Design and Synthesis of Transition State Analogs for Induction of Hydride Transfer Catalytic Antibodies

Article abstract of DOI:10.1021/jo962179v

Alcohol dehydrogenases and related aldehyde reductase enzymes catalyze the oxidation of alcohols to aldehydes and the simultaneous reduction of a nicotinamide derivative (NAD⁺ or NADP⁺) to the corresponding 1,4-dihydronicotinamide.Herien we report the design and synthesis of a stable transition state analog for this hydride transfer process.Compound 1 is a rigid <3.2.2> bicyclic structire containing 3-piperidone oxime as a mimic for 1,4-dihydronicotinamide.The piperidone is held in the boat conformation corresponding to the transition state by a three-atom lactam bridge between N(1) and C(4).The oxime function mimics the carboxamide group in nicotinamide.The lactam nitrogen serves as an attachment point for the alkyl group of the alcohol substrate , and the amide oxygen atom mimics its hydroxyl group.Compound 1 was prepared in in 10 steps from N-benzylpiperidone, functionalized with substrate and cofactor recognition elements into transition state analogs 2 and 3 and conjugated to carrier proteins for immunization.These novel analogs open the way for the exploration of the dehydrogenase reaction using catalytic antibodies.

Full text of DOI:10.1021/jo962179v

3220
J . Org. Chem. 1997, 62, 3220-3229
Design a n d Syn th esis of Tr a n sition Sta te An a logs for In d u ction of
Hyd r id e Tr a n sfer Ca ta lytic An tibod ies
J osef Schro¨er, Michel Sanner, J ean-Louis Reymond,*,† and Richard A. Lerner*
Departments of Molecular Biology and Chemistry, The Scripps Research Institute,
10666 North Torrey Pines Road, La J olla, California 92037
Received November 21, 1996X
Alcohol dehydrogenases and related aldehyde reductase enzymes catalyze the oxidation of alcohols
to aldehydes and the simultaneous reduction of a nicotinamide derivative (NAD+ or NADP+) to
the corresponding 1,4-dihydronicotinamide. Herein we report the design and synthesis of a stable
transition state analog for this hydride transfer process. Compound 1 is a rigid [3.2.2] bicyclic
structure containing 3-piperidone oxime as a mimic for 1,4-dihydronicotinamide. The piperidone
is held in the boat conformation corresponding to the transition state by a three-atom lactam bridge
between N(1) and C(4). The oxime function mimics the carboxamide group in nicotinamide. The
lactam nitrogen serves as an attachment point for the alkyl group of the alcohol substrate, and the
amide oxygen atom mimics its hydroxyl group. Compound 1 was prepared in 10 steps from
N-benzylpiperidone, functionalized with substrate and cofactor recognition elements into transition
state analogs 2 and 3 and conjugated to carrier proteins for immunization. These novel analogs
open the way for the exploration of the dehydrogenase reaction using catalytic antibodies.
Sch em e 1. Hyd r id e Tr a n sfer betw een
In tr od u ction
1,4-Dih yd r on icotin a m id e a n d a n Ald eh yd e
Stable transition state analogs of chemical reactions
play a central role in enzymology. These analogs are
often tight binding inhibitors of enzymes, and structural
analysis of their complexes with enzymes helps resolve
the site and geometry of catalytic sites and leads to
structural understanding of enzymatic reaction mecha-
nisms.1 Transition state analogs can also be used in an
“active” manner to prepare catalytic antibodies.2 This
approach provides a practical route to novel protein
catalysts and thereby allows one to test in practice
theoretical models for enzymatic catalysis.
Alcohol dehydrogenases are extremely important en-
zymes which catalyze hydride transfer between position
4 of a 1,4-dihydronicotinamide of a cofactor NAD (nico-
tinamide adenine dinucleotide) or NADP (nicotinamide
adenine dinucleotide phosphate) and the carbonyl group
of an aldehyde or ketone (Scheme 1).3 The thermody-
namic balance of the process in water normally favors
alcohol and pyridinium cation but is brought near equi-
librium within the active site of dehydrogenases.4 The
origin of catalysis and the geometry of the transition state
have been the object of intense scientific debate in recent
years. This debate is all the more remarkable given the
fact that NADH was the first enzymatic cofactor discov-
ered nearly 100 years ago.5
Ch a r t 1
necessity to bring together a reactive cofactor and a
substrate in a precise geometry within a hydrophobic
environment. Herein we report the design and synthesis
of a conformationally stable transition state analog 1 and
its functionalization into haptens 2 and 3 for the prepa-
ration of catalytic antibodies (Chart 1). The model
reaction is a hydride transfer between an N-benzyl-1,4-
dihydronicotinamide 4 and an aldehyde (Scheme 1). This
is an ideal reaction for study because the structure of
the transition state has been defined by calculations.
One of the fundamental difficulties in designing a
transition state analog for such a hydride transfer is the
† Present address: Department of Chemistry
University of Bern, 3012 Bern, Switzerland.
& Biochemistry,
X Abstract published in Advance ACS Abstracts, April, 15, 1997.
(1) For example: Haynes, M. R.; Stura, E. A.; Hilvert, D.; Wilson,
I. A. Science 1994, 263, 646.
(2) (a) Schultz, P. G.; Lerner, R. A.; Science 1995, 269, 1835. (b)
Lerner, R. A.; Benkovic, S. J .; Schultz, P. G. Science 1991, 252, 659.
(3) Rossman, M. G.; Liljas, A.; Bra¨nde´n, C.-I.; Banaszak, L. J .
Evolutionary and structural relationship among dehydrogenases. In
The Enzymes; Academic Press, Inc.: New York, 1975; Vol. 11, p 61.
(b) Oppenheimer, N. J .; Handlon, A. L. Mechanism of NAD-dependent
enzymes. In The Enzymes; Academic Press, Inc.: New York, 1992; Vol.
20, p 453.
(4) Nambiar, K. P.; Stauffer, D. M.; Kolodziej, P. A.; Benner, S. A.;
J . Am. Chem. Soc. 1983, 105, 5886.
(5) Harden, A.; Young, W. J . Proc. R. Soc. B, 1906, 77, 405 and 78,
369.
Resu lts a n d Discu ssion
Hundreds of inhibitors of alcohol dehydrogenases and
of the related aldehyde reductases have been discovered
in research programs directed at the inhibition of these
enzymes for treatment of alcohol abuse and acute dia-
S0022-3263(96)02179-2 CCC: $14.00 © 1997 American Chemical Society

Design and Synthesis of Transition State Analogs
J . Org. Chem., Vol. 62, No. 10, 1997 3221
computations.11 A detailed geometrical description of the
transition state was obtained for the hydride transfer
process between methanol and NAD+ within the active
site of several dehydrogenases. The problem has also
been treated independently in gas phase calculations of
the hydride transfer from 1,4-dihydropyridine to form-
aldehyde,12 to a methylene imminium cation13 or to
formaldehyde hydrogen bonded to an imidazole,14 as well
as for the symmetrical transfer between 1,4-dihydroni-
cotinamide and protonated nicotinamide.15 Despite widely
varying starting points and substrates, these studies
agree well as to the overall geometry of the process.
Ch a r t 2. Exa m p les of Alcoh ol Deh yd r ogen a se
In h ibitor s
In the calculated transition state of the enzymatic
reaction, the 1,4-dihydronicotinamide in NAD is present
in a puckered boat conformation with RN ≈ 5° and RC ≈
15° (Scheme 2). These values are slightly smaller in
calculations involving 1,4-dihydropyridine because puck-
ering strongly depends on the presence of an alkyl
substitutent at the pyridine nitrogen. The boat confor-
mation can be reached by deformation of the planar 1,4-
dihydropyridine ring in the ground state at little ener-
getical cost (≈1.8 kcal/mol), and experimental evidence
from Raman spectroscopic studies with stereospecifically
deuterated NADH suggest that this is indeed the case
for this cofactor in the enzyme-bound state.16 According
to the calculation by Bruice et al.,11 this boat conforma-
tion selectively lowers the transition state energy by as
much as 6 kcal/mol, which makes it a key feature to be
taken into account for transition state mimicry.
All the different calculation sources position the sub-
strate carbonyl group approximately 3 Å above the
dihydropyridine ring at the transition state. The results
of Houk for dihydropyridine13,15 are representative: the
breaking C(4)-H bond is elongated to approximately 1.21
Å, and the hydride attacks the carbonyl of the substrate
at an angle of approximately 107° with a bond length of
1.46 Å, the C(4)-H-C angle being 158°. In these calcula-
tions the substrate CdN bond is placed in a symmetrical
syn orientation above the pyridine ring. In the calcula-
tions for dogfish lactate dehydrogenase,11 the orientation
of the carbonyl is determined by a strong hydrogen-
bonding interaction with a histidine residue on the
enzyme. These calculations suggest that general acid-
base catalysis by this residue plays a key role in enabling
the hydride transfer to take place, and probably governs
the orientation of the carbonyl group at the transition
state.
