Introduction

Rotations about C-C single bonds (referred to as σ bonds) can affect three-dimensional structure of biologically important molecules, including DNA and proteins1,2,3,4,5,6,7, and often influence their physiological activity. At present, the molecular motions associated with simple σ bond rotations are reasonably well understood8,9,10,11,12,13,14,15. However, cooperative σ bond rotations involving more than one bond, as often observed in chemical and biological systems1,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30, are less well explored31. Although attracting attention in the context of, e.g., molecular rotors, control over multi-bond rotations in synthetic systems often rely on responsive behavior induced by external chemical32,33,34,35,36,37,38, electrochemical39,40,41,42,43, or photochemical stimuli44,45,46. Furthermore, there have been reported of convertibly atropisomeric macrocycles that can alter their conformation to a more stable state when subjected to an external stimulus. Examples of such macrocycles include cycloarylenes47,48, resorcinarenes49, biphen[n]arene50, polycyclic peptide51, and amide naphthotube52. Fast conformational conversion is widely recognized to be more common in synthetic macrocycles, which poses challenges in studying the process using common kinetic techniques. Although the conformational interconversions of artificial macrocycles have been investigated in the molecular recognition process involving guest molecules53,54,55, reports on slow conversion processes are still rarely reported52. The development of slow conformational exchange systems in macrocycles has multiple applications, including the design and synthesis of molecular machines and devices. Additionally, it serves as a valuable tool for analyzing host–guest binding mechanisms50. In fact, achieving reversible conformational transformations in the absence of an added species has proved challenging. Moreover, reversible σ bond rotation-based conversions between stable and metastable states of atropisomers are also rarely discussed. This has limited our understanding and precluded potential applications of molecular constructs whose three-dimensional structures can be modulated through conformational motion.

Here we report a macrocycle, octamethyl cyclo[4](1,3-(4,6)-dimethylbenzene)[4]((4,6-benzene)(1,3-dicarboxylate) (OC-4), that exists in stable C2v or C4v symmetric atropisomeric forms, C2v-OC-4 and C4v-OC-4, respectively, at 298 K. Different supramolecular chemical properties were seen for C2v- and C4v-OC-4. Whereas C4v-OC-4 acts as a fullerene receptor and could be used as a building block to construct pseudo-rotaxane structures via the binding of appropriately selected linear guests, no appreciable recognition chemistry was seen for C2v-OC-4. At 393 K, the conversion in 1,1,2,2-tetrachloroethane-d2 (TCE-d2) from C2v- to C4v-OC-4 produces an intermediate with Cs symmetry that is not present in detectable quantities in the original atropisomeric mixture (Fig. 1). This intermediate, Cs-OC-4, was isolated and characterized via a single crystal X-ray diffraction analysis in the solid state and by NMR spectroscopic methods in solution. To the best of our knowledge, this constitutes the limited example of an intermediate species captured in an experiment involving multiple σ bond rotations. Moreover, the transition from C2v- to C4v-OC-4 exhibited a notable influence from the solvent. The use of toluene-d8 as the solvent can promote the complete conversion (conversion percent larger than 99%) at a lower temperature of 373 K. Significantly, C4v-OC-4 has the ability to transform into C2v-, Cs-, and other forms of OC-4 in the TCE-d2 solution at temperatures of 373 or 393 K, or by further chemical reactions. However, the conversion of C4v-OC-4 did not occur when toluene-d8 was utilized, even at 373 K. The present work is thus expected to enhance our understanding of the dynamic features and structure—activity relationships of conformational mobile molecular systems, which should be beneficial for the development of stimuli-responsive molecules, supramolecular systems (such as molecular motors and devices), and stimuli-responsive smart materials.

Fig. 1: Schematic representation of intermediate-promoted reversible atropisomerism via heating or chemical reactions and related property changes.
figure 1

Also shown is the production via heating of an isomer that can serve as a building block for pseudrotaxane construction or as a fullerene receptor (left). Solvent effect induced irreversible conversion between stable and metastable species of OC-4 in toluene-d8 (right).

Full size image

Results

Synthesis and characterization of macrocycle octamethyl cyclo[4](1,3-(4,6)-dimethylbenzene)[4]((4,6-benzene)(1,3-dicarboxylate) (OC-4)

The macrocycle of this study, octamethyl cyclo[4](1,3-(4,6)-dimethylbenzene)[4]((4,6-benzene)(1,3-dicarboxylate) (OC-4), was generated via a combined fragment coupling and Suzuki-Miyaura reaction sequence (Fig. 2a). Various reaction conditions were evaluated sequentially (Supplementary Table 1). Cyclization in toluene at 373 K (condition i) gave C4v-OC-4 in a yield of 31% with little if any C2v-OC-4 being observed in this case. In acetonitrile at 333 K (condition ii), the atropisomers with C2v and C4v symmetry (i.e., C2v- and C4v-OC-4) were isolated from the reaction in 7% or 8% yield, respectively. Theoretical analyzes using semiempirical methods (PM7) were carried out. The results revealed that C4v-OC-4 is more stable than C2v-OC-4 (Supplementary Fig. 23 and Table 2), thus providing a rational preference for the isomer with C4v symmetry observed under both experimental conditions. Meanwhile, consider the corresponding cyclization Pd-catalyst contained intermediate (i.e., C4v-OC-4-im and C2v-OC-4-im), formation heat of C4v-OC-4-im was calculated and larger than that of C2v-OC-4-im. The finding implied that C2v-OC-4 as a kinetic product predominantly present under low-temperature condition (e.g., reaction condition ii).

