Making Molecular and Macromolecular Helical Tubes: Covalent and Noncovalent Approaches.

Aromatic foldamers possess well-defined cavity that can be stabilized by discrete intramolecular interactions including hydrogen bonding, solvophobicity, electrostatic repulsion, or coordination. Long foldamers can form dynamic deep helical tubular architectures that are not only structurally attractive but also useful hosts for guest encapsulation, chirality induction, delivery, and catalysis. This kind of helical tubular structures can be formed by single molecules or macromolecules or by connecting short-folded or helical segments through noncovalent or covalent forces. This perspective summarizes the recent advances on the construction of helical tubes and their properties and functions.

Introduction

Tubular structures are attractive targets in chemical and materials sciences because they are important scaffolds for the development of new functions and applications.[1−3] For carbon nanotubes and many other kinds of inorganic or composite nanotubes that exist only in the solid state, studies on their properties and functions are mainly limited to surfaces or interfaces. In contrast, molecular and supramolecular tubular structures are relatively flexible and more processible, and their cavity size can also be regulated much more easily. Moreover, modification and functionalization can be carried out from both inside and outside of the tubes. Therefore, in the past 2 decades, the construction of such family of “soft” tubular structures has received great attention. One classic family of such molecular tubes is that formed by covalent connection of cyclodextrins in the presence of linear templates.[1−3] However, the cavity of this kind of tubular structures is defined by the macrocycles which does not allow for further modulation. Uni(macro)molecular and multicomponent self-assembly on the basis of folded and helical conformations of their backbones provide another efficient approach for the generation of tubular structures. The diversity of the repeat segments and/or molecular components and accurate design of the backbones well enable the regulation and variation of the cavity diameter and depth of such kind of tubular structures. As a result, a large number of helical molecular and macromolecular tubes have been developed in the past decades.

α-Helix is the most important secondary structure of natural peptides and proteins. In 1974, Nolte et al. reported that polyisocyanides could form rigid helical conformations.[4] In 1980s, Seebach and co-workers developed efficient methods to synthesize chiral β-amino acids from which they could build short unnatural peptides that formed defined secondary structures.[5,6] Subsequent works by Gellman on β- and α/β-peptides revealed the formation of many elaborate artificial helical structures,[7,8] which led to the research field of foldamer chemistry.[8] However, because of their intrinsic structural nature, helical backbones formed by either natural or unnatural aliphatic amino acids are not able to generate hollow conformations. In contrast, aromatic backbones are more rigid to allow for the formation of hollow cavities.[9,10] In 1997, Moore et al. demonstrated this concept by using oligo(meta-phenylene ethynylene)s which formed a well-defined tubular conformation of ca. 0.7 nm diameter.[9] In 2001, Gong proposed that larger hollow cavity could be generated from folded aromatic amide oligomers by hybridizing meta- and para-substituted benzene rings in the backbones.[11] These seminal works remarkably inspired extensive studies for the construction of helical molecular tubes.

Long molecular tubes not only are structurally fascinating and important but also may exhibit new unique functions because they can produce microenvironments that are different from those formed by short foldamers or macrocycles of the identical repeat segments. Currently, such long helical tubes have been generated by single long oligomers and polymers or by the self-assembly of short oligomers driven by intermolecular noncovalent interactions. Many of these molecular and polymeric tubes have been revealed to display interesting properties or functions in, such as, molecular recognition and entrapment, chirality induction and transfer, ion transport, and reaction promotion or catalysis. This perspective summarizes the advances that concern the sequences that have more than one turn.

Helical Single Molecular Tubes

Aromatic oligomers can fold into helical conformations driven by intramolecular π–π stacking driven by solvophobicity or intramolecular hydrogen bonding.[12−15] Long oligomers can produce helices with ≥1 turn that form a deep tubular cavity for hosting guests of various sizes and shapes. Since 1997, several aromatic systems have been designed and developed as long helical foldamers. Their cavity diameter is typically well-defined, but the depth may vary considerably, depending on the backbone length and units per turn.

