Hydrindacenes as Versatile Supramolecular Scaffolds

0 downloads 4 Views 6MB Size Report
Hidetoshi Kawai. Department of Chemistry, Faculty of Science, Tokyo University of Science, Shinjuku-ku, Tokyo 162-8601. E-mail: [email protected]

Award Accounts

The Chemical Society of Japan Award for Young Chemists for 2007

Hydrindacenes as Versatile Supramolecular Scaffolds# Hidetoshi Kawai Department of Chemistry, Faculty of Science, Tokyo University of Science, Shinjuku-ku, Tokyo 162-8601 E-mail: [email protected] Received: October 20, 2014; Accepted: December 3, 2014; Web Released: March 15, 2015

This report describes the synthesis and structural features of supramolecular assemblies of s-hydrindacene (1,2,3,5,6,7-hexahydro-s-indacene), such as macrocycles with directionally persistent peripheral functionality, allosteric receptors, adrenaline receptors, and rotaxane molecular shuttles. These hydrindacene-based assemblies also exhibit allostericity due to induced polarization of amides or to entropy-driven switching between the imine and hydrogen bonds.

1. Introduction Supramolecular chemistry1 involves the design and synthesis of supramolecular building blocks that assemble into ordered superstructures with tunable functions through noncovalent interactions. A fundamental design principle that allows supramolecular building blocks1­10 to assemble into the desired functional structures is the ability to align both interacting units and functional units (preorganized) on a suitable molecular platform with rigid/flexible linker units (Chart 1). In many supramolecular building blocks, an inherent structure®planar, concave, cleft, or macrocyclic shape®combines with interacting units and functional units to produce properties such as self-assembly, guest binding, modulated spectroscopic properties, electron conductivity, redox activity, catalytic activity, and mobility. Recently, ureido-pyrimidones,3 1,3,5benzenetricarboxamides (BTAs),4 oligo Tröger’s bases,5 hexabenzocoronenes,6 perylene bisimides,7 cucurbiturils,8 and pillararenes9 have been reported as representative examples that have attracted attention across scientific areas. A molecular scaffold, which is a pure structural unit consisting of relatively rigid hydrocarbon frameworks10 with a welldefined alignment of substituents, does not possess outstanding molecular properties except for π-conjugated polycyclic aromatic hydrocarbons (PAHs).6,11 However, the featureless

Chart 1. Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

physical character of these molecular scaffolds can be useful in conjunction with various properties. For example, 1,3,5substituted triethylbenzenes,12 adamantanes,13 iptycenes14 such as triptycenes and pentiptycenes, and oligomethylene-bridged arenes15 have been used for supramolecular design (Scheme 1). Requirements of supramolecular scaffolds include rigidity or moderate flexibility of the framework, structural modification, directionality of substituents, and quantitative accessibility. Development of a new molecular scaffold could generate unprecedented new supramolecular properties. Thus, hydrindacene16 (1,2,3,5,6,7-hexahydro-s-indacene), which has welldefined geometrical features that create supramolecular assemblies (Scheme 2), is an attractive structural motif, and has been investigated along with rigid α-amino acids (AAA),17 bulky protecting groups,18 rigid rods,19 and boronate cage compounds.20 The present report describes the design of hydrindacenebased supramolecular assemblies, such as hydrindacene-based macrocycles,21 an allosteric receptor,22 an adrenaline receptor,23 and imine-bridged rotaxanes,24 as well as a new collaboration between interactive units, such as allosteric functionality in amide hydrogen bonds22 and entropy-driven switching in dynamic imine-hydrogen-bonding systems.24 2. Structural Features in s-Hydrindacenes As shown in Scheme 2, the s-hydrindacene (E/Z)-1 is composed of a rigid aromatic ring with a five-membered alicyclic ring on either side. These assemblies result from the multidimensional modification of hydrindacene through the addition of substituents X, Y, and Z (Scheme 2). Substituents X and Y, at the 2,6-positions of the five-membered rings, and substituent Z at the 4,8-positions of the aromatic ring, are mutually perpendicular. The presence of the two X,Y substituents results in two possible diastereomers of the resulting hydrindacenes. Thus, the (Z)-isomer has the two Y substituents on the same

© 2015 The Chemical Society of Japan | 399

Scheme 1. Hydrocarbon-based scaffolds.

Scheme 2. Structural features of hydrindacene derivatives.

Scheme 3. Preparation of hydrindacene derivatives.

side of the molecular plane in parallel, while the (E)-isomer has the substituents at the opposite side in an antiparallel fashion. The less bulky Y groups at the 2,6-positions preferentially occupy pseudoaxial positions, although the five-membered rings to which they are attached can be flipped. Either of these isomers is suitable for adjusting the geometry of a supramolecular building block. The (Z)-isomers are suitable for the construction of macrocycles21 and concave molecular receptors23 (endoreceptors), whereas the (E)-isomers are suitable for multivalent receptors22 (exoreceptors) and rotaxane axle units24 with both interactive portions appropriate for a macrocycle and the attachment portions of end-cap groups. 3. Preparation of Hydrindacene Derivatives In general, 2,6-substituted hydrindacenes are prepared from double-ring condensations between tetrakis(bromomethyl)400 | Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

