Sawaya MR, Sambashivan S, Nelson R, Ivanova MI, Sievers SA, Apostol MI, Thompson MJ, Balbirnie M, Wiltzius JJ, McFarlane HT, Madsen AØ, Riekel C, Eisenberg D.
Atomic structures of amyloid cross-beta spines reveal varied steric zippers.
Nature. 2007 May 24;447(7143):453-7.
Please login to recommend the paper.
To make a comment you must login or register.
New X-ray Crystallographic Data Reveal Variable Structures in Generic Amyloids
Among the “holy grails” in science has been the atomic-resolution determination of amyloid structure. Amyloids, the proteinaceous polymers whose accumulation in the extracellular spaces of tissues causes over 20 kinds of progressive local and systemic diseases, have been notoriously difficult to study. This difficulty derives in large part from the propensity of amyloid proteins to form fibrils that are not amenable to classical X-ray crystallographic study. Until recently, the best models of amyloid structure had come from solid-state NMR studies and low-resolution X-ray diffraction experiments. However, in 2005, David Eisenberg’s group at UCLA reported the successful growth and X-ray study of microcrystals (~1 µm) formed by small amyloidogenic peptide repeats, GNNQQNY and NNQQNY, found in the yeast prion Sup35 (1). Reflections were observed that were consistent with the known “cross-β” organization of amyloid fibrils, namely β-strands ~5 Å apart that are perpendicular to the fibril axis, and extended β-sheets ~10 Å apart that are parallel to the fibril axis. The quantum leap in knowledge provided by these studies was the determination of the fine-structure of the crystalline assembly, a pair of β-sheets whose side-chains interdigitated in an anhydrous crystal core. This structure resembled a zipper in which the interdigitating teeth were the side-chains of each β-sheet; hence, the structure was termed a “steric zipper.”
(Students of recent science history may find it illuminating to compare the steric zipper nomenclature and model to the “knobs in holes” packing postulated by the late Francis Crick in 1953 to explain helix packing in α-keratin .)
Now, in exceptionally broad follow-up work, Eisenberg’s group reports the results of similar studies of 13 amyloid-forming peptide segments derived from Aβ, tau, prion protein, insulin, islet amyloid polypeptide, lysozyme, myoglobin, α-synuclein, and β2-microglobulin (3). Segment length ranged from four (NNQQ from Sup35) to 29 (Aβ[12-40]) amino acids, and in the majority of cases, these peptides formed both microcrystals and fibrils. Interestingly, fibrils could be observed emanating from crystals, lending support to the relevance of the microcrystal structure for understanding fibril formation.
High-resolution structures obtained from the microcrystals all resembled the one observed in the GNNQQNY and NNQQNY microcrystals, i.e., a steric zipper (1). This definition of this common motif extends our understanding of the atomic organization of the classic cross-β structure and reveals the structural elements comprising the “generic” core (spine) of amyloid assemblies (4).
It should be noted that some have argued that the non-homologous nature of amyloid proteins, which produces sequence dissimilarity among amyloid-forming protein segments, should preclude a generic amyloid structure. The results of Sawaya et al. also may explain this apparent contradiction as they show that the fine-structure of the cross-β motif is remarkably variable. In fact, eight different classes of steric zipper may be envisioned. These differ in the orientation of strands within and between each β-sheet and how the sheets themselves are oriented with respect to the fibril axis. Five of the proposed classes were observed experimentally. Not only do multiple classes of steric zipper exist, but a number of peptides were identified that could populate different structural classes. Closely related sequences (e.g., overlapping peptide segments) also populated different classes. Thus, when analyzed at sufficient detail, one can appreciate that not only can different peptide provenance produce different (non-generic) assembly structures, but even the same peptide can form different structures.
Furthermore, structural variability in steric zippers may explain one of the most perplexing questions in an already perplexing field, i.e., the existence of prion “strains.” How can a “protein-only” pathogen with a single primary structure perpetuate itself in different pathologic forms that “breed true?” Formation of different classes of steric zippers by the same peptide segment offers an explanation.
In summary, the work of Eisenberg et al. represents a quantum leap in understanding amyloid assembly structure and provides a foundation for future biophysical and biological experiments. It is important to note, however, that the amyloid “atom” has an indeterminate number of quantum levels. Their complete traversal will be required if a full understanding of this difficult and clinically important protein assembly problem is to be obtained. For example, amyloid assembly in vivo does not involve the self-association of naked tetrapeptides in PBS but is a much more complicated process. Many amyloid proteins undergo post-translational modification, producing, for example, glycosylated or phosphorylated amyloid precursors. Such primary structure components may have significant effects on assembly kinetics and structure. It also should be emphasized that amyloid fibrils comprise more than a “spine.” One would like to define the structure of amyloid assemblies formed by the holoprotein, not just the core, as it is these non-core components that may mediate intermolecular and cellular interactions of relevance to disease, including control of fibril or prion formation by chaperone proteins and cellular toxicity.
References:1. Nelson R, Sawaya MR, Balbirnie M, Madsen AO, Riekel C, Grothe R, Eisenberg D. Structure of the cross-beta spine of amyloid-like fibrils.
Nature. 2005 Jun 9;435(7043):773-8.
2. Crick FHC. The Packing of α-Helices - Simple Coiled-Coils. Acta Crystallogr. 1953;6:689-697.
3. Sawaya MR, Sambashivan S, Nelson R, Ivanova MI, Sievers SA, Apostol MI, Thompson MJ, Balbirnie M, Wiltzius JJ, McFarlane HT, Madsen AO, Riekel C, Eisenberg D. Atomic structures of amyloid cross-beta spines reveal varied steric zippers.
