We report a one-step enantioselective synthesis of mechanically planar chiral [2]rotaxanes. Previous studies of such molecules have generally involved the separation of enantiomers from racemic mixtures or the preparation and separation of diastereomeric intermediates followed by post-assembly modification to remove other sources of chirality. Here, we demonstrate a simple asymmetric metal-free active template rotaxane synthesis using a primary amine, an activated ester with a chiral leaving group, and an achiral crown ether lacking rotational symmetry. Mechanically planar chiral rotaxanes are obtained directly in up to 50% enantiomeric excess. The rotaxanes were characterized by NMR spectroscopy, high-resolution mass spectrometry, chiral HPLC, single crystal X-ray diffraction, and circular dichroism. Either rotaxane enantiomer could be prepared selectively by incorporating pseudoenantiomeric cinchona alkaloids into the chiral leaving group.
We report a one-step enantioselective synthesis of mechanically planar chiral [2]rotaxanes. Previous studies of such molecules have generally involved the separation of enantiomers from racemic mixtures or the preparation and separation of diastereomeric intermediates followed by post-assembly modification to remove other sources of chirality. Here, we demonstrate a simple asymmetric metal-free active template rotaxane synthesis using a primary amine, an activated ester with a chiral leaving group, and an achiral crown ether lacking rotational symmetry. Mechanically planar chiral rotaxanes are obtained directly in up to 50% enantiomeric excess. The rotaxanes were characterized by NMR spectroscopy, high-resolution mass spectrometry, chiral HPLC, single crystal X-ray diffraction, and circular dichroism. Either rotaxane enantiomer could be prepared selectively by incorporating pseudoenantiomeric cinchona alkaloids into the chiral leaving group.
Mechanical planar chirality
arises in rotaxanes with achiral components
when an unsymmetrical axle is threaded through a macrocycle lacking
rotational symmetry (Figure ).[1−4] Although lacking classical elements of chirality, studies on mechanically
planar chiral rotaxanes suggest their asymmetry can be well expressed
for applications.[5−7] However, despite mechanically planar chiral rotaxanes
being known for nearly 50 years, their enantioselective synthesis
remains challenging.[1d,8] Most studies on these systems
rely on the separation of enantiomers from racemic mixtures by chiral
stationary phase HPLC, limiting the scale of enantioenriched material
that can readily be obtained.[9]
Figure 1
Enantioselective
synthesis of mechanically planar chiral rotaxanes
through metal-free active template N-acylation using
a macrocycle lacking rotational symmetry and an electrophile with
a point-chiral leaving group.
Enantioselective
synthesis of mechanically planar chiral rotaxanes
through metal-free active template N-acylation using
a macrocycle lacking rotational symmetry and an electrophile with
a point-chiral leaving group.Goldup et al. have addressed this synthetic problem through a chiral
auxiliary approach that forms intermediate diastereomeric rotaxanes
having both point chirality and mechanically planar chirality.[10,11] Separation of these diastereomeric intermediates by flash chromatography,
followed by removal of the point chirality by either substitution[10] or symmetrization,[11] afforded enantioenriched mechanically planar chiral rotaxanes. The
only single-step synthesis of enantioenriched mechanically planar
chiral rotaxanes to date used a chiral catalyst to resolve the interconverting
enantiomers of a crown ether-ammonium pseudorotaxane by capping.[12] Despite attempts to optimize this method, it
produced rotaxanes in just 4% enantiomeric excess (e.e.). Here we
report a simple, single-step, enantioselective synthesis of mechanically
planar chiral rotaxanes that produces either enantiomer in up to 50%
e.e.Metal-free active template reactions have recently been
developed
in which rotaxanes[13] are spontaneously
assembled under kinetic control in a single step by combining a primary
amine, electrophile, and crown ether[14] in
apolar solvents. Crown ethers stabilize the transition states of various
nucleophilic substitution reactions through the cavity by C–H
hydrogen bonding, thereby favoring the formation of rotaxanes over
the unthreaded axle. Different reactions, amines, and leaving groups
result in different degrees of accelerated reaction through the ring,
affording different rotaxane:thread selectivities. We chose crown
ether-stabilized N-acylation for the present study
(Scheme ), as this
active template reaction often results in a particularly high ratio
of rotaxane:thread.[14] This suggested the
reaction might be tolerant of the additional functionality necessary
in the macrocycle (to break rotational symmetry) and axle building
blocks (to provide a chiral leaving group).
