Oligomers of amyloid-β (Aβ) have created quite a stir since they were discovered to be perhaps the most toxic forms of the peptide. But oligomers are notoriously difficult to study. In fact, without much stirring, they spontaneously morph into larger and larger molecules that may be less germane to Alzheimer disease than their smaller counterparts. How is the researcher to keep tabs on what’s in the test tube? A little protein engineering may help. In the August 13 PNAS online, researchers led by Torleif Hard at the Swedish University of Agricultural Sciences, Uppsala, report that covalently constraining Aβ monomers into a specific β-hairpin structure accelerates formation of toxic oligomers. “These oligomers are completely stable and never convert to amyloid,” Hard told ARF. With about 50 times the toxic potency of natural Aβ monomers or fibrils, the stable, engineered peptides, Hard believes, could prove useful for studying oligomer dynamics and toxicity, and may even provide better epitopes for vaccination or raising antibodies for therapeutics.

Previously, Hard and colleagues found that certain proteins with antibody-like properties, called affibodies, stabilized Aβ peptides with a specific β-hairpin structure between amino acids 17 and 36 (see ARF related news story). The β structure resembled, but was not quite the same as, the β-sheet structures seen in amyloid fibrils. In the latter, the bonds between the β-sheets are intermolecular, whereas in affibody-stabilized peptides, they are intramolecular. Nevertheless, that work led to a proposed model of Aβ fibrillization where peptides with intramolecular β-sheets stacked upon one another before undergoing a conformation flip that resulted in intermolecular bonding. “We postulated that that β-sheet hairpin is a preferred conformation, and may be present in toxic oligomers,” said Hard.

To test that idea, first author Anders Sandberg and colleagues introduced two cysteine residues into the Aβ peptide, replacing alanine 21 and alanine 30, which are juxtaposed in the β-sheet hairpin. The theory was that a disulphide bond between the cysteines would stabilize that conformation. The researchers expressed the engineered peptide, dubbed Aβcc, in Escherichia coli. Aβcc readily formed oligomers with a range of conformations as judged by size exclusion chromatography, and formed protofibrils visible in the transmission electron microscope, but it failed to form fibrils unless it was first treated with a reducing agent to break the disulphide bond. The findings are in keeping with the need for a conformational change and rearrangement of intramolecular bonding for Aβ to form amyloid fibrils.

The covalently stabilized Aβcc may form oligomers, but how do those oligomers relate to the toxic ones that are released by cells (see ARF related news story), synthesized (see Lambert et al., 1998), and isolated from human tissue (see ARF related news story)? Sandberg found that various Aβcc species were highly toxic to SH-SY5Y neuroblastoma cells. In this assay, the potency of mid-molecular-weight oligomers and protofibrils of Aβcc was on par with that of wild-type Aβ42 oligomers and about 50 times higher than that of wild-type monomers or Aβ42 fibrils. “We think that the Aβcc oligomers, or the protofibrils formed by Aβcc oligomers, are very close to the naturally toxic form of Aβ,” said Hard.

“This was an ambitious project, the results of which support prior suggestions that oligomeric Aβ assemblies are particularly neurotoxic relative to monomeric and fibrillar forms of the peptide, and, therefore, that oligomers are attractive therapeutic targets,” suggested David Teplow, University of California, Los Angeles (see full comment below). Hard told ARF that some of those therapeutics could include monoclonal antibodies or vaccines. MIVAC Development AB, a biotechnology company in Gothenburg, Sweden, has begun looking into these possibilities. Hard is a consultant for, and holds stock in, this company.—Tom Fagan


