NMR spectroscopic study of cyclodextrin inclusion complexes with A-007 prodrugs

Article abstract of DOI:10.1016/j.carres.2008.03.014

One- and two-dimensional NMR spectroscopy was used to demonstrate the formation of inclusion cyclodextrin complexes with several A-007 prodrugs. These complexes are comprised from the encapsulation of the two phenol moieties of the A-007 prodrugs within the cyclodextrin cavity. Considering the size of the two phenol moieties of the A-007 prodrugs compared to the sizes of α-, β-, and γ-cyclodextrin cavities, we observed complementary binding of the A-007 prodrug with only β-cyclodextrin, which was also demonstrated spectroscopically. The β-cyclodextrin inclusion complexes increased the prodrug solubility and modified the prodrug half-life in water. Therefore, β-cyclodextrin inclusion complexes can be used as an essential form of A-007 prodrug delivery.

Full text of DOI:10.1016/j.carres.2008.03.014

Available online at www.sciencedirect.com
Carbohydrate Research 343 (2008) 1180–1190
NMR spectroscopic study of cyclodextrin
inclusion complexes with A-007 prodrugs
Sarada Sagiraju and Branko S. Jursic*
Department of Chemistry, University of New Orleans, New Orleans, LA 70148, United States
Received 17 December 2007; received in revised form 6 March 2008; accepted 9 March 2008
Available online 14 March 2008
Abstract— One- and two-dimensional NMR spectroscopy was used to demonstrate the formation of inclusion cyclodextrin
complexes with several A-007 prodrugs. These complexes are comprised from the encapsulation of the two phenol moieties of
the A-007 prodrugs within the cyclodextrin cavity. Considering the size of the two phenol moieties of the A-007 prodrugs compared
to the sizes of a-, b-, and c-cyclodextrin cavities, we observed complementary binding of the A-007 prodrug with only b-cyclo-
dextrin, which was also demonstrated spectroscopically. The b-cyclodextrin inclusion complexes increased the prodrug solubility
and modified the prodrug half-life in water. Therefore, b-cyclodextrin inclusion complexes can be used as an essential form of
A-007 prodrug delivery.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Cyclodextrin; Inclusion complexes; A-007 prodrugs; Molecular aggregation; NOESY
1. Introduction
HO
OH
electron rich (phenol moiety)
4,40-Dihydroxybenzophenone-2,4-dinitrophenylhydraz-
one (A-007, Fig. 1) has recently completed a phase I
clinical trial, where it was used as a therapeutic with
the targeted treatment of advanced cancer, with minimal
toxicity.1 It was speculated that A-007 activity comes
from complementary binding to lymphocyte receptors.
This speculation was supported by experimental findings
that indicate that A-007 interacts, presumably through
the electron-rich (phenol) and the electron-poor (di-
nitrophenyl) moieties, with the CD45 receptor through
complementary interactions.2
N
H
N
A-007
NO2
electron poor (dinitrophenyl moiety)
NO2
1
Figure 1. Chemical structure of the anticancer compound A-007.
water. Two possibilities exist to overcome the water
solubility problem: the first is to make a hydrolyzable
prodrug; and the second is to make an A-007 complex
with a water soluble host, such as cyclodextrin.4 Consid-
ering the complex structure of A-007, we hypothesized
using a combination of these two previously described
methods, which would utilize the chemical transforma-
tion of A-007 into a more water soluble prodrug and
then further increase the water solubility of this newly
formed prodrug through the formation of cyclodextrin
inclusion complexes. Recently, we explored the influence
Despite promising results in the clinical trials, there is
a major disadvantage to using A-007 as a broad scale
therapeutic. A-007 has low water solubility and in its
current formulation, is used only as a topically applied
0.25% gel.3 To make use of this promising anticancer
drug orally or intravenously, the short-term obstacle
must be to overcome the limited solubility of A-007 in
*
Corresponding author. Tel.: +1 504 280 7090; fax: +1 504 280 6860;
e-mail: bjursic1@uno.edu
0008-6215/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2008.03.014

S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
1181
of cyclodextrins on the formation of chiral molecular
associates using amino acid derivatives.5 The applica-
tion of cyclodextrins in drug delivery is well documented
in the literature.6 For example, the toxicity and compati-
bility of using cyclodextrins in drug delivery has been
previously explored and utilized for several commerciali-
zed drugs on the market today.6,7 Here we spectroscopi-
cally explore both the stability of our recently
synthesized A-007 prodrug in aqueous solution, as well
as the ability of our A-007 prodrug to form cyclodextrin
inclusion complexes.
acid salt 4 was prepared in a similar manner as the triflu-
oroacetic acid salt, except that the Boc-deprotection was
carried out with methanolic HCl instead of CF3CO2H–
CH2Cl2.
Glucosylation of A-007 was performed in two steps.
The first step involved Schmidt’s glucosylation10 of A-
007 with 2,3,4,6-tetra-O-acetyl-a-D-glucopyranosyl-tri-
chloroacetimidate,11 followed by sodium hydroxide
hydrolysis in CH2Cl2–MeOH (Scheme 1). The structural
properties of these A-007 prodrugs are ideal for testing
the nature of their biological activity as well as the
administration to patients in combination with
cyclodextrins.12
The simplest possible prodrug of A-007 is its sodium
salt (1Na). This compound will be converted into free
A-007 under physiological conditions. Unfortunately,
even this double sodium salt is not soluble in water in
high concentrations. While at an elevated temperature
(50 °C) a 10 mM water solution can be formed, upon
cooling, a red precipitate forms at room temperature.
It is reasonable to speculate that because A-007 is a sur-
factant by nature, molecular aggregates could form in
aqueous media. This should be reflected in chemical shift
of the aromatic protons of 1Na versus its concentration.
