An investigation of antibody acyl hydrolysis catalysis using a large set of related haptens

Article abstract of DOI:10.1016/S0968-0896(99)00302-8

An aspect of catalytic antibody research that receives little attention in the literature involves hapten systems that fail to elicit antibody catalysts despite a high affinity immune response and hapten designs that resemble those known to elicit catalysts. We have investigated a series of 12 phosphate and phosphonate haptens in a total of three animal systems. Dramatic and reproducible differences were observed in the catalytic activities of polyclonal antibodies elicited by the different haptens. A phosphate hapten with a phenyl ring on the side of the hapten opposite the linker elicited reproducibly high levels of polyclonal antibody catalytic activity. The other 11 haptens, most with benzyl groups on the side of the hapten opposite the linker, elicited immune responses in which catalytic activity was significantly weaker in terms of the level of observed catalytic activity, as well as frequency of elicited catalysts. Our results indicate that subtle features of transition state analogue hapten structure can have a dramatic and reproducible influence over the catalytic activity of elicited antibodies in related haptens. Whatever the explanation, subtle changes in mechanistic features due to altered leaving group ability/location or overall hapten flexibility, the comprehensive data presented here indicate that phenyl or 4-nitrophenyl leaving groups located opposite the hapten linker are to be preferred in order to elicit highly active antibody catalysts for acyl hydrolysis reactions. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(99)00302-8

Bioorganic & Medicinal Chemistry 8 (2000) 413±426
An Investigation of Antibody Acyl Hydrolysis Catalysis Using a
Large Set of Related Haptens
Amy L. Odenbaugh, Eric D. Helms and Brent L. Iverson*
The Department of Chemistry and Biochemistry, The University of Texas at Austin, Austin, TX 78712, USA
Received 16 July 1999; accepted 8 November 1999
AbstractÐAn aspect of catalytic antibody research that receives little attention in the literature involves hapten systems that fail to
elicit antibody catalysts despite a high anity immune response and hapten designs that resemble those known to elicit catalysts.
We have investigated a series of 12 phosphate and phosphonate haptens in a total of three animal systems. Dramatic and repro-
ducible di€erences were observed in the catalytic activities of polyclonal antibodies elicited by the di€erent haptens. A phosphate
hapten with a phenyl ring on the side of the hapten opposite the linker elicited reproducibly high levels of polyclonal antibody
catalytic activity. The other 11 haptens, most with benzyl groups on the side of the hapten opposite the linker, elicited immune
responses in which catalytic activity was signi®cantly weaker in terms of the level of observed catalytic activity, as well as frequency
of elicited catalysts. Our results indicate that subtle features of transition state analogue hapten structure can have a dramatic and
reproducible in¯uence over the catalytic activity of elicited antibodies in related haptens. Whatever the explanation, subtle changes
in mechanistic features due to altered leaving group ability/location or overall hapten ¯exibility, the comprehensive data presented
here indicate that phenyl or 4-nitrophenyl leaving groups located opposite the hapten linker are to be preferred in order to elicit
highly active antibody catalysts for acyl hydrolysis reactions. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
At least two conditions must theoretically be met for a
transition state analogue hapten to elicit antibody cat-
alytic activity by design. First, it is assumed to be
necessary, but not sucient, that the transition state
analogue has some structural features in common with
the rate-limiting transition state, such as charge dis-
tribution and geometry, that are not found in starting
material(s) or product(s). Second, it is also assumed that
the binding energy within elicited antibody binding
pockets is partitioned in such a way that substrate
binds, then is facilitated along the entire reaction path.
Failure to meet this second condition, either by failing
to allow substrate binding or by binding substrate in
such a way that catalytic product formation is not
favored, is presumably the reason so many high anity
antibodies elicited by transition state analogue haptens
fail to a€ect catalysis of a desired reaction with sub-
strate. Gaining a predictive understanding of how
overall hapten structure a€ects this second condition
will require an understanding of how di€erent struc-
tural features of a hapten are recognized by elicited
antibodies.
An aspect of catalytic antibody research that receives
little attention in the literature involves those hapten
systems that fail to elicit antibody catalysts despite a
high anity immune response and hapten designs that
resemble those known to elicit catalysts. A better
understanding of hapten structural features that lead to
a reduced or complete lack of elicited catalytic activity
would dramatically aid in the design of future haptens.
Keywords: polyclonal catalytic antibodies; acyl hydrolysis.
Abbreviations: kcat, catalytic rate constant, calculated as Vmax/2[IgG]
where [IgG] is measured through quantitative inhibition experiments
with hapten while Vmax and KM are the maximum catalytic velocity
and Michaelis constant, respectively, derived from the Lineweaver±
Burk analysis; kuncat, rate of reaction in the absence of added catalyst;
KLH, keyhole limpet hemocyanin; BSA, bovine serum albumin;
ELISA, enzyme-linked immunosorbent assay; PBS, phosphate buf-
fered saline; EDC, [1-ethyl-3-(dimethylamino)propyl]carbodiimide;
DMF, N,N-dimethylformamide; THF, tetrahydrofuran; MS, mass
spectral analysis; FAB, fast atom bombardment; DCC, dicyclohex-
ylcarbodiimide; HOBt, 4-hydroxybenzotriazole; KA, equilibrium asso-
ciation constant; KD, equilibrium dissociation constant calculated as
1/KA; nm, not measured; nd, none detected; Tris, tris(hydroxy-
methyl)aminomethane.
The advent of polyclonal catalytic antibody studies
involving immunization with a transition state analogue
hapten makes possible the systematic study of hapten
structure in the context of elicited antibody recognition
*Corresponding author. Tel.: +1-512-471-5053; fax: +1-512-417-
8696.
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00302-8

414
A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
and catalytic activity for an entire immune response.1±21
For example, a systematic study from our laboratory
involving the replacement of a single atom in a transi-
tion state analogue hapten±substrate system found that
optimum catalysis was observed when the hapten and
substrate were most similar in structure.12 A di€erent
study uncovered a striking relationship between hapten
size/hydrophobicity and catalysis.13 Speci®cally, a series
of phosphate haptens with large aromatic substituents
elicited antibodies that were not catalytic, but bound
strongly to the aromatic portions of their haptens and
substrates. Similar phosphate haptens with smaller aro-
matic groups elicited antibodies with catalytic activity
that was inversely correlated to the size of the aromatic
group. This latter study has veri®ed by systematic
experiment that the level of elicited catalytic activity can
be modulated in dramatic fashion by structural features
distinct from the portion of the hapten intended to
mimic uniquely a rate-limiting transition state.
suggest that increased ¯exibility and/or extremely subtle
mechanistic di€erences may be operating to limit sig-
ni®cantly elicited catalytic activity.
Results
The six (three enantiomeric pairs) phosphonate and
phosphate haptens 1±6 shown in Scheme 1 were syn-
thesized and coupled to keyhole limpet hemocyanin
(KLH). These haptens were designed to elicit catalysts
for the BAC2 hydrolysis of the corresponding ester and
carbonate substrates 8±13. The hapten±KLH conjugates
were used to immunize New Zealand white rabbits ®ve
times at 21-day intervals using standard protocols. In
each case, no catalysis of the corresponding substrates
was observed in the elicited polyclonal response in any
of the serum samples that were removed 10 days fol-
lowing both the third and ®fth immunizations. Impor-
tantly, as indicated by the relative dissociation constants
determined after the ®fth immunization for each sample
and listed in Table 1, high anity antibodies were eli-
cited by these immunizations. These relative dissocia-
tion constants were determined using competition
ELISA, a technique that accurately ranks relative an-
ities, although the absolute anities may be under-
estimated.22±24 As expected, each sample retained the
ability to discriminate between enantiomers, and the
immunizing hapten was always substantially preferred
relative to its enantiomer. Substrate binding in the range
of 1Â102 to 8Â103 M 1 was observed for the antibodies
elicited against haptens 5 and 6 by competition ELISA,
but limited polyclonal sample precluded the measure-
ment of substrate binding for the antibodies raised
against haptens 1±4.
The present study involves a series of 12 phosphate and
phosphonate haptens, all of which have structural fea-
tures that should have made them capable of eliciting
active polyclonal catalytic antibodies. We found that six
haptens, in the form of three pairs of enantiomers, failed
to induce any catalytic activity in rabbit polyclonal
antibodies despite the presence of high anity and
highly speci®c immune responses. This surprising result
led to the investigation of a series of benzyl phosphate
haptens in three di€erent animal models. For the entire
series of haptens, the elicited catalytic activity was sig-
ni®cantly weaker in terms of the level of observed cata-
lytic activity, as well as frequency of elicited catalysts,
compared to a phenyl phosphate hapten we have used
previously. In summary, our results indicate that only a
hapten with a phenyl group on the side of the hapten
opposite the linker was able to elicit reproducibly high
levels of polyclonal catalytic activity. Taken with
mechanistic investigations also described, our ®ndings
A positive control experiment was carried out using the
rabbit originally immunized with hapten 3 and the
rabbit originally immunized with hapten 4. Six months
Scheme 1. Structures of haptens 1±7 and substrates 8±14.