  1. Less Is More...When You’re Talking About Aβ Toxicity
    In 1906, Alois Alzheimer reported an interesting case of dementia at a meeting of the South-West German Society of Alienists (1). Four years later, Emil Kraepelin termed the disease “Alzheimer’s disease” (2). Since that time, a century of effort had been expended to elucidate the pathobiology of AD. For the last 100 years, only the histopathology was clear, namely plaques and tangles. It has only been in the last 25 years that the gene encoding the protein component of plaques, the amyloid-β protein (Aβ), was cloned and studies of the contribution of Aβ to AD have been undertaken. During this time, two successive “cascade” models have been embraced. The first, proposed by Hardy and Higgins in 1992 (3), comprised a cascade of pathogenetic events that began with amyloid fibril formation. The second, and current paradigm, I have referred to as the “oligomer cascade hypothesis” (4). This hypothesis suggests that Aβ oligomers, rather than fibrils, initiate the pathogenetic cascade leading to the neuronal injury and death characterizing AD.

    The oligomer cascade hypothesis has its roots in studies seeking to establish the pathway of formation of Aβ fibrils. The first of these studies identified the immediate precursor of fibrils, protofibrils (5,6). Subsequent work has revealed successively earlier steps in the assembly of Aβ (for a review, see Roychaudhuri et al. [7]). Linking Aβ assembly to Aβ-induced pathogenesis is critical for validating the oligomer cascade hypothesis, and in so doing, for determining appropriate targets for therapeutic agents. However, structure- activity studies of Aβ are difficult because nascent Aβ exists primarily as a statistical coil in rapid equilibrium with low-order oligomers (8,9). Incubation of Aβ produces higher-order oligomers, protofibrils, and fibrils, all of which also are in equilibrium.

    To overcome the problems of metastability and polydispersity, Ono et al. used a zero-length cross-linking method, coupled with SDS-PAGE and solvent extraction, to produce pure populations of stable monomers, dimers, trimers, and tetramers of human Aβ40 (4). This cross-linking method involves the native Tyr10 residue of Aβ. Biophysical and biological studies of these assemblies revealed a direct, but non-linear, relationship between order and β-sheet structure, and between order and neurotoxicity. Tetramers and trimers were three- to 10-fold more toxic than were dimers or monomers.

    Now, Sandberg et al. have reported results of studies in which the Aβ monomer conformational space is constrained by an intramolecular disulfide bond (10). These authors have used genetic engineering to produce Aβ40 and Aβ42 that contain an N-terminal Met residue as well as Cys at positions 21 and 30 (which are Ala in the native protein). These latter substitutions provide the means to form intramolecular disulfide bonds that stabilize a hairpin-like Aβ conformer that has been proposed to be the Aβ monomer folding nucleus (11). The disulfide-stabilized analogues of Aβ40 and Aβ42 were designated Aβ40CC and Aβ42CC, respectively.

    Peptide assembly experiments show that the disulfide form of Aβ40 (Aβ40CC) forms few, if any, ThT-positive assemblies during a 60-minute incubation. Instead, large numbers of oligomeric structures are observed. In fact, AβCC self-assembly is quite rapid. Injecting guanidine HCl-treated Aβ40CC or Aβ42CC (putative monomer populations) onto a size exclusion column produced a continuum of peaks corresponding in Stoke’s radii to assemblies of ~10, 15, 30, 50, and 100 kDa in molecular mass, along with higher-molecular-weight species. The distribution of Aβ42CC was shifted towards higher mass relative to that of Aβ40CC. Various fractions obtained by SEC were studied then using CD, EM, ELISA, immunoblotting, and caspase activation assays. The “take-home messages” derived from these experiments, according to the authors, were that the Aβ40CC and Aβ42CC peptides maintain wild-type peptide conformation, do not form fibrils, are 50-fold more toxic than are peptide monomers or fibrils, and may assemble through at least two different pathways (defined by differential binding of different assemblies to antibodies mAb158 and A11).

    This was an ambitious project, the results of which support prior suggestions that oligomeric Aβ assemblies are particularly neurotoxic relative to monomeric and fibrillar forms of the peptide, and, therefore, that oligomers are attractive therapeutic targets.

