Mechanistic studies of an antibody-catalyzed pericyclic rearrangement

Article abstract of DOI:10.1021/ja962933u

We report NMR and kinetic studies of antibody AZ-28, which was generated against the diaryl-substituted cyclohexanol derivative 3 and catalyzes the oxy-Cope rearrangement of the corresponding hexadiene 1 to aldehyde 2 (Braisted, A. C., Schultz, P. G. J. A

Full text of DOI:10.1021/ja962933u

VOLUME 120, NUMBER 9
MARCH 11, 1998
© Copyright 1998 by the
American Chemical Society
Mechanistic Studies of an Antibody-Catalyzed Pericyclic
Rearrangement
Edward M. Driggers,† Ho S. Cho,† Corey W. Liu,† Catherine P. Katzka,†
Andrew C. Braisted,†,‡ Helle D. Ulrich,† David E. Wemmer,*,§ and Peter G. Schultz*,§
Contribution from the Department of Chemistry, UniVersity of California, and Structural Biology DiVision,
Lawrence Berkeley National Laboratory, Berkeley, California 94720
ReceiVed August 20, 1996. ReVised Manuscript ReceiVed March 12, 1997
Abstract: We report NMR and kinetic studies of antibody AZ-28, which was generated against the diaryl-
substituted cyclohexanol derivative 3 and catalyzes the oxy-Cope rearrangement of the corresponding hexadiene
1 to aldehyde 2 (Braisted, A. C., Schultz, P. G. J. Am. Chem. Soc. 1994, 116, 2211-2212). Conformational
studies of free substrate and the antibody-substrate complex using transferred-NOE experiments demonstrate
that the antibody binds the substrate in a cyclic conformation, consistent with the chair-like geometry of 3. In
contrast, free substrate adopts an extended conformation in solution. In spite of the conformational restriction
of the bound substrate revealed by NMR, the temperature dependence of the reaction indicates that the antibody
functions primarily by lowering ∆Hq, offsetting a decrease in ∆Sq; the values of ∆Hq and ∆Sq are 15.4 ( 2.4
kcal/mol and -23 ( 8 cal/mol K, respectively. A secondary kinetic isotope effect (kcatH/kcatD) of 0.61 ( 0.1
was measured for substrate 23 in which both allylic termini are dideuterated, indicating that the chemical
rearrangement step is wholly or partially rate limiting under saturating conditions. The magnitude of this
secondary isotope effect is consistent with a significant degree of bond formation between C1 and C6 of the
substrate in the transition state. These results, together with the recently reported three-dimensional crystal
structure for the antibody-hapten 3 complex (Ulrich, H.; Mundroff, E.; Santarsiero, B. D.; Driggers, E. M.;
Stevens, R. C.; Schultz, P. G. Nature 1997, 389, 271-275), provide a detailed mechanistic model for this
antibody-catalyzed reaction.
Introduction
aromatic amino acid biosynthesis.3 More recently, a “Diels-
Alderase” has been identified, broadening the scope of known
enzyme catalyzed pericyclic reactions to include cycloaddition
reactions.4 In view of the small number of natural catalysts
for study, the generation and characterization of catalytic
antibodies provides an opportunity to investigate the mechanisms
by which proteins can utilize binding energy to catalyze this
important class of reactions. Indeed, examples of antibody-
Despite the importance of pericylic transformations in
synthetic and mechanistic organic chemistry, there are few
examples of enzymes that catalyze this class of reactions. Of
the few enzymes that have been identified, the best studied is
chorismate mutase, which catalyzes the Claisen rearrangement
of prephenate to chorismate in the shikimate pathway for
† University of California.
‡ Current address: Department of Protein Engineering, Genentech, 460
Pt. San Bruno Blvd., South San Francisco, CA 94080.
§ Lawrence Berkeley National Laboratory.
(1) Braisted, A. C.; Schultz, P. G. J. Am. Chem. Soc. 1994, 116, 2211-
2212.
(2) Ulrich, H.; Mundroff, E.; Santarsiero, B. D.; Driggers, E. M.; Stevens,
R. C.; Schultz, P. G. Nature 1997, 389, 271-275.
(3) Romero, R. M.; Roberts, M. F.; Phillipson, J. D. Phytochemistry 1996,
40, 1015-1025.
(4) Laschat, S. Angew. Chem., Int. Ed. Engl. 1996, 35, 289-291.
S0002-7863(96)02933-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/24/1998

1946 J. Am. Chem. Soc., Vol. 120, No. 9, 1998
Driggers et al.
Figure 1. Antibody-catalyzed oxy-Cope rearrangement of substrate 1 to product 2; structure 3 is the hapten.
catalyzed Diels-Alder reactions,5,6 Claisen7,8 and Cope rear-
rangements,1 and a [2,3]-sigmatropic rearrangement9 have been
reported.
NMR spectra for compound characterization were recorded at either
400 or 500 MHz at the UC Berkeley Department of Chemistry NMR
facility. Unless otherwise noted, spectra were determined in CDCl3
with tetramethylsilane as an internal standard at 0.00 ppm. Infrared
spectra were recorded on a Perkin-Elmer Paragon 500 FTIR or a
Mattson Polaris FTIR. Mass spectra were recorded at the UCB mass
spectroscopy facility on either a Kratos MS-50 (EI) or VG 70-SE (FAB)
mass spectrometer. Elemental analyses were performed by the mi-
croanalytical lab at UC Berkeley.
We have previously shown that antibody AZ-28, which was
generated against hapten 3 (Figure 1), catalyzes the oxy-Cope
rearrangement of the 2,5-diaryl hexadiene 1 to aldehyde 2 with
a rate enhancement (kcat/kuncat) of 5300-fold over the uncatalyzed
reaction.1 More recently, a series of related antibodies and their
germline precursors have been characterized, one of which
catalyzes the rearrangement with a rate acceleration of 3.0 ×
105.2 In addition, the crystal structure of the AZ-28 Fab-hapten
complex 3 has been determined at 2.6 Å resolution.2 As part
of a detailed investigation into the mechanism, structure, and
evolution of this catalytic antibody, we now report the results
of transferred-NOE (tr-NOE) experiments on the antibody-
substrate complex, as well as temperature dependence and
kinetic isotope effect measurements. Taken together, this
information provides a detailed mechanistic picture of this
antibody-catalyzed reaction, in which a combination of binding
interactions programmed by hapten 3 control the stereoelectronic
course of the rearrangement. Furthermore, insights from this
study provide a basis for the design of second generation haptens
and directed mutagenesis experiments to generate antibodies
with enhanced catalytic properties.
Vinyl Triflate of 4-Phenylcyclohexanone (4). To a solution of
4-phenylcyclohexanone (2.2 g, 12.6 mmol) in 40 mL of tetrahydrofuran
at -78 °C was added 27.5 mL of a 0.5 M solution of potassium
hexamethyldisilazide in toluene over the course of 10 min. The solution
was allowed to warm slowly to room temperature and then stirred for
an additional hour. The reaction was then cooled to -78 °C, and a
solution of N-phenyltriflimide (5.0 g, 14.0 mmol) in 20 mL of THF
was added. The reaction was again allowed to warm slowly to room
temperature and then stirred for 3 h at which time TLC analysis (10:1
hexanes/ ethyl acetate) showed the reaction to be complete. The solvent
was removed in vacuo, and the residue was purified by two rounds of
flash chromatography (4 × 20 cm; first, 3% ethyl acetate in hexanes;
second, 3% diethyl ether in hexanes) to yield 2.75 g (71.4%) of the
desired product 4 as a colorless oil. IR: 1414, 1211, 1142, 1026, cm-1
.
1H NMR (400 MHz): δ 1.90-1.99 (m, 1H), 2.02-2.08 (m, 1H), 2.31-
2.35 (m, 1H), 2.38-2.44 (m, 1H), 2.46-2.54 (m, 1H), 2.80-2.84 (m,
1H), 5.83 (dd, 1H, J ) 2.0, 2.9 Hz), 7.18-7.24 (m, 3H), 7.31 (d, 2H,
J ) 5.4 Hz). 13C NMR (101 MHz): δ 27.8, 29.6, 31.5, 41.0, 116.9,
118.0, 126.6, 128.6, 129.9, 144.5, 148.5. LRMS (EI+): m/z 306 (M+).
HRMS: calcd for C13H13O3F3S 306.0537, found 306.0535.
Experimental Section
General Procedures and Methods. Unless otherwise noted,
materials were obtained from commercial suppliers and used without
further purification. Solvents were dried and distilled under nitrogen
using standard procedures. All aqueous solutions were prepared from
distilled H2O purified on a Milli-Q purification system. Unless
otherwise noted, all reagents were added to reactions vessels via syringe
and stirred under an inert (N2 or argon) atmosphere. Flash chroma-
tography was performed on Merck Kieselgel 60, 230-400 mesh, and
unless otherwise noted, all chromatographic solvent systems were
chosen such that the compound of interest eluted with an Rf of
approximately 0.25.
4-Trimethylstannane Methyl Benzoate (5). To a solution of
methyl 4-bromobenzoate (3.0 g, 14.0 mmol) in toluene (50 mL) at 25
°C was added to 5.0 g (15.4 mmol) of hexamethylditin followed by
1.70 g (1.54 mmol) of tetrakis(triphenylphosphino)palladium(0). The
resulting solution was heated at reflux for 5 h, by which time it had
turned black. The reaction was allowed to cool to room temperature,
and 25 mL of a saturated aqueous potassium fluoride solution was
added, followed by 25 mL of hexanes, and the mixture was stirred
vigorously at room temperature for 30 min. After filtration through
Celite, the resulting pale yellow, biphasic solution was separated, and
the organic portion was washed twice with 50 mL of saturated aqueous
potassium fluoride and once with 50 mL of brine and then dried over
MgSO4, filtered, and concentrated. The residue was purified by flash
chromatography on a 5 × 20 cm silica column (3% ethyl acetate in
hexanes) yielding 2.67 (64%) of the product 5 as a colorless oil. IR:
(5) Braisted, A. C.; Schultz, P. G. J. Am. Chem. Soc. 1990, 112, 7430-
7431.
(6) Hilvert, D.; Hill, K. W.; Nared, K. D.; Auditor, M. T. M. J. Am.
Chem. Soc. 1989, 111, 9261-9262.
(7) Jackson, D. Y.; Jacobs, J. W.; Sugasawara, R.; Reich, S. H.; Bartlett,
P. A.; Schultz, P. G. J. Am. Chem. Soc. 1988, 110, 4841.
(8) Hilvert, D.; Carpenter, S. H.; Nared, K. D.; Auditor, M. T. M. Proc.
Natl. Acad. Sci. U.S.A. 1988, 85, 4953.
(9) Yoon, S. S.; Oei, Y.; Sweet, E.; Schultz, P. G. J. Am. Chem. Soc.
1996, 118, 11686-11687.
2983, 2950, 1726, 1591, 1434, 1279, 1114, 745 cm-1 1H NMR (400
.
MHz): δ 0.31 (s, 9H), 3.90 (s, 3H), 7.57 (d, 2H, J ) 6.8 Hz), 7.97 (d,
2H, J ) 7.5 Hz). 13C NMR (101 MHz): δ -9.6, 51.9, 128.4, 129.7,
135.7, 149.5, 167.3.
1-(4-Methyl benzoate)-4-phenyl-1-cyclohexene (6). To a mixture