A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
Table 1. Apparent dissociation constants, KD (mM), for the indicated
415
haptens and substrates measured for the various samples of New
Zealand white rabbit polyclonal antibodies as determined by compe-
tition ELISAa
Rabbit #/antibodies Hapten
KD (mM)
Substrate KD (mM)
2403/anti-1
2401/anti-2
2509/anti-3
2510/anti-4
2164/anti-5
2165/anti-5
2162/anti-6
2163/anti-6
1
2
2
1
3
4
4
3
5
6
5
6
6
5
6
5
<0.1 (‹0.4)
20 (‹2)
0.6 (‹0.2)
80 (‹9)
0.7 (‹0.3)
70 (‹10)
0.4 (‹0.1)
50 (‹6)
10 (‹2)
12
12
13
13
8000
1000
200
27
800 (‹80)
80 (‹10)
600 (‹60)
100 (‹10)
3000 (‹700)
0.2 (‹0.9)
300 (‹50)
Scheme 2. Structures of haptens 15±19 and substrates 20±24.
of the sera is required for catalytic polyclonal antibody
studies carried out in mice because of the small volume
of serum present in a single mouse. Pooling of sera also
reduces experimental artifacts derived from di€erences
in the immune responses of individual animals.
aStandard deviations of repetitive ELISA measurements are shown in
parentheses.
following the last immunization with either 3 or 4, each
rabbit was immunized ®ve times at 21-day intervals with
the phenyl phosphate hapten 7. Polyclonal antibodies
isolated from the sera of both rabbits 10 days following
the ®fth immunization displayed considerable catalytic
activity with the carbonate substrate 14. Lineweaver±
Burk analyses were used to determine apparent KM and
Vmax values.4 These values are termed `apparent' to
emphasize that they characterize the polyclonal samples
as a whole rather than any one catalytic antibody spe-
cies. The fraction of high anity, hapten-speci®c anti-
bodies were measured for each sample by carrying out
quantitative inhibition studies with hapten.4 These
fractions were used to estimate catalyst concentration
in order to calculate the values of apparent kcat and
apparent kcat/kuncat that are listed in Table 2.
For the polyclonal antibody samples that were catalytic
for hydrolysis of their most homologous substrates,
apparent KM, Vmax and kcat/kuncat were determined
(Table 3). Overall, the samples elicited by 15±17 dis-
played similar levels of catalytic activity, with apparent
kcat/kuncat values ranging from 420 to 1250.
While the anti-18 and anti-19 samples were not cataly-
tic for hydrolysis of their most homologous substrate,
these samples were capable of catalyzing the hydrolysis
of substrates 20±22. This unusual substrate speci®city
likely results from a change in mechanism for substrates
23 and 24, so that the haptens 18 and 19 no longer
resemble a rate-limiting transition state for cleavage.
The pH dependence of hydrolysis of all the benzyl
carbonate substrates was investigated in the absence of
any polyclonal antibody catalysts. Carbonates 20±22
hydrolyzed in a base-dependent fashion, as expected for
the BAC2 mechanism. Substrates 23 and 24 hydrolyzed
in a pH independent fashion, indicating that these sub-
strates hydrolyze by a mechanism di€erent from BAC2.
Although more detailed mechanistic investigation
would be required to establish the exact cleavage
mechanism for 23 and 24, it is tempting to propose a
rate-limiting step involving alkyl-oxygen heterolysis, in
which an intermediate benzylic cation is stabilized by
the electron-donating para substituent. A unimolecular
®ssion pathway has been previously suggested for the
thermal decomposition of a p-methoxybenzyl carbonate.25
The para-substituted benzyl haptens 15±19 were pre-
pared and coupled to KLH (Scheme 2). These haptens
were designed to elicit catalysts for the hydrolysis of the
correspondingly substituted benzyl carbonate substrates
20±24. Each hapten±KLH conjugate was used to
immunize a di€erent set of 10 BALB/cJ mice a total of
®ve times at 21-day intervals. As a control, polyclonal
antibodies were also elicited from a separate set of 10
BALB/cJ mice using underivatized KLH. Serum was
removed 10 days after the ®fth immunization from each
of the ten animals immunized with the same hapten.
The 10 serum samples were combined and the poly-
clonal antibodies isolated from the pooled sera. Pooling
Table 2. Kinetic parameters for rabbits 2509 and 2510 immunized ®ve times with hapten 7a
Rabbit #
Apparent KM (mM)
Apparent Vmax (mM/min)
% Hapten speci®c antibody
Apparent kcat
Apparent kcat/kuncat
2
1
1
2509
2510
34
85
0.27
0.59
10
5.5
9.6Â10 min
1430
4700
1
3.16Â10 min
aCatalytic reactions utilized substrate 14. The background rate for hydrolysis of 14 measured as kuncat in the absence of polyclonal antibodies was
5
1
6.7Â10 min
.

416
A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
Table 3. Immunological and catalytic parameters determined for the polyclonal antibodies isolated from the BALB/cJ mice
Hapten used to elicit BALB/cJ murine polyclonal antibodies
15
16
17
18
19
Titerb
1:800
0.12
73
0.11
0.045
1250
1:800
0.068
36
0.17
0.017
630
1:800
0.0056
11
0.09
0.016
420
1:1600
nda
1:800
nd
nm
nm
nm
nm
Apparent Vmax (mM/min)c
Apparent KM (mM)c
nma
nm
Hapten speci®c antibodyd total IgG
Apparent kcat (min
Apparent kcat/kuncat
1 e
)
nm
nm
f
anm, not measured; nd, none detected.
bMeasured from standard ELISA using puri®ed IgG at 1 mg/mL.
cDetermined by Lineweaver±Burk analysis.
dDetermined by quantitative inhibition studies.
eCalculated as apparent Vmax divided by the amount of high-anity, hapten speci®c antibody binding sites in the reactions.
fThe values of kuncat were determined under identical bu€er conditions in the absence of added antibody. The kuncat values of substrates 20, 21, and
5
5
5
22 were 3.6Â10 min 1, 2.7Â10 min 1, and 3.8Â10 min 1, respectively.
Apparent relative hapten dissociation constants (KD)
were determined using competition ELISA (Table 4). As
expected, in each case, the polyclonal samples bound
with highest anity the hapten used to elicit them. By
far the greatest speci®city for hapten was seen with the
antibodies elicited by hapten 15, the 4-nitrobenzyl deriv-
ative. This result was not unexpected, as early in this
century the pioneering studies of antibody cross-reactivity
by Landsteiner26 and Pauling and Pressman27,28 showed
that more electronegative groups, such as the nitro group,
can dominate a high anity antibody response.
murine samples. In contrast, none of the rabbits immu-
nized with 15 or 17 produced catalytic antibodies fol-
lowing the full immunization schedule.
The KD values obtained from the immunized rabbits
and ICR mice are shown in Tables 6 and 7. Again,
within a given sample, the highest anity was seen for
the hapten with which it was elicited. In general, the
anities seen with the polyclonal antibodies isolated
from the ICR mice were similar or slightly better than
those seen with the BALB/cJ mice, while the anities
measured for the samples from the rabbits were 3- to
40-fold higher than the murine samples.
To probe further the in¯uences of animal species and
strain upon the catalytic immune response, the KLH
conjugates of haptens 15±18 were also used to immunize
sets of 10 outbred ICR mice and pairs of New Zealand
white rabbits. The immunization schedule followed was
identical to that used for the BALB/cJ mice and the
polyclonal samples were isolated from sera taken 10
days after the ®fth and ®nal immunization. A sample
was also removed from the rabbits 10 days following the
third immunization. The catalytic parameters of the
polyclonal antibody samples for their most hapten-
homologous substrates were determined as before and
are shown in Table 5. The apparent kcat/kuncat values for
the outbred ICR mice are about half as large, yet show
the same trend of relative catalytic activities, as was seen
with inbred BALB/cJ mice. Both rabbits immunized
with 16 produced catalytic antibodies, and these anti-
bodies displayed kcat/kuncat values similar to those of the
Despite their lack of catalytic ability, the polyclonal
antibodies elicited by haptens 15 and 17 in the rabbits
did show a high degree of hapten anity and speci®city.
Importantly, these polyclonal samples failed to show
any signi®cant binding to either their most hapten-
homologous substrates, 20 and 22, or the benzyl alcohol
products that would result from the hydrolysis of these
substrates (data not shown). Thus, it appears that the
lack of observed catalysis elicited in these animals is due
to a lack of substrate binding.
Catalytic parameters were also determined for the anti-
bodies obtained from the rabbits 10 days following
the third injection with haptens 15±17. The apparent
kcat/kuncat values of the anti-16 rabbit samples were
lower than the values observed with antibodies obtained
Table 4. Apparent dissociation constants, KD (mM), for various haptens of BALB/cJ murine polyclonal antibodies as determined by competition
ELISAa
BALB/cJ murine antibodies
Hapten
15
16
>1000
17
18
19
Anti-15
Anti-16
Anti-17
Anti-18
Anti-19
20 (‹11)
50 (‹10)
300 (‹180)
100 (‹67)
>1000
>1000
>1000
>1000
400 (‹170)
20 (‹8.1)
30 (‹13)
30 (‹16)
400 (‹180)
5 (‹1.8)
30 (‹21)
150 (‹22)
240 (‹69)
0.5 (‹0.3)
160 (‹38)
5 (‹1.9)
1 (‹0.48)
>1000
700 (‹260)
>1000
aStandard deviations of repetitive ELISA measurements are shown in parentheses.

A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
417
Table 5. Immunological and catalytic parameters determined for the polyclonal antibodies isolated from the immunized ICR mice and the New
Zealand white rabbits
Hapten used to elicit polyclonal antibodies
15 (ICR)
16 (ICR)
17 (ICR)
16 (NZW#3)
16 (NZW#4)
Titera
1:500
0.12
28
0.24
0.021
580
1:1000
0.056
23
0.32
0.0073
270
1:800
0.043
52
0.23
0.0078
200
1:800
0.066
9.3
0.10
0.028
1000
1:900
0.053
13
0.13
0.017
630
Apparent Vmax (mM/min)b
Apparent KM (mM)b
Hapten speci®c antibodyc total IgG
Apparent kcat (min
Apparent kcat/kuncat
1 d
)
e
aMeasured from standard ELISA using puri®ed IgG at 1 mg/mL.
bDetermined by Lineweaver±Burk analysis.
cDetermined by quantitative inhibition studies.
dCalculated as apparent Vmax divided by the amount of high-anity, hapten speci®c antibody binding sites in the reactions.
eThe values of kuncat were determined under identical bu€er conditions in the absence of added antibody. The kuncat values of substrates 20, 21, and
5
5
5
22 were 3.6Â10 min 1, 2.7Â10 min 1, and 3.8Â10 min 1, respectively.
after ®ve immunizations (Table 8). The KD values for
the same antibodies were slightly higher than those fol-
lowing the ®fth injection. Surprisingly, the polyclonal
antibodies isolated after the third injection from one of
the two rabbits immunized with hapten 15 were slightly
catalytic. The pseudo-®rst-order rate constant for the
hydrolysis reaction of carbonate 20 in the presence of
12 mM anti-15 polyclonal antibody was 3.9-fold higher
than the pseudo-®rst-order rate constant found for the
reaction in the absence of polyclonal antibodies.
Unfortunately, this modest level of catalytic activity was
not sucient to allow a reliable determination of
apparent catalytic parameters for the sample due to a
low signal-to-noise ratio in the cleavage assay.
Discussion
The initial phase of the present work involved a series of
enantiomeric hapten pairs, 1±6, that were used to
immunize New Zealand white rabbits. Enantiomeric
pairs are ideal for use in systematic hapten structure±
activity studies, because the two enantiomers have
identical physical properties in achiral environments.
Nevertheless, the stereoselectivity inherent in the
immune response generally leads to two di€erent
immunological solutions to recognition, one for each
enantiomer. In this way, the generality of any structure±
activity relationship in the immune response can be
gauged by comparing the results obtained with the pair
of enantiomers.
18O labeling studies were used to investigate the rate-
limiting step of the benzyl carbonate hydrolysis. If
formation of the presumed tetrahedral intermediate is
rate-limiting, then 18O incorporation into substrate will
not be observed, while if breakdown of the presumed
tetrahedral intermediate is rate-limiting, then some sub-
strate should be recovered that has 18O incorporated as
the carbonyl oxygen atom. Even after a 66 h incubation
at pH 8.2 in 50% H2 18O, mass spectral analysis of
recovered substrate failed to show any incorporation of
solvent oxygen into substrate 20 over the background
level of isotopic oxygen, either in the presence or in the
absence of rabbit anti-15 antibodies. Using this tech-
nique, as little as 2% 18O incorporation would have
been detected.
For haptens 1±6, the results of the immunizations were
entirely self-consistent between enantiomers as well as
among all the members of the set. In each case, a high
anity, stereoselective immune response was obtained
judging from the ELISA studies (Table 1). The signi®cant
stereoselective preference for binding the immunizing
hapten, as opposed to its enantiomer, indicates that the
polyclonal responses were distinct for each enantiomer
and that the hapten remained intact during the course
of the immune response. Nevertheless, no catalytic
activity was observed. These results are surprising,
because 1±6 are similar in size to other phosphate and
Table 7. Apparent dissociation constants, KD (mM), for various
haptens of New Zealand white rabbit polyclonal antibodies as deter-
mined by competition ELISAa
Table 6. Apparent dissociation constants, KD (mM), for various
haptens of ICR murine polyclonal antibodies as determined by com-
petition ELISAa
Hapten
Rabbit #/antibodies
15
16
17
ICR murine antibodies
Hapten
NZW#1/anti-15
NZW#2/anti-15
NZW#3/anti-16
NZW#4/anti-16
NZW#5/anti-17
NZW#6/anti-17
0.05 (‹0.009) 0.9 (‹0.22) >1000
0.8 (‹0.49)
10 (‹3.7)
1 (‹0.32)
5 (‹3.7)
40 (‹26)
300 (‹180)
200 (‹110)
80 (‹40)
0.5 (‹0.28)
0.5 (‹0.14)
15
16
17
0.6 (‹0.24)
0.4 (‹0.27)
2 (‹0.86)
2 (‹1.4)
Anti-15
Anti-16
Anti-17
2 (‹0.71) 20 (‹7.1) >1000
20 (‹6.9)
120 (‹40)
2 (‹1.1)
5 (‹1.9)
200 (‹92)
1 (‹0.55)
20 (‹8.8)
aStandard deviations of repetitive ELISA measurements are shown in
parentheses.
aStandard deviations of repetitive ELISA measurements are shown in
parentheses.

