Nucleic acid triggered catalytic drug and probe release: A new concept for the design of chemotherapeutic and diagnostic agents

Article abstract of DOI:10.1016/S0968-0896(01)00245-0

Recently, we described a new concept for the design of highly selective antiviral and anticancer chemotherapeutic agents that makes use of a disease-specific nucleic acid sequence to template the association of a prodrug with a catalyst which catalyzes th

Full text of DOI:10.1016/S0968-0896(01)00245-0

Bioorganic & Medicinal Chemistry 9 (2001) 2501–2510
Nucleic Acid Triggered Catalytic Drug and Probe Release:
A New Concept for the Design of Chemotherapeutic and
Diagnostic Agents
Zhaochun Ma and John-Stephen Taylor*
Department of Chemistry, Campus Box 1134, Washington University, One Brookings Drive, St. Louis, MO 63130, USA
Received 2 February 2001; accepted 5 July 2001
Abstract—Recently, we described a new concept for the design of highly selective antiviral and anticancer chemotherapeutic agents
that makes use of a disease-specific nucleic acid sequence to template the association of a prodrug with a catalyst which catalyzes
the release of the drug. Herein, we report on the effect of mismatches, sterics, and electronics on the rate and specificity of drug and
probe release in both two and three component model systems, and on the stability of the prodrug linker in human serum. # 2001
Elsevier Science Ltd. All rights reserved.
Introduction
Most notable among such approaches is ADEPT (anti-
body directed enzyme prodrug therapy), in which an
antibody that recognizes a disease-specific antigen is
linked to a prodrug metabolizing enzyme which leads to
the release of a cytotoxic agent outside the cell. A rela-
ted approach involves targeting of a gene coding for a
prodrug metabolizing enzyme (GDEPT) by either che-
mical or viral methods to activate the prodrug within
the cell. Unfortunately, the success of these types of
approaches also depends on the existence of significant
biochemical differences between normal and diseased
cells that can be used for targeting, and would likewise
be susceptible to drug resistance.
The goal of chemotherapy is to design or discover drugs
that are selectively toxic to the diseased cell or the dis-
ease causing organism.1,2 This is a quite difficult chal-
lenge for cancer chemotherapy, however, because there
is often little, biochemically, to distinguish a normal cell
from a cancerous cell. Most chemotherapeutic drugs
found by traditional screening approaches have been
found to interfere with replication and cell division and
owe their selectivity to the fact that cancer cells divide
more rapidly than normal cells. Unfortunately, the
chemotherapeutic indices for these drugs are often quite
low. More recently, new approaches to chemotherapy
have sought to take advantage of what has been learned
about the biochemistry of cancer cells to design more
effective drugs.3ꢀ5 Promising as these approaches are,
individual drugs would have to be developed for each
type of cancer, and would still be susceptible to drug
resistance through mutations in the target enzymes or
proteins that are acquired by the rapidly dividing cancer
cells.
An ideal type of prodrug chemotherapy would involve
activation of the prodrug specifically within the diseased
cell without the need for targeting, and without the need
to knowthe biochemical basis of the disease. It is with
these criteria in mind that we have recently proposed a
newand general concept for the design of easily pro-
grammable and highly selective chemotherapeutic drugs
that we have termed nucleic acid triggered catalytic drug
release.13 Our basic idea is to convert a disease-specific
nucleic acid sequence into a prodrug metabolizing cata-
lyst specifically within a diseased cell by way of the high
specificity and simplicity of Watson–Crick base pairing.
In one embodiment of this idea (Fig. 1), the prodrug
metabolizing catalyst is created by a catalytic compo-
nent consisting of a catalytic group attached to an oli-
gonucleotide analogue which binds tightly and
specifically to a unique site on a disease-specific nucleic
Another approach to obtaining highly selectively
chemotherapeutic agents is to further increase the selec-
tivity of known agents by selectively targeting prodrugs,
or prodrug metabolizing enzymes to diseased cells.4ꢀ12
*Corresponding author. Tel.: +1-314-935-6721; fax: +1-314-935-
4481; e-mail: taylor@wuchem.wustl.edu
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00245-0

2502
Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
acid sequence, such as a unique or overexpressed
mRNA sequence. The prodrug in turn consists of a
cytotoxic drug that is attached via a cleavable linker to
an oligonucleotide analogue that binds reversibly to the
site adjacent to the catalytic component binding site.
When the catalytic component binds to the disease-spe-
cific nucleic acid sequence, a prodrug metabolizing
enzyme-like species is created which contains both a
prodrug binding site and a catalytic site. This enzyme-
like catalyst then catalyzes multiple releases of a cyto-
toxic drug from the prodrug. In a normal cell, the dis-
ease-specific nucleic acid sequence is either absent, or in
lowcopy number, and the drug will be inefficiently
released. The beauty of this approach, if it could be
realized, is that one only needs to be able to identify the
unique or over expressed nucleic acid sequences that are
unique to the diseased cell and not their biological
function, something which can now be readily deter-
mined by DNA chip technology.14,15 In the event of a
mutation in the triggering sequence, it would be quite
simple to synthesize a newcomplementary pair of pro-
drug and catalytic components. In the event of acquired
resistance to the effects of the cytotoxic drug used, a
different drug could be readily attached to the prodrug
component.
mer triggering ODN. Herein, we will describe some
further studies of this three component system, as well
as studies of a simpler hairpin model system, in an effort
to better understand the effects of mismatches, sterics,
and electronics on the rate and specificity of drug
release. We will also report on the effect of sterics and
electronics on the stability of the aryl ester subunit of
the prodrug component in human serum.
Results and Discussion
Hairpin model system
When we originally began our studies of this new
approach to chemotherapeutic agents, we made use of a
simpler two component model system in which the cat-
alytic component was directly fused to the triggering
nucleic acid sequence via a hairpin loop (Fig. 1). This
was done so as not to complicate the interpretation of
the kinetics by reversible binding of the catalytic com-
ponent to the triggering nucleic acid, and to insure a
stoichiometric association between the catalytic compo-
nent and the triggering component. A d(CTTG) tetra-
loop flanked by dG and dC was chosen for the hairpin
loop based on a previous study which showed this type
of sequence to form a very stable hairpin structure when
fused to a five base-pair stem.16 The catalytic group was
chosen to be imidazole because it is well known to cat-
alyze the hydrolysis of p-nitrophenylesters17ꢀ19 as well
as other arylesters.20 In addition, release of a phenol is
the key step in the release of cytotoxic drugs such as
daunorubicin, phenol mustards, and fluorouracil from
hydroxymethylphenyl-based prodrugs,21ꢀ23 and taxol
from trimethylene lock-based prodrugs24 (Fig. 2). The
imidazole was attached to the hairpin by a previously
described procedure25 involving coupling of the thio-
benzylester phosphoramidite building block 1 to CPG-
supported ODN in the last synthesis cycle to give the
intermediate thiobenzylester 2 followed by treatment
with excess histamine to give 3a (Fig. 3). Temperature
dependent UV absorbance spectroscopy demonstrates
We have recently validated this newapproach to pro-
grammable chemotherapeutic agents with an in vitro
three component system based on the imidazole cata-
lyzed hydrolysis of p-nitrophenyl esters.13 In that study,
we showed that p-nitrophenol could be released cataly-
tically and with turnover from a prodrug component
consisting of a d-valine-p-nitrophenylester linked octa-
mer ODN and a catalytic component consisting of an
imidazole linked 15-mer ODN in the presence of a 23-
Figure 1. Nucleic acid triggered catalytic drug release. The idea is to
convert a disease specific nucleic acid sequence into a drug releasing
enzyme-like catalyst by complex formation with a complementary
catalyst-bearing nucleic acid or analogue. To the right is a simpler
two-component model system that can be used to evaluate the cataly-
tic efficiency of various catalyst drug combinations.