Scheme 1
: (a) Achiral Rotaxane
Synthesis by Active Template N-Acylation Using Rotationally
Symmetrical Crown Ethers; (b) Racemic
and Unoptimized Enantioselective Synthesis of a Mechanically Planar
Chiral Rotaxane
An example of an active
template N-acylation is
the reaction of 24-crown-8 1, primary amine 2, and electrophile 3 in toluene at room temperature,
producing amide [2]rotaxane 4 in 84% yield (Scheme a).[14b] The rate-determining step of crown ether catalyzed N-acylation reactions is the collapse of the tetrahedral
intermediate formed on addition of the amine to the activated ester.[15] The nitro-phenol ester used in the reaction
of 1, 2, and 3 thus provides
an opportunity for a chiral directing group to be incorporated into
the leaving group that could interact with a rotationally unsymmetrical
macrocycle in the transition state (Figure ).[16]
Results and Discussion
Development
of an Enantioselective Rotaxane Synthesis
To establish
that functionalized crown ethers could take part in
the active template reaction, amine 2 and activated ester 3 were treated with commercially available dibenzo-24-crown-8
(5) in toluene, yielding the corresponding [2]rotaxane, 6, in 73% yield (Scheme a). However, although the rotaxane axle is unsymmetrical,
dibenzo-24-crown-8 (5) is D2 symmetric and so rotaxane 6 is achiral.[1] Macrocycle 7, containing two different
aromatic rings, lacks rotational symmetry (it has C1 symmetry, alternatively referred to
as C). Reaction of 7 with 2 and 3 furnished racemic
mechanically planar chiral rotaxane 8 in 78% yield (Scheme b). The enantiomers
of 8 could be separated by chiral stationary phase HPLC
(see Supporting Information).Next,
we investigated the structure and location for an effective chiral
leaving group in the electrophile. Preliminary screening studies identified
nitrophenol ester 9, in which the chiral information
stems from an O-alkylated cinchonidine unit adjacent
to the nitro-group (Scheme b). This electrophile was reactive under the rotaxane-forming
conditions despite the introduction of the deactivating electron-donating
ether linkage. Combining 2, 7, and 9 in a 1:1:1 stoichiometry in toluene at room temperature
afforded rotaxane 8 in 43% yield (Scheme b). Under similar conditions, electrophiles
based on alkyl (thio)esters or with the cinchonidine unit positioned
at the ortho position of the nitrophenol ring were either unreactive
or generated less rotaxane (see Supporting Information). HPLC analysis of rotaxane 8 (isolated by flash chromatography)
obtained from electrophile 9 revealed that the (+)-enantiomer
(determined by polarimetry) had been formed in 12% e.e., confirming
that a point-chiral leaving group was able to induce enantioselectivity
of a mechanically planar rotaxane product.Increasing the
electronic difference between the two aromatic substituents
within the macrocycle improved the enantioselectivity of the active
template reaction. Macrocycle 10, with a nitro group
on the catechol unit (see Supporting Information for its synthesis), afforded rotaxane (+)-11 in 23%
e.e. at room temperature, which increased to 40% e.e. (55% yield)
when the rotaxane-forming reaction was performed at −40 °C
(Figure a). Lowering
the reaction temperature beyond −40 °C did not result
in further improvements in enantioselectivity.[17]
Figure 2
(a) Enantioselective synthesis of mechanically planar chiral rotaxane 11. Reaction conditions: 2 equiv of amine 2,
1 equiv each of electrophile and crown ether 10, toluene,
[0.14 M], –40 °C, 24 h. (b) Partial 1H NMR
spectra (600 MHz, CDCl3, 295 K) of macrocycle 10 (top), rotaxane 11 (middle) and the corresponding unthreaded
axle (bottom).
