Versatile control of the submolecular motion of di(acylamino)pyridine-based [2]rotaxanes† †Electronic supplementary information (ESI) available: Experimental procedures, spectroscopic and mass spectrometry data for all new compounds, electrochemical studies, and full crystallographic details of 1c a

Di(acylamino)pyridine motifs enable the building of switchable interlocked systems in which their dynamics can be exchanged between different states.


General procedure for the preparation of benzylic amide macrocycle containing [2]rotaxanes
The thread (1 mmol) and Et 3 N (24 equiv.) in anhydrous CHCl 3 (300 mL) were stirred vigorously whilst solutions of p-xylylene diamine (12 equiv.) in anhydrous CHCl 3 (20 mL) and the corresponding acid dichloride (12 equiv.) in anhydrous CHCl 3 (20 mL) were simultaneously added over a period of 5 h using motor-driven syringe pumps. After a further 4 h the resulting suspension was filtered through a Celite ® pad, washed with water (2 x 50 mL), a saturated solution of NaHCO 3 (2 x 50 mL) and brine (2 x 50 mL).
The organic phase was dried over MgSO 4 and the solvent removed under reduced pressure. The resulting solid was subjected to column chromatography (silica gel) to yield unconsumed thread, [2]rotaxane and [2]catenane. The formation of the [3]rotaxane was not observed in any case.

Rotaxane 8
Rotaxane 8 was obtained following the described method from thread 6 (0.60 g, 0.77 mmol).  To a solution of the corresponding thread or rotaxane (1 equiv.) in CHCl 3 (2 mL) was added m-CPBA (2 equiv.). The reaction mixture was stirred overnight at room temperature after which time Amberlyst A-21 ® basic resin (1 gr/mmol RTX or Thread ) was added. The mixture was stirred for 1 hour and then filtered.
The residue was washed with 1 mL of CHCl 3 . The organic phase was concentrated under reduced pressure, yielding the titled compounds.

Rotaxane 12
Rotaxane 12 (50.6 mg, 0.050 mmol, 99%) was obtained as a white solid following the described method  6. General procedure for the reduction of pyridine N-oxide derivatives 1 To a solution of the corresponding pyridine N-oxide (1 equiv.) in MeCN, PPh 3 (polymer-supported, 3 mmol PPh 3 /g polymer) (15 equiv.) and MoO 2 Cl 2 (DMF) 2 (10 mol%) were added. The reaction mixture was heated for 2 hours in a closed-vial after which time the reaction was complete. The solution was diluted with a solution of CHCl 3 :MeOH (95:5) (2 mL) and filtered through a pad of silica. The residue was concentrated under reduced pressure to give the corresponding di(acylamino)pyridine derivative.
The reaction mixture was stirred 3 h at room temperature after this time the solution was concentrated under reduced pressure and dried under vacuum, yielding the corresponding salt quantitatively.

General procedure for the deprotonation of picrate salts
To a solution of the corresponding rotaxane/thread picrate salt (1 equiv) in CHCl 3 (1 mL), Amberlyst A21 ® resin (1 gr/mmol salt ) was added. The reaction mixture was stirred 1 h at room temperature after which time the solution was filtered and the filtrate was concentrated under reduced pressure and dried under vacuum, yielding quantitatively the corresponding (diacylaminopyridine) derivative.  In the case of the N-oxide rotaxane 2b the range of temperature oscillated between 393 K (in C 2 D 2 Cl 4 ) and 233 K. In this case we can observe two temperatures where different signals coalesced. At 298 K the signal associated with the methylene protons of the macrocycle (H E ) coalesced, associated with an energy barrier for the macrocycle pirouetting of 13.6 kcal·mol −1 . Again a splitting of the signals for the stoppers (H a and H b ) was observed at 318 K. Similar experiments were carried out for picrate salt 2c. In this case signals referred to the stoppers H a and H b coalesced at the temperature of 238 K, observing a splitting of those signals below that temperature. Interestingly two signals referred to the amide proton NH c also appeared. At 231 K the signal associated with the methylene protons of the macrocycle (H E ) coalesced, associated with an energy barrier for the macrocycle pirouetting of 10.7 kcal·mol −1 . In order to estimate the occupancy of the tetraamide macrocycle 3 over the diacylaminopyridine binding site in the synthesized rotaxanes, we focused our attention on the upshifting of the hydrogen atom at 4position of the pyridine ring (H e ) following the rotaxane formation. These values were compared with the shift of this proton in rotaxanes 2a-c, where the occupation of the binding site is complete (Table S1). The examination of the chemical shift variations in the 1 H NMR of the rotaxanes revealed that the occupation of the DAP station in 7 (68%) is notably higher than in 8 (28%) ( Table S1, c .f. entries 4 and 6). In rotaxane 8, the succinic ester station has stronger affinity for the macrocycle than the DAP moiety. In similar manner, the level of occupancy over the pyridine N-oxide function of rotaxane 12 was also calculated, being 100%.