Ring puckering and positioning of the carbonyl define
the position that substrate and cofactor should adopt in
space relative to one another. Together with the require-
ment for general acid-base catalysis for activation of the
carbonyl substrate, these parameters are the most im-
portant to incorporate in a transition state mimic.
Further important parameters include the torsion
angle of the carboxamide substituent at position 3 of the
dihydropyridine, øam. It is found that the amide oxygen
betes mellitus.6 Selected examples include disulfiram,7
sorbinil,8 and ponalrestat (Chart 2).9 Although many of
these inhibitors are very potent, with inhibition constants
in the nanomolar range, none are of the transition state
analog type or even bear any structural relation to the
cofactor and substrate involved in the enzymatic process.
A better structural resemblance is found with pyr-NADH,
a substrate-cofactor construct which undergoes conver-
sion to L-lac-NAD+ under the action of lactate dehydro-
genase, and competitively inhibits this enzyme with
Ki ) 1.35 × 10-7 M.10 However, despite being clearly
related to the actual enzymatic reaction, this substrate
construct is not formally a transition state analog, and
its conformational flexibility makes it an unlikely can-
didate for use as a template for a catalytic site in
immunization.
Geom etr y of th e Tr a n sition Sta te. With the aim
of inducing catalytic antibodies by immunization, we
sought to design an accurate and conformationally stable
transition state analog for the alcohol dehydrogenase
reaction. Crystal structures of alcohol dehydrogenases
as ternary complexes with both NAD+ and a substrate
are available from the protein database. Although these
do not provide a precise picture of the transition state,
the coordinates have been used as a starting point for
(11) (a) Almarsson, O¨ .; Karaman, R.; Bruice, T. C.; J . Am. Chem.
Soc. 1992, 114, 8702. (b) Almarsson, O¨ .; Bruice, T. C.; J . Am. Chem.
Soc. 1993, 115, 2125.
(12) Tapia, O.; Cardenas, R.; Colonna-Cesari F. J . Am. Chem. Soc.
1988, 110, 4046.
(13) Wu, Y.-D.; Houk, K. N. J . Am. Chem. Soc. 1987, 109, 2226.
(14) Wilkie, J .; Williams, I. H. J . Am. Chem. Soc. 1992, 114, 5423.
(15) Wu, Y.-D.; Lai, D. K. W.; Houk, K. N. J . Am. Chem. Soc. 1995,
117, 4100.
(16) (a) Deng, H.; Zheng, J .; Sloan, D.; Burgner, J .; Callender, R.
Biochemistry 1992, 31, 5085. (b) Deng, H.; Burgner, J .; Callender, R.
J . Am. Chem. Soc. 1992, 114, 7997.
(6) Zollner, H. In Handbook of Enzyme Inhibitors; VCH: Weinheim,
Germany, 1993; Part A, pp 33-43.
(7) Kitson, T. M. Biochem. J . 1991, 278, 189 and refs cited therein.
(8) Sarges, R.; Schnur, R. C.; Belletire, J . L.; Peterson, M. J . J . Med.
Chem. 1988, 31, 230.
(9) Mylari, B. L.; Larson, E. R.; Beyer, T. A.; Zembrowski, W. J .;
Aldinger, C. E.; Dee, M. F.; Siegel, T. W.; Singleton, D. H. J . Med.
Chem. 1991, 34, 108.
(10) Kapmeyer, H.; Pfeiderer, G.; Trommer, W. E. Biochemistry
1976, 15, 5024.

3222 J . Org. Chem., Vol. 62, No. 10, 1997
Schro¨er et al.
Sch em e 2. Ca lcu la ted Tr a n sition Sta te Geom etr y
for Hyd r id e Tr a n sfer a n d th e P r op osed An a log
Sch em e 3. Retr osyn th esis of Tr a n sition Sta te
An a log 1
with the lactam carbonyl might provide a spatial ar-
rangement favorable for catalysis.
Replacement of the three olefinic CH groups of the
dihydropyridine group in the transition state by aliphatic
methylenes in the analog was necessary to obtain a
synthetically feasible structure. This replacement was
judged acceptable since the van der Waals volume and
polarity of both groups are quite similar. However,
molecular modeling showed that unless at least one sp2-
center was maintained on the piperidine ring corre-
sponding to the dihydropyridine, the symmetrical boat
conformation corresponding to the transition state would
flip to a strongly tilted twist-boat conformation in the
bicyclic analog (see below, Figure 1B). The required sp2-
center was introduced in the form of an oxime as a
replacement for the carboxamide group at position 3 of
the dihydropyridine. Despite of the fact that the orienta-
is trans to C(2) in enzyme-NAD complexes (øam ) (160°)
and that an out-of-plane orientation lowers the transition
state energy and thereby directs the stereospecificity of
hydride transfer.17 Finally the torsion angle around the
C-N bond between the ribose and the nicotinamide øn
may adopt either an anti (A-specific dehydrogenases,
-170° < øn < -80°) or a syn orientation (B-specific
dehydrogenases, 43° < øn < 86°)18 and is correlated with
the redox potential of the substrates for a number of
dehydrogenases.19
Design of th e Tr a n sition Sta te An a log. We chose
to construct a transition state analog for the hydride
transfer between 4 and an aliphatic aldehyde (Scheme
1). The choice of 4 instead of the natural cofactor NADH
would allow us to distinguish specific antibody catalysis
from contaminant dehydrogenases in antibody cell cul-
tures and was supported by model studies on the redox
properties of N-alkylnicotinamide derivatives showing
that such compounds behave similarly to the naturally
occurring cofactor NAD.20
We envisioned that a rigid analog closely matching the
transition state could be realized by installing a lactam
bridge between position 1 and 4 of 3-piperidone oxime
(Scheme 2). This lactam bridge enforced the reactive
boat conformation on the pyridine ring while simulta-
neously providing an anchor for the aliphatic chain of
an aldehyde substrate at the appropriate site in the form
of the lactam nitrogen. In addition, the oxygen atom of
the carbonyl was properly positioned to serve as a polar
mimic for the carbonyl oxygen of the aldehyde. Quater-
nization of the piperidine nitrogen would install a
permanent positive charge in agreement with the partial
charge of that atom at the transition state. This positive
charge would also induce the key general acid-base
residue (BH in Scheme 2) in the antibody upon im-
munization.21 Hydrogen-bonding of this catalytic residue
tion of the carboxamide at the transition state (øam
)
(160°, see above) was not programmed by this replace-
ment, the oxime would induce a polar, hydrogen-bonding
environment capable of accommodating the carboxamide
substituent of the dihydropyridine cofactor. This led to
the bicyclic lactam 1 (Chart 1) as the primary synthetic
target.
Derivatives 2 and 3 were planned as haptens for
conjugation to carrier proteins and immunization, to
generate catalytic antibodies for the reduction of simple
aliphatic aldehydes by the 1,4-dihydropyridine 4 (Scheme
1). The aldehyde carbon in the transition state was
mimicked by a methylene group in hapten 2. Alterna-
tively, this carbon was mimicked by a sulfonyl group in
hapten 3, thereby providing two additional oxygen atoms
as mimics for the carbonyl oxygen in orientations com-
patible with the overall geometry of the process. These
might enable additional H-bonding residues to be induced
in the antibodies during the immune response.
Syn th esis. In the retrosynthetic analysis of target
lactam 1, the oxime function would derive from secondary
alcohol 5, which could be obtained by cyclization from
piperidine precursor 6 (Scheme 3).
cis-aminohydroxypiperidine 7 was prepared from N-
benzylpiperidone following published procedures (Scheme
4).22 BOC protection of the primary amine followed by
debenzylation and alkylation of the piperidine nitrogen
with ethyl bromoacetate gave 8. Removal of the BOC
group under standard conditions (CF3COOH in CH2Cl2
or TMSCl-phenol in CH2Cl2)23 gave very low yield,
presumably due to formation of an urethane with the
(17) (a) Donkersloot, M. C. A.; Buck, H. M. J . Am. Chem. Soc. 1981,
103, 6554. (b) de Kok, P. M. T.; Bastiaansen, L. A. M.; van Lier, P. M.;
Vekemans, J . A. J .M.; Buck, H. M. J . Org. Chem. 1989, 54, 1313.
(18) Wu, Y.-D.; Houk, K. N. J . Am. Chem. Soc. 1991, 113, 2353.
(19) Benner, S. A.; Nambiar, K. P.; Chambers, G. K. J . Am. Chem.
Soc. 1985, 107, 5513.
(20) (a) Mauzerall, D.; Westheimer, F. H. J . Am. Chem. Soc. 1955,
77, 2261. (b) Ohnishi, Y.; Tanimoto, S. Tetrahedron Lett. 1977, 22, 1909.
(c) J ohnston, C. C.; Gardner, J . L.; Suelter, C. H.; Metzler, D. E.
