|Home | About | Journals | Submit | Contact Us | Français|
Complexins and synaptotagmins regulate calcium-dependent exocytosis. A central helix in complexin confers specific binding to the SNARE fusion machinery. An accessory helix in the amino-terminal region inhibits membrane fusion by blocking SNAREpin zippering. We now show that an amphipathic helix in the carboxy-terminal region of complexin I binds lipid bilayers and affects SNARE-mediated lipid mixing in a liposome fusion assay. The substitution of a hydrophobic amino acid within the helix by a charged residue abolishes the lipid interaction and the stimulatory effect of complexin I in liposome fusion. In contrast, the introduction of the bulky hydrophobic amino acid tryptophan stimulates lipid binding and liposome fusion. This data shows that local complexin-lipid interactions can play a role in membrane fusion.
Intracellular membrane fusion is mediated and controlled by an obligatory set of conserved protein families . The core membrane fusion machinery consists of cognate v(vesicle)-SNAREs and t(target)-SNAREs, which can spontaneously fuse liposomes or entire cells, when located on opposite membranes [2,3]. The formation of trans-SNARE complexes (SNAREpins) between two membranes starts at the amino-terminal, membrane-distal ends of the so-called SNARE motifs, which are conserved coiled-coil regions of 50–60 amino acids. A functional SNARE complex always contains four SNARE motifs, classified as Qabc- and R-SNAREs [4–7]. SNARE motifs assemble in a zipper-like fashion leading to a progressive approach of the membranes . Finally, fully assembled SNARE motifs form an extremely stable four-helix bundle and this protein folding reaction suffices to spontaneously merge the two lipid bilayers [3,9].
Synaptic transmission and other forms of regulated exocytosis require additional components that ensure signal-dependent transmitter release. Two proteins - complexin and the calcium sensor synaptotagmin - are intimately coupled to the triggering process . Complexins, small cytosolic proteins, regulate distinct steps of the assembly and fusion reaction by contributing both stimulatory and inhibitory activities [11–23]. Complexins bind with low affinity to t-SNAREs, consisting of syntaxin 1 (stx1) and SNAP-25 on the plasma membrane . Using t-SNAREs reconstituted into liposomes, it has been shown that the complexin–t-SNARE interaction can interfere with the docking of v-SNARE vesicles (containing VAMP2/synaptobrevin2) and SNAREpin formation . In contrast, when complexin is incubated with liposomes, which are already bridged by SNAREpins, complexin stabilizes SNAREpins and stimulates membrane fusion [20,21]. Structural and functional studies revealed a potential molecular mechanism for the dual roles of complexins. Complexins contain an unstructured amino-terminal domain (amino acids 1–25), followed by a so-called accessory inhibitory helix (aa 26–47), a central helix (aa 48–70) and an unstructured carboxy-terminal domain (aa 71–134). The central helix of complexin binds with high affinity in an anti-parallel manner to the assembled four-helix SNARE bundle . The binding site is provided by a groove formed by VAMP2 and syntaxin 1 and covers the central portion of the SNARE bundle. In such fully assembled SNARE complexes, the accessory helix of complexin, which is located close to the membrane proximal region of the SNAREs, does not directly interact with the SNARE four-helix bundle. In contrast, in a partially assembled SNAREpin the accessory helix apparently binds the membrane proximal part of the t-SNARE and thereby competes with VAMP2 for t-SNARE binding [14,19]. In this scenario complexin inhibits SNAREpin zippering. This model can explain the incubation-step dependent inhibitory and stimulatory functions of complexin. Concerning the first 25 amino acids of CpxI, in vivo studies revealed that this highly conserved region plays a stimulatory role, but the molecular mechanism still remains to be elucidated [19,22]. Recent in vitro liposome fusion experiments showed that the carboxy-terminal domain of CpxI stimulates SNARE-mediated liposome fusion . Mutations in S115, a putative phosphorylation site, abolish this stimulation. We now demonstrate that the amino acids flanking S115 are part of a conserved sequence, containing an amphipathic helix, which confers lipid binding. Mutations in the hydrophobic face of this helix alter the lipid binding and liposome fusion-stimulating properties of complexin I.