Studies of Antibody-Catalyzed Pericyclic Rearrangement
J. Am. Chem. Soc., Vol. 120, No. 9, 1998 1947
of LiCl (2.50 g, 58.8 mmol) and tetrakis(triphenylphosphino)palladium-
(0) (220 mg, 0.20 mmol) was added a solution of organostannane 5
(2.35 g, 7.84 mmol) in 25 mL of THF followed by a solution of vinyl
triflate 4 (2.40 g, 7.84 mmol) in 25 mL of THF. The mixture was heated
at reflux for 20 h. After cooling to room temperature, the reaction
mixture was diluted with 75 mL of diethyl ether and washed with three
50 mL portions of aqueous 10% NH4OH and once with 50 mL of brine.
The combined aqueous washes were re-extracted once with 50 mL of
diethyl ether, and the combined organic layers were then dried over
MgSO4, filtered, and concentrated. The crystalline residue was purified
by flash chromatography on a 5 × 20 cm column of silica gel (30:1
hexanes/ethyl acetate) to provide 1.10 g (48%) of the desired product
6 as a white crystalline solid (mp 125-130 °C). IR: 1714, 1277, 1109,
766, 754, 700 cm-l. 1H NMR (400 MHz): δ 1.87-1.97 (m, 1H), 2.12-
2.17 (m, 1H), 2.34-2.41 (m, 1H), 2.53-2.63 (m, 2H), 2.86-2.91 (m,
1H), 3.91 (s, 3H), 6.33-6.35 (m, 1H), 7.20-7.35 (m, 5H), 7.47 (d,
2H, J ) 8.5 Hz), 7.99 (d, 2H, J ) 8.5 Hz). 13C NMR (101 MHz): δ
28.0, 30.2, 34.3, 39.8, 52.2, 125.0, 126.4, 126.8, 127.1, 128.5, 128.7,
130.0, 136.0, 146.7, 146.8, 167.3. LRMS (FAB+): m/z 293 (MH+).
Anal. Calcd for C20H20O3: C, 82.16; H, 6.90. Found: C, 82.03; H,
6.92.
The solvent was removed in vacuo, and the residue was purified by
preparative TLC (90:10:1 acetone/methanol/ammonium hydroxide) to
give 8 mg of the free amine as a colorless oil. The free amine was
dissolved in CH2Cl2 (5 mL) with 10 µL of diisopropylethylamine.
Thiophosgene (1.8 µL, 240 µmol) was added, and the reaction mixture
was stirred at room temperature for 5 min, at which time TLC (5%
MeOH in CH2Cl2) showed a single product to have formed (Rf ) 0.75).
The solvent was removed in vacuo, and the product 10 was obtained
as a colorless oil without further purification. IR: 3415, 2990, 2930,
2706, 2117, 1638, 1132 cm-1 1H NMR (400 MHz): δ 1.23-1.58
.
(m, 2H), 1.71 (t, 2H, J ) 12.2 Hz), 1.99 (br d, 2H, J ) 12.8 Hz), 2.33
(br d, 1H, J ) 12.8), 2.62 (td, 1H, J ) 3.3, 12.9 Hz), 2.80 (tt, 1H, J )
3.2, 12.4 Hz), 3.02 (br d, 2H, J ) 6.7), 3.58 (br s, 2H), 3.66-3.73 (m,
4H), 3.90 (td, 1H, J ) 4.0, 10.7 Hz), 6.72 (br s, 1H), 7.20-7.35 (m,
5H), 7.38 (d, 2H, J ) 8.3 Hz), 7.84 (d, 2H, J ) 8.3 Hz). LRMS
(FAB+): m/z 425 (MH+).
Keyhole Limpet Hemocyanin (KLH) and Bovine Serum Albumin
(BSA) Conjugates of 10. A solution of the protein (15 mg/mL for
BSA, 10 mg/mL for KLH) was separated from any insoluble material
by centrifugation at 14 000 rpm for 10 min, followed by filtration
through a 0.2 µm filter. The solution was diluted to 5 mL total volume
with water, and the pH was adjusted to 8.5-9.0 with 0.1 M NaOH.
Dimethylformamide (0.5 mL) was added, followed by slow addition
with gentle stirring of 1 mL of a 2 mg/mL solution of isocyanate 10.
Some cloudy precipitate formed during the addition, after which the
pH was readjusted to 9.0 with 0.1 M NaOH. Considerably more
precipitation occured with the KLH solution. After 3 h, the mixture
was diluted slowly with 1 mL of phosphate-buffered saline solution
(PBS, 138 mM NaCl, 10 mM NaH2PO4, 5 mM KCl, 1.8 mM KH2PO4,
pH 7.4) and filtered through a 0.45, µm filter. The protein conjugate
was then dialyzed exhaustively against PBS. Protein concentration was
determined by Bradford assay (based on a molecular weight of 59 000
for BSA, 77 000 for KLH), giving an epitope density of 15 for BSA
and 10 for KLH (UV difference spectroscopy at 250 nm; ꢀ250 (10) )
6500 M-1 cm-1).
(1S,2S,5S)-2-(4-Methyl benzoate)-5-phenyl-1-cyclohexanol (7).
To a solution of olefin 6 (600 mg, 2.05 mmol) in 20 mL of
tetrahydrofuran at 0 °C was added 6.15 mL of a 1 M solution of
borane-tetrahydrofuran complex. The solution was allowed to warm
to room temperature slowly, and after 14 h, TLC (5:1 hexanes/ethyl
acetate) indicated that all of the olefin (Rf ) 0.90) had been converted
to baseline material. The reaction was cooled to 0 °C, and 3 mL of
water was added cautiously, followed by 3 mL of 3 N NaOH and then
3 mL of 30% hydrogen peroxide. After 2 h the reaction mixture was
diluted with 50 mL of diethyl ether, washed with 50 mL of water and
50 mL of brine, and then dried over MgSO4, filtered, and concentrated.
The residue was purified by flash chromatography on a 2.5 × 20 cm
column of silica (3:1 hexanes/ethyl acetate) to provide 340 mg (53%)
of the product as a mixture of diastereomers (mp 85-92 °C). The
diastereomers were separated by preparative TLC: 40 mg of the mixture
was applied to a 1 mm × 20 cm × 20 mm silica gel plate, and after
repeated elution with 3:1 hexanes/ethyl acetate, two distinct bands were
visible. The less polar of the two corresponded to the desired 5S
diastereomer 7 and was isolated a white solid. IR: 3441, 2929, 2867,
2-(4-Methyl benzoate)-1-propene (12a).10 Freshly distilled DMSO
(48 mL, 617 mmol) was added to NaH (0.713 g, 29 mmol), and the
mixture was heated to 80 °C for 45 min, during which time H2 evolved.
The reaction was cooled to room temperature and CH3P(Ph)3Br (10.6
g, 29.7 mmol) was added in 10 mL of DMSO; the resulting solution
became dark green on stirring. Ketone 11 (2.14 g, 11.9 mmol) in 17
mL of DMSO was added, and the mixture was allowed to react for 20
min, at which time all of the ketone had been consumed. The reaction
mixture was washed with 5 × 200 mL of hexanes. Water (30 mL)
was added slowly to the hexanes with vigorous stirring, after which
the hexanes were washed twice with 200 mL of brine, dried, and then
concentrated. The resulting oil was chromatographed on silica gel (7
× 17 cm, 2% ethyl acetate in hexanes) to yield 2.06 g (98%)of 12a as
a white crystalline material (mp 48-51 °C). IR: 1716, 1276, 1123,
1720, 1610, 1435, 1281, 1113, 757, 701 cm-1 1H NMR (400 MHz):
.
δ 1.52-1.76 (m, 4H), 1.98 (dd, 2H, J ) 12.6, 3.2 Hz), 2.31 (dd, 1H,
J ) 12.8, 2.9 Hz), 2.61 (td, 1H, J ) 11.0, 3.3 Hz), 2.79 (td, 1H, J )
12.4, 3.0 Hz), 3.90 (s, 3H), 7.20-7.34 (m, 5H), 7.36 (d, 2H, J ) 8.3
Hz), 8.01 (d, 2H, J ) 8.3 Hz). 13C NMR (101 MHz): δ 32.8, 33.8,
41.8, 42.7, 52.0, 52.6, 74.0, 126.3, 126.7, 127.9, 128.5, 128.7, 130.0,
145.6, 148.6, 166.9. LRMS (FAB+): m/z 311 (MH+). HRMS: calcd
for C20H23O3 311.1647, found 311.1645.
(1S,2S,5S)-2-((4-(tert-Butoxycarbamoyl)diethylene glycol)benza-
mide)-5-phenyl-1-cyclohexanol (9). To a solution of the carboxylic
acid 7 (38 mg, 128 µmol) in 10 mL of CH2Cl2 was added the amine 8
(32 mg, 157 µmol) prepared by reaction of 2-(2-aminoethoxy)ethanol
with 2-(((tert-butoxycarbonyl)oxy)imino)-2-phenylacetonitrile, followed
by 21 mg (157 µmol) of 1-hydroxybenzotriazole and 32 mg (157 µmol)
of dicyclohexylcarbodiimide. After 14 h, a thick precipitate had formed,
and the reaction mixture was filtered through Celite and concentrated.
The residue was purified by flash chromatography (2 × 20 cm, 2%
MeOH in CH2Cl2) to yield 55 mg (89%) of 9 as a white foam. IR:
1105 cm-1 1H NMR (400 MHz): δ 2.16 (s, 3H), 3.91 (s, 3H), 5.19
.
(s, 1H), 5.46 (s, 1H), 7.51 (d, 2H, J ) 8.6 Hz), 7.99 (d, 2H, J ) 8.6
Hz). LRMS (EI+): m/z 176 (M+). Anal. Calcd for C11H12O2: C,
74.97; H, 6.87. Found: C, 74.84; H, 6.92.
2-(4-Methyl benzoate)-1-propene-1,1-d2 (12b). To a well stirred
slurry of methyltriphenylphosphonium iodide-d3 (11.9 g, 29.4 mmol)
in 184 mL of THF at 0 °C was slowly added 29 mL of 1 M
n-butyllithium. The resulting translucent orange solution was stirred
at 0 °C for 45 min. A solution of ketone 6 (4.4 g, 24.6 mmol) in 28
mL of THF was added slowly, resulting in a green solution which
formed a white precipitate as it warmed to room temperature. After
0.5 h, the reaction was quenched by addition of 200 mL of water and
extracted three times with 200 mL of ethyl acetate. The organic solvent
was dried over Na2SO4 and concentrated into a red gum which was
washed with approximately 1 L of hexanes. These washings were
concentrated and chromatographed on silica gel (7 × 17 cm, 2% ethyl
acetate in hexanes) to give 1.49 g (34%) of 12b as a white solid (mp
3347, 2924, 1690, 1638, 1541, 1507, 1281, 1170, 1120 cm-1 1H NMR
.
(400 MHz): δ 1.42 (s, 9H), 1.52-1.74 (m, 3H), 1.98 (br d, 2H, J )
12.4 Hz), 2.32 (br d, 1H, J ) 12.7 Hz), 2.79 (t, 1H, J ) 12.3 Hz), 3.29
(br d, 2H, J ) 5.0 Hz), 3.51 (t, 2H, J ) 5.1 Hz), 3.60 (br s, 4H), 3.87
(td, 1H, J ) 10.1, J ) 2.8 Hz), 5.02 (br s, 1H), 6.81 (br s, 1H), 7.20-
7.34 (m, 5H), 7.75 (d, 2H, J ) 8.1 Hz). 13C NMR (101 MHz): δ
32.9, 33.8, 39.6, 40.3, 41.9, 42.7, 52.4, 69.6, 70.0, 73.9, 79.3, 126.3,
127.4, 127.9, 128.4, 132.9, 145.6, 147.0, 156.0, 167.4. LRMS
(FAB+): m/z 483 (MH+). HRMS: calcd for C29H39N2O5: 483.2859,
found 483.2861.
53-56 °C). IR: 3420, 1637 cm-1 1H NMR (400 MHz): δ 2.15 (s,
.
3H), 3.90 (s, 3H), 7.50 (d, 2H, J ) 8.5 Hz), 7.98 (d, 2H, J ) 8.5 Hz).
13C NMR (101 MHz): δ 21.5, 52.0, 125.4, 128.9, 129.6, 142.2, 166.8,
Isothiocyanate Derivative of 9 (10). A solution of 9 (10 mg, 20.7
µmol) in 5 mL of CH2Cl2 was stirred with 500 µL of trifluoroacetic
acid at 0 °C for 1 h to remove the tert-butoxycarbonyl protecting group.
(10) Alexander, J.; Rao, K. Ind. J. Chem. 1973, 11, 619.