Biochemistry 1963, 2, 689. (d) Kim, C. S. Y.; Chaykin, S. Biochemistry
1968, 7, 2339. (e) van Eikeren, P.; Grier, D. L.; Eliason, J . J . Am. Chem.
Soc. 1979, 101, 7402. (f) Ohno, A.; Ushida, S.; Oka, S. Bull. Chem.
Soc. J pn. 1984, 57, 506. (g) Ohnishi, Y. Tetrahedron Lett. 1977, 2109.
(21) Koch, T.; Reymond, J .-L.; :Lerner, R. A.; J . Am. Chem. Soc.
1995, 117, 9383 and refs cited therein.
(22) Ghosh, A. K.; McKee, S. P.; Sanders, W. M. Tetrahedron Lett.
1991, 32, 711.

Design and Synthesis of Transition State Analogs
J . Org. Chem., Vol. 62, No. 10, 1997 3223
Sch em e 4. Syn th esis of Am in o Alcoh ol 6
Sch em e 6. Syn th esis of Mod el Com p ou n d s 17a /b
of the bicyclic intermediate 11. Alcohol 5 was then
oxidized with N-methylmorpholine oxide (NMO) and
catalytic tetrapropylammonium perruthenate24 (TPAP)
in DMF/acetonitrile to give ketone 12. This ketone was
highly sensitive to hydrolysis in contact with silica gel
and was directly converted to the stable oxime 1 by
treatment with excess hydroxylamine in pyridine.25
Sch em e 5. Syn th esis of Com p ou n d 1
Functionalization of compound 1 was first explored
starting from amide 13, easily prepared by Beckman
rearrangement of quinuclidone oxime (Scheme 6).26 Alkyl-
ation of the bridge amide nitrogen with propanesulfonyl
chloride or pentanesulfonyl chloride yielded sultams 14a
and 14b, which were alkylated with iodide 15 to am-
monium salts 16a and 16b. Protection of the carboxyl
function in 15 was required due to polymerization of the
free acid during the reaction. Cleavage of the allyl ester
under neutral and nonreducing conditions was achieved
using potassium 2-ethyl hexanoate and palladium ac-
etate.27 The resulting free acids 17a and 17b were finally
converted to the corresponding N-hydroxysuccinimide
esters 18a and 18b with EDC (3-(N-ethylN-diethylami-
no)propylcarbodiimide) in aqueous DMF.
Compound 1 was similarly functionalized to haptens
2 and 3 (Scheme 7). Despite protection of the oxime,
alkylation with iodohexane gave the double alkylated
derivative 19. Fortunately, quaternization with iodide
15 led to quaternary ammonium salt 20, in which the
oxime hexyl substituent had been removed, presumably
by nucleophilic substitution with iodide in analogy to
known ether and ester cleavage procedures.28 Cleavage
of the allyl ester as above gave 2, which was then
converted to an activated N-hydroxysuccinimide ester 21
and conjugated to carrier proteins KLH (keyhole limpet
hemocyanin) and BSA (bovine serum albumin) for im-
munizations.
adjacent hydroxyl group. Quantitative conversion of 8
to 6 was accomplished by hydrolysis in aqueous HCl to
the corresponding amino acid 9, followed by reesterifac-
tion in methanolic HCl.
Treatment of 6 in refluxing anhydrous methanol with
catalytic potassium cyanide gave the bicyclic alcohol 5,
which was isolated in 63% yield (Scheme 5). Dimer 10
(mixture of stereoisomers), amino acid 9, and starting
material were obtained as side products and recycled to
6 by acidic hydrolysis and reesterification. Attempts to
cyclize the trans-amino alcohol corresponding to 6 gave
only small amounts of dimer, suggesting that the in-
tramolecular cyclization to 5 was facilitated by formation
(24) Ley, S. V.; Norman, J .; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639.
(25) Sternbach, L. H.; Kaiser, S. J . Am. Chem. Soc. 1952, 74, 2215.
(26) Rubsov, M. V.; Mikhlina, E. E.; Vorob'eva, V. Y.; Yanina, A. D.
J . Gen. Chem. 1964, 34, 2232.
(23) Kaiser, E., Sr.; Tam, J . P.; Kubiak, T. M.; Merrifield, R. B.
Tetrahedron Lett. 1988, 29, 303.
(27) J effrey, P. D.; McCombie, S. W. J . Org. Chem. 1982, 47, 587.
(28) Bhatt, M. V.; El-Morey, S. S. Synthesis 1982, 1048.

3224 J . Org. Chem., Vol. 62, No. 10, 1997
Schro¨er et al.
Ch a r t 3. Str u ctu r e of Mod els A, B, a n d C# a
Sch em e 7. F u n ction a liza tion of 1 to Ha p ten 2
a
Substituents added or modified for modeling purposes are in
italics.
possible to obtain the activated ester 25 in moderate yield
and prepare conjugates with carrier protein KLH and
BSA as above.
Sch em e 8. F u n ction a liza tion of 1 to Ha p ten 3
Str u ctu r a l Com p a r ison s betw een Tr a n sition
Sta te a n d Its An a logs
Suitable crystals could be obtained for compounds 1
and 17b and their structure determined by X-ray crystal-
lography. As model for the corresponding transition state
we selected Houk’s transition state for hydride transfer
between 1,4-dihydropyridine and the methylene im-
minium cation.13
For comparison purposes the substitution patterns of
both analogs and the transition state were unified. As
common replacement for either the ribosyl or the benzyl
substituent, a methyl substituent was attached to the
bridgehead nitrogen in 1 and 17b and to the dihydropy-
ridine N(1) nitrogen in the calculated transition state. A
methyl group was also attached to the bridge lactam
nitrogen in 1 and 17b as a match for the carbon of the
methylene imminium cation in the transition state,
leading to model structure A from 1 and B from 17b
(Chart 3). Finally a carboxamide substituent was at-
tached to C(3) of the dihydropyridine in the transition
state, with the oxygen trans to C(2) (øam ) 160°), as seen
in dehydrogenase-NAD complexes, leading to transition
state model C#. These modifications were performed
using the modeling package “Insight” from Biosym.
An atom-by-atom superimposition of model A (from 1)
and the transition state C# is shown in Figure 1A and
illustrates the close similarities between the two struc-
tures.29 The boat conformation of the dihydropyridine
mimic in model A is too strongly puckered compared to
the transition state. However the bicyclic arrangement
allows one to place the substrate carbonyl mimic, repre-
sented by the N-methyl substituent and the oxygen of
the bridge lactam, in the correct position above the
pyridine ring. In addition, the oxime substituent at C(3)
has an excellent overlap with the carboxamide substitu-
ent in the transition state.
Sulfonylation of silylated 1 using propanesulfonyl
chloride gave sultam 22 without loss of the oxime silyl
group (Scheme 8). Reaction with iodide 15 as above
followed by deprotection of the silyl group with aqueous
acid gave the quaternary ammonium salt 23. Deprotec-
tion of the allyl group using triethylammonium 2-ethyl-
hexanoate under palladium catalysis gave hapten 3.
Interestingly, an attempted deprotection using the more
basic potassium 2-ethylhexanoate and palladium gave
the ring-opened product 24, resulting from elimination
of the sultam group. Hapten 3 and its derivatives were
highly sensitive to aqueous basic conditions due to
hydrolysis of the bridge sultam. Nevertheless it was
This oxime substituent at C(3) also plays a further role
in controlling conformation. The piperidine ring in A
adopts a symmetrical boat conformation mimicking the
(29) Upson, C.; Faulhaber, T.; Kamins, D.; Laidlaw, D.; Schlegel,
D.; Vroom, J .; Gurwitz, R.; vanDam, A. IEEE Comput. Graphics Appl.
1989, 9, 30-42.

Design and Synthesis of Transition State Analogs
J . Org. Chem., Vol. 62, No. 10, 1997 3225
F igu r e 1. (A, Top left) superimposition of model A (green), derived from the X-ray coordinates of 1, and model C# (white), derived
from the calculated transition state for hydride transfer. All atoms in the transition state C#, except the hydrogen atom and the
oxygen atom of the carboxamide substituent at C(3), were used in the superimposition. The pictures in parts A-C were rendered
on a Dec Alpha with a Denali graphics board using the AVS graphics software. (B, Top right) superimposition of models A (white)
and B (green), derived from the X-ray coordinates of 1 and 17b. Atoms N(1), C(3), and C(4) were used in the superimposition. The
sp2 hybridization of C(3) allows for a symmetrical boat conformation in A which fits the transition state C# (see part A). The
simpler model B with an sp3 hybridization at C(3) adopts a twist-boat conformation which also brings the three-atom bridge out
of symmetry. (C, Bottom) solvent-excluded surfaces of models A (left, derived from the X-ray coordinates of 1), B (right, derived
from the X-ray coordinates of 17b), and C# (center, derived from the calculated transition state for hydride transfer) using the
program MSMS.29 The surfaces are colored according to atom type: carbon or hydrogen in white, oxygen in red, nitrogen in blue.
