Conformationally adaptive macrocycles with flipping aromatic sidewalls

Xiaoping Wang a, Fei Jia ab, Liu-Pan Yang a, Hang Zhou a and Wei Jiang *a
aShenzhen Grubbs Institute, Guangdong Provincial Key Laboratory of Catalysis, and Department of Chemistry, Southern University of Science and Technology, Shenzhen, 518055, China. E-mail:
bInstitut für Chemie und Biochemie, Freie Universität Berlin, Arnimallee 20, 14195 Berlin, Germany

Received 8th April 2020

First published on 27th May 2020

Conformationally adaptive macrocycles possess multiple well-defined conformations through quickly flipping their aromatic sidewalls. The macrocycles combine the binding power of all the conformations. Upon binding a guest, one or a combination of conformations are selected to achieve the maximized binding affinity. In addition, the complex conformational network is responsive to changes in temperature or solvent. It has been demonstrated that these macrocycles have unique properties in chirality sensing, stimuli-responsive self-assembly, and molecular switches. In this tutorial review, we summarize recent advances on conformationally adaptive macrocycles with an emphasis on our own research. We believe that this class of macrocycles will have a bright future in supramolecular chemistry and beyond.

image file: d0cs00341g-p1.tif

Xiaoping Wang

Xiaoping Wang obtained his PhD degree from Xiamen University in 2016 with Prof. Hongping Zhu and worked as a postdoctoral fellow (2017–2019) at Southern University of Science and Technology (SUSTech) with Prof. Wei Jiang. He is currently a senior research scientist at SUSTech. His research is focused on stimuli-responsive macrocycles.

image file: d0cs00341g-p2.tif

Fei Jia

Fei Jia obtained his MSc degree from Jinan University (Guangzhou, China) in 2012. From 2013 to 2016, he worked with Prof. Wei Jiang at SUSTech as a research assistant, where he developed oxatubarene. Since October 2016, he has been pursuing his PhD degree at the Freie Universität Berlin under the joint supervision of Prof. Wei Jiang and Prof. Christoph A. Schalley. Currently, he is working on naphthocages.

image file: d0cs00341g-p3.tif

Liu-Pan Yang

Liu-Pan Yang is currently a research assistant professor at the Academy for Advanced Interdisciplinary Studies, SUSTech. He received his PhD degree in 2014 at Beijing Institute of Technology with Prof. Jia-Rong Li and worked as a postdoctoral fellow (2014–2017) at SUSTech with Prof. Wei Jiang. His current research focuses on supramolecular materials based on novel macrocyclic hosts.

image file: d0cs00341g-p4.tif

Hang Zhou

Hang Zhou received his BSc degree in Chemistry in 2017 from Beijing Normal University. Since July 2017, he has been a research assistant with Prof. Wei Jiang at SUSTech and will enroll into the joint PhD program between SUSTech and the University of Hong Kong in September 2020 under the guidance of Prof. Wei Jiang and Prof. David Lee Phillips. His research centers on the physical aspects of molecular recognition.

image file: d0cs00341g-p5.tif

Wei Jiang

Wei Jiang received his BSc degree in 2004 from Xi’an Jiaotong University and MSc degree in 2007 from Nankai University with Prof. Yu Liu. He was awarded the Schering prize for his PhD work which he finished in 2010 at Freie Universität Berlin with Prof. Christoph A. Schalley. After postdoctoral research with Prof. Julius Rebek Jr. at the Scripps Research Institute, he started his independent research career at the Department of Chemistry, SUSTech as a tenure-track associate professor and was promoted to full professor with tenure in 2018. His research is focused on biomimetic molecular recognition and its applications.

Key learning points

(1) Three limiting mechanisms for molecular recognition: lock-and-key, induced fit, and conformational selection.

(2) Conformational analysis and calculation of conformation numbers of macrocyclic hosts with flipping aromatic sidewalls.

(3) Conformational response of macrocyclic hosts with flipping aromatic sidewalls to guest bindings and environmental changes.

(4) The effect of conformational complexity on molecular switches, chiroptical sensing, and self-assembly.

1. Introduction

Molecular recognition is the basis of all biological processes and functions. Understanding the mechanism of molecular recognition is the key to understanding biology at the molecular level. More than a century ago, Fischer1 proposed the “lock-and-key” model to explain the specific catalytic property of enzymes. In this model, enzymes undergo a non-essential structural change upon binding a substrate. With more experimental results, it is clear that conformational changes are involved in molecular recognition. The “induced-fit” model, which suggests binding induces a new conformation, was then proposed by Koshland2 in 1958 to explain the mechanism. Soon after, this model was widely accepted and became a textbook mechanism for molecular recognition. Nevertheless, more and more recent experiments support an alternative model – conformational selection, which was proposed by Monod, Wyman and Changeux3 in 1965. In the “conformational selection” model, proteins exist as a large conformational ensemble (Fig. 1a). A ligand selects one best-fit conformation, and then the system undergoes a population shift towards this conformation. This model has been found very useful in understanding allostery4 in nature. However, both the “induced-fit” and “conformational selection” models are limiting in some aspects. In reality, conformational selection is often followed by minor conformational adjustment (that is, induced fit).
image file: d0cs00341g-f1.tif
Fig. 1 (a) Ligand binding through selecting one conformation from a conformational ensemble; binding models of (b) a flexible host or (c) a rigid host upon binding a rigid guest.

Conformational changes have been appreciated and used in molecular switches and machines, allosteric receptors, and controlling catalyst activity in supramolecular chemistry.5 However, understanding molecular recognition in synthetic systems seems to take the opposite path as in biology. Flexible hosts often have many conformations, but these conformations are not well-defined. The bound conformation is drastically different from the pre-existing conformations. Therefore, induced fit is often invoked to explain the binding mechanism of flexible hosts (Fig. 1b). Upon binding a guest, the conformational freedom of flexible receptors will be severely restricted, and an entropic penalty has to be paid with the binding free energy. This leads to low binding affinity and selectivity. To solve this problem, preorganization has been emphasized in designing molecular receptors.6 Highly preorganized and rigid hosts (Fig. 1c) indeed show extremely high binding affinity. Obviously, these receptors adopt the “lock-and-key” mechanism, and undergo essentially no structural change during binding. However, a perfect design of a rigid host for a specific target requires very high precision in the bonding angle and length, which is not always available in the chemical space. Consequently, a balance in conformational flexibility and preorganized structure has been advocated in designing efficient and selective molecular receptors.7

