A comparison between immobilized pyrimidine nucleoside phosphorylase from Bacillus subtilis and thymidine phosphorylase from Escherichia coli in the synthesis of 5-substituted pyrimidine 2′-deoxyribonucleosides

Article abstract of DOI:10.1016/j.molcatb.2013.05.007

Pyrimidine nucleoside phosphorylase from Bacillus subtilis (BsPyNP, E.C. 2.4.2.3) and thymidine phosphorylase from Escherichia coli (EcTP, E.C. 2.4.2.4) were used, as immobilized enzymes, in the synthesis of 5-halogenated pyrimidine 2′-deoxyribonucleosides (14-18) by transglycosylation in fully aqueous medium. From the comparative study of the two biocatalysts, no remarkable differences emerged about their substrate specificity, bioconversion yield, stability in organic cosolvents (DMF and MeCN). Moreover, both biocatalysts could be recycled for at least 5 times with no loss of the productivity. Both enzymes do not accept arabinonucleosides and 2′,3′- dideoxynucleosides as substrates, whereas they catalyze bioconversions involving 5′-deoxyribonucleosides and 5-halogenated uracils. The synthesis of compounds 14-18 proceeded at a similar conversion (33-68% for BsPyNP and 25-62% for EcTP, respectively). Immobilization was found to exert, for both the biocatalysts, a dramatic enhancement of stability upon incubation in MeCN. Optimization of 5-fluoro-2′-deoxyuridine (14) synthesis (pH 7.5, 10 mM phosphate buffer, nucleoside/nucleobase 3:1 molar ratio) and subsequent scale-up afforded the target compound in 73% (EcTP) or 76% (BsPyNP) conversion (about 9 g/L).

Full text of DOI:10.1016/j.molcatb.2013.05.007

Journal of Molecular Catalysis B: Enzymatic 95 (2013) 16–22
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
A comparison between immobilized pyrimidine nucleoside phosphorylase from
Bacillus subtilis and thymidine phosphorylase from Escherichia coli in the
ꢀ
synthesis of 5-substituted pyrimidine 2 -deoxyribonucleosides
a
a
a,b,1
, Simona Dalyc,
Immacolata Serra , Teodora Bavaro , Davide A. Cecchini
b
a
a,∗
Alessandra M. Albertini , Marco Terreni , Daniela Ubiali
Department of Drug Sciences and Italian Biocatalysis Center, University of Pavia, via Taramelli 12, I-27100 Pavia, Italy
Department of Biology and Biotechnology “L. Spallanzani”, University of Pavia, via Ferrata 1, I-27100 Pavia, Italy
Gnosis S.p.A., Via Lavoratori Autobianchi 1, I-20832 Desio, MB, Italy
a
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Pyrimidine nucleoside phosphorylase from Bacillus subtilis (BsPyNP, E.C. 2.4.2.3) and thymidine phos-
phorylase from Escherichia coli (EcTP, E.C. 2.4.2.4) were used, as immobilized enzymes, in the synthesis
of 5-halogenated pyrimidine 2 -deoxyribonucleosides (14–18) by transglycosylation in fully aqueous
Received 14 January 2013
Received in revised form 27 April 2013
Accepted 15 May 2013
Available online xxx
ꢀ
medium. From the comparative study of the two biocatalysts, no remarkable differences emerged about
their substrate specificity, bioconversion yield, stability in organic cosolvents (DMF and MeCN). Moreover,
both biocatalysts could be recycled for at least 5 times with no loss of the productivity.
Keywords:
Nucleoside phosphorylase
Transglycosylation
ꢀ
ꢀ
Both enzymes do not accept arabinonucleosides and 2 ,3 -dideoxynucleosides as substrates, whereas
ꢀ
they catalyze bioconversions involving 5 -deoxyribonucleosides and 5-halogenated uracils. The synthe-
sis of compounds 14–18 proceeded at a similar conversion (33–68% for BsPyNP and 25–62% for EcTP,
respectively). Immobilization was found to exert, for both the biocatalysts, a dramatic enhancement of
stability upon incubation in MeCN.
5
2
-Halogenated pyrimidine
-deoxynucleosides
ꢀ
Floxuridine
ꢀ
Optimization of 5-fluoro-2 -deoxyuridine (14) synthesis (pH 7.5, 10 mM phosphate buffer, nucleo-
side/nucleobase 3:1 molar ratio) and subsequent scale-up afforded the target compound in 73% (EcTP)
or 76% (BsPyNP) conversion (about 9 g/L).
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
halogenating agents and their toxicity frequently discourage this
approach which is also plagued by harsh experimental conditions,
long reaction times and complex work-up. On the other hand,
following a convergent approach, silylated uracil derivatives can
be chemically coupled with protected 2-deoxyribose or ribose
[9–11]. In the former case, a mixture of the ␣- and ␤-anomers
is produced which must be chromatographically separated and
By the late 1970s, when acyclovir came to the market, nucleo-
side analogues have been extensively investigated giving a new
impetus to the search for more active and selective molecules. Such
efforts have provided a large arsenal of antiviral and antitumor
chemotherapeutic agents still used nowadays in clinical practice. 5-
ꢀ
ꢀ
Substituted pyrimidine 2 -deoxynucleosides with a halogen (F, Br,
finally deprotected. In the latter approach, the 2 -O-protecting
I) or a halogenated group (CF , CH CHBr) constitute a small class
group on the ribose unit controls the stereochemistry of the
glycosidation to a single product. Notwithstanding, selective
2 -deoxygenation and final deprotection are eventually necessary
3
of drugs used for the topical treatment of local Herpes simplex virus
and Varicella zoster virus infections (15, 17, 18), or of certain types
of cancer (14, 16) (see Scheme 2) [1–6].
