Tetrameric b-Peptide Bundles

There is considerable current interest in the design of nonACHTUNGTRENNUNGproteinaceous quaternary structures with defined oligomeric states because these materials have potential as nanomaterial scaffolds, drug delivery tools and enzymatic platforms. We recently described a series of b-peptides, which are exemplified by the sequences of Zwit-1F and Acid-1Y (Figure 1A), that assemble into octameric b-peptide bundles of known structure and high stability. The 314 helices that comprise these octamers form three distinct faces: a leucine face whose side chains form the bundle core, a salt bridge face of alternating b-ornithine and b-aspartate resides, and an aromatic face that contains two b-tyrosine or b-phenylalanine residues. It is well established that natural coiled-coil–bundle stoichiometries are controlled by the identity and conformation of the side chains that are buried at the bundle interface. For example, substitution of valine or isoleucine for leucine at the GCN4 dimer interface leads to trimeric and tetrameric bundles. Here we show that the b-peptide bundle stoichiometry is also controlled by the side-chain identity within the bundle core. Specifically, by replacing the leucine residues at the octamer interfaces of Zwit-1F and Acid-1Y with valine generates valine derivatives, Zwit-VY and Acid-VY. These secondgeneration b-peptides assemble into discrete and stable tetrameric bundles that were characterized by analytical ultracentrifugation (AU), circular dichorism spectroscopy (CD), 1-anilino8-naphthalenesulfonate (ANS) binding, and deuterium exchange NMR spectroscopy. First we used analytical ultracentrifugation (AU) to determine whether Zwit-VY and Acid-VY formed bundles that possess a discrete stoichiometry in solution (Figure 1). The Zwit-VY AU data fit best to a monomer–n-mer equilibrium in which n= 4.07 with an RMSD of 0.00670 (Figure 1B). A comparable fit (RMSD=0.00672) resulted when n was set to equal 4 (Figure S2A in the Supporting Information). The lnKa value that was calculated from these two fits are 37.70 0.07 and 38.2 0.8, respectively. Significantly poorer fits (larger RMSD values and residuals with systematic errors) resulted when n was set to equal 5, 6 or 8 (Figure S2). The AU data that were collected for Acid-VY fit optimally to a tetramer where n=3.94 (n was allowed to vary) with an RMSD value of 0.00861 (Figure 1C). The Acid-VY data were fit to a variety of other oligomeric assemblies, all of which afforded poorer fits to the AU experimental data, and showed increased RMSD values and systematic errors in the residuals (Figure S2). We next used wavelength-dependent circular dichroism (CD) spectroscopy to characterize the concentration and temperature-dependent changes in the secondary structures of ZwitVY and Acid-VY (Figures 2 and S3). Zwit-VY underwent a ACHTUNGTRENNUNGconcentration-dependent increase in 314-helical structure (as judged by the molar residue ellipticity at 209 nm, MRE209 ) between 12 and 100 mm (Figure 2A), which is consistent with a concentration-dependent equilibrium between a partially structured monomer and a folded oligomer. This behavior mimics the behavior of previously characterized b-peptide bundles Zwit-1F, Zwit-1F*, and Acid-1Y. A plot of MRE209 as a Figure 1. A) Helical net representation of Zwit-1F, Acid-1Y, Zwit-VY and AcidVY. B, C) Zwit-VY and Acid-VY self-association monitored by analytical ultracentrifugation and fit to monomer-n-mer equilibria, where n was allowed to vary during fitting. Samples were prepared in 10 mm NaH2PO4, 200 mm NaCl (pH 7.1) and centrifuged to equilibrium at 25 8C at the indicated speed. The experimental data points are shown as open circles; lines indicate fits to the indicated monomer-n models.