1948 J. Am. Chem. Soc., Vol. 120, No. 9, 1998
Driggers et al.
14b as a colorless oil. IR: 3368, 1721, 1288, 908 cm-1 1H NMR
(400 MHz): δ 3.92 (s, 3H), 7.53 (d, 2H, J ) 8.7 Hz), 8.05 (d, 2H, J
) 8.6 Hz), 9.82 (s, 1H). 13C NMR (101 MHz): δ 52.0, 128.0, 129.4,
130.1, 137.9, 147.3, 166.5, 192.3. LRMS (EI+): m/z 192. HRMS:
calcd for C11H8D2O3 (M+) 192.0755, found 192.0750.
170.0. LRMS (EI+): m/z 178. HRMS: calcd for C11H10D2O2 (M+)
178.0962, found 178.0966.
.
2-4-Methyl-13C benzoate)-1-propene (12c). Using the procedure
described for 12b, isotopically labeled olefin 12c was produced from
the reaction of ketone 11 with 13CH3P(Ph)3I in 50% yield as a white
solid (mp 53-56 °C). IR: 3417, 1718, 1288, 908 cm-1
.
1H NMR
2-(4-Methyl benzoate)-1-propen-3-al-1-13C (14c). Using the pro-
cedure described for 14a, a solution of 13c (0.5 g, 2.5 mmol) was
oxidized and chromatographed, providing 420 mg (85%) of aldehyde
(400 MHz): δ 2.15 (d, 3H, J ) 7.10 Hz), 3.90 (s, 3H), 5.18 (d, 1H, J
1
) 156.8 Hz), 5.31 (d, H, J ) 155.7 Hz), 7.50 (d, 2H, J ) 8.5 Hz),
7.98 (d, 2H, J ) 8.5 Hz). 13C NMR (101 MHz): δ 21.2, 51.6, 114.6,
125.1, 129.3, 130.2, 142.4, 145.3, 166.5. LRMS (EI+): m/z 177.
HRMS: calcd for C1013C1H12O2 (M+): 177.0871, found 177.0874.
2-(4-Methyl benzoate)-1-propen-3-ol (13a). A solution of t-
BuOOH (5.5 M in 2,2,4-trimethylpentane, 3.5 mL) was added slowly
via syringe to a stirred slurry of SeO2 (0.538 g, 4.9 mmol) in 8 mL of
CH2Cl2 at room temperature. The mixture was allowed to stir for 5
min. A solution of 12a (1.71 g, 9.7 mmol) in 7 mL of CH2Cl2 was
added via syringe, and the mixture was stirred for 36 h at room
temperature after which time TLC (3:1 hexanes/ethyl acetate) showed
that the mixture contained primarily alcohol as product, along with
aldehyde 14a and unreacted starting material. Methylene chloride was
added to bring the reaction volume up to 50 mL, and the solution was
washed once with 50 mL each of H2O, concentrated Na2CO3, 1 M
Na2SO3, and finally brine. The organic solution was dried over Na2-
SO4 and concentrated in vacuo to a colorless oil. The crude mixture
was chromatographed on silica gel (4.5 × 17 cm, 3:1 hexanes/ethyl
acetate) to give 1.45 g of alcohol 13a (77%) as a white solid (mp 72-
75 °C) and 130 mg of aldehyde 14a (7%) as a clear oil, with the
remainder recovered as unreacted starting material. IR: 3436, 1659,
14c as a colorless oil. IR: 2954, 1720, 1612, 1457, 909 cm-1 1H
.
NMR (400 MHz): δ 3.91 (s, 3H), 6.28 (d, 1H, J ) 173 Hz), 6.68 (d,
1H, J ) 175 Hz), 7.52 (d, 2H, J ) 8.4 Hz), 8.04 (d, 2H, J ) 8.4 Hz),
9.81 (s, 1H). 13C NMR (101 MHz): δ 52.0, 128.0, 129.4, 130.1, 137.9,
147.3, 166.5, 192.3. LRMS (EI+): m/z 191. HRMS: calcd for
C1113C1H10O3 (M+): 191.0663, found 191.0669.
3-Bromo-2-phenyl-1-propene (15). Using standard procedures,12
15 was prepared by reaction of R-methyl-styrene (47.2 g, 400 mmol)
with N-bromosuccinimide (45 g, 240 mmol) using 5 mg of AIBN as a
radical initiator. The compound was partially purified by fractional
distillation (70-75 °C, 0.3 mmHg) to give material which was 70%
pure by NMR. A portion of this material (5 g) was purified by silica
gel chromatography (5 × 17 cm, 100% hexanes) to give 3 g of 15 as
a colorless oil. 1H NMR (400 MHz): δ 4.38 (s, 2H), 5.49 (d, 1H, J )
0.6 Hz), 5.55 (s, 1H), 7.27-7.40 (m, 3H), 7.49 (dd, 2H, J ) 8.0, 1.6
Hz). LRMS (EI+): m/z 196 (M+).
2-(4-Methyl benzoate)-3-hydroxy-5-phenyl-1,5-hexadiene (16a).
Aldehyde 14a (0.720 g, 3.8 mmol) and bromide 15 (0.955 g, 4.8 mmol)
were mixed in 12 mL of THF and cooled to 0 °C in an ice-water bath.
An equal volume of freshly prepared, saturated aqueous NH4Cl was
added with stirring. Zinc dust (0.361 g, 5.6 mmol) was added in small
portions over 1 min to the vigorously stirred solution. The mixture
was stirred for 1.25 h at 0 °C, diluted with 10 mL of ethyl acetate, and
warmed to room temperature. After filtering the solution through Celite
and washing the filtrate once with 10 mL brine, the organic solution
was dried over Na2SO4 and concentrated to a yellow oil. Silica gel
chromatography (1.5 cm × 17 cm, 5:1 hexanes/ethyl acetate) gave 965
mg (82%) of 16a as a colorless oil. IR: 3472, 3081, 2951, 1721, 1608,
1642 cm-1 1H NMR (400 MHz): δ 3.91 (s, 3H), 4.56 (br s, 2H),
.
5.46 (s, 1H), 5.57 (s, 1H), 7.51 (d, 2H, J ) 8.6 Hz), 8.01 (d, 2H, J )
8.6 Hz). 13C NMR (101 MHz): δ 52.1, 64.9, 114.7, 126.0, 129.8,
143.1, 146.4, 166.6. LRMS (EI+): m/z 192. HRMS: calcd for
C11H12O3 (M+) 192.0912, found 192.0914.
2-(4-Methyl benzoate)-1-propen-3-ol-1,1-d2 (13b). Using the
procedure described for 13a, 1.51 g of 12b was oxidized with SeO2 to
give 914 mg of alcohol 13b (56%) as a white solid (mp 73-76 °C)
and 100 mg of aldehyde 14b (6%) as a clear oil, with the remainder
1436, 1283, 1189, 1110 cm-1 1H NMR (400 MHz): δ 2.02 (br s,
.
recovered as unreacted starting material. IR: 3400, 2091, 1641 cm-1
.
1H), 2.51 (dd, 1H, J ) 9.2, 14.4 Hz), 2.88 (dd, 1H, J ) 2.3, 14.4 Hz),
3.92 (s, 3H), 4.66 (br d, 1H J ) 8.7 Hz), 5.14 (s, 1H), 5.377 (s, 1H),
5.382 (s, 1H), 5.387 (s, 1H), 7.27-7.33 (m, 5H), 7.40 (d, 2H, J ) 8.5
Hz), 7.99 (d, 2H, J ) 8.5 Hz). 13C NMR (101 MHz): δ 42.8, 52.1,
71.0, 114.5, 115.8, 126.2, 126.9, 127.8, 128.4, 129.2, 129.6, 140.0,
144.4, 145.0, 150.2, 166.8. LRMS (EI+): m/z 308. HRMS: calcd
for C20H20O3 (M+) 308.1410, found 308.1412.
1H NMR (400 MHz): δ 3.91 (s, 3H), 4.56 (d, 2H, J ) 6.1 Hz), 7.51
(d, 2H, J ) 8.7 Hz), 8.01 (d, 2H, J ) 8.7 Hz). 13C NMR (101 MHz):
δ 52.2, 64.8, 126.1, 129.5, 143.0, 146.4, 166.9. LRMS (EI+): m/z
194. HRMS: calcd for C11H10D2O3 (M+) 194.0912, found 194.0914.
2-(4-Methyl-13C benzoate)-1-propen-3-ol (13c). Using the proce-
dure described for 13a, 954 mg of 12c was oxidized with SeO2 to give
511 mg of alcohol 13c (50%). IR: 3418, 1718, 908 cm-1 1H NMR
.
2-(4-Methyl benzoate)-3-hydroxy-5-phenyl-1,5-hexadiene-1,1-d2
(16b). Using the procedure described for 16a, aldehyde 14b (0.100 g,
0.52 mmol) and bromide 15 (0.130 g, 0.66 mmol) were reacted and
chromatographed on silica gel (1.5 cm × 17 cm, 5:1 hexanes/ethyl
(400 MHz): δ 3.90 (s, 3H), 4.55 (s, 2H), 5.45 (d, 1H, J ) 158.6 Hz),
5.56 (d, 1H, J ) 157.5), 7.50 (d, 2H, J ) 8.4 Hz), 8.00 (d, 2H, J ) 8.4
Hz). 13C NMR (101 MHz): δ 52.1, 64.8, 114.7, 117.3, 117.7, 126.0,
129.8, 166.6. LRMS (EI+): m/z 193. HRMS: calcd for C1113C1H10O3
(M+) 193.0820, found 193.0819.
acetate) to give 129 mg (80%) of product as a colorless oil. IR: 3456,
-1
3081, 2951, 1722, 1608, 1436, 1282, 1184, 1113 cm
.
1H NMR
1
(400 MHz): δ 1.97 (d, H, J ) 2.5 Hz), 2.51 (dd, 1H, J ) 9.1, 14.5
Hz), 2.88 (dd, 1H, J ) 2.5, 14.5 Hz), 3.92 (s, 3H), 4.66 (m, 1H), 5.14
(s, 1H), 5.38 (s, 1H), 7.28-7.34 (m, 5H), 7.40 (d, 2H, J ) 8.5 Hz),
7.99 (d, 2H, J ) 8.5 Hz). 13C NMR (101 MHz): δ 42.8, 52.0, 71.0,
115.8, 126.2, 126.9, 127.8, 128.4, 129.2, 129.6, 140.0, 144.3, 145.0,
150.0, 166.8. LRMS (EI+): m/z 310. HRMS: calcd for C20H18D2O3
(M+) 310.1538, found 310.1531.
2-(4-Methyl benzoate)-1-propen-3-al (14a). Dess-Martin perio-
dinane (2.39 g, 5.6 mmol)11 was added as a solid to a solution of alcohol
13a (0.903 g, 4.7 mmol) in 10 mL CH2Cl2 at 0 °C. The reaction was
stirred for several hours, at which time the reaction was complete by
TLC (3:1 hexanes/ethyl acetate). The reaction mixture was diluted
with 20 mL of diethyl ether, and the reaction was quenched by addition
of 15 mL of concentrated NaHCO3. The emulsion was stirred
vigorously while 9.8 mmol of Na2S2O3 was added slowly. The organic
layer was separated, dried over Na2SO4, and concentrated in vacuo.
Silica gel chromatography (1.5 × 17 cm, 5:1 hexanes/ethyl acetate)
provided 653 mg (73%) of aldehyde 14a as a colorless oil. IR: 3438,
2-(4-Methyl benzoate)-3-hydroxy-5-phenyl-1,5-hexadiene-6-13C
(16c). Using the procedure described for 16a, aldehyde 14c (0.42 g,
2.2 mmol) and bromide 15 (0.56 g, 2.8 mmol) were reacted in the
presence of zinc dust (0.210 g, 3.2 mmol). Silica gel chromatography
(1.5 cm × 17 cm, 5:1 hexanes/ethyl acetate) provided 116 mg (17%)
1720, 1288, 909 cm-1 1H NMR (400 MHz): δ 3.91 (s, 3H), 6.26 (s,
.
16c as a colorless oil. IR: 3416, 1645, 1471, 1381, 1095, 908 cm-1
.
1H), 6.70 (s, 1H), 7.52 (d, 2H, J ) 8.4 Hz), 8.04 (d, 2H, J ) 8.4 Hz),
9.81 (s, 1H). 13C NMR (101 MHz): δ 52.1, 53.4, 123.8, 125.6, 128.0,
129.5, 130.1, 136.8, 147.6, 166.6, 192.4. LRMS (EI+): m/z 190.
HRMS: calcd for C11H10O3 (M+) 190.0630, found 190.0626.
2-(4-Methyl benzoate)-1-propen-3-al-1,1-d2 (14b). Using the
procedure described for 14a, a solution of 13b (0.5 g, 2.5 mmol) was
oxidized and chromatographed, providing 321 mg (68%) of aldehyde
1H NMR (400 MHz): δ 2.03 (s, 1H), 2.52 (dd, 1H, J ) 9.1, 14.5 Hz),
2.88 (dd, 1H, J ) 2.5, 14.5 Hz), 3.93 (s, 3H), 4.66 (m, 1H), 5.14 (s,
1H), 5.37 (d, 1H, J ) 157.6), 5.38 (s, 1H), 5.48 (d, 1H, J ) 159.5),
7.28-7.33 (m, 5H), 7.39 (d, 2H, J ) 8.3 Hz), 7.98 (d, 2H, J ) 8.3
Hz). 13C NMR (101 MHz): δ 29.9, 42.8, 52.0, 70.9, 114.4, 126.2,
126.9, 127.8, 128.3, 129.2, 129.6, 136.7, 140.0, 144.3, 145.0, 166.8.
(11) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4156-4157.
(12) Reed, S. F. J. Org. Chem. 1965, 30, 3258-3258.

