|Home | About | Journals | Submit | Contact Us | Français|
We identify a distinctive circular dichroism (CD) signature for self-assembled 14-helical β-peptides. Our data show that self-assembly leads to a mimimum at 205 nm, which is distinct from the well-known minimum at 214 nm for a monomeric 14-helix. The onset of assembly is indicated by [θ]205/[θ]214 > 0.7. Our results will facilitate rapid screening for self-assembling β-peptides and raise the possibility that far-UV CD will be useful for detecting higher-order structure for other well-folded oligoamide backbones.
The development of unnatural oligomers that can adopt specific folding patterns (“foldamers”)1 represents a profound challenge in molecular design. Considerable success has been achieved at the level of secondary structures, especially for helices, and attention has begun to turn toward tertiary structures.2 A helix-bundle tertiary structure, in which two or more helices associate with their long axes approximately aligned, is common among globular proteins and constitutes an attractive goal for foldamer-based efforts. Folding in water can be driven by burial of stripes of hydrophobic side chains that are projected from one side of each helix.3 The design of α-peptide sequences that adopt globally amphiphilic α-helical conformations and assemble to form helix-bundle quaternary structure has proven to be a useful step toward de novo protein (i.e. tertiary structure) design,4 and several groups have sought to recapitulate this approach with foldamers.2 We reported the first example of β-peptide helix-bundle quaternary structure formation in aqueous solution.2a Subsequent studies by the groups of DeGrado2b and Schepartz2c–g have advanced this field. Each of these efforts has involved the β-peptide 14-helix, which is defined by backbone i,i-2 C=O–H–N H-bonds (14-atom ring). Progress toward helix-bundle formation has been achieved with other foldamer backbones as well.2g–m
Here we report that formation of a helix-bundle quaternary structure by 14-helical β-peptides is accompanied by a characteristic change in the circular dichroism (CD) signature relative to a monomeric 14-helix. Discovery and optimization of α-peptides or foldamers that adopt a specific tertiary or quaternary structure are facilitated if formation of the target structure can be detected spectroscopically. CD provides a convenient method for monitoring helix-bundle quaternary assembly among α-peptides, which are typically unfolded as monomers in aqueous solution. Self-assembly leads to a dramatic concomitant increase in α-helicity that can be detected by CD and provides an indirect indication of assembly.5 In our initial studies of β-peptide quaternary structure formation, we concluded that CD was not useful for detecting self-assembly of 14-helices because the β-peptides we employed were already 14-helical in the monomeric state.2a This high level of secondary structure formation was achieved by using sequences rich in trans-2-aminocyclohexanecarboxylic acid (ACHC) residues, which strongly promote 14-helicity.6
The isolated 14-helix gives rise to a strong minimum around 214 nm, as established in many studies.1,7 Here, we show that a specific 14-helix quaternary structure is signaled by a distinctive minimum around 205 nm. This characteristic CD signature is observed whether or not the self-assembling β-peptides contain ACHC. Our findings rationalize previously unexplained features in reported CD data of β-peptides that form helix bundles.2c–g
We examined three β-peptides (1–3), which are designed to self-assemble in aqueous solution. These ACHC-rich β-peptides should adopt the 14-helical conformation, which has approximately three residues per turn. The lipophilic-lipophilic-hydrophilic triads common to 1–3 lead to segregation of the lipophilic ACHC residues on one side of the 14-helix and the hydrophilic β3-hLys residues on the other side (Figure 1). The global amphiphilicity of these 14-helices favors the assembly of 14-helices with lipophilic ACHC residues at the core. Previous work showed that 10-mer 1 begins to form discrete assemblies, in equilibrium with the monomeric state, at concentrations above ~1 mM in aqueous solution; analytical ultracentrifugation (AU) indicated a preference for either hexameric or tetrameric assembly depending upon pH.2a We expected 13-mer 2 and 16-mer 3 to have a higher propensity to self-associate than does 10-mer 1 because α-peptide self-association is known to increase as the hydrophobic surface area increases.8 Indeed, AU of 1 mM solutions of either 2 or 3 in aqueous buffer indicated that these β-peptides are both highly self-associated, while 1 mM 1 is mostly monomeric under similar conditions. β-Peptides 2 and 3 form much larger assemblies than does 1: for 1 mM solutions of 2 or 3, we estimate the average assembly to contain ~20 or ~40 molecules, respectively.9 1H NMR data support the conclusion that lengthening the globally amphiphilic 14-helix increases the degree of self-association. For 10-mer 1, as previously reported, amide proton resonances are sharp below 1 mM (100 mM acetate buffer, pH 3.8), but the resonances broaden at or above 1.4 mM which is consistent with the onset of self-association.9 In contrast, for 0.1 mM 13-mer 2 in aqueous solution, amide proton resonances are broadened beyond the point of detectability.9
The far-UV CD spectra of 0.1 mM aqueous solutions of 2 or 310 differ considerably from the canonical 14-helical signature (Figure 2A), and these differences can be shown to arise from a high degree of self-association. Both samples display an intense minimum at 205 nm, along with a shoulder at 214 nm (isolated 14-helices give rise to a minimum at 214 nm). Comparison of CD and AU data suggests that the [θ]205/[θ]214 ratio is a useful indicator for self-association of 14-helices (Figure 2, Table 1). [θ]205/[θ]214 ≈ 0.7 for 10-mer 1 under conditions that do not support self-association according to AU (0.1 mM in aqueous Tris buffer, pH 7). A similar [θ]205/[θ]214 value is observed for iso-1, an isomeric 10-mer that has been shown via AU not to self-associate.2a Therefore, we conclude that [θ]205/[θ]214 ≈ 0.7 is the signature for a β-peptide that is 14-helical but does not form helix bundles. β-Peptide 13-mer 2 (0.1 mM) displays [θ]205/[θ]214 ≈ 0.7 in pure methanol, a solvent expected to disfavor self-association. However, 0.1 mM 2 shows [θ]205/[θ]214 ≈ 1.8 in aqueous Tris buffer, pH 7.3, conditions demonstrated by AU to promote extensive self-assembly. Under similar conditions, [θ]205/[θ]214 ≈ 2.0 for β-peptide 16-mer 3. We conclude that these large [θ]205/[θ]214 values represent a high degree of self-association. Previously reported AU data for 1.6 mM 10-mer 1 in Tris buffer, pH 8, indicated partial self-assembly,2a and this sample displays [θ]205/[θ]214 ≈ 0.95, a value between the extremes.
