Synaptic vesicle cycling is not impaired in a glutamatergic and a cholinergic synapse that exhibit deficits in acidification and filling

The purpose of the present work was to investigate synaptic vesicle trafficking when vesicles exhibit alterations in filling and acidification in two different synapses: a cholinergic frog neuromuscular junction and a glutamatergic ribbon-type nerve terminal in the retina. These synapses display remarkable structural and functional differences, and the mechanisms regulating synaptic vesicle cycling might also differ between them. The lipophilic styryl dye FM1-43 was used to monitor vesicle trafficking. Both preparations were exposed to pharmacological agents that collapse ΔpH (NH4Cl and methylamine) or the whole ΔμH+ (bafilomycin), a necessary situation to provide the driving force for neurotransmitter accumulation into synaptic vesicles. The results showed that FM1-43 loading and unloading in neuromuscular junctions did not differ statistically between control and experimental conditions (P > 0.05). Also, FM1-43 labeling in bipolar cell terminals proved highly similar under all conditions tested. Despite remarkable differences in both experimental models, the present findings show that acidification and filling are not required for normal vesicle trafficking in either synapse.


INTRODUCTION
Neurotransmitters are packed in synaptic vesicles (SV) and released by the fusion of a vesicle with the presynaptic membrane, a synaptic step called exocytosis.During endocytosis, empty vesicles must be recycled, acidified and re-internalized with transmitters in a rapid fashion to sustain another SV cycle (Sudhof, 2004).The driving force for neurotransmitter accumulation in SV is provided by the generation of a transmembrane electrochemical gradient (ΔμH+) that has a chemical gradient (ΔpH, inside acidic) and an electrical potential across the vesicular membrane (ΔΨ, inside positive).ΔμH+ is generated in situ by the electrogenic vacuolar H+-ATPase (VH + -ATPase), which is responsible for acidification and rendering the vesicle lumen positive (Nelson, 1992).
The bioenergetics of vesicular amino acid transport differs dramatically from that of vesicular amine and acetylcholine transport (Liu, Edwards, 1997).While glutamate transport shows maximum activity under high ΔΨ conditions and inhibition by increasing concentrations of chloride or permeant anions that maximize ΔpH at the expense of ΔΨ (Maycox et al., 1988), acetylcholine transport depends largely on ΔpH (van der Kloot, 2003).Additionally, controversy remains over whether glutamate uptake is driven solely by ΔΨ (Hartinger & Jahn, 1993;Maycox et al., 1988) or by both ΔΨ and ΔpH of the ΔμH+ (Naito, Ueda, 1985;Tabb et al., 1992;Wolosker et al., 1996).It appears that ΔΨ plays a primary role in glutamate uptake but ΔpH may still be essential to antagonize glutamate efflux and retain glutamate inside the vesicles (Wolosker et al., 1996).
Some studies have attempted to investigate how vesicle trafficking occurs in SVs with alterations in neurotransmitter filling (Hong, 2001;Lima et al., 2010;Parsons et al., 1999;Van der Kloot et al., 2002;Zhou et al., 2000) and acidification (Cousin, Nicholls, 1997).They indicated that SV filling is not required for vesicle cycling although the physiological meaning of this data has yet to be elucidated.Unfortunately, the studies above used different experimental models (with distinct neurotransmitter internalization bioenergetics) and stimulus protocols, hampering comparison of results.Also, there is no data available on partially-filled or empty vesicles cycling in the non-conventional ribbon type synapse.
In the present work, using lipophilic styryl dye FM1-43 to monitor exo-endocytosis and pharmacological agents that collapse gradients of ΔμH+, we investigated whether filling and acidification are required for the SV cycle in the cholinergic frog neuromuscular junction and the glutamatergic ribbon type nerve terminal in the retina.

Material
FM1-43 ® was purchased from Invitrogen TM .Methylamine and bafilomycin were purchased from Sigma-Aldrich.All other chemicals and reagents were of analytical grade.All procedures were approved by the local animal care committee (CETEA-UFMG) and followed the guidelines for the Use and Care of Animals for Research issued by the NIH.

