A small molecule diacid with long-term chiral memory

Article abstract of DOI:10.1021/ol900955q

A small, axially chiral diacid was designed with chiral memory based on restricted rotation. Heating a racemic sample with a chiral alkaloid led to an enantiomeric excess of up to 40% ee. The guest-induced chirality was preserved on cooling to rt, which w

Full text of DOI:10.1021/ol900955q

ORGANIC
LETTERS
2009
Vol. 11, No. 12
2599-2602
A Small Molecule Diacid with Long-Term
Chiral Memory
Roger D. Rasberry, Xiangyang Wu, Brooke N. Bullock, Mark D. Smith, and
Ken D. Shimizu*
Department of Chemistry and Biochemistry, UniVersity of South Carolina,
Columbia, South Carolina 29208
shimizu@mail.chem.sc.edu
Received April 30, 2009
ABSTRACT
A small, axially chiral diacid was designed with chiral memory based on restricted rotation. Heating a racemic sample with a chiral alkaloid
led to an enantiomeric excess of up to 40% ee. The guest-induced chirality was preserved on cooling to rt, which was maintained even in the
absence of guest (t1/2 ) 14y). The chiral enrichment process was also reversible, allowing the diacid to be used as a chiral switch.
The efficient transfer of chiral information from one molecule
to another is an important process for many applications
including asymmetric transformations and enantiomeric
sensing.1 For example, flexible receptors have been devel-
oped with the ability to adopt chiral conformations in the
presence of chiral guests. This induced-fit mechanism has
been used to measure ee and to assign absolute stereochem-
istry.2 The utility of this approach, however, has been limited
by the inability of the flexible hosts to maintain their chiral
conformations in the absence of guest. Recently, a number
of receptors and supramolecular systems with chiral memory
have been developed.3 In this paper, we present a small
molecule diacid receptor 1 with excellent long-term chiral
memory (Scheme 1). Heating 1 with a chiral alkaloid guest
leads to enantiomeric enrichment, which is preserved on
cooling to rt due to restricted rotation about the central
Caryl-Nimide single bond.
Chiral memory systems represent a direct and potentially
efficient route to optically active molecules and chiral
molecular switches.1d,4 The chiral templation process does
not involve any bond-forming steps. In addition, the chiral
guest is not consumed in the process and can be recovered
and reused. The practical utility of chiral memory systems,
however, has been limited by their poor stability at rt and
by their large size. For example, most chiral memory systems
will racemize within minutes or days at rt.3a-o The receptors
are also usually large polymeric or self-assembled
structures.3d-r Chiral memory system 1 addresses both of
these limitations. The rotational barrier of diacid 1 is
(1) (a) Maeda, K.; Yashima, E. Top. Curr. Chem. 2006, 265, 47–88.
(b) Allenmark, S. Chirality 2003, 15, 409–422. (c) Tsukube, H.; Shinoda,
S. Chem. ReV. 2002, 102, 2389–2403. (d) Feringa, B. L.; van Delden, R. A.
Angew. Chem., Int. Ed. 1999, 38, 3419–3438.
(2) (a) Hembury, G. A.; Borovkov, V. V.; Inoue, Y. Chem. ReV. 2008,
108, 1–73. (b) Borovkov, V. V.; Inoue, Y. Top. Curr. Chem. 2006, 265,
89–146. (c) Amabilino, D. B.; Veciana, J. In Encyclopedia of Supramo-
lecular Chemistry; Atwood, J. L., Steed, J. W., Eds.; Marcel Dekker: New
York, 2004; p 245.
10.1021/ol900955q CCC: $40.75
Published on Web 05/20/2009
 2009 American Chemical Society

Alternatively, heating 1 with an alkaloid guest of opposite
chirality can directly yield the opposite enantiomer of diacid
1.