ꢀ
to complete the synthetic sequence.
5
-Substituted pyrimidine nucleosides can be synthesized by
Enzyme-based syntheses have a tremendous potential over
totally chemical methods. Nucleoside phosphorylases (NPs; E.C.
2.4.2), which participate by nature in the salvage pathway of
nucleoside biosynthesis [12,13], can be exploited for the stere-
C-5 halogenation of uracil-based nucleosides or their O-acetylated
derivatives [7]. In spite of some advances [8], difficult handling of
ꢀ
oselective synthesis of 2 -deoxynucleoside analogues [14–16].
NPs catalyze the reversible cleavage of the glycosidic bond of
∗ Corresponding author. Tel.: +39 0382 987889; fax: +39 0382 422975.
(deoxy)ribonucleosides in the presence of inorganic orthophos-
E-mail address: daniela.ubiali@unipv.it (D. Ubiali).
1
phate (Pi) as a second substrate to generate the nucleobase and
Present address: Département de Génie Biochimique et Alimentaire, INSA, 135
avenue de Rangueil, 31077 Toulouse Cedex 4, France.
␣-d-(deoxy)ribose-1-phosphate (see Scheme 1) [12]. If a second
1
381-1177/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcatb.2013.05.007

I. Serra et al. / Journal of Molecular Catalysis B: Enzymatic 95 (2013) 16–22
17
Scheme 1. Phosphorolysis of sugar donor nucleosides (1–6) catalyzed by immobilized BsPyNP or EcTP.
◦
◦
nucleobase is added to the reaction medium, the formation
of a new nucleoside can result (transglycosylation, Scheme 2).
Thus, following this pattern, the stereocontrolled synthesis
conditions (pH 10 and 45 C or 37 C), fundamental for biocatalytic
applications, i.e. to increase reagent solubility (and concentration,
thereof).
ꢀ
of 2 -deoxynucleosides can be achieved without the need for
Immobilized BsPyNP, coupled to immobilized purine nucleo-
side phosphorylase (PNP, E.C. 2.4.2.1) from B. subtilis, catalyzed
the synthesis of 2 -deoxyguanosine and 2 -deoxyinosine [19].
Immobilized BsPyNP also catalyzed the synthesis of doxifluridine
(5 -deoxy-5-fluorouridine), a prodrug of 5-fluorouracil [17,21].
Immobilized EcTP was tested in the synthesis of floxuridine by
using 2 -deoxyuridine or thymidine and 5-fluorouracil as reagents
[20]. Whereas the synthesis of purine 2 -deoxynucleosides cat-
protection/deprotection, in a “two-step, one-pot” reaction.
We have recently described the immobilization and a few syn-
thetic applications of purified uridine phosphorylase from Bacillus
subtilis (E.C. 2.4.2.3) [17–19], herein BsPyNP, and thymidine phos-
phorylase from Escherichia coli (EcTP, E.C. 2.4.2.4) [20]. According
to the classification proposed by Pugmire and Ealick [12], both
these enzymes belong to the NPII family, being characterized by
specificity for pyrimidine nucleosides.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
alyzed by immobilized BsPyNP and BsPNP was deeply investigated
by assaying the effect of pH, temperature, reagent concen-
tration till the achievement of 85–92% yield, “one-enzyme”
Immobilization of BsPyNP and EcTP [18,20] provided more sta-
ble biocatalysts than the soluble counterparts in non-physiological
Scheme 2. Synthesis of 5-substituted nucleosides 2 and 14–18 via transglycosylation catalyzed by immobilized BsPyNP or EcTP.

1
8
I. Serra et al. / Journal of Molecular Catalysis B: Enzymatic 95 (2013) 16–22
ꢀ
transglycosylations for the synthesis of pyrimidine 2 -
deoxyribonucleosides (i.e. floxuridine) [20] and congeners (i.e.
doxifluridine) [17] do lack such a systematic study.
pre-seed culture to a final concentration of 0.5% (v/v). The seed cul-
ture was incubated at 30 C for 16 h and was subsequently diluted
◦
to 5% into a stirred bioreactor containing 15 L of fermentative
medium ((NH ) SO 2 g/L, K HPO4 15 g/L, KH PO4 7.5 g/L, MgSO4
2 g/L, dextrose 50 g/L). The pH was controlled by automatic addi-
In this paper we report on a comparison between immobi-
lized BsPyNP and EcTP in the synthesis of 5-substituted pyrimidine
nucleosides 14–18 (Scheme 2). Floxuridine (14) was selected as the
target compound to optimize the experimental conditions. Apart
from its higher clinical relevance, this choice was addressed by
practical reasons: we aimed at developing a scalable synthetic pro-
cess in fully aqueous medium and floxuridine is the most water
soluble among the considered nucleosides. Bioconversion of 5-
fluorouracil (9, X = F, see Scheme 2 and Table 3) was carried out
till 50 mM. Halogenated nucleosides and nucleobases other than 5-
fluoro derivatives suffer from low solubility in aqueous solvents and
this property can be a shortcoming in scaling-up enzyme-catalyzed
transformations in fully aqueous medium. For this reason, we also
assayed the stability of immobilized BsPyNP and EcTP in the pres-
ence of acetonitrile (MeCN) and N,N-dimethylformamide (DMF) as
cosolvents in the view of preparative applications.