Figure 2. Prodrugs whose activation is triggered by the initial release
of a phenol group.

Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
2503
that the hairpin 3a is indeed stable at room temperature,
and has a TM of 65 ꢁC in 0.1 M NaCl, and 75 ꢁC in 1 M
NaCl. This compares favorably with the calculated
values of 60 and 71 ꢁC, respectively,26 and a TM of 71 ꢁC
at 1 M NaCl reported for d(GGAGCTTGCTCC).16
were 1:48:493. The observed relative rates are similar to
those of 1:188:340 that would be predicted based on the
initial rate being proportional to (Kd+[S])ꢀ1 if the sys-
tem is following a Michaelis–Menten mechanism (vide
infra), given that the concentration of imidazole hairpin
enzyme and substrate are the same. In addition, it is
assumed that kcat will be the same for all three sub-
strates. Attempts, however, to fit the rate of fluorescein
release from 7a–c in the presence of 3a to Michaelis–
Menten kinetics, or the rate of fluorescein release from
fluorescein dipivalate by imidazole to a one or two step
mechanism were not successful.
Fluorescein diester hairpin system
Because of the strong fluorescence and widespread use
of fluorescein as a fluorescent probe, we had first deci-
ded to investigate the rate of fluorescein release by imi-
dazole catalyzed hydrolysis of diacetyl fluorescein.
Diacetylated fluoresceins are non fluorescent, but upon
hydrolysis of the acyl groups become highly fluorescent.
Attempts to attach the N-hydroxysuccinimide derivative
of diacetylcarboxyfluorescein via amide bond formation
with the 30-amino linked oligodeoxynucleotide 5 (Fig. 4)
failed due to the high pH required for the reaction
which led to the premature hydrolysis of the acetate
groups. Attempts to conjugate 5-chloromethyl fluor-
escein diacetate with a 30-phosphorothioate labeled
ODN also failed at pH 7. Coupling could be achieved,
however, with the more base stable pivaloyl derivative 6
by a procedure that has previously been used to link it
to 50-amino linked oligodeoxynucleotide analogues.27
Three ODNs 7a–c, corresponding to a 4-mer, 6-mer and
8-mer were derivatized with bis-pivaloyl fluorescein to
test the effect of ODN length on the efficiency of
hydrolysis by the imidazole hairpin 3a. The TMs of 7a–c
were estimated from thermodynamic parameters to be
ꢀ20, 19, and 41 ꢁC under the conditions of the experi-
ment and have equilibrium dissociation constants of
3.4 mM, 8.1 mM, and 14 nM.28
As expected, the rate of fluorescein release was greatest
for the 8-mer 7c and roughly corresponded to the rate of
release observed in 0.01 mM imidazole buffer alone at
pH 7 (Fig. 5). The relative initial rates of fluorescein
release for 7a–c under otherwise identical conditions
Figure 4. Synthesis of the fluorescein dipivalate proprobes.
Figure 5. Effect of oligodeoxynucleotide length on the activation of
the fluorescein dipivalate ODNs (10 mM) by the imidazole hairpin 3a
(5 mM) in 0.1 M phosphate buffer (pH 7.0) at 20 ꢁC with excitation at
490 nm and detection at 525 nm. For comparison, the release of fluor-
escein from the fluorescein tetramer 7a by imidazole at 0.01 M imida-
zole alone is also shown.
Figure 3. Synthetic scheme for the preparation of the hairpin enzyme
for the two-component model system and the catalytic component for
the three component system.

2504
Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
ꢀ-Alanine-nitrophenol hairpin system
deoxynucleotide 14 (Fig. 8). The experimental KI and
KM values of 38 and 50 mM (Table 2) are more than 3
orders of magnitude higher than the expected value of
14 nM for the dissociation constant Kd of the prodrug
and inhibitor based on available thermodynamic para-
meters.28 One possibility for the large difference between
KM and the calculated Kd values is that kcat is much
greater than koff.31 Given that the on rate for duplex
formation is weakly temperature and sequence depen-
dent and within an order of magnitude of 106 Mꢀ1 sꢀ1
for many oligodeoxynucleotides,32 the sum of koff and
kcat would have to be about 50 sꢀ1 to account for the
observed KM. Given that the experimentally determined
To get more interpretable kinetic data, we shifted our
studies to monitoring p-nitrophenolate release. Because
of the greater sensitivity of p-nitrophenyl esters to
hydrolysis by base than the pivalate esters, we dis-
covered that we could not use the pH 9 conditions that
had worked successfully for coupling of the pivalate
esters by amide bond formation. Unfortunately, the
yields of amide bond formation at neutral pH by a
variety of methods were very poor and we had to
change the coupling strategy. One recently reported
method for coupling peptides to ODNs under neutral
conditions involves addition of a thiol derivatized ODN
to a maleimide derivatized peptide.29 To this end we
prepared a maleimide linked p-nitrophenyl ester 9 (Fig.
6) by condensing b-alanine with maleic anhydride to
form the maleamic acid derivative 8 which was then
refluxed with thionyl chloride followed by addition of p-
nitrophenol according to a general procedure.30 The
resulting maleimide p-nitrophenyl ester 9 was then
linked to an ODN 8-mer by reduction of the disulfide
protected ODN 14 that was prepared by automated
synthesis utilizing a commercially available support.
kcat value is 2.9Â10ꢀ3 sꢀ1, koff would have to be 50 sꢀ1
.
Thus, it would appear that koffꢂkcat and that p-nitro-
phenolate release is governed by a Michaelis–Menten
mechanism in which KM essentially equals Kd.