Fig. 2: Synthesis and characteristics of macrocycle OC-4.
figure 2

a Synthesis of macrocycle OC-4. Reaction conditions: (I) 1 (1 equiv.), 3 (1 equiv.), CsF (25 equiv.) and Pd(dppf)2Cl2• CH2Cl2 (0.20 equiv.) in toluene, 373 K, 12 h, C2v-OC-4, trace, C4v-OC-4, 31%; (ii) 1 (1 equiv.), 3 (1 equiv.), CsF (25 equiv.) and Pd(dppf)2Cl2• CH2Cl2 (0.20 equiv.) in acetonitrile, 333 K, 48 h, C2v-OC-4, 7%, C4v-OC-4, 8%. b Structures and conformational studies of macrocycle OC-4. (b1) Heating induced quantitative conformational change from C2v- to C4v-OC-4 in solid state; (b2) and (b3), or (b4) and (b5), top and side views of C2v- or C4v-OC-4 with a stick form in the single crystal of [C2v-OC-4•5CYH] or [C4v-OC-4•3CYH], respectively. The insert is the thin layer chromatography (TLC) analysis on silica gel plates (eluent: petroleum ether: ethyl acetate = 3:1, v/v) of C2v-, C4v-OC-4, and the conversion at 573 K. c1–c2 1H NMR spectra of 5.00 mM C4v-OC-4 (c1) and C2v-OC-4 (c2) recorded in TCE-d2 at 298 K (500 MHz). d Expanded view of the temperature-dependent 1H NMR spectra of C2v-OC-4 (5.00 mM) in TCE-d2 (500 MHz).

Full size image

Macrocycles C2v- and C4v-OC-4 were fully characterized in TCE-d2 solution at 298 K via 1H and 13C NMR spectroscopy, correlation spectroscopy (COSY) and nuclear Overhauser effect spectroscopy (NOESY) (Fig. 2c and Supplementary Figs. 11-14 and Figs. 17-20). Both OC-4 isomers gave rise to peaks when subjecting to matrix assisted laser desorption ionization-time of flight high resolution mass spectrometric analysis that were the same within error (MALDI-TOF HRMS) (m/z 1207.4080 or 1207.4083, respectively; calcd m/z 1207.4092) (Supplementary Figs. 16, 22).

Macrocycles C2v- and C4v-OC-4 were further confirmed by single crystal X-ray diffraction analysis. In both cases, diffraction-grade crystals were grown by subjecting a CH2Cl2 and cyclohexane (CYH) (1:1, v/v) solution of the respective pure isomer to slow evaporation. The resulting structures revealed macrocyclic frameworks with “C2v-” or “hourglass-like” C4v- symmetry, respectively (Fig. 2b). In [C2v-OC-4•5CYH], the meso-dimethylbenzene units on C2v-OC-4 exist as parallel sets, while the ester groups are found in the four corners and point to the outside of the cavity (diameter = 7.8 Å) (Supplementary Figs. 31–33). The torsion angles between neighboring benzene rings range from 57° to 63°. In [C4v-OC-4•3CYH], the aromatic rings on C4v-OC-4 adopt a 1,3-alternate-like conformation and the ester groups all point to the same side of the cavity. A near circular cavity with a diameter about 8.7 Å is seen (Supplementary Figs. 34–36), with inter-ring angles between neighboring aromatic rings ranging from 84° to 112° being observed.

Thermal conversion of OC-4 with intermediate process

The fact that C2v-OC-4 was mainly obtained as a minor product under the lower temperature (333–353 K) reaction condition is consistent with it being a kinetic product and C4v-OC-4 being the thermodynamic product. Support for this supposition came from the theoretical analysis noted above, as well as the finding that C2v-OC-4 could be converted quantitatively to C4v-OC-4 in the solid state at high temperature (573 K) (Fig. 2b, Supplementary Fig. 39). Temperature-dependent 1H NMR spectroscopic studies carried out in TCE-d2 solution (5.00 × 10-3 M) from 233 K to 373 K revealed that both the C2v and C4v isomers of OC-4 were stable on the laboratory time scale at the temperatures lower than 343 K. However, evidence of conversion between C2v- and C4v-OC-4 was seen at temperatures ≥343 K (Fig. 2d, Supplementary Figs. 24, 25, 27, 28). Considering the close relationship between molecular conformations and temperatures/solvents, the stability of OC-4 isomers in toluene-d8 was investigated. Through temperature-dependent 1H NMR spectroscopic studies, it was observed that the transformation of C2v-OC-4 into C4v-OC-4 began at a relatively low temperature of 318 K (Supplementary Fig. 26), which is significantly lower than in TCE-d2 solution. Furthermore, at temperatures above 373 K, almost all of C2v-OC-4 converted into C4v-OC-4 and other forms of OC-4. These findings suggest that toluene-d8 promotes the conversion from C2v- to C4v-OC-4.

