Synaptic vesicle recycling is enhanced by torsinA that harbors the DYT1 dystonia mutation


  • Yasuhiro Kakazu,

    1. Department of Molecular Physiology and Biophysics, University of Iowa Carver College of Medicine, Iowa City, Iowa
    Search for more papers by this author
  • Jin-Young Koh,

    1. Department of Molecular Physiology and Biophysics, University of Iowa Carver College of Medicine, Iowa City, Iowa
    Search for more papers by this author
  • K. W. David Ho,

    1. Department of Molecular Physiology and Biophysics, University of Iowa Carver College of Medicine, Iowa City, Iowa
    2. Medical Scientist Training Program, University of Iowa Carver College of Medicine, Iowa City, Iowa
    Search for more papers by this author
  • Pedro Gonzalez-Alegre,

    1. Medical Scientist Training Program, University of Iowa Carver College of Medicine, Iowa City, Iowa
    2. Department of Neurology, University of Iowa Carver College of Medicine, Iowa City, Iowa
    Search for more papers by this author
  • N. Charles Harata

    Corresponding author
    1. Department of Molecular Physiology and Biophysics, University of Iowa Carver College of Medicine, Iowa City, Iowa
    2. Medical Scientist Training Program, University of Iowa Carver College of Medicine, Iowa City, Iowa
    • Department of Molecular Physiology and Biophysics, University of Iowa Carver College of Medicine, 51 Newton Road, Iowa City, IA 52242, USA
    Search for more papers by this author


Early-onset generalized dystonia, DYT1, is caused by a mutation in the gene encoding the evolutionarily conserved AAA+ ATPase torsinA. Synaptic abnormalities have been implicated in DYT1 dystonia, but the details of the synaptic pathophysiology are only partially understood. Here, we demonstrate a novel role for torsinA in synaptic vesicle recycling, using cultured hippocampal neurons from a knock-in mouse model of DYT1 dystonia (ΔE-torsinA) and live-cell imaging with styryl FM dyes. Neurons from heterozygous ΔE-torsinA mice released a larger fraction of the total recycling pool (TRP) during a single round of electrical stimulation than did wild-type neurons. Moreover, when the neurons were subjected to prior high activity, the time course of release was shortened. In neurons from homozygous mice, these enhanced exocytosis phenotypes were similar, but in addition the size of the TRP was reduced. Notably, when release was triggered by applying a calcium ionophore rather than electrical stimuli, neither a single nor two ΔE-torsinA alleles affected the time course of release. Thus, the site of action of ΔE-torsinA is at or upstream of the rise in calcium concentration in nerve terminals. Our results suggest that torsinA regulates synaptic vesicle recycling in central neurons. They also indicate that this regulation is influenced by neuronal activity, further supporting the idea that synaptic abnormalities contribute to the pathophysiology of DYT1 dystonia. Synapse, 2012. © 2011 Wiley Periodicals, Inc.


Following exocytosis, synaptic vesicles reform by endocytosis and are reloaded with neurotransmitters for subsequent rounds of transmitter release (Rizzoli and Betz,2005; Schweizer and Ryan,2006). Synaptic transmission relies on the efficiency of recycling, which is influenced by the number of recycling vesicles, the rate of exocytosis, and the rate of endocytosis. These parameters can be influenced by neuronal activity (Zakharenko et al.,2001, 2002) and pathological conditions (Waites and Garner,2011). Numerous proteins regulate synaptic-vesicle recycling (Chua et al.,2010; Sudhof,2004), such that defects in these proteins or their interactions with one another may cause aberrant synaptic transmission and, subsequently, neurological disease.

TorsinA is an evolutionarily conserved member of the family of ATPases associated with diverse cellular activities (AAA+) (Breakefield et al.,2001; Jungwirth et al.,2010). These proteins generally perform chaperone-like functions, assisting in protein unfolding, protein complex disassembly, membrane trafficking, and fusion (Hanson and Whiteheart,2005). An in-frame deletion of a glutamate codon in the torsinA gene (ΔE-torsinA) has been linked to DYT1 dystonia (Ozelius et al.,1997), an autosomal-dominant movement disorder that is the most common hereditary form of early onset dystonia (Ozelius and Bressman,2011). Clinical studies have shown that neuronal excitability and metabolism are enhanced in these patients (Argyelan et al.,2009; Carbon et al.,2010), despite a lack of neurodegeneration or inflammation (Rostasy et al.,2003). These findings support the hypothesis that the pathophysiology of DYT1 dystonia may result from dysfunction at synapses or abnormalities in brain circuits (Breakefield et al.,2008).

