Optimization of the kinetic resolution of the DL-phospho-monoesters of threonine and serine by random mutagenesis of the acid phosphatase from Salmonella enterica

Article abstract of DOI:10.1002/adsc.200700041

Acid phosphatases are enzymes with a broad substrate specificity showing hydrolytic activity towards several different organic phosphate monoesters, such as nucleotides and sugar phosphates. The acid phosphatase from Salmonella enterica ser. typhimurium LT2 (PhoN-Se) is able to hydrolyze O-phospho-DL-threonine to yield L-threonine with a very high enantioselectivity (E > 200). When O-phospho-DL-serine was hydrolyzed by PhoN-Se, D-serine was formed, however, the ee values rapidly dropped to 50 %. Random mutagenesis by error-prone PCR was performed on the phosphatase in order to increase its enantioselectivity in the formation of D-serine. Two variants with increased selectivity from a library of 9600 mutants have been found, N151D and V78L showing E values of 18.1 and 4.1, respectively, compared to 3.4 for the wild-type (WT) enzyme.

Full text of DOI:10.1002/adsc.200700041

COMMUNICATIONS
DOI: 10.1002/adsc.200700041
Optimization of the Kinetic Resolution of the dl-Phospho-
monoesters of Threonine and Serine by Random Mutagenesis
of the Acid Phosphatase from Salmonella enterica
Teunie van Herk,a Aloysius F. Hartog,a Harald J. Ruijssenaars,a Richard Kerkman,b
Hans E. Schoemaker,a,c and Ron Wevera,*
a
Van’t Hoff Institute for Molecular Sciences, University of Amsterdam, Nieuwe Achtergracht 129, 1018 WS Amsterdam,
The Netherlands
Phone: (+31)-20-525-5110; fax: (+31)-20-525-5670; e-mail: rwever@science.uva.nl
Dutch States Mines, Anti-Infectives B.V., Delft, The Netherlands
Dutch States Mines, Life Sciences, Advanced Synthesis and Catalysis B.V., Geleen, The Netherlands
b
c
Received: January 23, 2007
Supporting information for this article is available on the WWW under http://asc.wiley-vch.de/home/.
Abstract: Acid phosphatases are enzymes with a
broad substrate specificity showing hydrolytic activ-
ity towards several different organic phosphate
monoesters, such as nucleotides and sugar phos-
phates. The acid phosphatase from Salmonella en-
terica ser. typhimurium LT2 (PhoN-Se) is able to
hydrolyze O-phospho-dl-threonine to yield l-threo-
nine with a very high enantioselectivity (E>200).
When O-phospho-dl-serine was hydrolyzed by
PhoN-Se, d-serine was formed, however, the ee
values rapidly dropped to 50%. Random mutagene-
sis by error-prone PCR was performed on the phos-
phatase in order to increase its enantioselectivity in
the formation of d-serine. Two variants with in-
creased selectivity from a library of 9600 mutants
have been found, N151D and V78L showing E
values of 18.1 and 4.1, respectively, compared to 3.4
for the wild-type (WT) enzyme.
phosphate monoesters. These enzymes are also capa-
ble of transferring a phosphate group from a donor
(phosphomonoester and pyrophosphate) to a wide
range of acceptors (alcohols).[5,6] The acid phospha-
tase from Shigella flexneri (PhoN-Sf)[7] shows enantio-
selectivity in the phosphorylation of d- and l-glucose
by pyrophosphate having a higher affinity for d-glu-
cose.[6] To broaden the application range of phospha-
tases, we investigated the possible use of PhoN-Sf and
PhoN-Se (the acid phosphatase from Salmonella en-
terica ser. typhimurium LT2)[8] in the kinetic resolu-
tion of phosphomonoesters and optimized the hydrol-
ysis reaction by applying random mutagenesis.
Here we report on the formation of l-threonine
and d-serine from O-phosphorylated dl-threonine
and dl-serine by PhoN-Sf and PhoN-Se. PhoN-Se
shows reasonable reaction rates in dephosphorylation
of both O-phospho-dl-threonine and O-phospho-dl-
serine as monitored by HPLC (Figure 1). Therefore
the pH dependency of the enantioselectivity of this
enzyme was investigated in more detail. Hydrolysis
rates are not affected to a large extent between
pH 3.9 and 5.0, but they clearly diminish at pH 6.25
and no activity is found at pH 3.3 (not shown). To our
surprise, only l-threonine was detected within 5 h of
incubation for pH values between 3.9 and 5.0 (~32%
Keywords: acid phosphatase; directed evolution;
enantioselectivity; enzyme catalysis; phosphomo-
noester hydrolysis; random mutagenesis
Enzymes are attractive tools in catalysis and efficient- conversion) and O-phospho-d-threonine was not hy-
ly complement (traditional) chemical methods. Some drolyzed, giving an ee value of 100%. [The detection
enzymes are commercially available[1,2] but the sub- limit of d-threonine in this system is 0.75 mM. In this
strate specificity and enantioselectivity are not always case, no d-threonine detected means that less than
sufficient. Finding an enzyme that shows high stereo- 1 mM of d-threonine is formed. This leads to an ee
selectivity towards a broad range of synthetically value of >97% at 30% conversion after 5 h of reac-
useful molecules is desirable. Bacterial non-specific tion time.] The E value of this reaction, calculated
acid phosphatases (NSAPs)[3,4] are a group of secreted from conversion and ee of the product according to
enzymes with a broad substrate specificity that shows Chen et al.,[9] is >200, which means that this dephos-
hydrolytic activity towards several different organic phorylation reaction can be used for preparative syn-
Adv. Synth. Catal. 2007, 349, 1349 – 1352
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1349

