588265
research-article2015
NROXXX10.1177/1073858415588265The NeuroscientistLeitz and Kavalali
Review
2+
Ca Dependence of Synaptic Vesicle Endocytosis
The Neuroscientist 1–13 © The Author(s) 2015 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1073858415588265 nro.sagepub.com
Jeremy Leitz1 and Ege T. Kavalali1,2
Abstract Ca2+-dependent synaptic vesicle recycling is essential for structural homeostasis of synapses and maintenance of neurotransmission. Although, the executive role of intrasynaptic Ca2+ transients in synaptic vesicle exocytosis is well established, identifying the exact role of Ca2+ in endocytosis has been difficult. In some studies, Ca2+ has been suggested as an essential trigger required to initiate synaptic vesicle retrieval, whereas others manipulating synaptic Ca2+ concentrations reported a modulatory role for Ca2+ leading to inhibition or acceleration of endocytosis. Molecular studies of synaptic vesicle endocytosis, on the other hand, have consistently focused on the roles of Ca2+calmodulin dependent phosphatase calcineurin and synaptic vesicle protein synaptotagmin as potential Ca2+ sensors for endocytosis. Most studies probing the role of Ca2+ in endocytosis have relied on measurements of synaptic vesicle retrieval after strong stimulation. Strong stimulation paradigms elicit fusion and retrieval of multiple synaptic vesicles and therefore can be affected by several factors besides the kinetics and duration of Ca2+ signals that include the number of exocytosed vesicles and accumulation of released neurotransmitters thus altering fusion and retrieval processes indirectly via retrograde signaling. Studies monitoring single synaptic vesicle endocytosis may help resolve this conundrum as in these settings the impact of Ca2+ on synaptic fusion probability can be uncoupled from its putative role on synaptic vesicle retrieval. Future experiments using these single vesicle approaches will help dissect the specific role(s) of Ca2+ and its sensors in synaptic vesicle endocytosis. Keywords synaptic transmission, synaptic vesicle recycling, FM1-43, SynaptopHluorin, endocytosis, calcium signaling
Introduction Synaptic vesicle recycling is critical for the maintenance and proper function of neurotransmission. After Ca2+ enters the synaptic terminal, it triggers synaptic vesicle fusion leading to exocytosis of neurotransmitter substances. The synapse is then faced with several problems: there are now fewer synaptic vesicles in the terminal, synaptic vesicle constituents are now present on the plasma membrane of the neuron, and the surface area of the synapse has grown by ~5000 nm2 per vesicle (assuming a vesicle radius of ~20 nm). These problems are overcome at the synapse by local retrieval of synaptic vesicles from the plasma membrane via endocytosis and ultimately, their reuse (Alabi and Tsien 2013; Kavalali 2006). Despite our growing understanding of the precise role of Ca2+ in triggering and regulation of synaptic vesicle fusion, its role in synaptic vesicle retrieval remains unclear (Yamashita 2012). Unlike exocytosis, endocytosis is largely inaccessible by electrophysiological methods in small synapses of the central nervous system (CNS) such as those of the cortex and hippocampus.
Therefore, electron microscopy and live cell imaging have proved invaluable in the study of endocytosis in the CNS. In addition to the technical challenges to directly assessing endocytosis, strong coupling of endocytic processes to prior exocytosis at the synapse hinders study of endocytosis in isolation. To maintain homeostasis of synaptic surface membrane area and integrity at steady state, the processes of exocytosis and endocytosis at synaptic terminals must balance each other. Indeed in experiments that monitor changes in membrane surface area, the extent of exocytosis appears to impact endocytosis indirectly by determining the number of synaptic vesicles and/or the extent of plasma membrane to be subsequently 1
Department of Neuroscience, UT Southwestern Medical Center, Dallas, TX, USA 2 Department of Physiology, UT Southwestern Medical Center, Dallas, TX, USA Corresponding Author: Ege T. Kavalali, UT Southwestern Medical Center, Dallas, TX 753909111, USA. Email: [email protected]
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
2
The Neuroscientist
endocytosed. This apparent balance of exocytosis and endocytosis also forms a key obstacle in studying the Ca2+ dependence of endocytosis in isolation from Ca2+’s direct role in exocytosis. Therefore, resolving any potential direct effect of Ca2+ on synaptic vesicle endocytosis has been challenging.
Early Insight into the Role of Ca2+ in Endocytosis The first indication that Ca2+ is a key component of the endocytic machinery came from electrophysiological experiments followed by post-stimulation fixation and electron microscopy (Hurlbut and Ceccarelli 1974). Using black widow spider venom, Ceccarelli and colleagues induced Ca2+-independent exocytosis and observed that in the absence of extracellular Ca2+ synaptic vesicle exocytosis caused a physical enlargement of the frog neuromuscular junction, and ultimately a rundown of neurotransmission. But when Ca2+ was included in the extracellular solution, neurotransmission could be maintained and cell swelling was reduced. This observation indicated that synaptic vesicles at the frog neuromuscular junction could only be recycled in the presence of Ca2+ (Ceccarelli and others 1979). These initial electron microscopic data established the executive role of Ca2+ in endocytosis, but electron microscopy, while unrivaled in its ability to visualize nanometer-scale structures, is difficult to apply toward the kinetics of a given process. To extract kinetic information one has to examine many preparations at different time points, and since each preparation must be fixed, it is impossible to monitor processes before and after a certain manipulation within the same cell. Therefore, methods that could be applied to live cells, and that enabled real-time measurements, would be better suited to examine kinetics. So while Heuser, Reese, and Ceccarelli were characterizing the structural intermediates of endocytosis (Ceccarelli and others 1979; Heuser and Reese 1973, 1981; Hurlbut and Ceccarelli 1974), electrophysiologists were working to refine the resolution of real-time recordings to monitor cellular capacitance changes that are proportional to the plasma membrane surface area in living cells. In 1982, Neher and Marty succeeded in this endeavor, resolving membrane capacitance changes on the order of femtofarads (10−15 farads) in adrenal chromaffin cells (Neher and Marty 1982). This allowed them to observe, in real-time, single chromaffin granule exocytosis (as an increase in cell capacitance) and endocytosis (as a decrease in capacitance). While exocytosis occurred as an abrupt increase, the observed time course of endocytosis varied from the limit of their temporal resolution (30 ms) to seconds. Moreover, from these experiments it was immediately evident that Ca2+ not only played an executive role in endocytosis but also a modulatory role on the
rate of endocytosis of large granules. Further refinement of the same technique permitted measurement at the level of small vesicles and also showed a clear Ca2+ dependence of endocytosis (Almers and Neher 1987). Yet despite the strength of electrophysiological measurements, a major limitation of this technique is the requisite use of large secretory cells—such as neuroendocrine cells—or giant synapses, thus precluding measurements of endocytosis in small glutamatergic and GABAergic synapses that make up the majority of the CNS (but see Hallermann and others 2003; Sun and Wu 2001). Therefore optical approaches became the method of choice in these cell types. Initially styryl dyes, such as FM1-43, that embed in the outer lipid leaflet of the plasma membrane but also easily dissociate and can be washed away were used in pulse-chase experiments (Betz and Bewick 1992); for review, see Kavalali and Jorgensen (2014) (Fig. 1). In these experiments, neurons were strongly stimulated in the absence of dye, once stimulation ceased, the extracellular solution was exchanged at varying time points with solution-containing dye, thus providing an indirect measure of endocytosis (Ryan and others 1993). Initially, these experiments suggested that endocytosis was independent of extracellular Ca2+ concentration (Ryan and others 1993; Ryan and others 1996). But similar experiments in drosophila using black widow spider venom as a stimulation again showed Ca2+ was required for endocytosis (Ramaswami and others 1994). Although FM dye–based measurements were successful in providing direct dynamic insight into the process of synaptic vesicle recycling, estimation of endocytosis kinetics using FM dye–based measurements suffered from lack of specificity and difficulty in repeated measurements due to increasing fluorescence background during each cycle of dye application and wash-out (Kavalali and Jorgensen 2014). These shortcomings could be overcome with the development of pHluorin, a pH-sensitive green fluorescent protein (GFP) (Miesenbock and others 1998). With pHluorin synaptic vesicle cycling could now be monitored directly and in a molecularly specific manner by targeting pHluorin to the luminal domains of synaptic vesicle proteins (Fig. 2). Each of these techniques: electron microscopy, electrophysiology, and live-cell imaging have their strengths and weaknesses, and each has in turn provided valuable, albeit at times conflicting, insights into the molecules and mechanisms governing endocytosis (Kavalali and Jorgensen 2014).