When the prodrug 15a was incubated with the imidazole
hairpin 3a, catalytic release of p-nitrophenolate was
observed that depended on the presence of the imida-
zole group (Fig. 7). As one can see, the half life for
release of a single p-nitrophenolate is only about 10 min,
and four turnovers are complete in about 4 h. We found
that the initial rate of p-nitrophenolate release followed
simple Michaelis–Menten kinetics,31 and was subject to
competitive inhibition by the disulfide linked oligo-
Figure 7. Kinetics of p-nitrophenolate release from 20 mM of the mat-
ched b-alanine 8-mer prodrug 15a (X=G) in the presence of 5 mM of
the imidazole hairpin 3a (+imidazole, matched) or the hairpin lacking
the imidazole group (ꢀimidazole, matched) in 10 mM phosphate, 1 M
NaCl, pH 7, at 20 ꢁC. Also shown are the kinetics of p-nitrophenolate
release from the mismatched prodrug component 15a (X=C) in the
presence of the imidazole hairpin 3a (+imidazole, mismatched).
Figure 8. Eadie–Hofstee plots of the kinetic data for p-nitrophenolate
release from the matched b-alanine 8-mer prodrug 15a (X=G) by the
imidazole hairpin in the presence and absence of the 8-mer inhibitor
14.
Figure 6. Synthesis of the p-nitrophenol-based prodrugs and hydro-
xycourmarin-based proprobes.

Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
2505
The substantially higher KI and KM values than pre-
dicted is somewhat perplexing, and might be due to
destabilizing steric and electrostatic interactions caused
by close proximity of the functional groups appended to
the 50- and 30-ends of the ODNs through phosphodiester
linkages. It would not appear that much of the destabi-
lization is due to electrostatics, as only a modest
increase in KM from 50 to 80 mM was observed on
decreasing the salt concentration from 1 M NaCl to
0.1 M (Table 1). The difference between KM and the
calculated Kd might also be due to the inaccuracy in
using thermodynamic parameters obtained close to the
TM to predict equilibrium constants at temperatures
much different than the TM value.33
rate was observed in 1 M NaCl, which could be increased
to a 7.5-fold difference in 0.1 M salt. Again, these unex-
pectedly lowdifferences in rates between matched and
mismatched substrates might result from unfavorable
interactions between the ends of the hairpin and sub-
strate, as well as the inaccuracy of predicting free energy
differences at temperatures remote from the TM.33
D-Valine-p-nitrophenol hairpin system
To be effective as a prodrug, the chemical linkage
between the nucleic acid and the drug must be labile to
the catalytic component but at the same time be stable
to endogenous enzymes. To increase the stability of the
ester linkage between the nucleic acid and the drug
activating component we decided to examine the effect
of substituting the b-alanine linkage with the sterically
more demanding d-valine linkage. Unnatural d-amino
acid esters have been shown to decrease the rate of ester
hydrolysis by esterases, proteases and lipases.35ꢀ38 The
d-valine linkage could be introduced by the same meth-
odology used for incorporating the b-alanine ester, as
we have recently reported.13 When the d-valine ester
16a was used in place of the b-alanine ester 15a with the
imidazole hairpin 3a, the rate of p-nitrophenolate
release dropped significantly. Analysis of the kinetic
data by Michaelis–Menten kinetics showed that the loss
could be attributed primarily to a 8.3-fold drop in kcat
(Table 1). Such a drop in the rate constant is consistent
with what has been previously observed to occur to the
rate of imidazole catalyzed hydrolysis of p-nitropheny-
lacetate upon alkyl substitution.39 When compared to
the rate of hydrolysis by imidazole alone (Table 2), the
imidazole hairpin was found to accelerate the rate of
hydrolysis of the b-alanine and d-valine p-nitrophenyl-
esters to the same extent (446 versus 429, respectively)
as calculated by (kcat/KM)/kIm (Table 1).
The imidazole hairpin system was found to enhance the
rate of ester hydrolysis 446-fold relative to imidazole by
comparing the bimolecular rate constant given by kcat
/
KM to the bimolecular rate constant for imidazole-cata-
lyzed hydrolysis of p-nitrophenylacetate, kIm (Table 2).
We also found that that there was an 11-fold difference
in the initial rates of p-nitrophenolate release at 1 M
NaCl between the completely complementary substrate-
hairpin 3aꢃ 15a (X¼G) and a single CC base-pair mis-
match resulting from a G!C transversion in the sub-
strate component 15a (X¼C) (Fig. 7). Assuming that
the rates are proportional to (Kd+[S])ꢀ1 for the reasons
cited above, the observed 11-fold rate difference would
have to be accounted for by a 15-fold increase in the Kd
of the mismatch substrate. This is more than 3 orders of
magnitude less than that calculated from standard ther-
modynamic parameters which predict a 5Â104 differ-
ence in Kd for the corresponding unmodified ODNs.28,34
A similar insensitivity to a mismatch was also observed
in our previous study of a three component system
involving a d-valine ester linkage and a TC mismatch in
place of TA.13 In that case only a 1.4-fold difference in
Table 1. Kinetic parameters for two component system consisting of the 8-mer prodrugs 15a (X=G), 16a, or 16c and the hairpin enzyme 3a at
20 ꢁC, in 10 mM phosphate buffer pH 7.0
b-Alanine-p-
nitrophenyl ester
15a (X=G)
b-alanine-p-
nitrophenyl ester
15a (X=G)
d-Valine-p-
nitrophenyl ester 16a
1.0 M NaCl
d-Valine hydroxy
coumarin ester 16c
1.0 M NaCl
1.0 M NaCl
0.1 M NaCl
Vmax (mM/s)
Km (mM)
1.4Æ0.1Â10ꢀ2
50Æ7
1.5Æ0.1Â10ꢀ2
82Æ15
1.7Æ0.2Â10ꢀ3
57Æ13
2.22Æ0.05Â10ꢀ4
16Æ1
kcat (sꢀ1
)
2.9Æ0.2Â10ꢀ3
38Æ3
3.0Æ0.2Â10ꢀ3
nd
3.5Æ0.4Â10ꢀ4
nd
4.4Æ0.1Â10ꢀ5
nd
KI (mM)
kIm(Mꢀ1ꢃ sꢀ1
Enhancement
)
1.3aÂ10ꢀ1
446
1.3aÂ10ꢀ1
1.43Æ0.04Â10ꢀ2
1.13Æ0.07Â10ꢀ3
281
429
2433
aFor p-nitrophenylacetate.
Table 2. Kinetic parameters for the imidazole catalyzed hydrolysis of p-nitrophenyl and 3-hydroxycoumarin esters at pH 7 in 0.5 M NaCl at 20 ꢁC,
where ko is the background hydrolysis rate constant, and kIm is the imidazole catalyzed rate constant
Ester
ko (sꢀ1
)
kIm (Mꢀ1 sꢀ1
)
p-Nitrophenylacetate
4.4Â10ꢀ5
0.13
Mal-d-valine-p-nitrophenylacetate, 11a
Mal-d-glycine-coumarin, 12c
Mal-b-alanine-coumarin, 9c
Mal-d-valine-coumarin, 11c
CBZ-glycine-coumarin, 13c
4.1Æ2Â10ꢀ6
1.3Æ6Â10ꢀ4
2.1Æ2Â10ꢀ5
4.9Æ4Â10ꢀ7
1.0Æ5Â10ꢀ4
1.4Æ0.04Â10ꢀ2
7.9Æ0.1Â10ꢀ2
5.7Æ0.05Â10ꢀ2
1.13Æ0.07Â10ꢀ3
9.2Æ0.5Â10ꢀ2

2506
Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
D-Valine-7-hydroxycoumarin ester hairpin system
Stability of the maleimide esters in human serum
Whereas the p-nitrophenolate may be a suitable system
for drug release, it is not very suitable for diagnostic
purposes because of the lowsensitivity of absorbance
spectroscopy used to detect p-nitrophenolate release.
