Search tips
Search criteria 


Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
Angew Chem Int Ed Engl. Author manuscript; available in PMC 2008 January 14.
Published in final edited form as:
PMCID: PMC2195894

Dynamic Covalent Chemistry on Self-Templating Peptides: Formation of a Disulfide-linked β-Hairpin Mimic**

The self-assembly of supramolecular structures is usually dependant on reversible noncovalent interactions.[1] Advantages of reversible self-assembly include mechanisms for error-correction through dynamic equilibration leading to the most stable structure and minimization of synthetic effort.[1b] However, thermodynamically stable assembled states often exhibit low kinetic stability. One approach to generate thermodynamic constructs that are kinetically stable is the concept of dynamic covalent chemistry where noncovalent interactions can be used to template covalent bond formation.[2] Dynamic covalent chemistry can generate an equilibrium mixture of interconverting molecules. The exchange mechanism may be stopped by changing the reaction conditions. The distribution of molecules shifts upon inclusion of a template to favor individual library members that bind to the template.[3] A specialized case of template-directed synthesis where the template is an integral part of the structure it helps to form has been termed “covalent capture”.[4] Reversible covalent capture using dynamic chemistry to form stable peptide-based assemblies has been demonstrated for the oligomerization of helical peptide bundles.[5] Herein, we describe the first studies on the dimerization of β-sheet-forming peptides by covalent capture under both reversible and irreversible conditions.

The objective was to explore whether self-templating of β-sheet-forming peptides occurs during the process of dimer formation by dynamic covalent chemistry. We designed two peptides ASH and BSH of length 4 mer and 10 mer respectively (Figure 1) containing (Leu–Lys)n repeats[6] known to predispose the peptide to form β-sheets.[7] The cationic nature of the (Leu–Lys)n motif also disfavors aggregation.[8] As the length of the peptide would dictate the degree of noncovalent interactions[9] two different values of n were chosen; n = 1 (ASH) and n = 4 (BSH). Of the several possible covalent coupling motifs, we have chosen to exploit the coupling of the thiol–disulfide system since it is water compatible, relatively fast, and can be switched on or off by changing the pH value.[10] Thiols were incorporated into the peptides as N-terminal Cys residues with a Gly spacer to offer some conformational flexibility for disulfide bond formation.[5]

Figure 1
Sequences of peptides ASH and BSH and the various dimers potentially formed by covalent capture.

A mixture of ASH and BSH can dimerize to form two possible homodimers (ASSA or BSSB) or a heterodimer (ASSB; Figure 2). We have investigated dimerization of equimolar solutions of ASH and BSH under irreversible and reversible conditions of covalent capture. Product distributions were analyzed from UV–HPLC traces of aliquots of the reaction mixtures (Figure 3) and quantified from a comparison of their areas under the traces with those of purified standards.[11] When a solution containing equimolar ASH and BSH in buffer was subjected to air oxidation[12] dimers ASSA:ASSB:BSSB were obtained in a 1.0:2.4:1.1 molar ratio. This essentially statistical distribution reflects that dimer formation is not influenced by noncovalent interactions. Even when the completely oxidized reaction mixture was left stirring for 72 h, the change in product distribution was negligible. The same results were obtained when we oxidized the mixture of thiols using K3[Fe(CN)6] in buffer.

Figure 2
Schematic representation of reversible covalent capture in simple linear peptides leading to preferential formation of covalently linked peptide partners.
Figure 3
HPLC traces of aliquots of reaction mixtures where equimolar amounts of ASH and BSH were a) oxidized by air, b) equilibrated in buffer containing GSH and GSSG (where G = glutathione). Note that these raw data peaks have not been normalized for the difference ...

