Med Lasers Sci 2001, 16:149158 2001 Springer-Verlag London Limited
REVIEW Imaging Transmitter Release. I. Peeking at the Steps Preceding Membrane Fusion
M. Oheim
Department of Molecular Biology of Neuronal Signals, Max-Planck Institute for Experimental Medicine, Gottingen, Germany
Abstract. Over the recent year we have witnessed considerable advances in the study of neurotransmitter release. This progress has been severalfold as di#erent techniques have allowed us to characterise many steps along the process of exocytosis, membrane fusion, formation of the fusion pore, and have given insight in the kinetics of release and membrane re-uptake. Patch clamping provided quantitative measurements of the capacitance changes as the membrane of the secretory vesicle is added to the surface of the cell during secretion, and the change in the opposite direction when membrane is retrieved back into the cell during exocytosis. Carbon-fibre microelectrodes have measured electrochemically the release of oxidisable transmitters into the extracellular space. Di#erential interference contrast microscopy has given us spatially resolved images of the cell surface during exocytosis; real-time images that are suggestive of bubbles breaking the surface of a boiling pot of water. The interest in novel techniques stems from the fact that existing approaches can provide only indirect evidence on the steps preceding membrane fusion. The vesicular dynamics just beneath the plasma membrane are out of the reach of capacitance measurements or amperometric detection. What we have needed is a tool that would allow us to look just below the cell surface. This much-needed tool appears to be evanescent-wave microscopy. This review describes how laser microscopy can be used to study exocytosis at single-vesicle resolution. A companion paper deals with the practical aspects of evanescent-wave imaging. Keywords: Evanescent-wave microscopy; Exocytosis; Fluorescence; Total internal reflection; Vesicle dynamics
INTRODUCTION Synaptic transmission relies on the release of vesicles, membrane-delimited transport organelles that are filled with transmitter and shuttled to their fusion sites [1,2]. On elevation of the near-membrane intracellular free calcium concentration ([Ca2+ ]i) the vesicle fuses with the target membrane and transmitters, peptides or hormones are liberated into the extracellular space (Fig. 1). The regulated release of vesicles involves a tightly controlled sequence of events, including the sorting, packaging and shuttling of transmitter substances. Vesicles are directed to specific sites at the plasma membrane where they
Correspondence to: Dr M. Oheim, Laboratoire de Neurophysiologie et Nouvelles Microscopies, INSERM Epi00-02, Ecole Superieure de Physique et Chimie Industrielles ESPCI, 10 rue Vauquelin, F-75005 Paris, France. Tel.: [phone], Fax: [phone], E-mail: [e-mail]
dock and mature through a sequence of biochemical activation steps before they are released [3,4]. The comprehension of the underlying molecular processes would benefit greatly from a method that allows the study of the kinetics and the amount of release in isolation from the response of the postsynaptic target cell. Synaptic terminals, by virtue of their small size have proved di$cult to study directly. Exceptions are the giant synaptic nerve terminals [5,8], that are accessible to experimental manipulation. Neuroendocrine cells like the chroma$n cell [9] of the adrenal medulla or the pituitary cell [10] have been used as `models' for the presynaptic terminal. These hormone and peptide-releasing cells are about ten times bigger than synapses and are easily maintained in culture. They are densely packed with 25 00030 000 granules (see Fig. 2) that are 10-fold larger in diameter than synaptic vesicles. Neuroendocrine cells are lacking a post-`synaptic' target and, upon
150
M. Oheim
Fig. 1. Release of transmitter and hormones by regulated exocytosis. The diagram shows the sequence of events that leads to the liberation of substances into the extracellular space. Fusion of vesicles causes a transient increase in membrane area that is counterbalanced by the subsequent re-uptake of the emptied vesicle. Depending on the experimental conditions, the time course of secretion can be governed by one of the earlier steps of the sequential reaction rather than the final fusion reaction (`rate-limiting step').
Fig. 3. Single-vesicle exocytosis studied with membranecapacitance, amperometry and optical techniques. Several experimental approaches have played key roles in monitoring secretion at the single-vesicle level: the patch clamp technique measuring transmembrane ionic currents and patchclamp cell-capacitance measurements, its variant for monitoring changes in membrane surface area, electrochemical techniques directly detecting the release of transmitters and optical methods for imaging fluorescently stained membranes or vesicles. See main text for reviews of the different aspects of these techniques.
