Reconstructing aldolase antibodies to alter their substrate specificity and turnover

Full text of DOI:10.1021/ja0005441

J. Am. Chem. Soc. 2000, 122, 4835-4836
4835
Reconstructing Aldolase Antibodies to Alter Their
Substrate Specificity and Turnover
Although these antibodies possess unusually promiscuous active
sites capable of catalyzing a wide variety of aldol and retro-aldol
reactions, the efficiency with which any given aldol is processed
can vary significantly. For example, retro-aldol reactions of
cyclohexanone-aldols 2 are relatively slow compared to those
involving acetone-aldols 3. The keto functionality appears to be
key in determining the relative efficiency of processing by these
catalysts since 3 with a wide variety of R groups are processed
efficiently. Therefore, we focused on the alteration of substrate
specificity using this strategy with the aim of preparing catalysts
that would more efficiently process 2.
Fujie Tanaka, Richard A. Lerner,* and Carlos F. Barbas, III*
The Skaggs Institute for Chemical Biology and the
Department of Molecular Biology,
The Scripps Research Institute
0550 North Torrey Pines Road, La Jolla, California 92037
1
ReceiVed February 15, 2000
To retain the catalytic function of the residue LysH93, we took
Many enzymes use common catalytic mechanisms in the
catalysis of analogous chemical transformations, for example in
the hydrolysis of esters and amides. Conservation of catalytic
mechanism is observed not only among evolutionarily related and
highly homologous enzymes but also between structurally dif-
advantage of information provided by the crystal structure of
a
3
3F12.6 The structure showed the reactive lysine at the bottom
of a deep binding pocket where most of the residues within a 4
Å radius of the ꢀ-amino group of LysH93 are hydrophobic and
are thus likely key in tuning the pK of this amine group. These
a
residues are SerH35, ValH37, TrpH47, TyrH95, TrpH103, and
PheL98. TyrH95 and TrpH103 are found in HCDR3 (heavy chain
complementarity determining region 3), and PheL98 is in LCDR3.
Therefore, the sequences of the LysH93, HCDR3, and LCDR3
of the aldolase antibodies were retained in the library. A naive
1
ferent and evolutionarily unrelated enzymes. Such observations
provide evidence for evolutionary convergence at the level of
chemical mechanism. It is anticipated that structural heterogeneity
may provide for distinct opportunities for the optimization of
catalytic efficiency with different substrates where chemical
mechanism is conserved. In the case of catalytic antibodies,
immunization of mice with transition-state analogues designed
for the preparation of hydrolytic antibodies has provided a variety
of catalytic antibodies that demonstrate high levels of homology
H
antibody heavy chain variable domain (V ) library was generated
using human bone marrow cDNA and fused to the parental heavy
8
chain sequences at H93. The LCDR3 sequences of the parental
2
antibodies were placed in the context of an unrelated human light
chain, that possessed a radically different amino acid sequence
at both structural and mechanistic levels. While the immune
repertoire provides a highly diverse array of antibody genes from
which to select catalysts, immune responses to haptens are often
9
as compared to the parental aldolase antibodies. The phage
displayed libraries8 were selected by three rounds of panning
against 1,3-diketones 4- and 1-BSA (bovine serum albumin) in
order to select antibodies that would accept both 2 and 3. In a
,10
3
highly restricted to a few favored V genes. This fact acts to
experimentally limit our opportunities to probe the structural
repertoire available to antibodies more completely. To search for
novel aldolase antibodies we sought a strategy that would not be
limited by the need to reimmunize animals but one that would
take advantage of insights gained by the study of existing catalysts.
Here we examine a new strategy to obtain improved catalytic
antibodies by recombining the catalytic machinery of parental
antibodies with a naive V gene repertoire. Antibody libraries of
this type may provide new evolutionary opportunities for catalysis
not accessible through the relatively small, five or six amino acid
4
residues changes in protein sequence that are typically probed
to prepare modified or enhanced catalysts.
Aldolase antibodies 38C2 and 33F12 were previously generated
5
by reactive immunization with 1,3-diketone 1. The antibodies
subsequent diversification step, V libraries were combined with
L
are highly homologous with respect to sequence, structure, and
the selected V libraries, and three additional rounds of selection
H
catalytic mechansim.5 Both antibodies possess a highly reactive
,6
were performed.