Model A mimicks nicely the shape and polarity of transition state C#.
dihydronicotinamide in C#. As shown by a superimposi-
tion of models A and B (Figure 1B), the piperidine ring
in B, which lacks this substituent, adopts a strongly tilted
twist-boat conformation. This tilted conformation in B
also displaces the three-atom lactam bridge out of sym-
metry, leading to further dissimilarity between B and
C#.
The solvent-excluded surface of each model is shown
in Figure 1C. The surfaces were calculated using the
program MSMS.30 The surface are colored according to
atom type to distinguish nonpolar surfaces induced by
hydrocarbon parts of the structures (white) from polar
surfaces with hydrogen-bonding capacity with oxygen
(red) or nitrogen (blue). Such distinction is meaningful
with respect to the planned interactions with an antibody-
binding pocket in an aqueous environment, where hy-
drophobic and hydrogen-bonding interactions are key
determinants of the binding energy.31 While model B
(right in Figure 1C) is clearly not a suitable analog of C#
(center in Figure 1C), the molecular surfaces of A (left
in Figure 1C) and C# are quite similar. Here again the
oxime substituent at C(3) plays a key role in providing
model A with an overall shape similar to that of transi-
tion state C#. By its volume and polarity, this oxime is
an excellent match for the carboxamide substituent at
C(3) of dihydronicotinamide.
(31) We have not used electrostatic potential comparisons between
analog and transition state because several examples in the literature
(the “bait and switch” strategy, where a positively charged transition
state analog mimics a negatively charged transition state, cited in ref
2b) are in clear contradiction with the notion that close electrostatic
analogy is a condition for inducing catalytic antibodies.
(30) Sanner, M.; Olson, A.; Spehner, J . C. Biopolymers 1996, 38,
305.

3226 J . Org. Chem., Vol. 62, No. 10, 1997
Schro¨er et al.
1.70-1.56 (m, 2H). 13C-NMR (125 MHz, CDCl3): 170.9, 67.7,
58.3, 57.1, 51.3, 50.7, 28.8. HRMS: C8H16N2O3 (M + H+) calcd
189.1239, found 189.1241. IR (CDCl3): 3373, 2952, 2252, 1739,
1653, 1203.
Con clu sion
The design and synthesis of an accurate and confor-
mationally stable transition state analog for the alcohol
dehydrogenase reaction have been reported. This transi-
tion state analog possesses a rigid bicyclic structure that
presents all crucial elements of the transition state for
hydride transfer in the appropriate geometry. Compound
1 was readily prepared in 10 steps from N-benzylpiperi-
done and functionalized to haptens 2 and 3 and their
carrier protein conjugates. These conformationally rigid
analogs present an excellent structural fit to the transi-
tion state for hydride transfer both in terms of shape and
polarity. The production of monoclonal antibodies against
these haptens and their characterization for catalysis will
be reported separately.
1,4-Dia za -6-h yd r oxybicyclo[3.2.2]n on a n -3-on e (5).
A
solution of amino alcohol 8 (0.497 g, 2.64 mmol) and sodium
cyanide (0.0135 g, 0.276 mmol) in anhydrous methanol (50 mL)
over molecular sieve (3 Å, 3 g) was refluxed for 3 h. The
molecular sieve was filtered off and the methanol evaporated
in vacuo. The residue was dissolved in a small amount of
methanol and rapidly filtered through
a silical gel plug
(CH2Cl2/MeOH/aqueous NH3 7:1:0.2; TLC 3:1:0.3, Rf ) 0.25)
to give 5 as a colorless crystalline product. Mp: 237-238 °C
(subl) (0.26 g, 63%). Small amounts of 6 were removed by
washing the solid extensively with ethyl acetate. Dimer 10
(Rf ) 0.42) and amino acid 9 (CH2Cl2/MeOH/aqueous NH3 3:2:
1, Rf ) 0.4) were obtained as side products. These were
recycled to starting material 6 by hydrolysis in 3 N hydro-
chloric acid followed by esterification with methanol. 1H-NMR
(300 MHz, CD3OD): 4.00 (td, 1H, J ) 4.5, 8); 3.63 (br s, 2H);
3.36 (m, 1H); 3.30 (m, 1H); 3.03-2.86 (m, 2H); 2.71 (ddd, 1H,
J ) 14.6, 4.4, 2.1); 2.22 (dddd, 1H, J ) 14.5, 8.9, 6.3, 2.5); 1.92
(dddd, 1H, J ) 19.3, 5.2, 3.8, 1.0).
Exp er im en ta l Section
Reagents were purchased from Aldrich or Fluka. Solvents
were A.C.S. grade from Fisher. All chromatographies (flash)
were performed with Merck Silica gel 60 (0.040-0.063 mm).
Preparative HPLC was done with Fisher Optima grade ac-
etonitrile and ordinary deionized water using a Waters prepak
cartridge (500 g) installed on a Waters Prep LC 4000 system
from Millipore, (flow rate 100 mL/min, gradient +0.5%/min
CH3CN, detection by UV at 254 nm). TLC analyses were
performed with fluorescent F254 glass plates. MS, HRMS, and
combustion analyses were provided by the Scripps Research
Institute facility (Gary Szuidak). X-ray crystallography was
performed by Dr. Raj K. Chadha at Scripps.
13C-NMR (125 MHz, CD3OD): 180.7, 72.0, 64.8, 58.4, 52.6,
48.1, 31.8. HRMS: C7H12N2O2 (M + H+) calcd 157.0977, found
157.0979. IR (KBr): 3397, 2907, 2861, 2461, 2265, 1597, 1466,
1351, 1087, 1052. Anal. Calcd: C, 53.83; H, 7.74; N, 17.94.
Found C, 53.74; H, 7.88; N, 17.86.
Dimer 10: 1H-NMR (300 MHz, CDCl3): 7.74, 7.58 (2 d, 1H,
J ) 9); 3.95-3.80 (m, 1H); 3.76 (br s, 2H); 3.70 (s, 3H); 3.20
(m, 2H); 3.15-2.60 (m, 11 H); 2.55 (m, 1H); 2.40-2.15 (m, 3H);
1.85-1.62 (m, 4H). MS (electrospray): (M + H+) 345, (M +
Na+) 367.
Amino acid 9: 1H-NMR (300 MHz, D2O): 3.80 (br s, 1H);
3.08 (dd, 1H, J ) 16.7, 11.7, 8.5); 2.98 (s, 2H); 2.91 (br d, 1H,
J ) 13); 2.78 (br d, 1H, J ) 12); 2.43 (dd, 1H, J ) 13, 1); 2.27
(td, 1H, J ) 11.7, 3.1); 1.78 (ddd, 1H, J ) 13.3, 11.4, 4.1); 1.67
(ddd, 1H, J ) 13, 9, 4). MS (FAB): (M + H+) 175.
ter t-Bu tyl (3′RS,4′SR)-N-(1′-((Eth oxyca r bon yl)m eth yl)-
3′-h yd r oxyp ip er a d in -4′-yl)ca r ba m a te (8). A solution of
amino alcohol 7 (6.14 g, 29.8 mmol) in ethyl acetate/water
(each 80 mL) containing sodium hydrogen carbonate (7.50g,
89.3 mmol) and di-tert-butyl dicarbonate (10.8g, 49.6 mmol)
was stirred vigorously for 15 h at 25 °C. The phases were
seperated, and the product obtained upon evaporation of the
organic phase was purified by flash chromatography (hexane/
acetone 3:1, Rf ) 0.25) to yield the N-BOC derivative of 7
(8.37g, 92%). This compound (8.37 g, 27.3 mmol) was dissolved
in 96% ethanol (200 mL) and vigorously stirred with a catalytic
amount of palladium hydroxide (10% on charcoal) under 1 atm
of hydrogen for 16 h at 25 °C. The catalyst was filtered off
over Celite, and the filtrate was concentrated to 100 mL and
refluxed for 16 h in the presence of ethyl bromoacetate (3.6
mL, 5.4 g, 32 mmol) and sodium hydrogen carbonate (3.467 g,
41.27 mmol). Solvent removal in vacuo, aqueous workup
(aqueous NaCl/CH2Cl2), and flash chromatography (hexane/
acetone 3:1, Rf ) 0.20) gave 8 as colorless solid. Mp: 77-78
°C (7.19 g, 87%).
1H-NMR (300 MHz, CDCl3): 5.00 (d, 1H, J ) 8.8); 4.13 (q,
2H, J ) 7.2); 3.74 (br s, 1H); 3.48 (m, 1H); 3.21 (2, 2H); 2.92
(ddd, 1H, J ) 11.7, 3.6, 2.5); 2.78 (dq, 1H, J ) 9.6, 2.5); 2.50
(d, 1H, J ) 11.7); 2.35 (td, 1H, J ) 11.7, 3); 1.77 (m, 1H); 1.67
(qd, 1H, J ) 11.7, 4.5); 1.40 (s, 9H); 1.24 (t, 3H, J ) 7.2).
