Aldolase antibodies of remarkable scope

Article abstract of DOI:10.1021/ja973676b

This paper describes the substrate specificity, synthetic scope, and efficiency of aldolase catalytic antibodies 38C2 and 33F12. These antibodies use the enamine mechanism common to the natural Class I aldolase enzymes. Substrates for these catalysts, 23 donors and 16 acceptors, have been identified. The aldol acceptor specificity is expected to be much broader than that defined here since all aldehydes tested, with the exception polyhydroxylated aldehydes, were substrates for the antibodies. 38C2 and 33F12 have been shown to catalyze intermolecular ketone-ketone, ketone- aldehyde, aldehyde-ketone, and aldehyde-aldehyde aldol addition reactions and in some cases to catalyze their subsequent dehydration to yield aldol condensation products. Substrates for intramolecular aldol reactions have also been defined. With acetone as the aldol donor substrate a new stereogenic center is formed by attack on the si-face of the aidehyde with ee's in most cases exceeding 95%. With hydroxyacetone as the donor substrate, attack occurs on the re-face, generating an (α,β-dihydroxy ketone with two stereogenic centers of the α-syn configuration in 70 to >98% ee. With fluoroacetone donor reactions, the major product is a syn α-fluoro-β- hydroxy ketone with 95% ee. Studies of retroaldol reactions demonstrate that the antibodies provide up to 10⁸-fold enhanced efficiency relative to simple amine-catalyzed reactions.

Full text of DOI:10.1021/ja973676b

2768
J. Am. Chem. Soc. 1998, 120, 2768-2779
Aldolase Antibodies of Remarkable Scope
Torsten Hoffmann, Guofu Zhong, Benjamin List, Doron Shabat, James Anderson,
Svetlana Gramatikova, Richard A. Lerner,* and Carlos F. Barbas III*
Contribution from The Skaggs Institute for Chemical Biology and the Department of Molecular Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
ReceiVed October 23, 1997
Abstract: This paper describes the substrate specificity, synthetic scope, and efficiency of aldolase catalytic
antibodies 38C2 and 33F12. These antibodies use the enamine mechanism common to the natural Class I
aldolase enzymes. Substrates for these catalysts, 23 donors and 16 acceptors, have been identified. The aldol
acceptor specificity is expected to be much broader than that defined here since all aldehydes tested, with the
exception polyhydroxylated aldehydes, were substrates for the antibodies. 38C2 and 33F12 have been shown
to catalyze intermolecular ketone-ketone, ketone-aldehyde, aldehyde-ketone, and aldehyde-aldehyde aldol
addition reactions and in some cases to catalyze their subsequent dehydration to yield aldol condensation
products. Substrates for intramolecular aldol reactions have also been defined. With acetone as the aldol
donor substrate a new stereogenic center is formed by attack on the si-face of the aldehyde with ee’s in most
cases exceeding 95%. With hydroxyacetone as the donor substrate, attack occurs on the re-face, generating
an R,â-dihydroxy ketone with two stereogenic centers of the R-syn configuration in 70 to >98% ee. With
fluoroacetone donor reactions, the major product is a syn R-fluoro-â-hydroxy ketone with 95% ee. Studies of
retroaldol reactions demonstrate that the antibodies provide up to 108-fold enhanced efficiency relative to
simple amine-catalyzed reactions.
Introduction
synthesis.3 Since no asymmetric catalyst exhibits the scope of
reactivity required to meet every synthetic challenge, there is a
need for methodologies that allow for the development of
asymmetric catalysts. This is true of both, transition metal based
as well as enzyme-based catalysts. For example, while the
Carreira Ti(IV) complex is limited in scope to the use of the
enolate equivalent 2-methoxypropene, fructose 1,6-diphosphate
aldolase is limited to the use of dihydroxyacetone phosphate as
the aldol donor substrate.3a
To address the problem of the de noVo generation of protein
catalysts of the aldol reaction, we recently described the
development of two aldolase catalytic antibodies 38C2 and
33F12.4
These antibodies were raised against the â-diketone hapten
1 which served as a chemical trap to imprint the lysine-
dependent class I aldolase mechanism in the active site of the
antibody. The suggested mechanism for the selection process
of antibodies 38C2 and 33F12 during immunization is shown
in Scheme 1. The ꢀ-amino group of the lysine residue reacts
with a carbonyl function of the â-diketone moiety of 1 to form
a â-keto hemiaminal followed by dehydration to give a â-keto
The aldol reaction is arguably one of the most important C-C
bond-forming reactions employed in synthetic transformations.
Traditionally, the aldol reaction has been a proving ground for
the development of asymmetric synthetic strategies. In the
1980s the aldol reaction experienced a renaissance with the
development of numerous strategies to effect highly stereose-
lective aldols.1 Generally, this has been most successfully
achieved through the use of stoichiometric quantities of chiral
auxiliaries. In recent years the design of stereoselective catalysts
of the aldol reaction has become a topic of interest.2 Most
notable of these approaches is the Carreira aldol reaction where
a chiral Ti(IV) complex (2-10 mol %) catalyzes the enantio-
selective addition of 2-methoxypropene to aldehydes with 66-
98% ee.2b As a challenge to traditional organic methodology,
the application of natural aldolase enzymes as synthetic catalysts
has yielded numerous efficient syntheses of stereochemically
complex molecules, particularly in the area of carbohydrate
(1) For reviews of the aldol reaction, see: (a) Heathcock, C. H. In
Asymmetric Synthesis; Morrison, J. D., Ed.; Academic Press: New York,
1984; Vol. 3. (b) Evans, D. A.; Nelson, J. V.; Taber, T. R. Top. Stereochem.
1982, 12, 1. (c) Masamune, S.; Choy, W.; Peterson, J. S.; Sita, L. R. Angew.
Chem., Int. Ed. Engl. 1985, 24, 1. (d) Heathcock, C. H. Aldrichim. Acta
1990, 23, 99. Heathcock, C. H. Science 1981, 214, 395. Evans, D. A. Ibid.
1988, 240, 420. Masamune, S.; Choy, W.; Peterson, J.; Sita, L. R. Angew.
Chem., Int. Ed. Engl. 1985, 24, 1. Evans, D. A.; Nelson, J. V.; Taber, T. R.
Top. Stereochem. 1982, 13, 1. Heathcock, C. H. et al. In ComprehensiVe
Organic Synthesis; Trost, B. M., Ed.; Pergamon: Oxford, 1991; Vol. 2, pp
133-319. Peterson, I. Pure Appl. Chem. 1992, 64, 1821.
(2) (a) Yanagisawa, A.; Matsumoto, Y.; Nakashima, H.; Asakawa, K.;
Yamamoto, H. J. Am. Chem. Soc. 1997, 119, 9319 and references therein.
Bach, T. Angew. Chem., Int. Ed. Engl. 1994, 33, 417 and references therein.
Yamada, Y. M. A.; Yoshikawa, N.; Sasai, H.; Shibasaki, M. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1871. (b) Carreira, E. M.; Lee, W.; Singer, R. A.
J. Am. Chem. Soc. 1995, 117, 3649.
(3) (a) Gijsen, H. J. M.; Qiao, L.; Fitz, W.; Wong, C.-H. Chem. ReV.
1996, 96, 443. (b) Wong, C.-H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto,
T. Angew. Chem., Int. Ed. Engl. 1995, 34, 412-432. (b) Henderson, I.;
Sharpless, K. B.; Wong, C.-H. J. Am. Chem. Soc. 1994, 116, 558. (c) Wong,
C.-H.; Whitesides, G. M. Enzymes in Synthetic Organic Chemistry;
Pergamon: Oxford, 1994. Bednarski, M. D. In ComprehensiVe Organic
Synthesis; Trost, B. M., Ed.; Pergamon: Oxford, 1991; Vol. 2, p 455. Gijsen,
H. J. M., Wong, C.-H. Ibid. 1995, 117, 2947. Wong, C.-H. et al. Ibid. 1995,
117, 3333. Chen, L.; Dumas, D. P.; Wong, C.-H. Ibid. 1992, 114, 741.
(4) (a) Wagner, J.; Lerner, R. A.; Barbas III, C. F. Science 1995, 270,
1797. (b) Zhong, G.; Hoffmann, T.; Lerner, R. A.; Danishefsky, S.; Barbas
III, C. F. J. Am. Chem. Soc. 1997, 119, 8131. (c) Barbas III, C. F.; Heine,
A.; Zhong, G.; Hoffmann, T.; Gramatikova, S.; Bjo¨rnestedt, R.; List, B.;
Anderson, J.; Stura, E. A.; Wilson, E. A.; Lerner, R. A. Science 1997, 278,
2085.
S0002-7863(97)03676-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/07/1998