Time-dependent 1H NMR spectroscopic analyzes of C2v-OC-4 (5.00 × 10−3 M) carried out at 393 K in TCE-d2 to study the isomerization process (Figs. 3a, b, Supplementary Fig. 40). After 2 h, the signals corresponding to C4v-OC-4 were integrated to ca. 71.4% of the total whereas those for C2v-OC-4 accounted for only 5.6% of the total (Fig. 3c)56. Moreover, signals ascribable to new species were observed. Their overall ratio increased to a maximum of 49.2% with time before decreasing to a final equilibrium value of ca. 23.0%. These results lead us to propose that one or more intermediates are involved in the conversion process. Further analysis of the NMR data provided support for the possible formation of three intermediates, namely isomers of OC-4 with C1-, Cs-, and C2-symmetry, respectively. Similarly, the transformation from C2v- to C4v-OC-4 was carried out at 373 K in toluene-d8 (note: the concentration was 1.00 × 10−3 M due to its low solubility in toluene-d8 (about 1.00 × 10−3 M)) (Supplementary Fig. 50). After 3 h, nearly all C2v-OC-4 are converted to C4v- (97.5%) and various forms of OC-4 (2.5%). The signals corresponding to three intermediates reached a maximum total proportion of 70.2% after 0.25 h, before dropping to a final equilibrium value of 2.5% (Supplementary Figs. 50, 51). For comparison purposes, the conversion process was also conducted in TCE-d2 using the same concentration (1.00 × 10−3 M) and temperature (373 K). In this case, the percent conversion of C2v-OC-4 was lower, reaching 86.9% after 3 h (Supplementary Figs. 53, 54). The decrease in conversion temperature to 373 K and the increase in percent conversion to more than 99% of C2v-OC-4 in toluene-d8 suggest that the transformation from C2v- to C4v-OC-4 may be affected by the choice of solvents.

Fig. 3: Atropisomer conversion process of OC-4.
figure 3

a Conversion of C2v- to C4v-OC-4 at 393 K in TCE-d2. b Expanded view of time-dependent 1H NMR (700 MHz) spectra of C2v-OC-4 (5.00 × 10−3 M) in TCE-d2 after warming at 393 K. c Time-dependent speciation plots for C2v- (red dot and line), C1- (purple dot and line), Cs- (green dot and line), C2- (orange dot and line) and C4v-OC-4 (blue dot and line) after heating C2v-OC-4 (5.00 × 10−3 M) at 393 K. d Schematic representation of the method used to obtain single crystals of intermediate Cs-OC-4. e Top and side views of Cs-OC-4 of the single crystal structure of [Cs-OC-4•n-hexane] in stick form.

Full size image

Efforts were made to capture the presumed intermediate(s) generated during the C2v- to C4v-OC-4 isomerization process. Toward this end, C2v-OC-4 (10 mg) was dissolved in TCE-d2 (0.6 mL) and placed into a NMR tube (5 mm) (Fig. 3d). A relatively large integration for the presumed intermediates (a relative ratio of up to 49.2%) is seen after heating the solution at 393 K for 15 min. After allowing the mixture to cool to room temperature, the solution was transferred to a clean vial. A super silent adjustable air pump was then used to remove the TCE-d2 solvent and any other volatiles while keeping the temperature constant. The resulting solid residue was dissolved in a mixture of CH2Cl2 and CH3OH (1.5 mL; 1:1, v/v). After slow evaporation at 298 K for 1 day, diffraction-grade single crystals were obtained. An ensuing X-ray diffraction analysis revealed the presence of three sets of crystals corresponding to C2v-OC-4, C4v-OC-4, and a new OC-4 species with Cs symmetry (i.e., Cs-OC-4) (Fig. 3e and Supplementary Figs. 57–59). In Cs-OC-4, the benzene units adapt a “down-up-up-down-up” orientation. The torsion angles between adjacent benzene units range from 55° to 110° (Supplementary Fig. 58). To the best of our knowledge, Cs-OC-4 constitutes the limited structurally characterized intermediate to be captured during an atropisomerization process involving presumably cooperative rotations about single C-C σ bond.

We further attempted to collect pure sample of Cs-OC-4 from the crystalline mixture of Cs-, C2v- and C4v-OC-4 for solution phase NMR spectral analysis. To ensure purity, the Cs-OC-4 crystals were separated mechanically and their integrity were checked by measuring the cell parameters crystal by crystal by X-ray diffraction analysis. A small amount (less than 0.1 mg) of Cs-OC-4 was collected in this way. The resulting sample was washed with petroleum ether and CYH three times, respectively, and dried at room temperature for 24 h to remove residue solvents (e.g., CH2Cl2, CYH and petroleum ether). The Cs-OC-4 samples were then dissolved in TCE-d2 and subjected to 1H, COSY and NOESY spectral analysis at 298 K (Supplementary Figs. 60–63). The main peak pattern proved consistent with the structure of Cs-OC-4 elucidated via the single crystal diffraction analysis. However, two small sets of signals were observed with the same integral ratios as 9.7% (Supplementary Fig. 61). On this basis we suggest that isomers with pseudo-C2 and C1 symmetry coexist with Cs-OC-4 in TCE-d2 solution. Support for this proposed coexistence came from theoretical calculation (cf. Fig. 4 and Supplementary Figs. 52, 55, 56).

Fig. 4: Potential energy analysis of atropisomer conversion corresponding to OC-4.
figure 4

Potential energy diagram for the conversion of C2v-OC-4 to C4v-OC-4 from the time-dependent 1H NMR spectral studies of C2v-OC-4 (5.00 × 10−3 M) in TCE-d2 at 393 K. The transition states, as well as those for C1-OC-4 and C2-OC-4, are listed based on the theoretical calculations.