Studies of Antibody-Catalyzed Pericyclic Rearrangement
J. Am. Chem. Soc., Vol. 120, No. 9, 1998 1949
LRMS (EI+): m/z 309. HRMS: calcd for C1913CH20O3 (M+) 309.1446,
found 309.1443.
The tube was then heated to 180 °C in an oil bath for 4 h. After cooling,
the solvent was removed in vacuo and the resulting oil was chromato-
graphed on a silica gel column (2.5 × 17 cm, 2:1 hexanes/ethyl acetate)
yielding 138 mg (95%) of aldehyde 2 as a colorless oil. IR: 2948,
2-(4-Benzoic acid)-3-hydroxy-5-phenyl-1,5-hexadiene (17a). Meth-
yl ester 16a (0.137 g, 0.5 mmol) was dissolved into dioxane (5 mL)
and placed into an ice-water bath while it was vigorously stirred. After
the addition of 3 mL of 1 N aqueous KOH over 1 min, the reaction
was stirred for 3 h at room temperature. The mixture was diluted by
the addition of 20 mL of H2O and washed once with 20 mL of diethyl
ether. The aqueous solution was acidified to pH 1 by the addition of
1 M NaHSO4 and then extracted three times with 50 mL of ethyl acetate.
The organic extracts were combined, dried over Na2SO4, and concen-
trated to yield 129 mg (97%) of the product as a light-yellow oil. IR:
1692, 1610, 1424, 1289 cm-1 1H NMR (400 MHz): δ 2.10 (m, 2H),
.
2.48 (m, 2H), 3.65 (t, 1H, J ) 6.8 Hz), 5.29 (s, 1H), 5.30 (s, 1H), 7.30
(m, 7H), 8.10 (d, 2H, J ) 8.1 Hz), 9.66 (s, 1H). 13C NMR (101
MHz): δ 28.2, 32.5, 58.3, 113.4, 126.0, 127.6, 128.4, 129, 130.9, 140.4,
142.2, 147.1, 171.8, 199.7. LRMS (EI+): m/z 294. HRMS: calcd
for C19H18O3 (M+) 294.1256, found 294.1256.
2-(4-Benzoic acid)-5-phenyl-3-13C-5-pentenal (19). Using the
procedure for 2, a solution of 17c (90 mg, 305 µmol) in a sealed tube
was degassed and then heated for 4 h at 180 °C. After cooling and
removal of the solvent, the oil was chromatographed on silica gel (2
× 17 cm, 2:1 hexanes/ethyl acetate), yielding 70 mg (77%) of the
aldehyde product 19 as a colorless oil. IR: 3422, 1705, 1638, 1288
3435, 2254, 1693, 1638, 1467 cm-1 1H NMR (400 MHz, CDCl3): δ
.
2.53 (dd, 1H, J ) 9.3, 14.4 Hz), 2.89 (dd, 1H, J ) 3.5, 14.4 Hz), 4.67
(dd, 1H, J ) 3.5, 9.4 Hz), 5.15 (s, 1H), 5.40 (s, 1H), 5.41 (s, 1H), 5.51
(s, 1H), 7.28-7.32 (m, 5H), 7.44 (d, 2H, 8.5 Hz), 8.06 (d, 2H, 8.5
Hz). 1H NMR (500 MHz, D2O): δ 2.75 (dd, 1H, J ) 8.2, 14.4 Hz),
2.97 (dd, 1H, J ) 4.7, 14.4 Hz), 4.76 (dd, 1H, J ) 4.9, 7.9 Hz), 5.20
(s, 1H), 5.39 (s, 1H), 5.43 (s, 1H), 5.44 (s, 1H), 7.37-7.43 (m, 7H),
7.86 (d, 2H, 7.9 Hz). 13C NMR (101 MHz): 29.6, 42.7, 53.4, 71.1,
115.9, 126.2, 127.0, 127.8, 128.4, 128.5, 130.2, 140.0, 144.8, 145.2,
150.0, 171.6. LRMS (EI+): m/z 294 (M+). HRMS: calcd for C19H18O3
(M+) 294.1256, found 294.1258.
2-(4-Benzoic acid)-3(S)-hydroxy-5-phenyl-1,5-hexadiene (1). Enan-
tiomerically pure substrate for kinetic assays was isolated using chiral
reverse-phase HPLC and assigned absolute stereochemistry by deriva-
tizing as the 3-hydroxy esters of phenyl (R)- and (S)-R-methoxy-R-
(trifluoromethyl)acetate, as previously described.13
cm-1 1H NMR (400 MHz): δ 1.93 (m, 2H), 2.48 (m, 2H), 3.65 (m,
.
1H), 5.04 (s, 1H), 5.32 (s, 1H), 7.30 (m, 7H), 8.10 (d, 2H, J ) 8.0
Hz), 9.66 (s, 1H). 13C NMR (101 MHz): δ 28.1, 32.7, 58.4, 126.0,
127.5, 128.3, 128.6, 129.1, 130.7, 140.4, 169.9, 199.7. LRMS (EI+):
m/z 295. HRMS: calcd for C1813CH18O3 (M+) 295.1289, found
295.1286.
Compounds 20a, 20b, and 20c were obtained as a mixtures of
inseparable diastereomers.
2-(4-Benzoic acid)-5-phenyl-5-pentenal Oxime (20a, 20a′). To a
solution of the aldehyde 19a (90 ma, 305 µmol) in 10 mL of ethanol
were successively added 400 mg each of hydroxylamine hydrochloride
(0.4 g, 5.9 mmol) and sodium acetate (0.4 g, 4.87 mmol). The resulting
slurry was stirred vigorously and diluted with water after 2 h. The
mixture was extracted with diethyl ether and then concentrated and
dried over Na2SO4. The oil obtained was HPLC purlfied (C18 reverse-
phase column; 43-60% gradient of CH3CN (0.1% TFA) in H2O (0.1%
TFA)), providing 87 mg (91%) of a mixture of 20a and 20a′ as a
2-(4-Benzoic acid)-3-hydroxy-5-phenyl-1,5-hexadiene-1,1-d2 (17b).
Using the procedure described for 17a, methyl ester 16b (0.250 g, 0.81
mmol) was hydrolyzed to yield 219 mg (91%) of 17b as a light-yellow
oil. IR: 3417, 2360, 2253, 1636, 913 cm-1 1H NMR (400 MHz): δ
.
2.15 (s, 1H), 2.53 (dd, 1H, J ) 9.1, 14.4 Hz), 2.86 (dd, 1H, J ) 3.5,
14.4 Hz), 4.64 (dd, 1H, J ) 3.6, 9.0 Hz), 5.12 (s, 1H), 5.36 (s, 1H),
7.25-7.30 (m, 5H), 7.39 (d, 2H, 8.3 Hz), 8.00 (d, 2H, 8.3 Hz). 13C
NMR (101 MHz): δ 42.8, 67.0, 71.0, 115.8, 126.3, 127.0, 127.8, 128.2,
128.4, 130.1, 140.1, 145.0, 150.0. LRMS (FAB-): m/z 295 (M - H).
HRMS: calcd for C19H15D2O3 (M - H) 295.1303, found 295.1298.
2-(4-Benzoic acid)-3-hydroxy-5-phenyl-1,5-hexadiene-1-13C (17c).
Using the procedure described for 17a, methyl ester 16c (0.11 g, 0.4
mmol) was hydrolyzed to yield 90 mg (86%) of 17c as a light-yellow
colorless foam. IR: 3420, 1687, 1421, 911 cm-1
.
1H NMR (400
MHz): (20a) δ 2.1 (m, 2H), 2.53 (m, 2H), 3.61 (m, 1H), 5.15 (d, 2H,
J ) 12.6), 6.84 (d, 1H, J ) 6.51 Hz), 7.30 (m, 7H), 8.10 (d, 2H, J )
8 Hz); (20a′) δ 2.1 (m, 2H), 2.53 (m, 2H), 4.46 (m, 1H), 5.30 (d, 2H,
J ) 6.28), 7.30 (m, 7H), 7.53 (d, 1H, J ) 6.67 Hz), 8.10 (d, 2H, J )
8 Hz). 13C NMR (101 MHz): δ 33.6, 34.0, 113.3, 113.5, 127.5, 128.5,
129.1, 129.4, 131.2, 142.2, 148.2, 149.4, 153.6, 170.0. LRMS
(FAB+): m/z 310 (MH+). HRMS: calcd for C19H20NO3 310.1443,
found 310.1443.
oil. IR: 3430, 1638, 908 cm-1 1H NMR (400 MHz): δ 2.04 (s, 1H),
.
2-(4-Methyl-13C benzoate)-5-phenyl-5-pentenal Oxime (20b, 20b′).
To a heated solution of KOH pellets (500 mg, 8.9 mmol) dissolved in
0.8 mL of water and 1 mL of ethanol was gradually added 500 mg of
N-methyl-13C-N-nitroso-p-toluenesulfonamide dissolved in 4.5 mL of
diethyl ether. The resulting 13C isotopically labeled diazomethane was
added to a stirred solution of the hydroxylamine oxime 20a (12.5 mg,
40 µmol) in CH2Cl2 until the solution retained a light-yellow color.
The mixture was then quenched with a solution of acetic acid in diethyl
ether and concentrated in vacuo to yield 15.3 mg (>99%) of methyl
esters 20b and 20b′. IR: 3405, 2946, 2360, 1721, 1610, 1432, 1282,
2.54 (dd, 1H, J ) 9.1, 14.4 Hz), 2.89 (dd, 1H, J ) 3.5, 14.4 Hz), 4.67
(m, 1H), 5.15 (s, 1H), 5.36 (s, 1H), 5.40 (d, 1H, J ) 157.6), 5.51 (d,
1H, J ) 159.6), 7.25-7.32 (m, 5H), 7.43 (d, 2H, 8.4 Hz), 8.06 (d, 2H,
8.4 Hz). 13C NMR (101 MHz): δ 30.0, 42.9, 71.0, 114.9, 116.0, 126.3,
127.1, 127.9, 128.4, 128.5, 129.7, 130.2, 140.0, 145.0, 171.0. LRMS
(EI+): m/z 295. HRMS: calcd for C1813C1H18O3 (M+) 295.1289, found
295.1293.
2-(4-Benzoic acid methyl-13C ester)-3-hydroxy-5-phenyl-1,5-hexa-
diene-1,1-d2 (18). Isotopically labeled diazomethane, produced by
reaction of N-methyl-13C-N-nitroso-p-toluenesulfonamide with KOH,
was distilled directly into a stirred solution of 20 mg (0.068 mmol) of
acid 17b in 3 mL of methanol at 0 °C. Distillation was continued
until the solution retained a light-yellow color, at which point the
reaction was quenched with acetic acid. Solvent was removed in vacuo
to give 21 mg (99% yield) of 18 as a colorless oil. IR: 3446, 2947,
1108 cm-1 1H NMR (400 MHz): (20b) δ 2.0 (m, 2H), 2.48 (m, 2H),
.
3.60 (m, 1H), 3.86 (d, 3H, J ) 38 Hz), 5.04 (d, 2H, J ) 12.68 Hz),
6.81 (d, 1H, J ) 6.51), 7.30 (m, 7H), 7.99 (d, 2H, J ) 8 Hz); (20b′)
δ 2.0 (m, 2H), 2.48 (m, 2H), 3.86 (d, 3H, J ) 38 Hz), 4.39 (m, 1H),
5.29 (d, 2H, J ) 6.28 Hz), 7.30 (m, 7H), 7.48 (d, 1H, J ) 6.67 Hz),
7.99 (d, 2H, J ) 8 Hz). 13C NMR (101 MHz): δ 31.9, 32.2, 32.8,
32.9, 45.7, 52.1, 113.1, 127.5, 127.8, 128.0, 128.3, 128.7, 130.0, 130.1,
153.2, 160.0, 172.0. LRMS (EI+): m/z 324. HRMS: calcd for
C1913C1H21NO3 (M+): 324.1555, found 324.1559.
2-(4-Benzoic acid)-5-phenyl-3-13C-5-pentenal oxime (20c, 20c′).
To a solution of aldehyde 19b (30 ma, 101 µmol) in 9 mL of ethanol
were successively added 300 mg each of hydroxylamine hydrochloride
(4.65 mmol) and sodium acetate (3.65 mmol). The resulting slurry
was diluted with 25 mL of water after 2 h and extracted with ethyl
acetate. The phases were separated, and the aqueous portion was
acidified with 1 M NaHSO4 and then extracted three times with 50
mL portions of ethyl acetate. The organic phase was washed with brine,
dried over Na2SO4, and concentrated in vacuo, to give 29 mg (99%)
of the 13C-labeled oxime 20c as a mixture of diastereomers. 1H NMR
2360, 1721, 1608, 1432, 1281, 1107 cm-1 1H NMR (400 MHz): δ
.
2.46 (dd, 1H, J ) 9.2, 14.4 Hz), 2.83 (dd, 1H, J ) 3.4, 14.4 Hz), 3.87
(d, 3H, J ) 147.1 Hz), 4.60 (dd, 1H, J ) 3.4, 9.2 Hz), 5.09 (s, 1H),
5.34 (s, 1H), 7.08-7.27 (m, 5H), 7.35 (d, 2H, J ) 8.5 Hz), 7.94 (d,
2H, J ) 8.5 Hz). 13C NMR (101 MHz): 42.9, 52.1, 53.3, 70.9, 89.2,
115.9, 126.3, 126.9, 127.8, 128.5, 129.6, 140.0, 145.1, 150.0, 160.0,
171.0. LRMS (EI+): m/z 311. HRMS: calcd for C1913C1H18D2O3 (M+)
311.1572, found 311.1578.
2-(4-Benzoic acid)-5-phenyl-5-pentenal (2). A solution of substrate
17a (145 mg, 492, µmol) in 10 mL of distilled benzene in a sealed
tube was degassed by three rounds of freeze-thawing under nitrogen.
(13) Ulrich, H. D.; Driggers, E. M. G.; Schultz, P. G. Acta Chem. Scand.
1996, 50, 328-332.

