Note to users. If you're seeing this message, it means that your browser cannot find this page's style/presentation instructions -- or possibly that you are using a browser that does not support current Web standards. Find out more about why this message is appearing, and what you can do to make your experience of our site the best it can be.

Science 18 September 2009:
Vol. 325. no. 5947, p. 1499
DOI: 10.1126/science.1176007
Prev | Table of Contents | Next

Technical Comments

Response to Comment on "The Dynamic Control of Kiss-And-Run and Vesicular Reuse Probed with Single Nanoparticles"

Qi Zhang, Yulong Li, Richard W. Tsien*

Granseth et al. argue that vesicle retrieval at hippocampal synapses is fully accounted for by a single mode of endocytosis. However, their assay focused on readily releasable pool vesicles (RRP), not RRP + reserve pool vesicles in tandem, and therefore cannot detect pool-dependent changes in vesicle-retrieval kinetics during a stimulus train. Using a probe similar to theirs, we observed rapid vesicle retrieval consistent with Kiss-And-Run fusion.

Department of Molecular and Cellular Physiology, Stanford University, Stanford, CA 94305, USA.

* To whom correspondence should be addressed. E-mail: rwtsien{at}stanford.edu

Vesicle recycling is critical for continued synaptic transmission at nerve terminals. We used single quantum dots (Qdots) to investigate synaptic vesicle behavior and found that Kiss-And-Run (K&R) vesicle fusion dominates at the beginning of stimulus trains and then gradually gives way to classical full-collapse fusion (FCF) (1). Based on previously reported data (2), Granseth et al. (3) argue that the pattern of vesicle recycling using pooled signals from a pH-sensitive green fluorescent protein (GFP) probe (synaptophysin-pHluorin or sypHy) is best explained by FCF only. They averaged the sypHy signals to the first and the 8th to 10th stimulus pulses during the 0.02-Hz stimulation train and observed no detectable difference between the two averages. We welcome this opportunity to clarify an advantage of the Qdot method: the ability to monitor single-vesicle behavior across the total recycling vesicle pool [readily releasable pool (RRP) + reserve pool]. In our study (1), changes in the dynamics of fusion/retrieval that originate from differences in behavior of RRP and reserve pool vesicles were evident during stimulus trains as the minority Qdot-loaded RRP vesicles were quickly depleted, giving way to the majority Qdot-loaded reserve pool vesicles, which are less likely to support K&R. In contrast, the sypHy signals of Granseth et al. (3) would arise almost entirely from RRP vesicles. At the low-frequency stimulation they used (0.02 Hz), each and every stimulus recruits vesicles with high vesicle release probability Pr,v—those residing in the RRP. Because all vesicles were sypHy-labeled, the RRP was continually replenished with labeled vesicles. Thus, the lack of deviation between their first and 8th to 10th responses is entirely to be expected. Our own experiments with a pHluorin-based probe confirm this expectation.

To explore whether signals corresponding to K&R can be obtained with a proteinaceous probe, we used synaptophysin tagged with four copies of pHluorin (SypH4X) (4). We were able to detect single-vesicle fusion events reliably, using 0.3 Hz stimulation to increase the yield of unitary signals (examples in Fig. 1B). We observed at least three distinct patterns of decay of the SypHy4X signal (Fig. 1B): (i) fast, complete decay (4, 5); (ii) staggered decay, with an early partial downstep followed by a later recovery (4); and (iii) a simple, late recovery (2, 6). The overall time course (averaged from 244 records) was not consistent with a single slow mode of retrieval but was well-fitted by a combination of fast and slow modes, using the same equation as Granseth et al. (2) (Fig. 1C). The best fit by {chi}2-test (Fig. 1C, inset) was obtained with a 52% contribution of K&R events. A similar biphasic time course was seen with unitary transients at the beginning of the stimulus train, after a long quiescence. Acute exposure to a vesicular H+-pump blocker, bafilomycin, abolished the rapid phase (Fig. 1D), indicating that it corresponds to internalization of probe into a reacidifying compartment, the lumen of retrieved vesicles. The reacidification time constant for brief events, 1.45 ± 0.18 s, was not significantly different from that obtained with Qdots 0.95 ± 0.18 s (P > 0.05, unpaired t test), which also agrees with the results from (5) (Fig. 2A).


Figure 1
View larger version (26K):
[in this window]
[in a new window]

  		Fig. 1. Single-vesicle fusion events probed with SypH4X. (A) Cartoon and hypothetical traces for SypH4X- and Qdot-based measurements. (B) Representative traces of SypH4X fluorescence signals reporting single fusion events (0.3 Hz stimulation). (C) Ensemble average (black trace) and the best fit with the same equation used by Granseth et al. (2) (solid purple line), consisting of a fast K&R component (red dashed line, 52%, {tau}KR = 0.53 s) and a slow FCF component (blue dashed line, 48%, {tau}FCF = 49 s). We fixed the following parameters according to our experiments: kr = 1/{tau}r = 1/1.45s = 0.70s–1, kKR = 1/{tau}KR = 1/0.53 s = 1.89s–1, and allowed {tau}FF = 1/kFF to vary between 10 and 60 s. The values of {tau}KR and {tau}r are faster than those estimated in (4) (both 2.6 s), though the interpretations are otherwise compatible. (Inset) The {chi}2 value falls to a minimum at 52% K&R. (D) Same ensemble average (black circles and line: mean ± SEM), compared with pooled SypH4X signals in 1µM bafilomycin A1 (applied 30 s before the first stimulus pulse). The remaining slow decay is likely due to the post-FCF lateral diffusion of SypH4X out of the region of interest. Cultured rat hippocampal CA3-CA1 neurons (1) transfected with SypH4X (4) at 8 to 11 days in vitro (DIV) and studied at 14 to 20 DIV. Neurons stimulated at 0.3 Hz for 2 min and imaged at 3 Hz with a Cascade 512B camera, standard GFP filter cube (Chroma), and a 40x, 1.3 NA objective with a 4x projection lens. Positive events were identified in kymographs with a moveable region of interest (12) to track an individual SypH4X cluster over time while minimizing the effect of its relocation. The spatial distribution of fluorescence was fitted to a two-dimensional Gaussian to define the center of a 2 µm-diameter circle for integrating total fluorescence.

