. Role of amyloid-beta glycine 33 in oligomerization, toxicity, and neuronal plasticity. J Neurosci. 2009 Jun 10;29(23):7582-90. PubMed.


Please login to recommend the paper.


Make a Comment

To make a comment you must login or register.

Comments on this content

  1. Sex and the Single Amino Acid—A Devilish Problem
    In an article published in the 10 June issue of Journal of Neuroscience (1), Harmeier et al. report results of structure-activity studies of the 42-residue form of the amyloid-β protein, Aβ42. The work seeks to understand the role of single amino acids within Aβ on the folding dynamics, structure, and cellular activity of the peptide. The work reveals that glycine 33 (Gly33) may have a particularly significant role. (Actually, this role has nothing to do with sex, but the title apparently did induce you to read the commentary!)

    First, the stipulations of fact: 1) the Multhaup group has done, and continues to do, beautiful, interesting, and significant work; 2) the work of Harmeier et al. continues this tradition; 3) notwithstanding these facts, this commentator believes that it is fun, stimulating, and valuable for the field to play “Devil’s advocate” at times (this being one of them).

    The experimental work of Harmeier et al. is quite compelling. Substitution of the hydrophobic amino acids Ala or Ile for Gly33 produces a peptide that has a high propensity to self-associate, as determined by size exclusion chromatography (SEC). In contrast, substitution of Ala for Gly29, has little effect. Interestingly, both glycines participate in the formation of GxxxG motifs that have been found to mediate APP self-association and affect Aβ production (2). The data thus show how single amino acids in different contexts (APP or Aβ) may produce disparate structural effects. Interestingly, the corresponding doubly substituted peptide, [Gly29Ala/Gly33Ala]Aβ42, displayed aggregation properties (SEC) intermediate between those of the singly substituted peptides. One might predict, based on the aggregation characteristics of the singly substituted peptides, that [Gly29Ala/Gly33Ala]Aβ42 would assemble like [Gly33Ala]Aβ42. This was not the case, an observation again emphasizing the context-behavior relationship, namely the effect of a Gly33Ala substitution alone (i.e., in wild-type Aβ42) versus its behavior in the context of [Gly29Ala]Aβ42.

    Similar results were observed in experiments designed to probe the structural dynamics of Aβ42 and its “mutants.” In these experiments, limited trypsin proteolysis was used to evaluate the protease accessibility of the Lys peptide bonds (Lys16 and Lys28) in Aβ42. Conformers in which monomer folding, or monomer self-association, sequester these susceptible peptide bonds show increased protease resistance. In contrast, destabilizing substitutions increase proteolysis rates (e.g., see 3,4). Both Gly33 substitutions significantly decreased proteolysis, whereas the presence of Ala29, whether alone or with Ala33, increased protease susceptibility.

    Mechanistic explanations for the conformational dynamics data illustrate the complexity of the Aβ system and emphasize the difficulty in establishing simple models of peptide behavior. Harmeier et al. suggest that the data illustrate how the Gly33 substitution “further enhances the stability of the folding nucleus around this lysine residue with the turn region (Lazo et al., 2005; Grant et al., 2007)” (3,4). However, the data also could be interpreted to show that while substitution at Gly29 decreases turn stability, homologous substitution at Gly33 does not affect turn stability per se but rather facilitates peptide monomer self-association, which itself sequesters the Lys28 peptide bond, a bond that otherwise would be just as labile as it is within Aβ42.

    With respect to efforts to elucidate atomic-resolution characteristics of the system, one must suggest caution in the use of computational modeling of substituted Aβ42 peptides, especially when they involve framework structures from fibrils. There is no reason, a priori, to assume that identical structures exist within monomers, oligomers, and fibrils. In fact, much evidence exists that substantial conformational change occurs in the monomer→oligomer, oligomer→protofibril, and even the nascent fibril→mature fibril transitions. Simulations of monomer conformational dynamics and early oligomerization steps reveal a constellation of structures without a predominant conformer. This is a critical observation, one that must be kept in mind when considering Aβ assembly—the Aβ system should be considered a statistical distribution, not a linear assembly pathway. The authors’ modeling may be correct, but experimental confirmation of the ideas is necessary.

