Antibody catalyzed cleavage of an amide bond using an external nucleophilic cofactor

Full text of DOI:10.1021/ja9732542

J. Am. Chem. Soc. 1998, 120, 817-818
817
Antibody Catalyzed Cleavage of an Amide Bond
Using an External Nucleophilic Cofactor
Oguz Ersoy,†,‡ Roman Fleck,† Anthony Sinskey,§ and
Satoru Masamune*,†
Departments of Chemistry and Biology
Massachusetts Institute of Technology
Cambridge, Massachusetts 02139
Figure 1.
ReceiVed September 16, 1997
The antibody-mediated hydrolysis of amides constitutes an
important step toward the development of biocatalysts for the
sequence-specific cleavage of peptides.1 Previous efforts toward
this end, including our own, used the selective binding of
transition-state analogs,2,3a proximity effects supplied by a metal
cofactor,3b intramolecular rearrangements,3c,d and complementary
acid-base catalysis4 similar to natural hydrolytic enzymes. As
these efforts did not seem to provide a reliable strategy to generate
amide-hydrolyzing antibodies, we decided to explore a different
mode of catalysis, namely nucleophilic catalysis. This mechanism
is well understood for serine and cysteine proteases.5 Ideally,
the nucleophile of the hydrolytic reaction is programmed into the
antibody binding site in the form of an amino acid side chain
residue. In practice, it is not a simple task to generate such a
precisely placed residue even with prior knowledge of the binding
site geometry.6 On the other hand, an auxiliary nucleophile can
be tightly bound in an appropriately created pocket of the antibody
binding site, and it should prove equally effective compared with
an internal nucleophilic residue. In this present study, phenol
was chosen as the auxiliary nucleophile since it is a readily water-
soluble compound, incorporating a phenyl ring that can take
advantage of hydrophobic binding interactions to optimally place
it in the antibody binding pocket. Furthermore, it is a good
nucleophile that would generate a water-labile phenyl ester 7
through the reaction with the target amide 1. The phenol-assisted
cleavage of propionyl p-nitroanilide 1 is schematically shown in
Figure 1.
Figure 2.
bond in the form of an intramolecular N f O acyl transfer
reaction of substrate 2 (Figure 3). Here, we report that these three
antibodies also catalyze the intermolecular N f O acyl transfer
reaction, the cleavage of propionyl p-nitroanilide 1 using phenol
as an external nucleophilic cofactor.
The design features of haptens O1 and O2 have been already
discussed in detail.7 In short, both haptens were designed to
program specific antibody binding pockets for the transition state
of the amide cleavage and the phenol cofactor. Furthermore, the
hapten designs sought to generate acidic and/or basic catalytic
residues in the antibody binding sites, via charge complementarity.
These residues would then augment the nucleophilic catalysis
provided by the phenol. The two haptens were used separately
(homologous immunization) to immunize Balb/C mice, as well
Recently, we reported the generation of three antibody catalysts,
6-17, 3-49 and 14-10, raised against the haptens O1, O2, and
both, respectively (Figure 2).7 In a preliminary screen, these
antibodies were shown to accelerate the cleavage of an amide
as in sequence (heterologous immunization).8
The details of the
immunization protocol have been reported,7 and the monoclonal
antibodies were generated according to standard procedures.9
The three antibody catalysts were screened for the acceleration
of the cleavage of propionyl p-nitroanilide 1 at varying concentra-
tions of phenol and at different pH values. All three were found
to catalyze this reaction, and the largest rate accelerations were
observed at pH 8.0 and 10-fold excess of phenol to propionyl
p-nitroanilide 1. The Michaelis-Menten parameters were de-
termined for the three antibody catalysts by varying either the
phenol or the propionyl p-nitroanilide concentration and holding
the other one in excess. These values are listed in Table 1. As
previously observed for the intramolecular N f O transfer
reaction, the heterologously generated antibody 14-10 was again
found to be an almost 7-fold more efficient catalyst than the two
homologously generated antibodies 6-17 and 3-49. More im-
portantly, the catalytic activity of the (heterologous) antibody 14-
10 was competitively inhibited by both haptens, while 6-17
(elicited against O1) was only inhibited by O1 and 3-49 (elicited
against O2) was only inhibited by O2.
† Department of Chemistry.
‡ Present address: Center for Molecular Medicine L8:01, Karolinska
Institute, S-17176 Stockholm, Sweden.
§ Department of Biology.
(1) For recent reviews on catalytic antibodies, see: (a) Lerner, R. A.;
Benkovic, S. J.; Schultz, P. G. Science 1991, 252, 659. (b) Schultz, P. G.;
Lerner, R. A. Acc. Chem Res. 1993, 26, 391. (c) Schultz, P. G.; Lerner, R. A.
Science 1995, 269, 1835. (d) MacBeath, G.; Hilvert, D. Chem. Biol. 1996, 3
(6), 433.
(2) (a) For an encouraging early effort, see: Janda, K. D.; Schloeder, D.;
Benkovic, S. J.; Lerner, R. A. Science 1988, 241, 1188.
