Nature Structural Biology
9, 78 - 80 (2002)
doi:10.1038/nsb0202-78
Irreversible assembly of membrane fusion machinesDoug Barrick1
& Frederick M. Hughson21 Doug Barrick is in the Department of Biophysics, Johns Hopkins University, Baltimore, Maryland 21218, USA. 2 Frederick M. Hughson is in the Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544 USA.
Correspondence should be addressed to Doug Barrick barrick@jhu.edu or Frederick M. Hughson hughson@princeton.eduA large kinetic barrier prevents spontaneous disassembly of SNARE complexes, complicating thermodynamic measurements and resulting in hysteresis. Nonetheless, this kinetic barrier is probably crucial for SNARE function.Transport within eukaryotic cells makes use of vesicles that bud from one membrane and fuse with another1. The proteins most centrally involved in the attachment and fusion of vesicles with their membrane targets are called SNAREs2. SNARE proteins associated with a vesicle membrane (v-SNAREs) bind to SNAREs associated with a membrane target (t-SNAREs) to form complexes that bridge the two membranes (Fig. 1a). There is still some question as to whether these v/t-SNARE complexes are directly responsible for membrane fusion itself. Nonetheless, the structural similarity between SNARE complexes and viral membrane fusion proteins3, and the ability of purified v- and t-SNAREs to cause liposome fusion4, have led many in the field to conclude that SNARE complexes constitute at least a minimal machinery for membrane fusion.
 | | |
The assembly and disassembly of SNARE complexes has emerged as a theme in studies aimed at understanding the regulation of intracellular membrane fusion5. The architecture of SNARE complexes is simple6. Each SNARE protein has one, or occasionally two, 'SNARE motifs' of  60 residues, characterized by heptad repeats of hydrophobic residues that confer a high propensity to form  -helical coiled-coil structures7. Functional SNARE complexes have at their core a four-helix bundle of SNARE motifs. In the majority of intracellular fusion reactions, the four SNARE motifs are provided by four different SNARE proteins, at least one of which is anchored in each of the two fusing membranes.
Neurotransmitter release, which entails the fusion of synaptic vesicles with the plasma membrane, requires three SNAREs: synaptobrevin, syntaxin, and SNAP-25 (Fig. 1a). Synaptobrevin is a v-SNARE localized on synaptic vesicles whereas syntaxin and SNAP-25 are both t-SNAREs localized to the neuronal plasma membrane. The SNARE complex required for neurotransmitter release has at its core four -helical SNARE motifs, two from SNAP-25 and one each from synaptobrevin and syntaxin. The resulting complex has long been known to be exceptionally stable, based on the observation that its dissociation in vitro requires high temperatures and/or detergents8. The stability of the complex seems well suited to a role in bringing membranes into close apposition prior to fusion9. Disassembly, because it is energetically costly, requires a hexameric ATPase that acts after membrane fusion to recycle 'spent' SNAREs10.
Now, in work published on page 144 of this issue of Nature Structural Biology, Fasshauer and colleagues have used recombinant SNARE motifs from synaptobrevin and syntaxin, together with SNAP-25, to study SNARE assembly and disassembly in vitro (Fig. 1b). A major aim of these studies was to use thermodynamic methods to quantify the energetics of SNARE complex formation, thus getting a precise measure of the amount of energy available to bring membranes together in the process of fusion, and to learn about the energetic requirements for SNARE recycling.
Thermodynamic versus kinetic control
In quantitative studies of protein folding and assembly, many of us are used to (or at least talk about) systems that are under thermodynamic control — that is, systems that come to equilibrium in an experimentally convenient time range. Reaching equilibrium allows us to connect our data to thermodynamic parameters such as free energies of folding and binding. In turn, these parameters provide information about what populations are to be found under equilibrium conditions and how much work a system can do — important, for example, in considering the work of bringing two repulsive membranes together. Unfortunately, as biophysicists tackle bigger and more complex folding and assembly problems, it seems that the luxury of equilibrium thermodynamics may have to be sacrificed in exchange for functional and biological insight. As described in the paper by Fasshauer and colleagues11, this sacrifice seems unavoidable, at least for now, in studies of SNARE complex assembly and disassembly.