Racemic diacid 1 was prepared in a single step (Scheme
2) by thermal condensation of an aniline (2-amino-3-
Scheme 1. Three-Step Chiral Memory Cycle for Diacid 1 Using
the Alkaloids Quinine or Quinidine (Blue Ovals)
Scheme 2. Synthesis of rac-Diacid 1
methylbenzoic acid) with a cyclic anhydride (3-carboxyph-
thalic anhydride). The reaction proceeds quantitatively in the
solid state at 150 °C for 24 h. Structural characterization of
rac-1 was established by X-ray crystallographic analysis,
which verified the formation of the cyclic imide ring and
the twisted axially chiral structure (Supplementary Info).7
The steric interaction of the aryl methyl and carboxylic
groups with the opposing imide carbonyls forces the aryl
and phthalimide surfaces out of plane (57°), forming enan-
tiomeric rotamers.
sufficiently high (29.6 kcal/mol) that it can maintain its
optical activity for months at rt without measurable isomer-
ization. Diacid 1 is also a small organic molecule, which
opens up applications as a chiral building block, auxiliary,
or catalyst.5
The chiral memory properties of diacid 1 are a conse-
quence of restricted rotation about its central Caryl-Nimide
single bond,6 which generates enantiomeric rotamers (atro-
pisomers) that are stable at room temperature. Thus, at
elevated temperatures, the axially chiral rotamers of 1 are
in equilibrium, and the equilibrium can be biased by
complexation with a chiral guest. On cooling to rt, restricted
rotation is reestablished, which “locks in” the guest-induced
enantiomeric enrichment and allows the guest to be removed.
The entire process is also reversible, and heating diacid 1 in
the absence of guest “erases” the enantiomeric excess.
The chiral memory properties of diacid 1 were initially
studied using CD spectroscopy (Figure 1). First, rac-1 was
(3) (a) Furusho, Y.; Kimura, T.; Mizuno, Y.; Aida, T. J. Am. Chem.
Soc. 1997, 119, 5267–5268. (b) Sugasaki, A.; Ikeda, M.; Takeuchi, M.;
Robertson, A.; Shinkai, S. J. Chem. Soc., Perkin Trans. 1 1999, 3259–
3264. (c) Mizuno, Y.; Aida, T.; Yamaguchi, K. J. Am. Chem. Soc. 2000,
122, 5278–5285. (d) Bellacchio, E.; Lauceri, R.; Gurrieri, S.; Scolaro, L. M.;
Romeo, A.; Purrello, R. J. Am. Chem. Soc. 1998, 120, 12353–12354. (e)
Prins, L. J.; De Jong, F.; Timmerman, P.; Reinhoudt, D. N. Nature 2000,
408, 181–184. (f) Rivera, J. M.; Craig, S. L.; Martin, T.; Rebek, J. Angew.
Chem., Int. Ed. 2000, 39, 2130–2132. (g) Ishi-i, T.; Crego-Calama, M.;
Timmerman, P.; Reinhoudt, D. N.; Shinkai, S. J. Am. Chem. Soc. 2002,
124, 14631–14641. (h) Aimi, J.; Oya, K.; Tsuda, A.; Aida, T. Angew. Chem.,
Int. Ed. 2007, 46, 2031–2035. (i) Rosaria, L.; D’Urso, A.; Mammana, A.;
Purrello, R. Chirality 2008, 20, 411–419. (j) Onouchi, H.; Miyagawa, T.;
Morino, K.; Yashima, E. Angew. Chem., Int. Ed. 2006, 45, 2381–2384. (k)
Morino, K.; Watase, N.; Maeda, K.; Yashima, E. Chem.sEur. J. 2004, 10,
4703–4707. (l) Ishikawa, M.; Maeda, K.; Mitsutsuji, Y.; Yashima, E. J. Am.
Chem. Soc. 2004, 126, 732–733. (m) Mammana, A.; De Napoli, M.; Lauceri,
R.; Purrello, R. Biorg. Med. Chem. 2005, 13, 5159–5163. (n) Welch, C. J.
