Antibody-catalyzed decarboxylation and aldol reactions using a primary amine molecule as a functionalized small nonprotein component

Article abstract of DOI:10.1016/j.bmc.2013.09.015

Catalytic antibody 27C1 bears binding sites for both a substrate- and a functionalized small nonprotein component in the active site. We investigated the possibility of exploiting imine and enamine intermediates using a primary amine molecule into the act

Full text of DOI:10.1016/j.bmc.2013.09.015

Bioorganic & Medicinal Chemistry 21 (2013) 7011–7017
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Antibody-catalyzed decarboxylation and aldol reactions
using a primary amine molecule as a functionalized small
nonprotein component
Fumihiro Ishikawa a,b, Kouki Uno a, Masao Nishikawa a, Takeshi Tsumuraya a, Ikuo Fujii a,
⇑
a Department of Biological Science, Graduate School of Science, Osaka Prefecture University, 1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8531, Japan
b Department of System Chemotherapy and Molecular Science, Division of Bioinformatics and Chemical Genomics, Graduate School of Pharmaceutical Sciences, Kyoto University,
Sakyo, Kyoto 606-8501, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Catalytic antibody 27C1 bears binding sites for both a substrate- and a functionalized small nonprotein
component in the active site. We investigated the possibility of exploiting imine and enamine interme-
diates using a primary amine molecule into the active site of antibody 27C1. The antibody catalyzed b-
keto acid decarboxylation with a rate enhancement (kcat/Km/kuncat) of 140,000, as well as highly regiose-
lective cross-aldol reactions of ketones and p-nitrobenzaldehyde. These studies provide new strategies
for the generation of catalytic antibodies possessing binding sites for functionalized components.
Ó 2013 Elsevier Ltd. All rights reserved.
Received 11 August 2013
Revised 5 September 2013
Accepted 6 September 2013
Available online 19 September 2013
Keywords:
Catalytic antibody
Decarboxylation reaction
Regioselective aldol reactions
Enamine mechanism
1. Introduction
catalytic antibodies has thus yielded numerous efficient syntheses
of stereochemically complex molecules.21–24
The study of amine-catalyzed decarboxylation and aldol reac-
tions remains a special area of research in synthetic chemistry, bio-
Recently, we developed catalytic antibody 27C1, which bears an
antigen-combining site that functions as an apoprotein for binding
functionalized small nonprotein components.25 This antibody was
elicited by immunization with the haptenic phosphonate diester 1.
The p-nitrophenyl and N-acetylphenyl groups in hapten 1 were de-
signed to elicit binding sites for the substrate and the functional-
ized component, respectively, in the antigen-combining site
(Fig. 1a). With a simple exchange of the functionalized component,
antibody 27C1 is capable of catalyzing a wide range of chemical
transformations including acyl-transfer, b-elimination, decarboxyl-
ation, and aldol reactions. Particularly, antibody 27C1 catalyzed
the aldol reaction of acetone and p-nitrobenzaldehyde 4 in the
presence of the functionalized component 2 (Fig. 1b).25 Catalysis
by 27C1 was efficient compared with the non-catalyzed reaction,
organic chemistry and enzymology.1–4
A number of elegant
enzymatic studies elucidated the mechanism by which acetoace-
tate decarboxylase catalyzes the decarboxylation of acetoacetic
acid.5–14 These studies demonstrated that the reaction proceeds
by the formation of a Schiff base between the e-amino group of a
lysine residue in the enzyme and acetoacetate, followed by decar-
boxylation to form an enamine, which is then tautomerized to a
Schiff base and subsequently hydrolyzed to release acetone and
the free enzyme. The imine and enamine intermediates developed
along the reaction catalyzed by acetoacetate decarboxylase are also
found along the reaction coordinate of reactions between aldolases
and aldol donors. In fact, natural class I aldolases and the pro-
grammed catalytic antibodies 38C2 and 33F12, which both have
an active-site lysine residue are bifunctional catalysts. In addition
to catalyzing the decarboxylation of b-keto acids, they also catalyze
the aldol reaction of aldehydes and ketones.15–19 Unlike natural en-
zymes, antibodies 38C2 and 33F12 were found to accept a variety
of ketones and aldehydes as aldol donors and acceptors to achieve
regio- and enantioselevtive aldol reactions.20 The application of
showing
a
rate enhancement [(kcat/Km 4)/kuncat of 4.4 Â 104
(27C1: kcat = 25.8 minÀ1, Km for 4 = 958
lM, Km for 2 = 67 mM). In
the aldol reaction, the direct precursor of the enamine is an immin-
ium ion; the presence of the imminium ion was established by iso-
lation of the reduction product following NaBH4 treatment.
Examination of enamine formation with acetone and amine 2 in
the presence of NaBH4 showed that antibody 27C1 catalyzed the
formation of an isopropylation product, providing evidence for an
enamine mechanism for the antibody-catalyzed reaction. This en-
amine formation was inhibited by addition of hapten 1, showing
that the enamine was formed in the antigen-combining site. In this
⇑
Corresponding author. Tel./fax: +81 72 254 9834.
E-mail address: fujii@b.s.osakafu-u.ac.jp (I. Fujii).
