Modifications in the C Terminus of the Synaptosome-associated Protein of 25 kDa (SNAP-25) and in the Complementary Region of Synaptobrevin Affect the Final Steps of Exocytosis*

Fusion proteins made of green fluorescent protein coupled to SNAP-25 or synaptobrevin were overexpressed in bovine chromaffin cells in order to study the role of critical protein domains in exocytosis. Point mutations in the C-terminal domain of SNAP-25 (K201E and L203E) produced a marked inhibition of secretion, whereas single (Q174K, Q53K) and double mutants (Q174K/Q53K) of amino acids from the so-called zero layer only produced a moderate alteration in secretion. The importance of the SNAP-25 C-terminal domain in exocytosis was also confirmed by the similar effect on secretion of mutations in analogous residues of synaptobrevin (A82D, L84E). The effects on the initial rate and magnitude of secretion correlated with the alteration of single vesicle fusion kinetics since the amperometric spikes from cells expressing SNAP-25 L203E and K201E and synaptobrevin A82D and L84E mutants had lower amplitudes and larger half-width values than the ones from controls, suggesting slower neurotransmitter release kinetics than that found in cells expressing the wild-type proteins or zero layer mutants of SNAP-25. We conclude that a small domain of the SNAP-25 C terminus and its counterpart in synaptobrevin play an essential role in the final membrane fusion step of exocytosis.

The SNAP 1 (soluble NSF attachment protein) receptor (SNARE) hypothesis (1) has been crucial to our understanding of the molecular machinery responsible for exocytosis. The molecular events taking place during the exocytotic fusion of cellular and vesicular membranes should provide fundamental information about the mechanism of membrane fusion common to different membrane trafficking processes (2). The assembly of a ternary complex formed by the plasma membrane proteins syntaxin (3) and SNAP-25 (4) and the vesicle-associated protein synaptobrevin (5) is considered to be one of the molecular events driving vesicle priming, involving maturation steps needed to promote the apposition and final fusion of membranes (6). A heptad repeat structural motif typical of coiledcoils forming proteins is present and could be the basis for the formation of the core complex. This was confirmed by the observed increase in ␣-helix content following assembly of the complex (7,8). In recent years, precise structural studies on the nature of the complex (9,10) have shown that single fragments of 60 -70 residues of syntaxin and synaptobrevin together with two segments of SNAP-25 form helices in a four-stranded coil. Although most of the interactions between these helices are hydrophobic, there is a polar layer embedded in the middle of this rod-shaped structure formed by three glutamines and one arginine (zero layer), which is believed to be critical for SNARE complex formation. In addition, functional studies using specific neurotoxins (11), antibodies against critical domains (12), peptides imitating regions of SNAREs (13), and overexpression of altered forms of these proteins (14,15) have revealed important details about the participation of SNAREs in the exocytotic process. These and other studies suggest that SNARE interaction results in a tighter structure reducing the repulsive energy barrier between vesicular and plasma membranes. Assembly of the complex may proceed from the distal N-terminal domains of SNAREs assembled in parallel and, in "zipperlike" fashion, end with the interaction of C-terminal domains relatively close to the fusion pore. Even though there is not direct evidence for this model, a number of authors suggest that SNAREs assembly is directly linked to membrane fusion (16,17).
We have overexpressed several mutants of SNAP-25 and synaptobrevin critical domains in adrenal chromaffin cells and studied their effect on catecholamine secretion. Our results indicate that specific residues at the SNAP-25 C-terminal domain and the equivalent synaptobrevin domain influence vesicle fusion kinetics, suggesting a role in the very last events of the exocytotic process. By contrast, zero layer amino acid mutants, despite their ability to negatively affect the formation and stability of SNAREs complexes assembled in vitro, did not produce strong alterations.

