Substrate-assisted antibody catalysis

Article abstract of DOI:10.1039/b314264g

A new strategy in transition-state analog design is demonstrated to elicit catalytic antibodies. The strategy is based on substrate-assisted antibody catalysis and utilizes analogs designed to mimic the transition-state for intramolecular catalysis and thereby favor antibodies that can recruit catalytic groups from substrate. The hydrolysis of the benzoyl ester of cocaine provides an illustration. The benzoyl ester of cocaine is distant from the protonated nitrogen in the stable chair conformer but proximate in the strained boat form. An antibody stabilizing the boat form and approximating ester and amine could catalyze ester hydrolysis. To mimic the transition-state for the intramolecular catalysis, we synthesized a cocaine analog that replaces this ester with a methylenephenylphosphinate bridge to the tropane nitrogen. This bridged analog elicited 85 cocaine esterases out of 450 anti-analog antibodies-a performance markedly superior to that of a simple phosphonate ester-based analog with an identical tether. The correspondence of the analog to a "high energy" conformer eliminated product inhibition. For certain polyfunctional targets, substrate assistance can be an effective strategy for eliciting catalytic antibodies.

Full text of DOI:10.1039/b314264g

View Online / Journal Homepage / Table of Contents for this issue
Substrate-assisted antibody catalysis
Shixian Deng, Narine Bharat, Paloma de Prada and Donald W. Landry*
Division of Clinical Pharmacology and Experimental Therapeutics, Department of Medicine,
Columbia University, Box 84, 630 W168th Street, New York, NY 10032, USA.
E-mail: DWL1@columbia.edu.; Fax: (212)3053475; Tel: (212)3051152
Received 7th November 2003, Accepted 17th December 2003
First published as an Advance Article on the web 24th December 2003
A new strategy in transition-state analog design is demon-
strated to elicit catalytic antibodies. The strategy is based
on substrate-assisted antibody catalysis and utilizes
analogs designed to mimic the transition-state for intra-
molecular catalysis and thereby favor antibodies that can
recruit catalytic groups from substrate. The hydrolysis of
the benzoyl ester of cocaine provides an illustration. The
benzoyl ester of cocaine is distant from the protonated
nitrogen in the stable chair conformer but proximate in
the strained boat form. An antibody stabilizing the boat
form and approximating ester and amine could catalyze
ester hydrolysis. To mimic the transition-state for the
intramolecular catalysis, we synthesized a cocaine analog
that replaces this ester with a methylenephenylphos-
phinate bridge to the tropane nitrogen. This bridged
analog elicited 85 cocaine esterases out of 450 anti-analog
antibodies–a performance markedly superior to that
of a simple phosphonate ester-based analog with an
identical tether. The correspondence of the analog to a
“high energy” conformer eliminated product inhibition.
For certain polyfunctional targets, substrate assistance
can be an effective strategy for eliciting catalytic
antibodies.
Catalytic antibodies1 have unique potential as therapeutic
artificial enzymes based on their biocompatibility and extended
plasma half-life. As a therapeutic approach to cocaine (1)
abuse,2 we prepared the phosphonate monoester transition-
state analog (TSA) 2 and obtained antibodies that cleave
cocaine into the non-addicting, non-toxic products ecgonine
methyl ester (3) and benzoic acid (4) (Scheme 1, path a).3,4
However, this analog of simple alkaline hydrolysis yielded
relatively few catalysts. New strategies favoring more complex,
and potentially more efficient, hydrolytic mechanisms have
focused on eliciting catalytic groups in the antibody.5–7 A com-
plementary strategy to recruit catalytic groups from substrate
was suggested by our observation that the protonated tertiary
amine of cocaine internally promotes hydrolysis of its methyl
ester but not its benzoyl ester.8 We hypothesized that for poly-
functional substrates a properly designed analog could elicit
antibodies that favor participation of one or more functional
groups from the substrate in the stabilization of an otherwise
unfavored transition state. We now report the synthesis of
TSA 5, a transition-state analog designed to promote substrate
assistance in the hydrolysis of the benzoyl ester of cocaine
(Scheme 1, path b), and the development of the first catalytic
antibodies using this strategy.
Scheme 1 Alkaline hydrolysis of cocaine (1) to ecgonine methyl ester (3) and benzoic acid (4). Path a: hydrolysis mimicked by TSA 2; Path b:
intramolecular catalysis mimicked by TSA 5.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 8 8 – 2 9 0
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
288