4
2
4
2
2
tion of diluted NH OH (25%, v/v), the stirring rate was adjusted to
4
500 rpm and the pO2 was maintained automatically at 50% with
aeration and agitation. The strain was grown to mid-exponential
◦
phase at 30 C until OD
reached 50, the temperature was then
600
◦
lowered to 22 C and the protein expression was induced by adding
0.5 mM isopropylthio-␤-d-galactoside (IPTG). After a 24 h expres-
sion phase, the final OD600 of the culture was 80.
Cells were disrupted by two cycles of homogenization at
1000 bar and the suspension was clarified by adding 0.5% qua-
ternary ammonium salt followed by centrifugation. The enzyme
was purified by loading the crude extract onto a Toyopearl Giga-
Cap Q-650 M column previously equilibrated with Tris–HCl buffer
pH 7.5. The adsorbed enzyme was then eluted with a linear gradi-
ent (0–0.5 M) of NaCl in 20 mM Tris–HCl pH 7.5 and 1 mM MgCl2.
Fractions containing the active enzyme were collected and dialyzed
An example of floxuridine synthesis by transglycosylation has
been recently reported [22]. However, a cell-based bioconversion
was studied whilst our attention was here focused on purified and
immobilized enzymes as biocatalysts which are more viable when
running pharmaceutical syntheses.
against 10 mM Tris–HCl pH 7.5 containing 1 mM MgCl and 250 mM
2
◦
NaCl and stored at −20 C.
After purification, 60 mL of protein extract were obtained; pro-
tein concentration was 5 mg/mL with a specific activity towards
ꢀ
2
-deoxyuridine (1) of 15 IU/mg (spectrophotometrically deter-
mined). The purity of the protein (>90%) was assessed by SDS-PAGE
electrophoresis (4–12%).
2
. Experimental
2.1. General
2.3. Spectrophotometric enzyme activity assay
Nucleosides (1–6 and 14–18), nucleobases (7–13), dextran (MW
0 kDa) and PEI (polyethylenimine, branched, MW 25 kDa) were
Enzyme activity assay was performed as previously reported
19]. Briefly, the activity of both BsPyNP and EcTP was deter-
2
[
purchased from Sigma–Aldrich and/or VWR International (Milano,
ꢀ
mined by measuring the absorbance at 297 nm of uracil produced
Italy). 2 -Deoxyuridine (1) was supplied by Pro.Bio.Sint. (Varese,
ꢀ ◦
from 2 -deoxyuridine (1) at 37 C in a cuvette of 1 cm light
ꢀ
Italy). 5 -Deoxyuridine (4) was prepared as previously reported
−
1
−1
®
path (ε = 1912 M cm ). One IU corresponds with an amount of
[
17]. Sepabeads EC-EP were kindly provided by Resindion Mit-
ꢀ
enzyme that liberates 1 ␮mol of uracil (7) from 2 -deoxyuridine (1)
subishi Chemical Co. (Binasco, Milano, Italy). All solvents were of
HPLC grade.
per minute.
Enzymatic reactions were monitored by using a HPLC Merck
Hitachi L-7100 equipped with a UV detector L-7400 and a column
oven L-7300 (Darmstadt, Germany). The column was a Lichrocart
2.4. Enzyme immobilization
◦
RP18 (5 ␮m, 250 mm × 4.6 mm, Merck) kept at 35 C; flow rate was
Immobilization of both BsPyNP and EcTP was performed accord-
ing to the procedure previously described [20]. Briefly, the enzyme
(EcTP, load: 1.3 mg/g; loaded activity: 60 IU/g; BsPyNP, load:
3 mg/g; loaded activity: 45 IU/g) was added to a suspension of
the matrix (Sepabeads EC-EP® derivatized with PEI) [18] in 5 mM
phosphate buffer pH 7.5 and gently stirred for 20 min at room
temperature. A freshly prepared solution of aldehyde dextran was
added to the immobilization suspension. After stirring for 1 h, pH
was adjusted to 10.05 and NaBH4 was added keeping the reaction
under mechanical stirring for 30 min. The immobilized preparation
was then filtered and washed with 10 mM potassium phosphate
buffer pH 5 and deionized water.
1
mL/min and analyses were detected at ꢀ = 260 nm.
Upon 15 L fermentation, cells were disrupted using a GEA Niro
Soavi Homogenizer (Parma, Italy) and cell debris was precipitated
by centrifugation using an Avant J25 Beckman Coulter (Cassina De’
Pecchi, Milan, Italy). The column for protein purification was from
Tosoh Bioscience (Rivoli, Turin, Italy). Enzyme activity [18,20] and
protein concentration [23] were evaluated on a Shimadzu spec-
trophotometer UV 1601 (Milano, Italy).