1950 J. Am. Chem. Soc., Vol. 120, No. 9, 1998
Driggers et al.
carboxylic acid 23 (83% yield). IR: 3649, 1792, 1472, 1381, 911 cm-1
1H NMR (400 MHz): δ 2.55 (dd, 1H, J ) 9.1, 14.4 Hz), 2.87 (dd, 1H,
J ) 2.2, 14.4 Hz), 4.67 (dd, 1H, J ) 2.2, 9.1 Hz), 7.32 (m, 5H), 7.43
(d, 2H, J ) 8.3 Hz), 8.06 (d, 2H, J ) 8.3 Hz). 13C NMR (101 MHz):
δ 14.1, 21.0, 42.7, 60.4, 71.0, 126.4, 127.0, 127.8, 128.4, 130.2, 140.0,
144.8, 149.2, 170.1. LRMS (EI+): m/z 298. HRMS: calcd for
C19H14D4O3 (M+) 298.1507, found 298.1508.
(400 MHz): (20c) δ 2.1 (m, 2H), 2.48 (m, 2H), 3.60 (m, 1H), 5.28 (d,
2H, J ) 3.3 Hz), 7.29 (m, 7H), 7.50 (d, 1H, J ) 6.7 Hz), 8.02 (d, 2H,
J ) 3.08 Hz); (20c′) δ 2.1 (m, 2H), 2.48 (m, 2H), 3.60 (m, 1H), 5.04
(d, 2H, J ) 8.2 Hz), 6.83 (d, 1H, J ) 7.51 Hz), 7.29 (m, 7H), 8.02 (d,
2H, J ) 3.08 Hz). 13C NMR (101 MHz): δ 31.0, 32.0, 32.6, 45.8,
113.2, 127.5, 127.9, 128.0, 128.3, 128.5, 130.6, 130.7, 140.6, 146.2,
147.3, 152.9, 170.9. LRMS (FAB+): m/z 311 (MH+). HRMS: calcd
for C1813C1H20NO3 (MH+) 311.1477, found 311.1476.
.
Antibody Production and Purification. Monoclonal antibodies
were generated against hapten 3 using standard hydridoma technol-
ogy.14,15 Swiss Webster mice were injected intraperitoneally with
KLH-3 (100 mg) emulsified in complete Freund’s adjuvant. After 2
weeks, the mice were boosted with KLH-3 emulsified in incomplete
Freund’s adjuvant, followed 2 weeks later by intravenous injection of
KLH-3 in phosphate-buffered saline solution (PBS, 138 mM NaCl, 10
mM NaH2PO4, 5 mM KCl, 1.8 mM KH2PO4, pH 7.4). Three days
after this final immunization, cells from a single spleen were fused
with 108 P3 myeloma cells in a solution of 50% polyethylene glycol,
plated into 40 96-well plates, and grown under standard HAT selective
conditions. Screening by enzyme-linked immunosorbent assay (ELISA)
for binding to BSA-3 identified 21 cell lines specific for the hapten,
which were then cloned by limiting dilution. Antibody containing
ascites fluid was produced for each of these monoclonal cell lines in
irradiated, pristane primed Swiss Webster mice using standard proto-
cols.16
4-Phenyl-3-hydroxy-1-(4-methyl benzoate)-4-penten-1-one-5,5-d2.
To 1.5 mL of a 1 M solution of LiN(Si(CH3)3)2 in THF cooled to -78
°C was slowly added a solution of 364 µL (3.12 mmol) of acetophenone
in 1.5 mL of THF. After 45 min, 14b dissolved in 1 mL of THF was
added dropwise to the mixture. The reaction was quenched after 20
min by the addition of 3 mL of methanol. The solution was washed
with brine, and the aqueous phase was extracted three times with ethyl
acetate. The combined organic phases were dried over Na2SO4 and
concentrated. The resulting oil was chromatographed on silica gel (2
× 17 cm, 4:1 hexanes/ethyl acetate), yielding 0.190 g (58%) of the
hydroxy ketone as a white solid (mp 102-106 °C). IR: 3063, 1719,
1681, 1448, 1436, 1184 cm-1 1H NMR (400 MHz): δ 3.13 (m, 2H),
.
3.63 (d, 1H, J ) 3.55 Hz), 3.90 (s, 3H), 5.30 (m, 1H), 7.49 (m, 5H),
7.84 (d, 2H, J )9Hz), 7.99 (d, 2H, J )8.6 Hz). 13C NMR (101 MHz):
δ 44.1, 52.1, 68.9, 126.8, 128.0, 128.4, 128.6, 129.8, 133.6, 136.0,
145.0, 149.0, 166.5, 200.0. LRMS (EI+): m/z 312. HRMS: calcd
for C19D2H16O4 312.1331, found 312.1339.
Antibodies were purified by protein-G affinity chromatography using
a procedure adapted from that of Fredriksson et al.17 A given volume
of ascites fluid was diluted 2:1 with binding buffer (aqueous 1.5 M
glycine, 3.0 M NaCl, pH 8.9), filtered through a glass wool plug to
eliminate particulate matter, and applied at 0.5 mL min-1 to an 8 mL
bed of agarose-linked protein-G already equilibrated in binding buffer.
Nonspecific binders were eluted from the column with binding buffer
until the OD280 of the eluent was <0.05 absorbance units. Antibody
was eluted from the column with 100 mM sodium citrate, pH 3.0.
Column fractions were collected, neutralized with 1.0 M Tris, pH 9.0,
and the protein concentration determined by absorbance at 280 nm (1
mg mL-1 ) 1.37 absorbance units, with a molecular weight of 150 000
for immunoglobulin-G). Approximately 4 mg of AZ-28 were recovered
per milliliter of ascites fluid. The eluted antibody was dialyzed
exhaustively against PBS and stored at 4 °C following sterile (0.2 µm)
filtration. Antibodies were judged to be >95% pure by polyacrylamide
gel electrophoresis.
4-Phenyl-3-((triisopropylsilyl)oxy)-1-(4-methyl benzoate)-4-penten-
1-one-5,5-d2 (21). Triisopropyl triflic acid (0.574 mL, 2.13 mmol) and
2,6-lutidine (0.478 mL, 4.11 mmol) were added to a stirred solution of
4-phenyl-3-hydroxy-1-(4-methyl benzoate)-4-penten-1-one-5,5-d2 in 2
mL of CH2Cl2. The reaction mixture was washed with 10 mL of brine
and dried over Na2SO4. Silica gel chromatography (3.5 × 17 cm, 2%
ethyl acetate in hexanes) provided 0.197 g (76%) of 21 as a colorless
oil. IR: 2945, 2867, 1716, 908 cm-1 1H NMR (400 MHz): δ 1.09
.
(m, 21H), 3.02 (dd, 1H, J ) 4.9, 11.2 Hz), 3.29 (dd, 1H, J ) 6.7, 9.4
Hz), 5.51 (m, 1H), 7.36 (d, 2H, J ) 7.5 Hz), 7.51 (d, 2H, J ) 7.27
Hz), 7.55 (d, 2H, J ) 8.3 Hz), 7.80 (d, 2H, J ) 7.27 Hz), 7.96 (d, 2H,
J ) 8.3 Hz). 13C NMR (101 MHz): δ 11.9, 12.2, 12.4, 12.5, 17.6,
18.0, 46.8, 52.0, 126.9, 128.1, 128.3, 129.1, 129.6, 133.0, 137.1, 143.4,
149.7, 166.9, 198.2. LRMS (EI+): m/z 468. HRMS: calcd for
C28D2H36O4Si 468.2665, found 468.2666.
2-(4-Methyl benzoate)-3-((triisopropylsilyl)oxy)-5-phenyl-1,5-
hexadiene-1,1,6,6-d4 (22). To a well stirred slurry of methyl-d3-
triphenylphosphonium iodide (0.208 g, 0.51 mmol) in 4 mL of THF at
0 °C was slowly added 2.05 mL of n-butyllithium (2.5 M). The solution
was stirred at 0 °C for 0.5 h, at which time a solution of ketone 21
(0.20 g, 0.427 mmol) in 3 mL of THF was slowly added and the mixture
allowed to stir for 0.5 h as it slowly warmed to room temperature.
Water (6 mL) was added to quench the reaction. The aqueous mixture
was extracted twice with 15 mL of ethyl acetate, washed with 15 mL
of brine, dried over Na2SO4, and concentrated. The resulting oily
residue was chromatographed on silica gel (2 × 17 cm, 1.5% ethyl
acetate in hexanes) to give 78 mg (39%) of 22 as a light oil. IR: 2948,
NMR Sample Preparation. For production of large quantities of
antibody AZ-28, ascites was produced in 30 Swiss Webster mice using
standard protocols.16 Antibody was purified from ascites as described
above and then exchanged into deuterated PBS buffer (dPBS, produced
by three successive lyophilizations from D2O of a known volume of
PBS). It was found that lyophilization inactivated antibody AZ-28, so
exchange into deuterated buffer was accomplished by four sequential
concentration and dilution (1:10) steps using a centriprep-10 concentra-
tion apparatus (Amicon). The final antibody concentration of the
resulting antibody stock solution was determined by UV absorbance
at 280 nm and typically ranged from 250 to 700 µM in binding sites.
In the NMR experiments, Girard’s Reagent T was used as an aldehyde
trapping agent to minimize product affinity for the antibody combining
site. Stock solutions of substrate 1 and trapping agent Girard’s Reagent
T were prepared separately by repeated lyophilization from D2O solution
at 10× the desired final concentration. Immediately prior to NMR
experiments, these stock solutions were combined with the antibody
stock solution and diluted with dPBS to give samples containing 50
µM whole antibody, 2 mM substrate, and 5 mM trapping agent (pD
7.0) in a volume of 500 µL. Kinetic competence of these samples
was independently confirmed using the described HPLC assay.
1381, 1283, 909, 735 cm-1 1HNMR(400 MHz): δ 1.02 (m, 21H),
.
2.80 (dd, 2H, J )7.7, 6.2 Hz), 3.91 (s, 1H), 4.72 (dd, 1H, J )5.5, 2.1
Hz), 7.20 (m, 5H), 7.47 (d, 2H, J ) 8.5 Hz), 7.92 (d, 2H, J ) 6.7 Hz).
13C NMR (101 MHz): δ 12.3, 18.0, 43.6, 51.9, 126.3, 127.1, 127.3,
127.9, 128.7, 129.1, 140.7, 144.1, 144.2, 149.4, 166.9. LRMS (EI+):
m/z 468. HRMS: calcd for C29H36D4O3Si (M+) 468.2998, found
468.2988.
2-(4-Benzoic acid)-3-hydroxy-5-phenyl-1,5-hexadiene-1,1,6,6-d4
(23). To a stirred solution of 22 (0.08 g, 1.3 mmol) in 10 mL of THF
at 0 °C was added 148 µL of a tetra-N-butylammonium fluoride solution
(1 M in THF). Upon completion, 2 mL of a saturated NH4C1 solution
was added and the aqueous mixture was extracted with CH2Cl2 (4 ×
10 mL). The organic extractions were combined, washed twice with
20 mL of brine, dried over Na2SO4, and concentrated to an oily residue
which was chromatographed on silica gel (2 × 17 cm, 3:1 hexanes/
ethyl acetate) to give 29 mg (49%) of the deprotected alcohol as a
colorless oil. This material was combined with the product of a second
identical reaction to give 59 mg of compound, which was hydrolyzed
in a basic aqueous solution as described for compounds 17a-c to give
(14) Harlow, E.; Lane, D. Antibodies: A Laboratory Manual; Cold Spring
Harbor: Cold Spring Harbor: Plainview, NY, 1988.
(15) Goding, J. W. MonoclonalAntibodies: Principles andPractice, 3rd
ed.; Academic Press, Ltd.: San Diego, CA, 1996; pp 141-180.
(16) Shokat, K. M. Ph.D. Thesis, University of California, Berkeley, CA,
1991.
(17) Fredriksson, U. B.; Fagerstam, L. G.; Cole, A. W. G. Protein
A-Sepharose C1-4B Affinity Purification of Ig G Monoclonal Antibodies
from Mouse Ascites; Pharmacia AB: Uppsala, Sweden, 1986.