View Online
Scheme 2 Synthesis of TSA 5: a) (Ph)(TfOCH2)PO(OMe), Et3N, DMF; b) TMSBr; c) PMe3; d) succinic anhydride; e) DCC–DMAP, PhCH2OH;
f ) TMSBr; g) DCC, pyridine, reflux; h) CH3I; i) CH3I; j) H2–Pd-C; k) EDC–NHS, BSA; l) H2O, 14 days.
Bridged analog 5 was designed to elicit antibodies that would
stabilize the boat conformation of cocaine, bring the tropane
nitrogen, protonated under physiological conditions, into
proximity of the benzoyl ester group and favor stabilization of
the transition-state for alkaline hydrolysis through protonation
of the incipient oxy-anion. Also, the tetrahedral phenylphos-
phinate group would mimic the geometry of the transition state
for benzoyl hydrolysis.1 Because a single analog can yield a
multi-step catalyst,1c we did not consider further the proton
transfer steps required to yield product. Finally, the present
strategy is distinct from the “bait-and-switch” strategy because
there is no “switch”, i.e., the positive charge in the analog is
also present in the substrate.6 Nonetheless, the quarternized
nitrogen could elicit a negatively charged group for par-
ticipation in the hydrolysis. A preferred attack of hydroxide anti
to the carbomethoxy group (Scheme 1, path b) might be
expected but either phosphinate diasteromer would be useful.
As small molecules do not elicit an immune response, we
required both free TSA 5a and the immunogenic conjugate to
carrier protein 5b.
the interaction of the dipoles of the two esters.9b N-methylation
of 9a with excess CH3I in DMF at 40–45 ЊC provided free
TSA 5a in 85% yield. Quarternization of the nitrogen rendered
the tropane mildly unstable at room temperature to decom-
position via a ring opening elimination to yield 10 (t1/2 ≈ 2 weeks
in H2O, pH 7.4). The synthesis of TSA 5b with bovine serum
albumin as carrier protein was similarly executed (Scheme 2,
Path B).
Balb/C mice were immunized with TSA 5b, and high-titer
antisera were elicited. Monoclonal antibodies were prepared by
standard protocols.4 Anti-analog antibodies were identified
by enzyme-linked immunosorbent assay (ELISA) against
immobilized 5b (ovalbumin conjugate). Protein A affinity-
purified antibodies were screened for the capacity to release
[3H]-benzoic acid from [3Hphenyl]-cocaine. Catalytic antibodies
further purified by DEAE anion exchange chromatography
retained activity. Catalysis of cocaine hydrolysis was confirmed
by HPLC detection of benzoic acid; the absence of anhydro-
ecgonine methyl ester by 1H NMR excluded production of
benzoic acid by elimination. The antibodies derived from TSA
5b were inhibited completely by 10 µM free TSA 5a and not
inhibited by elimination product 10 at 100 µM, confirming 5b as
the true immunogen. Kinetic and binding (Biacore surface
plasmon resonance) parameters were measured for several
antibodies (Table 1).
In our route to 5a, norecgonine methyl ester (6a) was
N-alkylated with the triflate of methyl (hydroxymethyl)phenyl-
phosphinate at room temperature to provide phosphinate ester
7a in 81% yield (Scheme 2, Path A). 1H and 13C NMR revealed
a 1 : 1 mixture of diastereomers at phosphorus. Selective
demethylation at the phosphinate methyl ester of 7a with
TMSBr afforded phosphinic acid 8a in 99% yield. Reflux in
pyridine with excess DCC provided cyclic phosphinate ester 9a,
in 68% yield, as a single diastereomer by 1H, 13C and 31P NMR
Table 1 Parameters for some antibodies elicited by TSA 5 at pH 7.4
Ϫ1
Mab
kcat/min
Km/µM
kcat/k0
Ki (5a)/nM
1
1
spectroscopy. H NMR COSY and H–13C correlation spectra
did not permit a definitive structure assignment and we
ultimately confirmed formation of the desired diastereomer by
X-ray crystallographic analysis of the HCl salt.9a The fortuitous
formation of the desired diasteromer could be rationalized by
13F12
105G7
8H1C11
19E2
0.046
0.030
0.037
0.029
500
615
263
108
573
370
460
362
29
93
65
100
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 8 8 – 2 9 0
289