HRMS: C14H26N2O5 (M + H+) calcd 303.1920, found 303.1925.
IR(CCl4): 3446, 2979, 1743, 1713, 1495, 1178. Anal. Calcd:
C, 55.61; H, 8.67; N, 9.26. Found: C, 55.55; H, 8.81; N, 9.08.
(3RS,4SR)-N-((Met h oxyca r b on yl)m et h yl)-4-a m in o-3-
h yd r oxyp ip er id in e (6). A solution of 8 (7.65 g, 25.3 mmol)
in 200 mL of 3 N aqueous HCl was heated to 80 °C for 16 h.
The hydrochloric acid was removed in vacuo, and the residue
was taken up in absolute methanol (200 mL). After addition
of 3 Å molecular sieve (6 g), the solution was saturated with
hydrogen chloride gas and refluxed for 5 h. The solvent was
removed in vacuo and the white solid residue dissolved in
water (150 mL). The solution was adjusted with aqueous
sodium carbonate to pH 8. Extraction with ethyl acetate and
1,4-Dia za -6-(h yd r oxylim in o)b icyclo[3.2.2]n on a n -3-
on e (1). Tetrapropylammonium perruthenate (0.33 g, 0.95
mmol) was added to a suspension of 5 (1.0 g, 6.4 mmol),
N-methylmorpholine N-oxide (1.22 g, 10.4 mmol), and pow-
dered molecular sieve (4 Å, 6.5 g) in acetonitrile/N,N-dimeth-
ylformamide 1:1 (30 mL). This mixture was stirred at 25 °C
for 4 h. The acetonitrile was evaporated and the residue was
rapidly passed through a plug of silica gel (CH2Cl2/MeOH 4:1,
TLC CH2Cl2/MeOH/aqueous NH3 4:1:0.3, Rf ) 0.4). The
product-containing fractions were evaporated and dried for
several minutes in vacuo. The sensitive colorless crystalline
ketone 12 was then dissolved in pyridine/ethanol 3:2 (25 mL)
and stirred at 25 °C for 16 h with molecular sieves (4 Å, 2 g)
and hydroxylamine hydrochloride (0.92 g, 13.2 mmol). The
molecular sieves were filtered off, and the solvents were
evaporated in vacuo. The residue was purified by chromatog-
raphy (methylene chloride/methanol/aqueous NH3 7:1:0.15;
TLC, CH2Cl2/MeOH/aqueous NH3 4:1:0.25, Rf ) 0.40) to give
1 (0.49 g, 45%) as a crystalline solid. Mp: 259-260 °C (dec.).
1H-NMR (300 MHz, D2O): 3.82 (dd, 1H, J ) 5, 2); 3.80, 3.70
(2d, 2 × 1H, J ) 19); 3.62, 3.52 (2d, 2 × 1H, J ) 19); 3.05-
2.80 (m, 2H); 2.32 (m, 1H); 2.00 (dddd, 1H, J ) 20, 15, 10, 5).
13C-NMR (125 MHz, D2O): 179.1, 161.6, 62.0, 50.9, 49.2, 47.1,
31.7. MS (FAB+): (M + H+) 170. IR (KBr): 3422, 2345, 1654,
1636, 1324, 1137, 931. Anal. C7H11N3O2 Calcd: C, 49.70; H,
6.55; N, 24.83. Found C, 49.47; H, 6.83; N, 25.11.
1,4-Dia za -4-(p r op ylsu lfon yl)b icyclo[3.2.2]n on a n -3-
on e (14a ) a n d 1,4-Dia za -4-(p en tylsu lfon yl)bicyclo[3.2.2]-
n on a n -3-on e (14b). Sodium hydride (60% in mineral oil, 0.12
g, 3.0 mmol) was added to a suspension of 13 (0.20 g, 1.43
mmol) in tetrahydrofuran (20 mL). After 1 h at 25 °C
1-propanesulfonyl chloride (0.60 mL, 5.3 mmol) was added
slowly. The reaction was quenched after 2 h with acetic acid
(0.2 mL). The solvent was removed in vacuo, and the residue
was purified by chromatography (CH2Cl2/MeOH/aqueous NH3
10:1:0.1, Rf ) 0.40) to give sultam 14a , a colorless solid. Mp:
58-59 °C (0.208 g, 59%). 1H-NMR (500 MHz, CDCl3): 4.57
(m, 1H); 3.72 (s, 2H); 3.44 (m, 2H); 3.02 (m, 2H); 2.92 (m, 2H);
chromatography (CH2Cl2/MeOH/aqueous NH3 5:1:0.2, Rf
)
0.40) gave 6 as colorless liquid (4.63 g, 97%). 1H-NMR (500
MHz, CDCl3): 3.69 (br s, 1H); 3.59 (s, 3H); 3.56 (br s, 1H);
3.18, 3.14 (2d, 2 × 1H, J ) 17); 2.80 (br d, 1H, J ) 12); 2.74-
2.65 (m, 1H); 2.41 (br d, 1H, J ) 12); 2.26 (td, 1H, J ) 11, 3);

Design and Synthesis of Transition State Analogs
J . Org. Chem., Vol. 62, No. 10, 1997 3227
2.06-1.92 (m, 4H); 1.70 (m, 2H); 0.98 (t, 3H, J ) 7.5). 13C-
NMR (125 MHz, CDCl3): 176.9, 62.7, 56.0, 47.7, 45.2, 28.5,
16.9, 12.5. HRMS: C10H18N2O3S (M + Cs+) calcd 379.0092,
found 379.0096. IR (CCl4): 2938, 2878, 1681, 1355, 1249,
1150. Anal. Calcd: C, 48.76; H, 7.37; N, 11.37; S, 13.02.
Found C, 48.62; H, 7.65; N, 11.48; S, 12.77.
Ta ble 1. HP LC Con d ition s for Qu a ter n a r y Am m on iu m
Der iva tivesa
comp
functionalityb
% A
% B
tR, min
16a
16b
17a
17b
18a
18b
20
2
21
23
24
CO2allyl
CO2allyl
COOH
40
30
60
40
60
40
30
30
30
60
80
60
60
60
70
40
60
40
60
70
70
70
40
20
40
40
5.4
8.2
4.1
5.4
6.6
8.2
9.8
3.9
5.6
5.2
8.7
3.6
5.4
A similar procedure using pentanesulfonyl chloride gave
sultam 14b (Rf ) 0.40), a waxy solid. Mp: 23-24 °C (0.302 g,
77%). 1H-NMR (500 MHz, CDCl3): 4.62 (m, 1H); 3.77 (s, 2H);
3.50 (m, 2H); 3.06 (m, 2H); 2.97 (m, 2H); 2.07 (m, 2H); 2.00
(m, 2H); 1.69 (m, 2H); 1.40-1.28 (m, 4H); 0.86 (t, 3H, J ) 7.5).
13C-NMR (125 MHz, CDCl3): 177.0, 62.8, 54.4, 47.7, 45.2, 29.8,
28.6, 22.7, 21.9, 13.5. HRMS: C12H22N2O3S (M + Cs+) calcd
407.0405, found 407.0409. IR (CCl4): 2957, 2875, 1683, 1354,
1249, 1150. Anal. Calcd: C, 52.53; H, 8.08; N, 10.21; S, 11.68.
Found C, 52.24; H, 8.48; N, 10.07; S, 12.20.
COOH
CO(NOS)c
CO(NOS)
CO2allyl
COOH
CO(NOS)
CO2allyl
COOH
COOH
CO(NOS)
3
25
a
N-(2′-((Allyloxy)ca r b on yl)et h yl)-4-(iod om et h yl)b en z-
a m id e (15). A suspension of â-alanine (1.50 g, 16.9 mmol) in
allyl alcohol (12 mL) was saturated with gaseous hydrogen
chloride and stirred for 1 h at 25 °C. Removal of the solvent
in vacuo yielded â-alanine allyl ester hydrochloride (2.77 g,
99%) as a colorless solid. A suspension of 4-(chloromethyl)-
benzoic acid (3.62 g, 21.2 mmol) in 100 mL of CH2Cl2 and N,N-
dimethylformamide (10 drops) was treated with oxalyl chloride
(4.0 mL, 47 mmol) at 25 °C until complete dissolution. Then
the solvent and excess reagent were removed in vacuo, and
the residue was dissolved in dry methylene chloride (50 mL).
Triethylamine (6.0 mL, 43 mmol) and â-alanine allyl ester
hydrochloride (2.77 g, 16.7 mmol) were added slowly at 0 °C,
and the mixture was stirred for 1 h at 25 °C. Aqueous workup
(NaCl/Na2CO3) and purification by chromatography (hexane/
acetone 3:1, Rf ) 0.15) gave N-(2′-((allyloxy)carbonyl)-ethyl)-
4-(chloromethyl)benzamide as a colorless crystalline product.
