An oxazaphospholidine approach for the stereocontrolled synthesis of oligonucleoside phosphorothioates

Article abstract of DOI:10.1021/ja034502z

The stereocontrolled synthesis of oligodeoxyribonucleoside phosphorothioates (PS-ODNs) using nucleoside 3′-O-oxazaphospholidine derivatives as monomer units is described. 2-Chloro-1,3,2-oxazaphospholidine derivatives were prepared from six kinds of enantiopure 1,2-amino alcohols and used for the phosphitylation reactions of 5′-O-protected nucleosides. A detailed study of these reactions revealed that the diastereoselectivity of the reaction depended on the structure of the enantiopure 1,2-amino alcohol, the reaction temperature, and the amine used as a scavenger of HCI. In addition, ab initio molecular orbital calculations for the 2-chlorooxazaphospholidine derivatives were carried out to elucidate the mechanism of these diastereoselective phosphitylation reactions. The LUMO of the 2-chloro-5-phenyloxazaphospholidine derivatives on the phosphorus atom was found to be almost orthogonal to the P-CI bond. This LUMO may be involved in the phosphitylation reactions with predominant retention of the P-configuration. A series of dialkyl(cyanomethyl)ammonium salts were developed and used as activators for the condensation reactions of the diastereopure nucleoside 3′-O-oxazaphospholidines with 3′-O-protected nucleosides. In the presence of the new activators, the reactions proceeded rapidly to give the corresponding dinucleoside phosphite triesters. The diastereoselectivity of the condensation reaction did not depend on the counteranion but on the structure of the dialkyl(cyanomethyl)amine. In the presence of the activator, which consists of a relatively small dialkyl(cyanomethyl)amine, the condensation proceeded with excellent diastereoselectivity. After sulfurization and deprotection, diastereopure (R_p)- and (S_p)-dinucleoside phosphorothioates were obtained in excellent yields. The present methodology was also applied to the solid-phase synthesis of stereoregulated PS-ODNs. all-(R_p)-[T_PS]₃T, all-(S_p)-[T_PS]₃T, all-(R_p)-d[G_PSA_PSC_PS]T, and all-(Rp)-[T_PS]₉T were synthesized on a highly cross-linked polystyrene resin.

Full text of DOI:10.1021/ja034502z

Published on Web 06/14/2003
An Oxazaphospholidine Approach for the Stereocontrolled
Synthesis of Oligonucleoside Phosphorothioates
Natsuhisa Oka, Takeshi Wada,* and Kazuhiko Saigo
Contribution from the Department of Integrated Biosciences, Graduate School of Frontier
Sciences, The UniVersity of Tokyo, Bioscience Building 702, Kashiwa, Chiba 277-8562, Japan
Received February 5, 2003; E-mail: wada@k.u-tokyo.ac.jp
Abstract: The stereocontrolled synthesis of oligodeoxyribonucleoside phosphorothioates (PS-ODNs) using
nucleoside 3′-O-oxazaphospholidine derivatives as monomer units is described. 2-Chloro-1,3,2-oxaza-
phospholidine derivatives were prepared from six kinds of enantiopure 1,2-amino alcohols and used for
the phosphitylation reactions of 5′-O-protected nucleosides. A detailed study of these reactions revealed
that the diastereoselectivity of the reaction depended on the structure of the enantiopure 1,2-amino alcohol,
the reaction temperature, and the amine used as a scavenger of HCl. In addition, ab initio molecular orbital
calculations for the 2-chlorooxazaphospholidine derivatives were carried out to elucidate the mechanism
of these diastereoselective phosphitylation reactions. The LUMO of the 2-chloro-5-phenyloxazaphospholidine
derivatives on the phosphorus atom was found to be almost orthogonal to the P-Cl bond. This LUMO
may be involved in the phosphitylation reactions with predominant retention of the P-configuration. A series
of dialkyl(cyanomethyl)ammonium salts were developed and used as activators for the condensation
reactions of the diastereopure nucleoside 3′-O-oxazaphospholidines with 3′-O-protected nucleosides. In
the presence of the new activators, the reactions proceeded rapidly to give the corresponding dinucleoside
phosphite triesters. The diastereoselectivity of the condensation reaction did not depend on the counteranion
but on the structure of the dialkyl(cyanomethyl)amine. In the presence of the activator, which consists of
a relatively small dialkyl(cyanomethyl)amine, the condensation proceeded with excellent diastereoselectivity.
