, 1999 and Sutton et al., 2006). The identity of the vesicles supporting these two modes of neurotransmission remains, however, highly debated (Chung et al., 2010, Fredj and Burrone, 2009, Groemer and Klingauf, BKM120 order 2007, Hua et al., 2010, Hua et al., 2011, Sara et al., 2005 and Wilhelm et al., 2010). One current view is that spontaneous events represent the stochastic fusion of vesicles

that are already docked and primed for release (Murthy and Stevens, 1999) and are driven by the same molecular machinery that supports evoked vesicle fusion (Sudhof, 2004). These “spontaneous” vesicles normally have a very low probability of fusion, which is heightened upon stimulation due to calcium influx, giving rise to “evoked” fusion. In the context GS-1101 in vivo of this theory, there would be no differences in the identity of the evoked and spontaneous vesicles except for the circumstances under which they happened to have fused. Although numerous studies support this hypothesis (Groemer and Klingauf, 2007, Hua et al., 2010 and Wilhelm et al., 2010), equally numerous experiments indicate that evoked and spontaneous vesicles form nonoverlapping pools with potentially different molecular signatures (Chung et al., 2010, Fredj and Burrone, 2009, Hua et al., 2011 and Sara et al., 2005). Despite the differing and sometimes contradictory conclusions

drawn from the previous studies, all of

them have primarily focused on the characterization of vesicle properties based upon the bulk dynamics of exo- and endocytosis, such as the kinetics of styryl (FM) dye destaining or changes in pHluorin fluorescence upon 3-mercaptopyruvate sulfurtransferase stimulation. Here, we sought to address this controversy by taking a different route toward understanding the properties of spontaneous and evoked vesicles. In particular, we performed nanometer-precision tracking of individual spontaneous and evoked vesicles in order to investigate whether these two functionally different vesicle categories could also be distinguished by their motional behavior. To reliably detect the position of a single fluorescently labeled vesicle, we implemented an approach similar in principles to the proven technique of fluorescence imaging with one nanometer accuracy (FIONA) (Yildiz et al., 2003), which has been demonstrated for other systems. Our strategy was first to use a new variant of FM dye, SGC5, which was previously shown to have similar lipid-binding properties as FM1-43 but has several-fold brighter fluorescence (Wu et al., 2009). The consequently high signal-to-noise ratio allowed the individual stained vesicles to be clearly distinguished above the background (Figures 1B and 1C; see also Figures S1A–S1C available online). Next, we ensured sparse labeling of vesicles.