Consistent with this hypothesis, overexpression of the mutant ΔE-torsinA protein in cultured neuroblastoma cells and hippocampal neurons was found to affect the recycling of synaptic vesicles (Granata et al.,2008, 2011). Although these reports demonstrated that synaptic vesicles in nerve terminals can be a target of new research, these results might not apply directly to autosomal-dominant DYT1 dystonia, where the carriers (who are heterozygous for ΔE-torsinA) do not overexpress torsinA (Goodchild et al.,2005). In addition, only 30–40% of the mutation carriers show overt neurological symptoms, which suggests that additional genetic (Kock et al.,2006; Risch et al.,2007) or environmental factors (Edwards et al.,2003a; Gioltzoglou et al.,2006) contribute to the onset of symptoms as second hits (Breakefield et al.,2008).

To evaluate the condition of synapses in ΔE-torsinA mutation carriers, we evaluated the presynaptic properties of neurons obtained from the ΔE-torsinA knock-in mouse (Dang et al.,2005; Goodchild et al.,2005). This is an animal model that does not show overt motor symptoms yet reproduces abnormalities in the central nervous system (with respect to the circuitry and metabolic activity) that have been observed in the human, nonmanifesting (i.e., without motor symptoms) carriers of ΔE-torsinA (Ulug et al.,2011). Using this system, we tested whether exocytosis and endocytosis of synaptic vesicles are affected in central neurons, and whether the change is induced by strong neuronal activity. Our findings support the hypothesis that torsinA is involved in synaptic functions. They further provide the first evidence for activity-induced changes in synaptic vesicle exocytosis in the carriers of ΔE-torsinA mutation.



Animal care and procedures were approved by the University of Iowa Animal Care and Use Committee and performed in accordance with the standards set by the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80-23), revised 1996. On postnatal days 0–1, pups of either sex of the knock-in model of DYT1 dystonia (Goodchild et al.,2005) were genotyped according to the fast-genotyping procedure (within ∼5 h, EZ Fast Tissue/Tail PCR Genotyping Kit, EZ BioResearch LLC, St, Louis, MO), using a published protocol (Goodchild et al.,2005).


Primary hippocampal neurons were cultured by a method in our previous work (Harata et al.,2006), with some modification. Briefly, the CA3-CA1 regions of the hippocampus were dissected on postnatal days 0–1, trypsinized and dissociated. The cells were plated at a density of 12,000 cells per well of a 24-well dish, on 12-mm coverslips preseeded with a rat glial feeder layer (Garcia-Junco-Clemente et al.,2010). The feeder layer had been seeded in a plating medium with the following composition: MEM (Invitrogen, Carlsbad, CA) plus 5 g/L glucose, 0.2 g/L NaHCO3, 100 mg/L bovine transferrin (EMD Chemicals, Gibbstown, NJ), 2 mM GlutaMAX (Invitrogen), 25 mg/L insulin, and 10% fetal bovine serum (Invitrogen). Feeder layers were maintained in a 1:1 mixture of plating medium and growth medium. The latter had the following composition: MEM plus 4 μM cytosine β-D-arabinofuranoside, 0.5 mM GlutaMAX, NS21 (Chen et al.,2008), and 5% FBS. The hippocampal neurons were used on day 11–13 of culture. The experimental data were obtained from cultures derived from 3 to 5 distinct animals per genotype (wild-type, heterozygous, and homozygous littermates).

Fluorescence imaging of FM dye

Nerve terminals (boutons) were stained with FM4-64 (Invitrogen; Gaffield and Betz,2006) by one of two methods. No antagonists were present during staining.

Basal activity staining method: neurons were immersed in 2.5 μM FM4-64 in MEM for 10 min in the culture incubator (37°C).

High-activity staining method: neurons were immersed in 2.5 μM FM4-64 in a high-K+ solution (45 mM KCl) for 1 min at room temperature (23–25°C). The high-K+ solution was prepared by equimolar substitution of KCl for NaCl in Tyrode solution (see below for composition).

After staining by either method, neurons were transferred to an imaging chamber with stimulation electrodes (RC-21BRFS, Warner Instruments, Hamden, CT). Neurons were washed with dye-free Tyrode solution (in mM: 125 NaCl, 2 KCl, 2 CaCl2, 2 MgCl2, 30 glucose, 25 HEPES, 310 mOsm, and pH 7.4) for 6–7 min at room temperature. Spontaneous loss of FM dye was suppressed by AMPA receptor antagonist CNQX (10 μM, Tocris Bioscience, Ellisville, MO), NMDA receptor antagonist D,L-AP5 (50 μM, Tocris) and the voltage-gated Na+ channel blocker tetrodotoxin (TTX, 0.5 μM, Tocris).

After washing, the boutons were destained by one of two methods at room temperature.