0968-0896/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2013.09.015

7012
F. Ishikawa et al. / Bioorg. Med. Chem. 21 (2013) 7011–7017
Figure 1. (a) Decarboxylation and aldol reactions catalyzed by antibody 27C1 and structure of the immunized hapten 1. (b) The mechanism of the antibody-catalyzed aldol
reaction.
work, we have investigated the possibility of exploiting the com-
mon imine and enamine intermediates using a primary amine mol-
ecule in the antibody-combining site. The primary amine mimics a
lysine residue correctly aligned in the active sites of natural en-
zymes and the catalytic antibodies 38C2 and 33F12. Here, we dem-
onstrate the utility of the amine molecule for catalyzing the
decarboxylation of a b-keto acid, as well as aldol reactions using
structurally related aldol donors (Fig. 1a).
component concentration to provide the true maximum rate Vmax
(12.9
M minÀ1) and the Michaelis constant, Km (Fig. 2). The uncat-
l
alyzed reaction of b-keto acid 3 with amine 2 was monitored by
high performance liquid chromatography (HPLC) as a pseudo-
first-order reaction. Product formation was assayed from 0 to
540 min and the second-order rate constant was calculated
(kuncat = 4.55 Â 10À6 mMÀ1 minÀ1). The kuncat can be considered
as kamine, amine 2-catalyzed decarboxylation of b-keto acid 3.
Catalysis by 27C1 was remarkably efficient compared with the
non-catalyzed reaction, showing a large rate enhancement [(kcat
/
2. Results
Km 3)/kuncat (the relative efficiencies over amine catalysis)] of
140,000; (kcat (per binding site) = 1.29 min–1, Km (for 3) = 2.1 mM,
Km (for 2) = 6.7 mM, pH 8.0) (Fig. 2).
2.1. Decarboxylation reaction
Antibody 27C1 catalyzed the decarboxylation reaction of b-keto
acid 3 to afford (p-nitrophenyl)acetone. Lineweaver–Burk plots
were constructed by holding the either the substrate or the func-
tionalized component constant while varying the concentration
of the other (Fig. 2). The slopes and y-intercepts obtained from this
analysis were replotted as a function of substrate or functionalized
2.2. Cross-aldol reactions
To demonstrate the versatility and define the scope of antibody
27C1 for the cross-aldol reactions, we tested a variety of commer-
cially available ketones 5–10 as donors and p-nitrobenzaldehyde

F. Ishikawa et al. / Bioorg. Med. Chem. 21 (2013) 7011–7017
7013
Figure 2. (A) Lineweaver–Burk plot with 3 held at four fixed concentrations while 2 was varied over concentrations ranging from 1.5 to 10 mM (d, [3] = 0.5 mM; j,
[3] = 1 mM; ꢀ, [3] = 2 mM; N, [3] = 3 mM); v, velocity. (B) Replot of the y-intercepts and slopes of the Lineweaver–Burk plot as a function of [3]À1. (C) An analogous
Lineweaver–Burk plot was constructed with 2 held at four fixed concentrations while 3 was varied over concentrations from 0.5 to 3 mM (d, [2] = 1.5 mM; j, [2] = 2 mM; ꢀ,
[2] = 5 mM; N, [2] = 10 mM). (D) An analogous plot was constructed to provide the kinetic constants for 2. The reaction mixtures contained 5
DMSO, and 50 mM Tris–HCl pH 8.0. The mixtures were incubated at 25 °C. All kinetic assays were measured in duplicate.
lM antibody 27C1, 10% (v/v)
4 as an acceptor (Fig. 1a and Table 1). Aldehyde 4 was chosen as the
standard acceptor because, like hapten 1, it bears a p-nitrophenyl
group and should be specifically recognized by the antibody. To as-
sign regioselectivity, the aldol products were synthesized by stan-
dard methods (see Section 5). To compare the reactivity of aldol
donors, we determined specific rates for the antibody-catalyzed
reactions under the following defined conditions: 5% donor,
1 mM of p-nitrobenzaldehyde 4, 10 mM of the primary amine 2,
isomer 15 and branched isomer 15a (44:56) (Table 2). These re-
sults suggest that under these conditions, the non-catalyzed reac-
tions are exclusively under thermodynamic control. Using the
antibody, we found that the product distribution is a function of
the catalyst, the acceptor and the donor. Antibody 27C1 predomi-
nantly afforded the linear regioisomers. The reaction of 2-butanone
(6) with 4 produced the linear isomer 12 exclusively (12:12a, >99).
The reaction of 2-pentanone (7) with 4 predominantly produced
the linear isomer 13 (13:13a, 90:10). The reaction of 4-hydroxy-
2-butanone (9) with 4 exclusively provided linear isomer 15
(15:15a, >99). The reaction of 2-hexanone (8) with 4 was moder-
and 10 lM of 27C1 in 50 mM Tris–HCl pH 8.0. As shown in Table 1,
a variety of different ketones underwent the antibody-catalyzed al-
dol reactions with aldehyde 4, as proven by comparison of the
HPLC retention times of the aldol products with those of indepen-
dently chemically synthesized standards: aliphatic open chain
(acetone, 2-butanone, 2-pentanone, and 2-hexanone), and func-
tionalized open chain (4-hydroxy-2-butanone). All antibody-cata-
lyzed aldol reactions were inhibited by addition of hapten 1 or
removal of functionalized molecule 2. Cyclobutanone was not ac-
cepted by antibody 27C1, presumably because of its bulkiness.
ately regioselective, producing
a mixture of the linear and
branched isomers 14 and 14a (14:14a, 62:38).
