Synthesis, properties, and reactivity of cocaine benzoylthio ester possessing the cocaine absolute configuration

Article abstract of DOI:10.1021/ja012376y

One aspect of immunopharmacotherapy for cocaine abuse involves the use of a catalytic monoclonal antibody (mAb) to degrade cocaine via hydrolysis of the benzoate ester. A cocaine benzoylthio ester analogue provides a means to implement high-throughput selection strategies to potentially isolate mAbs with high activity. The required analogue was synthesized starting from (-)-cocaine hydrochloride and possessed the cocaine absolute configuration. Key points in the preparation were the introduction of the sulfur atom at C-3 via a bromomagnesium thiolate addition to the exo face of anhydroecgonine, separation of C-2 diastereomers, recycling of a C-2 thio ester byproduct, and formation of the necessary C-2 methyl and C-3 benzoylthio esters. Effects resulting from the lower electronegativity and greater hydrophobicity of sulfur compared to oxygen were observed. These characteristics could result in interesting drug properties. Furthermore, the analogue was found to be a substrate for catalytic mAbs that hydrolyze cocaine as monitored by HPLC and also spectrophotometry by coupling cleavage of the benzoylthio ester to the disulfide exchange with Ellman's reagent. Screening antibody libraries with the new cocaine analogue using the spectroscopic assay provides an avenue for the high-throughput identification of catalysts that efficiently breakdown cocaine.

Full text of DOI:10.1021/ja012376y

Published on Web 03/14/2002
Synthesis, Properties, and Reactivity of Cocaine Benzoylthio
Ester Possessing the Cocaine Absolute Configuration
Shigeki Isomura, Timothy Z. Hoffman, Peter Wirsching,* and Kim D. Janda*
Contribution from the Department of Chemistry BCC-582, The Scripps Research Institute,
10550 North Torrey Pines Road, La Jolla, California 92037
Received October 16, 2001
Abstract: One aspect of immunopharmacotherapy for cocaine abuse involves the use of a catalytic
monoclonal antibody (mAb) to degrade cocaine via hydrolysis of the benzoate ester. A cocaine benzoylthio
ester analogue provides a means to implement high-throughput selection strategies to potentially isolate
mAbs with high activity. The required analogue was synthesized starting from (-)-cocaine hydrochloride
and possessed the cocaine absolute configuration. Key points in the preparation were the introduction of
the sulfur atom at C-3 via a bromomagnesium thiolate addition to the exo face of anhydroecgonine,
separation of C-2 diastereomers, recycling of a C-2 thio ester byproduct, and formation of the necessary
C-2 methyl and C-3 benzoylthio esters. Effects resulting from the lower electronegativity and greater
hydrophobicity of sulfur compared to oxygen were observed. These characteristics could result in interesting
drug properties. Furthermore, the analogue was found to be a substrate for catalytic mAbs that hydrolyze
cocaine as monitored by HPLC and also spectrophotometry by coupling cleavage of the benzoylthio ester
to the disulfide exchange with Ellman’s reagent. Screening antibody libraries with the new cocaine analogue
using the spectroscopic assay provides an avenue for the high-throughput identification of catalysts that
efficiently breakdown cocaine.
Cocaine abuse continues to be prevalent. Recent surveys for
the United States indicate that more than 23 million people have
tried cocaine and approximately 2.5 million are considered
chronic abusers.1 Numerous medical problems, including car-
diovascular toxicity, brain damage, and death, often accompany
cocaine use, and the association of the drug with the spread of
AIDS is of concern.2 Despite intensive efforts, the development
of effective therapies for cocaine craving and addiction remain
elusive. New strategies are necessary to increase the effective-
ness of rehabilitative programs and for the treatment of overdose
cases. We, and others, have shown that one aspect of immuno-
pharmacotherapy, the antibody-mediated binding of cocaine to
impede passage of the drug into the central nervous system,
resulted in a suppression of its characteristic actions.3,4 Admin-
istration of a monoclonal antibody (mAb) endowed with not
only binding but also catalytic activity to metabolize cocaine
would have enhanced therapeutic effects if the kinetic properties
of the mAb were sufficient.
Both spontaneous5 and esterase-catalyzed5c,6 hydrolysis of
cocaine contribute to the short in vivo half-life of ∼30 min in
human blood. Yet, for enzyme or catalytic mAb therapy to be
effective, extensive clearance of cocaine must take place within
seconds. The injection of purified human plasma cholinesterase
reduced cocaine toxicity in mice,7 but the efficiency would not
be suitable to achieve cocaine clearance in the human condition
at clinically manageable concentrations of enzyme. Both our
laboratory and that of Landry et al. used a transition-state (TS)
analogue approach for hapten design and reported the direct
isolation of several cocaine-hydrolyzing mAbs.8,9 In this model,
the benzoyl ester at C-3 of the cocaine framework is replaced
by a phenylphosphonate that approximates the TS for ester
* To whom correspondence should be addressed. E-mail: kdjanda@
scripps.edu.
(1) (a) Cornish, J. W.; O’Brien, C. P. Annu. ReV. Pub. Health 1996, 17, 259-
273. (b) Withers, N. W.; Pulvirenti, L.; Koob, G. F.; Gillin, J. C. J. Clin.
Psychopharmacol. 1995, 15, 63-78. (c) Blaine, J. D.; Ling, W. Psychop-
harmacol. Bull. 1992, 28, 11-14.
(2) (a) Goodkin, K.; Shapshak, P.; Metsch, L.; McCoy, C. B.; Crandall, K.
A.; Kumar, M.; Fujimura, R. K.; McCoy, V.; Zhang, B. T.; Reyblat, S.;
Xin, K.-Q.; Kumar, A. M. J. Neuroimmunol. 1998, 83, 88-101. (b) Isnen,
J. M.; Chokshi, S. K. Annu. ReV. Med. 1991, 42, 133-138. (c) Rich, J. A.;
Singer, D. E. Ann. Emerg. Med. 1991, 20, 616-621. (d) Brody, S. L.;
Slovis, C. M.; Wrenn, K. D. Am. J. Med. 1990, 88, 325-331. (e) Des
Jarlais, D. C.; Friedland, S. R. Science 1989, 245, 578-579. (f) Klonoff,
D. C.; Andres, B. T.; Obana, W. G. Arch. Neurol. 1989, 46, 989-993. (g)
Des Jarlais, D. C.; Friedland, S. R. Science 1989, 245, 578-579.
(3) (a) Carrera, M. R. A.; Ashley, J. A.; Wirsching, P.; Koob, G. F.; Janda, K.
D. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 1988-1992. (b) Carrera, M. R.
A.; Ashley, J. A.; Wirsching, P.; Koob, G. F.; Janda, K. D. Proc. Natl.
Acad. Sci. U.S.A. 2000, 97, 6202-6206. (c) Carrera, M. Rocio A.; Ashley,
J. A.; Parsons, L. H.; Wirsching, P.; Koob, G. F.; Janda, K. D. Nature
1995, 378, 727-730.
(4) Fox, B. S.; Kantak, K. M.; Edwards, M. A.; Black, D. M.; Bollinger, D.
K.; Botka, A. J.; French, T. L.; Thompson, T. L.; Schad, V. C.; Greenstein,
J. L.; Gefter, M. L.; Exley, M. A.; Swain, P. A.; Briner, T. J. Nat. Med.
1996, 2, 1129-1132.