benzenes and active methylene compounds, such as diethyl malonate or ethyl cyanoacetate, under basic conditions (Scheme 3).25 As reported by Závada,25 the selectivity for cyclization to five-membered rings toward polysubstitution or oligomerization is influenced by the acidity of the active methylene compounds and polarity of the reaction media. Active methylene compounds with a small pKa, such as malononitrile, induce efficient cyclization even in polar solvents such as DMF or DMSO. In contrast, high pKa reagents, such as phenylacetonitrile, tend to produce polysubstituted products in polar solvents. In these cases, a strong base in a two-phase system of CH2Cl2/PTC/NaOHaq or toluene/PTC/ NaOHaq is necessary for efficient cyclization of the fivemembered ring.26 This approach is practical for 4,8-unsubstituted hydrindacenes (Z = H), but 4-substituted or 4,8-disubstituted derivatives (e.g., Z = Br, I, OMe) can be produced only © 2015 The Chemical Society of Japan

Figure 1. X-ray structures of the (E)- and (Z)-isomers of hydrindacene derivatives (X = 4-bromophenyl, Y = CN).

Scheme 4. Preparation of hydrindacene macrocycles.

Figure 2. X-ray structures of a) dimer 2a, c) trimer 3a¢(hexane), d) tetramer 4¢(CH2Cl2)4. CH2Cl2 molecules are omitted for clarity. b) Pictorial presentation of HOMO in dimer 2a calculated at the B3LYP/6-31G* level.

in low yield under these conditions, presumably due to steric hindrance that prevents cyclization. The 2,6-substituted hydrindacenes do not have strong regioselectivity. In almost all cases, the (E)- and (Z)-isomers are produced in a 1:1 ratio under homogeneous reaction conditions. However, the (Z)-isomer is preferentially produced in a ratio of ca. 3:2 under two-phase systems, presumably due to a difference in hydrophilicity between substituents. These structures require single-crystal X-ray crystallography to differentiate between the orientation of C2h in the (E)-isomer and C2v in the (Z)-isomer, which cannot be determined from NMR spectra (Figure 1). The high Rf value obtained from TLC for the (E)-isomers due to the offset of dipole moment also can be applied to the qualitative determination of the stereostructure. A variety of substituents have been introduced at the 2,4,6,8positions. At the 2,6-positions, aryl groups, esters, and CN groups were introduced directly through cyclization with active methylene compounds. These groups also can be converted to ethynyl, CHO, and amide groups, respectively. At the 4,8positions, Br, I, and MeO groups could be introduced through 1,4-disubstituted 2,3,5,6-tetrakis(bromomethyl)benzenes as a cyclization precursor. In addition, 4,8-dibromohydrindacenes21,22 could be transformed into 4,8-disubstituted derivaBull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

tives through Suzuki­Miyaura coupling (Z = Ar, styryl),21 Stille coupling (Z = thienyl),21 Rosenmund­von Braun synthesis (Z = CN), and the Ritter reaction (Z = amide).22,23 4. Hydrindacene-Based Macrocycles Hydrindacene derivatives have been investigated as molecular scaffolds for various supramolecular building blocks. A series of novel macrocycles 2­4 have been prepared, all of which are functionalized in both the horizontal and vertical directions relative to the plane defined by the macrocyclic backbone (Scheme 4).21 Substituents X and Y, at the 2,6positions on the five-membered rings, and substituent Z, at the 4,8-positions on the aromatic ring, are mutually perpendicular. The two Y substituents at the 2- and 6-positions of the (Z)hydrindacenes are designed as linking groups to produce the macrocycle, whereas the two X groups at the 2- and 6-positions and the Z groups at the 4- and 8-positions remain perpendicular and are used as peripheral functionalities. Accordingly, the 2,6diethynylhydrindacene (Z)-1 was expected to function as a suitable monomer unit for the construction of macrocycles with directional-persistent peripheral functionality. The macrocycles were prepared successfully through Eglinton coupling of 2,6diethynylhydrindacenes 1a­1d. Macrocyclization of the mono-

© 2015 The Chemical Society of Japan | 401

Scheme 5. Expected complexation patterns for hydrindacene-based allosteric receptors.

mer (Z)-1 (5 or 10 mM) under modified Eglinton coupling conditions using Cu(OAc)2 in MeCN produced a series of macrocycles, including dimer 2 (39­64%), trimer 3 (13­35%), and tetramer 4 (5­12%), as well as higher oligomers (8­13%) in good total yields (77­91%).27 X-ray analyses of macrocycles 2a, 2b, 3a, 3b, and 3c provided unambiguous confirmation of their structures and demonstrated a well-ordered array of functionalities that consisted of peripheral ester groups (X = CO2R) along with Z-substituents (Z = MeO, Br, styryl, or thienyl) vertical to the macrocyclic ring (Figure 2). The resulting cavity in 2 was too small to accept a guest solvent molecule, while the two hydrindacene units in 2 interacted electronically through the butadiyne moieties, as shown by the UV­vis spectra, cyclic voltammograms, and theoretical calculations. Trimer 3 can capture one or more solvent molecules, such as hexane, CHCl3, or cyclohexane, within the rectangular cavity in the solid state. Tetramer 4 has a tub-shaped structure, in which all butadiyne moieties are located at the pseudoaxial position on the cyclopentene rings (Figure 2d). Four molecules of CH2Cl2 were trapped in the cavity of the tub (not shown in the figure for clarity). Furthermore, the peripheral ester groups of these macrocycles could be modified to introduce hydrophobic or hydrophilic chains through the use of Otera’s distannoxane catalyst.28 Adjusting the solubility of some of these modified macrocycles allows the subsequent assembly into tube-like structures. The introduction of π-conjugated groups, such as styryl and thienyl groups (as in 2c, 2d and 3c, 3d), or hydrogen-bonding groups such as amides, to the 4,8-positions of the hydrindacene unit provides the possibility of π-conjugated or self-assembled nanotubes. Further studies are underway to assemble these macrocycles into tubular structures. 5. Hydrindacene-Based Allosteric Receptors The geometry of (Z)-isomers of hydrindacenes is suitable for the construction of concave molecular receptors (endoreceptors),23 whereas that of (E)-isomers is suitable for multivalent receptors (exoreceptors).24 In addition, amides as hydrogen-bonding functional groups can form cooperative hydrogen-bonding networks due to the dual ability to act as a hydrogen-bonding donor (NH side), hydrogen-bonding acceptor (C=O side), or both.29 Thus, the 402 | Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

