Selectively Monomodified Cyclodextrins. Synthetic Strategies

Article abstract of DOI:10.1021/jo991347r

Monomodifications of cyclodextrins to give selectively 2-, 3-, or 6-substituted product is a challenging task because of the number of hydroxyl groups that can potentially react with the incoming reagent. The principles and the methods involved in manipul

Full text of DOI:10.1021/jo991347r

2624
J . Org. Chem. 2000, 65, 2624-2630
Selectively Mon om od ified Cyclod extr in s. Syn th etic Str a tegies
Shengping Tian, Henghu Zhu, Peter Forgo, and Valerian T. D’Souza*
Department of Chemistry, University of Missouri-St. Louis, St. Louis, Missouri 63121
Received August 25, 1999
Monomodifications of cyclodextrins to give selectively 2-, 3-, or 6-substituted product is a challenging
task because of the number of hydroxyl groups that can potentially react with the incoming reagent.
The principles and the methods involved in manipulations of the differences in the chemistry of
these hydroxyl groups to control the outcome of an electrophilic reaction with them to produce
monoalkylated (ether-linkaged) cyclodextrin derivatives are discussed and illustrated.
Cyclodextrins, the torus-shaped cyclic oligomers of R-D-
ization at a desired (2-, 3-, or 6-) position is a challenging
task, the differences in the chemistry among these sites
can be exploited to yield a preponderance of a particular
product.
glucopyranose,1 have have gained prominence in recent
years in diverse fields such as artificial enzymes,2-5 drug
delivery systems,6 and chiral separation media.7 Severe
constraints, such as availability of limited functional
groups for useful chemical processes and the rigid shape
and sizes, of these “structural and functional straight
jackets”8 had initially stymied the progress in these fields.
However, valiant efforts by several distinguished groups
Among the linkages (e.g., sulfide,17,18 amine,19 ester,11
ether,20 etc.) that are available for attachment of func-
tional groups to cyclodextrins, ether bonds are perhaps
the most desirable. They are less susceptible to degrada-
tion by oxidation or hydrolysis and can conceivably be
obtained by a single nucleophilic reaction of a hydroxyl
group of cyclodextrin on an appropriate electrophile. The
principles and the methods involved in manipulation of
the differences in the chemistry of the hydroxyl groups
to control the outcome of an electrophilic reaction with
cyclodextrin to produce monoalkylated (ether linkaged)
cyclodextrin derivatives are illustrated in this paper.
4-Methylamino-3-nitrobenzyl chloride is used as the
electrophilic reagent in this investigation because the
resulting cyclodextrin derivative, as described in the
literature,14 is an important intermediate in the synthesis
of artificial redox enzymes. The synthesis of these
artificial redox enzymes using the cyclodextrin deriva-
tives reported here are the subject of a future paper.
9-11
have ameliorated this situation to the extent that
now cyclodextrins with a variety of shapes, sizes, and
functional groups are available.12 Our efforts in this field
have been directed toward conversion of hydroxyl groups
of cyclodextrins to other useful functional groups.13-15
Among several possible types of modifications12 (mono-,
di-, tri-, and per-), monomodified cyclodextrins have been
particularly notable in the exciting field of artificial
enzymes.16 These monomodifications can conceivably be
achieved by a reaction of the hydroxyl group on an
electrophilic reagent. However, the large number of
hydroxyl groups at three different positions of cyclodex-
trins do not easily lend themselves to modification at a
single desired place. Although selective monofunctional-
(1) Bender, M. L.; Komiyama, M. Cyclodextrin Chemistry; Springer-
Verlag: New York, 1978.
Resu lts a n d Discu ssion
(2) (a) Murakami, Y.; Kikuchi, J .; Hisaeda, Y.; Hayashida, O. Chem.
Rev. 1996, 96, 721. (b) Breslow, R. Acc. Chem. Res. 1995, 28, 146.
(3) Ye, H.; Rong, D.; Tong, W.; D’Souza, V. T. J . Chem. Soc., Perkin
Trans. 2 1992, 2071-2076.
Selective Mon oa lk yla tion a t th e P r im a r y Sid e of
r-Cyclod extr in . It has been noted that the hydroxyl
groups at the primary side of cyclodextrin are the most
basic and very often most nucleophilic.1 Thus, normal
reactivity of the hydroxyl groups dictates that these
should be alkylated in the presence of an appropriate
electrophile in a weakly basic solvent. As described below,
although this is found to be true in the case of R-cyclo-
dextrin, it does not apply to â-cyclodextrin.
A simple method to selectively connect an alkyl group
(4-methylamino-3-nitrobenzyl group) to the primary side
of R-cyclodextrin (1) consists of (Scheme 1) a reaction of
compound 1 with 4-methylamino-3-nitrobenzyl chloride
2 in 2,6-lutidine to give 6-O-(4-methylamino-3-nitroben-
zyl)-R-cyclodextrin (3). During the reaction, a reddish
(4) Ye, H.; Tong, W.; D’Souza, V. T. J . Am. Chem. Soc. 1992, 114,
5470-5472.
(5) Ye, H.; Tong, W.; D’Souza, V. T. J . Chem. Soc., Perkin Trans. 2
1994, 2431-2437.
(6) Uekama, K.; Hirayama, F.; Irie, T. New Trends in Cyclodextrins
and Derivatives; Duch’ne, D., Ed.; Editions de Sante`: Paris, France,
1991; chapter 12.
(7) Hilton, M. L.; Armstrong, D. W. New Trends in Cyclodextrins
and Derivatives; Duch’ne, D., Ed.; Editions de Sante`: Paris, France,
1991; chapter 15.
(8) Ashton, P. R.; Ellwood, P.; Staton, I.; Stoddart, J . F. Angew.
Chem., Int. Ed. Engl. 1991, 30, 80.
(9) Comprehensive Supramolecular Chemistry; Atwood, J . L., Lehn,
J .-M., Eds.; Vol. 3, Cyclodextrins; Szejtli, J ., Osa,T., Eds; Pergamon:
Oxford, 1996.