In contrast to biological systems, the importance of conformational ensembles in molecular recognition has rarely been systematically discussed in supramolecular chemistry.8 There are several macrocyclic hosts with multiple well-defined conformations, for example, calix[n]arenes,9 pillar[n]arenes,10 and oxatub[4]arene.11 These macrocycles have aromatic sidewalls, and flipping these sidewalls leads to different conformations. These conformations have different binding properties and constitute a very complex conformational network. The macrocycles are thus conformationally adaptive to guests or environmental changes (temperature or solvent). Therefore, we call them conformationally adaptive macrocycles. In this review, we discuss the conformational properties and conformational responsiveness of conformationally adaptive macrocycles with representative examples. With host–guest chemistry in mind, the selected macrocycles are restricted to the ones with aromatic sidewalls (ideally) parallel to their cavity space especially when a guest is complexed. Two cage molecules will also be discussed because they possess similar properties to these macrocycles. We first perform general conformational analysis on macrocyclic hosts with flipping aromatic sidewalls and then discuss each class of conformationally adaptive macrocycles with different types of linkers. With this review, we hope to draw the attention of chemists to the conformational ensemble properties of macrocyclic hosts.

2. Conformational analysis of macrocyclic hosts with flipping aromatic sidewalls

There are three scenarios for macrocyclic hosts with flipping aromatic sidewalls. Macrocyclic hosts with only two repeating aromatic sidewalls are shown in Fig. 2 for each case as an example. In model A, the linking on the aromatic sidewalls is symmetric and flipping these sidewalls will not lead to new conformations. In model B, the linking is symmetric but the sidewalls are not the same at the two portals of the cavity. Flipping the aromatic sidewalls leads to new conformations. For model C, the linking on the low symmetric aromatic sidewalls is at the two centrosymmetric positions. Flipping one sidewall not only leads to new conformations, but also creates a chiral space and a pair of enantiomers coexists. The interconversion of the two planar-chiral conformations (II and III) has to go through the achiral one (I). In terms of conformational complexity, models B and C are the focus of this review.
image file: d0cs00341g-f2.tif
Fig. 2 Conformational analysis of three types of macrocyclic hosts with flipping aromatic walls. Macrocycles with only two repeating units are used here to demonstrate the basic idea.

How can one calculate the number of stereoisomers (conformations) for a macrocycle with n aromatic sidewalls that follows model B or model C? The situation in model B is similar to that of inositol homologs.12 Therefore, the corresponding equations can be adopted to calculate the numbers of stereoisomers (S(n)), diastereomers (D(n)), and enantiomer pairs (E(n)). In the case of model C, the solution for the conformational numbers can be found from the binary necklace problem in the field of mathematics. Isobe and coworkers13 have modified the mathematic equations to calculate the conformational numbers of cyclonaphthylenes. All the equations and worked examples can be found in the ESI. According to these equations, the calculated numbers of stereoisomers (S(n)), diastereomers (D(n)), and enantiomer pairs (E(n)) are listed in Table 1. Generally, model B and model C with the same number of aromatic sidewalls have the same numbers of diastereomers. But model C has more enantiomer pairs and thus more overall stereoisomers. For n < 6, there is no chiral conformation for model B.

Table 1 Numbers of stereoisomers (S(n)), diastereomers (D(n)), and enantiomer pairs (E(n)) of model B and model C with n repeating aromatic sidewalls
n Model B Model C
S (n) D (n) E (n) S (n) D (n) E (n)
2 2 2 0 3 2 1
3 2 2 0 4 2 2
4 4 4 0 6 4 2
5 4 4 0 8 4 4
6 9 8 1 13 8 5
7 10 9 1 18 9 9
8 22 18 4 30 18 12
9 30 23 7 46 23 23
10 62 44 18 78 44 34
11 94 63 31 126 63 63

Aromatic sidewalls define the cavity of macrocyclic hosts. Conformational interconversion is realized by flipping the sidewalls through the cavity. Therefore, the kinetics of conformational interconversion is determined by the size of the aromatic walls and the size of the cavity. To maintain quick conformational response, fast interconversion kinetics at ambient conditions is ideal. That is, it is better to have a low barrier for conformationally adaptive macrocycles. In addition, the thermodynamic stability of the conformations should be similar. Thus, these conformations coexist or are thermally accessible in response to guests or environmental changes.

The conformations have different symmetry and may show different 1H NMR peak patterns. But they often interconvert very quickly at the free state and cannot be distinguished at room temperature in most cases. Upon binding a guest, the cavity is occupied and flipping aromatic sidewalls is stopped. Consequently, the conformational interconversion kinetics is coupled to the guest exchange kinetics. When the guest exchange kinetics is slow on the NMR timescale, the well-defined 1H NMR peak patterns may be used to identify the guest-selected conformation(s) according to their different symmetries. In addition, 2D NMR experiments may be further used to distinguish the conformations with the same expected peak patterns but different structural arrangements. When it is not possible to characterize the conformations with 1D and 2D NMR, X-ray single crystallography would be the last resort.

Multiple conformations provide larger accessible binding space. This would significantly expand the binding scope by combining the binding ability of all the conformations. For a specific guest, one or a combination of best-fit conformations are selected to optimize the binding interactions. When the environmental parameters change, the conformational ensemble may respond to minimize the changes in binding affinities. For model C, a chiral guest may also induce the predominant existence of one of the conformational enantiomers. This leads to a chiroptical response to the chiral guest, which may be used in chiroptical sensing.

All the above properties, including conformational kinetics, guest binding ability and conformational adaptivity, are affected by the structural rigidity of macrocyclic hosts. For very rigid macrocyclic hosts, conformational kinetics may be restricted and conformational adaptivity is limited. Flexible macrocyclic hosts are conformationally responsive to a guest not only through selecting a conformation in the pre-existing conformational ensemble but also inducing a local conformational change after binding. The structural flexibility or rigidity is significantly affected by the linkers that connect the aromatic sidewalls together. Therefore, these macrocycles can be further divided into three classes according to their linkers: long (aromatic) linkers; direct linkage or methylene linkers; CH2–O–CH2 linkers. In addition, the naphthocages14,15 with the aromatic sidewalls joined together from top and bottom have similar properties to these macrocycles, representing another class. This classification is used in the following discussion. The consequence of conformational ensemble in molecular recognition and self-assembly is addressed as well.

