REVIEW ARTICLE Well-defined secondary structures Information-storing molecular duplexes and helical foldamers based on unnatural peptide backbones Adam R. Sanford, Kazuhiro Yamato, Xiaowu Yang, Lihua Yuan, Yaohua Han and Bing Gong Department of Chemistry, University at Buffalo, State University of New York, Buffalo, NY, USA Molecules and assemblies of molecules with well-defined secondary structures have been designed and characterized by controlling noncovalent interactions. By specifying intermolecular interactions, a class of information-storing molecular duplexes have been successfully developed. These H-bonded molecular duplexes demonstrate programmable, sequence-specificity and predictable, tunable stabilities. Based on these highly specific molecular zippers (or glues), a systematic approach to designing self-assembled structures is now feasible. Duplex-directed formation of b-sheets, block copolymers and templated organic reactions have been realized. By specifying intramolecular noncovalent interactions, a backbone-rigidification strategy has been established, leading to unnatural molecular strands that adopt well-defined, crescent or helical conformations. The generality of this backbone-rigidification strategy has been demonstrated in three different classes of unnatural oligomers: oligoaramides, oligoureas and oligo(phenylene ethynylenes). Large nanosized cavities have been created based on the folding of these helical foldamers. Tuning the size of the nanocavities has been achieved without changing the underlying helical topology. These helical foldamers can serve as novel platforms for the systematic design of nano- structures. Keywords: backbone rigidification; duplex; foldamer; folding; helix; hydrogen bond; nanocavity; noncovalent; self-assembly; template. Introduction The assembly and folding of biomolecules are arguably two of the most important features in nature. There is no doubt that without the ability of nature to form very stable aggregates of small molecules, or to form well-defined secondary, tertiary and even quaternary structures of macromolecules, life could not exist as we know it. For example, the formation of duplex DNA represents one of the most elegant and best known examples of both self- assembly and folding of biomacromolecules [1]. The folding of polypeptide chains into secondary and eventually a bewildering array of tertiary structures results in protein molecules that are responsible for most of the biological interactions and functions found in nature. A logical next step is to mimic nature and create nonbiologically derived molecules that either fold into well-defined secondary structures, or assemble into larger architectures. Since the early 1990s there has been a great deal of literature devoted to these types of biomimetic structures that involve intermolecular self-assembly and/or intramolecular folding of unnatural molecules, and their potential applications [2–15]. The focus of a growing number of research groups including our own is to achieve structures whose properties may lead to multitudes of applications both in and out of the biological realm. The development of most of the unnatural assembling and folding structures has been inspired by peptide and protein structures. For example, the pioneering studies of Gellman and Seebach on b-peptides, a class of unnatural peptidomi- mimetic folding oligomers, have demonstrated the feasibility of folding unnatural oligomers into well-defined conforma- tions. Extending the concept of b-peptideshasledtoother peptidomimetic foldamers such c-peptides [16,17], d-pep- tides [18,19] and oxa-peptides [20]. The early investigation on molecular association of nucleobases by Jorgensen and Zimmerman has led to the development of highly specific H-bonded pairs that have been used in specifying inter- molecular interactions [21–23]. Although there has been a great deal of work in both molecular self-assembly and folding, the focus of this review primarily covers the work completed and ongoing in our laboratory. For a more com- prehensive view of these two fields there have been a number of excellent reviews published in recent years concerning molecular self-assembly and folding [2–15,25–29,39]. Hydrogen-bonded duplexes Nature has, of course, a monopoly on the most complex self-assemblies of molecules. In nature, the cooperative action of many noncovalent attractions often leads to highly specific recognition events, resulting in thermodynamically stable assemblies. The unfavorable loss in entropy is usually Correspondence to B. Gong, Department of Chemistry, 811 Natural Sciences Complex, University at Buffalo, State University of New York, Buffalo, NY 14260, USA. Fax: + 1 716 6456963, Tel.: + 1 716 6456800 ext 2243, E-mail: bgong@chem.buffalo.edu Abbreviations: A, acceptor; D, donor; m-PE, oligo(meta-phenylene ethynylene). (Received 17 December 2003, revised 6 February 2004, accepted 2 March 2004) Eur. J. Biochem. 271, 1416–1425 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04062.x offset by the large numbers of cooperative enthalpic interactions within the assembled structures, such as in duplex DNA. Much effort has been expended recently to produce self- assembling molecular assemblies between two or more sep- arate molecules. Inspired by nucleic acids, a powerful approach to self-assembly involves the design of recognition units (modules) that store and retrieve information in a digital fashion[2]. Modules withhigh specificity and strength, when tethered to various structural components, can serve as information carriers for instructing the formation of a variety of self-assembling structures. Compared to systems whose assembly depends directly on the structural features of the corresponding molecular components, such an approach is more versatile because structural units that are