Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Jul 3;114(27):E5343-E5351.
doi: 10.1073/pnas.1702542114. Epub 2017 Jun 20.

Structural characterization of the Rabphilin-3A-SNAP25 interaction

Affiliations

Structural characterization of the Rabphilin-3A-SNAP25 interaction

Cristina Ferrer-Orta et al. Proc Natl Acad Sci U S A. .

Abstract

Membrane fusion is essential in a myriad of eukaryotic cell biological processes, including the synaptic transmission. Rabphilin-3A is a membrane trafficking protein involved in the calcium-dependent regulation of secretory vesicle exocytosis in neurons and neuroendocrine cells, but the underlying mechanism remains poorly understood. Here, we report the crystal structures and biochemical analyses of Rabphilin-3A C2B-SNAP25 and C2B-phosphatidylinositol 4,5-bisphosphate (PIP2) complexes, revealing how Rabphilin-3A C2 domains operate in cooperation with PIP2/Ca2+ and SNAP25 to bind the plasma membrane, adopting a conformation compatible to interact with the complete SNARE complex. Comparisons with the synaptotagmin1-SNARE show that both proteins contact the same SNAP25 surface, but Rabphilin-3A uses a unique structural element. Data obtained here suggest a model to explain the Ca2+-dependent fusion process by membrane bending with a myriad of variations depending on the properties of the C2 domain-bearing protein, shedding light to understand the fine-tuning control of the different vesicle fusion events.