All lipids were purchased from Avanti Polar Lipids. The chloroform stock solutions of all lipids were stored under N2 at −20 °C. Lipid stock solutions were mixed in round bottom glass flasks to obtain a lipid mixture which is composed of 65 mol% POPC (1-palmitoyl-2-oleoyl-SN-glycero-3-phosphatidylcholine) and 35 mol% DOPS (1,2-dioleoyl-SN-glycero-3-phosphatidylserine). Evaporating chloroform with a stream of N2 gas formed a homogenous lipid film. To remove residual solvent, the sample was further dried under vacuum for 1 hour. The lipid film was rehydrated in reconstitution buffer (25 mM HEPES (2-[4-(2-Hydroxyethyl)-1-piperazinyl]-ethanesulfonic acid) pH 7.4, 100 mM KCl) at 37 °C for 1 hour to obtain a lipid concentration of 10 mM. Following 10 cycles of freezing (dry ice in isopropanol) and thawing (37 °C water bath), the liposomes were passed 8 times through 50 nm filters (Nuclepore track-etched polycarbonate membrane, Whatman) using a LIPEX Extruder (Northern lipids).
Liposomes were prepared as described above. 14 µM CpxI was incubated with liposomes, containing 4 mM lipids in flotation buffer (25 mM HEPES pH 7.4, 500 mM KCl, 1 mM EGTA (Ethylene glycol-bis(2-aminoethylether)-N,N,N’,N’-tetraacedic acid), 1 mM DTT (1,4-Dithiothreitol)) for 1 hour at room temperature. After the incubation, 125 µl of the sample were mixed with 125 µl 80% Nycodenz (5-(N-2,3-dihydroxypropylacetamido)-2,4,6-tri-iodo-N,N’-bis(2,3-dihydroxypropyl)isophthalamide, Axis-Shield) in flotation buffer to yield a final concentration of 40% Nycodenz. The sample was overlayed with 400 µl 35% Nycodenz, 400 µl 30% Nycodenz and 200 µl buffer. After ultracentrifugation in a TLS-55 rotor (Beckman Coulter) at 55.000 rpm for 3 hours, the gradient was fractionated (100 µl fractions) from the top.
Fractions were subjected to 15% (wt/vol) SDS-PAGE, followed by Western blotting. Complexin I constructs were detected with the polyclonal rabbit antibody SM193 and were visualized with an Alexa688 coupled secondary antibody (Invitrogen) using an Odyssey infrared imaging system (LI-COR Bioscience).
Liposomes were prepared as described above. 4 µM complexin and liposomes, containing 1 mM lipids, were mixed and incubated for 30 min at 25 °C in quenching buffer (25 mM HEPES pH 7.4, 100 mM KCl, 250 mM acrylamide, 1 mM DTT). Samples were excited at 280 nm and the emission spectra were taken with a FP-6500 fluorescence spectrometer (JASCO) between 300 and 500 nm. After the first measurement, 1 mM CaCl2 was added directly to the samples and a second data set was subsequently taken. In order to normalize the data, the spectrum of the buffer was subtracted from the complexin containing spectra and the spectrum of the liposomes alone was subtracted from the spectra of the complexin/liposome mixture. Spectra displayed show one representative experiment. Experiments were repeated several times independently, yielding virtually identical results.
For donor liposomes 62 mol% POPC, 35 mol% DOPS, 2 mol% Rhodamine-DPPE (N-(lissamine rhodamine B sulfonyl) 1,2-dipalmitoyl phosphatidylethanolamine) and 1 mol% NBD-DPPE (N-(7-nitro-2,1,3-benzoxadiaziole-4-yl)-1,2-dipalmitoyl phosphatidylethanolamine) were mixed resulting in 3 mM total lipid. The 15 mM total lipid acceptor lipid mix contains 65 mol% POPC, 35 mol% DOPS. Final lipid concentrations after flotation were determined by using parallel reconstitutions, which contained 1 mol% Rhodamine-DPPE in the acceptor and donor liposomes.
Liposomes were formed in the presence of VAMP2 (0.3 mg/ml) or t-SNARE complex (1 mg/ml) in 1% Octyl β-D-glucopyranoside using the donor and acceptor lipid mixes defined above, and the previously described technique of dilution and dialysis followed by a Nycodenz gradient centrifugation [3,20]. Protein lipid ratios were 1/150, and 1/210 for VAMP2 and stx1/SNAP-25 liposomes, respectively. Protein expression and purification are described in detail in the supplementary material.
