1 Introduction

Regular secondary structures such as α-helices are essential components of protein architectures and are frequently found at protein–protein interfaces [1]. The coiled-coil motif which consists of two or more amphiphilic α-helical segments interacting together by interdigitation of their hydrophobic side chains is also common among protein tertiary structures and protein oligomerization domains [2]. The knowledge of these protein–protein interactions (PPIs) at atomic resolution has provided the basis for diverse applications of helically folded synthetic peptides including, for example, the modulation of therapeutically relevant PPIs [3,4] or the construction of bioinspired self-assembled (one-dimensional or three-dimensional) nanostructures [5,6]. However, in the absence of additional stabilizing effects (packing, favorable electrostatic interactions, and N- and C-capping motifs) [7,8], short peptide helices are only weakly populated in water thus limiting the development of their therapeutic applications. Today, a number of synthetic strategies are available to further increase the stability of short-chain α-helices, among which are the insertion of conformationally constrained amino acids (e.g., Cα-tetrasubstituted amino acids such as α-amino isobutyric acid, Aib [9,10], and β-amino acids [11]), side chain crosslinking [12,13], and helix nucleating templates [14–16]. The latter approach is aimed at pre-organizing the first amide bonds through the use of a rigid template generally placed at the N-terminus of the peptide chain (e.g., Kemp tricyclic templates [17,18]) or a hydrogen bond surrogate (HBS) in which the N-terminal intramolecular i,i+4 hydrogen bond is substituted with a covalent linkage (e.g., Arora HBS helices [19,20]). More recently, chiral templates have also been used to control helix screw sense preference in α-peptides made of helicogenic achiral Aib residues [21,22].

Surprisingly and despite potential interest, the use of a helical foldamer (synthetic folded oligomer) [23–25] as a template to nucleate/stabilize peptide helices has hardly been investigated; (αβ + α) peptides disclosed by the Gellman group represent to our knowledge the only case in the literature of a foldamer/α-peptide chimera mimicking an α-helix [26,27]. Most of the published studies are rather related to hybrid foldamer sequences whereby α-amino acid residues alternate with non-natural monomer units [28–30]. We [31] and others [32] have recently started to explore the concept of foldamer/α-peptide chimeras (or block co-foldamers as termed by Clayden [32]) with aliphatic oligoureas, a class of foldamers that adopt well-defined and stable helical structures akin to α-helices (similar helicity, polarity and pitch but a more complex three centred H-bond network and a larger diameter, Fig. 1a) [33–38]. In a previous study, we have shown that oligourea/peptide chimeras form well-defined helical structures in polar organic solvents with the propagation of a continuous intramolecular H-bond network spanning the entire sequence and connecting two geometrically distinct helices (α-helix and oligourea 2.5-helix) [31]. Furthermore, our results in solution also suggest that short triurea sequences have the ability to induce α-helicity when fused to either the N- or C-terminus of a peptide segment.

To gain additional insight into the α-helix stabilization properties of oligoureas, we have now investigated the folding properties of peptide-oligourea-peptide and oligourea-peptide-oligourea chimeras 1 and 2 containing 14 and 13 residues, respectively (Fig. 1b). Sequence 1 which contains a 6-mer oligourea motif flanked by two tetrapeptide units is aimed at testing the ability of a central oligourea chain to induce helix formation concurrently in the two peptide regions. Oligomer 2 was designed to evaluate whether two triurea units fused at the N- and C-termini of a peptide segment can work synergistically to promote α-helicity. In both sequences, the peptide segment is composed of alternating Ala and Leu residues, both known for their helix forming tendency [39,40].

