Multi-Responsive Molecular Encapsulation and Release Based on Hydrogen-Bonded Azo-Macrocycle

Research on stimuli-responsive host–guest systems is at the cutting edge of supramolecular chemistry, owing to their numerous potential applications such as catalysis, molecular machines, and drug delivery. Herein, we present a multi-responsive host–guest system comprising azo-macrocycle 1 and 4,4′-bipyridinium salt G1 for pH-, photo-, and cation- responsiveness. Previously, we reported a novel hydrogen-bonded azo-macrocycle 1. The size of this host can be controlled through light-induced E↔Z photo-isomerization of the constituent azo-benzenes. The host is found in this work to be capable of forming stable complexes with bipyridinium/pyridinium salts, and implementing guest capture and release with G1 under light in a controlled manner. The binding and release of the guest in the complexes can also be easily controlled reversibly by using acid and base. Moreover, the cation competition-induced dissociation of the complex 1a2⊃G1 is achieved. These findings are expected to be useful in regulating encapsulation for sophisticated supramolecular systems.


Introduction
Uncovering the mystery of biomolecules such as proteins for controlling the capture and release process has aroused extensive research interests [1,2]. However, the complexity of natural macrocycles renders chemists to design simple artificial systems for exploring the mechanism behind them [3]. In this regard, macrocycle-based host-guest systems are considered as one of the options that can fulfill this task more or less in functioning as encapsulation and release modules [4]. Multiple-responsive host-guest systems are particularly interesting in terms of their versatility in applications under various operative conditions with the aid of acid/base [5][6][7], light [8,9], and competitive guests [10,11]. Macrocyclic hosts, one of the components for such modular systems, have emerged in increasing numbers in recent years as recognition motifs for guest binding and for guest capture and release [12] owing to their easy synthesis and programmable, adjustable cavity size and binding affinity. Typical examples include cyclodextrin [13,14], crown ether [15][16][17], calixarene [18][19][20], pillararene [5,[21][22][23][24], and other macrocycles [25,26], which are widely utilized for constructing multiple-responsive host-guest systems [27][28][29][30][31][32] to implement binding and releasing functions.
There are mainly three types of host-guest-responsive systems used in the field of supramolecular chemistry according to external stimuli: (a) one relying on guest molecules that are switched between capture and release according to chemical actions (e.g., pH switching and substrate competition), (b) the second type accomplishing the capture and release process using light irradiation-a clean, easily manipulated approach-for controlling complexation, and (c) the third type modulating the host-guest binding ability using electrochemical methods to change the charge and electron distributions by implementing electron transfer reactions [33]. Most of the known light-switchable azo-benzene systems Scheme 1. (a) Schematic representation of the multi-responsive encapsulation and/or release of 4,4 -bipyridinium salts by azo-macrocycle 1, through photo-, acid/base-, and cation-competitive complexation-controlled process. (b) Chemical structures of azo-macrocycle 1 and guests G1-G6.

