Comparative investigations on thermostable pyrimidine nucleoside phosphorylases from Geobacillus thermoglucosidasius and Thermus thermophilus

Article abstract of DOI:10.1016/j.molcatb.2012.02.006

The recombinant expression and biocatalytic characterization of two thermostable pyrimidine nucleoside phosphorylases (PyNP), isolated from Geobacillus thermoglucosidasius (Gt) and Thermus thermophilus (Tt) is described. Both enzymes are highly thermostable (half life of GtPyNP is 1.6 h at 70 °C, half life of TtPyNP is >24 h at 80 °C). Kinetic parameters for the phosphorolysis of natural substrates were determined for GtPyNP at 60 °C (K_m for uridine 2.3 mM, K_m for thymidine 1.3 mM) and TtPyNP at 80 °C (K_m for uridine 0.15 mM, K_m for thymidine 0.43 mM). The k_cat values for uridine are almost identical for both enzymes (ca. 277 s^-1), while the k_cat value for thymidine is about 8 times higher for TtPyNP than for GtPyNP (679 s^-1 vs. 83 s^-1). Both enzymes were tested towards the ability to catalyze the phosphorolytic cleavage of 2′-fluorosubstituted pyrimidine nucleosides - a prerequisite for the efficient synthesis of a number of relevant purine nucleoside analogues. GtPyNP showed poor activity towards 2′-deoxy-2′-fluorouridine (dUrd_2′F; 0.4% substrate conversion after 30 min), and the phosphorolysis of the epimeric counterpart 1-(2-deoxy-2-fluoro-β-d-arabinofuranosyl)uracil (dUrd^2′F) could not be detected at all. By contrast, TtPyNP showed dramatically higher conversion rates (15.6% and 1.6% conversion in 30 min of both substrates, respectively). The amount of converted pyrimidine nucleosides increased significantly with time. After 17 h 65% of dUrd_2′F and 46% of dUrd^2′F was phosphorolytically cleaved. Our results demonstrate the potential of TtPyNP as a biocatalyst in transglycosylation reactions aiming at the production of 2′-fluorosubstituted purine nucleosides that are highly bioactive but hardly accessible by chemical methods.

Full text of DOI:10.1016/j.molcatb.2012.02.006

Journal of Molecular Catalysis B: Enzymatic 84 (2012) 27–34
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Comparative investigations on thermostable pyrimidine nucleoside
phosphorylases from Geobacillus thermoglucosidasius and Thermus thermophilus
Kathleen Szekera,∗, Xinrui Zhoua, Thomas Schwabb, Ana Casanuevac, Don Cowanc,
Igor A. Mikhailopulod, Peter Neubauera
a Department of Bioprocess Engineering, Institute of Biotechnology, Technische Universität Berlin, Ackerstr. 71-76, ACK24, 13355 Berlin, Germany
b Institute of Biophysics and Physical Biochemistry, University of Regensburg, Universitätsstrasse 31, 93053 Regensburg, Germany
c Institute for Microbial Biotechnology and Metagenomics, University of the Western Cape, Private Bag X17, Bellville 7535, Cape Town, South Africa
d Institute of Bioorganic Chemistry, National Academy of Sciences of Belarus, Acad. Kuprevicha 5/2, 220141 Minsk, Belarus
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 25 February 2012
The recombinant expression and biocatalytic characterization of two thermostable pyrimidine nucleoside
phosphorylases (PyNP), isolated from Geobacillus thermoglucosidasius (Gt) and Thermus thermophilus (Tt)
is described. Both enzymes are highly thermostable (half life of GtPyNP is 1.6 h at 70 ◦C, half life of TtPyNP
is >24 h at 80 ◦C). Kinetic parameters for the phosphorolysis of natural substrates were determined for
GtPyNP at 60 ◦C (Km for uridine 2.3 mM, Km for thymidine 1.3 mM) and TtPyNP at 80 ◦C (Km for uridine
0.15 mM, Km for thymidine 0.43 mM). The kcat values for uridine are almost identical for both enzymes (ca.
277 s−1), while the kcat value for thymidine is about 8 times higher for TtPyNP than for GtPyNP (679 s−1
vs. 83 s−1).
Keywords:
Biocatalysis
Thermostable pyrimidine nucleoside
phosphorylase
Pyrimidine nucleoside
2ꢀ-Deoxyfluoro nucleosides
Thermus thermophilus
Geobacillus thermoglucosidasius
Both enzymes were tested towards the ability to catalyze the phosphorolytic cleavage of 2ꢀ-
fluorosubstituted pyrimidine nucleosides – a prerequisite for the efficient synthesis of a number of
relevant purine nucleoside analogues. GtPyNP showed poor activity towards 2ꢀ-deoxy-2ꢀ-fluorouridine
(dUrd2ꢀF; 0.4% substrate conversion after 30 min), and the phosphorolysis of the epimeric counterpart 1-
(2-deoxy-2-fluoro-␤-d-arabinofuranosyl)uracil (dUrd2ꢀF) could not be detected at all. By contrast, TtPyNP
showed dramatically higher conversion rates (15.6% and 1.6% conversion in 30 min of both substrates,
respectively). The amount of converted pyrimidine nucleosides increased significantly with time. After
17 h 65% of dUrd2ꢀF
and 46% of dUrd2ꢀF was phosphorolytically cleaved.
Our results demonstrate the potential of TtPyNP as a biocatalyst in transglycosylation reactions aiming
at the production of 2ꢀ-fluorosubstituted purine nucleosides that are highly bioactive but hardly accessible
by chemical methods.
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
of the glycosidic bond of pyrimidine nucleosides or closely related
derivatives thereof according to the following general reaction
Pyrimidine nucleoside phosphorylases (PyNP; EC 2.4.2.2) are
homodimeric enzymes that are involved in the nucleotide salvage
pathway of some lower organisms [1–3]. In the presence of phos-
phate ions, PyNP catalyzes the reversible phosphorolytic cleavage
scheme (Scheme 1).