However, our spectroscopic studies revealed no notice-
able difference in chemical shift of 1 mM and 10 mM
concentration of 1Na in water media (Fig. 2), suggesting
that A-007 does not form molecular aggregates, such as
micelles.13 The formation of micelles in aqueous media is
a desirable property for these prodrugs, because micellar
stability can be a function of concentration. Ideally, a
2. Results and discussion
To explore cyclodextrin inclusion complexes with A-007
prodrugs, we prepared neutral, anionic, and cationic
prodrugs (Scheme 1). In all of these preparations, the
starting material used was the previously synthesized
compound, A-007, and was prepared from commercially
available starting materials following our previously
described method.1
Acetic acid derivatives of salt 2 were prepared from
A-007 with ethyl 2-bromoacetate, followed by hydroly-
sis and acidification according to previously reported
methods of preparation for similar phenol derivatives.8
The trifluoroacetic acid salt 3TFA was prepared by the
acylation of A-007 with Boc-glycine [(CH3)3COCO-
NHCH2CO2H] in the presence of dicyclohexylcarbo-
diimide (DCC), followed by trifluoroacetic acid
(CF3CO2H) deprotection in CH2Cl2.9 The hydrochloric
O
O
O
O
R
R
O
O
O
O
RO
OR
N
H
N
N
H
NO2
N
2: R=H
2Na: R=Na
2Et: R=Et
NO2
HO
OH
NO2
3TFA: R=NH3+-O2CCF3
NO2
e
N
H
f
N
A-007 (1)
NO2
3BOC: R=NHCOOC(CH3)3
i
RO
RO
OR
OR
OR
RO
O
O
OR
O
O
NO2
O
O
ClH3N
NH3Cl
RO
O
O
N
H
N
H
4
N
N
5: R=H
5Ac: R=OCCH3
NO2
NO2
NO2
NO2
Scheme 1. Synthetic routes of preparation of A-007 into the prodrug formulation. Reagents and conditions: (a) NaOH/CH3OH; (b) BrCH2-
CO2C2H5; (c) NaOH/H2O–CH3OH; (d) HCl/H2O; (e) (CH3)3COCONHCH2COOH/DCC/CH2Cl2; (f) CF3CO2H/CH2Cl2; (g) 2,3,4,6-tetra-O-
acetyl-a-D-glucopyranosyl trichloroacetimidate/BF3–THF/CH2Cl2; (h) NaOH/CH2Cl2–CH3OH; (i) (COCl)2/CH3OH/CH2Cl2.

1182
S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
Figure 2. 1H NMR spectra of aqueous 1Na.
prodrug should be stable at higher aqueous concentra-
tions, above the critical micellar concentration (CMC)
and it hydrolyzes easily at lower or below CMC concen-
trations. In this way, the prodrug can be stored at higher
concentrations and upon administration it slowly hydro-
lyzes in the blood stream.
This can be explained by the fact that the a-cyclodextrin
cavity is too small to encapsulate the phenol moieties of
1Na, while the c-cyclodextrin cavity is too large to forms
strong inclusion complexes with 1Na. On the other hand
it is well documented that b-cyclodextrin forms a strong
inclusion complex with phenol groups,14 and this obser-
vation corresponds to our spectroscopic data of 1Na
and b-cyclodextrin. We observed a significant chemical
shift in the NMR spectra of free 1Na and 1Na in aque-
ous b-CD (Fig. 3). The major chemical shift changes for
1Na were observed with a 2:1 ratio between b-CD and
1Na. This implies the possibility of a ternary cyclodex-
trin complex formation between one molecule of 1Na
and two molecules of b-CD.24
Because our one-dimensional NMR spectroscopic
study of b-CD complexation with 1Na is not conclusive,
additional spectroscopic investigation is essential. There
are two additional spectroscopic methods that can shed
more light on the nature of the complex, electron spray
mass spectroscopy (ES-MS) and the two-dimensional
nuclear Overhauser spectroscopy (NOESY). Electron
spray mass spectroscopy (ES-MS)15 is an important tool
used for the characterization of non-covalent binding
and using this spectroscopic technique, the ternary
complex signal at 1331.1 m/z was observed in the
b-cyclodextrin and 1Na water solution. This signal
corresponds to the molecular mass of two cyclodextrins
and one 1Na (Fig. 4).
Ideally, compounds to be used as prodrugs should be
water soluble or have the ability to form cyclodextrin
inclusion complexes. To test the capability of 1Na to
be used as a prodrug of A-007, we analyzed the capacity
of 1Na to form a cyclodextrin inclusion complex with a-,
b-, and c-cyclodextrin. We prepared a 10 mM solution
of 1Na with b-cyclodextrin, and observed that the solu-
tion stayed clear for several days, a dramatic improve-
ment considering that the precipitate forms in
approximately 30 min from 10 mM 1Na without b-
cyclodextrin. (A 10 mM solution was made by mixing
together A-007 with sodium hydroxide in D2O and son-
icated at 50 °C for 5 min. A solution that is 0.1 mM is
the highest water concentration of 1Na that remains
stable for several days.)
To determine whether a strong cyclodextrin inclusion
complex is formed, it is necessary to observe changes in
the spectroscopic pattern of the compound/complex,
where these changes should be reflected by changes in
the NMR chemical shifts. In our analyses, we found
no noticeable change in chemical shifts of aromatic pro-
tons of 1Na in combination with a- or c-cyclodextrins.
Figure 3. 1H NMR spectra of aqueous b-cyclodextrin of 1Na (10 mM).

S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
1183
summary, all spectroscopic studies, including the 1D
NMR, 2D NOESY, and ES-MS, indicate that there is
the formation of a ternary inclusion complex that sub-
stantially increases the water solubility of the simple
A-007 prodrug, 1Na.
