Two papers in PNAS add to the growing body of work aimed at understanding the molecular interactions involved in the formation of amyloid fibrils (see related news story). Writing in last week's early online edition, researchers from the European Molecular Biology Laboratory, Heidelberg, Germany, together with colleagues from the University of Cambridge, England, describe a molecular modeling approach to understanding fibril structure. Led by principal author Louis Serrano, this group used mathematical algorithms to study the forces driving the formation of polymeric β-sheet structures by hypothetical hexapeptides.

First author Manuela Lopez de la Paz et al. then put theory into practice by empirically measuring the ability of the peptides to self-associate into β-sheet structures in solution. The authors found an excellent fit between model and test tube, with peptides predicted to form stable β-sheets exhibiting circular dichroism spectra typical for this secondary structure. However, formation of β-sheets is not a guarantee that fibrils will form. For example, when the authors used the electron microscope (EM) to analyze charged peptides, they found that fibrils only formed when the net charge of the peptide was +/-1.

Taking EM images together with x-ray diffraction data, the authors constructed a model in which the protofilaments that stack together to make the fibril are composed of four antiparallel β-sheets.

In contrast, researchers from the NIH, Bethesda, Maryland, and Lulea University of Technology, Sweden, led by Robert Tycko, report in this week's online edition that the β-amyloid of Alzheimer's disease comprises hairpin-shaped β-sheet structures that are stacked upon each other in parallel.

Aneta Petkova et al. used solid-state NMR measurements of the human Aβ1-40 sequence to arrive at this conclusion. The straight parts of the hairpin, residues 12-24 and 30-40, are predicted to be β-sheet structures, while the bend itself (residues 25-29) contains amino acids that bring different Aβ molecules together through side-chain interactions.

The NMR data predict that each hairpin unit has a core between the β-sheets that contains only mostly neutral and hydrophobic amino acids. Two charged residues in this core, aspartic acid and lysine, are predicted to form a salt bridge between each length of the hairpin.—Tom Fagan


  1. An increasing number of neurodegenerative diseases are associated with the formation of amyloid fibrils. Understanding the mechanism of fibril formation thus has obvious clinical value. In addition, knowledge gained in the study of amyloid fibril assembly would increase our understanding of basic aspects of protein folding. Unfortunately, the propensity of amyloid proteins to polymerize into noncrystalline structures has complicated the determination of their atomic structure. Now, in a seminal study by the Tycko group, solid state NMR has been used to provide the most extensive and detailed model of fibril structure yet obtained for the amyloid β-protein (Aβ). The model is attractive because of its consistency with the large body of existing data and its ability to solve the difficult conundrum, based both on NMR and IR data, of how both parallel and antiparallel β-sheets could exist in the same structure. The model predicts coulombic interaction between Asp-23 and Lys-28, a peptide segment long hypothesized to contain a β-turn. Interestingly, substitution of Asp-23 by Asn is associated with a familial form of early onset cerebral amyloid angiopathy in an Iowan kindred. The importance of the central hydrophobic cluster (Leu-17 to Ala-21) and of Met-35 is apparent through inspection, which shows intramolecular packing of the Leu-17, Phe-19, and Met-35 side-chains at the interface between the two β-strands. Similarly, Ile-41, now shown to be critical for the formation of peptide oligomers (see Bitan et al., 2002) would be predicted to interact with His-13 in the anti-parallel β-sheet of the Aβ monomer. Predicted intermolecular associations offer explanations of how mutations producing amino acid substitutions at Glu-22 (the Arctic (Gly), Dutch (Gln), and Italian (Lys) mutations) might affect peptide assembly—there is a disruption of salt-bridges formed between Glu-22 and Lys-16 on each of two antiparallel β-strands in the N-termini of adjoining Aβ monomers. The Tycko model provides a wonderful theoretical foundation for the design of experiments to test the roles of specific amino acids in controlling Aβ assembly. It also offers an "end-stage" structure for simulations of the folding of assembly intermediates.


    . Amyloid beta -protein (Abeta) assembly: Abeta 40 and Abeta 42 oligomerize through distinct pathways. Proc Natl Acad Sci U S A. 2003 Jan 7;100(1):330-5. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Heads or Tails—What Makes an Amyloid Fibril?

Further Reading

No Available Further Reading

Primary Papers

  1. . De novo designed peptide-based amyloid fibrils. Proc Natl Acad Sci U S A. 2002 Dec 10;99(25):16052-7. PubMed.
  2. . 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. PubMed.