(3) For other efforts toward amide hydrolysis, see: (a) Pollack, S. J.; Hsiun,
P.; Schultz, P. G. J. Am. Chem. Soc. 1989, 111, 5961. (b) Iverson, B. L.;
Lerner, R. A. Science 1989, 243, 1184. (c) Gibbs, R. A.; Tayler, S.; Benkovic,
S. J. Science 1992, 258, 803. (d) Liotta, L. J.; Benkovic, P. A.; Miller, G. P.;
Benkovic, S. J. J. Am. Chem. Soc. 1993, 115, 350. (e) Martin, M. T.; Angeles,
T. S.; Sugasawara, R.; Aman, N. I.; Napper, A. D.; Darsley, M. J.; Sanchez,
R. I.; Booth, P.; Titmas, R. C. J. Am. Chem. Soc. 1994, 116, 6508.
(4) (a) Suga, H.; Ersoy, O.; Tsumuraya, T.; Lee, J.; Sinskey, A. J.;
Masamune, S. J. Am. Chem. Soc. 1994, 116, 487. (b) Suga, H.; Ersoy, O.;
Williams, S. F.; Tsumuraya, T.; Margolies, M. N.; Sinskey, A. J.; Masamune,
S. J. Am. Chem. Soc. 1994, 116, 6025. (c) Tsumuraya, T.; Suga, H.; Meguro,
S.; Tsukanawa, A.; Masamune, S. J. Am. Chem. Soc. 1995, 117, 11390.
(5) (a) Carter, P.; Wells, J. A. Nature 1988, 332, 564. (b) Fink, A. L.;
Enzyme Mechanisms; Page, M. I., Williams, A., Eds.; The Royal Society of
Chemistry; London, 1987; p 159.
(7) Ersoy, O.; Fleck, R.; Sinskey, A. J.; Masamune, S. J. Am. Chem. Soc.
1996, 118, 13077.
(8) The underlying genetic principles of heterologous immunization are
currently under investigation in our laboratories and will be published in due
course.
(9) (a) Goding, J. Monoclonal Antibodies: Principles and Practice, 2nd
ed.; Academic Press: New York, 1986. (b) Harlow, E.; Lane, D. Antibodies.
A Laboratory Manual; Cold Spring Harbor Labratory: New York, 1988.
(6) (a) Baldwin, E.; Schultz, P. G. Science 1989, 245, 1104. (b) For a novel
protocol to elicit reactive residues in antibody binding sites by “Reactive
Immunization”, see: Wagner, J.; Lerner, R. A.; Carlos, F. B. Science 1995,
270, 1797.
S0002-7863(97)03254-X CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/14/1998

818 J. Am. Chem. Soc., Vol. 120, No. 4, 1998
Communications to the Editor
Table 1
the linker site of the haptens. Therefore, it seems likely that
phenol is bound in a deep, hydrophobic pocket in the antibody
binding sites while the binding pocket for amide 1 is more solvent
accessible. It is also interesting to note that the antibody-catalyzed
bimolecular reaction between 1 and phenol is only about 2 orders
of magnitude slower than the antibody-catalyzed lactonization of
2.7,12
a
phenol
substrate 1
catalytic
antibody haptens
immuniza-
tion
app kcat
Km
Km
(10-5/min-1
)
(µM)
(µM)
6-17 O1/O1/O1 homologous
3-49 O2/O2/O2 homologous
14-10 O1/O1/O2 heterologous
2.1
1.3
13.3
78
14
136
310
77
370
a The kcat values were measured by holding the phenol concentration
Given that the product phenyl ester 7 appears to freely diffuse
out of the binding pocket (i.e. no product inhibition is observed),
and its rate of hydrolysis in the buffer employed (k ) 2.0 × 10-4
min-1 at pH 8.0) is 3 orders of magnitudes faster than the buffer-
catalyzed rate of hydrolysis of amide 1 at this pH, it can be said
that all three antibodies have, in fact, achieved the cleavage of
the amide bond in substrate 1. This represents a formal, two step
hydrolysis of an amide bond where the initial transfer of the acyl
group to phenol is followed by the relatively facile uncatalyzed
hydrolysis of the formed phenol ester 7. It should be pointed
out that this is the first time that a comparatively weak phenolic
nucleophile is turned into a powerful cofactor by a catalytic
antibody to effect the cleavage of an amide bond. Previously, a
tyrosine residue within an antibody combining site was implicated
to assist the hydrolysis of ester bonds.13 This concept should
prove generally applicable to other hydrolytic reactions of
biological significance such as glycolysis or phosphodiester
hydrolysis.14 Our future efforts will seek to improve on this
approach for amide cleavage by using different external nucleo-
philes that are more potent without sacrificing the lability of the
subsequently formed acyl intermediates.
in excess (pseudo-first-order rate constants).
Figure 3.
The participation of phenol in antibody catalysis was clearly
established with a number of simple experiments. First and
foremost, in the absence of phenol, none of the three antibodies
showed any rate acceleration of the cleavage of amide 1.