(a) Enantioselective synthesis of mechanically planar chiral rotaxane 11. Reaction conditions: 2 equiv of amine 2,
1 equiv each of electrophile and crown ether 10, toluene,
[0.14 M], –40 °C, 24 h. (b) Partial 1H NMR
spectra (600 MHz, CDCl3, 295 K) of macrocycle 10 (top), rotaxane 11 (middle) and the corresponding unthreaded
axle (bottom).The opposite enantiomer of the
rotaxane, (−)-11, could be selectively accessed
using electrophile 12, derived from (+)-cinchonine, a
pseudoenantiomer of cinchonidine
(see Supporting Information for synthesis).[18] Combining 2, 10, and 12 at −40 °C gave rotaxane (−)-11 in 50% e.e. and 51% yield (Figure a). The difference in enantioenrichment is a consequence
of electrophiles 9 and 12 being diastereomers
rather than true enantiomers.
Characterization of Rotaxanes
Comparison of the 1H NMR spectra of macrocycle 10, rotaxane 11, and the unthreaded axle (see Supporting Information for synthesis) in CDCl3 at 298 K (Figure b) confirmed the
interlocked structure of 11. The geminal protons of the
crown ether display twice the number of environments in rotaxane 11 as in unthreaded 10 due to desymmetrization
of the two macrocycle faces upon rotaxane formation, while H3 and H5 of the axle (hydrogen labeling shown in Figure a), which are situated
either side of the amide group, display significant diastereotopic
splitting (Δδ = 0.39 and 0.22 ppm respectively) within
the chiral environment of rotaxane 11 which, as
would be expected, is absent for the corresponding achiral noninterlocked
axle. Upfield shifts of H6 and H7 (Δδ
= −0.32 and −0.34 ppm) in the threaded axle and HA, HB, and HC (Δδ = −0.49,
−0.21, and −0.37 ppm) of the nitrocatechol unit of the
threaded macrocycle result from π–π interactions
involving these moieties. These intercomponent interactions may play
a role in rigidifying the transition state of the collapsing tetrahedral
intermediate. The large downfield shift of the amide N–H proton
H4 (Δδ = +1.74 ppm) in 11 is indicative
of intercomponent hydrogen bonding between the amide and the glycol
chain of the macrocycle. An upshield shift of HD (Δδ
= −1.29 ppm) results from hydrogen bonding with the amideoxygen
atom.[14b]Enantioenriched samples
of rotaxane 11 (40% e.e. for the (+) enantiomer and 50%
e.e. for the (−) enantiomer) were compared by circular dichroism
(Figure a). The CD
spectra of the mechanically planar chiral rotaxane enantiomers are
symmetrical in terms of curve shape and have exciton couplings of
opposite sign with maxima at 243 nm. The difference in intensity (normalized
for absorption) of the spectra in Figure a corresponds to the difference in enantioenrichment
of the samples.
Figure 3
(a) Circular dichroism spectra (1.0 × 10–4 M, CH2Cl2, 298 K) of (+)-11 (red)
and (−)-11 (blue), baseline corrected. (b) Chemical
structure of racemic rotaxane 13. (c) X-ray crystal structure
of racemic rotaxane 13, side-on view showing intercomponent
hydrogen bonds (in green). Hydrogen bond lengths: N47H—O16,
2.20 Å; O49—HC6, 2.63 Å. Hydrogen bond angles: N47—H–O16,
158.4°; O49—H–C6, 161.8°. (d) X-ray crystal
structure of 13 viewed along the axle showing π-stacking
between the macrocycle 1,2-dihydroxynaphthalene and axle bis(trifluoromethyl)phenyl
rings. Centroid–centroid distance, 3.67 Å. Angle described
by C40 and centroids, 97.6°. Solvate molecules and other hydrogen
atoms omitted for clarity.
