Chemoenzymatic preparation of nucleosides from furanoses

Article abstract of DOI:10.1016/j.tetlet.2008.02.087

Chemoenzymatic preparation of ribose, deoxyribose and arabinose 5-phosphates was accomplished. These compounds were tested as starting materials in the enzymatic preparation of natural and modified purine and pyrimidine nucleosides, using an overexpressed Escherichia coli phosphopentomutase.

Full text of DOI:10.1016/j.tetlet.2008.02.087

Available online at www.sciencedirect.com
Tetrahedron Letters 49 (2008) 2642–2645
Chemoenzymatic preparation of nucleosides from furanoses
Marisa Taverna-Porro a, Leon A. Bouvier b, Claudio A. Pereira b,
Javier M. Montserrat a,c,*, Adolfo M. Iribarren a,d
a INGEBI (CONICET), Vuelta de Obligado 2490-(1428), Buenos Aires, Argentina
b Instituto de Investigaciones Me´dicas Alfredo Lanari (CONICET-UBA), Av. Combatientes de Malvinas 3150 (1427), Cap. Fed., Argentina
c Instituto de Ciencias, Universidad Nacional de Gral. Sarmiento, J. M. Gutierrez 1150, Los Polvorines (B1613GSX), Prov. de Bs. As., Argentina
d Laboratorio de Biotransformaciones, Universidad Nacional de Quilmes, Roque Saenz Pen˜a 352 (1876) Bernal, Prov. de Bs. As., Argentina
Received 14 September 2007; revised 13 February 2008; accepted 14 February 2008
Available online 20 February 2008
Abstract
Chemoenzymatic preparation of ribose, deoxyribose and arabinose 5-phosphates was accomplished. These compounds were tested
as starting materials in the enzymatic preparation of natural and modified purine and pyrimidine nucleosides, using an overexpressed
Escherichia coli phosphopentomutase.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
proceeds via the formation of the corresponding a-fura-
nose 1-phosphate, that is, subsequently used by the purine
NPs (PNP) as substrate. This strategy has been applied to
the preparation of sugar and base modified nucleosides,
using isolated enzymes or whole cells.8 A limitation of this
approach, for sugar modified nucleosides, is that the corre-
sponding pyrimidine analogue must be available. A way to
overcome this problem is to employ furanose 1-phosphates
as starting materials,9 which unfortunately are limited by
the synthetic difficulty and intrinsic unstability10 of this
class of compounds. Alternatively, furanose 1-phosphates
can be enzymatically obtained from furanose 5-phosphates
using phosphopentomutase (EC 5.4.2.7, PPM), an enzyme
of the pentose pathway. PPM catalyses the reversible trans-
fer of a phosphate group between the hydroxyls of posi-
tions 5 and 1 of ribose and deoxyribose, in bacteria and
in mammal tissues. It showed to be a monomeric metallo-
enzyme,11 Mn2+ dependent12 and activated by glucose 1,6-
diphosphate. PPM was originally overexpressed along with
thymidine phosphorylase (TP) by Valentin-Hansen et al.,13
but it was Wong14 who first compared relative rates of
overexpressed PPM catalysed conversion of D-pentose
5-phosphates to a-D-pentose 1-phosphates based on a cou-
pled enzymatic assay with a PNP. In this work, they estab-
lished that Escherichia coli PPM accepted as substrates
The synthesis of modified nucleosides has an important
impact in human health since many antiviral and antitu-
moral agents belong to this type of compounds.1
Traditionally, nucleoside analogues have been chemi-
cally prepared,2 which often requires difficult and time con-
suming multistep processes, including selective protection–
deprotection reactions. In particular, the stereoselective
glycosylation of the pentose moiety is usually a challenging
step.3
An alternative to chemical synthesis is the biotechnolog-
ical preparation of nucleosides. This approach presents
some advantages, usually the reaction conditions are
smoother, environmentally friendly and principally regio-
and stereoselective.4 Along the last two decades, enzymatic
protection–deprotection schemes5 as well as biocatalyzed
glycosylation have been reported.6 Related to this last reac-
tion, nucleoside phosphorylases (NPs) have been exten-
sively used in the preparation of purine nucleoside from
pyrimidine ones (transglycosylation).7 The reaction
*
Corresponding author. Tel.: +54 11 4783 2871; fax: +54 11 4786 8578.