One of the hallmarks of systems under thermodynamic control is that as a system variable such as temperature is slowly changed, the system remains at equilibrium. The relative populations of different forms of the molecules may change, but these changes are prescribed to be the ones that minimize the free energy of the system. As a result, the same state of a system should be obtained irrespective of the history of the system. For example, in a protein assembly reaction, the same state of assembly should be produced whether the sample was heated or cooled to reach the specified temperature. In contrast, the extent of synaptic SNARE complex formation appears to be path-dependent11. When fully formed synaptic SNARE complexes are heated, they resist dissociation and unfolding until very high temperatures, but when these same thermally unfolded SNAREs are cooled back down, they remain unfolded and dissociated until temperatures are well below those required for unfolding (Fig. 2). Thus, the heating and cooling curves give rise to an open loop, such that at intermediate temperatures, the system can exist in either of two metastable states — one folded and assembled and one unfolded and dissociated — depending on whether the system is being heated or cooled. Such open loops are often referred to as 'hysteresis loops' and are observed in two different SNARE complexes with distinct sequences, stoichiometries and trafficking roles11.
| | |
Hysteresis
Hysteresis is seen in a wide variety of different systems and processes. Examples of simple chemical transformations that display hysteresis are magnetization of some metals, including iron (Fig. 2), and elastoplastic deformation of solids12,
13. Several biological systems also show hysteresis under various transformations. One of the first biological hysteresis loops identified was that associated with proton binding in myoglobin14. Hysteresis explains the effect of antifreeze peptides on the melting and freezing points of ice: although common sense (and equilibrium thermodynamics) would predict these two temperatures to be the same, in the presence of antifreeze peptide, ice consistently freezes at a lower temperature than it melts15. Other examples of hysteresis in biological macromolecules include thermally driven melting and annealing of DNA, and force-induced folding and unfolding transitions of proteins and RNA16,
17.
Although the molecular origins of these hysteretic processes are obviously quite different, there is one common feature: the existence of large kinetic barriers that separate microscopic configurations. In all cases, the molecules and their assemblies lack sufficient kinetic energy to pass over these barriers in an accessible time frame, and appear to be frozen or trapped, even though they may not exist in the lowest energy state. During iron magnetization, the kinetic barriers result from impurities that prevent domain walls from moving freely. During conformational transitions in macromolecules such as SNARE complexes, these kinetic barriers may correspond to high energy ensembles through which molecules must pass to get back and forth between folded, assembled conformations and unfolded, dissasembled ones.
There is an important difference between the hysteresis loops of simple chemical systems and SNARE assembly. In simple processes such as iron magnetization, the hysteresis loops are often referred to as 'rate-independent', which means that the same curve is drawn out if the loop is cycled at a slower rate13. In contrast, for the loop associated with SNARE assembly, it appears that slower heating and cooling rates narrow the loop, decreasing the hysteresis effect11. Technically speaking, the hysteresis loops of simple processes such as magnetization would also become narrower and eventually disappear as the rate of loop cycling approached zero18, but this effect is experimentally unobservable. The apparent assembly hysteresis observed by Fasshauer et al.11, although it lacks the time independence characteristic of this phenomenon in simpler systems, is nonetheless striking in its persistence, and important for design and interpretation of experiments on SNARE complex assembly.
Origins of apparent SNARE complex hysteresis
Since the origins of hysteresis are rooted in rates of interconversion, one approach to understanding hysteresis is to investigate the kinetics of interconversion. Fasshauer and colleagues11 made a series of kinetic measurements of SNARE complex unfolding at very high guanidine hydrochloride concentrations, and of folding at low guanidine hydrochloride concentrations. Indeed, results from the unfolding studies suggest a very large kinetic barrier to disassembly. Extrapolating to zero guanidine concentration, the authors estimate an unfolding rate constant of 4  10-18 s-1, indicating that billions of years would be required for significant unfolding. As the authors point out, the guanidine extrapolation is a long one, and it is possible that other pathways with lower barriers would produce faster kinetics in the absence of guanidine. Nonetheless, this potentially large barrier could easily prohibit equilibration on the laboratory timescale, resulting in apparent hysteresis. The existence of this barrier also provides a rationale for the hexameric ATPases that are required for disassembly of SNARE complexes in vivo10.
Assuming the disassembly barrier puts SNARE complex formation under kinetic control at low denaturant concentrations, it seems reasonable to ask what the molecular origins of this slow step are. Dissociation rate constants for short (33 residues), dimeric coiled coils have been determined by several groups, and are by comparison blazingly fast (0.4−10 s-1, refs 19,
20,
21). However, disruption of the coiled coils within the SNARE complex is likely to be significantly slowed by their greater length of interaction and by the presence of two coiled-coil interactions per helix rather than one.
In addition to the slow kinetics displayed for SNARE complex disassembly, Fasshauer and colleagues11 find evidence for complexity in SNARE complex assembly. The progress curves for assembly appear to show two distinct phases associated with helix formation, each with a unique dependence on guanidine and SNARE concentration. The authors suggest that the fast phase in the assembly reaction may result in the formation of a bundle made of only two of the three types of polypeptides (both t-SNARES) found in the intact SNARE complex (Fig. 1b). The more modest stability of this kinetic intermediate explains why ternary complexes fail to form in a reasonable amount of time under conditions in which, although the ternary complexes are stable, binary complexes are not. Although the details of assembly kinetics are far from understood, it seems likely that a complete analysis of the kinetics of SNARE assembly will provide much insight into the mechanisms by which these fusogenic complexes are built.