3. Macrocycles with long (aromatic) linkers

Cyclobis(paraquat-p-phenylene) (14+, Fig. 3a),16 the so-called “blue box”, possesses flipping aromatic sidewalls. However, 14+ is a model A-like macrocycle, and does not have the kind of conformational isomers as discussed above. When replacing the phenylene groups in 14+ with 1,5-disubstituted naphthalenes,17 the resulting structure 24+ is a model C-like macrocycle. The two bipyridiniums can be considered as the long aromatic linkers. As shown in Fig. 3a, it has three conformations: one achiral conformation (I) and a pair of planar-chiral ones (II and III). In solution, the diastereomeric ratio (I[thin space (1/6-em)]:[thin space (1/6-em)]II/III) is 1[thin space (1/6-em)]:[thin space (1/6-em)]3 in favour of the planar-chiral conformations. This ratio remains unchanged over a wide temperature range (−63–60 °C), indicating that the conformational interconversion is slow at room temperature. However, isolation of the diastereomers is unsuccessful. When incorporating 24+ into a catenane, the diastereomeric ratio (I[thin space (1/6-em)]:[thin space (1/6-em)]II/III) becomes 1[thin space (1/6-em)]:[thin space (1/6-em)]4, suggesting a conformational response to guest binding.
image file: d0cs00341g-f3.tif
Fig. 3 (a) Chemical structures and conformational analysis of cyclophanes 14+ and 24+; (b) chemical structures of macrocycles 3–6.

Macrocycles 3–6 (Fig. 3b)18–20 with long diyne, isophthalamide, or 2,6-pyridine dicarboxamide as linkers, also have conformational complexity. Macrocycles 3–5 follow model B and should have two conformations. Macrocycle 6 is analogous to 24+, and has one achiral and two planar-chiral conformations. For the tetralactam macrocycles 4–6, the conformational interconversion is very fast and no ratio can be determined in the absence of a guest. Upon guest binding, the high-symmetry conformations (the ones shown in Fig. 3b) are usually selected. For macrocycle 3 (R = CH2OCOCH3), the conformation ratio (I[thin space (1/6-em)]:[thin space (1/6-em)]II) is ca. 1[thin space (1/6-em)]:[thin space (1/6-em)]14.4 in favour of the anti conformation (II) at −13 °C. The activation barrier of the conformational interconversion from the anti to syn conformers is estimated to be 11.7 kcal mol−1. However, upon binding to 2-naphthalenesulfonic acid in water, the syn conformation (I) of 3 (R = OCOCH2NHMe2+) is selected to provide a more hydrophobic environment for the guest.

These macrocycles (24+6) possess only two aromatic sidewalls that can flip to form different conformations. The bipyridinium, diyne, and isophthalamide are considered as linkers and do not contribute to the comformational complexity due to their high symmetry and/or the fact that they are not parallel to the cavity space. Higher comformational complexity can be generated with large numbers of such aromatic sidewalls. In addition, the linkers may be shorter to maximize the contribution of the aromatic sidewalls in the binding.

4. Macrocycles with direct linkage or methylene linkers

Cycloparaphenylenes (7n, n = 8–14, Fig. 4a)21 consist solely of benzene rings which are directly linked at the para positions. These benzene rings can be flipped through rotating around single bonds. However, they follow model A in Fig. 2, and no new conformation will be produced through flipping the aromatic sidewalls. In contrast, belt-shaped cyclonaphthylenes have multiple conformations through flipping the naphthalene panels. Cyclonaphthylene 8 with 1,4-linkages (Fig. 4a) has been reported by Itami and coworkers.22 This compound follows model B. According to Table 1, 30 conformations can be expected. However, due to the steric hindrance, the ground-state structure of 8 has low symmetry and adopts a twisted conformation with one naphthalene parallel to the nanoring plane as suggested by DFT calculation and the 1H NMR spectrum. This conformation is inherently chiral with a racemization barrier of 19.9 kcal mol−1 (calculated). The conformational interconversion is slow on the NMR timescale at room temperature. Conformational response and host–guest chemistry have not been reported with this macrocycle.
image file: d0cs00341g-f4.tif
Fig. 4 (a) Chemical structures of cycloparaphenylenes, cyclonaphthylenes and cyclochrysene; (b) relative energetics of the diastereomeric structues of 97 obtained with DFT calculations. Reproduced with permission from ref. 13. Copyright 2016, National Academy of Sciences; (c) four crystal structures of 97 in a set of disordered structures.

Cyclonaphthylenes (9n) with 2,6-linkages reported by Isobe and coworkers13,23 follow model C. Due to the different linkages, the steric demand is less severe in 9n than that in 8. The calculated barrier of conformational interconversion is relatively low (for example, n = 6, 15.4 kcal mol−1; n = 7, 12.4 kcal mol−1) with a nonlinear decay toward the larger ring size. Therefore, they can undergo very fast conformational interconversion at room temperature. The numbers of conformers, enantiomers, and diastereomers of 9n can be found in Table 1. For example, 97 has 18 conformational isomers with 9 pairs of enantiomers. The relative stability of these isomers was calculated and is shown in Fig. 4b. The two most stable conformations a7 and a6b are also the ones found in the crystal structures (Fig. 4c). However, this is not necessarily always true. For example, the existing conformer a6b2 in the crystal structure23 of 98 is 2 kcal mol−1 less stable than the most stable conformer. With larger aromatic panels, the conformational interconversion is severely slowed down and even stopped. For example, chrysene-based macrocycle 10 cannot undergo conformational interconversion at room temperature.24 An equilibrium can only be reached after heating its solution in toluene for 8 hours at 100 °C. This interconversion kinetics is too slow for conformationally adaptive macrocycles.

Conformational interconversion kinetics of 97 is very fast with a half-life of 0.006 s at −10 °C, and its cavity size is complementary to C70.25 Depending on the solvents, the binding constant between 97 and C70 varied in the range of 107–109 M−1. The conformational interconversion kinetics is coupled to guest exchange kinetics which is slow (guest exchange barrier: 18 kcal mol−1) on the NMR timescale. Therefore, the bound conformation of 97 in the presence of C70 can be easily identified by an NMR peak pattern according to the symmetry. Among 9 possible diastereomers, only conformers a7/b7 were selected to bind C70, which is consistent with the crystal structure.

