Stereoselective synthesis of dinucleoside boranophosphates by an oxazaphospholidine method

Article abstract of DOI:10.1016/j.bmcl.2006.03.076

A stereoselective synthesis of dinucleoside boranophosphates by using nucleoside 3′-oxazaphospholidine derivatives is described. The diastereoselectivity of the internucleotidic bond formation reactions varied with the nucleobase used. (Rp)- and (Sp)-dith

Full text of DOI:10.1016/j.bmcl.2006.03.076

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3111–3114
Stereoselective synthesis of dinucleoside boranophosphates
by an oxazaphospholidine method
Takeshi Wada,* Yukihiro Maizuru, Mamoru Shimizu,
Natsuhisa Oka and Kazuhiko Saigo
Department of Medical Genome Sciences, Graduate School of Frontier Sciences, The University of Tokyo,
5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8562, Japan
Received 27 February 2006; revised 16 March 2006; accepted 20 March 2006
Available online 18 April 2006
Abstract—A stereoselective synthesis of dinucleoside boranophosphates by using nucleoside 30-oxazaphospholidine derivatives is
described. The diastereoselectivity of the internucleotidic bond formation reactions varied with the nucleobase used. (Rp)- and
(Sp)-dithymidine boranophosphates were synthesized with excellent diastereoselectivity both in solution and on a solid-support,
whereas a loss of diastereopurity was observed for the 20-deoxycytidine derivative having an unprotected nucleobase amino group.
On the other hand, complete chemoselectivity of the 30-oxazaphospholidine derivatives toward hydroxy groups over amino groups
was serendipitously found during the study. This unique chemoselectivity of the 30-oxazaphospholidine derivatives was investigated
by comparing them with the conventional nucleoside 30-phosphoramidite.
Ó 2006 Elsevier Ltd. All rights reserved.
Boranophosphate DNA is a new class of modified
nucleic acids, in which one of the non-bridging oxygens
of the phosphodiester linkage is replaced by a BH3
group.1 Incorporation of the boranophosphate linkages
into oligonucleotides results in significant increase of
their nuclease resistance and lipophilicity.2,3 In addition,
a duplex consisting of a boranophosphate DNA and its
complementary RNA is a good substrate for RNase
H.2,4 Thus, boranophosphate DNA is regarded as a
promising candidate for therapeutic agents applicable
to antisense and antigene approaches as well as boron
neutron capture therapy (BNCT).5
mers is virtually impossible in the cases of long oligo-
mers. Just and co-workers have demonstrated the
stereocontrolled synthesis of (Sp)-dithymidine borano-
phosphate with excellent diastereoselectivity by using
(S)-3-hydroxy-4-(2-indolyl)butyronitrile as a chiral aux-
iliary, though it has not been applied to the synthesis of
boranophosphate DNA oligomers.8 In contrast, fully
(Sp)-stereoregulated boranophosphate DNA can be ob-
tained by the enzymatic method using nucleoside 50-O-
a-boranotriphosphates.9 However, (Rp)-boranophos-
phate DNA cannot be obtained by the enzymatic meth-
od because of the substrate specificity of the enzyme.
Under these circumstances, the development of an effi-
cient method for the chemical synthesis of stereodefined
boranophosphate DNA is of great importance.
Boranophosphate DNA has chiral internucleotidic link-
ages, and it has been reported that the chirality affects
the properties of boranophosphate DNA.3,4b Shaw
and co-workers and Just and co-workers have reported
the preparations of (Rp)- and (Sp)-dithymidine borano-
phosphates based on chromatographic separations.6,7
These approaches can be applicable only to dinucleoside
boranophosphates, because the separation of diastereo-
Recently, we have developed an oxazaphospholidine
approach for the stereocontrolled synthesis of phosp-
horothioate DNA by the use of nucleoside 30-O-oxaza-
phospholidine monomer units and less-nucleophilic acid
activators.10 The method enables us to synthesize (Rp)-
and (Sp)-phosphite triester intermediates in high yields
and with excellent diastereoselectivity. The resulting dia-
stereopure phosphites are expected to be converted to the
corresponding diastereopure boranophosphates. In this
paper, we wish to describe a novel approach for the dia-
stereoselective synthesis of dinucleoside boranophos-
phates by the oxazaphospholidine method.
Keywords: Boranophosphate DNA; Stereoselective synthesis;
Antisense DNA; Solid-phase synthesis; Oxazaphospholidine;
Chemoselectivity.