4,8-hydrindacanedicarboxamide derivative 5 was designed as a multifaceted receptor, in which amide groups at the 4,8positions of hydrindacenes were twisted against the plane of the hydrindacene scaffold and functioned as both hydrogenbond donors and hydrogen-bond acceptors. In receptor 5, the syn- and anti-conformers were converted easily by rotation around the Caromatic­Camide bonds without a guest molecule (Scheme 5). The anti- and syn-conformers possessed two Hbonding sites on both sides of the molecular plane (DA/DA for the anti-, and AA/DD for the syn-conformer). Complexes of receptor 5 with guest molecules could induce two types of allosterism.30 For guests with a DA-type of H-bonding, receptor 5 adopted an anti-conformation and complexed with two guests in a positive homotropic allosteric manner, with the first guest enhancing the binding strength of the receptor toward a subsequent guest. In contrast, toward DD- or AA-type guests, receptor 5 adopted a syn-conformation and complexed with the guest on one side of the receptor. The complex could bind successively with another guest having opposite Hbonding functionality in a positive heterotropic allosteric manner or could inhibit binding of the guest with mismatched H-bonding functionality, such as a DA-type guest in a negative heterotropic allosteric manner. 5.1 Positive Homotropic Allosteric Receptors of Benzenediols. The cooperativity of amide hydrogen bonds plays an important role in stabilizing folded protein structures and in producing allosterism in natural products.31 The origin of this cooperativity has been proposed to be the restriction of internal motion (entropic contribution) as well as inductive polarization (enthalpic contribution) of the amide groups upon hydrogenbond formation. Thus, 4,8-hydrindacanedicarboxamide receptors 5a and 5b were designed and prepared according to these design concepts.22 The 4,8-hydrindacanedicarboxamides 5a and 5b behaved as an exoditopic receptor, which exhibits a positive homotropic allosteric binding process toward benzenediols and forms 1:2 complexes with resorcinols and catechol with a cooperative factor K2/K1 = 3­33 in CDCl3 (Scheme 6 and Table 1). For example, the microscopic binding constants K1 and K2 were 116 and 2800 M¹1 (K2/K1 = 24), respectively, for 1:2 complexation of 5b with resorcinol. High allostericity (K2/K1 = 12) also was observed for complexation of 5a and resorcinol, although the bulkiness of the trityl groups

© 2015 The Chemical Society of Japan

Scheme 6. Proposed mechanism for the positive homotropic allosterism (K1 < K2) observed in 1:2 complexation of receptor 5 with resorcinols. Table 1. Microscopic Binding Constants K (M¹1) and Cooperativity Parameters K2/K1 for Complexation of Receptors 5a and 5b with Guests in CDCl3 at 25 °C Complex 5a¢(resorcinol)2 5a¢(5-chlororesorcinol)2 5a¢(catechol)2 5a¢(3-hydroxybenzyl alcohol)2 5a¢(phenol)2

R¤ H Cl ® ® ®

5b¢(resorcinol)2 5b¢(5-methylresorcinol)2 5b¢(5-methoxyresorcinol)2 5b¢(5-chlororesorcinol)2 5b¢(catechol)2 5b¢(3-hydroxybenzyl alcohol)2 5b¢(2-hydroxybenzyl alcohol)2

H Me OMe Cl ® ® ®

·m 0.00 0.37 ® ® ® 0.00 ¹0.07 0.12 0.37 ® ® ®

K1/M¹1 43 83 49 18 ca. 1

K2/M¹1 520 1160 112 68 ca. 1

116 92 100 152 50 40 13

2800 2400 3000 5000 120 120 12

K2/K1 12 14 ca. 3 4 ca. 1 24 26 30 33 ca. 3 3 ca. 1

a) Estimated errors are within 15%. b) Estimated errors are within 30%.