Ca2+ Accelerates Endocytosis A series of electrophysiological studies that directly manipulated internal Ca2+ by photo-uncaging in both chromaffin cells (Heinemann and others 1994; Neher and Zucker 1993) and melanotrophs (Thomas and others
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
3
Leitz and Kavalali Residual dye? Exo
Endo
Exo
Endo
Wash
FM1-43
Fluorescence
Ca2+
Figure 1. Optical detection of synaptic fusion and endocytosis using FM (Fei Mao) dyes. FM1-43 and its analogs are amphipathic dyes which fluorescence increases almost 100-fold when they are incorporated into membranes. During a typical experiment, stimulation in the presence of extracellular dye leads to uptake of dye molecules into fused vesicles followed by endocytosis. Subsequent wash out of extracellular dye uncovers fluorescent staining specific to actively recycling vesicles during the stimulation paradigm. Stimulation results in fusion of dye-loaded vesicles and loss of fluorescence trapped in synapses. This fluorescence loss is due to departitioning of the dye into aqueous solution or lateral diffusion of dye in neuronal membrane. In some circumstances, such as kiss-and-run endocytosis, endocytosed synaptic vesicle may still contain residual dye.
A
Single synaptic vesicle fusion and retrieval H+ H+
Fluorescence
Corresponding pHluorin Signal
Time
B
Multiple synaptic vesicle fusion and retrieval H+
H+
H+ H+
Fluorescence
Corresponding pHluorin Signal
Time
Figure 2. (A) SynaptopHluorin is a fusion construct of the synaptic vesicle protein synaptobrevin with a pH-sensitive enhanced green fluorescent protein (EGFP) at its C-terminal (located in the synaptic vesicle lumen). Synaptic vesicle lumen normally has an acidic pH of approximately 5.5 at which SynaptopHluorin fluorescence is quenched. When vesicles fuse, lumenal EGFP is exposed to the extracellular pH, which results in a marked increase in its fluorescence. Fluorescence signal remains elevated for the duration when synaptic vesicle lumen is exposed to extracellular pH (assuming SynaptopHluorin molecules do not diffuse away and remain clustered). During endocytosis, pHluorin fluorescence is requenched as vesicle lumen becomes acidic. Panel A depicts single vesicle exocytosis-endocytosis and the corresponding fluorescence trace. Panel B depicts simultaneous fusion and endocytosis of multiple synaptic vesicles and the corresponding fluorescence signal.
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
4
The Neuroscientist
1993; Thomas and others 1994) found endocytosis proceeded with two kinetic components: a rapid component with time constant of less than 5 seconds that dominated when internal Ca2+ concentrations were high (≥10 μM and 50 μM in melanotrophs and chromaffin cells, respectively) and a slower component that was on the order of tens of seconds to minutes at low intracellular Ca2+ concentrations. Thus in these studies, increasing intracellular Ca2+ accelerated the rate of endocytosis. In agreement with this premise, later electrophysiological studies found that low concentrations of Ca2+ inhibited endocytosis in chromaffin cells (Ales and others 1999; Artalejo and others 1995; Burgoyne 1995; Smith and Neher 1997), in pancreatic β-cells (Eliasson and others 1996), and in goldfish bipolar cells (Neves and others 2001). Interestingly in two of these studies (Eliasson and others 1996; Smith and Neher 1997) the slow compensatory mode of endocytosis was lost following dialysis of the cell while in whole cell recording configuration—an early indication that Ca2+ acts on diffusible factors to control vesicle retrieval. The Ca2+/calmodulin-dependent phosphatase calcineurin was later identified as such a possible factor controlling endocytosis (Engisch and Nowycky 1998), which may then act on dynamin (Artalejo and others 1995; Cousin and Robinson 2001; Marks and McMahon 1998) to accelerate endocytosis. Meanwhile, optical evidence supporting an acceleration of endocytosis by Ca2+ in hippocampal neurons was obtained using various styryl dyes (FM1-43, FM2-10 and FM1-84) with distinct membrane dissociation kinetics (Kavalali and others 1999; Klingauf and others 1998). FM dye-labeling experiments in drosophila using black widow spider venom as stimulation showed that Ca2+ not only triggered but also could accelerate endocytosis (Ramaswami and others 1994). In experiments with expression of vesicular proteins tagged with pH-sensitive GFP (pHluorin) (Sankaranarayanan and Ryan 2001) also found endocytosis accelerated by increasing Ca2+. Labeling vesicle recycling with sulforhodamine uptake in snake motor terminals indicated that increasing Ca2+ shifted endocytosis toward a faster rate (Teng and Wilkinson 2005). A variety of electrophysiological and optical experiments in the calyx of Held—a large synaptic terminal in the auditory brain stem amenable to direct capacitance measurements—also reported a Ca2+dependent acceleration of endocytosis (Hosoi and others 2009; Sun and others 2010; Wu and others 2005; Wu and others 2009). Additionally, these studies further supported the Ca2+/calmodulin sensitive phosphatase calcineurin (Sun and others 2010; Wu and others 2009) as a key Ca2+-sensitive component of the endocytic machinery. Although these studies agree on the general Ca2+dependent acceleration in endocytic kinetics, there was significant variation in the reported absolute rate of
endocytosis and the number of kinetic steps involved in the process (see Table 1). In some instances, these differences were attributed to cell type and in other cases, the technique employed, but these studies agreed that endocytosis is accelerated by increasing Ca2+ concentrations.
Ca2+ Decelerates Endocytosis In parallel with growing evidence that Ca2+ accelerates endocytosis, there is also substantial evidence that Ca2+ decelerates endocytosis. Just over a year after the initial studies by Neher and Zucker reported an acceleration of endocytosis by Ca2+, the first study to systematically examine the relationship of Ca2+ and endocytosis, produced the opposite finding in goldfish retinal bipolar cells that possess synaptic terminals large enough to be accessible to capacitance measurements (von Gersdorff and Matthews 1994). Here endocytosis was slowed when intracellular Ca2+ was increased—as monitored by the Ca2+ indicator Fura-2. In this configuration, endocytosis proceeded on the order of seconds and was steeply dependent on intracellular Ca2+, with a Hill coefficient of 4— strikingly similar to the Ca2+ coordination of synaptotagmin in exocytosis (Fernandez-Chacon and others 2001). In retinal bipolar cells a model using two relatively fixed kinetic components of endocytosis was proposed (Neves and Lagnado 1999) where increasing Ca2+ increased the proportion of vesicles retrieved by the slow kinetic process. Although this model agrees that endocytosis slows with increasing Ca2+, it is important to note that this contrasts with a model where the kinetics are themselves are subject to change by Ca2+. More recent studies again strongly supported the role of calmodulin in regulating the Ca2+-dependent deceleration of endocytosis in mouse retinal bipolar cells (Wan and others 2008; Wan and others 2010) and dynamin in cortical synaptosomes (Cousin and Robinson 2000). These studies were bolstered by experiments in dissociated hippocampal neurons using pHluorin-tagged synaptic vesicle proteins to monitor synaptic vesicle retrieval (Armbruster and others 2013).
Ca2+ Does Not Affect the Rate of Endocytosis Finally, in addition to positive or negative regulation of endocytosis by Ca2+ a third possibility also emerged: that synaptic vesicle recycling is independent of Ca2+ and instead controlled by the number of synaptic vesicles that undergo fusion, or exocytic load (Ryan and others 1993; Ryan and others 1996; Wu and Betz 1996). This hypothesis emerged from studies using styryl dye uptake and release to quantify the rate of endocytosis. In these experiments, at various times after stimulation, the extracellular solution
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
5
Leitz and Kavalali Table 1. Brief chronological summary of experiments investigating the role of Ca2+ in endocytosis. A variety of experimental procedures, preparations and stimulations have produced a range of kinetic parameters for endocytosis.