Our second compound, the sodium salt 2Na, does not
strictly belong to the group of A-007 prodrugs because it
cannot be directly converted into A-007 itself. Neverthe-
less, we chose to include this compound in our study due
to the fact that phenol ethers metabolize via oxidative
O-dealkylation, resulting in the formation of a free
phenol group, or in our case to A-007.23 In addition,
the corresponding acid 2 also has anticancer activity that
is similar to the original A-007. Unfortunately, the low
solubility of this acid also hampers its oral or intravenous
administration. Using carboxylic acid salts as a polar
group of 2Na, we expected this compound to possess
physical properties similar to surfactants, lending it
capable of forming molecular assemblies similar to taxol
prodrugs reported by Nicolaou and coworkers.17 First,
we explored the change of the NMR chemical shift of
2Na versus its concentration in pure water. There was
a substantial downfield chemical shift observed in the
spectra when the concentration of 2Na increased from
1 mM to 10 mM, while the saturation was achieved after
the 10 mM concentration (after CMC).
Figure 4. Negative ES-MS spectra of 1Na (0.01 mM) and b-cyclo-
dextrin (0.03 mM).
Two-dimensional NOESY (nuclear Overhauser spec-
troscopy) can be used to demonstrate that two protons
or groups of protons are in proximity, as the protons
must be within 3.5 A of each other.16 This spectroscopic
˚
method is a valuable tool to study b-cyclodextrin inclu-
sion complexes with A-007 prodrugs. Using NOESY, we
found that the 2D NOESY spectra of our b-cyclodextrin
complex with 1Na showed intense cross peaks indicating
interactions between b-cyclodextrin and phenol moiety
of 1Na (Fig. 5). These two cross peaks indicate the pres-
˚
ence of these groups in the vicinity of 2–5 A. However
there are no crossing signals between b-cyclodextrin
hydrogens and the hydrogens of the 2,4-dinitrophenyl
moiety of 1Na, suggesting that this part of the 1Na mole-
cule is not encapsulated in the b-cyclodextrin cavity. In
In the aqueous cyclodextrin solution of 2Na we
expected to observe the appropriate NMR chemical
shift changes when the cyclodextrin inclusion complex
Figure 5. 2D NOESY spectra of 1Na (10 mM) in aqueous b-cyclodextrin (30 mM).

1184
S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
is formed. There were no noticeable changes when either
a-cyclodextrin or c-cyclodextrin was added into aqueous
2Na (10 mM). This observation did not come as a sur-
prise as the a-cyclodextrin cavity is too small (4.7–
Two-dimensional NOESY spectra clearly support our
assumption that the ternary b-cyclodextrin inclusion
complex with 2Na is formed (Fig. 7). There are several
NOE cross couplings between the two phenol moieties
of prodrug 2Na and the cyclodextrin hydrogens. How-
ever, we observed no cross coupling between hydrogens
from the 2,4-dinitrophenyl moiety of 2Na and the
b-cyclodextrin hydrogens. Furthermore, cross coupling
between the methylene hydrogens of the acetic acid
moieties and b-cylcodextrin hydrogens was observed as
well. All of these cross couplings strongly suggest that
the formation of a ternary b-cyclodextrin inclusion com-
plex with 2Na is formed in water media as demonstrated
in Figure 7.
˚
˚
5.3 A) while c-cyclodextrin cavity is too big (7.5–8.3 A)
to form strong inclusion complexes with 2Na.18 How-
ever, there was a substantial difference with b-cyclodex-
trin (Fig. 6). This difference is profound on the chemical
shift of the phenol moiety of the prodrug, indicating that
this part of the molecule interacts with b-cyclodextrin.
As was the case with 1Na, the b-cyclodextrin titration
of 10 mM solution of 2Na suggests that a ternary (one
molecule of 2Na and two molecules of b-cyclodextrin)
complex is formed. This was based on the fact that after
a ratio of 1:2 was reached, further increases of b-cyclo-
dextrin did not produce noticeable chemical shift
changes in the NMR spectra of 2Na.
Both of the previously described prodrugs (1Na and
2Na) are sodium salts and therefore are negatively
charged organic molecules that revert to the corresponding
Figure 6. 1H NMR spectra of 2Na (10 mM) in aqueous b-cyclodextrin.
Figure 7. 2D NOESY of 2Na (10 mM) in aqueous b-cyclodextrin (30 mM).

S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
1185
target low water soluble drugs upon acidification
(1 and 2, respectively). At physiological pH (7.4) both
of these prodrugs must be used in combination with
b-cyclodextrin. To ensure a proper drug delivery using
prodrugs, it is of crucial importance to explore ‘neutral’
(less pH sensitive) A-007 prodrugs. A second very
important point in the design of the A-007 prodrug is
its timely release to the targeted area. Therefore, the
half-life of drug released from the prodrug is an impor-
tant factor in drug design.19 The ideal prodrug is the
one that is instantly transformed into the target drug at
the site in which biological activity is desired.
Glucosides of active biological compounds are very
often used in ADEPT (antibody-directed enzyme pro-
drug therapy) strategy.20 This strategy provides the
means to selectively deliver an active drug to its intended
target from an inactive prodrug precursor. In the first
step, the enzyme (b-glucuronidase) is delivered to the
tumor site by monoclonal antibody–antigen recognition,
and in the second step, enzyme cleavage of the prodrug
liberates the active drug to the tumor cell surface. b-Glu-
curonidase is an appropriate enzyme for glycosylated
anticancer drugs. In normal tissue, glucosidase is located
in the liposome of the cell and it is found in very low
concentrations in the blood. High concentrations of this
enzyme are present extracellularly in necrotic areas.
