An antibody-catalyzed selenoxide elimination

Full text of DOI:10.1021/ja963748j

J. Am. Chem. Soc. 1997, 119, 3623-3624
3623
Scheme 1
An Antibody-Catalyzed Selenoxide Elimination
Zhaohui S. Zhou, Ning Jiang, and Donald Hilvert*
Departments of Chemistry and Molecular Biology
The Skaggs Institute of Chemical Biology
The Scripps Research Institute
10550 North Torrey Pines Road
La Jolla, California 92037
ReceiVed October 28, 1996
Despite their importance in chemical synthesis, pericyclic
reactions are rare in cellular metabolism.1 Engineered proteins
that catalyze cycloadditions2 and sigmatropic rearrangements3
with high rates and selectivities are consequently of considerable
interest for synthetic and mechanistic studies. Here we describe
three tailored antibody catalysts for a [2,3]-sigmatropic reaction,
the selenoxide elimination depicted in Scheme 1, and their
distinct selectivities.
Selenoxide syn elimination affords a convenient method of
introducing olefins into many molecules.4-7 The reaction is
believed to proceed via a planar, 5-membered, pericyclic
transition state which is less polar than the initial state.8,9 We
reasoned that antibodies raised against proline derivatives such
as 1a and 1b would provide a relatively low dielectric
environment capable of constraining the flexible alkyl aryl
selenoxides 2 in a reactive conformation, with the relative
disposition of the hapten’s carboxylate and 3-aryl moieties
dictating the orientation of the corresponding substituents at the
transition state. Although formation of trans-olefin products
from acyclic secondary selenoxides is favored for steric reasons,
antibodies prepared with the cis-hapten 1a (Scheme 1) could
conceivably provide sufficient binding energy to overcome the
unfavorable eclipsing interactions encountered in the transition
state leading to cis-olefins.
by the trans-hapten were purified and screened for catalytic
activity using the selenoxide derivatives 2a-e.12,13 Three
catalysts (SZ-cis-39C11, SZ-cis-42F7, and SZ-trans-28F8) were
identified and subjected to further characterization. In each case,
compounds 1a and 1b (Y ) NO2) are potent inhibitors of their
respective antibodies,14 indicating that catalysis is associated
with the induced active site. The antibodies raised against the
cis-hapten appear to be enantioselective, requiring 2 equiv of
racemic 1a per binding site to abolish activity, but SZ-trans-
28F8 accommodates both enantiomers of 1b, as judged by the
1:1 stoichiometry of inhibition.
For each of the antibodies, catalytic efficiency generally
increases with decreasing size of the substituent R to the
selenoxide moiety (R ) CO2H < CH2OH < CH3 j H). Steady
state kinetic parameters were determined for the best substrates
(2c-e) from plots of initial rates versus substrate concentration
and are presented in Table 1. The low Km values suggest
significant contributions to binding from the two aryl rings,
while the rate enhancements over the corresponding uncatalyzed
reactions (kcat/kuncat e 103) are similar in magnitude to those
observed for other antibody-catalyzed sigmatropic processes.3
Although the selenoxide moiety itself rapidly epimerizes
under the reaction conditions,15 the additional chiral center in
the secondary selenoxides influences the course of the antibody-
catalyzed reactions in dramatically different ways. Thus, SZ-
trans-28F8 evinces no chiral discrimination with substrate 2c,
converting 100% of the racemic starting material to anethole 3
(R ) CH3, >90% trans) (Figure 1). A small amount of cis-
olefin is formed (≈8%), as in the uncatalyzed reaction,
indicating more than a single binding mode for the flexible
substrate. Together, these results correlate well with this
antibody’s ability to recognize both racemic haptens.14 In
contrast, SZ-cis-39C11 and SZ-cis-42F7 appear to be highly
stereoselective as judged by catalytic conversion of only 50%
of the racemic substrate (Figure 1A). trans-Anethole is the
exclusive product of the SZ-cis-42F7-catalyzed reaction, but SZ-
cis-39C11 affords a 45:55 mixture of cis- and trans-olefin
(Figure 1B). The comparable energies of the SZ-cis-39C11-
Racemic haptens 1a and 1b (Y ) NHC(O)CH2Br) were
synthesized, coupled to carrier proteins, and used to produce
monoclonal antibodies by standard methods.10,11 Twenty-eight
antibodies elicited by the cis-hapten and 20 antibodies elicited
(1) Pindur, U.; Schneider, G. H. Chem. Soc. ReV. 1994, 409-415.