either incompatible or only randomly associate with one another can now be forced to assemble. The corresponding modules thus act as templates fororganizingand assembling structural components in specific sequential and spatial arrangements. A great majority of themodules have usedthe hydrogen bond as the primary stabilizing force due to the predictable directionality and strength of this noncovalent interaction. During the last decade, hydrogen-bonded complexes based on rigid heterocycles with multiple H-bonding donor (D) and acceptor (A) sites have received the most attention as recognition modules [2]. Systems based on various H-bonded modules have been reported. For example, Whitesides et al. described multicomponent structures based on the cyanuric acid–melamine motif [24–26]. The groups of Zimmerman [30–32] and Meijer [33–35] reported heterocyclic complexes with arrays of H-bond donors and acceptors. Krische [36] and Hunter [37] developed H-bonded duplexes. Ghadiri et al. [38,39] reported the assembly of modified cyclic peptides utilizing eight hydro- gen-bonds while Rebek et al. reported on the assembly of curved monomers that assembled into a form reminiscent of a tennis ball [40,41]. In spite of their increasingly wide applications, the use of heterocycle-based modules may suffer from several compli- cations. Most heterocyclic systems are accompanied by tautomerism [30], resulting in loss of information and reduction of the observed association constants. In addition, one must consider that in heterocycle-based modules, secondary electrostatic interactions, due to the proximity of adjacent H-bond donors and acceptors, often complicate the predictability of the strength of the designed complexes. We have developed our own approach for designing molecular recognition modules using molecular duplexes that are not only sequence-specific, but also with program- mable sequences and tunable stabilities [3]. Molecular duplexes: programmable, information-storing molecules We recently developed a class of highly stable molecular duplexes that are characterized by their by sequence- specificity and tunable stability: one single strand carrying an arrangement (sequence) of H-bond donors and acceptors specifically recognize another strand with a corresponding complementary H-bonding sequence (Fig. 1) [3]. These duplexes adopt an extended, tape-like conformation. The H-bonding sequences of these duplexes are readily programmable. Single strands with almost any donor– acceptor arrangement can be designed and prepared based on straightforward amide/peptide chemistry. These duplexes also showed regio-specificity and can thus act as molecular manipulators. Specifically, the molecular strands involve oligoamides consisting of meta-substituted benzene rings linked by glycine residues. The amide O and H atoms act as H-bond donor and acceptor sites. Various arrangements of the amide O and H atoms lead to Ôhydrogen bonding sequencesÕ that allow the specific association of two single strands carrying complementary seqences [3]. The structure of a six-H-bonded duplex is shown in Fig. 2 to illustrate the specific design: the duplex consists of two identical oligoamide strands carrying self-complementary H-bonding arrays of AADADD. Connecting the aromatic building blocks with glycine linkers in different order leads to oligoamide strands carrying different arrays (sequences) of H-bond donors and acceptors. As a result, duplexes with both self-complementary and complementary H-bonding sequences have been designed and characterized. Our study has demonstrated that these molecular duplexes are char- acterized by predictable, tunable affinity, programmable sequence-specificity and convenient synthetic availability, making them ideal as recognition modules for the instructed assembly of various structural units. Four H-bonded duplexes [3,42]. The first generation duplexes were designed around a four-H-bond platform with self-complementary H-bonding sequences of DDAA and DADA (Fig. 3). Of note are the six-membered inter- molecular hydrogen bonds that not only serve to block any undesirable intermolecular H-bond interactions, but also force a molecular conformation conducive to dimerization. Dimerization for both complexes was studied and confirmed through 1D, 2D and variable temperature Fig. 1. A schematic illustration of a H-bonded molecular duplex (molecular zipper). Fig. 2. A six-H-bonded, self-complementary duplex. Ó FEBS 2004 Well-defined secondary structures of unnatural oligomers (Eur. J. Biochem. 271) 1417 NMR experiments, and vapor pressure osmometry studies. Association constants between these homodimers were determined based on concentration-dependent chemical shifts of the amide protons. The dimerization constant for the DADA homodimer was found to be > 4.4 · 10 4 M )1 while the dimerization constant for the DDAA dimer was found to be  6.5 · 10 4 M )1 . The amino acid linker showed very little effect on the strength of the dimerization as a whole. For example, when the glycine used in initial tests was replaced with a bulkier phenylalanine, the dimerization constant was found to be  4 · 10 4 M )1 , within the ±10% error associated with the NMR method used in determining the constant. All three types of compounds, regardless of D–A sequence and amino acid spacer, exhibit very similar association constants and can be considered to have roughly the same stabilities. This is in contrast to H-bonded complexes based on rigid heterocycles, whose stabilities are influenced by secondary electrostatic interactions and are thus not only determined by the number of H-bonds, but also depend on the specific arrangement of the H-bond donors and acceptors. The sequence-independent stabilities of our duplexes are easily explained by the larger distance (>5 A ˚ ) between their adjacent intermolecular H-bonds