Host-Guest Complexation
Hydrogen-bonded aramide macrocycle 1 used in the present work adopts a nearly planar conformation with six carbonyl oxygen atoms pointing inward as recognition sites. The cavity size measures 8.3 Å in diameter, which renders it suitable for enclosing such organic cations as pyridinium or bipyridinium (G1-G5) (4.1 Å in width) [68]. The host-guest complexation in solution was studied using 1 H NMR, HRMS, and UV-vis spectroscopy.
First, we set out to examine the complexation between host 1a and guest G1 in solution by 1 H NMR spectroscopy. At room temperature, the 1 H NMR spectra of a mixture of host 1a and guest G1 recorded in CDCl 3 -CD 3 CN show signals corresponding to the formation of a host-guest complex ( Figure 1 and Figure S3). The signals of protons H 1 and H 2 in G1 experience distinct downfield shifts, especially for the proton H 2 with a shift as large as ca. 0.5 ppm at a 2:1 (H:G) molar ratio (Figure 1c) with respect to free guest (Figure 1a). Substantial upfield shifts for the aromatic protons c and d of the complex relative to the macrocycle (Figure 1c,d) were also observed. The signal evolution of these protons in the presence of 1 equiv. and 2 equiv. 1a suggests the production of the complex 1a 2 ⊃G1. The Job plot experiment indicates the formation of a 2:1 complex ( Figure S4), which is in accordance with the results from positive-ion electrospray ionization mass spectrometry (ESI-MS, Figure S5). UV-vis titration experiments in CHCl 3 -CH 3 CN (5:1, v/v) provided binding constants of K 1 = 6.55 × 10 4 M −1 , K 2 = 7.20 × 10 4 M −1 , and K 12 = 4.72 × 10 9 M −2 for 1a 2 ⊃G1 (Figures S7 and S8). Notably, only a single set of signals from the bipyridinium ion was detected in the presence of 1.0 equiv. of 1a, indicating a fast exchange on the NMR time scale for the complex formation ( Figure 1b). When using G2 with PF 6 − as a different counterion to replace CF 3 COO − , a similar trend was observed (Figures S9-S13). The Job plot experiments provided a 2:1 stoichiometry for 1b 2 ⊃G2 complexes in solution (Figures S11 and S12), which are all consistent with the results from the ESI-MS spectra ( Figure S13). The binding constants of K 1 = 5.63 × 10 4 M −1 , K 2 = 8.24 × 10 3 M −1 , and K 12 = 4.64 × 10 8 M −2 (G2) were obtained in CHCl 3 -CH 3 CN (1:1, v/v) due to the low solubility in CHCl 3 -CH 3 CN (5:1, v/v) by fitting the experimental data from the UV-vis spectroscopy (Figures S14 and S15). We then studied the complexation ability of 1a with 2,2 -bipyridinium salt G3, a structural isomer of G1/G2. G3 with 1a also exhibited a fast exchange process on the NMR time scale ( Figure S16). According to the results of the Job plot experiment (Figures S19 and S20), a host-guest complex with 2:1 stoichiometry was formed, which is in agreement with the HR-MS results ( Figure S25). The association constants for 1a 2 ⊃G3 were found to be K 1 = 9.91 × 10 5 M −1 , K 2 = 1.02 × 10 5 M −1 , and K 12 = 1.01 × 10 11 M −2 in CHCl 3 -CH 3 CN (5:1, v/v) ( Figures S28 and S29). Since π-π stacking interactions play an important role along with multiple C-H···O hydrogen-bonding and/or N + ···O ion-dipole interactions to form pseudorotaxanes [62], G4/G5 with a single positive charge was selected to test if 2:1 stoichiometry could be retained under the same condition. G4, 4-phenylpyridinium salt, differs from G5, 2-phenylpyridinium salt, only in the position of nitrogen atoms. In the 1 H NMR spectra, significant chemical shifts of 1a and G4/G5 could be observed, indicating the formation of host-guest complexes (Figures S17 and S18). The Job plot experiments only disclose the presence of 1:1 complexes between 1a and G4/G5 in solution ( Figures S21-S24). HR-MS studies provided further evidence for the formation of the 1:1 complexes between 1a and G4/G5 ( Figures S26 and S27), but no peaks relating to 2:1 species were observed. The binding constants for the 1:1 complex between 1a and guest G4/G5 were calculated to be K a = 5.50 × 10 3 M −1 (G4) and K a = 4.02 × 10 4 The complexation of G1 and 1a was probed by two-dimensional nuclear overhauser effect spectroscopy (2D NOESY) in CDCl 3 -CD 3 CN. The sizable cross-peaks, (c, H 1 ) and (c, H 2 ), are apparently a consequence of the interactions through space between the macrocycle and guest G1 ( Figure S6). The additional NOE correlation signals could also be discerned between internal aromatic protons a of 1a and protons H 1 /H 2 in each pyridinium unit of G1. Since the two faces of the H-bonded aramide macrocycle are diastereotopic, NOEs were observed with protons on only one of its faces. The NOESY spectrum does not reveal any correlations associated with the interaction of side-chain protons of the host and internal aromatic protons. The through-space contact observed implies that the guest was trapped in the cavity of the macrocycle to form a 1a 2 ⊃G1 pseudo [3]rotaxane. The results from these two cases remain coherent with the complexation process, in which the guest tends to reside in the cavity of macrocyclic host 1a.