3. Discussion
The catalytic efficiency of antibody 27C1 as a decarboxylase can
be assessed by comparing decarboxylation reactions catalyzed by
the antibody, simple amines,26–29 catalytic antibodies 38C2 and
33F12,18 and the enzyme acetoacetate decarboxylase (AAD).5–
2.3. Regioselectivity of the aldol reactions
13,30
The most effective amine catalyst for the decarboxylation of
acetoacetate is aminoacetonitrile, AAN.28 This amine molecule
has been used as a model for AAD-catalyzed decarboxylations be-
cause its pKa value approximates that of the active-site lysine of
AAD. The rate of AAD-catalyzed decarboxylation of acetoacetate
is 156 times the rate of decarboxylation catalyzed by AAN, suggest-
ing that most of AAD’s catalytic ability can be attributed to imine
formation with the amino group of the active site lysine.27,30
AAD-, 38C2-, and 33F12-catalyzed decarboxylation reactions pro-
ceeded in single-substrate manner. In contrast, antibody 27C1 cat-
alyzed reactions between two substrates (b-keto acid 3 and amine
We set out to study the degree of regiocontrol that antibody
27C1 could exercise on the aldol reactions. For all experiments,
the reaction products were proven by comparison of the retention
times of the products and independently chemically synthesized
standards using HPLC. We performed the reactions under the fol-
lowing defined conditions: 5% donor, 500 mM 4, 500 mM amine
2, and 10 mM antibody in 50 mM Tris–HCl pH 8.0 at 25 °C. Control
experiments revealed that, in the absence of antibody, the products
were the syn and anti stereoisomers resulting from an addition of
the more substituted carbon of 2-butanone, 2-pentanone, and 2-
hexanone to the aldehyde (Table 2). When 4-hydroxy-2-butanone
was used as the donor, the products were a mixture of the linear
2). Thus, we could not directly compare the rate enhancement, kcat
/
kuncat for 27C1-catalyzed decarboxylation reaction with those of

7014
F. Ishikawa et al. / Bioorg. Med. Chem. 21 (2013) 7011–7017
Table 1
amine 2 in the antigen-combining site of 27C1 has a pKa value sim-
ilar to those of the active-site lysines of AAD and 38C2. The anti-
Donor promiscuity of antibody 27C1: specific rates (l l
mol product hÀ1 molÀ1
antibody) for antibody-catalyzed cross-aldol reactions of a variety of ketones with
p-nitrobenzaldehyde 4 under the following defined conditions: 5% donor, 1 mM 4,
bodies 38C2 and 33F12 provide
a 10,000–25,000-fold rate
enhancement (kcat/Km/kamine) for the decarboxylation of structur-
ally related b-keto acid. These catalytic activities are 10-fold lower
than that of 27C1 (Table 3). The effectiveness of 27C1 results from
the higher effective molarity of the active-site amine molecule
(2.8 Â 102 M) compared with the active-site lysines of 38C2
(24 M) and 33F12 (18 M) (Table 3). Efficient decarboxylation in
antibody 27C1 presumably reflects the entropic advantage derived
from correctly alignment of the perturbed amine molecule with
the carbonyl carbon of b-keto acid 3.
Antibody 27C1 is capable of accepting a variety of ketones for
intermolecular aldol reactions with p-nitrobenzaldehyde. Previous
biochemical studies of this catalyst suggested an enamine mecha-
nism shared with the natural class I aldolase enzymes except for
the presence of a primary amine molecule instead of a lysine resi-
due. A primary amine molecule in the antigen-combining site func-
10 mM 2, and 10
l
M antibody. R = p-nitrophenyl
Donor
Product(s)
Specific rate
55.0
7.6
6.6
tions as a chemically unique lysine residue bearing an e-amino
0.2a
group, which has a highly perturbed pKa allowing for efficient
amine-based catalysis under conditions where a more typical
amine would be protonated and ineffective in this chemistry.
Amang the aldol reactions, the reaction involving 4-hydroxy-2-
butanone as the donor was significantly more efficient than those
involving the other ketones except acetone. Neverthless, these
studies suggest that the donor specificity of 27C1 is broad, albeit
limited to open-chain ketones. These experiments also indicate
an active site geometry which strictly recognizes substitutions
such as cyclobutanone along with the designed hapten molecule.
The donor specificity may reflect the alkyl linker moiety in the hap-
ten, which generates antibody catalysts with broad substrate spec-
ificity.31 We found that unlike antibodies 38C2 and 33F12,18
antibody 27C1 favored formation of the linear regioisomer regard-
less of the structure of the aldol donors. The reactions involving
ketones 6–9 with p-nitrobenzaldehyde (4) all provided predomi-
nantly or exclusively the aldol product resulting from the less
substituted carbon. Although antibody 27C1 preferentially cata-
lyzed regioselective formation of the linear isomer, the antibody-
catalyzed aldol reactions gave no enantiomeric excess (ee). When
acetone was the aldol donor substrate in the p-nitrobenzalde-
hyde–ketone crossed aldol reaction, the product was a racemic
22.1
ndb
All kinetic assays were measured in duplicate.
a
Specific rate was determined with 2.5 mM 2.
b
Not determined.
natural enzyme and specialized catalytic antibodies. For ease, we
evaluated the relative efficiencies over amine catalysis, kcat/Km/
kamine. The large rate enhancement [(kcat/Km 3)/kamine] of 140,000
for 27C1-catalyzed decarboxylation is comparable to the rate
acceleration for 38C2-catalyzed decarboxylation of structurally re-
lated b-keto acids (Table 3). This result suggests that the primary
Table 2
Regioselectivity of antibody-catalyzed cross-aldol reactions of a variety of ketones (6–9) with p-nitrobenzaldehyde 4
Donor
Product (s)
Regioselectivity
Uncatalysed
Catalyzed
<99
96:4
83:17
10:90
38:62
92:8
56:44
<99
All kinetic assays were measured in duplicate.

F. Ishikawa et al. / Bioorg. Med. Chem. 21 (2013) 7011–7017
7015
Table 3
was performed on silica gel 60 F254 plates (Merck). Flash chroma-
tography was performed on silica gel 60 (230–240 mesh) (Merck).
HPLC was performed on a Hitachi L-2130 system equipped with an
L-2400 UV detector. 1H and 13C NMR spectra were taken on a JEOL
NMR spectrometer. Chemical shifts (d) are reported in ppm down-
field from tetramethylsilane. Mass spectral data were collected
with electrospray ionization (ESI) and electron ionization (EI) mass
spectrometers.