than those (<2.3 A ˚ )of the H-bonded pairs of rigid heterocycles. Thus, by adjusting the number of intermolecular H-bonds, the stability of a duplex can be controlled accordingly. For example, compared to the four-H-bonded duplexes, the dimerization constant of a two-H-bonding molecule was found to be dramatically lower ( 25 M )1 ) [42]. Thus, if the number of H-bond D/A sites is increased, the overall stability of the corresponding duplex should also increase. This led to the investigation of even longer, more complex duplexes. The investigations also focussed on the creation of duplexes of consisting of two different but complementary strands. Six H-bonded Duplexes [3,43]. Thenextstepinthe course of investigation involved six-H-bonded molecular duplexes. Initially, compounds 1 and 2 were designed and synthesized as complementary pairs with only one allowable mode of dimerization (Fig. 4). Initial indications of dime- rization were apparent as the solubilities of separate solutions of 1 (< 1 m M )and2 ( 10 m M ) were low but that of the 1 : 1 mixture of the two in the same solvent was much higher ( >>100 m M ). This phenomenon presuma- bly was due to the shielding of the highly polar amide groups on the Ôsticky edgeÕ of each single strand. The formation of the six-H-bonded duplex was conclu- sively confirmed by 2D NMR data showing critical interstrand NOEs in chloroform. In addition, the duplex could even be detected by straight phase (SiO 2 ) TLC. Using 10% dimethylformamide in chloroform as the eluant, differences in R f values between 1 (R f ¼ 0.00), 2 (R f ¼ 0.10) and the 1 : 1 mixture (R f ¼ 0.96) clearly indicated the formation and high stability of the duplex. Further NMR studies showed significant downfield shifts of the amide protons of the 1 : 1 mixture of 1 and 2 in comparison to the separate solutions of the single strands. Attempts to determine the dimerization constant by NMR failed as no upfield shift of the amide protons was detected upon dilution in CHCl 3 to a concentration as low as 1 l M . Isothermal titration calorimetry was employed to deter- mine the association constant via titration of 1 with 2 in chloroform. The result can only be estimated to be ¼ 1 · 10 9 M )1 in chloroform, in agreement with NMR results. A more accurate value of K a was determined when the titration was carried out in chloroform containing 5% dimethyl sulfoxide, resulting in an association constant of 3.5 · 10 6 M )1 . In addition to the above complementary pairs of duplexes, a self-complementary, six-H-bonded duplex were also studied (see Fig. 2). Again, as with other six-H-bonded duplex, NMR experiments failed to determine an associ- ation constant due to the high strength of the association. By assuming a 10% dissociation, a lower limit of 4.5 · 10 7 M )1 wasestimatedfortheK a . Interstrand NOEs were observed under a wide range of concentration conditions, indicating little dissociation even in highly dilute solution. Simple observation on a TLC plate indica- ted even greater stability than that of a similar non-self- complementary six-H-bond duplex. Virtually no tailing, indicative of dissociation, was observed on the TLC medium for the self-complementary duplex while the non- self-complementary strand showed significant tailing. To determine a more accurate association constant, a pyrene- labeled derivative of this self-complementary system was studied by a fluorescence method. The dimerization con- stant was determined to be (6.8 ± 4.1) · 10 9 M )1 . Such a strong stability and the feature of self-complementarity gives this compound an interesting potential for the synthesis of supramolecular, high molecular mass homo- polymers [44]. The sequence-specificity of the six-H-bond duplex was probed with the inclusion of mismatched binding sites along the backbone of a duplex (Fig. 5) [45]. This study provided a unique look at the importance of sequence-specificity on these unnatural self-assembly systems. The mismatched pairs were studied with 1D and 2D NMR and isothermal titration calorimetry. The results were compared to the Fig. 3. Self-complmentary, four-H-bonded duplexes. Fig. 4. Example of six-H-bond, complementary duplex 1•2. 1418 A. R. Sanford et al. (Eur. J. Biochem. 271) Ó FEBS 2004 parent complementary duplex. The ÔmismatchedÕ strands still assemble as do the ÔmatchedÕ strands, but with much less stability. Isothermal titration calorimetry experiments revealed that the mismatched duplexes exhibited stabilities 40 times less than that of the corresponding matched pairs. Duplex foldamers [46] Progress with our four- and six-H-bonding duplexes led to the development of a new class of compounds with eight H-bonding sites. Single strands 4 and 5, one with two DDAD modules and the other with two AADA modules linked in a head-on fashion, were originally designed to form supramolecular polymers. Similarly, a self-comple- mentary strand consisting of two ADAD modules linked head-to-head was also designed. If these strands adopted an extended conformation, the directionality of the unsym- metrical four-H-bond module would enforce a partial overlap of the molecular strands, leading to the formation of supramolecular polymers. Instead, it was discovered that each eight-H-bonding strand associated with one separate complementary strand; however, like many biomacromole- cules, the molecular strands adopted a well-defined, folded conformation (Fig. 6). Noncovalent, templated formation of b-sheets [3,47] The duplexes described above can be viewed as mimics of two-stranded b-sheets consisting of extended b-strand mimetics. The interstrand distance in these duplexes should be nearly the same ( 5A ˚ ) as that found in b-sheets as both involve backbone amide groups in forming interstrand H-bonds. The duplexes, when tethered to natural peptide strands, may serve as templates to bring the