Figure 3. X-ray structures of 5a¢(resorcinol)2.

in 5a decreased K1 and K2. Similarly, large values for K2/ K1 = 26­33 were observed for the combination of 5b and 5-substituted resorcinols. Other than resorcinol, 5a and 5b formed complexes with catechol and 3-hydroxybenzyl alcohol, although the cooperativity was weak (K2/K1 = 4). No meaningful associations with hydroquinone or phenol were observed for 5a and 5b. The large selectivity for complexation toward aromatic diols containing similar functionality suggests that the degree of cooperativity (K2/K1) and the affinity of individual functionalities are important for the recognition properties of 5. To obtain information on the manner of allosteric binding, a series of receptors and their complexes were examined using crystallographic structural analyses and titration studies Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

(Figure 3). Results revealed that the observed positive allostericity was due to rotational restriction around the amide bond (¦S), which induced polarization in the electronic structure of the amide (¦H). Titration experiments for 5b at several temperatures revealed that the degree of allostericity (K2/K1) depended on temperature (¦¦H2-1 = ¹4.6 « 0.7 kJ mol¹1, ¦¦S2-1 = +11.9 « 2.4 J mol¹1 K¹1), indicating the importance of the contribution of the enthalpy term as well as the entropy term. Crystallographic studies revealed that both the amide C=O and C­N bonds synergistically changed bond length and simultaneously polarized upon hydrogen-bond formation at the either side of the amide NH or C=O. These changes were supported by results showing that the cooperative factor K2/K1 © 2015 The Chemical Society of Japan | 403

ca. Scheme 7. Multipoint recognition of catecholamines by hydrindacene-based receptor 5 accompanied by complexation-induced conformational switching.

for complexation of 5b with 5-substituted resorcinols correlated with the Hammett ·m values of substituents on the resorcinol [Me: ·m ¹0.07, K2/K1 = 26; H: ·m 0, K2/K1 = 24; MeO: ·m 0.12, K2/K1 = 30; Cl: ·m 0.37, K2/K1 = 33]. Thus, as the acidity of the resorcinol hydroxy groups increased, the amide groups became more polarized, which promoted subsequent binding with the second guest. These results provide the first clear evidence that cooperativity in amide hydrogen bonding by polarization32 leads to positive homotropic allosteric binding. 5.2 Functionality Selective Catecholamine Receptors. Although the anti-conformation of 5 is preferred in terms of polarization of the amide groups (Scheme 5), a structural change to the syn-conformer could be induced by multipoint hydrogen-bond formation upon complex formation with catecholamines (Scheme 7). The receptors 5b and 5c selectively bound adrenaline and dopamine salts at a receptor-guest ratio of 1:1 in 2% CD3CN­ CDCl3.23 The complexation-induced chemical shifts (CISs) derived from NMR titration of the catecholamine salt with 5b or 5c revealed guest binding based on multipoint hydrogen bonding via the ammonium group and 3-hydroxy group on the aromatic ring of the guest. With adrenaline, an additional hydrogen bond with a benzylic hydroxy group was formed. The adrenaline molecule was located above the hydrindacene, which adopted a syn-conformation for the two amide groups. The association constants toward adrenaline and dopamine were on the order of 104 M¹1, which is much larger than that with guests without a 3-hydroxy group (103 M¹1). In the absence of a guest molecule, the syn-conformer of 5c was less stable than the anti-conformer (anti:syn = 85:15 in CDCl3). Upon complex formation with adrenaline, the syn-conformer became dominant. Interestingly, the complexation of 5c with a phenethylamine salt also induced a conformational change to the syn-conformer, even though the guest did not contain a 3hydroxy group, indicating that conformational switching was caused not only by multiple hydrogen bonds, but also through a dipole-reversal effect (Scheme 7). Rotation of the amide group 404 | Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

on one side of the anti-conformer partially offset the reversed dipole generated by hydrogen bonding between the ammonium and amide groups on the other side. During complexation of 5c with adrenaline or dopamine, such conformational switching accompanied by binding of the guest ammonium group has the advantage of preorganizing the amide group for recognition of the 3-hydroxy group. 5.3 Allosteric Inhibitor Modulating from 1:2 Allosteric Binding to 1:1 Binding of Resorcinols. This conformational change to the syn-conformation accompanying 1:1 complexation of receptor 5 with a dopamine salt was applied to inhibit 1:2 allosteric binding with resorcinols (Scheme 8).22b Upon addition of 5-chlororesorcinol to the 1:1 complex with receptor 5b and dopamine salt in CDCl3, 5-chlororesorcinol weakly bound to the receptor at a 1:1 ratio with a binding constant of 160 M¹1. NMR titration showed a typical hyperbolic curve, instead of the positive allosteric 1:2 binding of resorcinol found without dopamine, which produced a sigmoid curve (Figure 4). Furthermore, upon titration of 5-chlororesorcinol to a 1:0.5 mixture of receptor 5b and dopamine salt, 5-chlororesorcinol was bound and showed a weak allosteric effect at a degree of saturation ratio of 1.5, compared to a degree of saturation ratio of 2.0 without dopamine. These results indicate that release of resorcinol from the 1:2 complex was induced by addition of dopamine as an allosteric inhibitor. Thus, the present system reveals that allosteric modulation by the effector molecule leads to the cooperative release of the bound guests. The hydrindacenediamide receptor has exhibited dual functions as a homotropic allosteric receptor toward resorcinols and a heterotropic allosteric receptor toward catecholamines. The self-assembling positive homotropic allosteric receptors are candidates for cooperative supramolecular polymers through a nucleation­elongation process.3b,33 Furthermore, the introduction of oligopeptide chains and hydrophilic side-chains to the 2,6-positions of hydrindacenediamides may allow specific recognition toward peptide sequences or bioactive compounds in aqueous solution.34 Further work on these approaches is currently underway.