MONITORING SECRETION IN SINGLE LIVE CELLS The patch-clamp technique [11] provides an electrical measure of the surface area of a cell by tracing changes in the membrane capacitance [12]. Electrochemically, the release of oxidisable molecules can be estimated by measuring their oxidation or reduction at the surface of carbon-fibre microelectrodes placed in vicinity of the sites of release [13,14]. More recently, optical techniques were used to directly visualise exocytosis and endocytosis [1519].
Fig. 2. Electron freeze-fracture micrograph of a bovine chromaffin cell. An isolated chromaffin cell was shock-frozen and subsequently fractured. Application of a thin film of platinum and carbon to the surface results in a replica of the cell's surface topography that can be studied under the electron microscope (EM). The total magnification is 15 300, the cell's diameter in the order of 12 m. Chromaffin cells contain 25 00030 000 300-nm large granules, membranedelimited storage organelles that are filled with adrenaline, noradrenaline and a variety of peptides and proteins, concentrated in an intragranular storage-matrix or dense core. The term `chromaffin' was coined to highlight the cell's property to be easily stained by chrome salts. The dark oval structure is the nucleus. The bubble-like structures represent individual chromaffin granules, cut at different heights (photograph by Wolfgang Schmitt, University of Innsbruck, Austria).
Membrane Capacitance Measurements Membrane capacitance is proportional to the cell's surface area. Vesicle fusion and re-uptake during exocytosis and endocytosis lead to an increase or decrease in the membrane area and, correspondingly, in-cell surface capacitance (Cm). Patch-clamp measurements of secretion rely on these changes in Cm, by determining the parameters of the equivalent circuit (Fig. 3) in response to AC voltages (see ref. [20] for review).
stimulation, release their substances into capillaries and blood vessels. The study of neuroendocrine secretion necessitates a detector for hormone or peptide release.
Imaging Transmitter Release
151
Results from Electrophysiological Studies Whole-cell techniques provide single-cell assays of exocytosis at single-granule resolution with millisecond-temporal resolution. In excitable cells, voltage-clamp depolarisations of the membrane potential can be used to activate voltage-dependent Ca2+ -channels and thereby stimulate secretion. Changing the pulse duration allows one to adjust the amount of Ca2+ injected by each depolarisation. This approach has been used to establish a correlation between the Ca2+ -current integral and the secretory response [21,22], indicating that the binding of three to four Ca2+ ions is required to trigger the release of a vesicle. Repetitive depolarisations lead to depression of the secretory response that, in the absence of further stimulation, recovers later on (measured by application of second pulse train after a variable time delay). This finding has been interpreted as the depletion and refilling of a small population of granules, in a docked and release-ready state in very close vicinity of the plasma membrane (see, e.g. ref. [23] for review). The Cm technique equally provides estimates for the conductance and time-courseof-expansion of the fusion pore, a water-filled channel that initiates the connection between vesicle interior and the cell exterior [24]. Limitations of Electrophysiological Studies Estimates of Cm depend on the equivalent electrical model circuit and rely on the accuracy of the electrical parameters, which assume spherical cell-models. When endocytic uptake of membrane sets in, Cm measurements tend to underestimate the amount of secretion. Capacitance changes can result from other factors than membrane addition and retrieval, e.g. charge mobilisation in voltage-gated ion channels or membrane ru%ing. Another limitation of the Cm technique is that it does not provide much information on the kinetics of the individual vesicle event. The subvesicular details of release have been studied with carbon-fibre microelectrodes. Electrochemical Detection of Released Molecules The oxidation or reduction of transmitter substances on the surface of microelectrodes placed in the close vicinity of the cell allows sensitive measurements of secretion [25]. The oxidation current is a direct measure for the
release of oxidisable molecules. Being directly related to the number of released molecules, electrochemical measurements are not modeldependent nor are they susceptible to interference from endocytosis. The recent observation that many cell types that are bathed in high concentrations of serotonin take up this transmitter which is consequently co-released when exocytosis is stimulated, make amperometric detection feasible for a larger variety of cell types that do not endogenously release oxidisable molecules (W. H. Betz, personal communication). Results from Electrochemical Studies Stimulation of the cell leads to a shower of spiking oxidative transients. These current traces often display spikes preceded by a slow pedestal that is smaller in magnitude. Sometimes, `stand-alone' events of this type are observed. The slower pedestal signals have been interpreted as the rate-limiting e#usion of molecules through the early fusion pore [26,27]. The `stand-alone' foot