Year
Authors
Preparation
Measurement Method
1974 Hurlbut and Cecarelli
Frog neuromuscular Rapid freeze junction fixation followed by electron microscopy (EM) 1984 Miller and Heuser Frog neuromuscular Rapid freeze fixation junction followed by EM 1985 Torri-Tarelli and Frog neuromuscular Rapid freeze fixation others junction followed by EM 1993 Neher and Zucker Bovine adrenal Capacitance chromaffin 1993 Ryan and others
Hippocampal FM 1-43 cultures Goldfish bipolar cell Capacitance
1994 von Gerhersdorff and Matthews 1994 Thomas and others Melanotrophs
1994 Heinemann and others
Capacitance
Bovine adrenal chromaffin
1994 Ramaswami and others
Capacitance
FM 1-43
Drosophila neuromuscular junction 1995 Burgoyne Bovine adrenal chromaffin 1995 Artalejo and others Bovine adrenal chromaffin 1996 Eliasson and others Pancreatic β-cells
Capacitance
1996 Ryan and others
FM 1-43
1996
FM 1-43
1997 1998 1998 1998 1998 1999
Hippocampal cultures Wu and Betz Frog neuromuscular junction Smith and Neher Bovine adrenal chromaffin Engisch and Bovine adrenal Nowycky chromaffin Klingauf and others Hippocampal neurons Marks and Rat synaptosomes McMahon Rouze and Goldfish bipolar cell Schwartz Neves and Lagnado Goldfish bipolar cell
1999 Ales and others 2000 Cousin and Robinson 2000 Sankaranarayanan and Ryan 2001 Heidelberger 2001 Neves et al
Capacitance Capacitance
Capacitance Capacitance FM dyes FM 2-10 Capacitance and FM 4-64 Capacitance, FM
Rat chromaffin cells Capacitance and amperometry Synaptosomes FM2-10 uptake Hippocampal cultures
SynaptopHluorin (Syb2/VAMP2pHluorin) Capacitance
Mouse rod bipolar cell Goldfish bipolar cell Capacitance
Stimulation Strength Multivesicle (Black Widow spider venom)
Number of Kinetic Steps N/A
Approximate Time-Constant of Endocytosisa
Role of Ca2+
Single Vesicle?
N/A
Required
No
Required
No
Required
No
Single AP
2
Single AP
1
Ultrafast and Slow Ultrafast
Multivesicle (flash photolysis) Multivesicle (high K+) Multivesicle (evoked) Multivesicle (flash photolysis) Multivesicle (flash photolysis) Multivesicle (high K+)
1
Fast
Accelerates
No
1
Fast
Independent
No
1
Fast
Decelerates
No
2
Ultrafast and Fast
Accelerates
No
2
Ultrafast and Slow
Accelerates
No
N/A
N/A
Independent
No
2
Fast
Accelerates
No
3
Accelerates
No
2
Ultrafast and Fast Fast and Slow
Accelerates
No
1
Fast
Independent
No
1
Fast and slow
Independent
No
2
Fast
Accelerates
No
3
Ultrafast and Fast Fast
Accelerates
No
Accelerates
No
N/A
Independent
No
N/A
Both
No
Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (high K+) Multivesicle (high K+) Multivesicle (high K+) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (high K+) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked)
1
2
Fast
Decelerates
No
1
Ultrafast
Accelerates
No
Decelerates
No
N/A 1
Fast and slow
Independent
No
1
Fast
Decelerates
No
1
Ultrafast and Fast
Accelerates
No (continued)
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
6
The Neuroscientist
Table 1. (continued)
Year
Authors
Preparation
2001 Sankaranarayanan and Ryan
Hippocampal cultures
2002 Sun and others
Calyx of Held
2003 Gandhi and Stevens Hippocampal cultures 2003 Aravanis and others 2005 Teng and Wilkinson
Measurement Method
Stimulation Strength
Number of Kinetic Steps
Multivesicle SynaptopHluorin (evoked) (Syb2/VAMP2pHluorin) Capacitance (of minis) Spontaneous/ single AP Single AP SynaptopHluorin (Syb2/VAMP2pHluorin) FM 1-43 Single AP
1
Approximate Time-Constant of Endocytosisa
Role of Ca2+
Single Vesicle?
1
Fast
Accelerates
No
3
Ultrafast and Fast Fast and Slow (stranded)
Independent
Yes
Decelerates
Yes
N/A
Yes
Accelerates
No
3
Sulforhodamine 101
Multivesicle (evoked)
N/A
2005 Wu and others
Hippocampal cultures Snake neuromuscular junction Calyx of Held
Capacitance
2
Fast
Accelerates
No
2006 Granseth and others
Hippocampal cultures
SynaptophysinpHluorin (SypHy)
1
Fast
Independent
Yes
2007 Balaji and Ryan
vGlut-pHluorin
1
Fast
1
Fast
Decelerates
No
2009 Hosoi and others
Hippocampal cultures Mouse rod bipolar cell Calyx of Held
Multivesicle (evoked) Single AP/ multivesicle (evoked) Single AP
Capacitance
2
Fast
Accelerates
No
2009 Wu and others
Calyx of Held
Capacitance
2
Fast
Accelerates
No
2009 Zhu and others
Hippocampal cultures Mouse rod bipolar cell Calyx of Held and hippocampal cultures
SynaptophysinpHluorin (4x) Capacitance
2
Fast
Accelerates
Yes
1
Fast
Decelerates
No
1
Fast
Accelerates
No
1
Fast
Accelerates
No
1
Fast
Decelerates
Yes
1
Fast
Decelerates
No
1
Ultrafast
Independent
Yes
2008 Wan and others
2010 Wan and others 2010 Sun and others
2010 Yamashita and others 2011 Leitz and Kavalali 2013 Armbruster and others 2014 Leitz and Kavalali
Capacitance
Capacitance; synaptopHluorin (syb2/VAMP2pHluorin)
Calyx of Held
Capacitance
Hippocampal cultures Hippocampal neurons Hippocampal cultures
vGlut-pHluorin vGlut-pHluorin vGlut-pHluorin/ synaptophysinpHtomato (SypHTo)
Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Single AP Multivesicle (evoked) Capacitance steps/ extracellular stimulation 20 Hz Multivesicle (evoked) Single AP Multivesicle (evoked) Spontaneous
Fast
Yes
a
“Ultrafast” is a subsecond to second range, “fast” is 1 second to tens of seconds, and “slow” is tens of seconds to minutes.
was exchanged with solution containing dye. The amount of dye uptake would then quantify the amount of endocytosis. These experiments concluded that endocytosis proceeded with a t1/2 of 60 seconds (Ryan and others 1993). Additional experiments demonstrated that when extracellular solution was changed to a Ca2+-free solution, dye could still be taken up to the same degree as in Ca2+containing solution (Ryan and others 1993; Ryan and others 1996) Together, these early imaging experiments
generated the hypothesis that the rate of endocytosis was independent of Ca2+ but dependent on the number of synaptic vesicles that fused during stimulation. Similar styryl dye experiments in the frog neuromuscular junction further supported this hypothesis (Wu and Betz 1996). The assumption that endocytosis and exocytosis were temporally segregated but coupled via their mutual dependence on Ca2+ influx was supported by experiments in cortical synaptosomes (Marks and McMahon 1998). Further
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
7
Leitz and Kavalali
Capacitance = Cm
n
Capacitance = Cm + (Σ ∆Cvi) i=1
n
Capacitance = Cm - (Σ ∆Cvi) i=1
+∆Cv1 +∆Cv2
Cv1 Cv2 Cv3
-∆Cv1 -∆Cv2 -∆Cv3
+∆Cv3
Cvn
+∆Cvn
Cm
-∆Cvn
Cm
Cm
Capacitance = Cm Balanced exocytosis and endocytosis No net change in Cm
-∆Cv +∆Cv Cm
Figure 3. Capacitance measurements. Top panels: Synaptic vesicles are membranous organelles, which release their neurotransmitter content after fusion with the plasma membrane. This process can be directly monitored via capacitance measurements. Alteration in total membrane capacitance can be used as reporter of membrane fusion or endocytosis as long as the magnitude of the capacitance change is above the background noise. A critical requirement for employment of this method is direct electrical access to the secretory cellular component. This requirement significantly impairs the applicability of this method to small synapses of the central nervous system. Therefore, most capacitance measurements have been limited to secretion from cell bodies of mast cells, adrenal chromaffin cells, large pituitary nerve terminals or large presynaptic terminals such as the calyx of Held, located within the brain stem. Lower panel: However, under some circumstances when exocytosis and endocytosis processes are tightly balanced, capacitance measurements may not report a net change in membrane area despite extensive vesicle trafficking.
studies using pHluorin imaging in hippocampus cultures (Sankaranarayanan and Ryan 2000; but see Sankaranarayanan and Ryan 2001) and capacitance recordings in the calyx of Held (Sun and others 2002) have indicated that the degree of exocytosis impacts the kinetics of endocytosis. These seemingly contradictory results could be reconciled in a study performed in goldfish retinal bipolar cells where both very low and very high levels of Ca2+ produced no change in membrane capacitance, while the rate of endocytosis rose and fell as Ca2+ deviated from the extremes (Rouze and Schwartz 1998). This work highlighted the important point that capacitance measurements can only report a net change in membrane surface area (Fig. 3). Therefore no change in capacitance could be the result of either no membrane addition or subtraction, or of membrane addition and subtraction occurring simultaneously at equal rates. The simultaneous occurrence of exocytosis and endocytosis during
activity was also evident in pHluorin based measurements in hippocampal synapses where application of vacuolar ATPase inhibitors such as bafilomycin or folimycin inhibited synaptic vesicle re-acidification and revealed a prominent rapid component of endocytosis occurring in synchrony with exocytosis (Ertunc and others 2007; Fernandez-Alfonso and Ryan 2008; Sankaranarayanan and Ryan 2001). Strong coupling between exocytosis and endocytosis complicates the interpretation of most experiments because while all of these studies use a variety of cell types and experimental procedures, the one factor they have in common is the use of strong stimulation (Fig. 4). So this issue raises the question whether endocytosis can be investigated independent of exocytosis. One way to decouple these processes is to know the exact number of synaptic vesicles that participated in fusion during stimulation. To this end, several groups sought to monitor exocytosis and endocytosis of a single synaptic vesicle.