Studies of Antibody-Catalyzed Pericyclic Rearrangement
J. Am. Chem. Soc., Vol. 120, No. 9, 1998 1951
Transferred NOE Spectroscopy. NMR studies were performed
on a Bruker AMX-600 spectrometer at a frequency of 600.14 MHz
and a GE Omega500 spectrometer at a frequency of 500.13 MHz, using
of the two substrates. This stock solution was added to a stock
concentration of antibody AZ-28 such that the final solution consisted
of 10 µM Ig, 500 µM each of 18 and 16a, 20 mM MES, 100 mM
NaCl, 5 mM NH2OH, 15% CH3CN, pH 6.0. The additional acetonitrile
was added to solubilize the methyl ester substrates. Total reaction
volumes were 2 mL. The antibody-catalyzed reaction proceeded for
16 h, at which time the solutions were frozen at -78 °C. Preparative
reverse-phase HPLC was used to isolate reaction products using a
mobile phase gradient of 35-100% CH3CN in unbuffered H2O, which
allowed recovery of both substrate and aldehyde product, trapped as
the oxime. Preparative samples were concentrated by lyophilization,
resuspended in 1:1 CH3CN/H2O, 1% AcOH, and analyzed by electron
impact mass spectrometry (EIMS) to determine the ratio of labeled to
unlabeled substrate and product. Six such reactions were analyzed for
the catalyzed and uncatalyzed reactions, and the catalyzed isotope effect
calculated as R-D(V/K) ) (PH/PD)/(SH/SD). Isotope effect reported is
the average ( standard deviation to 95% confidence. b. Vmax Effect.
For determination of the secondary deuterium kinetic isotope effect
1
5 mm H probes maintained at 18 °C by variable temperature units.
Spectra were processed using the Felix 95.0 (Molecular Simulations
Inc.) program on Silicon Graphics Inc. workstations. Chemical shift
assignments are referenced to the residual water present in the samples
(4.88 ppm at 18 °C).
Spectra for resonance assignments were acquired on a 3 mM sample
of 1 on the GE Omega500 spectrometer. Assignments were made using
a combination of 1H lD presaturation and 1H NOESY experiments. 1H
lD spectra were acquired with 4096 complex points, 5000 Hz spectral
width, 1s recycle delay, and 3 s water presaturation for 64 scans. 1H
NOESY spectra were acquired with 4096 × 256 complex points, 6024.1
Hz spectral width, 1 s recycle delay, and 2 s water presaturation for 32
scans per t1 increment.
For transferred-NOE experiments, a steady-state nuclear Overhauser
effect (NOE) difference pulse sequence was used, with a 1-1 jump-
return to achieve water suppression. A 1 s preirradiation pulse was
applied to selected proton frequencies. Reference spectra (no preir-
radiation) were interwoven with preirradiated spectra to average out
effects due to substrate turnover during experiment acquisition. Spectra
were acquired with 4096 complex points, 5555.56 Hz spectral width,
and 1 s recycle delay for a total of 3520 transients.
For CAMELSPIN experiments, a lD ROE pulse sequence was used
with a 1-1 jump-return to achieve water suppression. Spectra were
acquired with 4096 complex points, 6024.10 Hz spectral width, and
4.0 s recycle delay for a total of 2560 transients per resonance. A 50
ms DANTE pulse was utilized for the selective inversion followed by
a 50 ms spin lock at 6.19 kHz.
Kinetic Assays. Catalyzed and uncatalyzed reaction rates for the
rearrangement of 1 to 2 were measured using a reverse-phase HPLC
assay. Reactions were initiated by addition of a concentrated stock
solution of substrate 1 or the racemate 17a in DMSO or acetonitrile to
an aqueous solution, such that the final solution contained 20 mM MES,
100 mM NaCl, 5 mM NH2OH, 5% CH3CN or DMSO, pH 6.0, (500
nM AZ-28. Reaction volumes were typically 200 µL. Reaction
progress was monitored by periodic injection of a 30 µL sample from
the reaction mixture onto a reverse-phase C-18 HPLC column (Rainin
Microsorb) with UV detection at 240 nm, using an elution gradient of
43-60% CH3CN (0.1% TFA) in H2O (0.1% TFA) over 17 min at a
flow rate of 1 mL min-1 to resolve the product. Hydroxylamine was
introduced into the reaction buffer as a fast trapping agent for aldehyde
2; therefore, product was quantified as the hydroxylamine oxime of 2
(20a). All rates are initial rates, determined at less than 4% product
conversion based on product peak area, and calibrated to either
3-nitrobenzoic acid or o-nitroanisole as internal standards.
Determination of Activation Parameters. The kinetic assay
described above was carried out at a variety of temperatures (13-47
°C uncatalyzed, 5-25 °C catalyzed) in order to determine energies of
activation for the conversion of 1 to 2. For consistency, the catalyzed
and uncatalyzed temperature ranges were overlapped as much as
possible. Solutions were pre-equilibrated to the desired temperature
using a recirculating water bath, and maintained at that temperature
during the course of the assay. Temperature was monitored using a
dual channel thermometer (0.3% accuracy, Fisher Scientific) referenced
to 0 °C with a well-stirred ice-water bath. Antibody-catalyzed
temperature dependence assays were run with enantiomerically pure
(S)-1 at a concentration of 2 mM to insure Vmax conditions (Km ) 75
µM). Uncatalyzed temperature dependence assays were carried out
using 2 mM 17a. All reactions were carried out in quadruplicate, and
the data converted to enthalpies and entropies of activation by fitting
it to the standard Eyring relationships.18
on Vmax (
R-DV), a 2 mM solution of 23 was converted to hydroxylamine-
trapped product in the presence of 500 nM AZ-28 and the rate
determined using the HPLC assay described above. This Vmax rate was
compared to the Vmax rate for conversion of 17a with the same antibody
preparation. Rates were determined in quadruplicate and the isotope
effect calculated as R-DV ) kcatH/kcatD, reported as the average (
standard deviation to 95% confidence.
Results and Discussion
Synthesis. Hapten 3 was designed to mimic the chairlike
transition state for the oxy-Cope rearrangement of diene 1 to
aldehyde 2.1 An antibody generated against this hapten was
expected to catalyze this reaction by overcoming the negative
entropy of activation typically associated with a [3,3]-sigmat-
ropic rearrangement.18,19 Incorporation of the 2,5-diaryl sub-
stituents was expected to provide a large hydrophobic binding
surface for the antibody. In addition, phenyl substituents at the
2 and 5 positions of the corresponding substrate have been
shown to substantially lower the activation energy of this
reaction, facilitating kinetic analysis of the reaction at temper-
atures compatible with a protein catalyst.21 The 3-hydroxyl
group in both hapten 3 and substrate 1 might be expected to
participate in hydrogen-bonding interactions in the antibody
combining site that result in increased electron density on the
oxygen and, as a consequence, an increased the rate of
rearrangement.22 A linker was incorporated at the para position
of the 2-phenyl substituent for attachment of the hapten to the
carrier proteins bovine serum albumin (BSA) and keyhole limpet
hemocyanin (KLH). Linkage at this position to the protein,
via a thiourea group, was expected to maximize binding inter-
actions with the 2,5-diarylcyclohexyl core of hapten 3. Product
inhibition was expected to be minimized as a result of the rapid
conversion of the enol group in the product to the aldehyde.
Trapping of the aldehyde with hydroxylamine further minimizes
product inhibition and prevents addition of the aldehyde product
to lysine ꢀ-amino groups present on the antibody.
The synthesis of hapten 3 is illustrated in Scheme 1. The
disubstituted cyclohexanol core of 3 was generated by a Stille
coupling23 between aryl tin derivative 5 and vinyl triflate 4.
Subsequent hydroboration of 6 establishes a trans relationship
(19) Doering, W. V. E.; Roth, W. R. Tetrahedron 1962, 18, 67-74.
(20) Although the entropies determined by Evans and Golob are on the
order of -1 cal mol-1 K-1, these values are for a constrained norbornene
system. A more appropriate comparison would be to Foster et al.17
(21) Dewar, M. J. S.; Wade, L. E. J. J. Am. Chem. Soc. 1977, 99, 4417-
4424.
(22) Steigerwald, M. L.; Goddard, W. A. I.; Evans, D. A. J. Am. Chem.
Soc. 1979, 101, 1994-1997.
(23) Scott, W. J.; Crisp, G. T.; Stille, J. K. J. Am. Chem. Soc. 1984,
106, 4630-4632.
Kinetic Isotope Effect (KIE). a. Competition Experiments. The
secondary deuterium kinetic isotope effects on the antibody-catalyzed
reaction (R-D(V/K)) was measured by the method of internal competi-
tion. A solution containing approximately a 1:1 ratio of substrates 18/
16a was prepared by mixing equal volumes (20 mM each in CH3CN)
(18) Foster, G. E.; Cope, A. C.; Daniels, F. J. J. Am. Chem. Soc. 1947,
69, 1893-1896.

1952 J. Am. Chem. Soc., Vol. 120, No. 9, 1998
Driggers et al.
Scheme 1
between the hydroxyl group at C1 and the aryl substituent at
C2 due to the syn, anti-Markovnikov addition of borane to
olefins.24 While the C5 stereochemistry was not set, separation
of the diastereomers was found to be possible by repeated elution
on a preparative TLC plate. Stereochemical assignments were
made through a combination of 1H NMR and synthetic methods
following separation of the diastereomers. First, NMR assign-
ments were made on material which had been equilibrated in
DBU to the equatorially substituted ring system. Stereochemical
assignments of the separated diastereomers were then made by
NMR proton decoupling and comparison of coupling constants
with the equilibrated material.25 Hydrolysis of the methyl ester
7 and condensation of the resulting free acid with monoprotected
diamine 8 using 1,3-dicyclohexylcarbodiimide afforded hapten
9 in protected form. Acid-catalyzed deprotection of 9, followed
by condensation with thiophosgene, and coupling to carrier
proteins in aqueous solution (pH 9) produced the final protein
conjugates with epitope densities of 15 (BSA) and 10 (KLH).
Coupling of 14a to bromide 15 using zinc dust28 followed by
base hydrolysis provided racemic substrates 17a-c; 17a was
resolved by chiral reverse-phase HPLC to give substrate 1 as
reported.13 Racemichexadiene 17a was heated in a sealed tube
to give aldehyde 2 by oxy-Cope rearrangement, and the aldehyde
was trapped as the hydroxylamine oxime by reaction in sodium
acetate buffered ethanol to afford compounds 20a and 20a′
(inseparable diastereomers), the final assayed product of the
antibody catalysis reaction mixture. Preparative reverse-phase
HPLC purification of 20a provided material of sufficient purity
for kinetic assays.
In preparing deuterated substrate 18 for the measurement of
kinetic isotope effects, we found it necessary to introduce the
deuterium label in the initial step of the synthesis. While the
procedure for the Wittig reaction of 11 to 12a had provided
good yields using the unlabeled CH3P(Ph)3Br salt, it was found
that this method produced scrambling of the deuterium atoms
in reaction with CD3P(Ph)3I. Therefore, the phosphorus ylide
was generated using n-BuLi in THF, giving >99% specific
incorporation of deuterium. The same procedure gave specific
incorporation of 13C in 12c as well. Subsequent transformations
were as described for the unlabeled material and proceeded with-
out measurable scrambling to give completely specific deuterium
and 13C incorporation in the final substrates and products as
determined by integration of the relevant NMR resonances.
The syntheses of substrate 1 and isotopically substituted
derivatives are illustrated in Schemes 2 and 3. 4-Acetylbenzoic
acid was converted to olefin 12a by means of a Wittig reaction
with the ylide of methyltriphenylphosphonium bromide in
26
DMSO. Oxidation to the allylic alcohol with SeO2 gave a
mixture of alcohol 13a and aldehyde 14a. Purification of the
alcohol was followed by Dess-Martin oxidation11,27 to give
aldehyde 14a; recovery and reoxidation of starting material
resulted in 70% overall conversion of compound 12a to 14a.
In order to generate material which would give a sufficient
secondary deuterium kinetic isotope effect for direct detection
in kinetic assays, tetra-deuterated substrate 23 was synthesized
as shown in Scheme 4. Aldehyde 14b was coupled to
acetophenone using an aldol condensation. Following protection
of the alcohol with triisopropyl triflate, the ketone was easily
(24) Brown, H. C.; Zweifel, G. J. Am. Chem. Soc. 1961, 83, 2544-
2551.
(25) Braisted, A. B. Ph.D. Thesis, University of California, Berkeley,
CA, 1994.
(26) Warpehoski, M. A.; Chabaud, B.; Sharpless, K. B. J. Org. Chem.
1982, 47, 2897-2900.
(27) Ireland, R. E.; Liu, L. J. Org. Chem. 1993, 58, 2899-2900.
(28) Wilson, S. R.; Guazzaroni, M. E. J. Org. Chem. 1989, 54, 3087-
3091.