Field-stimulation destaining method: Extensive washing was first carried out for 1 min to remove TTX. The effectiveness of TTX washout was confirmed in separate experiments with patch clamp recording of voltage-gated Na+ current serving as the readout. Field stimulation was then applied (1-msec constant current at 30 mA, repeated at 10 Hz for 120 sec), using a pulse generator (Master-8, AMPI, Jerusalem, Israel) and an isolated stimulator (DS3, Digitimer, Hertfordshire, UK). This destaining round was repeated three times at 80-sec intervals to ensure that most of the stained synaptic vesicles released FM dye (Harata et al.,2006). CNQX and AP5 were present throughout the washing and imaging periods.

Ionomycin destaining method: The Ca2+ ionophore, ionomycin (5 μM), was applied for 120 sec in the continued presence of CNQX, AP5, and TTX.

For each coverslip, only a single imaging field was analyzed; this eliminates the possibility that any of the fields imaged are subject to presynaptic changes induced by high activity (Fig. 3 Bc) during an earlier experiment.


The structures labeled by FM dye during live-cell imaging were identified as nerve terminals by retrospective immunocytochemistry for the presynaptic marker vesicular glutamate transporter 1 (VGLUT1). After live-cell FM imaging, the cultured neurons were fixed with 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) and 4% sucrose in Tyrode solution for 30 min at 4°C. After being rinsed with Tyrode solution twice for 5 min each at 4°C, the cells were blocked and permeabilized with 2% normal goat serum (Jackson ImmunoResearch Laboratories, West Grove, PA) and 0.4% saponin in phosphate-buffered saline (PBS, Invitrogen) (blocking solution), for 60 min at room temperature. Thereafter, they were treated with polyclonal, guinea pig anti-VGLUT1 antibody (AB5905, Chemicon-Millipore, Billerica, MA; diluted 1,000 times in the blocking solution), overnight (15–21 h) at 4°C. Following rinsing with PBS, three times for 7 min each, the neurons were incubated with goat anti-guinea pig IgG antibody conjugated with Alexa Fluor 594 (Invitrogen; diluted 1,000 times in the blocking solution), for 60 min at room temperature. They were rinsed with PBS at least five times for 20 min each and observed directly in PBS. The same neurons were identified for live-cell FM imaging and fixed-cell immunocytochemistry, using differential interference contrast (DIC) images as a reference.

Fluorescence imaging system

Cells were imaged using an inverted microscope (Eclipse-TiE, Nikon, Melville, NY) equipped with an EMCCD camera (DU-860, Andor Technology, Belfast, UK). The camera was continuously perfused with chilled water (Oasis 160 liquid recirculating chiller, Solid State Cooling Systems, Wappingers Falls, NY) to maintain a temperature of −80°C, and to thereby reduce noise.

FM dye was excited using a 490-nm LED (CoolLED-Custom Interconnect, Hampshire, UK) with 10% intensity, and imaged with an objective lens (Plan Fluor, 40×, NA1.30, Nikon), a filter cube (490/20-nm ex, 510-nm dclp, and 650-nm-LP em), and 0.7× coupler. The 16-bit images were acquired at 1 frame/sec, 20 msec exposure, EM gain of 50, without binning, using the Solis software (Andor). Neuronal exposure to excitation light was minimized by turning the LED on only during image capture (triggered by “Fire” output of the camera).

For immunocytochemical observation, Alexa Fluor 594 visualization was carried out using largely the same set of imaging parameters used for FM4-64; exceptions were use of a 595-nm LED (CoolLED-Custom Interconnect), use of a filter cube (590/55-nm ex, 625-nm dclp, and 665/65-nm em), and 100 msec exposure. FM4-64 emits in the red visible range, but its fluorescence emission was lost during the fixation procedure (data not shown). Thus, FM4-64 did not interfere with the observation of Alexa Fluor 594 after fixation.

Image analysis of FM experiments

FM signal was quantified using Image J (v1.43 m, W. S. Rasband, NIH) and the associated plug-ins. Acquired images in a time series (“Stack”) were aligned, using the ImageJ-Image Stabilizer plug-in (Kang Li) to correct for small movements in FM signals. Functional boutons were identified based on differences in fluorescence intensity (ΔFM) between the five-frame averages taken immediately before and after the destaining series. Regions-of-interest (ROIs, 3 × 3 pixels, 2.4 × 2.4 μm2) were assigned on isolated, fluorescent puncta in the ΔFM image, if five or more of the nine pixels showed intensity above 30 arbitrary units (a.u.), which was approximately the standard deviation of the background intensity obtained from the bare coverslip area. Changes in ROI intensity were measured using ImageJ-Time Series Analyzer V2.0 (Balaji Jayaprakash) and were exported to Microsoft excel. ROIs were excluded if they exhibited any of the following changes in intensity: an increase during recording, a sudden decrease before stimulation or a long latency after stimulation. Photobleaching of FM dyes was minimal in our system (see prestimulation period in Fig. 1 C), and was not corrected. The nonparametric, Kolmogorov-Smirnov test was used to assess the statistical significance at P < 0.0001.