Considering this, our glucoside 5 should be an ideal
prodrug for A-007. After delivery of the A-007 deriva-
tive, glucoside 5, to the vicinity of the cancer cell, the
free drug, A-007, should be delivered by enzymatic
glycoside bond cleavage. To further support this
hypothesis, it was also determined in vitro that prodrug
5 has slightly higher anticancer activity than the original
A-007 (the results of this study will be published
elsewhere). However, the solubility of 5 is still relatively
low (ꢀ2 mg/1 mL of water). The solubility of this com-
pound is fivefold higher in the presence of b-cyclodextrin
(from ꢀ2 mg/1mL of water to 10 mg/1 mL of 30 mM b-
CD). Even so, we could not observe substantial change
of aromatic NMR chemical shift of 5 with any of the
three cyclodextrins. However, in the ES-MS spectra of
5 in aqueous b-cyclodextrin, there is a weak signal that
corresponds to the complex between one b-CD and
one molecule of 5 (Fig. 8). The only plausible explana-
tion is that inter-molecular interactions between 5 and
b-cyclodextrin occurred between glucose moieties of 5
and the outside wall of b-cyclodextrin. That will increase
the solubility of 5 in b-CD, show the molecular associate
signal in ES-MS, but not have influence in the chemical
shift of aromatic signals that are relatively far away
from the molecular interactions.
To explore the influence of cationic A-007 prodrugs
on delivery (half-life), solubility, and anticancer activ-
ity, we prepared and studied the physical properties
of glycine derived prodrugs 3TFA and 4. These com-
pounds are hydrolyzable prodrugs sensitive to the
moisture and their stability in aqueous media might
be questionable, based on the shelf-life of the com-
pounds.21 Considering that these compounds have sur-
factant-like properties, it is reasonable to expect that
they might be able to form molecular aggregates simi-
lar to micelles in aqueous media. Aggregation was
nicely demonstrated by following the NMR chemical
shift for aromatic protons versus concentration of 4
in water media (Fig. 9). We observed a substantial
downfield chemical shift with increasing concentration,
with the maximal effect observed at 10 mM. At this
concentration, aqueous 4 is stable at room tempera-
ture for long periods (at least three months). We as-
sume that at this concentration, 4 forms molecular
Figure 8. Negative ES-MS of diglucoside 5 (0.01 mM) and aqueous b-CD (0.03 mM).

1186
S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
Figure 9. Dependence of concentration of 4 on the chemical shift.
aggregates similar to micelles (Fig. 9), which slow
down the hydrolysis of the ester group.
mental evidence indicates that a-cyclodextrin forms
complexes with trifluoroacetic acid.22 If this is true, then
the trifluoroacetate anion of 3 should be complexed by
a-cyclodextrin, resulting in a slight downfield chemical
shift for all aromatic protons. We did not observe any
evidence that supports the formation of the a-cyclodex-
trin inclusion complex with the aromatic moieties of
3TFA (Fig. 11).
On the other hand, there is a noticeable difference in
the chemical shift of the 3TFA protons in aqueous
b-cyclodextrin, indicating strong interactions between
b-cyclodextrin and 3TFA. Unfortunately, we were not
able to record a reliable two-dimensional NOESY
NMR spectra due to fact that 3TFA hydrolyzed in water
media. In the NMR sample there are three components
(3TFA, monoester of A-007, and A-007) in complex
with b-cyclodextrin. The molar ratio of these three
components varies with the progression of the NMR
recording time. This is clearly demonstrated on the
one-dimensional spectrum of 3TFA in water and b-
cyclodextrin (Fig. 11). Even for the freshly prepared
sample of 3TFA in water, the NMR spectra (acquired
after 10 min) shows the presence of partially hydrolyzed
3TFA. Full hydrolysis of 3TFA generates A-007 that
immediately precipitates from the solution. Aqueous
To confirm this assumption, we also explored the
water stability of 4 at 1 mM concentration (Fig. 10).
Almost instantly, hydrolysis of the two ester groups of
4 started and the precipitation of A-007 could be
observed. Due to the exceptionally low water solubility
of A-007, we were not able to obtain the NMR spectra
of pure A-007 in water, and the additional signals in the
NMR spectra of 4 after one and two days correspond to
the mono ester of A-007 (partially hydrolyzed product
of 4). Neither the NMR spectra of 4 after one day nor
the NMR spectra of 4 in water after two days can be
used for computing the half-life of this drug because
both of the NMR spectra contain substantial amounts
of precipitate and their NMR acquisition time is pro-
gressively longer. The half-life was determined by evalu-
ating the amount of precipitate versus time. It was
determined that the half-life of 4 as a 1 mM concentra-
tion at room temperature was around 140 min. As men-
tioned above, at a 10 mM concentration the formation
of the precipitate was not observed even after a few
months at 5 °C.
There are some interesting observations regarding
cyclodextrin inclusion complexes with 3TFA. Experi-
Figure 10. 1H NMR spectra of 4 at 1 mM at different time points.

S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
1187
Figure 11. 1H NMR spectra of 3TFA (10 mM) in a-CD (100 mM) and b-CD (100 mM).
solution of 3TFA and b-cyclodextrin (with 1:10 molar
ratio) stays clear for a few hours, but the NMR spec-
trum indicates the presence of all three molecular
species.
Regardless of the fact that the characterization of the
b-cyclodextrin inclusion complex with 3TFA is hard to
evaluate by two-dimensional NMR spectroscopy, we
believe that the 1H NMR spectra supports the formation
of the b-cyclodextrin inclusion complex. Furthermore,
this prodrug half-life (ꢀ80 min in aqueous b-cyclodex-
trin) seems to be ideal for pharmacological study of this
prodrug.
The 1H and 13C NMR spectra were run on Varian
300 MHz Gemini2000 and on Varian 400 MHz Unity
in CDCl3, DMSO-d6, or D2O as solvent and internal
standards. Two-dimensional NOESY spectra were
recorded on Varian INOVA 500 MHz spectrophotometer
with D2O as a solvent and internal standard (4.80 ppm).