We therefore decided to examine the ability of our imi-
dazole based system to release the well known fluor-
escent probe 7-hydroxycoumarin otherwise known as
umbelliferone.40 Hydroxycoumarin was linked to the 8-
mer through a d-valine linkage to give 16c in the same
way used to link the p-nitrophenol (Fig. 6). When incu-
bated with the imidazole hairpin 3a, release of the
hydroxycoumarin was found to be about 10-fold slower
than for p-nitrophenol. Analysis of the kinetics by
Michaelis–Menten kinetics established that kcat was
about 10-fold lower than for p-nitrophenol release, and
that KM was also lower (Table 1). The 10-fold decrease
in kcat was also observed for kIm (Table 2) and is con-
sistent with the higher pKa of 7-hydroxycoumarin than
p-nitrophenol (7.8 versus 7.15) and the known effect of
increasing pKa of the phenol group on decreasing the
rate of imidazole catalyzed hydrolysis.18,20 The lower
KM for hydroxycoumarin release may be due to addi-
tional stability imparted by intercalation of the cou-
marin into the DNA, though this difference in KM is not
apparent for the three component system to be dis-
cussed later. When compared to hydrolysis by imidazole
alone, a 2433-fold rate acceleration was calculated
(Table 1) which is greater than the rate acceleration of
1100 observed for hydrolysis of a related hydro-
xycoumarin by a semisynthetic catalytic antibody.41
To determine whether or not prodrugs or proprobes
based on the d-valine ester linkages would be suitable
for use in humans we investigated the stability of the
maleimide esters of o-nitrophenol and 3-hydroxy-
coumarin D-11b and D-11c in human serum. We found
that these two esters are hydrolyzed at a significant rate
at 25 ꢁC with a half life of about 3 h (Fig. 9). Interest-
ingly, the two rates appear to be similar, suggesting that
the rate determining step for hydrolysis in serum may
not be the initial release of the p-nitrophenol or the
hydroxy coumarin, but may instead be the hydrolysis of
a rapidly formed acyl enzyme intermediate. We also
compared the stability of the d-valine esters to the mal-
eoyl-b-alanine, maleoyl-glycine and CBZ-glycine cou-
marin esters, 9c, 12c and 13c, and the l-valine ester L-
11c. We found that the maleoyl-b-alanine and maleoyl-
glycine esters were more stable than the CBZ-glycine
ester, but much less stable than the maleoyl-d-valine
esters to hydrolysis by human serum. Presumably, the
maleoyl group makes these derivatives poorer substrates
for the enzymes involved in the hydrolysis reaction, as
there is not a substantial difference in reactivity between
these substrates with respect to imidazole catalyzed
hydrolysis (Table 2). The stereochemistry at the a-car-
bon also plays a role, as the d-valine esters are more
slowly hydrolyzed by about of factor of about two than
the l-valine ester L-11c.
Implications of the kinetics for drug or probe selectivity
in vivo
There would appear to be two parameters of impor-
tance in determining the maximum degree of selectivity
achievable from the nucleic acid triggered drug and
probe release system described. The first is the number
of copies of the triggering nucleic acid in the diseased
cell relative to the normal cell, and the second is whether
or not the prodrug binding site on the triggering nucleic
acid differs in sequence, and hence KM between the
normal and diseased cells. From a simple consideration
of the Michaelis–Menten mechanism and the results of
this study the relative rates would appear to be given by:
D-Valine-7-hydroxycoumarin 3-component system
Having established that the imidazole hairpin could
catalyze the release of both p-nitrophenol and hydro-
xycoumarin that were linked to an 8-mer substrate, we
investigated the ability of the three component system
to release hydroxycoumarin. We have previously repor-
ted on the kinetics of the three component d-valine-
linked p-nitrophenol system13 in which the imidazole
hairpin component is replaced by a imidazole 15-mer 3b
which binds to a complementary 23-mer 4 (X=A) that
corresponds to a disease-specific triggering sequence. In
this case, the complex formed between the catalytic
component 3b and the triggering ODN 4 (X=A) was
designed to be highly stable and to function as an
enzyme. This is justified considering the predicted
kinetics of dissociation of the complex. The 15-mer was
calculated to have a Kd’s at 20 ꢁC of 4.3 fM and 0.012
fM in 0.1 and 1 M NaCl,28 respectively, and duplex half
lives of much greater than a year based on an on rate of
ꢀdiseased ½NAŠdiseasedðKd þ ½prodrugŠÞnormal
¼
ꢀnormal
½NAŠnormalðKd þ ½prodrugŠÞdiseased
Table 3. Kinetic parameters for three component system consisting
of the 8-mer prodrug 16a or proprobe 16c, 15-mer catalytic compo-
nent 3b, and 23-mer template 4 (X=A) in 10 mM sodium phosphate
buffer, pH 7.0 at 20 ꢁC
106 Mꢀ1 sꢀ1 32
As was the case for the three component
.
d-Valine-nitrophenylester
16aa 1.0 M NaCl
d-Valine-coumarin
ester 16c 1.0 M NaCl
p-nitrophenol system, release of hydroxycoumarin fol-
lowed Michaelis–Menten kinetics. As we had observed
for the imidazole hairpin system, replacing p-nitrophe-
nol with coumarin in the three component system
resulted in about a 10-fold decrease in rate that could
be attributed to a 10-fold drop in kcat (Table 3). Like-
wise, the kinetic parameters for both the hairpin and
three component systems were also found to be quite
similar.
Vmax (mM/s)
Km (mM)
1.5Æ0.02Â10ꢀ3
22Æ1.2
1.54Æ0.04Â10ꢀ4
18Æ1
kcat (sꢀ1
)
3.0Æ0.04Â10ꢀ4
1.7Æ
3.08Æ0.08Â10ꢀ5
9Æ1
KI (mM)
kIm (Mꢀ1/s)
Enhancement
1.4Æ0.04Â10ꢀ2
974
1.13Æ0.07Â10ꢀ3
1514
aFrom Ma and Taylor.13

Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
2507
Conclusion
We have begun to get a better understanding of the
scope and limitations of the imidazole-based nucleic
acid triggered catalytic drug and probe releasing system
and have found that the rate of their release is sensitive
to both steric and electronic effects. We have also found
that the Kms are higher and mismatch discrimination
lower than expected from standard thermodynamic
parameters, which may have to do with the nature of
the linkers used to tether the catalysts and probes to the
oligodeoxynucleotides. It still remains to be determined
whether the imidazole system can be made to be suitable
for in vivo drug or probe release purposes, as even the
d-valine esters of the o-nitrophenyl and 7-hydro-
xycoumarin esters are hydrolyzed at a significant rate
within human plasma. To work well, the prodrug and
proprobe components will have to be much more resis-
tant to hydrolysis by endogenous enzymes but not to
the drug releasing catalyst. Unfortunately, we found
that increased substitution of the ester linkage causes a
decrease in the rate of imidazole catalyzed hydrolysis as
does an increase in the pKa of the displaced phenol.