Macrocyclic arenes are a class of macrocyclic hosts in which aromatic sidewalls are connected by methylene groups. These macrocycles are often synthesized by reacting phenols or other related compounds with paraformaldehyde.26,27 Due to the methylene linkers, macrocyclic arenes are structurally more flexible than the above-discussed cyclophane compounds.

Calix[n]arenes (Fig. 5a)9 are one of the oldest macrocycles in the family of macrocyclic arenes. It is also the most classic conformationally adaptive macrocycle. The phenol groups are tilted to define a cavity, and can flip with a low barrier, resulting in multiple conformations. They belong to model B. Four typical conformations exist for calix[4]arene and calix[5]arene: cone, partial cone, 1,2-alternate, and 1,3-alternate (Fig. 5b). For calix[4]arene, only the cone conformation has a cavity which is able to host an organic guest; the cavities of the other three conformations are collapsed and cannot well accommodate organic guests. However, more conformations of larger calix[n]arenes have a cavity to accommodate organic cations. For example, the cone and partial cone conformations of calix[5]arene 1128 (Fig. 5c) are selected to complex secondary ammonium G1+ and the two conformations exist in a 55[thin space (1/6-em)]:[thin space (1/6-em)]45 ratio (cone[thin space (1/6-em)]:[thin space (1/6-em)]partial cone). This shows that calix[n]arenes are able to show conformational response to a guest. However, the cone conformation is more often desired to provide a well-defined cavity. This can be achieved through fixing the conformation by attaching large groups on the hydroxyl groups to completely block conformational interconversion.9 Other calix-shaped macrocycles have similar conformational properties as calix[n]arenes.9

image file: d0cs00341g-f5.tif
Fig. 5 Chemical structures of (a) calix[n]arene and (b) the four conformations of calix[4]arene; (c) conformational distribution of calix[5]arene 11 upon binding guest G1+ as determined by 1H NMR experiments (400 MHz, CDCl3, 25 °C).

Pillar[n]arenes (Fig. 6a)10,29 are very close relatives of calix[n]arenes, and are synthesized by reacting 1,4-dialkoxylbenzene with paraformaldehyde in the presence of Lewis acid. Pillar[n]arenes follow model C. The number of their conformations can be found in Table 1. For pillar[5]arene, 8 conformations exist with 4 pairs of enantiomers (Fig. 6a). There are 13 conformations for pillar[6]arene with 5 pairs of enantiomers and 8 diastereomers. However, most of these conformations are not detected due to steric hindrance of the substituents on the pillar[n]arenes. Instead, substituted pillar[n]arenes mainly exist as a racemate of planar-chiral conformers pR and pS (Fig. 6b). These two conformers of pillar[5]arenes are interconverted even with dodecyl groups as substituents (activation free energy: 15.1 kcal mol−1).30 But cyclohexylmethyl groups are bulky enough to stop the conformational interconversion, resulting in separation of the enantiomers.31

image file: d0cs00341g-f6.tif
Fig. 6 (a) Chemical structures of pillar[n]arenes and all the possible conformations of pillar[5]arene; (b) guest-induced population shift of the two enantiomers of pillar[5]arene 12 and 13, and conformational change of pillar[5]arene with chiral substituents 14 and the enantioseparated bicyclic pillar[5]arenes 15–17 in response to temperature, solvent, or guests; (c) chemical structures of S6-corona[3]arene[3]pyridazines 18 and guests G22+, G3+ and G42+ and X-ray crystal structures of the complexes of 18 with G22+, G3+ or G42+.

The interconversion between the enantiomeric pairs of pillar[5]arenes has been used in chiroptical sensing of chiral guests. Yang et al.32 and Lin et al.33 showed that pillar[5]arenes 12 and 13 (Fig. 6b) are able to detect the chirality and enantiomeric compositions of amino acid esters through guest-induced conformational shift and generating induced circular dichroism (CD) signals. With attaching chiral substituents on the pillar[5]arene (14), the distribution of the two diastereomers is biased and the chiral conformation distribution can be changed by altering the temperature, solvent, and guest.34 The Yang group35 and the Ogoshi group36 also reported bicyclic pillar[5]arenes (15–17, Fig. 6b) which are chirally resolved to provide enantiopure samples. These enantiopure compounds show two states: the cyclic side chain sits inside or outside the cavity. This would lead to the chirality inversion and thus inversion in the CD signals. Both research studies show that this conformational interconversion is responsive to solvent, temperature, and guest. All these stimuli worked through altering the binding affinity with the cyclic side chains or competing with this binding. The chiral conformational adaptivity of pillar[5]arenes is very unique, and may be used as a general principle in chiroptical sensing. Other related conformationally adaptive macrocycles27,37 may have such conformational properties as well.

As shown in Fig. 6c, coronarene 18[thin space (1/6-em)]38 has three lying aromatic walls and three standing aromatic walls. Flipping the three standing aromatic walls would lead to two pairs of planar-chiral conformations according to model C in Table 1. As shown in the three crystal structures with guests G22+G42+, different conformations are selected by different guests through flipping the 1,4-dicarboxylate phenylene walls.39 This shows that coronarene 18 is conformationally adaptive to guests.

5. Macrocycles with CH2–O–CH2 linkers

Although the macrocycles discussed above show a conformational response to guests and environmental changes, the advantage of conformationally adaptive macrocycles requires further demonstration. Ideally, all the conformations of the macrocycles have good guest-binding ability. In this way, a rich conformational response can be expected. Zorbarenes and oxatubarenes (Fig. 7) with flexible CH2–O–CH2 linkers are exactly such kind of conformationally adaptive macrocycles.
image file: d0cs00341g-f7.tif
Fig. 7 Chemical structures of zorb[4]arenes (19–20), oxatub[4]arenes (21–25) and the hybrid macrocycle (26) and their conformations resulting from naphthalene flipping.