© 2015 The Chemical Society of Japan

Scheme 8. Regulation of allosteric binding of receptor 5b with 5-chlororesorcinol by addition of the “allosteric effector” dopamine salt.

(a) no dopamine salt


(b) (0.5 equiv)

(c) dopamine salts (1 equiv)


Figure 4. NMR titration curves for 5b with 5-chlororesorcinol in CDCl3 at 25 °C with [5b] and [dopamine] of (a) 1.0 and 0 mM; (b) 1.0 and 0.5 mM; (c) 1.0 and 1.0 mM.

6. Hydrindacene-Based Rotaxane Systems In hydrindacene scaffolds, X and Y substituents at the 2,6positions are perpendicular. The less bulky substituent (Y) at the 2,6-positions in the (E)-isomer preferentially occupied a pseudoaxial position at opposing sides of the molecular plane. Thus, the scaffold also can function as a guest (i.e., an axle for rotaxanes) toward an appropriately designed macrocycle and can lead to novel rotaxanes by attaching endcaps to the substituents (X) occupying pseudo-equatorial positions at the 2,6positions. This was realized using dynamic imine-bond formation35 between carbaldehyde groups on the hydrindacene axle

and amino groups on the macrocycle (Scheme 9).24 In addition, the entropy-driven dynamic covalent character of imine bonds provided this rotaxane system with thermo- and pH-responsive fixation/movement controls. 6.1 First-Generation Imine-Bridged Rotaxanes. The bromo-terminated dicarbaldehyde axle 6a and TBSO-terminated axle 6b were prepared from 1,2,4,5-tetrakis(bromomethyl)benzene. Macrocyclic diamines 7a and 7b were prepared from 2,6-diarylaniline through Eglinton coupling. Threaded imines 8a­8c were prepared by threading and the formation of an imine bond between axles 6a or 6b and macrocycles 7a and 7b under acidic dehydrating conditions (i.e., in benzene at reflux with 4A MS with addition of trifluoroacetic acid) (Scheme 10).24a,24e Xray analysis of imine 8a with a bromo-terminated hydrindacene unit unambiguously revealed that the hydrindacene axle passed through the macrocycle. The threaded imines 8a, 8b, and 8c could be hydrolyzed easily to starting materials 6a, 6b and 7a, 7b, whereas the imine-bridged rotaxane 9a with a hexadiyne-type macrocycle derived from 8b did not produce the hydrolyzed [2]rotaxane 10a under similar acidic hydrolytic conditions (Scheme 11). The hydrolyzed [2]rotaxane 10b with a hexamethylene-type macrocycle was generated from imine-bridged rotaxane 9b as a minor product along with mono-imine 11b in a 9b/11b/ 10b ratio of 72/24/4 at 293 K. These results suggest that the flexibility of the macrocycle was an important factor for entropy gain in the hydrolyzed [2]rotaxane 10b, otherwise the imine-bridged form, such as the hexadiyne-type 9a, was heavily preferred due to release of two water molecules

Scheme 9. Preparation and interconversion of imine-bridged rotaxane and [2]rotaxane. Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

© 2015 The Chemical Society of Japan | 405

Scheme 10. Preparation of threaded imine 8a­8c and the X-ray structure of 8a.

Scheme 11. Interconversion between imine-bridged rotaxane 9 and [2]rotaxane 10.

accompanying imine-bond formation. This result was confirmed by the increase in population of the [2]rotaxane 10b as the temperature was lowered (9b/11b/10b = 72/24/4 at 293 K and 42/36/22 at 233 K), indicating that imine-bond hydrolysis was an enthalpically favored process, whereas imine-bond formation was an entropically favored process in this rotaxane system. 6.2 Second-Generation Imine-Bridged Rotaxanes with Novel imine-bridged Efficient Switching Capability. molecular shuttles that can be completely transformed to [2]rotaxane under hydrolytic conditions are desirable because the first-generation rotaxanes produce a low equilibrium ratio of [2]rotaxane under acidic hydrolytic conditions. Thus, iminebond cleavage under dynamic acidic hydrolytic conditions was expected to be promoted by the formation of hydrogen bonds of the resulting primary ammoniums (NH3+) with newly attached hydrogen-bond acceptors. The second-generation 406 | Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

novel imine-bridged molecular shuttle 9c attached to a triethylene glycol (TEG)-station was designed and prepared. This shuttle could be completely transformed to hydrolyzed [2]rotaxane 10c under hydrolytic conditions due to hydrogenbond formation between the macrocycle and TEG-station.24b Hydrolysis of the imine-bonds of the imine-bridged molecular shuttle 9c with TEG-stations gave the [2]rotaxane 10c exclusively with the macrocycle hydrogen-bonded to the TEGstation (Figure 5). Equilibrium between 9c and 10c, which controlled the position of the macrocycle, could be switched to either side by applying acidic hydrolytic or dehydrating conditions. Furthermore, the equilibrium was biased toward [2]rotaxane 10c under acidic hydrolytic conditions and could be reversed in favor of bis-imine 9c upon heating. For thermodynamic contributions, imine-bonding and hydrogen-bonding have to act synergistically for entropy-driven positional switching of the macrocycle in a rotaxane structure © 2015 The Chemical Society of Japan

Figure 5. 1H NMR spectra (300 MHz, wet CDCl3) of imine-bridged rotaxane 9c and [2]rotaxane 10c generated upon addition of TFA to a solution of 9c at 298 K.