Studies of Antibody-Catalyzed Pericyclic Rearrangement
J. Am. Chem. Soc., Vol. 120, No. 9, 1998 1953
Scheme 2
Scheme 3
converted into the desired deutero-olefin 22 by Wittig reaction
with CD3P(Ph)3I using conditions identical to those for the
synthesis of 12b. The Wittig reaction resulted in approximately
90% specific deuterium incorporation at the terminal olefin.
Deprotection with tetra-n-butylammonium fluoride (TBAF) in
THF, followed by base hydrolysis, afforded tetra-deutero
substrate 23, which was used as the racemate after HPLC
purification.
in ELISA assays using BSA-3 conjugate. Kinetic screening
revealed nine of the antibodies to have measurable activity in
the assays of oxy-Cope rearrangement of 1 to 2; as the most
efficient catalyst, antibody AZ-28 was chosen for further
characterization. As previously reported,1 AZ-28 was found
to catalyze the rearrangement of substrate 1 to aldehyde 2 with
a kcat of 0.02 min-1 and a Km (1) of 74 µM, corresponding to
a rate enhancement (kcat/kuncat) of 5300 over the uncatalyzed
reaction (kuncat ) 4.9 × 10-6 min-1). The antibody catalyzed
reaction was competitively inhibited by hapten 3 with a Ki of
Catalytic Properties. Following immunization, a total of
21 monoclonal antibodies were isolated which bound hapten 3

1954 J. Am. Chem. Soc., Vol. 120, No. 9, 1998
Driggers et al.
Scheme 4
17 nM. In addition, the antibody shows a kinetic preference
for the S configuration at C3 of the substrate (kcatS/kcatR ) 15.1:
1). In the development of a robust HPLC kinetic assay, it was
found to be necessary to trap the aldehyde product of the
reaction 2 with hydroxylamine to achieve reliable integration
of the product. It was subsequently determined in NMR studies
that catalysis in the absence of the trapping agent resulted in a
time-dependent inactivation of the antibody, presumably due
to imine formation of the product aldehyde with surface lysines.
The exchange rate of oxime-trapped product was determined
using 13C NMR with isotopically labeled product 20c and shown
to be at least 45 s-l, consistent with the lack of product
inhibition.1 Among the less-efficient oxy-Cope catalysts which
were generated by immunization with hapten 3 and characterized
kinetically, the values of kcat ranged from 4.4 × 10-4 to 9.4 ×
10-4 min-1, with Km values varying from 15 to 66 µM, making
AZ-28 by far the most efficient of the isolated antibodies.
Substrate Conformational Analysis. It was expected that
immunization with the equatorially substituted cyclohexanol
derivative 3 would result in an antibody that restricts the acyclic
substrate in a conformation resembling the pericyclic transition
state (Figure 1). However, geometric differences exist between
the transition state and hapten 3 that could preclude binding of
the substrate in the desired conformation. For example, C2 and
C5 in the transition state (for either a concerted or stepwise
reaction) are sp2 hybridized, while the C2 and C5 positions in
the hapten are sp3 hybridized. In addition, studies on the
solution structure of equatorially substituted phenylcyclohexanes
have shown that in the most stable conformation, the phenyl
ring lies in the symmetry plane bisecting the cyclohexyl ring,29
again geometrically different from the transition state. Conse-
quently, the antibody bound conformation of the substrate was
examined by NMR in order to demonstrate that the antibody
does indeed bind the substrate in the cyclic conformation
dictated by the hapten.
(tr-NOE),30 and can be compared with NMR data for the
unbound substrate or ligand. For example, in a system similar
to AZ-28 an antibody that catalyzes a Claisen rearrangement
was shown to bind its substrate in a “high energy” conforma-
tion.31 The use of tr-NOE spectroscopy to probe the conforma-
tion of the antibody-substrate complex requires that the ligand
be in fast exchange relative to the NOE relaxation rate (τexchange
/
T1bound < 0.2).30 To determine if the fast exchange requirement
was satisfied by the AZ-28 system, line-broadening experiments
were performed at low substrate/antibody stoichiometries (1:1
to 4.4:1). A titration series shows the sharpening and shifting
of the substrate resonances as the substrate:antibody ratio is
increased (Figure 2), confirming rapid exchange on the NMR
time scale. Addition of hapten 3 (1.1 equiv relative to antibody)
to displace the substrate from the antibody binding site gave
the final titration point. The difference in shift between the
free and bound resonances is 12 Hz, providing a lower limit of
12 s-1 for the substrate exchange rate.
Initial resonance assignments and studies on the feasibility
of applying tr-NOE to the AZ-28 system employed 2D-NOESY
spectroscopy. Spectra (τm ) 600 ms) of the free substrate 1 in
the absence of antibody AZ-28 exhibit observable crosspeaks
between olefinic protons and their nearest aromatic neighbors,
allowing complete assignment of each olefin proton resonance.
Additionally, the NOE crosspeaks are of opposite sign from
the diagonal peaks, characteristic of the short correlation times
observed in small molecule NMR spectroscopy. No crosspeaks
were observed between the terminal olefins pairs (Ha,b and Hc,d;
see Figure 1), indicating the solution structure of 1 is ap-
proximated by an extended conformation. NOESY spectra (τm
) 200 ms) of the substrate in the presence of antibody exhibit
crosspeaks of the same sign as the diagonal peaks, verifying
that AZ-28 binds the substrate and greatly increases its cor-
relation time. However, attempts to observe an NOE crosspeak
between the terminal olefin pairs of the substrate in the bound
Structural information on rapidly exchanging bound com-
plexes is readily available through transferred NOE spectroscopy
(30) Sykes, B. D. Curr. Opin. Biotech. 1993, 4, 392-396.
(31) Campbell, A. P.; Tarasow, T. M.; Massefski, W.; Wright, P. E.;
Hilvert, D. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 8663-8667.
(29) Hodgson, D. J.; Rychlewska, U.; Eliel, E. L.; Manoharan, M.; Knox,
D. E.; Olefirowicz, E. M. J. Org. Chem. 1985, 50, 4838-4843.

Studies of Antibody-Catalyzed Pericyclic Rearrangement
J. Am. Chem. Soc., Vol. 120, No. 9, 1998 1955
Figure 2. Spectra, from top to bottom, following the progress of a
titration series of substrate 1 into antibody AZ-28. Indicated on the
spectra are the equivalents of substrate present with respect to antibody
concentration. The gradual appearance and steady shift of substrate
resonances over the duration of the titration series confirms the fast-
exchange nature of this system. The final titration point also included
the addition of 1.1 equiv (with respect to antibody) of hapten 3, to
fully compete substrate out of the antibody binding site. Resonances
growing in at 5.36, 5.17, and 5.15 ppm are those of product trapped
by Girard’s Reagent T.
state via NOESY were complicated by antibody catalysis of
the rearrangement, limiting useful sample lifetime to under 5
h.
To overcome the extended data acquisition times required
for 2D-NOESY analysis, 1D saturation tr-NOE experiments
were performed. One-dimensional spectra of substrate 1 in the
presence of antibody (10:1 ratio) clearly show an NOE between
the Ha,b and Hc,d olefin pairs (Figure 3d,e,f). This result indicates
that the terminal olefins of the substrate are in close proximity
to each other (within 5 Å) in the antibody-substrate complex.
In order to achieve such close proximity, the substrate must
adopt a cyclic conformation, in accord with the structure of
hapten 3. The NOE intensities indicate that Hb is closer in
proximity to Hc,d than is Ha, with the possibility that transfer
from Ha is mediated by spin diffusion through Hb. Saturation
tr-NOE experiments were performed between the terminal
olefins in both directions (Ha,b to Hc,d and Hc,d to Ha,b) to rule
out the possiblity that the observed NOEs are due to magnetiza-
tion transfer from amino acid side chain protons within the
saturation frequency envelope. One-dimensional tr-NOE ex-
periments performed on the substrate in the presence of
antibody, 20 equiv of substrate, and 1.1 equiv of hapten 3
revealed no tr-NOEs between the Ha,b and Hc,d olefin pairs
(Figure 3c), confirming that the terminal olefins of the hexadiene
are spatially distant in the unbound state and that nonspecific
binding does not contribute to any NOEs observed for the
antibody-substrate complex. Additionally, it was noted that
the preirradiation was selective enough to avoid spillover
excitation of the Hb proton when selectively saturating Hd.
To assess the contribution of spin diffusion to the observed
NOEs, 1D ROE experiments were conducted under the same
solution conditions used for the NOE experiments. Quantitation
of the spin diffusion in this experiment was hindered by limited
selectivity in excitation and T1F relaxation during the spin lock.
Figure 3. 1H NMR spectra at 600 MHz of the olefin region of substrate
1. (A): reference spectrum. (B,C): difference NOE spectra of substrate
competitively displaced from the active-site by hapten 3 with Ha and
Hd irradiated, respectively. No signal transfer is observed between olefin
pairs. (D-F): difference NOE spectra of bound substrate with Ha, Hb,
and Hd irradiated, respectively. Observed transfer between olefin pairs
indicate the allylic termini of the substrate molecule are in close
proximity while bound by the antibody. Relative NOE intensities
indicate that Hb is closer in proximity to Hc,d than is Ha. The positive
amplitude of the Hb peak in B is indicative of an NOE attributable to
a small molecule, confirming that substrate is effectively displaced from
the binding site by hapten. In contrast, the negative amplitude of the
transferred NOE’s in (D), (E), and (F) result from antibody binding
the substrate molecule and lowering its overall correlation time. The
inset in B displays a dispersive subtraction artifact, while the inset in
D highlights a tr-NOE of interest. Spectra B-F represent the difference
between spectra which were acquired with 1 s of preirradiation of the
indicated protons and spectrum A. The preirradiation pulse is not
selective enough to avoid perturbing Hc while irradiating Hd (as
evidenced by the inverted Hc peak in C). However, the pulse does
resolve Hb as no spillover signal is observed in C.
However, the 1D ROE results support the conclusion that the
Hb to Hc,d transfer is direct, while that from Ha is partially via
spin diffusion through Hb. In conclusion, saturation tr-NOE
and tr-ROE results clearly show that the substrate is preorga-
nized into a cyclic conformation in the antibody-substrate
complex, as intended by the hapten design.
Activation Parameters. To determine the extent to which
conformational restriction of the substrate in the antibody-
substrate complex contributes to the energetics of catalysis, the
temperature dependence of both the uncatalyzed and the

1956 J. Am. Chem. Soc., Vol. 120, No. 9, 1998
Driggers et al.
antibody-catalyzed reactions was investigated. Rates of the
uncatalyzed rearrangement in 20 mM MES, 100 mM NaCl, pH
6.0, were measured at elevated temperatures between 37 and
50 °C, whereas rates of the antibody-catalyzed oxy-Cope
rearrangement were determined in the same buffer at temper-
atures between 5 and 25 °C. The temperature dependence of
the uncatalyzed rearrangement was fit to the standard Eyring
relationship, affording values for the enthalpy and entropy of
activation (∆Hq and ∆Sq) of 27.4 ( 1.3 kcal/mol and -3 ( 8
cal/mol K, respectively. This value of ∆Hq is in reasonable
agreement with reported values (32-38 kcal/mol) for similar
oxy-Cope rearrangements.32 The value for ∆Sq, however, is
higher than those found in previous studies using less polar
solvents20 and may reflect a decrease in ordered solvent around
the substrate (including water hydrogen bonded to the hydroxyl
group of 1) as the electron density in 1 becomes more
delocalized in the transition state.
The values of ∆Hq and ∆Sq for the antibody-catalyzed
reaction were determined to be 15.4 ( 2.4 kcal/mol and -23
( 8 cal/mol K, respectively. These values were determined
under conditions of saturating substrate (500 µM) and therefore
correspond to the kinetic constant kcat. This result indicates that
the antibody functions primarily by lowering ∆Hq rather than
-T∆Sq, a surprising result in light of the NMR structural data
which indicate that the substrate is preorganized into a cyclic
conformation upon binding to the antibody. These data argue
that the mechanism of catalysis is more complex than the simple
conformational restriction of substrate revealed by the NMR
experiments.
bond and little bond making at the C1 and C6 termini.35 The
differences between our results and the previously reported
isotope effects likely reflect the ability of the diphenyl substit-
uents in 1 to stabilize a biradicaloid transition state, versus
anionic weakening of the C3-C4 bond alone. Thus, it appears
in the case of substrate 1 that the antibody-catalyzed reaction
involves significant bond making in the transition state,
consistent with an activated complex having considerable
biradical character.
Kinetic isotope effects should also allow us to determine
whether the chemical rearrangement step is partially or wholly
rate limiting in the antibody-catalyzed reaction, as no significant
kinetic isotope effect would be expected for either isomerization
of the enol product to the aldehyde or the release of product
from the binding site. In order to directly measure the effects
of a secondary isotope effect on the rate of the antibody-
catalyzed reaction, substrate 23 with deuterium substitution on
both allylic termini was synthesized. The R-DV KIE was
determined to be 0.61 ( 0.1 under conditions of saturating
substrate (1 mM) at room temperature, again similar to values
determined by Gajewski for the analogous Cope rearrangement34
(this value normalizes to 0.78 ( 0.1 per D2 for comparison with
the R-D(V/K) KIE determined in the competition experiments).
The similarity of the R-DV and R-D(V/K) KIEs indicates that
the transition state is the same for both kcat and kcat/Km, further
confirming that the chemical rearrangement step contributes
significantly to the observed rate of the antibody-catalyzed
reaction. Thus, the measured activation parameters correspond
to the energetics of the pericyclic rearrangement step rather than
to a product dissociation step, a result consistent with the
observed exchange kinetics (given the substrate and oxime-
trapped product exchange rates, >12 s-1 and >45 s-1, respec-
tively, it would be surprising if product release was rate
determining in this reaction).
Mechanistic Considerations. The three-dimensional crystal
structure of the recombinant AZ-28 Fab-hapten 3 complex
determined to 2.6 Å resolution has recently been reported,
contributing to a complete picture of catalysis in antibody AZ-
28.2 The active site is shown in Figure 4 and described briefly
below as a framework for the mechanistic discussion that
follows. The structure reveals that the hapten is bound in a
deep cylindrical cavity in the antibody combining site and retains
the chair-like geometry expected in the ground state of an
equatorially substituted cyclohexanol. The phenyl substituents
of hapten 3 are rotated with respect to each other by a dihedral
angle of 17 ° and make extensive contacts with the active site
residues. The 5-phenyl group of hapten 3 is tightly packed at
the bottom of the cavity and surrounded by a large number of
aromatic and hydrophobic side chains. The 2-phenyl substituent
is located near the opening of the binding pocket and its
orientation is fixed by a π-stacking interaction with the
imidazole ring of HisH96 (heavy-chain histidine 96) and van
der Waals interactions with the side chain of TyrL91 (light-
chain tyrosine 91). The cyclohexyl ring of hapten 3 is rotated
out of the planes of the 5- and 2-phenyl rings by 81° and 85°,
respectively, an expected conformation based on known struc-
tures of phenylcyclohexanes.29 Its position is fixed by hydrogen-
bonding interactions between the hydroxyl substituent of hapten
3 and both the imidazole ring and backbone amide NH group
of HisH96. There are also hydrogen-bond contacts with the
carboxylate groups of GluH35 and GluH50, with the latter
interaction bridged by a water molecule. The cyclohexyl ring
Kinetic Isotope Effects. To gain greater insight into the
mechanism of the antibody-catalyzed reaction, secondary deu-
terium kinetic isotope effects (KIE’s) on both kcat (DV) and kcat/
Km (DV/K) were measured. Due to the rapid tautomerism of
enol to keto form in oxy-Cope rearrangement products, the
reaction is generally treated as irreversible, and to date only a
single thermal retro-oxy-Cope rearrangement has been ob-
served.33 This irreversibility allows an R-D(V/K) kinetic isotope
effect (KIE) determined under Vmax conditions to be interpreted
as a specific measure of the isotope effect on the pericylic
rearrangement, the last step prior to the tautomerization.
Competition experiments to determine R-D(V/K) were carried
out between the deuterated and nondeuterated substrates 18 and
16a, respectively, under conditions of saturating substrate in
the presence of 10 µM antibody. Electron impact mass spectral
analysis afforded a value of R-D(V/K) ) 0.72 ( 0.03. The
magnitude of the isotope effect indicates a significant change
in hybridization at the allylic termini in the transition state and
is comparable to the value 0.64 ( 0.04 determined previously
for the Cope rearrangement of 1,1,6,6-tetradeuterio-2,5-diphenyl-
1,5-hexadiene34 (this value is 0.80 ( 0.03 if normalized to the
dideuterio compound). In the latter case, the size of these
isotope effects have been interpreted as supporting a mechanism
in which there is a significant degree of bonding between
the allylic termini at the transition state, with the possibility of
a marginally stable biradicaloid intermediate. In contrast,
secondary deuterium isotope effects measured for the rearrange-
ment of the potassium alkoxide of 3-methyl-1,5-hexadien-3-ol
in organic solvents have been interpreted in terms of a
mechanism involving substantial bond breaking of the C3-C4
(32) Evans, D. A.; Golob, A. M. J. Am. Chem. Soc. 1975, 97, 4765.
(33) Elmore, S. W.; Paquette, L. A. Tetrahedron Lett. 1991, 32, 319-
322.
(34) Gajewski, J. J.; Conrad, N. D. J. Am. Chem. Soc. 1979, 101, 6693-
6704.
(35) Gajewski, J. J.; Gee, K. R. J. Am. Chem. Soc. 1991, 113, 967-
971.