(5) (a) Li, P.; Zhao, K.; Deng, S.; Landry, D. W. HelV. Chim. Acta 1999, 82,
85-89. (b) Garrett, E. R.; Seyda, K. J. Pharm. Sci. 1983, 72, 258-271.
(c) Stewart, D. J.; Inaba, T.; Lucassen, M.; Kalow, W. Clin. Pharmacol.
Ther. 1979, 25, 464-468.
(6) (a) Brzezinski, M. R.; Abraham, T. L.; Stone, C. L.; Dean, R. A.; Bosron,
W. F. Biochem. Pharmacol. 1994, 48, 1747-1755. (b) Boyer, C. S.;
Petersen, D. R. J. Pharmacol. Exp. Ther. 1992, 260, 939-946. (c)
Matsubara, K.; Kagawa, M.; Fukui, Y. Forensic Sci. Int. 1984, 26, 169-
180.
(7) Hoffman, R. S.; Morasco, R.; Goldfrank, L. R. Clin. Toxicol. 1996, 34,
259-266.
(8) Matsushita, M.; Hoffman, T. Z.; Ashley, J. A.; Zhou, B.; Wirsching, P.;
Janda, K. D. Bioorg. Med. Chem. Lett. 2001, 11, 87-90.
9
10.1021/ja012376y CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 3661-3668
3661

A R T I C L E S
Isomura et al.
hydrolysis. However, similar to the enzymatic esterases, none
of the mAbs studied thus far are endowed with the catalytic
power required to meet the demands of hydrolyzing cocaine
rapidly enough to alter its pharmacokinetic profile and psycho-
active effects in humans.
In an effort to isolate more efficient mAbs that degrade
cocaine, Cashman and co-workers initiated a novel program for
high-throughput screening of hybridoma cultures obtained from
immunizations with a TS analogue.10 The key component in
their approach is a benzoylthio ester analogue of cocaine that
liberates a thiol-containing product upon hydrolysis, which then
reacts with 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB, Ellman’s
reagent) to produce a chromophoric analyte.10,11 This method
allows for the rapid and continuous spectrophotometric assay
of many reactions without the need for extraction or derivati-
zation. Although the assay itself proved viable, the mAbs
selected from large panels of hybridomas were not significantly
better catalysts when tested with cocaine than those obtained
directly using TS analogues and standard protocols.8,9 The
problem likely resides in the fact that the benzoylthio ester
substrate utilized by Cashman does not correspond to the natural
isomer of cocaine, but rather to allococaine, which has the
opposite configuration of the benzoyl ester at the C-3 position.
With the aim of developing a high-throughput selection of
efficient cocaine catalytic mAbs from cell cultures and from
phage-display libraries, we herein report the synthesis and
properties of cocaine benzoylthio ester with the correct stereo-
chemistry at C-3 and in optically pure form, as well as its
reactivity with respect to interaction with catalytic mAbs.
Figure 1. Previously reported cocaine analogues containing a nitrogen or
sulfur heteroatom at C-3.
also in achieving the correct stereochemistries at C-2 and C-3,
as well as isolation of the appropriate 2â,3â isomer. For the
installation of a sulfur-bonded functionality, comparable prob-
lems might be expected. A concise strategy was used by Carroll
et al. in synthesizing 3-(phenylthio) analogues 3 by nucleophilic
addition/elimination of sodium thiophenoxide to the enol triflate
2, followed by samarium iodide reduction, which resulted in
the 2â,3â, 2R,3â and 2â,3R isomers.14 Cashman subsequently
employed an amine-catalyzed addition of 4-methoxy-R-toluene-
thiol to anhydroecgonine methyl ester 4 which gave an endo
mode of attack and a mixture of 2â,3R and 2R,3R isomers.11
The isomers were separable, and the free thiol was unmasked
and benzoylated to yield the allo (2â,3R) and allopseudo (2R,3R)
thio ester analogues 5.
To obtain the cocaine benzoylthio ester corresponding to the
natural 2â,3â configuration, we explored several strategies and
found one to be successful. The methodology of Carroll was
first examined by using the sodium salts of 4-methoxy-R-
toluenethiol or thiobenzoic acid for the conjugate addition, as
noted above, to 2 (Scheme 1). Both reactions proceeded
favorably to give 7 and 8 containing sulfur at an unsaturated
C-3 position. We were particularly satisfied with the result in
obtaining 7, since this compound already contained the desired
thio ester group. However, in neither case was the SmI2
reduction successful. The alkylthio enolether functionality in 8
apparently lacked sufficient reactivity, and the acylated coun-
terpart in 7, on the other hand, was too labile with the evidence
suggesting predominant C-S bond cleavage. In an attempt to
make further use of 8, the 4-methoxybenzyl group was cleaved
with trifluoroacetic acid, affording the thio analogue 9 of
2-(carbomethoxy)-3-tropinone 6. It has long been known that
6 could be reduced with sodium amalgam under acidic,
equilibrating conditions to provide ∼28% of ecgonine methyl
ester (2â,3â stereochemistry), together with a similar amount
of the 2R,3â pseudoconfigured isomer.15 The allo isomers, which
Results and Discussion
In recent years, a vast number of cocaine analogues bearing
a substituted benzoyl ring, another modified aryl substituent,
or an alkyl substitutent at C-3 were prepared for structure-
activity studies.12 However, heteroatoms, other than oxygen,
directly attached to the tropane nucleus at C-3 were unknown
until our laboratory introduced a nitrogen atom in the preparation
of a cocaine diamide hapten 1 (Figure 1).13 This unique example
reflected the difficulties not only in forming the C-N bond but
(9) (a) Yang, G. X.-Q.; Chun, J.; Arakawa-Uramoto, H.; Wang, X.; Gawi-
nowicz, M. A.; Zhao, Z.; Landry, D. W. J. Am. Chem. Soc. 1996, 118,
5881-5890. (b) Landry, D. W.; Zhao, K.; Yang, G. X.-Q.; Glickman, M.;
Georgiadis, T. M. Science 1993, 259, 1899-1901.
(10) Cashman, J. R.; Berkman, C. E.; Underiner, G. E. J. Pharmacol. Exp. Ther.
2000, 293, 952-961.
(11) Cashman, J. R.; Berkman, C. E.; Underiner, G.; Kolly, C. A.; Hunter, A.
D. Chem. Res. Toxicol. 1998, 11, 895-901.
(12) (a) Deng, S. X.; Huang, D. W.; Landry, D. W. Tetrahedron Lett. 2001, 42,
6259-6261. (b) Zhu, N.; Harrison, A.; Trudell, M. L.; Klein-Stevens, C.
L. Struct. Chem. 1999, 10, 91-103. (c) Xu, L.; Trudell, M. L. J. Heterocycl.
Chem. 1996, 33, 2037-2039. (d) Neumeyer, J. L.; Tamagnan, G.; Wang,
S.; Gao, Y.; Milius, R. A.; Kula, N. S.; Baldessarini, R. J. J. Med. Chem.
1996, 39, 543-548. (e) Kozikowski, A. P.; Saiah, M. K. E.; Johnson, K.
M., Bergman, J. S. J. Med. Chem. 1995, 38, 3086-3093. (f) Keverline, K.