Nature. 2007 Apr 29; [Epub ahead of print]
4. Dobson CM. Protein folding and misfolding.
Nature. 2003 Dec 18;426(6968):884-90. Review.
Assembly of amyloid proteins into potentially neurotoxic
aggregates is believed to be the leading cause of a number
of neurodegenerative disorders, including Alzheimer's,
Parkinson's, Creutzfeldt-Jakob, and others. All
amyloid proteins form extracellularly deposited amyloid
fibrils and bind Congo red. Amyloid and amyloid-like
fibrils also share cross-β X-ray diffraction pattern
with β-sheets parallel to the fibril axis and extended
protein strands perpendicular to the axis. Sawaya et al.
apply X-ray microcrystallography to 30 short segments of
fibril-forming proteins, including segments from
Alzheimer's amyloid-β and tau proteins, to reveal a
structure with each pair of β-sheets having facing side
chains arranged into a steric zipper. These “dry'' steric
zippers present the common structural feature shared by
The authors build this study on the hypothesis that
microcrystal structures reveal structural features of
amyloid fibrils. They use microcrystals of short peptides
because these are capable of revealing detailed atomic
structures. The question of how closely the microcrystalline
structure resembles the atomic structure of the fibril
still remains to be fully addressed. Sawaya et al.
presented evidence of structural similarity of the two
X-ray diffraction patterns and similarity in nucleation
kinetics of microcrystal and fibril formation. In
addition, they showed that mutations in fibril-forming
segments diminish fibril formation of full-length proteins,
indicating that microcrystals and fibrils share some key
structural and kinetic features.
Sawaya et al. classified steric zippers of about 30 protein
segments into eight classes, depending on the arrangement of
strands within the β-sheets and on the packing and
relative orientations of the two neighboring β-sheets.
The segments of several full-length proteins were
identified using a combination of bioinformatics and
experimental procedures. Two fibril-forming segments of
Alzheimer's amyloid β-protein (Aβ) were identified:
(a) GGVVIA, i.e., Aβ(37-42), and
(b) MVGGVV, i.e., Aβ(35-40), both six amino acids long. According to the
fibril model of Petkova et al. , the amino acids of these
two segments are likely involved in close intermolecular
packing. Sawaya et al. showed that these two segments
belong to different classes, with GGVVIA forming parallel
and MVGGVV forming antiparallel sheets. Consequently, it
was concluded that the fibril structure of a full-length
protein may contain more than a single type of β-sheet. In addition, polymorphic fibrils of Aβ .
and other proteins could according to the steric zipper
hypothesis differ by the amino acids involved in the steric
It is puzzling to me that the authors did not select the
central hydrophobic cluster KLVFFAE, i.e., Aβ(16-22), in
their study. This fibril-forming segment is widely
perceived to be a key Aβ segment involved in Aβ
fibril formation. In support to that perception, there are
several naturally occurring mutations (F19P, A21G, E22Q,
E22K, E22G) which affect this segment and lead to altered
aggregation and fibril formation . It would be interesting
to determine if and how these mutations change the atomic
structure of microcrystals.
In the model of Petkova et al. , each Aβ molecule has a
turn or loop centered at G25-S26 with two intramolecular
β-strands interacting through the side chains. Petkova
et al. hypothesized that a salt bridge D23-K28
considerably contributes to the fibril stability. Indeed,
Sciarretta et al.  have shown that for Aβ(1-40) with
a lactam bridge D23-K28, fibrillogenesis occurs without the
typical lag period and at a rate 1,000-fold greater than in
Aβ(1-40) wild-type. If the dry steric zipper hypothesis
of Sawaya et al. is correct and formation of the zipper is
a rate-limiting process in fibril formation, then maybe
there are in addition to intermolecular steric zippers also
intramolecular zippers that contribute to formation and
stability of the Aβ fibril structure. Existence of the
lactam bridge or salt bridge D23-K28 in Aβ(1-40) could
then be understood as a trigger of the formation of the
intramolecular steric zipper leading to a highly increased
propensity for fibril formation. Hopefully, future work
will shed more light on these structural characteristics.
References: Petkova AT, Ishii Y, Balbach JJ, Antzutkin ON, Leapman RD, Delaglio F, Tycko R. A structural model for Alzheimer's beta -amyloid fibrils based on experimental constraints from solid state NMR.
Proc Natl Acad Sci U S A. 2002 Dec 24;99(26):16742-7. Epub 2002 Dec 12.
 Petkova AT, Leapman RD, Guo Z, Yau WM, Mattson MP, Tycko R. Self-propagating, molecular-level polymorphism in Alzheimer's beta-amyloid fibrils.
Science. 2005 Jan 14;307(5707):262-5.
 DB Teplow, A Lomakin, GB Benedek, DA Kirschner, and
DM Walsh. Effects of beta-protein mutations on amyloid
fibril nucleation and elongation. Alzheimer's Disease:
Biology, Diagnosis and Therapeutics, K Iqbal, B Winblad,
T Nishimura, M Takeda, and HM Wisniewski, Editors, John
Wiley & Sons Ltd., Chichester, England, 311-319 (1997).
 Sciarretta KL, Gordon DJ, Petkova AT, Tycko R, Meredith SC. Abeta40-Lactam(D23/K28) models a conformation highly favorable for nucleation of amyloid.
Biochemistry. 2005 Apr 26;44(16):6003-14.