Keywords: C2 domains; Rabphilin-3A; SNAP-25; X-ray crystallography; membrane fusion.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.
Fig. 1.
Crystal structure of the Rph3A C2B–SNAP25 complex. (A) The asymmetric unit of the C2 crystals contains two SNAP25 molecules (-N1, -C1, and -N2, -C2 helices shown in light and dark green ribbons, respectively) organized in an antiparallel arrangement and bound to three Rph3A C2B molecules. Two of them (C2B-A, red, and C2B-B, yellow) contact through their bottom helices with the SNAP25 -N1 and -N2 regions, from K40 to R59. The third C2B-C molecule (orange) binds the N-terminal–most region of only the SNAP25-N1 molecule. (B) Open-book view of the SNAP25-C2B contact interfaces, for the C2B-A and C2B-B molecules, colored according to its electrostatic potential. As both interacting surfaces are quasiequivalent, only C2B-A was shown in panel. The two complementary electrostatic patches (black boxes) at each side of the central hydrophobic cluster are apparent in all interfaces. The black boxes indicate the SNAP-25 and Rabphilin C2B regions that are involved in complementary electrostatic patches. (C) Close-up view of the interactions between the C2B bottom helices and SNAP25-N region K40–R59, depicted only for molecules A. (D) Overall view of the Syt1 C2AC2B–SNARE complex (PDB ID 5CCG; Left Inset) and close-up view of the C2B–SNARE primary contact interface, region I that involves the SNAP-25-N moiety. Structural comparisons show that, although the Rph3A and Syt1 C2B regions involved in interactions with SNAP25 are not related, the SNAP25 surface participating in contacts is essentially coincident in the two complexes.
Fig. 2.
Fig. 2.
The Rph3A C2B-C molecule also binds SNAP25 using their bottom α-helices. (A) Overall view of the main C2B-C–SNAP25 contact interface, showing the C2 domain in orange and the SNAP25 N1 colored in light green. (B) Close-up view of the interactions between the C2B-C bottom helices contacting the N-terminal–most region of N1 helix. (C) Open-book view of the SNAP25–C2B-C contact interface, colored according to its electrostatic potential.
Fig. S1.
Fig. S1.
Electron density maps. (A) σA-weighted 2Fo−Fc electron density map, contoured at 1.0 σ, around the SNAP25–C2B complex that forms the asymmetric unit of the C2 crystal form. For clarity, the model is placed inside in ribbon representation; the four SNAP25 helices are shown in light (-N1, -N2) and dark (C1, - C2) green, respectively, with the bound C2B domains depicted in red C2B-A, yellow C2B-B, and orange C2B-C. (B) Stereoview of a σA-weighted 2Fo−Fc map (1.3 σ) around one of the contact interfaces involving the C2B bottom helices (atom-type color sticks with carbon atoms in yellow) in the P21 crystal form. The SNAP25-interacting residues are also as atom sticks with carbon atoms in green.
Fig. S2.
Fig. S2.
Organization of the four-helix bundle in the SNAP25–C2B complexes. (A) The -N and -C α-helices of each SNAP25 molecule assemble in a parallel way, whereas the two independent SNAP25 molecules were arranged in an antiparallel orientation. Besides the hydrophobic contacts, a number of polar and ionic interactions contribute to the stabilization of bundle. This arrangement contrasts to that of the SNARE complex (B) where the four components of the heterotrimer are organized in parallel, with the N termini at one end of the bundle (4). The two close-ups show the so-called “0” layer composed of R56 from VAMP2, Q226 from STX1A, Q53 from SNAP25-N, and Q174 from SNAP25-C in the SNARE complex (Bottom Inset in B). In this complex, the positively charged guanidinium groups of arginine interact with the carboxyl groups from each of the three glutamine residues. The Bottom Inset in A shows the equivalent region in the SNAP25 four-helix bundle. In absence of the VAMP2 R56 residue, the two SNAP25 glutamine side chains appear rearranged, interacting to each other. (C) It is important to note that, in absence of the full-length SNAP25 structure, the arrangement of the SNAP25-N and -C helices seen in our crystal structures would be compatible with two types of connections in which the two -N and -C helices can arrange either in the parallel mode described in A (Left) or in an antiparallel orientation (Right).
Fig. S3.
Fig. S3.
Asymmetric unit of the SNAP25–C2B complex in the P21 unit cell. The color code is the same as in Fig. 1A.
Fig. S4.
Fig. S4.
The Rph3A C2B–SNAP25 interactions formed by crystal packing, in the P21 (A) and C2 (B) space groups. The content of the crystal asymmetric unit is colored as in Fig. 1A and Fig. S3, and the neighboring molecules in the crystal are in gray. The Left Upper panels, in A and B, show a close-up of the largest contacting surfaces (contact area, >300 A2). The details of the interacting residues are shown in the Lower. Besides the interactions mediated by the C2B domain α-helices, the C2B-A molecule interacts with two additional SNAP25 molecules using two distinct surfaces: the first involves ionic contacts mediated by residues within a long loop located at bottom face of the domain (gray square), and the second involves residues of CBR2 (red square). Molecule C2B-C was also involved in additional crystal contacts, using the polybasic region of strands β3 and β4 and part of the CBR2 loop. The C2B-B molecule does not participate in any addition interface of contact with SNAP25 but is involved C2B–C2B interaction in the crystal packing.
