Biological Sciences - Neuroscience
Loss of AP3 function affects spontaneous and evoked relea at hippocampal mossy fiber synaps
Anita Scheuber*¶, Rachel Rudge¶, Lydia Danglot¶, Graca Raposo, Thomas Binz§, Jean-Christophe Poncer* and Thierry Galli.
* INSERM, Université Pierre et Marie Curie-Paris 6, UMR739, “Cortex & Epilepsy”, Paris, F-75013, France.
“Membrane Traffic in Neuronal and Epithelial Morphogenesis”, INSERM Avenir Team, Paris, F-75005 France; Institut Jacques Monod, CNRS UMR7592, Universitities Paris 6 & Paris 7, Paris, F-75005 France.
“Structure and Membrane Compartments”, CNRS UMR 144, Institut Curie, Paris, F-75005, France.
§ Institute of Biochemistry, Medical School Hannover, D-30625 Hannover, Germany.
¶The authors contributed equally to this work and are listed in rever alphabetical order.肉丝番号>家常糖醋鱼的做法
Corresponding author: Thierry Galli, Institut Jacques Monod, 2, place Jussieu, F-75251 Paris Cedex 05, France, mailto:thierry@tgalli, tel: +33 144 278 211, fax: +33 144 278 210, thierry.galli.free.fr/
Abbreviations footnote: Impact of loss of AP3 in Neurotransmitter Relea Abstract
Synaptic vesicle exocytosis mediating neurotransmitter relea occurs spontaneously at low intraterminal calcium concentrations and is stimulated by a ri in intracellular calcium. Exocytosis is compensated for by the reformation of vesicles at plasma membrane and endosomes. Although the adaptor complex AP3 was propod to be involved in the formation of synaptic vesicles from endosomes, whether its function has an indirect effect on exocytosis remains unknown. Using mocha mice, which are deficient in functional AP3, we identify an AP3-dependent, tetanus neurotoxin-resistant, asynchronous relea that can be evoked at hippocampal mossy fiber synaps. Presynaptic targeting of the tetanus neurotoxin-resistant vesicle SNARE TI-VAMP is lost in mocha hippocampal mossy fiber terminals while the localization of synaptobrevin 2 is unaffected. In addition, quantal relea in mocha cultures is more frequent, and more nsitive to sucro. We conclude that the lack of AP3 results in more constitutive cretion and the loss of an asynchronous evoked relea component, suggesting an important function of AP3 in regulating SV exocytosis at
mossy fiber terminals.
Introduction
The relea of neurotransmitter at the synap requires the fusion and recycling of synaptic vesicles (SVs). The fusion of SVs with the plasma membrane depends on the formation of SNARE complexes between vesicle SNAREs and plasma membrane SNAREs, as demonstrated by the striking nsitivity of neurotransmitter relea to clostiridial neurotoxins particularly Tetanus Neurotoxin (TeNT, for a review e (1)). Several models of the recycling of synaptic vesicles have been propod: endosomal recycling, SV budding from the plasma membrane (2), kiss-and-run and kiss-and-stay (for a review e (3)). Endosomal recycling involves the molecular coat AP3 as suggested from experiments in neuroendocrine cells (4) but the importance of AP3 in neurotransmitter relea is still unclear. AP3 is compod of four subunits and two different AP3 complexes are expresd in brain: the ubiquitous AP3A, compod of the į,ı3,ȕ3A and ȝ3A subunits, and the neuronal specific AP3B, compod of the į,ı3,ȕ3B and ȝ3B subunits (5, 6). Mocha mice are deficient for the į subunit and therefore lack both AP3A and AP3B complexes. The mice have neurological disorders including hyperactivity and spontaneous izures. In this paper, we t out to understand the importance of AP3 function in neurotransmitter relea by characterizing basal and e
voked relea neurotransmitter relea in mocha mice.