Kinetic parameters for antibody-catalyzed decarboxylation reactions
Catalyst
kcat/kuncat
kcat/Km/kuncat
kcat/kamine (M)
a
c
c
b
AAD
5.2 Â 109
1.4 Â 104
1.1 Â 104
9.5 Â 106
7.8 Â 107
c
38C2
33F12
27C1
2.5 Â 104
23.8c
c
1.0 Â 104
18.1c
1.4 Â 105
2.8 Â 102
a
kcat and Km of acetoacetate decarboxylase (AAD)-catalyzed decarboxylation of
acetoacetate from Highbarger et al.,30 kuncat from Westheimer.14
b
kamine
, aminoacetonitrile-catalyzed decarboxylation of acetoacetate from
5.2. Preparation of aldol products
Westheimer.14
c
kcat and Km of the 38C2 and 33F12-catalyzed decarboxylation from Björnestedt
et al.19 Amine = phenylalanine ethyl ester for 38C2 and 33F12.
5.2.1. Compounds 12, syn-12a, and anti-12a
To a solution of p-nitrobenzaldehyde (510 mg, 3.37 mmol) in
6 mL 2-butanone was added 600 lL of a 1% (w/v) aqueous NaOH
solution at 0 °C. Stirring was continued for 19 h at 0 °C. The reac-
tion mixture was then neutralized by addition of 1 M aqueous
HCl and concentrated in vacuo. The residue was extracted with
ether. The combined organic layers were dried over MgSO4, fil-
tered, and concentrated in vacuo. The residue was purified by HPLC
mixture of 11. When 2-butanone was used as the donor substrate,
the aldol product possessed the R-configuration with moderate
48% ee. These results suggest that the enamine formed by the reac-
tion between 2 and 2-butanone attacks on both the re- and si-faces
of p-nitrobenzaldehyde owing to the lack of additional interactions
to constrain the binding mode of the acceptor molecule in the anti-
gen-combining site. Further insight into the low enantioselectivity
of 27C1 will be provided by the X-ray crystal structure.
(YMC-Pack ODS-AM AM323: C-18 reverse-phase column,
u
10 mm  250 mm, acetonitrile/aqueous TFA (0.1%, 25:75), 3.0 mL/
min, 254 nm, tR: 12 = 44.0 min, 12a (syn and anti mix-
ture) = 38.5 min to give 12 as a white powder (193 mg, 26%) and
12a (syn and anti mixture) as a white powder (242 mg, 32%). The
syn and anti mixture 12a was continuously purified by HPLC
4. Conclusion
(YMC-Pack ODS-AM AM323: C-18 reverse-phase column,
u
These studies highlight the ability of the antigen-combining site
of antibody 27C1, elicited by immunization with the haptenic
phosphonate diester 1, to function as an apoproteins for binding
functionalized small nonprotein components. The programming
of the binding site for the functionalized component involved in
catalysis constitutes a unique feature of this approach and is as sig-
nificant as the use of transition state analogs to induce catalytic
antibodies.32 Ultimately, one would wish to combine both meth-
ods, enabling the experimenter to control both the global aspects
and the mechanistic details of the reaction. This approach offers
significant advantages over other methods for producing ‘special-
ized’ catalytic antibodies. It is possible to introduce additional cat-
alytic components (transition state stabilization, acid, and base
catalysis) by mutations of amino acids in the active site. Although
27C1 is the best catalytic antibody for decarboxylation, the rate
enhancement [(kcat/Km)/kuncat = 140,000] for the 27C1-catalyzed
decarboxylation of 3 is 104-fold lower than that reported for aceto-
acetate decarboxylase (Table 3). The large rate acceleration of 27C1
may shed light on the catalytic power of the amine molecule by the
proximity effect responsible for the enormous rate enhancements
observed for natural enzymes. As natural enzymes display the rate
enhancements of 106–1017 by a combination of several catalytic
mechanisms, it is expected that additional rate accelerations will
be realized by introducing not only functionalized small molecules
but also specific amino acid residues that function as transition-
state stabilization and general acid–base catalysis into the anti-
gen-combining site, using phage-displayed libraries33 or site-direc-
ted mutagenesis.34 The low enantioselectivity of aldol reactions
could be improved by altering the binding mode of aldol donor
and acceptor substrates through site-directed mutagenesis. Our re-
sults demonstrate that antibodies incorporating newly devised
components acting as ‘chemical teeth’ will enlarge the scope and
broaden the reaction boundaries of catalytic antibodies.
10 mm  250 mm, acetonitrile/aqueous TFA (0.1%, 20:80), 3.0 mL/
min, 254 nm, tR: syn-12a = 62.9 min, anti-12a = 66.0 min). 12: 1H
NMR (400 MHz, CDCl3):
d 8.21 (d, J = 8.8 Hz, 2H), 7.54 (d,
J = 8.8 Hz, 2H), 5.27 (dd, J = 4.0 Hz, 8.1 Hz, 1H), 2.81–2.84 (m, 2H),
2.49 (q, J = 7.3 Hz, 2H), 1.09 (t, J = 7.3 Hz, 3H). syn-12a: 1H NMR
(400 MHz, CDCl3): d 8.22 (d, J = 8.8 Hz, 2H), 7.51 (d, J = 8.8 Hz,
2H), 5.28 (d, J = 2.7 Hz, 1H), 2.83 (qd, J = 2.7 Hz, 7.3 Hz, 1H), 2.25
(s, 3H), 1.05 (d, J = 7.3 Hz, 3H). anti-12a: 1H NMR (400 MHz, CDCl3):
d 8.22 (d, J = 8.8 Hz, 2H), 7.52 (d, J = 8.8 Hz, 2H), 4.87 (d, J = 7.3 Hz,
1H), 2.91 (quin, J = 7.3 Hz,1H), 2.21 (s, 3H), 1.03 (d, J = 7.3 Hz, 3H).