Aldolase Antibodies of Remarkable Scope
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2769
Scheme 1. Mechanism of Trapping the Essential ꢀ-Amino Group of a Lysine Residue in the Antibody Binding Site Using the
â-Diketone Hapten 1
imine that finally tautomerizes into a stable enaminone 2.
Consequently, the hapten is now covalently bound in the binding
pocket. The mechanistic similarity between this stoichiometric
reaction and the accepted enamine mechanism of class I aldolase
enzymes has been discussed in detail elsewhere.4 The formation
of the enaminone has been monitored by UV spectroscopy (with
hapten 1: λmax ) 318 nm, ꢀ ∼ 15000) and is complete within
seconds to a few minutes, depending on whether antibodies were
incubated with hapten 1, or other diketones such as 2,4-
pentanedione or 3-methyl 2,4-pentanedione. Antibodies 38C2
and 33F12 have been previously shown to catalyze aldol
reactions of some aliphatic ketone donors with two different
aldehyde acceptors having a 4-acetanilide substituent in the
â-position as well as intramolecular aldol reactions that allowed
for our recent antibody-catalyzed synthesis of the Wieland-
Miescher ketone.4b Moreover, both antibodies were found to
catalyze the decarboxylation reactions of aromatic â-keto acids
by the formation of a Schiff base between the ꢀ-amino group
of the lysine residue and the keto group of the substrate.5
With antibodies 38C2 and 33F12 we have now addressed
four issues: (i) scope and limitations of substrates for inter-
molecular crossed and self-aldols as well as intramolecular
aldols, (ii) their stereoselectivity, (iii) kinetic parameters for these
reactions to understand the nature of the binding pocket, and
(iv) additional mechanistic studies to gain further insight into
these catalysts.
Figure 1. Aldehydes 3-8 as acceptors in the antibody-catalyzed cross-
aldol reactions.
Table 1, a wide variety of different ketones undergo the aldol
reaction with aldehyde 4 as proven by comparison of the
retention times of the products using high performance liquid
chromatography (HPLC) with independently chemically syn-
thesized standards: aliphatic open chain (acetone, butanone,
2-pentanone, and 3-pentanone), aliphatic cyclic (cyclopentanone
and cyclohexanone), functionalized open chain (hydroxyacetone
and fluoroacetone), and functionalized cyclic (2-hydroxy cy-
clohexanone) ketones. A large variety of other ketones were
also studied with aldehyde 4 with monitoring of the consumption
of 4 and appearance of a new peak tentatively assigned as the
â-hydroxy ketone product. Substrate ketones from all studies
are summarized in Table 2. For structures of ketones that were
tested and determined not to be substrates for these catalysts,
see Supporting Information. The antibody-catalyzed aldol
reactions were inhibited by addition of equimolar amounts of
the hapten 1 or 2,4-pentanedione. Further control experiments
were carried out using lysine or bovine serum albumin instead
of the antibodies. No catalysis of the aldol reactions studied
here was observed in these cases. A number of ketones, e.g.
octanone, demonstrated limited solubility in aqueous medium
perhaps precluding their availability to the catalysts. Further
study of mixed solvent systems may allow conditions to be
defined where substrates such as these are accepted by the
catalyst. Of the ketones studied, reactions involving hydroxy-
acetone as the donor were the most efficient. The astonishingly
high promiscuity of ab 38C2 for these physicochemically very
different ketones is in remarkable contrast to the natural
aldolases which tolerate minor, if any, change in the structure
of the donor.3
Results and Discussion
Cross-Aldol Reactions. To define the scope of antibody
catalysts 38C2 and 33F12 for the cross-aldol reaction we tested
a wide variety of commercially available ketones as donors and
a set of six different aldehydes (4-acetamidobenzaldehyde (3),
3-(4′-acetamidophenyl)propanal (4), 3-(4′-acetamidophenyl)-2-
methylpropanal (5), 4-(4′-acetamidophenyl)butanal (6), 4-ni-
trobenzaldehyde (7), and 4-nitrocinnamaldehyde (8); Figure 1)
as acceptors. To screen for donor activity, aldehyde 4 was
chosen as the standard acceptor aldehyde since it bears a
4-acetamidophenyl group at C3. This portion of the molecule
is closely related to the hapten structure 1 and may be
specifically recognized by the antibodies. For ease of com-
parison we determined specific rates for these reactions under
the following defined conditions: 1 M of donor, 500 µM of
aldehyde 4, and 0.4 mol % of antibody. As can be seen from
Next, we focused on defining the specificity for acceptor
aldehydes. We selected acetone, cyclopentanone, and hydroxy-
acetone as representative donors from the variety of ketones
(5) Bjo¨rnestedt, R.; Zhong, G.; Lerner, R. A.; Barbas III, C. F. J. Am.
Chem. Soc. 1996, 118, 11720.

2770 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Hoffmann et al.
Table 1. Donor Promiscuity: Specific Rates [µmol product × d-1
× µmol-1 ab] for Antibody-Catalyzed Cross-Aldol Reactions of a
Variety of Ketones with Aldehyde 4 under the Following Defined
Conditions: 1 M Donor, 500 µM 4, and 0.4 Mol % Antibody (2
µM). R ) 4-Acetamidophenyl
we chose some of the described reactions and studied initial
rates, varying the substrate concentrations of either the donor
or the acceptor while keeping the concentrations of the second
reactant constant. All reactions with aldehydes 4-8 followed
typical Michaelis-Menten kinetics by treating the data accord-
ing to pseudo-first-order kinetics. Examples for antibodies 38C2
and 33F12 are provided in Supporting Information. Generally,
no substrate or product inhibition was observed. Using 4-ac-
etamidobenzaldehyde 3 at higher concentrations (2 mM and
higher) substrate inhibition was found, presumably because of
a reversible Schiff base formation between the active site lysine
and the aldehyde. Michaelis-Menten kinetic data are sum-
marized in the Supporting Information section. Typical values
for the Michaelis constants KM of the donors range from 1 mM
to 1 M, reflecting the ability of the antibody to accept a wide
variety of different ketones. Values for KM of the acceptors
3-8 range from 10 to 500 µM. Obviously, the aromatic portion
of these molecules is responsible for an increased recognition
by the active site. Characteristic values for kcat range from 10-3
to 1 min-1 and show a ratio of kcat/kuncat of 105 to 107 (see below
for additional kinetic data). All data are reported per antibody
active site where the antibody molecule has two active sites.
Self-Aldol Reactions. During the course of our investiga-
tions, we also addressed the question whether antibodies 38C2
and 33F12 are capable of catalyzing self-aldol reactions. Self-
aldolization was observed with propionaldehyde, acetone, and
cyclopentanone but not for compounds 4-6 (Table 4). Again,
for ease of comparison, specific rates were determined using
the following defined conditions: 0.1 M of ketone or aldehyde
and 0.005 mol % of antibody. With these substrates the
antibodies catalyzed the aldol condensation reaction which
consists of two consecutive steps: the aldol addition and the
subsequent elimination of water. Using propionaldehyde as
substrate, the aldol addition product, 3-hydroxy-2-methylpen-
tanal, was not detected as an intermediate but exclusively trans-
2-methyl-2-pentenal 33 was detected, suggesting that the
elimination of water is also catalyzed (vide infra). Compound
33 was obtained as single product and is not a substrate for a
consecutive aldol addition reaction with a third propionaldehyde
molecule. Interestingly, compound 33 acts as an acceptor if
acetone is used as donor (see paragraph Miscellaneous Aldol
Reactions). Acetone itself and cyclopentanone undergo the self-
aldol condensation to give mesityl oxide 34 and compound 35,
respectively, if no aldehyde acceptor for a cross-aldol reaction
is present which binds preferentially to the active site. These
ketone self-aldol reactions are substantially slower, typically
more than 500-fold slower, than the cross-aldol reactions
involving these donors described above. In the cross-aldol
reactions described above with acetone and cyclopentanone as
donors, the acetone-acetone and cyclopentanone-cyclopen-
tanone self-condensation products were not observed. Thus,
self-aldolization activity in the cross-aldol reaction does not
compromise isolation of the cross-aldol product. To elucidate
the general mechanism of the self-aldol condensation reactions
described above, we chose cyclopentanone as substrate and
monitored the appearance of the aldol addition product 36 and
its elimination product 35 by HPLC and gas chromatography
(GC). In addition, ab 38C2 was incubated with independently
synthesized intermediate 36, and the appearance of 35 was
monitored by HPLC. Indeed, both reaction steps are catalyzed
with the aldol addition occurring faster than the elimination step.
Kinetic study of the overall transformation of cyclopentanone
a Specific rate was determined with 200 mM donor. b Not deter-
mined.
listed in Table 1 and tested these in the aldol reaction with
aldehydes 3 and 5-8 (Figure 1). Again, identical conditions
(as described above) were used for the determination of the
corresponding specific rates (Table 3). All substrate combina-
tions gave the expected products. Simple aliphatic aldehyde
acceptors are described below (see Miscellaneous Aldol Reac-
tions). A variety of polyhydroxylated acceptor aldehydes,
glyceraldehyde, glucose, and ribose were also studied. These
substrates were not accepted by these antibodies presumably
due to the hydrophobic nature of the active site. These studies,
together with those described below, suggest an acceptor
specificity that is very broad, albeit limited to hydrophobic
aldehydes.
to 35 revealed KM ) 845 µM, kcat ) 2 × 10-4 min-1
.
The
To gain more detailed information about the catalytic system
elimination reaction from 36 to 35 also followed Michaelis-