EXPERIMENTAL PROCEDURES
Generation of Constructs of GFP Coupled to SNAP-25 or Synaptobrevin-To produce an in-frame fusion of SNAP-25 to the C terminus of GFP, the cDNA corresponding to the SNAP-25a isoform (18) was cloned into the expression vector pEGFP-C3 (CLONTECH, Palo Alto, CA) as previously described (14). The generation of the nine amino acids Cterminal deletion (⌬9) and the point mutation L203E has also been described (14). The strategy for generating the other C-terminal mutants was based in the presence of a TaiI site at the 5Ј-end of the region containing the nucleotides to be mutated and a BamHI site at the 3Ј-end. Complementary oligonucleotides carrying the desired mutations and the mentioned restriction sites were annealed and cloned into the corresponding construct in place of the original sequence. The point mutation Q53K was generated by PCR of a fragment corresponding to the N terminus of SNAP-25 using the GFP-SNAP-25 construct as template, a sense primer (5Ј-CGGCATGGACGAGCTGTACA-3Ј) corresponding to the C terminus of GFP and close to the junction between GFP and SNAP-25 and an antisense primer, which included the sequence to be mutated (indicated in lower characters) (5Ј-CGACAC-GATCGAGTTGTTCTCCtttTTCAT-CCAACATAACC-3Ј). The PCR product was digested with XhoI and PvuI and used to replace the equivalent fragment into the GFP-SNAP-25 construct, which had been digested with the same enzymes. The point mutation Q174K was generated by PCR of a fragment corresponding to the C terminus of SNAP-25 with a sense primer containing the sequence to be mutated (5Ј-CAATGAGATCGATACAaAGAATCGCCAGATCG-3Ј) and an antisense primer downstream of the C terminus of SNAP-25 (5Ј-CTA-CAAATGTGGTATGGCTG-3Ј). The amplified DNA carried an internal ClaI site, close to the 5Ј-end, and a BamHI site at the 3Ј-end. These enzymes were used to substitute the original SNAP-25 sequence by the modified one in the GFP-SNAP-25 construct. The double mutant Q53K/ Q174K was obtained by combining the single ones through enzyme restriction and ligation reactions.
To produce an in-frame fusion of synaptobrevin to the N terminus of GFP, the coding region corresponding to synaptobrevin II (19) was amplified by PCR with the following primers: 5Ј-GCCGAATTCCCGC-CATGTCGGCTACC-3Ј (sense) and 5Ј-GCCGGATCCGAGC-TGAAGTA-AACGATGATG-3Ј (antisense). The PCR product was digested with EcoRI and BamHI and cloned into the same sites of the expression vector pEGFP-N1 (CLONTECH, Palo Alto, CA). The strategy for generating the synaptobrevin mutants (A81D, A82D and L84E) was based in the presence of an SpeI site at the 5Ј-end of the region containing the nucleotides to be mutated and the previously mentioned BamHI site at the 3Ј-end. Adequate sense primers carrying the desired mutations and the SpeI site as well as the above antisense primer with the BamHI site were used in PCR amplifications. The corresponding products were used to substitute the original synaptobrevin sequence. All the introduced cassettes were sequenced to check that only the desired mutations have been produced.
Chromaffin Cell Preparation, Culture, and Infection-Chromaffin cells were prepared from bovine adrenal glands by collagenase digestion and further separated from debris and erythrocytes by centrifugation on Percoll gradients as described elsewhere (13). Cells were maintained in monolayer cultures using Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 10 M cytosine arabinoside, 10 M 5-fluoro-2Ј-deoxyuridine, 50 IU/ml penicillin, and 50 g/ml streptomycin. Cells were harvested at a density of 25,000 cells/cm 2 in 35-mm Petri dishes.
Primary cultures of chromaffin cells were infected with a Herpes Simplex virus (HSV-1) amplicon containing the constructs described above. For this purpose they were transferred from the pEGFP-C3 or pEGFP-N1 vectors to the pHSVpUC vector (20). The packaging of the different helper viruses (HSV-1 IE2 deletion mutant 5dl 1.2) was carried out as previously described (21). Efficiency of virus infection was determined by fluorescent microscopy, using serial dilutions of purified virus. The dilution producing 15-20% infection efficiency (usually 30 -40 l virus per 35-mm plate containing 1 ml of medium) was chosen in further experiments. GFP fluorescence was observed 1 day after infection and persisted, at least, during the two following days. Amperometric measurements were carried out during this time.