event could be related to conductance flickering, a rapid succession of open/closed transitions, of the fusion pore in Cm measurements [28]. Indeed, the complete collapse of the vesicle into the plasma membrane may not be necessary. The formation of a small pore allows the entire contents to be released and the vesicle membrane to be recycled without ever undergoing full fusion. A recent study on rat chroma$n cells provided evidence that high [Ca2+ ]i can shift the mode of exocytosis from the `classical' mode to a `kissand-run' mechanism [29]. Another result from electrochemical measurements of secretion has been that the time-course of release is much slower than expected for free e#usion of transmitter from the vesicle [14], and is influenced by intracellular pH and [Ca2+ ]i [30]. Granules are known to contain an intravesicular storage matrix to accumulate transmitter at high concentrations. Upon opening of the fusion pore, pH changes may be required to dissolve this `sponge' for the granular contents to be completely released. Similar findings of a postfusion regulation of secretion have recently been reported in lactotrophs [31]. Limitations of Electrochemical Studies As a consequence of the separation of the electrode tip and the site of release (Fig. 3), the amperometric signal is temporally `filtered' by di#usion. The functional dependence of the
152
M. Oheim
current response on the separation distance has been used together with a moving microelectrode inside a patch pipette [32] to deconvolve the kinetics of di#usion and release. As the electrode only covers a small fraction of the cell's surface area the oxidation current may not be representative of secretion elsewhere. The relatively large tip diameter of conventional carbon fibres (up to 10 m) makes it di$cult to relate the measurement to the topography of individual release sites. For the same reason, it has been di$cult to probe synaptic release (see [33]). This drawback can be used to advantage to probe the secretory response with some spatial resolution: scanning 2- m diameter carbon-fibre electrodes over the surface of the cell, areas of enhanced secretion were detected [34]. Access to the Prefusion Steps Has Been Difficult Despite considerable progress in understanding the mechanisms and molecular basis of membrane fusion and release, much less is known about the transport of granules and vesicles to their docking and fusion sites at the plasma membrane and the intracellular maturation of the release-steady state. This is because the intracellular steps preceding fusion and release have been largely inaccessible in live cells. Electrophysiological studies detect membrane addition or release, and thus provide only indirect evidence on earlier steps in the cascade of events, for example by the exhaustion of the secretory response upon successive membrane depolarisation (see [35] for review). Uniform rapid Ca2+ -elevation by photochemical release has revealed di#erent kinetic phases of the secretory response [3638]. These phases have been interpreted by the existence of large clusters of vesicles in apposition to the site of membrane fusion that are sequentially mobilised upon stimulation of exocytosis to sustain the rate of release during repetitive stimulation (see [23,39] for review). Upon maintained stimulation, these reserve pools of vesicles get successively depleted, and the rate of release decreases. Recent optical studies have begun to functionally identify these intermediates. Optical Methods to Study Secretion: The Contribution of Laser Microscopy Imaging individual vesicles en route for release is not an easy task. The microscope's objective
spreads out their image so that the individual vesicle is not a point of light, but a threedimensional distribution of light known as the point-spread function. Two neighbouring vesicles are resolved if the overlap between their images is not too big. Whether they are seen as one or are resolved as two depends on the image contrast, i.e. on the magnitude of the intensity saddle between the two maxima relative to the peak intensity, and the noise present in the measurement. Above and below the focal plane this pattern is increasingly blurred, its envelope resembling a double cone of light. The central intensity-maximum of this three-dimensional distribution of light has the shape of a cigar with its long axis along the optical axis of the microscope. Otherwise stated, the microscope laterally resolves vesicles that are separated by a distance in the order of the wavelength of light and axially integrates over micrometre distances. Although it is possible to `see' individual vesicles, vesicles of 30 nm and even 300-nm granules are only resolved if they are sufficiently remote from each other. From electron micrographs it is known that the density of vesicles and granules is much higher than that needed for single-vesicle resolution (Fig. 2). The task is thus to pick out a few points of light from a hazy background. In terms of the optics involved, this can be done by either reducing the volume of fluorescence readout or confining the volume of fluorescence excitation (Fig. 4AC). Another practical approach is to label only a few vesicles out of many, e.g. after endocyte uptake of a fluorescent marker [40,41] (Fig. 4D) or to work on isolated granules [42].
