Developing a collection of immobilized nucleoside phosphorylases for the preparation of nucleoside analogues: Enzymatic synthesis of arabinosyladenine and 2',3'-dideoxyinosine

Article abstract of DOI:10.1002/cplu.201200278

The use of nucleoside phosphorylases (NPs; EC 2.4.2.n) represents a convenient alternative to the chemical route for the synthesis of natural and modified nucleosides. We purified four recombinantly expressed nucleoside phosphorylases from the bacterial pathogens Citrobacter koseri, Clostridium perfringens, and Streptococcus pyogenes (CkPNPI, CkPNPII, CpUP, SpUP) and their substrate specificity was investigated towards either natural pyrimidine or purine nucleosides and some analogues, namely, arabinosyladenine (araA) and 2',3'-dideoxyinosine (ddI). A 2-3 % activity towards these latter compounds (compared to the natural substrates) was observed. Enzyme activities were compared to the specificities obtained for the enzymes pyrimidine nucleoside phosphorylase from Bacillus subtilis (BsPyNP) and purine nucleoside phosphorylase from Aeromonas hydrophila (AhPNPII) previously reported by some of the authors. The enzymes displaying the suitable specificity for the synthesis of araA and ddI were immobilized on aldehyde-agarose. The immobilized preparations were highly stable at alkaline pH and in the presence of methanol or acetonitrile as cosolvent. They were used in the synthesis of araA and ddI by a one-pot, bienzymatic transglycosylation achieving 74 and 44 % conversion, respectively. Something different: Nucleoside phosphorylases are a convenient alternative to the chemical route for the synthesis of natural and modified nucleosides. Four new nucleoside phosphorylases have been prepared, characterized, and tested for their use in biocatalyzed syntheses of araA and ddI (see scheme). A generally applicable immobilization technique has been found to provide active and stable biocatalysts.

Full text of DOI:10.1002/cplu.201200278

CHEMPLUSCHEM
FULL PAPERS
DOI: 10.1002/cplu.201200278
Developing a Collection of Immobilized Nucleoside
Phosphorylases for the Preparation of Nucleoside
Analogues: Enzymatic Synthesis of Arabinosyladenine and
2
’,3’-Dideoxyinosine
[a]
[a]
[b]
[b]
[d]
Immacolata Serra,* Daniela Ubiali, Jure Pi sˇ kur, Stig Christoffersen,
Elizabeth S. Lewkowicz, Adolfo M. Iribarren, Alessandra M. Albertini, and Marco Terreni
[c]
[c]
[a]
The use of nucleoside phosphorylases (NPs; EC 2.4.2.n) repre-
sents a convenient alternative to the chemical route for the
synthesis of natural and modified nucleosides. We purified four
recombinantly expressed nucleoside phosphorylases from the
bacterial pathogens Citrobacter koseri, Clostridium perfringens,
and Streptococcus pyogenes (CkPNPI, CkPNPII, CpUP, SpUP) and
their substrate specificity was investigated towards either natu-
ral pyrimidine or purine nucleosides and some analogues,
namely, arabinosyladenine (araA) and 2’,3’-dideoxyinosine (ddI).
A 2–3% activity towards these latter compounds (compared to
the natural substrates) was observed. Enzyme activities were
compared to the specificities obtained for the enzymes pyrimi-
dine nucleoside phosphorylase from Bacillus subtilis (BsPyNP)
and purine nucleoside phosphorylase from Aeromonas hydro-
phila (AhPNPII) previously reported by some of the authors.
The enzymes displaying the suitable specificity for the synthe-
sis of araA and ddI were immobilized on aldehyde–agarose.
The immobilized preparations were highly stable at alkaline pH
and in the presence of methanol or acetonitrile as cosolvent.
They were used in the synthesis of araA and ddI by a one-pot,
bienzymatic transglycosylation achieving 74 and 44% conver-
sion, respectively.
Introduction
[
1]
Nucleoside phosphorylases (NPs; EC 2.4.2.n) are important in-
tracellular enzymes distributed widely in eukaryotes and pro-
karyotes. They act in the salvage of nucleobases and catalyze
the reversible conversion of (deoxy)ribonucleosides to their
corresponding free base and alfa-d-(deoxy)ribose-1-phosphate
in the presence of inorganic orthophosphate (phosphorolysis).
NPs are divided into two main structural classes: NPI and
NPII. Enzymes of the NPI family share a conserved structural
fold and include both the trimeric purine nucleoside phosphor-
ylase (PNPI) and the hexameric purine nucleoside phosphory-
lase (PNPII), and uridine phosphorylase (UP). Trimeric PNPI is
specific for guanine and hypoxanthine (2’-deoxy)ribonucleo-
sides, whereas hexameric PNPII also accepts adenine nucleo-
sides. The members of the NPII family (thymidine phosphory-
lase, TP, and pyrimidine nucleoside phosphorylase, PyNP) are
structurally distinct from the NPI class and are specific for pyri-
[2]
midine nucleosides.
The potency of NPs for the biocatalytic synthesis of nucleo-
side analogues is primarily due to the advantages that these
enzymes provide over the commonly used complex and multi-
stage chemical methods. In fact, the reaction reversibility of
NPs can be exploited to perform a transglycosylation reaction:
if a second nucleobase (sugar acceptor), either natural or not,
is added to the reaction medium, the formation of a new nu-
[3]
cleoside can result (Scheme 1). Most examples of nucleosides
synthesized through biocatalytic routes employ whole cells as
[3–7]
catalysts.
However, it has been demonstrated that isolated
and immobilized UP/PyNP and PNP, simultaneously used in
a so-called one-pot cascade reaction, show such an efficiency
that it could form the basis of nucleoside-production strat-
[
a] Dr. I. Serra, Dr. D. Ubiali, Prof. M. Terreni
Department of Drug Sciences and Italian Biocatalysis Center
University of Pavia
Via Taramelli 12, 27100 Pavia (Italy)
E-mail: immacolata.serra@unipv.it
[8,9]
egies.