Full size image

When Cs-OC-4 was allowed to sit in TCE-d2 solution at room temperature for 12–40 h (Supplementary Fig. 64), relatively weak signals corresponding to both C2v- and C4v-OC-4 were observed in the 1H NMR spectrum. Time-dependent 1H NMR spectroscopic studies of Cs-OC-4 were then carried out at 393 K in TCE-d2 solution. As shown in Supplementary Fig. 66, the Cs-OC-4 ratio decreased to 15.8% at final equilibrium within 30 min. Meanwhile, the ratio of C2v-OC-4 increased to a maximum of 14.3% at 4 min, and then decreased to 4.1% at equilibrium. Over the course of these studies, the proportion of C1-OC-4 and C2-OC-4 stayed roughly identical with 1.9% at final equilibrium, while the population of C4v-OC-4 increased to 76.4% by the time equilibrium was reached. In a separate experiment, C4v-OC-4 in TCE-d2 solution was held at 393 K for 12 h (Supplementary Fig. 67). Based on 1H NMR spectral integrations, the initially pure sample was converted to a mixture containing the C2v-OC-4 (5.1%), Cs-OC-4 (17.5%), C1-OC-4 (2.9%) and C2-OC-4 (2.9%) forms, in addition to the thermodynamically favored C4v-OC-4 isomer (Supplementary Fig. 68). A similar final ratio of each OC-4 form was seen when C2v-OC-4 or Cs-OC-4 were warmed under identical conditions. However, there is no if little change observed when C4v-OC-4 is heated at 373 K for 12 h in a toluene-d8 solution (Supplementary Fig. 69). This suggests that the conversion from C2v- to C4v-OC-4 in the toluene-d8 is irreversible.

With time-dependent 1H NMR spectroscopic detailed analyzes of C2v-OC-4 (5.00 × 10−3 M) carried out in TCE-d2 at 393 K (Fig. 3b), the relationship between the concentration ratio of each OC-4 form and time (t) could be expressed as below (for details see the Supplementary equations 1–25 and Figs. 41–48):

$$[{{{{\boldsymbol{C}}}}}_{2{{{\boldsymbol{v}}}}}-{{{\bf{O}}}}{{{\bf{C}}}}-{{{\mathbf{\ 4}}}}]\%=5.6205+94.3795\times \exp \left(\frac{-0.142{{{\rm{t}}}}}{94.3795}\right)$$
(1)
$$[{{{{\boldsymbol{C}}}}}_{{{{\boldsymbol{s}}}}}-{{{\bf{O}}}}{{{\bf{C}}}}-{{{\mathbf{\ 4}}}}]\%=0.736\times \frac{94.37\times (1-{e}^{{{{\rm{\gamma }}}}})\times ({{{\rm{\alpha }}}}+{{{\rm{\beta }}}}+0.31754)}{{{{\rm{\alpha }}}}+{{{\rm{\beta }}}}+1.31754}$$
(2)
$$[{{{{\boldsymbol{C}}}}}_{{{{\mathbf{\ 1}}}}}-{{{\bf{O}}}}{{{\bf{C}}}}-{{{\mathbf{\ 4}}}}]\%=[{{{{\boldsymbol{C}}}}}_{{{\mathbf{\ 2}}}}-{{{\bf{O}}}}{{{\bf{C}}}}-{{{\mathbf{\ 4}}}}]\%=0.132\times \frac{94.37\times (1\,-\,{{{{\rm{e}}}}}^{{{{\rm{\gamma }}}}})\times ({{{\rm{\alpha }}}}+{{{\rm{\beta }}}}+0.31754)}{{{{\rm{\alpha }}}}+\,{{{\rm{\beta }}}}+1.31754}$$
(3)
$$[{{{\boldsymbol{C}}}}{{{\mathbf{\ 4}}}}_{{{\boldsymbol{v}}}}-{{{\boldsymbol{OC}}}}-{{{\mathbf{\ 4}}}}]\%=94.37\times (1-{{{{\rm{e}}}}}^{{{{\rm{\gamma }}}}})-[{{{{\boldsymbol{C}}}}}_{{{{\boldsymbol{s}}}}}-{{{\mathbf{\ OC}}}}-{{{\mathbf{\ 4}}}}+{{{{\boldsymbol{C}}}}}_{{{{\mathbf{\ 1}}}}}-{{{\mathbf{\ OC}}}}-{{{\mathbf{\ 4}}}}+{{{{\boldsymbol{C}}}}}_{{{{\mathbf{\ 2}}}}}-{{{\mathbf{\ OC}}}}-{{{\mathbf{\ 4}}}}]$$
(4)

Here, α, β and γ are functions of heating time (t) as below:

$$\alpha=1.6161\times \exp \left(\frac{-{{{\rm{t}}}}}{1102.30398}\right)$$
(5)
$${{{\rm{\beta }}}}=19.55909\times {{{\rm{exp}}}}\left(\frac{-{{{\rm{t}}}}}{326.3319}\right)$$
(6)
$${{{\rm{\gamma }}}}=\frac{-0{{{\rm{.0014t}}}}}{0.9437}$$
(7)

Thus, the thermodynamic and kinetic parameters of the conversion could be calculated in accord with the equilibria shown below (equation 8). The equilibrium constants (K), rate constants (k), Gibbs free energies (∆Gθ) and Gibbs activation energy (-∆G≠) between each other are given in Table 1. Setting the formation energy of C2v-OC-4 at 0.0 kcal, an energy diagram for the interconversion between the C2v- and C4v-OC-4 forms in TCE-d2 at 393 K could be constructed. It is shown in Fig. 4.

