The three proteins synaptosomal-associated protein 25 (SNAP-25), Syntaxin-1a and vesicle-associated membrane protein (VAMP-2) are collectively called SNAREs (soluble N-ethylmaleimide-sensitive factor attachment protein receptors). By assembling into an exocytic complex, the three SNAREs help in catalyzing membrane fusion. Due to lack of probes that adequately reconstitute the intracellular behavior of endogenous SNAREs, the dynamics of SNARE complexes in living cells is poorly understood. Here we describe a new FRET-based probe, called Cerulean-SNAP-25-C4 (CSNAC), that can track the conformational changes undergone by SNAP-25 as exocytic complexes assemble. The fluorescent protein Cerulean was attached to the N terminus and served as a FRET donor. The biarsenical dye FlAsH served as a FRET acceptor and was attached to a short tetracysteine motif (C4) motif inserted into the so-called linker domain of SNAP-25. CSNAC reported successive FRET changes when first Syntaxin-1a and then VAMP-2 were added in vitro. Small tetracysteine insertions used as a FRET acceptor are expected to have less steric hindrance than previously used GFP-based fluorophores. We propose that genetically-encoded tetracysteine tags can be used to study regulated SNARE complex assembly in vivo.

The soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptors (SNAREs) Syntaxin-1a, VAMP-2 and SNAP-25 form the complex required for exocytosis (1). SNAP-25 is unstructured in solution, but when the protein binds Syntaxin, or combines with both Syntaxin and VAMP to form an exocytic SNARE complex, then the two SNARE motifs of SNAP-25 (H1 and H2 in Figure 1A) form parallel α-helices and their two N-terminals are brought together by intra-molecular folding (2,–3). Because Syntaxin-1a and SNAP-25 reside on the target membrane and VAMP-2 resides on the vesicular membrane, the assembly of the three SNARE proteins in an exocytic SNARE complex provides the free energy that pulls two membranes together and drives membrane fusion in the cell (4). Exocytic complex assembly is subject to control by regulatory proteins, such as Munc-18, synaptotagmin and complexin (5-7).

Figure 1. CSNAC design and labeling with FlAsH. (Click to enlarge)

Although SNARE assembly has been extensively studied by structural and biochemical methods, little is known about the dynamics of SNARE complex in vivo. By attaching appropriate fluorophores near the two terminals of SNAP-25 α-helices, SNARE complex assembly can be studied by intra-molecular fluorescence resonance energy transfer (FRET). This strategy was employed by labeling of a pair of cysteines in the C-terminal ends of the two SNARE domains of SNAP-25 with Cy3, a donor, and Cy5, an acceptor of fluorescence (8). Because the mixture of Cy3 and Cy5 was used to label a pair of cysteines located on the same molecule of SNAP-25, a random incorporation of a donor or an acceptor in SNAP-25 is possible, limiting the use of malemide-functionalized dyes when many rather than single molecules are being studied. Furthermore, the introduction of this probe in live cells required microinjections (8). Another FRET reporter, SCORE, used a brighter version of the cyan fluorescent protein Cerulean (9,–10) as a FRET donor and the yellow fluorescent protein Venus as a FRET acceptor. A similar reporter used Citrine in place of Venus (11). However, we found that the use of SCORE in vitro is limited by the tendency of bacterially-expressed SCORE proteins to quickly lose their FRET response after storage and freeze-thaw. Slow inactivation of Venus seen in vitro may also impose a potential problem of inaccurate FRET measurements in vivo.

Given that SNAREs interact with multiple regulatory factors, the insertions of large fluorophores such as Venus and Citrine in the middle of the coding region of SNAP-25 may disrupt protein-protein interactions and result in accumulation of aberrant SNARE complexes in vivo. By replacing Venus with a smaller chromophore, we hoped to improve the stability and reduce the size of the probe. Here we used genetically encoded Cerulean as a FRET donor and the membrane-permeable biarsenical dye FlAsH, a small fluorophore that is expected to have less steric hindrance than Venus, as a FRET acceptor. We propose that this SNAP-25 probe can be used for real-time studies of SNARE complex assembly in vitro and discuss its potentially unique application in living cells.