Aldolase Antibodies of Remarkable Scope
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2771
Table 2. Donor Substrates of Aldolase Antibodies 38C2 and 33F12
Menten kinetics (KM ) 750 µM, kcat ) 9 × 10-4 min-1 and
kcat/kuncat ) 2240). From these experiments we suggest a
mechanism for the self-aldol condensation reaction of cyclo-
pentanone as shown in Scheme 2 (pathway 1). The â-hydroxy
iminium cation (formed in the active site after attack of the
enamine of cyclopentanone at the carbonyl C-atom of a second
molecule cyclopentanone) tautomerizes after loss of a proton
into a â-hydroxy enamine which then loses water. This process
is assisted by the electron donation from the enamine nitrogen
atom.6 The R,â-unsaturated iminium cation is finally hydro-
lyzed, and the aldol condensation product 35 is released while
the lysine residue of ab 38C2 re-enters the catalytic cycle. An
alternate mechanism in which the elimination step occurs in
the background via free 36 is also shown in Scheme 2 (pathway
2) but seems less likely, since the elimination was shown to be
catalyzed by ab 38C2 and inhibited by acetylacetone. No
background reaction for the overall conversion of cyclopen-
tanone to 35 in the absence of antibody could be observed
following extended incubation for 2 months.
pheromone of the Douglas Fir beetle8swas monitored by HPLC.
The reaction followed Michaelis-Menten kinetics (Table 5) and
showed an excellent rate acceleration. The antibody was also
incubated with the corresponding intermediate, 3-hydroxy-3-
methylcyclohexanone. As seen in the self-aldol condensation
of cyclopentanone, the elimination step leading to 40 was also
catalyzed by ab 38C2. In reported experiments,4b we tested
the triketone 43 and the enantiomerically pure diketones (S)-
45 and (R)-45 as substrates (Scheme 4). In all cases product
formation (Wieland-Miescher ketone (S)-44 and the ketones
(S)-46 and (R)-46), respectively) was catalyzed by 38C2 and
proceeded with a rate which is comparable to the one found
using the simple diketone 39 as substrate. All reactions followed
Michaelis-Menten kinetics, the parameters of which are sum-
marized in Table 5. We also examined 2-methyl-2-(3′-oxo-
pentyl)-1,3-cyclohexanedione, 2-methyl-2-(3′-oxobutyl)-1,3-
cyclopentanedione, and 2-methyl-2-(4′-oxopentyl)-1,3-cyclo-
hexanedione as potential substrates for these antibodies; how-
ever, no reactions with these substrates were observed.
Miscellaneous Aldol Reactions. As described above, the
self-aldol condensation of two molecules propionaldehyde
terminates after the condensation step. The product, trans-2-
methyl-2-pentenal (33) is not a substrate for a subsequent aldol
addition (or condensation) of a third molecule propionaldehyde.
In contrast, aldehyde 33 is a substrate if acetone is present as
donor (Scheme 5). Surprisingly, the elimination of water from
product 47 is not catalyzed by the antibody, although a
thermodynamically favored R,â,γ,δ-unsaturated ketone could
result. The binding pocket also accepts chain elongation in the
acceptor structure as rationalized in 2,4-hexadienal (48). As
seen with product 47, the antibody does not catalyze the
elimination of water from the aldol addition product 49 either.
We also investigated whether propionaldehyde might be a donor
for other aldehydes, which would bind preferentially to the
binding pocket and would give a â-hydroxy aldehyde as product
or, after elimination of water, the R,â-unsaturated aldehyde. To
Intramolecular Aldol Condensations. To study intramo-
lecular aldol reactions, 38C2 was incubated with three different
aliphatic diketones: 2,4-hexanedione (37), 2,5-heptanedione
(39), and 2,6-octanedione (41) (Scheme 3). No catalysis was
observed in the reaction pathway from 37 to 3-methylcyclopent-
2-enone 38, presumably because of the Baldwin7 disfavored 5-
(enolendo-trig) process involved in the attack of the enamine
at C2 in substrate 37. Although the corresponding ring closure
(followed by water elimination) of 41 to give 3-methylcyclohept-
2-enone 42 is Baldwin favored (a 7(enolendo-trig) process) also
in this case no product formation was observed. In contrast,
the Baldwin favored ring closure reaction of 2,5-heptanedione
(39) (a 6(enolendo-trig) process) followed by elimination of
water and giving 3-methylcyclohex-2-enone (40) was catalyzed
by ab 38C2. Appearance of the product 40swhich is the sex
(6) Hupe, D. J.; Kendall, M. C. R.; Spencer, T. A. J. Am. Chem. Soc.
1973, 95, 2271.
(7) Baldwin, J. E.; Lusch, M. J. Tetrahedron 1982, 38, 2939.
(8) Dickens, J. C.; Gutman, A.; Payne, T. L.; Ryker, L. C.; Rudinsky, J.
A. J. Chem. Ecol. 1983, 9, 1383. Libbey, L. M.; Oehlschlager, A. C.; Ryker,
L. C. J. Chem. Ecol. 1983, 9, 1533.

2772 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Hoffmann et al.
Table 3. Acceptor Promiscuity: Specific Rates [µmol product × d-1 × µmol-1 ab] for Antibody-Catalyzed Cross-Aldol Reactions of
Acetone, Cyclopentanone, and Hydroxyacetone with Aldehydes 3 and 5-8 under the Following Defined Conditions: 1 M Donor, 500 µM
Aldehyde, and 0.4 Mol % Antibody (2 µM). R ) 4-Acetamidophenyl, R′ ) 4-Nitrophenyl
suppress the self-aldol addition, a lower concentration of the
donor and a higher concentration of the acceptor aldehyde was
used. An interesting situation was observed with acetaldehyde.
It was found to act solely as an acceptor to give 2-methyl-2-
butenal (50) with propionaldehyde as donor. 2-Hexenal, the
cross-aldol condensation product of the reversed reactivity
(acetaldehyde as donor and propionaldehyde as acceptor), was
not detected. As shown in Table 6, the antibodies catalyze
cross-aldol reactions where simple aliphatic aldehydes act as
acceptor substrates. The cross-aldol reaction between cyclo-
pentanone and pentanal yielding 53 proved to be very efficient
and kinetic studies revealed a kcat of 1.1 min-1 and a KM for
pentanal of 3.9 mM. Pentanal was a very efficient acceptor
substrate when paired with hydroxyacetone as the donor.
Retroaldol reactions involving the pentanal derived products 55
and 56 were also catalyzed.
Stereoselectivity and Absolute Configurations of the Aldol
Products. Both aldolase antibodies 38C2 and 33F12 catalyze
highly stereoselective aldol reactions. As a general rule, acetone
adds to aldehydes with si-facial selectivity whereas with
hydroxyacetone a reversal of enantioface selectivity results in
addition to the re-face. The products were formed in ee’s up
to >99% (Table 7). The ee’s were determined by HPLC and
using chiral stationary phases. To assign absolute configura-
tions, the products were synthesized by either of two ways. In
case of the acetone products, we used the method developed
by Paterson et al.,9 utilizing the diisopinocamphenyl enol
borinate, prepared in situ from (-)-Ipc2BOTf and acetone. The
(9) Paterson, I.; Goodman, J. M.; Lister, M. A.; Schumann, C.; McClure,
C. K.; Norcross, R. D. Tetrahedron 1990, 46, 4663.