Oligo(arylene ethynylene) Backbones

Oligo(m-phenylene ethynylene)s developed by Moore et al. represent the first series of aromatic backbones that folded into helical conformations driven solvophobically in polar media, typically acetonitrile.[9] For this kind of backbones, at least seven repeat phenylene ethynylene units are required to allow for folding into one turn, i.e., the two benzene rings at the ends can stack each other intramolecularly. Thus, for the formation of stable helical conformations, relatively long backbones (≥10-mer) are required, which means that at least four stacking interactions are needed to stabilize the helical conformation. Given the rigidity of the backbones, this kind of helical tubes has a fixed inner diameter of approximately 0.7 nm. For 18-mer oligomer 1 (Figure ), the compact helical state produces a tubular structure of ca. 1.1 nm depth. The triglyme monomethyl ether group provided good solubility in organic solvents of discrete polarity. In benign chloroform of low polarity, aromatic stacking was very weak, and thus, the oligomers were conformationally flexible. Chiral triglyme monomethyl ether group induced 18-mer 2 to exhibit a strong Cotton effect with an exciton couplet in acetonitrile,[16] indicating a highly ordered helical conformation in which the chiral side chains biased the handedness of the backbone. In contrast, in chloroform an 18-mer analogue that bore no chirality centers showed no optical activity, pointing to a random conformation. Moore et al. also demonstrated that the hydrophobic cavity of helical oligo(m-phenylene ethynylene)s are good receptors for monoterpene guests. For rod-like chiral guests like 3a and 3b, the helices showed significant length- and shape-dependence.[17] In 40% aqueous acetonitrile, 3a had a complementary length to the cylindrical cavity of 20-mer 4a and 22-mer 4b, whereas for 3b, these two oligomers displayed even higher specificity. In contrast, longer 24-mer 4c was found to have a lower affinity for 3b relative to 4a and 4b because of the steric hindrance of the triphenylmethyl capping groups of 3b.

Figure 1

Helical conformation of the aromatic backbone of 18-mer 1, highlighting an inner diameter of 0.7 nm and depth of 1.1 nm. The side chains and end groups have been removed.

The benzene subunits of the above oligomers have been replaced with pyridine, (bi)indole, or indolocarbazole segments, which led to the generation of several new sequences.[18−20] In this category, Inouye and Abe had prepared oligo(m-ethynylpyridine)s 5a and 5b that had 18 and 24 repeat units,[18] respectively, which corresponded to three and four turns of their respective helical backbone. In solution, this kind of aromatic sequences has been revealed to adopt unfolded conformations because the pyridine nitrogen atoms are favorably located on the opposite sides of the C≡C bonds which could attenuate the dipoles of the pyridine rings. Both 5a and 5b showed induced circular dichroism (CD) through binding chiral monosaccharides in less polar organic solvents such as dichloromethane, whereas shorter analogues did not, showing that these long oligomers possessed stronger binding affinity through forming helical conformations, which exhibited helicity bias due to the induction of the chiral guests. Compared with that of the benzene- and pyridine-based oligomers, the repeat degrees of the larger indole (6a and 6b) and indolocarbazole (7a and 7b) oligomers, all developed by Jeong et al., are lower. Both series of backbones are inherently flexible. However, their high aspect ratio still enabled deep helical conformations. Thus, in acetonitrile, sulfate anion-induced 6a and 6b to fold into helical conformations that had 2 or 2.5 turns.[19] The carboxylate group of 7a and 7b provided solubility in water.[20] In water, the two oligomers could encapsulate chloride anion through helical conformation that had 1 or 1.5 turns.