J. Chromatogr. A 1995, 689, 189–193. (o) Mateos-Timoneda, M. A.; Crego-
Calama, M.; Reinhoudt, D. N. Chem. Soc. ReV. 2004, 33, 363–372. (p)
Yashima, E.; Maeda, K.; Okamoto, Y. Nature 1999, 399, 449–451. (q)
Ziegler, M.; Davis, A. V.; Johnson, D. W.; Raymond, K. N. Angew. Chem.,
Int. Ed. 2003, 42, 665–668. (r) Lauceri, R.; Raudino, A.; Scolaro, L. M.;
Micali, N.; Purrello, R. J. Am. Chem. Soc. 2002, 124, 894–895.
(4) (a) Feringa, B. L.; van Delden, R. A.; Koumura, N.; Geertsema,
E. M. Chem. ReV. 2000, 100, 1789–1816. (b) Guo, P. Z.; Zhang, L.; Liu,
M. H. AdV. Mater. 2006, 18, 177–180.
Figure 1. CD spectra of rac-1 in CH3CN after heating (neat, 125
°C, 24 h) with 2 equiv of quinine (red line) or 2 equiv of quinidine
(blue line). The CD spectra were taken after removal of the chiral
guest by extraction with 0.25 N HCl. The green line is the spectrum
of quinine-templated 1 after heating in the absence of guest (neat,
125 °C, 24 h).
heated neat with quinine or quinidine (2 equiv, 125 °C, 24 h).
The mixture was cooled to rt, dissolved in EtOAc, and
washed with 0.25 N HCl to remove the basic alkaloid guest.
Figure 1 shows the CD spectra of samples of 1 heated neat
with quinine (red line) and quinidine (blue line), respectively.
The Cotton effects in both spectra were assigned to diacid
1, as the CD signals cross from positive to negative ellipticity
(5) Chin, J.; Chong, Y. S.; Bobb, R.; Studnicki, L.; Hong, J. I. Chem.
Commun. 2007, 120–122.
(6) (a) Kishikawa, K.; Yoshizaki, K.; Kohmoto, S.; Yamamoto, M.;
Yamaguchi, K.; Yamada, K. J. Chem. Soc., Perkin Trans. 1 1997, 1233–
1239. (b) Verma, S. M.; Singh, N. B. Aust. J. Chem. 1976, 29, 295–300.
(7) Barooah, N.; Baruah, J. B. Mini-ReV. Org. Chem. 2007, 4, 292–
309.
2600
Org. Lett., Vol. 11, No. 12, 2009

at the adsorption maximum of diacid 1 (234 nm). Quinine
and quinidine favor opposing enantiomers of diacid 1, which
is consistent with opposing chirality of the two key chiral
centers in these diastereomeric guests. This could be seen
by the near mirror image Cotton effects in the CD spectra
of the quinine and quinidine heated samples of 1.8 The
reversibility of the chiral templation process was demon-
strated by heating samples of quinine-templated 1 in the
absence of guest (neat, 125 °C, 24 h), which led to a complete
loss of optical activity (Figure 1, green line).
Table 1. Chiral Imprinting Studies Monitored by Chiral HPLC
entry
chiral guesta
conditionsb
% ee of 1c (major)
1
2
3
4
5
6
7
quinine
neat
neat
neat
neat
neat
neat
5 mM TCE
40 (-)
40 (+)
2 (-)
8 (+)
0
quinidine
cinchonidine
cinchonine
nicotine
(-)-sparteine
quinine
0
0
The excellent kinetic stability of the rotamers of 1 provided
the opportunity to more accurately monitor the chiral memory
properties using chiral HPLC. This is in contrast to most of
the previous chiral memory systems that could only be
monitored by in situ methods such as CD and chiral
NMR.3a-c Diacid 1 was too polar for direct chromatographic
analysis. Thus, samples of diacid 1 were esterified with an
excess of phenyldiazomethane (1 mM, toluene) prior to
analysis by chiral HPLC. The resulting dibenzyl ester 2
displayed clean baseline resolution (Figure 2). In this manner,
a Two equivalents of guest. b Experiments were conducted at 125 °C
(24 h). c The % ee of 1 was determined by chiral HPLC (Chiralpak IC)
after treatment with phenyldiazomethane.
differ by only a methoxy group. One possible explanation
is that these small structural differences are magnified in the
solid-state environment. Support for this hypothesis was
provided by the low ee of the analogous studies carried out
in solution (1,1,2,2-tetrachloroethane, TCE) (Table 1, entry
7).