Amperometric Studies of Secretion in the Infected Cells-To study secretory activity from control non-infected and fluorescence-emitting cells expressing the different constructs, culture media was replaced by Krebs/HEPES basal solution with the following composition (in mM): NaCl 134, KCl 4.7, KH 2 PO 4 1.2, MgCl 2 1.2, CaCl 2 2.5, Glucose 11, ascorbic acid 0.56, and Hepes 15, and the pH value was adjusted to 7.4 using a NaOH solution. Carbon-fiber electrodes insulated with polypropylene and with 11-m diameter tips were used to monitor catecholamine release from individual chromaffin granules in cells under superfusion (22). Electrodes were carefully positioned in close apposition to the cell surface using high precision hydraulic micromanipulation and assessing cell membrane deformation with an Axiovert 135 inverted-stage microscope (Zeiss, Oberkochen, Germany) mounting Hoffman optics (Modulation Optics, Greenvale, NY). Electrical connection was accomplished with mercury. An amperometric potential of ϩ650 mV versus an Ag/AgCl bath reference electrode was applied using an Axopatch 200A amplifier (Axon Instruments Inc., Foster City, CA). Current product of the catecholamine oxidation was digitized with an analogical/digital converter (ITC-16, Instrutech Corp., Great Neck, NY) and recorded at 200 s/point using the program PulseControl (23) running on top of the graphical software Igor Pro (Wavematrics, Lake Oswego, OR) in a PowerMac 7100 computer. Experiments were performed in cells stimulated by superfusion with depolarizing 59 mM high potassium (obtained by replacing isosmotically NaCl by KCl) and applied through a valve-controlled puffer tip commanded by the acquisition software and located near the studied cells.
Individual spike characteristics were studied using Igor-Pro macros called "Spike" versus 1.0 (24) supplied by Dr. Ricardo Borges (Unidad de Farmacología, Universidad de la Laguna, Tenerife, Spain), allowing for peak detection, integration, and kinetic parameter calculations. Oxidized current was filtered at a corner frequency of 400 Hz using an 8-poles low-pass Bessel filter and acquired at 0.2 ms/point. Only well defined narrow peaks with amplitude higher than 5 pA and the width at the half-height (half-width) lower than 75 ms were taken to build event histograms. Electrode to electrode variations were alleviated by using the same electrodes for measurements in control (non-infected) and infected cells (electrode tip was frequently cleaned with isobutanol). Controls were taken before and after (post-control) performing measurements in infected cells, and peak analysis included experiments performed in 23-66 cells in each condition. After obtaining the mean spike parameters for each individual cell, they were average for the number (n) of cells analyzed in each experimental condition (25). The analysis of variance test implemented in the program Graphpad Instat was used for statistical analysis comparing SNAP-25 and synaptobrevin control and post-control parameters with the values obtained for the different mutants. All data was expressed as mean Ϯ S.E. from experiments performed in a number (n) of individual cells. Data represent experiments performed with cells from at least three different cultures.
Confocal Microscopy Studies of the Distribution of GFP-Synaptobrevin Constructs-Briefly, cells overexpressing GFP-synaptobrevin constructs were fixed using a 4% paraformaldehyde in phosphate-buffered saline solution during 20 min. Then cells were washed with phosphatebuffered saline solution and mounted in 80% glycerol in the same solution. Fluorescence was investigated using a Laser Scanning confocal TCS Leika microscope. Usually, eight confocal layers covering the total cell volume were obtained (1.25-m thickness (z)), and individual layers or projections were used to study fluorescence distribution.