*
Corresponding author. Tel.: +81 4 7136 3612; fax: +81 4 7136
3612; e-mail: wada@k.u-tokyo.ac.jp
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.03.076

3112
T. Wada et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3111–3114
As we previously reported, the diastereopure 50-O-(tert-
butyldiphenylsilyl)thymidine 30-O-oxazaphospholidine
monomer units (Sp)-1 and (Rp)-1 were obtained from
the corresponding enantiopure 1,2-amino alcohols with
the diastereomer ratios of >99:1 and 99:1, respectively.10
The monomer (Sp)-1 was condensed with 30-O-(tert-
butyldimethylsilyl)thymidine 2 in the presence of the
activator 3 to give the diastereopure phosphite interme-
diate (Rp)-4 (dr > 99:1).10 The resultant phosphite was
then boronated by treatment with 1 M BH3ÆTHF at rt
for 10 min to afford the boranophosphate triester
(Rp)-5 (Scheme 1). The chiral auxiliary in (Rp)-5 could
be easily removed by treatment with DBU for 30 min
at 50 °C to give the 50-O- and 30-O-silylated dithymidine
boranophosphate (Sp)-6. Finally, the 50-O- and 30-O-si-
lyl groups were removed by treatment with 3HFÆEt3N,11
and purification by reverse-phase column chromatogra-
phy gave the fully deprotected dimer (Sp)-7a in 66% iso-
lated yield (4 steps). The resultant dimer was almost
diastereopure (dr = 98:2), which was confirmed by re-
verse-phase HPLC. The P-configuration of the dimer
HO
Th
B
DMTrO
B
DMTrO
O
O
O
N
O
O
O
O
P
P
+
Th
O
9
N
O
O
H2
Ph
(i)
Ph
O
B = Th: (Rp)-8a
B = Cy: (Rp)-8b
B = Th: (Sp)-10a
B = Cy: (Sp)-10b
(ii) boronation
B
HO
O
B
RO
O
H3B
–O
O
O
(iv) deprotection
P
H3B
O
Th
P
O
+
Th
O
O
N
O
H2
Ph
O
O
B = Th: (Rp)-13a
B = Cy: (Rp)-13b
R = DMTr,
(v)
B = Th: (Sp)-11a,
B = Cy: (Sp)-11b
(iii)
B
HO
O
R = H,
1
B = Th: (Sp)-12a,
B = Cy: (Sp)-12b
was determined to be Sp by H, 31P NMR and RP-
H3B
–O
O
O
HPLC analysis.3,6,12 Previously, we have found that
the condensation of (Sp)-1 with 2 in the presence of 3
proceeded with inversion of the configuration at the
phosphorus atom and that the removal of the chiral aux-
iliary by DBU treatment took place with retention of the
configuration.10 Moreover, it is known that the conver-
sion of a phosphite triester to a boranophosphate by
BH3ÆTHF proceeds with retention of the P-configura-
tion. Therefore, the absolute P-configuration of the
resultant dithymidine boranophosphate is consistent
with those reported in the previous studies.10 In a similar
manner, (Rp)-7a could be obtained from (Rp)-1 in 63%
isolated yield (4 steps) with the dr of 96:4.
P
Th
O
OH
B = Th: (Rp)-7a, Rp : Sp = 98 : 2
B = Cy: (Rp)-7b, Rp : Sp = 76 : 24
Scheme 2. Solid-phase synthesis of (Rp)-dinucleoside boranophos-
phates (Rp)-7a,b. Reagents and conditions: (i) 3, CH3CN, rt, 3 min; (ii)
see Table 1; (iii) 3% DCA, Et3SiH–CH2Cl2 (1:1, v/v), rt, 30 s; (iv) see
Table 1; (v) concd NH3, 50 °C, 30 min.
trolled-pore glass (CPG) (9) in the presence of 3, and
the resultant phosphite triester (Sp)-10a was boronated
under various conditions (Table 1). The 50-O-DMTr
group of (Sp)-11a was then removed by treatment with
3% DCA in CH2Cl2 in the presence of Et3SiH as a trityl
cation scavenger.13 After removal of the chiral auxiliary
in (Sp)-12a, the dimer was cleaved from the solid-sup-
port by treatment with concd NH3 at 50 °C for
30 min, and analyzed by reverse-phase HPLC. When
the boronation of the phosphite intermediate (Sp)-10a
and the deprotection of the chiral auxiliary in (Sp)-12a
were carried out under similar conditions to those for
the solution-phase synthesis, the HPLC profile of the
crude product was very complicated (Table 1, entry 1).
