A versatile synthesis of 5'-fenctionalized nucleosides through regioselective enzymatic hydrolysis of their peracetylated precursors

Article abstract of DOI:10.1002/ejoc.200801096

We describe a chemo-enzymatic synthesis of modified nucleosides through lipase-catalyzed hydrolysis of their peracetylated precursors. It was found from screening of a large number of substrates that these enzymesregioselectivities were affected by the sugar and the nucleobase structures. By selecting the best enzyme for each substrate in terms of activity and regioselectivity, we prepared a small library of differently monodeprotecled purine and pyrimidine nucleosides useful as intermediates for the synthesis of high-value nucleosides and mononucleotides. By this approach, the chemo-enzymatic preparation of doxifluridine (14) and uridine 5'-monophosphate (5'-UMP, 15) from peracetylated uridine 1 was carried out. Elimination of many of the processing stages associated with existing methods was achieved, and higher yields and products of increased purity were generated. Wiley-VCH Verlag GmbH & Co. KGaA.

Full text of DOI:10.1002/ejoc.200801096

FULL PAPER
DOI: 10.1002/ejoc.200801096
A Versatile Synthesis of 5Ј-Functionalized Nucleosides Through Regioselective
Enzymatic Hydrolysis of Their Peracetylated Precursors
Teodora Bavaro,[a,b,c] Silvia Rocchietti,[c,b] Daniela Ubiali,*[a,b] Marco Filice,[a,b]
Marco Terreni,[a,b] and Massimo Pregnolato*[a,b]
Keywords: Immobilization / Lipases / Nucleosides / Regioselectivity / Enzymes
We describe a chemo-enzymatic synthesis of modified nu-
cleosides through lipase-catalyzed hydrolysis of their per-
acetylated precursors. It was found from screening of a large
number of substrates that these enzymes’ regioselectivities
were affected by the sugar and the nucleobase structures. By
selecting the best enzyme for each substrate in terms of ac-
tivity and regioselectivity, we prepared a small library of dif-
ferently monodeprotected purine and pyrimidine nucleosides
useful as intermediates for the synthesis of high-value nu-
cleosides and mononucleotides. By this approach, the
chemo-enzymatic preparation of doxifluridine (14) and
uridine 5Ј-monophosphate (5Ј-UMP, 15) from peracetylated
uridine 1 was carried out. Elimination of many of the pro-
cessing stages associated with existing methods was
achieved, and higher yields and products of increased purity
were generated.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
ities, selective methods that do not produce isomers would
be desirable. This would also allow better control over by-
product content in the final compounds.
Introduction
The importance of nucleosides and nucleotides in dif-
ferent market sectors (pharmaceutical, food, biotechnol-
ogy…) has led to various approaches for their synthesis.
Current methodologies either employ natural nucleosides as
starting materials or are based on convergent approaches
involving condensation of the carbohydrate precursors and
the heterocyclic bases. In both cases chemo-, regio-, and
stereoselectivities are ensured through the use of protecting
groups and specific deprotection reactions. As a result,
chemical syntheses of nucleoside analogues involve many
stages and are often low-yielding or unreliable, particularly
for modified nucleosides.[1,2] Additionally, the resultant
products may not be sufficiently clean for downstream reac-
tions in spite of complex purification processes.
One of the main fields of nucleoside research explored in
the past has involved nucleoside analogues as antitumor
and antiviral agents.[3,4] However, interest in modified nu-
cleosides, also used as synthons for the preparation of these
drugs, is still growing, so it is not surprising that novel and
more efficient syntheses are being sought.[5] Since nucleo-
sides possess several functional groups with similar reactiv-
Enzymatic syntheses have been generally shown to be an
advantageous alternative both to chemical methods[6] and
to the coupling of chemical methods and biochemical trans-
formations. This strategy often provides synthetic routes
with fewer steps and improved overall synthetic efficiency in
yields and regio- and stereoselectivities. Enzymes’ inherent
selectivities, as well as their ability to function under mild
conditions, can assist easier obtainment of pure products.
Enzymes can be successfully exploited both for nucleoside
functionalization[5] and for glycosidic bond formation.[6,7]
Suitable introduction and removal of protecting groups
is often critical. In nucleoside chemistry, particularly in the
sugar moieties, this problem is accentuated by the presence
of multiple hydroxy functions with similar reactivities. In
this case, stereo- and regioselective acylation and deacyla-
tion of the hydroxy groups can be achieved by using lipases
as biocatalysts. Lipases (glycerol ester hydrolases,
E.C. 3.1.1.3) are extremely versatile thanks to their effi-
ciency, easy workup, and stability in aqueous and organic
solvents.[8,9] Indeed, many examples of lipase-catalyzed de-
acylation of protected nucleosides are known.[5,10] In par-
ticular, when peracetylated nucleosides are used as sub-
strates, it has been reported that ribonucleosides are gen-
erally deprotected at the primary hydroxy group, whereas
in the case of the 2Ј-deoxyribo counterparts the 3Ј-position
is preferentially hydrolyzed. However, only a few reports
have so far been published on the influence either of non-
natural sugars (e.g., arabinose) or of the nucleobase (mostly
pyrimidines) on the enzymes’ regioselectivities. Moreover,
[a] Dipartimento di Chimica Farmaceutica, Pharmaceutical Bioca-
talysis Laboratories, Università degli Studi di Pavia,
12 via Taramelli, 27100 Pavia, Italy
Fax: +39-0382-422975
E-mail: maxp@ibiocat.eu
dany@ibiocat.eu
[b] Italian Biocatalysis Center, Dipartimento di Chimica Farma-
ceutica, 12 via Taramelli, Università degli Studi di Pavia, 12 via
Taramelli, 27100 Pavia, Italy
[c] Innovate Biotechnology srl, Parco Scientifico Tecnologico,
Strada Savonesa 9, 15050 Rivalta Scrivia (AL), Italy
Eur. J. Org. Chem. 2009, 1967–1975
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1967

Daniela Ubiali, Massimo Pregnolato et al.
FULL PAPER
all those studies, with few exceptions,[11] have been per- domonas cepacia lipase (PCL), and porcine pancreas lipase
formed with the native enzymes. (PPL)] were immobilized on a hydrophobic support (octyl-
In this work, lipases from different biological sources sepharose) (see Materials and Methods). The enzyme prep-
were immobilized on hydrophobic supports[12] and used for arations were then tested in the hydrolysis of different nucle-
the regioselective hydrolysis of several peracetylated nucleo- oside esters, by studying the influence of the sugar and the
sides, including some substrates never tested before, such as base on each enzyme’s regioselectivity. For this purpose, the
arabinosyl derivatives. As a protective group, the acetyl screening of the lipases was performed with the aim of iden-
moiety answers all the requisites for the development of a tifying the best catalysts in terms of activity and regioselec-
facile and scalable synthesis of monodeprotected nucleo- tivity in the hydrolysis of peracetylated nucleosides, pre-
sides: it is recognized as a substrate by lipases and it can be pared by standard procedures (see Exp. Sect. for details).
introduced by a cheap, high-yielding, and quick reaction. In order to investigate the role of the sugar components,
Moreover, the use of immobilized enzymes increases the en- the peracetylated substrates uridine (1), arabinosyluracil
zymes’ stabilities under a wide range of experimental condi- (2), and 2Ј-deoxyuridine (3) were considered first
tions and enables the reuse of the catalysts.
(Scheme 1).
These compounds share the same base (uracil) but have
different sugar components. Enzymatic hydrolysis was car-
ried out at room temperature in phosphate buffer (pH 7)
containing acetonitrile (10% v/v).
An extensive screening of substrates and immobilized en-
zymes provided both a small “enzymatic library” of easy to
handle biocatalysts characterized by different regioselectivi-
ties and a small library of differently monodeprotected pu-
rine and pyrimidine nucleosides useful as intermediates for
the synthesis of high-value modified nucleosides and
mononucleotides.
