Stereoselectivity and expanded substrate scope of an engineered PLP-dependent aldolase

Article abstract of DOI:10.1002/anie.200602529

Optimizing the active site: A pyridoxal 5′-phosphate(PLP)-dependent alanine racemase was converted into a retro-aldolase by a single active-site mutation. This new catalyst reacts efficiently with quaternary β-hydroxy α-amino acids, a class of molecules that natural aldolases have not been evolved to handle. Computational docking experiments suggest a novel mechanism for this process and rationalize the stereochemical preferences of the enzyme. (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200602529

Communications
selective CꢀC bond making and breaking. Indeed,natural
aldolases have been used to catalyze a wide range of aldol
[
2]
reactions on synthetically useful scales.
Furthermore,
enzyme-engineering methods have been applied to generate
catalysts with higher activities or altered substrate specif-
icities to match a growing demand for molecular diversity in
[
3]
the chemical and pharmaceutical industries.
In a previous report,we showed that aldolase activity can
also be recruited from unexpected origins: a single active-site
mutation (Y265A) in a pyridoxal 5’-phosphate (PLP)-depen-
dent alanine racemase from Geobacillus stearothermophillus
(
alr) was found to increase retro-aldolase activity against
[
4]
phenylserine 1 (Scheme 1) by five orders of magnitude. The
Engineered Aldolase
DOI: 10.1002/anie.200602529
Stereoselectivity and Expanded Substrate Scope
of an Engineered PLP-Dependent Aldolase**
Florian P. Seebeck, Angelo Guainazzi, Celine Amoreira,
Kim K. Baldridge, and Donald Hilvert*
Scheme 1. Mechanism of the alrY265A-catalyzedretroal do l reaction of
the b-phenylserine derivatives 1 and 2. The substrate initially reacts
with the enzyme-boundPLP 3 to form the external aldimine adduct 4.
Cleavage of the CaꢀCb bond affords benzaldehyde and quinonoid
Biocatalysis has become an invaluable tool for industrial scale
chemical synthesis. Aldolases are particularly interesting for
practical applications owing to their proficiency at enantio-
[
1]
intermediate 5. Subsequent protonation to give 6, followedby hy dr ol-
ysis, liberates glycine or alanine andregenerates the catalyst. The
cofactor phosphate is the likely base (B) for the retroaldol reaction
(see text), andeither Lys39 or water may protonate 5 (B’H).
[
*] Dr. F. P. Seebeck, A. Guainazzi, Prof. Dr. D. Hilvert
Laboratorium für Organische Chemie
ETH Zürich
Hönggerberg HCI F339, 8093 Zürich (Switzerland)
Fax: (+41)44-632-1486
E-mail: hilvert@org.chem.ethz.ch
conversion of the racemase into an aldolase profited from the
similarity of the central PLP-quinonoid intermediates formed
in the course of both reactions. Interestingly,natural evolu-
tion appears to have exploited this chemical relationship as
well. Genetic analyses have demonstrated that alanine race-
mases and natural d-threonine aldolases share a common
C. Amoreira, Prof. Dr. K. K. Baldridge
Organisch-Chemisches Institut
Universität Zürich
Winterthurstrasse 190, 8057 Zürich (Switzerland)
[
5]
evolutionary origin. However,alanine racemase evolved to
accommodate an alanyl-PLP quinonoid intermediate (5, R =
[
**] We thank the ETH Zürich, the University of Zürich, andthe
Schweizerischer Nationalfonds for financial support. PLP=pyri-
doxal 5’-phosphate.
CH ),whereas d-threonine aldolases stabilize a glycyl-PLP
3
quinonoid species (5, R = H). In the case of our alrY265A
variant,one could imagine that the residual binding pocket
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
6
824
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 6824 –6826

Angewandte
Chemie
for the alanine methyl group might be exploited together with
the new aldolase activity to catalyze the cleavage (or,in the
synthetic direction,formation) of quaternary a-amino acids
like 2 (Scheme 1). Such activity has not been reported for
natural aldolases but would have considerable synthetic
alanine is consistent with the participation of Lys39,the
active-site residue that protonates the quinonoid intermedi-
ate from the si face in native alr,and the absence of Tyr265,
which normally protonates 5 from the opposite face. The
small amount of l-alanine that is formed presumably arises by
direct protonation of the intermediate by water from the same
face as the departing benzaldehyde.