A prolonged boronation reaction was found to be less
effective (entry 2). Milder deprotection conditions result-
ed in the improved yield of the dimer (entry 3), though
many unidentified byproducts were still observed. When
BH3ÆMe2S was used as a boronating agent in place of
BH3ÆTHF, the yield of the dimer was appreciably im-
proved (entry 4). During the deprotection of (Sp)-12a,
the 50-hydroxy group would attack the neighboring
phosphorus atom to decompose the product. In order
to avoid this undesirable side reaction, the 50-hydroxy
group was acetylated prior to the DBU treatment (entry
5). This treatment was found to be very effective to avoid
the decomposition of the product. Under the optimized
conditions, (Rp)- and (Sp)-dithymidine boranophos-
phates were synthesized with the dr of 98:2 (94% yield)
and 99:1 (92% yield), respectively.14
On the basis of these results, we applied this method to
the solid-phase synthesis of diastereopure dithymidine
boranophosphates (Scheme 2). The monomer (Rp)-
8a10 was condensed with thymidine anchored to a con-
CN
H –OTf
N+
TBDPSO
Th
TBDPSO
Th
O
O
N
3
O
O
O
P
3'-O-(TBDMS)thymidine 2
P
+
N
Th
O
CH3CN
O
O
H2
Ph
Ph
OTBDMS
(Rp)-4
(Sp)-1
BH3•THF
R1O
H3B
Th
O
Th
TBDPSO
H3B
O
DBU
O
O
O
P
CH3CN
–O
O
P
Th
+
N
Th
O
O
O
H2
Ph
OR2
OTBDMS
(Rp)-5
R1 = TBDPS, R2 = TBDMS:
(Sp)-6
Et3N•3HF
R1 = OH, R2 = OH:
(Sp)-7a
Scheme 1. Solution-phase synthesis of (Sp)-dithymidine boranophos-
phate (Sp)-7a.

T. Wada et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3111–3114
Table 1. Reaction conditions for boronation and deprotection
3113
Entry
Boronation
Deprotection
T
PBT:Ta
1
1 M BH3ÆTHF/THF/10 min
1 M BH3ÆTHF/THF/30 min
1 M BH3ÆTHF/THF/30 min
1 M BH3ÆMe2S/CH2Cl2/30 min
1 M BH3ÆMe2S/CH2Cl2/30 min
0.2 M DBU/50 °C/30 min
0.2 M DBU/50 °C/30 min
0.2 M DBU/25 °C/overnight
0.2 M DBU/25 °C/overnight
0.2 M DBU/25 °C/overnight
53:47b
55:45b
70:30b
89:11
93:7
2
3
4
5c
a The ratios were determined by HPLC.
b Many side products were observed other than TPBT.
c Acetylation was carried out before the DBU treatment.
Next, we attempted to apply this method to the synthe-
sis of dinucleoside boranophosphates having nucleobas-
es other than thymine. It has been reported that
boronating reagents, such as BH3ÆMe2S, reduce the con-
ventional acyl protecting groups for nucleobase amino
groups into the corresponding alkyl groups (e.g., benzo-
yl to benzyl). The resultant alkyl groups on the nucleo-
base amino groups would be difficult to remove from the
products.3,15,16 On the other hand, the reactions between
the unprotected nucleobases and boronating reagents
have been reported to be reversible coordinations of
the nucleobase nitrogen atoms to BH3 rather than irre-
versible reductions of the nucleobases.17,18 These reports
suggest that nucleoside 30-oxazaphospholidine mono-
mers having an unprotected nucleobase should be used.
First, a monomer unit having N4-unprotected cytosine
[(Rp)-8b] was synthesized. The phosphitylation of the
30-OH was carried out at ꢀ78 °C to minimize undesired
phosphitylations at the nucleobase to afford the diaste-
reopure oxazaphospholidine monomer (Rp)-8b in 44%
yield after silica gel column chromatography. (Rp)-8b
was then applied to the synthesis of CPBT on the CPG
served (Fig. 1A).10b,20 On the contrary, the experiment
using the conventional thymidine 30-phosphoramidite
15 in the place of 8a gave a complex mixture, in which
several broad signals around 132–123 ppm were ob-
served by 31P NMR probably due to the N-phosphityla-
tion (Fig. 1B). Similarly, 8a did not give any observable
adducts with the adenine and guanine amino groups
either, whereas 15 gave broad 31P NMR signals around
132–123 ppm. Our first assumption was that the acid
activator 3 mediated the chemoselective phosphityla-
tions because the chemoselectivity of the O-selective
phosphitylations reported in the literature was depen-
dent on activators.21 However, we found that 8a did
not give any N-phosphitylated products even when the
conventional acidic activator for the phosphoramidite
method, 1H-tetrazole, was used in the place of 3, whereas
Th
DMTrO
NH2
N
O
N
O
P
CN
N+
H –OTf
TBDMSO
N
O
+
+
O
(Scheme 2). The HPLC analysis showed that CPB
T
O
was generated in 86% yield without any base modifica-
tions. However, a significant loss of diastereopurity
was observed (dr = 76:24). Although the exact mecha-
nism is still not clear, the contributor to this low dia-
stereoselectivity must be the unprotected cytosine
amino group judging from the fact that TPBT [(Rp)-
and (Sp)-7a] were obtained with little loss of diastereo-
purity. This implies the necessity for the development
of a new base-protecting group for the nucleobase ami-
no groups, which are compatible with boronating
reagents, for the synthesis of stereoregulated borano-
phosphate DNA having nucleobases other than thy-
mine. A study based on this consideration is currently
in progress.
