Hypothesis - buttressed rings assemble, clamp, and release SNAREpins for synaptic transmission

Abstract

Neural networks are optimized to detect temporal coincidence on the millisecond timescale. Here, we offer a synthetic hypothesis based on recent structural insights into SNAREs and the C2 domain proteins to explain how synaptic transmission can keep this pace. We suggest that an outer ring of up to six curved Munc13 'MUN' domains transiently anchored to the plasma membrane via its flanking domains surrounds a stable inner ring comprised of synaptotagmin C2 domains to serve as a work-bench on which SNAREpins are templated. This 'buttressed-ring hypothesis' affords straightforward answers to many principal and long-standing questions concerning how SNAREpins can be assembled, clamped, and then released synchronously with an action potential.

Keywords: SNARE; membrane fusion; synaptic transmission.

Figure 1

Recent structural insights into the protein machinery involved in synaptic transmission. (A) 3D‐reconstruction of the ring‐like oligomers formed by Syt1, the primary Ca²⁺ sensor for neurotransmitter release at the synapse 21, 22, 23. These oligomers, which are typically 20–35 nm in diameter (12–20 copies), assemble based on the interactions between the C2B domains (gray), with the C2A (cyan) domain locating outside the ring structure. The Ca²⁺ binding loops (red dots) are involved in oligomer formation and are located within the C2B‐C2B interfaces and thus render the oligomers sensitive to Ca²⁺. (B) Crystal structure of a prefusion SNARE‐Cpx‐Syt1 complex revealed that two Syt1 molecules can bind to opposite sides of the same SNAREpin, each via its C2B domain 24. One Cpx‐independent (termed ‘primary’) interface involves contacts with both SNAP25 helices (green); the other Cpx‐dependent interface (termed ‘tripartite’) involves contacts with helices derived from Cpx (cyan), syntaxin (red), and VAMP2 (blue). (C) Electron microscopy‐derived structure of a purified synaptophysin‐VAMP2 complex reveals a hexameric ring architecture wherein six synaptophysin (magenta) molecules bind and organize six VAMP2 (blue) dimers such that they are directionally oriented toward the target membrane 26. The ring is stabilized by contacts among transmembrane helices within the bilayer. (D) The domain arrangement of Munc13‐1 protein, which involves C1, C2B, and C2C domain flanking its widely conserved MUN domain. The amino terminal C2A domain is not shown. The crystal structure of the Munc13‐1 MUN domain consists of an elongated, arch‐shaped structure formed by α‐helical bundles, with a highly conserved hydrophobic pocket approximately in the middle (yellow highlights) 27. Mutations in this region (residues in red) compromise Munc13 ability to chaperone SNARE assembly 28, 29. Figure adapted from Ref. 24, 26, 27.

Figure 2

Buttressed rings of Munc13 and Syt are proposed to act as work stations for SNAREpin assembly, clamping, and release. (A) Proposed hexameric end‐to‐end organization of MUN domain of Munc13 to form a flat, ring‐like structure. The view from the top with the PM (blue) below is shown. (B) The hypothetical MUN domain ring closely encloses a Syt ring‐oligomer formed by 18 Syt C2B domains. This concentric ring organization naturally aligns the hydrophobic pocket in the middle of MUN domain with the primary interface of every third Syt1 C2B (both shown in yellow highlights). (C) When the SV (not shown for clarity) approaches the PM, the hexagonally arranged VAMPs (blue) organized by synaptophysin (magenta) are suitably positioned to reach the proposed active surface of each MUN subunit to template the SNARE assembly. (D) As a result, a total of six SNAREpins are proposed to be assembled spanning the outer and inner rings with their carboxy‐terminal transmembrane domains (marked by ‘C’) inside the rings inserted above and below into SV and PM, respectively.

Figure 3

Clamping of SNAREpin terminal zippering by radial retention according to the buttressed‐ring hypothesis. The assembling SNAREpins are retained on the inner Syt ring via the ‘primary’ SNARE‐C2B domain interaction. In the ring oligomers, the second independent C2B domain (magenta) sits above the SNAREpin bound via the ‘tripartite’ binding site in conjunction with Cpx. Such an arrangement would allow the ‘tripartite’ C2B to bind the SV membrane, likely via lipid interaction. As a result, each SNAREpin is held in a vice‐like clamp between the two C2B domains, each bound to and buttressed by the opposing membranes preventing further zippering. Cpx likely strengths this clamp by additionally anchoring its SNAREpin to the SV membrane to which it is bound via its C‐terminal domain (not shown). Note that in the ring arrangement, the conserved helical extension in Syt C2B (red arrow), which is the basis of tripartite binding faces the PM. Both the side view (A) and the top view (B) are shown. Note: for clarity, the outer ring of Munc13 is not shown.