View Online
Several lines of evidence support substrate assistance. First,
the antibodies showed a decline in activity at pH > 8 (pKa for
cocaine is 8.6) as illustrated in Fig. 1 and failed to accept
N-acetyl norcocaine as a substrate. These results support
intramolecular catalysis by the protonated tropane nitrogen.
However, we cannot exclude the possibility that the positively
charged TSA elicits antibodies with a complementary carb-
oxylate group that participates in catalysis. None of the anti-
bodies elicited by TSA 5b were inhibited by benzoic acid on
competitive ELISA (Fig. 2). In contrast, product inhibition was
found for antibodies from TSA 2. This observation is consistent
with the reaction of TSA 5b-antibodies through a more highly
strained conformer.10 Lastly, catalytic antibodies elicited by
TSA 5b were potently inhibited by TSA 2. The boat conform-
ation of TSA 2 is more easily accessible then that of cocaine
due to internal stabilization of the zwitterions and X-ray
crystallographic analysis of an antibody-TSA 2 complex will be
of interest.
observation is important because if shape complementarity by
antibodies is more easily achieved than complex functional
group arrangements, then analogs targeting the former would
be more successful generally.
In conclusion we have successfully implemented a new
strategy in transition-state analog design applicable to sub-
strates with potentially catalytic, conformationally accessible,
functional groups. The antibodies from these initial screenings
were not more active than the most potent antibodies elicited
by TSA 2 (kcat/k0>103),4 but analog 5 favors only substrate-
mediated stabilization of the developing oxy-anion and
additional antibody stabilization through mutagenesis is
achievable. Tertiary amines are only modest intramolecular
catalysts for acyl hydrolysis,12 but substrate-assistance could
be more potently applied to targets containing a consensus
sequence that would correspond to elements of a catalytic
triad.13 Thus, histidine and serine could be predictably recruited
for a serine protease-like mechanism by antibody binding. We
are now testing this hypothesis.
Acknowledgements
Support from the Office of National Drug Control Policy is
gratefully acknowledged.
References
1 (a) A. Tramontano, K. D. Janda and R. A. Lerner, Science, 1986,
234, 1566; (b) S. J. Pollack, J. W. Jacobs and P. G. Schultz, Science,
1986, 234, 1570; (c) S. J. Benkovic, J. A. Adams, C. L. Borders,
K. D. Janda and R. A. Lerner, Science, 1990, 250(4984), 1135;
(d ) G. M. Blackburn and A. Garcon, Biotechnology, 2000, 8b, 403;
(e) T. Fujie, Chem. Rev., 2002, 102(12), 4885.
2 (a) B. Mets, G. Winger, C. Cabrera, S. Seo, S. Jamdar, G. Yang,
K. Zhao, R. J. Briscoe, R. Almonte, J. H. Woods and D. W. Landry,
Proc. Natl. Acad. Sci. USA, 1998, 95, 10176; (b) T. J. Baird,
S. X. Deng, D. W. Landry, G. Winger and J. H. Woods, J. Pharmacol.
Exp. Ther., 2000, 295(3), 1127; (c) R. J. Briscoe, P. M. Jeanville,
C. Cabrera, T. J. Baird, J. H. Woods and D. W. Landry,
Int. Immunopharmacol., 2001, 1(6), 1189.
Fig. 1 The effect of pH on relative rate of hydrolysis of cocaine by
illustrative catalytic antibodies: 8H1C11(᭜), 13F12(᭿), 19E2(ꢀ) and
105G7(*). Rate is corrected for background hydrolysis at each pH.
3 D. W. Landry, K. Zhao, G. X. Q. Yang, M. Glickman and
M. Georgiadis, Science, 1993, 259, 1899.
4 G. Yang, J. Chun, U. H. Arakawa, X. Wang, M. A. Gawinowicz,
K. Zhao and D. W. Landry, J. Am. Chem. Soc., 1996, 118, 5881.
5 H. Suga, O. Ersoy, S. F. Williams, T. Tsumuraya, M. N. Margolies,
A. J. Sinskey and S. Masamune, J. Am. Chem. Soc., 1994, 116(13),
6025.
6 K. D. Janda, M. I. Weinhouse, D. M. Schloeder, R. A. Lerner and
S. J. Benkovic, J. Am. Chem. Soc., 1990, 112(3), 1274.
7 P. Wirsching, J. A. Ashley, C. H. Lo, K. D. Janda and R. A. Lerner,
Science, 1995, 270(5243), 1775.
8 P. Li, K. Zhao, S. Deng and D. W. Landry, Helvetica Chim. Acta,
1999, 82, 85.
9 (a) S. X. Deng, D. W. Landry, C. G. Zhan, manuscript in
preparation; (b) R. Y. Chen and S. C. Dan, Phosphorus, Sulfur
Silicon Relat. Elem., 1991, 56(1–4), 39.
10 A similar absence of product inhibition was observed previously
for antibodies elicited by an analog that mimicked a high-energy
conformer of a chloramphenicol ester. H. Miyashita, Y. Karaki,
M. Kikuchi and I. Fujii, Proc. Natl. Acad. Sci. USA, 1993, 90(11),
5337.
Fig. 2 Competitive ELISA for catalytic antibodies against the eliciting
TSA and products benzoic acid (BA) and ecgonine methyl ester (EME).
Antibodies from TSA 5 (᭹; 58 most productive clones) and illustrative
antibody mAb 15A10 from TSA 2 (᭺).
Based on a preliminary screen of unpurified supernatants,
the efficiency of TSA 5 (85 catalysts among 450 anti-analog
antibodies) was much greater than that of TSA 2 (9/670). The
tether length, a major determinant of analog efficiency,11 was
identical for each. The superior effectiveness of TSA 5 is con-
sistent with the premise that the structural requirements for
antibody catalysts are reduced by substrate assistance. This
11 M. Matsushita, T. Z. Hoffman, J. A. Ashley, B. Zhou, P. Wirsching
and K. D. Janda, Bioorg. Med. Chem. Lett., 2001, 11(2), 87.
12 (a) M. L. Bender, E. J. Pollock and M. C. Neveu, J. Am. Chem. Soc.,
1962, 84, 595; (b) R. S. Brown and A. Aman, J. Org. Chem., 1997,
62(14), 4816.
13 P. Carter and J. A. Wells, Science, 1987, 237(4813), 394.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 2 8 8 – 2 9 0
290