Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 2013 Jan 22; 110(4): 1333–1338.
Published online 2013 Jan 8. doi:  10.1073/pnas.1218818110
PMCID: PMC3557091
Biophysics and Computational Biology

Fusion pore formation and expansion induced by Ca2+ and synaptotagmin 1


Fusion pore formation and expansion, crucial steps for neurotransmitter release and vesicle recycling in soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE)-dependent vesicle fusion, have not been well studied in vitro due to the lack of a reliable content-mixing fusion assay. Using methods detecting the intervesicular mixing of small and large cargoes at a single-vesicle level, we found that the neuronal SNARE complexes have the capacity to drive membrane hemifusion. However, efficient fusion pore formation and expansion require synaptotagmin 1 and Ca2+. Real-time measurements show that pore expansion detected by content mixing of large DNA cargoes occurs much slower than initial pore formation that transmits small cargoes. Slow pore expansion perhaps provides a time window for vesicles to escape the full collapse fusion pathway via alternative mechanisms such as kiss-and-run. The results also show that complexin 1 stimulates pore expansion significantly, which could put bias between two pathways of vesicle recycling.

Keywords: membrane fusion, synaptic vesicle, fusion pore dynamics, DNA probe, single molecule

Soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNAREs) mediate intracellular vesicle fusion in a wide variety of cellular activities such as neurotransmitter release. The fast synaptic vesicle fusion for neurotransmitter release is regulated with precision by various proteins including synaptotagmins, complexins, and SM proteins (1, 2). During this process, an initial fusion pore between two membranes can either close back or expand to a larger pore. Fusion pore expansion to the point where the vesicle membrane flattens on the plasma membrane surface, leading to the complete luminal contents release, is thought to be the final step in the fusion process (3, 4). SNAREs and accessary proteins may then be recycled to make fresh vesicles through endocytosis. Without pore expansion, however, the vesicles may be used again through the mechanism known as “kiss-and-run” (5). Therefore, pore expansion is an important event that determines how synaptic vesicles are regenerated.

To dissect the SNARE-mediated membrane fusion process, we and others developed in vitro single-vesicle assays based on lipid mixing of proteoliposomes reconstituted with SNARE proteins and content mixing of small cargoes (610). However, these assays are blind to the expansion of the fusion pore and therefore unable to tell how the regulatory proteins are involved in this final step of the full-collapse fusion pathway, in which the small opening of the pore continues to expand to a large pore.

To monitor fusion pore expansion, we developed a single-molecule/vesicle content-mixing assay based on vesicle-encapsulated DNA molecules (11, 12). This assay can detect expansion of the fusion pore that is large enough to pass ∼11-kDa DNA probes between two apposed proteoliposomes. With this method, we showed that yeast SNAREs alone can efficiently drive expansion of the fusion pore (12). In this work, we systematically dissect lipid mixing, fusion pore opening, and fusion pore expansion steps in neuronal SNARE-dependent membrane fusion. To this end, we use the combined methods of single-vesicle lipid mixing, content mixing of small cargoes, and content mixing of large DNAs and examine the requirements of regulatory factors including a major Ca2+-sensor synaptotagmin 1 (Syt1), a soluble presynaptic protein complexin 1, and Ca2+.


SNARE-Dependent Lipid Mixing and Content Mixing of Small Cargoes.

First, we revisited lipid mixing based on lipophilic probes (7) and content mixing based on small sulforhodamine B (9) in the single-vesicle fusion assay. For lipid mixing, we prepared a population of vesicles with v-SNARE VAMP2 (lipid-to-protein ratio L/P of 200) doped with a fluorescence acceptor DiD [1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine perchlorate, DiIC18(5)] (1 mol%) (v-vesicles) and a separate population of vesicles with t-SNARE (a 1:1 precomplex of Syntaxin 1A and SNAP-25, L/P = 200) doped with a 1 mol % fluorescence donor DiI [1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate, DiIC18(3)] (t-vesicles). In a flow cell, v-vesicles were tethered to the PEG-coated imaging surface using the biotin-neutravidin conjugation and t-vesicles were then flown in to dock and fuse with surface-bound v-vesicles.

One minute after the reaction was started at room temperature, we observed lipid mixing for 31% of docked vesicle pairs (Fig. 1A). When Syt1 was incorporated in the v-vesicles (L/P = 200) we observed lipid mixing for as much as 50% of docked vesicle pairs, indicative of some stimulation of lipid mixing by Syt1. However, we observed a factor of 3.1 increase of the docking number when Syt1 was present in the v-vesicles (Table S1), most likely due to the t-SNARE–Syt1 interaction that is thought to assist vesicle docking (1315).

Fig. 1.
(A) Lipid mixed fractions of docked vesicle pairs 1 min after docking. The first bar is with SNAREs only, and the second bar is with SNAREs and Syt1. The error bars represent SD. (B) A typical time trace of fluorescence dequenching of small sulforhodamine ...