Mp: 55-56 °C (4.24 g, 90% over all steps). A solution of this
chloride (0.50 g, 1.78 mmol) and sodium iodide (2.22 g, 14.8
mmol) in acetone (18 mL) was refluxed for 2 h. Removal of
the solvent in vacuo, aqueous workup (aqueous sodium sulfite/
CH2Cl2), and flash chromatography (hexane/acetone 5:2, Rf )
0.30) gave 15. Mp: 100-101 °C (0.5908 g, 89%). 1H-NMR
(300 MHz, CDCl3): 7.67, 7.40 (2d, 2 × 2H, J ) 8.5); 6.81 (br t,
1H, J ) 5.8); 5.90 (ddt, 1H, J ) 17, 10.5, 5.8); 5.31 (dq, 1H,
J ) 17, 1.5); 5.24 (dq, 1H, J ) 10.5, 1.5); 4.60 (dt, 2H, J ) 5.8,
1.5); 4.44 (s, 2H); 3.70 (q, 2H, J ) 5.8); 2.66 (t, 2H, J ) 5.8).
HRMS: C14H16NO3I (M + Cs+) calcd 505.9229, found 505.9235.
IR (CCl4): 3388, 1726, 1546, 1499, 1182. Anal. Calcd: C,
45.06; H, 4.32; N, 3.75. Found C, 45.10; H, 3.95; N, 3.46.
Isocratic elution at 1.5 mL/min; detection by UV at 240 nm;
A ) 0.1% TFA in H2O, B ) CH3CN/H2O 50:50; analytical column,
Microsorb-MV 86-203-C5, 0.45 × 22 cm. Preparative HPLC was
performed at 100 mL/min on a Waters prepak cartridge 500 g,
gradient +1% B/min starting with 10-15% less B than the
isocratic conditions. b â-alanine linker. c NOS ) N-oxysuccinimide.
0.86 (t, 3H, J ) 7). HRMS: C26H38N3O6S+ (M+) calcd 520.2481,
found 520.2489. IR (CDCl3): 2932, 2253, 1687, 1549, 1358.
1-((4′-((2-Ca r bon yleth yl)ca r ba m oyl)p h en yl)m eth yl)-4-
a za -3-oxo-4-(p r op ylsu lfon yl)b icyclo[3.2.2]n on a n -1-a zo-
n iu m Tr iflu or oa ceta te (17a ) a n d 1-((4′-((2-Ca r bon yleth -
y l)c a r b a m y l)p h e n y l)m e t h y l)-4-a za -3-o x o -4-(p e n t y l-
su lfon yl)bicyclo[3.2.2]n on a n -1-a zon iu m Tr iflu or oa ceta te
(17b).32 A solution of 16a (0.100 g, 0.165 mmol) in 2 mL of
ethyl acetate was treated with triphenylphosphine (0.25 g, 0.94
mmol), tetrakis(triphenylphosphine)palladium(0) (0.20 g, 0.17
mmol), 2-ethylhexanoic acid (1.32 mL, 8.25 mmol), and tri-
ethylamine (0.575 mL, 4.12 mmol) at 45 °C for 14 h. Ethyl
acetate (15 mL) was then added and the mixture was extracted
with water (containing 0.1% of trifluoroacetic acid). The
aqueous phase was purified by preparative reverse phase
HPLC (see Table 1) to give hapten 17a , a lyophilized colorless
powder. Mp: 175-180 °C (68 mg, 73%). 1H-NMR (500 MHz,
D2O): 7.86, 7.70 (2d, 2 × 2H, J ) 8.5); 4.89 (br s, 1H); 4.73 (s,
2H); 4.59 (s, 2H); 3.86 (m, 4H); 3.65 (m, 4H); 2.70 (m, 2H);
2.53 (m, 2H); 2.36 (m, 2H); 1.79 (m, 2H); 1.01 (t, 3H, J ) 7.5).
13C-NMR (125 MHz, D2O): 178.3, 171.7, 167.8, 138.4, 135.7,
131.1, 130.0, 73.5, 68.8, 58.1, 56.5, 48.3, 38.0, 35.8, 27.4, 18.4,
13.8. HRMS: C21H30N3O6S+ (M+) calcd 452.1855, found
452.1860. IR (KBr): 3418, 3000, 2985, 1670, 1541, 1507, 1362,
1196.
1-((4′-((2-((Allyloxy)ca r b on yl)et h yl)ca r b a m oyl)p h en -
yl)m et h yl)-4-a za -3-oxo-4-(p r op ylsu lfon yl)b icyclo[3.2.2]-
n on a n -1-a zon iu m Tr iflu or oa ceta te (16a ) a n d 1-((4′-((2-
((Allyloxy)ca r b on yl)et h yl)ca r b a m yl)p h en yl)m et h yl)-4-
a za -3-oxo-4-(p en t ylsu lfon yl)b icyclo[3.2.2]n on a n -1-a zo-
n iu m Tr iflu or oa ceta te (16b). A solution of the sultam 14a
(0.10 g, 0.41 mmol) and iodide 15 (0.23 g, 0.62 mmol) in N,N-
dimethylformamide (1 mL) was gently heated to 45 °C for 14
h. The reaction mixture was diluted with acetonitrile (5 mL)
and water (containing 0.1% trifluoroacetic acid, 40 mL). The
precipitate was filtered off and washed with water (5 × 40
mL containing 0.1% trifluoroacetic acid). The filtrate was
purified by preparative reverse phase HPLC (see Table 1) to
give 16a as a colorless oil (0.210 g, 85%). 1H-NMR (300 MHz,
CDCl3): 7.77, 7.56 (2d, 2 × 2H, J ) 8.5); 7.60 (t, 1H, J ) 6);
5.90 (ddt, 1H, J ) 17, 10.5, 5.8); 5.30 (dd, 1H, J ) 17, 1.4);
5.22 (dd, 1H, J ) 10.5, 1.4); 4.90 (s, 2H); 4.76 (br s, 1H); 4.73
(s, 2H); 4.58 (dd, 2H, J ) 6, 1); 3.90 (br s, 4H); 3.69 (q, 2H,
J ) 6); 3.45 (m, 2H); 2.70 (t, 2H, J ) 6); 2.50-2.28 (m, 4H);
1.75 (hex, 2H, J ) 7.5); 1.02 (t, 3H, J ) 7.5). HRMS:
A similar procedure starting with 16b (0.100 g, 0.159 mmol)
gave hapten 17b, a colorless lyophilized powder. Mp: 201-
203 °C (recrist from water) (88 mg, 94%). 1H-NMR (500 MHz,
D2O): 7.86, 7.70 (2d, 2 × 2H, J ) 8); 7.68 (s, 1H); 4.89 (br s,
1H); 4.73 (s, 2H); 4.60 (s, 2H); 3.86 (m, 4H); 3.66 (q, 4H, J )
7); 2.68 (t, 2H, J ) 7); 2.53 (m, 2H); 2.36 (m, 2H); 1.76 (quint,
2H, J ) 7); 1.30 (quint, 2H, J ) 7); 0.85 (t, 3H, J ) 7). 13C-
NMR (125 MHz, D2O): 167.8, 135.6, 129.9, 73.5, 66.8, 56.4,
48.3, 38.1, 36.0, 31.1, 27.3, 23.9, 23.2, 14.8. HRMS: C23H33
-
N3O6S+ (M+) calcd 480.2168, found 480.2172. IR (KBr): 3410,
2960, 1686, 1544, 1356, 1194.
1,4-Dia za -6-[(h exyloxy)im in o]-4-h exylb icyclo[3.2.2]-
n on a n -3-on e (19). A suspension of 1 (0.453 g, 2.68 mmol) in
acetonitrile (15 mL) was gently heated with N-(tert-butyldim-
ethylsilyl)-N-trifluoroacetamide (3.2 mL, 13.7 mmol) until
complete dissolution. After an additional 15 min at 25 °C,
removal of the solvent and excess silylating agent in vacuo
and purification by chromatography (CH2Cl2/MeOH 30:1, Rf
) 0.28) gave the silylated oxime (0.74 g, 98%). A solution of
this silylated oxime (0.202 g, 0.714 mmol) in anhydrous DMF
(2.5 mL) was treated with NaH (60% in mineral oil, 75 mg,
1.88 mmol) for 10 min at 20 °C. 1-Iodohexane (0.158 mL, 1.07
C
24H34N3O6S+ (M+) calcd 492.2168, found 492.2175. IR
(CDCl3): 2253, 1682, 1468, 1381.
A similar procedure starting with sultam 14b (0.10 g, 0.38
mmol) gave 16b as a colorless oil (0.202 g, 85%). 1H-NMR (300
MHz, CDCl3): 7.72 (m, 3H); 7.51 (m, 2H); 5.89 (ddt, 1H, J )
17, 10.5, 5.8); 5.30 (dd, 1H, J ) 17, 1.4); 5.22 (dd, 1H, J )
10.5, 1.4); 4.92 (br s, 2H); 4.75 (br s, 3H); 4.56 (d, 2H, J ) 6);
3.90 (br s, 4H); 3.67 (q, 2H, J ) 6); 3.47 (m, 2H); 2.69 (t, 2H,
J ) 6); 2.50-2.10 (m, 4H); 1.68 (m, 2H); 1.42-1.21 (m, 4H);
(32) The author has deposited atomic coordinates for this structure
(17b) with the Cambridge Crystallographic Data Centre. The coordi-
nates can be obtained, on request, from the Director, Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge, CB2 1EZ,
UK.