Screening of the phage selected clones by ELISA for the
production of soluble Fab capable of binding to both 4- and
7
lysine residue (LysH93) in their active sites that is essential to
their catalytic mechanism. The ꢀ-amino group of the lysine residue
is key in the formation of Schiff base and enamine intermediates
that appear along the reaction coordinate of the aldol reaction.5,6
(6) (a) Barbas, C. F., III; Heine, A.; Zhong, G.; Hoffmann, T.; Gramatikova,
S.; Bj o¨ rnestedt, R.; List, B.; Anderson, J.; Stura, E. A.; Wilson, I. A.; Lerner,
R. A. Science 1997, 278, 2085. (b) 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. (c) Zhong, G.; Hoffmann, T.; Lerner, R. A.;
Danishefsky, S.; Barbas, C. F., III. J. Am. Chem. Soc. 1997, 119, 8131. (d)
List, B.; Shabat, D.; Barbas, C. F., III; Lerner, R. A. Chem. Eur. J. 1998, 4,
881. (e) Zhong, G.; Shabat, D.; List, B.; Anderson, J.; Sinha, R. A.; Lerner,
R. A.; Barbas, C. F., III. Angew. Chem., Int. Ed. 1998, 110, 2609. (f) List, B.;
Shabat, D.; Zhong, G.; Turner, J. M.; Li, A.; Bui, T.; Anderson, J.; Lerner, R.
A.; Barbas, C. F., III. J. Am. Chem. Soc. 1999, 121, 7283. (g) Shabat, D.;
Rader, C.; List, B.; Lerner, R. A.; Barbas, C. F., III. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 6925. (h) List, B.; Lerner, R. A.; Barbas, C. F., III. Org.
Lett. 1999, 1, 59. (i) Shabat, D.; List, B.; Lerner, R. A.; Barbas, C. F., III.
Tetrahedron Lett. 1999. 40, 1437. (j) Sinha, S.; Barbas, C. F., III; Lerner, R.
A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 14603.
(7) The numbering is according to Kabat et al. Kabat, E. A.; Wu, T. T.;
Perry, H. M.; Gottesman, K. S.; Foeller, C. Sequences of Proteins of
Immunological Interest, 5th ed.; U.S. Public Health Service, National Institute
of Health: Bethesda, MD, 1991.
(8) Barbas, C. F., III, Burton, D. R., Scott, J., Silverman, G., Eds. Phage
Display of Proteins and Peptides: A Laboratory Manual; Cold Spring Harbor
Laboratory Press: Cold Spring Harbor, New York, 2000.
(
1) (a) Branden, C.; Tooze, J. Introduction to Protein Structure; Garland
Publishing: New York, 1991; p 236. (b) Russell, R. B. J. Mol. Biol. 1998,
2
79, 1211. (c) Rao, J. K.; Erickson, J. W.; Wlodawer, A. Biochemistry 1991,
0, 4663.
3
(
2) (a) Angeles, T. S.; Smith, R. G.; Darsley, M. J.; Sugasawara, R.;
Sanchez, R. I.; Kenten, J.; Schultz, P. G.; Martin, M. T. Biochemistry 1993,
3
2, 12128. (b) Miyashita, H.; Hara, T.; Tanimura, R.; Tanaka, F.; Kikuchi,
M.; Fujii, I. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 6045. (c) Fujii, I.; Tanaka,
F.; Miyashita, H.; Tanimura, R.; Kinoshita, K. J. Am. Chem. Soc. 1995, 117,
6
199. (d) MacBeath, G.; Hilvert, D. Chem. Biol. 1996, 3, 433. (e) Buchbinder,
J. L.; Stephenson, R. C.; Scanlan, T. S.; Fletterick, R. J. J. Mol. Biol. 1998,
2
82, 1033. (f) Charbonnier J.-B.; Golinelli-Pimpaneau, B.; Gigant, B.; Tawfik,
D. S.; Chap, R.; Scindler, D. G.; Kim, S.-H.; Green, B. S.; Eshhar, Z.;
Knossow, M. Science 1997, 275, 1140.
(
3) Kaartinen, M.; Solin, M.-L.; M a¨ kel a¨ , O. Eur. J. Immunol. 1991, 21,
2
863.
(
4) (a) Baca, M.; Scanlan, T. S.; Stephenson, R. C.; Wells, J. Proc. Natl.
Acad. Sci. U.S.A. 1997, 94, 10063. (b) Fujii, I.; Fukuyama, S.; Iwabuchi, Y.;
Tanimura, R. Nat. Biotechnol. 1998, 16, 463.
(
5) Wagner, J.; Lerner, R. A.; Barbas, C. F., III. Science 1995, 270, 1797.