peptide strands into close proximity. In addition, the programmable sequence-specificity of the duplexes allows the design of unsymmetrical H-bonding sequences, which ensures the precise registration of the amino acid residues of the attached natural peptide strands, leading to the formation of two-stranded b-sheets. Therefore, the organizational stability of the H-bonded duplexes provided an opportunity for nucleating and stabilizing b-sheets, a feature that could help provide critical insight into such structures. As opposed to studies on b-sheets based on b-hairpins, the nature of such an assembly is intermolecular, making it possible to pair peptides of various lengths and sequences by simply mixing the corresponding templated peptides (Fig. 7). The resultant mixing of complementary hybrid duplex strands would force the otherwise flexible peptide chains to form b-sheets. To evaluate the strategy of nucleation and stabilization of b-sheets, we designed four hybrid chains, each consisting of a tripeptide segment coupled with a region capable of forming a four-H-bonded duplex (Fig. 8). The H-bonding sequence of the duplex was designed to be unsymmetrical to direct the peptide chains to the same (or different) end of the template. Hybrids 6a and 6b were paired with hybrids 7a and 7b. As a control, hybrid 6c was designed. Pairing 6c to the corresponding hybrid strands led to the attachment of the tripeptide chains to the opposite end of the duplex template. H 1 NMR results of duplexes of 6a with 7a and 7b,aswell as 6b with 7a and 7b exhibited very sharp, well-defined resonances indicative of well-defined overall structures. Mixtures of 6c with its corresponding partner showed no such resolution, exhibiting broad, nearly indistinguishable resonances. Also observed was the noticeable chemical shift of the H 1 NMR resonances in only the complementary hybrid strands in comparison to single hybrid strands and the tripeptide strands alone. Two-dimensional NMR (NOESY) studies provided conclusive evidence of duplex and b-sheet formation. NOE contacts were observed between opposite amino acid residues. In the case of ‘plain’ tripeptide mixtures or duplexes with ‘wrongly’ attached tripeptides, no NOE contacts were observed. Fig. 5. An example of a duplex containing a mismatched ‘binding’ site. Fig. 6. Single strands 4 and 5. Each strand consisting of two identical four-H-bonding halves linked in a head-on fashion, were originally designed to (A) form supramolecular polymer, but were found to (B) adopt folded (stacked) conformations upon associating with each other. Fig. 7. Schematic representation of b-sheet nucleation and stabilization by a four H-bond duplex template. Ó FEBS 2004 Well-defined secondary structures of unnatural oligomers (Eur. J. Biochem. 271) 1419 Further investigations based on H-bonded duplexes Supramolecular block copolymers. The extraordinary spe- cificity and stability of our H-bonded duplex is again demonstrated recently in the design of supramolecular block copolymers. By attaching three polystyrene and three poly(ethylene glycol) chains to the two strands of a duplex derived from 1 and 2, a total of nine block coploymers were created by simply mixing the polymer-tethered templates (Fig. 9). NOESY study confirmed that the duplex template was precisely matched as expected from the H-bonding sequences. The successful noncovalent linking of the polystyrene and poly(ethylene glycol) chains was confirmed by size exclusion chromatography. More exciting were the results from the atomic force microscopy that revealed microphase separation typical of covalent block copolymers by these supramolecular block copolymers. Figure 9B shows one such atomic force microscopy image of a pair carrying polystyrene 21 000 and poly(ethylene glycol) 6000 chains. A duplex-templated organic reaction Olefin metathesis reactions have found applications in a wide variety of fields [48,49]. Intermolecular olefin meta- thesis involving two different olefins can be complicated by the fact that a mixture of three products can result when the two reacting olefins have similar stabilities. The use of our duplex strands can solve this problem. By tethering two separate olefins on two complementary strands, the olefins Fig. 8. Hybrid duplex strands. Fig. 9. Block copolymers. (A) Design of supramolecular block copolymers based on the six-H-bonded hetero-duplex 1•2. Mixing three templated polystyrene chains with three complementarily templated poly(ethylene glycol) chains leads to nine block copolymers. (B) The AFM image showing microphase separation of one of the supramolecular block copolymers. 1420 A. R. Sanford et al. (Eur. J. Biochem. 271) Ó FEBS 2004 were brought into close proximity by the pairing of the duplex strands. Metathesis reactions have been carried out between two olefins at concentrations (1 or 2 m M )thatare otherwise too low for intermolecular metathesis reactions to occur. As a result, only the desired, hetero-crosslinked products were formed in high (>90%) yields [40]. Intramolecular self-assembly: helical foldamers In recent years there has been intense interest in creating oligomers and/or polymers with unnatural backbones that display stable, well-defined conformations. Pioneering reports from Gellman [7,50] and Seebach [51,52] on helically folding b-peptides opened the floodgates for a rush of reports on other unnatural helical structures. From the c-peptides reported by Hanessian [53] and Seebach to the oligo(pyridine dicarboxamides) reported by Huc and Lehn [54] to the folding oligo(m-phenylene ethynylenes) reported by Moore [55,56] and the helical aromatic oligoureas by Tanatani [57], folding structures are obviously not monopolized by nature itself. We have reported novel helical foldamers based on the enforced folding of oligoarylamides and oligo(phenylene ethyny- lenes) that