2.2. Enzymes
Thymidine phosphorylase from E. coli (E.C. 2.4.2.4) was pur-
2
.5. Enzyme stability assay
chased from Sigma-Aldrich (Milano, Italy). Protein concentration
ꢀ
was 33 mg/mL with a specific activity towards 2 -deoxyuridine (1)
Immobilized enzyme preparation (200 mg) was added to a solu-
of 46 IU/mg (spectrophotometrically determined).
tion (1.5 mL) containing 50 mM phosphate buffer and 20% of MeCN
or DMF (v/v) at pH 7.5. The mixture was stirred at room tempera-
ture for 24 h. At fixed times, 40 ␮l of suspension was withdrawn and
activity was tested by standard spectrophotometric activity assay.
Expression and purification of BsPyNP (E.C. 2.4.2.2) were carried
out at 15 L fermentation volume following the procedure previ-
ously reported [19], with some modifications to maximize the
protein expression.
Thus, a three-phase process (pre-seed, seed and fermenta-
tion) was carried out. The pre-seed cells were prepared in 50 mL
LB medium and incubated at 37 C. After 6 h, 500 mL of pre-
2.6. Phosphorolysis: general protocol
◦
fermentative medium (yeast extract 1 g/L, (NH ) SO 1 g/L, KH PO
To a solution of the nucleoside (1–6) (10–12 mg, 5 mM) in 50 mM
phosphate buffer pH 7.5 (10 mL), immobilized EcTP or BsPyNP
4
2
4
2
4
1
5 g/L, MgSO4 5 g/L, dextrose 10 g/L) were inoculated with the

I. Serra et al. / Journal of Molecular Catalysis B: Enzymatic 95 (2013) 16–22
19
(
1 IU) was added and the mixture was stirred at room tempera-
2.8. Recycling of the biocatalysts
ture. Aliquots (0.2 mL) of the reaction mixture were withdrawn at
fixed time intervals (5, 10, and 15 min), filtered through a pipette
filter device to remove the biocatalyst and analyzed by HPLC (see
Recycling of immobilized EcTP or BsPyNP was performed by
evaluating the synthesis of thymidine (2) by transglycosylation, in
the same conditions described in Section 2.7. When the highest
conversion was achieved, the reaction mixture was filtered under
reduced pressure and the immobilized biocatalyst was re-used for
the next reaction.
below for chromatographic conditions and Rt).
ꢀ
2
-Deoxyuridine (1). Mobile phase: 0.01 M KH PO buffer
2 4
ꢀ
pH 4.6/methanol 90% (97:3). Uracil (7, X = H), Rt = 4.92 min, 2 -
deoxyuridine (1), Rt = 12.19 min.
Thymidine (2). Mobile phase: 0.01 M KH PO4 buffer pH
2
4
.6/methanol 90% (93:7). Thymine (8, X = CH ), Rt = 4.36 min,
3
3. Results and discussion
thymidine (2), Rt = 12.27 min.
Uridine (3). Mobile phase: 0.01 M KH PO4 buffer pH
.6/methanol 90% (97:3). Uracil (7, X = H), Rt = 4.92 min, uridine
2
The aim of this work was the synthesis of some 5-halogenated
pyrimidine nucleosides of medicinal interest (14–18) through
a “one-enzyme, one-pot” transglycosylation in fully aqueous
medium. We selected immobilized BsPyNP from B. subtilis and EcTP
from E. coli as the biocatalysts, which have been shown to display
promising properties in terms of activity and stability [17,19,20]
when assayed in nucleoside biotransformation.
4
(
3), Rt = 5.86 min.
ꢀ
5
-Deoxyuridine (4). Mobile phase: 0.01 M KH PO buffer pH
2 4
ꢀ
4
.6/methanol 90% (85:15). Uracil (7, X = H), Rt = 2.10 min, 5 -
deoxyuridine (4), Rt = 10.23 min.
ꢀ
ꢀ
Arabinosyluracil (5) and 2 ,3 -dideoxyuridine (6). Mobile phase:
0
(
.01 M KH PO buffer pH 4.6/methanol 90% (90:10). Uracil
2
4
ꢀ
ꢀ
7, X = H), Rt = 2.60 min, arabinosyluracil (5), Rt = 5.25 min, 2 ,3 -
3.1. Screening of sugar donor (1–6) (phosphorolysis)
dideoxyuridine (6), Rt = 8.67 min.
ꢀ
Target compounds of transglycosylation reactions are all 2 -
deoxyribonucleosides (14–18, see Scheme 2). Therefore, the first
2
.7. Transglycosylation: general protocol
ꢀ
issue was to select a pyrimidine 2 -deoxynucleoside as the sugar
ꢀ
donor. 2 -Deoxyuridine (1) and thymidine (2) were thus assayed in
To a solution of phosphate buffer (10 mM) at pH 7.5 (10 mL) con-
phosphorolysis (Scheme 1) since they are the natural substrates of
BsPyNP and EcTP, respectively. Besides, both these compounds can
be easily synthesized at a moderate price resulting in affordable
starting material for synthetic applications.
ꢀ
taining 2 -deoxyuridine (1, 45.6 mg, 20 mM) and the 5-substituted
base (8–13, 13–22 mg, 10 mM), 10 IU of immobilized enzyme
(
EcTP or BsPyNP) were added and the suspension was kept
under mechanical stirring at room temperature. In the case of
-bromovinyluracil (13), pH was set to 10 by using a 10 mM
ꢀ
As expected, BsPyNP and EcTP are able to convert both 2 -
5
deoxyuridine (1) and thymidine (2) into the parent nucleobases.