I.; Abraham, P.; Lewin, A. H.; Carroll, F. I. Tetrahedron Lett. 1995, 36,
3099-3102. (g) Carroll, F. I.; Kotian, P.; Dehghani, A.; Gray, J. L.;
Kuzemko, M. A.; Parham, K. A.; Abraham, P.; Lewin, A. H.; Boja, J. W.;
Kuhar, M. J. J. Med. Chem. 1995, 38, 379-388. (h) Carroll, F. I.;
Mascarella, S. W.; Kuzemko, M. A.; Gao, Y.; Abraham, P.; Lewin, A. H.;
Boja, J. W.; Kuhar, M. J. J. Med. Chem. 1994, 37, 2865-2873. (i) Davies,
H. M. L.; Saikali, E.; Huby, N. J. S.; Gilliartt, V. J.; Matasi, J. J.; Sexton,
T.; Childers, S. R. J. Med. Chem. 1994, 37, 1262-1268. (j) Carroll, F. I.;
Lewin, A. H.; Kuhar, M. J. J. Med. Chem. 1992, 35, 969-981. (k)
Kozikowski, A. P.; Roberti, M.; Xiang, L.; Bergman, J. S.; Callahan, P.
M.; Cunningham, K. A.; Johnson, K. E. J. Med. Chem. 1992, 35, 4764-
4766. (l) Clarke, R. L.; Daum, S. J.; Gambino, A. J.; Aceto, M. D.; Pearl,
J.; Levitt, M.; Cumiskey, W. R.; Bogado, E. F. J. Med. Chem. 1973, 16,
1260-1267.
(14) Thiruvazhi, M.; Abraham, P.; Kuhar, M. J.; Carroll, F. I. Chem. Commun.
1997, 555-556.
(15) (a) Lewin, A. H.; Naseree, T.; Carroll, F. I. J. Heterocycl. Chem. 1987,
24, 19-21. (b) Willsta¨tter, R.; Wolfes, O.; Ma¨der, H. Ann. Chem. 1923,
434, 111-139.
(13) Sakurai, M.; Wirsching, P.; Janda, K. D. Tetrahedron Lett. 1996, 37, 5479-
5482.
9
3662 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002

Cocaine Benzoylthio Ester
A R T I C L E S
Scheme 1. Unsuccessful Strategies toward the Preparation of Cocaine Benzoylthio Estera
a Conditions: (a) sodium bis(trimethylsilylamide), N-phenyltrifluoromethanesulfonamide, THF (90% yield); (b) NaH, PhCOSH, DMF (26% yield); (c)
sodium bis(trimethylsilylamide), p-MeOPhCH2SH (89% yield); (d) SmI2, MeOH, THF; (e) TFA, reflux (95% yield); (f) Na(Hg), H2SO4, pH 3.4, 0 °C.
have the axially disposed C-3 hydroxyl, are not formed.
Unfortunately, reduction of 9, existing as a mixture of tautomers
as in the case of 6, yielded only starting material and an
unidentifiable byproduct. A number of other reducing agents
were also found to be unsuccessful in reactions with 8 and 9,
resulting in either no reaction, decomposition, the incorrect
stereochemistry at C-3, or some combination of these outcomes.
From our earlier work, we were aware that attempted
substitution of an activated hydroxyl group at C-3 of alloecgo-
nine methyl ester by SN2 displacement would result only in
elimination to 4 promoted by rapid epimerization at C-2.13,16
At this point, our only recourse was a thorough reevaluation of
Figure 2. Possible rationale for exo selectivity in the conjugate addition
of magnesium-containing nucleophiles to anhydroecgonine methyl ester.
what was known in the literature concerning the chemistry in
forming tropane derivatives. It occurred to us that although
Cashman11 observed only endo attack of an amine thiolate on
of the 2â,3â isomer 13. Our observation was that the first
4, previous investigations utilizing various aryl Grignard
equivalent of thiolate underwent a rapid transesterification to
reagents afforded exclusively exo addition and a mixture of
give 15 but that the 1,4-conjugate addition of the second
2â,3â and 2R,3â isomers.12h,l In a related context, tertiary amines
equivalent was rather sluggish. Although the yield of 13 on
moderate scales (10 mmol) was only ∼10%, when conducted
were reported to have a favorable effect on the rates of some
Grignard reactions and so perhaps interact with this type of
on a 1-mmol scale ∼30% was produced. Significantly, none of
the allo isomers were detected. Furthermore, 13, 14, and 15
were readily separated by silica gel chromatography and the
derivative 15 could be recycled after each reaction by addition
of the starting thiolate reagent to eventually obtain a ∼60:40
ratio of pure, isolated 13 and 14 in ∼50% overall yield.
Attempted conversion of these compounds to 16 and 18,
respectively, to provide the required C-2 methyl ester initially
proved problematic. The use of sodium methoxide/methanol for
ester exchange resulted only in epimerization at C-2 for 13 and
no reaction in the case of 14. Subsequently, we found that a
one-pot procedure of aqueous Na2CO3/methanol followed by
diazomethane effected a smooth transformation. However, it
was necessary to conduct the hydrolysis reaction of 13 at room
temperature to avoid epimerization, but for 14, reflux was
required to cleave the thio ester and no epimerization occurred.
Finally, the 4-methoxybenzyl group of 16 and 18 was cleaved
with TFA and the free thiol immediately treated with benzoyl
chloride to afford, after purification, the free-base form of (-)-
organometallic compound.17 Hence, we speculated that the
bridgehead nitrogen atom of 4 might participate in chelation
control or, to some degree in just complexation, and direct or
bias the magnesium-containing Grignard species to the exo face
(Figure 2). Even if the interaction were only weak, it could be
enough to overcome the presumably greater steric barrier
imposed by the N-Me bridge compared to the endo ethylene
bridge, as was apparently the case with the aryl Grignard
additions.
With this hypothesis in hand, 4 was prepared18 from (-)-
cocaine hydrochloride (12) and reacted with the bromomagne-
sium salt of 4-methoxy-R-toluenethiol (Scheme 2). A mixture
of three compounds was obtained and, to our delight, evidence
(16) (a) Casale, J. F.; Lewin, A. H.; Bowen, J. P.; Carroll, F. I. J. Org. Chem.
1992, 57, 4906-4912. (b) Findlay, S. P. J. Am. Chem. Soc. 1954, 76, 2855-
2862. (c) Findlay, S. P. J. Am. Chem. Soc. 1953, 75, 4624-4625.
(17) Wotiz, J. H.; Hollingsworth, C. A.; Simon, A. W. J. Org. Chem. 1959, 24,
1202-1205.
(18) Zhang, C.; Lomenzo, S. A.; Ballay, C. J., II.; Trudell, M. L. J. Org. Chem.
1997, 62, 7888-7889.
9
J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002 3663

A R T I C L E S
Isomura et al.
Scheme 2. Synthesis of (-)-Cocaine Benzoylthio Estera
a(a) Conditions: concentrated HCl, reflux, 20 h; (b) p-MeOPhCH2SMgBr (prepared from p-MeOPhCH2SH, EtMgBr, THF); (c) (i) 5% Na2CO3, MeOH,
room temperature, 72 h; (ii) CH2N2, ether, MeOH; (d) (i) TFA, reflux, 3 h; (ii) benzoyl chloride, NEt3, THF; (e) (i) 5% Na2CO3, MeOH, reflux, 72 h; (ii)
CH2N2, ether, MeOH.
cocaine benzoylthio ester 17 corresponding to natural (-)-
cocaine, as well as the 2R,3â-isomer 19 analogous to pseudo-
cocaine (Scheme 2). Surprisingly, contrary to our previous study
of the pseudococaine benzamide analogue,13 refluxing 19 in
water did not result in any epimerization. Notably, the isomer
17 was also stable in this respect. Hence, the C-2 proton in
pseudococaine benzoylthio ester 19 is less labile apparently due
to the lower electronegativity of sulfur and electron-withdrawing
capacity of the benzoylthio group.