Imaging Single Vesicles Using Confocal Detection Individual vesicle events have been imaged using confocal laser scanning microscopy in neuroendocrine cells [43,44] and in hippocampal neurons [40,45]. Several drawbacks of confocal microscopy have limited its popularity for vesicle tracking (Fig. 4A). First the thickness of the confocal section cannot be any better than 1/2 times the axial resolution of the microscope [46], and is still in the micrometre range. Thus, on a length-scale important for local Ca2+ -signalling and for the approach of vesicles to their docking sites confocal imaging is insu$cient. More critically, in the choice of the dwell-time-per-pixel the
Imaging Transmitter Release
153
Fig. 4. Visualising individual-vesicle events. (A) Two dyes, a vesicle dye (1) and a Ca2+ indicator (2) serve as markers for vesicle positions and the intracellular free calcium concentration, [Ca2+ ]i. Due to its limited axial resolution, wide-field microscopy does not resolve single vesicles out of a blur of fluorescence. While the out-of-focus signal is suppressed when using confocal detection schemes, photobleaching and -damage are still limiting the recording of long image series which are advantageous for tracking vesicle movement. (B) Two-photon absorption confines the fluorescence excitation to a small focal region and reduces photodamage, but requires scanning of the spot through the specimen to reconstruct a 3-D image. (C) Evanescent-wave microscopy selectively excites dye molecules fine excitation to small volumes, at or near an interface. The near-field decays exponentially within 100 nm from the interface. Although providing a powerful means for the investigation of near-membrane events, this method precludes a deeper look into the cell. (D) Unlike optical techniques that seek to confine excitation light, amphiphilic (fm-) dyes have been used to confine dye loading to only a few vesicles. With fewer vesicles labelled, individual events translate into larger relative fluorescence changes.
experimenter has to trade o# the time resolution against signal. During the image scan, the hourglass-shaped double cone of excitation light illuminates a large volume of the specimen, but only photons emerging from the confocal plane are detected and signal photons scattered during their way out of the specimen are equally rejected by the pinhole. In addition, the ine$cient use of excitation light in confocal microscopes requires either (a) to image at low excitation intensity but also low frame rate to integrate su$cient photons per pixel or (b) to tolerate relatively high photodamage per signal photon.
Multiphoton Excitation of Vesicular Fluorescence The risk of photodamage is reduced when using multiphoton excitation (Fig. 4B). Two [47] or three photons [16] of roughly twice or threefold, respectively, the wavelength needed for one-photon excitation combine to
excite the fluorophore. But similar to confocal detection, the scanning requirement and integration over a relatively large axial optical section limit its use for imaging dynamic single-vesicle events. Multifocal microscopy [48,49], resonant galvanometric scanning [50], bilateral scanning [51], or a supercharging technique for driving the galvanometric mirrors [52] have overcome the temporal limitations of scanning microscopy. An alternative is selectively to illuminate a laterally extended but axially confined optical section just beneath the plasma membrane (Fig. 4C). In this case, wide-field collection of fluorescence can be used and, due to the confinement of excitation, no unwanted fluorescence is excited in cytosolic regions distant from the membrane.
Evanescent-field Microscopy This technique is based on the generation of a decaying near field close to the cellsubstrate
154
M. Oheim
interface. When light hits an interface between two media of di#erent refractive index light is refracted and reflected. When the angle of incidence exceeds a certain critical angle, total internal reflection occurs so that virtually all light bounces o# the interface and is reflected back into the denser optical medium. But, total internal reflection of light also sets up a decaying near field just beneath the glass substrate. This near-field component, termed the evanescent wave (EW), is, and can be, often ignored since it is not seen on the micrometre-length scales typically relevant in far-field light microscopy. The intensity maximum of the EW is located at the reflecting interface. In the axial direction, the EW field intensity decays, roughly as a falling exponential function, within less than an optical wavelength [53] so that fluorescence changes magnify displacements of fluorophores in the axial direction. If cells are grown on the glass, only the bottom of the cell is illuminated. Thus, even if all vesicles are labelled, only those that are su$ciently close to the reflecting interface to be illuminated by the EW are seen in the image (Fig. 5A). The distribution of excitation light is only determined by the wavelength, polarisation, the incidence angle and diameter of the light, and is not determined by the di#raction limited objective lens. It is thus fundamentally di#erent from the hourglass shape in wide-field illumination modes. The confinement of fluorescence excitation to the near-membrane region eliminates out-offocus fluorescence present in conventional fluorescence excitation. Although the generation of EWs by total internal reflection has long been known its use for imaging purposes is fairly recent [54,55]. Axelrod and co-workers pioneered EW-microscopy for biological applications [56]. A quantitative use of EW microscopy started in the mid-1980s [5760]. As applied to cells cultured on glass substrates, EW microscopy allows the selective visualisation of cell/substrate contact regions [56]. It can be used to visualise the position, extent and motion of these contact regions [6163], or to determine the dynamics [64], concentrations and kinetics [65,66] of fluorophores close to the interface. EW excitation allows sensitive highresolution in the near-membrane region of live cells and has been used to image single molecular interactions with a dielectric interface [67], single kinesin moving molecules along microtubules, and actin/myosin interactions [68].
