Post-synthetic modification of a macrocyclic receptor via regioselective imidazolium ring-opening

Shown in graphic form is the use of mild basic conditions to effect the post-synthetic, ring-opening based modification of a tetraimidazolium macrocycle known as the “Texas box”. This ring opening modulates the intrinsic binding characteristics of the system.

In a typical procedure, a 25ml small vial was charged with 50 mg (0.041 mmol) 1 4+ ·4PF 6 − (1 4+ : cyclo [2](2,6-di(1H-imidazol-1-yl)pyridine) [2](1,4-dimethylenebenzene)) in a mixture containing 1 mL aqueous ammonia (between 25-28% by mass), 3.5 mL water and 4.5 mL acetonitrile. This solution was stirred hermetically at 313K for 24h. The reaction was then cooled to room temperature and acetonitrile was removed via blowing with compressed air, which gave rise to a light yellow precipitate (the crude product). The product was then collected via filtration and washing with 2 ml of water. Herein, reference to %NH 3 ·H 2 O is meant to represent the volume fraction of NH 3 ·H 2 O (25-28% by mass) used in the final solution. The reaction solvent consists of acetonitrile (4.5 mL) and dilutes aqueous ammonia (4.5 mL). The latter was obtained by mixing concentrated aqueous ammonia (25-28% by weight) with water to give the final stated concentration of ammonia. The ratio of the initial organic and aqueous phases was 1:1 in all experiments. It is suggested that the crude product mainly contains 1 4+ ·4PF 6 − , the single ringopened product I 3+ ·3PF 6 − , the two ring opened product II 2+ ·2 PF 6 − (i.e., 2 2+ ·2PF 6 − and 3 2+ ·2PF 6 − ) (cf. Figure S2 and S3). Figure S1. 1 H NMR spectra recorded in DMSO-d 6 (400 MHz, 300K) representing the crude products obtained from the ring-opening reaction of 1 4+ ·4PF 6 − using different concentrations of NH 3 ·H 2 O at 288K for 45h. Here NH 3 ·H 2 O% represents the volume fraction of NH 3 ·H 2 O (25-28% by weight) in the final solution. The pH value shown is that of the initial solvent system. Figure S2. The expanded 1 H NMR spectrum recorded in DMSO-d 6 (400 MHz, 300K) of the crude product obtained from the reaction wherein the initial solvent pH was 11.64 and the reaction was run at 288K for 45 h. The proton signals of the C-H protons between the two nitrogen atoms on the imidazolium rings of 1 4+ (labeled as "▼"), I 3+ (labeled as "■") and II 2+ (labeled as "•"). The relative amounts of each independent component in the crude product were obtained from the integrated areas. Figure S3. The full view (a) and expanded view (b) of the ESI high resolution mass spectrum of the crude product obtained from the ring-opening reaction of 1 4+ ·4PF 6 − which was conducted at 288K for 45h at an initial pH of 11.64. It is implied that the crude product contained at least three components 1 4+ , I 3+ and II 2+  S8 Figure S4. 1 H NMR spectra recorded in DMSO-d 6 (400 MHz, 300K) representing the crude products obtained from the ring-opening reaction of 1 4+ ·4PF 6 − using different concentrations of NH 3 ·H 2 O at 288K for 96 h. Here, NH 3 ·H 2 O% represents the volume fraction of NH 3 ·H 2 O (25-28% by weight) in the final solution. The pH value shown is that of the initial solvent system. Figure S5. 1 H NMR spectra recorded in DMSO-d 6 (400 MHz, 300K) representing the crude products obtained from the ring-opening reaction of 1 4+ ·4PF 6 − using different concentrations of NH 3 ·H 2 O at 288K for 120h. Here NH 3 ·H 2 O% represents the volume fraction of NH 3 ·H 2 O (25-28% by weight) in final solution. The pH value shown is that of the initial solvent system. S9 Figure S6. 1 H NMR spectra recorded in DMSO-d 6 (400 MHz, 300K) representing the crude products obtained from the ring-opening reaction of 1 4+ ·4PF 6 − using different concentrations of NH 3 ·H 2 O at 298K for 45 h. Here NH 3 ·H 2 O% represents the volume fraction of NH 3 ·H 2 O (25-28% by weight) in the final solution. The pH value shown is that of the initial solvent system. Figure S7. Full view (a) and expanded view (b) of the ESI high resolution mass spectrum of the crude product obtained from the ring-opening reaction of 1 4+ ·4PF 6 − , which was conducted at 298K for 45 h at an initial pH of 10.40. It was found that signals for 1 4+   [a] Volume fraction of NH 3 ·H 2 O used in the final solution. The mixture of reaction solvents consisted of acetonitrile (4.5 mL) and dilute aqueous ammonia (4.5 mL). The latter was obtained by mixing concentrated aqueous ammonia (25-28% by weight) with water to give the final stated concentration of ammonia. The ratio of the initial organic and aqueous phases was 1:1 in all experiments.