418
A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
Table 8. Immunological and catalytic parameters determined for the
polyclonal antibodies isolated from the New Zealand white rabbits
immunized with hapten 16 following the third injection of hapten
Anti-16 third bleed antibody
Rabbit #3
Rabbit #4
Titera
1:1250
1 (‹0.42)
0.032
9.3
8.0
0.017
620
1:2100
1.8 (‹0.18)
0.032
39
KD (hapten, mM)b
Apparent Vmax (mM/min)c
Apparent KM (mM)c
Hapten speci®c antibodyd total IgG
Apparent kcat (min
Apparent kcat/kuncat
12
0.011
410
Scheme 3. (A) The structures of the hapten used by Gallacher and
Brockelhurst et al.1±3,5,10,11 to immunize sheep. (B) The structure of
the hapten used by Suzuki et al.7 to immunize rabbits.
1 e
)
f
aMeasured from standard ELISA using puri®ed IgG at 1 mg/mL.
bDetermined by competition ELISA.
group ability/structure and/or location may be respon-
sible for the observed di€erences in elicited catalytic
activities. Haptens 1±6 and the haptens used to elicit
successfully catalytic activity are of a similar size, and
haptens 1, 2 and 5, 6 have phenyl rings on both sides of
the phosphonate or phosphate moiety, respectively, just
like the successful haptens. One di€erence between suc-
cessful and unsuccessful haptens involves the nature of
the leaving group on corresponding substrates. Corre-
sponding substrates for the successful haptens have
relatively good leaving groups, phenyl or 4-nitrophenyl,
while corresponding substrates for the unsuccessful
haptens have 4-alkylphenyl (more electron rich com-
pared to phenyl or 4-nitrophenyl) or benzyl leaving
groups. Perhaps these di€erences in leaving group lead
to di€erences in rate-limiting transition state structure
(see `Mechanistic considerations' below). Another clear
di€erence is that the successful haptens have phenyl or
4-nitrophenyl groups located on the side of the hapten
opposite the linker, while the unsuccessful haptens 1±6
have more ¯exible alkyl or benzyl groups in the analo-
gous position. Note that attention is being paid to the
location of the group with respect to the linker, since it
is generally observed that groups farthest from the lin-
ker dominate the immune response and are buried most
deeply in the elicited antibody binding pockets. This is
not surprising given that linkage to a carrier protein
introduces obvious steric hindrance to binding of
nearby groups, thereby limiting antibody interactions in
this region of the hapten.
cDetermined by Lineweaver±Burk analysis.
dDetermined by quantitative inhibition studies.
eCalculated as apparent Vmax divided by the amount of high-anity,
hapten speci®c antibody binding sites in the reactions.
fThe values of kuncat were determined under identical bu€er conditions
in the absence of added antibody. The kuncat values of substrates 20,
5
5
5
21, and 22 were 3.6Â10 min 1, 2.7Â10 min 1, and 3.8Â10
min 1, respectively.
phosphonate haptens that have been used previously to
elicit catalytic antibodies.
It is unlikely that serendipity alone can explain the lack
of catalytic activity elicited by haptens 1±6, especially
since pairs of enantiomers yielded the same results.
Rather, there are probably structural features present in
these six haptens that somehow prevent catalysis in the
responding polyclonal antibodies. To emphasize this
point, the rabbits originally immunized with haptens 3
and 4 were each immunized with hapten 7 as a positive
control. Hapten 7 has been shown previously to elicit
catalytically active polyclonal antibodies in rabbits as
well as BALB/cJ mice. In both rabbits originally
immunized with 3 and 4, catalytic activity was present
in the elicited polyclonal antibodies, and the observed
apparent kcat/kuncat values of 1430 and 4700 compare
well with the values obtained for other polyclonal sam-
ples immunized with this hapten. To the best of our
knowledge, these are the ®rst examples of animals being
successfully reused for catalytic antibody studies.
If altered leaving group ability/location and/or hapten
¯exibility are important, then benzyl haptens 15±19
would also be predicted to be less likely to elicit anti-
body catalysts than a phenyl hapten such as 7; and,
indeed, both in terms of the level of observed catalytic
activity, as well as frequency of elicited catalysts, the
haptens 15±19 proved to be signi®cantly less e€ective
than hapten 7. Benzyl haptens 15±17 elicited catalysts
The phosphate and phosphonate haptens investigated in
this and previous polyclonal catalytic antibody studies
can be divided into two distinct groups. Those haptens
that have reproducibly generated antibody catalysts
when used to immunize di€erent animals, and those
haptens that failed to elicit catalysts in repeated
attempts, despite the presence of a high anity immune
response. Haptens in the ®rst category include hapten
7, used in this and previous studies,9,13 the nitro-
phenyl phosphate hapten of Gallacher and Brocklehurst
et al.1±3,5,10,11 (hapten `A' in Scheme 3), and the nitro-
phenyl phosphate hapten of Suzuki et al.7 (hapten `B' in
Scheme 3). Haptens in the latter category include hap-
tens 1±6.
that were only moderately active, having apparent kcat
/
kuncat values between 0 and 1250 in all the animal sys-
tems examined. This compares to apparent kcat/kuncat
values between 1400 and 40,000 seen with the phenyl
hapten 7.9,13 Importantly, out of the three separate
studies using a total of ®ve rabbits and 10 mice immunized
with hapten 7 that have now been completed, catalytic
activity was observed in each case, while the benzyl
haptens 3, 4, 15, and 17 failed to elicit catalytic activity
after ®ve immunizations in any of the rabbits used. At
It can be argued from the negative catalytic results for
haptens 1±6 that a change associated with altered leaving

A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
419
this time, it is not clear why the 4-chlorobenzyl hapten
16 should elicit catalysts in the rabbits, while the 4-
nitrobenzyl (15) and underivatized benzyl (17) haptens
did not.
Mechanistic considerations
One possible mechanistic explanation for the di€erences
seen between the successful and less successful hapten
systems involves a di€erence in rate-limiting step for
phenyl carbonate substrates such as 14 compared to the
benzyl carbonate substrates such as 20±22. The pheno-
late anion is a better leaving group than a benzyloxy
anion, judging from the rate constants for the hydroly-
ses of the benzyl carbonate substrates measured in the
absence of antibody. The benzyl carbonates 20±22
Figure 1. pH Dependence of the hydrolysis reactions of carbonates 22
and 23. The open squares represents points for carbonate 23, the dar-
kened diamonds represent points for carbonate 22. The pH indepen-
dence of hydrolysis for carbonate 23 was taken as evidence for a
mechanism other than BAC2. Substrates 20 and 21 behaved in a man-
ner analogous to 22, while substrate 24 behaved in a manner analo-
gous to 23. These were omitted for clarity.
5
hydrolyzed with rate constants ranging from 3.8Â10
1
5
min
to 2.7Â10
min 1, while the hydrolysis rate
1
constant for the phenyl carbonate 14 was 5Â10 5 min
when measured at the same pH. Thus, even though
these rate di€erences are small, it is theoretically possi-
ble that for a phenyl carbonate such as 14, formation of
the tetrahedral intermediate could be rate-limiting,
while the more sluggish benzyloxy leaving groups could
change the rate-limiting step to breakdown of the tetra-
hedral intermediate for 20±22. Such a change in rate-
limiting step could explain di€erences in behavior
between antibodies elicited by phenyl phosphates versus
benzyl phosphates. Yet, the 18O labeling studies found
no incorporation of 18O into substrate 20 even after
extended incubation, sucient for substantial hydroly-
sis, with or without antibody. Thus, there is no reason to
believe that the rate-limiting step of the hydrolysis
reaction has changed to breakdown of the tetrahedral
intermediate for the benzyl systems, at least for 20,
either in solution or in the antibody binding pocket. Of
course, there is no way to rule out more subtle di€er-
ences in mechanism such as the detailed geometry of the
rate-limiting transition state that may be di€erent
between phenyl and benzyl carbonates, leading to dif-
ferences in antibody catalysis.
immune responses were of signi®cantly lower hapten
anity compared to those seen with the rabbits. In
addition, with haptens 15 and 17, the rabbits displayed
high hapten anity, but no catalysis after ®ve immuni-
zations in all of the rabbits used. Both species of mice
produced moderately active catalysts with these hap-
tens. We were unable to analyze catalytic activity in
individual mice because of small serum volumes, so it is
currently unknown how catalytic activity varied in the
di€erent mice.
Flexibility
While other explanations such as pure serendipity (pre-
sumably unlikely given the large number of experiments
described here) or subtle changes in transition-state
structure cannot and should not be ruled out, an intri-
guing idea is that ¯exibility in the region of the hapten
opposite the linker decreases e€ectiveness in the less
successful haptens 1±6 and 15±17. An aspect of ¯exible
systems that could decrease catalytic antibody activity is
that antibody-bound conformations of the ¯exible hap-
tens may be di€erent than any relatively low-energy
conformation obtainable by a ¯exible substrate. This
appears to be the case for at least the rabbit antibodies
elicited by haptens 15 and 17, since these antibodies
displayed no binding to their most homologous sub-
strates within the limits of the competition ELISA.
Interestingly, these non-catalytic rabbit antibodies to 15
and 17 were of higher anity to the immunizing hapten
than the catalytic responses generated in mice, possibly
indicating that the higher anity antibodies are less
`forgiving.' Nevertheless, other non-catalytic polyclonal
samples in our studies did bind substrate, indicating
that there is likely more to be considered than a lack of
substrate binding.
Another mechanistic question that deserves comment
involves substrates 23 and 24. The lack of pH depen-
dence for the hydrolysis of substrates 23 and 24 in
the absence of antibody (see Fig. 1) indicates that
these substrates react by a mechanism di€erent from
the BAC2 mechanism presumed for the other carbonate
substrates of this study. It is gratifying that no cataly-
sis was observed with these substrates for any of the
antibodies elicited in BALB/cJ mice by 15±19. Inter-
estingly, the antibodies elicited by haptens 18 and 19
were catalytic for hydrolysis of substrates 20±22, the
substrates that showed the pH dependence of reaction
expected for the BAC2 mechanism. Thus, the impor-
tance of transition state stabilization has been clearly
identi®ed within this system of haptens and sub-
strates.
Flexibility of the hapten could conceivably disrupt cat-
alysis of a bound substrate in other ways. For example,
a ¯exible hapten could elicit an antibody binding pocket
that can bind substrates, but in a ground state con-
formation that cannot proceed to transition state(s)
Di€erent species
The di€erences seen between species immunized with
haptens 15±17 are interesting. In general, the murine