(10) Gattuso, G.; Nepogodiev, S. A.; Stoddart, J . F. Chem. Rev. 1998,
98, 1959.
(11) Croft, A. P.; Bartsch, R. A. Tetrahedron 1983, 39, 1417.
(12) Khan, R. F.; Forgo, P.; Stine, K.; D’Souza, V. T. Chem. Rev.
1998, 98, 1977.
(13) Rong, D.; D’Souza, V. T. Tetrahedron Lett. 1990, 31, 4275-4278.
(14) Rong, D.; Ye, H.; Boehlow, T. R.; D’Souza, V. T. J . Org. Chem.
1992, 57, 163-167.
(15) Tian, S.; D’Souza, V. T. Tetrahedron Lett. 1994, 50, 9339-9342.
(16) (a) D’Souza, V. T.; Bender, M. L. Acc. Chem. Res. 1987, 20, 146-
152. (b) Breslow, R. Acc. Chem. Res. 1995, 28, 146. (c) Breslow, R.,
Dong, S. Chem. Rev. 1998, 98, 1997.
(17) Sallas, F.; Leroy, P.; Marsura, A.; Nicolas, A. Tetrahedron Lett.
1994, 35, 6079-6082.
(18) Nelles, G.; Weisser, M.; Back, R.; Wohlfart, P.; Wenz, G.; Mitter-
Neher, S. J . Am. Chem. Soc. 1996, 118, 5039-5046.
(19) Petter, R. C.; Salek, J . S.; Sikorsky, C. T.; Kumaravel, G.; Lin,
F.-T. J . Am. Chem. Soc. 1990, 112, 3860.
(20) Takeo, K.; Uemura, K.; Mitoh, H. J . Carbohydr. Chem. 1988,
7, 293.
10.1021/jo991347r CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/31/2000

Selectively Monomodified Cyclodextrins
J . Org. Chem., Vol. 65, No. 9, 2000 2625
Sch em e 1. A sim p le m eth od for con n ectin g a n
a lk yl gr ou p to th e p r im a r y sid e of r-cyclod extr in .
suggest that these three signals belong to C6′, C5′, and
C4′ respectively.
Selective Mon oa lk yla tion a t th e P r im a r y Sid e of
â-Cyclod extr in . As often is the case with well laid out
strategies for modification of cyclodextrins, the successful
method described above for alkylation of the primary side
R-cyclodextrin cannot be applied for a similar transfor-
mation of â-cyclodextrin. Attempts to alkylate the pri-
mary side of â-cyclodextrin (7) by following the above
approach were unsuccessful because 7 is not very soluble
in lutidine, which makes the reaction extremely slow.
TLC of the reaction mixture indicates that more than one
hydroxyl group is modified, which can be attributed to
the reaction of 2 with monomodified product that remains
in solution. Pyridine is known to dissolve â-cyclodextrin
and direct its modification to the primary side in sul-
fonation and silylation reactions.15c However, in pyridine,
compound 2 reacts with the solvent to form pyridinium
salt, which does not alkylate 7 even after an overnight
reflux (bp, 110 °C). The reaction was carried out in a
mixture of solvents containing DMF and 2,6-lutidine (1:1
v/v) in which cyclodextrin is very soluble. However, upon
workup, this reaction produced a mixture of the desired
compound 8 and its regioisomer 9 (Scheme 3). The
formation of these two products can be rationalized by
assuming that 2 forms a complex with 6 in such a way
that the chloromethyl group of 2 is oriented toward the
secondary side of the host as shown in Scheme 3. Thus,
whereas this complex leads to the formation of 9, the
normal reactivity of cyclodextrin leads to the formation
of 8. It was anticipated that the separation of these
isomers would be very difficult, and therefore an indirect
approach via protection of the secondary side was con-
sidered a reasonable choice.23
Sch em e 2. Rea ction m ech a n ism for th e
for m a tion of 6-O-(4-m eth yla m in o-3-
n itr oben zyl)-r-cyclod extr in (5).
intermediate is first generated, which finally converts to
lutidinium hydrochloride (6) as a white precipitate. To
determine the identity of this intermediate, 2 was reacted
with 2,6-lutidine (4) at 120 °C to give the reddish
compound. This compound is highly soluble in water and
gives a white precipitate instantly with silver nitrate,
indicative of the presence of free chloride ions. In a
separate reaction, compound 2 is reacted with 4 at 120
°C in the presence of 1 to form the reddish intermediate,
which reacts further to give 3 only when the temperature
is raised to 130 °C. The formation of 6 in the reaction
mixture is confirmed by comparison of the TLC with an
authentic sample. These observations are consistent with
literature precedence21 and support a reaction mecha-
nism in which 2 reacts with lutidine to give a lutidinium
salt 5, which subsequently alkylates the cyclodextrin
(Scheme 2).
Tentative assignment of the regiochemistry of the
product is made by an examination of the carbohydrate
region in the 13C NMR spectrum (Supporting Informa-
tion, spectrum 1). Five intense peaks in the carbohydrate
region, upon a comparison with the spectrum of R-cyclo-
dextrin and literature precedence, are assigned to un-
modified glucose units at 101.87 (C1), 82.02 (C4), 73.21
(C3), 72.03 (C2 and C5), and 59.90 (C6) ppm. The three
less intensive peaks of the modified glucose unit, which
are indicative of the substitution pattern, are observed
at 68.56, 71.18, and 82.31 ppm. The first signal is 8.66
ppm downfield from the chemical shift of C6 of the
unmodified glucose unit. The other two are 0.85 ppm
upfield and 0.29 ppm downfield from the signals of C5
and C4, respectively, of unmodified glucose units. These
data suggest that 3 is modified at the 6-position of
R-cyclodextrin since the chemical shifts of its CR and Cγ
are expected to move downfield by about 10 and 1 ppm,
respectively, and Câ is expected to move upfield by about
2 ppm when an alcohol is alkylated.22 They further
Initially, heptakis(2,3-di-O-methyl)-â-cyclodextrin (10),
a cyclodextrin derivative whose secondary side is methy-
lated, was selected for further attachment of the desired
group to the primary side. The synthesis 10 is described
in the literature, and methyl ethers are expected to be
stable under our reaction conditions. Compound 2 suc-
cessfully reacts with 10 (Scheme 4) under conditions
similar to those discussed for reaction with R-cyclodextrin
to give heptakis(2,3-di-O-methyl)-6-O-(4-methylamino-3-
nitrobenzyl)-â-cyclodextrin 11. Since the secondary side
of 10 is occupied with methoxy groups, the modification
must take place on its primary side. An analysis of the
13C spectrum further supports this suggestion (Support-
ing Information, spectrum 2). The three less intense
peaks due to the modified glucose unit are at 68.64, 70.45,
and 79.69 ppm. The first one is downfield shifted by 8.56
ppm from the C6, and the second one is upfield shifted
by 1.41 ppm from the C5 of the unmodified glucose units.