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
8
The Neuroscientist
Single vesicle exocytosis and endocytosis
Multivesicle exocytosis and endocytosis
Retrograde inhibition
Membrane infoldings
Figure 4. Single versus multivesicle exo-endocytosis. Most studies probing the role of Ca2+ in endocytosis have relied on measurements of synaptic vesicle retrieval after strong stimulation. Strong stimulation paradigms elicit fusion and retrieval of multiple synaptic vesicles. Therefore, interpretation of these experiments is confounded by several factors besides the kinetics and duration of Ca2+ signals. These include the number of exocytosed vesicles, accumulation of released neurotransmitters which all alter fusion and retrieval processes indirectly via retrograde signaling. Studies monitoring single synaptic vesicle endocytosis, on the other hand, allow better dissociation of fusion and retrieval processes as in these settings the impact of Ca2+ on synaptic fusion probability can be uncoupled from Ca2+’s putative impact on synaptic vesicle retrieval.
Endocytosis of Single Synaptic Vesicles Monitoring retrieval of single synaptic vesicles following fusion provides a clear way to uncouple regulation of synaptic vesicle fusion and retrieval. As fusion occurs in a quantal, all-or-nothing fashion, once a vesicle is fused its subsequent trajectory can be kinetically isolated from its fusion probability. The isolation of exocytosis and endocytosis kinetics is hard to achieve if a large number of vesicles fuse in rapid succession (Fig. 4). The binary nature of single vesicle fusion, therefore, eliminates the requirement to normalize the extent of endocytosis to exocytosis to isolate potential direct effects on endocytosis. Initial attempts at detecting single synaptic vesicle exo-endocytosis using capacitance measurements heavily relied on signal averaging and was confounded by susceptibility to contamination by capacitance changes
unrelated to synaptic vesicle exocytosis (Sun and others 2002; Yamashita and others 2005). Therefore, to resolve vesicle recycling at the single vesicle level optical techniques again became the method of choice. However, to employ optical techniques in hippocampal neurons, several approaches were developed to overcome the low signal elicited by single synaptic vesicle fusion. Experiments sought to minimize background fluorescence by pre-photobleaching—which damages the cells—or fluorescent trace averaging, that occludes subtle differences between fusion events and assumes that the components of vesicle cycling (i.e., the kinetics the amplitude) are normally distributed (Gandhi and Stevens 2003; Granseth and others 2006). Another complication in optical experiments is the lack of accessible and practical postsynaptic confirmation of synaptic vesicle fusion (but see Leitz and Kavalali 2014; Richards 2009). Therefore, measurements at single vesicle resolution rely on several corollary indicators for validation: such as vesicle fusion probability distributions and quantal fluorescence amplitude distributions. Both FM dyes and pHluorin-tagged proteins have been applied toward single vesicle imaging, but each study that utilized pHluorin yielded different endocytic kinetics depending on which synaptic vesicle protein was tagged with pHluorin. First, synaptobrevin-pHluorin (SynaptopHluorin) revealed three distinct kinetic components of endocytosis: a fast mode consistent with “kiss-and-run” kinetics (with decay time of less than 3 seconds), a slower mode (with decay time in the tens of seconds) and a stranded mode where fluorescence did not decay (Gandhi and Stevens 2003). Although the authors did not investigate directly the effect of Ca2+ on endocytosis, they did show that in higher Ca2+ the rapid component of decay was absent, suggesting that as Ca2+ increases, endocytosis slows. In these experiments, single vesicle resolution was determined by a recapitulation of vesicle fusion probability observed in classical literature, as well as quantal amplitude distribution analysis in elevated Ca2+ concentration. Later, other experiments supported a predominance of rapid kiss-andrun exo-endocytosis at low intracellular Ca2+ as determined by FM dye destaining kinetics at the single vesicle level (Aravanis and others 2003; Richards and others 2005). Another study took advantage of both SynaptopHluorin and pHluorin fused to the luminal domain of synaptophysin and reported a single slow form decay time-constant in the range of 14 to 20 seconds extracted from signal averaging the fluorescence changes from several putative single vesicle fusion events when a stimulation was applied (Granseth and others 2006). Furthermore, directly comparing normalized average fluorescence traces of single vesicle events with high frequency stimulation produced nearly identical decay traces leading to the conclusion that the rate of endocytosis might be constant
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
9
Leitz and Kavalali (Granseth and others 2006). A subsequent study took advantage of pHluorin fused to the vesicular glutamate transporter1 (vGlut-pHluorin), which showed significantly reduced background fluorescence and reported a similar single stochastic kinetic component of endocytosis with average time constant of ~14 seconds (Balaji and Ryan 2007). In these experiments, single vesicle resolution was verified by the monophasic distributions of event amplitudes used for analysis. The same group later monitored the rate of endocytosis while varying stimulation frequency or Ca2+ concentration (Balaji and others 2008). Increasing stimulation frequency or Ca2+ concentration slowed endocytosis during 10-Hz stimulation but at the single vesicle level, stimulation frequency had no effect, leading to the conclusion that the exocytic load — the number of vesicles fused—determines the endocytic rate. While vesicle fusion probability was not explicitly examined in either of these studies, examination of example traces suggests a very high release probability in these synapses (Balaji and Ryan 2007; Balaji and others 2008). A subsequent study that utilized an improved synaptophysin-pHluorin construct with four pHluorin tags to increase signal-to-noise ratio identified two kinetic components to endocytosis; one fast (~3 seconds) and one slow (~10 seconds) (Zhu and others 2009). Here, increasing Ca2+ decreased the dwell time of fluorescence—indicating shorter residence of synaptophysin-pHluorin molecules at the surface membrane on fusion—while the number of vesicles that fused increased the decay time of only the slow component of endocytosis, the fast rate was unchanged. Together, this slowing of decay time and decrease in dwell time resulted in a net acceleration of endocytosis and led the authors to propose a model where two vesicles are retrieved after one vesicle fuses. These experiments did not show synaptic vesicle fusion probability but instead relied on quantal amplitude distributions as validation of single vesicle resolution. Some support for this two vesicle model could be found in recent “flash-and-freeze” electron microscopy data that shows the ultrafast internalization (less than 100 ms) of large compartments with surface area four times that of a synaptic vesicle (Watanabe and others 2013). However, it is not known if these “compensatory” endocytic events are the product of vesicles that immediately underwent fusion, previously underwent fusion (stranded vesicles retrieved by subsequent rounds of stimulation as previously demonstrated by Thomas and others (1994)) or instead reflect retrieval of surface membrane and not fused synaptic vesicles. Although each of the above studies has visualized synaptic vesicle fusion at the single vesicle level, they produced different estimates for the rate of single synaptic vesicle endocytosis. Therefore, in recent work we sought
to systematically investigate the role of Ca2+ in endocytosis at the single vesicle level (Leitz and Kavalali 2011, 2014). Using vGlut-pHluorin to monitor single vesicle fusion we generated quantal amplitude distributions to validate single-vesicle resolution. In this setting, we found that in near physiological (2 mM) extracellular Ca2+, fluorescence decay after single action potential stimulation was a rapid process well described by a single exponential with average decay time of ~3 seconds that proceeded immediately after fluorescence increase, although some fluorescence traces did manifest a detectable dwell time likely indicating a surface residency period of
research-article2015
NROXXX10.1177/1073858415588265The NeuroscientistLeitz and Kavalali
Review
2+
Ca Dependence of Synaptic Vesicle Endocytosis
The Neuroscientist 1–13 © The Author(s) 2015 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1073858415588265 nro.sagepub.com
Jeremy Leitz1 and Ege T. Kavalali1,2