Expression of SNAP-25 and Synaptobrevin Constructs and Analysis of SNARE Complex Formation and Stability-GST fusion proteins encoding SNAP-25, syntaxin, and synaptobrevin were produced by cloning their respective cDNAs in the pGEX-KG vector (26). The different SNAP-25 and synaptobrevin mutants were transferred to pGEX-KG either through the NcoI site present in the initial methionine of SNAP-25 or using an EcoRI site preceding the coding region of synaptobrevin, thus ensuring the proper reading frame in the GST fusions. Recombinant proteins were purified from expressing bacteria (Escherichia coli BL21 strain for SNAP-25 and synaptobrevin and C43 strain for syntaxin) after expression induction with isopropyl-␤-D-thiogalactopyranoside during 5 h at 37°C. The constructs were bound to affinity columns of glutathione-Sepharose 4B (Amersham Biosciences) in overnight incubations, and after extensive washing the proteins were eluted by proteolysis with thrombin, followed by inactivation of this protein with 2 mM phenylmethylsulfonyl fluoride.
Formation of ternary complexes by the purified proteins at a 4-M concentration and room temperature was assessed in buffer: 100 mM NaCl, 0.005% octylglucoside, and 5 mM dithiothreitol in 25 mM HEPES, pH 7.5. After incubation periods comprised between 10 s and 60 min, 10-l aliquots were removed, and complex formation was stopped by adding an equal volume of SDS-PAGE buffer (0.1% SDS, 10% (v/v), glycerol, 0.01% ␤-mercaptoethanol, 0.01% bromphenol blue in 0.062 M Tris, pH 6.8). Complex thermostability was assayed by forming complexes during 1-h incubations in the conditions described above and heating samples during 3 min at temperatures ranging from 40 to 90°C after addition of electrophoresis buffer. Complex dissociation was stopped by cooling the samples to 4°C before complexes were analyzed by SDS-PAGE.

GFP-SNAP-25 Constructs Are Heavily Expressed and Functional in Chromaffin Cells Infected with Herpes Simplex Modified Virus (Amplicons)-Overexpression of native and mutant
forms of SNAP-25 in chromaffin cells is a valuable tool when studying its role in neurosecretion (14,15). In initial studies we used calcium phosphate and electroporation transfection methods, resulting in low levels of transfection (less than 1%). We therefore decided to use amplicons, modified HSV (27). Viral infection produced a greater proportion (15-30%) of cells expressing high amounts of GFP-SNAP-25 constructs, as indicated by their brilliant green fluorescence. We used single cell amperometry (11 micron carbon fiber electrodes) to test the ability of these cells to secrete catecholamines in response to a depolarizing stimulus. Upon stimulation, characteristic spikes depicting single fusion events were obtained for a number of individual cells (Fig. 1A). The secretory profile was obtained by the integration and averaging of amperograms from 22 to 60 cells under each experimental condition (Fig. 1B). HSV-infected cells incorporating the GFP-SNAP-25 construct and cells expressing GFP alone (Fig. 1, A and B) produced about 50% of the secretory activity normally found in uninfected cells (not shown). The close match of GFP and GFP-SNAP-25 secretory rates when normalized, indicates that secretory kinetics were unaffected by expression of the GFP-SNAP-25 construct and that the reduction in the number of released vesicles observed in infected cells was due to the viral infection itself rather than to a possible inhibition of the secretory machinery by overexpression of SNAP-25. Validity of the HSV amplicon system for testing SNAP-25 function was further confirmed by the catecholamine release in cells overexpressing a GFP-SNAP-25 construct incorporating the mutation of leucine 203 to glutamic acid (L203E) that had previously been shown to greatly affect secretion kinetics (14). The expression of this construct partially abrogated the overall extent of the secretory response ( Fig. 1A shows a 70% reduction in the sustained response compared with the control GFP-SNAP-25 construct), and initial rate of release (Fig. 1B, about 0.2 vesicles/s, 32% of the rate found for the GFP-SNAP-25 construct). These results are consistent with those obtained by conventional transfection methods (14). Similar results (Fig. 1B) were observed using a construct lacking the last nine C-terminal residues of this SNARE (GFP-SNAP-25 ⌬9) and therefore equivalent to the proteolyzed form produced by the action of botulinum toxin A (28,29). These results indicated that the amino acid L203, which is absent in this construct as well as in the SNAP-25 polypeptide cleaved by botulinum neurotoxin A, is critical for exocytosis.