The mass spectra were recorded on a Micromass Quat-
tro 2 Triple Quadropole Mass Spectrometer. The pre-
pared ammonium A-007 prodrugs are hydroscopic and
decompose by ester hydrolysis. However under dry con-
ditions they can be stored for at least several months. All
compounds have melting points higher than 200 °C and
decompose before melting. For ammonium compounds
3TFA and 4 carbon NMR spectra were not reported
because they decompose in solution during NMR acqui-
sition time. For these compounds 1H NMR samples
were prepared immediately before recording the
spectrum.
3. Conclusion
Several interesting A-007 prodrugs were prepared
through ester, ether, and acetal linkages of polar mole-
cules to A-007. All the prodrugs have significantly higher
water solubility (10–50 mM) as compared to original
A-007. To further increase their water solubility, alter
their water half-life, and biodelivery, cyclodextrin inclu-
sion complexes were studied. Through one-dimensional
NMR, two-dimensional NOESY, and negative ES-MS
spectroscopy studies, it was demonstrated that inclusion
complexes with b-cyclodextrin were formed for all but
A-007 glucoside 5. It seems that the a-cyclodextrin
cavity is too small and the c-cyclodextrin cavity is too
large to form stable inclusion complexes with these
studied A-007 prodrugs. It was postulated that b-cyclo-
dextrin forms a 2:1 inclusion complex with the studied
A-007 prodrugs. None of the three studied cyclodextrins
form inclusion complexes with the diglucoside of A-007,
although they slightly increase the prodrug water solu-
bility. This prodrug was prepared for targeted cancer
therapy (ADEPT).
4.1. Preparation of ethyl {4-[[(2,4-dinitrophenyl)hydra-
zono]-(4-ethoxycarbonyl-methoxyphenyl)-methyl]-phen-
oxy}-ethanoate (2Et)
An acetone (500 mL) suspension of A-007 (3.94 g,
0.01 mol), potassium carbonate monohydrate (3.12 g,
0.01 mol), and ethyl 2-bromoacetate (3.34 g, 0.02 mol)
was sonicated at room temperature for 2 h and heated
at reflux overnight. The solvent from dark red suspen-
sion was evaporated to the solid residue. The solid resi-
due was mixed with CH2Cl2 (200 mL), stirred at room
temperature for 2 h, and the insoluble material was sep-
arated by filtration. The filtrate was washed with 10%
potassium carbonate, dried over anhydrous magnesium
sulfate and evaporated to deep red solid residue (5.3 g,
94%). According to the NMR spectroscopy product
was more than 96% pure and was used in next step with-
1
out further purification. H NMR (300 MHz, CDCl3) d
11.30 (1H, NH), 9.09 (1H, d, J = 2.4 Hz), 8.35 (1H, dd,
J1 = 9.3, J2 = 2.4 Hz), 8.18 (1H, d, J = 9.3 Hz), 7,64
(2H, d, J = 9 Hz), 7.32 (2H, d, J = 8.4 Hz), 7.30 (2H,
d, J = 8.7 Hz); 7.17 (2H, d, J = 8.4 Hz), 6.94 (2H, d,
J = 9.0 Hz), 4.76 (2H, s), 4.69 (2H, s), 4.37 (2H, q,
4. Experimental
Melting points were taken on an Electrothermal IA 9000
Digital Melting Point Apparatus and are uncorrected.

1188
S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
J = 6.9 Hz), 4.31 (2H, q, J = 6.9 Hz), 1.36 (3H, t,
J = 6.9 Hz), 1.34 (3H, t, J = 6.9 Hz); 13C NMR
(300 MHz, CDCl3) d 168.7, 159.8, 159.4, 155.2, 144.7,
138.0, 130.6, 130.1, 129.9, 129.6, 124.9, 123.7, 116.8,
116.3, 114.8, 65.6, 65.5, 14.4. Anal. Calcd for
C27H26N4O10 (MW 566.52): C, 57.24; H, 4.63; N, 9.89.
Found: C, 57.08; H 4.77, N 9.72.
1.48 (9H, s), and 1.47 (9H, s); 13C NMR (CDCl3,
400 MHz) 169.1, 168.9, 155.9, 153.6, 152.2, 152.0,
144.5, 136.4, 134.3, 130.1, 129.9, 129.8, 129.3, 123.5,
121.8, 116.7, 80.5, 42.8, 28.4. Anal. Calcd for
C33H36N6O12 (MW 708.67): C, 55.93; H, 5.12; N,
11.86. Found: C, 55.73; H, 5.12; N, 11.77.
4.4. Preparation of 3,30-(4,40-((2-(2,4-dinitrophenyl)-
hydrazono)methylene)bis(4,1-phenylene)bis(oxy))bis-
(2-oxoethylammonium trifluoroacetate) (3TFA)
4.2. Preparation of (4-{(4-carboxymethoxy-phenyl)-[(2,4-
dinitro-phenyl)-hydrazono]-methyl}-phenoxy)-acetic acid
(2)
A mixture of CH2Cl2–trifluoroacetic acid (9 mL:1 mL)
wa cooled to 0 °C and 3BOC was added (354 mg,
0.5 mmol). The reaction mixture was stirred at 0 °C
for 1 h and solvent was evaporated under a stream of
nitrogen resulting in red crystalline product (361 mg;
Water (150 mL), sodium hydroxide (1.6 g, 0.04 mol),
and MeOH (150 mL) were mixed with CH2Cl2
(150 mL) solution of crude A-007 ester (5.3 g,
9.3 mmol). This solution was stirred at room tempera-
ture for 4 h. The reaction mixture was concentrated to
1/3 of its original volume at 40 °C and reduced pressure,
diluted with water (100 mL), and neutralized with con-
centrated hydrochloric acid (10 mL). The solid material
was separated by filtration, washed with water
(3 ꢁ 100 mL), and dried at 110 °C for a few hours to
afford a product that was pure by NMR spectroscopy
1
98%). H NMR (DMSO-d6, 400 MHz) d 11.04 (1H, s),
8.83 (1H, d, J = 2.4 Hz), 8.48 (3H, m), 8.83 (3H, m),
8.26 (2H, d, J = 9.3 Hz), 7.78 (2H, d, J = 8.8 Hz), 7.66
(2H, d, J = 8.4 Hz), 7.55 (2H, d, J = 8.4 Hz), 7.32 (2H,
d, J = 8.8 Hz), 4.21 (2H, m), 4.16 (2H, m). Anal. Calcd
for C27H22F6N6O12ꢂ2H2O (MW 772.52): C, 41.98; H,
3.39; N, 10.88. Found: C, 42.02; H, 3.37; N, 10.75.
