Discussion
Our goal has been to analyze whether most experimental results concerning Synuclein oligomers, protofibrils, and transmembrane channels are consistent with relatively simple β-barrel assemblies in which all monomers have approximately identical conformations and interactions. We are as concerned with understanding the overall assembly process as with modeling any specific oligomer structure. Figs. 17 and 25 outline a plausible, if somewhat speculative, series of transitions leading from trimers and tetramers, to larger oligomers, lipoproteins, protofibrils, fibrils, and transmembrane channels.
Although we attempt to make our models and hypotheses as simple as possible, they are more complicated and ambiguous than we would have preferred. Nonetheless, they are no more complex than required to explain experimental results; if anything they are too simple. Our models typically have a higher content of secondary structure than reported experimentally. As previously described for the helical trimers and tetramers of Figs. 5 and 18, we suspect that the secondary structures of these more complex models are also only fractionally occupied. A major point of Figs. 17 and 25 is that the assemblies are dynamic and unlikely to be as structured as the models imply, especially as they merge or morph into other assemblies or break down. Thus we view our models as idealized hypothetical approximations of time-averaged structures for some of the more stable stages of synuclein interactions; i.e. for specific assemblies that can be identified microscopically and/or isolated biochemically.
Tight packing between adjacent β-barrels in our models may account for some features of synuclein sequences that, to our knowledge, have not been explained previously: e.g., the prevalence of small apolar side-chains, especially glycine and alanine, but also valine, and small ambivalent threonine and serine within putative β-strands and the preference for valine over the larger isolucine and leucine side-chains. The high degree of conservation of these residues among the three synuclein families suggests a vital role for these residues, but that role is unlikely to involve fibril formation since β-Syn does not form fibrils. A novel feature of our concentric β-barrel hypothesis is the concept that intermeshing-pleats facilitates tight packing between adjacent concentric β-barrels in some assemblies. The presence of glycines may reduce steric clashes when this type of packing does not occur; i.e., if pleates cross one another. Another striking feature of synuclein sequences is the distribution of charged residues. Although there is evidence that the negatively charged Ct domain is important to some interactions and assemblies, its purpose has not been explained. Here we suggest that its negative charges interact with positively charged components of some zitterionic lipid headgroups, particularly in high-density lipoproteins, but possibly in membranes as well. Highly conserved positively charged lysines of the putative signature turn regions are proposed to interact with negatively charged components of lipids and fatty acids.
The major motivation for analyzing amyloid structures is to develop effective ways to prevent, treat, and/or cure the devastating diseases they cause. Those of us who study molecular structure want to know which assemblies are functional, which are benign, and which are toxic. Then perhaps we can eliminate or neutralize toxic culprits without hindering vital functions. Low resolution EM images and hypothetical models will not suffice; but they may help achieve these goals.
Lysenin and Aerolysin toxins share several properties with the concentric β-barrel models we have proposed previously for Aβ42 and propose here for synuclein amyloid oligomers: 1) they are composed of multiple identical subunits related by radial symmetry, 2) they contain concentric β-barrel structures, 3) diameters and tilt angles of their β-barrels are consistent with β-barrel theory, 4) they exist both in soluble and transmembrane channel conformations, and (5) AFM images of the channel conformations are similar. However, amyloids are substantially more polymorphic and less stable than the two toxins; so much so that they are often said to be intrinsically disordered. These factors complicate experimental determination of their oligomeric structures. In the Appendix we list some examples of experimental approaches suggested by our models. These include experiments to: reduce polymorphism and disorder by stabilizing specific oligomer structures, determine which assemblies are toxic and which portions of those assemblies are responsible for the toxicity, to obtains additional structural images, and to develop techniques and/or drugs to treat PD and other Synuclein-related diseases.
We hope our hypotheses and models will assist efforts to obtain the detailed evidence that can lead to medical breakthroughs.