After sulfurization and deprotection, diastereopure (Rp)- and (Sp)-dinucleoside phosphorothioates were
obtained in excellent yields. The present methodology was also applied to the solid-phase synthesis of
stereoregulated PS-ODNs. all-(Rp)-[TPS]3T, all-(Sp)-[TPS]3T, all-(Rp)-d[GPSAPSCPS]T, and all-(Rp)-[TPS]9T were
synthesized on a highly cross-linked polystyrene resin.
methods.4 The drawback arises from the lack of an efficient
method to produce sufficient quantities of stereodefined PS-
Introduction
Oligodeoxyribonucleoside phosphorothioates (PS-ODNs) have
been recognized as antisense drugs, and further studies and
clinical applications have also been tried for various diseases.1
Recent in vitro studies have shown that the properties of PS-
ODNs as antisense oligonucleotides, such as binding affinity
to the complementary RNA, stability to nucleases, and RNase
H activity, are affected by the configurations of the phosphorus
atoms and that PS-ODNs with properly arranged Rp and/or Sp
phosphorothioate internucleotidic linkages have much potential
as antisense molecules.2 However, a drawback in the synthesis
of PS-ODNs by conventional methods is the possible formation
of 2n diastereoisomers due to the chirality of the phosphorus
atoms (n is the number of internucleotidic phosphorothioate
linkages);3 almost all PS-ODNs currently used have been
synthesized by nonstereoselective automated phosphoramidite
ODNs. Therefore, it is highly required to develop a method to
produce stereodefined PS-ODNs with productivity comparable
to that of PS-ODNs produced by the conventional phosphor-
amidite method.
Thus, the stereoselective syntheses of PS-ODNs have been
extensively studied in recent years.5,6 Among the studies, the
oxathiaphospholane method,5a which has been developed by Stec
et al., and the method utilizing nucleoside 3′-O-(3-N-acyl)oxaza-
phospholidine derivatives as monomer units, which has been
developed by Beaucage et al.,5b afforded fully stereoregulated
PS-ODNs to date. In these methods, the diastereopure monomer
units are condensed with nucleosides in the presence of strong
(3) Stec, W. J.; Wilk, A. Angew. Chem., Int. Ed. Engl. 1994, 33, 709-722
and references therein.
(4) For a review, see: Beaucage, S. L.; Iyer, R. P. Tetrahedron 1992, 48, 2223-
(1) For a review, see: Levin, A. A. Biochim. Biophys. Acta 1999, 1489(1),
69-84.
2311.
(5) (a) Stec, W. J.; Karwowski, B.; Boczkowska, M.; Guga, P.; Koziolkiewicz,
M.; Sochacki, M.; Wieczorek, M. W.; Blaszczyk, J. J. Am. Chem. Soc.
1998, 120, 7156-7167. (b) Wilk, A.; Grajkowski, A.; Phillips, L. R.;
Beaucage, S. L. J. Am. Chem. Soc. 2000, 122, 2149-2156. (c) Lu, Y.;
Just, G. Angew. Chem., Int. Ed. 2000, 39, 4521-4524.
(2) (a) Koziolkiewicz, M.; Krakowiak, A.; Kwinkowski, M.; Boczkowska, M.;
Stec, W. J. Nucleic Acids Res. 1995, 23, 5000-5005. (b) Koziolkiewicz,
M.; Wojcik, M.; Kobylanska, A.; Karwowski, B.; Rebowska, B.; Guga,
P.; Stec, W. J. Antisense Nucl. Acid Drug DeV. 1997, 7, 43-48. (c) Yu,
D.; Kandimalla, E. R.; Roskey, A.; Zhao, Q.; Chen, L.; Chen, J.; Agrawal,
S. Bioorg. Med. Chem. 2000, 8, 275-284 (d) Inagawa, T.; Nakashima,
H.; Karwowski, B.; Guga, P.; Stec, W. J.; Takeuchi, H.; Takaku, H. FEBS
Lett. 2002, 528, 48-52.