Aldolase Antibodies of Remarkable Scope
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2773
Table 4. Specific Rates [µmol product × d-1 × µmol-1 ab] for
Antibody-Catalyzed Self-Aldol Reactions of Propionaldehyde,
Acetone, and Cyclopentanone and for the Elimination of Water
from 36 to 35 under the Following Defined Conditions: 0.1 M of
Ketone or Aldehyde and 0.005 Mol % Antibody (5 µM). These
Reactions Take Place Predominantly in the Absence of an Aldehyde
Acceptor for Cross-Aldol Reactions
Scheme 3. Diketones 37, 39, and 41 Tested as Substrates in
the Antibody-Catalyzed Intramolecular Aldol Condensation.
Only the 6(enolendo-trig)-exo-trig Process, Leading to 40
from 39, Is Catalyzed
Table 5. Kinetic Parameters (Michaelis-Menten Kinetics) for
Intramolecular Aldol Condensations of Substrates 39, 43, and 45
substrate
KM [mM]
kcat [min-1
]
kuncat [min-1
]
kcat/kuncat
39
43
(S)-45
(R)-45
2.05
2.34
12.4
2.45
0.082
0.086
0.186
0.126
6.7 × 10-9
2.4 × 10-8
nda
1.2 × 107
3.6 × 106
nda
nda
nda
a Not determined.
Scheme 4. Antibody-Catalyzed Synthesis of Steroid Partial
Structures (S)-44 (Wieland-Miescher ketone), (S)-46, and
(R)-46
Scheme 2. Proposed Mechanism (pathway 1) for the
Antibody-Catalyzed Self-Aldol Condensation of
Cyclopentanone. Both Steps, the Addition and the
Elimination, Are Catalyzed by the Antibodies. The Less
Favored Mechanism (pathway 2) in Which Water
Elimination Occurs via Free 36 Is Also Shown
Scheme 5. Aldehydes 33, 48, and Acetaldehyde as
Acceptors in the Antibody-Catalyzed Cross-Aldol Reactions
hydroxyacetone products were prepared from the corresponding
R,â-unsaturated ketones via Sharpless asymmetric dihydroxy-
lation.10
The highest enantioselectivities were observed for aldol
reactions where conjugated aldehydes, allylic or benzylic, served
as acceptors with acetone as donor. In these cases the observed
enantioselectivities were always over 98% ee with either
antibody catalyst. Lower enantioselectivities were obtained with
aldol acceptors containing an sp3 center in the R-position, though
enantioselectivity may be increased by addition of steric bulk
at this carbon center as suggested with compound 21. With
acceptors of this class, the two catalysts exhibited differential
degrees of stereoselectivity. The lowest degree of enantiose-
lectivity is obtained with 4-(4′-acetamidophenyl)butyraldehyde
(10) (a) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem.
ReV. 1994, 94, 2483. Walsh, P. J., Sharpless, K. B. Synlett 1993, 605.

2774 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Hoffmann et al.
Table 6. Specific Rates [µmol product × d-1 × µmol-1 ab] for Antibody-Catalyzed Cross-Aldol Reactions of Cyclopentanone and
Hydroxyketones with Aliphatic Aldehydes under the Following Defined Conditions: 1 M Donor, 500 µM Aldehyde, and 0.4 Mol % Antibody
(2 µM)
a Not determined.
as an acceptor and acetone as a donor to yield compound 24.
The ee in this case is 20% and 3%, with catalysts 38C2 and
33F12, respectively. Donor, acceptor, and catalyst all exert an
effect on the enantioselectivity of the reaction since use of
hydroxyacetone as an aldol donor instead of acetone in the
reaction with 4-(4′-acetamidophenyl)butyraldehyde provides
product 58 in 77% ee. The regioselectivity of this reaction with
respect to bond formation at either R position on hydroxyacetone
is perfect. Only product formed by reaction at the R-hydroxy
bearing side of hydroxyacetone was detected. Study of the
reaction of fluoroacetone with aldehyde 4 catalyzed by 38C2
also demonstrated regioselectivity of bond formation toward the
R-substituted carbon of fluoroacetone; however, 7% of the other
regioisomer 60 was also isolated (Scheme 6). The syn-16,
constituted 72% of the isolated product and was formed in 95%
ee, while the anti-16, 21% of the isolated product, was formed
with 34% ee. The absolute stereochemistry of these products
remains to be assigned. The regioselectivity of the aldol
addition of fluoroacetone catalyzed by these antibodies is
opposite that observed with the natural aldolase, deoxyribose-
5-phosphate aldolase, where only addition to the unsubstituted
side of fluoroacetone was observed.11 There are no known
catalysts or general methodologies for the asymmetric synthesis
of R-fluoro-â-hydroxy ketones such as syn-16.12 Stereoselec-
tivity involving reactions of R-aliphatic substituted ketones, for
example, cyclopentanone, were not studied in detail in this case
due to the stereochemical lability at the R-position of products
formed with these donors in buffered aqueous solvent.13 The
anti/syn selectivity of reactions with cyclopentanone as the donor
in reaction with aliphatic ketones are given in Supporting
Information as compared to the product distribution obtained
with kinetically controlled synthesis (LDA, -78 °C). With
n-pentanal as the acceptor, 53 is produced via antibody catalysis
with an anti/syn ratio nearly identical to that obtained under
kinetically controlled synthesis. A reversal of selectivity is noted
with heptanal as the acceptor where the syn-product predomi-
nates.
Retroaldolization, Catalytic Efficiency, and Proposed
Mechanism. Previous biochemical studies of both catalysts and
structural studies of 33F12 are all consistent with an enamine
mechanism shared with the natural class I aldolase enzymes.4
Central to this mechanism is a chemically unique lysine residue
bearing an ꢀ-amino group with 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. The catalytic efficiency of antibody 38C2 as an
aldolase is most readily compared with simple amine catalysis
by study of the retroaldol reaction since the second-order rate
constant of the amine-catalyzed reaction can be directly related
to kcat/Km of the antibody-catalyzed reaction. Kinetic studies
of amine-catalyzed aldol addition reactions have been reported
and are approximately 102-fold slower than the amine-catalyzed
retroaldol reaction.14 We studied the retroaldol reaction of
substrate 59 under antibody and amine catalysis where amine
catalysis was studied in both, buffered aqueous solvent and
organic solvent (Scheme 7). Antibody 38C2 catalyzed the
retroaldolization of 59 following Michaelis-Menten kinetics-
(kcat ) 1.4 min-1, Km ) 270 µM). The background rate of this
retroaldol reaction (100 mM MOPS buffer, pH 7) was deter-
mined to be 8.3 × 10-8 min-1. The relative rate enhancement
over background provided by the antibody for this reaction (kcat/
(11) Barbas III, C. F., Wang, Y.-F., Wong, C.-H. J. Am. Chem. Soc.
1990, 112, 2013.
(12) Welch, J. T.; Seper, K.; Eswarakrishnan, S.; Samartino, J. J. Org.
Chem. 1984, 49, 4720. House, H. O.; Fischer, W. F.; Gall, M.; McLaughlin,
T. E.; Peet, N. P. J. Org. Chem. 1971, 36, 3429-3437. Kowalski, C.; Creary,
X.; Rollin, A. J.; Burke, M. C. J. Org. Chem. 1978, 43, 2601.
(13) Iriye, R.; Takai, K.; Noguchi, M. Bioorg. Med. Chem. Lett. 1997,
7, 199. Konieczny, M. T.; Toma, P. H.; Cushman, M. J. Org. Chem. 1993,
58, 4619.
k
uncat) is 1.7 × 107, and the specificity constant (kcat/Km) is 5.2
× 103 min-1 M-1. To date, the most efficient substrate for the
38C2-catalyzed retroaldol reaction is 6-(4′-(dimethylamino)-
(14) Reymond, J.-L.; Chen, Y. J. Org. Chem. 1995, 60, 6979.