12, LRMS (EI): m/z calcd for C11H13NO4, 223.08; found, 223. 12a,
LRMS (EI): m/z calcd for C11H13NO4, 223.08; found, 223.
5.2.2. Compound 13
2-Pentanone (110 lL, 1.0 mmol) was added to a freshly pre-
pared solution of LDA (10% (w/w) suspension in hexanes; 3.1 mL,
2.2 mmol) in 5 mL of THF at À78 °C. After stirring at À78 °C for
30 min, a solution of p-nitrobenzaldehyde (151 mg, 1.0 mmol) in
THF (5 mL) was added over a period of 1 min. After stirring at
À78 °C for 3 h, saturated NH4Cl solution was added, and the reac-
tion mixture was allowed to warm to room temperature. The prod-
uct was extracted with ethyl acetate, dried over MgSO4, and
evaporated in vacuo. The residue was purified by silica gel column
chromatography (1:4 to 3:7 EtOAc/hexane) to afford compound 13
as a yellow solid (105 mg, 58%). 1H NMR (400 MHz, CDCl3): d 8.22
(d, J = 8.8 Hz, 2H), 7.54 (d, J = 8.8 Hz, 2H), 5.25–5.29 (m, 1H), 3.64
(d, J = 3.2 Hz, 1H), 2.75–2.86 (m, 2H), 2.43 (t, J = 7.3 Hz, 2H), 1.63
(sext, J = 7.3 Hz, 2H), 0.93 (t, J = 7.3 Hz, 3H); LRMS (EI): m/z calcd
for C12H15NO4, 237.10; found, 237.
5.2.3. Compounds syn-13a and anti-13a
To a solution of p-nitrobenzaldehyde (200 mg, 1.32 mmol) in
3 mL of 2-pentanone was added 500 lL of a 1% (w/v) aqueous
NaOH solution at 0 °C. Stirring was continued at 0 °C for 15 h.
The reaction mixture was then neutralized by addition of 1 M
aqueous HCl and concentrated in vacuo. The residue was extracted
with ethyl acetate. The combined organic layers were dried over
MgSO4, filtered, and concentrated in vacuo. The residue was puri-
fied by HPLC (YMC-Pack ODS-AM AM323: C-18 reverse-phase col-
5. Experimental
5.1. General methods
All oxygen- or moisture-sensitive reactions were carried out
under N2. Analytical and preparative thin-layer chromatography
umn,
u
10 mm  250 mm, acetonitrile/aqueous TFA (0.1%, 35:65),

7016
F. Ishikawa et al. / Bioorg. Med. Chem. 21 (2013) 7011–7017
3.0 mL/min, 254 nm, tR: syn-13a = 24.8 min, anti-13a = 26.6 min)
to give syn-13a as a white powder (59 mg, 19%) and anti-13a as
a white powder (74 mg, 24%). syn-13a: 1H NMR (400 MHz, CDCl3):
d 8.21 (d, J = 8.8 Hz, 2H), 7.53 (d, J = 8.8 Hz, 2H), 5.10 (d, J = 4.4 Hz,
1H), 2.83 (dt, J = 4.4 Hz, 8.8 Hz, 1H), 2.17 (s, 3H), 1.68–1.79 (m, 1H),
1.53–1.63 (m, 1H), 0.85 (t, J = 7.3 Hz, 3H). anti-13a: 1H NMR
(400 MHz, CDCl3): d 8.22 (d, J = 8.8 Hz, 2H), 7.51 (d, J = 8.8 Hz,
2H), 4.92 (d, J = 6.8 Hz, 1H), 2.86 (dt, J = 6.8 Hz, 7.6 Hz, 1H), 2.13
(s, 3H), 1.58–1.69 (m, 1H), 1.43–1.54 (m, 1H), 0.91 (t, J = 7.0 Hz,
3H); LRMS (EI): m/z calcd for C12H15NO4, 237.10; found, 237.
2.19 (s, 3H). 15, LRMS (ESI–): [M–H]– calcd for C11H12NO4,
238.07; found, 238. 15a, LRMS (ESI–): [M–H]– calcd for
C11H12NO4, 238.07; found, 238.
5.2.6. Compounds syn-16 and anti-16
To a solution of p-nitrobenzaldehyde (151 mg, 1.0 mmol) in
1 mL of cyclobutanone was added 300 lL of a 1% (w/v) aqueous
NaOH solution at 0 °C. Stirring was continued at 0 °C for 15 min.