Fig. 3.
Fig. 3.
Comparison of the SNAP25 binding interfaces of Rph3A C2B and Syt1 C2B. (A) Structure-based sequence alignment of the Rph3A and Syt1 C2B domains. The strictly conserved residues are in red blocks, and similar residues are in red characters and blue boxes. The Rph3A C2B residues interacting with the SNAP25 are marked with olive blocks, and that of syt1 C2B with salmon blocks. (B) BLAST alignment (18), using the sequences around the C2B α1- and α2-helices (red rectangles) that form the main interface of the C2B–SNAP25 complex. Blue rectangles show residues interacting with SNAP25, with red stars indicating amino acids that contact the Nt-region II (D51, E52, E55, and R59), and yellow stars highlighting residues that contact Nt-region I (E38, K40, and D41).
Fig. 4.
Fig. 4.
Structures of Rph3A C2B domain bound to PIP2. Panels show the two C2B–Ca2+–PIP2 complexes contained in the crystal asymmetric unit. (A) The C2B-A molecule is shown in orange with the side chains of amino acids directly in contact with PIP2 depicted as sticks in atom-type color. (B) The C2B-B molecule is shown in deep olive with the side chains of amino acids, interacting with the phosphoinositide also shown in atom-type sticks. In both panels, the bound PIP2 molecules are depicted as sticks in atom-type color, with the weighted 2|Fo|−|Fc| electron density maps (contoured at 1.0 σ) shown as a mesh in light blue. The calcium ions are shown as light blue spheres. In A, the SO42− ion bound to Ca1 is shown in atom-type sticks.
Fig. S5.
Fig. S5.
Structure and interactions of the PIP2 binding pocket in different C2 domains. (A) Superimposition of the two of Rph3A C2B molecules contained in the asymmetric unit of the Rph3A C2B C2B–Ca2+–PIP2 complex. The C2B-A molecule is shown in orange with the bound PIP2 and PIP2-interacting side chains depicted as sticks in atom-type color (carbons in orange). The C2B-B domain is represented in deep olive and semitransparent tracing. (B) The Rph3A C2A–PIP2 complex (PDB ID 4NS0; ref. 12). The C2A domain in shown in green with the bound PIP2 and PIP2-interacting side chains represented in atom-type sticks (carbon atoms in white and green, respectively). (C) The PKCa C2 domain in complex with I3P (PDB ID 3GPE; ref. 24). The C2 domain in shown in slate blue with the bound PIP2 and PIP2-interacting side chains represented in atom-type sticks (carbon atoms in white and slate blue, respectively).
Fig. 5.
Fig. 5.
Characterization of the Rph3A/SNAP25 main interface. (A) Representative lipid sedimentation assay to measure the Ca2+/PIP2-dependent membrane binding of the Rph3A–SNAP25 complex. Increasing concentrations of Rph3A C2AB domain (2–20 μM) were incubated with 20 μM SNAP25 before adding the lipid mixture. The free (S) and bound (P) Rph3A C2AB domain and SNAP25 were analyzed by SDS/PAGE (Left). The Right panel shows the relative binding of Rph3A C2AB (dark gray bars) and SNAP25 (light gray bars) to liposomes as a function of the total Rph3A C2AB. Values are normalized for the total protein added to each complete assay. (B) Lipid-binding assays performed for five protein mutants, based on X-ray data: C2AB-M1 and C2AB-M2, located in the C2B α-helices and those involving the SNAP25 Nt-regions (SNAP25-M1, SNAP25-M2, and SNAP25-M3). The Right panel shows the ratio of membrane-bound SNAP25/C2AB at 20 μM concentration. Data shown correspond to the mean of three independent experiments ±SD (n = 3). (C) Docking model of the Rph3A C2B–SNAP25 complex on a membrane (http://people.ucalgary.ca/∼tieleman/download.html, popc128a), based on the structures determined in this work. The SNAP25 helices are shown in green with the approximate position of the palmitoylation sites highlighted with a yellow circle. The C2B domain is shown in orange with the bound PIP2 in stick representation. a.u., asymmetric unit.
Fig. S6.
Fig. S6.
We produced recombinant C2AB-domain of Raph3A, SNAP25 (full length), and their corresponding mutants. (A) To discard direct SNAP25 interaction with membranes or protein aggregation effects interfering in the interpretation of the sedimentation assays, 20 μM of each of the SNAP25 constructs were incubated in the presence of 200 μM CaCl2 and lipid vesicles containing POPC/POPS/PIP2 (65:25:10 mol/mol). They were microcentrifuged to separate the soluble (S) and precipitated protein (P). Neither WT nor the mutants precipitate under the experimental conditions, indicating that they are properly folded and do not interact with the membrane. (B) To discard aggregation when both proteins are incubated simultaneously, we performed the lipid-binding assay under the same conditions with 200 μM CaCl2 and lipids containing POPC/POPS (75:25 mol/mol), a mixture known to have lower affinity for the C2AB domain than PIP2-containing vesicles. (C) To discard C2AB domain and mutant aggregation effects interfering in the interpretation of the sedimentation assays, 20 μM of each of