Teunie van Herk et al.
COMMUNICATIONS
Thus, PhoN-Se shows a slight preference for the hy-
drolysis of the d-isomer of serine, but surprisingly a
high preference for the l-isomer of threonine. The
same specificity has been observed for wheat germ
acid phosphatase.[12] Since it is known that the ee of
an enantioselective reaction may strongly depend on
the temperature, the temperature dependency of the
O-phospho-dl-threonine
dephosphorylation
by
PhoN-Se was investigated (T=5–758C). However, no
effect was seen on the enantioselectivity of the reac-
tion.
PhoN-Sf was virtually unable to dephosphorylate
O-phospho-dl-threonine as shown by a very low con-
version of 0.3% after 5.5 h at pH 5.0 (Figure 1). The
observed ee value of 100% in formation of l-threo-
nine is therefore not representative for the whole re-
action. O-phospho-dl-serine conversion by PhoN-Sf
after 24 h at pH 5.0 was 10% and at this low conver-
sion, the ee value had already dropped to 13.7% in
favour of the d-isomer. Considering this result, this
enzyme was not studied further.
To improve the enantioselectivity in hydrolysis of
O-phospho-dl-serine, error prone PCR mutagenesis
was used to create a PhoN-Se mutant library. High-
throughput screening of the mutants was performed
in which the enzymatic hydrolysis of both d- and l-O-
phosphoserine was followed in separate incubations
using identical enzyme lysates and subsequent assay
for inorganic phosphate to identify variant enzymes
with changed stereospecific hydrolysis. After the first
screening of 9600 mutants, 61 variants were selected
and screened with a more sensitive Biomol GreenTM
Pi assay. The three most promising variants were
Figure 1. Time course of the dephosphorylation of 100 mM
O-phospho-dl-threonine (squares) and O-phospho-dl-
serine (circles) by 5 mM PhoN-Sf (closed symbols) and
PhoN-Se (open symbols) at pH 5.0; Panel A: Extent of the
reaction. Panel B: ee values of the reaction products.
thesis of l-threonine. In principle, the remaining O- grown in larger amounts, purified and assayed in
phospho-d-threonine can be hydrolyzed and subjected time-dependent reactions and their stereoselectivity
to a threonine racemase[10,11] and phosphorylated was determined by chiral HPLC measurements. Se-
again for further resolution and higher yields. For O- quencing revealed that two of the variants carried the
phospho-dl-serine, the rate of the pH dependent de- same mutation (N151D) indicating that this site is a
phosphorylation was equal to that seen with O-phos- ꢁhot spotꢂ. The other variant carried a V78L mutation.
pho-dl-threonine. However, after 0.5 h of incubation
The best mutant generated, N151D(1), has a 5.3
the ee values were already significantly lower times increased E value compared to the WT enzyme
(Figure 1) and further slowly decreased to about 50% in the hydrolysis of O-phospho-dl-serine under our
at 24 h. The formation of d-serine was faster com- screening conditions as shown in Table 1 and Figure 2.
pared to its l counterpart. This results in an E value This variant appears to be quite active, reaching a
of 3.7 for the hydrolysis of O-phospho-dl-serine. conversion of 61% compared to 31% for the WT
Table 1. Summary of enantioselectivity of wild-type PhoN-Se and improved variants as catalysts in the formation of d-serine
from O-phospho-dl-serine.
0.5 h
Conversion [%]
3 h
6 h
Enzyme
eep [%]
Conversion [%]
eep [%]
Conversion [%]
eep [%]
E value[a]
WT
8.9
49.6
83.0
86.4
56.1
30.7
61.2
49.6
43.5
48.0
65.8
75.3
48.3
40.0
68.8
58.1
55.4
44.5
55.0
67.5
43.1
3.4Æ0.1
18.1Æ4.5
16.0Æ1.6
4.1Æ0.1
N151D (1)
N151D (2)
V78L
33.8
20.0
17.5
[a]
The selectivity factor E reflects the ratio of the relative rates of reaction of two enantiomers.
1350
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1349 – 1352