E-mail address: jmonstser@ungs.edu.ar (J. M. Montserrat).
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.02.087

M. Taverna-Porro et al. / Tetrahedron Letters 49 (2008) 2642–2645
2643
deoxyribose 5-phosphate, ribose 5-phosphate and arabi-
nose 5-phosphate but not 2,3-dideoxyribose 5-phosphate;
unfortunately no indication of yields or absolute activities
were reported. For the particular case of deoxynucleosides,
a synthetic use of PPM has been reported by Ouerkerk
et al.15 for the preparation of 15N and 13C radiolabelled
thymidine and 20-deoxyuridine. In the same sense, Shimizu
group16 developed a one-pot microbial synthesis of 20-
deoxyribonucleoside in three steps.
Taking into account these antecedents, we decided to
explore the chemoenzymatic production of ribose-, deoxy-
ribose- and arabinose 5-phosphate, in order to evaluate the
production of natural and modified nucleosides by the use
of an overexpressed PPM, emphasizing on the generality of
the method.
82% total yield starting from 1c. Methyl riboside and
methyl arabinoside were also peracetylated using previ-
ously described conditions.18
Then, the 5-acetyl groups of methyl 2,3,5-tri-O-acetyl-
a,b-D-riboside, methyl 2,3,5-tri-O-acetyl-a,b-D-arabinoside
and 1,3,5-tri-O-acetyl-2-deoxy-a,b-D-ribofuranose were
regioselectively removed using lipase from Candida rugosa
(CRL) or Candida antartica B lipase (CAL B) under alco-
holysis conditions19 to obtain products 2a, 2b and 2c
(Fig. 1).
Methyl 2,3-di-O-acetyl-a,b-D-riboside (2a, Fig. 1) was
phosphorylated with phosphorous oxychloride in acetoni-
trile and pyridine. Total deprotection was achieved by the
simple addition of water to the reaction medium. After
neutralisation, inorganic phosphate was precipitated with
aqueous BaCl2 and filtered. The subsequent addition of
ethanol to the solution afforded 5a (Fig. 1) as its barium
salt precipitate. This solid was transformed to the sodium
form using ion exchange chromatography (Dowex
50WX2-200 (H)) in 75% total yield.
When the previously described reaction was applied
to compounds 2b and 2c, the separation of inorganic
phosphate from furanose 5-phosphate was not possible.
As consequence that inorganic phosphate could cause
PPM inhibition, a different phosphorylation strategy was
assessed. Compounds 2b and 2c were phosphorylated using
dibenzyl N,N-diisopropylphosphoramidite in THF with
tetrazole activation and directly oxidised to phosphate by
the action of tert-butyl hydroperoxide (Fig. 1). Phosphates
3b and 3c were purified by column chromatography and
further debenzylated by hydrogenolysis catalysed with pal-
ladium hydroxide (4b, 4c, Fig. 1). Total deprotection was
achieved by chemical acid hydrolysis and the furanose 5-
phosphates were precipitated as barium salts, isolated by
centrifugation and transformed to the corresponding
sodium salts using ion exchange chromatography. In this
way, arabinose 5-phosphate (5b, Fig. 1) and 2-deoxyribose
5-phosphate (5c, Fig. 1) were obtained as pure products.
In order to determine the better conditions for the PPM
catalysed reaction, the dependence of enzyme activity and
reaction yield on pH, temperature, glucose 1,6-diphos-
phate, b-mercaptoethanol and phosphate concentrations
was studied. For this purpose, the conversion of ribose 5-
phosphate to adenosine, catalysed by the overexpressed
PPM and commercial PNP, was selected as the model sys-
tem. The enzyme activity and yield results are presented at
Figure 2.