Table 1 Reaction equilibrium constants (K)a calculated from the experimental data, forward rate constants (k), Gibbs free energies (∆Gθ)a, Gibbs activation free energies (-∆G≠)a; and theoretical computationalb,c formation energy values (E and ΔE) for the proposed equilibria involved in the conversion of C2v- to C4v-OC-4 at 393 K
Full size table

Further support for the observed conformational conversions came from theoretical studies involving PM7 analyzes. These analyzes revealed that 1) Cs-OC-4 and C4v-OC-4 are energetically more stable than C2v-OC-4 and 2) a small energy barrier (≤10 kcal mol−1) persists between Cs-OC-4 and C2v-OC-4 or C4v-OC-4 (Supplementary Fig. 56).

The calculations also confirmed that C1- and C2-OC-4, whose existence as minor intermediates was inferred from the spectral studies, are less stable than C2v-OC-4. Furthermore, the conformation optimization provided the possible structures for C1-OC-4, C2-OC-4 and other OC-4 transient state forms. These simulated structures and the results calculated from the experimental data shown above allowed the profiles corresponding to the proposed conversion of C2v-OC-4 to C4v-OC-4 to be obtained (Fig. 4).

Conversion of C 4v - to C 2v -OC-4 or C s -OC-4 promoted by chemical reactions

The reversible interconversion between OC-4 forms led us test whether C2v-OC-4 could be produced from C4v-OC-4 more efficiently than by simple heating in TCE-d2 solution (Fig. 5). With this goal in mind, pure C4v-OC-4 was subjected to hydrolysis to give the corresponding acid, cyclo[4](1,3-(4,6)-dimethylbenzene)[4]((4,6-benzene)(1,3-dicarboxylic acid) (CA-4). It was found that the acid product produced in this way included several forms (i.e., C4v-, C2v-CA-4, etc.) (Fig. 5a and Supplementary Figs. 70–72). Herein, different MOH (M = Li, Na, K, Rb or Cs) salts and tetra-n-butylammonium hydroxide (TBAH) were tested in the initial hydrolysis step prior to reesterification. Significant countercation effects were observed (Supplementary Fig. 70). The generated CA-4 mixture without further purification was then treated with tetra-n-butylammonium fluoride (TBAF) and iodomethane at room temperature for 48 h57. Specifically, the final C2v-, Cs- and C4v-OC-4 ratio proved highly dependent on the countercation species employed (Fig. 5b). The overall conversion ratio of C4v-OC-4 was relatively high in the case of NaOH (70.5%), KOH (63.1%) or RbOH (65.1%) (Fig. 5d). The tested organic base TBAH used in hydrolysis promoted largest conversion ratio of C4v-OC-4 as 90.0% (including C2v- and Cs-OC-4, 33.6% and 50.3% yield, respectively). The total conversion rates of C4v-OC-4 are much higher when compared to direct heating the TCE-d2 solution of C4v-OC-4 at 393 K (conversion rate as 28.4%). This suggests that the chemical reactions may facilitate the reversible conversion of C4v-OC-4 into different forms.

Fig. 5: Chemical reaction promoted reversible atropisomer transformation of OC-4.
figure 5

a Schematic representation of molecular structure conversion from C4v- to C2v-OC-4 via subjecting to hydrolysis and esterification with different alkali hydroxide salts; b Expanded view of 1H NMR spectra of OC-4 recorded in TCE-d2 after esterification following treatment of CA-4 with different alkali hydroxide salts (500 MHz); c Structure of [C2v-CA-4-3H+]3-•Na+•K+•3Cl-•DMF and C4v-CA-4 shown in stick form obtained from a single crystal structural analysis of [[C2v-CA-4-3H+]3-•Na+•K+•3Cl-•DMF]•4H3O+•9.5H2O•5DMF and [C4v-CA-4•2DMF•H2O], respectively; d Isomer fraction of various OC-4 forms (red column: C2v-OC-4; purple column: C1-OC-4; green column: Cs-OC-4; orange column: C2-OC-4; blue column: C4v-OC-4) after reesterification following C4v-OC-4 hydrolysis with different alkali hydroxide salts, error bars correspond to S.D.