1
(4.5 g, 95%). H NMR (DMSO-d6, 400 MHz) d 11.12
(1H, s), 8.79 (1H, d, J = 2.4 Hz), 8.34 (1H, dd,
J1 = 9.6 Hz, J2 = 2.4 Hz), 8.17 (1H, d, J = 9.6 Hz),
7.54 (2H, d, J = 8.8 Hz), 7.38 (2H, d, J = 8.4 Hz), 7.20
(2H, d, J = 8.4 Hz), 6.99 (2H, d, J = 8.4 Hz), 4.81 (2H,
s), 4.73 (2H, s); 13C NMR (DMSO-d6, 400 MHz) d
169.93, 169.90, 159.6, 159.0, 154.6, 144.1, 137.1, 130.1,
130.0, 129.6, 129.3, 123.8, 122.9, 116.6, 115.8, 114.7,
64.7, 64.6. Anal. Calcd for C23H18N4O10 (MW
510.41): C, 54.12; H, 3.55; N, 10.98. Found: C, 53.98;
H, 3.63; N, 10.87.
4.5. Preparation of 3,30-(4,40-((2-(2,4-dinitrophenyl)-
hydrazono)methylene)bis(4,1-phenylene)bis(oxy))bis-
(2-oxoethylammonium chloride) (4)
A
CH2Cl2 (0.2 mL) solution of 3BOC (45 mg;
0.06 mmol) was injected into an ice cold MeOH
(2 mL) solution of hydrochloric acid (made by the
addition of 0.2 mL of oxalyl chloride). A red hydro-
scopic precipitate was immediately formed. The product
was separated by filtration, washed with CH2Cl2
(3 ꢁ 0.5 mL), and dried under nitrogen to produce a
product that was pure by NMR spectroscopy (33 mg,
4.3. Preparation of tert-butoxycarbonylamino-acetic acid
4-{[4-(2-tert-butoxycarbonylamino-acetoxy)-phenyl]-
[(2,4-dinitro-phenyl)-hydrazono]-methyl}-phenyl ester
(3BOC)
1
95%). H NMR (DMSO-d6, 400 MHz) d 11.04 (1H, s),
8.33 (1H, d, J = 2 Hz), 8.45 (1H, dd, J1 = 9.6 Hz,
J2 = 2.0 Hz), 8.26 (1H, d, J = 9.6 Hz), 7.77 (2H, d,
J = 8.4 Hz), 7.64 (2H, d, J = 8.0 Hz), 7.56 (2H, d,
J = 8.4 Hz), 7.32 (2H, d, J = 8.4 Hz), 4.18 (2H, s), 4.13
(2H, s). Negative ES-MS m/z 507.4 (Mꢃ2HClꢃH+),
450.3 (Mꢃ3HClꢃCOCH2NH2þ), 393.4 (Mꢃ2HClꢃ
2COCH2NH2ꢃH+); Anal. Calcd for C23H22Cl2N6O8ꢂ
2H2O (MW 617.39): C, 44.74; H, 4.24; N, 13.61. Found:
C, 44.56; H, 4.31; N, 13.45.
tert-Butoxycarbonylamino-acetic acid (175 mg, 1 mmol),
A007 (197 mg, 0.5 mmol), and dicyclohexylcarbodiimide
(248 mg, 1.2 mmol) were taken in 20 mL of dry CH2Cl2.
The reaction mixture was stirred at 0 °C for 1 h and then
at room temperature overnight. The progress of the reac-
tion was monitored by proton NMR spectroscopy. The
reaction mixture was concentrated to half of the original
volume and filtered. The white residue was discarded and
the filtrate was further concentrated and subjected to col-
umn chromatography on a silica gel column with CH2Cl2
and EtOAc (10:1) producing pure product (71 mg; 20%).
1H NMR (CDCl3, 400 MHz) 11.20 (1H, s), 9.01 (1H, d,
J = 2.8 Hz), 8.33 (1H, dd, J1 = 9.6 Hz, J2 = 2.0 Hz),
8.14 (2H, d, J = 9.6 Hz), 7.43 (2H, d, J = 8.8 Hz), 7.38
(2H, d, J = 8.4 Hz), 7.15 (2H, d, J = 8.8 Hz), 5.28 (2H,
m), 4.22 (2H, d, J = 4.8 Hz), 4.18 (2H, d, J = 4.4 Hz),
4.6. Preparation of 4,40-dihydroxybenzophenone-2,4-di-
nitrophenylhydrazone bis(2,3,4,6-tetra-O-acetyl-b-D-
glucopyranoside) (5Ac)
2,3,4,6-Tetra-O-acetyl-a-D-glucopyranosyl-trichloro-
acetimidate (370 mg, 0.75 mmol) and A-007 (118 mg,
0.3 mmol) were stirred under an atmosphere of nitrogen
˚
in anhydrous CH2Cl2 (20 mL) with 3 A molecular sieves
for 1 h. The solution was cooled in an ice-bath for

S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
1189
2. Morgan, L. R.; Jursic, B. S.; Hooper, C. L.; Neumann, D.
M.; Thangaraj, K.; LeBlanc, B. Bioorg. Med. Chem. Lett.
