Stereoselective Synthesis of Mechanically Planar Chiral Rotaxanes

Abstract Chiral interlocked molecules in which the mechanical bond provides the sole stereogenic unit are typically produced with no control over the mechanical stereochemistry. Here we report a stereoselective approach to mechanically planar chiral rotaxanes in up to 98:2 d.r. using a readily available α‐amino acid‐derived azide. Symmetrization of the covalent stereocenter yields a rotaxane in which the mechanical bond provides the only stereogenic element.


Entry 1 -axle and rotaxanes derived from alkyne 2a and azide 3a
These have been reported previously. See ref 3. Entry 2 -axle S12 and rotaxanes S13 derived from alkyne 2b and azide 3b Axle (D)-S12 A dry sealed vessel was charged with (D)-2b (22.4 mg, 0.075 mmol), 3b (17.4 mg, 0.075 mmol), [Cu(MeCN)4]PF6 (26.8 mg, 0.073 mmol) and anhydrous CH2Cl2 (1.25 mL). The reaction mixture was stirred at rt for 16 h, protected by an argon atmosphere. Saturated EDTA-NH3 (10 mL) was added, and the aqueous layer was extracted with CHCl3 (3 × 20 mL), dried over MgSO4, filtered, and had the solvent removed in vacuo. The residue was purified by chromatography (75% CH2Cl2-MeCN), to yield axle (D)-S12 as a yellow oil (37.7 mg, 95%); 1     *The signals arising from the two diastereomers designated HX and Hx`. Careful examination of 2D NMR data allowed the signals from each diastereomer to grouped accurately within the axle and macrocycle but not between them. Proton counts are provided for each signal and represent the expected integration of that environment. Where the signals of both diastereoisomers are coincident, no HX/Hx' label is provided and the proton count indicated refers the expected integration of that signal in each of the stereoisomers that contributes to the multiplet.       +0.2% formic acid →1 : 0 MeCN-H2O +0.2% formic acid), UV 254 nm), of (D,Rmp/Smp)-S13 following purification by chromatography.

Entry 8 -axle (S)-S23 and rotaxane (S,Rmp/Smp)-S24 derived from alkyne 2e and azide (S)-3e
Axle (    *As stereochemistry could not be unambiguously assigned, the signals are simply designated (major) or (minor). Proton counts are provided for each signal and represent the expected integration of that environment. Where the major and minor diastereoisomer signals are coincident, no (major)/(minor) label is provided and the proton count indicated refers the expected integration of that signal in each of the stereoisomers that contributes to the multiplet.

Entry 9 -axle (S)-S25 and rotaxane (S,Rmp/Smp)-S26 derived from alkyne 2f and azide (S)-3e
Axle (S)-S25 A 10 mL round bottomed flask was charged with 2f (75.0 mg, 0.264 mmol), (S)-3e (28.1 mg, 0.120 mmol), CuSO4.5H2O (29.29 mg, 0.120 mmol), sodium L-ascorbate (31.1 mg, 0.157 mmol), and DMF (3 mL). The reaction mixture was stirred at rt for 16 h. Saturated EDTA-NH3 solution (20 mL) was added, and the aqueous layer was extracted with EtOAc (3 × 20 mL). The combined organic extracts were washed with 5% w/v LiCl (5 × 20 mL), brine (20 mL), were dried over MgSO4, filtered, and had the solvent removed in vacuo. The residue was purified by chromatography (CH2Cl2 with 0→10% EtOH), to yield axle (S)-S25 as a yellow oil (16.7 mg, 28%); 1 13  *As stereochemistry could not be unambiguously assigned, the signals are simply designated (major) or (minor). Proton counts are provided for each signal and represent the expected integration of that environment. Where the major and minor diastereoisomer signals are coincident, no (major)/(minor) label is provided and the proton count indicated refers the expected integration of that signal in each of the stereoisomers that contributes to the multiplet.         *As stereochemistry could not be unambiguously assigned, the signals are simply designated (major) or (minor). Proton counts are provided for each signal and represent the expected integration of that environment. Where the major and minor diastereoisomer signals are coincident, no (major)/(minor) label is provided and the proton count indicated refers the expected integration of that signal in each of the stereoisomers that contributes to the multiplet.           * As stereochemistry could not be unambiguously assigned, the signals are simply designated (major) or (minor). Proton counts are provided for each signal and represent the expected integration of that environment. Where the major and minor diastereoisomer signals are coincident, no (major)/(minor) label is provided and the proton count indicated refers the expected integration of that signal in each of the stereoisomers that contributes to the multiplet. S103 Figure S142. 1

Single crystal X-ray crystallographic data of (S,Smp)-4
Single crystals of (S,Smp)-4 were grown by slow evaporation of 11 Petrol-CHCl3. Data was collected at 100 K using a Rigaku 007 HF diffractometer equipped with a Saturn724+ enhanced sensitivity detector. Cell determination, data collection, data reduction, cell refinement and absorption correction were performed with CrysalisPro. Using Olex2 the structure was solved with the SHELXT program using charge flipping, [18] and refined with the SHELXL refinement package. [18] H atoms were placed in calculated positions and refined using a riding model. Figure S181. Single crystal X-ray structure of (S,Smp)-4 . Ellipsoids shown at 50% probability S127 Figure S184. Chiral SCFC chromatogram of enantiomerically pure (Rmp)-5, following purification by chromatography.