2 Results and discussion

Both oligomers 1 and 2 have been synthesized by a solid-phase methodology starting from a Sieber amide resin [41] as shown in Scheme 1. Oligourea segments were assembled by coupling succinimidyl (2-azidoethyl)carbamate derivatives 3 (with side chains of Val, Ala and Leu) using microwave assistance followed by the reduction of azido groups under Staudinger conditions as previously described [42]. The peptide segments were built using standard Fmoc chemistry. Only the last Ala residue in 1 and the last urea unit in 2 were coupled as N-Boc protected monomers. Cleavage from the resin was performed by treatment with 1% TFA in CH₂Cl₂ thus allowing the N-terminal Boc protecting group to be preserved. The two oligomers were purified to homogeneity by C₁₈ RP-HPLC and their identity was confirmed by mass spectrometry and NMR in CD₃OH. The frequencies of all ¹H, ¹³C and ¹⁵N atoms were assigned by using homonuclear COSY, TOCSY and NOESY experiments and heteronuclear experiments (Tables 1 and 2). The assignment of spin systems and conformational investigation of the two chimeras were facilitated by the dispersion and by the distinct ranges of chemical shifts of amide and urea NH resonances, amide NHs appearing systematically downfield to urea NHs (in the range of 7.2–9.1 ppm in 1 and 7.2–8.9 ppm in 2).

Oligomer 1 consists of a 6 mer oligourea chain prolonged at both ends by a Ala-Leu-Ala-Leu tetrapeptide sequence. Inspection of the oligourea NH/CH fingerprint region in the ¹H NMR spectra reveals features typical of a helically folded oligourea chain, i.e., the large vicinal coupling constants between NH and CH(R) protons (³J > 9 Hz) and a high degree of anisochronicity (Δδ) of main chain methylene protons (0.81 < Δδ < 1.35 ppm) [43–45]. Furthermore, the presence of unambiguous non-sequential nuclear Overhauser enhancements (NOEs) along the oligourea backbone such as βN(i,i + 2) and βN′(i,i + 2) are fully consistent with helix formation in the oligourea segment [43]. The observation in the two peptide regions of an uninterrupted NN(i, i + 1) pattern of NOEs indicates that the peptide backbone also adopts a helical conformation (Fig. 2). This is supported by the presence of non-sequential medium range (i,i + 2), (i,i + 3) NOEs in the peptide regions and at the junction between the peptide and oligourea segments (e.g., βN(9,12), βN(9,13), and α2N(9,13) connectivities, Fig. 2b). However, because medium range NOEs between peptide and oligourea protons are non canonical, it is difficult to discriminate between a possible α-helical conformation and a 3₁₀-helical conformation for the short peptide segments. The values of the vicinal ³J (NH, ^αCH) coupling constants in the peptide segments are slightly above those reported for stable α- and 3₁₀ helical structures (<6 Hz) [46] thus suggesting helix fraying. Nevertheless, this NMR study tends to suggest that a helical structure develops without a break between the three segments, thus confirming the compatibility of both helix geometries and related H-bonding networks.

The ¹H NMR spectrum of triblock chimera 2 shares a number of features with that of chimera 1, including large >9.5 Hz ³J(NH, ^βCH) and anisochronicity of main chain methylene protons along the oligourea backbone, which altogether are indicative of a helical conformation [43]. Remarkably, five out of the seven vicinal ³J (NH, ^αCH) coupling constants in the peptide segment are below 6 Hz (3–5.8 Hz), supporting the view that the peptide region is also helical in solution. The non-sequential medium range βN(i,i + 2) and βN′(i,i + 2) in the two oligourea segments as well as NN(i, i + 1), αN(i,i + 3), αN(i,i + 2) and αN(i,i + 4) NOEs along the peptide chain observed in the NOESY spectrum of 2 are also consistent with the formation of a seamless helical conformation with no apparent interruption at the two peptide/oligourea junctions (Fig. 3).

Circular dichroism also proved useful to characterize the conformational preference of chimera 2. The spectrum recorded at 50 μM in MeOH (see Supporting information) was found to combine the hallmarks of both helically folded oligourea and α-peptide backbones. The spectrum is largely dominated by the contribution of the urea chromophore with a strong positive maximum at ca 203 nm characteristic of the oligourea 2.5-helical structure. A negative maximum at ca 222 nm is also observed that is not present in the spectra of the cognate homooligourea 5 and chimera 1 (see Supporting information). This signal can thus be tentatively ascribed to the peptide α-helical conformation and the corresponding molar ellipticity at 222 nm is likely to reflect the degree of α-helicity in chimera 2. However, due to the presence of two distinct chromophores and helix types in the main chain, the spectra of chimeras should be analyzed with caution and our interpretation at that stage remains largely tentative. Nevertheless, the ellipticity value at 222 nm for 2 in MeOH surpasses that measured for the parent heptapeptide 6 in both MeOH or TFE, supporting a stabilizing effect of the two oligourea domains in 2.