420
A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
and/or products due to conformational constraints. For
example, ¯exible haptens may `collapse' due to hydro-
phobic factors, etc., into low energy structures that are
then recognized by antibodies. These collapsed con-
formations recognized by the antibodies may not be
compatible with catalysis along the entire reaction
coordinate. Note that these ¯exibility arguments may ®t
nicely with those presented by Linder et al.29 in which
rearrangment of the binding pocket to an `active' form
was required for catalysis in a series of benzyl ester
hydrolyzing antibodies.
stirred at 20 ꢀC for 5 min. The reaction mixture was
allowed to warm to 5 ꢀC and stirred for an additional
20 min followed by stirring at room temperature for
20 min.
Methyl-4-hydroxymethylbenzoate
(0.25 g,
1.54 mmol) dissolved in 2 mL DMF was added and the
reaction was allowed to warm to room temperature
where it was stirred for 4 h. Saturated sodium bicarbo-
nate (10 mL) was added and the reaction mixture was
washed twice with diethyl ether. The resulting aqueous
layer was puri®ed by ion anity chromatography on
DEAE Sephadex A-25, using 100 mM potassium chlor-
ide as eluent. The aqueous solvents were removed in
vacuo and the resulting white solid was puri®ed on a
short column of silica gel, eluting with 10% methanol in
ethanol. The solvents were removed on the rotary evap-
orator, resulting in a white solid. This solid, N-CBZ-1-
aminoethylphosphonic acid, (4-methoxycarbonyl)-ben-
zyl ester, potassium salt (0.105 g, 0.26 mmol) was dis-
solved in water/THF/methanol. The ratio of solvents
was adjusted until the solution was homogeneous. One
equivalent of lithium hydroxide (0.012 g, 0.26 mmol)
was added and the reaction stirred at 60 ꢀC for 4 h. The
solvents were removed on the rotary evaporator. The
resulting crude product was puri®ed by reversed-phase
liquid chromatography, eluting with a 1-h gradient from
100% water to 100% acetonitrile. The solvents were
Conclusion
Antibody catalysis is a complex phenomenon, and stud-
ies with numerous haptens, such as those described
here, are required to establish important trends that
may not be obvious from any one study. Taken together,
our data indicate that some haptens produce immune
responses that are reproducibly highly catalytic, while
seemingly similar haptens produce immune responses
that are reproducibly less catalytically active. Whatever
the explanation, subtle changes in mechanistic features
due to altered leaving group ability/location or overall
hapten ¯exibility, the comprehensive data presented
here indicate that phenyl or 4-nitrophenyl leaving
groups located opposite the hapten linker are to be
preferred in order to elicit highly active antibody cata-
lysts for acyl hydrolysis reactions. This work comple-
ments our previous studies which demonstrated that
large, hydrophobic groups should be avoided when
designing transition state analogue haptens,13 and that
homology between hapten and substrate is important.12
Future studies will hopefully validate and re®ne these
emerging rules for the production of optimized transi-
tion state analogue haptens.
1
removed in vacuo to yield a white powder. H NMR
(300 MHz, D2O) dH 1.30 (3H, dd, J=17.9, 8.5 Hz,
CHCH3), 3.85 (1H, m, CHCH3), 4.95 (4H, complex
multiplet, both ArCH2), 7.35, (5H, m, C6H5O), 7.41
(2H, d, J=12.6 Hz, ArH), 7.85 (2H, d, J=9.5 Hz, ArH);
13C NMR (300 MHz, D2O) dC 11.86, 16.13, 52.57,
66.99, 67.07, 126.42, 127.25, 127.46, 128.58, 129.98,
130.36, 139.75, 142.20, 142.30, 156.92, 166.88; high-
resolution FABMS calcd for C18H18NO7P: 391.0821,
found 391.0204; mp > 250 ꢀC (uncorr).
N-Methoxycarbonyl-1-aminoethylphosphonic acid, (4-
carboxy)benzyl ester (3 and 4). Both enantiomers were
prepared according to the following procedure. 1-Amino-
ethylphosphonic acid (0.33 g, 2.64 mmol) was dissolved
in an excess of 10% Na2CO3 (aq). An excess of methyl
chloroformate (0.38 g, 4 mmol) in 2 mL of tetra-
hydrofuran (THF) was added and the reaction was
allowed to stir at room temperature for 4 h. Vigorous
bubbling of the solution was observed. The excess
methylchloroformate and solvents were removed in
vacuo resulting in a crystalline, white solid. N-methoxy-
carbonyl-1-aminoethylphosphonic acid, disodium salt
(0.6 g, 2.64 mmol) was dissolved in 9 mL of N,N-di-
methylformamide (DMF), blanketed with argon, and
chilled to 20 ꢀC. Thionyl chloride (0.38 g, 3.17 mmol)
was added via syringe and the reaction stirred at 20 ꢀC
for 5 min. The reaction mixture was allowed to warm to
5 ꢀC and stirred for an additional 20 min followed by
stirring at room temperature for 20 min. Methyl-4-
hydroxymethylbenzoate (0.44 g, 2.64 mmol) dissolved in
2 mL DMF was added and the reaction was allowed to
warm to room temperature where it was stirred for 4 h.
Saturated sodium bicarbonate (10 mL) was added and
the reaction mixture was washed twice with diethyl
ether. The resulting aqueous layer was puri®ed by ion
anity chromatography on DEAE Sephadex A-25,
Experimental
Syntheses
N-CBZ-1-aminoethylphosphonic acid, (4-carboxy)benzyl
ester (1 and 2). Both enantiomers were prepared
according to the following procedure. 1-aminoethyl-
phosphonic acid (0.5 g, 4 mmol) was dissolved in 10 mL
of 10% Na2CO3 (aq). An excess of benzylchloroformate
(1.02 g, 6 mmol) in 3 mL of tetrahydrofuran (THF) was
added and the reaction was allowed to stir at room
temperature for 4 h. Vigorous bubbling of the solution
was observed. The excess benzylchloroformate and
impurities were removed by extraction into diethyl ether
and the resulting aqueous layer acidi®ed to pH 2 with
hydrochloric acid. The aqueous layer was extracted
three times with diethyl ether. The combined organic
extracts were dried over sodium sulfate, ®ltered, and the
solvents removed on the rotary evaporator to yield col-
orless crystals of N-CBZ-1-aminoethylphosphonic acid.
A portion of the acid (0.4 g, 1.54 mmol) was dissolved in
9 mL of N,N-dimethylformamide (DMF), blanketed
with argon, and chilled to 20 ꢀC. Thionyl chloride
(0.38 g, 3.17mmol) was added via syringe and the reaction