exhibit well defined helical secondary structures that have great potential in both materials science and biological application. Oligoamide foldamers Our approach to the design of helical foldamers involves the development of oligoarylamides with rigid, crescent backbones [58,59]. These oligoamides can also be viewed as aromatic c-peptides. The initial focus was placed on oligomers consisting of building blocks derived from 2,4-dihydroxy-5-nitrobenzoic acid (or meta-disubstituted building block). When derivatives of this molecule were connected via an amide bond, it was reasoned that the bifurcated (three-centered) H-bond, consisting of the two intramolecularly H-bonded five and six membered rings, would limit the rotational freedom of the aryl-amide bonds. The three-centered H-bonds rigidify the overall backbone and thus force a crescent shape, for example, on the tetramer shown in Fig. 10. The resulting three-centered hydrogen bond was found to be highly stable from both theoretical and experimental evidence [59]. The crescent and/or helical conformations should be reinforced by the interplay of multiple factors such as the rigidity of the benzene ring and amide groups, the propen- sity of the amide bonds to adopt the trans conformation, and eventually (as chain length increased to over one turn) favorable p–p stacking interactions between overlapping phenyl rings. This indeed was the case as indicated by 2D NMR and X-ray crystallographic data from oligomers consisting of two, three, four, five and six residues. The structure proved itself to be very stable in both the solid state and in solution. From this data it is clear that the orientation of the amide oxygens yielded an interior cavity that is electronegative and hydrophilic. Once the conforma- tion of the amide backbone was confirmed, the extension of the backbone beyond the length of a single turn, and thus a helix was attempted [60]. Indeed, combining two tetramers with a symmetrical residue derived from 4,6-dihydroxyisophthalic acid, resulted in nonamers of more than one full helical turn [60]. Side chains, however, were found to play a critical role in the solubility of the corresponding oligomers. It was found that oligomers carrying short side chains such as methyl or isopropyl groups had rather limited solubility. Long alkyl chains, such as the linear octyl or dodecyl groups, were adopted to impart solubility. However, it was discovered that when side chains longer than a methyl group were placed on adjacent residues of an oligomer, this led to a distortion/twisting of the helix, primarily from steric inter- actions between side chains. This problem was solved by designing oligomers containing methyl side chains on every other residue. Recently, a building block carrying one long alkyl and one methyl side chain became available, making it possible to construct oligomers using a single building block. The folded structures were rigorously characterized by both 2D and X-ray crystallography. The helices show excellent stability in organic solvents and our recent results show that folding also occurred in very polar (and hydrogen bond disrupting) solvents such as water and dimethyl sulfoxide. The persistence of the highly favorable three- centered H-bonds, which act to rigidify the backbone and lead to the overall folded conformations of the oligoamides, was further demonstrated by extremely slow amide proton– deuterium exchange rates. While in nature, large cavities of nanometer scale are usually found at the tertiary or quarternary structural level of proteins, we have been able to create and tune the nanocavities while maintaining the same helical topology of our foldamers [60]. One of the highly attractive features of oursystemistherelativeeaseinwhichtheinteriorcavity Fig. 10. A tetrameric crescent oligoamide (A) and its crystal structure (B). For clarity, the octyl groups are replaced with green dummy atoms in the crystal structure. Ó FEBS 2004 Well-defined secondary structures of unnatural oligomers (Eur. J. Biochem. 271) 1421 can be altered. It was initially envisioned that by merely incorporating a certain proportion of residues derived from 2,3-hydroxy-4-nitrobenzoic acid (or simply para-disubsti- tuted building block) into the oligomer, the curvature of the corresponding backbone would be decreased, leading to an increase in the size of the interior cavity (Fig. 11). For example, a 21-mer (Fig. 12) consisting of alternating meta- and para-building blocks was found to fold into a helical conformation of slightly more than one spiral turn. The existence of a helical conformation was supported by end- to-end NOEs and by NOEs between each amide proton and the protons on two of its neighboring side chains. A computer model of the 21-mer constructed based on parameters from the X-ray structures of short oligomers of the same system revealed a helix with an interior cavity of over 30 A ˚ across, the largest thus far created by unnatural foldamers [60]. Extension of this porous foldamer system beyond our original explorations is at the forefront of our current focus. As indicated by NMR and X-ray crystallographic data, the interior channel lined by amide oxygens has potential for application. Short oligomers may act as membrane-bound carriers for ions and small molecules; longer oligomers may fold into nanotubes with hydrophilic channels. Oligomers with lengths matching the thickness of the lipid bilayer could act as channels for transporting ions and small molecules (Fig. 13). The side chains, which point radially outward on the exterior of a helix, should help the integration of our porous foldamers into lipid bilayers. Probing the stability of our backbone-rigidified helical foldamers in polar solvents will be one of our future studies. Oligomers that fold stably in aqueous media are of great interest due to their biological significance and their potential for the development a variety of bio-related materials. The incorporation of triethylene glycol side chains and aliphatic chains terminated with carboxyl groups are the strategies we intend to use to render the corresponding oligomers reasonably soluble in aqueous media. If an oligoamide backbone could fold into well-defined and robust conformation based on backbone rigidification, could the same strategy of backbone rigidification be applied to different unnatural backbones? Initial work in our laboratory involving oligoureas have shown promise. Fig. 11. Adjusting cavity size by tuning the curvature of a backbone. Fig. 12. A 21mer (approximately half is shown) with an interior cavity of > 30 A ˚ . Fig. 13. Schematic representation of ion channel based on our porous helix. 