Figure 1.

Experimental system. A: Protocols used for staining and destaining of cultured neurons with FM4-64. (a) Staining by basal activity method (incubation with FM4-64 at 37°C for 10 min), and destaining by electrical field-stimulation method (three rounds of 10 Hz stimulation for 2 min). (b) Staining by the high-activity method (application of high K+, i.e., 45 mM KCl), and destaining by field stimulation. (c) Staining by basal activity method, and destaining by ionomycin method (5 μM for 2 min). (d) Staining by high-activity method and destaining by ionomycin method. The high-activity staining method served two purposes: one was to stain the neurons, and the other was to introduce strong neuronal stimulation before examining the destaining phase for the readout of activity-dependent change in synaptic properties. The basal-activity staining was used as a control staining method without additional stimulation. B: Representative image of a cultured neuron observed using DIC optics (a). The boxed area is enlarged (b–f) and imaged by DIC (b), and by epifluorescence optics, both before (c) and after (d) three rounds of field stimulation for destaining. Images illustrating the difference in fluorescence intensity (ΔFM; c–d) are also shown (e and f), with regions of interest (ROIs) indicated in (f). C: Changes in the fluorescence intensity of six representative ROIs. Wild-type neurons were stained using the high-activity method and destained by the field-stimulation method (protocol in Fig. 1 Ab). D: Colocalization of FM signal with the presynaptic marker VGLUT1. A dendritic region (visualized with DIC) of a homozygous neuron was stained (FM before Stim) and destained with field stimulation. A difference image (ΔFM) is similar to the stained image (FM before Stim), indicating that the FM-labeled structure can release the FM dye efficiently. This is consistent with the structure being the synaptic vesicles. After the specimen was fixed, the VGLUT1 immuno-fluorescence was localized at the same region.

The absolute value of initial slope (a.u./s) was calculated according to the following formula (Daniel et al.,2009) with slight modification. equation image where a.u. represents arbitrary unit of fluorescence intensity; ΔFM1 round the amount of FM destaining during a single round of electrical stimulation; τ the time constant of destaining during a single round of electrical stimulation; t50 the time required for FM signal to decrease to 50% of its initial value; and ln the natural logarithm. This formula assumes that the majority of FM release events follow the following rules: (1) the vesicles release the same amount of FM during a given exocytotic event, (2) a given nerve terminal has only one release site, and (3) each successful release event involves only a single vesicle.


All chemical reagents were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise specified. Reagents were applied by bath perfusion (600 μL/min) and the “Y-tube” method, a fast application system that allows exchange of the external solution within 30 msec (Harata et al.,1999).


Synaptic vesicle recycling was assessed in cultured neurons from the mouse hippocampus, a site at which torsinA transcripts and protein are abundantly expressed (Allen_Mouse_Brain_Atlas,2009; Augood et al.,1999; Walker et al.,2001). In addition, mechanisms underlying presynaptic signaling have been well characterized in these neurons.

Live-cell imaging of FM dye, a marker of recycling (functional) synaptic vesicles

Due to spontaneous activity in resting neurons, it was possible to load FM dye into recycling synaptic vesicles by simply incubating the neurons in FM dye (basal activity staining method, Fig. 1 Aa). However, to induce extensive neuronal stimulation, we took a second approach whereby neurons were exposed to high K+, which causes depolarization (high-activity staining method, Fig. 1 Ab). Following staining by either method, the FM dye was washed out and exocytosis was monitored over the course of three rounds of electrical field stimulation (Fig. 1 B). The decrease in FM fluorescence (destaining) due to fusion of the labeled vesicles with the plasma membrane (Fig. 1 Bc, d) provided a measure of vesicle exocytosis. The change in FM intensity (ΔFM) before and after electrical stimulation (Fig. 1 Be) was used to assign regions of interest (ROIs) that correspond to individual nerve terminals (Fig. 1 Bf). The fluorescence intensity of the ROIs was decreased in response to the stimuli until only a negligible amount of fluorescence loss was observed by the third round of electrical stimulation (Fig. 1 C). This suggests that three rounds of stimulation were sufficient to destain all the recycling vesicles that were loaded with FM during the staining phase.