Despite the efforts made so far extending the application of
NPs to large-scale production, nucleoside analogues are still
largely synthesized by “conventional” chemical methods. In
particular, limitations of NPs are currently related to their strict
substrate specificity (that narrows the number of possible ap-
[
b] Prof. J. Pi sˇ kur, Dr. S. Christoffersen
Department of Biology, Lund University
Soelvegatan 35, 22362 Lund (Sweden)
[c] Prof. E. S. Lewkowicz, Prof. A. M. Iribarren
Biotransformation Laboratory, Universidad Nacional de Quilmes
R. S. PeÇa 352, 1876 Bernal, Buenos Aires (Argentina)
[10]
plications), but protein stability is also a crucial issue: reac-
[d] Prof. A. M. Albertini
tion productivity greatly improves upon high temperatures,
high pH, and the presence of cosolvents ensuring solubiliza-
Department of Biology and Biotechnology, University of Pavia
Via Ferrata 1, 27100 Pavia (Italy)
[9]
tion of a large amount of substrates and products. For these
reasons, the research in this field focuses mainly on the iden-
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cplu.201200278.
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 157

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
uridine phosphorylase (CpUP), Streptococcus pyogenes
uridine phosphorylase (SpUP), Citrobacter koseri
purine nucleoside phosphorylase I (CkPNPI), and Cit-
robacter koseri purine nucleoside phosphorylase II
(
CkPNPII).
The immobilization on solid carriers of the most
promising enzymes has been investigated to obtain
stable biocatalysts suitable for easy and convenient
use in preparative processes. The immobilized biocat-
alysts were then exploited for the synthesis of the
two antiviral drugs arabinosyladenine (araA) and
Scheme 1. General scheme of the transglycosylation reaction catalyzed by NPs.
2
’,3’-dideoxyinosine (ddI) through one-pot transgly-
tification of new enzymes with suitable specificity and their
subsequent stabilization.
cosylation.
Conventional chemical syntheses of araA and ddI require
several protection and deprotection steps that result in tedious
procedures and moderate to poor yields of the target com-
Generally speaking, microorganisms are preferred to mam-
malian NP sources because of their higher substrate promiscui-
ty and, in recent years, a number of microbial NPs have been
[
22,23]
pounds.
In the case of ddI, for example, this nucleoside
[
1,11,12]
identified, cloned, and expressed.
In this context, microor-
can be synthesized by direct 5’-O-benzoylation of 2’-deoxyino-
sine, followed by deoxygenation at the 3’-position and benzo-
yl-group removal in about 20% yield overall. Alternative syn-
thetic routes have been developed over the years with the aim
to achieve more straightforward methods as well as safer and
ganisms from extreme environments are also being sought; it
is hardly surprising, for example, that thermostable enzymes
are attractive alternatives to their thermolabile counterparts in
[
22]
[
13–16]
many biocatalytic applications.
[
24–26]
However, the most widespread procedure for enzyme stabili-
zation is immobilization on solid supports. Immobilization of
biocatalysts also allows many other enzyme limitations to be
addressed, such as their solubility and reuse. In addition, when
properly designed, immobilization may be effective in modu-
cleaner reaction conditions.
In spite of a few advance-
ments, ddI is still synthesized through laborious schemes in
moderate to good yields (30–60%).
[
17]
lating activity or selectivity.
Results and Discussion
Only a few examples of the immobilization of NPs have
been reported so far. One of the most efficient techniques for
NP immobilization is ionic adsorption on carriers coated with
polyethylenimine (PEI) followed by cross-linking with aldehyde
dextran. This approach, which prevents enzyme desorption,
was successful in the case of thymidine phosphorylase from Es-
cherichia coli (EcTP) and pyrimidine nucleoside phosphorylase
from Bacillus subtilis (BsPyNP), both belonging to the NPII
Enzyme production and substrate specificity
Hypothetic genes coding four new nucleoside phosphorylases
from bacterial sources (Clostridium perfringens, Streptococcus
pyogenes, and Citrobacter koseri) were amplified from the ge-
nomic DNA using primers designed on the basis of the corre-
sponding genome sequences reported in the National Center
for Biotechnology Information (NCBI) database. After expres-
sion, all the proteins were purified by affinity chromatography.
Enzyme purity (>90%) was estimated by sodium dodecyl sul-
fate polyacrylamide gel electrophoresis (SDS-PAGE).
[
18,19]
family.
The adsorption on the highly flexible PEI-coated
carrier does not hamper the domain movement required
during catalysis and the multimeric quaternary structure is pre-
[18,20]
served by the final cross-linking with oxidized dextran.
Im-
The four recombinant enzymatic activities were evaluated
versus a range of nucleoside compounds (Scheme 2). Table 1
shows the enzymatic activities of the Clostridium perfringens
mobilized EcTP and BsPyNP have been conveniently assayed in
[
9,19]
enzymatic transglycosylation.
For many years, our group has been involved in research
aimed at developing a collection of solid and stable biocata-
lysts endowed with substrate specificities suitable for the syn-
thesis of nucleoside analogues. In this vein, the use of a PNP
from Aeromonas hydrophila (AhPNPII) in the preparative syn-
thesis of 6-modified purine ribonucleosides has been recently
Table 1. The activity of CpUP, SpUP, and BsPyNP towards pyrimidine sub-
strates. The percentage of conversion relative to the values for 2’-deoxy-
uridine are given in parentheses.
[a]
À1
À1
À1
Substrate
CpUP [IUmg ] SpUP [IUmg ] BsPyNP [IUmg ]
[
21]
reported.
(
%)
(%)
(%)
In the present study, the specificity of BsPyNP and AhPNPII
has been investigated further by considering nucleosides bear-
ing unnatural sugars useful for the synthesis of some chemo-
therapeutic drugs. In addition, the search for new enzymes
active on modified nucleosides has been considered. Thus,
four more microbial NP genes have been isolated, cloned, and
the recombinant enzymes expressed and studied. In particular,
the following enzymes were prepared: Clostridium perfringens
uridine (1)
60 (88)
68 (100)
31 (46)
3.0 (4.0)
11 (50)
22 (100)
18 (82)
0.4 (1.8)
1.7 (7.8)
n.a
20 (133)
15 (100)
6 (40)
n.a.
n.a.
n.a.
2’-deoxyuridine (2)
thymidine (3)
arabinosyluracil (4)
2’,3’-dideoxyuridine (5) 2.0 (3.0)
cytidine (6) n.a.
À1
[
a] IU: mmolmin . Experimental conditions: 50 mm phosphate buffer
pH 7.5, RT, 10 mL, [substrate] 5 mm; n.a.: not active.