Results
Asynchronous relea evoked at mossy fiber terminals is lost in mocha mice动态图片gif
AP3 is particularly concentrated in the hilus and CA3 region of the hippocampus in heterozygous control (+/-) mice (Fig. S1a). Therefore, in order to asss the role of AP3 in neurotransmitter relea, we compared synaptic transmission at mossy fiber (MF)-CA3 synaps in organotypic hippocampal cultures from mocha (-/-) and heterozygous control (+/-) littermates (Fig. 1a). We first examined Ca2+-dependent relea evoked by MF stimulation. A stimulating electrode placed at the hilar border of the granule cell layer reliably evoked large post-synaptic currents (PSCs) that were specifically suppresd by the group 2 metabotropic glutamate receptor agonist DCG-IV (Fig. 1b;
(7)). The amplitude of the PSCs gradually incread with stimulation intensities ranging from 15 to 600 V.μs. The correlation between the average charge of the PSCs and the stimulation intensity was not significantly different in cultures prepared from control vs. mocha mice (Fig. 1c). Cleavage of the SV SNARE synaptobrevin 2 (Syb2)by preincubation of the cultures with tetanus neurotoxin (TeNT) for >72h caud a dramatic reduction in transmitter relea evoked by MF stimulation. In cultures pr
epared from control mice, stimuli up to 2000 V.μs evoked small, unreliable PSCs, reminiscent of evoked transmission in cultured hippocampal neurons from Syb2 knockout mice (8). Relea at stimulated synaps was usually asynchronous and occurred within ~200 ms after stimulation with an average probability of 0.26±0.05 at the highest stimulation intensities (n=7). In contrast, no PSC could be evoked by MF stimulation in cultures prepared from mocha littermate mice.
Loss of presynaptic TI-VAMP in mossy fiber terminals
Our obrvations may reflect the contribution of a TeNT-innsitive, AP3-dependent pathway of transmitter relea at MF terminals. We have previously shown that TI-VAMP (i) is prent at a high level in SVs at MF terminals (9), (ii) interacts with the įsubunit of the AP3 complex and (iii) is mistargeted in mocha fibroblasts (10). TI-VAMP is therefore the best candidate v-SNARE to support TeNT-resistant vesicle exocytosis at MF terminals. Since TI-VAMP is not expresd at Schaffer collateral
琴伤
terminals onto rat CA1 pyramidal cells ((9), Fig. S1a), we anticipated this form of exocytosis may not be obrved at the synaps. Consistent with this prediction, evoked relea was entirely impaired by TeNT at Schaffer collateral synaps onto CA1 pyramidal cells in control mice (Fig. 1a,b)
. We therefore compared TI-VAMP localization in hippocampal ctions from control and mocha mice. In the CA3 and dentate gyrus areas of control mice, TI-VAMP was localized in the MF terminals, as demonstrated by its colocalization with the presynaptic marker synaptophysin (Syp, Fig. 2a, S1a). This was further confirmed in cultured granule cells becau we found that 67.20+/-6.65% of Syp positive punctae were also TI-VAMP positive (Fig. S2a). However in mocha ctions, TI-VAMP labeling in MFs was completely lost (Fig. 2a,
S1b). Thus, AP3 is required for the presynaptic targeting of TI-VAMP to MF terminals. In contrast, Syb2 localization was unchanged in brain ctions from mocha mice as compared to control (Fig. 2b) suggesting that the presynaptic targeting of Syb2 is independent of AP3. Other SV proteins, including Rab3a, synaptotagmin 1, showed a presynaptic targeting similar in control and mocha mice (Fig. S1b, and our unpublished obrvations). Together, the results suggest that a form of TeNT-innsitive, AP3-dependent, evoked relea exists at MF-CA3 synaps, which likely involves TI-VAMP as a v-SNARE.
TI-VAMP is blocked in the cell body of mocha neurons
风衣怎么搭配好看We then examined the subcellular localization of TI-VAMP in mocha hippocampal granule cells from
which mossy fibers originate. We found that TI-VAMP accumulated in the cell bodies of granule cells as shown by colocalization with VAMP4, a vesicular SNARE located in early endosomes (11, 12) and the TGN (13) (Fig. 2d). However, the AP-3 dependent sorting of TI-VAMP in the perinuclear VAMP4-positive compartment may not be specific to granule cells since a strong colocalization of TI-VAMP and VAMP4 was also found in the mocha CA3 pyramidal cells (Fig. 2c). The accumulation of TI-VAMP in cell bodies in mocha neurons was also obrved in cultured hippocampal pyramidal neurons by immunolabeling (Fig. S2b) as well as immunogold labeling in ultrathin cryoctions analyzed by electron microscopy (Fig. S2c). TI-VAMP labeling in mocha neurons was found restricted to the cytosolic side of Golgi cisternea whereas in control pyramidal neurons TI-VAMP labeled vesicles and the plasma membrane, consistent with its function as a cretory v-SNARE in control but not mocha neurons. Therefore, an AP-3 dependent sorting of TI-VAMP at the level of a VAMP4-positive perinuclear compartment is required for the proper targeting of TI-VAMP.