Abstract Ca2+-dependent synaptic vesicle recycling is essential for structural homeostasis of synapses and maintenance of neurotransmission. Although, the executive role of intrasynaptic Ca2+ transients in synaptic vesicle exocytosis is well established, identifying the exact role of Ca2+ in endocytosis has been difficult. In some studies, Ca2+ has been suggested as an essential trigger required to initiate synaptic vesicle retrieval, whereas others manipulating synaptic Ca2+ concentrations reported a modulatory role for Ca2+ leading to inhibition or acceleration of endocytosis. Molecular studies of synaptic vesicle endocytosis, on the other hand, have consistently focused on the roles of Ca2+calmodulin dependent phosphatase calcineurin and synaptic vesicle protein synaptotagmin as potential Ca2+ sensors for endocytosis. Most studies probing the role of Ca2+ in endocytosis have relied on measurements of synaptic vesicle retrieval after strong stimulation. Strong stimulation paradigms elicit fusion and retrieval of multiple synaptic vesicles and therefore can be affected by several factors besides the kinetics and duration of Ca2+ signals that include the number of exocytosed vesicles and accumulation of released neurotransmitters thus altering fusion and retrieval processes indirectly via retrograde signaling. Studies monitoring single synaptic vesicle endocytosis may help resolve this conundrum as in these settings the impact of Ca2+ on synaptic fusion probability can be uncoupled from its putative role on synaptic vesicle retrieval. Future experiments using these single vesicle approaches will help dissect the specific role(s) of Ca2+ and its sensors in synaptic vesicle endocytosis. Keywords synaptic transmission, synaptic vesicle recycling, FM1-43, SynaptopHluorin, endocytosis, calcium signaling
Introduction Synaptic vesicle recycling is critical for the maintenance and proper function of neurotransmission. After Ca2+ enters the synaptic terminal, it triggers synaptic vesicle fusion leading to exocytosis of neurotransmitter substances. The synapse is then faced with several problems: there are now fewer synaptic vesicles in the terminal, synaptic vesicle constituents are now present on the plasma membrane of the neuron, and the surface area of the synapse has grown by ~5000 nm2 per vesicle (assuming a vesicle radius of ~20 nm). These problems are overcome at the synapse by local retrieval of synaptic vesicles from the plasma membrane via endocytosis and ultimately, their reuse (Alabi and Tsien 2013; Kavalali 2006). Despite our growing understanding of the precise role of Ca2+ in triggering and regulation of synaptic vesicle fusion, its role in synaptic vesicle retrieval remains unclear (Yamashita 2012). Unlike exocytosis, endocytosis is largely inaccessible by electrophysiological methods in small synapses of the central nervous system (CNS) such as those of the cortex and hippocampus.
Therefore, electron microscopy and live cell imaging have proved invaluable in the study of endocytosis in the CNS. In addition to the technical challenges to directly assessing endocytosis, strong coupling of endocytic processes to prior exocytosis at the synapse hinders study of endocytosis in isolation. To maintain homeostasis of synaptic surface membrane area and integrity at steady state, the processes of exocytosis and endocytosis at synaptic terminals must balance each other. Indeed in experiments that monitor changes in membrane surface area, the extent of exocytosis appears to impact endocytosis indirectly by determining the number of synaptic vesicles and/or the extent of plasma membrane to be subsequently 1
Department of Neuroscience, UT Southwestern Medical Center, Dallas, TX, USA 2 Department of Physiology, UT Southwestern Medical Center, Dallas, TX, USA Corresponding Author: Ege T. Kavalali, UT Southwestern Medical Center, Dallas, TX 753909111, USA. Email: [email protected]
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
2
The Neuroscientist
endocytosed. This apparent balance of exocytosis and endocytosis also forms a key obstacle in studying the Ca2+ dependence of endocytosis in isolation from Ca2+’s direct role in exocytosis. Therefore, resolving any potential direct effect of Ca2+ on synaptic vesicle endocytosis has been challenging.
Early Insight into the Role of Ca2+ in Endocytosis The first indication that Ca2+ is a key component of the endocytic machinery came from electrophysiological experiments followed by post-stimulation fixation and electron microscopy (Hurlbut and Ceccarelli 1974). Using black widow spider venom, Ceccarelli and colleagues induced Ca2+-independent exocytosis and observed that in the absence of extracellular Ca2+ synaptic vesicle exocytosis caused a physical enlargement of the frog neuromuscular junction, and ultimately a rundown of neurotransmission. But when Ca2+ was included in the extracellular solution, neurotransmission could be maintained and cell swelling was reduced. This observation indicated that synaptic vesicles at the frog neuromuscular junction could only be recycled in the presence of Ca2+ (Ceccarelli and others 1979). These initial electron microscopic data established the executive role of Ca2+ in endocytosis, but electron microscopy, while unrivaled in its ability to visualize nanometer-scale structures, is difficult to apply toward the kinetics of a given process. To extract kinetic information one has to examine many preparations at different time points, and since each preparation must be fixed, it is impossible to monitor processes before and after a certain manipulation within the same cell. Therefore, methods that could be applied to live cells, and that enabled real-time measurements, would be better suited to examine kinetics. So while Heuser, Reese, and Ceccarelli were characterizing the structural intermediates of endocytosis (Ceccarelli and others 1979; Heuser and Reese 1973, 1981; Hurlbut and Ceccarelli 1974), electrophysiologists were working to refine the resolution of real-time recordings to monitor cellular capacitance changes that are proportional to the plasma membrane surface area in living cells. In 1982, Neher and Marty succeeded in this endeavor, resolving membrane capacitance changes on the order of femtofarads (10−15 farads) in adrenal chromaffin cells (Neher and Marty 1982). This allowed them to observe, in real-time, single chromaffin granule exocytosis (as an increase in cell capacitance) and endocytosis (as a decrease in capacitance). While exocytosis occurred as an abrupt increase, the observed time course of endocytosis varied from the limit of their temporal resolution (30 ms) to seconds. Moreover, from these experiments it was immediately evident that Ca2+ not only played an executive role in endocytosis but also a modulatory role on the
rate of endocytosis of large granules. Further refinement of the same technique permitted measurement at the level of small vesicles and also showed a clear Ca2+ dependence of endocytosis (Almers and Neher 1987). Yet despite the strength of electrophysiological measurements, a major limitation of this technique is the requisite use of large secretory cells—such as neuroendocrine cells—or giant synapses, thus precluding measurements of endocytosis in small glutamatergic and GABAergic synapses that make up the majority of the CNS (but see Hallermann and others 2003; Sun and Wu 2001). Therefore optical approaches became the method of choice in these cell types. Initially styryl dyes, such as FM1-43, that embed in the outer lipid leaflet of the plasma membrane but also easily dissociate and can be washed away were used in pulse-chase experiments (Betz and Bewick 1992); for review, see Kavalali and Jorgensen (2014) (Fig. 1). In these experiments, neurons were strongly stimulated in the absence of dye, once stimulation ceased, the extracellular solution was exchanged at varying time points with solution-containing dye, thus providing an indirect measure of endocytosis (Ryan and others 1993). Initially, these experiments suggested that endocytosis was independent of extracellular Ca2+ concentration (Ryan and others 1993; Ryan and others 1996). But similar experiments in drosophila using black widow spider venom as a stimulation again showed Ca2+ was required for endocytosis (Ramaswami and others 1994). Although FM dye–based measurements were successful in providing direct dynamic insight into the process of synaptic vesicle recycling, estimation of endocytosis kinetics using FM dye–based measurements suffered from lack of specificity and difficulty in repeated measurements due to increasing fluorescence background during each cycle of dye application and wash-out (Kavalali and Jorgensen 2014). These shortcomings could be overcome with the development of pHluorin, a pH-sensitive green fluorescent protein (GFP) (Miesenbock and others 1998). With pHluorin synaptic vesicle cycling could now be monitored directly and in a molecularly specific manner by targeting pHluorin to the luminal domains of synaptic vesicle proteins (Fig. 2). Each of these techniques: electron microscopy, electrophysiology, and live-cell imaging have their strengths and weaknesses, and each has in turn provided valuable, albeit at times conflicting, insights into the molecules and mechanisms governing endocytosis (Kavalali and Jorgensen 2014).
