Selective modification of β-cyclodextrin: An unexpected tandem reaction enables the cross-linking of C2A and C2B via a sulfur atom

Article abstract of DOI:10.1039/b703943c

2^A,3^A-Alloepithio-2^B-sulfonyl-β- cyclodextrin undergoes a tandem reaction to generate an unprecedented C2 ^A-S-C2^B-bridged glucosyl-3^A,6 ^A-anhydroglucoside segment. The Royal Soci

Full text of DOI:10.1039/b703943c

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Selective modification of b-cyclodextrin: an unexpected tandem
reaction enables the cross-linking of C2A and C2B via a sulfur atom
Makoto Fukudome,a Kazuki Yoshikawa,a Kazutaka Koga,b De-Qi Yuan*a and Kahee Fujita*a
Received (in Cambridge, UK) 15th March 2007, Accepted 3rd May 2007
First published as an Advance Article on the web 22nd May 2007
DOI: 10.1039/b703943c
2A,3A-Alloepithio-2B-sulfonyl-b-cyclodextrin undergoes a tan-
of residues A and B (Table 1), indicated that the two peaks at d
48.2 and 54.1 ppm correspond to the C2A of the residue A and the
C2B of the residue B, respectively. The formation of a single
S-bridge between the C2A and C2B was strongly suggested by the
appearance of strong HMBC signals between C2A and H2B as well
as between H2A and C2B (Fig. 1). The bridging of C3Aand C6A by
an ether bond was suggested by the large downfield shifts of the
two carbons and the strong HMBC signal between C6A and H3A.
The residue A closely resembles the modified residue of 3A,6A-
anhydro-2A-deoxy-2A-thio-b-CD6 both in chemical shifts and
coupling patterns (H1A y 6A: d 5.30, 3.48, 4.53, 3.99, 4.50, 3.92
and 4.22; J1,2, J2,3, J3,4, J4,5, J5,6, J5,69, and J6,69 = ca. 4.8, 3.5, 2.2,
2.4, 0, and 11.2 Hz, respectively). Therefore, the residue A should
be a 3A,6A-anhydro-2A-deoxy-2A-thioglucoside with the S-atom
linked to C2B of the adjacent residue. Finally, the structure of the
residue B was deduced on the basis that the chemical shifts of C3B
(d 70.5 ppm) and C2B (d 54.1 ppm) of 5 are very close to the
corresponding values d 72.0 and 51.5 ppm of 2-benzylthio-2-
dem reaction to generate an unprecedented C2A–S–C2B-
bridged glucosyl-3A,6A-anhydroglucoside segment.
Hetero-bifunctionalization of cyclodextrin (CD) has attracted
much attention in artificial enzyme development.1 However, even
CDs hetero-bifunctionalized on the primary hydroxyl sides are
very difficult to access,2 except in only one case.3 Moreover, no
hetero-bifunctionalization of the secondary hydroxyl sides had
been reported until we recently demonstrated the efficient synthesis
of 2 from 2A,2B-O,O-di(mesitylenesulfonyl)-b-CD 1 (Scheme 1).4
Herein we report the selective synthesis of alloepithio-sulfonate
3 and an unexpected tandem reaction of 3 to generate the first
tetracyclic residue within the macrocycle belt.
As part of a continuing project into the hetero-bifunctionaliza-
tion of CDs, we attempted to synthesise the target structure 4.
Compound 3 can be obtained from 2 in 67.8% yield by stirring 2
and thiourea in 0.1 HCl solution at 70 uC, followed by treatment
with NaHCO3 (cf. Table 1 for the structural data).5{ When treated
by 1 M NaOH at 70 uC, 3 underwent a very clean reaction,
affording a pure product in 73% yield.{ However, the structural
assignment revealed that this product takes the structure of 5
rather than the expected 4!
deoxy-b-CD.7 The small J1,2 and large J2,3 coupling constants
(2.5 and 9.6 Hz) are consistent with equatorial–axial and trans-
diaxial couplings, respectively, ruling out the formation of a
mannoside structure via direct displacement of the mesitylenesul-
fonate group by the sulfur atom, which would result in a
configuration inversion of C2.