A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
421
using 100 mM potassium chloride as eluent. The aqu-
eous solvents were removed in vacuo and the resulting
white solid was puri®ed on a short column of silica gel,
eluting with 10% methanol in ethanol. The solvents
were removed on the rotary evaporator, resulting in a
white solid. This salt, N-methoxycarbonyl-1-amino-
ethylphosphonic acid, (4-methoxycarbonyl)benzyl ester,
potassium salt (0.48 g, 1.3 mmol) was dissolved in water/
THF/methanol. The ratio of solvents was adjusted until
the solution was homogeneous. One equivalent of
lithium hydroxide (0.055 g, 1.3 mmol) was added and
the reaction stirred at 60 ꢀC for 4 h. The solvents were
removed on the rotary evaporator. The resulting crude
product was puri®ed by reversed-phase liquid chroma-
tography, eluting with a 1 h gradient from 100% water
to 100% acetonitrile. The solvents were removed in
2 h. The reaction mixture was poured into saturated
NaCl and washed twice. The organic layer was dried
over sodium sulfate, ®ltered, and the solvents removed
on the rotary evaporator (procedure in ref 8) The crude
phosphates were puri®ed on silica gel, eluting with 20%
ethyl acetate:80% hexane to yield a colorless oil. 1.0 M
tetrabutyl ammonium ¯uoride in THF (1.1 equiv,
according to the weight of the phosphate recovered) was
added to a solution of the puri®ed phosphate and
allowed to stir at room temperature until bubbling
ceased. To this solution was added one equivalent of
LiOH in 1:1 methanol:water and the reaction stirred
overnight at room temperature. Solvents were removed
1
to yield the hapten as a light-yellow solid. H NMR
(250 MHz, D2O) dH 0.70 (3H, t, J=8.6 Hz, CH3-
CH2CH), 0.81 (12H, t, J=8.6 Hz, CH3CH2CH2CH2),
1.23 (8H, m, J=11.5 Hz, CH3CH2CH2CH2), 1.49 (8H,
m, CH3CH2CH2CH2), 1.75 (2H, m, J=5.7 Hz, CH3-
CH2CH), 3.04 (8H, t, CH3CH2CH2CH2), 4.90 (1H, q,
J=8.6 Hz, CH3CH2CH), 6.75 (2H, d, J=11.5 Hz,
ArH), 6.95 (2H, d, J=9.8 Hz, ArH), 7.24 (5H, m,
C6H5); 13C NMR (300 MHz, D2O) dC 9.55, 14.0, 20.2,
24.7, 33.2, 41.2, 59.5, 70.0, 121.1, 121.3, 126.2, 127.2,
127.6, 128.5, 128.9, 130.1; high-resolution FABMS
calcd for C33H53NO6PLi: 597.3770, found 597.7062; mp
>280 ꢀC (uncorr).
1
vacuo to yield a white powder. H NMR (300 MHz,
D2O) dH 1.34 (3H, dd, J=16.7, 7.3 Hz, CHCH3), 3.80
(3H, s, CH3O), 3.85 (1H, m, CHCH3), 5.06 (2H, d,
J=7.22 Hz, ArCH2), 7.38 (2H, d, J=9.81 Hz, ArH),
7.93 (2H, d, J=8.26 Hz, ArH); 13C NMR (300 MHz,
D2O) dC 11.23, 16.13, 52.32, 52.57, 66.99, 67.07, 127.46,
129.98, 130.36, 142.20, 142.30, 156.83, 156.92, 166.88;
high-resolution FABMS calcd for C12H14 NO7P:
315.0508, found 315.0996; mp >250 ꢀC (uncorr).
O-1-Phenylpropyl-O-(4-carboxymethylphenyl)-phosphate,
tetrabutylammonium salt, lithium salt (5 and 6). Both
enantiomers were prepared using a procedure adopted
from Martin and Josey30 using (trimethylsilyl)ethyl
phosphite, prepared according to the method of Poulter
et al.31 7.55 g (55 mmol) of PCl3 was dissolved in 60 mL
dry ether and chilled to 78 ꢀC under argon. 5 g
(42 mmol) of 2-(trimethylsilyl)ethanol in 80 mL dry
ether were added slowly using a cannula. The reaction
was stirred at 78 ꢀC for 1 h and then allowed to come
to room temperature. The solvents were removed at
reduced pressure on the rotary evaporator, care being
taken to exclude moisture. The clear residue was dis-
tilled under high vacuum. The fraction boiling at 38±
40 ꢀC was collected in a chilled ¯ask ( 78 ꢀC). The ¯ask
was sealed with a septum and blanketed with argon.
The clear, liquid 2-(trimethylsilyl)ethyl dichloropho-
sphite was stored at 4 ꢀC 1.15 g (5.65 mmol) 2-(tri-
methylsilyl)ethyl dichlorophosphite were added to a dry
¯ask using a tared syringe. Dry THF was added and the
solution chilled to 78 ꢀC under argon. 1.83 g (14 mmol)
of dry diisoproplyethylamine were added to the solution
followed by 0.83 g (5 mmol) of methyl-4-hydroxy-
phenylacetate dissolved in dry THF. A white precipitate
formed. The reaction was stirred for 10 min. 0.77 g
(5.65 mmol) of 1-phenyl-1-propanol dissolved in dry
THF were added. The reaction was stirred at 78 ꢀC for
2 h, removed from the cooling bath, and stirred for an
additional hour. Solvents were removed on the rotary
evaporator and the residue taken up in ethyl acetate.
This solution was ®ltered through Celite 545, dried over
sodium sulfate, ®ltered, and the solvents removed on the
rotary evaporator. The resulting crude phosphite was
puri®ed on silica gel (hexane:ethyl acetate, 80:20). The
phosphite was dissolved in dichloromethane and 0.29 g
(1 mL, 8.48 mmol) of 30% H2O2 was added slowly. The
reaction was allowed to stir at room temperature for
O-4-Carboxybenzyl-O-phenylphosphate, lithium salt, tetra-
butyl ammonium salt (7). The procedure above was
followed using phenol and methyl-4-hydroxymethyl-
benzoate as the two alcohols added to the 2-(tri-
methylsilyl)ethyl dichlorophosphite to yield the hapten
as a white solid. 1H NMR (300 MHz, CD3OD) dH
1.02 (12H, t, J=9 Hz, CH3CH2CH2CH2) 1.38 (8H, sex-
tet, J=9 Hz, CH3CH2CH2CH2), 1.64 (8H, quintet,
J=10 Hz, CH3CH2CH2CH2), 3.23 (8H, t, J=12 Hz,
CH3CH2CH2CH2), 5.00 (2H, d, J=9 Hz, ArCH2), 7.00±
7.26 (5H, m, C6H5), 7.32 (2H, d, J=9 Hz, ArH), 7.89
(2H, d, J=9 Hz, ArH); 13C NMR (300 MHz, CD3OD)
dC 13.98, 20.64, 24.72, 59.38, 68.22, 68.28, 121.23,
121.30, 124.11, 127.51, 130.16, 130.30, 138.25, 141.45,
141.50, 154.46, 154.51, 174.73; high-resolution FABMS
calcd for C14H11O6PLi: 313.0453, found 313.0443; mp
>280 ꢀC (uncorr).
N-(2-diethylaminoethyl)-4-hydroxymethylbenzamide. To
a stirred solution of 4-hydroxymethyl benzoic acid
penta¯uorophenyl ester (Sigma) (1.5 g, 4.7 mmol) in
dichloromethane at 0 ꢀC were added 0.66 g (5.7 mmol)
of 2-(diethylamino)ethylamine. The resulting solution
was stirred at 0 ꢀC for 30 min and allowed to warm to
room temperature for an additional 30 min. The reac-
tion mixture was washed three times with 1 N NaOH,
once with saturated NaCl, dried over sodium sulfate,
and ®ltered. The solvents were removed on the rotary
evaporator and the resulting pale, brown oil was placed
under high vacuum overnight. The oil was chromato-
graphed on silica (10% methanol:methylene chloride) to
1
yield a pale-yellow oil. H NMR (300 MHz, CD2Cl2)
dH 1.01 (6H, t, J=7.2 Hz, CH3CH2), 2.57 (4H, q,
J=7.2 Hz, CH3CH2), 2.65 (2H, t, J=6.1 Hz, NCH2
CH2NH), 3.42 (2H, q, J=5.8 Hz, NCH2CH2NH), 4.65
(2H, s, ArCH2), 7.25 (2H, d, J=8.1 Hz, ArH), 7.65 (2H,

422
A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
d, J=7.8 Hz, ArH); 13C NMR (300 MHz, CD2Cl2) dC
11.62, 37.64, 47.04, 51.97, 63.89, 126.77, 127.25, 133.55,
145.63, 167.70; high-resolution MS (CI+) calcd for
C14H23N2O2 (MH+): 251.1760, found 251.1764.
ethyl acetate. The combined organic extracts were dried
over sodium sulfate, ®ltered, and the solvents removed
on the rotary evaporator to yield a colorless oil. The N-
methoxycarbonylalanine (0.5 g, 3.4 mmol) was dissolved
in DMF/dichloromethane and HOBt (0.46 g, 3.4 mmol)
in DMF was added followed by DCC (0.7 g, 3.4 mmol).
The reaction was stirred under argon for 1 h at room
temperature. A solution of N-(2-diethylaminoethyl)-4-
hydroxymethylbenzamide (0.85 g, 3.4 mmol) in dichloro-
methane was added. The reaction mixture was allowed
to stir at room temperature for four hours and then ®l-
tered to remove the precipitated urea. The solvents were
removed by vacuum distillation and the resulting resi-
due extracted with ethyl acetate. The combined ethyl
acetate layers were dried over sodium sulfate, ®ltered,
and the solvents removed on the rotary evaporator. The
substrate was puri®ed by column chromatography on
silica gel using 2% methanol in dichloromethane as
eluent, giving a colorless oil. 1H NMR (300 MHz,
CD2Cl2) dH 1.01 (6H, t, J=7 Hz, CH3CH2), 1.41 (3H,
d, J=7.2 Hz, CH3CH), 2.57 (4H, q, J=5.7 Hz,
CH3CH2), 2.65 (2H, t, J=6 Hz, NCH2CH2NH), 3.42
(2H, q, J=6 Hz, NCH2CH2NH), 3.64 (3H, s, CH3O),
4.40 (1H, m, CH3CH), 5.22 (2H, s, ArCH2), 5.39, (1H,
d, J=7 Hz, NH), 7.25 (2H, d, J=8 Hz, ArH), 7.65 (2H,
d, J 8.2 Hz, ArH); 13C NMR (300 MHz, CD2Cl2) dC
11.62, 18.26, 37.64, 47.04, 49.37, 51.97, 52.38, 63.89,
126.77, 127.25, 133.55, 145.63, 156.61, 167.70, 177.03;
High-resolution MS (CI+) calcd for C19H29N3O5:
379.2107, found 379.4600.
N-CBZ-alanine, N-(2-diethylaminoethyl)-4-hydroxymethyl-
benzamide ester (8 and 9). Both enantiomers were pro-
duced as follows. Alanine (3 g, 34 mmol) was dissolved
in an excess of 10% Na2CO3 (aq). An excess of benzyl-
chloroformate in tetrahydrofuran (THF) was added and
the reaction was allowed to stir at room temperature for
4 h. Vigorous bubbling of the solution was observed.
The excess benzylchloroformate and impurities were
removed by extraction into ethyl acetate and the result-
ing aqueous layer dried in vacuo, resulting in a crystal-
line, white solid. The solid was dissolved in water and
the resulting solution made acidic with 0.1 M HCl and
extracted three times with ethyl acetate. The combined
organic extracts were dried over sodium sulfate, ®ltered,
and the solvents removed on the rotary evaporator
to yield a white solid. The N-CBZ-alanine (0.52 g,
23.5 mmol) was dissolved in DMF/dichloromethane and
4-hydroxybenzotriazole (HOBt) (0.32 g, 23.5 mmol) in
DMF was added followed by dicyclohexylcarbodiimide
(DCC) (0.49 g, 23.5 mmol). The reaction was stirred
under argon for 1 h at room temperature. A solution of
N-(2-diethylaminoethyl)-4-hydroxymethylbenzamide
(0.59 g, 23.5 mmol) in dichloromethane was added. The
reaction mixture was allowed to stir at room tempera-
ture for 4 h and then ®ltered to remove the precipitated
dicyclohexylurea. The solvents were removed by
vacuum distillation and the resulting residue extracted
with ethyl acetate. The combined ethyl acetate layers
were dried over sodium sulfate, ®ltered, and the solvents
removed on the rotary evaporator. The substrate was
puri®ed by column chromatography on silica gel using
2% methanol in dichloromethane as eluent to yield a
(4-(N-Butyl) carboxamidomethyl) phenyl (1-phenyl)-1-
propyl carbonate (12 and 13). Both enantiomers were
prepared as follows. 4-Nitrophenylchloroformate (1.9 g,
6.3 mmol) was dissolved in dry dichloromethane and
blanketed with argon while being chilled to 0 ꢀC in an
ice/water bath. 0.51 mL (6.3 mmol) of dry pyridine was
added via syringe and a precipitate formed. This was
followed by the dropwise addition of 0.77 g (5.65 mmol)
of 1-phenyl-1-propanol. The reaction mixture cleared.
The reaction was stirred for 4 h while warming to room
temperature. During this time a precipitate formed. The
solvents were removed in vacuo. The crude reaction
mixture was chromatographed on silica gel (CH2Cl2)
yielding a tan colored solid. N-Butyl-4-hydroxyphenyl-
acetamide (0.43 g, 2.2 mmol) was dissolved in dry THF
and blanketed with argon while chilling to 0 ꢀC. 0.053 g
(2.2 mmol) of NaH was added. The reaction mixture
bubbled vigorously. After the bubbling subsided, 0.66 g
(2.2 mmol) 1-phenyl-1-propyl 4-nitrophenyl carbonate
in THF was added via syringe and the reaction mixture
allowed to warm to room temperature over 2 h, during
which time the reaction mixture turned bright yellow.
The solvents were removed on the rotary evaporator
and the yellow residue was dissolved in ethyl acetate
and washed with 1% NaOH until the aqueous extracts
were colorless. The organic layer was washed with water
followed by saturated NaCl. The organic layer was
dried over sodium sulfate, ®ltered, and the solvents
removed on the rotary evaporator. The resulting oil was
chromatographed on silica gel (2% methanol/dichloro-
1
thick, tan colored oil. H NMR (300 MHz, CD2Cl2) dH
1.01 (6H, t, J=7.2 Hz, CH3CH2), 1.41 (3H, d,
J=7.2 Hz, CH3CH), 2.54 (4H, q, J=5.7 Hz, CH3CH2),
2.61 (2H, t, J=6.0 Hz, NCH2CH2NH), 3.43 (2H, q,
J=5.7 Hz, NCH2CH2NH), 4.39, (1H, m, CH3CH), 5.07
(2H, s, ArCH2), 5.18 (2H, s, ArCH2), 6.17 (1H, d,
J=7.5 Hz, NH), 7.25 (1H, m, NH), 7.32 (5H, m, C6H5),
7.37 (2H, d, J=8.1 Hz, ArH), 7.78 (2H, d, J=8.2 Hz,
ArH); 13C NMR (300 MHz, CD2Cl2) dC 13.5, 18.8,
38.7, 46.6, 50.2, 52.3, 67.5, 68.3, 127.0, 127.5, 128.2,
128.7, 129.0, 129.4, 135.1, 139.9, 163.5, 166.3, 174.7;
high-resolution MS (CI+) calcd for C25H33N3O5:
455.2420, found 455.5589.
N-Methoxycarbonylalanine, N-(2-diethylaminoethyl)-4-
hydroxymethylbenzamide ester (10 and 11). Both enan-
tiomers were produced as follows. Alanine (2 g,
22 mmol) was dissolved in an excess of 10% Na2CO3
(aq). An excess of methyl chloroformate (4.2 g,
45 mmol) in THF was added and the reaction was
allowed to stir at room temperature for 4 h. Vigorous
bubbling of the solution was observed. The excess
methylchloroformate and solvents were removed in
vacuo resulting in a crystalline, white solid. The solid
was dissolved in water and the resulting solution made
acidic with 0.1 M HCl and extracted three times with
1
methane) to yield a colorless oil. H NMR (300 MHz,
CD2Cl2) dH 0.92 (6H, complex multiplet, CH3), 1.39