Ca2+ Accelerates Endocytosis A series of electrophysiological studies that directly manipulated internal Ca2+ by photo-uncaging in both chromaffin cells (Heinemann and others 1994; Neher and Zucker 1993) and melanotrophs (Thomas and others
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
3
Leitz and Kavalali Residual dye? Exo
Endo
Exo
Endo
Wash
FM1-43
Fluorescence
Ca2+
Figure 1. Optical detection of synaptic fusion and endocytosis using FM (Fei Mao) dyes. FM1-43 and its analogs are amphipathic dyes which fluorescence increases almost 100-fold when they are incorporated into membranes. During a typical experiment, stimulation in the presence of extracellular dye leads to uptake of dye molecules into fused vesicles followed by endocytosis. Subsequent wash out of extracellular dye uncovers fluorescent staining specific to actively recycling vesicles during the stimulation paradigm. Stimulation results in fusion of dye-loaded vesicles and loss of fluorescence trapped in synapses. This fluorescence loss is due to departitioning of the dye into aqueous solution or lateral diffusion of dye in neuronal membrane. In some circumstances, such as kiss-and-run endocytosis, endocytosed synaptic vesicle may still contain residual dye.
A
Single synaptic vesicle fusion and retrieval H+ H+
Fluorescence
Corresponding pHluorin Signal
Time
B
Multiple synaptic vesicle fusion and retrieval H+
H+
H+ H+
Fluorescence
Corresponding pHluorin Signal
Time
Figure 2. (A) SynaptopHluorin is a fusion construct of the synaptic vesicle protein synaptobrevin with a pH-sensitive enhanced green fluorescent protein (EGFP) at its C-terminal (located in the synaptic vesicle lumen). Synaptic vesicle lumen normally has an acidic pH of approximately 5.5 at which SynaptopHluorin fluorescence is quenched. When vesicles fuse, lumenal EGFP is exposed to the extracellular pH, which results in a marked increase in its fluorescence. Fluorescence signal remains elevated for the duration when synaptic vesicle lumen is exposed to extracellular pH (assuming SynaptopHluorin molecules do not diffuse away and remain clustered). During endocytosis, pHluorin fluorescence is requenched as vesicle lumen becomes acidic. Panel A depicts single vesicle exocytosis-endocytosis and the corresponding fluorescence trace. Panel B depicts simultaneous fusion and endocytosis of multiple synaptic vesicles and the corresponding fluorescence signal.
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
4
The Neuroscientist
1993; Thomas and others 1994) found endocytosis proceeded with two kinetic components: a rapid component with time constant of less than 5 seconds that dominated when internal Ca2+ concentrations were high (≥10 μM and 50 μM in melanotrophs and chromaffin cells, respectively) and a slower component that was on the order of tens of seconds to minutes at low intracellular Ca2+ concentrations. Thus in these studies, increasing intracellular Ca2+ accelerated the rate of endocytosis. In agreement with this premise, later electrophysiological studies found that low concentrations of Ca2+ inhibited endocytosis in chromaffin cells (Ales and others 1999; Artalejo and others 1995; Burgoyne 1995; Smith and Neher 1997), in pancreatic β-cells (Eliasson and others 1996), and in goldfish bipolar cells (Neves and others 2001). Interestingly in two of these studies (Eliasson and others 1996; Smith and Neher 1997) the slow compensatory mode of endocytosis was lost following dialysis of the cell while in whole cell recording configuration—an early indication that Ca2+ acts on diffusible factors to control vesicle retrieval. The Ca2+/calmodulin-dependent phosphatase calcineurin was later identified as such a possible factor controlling endocytosis (Engisch and Nowycky 1998), which may then act on dynamin (Artalejo and others 1995; Cousin and Robinson 2001; Marks and McMahon 1998) to accelerate endocytosis. Meanwhile, optical evidence supporting an acceleration of endocytosis by Ca2+ in hippocampal neurons was obtained using various styryl dyes (FM1-43, FM2-10 and FM1-84) with distinct membrane dissociation kinetics (Kavalali and others 1999; Klingauf and others 1998). FM dye-labeling experiments in drosophila using black widow spider venom as stimulation showed that Ca2+ not only triggered but also could accelerate endocytosis (Ramaswami and others 1994). In experiments with expression of vesicular proteins tagged with pH-sensitive GFP (pHluorin) (Sankaranarayanan and Ryan 2001) also found endocytosis accelerated by increasing Ca2+. Labeling vesicle recycling with sulforhodamine uptake in snake motor terminals indicated that increasing Ca2+ shifted endocytosis toward a faster rate (Teng and Wilkinson 2005). A variety of electrophysiological and optical experiments in the calyx of Held—a large synaptic terminal in the auditory brain stem amenable to direct capacitance measurements—also reported a Ca2+dependent acceleration of endocytosis (Hosoi and others 2009; Sun and others 2010; Wu and others 2005; Wu and others 2009). Additionally, these studies further supported the Ca2+/calmodulin sensitive phosphatase calcineurin (Sun and others 2010; Wu and others 2009) as a key Ca2+-sensitive component of the endocytic machinery. Although these studies agree on the general Ca2+dependent acceleration in endocytic kinetics, there was significant variation in the reported absolute rate of
endocytosis and the number of kinetic steps involved in the process (see Table 1). In some instances, these differences were attributed to cell type and in other cases, the technique employed, but these studies agreed that endocytosis is accelerated by increasing Ca2+ concentrations.
Ca2+ Decelerates Endocytosis In parallel with growing evidence that Ca2+ accelerates endocytosis, there is also substantial evidence that Ca2+ decelerates endocytosis. Just over a year after the initial studies by Neher and Zucker reported an acceleration of endocytosis by Ca2+, the first study to systematically examine the relationship of Ca2+ and endocytosis, produced the opposite finding in goldfish retinal bipolar cells that possess synaptic terminals large enough to be accessible to capacitance measurements (von Gersdorff and Matthews 1994). Here endocytosis was slowed when intracellular Ca2+ was increased—as monitored by the Ca2+ indicator Fura-2. In this configuration, endocytosis proceeded on the order of seconds and was steeply dependent on intracellular Ca2+, with a Hill coefficient of 4— strikingly similar to the Ca2+ coordination of synaptotagmin in exocytosis (Fernandez-Chacon and others 2001). In retinal bipolar cells a model using two relatively fixed kinetic components of endocytosis was proposed (Neves and Lagnado 1999) where increasing Ca2+ increased the proportion of vesicles retrieved by the slow kinetic process. Although this model agrees that endocytosis slows with increasing Ca2+, it is important to note that this contrasts with a model where the kinetics are themselves are subject to change by Ca2+. More recent studies again strongly supported the role of calmodulin in regulating the Ca2+-dependent deceleration of endocytosis in mouse retinal bipolar cells (Wan and others 2008; Wan and others 2010) and dynamin in cortical synaptosomes (Cousin and Robinson 2000). These studies were bolstered by experiments in dissociated hippocampal neurons using pHluorin-tagged synaptic vesicle proteins to monitor synaptic vesicle retrieval (Armbruster and others 2013).
Ca2+ Does Not Affect the Rate of Endocytosis Finally, in addition to positive or negative regulation of endocytosis by Ca2+ a third possibility also emerged: that synaptic vesicle recycling is independent of Ca2+ and instead controlled by the number of synaptic vesicles that undergo fusion, or exocytic load (Ryan and others 1993; Ryan and others 1996; Wu and Betz 1996). This hypothesis emerged from studies using styryl dye uptake and release to quantify the rate of endocytosis. In these experiments, at various times after stimulation, the extracellular solution
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
5
Leitz and Kavalali Table 1. Brief chronological summary of experiments investigating the role of Ca2+ in endocytosis. A variety of experimental procedures, preparations and stimulations have produced a range of kinetic parameters for endocytosis.