OTBDMS
14
Ph
3
trans–8a
4 equiv
2 equiv
trans–8a
trans–8a
30 min
cis–8a
145
TBDMSO
15
140
135
130
125
ppm
145
140
135
130
125
ppm
DMTrO
Th
O
N
NH2
N
O
CN
N+
H –OTf
N
O
NC
P
+
+
O
O
On the other hand, it is noteworthy that the CPBT ob-
tained by this method was not accompanied with any
N-phosphorylated products, because it is well-known
that the nucleobase amino group of cytosine, as well as
that of adenosine, is reactive enough to be phosphitylat-
ed by the conventional phosphoramidites.19 To confirm
that the oxazaphospholidine monomer units do not give
any adducts with the nucleobase amino groups, trans-8a
(a mixture of (Rp)-8a and (Sp)-8a) was mixed with 30,50-
bis(tert-butyldimethylsilyl)cytidine 14 and 3 in CH3CN–
CD3CN (4:1, v/v), and analyzed by 31P NMR. No ad-
ducts between trans-8a and 14 were observed; only the
P-epimerization from the trans-8a to the cis-8a was ob-
OTBDMS
14
3
15
2 equiv
4 equiv
30 min
150
145
140
135
130
125 ppm
150
145
140
135
130
125 ppm
Figure 1. 31P NMR spectra of the reaction mixtures obtained by the
reactions of 14 with trans-8a (A) or with 15 (B) in the presence of 3 (4
equiv) in CH3CN–CD3CN (4:1, v/v) at rt.

3114
T. Wada et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3111–3114
2. (a) Zhang, J.; Terhorst, T.; Matteucci, M. D. Tetrahedron
RO
Lett. 1997, 38, 4957; (b) Higson, A. P.; Sierzchala, A.;
Brummel, H.; Zhao, Z.; Caruthers, M. H. Tetrahedron
Lett. 1998, 39, 3899; (c) Sergueev, D. S.; Shaw, B. R.
J. Am. Chem. Soc. 1998, 120, 9417; (d) Brummel, H. A.;
Caruthers, M. H. Tetrahedron Lett. 2002, 43, 749.
3. Sergueeva, Z. A.; Sergueev, D. S.; Ribeiro, A. A.; Summers,
J. S.; Shaw, B. R. Helv. Chim. Acta 2000, 83, 1377.
4. (a) Rait, V. K.; Shaw, B. R. Antisense Nucleic Acid Drug
Dev. 1999, 9, 53; (b) Wang, X.; Dobrikov, M.; Sergueev,
D.; Shaw, B. R. Nucleosides Nucleotides 2003, 22, 1151.
5. Hawthorne, M. F. Angew. Chem. Int. Ed. Engl. 1993, 32,
1044.
OR
P
HN
NH2
+
P
H2N
O
O
N+
N
+
N
Ph
H
intramolecular
re-cyclization
N
O
Ph
N
O
protonated
trans-8a
14
Not detectable by 31P NMR
Figure 2. Plausible mechanism of the intermolecular nucleophilic
addition between 14 and trans-8a, and the intramolecular re-
cyclization.
15 gave the N-phosphitylated nucleosides.14 To the best
of our knowledge, this is the first example of an O-selec-
tive phosphitylation the chemoselectivity of which is
independent of activators. This chemoselectivity peculiar
to the oxazaphospholidine derivatives can be explained
by the intramolecular re-cyclization of the oxazaphosp-
holidine ring, which would be much faster than the inter-
molecular nucleobase phosphitylations (Fig. 2). This new
O-selective phosphitylation may lead to a new synthetic
method for nucleic acid analogs without nucleobase
amino protection.
6. Sergueeva, Z. A.; Sergueev, D. S.; Shaw, B. R. Tetrahe-
dron Lett. 1999, 40, 2041.