    The “activity” portion of the structure-activity analyses was very compelling. The authors show clearly that the Gly33 “mutants” have diminished, or no, toxic activity in assays on SH-SY5Y neuroblastoma cells, primary hippocampal neurons, LTP, and Drosophila eye development. These data were beautiful! Now comes interpretation.

    Is Gly33 “the key amino acid in the toxic activity of Aβ”? How does one know unless the substitution strategy executed by the authors is applied to each amino acid in Aβ? If Gly33 is the key amino acid, why is its activity, both with respect to conformational dynamics and toxicity, altered significantly by sister substitutions at Gly29? I bet one triple Americano at Starbucks (very meaningful to this commentator) that if these experiments are done, an interesting distribution of conformational and toxicity effects will be observed. Gly33 will not be the sole site at which structural changes produce large effects. Rather, this amino acid will be one of a number that have the potential to affect both assembly and toxicity.

    A distribution of toxic activity versus oligomer order also is expected. Harmeier et al. argue that among “different SEC fractions, tetramers of Aβ42 WT [and [Gly29Ala]Aβ42] exhibit the highest toxicity.” In part, these conclusions are based on Western blots of SEC fractions. This technique does not provide accurate assessments of oligomer state within a population of non-covalently associated assemblies (5). The SEC itself, especially for assemblies larger than dimer, cannot produce pure oligomer populations (e.g., see Fig. S1) and the toxicities of oligomers of different orders sometimes were similar, depending on the assay (MTT, MTS, or Live-Dead). The determination of “the most toxic” Aβ oligomer, akin to the “most infectious prion” oligomer (6), thus remains an open question.

    Moot is also the notion that “aggregation of Aβ42 peptides is uncoupled from toxicity.” My bet (the same as above) is that some oligomers will be toxic and others will not. Why? Because one must consider in great detail the question of structure-activity. A change of a single atom in Aβ, e.g., as in the Iowa Asp23Asn substitution that replaces an oxygen atom with a nitrogen atom, can cause disease (in this case, the Iowa form of cerebral amyloid angiopathy). However, other substitutions have little or no effects. Does one say, based on the latter observation, that primary structure is uncoupled from effect? I think not. Rather, one must consider each change, and the effects of that change, on an ad hoc basis.

    Finally, no evidence exists for Aβ “strains.” Strains are microbiological constructs that have to do with organismal characteristics that are perpetuated during organism replication. Although used appropriately in the prion case, no evidence exists for the existence of strains in AD.

    In conclusion, the work of Harmeier et al. should stimulate the field to dig deeper into what may be the Holy Grail of AD research, the determination of the frequency distributions of oligomer order and toxicity. Through this process, it is hoped that key therapeutic targets can be identified.


    . Role of amyloid-beta glycine 33 in oligomerization, toxicity, and neuronal plasticity. J Neurosci. 2009 Jun 10;29(23):7582-90. PubMed.

    . GxxxG motifs within the amyloid precursor protein transmembrane sequence are critical for the etiology of Abeta42. EMBO J. 2007 Mar 21;26(6):1702-12. PubMed.

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

    . Familial Alzheimer's disease mutations alter the stability of the amyloid beta-protein monomer folding nucleus. Proc Natl Acad Sci U S A. 2007 Oct 16;104(42):16522-7. PubMed.

    . Neurotoxic protein oligomers--what you see is not always what you get. Amyloid. 2005 Jun;12(2):88-95. PubMed.

    . The most infectious prion protein particles. Nature. 2005 Sep 8;437(7056):257-61. PubMed.

This paper appears in the following:


  1. The Toxic Fold? Aβ Dodecamers, Tetramers Show Their Conformations