Spectral and physical analyses corroborated the structure of
17 and revealed further influences of the sulfur substituent. The
coupling constants between H-2 and H-3, and between H-3 and
the two H-4 protons, are characteristic for each of the four
cocaine structural isomers.19 In particular, the H-3 coupling
pattern is diagnostic and for natural cocaine shows a sharp five-
line signal arising from doublet-doublet-doublet couplings (J
) 12.0, 6.0, 6.0 Hz) indicative of one axial-axial (H-3-H-
4ax) and two axial-equatorial (H-2-H-3, H-3-H-4eq) inter-
actions. For cocaine benzoylthio ester, the central overlap is
slightly spaced due to a small change in coupling values (J )
12.6, 5.6, 5.6 Hz) that suggest an increased puckering in the
tropane ring system which brings H-3 and H-4ax somewhat
closer to the Karplus optimum 180° dihedral angle. As expected
from the lower electronegativity of sulfur compared to oxygen,
there is also a marked upfield chemical shift of H-3 from 5.3
ppm in cocaine to 4.16 ppm in the thio analogue. The effect is
also dramatically evident from 13C NMR and C-H COSY
experiments in which C-3 in the thio compound is found at
34.2 ppm, as opposed to 66.8 ppm in cocaine. However, the
lower electronegativity of sulfur influences the carbonyl group
differently.20 The reduced double-bond character of the benzoyl
group in 17 and increased partial positive charge on the carbonyl
carbon (lower electronic excitation energy of the carbonyl)
causes a large downfield shift from 166.0 to 192.4 ppm. With
regard to physicochemical properties, the presence of sulfur also
imparts enhanced hydrophobicity to the analogue 17. The log
value of the ion-corrected octanol/PBS partition coefficient (P),
determined by the shake-flask method,21 was found to be log P
) 3.47 for 17 compared to log P ) 2.68 for cocaine. Consistent
with experiment, a calculated log P for 17 using 2.68 plus the
average difference in π-hydrophobicity constants for acylthio
versus acyloxy groups (∆π ) 0.74) affords log P ) 3.42.22
Given the structural similarity between the two compounds, the
data suggest that cocaine benzoylthio ester should be more
effective in crossing the blood-brain barrier than cocaine
itself,23 and perhaps also in binding to the dopamine transporter,
and, therefore, could have interesting agonist/antagonist proper-
ties.24
(20) Breitmaier, E.; Voelter, W. Carbon-13 NMR Spectroscopy: High-Resolution
Methods and Applications in Organic Chemistry and Biochemistry; VCH:
New York, 1987.
(21) Leo, A.; Hansch, C.; Elkins, D. Chem. ReV. 1971, 71, 525-616.
(22) Hansch, C.; Leo, A. Substituent Constants for Correlation Analysis in
Chemistry and Biology; John Wiley & Sons: New York, 1979.
(23) (a) Fischer, H.; Gottschlich, R.; Seelig, A. J. Membr. Biol. 1998, 165, 201-
211. (b) Kai, J.; Nakamura, K.; Masuda, T.; Ueda, I.; Fujiwara, H. J. Med.
Chem. 1996, 39, 2621-2624. (c) Hansch, C.; Bjoerkroth, J. P.; Leo, A. J.
Pharm. Sci. 1987, 76, 663-687.
(19) Carroll, F. I.; Coleman, M. L.; Lewin, A. H. J. Org. Chem. 1982, 47, 13-
19.
9
3664 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002

Cocaine Benzoylthio Ester
A R T I C L E S
Table 1. Kinetic Data for GNL Catalytic MAbs That Hydrolyze
Cocaine and 17 as Substrates.a,b
parameter
cat (min-1
GNL3A6
GNL23A6
GNL4D3
k
)
1.37 × 10-2
4.70 × 10-4
53
4.91 × 10-3
5.00 × 10-4
430
1.31 × 10-3
0.750
Km (µM)
42
81
35
157
0.183
k
k
cat/Km (M-1 s-1
cat/kuncat
)
4.35
0.517
2.50 × 10-2
1.36 × 103
36
5.33 × 10-2
4.86 × 102
38
1.54 × 102
1.30 × 102
5.70 × 104
a Determined in 100 mM PB, pH 7.4, 5% DMSO cosolvent, 23 °C.
Values in first row for each parameter correspond to reactions with cocaine
and values in second row to reactions with 17. b Spontaneous pseudo-first-
order k (cocaine) ) 1.01 × 10-5 min-1; k (17) ) 1.32 × 10-5 min-1
Since cleavage of the benzoate ester of cocaine produces the
nonpsychoactive metabolite ecgonine methyl ester,25 the reaction
is an excellent target for an immunopharmacological strategy.
The outstanding features of 17 reside in its cocaine-like
stereochemistry and the ability to use catalysis, rather than
merely binding of a TS analogue, as a selectable trait from large
libraries of antibodies. To provide support for the premise that
catalytic mAbs derived from a selection process using 17 must
then also match with cocaine as a substrate, the converse pairing
was examined. Hence, three of our cocaine-hydrolyzing mAbs
elicited using the TS analogue approach8 were tested for activity
with 17, assaying the formation of benzoic acid by HPLC (Table
1). Although catalysis was not entirely unexpected, the profi-
ciency of one pairing, GNL4D3 and 17, was quite surprising.
The mAb afforded a 57 000-fold initial rate enhancement over
background with an apparent second-order rate constant kcat/
Km > 102 M-1 s-1. Operationally, 5 mol % of GNL4D3
completely degraded an offered sample of 17 in 40 min. The
activity was far greater than that of both GNL3A6 and
GNL23A6.
Spectrophotometric monitoring of the benzoylthio ester
cleavage of 17, when coupled to the disulfide exchange with
Ellman’s reagent,26 was consistent with the differences in
activities observed by HPLC. However, the GNL4D3-catalyzed
reaction revealed “burst kinetics”, whereas GNL3A6 and
GNL23A6 showed only typical, linear initial rates (Figure 3,
Scheme 3). In light of the unique behavior of GNL4D3, we
initially surmised that the burst was indicative of the formation
of an acyl-antibody intermediate, as previously observed for a
catalytic mAb in our laboratory.27 Indeed, the experiments
demonstrate that there is an intial rapid, exponential phase to
liberate 2 equiv of 20 (the mAb has two active sites), which
subsequently react with DTNB and form the equivalent con-
centration of TNB, followed by a much slower steady-state
phase turnover. Yet, closer inspection showed that the apparent
steady-state rates were very similar and not proportional to the
mAb concentration as required for progress curves invoking a
burst-derived covalent complex. Also, curvature of this phase
of the reaction continued with time and approached a shutdown
Figure 3. Visible spectrophotometry of the liberation of 2-nitro-5-
thiobenzoic acid (TNB) during the mAb-catalyzed hydrolysis of 17 in the
presence of Ellman’s reagent.
of the velocity. Furthermore, a subsequent radiolabeling experi-
ment of GNL4D3 using [7-14C-benzoyl]-17 did not provide
evidence for the existence of a labeled mAb. No radioactivity
remained with the antibody after dialysis at low pH designed
to trap the benzoyl intermediate. The data suggested that an
alternative phenomenon, namely, rapid onset of product inhibi-
tion, was responsible for the burst kinetics. In fact, for GNL4D3,
single-turnover inhibition would be the cause of the effect. Since
HPLC analysis showed complete hydrolysis, the degradation
products of 17 were ruled out as inhibitors. On the other hand,
HPLC analysis of the reaction used for spectrophotometry,
which is GNL4D3 and 17 in the presence of DTNB, correlated
with the spectrophotometric result showing the formation of only
a stoichiometric amount of benzoic acid followed by a halt in
reaction progress. Consequently, the disulfide interchange
product 21 must act as a potent inhibitor of the reaction (Scheme
3).