Immunocytochemistry was used to analyze the structure labeled by FM dye during live-cell imaging (Fig. 1 D). Neurons were stained by the high-activity method (outlined in Fig. 1 Ab), and a dendritic region was imaged before and after destaining by field stimulation. The similarity between the stained image (FM before Stim) and ΔFM image (ΔFM) indicates that the structure labeled by the FM dye can actively release the FM dye with minimal trapping. Retrospective immunocytochemistry of the same regions demonstrated that the FM label colocalized mostly with the presynaptic marker vesicular glutamate transporter 1 (VGLUT1). These results confirm that FM dye labeled the nerve terminals.

Size of total recycling pool (TRP)

After neurons were stained using the basal activity method, the ROI intensity was monitored in neurons obtained from different genotypes (Fig. 2 Aa). The size of the TRP was defined as the difference in the absolute FM intensities before and after three rounds of electrical stimulation at each bouton (ΔFM3 rounds). A black arrow in Figure 2 Aa indicates a sample measurement of ΔFM3 rounds from wild-type neurons. The TRP size correlates with the capacity of nerve terminals to respond to high-frequency stimulation. To compare the TRP size of the mutants with that of wild-type neurons, we plotted TRP sizes (ΔFM3 rounds values) of individual ROIs in a cumulative format (Fig. 2 Ab). Each curve of the cumulative histogram was obtained by averaging the cumulative histograms obtained from individual coverslips of the same genotype (insets), to eliminate different weights of coverslips due to different numbers of ROIs. The TRP size in heterozygous neurons (Hetero, red) did not differ from that in wild-type neurons (WT, black). In contrast, TRP size was significantly smaller in homozygous ΔE-torsinA neurons (Homo, green) than in wild-type neurons (P < 10−10; Kolmogorov-Smirnov test). A similar result was obtained for neurons stained by the high-activity method (P < 10−4; Fig. 2 Ac, d).

Figure 2.

Genotypic differences observable in the absolute intensity of FM dyes. Changes in the absolute intensity of FM4-64 before and after three rounds (A, ΔFM3 rounds) or one round (B, ΔFM1 round) of field stimulation. Results were obtained following staining by the basal-activity (left column) or high-activity (right column) method. (a and c), Averaged time-course traces (mean ± SEM). Filled bars indicate timing of field stimulation (10 Hz, 120 s). Double-ended arrows indicate the measurements of ΔFM3 rounds = TRP (Aa,c) or ΔFM1 round (Ba,c), with a wild-type case (black traces) serving as an example. Numbers in parentheses indicate those of analyzed boutons. (b and d), Cumulative histograms of ΔFM in all ROIs obtained in corresponding experiments (a, c). Insets show the overlays of cumulative histograms from ROIs on individual coverslips. Asterisk, significant difference with respect to wild-type (P < 10−4, Kolmogorov-Smirnov test), color-coded for a genotype; NS, not significant. Figures 2–4 have the following, common format of presentation. Left column represents the destaining data after staining by the basal-activity method, and right column represents the destaining data after staining by the high-activity method. Top panels (a and c) show the FM destaining traces, obtained by averaging all ROIs (boutons) in a given genotype. Bottom panels (b and d) show the cumulative histograms of measurements from individual ROIs. The individual measurement is exemplified by arrows in panels a and c (although panels a and c show averaged traces). Thus the lengths of arrows in panels a and c correspond to the x-axes in panels b and d.[Color figure can be viewed in the online issue, which is available at]

Amount of FM dye release during one round of electrical stimulation

Exocytosis in different genotypes was evaluated by performing several analyses. In one, we measured the amount of FM destaining during a single round of electrical stimulation (ΔFM1 round, with black arrow representing that for wild-type neurons in Fig. 2 Ba). This value was significantly larger in heterozygous than in wild-type neurons (P < 10−6; Fig. 2 Bb). This large ΔFM1 round occurred in heterozygous neurons, also when the neurons were stained by the high activity method (P < 10−4, Fig. 2 Bc, d). Notably, homozygotes did not exhibit this property when stained by either method (Fig. 2 Bb, d).

Fraction of total recycling-vesicle pool that released FM during one round of stimulation

The above-described results can be explained qualitatively by the release of a higher fraction of the TRP (Releasable %) in both heterozygous and homozygous ΔE-torsinA neurons, since TRP × Releasable % (per 1-round stimulus) = ΔFM1 round. To test this possibility, we normalized ΔFM values to the TRP size of each ROI (Fig. 3 A). As expected, the releasable % was significantly higher in both heterozygous and homozygous neurons than in wild-type neurons (P < 10−7), when they were stained by either the basal activity (Fig. 3 Aa, b) or high activity (Fig. 3 Ac, d) method.

Figure 3.