Full size image

Deeper insight into the countercation effects on the above conversion came from isolation of a putative intermediate. Here, C4v-OC-4 was subjected to hydrolysis with NaOH, followed by protonation with HCl (10 equiv.), and then redissolved in KOH solution (DMF/H2O (1: 1, v/v)) or dissolved directly in DMF/H2O (1: 1, v/v). The resulting products were allowed to undergo slow evaporation at room temperature. After two weeks, diffraction-grade single crystals were obtained. One structure is [C4v-OC-4•2DMF•H2O]. The X-ray diffraction data of another structure revealed a ratio of 1:1:1:3 between C2v-CA-4, Na+, K+, and Cl-. When considering the charge balance and the formation of carboxylate anions in metal complexation, it is recommended that four protons should distribute among the solvent water molecules. This is because water is a stronger Brønsted-Lowry base compared to Cl-. However, it should be noted that the protons on hydronium ions cannot be precisely located due to their disorder and technical limitations. Finally, this single crystal structure was suggested as a charge-neutral ionic form with [[C2v-CA-4-3H+]3-•Na+•K+•3Cl-•DMF]•4H3O+•9.5H2O•5DMF, rather than the conventional form as C2v-CA-4•NaCl•KCl•HCl•13.5H2O•6DMF. This is because the former representation provides more detailed information about the structure and coordination (Fig. 5c). In [[C2v-CA-4-3H+]3-•Na+•K+•3Cl-•DMF]•4H3O+•9.5H2O•5DMF, the resulting structure showed that K+ could effectively bridge the O atoms of neighboring carboxyl groups on C2v-CA-4. The associated stabilization was by theoretical calculations (for details see the Supplementary Tables 9–16). The putative stabilization of other alkali metal cations was studied by theory and, as expected, revealed differing degrees of thermodynamic stability for each cation-CA-4 isomer pair that as a general rule were consistent with the relative ratios seen by experiment. This finding could be advantageous in the development of molecular machines, devices, as well as stimuli-responsive smart materials.

Host–guest properties of OC-4

The different shapes seen for C4v-OC-4 and C2v-OC-4 led us to explore whether differences in their molecular recognition features might be observed. Initial studies involved probing the interactions between these two OC-4 forms and the linear guests 1,8-dibromooctane (G1), octane-1,8-di-thiol (G2) and 1,9-decadiyne (G3). For this study, mixtures containing C2v-OC-4 (20 mM) and 1 molar equal of the guest in question, either G1, G2 or G3, were studied via 1H NMR spectroscopy in CDCl3/CD3OD (1:2, v/v) (note: In all of the tested cases, the discussed solution system showed minimal solvent effects, meaning that the host–guest response was at its highest level.) (Supplementary Figs. 80–82). Little or no changes in the proton signals, either of the guests or the macrocycle, were seen in the case of C2v-OC-4. In contrast, notable shifts in key signals were seen in the case of C4v-OC-4 (Supplementary Figs. 83, 86, 89). 1H NMR spectroscopic Job-plot analyzes provided support for a 1:1 ([H]/[G]) binding stoichiometry for all three substrates (Supplementary Figs. 85, 88, 91). Further evidence for the formation of 1:1 complexes, C4v-OC-4⊃guest (where guest = G1, G2 or G3), came from electrospray ionization high-resolution mass spectrometric (ESI-HRMS) analyzes, the expected peaks were observed (Supplementary Figs. 92–94). Association constants (Ka) of (4.7 ± 0.5) M−1, (6.2 ± 0.6) × 10 M−1 or (6.0 ± 0.6) × 10 M−1 were seen for C4v-OC-4 and guests G1, G2 or G3, respectively.

The above pseudo-rotaxane complexes were further characterized via single crystal X-ray diffraction analyzes (Fig. 6a). Diffraction-grade single crystals were obtained by slow evaporation of mixtures containing C4v-OC-4 and guests G1, G2 or G3 in CH2Cl2/CH3OH (1:2, v/v), respectively. In the solid state, these linear guest thread through the cavity of C4v-OC-4 (Supplementary Figs. 95–97). A particularly notable [3]pseudo-rotaxane complex, 2C4v-OC-4⊃n-eicosane, was obtained using n-eicosane (G4) as guest. Short separations (around 3.8 Å) between the carbon atoms of the linear guests and the carbon atoms located on the benzene units of C4v-OC-4 were observed. These findings are consistent with the pseudo-rotaxane structures being stabilized via C–H···π interactions.

Fig. 6: Single crystal X-ray structures of the complexes involving C4v-OC-4 and linear guests or fullerenes.
figure 6

a Single crystal X-ray molecular structure of the complexes formed from C4v-OC-4 and 1,8-dibromooctane (G1), octane-1,8-dithiol (G2), 1,9-decadiyne (G3), and n-eicosane (G4). These species crystallize as [C4v-OC-4•1,8-dibromooctane], [C4v-OC-4•octane-1,8-dithiol], [C4v-OC-4•1,9-decadiyne], and [2C4v-OC-4⊃n-eicosane], respectively. b Single crystal X-ray diffraction structure of C4v-OC-4כC60 and C4v-OC-4כC70, species that crystalize as [C4v-OC-4•C60•3toluene•2THF] and [C4v-OC-4•C70•2toluene], respectively.

Full size image

In a further study, explored whether C2v-OC-4 or C4v-OC-4 would act as a fullerene receptor. In fact, 1H NMR spectral responses were observed after adding 1 molar equiv. of C60 or C70 to TCE-d2 solutions of C4v-OC-4 (note: the TCE-d2 was used due to its preferable solubility for OC-4 and fullerenes) (Supplementary Fig. 100). In contrast, little or no response was seen in the case of C2v-OC-4 (Supplementary Fig. 99). We rationalize this difference in binding propensity to the fact that C4v-OC-4 has a more open cavity and is thus better able to act as an endoreceptor for the fullerene guest. In TCE-d2, 1H NMR spectral titrations provided support for a 1:1 binding stochiometry and Ka values of (5.9 ± 0.6) × 103 M−1 and (5.2 ± 0.5) × 103 M−1 for the interaction between C4v-OC-4 and C60 and C70 (Supplementary Figs. 102, 103), respectively. These results basically coincided with the calculated (Ka) values (2.4 ± 0.2) × 103 M−1 and (2.2 ± 0.2) × 103 M−1 between C4v-OC-4 and C60 or C70 via UV-vis titrations (Supplementary Figs. 104, 105). Further evidences for the formation of 1:1 complex between C4v-OC-4 and C60 or C70 came from MALDI-TOF HRMS analysis (Supplementary Figs. 106, 107). Single crystal structures of C4v-OC-4כC60 and C4v-OC-4כC70 were determined via single crystal X-ray diffraction analysis (Fig. 6b and Supplementary Figs. 108–111). Both structures revealed 1:1 composition similar to what was inferred from the solution phase studies. Based on the metric parameters (i.e., the ≤3.8 Å separation between the carbon atoms of fullerenes C60 or C70 and those of C4v-OC-4), we suggest that complex formation is driven in part by stabilizing π···π donor-receptor interactions.

