Chemo-enzymatic synthesis of 2′,3′-dideoxy-3′-fluoro-β-D-guanosine via 2,3-dideoxy-3-fluoro-α-D-ribose 1-phosphate

Article abstract of DOI:10.1016/S0040-4039(03)00428-3

2,3-Dideoxy-3-fluoro-α-D-ribose 1-phosphate 2 was stereoselectively synthesized and converted to 2′,3′-dideoxy-3-fluoro-β-D-guanosine 1 by enzymatic reaction using purine nucleoside phosphorylase. This chemo-enzymatic strategy was first applied to the synthesis of 1.

Full text of DOI:10.1016/S0040-4039(03)00428-3

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 2899–2901
Chemo-enzymatic synthesis of
2%,3%-dideoxy-3%-fluoro-b-
D
-guanosine via
2,3-dideoxy-3-fluoro-a-
D
-ribose 1-phosphate
Hironori Komatsua,* and Tadashi Arakib
aCatalysis Science Laboratory, Mitsui Chemicals, Inc., 580-32 Nagaura, Sodegaura-shi, Chiba 299-0265, Japan
bLife Science Laboratory, Mitsui Chemicals, Inc., 1144 Togo, Mobara-shi, Chiba 297-0017, Japan
Received 26 November 2002; revised 12 February 2003; accepted 14 February 2003
Abstract—2,3-Dideoxy-3-fluoro-a-
fluoro-b- -guanosine 1 by enzymatic reaction using purine nucleoside phosphorylase. This chemo-enzymatic strategy was first
applied to the synthesis of 1. © 2003 Elsevier Science Ltd. All rights reserved.
D-ribose 1-phosphate 2 was stereoselectively synthesized and converted to 2%,3%-dideoxy-3-
D
Stereoselective glycosylation has attracted considerable
attention in the field of nucleoside chemistry. For
medicinal use, by-product content must be strictly con-
trolled. Therefore, synthetic methods that do not pro-
duce isomers are preferable. A number of methods have
been developed;1 however, glycosylation with a guanine
base 3 has been an obstacle for application to scalable
chemical preparations. In contrast, enzymatic reaction
a stereoselective manner and is then enzymatically con-
verted to 1.
First, chloro sugar 6 was prepared starting from methyl
furanoside 4 (Scheme 2).5 Because 6 is extremely labile
under acidic conditions and prone to decomposition via
furan derivatives by losing HCl and HF even at room
temperature, direct conversion of 4 to 6 was avoided.
Reaction of 4 with Ac2O in AcOH in the presence of
H2SO4 gave an anomeric mixture of 5 (a:b=2:1) fol-
lowed by conversion to 6 (a:b=2:1) using the stoichio-
metric amounts of HCl. Both of the reactions had to be
maintained at 0°C to avoid decomposition of 5 and 6.
with
3 affords b-nucleosides when using purine
nucleoside phosphorylase (PNPase) as the enzyme with
a-D-furanose 1-phosphate as the substrate for the reac-
tion.2 This strategy has not been very successful, how-
ever, due to difficulties in preparing the furanose
1-phosphate. We previously reported chemo-enzymatic
synthesis of 2%-deoxy-b-
synthesis of 2-deoxy-a-
shown in Scheme 1, this strategy is applicable to the
D
-nucleoside via the chemical
Stereoselective phosphorylation of 6 was then investi-
gated. Even though b-isomer of 2 was not likely to be
recognized by PNPase and would not give an a-isomer
of 1, high stereoselectivity of the phosphorylation was
required to increase the isolated yield of 1. A coupling
D
-ribose 1-phosphate.3 As
synthesis of 2%,3%-dideoxy-3%-fluoro-b- -guanosine 1 that
D
has prominent antiviral activity.4 Here we report
n
chemo-enzymatic synthesis of 1, in which 2,3-dideoxy-
reaction of 6 with o-H3PO4 in the presence of Bu3N
3-fluoro-a-
D
-ribose 1-phosphate 2 is first synthesized in
was performed. In the first attempt, the reaction gave 7
Scheme 1. Retrosynthesis of 2%,3%-dideoxy-3%-fluoro-b-D-guanosine 1.