1
0.1021/ja0005441 CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/29/2000

4836 J. Am. Chem. Soc., Vol. 122, No. 19, 2000
Communications to the Editor
Table 1. Kinetic Parameters of the Retro-Aldol Reactions of 5a
anti-5
syn-5
kcat, min-
1
kcat/kuncat b
Km, µM
kcat, min-1
kcat/kuncat b
Km, µM
2
2
-1
-2
4
2
-1
-1
-2
-2
4
Fab 28
Fab 28
(
2R,1′S
(
1.8 × 10
1.0 × 10
36
1.0 × 10
9.5 × 10
3.4 × 10
1.1 × 10
2.3 × 10
(
2S,1′Sd
(
1.1 × 10
1.1 × 10
1.3 × 10
2.9 × 10
1.2 × 10
c
d
4
4
2.1 × 10
87
2.9 × 10
1.2 × 10
-
-
2
2
3
2
2
3
3
3
8C2
7.5 × 10
9.7 × 10
4.0 × 10
3
3
3F12
(
97
2.5 × 10
(
4.7 × 10
1.9 × 10
a
b
Reaction conditions: 5% DMSO/PBS (pH 7.4), 25 °C except as noted. The first-order kinetic constant of the background reaction (kuncat) was
-
6
-1
-5
-1
c
4
.5 × 10 min for (()-anti-5 and 2.4 × 10 min for (()-syn-5. Reaction was performed in 5% CH
3
CN/PBS (pH 7.4). The kuncats were the
d
same as those observed in 5% DMSO/PBS (pH 7.4). See ref 15.
1
-BSA, identified 23 clones. The five clones that showed the
nm was observed. This result is consistent with enaminone
formation5 between Fab 28 and 10, and with the reaction
mechanism of the parental antibodies. The amino acid sequences
,6
strongest ELISA signal were chosen for further study, and the
antibody Fab fragments were purified.11 Screening for the ability
to catalyze the retro-aldol reaction of fluorogenic substrate 512
using 0.5 µM of Fab and a mixture of (()-anti-5 (125 µM)-
L H
of V and V of Fab 28 and of the parental antibodies are shown
in Figure 1 in the Supporting Information. As expected from the
library construction, Fab 28 retained sequence elements of the
parental antibodies as was the design of the library. The remaining
protein sequence is of human origin and is not related in primary
sequence to the parental antibodies. To further evaluate the role
of LysH93 in Fab 28, it was mutated to alanine and methionine.
The mutants did not catalyze any of the reactions studied above
(()-syn-5 (125 µM) in 5% DMSO/PBS (pH 7.4) at 25 °C,
identified antibody Fab 28 as the best catalyst, and this antibody
was studied in detail. Fab 28-catalyzed reactions displayed
saturation kinetics described by the Michaelis-Menten equation
in the steady-state of the retro-aldol reaction. The kinetic
parameters are shown in Table 1. The kcat values of Fab 28 were
superior to those of the parental antibodies for these substrates
by approximately 3 to 10-fold. In addition, Fab 28 catalyzed the
1
9
and displayed reduced binding affinities for diketones. Further,
enaminone formation between the mutants and diketones was not
detected. These results are consistent with an essential role of
LysH93 in the catalytic mechanism of Fab 28.
13
retro-aldol reaction of (()-6. The kcat value is similar to that of
antibody 33F12.13 Fab 28-catalyzed reactions were inhibited by
the addition of diketone 7. The K
i
was determined by Dixon
In this work, we have demonstrated that a V gene shuffling
strategy can be used to reconstruct aldolase antibodies while
retaining the catalytic residues and mechanism of the parental
antibodies. This strategy that provides meaningful structural
analysis to be 1.0 µM in 28-catalyzed retro-aldol reaction of (()-
syn-5. To determine the enantioselectivity of the catalyzed reac-
tion, each of the enantiomers of these substrates were separately
studied.14 Fab 28 catalyzed the retro-aldol reactions of (2R,1′S)-
diversity is complementary to the existing lower-complexity
15
4,11
5, (2S,1′S)-5 (Table 1), and of (S)-6. Catalysis of retro-aldol-
randomization procedures
and is analogous to family-based
DNA shuffling strategies.20 Since the aldolase antibody selected
here is now humanized with respect to its primary sequence, it
has potential utility in the activation of designed prodrugs for
ization of the opposite optical isomers was not detected. The
stereochemistries of the preferred substrate enantiomers of Fab
28 are the same as those of the parental antibodies. Fab 28 also
15-17
6g
catalyzed the retro-aldol reaction of (2S,1′S)-8.
This reaction
cancer therapy. In addition to selection with novel diketones,
was also stereoselective, and the ratio of the initial velocities of
8-catalyzed reactions using each of the enantiomers of syn-815
200 µM) was >99:1. This contrasts the results obtained for 38C2-
the libraries in this report might also allow for the selection of
catalysts not accessible though immunization, for example, novel
catalysts by posttranslational modification at LysH93 within the
2
(
2
1
catalyzed reactions of the enantiomers of syn-8 (200 µM) where
an insignificant velocity ratio of 2.4:1 was observed. Fab 28 also
catalyzed the retro-aldol reactions of tertiary aldol (()-9.6
library to prepare cofactor-dependent catalytic antibodies.