(6) (a) Iyer, R. P.; Yu, D.; Ho, N.-H.; Tan, W.; Agrawal, S. Tetrahedron:
Asymmetry 1995, 6, 1051-1054. (b) Iyer, R. P.; Guo, M.-J.; Yu, D.;
Agrawal, S. Tetrahedron Lett. 1998, 39, 2491-2494. (c) Lu, Y.; Just, G.
Tetrahedron 2001, 57, 1677-1687.
9
10.1021/ja034502z CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 8307-8317
8307

A R T I C L E S
Oka et al.
Table 1. Synthesis of 5′-O-(TBDPS)thymidine
3′-O-Oxazaphospholidines 5a-f
bases, such as 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) and
N,N,N′,N′-tetramethylguanidine (TMG), as promoters with
nearly complete stereospecificity. However, the diastereopure
monomer units have to be prepared through the time-consumed
separation of a mixture, containing essentially an equal amount
of diastereomers, by column chromatography.
The methods utilizing nucleoside 3′-O-oxazaphospholidine
or oxazaphosphorine derivatives derived from enantiopure amino
alcohols, such as (1R,2S)-ephedrine and prolinol, as monomer
units have been reported in recent years.6 Contrary to the
nucleoside 3′-O-(3-N-acyl)oxazaphospholidine derivatives, these
monomer units with an alkyl group on the nitrogen atom of the
oxazaphospholidine or oxazaphosphorine ring can be activated
by weak acids, such as 1H-tetrazole, to condense with nucleo-
sides. The most promising aspect of these methods is that
diastereopure monomers can be obtained from appropriate
enantiopure amino alcohols with excellent or complete dia-
stereoselectivity. Despite this advantage, the condensation
reactions with 3′-O-protected nucleosides are more or less
nonstereospecific. The nonstereospecificity would be attributed
to the repetitive attacks of a nucleophilic activator, such as 1H-
tetrazole, on the phosphorus atom.
Under these circumstances, we developed a new class of
activators, dialkyl(cyanomethyl)ammonium salts 1, and pre-
sented a preliminary result in the previous communication.7 In
this paper, we report the results of detailed studies on the
stereocontrolled synthesis of dideoxyribonucleoside phos-
phorothioates by an oxazaphospholidine method and on its
application to a solid-phase synthesis of PS-ODNs. We also
describe a possible mechanism for the diastereoselective forma-
tion of nucleoside 3′-O-oxazaphospholidine derivatives on the
basis of ab initio molecular orbital calculations.
isolated
yielda,b
(%)
reaction
3
4
5
entry
5
R
R
R
amine
Et3N
Et3N
i-Pr2NEt rt, 30 min
conditions
trans:cis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
a
Me
H
Ph
rt,c 30 min 95:5
77
reflux, 1 h
59:41
88:12
70:30
72:28
87:13
82:18
58:42
92:8
b
c
Me Me Ph
Et3N
Et3N
Et3N
rt, 30 min
rt, 12 h
reflux, 1 h
i-Pr2NEt rt, 30 min
Et3N rt, 30 min
i-Pr2NEt rt, 30 min
i-Pr2NEt reflux, 1 h
64
Me Ph
Ph
Ph
97:3
93:7
96
64
d
e
i-Pr
H
H
Et3N
Et3N
rt, 30 min
reflux, 1 h
rt, 30 min
reflux, 1 h
67:33
85:15
64:38
71:29
20:80
76:24
44:56
Me
i-Pr Et3N
Et3N
i-Pr2NEt rt, 30 min
f
Me i-Pr
H
Et3N
Et3N
rt, 30 min
reflux, 1 h
i-Pr2NEt rt, 30 min
a Diastereopure trans-5a-d were isolated by silica gel column chroma-
tography. b Isolation of trans-5 was not attempted in entries 2-6, 8, 9, and
12-18. c rt ) room temperature.