2. Results and discussion
In an attempt to develop a general strategy for the prep-
aration of furanose 5-phosphates starting from ribose,
arabinose and 2-deoxyribose (1a, b, c, Fig. 1), the corre-
sponding methyl glycosides were prepared. While methyl
riboside and methyl arabinoside were obtained mainly in
the furanoside form, the methyl deoxyriboside was
obtained as a 1:0.75 mixture of the pyranoside and furan-
oside forms that could not be separated. To overcome this
problem, deoxyribose was locked in the furanose form
using a different strategy that involved the regioselective
acetylation of the 5-position using acetic anhydride and
Candida Antartica lipase (CAL B) as biocatalyst.17 Then,
chemical acetylation using acetic anhydride in pyridine
was performed to protect positions 1 and 3, obtaining per-
acetylated deoxyribose exclusively in the furanose form in
HO
HO
1. CH3OH
O
O
H2SO4(conc)
OH
OCH3
Y
Z
2. Ac2O, Py
3. CRL, EtOH
HO
X
AcO
2a, b
43-47% yield
W
1a, b
HO
HO
AcO
O
O
O
OH
OAc
OH
OH
Ac2O
1. Ac2O, Py
CAL B
2. CAL B, EtOH
HO
AcO
HO
2c
1c
O
PO
91% yield
HO
HO
1. POCl3/Py
O
2a
2. H2O (pH=1)
HO OH
5a 75 % yield
O
PO
O
PO
BzO
HO
HO
BzO
-Pr)2
Tetrazole, THF
O
O
Z
Recombinant PPM was further tested applying the opti-
mised conditions previously discussed (Fig. 2), coupled to
two different commercial NPs, PNP and TP. Although
the commercial supplier (Sigma–Aldrich) indicated that
the origin of the TP was E. coli, no data about the bacterial
origin of PNP was available. The estimation of its molecu-
lar weight by gel electrophoresis (data not shown) and the
preference for hypoxanthine over adenine (Table in Fig. 3)
suggest that it belongs to the low molecular mass trimeric
PNP family.6
1. (BzO)2PN(
i
OR
OR
H2(g)
Pd(OH)2/C
Z
2b, c
2. t-BuOOH
AcO
W
AcO
4b, c
W
3b, c
74-79% yield
O
HO
HO
PO
Compound
X
OH
H
Y
W
Z
R
O
a
b
c
H
OH
H
OAc
H
H
CH3
OAc CH3
OAc
H2O
OH
Y
H
H
H
(pH=1)
HO
5b, c
68-72% yield
X
Fig. 1. Preparation of furanose 5-phosphates.

2644
M. Taverna-Porro et al. / Tetrahedron Letters 49 (2008) 2642–2645
O
PO
NH2
HO
NH2
HO
-
O
B
N
N
N
N
O
B
O
N
O
OP
OH
HO
O-
HO
X
X
N
PPM, NP
Tris-Buffer
N
H
N
O
O
HO
Y
HO
Y
OH
5a, b, c
PPM, PNP
Tris-Buffer
OH OH
OH OH
Comp.
Nº
6
7
8
X
Y
NP
B
Rel.
Act.