1422 A. R. Sanford et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Folding oligoureas The backbones of this class of oligomers involve benzene rings linked by N,N¢-disubstituted urea groups (Fig. 14). The presence of ester groups ortho to the urea N atoms leads to the formation of a intramolecularly H-bonded, six- memberd ring, which, combined with the preferred (cis, cis) conformation of the urea group, leads to the rigidification of the oligourea backbone. The meta-disubstituted benzene rings, in combination with the nonlinear urea groups, enforce a curved conformation. Molecular modeling studies showed that, for oligomers with more than four residues, a helical conformation results. (Fig. 14A) The interior of the helix is characterized by oxygen atoms from the urea functionality, which lead to hydrophilic cavities of  4A ˚ across. Urea linkages have the advantage of being chem- ically robust and resistant to most natural enzymes. In contrast to the large cavity of the crescent oligoamides mentioned above, the much smaller sizes of the neutral, electrostatically negative cavities of the oligoureas are ideal for binding and/or transporting ions. Our ultimate goal is to develop these foldamers into novel ion carriers, and eventually, into ion channels. Oligoureas from the dimer to the hexamer have been prepared. Extensive studies based on computational, 1D and 2D NMR studies have con- firmed the design principle. Our latest preliminary studies on metal ion binding have identified a tetramer that selectively binds sodium ions. During the revision of this manuscript, a paper published by Meijer, et al. came to our attention that reports helical polyureas based on design principles very similar to those described here [61]. As shown in Fig. 14B, replacing one of the ester group on a residue with an ether group results in five-membered hydrogen-bonded rings. The backbone of the corresponding oligomer should be more flexible due to the weaker H-bond in the five-membered ring. The driving forces that facilitate the folding of these oligoureas are virtually identical to those for the folding of oligoamides mentioned above: (a) two localized intramole- cular hydrogen bonds that lead to the rigidification of each of the cis, cis-urea linkages and (b) the aromatic stacking interaction that further stabilized the helical conformation. Folding oligo(phenylene ethynylenes) [62] The strategy of backbone rigidification has also been also applied to oligo(meta-phenylene ethynylenes) (m-PEs). Solvent-driven folding of oligophenylacetylenes carrying Fig. 14. Folding oligourea (general structures). Fig. 15. A backbone-rigidified hexa (phenylene ethynylenes). Fig. 16. Tuning the cavity of backbone-rigidified m-PE foldamers. A larger cavity can be produced by incorporating one (left) or two (right) para substituted residues. Ó FEBS 2004 Well-defined secondary structures of unnatural oligomers (Eur. J. Biochem. 271) 1423 polar side chains has been well established by Moore and coworkers [56,57]. The Moore system relied on polar side chains to effect a hydrophobic collapse of the PE backbone, leading to helical structures that were denatured in nonpolar solvents. Based on the m-PE system, a completely different folding strategy was achieved by introducing an intra- molecular H-bond that restrict the rotational freedom of the backbone (Fig. 15). Folded m-PE oligomers, from dimers to heptamers, were observed in nonpolar solvents such as chloroform. The folded conformations were confirmed by X-ray and 2D-NMR studies. As anticipated, the 2D 1 HNMR (NOESY) spectrum of a hexamer showed end-to-end NOEs that could only be explained by a helical confirmation of this oligomer. Similar to the crescent oligoamides, the cavity size of the backbone-rigidified m-PEs can be tuned by changing the connectivity of some of the backbone units (Fig. 16). Placing the H-bonds and the corresponding side chains inward should lead to functional cavities. Conclusions While progress has been made in mimicking the basic features of biological systems, there is still a great deal of work to be done. For all intents and purposes, molecular self-assembly and molecular folding are still in their infancy. Nearly all unnatural systems, while elegant and fascinating in their own regard, still fall well short of the shear complexity and grandeur of constructs seen in nature. The recognition, interaction and folding of bio- molecules, particularly biomacromolecules, have inspired most of the currently known unnatural systems. By mimicking natural systems, the scientific community has witnessed the breathtaking progress made the relevant fields over the past decade. The advantage of unnatural systems, however, is not limited to mimicking the structures and functions of biological molecules. The potential of unnatural self-assembling or folding molecules may ultimately lie with the scientists, who are only limited by their imagination. Along these lines of thinking, we have succeeded in synthesizing and characterizing novel, and potentially useful self-assembly and folding systems that draw inspirations from nature, which may be applied to both natural and unnatural settings. Both our information-storing duplexes and helical foldamers have already shown exciting poten- tials in a variety of different applications, and are motivating us toward the development of new systems. Acknowledgements We thank the NASA (NAG5-8785), NSF (CHE-0314577), NIH (R01G M 63223), ONR (N000140210519) and ACS-PRF (37200-AC4) for financial support. References 1. Saenger, W. (1984) Principles of Nucleic Acid Structure,Springer, New York. 2. Zimmerman, S.C. & Corbin, P.S. (2000) Heteroaromatic modules for self-assembly using multiple hydrogen bonds. Struct. Bond. 96, 63–94. 