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 158

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
CkPNPI can be described as specific towards inosine/guano-
sine as it was inactive with adenine nucleosides. This enzyme
showed the highest activity in the phosphorolysis of inosine
(
12) and a progressive reduction of its activity was registered
when first the hydroxyl group in the C-2’ and then in the C-3’
positions were removed. However, this enzyme retains a certain
activity also towards 2’,3’-dideoxyinosine (14; Table 2).
Immobilization of AhPNPII: A key study
Since AhPNPII had been proven already to have an outstand-
ing potential for synthetic purposes owing to its activity to-
[21]
wards modified purine ribonucleosides, and herein towards
arabino- and dideoxynucleoside analogues, it was selected for
the first immobilization trial. The immobilization techniques re-
sulting in a balance between activity and stability were then
applied also to CkPNPI and CpUP.
Scheme 2. Substrates (1–14) submitted to phosphorolysis catalyzed by NPs.
and the Streptococcus pyogenes UPs together with the values
obtained for the Bacillus subtilis PyNP enzyme.
Stability of the soluble enzyme
As expected, none of the phosphorylases tested could con-
vert cytidine (6). The CpUP and SpUP enzymes accepted all the
remaining naturally available pyrimidine nucleoside substrates.
It is worth noting that these enzymes also converted 2’,3’-di-
deoxyuridine (5) and arabinosyluracil (4), which bear unnatural
sugars. In comparison with 2’-deoxyuridine (2), the activity de-
creased by thirtyfold presumably as a consequence of the in-
version of the configuration at the C-2’ position or of the lack
of a hydroxyl group in C-3’ (Table 1).
The stability of the soluble enzyme was determined under sev-
eral conditions useful for immobilization, including the pres-
ence of stabilizing agents.
In the case of immobilization on aldehyde carriers, alkaline
conditions (pH 10) are required to promote an efficient
[27]
enzyme-carrier interaction. This procedure relies on the for-
mation of Schiff bases between the enzyme amino groups
(
that is mostly surface lysine (Lys)) and the support aldehydes,
In Table 2, the activities obtained for the purine NPs relative
to inosine (12) conversion are shown together with the values
for Aeromonas hydrophila PNPII. All PNPs showed high activi-
ties with ribo- and 2’-deoxyribonucleosides. PNPII from A. hy-
drophila and C. koseri presented very similar profiles for their
activities towards all the purine substrates tested and both en-
zymes accepted arabinosyladenine (9). These activities were
about forty times less than the activity obtained for adenosine
followed by reduction of the weak imino bonds with NaBH4
into stable CÀN bonds. When working at high pH, Lys residues
are not ionized and can thus react with the aldehyde groups
of the carrier. The results obtained are shown in Figure 1: at
pH 10 and room temperature the soluble enzyme loses 100%
of its activity after 1 h but, by lowering the temperature to
4
8C, an increase of stability was achieved.
The influence of glycerol, known as a stabilizing and preser-
(
7). The highest turnover for these enzymes was observed
[28,29]
vative compound,
was evaluated. The enzyme stability im-
using 2’-deoxyinosine (13).
proved in the presence of an increasing concentration of this
cosolvent (Figure 1). The presence of 20% glycerol (v/v) en-
sured the almost complete stability of AhPNPII at pH 10 and
room temperature.
Table 2. The activity of AhPNPII, CkPNPII, and CkPNPI towards purine sub-
[
a]
strates. The percentage of conversion relative to the values for inosine
are given in parentheses.
Substrate
AhPNPII
CkPNPII
[IUmg
(%)
CkPNPI
[IUmg
(%)
À1
À1
À1
[IUmg
%)
]
]
]
(
adenosine (7)
35 (46)
28 (37)
15 (45)
21 (63)
n.a.
n.a.
2’-deoxyadenosine (8)
arabinosyladenine (9)
0.9 (1.2)
26 (35)
28 (37)
75 (100)
137 (182)
0.8 (2.4)
11 (33)
46 (139)
33 (100)
108 (327)
0.16 (0.5)
n.a.
[
b]
guanosine (10)
’-deoxyguanosine (11)
inosine (12)
14 (32)
23 (50)
46 (100)
36 (78)
[
b]
2
2’-deoxyinosine (13)
2’,3’-dideoxyinosine (14)
0.02 (0.03)
1.6 (3.4)
Figure 1. Stability of soluble AhPNPII at pH 10: influence of the temperature
and glycerol. [AhPNPII]=0.06 mgmL (corresponding to 0.3 mg of enzyme
in the assay and 22.5 IU).
À1
[
a] IU: mmolmin . n.a.: not active. Experimental conditions: 50 mm phos-
À1
phate buffer pH 7.5, RT, 10 mL, [substrate] 5 mm. [b] 1 mm.
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 159

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
Immobilization
mobilized after 1.5 h at 48C (to ensure suitable enzyme stabili-
ty), but the expressed activity was very poor (7% yield; entry 4,
Table 3).
We explored different immobilization carriers and techniques,
while bearing in mind the multimeric structure expected for
AhPNPII. In fact, other NPs characterized by the same substrate
specificity (for adenine, hypoxanthine, and guanine nucleo-
Aldehyde–agarose was considered as the support and differ-
ent immobilization conditions, such as temperature and the
presence of stabilizing cosolvents, were assayed (Table 3). In all
cases, immobilization proceeds very fast and after 1.5 h almost
100% of the protein was immobilized. When immobilization
was performed at 48C for 1.5 h, a yield of 21% was achieved
(entry 5, Table 3). This value was threefold higher than that ob-
tained after immobilization on aldehyde–Sepabeads under the
same conditions. The carrier activation is equal, which high-
lights the detrimental effect of the hydrophobic nature of Se-
pabeads on the enzyme stability.
[
2]
sides) are reported to be hexameric. Both ionic adsorption
and covalent binding were taken into account for immobiliza-
tion. Moreover, post-immobilization cross-linking was consid-
ered to induce the stabilization of the multimeric structure of
the protein.
Adsorption on an ionic support followed by cross-linking
with aldehyde polymers was considered first, according to the
results previously obtained on the stabilization of other multi-
[
18,20]
meric enzymes.