 


Figure 2
View larger version (27K):
[in this window]
[in a new window]

  		Fig. 2. Interpretation of Qdot signals from vesicle fusion experiments. (A) Time constants of reacidification measured by Qdots (red symbols) are consistent with pHluorin-based data of (5) (black line and symbols). (B) Movements of single Qdots during fusion pore open time are usually < 50 nm, and almost always < 100 nm. (Inset) Scenarios of hypothetical Qdot reuptake. (C) Contradicting Qdot reuptake scenario, lack of upstep, documented with analysis 4 s (dashed line) after stimulation under normal conditions. Complete absence of upstep (open arrowhead), contrasts with clear upstep in acute bafilomycin (filled arrowhead).

 
There are many potential explanations for the discrepancy between Granseth et al. (2, 3) and data obtained with Qdots (1) and with SypH4X (4). One may consider differences in cultures, experimental conditions, cytosolic calcium concentration, or the degree of G-protein activation, known to affect the balance between K&R and FCF (7). A broadly accommodating scenario is that while the Qdot (15 nm in diameter) remained trapped inside the vesicle and reported a K&R event, the fusion pore could dilate to a lesser degree (0.5 to 14 nm), perhaps even allowing escape of a subset of vesicular protein markers by lateral diffusion within the membrane. Thus, pHluorin-labeled vesicle proteins might escape during an event identified as K&R by Qdot analysis, akin to "cavicapture" (8).

Regarding the concern that Qdots induced fast vesicle retrieval, our results indicated that even maximal loading of the total recycling pool with Qdots does not affect vesicle trafficking as monitored with the styryl reporter dye FM 4-64 (9), electrophysiology (9), and a pHluorin-based probe (1). Granseth et al. (3) question whether the uptick equates with K&R rather than full fusion followed by retrieval of a Qdot. We can exclude the latter on two grounds. Co-occurring uptake of a fresh Qdot can be firmly excluded; there were no Qdots on the surface during destaining (removed by the 20-min wash), and only a single, vesicle-enclosed Qdot remained visible in the region of interest. The alternative scenario, immediate recapture of the exocytosed Qdot by an adjacent clathrin-coated, endocytosis-ready vesicle, was disfavored by the limited movement of Qdots during the deacidification period, which was mostly less than the minimal center-to-center distance (100 nm) between a fused vesicle and a putative endocytosis-ready vesicle next to it (Fig. 2B). Finally, we considered whether Qdots might be recaptured by endocytosis of the original vesicle itself (Fig. 2C). This scenario predicts that Qdot photoluminescence signal would generate an "upstep," reflecting Qdot externalization after FCF and brightening before re-internalization. Evidence on clathrin-mediated endocytosis predicts an upstep with a lifetime of many seconds [{tau} = 14 s, according to (3)], but nothing like this was ever observed (Fig. 2C). Instead, the plateau of the uptick lasted on average ~0.5 to 1 s (1), consistent with capacitance measurements (10) and much too fast for conventional clathrin coat formation (11). Taken together, our results preclude the idea that the same Qdot was recaptured, either by an adjacent endocytotic vesicle (Fig. 2B) or by endocytosis of the original vesicle (Fig. 2C).

A final scenario, inconsistent with the proposal of Granseth et al. (3) but worth considering, is that K&R occurs but is supported by a hypothetical clathrin precoat rather than de novo clathrin assembly. Clathrin-coated vesicles have rarely been seen at the active zone of quiescent nerve terminals, nor is it clear how a clathrin precoat could coexist with the SNARE proteins of the fusion machinery. Nonetheless, if K&R arises because of mechanical constraints on what would otherwise be FCF, clathrin precoating is a possible candidate (2) that would be consistent with Zhu et al. (4).

Each of the approaches used to study vesicle recycling has its own virtues and drawbacks. In addition to their favorable optical properties, Qdots can selectively label one pool of vesicles (1), whereas pHluorin-based indicators cannot. Single vesicles take up only one Qdot at most (1, 9), allowing unambiguous tracking of the vesicle that harbors it, even if the vesicle were to exchange some proteins or lipids with the plasma membrane. However, Qdots do not yet track vesicle proteins or lipids. Ultimately, and perhaps already (Fig. 2A), Qdot- and pHluorin-based measurements will provide complementary insights into intriguing aspects of vesicle fusion, retrieval, and reuse.

References and Notes

Received for publication 1 June 2009. Accepted for publication 27 August 2009.

To Advertise     Find Products

ADVERTISEMENT

Featured Jobs

Science. ISSN 0036-8075 (print), 1095-9203 (online)