phenylsilyl)thymidine [5′-O-(TBDPS)thymidine] (4) under vari-
ous reaction conditions to prepare the nucleoside 3′-O-oxaza-
phospholidine derivatives 5a-f (Table 1). Excellent diastereo-
selectivity was achieved when 3a,c were used, although the
suitable reaction conditions were different; in the case of 3a,
the reaction under kinetic conditions (see the Experimental
Section) resulted in excellent diastereoselectivity (trans-5a:cis-
5a ) 95:5, entry 1),8 and in the case of 3c, the reaction under
thermodynamic conditions (see the Experimental Section) gave
the best result (trans-5c:cis-5c ) 97:3, entry 10). Among the
resultant compounds, diastereopure trans-5a-d were very easily
isolated by simple silica gel column chromatography in modest
to excellent yields, and used as the monomer units for a study
on the condensation reactions with 3′-O-protected nucleosides
(entries 1, 7, 10, and 11).
Stereochemistry of the Diastereoselective Phosphitylation
Reactions. The results summarized in Table 1 show that the
diastereoselectivity for the formation of 5 dramatically varied
with the structure of the phosphitylating agent 3, the reaction
temperature, and the amine used as a scavenger of HCl. For
example, cis-5f was mainly obtained under kinetic conditions
(trans-5f:cis-5f ) 20:80, entry 16), while trans-5f was the major
isomer under thermodynamic conditions (trans-5f:cis-5f ) 76:
24, entry 17). This result is in good agreement with our
assumption that the trans-isomer, which may be thermodynami-
cally more stable than the corresponding cis-isomer because of
the steric repulsion between the chlorine atom and the substituent
(R4) on the oxazaphospholidine ring, would be generated
preferentially. In contrast, 5a was obtained with excellent
diastereoselectivity under kinetic conditions (trans-5a:cis-5a )
Results and Discussion
Synthesis of Oxazaphospholidine Monomer Units. In the
present strategy, enantiopure amino alcohols have to work as
chiral auxiliaries for the diastereoselective synthesis of monomer
units and also as protecting groups for internucleotidic linkages.
This means that these amino alcohols must be effective chiral
auxiliaries and be easily removed without the epimerization of
the resultant PS-ODN. In addition, both enantiomers of the
amino alcohols should be available for the formations of both
Rp and Sp internucleotidic phosphorothioate linkages. From these
viewpoints, enantiopure 1,2-amino alcohols are considered to
be suitable candidates, since it has been reported that the
reactions of 5′-O-protected nucleosides with 2-chloro-1,3,2-
oxazaphospholidines derived from some enantiopure 1,2-amino
alcohols afford the trans-nucleoside 3′-O-oxazaphospholidine
monomers diastereoselectively6a,b and that 2-aminoethyl moieties
derived from 1,2-amino alcohols can be removed without
epimerization of the resultant phosphorothioates.5b Thus, we
selected six kinds of enantiopure 1,2-amino alcohols (2a-f),
both enantiomers of which can be easily obtained,8 and carried
out an extensive study on the synthesis of nucleoside 3′-
O-oxazaphospholidine monomer units. 2-Chlorooxazaphos-
pholidine derivatives 3a-f were synthesized from 2a-f and
phosphorus trichloride according to the procedure described in
the literature,6a and allowed to react with 5′-O-(tert-butyldi-
(7) Oka, N.; Wada, T.; Saigo, K. J. Am. Chem. Soc. 2002, 124, 4962-4963.
(8) See the Supporting Information.
9
8308 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Stereocontrolled Synthesis of PS-ODNs
A R T I C L E S
Table 2. Synthesis of 2-Chloro-1,3,2-oxazaphospholidine
Derivatives 3a-f
isolated
yieldb,c
(%)
31P chemical shifts
of the crude 3a−fa
3
4
5
entry
5
R
R
R
1
2
3
4
5
6
a
b
c
d
e
f
Me
Me
Me
i-Pr
Me
Me
H
Me
H
Me
H
i-Pr
Ph
Ph
Ph
Ph
i-Pr
H
172.6 (172.4, 171.4)d
171.4 (171.5)
171.8e
65
52
Figure 1. LUMO of the 2-chloro-1,3,2-oxazaphospholidines (A) trans-3a
and (B) cis-3a calculated at the HF/6-31G* level.
