Programming permanent and transient molecular protection via mechanical stoppering

A macrocycle (permanently or transiently) protects a viologen from heterogenous reduction, all thanks to bespoke mechanical stoppering.


General
All commercially available chemicals were purchased from Sigma-Aldrich and Oakwood Chemicals and used without further purification. Dry dichloromethane (CH2Cl2) and tetrahydrofuran (THF) were collected from an Inert PureSolv MD5 purification system, whereas nitromethane (MeNO2) and acetonitrile (MeCN) were dried with freshly activated 5 Å and 3 Å molecular sieves, respectively. Deuterated solvents (CD3CN, CDCl3, and DMSO-d6) were purchased from Cambridge Isotope Laboratories and Sigma-Aldrich. Flash column chromatography was carried out using SiliCycle (230-400 mesh) silica gel as the stationary phase. Nuclear magnetic resonance (NMR) experiments were recorded on Bruker AVIII HD 400 MHz and Bruker Avance 400 MHz spectrometers; 1 H and 13 C NMR chemical shifts () are given in parts per million (ppm) relative to TMS, using the residual solvent signal for calibration. J values are reported in Hz, and signal multiplicity is denoted as s (singlet), d (doublet), t (triplet), m (multiplet) and br (broad signal). UV-vis spectra were recorded on a Cary 5000 UV-vis-NIR spectrometer, employing 1 mm pathlength quartz cuvettes. High-resolution mass spectra (HRMS) were recorded on an ESI-TOF Waters Micromass LCT spectrometer. Cyclic voltammograms were acquired using an AFCBP1 potentiostat from Pine Instrument company. Macrocycle DN32C8, 1 and precursors C-21 2 and C-22, 3 were synthesized following reported methodologies and all spectroscopic characterization matched with the published data. Figure S1 shows the macrocycles trans-21C6, trans-22C6 and DN32C8 as found in reported [2]rotaxane structures. 1,2 Figure S1. Solid-state structures of the chosen macrocyclic scaffolds. The thread species hosted in their cavities have been omitted for clarity.
Changes in the NMR data (Figure S3a), for both the host and guest resonances, suggest the formation of the [2]pseudorotaxane [1•H2⸦(21C6)2] 4+ , which is stabilized by - stacking interactions. UV-vis spectroscopy reveals an absorption band centered at 475 nm (Figure S3b
The resulting isotherms were fit by a nonlinear least-squares method using the BindFit platform (supramolecular.org), and using a 1:1 global fitting model (Nelder-Mead method). Gasso was determined by applying Eq. 1. Table S1 contains the obtained thermodynamic data.
Eq. 1 ∆ = −  Figure S5 shows a representative example for the system assembled with C-22.
Mixing [1•H2][PF6]4 with DN32C8, in a 1:1 ratio (5 × 10 -3 M), resulted in a red solution, a color change ascribed to the formation of a charge-transfer complex with an absorbance band at 475 nm in the UV-vis spectrum (Figure S5c). In the 1 H NMR spectrum, only the bipyridinium resonances ( and ) showed significant changes (Figure S5a, ii), e.g.  moves to lower frequency and gets broader, indicating that the ring encircles only the central portion of [1•H2] 4+ .
In a separate experiment, [1•H2][PF6]4 (5 × 10 -3 M) was mixed with two equiv of C-22 and no color change was registered. In the corresponding NMR spectrum (Figure S5a,iv), only the + NH2 resonance registered a change while the pyridinium proton signals remained unaltered; this suggests that C-22 wraps exclusively around the ammonium stations.
DN32C8 sits on the central recognition site of [1•H2] 4+ while C-22 interacts with the ammonium end-groups, this was confirmed by the observed matching of all chemical shifts in the 1 H NMR (Figure S5a,iii) with respect to the two previously analyzed systems (vide supra). Moreover, the UV-vis band at 475 nm was detected with no detriment ( Figure S5b).   Figure   S11.  An alkaline solution (1 M, 1 equiv, NaOH(aq)) was added to a CD3CN solution of p-H [4]R and stirred for 10 min at room temperature. The 1 H NMR data collected at this point is shown in Figure   S12. From the experiment, it is evident that the + NH2 resonance is present after the addition of base.

Unstoppering attempts on p-H[4]R
The chemical shifts of the remaining signals suggest that the hetero [4]rotaxane stays (in both isomeric forms) unaltered in solution. Only slight changes were noticed that might be attributed to the presence of water and Na + ions coming from the added alkaline solution.