Electrochemical Studies.
Electrochemical measurements were performed on a CH Instruments 760D Electrochemical Workstation using CHI Version 10.03 software. Electrochemical experiments for obtaining the cyclic voltammograms (CVs) in the main text were conducted at 298 K using a CH Instruments glassy carbon button working electrode (area = 0.071 cm 2 ), BASi Ag/AgNO 3 pseudo reference electrode, and Pt mesh counter electrode.
All electrode potentials were referenced to the ferrocene/ferrocenium couple by doping in samples of ferrocene to the electrolyte. All electrochemical experiments were conducted in electrolyte solutions prepared using HPLC grade DCM, and were thoroughly degassed with argon. The supporting electrolyte was tetrabutylammonium hexafluorophosphate (TBAPF 6 ) at a concentration of 1 mol dm -3 . Solutions were agitated between acquisition of individual CVs and CVs were corrected for resistance, using the iR compensation function of the potentiostat. All data points were repeated at least twice and gave very similar results. Figure 5 in the main paper shows representative cyclic voltammograms and the curve shown in Figure S7 is based on the average E ½ values obtained from these multiple repeats.

ESI-26
A typical methodology is as follows: To a solution of rotaxane 7 (5 mM) in DCM (with 1 M TBAPF 6 as supporting electrolyte), aliquots of a 2.5 mM solution of 1,8-naphthalimide in dry DCM were added. Care was taken to ensure that the overall volume of solvent remained the same in all measurements by allowing added solvents to evaporate in a stream of Ar such that a fixed volume was achieved. All samples were kept under Ar during measurements (CVs were much less reversible under air). CVs were recorded at room temperature and a scan rate of 100 mV/s. Shifts relative to the pseudo reference were converted to shifts vs. ferrocene/ferrocenium by adding ferrocene to the solutions after all other data had been collected.
Rotaxane 7 did not display any redox waves within the window probed (0 to -2 V vs.  Figure S7 (where 0 equivalents of 7 corresponds to the reduction wave for 1,8-naphthalimide on its own). This shows that maximal shift is achieved in the presence of around 6 equivalents of rotaxane, and that the presence of further equivalents of rotaxane makes little difference to the shift.  Figure S8). The spectrum also shows crosspeaks relating protons H F -H c and H F -H g . Figure S8. Partial 1 H, 1 H-NOESY spectrum crosspeaks for rotaxane 7

ESI-28
Additionally, the NOESY spectrum reveals the proximity between the aromatic protons H F placed at the macrocycle and several aliphatic protons close to the diacylaminopyridine station (H b , H h , H i , H j and H k ) ( Figures S9 and S10). Weak crosspeaks relating protons H F -H s and H q support the 32% of occupancy over the amide function.  The 1 H, 1 H-NOESY spectrum does not show crosspeaks relating protons from the macrocycle and those of the diacylaminopyridine station ( Figure S11). However, the spectrum shows intense crosspeaks between H F -H q , H F -H m,n , H F -H i and H C -H h compatible with translocation of the macrocycle. Also weak crosspeaks relating protons H F -H p and H F -H j have been found ( Figure S12). On the contrary, the signal for proton H s appears upfield (∆δ  = -0.16 ppm) as well as all resonances for the aliphatic chain. All these data agree well with a decrease in the occupancy percentage of the macrocycle at the diacylaminopyridine site, due to the association of this moiety to N-hexylthymine. In fact, the ratio of occupancy determined as usual takes a value of 32:68.

ESI-31
As expected, the 1 H, 1 H-NOESY spectrum does not show crosspeaks relating protons from the macrocycle and those of the diacylaminopyridine station (all resonances, except those for protons l-o placed at the aliphatic chain, could be assigned on the basis of the COSY and 1 H, 1 H-NOESY spectra). On the contrary, the spectrum shows intense crosspeaks relating protons of the macrocycle and those close to the amide site, i.e. H C -H r , H C -H s , H F -H q and H F -H s ( Figure S13). The NOESY spectrum also displays a crosspeak between H C and H h . Moreover, strong signals are observed between protons H F and those of the aliphatic chain H i-p ( Figure S14). Finally, the NOESY spectrum displays also weak crosspeaks between H D -H q and H D -H s ( Figure S15). The structure was solved and refined using the Bruker SHELXTL Software Package, using the space group P -1, with Z = 4 for the formula unit, C 42 H 34 N 6 O 10 . The final anisotropic full-matrix least-squares refinement on F 2 with 1063 variables converged at R1 = 11.87%, for the observed data and wR2 = 27.48% for all data. The goodness-of-fit was 1.173. The largest peak in the final difference electron density synthesis was 0.827 e -/Å 3 and the largest hole was -0.737 e -/Å 3 with an RMS deviation of 0.105 e -

ESI-33
/Å 3 . On the basis of the final model, the calculated density was 1.383 g/cm 3 and F(000), 1632 e -.The poor overall precision resulted from the presence of a great deal of disordered picrate anions and methanol cosolvate molecules, coupled with poor overall crystal quality as a result of decomposition.

ESI-34
Figure S16. Molecular structure of salt 1c (CCDC 1051908) with thermal ellipsoids drawn at 50% probability. For clarity, the picrate anion and a methanol co-solvate molecule have been omitted. The structure was solved and refined using the Bruker SHELXTL Software Package, using the space group P -1, with Z = 2 for the formula unit, C 67 H 57.75 N 7 O 7 . The final anisotropic full-matrix least-squares refinement on F 2 with 773 variables converged at R1 = 5.68%, for the observed data and wR2 = 13.16%

ESI-35
for all data. The goodness-of-fit was 1.104. The largest peak in the final difference electron density synthesis was 0.715 e -/Å 3 and the largest hole was -0.504 e -/Å 3 with an RMS deviation of 0.06 e -/Å 3 . On the basis of the final model, the calculated density was 1.345 g/cm 3 and F(000), 1130 e -.  Figure S17. Molecular structure of rotaxane 2b (CCDC 1051909) with thermal ellipsoids drawn at 50% probability.