Next, we prepared v-vesicles encapsulating sulforhodamine B in a high enough concentration (50 mM) to have fluorescence self-quenching and empty t-vesicles separately. We used the same experimental strategies that Brunger and coworkers implemented in their study (9). We expected to observe a sudden increase of the sulforhodamine B fluorescence signal due to dilution-induced fluorescence dequenching when content mixing happened (Fig. 1B).

Content mixing happened for only less than 1% of the docked vesicle pairs (8 out of 818, see Table S2) in the first 60 s at room temperature when L/P = 200, which is far less than the 31% yield in lipid mixing. Increasing the SNARE density in the vesicles promoted content mixing but only mildly (Fig. 1C). At L/P = 50 the yield was still less than 6% (Fig. 1C), indicating that SNAREs alone are ineffective in opening the fusion pore that is big enough for a small probe such as sulforhodamine B. Furthermore, incorporation of Syt1 in the v-vesicles virtually had no effect on content mixing of sulforhodamine B (Fig. 1D). Therefore, the results show that SNAREs alone can promote lipid mixing fairly well but are not effective in driving opening of small fusion pore. The hemifusion state is considered a bona fide fusion intermediate (16) and it is operationally defined as a state that allows lipid mixing but no content mixing (17). Thus, the results suggest that SNAREs are capable of bringing about membrane hemifusion but hardly effective in driving fusion pore opening.

Next, we explored whether Ca2+ is necessary for fusion pore opening. When 500 µM Ca2+ was added to the fusion reaction between SNARE-reconstituted vesicles without Syt1, there was no change in the fusion kinetics as shown in a cumulative plot (Fig. 1D). However, when Syt1 was present in the v-vesicles, there was a significant increase in the fusion rate. The half-time of the fusion reaction was about 12 s, which is more than 10 times faster than that in the absence of Syt1. The fusion efficiency at t = 60 s reaches as much as 15%, an order of magnitude more than that with SNAREs only (Fig. 1D). With varying Ca2+, we observed a monotonic increase of fusion efficiency for content mixing of sulforhodamine B (Fig. S1), which is different from the Ca2+-dependent lipid mixing that tails off at higher Ca2+ concentration (13, 14). Thus, the results show that Syt1 and Ca2+ are required for efficient fusion pore formation. Furthermore, through labeled SNAREs, we verified that the Ca2+/Syt1-induced content mixing was accompanied by enhanced full SNARE assembly (Fig. S2).

Fusion Pore Expansion Examined by Content Mixing of Large DNA Probes.

Fig. 2A shows the large-probe content-mixing assay. A DNA hairpin composed of 6-bp stem and poly-thymidine loop, labeled with a donor (Cy3) and an acceptor (Cy5) at the ends of the stem (Fig. S3), was encapsulated in the v-SNARE vesicle. We separately encapsulated multiple copies of unlabeled poly-adenosine target DNA strands in the t-SNARE vesicle. If the two vesicles form a sufficiently large fusion pore between them to pass DNA molecules, the two DNA molecules would hybridize, opening up the stem region of the hairpin and switching the Förster resonance energy transfer (FRET) efficiency (E) between Cy3 and Cy5 from a high to a low value. DNase treatment to the outside showed FRET signals are from DNA probes encapsulated in the vesicles (Fig. S5).

Fig. 2.
Single-molecule content-mixing assay using large DNA cargoes. (A) Schematics of the assay. Vesicles reconstituted with VAMP2 proteins (v-vesicles) and encapsulating dual-labeled DNA hairpins are immobilized on the surface of the flow cell. Vesicles reconstituted ...

To estimate the minimally required fusion pore size that allows the passage of our DNA probes, we performed a molecular dynamics simulation. The average radius of gyration (Rg) over the last 25 ns of simulation is 1.91 ± 0.53 nm for the DNA hairpin and is 1.80 ± 0.28 nm for the target DNA (Fig. 2B). The snapshots of resulting equilibrated systems are shown in Fig. 2C. From this simulation, we conclude that the radius of the pore must be at least ∼2 nm for either DNA molecule to pass through.

Fusion Pore Expansion Induced by Syt1 and Ca2+.

We now investigate expansion of fusion pore with content mixing of DNA probes. We monitored the real-time dynamics of fusion pore expansion. If fusion pore expansion happens, we would then observe a sudden increase of the donor fluorescence signal and a corresponding decrease of the acceptor fluorescence signal due to the hybridization of DNA pairs accompanying content mixing (Fig. 3A). Such anticorrelated changes of fluorescence signals could also happen when the acceptor was photobleached, which could be easily discerned by the fluorescence signal in response to the red laser illumination near the end of the experimental runs (Fig. S7).

Fig. 3.
Single-vesicle content mixing of DNA cargoes. (A) Typical time traces of fluorescence intensities (green curve for the donor and red curve for the acceptor) and the corresponding FRET efficiency (blue curve) for the DNA probes that show a content-mixing ...