2002, 12, 3407–3411.
3. Eilender, D.; LoRusso, P.; Thomas, L.; McCormick, C.;
Rodgers, A. H.; Hooper, C. L.; Tornyos, K.; Krementz, E.
T.; Parker, S.; Morgan, L. R. Cancer Chemother. Phar-
macol. 2006, 57, 719–726.
30 min before BF3ꢂOEt2 (2 mL) was added and stirred
for 1 min the ice-bath followed by 1 h at room tempera-
ture. The solution was added to an ice cold satd aq
NaHCO3 solution (100 mL) with vigorous stirring and
then extracted with ether (2 ꢁ 75 mL). The organic lay-
ers were combined and dried over anhydrous sodium
sulfate. The solvent was removed under reduced pres-
sure and the residue was purified by flash column chro-
matography (CH2Cl2–CH3CO2Et 10:1) to give 5Ac as
4. Anderson, B. D. Adv. Drug Delivery Rev. 1996, 19, 171–
202.
5. For instance see: (a) Jursic, B. S.; Patel, P. K. Carbohydr.
Res. 2006, 341, 2858–2866; (b) Jursic, B. S.; Patel, P. K.
Tetrahedron 2005, 61, 919–926; (c) Jursic, B. S.; Patel, P.
K. Carbohydr. Res. 2005, 304, 1413–1418; (d) Kobetic, R.;
Jursic, B. S.; Bonnette, S.; Salamone, S. J. Tetrahedron
Lett. 2001, 42, 6077–6082.
6. For instance see: (a) Challa, R.; Ahuja, A.; Ali, J.; Khar,
R. K. AAPS PharmSciTech 2005, 6, E329–E357; (b)
Holvoet, C.; Vander Heyden, Y.; Plaizier-Vercammen, J.
Pharmazie 2007, 62, 510–514; (c) Loftsson, T.; Duchene,
D. Int. J. Pharm. 2007, 329, 1–11; (d) Uekama, K.;
Hirayama, F.; Irie, T. Chem. Rev. 1998, 98, 2045–2076; (e)
Szejtli, J. Med. Res. Rev. 1994, 14, 353–386.
7. (a) For general information about cyclodextrins see:
Comprehensive supramolecular Chemistry; Szejtli, J., Osa,
T., Eds.; Pergamon: Oxford, UK, 1996; Vol. 3; (b) Szejtli,
J. Chem. Rev. 1998, 98, 1743–1754.
8. Lee, K.; Lee, J. H.; Boovanahalli, S. K.; Jin, Y.; Lee, M.;
Jin, X.; Kim, J. H.; Hong, Y.-S.; Lee, J. J. J. Med. Chem.
2007, 50, 1675–1684.
1
red crystalline solid (43 mg; 20%). H NMR (CDCl3,
400 MHz) d 11.28 (1H, s), 9.09 (1H, d, J = 2.8 Hz),
8.37 (1H, dd, J1 = 2.8 Hz, J2 = 9.6 Hz), 8.17 (1H, d,
J = 9.2 Hz), 7.61 (2H, d, J = 8.8 Hz), 7.31 (2H, d,
J = 8.8 Hz), 7.25 (2H, d, J = 8.8 Hz), 5.40–5.10 (6H,
m), 4.30 (2H, m), 4.20 (2H, t), 3.9 (2H, m), 2.13 (3H,
s), 2.09 (3H, s), 2.08 (3H, s), 2.075 (6H, s), 2,071 (3H,
s), and 2.065 (3H, s). Negative ES-MS 1011.8 [MꢃH]ꢃ;
positive ES-MS 1035.8 [M+Na]+. Anal. Calcd for
C45H48N2O23 (MW 1012.88): C, 53.36; H, 4.78; N,
5.53. Found: C, 53.25; H, 4.88; N, 5.42
4.7. Preparation of 4,40-dihydroxybenzophenone-2,4-di-
nitrophenylhydrazone bis(b-D-glucopyranoside) (5)
9. Procedure was adapted from HCl/Pd–C deprotection of
N-(benzyloxycarbonyl)glycine: Karrer, P.; Heynemann,
H. Helv. Chim. Acta 1948, 31, 398–404.
10. Schmidt, R. R.; Kinzy, W. Adv. Carbohydr. Chem.
Biochem. 1994, 50, 21–123.
Octaacetate 5Ac (101 mg, 0.1 mol) was dissolved in 2:1
MeOH–CH2Cl2 (30 mL). Sodium hydroxide (1 M) was
added to a pH ꢀ 9–10. The solution was stirred at room
temperature overnight, neutralized with acidic Dowex
resin, and evaporated. The solid residue was dried under
reduced pressure at room temperature to give 68 mg
11. Schmidt, R. R.; Michel, J.; Roos, M. Liebigs Ann. Chem.
1984, 1343–1357.
1
12. For recent review on cyclodextrins as drug delivery see: (a)
Challa, R.; Ahuja, A.; Ali, J.; Khar, R. AAPS PharmSci-
Tech 2005, 6, E329–E357; (b) Uekama, K.; Hirayama, F.;
Arima, H. J. Inclusion Phenom. Macro. 2006, 56, 3–8.
13. For definition and determination of micellar characteristic
of organic molecules see: Furton, K. G.; Norelus, A. J.
Chem. Educ. 1993, 70, 254–257.
14. For order of the cyclodextrins partition coefficient
between solution phase and phenyl-sepharose CL-48 see:
(a) Janado, M.; Yano, Y.; Umura, M.; Kondo, Y. J. Sol.