* Corresponding author. Tel.: +81-438-64-2313; fax: +81-438-64-2371; e-mail: hironori.komatsu@mitsui-chem.co.jp
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)00428-3

2900
H. Komatsu, T. Araki / Tetrahedron Letters 44 (2003) 2899–2901
Scheme 3. Reagents and conditions: (a) aq. KOH, 48 h, 50°C;
(b) 3, CaCl2, PNPase, 40–60°C, 5 h (63% from 8).
(Hb). The a-7 was stabilized in solvents with a large
dielectric constant (CH3CN and PC) and the stabiliza-
tion led to the high stereoselectivity (compare entries
2–5 with entry 1). The relatively larger dipole moment
of a-7 (3.72 Debye) than of b-7 (3.62 Debye) was
attributable to the stabilization. The largest value of the
dielectric constant of PC, however, did not directly
affect the heat of formation energy (entry 5). The
accurate ionized state of phosphate 7 in the reaction
mixture should be used to calculate the heat of forma-
tion for a more precise explanation of the effect of PC.
A synthetic sample of 7 was prepared in a CH3CN/4-
methyl-2-pentanone (MIBK) (10:1) mixture solvent.
HPLC analysis of the reaction revealed that the solvent
system affords a stereoselectivity of 78:22 (a-7:b-7).
Phosphate 8 was isolated as cyclohexylammonium salt
in 59% yield from 4 and the stereoselectivity was 80:20
(a:b). The structure of 8 was confirmed by spectro-
scopic methods (1H, 13C, and 31P NMR, and IR) and
mass spectral analysis.9 Without further purification,
deprotection was performed in MeOH in the presence
of cyclohexylamine. Because the reaction was too slow
(50 days at room temperature for completion) to scale
up this reaction condition, saponification in aqueous
KOH was performed. The potassium salt of the desired
compound 10 was obtained as an aqueous solution
after filtration to remove potassium 4-phenylbenzoate
(Scheme 3). The structure of the substrate for the key
enzymatic reaction was confirmed as cyclohexylammo-
nium salt 9 (a:b=87:13) by spectroscopic methods (1H,
13C, and 31P NMR, and IR) and mass spectral
analysis.10
Scheme 2. Reagents and conditions: (a) Ac2O, AcOH, H2SO4,
0°C, 2.5 h; (b) 4N HCl/1,4-dioxane, c-C6H12/PhCH3, 0°C, 8
,
h; (c) o-H3PO4, nBu3N, 4 A MS, CH3CN/MIBK, 0°C, 6 h,
then c-C6H11NH2 (59% from 4); (d) c-C6H11NH2, MeOH, rt,
50 days (79%).
with no stereoselectivity at the anomeric C1-position
(a:b=ca. 1:1). Under acidic conditions in the presence
of excess amounts of o-H3PO4, however, equilibration
between a- and b-7 occurred and gradually shifted to
the thermodynamically more stable a-7.
The selectivity was dependent on the solvent used for
the reaction. As depicted in Table 1, a solvent with a
large dielectric constant led to high stereoselectivity at
the state of equilibrium saturation (entries 2–5). The
reaction in propylene carbonate (PC) with the largest
dielectric constant resulted in the highest stereoselectiv-
ity (a-7:b-7=81:19) in our experiments (entry 5). To
investigate the relation between the solvent effect and
stereoselectivity, the heats of formation in each solvent
were calculated with MOPAC/PM37 using the conduc-
tor-like screen model (COSMO) routine;8 the results are
shown in Table 1 as the difference (DH=Ha−Hb)
between the heats of formation of a-7 (Ha) and b-7
Finally, the enzymatic glycosylation was performed by
using the aqueous solution of 10 and guanine 3. Con-
ventional procedure2 was modified in two ways. First, a
smaller amount of 3 was used in this reaction since 3
was difficult to remove by recrystallization in the work-
Table 1. Solvent effect in phosphorylation
Entry
Solvent
a-7:b-7a
Dielectric constantb
Heat of formation DH (kJ/mol)c
1
2
3
4
5
PhCH3
Acetone
2-Butanone
CH3CN
PCd
49:51
71:29
73:27
79:21
81:19
2.4
20.7
18.5
37.5
64.4
1.7
−5.7
−6.2
−8.1
−6.9
a All amomer ratios were analyzed by HPLC.
b All data were taken from Handbook of Organic Chemistry.6
c All values were calculated with MOPAC/PM3 using the COSMO routine and shown as DH (DH=Ha−Hb) where Ha and Hb are the heats of
formation of a-7 and b-7, respectively.
d PC, propylene carbonate.