Polarity and temperature effect
When a solution of rotaxane p-H [4]R was prepared in DMSO-d6 and analyzed at room temperature over a period of one month, no changes were observed. A representative 1 H spectrum is shown in Figure S14. As observed in CD3CN, two isomers are distinguished in DMSO-d6; however, their relative concentrations correspond to 2:1 according to the 1 H NMR integrals. Heating the DMSO-d6 sample did not produce significant effects. Elevated temperatures only caused faster isomer interconversion as suggested by the 1 H NMR spectra collected from 25 ºC to 125 ºC ( Figure S15). S17 Figure S15. VT-NMR spectra (400 MHz, DMSO-d6) for compound p-H [4]R. Selected resonances are labeled.
To estimate the energy barriers between isomers A and B (interconverted by the rotation of one naphthalene motif (Np) along the RO-Np-OR axis), we employed the coalescence temperature method. 4 Resonances N1a and N1b were monitored by 1 H NMR in DMSO-d6 from 25 to 125 ºC ( Figure S16). At the coalescence temperature (115 ºC) the exchange rate of the isomers is given by where k is the exchange rate constant and  is the limiting chemical shift between the exchanging resonances. For our system a k of 110 s -1 was found. The values of the free activation energies for each isomer were obtained using Eyring's equation: where Tc is the coalesce temperature (388 K), R is the gas constant ( Figure S17a for reference), X = 2.09 at P = 0.33. The interconversion found barrier energies correspond to 19.5 kcal mol -1 and 19.0 kcal mol -1 , which implies a free energy difference (G) of 0.5 kcal mol -1 ( Figure S17b).

S20
Although other dynamic processes occurring within the hetero [4]rotaxanes cannot be excluded (such as short-rage shuttling and pirouetting), we rationalize that the anti/syn isomerization would be the only solvent-dependent process yielding different ratios of A and B in CD3CN (1:1) and .
Further computations demonstrated that the anti-DN32C8 ring renders a more stable interlocked species when compared to the corresponding syn isomer. The computed energy difference for both isomers, using a DN32C8/n-butyl viologen system as a model, corresponds to 0.94 kcal mol -1 , in close agreement with the experimental observations in DMSO. Computed data in MeCN, as explicit solvent, produced similar results (1.43 kcal mol -1 ), see Figure S18. These results were obtained with the Gaussian suite of program G016.revB01. 5 Geometries have been computed with the range separated dispersion corrected (wB97xD) 6 xc functional. Basis set for all atoms was 6-31g(d). 7 All structures were checked for being minima by computing the Hessian matrix at the same level of theory. All integrals were computed using the "Ultrafine" integration grid. Calculations were performed in solvent (MeCN and DMSO) using the SCRF (Self Consistent Reaction Field) approach for continuum solvent model simulation included in the Gaussian suite of programs (see data in Table S2-Table S4). 8 Reported SCXRD geometries 1,9 were used as input guesses for the optimization, increasing all the bond distances involving hydrogen atoms to 1.07 Å and adding the n-butyl side chains to the N and N' positions of the viologen moiety.

Scheme S5. Deprotection attempted on rotaxane p-H[4]R via polarity effect.
A sample of p-H [4]R in DMSO (1 mL, 2 × 10 -3 M) was loaded with zinc dust (50 mg) and stored over 6 months at room temperature in a N2 glove box. Figure S19 shows the UV-vis spectra for the initial and final stages, where no significant differences are observed. This confirms that the hetero [4]rotaxane remains assembled, with the substrate permanently protected.
The ferrocene (Fc) / ferrocenium (Fc + ) redox couple was used to reference all measurements. As shown in Figure S20, all cyclic voltammograms showed a reversible two-electron process for the reduction of the viologen substrate. As expected, the half-wave potentials (E 1 1/2 and E 2 1/2, summarized in Table S5) of the unprotected species, thread [1•H2] 4+ and [3]rotaxane [1•H2⸦(21C6)2] 4+ , are equivalent; i.e. E 1 1/2 is -0.86 V in both cases (as for the formation of the corresponding radical trications). In contrast, the presence of the protecting unit (DN32C8) in

p-H[4]R and t-H[4]
R produces a cathodic shift of ~140 mV for the same redox pair. This can be ascribed to the stabilization effect of DN32C8 to the viologen as result of a charge transfer process within the rotaxane structures, i.e. from the electron-rich ring to the electron-poor substrate. This stabilization effect was calculated to be ~13 kJ mol -1

Solutions of rotaxane t-H[4]
R at 1 × 10 -3 M concentration were analyzed towards three different stimuli: base, polarity, and heat.

Base trigger Scheme S7. Unstoppering process on t-H[4]R triggered by base addition.
A base solution (1 M, NaOH(aq), 1 equiv) was added to a CD3CN solution of t-H [4]R and stirred for 10 min at room temperature, then a 1 H NMR spectrum was collected ( Figure S22).   Figure S66, was no longer identified in the spectrum ( Figure S23a). In addition, just after mixing with base, the solution turned from red to pale yellow and the charge transfer band (490 nm) in the UV-vis spectrum was no longer detected (Figure S23b), suggesting that t-H [4]R was successfully disassembled. Polarity and temperature stimuli Scheme S8. [4]Rotaxane disassembly triggered by heat and polarity.
In contrast to other non-competitive solvents, such as CH2Cl2 or MeCN, DMSO produced unstoppering at room temperature, releasing the protecting macrocycle DN32C8. Heating the system causes fast unstoppering that renders all free components in solution, as clearly detected by 1 H NMR spectroscopy ( Figure S24).   Figure S66, was no longer detected ( Figure S25).  Figure S26 shows the UV-vis spectrum of the system after deprotection, this experiment matches with the one reported for species [1‧H2] (3+)• , shown in Figure   S2a.

Synthesis
Scheme S11. Synthesis of the targeted hetero [4]rotaxanes and their precursors.