the C2AB domain constructs were incubated in the presence of 200 μM CaCl2 in the absence of lipid vesicles and microcentrifuged to separate the soluble (S) and precipitated protein (P). Neither WT nor the mutants precipitate under these conditions, suggesting they are properly folded. (D) To check the binding affinities for lipid vesicles of independent C2AB and mutants, they were incubated in the presence of CaCl2 and lipid vesicles containing POPC/POPS/PIP2 (65:25:10 mol/mol). (E) Data show that all of them exhibit very similar affinities to bind PIP2-containing vesicles, confirming that the mutants are functional with respect to their PIP2 binding site at the polybasic region. Quantification of the bound proteins was performed with the Analyze/Gels plugging of Fiji ImageJ (40); values were normalized for the total protein added to each complete assay [dark gray bars correspond to the WT C2AB domain, very light gray to the C2AB-M1 (K651A/K656A/K663A), and light gray to the C2AB-M2 (K651A/K656A/K663A/H617A) mutants]. Data shown are mean of at least three independent experiments ±SD (n = 3). a.u., asymmetric unit, PC, phosphatidyl choline; PS, phosphatidyl serine.
Fig. S7.
Fig. S7.
The C2AB domain of Rph3A also binds to the SNARE complex through its α2-helix. (A) We produced recombinant SNAP25/STX1A, and its ability to interact with the C2AB and C2AB-M1 (K651A/K656A/K663A) mutant domains was tested by sedimentation assay. Twenty micromolar concentration of each construct was incubated in the presence of 200 μM CaCl2 and lipid vesicles containing POPC/POPS/PIP2 (65:25:10 mol/mol). Control experiments were performed in the presence of Ca2+ and lipid vesicles with either the complex or the C2AB construct alone to demonstrate the complex is not able to bind to the lipid vesicles under these conditions. (B) The same experiment was performed with other construct containing the complete SNARE complex (SNAP25/STX1A/VAMP2). (C) Relative quantification of the SNAP25/STX1A and SNARE complexes bound to the C2AB domain at 20 μM conditions. Data shown are mean of at least three independent experiments ± SD (n = 3). a.u., asymmetric unit.
Fig. 6.
Fig. 6.
Comparative binding models of the Rph3A C2B–SNARE and Sytl C2B–SNARE complexes on the plasma membrane. (A) Docking models of the Rph3A–SNARE (Left) and Sytl–SNARE (PDB ID 5CCG; Right) complexes on the membrane surface built considering the structural information available (this work and refs. , , and 24). The C2B domains (orange and yellow for the Rph3A and Sytl, respectively) have been accommodated on Top of a half-bilayer model membrane (PDB ID popc128a) with a PIP2 molecule (atom-type color sticks, carbon atoms in cyan) bound to their polybasic regions, using the Rph3A C2B–PIP2 complex as a reference. In Raph3A, this membrane-binding model fully exposes the α-helices at the bottom face of the C2B domain, interacting with SNAP25 (green). The STX1A (blue) and VAMP2 (red) helices have been modeled by structural alignment with the SNARE complex (PDB ID 1SFC). (B and C) Cartoon representations of two symmetric Rph3A C2B–SNAP25–STX1A–VAMP2 complexes docked to the plasma membrane by PIP2, SNAP25 palmitoylation (black line), and the STX1A transmembrane helix (blue line), before (B) and after (C) Ca2+ binding. For simplicity of the scheme, the C2A domain of Rph3A has not been represented. However, it is important to note that C2A also interacts with another PIP2 molecule and responds to Ca2+ at the membrane surface providing an additional anchorage and potential deformation point. (D and E) Comparative representation of the Sytl C2B–SNARE complex; in both cases, the Ca2+-induced membrane curvature is different but compatible with the formation of the fusion pore. For simplicity, many other proteins involved in the fusion process of synaptic vesicles like Munc18, Munc13, full-length STX1A, and complexin have not been represented; the C2A domain of Syt1 is not shown, but it is important to remark that it does not bind PIP2 (12) and permanently engages the synaptic vesicle through the transmembrane region making almost impossible its approach to the plasma membrane. Vesicle size for the main function of each protein has also been represented. In B, dcv refers to a dense-core vesicle in PC12 cells for Rph3A (14). In D, sv refers to a synaptic vesicle in hippocampal neurons for Syt-1 (3).

Similar articles

Cited by

References

    1. Brunger AT, Weninger K, Bowen M, Chu S. Single-molecule studies of the neuronal SNARE fusion machinery. Annu Rev Biochem. 2009;78:903–928. - PMC - PubMed
    1. Jahn R, Fasshauer D. Molecular machines governing exocytosis of synaptic vesicles. Nature. 2012;490:201–207. - PMC - PubMed
    1. Südhof TC. The molecular machinery of neurotransmitter release (Nobel lecture) Angew Chem Int Ed Engl. 2014;53:12696–12717. - PubMed
    1. Sutton RB, Fasshauer D, Jahn R, Brunger AT. Crystal structure of a SNARE complex involved in synaptic exocytosis at 2.4 Å resolution. Nature. 1998;395:347–353. - PubMed
    1. Corbalan-Garcia S, Gómez-Fernández JC. Signaling through C2 domains: More than one lipid target. Biochim Biophys Acta. 2014;1838:1536–1547. - PubMed

Publication types