[
6]
utility.
To examine the ability of alrY265A to process b-hydroxy
amino acids bearing an additional a substituent,we synthe-
sized (2R,3S)- and (2R,3R)-a-methyl-b-phenylserines (2a and
To rationalize the stereochemical preferences of
alrY265A,we performed hybrid docking and ab initio
[
7]
[10]
2
b,respectively) following published procedures,
and
investigations with the aldimines of all four a-methyl-b-
measured the kinetic parameters associated with their
alrY265A-catalyzed conversion into alanine and benzalde-
hyde. As shown in Table 1,the two diastereomers are
phenylserine diastereomers (2a and 2b,and the correspond-
[
11]
ing l-amino acids) bound at the enzyme active site.
In
accord with our experimental findings,the modified binding
pocket readily accommodates both 2R diastereomers,
whereas neither of the 2S isomers fits without clashes with
active-site residues. Moreover,energy minimization of the
models reveals that the 2R aldimines exploit very similar
interactions with the binding pocket and the cofactor
Table 1: Steady-state parameters for the conversion of b-phenyl-serines
[
a]
into benzaldehyde and glycine/alanine.
[
c]
Substrate kcat
Km
kcat/Km
kuncat
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
[min
]
[mm] [m min
]
[min ]
(Figure 1). As expected,the carboxylate and
a-methyl
(
(
(
(
(
(
2R,3S)-b-phenyl-serine
1)
2R,3S)-a-methyl-b-phenylserine 2.8
2a)
2R,3R)-a-methyl-b-phenylserine 0.16
2b)
5.7
8.5
2.8
670
0.0015
0.00033
–
groups lock these substrates in an orientation that places
the benzyl side chain perpendicular to the plane of the PLP
cofactor,with the aryl ring inserted into the cavity created by
the Y265A substitution. The preferred conformations of the
two diastereomers differ by a 308 rotation around the Caꢀb
[
b]
1000
0.14 1100
bond,so that the aryl ring of the 2 R,3R isomer (2b) is a little
closer to the side chain of His166. As a consequence of this
difference in dihedral angle,the b-hydroxy group of both
isomers points toward the cofactor phosphate. The resulting
O···O distance is 2.70 and 3.05 Š for 2a and 2b,respectively,
which corresponds to a favorable hydrogen bond.
[
1
a] The enzymatic reactions were performedwith alrY265A at 30 8C in
00 mm HEPES buffer (pH 8). [b] See Ref. [4]. [c] The rate constant for
the pyridoxal-dependent retroaldol reaction of b-phenylserines in the
absence of enzyme was determined following procedures described in
Refs.[4,13]; the standard error on all kinetic parameters is ꢂ20%.
Although His166 was originally proposed as a possible
candidate for the base that initiates proton abstraction from
consumed with an even higher catalytic efficiency (k /K >
cat
m
3
ꢀ1
ꢀ1
[4]
1
0 m min ) than the previously examined (2R,3S)-b-phen-
the substrate hydroxy group (Scheme 1, 4!5), the modeling
ꢀ
1
ꢀ1
ylserine (1; 670m min ). Although the individual steady-
state parameters kcat and Km are likely to be complex
amalgams of different microscopic rate and binding constants,
results indicate that its imidazole ring adopts the wrong
tautomeric form and is too far from the alcohol to assume this
role. Instead,as shown in Figure 1,the PLP phosphate group
is well positioned to serve as the catalytic base for both
diastereomers. The postulated recruitment of the cofactor
phosphate for catalysis has some precedent in the reaction
the fact that the K values observed for 2a and 2b are lower
m
than that for 1 can be attributed to favorable interactions of
the extra a-methyl group with the binding pocket. Further-
more,while its kcat parameter is only twofold lower than that
[
12]
catalyzed by serine dehydratase.
If this mechanistic
of 1,substrate 2a shows a k /kuncat ratio that is twofold higher
hypothesis can be confirmed,it may be possible to exploit
subtle steric changes within the active site to disrupt the
crucial hydrogen bond for one but not the other substrate
epimer,thereby enhancing the C b stereoselectivity of the
enzyme.
cat
(
Table 1).
