, which is necessary for the facilitation of evoked ACh release. ACKNOWLEDGMENTS This study was supported by the Russian Founda tion for Basic Research, project no. 130400413a. REFERENCES 1. Merrill, M., Chen, Y., Strack, S., and Hell, J., Trends Pharmacol. Sci., 2005, vol. 26, pp. 645–653. 2. Wheeler, D., Barrett, C., Groth, R., Safa, P., and Tsien, R., J. Cell Biol., 2008, vol. 183, pp. 849–863. 3. Jiang, X., Lautermilch, N., Watari, H., Westenbroek, R., Scheuer, T., and Catterall, W., Proc. Natl. Acad. Sci. USA, 2008, vol. 105, pp. 341–346. 4. Shakiryanova, D., Morimoto, T., Zhou, C., Chouhan, A., Sigrist, S., Nose, A., Macleod, G., Deitcher, D., and Levitan, E., J. Neurosci., 2011, vol. 31, pp. 9093–9100. 5. Nudler, S., Piriz, J., Urbano, F., RosatoSiri, M., Renteria, E., and Uchitel, O., Ann. NY Acad. Sci., 2003, vol. 998, pp. 11–17. 6. Pagani, R., Song, M., McEnery, M., Qin, N., Tsien, R., Toro, L., Stefani, E., and Uchitel, O., Neuroscience, 2004, vol. 123, pp. 75–85. 7. Flink, M. and Atchison, W., J. Pharmacol. Exp. Ther., 2003, vol. 305, pp. 646–652. 8. Balezina, O., Bogacheva, P., and Orlova, T., Bull. Exp. Biol. Med, 2007, vol. 143, pp. 171–174. 9. McLachlan, E. and Martin, A., J. Physiol., 1981, vol. 311, pp. 307–324. 10. Gaydukov, A., Tarasova, E., and Balezina, O., Neuro chemical Journal, 2013, vol. 7, pp. 29–33. 11. Gaydukov, A., Melnikova, S., and Balezina, O., Bull. Exp. Biol. Med, 2009, vol. 148, pp. 163–166. 12. Gaydukov, A., Marchenkova, A., and Balezina, O., Bull. Exp. Biol. Med, 2012, vol. 153, pp. 415–418. 13. Dittman, J., Kreitzer, A., and Regehr, W., J. Neurosci., 2000, vol. 20, pp. 1374–1385. 14. Mukhamedyarov, M., Kochunova, J., Yusupova, E., Haidarov, B., Zefirov, A., and Palotas, A., Brain Res. Bull., 2010, vol. 81, pp. 613–616.
NEUROCHEMICAL JOURNAL
Vol. 9
No. 2
2015
107
15. Shakiryanova, D., Klose, M., Zhou, Y., Gu, T., Deitcher, D., and Atwood, H., Hewes, R., Levitan, E., J. Neurosci., 2007, vol. 27, pp. 7799–7806. 16. Winder, D. and Sweatt, J., Nat. Rev. Neurosci., 2001, vol. 2, pp. 461–474. 17. Pellicena, P. and Schulman, H., Front. Pharmacol., 2014, vol. 5, p. 21. 18. Pitt, G., Cardiovasc. Res., 2007, vol. 73, pp. 641–647. 19. Dai, S., Hall, D., and Hell, J., Physiol. Rev., 2009, vol. 89, pp. 411–452. 20. Urbano, F., Depetris, R., and Uchitel, O., Pflugers Arch., 2001, vol. 441, pp. 824–831. 21. Tsui, J., Inagaki, M., and Schulman, H., J. Biol. Chem., 2005, vol. 280, pp. 9210–9216. 22. Tsui, J. and Malenka, R., J. Biol. Chem., 2006, vol. 281, pp. 13794–13804. 23. Grueter, C., Abiria, S., Dzhura, I., Wu, Y., Ham, A., Mohler, P., Anderson, M., and Colbran, R., Mol. Cell, 2006, vol. 23, pp. 641–650. 24. Abiria, S. and Colbran, R., J. Neurochem., 2010, vol. 112, pp. 150–161. 25. Kubota, M., Narita, K., Murayama, T., Suzuki, S., Soga, S., Usukura, J., Ogawa, Y., and Kuba, K., Cell Calcium, 2005, vol. 38, pp. 557–567. 26. Kolarow, R., Brigadski, T., and Lessmann, V., J. Neurosci., 2007, vol. 27, pp. 10350–10364. 27. CalinJageman, I. and Lee, A., J. Neurochem., 2008, vol. 105, pp. 573–583. 28. TaoCheng, J., Dosemeci, A., Winters, C., and Reese, T., Brain Cell Biol., 2006, vol. 35, pp. 117–124. 29. Wang, J. and Kelly, P., Learn Mem., 1996, vol. 3, pp. 170–181. 30. Hudmon, A., Schulman, H., Kim, J., Maltez, J., Tsien, R., and Pitt, G., J. Cell Biol., 2005, vol. 171, pp. 537–547. 31. Tang, Q., Bangaru, M., Kostic, S., Pan, B., Wu, H., Koopmeiners, A., Yu, H., Fischer, G., McCallum, J., Kwok, W., Hudmon, A., and Hogan, Q., J. Neurosci., 2012, vol. 32, pp. 11737–11749. 32. Lee, T., Ono, K., Miyamoto, S., Hadama, T., and Arita, M., J. UOEH, 2006, vol. 28, pp. 277–286. 33. Oliveria, S., Dittmer, P., Youn, D., Dell’Acqua, M., and Sather, W., J. Neurosci., 2012, vol. 32, pp. 15328– 15337. 34. Subramanian, J., Dye, L., and Morozov, A., J. Neurosci., 2013, vol. 33, pp. 7245–7252. 35. Erxleben, C., Liao, Y., Gentile, S., Chin, D., Gomez Alegria, C., Mori, Y., Birnbaumer, L., and Armstrong, D., Proc. Natl. Acad. Sci. USA, 2006, vol. 103, pp. 3932– 3937.