Year
Authors
Preparation
Measurement Method
1974 Hurlbut and Cecarelli
Frog neuromuscular Rapid freeze junction fixation followed by electron microscopy (EM) 1984 Miller and Heuser Frog neuromuscular Rapid freeze fixation junction followed by EM 1985 Torri-Tarelli and Frog neuromuscular Rapid freeze fixation others junction followed by EM 1993 Neher and Zucker Bovine adrenal Capacitance chromaffin 1993 Ryan and others
Hippocampal FM 1-43 cultures Goldfish bipolar cell Capacitance
1994 von Gerhersdorff and Matthews 1994 Thomas and others Melanotrophs
1994 Heinemann and others
Capacitance
Bovine adrenal chromaffin
1994 Ramaswami and others
Capacitance
FM 1-43
Drosophila neuromuscular junction 1995 Burgoyne Bovine adrenal chromaffin 1995 Artalejo and others Bovine adrenal chromaffin 1996 Eliasson and others Pancreatic β-cells
Capacitance
1996 Ryan and others
FM 1-43
1996
FM 1-43
1997 1998 1998 1998 1998 1999
Hippocampal cultures Wu and Betz Frog neuromuscular junction Smith and Neher Bovine adrenal chromaffin Engisch and Bovine adrenal Nowycky chromaffin Klingauf and others Hippocampal neurons Marks and Rat synaptosomes McMahon Rouze and Goldfish bipolar cell Schwartz Neves and Lagnado Goldfish bipolar cell
1999 Ales and others 2000 Cousin and Robinson 2000 Sankaranarayanan and Ryan 2001 Heidelberger 2001 Neves et al
Capacitance Capacitance
Capacitance Capacitance FM dyes FM 2-10 Capacitance and FM 4-64 Capacitance, FM
Rat chromaffin cells Capacitance and amperometry Synaptosomes FM2-10 uptake Hippocampal cultures
SynaptopHluorin (Syb2/VAMP2pHluorin) Capacitance
Mouse rod bipolar cell Goldfish bipolar cell Capacitance
Stimulation Strength Multivesicle (Black Widow spider venom)
Number of Kinetic Steps N/A
Approximate Time-Constant of Endocytosisa
Role of Ca2+
Single Vesicle?
N/A
Required
No
Required
No
Required
No
Single AP
2
Single AP
1
Ultrafast and Slow Ultrafast
Multivesicle (flash photolysis) Multivesicle (high K+) Multivesicle (evoked) Multivesicle (flash photolysis) Multivesicle (flash photolysis) Multivesicle (high K+)
1
Fast
Accelerates
No
1
Fast
Independent
No
1
Fast
Decelerates
No
2
Ultrafast and Fast
Accelerates
No
2
Ultrafast and Slow
Accelerates
No
N/A
N/A
Independent
No
2
Fast
Accelerates
No
3
Accelerates
No
2
Ultrafast and Fast Fast and Slow
Accelerates
No
1
Fast
Independent
No
1
Fast and slow
Independent
No
2
Fast
Accelerates
No
3
Ultrafast and Fast Fast
Accelerates
No
Accelerates
No
N/A
Independent
No
N/A
Both
No
Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (high K+) Multivesicle (high K+) Multivesicle (high K+) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (high K+) Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked)
1
2
Fast
Decelerates
No
1
Ultrafast
Accelerates
No
Decelerates
No
N/A 1
Fast and slow
Independent
No
1
Fast
Decelerates
No
1
Ultrafast and Fast
Accelerates
No (continued)
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
6
The Neuroscientist
Table 1. (continued)
Year
Authors
Preparation
2001 Sankaranarayanan and Ryan
Hippocampal cultures
2002 Sun and others
Calyx of Held
2003 Gandhi and Stevens Hippocampal cultures 2003 Aravanis and others 2005 Teng and Wilkinson
Measurement Method
Stimulation Strength
Number of Kinetic Steps
Multivesicle SynaptopHluorin (evoked) (Syb2/VAMP2pHluorin) Capacitance (of minis) Spontaneous/ single AP Single AP SynaptopHluorin (Syb2/VAMP2pHluorin) FM 1-43 Single AP
1
Approximate Time-Constant of Endocytosisa
Role of Ca2+
Single Vesicle?
1
Fast
Accelerates
No
3
Ultrafast and Fast Fast and Slow (stranded)
Independent
Yes
Decelerates
Yes
N/A
Yes
Accelerates
No
3
Sulforhodamine 101
Multivesicle (evoked)
N/A
2005 Wu and others
Hippocampal cultures Snake neuromuscular junction Calyx of Held
Capacitance
2
Fast
Accelerates
No
2006 Granseth and others
Hippocampal cultures
SynaptophysinpHluorin (SypHy)
1
Fast
Independent
Yes
2007 Balaji and Ryan
vGlut-pHluorin
1
Fast
1
Fast
Decelerates
No
2009 Hosoi and others
Hippocampal cultures Mouse rod bipolar cell Calyx of Held
Multivesicle (evoked) Single AP/ multivesicle (evoked) Single AP
Capacitance
2
Fast
Accelerates
No
2009 Wu and others
Calyx of Held
Capacitance
2
Fast
Accelerates
No
2009 Zhu and others
Hippocampal cultures Mouse rod bipolar cell Calyx of Held and hippocampal cultures
SynaptophysinpHluorin (4x) Capacitance
2
Fast
Accelerates
Yes
1
Fast
Decelerates
No
1
Fast
Accelerates
No
1
Fast
Accelerates
No
1
Fast
Decelerates
Yes
1
Fast
Decelerates
No
1
Ultrafast
Independent
Yes
2008 Wan and others
2010 Wan and others 2010 Sun and others
2010 Yamashita and others 2011 Leitz and Kavalali 2013 Armbruster and others 2014 Leitz and Kavalali
Capacitance
Capacitance; synaptopHluorin (syb2/VAMP2pHluorin)
Calyx of Held
Capacitance
Hippocampal cultures Hippocampal neurons Hippocampal cultures
vGlut-pHluorin vGlut-pHluorin vGlut-pHluorin/ synaptophysinpHtomato (SypHTo)
Multivesicle (evoked) Multivesicle (evoked) Multivesicle (evoked) Single AP Multivesicle (evoked) Capacitance steps/ extracellular stimulation 20 Hz Multivesicle (evoked) Single AP Multivesicle (evoked) Spontaneous
Fast
Yes
a
“Ultrafast” is a subsecond to second range, “fast” is 1 second to tens of seconds, and “slow” is tens of seconds to minutes.
was exchanged with solution containing dye. The amount of dye uptake would then quantify the amount of endocytosis. These experiments concluded that endocytosis proceeded with a t1/2 of 60 seconds (Ryan and others 1993). Additional experiments demonstrated that when extracellular solution was changed to a Ca2+-free solution, dye could still be taken up to the same degree as in Ca2+containing solution (Ryan and others 1993; Ryan and others 1996) Together, these early imaging experiments
generated the hypothesis that the rate of endocytosis was independent of Ca2+ but dependent on the number of synaptic vesicles that fused during stimulation. Similar styryl dye experiments in the frog neuromuscular junction further supported this hypothesis (Wu and Betz 1996). The assumption that endocytosis and exocytosis were temporally segregated but coupled via their mutual dependence on Ca2+ influx was supported by experiments in cortical synaptosomes (Marks and McMahon 1998). Further
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
7
Leitz and Kavalali
Capacitance = Cm
n
Capacitance = Cm + (Σ ∆Cvi) i=1
n
Capacitance = Cm - (Σ ∆Cvi) i=1
+∆Cv1 +∆Cv2
Cv1 Cv2 Cv3
-∆Cv1 -∆Cv2 -∆Cv3
+∆Cv3
Cvn
+∆Cvn
Cm
-∆Cvn
Cm
Cm
Capacitance = Cm Balanced exocytosis and endocytosis No net change in Cm
-∆Cv +∆Cv Cm
Figure 3. Capacitance measurements. Top panels: Synaptic vesicles are membranous organelles, which release their neurotransmitter content after fusion with the plasma membrane. This process can be directly monitored via capacitance measurements. Alteration in total membrane capacitance can be used as reporter of membrane fusion or endocytosis as long as the magnitude of the capacitance change is above the background noise. A critical requirement for employment of this method is direct electrical access to the secretory cellular component. This requirement significantly impairs the applicability of this method to small synapses of the central nervous system. Therefore, most capacitance measurements have been limited to secretion from cell bodies of mast cells, adrenal chromaffin cells, large pituitary nerve terminals or large presynaptic terminals such as the calyx of Held, located within the brain stem. Lower panel: However, under some circumstances when exocytosis and endocytosis processes are tightly balanced, capacitance measurements may not report a net change in membrane area despite extensive vesicle trafficking.
studies using pHluorin imaging in hippocampus cultures (Sankaranarayanan and Ryan 2000; but see Sankaranarayanan and Ryan 2001) and capacitance recordings in the calyx of Held (Sun and others 2002) have indicated that the degree of exocytosis impacts the kinetics of endocytosis. These seemingly contradictory results could be reconciled in a study performed in goldfish retinal bipolar cells where both very low and very high levels of Ca2+ produced no change in membrane capacitance, while the rate of endocytosis rose and fell as Ca2+ deviated from the extremes (Rouze and Schwartz 1998). This work highlighted the important point that capacitance measurements can only report a net change in membrane surface area (Fig. 3). Therefore no change in capacitance could be the result of either no membrane addition or subtraction, or of membrane addition and subtraction occurring simultaneously at equal rates. The simultaneous occurrence of exocytosis and endocytosis during
activity was also evident in pHluorin based measurements in hippocampal synapses where application of vacuolar ATPase inhibitors such as bafilomycin or folimycin inhibited synaptic vesicle re-acidification and revealed a prominent rapid component of endocytosis occurring in synchrony with exocytosis (Ertunc and others 2007; Fernandez-Alfonso and Ryan 2008; Sankaranarayanan and Ryan 2001). Strong coupling between exocytosis and endocytosis complicates the interpretation of most experiments because while all of these studies use a variety of cell types and experimental procedures, the one factor they have in common is the use of strong stimulation (Fig. 4). So this issue raises the question whether endocytosis can be investigated independent of exocytosis. One way to decouple these processes is to know the exact number of synaptic vesicles that participated in fusion during stimulation. To this end, several groups sought to monitor exocytosis and endocytosis of a single synaptic vesicle.