Natural d-threonine aldolases exhibit stringent stereo-
selectivity at Ca in the cleavage of b-hydroxy amino acids but
[
8]
relaxed selectivity at Cb. alrY265A is similarly selective
with respect to Ca,accepting only d-configured amino
In conclusion,we have shown that an engineered alr
catalyzes a biologically unprecedented retroaldol reaction of
a-substituted b-phenylserines with high stereoselectivity at
the a carbon but low stereochemical control at the b carbon.
[
4]
acids. However,the fact that 2a and 2b afford similar kcat/
K values (Table 1) indicates poor stereochemical control at
m
the b carbon. Nor does the enzyme exhibit strict stereochem-
ical fidelity in the final stages of the reaction. In principle,the
4
Not only is this reaction catalyzed some 10 -fold more
quinonoid intermediate 5 (R = CH ) could be protonated
efficiently than by the cofactor in solution,but our computa-
tional model provides a basis for further optimization of the
alrY265A active site by means of site-directed mutagenesis or
directed evolution. Such improvement may yield practical
catalysts for the enantioselective synthesis of novel amino
acids and their derivatives.
3
either from the si face of the imine to give d-alanine,or from
[
9]
its re face to give l-alanine. The partitioning of 5 (R = CH3)
was quantified by measuring the amount of d-alanine formed
in the alrY265A-catalyzed cleavage of 2b relative to benzal-
dehyde. d-Alanine-dependent oxidation of NADH,using d-
amino acid oxidase and lactate dehydrogenase as coupling
enzymes,showed that 80% ( ꢁ 10%) of the alanine product
has the d configuration. Preferential formation of d- over l-
Received: June 23,2006
Published online: September 26,2006
Angew. Chem. Int. Ed. 2006, 45, 6824 –6826
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6825

Communications
[
3] S. Fong,T. D. Machajewski,C. C. Mak,C.-H. Wong, Chem. Biol.
000, 7,873; N. Wymer,L. V. Buchanan,D. Henderson,N.
2
Mehta,C. H. Botting,L. Pocivavsek,C. A. Fierke,E. J. Toone,
J. H. Naismith, Structure 2001, 9,1; G. DeSantis,J. J. Liu,D. P.
Clark,A. Heine,I. A. Wilson,C.-H. Wong, Bioorg. Med. Chem.
2
003, 11,43; J. S. Griffiths,M. Cheriyan,J. B. Corbell,L.
Pocivavsek,C. A. Fierke,E. J. Toone, Bioorg. Med. Chem. 2004,
2,4067; G. J. Williams,S. Domann,A. Nelson,A. Berry, Proc.
1
Natl. Acad. Sci. USA 2003, 100,3143; D. Franke,C.-C. Hsu,C.-
H. Wong, Methods Enzymol. 2004, 388,224.
[
[
4] F. P. Seebeck,D. Hilvert, J. Am. Chem. Soc. 2003, 125,10158.
5] A. Paiardini,R. Contestabile,S. DꢀAguanno,S. Pascarella,F.
Bossa, Biochim. Biophys. Acta 2003, 1647,214.
[
6] C. Cativiela,M. D. Diaz-de-Villegas, Tetrahedron: Asymmetry
1998, 9,3517.
[
7] U. Schöllkopf, Tetrahedron 1983, 39,2085; U. Schöllkopf,U.
Groth,W. Hartwig, Liebigs Ann. Chem. 1981,2407; U. Schöll-
kopf,U. Groth,K.-O. Westphalen,C.-Z. Deng, Synthesis 1981,
969.
[
8] M. Kataoka,M. Ikemi,T. Morikawa,T. Miyoshi,K. Nishi,M.
Wada,H. Yamada,S. Shimizu, Eur. J. Biochem. 1997, 248,385;
J. Q. Liu,M. Odani,T. Yasuoka,T. Dairi,N. Itoh,M. Kataoka,S.
Shimizu,H. Yamada, Appl. Microbiol. Biotechnol. 2000, 54,44.
9] M. A. Spies,J. J. Woodward,M. R. Watnik,M. D. Toney, J. Am.
Chem. Soc. 2004, 126,7464.