Structurally, PyNP is closely related to human thymidine phos-
phorylases (TP; EC 2.4.2.4) – an angiogenic enzyme that was found
to be identical to the platelet-derived endothelial cell growth fac-
tor in humans [4,5]. Together, TP and PyNP form the nucleoside
phosphorylase-II family, which share a common two-domain sub-
unit fold and a high level of sequence identity [3]. Despite the
similarity of the reaction catalyzed, uridine phosphorylase (UP;
EC 2.4.2.3) belongs to the phosphorylase-I family with distinct
structural characteristics. From the catalytical point of view, TP is
distinguished from UP due to its high specificity for the 2ꢀ-deoxy-
d-ribofuranose moiety of pyrimidine nucleosides [3]. By contrast,
PyNP does not discriminate between uridine (1) and thymidine (2)
and accepts both compounds as natural substrates [6].
Abbreviations: PyNP, pyrimidine nucleoside phosphorylase; PNP, purine
nucleoside phosphorylase; NP, nucleoside phosphorylase; TP, thymidine phos-
phorylase; UP, uridine phosphorylase; Tt, Thermus thermophilus; Gt, Geobacillus
ꢀ
ꢀ
thermoglucosidasius; dUrd2 F
, , 1-(2-deoxy-2-
2ꢀ-deoxy-2ꢀ-fluorouridine; dUrd2 F
fluoro-␤-d-arabinofuranosyl)uracil.
∗
Corresponding author. Tel.: +49 30 314 72521; fax: +49 30 314 27577.
E-mail addresses: k.szeker@mailbox.tu-berlin.de (K. Szeker),
xinrui.zhou@gmail.com (X. Zhou), thomas.schwab@biologie.uni-regensburg.de (T.
Schwab), acasanueva@uwc.ac.za (A. Casanueva), dcowan@uwc.ac.za (D. Cowan),
igormikh@iboch.bas-net.by (I.A. Mikhailopulo), peter.neubauer@tu-berlin.de
(P. Neubauer).
This rather broad substrate specificity makes PyNP a versatile
biocatalyst suitable for certain synthetic applications, e.g. for the
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2012.02.006

28
K. Szeker et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 27–34
thermophilus [17]. With respect to practical use as biocatalyst, only
the G. stearothermophilus enzyme has been described [18].
In the present study, we report on the recombinant expres-
sion and characterization of two additional thermostable PyNPs,
derived from G. thermoglucosidasius 11955 (GtPyNP) and T.
thermophilus HB27 (TtPyNP), respectively. Albeit the expres-
sion and crystallization of PyNP from T. thermophilus HB8 has
been reported [17], information about the biocatalytic char-
acterization is not available. Here, we investigate thermal
characteristics and kinetic parameters as well as the ability to
phosphorolyze 2ꢀ-fluorosubstituted pyrimidine nucleosides, viz.,
Scheme 1.
2ꢀ-deoxy-2ꢀ-fluorouridine (dUrd2 F) and 1-(2-deoxy-2-fluoro-␤-d-
ꢀ
ꢀ
arabinofuranosyl)uracil (dUrd2 F) (Scheme 2), which are of special
enzymatic synthesis of nucleosides. Of particular interest is the
synthesis of modified nucleosides, that can be used as pharmaceu-
tical agents for the treatment of viral infections and cancer, as well
as in molecular biological techniques and diagnostics [7–10]. For
their synthesis, efficient chemo-enzymatic approaches have been
developed, among them the enzymatic transglycosylation of chem-
ically modified nucleosides [11]. Scheme 2 illustrates the enzymatic
synthesis of modified purine nucleosides as an example: by the
concerted action of a PyNP and a purine nucleoside phosphory-
lase (PNP; EC 2.4.2.1), a pentofuranosyl moiety is transferred from
a pyrimidine nucleoside that serves as pentofuranosyl donor to a
heterocyclic purine base that serves as pentofuranosyl acceptor. It
is obvious that the lack of enzymes with broad substrate specifici-
ties is limiting the application spectrum of this strategy. In addition,
the efficiency of such processes, when run at high temperatures, is
often reduced due to the thermal lability of the biocatalysts.
The search for enzymes with improved properties is hence of
prime importance in this field of research and prompted us to
investigate biocatalytical characteristics of novel nucleoside phos-
phorylases. In contrast to PNPs that have been extensively studied
from a multitude of mesophilic and thermophilic (micro) organ-
isms [12–15], PyNPs have been hardly studied in detail. Only few
examples can be found in the scientific literature: PyNP from Bacil-
lus subtilis [16], Geobacillus stearothermophilus [1,2,6] and from T.
interest, as sugar donors, for purine nucleoside synthesis. Our data
reveal striking differences of the conversion rates depending on the
substrate analogue and the biocatalyst applied.
2. Experimental
2.1. Molecular biology
G. thermoglucosidasius 11955 was grown for 1.5 days at 52 ◦C
in Luria Broth (10 g l−1 tryptone, 5 g l−1 yeast extract, 10 g l−1
NaCl, pH 7.0), whereas T. thermophilus HB27 was grown as
previously described [19]. Genomic DNA of both strains was iso-
lated using standard protocols [20]. The genes coding for PyNP
in G. thermoglucosidasius (GenBank accession no. ZP 06809030)
and T. thermophilus (GenBank accession number AAS81754.1)
were amplified using Pfu DNA polymerase (Fermentas, Lithua-
nia). The following primer pair was used for the isolation of
the PyNP gene from the G. thermoglucosidasius genome: 5ꢀ
ACTAGGGATCCATGGTCGATTTAATTGCGA 3ꢀ (BamHI site under-
lined) and 5ꢀ AGCATGCGGCCGCTTATGAAATGGTTTCGTATATA 3ꢀ
(NotI site underlined). The primer pair used for the T. thermophilus
gene was: 5ꢀ ACTAGGGATCCAACCCCGTGGTCTTCATC 3ꢀ (BamHI
Scheme 2.