Additional insight into the conformation of triblock chimeras was obtained by high resolution crystallographic structural studies. Single crystals of chimera 2 suitable for X-ray diffraction were grown from a DMF solution and the structure was solved in the P2₁2₁2₁ space group (Table 3). The structure shows a continuous and regular helix that spans the entire sequence with no apparent distortion and with all intramolecular complementary H-bonding sites being satisfied. (Fig. 4). Closer inspection of the structure revealed useful details about the propagation of the H-bonding network and the way the oligourea and peptide helices communicate. The geometry of the helical peptide segment in the crystal structure of 2 is particularly notable. The Φ and Ψ dihedral angles of amino acid in the structure of 2 (mean Φ and Ψ values for the seven residues : −69°, −41°) match almost perfectly those of a canonical α-helix (−63°, −42°) [47]. The two terminal oligourea segments adopt a canonical 2.5-helical conformation [37,44]. A series of specific, non canonical bifurcated H-bonds are observed at the junction between the oligourea and peptide parts at both ends of the peptide sequence. In particular, urea carbonyls of Ala^u2 and Leu^u3 are both engaged in H-bonds with i+3 and i+4 amide protons (i.e., NHs of L5, A6 and L7). In a similar manner, amide carbonyls of Leu7 and Ala8 form a continuous network of four H-bonds with NHs of Leu^u11 and Ala^u12. In contrast, amide carbonyls of Leu9 and Ala10 form two centered i,i+3 H-bonds with N′H and NH of Ala^u12 and Leu^u13, closing 15- and 14-membered pseudocycles, respectively. Overall, the medium range NOEs between main chain NH and CH protons of the junction region (see for example αN(2,5), βN(2,5), αN(7,11), and αN(8,11) connectivities in Fig. 3b) are in good agreement with the short distances measured in the crystal structure, thus further supporting the conclusion that the molecule is largely helical in solution.

3 Conclusion

Herein, we have studied helix propagation and α-helix nucleation in peptide-oligourea-peptide and oligourea-peptide-oligourea triblock oligomers. Our results obtained in a polar organic solvent (methanol) and in the crystal state confirm the ability of oligourea segments as short as triureas to serve as nucleation templates when located either at the N- or C-terminus [31,32] or at both ends of a short peptide sequence. The crystal structure of the 13-mer chimera 2 reveals the formation of a regular helix that spans the entire sequence. In this structure, the peptide part in the middle of the sequence adopts a canonical α-helical conformation which is locked at both ends by two short 2.5 oligourea helices. This observation suggests a synchronized effect of both accessory foldamers in nucleating the α-helical structure and a possible strategy to stabilize α-helical peptides targeting protein surfaces. The oligourea backbone which is strongly biased towards helix formation provides a pre-organized cap for the initial four peptide NHs and final four carbonyl groups. It is noteworthy that most synthetic α-helical templates [8,14,15] reported in the literature are designed to engage H-bonds in the initial turn of the helix but relatively few molecules have been shown to effectively cap the carbonyl groups in the final turns of the helix [48]. Similar α-helix capping effects also take place in proteins [49]. To some extent, the complementary H-bonded network observed in 2 can be compared to that of native capping motifs found at the N- and C-termini of α-helices in proteins such as the N-terminal capping box [50,51], or the C-terminal Schellman motif [52,53]. Overall these findings and the high resolution structural data reported here may suggest a possible strategy based on oligourea foldamers to stabilize helical peptide epitopes for protein ligand design. Our progress in this direction will be reported in due course.

4 Experimental

Acknowledgements

This work was supported by the CNRS, Univ. Bordeaux, the Conseil Regional d’Aquitaine (Project #20091102003), ANR (Project #ANR-12-BS07-0019) and UREkA. CIFRE support from UREkA and ANRT to L. M. is gratefully acknowledged. We thank Jonathan Clayden for sharing results with us and for fruitful discussions.