The chemo-enzymatic approach described here has been
successfully shown to be efficient and versatile. In fact, after
selection of the best enzyme in terms of activity and re-
gioselectivity for the 5Ј-position of peracetylated uridine (1,
Scheme 1), this biocatalyst was used in the preparative
synthesis of 5Ј-deoxy-5-fluorouridine (doxifluridine, 14,
Scheme 4, below) and 5Ј-uridinemonophosphate (5Ј-UMP,
15, Scheme 5, below). Doxifluridine (14) is used in cancer
chemotherapy as a 5-fluorouracil pro-drug,[13] whereas 5Ј-
UMP (15) and the other natural 5Ј-monophosphates have
attracted considerable attention on the food market as addi-
tives in infant dietary formulas.[14]
As reported in Table 1, in the enzymatic deprotection of
peracetylated uridine (1), CRL and PFL each exhibited a
marked preference for the 5Ј-position. After 19 h, PFL had
afforded an almost quantitative yield of 2Ј,3Ј-O-acetylurid-
ine (1a, 93%), whereas the yield of the CRL-catalyzed hy-
drolysis was 85%. The relevance of 5Ј-monodeprotected
peracetylated uridine 1a lies in the use of this molecule as
an intermediate for the synthesis of doxifluridine (14,
Scheme 4, below) by 5Ј-deoxygenation and for its 5Ј-phos-
phorylation to give 5Ј-uridinemonophosphate (15, 5Ј-UMP,
Scheme 5, below).
In the enzymatic hydrolysis of peracetylated arabinosyl-
uracil (2), the lipases under investigation all displayed low
regioselectivities with the exception of CRL, which afforded
2Ј,3Ј-O-acetylarabinosyluracil (2a) in 89% yield. It is inter-
esting to note that both ribosyl- and arabinosyluracil are
preferentially hydrolyzed at the primary hydroxy group.
Some other data relating to the 5Ј-regioselectivities of free
lipases with regard to peracetylated uridine (1) have been
published;[5,10b] these data are in agreement with the results
obtained with immobilized enzymes reported here. In con-
trast, to the best of our knowledge, few reports relating to
Results and Discussion
Regioselective Enzymatic Hydrolysis of Pyrimidine
Nucleoside Derivatives
Lipases from different biological sources [Candida rugosa the study of lipase regioselectivity on arabinosyl derivatives
lipase (CRL), Pseudomonas fluorescens lipase (PFL), Pseu- have so far been published.[15]
Scheme 1. Enzymatic hydrolysis of peracetylated nucleosides 1–6; experimental conditions: immobilized lipase, CH3CN (10%) in KH2PO4
buffer (pH 7, 25 m), room temp.
1968
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1967–1975

Enzymatic Synthesis of 5Ј-Functionalized Nucleosides
Table 1. Enzymatic screening of peracetylated pyrimidine nucleo- PFL: almost quantitative conversion (93%) of 6a was
sides 1–6.[a]
achieved with a remarkable increase in the biotransforma-
tion rate (4 h). In contrast, 2Ј,3Ј,5Ј-tri-O-acetylcytidine (5)
was poorly hydrolyzed by PFL (19% of 5a after 72 h). The
results relating to cytidine derivatives 5, 6 and 2Ј-deoxyri-
bonucleosides 3, 4 reported here clearly show that the re-
gioselectivity of a biocatalyst may strongly depend on the
nature of the substrate. Even a minimal structural variation
can play a crucial role in affecting the outcome of the reac-
tion.
En-
zyme
%
Conv.
Substrate
t [h]
Vh[b]
Product (% yield)
5Ј-OH 3Ј-OH
1
CRL
PFL
PCL
PPL
CRL
PFL
PCL
PPL
CRL
PFL
PCL
PPL
CRL
PFL
PCL
PPL
CRL
PFL
PCL
PPL
CRL
PFL
PCL
PPL
24
19
29
48
24
24
24
24
4
30
24
48
24
2
24
48
120
72
98
48
22
4
0.09
0.24
0.04
0.02
0.07
Ͻ0.01
0.04
Ͻ0.01
0.34
0.02
0.12
0.01
0.10
1.07
0.28
0.02
0.02
Ͻ0.01
Ͻ0.01
0.03
0.17
0.30
0.12
0.02
Ͼ98
100
95
64
97
92
100
39
95
6
72
8
Ͼ98
96
96
34
97
39
61
1a (85)
1a (93)
1a (73)
1a (59)
2a (89)
2a (2)
n.i.[c]
2
3
4
5
6
n.i.[c]
2a (66)
2a (6)
Products 1a–3a, 3b, 4a, 4b, 5a, and 6a were isolated, fully
1
characterized by H and COSY NMR, and then used as
3a (24) 3b (65)
3a (33) 3b (28)
3a (6) 3b (59)
3a (32) 3b (5)
4a (4) 4b (91)
4a (11) 4b (46)
4a (13) 4b (73)
4a (26) 4b (5)
analytical standards. All the results are summarized in
Table 1.
Regioselective Enzymatic Hydrolysis of Purine Nucleoside
Derivatives
5a (77)
5a (19)
5a (39)
5a (34)
6a (86)
6a (93)
6a (72)
6a (16)
n.i.[c]
Regioselective lipase-catalyzed hydrolysis of the peracety-
lated purine nucleosides 7–9 was also investigated (see
Schemes 2 and 3).
43
100
95
Ͼ98
19
n.i.[c]
24
48
[a] Experimental conditions: CH3CN (10%) in KH2PO4 buffer
(pH 7, 25 m), immobilized lipase (50 IU), reaction volume
2.5 mL, [substrate]: 10 m, room temp. [b] µmolϫmin/IU. [c] n.i.:
not isolated.
Enzymatic deacetylation of 3Ј,5Ј-di-O-acetyl-2Ј-deoxy-
uridine (3) was achieved only with low regioselectivity in all
cases. The 3Ј-deprotected nucleoside, however, was formed
as the major product, although only in moderate yield in
comparison with its ribosyl and arabinosyl counterparts. In
fact, the most efficient biocatalysts (CRL and PCL) af-
forded 5Ј-O-acetyl-2Ј-deoxyuridine (3b) in 65% and 59%
yields, respectively. The other tested lipases showed poor
activity with this substrate.
Scheme 2. Enzymatic hydrolysis of peracetylated nucleosides 7–8;
experimental conditions: immobilized lipase, CH3CN (10%) in
KH2PO4 buffer (pH 7, 25 m), room temp.
Surprisingly, when 3Ј,5Ј-di-O-acetylthymidine (4) was
subjected to enzymatic hydrolysis by CRL the yield of the
3Ј-monodeprotected compound 4b increased from 65% (for
3b) up to 91%. The regioselectivity for the 3Ј-position gen-
erally observed for 2Ј-deoxynucleosides was preserved. This
selectivity might be exploitable for easy preparation of
starting materials for 3Ј-functionalized thymidines such as
the well known anti-HIV drug 3Ј-azido-3Ј-deoxythymidine
(zidovudine, AZT).[16]
Scheme 3. Enzymatic hydrolysis of peracetylated ribavirin (9);
experimental conditions: immobilized lipase, CH3CN (10%) in
KH2PO4 buffer (pH 7, 25 m), room temp.
For this purpose, adenosine, arabinosyladenine, and riba-
virin were considered. The results are shown in Table 2.
The enzymatic deacylation of 2Ј,3Ј,5Ј-tri-O-acetyladeno-
Chemical acetylation of cytidine as substrate afforded the
tri- and the tetraacetylated derivatives 5 and 6, respectively, sine (7), consistently with the results reported above for ri-
due to the presence of the exocyclic amino group. In the bonucleosides, was more selective for the 5Ј-hydroxy group.
enzymatic hydrolysis of 2Ј,3Ј,5Ј-tri-O-acetylcytidine (5), In fact, 2Ј,3Ј-O-acetyladenosine (7a) was synthesized in
CRL was identified as the best enzyme for the 5Ј-position 89% yield after 24 h when PFL was used as biocatalyst.
(77% yield of 5a, 120 h). Poorer regioselectivities were ob-
2Ј,3Ј,5Ј-Tri-O-acetylarabinosyladenine (8) was selectively
served for all the other enzymes under the same experimen- hydrolyzed at C-3Ј (to afford 8b) by CRL in up to 77%
tal conditions. Interestingly, in the case of the tetraacety- yield in 24 h. This result is very interesting in comparison
lated cytidine 6 the performance of CRL was even better with the data obtained for arabinosyluracil (Table 1): de-
(86%), but what is noteworthy is the result achieved with pending on the nucleobase, the regioselectivity of CRL
Eur. J. Org. Chem. 2009, 1967–1975
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1969

Daniela Ubiali, Massimo Pregnolato et al.