Kinetic Resolution of the dl-Phosphomonoesters of Threonine and Serine
COMMUNICATIONS
Figure 2. Conversion vs. ee values for wild type PhoN-Se
and 3 variants in the hydrolysis of O-phospho-dl-serine. Re-
action mixture contained 5 mM of enzyme, 100 mM O-phos-
pho-dl-serine in 100 mM acetate buffer (pH 5.0) at 308C.
Measurements were time dependent starting at 30 min with
30 min intervals. Conversions and ee values were determined
by chiral HPLC.
Figure 3. Model of the active site of WT PhoN-Se based on
the acid phosphatase from E. blattae. Molecular graphics
created with YASARA (www.yasara.org) and PovRay
(www.povray.org).
enzyme after 3 h of incubation. The variant V78 L,
with a slightly improved E value of 4.14 also shows a formational change may change the stereoselectivity.
higher catalytic activity, comparable to the N151D Surprisingly, residue 78 has been described to be in-
mutant. Figure 2 fits well with simulations described volved in the pH dependency of the dephosphoryla-
by Chen et al.[9]
tion of pNPP by the acid phosphatase form Salmonel-
The acid phosphatases form Salmonella enterica la enterica ser. typhymurium MD6001.[13] In this pro-
ser. typhimurium MD6001[13] and Escherichia blat- tein, exchanging the isoleucine 78 for an aspartic acid
tae[14] have been crystallized and their 3D structures resulted in a pH shift towards more basic pH. The au-
have been resolved. Our PhoN-Se shows 94% and thors explain this by an increase in pKa of the H158
38% identity, respectively, and 3D models have been involved in the protonation of the leaving group due
build. The class of NSAPs to which PhoN-Se belongs to electrostatic effects from the D78, which stabilized
share some conserved amino acid residues, which are the protonated form of H158.
divided into three domains.[3] These conserved regions
The effect of the mutations on the hydrolysis of O-
play very important roles in catalysis. They contain phospho-dl-threonine was also tested. The wild type
the active site residues that participate in the binding enzyme already shows a selectivity factor of E>200
of the phosphate, act as a nucleophile, stabilize the in the formation of l-threonine. The variants did not
pentacoordinated transition state and play a role in show significant difference in the stereoselective hy-
leaving group protonation. The N151D variant is situ- drolysis compared to the WT enzyme.
ated closely to the active site. It is changed from an
In conclusion, we investigated the applicability of
asparagine which has an amide group as side chain PhoN-Se in the kinetic resolution of O-phospho-dl-
into an aspartic acid, which has a carboxylic acid threonine and serine. We show that PhoN-Se may be
group as side chain. This changes not only the charge used for the preparative synthesis of enantiomerically
but also the volume of the residue and this may ex- pure l-threonine. The ee value was 100% for l-threo-
plain the observed difference in enantioselectivity as nine up to 60% conversion of O-phospho-l-threo-
compared to the WT enzyme. It can be seen in nine. It is shown that the enantioselectivity for hydrol-
Figure 3 that residue N151 is in close proximity to the ysis of O-phospho-dl-serine by PhoN-Se is much less
residues R130 and K123 which are involved in stabi- and reversed; d-serine is formed preferentially. The
lizing the pentacoordinated transition state. N151 other optical antipodes may be obtained by hydrolyz-
forms two hydrogen bonds with R130 and changing ing the remaining phosphate esters using an unspecif-
this residue into an aspartic acid deletes one of these ic alkaline phosphatase after the resolution and subse-
hydrogen bonds. This may also result in a slightly al- quent separation. We have used a directed evolution
tered active site, changing the stereoselectivity. The approach and obtained two variants with improved
other variant, V78L, also contacts the active site. The stereoselectivity in the hydrolysis of O-phospho-dl-
side chain is in close proximity to H158 which is in- serine. A better E value for the synthesis of d-serine
volved in leaving group protonation and a subtle con- may be obtained by saturation mutagenesis of the
Adv. Synth. Catal. 2007, 349, 1349 – 1352
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1351