Acknowledgment. We thank Marikka Elia, Sally Sweeney, and
Roberta Fuller for technical assistance. This study was supported in part
by the NIH Grant CA27489.
f,18
Supporting Information Available: Alignment of the amino acid
sequences, synthesis of 5, assays, enaminone formation, construction of
the libraries, panning selection, protein production and purification, and
preparation of mutants (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA0005441
(15) The purity of the enantiomeric substrates that were used in the reactions
is as follows: (2R,1′S)-5, 99.0% ee; (2S,1′R)-, (2S,1′S)-, and (2R,1′R)-5,
>99.5% ee; (S)-6, 98% ee; (R)-6, 99.5% ee; (2S,1′S)- and (2R,1′R)-8, >99.5%
ee. The enantiomers were purified using chiral phase HPLC, and the ee was
determined by the HPLC analysis. The absolute stereochemistry of syn-5 was
determined by the analogy of the reactivity in the 28-catalyzed retro-aldol
reactions of known enantiomers, (2S,1′S)- and (2R,1′R)-8. For the determi-
nation of absolute stereochemistry of anti-5, an enantiomer of anti-5 (>99.5%
The key elements of the catalytic mechanism of the parental
antibodies appear to be conserved in Fab 28-catalyzed reactions.
When Fab 28 (2.6 µM) was mixed with 2,4-pentanedione (10)
1
6
(500 µM) in 5% CH
3
CN/PBS (pH 7.4), a new absorption at 318
1
6
(9) The antibody constant regions were taken from the human anti-tetanus
2 3
ee) was epimerized by treating with K CO in MeOH to provide syn-5.
toxoid Fab p313. The antibody light chain of this clone was also used as the
recipient of the LCDR3s of 38C2 and 33F12. Barbas, C. F., III; Kang, A. S.;
Lerner, R. A.; Benkovic, S. J. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 7978.
(16) Denmark, S. E.; Stavenger, R. A.; Wong, K.-T.; Su, X. J. Am. Chem.
Soc. 1999, 121, 4982.
2
(17) The kinetic parameters for (2S,1′S)-8: K
m
2.3 × 10 µM, kcat 4.6 ×
10- min , and kcat/kuncat 1.5 × 10 in 10% CH
2
-1
4
CN/PBS (pH 7.4). The reaction
(
10) Phage display approach for humanization of binding antibodies: Rader,
C.; Cheresh, D.; Barbas, C. F., III. Proc. Natl. Acad. Sci. U.S.A. 1998, 95,
910.
11) Yang, W.-P.; Green, K.; Pinz-Sweeney, S.; Briones, A. T.; Burton,
D. R.; Barbas, C. F., III. J. Mol. Biol. 1995, 254, 392.
3
was monitored by HPLC analysis.
2
-1
8
(18) The kinetic parameters for (()-9: K
m
2.0 × 10 µM, kcat 2.9 × 10
-
1
5
(
min , and kcat/kuncat 2.0 × 10 in 5% CH
(19) The K
competitive ELISA while the K
3
CN/PBS (pH 7.4).
d
of Fab 28 for diketone 7 was determined to be 0.8 µM by
s of the alanine and methionine mutants were
(
12) List, B.; Barbas, C. F., III; Lerner, R. A. Proc. Natl. Acad. Sci. U.S.A.
d
2
2
1
998, 95, 15351.
13) The kinetic parameters for (()-6: Fab 28; K
determined to be 6 × 10 and 8 × 10 µM, respectively. The K
d
was
2
(
m
6.5 × 10 µM, kcat 9.5
determined as the inhibitor concentration required to inhibit 50% of the
maximal binding in the competitive ELISA using 1-BSA coated well.
(20) Crameri, A.; Raillard, S. A.; Bermudez, E.; Stemmer, W. P. Nature
1998, 391, 288.
-2
-1
5
×
10 min , and kcat/kuncat 2.1 × 10 . Antibody 33F12; K
m
43 µM, kcat 1.1
-1
-1
5
×
10 min , and kcat/kuncat 2.4 × 10 (per active site).
(
14) Tanaka, F.; Kinoshita, K.; Tanimura, R.; Fujii, I. J. Am. Chem. Soc.
1
996, 118, 2332 and references therein.
(21) Tanaka, F.; Lerner, R. A.; Barbas, C. F., III. Chem. Commun. 1999,
1383.