These data indicate that the modification has taken place
at the 6-position, The antiphase appearance of the signal
at 68.64 ppm in APT-NMR (Supporting Information,
spectrum 3) suggests that this signal belongs to the
methylene group at the 6-position of the modified glucose
unit.
(22) These data are obtained by manipulating the chemical shift
parameters listed in a text book by Sillvertein et al. and are confirmed
with assignments of 13C NMR spectra of compounds 58 and 59.
Silverstein, R.; Bassler, G. C.; Morrill, T. C. Spectral Identification of
Organic Compounds, 5th ed.; J ohn Wiley & Sons: New Yoek, 1991; p
236.
(21) Clarke, K.; Rothwell, K. J . Chem. Soc. 1960, 1885.
(23) Tian. S.; D’Souza, V. T. Tetrahedron Lett. 1994, 35, 9339.

2626 J . Org. Chem., Vol. 65, No. 9, 2000
Sch em e 3. Rea ction of th e sa m e electr op h ile 2 w ith â-cyclod extr in .
Tian et al.
Sch em e 4. Alk yla tion of th e p r im a r y sid e w ith 2
Sch em e 5. Str a tegy for m on oa lk yla tion of th e
u sin g secon d a r y-sid e-p r otected â-cyclod extr in .
p r im a r y sid e of â-cyclod extr in .
Because methoxy groups of 11 cannot be easily con-
verted to hydroxyl groups to give the desired compound,
the protection at the secondary side of â-cyclodextrin with
removable groups was pursued. Available procedures for
the protection of the secondary side of â-cyclodextrin
involve the protection of its primary side first and then
its secondary side with a different group, followed by the
removal of the protecting group at its primary side.20
Because this procedure is tedious, a new approach to
protect its secondary side directly from cyclodextrin using
TBDMS groups was investigated. The precedence for a
direct reaction of the hydroxyl groups at the secondary
side with an electrophile comes from the methylation of
the secondary side of cyclodextrin with NaH in anhydrous
DMF.24 The TBDMS group has an advantage because it
can be removed easily after appropriate modification on
the primary side.
As shown in Scheme 5, compound 7 is stirred with
NaH, and TBDMS chloride is added to the gel that is
formed. Simultaneously, the temperature is raised to 90
°C in the first 2 h and then kept at 90 °C until the
reaction is completed. This addition sequence is used to
give maximal selectivity at the beginning of the reaction
at low temperature, and high temperature later is used
to push the reaction further. TLC of the crude material
12 suggests that it is a mixture of products with varying
degrees of substitution. A strong peak centered at 74.00
ppm and a very weak absorption at 61.8 ppm in its 13C
NMR spectrum (Supporting Information, spectrum 4)
indicate that the silylation reaction has predominantly
taken place on the secondary side. The integration ratio
between signals at 4.82-4.72 ppm (anomeric protons)
and 1.00-0.68 ppm (tert-butyl groups) in its 1H NMR
spectrum (Supporting Information, spectrum 5) indicates
that an average of 6.4 silyl groups are attached to a
cyclodextrin molecule. Attempts to improve this situation
to obtain a homogeneous product by manipulation of
reagent ratios, solvents, and other reaction conditions
failed.
Although this is not an ideal situation, it can be
rationalized that the complete silylation on the secondary
side is unfavorable presumably because the steric hin-
drance of bulky tert-butyldimethylsilyl groups increases
as modification continues. If this assumption is correct,
then it stands to reason that any further electrophilic
reactions on the secondary side would be inhibited and
this would enhance the possibility of our original goal of
selective modification on the primary side. This assump-
tion is supported if 12 reacts with the electrophile 2 to
give exclusively 6-substituted product because, as de-
scribed earlier, native â-cyclodextrin does not do so with
this electrophile. When compound 12 is first reacted with
2 and then deprotected with tetrabutylammonium fluo-
ride (TBAF) without the isolation24 of the intermediate
13, the desired compound 8 is obtained as the major
product (23.8% after appropriate workup of the reaction).
A small amount of unreacted 7, and its higher substituted
derivatives obtained in the crude product are separated
by Sephadex.
(24) In a separate experiment, the intermediate 13 was isolated by
rotary evaporation of the reaction mixture to remove most of the
lutidine, poured into ice water, washed with water, and dried in the
vacuum oven overnight at 90 °C. The 1H and13C NMR of the crude
product indicate that the 4-methylamino-3-nitrobenzyl group is at-
tached to the primary side of the cyclodextrin. Since this is a mixture
of varying degrees of silylation on the secondary side of cyclodextrin
as indicated by TLC, a complete characterization of this product is
not possible. This intermediate, when refluxed with 8 equiv of TBAF
in THF overnight, gives the same final product 8 after appropriate
workup.