For SNARE-reconstituted vesicles, regardless of the presence or absence of Syt1, we observed very few pore expansion events (1–3 out of 1,000 docked vesicle pairs at t = 60 s, see Table S3). However, when we added 500 µM Ca2+ to the preassembled t-vesicle/v-vesicle complex in the presence of Syt1, we observed appreciable content-mixing events. In the cumulative plot, the reaction occurred with a slow start (Fig. 3B, magenta curve), which was reproducible over multiple experiments (n = 6). This slow start is not due to the time it takes to open the hairpin upon hybridization to the poly-adenosine strand, which is at most 1 s (12). At 2 min, we observed 195 content-mixing events among 5,054 docked vesicle pairs (∼4.0%), about 13 times more than those without Ca2+, where we observed only 3 content-mixing events out of 1,000 docked vesicle pairs.

Compared with small pore opening detected by sulforhodamine B, pore expansion probed by content mixing of DNAs was much slower. We estimate that tens of small pore formation events are needed before allowing one pore expansion event (see Fig. 3B). The results suggest that the fusion pore opens but stays small or narrow for a prolonged period before expanding to a larger pore. It is also possible that the small pore opens and closes repetitively before finally expanding to a larger pore. Presuming that the small pore is an obligatory step toward the large expanded pore and all small pores advance to the large pore eventually, the half-time for the creation of small pore is about 13 times faster than that for pore expansion, which translates into the activation energy of ∼1.5 kcal/mole for pore expansion over small pore (Fig. 3C). Control experiments showed that the presence of DNA probes does not interfere with the kinetics of sulforhodamine B content mixing or the docking number (Figs. S8 and S9). With both probes, we found that the probability of pore opening and expansion is very low without Syt1 or Ca2+ in the early time (Fig. 3D), strongly supporting that Syt1/Ca2+ is necessary for both opening and expansion of the fusion pore.

Complexin 1 Promotes Ca2+-Triggered Fusion Pore Expansion.

Complexin 1 is a soluble presynaptic protein known to bind the SNARE complex. It is known to play activating and inhibiting roles for fast synchronous and spontaneous neurotransmitter release, respectively (6, 18, 19). To find the influence of complexin 1 on Ca2+-triggered pore opening and expansion, we included 5 µM complexin 1 in the content-mixing assays. Complexin 1 caused a 21% increase in the reaction rate of the content mixing of sulforhodamine B (Fig. 4 A and C). With DNA probes, there is overall about a factor of 2 increase of content mixing with complexin 1 (Fig. 4 B and C). Most interestingly, however, the slow phase at the early time (t < 5 s) was replaced with the fast rise of the content-mixing events. At t = 5 s, we observed about 5 times more DNA content-mixing events in the presence of complexin 1. Thus, the results show that complexin 1 is actively involved in pore formation but more extensively at the early phase of pore expansion. We note that the complexin 1 effect on membrane fusion is specific to neuronal SNAREs as it shows no change on the kinetics of yeast SNARE-dependent liposome fusion (Fig. S10). Thus, we rule out the possibility of artifacts stemming from nonspecific interaction of complexin 1 with phospholipids. Interestingly, the complexin 1 truncation mutant lacking regulatory N-terminal 40 residues (18) showed little stimulatory effect (Fig. 4D), suggesting that the N-terminal region plays a role in stimulating pore expansion.

Fig. 4.
Complexin 1 accelerates fusion pore expansion for Ca2+-triggered fusion mediated by SNARE and Syt1. Complexin 1 promotes fusion pore opening probed by small sulforhodamine B (A) only modestly, but it stimulates fusion pore expansion probed large DNA probes ...


Although lipid mixing reaches 30–50% at t = 60 s even without Ca2+ content mixing of both small sulforhodamine B and large DNAs happens for only less than a few percent of total vesicle pair population. The results suggest that the SNARE complex is a kind of incomplete or self-clamping fusogen that has the capacity to carry the fusion up to the hemifusion state or a hemifusion diaphragm (20), in line with the recent finding by Brunger and coworkers (21). These results for neuronal SNAREs are in sharp contrast with those for yeast SNAREs Sso1p, Sec9, and Snc2p, constitutively active for trafficking from golgi to plasma membrane, which are capable of carrying out pore formation and expansion highly efficiently even in the absence of other factors (12). Thus, it appears that neuronal SNAREs have evolved in such a way that other regulators including Syt1, Ca2+, and complexin 1 can control pore opening and expansion. A few in vivo and in vitro studies have suggested that the vesicles are trapped at the hemifusion state before the Ca2+ trigger (2224), consistent with our results. Meanwhile, we cannot rule out the possibility that hemifusion is an off-pathway product for fast fusion, as shown in a recent study involving both lipid and content indicators (21).