Nonetheless, if nucleic acid triggered catalytic drug
release could be realized in vivo, it would be superior to
other approaches because it would only depend on the
presence of a disease specific nucleic acid sequence and
not on a biochemical targeting or activating mechanism.
The additional beauty of this system would appear to be
the potential to use thermodynamic parameters to opti-
mize the sequence of the prodrug component and hence
the selectivity for a specific triggering sequence, and
thus the ability to tailor a drug system for an individual
patient.
Figure 9. Stability of various esters that could be used as prodrug lin-
kers in human serum at 20 ꢁC temperature at pH 8.3.
where [NA] represents the concentration of the nucleic
acid trigger, Kd is the dissociation constant for the drug
component binding to the triggering sequence, [pro-
drug] is the concentration of the prodrug. In deriving
this equation, the assumption is made that enough cat-
alytic component is present with a high enough binding
constant to saturate the nucleic acid trigger and hence
the enzyme concentration would equal the concentra-
tion of the nucleic acid trigger. The second assumption
is that, independent of any specific targeting system, the
concentration of drug component would be the same in
all cells. A third assumption is that kcat will be the same
for any catalytic complex that assembles in either the
normal or the diseased cells. Thus, selective drug release
could be achieved by either or both the presence of an
overexpressed triggering sequence, such as an over-
expressed mRNA sequence, or by having the prodrug
binding site differ in sequence, and hence KM. For the
latter two cases, it is clear from the derived expression
for selectivity, that it would be important to have
Kdꢂ[prodrug] to achieve maximum selectivity.
Experimental
Materials and methods
Dichloromethane and triethylamine were dried by
refluxing with CaH2 overnight followed by distillation.
Benzyl mercaptan, dimethylaminopyridine (DMAP),
1,3-dicyclohexylcarbodiimide (DCC), dichloroacetic acid,
2 - (cyanoethyl) - N,N,N0,N0 - tetraisopropylphosphorodi-
amidite, histamine, maleic anhydride, thionyl chloride
and b-alanine were purchased from Aldrich (Milwau-
kee, WI). Tris(2-carboxyethyl)phosphine hydrochloride
(TCEPꢃ HCl) was purchased from Pierce (Rockford,
IL). Reagents for automatic oligonucleotides synthesis
were purchased from Glen Research (Sterling, VA).
Oligonucleotides were assembled on an Applied Bio-
systems 380 DNA synthesizer using phosphoramidite
chemistry and recommended protocols (DMTr off
synthesis). Oligonucleotides were purified by reversed
Another important consideration in designing the
nucleic acid catalyzed prodrug releasing components is
that they are only activated by the disease-specific
nucleic acid sequence, and that drug release is efficient.
In humans, a minimum sequence length of 15–17
nucleotides has been suggested to be required to
uniquely recognize a specific RNA transcript. In our
system, the required specificity is embedded within the
catalytic component, which by design is composed of a
long sequence to anchor it to the RNA. Also, as part of
the design, the prodrug component must have a suffi-
cient off rate to allowfor turnover. If the length of the
sequence is too short and hence not specific, the rate of
drug release could be inhibited by non-productive bind-
ing of the prodrug component to other accessible sites.
Thus it would also be important to design a sequence of
sufficient length to insure specificity of prodrug binding,
and of sufficiently lowbinding affinity to insure that it
would have a fast enough off rate to allow for rapid
turnover. If necessary, lowbinding affinity could be
engineered into the sequence by use of an affinity low-
ering backbone analogue, or appropriate substituents.
phase HPLC as described belowin detail. 1H NMR, 13
C
NMR and 31P NMR spectra were obtained on either a
Varian UnityPlus-300 (300 MHz) or Varian Mercury-
300 (300 MHz) spectrometer. The chemical shifts are
expressed in ppm downfield from residual chloroform
(d=7.24) and acetone (d=2.04) as an internal standard.
31P NMR spectra were referenced against 85% H3PO4 in
a coaxial insert. Flash chromatography was performed

2508
Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
on Selecto Scientific silica gel. Thin layer chromato-
graphy (TLC) was run on pre-coated 254 nm fluorescent
silica gel sheets manufactured by Alltech (Dearfield, IL).
UV spectral data were acquired on a Bausch and Lomb
Spectronic 1001 spectrophotometer or Varian Cary 1E
UV–vis Spectrophotometer. Fluorescence measure-
ments were carried out on a SPEX Fluoromax instru-
ment. MALDI mass spectra of oligodeoxynucleotides
were measured on PerSeptive Voyager RP MALDI-
TOF mass spectrometer.
washed with water (3Â10 mL) followed by anhydrous
ethanol (3Â10 mL), and then anhydrous ether
(3Â10 mL). After drying, 11.6 g (69%) of the maleamic
1
acid was obtained. H NMR (300 MHz, D2O) d 2.45 (t,
J=6.4 Hz, 2H), 3.32 (t, J=6.4 Hz, 2H), 6.06 (d,
J=12.4 Hz, 1H), 6.26 (d, J=12.4 Hz, 1H);
HRMS(FAB) calcd for C7H10NO5 (M+H+) 188.0559,
found 188.0558.
Synthesis of N-maleoyl-ꢀ-alanine-4-nitrophenol ester
(9a). The maleic acid 6 was dissolved in 20 mL of thio-
nyl chloride and heated at reflux until gas evolution had
ceased. The excess thionyl chloride was evaporated
under reduced pressure. Carbon tetrachloride was
added to the remaining material, and the resulting
solution was evaporated under reduced pressure to
insure complete removal of thionyl chloride. The
resulting product was dissolved in 20 mL of CH2Cl2 and
was slowly added to a stirred mixture of 4-nitrophenol
(0.82 g, 5.9 mmol) and triethylamine (1.6 mL, 12 mmol)
in 20 mL of CH2Cl2 at 0 ꢁC. The reaction mixture was
allowed to warm to room temperature. After stirring an
additional half hour at room temperature, the reaction
mixture was diluted with 200 mL CH2Cl2, washed with
brine, water, dried over Na2SO4 and concentrated in
vacuo. The residue was recrystallized in ethyl acetate
and hexane mixture (v/v=1:1) to afford the ester 9a,
1.5 g (88%). 1H NMR (300 MHz, CDCl3) d 2.92 (t,
J=6.9 Hz, 2H), 3.98 (t, J=6.9 Hz, 2 H), 6.74 (s, 2H),
7.30 (d, J=9.0 Hz, 2H), 8.26 (d, J=9.0 Hz, 2H); 13C
NMR (300 MHz, CDCl3) d 33.3, 33.4, 122.5, 125.2,
134.3, 145.4, 155.0, 168.4, 170.2; HRMS calcd for
C13H10N2O6Na (M+Na+) 313.0436, found 313.0439.