Scheme 12. Thermodynamic interplay of imine- and hydrogen-bonding in a rotaxane-based molecular shuttle exhibiting entropydriven translational isomerism. Right: temperature dependence of the equilibrium ratio of 9c and 10c in the hydrolyzed mixture.

(Scheme 12): hydrogen-bond formation is an enthalpically favored interaction accompanied by a loss of entropy, whereas imine-bond formation is an entropically favored process in the rotaxane system described here. Thus, imine-bond cleavage and hydrogen-bond formation are thermodynamically matched processes (i.e., enthalpically favored and entropically disfavored processes). The reverse processes of hydrogen-bond cleavage and imine bonding also are matched processes (i.e., enthalpically disfavored and entropically favored processes). The van’t Hoff plots for imine hydrolysis of 9b to [2]rotaxane 10b revealed that the enthalpy and entropy differences between 9b and 10b were both negative (¦H = ¹6.0 « 0.4 kcal mol¹1 and ¦S = ¹26.0 « 1.4 cal mol¹1 K¹1), confirming that hydrolysis of the imine bonds was an enthalpically favored and entropically disfavored process. Furthermore, the changes in both enthalpy and entropy from 9c to 10c (¦H = ¹16.7 « 0.7 kcal mol¹1 and ¦S = ¹52.2 « 2.1 cal mol¹1 K¹1) were larger than those in the change from 9b to 10b. This large difference in the enthalpy and entropy changes originates from the incorporation of a TEG-station in place of an XYL-station (i.e., the additional enthalpic stabilization and loss of entropy due to hydrogen bonding of the macrocycle with the TEG station in 10b). Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

Consequently, incorporation of a TEG station with hydrogen-bonding ability into imine-bridged rotaxanes not only biased the equilibrium between 9b and 10b under acidic hydrolytic conditions in favor of hydrolyzed [2]rotaxane, but also enabled entropy-driven positional switching of the macrocycle by supplying additional enthalpic and entropic changes. Furthermore, imine-bond hydrolyses and the formation of hydrogen bonds between the macrocycle and station were thermodynamically matched processes, because both processes were enthalpically favored and accompanied by a loss of entropy. Conclusion The s-hydrindacene molecule was used as a versatile scaffold for the synthesis of supramolecular assemblies, such as macrocycles with directionally persistent peripheral functionalities, allosteric receptors, adrenaline receptors, and rotaxane molecular shuttles. These assemblies resulted from the multidimensional modification of hydrindacene based on addition of substituents X, Y, and Z, all of which were mutually perpendicular. This predictable structural feature will facilitate the design of supramolecular assemblies. In addition, the featureless physical character of the hydrindacene skeleton itself © 2015 The Chemical Society of Japan | 407

allows the unique properties of the assemblies to take precedence. The hydrindacenediamide receptors clearly demonstrated cooperativity for amide hydrogen-bond formation, induced polarity, and dipole reversal effect upon cation binding. In the imine-bridged rotaxane, the combination of synergistic imine bonding and hydrogen bonding allowed entropy-driven positional switching due to thermodynamically matched processes. Applications of hydrindacenes as a versatile platform for the development of novel supramolecular assemblies are being explored further to develop novel assembly methods for supramolecular chemistry. The author expresses his sincerest appreciation to Prof. Takanori Suzuki and Prof. Emeritus Takashi Tsuji (Hokkaido University) for their continuous support and guidance. The author thanks all co-workers and collaborators listed in the references, in particular Dr. Ryo Katoono, Dr. Takeshi Umehara, and Dr. Hiroyoshi Sugino for their considerable experimental contributions in this work. This work was supported by Grants-in-Aid for Scientific Research for Young Scientists (B) (Nos. 16750024, 18750024, and 20750024) and (A) (No. 24685008), and by PRESTO, Japan Science and Technology Agency (JST). References # Dedicated to Professor Takashi Tsuji on the occasion of his 75th birthday. 1 a) Supramolecular Chemistry: From Molecules to Nanomaterials, ed. by P. A. Gale, J. W. Steed, Wiley-VCH, Weinheim, 2012. doi:10.1002/9780470661345. b) J.-M. Lehn, Supramolecular Chemistry: Concepts and Perspectives, Wiley-VCH, Weinheim, 1995. doi:10.1002/3527607439. 2 a) E. Elacqua, D. S. Lye, M. Weck, Acc. Chem. Res. 2014, 47, 2405. b) Z. Dong, Q. Luo, J. Liu, Chem. Soc. Rev. 2012, 41, 7890. c) L. J. Prins, D. N. Reinhoudt, P. Timmerman, Angew. Chem., Int. Ed. 2001, 40, 2382. d) D. Philp, J. F. Stoddart, Angew. Chem., Int. Ed. Engl. 1996, 35, 1154. e) D. S. Lawrence, T. Jiang, M. Levett, Chem. Rev. 1995, 95, 2229. 3 a) R. P. Sijbesma, E. W. Meijer, Chem. Coummun. 2003, 5. b) T. F. A. De Greef, M. M. J. Smulders, M. Wolffs, A. P. H. J. Schenning, R. P. Sijbesma, E. W. Meijer, Chem. Rev. 2009, 109, 5687. 4 S. Cantekin, T. F. A. de Greef, A. R. A. Palmans, Chem. Soc. Rev. 2012, 41, 6125. 5 a) B. Dolenský, M. Havlík, V. Král, Chem. Soc. Rev. 2012, 41, 3839. b) Ö. V. Rúnarsson, J. Artacho, K. Wärnmark, Eur. J. Org. Chem. 2012, 7015. 6 J. Wu, W. Pisula, K. Müllen, Chem. Rev. 2007, 107, 718. 7 a) F. Würthner, Pure Appl. Chem. 2006, 78, 2341. b) T. Seki, X. Lin, S. Yagai, Asian J. Org. Chem. 2013, 2, 708. 8 a) K. Kim, N. Selvapalam, Y. H. Ko, K. M. Park, D. Kim, J. Kim, Chem. Soc. Rev. 2007, 36, 267. b) E. Masson, X. Ling, R. Joseph, L. Kyeremeh-Mensah, X. Lu, RSC Adv. 2012, 2, 1213. 9 a) T. Ogoshi, J. Inclusion Phenom. Macrocyclic Chem. 2012, 72, 247. b) M. Xue, Y. Yang, X. Chi, Z. Zhang, F. Huang, Acc. Chem. Res. 2012, 45, 1294. c) H. Zhang, Y. Zhao, Chem.® Eur. J. 2013, 19, 16862. d) N. L. Strutt, H. Zhang, S. T. Schneebeli, J. F. Stoddart, Acc. Chem. Res. 2014, 47, 2631. 10 H. Hopf, Classics in Hydrocarbon Chemistry, Wiley-VCH, Weinheim, 2000. 408 | Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