FULL PAPER
Table 2. Enzymatic screening of peracetylated nucleosides 7–9.[a]
Comparison between Immobilized and Native Enzymes
Substrate Enzyme t [h]
Vh[b] % Conv.
Product (% yield)
Figure 1 shows enzymatic hydrolyses of 1 and 2 with im-
mobilized and native CRL. In both cases the use of the
immobilized enzyme positively affected the reaction rate
and the yield. In fact, the hydrolysis of 1 and 2 by immobi-
lized CRL produced 1a and 2a in about 90% yields (in
24 h), whereas when native CRL was used the yields of the
monodeprotected compounds (1a and 2a, respectively) were
30% after the same reaction time. Additionally, it is well
known that many applications can benefit from use of im-
mobilized enzymes rather than their soluble counterparts –
as reusable heterogeneous biocatalysts, for instance –
through reduction of production costs by recycling and
control of the process. Moreover, enzyme immobilization
can result in improvement in enzyme performance: for ex-
ample, in stability, selectivity, and activity. It is currently
possible to draw the conclusion that immobilized enzymes,
as in this case, can perform better than the native enzymes
(if the immobilization method is correctly selected).[18]
5Ј-OH
3Ј-OH
7
8
9
CRL
PFL
PCL
PPL
CRL
PFL
PCL
PPL
CRL
PFL
PCL
PPL
6
0.25
0.14
0.09
0.01
0.10
0.01
0.03
0.04
0.06
0.58
0.09
0.18
Ͼ98
Ͼ98
Ͼ98
38
91
7
28
48
98
97
7a (59)
7a (89)
7a (64)
7a (28)
n.i.
n.i.[c]
24
24
48
24
48
48
48
24
3
8b (77)
8b (6)
8b (18)
8b (43)
n.i.[c]
9a (13)
9a (93)
9a (44)
9a (7)
24
24
Ͼ98
86
[a] Experimental conditions: CH3CN (10%) in KH2PO4 buffer
(pH 7, 25 m), immobilized lipase (50 IU), reaction volume
2.5 mL, [substrate]: 10 m, room temp. [b] µmolϫmin/IU. [c] n.i.:
not isolated.
switched from C-5Ј for the peracetylated arabinosyluracil 2
to C-3Ј for the peracetylated arabinosyladenine 8.
Ribavirin, used for the treatment of hepatitis C infec-
tion,[17] is an example of a base-modified purine ribonucleo-
side. We studied the enzymatic hydrolysis of peracetylated
ribavirin (9) because of the presence of the unnatural nu-
cleobase. The regioselectivity of CRL, the most efficient
biocatalyst (93% yield in 3 h), was not affected by the base.
In fact, and consistently with most of the peracetylated ri-
bonucleosides considered here (uridine, cytidine, adeno-
sine), which were selectively deprotected at C-5Ј by CRL,
2Ј,3Ј-di-O-acetylribavirin (9a) was also generated as the
main product in the case of ribavirin.
Products 7a, 8b, and 9a were isolated, fully characterized
by H and COSY NMR, and then used as analytical stan-
dards. All the results are summarized in Table 2.
Figure 1. Enzymatic hydrolysis of peracetylated uridine, (1, black)
and arabinosyluracil (2, gray) catalyzed by free CRL (triangle) and
by immobilized CRL (circle).
1
Scheme 4. Chemo-enzymatic synthesis of doxifluridine (14); experimental conditions: a) MeSO2Cl, pyridine, room temp. and 0 °C (yield
91%); b) Br–N+Bu4, DMF, 130 °C (yield 60%); c) (Bu)3SnH, AIBN, toluene, room temp. (yield 65%); d) MeOH/NaOMe, room temp.
(yield 92%); e) immobilized uridine phosphorylase from Bacillus subtilis (UP),[21] KH2PO4 buffer (pH 7.5, 10 m), room temp. (conv.:
38%, not optimized).
1970
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1967–1975

Enzymatic Synthesis of 5Ј-Functionalized Nucleosides
Chemo-Enzymatic Synthesis of 5Ј-Deoxy-5-fluorouridine
(Doxifluridine, 14) and 5Ј-Uridinemonophosphate (5Ј-UMP,
15)
small library of monohydroxy acetylated nucleosides was
prepared in good overall yields and high purities. The syn-
thesis of those synthons by standard chemical procedures
would require multi-stage protocols characterized by time-
consuming protection/deprotection reactions and tedious
purification steps.
Some conclusions can be drawn from the enzymatic
screening described here, and these are summarized in
Table 3.
2Ј,3Ј,5Ј-Tri-O-acetyluridine (1) was subjected to PFL-cat-
alyzed regioselective enzymatic hydrolysis to give 1a. This
nucleoside, bearing only one free hydroxy group, at 5Ј, was
directly used without purification as an intermediate for
functionalization (Scheme 4).
2Ј,3Ј-Di-O-acetyl-5Ј-O-mesyluridine (10) was prepared in
91% yield by treatment of 2Ј,3Ј-di-O-acetyluridine (1a) with
mesyl chloride in pyridine.[19]
Table 3. Enzymatic library for the regioselective deacylation of per-
acetylated nucleosides 1–9.
2Ј,3Ј-Di-O-acetyl-5Ј-bromouridine (11) was prepared in
60% yield from 10 by replacement of the mesyloxy group
by bromine, accomplished by treatment with tetrabutylam-
monium bromide in DMF at 130 °C.[19] Subsequent tribu-
tyltin hydride reduction of 11 in a mixture of toluene and
ethanol at reflux with initiation with AIBN afforded 2Ј,3Ј-
di-O-acetyl-5Ј-deoxyuridine (12) in 65% yield.[19] Com-
pound 12 was deprotected with sodium methoxide in anhy-
drous MeOH to afford 5Ј-deoxyuridine (13) in 92%
yield.[20] This compound was used, after purification, for
enzymatic (pyrimidine/pyrimidine) transglycosylation with
5-fluorouracil. This bioprocess, catalyzed by immobilized
UP (E.C. 2.4.2.1) from Bacillus subtilis,[21] gave doxiflurid-
ine (14) in about 40% conversion (not optimized).
The phosphorylation of the intermediate 1a (Scheme 5)
was performed with phosphorus oxychloride (POCl3) in tri-
ethyl phosphate (TEP)[22] to give the fully deprotected nu-
cleoside 5Ј-monophosphate directly with 97% conversion.
Uridine 5Ј-monophosphate (15, 5Ј-UMP) was obtained by
a chemo-enzymatic two-step process starting from peracety-
lated uridine (1).
Substrate
Enzyme
Deprotected position
% Yield
1
2
3
4
5
6
7
8
9
PFL
CRL
CRL
CRL
CRL
PFL
PFL
CRL
PFL
5Ј-OH
5Ј-OH
3Ј-OH
3Ј-OH
5Ј-OH
5Ј-OH
5Ј-OH
3Ј-OH
5Ј-OH
93
89
65
91
77
93
89
77
93
Out of the enzymes tested, CRL and PFL were the most
efficient biocatalysts in terms of activity, whereas PPL dis-
played insignificant activity with respect to all the consid-
ered substrates.
With regard to the substrates, generally speaking, per-
acetylated ribonucleosides 1, 6, 7, and 9 were preferentially
hydrolyzed at C-5Ј in very high yields by PFL, with the
exception of 2Ј,3Ј,5Ј-tri-O-acetylcytidine (5), which was a
better substrate for CRL. In the hydrolysis of peracetylated
2Ј-deoxyribonucleosides 3–4 the 3Ј-monodeprotected prod-
ucts were prevalently generated. Finally, in the case of the
arabinonucleosides 2 and 8, the same catalyst (CRL) dis-
played different regioselectivities depending on the nucleo-
base: the peracetylated arabinosyluracil 2 was deprotected
at its 5Ј-position, peracetylated arabinosyladenine 8 at its
3Ј-position.