Incread basal relea in mocha and BFA-treated mossy fiber terminals
We then asked whether the lack of AP3 and TI-VAMP may impact Ca2+-independent, constitutive relea at MF terminals from mocha mice. Miniature EPSCs (mEPSCs) were recorded from CA3 pyramidal cells in slice cultures prepared from control and mocha littermates. In cultures from mocha
mice, the frequency of mEPSCs was ~2-fold higher than in control cultures (3.22±0.55 vs. 1.66 ± 0.72 Hz, p<0.03, Fig. 3a). However, their mean amplitude was unchanged (19.44 ± 1.73 vs. 18.97 ± 3.32 pA,
p=0.91) as were their ri time (10-90% of peak, 1.59 ± 0.09 vs. 1.45 ± 0.19 ms,
p=0.32) and decay time constant (3.86 ± 0.12 vs. 3.37 ± 0.24 ms, p=0.14; Fig. 3b,c). The results suggest the lack of AP3 did not alter the rate of fusion pore opening or the number of postsynaptic receptors at excitatory synaps onto CA3 pyramidal cells. In addition, the higher mEPSC frequency in mocha cultures was not affected by the NMDA receptor antagonist APV (our unpublished obrvations), suggesting it did not reflect a greater activation of presynaptic NMDA receptors (14, 15) due to the lack of Zn relea from MF terminals in mocha mice (16, 17). After 72h incubation with TeNT, mEPSC frequency in control cultures was reduced by ~84 % (to 0.27 ± 0.04 Hz, n=9; Fig. 3d), with no apparent change in their mean amplitude (19.18 ± 0.99 vs. 18.97 ±
3.32 pA, p=0.89), again consistent with obrvations in cultured hippocampal neurons from Syb2 knockout mice (8). In contrast, quantal relea in mocha cultures was more resistant to TeNT and was reduced in frequency by only ~44 % (to 1.79 ± 0.56 Hz,
n=8). The obrvations suggest the abnce of AP3 not only increas Ca-independent quantal relea but also reduces the effect of TeNT. Apart from TI-VAMP and Syb2, no other v-SNARE protein is known to be prent at excitatory synaps on CA3 pyramidal cells. Therefore, we analyzed the penetration and cleavage efficiency of TeNT in the slice cultures. We labeled TeNT-treated slice cultures with DAPI and antibodies against TeNT and SNAP-25. We found that TeNT penetrated throughout
大三元牌型
mocha (Fig. 3e) as well as control explants (our unpublished obrvations). Furthermore, although the vast majority of Syb2 was cleaved in control and mocha explants after 72h incubation with TeNT, a small fraction of Syb2 was still detected by western blotting. Quantification of the blots revealed twice as much TeNT-resistant Syb2 protein in mocha cultures as compared to control (Fig. 3f). Treatment with TeNT drastically reduced Syb2 labeling, but the remaining signal corresponded largely to Syb2 prent at synaps, as revealed by colocalization with synaptophysin (Fig. 3g) thus the numerous TeNT-resistant mEPSCs obrved in mocha are likely to be mediated by a presynaptic pool of Syb2 that resisted the treatment with TeNT.
Newly formed vesicles derived from endosomes disappear with a half-life of ~36 min after treatment with the fungal drug Brefeldin A (BFA) which specifically targets the AP3-dependent pathway in PC1
2 cells (4, 18-20). In order to test the effect of acute inactivation of the AP3 pathway, we therefore incubated control hippocampal slices with BFA for ~2-4 hours. In BFA-treated slices, mEPSC frequency was incread by
~127 % as compared to control (2.59 ± 0.33 vs. 1.14 ± 0.25 Hz, n=14 and 12 cells, respectively, p<0.005, Fig. S4a-c). The ratio of synaptic TI-VAMP/synaptic Syp measured by immunolabeling was not significantly altered by BFA neither in the hilus nor in the st. lucidum (Fig. S4d). This is consistent with the fact that longer times may be required to clear TI-VAMP from MF terminals and demonstrate that the mocha phenotype can be reproduced by acute pharmacological inactivation of AP3-dependent SV formation.