However, it is worth noting that for both substrates the rate of the
phosphorolysis (vp) of EcTP is 2.5-fold higher than that of BsPyNP:
being the equilibrium-controlled conversion comparable (1) or the
same (2), the higher bioconversion rate of EcTP can be advantageous
in terms of reaction time and enzyme stability.
phosphate buffer containing 13.8 mg of potassium carbonate (final
concentration 10 mM). Aliquots (0.2 mL) were periodically with-
drawn, filtered through a pipette filter device to remove the
biocatalyst and analyzed by HPLC (see below for chromatographic
conditions and Rt). When the highest conversion was achieved,
the reaction was stopped by filtration of the immobilized bio-
catalyst on a Büchner funnel with a sintered glass disc under
reduced pressure. The produced nucleosides (2 and 14–18) were
identified by comparison of their HPLC Rt with that of authentic
samples.
Unnatural sugar-modified nucleosides (4–6) were also tested
ꢀ
as substrates in the phosphorolysis reaction. 5 -Deoxyuridine (4),
ꢀ
ꢀ
arabinosyluracil (5) and 2 ,3 -dideoxyuridine (6) are valuable sugar
donor to synthesize nucleoside analogues used as antiviral or anti-
tumour drugs, either through a “one-enzyme” transglycosylation,
or following a “two-enzyme” synthetic sequence. However, both
BsPyNP and EcTP displayed a narrow substrate specificity since
Thymidine (2). Mobile phase: 0.01 M KH PO buffer pH
.6/methanol 90% (90:10). Uracil (7), Rt = 3.30 min, 2 -deoxyuridine
2
4
ꢀ
4
(
1), Rt = 4.84 min, thymine (8, X = CH ), Rt = 5.54 min, thymidine (2),
3
ꢀ
their ability to cleave the glycosidic bond was restricted to 2 -
deoxyribonucleosides (1 and 2), uridine (3) and 5 -deoxyuridine
Rt = 9.74 min.
ꢀ
ꢀ
ꢀ
5
-Fluoro-2 -deoxyuridine (14) and 5-trifluoromethyl-2 -
(4) (Table 1). Inversion of the rate of phosphorolysis (vp) of BsPyNP
deoxyuridine (15). Mobile phase: 0.01 M KH PO4 buffer
pH 4.6/methanol 90% (97:3). Uracil (7), Rt = 4.92 min, 2 -
2
and EcTP for substrates 3 and 4 is consistent with the evidence that,
ꢀ
deoxyuridine (1), Rt = 12.19 min, 5-fluorouracil (9, X = F),
Table 1
Rt = 5.65 min, 5-trifluoromethyluracil (10, X = CF ), Rt = 10.89 min,
3
ꢀ
Phosphorolysis of sugar donor nucleosides (1–6) catalyzed by immobilized BsPyNP
or EcTP.
5
2
-fluoro-2 -deoxyuridine (14), Rt = 15.89 min, 5-trifluoromethyl-
ꢀ
-deoxyuridine (15), Rt = 16.91 min.
ꢀ
5
-Bromo-2 -deoxyuridine (16). Mobile phase: 0.01 M KH PO
Sugar donor
BsPyNP (vp)
EcTP (vp)
2
4
buffer pH 4.6/methanol 90% (95:5). Uracil (7), Rt = 4.07 min,
1
2
3
4
5
6
70(3.6)
62(2.6)
66(3.7)
50(3.4)
0(–)
60(9.0)
60(7.1)
51(0.5)
50(0.7)
0(–)
ꢀ
2
-deoxyuridine (1), Rt = 6.7 min, 5-bromouracil (11, X = Br),
ꢀ
Rt = 8.79 min, 5-bromo-2 -deoxyuridine (16), Rt = 21.19 min.
ꢀ
5
-Iodo-2 -deoxyuridine (17). Mobile phase: 0.01 M KH PO
2
4
buffer pH 4.6/methanol 90% (90:10). Uracil (7), Rt = 2.60 min,
0(–)
0(–)
ꢀ
2
-deoxyuridine (1), Rt = 5.41 min, 5-iodouracil (12, X = I),
Percentage of produced nucleobases after 24 h and reaction rate (vp,
mol min
buffer, room temperature, substrate (1–6) = 5 mM, volume = 10 mL, BsPyNP or
EcTP = 1 IU.
, 2 -deoxyuridine; 2, thymidine; 3, uridine; 4, 5 -deoxyuridine; 5, arabinosyluracil;
ꢀ
Rt = 9.35 min, 5-iodo-2 -deoxyuridine (17), Rt = 15.29 min.
−1 −1
) are reported. Experimental conditions: 50 mM phosphate
␮
g
ꢀ
5
-Bromovinyl-2 -deoxyuridine (18) Mobile phase: 0.01 M
KH PO buffer pH 4.6/acetonitrile, gradient elution, 0 min-5 min
2
4
ꢀ ꢀ
ꢀ ꢀ
2 ,3 -dideoxyuridine; BsPyNP, pyrimidine nucleoside phosphorylase from
1
6
(
5% acetonitrile), 5.1 min-22 min (15% acetonitrile). Uracil (7),
,
ꢀ
Rt = 4.07 min, 2 -deoxyuridine (1), Rt = 6.1 min, 5-bromovinyluracil
(
(
Bacillus subtilis; EcTP, thymidine phosphorylase from Escherichia coli; vp, rate of
phosphorolysis.