Our results from single-vesicle content-mixing assays show that cooperative action of Syt1 and Ca2+ are the main thrust for the pore opening and expansion. However, pore expansion is as much as 13 times slower than small pore opening, indicating that pore expansion is the rate-limiting step for full-collapse SNARE-dependent membrane fusion. Neurotransmitters are small molecules, even smaller than sulforhodamine B. Thus, their release would occur immediately upon small pore formation. Our results suggest that the pore size remains small for a prolonged period. We speculate that the fate of the vesicle, whether it collapses completely or return back to a free spent vesicles (kiss-and-run) to be refreshed for the next round, is determined in this period. Such a time lapse between pore opening and expansion is necessary for the coexistence of two recycling pathways during this time.

Under current experimental conditions, the yield of the DNA content-mixing reaction is as low as 4% of docked vesicles at room temperature even when activated with Ca2+ in presence of Syt1. However, these are not the consequence of random fusion events or some experimental artifacts, because without Ca2+ we observe only one to three fusion events out of 1,000 vesicle pairs. Furthermore, when the flow cell harboring docked vesicle pairs was incubated at 37 °C for 30 min in the presence of Ca2+ the yield was improved to as much as 50%, indicating that more than 50% vesicle pairs are fusion-competent (Fig. S11). We were under experimental conditions in which fusion pore opening and expansion activities are somewhat suppressed but sensitive to the changes, whereas the lipid mixing activity is efficient.

Complexin 1 is known to function in the early steps of SNARE-dependent membrane fusion. Complexin 1 can clamp the fusion at the hemifusion state (25) or a step before lipid mixing by interfering with SNARE complex formation (26), which might be de-clamped by Syt1 and Ca2+ (18, 27). We found that complexin 1 might have an additional function to modulate pore expansion. Our results show that complexin 1 has the capacity to accelerate fusion pore expansion, perhaps to favor the complete collapse fusion pathway. Understanding the detailed mechanism, at a molecular level, however, requires further investigation.

In summary, results from single-vesicle fusion assays using lipophilic probes, small cargo, and large cargoes paints unprecedentedly detailed picture of the regulation of individual fusion steps including hemifusion, fusion pore formation, and pore expansion by Syt1, Ca2+, and complexin 1. It appears that SNAREs alone are sufficient in promoting membrane hemifusion, although Syt1 and Ca2+ are required to carry over toward pore formation and expansion. Pore expansion is the late-determining step for entire membrane fusion process, which might play an important role in allowing the kiss-and-run mechanism for synaptic recycling.


Plasmid Constructs and Site-Directed Mutagenesis.

DNA sequences encoding Syntaxin 1A (Syn1A, amino acids 1–288 with three cysteines replaced by alanines), VAMP2 (amino acids 1–116 with C103 replaced by alanines), SNAP-25 (amino acids 1–206 with four native cysteines replaced by alanines), complexin 1 (amino acids 1–134), Sso1p (amino acids 185–290), Sec9 (amino acids 401–651), Snc2p (amino acids 1–115), complexin 1 N-terminal truncation (amino acid 41–134), and soluble VAMP2 (amino acids 1–96) were inserted into the pGEX-KG vector as N-terminal GST fusion proteins. We used the Quick Change site-directed mutagenesis kit (Stratagene) to generate all cysteine mutants, including Syn1A I187C, Syn1A G288C, VAMP2 G18C, and VAMP2 T116C; DNA sequences were confirmed by the Iowa State University DNA Sequencing Facility.

Protein Expression, Purification, and Labeling.

Recombinant neuronal SNARE proteins from rat: Syn1A (amino acids 1–288), SNAP-25 (amino acids 1–206), VAMP2 (amino acids 1–116), Syn1A I187C, Syn1A G288C, VAMP2 G18C, VAMP2 T116C, complexin 1 (amino acid 1–134), Sso1p (amino acids 185–290), Sec9 (amino acids 401–651), Snc2p (amino acids 1–115), complexin 1 N-terminal truncation (amino acid 41–134), and soluble VAMP2 (amino acids 1–96) were expressed as N-terminal GST fusion proteins. These fusion proteins expressed in Escherichia coli Rosetta (DE3) pLysS (Novagene). The cells were grown at 37 °C in LB medium with 100 µg/mL ampicillin until the absorbance at 600 nm reached 0.6–0.8. The cells were further grown for overnight after adding IPTG (0.5 mM final concentration) at 16 °C. We purified the proteins using glutathione–agarose chromatography. Cell pellets were resuspended in 20 mL of PBS, pH 7.4, containing 0.2 v% Triton X-100, with final concentrations of 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF), 2 mM DTT. Cells were broken by sonication in an ice bath and centrifuged at 27,000 × g for 30 min at 4 °C. The supernatant was mixed with 2 mL of glutathione-agarose beads in PBS and nutated in the cold room (4 °C) for 2 h. The proteins were then cleaved by thrombin in cleavage buffer (50 mM Tris⋅HCl/150 mM NaCl, pH 8.0) with 0.8 wt% n-octyl-D-glucopyranoside (OG) for neuronal SNAREs, or cleaved by thrombin in cleavage buffer for complexin 1, Munc18, and soluble VAMP2. The full-length Syt1 (amino acids 51–421) were expressed as C-terminal 6*Histidine-tag in an Escherichia coli BL21 Rosetta (DE3) pLysS (Novagen) and purified with the same protocol above by using Ni-NTA column. His6-tagged full-length Syt1 was eluted by elution buffer (25 mM Hepes/400 mM KCl/500 mM immidazole, 0.8 wt% OG). After removing immidazole with PD10 desalting column (GE Healthcare), the eluted Syt1 was kept in the cleavage buffer containing 0.8 wt% OG and 1 mM EDTA. DNA sequences were confirmed by the Iowa State University DNA Sequencing Facility. Purified proteins were examined with 15% (wt/vol) SDS/PAGE, and the purity was at least 85% for all proteins.