Aldolase Antibodies of Remarkable Scope
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2775
butylamine and aminoacetonitrile in water are 10.61 and 5.34.17
The relative efficiency of 38C2 over amine catalysis (kcat/Km)/
kNH and the effective molarity (kcat/kNH ) of the active site lysine
Table 7. Stereochemical Purity of Some Products As Determined
by Chiral Phase HPLC and GC
2
2
of 38C2 are given in Table 8. Antibody 38C2 provides a 106-
108-fold enhancement of the efficiency of the retroaldol reaction
of 59 as compared to the nonenzymic amine-catalyzed reactions
in either aqueous or organic media. The relative efficiency of
38C2 over simple amine-catalyzed retroaldolization of 59
compares favorably with the efficiency of the enzyme acetoac-
etate decarboxylase that has been compared with aminoaceto-
nitrile-catalyzed decarboxylation of the same substrate, ace-
toacetate, where (kcat/Km)/kNH is 9.5 × 106.18 Acetoacetate
2
decarboxylase is the most studied of enzymes whose mechanism
centers around an activated ꢀ-amino group of lysine. The
effectiveness of the active site amine of 38C2 is also indicated
by effective molarities between 560 and 35000 M depending
on the amine and solvent system studied. As shown in Table
8, n-butylamine catalysis in n-octanol is increased 63-fold
compared to catalysis in aqueous solution since the amine is in
its reactive unprotonated state in n-octanol. Aminoacetonitrile
exhibits similar efficiency in both solvents due to its low pKa.
The activation energy for this reaction should be lower in
nonpolar solvents such as the active site of the antibody and
n-octanol since the reaction involves charge dispersal in the
transition state. A polar medium would be expected to stabilize
the cationic iminium intermediates to a greater extent than the
activated complex.19
The crystal structure of unliganded 33F12 shows LysH93
within hydrogen bonding distance to a water molecule that is
also within hydrogen bonding distance to the hydroxyl group
of TyrL41. With the caveat that antibodies may undergo large
conformational changes on binding, a mechanism analogous to
that proposed for class I aldolases is shown in Figure 2 and
appears most likely. In the proposed mechanism, TyrL41 may
act as the base. Acceptor activation appears to be essential since
we have failed to trap the enamine with addition of alkylating
substrates such as 4-nitrophenethyl bromide or R,â-unsaturated
substrates such as methyl vinyl ketones according to Stork
enamine20 or Michael reactions.21 Acceptor activation and the
microenvironment in which it occurs, however, must be
balanced so as not to facilitate formation of the unreactive gem-
diol form of the aldehyde. Enamine formation and or the C-C
bond-forming/breaking step are rate-limiting with these catalysts.
Study of the 38C2-catalyzed retroaldolization of 59 in 18O-
labeled water by electrospray mass spectrometry showed rapid
antibody-catalyzed incorporation of 18O into substrate 59 (Figure
3). Within 6 min, 1 mol % 38C2 catalyzed 18O incorporation
into more than 50% of the substrate 59 molecules where no
incorporation in the control reaction not containing antibody
*Reference 4a.
phenyl)-4-hydroxy-5-hexen-2-one. The specificity constant for
this substrate is 2.0 × 105.4c The specificity constant of antibody
38C2 for this reaction exceeds that previously reported for an
amine cofactor-dependent antibody aldolase by a factor greater
than 106.15
Earlier studies have suggested that the pKa of the ꢀ-amino
group of the active site lysine, Lys H93, that is central to the
chemistry of these catalysts, is highly perturbed by a hydro-
phobic microenvironment that disfavors protonation and charge
development in the unliganded antibodies. The pKa’s of the
active-site lysines of 38C2 and 33F12 have been estimated to
be 6.0 and 5.5, respectively.4c The pKa of the ꢀ-amino group
of lysine free in solution is 10.5.16 These studies also suggested
an active-site hydrophobicity that approximates that of n-octanol.4c
Therefore the second-order rate constants for n-butylamine- and
aminoacetonitrile-catalyzed retroaldolization of 59 were deter-
mined in aqueous buffer and n-octanol. N-Butylamine was
studied because of its structural similarity to the side-chain of
lysine that is key in the antibody-catalyzed reaction, and
aminoacetonitrile was studied since its pKa approximates that
of the active-site lysine of the antibody. The pKa’s of n-
(17) Hupe, D. J.; Kendall, M. C. R.; Spencer, T. A. J. Am. Chem. Soc.
1972, 94, 1254.
(18) Westheimer, F. H. Proceeding of the Robert A. Welch Found. Conf.
Chem. Res. 1971, 15, 7. Westheimer, F. H. Tetrahedron 1995, 51, 3.
Highbarger, L. A.; Gerlt, J. A.; Kenyon, G. L. Biochemistry 1996, 35, 41.
(19) Reichardt, C. In SolVents and SolVents Effects in Organic Chemistry,
2nd ed.; Ebel, H. F., Wentrup-Byne, E., Eds.; New York, 1990; pp 1-511.
Heathcock, C. H. Stereoselective Aldol Condensations. In ComprehensiVe
Carbanion Chemistry; Buncel, E., Durst, T., Eds.; Elsevier: Amsterdam,
1984; part B, p 177ff (particularly p 198).
(20) Stork, G.; Brizzolara, A.; Landesman, H.; Szmuszkovicz, J.; Terrell,
R. J. Am. Chem. Soc. 1963, 85, 207. For a general review including
experimental procedures, see: Whitesell, J. K., Whitesell, M. A. Synthesis
1983, 517-536.
(21) Lucero, M. J.; Houk, K. N. J. Am. Chem. Soc. 1997, 119, 826. For
leading references, see: Pfau, M.; Revial, G.; Guingant, A.; d’Angelo, J. J.
J. Am. Chem. Soc. 1985, 107, 273. d’Angelo, J.; Desmae¨le, D.; Dumas, F.;
Guingant, A. Tetrahedron: Asymmetry 1992, 3, 459. Jabin, I.; Revial, G.;
Tomas, A.; Lemoine, P.; Pfau, M. Tetrahedron: Asymmetry 1995, 6, 1795.
Cave´, C.; Desmae¨le, D.; d’Angelo, J. J. Org. Chem. 1996, 61, 4361.
(15) Reymond, J.-L. Angew. Chem., Int. Ed. Engl. 1995, 34, 2285.
(16) Dean, J. A. In Lange’s Handbook of Chemistry; McGraw-Hill: San
Francisco, CA, 1992; pp 8.19-8.71.

2776 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Hoffmann et al.
Scheme 6. Preparation of Fluorinated Aldols (syn-16, anti-16, and 60) by Antibody-Catalyzed Aldol Reaction
Scheme 7. Antibody- or Amine-Catalyzed Retroaldol
Reaction of Substrate 59
Table 8. Comparison of Antibody 38C2-Catalyzed
Retroaldolization of 59 with Amine-Catalyzed Retroaldolization in
Aqueous and Organic Solvents
amine/solvent
(kcat/Km)a/kamine
kcat/kamine (M)
n-butylamine/MOPSb
n-butylamine/n-octanol
aminoacetonitrile/MOPSb
aminoacetonitrile/n-octanol
1.3 × 108
2.1 × 106
2.9 × 107
1.6 × 107
3.5 × 104
5.6 × 102
7.8 × 103
4.2 × 103
a kcat and Km of antibody 38C2-catalyzed retroaldolization (Scheme
7). b 100 mM MOPS buffered water, pH 7.0, 25 °C.
was seen at this time point. This result contrasts what we
observed in similar studies of 38C2-catalyzed decarboxylation
of a â-keto acid where imine formation is rate-limiting and
catalysis of 18O exchange back into the substrate was not
observed.5 Reactions involving hydroxyacetone as donor are
faster than those involving acetone. This observation is also
in support of enamine formation being at least partially rate-
determining since ab initio molecular orbital calculations predict
a lower energy of activation for the imine to enamine inter-
conversion for hydroxyacetone relative to acetone by ap-
proximately 2 kcal/mol due to the electron-withdrawing char-
acter of the hydroxy group on hydroxyacetone that facilitates
proton abstraction at the R-carbon.22 This together with an
increased relative stability of the enamine formed at the hydroxy-
bearing side of hydroxyacetone explains the perfect control of
the regiochemistry of hydroxyacetone aldols where bond
formation has only been detected at the R-position bearing the
hydroxyl group. By similar arguments, the observed regiose-
lectivity of fluoroacetone addition is also in accord with these
reported ab initio molecular orbital calculations.22 The reversal
in the enantiofacial selectivity of the addition of hydroxyacetone
remains to be completely explained. The R-syn stereochemistry
of the hydroxyacetone-derived products 54 and 58 can be
rationalized by the preferential formation of the Z-enamine of
hydroxyacetone, stabilized over the E-configuration via in-
tramolecular hydrogen bonding, and subsequent attack on the
Figure 2. Detailed representation of the suggested mechanism of ab
38C2 and 33F12 catalysis of the aldol reaction. The rate-determining
step is presumably the C-C bond formation in step 5.
re-face of the acceptor as shown in Figure 4. Branching at the
R-position of the enamine may result in a reorganization of an
activating water molecule or another amino acid side-chain that
serves this function, altering the enantiofacial selectivity. The
observation of both, syn and anti products, in additions of
cyclopentanone to aldehydes, products 51-53, supports the
availability of both faces of the acceptor aldehyde toward attack
with this enamine which is limited to formation of an E-enamine.
Diminished enantioselectivity with substrates such as 6 over
the high enantioselectivities observed with benzylic or allylic
acceptors 3, 7, and 8 is consistent with more efficient enantioface
selectivity with acceptor molecules with fewer degrees of
(22) Lin, J.-F.; Wu; C.-C.; Lien, M.-H. J. Phys. Chem. 1995, 99, 16903.