Additionally, the use of immobilized enzymes in place
of the native ones under the same experimental conditions
(Figure 1) clearly showed that the former give improve-
ments in the enzymatic performances, such as activity and
stability.[23]
Scheme 5. Chemo-enzymatic synthesis of 5Ј-UMP (15).
In conclusion, as a result of the study reported here, a
“ready to use” toolset of immobilized enzymes, fully char-
acterized in terms of their regioselectivities with several per-
Conclusions
Four immobilized lipases (CRL, PFL, PCL, PPL) were acetylated nucleosides, has been developed. This chemo-en-
tested in the hydrolysis of a set of peracetylated nucleosides zymatic strategy can provide an easy and versatile synthetic
1–9 to study the influence of different sugars and bases on route for the preparation of various selectively protected
the enzymes’ regioselectivities.
nucleosides to be used for further functionalization. This
The resulting products, generally deprotected at either 3Ј has been successfully demonstrated by the application of
or 5Ј, can be used as starting materials for the preparation this approach to the syntheses of doxifluridine (14) and 5Ј-
of 3Ј- or 5Ј-functionalized nucleosides and mononucleo- UMP (15). In the case of doxifluridine (14, Scheme 4), it is
tides. All these molecules are important in pharmaceuticals, worth mentioning that the synthetic process involves two
the food industry, and biological fields.
enzymatic steps: the lipase-catalyzed hydrolysis of peracety-
This screening made it possible to find catalysts capable lated uridine (1) to give 2Ј,3Ј-di-O-acetyluridine (1a), and
of regioselective deprotection of two crucial positions in nu- the transglycosylation between 5Ј-deoxyuridine (13) and 5-
cleosides. By this chemo-enzymatic approach, therefore, a fluorouracil. Transglycosylation was catalyzed in fully aque-
Eur. J. Org. Chem. 2009, 1967–1975
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1971

Daniela Ubiali, Massimo Pregnolato et al.
FULL PAPER
purified by silica gel column chromatography (CH2Cl2 100% to
CH2Cl2/MeOH, 97:3) to afford the peracetylated nucleosides 1–8.
ous medium by immobilized uridine phosphorylase (UP,
E.C. 2.4.2.1) from Bacillus subtilis by a procedure pre-
viously developed by our group for the synthesis of some
purine 2Ј-deoxynucleosides.[24] In the case of doxifluridine
(14), a further advantage of the transglycosylation is the
obtainment of the fluorinated nucleoside 14 only in the fi-
nal step: this minimizes the need for manipulation of toxic
materials such as 5-fluorouracil and doxifluridine itself.
This aspect should not be underestimated in view of process
hazard evaluation.
2Ј,3Ј,5Ј-Tri-O-acetyluridine (1): Yield 96% (7.10 g); chromato-
graphic conditions (CH2Cl2 100% to CH2Cl2/MeOH, 98:2); TLC
(CH2Cl2/MeOH, 9:1): Rf = 0.61. 1H NMR (400 MHz, [D6]DMSO,
25 °C): δ = 11.43 (s, 1 H, 3-NH), 7.90 (d, 1 H, 5-H), 6.01 (d, 1 H,
1Ј-H), 5.73 (d, 1 H, 6-H), 5.43–5.31 (m, 2 H, 2Ј-H, 3Ј-H), 4.20–
4.10 (m, 3 H, 4Ј-H, 5Ј-H), 2.10–2.01 (s, 9 H, 3 OAc) ppm. MS:
calcd. for [M + 1]+: 371.30; found 370.00. HPLC: Rt = 25.64 min
(method: 10 m KH2PO4 buffer 80%/CH3CN 20%, pH 4.2, flow:
1 mL min–1, λ = 260 nm, RP-18 LiChrosphere select column).
2Ј,3Ј,5Ј-Tri-O-acetylarabinosyluracil (2): Yield 96% (4.20 g); chro-
matographic conditions (CH2Cl2 100% to CH2Cl2/MeOH, 98:2);
1
TLC (CH2Cl2/MeOH, 9:1): Rf = 0.62. H NMR (400 MHz, [D6]-
Experimental Section
DMSO, 25 °C): δ = 11.50 (s, 1 H, 3-NH), 7.60 (d, 1 H, 5-H), 6.65
(d, 1 H, 1Ј-H), 5.75 (d, 1 H, 6-H), 5.37 (t, 1 H, 2Ј-H), 5.15 (t, 1 H,
3Ј-H), 4.25–4.50 (m, 3 H, 4Ј-H, 5Ј-H), 2.10–2.00 (s, 9 H, 3
OAc) ppm. MS: calcd. for [2M + Na]+: 763.61; found 763.00.
HPLC: Rt = 17.31 min (A: 10 m KH2PO4 buffer 90%/CH3CN
10%, B: CH3CN 90%/H2O 10%, pH spontaneous; method: 0–
2 min 100% A, 2–8 min 80% A to 20% B, 8–11 min 70% A to
30% B, 11–18 min 100% A, flow: 1 mL min–1, λ = 260 nm, RP-18
Shiseido Capcell Pak column).
Materials and Methods: Lipases from Candida rugosa (CRL,
406 units g–1), Porcine pancreas (PPL, 139 units g–1), Pseudomonas
fluorescens (PFL, 136 units g–1), Pseudomonas cepacia (PCL,
250 units g–1) and uridine phosphorylase (UP) from Bacillus subtilis
were immobilized by Innovate Biotechnology s.r.l. Rivalta Scrivia
(AL), Italy.[12,21] Uridine was kindly donated by Adorkem Technol-
ogy s.p.a., Costa Volpino (BG), Italy, arabinosyluracil, 2Ј-deoxyuri-
dine, and peracetylated ribavirin were kindly donated by
Pro.bio.sint s.r.l. Varese, Italy, and cytidine, adenosine, doxiflurid-
ine (14), and uridine 5Ј-monophosphate (5Ј-UMP, 15) were pur-
chased from Sigma–Aldrich, Milano, Italy. Chromatographic puri-
fications were performed on silica gel (Merck 60, 40–63 µm) with
the solvent system indicated; TLC analyses were run on silica plates
(Merck 60 F254) and visualized with UV light (254 nm). HPLC
analyses were run with a HPLC Merck Hitachi L-7100 (HPLC
Multi HSM Manager Merck Hitachi D-7000) fitted with an L-7400
detector and an L-7300 oven.
3Ј,5Ј-Di-O-acetyl-2Ј-deoxyuridine (3): Yield 94% (3.50 g); chroma-
tographic conditions (CH2Cl2 100% to CH2Cl2/MeOH, 98:2); TLC
(CH2Cl2/MeOH, 9:1): Rf = 0.53. 1H NMR (400 MHz, [D6]DMSO,
25 °C): δ = 9.10 (s, 1 H, 3-NH), 7.50 (d, 1 H, 5-H), 6.10 (dd, 1 H,
1Ј-H), 5.60 (d, 1 H, 6-H), 5.15 (m, 1 H, 3Ј-H), 4.15 (m, 3 H, 4Ј-H,
5Ј-H), 2.25–2.35 (m, 2 H, 2Ј-H), 2.10–2.00 (s, 6 H, 2 OAc) ppm.
MS: calcd. for [M + Na]+: 335.27; found 335.00. HPLC: Rt =
15.92 min (A: 10 m KH2PO4 buffer 90%/CH3CN 10%,
B: CH3CN 90%/H2O 10%, pH 4.2; method: 0–6 min 100% A, 6–
14 min 85% A to 15% B, 14–22 min 100% A, flow: 1.3 mL min–1,
λ = 260 nm, RP-18 Zorbax SB-AQ column).
1H NMR spectra were recorded on a Bruker AC 400 MHz instru-
ment with tetramethylsilane (Me4Si, δ = 0.00 ppm) as internal stan-
dard. The chemical shifts are expressed in parts per millions (ppm)
relative to the signal of the [D6]DMSO at δ = 2.49 ppm as internal
3Ј,5Ј-Di-O-acetylthymidine (4): Yield 90% (3.50 g); TLC (ethyl ace-
tate/hexane, 7:3): Rf = 0.32. 1H NMR (400 MHz, [D6]DMSO,
25 °C): δ = 10.70 (s, 1 H, 3-NH), 7.50 (s, 1 H, 6-H), 6.10 (dd, 1 H,
1Ј-H), 5.10 (m, 1 H, 3Ј-H), 4.20–4.00 (m, 3 H, 4Ј-H, 5Ј-H), 2.80
(m, 2 H, 2Ј-H), 2.10–2.00 (s, 6 H, 2 OAc), 190 (s, 3 H, 5-H) ppm.