Fig. 5. Principle of evanescent-wave excitation. (A) A light beam is directed to a dielectric interface. Total internal reflection sets up a decaying near-field that excites fluorophores, e.g. dye-loaded chromaffin granules, in cells grown on glass coverslips. Only granules close to the reflecting interface light up (clear spots), whereas those located deeper in the cytoplasm are not excited by the evanescent field (grey shading; drawing not to scale). (B) Evanescent-wave image at a calculated penetration depth of the evanescent wave of 120 nm. Individual vesicles show up as bright fluorescent spots in front of a dim background that is probably due to scattered excitation light and unspecific staining. The scale bar is 1 m. ( Springer-Verlag, Eur. Biophys. J. (2000).)
The introduction of EW microscopy to studies of secretion [17,69,70] has made possible the direct observation of individual dye-loaded granule trajectories in live neuroendocrine cells [7173] (Fig. 5B). Recently, EW-imaging of synaptic vesicles at a ribbon-type synapse has been reported [41].
Results from Optical Studies of Secretion Di#erential-interference contrast microscopy yields image series of chroma$n granules that are reminiscent of bubbles of boiling water [74]. Yet, these images are hard to quantitate so that most authors have preferred to use fluorescent labelling. Specific labelling of synaptic vesicles and granules has been achieved using a variety of techniques. In many types of secretory cells, vesicles contain ATP in addition to the specific transmitters or hormones. The ATP-dependent luciferase-catalysed luminescent oxidation of luciferin can be used for a photometric assay of ATP secretion [75]. The same mechanism was more recently used to generate fusion proteins of luciferase and the vesicle protein synaptotagmin-I or VAMP-2/synaptobrevin. Upon membrane fusion, these `synaptolucins' form light-emitting complexes with their cognate luciferin, when added to the extracellular liquid. Real-time optical monitoring of synaptic vesicle recycling has become possible with the advent of styryl dyes [15]. The amphiphilic
Imaging Transmitter Release
155
dye molecules are water-soluble but preferentially partition into lipid membranes where they fluoresce intensely. Upon stimulation of exocytosis and subsequent wash of fm from the extracellular membrane, fm is trapped in synaptic vesicles. Further stimulation triggers vesicle cycling and causes preloaded vesicles to release their dye into the bathing medium, so that synaptic fluorescence declines [44,76]. Measurements of fluorescence intensity during de-staining, especially when combined with electrophysiological recordings, allow the kinetics of vesicle recycling to be derived. A quantitative analysis of fm-dye fluorescence was used to study the integrity of the vesicle through the endocytic limb of the cycle [77]. Although fm-dyes are somewhat notorious for non-specific staining of lipidic debris [78] imaging of synaptic activity in intact brain and brain slices has been demonstrated when fm was used together with a cyclodextrin-derivate [79]. Several cationic monoamines or diamines, among these the acidotropic dyes acridine orange (AO) [80], quinacrine, and the dyes of the Lysotracker/LysoSensor families, move freely across membranes in their unprotonated form, and accumulate on the side of the membrane with lower pH after protonation. Some of these dyes respond to pH-gradients by changing absorbency and emission properties. The pH-dependent accumulation of dye molecules renders these acidotropic dyes suitable for labelling acidic compartments, including lysosomes, granules and vesicles. Using acidotropic dyes, the restriction of granule movement in a plane close to the plasma membrane has been demonstrated [7173]. Upon Ca2+ -influx through voltage-activated ion channels, individual vesicles are seen to liberate their contents in a cloud of released dye molecules [17,20]. Stimulation of exocytosis causes an increase in mobility of a previously virtually immobile population of vesicles [7173]. Among the disadvantages of these dyes are their low specificity, and unknown final concentration. The complicated photochemistry of acridine orange makes a quantitative interpretation of fluorescence intensities somewhat di$cult. Genetically encoded fluorophores o#er control over some of these parameters. Fusion proteins of variants of the green fluorescent protein (GFP) with vesicle proteins or proteins involved in regulating the vesicle life cycle