The reaction mixture was then neutralized by addition of 1 M
aqueous HCl. The mixture was extracted with EtOAc. The com-
bined organic layers were dried over MgSO4, filtered, and concen-
trated in vacuo. The residue was purified by HPLC (YMC-Pack ODS-
5.2.4. Compounds 14, syn-14a, and anti-14a
2-Hexanone (124
lL, 1.0 mmol) was added to a freshly pre-
AM AM323:C-18 reverse-phase column,
u
10 mm  250 mm, ace-
pared solution of LDA (10% (w/w) suspension in hexanes; 3.1 mL,
2.2 mmol) in 3 mL of THF at À78 °C. After stirring at À78 °C for
30 min, p-nitrobenzaldehyde (151 mg, 1.0 mmol), dissolved in
5 mL of THF, was added over a period of 1 min. After stirring for
3 h at À78 °C, saturated NH4Cl solution was added, and the reaction
mixture was allowed to warm to room temperature. The product was
extracted with ethyl acetate, dried over MgSO4, and evaporated in va-
cuo. The residue was purified by HPLC (YMC-Pack ODS-AM AM323:
tonitrile/aqueous TFA (0.1%, 35:65), 3.0 mL/min, 254 nm, tR: syn-
16 = 14.7 min, anti-16 = 16.5 min) to give syn-16 as a colorless oil
(26 mg, 12%) and anti-16 as a white powder (60 mg, 27%). syn-16:
1H NMR (400 MHz, CDCl3): d 8.22 (d, J = 8.8 Hz, 2H), 7.53 (d,
J = 8.8 Hz, 2H), 5.30 (d, J = 3.4 Hz, 1H), 3.66–3.72 (m, 1H), 2.96–3.06
(m, 2H), 2.19–2.24 (m, 1H), 1.89–2.01 (m, 1H). anti-16: 1H NMR
(400 MHz, CDCl3): d 8.23 (d, J = 8.8 Hz, 2H), 7.56 (d, J = 8.8 Hz, 2H),
5.00 (d, J = 8.1 Hz, 1H), 3.58–3.65 (m, 1H), 3.09–3.18 (m, 1H), 2.94–
3.04 (m, 1H), 2.07–2.17 (m, 1H), 1.88–1.98 (m, 1H); LRMS (ESI–):
[M–H]– calcd for C11H10NO4, 220.06; found, 220.
C-18 reverse-phase column,
u
10 mm  250 mm, acetonitrile/aque-
ous TFA (0.1%, 35:65), 3.0 mL/min, 254 nm, tR: 14 = 61.9 min, syn-
14a = 42.4 min, and anti-14a = 44.5 min) to give 14 as a white pow-
der (41 mg, 20%), syn-14a as a colorless oil (5.7 mg, 3%), and anti-
5.3. Preparation of antibody 27C1
14a (4.6 mg, 2%) as
a
colorless oil. Compound 14: 1H NMR
(400 MHz, CDCl3): d 8.21 (d, J = 8.8 Hz, 2H), 7.54 (d, J = 8.8 Hz, 2H),
5.26–5.28 (m, 1H), 3.65 (d, J = 2.8 Hz, 1H), 2.76–2.87 (m, 2H), 2.45
(t, J = 7.6 Hz, 2H), 1.55–1.62 (m, 2H), 1.32 (sext, J = 7.6 Hz, 2H), 0.91
(t, J = 7.6 Hz, 3H). syn-14a: 1H NMR (400 MHz, CDCl3): d 8.21 (d,
J = 8.8 Hz, 2H), 7.52 (d, J = 8.8 Hz, 2H), 5.09 (d, J = 4.4 Hz, 1H), 3.24
(br s, 1H), 2.88 (dt, J = 4.4 Hz, 8.8 Hz, 1H), 2.18 (s, 3H), 1.62–1.68
(m, 1H), 1.42–1.51 (m, 1H), 1.25–1.28 (m, 1H), 1.09–1.13 (m, 1H),
0.82 (t, J = 7.3 Hz, 3H). anti-14a: 1H NMR (400 MHz, CDCl3): d 8.22
(d, J = 8.8 Hz, 2H), 7.51 (d, J = 8.8 Hz, 2H), 4.90 (d, J = 6.6 Hz, 1H),
2.91–2.96 (m, 1H), 2.12 (s, 3H), 1.25–1.63 (m, 4H), 0.88 (t,
J = 7.2 Hz, 3H). 14, LRMS (EI): m/z calcd for C13H17NO4, 251.28; found,
251. 14a, LRMS (EI): m/z calcd for C13H17NO4, 251.28; found, 251.
Hybridoma cells for antibody 27C1 were grown to 4 L, and the
supernatants were purified by anti-mouse IgG + IgM affinity chro-
matography (CHROMATOP) (NGK, loaded in PBS and eluted with
0.2 M Gly-HCl pH 2.5) to yield purified antibody.
5.4. Antibody assays for cross-aldol reactions
All antibody-catalyzed or background reactions were performed
in 50 mM Tris–HCl pH 8.0 at 25 °C. The reaction products were
monitored by HPLC. The analytical HPLC was performed on a Hit-
achi L-2130 unit equipped with a Hitachi L-2400 UV detector using
a YMC-Pack ODS-AM AM303 column (250 Â 4.6 mm) and acetoni-
trile/water mixtures (containing 0.1% TFA) as eluent at a flow rate
of 1.0 mL/min.
5.2.5. Compounds 15, syn-15a, and anti-15a
4-Hydroxy-2-butanone (110 lL, 1.1 mmol) was added to a
freshly prepared solution of LDA (10% (w/w) suspension in hex-
anes; 3.1 mL, 2.2 mmol) in 3 mL of THF at À78 °C. After stirring
at À78 °C for 30 min, a solution of p-nitrobenzaldehyde (151 mg,
1.0 mmol) in THF (5 mL) was added over a period of 1 min. After
stirring at À78 °C for 3 h, saturated NH4Cl solution was added,
and the reaction mixture was allowed to warm to room tempera-
ture. The product was extracted with ethyl acetate, dried over
MgSO4, and evaporated in vacuo. The residue was purified by silica
gel column chromatography (EtOAc) to afford compound 15 as a
pale yellow oil (13 mg, 6%). The syn and anti mixture 15a was con-
tinuously purified by HPLC (YMC-Pack ODS-AM AM323: C-18 re-
5.4.1. Specific rates of aldol reactions
The specific rates of cross-aldol reactions were determined be-
fore 10% completion of the reactions using 1 mM acceptor sub-
strate 4, 5% donor ketones, 10 mM primary amine 2, and 10 lM
antibody. The specific rate for the cross-aldol reaction of 2-hexa-
none and p-nitrobenzaldehyde was determined with 2.5 mM 2.