(2) (a) Hilvert, D.; Hill, K. W.; Nared, K. D.; Auditor, M.-T. M. J. Am.
Chem. Soc. 1989, 111, 9261-9262. (b) Braisted, A. C.; Schultz, P. G. J.
Am. Chem. Soc. 1990, 112, 7430-7431. (c) Gouverneur, V. E.; Houk, K.
N.; Pascual-Teresa, B.; Beno, B.; Janda, K. D.; Lerner, R. A. Science 1993,
262, 204-208.
(3) (a) Hilvert, D.; Carpenter, S. H.; Nared, K. D.; Auditor, M. T. M.
Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 4953-4955. (b) Jackson, D. Y.;
Jacobs, J. W.; Sugasawara, R.; Reich, S. H.; Bartlett, P. A.; Schultz, P. G.
J. Am. Chem. Soc. 1988, 110, 4841-4842. (c) Braisted, A. C.; Schultz, P.
G. J. Am. Chem. Soc. 1994, 116, 2211-2212.
(4) Sharpless, K. B.; Young, M. W.; Lauer, R. F. Tetrahedron Lett. 1973,
1979-1982.
(5) Sharpless, K. B.; Young, M. W. J. Org. Chem. 1975, 40, 947-949.
(6) Reich, H. J.; Wollowitz, S.; Trend, J. E.; Chow, F.; Wendelborn, D.
F. J. Org. Chem. 1978, 43, 1697-1705.
(7) Reich, H. J. Acc. Chem. Res. 1979, 12, 22-30.
(12) Substrates 2a-e were prepared immediately prior to use by in situ
oxidation of the corresponding selenides with excess hydrogen peroxide.
The selenides were synthesized from the corresponding alcohols, activated
as the mesylate, by nucleophilic displacement with a substituted arylsele-
nolate. All new compounds gave satisfactory spectroscopic data.
(13) All kinetic measurements were performed in aqueous buffer (60
mM Tris, 100 mM NaCl, pH 8.00) at 25 °C unless otherwise indicated.
Reactions were monitored spectrophotometrically (at 275 nm for 2a, 260
nm for 2b and 2c, and 258 nm for 2d and 2e) and/or by HPLC. HPLC
assays were performed on a LiChrosorb C-18 reverse-phase column (10
mm × 25 cm, eluted isocratically with mixtures of water and acetonitrile
containing 0.05% trifluoroacetic acid at 1.2 mL/min). Reaction products
were verified by HPLC by coinjection with authentic samples.
(8) Kwart, H. Acc. Chem. Res. 1982, 15, 401-408.
(9) Kwart, L. D.; Horgan, A. G.; Kwart, H. J. Am. Chem. Soc. 1981,
103, 1232-1234.
(10) Haptens 1a and 1b were synthesized in racemic form from 3-(4-
methoxyphenyl)proline (Chung, J. Y. L.; Wasicak, J. T.; Arnold, W. A.;
May, C. S.; Nadzan, A. M.; Holladay, M. W. J. Org. Chem. 1990, 55,
270-275) by N-arylation with 4-fluoronitrobenzene, chromatographic
separation of the cis and trans isomers, followed by reduction of the nitro
group and acylation of the resulting amines with bromoacetyl bromide. All
new compounds gave satisfactory spectroscopic data. Protein conjugates
were prepared by alkylation of the thiols on thyroglobulin and bovine serum
albumin previously modified with 2-iminothiolane. Epitope density ranged
from 8 to 26 haptens per protein molecule.
(14) Inhibition constants, determined as previously described (Tarasow,
T. M.; Lewis, C.; Hilvert, D. J. Am. Chem. Soc. 1994, 116, 7959-7963),
show that the anti-1a antibodies bind the cis-hapten more than 3 orders of
magnitude more tightly than the trans. For SZ-cis-39C11, for example, Ki
values of 47 nM and 96 µM were obtained for 1a and 1b (Y ) NO2),
respectively. In contrast, SZ-trans-28F8 binds both hapten isomers with
comparable affinity (Ki ) 84 nM, 1a; 82 nM, 1b).