Theoretical calculations were further performed to obtain additional insight into the interactions between C4v-OC-4 and representative linear guests, C60, and C70 (Supplementary Tables. 19-22). The lowest energies of the limiting outside and interpenetrated binding modes involving C4v-OC-4 and each guest were considered in vacuum using the MM+ methods included in the HyperChem 8.0 program or the PM7 method available in the MOPAC program. The ΔEin−out values provide support for the notion that the insert mode is more stable than various hypothetical outside binding modes.

Discussion

In summary, the macrocyclic atropisomers, C2v-OC-4 and C4v-OC-4, were studied in an effort to understand synergetic rotation processes involving multiple single σ bonds. A thermodynamic metastable state, C2v-OC-4, was found to convert to a more stable state, C4v-OC-4, at 373 or 393 K. Under the conditions of conversion, Cs-OC-4, could be isolated as an active intermediate. This intermediate was characterized via a combination of solid state single crystal X-ray diffraction and NMR spectroscopic analyzes. Reversible atropisomerization between C2v- and C4v-OC-4 could induced by heating. Moreover, simple chemical reactions could be employed to induce the otherwise thermodynamically unfavorable conversion of C4v-OC-4 to C2v- and Cs-OC-4. To the best of our knowledge, the present study provides the limited example where a stable rigid molecular species is transformed into metastable products through simple heating or chemical reactions. We thus believe that further studies of these and other model systems might provide insights into σ bond rotations and the changes in structure and function they can induce.

Methods

General considerations

All reagents were purchased commercially (Aldrich, Acros, or Fisher) and used without further purification. Deuterated solvents were purchased from Cambridge Isotope Laboratory (Andover, MA). NMR spectra were recorded on Bruker AVANCE 400, Bruker AVANCE III 500, Bruker NEO 700 or 600 JNM-ECZ 400 S spectrometers. The 1H and 13C NMR chemical shifts are reported relative to residual solvent signals (1H: CDCl3 at 7.26 ppm, TCE-d2 at 5.95 ppm, methanol-d4 at 3.31 ppm, or dimethyl sulfoxide-d6 (DMSO-d6) at 2.50 ppm. 13C: CDCl3 at 77.2 ppm, TCE-d2 at 74.10 ppm. High resolution mass spectra were detected on Bruker Solarix FT-ICR-MS (ESI, EI), Bruker autoflex speed MALDI-TOF or AB SCIX Triple TOF 5600 apparatus. UV-vis spectra were collected on a Shimadzu UV-2450 instrument.

Single crystal X-ray diffractions

Unless otherwise noted, all single crystals used to obtain the X-ray diffraction structures grew as colorless prisms or blocks. The crystals used for single crystal X-ray diffraction analyzes were cut from clusters of the corresponding crystals. The data were collected on Saturn724+ (2 × 2 bin mode) or SuperNova, Dual, Cu at Home/Near, AtlasS2 diffractometers. Data reduction was performed using CrystalClear or CrysAlispro (Rigaku OD) software packages. The structures were refined by full–matrix least–squares on F2 with anisotropic displacement parameters for the non-H atoms using SHELXL-2014 or SHELXL-201858. The hydrogen atoms were calculated in idealized positions with isotropic displacement parameters set to 1.2 × Ueq of the attached atom (1.5 × Ueq for methyl hydrogen atoms). Definitions used for calculating R(F), Rw(F2) and the goodness of fit, S, are given below and in the.cif documents. Neutral atom scattering factors and values used to calculate the linear absorption coefficient are from the International Tables for X-ray Crystallography (1992). All ellipsoid figures were generated using SHELXTL/PC.