Fusion reactions and data analysis were performed as described previously  with the following modifications: (a) In all cases, 30 µl of acceptor (unlabelled) and 5 µl of donor (labeled) liposomes were used, (b) acceptor and donor liposomes were preincubated on ice for 15 min in the presence or absence of different complexin constructs. The cold microtiter plate was transferred into a Synergy4 plate reader (Biotek) and samples were measured at intervals of 30 seconds for 30 minutes at 37 °C. (c) The fusion-dependent fluorescence is normalized to the maximal fluorescence signal obtained in the presence of 0.4% dodecyl-β-D-maltoside. Fusion kinetics displayed show one representative experiment. For a quantitative comparison of the various complexin mutants, subtracting the signals of the VAMP CD-containing reaction from the other fusion reactions. The NBD fluorescence after 10 min was used as an indicator for the fusion kinetics. Fluorescence intensity of the control reaction was set as 100% and the other reactions were calculated accordingly. The error was calculated as standard error of the mean from three independent experiments.
To identify the molecular mechanism that underlies the stimulatory role of the CpxI carboxy-terminus in liposome fusion, we analyzed the region flanking amino acid serine 115. The sequence alignment of complexin homologues shows a cluster of negatively charged amino acids that precedes S115 (Fig. 1A), followed by a short amphipathic helix, which covers approximately 3 helix turns and is terminated by a proline that is conserved in CpxI-III (Fig. 1B). Drosophila Cpx apparently lacks the conserved serine and the amphipathic helix in this region.
Since amphipathic helices are known to interact with membranes, locally perturb the lipid bilayer, change the membrane curvature, and can participate in membrane fusion processes, it was tested if complexin I has the capability to bind membranes. Since the amphipathic helix is short, a potential complexin-lipid interaction might be weak. To test if the amphipathic helix indeed binds lipid bilayers, wild type and mutant CpxI constructs were incubated with liposomes containing 65 mol% PC and 35 mol% PS. Liposomes were reisolated by flotation on a Nycodenz gradient, containing 500 mM KCl. High ionic strength was employed to focus primarily on the hydrophobic interactions and to suppress ionic binding, which could occur between the highly charged complexins and the lipid bilayers. The recovered CpxI in the top fractions provides a measurement for membrane binding (Fig. 2A). Wild type CpxI binds to liposomes, and a point mutation at position 117, with leucine replaced by the bulky hydrophobic amino acid tryptophan (L117W), showed an enhanced membrane interaction (Fig. 2C). In contrast, mutations that convert this leucine into a charged amino acid (L117K), or the introduction of the helix breaker proline (V120P), dramatically reduce lipid binding. These results demonstrate that the amphipathic helix confers lipid binding.
As an independent method to detect the lipid interaction, we made use of the L117W mutant. Tryptophans are characterized by an absorption maximum at 280 nm, but their emission peak varies from 330–350 nm dependent on the local environment. In a hydrophilic buffer, tryptophan emits at 350 nm, in a hydrophobic environment the peak is usually blue-shifted and the fluorescence intensity is increased. Since at steady state only a small fraction of the complexin population interacts with liposomes, acrylamide was added to quench the tryptophan fluorescence in the aqueous environment . In this assay, we employed physiological salt concentrations, which also allow ionic interactions. In addition it was tested if calcium affects the CpxI-lipid binding, because the cluster of negatively charged amino acids preceding the amphipathic helix might interfere with the binding to the negatively charged PS in the liposomes. When incubated with liposomes in the absence of calcium a small blue-shift and an increase in the emission peak of the W117 was detected (Fig. 2D). In the presence of calcium, the fluorescence intensity was further increased. When PS was omitted from the liposomes, or when 500 mM KCl was added during the incubation, the increase and blue-shift of the W117 fluorescence was prominent, but largely calcium-independent (data not shown). A control tryptophan introduced at position 101 (I101W), which is not part of the amphipathic helix, did not show any significant emission increase or blue-shift upon liposome addition (Fig. 2E). Thus, the tryptophan residue at position 117 dips into the hydrophobic core of the lipid bilayer, in a manner that depends on the local charge environment.