Materials and methods

Plasmid construction

To create the mammalian CMV-driven CSNAC-expressing plasmid, the coding region of rat SNAP-25 was amplified by PCR and subcloned into BsrG1/XhoI sites of pECFP-C1 vector (Clontech). CFP sequence was replaced with the Cerulean sequence by PCR. The nucleotide linker encoding the tetracysteine motif (EFDLAGGCCPGCCGGGEF) was subcloned into pre-designed EcoR1 sites at the following positions of the SNAP-25 coding region: R142 (CSNAC1), G132 (CSNAC2), E125 (CSNAC3) and A99 (CSNAC4). For bacterial expression of GST–CSNAC fusion proteins, the sequence encoding CSNAC proteins was amplified by PCR and subcloned into NcoI/XhoI sites of the pNGST vector (a gift of Dr. Eric Gouaux). To minimize the nonspecific binding of FlAsH, four endogenous cysteines of SNAP-25 were replaced with serines using the QuickChange mutagenesis kit (Stratagene, USA). The deltaH2 deletion mutant of the CSNAC2 construct was created by replacing of the nucleotide sequence GTA encoding Val-138 with a stop codon TAA. s-VAMP and s-Syntaxin were previously described (9).

Protein expression, purification and labeling with FlAsH

GST- Cerulean-SNAP-25-C4 (CSNAC) and GST-SCORE proteins were expressed in the BL21 bacterial strain (Stratagene, USA). Briefly, 200 mL of bacterial culture was grown to OD₆₀₀ 0.8 at 37°C, chilled to 20°C and induced overnight with 0.5 mM IPTG at 20°C. The cells were disrupted in the ice-cold breaking buffer (25 mM HEPES, pH 7.4, 0.1 M KCl, 10% glycerol, 1 mM β-mercaptoethanol, protease inhibitors), using the EmulsiFlex cell disruptor (Avestin, Canada). Unbroken cells were removed by centrifugation for 1 h at 20,000xg, and supernatants were incubated for 2–4 h at 4°C with 0.5 mL of glutathione beads (Amersham, USA). Beads were washed five times with 10 mL of the breaking buffer containing 1% octyl-β-D-Glucopyranoside (Anatrace, USA). To cleave off CSNAC from the GST moiety, beads containing GST-CSNAC were incubated for 2 h at room temperature in 1 mL breaking buffer containing containing 0.5% octyl-β-D-Glucopyranoside and 4 µL of HTC-thrombin (Haematologic Technologies Inc, USA). HTC-thrombin was inactivated with 0.5 mM AEBSF (Calbiochem, USA), and protein concentration was determined by Bradford analysis (Bio-Rad, USA). To label CSNAC with FlAsH, 200 µL of 40 µM protein solution (2 mg/mL) were incubated at room temperature for 1 h with 40 µM FlAsH (Invitrogen, USA) in the final volume of 200 µL. The labeling reaction was transferred to 4°C and continued overnight. Labeling was stopped by addition of 25 µM 2,3-dimercaptopropanol and dialyzed three times against the breaking buffer containing 0.25% octyl-β-D-Glucopyranoside and 0.5 mM Tris-[2-carboxyethyl]phosphine (TCEP). CSNAC protein solution was aliquoted, snap-frozen in liquid nitrogen and stored at -80°C. The presence of 0.1%–0.25% octyl-β-D-Glucopyranoside in the stock solution of CSNAC proteins is essential for FRET activity. It is possible that the presence of detergent inhibits self-aggregation of CSNAC (and SNAP-25). FlAsH-labeled recombinant proteins remain stable at 4°C for at least 24 h, and at -80°C for up to 6 months, although a partial loss of FRET activity may occur after extended storage and freeze-thaw. Fluorescence measurements were conducted as described in Supplementary Materials.