The amplicon system was further used to examine the secretory response of cells expressing SNAP-25 proteins mutated at amino acids located in critical areas. It has been suggested that amino acid residues in the zero layer (Gln-53 and Gln-174 of SNAP-25) are critical for SNARE complex formation (9,30). The single mutation in Q174K showed a relatively small effect on secretion when compared with the L203E mutant (both magnitude and initial rate of catecholamine release were reduced 35% in comparison to control values, see Fig. 1, A and B). A slight effect was also observed with the Q53K mutation (also present in the zero layer but forming part of the SNAP-25 N-terminal domain), which released 15ϩ2 vesicles (n ϭ 20) during prolonged depolarization (52% inhibition, data not shown) Double mutations in the previously mentioned glutamines (Q53K/Q174K) produced lower inhibition (50% in extent and initial rate of secretion, Fig. 1B) than that obtained with L203E. Although these data imply a role for the zero layer in secretion, it appears to be less important than the part played by Leu-203 at the SNAP-25 C terminus.
Mutational Study of Other Amino Acids Near Leu-203-To assess the importance of regions flanking Leu-203 in molecular events leading to membrane fusion we studied the possible role of residues at the SNAP-25 C-terminal domain in this process. To this end we mutated Ala-199 and Met-202, which have been, respectively, assigned to the seventh and eighth hydrophobic layers of the four-helix bundle in the synaptic fusion complex (30), as well as Lys-201, which is located in between these two. The secretory behavior of cells expressing mutants A199E and M202E was slightly affected with only the initial rate of vesicle release being slower in the M202E mutant (Fig. 2). Mutant K201E had a greater inhibitory effect on the total number of vesicles secreted during cell depolarization, and the initial rate of vesicle release was also retarded (Fig. 2). These results suggest that hydrophobic amino acids in the C-terminal domain of SNAP-25, such as Leu-203 and to a lesser extent Met-202, play a relevant role in membrane fusion events leading to exocytosis and that electrostatic interactions through residue Lys-201 could also be involved in this process.
Mutations in Synaptobrevin Residues Close to the SNAP-25 C-terminal Domain Affect Exocytosis-Current structural data suggest that several hydrophobic synaptobrevin amino acids such as Ala-81, Ala-82, and Leu-84 could interact with Leu-203 of SNAP-25. We therefore decided to mutate these synaptobrevin residues in GFP constructs linked to the synaptobrevin II C-terminal domain, which were also overexpressed following amplicon infection of bovine chromaffin cells (Fig. 3). Constructs with wild-type synaptobrevin (Fig. 3A) as well as the A81D (Fig. 3B), A82D (Fig. 3C), and L84E (Fig. 3D) mutants, were expressed at similar levels in a punctate pattern suggesting their location in chromaffin vesicles. The secretory activity of cells expressing these constructs was studied using depolarizing stimuli. Cells expressing wild-type synaptobrevin coupled to GFP (SV WT) had an exocytotic behavior resembling that observed in cells expressing GFP and the GFP construct coupled to wild-type SNAP-25 ( Fig. 4) in that neither the kinetics nor the level of secretion obtained after 1 min of continuous depolarization was modified. Expression of mutant A81D had a relatively low impact on the release of catecholamines, which showed a 35% inhibition in both overall extent and initial rate of release (Fig. 4). Mutants A82D and L84E had a more pronounced effect on secretory behavior, reaching levels close to 70% inhibition (Fig. 4) similar to that observed in the mutant L203E of SNAP-25. These data strengthen the notion that these synaptobrevin and SNAP-25 hydrophobic residues have a marked and perhaps related influence on secretory behavior.