All kinetic assays were measured in duplicate.
5.4.2. Regioselectivity of cross-aldol reactions
The regioselectivity of antibody-catalyzed cross-aldol reactions
was determined before 10% completion of the reactions using
verse-phase column,
u
10 mm  250 mm, acetonitrile/aqueous
500
l
M acceptor substrate 4, 5% donor ketones, 500
lM primary
TFA (0.1%, 35:65), 3.0 mL/min, 254 nm, tR: syn-15a = 18.4 min,
anti-15a = 15.9 min) to afford syn-15a as a pale yellow oil (5 mg,
2%) and anti-15a as a pale yellow oil (14 mg, 6%). Compound 15:
1H NMR (400 MHz, CDCl3): d 8.22 (d, J = 8.5 Hz, 2H), 7.55 (d,
J = 8.5 Hz, 2H), 5.29–5.35 (m, 1H), 3.89–3.94 (m, 1H), 3.85 (q,
J = 6.1 Hz, 1H), 3.41 (d, J = 4.0 Hz, 1H), 2.87–2.90 (m, 2H), 2.69–
2.75 (m, 2H). syn-15a: 1H NMR (400 MHz, CDCl3): d 8.25 (d,
J = 8.8 Hz, 2H), 7.56 (d, J = 8.8 Hz, 2H), 5.50–5.53 (m, 1H), 4.01
(dd, J = 3.7 Hz, 12.2 Hz, 1H), 3.82 (dd, J = 3.7 Hz, 12.2 Hz, 1H), 3.79
(br s, 1H), 2.90 (q, J = 3.7 Hz, 1H), 2.68 (br s, 1H), 2.29 (s, 3H).
anti-15a: 1H NMR (400 MHz, CDCl3): d 8.24 (d, J = 8.8 Hz, 2H),
7.59 (d, J = 8.8 Hz, 2H), 5.23 (d, J = 6.4 Hz, 1H), 3.93 (dd, J = 4.1 Hz,
11.2 Hz, 1H), 3.71 (dd, J = 11.2 Hz, 4.8 Hz, 1H), 3.06–3.09 (m, 1H),
amine 2, and 10
l
M antibody. The uncatalyzed reactions were per-
formed under the same conditions except for the absence of anti-
body. The reactions were incubated for 24 h. The ratios of
isomers were determined by HPLC analysis. The conversion rates
of the uncatalyzed reactions were estimated to be 73% (2-buta-
none), 19% (2-pentanone), 29% (2-hexanone), and 58% (4-hydro-
xy-2-butanone). All kinetic assays were measured in duplicate.
5.4.3. Determination of enantiomeric excess of an aldol product
12
The enantiomeric excess (ee) of 12 catalyzed by 27C1 was
determined under the following conditions: 0.5 mM p-nitrobenzal-
dehyde 4, 5% 2-butanone, 0.5 mM primary amine 2, and 20 lM

F. Ishikawa et al. / Bioorg. Med. Chem. 21 (2013) 7011–7017
7017
antibody in 50 mM Tris–HCl pH 8.0. Total DMSO concentration was
kept at 4%. The reaction mixture was incubated at 25 °C for 2 h. The
reaction mixture was purified by HPLC (YMC-Pack ODS-AM
catalyzed rates were measured at fixed concentrations of amine
2 (1.5 mM 6 [2] 6 10 mM) and varying concentrations of b-keto
acid 3 (0.5 mM 6 [3] 6 3 mM). Analogous plots were constructed
to provide kinetic constants for amine 2. All kinetic assays were
measured in duplicate.
AM323: C-18 reverse-phase column,
u
10 mm  250 mm, acetoni-
trile/aqueous TFA (0.1%, 25:75), 3.0 mL/min, 254 nm, tR:
12 = 44.0 min, 12a (syn and anti mixture) = 38.5 min to give pure
aldol product 12. The aldol product was redissolved in 2-propanol
and the ee was determined by normal phase HPLC using Daicel col-
Acknowledgement
umn (CHIRALPAK AS-H,
u
4.6 mm  250 mm, 2-propanol/hexane
This work was supported by JSPS KAKENHI Grant Number
16350091.
(50:50), 0.5 mL/min, 275 nm, tR: (R)-12 = 12.5 min and (S)-
12 = 16.8 min). Chemically synthesized standard was prepared
according to the literature procedure to provide (R)-12.35 Briefly,
to a mixture of anhydrous DMSO (4 mL) and 2-butanone (1 mL)
was added p-nitrobenzaldehyde (100 mg, 0.66 mmol) followed
References and notes
1. Nelson, S. G. Tetrahedron: Asymmetry 1998, 9, 357.
2. Machajewski, T. D.; Wong, C.-H. Angew. Chem., Int. Ed. 2000, 39, 1352.
3. Notz, W.; Tanaka, F.; Barbas, C. F., III Acc. Chem. Res. 2004, 37, 580.
4. Gijsen, H. J. M.; Qiao, L.; Fitz, W.; Wong, C.-H. Chem. Rev. 1996, 96, 443.
5. Hamilton, G. A.; Westheimer, F. H. J. Am. Chem. Soc. 1959, 81, 6332.
6. Fridovich, I.; Westheimer, F. H. J. Am. Chem. Soc. 1962, 84, 3208.
7. Laursen, R. A.; Westheimer, F. H. J. Am. Chem. Soc. 1966, 88, 3426.
8. Takagi, W.; Guthrie, J. P.; Westheimer, F. H. Biochemistry 1968, 7, 905.
9. Autor, A. P.; Fridovich, I. J. Biol. Chem. 1970, 245, 5214.
10. Frey, P. A.; Kokesh, F. C.; Westheimer, F. H. Biochemistry 1971, 14, 7266.
11. Kokesh, F. C.; Westheimer, F. H. J. Am. Chem. Soc. 1971, 93, 7270.
12. Schmidt, D. E., Jr.; Westheimer, F. H. Biochemistry 1971, 10, 1249.
13. Westheimer, F. H. Tetrahedron 1995, 51, 3.
14. Westheimer, F. H. Proceedings of the Robert A. Welch Fundation Conferences
on Chemical Research; Robert A. Welch Foundation, Huston, 1971; Vol. 15, pp
7–50.