Photo-Switchable Complexation
Following the study on the complexation between 1a and protonated bipyridines/ phenylpyridines, we carried out 1 H NMR experiments to probe the controlled guest release and encapsulation by different stimuli. Among the reported stimuli, the use of light for manipulating guest binding and release is recognized as the most convenient means owing to its clean and benign process, fast response, and ease of spatio-temporal control [8]. H-bonded azo-macrocycle 1 used in this work features a molecular structure with two azo-benzene units incorporated into the macrocyclic backbone. Our previous work demonstrated that 1 is able to isomerize reversibly between E and Z configurations upon light irradiation [40]. The persistent shape due to the inherent nature of molecular rigidity and adaptable flexibility provides the macrocycle with an ability to quantitatively capture and release cationic guests by photo-irradiation. Protonated 4,4 -bipyridine G1 was selected in this work for the following capture and release experiments.
Before exploring switchable complexation with light, we first examined photoisomerization behavior under UV/BL light using 1 H NMR spectroscopy ( Figures S34 and S35). In the mixed solvent system of CDCl 3 and CD 3 CN (5:1, v/v), azo-macrocycle 1a exhibited pronounced light-induced E→Z isomerization. At the photo-stationary state, three isomers were identified as E,E-1a (10%), E,Z-1a (25%), and Z,Z-1a (65%), and the relative percentages were calculated by integrating the area of different stereoisomers ( Figure S34). This result is consistent with prior observations in a more polar solvent system (CDCl 3 -CD 3 CN, [40]. The light-induced isomerization process was finalized within 25 min as indicated by UV-vis (UV, 365 nm). A reverse process from Z,Z-1a to E,E-1a can be completed in less than 10 min when the sample solution is irradiated with blue light (BL, 450 nm) ( Figure S35), indicating the high photo-sensitivity for such an azo-macrocycle. It is unlikely to convert all E,E-1a to Z,Z-1a under the experimental conditions, probably due to the rigidity of the macrocyclic backbone. According to the DFT calculations (B3LYP(PCM, chloroform)/6-31G (d,p)), the size of the macrocyclic cavity decreases substantially upon azo-benzene photo-isomerization ( Figures S55-S57), which lays the foundation for the exclusion of guests through a reduction in cavity size, thereby ensuring the controlled binding and release of bipyridinium salts by using UV and BL irradiation.
Indeed, when a mixture of 1a and G1 (1a 2 ⊃G1 complex) was exposed to photoirradiation, the signals from protons H 1 and H 2 of E,Z-1a and Z,Z-1a appeared upfield (Figures 2b,c and S36). The changes in the chemical shifts of signals from aromatic protons of 1a were also observed. Specifically, the signal from the proton H 1 experienced a pronounced upfield shift (∆δ = −0.23 ppm) relative to the proton resonance of the 1a 2 ⊃G1 complex before UV irradiation, and the signal from the other proton H 2 also shifted upfield to a great extent (∆δ = −0.39 ppm), strongly suggesting that the guest was released (Figure 2c). Reversely, when the sample solution was irradiated with BL, the guest G1 was bound to the macrocycle again. This is indicated by the same signals reappearing at the same position, pointing to the regeneration of the E,E-1a 2 ⊃G1 complex. In fact, the signals from protons H 1 and H 2 had a pronounced downfield shift (∆δ = +0.14 ppm and +0.23 ppm, respectively) under blue light irradiation with respect to the same protons in solution after irradiation. It should be noted that not all the signals of protons H 1 and H 2 returned to the original resonances because the conversion from Z,Z-1a to E,E-1a was incomplete, which is normal in photo-isomerization for most azo-based compounds [69]. On the basis of these experimental observations, we conclude that neither E,Z-1a nor Z,Z-1a has an appreciable affinity for G1 via complexation within the cavity. Upon the exposure of the 1a 2 ⊃G1 complex to UV light irradiation, the E→Z transformation of the macrocycle occurred, leading to the reduced size of the cavity and, thus, the release of G1. Exposure of the irradiated mixture to light at 450 nm reset the system back to the guest-trapped E,E-1a 2 ⊃G1-enriched state. With a view to gaining further understanding of the detected binding properties, DFT-technique-based computational simulations were performed for the host-guest complex system comprising azo-benzene macrocycle 1c and protonated 4,4 -bipyridine at the B3LYP(PCM, chloroform)/6-31(d,p) level ( Figure 3 and Figures S55-S57). The results of the binding energies disclosed that each guest molecule resides in the macrocyclic cavity through multiple hydrogen-bonding interactions (Figure 3a). For example, in the case of G1, there exist six C-H···O hydrogen bonds and two + N-H···O hydrogen bonds in the complex (Figure 3a). An independent gradient model (IGM) analysis (Figure 3b, Figures S58 and  S59) revealed the multiple H-bond interactions between 1c and G1. DFT calculations led to the conclusion that the lowest energy for the molecular configuration of Z,Z-1c 2 ⊃G1 is 26.69 kcal/mol higher in energy than that of E,Z-1c 2 ⊃G1, which lies 39.49 kcal/mol above E,E-1c 2 ⊃G1 (Figure 3c). These findings are consistent with the notion that E,E-1a 2 ⊃G1 is more stable than E,Z-1a 2 ⊃G1 or Z,Z-1a 2 ⊃G1.
The results above demonstrate the reversibility of the present complexation system working under UV and BL light irradiation to a greater extent. Interestingly, the release and encapsulation of guests from the complexes 1a 2 ⊃G2, 1a 2 ⊃G3, 1a⊃G4, and 1a⊃G5 were not observed under photo-responsive conditions (Figures S37-S40). We hypothesized that the proton transfers from G2-G5 to the azo-benzene groups of 1a may occur, thereby blocking photo-isomerization. Protonation experiments were carried out to provide an important piece of evidence for the claim that protonated azo groups are unable to isomerize. Indeed, macrocycle 1a failed to isomerize in the presence of TFA ( Figure S45). We noticed that guests G2-G5 were in a state of dynamic equilibrium. When mixing bipyridinium (or pyridinium) salt (G2-G5) with bipyridine (or pyridine), only a signal set of signals ( Figures S41-S44) appear, indicative of the occurrence of proton transfer. Therefore, the observation for the failure in guest capture and uptake with G2-G5 under light irradiation is attributable to the rigidification of the molecular backbone through the protonation of azo groups in the macrocycle. Structure, relative energies of E,E-1c 2 ⊃G1, E,Z-1c 2 ⊃G1, and Z,Z-1c 2 ⊃G1. The energy is measured in kcal/mol and compared to the lower energy conformation of E,E-1c 2 ⊃G1, which is set to 0 kcal/mol.