Oligo(aromatic amide and hydrazide) Backbones

Hydrogen bonding-driven aromatic amide and hydrazide oligomers constitute another large family of molecular tubes which have a cavity of different inner diameters. Compounds 8–15 are the longest oligomers of the representative sequences in this category.[11,21−27] Chen et al. reported that the two adjacent amide units of 12 did not form intramolecular hydrogen bonding. Instead, they produced large torsion but still allowed the formation of intramolecular five-membered N–H···N hydrogen bonding, which induced the backbone to stack compactly to produce a helix of four turns with no cavity (Figure ).[26] All other oligomers are induced by successive intramolecular hydrogen bonding to form a tubular helical conformation. One of us found that in chloroform the cavity of 9 and 15 could bind alkylated chiral saccharides by forming intermolecular multiple N–H···O hydrogen bonding, which also led to important helicity bias of the hosts.[21,27] As expected, the binding of 13-mer 15 toward saccharide guests was stronger because of its deeper cavity (ca. 2.5 turns). We also found that the cavity of 10 and 11 was smaller but could still host aliphatic ammoniums.[22,23] Zeng reported the crystal structure of 13 and 14 which revealed that the contractive geometry of the pyridine ring and the intramolecular hydrogen bonding caused that a little more than four repeat units formed one turn.[24,25] In this helical conformation, the two farthest pyridine rings underwent partial stacking to produce a small cavity that hosted water molecules stabilized by intermolecular hydrogen bonding (Figure ).[25] Pyridine-derived amide oligomers can also form double, triple, or quadruple helices. However, these more complicated helical structures typically produce even smaller cavity which is not able to host any guest.[14,28−30] 8-Aminoquinoline-2-carboxylic acid-derived amide oligomers can produce one turn with only three repeat units, and a 13-turn helix has been constructed from this segment, which has a length of 5.1 nm.[31]

Figure 2

Crystal structure of oligomer 12, showing the small cavity of the helical conformation that hosts water molecules through intermolecular hydrogen bonding.

Figure 3

Crystal structure of oligomer 14, showing no cavity of the helical conformation.

Huc and co-workers introduced a general approach of designing hollow helical tubes by incorporating different heterocycles, including pyridine, quinoline, pyrido[3,2-g]quinoline, 1,8-naphthyridine, or pyridazine, into one sequence.[32] The resulting single molecular capsules are able to entrap small solvents, alkali and alkaline earth metal ions, or organic acid guests.[33−35] With rationally designed linear molecules as templates, a number of single and double helices have been induced to form long tubular structures to thread the guests to form unique dynamic interlocked supramolecular systems.[32]

Oligo(aromatic triazole) and Hybrid Backbones

The C5–H atom of 1,2,3-triazole is a weak hydrogen donor because of the electron-withdrawing effect of the three nitrogen atoms. Thus, 1,2,3-triazole can form both intermolecular and intramolecular hydrogen bonding. The intermolecular hydrogen bonding has been used to induce benzene-triazole-alternate oligomers to fold into helical conformations.[36] For example, Craig et al. reported that 16 folded into one-turn helix by forming intermolecular multiple C–H···Cl– hydrogen bonds.[37] With halide anions as the template, longer oligomers incorporating benzene and pyridine units formed tubular helices with deeper cavity, and the binding may occur in organic and aqueous media, as demonstrated by Hecht and Jiang et al.[38−40] We found that intramolecular C–H···OR or C–H···F hydrogen bonding can also induce the same kind of oligomers to fold into helical conformations.[41,42] This kind of foldamers has a cavity of approximately 1.8 nm, and oligomer 17 is the longest one that forms more than one turn of helical conformation.[41] All the N-2 and N-3 atoms of the triazole rings are orientated inward and they were revealed to form intermolecular N···ICF2 halogen bonds with organohalogens, such as triiodide 18, in dichloromethane or its mixture with n-hexane.

In the absence of intramolecular and intermolecular hydrogen bonding, aromatic triazole oligomers are usually conformationally flexible. Our group illustrated that when one or two such inherently flexible segments were attached to a hydrogen-bonded folded amide segment of at least four repeat units, the resulting hybrids, such as 19–21, underwent folding-inducing-folding to give rise to a new kind of hybridized tubular helices in apolar solvents like benzene and n-hexane (Figure ).[43] However, in benign chloroform or dichloromethane, this process did not occur. Thus, we attributed this induction process to solvophobically driven intramolecular stacking between the folded amide segment and the attached triazole segment(s). Moreover, for enantiomers R-20 and S-20, the chiral group was able to induce the whole framework to generate strong helicity bias, suggesting the formation of chiral helical tubes. Molecular dynamic modeling showed that the longest oligomer 21 produced a helical tube that has about four turns with a depth of 1.1 nm.