MS: calcd. for [2M + Na]+: 675.21; found 674.60. HPLC: Rt =
20.67 min (A: 10 m KH2PO4 buffer 90%/CH3CN 10%,
B: CH3CN 90%/H2O 10%, pH 4.2; method: 0–6 min 100% A, 6–
14 min 85% A to 15% B, 14–22 min 100% A, flow: 1.3 mL min–1,
λ = 260 nm, RP-18 Zorbax SB-AQ column).
1
reference. The coupling constants H,1H (J, Hz) are in agreement
with the proposed structures. The products obtained in the enzy-
matic hydrolyses were characterized by COSY 2D NMR spec-
troscopy.
Mass spectra were recorded on a LCQ-DECA Thermo Finnigan
Spectrometer by the ESI (Electron Spray Ionization) ionization
method with an ionic source and with use of Xcalibur 1.3 software
(Thermo-Finnigan, San Jose, CA, USA). For sample injection, a
flow of 5 µLmin–1 was used. Analyses were run under positive mo-
dality and the experimental conditions were: voltage of the source
5.0 kV, voltage of the capillary 14 V, flow of the gas 35 (arbitrary
units), temperature 200 °C.
2Ј,3Ј,5Ј-Tri-O-acetylcytidine (5): Yield 86% (3.80 g); chromato-
graphic conditions (CH2Cl2 100% to CH2Cl2/MeOH, 97:3); TLC
(CH2Cl2/MeOH, 9:1): Rf = 0.60. 1H NMR (400 MHz, [D6]DMSO,
25 °C): δ = 7.70 (d, 1 H, 5-H), 7.30 (s, 2 H, NH2), 5.90 (d, 1 H, 1Ј-
H), 5.80 (d, 1 H, 6-H), 5.40 (t, 1 H, 2Ј-H), 5.20 (t, 1 H, 3Ј-H), 4.50–
4.40 (m, 3 H, 4Ј-H, 5Ј-H), 2.30–2.10 (s, 12 H, 3 OAc) ppm. MS:
calcd. for [M + Na]+: 411.35; found 411.10. HPLC: Rt = 17.40 min
(A: 10 m KH2PO4 buffer 90%/CH3CN 10%, B: CH3CN 90%/
H2O 10%, pH spontaneous; method: 0–3 min 100% A, 3–11 min
80% A–20% B, 11–14 min 70% A–30% B, 14–21 min 100% A,
flow: 1 mL min–1, λ = 260 nm, RP-18 Shiseido Capcell Pak col-
umn).
The pH was kept constant during the enzymatic hydrolyses by
automatic titration and the enzymatic activities were measured with
a Metrohm pH-stat 718 Stat Tritino instrument (Herisau, Switzer-
land).
Synthesis of Peracetylated Nucleosides 1–8: A suspension of the
nucleoside (12–20 mmol) in acetonitrile (1 mL/0.2 mmol of nucleo-
side) was treated with triethylamine (TEA, 4 equiv.) and acetic an-
hydride (4 equiv.) in the presence of a catalytic amount of 4-(di-
methylamino)pyridine (DMAP). The resulting mixture was stirred
at room temp. until the reaction was complete (TLC analysis) and
was then diluted with chloroform and water (1:1). The organic
phase was separated, washed with water (4ϫ20–50 mL) and dried
(Na2SO4), filtered, and concentrated in vacuo The residue was
4-N-Acetyl-2Ј,3Ј,5Ј-tri-O-acetylcytidine (6): Yield 90% (4.45 g);
TLC (CH2Cl2/MeOH, 9:1): Rf = 0.76. 1H NMR (400 MHz, CDCl3,
25 °C): δ = 8.90 (s, 1 H, NHAc), 7.95 (d, 1 H, 5-H), 7.50 (d, 1 H),
6.15 (d, 1 H, 1Ј-H), 5.40 (t, 1 H, 2Ј-H), 5.30 (t, 1 H, 3Ј-H), 4.40
(m, 1 H, 4Ј-H) 4.20 (m, 2 H, 5Ј-H) 2.05 (s, 9 H, 4 OAc) ppm. MS:
1972
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1967–1975

Enzymatic Synthesis of 5Ј-Functionalized Nucleosides
calcd. for [M + Na]+: 392.32; found 393.00. HPLC: Rt = 19.81 min
(A: 10 m KH2PO4 buffer 90%/CH3CN 10%, B: CH3CN 90%/
H2O 10%, pH spontaneous; method: 0–3 min 100% A, 3–11 min
80% A to 20% B, 11–14 min 70% A to 30% B, 14–21 min 100%
A, flow: 1 mL min–1, λ = 260 nm, RP-18 Shiseido Capcell Pak col-
umn).
HPLC: Rt = 8.70 min (A: 10 m KH2PO4 buffer 90%/CH3CN
10%, B: CH3CN 90%/H2O 10%, pH spontaneous; method: 0–
2 min 100% A, 2–8 min 80% A to 20% B, 8–11 min 70% A to 30%
B, 11–18 min 100% A, flow: 1 mL min–1, λ = 260 nm, RP-18 Shise-
1
ido Capcell Pak column). H NMR of products 3a and 4a were in
agreement with those previously reported.[25].
2Ј,3Ј,5Ј-Tri-O-acetyladenosine (7): Yield 88% (4.15 g); chromato-
graphic conditions (CH2Cl2 100% to CH2Cl2/MeOH, 97:3); TLC
(CH2Cl2/MeOH, 9:1): Rf = 0.65. 1H NMR (400 MHz, [D6]DMSO,
25 °C): δ = 8.40 (s, 1 H, 2-H), 8.10 (s, 1 H, 6-H), 7.50 (s, 2 H,
NH2), 6.12 (d, 1 H, 1Ј-H), 6.05 (t, 1 H, 2Ј-H), 5.60 (t, 1 H, 3Ј-H),
4.30–4.20 (m, 3 H, 4Ј-H, 5Ј-H), 2.00–1.90 (s, 9 H, 3 OAc) ppm. MS:
calcd. for [M + Na]+: 416.34; found 416.00. HPLC: Rt = 19.66 min
(A: 10 m KH2PO4 buffer 90%/CH3CN 10%, B: CH3CN 90%/
H2O 10%, pH spontaneous; method: 0–3 min 100% A, 3–11 min
80% A to 20% B, 11–14 min 70% A to 30% B, 14–21 min 100%
A, flow: 1 mL min–1, λ = 260 nm, RP-18 Shiseido Capcell Pak col-
umn).
5Ј-Acetyl-O-2Ј-deoxyuridine (3b): TLC (CH2Cl2/MeOH, 9:1): Rf =
0.50. H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 11.50 (s, 1 H,
1
3-NH), 7.90 (d, 1 H, 5-H), 6.10 (dd, 1 H, 1Ј-H), 5.80 (d, 1 H, 6-
H), 5.20 (s, 1 H, OH in 3Ј), 5.15 (m, 1 H, 3Ј-H), 4.10 (m, 1 H, 4Ј-
H), 3.60 (m, 2 H, 5Ј-H), 2.25 (m, 2 H, 2Ј-H), 2.00 (s, 3 H,
OAc) ppm. MS: calcd. for [M + Na]+: 293.23; found 293.00.
HPLC: Rt = 4.93 min (A: 10 m KH2PO4 buffer 90%/CH3CN
10%, B: CH3CN 90%/H2O 10%, pH 4.2; method: 0–6 min 100%
A, 6–14 min 85% A to 15% B, 14–22 min 100% A, flow:
1.3 mL min–1, λ = 260 nm, RP-18 Zorbax SB-AQ column).