Frog neuromuscular junction experiments and imaging analysis
Frog cutaneous pectoris nerve muscle preparations were dissected from Rana catesbeiana (~60g) and mounted in a sylgard-lined chamber.FM1-43 (4 μм) was used to stain the recycling pool of synaptic vesicles (Betz et al., 1992).This constitutes a styryl dye widely used to track exocytosis, endocytosis and recycling of secretory granules and vesicles (revised by Amaral et al., 2011).The dye presents a hydrophobic tail that reversibly binds to membranes, a positive head that prevents membrane permeation, and a body with aromatic rings and a double bond that determines spectral fluorescence properties.Also, FM1-43 is weakly fluorescent in water but its fluorescence intensity is higher in lipid environments.Thus, FM1-43 labels membranes without trespassing them.When the nerve terminal is submitted to a stimulus that evokes exocytosis, the compensatory endocytosis internalizes the dye and results in terminal staining.If the nerve terminal is submitted to a new round of stimulus, the exocytosis of the recycling pool would produce terminal destaining.
The muscles were incubated with d-tubocurarine (16 μм) to prevent contractions during stimulation or image acquisition.Destaining in the absence of stimulus (photobleaching) was used as a control.Experimental neuromuscular preparations were incubated with NH 4 Cl (30 mм), methylamine (10 mм) and bafilomycin (1 µм) for 1 hour.After labeling with FM1-43, exocytosis was investigated by the exposure of neuromuscular preparations to a high K + solution (60 mм KCl) alone or associated to NH 4 Cl, methylamine or bafilomycin.Images were acquired at 5-min intervals until the end of the experiments using a fluorescence microscope (Zeiss Axioskop) coupled to a CCD camera (12 bits, Micromax) and visualized on a computer screen.The microscope was equipped with water immersion objectives (63X, 0.95 NA and 40X, 0.75 NA) and standard fluorescein optics (excitation 480 nm, dichroic 505 nm, emission 535 nm long pass).Excitation light was provided by a 100 W Hg lamp.Image analysis was performed using software Image J, which allows the brightness levels emitted from regions of interest to be measured.Mean fluorescence intensity was determined for each group of spots and plotted as a percentage of its mean initial fluorescence using Microsoft Excel and Sigma Plot 10.0 software.

Goldfish ribbon synapse experiments and imaging analysis
Bipolar cells were acutely dissociated from the retina of goldfish (Carassius auratus auratus), according to von Gersdorff and Matthews (1994).Goldfish were dark-adapted for 1 h and killed by decapitation followed immediately by destruction of the brain and spinal cord.The eyeballs were removed and retinae dissected out and treated for 25 min with hyaluronidase to remove vitreous humor.Each retina was quartered, placed ganglion cell layer down on filter paper and kept at 4°C in medium comprising (mm): NaCl 127, KCl 2.5, MgCl 2 1.0, CaCl 2 0.5, Hepes 5 and glucose 12, adjusted to pH 7.45 with NaOH.Slices were cut into smaller pieces and mounted in a perfusion chamber.FM1-43 fluorescence was imaged with a Zeiss water immersion objective (40x, 1.2 NA) coupled to a Zeiss LSM 510 laser scanning confocal microscope located at CEMEL-UFMG.A 488 argon laser was used to excite the preparation using FITC optics.Dissociated bipolar cells were kept in Fish Ringer associated with methylamine (10 mм), NH 4 Cl (30 mм) or bafilomycin (500 nм) for 5 min before FM1-43 staining.Based on Lagnado and colleagues (1996), bipolar cell membranes were first stained with FM1-43 (4 µм), 0 mм Ca 2+ and KCl 50 mм (iso-osmotic replacement with NaCl) but in association with methylamine, NH 4 Cl or bafilomycin.After 120 s, bipolar cells were perfused with fish Ringer containing FM1-43 (4 µм), Ca 2+ (2.5 mм) and KCl (50 mм) for 360 s in the continued presence of each pharmacological agent, resulting in an image.Fluorescence measurements were performed using the Metamorph Imaging System 3.6 software (Universal Imaging Corporation).Briefly, cell terminal areas at the best focus were quantified as gray level index for FM1-43 fluorescence (control) and after drug exposure.Values were plotted on a graph (Sigma Plot 10.0).

Statistical analysis
Statistical analysis was performed using the paired students t-test or ANOVA.*P values < 0.05 were considered statistically significant.