3228 J . Org. Chem., Vol. 62, No. 10, 1997
Schro¨er et al.
mmol) was added. After 3 h at 20 °C, the reaction was
quenched with acetic acid (0.2 mL). Aqueous workup (CH2Cl2/
water) and chromatography (CH2Cl2/MeOH 10:1, Rf ) 0.53)
gave 19 as colorless solid. Mp: 48-49 °C (0.163 g, 68%). 1H-
NMR (500 MHz, CDCl3): 4.01 (t, 2H, J ) 7); 3.91 (dd, 1H,
J ) 6, 1.6); 3.77 (d, 1H, J ) 18); 3.74 (s, 2H); 3.71 (d, 1H, J )
18); 3.49 (ddd, 1H, J ) 13.5, 9, 6); 3.39 (ddd, 1H, J ) 13.5,
8.5, 6.5); 3.02 (m, 1H); 2.96 (m, 1H); 2.37 (m, 1H); 2.20 (m,
1H); 1.61 (m, 2H); 1.55 (m, 2H); 1.35-1.25 (m, 12H); 0.86 (m,
6H). 13C-NMR (125 MHz, CDCl3): 171.8, 157.0, 74.4, 62.4,
54.1, 50.3, 48.9, 45.9, 31.6, 30.9, 29.0, 27.7, 26.4, 25.5, 22.5,
14.0. HRMS: C19H35N3O2 (M + Cs+) calcd 470.1784, found
470.1789. IR (CCl4): 2957, 2931, 2854, 1646, 1467, 1138.
1-((2-((4′-((Allyloxy)ca r b on yl)et h yl)ca r b a m yl)p h en -
yl)m et h yl)-4-a za -3-oxo-6-(h yd r oxylim in o)-4-(p r op ylsu l-
fon yl)b icyclo[3.2.2]n on a n -1-a zon iu m Tr iflu or oa cet a t e
(23). A solution of the sultam 22 (0.301 g, 0.772 mmol) and
iodide 15 (0.183 g, 0.491 mmol) in N,N-dimethylformamide (1
mL) was gently heated to 45 °C. Three additional portions of
0.14 g of iodide 15 were added after 1.5, 4, and 7 h. After a
total reaction time of 14 h at 45 °C, the reaction was diluted
with acetonitrile (10 mL) and water (containing 0.1% TFA, 40
mL), thoroughly shaken, and kept at 25 °C for 3 h. The
precipitate was filtered off and washed with water (5 × 40
mL containing 0.1% CF3COOH), and the filtrate purified by
preparative reverse phase HPLC (see Table 1) to give 23, a
lyophilized colorless powder. Mp: 114-116 °C (0.40 g, 82%).
1H-NMR (300 MHz, D2O): 7.68, 7.55 (2 d, 2 × 2H, J ) 8); 5.74
(ddt, 1H, J ) 17, 10.5, 5); 5.27 (dd, 1H, J ) 5, 1); 5.13 (dd, 1H,
J ) 17); 5.04 (br d, 1H, J ) 10.5); 4.72 (s, 2H); 4.55 (m, 4H);
4.43 (d, 2H, J ) 5); 3.88 (m, 1H); 3.72 (m, 1H); 3.56-3.45 (m,
4H); 2.62-2.46 (m, 4H); 1.58 (quint, 2H, J ) 7.5); 0.82 (t, 3H,
J ) 7.5). HRMS: C24H33N4O4S+ (M+) calcd 521.2070, found
521.2076. IR (CH3CN): 3368, 1736, 1694, 1538, 1200, 1166,
1127.
1-((4′-((2-((Allyloxy)ca r b on yl)et h yl)ca r b a m yl)p h en -
yl)m e t h yl)-4-a za -3-oxo-6-(h yd r oxylim in o)-4-h e xylb i-
cyclo[3.2.2]n on a n -1-a zon iu m Tr iflu or oa ceta te (20).
A
solution of 23 (28 mg, 0.082 mmol) and iodide 15 (162 mg, 0.43
mmol) in dry DMF (0.5 mL) was gently heated at 40-45 °C
for 16 h. Acetonitrile (10 mL) and water (containing 0.1% TFA,
30 mL) were added. The precipitate was filtered off and
washed with 5 × 40 mL of water/acetonitrile 3:1. Purification
of the filtrate by preparative reverse-phase HPLC (see Table
1) gave 20 (30.5 mg, 61%). 1H-NMR (300 MHz, CD3OD): 8.72
(t, 1H, J ) 7); 7.97, 7.70 (2d, 2 × 2H, J ) 8); 5.93 (tdd, 1H, J
) 17, 10.5, 5.8); 5.30 (dq, 1H, J ) 17, 1.4); 5.20 (dq, 1H, J )
10.5, 1.4); 4.82 (s, 2H); 4.88, 4.72 (2d, 2 × 1H, J ) 16); 4.60
(dt, 2H, J ) 6, 1.4); 4.42 (s, 2H); 4.20 (dd, 1H, J ) 5.4, 2.2);
4.13 (t, 2H, J ) 7); 4.02, 3.86 (2m, 2 × 1H); 3.66 (m, 2H); 2.70
(t, 2H, J ) 7); 2.55 (m, 2H); 1.66 (m, 2H); 1.42-1.21 (m, 6H);
0.90 (t, 3H, J ) 7). MS (electrospray): (M+) 499.
1-((4′-((2-Ca r b on ylet h yl)ca r b a m yl)p h en yl)m et h yl)-4-
a za -3-oxo-6-(h yd r oxylim in o)-4-h exylbicyclo[3.2.2]n on a n -
1-a zon iu m Tr iflu or oa ceta te (2). A solution of 20 (30.5 mg,
0.050 mmol), triphenylphosphine (88 mg, 0.33 mmol), tetraki-
s(triphenylphosphine)palladium(0) (58 mg, 0.050 mmol), 2-eth-
ylhexanoic acid (0.40 mL, 2.5 mmol), and triethylamine (0.17
mL, 1.24 mmol) in ethyl acetate (1.5 mL) was stirred at 45-
50 °C for 2.5 h, diluted with ethyl acetate (25 mL), and
extracted with dilute aqueous NaCl (containing 0.1% TFA,
6 × 25 mL). The combined aqueous phase was purified by
preparative reverse-phase HPLC (see Table 1) to yield 2, a
colorless lyophilized powder. Mp: 113-115 °C (21 mg, 74%).
1H-NMR (500 MHz, CD3OD): 7.97, 7.71 (2d, 2 × 2H, J ) 8.5);
4.82 (s, 2H); 4.86, 4.74 (2d, 2 × 1H, J ) 18); 4.45, 4.41 (2d,
2 × 1H, J ) 17); 4.19 (dd, 1H, J ) 5, 2); 4.13 (t, 2H, J ) 7);
3.98, 3.86 (2m, 2 × 1H); 3.64 (t, 2H, J ) 7); 2.63 (t, 2H, J ) 7);
2.59, 2.53 (2m, 2 × 1H); 1.68 (quint, 2H, J ) 7); 1.40-1.25
(m, 6H); 0.91 (t, 3H, J ) 7). 13C-NMR (125 MHz, CD3OD):
168.8, 165.5, 148.0, 138.3, 134.8, 130.6, 129.4, 76.6, 72.9, 66.3,
58.4, 57.3, 46.0, 37.3, 34.8, 32.8, 30.0, 28.2, 26.5, 23.6, 14.4.
HRMS: C24H35N4O5 (M+) calcd 459.2607, found 459.2613. IR
(CH3CN): 3628, 2939, 1692, 1180, 1126.
1-((4′-((2-Ca r b on ylet h yl)ca r b a m yl)p h en yl)m et h yl)-4-
a z a -3 -o x o -6 -(h y d r o x y l i m i n o )-4 -(p r o p y l s u l fo n y l )-
bicyclo[3.2.2]n on a n -1-a zon iu m Tr iflu or oa ceta te (3).