H NMR and ESI mass spectrometric analysis of the synthesis of [2 2+ ·2PF 6 − ] and [3 2+ ·2PF 6 − ] under various reaction conditions
[b] pH of the initial solvent system. Figure S8. 1 H NMR spectra recorded in DMSO-d 6 (400 MHz, 300K) representing the crude products obtained from the ring-opening reaction of 1 4+ ·4PF 6 − using different concentrations of NH 3 ·H 2 O at 333K for 2 h. Here NH 3 ·H 2 O% represents the volume fraction of NH 3 ·H 2 O (25-28% by weight) in final solution. The pH value shown is that of the initial solvent system. Figure S9. 1 H NMR spectra recorded in DMSO-d 6 (400 MHz, 300K) representing the crude products obtained from the ring-opening reaction of 1 4+ ·4PF 6 − using different concentrations of NH 3 ·H 2 O at 313K for 24h. Here NH 3 ·H 2 O% represents the volume fraction of NH 3 ·H 2 O (25-28% by weight) in the final solution. The pH value shown is that of the initial solvent system. Figure S10. 1 H NMR spectra recorded in DMSO-d 6 (400 MHz, 300K) representing the crude products obtained from the ring-opening reaction of 1 4+ ·4PF 6 − using different concentrations of NH 3 ·H 2 O at 313K for 48h. Here NH 3 ·H 2 O% represents the volume fraction of NH 3 ·H 2 O (25-28% by weight) in the final solution. The pH value shown is that of the initial solvent system.

Simulated ratio of 2 2+ ·2PF 6
− and 3 2+ ·2PF 6 − in the mixture Figure S11. Expanded view of signals for H(1) (the proton signals with most lowest chemical shift on 2 2+ ·2PF 6 − and 3 2+ ·2PF 6 − ) in the crude product (i.e., the mixture of 2 2+ and 3 2+ obtained via the general reaction conditions) in DMSO-d 6 at 300K (400 MHz). The black curve is the assynthesized peak data for H(1) in 2 2+ and 3 2+ . The red curve is the simulated peak data for H(1) in the mixture of 2 2+ and 3 2+ . The green curves is the simulated peak data for the H(1) signals in 2 2+ . The blue curves is the simulated peak data for the H(1) signals of 3 2+ . The mathematical area ratio, consistent with the molar ratio of 2 2+ ·2PF 6 − and 3 2+ ·2PF 6 − in the crude product, is approximately 3:1.

Section S4: Solution binding studies and characterization of the host-guest complexes formed between [2 2+ ·2PF 6
− ] and dianionic guests 2,6-naphthalenedicarboxylate (4), 2,6naphthalenedisulfonate (5) or 1,5-naphthalenedisulfonate (6) Solution binding studies and characterization of the host-guest complex formed between 2 2+ and dianionic guest 2,6-naphthalene dicarboxylate (4):       Figure S48. 1 H NMR spectroscopic titration of 2 2+ ·2PF 6 − (1.00 × 10 -3 M) with 2,6naphthalenedisulfonic acid in the presence of 2 molar equiv. of TMA + ·OHcollected in DMSOd 6 at 300K (600 MHz). Figure S49. 1 H NMR binding isotherms that correspond to the interaction between 2 2+ ·2PF 6 − and 5 collected in DMSO-d 6 at 300K. The chemical shift changes of (a) H(1), H(4) and H(7) on 2 2+ were used for the calculation of K a = (1.7 ± 0.2) ×10 2 M -1 for [2 2+ ·5] formation using the Hyperquad 2003 program. 9 The changes in chemical shift of other protons present in 2 2+ and 5 overlapped and thus could not be accurately fit with either a Guassian or Lorentzian function. Also in some cases the chemical shift change was too small (less than 8.0 Hz); these signals were therefore not used in the calculation of the association constant. 10 The red dashed lines show the non-linear curve fit of the experimental data to the appropriate equation.    7) and H(8) on 2 2+ were used for the calculation of (4.0 ± 0.5) ×10 2 M -1 using the Hyperquad 2003 program. 9 The changes in chemical shift of the other protons present in 2 2+ and 6 overlapped and thus could not be accurately fit with either a Guassian or Lorentzian function. Also in some cases the chemical shift change was too small (less than 8.0 Hz); these signals were therefore not used in the calculation of the association constant. 10 The red dashed lines show the non-linear curve fit of the experimental data to the appropriate equation.

Section S5:
Single crystal X-ray analysis of the complexes formed between 2 2+ and dianionic guest 5 or 6

S60
Section S6: Mass spectrometric analysis of the complexes formed between 2 2+ and dianionic guest 4, 5 or 6 Figure S66. Full (a) and expanded view (b) of the ESI high resolution mass spectrum obtained from a mixture of 2 2+ and guest 4. Figure S67. Full (a) and expanded view (b) of the ESI high resolution mass spectrum obtained from a mixture of 2 2+ and guest 5. Figure S68. Full (a) and expanded view (b) of the ESI high resolution mass spectrum obtained from a mixture of 2 2+ and guest 6.              11 It is found that the top width of 2 2+ in its clip conformation is 7.487 Å, which is smaller than for 1 4+ with a clip conformation of 9.370 Å. Table S5. Summary of the limiting binding modes for the complexes formed between 2 2+ ·2PF 6 − , 1 4+ ·4PF 6 − and dianions (4, 5, or 6).

Guest
Stoichiometric ratio and Association constants (K a ) with 1 4+ Stoichiometric ratio and Association constants (K a ) with 2 2+