170.7f
173.1 (173.7, 172.5)
176.8 (176.7)
60
70
a In parentheses, the 31P chemical shifts of distilled 3 are given. b 3a,b,e,f
were purified by distillation under reduced pressure. c 3c,d were not purified.
d Small broad signals (ca. 5%) were observed at around 165 ppm. e A small
broad signal (ca. 5%) was observed at 175.4 ppm. f The obviously broad
signal indicated that the product was a mixture of the trans- and cis-isomers.
Table 3. Stereocourse of the Phosphitylations under Kinetic
Conditions
3
stereocourse
5
3a,d,e (mixture)
3b,c (trans-rich)
3f (trans-rich)
5a,d,e (trans-rich)
5b,c (trans-rich)
5f (cis-rich)
retention
inversion
95:5, entry 1), whereas thermodynamic conditions gave 5a with
miserable diastereoselectivity (trans-5a:cis-5a ) 59:41, entry
2). This phenomenon cannot be explained by our assumption.
Thus, we tried to elucidate the mechanism of the present
diastereoselective phosphitylation reaction.
Figure 2. Possible pathways and the activation energies for the reactions
of trans-3a and cis-3a with an alcohol.
At first, the P-configurations of 3a-f were justly determined
to make the stereocourses of the reactions clear. 31P NMR
spectra of crude 3a-f, containing a slight amount of the HCl
salt produced during the synthesis of 3, fundamentally showed
a broad signal in a range of 180-170 ppm (Table 2). In contrast,
in the 31P and 1H NMR spectra of the distilled 3a,e, the signals
of the phosphorus atom and the proton at the 5-position in the
oxazaphospholidine ring were split into two broad singlets (ca.
1:1, entries 1 and 5). These observations indicate that 3a,e are
a ca. 1:1 mixture of the trans- and cis-isomers, and that Cl-
quickly and repetitively attacked the phosphorus atom. However,
when the concentration of Cl- was reduced by distillation, the
rate of the isomerization between the trans- and cis-isomers
decreased to be observed individually by NMR spectroscopies.
Similarly, 3d was assumed to be a mixture of trans- and cis-
isomers because the signal of the 5-H in the oxazaphospholidine
ring of 3d was obviously broad, even though the signal was
not split distinctly into two singlets.8 This explanation is strongly
supported by the fact that the corresponding NMR signals
coalesced to a broad singlet again upon addition of a small
amount of N-methylmorpholinium chloride to the solutions of
distilled 3a,e. On the other hand, 3b,c,f were considered to exist
as a single diastereomer on the basis of their 1H and 31P NMR
spectra.
3. Despite 3a being a mixture of trans- and cis-isomers, excellent
diastereoselectivity was achieved when 3a was used as a
phosphitylating agent as shown in Table 1 (trans-5a:cis-5a )
95:5). To explain this anomalous phenomenon of the diastereo-
enrichment from ca. 1:1 to 95:5, we carried out ab initio MO
calculations for 3a.
The optimized geometries and the LUMOs of trans-3a and
cis-3a were calculated at the HF/6-31G* level (Figure 1). The
calculations surprisingly showed that the LUMO on the
phosphorus atom of 3a is almost orthogonal to the P-Cl bond,
even though such an unoccupied molecular orbital (UMO),
orthogonal to the P-Cl bond, usually has higher energy than
the LUMO. This unique LUMO of 3a is considered to be
stabilized by the interaction of the UMO orthogonal to the P-Cl
bond with the UMO of the phenyl group at the 5-position in
the oxazaphospholidine ring. The resultant LUMO would induce
the nucleophilic attack of an alcohol from the backside of the
P-N or P-O bond. Thus, four reaction pathways via trigonal
bipyramidal transition states would be possible (Figures 2 and
3). The activation energy of each pathway was then calculated
at the B3LYP/6-31G*//HF/6-31G* level, using methanol as a
nucleophile in place of the nucleoside 4 (Figure 2). The
calculations showed that the activation energy of pathway A,
which gives the trans-isomer with retention of configuration
by the nucleophilic attack of methanol from the backside of
the P-N bond, is the smallest (17.08 kcal/mol). Taking into
account the existence of fast equilibrium between trans-3a and
Combining the results of the phosphitylations and the
elucidation of the stereochemistry of the phosphitylating agents
3, the stereocourses of the phosphitylation reactions, carried out
under kinetic conditions, can be summarized as shown in Table
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8309

A R T I C L E S
Oka et al.