RESULTS AND DISCUSSION
Figure 1A shows an ex-vivo preparation of frog neuromuscular junction stained with FM1-43 (4 µм) during a high K + stimulus (KCl 60 mм for 10 min).Each fluorescent spot represents clusters of SVs labeled with the dye.NH 4 Cl and methylamine are membrane-permeable weak bases that decrease acidification by disrupting the ΔpH (Abreu et al., 2008;Cousin, Nicholls, 1997;van der Kloot, 1987;Zoccarato et al., 2003).Since ΔpH is required for vesicle loading with acetylcholine, its disruption by NH 4 Cl or methylamine decreases SV filling (Cousin, Nicholls, 1997;Van der Kloot, 1987).Even in the presence of NH 4 Cl or methylamine, motor terminals were labeled with FM1-43 during a high K + stimulus with a similar intensity to that observed under the control condition (Figure 1B, 1C and 1E).Moreover, stimulation with high K + solution in the presence of bafilomycin, a potent VH + -ATPase inhibitor that completely blocks vesicle loading due to entire ΔμH+ collapse (Moriyama et al., 1990), produced FM1-43 labeling equivalent to the control condition (Figure 1D and 1E).
To determine whether SVs with reduced quantal content can be released, FM1-43 destaining induced by high K + solution in the control condition was compared to destaining in the presence of NH 4 Cl, methylamine or bafilomycin.Figure 2 (panels A, C, E, G and I) shows frog motor nerve terminals labeled with FM1-43.Figure 2 also depicts fluorescence loss due to photobleaching (panel 2B), FM1-43 destaining induced by high K + in the control condition (panel 2D) and the destaining caused by high K + in the presence of NH 4 Cl, methylamine and bafilomycin (panels 2F, 2H and 2J, respectively).As can be seen in Figure 2K, inhibition of vesicle filling with neurotransmitter had no influence on FM1-43 destaining.
Next, we tested whether SV cycling occurs normally at bipolar cell terminals in the retina.Synaptic terminals Observe that agents which collapse the electrochemical gradient and lead to a decrease in quantal content did not interfere with endocytosis of synaptic vesicles and FM1-43 uptake .Error bars: ± SEM of 1.110 spots from 237 nerve terminals.
perfused with high K + solution, FM1-43 and 0 mм Ca 2+ in the presence of methylamine (Figure 3A), NH 4 Cl (Figure 3C) and bafilomycin (Figure 3E), exhibit fluorescence only at the plasma membrane.When these terminals were perfused with the same solution but in the presence of 2.5 mм Ca 2+ , the fluorescence was internalized during depolarization in the synaptic terminals for all the experimental conditions tested (Figure 3B, D, F).
FM1-43 loading represents endocytosis of SVs and the endocytic rate is highly related to the previous exocytosis (Wu et al., 1996).Thus, this enabled investigation of vesicle exo and endocytosis by observing the amount of FM1-43 loading in the bipolar cell synaptic terminals.As can be observed in Fig 3G, the quantification of FM1-43 fluorescence was the same for all substances tested.
Despite controversy about glutamate filling, it seems that ΔΨ drives glutamate uptake but ΔpH may avoid neurotransmitter efflux (Wolosker et al., 1996).Thus, SV filling was blocked in experiments with bafilomycin although methylamine and NH 4 Cl might also have caused filling alterations.
The pH levels inside secretory vesicles may modulate many aspects of synaptic function, such as vesicle matrix organization, transport activity of small molecules, activity and maturation of prohormones (Blackmore et al., 2001) and reacidification of SV after a round of exocytosis.Therefore, the pH levels inside vesicles could be a limiting factor during the entire synaptic cycle (Melnik et al., 2001).
On the other hand, an appropriate number of SV filled with a discrete amount of neurotransmitters is needed to sustain synaptic transmission during prolonged depolarization (Edwards, 2007).
In the present work, experiments were performed using the styryl dye FM1-43, a powerful tool for visualizing exocytosis and endocytosis in neuronal and non-neuronal  cell types (Amaral et al., 2011).Using a very similar stimulus protocol for comparison, results showed that vesicle acidification and filling are not required for normal vesicle trafficking in two very distinct synapses: a cholinergic frog neuromuscular junction and a glutamatergic central ribbon type synapse in the retina.We looked for vesicle trafficking impairments when SVs of both synapses show acidification deficits acquired with membrane-permeable weak bases.Specifically in the case of the neuromuscular junction, ΔpH collapse by NH 4 Cl and methylamine also leads to a decrease in SV filling.Furthermore, we studied whether ΔμH+ disruption produced by bafilomycin incubation could affect vesicle cycling.