constitute a powerful tool to specifically label the secretory organelles or fusion sites. Examples include GFP-chromogranin A or GFP-neuropeptide-Y [69,81]. GFP has been used for the observation of vesicle movement [43], protein translocation [82], or pH changes during exocytosis. The simultaneous use of spectrally di#erent GFP-variant pairs, e.g. enhanced eGFP and blue (BFP) [83], or cyan (CFP) and yellow (YFP) [84,85] protein, linked to di#erent vesicle and membrane proteins, o#ers the possibility of fluorescence resonance energy transfer (FRET) measurements to study reactions on a molecular scale and probe molecular interactions. Already, GFP and its mutants have been extremely powerful to gain insight into the mechanisms underlying the spatial organisation and dynamics of vesicles on their way from the endoplasmic reticulum (ER)/Golgi system to their target membranes [8587]. Using GFP-fusion proteins, it has been demonstrated that actin tails propel endocytosed vesicles into the cytoplasm [88]. In line with these findings, the manipulation of the actin-based meshwork has recently been shown to a#ect vesicle mobility and regulate the rate of secretion and membrane re-uptake [72,89].
Limitations of Optical Studies Imaging techniques have proven very powerful for our understanding of cell biology, but few approaches can provide the resolution and signal-to-background ratio required to quantify unitary events in secretion control. Examples include the di$culty of precisely mapping dynamic single vesicle events [42] or the localised spatio-temporal Ca2+ -profiles (`microdomains') at the mouth of individual ion channels that are potentially associated with a small cluster of release sites [90]. Resolving individual synaptic vesicles out of a crowd is limited by the fundamental laws of di#raction so that the spatial confinement of fluorescence excitation, the use of high-NA optics for detection [41,91] and labelling a few vesicles instead of all have to go together. The millisecond time-scale typical for synaptic release and the rapid turnover of large clusters of vesicles, together with the available signal make that the signal is at the limits of detection of present imaging systems. Equally, the use of imaging techniques has been limited for experiments that address problems of subvesicular dimension, e.g. partial release [92] or
156
M. Oheim 2. DelCastillo J, Katz B. Quantal components of the end-plate potential. J Physiol (Lond.) 1954;124:56073. 3. Sudhof TC. The synaptic vesicle cycle: a cascade of protein-protein interactions. Nature 1995;375:64553. 4. Rothman JE, Sollner TH. Throttles and dampers: con trolling the engine of membrane fusion. Science 1997;276:121213. 5. Borst JGG, Helmchen F, Sakmann B. Pre- and postsynaptic whole-cell recordings in the medial nucleus of the trapezoid body of the rat. J Physiol 1995;489: 82540. 6. Hori T, Takai Y, Takahashi T. Presynaptic mechanism for phorbol-ester induced synaptic potentiation. J Neurosci 1999;19:72627. 7. Llinas R, Steinberg I, Walton K. Relationship between presynaptic calcium current and postsynaptic potential in squid giant synapse. Biophys J 1981;33:32351. 8. Zucker RS. The role of calcium in regulating neurotransmitter release in the squid giant synapse. In: Feigenbaum J, Hanani M (eds) Presynaptic Regulation of Neurotransmitter Release: a Handbook. London: Freund, 1991:15395. 9. Burgoyne RD, Morgan A. Regulated exocytosis. Biochem J 1993:293:30516. 10. Zorec R. Exocytosis in pituitary cells. Acta Pharmacol 1992;42:2816. 11. Hamill OP et al. Improved patch-clamp techniques for high-resolution current recording from cells and cellfree membrane patches. Pflugers Arch 1981;391: 85100. 12. Neher E, Marty A. Discrete changes in of cell membrane capacitance observed under conditions of enhanced secretion in bovine adrenal chroma$n cells. Proc Natl Acad Sci USA 1982;79:671216. 13. Wightman RM et al. Temporarily resolved catecholamine spikes correspond to single vesicle release from individual chroma$n granules. Proc Natl Acad Sci USA 1991;88:107548. 14. Chow RH, Ruden Lv, Neher E. Delay in vesicle fusion revealed by electrochemical monitoring of single secretory events in adrenal chroma$n cells. Nature 1992;356:603. 