Incread nsitivity of relea to osmotic stimulation in mocha MF terminals
Syb2 is resistant to TeNT (21) in SNARE complexes (21) that may be clamped by complexin and synaptotagmin before calcium ri (22). TeNT-resistant Syb2 was associated with relea-competent SVs (23, 24). Our previous results could suggest that the lack of AP3 and TI-VAMP in MF SVs may thus increa the capacity of Syb2 to form TeNT-resistant clamped SNARE complexes thereby enhancing the probability of calcium-independent fusion at mocha MF terminals. In order to test this
hypothesis, we examined the rate of relea induced by focal application of a hypertonic solution becau previous studies showed that hypertonic solution specifically recruits readily releasable quanta at hippocampal synaps (25) and stimulates cretion in a calcium independent manner. We thus compared the effects of focal applications of a 0.5 M sucro solution through a patch pipette positioned in st. lucidum ~25 μm away from the somata of the recorded pyramidal cells. Since effective sucro concentration at relea sites may be difficult to control in slice cultures, we ud varying injection pressures to compare the sucro nsitivity of relea in mocha vs. control cultures. In cells recorded from control cultures, application of sucro with low pressure (0.25 psi) caud a ~2-fold increa in mEPSC frequency, whereas at a higher pressure (1.5 psi), a further ~7-fold increa was obrved (Fig. 4a, b, c). The nsitivity of the relea rate to sucro was significantly incread in recordings from mocha cultures: even low pressure application of sucro caud a ~9-fold increa in mEPSC frequency, which was further enhanced by another ~35 % at high pressure (Fig. 4a, b, c). Interestingly however, recruitment of readily releasable vesicles by sucro was disrupted in both control and mocha cultures by prior incubation with TeNT, and application of hypertonic solution failed to produce an increa in mEPSC frequency even at high pressure (Fig. 4a, b). Taken together, the results suggest that an AP3-dependent mechanism decreas the sucro nsitivity of constitutive cretion.
9 Discussion
Mocha mice are deficient for AP3į subunit and therefore lack both ubiquitous AP3A and neuronal AP3B complexes. Here, we have shown that presynaptic TI-VAMP, a well-established AP-3 į cargo (10, 26), is lost in mocha CA3 MFs. Other AP-3 cargos including the zinc transporter ZnT-3 (16, 17), the chloride channel ClC-3 (27), and the phosphatidylinositol-4-kina type IIα (28) are mislocalized in mocha CA3 MFs. A previous study in ZnT3 knock out mice showed the lack of vesicular zinc in MFs does not significantly affect the MF-associated excitability of CA3 pyramidal cells (29). In addition, we obrved that the higher mEPSC frequency in mocha cultures was not affected by blocking NMDA receptors, further suggesting that the lack of Zn relea from MF terminals in mocha mice does not explain the phenotype we obrved. Similarly, the lack of MF ClC-3 is unlikely to explain the incread quantal relea from mocha MF terminals mice. The loss of CIC-3 rather affects acidification of SVs resulting in a slight reduction of quantal size (30). Finally, recent data suggested AP3 may regulate the volume of large den-core vesicles in chromaffin cells (31). This however is unlikely to apply to MF terminals since i) we did not obrve any change in quantal size in mocha cultures and ii) the pathway of SV reformation largely differs from that of cretory granules, the latter maturing via the removal of material from immature cretory granules. Treatmen
t of hippocampal slice cultures with TeNT revealed an asynchronous component of cretion evoked in CA3 pyramidal cells by single stimulation of mossy fibers. This asynchronous relea is unlikely to be mediated by the low amount of TeNT-resistant Syb2 since it was not obrved in mocha cultures which showed more TeNT-resistant Syb2. Furthermore, this asynchronous component was obrved at MF-CA3 synaps but not at CA1-CA3 synaps, where TI-VAMP is not expresd. Since no other TeNT-resistant v-SNARE was ever detected at MF terminals, we suggest TI-VAMP likely mediates the asynchronous evoked relea unraveled in our experiments. In conclusion, although we cannot exclude that other AP3 cargos lost in mocha MF terminals may participate to the mocha phenotype, the loss of TI-VAMP ems most likely to explain the perturbation of evoked SV relea described in the prent study.