Synthesis of OC-4

Under Ar atmosphere, trimer 3 (0.50 g, 0.76 mmol), dimethyl 4,6-dibromoisophtalate 1 (0.27 g, 0.76 mmol), base (25.0 equiv., 19.1 mmol), Pd(dppf)2Cl2• CH2Cl2 (0.125 g, 0.153 mmol) and 100 mL solvent (toluene or acetonitrile) were added to a 250 mL three-necked, round-bottomed flask. The mixture was stirred at 333–373 K for 12–48 h as Supplementary Table 1. After cooling to room temperature, the resulting mixture was concentrated via reduced pressure with evaporator under corresponding reaction temperatures (Supplementary Table 1). The residue was dissolved in 150 mL CH2Cl2 and washed with water (50 mL × 3). The organic layer was collected and dried with anhydrous Na2SO4. After removing the solvent via rotary evaporation, the crude product was purified by column chromatography on silica gel with a mixed solvent (PE: EA = 4:1, v/v) as the eluent to give out C2v-OC-4 and C4v-OC-4 as white solid, respectively, which further recrystallized in CH2Cl2/cyclohexane (CYH) (1:1, v/v) to afford pure C2v-OC-4 (m. p. >573 K) and C4v-OC-4 (m. p. > 573 K), respectively. C2v-OC-4: 1H NMR (700 MHz, TCE-d2, 298 K) δ (ppm): 8.53 (s, 4H), 7.41 (s, 4H), 7.15 (s, 4H), 7.00 (s, 4H), 3.71 (s, 24H), 1.98 (s, 24H). 13C NMR (175 MHz, TCE-d2, 298 K) δ (ppm): 167.2, 145.8, 138.7, 135.3, 135.2, 131.9, 130.8, 129.8, 129.4, 52.5, 19.5. FT-IR (cm-1): 1726, 1604. MALDI-TOF HRMS (m/z): [M+Na]+• calcd for C72H64NaO16, 1207.4092; found, 1207.4080. C4v-OC-4: 1H NMR (700 MHz, TCE-d2, 298 K) δ (ppm): 8.38 (s, 4H), 7.03 (d, 8H), 6.68 (s, 4H), 3.66 (s, 24H), 2.03 (s, 24H). 13C NMR (175 MHz, TCE-d2, 298 K) δ (ppm): 167.0, 145.3, 137.4, 134.3, 134.2, 131.9, 130.9, 129.8, 127.8, 52.4, 19.9. FT-IR (cm-1): 3449, 1738, 1716, 1603. MALDI-TOF HRMS (m/z): [M+Na]+• calcd. for C72H64NaO16, 1207.4092; found, 1207.4083.

Intermediate capture and characterization

C2v-OC-4 (10 mg) were dissolved in TCE-d2 solution (0.6 mL) and placed into an NMR tube. A maximum number of intermediates were formed after heating at 393 K for 15 min. Then cooling the mixture to room temperature, the residual solution was transferred to a 5 mL vial and treated with room temperature airflow to remove TCE-d2 solvent. The dried sample was dissolved in the mixture of CH2Cl2/CH3OH (1.5 mL; 1:1, v/v). Diffraction grade single crystal of Cs-OC-4 was obtained via siting this mixture on the bench for 1 day at 298 K. Three kinds of crystals (Cs-, C2v- and C4v-OC-4) were generated at the same time. Cs-OC-4 crystals were further separated through one by one checking cell parameters of small crystal samples under the aid of a single crystal diffractometer. The separated Cs-OC-4 crystals were collected and further washed with PE and CYH three times, respectively. After drying at room temperature for 24 hours, Cs-OC-4 crystal samples (less than 0.1 mg) were immediately dissolved in TCE-d2 (0.6 mL) for NMR analysis at 298 K. Cs-OC-4: 1H NMR (700 MHz, TCE-d2, 298 K) δ (ppm): 8.49 (s, 2H), 8.42 (s, 2H), 7.25 (s, 2H), 7.15 (s, 2H), 7.09 (s, 2H), 7.01 (s, 1H), 6.99 (s, 1H), 6.96 (s, 1H), 6.91 (s, 1H), 6.79 (s, 2H), 3.73 (s, 6H), 3.69 (s, 6H), 3.66 (s, 6H), 3.56 (s, 6H), 2.18 (s, 6H), 2.02 (s, 6H), 1.99 (s, 6H), 1.94 (s, 6H).

Hydrolysis reaction of C 4v -OC-4 to obtain C 2v -, C s -, C 4v - and other CA-4 forms

Pure C4v-OC-4 (100 mg), 100 molar equiv. of MOH (M: Li, Na, K, Rb or Cs) or tetrabutylammonium hydroxide (TBAH), and 25 mL mixture of ethanol/H2O (v/v, 1/1) were added into 100 mL three-necked flask with magneton. The mixture was refluxed for 24 hours. After cooling to room temperature, 1 M HCl water solution was dropwise added to the mixture until pH as 3. The suspended solids were collected by centrifugation. After drying, CA-4 was obtained as white solid which is unpurified and direct used for further analysis and esterification reaction. The hydrolysis product was the mixture containing C2v-, C4v-, Cs- and other form of CA-4.

Esterification of CA-4 to obtain corresponding OC-4

CA-4 solid mixture without more treatment (50.0 mg), CH3I (66.1 mg), tetrabutylammonium fluoride (TBAF) (1 mol/L in THF, 466 μL) and THF (5 mL) were added into 25 mL flask with magneton. The mixture was stirred at room temperature for 48 h. The resulting solution was concentrated under reduced pressure via evaporator with temperature lower than 343 K. The crude product was filtered with silica plug to remove residual CH3I and TBAF with a mixed solvent (PE/EA = 4:1, v/v) as the eluent, collecting the eluent and then removing the solvent under reduced pressure with evaporator under 343 K give out OC-4 mixture as white solid.

Theoretical calculations

The molecular mechanics (MM + ) force field in the HyperChem 8.0 program and/or semiempirical methods (PM7) in the MOPAC software were used to calculate the energy values of different forms of OC-4, CA-4, corresponding transition states and complexes. Related single crystal data discussed in the main text were utilized as the starting point for the computational modeling studies.