The ionic interaction requires mild immo-
To enhance the formation of a high number of bonds be-
tween the support and the enzyme, two strategies were pur-
sued. First, the contact time at 48C was increased up to 3 h,
but afforded only 13% yield (entry 6, Table 3). This result can
be reasonably ascribed to the poor stability of the soluble
enzyme under these conditions, retaining the 60% of its initial
activity (Figure 1).
bilization conditions (5 mm phosphate buffer, pH 7.5) that do
not affect the enzyme activity. In particular, Sepabeads coated
with polyethylenimine (PEI) were selected as the ionic support
to combine very mild reaction conditions with the strong ionic
[
18]
interaction exerted by the polyamine. As reported in Table 3,
about 80% of the enzyme was adsorbed and the recovered ac-
Second, the immobilization
was performed at room temper-
[
a]
Table 3. Immobilization of AhPNPII.
ature in the presence of 20%
glycerol. This choice was ad-
Entry Support
Activation
T
pH
Time
Cross-linking Additives
[kDa]
h)
Immobilized Yield
[c]
dressed by the stability investi-
gation described above (see
Figure 1): the enzyme is com-
pletely stable after 3 h of incuba-
tion (that is, the immobilization
time) and retains 80% of its ini-
tial activity after 6 h. Following
the latter approach, a 20% im-
mobilization yield was achieved
[
8C]
protein
[%]
[%]
[
b]
(
1
2
3
4
5
6
7
8
9
1
Sepabeads
PEI
PEI
PEI
RT
7.5 20 min
7.5 20 min 100 (1)
7.5 20 min 500 (1)
10
10
10
20 (1)
–
78
83
74
13
21
12
7
Sepabeads
Sepabeads
Sepabeads
agarose
agarose
agarose
agarose
agarose
agarose
agarose
agarose
agarose
RT
RT
4
–
–
aldehyde
aldehyde
aldehyde
aldehyde
aldehyde
aldehyde
glutaraldehyde RT
glutaraldehyde RT
glutaraldehyde RT
1.5 h
1.5 h
3 h
3 h
3 h
3 h
3 h
3 h
3 h
–
–
93
4
4
–
–
100
96
21
13
20
30
8
–
–
RT 10
RT 10
RT 10
–
glycerol
100
100
100
100
96
[
[
d]
d]
–
glycerol, hpx
100 (6)
glycerol, hpx
(
entry 7, Table 3). However, the
0
8.5
7
8.5
8.5
–
–
–
–
–
2.5
0
crucial step is represented by
the final reduction with NaBH4
1
1
–
glycerol
–
1
1
2
3
100
100
3
17
glutaraldehyde
4
3 h
that is necessary, at the end of
À1
the process, to convert the
[
a] Load: 1 mgg . [b] Determined by the Bradford method. [c] On the basis of the expressed activity consider-
ing inosine as the reference substrate. [d] 5 mm.
imino bonds into stable CÀN
bonds. In fact, the enzyme re-
tained 45% of the initial activity
tivity after ionic adsorption was complete (results not shown).
The post-immobilization cross-linking was then performed
with aldehyde dextran. To ensure a full coating of all the
enzyme subunits we studied dextran with different molecular
weights. The best results were obtained with 100 kDa dextran
that afforded a 20% recovered activity after cross-linking
after 3 h of interaction with the activated carrier (Figure 1 in
the Supporting Information) but chemical reduction decreased
the final yield by half (Table 3).
Owing to the promising results achieved with aldehyde–
agarose, this immobilization protocol was subjected to a further
optimization. It is known that the addition of an inhibitor or
a substrate to the immobilization suspension can prevent pos-
sible distortions of the enzyme structure upon interaction with
the activated carrier. In fact, the ligand can shield the active
site by binding to it during immobilization. In this case, when
hypoxanthine was added to the immobilization medium, after
3 h, the yield increased to 30% (entry 8, Table 3). The cross-
linking of the resulting preparation with oxidized 100 kDa dex-
tran was not beneficial to the outcome of the process and the
yield was very low (8%; entry 9, Table 3).
(entry 2, Table 3), whereas with higher and lower molecular
weights the immobilization yields decreased (13 and 12% with
20 and 500 kDa dextran, respectively; entries 1 and 3, Table 3).
Among covalent linkages, the attention was focused on al-
dehyde activation since it was shown to be suitable for the im-
[
27,30]
mobilization/stabilization even of multimeric enzymes
in-
[
18]
cluding trimeric PNPI from Bacillus subtilis. Both hydrophobic
and hydrophilic matrices were considered. When using the hy-
drophobic aldehyde–Sepabeads, 93% of the protein was im-
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 160

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
As a final approach, agarose activated with glutaraldehyde,
which relies on a very mild immobilization procedure, was con-
sidered. Although the soluble enzyme is completely stable in
the immobilization buffer (data not shown), only the immobili-
zation performed at 48C gave positive results, with an immobi-
lization yield of 17% (entry 13, Table 3). Differently from alde-
hyde–agarose, the activity of the suspension, monitored
during immobilization on glutaraldehyde–agarose, drastically
decreased within the first hour of immobilization (Figure 1 in
the Supporting Information). This fact leads us to hypothesize
that the contact with the glutaraldehyde–agarose somehow in-
duces the enzyme inactivation.
Immobilization of CkPNPI and CpUP
The aldehyde–agarose was found to be the best matrix so its
applicability for the immobilization of other multimeric NPs
(CpUP and CkPNPI) was studied. Both enzymes in the native
form were stable, even in the absence of stabilizing agents, at
pH higher than 10 up to 24 h (data not shown). For CpUP, the
immobilization yield was 33% at room temperature and pH 10.
The presence of either uracil or 20% glycerol in the immobili-
zation medium did not have any effect on the immobilization
outcome. Also for this enzyme the glutaraldehyde activation
afforded very poor results (Table 4). Immobilization of CkPNPI
provided higher yields than the other NPs, affording 51 and
35% of recovered activity for aldehyde and glutaraldehyde ac-
tivation, respectively.
Stability of the immobilized preparations
The immobilized preparations were tested for their stability at
pH 10 (see Figure 2 and Table 3). In spite of what was expect-
ed, AhPNPII immobilized on Sepabeads–PEI (entry 2, Table 3)
and covered with 100 kDa dextran showed very poor stability,
Table 4. Immobilization of CpUP and CkPNPI on aldehyde- and glutaral-
dehyde-activated agarose.