15. Kobes, R. D.; Dekker, E. E. Biochem. Biophys. Res. Commun. 1967, 27, 607.
16. Nishihara, H.; Dekker, E. E. J. Biol. Chem. 1972, 247, 5079.
17. Vlahos, C. J.; Dekker, E. E. J. Biol. Chem. 1986, 261, 11049.
18. Wangner, J.; Lerner, R. A.; Barbas, C. F., III Science 1995, 270, 1797.
19. Björnestedt, R.; Zhong, G.; Lerner, R. A.; Barbas, C. F., III J. Am. Chem. Soc. 1996,
118, 11720.
20. Hoffmann, T.; Zhong, G.; List, B.; Shabat, D.; Anderson, J.; Gramatikova, S.;
Lerner, R. A.; Barbas, C. F., III J. Am. Chem. Soc. 1998, 120, 2768.
21. List, B.; Shabat, D.; Barbas, C. F., III; Lerner, R. A. Chem. Eur. J. 1998, 4, 881.
22. Sinha, S.; Barbas, C. F., III; Lerner, R. A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95,
14603.
23. Sinha, S.; Sun, J.; Miller, G.; Barbas, C. F., III; Lerner, R. A. Org. Lett. 1999, 1, 1623.
24. Shabat, D.; List, B.; Lerner, R. A.; Barbas, C. F., III Tetrahedron Lett. 1999, 40,
1437.
25. Ishikawa, F.; Tsumuraya, T.; Fujii, I. J. Am. Chem. Soc. 2009, 131, 456.
26. Guthrie, J. P.; Jordan, F. J. Am. Chem. Soc. 1972, 94, 9132.
27. Guthrie, J. P.; Jordan, F. J. Am. Chem. Soc. 1972, 94, 9136.
28. Hine, J.; Via, F. A. J. Am. Chem. Soc. 1972, 94, 190.
29. O’Leary, M. J.; Baughn, R. L. J. Am. Chem. Soc. 1972, 94, 626.
30. Highbarger, L. A.; Gerlt, J. A.; Kenyon, G. L. Biochemistry 1996, 35, 41.
31. Kurihara, S.; Tsumuraya, T.; Suzuki, K.; Kuroda, M.; Liu, L.; Takaoka, Y.; Fujii, I.
Chem. Eur. J. 2000, 6, 1656.
by L-Proline (20 mol %) and the resulting mixture was stirred at
room temperature for 4 h. The reaction mixture was treated with
saturated ammonium chloride solution, the layers were separated,
and the aqueous layer was extracted with EtOAc, dried over
MgSO4, and evaporated in vacuo. The residue was purified by flash
chromatography (1:4 EtOAc/hexane) to give (R)-12 as a yellow
powder (66 mg, 53%). Chiral-phase HPLC analysis revealed that
(R)-12 was formed in 59% ee. The 1H NMR, LRMS, and ee data were
identical to the literature.35
5.5. Kinetic parameters determination of the decarboxylation
reaction
The 27C1-catalyzed decarboxylation reaction was performed in
a Digital Uni Ace UA-100 water bath (Eyela) at 25 °C ( 0.1) in
50 mM Tris–HCl pH 8.0. In all experiments the total DMSO concen-
tration was kept at 10%. The reaction was followed by monitoring
the formation of (p-nitrophenyl)acetone by reversed-phase HPLC
at 280 nm. The analytical HPLC was performed on a Hitachi L-
2130 unit equipped with a Hitachi L-2400 UV detector, using a
YMC-Pack ODS-AM AM303 column eluted with acetonitrile/aque-
ous TFA (0.1%, 60:40) at a flow rate of 1.0 mL/min. The kinetic
parameters were caliculated for a rapid-equilibrium random sys-
tem.36 The antibody-catalyzed rates were measured at fixed con-
centrations of b-keto acid 3 (0.5 mM 6 [3] 6 3 mM) and varying
concentrations of amine 2 (1.5 mM 6 [2] 6 10 mM). Final antibody
concentration (5 lM) and temperature (25 °C) were maintained
throughout the assay. Kinetic parameters were determined by a
two-step analysis. First, Lineweaver–Burk (1/V vs 1/S) plots of the
raw data were constructed. These y-intercepts and slopes were
then replotted to yield the actual Vmax and Km (b-keto acid 3) val-
ues for the 27C1-catalyzed process. The kcat values were deter-
mined from the actual Vmax values. Similarly, the antibody-
32. Miyashita, H.; Karaki, Y.; Kikuchi, M.; Fujii, I. Proc. Natl. Acad. Sci. U.S.A. 1993, 90,
5337.
33. Fujii, I.; Fukuyama, S.; Iwabuchi, Y.; Tanimura, R. Nat. Biotechnol. 1998, 16, 463.
34. Miyashita, H.; Hara, T.; Tanimura, R.; Fukuyama, S.; Cagnon, C.; Kohara, A.;
Fujii, I. J. Mol. Biol. 1997, 267, 1247.
35. Sakthivel, K.; Notz, W.; Bui, T.; Barbas, C. F., III J. Am. Chem. Soc. 2001, 123, 5260.
36. Segal, I. H. Enzyme Kinetics; John Wiley & Sons, Inc.: New York, 1975.