Figure 4

Schematic representation of solvophobically driven folding-inducing-folding of oligomers 19, chiral R-20 and 21.

Helical Single Macromolecular Tubes

The preparation of long oligomers requires multistep synthesis and is usually time-consuming. Linear polymers are much easier to obtain by one-step coupling of rationally designed bifunctional precursors and thus have attracted great attention of chemists for the generation of macromolecular tubes of deep cavity.[44] Reported macromolecular backbones have been developed on the basis of their oligomeric analogues. However, the driving forces for the formation of helical conformations may be different.

Poly(arylene ethynylene) Backbones

Hecht and Khan developed the first example of polymeric tubular helices from m-phenylene ethynylene-based backbones.[45] Pd-catalyzed coupling of diiodine 22 and ethynyltrimethylsilane afforded polymer P23 (Scheme ), which has a number-average degree of polymerization (DP) of 60 and a typical polydispersity index (PDI) of 1.3. As expected, this polymer folded into the dynamic helical state in polar acetonitrile, as observed for its oligomeric analogues. Solvent denaturation experiments in binary chloroform and acetonitrile, by monitoring the population of random-coiled and helical conformations through differences in the transoid and cisoid absorptions, revealed the cooperative folding behavior of the backbone, and the helical conformation was more stable than that of shorter oligomers. Cross-chain [2 + 2] photodimerization of P23 in acetonitrile led to the formation of side chain-cross-linked helix P24 (Scheme ). The absorption spectrum of P24 in acetonitrile revealed no significant changes with the addition of increasing amount of the chloroform denaturant, which indicated that its locked helical conformation was stable and the addition of chloroform did not cause de-folding. By assuming one that turn had six repeat units, a DP of 60 suggested that polymeric helices P23 and P24 should have about ten turns and thus a depth of approximately 3.5 nm.

Scheme 1

Synthesis of Polymers P23 and P24 (DBU = 1,8-Diazabicyclo[5.4.0]undec-7-ene)

Hecht et al. also prepared poly(propylene oxide)-poly(m-phenylene ethynylene) blocks and graft copolymers P25 that had a low DP of 8–11 for the aromatic backbone.[46] These polymers were found to partially fold into helical conformation in methanol. Their CD spectrum in methanol or chloroform all exhibited a positive signal in the region of the polyethylene glycol backbone absorption. However, no bisignate exciton couplet was observed related to the helical conformation of the helical poly(m-phenylene ethynylene) segment, indicating that chirality transfer to the helical aromatic backbone did not occur, probably because of the fact that the backbones were too short.

Inouye and Abe and co-workers found that, as its oligomeric analogues (5a and 5b), poly(m-ethynylpyridine) P26 was also conformationally flexible in chloroform.[18] Upon binding chiral alkylated saccharides, both oligomers and polymer were induced to produce helicity bias for the resulting tubular helices. P26 could extract native monosaccharides into less polar organic solvent such as chloroform. By introducing hydrophilic poly(ethylene glycol) chains, the resulting polymers were found to be water-soluble and complex-native saccharides.[47] The complexation should take place also through forming helical conformations. Because this binding was relatively weaker, enhanced intramolecular π stacking might make greater contribution in stabilizing the helical conformations. When chiral hydrophilic side chains were introduced, the resulting water-soluble polymers R- and S-P27, which had different molecular weights and sizes depending on the fractions obtained by gel permeation chromatography (GPC), displayed different induced CD signals which related to the aromatic backbones, depending on,[48] while the heaviest fractions exhibited CD responses to chiral mannose, cyclodextrins, and polysaccharides, which was attributed to the formation of the single helix and its conversion to more complicated entangled intramolecular duplex. Polymer S-P28, which bore both chiral and achiral hydrophilic side chains, was soluble in both polar and apolar organic solvents.[49] This polymer could complex metal salts in water or aqueous ethanol. This binding led to positive Cotton effect and hypochromism around 360 nm, which had been ascribed to the coordination of the cations to the amide side chains. This coordination was proposed to enhance the intramolecular aromatic stacking and consequently the helicity bias.