    The results of Sandberg et al. also illustrate the aforementioned difficulties in establishing formal structure-activity relationships in the Aβ system. For example:

    1. Stabilization of the turn element between Ala21 and Ala30 does not stabilize particular assemblies significantly. The Stoke’s radius frequency distribution is complex.

    2. Statistical coil secondary structure decreases as assembly order increases. However, careful examination of the data suggests that the assembly process is not simply a coil to β-sheet process. Rather, the transition is from coil to mixed α-helix-β-sheet conformers.

    3. “β-sheet oligomers,” although more stable than their monomer precursors, do remain in equilibrium with lower-order assemblies (10 and 30 kDa).

    4. “Disordered monomers” can produce fibrils, either by an “A11” pathway or through a “mAb158” pathway. Are these two pathways really independent, or do the avidities of the antibodies simply change as the target antigen transforms from one conformational state into another? That is, are the antibodies recognizing discrete assemblies or discrete epitopes on the same assemblies, epitopes that may or may not be masked as assembly proceeds?

    For this reader, the “take-home” messages of the work are: 1) oligomers are substantially more toxic than are assemblies of lower or higher order; 2) methods to produce pure, structurally defined populations of biologically relevant oligomers are obligatory if formal structure-activity relationships are to be established; and 3) quantum leaps in understanding Aβ assembly will not occur unless the complexity of the system is controlled adequately.

    See also:

    1. Alzheimer, A. (1906) Neurologisches Centralblatt. 23, 1129-1136.

    2. Kraepelin, E. (1910) Psychiatrie: Ein Lehrbuch fu ̈r Studierende und A ̈rzte. (J. Barth, Leipzig) Vol. 2 (8th Edition).


    . Alzheimer's disease: the amyloid cascade hypothesis. Science. 1992 Apr 10;256(5054):184-5. PubMed.

    . Structure-neurotoxicity relationships of amyloid beta-protein oligomers. Proc Natl Acad Sci U S A. 2009 Sep 1;106(35):14745-50. PubMed.

    . Observation of metastable Abeta amyloid protofibrils by atomic force microscopy. Chem Biol. 1997 Feb;4(2):119-25. PubMed.

    . Amyloid beta-protein fibrillogenesis. Detection of a protofibrillar intermediate. J Biol Chem. 1997 Aug 29;272(35):22364-72. PubMed.

    . Amyloid beta-protein assembly and Alzheimer disease. J Biol Chem. 2009 Feb 20;284(8):4749-53. PubMed.

    . 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.

    . Preparation of amyloid beta-protein for structural and functional studies. Methods Enzymol. 2006;413:20-33. PubMed.

    . Stabilization of neurotoxic Alzheimer amyloid-beta oligomers by protein engineering. Proc Natl Acad Sci U S A. 2010 Aug 31;107(35):15595-600. PubMed.

    . On the nucleation of amyloid beta-protein monomer folding. Protein Sci. 2005 Jun;14(6):1581-96. PubMed.

    View all comments by David Teplow

Make a Comment

To make a comment you must login or register.


News Citations

  1. Affibodies—Putting the β in Aβ?
  2. Earliest Amyloid Aggregates Fingered As Culprits, Disrupt Synapse Function in Rats
  3. Paper Alert: Patient Aβ Dimers Impair Plasticity, Memory

Paper Citations

  1. . Diffusible, nonfibrillar ligands derived from Abeta1-42 are potent central nervous system neurotoxins. Proc Natl Acad Sci U S A. 1998 May 26;95(11):6448-53. PubMed.

External Citations

  1. MIVAC Development AB

Further Reading


  1. . Discriminating early stage A{beta}42 monomer structures using chirality-induced 2DIR spectroscopy in a simulation study. Proc Natl Acad Sci U S A. 2010 Sep 7;107(36):15687-92. PubMed.

Primary Papers

  1. . Stabilization of neurotoxic Alzheimer amyloid-beta oligomers by protein engineering. Proc Natl Acad Sci U S A. 2010 Aug 31;107(35):15595-600. PubMed.