Acid/Base-Switchable Complexation
Switchable complexation in the presence of acid/base is usually enabled by the preferential binding of one guest over another. Bipyridine and pyridine are known to be easily protonated, providing bipyridinium and pyridinium salts in solution, and a reverse process is induced by a base such as Et 3 N. The H-bonded azo-macrocycle 1a is capable of binding positively charged G1-G5 and has no or very low affinity for neutral bipyridine/pyridine. This constitutes the basis for the chemically controlled association and dissociation of the complexes by acid and base.
To this end, we first tested the interaction between 1a and 4,4 -bipyridine. The 1 H NMR spectrum shows slight changes (~0.02 ppm) for the 2:1 mixture of host and guest in CDCl 3 -CD 3 CN (5:1, v/v) (Figures 4a-c and S46). We attempted to determine the stoichiometry and the binding constant between 1a and 4,4 -bipyridine. The Job plot experiment indicates the formation of a 2:1 complex (Figures S47 and S48); however, the binding constant fitting based on the data from the 1 H NMR titration experiments failed (Figure S49), indicating that the binding affinity between 1a and 4,4 -bipyridine is too low to be determined [70]. Therefore, the effect of neutral bipyridine on the binding event is negligible. When 4.0 eq. of trifluoroacetic acid (TFA) was added to the solution, the aromatic ArH resonances (c, d, e, and f ) for 1a experienced an upfield shift by ∆δ = −0.49, −0.16, −0.18, and −0.11 ppm, respectively (Figure 4c,d), while protons H 1 and H 2 of guest G1 exhibited a downshift of ∆δ = +0.12 and +0.91 ppm, respectively (Figure 4c,f). Comparing the spectrum (Figure 4c) of the mixture with that of the complex E,E-1a 2 ⊃G1 (Figure 1c) indicates that a stable complex 1a 2 ⊃G1 was formed. As excess triethylamine (Et 3 N, 6 eq.) is added to the solution above, deprotonation occurs on G1, and as a consequence, the complex 1a 2 ⊃G1 dissociates (Figure 4e). This is clearly indicated by the reappearance of the resonance from 4 4-bipyridine at the original position. A similar change in signals is observed with 1a. These results demonstrate that the capture and release of 1a 2 ⊃G1 could be induced by adding in sequence TFA and Et 3 N. The complexes 1a 2 ⊃G2, 1a 2 ⊃G3, 1a⊃G4, and 1a⊃G5 can be regulated in a similar fashion (Figures S50-S53).