Poly(aromatic amide) Backbones and Analogues

Arylamide-based helical polymers represent a large family of backbones that can form helical tubes with tunable diameter and depth. The condensation of isophthalic acid and benzene-1,3-diamine has been studied in solution and the solid state,[50,51] which afforded poly(m-phenyleneisophthalamide). However, their insolubility did not allow any studies on the conformations of the backbones. As the extension of oligomer 11,[23] Guan and one of us designed the preparation of polymer P30 from the self-coupling reaction of 29 in N-methylpyrrolidone (NMP) (Scheme ).[52] Nevertheless, the reaction actually afforded polymer P30′ as a result of the decomposition of the methoxyl groups to hydroxyl groups (71%) during the reaction. Treatment of P30′ with methyl sulfate led to the formation of polymer P30. The molecular weight of P30 (Mn = 2.43 KD, PDI = 1.47) was lower than that of P30′ (Mn = 3.75 KD, PDI = 1.72). This difference was attributed to the flexible conformation of P30′ because of the changed hydrogen bonding motif of the hydroxyl groups and the helical conformation of P30, which displayed an underestimated molecular weight.[53] Polymer P30 was estimated to have about three helical turns, which corresponded to a cavity depth of 1.1 nm. This helical conformation possesses a very small cavity, whose binding for chiral 1-phenylethan-1-amine induced the polymer to exhibit an excess of one particular helical handedness through intermolecular hydrogen bonding.

Scheme 2

Synthesis of Polymers P30′ and P30

Zhu and Gong and co-workers also prepared polymers P31a–c to study the effect of the solvent on the folding of the backbones.[54] The two alkoxyl groups on the isophthalamide units formed stable intramolecular O–H···N hydrogen bonding to favor the helical state, whereas the 3,5-diaminobenzamide moieties could not. The Mn’s of P31a–c were determined to be 10.2, 18.6, and 35.1 KD, which corresponded to helical tubes having roughly 5, 9, and 18 turns, respectively. All of the polymers had low molecular weight dispersity (1.02 to 1.03) and thus very narrow distributed lengths. The side chain amide units were evidenced to form cross-turn hydrogen bonding to stabilize the helical conformation, whereas the chiral side chains transferred the chirality to the backbones to induce helicity bias in chloroform. Notably, the intensity of the resulting Cotton effects was correlated with their respective helix length, indicating that induction of the chiral side chains displayed cooperativity.

Typically, aromatic amide-based oligomers and polymers need to use intramolecular hydrogen bonding to stabilize their folded or helical conformations. Our study showed that in the absence of such intramolecular hydrogen bonding, oligomeric backbones prefer to form intermolecular hydrogen-bonded duplexes.[55] To further explore the possibility of using aromatic stacking to induce the folding of aromatic amide backbones, we also prepared polymer P32a.[56] This polymer had Mn and weight-average molecular weight (Mw) of 32.0 and 59.0 KD, respectively, which corresponded to a weight dispersity of 1.8 and a 9-turn helix with a depth of ca. 3.1 nm. Molecular dynamics modeling showed that the cavity had a size of 1.3 nm (Figure ). P32a was found to form tubular helices in water and many organic solvents of varying polarity. It was proposed that in water or polar organic solvents, solvophobicity was the main driving force, whereas in less polar organic solvents, both solvophobicity and cross-turn hydrogen bonding made important contributions. N-Methylated polymer P32b was not soluble in water and only formed the helical conformation in polar organic solvents such as methanol, supporting that solvophobicity played a key role in inducing the formation of the helical conformation. When the 1,8-naphthalimide unit was replaced with smaller naphthalene, the resulting polymer did not spontaneously fold into helical conformation. However, chiral natural amino acid anions could induce the formation and helicity bias of the helical conformation in chloroform through forming multiple intermolecular hydrogen bonding.[57] With 2,2′-bipyridine being introduced into the backbones, its binding toward Ni2+ could lead to the formation of two different helical tubes through the configuration overturn of the bipyridine units.[58]

Figure 5

(a) Side and (b) top views of the optimized right-handed helix formed by polymer P32a of 9 turns with a cavity diameter of 1.3 nm. The tetra(ethylene glycol) chains were replaced with methyl groups for clarity.