Tentative assignment of regiochemistry of the product
is made by an analysis of the 13C spectrum (Supporting
Information, spectrum 6). As described earlier, the six
intense signals in the carbohydrate region at 101.85,
81.44, 73.00, 72.25, 71.99, and 59.83 ppm are assigned
to C1, C4, C3, C5, C2, and C6 of the unmodified glucose
units, respectively. Four less intense peaks at 102.35,

Selectively Monomodified Cyclodextrins
J . Org. Chem., Vol. 65, No. 9, 2000 2627
Sch em e 6. Mon oa lk yla tion a t th e 3-p osition of â-cyclod extr in .
81.88, 71.08, and 68.78 ppm can be attributed to C1′, C4′,
C5′, and C6′ of the modified glucose unit, respectively.
The downfield shifts of C6′ and C4′ by 8.95 and 0.44 ppm,
respectively, and the upfield shift of C5′ by 0.91 ppm
suggests that the modification has taken place at the
primary side of â-cyclodextrin. The complete NMR as-
signments for this compound are reported elsewhere.25
Select ive Mon oa lk yla t ion a t t h e 3-P osit ion of
â-Cyclod extr in . It has been noted that the hydroxyl
groups at the 3-position of cyclodextrin are the least
accessible for electrophilic attack. The strategy that has
been successful in the literature for modification at this
position involves the choice of a reagent that binds the
cyclodextrin and orients its reactive site toward the
hydroxyl group at this position.26 Discussions in the
previous sections suggest that 2 binds to â-cyclodextrin
to orient the reactive benzyl chloride group toward the
secondary side. Usually hydroxyl groups at 2-positions
on the secondary side of cyclodextrin are more reactive
and are modified more easily than those at 3-positions.27-29
However, it is conceivable that the orientation of the
guest may change if the nature of the cavity is altered
by chemical modification of the native cyclodextrin. The
new orientation may direct the reactive site toward the
3-hydroxyl group to alkylate cyclodextrin at that position.
Since the cavity of heptkis(tert-butyldimethylsilyl)-â-
cyclodextrin (14) is known30 to be partly occupied by
TBDMS groups, changing its character and binding
properties, 14 was reacted with the electrophile 2 to
determine the effect of these changes in the reaction
product. We discovered that this reaction leads to the
alkylation of a hydroxyl group at the 3-position. Thus,
3-O-(4-methylamino-3-nitrobenzyl)-â-cyclodextrin (16) was
synthesized by reacting 14 with 2 in 2,6-lutidine, followed
by deprotection with TBAF (Scheme 6), without isolation
of the reaction intermediate. The crude product after
workup and purification by gel filtration followed by
HPLC affords 16 as major product along with a small
amount of 9. The formation of 16 can be explained based
on an intramolecular complex between 14 and 5 as
illustrated in Figure 1. Compound 2 is expected to react
with 2,6-lutedine to form 5 and this is proposed to bind
in a way that makes its reactive carbon atom closer to
the hydroxyl groups at 3-positions and reacts with one
of them to give 16.
F igu r e 1. Proposed complex between 14 and 2 that explains
the formation of 16.
A tentative assingment of the regiochemistry of the
product is made by an analysis of its 13C spectrum
(Supporting Information, spectrum 7). Compound 16
demonstrates a more dispersed carbon spectrum than its
regioisomers 9 and 8. Peaks at 101.64-100.89 and
81.30-80.50 ppm are assigned to C1 and C4 carbons of
unmodified glucose units. Peaks between 73.1 and 71.50
ppm are accredited to C2, C3, and C5, and those at
59.65-59.55 are assigned to C6 of unmodified glucose
units. The signal at 77.98 ppm is assigned to C3′ of the
modified glucose. Both the downfield shift of C3′ and the
upfield shift of C4′ suggest that the modification takes
place on the hydroxyl group at the 3-position. The
complete NMR assignments for this compound are re-
ported elsewhere.25
Mon oa lk yla tion a t th e 2-P osition of â-Cyclod ex-
tr in . The primary principle involved in monoalkylation
at the 2-position of cyclodextrins is that these hydroxyl
groups are the most acidic and can be preferentially
deprotonated by a strong base. The resulting oxyanion
is the most nucleophilic species in the resulting reaction
mixture and would preferentially react with the available
electrophile. This strategy for modification of the second-
ary side of cyclodextrin has been illustrated13 using a
benzyl chloride derivative as an electrophile. We further
illustrate this by using bromomethyl ketone derivative
1′-bromo-4-methylamino-3-nitroacetophenone (1).
Monoalkylation of â-cyclodextrin at the 2-position is
accomplished by the reaction of 7 with 1′-bromo-4-
methylamino-3-nitroacetophenone (18). Compound 18
was synthesized by refluxing 4-methylamino-3-nitroace-
tophenone (17)31 with 1.5 equiv of cupric bromide in a
mixture of CHCl3 and ethyl acetate (Scheme 7). 2-O-[2-
(25) Forgo, P. Ph.D. Dissertation, University of Missouri-St. Louis,
1999.
(26) (a) Fujita, K.; Nagamura, S.; Imoto, T.; Tahara, T. Koga, T. J .
Am. Chem. Soc. 1985, 107, 3233. (b) Fujita, K.; Tahara, T.; Imoto, T.;
Koga, T. J . Am. Chem. Soc. 1986, 108, 2030.
(27) Tekeo, L.; Uemura, K.; Mitoh, H. Carbohydr. Chem. 1988, 7,
29.
(28) Takeo, L.; Mitoh, H.; Uemura, K. Carbohydr. Res. 1989, 187,
20.
(29) Tian, S.; D’Souza, V. T. Tetrahedron Lett. 1994, 50, 9339-9342.
(30) Tabushi, I. Acc. Chem. Res. 1982, 15, 66.
(31) Levine, H. L.; Kaiser, E. T. J . Am. Chem. Soc. 1978, 100, 7670.

2628 J . Org. Chem., Vol. 65, No. 9, 2000
Tian et al.
performed on a Econosphere (C18, 10 µm; length, 250 mm; i.d.,
22 mm) column with acetonitrile in water as solvent. R- and
â-Cyclodextrin were dried overnight in a drying pistol in
refluxing 1-butanol under vacuum. DMF was dried over CaH2
overnight and fractionally distilled before use. 2,6-Lutidine
packaged in a sure/seal bottle was purchased from Aldrich and
used without further purification. Compounds 14,28,32 2,33 10,28
and 1734 were synthesized according to literature procedures.