H. Komatsu, T. Araki / Tetrahedron Letters 44 (2003) 2899–2901
2901
up. Second, CaCl2 was added to enhance the conver-
sion by removing H3PO4 as Ca salts from the reaction
solution. Consequently, 10 was converted to 1 by using
0.84 equiv. of guanine 3 in the presence of bacterial
PNPase2,11 in 71% HPLC yield from 8. Based on the
a-isomer content, the conversion yield was 89% from 8.
The desired 1 was the only product observed in the
HPLC assay and was isolated by crystallization directly
from the reaction mixture in pure form. The isolated
yield was 63% from 8. The structure of 1 was confirmed
by spectroscopic methods (1H and 13C NMR, and IR)
and mass spectral analysis.12 They were identical with
the reported data.4
9. Compound 8 (a:b=80:20): mp 146–150°C; [h]2D8=+2.8° (c
1
2.7, CH3OH); H NMR (CD3OD, 400 MHz): l 8.13 (d,
J=8.3 Hz, 1/5 of 2H), 8.08 (d, J=8.3 Hz, 4/5 of 2H),
7.75 (d, J=8.3 Hz, 2H), 7.67 (d, J=8.3 Hz, 2H), 7.47
(dd, J=8.3 and 7.3 Hz, 2H), 7.39 (m, 1H), 6.02 (m,
1/5H), 5.96 (dd, J=5.6 and 4.6 Hz, 4/5H), 5.34 (bd,
J=54.6 Hz, 1/5H), 5.26 (bd, J=55.6 Hz, 4/5H), 4.71 (bd,
J=25.7 Hz, 4/5H), 4.49 (dd, J=13.2, 4.1 Hz, 1H), 4.42
(dd, J=13.2, 4.1 Hz, 1H), 3.00 (m, 2H), 2.54–2.32 (m,
2H), 2.01 (m, 4H), 1.78 (m, 4H), 1.65 (d, J=12.4 Hz,
2H), 1.32 (m, 8H), 1.20 (m, 2H); 13C NMR (CD3OD, 100
MHz): l 167.58 (b), 167.53 (a), 147.42, 141.07 (a), 147.35
(b), 141.13 (b), 141.07 (a), 131.33 (b), 131.21 (a), 130.08
(a), 129.80 (b), 129.38 (a), 129.83 (b), 129.74 (a), 128.24
(a), 128.19 (a), 128.15 (b), 101.59 (b), 101.20 (d, J=4.1
Hz, a), 95.89 (d, J=177.6 Hz, b), 95.22 (d, J=179.5 Hz,
a), 83.72 (d, J=27.3 Hz, a), 83.23 (d, J=24.8 Hz, b),
66.04 (d, J=9.9 Hz, b), 65.36 (d, J=9.1 Hz, a), 51.25,
41.96 (dd, J=20.7, 6.2 Hz, a), 32.84, 26.11, 25.58; 31P
In summary, stereoselective synthesis of 2,3-dideoxy-3-
fluoro-a-
nium salt 9 and potassium salt 10) was achieved and its
enzymatic conversion to 2%,3%-dideoxy-3%-fluoro-b-
D
-ribose 1-phosphate 2 (its cyclohexylammo-
D
-
guanosine 1 is presented. This chemo-enzymatic strat-
egy produced no isomers of 1, thus providing an
alternative synthetic route for the preparation of vari-
ous unnatural nucleosides.
NMR (CD3OD, 162 MHz): l 1.57 (a), 1.42 (b); IR (cm−1
,
KBr): 3424, 2938, 2858, 1709, 1612, 1561, 1451, 1390,
1279, 1084, 980, 935, 748, 699; MS (APCI) m/z 395
(M−H)−.