Even though cocaine and 17 have similar spontaneous rates
of hydrolysis (Table 1), the TS mechanism invoked by GNL3A6
and GNL23A6 is better at accommodating cocaine-benzoate
cleavage by hydroxide. Apparently, hapten programming of the
cocaine hydrolytic TS enforces stereoelectronic effects which
are less congruent to the TS for hydrolysis of 17. However, for
GNL4D3, 17 is utilized better than cocaine and very efficiently,
even though this mAb was similarly derived from selection using
a TS analogue. This suggests that possible additional mecha-
nisms, such as general-acid/base catalysis, are invoked at the
mAb active site which favor 17 in the formation or breakdown
of the tetrahedral intermediate.28 In other studies, we observed
that a catalytic mAb designed solely on the basis of TS
stabilization also fortuitously combined more elaborate chemical
mechanisms to provide large rate enhancements.27 The use of
17 as a selection reagent to directly screen antibody libraries
for catalysis will prove invaluable in identifying subsets of mAbs
that operate via TS stabilization or other chemical mechanisms,
(24) (a) Wang, S.; Sakamuri, S.; Enyedy, I. J.; Kozikowski, A. P.; Deschaux,
O.; Bandyopadhyay, B. C.; Tella, S. R.; Zaman, W. A.; Johnson, K. M. J.
Med. Chem. 2000, 43, 351-360. (b) Lin, Z.; Wang, W.; Uhl, G. R. Mol.
Pharmacol. 2000, 58, 1581-1592.
(25) Misra, A. L.; Nayak, P. K.; Bloch, R.; Mule, S. J. J. Pharm. Pharmacol.
1975, 27, 784-786.
(26) (a) Riddles, P. W.; Blakeley, R. L.; Zerner, B. Anal. Biochem. 1979, 94,
75-81. (b) Ellman, G. L. Arch. Biochem. Biophys. 1959, 82, 70-77.
(27) Wirsching, P.; Ashley, J. A.; Benkovic, S. J.; Janda, K. D.; Lerner, R. A.
Science 1991, 252, 680-685.
(28) (a) Castro, E. A. Chem. ReV. 1999, 99, 3505-3524. (b) Jencks, W. P.
Catalysis in Chemistry and Enzymology; Dover Publications: New York,
1987.
9
J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002 3665

A R T I C L E S
Isomura et al.
Scheme 3. Pathway for the Formation of the C-3 Thiolate Hydrolysis Product and Trapping Using Ellman’s Reagent
were washed with brine, dried over Na2SO4, and evaporated to give
crude thioester 15. To a solution of the crude reaction product in THF
(30 mL) was added the bromomagnesium thiolate [prepared as described
above using THF (14 mL), 4-methoxytoluene-R-thiol (2.07 mL, 14.8
mmol), and EtMgBr (15.6 mL, 1.0 M solution in THF, 15.6 mmol)] at
0 °C with stirring, and then the mixture was allowed to stir at room
temperature for 40 h. After this time, the mixture was cooled to -78
°C and TFA (1.44 mL, 18.7 mmol) was added. The mixture was allowed
to warm to room temperature, poured into 1 M HCl, and extracted
with Et2O to remove unreacted thiol. The aqueous layer was brought
to pH ∼9 with NH4OH and extracted several times with CH2Cl2; the
CH2Cl2 layers were washed with brine, dried over Na2SO4, and
evaporated. Analysis by TLC [CHCl3/EtOAc/NH4OH (50:50:1)] showed
three major compounds of Rf ) 0.85, 0.60, 0.40. The residue was
purified by chromatography on silica gel, eluting with CHCl3/EtOAc/
NH4OH (60:40:1), to afford, in order of elution, the desired compound
13 (334 mg, 7.40% yield), the pseudoisomer 14 (205 mg, 4.50% yield),
and the thioester 15 (2.39 g, 79.8% yield) as pale yellow oils. The
thioester was recycled four times until 1.25 g of 13 (27.6% based on
4) and 0.940 g of 14 (20.7% based on 4) were obtained.
including covalent catalysis, some of which should be highly
active for cocaine hydrolysis. Spectrophotometry of large panels
of hybridoma cell cultures and visual colorimetric assessment
of phage-display libraries on agar plates will allow high-
throughput analysis and selection of catalysts. Finally, antibody
engineering using X-ray structure-based methods and mutagen-
esis techniques provides the means to modify critical amino
acid residues with the aim of obtaining a value of kcat/Km
104 M-1 s-1 necessary for therapy.8
∼
A catalytic mAb that metabolizes cocaine with a sufficient
rate would find use in human immunopharmacotherapy both
for rehabilitative purposes and in paramedic or emergency room
situations for drug overdose cases. In addition, cocaine benzoyl-
thio ester might serve as a novel pharmacotherapeutic in itself
by blocking the psychoactive effects of cocaine or as a
nonaddictive cocaine surrogate. A drug that interacts at the
dopamine transporter to prevent cocaine binding, yet does not
effect dopamine reuptake, or has less reinforcing cocaine-like
properties would be beneficial for the treatment of cocaine
addiction. Studies in these directions are underway in our
laboratory.
(1R,2R,3S,5S)-3-[[(4-Methoxyphenyl)methyl]thio]-8-methyl-8-
azabicyclo[3.2.1]octane-2-carbothioic acid S-[(4-methoxyphenyl)-
methyl] ester 13: 1H NMR (CDCl3) δ 7.24-7.18 (m, 4H), 6.85-6.79
(m, 4H), 4.13 (d, J ) 13.5 Hz, 1H), 4.12 (d, J ) 13.5 Hz, 1H), 3.80
(s, 3H), 3.78 (s, 3H), 3.66 (d, J ) 13.5 Hz, 1H), 3.64 (d, J ) 13.5 Hz,
1H), 3.40 (dd, J ) 2.1, 7.1 Hz, 1H), 3.15 (m, 1H), 2.72-2.63 (m, 2H),
2.19 (ddd, J ) 12.3, 2.6, 2.6 Hz, 1H), 2.12 (s, 3H), 2.05-1.94 (m,
2H), 1.55-1.50 (m, 1H), 1.40-1.29 (m, 2H); 13C NMR (CDCl3) δ
196.8, 158.52, 158.49, 130.3, 129.90, 129.87, 129.0, 113.8, 113.7, 65.8,
62.3, 60.4, 55.2, 55.1, 41.6, 38.4, 35.9, 35.8, 32.5, 25.6, 25.0; HRMS
(MALDI-FTMS) calcd for C25H32NO3S2 (MH+) 458.1818, found
Experimental Section
General Methods. 1H and 13C NMR spectra were measured at 400
MHz on a Brucker AMX-400 spectrometer. Chemical sifts (ppm) are
reported relative to internal CDCl3 (1H, 7.26 ppm and 13C, 77.0 ppm).