CD data indicate that the associated state of 13-mer 2 can be disassembled, without disruption of 14-helicity, by addition of methanol (Figure 2A). This cosolvent is known to promote helical secondary structure, relative to pure water, in both α- and β-peptides;6b,11 however, methanol disfavors tertiary and quarternary structure among proteins, because this cosolvent diminishes the drive for burial of hydrophobic surfaces.12 Figure 3 shows how [θ]205/[θ]214 changes when an increasing proportion of methanol is added to an aqueous solution of 0.1 mM 2. Small methanol proportions exert little effect, but self-assembly appears to be completely suppressed for >50 vol % methanol. The midpoint for the transition from self-assembled to nonassembled states occurs near 30 vol % methanol, which is similar to the methanol content necessary to denature small globular proteins such as α-chymotrypsinogen.12c NMR data support the conclusion that methanol prevents the self-association, because the amide resonances are sharp for 0.1 mM 2 in pure CD3OH, while, as noted above, these resonances are extensively broadened for 0.1 mM 2 in aqueous solution. Similar behavior was observed for 1 at higher concentrations.9
Our results show that a shift in CD minimum from 214 to 205 nm is correlated with a shift from monomeric to self-assembled states of β-peptides, but is there evidence to indicate that 14-helical folding persists upon self-assembly? Addressing this question is critical, because an alternative explanation for the CD shift is that self-assembly causes a change in β-peptide secondary structure. Indirect evidence for persistence of 14-helical folding in the self-assembled state may be found in recent work of Schepartz et al.,2d who observed a CD minimum at 205 nm for the self-associated state of a β-peptide containing exclusively β3-residues; this β-peptide crystallizes as a bundle of eight 14-helical molecules. To gain direct evidence against induction of alternative folding patterns upon self-association in solution, we conducted two-dimensional NMR studies of 10-mer 1 at 2.3 mM in aqueous buffer, conditions that allow partial self-association but that provide sufficiently resolved NMR resonances to enable conformational analysis. All backbone proton resonances of 1 could be assigned, and we could observe medium-range NOEs characteristic of the 14-helix (Figure 4, Table S2, Supporting Information). The strongest such NOEs occur between CβH(i) and CβH(i+3). Four of the seven possible NOEs of this type could be assigned for 1; the other three may have been present but were ambiguous because of resonance overlap. NH(i) → CβH(i+2) NOEs are characteristic of the 14-helix as well, although they are weaker than the CαH(i) →CβH(i+3) NOEs (Figure 4).13 Three of the seven possible NH(i) → CβH(i+2) NOEs were assigned, and two others were ambiguous. Additionally, one weak 14-helical NH(i) → CβH(i+3) was assigned. Most importantly, no NOEs inconsistent with 14-helical folding were detected, which suggests that self-association does not cause the population of a different secondary structure.
Our findings indicate that self-assembly of 14-helical β-peptides causes a diagnostic change in CD signature relative to monomeric 14-helices. This change seems to be characteristic of a specific mode of 14-helix assembly, because the formation of liquid crystalline phases by 14-helical β-peptides, which involves assembly into nanofi-bers,14 does not cause the characteristic shift from 214 to 205 nm. Schepartz et al. have identified β-peptides that do not contain ACHC but that form helical bundles in aqueous solution.2d,f,g These β-peptides display a CD minimum at 205 nm in the self-assembled state, which was attributed to 14-helix formation. However, our results show that the 205 nm minimum is not reporting on secondary structure, but instead is a distinctive indicator of helix bundle self-assembly. The structural changes that give rise to this CD shift appear to be relatively subtle, since 14-helicity is maintained. Lau and Hodges made related CD observations with α-helical coiled-coils.8 The ratio of the n–π * (~220 nm) and π–π * (~209 nm) transition minimum intensities is 0.8 for a monomeric α-helix, but after α-helical coiled-coil formation [θ]220/[θ]209 = 1.0. This α-peptide precedent supports our hypothesis that the electronic transitions underlying β-peptide far-UV CD are affected in a characteristic manner by 14-helix-bundle formation.
Our results will facilitate the screening of new β-peptide sequences for the propensity to form either a quaternary or tertiary structure because CD is a more rapid technique for initial analysis than is NMR or AU. In addition, these results raise the possibility that far-UV CD will be useful for detecting higher-order structure in other oligoamide backbones that are preorganized to display a high degree of secondary structure.
This research was supported by the UW-Madison NSEC (NSF DMR-042588) and NIH grant GM56414. T.L.R.G. was supported in part by a graduate fellowship from the National Science Foundation, and J.R.L. was supported in part by a PGS B scholarship from the Natural Sciences and Engineering Research Council (NSERC) of Canada. We thank PepTech for providing Fmoc β3-amino acids at discounted prices.