Aldolase Antibodies of Remarkable Scope
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2777
Figure 5. Consensus ketone-aldehyde cross-aldol substrates for
antibodies 38C2 and 33F12.
exceeding 95%. With hydroxyacetone as the donor substrate,
attack occurs at the re-face of the aldehyde generating an R,â-
dihydroxy ketone with the two stereogenic centers having a
R-syn configuration, (3R,4S). These reactions proceed with 70
to >99% ee. The major product of a cross-aldol reaction with
fluoroacetone as the donor substrate is a syn R-fluoro-â-hydroxy
ketone formed in 95% ee.
These studies highlight the ability of reactive immunization
to produce catalysts that are efficient yet broad in scope. Since
such antibodies are tailor-made and optimized by the immune
system to covalently bind the hapten, the binding pocket is not
necessarily refined with respect to noncovalent interactions with
the immunogen.4c Consequently, the biocatalyst can accept a
variety of substrates which differ enormously with respect to
their physicochemical properties. Antibodies 38C2 and 33F12
are useful additions to the repertoire of asymmetric catalysts
due to their tremendous scope, efficiency, and stereoselectivities.
This has been demonstrated herein with syntheses up to the 1
g scale and in the highly enantioselective total syntheses of 10
different brevicomins.23 Unlike current transition metal cata-
lysts, the antibodies reported here catalyze both, the aldol and
retroaldol reaction, allowing for the preparation of both enan-
tiomers of â-hydroxy ketones with high optical purity either
by enantioselective synthesis or kinetic resolution.24 The issue
of scale in antibody-catalyzed reactions can be addressed with
currently available technologies, particularly the heterologous
expression of antibodies in plants and algaes, where low-cost
production of these catalysts on a multiton scale could be
achieved to allow for the “green synthesis” of aldols on a
virtually unlimited scale.25 Antibody 38C2 is now commercially
available from the Aldrich Chemical Co.
Figure 3. Mass spectra from electrospray MS showing the antibody-
catalyzed O-exchange of the aldol product 59 in buffered 98% 18O-
labeled water. The m/z of 272 corresponds to the starting material
whereas the antibody-catalyzed exchange (presumably via a covalently
bound intermediate) incorporated 18O to produce m/z of 274. Negligible
18O incorporation into 59 was observed in the absence of ab 38C2 under
otherwise identical conditions.
Figure 4. Potential mechanistic origin of enantioselectivities where
acetone and hydroxyacetone serve as aldol donor substrates.
freedom. Further insight into the chemical mechanism of these
catalysts will come with the solution of the X-ray crystal
structure of antibody 33F12 bound to hapten 1. This study is
now underway.
Experimental Section
Conclusions. Antibodies 38C2 and 33F12 are capable of
efficiently catalyzing a wide variety of ketone-ketone, ketone-
aldehyde, aldehyde-ketone, and aldehyde-aldehyde intermo-
lecular aldol reactions, and in some cases to catalyze their
subsequent dehydration to yield aldol condensation products.
A number of intramolecular aldol reactions have also been
defined. Catalysis of all intramolecular aldol reactions examined
yields the corresponding condensation products. The consensus
donor and acceptor substrates are given in Figure 5. The
primary limitation in donor specificity appears to be an inability
of the catalysts to accept ketones that are branched at the non-
bond-forming R-position, whereas acceptor substrates are only
limited to the extent that they are relatively hydrophobic
aldehydes or ketones since no polyhydroxylated aldehydes have
yet been defined as substrates. The scope of these antibody
catalysts exceeds that observed with any known natural enzyme
aldolase or transition metal based aldol catalysts.2,3 When
acetone is the aldol donor substrate in a ketone-aldehyde
crossed aldol reaction, a new stereogenic center is formed by
attack of the si-face of the aldehyde with ee’s in most cases
General. All reactions requiring anhydrous conditions were per-
formed in oven-dried glassware under an Ar or N2 atmosphere.
Chemicals and solvents were either puriss p.A. or purified by standard
techniques. THF was distilled from sodium-benzophenone. Thin-
layer chromatography (TLC): silica gel plates Merck 60 F254, com-
pounds were visualized by irradiation with UV light and/or by treatment
with a solution of 25 g of phosphomolybdic acid, 10 g of Ce(SO4)2‚H2O,
60 mL of concd H2SO4, and 940 mL of H2O followed by heating and/
or by staining with a solution of 12 g of 2,4-dinitrophenylhydrazine in
60 mL of concd H2SO4, 80 mL of H2O, and 200 mL of 95% EtOH
followed by heating and/or by immersing in an iodine bath (30 g of I2,
2 g of KI, in 400 mL of EtOH/H2O 1:1) and warming. Flash
chromatography (FC): silica gel Merck 60 (particle size 0.040-0.063
(23) List, B.; Shabat, D.; Barbas III, C. F.; Lerner, R. A. Chem. Eur. J.
In press.
(24) Zhong, G. et al., submitted.
(25) Hiatt, A. Nature 1990, 334, 469. Hiatt, A.; Ma, J. K. Int. ReV.
Immunol. 1993, 10, 139. Ma, J. K.; Hiatt, A.; Hein, M.; Vine, N. D.; Wang,
F.; Stabila, P.; van Dolleweerd, C.; Mostov, K.; Lehner, T. Science 1995,
268, 716. Franken, E.; Teuschel, U.; Hain, R. Curr. Opin. Biotechnol. 1997,
8, 411.

