Amyloid plaques are one of the hallmarks of Alzheimer's disease and were first described almost one hundred years ago by Alois Alzheimer. Today, however, their formation is still something of a mystery. It has been suggested that amyloid β, the protein precursor to amyloid fibrils, undergoes partial denaturation forming a peptide that is "stickier" than the native molecule. These peptides can then bond together side-by-side into a long stable fibril. A similar model has been proposed for formation of fibrils comprising other amyloidogenic proteins including β2-microglobulin and transthyretin.

Though in their native states these amyloidogenic proteins have no common structure, they all form amyloid fibrils composed of extensive cross β fibers, suggesting that a common glue holds all these different fibrils together. Understanding the composition of this glue could help scientists devise a strategy to dissolve it, though to date, which parts of the native proteins contribute to the cross β structure remains uncertain.

In Monday's Nature Structural Biology online, papers from two independent labs describe efforts to determine which amino acids of fibrillary β2-microglobulin are the most structurally inert and therefore most likely to contribute to the glue holding these fibrils together.

Yuji Goto at the Institute for Protein Research, Osaka University, and colleagues allowed β2-microglobulin fibrils to equilibrate with a deuterated solution thus replacing protons with the heavier deuterons. Following equilibrium they measured the remaining protons, and thus the most inert amino acids, by NMR. Sheena Radford and colleagues at the University of Leeds also used NMR but in this case to probe the denaturing effects of urea on the fibrillary intermediate.

Both labs arrived at a similar conclusion. They found the N- and C-terminal ends of the molecule are relatively unstable but five of the seven β strands, which comprise the core of the molecule, are protected from the denaturing effects of urea and from proton-deuteron exchange. The results suggest that the hydrogen bond network in these fibrils is extensive, thus explaining the relative stability and resistance to proteolysis exhibited by amyloid, and why polar solvents, such as dimethylsulfoxide, can dissolve amyloid in vitro. It remains to be seen, however, whether dissolution of amyloid in vivo will be possible.—Tom Fagan


  1. Among the most difficult problems for those interested in understanding the structural biology of amyloidoses are how to determine the complete three-dimensional structure of amyloid fibrils and to monitor the protein folding events leading to formation of these end-stage assemblies. X-ray diffraction approaches have yielded only modest amounts of information because of the inability to crystallize amyloid fibrils. However, great strides recently have been made in the application of NMR techniques to this problem. In two papers published online in Nature Structural Biology on 22 April 2002, the Goto and Radford groups report exciting new results that promise to significantly extend our ability to elucidate the structural dynamics of amyloid formation.

    Both labs studied the amyloid formed by β2-microglobulin (β2m), a 99-residue protein associated with hemodialysis-related amyloidosis. The Goto group used H/D exchange to map protected residues in the amyloid fibril. The Radford group used urea denaturation to determine individual residue stabilities, establishing four distinct behavioral groups based on denaturation profiles. The results from both studies were remarkably consistent. A large portion of the β-sheet comprising the native β2m monomer was involved in forming the core of the amyloid fibril. Less stable regions in the native fold, including the N- and C-termini and two edge β-strands, had lower protection factors and lower stability in urea. These areas may thus be important in initiating the amyloidogenic transformation of the native protein. The experimental approaches used in these studies should be applicable to other amyloidogenic proteins, allowing new insights about the structural dynamics of their assembly.

    What about Aβ? Kheterpal et al., 2000 have shown that H/D exchange measurements with mass spectroscopy can provide information about the organization of the Aβ fibril core. Therefore, much could be learned about Aβ fibril structure using the new techniques. Whether an equivalent amount can be learned about individual stages of Aβ folding and assembly is unclear. Aβ is a peptide, not a protein. It has no disulfide bonds nor is it known to have a stable, native structure in vivo, as do β2m, transthyretin, and many other amyloidogenic proteins. The conformational transitions occurring during Aβ amyloid assembly thus are unlikely to involve the same degree of reorganization of existing, stable, secondary structure elements as seen in proteins. Nevertheless, it will be interesting and important to do the experiments to see just how much we can learn. Based on the Goto and Radford reports, there is every reason to be optimistic.


    . Abeta amyloid fibrils possess a core structure highly resistant to hydrogen exchange. Proc Natl Acad Sci U S A. 2000 Dec 5;97(25):13597-601. PubMed.

Make a Comment

To make a comment you must login or register.


No Available References

Further Reading

No Available Further Reading

Primary Papers

  1. . Mapping the core of the beta(2)-microglobulin amyloid fibril by H/D exchange. Nat Struct Biol. 2002 May;9(5):332-6. PubMed.
  2. . Structural properties of an amyloid precursor of beta(2)-microglobulin. Nat Struct Biol. 2002 May;9(5):326-31. PubMed.