Single Fusion Analysis Suggests That the SNAP-25 C Terminus and Complementary Region of Synaptobrevin, but Not the Zero Layer, Play a Role in the Very Last Events of Exocytosis-
The study of individual fusion kinetics could be potentially useful in elucidating the mechanisms behind the differential effects of SNAP-25 and synaptobrevin mutants. This is possible by analyzing the shape of single amperometric events (31) occurring in the proximity of the carbon fiber electrode. A semi-automatic analysis was performed using software (24) that measured spike amplitude, half-width, and event charge and selecting well separated spikes with amplitudes over 5 pA and half-width values lower than 75 ms. The collected data were binned to generate distribution histograms for means obtained from individual cell analysis in which parameters were obtained for each individual cell and later averaged, sim-ilar to the ones shown for the entire event population in Fig. 5 and summarized in Table I. Non infected cells and cells expressing GFP-SNAP-25 or GFP-synaptobrevin were characterized by very similar distributions, showing means taken from averaged individual cell amplitudes of 70, 59, and 50 pA, respectively (Table I). Amplitudes were also similar in cells expressing single (Q174K, Fig. 5D and Table I; Q53K, Table I) and double (Q174K/Q53K, Table I) mutants of the zero layer as well as A199E and M202E mutants of the SNAP-25 C terminus (Table I). However, the mean amplitudes obtained with L203E ( Fig. 5G and Table I) and K201E mutants (Table I)  tudes over 75 pA, which were relatively abundant in control cells were less frequent in these cases. These results are in good agreement with those reported for C-terminal deletions in transfected chromaffin cells (14). On the other hand, the alteration seen in mean amplitude was not associated with a change in the charge released per event, which remained relatively unaltered in control, zero layer, and C terminus mutants with average values ranging from 1.3 to 0.94 pC 1/3 (Fig. 5, B, E, and H and Table I). By contrast, statistically significant differences in the half-height width distribution (half-width) were observed for L203E, K201E and ⌬9 mutants (p Ͻ 0.002, analysis of variance test) when compared with values obtained for wildtype SNAP-25 (see Fig. 5, C, F, and I and Table I). The halfwidth distributions of wild-type SNAP-25 as well as mutants Q174K, Q53K, and A199E had a high proportion of short t 1/2 value events (ranging from 6 to 14 ms and mean values around 18 -22 ms), whereas cells expressing L203E, K201E, and ⌬9 mutants had longer and smaller events (24 -26 ms mean values), the presence of narrow spikes corresponding to vesicles fusing with short half-width values being less frequent. The amperometric events from synaptobrevin constructs were also analyzed. In this case cells expressing mutant A82D were characterized by lower average amplitude (Fig. 5J) and higher halfwidth (Fig. 5L) values (p Ͻ 0.001, Table I). The differences with respect to the controls were even greater than those observed with the SNAP-25 mutants. The effect was less pronounced with the L84E mutant (p Ͻ 0.05, Table I), despite a marked decrease in overall secretion (see Fig. 4). The A81D mutant showed no difference from the wild-type construct (Table I).
In summary, only certain amino acids of SNAP-25 and synaptobrevin, presumably in close proximity each other, appear to play a central role in the membrane fusion events immediately preceding the release of stored catecholamines.

Thermostability of SNARE Complexes Formed by Different Mutants in Vitro Does Not Fully Correlate with Their Effects on
the Exocytotic Process-The previously described forms of SNAP-25 and synaptobrevin together with syntaxin were used to study the in vitro stability of their ternary complexes in an attempt to understand the molecular basis for the alterations observed in exocytosis. Initially, complexes were formed using syntaxin, synaptobrevin lacking the membrane anchor domain, and the different SNAP-25 mutants. All mutants (apart from Q53K/Q174) retained the ability to form the complex at room temperature. Kinetic studies indicated that the mutant forms retarded the complex formation when compared with the wildtype (data not shown). Fully formed complexes were obtained over periods ranging from 4 to 10 min at room temperature. Different batches of purified protein produced slight variations in the total amount of complex formed.