Genotypic differences observable in fractional changes in FM intensity. Staining was by either the basal-activity (left column) or high-activity (right column) method. A: Data in Figure 2 A were normalized to the pre-destaining level of FM4-64 intensity (a and c). Double-ended arrow indicates the measurement used to assess the fraction of TRP that was released (Releasable %), with a wild-type case (black) serving as an example. Cumulative histograms plot the changes in the fraction of TRP that was released at the end of 1 round of stimulation (b and d). B: Data in Figure 2 B were normalized to the predestaining level of FM4-64 intensity (a and c). Inset shows one raw trace (gray) overlaid with a running average of 11 points (black), used for measuring the half decay time (t50). Cumulative histograms plot t50 (b and d). Filled bars indicate field stimulation (10 Hz; 120 s). As in Figure 2, the left column represents data after staining by the basal-activity method, and the right column represents the data after staining by the high-activity method. Top panels (a and c) show the averaged traces of FM destaining; bottom panels (b and d) show the cumulative histograms of a measurement from individual ROIs, as exemplified by arrows in panels a and c. [Color figure can be viewed in the online issue, which is available at]

Time course of FM release

Is the increase in percentage release associated with the ΔE-torsinA allele due to faster exocytosis during a single round of stimulation? To test this, the time course of FM destaining was assessed by measuring the time required for 50% decay of the FM signal with respect to the onset of stimulation (t50). Following staining by the basal activity method, both heterozygous and homozygous ΔE-torsinA neurons showed the same destaining time course as wild-type neurons (Fig. 3 Ba, b). However, following high-activity staining, FM destaining was significantly faster for both of the ΔE-torsinA genotypes than for wild-type neurons (P < 10−20; Fig. 3 Bc, d). This result indicates that in both mutants, prior, high neuronal activity led to the change in presynaptic phenotype. This effect, in principle, could have been due to changes in the dynamics of vesicle exocytosis, or to changes in an upstream event such as a rise in intracellular Ca2+ concentration that mediate exocytosis.

To clarify at which stage the ΔE-torsinA allele influences exocytosis, we bypassed the need for activities of Ca2+-raising mechanisms, such as a voltage-gated Ca2+ channel, by using the Ca2+ ionophore ionomycin to force Ca2+ influx (Piedras-Renteria et al.,2004). The neurons were stained with FM dye by either the basal activity (Fig. 1 Ac) or high activity (Fig. 1 Ad) method. Subsequent ionomycin application induced FM destaining without noticeably changing the time course across genotypes, regardless of staining method (Fig. 4). The lack of an acceleration in exocytosis following high-activity staining in this context suggests that ΔE-torsinA causes abnormal vesicle release, by acting on regulators of the intracellular Ca2+ concentration, such as voltage-gated Ca2+ channels, intracellular Ca2+ stores, or influencing upstream events that lead to their activation.

Figure 4.

Neurons of all genotypes exhibit similar sensitivity to intracellular Ca2+ influx. Experiments were similar to those depicted in Figure 3 B, but boutons were destained using the Ca2+ ionophore, ionomycin. Data were acquired after staining by the basal-activity method (left column, using the protocol in Fig. 1Ac) or the high-activity method (right column, using the protocol in Fig. 1Ad). Open bars indicate the timing of ionomycin application. As in Figure 2, the left column represents data after staining by the basal-activity method, and the right column represents the data after staining by the high-activity method. The top panels (a and c) show the averaged traces of FM destaining; bottom panels (b and d) show the cumulative histograms of a measurement from individual ROIs, as exemplified by arrows in panels a and c. [Color figure can be viewed in the online issue, which is available at]

Initial slope

Finally, we analyzed the initial slope of FM destaining. This parameter represents the amount of FM dye released during an initial, short period of time (see Methods for the formula and assumptions needed for this calculation). In our experiments, the neurons were electrically stimulated at 10 Hz, while the imaging rate was 1 frame/sec. Thus, the initial slope of FM destaining will reflect a sum of multiple release events over multiple electrical pulses. Although this parameter is similar to the concept of release probability in electrophysiology, they are different because the release probability in electrophysiology typically represents a probability of release in response to a single, rather than multiple, electrical stimulation (e.g., Chen et al.,2004).

The calculated slope was indistinguishable between mutants and wild-type neurons when the basal activity method was used for staining (Fig. 5 Aa). However, it was significantly higher in both heterozygous and homozygous neurons than in wild-type neurons when the high-activity method was used for staining (P < 10−20; Fig. 5 Ab). The results suggest that the synaptic vesicle exocytosis was enhanced in the mutant neurons after high neuronal activity.

Figure 5.