K. Szeker et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 27–34
29
site underlined) and 5ꢀ AGCATGCGGCCGCCTAGATGGCCTCCAGGA
3ꢀ (NotI site underlined). The PyNP encoding genes were cloned via
BamHI/NotI digestion (FastDigest restriction endonucleases, Fer-
mentas, Lithuania) and subsequent ligation (T4 DNA Ligase, Roche)
into a derivative of the expression vector pCTUT7 [21]. This vector
is characterized by an IPTG inducible lac promoter derivative and a
pBR322 origin of replication. We have modified this vector before
by substituting the chloramphenicol resistance cassette by the plas-
mid stabilizing parB locus [22] and an ampicillin resistance cassette.
In addition we introduced a sequence encoding a hexahistidine tag
connected to the 5ꢀ-end of target gene’s transcript. The resulting
expression vectors were transformed into the Escherichia coli BL21
strain (Novagen).
Protein unfolding: Thermal denaturation of 10 ␮M purified
TtPyNP protein dissolved in potassium phosphate (50 mM, pH 7.5)
was monitored with a Jasco J-815 circular dichroism (CD) spec-
trometer in a 0.1 cm cuvette by following the loss of ellipticity
at 220 nm. Unfolding was induced by raising the temperature in
0.1 ◦C increments at a ramp rate of 1 ◦C min−1 with a Peltier-effect
temperature controller. The measured ellipticity was normalized,
and the apparent melting temperature (TMapp) was determined. DSC
experiments were performed with 43 ␮M TtPyNP protein dissolved
in potassium phosphate (50 mM, pH 7.5) by heating the samples in
a CSC 5100 Nano differential scanning calorimeter with a scan rate
of 1 ◦C min−1. The DSC data were analyzed with the program CpCalc
(version 2.1, Calorimetry Sciences Corporation, 1995) to determine
TMapp at which half of the protein is unfolded. The irreversibility of
thermal denaturation precluded thermodynamic analysis of the CD
and DSC unfolding traces.
2.2. Bioinformatics and homology modeling
Amino acid sequence identities were assessed with the protein
basic local alignment tool (BLAST) [23] of the NCBI web server. For
multiple sequence alignments the ClustalW2 program from the EBI
web server was used [24]. Three-dimensional protein structures
of the target proteins were built by homology modeling using the
Swiss-model workspace [25]. These 3D models were superposed
and visualized with the CCP 4 mg molecular-graphics software [26].
2.6. Activity assay
If not otherwise stated the activity assay was performed in
50 mM potassium phosphate buffer, pH 7.0 containing 1 mM of
uridine as substrate. After 2 min of pre-heating, 1–2 ␮l of diluted
enzyme was added per 100 ␮l of reaction mixture. Samples were
withdrawn after defined time intervals and the reaction was imme-
diately stopped by adding 1 volume of reaction mixture to ½
volume of 10% trichloroacetic acid. Precipitated proteins were
removed by centrifugation (20,000 × g, 15 min) and the samples
were stored at −20 ◦C for later analysis by HPLC.
2.3. Overexpression in E. coli
Recombinant E. coli strains were cultivated in terrific broth
(TB) medium [27] or EnPresso® medium (BioSilta, Finland) using
Ultra Yield FlasksTM and AirOTopTM seals (Thomson Instrument
Company, USA). Recombinant protein expression was induced by
addition of IPTG to a final concentration of 100 ␮M, after 2.5 h cul-
tivation time (TB medium) or after overnight cultivation (EnPresso
medium). Cells were harvested (centrifugation at 16,000 × g, 5 min,
4 ◦C) 3 h after induction (TB medium) or 24 h after induction
(EnPresso medium) and stored at −20 ◦C.
From the HPLC results the concentration of residual pyrimidine
nucleoside (substrate) and liberated pyrimidine base (product of
the phosphorolytic cleavage) was retrieved. The substrate conver-
sion was calculated with the following formula:
base
Conversion (%) =
× 100
base + nucleoside
Only conversion rates that were linear with respect to time and
amount of PyNP added, were considered for further analysis. This
was usually the case if not more than 10% of the substrate was
converted.
2.4. Cell disruption and purification
Cell pellets were re-suspended in NPI-10 binding buffer (50 mM
NaH2PO4, 300 mM NaCl, 10 mM imidazole, pH 8.0) at 5 ml/g wet
weight. This cell suspension was sonicated on ice using a UP200S
sonicator (Hielscher Ultrasonics GmbH). The sonication was per-
formed twice with 30% power input for 3 min in 30 s intervals using
a sonotrode of 7 mm in diameter.
2ꢀ-Deoxy-2ꢀ-fluorouridine
was
purchased
from
TCI
Deutschland (Eschborn, Germany) and 1-(2-deoxy-2-fluoro-
␤-d-arabinofuranosyl)uracil obtained from Metkinen Chemistry
(Kuusisto, Finland). Other chemical reagents were purchased from
Sigma–Aldrich.
Cell lysates were centrifuged (20,000 × g, 15 min, 4 ◦C) to sepa-
rate soluble from insoluble fraction. The soluble portion of the cell
lysate was heated for 15 min at 60 ◦C (GtPyNP) or 80 ◦C (TtPyNP).
Coagulated proteins were removed by centrifugation (20,000 × g,
15 min, 4 ◦C). The target proteins were further purified by immo-
bilized metal ion affinity chromatography using a 5 ml Ni-NTA
Superflow cartridge (Qiagen), subsequently the excess of salt and
imidazole was removed by the use of a HiPrep 26/10 Desalting col-
umn (GE Healthcare). The purified protein solution was aliquoted,
rapidly frozen in liquid nitrogen and stored at −80 ◦C.