All new samples were dried at room temperature under high
vacuum for several days before being submitted to elemental
analysis.
Sch em e 7. Mon oa lk yla tion a t th e 2-p osition of
â-cyclod extr in .
All of the UV spectra were taken at 25.0 ( 0.1 °C using
stock solutions. Solutions were prepared by dissolving 8 (5.0
mg), 16 (5.0 mg),and 4-methylamino-3-nitrobenzyl alcohol (1.0
mg) in 50 mL of water, DMF, and DMSO. The final concentra-
tions were 7.7 × 10-5, 7.7 × 10-5, and 1.0 × 10-4 M,
respectively.
6-O-(4-Meth yla m in o-3-n itr oben zyl)-r-cyclod extr in (3).
4-Methylamino-3-nitrobenzyl chloride (2, 0.21 g, 1.04 mmol)
was added as a solid to a solution of R-cyclodextrin (1, 1.00 g,
1.03 mmol) in 2,6-lutidine (80 mL), and this solution was
heated to 150 °C under argon atmosphere for 3.5 h. A major
cyclodextrin-containing compound appeared at Rf ) 0.47 on
TLC. Crude product was concentrated to one-third in vacuo,
precipitated out in acetone (250 mL), filtered, washed with
acetone three times, and dried in a vacuum oven at 45 °C
overnight to yield a yellowish powder (1.15 g).The above crude
product (200 mg) was dissolved in 8 mL of water, loaded on
the Sephadex (G-15) column, and eluted with deionized water
at 15 mL/hour. The first orange band (retention volume 560
mL) was collected, concentrated, and dried under vacuum (0.3
mmHg, room temperature) to give compound 3 (70 mg as
reddish powder, 34.4%). 1H NMR (300 MHz, DMSO-d6):
(Supporting Information, spectrum 9) δ 8.21-8.11 (m, 1H),
7.99 (s, 1H), 7.52 (d, 1H), 6.98 (d, 1H), 5. 58-5.35, (m, 14H),
4.78 (s, 7H), 4.57-4.33 (m. 8H), 3.86-3.20 (m, 36 H), 2.95 (d,
3H). 13C NMR (75 MHz, DMSO-d6): (Supporting Information,
spectrum 10) δ145.31, 136.60, 130.11, 125.06, 124.98, 114.30,
102.21, 101.87, 82.31, 82.02, 73.31, 72.24, 72.03, 71.08, 70.45,
68.56, 59.91, 29.74. Anal. Calcd for C44H68N4O32‚1.5H2O: C,
45.40; H, 6.15; N, 2.41. Found: C, 45.37; H, 6.15; N, 2.52.
Rea ction of 4-Meth yla m in o-3-n itr oben zyl Ch lor id e (2)
w ith 2,6-Lu tid in e. A solution of 4-methylamino-3-nitrobenzyl
chloride (2, 0.21 g, 1.04 mmol) in 2,6-lutidine (8 mL) was
stirred and heated to 120 °C under argon atmosphere over-
night to give a reddish precipitate that contained a compound
appearing at Rf ) 0.05 on TLC (ethyl acetate/hexane, 1:1). The
precipitate was filtered under vacuum and easily dessolved
in water (2 mL), which gave a white precipitate under room
temperature instantly with the solution of silver nitrate in
ethanol.
(4-Methylamino-3-nitrophenyl)-2-oxyethyl]-â-cyclodex-
trin (19) was synthesized by the treatment of â-cyclo-
dextrin (7) with 1.2 equiv of NaH in dry DMF overnight,
followed by the reaction with 1′-bromo-4-methylamino-
3-nitroacetophenone. After a regular workup, the reaction
product was purified by Sephadex followed by HPLC to
yield 19 in 16.5%. Besides regular carbon peaks for
unmodified glucose units of â-cyclodextrin at 101.74,
81.30, 72.98, 72.34, 71.99, and 59.78 ppm, three less
intense peaks at 100.14 (C1′), 82.06 (C4′), and 80.31 (C2′)
ppm of modified glucose units are observed (Supporting
Information, spectrum 8). Downfield shift of C2′ and
upfield shift of C1′ for the modified glucose unit from the
corresponding carbons C2 and C1 of unmodified glucose
units indicate that substitution is on the hydroxyl group
at the 2-position of a â-cyclodextrin molecule.
The UV-vis spectra of representative members of
these derivatives in aqueous and organic solvents did not
exhibit significant differences as shown in spectrum 17
and Table 1 in the Supporting Information. This can be
attributed to lack of intermolecular complex formation
in these compounds at the concentrations employed in
this investigation.
Con clu sion
Subtle differences in the chemistry of the hydroxyl
groups at the 2-, 3-, and 6-positions of cyclodextrins can
be exploited to direct an electrophilic reagent to the
desired site. We have thus synthesized derivatives of
cyclodextrins in which the same substituent is inserted
into each of these positions. A detailed NMR spectral
investigation of these compounds will be the subject of a
future report.
Rea ction of â-Cyclod extr in (7) w ith 4-Meth yla m in o-
3-n itr oben zyl Ch lor id e (2) in P yr id in e. A solution of
4-methylamino-3-nitrobenzyl chloride (2, 0.177 g, 0.881 mmol)
and â-cyclodextrin (7, 1.00 g, 0.881 mmol) in pyridine (80 mL)
was heated to 150 °C under argon atmosphere overnight. TLC
of the reaction mixture indicated that the spot of 2 at Rf >
0.95 disappeared, and a red spot at Rf ) 0.05 appeared. Longer
reaction time (5 h) did give a compound that is not UV-visible
and chars by heat after the treatment of 50% methanolic H2-
SO4.