[
[
10] Computational methods: Binding of the four diastereomeric
aldimine adducts of a-methyl-b-phenylserine at the alrY265A
active site was investigated using ab initio and docking methods.
[
14]
Ab initio calculations were performed using GAMESS at the
[
15]
B3LYP/DZV+ (2d,p) level of theory (e = 4), with CHELPG
[
16]
charges and standard radii. Initial docking was performed with
[
17]
QMView, and energy maps were generated by varying the
torsional angles corresponding to rotation around the CaꢀCb
and CbꢀCg bonds of the intermediates in 308 increments.
[
18]
Electrostatic binding energies were calculated with APBS.
[
19]
Following molecular dynamics simulations with AMBER, the
most favorable conformers were optimized with ab initio
methods. See the Supporting Information for details.
[
11] A. Watanabe,T. Yoshimura,B. Mikami,H. Hayashi,H.
Kagamiyama,N. Esaki, J. Biol. Chem. 2002, 277,19166.
12] T. Yamada,J. Komoto,Y. Takata,H. Ogawa,H. C. Pitot,F.
Takusagawa, Biochemistry 2003, 42,12854.
[
Figure 1. Energy-minimizedstructures of (2 R,3S)-a-methyl-b-phenylser-
ine (top) and(2 R,3R)-a-methyl-b-phenylserine (bottom) boundto the
active site of alrY265A as external aldimines with PLP. The b-hydroxy
group of both substrates is within hydrogen-bonding distance of the
phosphate group of PLP. A hydrogen bond between Arg219 and
His166 (dotted line to the protein), which is also present in the parent
[
[
13] K. Tatsumoto,A. E. Martell, J. Am. Chem. Soc. 1978, 100,5549.
14] M. W. Schmidt,K. K. Baldridge,J. A. Boatz,S. T. Elbert,M. S.
Gordon,J. H. Jensen,S. Koseki,N. Matsunaga,K. A. Nguyen,
S. J. Su,T. L. Windus,M. Dupuis,J. A. Montgomery J, r.,
J.
[
20]
Comput. Chem. 1993, 14,1347.
enzyme, dictates the tautomeric form of the imidazole.
[
[
[
[
15] C. T. Lee,W. T. Yang,R. G. Parr, Phys. Rev. B 1988, 37,785.
16] C. M. Breneman,K. B. Wiberg, J. Comput. Chem. 1990, 11,361.
17] K.K. Baldridge,J. P. Greenberg, J. Mol. Graphics 1995, 13,63.
18] N. A. Baker,D. Sept,M. J. Holst,J. A. McCammon, IBM J. Res.
Dev. 2001, 45,427.
Keywords: b-hydroxy amino acids · aldolases · enzyme models ·
stereoselectivity
.
[
19] J. M. Wang,P. Cieplak,P. A. Kollman, J. Comput. Chem. 2000,
21,1049.
[
20] J. P. Shaw,G. A. Petsko,D. Ringe, Biochemistry 1997, 36,1329.
[
1] A. Schmid,J. S. Dordick,B. Hauer,A. Kiener,M. Wubbolts,B.
Witholt, Nature 2001, 409,258; H. E. Schoemaker,D. Mink,
M. G. Wubbolts, Science 2003, 299,1694.
[
2] T. Kimura,V. P. Vassilev,G.-J. Shen,C.-H. Wong, J. Am. Chem.
Soc. 1997, 119,11734; T. D. Machajewski,C.-H. Wong, Angew.
Chem. 2000, 112,1406; Angew. Chem. Int. Ed. 2000, 39,1352; J.-
Q. Liu,T. Dairi,N. Itoh,M. Kataoka,S. Shimizu,H. Yamada,
J.
Mol. Catal. B 2000, 10,107; W. D. Fessner,V. Helaine, Curr.
Opin. Biotechnol. 2001, 12,574; W. A. Greenberg,A. Varvak,
S. R. Hanson,K. Wong,H. J. Huang,P. Chen,M. J. Burk, Proc.
Natl. Acad. Sci. USA 2004, 101,5788; H. J. Lamble,M. J.
Danson,D. W. Hough,S. D. Bull, Chem. Commun. 2005,124.
6
826 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 6824 –6826