Zorb[4]arene (ZB4)40,41 belongs to model B and there should be four achiral conformations (Table 1). However, this macrocycle is very flexible and exists as a self-inclusion conformation in the X-ray crystal structure (Fig. 8b). Monte Carlo conformational search generates 3704 conformations and further DFT calculations identify 10 representative conformers within a 10.0 kcal mol−1 range. But only the four conformations shown in Fig. 7 have a cavity to accommodate a guest. The four conformers have different symmetry and may be distinguished according to different peak patterns in 1H NMR spectra provided their interconversion is slow on the NMR timescale. At room temperature, the NMR peaks are sharp and only two aromatic signals exist, suggesting either only conformer I/IV predominates in solution or the conformational interconversion is very fast on the NMR timescale. However, the diastereotopic methylene protons of the CH2–O–CH2 linkers exist as a single peak, supporting fast conformational interconversion at room temperature. The NMR peaks remain sharp above −40 °C and are split into a complex pattern at −80 °C, suggesting a very low rotational barrier. This is ideal for conformationally adaptive macrocycles.

image file: d0cs00341g-f8.tif
Fig. 8 (a) Chemical structures of organic cation guests for ZB4, TA4, and HM4; guest-selected conformations of (b) ZB4, (c) TA4, and (d) HM4 as confirmed by X-ray crystal structures (with filled, coloured aromatic panels) and by 1D/2D NMR spectra (calculated structures, with nonfilled, coloured aromatic panels). The butoxyl groups are shortened to methoxyl groups in the crystal structures and calculated structures for viewing clarity. The binding constants were determined by ITC titrations (1,2-dichloroethane[thin space (1/6-em)]:[thin space (1/6-em)]MeCN = 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 25 °C) or by 1H NMR titrations (400 MHz, CD2Cl2[thin space (1/6-em)]:[thin space (1/6-em)]CD3CN = 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 25 °C).

Oxatub[4]arene (TA4, Fig. 7)11 belongs to model C. 4 diastereomers and 2 planar-chiral enantiomer pairs can be obtained through flipping the aromatic sidewalls (Table 1). 1,5-Linkages in TA4 avoid the self-inclusion conformation as observed in ZB4. Similarly, the 1H NMR peaks only start to become broadened at the temperature below −60 °C and split into a complex pattern at −80 °C, indicating fast conformational interconversion at room temperature as well. The hybrid macrocycle (HM4, Fig. 7)42 with repeating units from both TA4 and ZB4 belongs to neither model B nor model C. 5 diastereomers and 2 enantiomer pairs exist due to the sidewall flipping. The 1H NMR peaks are sharp at room temperature and split into well-defined signals at −40 °C. This temperature is higher than those of TA4 and ZB4, suggesting a slightly higher flipping barrier.

Macrocycles TA4, ZB4, and HM4 have electron-rich cavities and are good receptors for organic cations (Fig. 8a) in nonpolar solvents. The three macrocycles have a wide guest-binding scope, and TA4 is often the best receptor among them.42 The binding constants range from 102 to 105 M−1 (Fig. 8b–d). This shows that flexible receptors can also achieve very decent binding affinities.

As discussed above, when guest exchange kinetics is slow on the NMR timescale, the conformational interconversion of the complexes becomes slow. The bound conformation(s) can be assigned based on the 1H NMR peak pattern and 2D NMR in solution. In addition, X-ray single crystallography provides more clear-cut evidence for the conformations. As shown in Fig. 8, three out of the four possible conformations are detected in the complexes of ZB4 (19);40,43 two out of five diastereomers are observed for HM4 (26);42 all four diastereomeric conformations of TA4 (21) are induced by guests.11,44 From another point of view, this suggests that different conformers show different binding affinities to the same guest. This was further demonstrated by DFT calculations on the complexes of oxatub[6]arene with C60/C70 using different conformers.45 These macrocycles are conformationally adaptive and use different conformations in binding different guests. The driving force of conformational change should be thermodynamic with the purpose to achieve maximized noncovalent interactions.

The corresponding complexes with TA4 (21) also show conformational response to the remote electronic substituents on guests (Fig. 9a). Remote electronic substituents on the para positions of guest G9+ can significantly affect their binding affinities with TA4 (21) through a field/inductive effect as revealed by a linear free energy relationship with σm.46 Surprisingly, the conformational distribution undergoes drastic changes with different substituents. Conformer I predominates for the guests with the electron-withdrawing groups (NO2 and CN); conformers I and II coexist at the similar percentage for the halogen substituents and CF3; conformer IV is favoured with the electron-donating substituents (H, OMe, Me, and C(Me)3); conformer II is the major conformation for SMe. This suggests that the conformational network of TA4 shows response to the remote electronic substituents on guests to maintain the linear free energy relationship. Nevertheless, such conformational response and linear free energy relationship were not observed for ZB4.43

image file: d0cs00341g-f9.tif
Fig. 9 (a) Conformational distributions of TA4 upon binding to guests G9+ with different electronic substituents at the para positions (1H NMR, 400 MHz, CD2Cl2[thin space (1/6-em)]:[thin space (1/6-em)]CD3CN = 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 25 °C). Reproduced with permission from ref. 46. Copyright 2017, American Chemical Society; (b) cartoon representation of the conformational response of TA4 to temperature as revealed through rotaxane synthesis in 1,2-dichloroethane. Reproduced with permission from ref. 48. Copyright 2019, Royal Society of Chemistry; (c) photo of ZB4 (R = n-dodecyl group, 20) and SEM images of its complexes with guests G13+, G9a+, or G10+ which induce conformation I, III, or IV of ZB4, respectively. Reproduced with permission from ref. 41. Copyright 2017, John Wiley and Sons; (d) SEM images of TA4 (R = n-octyl group, 22) before and after binding to guest G22+. Reproduced with permission from ref. 49. Copyright 2017, Royal Society of Chemistry.

The complexes of TA4 also show conformational response to solvent and temperature. For the complex between G9e+ and TA4 (tetraethylene glycol side chains, 25), conformers I, II, and IV coexist in a 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio in CD2Cl2/CD3CN = 1[thin space (1/6-em)]:[thin space (1/6-em)]1, which is changed to 5[thin space (1/6-em)]:[thin space (1/6-em)]2[thin space (1/6-em)]:[thin space (1/6-em)]2 in water.47 This shows the conformational response to solvent change. In order to study the conformational response to temperature, rotaxane synthesis was resorted to in order to fix the conformational distribution at different temperatures (Fig. 9b).48 In the rotaxane structure, conformer IV predominates at the temperatures below 40 °C, but conformer III starts to emerge at higher temperatures and prevails at 60–75 °C. This is surprising, because conformer III has never been detected for 1,4-diazabicyclo[2.2.2]octane-based organic cations at room temperature. In addition, the binding constants of TA4 (21) to guest 28+ do not change much at different temperatures. This is presumably due to the “buffer effect” of the complex conformational network of TA4.