10
Asynchronous relea was not obrved in TeNT-untreated explants, suggesting inactivation of Syb2 may be required for the expression of this asynchronous evoked relea. This obrvation strongly suggests TI-VAMP and Syb2 may both be prent on the same rather than distinct SVs. Consistent with this scenario, Syb2 was shown to have a higher rate of SNARE complex asmbly than TI-VAMP both in vitro and in vivo (10), predicting that evoked relea mediated by TI-VAMP wo
uld be detected only after cleavage of Syb2. In addition, cleavage of Syb2 by TeNT was shown to modify the coupling of intracellular calcium and relea-competent vesicles (32) suggesting that removal of TeNT-nsitive v-SNAREs allows for the expression of a cretory machinery that may be hard to obrve otherwi. Interestingly, Sr+ preferentially stimulates asynchronous relea (33) and different synaptotagmin isoforms show different nsitivities to Ca2+ and Sr+ (34). For instance, Synaptotagmin 1 and 7 have different nsitivities to calcium (35), the latter interacting with TI-VAMP in fibroblasts (36) and showing biochemical properties suitable for a Ca2+ nsor for asynchronous relea (35, 37, 38). Therefore, we speculate that MF SVs may be equipped with two distinct fusion machineries for exocytosis, one depending on Syb2 and the other on TI-VAMP as a v-SNARE, which may be recruited in different conditions.
We have shown a higher basal, calcium-independent relea in moch a cultures, with no apparent change in quantal size or current kinetics. This obrvation strongly argues for a presynaptic difference between control and mocha MF-CA3 synaps. This likely not due to a greater density of MF terminals onto mocha CA3 pyramidal cells becau of axonal sprouting induced in organotypic cultures (39) or a greater contribution of recurrent CA3-CA3 synaps to miniature EPSCs in mocha cultures. Indeed, we found an equal Syp staining in control and mocha brain ctions and the increa
d quantal relea was also obrved upon pharmacological disruption of AP3 by BFA in acute hippocampal slices. In addition, quantal relea evoked by hyperosmotic solution directly applied onto MF terminals was also incread in moch a cultures. Therefore, the most likely explanation is that more vesicles are ready-to-fu due to the abnce of AP3 and possibly TI-VAMP in mocha vs. control MF terminals. Since BFA specifically impairs the AP3-dependent formation of new vesicles from endosomes our数学手工制作
11
results suggest that AP3-dependent, newly-formed vesicles have a reduced capacity to fu. How AP3-dependent sorting may decrea basal relea remains to be explored.
In conclusion, our data suggest that the molecular mechanism of transmitter relea at MF terminals reaches a high degree of complexity with at least two exocytic SV-SNAREs: Syb2 and TI-VAMP. AP3 function is important for both constitutive as well as evoked relea, raising the possibility that specific forms of synaptic plasticity might occur at terminals expressing AP3 cargos like TI-VAMP (9).
12 Materials and methods
Animals
Heterozygous mocha mice were originally obtained from M. Robinson (CIMR, Cambridge) and then bred in-hou. The experiments were carried out in accordance with the European Communities Council Directive of 24 November 1986
(86/609/EEC). All efforts were made to minimize the number of animals ud and their suffering.
Immunofluorescence
Mocha and control (heterozygous littermates) 1-2 months old mice were anaesthetized with 35% chloral hydrate or pentobarbital and perfud through the heart with paraformaldehyde 4%. The discted brains were fixed overnight in 4% paraformaldehyde in phosphate buffered saline and cut on a vibratome to 30ȝm thick ctions and procesd for immunofluorescence as described previously (9). Confocal lar-scanning microscopy was performed using a SP2 confocal microscope (Leica Microsystems, Mannheim, FRG). Images were asmbled using Adobe Photoshop (Adobe Systems, San Jo, CA, USA).
Western Blotting
SDS-PAGE analysis was performed using 4-12% NuPAGE (Invitrogen, France) gradient gels and th股票短线炒股技巧
e manufacturer’s buffers and then procesd for Western blotting. Blots were quantified using ImageJ (NIH, Maryland) and statistical significance was estimated using Mann-Whitney rank sum tests performed under SigmaStat (SPSS Inc.)
Electrophysiological recordings
Organotypic hippocampal slices from 6 days old mice were maintained in culture as described previously (40, 41). After 2-3 days, cultures were preincubated 24 hours in rum-free medium and then grown another 3-4 days in fresh rum-free medium containing TeNT (500 ng/ml) or not. Electrophysiological recordings were carried out as described in supplementary methods. For mossy fiber stimulation, the stimulating electrode was positioned at the hilar border of the granule cell layer. For Schaffer