5Ј-O-Acetylthymidine (4b): TLC (ethyl acetate/hexane, 7:3). Rf =
1
0.20. H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 8.40 (s, 1 H, 3-
2Ј,3Ј,5Ј-Tri-O-acetylarabinosyladenine (8): Yield 86% (4.05 g);
NH), 7.30 (s, 1 H, 6-H), 6.30 (dd, 1 H, 1Ј-H), 5.80 (s, 1 H, OH in
3Ј), 5.50 (m, 1 H, 3Ј-H), 4.50–4.10 (m, 4 H, 4Ј-H, 5Ј-H e OH in
5Ј), 2.50 (m, 2 H, 2Ј-H), 2.20 (s, 3 H, OAc), 2.10 (s, 3 H, 5-H) ppm.
MS: calcd. for [M + Na]+: 591.52; found 591.00. HPLC: Rt =
9.08 min (A: 10 m KH2PO4 buffer 90%/CH3CN 10%, B: CH3CN
90%/H2O 10%, pH 4.2; method: 0–6 min 100% A, 6–14 min
85% A to 15% B, 14–22 min 100% A, flow: 1.3 mL min–1, λ =
260 nm, RP-18 Zorbax SB-AQ column).
chromatographic conditions (CH2Cl2 100% to CH2Cl2/MeOH,
1
97:3); TLC (CH2Cl2/MeOH, 9:1): Rf = 0.54. H NMR (400 MHz,
[D6]DMSO, 25 °C): δ = 8.20 (s, 1 H, 2-H), 8.10 (s, 1 H, 6-H), 7.30
(s, 2 H, NH2), 6.50 (d, 1 H, 1Ј-H), 5.25 (t, 1 H, 2Ј-H), 5.00 (t, 1 H,
3Ј-H), 4.50–4.00 (m, 3 H, 4Ј-H, 5Ј-H), 2.00–1.50 (s, 9 H, 3
OAc) ppm. MS: calcd. for [M + Na]+: 416.34; found 416.00.
HPLC: Rt = 11.20 min (A: 10 m KH2PO4 buffer 90%/CH3CN
10%, B: CH3CN 90%/H2O 10%, pH spontaneous; method: 0–
3 min 90% A to 10% B, 3–12 min 80% A to 20% B, 12–16 min
90% A to 10% B, flow: 1 mL min–1, λ = 260 nm, RP-18 Shiseido
Capcell Pak column).
2Ј,3Ј-Di-O-Acetylcytidine (5a): TLC (CH2Cl2/MeOH, 9:1): Rf
=
0.50. 1H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 7.80 (d, 1 H, 5-
H), 7.30 (s, 2 H, NH2), 6.00 (d, 1 H, 1Ј-H), 5.70 (d, 1 H, 6-H), 5.40
(m, 1 H, 2Ј-H, 3Ј-H, OH in 5Ј), 4.10 (m, 1 H, 4Ј-H), 3.70 (m, 2 H,
5Ј-H), 2.10–2.00 (s, 6 H, 2 OAc) ppm. MS: calcd. for [M + Na]+:
350.28; found 350.10. HPLC: Rt = 9.80 min (A: 10 m KH2PO4
buffer 90%/CH3CN 10%, B: CH3CN 90%/H2O 10%, pH sponta-
neous; method: 0–3 min 100% A, 3–11 min 80% A to 20% B, 11–
14 min 70% A to 30% B, 14–21 min 100% A, flow: 1 mL min–1, λ
= 260 nm, RP-18 Shiseido Capcell Pak column).
Enzymatic Hydrolysis of Peracetylated Nucleosides 1–9: A solution
of a peracetylated nucleoside (2–4 mmol) in acetonitrile (3.75 mL)
was added to a solution of potassium phosphate buffer (pH 7,
25 m, 33.75 mL). The pH was adjusted to 7.0 and the appropriate
amount of immobilized lipase was added. The suspension was
maintained under mechanical stirring at room temperature until
the maximum hydrolysis of the substrate. During the reaction the
pH was kept constant by automatic titration (Metrohm 718 STAT
Tritino). Samples of the reaction mixture were analyzed at different
times by TLC and HPLC. Finally, the enzyme was filtered off and
washed with deionized water and a solution of acetonitrile (10%),
and the filtrate was concentrated under reduced pressure and ex-
tracted with ethyl acetate (3ϫ20 mL). The collected organic phases
were dried with anhydrous Na2SO4, filtered, and dried under vac-
uum. The residue, when necessary, was further purified by silica gel
column chromatography (CH2Cl2 100% to CH2Cl2/MeOH, 97:3)
to afford the deprotected nucleosides 1a–3a, 3b, 4a, 4b, 5a–7a, 8b
and 9a.
4-N-Acetyl-2Ј,3Ј-di-O-acetylcytidine (6a): TLC (CH2Cl2/MeOH,
9:1): Rf = 0.61. 1H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 11.00
(s, 1 H, NHAc), 8.30 (d, 1 H, 5-H), 7.25 (d, 1 H, 6-H), 6.10 (d, 1
H, 1Ј-H), 5.20–5.30 (m, 1 H, 2Ј-H, 3Ј-H, OH in 5Ј), 4.30 (m, 1 H,
4Ј-H), 3.60–3.80 (m, 2 H, 5Ј-H), 2.10–2.00 (s, 9 H, 3 OAc) ppm.
MS: calcd. for [M + Na]+: 392.32; found 392.10. HPLC: Rt =
15.20 min (A: 10 m KH2PO4 buffer 90%/CH3CN 10%,
B: CH3CN 90%/H2O 10% pH spontaneous; method: 0–3 min
100% A, 3–11 min 80% A to 20% B, 11–14 min 70% A to 30% B,
14–21 min 100% A, flow: 1 mL min–1, λ = 260 nm, RP-18 Shiseido
Capcell Pak column).
2Ј,3Ј-Di-O-acetyladenosine (7a): TLC (CH2Cl2/MeOH, 9:1): Rf =
0.53. H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 8.40 (s, 1 H, 2-
2Ј,3Ј-Di-O-Acetyluridine (1a): TLC (CH2Cl2/MeOH, 9:1): Rf
=
1
1
0.51. H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 11.40 (s, 1 H,
3-NH), 7.89 (d, 1 H, 5-H), 6.00 (d, 1 H, 1Ј-H), 5.70 (d, 1 H, 6-H),
5.50 (s, 1 H, OH in 5Ј), 5.30 (m, 2 H, 2Ј-H, 3Ј-H), 4.14 (m, 1 H,
4Ј-H) 3.64 (m, 2 H, 5Ј-H) 2.10–2.02 (s, 6 H, 2 OAc) ppm. MS:
calcd. for [M + 1]+: 351.26; found 351.00. HPLC: Rt = 21.81 min
(method: 10 m KH2PO4 buffer 80%/CH3CN 20%, pH 4.2, flow:
1 mL min–1, λ = 260 nm, RP-18 LiChrosphere select column).
H), 8.20 (s, 1 H, 6-H), 7.50 (s, 2 H, NH2), 6.12 (d, 1 H, 1Ј-H), 5.80
(t, 1 H, 2Ј-H), 5.60 (m, 1 H, OH in 5Ј), 5.50 (t, 1 H, 3Ј-H), 4.20
(m, 1 H, 4Ј-H), 3.70–3.60 (m, 2 H, 5Ј-H), 2.20–1.90 (s, 6 H, 2
OAc) ppm. MS: calcd. for [M + Na]+: 374.30; found 374.00.
HPLC: Rt = 16.31 min (A: 10 m KH2PO4 buffer 90%/CH3CN
10%, B: CH3CN 90%/H2O 10%, pH spontaneous; method: 0–
3 min 100% A, 3–11 min 80% A to 20% B, 11–14 min 70% A to
30% B, 14–21 min 100% A, flow: 1 mL min–1, λ = 260 nm, RP-18
Shiseido Capcell Pak column).
2Ј,3Ј-Di-O-Acetylarabinosyluracil (2a): TLC (CH2Cl2/MeOH, 9:1):
1
Rf = 0.54. H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 11.00 (s,
1 H, 3-NH), 7.85 (d, 1 H, 5-H), 6.25 (d, 1 H, 1Ј-H), 5.60 (d, 1 H,
6-H), 5.40–5.20 (m, 2 H, 2Ј-H, 3Ј-H), 5.10 (s, 1 H, OH at 5Ј), 4.10 2Ј,5Ј-Di-O-acetylarabinosyladenine (8b): TLC (CH2Cl2/MeOH,
1
(m, 1 H, 4Ј-H), 3.60 (m, 2 H, 5Ј-H), 2.10–1.90 (s, 6 H,
9:1): Rf = 0.40. H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 8.10
OAc ϫ2) ppm. MS: calcd. for [M + 1]+: 351.26; found 351.20.