2.6.1. Thermal stability
Enzyme preparations were diluted 1:100 in 50 mM potassium
phosphate buffer (pH 7.0) yielding enzyme concentrations of about
46 ␮g ml−1. These diluted enzyme solutions were aliquoted and
incubated in 0.2 ml tubes in a thermocycler at the respective tem-
peratures. After defined time intervals, tubes were withdrawn and
the residual activity of the incubated enzyme solution was deter-
mined (at 60 ◦C for GtPyNP and at 80 ◦C for TtPyNP) and plotted over
the reaction time. In order to determine the half life, the resulting
curve was fitted (Sigma Plot) to the decay function:
2.5. Protein analytics
v = v0 · ek ·t
i
The protein concentration of the purified protein solution was
determined by measuring the absorption at 280 nm (Nanodrop,
Thermo Scientific). The absorption coefficients were theoretically
calculated from the amino acid sequence (Vector NTI software,
Invitrogen) and were 0.42 AU (GtPyNP) and 0.56 AU (TtPyNP) for
and subsequently the half life was calculated:
t1/2 = ln 0.5 · ki−1
a solution of 1 mg ml−1
.
2.6.2. Temperature optimum
The molecular weight marker for SDS-PAGE analysis was pur-
chased from Fermentas (Lithuania).
The reaction mixture was pre-heated for 2 min at respective
temperatures. The diluted enzyme solution was added and the

30
K. Szeker et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 27–34
reaction was stopped after 3 min. 50 mM potassium phosphate
buffer (pH 7.5) was used.
1-(␤-d-Arabinofuranosyl)uracil
Sigma–Aldrich,
was
purchased
from
O2,2ꢀ-anhydro-1-(␤-d-arabinofuranosyl)uracil
was synthesized as described in [29].
2.6.3. Kinetic parameters
Activity tests were performed in triplicates for at least 5
different substrate concentrations spanning 0.25–5 times of
the Michaelis–Menton constant. The initial reaction rates were
plotted over the substrate concentrations and fitted to the
Michaelis–Menton function (Sigma Plot). From the resulting for-
mula Km and vmax were directly retrieved, and kcat could be easily
calculated. For details see [28].
3. Results and discussion
3.1. Sequence analysis and homology modeling
The proteins recombinantly expressed and characterized in
this work are the pyrimidine nucleoside phosphorylases from T.
thermophilus HB27 (GenBank accession number AAS81754.1) and
from G. thermoglucosidasius 11955. The sequence of the latter
corresponded to the PyNP sequence from G. thermoglucosidasius
C56-YS93 (GenBank accession number ZP 06809030), with the
exception of one amino acid. Since the corresponding deviation
of the DNA sequence was suspected to be the result of the PCR
performed for cloning, the respective codon was changed by site-
directed mutagenesis to fit the data bank protein (Gln214 changed
to Arg).
2.6.4. Phosphorolysis of modified nucleosides
Reactions were performed in 1 mM pyrimidine nucleoside sub-
strate solution containing 10 mM sodium phosphate buffer (pH 6.5)
and an enzyme loading of ca. 0.1 mg ml−1 at either 60 ◦C or 80 ◦C.
2.6.5. HPLC analysis
The amount of pyrimidine nucleosides and corresponding
bases was determined by following the absorption at 260 nm
Three entries of solved crystal structures of PyNPs can be
found in the protein database (www.pdb.org). Both target enzyme
sequences were blasted against the amino acid sequences belong-
ing to these pdb entries. GtPyNP aligned best with the PyNP from
G. stearothermophilus ATCC 12980 (assigned here as GsPyNP, pdb
entry 1BRW) with 78% sequence identity. TtPyNP aligned best with
the PyNP from T. thermophilus HB8 (pdb entry 2DSJ), to which it is
almost identical (approx. 98% sequence identity) but showed also
a high degree of sequence identity (approx. 50%) to GsPyNP. The
amino acid sequence alignment of GsPyNP, GtPyNP, TtPyNP and
E. coli thymidine phosphorylase (EcTP) is shown in Fig. 1. Amino
during HPLC analysis using
a reversed phase C18 column
(Gemini-Nx 5u, 150 × 4.60 mm, Phenomenex) with the following
gradient: from 97% 20 mM ammonium acetate and 3% acetoni-
trile to 60% 20 mM ammonium acetate and 40% acetonitrile in
10 min. Under these conditions the following retention times
were determined: uridine (3.2 min), uracil (2.4 min), thymidine
ꢀ
(4.7 min), thymine (4.0 min), dUrd2 F (4.6 min), dUrd2 F (4.4 min),
ꢀ
O2,2ꢀ-anhydro-1-(␤-d-arabinofuranosyl)uracil (2.3 min), 1-(␤-d-
arabinofuranosyl)uracil (3.8 min).
Fig. 1. 1Multiple sequence alignment of PyNP from G. stearothermophilus (GsPyNP), G. thermoglucosidasius (GtPyNP), T. thermophilus (TtPyNP) and TP from E. coli (EcTP).
Shading represents the degree of sequence identity. Residues of the active site pocket are highlighted.

K. Szeker et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 27–34
31
Fig. 2. 3D structure models. The superposition of GtPyNP (purple), TtPyNP (yellow)
and GsPyNP (grey) 3D structures are shown. Structural models of GtPyNP and TtPyNP
were built based on homology modeling using the structure of GsPyNP as template.
(For interpretation of the references to color in this figure legend, the reader is
referred to the web version of the article.)
Fig. 3. SDS-PAGE of GtPyNP and TtPyNP samples. Recombinant expression of
GtPyNP under standard conditions (37 ◦C, TB medium) was very efficient (4), but
the expression of TtPyNP under the same conditions rather unsatisfactory (5).