Cysteine mutants Syn1A I187C, Syn1A G288C, VAMP2 G18C, and VAMP2 T116C were labeled with the fluorescence labels Cy3 and Cy5 maleimide (Amersham). The reaction mixture, with the dye to protein ratio of more than 5:1, was left overnight at 4 °C. The free dye was removed using the PD-10 desalting column (Amersham). All labeled proteins were analyzed by the SDS-polyacryamide gel, and the labeling efficiencies were over 80% for all proteins.

Membrane Reconstitution.

The lipid molecules used in this study are 1,2-dioleoyl-sn-glycero-3-phospho-L-serine (DOPS), 1-palmitoyl-2-oleoyl-snglycero-3-phosphocholine (POPC), phosphatidylinositol-4,5-bisphosphate (PIP2, from porcine brain), cholesterol, and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamin-N-(biotinyl) (biotin-DPPE). All lipids were obtained from Avanti Polar Lipids.

For the NN/CC FRET assay, the molar ratios of lipid species were 15:63:20:2 (DOPS:POPC:cholesterol:PIP2) for the t-vesicles, and 5:75:20 (DOPS:POPC:cholesterol) for the v-vesicles. The lipid mixture was first completely dried and then hydrated by fusion buffer (25 mM Hepes/100 mM KCl, pH 7.4). Protein-free large unilamellar vesicles (∼100 nm in diameter) were prepared by extrusion through polycarbonate filters (Avanti Polar Lipids).

To reconstitute labeled SNARE proteins into liposomes, Syn1A I187C-Cy3 or Syn1A G288C-Cy3 and SNAP-25, in a molar ratio of 1:1.5, were premixed, and the mixture was left at room temperature for 1 h to form the complex before the reconstitution. VAMP2 G18C-Cy5 or VAMP2 T116C-Cy5 and Syt1 were mixed at molar ratio of 1:1 containing 1 mM EDTA. For membrane reconstitution, SNARE proteins were mixed with vesicles at the lipid to protein molar ratio of 200 with ∼0.8 wt% OG in the buffer at 4 °C for 15 min. The mixture was diluted two times with dialysis buffer (25 mM Hepes/100 mM KCl, pH 8.0), and this diluted mixture was then dialyzed in 2 l dialysis buffer at 4 °C for overnight.

For the lipid-mixing assay, the molar ratios of lipid species were 15:61:20:2:2 (DOPS:POPC:cholesterol:PIP2:DiI) for the t-vesicles, and 5:73:20:2 (DOPS:POPC:cholesterol:DiD) for the v-vesicles, respectively. The above protein reconstitution process was used for unlabeled SNARE proteins.

For the large-DNA content-mixing assay, the molar ratios of lipid species were 15:63:20:2 (DOPS:POPC:cholesterol:PIP2) for the t-vesicles and 5:74.9:20:0.1 (DOPS:POPC:cholesterol: biotin-DPPE) for the v-vesicles. The dried lipid film was resuspended with T50 buffer (10 mM Tris⋅HCl/50 mM NaCl, pH 8.0) containing 5 μM target DNAs or 2 μM Cy3/Cy5 dual-labeled DNA probes (Integrated DNA Technologies), separately. After five freeze–thaw cycles, unilamellar vesicles were extruded through polycarbonate filters (100-nm pore size, Avanti Polar Lipids) at least 30 times. To reconstitute content-mixing vesicles, the membrane proteins and vesicles were mixed together with the desired lipid to protein ratio of 200. Then the mixture containing ∼0.8 wt% OG in the buffer was kept at 4 °C for 20 min. The mixture was diluted two times with dialysis buffer (25 mM Hepes/100 mM KCl, pH 7.4) for t-vesicles, and then the diluted t-vesicle mixture was dialyzed in 2 L of pH 7.4 dialysis buffer at 4 °C overnight; for v-vesicles, the mixture was diluted twice with pH 4.5 dialysis buffer and dialyzed in 2 L of pH 4.5 dialysis buffer at 4 °C for 4 h, then changed to pH 7.4 dialysis buffer overnight.