Mutant neurons show large values of initial slopes in FM destaining. A: Cumulative histograms of the slopes of destaining by the field-stimulation method, after staining by either the basal-activity method (left, using the protocol in Fig. 1Aa) or high-activity method (right, using the protocol in Fig. 1Ab). The initial slopes were calculated based on the formula in Materials and Methods. B: Summary of findings from the present study. Vertical arrows indicate the directions of change with respect to wild-type neurons (P < 10−4, Kolmogorov-Smirnov test). Horizontal arrows indicate that there is no difference with respect to wild-type neurons. [Color figure can be viewed in the online issue, which is available at]


This is the first report of synaptic vesicle recycling in an animal model that mimics the condition of nonmanifesting carriers of the ΔE-torsinA mutation (knock-in mice). FM dyes allowed us to distinguish general features of synaptic vesicle exocytosis and endocytosis from the features of processes that depend on the properties of endogenous neurotransmitters, for example, transport from the extracellular space to the cytosol, transport from the cytosol to the synaptic vesicle, and activation of postsynaptic receptors. Capitalizing on the usefulness of FM dyes, we have evaluated the presynaptic effects of ΔE-torsinA. Our results demonstrate (summarized in Fig. 5 B) that a single copy of ΔE-torsinA causes an increase in exocytosis without an accompanying change in TRP size. We also found that a single copy of ΔE-torsinA accelerates exocytosis following cellular exposure to high activity. These enhancements of synaptic vesicle exocytosis in heterozygous neurons suggest that wild-type torsinA inhibits synaptic transmission through a presynaptic mechanism, and that ΔE-torsinA relieves this brake, especially after high activity due to strong stimulation. It is of note that homozygous neurons exhibited not only an increase in exocytosis, but also a reduction in TRP size that was not detectable in heterozygous neurons. This raises the possibility that torsinA has more than one site of action in synaptic vesicle recycling.

Synaptic vesicle recycling in ΔE-torsinA knock-in mice

Prior to this study, Warner and coworkers first identified defective synaptic vesicle recycling in cell models of DYT1 dystonia, triggering this study to explore further the presynaptic abnormalities. However, the results from the previous study were different from ours. For example, it was claimed that the overexpression of ΔE-torsinA in neuroblastoma cells increased the uptake of FM dye (FM1-43), suggesting that vesicular endocytosis was enhanced (Granata et al.,2008). This contradiction from our results seems to have arisen from the complexity of synaptic vesicle recycling, in addition to the differences in cellular preparations (neuroblastoma vs. hippocampal neurons) and in expression levels of torsinA (overexpression vs. endogenous level). FM dyes can be loaded into not only recycling synaptic vesicles, but also intracellular compartments that do not allow immediate rerelease, and this leads to residual FM staining even after extensive stimulation (Granata et al.,2008). Given that this residual fraction does not represent the recycling vesicles, it should be subtracted from the FM signal, as in our analysis of ΔFM. However, collectively, our results and those from Warner and coworkers indicated the presence of abnormal presynaptic features in models of DYT1 dystonia.

The main, novel features of our findings are common to both heterozygous and homozygous ΔE-torsinA neurons. Specifically: (1) exocytosis in response to a single round of electrical stimulation involves a higher percentage of TRP of vesicles (Fig. 3 A) and (2) this release is accelerated when the neurons are stimulated prior to destaining (i.e., high activity staining method; Figs. 3 B and 5 B). These enhancements in exocytosis demonstrate that ΔE-torsinA induces fundamental changes in synaptic-vesicle recycling. These enhancements are expected to correlate with increased neurotransmitter release during high-frequency firing of action potentials, especially in heterozygous neurons as they released higher amount of FM dye (ΔFM1 round; Fig. 2 B). Thus, these enhancements are expected to have a strong impact on the efficiency of synaptic transmission.

An intriguing finding in the knock-in mice was that prior stimulation of the neurons (high activity staining method) accelerated FM destaining (Fig. 3 Bc, d). The fact that this phenotype was not evident with the basal activity staining method (Fig. 3 Ba, b) indicates that high activity can serve as a key trigger for the neuronal phenotype. The mechanism that leads to accelerated activity is expected to influence that controlling the intracellular Ca2+ concentration, as there was no difference in the sensitivity to ionomycin (Fig. 4). One possible explanation for this acceleration is that synaptic plasticity was induced by high activity staining method. A similar acceleration of FM destaining in nerve terminals was noted during long-term potentiation of synaptic transmission in the hippocampus of wild-type rodents (Zakharenko et al.,2001). Interestingly, the acceleration was observed only in forms of long-term potentiation whose induction relied on voltage-gated Ca2+ channels (L-type) (Zakharenko et al.,2001). The possibility that ΔE-torsinA might influence synaptic plasticity is consistent with the observation that the basal ganglia of transgenic mice expressing this protein exhibit enhanced long-term potentiation (Martella et al.,2009) and modified voltage-gated Ca2+ channel (N-type) activity (Pisani et al.,2006; Sciamanna et al.,2011). Another possible explanation for the apparent acceleration in exocytosis is that the two staining methods led to labeling of different pools of vesicles; this would be independent of synaptic plasticity, and consistent with the notion that spontaneous and evoked release in hippocampal neurons involves separate vesicle pools (Chung et al.,2010). However, it should be noted that the concept of such different vesicle pools remains controversial, because contradictory results have also been reported (Hua et al.,2010).