All experiments were performed as duplicate.
ꢀ
13, X = CHCHBr), Rt = 18.59 min, 5-bromovinyl-2 -deoxyuridine
18), Rt = 19.67 min.

2
0
I. Serra et al. / Journal of Molecular Catalysis B: Enzymatic 95 (2013) 16–22
differently from TPs, PyNPs do not discriminate at the 2-position of
the ribose [12].
3.2. Synthesis of 5-halogenated pyrimidine nucleosides 14–18
(
transglycosylation)
Substrate specificity of BsPyNP and EcTP against 5-halogenated
pyrimidines (9–13, see Scheme 2) was evaluated through the trans-
ꢀ
glycosylation reaction using 2 -deoxyuridine (1) as the sugar donor.
According to the results of Table 1, regardless the used biocata-
lyst, phosphorolysis of 1 was higher or similarly yielding than the
bioconversion of thymidine (2) and proceeded at a higher rate.
The reactions were performed by using a 2:1 molar ratio
ꢀ
of 2 -deoxyuridine (1) in order to generate a slight excess of
2
-deoxyribose-1-phosphate for the following coupling with the
nucleobase (9–13). Whereas in the phosphorolysis the phosphate
concentration was 10-fold higher than the nucleoside concentra-
tion (50 mM vs 5 mM), transglycosylations were performed by
using a equimolar ratio of phosphate and the nucleobase (10 mM).
In fact, in this case, phosphate acts as a substrate in the first step of
the bioconversion (phosphorolysis) but it then becomes a reaction
product being regenerated once the coupling between the nucle-
obase and the intermediate sugar-1-phosphate has occurred [24].
All the tested nucleobases (9–13) were converted into the corre-
ꢀ
sponding 2 -deoxyribonucleosides (14–18) (Scheme 2). No striking
differences between BsPyNP and EcTP were observed about the
percentage of conversion of (16) and (17), but for the fluorinated
compounds (14 and 15) and brivudin (18). BsPyNP-Catalyzed syn-
thesis of floxuridine (14) and trifluridine (15) afforded slightly
higher conversion (68% and 57%, respectively) than EcTP (62% and
Fig. 1. Cycles of transglycosylation reaction catalyzed by immobilized BsPyNP and
EcTP (panel A and B, respectively).
5
1%). The overall low conversion of 13, associated to both the bio-
catalysts, might be ascribed to the steric hinderance of the bulky
fermented till up 15 L and BsPyNP could be therefore readily avail-
able in large amount. The study of fermentation and purification
conditions allowed the production of BsPyNP on a pre-industrial
scale with low production cost and good results for that concerning
yield and quality of the purified enzyme.
5
-substituent, in agreement with previous reports [22].
The rate of synthesis (vs, Scheme 2) emerged as the main differ-
ence between the two biocatalysts. Bioconversions were generally
faster when the biocatalyst was EcTP.
As reported in Table 2, the highest conversions were achieved in
the pH range 7.5–10, whereas a weakly acid pH was detrimental for
the reaction outcome; even if the use of strongly basic pH appeared
to be eligible, the optimal compromise between percentage of con-
version, reaction rate and long term stability of the biocatalyst was
pH 7.5 which was therefore set in the next study on the effect of
substrate concentration (phosphate, nucleobase and nucleoside).
According to data reported in Table 3, a decrease of phosphate
concentration from 10 to 1 mM (entries 1–3) did not exert any effect
on the final conversion but exclusively on the initial rate of the
reaction. Neither the conversion, nor the rate of synthesis were
affected when substrate concentration was increased from 20 and
3
.3. Recycling of the biocatalysts
Immobilized enzymes offer both technical and economical
advantages over their soluble counterparts [25], one of them
being the possibility to develop a continuous process. Immobilized
BsPyNP and EcTP have been here tested in the in continuum syn-
thesis of thymidine (2) by transglycosylation. To this aim, each
ꢀ
biocatalyst was incubated in the reaction mixture containing 2 -
deoxyuridine (1) and thymine (8) in the standard conditions used
for the transglycosylation studies (2:1 molar ratio, 10 mM phos-
phate buffer pH 7.5, r.t., see Section 2). When the highest conversion
was achieved, the mixture was separated from the biocatalyst via
a simple filtration and the immobilized enzyme was then re-used
for a new bioconversion. As depicted in Fig. 1, both biocatalysts
retained all of their activity after 5 reactions showing optimal sta-
bility and recyclability.
1
0 mM to 80 and 40 mM for the sugar donor and the nucleobase,
respectively: as mentioned above (Section 3.2) the in situ phosphate
regeneration following the glycosylation step concurs to flatten the
influence of the buffer concentration (5 mM or 10 mM, entries 4–5).
No significant loss of activity was observed both for BsPyNP and
EcTP after 5 cycles thus indicating that these biocatalysts could be
re-used for additional reactions.
Table 2
Effect of pH on the synthesis of floxuridine (14) catalyzed by immobilized BsPyNP
(
see Scheme 2).
−
1
−1
g )
pH
vs (␮mol min
Conversion (%)
3.4. Optimization and scale-up of floxuridine (14)
6
7
0.07
1.6
1.7
1.5
1.3
40
63
68
70
70
The synthesis of floxuridine (14) was further investigated with
7
9
0
.5
the aim to highlight the influence of reaction parameters such as pH
and substrate concentration on the percentage of bioconversion.