3. Gong, B. (2001) Specifying non-covalent interactions: sequence- specific assembly of hydrogen-bonded molecular duplexes. Synlett 582–589. 4. Sijbesma, R.P. & Meijer, E.W. (2003) Quadruple hydrogen bon- ded systems. Chem. Commun., 5–16. 5. Prins, L.J., Reinhoudt, D.N. & Timmerman, P. (2001) Non- covalent synthesis using hydrogen bonding. Angew. Chem. Int. Ed. Engl. 40, 2383–2426. 6. Seebach, D. & Matthews, J.L. (1997) b-Peptides: a surprise at every turn. Chem. Commun., 2015–2022. 7. Gellman, S.H. (1998) Foldamers: a manifesto. Acc. Chem. Res. 31, 173–180. 8. Stigers, K.D., Soth, M.J. & Nowick, J.S. (1999) Designed mole- cules that fold to mimic protein secondary structures. Curr. Opin. Chem. Biol. 3, 714–723. 9. Gong, B. (2001) Crescent oligoamides: From acyclic ÔmacrocyclesÕ to folding nanotubes. Chem. Eur. J. 7, 4336–4342. 10. Hill, D.J., Mio, M.J., Prince, R.B., Hughes, T.S. & Moore, J.S. (2001) A field guide to foldamers. Chem. Rev. 101, 3893–4011. 11. Cheng, R.P., Gellman, S.H. & DeGrado, W.F. (2001) b-Peptides: From structure to function. Chem. Rev. 101, 3219–3232. 12. Cubberley, M.S. & Iverson, B.L. (2001) Models of higher-order structure: Foldamers and beyond. Curr. Opin. Chem. Biol. 5, 650–653. 13. Patch, J.A. & Barron, A.E. (2002) Mimicry of bioactive peptides via non-natural, sequence-specific peptidomimetic oligomers. Curr. Opin. Chem. Biol. 6, 872–877. 14. Schmuck, C. (2003) Molecules with helical structure: how to build a molecular spiral staircase. Angew. Chem. Int. Ed. Engl. 42, 2448–2452. 15. Sanford, A.R. & Gong, B. (2003) Evolution of helical foldamers. Curr. Org. Chem. 7, 1649–1659. 16.Hanessian,S.,Luo,X.H.,Schaum,R.&Michnick,S.(1998) Design of secondary structures in unnatural peptides: Stable helical c-tetra-, hexa-, and octapeptides and consequences of a-substitution. J. Am. Chem. Soc. 120, 8569–8570. 17. Seebach, D., Brenner, M., Rueping, M. & Jaun, B. (2002) c 2 -, c 3 -, and c 2,3,4 -Amino acids, coupling to c-hexapeptides: CD spectra, NMR solution and X-ray crystal structures of c-peptides. Chem. Eur. J. 8, 573–584. 18. Szabo, L., Smith, B.L., McReynolds, K.D., Parrill, A.L., Morris, E.R. & Gervay, J. (1998) Solid phase synthesis and secondary structural studies of (1 fi 5) amide-linked sialooligomers. J. Org. Chem. 63, 1074–1078. 19. Hungerford, N.L., Claridge, T.D.W., Watterson, M.P., Aplin, R.T., Moreno, A. & Fleet, G.W.J. (2000) Tetrahydrofuran amino acids: Secondary structure in tetrameric and octameric carbo- peptoids derived from a D -allo-5-(aminomethyl) tetrahydrofuran- 2-carboxylic acid. J. Chem. Soc., Perkin Trans. 1, 3666–3679. 20. Yang,D.,Qu,J.,Li,B.,Ng,F F.,Wang,X C.,Cheung,K K., Wang, D P. & Wu, Y D. (1999) Novel turns and helices in peptides of chiral a-aminoxy acids. J. Am. Chem. Soc. 121, 589–590. 21. Jorgensen, W.L. & Pranata, J. (1990) Importance of secondary interactions in triply hydrogen-bonded complexes – Guanine- cytosine vs uracil-2,6. J. Am. Chem. Soc. 112, 2008–2010. 22. Pranata, J., Wierschkem, S.G. & Jorgensen, W.L. (1991) OPLS potential functions for nucleotide bases – relative association constants of hydrogen-bonded base-pairs in chloroform. J. Am. Chem. Soc. 113, 2810–2819. 23. Murray, T.J. & Zimmerman, S.C. (1992) New triply hydrogen- bonded complexes with highly variable stabilities. J. Am. Chem. Soc. 114, 4010–4011. 24. Seto, C.T. & Whitesides, G.M. (1990) Self-assembly based on the cyanuric acid melamine lattice. J. Am. Chem. Soc. 112, 6409–6411. 1424 A. R. Sanford et al. (Eur. J. Biochem. 271) Ó FEBS 2004 25. Whitesides, G.M., Mathias, J.P. & Seto, C.T. (1991) Molecular self-assembly and nanochemistry – a chemical strategy for the synthesis of nanostructures. Science 254, 1312–1319. 26. Whitesides, G.M., Simanek, E.E., Mathias, J.P., Seto, C.T., Chin, D.N., Mammen, M. & Gordon, D.M. (1995) Noncovalent synthesis – using physical-organic chemistry to make aggregates. Acc. Chem. Res. 28, 37–44. 27. Conn, M.M. & Rebek, J. (1997) Self-assembling capsules. Chem. Rev. 97, 1647–1668. 28. Philp, D. & Stoddart, J.F. (1996) Self-assembly in natural and unnatural systems. Angew. Chem. Int. Ed. Engl. 35, 1155–1196. 29. Archer, E.A., Sochia, A.E. & Krische, M.J. (2001) The covalent casting of one-dimensional hydrogen bonding motifs: toward oligomers and polymers of predefined topography. Chem. Eur. J. 7, 2059–2065. 30. Corbin, P.S. & Zimmerman, S.C. (1998) Self association without regard to prototropy. A heterocycle that forms extremely stable quadruply hydrogen-bonded dimers. J. Am. Chem. Soc. 120, 9710–9711. 31. Murray, T.J. & Zimmerman, S.C. (1992) New triply hydrogen- bonded complexes with highly variable stabilities. J. Am. Chem. Soc. 114, 4010–4011. 32. Fenlon, E.E., Murray, T.J., Baloga, M.H. & Zimmerman, S.C. (1993) Convenient synthesis of 2-amino-1,8-naphthyridines, building-blocks for host-guest and self-assembling systems. J. Org. Chem. 58, 6625–6628. 33. Beijer, F.H., Sijbesma, R.P., Kooijman, H., Spek, A.L. & Meijer, E.W. (1998) Strong dimerization of ureidopyrimidones via quadruple hydrogen bonding Self-assembly mediated by the donor-donor-acceptoracceptor-acceptor-donor (DDA-AAD) hydrogen-bonding motif: Formation of a robust hexameric aggregate. J. Am. Chem. Soc. 120, 6761–6769. 