(a) Circular dichroism spectra (1.0 × 10–4 M, CH2Cl2, 298 K) of (+)-11 (red)
and (−)-11 (blue), baseline corrected. (b) Chemical
structure of racemic rotaxane 13. (c) X-ray crystal structure
of racemic rotaxane 13, side-on view showing intercomponent
hydrogen bonds (in green). Hydrogen bond lengths: N47H—O16,
2.20 Å; O49—HC6, 2.63 Å. Hydrogen bond angles: N47—H–O16,
158.4°; O49—H–C6, 161.8°. (d) X-ray crystal
structure of 13 viewed along the axle showing π-stacking
between the macrocycle 1,2-dihydroxynaphthalene and axle bis(trifluoromethyl)phenyl
rings. Centroid–centroid distance, 3.67 Å. Angle described
by C40 and centroids, 97.6°. Solvate molecules and other hydrogen
atoms omitted for clarity.Although we were unable to obtain high quality single crystals
of 11, single crystals of a racemic sample of 13 suitable for analysis by X-ray diffraction were grown by slow evaporation
of an isopropanol/hexane solution of 13 (Figure b). Rotaxane 13 contains the same macrocycle as 11 and an axle derived
from amine 2 and a different acyl stopper. The X-ray
crystal structure of 13 (Figure c), containing both rotaxane enantiomers
in the unit cell, shows similar intercomponent interactions to those
observed by 1H NMR for 11 in solution (Figure b). Hydrogen bonds
are present between an oxygen of the macrocycle glycol chain and the
amidehydrogen atom of the axle and between the amideoxygen and a
macrocycle C–H hydrogen atom (analogous to HD in 11).[14] The di(alkoxyl)naphthalene
ring of the macrocycle and bis(trifluoromethyl)benzene
unit of the axle π-stack (Figure d, closest centroid-centroid distance = 3.67 Å),[19] with the nitro-catechol moiety positioned so
as to cover one face of the amide group. A similar arrangement in
the transition state of the active template reaction would orient
the macrocycle with respect to the axle building blocks such that
one mechanically planar chiral enantiomer would be favored over the
other.
Origin of Enantioselectivity
A preliminary indication
of the origin of chiral transduction in these systems comes from the
relative energies of the tetrahedral intermediates preceding (+)-
and (−)-11, calculated at the PM6 level[20] using the Gaussian 09 software package[21] (Supporting Information and Figure ). The
collapse of similar tetrahedral intermediates has previously been
shown[15a] to be the rate-determining step
for the glyme catalysis of ester aminolysis. Following the Hammond
postulate, the differences between the diastereomeric tetrahedral
intermediates to (+)- and (−)-11 from 9 and 12 may resemble those between the transition states.
The lowest energy intermediate calculated for both pseudoenantiomeric
leaving groups featured an (S) stereocenter adjacent
to the ammonium unit, but with the macrocycle orientation inverted
for the two pseudoenantiomers (Figure ), meaning changing between the leaving groups of 9 and 12 favors the formation of a different
enantiomer of 11, as observed experimentally. The somewhat
surprising indication that the two chiral leaving groups both favor
an (S)-tetrahedral intermediate may reflect why the
pseudoenantiomers do not generate equal and opposite e.e.’s
in the active template reaction. The noncovalent interactions in the
intermediate (e.g., the stacking of the electron-rich naphthalene
unit with the electron-poor aryl group of the nucleophile, and the
hydrogen bonding of the glycoloxygens to the H–N atoms) are
reminiscent of those present in the X-ray crystal structure of rotaxane 13.