Teunie van Herk et al.
COMMUNICATIONS
the amino acids and OPA derivatization. The authors would
like to thank Prof. G. Vriend, University of Nijmegen, for
building the PhoN-Se models. This work was supported by
the Dutch National Research Combination (NRSC-Cataly-
sis).
N151 and V78, or a further round of random muta-
genesis using the mutants found as parent strains.
Experimental Section
General Information
Expression and purification of recombinant Shigella flexneri
PhoN (PhoN-Sf)[5] and Salmonella enterica ser. typhimurium
LT2 PhoN (PhoN-Se)[15] were as described. All chemicals
and other enzymes were purchased from commercial suppli-
ers and used without purification.
References
[1] K. Drauz, H. Waldmann, Enzyme Catalysis in Organic
Synthesis; A Comprehensive Handbook, Vols. I–III, 2nd
edn., VCH, Weinheim, 2002.
[2] K. Faber, Biotransformations in Organic Chemistry, 5th
edn., Springer, Berlin, 2004.
Dephosphorylation O-Phospho-dl-threonine and O-
Phospho-dl-serine
[3] J. Stukey, G. M. Carman, Protein Science 1997, 6, 469–
472.
[4] G. M. Rossolini, S. Schippa, M. L. Riccio, F. Berlutti,
L. E. Macaskie, M. C. Thaller, Cell. Mol. Life Sci. 1998,
54, 833–850.
[5] N. Tanaka, Z. Hasan, A. F. Hartog, T. van Herk, R.
Wever, Org. Biomol. Chem. 2003, 1, 2833.
[6] T. van Herk, A. F. Hartog, A. M. van der Burg, R.
Wever, Adv. Synth. Catal. 2005, 347, 1155–1162.
[7] K. Uchiya, M. Tohsuji, T. Nikai, H. Sugihara, C. Sasa-
kawa , J. Bacteriol. 1996, 178, 4548–4554.
[8] M. Kasahara, A. Nakata, H. Shinagawa, J. Bacteriol.
1991, 173, 6760–6765.
A mixture containing O-phospho-dl-threonine (100 mM) or
O-phospho-dl-serine (100 mM) in acetate buffer (100 mM)
at 308C was completed with either PhoN-Sf (5 mM) or
PhoN-Se (5 mM). pH-dependent dephosphorylation was car-
ried out at pH 3.3, 3.9, 4.3, 5.0 and 6.25. Reaction rates were
determined by time course HPLC measurements of the re-
leased phosphate. Chiral HPLC was used to determine the
ee values of the released amino acids.
Supporting Information
For analytical procedures, random mutagenesis, high-
throughput expression and high-throughput screening
assays, see the Supporting Information.
[9] C.-H. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J.
Am. Chem. Soc. 1982, 104, 7294–7299.
[10] H. Amos, J. Am. Chem. Soc. 1954, 76, 3858.
[11] B. Schnell, K. Faber, W. Kroutil, Adv. Synth. Caral.
2003, 345, 653–666.
Acknowledgements
[12] M. P. Scollar, G. Sigal, A. M. Klibanov, Biotech.
Bioeng. 1985, 247–252.
We gratefully acknowledge Dr. W. Hemrika and Dr. V.
Dumay for providing us the plasmid harbouring the S. enter-
ica ser. typhimurium LT2 PhoN gene and Dr. Z. Hasan for
optimizing the expression system of PhoN-Sf. The help of Dr.
A. L. Duchateau, DSM fine chemicals, Geleen, is greatly ap-
preciated in providing the HPLC method for separation of
[13] R. D. Makde, K. Dikshit, V. Kumar, Biomol. Eng. 2006,
23, 247–251.
[14] K. Ishikawa, Y. Mihara, K. Gondoh, E-I. Suzuki, Y.
Asano, EMBO J. 2000, 19, 2412–2423.
[15] N. Tanaka, V. Dumay, Q. Liao, A. J. Lange, R. Wever,
Eur. J. Biochem. 2002, 269, 2162–2167.
1352
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1349 – 1352