The reversibility of the chiral templation process and the
stability of the chiral products were also verified by chiral
HPLC. A 40% ee sample of (-)-diacid 1 (Figure 2, red line)
became racemic after heating in the absence of chiral guest
(green line). This racemic sample could also be switched to
the opposite enantiomer by heating with quinidine, yielding
40% ee of the (+) enantiomer (blue line).
The ability to quantitatively measure the ee of diacid 1
also enabled an accurate measure of its Caryl-Nimide rotational
barrier. A sample of enantiomerically enriched 1 was heated
neat at 100 °C, and the racemization was followed by chiral
HPLC. The data was fitted to a first-order kinetic analysis
(Supporting Information) to yield a rate constant of 6.5 ×
10-5 s-1, which calculates to a rotational barrier of 29.6 kcal/
mol.9 This high barrier was consistent with the observed
stability of the enantiomers of 1 at rt and equates to a half-
life of 14 years at 23 °C. At the same time, the barrier was
not too high, as isomerization could be facilitated by heating
at a moderate temperature (t1/2 ) 14 min at 125 °C). An
almost identical barrier was measured for diacid 1 (29.4 kcal/
mol) in solution (TCE). Thus, enantiomerically enriched 1
was kinetically stable both in solution and in the solid-state.10
Control studies confirmed that enriched samples of 1 were
stable at rt as they did not show any measurable change in
ee after 2 months in solution (CH2Cl2 or EtOAc) or in the
solid state. Similar studies established that the dibenzyl ester
2 also had a high rotational barrier (∆G ) 28.2 kcal/mol),
confirming that isomerization was not occurring during the
HPLC analyses.
Figure 2. Staggered chromatograms of the chiral HPLC analyses
of samples of rac-1 (black line), which was sequentially heated
(neat, 24 h, 125 °C) with 2 equiv of quinine (red line), without
guest (green line), and with 2 equiv of quinidine (blue line). Prior
to HPLC analysis, each sample of 1 was washed with 0.25 N HCl
to remove the alkaloid guests and derivatized with phenyldiaz-
omethane.
the chiral templating efficiencies of the quinine and quinidine
with diacid were quantitatively assessed. The results cor-
roborated the CD studies as quinine and quinidine templated
the formation of opposite enantiomers of 1. Samples of 1
heated with quinine had a 40% ee of the faster eluting (-)-
1. Samples of 1 heated with quinidine had a 40% ee of the
slower eluting (+)-1.
The templating abilities of other commercially available
chiral alkaloids were also tested (Table 1). However, they
all displayed lower chiral imprinting efficiencies than quinine
and quinidine. Nicotine and sparteine showed no chiral
induction. The cinchonidine and cinchonine diastereomers
displayed poor templating efficiencies with low ee’s (2% and
8% ee). This was surprising since cinchonidine and quinine
To study the mechanism of the chirality-transfer process,
samples of rac-1 were heated with varying stoichiometries
of quinine and quinidine (Figure 3a). Both alkaloids showed
very similar sigmoidal ee profiles, reaching a plateau at 2
(8) To verify that the CD spectra were due to enantiomeric enrichment
of 1 and not residual quinine or quinidine, the CD spectra of 1/quinine and
1/quinidine mixtures were measured (Supplementary Information). These
samples had very different spectra and did not display a Cotton effect.
(9) Stewart, W. E.; Siddall, T. H. Chem. ReV. 1970, 70, 517–551.
(10) (a) Colombini, D.; Merle, G.; Alberola, N. D. Macromolecules 2001,
34, 5916–5926. (b) Park, J. W.; Green, M. M.; Morawetz, H. Macromol-
ecules 2001, 34, 5719–5722.