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
8
The Neuroscientist
Single vesicle exocytosis and endocytosis
Multivesicle exocytosis and endocytosis
Retrograde inhibition
Membrane infoldings
Figure 4. Single versus multivesicle exo-endocytosis. Most studies probing the role of Ca2+ in endocytosis have relied on measurements of synaptic vesicle retrieval after strong stimulation. Strong stimulation paradigms elicit fusion and retrieval of multiple synaptic vesicles. Therefore, interpretation of these experiments is confounded by several factors besides the kinetics and duration of Ca2+ signals. These include the number of exocytosed vesicles, accumulation of released neurotransmitters which all alter fusion and retrieval processes indirectly via retrograde signaling. Studies monitoring single synaptic vesicle endocytosis, on the other hand, allow better dissociation of fusion and retrieval processes as in these settings the impact of Ca2+ on synaptic fusion probability can be uncoupled from Ca2+’s putative impact on synaptic vesicle retrieval.
Endocytosis of Single Synaptic Vesicles Monitoring retrieval of single synaptic vesicles following fusion provides a clear way to uncouple regulation of synaptic vesicle fusion and retrieval. As fusion occurs in a quantal, all-or-nothing fashion, once a vesicle is fused its subsequent trajectory can be kinetically isolated from its fusion probability. The isolation of exocytosis and endocytosis kinetics is hard to achieve if a large number of vesicles fuse in rapid succession (Fig. 4). The binary nature of single vesicle fusion, therefore, eliminates the requirement to normalize the extent of endocytosis to exocytosis to isolate potential direct effects on endocytosis. Initial attempts at detecting single synaptic vesicle exo-endocytosis using capacitance measurements heavily relied on signal averaging and was confounded by susceptibility to contamination by capacitance changes
unrelated to synaptic vesicle exocytosis (Sun and others 2002; Yamashita and others 2005). Therefore, to resolve vesicle recycling at the single vesicle level optical techniques again became the method of choice. However, to employ optical techniques in hippocampal neurons, several approaches were developed to overcome the low signal elicited by single synaptic vesicle fusion. Experiments sought to minimize background fluorescence by pre-photobleaching—which damages the cells—or fluorescent trace averaging, that occludes subtle differences between fusion events and assumes that the components of vesicle cycling (i.e., the kinetics the amplitude) are normally distributed (Gandhi and Stevens 2003; Granseth and others 2006). Another complication in optical experiments is the lack of accessible and practical postsynaptic confirmation of synaptic vesicle fusion (but see Leitz and Kavalali 2014; Richards 2009). Therefore, measurements at single vesicle resolution rely on several corollary indicators for validation: such as vesicle fusion probability distributions and quantal fluorescence amplitude distributions. Both FM dyes and pHluorin-tagged proteins have been applied toward single vesicle imaging, but each study that utilized pHluorin yielded different endocytic kinetics depending on which synaptic vesicle protein was tagged with pHluorin. First, synaptobrevin-pHluorin (SynaptopHluorin) revealed three distinct kinetic components of endocytosis: a fast mode consistent with “kiss-and-run” kinetics (with decay time of less than 3 seconds), a slower mode (with decay time in the tens of seconds) and a stranded mode where fluorescence did not decay (Gandhi and Stevens 2003). Although the authors did not investigate directly the effect of Ca2+ on endocytosis, they did show that in higher Ca2+ the rapid component of decay was absent, suggesting that as Ca2+ increases, endocytosis slows. In these experiments, single vesicle resolution was determined by a recapitulation of vesicle fusion probability observed in classical literature, as well as quantal amplitude distribution analysis in elevated Ca2+ concentration. Later, other experiments supported a predominance of rapid kiss-andrun exo-endocytosis at low intracellular Ca2+ as determined by FM dye destaining kinetics at the single vesicle level (Aravanis and others 2003; Richards and others 2005). Another study took advantage of both SynaptopHluorin and pHluorin fused to the luminal domain of synaptophysin and reported a single slow form decay time-constant in the range of 14 to 20 seconds extracted from signal averaging the fluorescence changes from several putative single vesicle fusion events when a stimulation was applied (Granseth and others 2006). Furthermore, directly comparing normalized average fluorescence traces of single vesicle events with high frequency stimulation produced nearly identical decay traces leading to the conclusion that the rate of endocytosis might be constant
Downloaded from nro.sagepub.com at FLORIDA ATLANTIC UNIV on July 17, 2015
9
Leitz and Kavalali (Granseth and others 2006). A subsequent study took advantage of pHluorin fused to the vesicular glutamate transporter1 (vGlut-pHluorin), which showed significantly reduced background fluorescence and reported a similar single stochastic kinetic component of endocytosis with average time constant of ~14 seconds (Balaji and Ryan 2007). In these experiments, single vesicle resolution was verified by the monophasic distributions of event amplitudes used for analysis. The same group later monitored the rate of endocytosis while varying stimulation frequency or Ca2+ concentration (Balaji and others 2008). Increasing stimulation frequency or Ca2+ concentration slowed endocytosis during 10-Hz stimulation but at the single vesicle level, stimulation frequency had no effect, leading to the conclusion that the exocytic load — the number of vesicles fused—determines the endocytic rate. While vesicle fusion probability was not explicitly examined in either of these studies, examination of example traces suggests a very high release probability in these synapses (Balaji and Ryan 2007; Balaji and others 2008). A subsequent study that utilized an improved synaptophysin-pHluorin construct with four pHluorin tags to increase signal-to-noise ratio identified two kinetic components to endocytosis; one fast (~3 seconds) and one slow (~10 seconds) (Zhu and others 2009). Here, increasing Ca2+ decreased the dwell time of fluorescence—indicating shorter residence of synaptophysin-pHluorin molecules at the surface membrane on fusion—while the number of vesicles that fused increased the decay time of only the slow component of endocytosis, the fast rate was unchanged. Together, this slowing of decay time and decrease in dwell time resulted in a net acceleration of endocytosis and led the authors to propose a model where two vesicles are retrieved after one vesicle fuses. These experiments did not show synaptic vesicle fusion probability but instead relied on quantal amplitude distributions as validation of single vesicle resolution. Some support for this two vesicle model could be found in recent “flash-and-freeze” electron microscopy data that shows the ultrafast internalization (less than 100 ms) of large compartments with surface area four times that of a synaptic vesicle (Watanabe and others 2013). However, it is not known if these “compensatory” endocytic events are the product of vesicles that immediately underwent fusion, previously underwent fusion (stranded vesicles retrieved by subsequent rounds of stimulation as previously demonstrated by Thomas and others (1994)) or instead reflect retrieval of surface membrane and not fused synaptic vesicles. Although each of the above studies has visualized synaptic vesicle fusion at the single vesicle level, they produced different estimates for the rate of single synaptic vesicle endocytosis. Therefore, in recent work we sought
to systematically investigate the role of Ca2+ in endocytosis at the single vesicle level (Leitz and Kavalali 2011, 2014). Using vGlut-pHluorin to monitor single vesicle fusion we generated quantal amplitude distributions to validate single-vesicle resolution. In this setting, we found that in near physiological (2 mM) extracellular Ca2+, fluorescence decay after single action potential stimulation was a rapid process well described by a single exponential with average decay time of ~3 seconds that proceeded immediately after fluorescence increase, although some fluorescence traces did manifest a detectable dwell time likely indicating a surface residency period of