We opted for BsPyNP as the biocatalyst because the conversion
of 9 into 14 was slightly higher compared to EcTP: percentage
of bioconversion is reasonably privileged over reaction rate in
the optimization of a synthetic process. Moreover, B. subtilis was
a
1
Experimental conditions: 10 mM phosphate buffer, room temperature, substrate
1) = 20 mM, 5-fluorouracil (9) = 10 mM, volume = 10 mL, BsPyNP = 1 IU. Conversions
were recorded after 24 h. All experiments were performed as duplicate.
(
a
Phosphate buffer was additioned with 10 mM K2CO3.

I. Serra et al. / Journal of Molecular Catalysis B: Enzymatic 95 (2013) 16–22
21
Table 3
Effect of substrate concentration on the synthesis of floxuridine (14) catalyzed by immobilized BsPyNP or EcTP (see Scheme 2).
−
1
−1
g )
Entry
Enzyme
1 (mM)
9 (mM)
Phosphate buffer (mM)
vs (␮mol min
Conversion (%)
1
2
3
4
5
6
7
8
BsPyNP
BsPyNP
BsPyNP
BsPyNP
BsPyNP
BsPyNP
EcTP
20
20
20
80
80
150
80
150
10
10
10
40
40
50
40
50
1
5
10
5
10
10
10
10
1.0
1.6
1.7
1.5
1.6
2.2
2.1
5.7
68
68
68
70
69
76
65
73
EcTP
Experimental conditions: phosphate buffer pH 7.5, room temperature, volume = 10 mL, BsPyNP or EcTP = 10 IU. The reaction was monitored till 10 h. All experiments were
performed as duplicate.
Mostly interesting for a preparative application is the effect
of a higher substrate concentration (150 mM) and molar ratio
buffer but are frequently not sufficient to achieve suitable concen-
trations (g of product per litre of reaction) for in batch product
downstream. 5-Halogenated nucleobases and nucleosides 15–18,
other than floxuridine (14), do possess this limitation. A moderate
content of organic solvent can assist in the reaction and work-up
thereof. Therefore, stability of immobilized BsPyNP and EcTP was
assayed in acetonitrile (MeCN) and N,N-dimethylformamide (DMF)
which are some of the most commonly used solvents for nucleo-
side reactions [8]. As depicted in Fig. 2, both immobilized enzymes
maintained more than 65% of their activity after 24 h incubation
regardless the solvent used.
Surprisingly, the residual activity in DMF of both the soluble
enzymes was slightly higher than that of the parent immobilized
biocatalyst. The sensitively improved stability of the immobilized
preparations with respect to the native enzymes emerged upon
incubation in MeCN: differently from DMF, MeCN exerted a detri-
mental effect on the soluble proteins, particularly in the case of
EcTP.
(
3:1) of the nucleoside donor (entries 6 and 8): as expected in a
controlled-equilibrium biotransformation, both the yield and the
rate of synthesis resulted positively affected. Immobilized BsPyNP
and EcTP, compared in these conditions, catalyzed the synthesis
of 14 with a similar conversion (76% and 73%, respectively) but
different rate of synthesis, being EcTP, once again, 2.5-fold faster
than BsPyNP. Starting from 50 mM of 9 (6.5 g/L), the theoretical
volumetric yield of 14 is about 9 g/L, produced in about 3 h in
fully aqueous medium and mild conditions (room temperature and
pH 7.5).
3.5. Stability of immobilized BsPyNP and EcTP in the presence of
cosolvents
The optimization of a synthetic process, mainly aimed at achiev-
ing the highest conversion and volumetric yield, requires that a
biocatalyst is sufficiently stable to tolerate a wide range of experi-
mental conditions. Immobilized BsPyNP and EcTP were shown to be
From an applicative viewpoint, however, both solvents seem
to be useful to perform a transglycosylation scale-up catalyzed by
immobilized BsPyNP or EcTP. In the case of DMF, being stability of
soluble and immobilized enzymes similar, the use of the immobi-
lized preparation is however to be preferred.
◦
◦
very stable at pH 10 and 45 C or 37 C, respectively [19,20]. These
conditions enhance the solubility of 5-halogenated bases (9–13)
ꢀ
and their 2 -deoxyribonucleoside counterparts (14–18) in aqueous
Fig. 2. Stability of immobilized BsPyNP (panel A) and EcTP (panel B) in presence of organic cosolvents.

2
2
I. Serra et al. / Journal of Molecular Catalysis B: Enzymatic 95 (2013) 16–22
4
. Conclusions
discussion and Dr. Caterina Temporini for English revision. This
work was funded by Regione Lombardia (Metadistretti Call 2007,
ID 4165).
On account of the overall simplicity, nucleoside synthesis
through transglycosylation appears superior to chemical methods.
Particularly, immobilized BsPyNP and EcTP were shown to cat-
alyze the synthesis of 5-halogenated-2 -deoxynucleosides (14–18)
References
ꢀ
[
[
[
1] J.P. Bégué, D. Bonnet-Delpon, J. Fluor. Chem. 127 (2006) 992–1012.
2] E. De Clercq, J. Clin. Virol. 30 (2004) 115–133.
3] E. De Clercq, Curr. Opin. Microbiol. 8 (2005) 552–560.
in fully aqueous medium in moderate to quite high conversion.