Synthesis of the imidazole hairpin (3a). The oligonucleo-
tides were assembled on commercial nucleoside derivatized
columns using standard protocols. Phosphoramidite1
(reference) was used in the last coupling step (0.1 M in
acetonitrile; 30 min coupling time). After oxidation with
0.1 M iodine, the protected ODN 2a was treated with
0.2 M histamine in water for 6 h. 1 Complete deprotec-
tion was carried out in concentrated aqueous ammonia
at 55 ꢁC for 8 h. The ammonia solution was evaporated
to dryness on a Savant Speedvac, first under water
aspirator pressure and then under high vacuum to yield
the crude oligomer 3a. It was then dissolved in doubly
distilled water and purified by reversed phase HPLC on
a Rainin Dynamax column (C-18, 5 mm, 4.6Â250 mm)
using two buffers: A [100 mM triethyl ammonium ace-
tate buffer pH 7.0(50%)/water (50%)] and B [100 mM
triethyl ammonium acetate buffer pH 7.0 (50%)/aceto-
nitrile (50%)]. A linear gradient was run from 0 to 30%
B in 30 min, flowrate=1.0 mL/min. The wavelength of
the detector was set at 260 nm. The pure fraction was
collected, concentrated and desalted by using the same
column, washed with excess doubly distilled water and
eluted with 50:50 acetonitrile/water. The desalted frac-
tions were combined and concentrated to dryness in
vacuo. The product 3a was analyzed by MALDI-TOF,
calcd (MꢀH+) 7622.0, found 7621.3.
Synthesis of N-maleoyl-ꢀ-alanine-coumarin ester (9c).
The maleamic acid of b-alanine 8 (475 mg, 2.54 mmol)
was dissolved in 6 mL of thionyl chloride and heated at
reflux until gas evolution had ceased. The excess thionyl
chloride was evaporated under reduced pressure. Car-
bon tetrachloride was added to the remaining material,
and the resulting solution was evaporated under
reduced pressure to ensure complete removal of thionyl
chloride. The resulting product was dissolved in 10 mL
of THF and was slowly added to a stirred mixture of 7-
hydroxycoumarin (411 mg, 2.54 mmol) and triethyla-
mine (0.74 mL, 5.36 mmol) in 20 mL of THF at 0 ꢁC.
The reaction mixture was allowed to warm to room
temperature. After stirring an additional half hour at
room temperature, the reaction mixture was diluted
with 200 mL ethyl acetate, washed with brine, water,
dried over Na2SO4 and concentrated in vacuo. The
residue was flash chromatographed on silica gel (1:1
ethyl acetate/hexane) to afford the N-maleoyl-b-alanine
Synthesis of fluorescein dipivalate oligodeoxynucleotide
conjugates (7). To a mixture of DMF (32 mL) and
0.1 M NaHCO3/Na2CO3 buffer (pH 9) (140 mL) were
added successively a solution of 30-amino-linked oligo 5
(1.2 mM in water, 36 nmol) and a solution of the acti-
vated ester 6 (12 mM in DMF, 1.44 mmol). The turbid
mixture was stirred vigorously at room temperature for
1.5 h in the dark. The oligodeoxynucleotide conjugates 7
were purified by preparative reverse-phased HPLC on a
Ranin Dynamax column (solvent A=0.05 M triethyl-
ammonium acetate, pH 7; solvent B=80% acetonitrile
in buffer A; linear gradient, from 7 to 63% of B over
20 min and then to 100% of B over 20 min; flow
rate=1 mL minꢀ1). Yield (measured by UV absorbance
at 260 nm) was 25% after purification. The conjugates
were analyzed by MALDI-MS m/z [MꢀH] 7a: calcd
1910.5, obsd 1909.2; 7b: calcd 2503.9, obsd 2502.5; 7c:
calcd 3097.3, obsd 3095.8.
1
ester 9c, 300 mg (38%). H NMR (300 MHz, CDCl3) d
2.92 (t, J=6.6 Hz, 2H), 3.97 (t, J=6.8 Hz, 2H), 6.41 (d,
J=9.6 Hz, 1H), 6.76 (s, 2H), 7.06–7.10 (m, 1H), 7.14 (d,
J=2.2 Hz, 1H), 7.50 (d, J=8.2 Hz, 1H), 7.70 (d,
J=9.3 Hz, 1H).
Synthesis of the maleamic acid of ꢀ-alanine (8). b-Ala-
nine (8.9 g, 100 mmol) was dissolved in 10 mL of water.
Maleic anhydride (9.8 g, 100 mmol) was added all at
once and the mixture was stirred for 4 h at ambient
temperature. After completion of the reaction, the mix-
ture was filtered and the white power obtained was
Synthesis of N-maleoyl-D-valine-2-nitrophenol ester (D-
11b). The maleamic acid of d-valine 10 (500 mg,
2.54 mmol) was dissolved in 10 mL of thionyl chloride
and heated at reflux until gas evolution had ceased. The

Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
2509
excess thionyl chloride was evaporated under reduced
pressure. Carbon tetrachloride was added to the
remaining material, and the resulting solution was eva-
porated under reduced pressure to insure complete
removal of thionyl chloride. The resulting product was
dissolved in 10 mL of CH2Cl2 and was slowly added to a
stirred mixture of 2-nitrophenol (323 mg, 2.32 mmol)
and triethylamine (0.64 mL, 4.64 mmol) in 10 mL of
CH2Cl2 at 0 ꢁC. The reaction mixture was allowed to
warm to room temperature. After stirring an additional
half hour at room temperature, the reaction mixture
was diluted with 200 mL CH2Cl2, washed with brine,
water, dried over Na2SO4 and concentrated in vacuo.