Differences in the thermostability of these complexes were also evident (Fig. 6). Control complexes started to dissociate at 75-80°C. However, complexes formed with the altered forms of SNAP-25 dissociated at lower temperatures, (60 -70°C). When fitted to simple sigmoid curves the data gave half-values of 78°C for control complexes and 65, 68, and 68°C for the SNAP-25 ⌬9, L203E, and Q174K constructs, respectively. This would indicate that mutations producing different effects on the secretory process of chromaffin cells, such as L203E and Q174K, have a similar effect on complex formation in vitro.
Similarly the Q53K mutation in the N terminus of SNAP-25, had a half-value of 63°C (not shown). Therefore, a direct correlation between the ability to inhibit exocytotic function and modify the in vitro formation of SNARE complexes does not seem to exist. Similarly, the temperature-dependent dissociations of the complexes formed by SV A82D and SV L84E synaptobrevin mutants were characterized by lower half-values (around 57-58°C) when compared with complexes formed by wild-type whole synaptobrevin (Fig. 7). Again, despite the dif-  ferences found in their individual fusion kinetics parameters, the thermostability of the ternary complexes formed by these synaptobrevin mutants is very similar.

DISCUSSION
The use of botulinum neurotoxin has been crucial to understand the key role played by the SNAP-25 C terminus in the secretory process of neuroendocrine chromaffin and PC12 cells (22,(32)(33)(34)(35). More recently, the use of constructs containing modifications in this domain (14) has facilitated access to protein segments different from those defined by toxin cleavage. Thus, the substitution of endogenous SNAP-25 by an overexpressed mutant in which the last six C-terminal amino acids are deleted, alters the single vesicle fusion kinetics analyzed by amperometry (36), suggesting the involvement of these residues in the very last events of granule exocytosis. The data presented here expands on our previous findings and provides new insights into the exocytotic role of single amino acids from critical domains of SNAP-25 and synaptobrevin. Our intention was to disturb the function of these domains by introducing drastic mutations. Despite these radical changes, and with the exception of the Q174K/Q53K SNAP-25 double mutant, all others mutants were able to form complexes in vitro, and therefore it is reasonable to think that they also associate in vivo as is suggested by their effects on exocytosis. These effects were not unspecific, and only some mutants affected exocytosis in a drastic way. Another concern, always present when using dominant negative mutants, is related to the potential function of the native wild-type protein remaining in the cells. Since the mutated forms of SNAP-25 are overexpressed, it is reasonable to assume that they take over the role of the native protein.
However, the possibility that native proteins may coexist with mutated forms during exocytosis cannot be discarded. If this were the case, we could predict one of two possible outcomes. Firstly, the native protein could perform its function undisturbed, independently of the type of mutant introduced. However, this does not seem to be the case, since a deleterious effect is clearly observed only with some mutants. Secondly, the expression of specific mutants could somehow influence the native protein behavior, thereby changing the fusion characteristics and amperometric parameters. This might imply complex relationships between the native and the mutant proteins that would be difficult to analyze but cannot be discarded. In general, our strategy of testing several closely located mutants in the same region and the subsequent observation of their differential effects on exocytosis support the following conclusions.
From the functional amperometric studies, we can see that L203E and K201E mutants as well as the ⌬9 deletion of the SNAP-25 C terminus induce a strong inhibition of secretion ( Figs. 1 and 2), which is correlated to the altered kinetics of catecholamine release from individual vesicles (Table I). These mutants, when compared with controls (either expressing GFP-SNAP-25 or not) were characterized by smaller and wider average amperometric spikes, indicating slower single vesicle kinetics. By contrast, the effect of single (Q174K and Q53K) and double (Q53K/Q174K) mutants of the zero layer as well as the other C-terminal mutant (A199E) on overall secretion was less pronounced (Figs. 1 and 2), and the kinetic characteristics of the spikes remained unaltered (Table I). Thus, these experiments strongly suggest that specific SNAP-25 C-terminal residues, such as L203 and K201, that had not been mapped onto the highly conserved layers of interacting amino acids in the four-helix bundle play a fundamental role in the very last membrane fusion steps of exocytosis and that the zero layer may participate in up-stream stages of this process. These conclusions are supported by a recent piece of work (15) in which the effects of the ⌬9 deletion and Q174L mutant were studied by combining flash photolysis and capacitance techniques. It was deduced that the zero layer is critical for the formation of SNARE complexes but, in contrast to the last nine C-terminal amino acids, plays no role in the dynamic equilibrium between the two exocytotic burst components. The significance of Leu-203 and Lys-201 at the SNAP-25 C terminus was further stressed when mutations in synaptobrevin residues such as Ala-82 and Leu-84, which have been mapped in their immediate proximity, (9, 10, and the corresponding entries of the Protein Data Bank 1SFC, 2BU0), produced a very similar impact on both overall secretory properties (Fig. 4) and the individual vesicle fusion characteristics (Table I).