To explore dimerization under reversible conditions we employed a redox buffer comprising oxidized (GSSG) and reduced (GSH) forms of glutathione.[5] As this peptide-based buffer was used in large excess over ASH and BSH, it also serves as a competitor to reduce the effect of nonspecific interactions between other peptide components. Equimolar solutions of ASH and BSH were equilibrated in GSH/GSSG redox buffer for 12 h. HPLC analysis of the products revealed only trace amounts of the statistically favored disulfide ASSB and a dramatic increase in the quantity of BSSB (Figure 3 b).[13] A new peak corresponding to ASSG where ASSG:ASSA was 13:1 was also observed. This feature is attributable to a concentration effect (initial [GSH]:[ASH] = 12.5:1) because of the relatively low stabilization resulting from self-recognition of ASH. All the oxidized BSH is present in the form of BSSB with no BSSG detected despite the 13-fold excess of GSH. This result is indicative of BSSB being a particularly stabilized structure. The molar ratio of ASSB:BSSB under reversible conditions was 1:86 which reflects a significant contribution of noncovalent stabilization energy resulting from self-recognition of BSH. The product distribution was unchanged when the reaction was left to equilibrate under argon for up to 2 weeks (data not shown), which suggests that this is a true thermodynamic distribution. To confirm that the system had indeed reached a thermodynamic equilibrium, a 1.0:2.4:1.1 mixture of ASSA:ASSB:BSSB generated by air oxidation in buffer was re-equilibrated in the presence of GSH/GSSG (500 μm/125 μm; Supporting Information). The product distribution was evaluated at various time intervals by HPLC. After equilibration for 36 h a product distribution identical to that shown in Figure 3 b was obtained, which confirms our model. The same results were obtained when the equilibration was effected by a neutral redox reagent employing oxidized and reduced forms of β-mercaptoethanol (Supporting Information). Thus, for this system, the introduction of reversible conditions allows an error-correction step in the covalent capture of a peptide partner.

To determine the origin of stabilization in BSSB, the structure of BSSB in water was studied using NMR spectroscopy. The influence of dimerization on structure was gauged by comparison with the monomer BSH, used as the reference compound. 1H NMR spectra were recorded in 10% (v/v) D2O and 90% (v/v) H2O and assigned using a combination of TOCSY and NOESY 2D data sets (See Supporting Information for details of experiments and spectra). A comparison of the 1D spectra of dimer BSSB (80 μm) and monomer BSH (80 μm) revealed a high degree of similarity. Moreover, TOCSY on BSSB clearly identified only eight Leu–Lys spin systems not 16 which indicates that the peptide backbones in BSSB have C2 symmetry (Supporting Information). Perturbations to the Hα and N–H chemical shifts are characteristic of secondary structure formation by peptides.[14,15] Substantial Hα upfield shifts (Δδ = 0.3 ppm) were observed at C10 and C11 residues at the center of BSSB which is indicative of a turn conformation at these residues. Keeping in mind the C2 symmetry of BSSB, this suggests that the disulfide linkage is at the center of a turn with both the peptide backbones propagating in the same direction. The 3J coupling constants of BSSB and BSH at 300 K are shown in Table 1 (see Figure 5 for numbering scheme). The 3J coupling constants for BSH are close to the ideal random coil value of 6.0 Hz. In contrast, for BSSB the values for the (Lys–Leu)4 residues are generally over 7.0 Hz some reaching even 9.0 Hz—characteristic of them being present in β-sheet conformation.[16] G9 and G12 showed ΔJ values of −0.7 Hz indicating that they form part of the turn in BSSB. More evidence for folded structure is provided by the observation of long-range NOEs between residues in the two strands of BSSB. A number of cross-strand NH–NH interactions between hydrogen-bonded residues (L2–K18, K16–L6, L6–K14, K14–L8, L8–G12) were also observed.[17] (Figure 4). Furthermore, the temperature coefficients (ΔδT) of N–H protons on Lys residues at various points on the sheet decreased progressively from the N-terminus to the C-terminus (Supporting Information). Similar results were obtained for the corresponding Leu residues. This result demonstrates that N–H protons nearer the turn are more strongly hydrogen bonded than those near the C-termini. This situation is completely consistent with BSSB being present in a hairpin-type conformation as shown in Figure 5. Though disulfide linkages are known to stabilize β-hairpins[13] and have been proposed to act as turn scaffolds between helices,[18] this is the first example of a disulfide as a turn scaffold in a synthetic peptide yielding a β-hairpin-type conformation in water.

Figure 4
Portion of the NOESY spectrum of BSSB at 300 K in 10% D2O/90% H2O illustrating cross-strand NH–NH NOEs that are consistent with the proposed folded structure.
Figure 5
NOEs between non-adjacent residues observed in NOE NMR spectroscopy analysis of BSSB at 300 K.
Table 1
Coupling constants (3J) of various residues on BSSB and the monomer BSH at 300 K.

Biological systems frequently employ reversible covalent capture to stabilize inter- and intramolecular assemblies.[19] This is the first example of dynamic covalent chemistry applied to peptides to produce a thermodynamically stabilized β-sheet assembly.


**YKG thanks the Royal Commission for the Exhibition of 1851 for a Research Fellowship. We thank Dr P. Grice for recording NMR spectra and Drs S. Otto and J. Christodoulou for helpful comments and suggestions.

Supporting information for this article is available on the WWW under or from the author.