Figure 5. Dialkyl(cyanomethyl)ammonium salts 1a-n.
Development of a New Class of Activators. For a new class
of activators it is essential to promote the condensation of
diastereopure nucleoside 3′-O-oxazaphospholidine monomers
with a nucleoside without any loss of the diastereopurity.
However, the loss of diastereopurity would be unavoidable in
the presence of a conventional, nucleophilic activator such as
1H-tetrazole, due to its repetitive attacks on the phosphorus
atom. Thus, we tried to devise a novel type of activator to solve
the problem; an ideal activator should consist of less nucleophilic
components and can activate a phosphoramidite only by the
protonation of the nitrogen atom of the phosphoramidite with a
strong proton-donating ability. Strong acids, however, tend to
protonate the phosphorus atom of a phosphoramidite to de-
activate the P-N bond.9 In addition, in the presence of a strong
acid, some side reactions, such as the cleavage of a 5′-O-DMTr
group used as a protecting group, would occur.
Figure 3. Transition-state structures for the reaction of 3a with methanol
calculated at the HF/6-31G* level.
Then, we focused our attention on ammonium salts as
activators,10 since the acidity and nucleophilicity of the com-
ponents can be precisely tuned by using the combinations of
various amines and acids. From some candidates, dialkyl-
(cyanomethyl)amines 6a-d were selected as amino components,
since their protonated forms have a little stronger acidity than
Figure 4. LUMO of the 2-chloro-1,3,2-oxazaphospholidines (A) 3b and
(B) 3f calculated at the HF/6-31G* level.
cis-3a, which is promoted and accelerated by a HCl salt formed
during the condensation, trans-5a would be produced predomi-
nantly from a mixture of trans-3a and cis-3a. This explanation
is consistent with the experimental result.
+
the conventional activator 1H-tetrazole [pKBH of N-(cyano-
methyl)piperidine (6c), 4.55; pKa of 1H-tetrazole, 4.90].11 As
acidic components, four kinds of acids, which generate a less
nucleophilic conjugate anion, were tested.9,12 All of the new
activators could be easily prepared by simple mixing of an
N-(cyanomethyl)amine with an acid and had good solubility in
CH3CN (>1 M). Compounds 1a,d,e,i (Figure 5) were very
deliquescent and had to be handled in an inert atmosphere. On
the other hand, 1f,j,m (Figure 5) were a little deliquescent and
could be handled in air, and 1b,c,g,h,k,l,n (Figure 5) were not
deliquescent. Using 1a-n, an extensive study on the condensa-
tion of the diastereopure nucleoside 3′-O-oxazaphospholidine
monomers 5 with a nucleoside was carried out.
The new theoretical model can also be applied for the
explanation of the diastereoselectivity of the reaction of 3b with
4. The calculation indicated that the LUMO on the phosphorus
atom of 3b is almost orthogonal to the P-Cl bond in analogy
with 3a (Figure 4A). Since 3b exists completely in a trans form
and since there is no equilibrium between the trans- and cis-
isomers, two reaction pathways, analogous to pathways A and
D in the case of 3a (Figure 2), would be considered. The
activation energies for these pathways were then calculated at
the B3LYP/6-31G*//HF/6-31G* level. The activation energy
for giving trans-5b is found to be smaller by 2.87 kcal/mol than
that for giving cis-5b. Therefore, the pathway giving trans-5b
would be more favorable. However, cis-5b may be formed to
some extent because the difference between the activation
energies in not enough to prevent the reaction through the
pathway analogous to pathway D in Figure 2. As a result, the
condensation of 3b with 4 gave a mixture of trans-5b and cis-
5b in a ratio of 70:30.
The predominant inversion of the P-configuration in the
reaction of 3f with 4 can also be explained on the basis of the
new model. In the case of 3f, there is a large LUMO on the
backside of the P-Cl bond in analogy with the common trivalent
organophosphorus compounds having a leaving group (Figure
4B). Thus, the reaction of 3f with an alcohol would be most
likely to proceed through an SN2 mechanism with inversion of
configuration; this is consistent with the experimental result
(trans-5f:cis-5f ) 20:80).