The expression level of TtPyNP could be significantly increased by using EnPresso
medium that provides fed-batch like expression conditions (6). Purified GtPyNP (2)
and TtPyNP (3) showed bands corresponding to the theoretically calculated molecu-
lar weight of the subunits (47.6 kDa and 46.9 kDa, respectively). Lane (1) represents
the molecular weight marker.
acids that have been described to be involved in substrate binding
or in the catalytic mechanism, respectively, are indicated [30,31].
The chain B of GsPyNP (pdb code 1BRW) was used as template
to model both TtPyNP and GtPyNP. The structure of GsPyNP was
elucidated in its closed conformation [30], uracil and the phosphate
ion can be seen in the active site pocket. The structural fold of the
models of GtPyNP and TtPyNP are almost identical to the template
structure GsPyNP, which can be seen in the superposition of the
secondary structure elements (Fig. 2).
3.3. Thermal characteristics
The effect of temperature on the activity of the enzymes, as well
as their thermal stability was investigated.
GtPyNP showed a temperature optimum of 60 ◦C, while the rel-
ative activity of TtPyNP increased with the reaction temperature,
up to the highest temperature tested (99 ◦C, Fig. 4A). An apparent
melting temperature of ≥102 ◦C and 103 ◦C was determined by cir-
cular dichroism and differential scanning calorimetry, respectively
(Fig. 5). Hence, we assume that the temperature optimum of TtPyNP
is in the range of 95–103 ◦C.
3.2. Expression and purification
The desired sequences were expressed with an artificial N-
terminal hexahistidine tag. In addition to
a simple protein
The stability half life of GtPyNP was estimated to be 1.6 h at
70 ◦C; while at 60 ◦C no significant loss of activity could be seen
within 16 h of incubation. The stability half life of TtPyNP at 80 ◦C
exceeds 23 h (Fig. 4B).
Compared to other reported enzymes with pyrimidine nucle-
oside phosphorylase activity, including UPs and TPs, the thermal
stability and the temperature optimum of TtPyNP are extremely
high (Table 1).
purification procedure this method offers the possibility to avoid
expression problems associated with secondary structure forma-
tion of the 5ꢀ-end of mRNA, which is a critical issue, especially for
genes derived from thermophiles [32]. Initially, the recombinant
expression in E. coli of both GtPyNP and TtPyNP was performed
under standard conditions (37 ◦C) in terrific broth (TB) medium.
While the expression level of GtPyNP was satisfactory from the
beginning (Fig. 3, lane 4), TtPyNP was expressed only very poorly
(Fig. 3, lane 5). Hence, we tested also EnPresso medium for TtPyNP
expression. This medium is based on an enzyme controlled sub-
strate delivery and was reported to lead to higher cell densities
and productivity per cell [33]. Indeed, the expression level of
TtPyNP increased significantly (Fig. 3, lane 6) and the final cell
density doubled (from OD600 = 11 to OD600 = 22). The volumet-
ric yield obtained with this condition for TtPyNP expression was
0.06 mg ml−1. Expression of GtPyNP in TB medium resulted in a
volumetric yield of 0.17 mg ml−1. Applied to SDS-PAGE gels, the
purified proteins showed bands representing molecular weights
that are consistent with the theoretically calculated values of the
monomers (Fig. 3, lanes 2 and 3).
3.4. Kinetic parameters
Kinetic parameters describing the biocatalytical properties were
determined at 60 ◦C for GtPyNP and at 80 ◦C for TtPyNP. The results
are summarized in Table 2. The estimated Km values of TtPyNP
for the natural substrates uridine and thymidine are slightly lower
than the Km values determined by Hori et al. [1] for GsPyNP. In con-
trast to TtPyNP, GtPyNP is characterized by extremely low substrate
affinities (high Km values) towards both natural substrates.
The catalytic efficiency of an enzyme is best described by
the ratio of kcat/Km [28]. These ratios are 15 times (substrate

32
K. Szeker et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 27–34
Table 1
Thermal characteristics of reported PyNP, UP and TP.
Enzyme
Organism
Thermal stability (t1/2
)
Temp. optimum
–
Reference
TP
UP
E. coli
E. coli
<10 min (55 ◦C)
9.9 h (60 ◦C)
1.6 h (70 ◦C)
3.3 h (70 ◦C)
1 week (60 ◦C)
25 min (70 ◦C)
–
[34]
[35]
This study
[35]
[36]
[1,2]
[37]
This study
40 ◦
60 ◦
60 ◦
65 ◦
70 ◦
70 ◦
>95 ◦
C
C
C
C
C
C
C
PyNP
UPengineered
UP
PyNP
UP
G. thermoglucosidasius
E. coli
Enterobacter aerogenes
G. stearothermophilus
Erwinia carotovora
T. thermophilus
PyNP
>24 h (80 ◦C)
Fig. 4. Thermal characteristics. Relative activity of GtPyNP and TtPyNP over the
reaction temperature (A), where the highest reaction rate determined was set to
100% for each enzyme. Thermostability (B) was investigated by incubating protein
samples for defined time intervals and subsequently determining the residual activ-
ity, where the activity of enzyme samples that were not thermo-treated was set to
100%.
Fig. 5. Thermal unfolding of TtPyNP. Thermal unfolding trace monitored by loss
of the far-UV circular dichroism signal at 220 nm provided an apparent melting
temperature of at least 102 ◦C (A). The apparent melting temperature determined
by differential scanning calorimetry (B) was 103 ◦C.
uridine) and 25 times (substrate thymidine) higher for TtPyNP than
for GtPyNP. The kcat/Km ratio is also a measure to compare an
enzyme’s specificity towards different substrates [28]. The results
of the present study show that both enzymes are more specific for
uridine than for thymidine, but the difference in specificity is more
pronounced for GtPyNP: the kcat/Km ratio is 2 fold higher for uridine
than for thymidine. In contrast, the kcat/Km ratio of TtPyNP is only
1.23 fold higher for uridine vs that of thymidine.