A
solution of 23 (48.4 mg, 76.3 µmol) in ethyl acetate/methylene
chloride 1:1 (12 mL) was treated with triphenylphosphine (170
mg, 0.65 mmol), tetrakis(triphenylphosphine)palladium(0) (96
mg, 0.082mmol), 2-ethylhexanoic acid (0.63 mL, 3.9 mmol), and
triethylamine (0.27 mL, 2.0 mmol) at 35 °C for 3 h. Ethyl
acetate (15 mL) was then added and the mixture was extracted
with water (containing 0.1% of trifluoroacetic acid). The
aqueous phase was purified by preparative reverse phase
HPLC (see Table 1) to give hapten 3, a colorless lyophilized
powder. Mp: 136-138 °C (33.5 mg, 74%). 1H-NMR (500 MHz,
D2O): 7.81, 7.69 (2 d, 2 × 2H, J ) 8.5); 5.37 (dd, 1H, J ) 6.5,
2); 4.86, 4.82 (2d, 2 × 1H, J ) 13); 4.79 (dd, 1H, J ) 18, 2.4);
4.72-4.66 (m, 2H); 4.60 (dd, 1H, J ) 18, 2.4); 3.99, 3.86 (2m,
2 × 1H); 3.59 (t, 2H, J ) 7); 3.56 (m, 2H); 2.73-2.56 (m, 2H);
2.64 (t, 2H, J ) 7.5); 1.68 (hex, 2H, J ) 7.5); 0.93 (t, 3H, J )
7.5). 13C-NMR (125 MHz, D2O): 178.1, 171.6, 166.1, 147.6,
138.5, 135.8, 130.1, 74.0, 67.1, 58.3, 58.0, 57.4, 50.2, 37.8, 35.4,
27.7, 18.3, 13.7. HRMS: C25H32N5O9S+ (M+) calcd 481.1751,
found 481.1788. IR (KBr): 3422, 1685, 1654, 1364, 1200, 1163,
1128.
1-(((P r op ylsu lfon yl)ca r ba m yl)m eth yl)-1-((4′-((2-ca r bo-
n yleth yl)ca r ba m yl)p h en yl)m eth yl)-3-{h yd r oxylim in o)cy-
cloh ex-4-en -1-a zon iu m Tr iflu or oa ceta te (24). A solution
of sultam 23 (5.4 mg, 0.0085 mmol) in CH2Cl2 (1 mL) and ethyl
acetate (0.5 mL) was heated at 40 °C for 3 h with tetraki-
s(triphenylphosphine) palladium(0) (13 mg, 0.0011 mmol),
triphenylphosphine (18.5 mg, 0.07 mmol), and potassium
2-ethylhexanoate (71 mg, 0.39 mmol). The reaction was
diluted with water (40 mL, containing 0.1% TFA) and washed
with CH2Cl2 (5 mL). Preparative reverse-phase HPLC (see
Table 1) of the aqueous solution yielded compound 24 (2.4 mg,
47%) as the only product. 1H-NMR (300 MHz, D2O): 7.64, 7.42
(2d, 2 × 2H, J ) 8); 6.43 (d, 1H, J ) 10.5); 6.10 (dt, 1H, J )
10.5, 1.5); 4.88, 4.82 (2d, 2 × 1H, J ) 16); 4.71, 4.42 (2d, 2 ×
1H, J ) 16); 4.26 (br s, 2H); 3.80, 3.72 (2d, 2 × 1H, J ) 16);
3.48 (t, 2H, J ) 6); 3.10 (t, 2H, J ) 8); 2.53 (t, 2H, J ) 6.5);
1.58 (hex, 2H, J ) 7.5); 0.86 (t, 3H, J ) 7.5). MS (FAB+) (M+)
481.
1,4-D ia za -6-[(t er t -b u t y ld im e t h y ls ily l)o x im in o ]-4-
(p r op ylsu lfon yl)bicyclo[3.2.2]n on a n -3-on e (22). The si-
lylated derivative of oxime 1 (0.222 g, 0.783 mmol) was
dissolved in 20 mL of THF and treated with NaH (60% in
mineral oil, 0.115 g, 2.88 mmol). After 30 min. at 25 °C,
1-propanesulfonyl chloride (0.86 mL, 7.6 mmol) was slowly
added. The reaction was stirred for 3 h at 25 °C and then
refluxed for 1 h. Quenching with acetic acid (0.2 mL), solvent
removal in vacuo, and chromatography (hexane/ethyl acetate
3:2, TLC 1:1, 2 isomers in 4 to 1 ratio, Rf ) 0.40 and 0.28)
gave 22 (0.29 g, 95%). The pure major isomer (Rf ) 0.40), a
colorless solid, was characterized. Mp: 96-97 °C. 1H-NMR
(500 MHz, CDCl3): 5.17 (dd, 1H, J ) 5, 2); 3.87 (dd, 1H, J )
19, 1); 3.84 (dd, 1H, J ) 19, 1); 3.69 (d, 1H, J ) 19); 3.63 (dd,
1H, J ) 19, 1); 3.57 (ddd, 1H, J ) 14, 10.6, 5.3); 3.35 (ddd,
1H, J ) 14, 10.5, 5.5); 3.07 (m, 2H); 2.45-2.37 (m, 1H); 2.22-
2.15 (m, 1H); 1.81-1.75 (m, 1H); 1.73-1.65 (m, 1H); 0.98 (t,
3H, J ) 7.5); 0.87 (s, 9H); 0.11, 0.09 (2s, 2 × 3H). 13C-NMR
(125 MHz, CDCl3): 174.6, 160.1, 63.3, 56.0, 50.9, 49.6, 45.3,
30.0, 25.7, 17.8, 16.7, 12.5. HRMS: C16H31N3O4SSi (M + Cs+)
calcd 522.6859, found 522.6865. IR (CCl4): 2931, 2858, 1690,
1462, 1364, 1160. Anal. Calcd: C, 49.33; H, 8.02; N, 11.08; S,
8.23. Found C, 49.34; H, 8.28; N, 11.02; S, 7.97.
P r ep a r a tion of th e NHS-Ester s a n d Cou p lin g to Ca r -
r ier P r otein s. A solution of hapten 17a (20 mg, 0.035 mmol)
in 0.5 mL of DMF was treated with N-hydroxysuccinimide
(0.02 mL of a 3.65 M solution in DMF, 0.073 mmol) and 1-(3-
(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride (0.04
mL of a 1.8 M solution in water, 0.072 mmol) at 20 °C for 3 h,
diluted with 15 mL water and rapidly purified by preparative
reverse-phase HPLC (see Table 1) to give 18a as a colorless
lyophilized powder (19 mg, 81%). Similar procedures gave 18b
(19 mg, 76%) from hapten 17b (21 mg, 0.035 mmol), 21 (6.3
mg, 95%) from 2 (5.7 mg, 0.010 mmol), and 25 (1.8 mg, 7%)
from 3 (22 mg, 0.037 mmol). These activated esters were
approximately 95% pure by HPLC, containing 5% starting acid

Design and Synthesis of Transition State Analogs
J . Org. Chem., Vol. 62, No. 10, 1997 3229
that formed during collection of the HPLC fractions. These
were used directly for coupling to carrier proteins as follows:
Ester 18a (19 mg, 0.029 mmol) was dissolved in 0.2 mL of
water/DMF 1:1. A 0.05 mL portion was added to a solution of
KLH (keyhole limpet hemocyanin, 9.7 mg) in 1.95 mL of PBS
(10 mM phosphate, 160 mM NaCl in water, pH 7.4) and 0.15
mL was added to a solution of BSA (bovine serum albumin,
30 mg) in 2.85 mL of PBS. These solutions were kept at 4 °C
for 24 h and then used directly for immunological work. Ester
18b (18.5 mg, 0.027 mmol) was similarily combined with KLH
(9.5 mg) and BSA (30 mg). Ester 21 (6.3 mg, 0.01 mmol) was
prepared directly from the hapten. Thus acid 3 (10 mg, 0.017
mmol) in 0.1 mL of DMF was treated with N-hydroxysuccin-
imide (0.036 mmol) and EDC (0.036 mmol) for 2.5 h at 20 °C,
and combined directly with BSA (11 mg in 0.87 mL PBS) at 4
°C for 2 h. This conjugate was purified by gel filtration over
a PD-10 column (Sephadex G-25) eluting with PBS.
Ack n ow led gm en t. This work was supported in part
by the Humboldt Stiftung, Germany (J .S.) and by the
US National Institutes of Health (GM 49736 to J .-L.R.).
dissolved in 0.5 mL of PBS containing 0.08 mL of DMF.
A
Su p p or tin g In for m a tion Ava ila ble: 1H-NMR spectra of
compounds 2, 3, 6, 9, 10, 16a , 16b, 17a , 17b, 19, 20, 23, 24
(13 pages). This material is contained in libraries on micro-
fiche, immediately follows this article in the microfilm version
of the journal, and can be ordered from the ACS; see any
current masthead page for ordering information.
0.21 mL portion of this solution was added to KLH (9.2 mg in
1.82 mL PBS), and 0.37 mL was added to BSA (16 mg in 3.16
mL PBS) and kept at 4 °C for 24 h. Ester 25 (1.6 mg, 0.0023
mmol) was dissolved in 0.017 mL of water/DMF 1:1 and
combined with KLH (3.2 mg in 0.32 mL PBS) at 4 °C. After
1.5 h, the coupling reaction was complete as judged by HPLC,
and the sample was frozen to -80 °C. The BSA conjugate was
J O962179V