Our group has prepared aromatic hydrazide oligomers and polymer (P33) that bear amphiphilic short peptide chains which could enhance the membrane-inserting ability.[59,60] The polymer had an Mn of 15.8 KD, which corresponded to the polymerization degree of about 10 and 3.5 helix turns. Both oligomers and polymer were found to insert lipid bilayers to form uni(macro)molecular channels (Figure ). The polymer exhibited an NH4+/K+ selectivity that was higher than that of gramicidin A as well as long open duration. Moreover, its different components could generate a trace of multiple currents because of their varying length.

Figure 6

Schematic presentation of the transmembrane channel formed by helical polymer P33.

Polymers P34a–c have Mn of 14.4, 12.2, and 7.2 KD and Mw of 31.2, 31.1, and 2.1 KD.[61] The Mn values corresponded to 26, 24, and 14 aromatic subunits and 4.3, 4.0, or 2.3 turns for the designed helical conformation. The polymers were soluble in water because of the attachment of hydrophilic oligo(ethylene glycol). Their repeated benzene rings bear three, two, or one chiral centers. The intramolecular hydrogen bonds formed by one of the repeated two isophthalamide units had been demonstrated to favor the formation of helical conformations.[27] In polar organic solvents such as methanol or acetonitrile, all the polymers folded into helical conformations, which, however, did not exhibit helicity bias. In more polar water, helicity bias was displayed, as evidenced by the formation of induced CD signals. This helicity bias was comparable for P34a and P34b, indicating that their chiral side chains on the 5-position of one of the repeated two benzene rings did not considerably contribute to the helicity induction.

Meijer et al. also prepared polymers P35a–c from the corresponding diamines and diisocyanic acids that bore chiral side chains.[62] Stable intramolecular six-membered C=O···H–N hydrogen bonding and the rigidity of the aromatic urea units enabled helical conformations for this kind of polymers that had a cavity of 1.4 nm. Crude polymer P35a had been separated by GPC into high, intermediate, and low Mw fractions. In THF, only the high Mw fraction (n = ∼30) folded into a stable chiral helical conformation, which was consistent with the formation of a deep tubular structure (ca. 5 turns). Polymer P35b, with a DP of about 20, could form chiral helical conformation in water.[63] The intramolecular hydrogen bonding should be weaker because of the competition of water molecules. Thus, cross-turn intramolecular π–π stacking would make important contribution for the formation of the helical conformation. The large oligo(p-phenylenevinylene) (OPC) chromophores of polymer P35c was found to enhance backbone stacking.[64] CD studies of the long-chain fraction of P35c revealed induced Cotton effect for the OPV chromatophores in THF but not in chloroform, again showing that π–π stacking played a key role in the formation of the helical conformation in a more polar medium.

Poly(aromatic triazole and oxadiazole) Backbones

Hecht and Meudtner prepared polymers P36a and P36b that consisted of alternating pyridine-triazole-benzene-triazole units.[65] It was found that the triazole-pyridine segments adopted an anti–anti configuration because of electrostatic repulsion of their nitrogen atoms,[66] which led both polymers to form a helical conformation even in the absence of anionic templates, although the rotation of the benzene-connected C–N bonds was not controlled. Moreover, the chiral side chains could induce helicity bias. Even though no molecular weights were given, both polymers should be obviously longer than previously developed oligomeric analogues.[66] Thus, π–π stacking should also be an important stabilizing factor for the helical conformations.