Figure 4

Clamping of SNAREpin terminal zippering by enforced spatial separation according to the buttressed‐ring hypothesis. (A) Terminology to describe the anatomy of the fully assembled SNARE complex four‐helix bundle 45. The hydrophobic layers (alternately consisting of 3 and 4 residues each) are numbered from −7 to +8. The locations of the residues contacting Syt1 C2B in the ‘primary’ Syt binding site (red dots) suggest that the SNAREpins have to be zippered up to or beyond the +3 layer to accommodate Syt binding. (B) Modeling shows that given the elevated positioning of the SNAREpin atop the Syt ring, the residues of the SNAP25 SN1 helix that assemble into layers +5 to +8 (red spots) to complete zippering and trigger fusion would instead need to be nearly fully extended in this geometry to enable the adjacent Cys palmitoylation sites (yellow dots) to be inserted into the PM bilayer (blue). This suggests that zippering beyond layer +4 will be impeded due to the spatial separation between the membranes imposed by the ring. (C) The SNAREpins on the Syt ring are held at an angle such that even if they were to fully zipper their four helix bundles to layer +8, the tips of these bundles would be positioned inside the ring too far out radially (~ 17 nm from +8 tip to +8 tip) from the center to enable fusion (see Appendix 1 for details). Thus, the Syt ring also radially restrains the full zippering of the SNARE complex. Upon the influx of Ca²⁺, the Syt ring oligomers are disrupted as Syt molecules rotate to insert into the PM. This frees the SNAREpins to complete zippering and move in radially to open the initial fusion pore.

Figure A1

(A) Short range repulsion between two DOPC/DOPE (60/40) lipid bilayers. The values are based on published data [47–49]. On the right axis is plotted the corresponding energy when the pressure is applied over a 13 nm² area, i.e. a 4 nm disk. Fusion occurs when the pressure reaches 440 atm, or equivalently 35 k _B T. (B) The ring model as viewed from top. Six SNAREpins are uniformly distributed over the C2B inner ring. The inner diameter of the ring is 19 nm and the outer diameter 25 nm. Layer −1 of the SNAREpin is located right at the mid‐point (22 nm diameter) of the ring. The SNAREpin is positioned at an angle α = 40° to the tangent of the ring and the inner extension between layer −1 and the end of the SNARE domain (layer +8) is at L = 7 nm. Layer +8 of the six SNAREpins are located on a circle with a diameter s ₀ = 17 nm as calculated in the text of the Appendix. (C) Side view. The two membranes (SV in red and PM in blue) are pulled in towards each other by the SNAREpins. At equilibrium, the inter‐membrane distance, defined as the water layer thickness, is h. (D) Energy landscape corresponding to the process of zippering of the last several layers of the SNAREpin. Only the values of the minima and the maximum have been reported [18]; their relative positions are unknown. Here, we assume this process occurs over 5 nm and that it is harmonic. The shape of the contour was chosen to achieve continuity in energy and force at the midpoint between two consecutive extremes. Note that this includes the zippering of the C‐terminal linker domains, immediately after the SNARE domains. The inter‐membrane distance is assumed equal to 0 when the linker domains are fully zippered. (E) Zoom‐in view of the short‐range repulsion energy (panel A) and the SNAREpin zippering for six SNAREpins (6× panel D) between the transition distance (0.6 nm) and 1.5 nm with energy values at 1.5 nm for clarity. The sum of the two curves provides the actual energy landscape produced by the SNAREpins and the short‐range repulsion. An intermediate state appears at 0.7 nm and the activation energy to transit from this intermediate state to the fused state is 4.6 k _B T.

Figure A2

Relative position of the synaptic vesicle (SV) and the plasma membrane (PM) prior to fusion (left). Upon the action of the SNAREpins, the synaptic vesicle moves towards (arrow) and coalesces with the plasma membrane. The positions after fusion are shown at right. The overlapping regions have mixed. Because they are up to 1000 times more viscous than water, the slowest movement comes from the displacement of lipids themselves. Hence, upon the collective action of SNAREpins, the limiting factor in the diffusion‐limited fusion process is the movement of the synaptic vesicle over the thickness of a bilayer (5 nm). The radius of the impacted region is r _fus = 15 nm. The black box represents the contour of the disk over which the calculation of the viscous force is made (see text).

Figure A3

Movements of a C2B domain after its detachment from the ring: (A) C2B monomer is modeled as a cylinder of diameter e = 5 nm and length l = 5 nm that rotates around its axis with an angular velocity ω. The direction of the axis is dictated by SNAREpin zippering, which orients the movement. (B) Loop insertion. The aliphatic loops flanking that Ca²⁺ binding site, located just above the cytoplasmic leaflet of the PM, are modeled as inserting by moving 1 nm down vertically into the PM after Ca²⁺ binds. The shape of the loops is reasonably modeled as a cylinder 1 nm long and 2.5 nm in diameter.

Figure A4

(A) Top view. A SNAREpin is released from the ring. Initially it was positioned on the ring (grey). As terminal zippering proceeds, the SNAREpin moves so as to position itself radially with its transmembrane domains (inserted above into the SV and below into the PM) towards the center of the ring (light orange). During the process, the movement of the transmembrane domain is approximated by a straight line, wherein x is the displacement from the initial position. (B) Side view (perpendicular to the SNAREpin). The contribution of the zippering force, F, to the movement corresponds to its projection tangentially to the SV, f.

Figure A5

The SNAREpin can be in two states: bound to C2B or free. The energy of the free‐state is 0 and that of the bound state −21 k _B T. An additional energy barrier, ΔE _a = 5 ± 3 k _B T suffices to explain the observed spontaneous release rate.