10. Compound 9 (a:b=87:13): mp 170–171°C (dec.); [h]2D8=
1
+26.1° (c 5.0, H2O); H NMR (D2O, 400 MHz): l 5.64
Acknowledgements
(dd, J=5.6, 5.6 Hz, 1/6H), 5.61 (dd, J=5.7, 5.4 Hz,
5/6H), 5.02 (dd, J=54.6, 6.1 Hz, 1/6H), 4.96 (dd, J=
54.9, 5.1 Hz, 5/6H), 4.30 (ddd, J=27.1, 5.1, 4.4 Hz,
5/6H), 4.06 (ddd, J=24.2, 5.4, 6.1 Hz, 1/6H), 3.5–3.3 (m,
2H), 2.90 (m, 2H), 2.2–2.1 (m, 2H), 1.74 (m, 4H), 1.56
(m, 4H), 1.41 (d, J=11.7 Hz, 2H), 1.10 (m, 8H), 0.95 (m,
2H); 13C NMR (D2O, 100 MHz): l 100.41 (d, J=4.0 Hz,
a), 95.52 (d, J=173.4 Hz, a), 95.31 (d, J=171.4 Hz, b),
85.83 (d, J=23.8 Hz, a), 85.43 (d, J=23.0 Hz, b), 62.30
(d, J=10.7 Hz, b), 62.02 (d, J=10.7 Hz, a), 51.03, 41.5
(dd, J=18, 5.7 Hz, b), 41.13 (dd, J=18.9, 4.9 Hz, a),
31.32, 25.10, 24.61; 31P NMR (D2O, 162 MHz): l 1.63
(b), 1.57 (a); IR (cm−1, KBr): 2936, 2858, 2560, 2209,
1625, 1561, 1453, 1392, 1243, 1066, 980, 933, 844, 732;
MS (APCI) m/z 215 (M−H)−.
We would like to thank the referees of this paper for
helpful comments. We thank Mr. Hironori Kamachi
for the calculation of the heats of formation, Mr.
Ichirou Ikeda, Mr. Junya Fujiwara and Mr. Hideki
Umetani for helpful discussions.
References
1. For a review, see: Zorbach, W. W. Synthesis 1970, 7, 329.
2. (a) Tarr, H. L. A. Can. J. Biochem. Physiol. 1958, 36,
517; (b) Zintchenko, A. I.; Eroshevskaya, L. A.; Barai, V.
N.; Mikhailopulo, I. A. Nucleic Acids Res. Symp. Ser.
1987, 18, 137.
3. (a) Komatsu, H.; Awano, H.; Tanikawa, H.; Itou, K.;
Ikeda, I. Nucleosides Nucleotides Nucleic Acids 2001, 20,
1291; (b) Komatsu, H.; Awano, H. J. Org. Chem. 2002,
67, 5419.
4. Herdewijn, P.; Balzarini, J.; Baba, M.; Pauwels, R.; Van
Aerschot, A.; Janssen, G.; De Clercq, E. J. Med. Chem.
1988, 31, 2040.
5. Abdel-Bary, H. M.; El-Barbary, A. A.; Khodair, A. I.;
Abdel Megied, A. E. S.; Pedersen, E. B.; Nielsen, C. Bull.
Soc. Chim. Fr. 1995, 132, 149.
11. For a preparation of bacterial PNPase, see: (a) Jensen, K.
F.; Nygaard, P. Eur. J. Biochem. 1975, 51, 253; (b)
Krenitsky, T. A.; Koszalka, G. W.; Tuttle, J. V. Biochem-
istry 1981, 20, 3615.
12. Compound 1: mp 253–255°C (dec.); [h]2D3=−32.7° (c
1
0.0585, H2O); H NMR (DMSO-d6, 270 MHz): l 10.6 (s,
1H), 7.92 (s, 1H), 6.44 (s, 2H), 6.15 (dd, J=9.2, 5.7 Hz,
1H), 5.37 (dd, J=53.7, 4.3 Hz, 1H), 5.11 (bs, 1H), 4.15
(ddd, J=27.0, 5.1, 4.6 Hz, 1H), 3.56 (m, 2H), 2.79 (dddd,
J=39.4, 14.6, 9.2, 4.3 Hz, 1H), 2.58 (ddd, J=21.1, 14.6,
5.7 Hz, 1H); 13C NMR (DMSO-d6, 67.5 MHz): l 156.67,
153.70, 151.00, 135.31, 94.94 (d, J=173.8 Hz), 85.17 (d,
J=22.4 Hz), 82.64, 60.98 (d, J=10.6 Hz), 36.87 (d,
J=20.7 Hz); IR (cm−1, KBr): 3440, 3320, 3166, 1691,
1636, 1600, 1402, 1105, 1058, 945, 780; MS (APCI) m/z
270 (M+H)+.
6. Dean, J. A. Handbook of Organic Chemistry; McGraw-
Hill: New York, 1987.
7. Stewart, J. J. P. QCPE Bull. 1990, 10, 86.
8. Klamt, A.; Schurman, G. J. Chem. Soc., Perkin Trans. 2
1993, 799.