HRMS spectra were measured using MALDI techniques. Infrared
spectra were measured using neat samples on a Nicolet Avatar 360
FT-IR equipped with a Nicolet Smart Golden Gate (ZnSe) by the ATR
technique. Glassware and solvents were dried by standard methods.
Flash chromatography was performed on silica gel 60 (230-400 mesh)
and thin-layer chromatography (TLC) on glass plates coated with a
0.25-mm layer of silica gel 60 F-254. The (-)-cocaine hydrochloride
was supplied by the National Institute on Drug Abuse (NIDA) and
other chemical reagents and solvents were from Aldrich Chem. Co.
and used without further purification. Nomenclature is in accord with
the Chemical Abstracts Registry Service (CAS).
Introduction of Sulfur onto the Cocaine Framework Using a
Thiolate Reagent: 13, 14, and 15. A pear-shaped flask was charged
with THF (11.3 mL) and 4-methoxytoluene-R-thiol (1.65 mL, 11.9
mmol) and then the Grignard reagent EtMgBr (12.5 mL, 1.0 M solution
in THF, 12.5 mmol) was added at room temperature. After bubbling
subsided, the resulting solution of bromomagnesium thiolate (0.5 M
in THF) was added dropwise via cannula over a period of ∼15 min to
a solution of 4 (1.79 g, 9.89 mmol) in THF (30 mL) at 0 °C with
stirring, and then the mixture was allowed to stir at room temperature
for 14 h. After this time, the mixture was cooled to -78 °C and TFA
(1.16 mL, 15 mmol) was added. The mixture was allowed to warm to
room temperature, poured into 1 M HCl, and extracted with Et2O to
remove unreacted thiol. The aqueous layer was brought to pH ∼9 with
NH4OH and extracted several times with CH2Cl2; the CH2Cl2 layers
458.1809; [R]25 ) -91.0° (c ) 1.042, CHCl3).
D
(1R,2S,3S,5S)-3-[[(4-Methoxyphenyl)methyl]thio]-8-methyl-8-
azabicyclo[3.2.1]octane-2-carbothioic acid S-[(4-methoxyphenyl)-
methyl] ester 14: 1H NMR (CDCl3) δ 7.23-7.17 (m, 4H), 6.84-6.79
(m, 4H), 4.14 (d, J ) 13.8 Hz, 1H), 4.09 (d, J ) 13.8 Hz, 1H), 3.79
(s, 3H), 3.76 (s, 3H), 3.66 (d, J ) 12.6 Hz, 1H), 3.62 (d, J ) 12.6 Hz,
1H), 3.28 (d, J ) 6.5 Hz, 1H), 3.15-3.08 (m, 3H), 2.31 (s, 3H), 1.99-
1.90 (m, 2H), 1.82-1.72 (m, 2H), 1.57-1.52 (m, 1H), 1.45-1.38 (m,
1H); 13C NMR (CDCl3) δ 199.3, 158.7, 158.5, 130.2, 130.0, 129.9,
129.4, 114.0, 113.7, 64.5, 60.7, 59.3, 55.2 (2), 38.3, 37.1, 36.1, 35.1,
32.8, 26.5, 22.9; HRMS (MALDI-FTMS) calcd for C25H32NO3S2 (MH+)
458.1818, found 458.1811.
8-Methyl-8-azabicyclo[3.2.1]oct-2-ene-2-carbothioic acid S-[(4-
methoxyphenyl)methyl] ester 15: 1H NMR (CDCl3) δ 7.25-7.21 (m,
2H), 6.85-6.81 (m, 2H), 6.79 (t, J ) 3.2 Hz, 1H), 4.11 (s, 2H), 3.84
(d, J ) 5.3 Hz, 1H), 3.78 (s, 3H), 3.24 (t, J ) 5.3 Hz, 1H), 2.62 (br d,
J ) 19.4 Hz, 1H), 2.33 (s, 3H), 2.17-2.14 (m, 2H), 1.86-1.80 (m,
2H), 1.52-1.48 (m, 1H); 13C NMR (CDCl3) δ 190.9, 158.7, 141.6,
134.4, 130.0, 129.6, 113.9, 58.4, 57.0, 55.2, 36.0, 34.4, 32.2, 31.4, 30.1;
HRMS (MALDI-FTMS) calcd for C17H22NO2S (MH+) 304.1366, found
304.1362.
9
3666 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002

Cocaine Benzoylthio Ester
A R T I C L E S
On small scales, the protocol used was as follows. A stock solution
of the bromomagesium thiolate was prepared as described above using
4-methoxytoluene-R-thiol (1.46 mL, 10.5 mmol), THF (8.54 mL), and
EtMgBr (10.0 mL, 1.0 M solution in THF, 10.0 mmol), but the addition
was carried out at 0 °C and the mixture stirred at room temperature
for 2 h. The bromomagesium thiolate (4.51 mL, 2.25 mmol) was added
dropwise via syringe over several minutes to a solution of 4 (136 mg,
0.751 mmol) in THF (1.0 mL) at room temperature and the resultant
mixture stirred for 75 h. After this time, the mixture was cooled to
-78 °C and TFA (208 µL, 2.70 mmol) was added. The mixture was
allowed to warm to room temperature, poured into 1 M HCl, and
extracted with Et2O to remove unreacted thiol. The aqueous layer was
brought to pH ∼9 with NH4OH and extracted several times with CH2-
Cl2; the CH2Cl2 layers were washed with brine, dried over Na2SO4,
and evaporated. The residue was purified by chromatography on silica
gel, eluting with CHCl3/EtOAc/NH4OH (60:40:1), to afford, in order
of elution, the desired compound 13 (109 mg, 31.8% yield), the
pseudoisomer 14 (11.1 mg, 3.2% yield), and the thioester 15 (87.0 mg,
38.2%) as pale yellow oils.
(1R,2R,3S,5S)-3-[[(4-Methoxyphenyl)methyl]thio]-8-methyl-8-
azabicyclo[3.2.1]octane-2-carbothioic Acid Methyl Ester 18. A
mixture of 14 (202 mg, 0.442 mmol), MeOH (4 mL), and Na2CO3 (70
mg, 0.656 mmol) was refluxed and stirred for 24 h. Then, H2O (1.5
mL) was added and the mixture was refluxed and stirred for 48 h. After
cooling to room temperature, the mixture was acidified to pH ∼3 with
1 M HCl and then treated with CH2N2. The excess CH2N2 was quenched
with AcOH, and then the mixture was evaporated until about half the
volume remained and was poured into 1 M HCl and washed with Et2O.