The conformational complexity of these macrocycles can influence their macroscopic assembly morphologies. With long alkyl groups, ZB4 and TA4 can form nanostructures upon evaporating solvents (Fig. 9c). At room temperature, dodecyl-substituted ZB4 (20) with a high molecular weight (2155) is liquid. This is presumably due to poor packing and intermolecular interactions.40 Guests G13+, G9a+, and G10+ induce conformers I, III, and IV of 20, respectively. It was found that the complexes of 20 with guests G13+ and G9a+ self-assemble into sheet-like structures, while complex G10+@20 forms a nanotube structure with a length up to 12.5 μm and a diameter of 750 nm. At different conformations, the distributions of the alkyl groups on the naphthalenes of ZB4 are different. This may affect the self-assembly structures of ZB4. Further experiments suggest that the nanotube structures are folded from nanosheets during solvent evaporation. However, the distribution of alkyl groups for TA4 is not drastically different and no very different structures are observed for the complexes of TA4 with different conformations.49 But binding a guest drastically changes the morphologies of TA4. For example, the leaf-like morphology of 22 was changed to flower ball after binding to guest G22+ (Fig. 9d).

6. Naphthocages

Although macrocycles TA4, ZB4, and HM4 have a relatively wide guest binding scope, the binding constants are relatively low (102–105 M−1). This should be caused by the high entropic penalty during binding. A question arises: can we achieve high-binding affinity with a conformationally flexible receptor? This problem may be solved by harnessing multiple noncovalent interactions and completely wrapping a guest with a cage structure. In this way, the non-covalent interactions are strong enough to compensate the entropic penalty, resulting in high binding affinity.

Naphthocages NC114 and NC215 (Fig. 10a), with three aromatic sidewalls linked by two 1,3,5-triethylbenzene groups from top and bottom, were then synthesized. They have extremely strong binding affinities to organic cations. For singly charged organic cations (Fig. 10b), the binding constants of NC1 and NC2 (Table 2) are generally over 106 M−1 with the best to be 1010 M−1. This is very surprising for such a flexible macrocycle. More surprisingly, the cage NC1 even adopts a self-inclusion conformation at the free state, as evidenced by X-ray single crystal structure (Fig. 10c) and 1H NMR spectra. The strong binding enables construction of an ion-selective electrode with a super-Nernstian response to acetylcholine chloride (Ach+) in water.

image file: d0cs00341g-f10.tif
Fig. 10 (a) Chemical structures and DFT calculated structures of naphthocages NC1 and NC2 and their conformers. Reproduced with permission from ref. 15. Copyright 2020, Royal Society of Chemistry; (b) chemical structures of organic cation guests for the naphthocages; (c) X-ray single crystal structure of NC1, G82+@NC1-I and G10+@NC2-II; cyclic voltammetry (ClCH2CH2Cl[thin space (1/6-em)]:[thin space (1/6-em)]CH3CN = 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 1.0 mM, 25 °C, 100 mV s−1, electrolyte: n-Bu4NPF6, 0.1 M) of (d) ferrocene (Fc) with or without NC1 and (e) NC1 alone. Potential is given against a silver wire pseudoreference electrode. The calculated structure of NC1˙+ is shown in (e). Reproduced with permission from ref. 14 and 50. Copyrights 2019 and 2020, American Chemical Society.
Table 2 Binding constants (Ka, M−1) and conformational distribution of NC1 and NC2 in the presence of guests G7+, G10+, or G14+–G16+a
Guest NC1 NC2
K a I[thin space (1/6-em)]:[thin space (1/6-em)]II K a I[thin space (1/6-em)]:[thin space (1/6-em)]II
a The binding constants were determined by ITC titrations (1,2-dichloroethane[thin space (1/6-em)]:[thin space (1/6-em)]MeCN = 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 25 °C); the ratios of the two conformations were determined through 1H NMR integrations (400 MHz, CD2Cl2[thin space (1/6-em)]:[thin space (1/6-em)]CD3CN = 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 25 °C).
G7+ 6.1 × 109 72[thin space (1/6-em)]:[thin space (1/6-em)]28 8.9 × 106 5[thin space (1/6-em)]:[thin space (1/6-em)]95
G10+ 3.7 × 107 72[thin space (1/6-em)]:[thin space (1/6-em)]28 3.5 × 106 22[thin space (1/6-em)]:[thin space (1/6-em)]78
G14+ 1.6 × 107 98[thin space (1/6-em)]:[thin space (1/6-em)]2 1.1 × 108 12[thin space (1/6-em)]:[thin space (1/6-em)]88
G15+ 1.8 × 107 95[thin space (1/6-em)]:[thin space (1/6-em)]5 7.5 × 107 7[thin space (1/6-em)]:[thin space (1/6-em)]93
G16+ 5.1 × 107 95[thin space (1/6-em)]:[thin space (1/6-em)]5 1.5 × 108 9[thin space (1/6-em)]:[thin space (1/6-em)]91

Both NC1 and NC2 have multiple conformations (Fig. 10a). NC1 follows model C, and there are four conformations with two planar-chiral enantiomeric pairs. Only two conformations are expected for NC2. The distribution of the conformations in the presence of a guest again reflects their conformational response. As shown in Table 2, conformer I is predominant for NC1, while conformer II is favoured for NC2 for all five guests. This suggests that the two naphthocages have different conformational responses. The X-ray single crystal structure G10+@NC2-II is shown in Fig. 10c. In addition, they also have different binding preference: NC1 binds similarly to tetramethylammonium G14+ and tetraethylammonium G10+, while NC2 prefers G14+ over G10+. Mixing G14+, G10+, NC1, and NC2 in 1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio leads to a self-sorting system with only complexes G10+@NC1 and G14+@NC2.