3
3
The TOF mass spectrum demonstrated a peak at m/z 1137
which is consistent with the [M + Na] ion of both 4 and 5. The
evidence for the structure of 5 was gathered from detailed NMR
spectral analysis. As shown in Fig. 1, only two carbon signals can
be recognized in the range d 30–60 ppm where the 2,3-epoxy- or
thio-bearing carbons should appear, one peak less than expected
The formation of 5 is somewhat astonishing because the C2
carbons of two adjacent saccharide residues in CD derivatives are
normally separated far beyond the reachable range of a single
atom! The bridging of C2A and C2B by a single sulfur atom should
induce significant distortion of the CD cavity, which has proved
interesting in defining the accommodation of guest molecules in
the CD cavity.8 The bridging of two adjacent methylene carbons
of a-CD and cross linking of the C3x and C2x+1 of b-CD by a
single atom have been reported.9 The present work represents the
first example of bridging two neighboring C2 carbons by a single
atom.
1
1
from the structure of 4. H–1H and H–13C COSY experiments,
which enabled the extraction of chemical shifts relating the nuclei
aDepartment of Molecular Medicinal Sciences, Graduate School of
Biomedical Sciences, Nagasaki University, Bunkyo-machi 1-14,
Nagasaki 852-8521, Japan. E-mail: fujita@nagasaki-u.ac.jp;
deqiyuan@nagasaki-u.ac.jp
bNihon Pharmaceutical University, Kita-adachi, Saitama 362-0806,
Japan
A plausible mechanism can be drawn for the formation of 5 by
taking 6 as the reaction intermediate whose formation is followed
Scheme 1 An unexpected tandem reaction enabled construction of a tetracyclic segment within the cyclodextrin belt (Mess = mesitylenesulfonyl).
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 3157–3159 | 3157

View Article Online
Table 1 1H and 13C NMR data relating the units A and B of 2, 3, 5 and 6 at 35 uC in D2Oa
1H, d/ppm (coupling constant 3J/Hz) 13C, d/ppm
Compound 2 Compound 3 Compound 5 Compound 6
Compound 2
Compound 3
Compound 5
Compound 6
Nuclei
1A
2A
3A
4A
5A
6A
5.15 s
5.39 (5.3)
ca. 3.55
3.21 dd (6.4, 4.3) 4.77 dd (3.5, 5.4)
4.00 dd (4.3, 9.2) 4.34 dd (5.4, 3.3)
5.30 d (4.8)
3.41 dd (4.8, 3.5)
5.49 d (5.2)
ca. 3.84
ca. 3.72
98.5
50.2
55.2
94.6
40.7
38.7
74.6
68.3
99.2
48.2
77.7
ca. 73.7
76.8
69.6
95.0
40.9
38.7
73.5
68.1
61.4
3.32 d (3.7)
2.29 d (3.7)
ca. 3.46
4.42 dd (4.4, 9.3) ca. 72.8
70.3
ca. 3.50
ca. 3.50
4.41 dd (3.3, 2.4)
3.92 dd (2.4, 11.2)
4.18 d (11.2)
ca. 3.66 m
1B
2B
3B
4B
5B
a
5.08 d (3.7)
5.20 d (3.4)
4.44 dd (3.7, 9.7) 4.50 dd (3.4, 9.8) 2.99 dd (2.5, 9.6)
5.32 d (2.5)
5.38 s
3.35 d (3.6)
ca. 3.52
99.9
79.2
70.9
79–82
98.1
78.6
71.3
82.2
97.9
54.1
70.5
79.0
95.6
50.4
54.2
4.01 t (9.7)
ca. 3.58
4.04 t (9.8)
ca. 3.60
4.25 t (9.6)
ca. 3.57
4.20 dbt (10.3, 3.4) ca. 3.66 m
ca. 72.0
70.3
The marks s, d, t, dd, dbt, and m denote singlet, doublet, triplet, double doublet, double broad triplet, and multiplet, respectively.