From the comparative investigation of BsPyNP and EcTP, a significa-
tive difference did not emerge, neither considering the substrate
specificity in phosphorolysis, nor the performance in the studied
transglycosylations.
[4] E. De Clercq, Antiviral Res. 67 (2005) 56–75.
[
5] A. Tamarit, J. Alberola, M.J. Galindo, D. Navarro, J. Virol. Methods 105 (2002)
47–251.
6] Y. Tsume, J.M. Hilfinger, G.L. Amidon, Mol. Pharm. 5 (2008) 717–727.
2
[
ꢀ
Both enzymes catalyze the phosphorolysis of uracil-based 2 -
[7] V. Kumar, J. Yap, A. Muroyama, S.V. Malhotra, Synthesis (23) (2009) 3957–3962
(for examples about direct halogenation of nucleosides, see Refs. [17–29]
therein).
ꢀ
deoxy-, 5 -deoxy- and ribonucleosides but are not able to cleave
ꢀ
ꢀ
ꢀ
the C1 N bond of arabino- or 2 ,3 -dideoxyribosides. However,
BsPyNP and EcTP complement one another if phosphorolysis rate
is considered: the higher is the reaction rate of EcTP against nat-
[
[
8] V. Kumar, V.S. Parmar, S.V. Malhotra, Biochimie 92 (2010) 1260–1265.
9] E. De Clercq, J. Descamps, P. de Somer, P.J. Barr, A.S. Jones, R.T. Walker, Proc.
Natl. Acad. Sci. U. S. A. 76 (1979) 2947–2951.
ꢀ
[10] H. Vorbruggen, Acc. Chem. Res. 28 (1995) 509–520.
11] H. Vorbrüggen, C. Ruh-Pohlenz, Handbook of Nucleoside Synthesis, John Wiley
Sons, New York, 2001.
[12] M.J. Pugmire, S.E. Ealick, Biochem. J. 361 (2002) 1–25.
ural 2 -deoxyribonucleosides (1 and 2), whilst the faster is the
BsPyNP-catalyzed bioconversion of uridine (3) and 5 -deoxyuridine
[
ꢀ
&
(
4).
Transglycosylations proceed at a higher rate when EcTP is used
[
[
[
13] P.A. Levene, M. Yamagawa, I. Weber, J. Biol. Chem. 60 (1924) 693–706.
14] I.A. Mikhailopulo, Curr. Org. Chem. 11 (2007) 317–335.
15] I.A. Mikhailopulo, A.I. Miroshnikov, Mendeleev Commun. 21 (2011) 57–68.
as biocatalyst. The same behaviour was registered when the syn-
thesis of floxuridine was studied and scaled-up. Both biocatalysts
catalyzed the synthesis of this nucleoside in more than 70% con-
version when a 3:1 molar ratio of nucleoside donor was used, at pH
[16] J.V. Sinisterra, A. Alcántara, M. Almendros, M.J. Hernáiz, J.M. Sanchez-Montero,
J. Trelles, Encyclopedia of Industrial Biotechnology, John Wiley and Sons, New
York, 2010.
[
[
[
17] T. Bavaro, S. Rocchietti, D. Ubiali, M. Filice, M. Terreni, M. Pregnolato, Eur. J. Org.
Chem. (2009) 1967–1975.
18] S. Rocchietti, D. Ubiali, M. Terreni, A.M. Albertini, R. Fernàndez-Lafuente, J.M.
Guisàn, M. Pregnolato, Biomacromolecules 5 (2004) 2195–2200.
19] D. Ubiali, S. Rocchietti, F. Scaramozzino, M. Terreni, A.M. Albertini, R. Fernàndez-
Lafuente, J.M. Guisàn, M. Pregnolato, Adv. Synth. Catal. 346 (2004) 1361–1366.
7
.5 and in fully aqueous medium.
Upon standard conditions used for transglycosylation (10 mM
phosphate buffer, pH 7.5, r.t., nucleoside donor/base acceptor ratio
:1), both biocatalysts remained highly productive after 5 cycles.
2
The optimal stability shown by both immobilized enzymes also in
DMF and MeCN as cosolvents will assist in further scaling-up the
synthesis of 5-halogenated pyrimidine 2 -deoxyribonucleosides.
[20] I. Serra, C.D. Serra, S. Rocchietti, D. Ubiali, M. Terreni, Enzyme Microb. Technol.
9 (2011) 52–58.
4
[
[
21] D.B. Longley, D.P. Harkin, P.G. Johnston, Nat. Rev. Cancer 3 (2003) 330–338.
22] C.W. Rivero, C.N. Britos, M.E. Lozano, J.V. Sinisterra, J.A. Trelles, FEMS Microbiol.
Lett. 331 (2012) 31–36.
ꢀ
[
[
23] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254.
24] G.E.R. Gordon, D.F. Visser, D. Brady, N. Raseroka, M.L. Bode, J. Biotechnol. 151
Acknowledgments
(
2011) 108–113.
[25] (a) L. Cao, Curr. Opin. Chem. Biol. 9 (2005) 217–226;
b) U. Hanefeld, L. Gardossi, E. Magner, Chem. Soc. Rev. 38 (2009) 453–468.
We thank Dr. Milena Schiraldi and Lucia Di Gennaro for
their help in the experimental, Dr. Giorgio Marrubini for helpful
(