Enzyme Activation
Load
Additives Immobilized Yield
À1
[b]
[
mgg
]
protein
[%]
[
a]
[
%]
CpUP
CpUP
CpUP
CpUP
CkPNPI
CkPNPI
aldehyde
aldehyde
aldehyde
1.4
1.5
1.2
1.0
0.9
1.2
glycerol
–
uracil
–
glycerol
–
94
23
33
35
4
51
35
100
100
100
100
98
[
c]
[
[
d]
d]
glutaraldehyde
aldehyde
glutaraldehyde
[
a] Determined by the Bradford method. [b] On the basis of the expressed
activity considering as reference substrates inosine and 2’-deoxyuridine
for PNP and UP, respectively. [c] 5 mm. [d] Experimental conditions:
100 mm phosphate buffer, pH 8.5 and 48C.
Figure 2. Stability at pH 10 of the soluble and immobilized AhPNPII (referring
to Table 3 entries).
The results obtained with aldehyde–agarose are consistent
with those of AhPNPII and agree with those previously report-
[
18]
slightly higher than that of the soluble enzyme; the aldehyde–
agarose derivative immobilized at 48C for 1.5 h (entry 5,
Table 3) displayed the same behavior. On the contrary, the
same derivative immobilized at room temperature for 3 h in
the presence of 20% glycerol (entry 8, Table 3) showed very
high stability, retaining completely its activity after 6 h.
ed for BsPNPI.
In terms of the stability in organic solvents, both aldehyde
preparations of CkPNPI and CpUP maintained 100% of their in-
itial activity after incubation for 24 h in the presence of 20%
MeCN and MeOH (results not shown).
The higher immobilization yields obtained for PNPI, relative
to CpUP and AhPNPII, may be ascribed to the simpler quater-
These data confirm that the length of time of contact and
the reaction temperature are important parameters that need
to be considered in the design of a covalent immobilization. In
fact, by increasing the temperature and reaction time a multi-
point attachment is induced, which leads to an increase in sta-
bility of the immobilized preparation.
[
1]
nary structure of the enzymes belonging to this class.
Overall, the high performance of agarose as immobilization
carrier for NPs regardless of the complexity of their quaternary
structure can be explained by invoking the internal morpholo-
gy of agarose beads. It can be assumed that this support is
formed by fibers with a diameter thicker than a protein and,
thus, it is able to offer both large surfaces for protein-support
multi-interactions and to enhance a better geometric congru-
Moreover, the stability of the best aldehyde–agarose prepa-
ration (see entry 8 in Table 3) was tested also in the presence
of organic solvents, 20% acetonitrile (MeCN) or methanol
[27]
(MeOH), and compared with that of the soluble counterpart.
ence between a large multimeric enzyme and the support.
The immobilized enzyme was found to be several times more
stable than the native one, in particular when incubated in the
presence of MeCN (Figure 2 in the Supporting Information).
However, in general, MeOH was a better solvent than MeCN
for enzyme stability.
The different strategies developed for NPI and NPII can be
tentatively explained by considering the diversity of catalytic
requirements between these classes. In fact, NPsII have been
successfully immobilized on a flexible ionic carrier followed by
[18,19]
cross-linking with aldehyde dextran.
This result may be in-
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 161

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
terpreted as being a consequence of the peculiar catalytic
mechanism of the enzymes belonging to this class. For in-
stance, the crystal structure of EcTP reveals an asymmetric
dimer in which one subunit is in an open conformation and
the second is in a closed substrate-bound form. The reaction
mechanism is reported to involve significant movement of
structural domains for the open-
Synthesis of araA (9) and ddI (14)
By selecting the enzymes with the suitable substrate specificity,
the one-pot syntheses of araA (9) and ddI (14) were carried
out using araU and ddU as sugar sources, respectively
(Scheme 3 and Table 6). The effect of a cosolvent on the maxi-
ing and closing of each subunit.
The immobilization approach on
Sepabeads–PEI likely allows the
preservation of this domain
movement (and the catalytic ac-
tivity thereof). On the other
hand, NPsI, which do not require
any domain movement, were ef-
ficiently immobilized and stabi-
lized through the more con-
straining covalent bonds on al-
dehyde agarose.
We also wondered whether
the His6 tag might influence the
immobilization yield and, toward
Scheme 3. Synthesis of araA and ddI by transglycosylation catalyzed by NPs.
this aim, new experiments were carried out. The His6 tag was
cleaved from the purified AhPNPII with Ac-TEV protease (see
the Experimental Section) and the recovered wild-type protein
Table 6. One-pot bienzymatic transglycosylation reaction for the synthe-
sis of araA and ddI.
[
a]
(
which had a specific activity comparable with that of the
[b]
Donor
(mm)
Acceptor
(mm)
Solvent
–
Product
% Conv.
(h)
tagged enzyme) was immobilized on aldehyde–agarose. The
cleaved protein showed similar percentages of immobilized
protein (determined by Bradford method) and final yield (per-
centage of expressed activity) to the tagged enzyme (Table 5).
4
(4)
Ade (2)
Ade (2)
Ade (2)
Ade (5)
Ade (5)
Hpx (2)
Hpx (2)
Hpx (2)
Hpx (5)
9
9
9
9
9
14
14
14
14
63 (48)
63 (48)
69 (48)
71 (48)
74 (48)
26 (24)
30 (24)
34 (24)
44 (24)
4 (4)
4 (4)
MeOH
MeCN
MeOH
MeCN
–
MeOH
MeCN
MeCN
4
4
5
5
5
5
(10)
(10)
(4)
(4)
(4)
Table 5. Optimization of the immobilization on aldehyde–agarose.
Enzyme
Load
Additives
Immobilized
protein [%]
Yield
[%]
(10)
À1
[a]
[b]
[mgg
]
[
a] Experimental conditions: 50 mm phosphate buffer, pH 7.5, RT, 10 mL;
CpUP (6 IU): in the synthesis of araA and ddI AhPNPII (12 IU) and CkPNPI
12 IU) were used, respectively; one IU is referred to inosine for AhPNPII
AhPNPII non-tagged
AhPNPII
AhPNPII
AhPNPII
CpUP non-tagged
CpUP
CpUP
CkPNPI
CkPNPI
0.1
0.9
1.0
5.0
1.5
1.0
5.0
1.0
5.0
–
–
77
100
100
100
100
100
100
100
96
17
21
30
17
33
41
19
51
43
(
glycerol, hpx
glycerol, hpx
and CkPNPI and to 2’-deoxyuridine for UP. [b] 10% v/v.