100
14
66
Yield/ Time/
%
98
90
89
25
40
100
46
8
h
2
48
2
32
96
1.5
46
96
24
72
2
a
H
H
H
H
H
OH PNP
OH PNP
OH PNP
OH PNP 6-Cl-2-NH2-Pu
OH PNP
Hyp
A
6-SH-Pu
100
80
60
40
20
0
9
1
4.1
10
11
12
13
14
15
16
17
18
19
a
2-F-Ad
TCA
Hyp
H
OH PNP
127
1.15
0.54
0.34
0.27
204
148
145
104
OH
OH
OH
OH
H
H
H
H
H
H
H
H
H
H
H
H
PNP
PNP
PNP
PNP
TP
TP
TP
TP
A
-7
-6
-5
-4
-3
-2
6-SH-Pu
TCA
T
U
5-F-U
5-Br-U
7
4
6
8
10
log[glu-1,6-diP]/M
19
75
74
60
100
pH
Rel. activity
Rel. yield
Rel. activity
Rel. yield
2
2.25
0.75
100
80
100
80
60
40
20
0
The relative activity value of 100 % correspond to 0.49 μmol.min-1.mg prot-1
60
40
O
NH
SH
Cl
O
2
N
N
N
N
N
20
NH
N
N
N
NH
2
N
0
0
N
H
N
H
N
H
N
H
N
H
N
N
N
N
NH
2
50
100
150
200
0
1
2
3
4
[mercaptoethanol]/mM
Hip
[Ribose 5-P]/mM
6-Cl-2-NH2-Pu TCA
A
6-SH-Pu
O
O
O
O
NH
Rel. Activity
Rel. Yield
2
Rel. activity Rel. yield
F
Br
N
NH
NH
NH
NH
N
N
H
N
H
O
N
H
O
N
H
O
N
H
O
N
F
110
100
90
100
80
60
40
20
0
2-F-Ad
U
T
5-F-U
5-Br-U
80
70
Fig. 3. Glycosylation reaction with different furanoses 5-phosphates and
bases.
60
50
40
25
30
35
40
45
50
0
0.5
1
1.5
2
[Na3PO4]/mM
T/ºC
Rel. activity
nine, 6-mercaptopurine and 1,2,4-triazole-3-carboxamide
were employed for the coupled glycosylation catalysed by
the commercial PNP (12–15, Fig. 3). It has been previously
established that arabinosides are poorer substrates than
ribosides for PNP,20 which may explain the low relative
activities and yields obtained. The best result (46% yield)
was achieved for compound 12.
Finally, deoxyribose 5-phosphate (5c, Fig. 1) was used
in combination with PPM and TP employing thymine,
uracil, 5-fluorouracil and 5-bromouracil as bases (16–19,
Fig. 3). In all cases, high activities and yields were
observed. In the case of the preparation of compound 19,
a chemical dehalogenation produced by the b-mercap-
toethanol present in the reaction medium was observed.
For this particular case, an activity and yield optimisation
experiment was performed against b-mercaptoethanol con-
centration, setting this value around 50 mM.
Rel. Yield
Rel. activity
Rel. yield
-4
Optimum conditions:
pH=7, [glucose-1,6-diphosphate] = 5.10 M,
[β-mercaptoethanol] = 100 mM, [R-5-P ]= 1 mM, T=45°C.
Enzyme activities and yields are indicated as relative values for
each experiment.
Fig. 2. Optimisation of PPM activity and yield using ribose 5-phosphate
and adenine as model.
Using the previously synthesised ribose 5-phosphate (5a,
Fig. 1), a set of natural and modified nucleosides were
obtained using PPM and PNP as biocatalysts. Hypoxan-
thine was rapidly converted to inosine (2 h, 6, Fig. 3) in
quantitative yield. When adenine was used, the relative
activity was lower (compound 7, Fig. 3) but high reaction
yields were achieved at longer reaction times. Then, a set
of different substituted purines: 6-mercaptopurine, 6-
chloro-2-aminopurine and 2-fluoropurine were assayed.
Although activity and yield were acceptable for the prepa-
ration of the 6-mercaptopurine ribonucleoside (8, Fig. 3),
results were not so positive in the case of 6-chloro-2-amino-
purine ribonucleoside (9, Fig. 3) and 2-fluoroadenosine (10,
Fig. 3). An interesting example was the synthesis of Ribavi-
rine (Virazole) from 5a and 1,2,4-triazole-3-carboxamide
(11, Fig. 3). The reaction proceeded quantitatively and with
high activity in 1.5 h.
In this work, the preparation conditions for the chemo-
enzymatic obtainment of arabino, ribo and deoxy nucleo-
sides have been explored, using different pyrimidine and
purine bases. Experiments with diverse NPs sources and
efforts to immobilise PPM are in progress.
Acknowledgements
This work was financially supported by ANPCyT (PICT
REDES 2003-00300, PICT 02: 1410582, PICT 2005 33431),
CONICET (PIP 5492) and UBA (UBACyT X073).