Studies of Antibody-Catalyzed Pericyclic Rearrangement
J. Am. Chem. Soc., Vol. 120, No. 9, 1998 1957
density at C2 or C5 with the aromatic rings in a transition state
with considerable biradicaloid character. In the crystal structure
of the antibody-hapten 3 complex, the two phenyl substituents
of 3 bind in an almost parallel orientation with respect to each
other, allowing both to conjugate with the 4π + 2σ, orbital
system simultaneously. However, if substrate 1 binds to AZ-
28 in the same conformation as the hapten in the crystal
structure, then the 1,5-hexadiene group must rotate with respect
to the phenyl substituents to achieve conjugative stabilization
in the transition state. This rotation would require reorganization
of the extended network of van der Waals and hydrogen-bonding
interactions between substrate and the active site residues and
is consistent with the large negative entropy of activation
observed for the kcat step.
This interpretation is consistent with immunological studies
on the affinity maturation of AZ-28 from its germline precursor.
The germline-encoded antibody binds hapten 3 with 40-fold
lower affinity and has a rate enhancement (kcat/kuncat) of
163 000.2 This difference in rate largely results from the somatic
mutation of active site serine 34L in the germline antibody to
an asparagine in AZ-28. Asparagine 34L directly contacts the
cyclohexyl ring and also interacts with a number of residues
that help fix the conformation of bound hapten by fixing the
configuration of complementarily determining region (CDR) 3
of the heavy chain. Thus, this residue might be expected to
play a significant role in restricting substrate rotation around
the C2 and C5 aryl bonds. Consequently, mutation of Asn 34L
in AZ-28 to serine may facilitate rotation about these bonds,
increasing the bound concentration of the favorable rotamer and
enchancing kcat in the germline antibody. Consistent with this
model, changes in the structure of CDR 3 of the heavy chain,
which plays an important role in fixing the orientation of the
2-phenyl substituent, also lead to large variations in catalytic
activity in a series of related antibodies.36 While it is unclear
whether the enhanced rate of the germline-encoded antibody
results from improved overlap between the hexadiene system
and one or both aryl substituents, reorganization in the case of
AZ-28 likely involves rotation of the hexadiene core and one
or both of the aryl rings relative to each other.
In addition to this π-overlap of the aryl substituents, electronic
effects arising from the 3-hydroxy substituent may lower the
enthalpy of the [3,3]-sigmatropic rearrangement reaction. The
anionic substituent effect accelerates the oxy-Cope rearrange-
ment through hyperconjugation of electron density on oxygen.22
Consequently, the side chains HisH96, GluH35, and GluH50,
which hydrogen bond to the 3-hydroxyl group of the substrate,
might influence the rate of the rearrangement by increasing or
decreasing the electron density on the oxygen substituent of
the substrate. In the case of anionic catalysis of the oxy-Cope
rearrangement of 3-methyl-1,5-hexadien-3-ol in solution, a small
secondary deuterium kinetic isotope effect is observed for
isotopic substitution at the allylic termini (approximately 0.96
( 0.05).35 However, the magnitude of the C1-C6 isotope effect
in a C2,C5-aryl-substituted hexadienol undergoing anionic
catalysis has never been studied and may be closer to the value
observed here as a result of the aryl substitution.
Figure 4. Active site of AZ-28 Fab-hapten 3 crystal structure showing
the orientation of the 2- and 5-phenyl substituents relative to the
cyclohexadienol core, as well as the position of key residues which
contact the hapten. Side chains and peptide backbone are brown, and
the hapten is lavender. Heteroatoms are color-coded blue for nitrogen
and red for oxygen.
of hapten 3 is also in van der Waals contact with the
hydrophobic side chains of the light chain and the hydrophilic
side chains of AsnL34 and AspH101.
In the context of this structural information, the NMR and
kinetic experiments described above allow one to formulate the
following mechanistic proposal for the antibody-catalyzed
reaction. The substrate, which exists predominantly in an
extended conformation in solution, is bound in a rapid, reversible
step in the active site and its conformation is fixed in a cyclic
geometry. This conformational restriction likely involves
extended interactions with the aryl substituents, which orient
the C2 and C5 carbons, as well as hydrogen-bonding interactions
with the hydroxyl group of 1. The latter interactions are
consistent with the kinetic preference of the antibody for the
(S)-substrate stereochemistry at C3. Binding interactions with
the active site thus lead to a cyclic alignment of the 4π + 2σ
orbital system of the hexadiene core, contributing to the overall
rate enhancement of the antibody-catalyzed reaction.
The Vmax kinetic isotope effect experiments indicate that the
[3,3]-sigmatropic rearrangement is rate limiting in the antibody-
catalyzed reaction, while the magnitude of the R-DV/K secondary
isotope effect points to a significant degree of bond formation
between carbons C1 and C6 in the transition state. As discussed
above, this effect can arise from either a biradicaloid transition
state leading to a resonance stabilized 1,4-biradical intermediate
or from a concerted reaction in which bond breaking and
formation occur simultaneously in a single transition state. In
the case of the Cope rearrangement, it is known that phenyl
substituents at the 2 and 5 positions of 1,5-hexadiene reduce
the activation energy by approximately 10 kcal/mol relative to
the unsubstituted diene,21 attributable to conjugation of electron
The final steps in the antibody-catalyzed reaction involve
tautomerism of the resulting enol group to an aldehyde and
release of the product, both of which are rapid in comparison
to the rate of the rearrangement reaction. It is necessary to trap
this aldehyde with an added nucleophile to prevent time-
dependent inactivation (see above), presumably due to imine
formation between antibody and product. It has not been
(36) Ulrich, H. D.; Schultz, P. G. J. Mol. Biol. In press.

1958 J. Am. Chem. Soc., Vol. 120, No. 9, 1998
Driggers et al.
determined whether the antibody contains a specific site for
hydroxylamine binding or whether the trapping reaction occurs
in the active site or after diffusion of substrate from the active
site. Ultimately, this trapping reaction could provide the
basis for a screen of AZ-28 mutants with enhanced catalytic
activity.
The mechanistic proposal presented here suggests a number
of experiments both to test this model and to enhance the
efficiency of the AZ-28-catalyzed reaction. Specifically, modi-
fications to the hapten structure which restrict the aryl substit-
uents to be coplanar with the cyclohexyl ring may afford
antibodies with increased rates. In addition, site-directed
mutagenesis experiments which convert HisH96, GluH35, and
GluH50 to residues with modified hydrogen-bonding capacities
should provide insight into the role of electron density on the
oxygen of the 3-hydroxyl substituent in catalysis. Determination
of activation parameters for the germline antibody would help
to validate the mechanistic model, as would the crystal structure
of the germline antibody. Several of these investigations are
currently underway.
Finally, the mechanism of antibody AZ-28 can be compared
in general to those of other proteins that catalyze [3,3]-
sigmatropic rearrangements. Structural and mechanistic studies
have been carried out on one antibody (1F7)37 and three enzymes
(Bacillus subtilis, Escherichia coli, and yeast chorismate mu-
tases)38,39 that catalyze the [3,3]-sigmatropic rearrangement of
chorismate to prephenate. Similar studies have been carried
out on an antibody which catalyzes a Diels-Alder reaction
(39A11).40 In each of these cases, the protein appears to
preorganize the substrate into a reactive conformation by a
network of hydrogen-bonding, electrostatic, and van der Waals
interactions. In the case of AZ-28, however, additional
unanticipated conformational adjustments appear to be required
to reach the transition state. In the active sites of the B. subtilis
chorismate mutase and antibody 1F7, hydrogen bonds from an
active site arginine to the enol ether oxygen of chorismate have
been postulated to stabilize the developing charge at this center
in the transition state. Similarly, the crystal structure of the
39A11 Fab-hapten complex shows a specific hydrogen-bonding
interaction which likely serves to reduce the electron density
on the substrate dienophile and accelerate the antibody-catalyzed
reaction. In the case of AZ-28, the role of active site hydrogen
bonding is less clear, and further studies will be required to
determine if hydrogen-bonding interactions with the 3-hydroxy
substituent of the substrate serve to accelerate or retard the
reaction. It appears, however, that, for all of these pericyclic
reactions, Diels-Alder, Claisen, and oxy-Cope, catalysis in-
volves both the restriction of rotational and translational entropy
in the substrate, as well as hydrogen-bonding interactions which
modulate electron densities on key substituents in the transition
state.
The studies reported here demonstrate the role antibodies can
play in understanding biological catalysis, particularly when
information from small-molecule NMR is combined with more
traditional kinetic and structural probes of enzyme mechanism.
They also provide a basis for further improvement in the rates
of antibody-catalyzed reactions.
Acknowledgment. We thank Professor Barry K. Carpenter
for generously providing the computer program used to generate
the activation parameters. P.G.S. is a Howard Hughes Medical
Institute Investigator. P.G.S., E.M.D.,C.P.K., and A.C.B. are
grateful for financial support for this work from the National
Institutes of Health (Grant no. R01 AI24695). D.E.W., H.S.C.,
and C.W.L. have been supported by the Director, Office of
Energy Research, Office of Biological and Environmental
Research, General Life Science division of the U.S. Department
of Energy under contract no. DE-AC03-76F00098. Instrumenta-
tion grants to D.E.W. were provided by the U.S. Department
of Energy (DE-FG05-86ER75281) and NSF (DMB86-09305
and BBS 87-20134).
(37) Haynes, M. R.; Stura, E. A.; Hilvert, D.; Wilson, I. A. Science 1994,
263, 646-652.
(38) Lee, A. Y.; Stewart, J. D.; Clardy, J.; Ganem, B. Chem. Biol. 1995,
2, 195-203.
(39) Guilford, W. J.; Copley, S. D.; Knowles, J. R. J. Am. Chem. Soc.
1987, 109, 5013-5019.
(40) Romesberg, F. R.; Spiller, B.; Schultz, P. G.; Stevens, R. C.
Submitted for publication.
JA962933U