Fig. 1 1H–13C HMBC spectrum of the tetracyclic segment-incorporated cyclodextrin 5 in D2O.
by a tandem reaction initiated by the attack of 6A-OH on the C3A
of the epithio residue. The ring opening of the latter generates an
thiolate anion (42) which attacks the C2B of the neighboring
2B,3B-mannoepoxy residue to form compound 5 (Scheme 2).
The intermediate 6 was isolated in 84.4% yield by performing
the reaction in a pH 12 buffer solution at rt (Table 1).{ Treatment
of 6 with a NaOH solution at 70 uC afforded 5, confirming that
the sulfur atom is introduced to the C2B via epoxide ring opening
of the unit B.
The attack of 2A-S2 on the 2B,3B-mannoepoxide in intermediate
42 was hardly supposed to occur because of the long distance
between the C2B and the S2 group. However, trapping this
intermediate was not successful, which may imply that 42 is highly
reactive and does not accumulate in the reaction. Molecular
model speculation indicated that the 3A,6A-anhydro-2A-deoxy-2A-
thioglucoside residue takes the 1C4 conformation while the normal
4
glucoside residues prefer the C1 conformation. The combination
of the 1C4 conformation of the residue A with the 4C1
conformation of the residue B generates a bridge consisting of
two axial C–O bonds between the two residues, which is different
from the normal axial–equatorial linkage that bridges two
neighboring glucoside residues in CD molecules. Such an axial–
axial linkage makes the AB disaccharide segment sharply bent,
allowing the 2A-SH, which is axially disposed inside the CD cavity,
to get close to the inner side of the residue B. This means that
the 2A-SH may not be as far from the residue B as it was
supposed. This inference is evidenced by the 1H-NMR spectrum of
Scheme 2 A plausible pathway for the formation of 5 from 3.
3158 | Chem. Commun., 2007, 3157–3159
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Fig. 2 ROESY NMR spectrum of 5 (23 mM) and sodium p-nitrophenolate (120 mM) in D2O.
3A,6A-anhydro-2A-deoxy-2A-thio-b-CD6 which demonstrated an
unexpected, large downfield shift (Dd ca. 0.3 ppm) and an
abnormal double triplet coupling pattern (J = 10, 2.7 Hz) for the
H5B of the unmodified residue B. Such influence should stem from
the 2A-SH of the residue A because it is similarly demonstrated by
5 (H5B at d 4.20 ppm, double triplet, J = 10.3, 3.4 Hz) but not by
3A,6A-anhydro-b-CD.10 Therefore, it is clear that the 2A-S2 of 42
should be located in close proximity to the residue B. As soon as it
is generated it attacks the C2B of the 2B,3B-mannoepoxy residue to
complete the tandem reaction.
the tandem reaction was demonstrated to have the ability to finely
control the orientation and location of the guest molecules bound
in its cavity.
Notes and references
{ Synthesis of 3: A modification of the previously reported procedure was
employed to synthesize 3 from 2 (67.8%).
Synthesis of 5: A solution of 3 (100 mg, 0.076 mmol) in 1 M NaOH
(10 ml) was stirred at 70 uC for 4 h. After cooling in a cold water bath, the
reaction mixture was neutralized with HCl, membrane filtered, and
chromatographed on a Lobar column (Rp-18, size B), eluting with water
(500 ml) and then a gradient elution from water to 30% aq. EtOH (500 ml
for each) to give 5 (62 mg, 73%).
Synthesis of 6: A solution of 3 (1.5 g, 1.1 mmol) in phosphate buffer
(150 ml, pH 12.0, 0.1 M) was stirred at rt. After the complete disappearance
of 3 (it took one day) was confirmed by TLC, the reaction was terminated
by neutralization with HCl. Routine work-up of the reaction solution
afforded 6 (1.07 g, 84.4%).