Supporting these results, a body of evidence suggests that empty SV undergo exo/endocytosis when vesicle acidification and filling are impaired in other experimental models (Cousin, Nicholl, 1997;Parsons et al., 1999;Zhou et al., 2000).
In contrast to the neuromuscular junction, a conventional synapse that encodes information by changing the rate of action potentials (Fagerlund, Eriksson, 2009), bipolar cells transmit their signals tonically and in a graded fashion (tom Dieck, Brandstätter, 2006;Tachibana, 1999).To accomplish this, the compound fusion of SV assures high release rates during strong repetitive stimulation in ribbon synapses (Matthews, Sterling, 2008), even though other rat and mouse synapses also show this mechanism (He et al., 2009).Additionally, glutamate and acetylcholine are accumulated into vesicles using different energetic components of the proton motive force set up across the vesicle membrane by the VH + -ATPase.Due to these remarkable differences, it is reasonable to propose that mechanisms regulating SV cycling might differ between these synapses.
Interestingly, the number of recycling vesicles and the SV cycle remain preserved at the neuromuscular junctions of mice with decreased VAChT protein expression (VAChT knockdown), which in itself is sufficient to reduce quantal size and post-synaptic events (Prado et al., 2006).In fact, even in the complete absence of SV transporters for acetylcholine (de Castro et al., 2009), dopamine and glutamate in knock-out mice, FM1-43 staining and destaining remain unaltered (Croft et al., 2005).The authors hypothesize that a distinct pool of empty SVs would undergo exocytosis and endocytosis, incorporating the vesicular transporter depending on their local availability (Prado et al., 2006).
We cannot rule out the possibility that only a small amount of SV that underwent recycling was truly empty and not acidified and that some fully-filled and acidic vesicles could cycle in a prolonged high K+ stimulus.
However, using electrophysiological techniques, Hong (2001) reported that bafilomycin A and Concanamycin A applied at mouse diaphragm neuromuscular junctions in concentrations and for a period of incubation equivalents to those used in our experiments resulted in twitch tension depression and reduction of quantal size, EPPs and MEEPs amplitude.Additionally, Zhou and colleagues (2000) also observed that incubation with bafilomycin A reduced the amplitude and frequency of miniature inhibitory postsynaptic currents (mIPSCs) in rat hippocampal slices.Furthermore, they also reported that similar results were seen in glutamatergic miniature excitatory postsynaptic currents (mEPSCs) and GABAergic mIPSCs from cultured neurons.Finally, our previous work demonstrated that bafilomycin, NH 4 Cl and methylamine (with the same concentrations and period of incubation used here) decreased SV acidification in the bipolar cell terminals in a rapid fashion (Abreu et al., 2008).In summary, our data with image techniques are in agreement with electrophysiological data, lending further support to the notion that the synaptic vesicle cycle is independent of vesicle filling or type of synapse.
Could the cycling of SVs with reduced or absent quantal content have any physiological relevance?The exo/endocytosis cycle permits fast molecular changes between the plasma membrane and endossome and SV membranes.This membrane flux could be important for the maintenance of plasma membrane lipid composition or for the trafficking of proteins that are mobilized from intracellular membranes when they are recruited for action on the presynaptic membrane.For example, since the high affinity acetylcholine transporter (CHT1) is predominantly present in SVs and its regulation is closely linked to the exocytotic release of acetylcholine, it has been proposed that SVs may play a role in delivering the choline transporter to the cell surface (Ribeiro et al., 2003;Ribeiro et al., 2006).Indeed, Ribeiro and colleagues (2007) demonstrated that CHT1 recycles back to the cell surface and that K + -induced depolarization increases CHT1 cell surface localization by enhancing the recruitment of intracellular transporters to the cell surface.The transport of these proteins, along the SV-cell membrane could explain how some neurons support neurotransmission signals during intense activity.