The residue was flash chromatographed on silica gel (1:4
ethyl acetate/hexane) to afford the N-maleoyl-d-valine
ester D-11b, 270 mg (37%). [a]D=+25.5ꢁ (c 0.94, ace-
tone); 1H NMR (300 MHz, CD3COCD3) d 0.93 (d,
J=6.9 Hz, 3H), 1.14 (d, J=6.9 Hz, 3H), 2.68–2.75 (m,
1H), 4.72 (d, J=8.0 Hz, 1H), 7.06 (s, 2H), 7.34–7.37 (m,
1H), 7.56–7.61 (m, 1H), 7.81–7.87 (m, 1H), 8.14–8.17
(m, 1H); 13C NMR (300 MHz, CD3COCD3) d 18.6,
20.2, 57.2, 125.3, 125.8, 127.6, 134.8, 135.3, 143.4, 166.5,
170.2.
buffer, pH 7.0, for 2 h at room temperature under
argon. Maleimide ester (1 mmol in 50 mL acetonitrile)
was added without elimination of the TCEP excess. The
oligodeoxynucleotide conjugates were purified by
reversed-phase HPLC on a Rainin Dynamax column
(C-18, 5 mm, 4.6Â250 mm) using two buffers: A (10%
methanol and 90% 75 mM sodium phosphate buffer pH
7.0) and B (50% methanol and 50% 75 mM sodium
phosphate buffer pH 7.0). A linear gradient was run
from 0 to 100% B in 30 min, flowrate=1.0 mL/min and
the effluent was monitored by its absorbance at 260 nm.
The desired fraction was collected, concentrated and
desalted by using the same column, washed with excess
doubly distilled water and eluted with 50:50 acetonitrile/
water. The desalted fractions were combined and con-
centrated to dryness in vacuo. The product was ana-
lyzed by MALDI-TOF, 15a calcd (MꢀH+) 2803.2,
found 2802.9; 16c calcd (MꢀH+) 2831.3, found 2832.6;
16c calcd (MꢀH+) 2853.4, found 2852.6.
Kinetics of p-nitrophenol and 7-hydroxycoumarin release
from the hairpin and three-component systems
For typical assays, phenol or coumarin oligodeox-
ynucleotide conjugates 15 or 16 and the imidazole hair-
pin 3a or the complex 3b/4 were incubated in 10 mM
sodium phosphate, pH 7.0, which contained 0.1 or
1.0 M NaCl. The reaction temperature was maintained
at 20 ꢁC, and the production of phenolate or 7-hydro-
xycoumarin was monitored by UV absorbance at
400 nm (e400=6.26Â103 at pH 7) or by fluorescence
(lEx=355 nm, lEm=452 nm). Initial velocities of the
reaction, obtained for each substrate concentration,
were fitted to the Michaelis–Menten equation by a non-
linear least squares method using KaleidaGraph soft-
ware. The inhibition constant KI was determined in the
presence of 25 mM 7 d(TCCTGTCA) in 10 mM sodium
phosphate buffer pH 7.0 containing 1.0 M NaCl by the
plotting 1/v versus 1/[S] and calculating KI from the
slope of the line according following equation:
Synthesis of N-maleoyl-D-valine coumarin ester (D-11c).
The maleamic acid of D-valine 8 (500 mg, 2.54 mmol)
was dissolved in 10 mL of thionyl chloride and heated at
reflux until gas evolution had ceased. The excess thionyl
chloride was evaporated under reduced pressure. Car-
bon tetrachloride was added to the remaining material,
and the resulting solution was evaporated under
reduced pressure to insure complete removal of thionyl
chloride. The resulting product was dissolved in 10 mL
of CH2Cl2 and was slowly added to a stirred mixture of
7-hydroxycoumarin (376 mg, 2.32 mmol) and triethyla-
mine (0.64 mL, 4.64 mmol) in 10 mL of CH2Cl2 at 0 ꢁC.
The reaction mixture was allowed to warm to room
temperature. After stirring an additional half hour at
room temperature, the reaction mixture was diluted
with 200 mL CH2Cl2, washed with brine, water, dried
over Na2SO4 and concentrated in vacuo. The residue
was flash chromatographed on silica gel (1:2 ethyl ace-
tate/hexane) to afford the N-maleoyl-d-valine ester D-
½IŠ
KMð1 þ
Vmax
Þ
11c, 390 mg (49%). [a]D=+85.2ꢁ (c 1.0, acetone); H
1
1
1
1
KI
¼
þ
ꢀ
NMR (300 MHz, CD3COCD3) d 0.93 (d, J=6.8 Hz,
3H), 1.11 (d, J=6.8 Hz, 3H), 2.60–2.64 (m, 1H), 4.79 (d,
J=6.9 Hz, 1H), 6.40 (d, J=9.6 Hz, 1H), 7.05–7.10 (m,
4H), 7.72 (d, J=8.2 Hz, 1H), 7.99 (d, J=9.6 Hz, 1H);
13C NMR (75 MHz, CD3COCD3) d 19.0, 20.8, 57.3,
110.8, 116.8, 118.0, 119.1, 130.0, 135.5, 144.0, 153.8,
155.5, 160.1, 167.8, 171.2, 205.9, 206.2; HRMS (FAB)
calcd for C18H16NO6 [M+H+] 342.0978, found
342.0964.
ꢀ
Vmax
½SŠ
where v is the velocity of the reaction, Vmax is the max-
imum velocity, KM is the substrate concentration at half
maximal reaction rate, [I] is the inhibitor concentration,
and [S] is the substrate concentration.
Kinetics of p-nitrophenol and 7-hydroxycoumarin release
by imidazole alone
Synthesis of N-maleoyl-L-valine coumarin ester (L-11c).
This was prepared in 44% yield by the same method
described for the d-valine derivative, D-11c.
[a]D=ꢀ83.5ꢁ (c 1.0, acetone).
For comparison purposes, the rate constant kIm for the
imidazole catalyzed hydrolysis of the various esters were
determined in imidazole buffer at pH 7.0, 0.5 M NaCl at
20 ꢁC as previously described.42 The imidazole buffers
were prepared in 0.5 M NaCl, and adjusted to pH 7 by
addition of 1 M HCl. The hydrolysis reaction was fol-
lowed by monitoring the absorbance of the phenolate
General method for synthesis of the ester oligodeoxynu-
cleotide conjugates 15 and 16. Oligodeoxynucleotide 14
(100 nmol) bearing a disulfide group4 was reduced with
TCEP (150 nmol) in 500 mL 0.1 M sodium phosphate

2510
Z. Ma, J.-S. Taylor / Bioorg. Med. Chem. 9 (2001) 2501–2510
ion at 400 nm or the fluorescence emission of 7-hydro-
xycoumarin at 452 nm (lEx=350 nm) as a function of
time. In a typical run, 5 mL of the ester (8 mM in aceto-
nitrile for the phenyl esters and 200 mM in acetonitrile for
the 7-hydroxycoumarin esters) were added to a cuvette
containing 400 mL of imidazole buffer (0.004–0.6 M),
capped and mixed by inverting several times. The pseudo
first-order rate constant for each concentration of imi-
dazole was obtained by linear least squares fitting of
ln(A ꢀA) or ln(F ꢀF) versus time. The rate constants
11. Senter, P. D.; Svensson, H. P. Adv. Drug Deliv. Rev. 1996,
22, 341.
12. Dubowchik, G. M.; Walker, M. A. Pharmacol. Ther.
1999, 83, 67.
13. Ma, Z.; Taylor, J.-S. Proc. Natl. Acad. Sci. U.S.A. 2000,
97, 11159.
