C
Synlett
K. Seio et al.
Letter
To test the applicability of the reaction using ADDP and
the best results were originally obtained using P(n-Bu)3,
DIAD, and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in
acetonitrile.9a On the other hand, our reaction proceeded
smoothly in the presence of P(n-Bu)3 and ADDP in DMF
without the need for a strong base such as DBU, because the
zwitterion generated from the addition of ADDP and P(n-
n-Bu P in DMF, we synthesized various derivatives of stan-
3
dard deoxynucleosides, and the results are shown in Figure
1
. The reaction of 2 and 6-chloropurine gave a 35:65 mix-
12
ture of 5α and 5β in 60% yield. The reaction with unpro-
13
tected thymine gave 6 in 18% yield with no stereoselectiv-
ity. Notably, a mixture of regioisomers in which the thy-
mine and deoxyribose moieties were connected at different
positions, such as N3, O2, and O4, was obtained, indicating
that N1 was the most reactive but not the only nucleophilic
position of thymine. On the other hand, the reaction of 2
with unprotected adenine gave a 30:70 mixture of 7α and
Bu) is sufficiently basic to abstract the proton from the nu-
3
cleobase. Moreover, recent study by Downey and co-work-
9b
ers reported the use of ADDP and P(n-Bu) also gave selec-
3
tively the β-anomers because of the formation of the 1′,2′-ep-
oxyribose intermediate. In our deoxynucleoside synthesis, a
mixture of α- and β-anomers was obtained in a ratio de-
pending on the solvent and the nucleobase structure be-
cause of the lack of the 2′-hydroxy group which is neces-
sary to form the epoxy intermediate. Thus, our method is
useful for the preparation of separable α- and β-anomers,
as in the case of compounds 1, 10, and 12.
In this paper, we present a systematic study of the syn-
thesis of deoxynucleosides by Mitsunobu reaction under
mild conditions. The reaction generated both α and β iso-
mers in a ratio dependent on the nucleobase structure and
the solvent. In particular, the reactions in DMF (Table 1, en-
try 6) gave the highest yield and β-selectivity. This ap-
proach is expected to be applied to the synthesis of unnatu-
ral deoxynucleosides useful for the pharmaceutical and bio-
logical applications.
1
4
7
β
in 65% yield, whereas the reaction with unprotected
cytosine or guanine to give 8 or 9, respectively, was not ef-
fective because of the low solubility of the nucleobases in
DMF.
The developed procedure was also tested for the prepa-
ration of nonstandard deoxynucleosides, namely, previous-
ly reported 1, 10 (Figure 1), and 12 (Scheme 1). Although
the α and β isomers of above natural nucleoside derivatives
could not be separated because of the similar physicochem-
ical properties of them, the isomers of the below unnatural
nucleosides could be separated by silica gel chromatogra-
phy. The reaction of 4,5-dicyanoimidazole with 2 gave im-
idazole nucleoside 10. Unlike in the case of standard deoxy-
nucleoside derivatives, the isomers of 10 were separated by
silica gel column chromatography to give 10α and 10β in
4
2% and 46% yield, respectively. As previously reported, the
reaction with a protected 7-deazaguanine derivative gave
α and 1β in 23% and 61%, respectively, after purification by
Funding Information
1
This work was supported by JSPS KAKENHI Grant Number 15H01062,
silica gel column chromatography. The Mitsunobu reaction
was also applicable to the synthesis of 5-fluorouridine de-
rivative 12. In this case, compound 2 was coupled with N -
benzoyl-5-fluorouracil (11) to give the fully protected de-
oxynucleoside, which was treated with aqueous ammonia
to afford 12α and 12β in 24% and 59% yield, respectively, af-
ter silica gel column chromatography.
15K13738, and 26288075
S
JP
S
K
A
K
E
N
H
I
Grant
1(
5
H
0
1
0
6
2SJ)P
S
K
A
K
E
N
H
I
Grant
1(
5
K
1
3
7
3
8SJ)P
S
K
A
K
E
N
H
I
Grant
2(
6
2
8
8
0
7
5)
3
Supporting Information
Supporting information for this article is available online at
https://doi.org/10.1055/s-0036-1588445.
S
u
p
p
o
nrtogI
i
f
rm oaitn
S
u
p
p
ortioIgnfmr oaitn
2
(3.0 equiv)
References and Notes
O
ADDP (3.0 equiv)
P(n-Bu)3 (3.0 equiv)
EtOH–28% NH4OH
(1:3, v/v)
F
NBz
(1) McGuigan, C.; Yarnold, C. J.; Jones, G.; Velázquez, S.; Barucki, H.;
Brancale, A.; Andrei, G.; Snoeck, R.; De Clercq, E.; Balzarini, J.
J. Med. Chem. 1999, 42, 4479.
DMF
r.t., 2 h
r.t., 9 h
N
H
O
F
11
(
2) (a) Tokugawa, M.; Masaki, Y.; Canggadibrata, J. C.; Kaneko, K.;
Shiozawa, T.; Kanamori, T.; Grøtli, M.; Wilhelmsson, L. M.;
Sekine, M.; Seio, K. Chem. Commun. 2016, 52, 3809. (b) Sandin,
P.; Lincoln, P.; Brown, T.; Wilhelmsson, L. M. Nat. Protoc. 2007,
2, 615.
TBSO
O
N
O
O
NH
OTBS
1
1
2α: 24%
2β: 59%
(3) (a) Okamoto, I.; Miyatake, Y.; Kimoto, M.; Hirao, I. ACS Synth.
Biol. 2016, 5, 1220. (b) Winiger, C. B.; Shaw, R. W.; Kim, M.-J.;
Moses, J. D.; Matsuura, M. F.; Benner, S. A. ACS Synth. Biol. 2017,
Scheme 1 Preparation and separation of the α and β isomers of a 5-
fluorodeoxyuridine derivative
6
, 194. (c) Malyshev, D. A.; Romesberg, F. E. Angew. Chem. Int.
Ed. 2015, 54, 11930. (d) Cavanagh, B. L.; Walker, T.; Norazit, A.;
Meedeniya, A. C. Molecules 2011, 16, 7980.
4) (a) Marfurt, J.; Parel, S. P.; Leumann, C. J. Nucleic Acids Res. 1997,
A comparison between our deoxynucleoside synthesis
and the Mitsunobu-based ribonucleoside synthesis report-
ed by Hocek et al. shows important differences in terms of
(
25, 1875. (b) Timofeev, E. N.; Goryaeva, B. V.; Florentiev, V. L.
9
J. Biomol. Struct. Dyn. 2006, 24, 183. (c) Kolganova, N. A.;
results and conditions. First, in the ribonucleoside synthesis
Shchyolkina, A. K.; Chudinov, A. V.; Zasedatelev, A. S.;
©
Georg Thieme Verlag Stuttgart · New York — Synlett 2017, 28, A–D