The aqueous layer was brought to pH ∼9 with NH4OH and extracted
several times with CH2Cl2; the combined CH2Cl2 layers were washed
with brine, dried over Na2SO4, and evaporated. The residue was purified
by chromatography on silica gel, eluting with CHCl3/EtOAc/NH4OH
(60:40:1) to afford 18 as a colorless oil (71.1 mg, 48.0% yield): 1H
NMR (CDCl3) δ 7.26-7.22 (m, 2H), 6.84-6.81 (m, 2H), 3.78 (s, 3H),
3.76 (d, J ) 12.9 Hz, 1H), 3.69 (d, J ) 12.9 Hz, 1H), 3.69 (s, 3H),
3.25 (dd, J ) 6.5, 2.6 Hz, 1H), 3.08 (m, 1H), 3.01 (dt, J ) 12.0, 5.9
Hz, 1H), 2.88 (dd, J ) 12.0, 2.6 Hz, 1H), 2.31 (s, 3H), 1.99-1.74 (m,
4H), 1.50 (ddd, J ) 13.2, 5.9, 2.9 Hz, 1H), 1.43-1.37 (m, 1H); 13C
NMR (CDCl3) δ 137.5, 158.5, 130.4, 130.0, 113.7, 63.4, 60.5, 55.2,
51.7, 50.6, 38.1, 36.8, 36.3, 35.0, 26.6, 23.1; HRMS (MALDI-FTMS)
(1R,2S,3S,5S)-3-[[(4-Methoxyphenyl)methyl]thio]-8-methyl-8-
azabicyclo[3.2.1]octane-2-carbothioic Acid Methyl Ester 16. A
mixture of 13 (853 mg, 1.87 mmol), MeOH (20 mL), and 5% Na2CO3
(5.76 mL, 2.72 mmol) was stirred at room temperature for 24 h. Then,
additional MeOH (5 mL) and 5% Na2CO3 (2.18 mL, 1.03 mmol) were
added, and the mixture stirred for 48 h. After this time, the mixture
was acidified to pH ∼3 with 1 M HCl and then treated with CH2N2.
The excess CH2N2 was quenched with AcOH, and the mixture was
evaporated until about half the volume remained and was poured into
1 M HCl and washed with Et2O. The aqueous layer was brought to pH
∼9 with NH4OH and extracted several times with CH2Cl2; the combined
CH2Cl2 layers were washed with brine, dried over Na2SO4, and
evaporated. The residue was purified by chromatography on silica gel,
eluting with CHCl3/EtOAc/NH4OH (60:40:1), to afford 16 as a pale
yellow oil (235 mg, 37.6% yield): 1H NMR (CDCl3) δ 7.24-7.20 (m,
2H), 6.86-6.82 (m, 2H), 3.80 (s, 3H), 3.70 (s, 3H), 3.64 (d, J ) 13.5
Hz, 1H), 3.63 (d, J ) 13.5 Hz, 1H), 3.42 (m, 1H), 3.10 (m, 1H), 2.62
(ddd, J ) 12.6, 5.3, 5.3 Hz, 1H), 2.47 (t, J ) 4.1 Hz, 1H), 2.19 (ddd,
J ) 12.6, 2.9, 2.9 Hz, 1H), 2.11 (s, 3H), 2.03-1.93 (m, 2H), 1.49-
1.44 (m, 1H), 1.37-1.29 (m, 2H); 13C NMR (CDCl3) δ 172.1, 158.5,
130.7, 129.9, 113.8, 65.4, 62.7, 55.2, 53.2, 51.4, 41.6, 38.3, 36.4, 35.7,
25.4, 25.0; HRMS (MALDI-FTMS) calcd for C18H26NO3S (MH+)
calcd for C18H26NO3S (MH+) 336.1628, found 336.1624; [R]25
-34.2° (c ) 0.556, CHCl3).
)
D
(1R,2R,3S,5S)-3-(benzoylthio)-8-methyl-8-azabicyclo[3.2.1]octane-
2-carboxylic Acid Methyl Ester (Pseudococaine Benzoylthio Ester)
19. A solution of 18 (66.3 mg, 0.198 mmol) in TFA (3.5 mL) was
refluxed and stirred for 1 h. After cooling, the mixture was coevaporated
with toluene (three times). The residue was dissolved in THF (2 mL)
and Et3N (44.1 µL, 0.317 mmol) and benzoyl chloride (29.9 µL, 0.257
mmol) were added at room temperature and stirred for 2 h. Although
TLC showed the reaction was incomplete, additional Et3N (44.1 µL,
0.317 mmol), benzoyl chloride (29.9 µL, 0.257 mmol), and 1.5 h of
stirring did not provide an improvement. Excess benzoyl chloride was
quenched with MeOH and the mixture evaporated. The residue was
purified twice by chromatography on silica gel, each time eluting with
CHCl3/EtOAc/NH4OH (60:40:1), to afford 19 as a pale yellow oil that
crystallized to an off-white solid upon standing in a cold room at 4 °C
(48.9 mg, 77.5% yield): 1H NMR (CDCl3) δ 7.91-7.88 (m, 2H), 7.54-
7.50 (m, 1H), 7.42-7.38 (m, 2H), 4.12 (dt, J ) 12.0, 7.0 Hz, 1H),
3.63 (s, 3H), 3.39 (dd, J ) 6.2, 2.9 Hz, 1H), 3.18 (m, 1H), 3.08 (dd,
J ) 12.3, 2.9 Hz, 1H), 2.39 (s, 3H), 2.10-1.88 (m, 5H), 1.85-1.77
(m, 1H); 13C NMR (CDCl3) δ 190.6, 172.4, 136.9, 133.3, 128.5, 127.2,
63.7, 60.7, 51.8, 48.2, 38.3, 36.4, 35.8, 26.4, 23.2; HRMS (MALDI-
336.1628, found 336.1627; [R]25 ) -38.2° (c ) 0.654, CHCl3).
D
FTMS) calcd for C17H22NO3S (MH+) 320.1315, found 320.1310; [R]25
(1R,2S,3S,5S)-3-(Benzoylthio)-8-methyl-8-azabicyclo[3.2.1]octane-
2-carboxylic Acid Methyl Ester (Cocaine Benzoylthio Ester) 17. A
solution of 16 (269 mg, 0.803 mmol) in TFA (8 mL) was refluxed and
stirred for 3 h. After cooling, the mixture was coevaporated with toluene
(three times). The residue was dissolved in THF (8 mL) and cooled to
0 °C, and Et3N (447 µL, 3.21 mmol) and benzoyl chloride (237 µL,
2.01 mmol) were added with stirring. After stirring at room temperature
for 2.5 h, MeOH was added to quench the excess benzoyl chloride
and the mixture was evaporated. The residue was purified twice by
chromatography on silica gel, each time eluting with CHCl3/EtOAc/
NH4OH (60:40:1), to afford 17 as a white solid (154 mg, 60.1%
yield): 1H NMR (CDCl3) δ 7.97-7.94 (m, 2H), 7.55-7.51 (m, 1H),
7.43-7.38 (m, 2H), 4.16 (ddd, J ) 12.6, 5.6, 5.6 Hz, 1H) (H-3), 3.74
(s, 3H) (OMe), 3.59 (m, 1H) (H-1), 3.22 (m, 1H) (H-5), 2.75 (t, J )
4.1 Hz, 1H) (H-2), 2.32 (dt, J ) 12.6, 2.9 Hz, 1H) (H-4ax), 2.18 (s,
3H) (NMe), 2.21-2.08, 1.88-1.78 (m, 4H) (H-6, H-7), 1.68-1.63 (m,
1H) (H-4eq); 13C NMR (CDCl3) δ 192.4 (SCOPh), 172.6 (MeOCO),
136.9 (1′-Ph), 133.3 (4′-Ph), 127.2 (2′,6′-Ph), 128.5 (3′,5′-Ph), 65.7
(C-1), 62.9 (C-5), 52.2 (C-2), 51.7 (OMe), 41.7 (NMe), 37.6 (C-4),
34.2 (C-3), 25.3 (C-6), 25.1 (C-7); HRMS (MALDI-FTMS) calcd for
C17H22NO3S (MH+) 320.1315, found 320.1312; [R]25D ) -42.8° (c )
0.694, CHCl3); IR (cm-1) 1717 (MeOCO), 1647 (SCOPh); mp 111-
113 °C.