–
–
–
glycerol
glycerol
mum conversion and reaction rate was evaluated: the addition
of organic solvents is indeed often necessary to ensure the
complete solubilization of high concentrations of reagents and
products. Thus, according to the stability of the different en-
zymes, MeOH and MeCN were tested; the former showed no
relevant influence on the percentage of product formed,
whereas MeCN induced a slight increase of the conversion,
which was more evident for ddI. Under these conditions, and
increasing the substrate concentration, the conversions ach-
ieved were 74 and 44% of araA and ddI, respectively (Table 6).
[
a] Determined by the Bradford method. [b] On the basis of the expressed
activity considering as reference substrates inosine and 2’-deoxyuridine
for PNP and UP, respectively.
Similar outcomes were obtained with the native CpUP ob-
tained by cleavage of the glutathione S-transferase (GST) tag
with thrombin. We can thus conclude that the presence of the
tag does not influence the immobilization process on alde-
hyde–agarose.
Conclusion
Afterwards, the load of protein per gram of support was in-
creased from 1 to 5 mg. As expected, in all cases, the ex-
pressed activity was sensitively reduced when a high amount
of protein was offered per each gram of carrier (Table 5).
New nucleoside phosphorylases have been prepared, partly
characterized, and tested for their use in the syntheses of araA
and ddI. Uridine phosphorylases from the bacterial pathogens
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 162

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
C. perfringens and S. pyogenes were shown to convert arabino-
syluracil and 2’,3’-dideoxyuridine. In addition, purine nucleo-
side phosphorylases from A. hydrophila and C. koseri showed
some activity for arabino- and 2’,3’-dideoxypurine nucleosides,
which indicates that transglycosylation reactions for the forma-
tion of nucleoside analogues can be performed by a mixture
of CpUP or SpUP and any of these PNPs. Pyrimidine nucleoside
phosphorylase from B. subtilis did not show activity on pyrimi-
dine nucleosides bearing arabinose and 2’,3’-dideoxyribose as
sugar moieties. The studied enzymes (with the exception of
CkPNPII and SpUP) were successfully immobilized and stabi-
lized by covalent attachment on aldehyde-activated agarose.
The immobilization protocol described here allowed the stabili-
zation of all the studied enzymes regardless of their multimeric
asset. Immobilization of NPI on aldehyde–agarose can be con-
sidered, therefore, a technique of general applicability within
this class of enzymes. The results reported in this paper, in ad-
proteases, respectively. The putative NP genes were identified by
[32]
using the basic local sequence alignment tool (BLAST). The cor-
responding open reading frames were amplified from the genomic
DNA by the polymerase chain reaction (PCR) using PfuUltra DNA
polymerase (Stratagene) and primers (listed in the Supporting In-
formation) with overhangs specific for cloning into the pET/TOPO
1
51 vector (Invitrogen).
The N-terminal His6 tag fusion proteins were purified according to
[21]
the procedure reported by Ubiali et al. The final concentration of
the protein preparation determined by means of the Bradford
À1
method was in the range of 2–5 mgmL .
In addition, CpUP was also cloned into a pGEX-2T vector (GE
Healthcare). The primers used are reported in the Supporting Infor-
mation.
Purification and cleavage with thrombin of GST-CpUP for
preparation of native CpUP
[
18,19,21]
CpUP over-expressed as the fusion protein with an N-terminal GST
fusion tag was purified by glutathione–Sepharose affinity chroma-
tography followed by thrombin cleavage. Purification was per-
formed by slightly modifying the procedure reported by Knecht
dition to our previous work,
provide a collection of
stable solid biocatalysts for the enzymatic synthesis of modi-
fied nucleosides by transglycosylation. By selecting an appro-
priate enzyme within this collection (depending on the struc-
ture of the substrates) sugar- or base-modified nucleosides can
be prepared. In the case of araA and ddI, the reaction optimi-
zation, scale-up, and product isolation are in progress.
[33]
et al.
Briefly, the crude extract was loaded onto a 1 mL GSH
column pre-equilibrated in binding buffer (20 mm Na PO , pH 7.3;
3
4
150 mm NaCl; 10% glycerol and 0.1% polyethylene glycol p-
1,1,3,3-tetramethylbutyl)phenyl ether (Triton X-100)); then the
(
column was washed with binding buffer (20 mL). Thrombin cleav-
age was performed by applying binding buffer (1 mL) containing
thrombin (50 IU) on the column and recirculating O/N at 48C. The
native CpUP was eluted from the column with binding buffer.
Experimental Section
General
All nucleosides, heterocyclic bases, solvents, polyethylenimine (PEI),
glutaraldehyde 25% (v/v), ethylendiamine (EDA), and dextran were
purchased from Sigma Aldrich and/or VWR International (Milano,
Italy). Sepabeads were a gift from Resindion (Mitsubishi, Binasco,
Italy). Agarose CL6B was purchased from Pharmacia Biotech AB
TM
Cleavage with Ac-TEV protease for preparation of native
AhPNPII
In a total volume of 1 mL, the tagged protein (0.6 mg) was sus-
pended in tris(hydroxymethyl)aminomethane hydrochloride (Tris
HCl; 50 mm), pH 8.0, containing ethylenediaminetetraacetic acid
(
Uppsala, Sweden). All solvents were HPLC grade. Enzymatic reac-
tions were monitored by using a HPLC Merck Hitachi L-7100 instru-
ment equipped with a UV detector L-7400 and column oven L-
(
EDTA; 0.5 mm) and dl-dithiothreitol (DTT; 1 mm). Then Ac-TEV
protease (20 IU) was added; the mixture was incubated under O/N
with gentle stirring at 48C. The His6-tag fragment and the Ac-TEV
protease were removed by loading the mixture on Protino Ni-TED
7
300 (Darmstadt, Germany). PfuUltra DNA polymerase was from
Stratagene (Agilent Technologies, Life Sciences and Chemical Anal-
ysis, Waldbronn, Germany). The pET/TOPO 151 vector was pur-
chased from Invitrogen (San Giuliano Milanese, Italy) and the re-
combinant proteins were purified using Protino Ni-TED 1000
packed columns (MACHEREY-NAGEL, Dꢁren, Germany) and a gluta-
thione (GSH) column (Amersham Pharmacia, Uppsala, Sweden).
The protein-concentration assay was performed on a Shimadzu
1
000 packed columns. The cleaved native protein was in the flow-
through fractions. The collected fractions were analyzed by SDS-
PAGE and the protein concentration was determined by the Brad-
ford method.