Competitive Ion Complexation
Tuning host-guest affinity may provide a useful alternative for the control of complicated switchable supramolecular systems. H-bonded azo-macrocycle 1a has a strong binding ability toward N,N-dialkyl-4,4 -bipyridinium salts. This promoted us to further examine the cation competition with a host-guest system comprising 1a and G2 (PF 6 − as a counter ion) in the presence of G6 as a competitive ion to release protonated 4,4 -bipyridine G2 ( Figure S54).
Firstly, when macrocycle 1a (2.0 mM) was mixed with guest G2 (1.0 mM) in CDCl 3 -CD 3 CN (1:1, v/v), a stable 1a 2 ⊃G2 complex was formed (Figure 5a). The 1 H NMR spectra show signals of protons H 3 and H 4 of G2, which shift upfield by ∆δ = −0.16 and −0.2 ppm, respectively. Accompanied by the change in chemical shifts is the appearance of signals corresponding to the formation of the complex 1a 2 ⊃G6 (Figures 5 and S54). The protons H 21 and H 22 show board signals because of the presence of rotational isomerization in this system [71]. Beyond 1.0 mM of G6, signals from protons H 3 and H 4 of guest G2 return by shifting upfield to the position where free G2 resonates (Figure 5e,g). These results indicate that the complex 1a 2 ⊃G6 is formed upon the dissociation of the complex 1a 2 ⊃G2, thereby accomplishing the release of guest G2 through a cation competition process. Unfortunately, this process is irreversible due to the much stronger binding affinity of paraquat G6 over G2 [40].

Experimental Methods
NMR Analytical NMR spectra were recorded on Bruker AVANCE AV II-400/600 MHz at a temperature of 298 K ( 1 H: 400 MHz; 2D: 600 MHz). Chemical shifts were reported in δ values in ppm using tetramethylsilane (TMS) or residual solvent as internal standard, and coupling constants (J) are denoted in Hz. Multiplicities are denoted as follows: s-singlet, d-doublet, t-triplet, dd-double doublet, and m-multiplet. High-resolution mass (HRMS) data were collected by WATERS Q-TOF Premier. UV-vis spectra were measured by SHIMADZU UV-2450 (Kyoto, Japan). 1 (1a and 1b) and Guests G1-G6

DFT Calculations
DFT calculations for the geometrical optimizations were performed in Gaussian 09 program package [74]. All substituents (R 1 and R 2 ) on the periphery of azo-macrocycle 1 were replaced by methyl groups to provide compound 1c for simplicity. The counterion CF 3 COO − of guest G1 was also omitted for clarity.

Visualization of Noncovalent Interactions
Independent gradient model (IGM) analysis is an approach [75] based on promolecular density (an electron density model prior to molecule formation) to identify and isolate intermolecular interactions. DFT optimize structures are used as input files. Strong polar attractions, van der Waals contacts, and repulsive forces are visualized as an isosurface with blue, green, and red colors, respectively. The binding surface was calculated by Multiwfn 3.8 program [76] and visualized using PyMOL [77].

Conclusions
In summary, we demonstrated a photo-responsive and acid/base-controllable hostguest system based on hydrogen-bonded azo-macrocycle 1 and protonated 4,4 -bipyridine. The complexation and dissociation of the complexes were explored under UV/BL irradiation and acid/base conditions. Switchable complexation was corroborated and rationalized by using the results from 1 H-NMR, HR-MS, binding constant, and titration experiments. Moreover, control can also be achieved using cation competition. The multi-responsive host-guest systems in this work may serve as a potential platform to construct multicomponent and complicated molecular machines. Correlational research is underway in our laboratory.