Table 2
Kinetic parameters of reported PyNP in comparison to E. coli enzymes TP and UP.
Enzyme
Km (␮M)
kcat (s−1
)
kcat/Km (s−1 ␮M−1
)
React. conditions
Reference
Uridine
Thymidine
Uridine
Thymidine
Uridine
Thymidine
GtPyNP
TtPyNP
GsPyNP
EcTP
2342
1282
435
460
300
270
275
279
–
83
679
–
198
5
0.12
1.92
–
0.06
1.56
–
0.66
0.02
60 ◦C, pH 7.0
80 ◦C, pH 7.0
60 ◦C, pH 7.0
25 ◦C, pH 6.5
25 ◦C, pH 7.5
This study
This study
[1]
[38,39]
[38,40]
145
190
60
<1 × 10−4
<1.7 × 10−6
1.22
EcUP
80
98
The estimation errors of the Km and kcat values determined in this study were not higher than 10%. (–) Data not indicated.

K. Szeker et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 27–34
33
Fig. 6. Phosphorolysis of natural and modified pyrimidine nucleosides. (A) The percentage of natural and 2ꢀ-fluorosubstituted pyrimidine nucleosides that were phospho-
rolytically cleaved by TtPyNP or GtPyNP after 30 min, (B) percentage of 2ꢀ-fluorosubstituted pyrimidine nucleosides phosphorolytically cleaved by TtPyNP after prolonged
ꢀ
reaction times at indicated temperatures and (C) percentage of dUrd2 F molecules that reacted to anhydro-Urd and ara-U during the TtPyNP catalyzed phosphorolytic cleavage
reaction in dependence of reaction time and temperature. The ara-U formation at 60 ◦C could not be accurately determined but was estimated not to be higher than 0.4%.
In synthetic applications, where high substrate concentrations
are used (cs ꢁ Km) the kcat value may be the most appropriate
parameter describing the efficiency of the biocatalyst. The turnover
numbers (kcat) of GtPyNP and TtPyNP are in similar range, with uri-
dine as substrate. By contrast, the turnover numbers for thymidine
differ significantly, in favor of thymidine phosphorolysis by TtPyNP.
The kcat value of TtPyNP for thymidine is also unusually high in com-
parison to alternative enzymes that are used for the phosphorolysis
of pyrimidine nucleosides, e.g. EcUP and EcTP (Table 2).
ꢀ
Indeed, it was reported that (i) dUrd2 F showed no detectable
substrate activity towards EcUP, (ii) EcTP catalyzed the phosphorol-
ꢀ
ysis of dUrd2 F but at an extremely low raꢀte, and (iii) the enzymatic
cleavage of the glycosidic bond of dUrd2 F equally afforded a high
amount of enzyme and prolonged reaction time (6 days) [42,46].
In this study we have investigated the phosphorolysis of these
challenging substrates by GtPyNP and TtPyNP (Fig. 6). Our results
indicate that TtPyNP might be a good alternative to the use of E. coli
enzymes, but the use of GtPyNP is apparently not suitaꢀble for the
applications discussed above: no activity towards dUrd2 F and only
3.5. Phosphorolysis of 2ꢀ-fluoro substituted pyrimidine
nucleosides
ꢀ
poor activity towards dUrd2 F (0.44% substrate conversion) was
detected with GtPyNP as biocatalyst after 30 min reaction time.
By contrast, the TtPyNP catalyzed reaction under the same
conditionꢀs resulted in the phosphorolytic cleavage of 0.65%
ꢀ
. Since the optimal reaction
Of particular interest is the potential of both thermostable
PyNPs as biocatalysts in the synthesis of modified nucleosides.
With this aim in view, the phosphorolysis of natural pyrimidine
nucleoside substrates (thymidine and uridine) and their sugar
of dUrd2 F and 7.0% of dUrd2 F
temperature of TtPyNP is significantly higher than 60 ◦C (see Sec-
tion 3.3), we repeated the same reaction also at 80 ◦C. Now,
modified analogues, viz., 2ꢀ-deoxy-2ꢀ-fluorouridine (dUrd2 F) and
the TtPyNP catalyzed reaction resulted in 1.4% phosphorolyzed
ꢀ
ꢀ
ꢀ
dUrd2 F and 15.6% of dUrd2 F
.
However, under these condi-
1-(2-deoxy-2-fluoro-␤-d-arabinofuranosyl)uracil (dUrd2 F), were
ꢀ
tions the formation of two new peaks were observed by HPLC
investigated. These substrates can be used as pentofuranosyl
donors in enzymatic transglycosylations aiming at the synthesis of
pharmaceuticallyvaluable 2ꢀ-fluorosubstitutedpurine nucleosides.
ꢀ
analysis of the reaction mixture that contained dUrd2 F. The
retention times coincide with those of authentic samples of O2,2ꢀ-
anhydro-1-(␤-d-arabinofuranosyl)uracil (anhydro-Urd) (2.3 min)
and 1-(␤-d-arabinofuranosyl)uracil (ara-U) (3.8 min). Hence, the
formation of anhydro-Urd resulting from HF release from the
ꢀ
With this strategy dUrd2 F served as a substrate for the enzymatic
synthesis of 2ꢀ-deoxy-2ꢀ-fluoroguanosine using the whole E. coli
cells as a biocatalyst [41] and a multitude of other purine 2ꢀ-deoxy-
2ꢀ-fluororibosides with antiviral activity using a combination of
ꢀ
dUrd2 F molecule and the subsequent hydrolysis of anhydro-
Urd resulting in ara-U appear to be a reasonable explanation
(Scheme 3). Phosphorolysis of both 2ꢀ-fluorosubstituted pyrimi-
dine nucleosides catalyzed by TtPyNP was also monitored over
EcTP and EcPNP as a biocatalyst [42].