2778 J. Am. Chem. Soc., Vol. 120, No. 12, 1998
Hoffmann et al.
mm), eluent given in parentheses. 1H NMR: Bruker AMX 300, Bruker
AMX 250. The chemical shifts are given in δ relative to TMS (δ )
0 ppm), the coupling constants J are given in Hz. The spectra were
recorded in CDCl3 as solvent at room temperature unless stated
otherwise. HR-MS: liquid secondary ionization (LSI-MS): VG ZAB-
ZSE with 3-nitrobenzyl alcohol matrix.
Antibody Stability. The antibodies 38C2 and 33F12 are stable at
room temperature for weeks dissolved in different buffer solutions (pH
5.5 to 8.5) and even pure water. They can be lyophilized and passed
over a Sephadex column with less than 5% activity loss. No detectable
activity loss was found if the antibodies were stored in stock solutions
of 10 to 20 mg/mL in phosphate-buffered saline (PBS) (10 mM
phosphate, 150 mM NaCl, pH 7.4) at -78 °C.
at this temperature for 30 min, the aldehyde (1.0 mmol), dissolved in
2 mL of THF, was added over a period of 1 min. After stirring for
5-30 min at -78 °C, saturated NH4Cl solution (1 mL) was added,
and the reaction mixture was allowed to warm to room temperature.
The product was extracted with ethyl acetate (3 × 10 mL), dried
(MgSO4), and evaporated. The pure aldol products were obtained by
FC. Spectroscopic data for the acetone addition products with
aldehydes 3-8 and 48 are given below as examples.
6-(4′-Acetamidophenyl)-4-hydroxy-2-hexanone (9). FC (ethyl
acetate/hexane 75:25) gave 0.21 g (85%) of pure aldol 9. 1H NMR
(300 MHz, CDCl3): δ 8.12 (s, br, 1 H), 7.37 (d, J 8.4, 2 H), 7.06 (d,
J 8.4, 2 H), 4.01 (m, 1 H), 3.42 (m, 1 H), 2.71 (m, 1 H), 2.62 (m, 3 H),
2.13 (s, 3 H), 2.10 (s, 3 H), 1.75 (m, 1 H), 1.64 (m, 1 H); HR-MS:
250.1450; C14H20O3N+ (calcd 250.1443); C14H19O3N (249.31).
4-(4′-Acetamidophenyl)-4-hydroxy-2-butanone (18). FC (ethyl
acetate/hexane 70:30) gave 0.18 g (82%) of pure aldol 18. 1H NMR
(250 MHz, DMSO-d6): δ 9.88 (s, 1 H), 7.49 (d, J 8.5, 2 H), 7.24 (d,
J 8.5, 2 H), 5.28 (d, J 4.4, 1 H), 4.91 (m, 1 H), 2.65 (m, 2 H), 2.09 (s,
3 H), 2.01 (s, 1 H); HR-MS: 244.0955; C12H15O3NNa+ (calcd
244.0949); C12H15O3N (221.26).
4-(4′-Isobutyramidophenyl)-4-hydroxy-2-butanone (59). FC (eth-
yl acetate/hexane 40:60) gave 0.23 g (91%) of pure aldol 59. 1H NMR
(250 MHz, CDCl3): δ 7.50 (d, J 8.5, 2 H), 7.30 (d, J 8.5, 2 H), 7.27
(s, br, 1 H), 5.12 (m, 1 H), 3.36 (s, 1 H), 2.84 (m, 2 H), 2.51 (pent, J
6.8, 1H), 2.20 (s, 3 H), 1.25 (d, J 6.8, 6 H); HR-MS: 272.1267;
C14H19O3NNa+ (calcd 272.1263); C14H19O3N (249.31).
6-(4′-Acetamidophenyl)-4-hydroxy-5-methyl-2-hexanone (21). FC
(ethyl acetate/hexane 70:30) gave 0.21 g (79%) of pure aldol 21. 1H
NMR (300 MHz, CDCl3): δ 7.85 (s, br, 1 H), 7.39 (d, J 8.3, 2 H),
7.07 (d, J 8.3, 2 H), 3.95 (m, 1 H), 3.08 (m, 1 H), 2.82 (m, 1 H), 2.66
(m, 3 H), 2.55 (m, 1 H), 2.37 (m, 1 H), 2.18 (s, 3 H), 2.15 (s, 3 H),
1.76 (m, 1 H), 0.85 (d, J 6.6, 3 H); HR-MS: 264.1608; C15H22O3N+
(calcd 264.1600); C15H21O3N (263.34).
7-(4′-Acetamidophenyl)-4-hydroxy-2-heptanone (24). FC (ethyl
acetate/hexane 70:30) gave 0.23 g (89%) of pure aldol 24. 1H NMR
(300 MHz, CDCl3): δ 7.39 (d, J 8.4, 2 H), 7.24 (s, br, 1 H), 7.13 (d,
J 8.4, 2 H), 4.05 (m, 1 H), 3.04 (m, 1 H), 2.57 (m, 4 H), 2.17 (s, 3 H),
2.16 (s, 3 H), 1.70 (m, 2 H), 1.45 (m, 2 H); HR-MS: 286.1411;
C15H21O3NNa+ (calcd 286.1419); C15H21O3N (263.34).
4-(4′-Nitrophenyl)-4-hydroxy-2-butanone (27). FC (ethyl acetate/
hexane 50:50) gave 0.18 g (86%) of pure aldol 27. 1H NMR (250
MHz, CDCl3): δ 8.21 (d, J 7.2, 2 H), 7.52 (d, J 7.2, 2 H), 5.25 (m, 1
H), 3.56 (d, J 3.2, 1 H), 2.83 (m, 2 H), 2.21 (s, 1 H); HR-MS: 232.0591;
C10H11O4NNa+ (calcd 232.0586); C10H11O4N (209.20).
6-(4′-Nitrophenyl)-4-hydroxy-5-hexen-2-one (30). FC (ethyl acetate/
hexane 50:50) gave 0.21 g (89%) of pure aldol 30. 1H NMR (250
MHz, CDCl3): δ 8.15 (d, J 7.3, 2 H), 7.46 (d, J 7.3, 2 H), 6.71 (d, J
15.9, 1 H), 6.35 (dd, J 15.9, J 5.3, 1 H) 4.78 (m, 1 H), 3.28 (d, J 3.7,
1 H), 2.73 (m, 2 H), 2.21 (s, 1 H); HR-MS: 258.0751; C12H13O4NNa+
(calcd 258.0742); C12H13O4N (235.24).
Preparation of 4-(4′-Acetamidophenyl)butyraldehyde (6). Al-
dehyde 6 was prepared in four steps starting from commercially
available 4-(4′-aminophenyl)butyric acid as follows.
(i) 4-(4′-Acetamidophenyl)butyric Acid. 4-(4′-Aminophenyl)-
butyric acid (4.0 g, 22 mmol) was added to 150 mL of mixed solvent
of acetonitrile and water (9/1). Acetic anhydride (7.4 g, 55 mmol, 2.5
equiv) was added at 0 °C. Then the reaction mixture was stirred at
room temperature for 4 h. The filtration of the reaction mixture was
followed by drying at 120 °C overnight to give 4.53 g of 4-(4′-
acetamidophenyl)butyric acid (93%). HR-MS: 222.1140; C12H16O3N+
(calcd 222.1130); C12H15O3N (221.26).
(ii) 4-(4′-acetamidophenyl)butyric Acid Methyl Ester. 4-(4′-
Acetamidophenyl)butyric acid (4.50 g, 20 mmol) and potassium
carbonate (2.90 g, 21 mmol) were added to 20 mL of dry DMF. The
reaction mixture was stirred at room temperature for 15 min. Then
methyl iodide (14.2 g, 0.10 mol) was added under nitrogen. The
reaction mixture was stirred at room-temperature overnight. After
evaporation of the solvent and methyl iodide, a solid mixture was
obtained, from which the methyl ester was isolated by extraction with
ethyl acetate (3 × 70 mL) to 4.7 g (>99%). 1H NMR (300 MHz,
CDCl3): δ 7.41 (d, J 8.4, 2 H), 7.34 (s, br, 1 H), 7.12 (d, J 8.4, 2 H),
3.67 (s, 3 H), 2.61 (t, J 7.3, 2 H), 2.32 (t, J 7.3, 2 H), 2.17 (s, 3 H),
1.93 (hept, J 7.3, 2 H); HR-MS: 236.1296; C13H18O2N+ (calcd
236.1287); C13H17O3N (235.28).
(iii) 4-(4′-Acetamidophenyl)butanol. 4-(4′-Acetamidophenyl)bu-
tyric acid methyl ester (4.70 g, 20 mmol) was dissolved in 50 mL of
dry THF, and DIBALH in methylene chloride (1.0 M, 40 mL) was
dropwise added at -30 °C. The mixture was kept stirring at this
temperature for 3 h. Saturated ammonium chloride (25 mL) was added
slowly. Extraction with 3 × 80 mL of ethyl acetate followed by
evaporation of the solvent gave a residue which was purified by FC
(hexane/ethyl acetate 4:1) to give 2.73 g (66%) of 4-(4′-acetamidophen-
yl)butanol. 1H NMR (300 MHz, CDCl3): δ 7.42 (d, J 8.4, 2 H), 7.23
(s, br, 1 H), 7.13 (d, J 8.4, 2 H), 3.71 (t, J 7.2, 2 H), 2.56 (t, J 7.2, 2
H), 2.17 (s, 3 H), 1.65 (m, 4 H); HR-MS: 208.1342; C12H18O2N+ (calcd
208.1338); C12H17O2N (207.27).
(iv) 4-(4′-Acetamidophenyl)butyraldehyde (6). A mixture of
methylene chloride (25 mL) and oxalyl chloride (1.0 mL, 11 mmol)
were placed in a flask. Dimethyl sulfoxide (1.7 mL, 22 mmol) diluted
with methylene chloride (5 mL) was added to the stirred solution at
-78 °C. The reaction mixture was stirred for 2 min, and 4-(4′-
acetamidophenyl)butanol (2.2 g, 10 mmol) in 10 mL of methylene
chloride was added within 5 min. The stirring was continued for
additional 15 min. Triethylamine (7.0 mL, 50 mmol) was added and
the reaction mixture was stirred for 5 min and then allowed to warm
to room temperature. Water (50 mL) was added, and the aqueous layer
was reextracted with methylene chloride (50 mL). The organic layers
were combined, washed with saturated sodium chloride solution (100
mL), and dried with magnesium sulfate. After concentration, 2.07 g
(94%) of pure 4-(4′-acetamidophenyl)butyraldehyde (6) was obtained
by FC (hexane/ethyl acetate 3:1). 1H NMR (300 MHz, CDCl3): δ
9.76 (s, br, 1 H), 7.42 (d, J 8.4, 2 H), 7.23 (s, br, 1 H), 7.13 (d, J 8.4,
2 H), 2.63 (t, J 7.2, 2 H), 2.45 (t, J 7.2, 2 H), 2.17 (s, 3 H), 1.94 (hept,
J 7.2, 2 H); HR-MS: 206.1185; C12H16O2N+ (calcd 206.1181);
C12H15O2N (205.26).
4-Hydroxynona-5,7-dien-2-one (49). FC (ethyl acetate/hexane 15:
85) gave 88 mg (57%) of pure aldol 49. 1H NMR (300 MHz, CDCl3):
δ 6.21 (m, 1 H), 6.05 (m, 1 H), 5.74 (m, 1 H), 5.54 (m, 1 H), 4.56 (m,
1 H), 2.93 (m, 1 H), 2.62 (d, J 6.2, 2 H), 2.18 (s, 3 H), 1.75 (d, J 6.5,
3 H); HR-MS: 177.0899; C9H14O2Na+ (calcd 177.0892); C9H14O2
(154.21).
Antibody Catalysis. Preparative Antibody-Catalyzed Reactions.
Example 1. Preparation of (S)-4-Hydroxy-6-(4-nitrophenyl)-5-
hexen-2-one (30). To a solution of 4-nitrocinnamaldehyde (110 mg,
0.61 mmol) in 15 mL of DMF and 31 mL of acetone was added PBS
buffer (571 mL, degassed and kept under argon) slowly to avoid
precipitation. Antibody 38C2 (8.0 mL of a 120 µM solution) was
added. The final concentrations of 4-nitrocinnamaldehyde and Ab38C2
were 1.0 mM and 1.9 µM, respectively, in a total volume of 625 mL
containing 5% (v/v) of acetone. The reaction mixture was kept in a
dark place at room temperature for 7 days under argon. The reaction
mixture was saturated with sodium chloride and extracted with 3 ×
150 mL of ethyl acetate. The extracts were dried over MgSO4 and
evaporated to yield 140 mg of crude product. Purification by FC (1:2,
ethyl acetate/hexane) gave 96 mg (67%) of pure aldol product (30)
with an ee of 91%.
General Procedure for the Preparation of Aldol Products. The
corresponding ketone (1.0 mmol) was added to a freshly prepared
solution of LDA (1.05 mmol) in 2 mL of THF at -78 °C. After stirring