A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
423
(2H, m, J=6.9 Hz, CH3CH2CH2CH2), 1.58 (2H, m,
J=7.1 Hz, CH3CH2CH2CH2), 2.00 (2H, complex multi-
plet, CHCH2CH3), 3.41 (2H, q, J=7.0 Hz, CH3CH2
CH2CH2), 3.61 (2H, s, ArCH2), 5.6 (1H, t, J=7.4 Hz,
CHCH2CH3), 6.20 (1H, s, NH), 7.09 (2H, d, J=9.5 Hz,
ArH), 7.25 (2H, d, J=9.7 Hz, ArH), 7.40 (5H, m,
C6H5); 13C NMR (300 MHz, CD2Cl2) dC 9.97, 29.70,
40.61, 52.28, 82.82, 121.48, 126.85, 128.60, 128.85,
130.70, 132.50, 140.1, 150.7, 154.0, 172.0; high-resolu-
tion MS calcd for C21H26NO5 (MH+): 372.1733, found
372.4371.
4.92 (4H, m, ArCH2), 7.33 (2H, m, ArH), 7.55 (2H, m,
ArH), 7.88 (2H, m, ArH), 8.15 (2H, m, ArH); 13C NMR
(250 MHz D2O) dC 67.0 (d, J=6 Hz); 68.0 (d, J=6 Hz);
124.3; 127.8; 128.0; 128.4; 129.7 (d, J=6 Hz); 136.5;
140.7; 145.8; 175.7; high-resolution FABMS calcd for
C15H13NO8P: 366.0379, found 366.0364.
4-(Carboxy)phenylmethyl 4-chlorophenylmethyl phos-
phate, dipotassium salt (16). The procedure described
above for 15 was followed using 4-chlorobenzyl alcohol
in place of the 4-nitrobenzyl alcohol, yielding the hapten
1
as a white solid. H NMR (250 MHz, CD3OD) dH 4.87
4-(N-(2-(Diethylamino)ethyl)carbamoyl)benzyl phenyl
carbonate (14). N-(2-Diethylaminoethyl)-4-hydroxy-
methylbenzamide (0.48 g, 1.93 mmol) was dissolved in
dry dichloromethane and blanketed with argon while
being chilled to 0 ꢀC in an ice/water bath. 0.15 mL
(1.93 mmol) of dry pyridine was added via syringe fol-
lowed by the dropwise addition of 0.28 mL (2.31 mmol)
of phenyl chloroformate. The reaction was stirred for
4 h while warming to room temperature. The reaction
mixture was then washed three times with 0.1 M NaOH
followed by one wash with saturated NaCl. The organic
phase was dried over sodium sulfate, ®ltered, and the
solvents removed at reduced pressure. The crude car-
bonate was chromatographed on silica gel using 2%
methanol in dichloromethane as eluent. The resulting
pale brown oil was stored in vacuo overnight: 1H NMR
(300 MHz, CD2Cl2) dH 0.92 (6H, t, J=7.2 Hz, CH3-
CH2), 2.55 (4H, q, J=7.0 Hz, CH3CH2), 2.54 (2H, t,
J=6.2 Hz, NCH2CH2NH), 3.36 (2H, q, J=5.7 Hz,
NCH2CH2NH), 5.17 (2H, s, ArCH2), 7.07 (2H, d,
J=7.6 Hz, ArH), 7.24 (5H, m, C6H5) 7.45 (2H, d,
J=8.1 Hz, ArH); 13C NMR (300 MHz, CD2Cl2) dC
11.97, 37.71, 47.10, 51.80, 69.78, 121.35, 126.42, 127.60,
128.42, 129.79, 135.42, 138.43, 151.47, 153.86, 166.80;
HRMS (CI+) calcd for C21H27N2O4 (MH+): 371.1971,
found 371.1963.
(4H, m, ArCH2), 7.33 (6H, m, ArH), 7.93 (2H, d,
J=4 Hz, ArH); 13C NMR (250 MHz, D2O) dC 67.3
(d, J=5 Hz); 68.0 (d, J=5 Hz); 127.9; 129.2; 129.8;
130.0; 134.0; 136.6; 140.8; 141.0; 175.9; high-resolution
FABMS for C15H13O6PCl: calcd 355.0138, found
355.0115.
4-(Carboxy)phenylmethyl phenylmethyl phosphate, dipo-
tassium salt (17). The procedure described for 15 was
followed using benzyl alcohol in place of the 4-nitro-
benzyl alcohol, a€ording the hapten as a white solid. 1H
NMR (250 MHz, CD3OD) dH 4.89 (4H, m, ArCH2),
7.32 (7H, m, ArH), 7.92 (2H, m, ArH); 13C NMR
(250 MHz, D2O) dC 68.0 (d, J=6 Hz); 69.0 (d, J=5 Hz);
128.1; 128.5; 129.0; 129.4; 129.8; 136.7; 141.0 (d, J= 7 Hz);
176.0; high-resolution MS (CI+) calcd for C15H14O6P:
321.0528, found 321.0542.
4-(Carboxy)phenylmethyl 4-(methoxy)phenylmethyl phos-
phate, dipotassium salt (18). The procedure described
for 15 was followed using 4-methoxybenzyl alcohol
instead of 4-nitrobenzyl alcohol. The resulting hapten
was isolated as a white solid. 1H NMR (250 MHz,
CD3CDOD) dH 3.74 (3H, s, ArOCH3), 4.85 (4H, m,
ArCH2), 6.79 (2H, m, ArH), 7.29 (4H, m, ArH), 7.95
(2H, m, ArH); 13C NMR (250 MHz, D2O) dC 56.2; 67.7
(d, J=6 Hz); 68.2 (d, J=6 Hz); 114.7; 127.9; 128.3;
129.8 (d, J=6 Hz); 130.5; 136.1 (d, J=6 Hz); 140.9 (d,
J=6 Hz); 159.2; 176.1; high-resolution FABMS calcd
for C16H16O7P: 351.0634, found 351.0670.
4-(Carboxy)phenylmethyl 4-nitrophenylmethyl phosphate,
dipotassium salt (15). The procedure for making
haptens 5 and 6 was followed, using 4-nitrobenzyl alco-
hol and methyl 4-(hydroxymethyl)benzoate as the two
alcohols added to the 2-(trimethylsilyl)ethyl dichloro-
phosphite, excepting that after removal of the silyl pro-
tecting group with tetrabutyl ammonium ¯uoride, the
resulting residue was dissolved in H2O and loaded onto
DEAE Sephadex A-25 anion exchange resin. The resin
was washed with H2O and then the phosphate was
eluted with 150 mM KCl. The H2O was lyophilized to
leave behind a copious amount of white powder. A
mixture of 10% methanol in ethanol was added to the
powder to make a slurry which was ®ltered through
silica gel to remove excess KCl. The concentrated ®l-
trate was taken up in ethanol and ®ltered through a
fritted glass funnel to remove any silica gel which may
have dissolved in the methanol/ethanol mixture. The
residue left after concentration of this second ®ltrate
was dissolved in 30 mL methanol, and 1 M aq KOH
(0.892 mL, 0.892 mmol) was added. The resulting mix-
ture was stirred at room temperature overnight. Solvent
was removed by rotary evaporator to a€ord the hapten
4-(Carboxy)phenylmethyl 4-(ethoxy)phenylmethyl phos-
phate, lithium salt, tetrabutylammonium salt (19). The
procedure described for synthesizing haptens 5 and 6
was followed, using 4-ethoxybenzyl alcohol and methyl
4-(hydroxymethyl)benzoate as the two alcohols added
to the 2-(trimethylsilyl)ethyl dichlorophosphite, yielding
the product as a white solid. 13C NMR (250 MHz,
CD3OD) dC 59.4, 64.4, 67.8 (d, J=6 Hz), 68.0 (d,
J=6 Hz), 115.2, 127.5, 130.2, 130.4, 131.5, 138.2, 141.9,
160.0, 174.4; high-resolution FABMS calcd for C17H18-
O7P1: 365.0790, found 365.0816.
N-[20-(Dimethylamino)ethyl] 4-(hydroxymethyl)benzamide.
4-(Hydroxymethyl)benzoic acid penta¯uorophenyl ester
(8.0 g, 25.12 mmol) was dissolved in a minimal amount
of dry dichloromethane. N,N-Diethylethylene diamine
(2.516 g, 21.68 mmol) was added, and the mixture was
stirred at room temperature for 48h. The reaction mix-
ture was washed with 5% aq NaOH and then with brine.
The organics were dried over sodium sulfate, ®ltered
1
as a olive green solid. H NMR (250 MHz, CD3OD) dH