15. Betz WJ, Bewick GS. Optical analysis of synaptic vesicle recycling at the frog neuromuscular junction. Science 1992;255:200203. 16. Maiti S et al. Measuring serotonin distribution in live cells with three-photon excitation. Science 1997;275: 5302. 17. Steyer JA, Horstmann H, Almers W. Transport, docking and exocytosis of single secretory granules in live chroma$n cells. Nature 1997;388:4748. 18. Neves G, Lagnado L. The kinetics of exocytosis and endocytosis in the synaptic terminals of goldfish retinal bipolar cells. J Physiol 1999;515:181202. 19. Ryan TA et al. The kinetics of synaptic vesicle recycling measured at single presynaptic boutons. Neuron 1993;11:71324. 20. Gillis KD. Techniques for membrane capacitance measurements. In: Sakmann B, Neher E (eds) Single Channel Recording. New York: Plenum, 15597. 21. Neher E. Spatial and temporal aspects of Ca changes in secretional control. Biochem Soc Trans 1992;21:4203. 22. Augustine GJ, Neher E. Calcium requirements for secretion in bovine chroma$n cells. J Physiol 1992;450:24771.
vesicular integrity during the exo-/endocytic cycle [77,93]. A combination of evanescentwave excitation for studying vesicle location and FRET for studying super resolution or molecular interactions seems promising [50]. Finally, as the overall photon e$ciency of most microscopes has remained surprisingly low and cost-e#ective detectors often display relatively low quantum e$ciencies, photo bleaching and damage is a major concern when long image series are recorded.
CONCLUSION Both electrophysiological and electrochemical methods have greatly improved our knowledge of transmitter release. They have enabled detailed studies on the time-course and the Ca2+ regulation of neuroendocrine secretion. However, inherent to their detection of membrane merger or release, experiments have given little information on the very steps preceding membrane merger. Both techniques are not readily applicable to individual synapses. Optical studies have made possible the direct observation of vesicles prior to fusion or after exocytosis. Their application, however, has so far been limited due to high cost, insufficient resolution and cell damage during image acquisition. Evanescent-wave microscopy based on the total internal reflection of light is an elegant technique to observe small fluorescence changes near a dielectric interface. Although having been known for a long time, the technique of evanescent-wave imaging has gained popularity with cell biologists and neuroscientists only recently, partly because no commercial apparatus is currently available. To make evanescent-wave imaging more accessible to a broader readership, the following companion paper reviews theoretical and technical aspects of evanescentwave microscopy. The recent observation of, at least clusters, of few synaptic vesicles and the fluorescence changes associated with individual-vesicle events [41] seems promising for future progress in imaging transmitter release.
REFERENCES
1. Ceccarelli B, Hurlbut WP. Vesicle hypothesis of the release of quanta of acetylcholine. Physiol Rev 1980;60(2):396441.
Imaging Transmitter Release 23. Neher E. Vesicle pools and Ca2+ microdomains: New tools for understanding their roles in neurotransmitter release. Neuron 1998;20:38999. 24. Almers W. Exocytosis. Annu Rev Physiol 1990;52: 607 24. 25. Chow RH, von Ruden L. Electrochemical detection of secretion from single cells. In: Sakmann B, Neher E (eds) Single-Channel Recording. New York: Plenum, 1995;24575. 26. Neher E. Secretion without full fusion. Nature 1993;363:4978. 27. Alvarez de Toledo G, Fernandez-Chacon R, Fernandez JM. Release of secretory products during transient vesicle fusion. Nature 1993;363:5547. 28. Breckenridge L, Almers W. Currents through the fusion pore that forms during exocytosis of a secretory vesicle. Nature 1987;328:81417. 29. Ales E et al. High calcium concentrations shift the mode of exocytosis to the kiss-and-run mechanism. Nature Cell Biol 1999;1:404. 30. Jankowski JA et al. Temporal characteristics of quantal secretion of catecholamines from adrenal medullary cells. J Biol Chem 1993;268:14694700. 31. Angleson JK et al. Regulations of dense core release from neuroendocrine cells revealed by imaging single exocytic events. Nature Neurosci 1999;2:4405. 32. Albillos A et al. The exocytotic event in ...