(s, 1 H, 2-H), 8.00 (s, 1 H, 6-H), 7.40 (s, 2 H, NH2), 6.50 (d, 1 H,
Eur. J. Org. Chem. 2009, 1967–1975
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1973

Daniela Ubiali, Massimo Pregnolato et al.
FULL PAPER
1Ј-H), 5.30 (t, 1 H, 2Ј-H), 5.00 (m, 1 H, OH in 3Ј), 4.50 (t, 1 H,
Synthesis of 2Ј,3Ј-Di-O-acetyl-5Ј-deoxyuridine (12): 2,2Ј-Azobis(2-
3Ј-H), 4.20–4.00 (m, 3 H, 4Ј-H, 5Ј-H), 2.00–1.50 (s, 6 H, 2 methylpropionitrile) (AIBN, 0.15 g) was added to 2Ј,3Ј-di-O-acetyl-
OAc) ppm. MS: calcd. for [M + Na]+: 374.30; found 374.00.
5Ј-bromouridine (11, 0.25 g, 0.639 mmol) at reflux in toluene
HPLC: Rt = 4.69 min (A: 10 m KH2PO4 buffer 90%/CH3CN (20 mL) and absolute ethanol (5 mL). A solution of tributyltin hy-
10%, B: CH3CN 90%/H2O 10%, pH spontaneous; method: 0–
3 min 90% A to 10% B, 3–12 min 80% A to 20% B, 12–16 min
90% A to 10% B, flow: 1 mL min–1, λ = 260 nm, RP-18 Shiseido
Capcell Pak column).
dride (0.794 g, 2.73 mmol) in toluene (5 mL) was then added and
the reaction mixture was heated at reflux for 2 h. The mixture was
filtered under reduced pressure, the filtrate was concentrated in
vacuo, and the residue was crystallized from absolute ethanol to
afford 12 in 65% yield (0.13 g). TLC (ethyl acetate/hexane, 7:3): Rf
= 0.50. 1H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 11.50 (s, 1 H,
3-NH), 7.80 (d, 1 H, 5-H), 5.80 (d, 1 H, 1Ј-H), 5.70 (d, 1 H, 6-H),
5.50 (m, 1 H, 2Ј-H), 5.00 (m, 1 H, 3Ј-H), 4.00 (m, 1 H, 4Ј-H), 2.00
(m, 6 H, 2 OAc), 1.30 (s, 3 H, 5Ј-H) ppm.
2Ј,3Ј-Di-O-N-acetylribavirin (9a): TLC (CH2Cl2/MeOH, 9:1): Rf =
1
0.50. H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 9.00 (s, 1 H, 2-
H), 6.30 (d, 1 H, 1Ј-H), 5.70 (t, 1 H, 2Ј-H), 5.50 (m, 1 H, 3Ј-H),
5.10 (m, 1 H, OH in 5Ј), 4.25 (m, 1 H, 4Ј-H), 3.90 (s, 3 H, NHAc),
3.75–3.50 (m, 1 H, 5Ј-H), 3.40 (s, 1 H, NH), 2.20–2.00 (s, 6 H, 2
OAc) ppm. MS: calcd. for [M + Na]+: 392.30; found 392.00.
HPLC: Rt = 12.35 min (A: 10 m KH2PO4 buffer 90%/CH3CN
10%, B: CH3CN 90%/H2O 10%, pH spontaneous, T = 30 °C;
method: 0–2 min 97% A to 3% B, flow 1 mL min–1, 2–7 min
80% A to 20% B, flow 1 mL min–1, 7–12 min 70% A to 30% B,
flow 1.2 mL min–1, 12–20 min 97% A to 3% B, flow 1.2 mL min–1,
λ = 220 nm, RP-18 Shiseido Capcell Pak column).
Synthesis of 5Ј-Deoxyuridine (13): Sodium methoxide (5 mL) was
added to a solution of 2Ј,3Ј-di-O-acetyl-5Ј-deoxyuridine (12, 0.08 g,
0.256 mmol) in dry MeOH (10 mL) until pH 9–10 was reached.
The mixture was kept at room temperature overnight and was then
neutralized with Amberlite IR-120 (H+) resin. The resin was fil-
tered off, and the filtrate was concentrated and coevaporated with
dry CH2Cl2 (2ϫ10 mL) to give 5Ј-deoxyuridine (13) in 92% yield
(0.054 g). TLC (CH2Cl2/MeOH, 9:1): Rf
=
0.36. 1H NMR
Synthesis of 2Ј,3Ј-Di-O-acetyluridine (1a): A solution of 1 (100 m,
37 g L–1) in acetonitrile (32 mL) was added to a solution of potas-
sium phosphate buffer (pH 7, 25 m, 128 mL). The pH was ad-
justed to 7.0 and the immobilized lipase (400 units) was added. The
suspension was kept under mechanical stirring at room temperature
until the maximum hydrolysis of the substrate. During the reaction,
the pH was kept constant by automatic titration (Metrohm 718
STAT Tritino). Samples of the reaction mixture were analyzed at
different times by TLC. Finally, the enzyme was filtered off and
washed with deionized water and a solution of acetonitrile (20%),
and the filtrate was extracted with ethyl acetate (3ϫ70 mL) and
dried in vacuo to give 1a in 90% yield (4.75 g, 29.7 g/L).
(400 MHz, [D6]DMSO, 25 °C): δ = 10.60 (s, 1 H, 3-NH), 7.80 (d,
1 H, 5-H), 5.90 (d, 1 H, 6-H), 5.80 (d, 1 H, 1Ј-H), 5.70–5.20 (m, 2
H, OH in 2Ј e 3Ј), 4.10 (m, 1 H, 2Ј-H), 3.90 (m, 1 H, 4Ј-H), 3.75
(m, 1 H, 3Ј-H), 1.30 (s, 3 H, 5Ј-H) ppm.
Synthesis of 5Ј-Deoxy-5-fluorouridine (Doxifluridine, 14): A solu-
tion of KH2PO4 buffer (pH 7.5, 10 m, 1 mL) containing 5-fluoro-
uracil (26 mg) and 5Ј-deoxyuridine (13, 9 mg) was maintained un-
der mechanical stirring at 25 °C. Immobilized UP (2 U) was added
to the solution. The reaction was monitored by HPLC (dilution
samples 1:20 in 10 m KH2PO4 buffer). The suspension was stirred
for 24 h and then stopped by filtration of the immobilized enzyme
under reduced pressure. Conversion 38%, Rt = 19.53 min (KH2PO4
buffer, 10 m, pH 6.8 99%/MeOH 1%, flow: 1 mL min–1, λ =
260 nm, T = 35 °C, LiChrocart RP-18 250 column).
Synthesis of 2Ј,3Ј-Di-O-acetyl-5Ј-O-mesyluridine (10): Mesyl chlo-
ride (1.2 mL) was added dropwise with stirring to an ice-cooled
solution of 2Ј,3Ј-O-diacetyluridine (1a, 0.697 g, 2.13 mmol) in dry
pyridine (50 mL). The solution was stirred at 25 °C for 1 h and
was then kept in an ice-box overnight. The reaction mixture was
concentrated and the residue was repeatedly coevaporated with a
mixture of toluene and ethanol. The resultant residue was dissolved
in ethyl acetate and washed with water (10 mL). The organic phase
was extracted with ethyl acetate (3ϫ15 mL), dried with anhydrous
Na2SO4, and filtered. The filtrate was concentrated in vacuo to
afford 10 in 90.5% yield (0.65 g). TLC (CH2Cl2/MeOH, 95:5): Rf
= 0.62. 1H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 11.29 (s, 1 H,
3-NH), 7.50 (d, 1 H, 5-H), 5.69 (d, 1 H, 1Ј-H), 5.49 (d, 1 H, 6-H),
5.40–5.00 (m, 2 H, 2Ј-H, 3Ј-H) 4.50–4.00 (m, 3 H, 4Ј-H, 5Ј-H) 3.10
(s, 3 H, CH3S), 1.82 (s, 6 H, 2 OAc) ppm.