For small sulforhodamine B content indicators, the lipid composition is the same as that used for the content-mixing assay for DNA probes. The lipid mixture was first completely dried and then hydrated by dialysis buffer, but the v-vesicles were in the presence of 50 mM sulforhodamine B (Invitrogen). After five freeze–thaw cycles, protein-free large unilamellar vesicles (∼100 nm in diameter) were prepared by extrusion through polycarbonate filters (Avanti Polar Lipids). Free sulforhdamine B was removed using the PD-10 desalting column (Amersham) after dialysis.

Ensemble N-Terminal End-to-N-Terminal End (NN)-FRET and C-Terminal End-to-C-Terminal End (CC)-FRET.

Reconstituted t- and v-vesicles were mixed at a ratio of 1:1. The fluorescence intensity was monitored in two channels with the excitation wavelength of 545 nm and emission wavelengths of 570 and 668 nm, respectively. Fluorescence changes were recorded with the Varian Cary Eclipse model fluorescence spectrophotometer using a quartz cell of 100 μL with a 2-mm path length. All measurements were performed at 35 °C.

Lipid Mixing FRET Assays.

Reconstituted t- and v-vesicles were mixed at a ratio of 1:1. The final lipid concentration in the reaction is 0.05 mM. The fluorescence intensity was monitored in two channels with the excitation wavelength of 530 nm and emission wavelengths of 570 and 670 nm for DiI and DiD fluorescence dye pair, respectively. Fluorescence changes were recorded with the Varian Cary Eclipse model fluorescence spectrophotometer using a quartz cell of 100 μL with a 2-mm path length. All measurements were performed at 35 °C.

Single-Vesicle Lipid and Content-Mixing Assay.

After surface coating of polyethylene glycol (PEG) molecules to eliminate nonspecific binding of vesicles and DNA molecules, the quartz slide was assembled into a flow chamber and coated with neutravidin (0.2 mg/mL). Following a 30-min incubation at room temperature, the v-SNARE vesicles were immobilized on the PEG-coated surface. After two rounds of washing in 200 µL of buffer, the t-SNARE vesicles (100∼200 nM) were injected into the flow chamber for the 30-min predocking at room temperature (∼25 °C), with or without accessory proteins. After washing out the free t-SNARE vesicle, the reaction was conducted at 37 °C in the buffer (25 mM Hepes/100 mM KCl, pH 7.4) for 30 min with or without 500 µM Ca2+. The detail of the single-vesicle lipid-mixing assay was reported in our previous work (11).

For content-mixing assay, the FRET measurements by a total internal reflection (TIR) fluorescence microscope were performed with an oxygen scavenger system. The details of TIR fluorescence microscope imaging and single molecule data analysis have been reported in our previous yeast content-mixing paper (11). In real-time imaging of large DNA probes, t-SNARE vesicles were injected into the channel immobilized by v-SNARE vesicles containing DNA probe. After 30-min incubation in room temperature (∼25 °C), 500 µM Ca2+ was injected at a speed of 33 µL/s by a motorized syringe pump. To exclude the possibility of photobleaching, acceptor Cy5 was excited by red laser both before Ca2+ injection and in the end of the detection. An oxygen scavenger system (0.1 mg/mL glucose oxidase, 0.02 mg/mL catalase, and 0.4 wt% β-D-glucose and Trolox) was used for eliminating single-molecule blinking events and reducing the likelihood of photobleaching. For the real-time imaging of small sulforhodamine B content release, the same PEG surface was used as above. After the surface was coated with neutravidin, the sulforhodamine B containing v-SNARE vesicles were immobilized on the PEG-coated surface. After two rounds washing in 1 mL of buffer, empty t-SNARE vesicles were injected into the channel incubating with v-SNARE vesicles. After 30-min incubation at room temperature (∼25 °C), Ca2+ was injected to trigger fusion with the same injection speed used for large DNA content mixing.

Molecular Dynamics (MD) Simulations.