11 a) Carbon-Rich Compounds: From Molecules to Materials, ed. by M. M. Haley, R. R. Tykwinski, Wiley-VCH, Weinheim, 2006. doi:10.1002/3527607994. b) Y.-T. Wu, J. S. Siegel, Chem. Rev. 2006, 106, 4843. c) F. Diederich, M. Kivala, Adv. Mater. 2010, 22, 803. 12 a) G. Hennrich, E. V. Anslyn, Chem.®Eur. J. 2002, 8, 2218. b) X. Wang, F. Hof, Beilstein J. Org. Chem. 2012, 8, 1. 13 a) N. Pannier, W. Maison, Eur. J. Org. Chem. 2008, 1278. b) M. Tominaga, K. Katagiri, I. Azumaya, CrystEngComm 2010, 12, 1164. 14 a) J. H. Chong, M. J. MacLachlan, Chem. Soc. Rev. 2009, 38, 3301. b) T. M. Swager, Acc. Chem. Res. 2008, 41, 1181. 15 a) F.-G. Klärner, B. Kahlert, Acc. Chem. Res. 2003, 36, 919. b) F.-G. Klärner, T. Schrader, Acc. Chem. Res. 2013, 46, 967. 16 a) M. Freund, K. Fleischer, E. Gofferjé, Justus Liebigs Ann. Chem. 1918, 414, 12. b) R. T. Arnold, R. A. Barnes, J. Am. Chem. Soc. 1944, 66, 960. 17 a) S. Kotha, Acc. Chem. Res. 2003, 36, 342. b) R. A. Weatherhead, M. D. Carducci, E. A. Mash, J. Org. Chem. 2009, 74, 8773. c) S. Kotha, E. Brahmachary, J. Org. Chem. 2000, 65, 1359. 18 a) T. Matsuo, M. Kobayashi, K. Tamao, Dalton Trans. 2010, 39, 9203. b) J. Nishigaki, R. Tsunoyama, H. Tsunoyama, N. Ichikuni, S. Yamazoe, Y. Negishi, M. Ito, T. Matsuo, K. Tamao, T. Tsukuda, J. Am. Chem. Soc. 2012, 134, 14295. c) T. Matsuo, K. Suzuki, T. Fukawa, B. Li, M. Ito, Y. Shoji, T. Otani, L. Li, M. Kobayashi, M. Hachiya, Y. Tahara, D. Hashizume, T. Fukunaga, A. Fukazawa, Y. Li, H. Tsuji, K. Tamao, Bull. Chem. Soc. Jpn. 2011, 84, 1178. 19 a) P. Wessig, K. Möllnitz, J. Org. Chem. 2012, 77, 3907. b) J. Barańska, J. Grochowski, J. Jamrozik, P. Serda, Org. Lett. 2000, 2, 425. 20 a) S. Ito, H. Takata, K. Ono, N. Iwasawa, Angew. Chem., Int. Ed. 2013, 52, 11045. b) S. Ito, K. Ono, N. Iwasawa, J. Am. Chem. Soc. 2012, 134, 13962. c) N. Iwasawa, H. Takahagi, J. Am. Chem. Soc. 2007, 129, 7754. d) H. Sakurai, N. Iwasawa, K. Narasaka, Bull. Chem. Soc. Jpn. 1996, 69, 2585. 21 H. Kawai, T. Utamura, E. Motoi, T. Takahashi, H. Sugino, M. Tamura, M. Ohkita, K. Fujiwara, T. Saito, T. Tsuji, T. Suzuki, Chem.®Eur. J. 2013, 19, 4513. 22 a) H. Kawai, R. Katoono, K. Nishimura, S. Matsuda, K. Fujiwara, T. Tsuji, T. Suzuki, J. Am. Chem. Soc. 2004, 126, 5034. b) H. Kawai, J. Synth. Org. Chem., Jpn. 2007, 65, 677. 23 H. Kawai, R. Katoono, K. Fujiwara, T. Tsuji, T. Suzuki, Chem.®Eur. J. 2005, 11, 815. 24 a) H. Kawai, T. Umehara, K. Fujiwara, T. Tsuji, T. Suzuki, Angew. Chem., Int. Ed. 2006, 45, 4281. b) T. Umehara, H. Kawai, K. Fujiwara, T. Suzuki, J. Am. Chem. Soc. 2008, 130, 13981. c) H. Sugino, H. Kawai, K. Fujiwara, T. Suzuki, Kobunshi Ronbunshu 2011, 68, 795. d) H. Sugino, H. Kawai, K. Fujiwara, T. Suzuki, Chem. Lett. 2012, 41, 79. e) H. Sugino, H. Kawai, T. Umehara, K. Fujiwara, T. Suzuki, Chem.®Eur. J. 2012, 18, 13722. 25 a) P. Holý, M. Havránek, M. Pánková, L. Ridvan, J. Závada, Tetrahedron 1997, 53, 8195. b) L. Ridvan, J. Závada, Tetrahedron 1997, 53, 14793. 26 a) G. Eck, M. Julia, B. Pfeiffer, C. Rolando, Tetrahedron Lett. 1985, 26, 4725. b) W. C. Wong, C. Gluchowski, Synthesis 1995, 139. 27 The calculated energies per diethynylhydrindacene unit for the unsubstituted dimer, trimer, and tetramer (+1.3, +0.0, and 1.0 kcal mol¹1, respectively) suggested that the trimer form was the most stable and the product distribution was kinetically controlled