Synthesis of Uridine 5Ј-Monophosphate (5Ј-UMP, 15): A mixture of
2Ј,3Ј-di-O-acetyluridine (1a, 0.40 g, 1.2 mmol) in water (0.025 mL)
suspended in triethyl phosphate (3.8 mL, 3.6 mmol) was stirred at
0 °C for 10 minutes. Phosphorus oxychloride (0.4 mL, 3.6 mmol)
was added to the cold mixture, which was maintained under stir-
ring for 24 h. The solution was adjusted to pH 1.5 with a sodium
hydroxide solution and heated at 70 °C for 2 h. Uridine 5Ј-mono-
phosphate (15) was formed in 97% conversion, as determined by
HPLC. Rt = 2.88 min [A: 2 g (NH4)H2PO4, 0.5 g (NH4)2HPO4 in
1 L and 30 mL of MeOH, B: CH3CN 90%/H2O 10%, method 0–
8 min 100% A, 8–22 min 70% A to 30% B, 22–28 min 100% A],
flow: 1 mL min–1, λ = 260 nm, T = 25 °C, LiChrocart RP-18 col-
umn.
Synthesis of 2Ј,3Ј-Di-O-acetyl-5Ј-bromouridine (11): A mixture of
2Ј,3Ј-di-O-acetyl-5Ј-O-mesyluridine (10, 0.65 g, 1.59 mmol), tetra-
butylammonium bromide (2.54 g, 7.95 mmol), and dimethylform-
amide (DMF, 27 mL) was heated at 130 °C for 1.5 h and concen-
trated, and the residue was dissolved in ethyl acetate (50 mL). The
solution was washed with water (40 mLϫ4), and the aqueous
washings were combined and extracted with ethyl acetate (50 mL).
The ethyl acetate solution was dried with anhydrous Na2SO4 and
filtered, and the filtrate was concentrated. The final residue was
chromatographed on a silica gel column (ethyl acetate/hexane, 70–
30%); yield 60% (0.25 g). TLC (ethyl acetate/hexane, 7:3): Rf =
Acknowledgments
We thank Dr. I. Nieto (Innovate Biotechnology s.r.l.) for her help
in the experimental work. This work was partially funded by the
subproject BIOCAT REGINS 2E0006R within the INTERREG
IIIC Regional Framework Operation (RFO) and by Sovvenzione
Globale INGENIO POR Ob. 3 F.S.E. 2000–2006.
1
0.40. H NMR (400 MHz, [D6]DMSO, 25 °C): δ = 11.20 (s, 1 H,
[1] H. Vorbruggen, C. Ruh-Pohlenz, Synthesis of Nucleosides in
Organic Reactions, Wiley, New York, 2000.
[2] H. Vorbruggen, C. Ruh-Pohlenz, Handbook of Nucleosides
Synthesis, Wiley, New York, 2001.
3-NH), 7.40 (d, J = 8.0 Hz, 1 H, 5-H), 5.70 (d, 1 H, 1Ј-H), 5.50 (d,
1 H, 6-H), 5.30–5.00 (m, 2 H, 2Ј-H, 3Ј-H), 4.30–4.00 (m, 3 H, 4Ј-
H, 5Ј-H), 1.90 (s, 6 H, 2 OAc) ppm.
1974
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1967–1975

Enzymatic Synthesis of 5Ј-Functionalized Nucleosides
[3] C. M. Galmarini, J. R. Mackey, C. Dumontet, Lancet Oncol.
2002, 3, 415–424.
[4] P. Herdewijn, Current Protocols in Nucleic Acid Chemistry,
John Wiley & Sons, New York, 2007, pp. 14.0.1–14.0.6.
[5] M. Ferrero, V. Gotor, Chem. Rev. 2000, 100, 4319–4347.
[6] C. H. Wong, G. M. Whitesides, Enzymes in Synthetic Organic
Chemistry Pergamon, Oxford, 1994.
[15]
[16]
F. Moris, V. Gotor, Tetrahedron 1993, 49, 10089–10098.
F. Barre-Sinoussi, J. C. Chermann, F. Rey, M. T. Nugeyre, S.
Chamaret, J. Gruest, C. Dauguet, C. Axler-Blin, F. Vezinet-
Brun, C. Rouzioux, W. Rozenbaum, L. Montagnier, Science
1983, 220, 868–871.
a) J. Y. N. Lau, R. C. Tam, T. J. Liang, Z. Hong, Hepatology
2002, 35, 1002–1009; b) R. T. Chung, M. Gale Jr., S. J. Polyak,
S. M. Lemon, T. J. Liang, J. H. Hoofnagle, Hepatology 2008,
47, 306–320.
L. Cao, Curr. Opin. Chem. Biol. 2005, 9, 217–226.
L. Hein, P. Draser, J. Beranek, Nucleic Acids Res. 1976, 3,
1125–1137.
F. Sajtos, J. Hajkò, K. E. Kövér, A. Lipták, Carbohydr. Res.
2001, 334, 253–259.
[17]
[7] A. K. Prasad, S. Trikha, V. S. Parmar, Bioorg. Chem. 1999, 27,
135–154.
[8] R. D. Schmid, R. Verger, Angew. Chem. Int. Ed. 1998, 37,
[18]
[19]
1609–1633.
[9] B. La Ferla, Monatsh. Chem. 2002, 133, 351–368.
[10] a) A. Uemura, K. Nozaki, J. Yamashita, M. Yasumoto, Tetra-
hedron Lett. 1989, 30, 3819–3820; b) H. K. Singh, G. L. Cote,
R. S. Sikorski, Tetrahedron Lett. 1993, 34, 5201–5204; c) A.
Zinni, A. Schmidt, M. Gallo, L. Iglesias, A. Iribarren, Mole-
cules 2000, 5, 533–534; d) D. I. Roncaglia, A. M. Schmidt,
L. E. Iglesias, A. M. Iribarren, Biotechnol. Lett. 2001, 23,
1439–1443; e) M. A. Zinni, S. D. Rodríguez, R. M. Pontiggia,
J. M. Montserrat, L. E. Iglesias, A. M. Iribarren, J. Mol. Catal.
B Enzym. 2004, 29, 129–132.
[11] S. Rocchietti, M. Terreni, M. Pregnolato, PCT Int. Appl.
WO03057894, EP1466004A1, AU2003235812A1; Chem. Abstr.
2003, 139, 116337.
[12] A. Bastida, P. Sabuquillo, P. Armisen, R. Femàndez-Lafuente,
J. Huguet, J. M. Guisàn, Biotechnol. Bioeng. 1998, 58, 486–493.
[13] D. B. Longley, D. P. Harkin, P. G. Johnston, Nat. Rev. Cancer
2003, 3, 330–338.
[20]
[21] S. Rocchietti, D. Ubiali, M. Terreni, A. M. Albertini, R. Fer-
nandez-Lafuente, J. M. Guisan, M. Pregnolato, Biomacromole-
cules 2004, 5, 2195–2200.
[22] M. Yoshikawa, T. Kato, T. Takenishi, Bull. Chem. Soc. Jpn.
1969, 42, 3505–3508.
[23] C. Mateo, J. M. Palomo, G. Fernandez-Lorente, J. M. Guisan,
R. Fernandez-Lafuente, Enzyme Microb. Technol. 2007, 40,
1451–1463.
[24] D. Ubiali, S. Rocchietti, F. Scaramozzino, M. Terreni, A. M.
Albertini, R. Fernandez-Lafuente, J. M. Guisan, M. Pregnol-
ato, Adv. Synth. Catal. 2004, 346, 1361–1366.
[25] V. Gotor, F. Morís, Synthesis 1992, 7, 626–628.
Received: November 6, 2008
[14] a) H. Jyonouchi, J. Nutr. 1994, 124, 138S–143S; b) C. T.
Van Buren, A. D. Kulkarni, F. B. Rudolph, J. Nutr. 1994, 124,
160S–164S.
Published Online: February 23, 2009
Eur. J. Org. Chem. 2009, 1967–1975
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1975