34. Beijer, F.H., Kooijman, H., Spek, A.L., Sijbesma, R.P. & Meijer, E.W. (1998) Self-complementarity achieved through quadruple hydrogen bonding. Angew. Chem. Int. Ed. Engl. 37, 75–78. 35. Folmer, B.J.B., Sijbesma, R.P., Kooijman, H., Spek, A.L. & Meijer, E.W. (1999) Cooperative dynamics in duplexes of stacked hydrogen-bonded moieties. J. Am. Chem. Soc. 121, 9001–9007. 36. Archer, E.A. & Krische, M.J. (2002) Duplex oligomers defined via covalent casting of a one-dimensional hydrogen-bonding motif. J. Am. Chem. Soc. 124, 5074–5083. 37. Bisson, A.P., Carver, F.J., Eggleston, D.S., Haltiwanger, R.C., Hunter,C.A.,Livingstone,D.L.,McCabe,J.F.,Rotger,C.& Rowan, A.E. (2000) J. Am. Chem. Soc. 122, 8856–8868. 38. Ghadiri, M.R., Kobayashi, K., Granja, J.R., Chadha, R.K. & McRee, D.E. (1995) The structural and thermodynamic basis for the formation of self-assembled peptide nanotubes. Angew. Chem. Int. Ed. Engl. 34, 93–95. 39. Hartgerink, J.D., Clark, T.D. & Ghadiri, M.R. (1998) Peptide nanotubes and beyond. Chem. Eur. J. 4, 1367–1372. 40. Szabo, T., Hilmersson, G. & Rebek, J. (1998) Dynamics of assembly and guest exchange in the tennis ball. J. Am. Chem. Soc. 120, 6193–6194. 41. Rebek, J. (1999) Reversible encapsulation and its consequences in solution. Acc. Chem. Res. 32 (4), 278–286. 42. Gong, B., Yan, Y., Zeng, H., Skrzypczak-Jankunn, E., Kim, Y.W., Zhu, J. & Ickes, H. (1999) A new approach for the design of supramolecular recognition units: Hydrogen-bonded molecular duplexes. J. Am. Chem. Soc. 121, 5607–5608. 43. Zeng, H., Miller, R.S., Flowers, R.A. & Gong, B. (2000) A highly stable, six-hydrogen-bonded molecular duplex. J. Am. Chem. Soc. 122, 2635–2644. 44. Zeng, H.Q., Yang, X.W., Brown, A.L., Martinovic, S., Smith, R.D. & Gong, B. (2003) An extremely stable, self-complementary hydrogen-bonded duplex. Chem. Comm. 1556–1557. 45. Zeng,H.,Ickes,H.,Flowers,R.A.&Gong,B.(2001)Sequence specificity of hydrogen-bonded molecular duplexes. J. Org. Chem. 66, 3574–3583. 46. Yang, X.W., Martinovic, S., Smith, R.D. & Gong, B. (2003) Duplex foldamers from assembly-induced folding. J. Am. Chem. Soc. 125, 9932–9933. 47. Zeng, H., Yang, X., Flowers, R.A. & Gong, B. (2002) A non- covalent approach to antiparallel b-sheet formation. J. Am. Chem. Soc. 124, 2903–2910. 48. Connon, S.J. & Blechert, S., (2003) Recent developments in olefin cross-metathesis. Angew. Chem. Int. Ed. Engl. 42, 1900–1923. 49. Trnka, T.M. & Grubbs, R.H. (2001) The development of L2X2Ru¼CHR olefin metathesis catalysts: an organometallic success story. Acc. Chem. Res. 34, 18–29. 50. Dado, G.P. & Gellman, S.H. (1994) Intramolecular hydrogen- bonding in derivatives of b-alanine and c-amino butyric-acid – model studies for the folding of unnatural polypeptide backbones. J. Am. Chem. Soc. 116, 1054–1062. 51. Seebach, D., Overhand, M., Kuhnle, F.N.M., Martinoni, B., Oberer, L., Hommel, U. & Widmer, H. (1996) b-Peptides: Syn- thesis by Arndt-Eistert homologation with concomitant peptide coupling. Structure determination by NMR and CD spectroscopy and by X-ray crystallography. Helical secondary structure of a b-hexapeptide in solution and its stability towards pepsin Helv. Chim. Acta 79, 913–941. 52. Seebach, D., Ciceri, P.E., Overhand, M., Jaun, B., Rigo, D., Oberer, L., Hommel, U., Amstutz, R. & Widmer, H. (1996) Probing the helical secondary structure of short-chain b-peptides. Helv. Chim. Acta 79, 2043–2066. 53.Hanessian,S.,Luo,X.H.,Schaum,R.&Michnick,S.(1998) Design of secondary structures in unnatural peptides: stable helical c-tetra-, hexa-, and octapeptides and consequences of a-substitu- tion. J. Am. Chem. Soc. 120, 8569–8570. 54.Berl,V.,Huc,I.,Khoury,R.G.,Krische,M.J.&Lehn,J.M. (2000) Interconversion of single and double helices formed from synthetic molecular strands. Nature 407, 720–723. 55. Hill, D.J. & Moore, J.S. (2002) Helicogenicity of solvents in the conformational equilibrium of oligo(m-phenylene ethynylene)s: implications for foldamer research. Proc. Natl Acad. Sci. USA 99, 5053–5057. 56. Nelson, J.C., Saven, J.G., Moore, J.S. & Wolynes, P.G. (1997) Solvophobically driven folding of nonbiological oligomers. Science 277, 1793–1796. 57. Tanatani, A., Kagechika, H., Azumaya, I., Fukutomi, R., Ito, Y., Yamaguchi, K. & Shudo, K. (1997) Helical aromatic urea and guanidine. Tetrahedron Lett. 38, 4425–4428. 58. Zhu, J., Parra, R.D., Zeng, H., Skrzypczak-Jankun, E., Zeng, X.C. & Gong, B. (2000) A new class of folding oligomers: Crescent oligoamides. J. Am. Chem. Soc. 122, 4219–4220. 59. Parra,R.D.,Zeng,H.,Zhu,J.,Zheng,C.,Zeng,X.C.&Gong,B. (2001) Stable three-center hydrogen bonding in a partially rigidi- fied structure. Chem.Eur.J.7, 4352–4357. 60.Gong,B.,Zeng,H.,Zhu,J.,Yua,L.,Han,Y.,Cheng,S., Furukawa, M., Parra, R.D., Kovalevsky, A.Y., Mills, J.L., Skrzypczak-Jankun, E., Martinovic, S., Smith, R.D., Zheng, C., Szyperski, T. & Zeng, X.C. (2002) Creating nanocavities of tunable sizes: Hollow helices. Proc. Natl Acad. Sci. USA 99, 11583–11588. 61. van Gorp, J.J., Vekemans, J.A.J.M., Meijer, E.W. (2004) Facile synthesis of a chiral polymeric helix; folding by intramolecular hydrogen bonding. Chem. Commun. 60–61. 62. Yang, X., Brown, A.L., Furukawa, M., Li, S., Gardinier, W.E., Bukowski,E.J.,Bright,F.V.,Zheng,C.,Zeng,X.C.&Gong, B. (2003) A new strategy for folding oligo(m-phenylene ethyny- lenes). Chem. Comm. 56–57. Ó FEBS 2004 Well-defined secondary structures of unnatural oligomers (Eur. J. Biochem. 271) 1425 . REVIEW ARTICLE Well-defined secondary structures Information-storing molecular duplexes and helical foldamers based on unnatural peptide backbones Adam R loss of information and reduction of the observed association constants. In addition, one must consider that in heterocycle -based modules, secondary electrostatic