Two independent MD simulations were performed on single-strand DNA with different sequences corresponding to Cy3/Cy5 dual-labeled DNA hairpin and unlabeled target DNA to probe their radius of gyration in equilibrium. Both starting structures were generated using 3D-DART (28) and VMD (29) with CHARMM27 topology (30). The DNA hairpin was solvated using 22043 TIP3P water molecules and neutralized with 100 mM KCl in a water box of 89 Å × 89 Å × 89 Å. The unlabeled target DNA was immerged in a simulation box of 100 Å × 100 Å × 100 Å including a POPC lipid patch, water, and 100 mM KCl. The DNA was kept in close with lipid using a virtual bond connecting the sugar ring oxygen of A30 at 3′ end of target DNA and the phosphate of a lipid. The all-atom MD simulations were performed using the developing version of program NAMD 2.7 (31) with the TIP3P model for explicit water and the CHARMM27 force field including the CMAP correction (30). In all simulations, periodic boundary condition was applied with a time step of 1 fs. Nonbonded energies were calculated using particle mesh Ewald full electrostatics and a smooth (1.0–1.2) cutoff of the van der Waals energy. Constant temperature was maintained using Langevin thermostat with a damping coefficient of 1 ps−1. A Nosé–Hoover Langevin piston barostat was used to maintain a constant pressure with a period of 200.0 fs and damping timescale of 100.0 fs. Each system was simulated for 75 ns, where the last 25 ns of data were taken to calculate the radius of gyration.

The DNA hairpin probe composes of a 6-bp stem and a poly-thymidine (T20) loop. The unlabeled target DNA has a complementary sequence of poly-adenosine (A30). The DNA hairpin stem was modeled with B-DNA-type geometry with a random conformation for the poly-T loop. K+ and Cl ions were added to the water box to obtain a 100 mM KCl solution. Additional K+ ions were added to neutralize the system. The resulting system contains 67,229 atoms including DNA, water, and ions. The starting structure of target DNA was modeled by removing the complementary strand from a B-form double-helix DNA. A patch of POPC with 274 lipids was added to the system with the membrane normal along the z axis. The distance between the oxygen of A30 at 3′ end of target DNA and the phosphate of one of lipids was harmonically restrained to be 10 Å through a virtual bond. Such restraints mimic the 3′-Cholesteryl TEG used in the experiment to prevent the target DNA from moving away from membrane while allowing the target DNA diffuse along with the bonded lipid. Solvent was then added to both sides of the membrane, and the system was neutralized with 100 mM KCl using the Solvate and Auto ionize plugins of VMD. The simulated system has about 183,000 atoms.

The simulated system of DNA hairpin was subject to 4,000 steps of conjugate gradient minimization and 500-ps NPT (constant pressure and temperature) equilibration with DNA harmonically restrained. The system then was subject to 10-ns annealing at 500 K with the 6-bp stem harmonically restrained in the NVT (constant volume and temperature) ensemble. After cooling the system down, the 75-ns production run was performed in the NVT ensemble at T = 310 K. The simulated system of unlabeled target DNA was subject to 4,000 steps of conjugate gradient minimization first. Then, lipid tails were melted in a 1-ns NVT simulation at 310 K during which all other atoms were fixed. The system was then equilibrated in the NPT simulation at 1 atm and 310 K for 1 ns with DNA harmonically restrained and followed by 1-ns simulation with all restraint released. Finally, the 75-ns production run was performed in the NPT ensemble with constant area of the lipid bilayer and a constant normal pressure of 1 atm.

Supplementary Material

Supporting Information:


We thank Dr. C. Joo for help with preparing illustrations, and Dr. E. Hui and Dr. M. B. Jackson for fruitful discussions. This work was supported by National Institutes of Health Grants R21 GM074526 (to T.H.) and R01 GM051290 (to Y.-K.S.). T.H. is an investigator of the Howard Hughes Medical Institute.


The authors declare no conflict of interest.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1218818110/-/DCSupplemental.