Effect of the Oxazaphospholidine Ring Structure on the
Diastereoselective Condensation. To investigate the effect of
the oxazaphospholidine ring structure on the diastereoselectivity
of the condensation, diastereopure 5′-O-(TBDPS)thymidine 3′-
O-oxazaphospholidine derivatives 5a-d were allowed to react
(9) (a) Nurminen, E. J.; Mattinen, J. K.; Lo¨nnberg, H. J. Chem. Soc., Perkin
Trans. 2 2001, 11, 2159-2165. (b) Nurminen, E. J.; Mattinen, J. K.;
Lo¨nnberg, H. J. Chem. Soc., Perkin Trans. 2 2000, 11, 2238-2240.
(10) Various ammonium salts have already been employed as activators for the
phosphoramidite method. However, these activators consist of nucleophilic
components, and would cause P-epimerization of the monomer unit. (a)
Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859-1862.
(b) Arnold, L.; Tocik, Z.; Bradkova, E.; Hostomsky, Z.; Paces, V.; Smrt,
J. Collect. Czech. Chem. Commun. 1989, 54, 523-532. (c) Hayakawa, Y.;
Kawai, R.; Hirata, A.; Sugimoto, J.; Kataoka, M.; Sakakura, A.; Hirose,
M.; Noyori, R. J. Am. Chem. Soc. 2001, 123, 8165-8176.
(11) Lange’s Handbook of Chemistry, 15th ed.; McGraw-Hill: New York, 1999.
(12) Nifantyev, E. E.; Gratchev, M. K.; Burmistrov, S. Y.; Antipin, M. Y.;
Struchkov, Y. T. Phosphorus Sulfur 1992, 70, 159-174.
9
8310 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Stereocontrolled Synthesis of PS-ODNs
A R T I C L E S
Table 4. Condensation of 5a-d with 7 in the Presence of 1d
Table 5. Condensation of (Rp)-5a with 7 in the Presence of 1a-n
1
entry
1
R
A-
(S )-8a:(R )-8aa
p p
reaction
3
4
entry
5
R
R
configurationa
time
<5 min >99:1 (140.1)
<5 min 96:4 (140.1, 138.1)
dr of 8b,c
-
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
a
b
c
d
e
CH3
BF4-
PF6
>99:1 (140.1)
>99:1 (140.3)
1
2
3
4
5
(Sp)-5a Me
(Rp)-5a Me
H
H
(2S,5S)
(2R,5R)
TfO-
>99:1 (139.9)
-
(CH2)4
BF4-
96:4 (140.1, 138.1)
>99:1 (140.3)
(Rp)-5b Me Me (2R,4S,5R) <15 min
95:5 (139.8, 137.4)
93:7 (140.4, 138.0)
98:2 (140.9, 138.5)
PF6
(Rp)-5c Me Ph (2R,4S,5R)
(Rp)-5d i-Pr (2R,5R)
5 h
<5 min
f
TfO-
TfO-
Tf2N-
>99:1 (139.8)
H
fb
g
h
i
96:4 (139.9, 138.2)
98:2 (140.4, 138.3)
96:4 (140.0, 138.1)
97:3 (140.2, 138.2)
99:1 (139.9, 138.1)
55:45 (140.0, 138.1)
56:44 (140.2, 138.2)
39:61 (139.8, 138.2)
34:66 (140.3, 138.3)
a The configuration of the oxazaphospholidine ring. b The dr was
determined by 31P NMR. c In parentheses, the 31P chemical shifts of 8 are
given.