Exp er im en ta l Section
Rea ction of â-Cyclod extr in (7) w ith 4-Meth yla m in o-
3-n itr oben zyl Ch lor id e (2) in 2,6-Lu tid in e. â-Cyclodextrin
(7, 1.00 g, 0.881 mmol) was added as a solid to a solution of
4-methylamino-3-nitrobenzyl chloride (2, 0.177 g, 0.881 mmol)
in 2,6-lutidine (80 mL), and the above suspended solution was
heated to 150 °C under argon atmosphere overnight. TLC of
the reaction solution gave four orange spots at 0.72, 0.64, 0.49,
and 0.05; the first three of these charred with heat after the
treatment with sulfuric acid, and the last one was not. TLC
of the precipitate at the bottom of the reaction flask gave a
Gen er a l Meth od . Proton and carbon chemical shifts are
reported in ppm (δ) using the residual signal of DMSO-d6 (2.49
ppm for 1H and 39.50 ppm for 13C) or TMS (0.00 ppm) as the
internal standard. TLC was performed on aluminum sheets
precoated with 0.2 mm silica 60 F254. The eluent for TLC of
cyclodextrin derivatives, unless mentioned otherwise, is 1-bu-
tanol/ethanol (95%)/water, 5:4:3 by volume, and the spots were
detected first by UV lamp if the product contained a chro-
mophore and then by charring with heat after spraying the
plate with 50% methanolic sulfuric acid. Charring is the
indication for the presence of cyclodextrin moieties. Gel
filtrations were performed on a Sephadex column (length, 125
cm; i.d., 2.8 cm) using deionized water as eluent. HPLC was
(32) Fu¨gedi, P. Carbohydr. Res. 1989, 192, 366.
(33) Rong, D.; D’Souza, V. T. Tetrahedron Lett. 1990, 31, 4275-4278.
(34) Levine, H. L.; Kaiser, E. T. J . Am. Chem. Soc. 1978, 100, 7670.

Selectively Monomodified Cyclodextrins
J . Org. Chem., Vol. 65, No. 9, 2000 2629
spot that is identical to that of â-cyclodextrin. The reaction
solution was cooled to room temperature, filtered, concentrated
to one-quarter, and poured into acetone (150 mL) to give an
orange precipitate that was filtered, washed with acetone ×
3, and dried under vacuum in an oven at 45 °C for 5 h. 1H
NMR spectrum (300 MHz, DMSO-d6) for the crude compound
was complicated, and the integration ratio of the signals in
the aromatic region to that of anomeric protons was 1:1.
Rea ction of â-Cyclod extr in (7) w ith 4-Meth yla m in o-
3-n itr oben zyl Ch lor id e (2) in DMF a n d 2,6-Lu tid in e. A
solution of 4-methylamino-3-nitrobenzyl chloride (2, 0.177 g,
0.881 mmol) and â-cyclodextrin (7, 1.00 g, 0.881 mmol) in
mixed solvent of DMF (20 mL) and pyridine (20 mL) was
heated to 150 °C under argon atmosphere overnight. TLC of
the reaction mixture showed three spots at 0.64, 0.55, and 0.49,
which were UV-visible and charred with heat after treatment
with 50% sulfuric acid, besides cyclodextrin. The reaction
mixture was concentrated to one-third by rotary evaporation
and poured into acetone (150 mL), and the precipitate was
filtered and dried in a vacuum oven at 45 °C to give 1.17 g of
an orange powder. A 150 mg portion of the solid was dissolved
in 8 mL of distilled water, loaded in a Sephadex column, and
eluted with deionized water at 30 mL/h, and the eluent
between 560 and 610 mL was collected and concentrated to
dryness. TLC of this eluent showed two spots at 0.55 and 0.49.
(75 MHz, DMSO-d6): (Supporting Information, spectrum 6) δ
145.34, 136.61, 130.15, 125.16, 124.96 114.29, 102.35, 101.85,
81.88, 81.44, 73.00, 72.25, 71.99, 71.08, 70.37, 68.78, 59.83,
29.76. Anal. Calcd for C50H78N2O37‚6H2O: C, 42.68; H, 6.45;
N, 1.99. Found: C, 42.72; H, 6.50; N, 1.87.
H ep t a k is(2,3-d i-O-m et h yl)-6-O-(4-m et h yla m in o-3-n i-
tr oben zyl)-â-cyclod extr in (11). The solution of heptakis(2,3-
di-O-methyl)-â-cyclodextrin (10,28 750 mg, 0.563 mmol) and
compound 236 (110 mg, 0.546 mmol) in 2,6-lutidine (30 mL)
was refluxed overnight under argon atmosphere. TLC of this
reaction mixture showed a new spot on TLC (in AcOEt/CH3-
OH, 1:1 by volume) at Rf ) 0.5, which was UV-visible and
charred with heat after the treatment of H2SO4. The reaction
mixture was evaporated under vacuum to dryness to give a
yellowish solid (880 mg). The crude product (200 mg) was
dissolved in deionized water (9 mL), loaded on a Sephadex
(G25-100) column, and eluted with deionized water at 20 mL/
h. The first orange band in the Sephadex column (retention
volume was 540 mL) was collected, concentrated, loaded on
the HPLC column, eluted with 30% acetonitrile in water, and
detected with an UV detector at 254 nm. The peak with 21
min retention time was collected, combined, and further
purified with HPLC under similar conditions to give 11 (64.91
1
mg, 35%). H NMR (300 MHz, DMSO-d6): (Supporting Infor-
mation, spectrum 12) δ 8.17 (q, 1H), 8.00 (d, 1H), 7.50 (dd,
1H), 6.98 (d, 1H), 5.14-5.02 (7H), 4.57-4.43 (m, 6H), 4.43-
4.31 (dd, 2H), 34.00-2.97 (m, 84), 2.94 (d, 3H). 13C NMR
(DMSO-d6): (Supporting Information, spectrum 2) δ 145.28,
136.38, 130.14, 125.03, 124.82, 114.24, 97.66, 97.27, 81.69,
81.62, 81.31, 81.19, 81.12, 79.69, 78.93, 78.79, 78.71, 78.66,
78.63, 78.58, 78.47, 71.81, 70.92, 70.45, 68.64, 60.77, 60.58,
60.55, 60.03, 60.00, 57.90, 59.88, 57.77, 57.51, 57.48, 29.72.