Figure 4
Tentative rationale for the transfer of chirality from Euclidean
point-chirality (of the leaving group) to mechanical planar chirality
(of the rotaxane). The lowest energy tetrahedral intermediates were
modeled (see Supporting Information) using
(a) electrophile 9 or (b) electrophile 12. The di(alkoxyl)naphthalene ring of the macrocycle and bis(trifluoromethyl)benzene
unit originating from the nucleophile π-stack, causing the nitro-catechol
ring to be positioned so as to cover one face of the tetrahedral center
of the intermediate. This thermodynamically favored arrangement of
components ensures the different handedness of the pseudoenantiomeric
leaving groups is well-expressed in the diastereomeric transition
states, resulting in enantioselectivity in the mechanically planar
chiral rotaxane product. Hydrogen bonds are indicated by black dotted
lines.
Tentative rationale for the transfer of chirality from Euclidean
point-chirality (of the leaving group) to mechanical planar chirality
(of the rotaxane). The lowest energy tetrahedral intermediates were
modeled (see Supporting Information) using
(a) electrophile 9 or (b) electrophile 12. The di(alkoxyl)naphthalene ring of the macrocycle and bis(trifluoromethyl)benzene
unit originating from the nucleophile π-stack, causing the nitro-catechol
ring to be positioned so as to cover one face of the tetrahedral center
of the intermediate. This thermodynamically favored arrangement of
components ensures the different handedness of the pseudoenantiomeric
leaving groups is well-expressed in the diastereomeric transition
states, resulting in enantioselectivity in the mechanically planar
chiral rotaxane product. Hydrogen bonds are indicated by black dotted
lines.Also consistent with the stacking
of the electron-rich naphthalene
unit with the electron-poor aryl group of the nucleophile providing
the driving force for organization of the transition state is the
experimental evidence that decreasing the electron density of the
other aromatic ring of the macrocycle increases the enantioselectivity
of rotaxane formation (i.e., 12% e.e. for (+)-8; 40%
e.e. for (+)-11). The less electron-rich the catechol
ring is, the less it competes with the naphthalene group for π-stacking
with the bis(trifluoromethyl)benzylamine and so
the greater the enantiodiscrimination in the transition state.
Conclusions
The examples presented demonstrate that mechanically planar chiral
rotaxanes can be directly accessed in up to 50% e.e. in a single synthetic
step. The chirality of the point-chiral leaving group is transferred
into mechanically planar chirality in the rotaxane through metal-free
active template N-acylation. Pseudoenantiomeric cinchona
alkaloids allow either rotaxane enantiomer to be accessed. X-ray crystallography
and molecular modeling suggest that the origin of the enantioselectivity
lies in π-stacking of an electron-rich aromatic ring on the
macrocycle with an electron-poor aryl group originating from the nucleophilic
axle building block. This positions the second aromatic ring of the
macrocycle in an orientation that blocks one face of the electrophile.
Simple methods for accessing enantioenriched mechanically planar chiral
rotaxanes should improve their availability for investigation in applications
such as asymmetric catalysis,[7] chiral (bio)molecule
sensing,[1,6,22] and novel
designs[23] of molecular machinery.
Authors: Marcel Dommaschk; Javier Echavarren; David A Leigh; Vanesa Marcos; Thomas A Singleton Journal: Angew Chem Int Ed Engl Date: 2019-09-10 Impact factor: 15.336
Authors: Chong Tian; Stephen D P Fielden; George F S Whitehead; Iñigo J Vitorica-Yrezabal; David A Leigh Journal: Nat Commun Date: 2020-02-06 Impact factor: 14.919
Authors: Arnau Rodríguez-Rubio; Andrea Savoini; Florian Modicom; Patrick Butler; Stephen M Goldup Journal: J Am Chem Soc Date: 2022-06-28 Impact factor: 16.383
Authors: Lorna Binks; Chong Tian; Stephen D P Fielden; Iñigo J Vitorica-Yrezabal; David A Leigh Journal: J Am Chem Soc Date: 2022-08-18 Impact factor: 16.383