© 2015 The Chemical Society of Japan

under pseudo-high dilution conditions. 28 J. Otera, Chem. Rev. 1993, 93, 1449. 29 a) G. T. R. Palmore, J. C. Macdonald, in The Amide Linkage, ed. by A. Greenberg, C. M. Breneman, J. F. Liebman, Wiley-Interscience, Hoboken, 2003, Chap. 10. b) J. C. MacDonald, G. M. Whitesides, Chem. Rev. 1994, 94, 2383. c) H. Kawai, D. Hosoda, CrystEngComm 2012, 14, 5717. 30 a) S. Shinkai, M. Ikeda, A. Sugasaki, M. Takeuchi, Acc. Chem. Res. 2001, 34, 494. b) M. Takeuchi, M. Ikeda, A. Sugasaki, S. Shinkai, Acc. Chem. Res. 2001, 34, 865. c) L. Kovbasyuk, R. Krämer, Chem. Rev. 2004, 104, 3161. d) S. Shinkai, M. Takeuchi, Bull. Chem. Soc. Jpn. 2005, 78, 40. e) C. Kremer, A. Lützen, Chem.®Eur. J. 2013, 19, 6162. 31 a) D. H. Williams, E. Stephens, M. Zhou, Chem. Commun.

2003, 1973. b) D. H. Williams, E. Stephens, D. P. O’Brien, M. Zhou, Angew. Chem., Int. Ed. 2004, 43, 6596. 32 a) B. W. Gung, Z. Zhu, B. Everingham, J. Org. Chem. 1997, 62, 3436. b) R. Ludwig, F. Weinhold, T. C. Farrar, J. Phys. Chem. A 1997, 101, 8861. c) A. P. Bisson, C. A. Hunter, J. C. Morales, K. Young, Chem.®Eur. J. 1998, 4, 845. 33 D. Zhao, J. S. Moore, Org. Biomol. Chem. 2003, 1, 3471. 34 a) B. E. Collins, E. V. Anslyn, Chem.®Eur. J. 2007, 13, 4700. b) H. Wennemers, M. Conza, M. Nold, P. Krattiger, Chem.® Eur. J. 2001, 7, 3342. 35 a) C. D. Meyer, C. S. Joiner, J. F. Stoddart, Chem. Soc. Rev. 2007, 36, 1705. b) P. C. Haussmann, J. F. Stoddart, Chem. Rec. 2009, 9, 136. c) M. E. Belowich, J. F. Stoddart, Chem. Soc. Rev. 2012, 41, 2003.

Hidetoshi Kawai, Associate Professor at Tokyo University of Science, was born in Sapporo, Japan, in 1972. He graduated from Hokkaido University in 1995 and received his Ph.D. in 2000 under the supervision of Professor Takashi Tsuji, focusing on the synthesis and kinetic stabilization of [1.1]paracyclophanes. He worked at Hokkaido University as a Research Fellow of the Japan Society for the Promotion of Science (JSPS) for 1998­1999 and as a Research Associate for 2000­2011. In 2007­2011, he was a researcher of PRESTO, Japan Science and Technology Agency (JST). In 2011, he moved to Faculty of Science, Tokyo University of Science as an Associate Professor. His research interests include highly strained molecules and self-assembly based on cooperative hydrogen bonding and dynamic covalent bonds.

Bull. Chem. Soc. Jpn. 2015, 88, 399–409 | doi:10.1246/bcsj.20140313

© 2015 The Chemical Society of Japan | 409

Suggest Documents