(11) Immunization with the thyroglobulin conjugate of 1a and 1b and
preparation of monoclonal antibodies were performed by standard methods
(Harlow, E.; Lane, D. Antibodies: A Laboratory Manual; Cold Spring
Harbor Lab.: New York, 1988). Hybridomas were subcloned twice and
propagated in mouse ascites. Antibodies were purified by ammonium sulfate
precipitation and sequential chromatography on DEAE-Sepharose, Protein
G, and MonoQ ion-exchange columns.
(15) Davis, F. A.; Reddy, R. T. J. Org. Chem. 1992, 57, 2599-2606.
S0002-7863(96)03748-1 CCC: $14.00 © 1997 American Chemical Society

3624 J. Am. Chem. Soc., Vol. 119, No. 15, 1997
Communications to the Editor
Table 1. Kinetic Parameters for the Antibody-Catalyzed Syn
Elimination of Selenoxide 2a
kcat/Km
kcat
antibody
substrate
(M-1 min-1) (min-1) kcat/kuncat
SZ-trans-28F8 2c (R ) Me, X ) NO2)
2d (R ) H, X ) NO2)
120000
1700
250
2800
1300
2400
17000
600
0.18
160
210
73
0.016
0.0012
0.039
0.040
0.035
0.067
0.020
0.0021
2e (R ) H, X ) H)
SZ-cis-39C11 2c (R ) Me, X ) NO2)b
2d (R ) H, X ) NO2)
36c
550
2200
62
270
130
2e (R ) H, X ) H)
Figure 2. Rate constants for the reaction of compound 2e (25 °C)
plotted against the solvent solvatochromic polarity parameter ET(30)
in cyclohexane, n-hexane, 1,4-dioxane, tetrahydrofuran, ethyl acetate,
chloroform, methyl acetate, dichloromethane, 1,2-dichloroethane (filled
diamond), tert-butyl alcohol, 2-butanol, isopropyl alcohol, n-butanol,
ethanol, ethylene glycol, and aqueous buffer (Tris-HCl 60 mM, NaCl
100 mM, pH 8.0) in the order of increasing ET(30) value.
SZ-cis-42F7
2c (R ) Me, X ) NO2)
2d (R ) H, X ) NO2)
2e (R ) H, X ) H)
39
a Assays were performed in aqueous buffer (60 mM Tris-HCl, 100
mM NaCl, pH 8.0) at 25 °C. To account for substrate depletion during
assays with high antibody concentrations (1-10 µM), the data were
fit to the equation ν ) {(kcat/2)[(E + S +Km) - [(E + S + Km)2 -
4ES]1/2]},20 where ν is the initial rate, kcat and Km are the steady-state
kinetic parameters, and E and S are the total concentrations of antibody
binding site and substrate respectively. b Severe substrate inhibition was
observed with this substrate, and the kinetic parameters were estimated
from data obtained at low substrate concentration (e30 µM). c As
described in the text, significant yields of cis-anethole are obtained in
the presence of antibody SZ-cis-39C11 but not in the uncatalyzed
elimination, thus the value of kcat/kun substantially underestimates
catalytic efficiency in this case.
lyzed reaction may reflect the release of ordered solvent
molecules as the transition state is approached. The selenoxide
substrate is a good hydrogen bond acceptor and will be solvated
to a much greater extent in aqueous buffer than the compara-
tively less-polar transition state. Differences in solvation of the
antibody-bound ground and transition states are likely to be
much less pronounced. Because selenoxide syn eliminations
are not subject to acid and base catalysis,6 partitioning the
substrate into the less-polar medium of the antibody binding
site alone could account for much of the observed rate
enhancement,17,18 while the conformational constraints of the
active site would dictate the observed selectivity. Indeed, the
elimination of 2e is dramatically accelerated by aprotic sol-
vents: log(kobsd) correlates with the solvatochromic polarity
measure ET(30) over a reactivity range of 104 (r ) 0.973)
(Figure 2). The rate achieved by SZ-cis-39C11 (kcat ) 0.036
min-1) is similar to that found with the aprotic solvent 1,2-
dichloroethane (ET(30) ) 41.3; kobsd ) 0.0440 min-1). Notably,
reaction in the latter solvent is also characterized by an
unfavorable activation entropy relative to aqueous buffer (∆Hq
) 20.3 ( 0.5 kcal/mol; ∆Sq ) -4.8 ( 1.7 eu).