D
) +27.8° (c ) 0.964, CHCl3); mp 61-68 °C.
Experimental Determination of log P. The log P values for cocaine
and cocaine benzoylthio ester were derived using the same conditions.
Briefly, the extinction coefficients at 254 nm (ꢀ254) for both compounds
in free-base form were measured in octanol [ꢀ254 (cocaine) ) 783 M-1
cm-1; ꢀ254 (17) ) 5960 M-1 cm-1]. A precisely weighed amount of
each compound (∼10 mg proved convenient) was partitioned between
octanol and 10 mM phosphate/150 mM NaCl, pH 7.4 (PBS) using
accurately dispensed volumes of the two solvents in a ratio such that
∼50-75% of the compound was extracted into the upper octanol phase.
Experiments were conducted in screw-capped glass vials or polypro-
pylene tubes and manually inverted for 10 min (∼300 inversions). An
aliquot of the octanol phase was then diluted to 1 mL of octanol in a
quartz cuvette and the absorbance measured at 254 nm. Back calcula-
tions afforded the concentration of compound in the octanol phase,
and subtraction from the known total amount of compound afforded
the concentration in the PBS phase. The empirical P ) [cmpd in
octanol]/[cmpd in PBS] resulted in log P (cocaine) ) 1.49 and log P
(17) ) 2.27. These values were corrected to account for only free-
base partitioning at pH 7.4 by using the known pKa ) 8.6 for cocaine,29
(29) Bailey, D. N.; Bessler, J. B.; Sawrey, B. A. J. Anal. Toxicol 1997, 21,
41-43.
9
J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002 3667

A R T I C L E S
Isomura et al.
and the assumption of the same value for 17, and are noted in the text
above. An uncorrected log P (cocaine) ) 1.05 in octanol/water at pH
7.4, which did not specify conditions, was previously reported.30
Kinetic Experiments. The kinetic parameters for the spontaneous
and mAb-catalyzed hydrolysis of cocaine and 17 were determined in
100 mM sodium phosphate (PB), pH 7.4, 5% DMSO cosolvent at 23
°C. Stock solutions of free-base cocaine and free-base 17 were prepared
in DMSO. The initial rates were determined by following the total
formation of benzoic acid by analytical reversed-phase HPLC (C-18
column, VYDAC 201TP54; isocratic mobile phase of 17% CH3CN,
83% H2O, 0.1% TFA; flow rate, 1.5 mL/min; detector setting, 254 nm;
retention times: benzoic acid 8.00 min, benzoylecgonine 5.90 min,
cocaine 11.70 min, thiobenzoylecgonine 12.70 min, 17 27.0 min). Based
on our previous measurements with no cosolvent,8 the mAb catalysis
of cocaine hydrolysis was adversely effected by the presence of DMSO,
particularly for GNL23A6 and GNL4D3. However, to achieve sufficient
solubility of 17, DMSO was required and so all reactions were
performed under identical conditions to allow comparisons. The
spontaneous and mAb-catalyzed hydrolysis of 17 and trapping of 20
was observed using spectrophotometry. The reactions were performed
in 100 mM PB, pH 7.4, 5% DMSO cosolvent at 23 °C in 1-mL (1 cm)
quartz cuvettes. The rates were monitored by the increase in absorbance
at 412 nm due to the formation of 2-nitro-5-thiobenzoic acid (TNB)
liberated upon reaction of 20 with Ellman’s reagent. Reactions were
done singly in a Shimadzu UV2100U spectrophotometer equipped with
an automatic cell changer, Peltier temperature control, and computer
interface. The absorption maximum (412 nm) and extinction coefficient
(ꢀ412 ) 12 900 M-1 cm-1) of TNB under the reaction conditions was
determined by reduction of a known amount of Ellman’s reagent with
a 20-fold excess of tris(2-carboxyethyl)phosphine (TCEP).31 Reactions
contained a final concentration of 100 µM 17 and 500 µM Ellman’s
reagent aliquoted from stock solutions in DMSO. The mAb-catalyzed
reactions contained various concentrations of GNL4D3 or 5 µM of
either GNL3A6 or GNL23A6. After temperature equilibration of buffer
containing the Ellman’s reagent and in the presence or absence of mAb,
a solution was first auto-zeroed and the reaction initiated by addition
of 17 and then mixing (Mini-Mix; Precision Cells, Inc.). The total time
from addition to the start of data accumulation was ∼7 s. Decomposition
of Ellman’s reagent was ∼0.03%/h within the time frame of the
experiments. The initial rate for the observed spontaneous hydrolysis
of 17 (∼0.7%/h) was subtracted from the mAb-catalyzed reactions.
Preparation of Radiolabeled 17. The procedure was analogous to
the preparation of unlabeled 17. The compound 16 (31 mg, 0.0925
mmol) was deprotected and the residue dissolved in THF (0.25 mL)
and Et3N (51.5 µL, 0.370 mmol). A solution of radiolabeled benzoyl
chloride [prepared in situ from 1 mCi of [7-14C]-benzoic acid at 55
mCi/mmol (American Radiolabeled Chemicals), unlabeled benzoic acid
(24 mg, 0.197 mmol), oxalyl chloride (20 µL, 0.225 mmol), and a
microdrop of DMF in THF (1.5 mL) for 1 h] was then added and the
reaction stirred until TLC showed completion. The crude material was
purified using column chromatography. A stock solution was prepared
in DMSO and scintillation counting used to determine the specific
activity (4.69 mCi/mmol).
Radiolabeling Experiments with mAb GNL4D3. A 500-µL
solution of GNL4D3 (67 µL of a 59.8 µM stock solution to give 4
µM) was prepared in 100 mM PB, pH 7.4 (408 µL), and DMSO (20
µL). An aliquot of radiolabeled 17 (5 µL of a 10 mM stock solution to
give 100 µM) was then added and the solution left at 23 °C for 30
min. After this time, the solution was transferred to a Slide-A-Lyzer
dialysis cassette (10 000 MWCO) (Pierce) and dialyzed against seven
portions of 200 mL of 100 mM ammonium acetate, pH 5.5 at 4 °C,
with changing of the buffer every hour. The radioactivity of the mAb
solution was measured using scintillation counting. Stoichiometric
incorporation would have yielded ∼83 000 counts/min, but <1000
counts/min were observed.
Acknowledgment. This work was supported by funding from
the Skaggs Institute for Chemical Biology, NIDA Grant
DA08590 (K.D.J.), Mr. Robert E. Martini, and the Joseph
Drown Foundation. K.D.J. is an Arthur C. Cope Scholar.
Supporting Information Available: 1H NMR and 13C NMR
spectra for all prepared compounds. This material is available
free of charge via the Internet at http://pubs.acs.org. See any
current masthead page for ordering information and Web access
instructions.
(30) Carmichael, F. J.; Israel, A. Y. J. Pharmacol. Exp. Ther. 1973, 186, 253-
260.
(31) Burns, J. A.; Butler, J. C.; Moran, J.; Whitesides, G. M. J. Org. Chem.
1991, 56, 2648-2650.
JA012376Y
9
3668 J. AM. CHEM. SOC. VOL. 124, NO. 14, 2002