The 3,6-bridge of sugar residue A distorts the cavity shape by
straightening the A–G disaccharide and bending the A–B
disaccharide sharply while the C2A–S–C2B bridge narrows
significantly the cavity portal of the secondary side. Both units
A and B turned their secondary side toward the cavity. Such
distortion of the cavity slightly decreases the binding affinity of 5
toward p-nitrophenol (PNP, Ka = 160 M21 in aqueous solution,
ca. half that of b-CD). However, 5 exhibits more refined binding:
both the orientation and binding depth of the guest in the cavity
are exactly controlled. Upon binding PNP, 5 demonstrated
1 R. Breslow, Artificial Enzymes, Wiley-VCH, Weinheim, 2005.
2 (a) E. Fasella, S. D. Dong and R. Breslow, Bioorg. Med. Chem., 1999, 7,
709–714; (b) K. Fujita, H. Yamamura, Y. Egashira and T. Imoto,
Tetrahedron Lett., 1992, 33, 3511–3514.
3 D.-Q. Yuan, T. Yamada and K. Fujita, Chem. Commun., 2001,
2706–2707.
1
significant changes in its H NMR spectrum (Fig. 1 for free 5
and Fig. 2 for the PNP?5 complex). Assignments of the signals
allowed the mapping of the anisotropic effects of PNP on 5:
residues F and B in the deshielding region and residues G, E and D
in the shielding region. The meta-H of the PNP exhibited strong
NOE correlations with the H5 and H3 protons of 5, while the
ortho-H of the PNP correlated only to the H3 protons of 5,
both with the strongest correlation going to residue F. These
observations clearly indicated that the PNP molecule was
accommodated in the cavity of 5 by directing its NO2 group to
the primary side and the two edges of the benzene ring toward the
residues F and B, respectively. The PNP molecule did not bind
very deeply in the cavity but at such a depth that the proton pairs
4 M. Fukudome, Y. Sugimoto, D.-Q. Yuan and K. Fujita, Chem.
Commun., 2005, 3168–3170.
5 M. Fukudome, T. Shiratani, S. Immel, Y. Nogami, D.-Q. Yuan and
K. Fujita, Angew. Chem., Int. Ed., 2005, 44, 4201–4204.
6 M. Fukudome, Y. Okabe, M. Sakaguchi, H. Morikawa, T. Fujioka,
D.-Q. Yuan and K. Fujita, Tetrahedron Lett., 2004, 45, 9045–9048.
7 D.-Q. Yuan, T. Tahara, W.-H. Chen, Y. Okabe, C. Yang, Y. Yagi,
Y. Nogami, M. Fukudome and K. Fujita, J. Org. Chem., 2003, 68,
9456–9466.
8 K. Fujita, W.-H. Chen, D.-Q. Yuan, Y. Nogami, T. Koga, T. Fujioka,
K. Mihashi, S. Immel and F. W. Lichtenthaler, Tetrahedron:
Asymmetry, 1999, 10, 1689–1696; W.-H. Chen, M. Fukudome,
D.-Q. Yuan, T. Fujioka, K. Mihashi and K. Fujita, Chem. Commun.,
2000, 541–542.
9 K. Fujita, T. Fujioka, H. Shimada, K. Ohta, A. Yoshino, Y. Okabe,
M. Fukudome and D.-Q. Yuan, Eur. J. Org. Chem., 2004, 3113–3118;
E. Engeldinger, L. Poorters, D. Armspach, D. Matt and L. Toupet,
Chem. Commun., 2004, 634–635; B. Benmerad, P. Clair, D. Armspach,
D. Matt, F. Balegroune and L. Toupet, Chem. Commun., 2004,
634–635.
…
…
of meta-H ortho-H and H5 H3 of 5 form an offset arrangement
with the meta-H being located between the H5 and H3.
In conclusion, we have described an unexpected tandem
reaction on a hetero-bifunctional b-CD, which allows the
construction of a tetracyclic structural segment within the CD
belt and a unique, distorted cavity. The reaction mechanism was
elucidated on the basis of experimental evidence. The product of
10 K. Fujita, H. Yamamura, T. Imoto and I. Tabushi, Chem. Lett., 1988,
543–546.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 3157–3159 | 3159