7. Jin, Y.; Just, G. Tetrahedron Lett. 1998, 39, 6429.
8. Jin, Y.; Just, G. Tetrahedron Lett. 1998, 39, 6433.
9. (a) Li, H.; Porter, K.; Huang, F.; Shaw, B. R. Nucleic
Acids Res. 1995, 23, 4495; (b) Sergueev, D.; Hasan, A.;
Ramaswamy, M.; Shaw, B. R. Nucleosides Nucleotides
1997, 16, 1533; (c) He, K.; Porter, K. W.; Hasan, A.;
Briley, J. D.; Shaw, B. R. Nucleic Acids Res. 1999, 27,
1788.
10. (a) Oka, N.; Wada, T.; Saigo, K. J. Am. Chem. Soc. 2002,
124, 4962; (b) Oka, N.; Wada, T.; Saigo, K. J. Am. Chem.
Soc. 2003, 125, 8307.
11. Gasparutto, D.; Livache, T.; Bazin, H.; Duplaa, A.-M.;
Guy, A.; Khorlin, A.; Molko, D.; Roget, A.; Teoule, R.
Nucleic Acids Res. 1992, 20, 5159.
12. Li, H.; Huang, F.; Shaw, B. R. Bioorg. Med. Chem. 1997,
5, 787.
13. Wada, T.; Shimizu, M.; Oka, N.; Saigo, K. Tetrahedron
Lett. 2002, 43, 4137.
14. See Supplementary data.
15. Sergueeva, Z. A.; Sergueev, D. S.; Shaw, B. R. Nucleosides
Nucleotides 2000, 19, 275.
16. Sergueeva, Z. A.; Sergueev, D. S.; Shaw, B. R. Nucleosides
Nucleotides 2001, 20, 941.
17. Sood, A.; Spielvogel, B. F.; Shaw, B. R. J. Am. Chem. Soc.
1989, 111, 9234.
18. Sergueev, D. S.; Sergueeva, Z. A.; Shaw, B. R. Nucleosides
Nucleotides 2001, 20, 789.
19. Wada, T.; Moriguchi, T.; Sekine, M. J. Am. Chem. Soc.
1994, 116, 9901.
In summary, the stereocontrolled synthesis of dithymi-
dine boranophosphates by the oxazaphospholidine
method afforded (Rp)- and (Sp)-dithymidine borano-
phosphates with excellent diastereoselectivity both in
solution- and solid-phase syntheses. On the contrary,
the application of this method to the synthesis of dinucle-
oside boranophosphate having an unprotected nucleo-
base amino group resulted in a significant loss of
diastereopurity. However, the analysis of the resultant
dinucleoside boranophosphate showed that the nucleo-
side 30-oxazaphospholidine monomers did not give any
nucleobase-phosphitylated byproducts. 31P NMR analy-
sis of the reactions demonstrated that the chemoselectivity
of the oxazaphospholidine derivatives toward hydroxy
groups over amino groups was independent of activators.
20. We have reported that the nucleoside 30-oxazaphosphol-
idine derivatives having a 3-methyl-5-phenyl-1,3,2-oxaza-
phospholidine ring, such as (Rp)-8a,b, epimerize in the
presence of acids, such as 3. However, the loss of
diastereopurity observed for CPBT is not attributable to
the epimerization of (Rp)-8b by 3, because 3 mediated the
condensations of (Rp)- and (Sp)-8a with little loss of
diastereopurity.
21. (a) Fourrey, J.-L.; Varenne, J. Tetrahedron Lett. 1985, 26,
2663; (b) Gryaznov, S. M.; Letsinger, R. L. J. Am. Chem.
Soc. 1991, 113, 5876; (c) Gryaznov, S. M.; Letsinger, R. L.
Nucleic Acids Res. 1992, 20, 1879; (d) Hayakawa, Y.;
Kataoka, M. J. Am. Chem. Soc. 1998, 120, 12395; (e)
Sekine, M.; Ohkubo, A.; Seio, K. J. Org. Chem. 2003, 68,
5478; (f) Ohkubo, A.; Seio, K.; Sekine, M. Tetrahedron
Lett. 2004, 45, 363; (g) Ohkubo, A.; Ezawa, Y.; Seio, K.;
Sekine, M. J. Am. Chem. Soc. 2004, 126, 10884; (h)
Dabkowski, W.; Ozarek, A.; Tworowska, I. New. J. Chem.
2005, 29, 1396.
Acknowledgment
This work was supported by Grants from the Ministry
of Education, Culture, Sports, Science and Technology,
Japan.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2006.03.076.
References and notes
1. Sood, A.; Shaw, B. R.; Spielvogel, B. F. J. Am. Chem. Soc.
1990, 112, 9000.