Structural Studies of Alzheimer's beta-Amyloid Fibrils

2007财政年度在以下领域取得了进展： 1.三重对称β-淀粉样纤维的结构模型：在早期的工作中（Petkova等人，Science 2005），我们发现40个残基的β-淀粉样肽可以形成多于一种的原纤维结构，主要的原纤维结构可以通过生长条件的细微变化来确定，在电子显微镜图像中具有不同形态的原纤维具有不同的潜在分子结构，并且当预先存在的原纤维片段用作新原纤维生长的种子时，形态和分子结构是自繁殖的。 在早期的工作中（Petkova等人，Biochemistry 2006），我们还开发了β-淀粉样蛋白原纤维的完整结构模型，所述β-淀粉样蛋白原纤维在β-淀粉样蛋白溶液的温和搅拌下形成，并且具有通过扫描透射电子显微镜（与R. D. Leapman，DBEPS，NIBIB），对应于交叉β结构基序中的两个分子层。 我们发现，在静态溶液条件下生长的原纤维具有对应于三个分子层的MPL值，但我们早期的固态NMR数据表明，两个或更多个静态原纤维结构可能共存。 现在，在2007财政年度，我们成功地制备了结构高度均匀的β-淀粉样蛋白原纤维，其MPL值对应于三个分子层，并且我们成功地从固态NMR中获得了足够的结构约束，以开发完整的分子模型。 而我们以前发表的模型搅拌原纤维有两重对称的长原纤维轴，新的模型（适用于不同的原纤维类型）有三重对称。 两重和三重对称结构中的β-淀粉样肽构象相似，具有基本相同的β-链区段和相同的对齐平行β-折叠三级结构，但非β-链区段的构象不同，β-折叠之间的接触也不同。 描述新的β-淀粉样蛋白纤维结构的手稿目前正在编写中。 对三重对称纤维的进一步测量正在进行中，包括与美国科学院合作的冷冻电镜研究。A.C.来自NIAMS的Steven 2.与R.B.合作在NIDDK的Wickner的小组中，我们已经进行了由酵母朊病毒蛋白Ure 2 p和Sup 35 p形成的淀粉样纤维的相关研究。 这是第一个固态核磁共振研究全长朊病毒域的酵母朊病毒。 在Sup 35 NM（Sup 35 p的N-末端和中间结构域）的实验中，我们发现Sup 35 NM原纤维是由对齐的平行β-折叠构成的，并且不具有β-螺旋结构，这已经由其他组的早期X射线纤维衍射和荧光标记实验提出。 有趣的是，我们的固态NMR数据表明Sup 35 NM中的β-折叠并不完全位于N-末端、富含Asn和Gln的结构域（预期其形成淀粉样蛋白原纤维核心）中，而是也在M结构域的部分中发现。 这项工作已发表在PNAS，第103卷，第19754页，2006年。 M结构域中的β-折叠形成可能是PSI朊病毒变体或毒株之间的区别性结构特征，其由细胞内Sup 35 p淀粉样蛋白形成诱导。 这种可能性得到了最近来自UCSF的Weissman小组的H/D交换数据的支持，该数据表明不同Sup 35 NM淀粉样蛋白变体在M结构域中具有不同程度的H/D交换保护。 我们还进行了Ure 2 p的朊病毒结构域（残基1-89）的固态NMR研究。 Ure 2 p（1-89）原纤维还含有对齐的平行β-折叠。 这项工作在一份手稿中进行了描述，目前正在为生物化学进行审查。

Progress in FY2007 has been in the following areas: 1. STRUCTURAL MODEL FOR THREE-FOLD SYMMETRIC BETA-AMYLOID FIBRILS: In earlier work (Petkova et al., Science 2005), we discovered that the 40-residue beta-amyloid peptide can form more than one fibril structure, that the predominant fibril structure can be determined by subtle variations in growth conditions, that fibrils with distinct morphologies in electron microscope images have distinct underlying molecular structures, and that morphology and molecular structure are self-propagating when pre-existing fibril fragments are used as seeds for the growth of new fibrils. In earlier work (Petkova et al., Biochemistry 2006), we have also developed a full structural model for beta-amyloid fibrils that form under gentle agitation of a beta-amyloid solution and that have mass-per-length (MPL) values, determined by scanning transmission electron microscopy (collaboration with R.D. Leapman, DBEPS, NIBIB), corresponding to two molecular layers in a cross-beta structural motif. We found that fibrils grown under quiescent solution conditions have MPL values that correspond to THREE molecular layers, but our earlier solid state NMR data indicated that two or more quiescent fibril structures might coexist. Now, in FY2007, we have succeeded in preparing highly structurally homogeneous beta-amyloid fibrils with MPL values corresponding to three molecular layers, and we have succeeded in obtaining sufficient structural constraints from solid state NMR to develop a full molecular model. Whereas our previously published model for agitated fibrils had two-fold symmetry about the long fibril axis, the new model (which applies to a different fibril type) has three-fold symmetry. The beta-amyloid peptide conformations in two-fold and three-fold symmetric structures are similar, with essentially the same beta-strand segments and the same in-register parallel beta-sheet tertiary structure, but the conformations of non-beta-strand segments are different, and the contacts between beta-sheets are also different. A manuscript describing the new beta-amyloid fibril structure is currently in preparation. Additional measurements on three-fold symmetric fibrils are in progress, including cryo-EM studies in collaboration with U. Baxa and A.C. Steven of NIAMS. 2. In collaboration with R.B. Wickner's group in NIDDK, we have carried out related studies of amyloid fibrils formed by the yeast prion proteins Ure2p and Sup35p. These are the first solid state NMR studies of full-length prion domains of yeast prions. In experiments on Sup35NM (the N-terminal and middle domains of Sup35p), we showed that Sup35NM fibrils are constructed from in-register parallel beta-sheets and do not have a beta-helical structure, which had been suggested by earlier x-ray fiber diffraction and fluorescence labeling experiments in other groups. Interestingly, our solid state NMR data indicate that the beta-sheets in Sup35NM are not localized exclusively in the N-terminal, Asn- and Gln-rich domain (which was expected to form the amyloid fibril core), but are also found in parts of the M domain. This work has been published in PNAS, vol. 103, p. 19754, year 2006. Beta-sheet formation in the M domain may be the distinguishing structural feature among PSI prion variants or strains, which are induced by intracellular Sup35p amyloid formation. This possibility is supported by recent H/D exchange data from the Weissman group at UCSF, which indicate different degrees of H/D exchange protection in the M domain for different Sup35NM amyloid variants. We have also carried out solid state NMR studies of the prion domain (residues 1-89) of Ure2p. Ure2p(1-89) fibrils also contain in-register parallel beta-sheets. This work is described in a manuscript that is currently under review for Biochemistry.