As expected from their marked effect on secretory cell behavior, complex thermostability decreased in the presence of C terminus-altered forms of SNAP-25 such as ⌬9 and L203E as well as in their functional synaptobrevin counterparts A82D and L84E. However, Q174K and Q53K mutants of the SNAP-25 zero layer, which did not affect secretion in a drastic way, had a strong effect on the in vitro formation and stability of the core complex. As such there is not a complete correlation between the functional effects of the mutants in vivo and their ability to modify the properties of the ternary complex in vitro. In this sense, the most extreme case was the double mutation in the zero layer arginines contributed by SNAP-25, which had a severe impact on the complex since we could not detect its formation at room temperature but which was able to interact in vivo since it showed an intermediate behavior between that of the controls and the most inhibitory C-terminal mutants. These data point to the fact that observations on the formation and stability of the synaptic fusion complex in vitro should be interpreted with caution and validated with functional data.
What could be the role of SNAP-25, Leu-203, and Lys-201 FIG. 7. Thermal stability of the complexes formed with synaptobrevin constructs. Experiments were performed as described in the previous figure but using the different synaptobrevin constructs. A, examples of ternary complex thermostability corresponding to fulllength synaptobrevin (SV WT) and to the mutants SV A82D and L84E. B, thermostability curves obtained in the conditions described in A for data corresponding to three experiments. Results were expressed as percentage of the maximum amount of formed complex and given as means Ϯ S.E. and synaptobrevin Ala-82 and Leu-84 amino acids during the final membrane fusion steps of exocytosis? As previously indicated, these amino acids had not been assigned to any of the conserved layers of amino acids whose side chains contribute to synaptic fusion complex formation. Nevertheless, the hydrophobic nature of SNAP-25 Leu-203 and synaptobrevin Ala-82 and Leu-84 suggests their involvement in leucine zipper mechanisms stabilizing protein-protein interactions. In addition, the importance of these hydrophobic residues may reside in their location within the SNAP-25 C terminus and the complementary region of synaptobrevin where they could: (a) reinforce the structural integrity of the zippering movement toward the Cterminal regions of the participating SNAREs (37), (b) help form a tight state of the complex after zippering and participate in its final closure, and (c) drive the membrane fusion by apposition of the membranes bellow a critical distance. Regarding Lys-201, and according to the actual crystalographical models, it may be located so that its side chain is oriented toward the outside of the four-helix bundle. Its positively charged nature suggest an involvement in electrostatic interactions with other protein factors that may be present as components of the basic machinery responsible for vesicle docking, maturation, and at least part of the final membrane fusion mechanism. An alternative hypothesis would be that Leu-203 and Lys-201 SNAP-25 mutations and Ala-82 and Leu-84 synaptobrevin mutations disrupt the overall structure of the corresponding core complex domain. However, this disruption should be specific for these amino acids and not for others located in the same area, such as SNAP-25 Ala-199 and synaptobrevin Ala-81. In either case, the experiments described in this paper demonstrate an important participation by specific amino acids of the SNAP-25 C-terminal domain and the corresponding synaptobrevin domain in the final membrane fusion steps of the exocytotic process.