ꢀ
However, dUrd2 F and dUrd2 F are very poor substrates in phos-
ꢀ
phorolysis reactions. This is presumably a result of increased
strength of the glycosyl bond as it follows from the crystallographic
prolonged reaction times (Fig. 6B). The results show that the con-
ꢀ
version of dUrd2 F at 80 ◦C could be increased to 46% after 17 h;
data for the N1–C1ꢀ bond length of uridine (aver. value 1.490 A
˚
ꢀ
˚
˚
side-product formation was not observed. The final amount of
[43]) and its 2 -deoxyfluoro analogues (1.454 A [44] and 1.460 A
[45], respectively). Moreover, introduction of a fluorine atom into
pentofuranose ring of nucleosides results in dramatic changes of
the conformation of such analogues precluding the formation of the
productive substrate-catalytic center complex (for more detailed
discussion, see [11]).
phosphorolyzed dUrd2 F after 17 h at 80 ◦C (65%) was in similar
ꢀ
range as the amount obtained at 60◦ (60%). By contrast, side prod-
uct formation, as discussed above, at 80 ◦C was significantly higher
than at 60 ◦C (Fig. 6C). After 17 h 8.3% of dUrd2 F reacted to anhydro-
ꢀ
Urd at 80 ◦C, whereas only 1.2% of dUrd2 F reacted to anhydro-Urd
ꢀ
at 60 ◦C.
4. Conclusion
PyNPs isolated from thermophilic microorganisms are promis-
ing biocatalysts for the efficient synthesis of modified nucleosides
[18]. Up to now, the biocatalytic characterization of PyNPs from
thermophilic microbes was restricted to PyNPs from G. stearother-
mophilus strains [1,2,6]. We have studied here the biocatalytical
properties of two additional thermostable PyNP, originating from
G. thermoglucosidasius and T. thermophilus.
Scheme 3.

34
K. Szeker et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 27–34
Our results indicate that both enzymes show excellent bio-
[13] A. Bzowska, E. Kulikowska, D. Shugar, Pharmacol. Ther. 88 (2000) 349–425.
[14] G. Cacciapuoti, S. Gorassini, M.F. Mazzeo, R.A. Siciliano, V. Carbone, V. Zappia,
M. Porcelli, FEBS J. 274 (2007) 2482–2495.
[15] D. Visser, F. Hennessy, K. Rashamuse, M. Louw, D. Brady, Extremophiles 14
(2010) 185–192.
[16] X.F. Gao, X.R. Huang, C.C. Sun, J. Struct. Biol. 154 (2006) 20–26.
[17] K. Shimizu, N. Kunishima, Acta Crystallogr. F 63 (2007) 308–310.
[18] S.A. Taran, K.N. Verevkina, S.A. Feofanov, A.I. Miroshnikov, Russ. J. Bioorg. Chem.
35 (2009) 739–745.
catalytical properties for applications with natural pyrimidine
nucleosides as substrates (thymidine, uridine) and a reaction tem-
perature of 60 ◦C. In addition, the unusually high thermal stability of
TtPyNP makes this biocatalyst also suitable for reactions requiring
an even higher reaction temperature of 80 ◦C.
We have further tested both thermostable PyNPs towards their
ability to phosphorolyze 2ꢀ-fluorosubstituted pyrimidine nucle-
osides that have been shown to be very poor substrates in
phosphorolysis reactions employing EcUP or EcTP as biocatalyst
[42,46]. Our results reveal striking differences of the substrate
specificities of GtPyNP and TtPyNP, in favor of the latter. These
findings make TtPyNP a candidate as powerful biocatalyst in the
transglycosylation reactions aiming at the synthesis of 2ꢀ-fluoro
substituted purine nucleosides. Expression of appropriate PNPs and
their use in synthetic applications are ongoing.
[19] T. Schwab, R. Sterner, Chembiochem 12 (2011) 1581–1588.
[20] D. Moore, D. Dowhan, in: F.M. Ausubel, et al. (Eds.), Purification and Concentra-
tion of DNA from Aqueous Solutions, John Wiley & Sons, Houston, 2002, Unit
2.1A.
ˇ
[21] J. Siurkus, J. Panula-Perälä, U. Horn, M. Kraft, R. Rimsˇeliene, P. Neubauer, Microb.
Cell Fact. 9 (2010) 35.
[22] K. Gerdes, Biotechnology 6 (1988) 1402–1405.
[23] S.F. Altschul, W. Gish, W. Miller, E.W. Myers, D.J. Lipman, J. Mol. Biol. 215 (1990)
403–410.
[24] M.A. Larkin, G. Blackshields, N.P. Brown, R. Chenna, P.A. McGettigan, H.
McWilliam, F. Valentin, I.M. Wallace, A. Wilm, R. Lopez, J.D. Thompson, T.J.
Gibson, D.G. Higgins, Bioinformatics 23 (2007) 2947–2948.
[25] L. Bordoli, F. Kiefer, K. Arnold, P. Benkert, J. Battey, T. Schwede, Nat. Protoc. 4
(2009) 1–13.
Acknowledgments
[26] S. McNicholas, E. Potterton, K.S. Wilson, M.E.M. Noble, Acta Crystallogr. D 67
(2011) 386–394.
This work is part of the Cluster of Excellence “Unifying Concepts
in Catalysis” coordinated by the Technische Universität Berlin.
Financial support by the Deutsche Forschungsgemeinschaft (DFG)
within the framework of the German Initiative for Excellence is
gratefully acknowledged (EXC 314). We would like to thank Prof.
Reinhard Sterner (Institute of Biophysics and Physical Biochem-
istry, University of Regensburg) for his support and providing us
the facilities for protein analytical measurements. 1-(2-Deoxy-2-
fluoro-␤-d-arabinofuranosyl)uracil was kindly provided by Prof.