NC1 shows essentially no binding to neutral ferrocene (Fc) but strong binding to the cobaltocenium (G7+) – an analogue of ferrocenium (Fc+) whose binding constant cannot be determined by isothermal titration calorimetry because of its intrinsic instability in air. Due to the strong binding, NC1 can significantly alter the cyclic voltammetry (CV) peaks of ferrocene (Fc) by cathodically shifting the anodic and cathodic peaks by ca. 80 and 810 mV, respectively (Fig. 10d). Digital simulation reveals a binding constant of 1.4 × 1010 M−1 (in CH2Cl2) between Fc+ and NC1 and an extremely large binding selectivity (ca. 1015) of Fc+ over Fc. NC1 also shows strong binding (up to 1010 M−1) to aromatic (di)cationic guests, for example, methyl viologen (G82+) and tetrathiafulvalene (TTF) dication, which are intercalated between two naphthalene walls inside the cage cavity (Fig. 10c).50 The redox-active guests can be switched into and out of the cage cavity with high binding selectivity (1010–1012) by oxidation and reduction processes, respectively. In addition, free cage NC1 is also redox-active. Empty NC1 can be oxidized via two steps (Fig. 10e). The first oxidation leads to a mono radical cation (NC1˙+) which is stabilized by a self-intercalated structure with the oxidized naphthalene sandwiched between two neutral ones (Fig. 10e). Further two-electron oxidation affords NC13(˙+) which adopts an expanded conformation due to charge repulsion. The redox property of NC1 was used as well to electrochemically control cationic guest release.50 The rich electrochemistry and host–guest chemistry may enable NC1 to be used in stimuli-responsive materials.

7. Conclusions and outlook

Macrocyclic hosts with flipping aromatic sidewalls possess multiple conformations, and are conformationally adaptive to guests and environmental changes (temperature and solvents). Cyclonaphthylenes, cyclophanes and tetralactam macrocycles with naphthalene sidewalls, calix[n]arenes, pillar[n]arenes, coronarenes, zorbarenes, oxatubarenes and naphthocages are all macrocycles of this kind. A complex conformational network is formed through the interconversion of multiple well-defined conformations. The macrocycle combines the binding power of all the conformations, and thus gives rise to strongest noncovalent interactions by selecting the best-fit conformation(s). This feature is analogous to that observed in proteins, but the whole system is only one synthetic macrocycle.

Conformationally adaptive macrocycles have the conformational ensemble features of bioreceptors. In view of the importance of conformational ensemble in allostery and signal transduction in nature, we believe conformationally adaptive macrocycles will find wide applications especially in chirality sensing, stimuli-responsive self-assembly, and molecular machines and devices. Some of these applications have been demonstrated in this review.

Conflicts of interest

The authors declare no conflicts of interest.


We are grateful to all collaborators and coworkers for their contributions. We thank the National Natural Science Foundation of China (No. 21572097 and 21822104), the Shenzhen Science and Technology Innovation Committee (JCYJ20180504165810828), Guangdong Provincial Key Laboratory of Catalysis (2020B121201002), and the Shenzhen Nobel Prize Scientists Laboratory Project (C17783101) for financial support. W. J. acknowledges Shenzhen Education Bureau for the support of “Pengcheng Scholar”.