14. Golub, T. R.; Slonim, D. K.; Tamayo, P.; Huard, C.;
Gaasenbeek, M.; Mesirov, J. P.; Coller, H.; Loh, M. L.;
Downing, J. R.; Caligiuri, M. A.; Bloomfield, C. D.; Lander,
E. S. Science 1999, 286, 531.
15. Gerhold, D.; Rushmore, T.; Caskey, C. T. Trends Bio-
chem. Sci. 1999, 24, 168.
16. Antao, V. P.; Tinoco, I., Jr. Nucleic Acids Res. 1992, 20,
819.
17. Bender, M. L. J. Am. Chem. Soc. 1957, 79, 1652.
18. Bruice, T. C.; Schmir, G. L. J. Am. Chem. Soc. 1957, 79,
1663.
1
1
were then plotted against total imidazole concentration
to get ko, the first order rate constant for the uncata-
lyzed hydrolysis reaction, and kIm, the second-order rate
constant for catalysis by imidazole buffer.
19. Bruice; Sturtevant, J. Am. Chem. Soc. 1959, 81, 2860.
20. Bender, M. L.; Turnquest, B. W. J. Am. Chem. Soc. 1957,
79, 1656.
21. Gesson, J. P.; Jacquesy, J. C.; Mondon, M.; Petit, P.;
Renoux, B.; Andrianomenjanahary, S.; Dufat-Trinh Van, H.;
Koch, M.; Michel, S.; Tillequin, F.; Florent, J. C.; Monneret,
C.; Bosslet, K.; Czech, J.; Hoffmann, D. Anti-C Drug Des.
1994, 9, 409.
22. Lougerstay-Madec, R.; Florent, J.-C.; Monneret, C.;
Nemati, F.; Poupon, M. F. Anti-Cancer Drug Des. 1998, 13,
995.
23. Madec-Lougerstay, R.; Florent, J.-C.; Monneret, C. J.
Chem. Soc., Perkin Trans. 1, 1369.
24. Ueda, Y.; Matiskella, J. D.; Mikkilineni, A. B.; Farina,
V.; Knipe, J. O.; Rose, W. C.; Casazza, A. M.; Vyas, D. M.
Bioorg. Med. Chem. Lett. 1995, 5, 247.
25. Hovinen, J.; Guzaev, A.; Azhayeva, E.; Azhayev, A.;
Lonnberg, H. J. Org. Chem. 1995, 60, 2205.
26. SantaLucia, J., Jr. Proc. Natl. Acad. Sci. U.S.A. 1998, 95,
1460.
Hydrolysis of esters in human serum
The hydrolysis of the coumarin esters 9c, 11c, 12c, and
13c was followed in a SPEX Fluoromax spectro-
fluorimeter at the following wavelengths: lEx=350 nm,
lEm=452 nm. In a typical run, 3 mL of coumarin ester
solution (7 mM in acetonitrile) was added to 400 mL of
Human serum, pH 8.3 (Innovative Research Inc.) in a
cuvette. The cuvette was capped, inverted several times
for thorough mixing, placed in a cell block and record-
ing began. The appearance of 7-hydroxycoumarin was
monitored at 452 nm at room temperature. The same
procedure was followed for monitoring the hydrolysis
of the o-nitrophenyl ester 11b except that the solution
was centrifuged in an Eppendorf centrifuge prior to
making absorbance measurements at 400 nm.
Acknowledgements
27. Laurent, A.; Debart, F.; Lamb, N.; Rayner, B. Bioconju-
gate Chem. 1997, 8, 856.
This work was supported in part by NIH grant
CA40463 and a Wheeler Fellowship for Z. Ma. The
assistance of the Washington University High Resolu-
tion NMR Facility, funded in part through NIH Bio-
medical Research Support Shared Instrument Grants
RR-02004, RR-05018, and RR-07155 is gratefully
acknowledged, as is the Washington University Mass
Spectrometry Resource, an NIH Research Resource
(Grant No. P41RR0954).
28. SantaLucia, J., John.; Allawi, H. T.; Seneviratne, P. A.,
Biochemistry 1996, 35, 3555.
29. Arar, K.; Monsigny, M.; Mayer, R. Tetrahedron Lett.
1993, 34, 8087.
30. Frazier, K. A. US Patent 4,980,482, 1990.
31. Fersht, A. Enzyme Structure and Mechanism; W. H. Free-
man: NewYork, 1985.
32. Bloomfield, V. A.; Crothers, D. M.; Inioco, I. Nucleic
Acids: Structures, Properties, and Functions; University
Science Books: Sausalito, CA, 1999.
33. Gelfand, C. A.; Plum, G. E.; Mielewczyk, S.; Remeta,
D. P.; Breslauer, K. J. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
6113.
34. Peyret, N.; Seneviratne, P. A.; Allawi, H. T.; SantaLucia,
J., Jr. Biochemistry 1999, 38, 3468.
References and Notes
1. Albert, A. Selective Toxicity. The Physico-chemical Basis of
Therapy, 6th ed.; Chapman and Hall: London, 1979.
2. Drews, J. Science 2000, 287, 1960.
35. Stoops, J. K.; Horgan, D. J.; Runnegar, M. T. C.; De
Jersey, J.; Webb, E. C.; Zerner, B. Biochemistry 1969, 8,
2026.
3. Gibbs, J. B. Science 2000, 287, 1969.
4. Denny, W. A. Curr. Pharm. Des 1996, 2, 281.
5. Sinhababu, A. K.; Thakker, D. R. Adv. Drug Deliv. Rev.
1996, 19, 241.
6. Melton, R.; Connors, T.; Knox, R. J. S.T.P. Pharma Sci.
1999, 9, 13.
7. Chari, R. V. J. Adv. Drug Deliv. Rev. 1998, 31, 89.
8. Sherwood, R. F. Adv. Drug Deliv. Rev. 1996, 22, 269.
9. Springer, C. J.; Niculescu-Duvaz, I. Adv. Drug Deliv. Rev.
1996, 22, 351.
10. Connors, T. A.; Knox, R. J. Stem Cells (Dayton) 1995,
13, 501.
36. Chakravarty, P. K.; Carl, P. L.; Weber, M. J.; Katze-
nellenbogen, J. A. J. Med. Chem. 1983, 26, 633.
37. Campbell, D. A.; Gong, B.; Kochersperger, L. M.; Yon-
kovich, S.; Gallop, M. A.; Schultz, P. G. J. Am. Chem. Soc.
1994, 116, 2165.
38. Miyazawa, T. Amino Acids 1999, 16, 191.
39. Milstien, J. B.; Fife, T. H. J. Amer. Chem. Soc. 1968, 90,
2164.
40. Fink, D. W.; Koehler, W. R. Anal. Chem. 1970, 42, 990.
41. Pollack, S. J.; Schultz, P. G. J. Am. Chem. Soc. 1989, 111,
1929.
42. Lombardo, A. J. Chem. Educ. 1982, 59, 887.