[31]
spectrophotometer UV 1601 by using the Bradford method
using bovine serum albumin as standard. PyNP from Bacillus subti-
HPLC standard activity assay (phosphorolysis)
[18,21]
A solution of potassium buffer pH 7.5 (50 mm, 10 mL) containing
the desired concentration of nucleoside substrate was prepared.
The reaction was started by the addition of the enzyme (soluble or
immobilized) and kept under mechanical stirring. At different
times, samples were withdrawn, filtered (either by centrifugation
with 10 kDa molecular weight cutoff (MWCO) Nanosep at 48C and
lis and AhPNPII were prepared as previously reported.
All ex-
periments were performed at least in duplicate. See the Support-
ing Information for details of the bacterial strains. C. koseri genom-
ic DNA was extracted according to the protocol previously used
[21]
for A. hydrophila.
1
3200 rpm for 2 min for the soluble enzymes, or by using pipette-
filter devices for the immobilized enzymes), and analyzed by HPLC.
The column was a Gemini 250ꢂ4.6 mm, 5 mm (Phenomenex, Cer-
nusco sul Naviglio, Italy); eluent: 0.01m potassium phosphate
buffer pH 4.6 and methanol 90% (95:5); flow: 1 mLmin ; T=
358C; l=260 nm.
Cloning, expression and purification of the recombinant
proteins
À1
AhPNPII, CpUP, SpUP, CkPNPI, and CkPNPII were prepared as fusion
proteins with an N-terminal His6 tag; CpUP was obtained also as
a GST fusion protein. In addition, AhPNPII and CpUP were prepared
as native proteins by cleavage of the tags with Ac-TEV or thrombin
One international unit (IU) of the enzymatic activity corresponds to
the amount of enzyme that hydrolyzes 1 mmol of substrate per
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 163

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
minute at room temperature. The specific activity is defined as
units of enzyme activity per milligram of protein.
and the mixture was kept at room temperature under mechanical
stirring until the highest conversion was achieved. The reaction
was monitored by HPLC (l=260 nm), which was used to identify
the products by comparing their retention times with those of au-
thentic samples. Mobile phase: 0.01m KH PO buffer pH 4.6/metha-
Preparation of the activated carriers
2
4
nol 90% (90:10). Uracil, t =3.85 min; hypoxanthine, t =4.68 min;
R
R
Preparation of aldehyde carriers: Aldehyde–agarose was prepared
adenine, t =7.42 min; arabinosyluracil (4), t =5.62 min; 2’,3’-di-
[34]
R
R
as previously reported. Aldehyde–Sepabeads were prepared by
oxidation of Sepabeads EC-HG following the previously reported
deoxyuridine (5), tR =11.55 min; arabinosyladenine (9), tR =
2.33 min; 2’,3’-dideoxyinosine (14), t =16.35 min. The percentage
1
[35]
R
protocol.
of conversion was calculated on the basis of the depletion of the
sugar acceptor compound (heterocyclic base) and monitoring the
formation of the nucleoside products: Conversion [%]=[product
area/(product area+base area)]ꢂ100. The reaction was stopped
by sub vacuum filtration of the immobilized biocatalyst.
Activation of agarose with glutaraldehyde: The amination of alde-
hyde–agarose with ethylendiamine (EDA) was performed as previ-
[36]
ously reported. The EDA-activated agarose was then suspended
in phosphate buffer pH 7 (0.2m, 3.4 mL) and a solution of 25% (v/
v) glutaraldehyde (5.1 mL) was added. The mixture was kept under
stirring for 16 h at room temperature in the darkness. The activat-
ed support was washed with deionized water and used immediate- Acknowledgements
ly.
This study was supported by Regione Lombardia-Metadistretti
Call 2007 (MD 2007 ID 4165).
Preparation of Sepabeads–PEI: Activation of epoxy Sepabeads with
PEI was performed as previously reported.
[18]
Keywords: biocatalysis
nucleosides · transglycosylation
·
enzymes
·
immobilization
·
Immobilization of enzymes
Immobilization on aldehyde-activated carriers: Immobilization on
carriers activated with aldehyde groups was performed by slightly
[
1] M. J. Pugmire, S. E. Ealick, Biochem. J. 2002, 361, 1–25.
[19,35]
modifying the procedure previously reported.
Briefly, the alde-
[2] M. J. Pugmire, W. J. Cook, A. Jasanoff, M. R. Walter, S. E. Ealick, J. Mol.
Biol. 1998, 281, 285–299.
[3] E. S. Lewkowicz, A. M. Iribarren, Curr. Org. Chem. 2006, 10, 1197–1215.
hyde support (1.4 mL) was suspended in potassium phosphate/car-
bonate buffer (50 mm) at pH 10.05. After the addition of the de-
sired amount of protein, the suspension (14 mL) was kept under
mechanical stirring. Chemical reduction of Schiff bases was carried
[
[
[
[
4] M. Nꢃbile, M. Terreni, E. Lewkowicz, A. M. Iribarren, Biocatal. Biotrans-
form. 2010, 28, 395–402.
5] C. W. Rivero, C. N. Britos, M. E. Lozano, J. V. Sinisterra, J. A. Trelles, FEMS
Microbiol. Lett. 2012, 331, 31–36.
6] E. Spoldi, D. Ghisotti, S. Calꢄ, M. Grisa, G. Orsini, G. Tonon, G. Zuffi, Nu-
cleosides Nucleotides Nucleic Acids 2001, 20, 977–979.
out over 30 min by adding NaBH (14 mg) to the mixture. The im-
4
mobilized enzyme was then filtered and washed with 10 mm po-
tassium phosphate buffer pH 5 and deionized water.
7] I. Votruba, A. Holy, H. Dvorakova, J. Gunter, D. Hockova, H. Hrebabecky,
T. Cihlar, M. Masojidkova, Collect. Czech. Chem. Commun. 1994, 59,
Immobilization on glutaraldehyde–agarose: The glutaraldehyde–
agarose gel (1 g), prepared as described above, was suspended in
potassium phosphate buffer (100 mm) at the desired pH. After the
addition of the desired amount of protein, the solution (7.1 mL)
was kept under mechanical stirring; after 3 h the pH was increased
2303–2330.
[
[
8] I. A. Mikhailopulo, Curr. Org. Chem. 2007, 11, 317–335.
9] D. Ubiali, S. Rocchietti, F. Scaramozzino, M. Terreni, A. M. Albertini, R.