  1. E. Fischer, Ber. Dtsch. Chem. Ges., 1894, 27, 2984–2993 Search PubMed .
  2. D. E. Koshland, Proc. Natl. Acad. Sci. U. S. A., 1958, 44, 98–104 CrossRef CAS PubMed .
  3. J. Monod, J. Wyman and J.-P. Changeux, J. Mol. Biol., 1965, 12, 88–118 CrossRef CAS PubMed .
  4. H. N. Motlagh, J. O. Wrabl, J. Li and V. J. Hilser, Nature, 2014, 508, 331–339 CrossRef CAS PubMed .
  5. P. C. Knipe, S. Thompson and A. D. Hamilton, Chem. Sci., 2015, 6, 1630–1639 RSC .
  6. D. J. Cram, Angew. Chem., Int. Ed. Engl., 1986, 25, 1039–1134 CrossRef .
  7. J. K. M. Sanders, Chem. – Eur. J., 1998, 4, 1378–1383 CrossRef CAS .
  8. L.-P. Yang, L. Zhang, M. Quan, J. S. Ward, Y.-L. Ma, H. Zhou, K. Rissanen and W. Jiang, Nat. Commun., 2020 DOI:10.1038/s41467-020-16534-9 .
  9. Calixarenes and Beyond, ed. P. Neri, J. L. Sessler and M.-X. Wang, Springer International Publishing, Basel, 2016 Search PubMed .
  10. T. Ogoshi, T.-a. Yamagishi and Y. Nakamoto, Chem. Rev., 2016, 116, 7937–8002 CrossRef CAS PubMed .
  11. F. Jia, Z. He, L.-P. Yang, Z.-S. Pan, M. Yi, R.-W. Jiang and W. Jiang, Chem. Sci., 2015, 6, 6731–6738 RSC .
  12. A. Yajima, Bull. Chem. Soc. Jpn., 2014, 87, 1260–1264 CrossRef .
  13. Z. Sun, T. Suenaga, P. Sarkar, S. Sato, M. Kotani and H. Isobe, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 8109–8114 CrossRef CAS PubMed .
  14. F. Jia, H. Hupatz, L.-P. Yang, H. V. Schröder, D.-H. Li, S. Xin, D. Lentz, F. Witte, X. Xie, B. Paulus, C. A. Schalley and W. Jiang, J. Am. Chem. Soc., 2019, 141, 4468–4473 CrossRef CAS PubMed .
  15. S.-B. Lu, H. Chai, J. S. Ward, M. Quan, J. Zhang, K. Rissanen, R. Luo, L.-P. Yang and W. Jiang, Chem. Commun., 2020, 56, 888–891 RSC .
  16. B. Odell, M. V. Reddington, A. M. Z. Slawin, N. Spencer, J. F. Stoddart and D. J. Williams, Angew. Chem., Int. Ed. Engl., 1988, 27, 1547–1550 CrossRef .
  17. P. R. Ashton, S. E. Boyd, S. Menzer, D. Pasini, F. M. Raymo, N. Spencer, J. F. Stoddart, A. J. P. White, D. J. Williams and P. G. Wyatt, Chem. – Eur. J., 1998, 4, 299–310 CrossRef CAS .
  18. S. P. Adams and H. W. Whitlock, J. Am. Chem. Soc., 1982, 104, 1602–1611 CrossRef CAS .
  19. L.-L. Wang, Y.-K. Tu, H. Yao and W. Jiang, Beilstein J. Org. Chem., 2019, 15, 1460–1467 CrossRef CAS PubMed .
  20. L.-L. Wang, Y.-K. Tu, A. Valkonen, K. Rissanen and W. Jiang, Chin. J. Chem., 2019, 37, 892–896 CrossRef CAS .
  21. E. R. Darzi and R. Jasti, Chem. Soc. Rev., 2015, 44, 6401–6410 RSC .
  22. A. Yagi, Y. Segawa and K. Itami, J. Am. Chem. Soc., 2012, 134, 2962–2965 CrossRef CAS PubMed .
  23. Z. Sun, P. Sarkar, T. Suenaga, S. Sato and H. Isobe, Angew. Chem., Int. Ed., 2015, 54, 12800–12804 CrossRef CAS PubMed .
  24. S. Hitosugi, W. Nakanishi, T. Yamasaki and H. Isobe, Nat. Commun., 2011, 2, 492 CrossRef .
  25. Z. Sun, T. Mio, T. Okada, T. Matsuno, S. Sato, H. Kono and H. Isobe, Angew. Chem., Int. Ed., 2019, 58, 2040–2044 CrossRef CAS PubMed .
  26. C.-F. Chen and Y. Han, Acc. Chem. Res., 2018, 51, 2093–2106 CrossRef CAS PubMed .
  27. J. R. Wu and Y.-W. Yang, Chem. Commun., 2019, 55, 1533–1543 RSC .
  28. M. De Rosa, C. Talotta, C. Gaeta, A. Soriente, P. Neri, S. Pappalardo, G. Gattuso, A. Notti, M. F. Parisi and I. Pisagatti, J. Org. Chem., 2017, 82, 5162–5168 CrossRef CAS PubMed .
  29. M. Xue, Y. Yang, X. Chi, Z. Zhang and F. Huang, Acc. Chem. Res., 2012, 45, 1294–1308 CrossRef CAS PubMed .
  30. T. Ogoshi, K. Kitajima, T. Aoki, S. Fujinami, T.-A. Yamagishi and Y. Nakamoto, J. Org. Chem., 2010, 75, 3268–3273 CrossRef CAS PubMed .
  31. T. Ogoshi, K. Masaki, R. Shiga, K. Kitajima and T. Yamagishi, Org. Lett., 2011, 13, 1264–1266 CrossRef CAS PubMed .
  32. J. Ji, Y. Li, C. Xiao, G. Cheng, K. Luo, Q. Gong, D. Zhou, J. J. Chruma, W. Wu and C. Yang, Chem. Commun., 2020, 56, 161–164 RSC .
  33. Y. Chen, L. Fu, B. Sun, C. Qian, R. Wang, J. Jiang, C. Lin, J. Ma and L. Wang, Org. Lett., 2020, 22, 2266–2270 CrossRef CAS PubMed .
  34. T. Ogoshi, R. Shiga, T. Yamagishi and Y. Nakamoto, J. Org. Chem., 2011, 76, 618–622 CrossRef CAS PubMed .
  35. J. Yao, W. Wu, W. Liang, Y. Feng, D. Zhou, J. J. Chruma, G. Fukuhara, T. Mori, Y. Inoue and C. Yang, Angew. Chem., Int. Ed., 2017, 56, 6869–6873 CrossRef CAS PubMed .
  36. T. Ogoshi, T. Akutsu, D. Yamafuji, T. Aoki and T.-A. Yamagishi, Angew. Chem., Int. Ed., 2013, 52, 8111–8115 CrossRef CAS PubMed .
  37. P. D. Sala, R. D. Regno, C. Talotta, A. Capobianco, N. Hickey, S. Geremia, M. D. Rosa, A. Spinella, A. Soriente, P. Neri and C. Gaeta, J. Am. Chem. Soc., 2020, 142, 1752–1756 CrossRef PubMed .
  38. M.-X. Wang, Sci. China: Chem., 2018, 61, 993–1003 CrossRef CAS .
  39. Q.-H. Guo, L. Zhao and M.-X. Wang, Chem. – Eur. J., 2016, 22, 6947–6955 CrossRef CAS PubMed .
  40. A. H. Tran, D. O. Miller and P. E. Georghiou, J. Org. Chem., 2005, 70, 1115–1121 CrossRef CAS PubMed .
  41. L.-P. Yang, F. Jia, Q.-H. Zhou, F. Pan, J.-N. Sun, K. Rissanen, L. W. Chung and W. Jiang, Chem. – Eur. J., 2017, 23, 1516–1520 CrossRef CAS PubMed .
  42. H. Chai, Z.-S. Pan, L.-P. Yang, S. He, F. Pan, K. Rissanen and W. Jiang, Chem. Commun., 2019, 55, 7768–7771 RSC .
  43. L.-P. Yang, S.-B. Lu, A. Valkonen, F. Pan, K. Rissanen and W. Jiang, Beilstein J. Org. Chem., 2018, 14, 1570–1577 CrossRef CAS PubMed .
  44. F. Jia, H.-Y. Wang, D.-H. Li, L.-P. Yang and W. Jiang, Chem. Commun., 2016, 52, 5666–5669 RSC .
  45. F. Jia, D.-H. Li, T.-L. Yang, L.-P. Yang, L. Dang and W. Jiang, Chem. Commun., 2017, 53, 336–339 RSC .
  46. F. Jia, L.-P. Yang, D.-H. Li and W. Jiang, J. Org. Chem., 2017, 82, 10444–10449 CrossRef CAS PubMed .
  47. L.-P. Yang, H. Liu, S.-B. Lu, F. Jia and W. Jiang, Org. Lett., 2017, 19, 1212–1215 CrossRef CAS PubMed .
  48. D.-H. Li, L.-P. Yang, H. Chai, F. Jia, H. Ke and W. Jiang, Org. Chem. Front., 2019, 6, 1027–1031 RSC .
  49. L.-P. Yang, F. Jia, F. Pan, Z.-S. Pan, K. Rissanen and W. Jiang, Chem. Commun., 2017, 53, 12572–12575 RSC .
  50. F. Jia, H. V. Schröder, L.-P. Yang, C. von Essen, S. Sobottka, B. Sarkar, K. Rissanen, W. Jiang and C. A. Schalley, J. Am. Chem. Soc., 2020, 142, 3306–3310 CrossRef CAS PubMed .


Electronic supplementary information (ESI) available: Equations and worked examples for calculating the numbers of stereoisomers. See DOI: 10.1039/d0cs00341g
These authors contributed equally.

This journal is © The Royal Society of Chemistry 2020