Mice homozygous for ΔE-torsinA die within 2–3 days after birth, whereas heterozygous mice are indistinguishable from wild-type mice (Goodchild et al.,2005). In this study, the synaptic phenotypes of the cultured homozygous neurons did not differ drastically from those of their heterozygous counterparts. These findings are consistent with the notion that the lethality in homozygous knock-in mice is associated with the reported abnormalities in the nuclear membranes of these neurons (Yokoi et al.,2011). However, it is also possible that the lethality is due to more severe functional abnormalities in other areas of the central nervous system, such as the spinal cord.

Implications for DYT1 dystonia

The ΔE-torsinA knock-in mice used in this study do not show overt motor symptoms, but do show abnormalities in the circuitry of and metabolic activity in the brain. As such, the phenotype resembles that in human, nonmanifesting, mutation carriers of DYT1 dystonia (Ulug et al.,2011). This is in sharp contrast to the phenotype of ΔE-torsinA transgenic mice that demonstrate motor symptoms (e.g., Grundmann et al.,2007; Shashidharan et al.,2005). Therefore, our results have implications for the pathophysiology of nonmanifesting DYT1 dystonia, with the caveat that the hippocampus may not be the main site affected in manifesting DYT1 dystonia (see the last paragraph).

First, the presence of synaptic phenotypes indicates that synaptic abnormalities are prevalent among carriers of the mutation, and this may underlie some of the endophenotypes (i.e., subclinical markers) for nonmanifesting human carriers. These include a number of deficits, for example, in electrophysiological responses (Edwards et al.,2003b), temporal processing of sensory stimuli (Fiorio et al.,2007), motor sequence learning (Carbon et al.,2011; Ghilardi et al.,2003), and sensorimotor cortical activity (Carbon et al.,2010), as well as an increase in susceptibility to recurrent major depression (Heiman et al.,2004).

Second, our results demonstrate that strong neuronal activity can alter the synaptic phenotype, and therefore may serve as a “second hit” in ΔE-torsinA neurons. The low penetrance and phenotypic variability of DYT1 dystonia imply that genetic polymorphisms (Kock et al.,2006; Risch et al.,2007) or environmental factors (Edwards et al.,2003a; Gioltzoglou et al.,2006) play important roles in the pathogenesis of DYT1 dystonia, as is the case for other CNS disorders such as schizophrenia (Ayhan et al.,2009). Our finding of activity-induced change may reflect one cellular mechanism whereby environmental insults contribute to the emergence of symptoms. The details of this mechanism are unclear at present, and it remains to be determined which other cellular responses (e.g., inflammation and ischemia) can serve as environmental insults and modifiers of neuronal phenotypes. Our experimental system represents a useful tool for dissecting the gene-environment interactions that influence dystonia.

Traditionally, the striatum (mostly the putamen) has been considered the primary site of dysfunction in dystonia. More recently, dystonia was hypothesized to be a network disorder that involves at least the cerebellum, thalamus, striatum, and cerebral cortex (Neychev et al.,2011). How torsinA or other dystonia-causing proteins alter this circuit remains unknown. We acknowledge that the hippocampus is unlikely to be the primary trigger of the pathogenic process. Nevertheless, there are several reasons to study synaptic physiology in this particular region of the brain. First, the hippocampus is a well-characterized experimental system that is widely used to explore synaptic physiology (e.g., Ruiz et al.,2010; Terashima et al.,2008; Zhou et al.,2000) and the effects of defective proteins (e.g., Deak et al.,2004; Feng et al.,2010). Second, the fact that hippocampal neurons express torsinA at high levels (Allen_Mouse_Brain_Atlas,2009; Augood et al.,1999; Walker et al.,2001) and upregulate its expression further as a consequence of ischemia (Zhao et al.,2008) suggests that torsinA plays a functional role in the context of neuronal stress in this region. Third, hippocampal cells also express other proteins, such as ε-sarcoglycan, associated with other forms of dystonia (Ritz et al.,2011). Fourth, the hippocampus is one of the highly activated sites after the systemic administration of L-type Ca2+ channel activators triggered dystonia in rodents (Jinnah et al.,2003). Thus, expanding studies of dystonia into this well-established mammalian neuronal system represents a step toward better understanding the roles of torsinA in mammalian synapses. However, it will fall to future studies to address whether this is a universal phenomenon or restricted to specific neuronal subtypes.


The authors thank Dr. Amy Lee for a detailed critique of the manuscript.