Racemisation of axle (S)-S20 under CuAAc conditions
In the synthesis of axle (S)-S20, partial racemization was observed. A small screen was performed to assess the effect of reaction conditions on the racemisation process. Strikingly, although no racemisation was observed in the AT-CuAAC synthesis of rotaxanes 4 or 5, when a macrocycle S9 [3] that is too large to be retained by the stoppers is employed (entry c), the racemisation process is exacerbated. Thus, it appears that the racemisation of the stereocentre α-to the ester is inhibited in rotaxane 4, presumably due to the steric hindrance provided by the mechanical bond.   Table S2, entry a, following purification by chromatography. Figure S186. Chiral SCFC chromatogram of Table S2, entry b, following purification by chromatography. S130 Figure S187. Chiral SCFC chromatogram of Table S2, entry c, following purification by chromatography.        Figure S195. ESI-MS isotopic pattern of (R/S,Smp)-S33; observed (top) and calculated (bottom). S136 Figure S196. LCMS trace (C18 column, gradient 5 minutes (1 : 4 MeCN+0.2% formic acid-H2O +0.2% formic acid →1 : 0 MeCN-H2O +0.2% formic acid), UV 254 nm), of (D,Rmp/Smp)-S13 following purification by chromatography.

Preliminary Molecular Modelling
Based on previous work, [19] the AT-CuAAC reaction of macrocycle 1, acetylene 2a and azide 3e is thought to proceed via Cu I -acetylides I and II which are irreversibly converted to Cu I -triazolides III and IV and ultimately, after protolytic work-up, to rotaxanes (S,Smp)-4 (major) and (S,Rmp)-4 respectively (Scheme S1). Based on this proposed mechanism, two obvious sources of diastereoselectivity can be identified; i) a significant energy difference between I and II which results in a biased pre-equilibrium (Cu I acetylide formation can be expected to be reversible in the presence of N i PrEt2) prior to irreversible covalent bond formation; ii) a significant difference in reaction rate for the conversion of IàIII and IIàIV.
Scheme S1. Schematic mechanism of the AT-CuAAC reaction via key acetylide and triazolide intermediates

S137
To probe the origin of stereoselectivity in this AT-CuAAC reaction we carried out preliminary calculations to determine the relative energies of I/II and of the reaction DG IàIII and IIàIV. It should be noted that, given the controversial nature of the cycloaddition mechanism and the size of the molecules concerned, accurately identifying the transition states energies for the cycloaddition step lies beyond the scope of this preliminary study. However, linear-free energy relationship considerations suggest that reactions proceeding via the same pathway but with a larger DG of reaction can be expected to have a lower reaction barrier, although the difference in reaction energies is likely to be significantly larger than the difference in activation energies as the reaction is predicted to have an early barrier (Hammond postulate). Thus, although these calculations cannot be used to derive the difference in reaction rates for IàIII and IIàIV, the comparison of the reaction energies IàIII and IIàIV gives an indication of whether a difference in reaction rates is to be expected. 2.0 kJmol -1 6.2 kJmol -1 a Modelling was carried out using Spartan '10 (Wavefunction). Molecular models (Figures S197-S200) of the intermediates (Cu I -acetylide and -triazolide) were prepared and subjected a conformer distribution (MMFF) search. The lowest energy conformation was selected and the energy minimised (PM6, gas phase). The energies obtained (Table S3) were compared.
Based on the results in Table S3 the origin of the stereoselectivity is expected to be a combination of a biased pre-equilibrium and the difference in reaction rates of the cycloaddition step, which is perhaps unsurprising. Furthermore, comparison of the models of I and II and III and IV reveals that, as might be expected, the stereoselectivity seems to be driven by avoiding steric clash between the more rigid aryl-pyridine motif and the benzyl group of the azide component. To examine the reaction further, we modelled the same intermediates (V-VIII) for the reaction macrocycle 1, azide 3e and phenyl acetylene (Table S4); our experimental results (Table 1, entry 11) suggest that the mt Bu groups of alkyne 2a are important for high stereoselectivity.  S204) of the intermediates (Cu I -acetylide and -triazolide) were prepared and subjected a conformer distribution (MMFF) search. The lowest energy conformation was selected and the energy minimised (PM6, gas phase).
Pleasingly, the results in Table S5 are in broad agreement with experiment in that the preequilbrium of acetylides derived from phenylacetylene is predicted to be essentially unbiased removing this source of stereoselectivity, whereas a significant difference in reaction free energy is predicted. Thus, the modelling predicts that the stereoselectivity in this reaction is determined by the difference in reaction rates in the reactions VàVII and VIàVIII without the benefit of a biased equilibrium between V and VI.  (Table S4) The results of this preliminary molecular modelling suggest that it may be possible to qualitatively predict the stereochemical efficiency of a directing group by identifying candidates that produce a biased pre-equilbrium of Cu I -acetylides. However, further comparisons of experiment and theory are required before this can be confirmed.