Thermodynamics from kinetics?
The experiments reported by Fasshauer and colleagues11 demonstrate the existence of a large kinetic barrier in the disassembly of SNARE complexes and suggest that binary syntaxin−SNAP-25 complexes are kinetic intermediates in the formation of ternary membrane-bridging complexes. What then about the thermodynamic stability of the ternary complex? This quantity is of key importance because the fusion activity of SNARE complexes probably depends on their stability. Given that SNARE complex assembly is under kinetic rather than thermodynamic control, the assembly free energy remains elusive. One way forward is through kinetics itself. For simple reactions, equilibrium constants can be expressed as ratios of forward and reverse rate constants. For complex reactions like SNARE assembly, connecting rates with equilibrium constants presents a greater challenge. Hurdles to be surmounted include the extrapolation of rate constants to the same conditions and accurate accounting for intermediates that form during assembly. In the meantime, Fasshauer and colleagues have been able to gain significant traction in studying a difficult and complex protein assembly reaction of clear biological significance.
REFERENCES
- Jahn, R. & Südhof, T.C. Annu. Rev. Biochem. 68, 863−911 (1999). | Article | PubMed | ISI | ChemPort |
- Söllner, T. et al. Nature 362, 318−324 (1993). | Article | PubMed | ISI | ChemPort |
- Skehel, J.J. & Wiley, D.C. Cell 95, 871−874 (1998). | Article | PubMed | ISI | ChemPort |
- Weber, T. et al. Cell 92, 759−772 (1998). | Article | PubMed | ISI | ChemPort |
- Waters, M.G. & Hughson, F.M. Traffic 1, 588−597 (2000). | Article | PubMed | ISI | ChemPort |
- Sutton, R.B., Fasshauer, D., Jahn, R. & Brunger, A.T. Nature 395, 347−353 (1998). | Article | PubMed | ISI | ChemPort |
- Weimbs, T. et al. Proc. Natl. Acad. Sci. U.S.A. 94, 3046−3051 (1997). | Article | PubMed | ChemPort |
- Hayashi, T. et al. EMBO J. 13, 5051−5061 (1994). | PubMed | ISI | ChemPort |
- Hanson, P.I., Roth, R., Morisaki, H., Jahn, R. & Heuser, J.E. Cell 90, 523−535 (1997). | Article | PubMed | ISI | ChemPort |
- May, A.P., Whiteheart, S.W. & Weis, W.I. J. Biol. Chem. 276, 21991−21994. (2001). | Article | PubMed | ISI | ChemPort |
- Fasshauer, D., Antonin, W., Subramaniam, V. & Jahn, R. Nature Struct. Biol. 9, 144−151 (2002). | Article | PubMed | ISI | ChemPort |
- Feynman, R.P., Leighton, R.B. & Sands, M. The Feynman Lectures on Physics , Vol. II (Addison-Wesley Publishing Co., Reading, Massachusetts; 1964).
- Brokate, M. & Sprekels, J. Hysteresis and Phase Transitions (Springer-Verlag, New York; 1996).
- Steinhardt, J. & Zaiser, E.M. Adv. Protein Chem. 10, 151−226 (1955). | PubMed | ISI | ChemPort |
- Knight, C.A., DeVries, A.L. & Oolman, L.D. Nature 308, 295−296 (1984). | Article | PubMed | ISI | ChemPort |
- Kellermayer, M.S., Smith, S.B., Granzier, H.L. & Bustamante, C. Science 276, 1112−1116 (1997). | Article | PubMed | ISI | ChemPort |
- Liphardt, J., Onoa, B., Smith, S.B., Tinoco, I.J. & Bustamante, C. Science 292, 733−737 (2001). | Article | PubMed | ISI | ChemPort |
- Adkins, C.J. Equilibrium Thermodynamics (Cambridge University Press, Cambridge; 1997).
- Wendt, H., Berger, C., Baici, A., Thomas, R.M. & Bosshard, H.R. Biochemistry 34, 4097−4107 (1995). | Article | PubMed | ISI | ChemPort |
- Ibarra-Molero, B., Makhatadze, G.I. & Matthews, C.R. Biochemistry 40, 719−731 (2001). | Article | PubMed | ISI | ChemPort |
- Sosnick, T.R., Jackson, S., Wilk, R.R., Englander, S.W. & DeGrado, W.F. Proteins 24, 427−432 (1996). | Article | PubMed | ISI | ChemPort |
- Carr, C.M. & Novick, P.J. Nature 404, 347−349 (2000). | Article | PubMed | ISI | ChemPort |
|