424
A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
and concentrated in vacuo. The resultant brown oil was
puri®ed by ¯ash chromatography (gradient elution with
CH2Cl2:MeOH, 49:1! CH2Cl2:MeOH, 9:1! CH2Cl2:
MeOH:triethylamine, 87:10:3) to provide the benzamide
The cloudy orange mixture was vacuum ®ltered through
a fritted glass funnel, and the yellow ®ltrate was con-
centrated in vacuo. Puri®cation by ¯ash chromato-
graphy (gradient elution CH2Cl2:EtOH, 49:1! CH2Cl2:
EtOH, 23:2) and chromatatron chromatography
(CH2Cl2:EtOH, 49:1!CH2Cl2:EtOH, 23:2) a€orded the
carbonate as a yellow oil. 1H NMR (250 MHz, CD3CN)
dH 1.03 (6H, m, CH2CH3), 2.63 (6H, m, NCH2), 3.42
(2H, m, CONHCH2), 5.12 (2H, s, ArCH2), 5.18 (2H, s,
ArCH2), 7.39 (6H, m, ArH), 7.79 (2H, m, ArH); 13C
NMR (250 MHz, CD3CN) dC 12.0, 38.1, 47.9, 52.7,
69.5, 69.7, 128.2, 128.8, 129.5, 130.8, 134.8, 135.5, 135.8,
139.9, 155.7, 167.2; high-resolution MS (CI+) calcd for
C22H28ClN304: 419.1738, found 419.1727.
1
as a yellow oil. H NMR (250 MHz, CDCl3) dH 1.01
(6H, m, CH2CH3), 2.61 (6H, m, NCH2), 3.45 (2H, m,
CONHCH2), 4.60 (2H, s, ArCH2OH), 7.23 (2H, d,
J=8 Hz, ArH), 7.51 (1H, s, NH), 7.65 (2H, d, J=8 Hz,
ArH); 13C NMR (250 MHz, CDCl3) dC 11.0, 37.0, 46.9,
51.8, 63.9, 126.61, 127.1, 133.1, 144.8, 167.6; high-
resolution MS (CI+) calculated for C14H23N2O2:
251.1760, found 251.17611.
40 -[N-200 -(Dimethylamino)ethyl]carbamoyl]phenylmethyl
4-nitrophenylmethyl carbonate (20). N-[20-(Dimethyl-
amino)ethyl]
4-(hydroxymethyl)benzamide
(0.30 g,
4 - [N - 20 - (Dimethylamino)ethyl]carbamoyl]phenylmethyl
phenymethyl carbonate (22). N-[20-(Dimethylamino)-
ethyl] 4-(hydroxymethyl)benzamide (0.2286 g, 0.913 mmol)
and NaH (0.0219 g, 0.913 mmol) were dissolved in
45 mL dry THF and stirred at 45 ꢀC for 1 h. The mixture
was cooled to room temperature, and a solution of
benzyl chloroformate (0.156 g, 0.913 mmol) in 10 mL
dry THF was added via syringe. The reaction was stir-
red at room temperature for 36 h. The resulting cloudy
orange solution was vacuum ®ltered through a fritted
glass funnel, and the ®ltrate was concentrated in vacuo.
Puri®cation of the residue by ¯ash chromatography
(gradient elution CH2Cl2:MeOH, 99:1! CH2Cl2:MeOH,
9:1) and chromatatron chromatography (CH2Cl2:MeOH,
1.2 mmol) and 4-nitrobenzyl chloroformate (0.31 g,
1.44 mmol) were dissolved in 1 mL pyridine and heated
to 100 ꢀC for 30 min. A white precipitate formed. The
mixture was cooled to 50 ꢀC and stirred for a further
12 h. During this time more 4-nitrobenzyl chloro-
formate (0.30 g) was added. The mixture was con-
centrated in vacuo, and the residue was puri®ed via
¯ash chromatography (gradient elution CH2Cl2:MeOH,
49:1! CH2Cl2:MeOH, 23:2) to give the crude product
as a brown oil. Further puri®cation by chromatatron
chromatography (gradient elution CH2Cl2! CHCl2:
MeOH, 49:1!CH2Cl2:EtOH, 19:1) a€orded the car-
1
bonate as a clear oil. H NMR (250 MHz, CD3CN) dH
1
1.22 (6H, m, CH2CH3), 3.03 (6H, m, NCH2), 3.67 (2H,
m, CONHCH2), 5.19 (2H, s, ArCH2), 5.25 (2H, s,
ArCH2), 7.43 (2H, d, J=8 Hz, ArH), 7.56 (2H, d,
J=8 Hz, ArH), 7.99 (2H, d, J=8 Hz, ArH), 8.17 (2H, d,
J=8 Hz, ArH), 8.64 (1H, broad s, CONH); 13C NMR
(250 MHz, CD3CN) dC 9.6, 36.4, 48.4, 53.4, 68.9, 69.9,
124.6, 128.7, 128.8, 129.4, 135.3, 140.0, 144.1, 148.5,
155.6, 167.3; high-resolution MS (CI+) calcd for
C22H28N3O6: 430.1978, found 430.1959.
49:1) a€orded the carbonate as a clear oil. H NMR
(250MHz, CD3CN) dH 1.02 (6H, m, CH2CH3), 2.64 (6H,
m, NCH2), 3.44 (2H, m, CONHCH2), 5.15 (2H, s, ArCH2),
5.18 (2H, s, ArCH2), 7.41 (7H, m, ArH), 7.81 (2H, d,
J=8 Hz, ArH); 13C NMR (250 MHz, CD3CN) dC 11.9,
38.1, 47.9, 52.7, 68.6, 70.4, 118.3, 128.2, 128.8, 129.1, 129.4,
129.5, 135.7, 140.0, 155.8, 167.3; high-resolution MS
(CI+) calcd for C22H29N2O4: 385.2127, found 385.2113.
4(-[N-20 -(Dimethylamino)ethyl]carbamoyl]phenylmethyl
4-methoxyphenylmethyl carbonate (23). N-[20-(Dimethyl-
amino)ethyl] 4-(hydroxymethyl)benzamide (0.161 g,
0.644 mmol) and NaH (0.0154 g, 0.644 mmol) were dis-
solved in 40 mL dry THF and stirred at 40 ꢀC for 2 h.
The cloudy yellow mixture was cooled to room tem-
perature, and a solution of 4-methoxybenzyl S-(4,6-di-
methylprimidin-2-yl)thiolcarbonate (0.098 g, 0.32 mmol)
in 20 mL dry THF was added via syringe. The reaction
was stirred at 35 ꢀC for 5 h and then at room tempera-
ture for 3 days. Solvent was removed in vacuo, and the
concentrate was puri®ed via ¯ash chromatography
(gradient elution CH2Cl2:EtOH, 49:1!CH2Cl2:EtOH,
9:1) and chromatatron chromatography (gradient elu-
tion CH2Cl2:EtOH, 49:2! CH2Cl2:EtOH, 23:2) to give
the carbonate as a clear oil. 1H NMR (250 MHz,
CD3CN) dH 1.01 (6H, m, CH2CH3), 2.59 (6H, m,
NCH2), 3.39 (2H, m, CONHCH2), 3.77 (3H, s, OCH3),
5.08 (2H, s, ArCH2), 5.17 (2H, s, ArCH2), 6.91 (2H, m,
ArH), 7.15 (1H, s, NH), 7.32 (2H, m, ArH), 7.42 (2H, d,
J=8 Hz, ArH), 7.76 (2H, d, J=8 Hz, ArH); 13C NMR
(250 MHz, CD3CN) dC 12.3, 38.3, 47.8, 52.6, 55.9, 69.5,
70.3, 114.8, 128.1, 128.8, 131.2, 140.0, 155.8, 160.0, 160.9,
167.2, 176.6; high-resolution MS (CI+) calculated for
C23H31N2O5: 415.2233, found 415.2219.
40 -[N-200 -(Dimethylamino)ethyl]carbamoyl]phenylmethyl
4-nitrophenyl carbonate. N-[20-(Dimethylamino)ethyl] 4-
(hydroxymethyl)benzamide (0.990 g, 3.96 mmol) was
dissolved in a minimal amount of dry dichloromethane.
Pyridine (0.43 mL, 5.2 mmol) and 4-nitrophenyl chloro-
formate (1.04 g, 5.15 mmol) were added, and the reac-
tion was stirred at room temperature for 24 h. The
mixture was concentrated in vacuo and puri®ed by ¯ash
chromatography (gradient elution with CH2Cl2:MeOH,
97:3! CH2Cl2:MeOH, 22:3) to give the product as a
1
white solid. H NMR (250 MHz, CDCl3) dH 1.42 (6H,
m, CH2CH3), 3.21 (6H, m, NCH2), 3.92 (2H, m,
CONHCH2), 5.31 (2H, s, ArCH2), 7.40 (2H, m, ArH),
7.50 (2H, d, J=8 Hz, ArH), 8.15 (2H, d, J=8 Hz, ArH),
8.28 (2H, m, ArH).
4-Chlorophenylmethyl
40-[N-200-(dimethylamino)ethyl]-
carbamoyl]phenylmethyl carbonate (21). 4-Chlorobenzyl
chloroformate (0.142 g, 0.997 mmol) and NaH (0.024 g,
0.997 mmol) were dissolved in 45 mL dry THF and
heated to 50 ꢀC for 1 h. The solution was cooled and 40-
[N-200 -(dimethylamino)ethyl]carbamoyl]-phenylmethyl
4-nitrophenyl carbonate (0.20 g, 0.48 mmol) was added.
The reaction was stirred at room temperature for 2 days.