Fernꢅndez-Lafuente, J. M. Guisꢅn, M. Pregnolato, Adv. Synth. Catal.
2004, 346, 1361–1366.
to 10.05 and NaBH (7 mg) was then added. The reduction time
4
was 30 min. Finally, the immobilized enzyme was filtered and
washed with 10 mm potassium phosphate buffer pH 5 and deion-
ized water.
[10] I. A. Mikhailopulo, A. I. Miroshnikov, Mendeleev Commun. 2011, 21, 57–
68.
[
11] N. H. Martins, A. N. Meza, C. R. Santos, P. O. de Giuseppe, M. T. Murakami,
Acta Crystallogr. Sect. F 2011, 67, 618–622.
Immobilization on Sepabeads–PEI and cross-linking with aldehyde
dextran: Immobilization on Sepabeads–PEI and stabilization by
cross-linking with 20% oxidized dextran (aldehyde dextran) was
[12] X. Xie, J. Xia, K. He, L. Lu, Q. Xu, N. Chen, Biotechnol. Lett. 2011, 33,
1107–1112.
[
13] B. Sonnleitner, A. Fiechter, Trends Biotechnol. 1983, 1, 74–80.
[19]
[14] J. A. Littlechild, J. Guy, S. Connelly, L. Mallett, S. Waddell, C. A. Rye, K.
performed according to the procedure previously described.
Line, M. Isupov, Biochem. Soc. Trans. 2007, 35, 1558–1563.
[
[
[
[
[
[
15] S. A. Taran, K. N. Verevkina, S. A. Feofanov, A. I. Miroshnikov, Russ. J.
Bioorg. Chem. 2009, 35, 739–745.
16] S. A. Taran, K. N. Verevkina, T. Z. Esikova, S. A. Feofanov, A. I. Miroshnikov,
Appl. Biochem. Microbiol. 2008, 44, 162–166.
17] C. Mateo, J. M. Palomo, G. Fernꢅndez-Lorente, J. M. Guisꢅn, R. Fernꢅn-
dez-Lafuente, Enzyme Microb. Technol. 2007, 40, 1451–1463.
18] S. Rocchietti, D. Ubiali, M. Terreni, A. M. Albertini, R. Fernꢅndez-Lafuente,
J. M. Guisꢅn, M. Pregnolato, Biomacromolecules 2004, 5, 2195–2200.
19] I. Serra, C. D. Serra, S. Rocchietti, D. Ubiali, M. Terreni, Enzyme Microb.
Technol. 2011, 49, 52–58.
Enzyme-stability assay
Following a general procedure, the enzyme was added to buffer
5 mL; under the desired conditions) and kept under mechanical
(
stirring at room temperature. At different times, samples were
withdrawn and the residual enzymatic activity was monitored by
a standard activity assay towards inosine.
20] R. Fernandez-Lafuente, Enzyme Microb. Technol. 2009, 45, 405–418.
General procedure of enzymatic transglycosylation
[21] D. Ubiali, C. D. Serra, I. Serra, C. F. Morelli, M. Terreni, A. M. Albertini, P.
Manitto, G. Speranza, Adv. Synth. Catal. 2012, 354, 96–104.
22] R. R. Webb, J. A. Wos, J. C. Martin, P. R. Brodfuehrer, Nucleosides Nucleo-
tides 1988, 7, 147–153.
[23] W. W. Lee, A. Benitez, L. Goodman, B. R. Baker, J. Am. Chem. Soc. 1960,
82, 2648–2649.
A solution of phosphate buffer (10 mL, 50 mm) with the appropri-
ate amount of solvent (v/v) at pH 7.5 containing araU or ddU (4
and 5) and the heterocyclic base at the desired concentration was
prepared. The enzyme preparations were added to the reaction
[
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 164

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
[
[
[
[
24] V. Bhat, E, Stocker, B. G. Ugarkar, Synth. Commun. 1992, 22, 1481–
486.
25] Z. Wang, D. R. Prudhomme, J. R. Buck, M. Park, C. J. Rizzo, J. Org. Chem.
000, 65, 5969–5985.
26] T. Torii, K. Izawa, D. H. Cho, D. O. Jang, Nucleosides Nucleotides Nucleic
Acids 2007, 26, 985–988.
27] C. Mateo, J. M. Palomo, M. Fuentes, L. Betancor, V. Grazu, F. Lꢆpez-Galle-
go, B. C. C. Pessela, A. Hidalgo, G. Fernꢅndez-Lorente, R. Fernꢅndez-La-
fuente, J. M. Guisꢅn, Enzyme Microb. Technol. 2006, 39, 274–280.
28] K. Ruan, C. Xu, T. Li, J. Li, R. Lange, C. Balny, Eur. J. Biochem. 2003, 270,
[31] M. M. Bradford, Anal. Biochem. 1976, 72, 248–254.
[32] S. F. Altschul, W. Gish, W. Miller, E. W. Myers, D. J. Lipman, J. Mol. Biol.
1990, 215, 403–410.
1
2
[33] W. Knecht, E. Rozpedowska, C. Le Breton, M. Willer, Z. Gojkovic, M. P. B.
Sandrini, T. Joergensen, L. Hasholt, B. Munch-Petersen, J. Piskur, Gene
Ther. 2007, 14, 1278–1286.
[34] J. M. Guisꢇn, Enzyme Microb. Technol. 1988, 10, 375–382.
[35] C. Temporini, P. Bonomi, I. Serra, A. Tagliani, T. Bavaro, D. Ubiali, G. Mas-
solini, M. Terreni, Biomacromolecules 2010, 11, 1623–1632.
[36] R. Fernandez-Lafuente, C. M. Rosell, V. Rodriguez, C. Santana, G. Soler, A.
Bastida, J. M. Guisꢇn, Enzyme Microb. Technol. 1993, 15, 546–550.
[
[
[
1
654–1661.
29] V. V. Mozhaev in Progress in Biotechnology (Ed.: A. Ballesteros), Elsevier,
998, pp. 355–363.
1
30] J. M. Bolivar, F. Cava, C. Mateo, J. Rocha-Martꢄn, J. M. Guisꢅn, J. Beren-
guer, R. Fernꢅndez-Lafuente, Appl. Microbiol. Biotechnol. 2008, 80, 49–
Received: October 22, 2012
58.
Published online on December 19, 2012
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 157 – 165 165