Klumperman et al. prepared polymers P37 and P38,[67] which were found to exist as random coils in DMF. In more polar binary DMF and water (9:1, v/v), S- and R-P37 formed helical conformations with helicity bias becasuse of the induction of the chiral side chains. Molecular modeling showed that the helical conformation had 14.5 repeat units per turn. Thus, the helices possessed a very large cavity of approximately 30.6 Å in diameter (Figure ). With the increase of the water content, the helices stayed intact but also further aggregated into tertiary nanotubes. Adding rigid and hydrophobic poly(γ-benzyl-l-glutamate) (PBLG) α-helix, which has a diameter of about 1.6 nm, to the solution of P38 caused the polymer to undergo coil-to-helix transition through threading into the helical cavity of the polymer. Again, this threaded helix exhibited helicity bias because of the chirality induction of PBLG.

Figure 7

Calculated helix of polymers P37 and P38, highlighting a large cavity diameter.

Dong and Liu and co-workers prepared polymers P39a and S- and R-P40 which consisted of alternating 1,3,4-oxadiazole and pyrido[3,2-g]quinolone segments.[68]P39a had a molecular weight of up to 12000, which was evidenced by MALDI-TOF mass spectrometry, and DP of up to 30. Its Mn and PDI were evaluated to be 6000 and 1.16, respectively. Scanning tunneling microscopy images confirmed that all of the polymers formed helical structures, with the diameter being only ca. 0.55 nm, and S- and R-P40 exhibited helicity bias in dichloromethane. Moreover, the helical structures could further bind to form double helices, which had been observed before only for heterocyclic amide-derived backbones.[28,29] The same group further prepared polymers P39b and P39c, which were more soluble in organic solvents because of the introduction of longer n-dodecyl chains.[69]P39b had Mn of 18700 and PDI of 1.21. Again, both polymers were found to form helical conformations in organic solvents. Molecular modeling predicted that this series of backbones had the helical pitch of 0.36 nm and 3.2 units per turn. Thus, the helix of P39b was estimated to have a depth of ca. 3.3 nm. P39b was revealed to insert into lipid bilayers to form unimacromolecular channels, which could mediate the transport of protons and cations. The shorter P39c also could form transmembrane channels but worked through a dimeric tubular structure. Mechanism studies supported that both channels transported ions through the interior of the helices.

Macromolecular Backbones Formed Through End-to-End Dynamic Covalent Bonding and Coordination

Partial replacement of the ethynyl linkers of the oligo- or poly(meta-phenylene ethynylene) backbones with imine units does not break the tendency of the backbones to stack into helical conformations. However, the dynamic nature of the imine bond allows for the study for the equilibrium distribution of different imine-derived backbones when multiple amine and aldehyde precursors are used as precursors for building the helical backbones. To test this, Moore and co-workers prepared oligomers 41 and 42 to study their imine metathesis, which would result in the formation of polymers of different lengths.[70,71] It was found that in benign solvents such as chloroform, equilibrium constants for different polymer products was close to unity. However, in polar media like acetonitrile, long polymer products could fold into more stable helical conformations through stacking-promoted nucleation, which caused the shifting of the equilibrium to longer, high-molecular weight products. Through this imine metathesis, long polymeric helical tubes can be produced under mild reaction conditions. The melt condensation of the corresponding diamine and dialdehyde only gave rise to polymers of low molecular weights,[71] indicating that no stacking-promoted nucleation event took place.

Moore and Wackerly also prepared short dipyridines 43a–c which themselves adopted flexible conformations.[72] In acetonitrile, their pyridines could be coordinated to Pd2+. The coordination of hexamer 43b gave rise to a 1 + 1 macrocycle which could aggregate into columnar tubes. In contrast, for 43a and 43c, this binding led to the formation of new coordination polymers that could fold into helical conformations (Figure ). As expected, a paralinked control oligomer only exhibited simple isodesmic step-growth polymerization. The introduction of imine bonds or Pd2+ ions should cause just a small deformation of the resulting helical tubes. Thus, these dynamic supramolecular tubular structures are also potentially useful hosts for studying molecular recognition for linear guests.

Figure 8

Pyridine-Pd2+ binding of 43a–c forms metal–ligand complexes to generate coordination supramolecular helical polymer or π-stacked columnar polymer.