Chem. 1995, 24, 587–600; for general review of cyclo-
dextrins application see: (b) Saenger, W. Angew. Chem.,
Int. Ed. 2003, 19, 344–362.
15. (a) Yang, C.; Tarr, M. A.; Xu, G.; Yalcin, T.; Cole, R. B.
J. Am. Soc. Mass. Spectrom. 2003, 14, 449–459 and
references cited therein; (b) Kadri, M.; Djemil, R.;
Abdaoui, M.; Winum, J.-Y.; Coutrot, F.; Montero, J. L.
Biorg. Med. Chem. Lett. 2005, 15, 889–894; (c) Kobetic,
R.; Jursic, B. S.; Bonnette, S.; Tsai, J. S.-C.; Salvatore, S.
J. Tetrahedron Lett. 2001, 42, 6077–6082.
(95%) of red crystalline product. H NMR (DMSO-d6,
400 MHz) d 11.13 (1H, s), 8.80 (2H, d, J = 2.4 Hz),
8.40 (2H, dd, J1 = 9.6 Hz, J2 = 2.4 Hz), 8.19 (2H, d,
J = 9.6 Hz), 5.54 (1H, d, J = 4.8 Hz), 5.48 (1H, d,
J = 4.8 Hz), 5.18 (4H, m), 4.99 (1H, d, J = 7.2 Hz),
4.92 (1H, D, J = 7.4 Hz), 4.74 (1H, t, J = 5.6 Hz), 4.67
(1H, t, J = 5.6 Hz), 3.2 (4H, m); 13C NMR (DMSO-
d6, 400 MHz) d 159.4, 159.0, 154.9, 144.5, 137.6, 130.6,
130.3, 129.7, 129.6, 125.0, 123.4, 117.8, 117.0, 116.6,
100.6, 100.3, 77.3, 76.8, 73.6, 73.5, 70.0, 61.0. Positive
ES-MS 731.5 [M+Na]+; Negative ES-MS 717.4 [MꢃH].
Anal. Calcd for C31H34N4O16 (MW 718.62) C, 51.81;
H, 4.77; N, 7.80. Found: C, 51.65; H, 4.88; N, 7.68.
Acknowledgment
16. For a few examples of cyclodextrins inclusion complexes
and 2D NOESY see: (a) Liu, Y.; Chen, G.-S.; Li, L.;
Zhang, H.-Y.; Cao, D.-X.; Yuan, Y.-J. J. Med. Chem.
2003, 4634–4637; (b) Aachmann, F. L.; Otzen, D. E.;
Larsen, K. L.; Wimmer, R. Protein Eng. 2003, 16, 905–
912; (c) Torne, S. J.; Torne, J. S.; Vavia, P. R.; Singh, S.
K.; Kishore, N. J. Inclusion Phenom. Macrocycl. Chem.
2007, 57, 689–697.
We thank the National Science Foundation for financial
support (CHE-0611902) for this work.
References
17. Nicolaou, K. C.; Guy, R. K.; Pitsinos, N. E.; Warasidlo,
W. Angew. Chem., Int. Ed. 2003, 33, 1583–1587.
18. Szejtli, J. Chem. Rev. 1998, 98, 1743–1753.
1. Morgan, L. R.; Thangaraj, K.; LeBlanc, B.; Rodgers, A.;
Wolford, L. T.; Hooper, C. L.; Fan, D.; Jursic, B. S. J.
Med. Chem. 2003, 46, 4552–4563.

1190
S. Sagiraju, B. S. Jursic / Carbohydrate Research 343 (2008) 1180–1190
19. For some pharmaceutically acceptable prodrug half-life
see: (a) AstridMayer, A.; Francis, R. J.; Sharma, S. K.;
Tolner, B.; Springer, C. J.; Martin, J.; Boxer, G. M.; Bell, J.;
Green, A. J.; Hartley, J. A.; Cruickshank, C.; Wren, J.;
Chester, K. A.; Begent, R. H. J. Clin. Cancer Res. 2006, 12,
6509–6516; (b) Leoa, E.; Contado, C.; Bortolotti, F.;
Pavand, B.; Scatturin, A.; Tosia, G.; Manfredini, S.;
Angusti, A.; Dalpiaz, A. Int. J. Pharm. 2006, 307, 103–113.
20. (a) Bagshawe, K. Br. J. Cancer 1987, 56, 531–532; (b)
Ghosh, A. K.; Khan, S.; Farquhar, D. Chem. Commun.
1999, 2527–2528.
gham, A.; Englund, E. E. Org. Biomol. Chem. 2005, 3,
3880–3882.
22. Gadre, A.; Connors, K. A. J. Pharm. Sci. 1997, 86, 1210–
1214.
23. For instance see: (a) D’yachkov, P. N.; Khalepo, A. I.;
Kirakosyan, G. A.; Pervukhina, I. V.; Ibatullina, R. B.
Gigiena Truda i Professional’nye Zabolevaniya 1990, 9, 55–
56; (b) Chem. Abstr. 1991, 114, 1908, Metabolism of
2-phenyloxy-1,1,2,2,-tetraflouro-1-ethanol was study by
19F NMR spectroscopy by analyzing the rats urine upon
the drug injection into the rats stomach..
21. For a recent similar approach of amino acid derived drugs
see: (a) Xu, F.; Hallop, M. A.; Sasikumar, V.; Dilip, U.
Serine amino acid derived prodrugs of propofol, compo-
sition, uses, and crystalline forms thereof, WO 2006/
071995; (b) Wipf, P.; Lynch, S. M.; Powis, G.; Birmin-
24. Ordinarily classical Job’s plot is used to determine the
ratio of components in the complex. However in our case
the plot does not produce conclusive results. For infor-
mation about Job’s plot see: Yannis, L.; Loukas, Y. L.
Analyst 1997, 122, 377–381.