Having identified a lipid bilayer-binding site in CpxI, the question arises if this interaction is of functional relevance. In the simplest case, the local interaction of the amphipathic helix with the lipid bilayer would change the membrane structure and affect membrane fusion. To test this hypothesis, we made use of a well-established reconstituted lipid-mixing assay [3,26]. v-SNARE liposomes contain VAMP2 and a quenched pair of NBD- and rhodamine-labeled lipids. The acceptor t-SNARE liposomes contain syntaxin 1 and SNAP-25 but lack the fluorophores. Fusion of these liposomes results in fluorescence dequenching and an increase of NBD-fluorescence. We employed conditions of a moderate stimulation of membrane fusion upon the addition of wild type CpxI (Fig. 3A). When the CpxI L117K mutant was added to the fusion reaction, the fusion kinetics was comparable to the control reaction lacking CpxI (Fig. 3A). In contrast, the addition of CpxI L117W significantly stimulated the fusion kinetics (Fig. 3A). Lipid mixing was SNARE-dependent, because the addition of the cytoplasmic domain of VAMP2 abolished the fusion signal. To further analyze if this stimulatory effect depends on the CpxI-SNARE interaction, we introduced point mutations in the central helix of CpxI (R48L/R59H) that abolish SNARE binding. This triple CpxI R48L/R59H/L117W mutant completely lost its stimulatory potential (Fig. 3B). As expected, the CpxI I101W had only moderate effects, in contrast to the CpxI V120P mutant, which completely abolished the stimulatory effect of CpxI (Fig. 3B). Figure 3C shows a quantitative analysis of the different complexin mutants 10 minutes after the start of the reactions. Fig. 3D demonstrates that equal amounts of the distinct CpxI constructs were added to the fusion reactions. Thus, the carboxy-terminal amphipathic helix of CpxI modulates lipid mixing in a SNARE- and lipid-interaction-dependent manner.
This study shows that at least some complexins contain an amphipathic helix at their carboxy-terminus. Our analyses using CpxI demonstrate that this amphipathic helix binds lipid bilayers. A cluster of negatively charged amino acids preceding the amphipathic helix can apparently modulate the lipid interaction as supported by our data. Thus, both ionic and hydrophobic interactions affect the lipid interaction. Indeed, dependent on the local lipid composition, calcium enhances the CpxI–membrane interaction. This result is consistent with a recent publication showing that CpxI stimulates liposome fusion in a calcium- and lipid-dependent manner . Our data directly demonstrates a role of the hydrophobic face of the amphipathic helix in lipid binding and SNARE-dependent liposome fusion, measured by a lipid mixing assay.
The overall interaction of CpxI with the lipid bilayer is relatively weak, but in the presence of SNAREpins, CpxI will be concentrated at fusion sites. Thus, the Cpx-lipid interaction will occur in a highly focused manner. Indeed, the simple presence of CpxI in the fusion assay does not affect membrane fusion, when the CpxI-SNARE interaction is abolished. Remarkably, the introduction of a tryptophan at position 117 significantly enhances the stimulatory effect of CpxI in membrane fusion. Since tryptophan is a bulky amino acid, it is likely that the insertion of tryptophan into the membrane induces positive membrane curvature. Recent publications showed that similar amino acid substitutions in the hydrophobic loops of the C2 domains of synaptotagmin I also enhance membrane fusion in vitro and in vivo [27–30]. Furthermore, mutations in Ser 115 also affect liposome fusion indicating that the hydrophilic face of the amphipathic helix might participate in complexin-protein interactions, linking the lipid interaction to the fusion apparatus [20,31].
Since membrane fusion is a multiple step process and the local lipid composition at fusion sites is unknown, it is difficult to predict the exact function of the Cpx-lipid interaction. For example, several SNAREpins accumulate at the fusion site, and coherently multiple Cpx-lipid interactions could create a unique lipid environment that stalls or accelerates distinct fusion intermediates (e.g. affect the local induction of positive membrane curvature, or hemifusion intermediates). In our assay, which uses high curvature liposomes, the CpxI-lipid interaction clearly stimulates lipid mixing. In the presence of appropriate lipids, the calcium-dependent CpxI-lipid binding together with calcium-dependent synaptotagmin/SNARE/lipid interactions could become an integral part of a triggering mechanism opening the fusion pore or at least, function in concert with synaptotagmins to deform the membrane. In addition, or alternatively, the CpxI-lipid interaction might have a modulatory role in the core fusion reaction.
This work was supported by the US National Institutes of Health (NS043391 to T.H.S.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.