Org. Lett., Vol. 11, No. 12, 2009
2601

ing in solution. Alternatively, the success in toluene may be
related to its lower polarity, which enhances the intermo-
lecular interactions.
Finally, the more thermally stable diacid 1 still retains its
dynamic properties, as the chiral memory can be “written”,
“erased”, and “rewritten.” This was demonstrated by its use
as a chiral molecular switch.4 A sample of 1 was repeatedly
heated with 2 equiv of quinidine and then with 2 equiv of
quinine (Figure 3b). After each heating cycle, the samples
were dissolved in EtOAc and washed with 0.25 N HCl to
remove the guest. An aliquot was removed for analysis by
chiral HPLC, and the remaining material was used in
subsequent heating cycles. After three cycles of heating with
quinidine then quinine, no significant loss in material or in
the efficiency of the chiral induction was observed. This
study demonstrates the excellent fidelity of the switching
process, which is a consequence of the simplicity of the
switching process that does not involve any covalent bond-
forming or -breaking steps.12
In summary, a chiral atropisomer 1 was designed with
chiral memory. The chiral templation process is efficient as
it is an example of a dynamic kinetic resolution where the
enantiomeric enrichment of a host by a chiral guest is
kinetically trapped by restricted rotation.13 The templation
process is also reversible, yielding products with dynamic
properties that can be used in applications such as chiral
molecular switches. The present system is distinguished from
previous chiral memory systems in two key respects. First,
atropisomeric receptor 1 is a small organic molecule, which
facilitates its application to the synthesis of chiral building
blocks and catalysts. Second, the guest-induced enantiomeri-
cally enriched state displays excellent kinetic stability.
Finally, the ee’s of this chiral templation strategy are still
moderate. Low ee’s were observed for the first reports of
asymmetric transition metal and organic catalysts, which have
now become well-established routes to chiral materials. We
expect that the ee’s will improve with continued optimization
and development of this strategy.
Figure 3. (a) Measured ee of diacid 1 when heated with varying
equivalents of quinine (blue line) and quinidine (red). (b) Percent
ee for three cycles of heating 1 (neat, 125 °C, 24 h) with quinidine
(2 equiv) and then with quinine (2 equiv).
equiv. This is consistent with the formation of a 2:1 (guest/
1) complex, in which the more basic bicyclic amine of two
different quinoline alkaloids forms a salt with each carboxylic
acid of diacid 1. The chiral induction appears to be highly
cooperative as stoichiometries lower than 1 equiv of guest
gave virtually no chiral induction.11 Thus, the corresponding
1:1 (alkaloid/1) complex is not very efficient in transferring
chiral information. The possibility that the chiral induction
was due to the formation of a highly ordered crystalline state
was examined by powder XRD and DTA analyses of neat
1·(alkaloid)2 mixtures (Supporting Information). All of the
samples were found to be amorphous and did not undergo
any phase changes on heating or cooling, suggesting that
the individual alkaloid/1 complexes were primarily respon-
sible for the chiral induction.
Most recently, we have been able to extend the chiral
imprinting process to the solution phase. Heating 1 with 2
equiv of quinine in toluene (100 °C, 3 d) yielded (-)-1 in
47% ee. One possible explanation for the successful chiral
induction in toluene and not in TCE may be the low solubility
of 1·(alkaloid)2, which only dissolves on heating in toluene.
On cooling, it is possible that the less soluble diastereomeric
salt precipitates while the more soluble salt is still equilibrat-
Acknowledgment. This work was supported by the
National Science Foundation (CHE 0616442).
Supporting Information Available: NMR, HPLC, DTA,
and crystallographic data for 1 and diester 2. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL900955Q
(12) Clayden, J.; Vallverdu, L.; Clayton, J.; Helliwell, M. Chem.
Commun. 2008, 561–563.
(11) Shinkai, S.; Ikeda, M.; Sugasaki, A.; Takeuchi, M. Acc. Chem. Res.
2001, 34, 494–503.
(13) Pellissier, H. Tetrahedron 2003, 59, 8291–8327.
2602
Org. Lett., Vol. 11, No. 12, 2009