Dong and Liu and co-workers prepared polymer P44a through the condensation of the corresponding dialdehyde and diacylhydrazine precursors.[73] Its Mn and PDI were estimated by GPC to be 31 KD and 1.46, respectively. The intramolecular hydrogen bonding of the pyrido[3,2-g]quinoline and the hydrazone units, together with the electrostatic repulsion between the imine and connected pyridine nitrogen atoms, forced the backbone to produce a tubular helix of approximately 1.0 cavity diameter in dichloromethane. Interestingly, the tubular helix could catalyze the oxidation of benzenethiol into diphenyldisulfide in the presence of hydrogen peroxide through the polar microenvironment inside the tube. The aldehyde units at the end of the backbone could be converted into imine units by treating with amines. When chiral 1-phenylethan-1-amines were used, the resulting polymers P44b and P44c exhibited induced CD signals, indicating the formation of chiral helical tubes.

Macromolecular Backbones Formed Through End-to-End Hydrogen Bonding

Although a large number of aromatic amide-derived oligomers and polymers have been developed to form tubular structures, the stacking of the tubular structures into long one-dimensional (1D) supramolecular structures has been a challenge. Orthogonal end-to-end noncovalent interactions would help to promote such 1D stacking. However there have not yet efficient approaches to realize this in solution.

Zeng and Huo developed a “sticky” strategy for short pyridine amide-based helices to form column-like aggregates in the solid state through end-to-end hydrogen bonding, which may be N–H···O or N–H···O motif.[74] For example, compound 45 formed such 1D tubes which entrapped methanol in the cavity (Figure ).[75] Water or dichloromethane molecules might also be entrapped in the cavity, depending on the conditions for growing the single crystals. For example, pentamer 46 afforded similar 1D supramolecular tubes which hosted a hydrogen-bonded water line (Figure ).[76] Tetramers generally could not form such column-like structures because they cannot produce a full helical turn. One hexamer of the same sequence was also found to be unable to produce stacking columns. Moreover, many pentamers that bear different end groups and/or side chains did not stack in a similar way. Clearly, the formation of such end-to-end intermolecular hydrogen bonding needs, among others, subtle balance of the length of the backbones, the size and electronic nature of substituents at different positions, and the steric effect of different segments.

Figure 9

Crystal structures of 45·MeOH. The neighboring helices were linked through end-to-end C=O···H–C(Bn) hydrogen bonding to produce column-like tube that hosted methanol molecules.

Figure 10

(a) Side view of the helical geometry taken up by 46 in the solid state, (b) top view of (46) illustrating two sets of complementary end atoms each containing two H atoms in gray and one O atom in red that participate in forming two weak intermolecular H-bonds of C=O···H–C type. These H-bonds help to create (c) 1D chiral helical stacks consisting of helices of the same handedness and (d) chiral aquapores of ∼2.8 Å across for accommodating (e) H-bonded 1D water chains.

Conclusion and Outlook

In this perspective paper, we have summarized the advance of using helical molecules or macromolecules to build tubular structures. The tubes may be formed by one or multiple molecules which are connected by dynamic covalent bonding, coordination, and intermolecular hydrogen bonding. The cavity size and depth of yet reported (macro)molecular tubes can be tuned from less than 1 nm to several nm. Currently, their functions have been investigated mainly in the category of molecular recognition. Moreover, examples of transmembrane transport and catalysis or reaction promotion have also been reported. We believe that further exploration of new structures for the development of new properties or functions will be crucially important in the future. As for the development of new structures, systems with a very large cavity size should be attractive for the design of unimolecular tubes, which would allow for the introduction of functional groups in the cavity for exploiting new functions such as catalysis. Up to now, many polymeric tubes have been reported. However, efficient methods for the preparation of very long polymeric backbones are still not available. As a result, the length of the macromolecular tubes reported yet is still limited. For the purpose of increasing the DP, it is necessary to find the ways to overcome the increasing stacking-initiated steric hindrance during the polymerization.[77] Extension of large tubular helices, particularly those driven by hydrogen bonding, is expected to produce tremendous deformation and impose great stress.[78] Suitable operation of this process may be utilized to modulate the mechanic property of new smart materials.