Design and characterization of a membrane permeable N-methyl amino acid-containing peptide that inhibits Abeta1-40 fibrillogenesis.

Alzheimer's disease, Huntington's disease and prion diseases are part of a growing list of diseases associated with formation of beta-sheet containing fibrils. In a previous publication, we demonstrated that the self-association of the Alzheimer's beta-amyloid (Abeta) peptide is inhibited by peptides homologous to the central core domain of Abeta, but containing N-methyl amino acids at alternate positions. When these inhibitor peptides are arrayed in an extended, beta-strand conformation, the alternating position of N-methyl amino acids gives the peptide two distinct faces, one exhibiting a normal pattern of peptide backbone hydrogen bonds, but the other face having limited hydrogen-bonding capabilities due to the replacement of the amide protons by N-methyl groups. Here, we demonstrate, through two-dimensional NMR and circular dichroic spectroscopy, that a pentapeptide with two N-methyl amino acids, Abeta16-20m or Ac-K(Me)LV(Me)FF-NH2, does indeed have the intended structure of an extended beta-strand. This structure is remarkably stable to changes in solvent conditions and resists denaturation by heating, changes in pH (from 2.5 to 10.5), and addition of denaturants such as urea and guanindine-HCl. We also show that this peptide, despite its hydrophobic composition, is highly water soluble, to concentrations > 30 mm, in contrast to the nonmethylated congener, Abeta16-20 (Ac-KLVFF-NH2). The striking water solubility, in combination with the hydrophobic composition of the peptide, suggested that the peptide might be able to pass spontaneously through cell membranes and model phospholipid bilayers such as unilamellar vesicles. Thus, we also demonstrate that this peptide is indeed able to pass spontaneously through both synthetic phospholipid bilayer vesicles and cell membranes. Characterization of the biophysical properties of the Abeta16-20m peptide may facilitate the application of this strategy to other systems as diverse as the HIV protease and chemokines, in which there is dimerization through beta-strand domains.