Then, arabinose 5-phosphate (5b, Fig. 1) was used as a
substrate of PPM and bases such as hypoxanthine, ade-

M. Taverna-Porro et al. / Tetrahedron Letters 49 (2008) 2642–2645
2645
7. See for example: Lewkowics, E. S.; Iribarren, A. M. Curr. Org. Chem.
2006, 10, 1197–1215.
J.M.M., L.A.B., C.A.P. and A.M.I. are CONICET
members.
´
8. See for example: Medici, R.; Lewkowicz, E. S.; Iribarren, A. M. J.
Mol. Catal. B: Enzym. 2006, 39, 40–44.
Supplementary data
9. Komatsu, H.; Araki, T. Tetrahedron Lett. 2003, 44, 2899–2901.
10. (a) MacDonald, D. L.; Fletcher, H. G., Jr. J. Am. Chem. Soc. 1961,
84, 1262–1265; (b) Friedkin, M.; Kalckar, H. M. J. Biol. Chem. 1950,
184, 437–448.
11. Bonney, R. J.; Weinfeld, H. J. Bacteriol. 1970, 103, 650–655.
12. Galperin, M. Y.; Bairoch, A.; Koonin, E. V. Prot. Sci. 1998, 7, 1829–
1835.
Detailed experimental procedures for the protein expres-
sion and purification, synthesis of furanose 5-phosphates
and analytical methods are available. Supplementary data
associated with this article can be found, in the online ver-
sion, at doi:10.1016/j.tetlet.2008.02.087.
13. Valentin-Hansen, P.; Hammer, K.; Lø Larsen, J. E.; Svendsen, I.
Nucleic Acids Res. 1984, 12, 5211–5224.
14. Barbas, C. F., III; Wong, C.-H. Biorg. Chem. 1991, 19, 261–269.
15. Ouerkerk, N.; Steenweng, M.; de Ruijter, M.; Brouwer, J.; Van
Boom, J. H.; Lugtenburg, J.; Raap, J. J. Org. Chem. 2002, 67, 1480–
1489.
16. Horinouchi, N.; Ogawa, J.; Kawano, T.; Sakai, T.; Saito, K.;
Matsumoto, S.; Sasaki, M.; Mikami, Y.; Shimizu, S. Biotechnol.
Lett. 2006, 28, 877–881.
References and notes
1. Ichikawa, E.; Kato, K. Curr. Med. Chem. 2001, 8, 385–423.
2. Agrofoglio, L. A.; Gillaizeau, I.; Saito, Y. Chem. Rev. 2003, 103,
1875–1916.
3. Vorbruggen, H. Acta Biochim. Pol. 1996, 43, 25–36.
4. Biotransformations in Organic Chemistry: A Textbook; Faber, K., Ed.,
5th ed.; Springer: Heilderberg, 2004.
17. Prasad, A. K.; Serensen, M. D.; Parmar, V. S.; Wengel, J. Tetrahedron
Lett. 1995, 36, 6163–6166.
5. See for example: (a) Ferrero, M.; Gotor, V. Chem. Rev. 2000,
100, 4319–4347; (b) Panero, J.; Trelles, J.; Rodano, V.; Montserrat, J.
M.; Iglesias, L. E.; Lewkowicz, E. Biotech. Lett. 2006, 28, 1077–
1081.
18. Inigo, S.; Taverna-Porro, M.; Montserrat, J. M.; Iglesias, L. E.;
˜
Iribarren, A. M. J. Mol. Catal. B: Enzym. 2005, 35, 70–73.
19. Hennen, W. J.; Sweers, H. M.; Wang, Y.-F.; Wong, C.-H. J. Org.
Chem. 1988, 53, 4939–4945.
6. See for example: Bzowska, A.; Kulilowska, E.; Shugar, D. Pharmacol.
Therap. 2000, 88, 349–425.
20. Bennet, E. M.; Li, C.; Allan, P. W.; Parker, P. W.; Ealick, S. E. Biol.
Chem. 2003, 278, 47110–47118.