Anal. Calcd for C64H106N2O37‚H2O: C, 50.79; H, 7.19; N, 1.85.
Found: C, 51.03; H, 7.18; N, 1.91.
3-O-(4-Meth ylam in o-3-n itr oben zyl)-â-cyclodextr in (16).
A lutidine solution (25 mL) of 2 (80.0 mg, 0.40 mmol) was
refluxed under argon atmosphere for 4 h when a spot at Rf )
0.80 diminished and new spot at Rf ) 0.05 appeared on TLC
(in hexane and ethyl acetate (1:1 v/v). The above solution was
cooled to 80 °C, compound 14 (1.0 g, 0.52 mmol) was added to
it in one portion, and the above solution was refluxed for 4 h
when the spot at Rf ) 0.05 on TLC disappeared and gave a
new spot at Rf ) 0.45 (ethyl acetate/ethanol/water, 50:1:1).
After the reaction mixture was cooled below 80 °C, an excess
amount of TBAF (2 mL, 1.0 N in THF) was added, and the
mixture was stirred at 80 °C for 3 h. most of the solvent was
removed by rotary evaporation, and the residue was dissolved
into 100 mL of water. The aqueous solution was washed with
ethyl acetate twice, concentrated to dryness, and further dried
under vacuum (0.2 mmHg) for overnight to afford a yellowish
solid (730 mg).
6-O-(4-Meth yla m in o-3-n itr oben zyl)-â-cyclod extr in (8).
â-Cyclodextrin (7, 5.0 g, 4.41 mmol) was stirred with 7.3 equiv
of NaH (1.29 g of 60% dispersion in oil, 31.5 mmol) in dry DMF
(250 mL) under argon atmosphere until a gel was formed (48
h). tert-Butyldimethylsilyl chloride (4.99 g, 33.1 mmol) in dry
DMF (50 mL) was added to the above gel dropwise for 4 h,
during which time the temperature was raised from 25 to 90
°C gradually for the initial 2 h and was kept at 90 °C until
the reaction was completed. The reaction mixture was cooled
to room temperature and approximately two-thirds of the DMF
was removed by rotary evaporator. The residue was poured
into 500 mL of ice water, and the precipitate was filtered,
washed with water, and dried in a vacuum oven at 90 °C
overnight to give 12 (6.7 g, 79%).35 TLC of the crude product,
eluted in a mixture of solvents containing AcOEt, EtOH, and
H2O (7:1:1 by volume), gave six peaks at 0.55, 0.50, 0.33, 0.17,
1
0.07, and 0.00 after the treatment with heat/H2SO4. H NMR
(300 MHz, DMSO-d6): (Supporting Information, spectrum 5)
δ 6.20-5.50 (b, 5H), 4.82-4.72 (s, 7H), 4.60-4.10 (b, 10H),
4.80-3.20 (m, 42H), 1.00-0.68 (2s, 45H), and 0.20-0.03 (d,
18H). 13C NMR (300 MHz, DMSO-d6 ): (Supporting Informa-
tion, spectrum 4) δ 101-102 (b), 83-81(b), 75-74 (b), 73.5-
71.2 (b), 60.2-59.0 (b), 26.1 (s), 25.8 (s), 18.8 (s), and -0.42
(d).
Compound 12 (1 g, 0.52 mmol) was added to a solution of 2
in lutidine and refluxed for 8 h under a dry argon atmosphere,
cooled to room temperature. An excess amount of tetrabutyl-
ammonium fluoride (4 mL, 1 N) was added to the above
reaction mixture, which was allowed to react at 80 °C
overnight, cooled to 0 °C, and maintained at that temperature
for 1 h when the product precipitated out. The product was
filtered, and its TLC showed two spots at Rf ) 0.61 and 0.49
that were UV-visible. After treatment of the TLC plate with
H2SO4/heat, these two spots charred, and a new spot appeared
at Rf ) 0.32 that was consistent with â-cyclodextrin. The crude
product was dissolved in 200 mL of deionized water and
filtered through a cation exchanger (Amberlite CG-120, sodium
form, Sigma company) column (length 25 cm, i.d. 2.8 cm), and
the filtrate was concentrated to 32 mL. One-quarter of this
solution (8 mL) was loaded onto a Sephadex (G25-100) column
and eluted with water at 35 mL/h, and the major orange band
that elutes between 580 and 660 mL was collected and
concentrated to dryness to yield 8 (Rf ) 0.49, 50.3 mg, 23.8%
The crude product (200 mg) was dissolved in water (7 mL),
loaded on a Sephadex (G25-100) column, and eluted with
deionized water at the rate of 10 mL/h. Two orange bands were
observed in the column. The first, which eluted between 500
and 520 mL, contained cyclodextrin as the major component
(Rf ) 0.33), a cyclodextrin-containing compound (UV-visible)
as minor product (Rf ) 0.55), and a cyclodextrin-free compound
(Rf ) 0.8, UV-visible). The second yellow band, which was
collected between 600 and 720 mL, consisted of a major product
at Rf ) 0.45 and a minor one at Rf ) 0.55. The second band
was concentrated to dryness and further dried under vacuum
(0.2 mmHg) for overnight to afford 40 mg. The sample was
further purified on HPLC with 17% aqueous acetonitrile as
the solvent, and the peak with retention times between 9 and
10 min was collected, combined, concentrated, and loaded on
HPLC for further purification under similar conditions to yield
1
16 (28.5 mg, 20.0%). H NMR (500 MHz, D2O): (Supporting
1
from cyclodextrin). H NMR (300 MHz, DMSO-d6): (Support-
Information, spectrum 13) δ 8.32 (d, 1H, J ) 1.8 Hz), 7.85 (dd,
1H, J ) 1.8, 8.8 Hz), 7.21 (d, 1H, J ) 8.8 Hz), 5.30-4.92 (m,
9H), 4.12-3.55 (m, 42 H), 3.23 (s, 3H). DEPT (125 MHz,
DMSO-d6): (Supporting Information, spectrum 7) δ 132.37,
ing Information, spectrum 11) δ 8.14 (m, 1H), 7.99 (d, 1H),
7.51 (dd, 1H), 6,98 (d, 1H), 5.82-5.58 (m, 14H), 4.80 (d, 7H),
4.55-4.30 (m, 8H), 3.80-3.17 (m, 42H), 2.95 (d, 3H). 13C NMR
(35) Calculated assuming all seven hydroxyl groups are substituted.