Alex Azhayev (Metkinen Chemistry). We further thank Prof. Xiaoda
Yang (School of Pharmaceutical Sciences, Peking University Health
Science Centre), Dr. Andy Maraite and Prof. Marion Ansorge-
Schumacher (Technische Universität Berlin) for fruitful discussions.
[27] J. Sambrook, D.W. Russell, Molecular Cloning: A Laboratory Manual, 3rd ed.,
Cold Spring Harbor Laboratory Press, New York, 2001.
[28] R.A. Copeland, ENZYMES a Practical Introduction to Structure, Mechanism, and
Data Analysis, 2nd ed, Wiley-VCH, New York, 2000.
[29] A.A. Akhrem, G.V. Zaitseva, I.A. Mikhailopulo, A.F. Abramov, J. Carbohydr. Nucl.
4 (1977) 43–63.
[30] M.J. Pugmire, S.E. Ealick, Structure 6 (1998) 1467–1479.
[31] J. Mendieta, S. Martin-Santamaria, E.M. Priego, J. Balzarini, M.J. Camarasa, M.J.
Perez-Perez, F. Gago, Biochemistry 43 (2004) 405–414.
[32] K. Szeker, O. Niemitalo, M.G. Casteleijn, A.H. Juffer, P. Neubauer, J. Biotechnol.
156 (2010) 268–274.
[33] K. Ukkonen, A. Vasala, H. Ojamo, P. Neubauer, Microb. Cell Fact. 10 (2011) 107.
[34] T.A. Krenitsky, J.V. Tuttle, Biochim. Biophys. Acta 703 (1982) 247–249.
[35] D.F. Visser, F. Hennessy, J. Rashamuse, B. Pletschke, D. Brady, J. Mol. Catal. B:
Enzym. 68 (2011) 279–285.
[36] T. Utagawa, H. Morisawa, Y. Shigeru, A. Yamazaki, F. Yoshinaga, Y. Hirose, Agric.
Biol. Chem. 49 (1985) 3239–3246.
[37] H. Shirae, K. Yokozeki, Agric. Biol. Chem. 55 (1991) 1849–1857.
[38] N.G. Panova, E.V. Shcheveleva, K.S. Alekseev, V.G. Mukhortov, A.N. Zuev, S.N.
Mikhailov, R.S. Esipov, D.V. Chuvikovskii, A.I. Miroshnikov, Mol. Biol. (Mosk) 38
(2004) 907–913.
[39] N.G. Panova, C.S. Alexeev, A.S. Kuzmichov, E.V. Shcheveleva, S.A. Gavryushov,
K.M. Polyakov, A.M. Kritzyn, S.N. Mikhailov, R.S. Esipov, A.I. Miroshnikov, Bio-
chemistry (Moscow) 72 (2007) 21–28.
[40] C.S. Alexeev, N.G. Panova, K.M. Polyakov, S.N. Mikhailov, Abstracts of XIX Inter-
national Round Table on Nucleosides, Nucleotides and Nucleic Acids IRT 2010,
2010, p. 94.
[41] G.V. Zaitseva, A.I. Zinchenko, V.N. Barai, N.I. Pavlova, E.I. Boreko, I.A. Mikhailop-
ulo, Nucleosides Nucleotides 18 (1999) 687–688.
[42] J.V. Tuttle, M. Tisdale, T.A. Krenitsky, J. Med. Chem. 36 (1993) 119–125.
[43] E.A. Green, R.D. Rosenstein, R. Shiono, D.J. Abraham, B.L. Trus, R.E. Marsh, Acta
Crystallogr. B 31 (1975) 102–107.
[44] C. Marck, B. Lesyng, W. Saenger, J. Mol. Struct. 82 (1982) 77–94.
[45] A. Hempel, N. Camerman, J. Grierson, D. Mastropaolo, A. Camerman, Acta Crys-
tallogr. C 55 (1999) 632–633.
References
[1] N. Hori, M. Watanabe, Y. Yamazaki, Y. Mikami, Agric. Biol. Chem. 54 (1990)
763–768.
[2] T. Hamamoto, T. Noguchi, Y. Midorikawa, Biosci. Biotechnol. Biochem. 60 (1996)
1179–1180.
[3] M.J. Pugmire, S.E. Ealick, J. Biochem. 361 (2002) 1–25.
[4] T. Furukawa, A. Yoshimura, T. Sumizawa, M. Haraguchi, S. Akiyama, K. Fukui,
M. Ishizawa, Y. Yamada, Nature 356 (1992) 668.
[5] S. Akiyama, T. Furukawa, T. Sumizawa, Y. Takebayashi, Y. Nakajima, S. Shi-
maoka, M. Haraguchi, Cancer Sci. 95 (2004) 851–857.
[6] P.P. Saunders, B.A. Wilson, G.F. Saunders, J. Biol. Chem. 244 (1969) 3691–3697.
[7] M.G. Vander Heiden, Nat. Rev. Drug Discov. 10 (2011) 671–684.
[8] E. De Clercq, Annu. Rev. Pharmacol. 51 (2011) 1–24.
[9] A.J. Berdis, Biochemistry 47 (2008) 8253–8260.
[10] P. Herdewijn, Modified Nucleosides in Biochemistry, Biotechnology and
Medicine, Wiley-VCH, Weinheim, 2008.
[11] I.A. Mikhailopulo, A.I. Miroshnikov, Mendeleev Commun. 21 (2011) 57–68.
[12] N. Hori, M. Watanabe, Y. Yamazaki, Y. Mikami, Agric. Biol. Chem. 53 (1989)
2205–2210.
[46] J.V. Tuttle, T.A. Krenitsky, European Patent Publication EP 0285432 B1
(1992).