1. Jahn R, Lang T, Südhof TC. Membrane fusion. Cell. 2003;112(4):519–533. [PubMed]
2. Rizo J, Rosenmund C. Synaptic vesicle fusion. Nat Struct Mol Biol. 2008;15(7):665–674. [PMC free article] [PubMed]
3. Chernomordik LV, Kozlov MM. Protein-lipid interplay in fusion and fission of biological membranes. Annu Rev Biochem. 2003;72:175–207. [PubMed]
4. Jackson MB, Chapman ER. The fusion pores of Ca2+-triggered exocytosis. Nat Struct Mol Biol. 2008;15(7):684–689. [PMC free article] [PubMed]
5. Zhang Q, Li Y, Tsien RW. The dynamic control of kiss-and-run and vesicular reuse probed with single nanoparticles. Science. 2009;323(5920):1448–1453. [PMC free article] [PubMed]
6. Yoon TY, et al. Complexin and Ca2+ stimulate SNARE-mediated membrane fusion. Nat Struct Mol Biol. 2008;15(7):707–713. [PMC free article] [PubMed]
7. Yoon TY, Okumus B, Zhang F, Shin YK, Ha T. Multiple intermediates in SNARE-induced membrane fusion. Proc Natl Acad Sci USA. 2006;103(52):19731–19736. [PMC free article] [PubMed]
8. Diao JJ, et al. Single-vesicle fusion assay reveals Munc18-1 binding to the SNARE core is sufficient for stimulating membrane fusion. ACS Chem Neurosci. 2010;1(3):168–174. [PMC free article] [PubMed]
9. Kyoung M, et al. In vitro system capable of differentiating fast Ca2+-triggered content mixing from lipid exchange for mechanistic studies of neurotransmitter release. Proc Natl Acad Sci USA. 2011;108(29):E304–E313. [PMC free article] [PubMed]
10. Diao JJ, Ishitsuka Y, Bae WR. Single-molecule FRET study of SNARE-mediated membrane fusion. Biosci Rep. 2011;31(6):457–463. [PubMed]
11. Diao JJ, et al. A single vesicle-vesicle fusion assay for in vitro studies of SNAREs and accessory proteins. Nat Protoc. 2012;7(5):921–934. [PubMed]
12. Diao JJ, et al. A single-vesicle content mixing assay for SNARE-mediated membrane fusion. Nat Commun. 2010;1:54. [PMC free article] [PubMed]
13. Lai Y, Shin YK. The importance of an asymmetric distribution of acidic lipids for synaptotagmin 1 function as a Ca2+ sensor. Biochem J. 2012;443(1):223–229. [PubMed]
14. Lee HK, et al. Dynamic Ca2+-dependent stimulation of vesicle fusion by membrane-anchored synaptotagmin 1. Science. 2010;328(5979):760–763. [PMC free article] [PubMed]
15. Kim JY, et al. Solution single-vesicle assay reveals PIP2-mediated sequential actions of synaptotagmin-1 on SNAREs. EMBO J. 2012;31(9):2144–2155. [PMC free article] [PubMed]
16. Xu Y, Zhang F, Su Z, McNew JA, Shin YK. Hemifusion in SNARE-mediated membrane fusion. Nat Struct Mol Biol. 2005;12(5):417–422. [PubMed]
17. Kemble GW, Danieli T, White JM. Lipid-anchored influenza hemagglutinin promotes hemifusion, not complete fusion. Cell. 1994;76(2):383–391. [PubMed]
18. Xue M, et al. Distinct domains of complexin I differentially regulate neurotransmitter release. Nat Struct Mol Biol. 2007;14(10):949–958. [PubMed]
19. Maximov A, Tang J, Yang XF, Pang ZPP, Südhof TC. Complexin controls the force transfer from SNARE complexes to membranes in fusion. Science. 2009;323(5913):516–521. [PMC free article] [PubMed]
20. Hernandez JM, et al. Membrane fusion intermediates via directional and full assembly of the SNARE complex. Science. 2012;336(6088):1581–1584. [PMC free article] [PubMed]
21. Diao JJ, et al. Synaptic proteins promote calcium-triggered fast transition from point contact to full fusion. eLife. 2012;1:e00109. [PMC free article] [PubMed]
22. Chicka MC, Hui EF, Liu HS, Chapman ER. Synaptotagmin arrests the SNARE complex before triggering fast, efficient membrane fusion in response to Ca2+ Nat Struct Mol Biol. 2008;15(8):827–835. [PMC free article] [PubMed]
23. Zhang XD, Kim-Miller MJ, Fukuda M, Kowalchyk JA, Martin TFJ. Ca2+-dependent synaptotagmin binding to SNAP-25 is essential for Ca2+-triggered exocytosis. Neuron. 2002;34(4):599–611. [PubMed]
24. Walter AM, Groffen AJ, Sørensen JB, Verhage M. Multiple Ca2+ sensors in secretion: Teammates, competitors or autocrats? Trends Neurosci. 2011;34(9):487–497. [PubMed]
25. Schaub JR, Lu X, Doneske B, Shin YK, McNew JA. Hemifusion arrest by complexin is relieved by Ca2+-synaptotagmin I. Nat Struct Mol Biol. 2006;13(8):748–750. [PubMed]
26. Giraudo CG, Eng WS, Melia TJ, Rothman JE. A clamping mechanism involved in SNARE-dependent exocytosis. Science. 2006;313(5787):676–680. [PubMed]
27. Tang J, et al. A complexin/synaptotagmin 1 switch controls fast synaptic vesicle exocytosis. Cell. 2006;126(6):1175–1187. [PubMed]
28. van Dijk M, Bonvin AM. 3D-DART: A DNA structure modelling server. Nucleic Acids Res. 2009;37(Web Server issue):W235–W239. [PMC free article] [PubMed]
29. Humphrey W, Dalke A, Schulten K. VMD: Visual molecular dynamics. J Mol Graph. 1996;14(1):33–38. 27–38. [PubMed]
30. Mackerell AD, Jr, Feig M, Brooks CL., 3rd Extending the treatment of backbone energetics in protein force fields: Limitations of gas-phase quantum mechanics in reproducing protein conformational distributions in molecular dynamics simulations. J Comput Chem. 2004;25(11):1400–1415. [PubMed]
31. Phillips JC, et al. Scalable molecular dynamics with NAMD. J Comput Chem. 2005;26(16):1781–1802. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...