A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
425
4-Ethoxyphenylmethyl 40[N-20-(Dimethylamino)ethyl]car-
bamoyl]-phenyl-methyl carbonate (24). 4-Ethoxybenzyl
chloroformate (0.159 g, 1.05 mmol) and NaH (0.025 g,
1.05 mmol) were dissolved in 50 mL dry THF and
heated to 50 ꢀC for 1 h. The solution was cooled and 40-
[N-200 -(dimethylamino)ethyl]carbamoyl]-phenylmethyl
4-nitrophenyl carbonate (0.21 g, 0.505 mmol) was
added. The mixture was stirred at room temperature for
6 days. The cloudy orange solution was vacuum ®ltered
through a fritted glass funnel, and the ®ltrate was con-
centrated in vacuo. Puri®cation by ¯ash chromato-
graphy (gradient elution CH2Cl2:EtOH, 49:1!CH2Cl2:
EtOH, 9:1) and chromatatron chromatography (gra-
dient elution CH2Cl2:EtOH, 49:1!CH2Cl2:EtOH, 23:2)
temperature incubation of this mix on an ELISA plate
coated with the immunizing hapten±BSA conjugate.
The absorbance values and their corresponding com-
petitor concentrations were fed into a table using
KaleidagraphTM 3.0 and plotted using the following
1:1 binding curve equation to generate association
constants:
ꢁ
ꢁ
Ao ꢀKA ‰IŠ ‡ KA ‰AbŠ ‡ 1†
q
2
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢀKA ‰IŠ ‡ KA ‰AbŠ ‡ 1†
4 KA ‰AbŠ ‰IŠ
A ˆ
ꢁ
ꢁ
2 KA Ao=‰AbŠ
where Ao=the maximal absorbance of uninhibited
antibody, KA=hapten:antibody equilibrium association
constant, [I]=concentration of competitor, and
[Ab]=concentration of antibody binding sites. The dis-
sociation constants reported in the tables are calculated
as KD ˆ 1=KA.
1
a€orded the product carbonate. H NMR (250 MHz,
CD3CN) dH 1.01 (6H, m, NCH2CH3), 1.34 (3H, m,
OCH2CH3), 2.61 (6H, m, NCH2), 3.40 (2H, m,
CONHCH2), 4.00 (2H, m, OCH2CH3), 5.07 (2H, s,
ArCH2), 5.16 (2H, s, ArCH2), 6.88 (2H, d, J=8 Hz,
ArH), 7.29 (2H, m, ArH), 7.42 (2H, d, J=8 Hz, ArH),
7.78 (2H, m, ArH); 13C NMR (250 MHz, CD3CN)C
12.2, 15.0, 38.3, 47.9, 52.6, 64.3, 69.5, 70.3, 115.3, 118.2,
128.1, 128.5, 128.8, 135.8, 140.0, 155.8, 160.2, 167.2;
high-resolution MS (CI+) calcd for C24H33N2O5:
429.2389, found 429.2392.
H218O incorporation experiments
All H218O incorporation experiments were run with
substrate 20 at a concentration of 125 mM in 1ÂTris±
HCl bu€er (pH 8.0). Antibody, if used, was post-®fth
injection rabbit anti-15 antibody at 6.3 mM. H218O
(95% H218O, Fluka), when used, was 50% of the total
reaction volume. Reactions were carried out in a dessi-
cator at room temperature. Chemical ionization high
resolution mass spectrometry was used to determine the
relative amount of 18O incorporation by comparing the
heights of the peak at 430.2 mass units, the peak corre-
sponding to substrate 20 with all O16, and the peak at
432.2 mass units, corresponding to 20 containing one
18O. The ratio of the two peak heights was compared to
that obtained from substrate 20 samples allowed to
hydrolyze in the absence of H2 18O. Two di€erent reac-
tion times were used for each sample: 20 h, which is the
time required for 10% of the substrate to be converted
to product in the presence of the antibodies, and 66 h,
the time required for 10% of the product to be con-
verted to substrate in the absence of antibody.
Immunogen synthesis
1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC)
(6 mg, 0.032 mmol) and sulfo-NHS (4 mg, 0.019 mmol)
were added to a solution of the hapten in 10 mM PBS.
The concentration of the hapten varied with molecular
weight to ensure approximately equal loading. After
stirring for 10 min, a solution of 10 mg of keyhole
limpet hemocyanin (KLH) in 10 mM PBS was added
and the reaction allowed to stir for 1 h. After this time,
the reaction mixture was diluted to 1 mg/mL protein
with 10 mM PBS and dialyzed for 3 days into 10 mM
PBS. After this time the immunogen was aliquoted,
frozen, and stored at 20 ꢀC. Another conjugate was
prepared as described, substituting bovine serum albu-
min (BSA) for KLH. This conjugate was used for
ELISA studies.
Acknowledgments
Rabbit immunizations and antibody puri®cation
We gratefully acknowledge ®nancial support from the
Searle Foundation (Chicago Community Trust), the
Welch Foundation (#F-1188), and an NSF PYI award
#CHE-9157440 to B.L.I.
The immunogens prepared were used to immunize
5 lb, male New Zealand White rabbits (Diamond B
Rabbitry, Stockdale, Texas), sets of 10 5-week old male
BALB/cJ mice (Jackson Laboratories, Bar Harbour,
Maine), or sets of 5-week old male outbred ICR mice
(Harlan Sprague Dawley, Indianapolis, IN) according
to published procedures4,12,13 with immunizations every
21 days. The polyclonal antibody samples were isolated
10 days following immunization, and puri®ed according
to published procedures.4
References and Notes
1. Gallacher, G.; Jackson, C. S.; Searcey, M.; Badman, T.;
Goel, R.; Topham, C. M.; Mellor, G. W.; Brocklehurst, K.
Biochem. J. 1991, 279, 871±881.
2. Gallacher, G.; Searcey, M.; Jackson, C. S.; Brocklehurst, K.
Biochem. J. 1992, 284, 675±680.
Anity constants by competition ELISA
3. Gallacher, G.; Searcey, M.; Jackson, C. S.; Goel, R.; Mel-
lor, G. W.; Brocklehurst, K. Eur. J. Biochem. 1993, 214, 197±
207.
4. Stephens, D.; Iverson, B. L. Biochem. Biophys. Res. Comm.
1993, 192, 1439±1444.
Competition ELISA experiments were carried out
according to standard procedures13 using a 1 h at room
temperature preincubation of antibodies with compe-
titor on a separate plate, followed by a 10 min at room

426
A. L. Odenbaugh et al. / Bioorg. Med. Chem. 8 (2000) 413±426
5. Gallacher, G. Biochem. Soc. Trans. 1993, 21, 1087±1090.
6. Stephens, D. B.; Wilmore, B. H.; Iverson, B. L. Bioorg.
Med. Chem. 1994, 2, 653±658.
7. Suzuki, H.; Sasaki, K.; Mukouyama, E.; Kawamura-
Konishi, Y.; Wada, Y.; Ono, M. Biochem. Mol. Biol. Int. 1994,
34, 27±35.
8. Iverson, B. L. CHEMTECH 1995, 25, 17±21.
9. Shreder, K. R.; Thomas, R.; Wallace, M. B.; Helms, E.;
Iverson, B. L. Isr. J. Chem. 1996, 36, 215±220.
10. Carter, S.; Resmini, M.; Simms, C.; Sreedharan, S. K.;
Gallacher, G.; Brocklehurst, K. Biochem. Soc. Trans. 1997, 25,
86S.
18. Mu, Y. Q.; Gibbs, R. A. Bioorg. Med. Chem. 1997, 5,
1327±1337.
19. Pontoni, G.; Lopalco, L.; De Maria, S.; Zambardino, U.;
Carteni Farina, M.; Siccardi, A. G.; Zappia, V. Amino Acids
(Vienna) 1997, 12, 299±308.
20. Stephens, D. B.; Thomas, R. E.; Stanton, J. F.; Iverson, B.
L. Biochem. J. 1998, 332, 127±134.
21. Resmini, M.; Gul, S.; Sonkaria, S.; Gallacher, G.; Brock-
lehurst, K. Biochem. Soc. Trans. 1998, 26, S170.
22. Butler, J. E. In The Structure of Antigens; Van Regen-
mortel, M. H., Ed.; CRC: Boca Raton, 1992; Vol. 1, pp 209±
259.
11. Resmini, M.; Vigna, R.; Simms, C.; Barber, N.; Hagi-
Pavli, E. P.; Watts, A. B.; Verma, C.; Gallacher, G.; Brockle-
hurst, K. Biochem. J. 1997, 326, 279±287.
12. Wilmore, B. H.; Iverson, B. L. J. Am. Chem. Soc. 1994,
116, 2181±2182.
23. Friguet, B.; Cha€otte, A. F.; Djavadi-Ohaniance, L.;
Goldberg, M. E. J. Immunol. Methods 1985, 77, 305±319.
24. Underwood, A. P. J. Immunol. Methods 1993, 164, 119±130.
25. Prokipcak, J. M.; Breckles, T. H. Can. J. Chem. 1972, 50,
1770±1774.
13. Wallace, M. B.; Iverson, B. L. J. Am. Chem. Soc. 1996,
118, 251±252.
14. Basmadjian, G. P.; Singh, S.; Sastrodjojo, B.; Smith, B. T.;
Avor, K. S.; Chang, F.; Mills, S. L.; Seale, T. W. Chem.
Pharm. Bull. 1995, 43, 1902±1911.
15. Liu, D.; Yu, Y. Ann. N. Y. Acad. Sci. 1996, 799, 46±49.
16. Carter, S.; Resmini, M.; Simms, C.; Sreedharan, S. K.;
Gallacher, G.; Brocklehurst, K. Biochem. Soc. Trans. 1997, 25,
86S.
17. Resmini, M.; Vigna, R.; Simms, C.; Barber, N. J.; Hagi-
Pavli, E. P.; Watts, A. B.; Verma, C.; Gallacher, G.; Brockle-
hurst, K. Biochem. J. 1997, 326, 279±287.
26. Landsteiner, K. The Speci®city of Serological Reactions,
2nd Ed.; Harvard University: Cambridge, 1945; pp 156±170.
27. Pauling, L.; Pressman, D.; Grossberg, A. L. J. Am. Chem.
Soc. 1945, 66, 784±792.
28. Nisono€, A.; Shaw, A. R.; Pressman, D. J. J. Am. Chem.
Soc. 1959, 81, 1418±1423.
29. Linder, A. B.; Eshar, Z.; Taw®k, D. S. J. Mol. Biol. 1999,
285, 421±430.
30. Martin, S. F.; Josey, J. A. Tetrahedron Lett. 1988, 29,
3631±3634.
31. Mautz, D. S.; Davisson, V. J.; Poulter, C. D. Tetrahedron
Lett. 1989, 30, 7333±7336.