(36) Rong, D.; D’Souza, V. T. Tetrahedron Lett. 1990, 31, 4275-4278.

2630 J . Org. Chem., Vol. 65, No. 9, 2000
Tian et al.
130.85, 124.46, 101.64, 101.57, 100.89, 81.37, 81.32, 81.26,
81.17, 80.93, 80.72, 77.98, 73.05, 73.01, 72.89, 72.80, 72.75,
72.48, 72.29, 72.19, 72.10, 72.03, 71.96, 71.78, 71.75, 59.65,
59.55, 29.96. 13C NMR (125 MHz, D2O): δ 149.05, 141.59,
132.37, 130.85, 124.46, 116.68, 104.93, 104.45, 104.31, 104.25,
103.85, 84.00, 83.87, 83.81, 83.55, 83.44, 83.37, 78.33, 76.91,
76.23, 75.83, 75.70, 75.39, 75.35, 75.29, 74.62, 74.54, 74.38,
74.30, 74.20, 74.08, 73.99, 62.11, 61.98, 31.50. Anal. Calcd for
C50H78N2O37‚3H2O: C, 44.38; H, 6.26; N, 2.07. Found: C, 44.43;
H, 6.18; N, 2.08.
acetone (100 mL) to precipitate the crude product. The
precipitate was filtered, washed thoroughly with acetone, and
dried under vacuum (0.2 mmHg) for 18 h to afford 640 mg of
orange powder. The solid (180 mg) was dissolved in water (6
mL) and eluted through Sephadex (G25-100) with deionized
water at 15 mL/h. The first yellow band (80 mL from 560 to
640 mL retention volume) was collected, concentrated by rotaty
evaporation, and dried under vacuum (0.2 mmHg) for 10 h to
afford 30 mg (26.3%) of 19. 1H NMR (300 MHz, DMSO-d6):
(Supporting Information, spectrum 15) δ 8.79 (b, 1H), 8.62 (s,
1H), 8.04 (d, 1H, J ) 8.0 Hz), 7.12 (d, 1H, J ) 9.0 Hz), 6.02-
5.59 (b, 14H), 5.34-5.09 (m, 2H), 4.88 (d, 6H, J ) 2.7 Hz),
4.72 (s, 1H), 4.67-4.31 (b, 7H), 3.99-3.23 (m, 76H, some from
water), 3.06 (d, 3H, J ) 4.8 Hz). 13C NMR (75 MHz, DMSO-
d6): (Supporting Information, spectrum 16) δ 192.46, 147.61,
133.66, 130.00, 126.68, 120.70, 114.05, 101.74, 100.14, 82.06,
81.30, 80.31, 73.65, 72.98, 72.34, 71.99, 59.78, 29.70. The
sample was further purified by HPLC with 9% aqueous
acetonitrile, and the band with 15 min retention time was
collected, combined, concentrated to dryness, and further dried
under vacuum to yield 18.8 mg (16.5%). Anal. Calcd for
C51H78N2O38‚6H2O: C, 43.78; H, 6.20; N, 2.00. Found: C, 43.76;
H, 5.90; N, 1.97.
2-O-[2-(4-Met h yla m in o-3-n it r op h en yl)-2-oxy-et h yl]-â-
cyclod extr in (19). 4-Methylamino-3-nitroacetophenone (17,
0.80 g, 4.14 mmol) was dissolved in a mixture of solvents
containing freshly distilled ethyl acetate (70 mL) and chloro-
form (70 mL) and refluxed with cupric bromide (CuBr2, 1.39
g, 6.21 mmol) for 1.5 h. A new spot appeared at Rf ) 0.45 on
TLC (in CH2Cl2) besides starting material at Rf ) 0.33. After
the solution was cooled to room temperature, the solid was
removed by filtration, and the filtrate was concentrated to
dryness by rotary evaporation and purified on chromatotron
(silica gel) with pure methylene chloride as an eluent to yield
1′-bromo-4-methylamino-3-nitroacetophenone (Rf ) 0.45, 0.520
1
g, 46.0%). H NMR (300 MHz, CDCl3): (Supporting Informa-
tion, spectrum 14) δ 8.82 (1H, d, J ) 2.0), 8.48 (1H), 8.09 (1H,
dd, J ) 2.0, 9.0), 6.90 (1H, d, J ) 9.0), 4.36 (2H, s), 3.09 (3H,
d, J ) 4.2). MS (ES, low resolution): calcd for C9H9Br1N4O3
273, found 273.
â-Cyclodextrin (2) (0.50 g, 0.44 mmol) was stirred with NaH
(21.1 mg, 0.530 mmol, 60% in oil) in DMF (30 mL) under dry
argon overnight. The above solution was transferred by
cannula into a solution of 1′-bromo-4-methylamino-3-nitroace-
tophenone (163.8 mg, 0.600 mmol) in dry DMF, and the
mixture was further stirred under dry argon until it became
neutral (3 h). Most of the solvent was then removed under
vacuum, and the residue was poured into the solution of
Ack n ow led gm en t. The authors gratefully acknowl-
edge financial support from the University of Missouri
Research Board and generous donations of cyclodextrins
from Cerestar USA Inc. and Wacker Chemicals (USA),
Inc.
Su p p or t in g In for m a t ion Ava ila b le: 1H and 13C NMR
and UV sprectra discussed in the paper. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O991347R