Aldolase Antibodies of Remarkable Scope
J. Am. Chem. Soc., Vol. 120, No. 12, 1998 2779
Example 2. Enantioselective Preparation of Fluorinated Aldols.
A solution of 3-(4′-acetamidophenyl)propanal 4 (15 mg, 0.078 mmol)
in 0.2 mL of DMF, 0.5 mL of fluoroacetone, and 8.0 mL of PBS buffer
was added to Ab38C2 (1.5 mL of a 120 µM solution). The final
concentrations of 3-(4′-acetamidophenyl)propanal and Ab38C2 were
7.6 mM and 17.5 µM, respectively, in a total volume of 10.2 mL
containing 5% (v/v) of fluoroacetone. The reaction mixture was kept
at room temperature for 21 days. Three aldol products syn-16 (ee 95%),
anti-16 (ee 34%), and 60 (ee 97%) (Scheme 6) were isolated by
semipreparative RP-HPLC (column: VYDAC protein and peptide C18,
λ ) 254 nm, 12% CH3CN/88% water with 0.1% TFA, 4.0 mL/min,
1.0 mL reaction mixture/injection) to give syn-16 (12 mg, tR ) 20.8
min, yield 61%), anti-16 (3.6 mg, tR ) 18.2 min, yield 21%), and 60
(1.2 mg, tR ) 16.3 min, yield 7%). The overall yield was 82% (15%
of aldehyde 4 was recovered).
Syn-isomer of 6-(4′-acetamidophenyl)-3-fluoro-4-hydroxy-2-hex-
anone (syn-16): 1H NMR (300 MHz, CDCl3): δ 7.41 (d, J 8.4, 2 H),
7.22 (s, br, 1 H), 7.15 (d, J 8.4, 2 H), 4.62 (dd, J 48.8, 4.9, 2 H), 3.97
(m, 1 H), 2.85 (m, 1 H), 2.67 (m, 1 H), 2.30 (d, J 5.1, 3 H), 2.18 (s,
3 H), 1.84 (m, 2 H).
Anti-isomer of 6-(4′-acetamidophenyl)-3-fluoro-4-hydroxy-2-hex-
anone (anti-16): 1H NMR (300 MHz, CDCl3): δ 7.42 (d, J 8.4, 2 H),
7.17 (s, br, 1 H), 7.16 (d, J 8.4, 2 H), 4.64 (dd, J 48.9, 2.4, 1 H), 4.00
(dm, J 24.5, 1 H), 2.77 (m, 2 H), 2.31 (d, J 5.0, 3 H), 2.18 (s, 3 H),
1.94 (m, 2 H).
as eluents at a flow rate of 1.5 mL/min or 1.0 mL/min. Formation of
products 51-53 was followed by gas chromatography (DB-5, J&W
Scientific, length: 30 m, i.d.: 0.32 mm, temp: 65 °C (2 min), rate:
10 °C/min, tR: (anti-51) ) 13.82 min, (syn-51) ) 14.05 min; (anti-
52) ) 12.56 min, (syn-52) ) 12.82 min; (anti-53) ) 11.24 min, (syn-
53) ) 11.52 min).
Specific Rates of Cross-Aldol Reactions. The specific rates of
cross-aldol reactions were determined before 10% completion of the
reactions using initial concentrations of the acceptor substrate (500 µM),
antibody (2 µM), and donor ketone (1.0 M).
Specific Rates of Self-Aldol and Intramolecular Aldol Reactions.
The specific rates of self-aldol and intramolecular aldol reactions were
determined before 10% completion of the reactions using initial
concentrations of the substrate (0.10 M) and antibody (5 µM).
Michaelis-Menten Kinetics. Product formation or percent conver-
sion of antibody-catalyzed reaction mixtures was monitored by HPLC
or GC. The experimental data was plotted using nonlinear regression
analysis with GraFit software to give kcat and KM of the reactions. All
data are reported per antibody active site. An IgG antibody possesses
2 active sites per MW of ∼150000 g/mol.
Determination of Enantiomeric Excess of Products 9, 18, 59, 21,
24, 27, 30, and 54. To a 6.25 mM solution of the aldehyde in 160 µL
of PBS were added 10 µL of acetone and 30 µL of a 135 µM solution
of the antibody in PBS. The final concentrations were 5 mM of
aldehyde and 20 µM of antibody in a total volume of 200 mL of PBS
containing 5% (v/v) of acetone. After 18 h, 12 mL of CH2Cl2 were
added, and the organic phase was dried (MgSO4) and evaporated. In
case of products 9, 18, 59, 21, 24, 27, and 30 the residue was redissolved
in ca. 1 mL of 2-propanol and the ee was determined by normal phase
HPLC using an appropriate Daicel column for enantiomer separation.
The ee of product 54 was determined using GC on a chiral capillary
column (Cyclodex-B, J&W Scientific, length: 30 m, i.d.: 0.25 mm,
temp: 150 °C, tR: (R) ) 15.48 min, (â) ) 16.01 min).
Regioisomer, 6-(4′-acetamidophenyl)-1-fluoro-4-hydroxy-2-hex-
anone (60): 1H NMR (300 MHz, CDCl3): δ 7.43 (d, J 8.4, 2 H), 7.17
(d, J 8.4, 2 H), 7.14 (s, br, 1 H), 4.83 (d, J 47.5, 2 H), 4.12 (m, 1 H),
2.73 (m, 4 H), 2.18 (s, 3 H), 1.81 (m, 2 H).
Example 3. Preparative Scale Synthesis of 2-[1′-(4′′-Isobutyra-
midophenyl)-1′-hydroxymethyl]cyclopentanone (61). Cyclopen-
tanone (1 mol, 88 mL) was dissolved in 912 mL of MOPS buffer (100
mM, pH ) 7.4). Next, Ab38C2 (1.3 µmol, 0.1 g) was added, and the
first addition of aldehyde, 4-isobutyramidobenzaldehyde (1.1 mmol,
213 mg), was made. The reaction was stirred for 24 h followed by a
second addition of aldehyde. Two subsequent additions of aldehyde
followed at 24 h time periods for a total of 852 mg, 4.4 mmol. The
reaction progress was monitored by HPLC (Hitachi HPLC system:
pump L-7100, UV detector L-7400, and integrator D-7500) using a
Rainin column (Microsorb-MV, C18, 300 Å, 5 mm; 250 × 4.6 mm)
and acetonitrile/water mixture (20% CH3CN/80% water containing 0.1%
trifluoroacetic acid) with a flow rate of 1.0 mL/min. The reaction
mixture was kept in a dark place at room temperature for 21 days under
argon. The reaction mixture was then saturated with NaCl. The
mixture was extracted with 3 × 500 mL of ethyl acetate, dried over
MgSO4, and evaporated to yield 1.4 g of crude product. Purification
by FC (60:40, EtAc/Hex) gave 0.9 g (72%) of pure product 61 with a
de of >95%. 1H NMR (300 MHz, CDCl3): δ 7.49 (d, J 8.4, 2 H),
7.31 (s, br, 1 H), 7.27 (d, J 8.4, 2 H), 5.24 (s, 1 H), 4.62 (m, 1 H),
2.6-1.6 (m, 7 H), 1.26 (d, J 6.8, 6 H).
Antibody Assays. All antibody-catalyzed reactions were performed
in phosphate buffered saline (10 mM phosphate, 150 mM NaCl, pH
7.4) except reactions with substrate 43 (Scheme 4, H2O, pH 6.5) and
substrates n-C4H9CHO, n-C5H11CHO, n-C6H13CHO (Table 6, (3-
morpholinopropanesulfonic acid) sodium salt buffer (MOPS buffer),
pH 7.0). All antibody-catalyzed reactions and background reactions
with substrates 3-8, 33, 37, 39, 41, 43, 45, and 48 were monitored by
high-pressure liquid chromatography (HPLC; Hitachi HPLC system
(pump L-7100, UV detector L-7400, and integrator D-7500) using a
Rainin column (Microsorb-MV, C18, 300 Å, 5 mm; 250 × 4.6 mm)
and acetonitrile/water mixtures (containing 0.1% trifluoroacetic acid)
18O Incorporation. The electron spray ionization (ESI) mass
spectrometry used to monitor 18O incorporation into 59 was performed
on an API III Perkin-Elmer SCIEX triple quadropole mass spectrometer.
In preparation of the sample, lyophilized antibody 38C2 was resus-
pended in 18O labeled water (18O, 95-98%, Cambridge Isotope
Laboratories, Andover, MA) to give a final concentration of 7.5 µM.
The reaction was started by addition of aldol product 59 (1.5 mM),
and aliquots were taken out for analysis of 18O incorporation over time.
Immediately before analysis the samples were diluted 10-fold in
methanol.
Acknowledgment. We are grateful to M. G. Finn for helpful
discussions, Brian Bothner and Gary Suizdak for assistance with
MS, Ron Lewis for studies of Michael reactions, and Gunther
Schlingloff as well as other members of the Sharpless laboratory
for their help with ee measurements. This study was supported
in part by the NIH (CA27489). T.H and B.L. thank the
Alexander von Humboldt Foundation, Germany, for a Feodor
Lynen fellowship. C.F.B. acknowledges an Investigator Award
from the Cancer Research Institute.
Supporting Information Available: Kinetic data, nondonor
substrate structures, and diastereoselectivity studies (7 pages).
See any current masthead page for ordering and Web access
instructions.
JA973676B