Furthermore, neither 2-naphthol (4) nor 4-tert-butylphenol (5) was
accepted as nucleophilic cofactors for catalysis by any of the three
antibodies (Figure 3). These results indicate the existence of a
specific binding pocket for the phenol auxiliary nucleophile.
Similarly, these catalysts were also specific for anilide 1. tert-
Butylacetyl p-nitroanilide 6 was not accepted as a substrate,
indicating that a dedicated and specific binding pocket also exists
for 1. Together with the observation that substrate 3 lacking the
nucleophilic hydroxyl group was not subject to antibody catalysis,
these results confirm that the participation of phenol is essential
for the antibody-catalyzed cleavage of amide 1. The product of
the reaction, phenyl propionate 7, was shown not to inhibit the
antibody-catalyzed cleavage of 1, even at relatively high con-
centrations (5 mM 7 and 0.5 mM 1). Furthermore, we found
that none of the three antibodies displayed any acceleration of
the rate of hydrolysis of 7. Thus, it can be stated that the product
7 readily diffuses out of the antibody binding pocket upon its
formation. Finally, the comparatively fast uncatalyzed rate of
hydrolysis of product 7 regenerates the phenol cofactor, thus
completing the catalytic cycle.
Acknowledgment. We thank Dr. Hiroaki Suga for helpful discussions
and comments. S.M. acknowledges generous support from the National
Science Foundation (CHE-9424283) and Kao Corporation, Japan. A.J.S.
thanks the National Science Foundation under the engineering research
center initiative to the Biotechnology Process Engineering Center
(Cooperative Agreement EDC-88-03014) for financial support.
JA9732542
(10) (a) Janda, K. D.; Shevlin, C. G.; Lerner, R. A. Science 1993, 259,
490. (b) Janda, K. D.; Shevlin, C. G.; Lerner, R. A. J. Am. Chem. Soc. 1995,
117, 2659. (c) Na, J.; Houk, K. N.; Shevlin, C. G.; Janda, K. D.; Lerner, R.
A. J. Am. Chem. Soc. 1993, 115, 8453. (d) Shabat, D.; Itzhaky, H.; Reymond,
J.-L.; Keinan, E. Nature 1995, 374, 143. (e) Schultz, P. G.; Lerner, R. A.
Acc. Chem. Res. 1993, 26, 391.
(11) Mechanistically, it is possible that phenol can attack the amide bond
in the presence of competing water (even though phenol is a rather poor
nucleophile). However, the breakdown of the formed tetrahedral intermediate
would most likely favor the expulsion of phenol due to the huge pKa difference
(greater than 6 units) beween the phenol and nitroaniline. The antibodies are
capable of rerouting this process in favor of nitroanilide expulsion.
(12) There could be steric repulsion between the o-hydrogen of phenol
and the â-hydrogens of propionamide 1 since the hapten structure represents
this as a carbon-carbon bond. As a result, phenol and amide 1 might not
assume an optimal orientation within the antibody in terms of the phenol attack
angle, phenol deprotonation, and anilide protonation. However, it seems that
antibodies do provide some flexibility in accommodating different substrates.
(13) (a) Angeles, T. S.; Smith, R. G.; Darsley, M. J.; Sugasawara, R.;
Sanchez, R. I.; Kenten, J.; Schultz, P. G.; Martin, M. T. Biochemistry 1993,
32, 12128. (b) Martin, M. T. Drug DiscoVery Today 1996, 1 (6), 239.
(14) (a) Suga, H.; Tanimoto, N.; Sinskey, A. J.; Masamune, S. J. Am. Chem.
Soc. 1994, 116, 11197. (b) Yu, J.; Hsieh, L. C.; Kochersperger, L.; Yonkovich,
S.; Stephans, J. C.; Gallop, M. A.; Schultz, P. G. Angew. Chem., Int. Ed.
Engl. 1994, 33, 339. (c) Reymond, J.-L.; Janda, K. D.; Lerner, R. A. Angew.
Chem., Int. Ed. Engl. 1991, 30, 1711. (d) Scanlan, T. S.; Prudent, J. R.; Schultz,
P. G. J. Am. Chem. Soc. 1991, 113, 9397.
The background rate of the hydrolysis of amide 1 was found
to be 2.1 × 10-7 min-1 at pH 8.0. Interestingly, this rate does
not vary with a change in the concentration of phenol in aqueous
solution, indicating that phenol does not participate in the
uncatalyzed hydrolysis of amide 1. Therefore, the kcat values in
Table 1 cannot be directly compared with this background
hydrolysis rate. Rather, the antibody-catalyzed reaction appears
to proceed by a pathway that is kinetically “disfavored” in aqueous
phenol
solution.10,11 An analysis of Km values shows that the Km
values of all three antibodies are smaller than the respective
substrate 1
Km
values (Table 1). This may be explained by the
consideration of the hapten designs. In both cases, the part of
the hapten that corresponds to phenol is positioned furthest from