In summary, these experiments expand the scope of antibody
catalysis to a new class of pericyclic reactions for which natural
enzymes are unknown.19 More importantly, they illustrate how
conformational constraints imposed through hapten structure can
be exploited together with medium effects to control chemical
selectivity and reactivity. Improvements in hapten design,
coupled with more extensive screening of the immune response,
may yield synthetically useful catalysts for a wide range of
solvent-sensitive processes.
Figure 1. (A) Total time course of the antibody-catalyzed and
uncatalyzed selenoxide elimination of 2c. The reactions were performed
as described13 with 24 µM substrate at 15 °C without antibody (bottom
trace), with 14 µM SZ-cis-39C11 (middle trace), and 2.4 µM SZ-trans-
28F8 (top trace). They were monitored at 248 nm (∆ꢀ ) 10 000 M-1
cm-1 for both cis- and trans-anethole). (B) HPLC traces13 of authentic
trans-anethole (11.8 µM, 21.4 min), an aliquot of the SZ-trans-28F8
reaction at t ) 260 min (1.6 µM cis 3, 20.2 min; 21.3 µM trans 3, 21.6
min), an aliquot of the SZ-cis-39C11 reaction at t ) 376 min (5.6 µM
cis 3, 20.0 min; 7.8 µM trans 3, 21.4 min), and authentic cis-anethole
(11.8 µM; 19.8 min). Acetophenone (5.0 min) was used as an internal
standard.
bound transition states leading to the cis- and trans-olefin
products is notable and might be explained by a bifurcated
binding pocket analogous to that of the crossreactive antibody
DB316 which exhibits high affinity recognition of conforma-
tionally distinct steroids. The production of significant amounts
of the cis-olefin must also account for the anomalously low
efficiency of SZ-cis-39C11 with 2c (Table 1), since the more
facile pathway leading to trans product dominates the uncata-
lyzed reference reaction.
To probe the origins of catalysis, the reaction of primary
selenoxide 2e with SZ-cis-39C11 was examined in greater detail.
Comparison of the activation parameters for the reaction in the
antibody pocket (∆Hq ) 19.7 ( 1.2 kcal/mol and ∆Sq ) -7.8
( 4.1 eu) and in aqueous buffer (∆Hq ) 26.3 ( 0.15 kcal/mol
and ∆Sq ) +0.014 ( 0.47 eu) shows that transition state
stabilization is achieved through enthalpic rather than entropic
means. The more favorable activation entropy for the uncata-
Acknowledgment. This work was supported in part by the National
Institutes of Health (GM38273 to D.H.).
JA963748J
(17) Lewis, C.; Kramer, T.; Robinson, S.; Hilvert, D. Science 1991, 253,
1019-1022.
(18) By analogy, other proteins with generic hydrophobic binding pockets
should also accelerate these transformations. In fact, BSA, which promotes
other medium-sensitive reactions (Kikuchi, K.; Thorn, S. N.; Hilvert, D. J.
Am. Chem. Soc. 1996, 118, 8184-8185; Hollfelder, F.; Kirby, A. J.; Tawfik,
D. S. Nature 1996, 383, 60-63), catalyzes the selenoxide elimination of
1e (kcat ) 0.0022 min-1, Km ) 170 µM), but depending on the substrate
concentration, its efficiency is 20 to 200-fold lower than that of SZ-cis-
39C11. In addition, BSA does not catalyze the formation of cis-olefins from
secondary selenoxides.
(19) Antibodies that catalyze a related [2,3]-sigmatropic elimination
reaction of an N-oxide were recently reported (Yoon, S. S.; Oei, Y.; Sweet,
E.; Schultz, P. G. J. Am. Chem. Soc. 1996, 118, 11686-11687). As in our
study, medium effects were identified as a major contributor to catalytic
efficiency.
(20) Segel, I. H. Enzyme Kinetics: BehaVior and Analysis of Rapid
Equilibrium and Steady-State Enzyme Systems; John Wiley & Sons: New
York, 1975; pp 72-77.
(16) Arevalo, J. H.; Taussig, M. J.; Wilson, I. A. Nature 1993, 365, 859-
863.