1. a) Philp D, Stoddart JF. Angew. Chem. 1996;108:1242.Angew. Chem. Int. Ed. Engl. 1996;35:1154.b) Lindsey JS. New J. Chem. 1991;15:152.
2. For selected reviews see: Rowan SJ, Cantrill SJ, Cousins GRL, Sanders JKM, Stoddart JF. Angew. Chem. 2002;114:938. Angew. Chem. Int. Ed. 2002;41:898. [PubMed] Lehn J-M. Chem. Eur. J. 1999;5:2455.
3. Huc I, Lehn J-M. Proc. Natl. Acad. Sci. USA. 1997;94:2106. [PubMed]Nicolaou KC, Hughes R, Cho SY, Winssinger N, Labischinski H, Endermann R. Chem. Eur. J. 2001;7:3824. [PubMed]Goral V, Nelen MI, Eliseev AV, Lehn J-M. Proc. Natl. Acad. Sci. USA. 2001;98:1347. [PubMed]Karan C, Miller BL. J. Am. Chem. Soc. 2001;123:1074. [PubMed]Furlan RLE, Ng YF, Otto S, Sanders JKM. J. Am. Chem. Soc. 2001;123:8876. [PubMed]
4. Clark TD, Kobayashi K, Ghadiri MR. Chem. Eur. J. 1999;5:782.Royo M, Contreras MA, Giralt E, Albericio F, Pons M. J. Am. Chem. Soc. 1998;120:6639.
5. Saghatelian A, Yokobayashi Y, Soltani K, Ghadiri MR. Nature. 2001;409:797. [PubMed]Bilgiçer B, Xing X, Kumar K. J. Am. Chem. Soc. 2001;123:11815. [PubMed]Case MA, McLendon GL. J. Am. Chem. Soc. 2000;122:8089.
6. DeGrado WF, Lear JD. J. Am. Chem. Soc. 1985;107:7684.
7. Xiong H, Buckwalter BL, Shieh H-M, Hecht MH. Proc. Natl. Acad. Sci. USA. 1995;92:6349. [PubMed]
8. Haris PI, Chapman CH. Biopolymers. 1995;37:251–263. circular dichroism (CD) spectra of BSH and BSSB did not show evidence of extended β-sheet formation up to 300 μm, which indicates that there is no perturbation of the system by aggregation at the concentrations used for covalent capture experiments or NMR spectroscopy experiments. [PubMed]
9. Stanger HE, Syud FA, Espinosa JF, Giriat I, Muir T, Gellman S. Proc. Natl. Acad. Sci. USA. 2001;98:12015. [PubMed]
10. Szajewski RP, Whitesides GM. J. Am. Chem. Soc. 1980;102:2011.
11. Concentrations of standards were determined by amino acid analysis. Errors in such estimation in our hands were within 3%.
12. Air oxidation can yield thermodynamically stable products. See Otto S, Furlan RLE, Sanders JKM. J. Am. Chem. Soc. 2000;122:12063.
13. Peaks for ASH and BSH obtained in Figure 3 b corresponded to the equilibrium concentrations of ASH or BSH alone when equilibrated in a redox buffer of GSH/GSSG. See Cochran AG, Tong RT, Starovasnik MA, Park EJ, McDowell RS, Theaker JE, Skelton NJ. J. Am. Chem. Soc. 2001;123:625. [PubMed]
14. Wüthrich K. NMR of Proteins and Nucleic Acids. New York: Wiley; 1986.
15. Williamson MP, Asakura T, Nakamura E, Demura M. J. Biomol. NMR. 1992;2:83. [PubMed]Wishart DS, Sykes BD, Richards FM. J. Mol. Biol. 1991;222:311. [PubMed]
16. Fiebig KM, Schwalbe H, Buck M, Smith LJ, Dobson CM. J. Phys. Chem. 1996;100:2661.Maynard AJ, Sharman GJ, Searle MS. J. Am. Chem. Soc. 1998;120:1996.
17. The NOEs could arise between sequential residues. However, these NOEs were not observed in the case of BSH suggesting that BSSB is structured and the observed NOEs are a result of the presence of the second chain.
18. Karle I, Ranganathan D, Lakshmi C. Biopolymers. 2001;59:301. [PubMed]
19. Kogama T. Microbiol. Mol. Biol. Rev. 1998;61:212. [PubMed]Kodadek T. Trends Biochem. Sci. 1998;23:79. [PubMed]Freedman RB. Curr. Opin. Struct. Biol. 1995;5:85. [PubMed]Zhang T, Alber R. Nat. Struct. Biol. 1994;1:434. [PubMed]