Formation of SDS-resistant SNARE complexes

4 µM unlabelled CSNAC2 was incubated overnight at 4°C with the increasing concentrations (5, 10, 15, 20, 40 µM) of s-Syntaxin and s-VAMP. The molar ratio of s-Syntaxin and s-VAMP were kept equal. The total volume of protein mixture was 50 µL and the incubation buffer contained 25 mM HEPES, pH 7.4, 10% glycerol, 0.1 M KCl and 0.5 mM TCEP. 50 µL protein samples were mixed with 17 µL of 4xSDS buffer (240 mM Tris, pH 6.8, 40% glycerol, 8% SDS, and 9% β-mercaptoethanol), incubated for 5 min at room temperature or at 95°C, and analyzed on a 4%–12% NuPAGE Novex Bis-Tris mini gel (Invitrogen). Proteins were stained with Coomassie Blue.

Results and discussion

Tsien and colleagues have introduced a method for a site-specific labeling of recombinant proteins with fluorescence by inserting a short amino acid sequence containing two pairs of cysteines, termed the tetracysteine motif (C4) that binds with high affinity a derivative of fluorescein FlAsH (12). Because the emission spectrum of Cerulean (peak at 475 nm) overlaps with the excitation spectrum of FlAsH (peak at 510 nm), Cerulean and FlAsH can be used as a FRET pair (Figure 1B). Cerulean is significantly brighter and has a higher quantum yield and a higher extinction coefficient than the cyan fluorescent proteins (eCFP) and hence can improve the sensitivity of FRET measurements. We fused Cerulean to the N terminus of SNAP-25, and inserted a C4 motif at different positions in the linker domain that connects the H1 and H2 SNARE domains of SNAP-25, (Figure 1A). Cerulean-SNAP-25-C4 (CSNAC) proteins were expressed in bacteria as N-terminal GST fusion constructs, GST was cleaved off, and the proteins were fluorescently labeled by incubation with FlAsH. The binding of FlAsH to CSNAC is concentration-dependent (Figure 1C-F). When all C4 motifs carry FlAsH, and if all Cerulean moieties are fluorescent, then each CSNAC carries one donor and one acceptor molecule.

We tested the effect of s-Syntaxin, the SNARE motif of Syntaxin-1a (corresponds to residues 180–266 in rat Syntaxin sequence) on FRET. Figure 2A shows the emission spectra of Cerulean-SNAP-25 that lacks the C4 motif incubated in the buffer alone (red line) or together with s-Syntaxin (black circles). When fluorescence was excited at 433 nm, Cerulean-SNAP-25 displayed the familiar emission spectrum of Cerulean with its major and minor peaks at 475 nm and 510 nm. Addition of s-Syntaxin had no effect on the shape of spectrum and the intensity of fluorescence. When the experiment was performed using CSNAC2 in place of Cerulean-SNAP-25, there was a third peak at 530 nm that coincided with the 530 nm emission peak of FlAsH. The addition of s-Syntaxin increased the 530 nm peak and decreased the 475 nm peak (Figure 2B, circles). We take the change to represent increased intra-molecular FRET between Cerulean and FlAsH. The result is consistent with the idea that Syntaxin brings the N-terminals of the two SNARE motifs of SNAP-25 into close proximity (2,–3,9). The opposite behavior was seen with CSNAC4 where the C4 motif was inserted near the C terminal of H1 (Figure 1A). After being labeled with FlAsH, this protein showed a pronounced fluorescence peak at 530 nm even in the absence of s-Syntaxin. Adding s-Syntaxin caused the 530 nm peak to diminish, and the 475 nm peak to increase (Figure 2C, circles). Clearly the fluorescence change induced by s-Syntaxin depends strongly on the position of the C4 motif. We suggest that s-Syntaxin forces the two fluorophores in CSNAC4 apart, possibly because the H1 and H2 SNARE motifs change from flexible random coils to fully extended rods (13) that keep N- and C-terminals “at arm's length” (Figure 2E). The graph in Figure 2D summarizes results over a range of s-Syntaxin concentrations and with C4 motifs at other positions. The magnitude of the FRET changes saturated as the s-Syntaxin concentration was raised. It was positive with CSNAC1–3, largest for CSNAC2 and negative with CSNAC4. These effects were reproduced in another, independent set of CSNAC protein preparations. In the absence of s-Syntaxin, CSNAC4 showed the highest FRET ratio among the four variants (Figure 2D, 0 µM s-Syntaxin), possibly because the peptide chain separating Cerulean and FlAsH was the shortest (Figure 1A).