-
(CH2)5
BF4-
PF6
j
k
l
m
n
TfO-
-
i-Pr
BF4-
with 3′-O-(tert-butyldimethylsilyl)thymidine [3′-O-(TBDMS)-
thymidine] (7) in the presence of 1d, and the reactions were
monitored by 31P NMR spectroscopy. The results are sum-
marized in Table 4. The rate of condensation strongly depended
on the substituent R4 in the oxazaphospholidine ring. In the cases
of 5a,d (R4 ) H), the condensations were completed within 5
min, whereas the reaction of 5b (R4 ) Me) was completed
within 15 min and the reaction of 5c (R4 ) Ph) required 5 h to
complete. This dependence on the substituent R4 can be
attributed to the bulkiness and the electron-withdrawing char-
acter of the substituent R;4 the large and/or electron-withdrawing
substituent R4 on the oxazaphospholidine ring would retard the
N3-protonation of the oxazaphospholidine derivatives to diminish
the reaction rate. On the contrary, the reaction rate was not
affected by the bulkiness of the substituent R3 at the 3-position
in the oxazaphospholidine ring (entry 5); the reaction of 5d (R3
) i-Pr) was completed within 5 min, which is comparable to
the reaction time of 5a (R3 ) Me). The N3-i-Pr group would
not retard the N3-protonation because of the free rotation of the
PF6
TfO-
Tf2N-
a In parentheses, the 31P chemical shifts of 8a are given. b (Rp)-5a
contaminated with 4 mol % of the corresponding cis-isomer [(2S,5R)-5a]
was used as a monomer unit.
diastereoselectivity; the diastereoselectivity varied in the fol-
lowing order: N-(cyanomethyl)dimethylamine g N-(cyano-
methyl)pyrrolidine g N-(cyanomethyl)piperidine . N-(cyano-
methyl)diisopropylamine. In the presence of 1a-c,e,f, the
condensations proceeded with complete diastereoselectivity to
afford only one diastereoisomer (entries 1-3, 5, and 6). In
contrast, the diastereoselectivities dramatically decreased when
1k-n were used as activators (entries 12-15). From the
viewpoints of easy preparation and handling, sufficient solubility
in CH3CN, and satisfactory diastereoselectivity, 1f was used in
the following studies.
Conversion to (Rp)- and (Sp)-Dithymidine Phosphorothio-
ates. Diastereopure dinucleoside phosphite 8a, obtained by the
condensation of (Rp)-5a (1.2 equiv) with 7 in the presence of
1f, was treated with acetic anhydride and pyridine for the
acetylation of the methylamino function, which was essential
to prevent side reactions, and then sulfurized by the Beaucage
reagent (Scheme 1).13 Although a pair of signals was observed
in the 31P NMR spectra of 9 and 10, the phenomenon would be
attributable to the existence of acetamide rotamers.5b The
acetylated chiral auxiliary could be removed by treatment with
10 equiv of DBU for 30 min at 50 °C to afford 5′-O- and 3′-
O-silylated dithymidine phosphorothioate 11 in excellent yield
without any epimerization.7 Alternatively, the chiral auxiliary
was found to be removed by treatment with 25% NH3-pyridine
for 30 min at 55 °C.5b The diastereomer ratio (dr) of 11 was
determined to be >99:1 by 31P NMR spectroscopy. Finally, the
5′-O- and 3′-O-silyl groups were removed by treatment with
3HF‚Et3N14 to afford fully deprotected dimer 12a. The resultant
dimer was almost diastereopure (dr ) >99.5:0.5), and the
P-configuration of the dimer was found to be Sp by RP-HPLC
N
3-i-Pr bond and/or because of the stronger electron-donating
ability of the isopropyl group than the methyl group.
The diastereoselectivity of the reaction of (Rp)-5a depended
on the substituent R;4 5 having a less bulky substituent as the
R4 showed better diastereoselectivity. As a result, 5a (R4 ) H)
gave the best result. The monomers, derived from (R)- and (S)-
2-methylamino-1-phenylethanol, gave different results; the
reaction of (Sp)-5a with 7 afforded only one diastereomer (entry
1). This means that the combination of (Sp)-5a and 7 is matching
whereas the combination of (Rp)-5a and 7 is mismatching for
the present diastereoselective reaction activated by 1d.
Effect of the Activators on the Diastereoselective Con-
densation. Next, the effect of the activators on the condensation
reaction was determined. Diastereopure (Rp)-5a was allowed
to condense with 7 in the presence of a series of the new
activators 1a-n, and the reactions were monitored by 31P NMR
spectroscopy. The results are summarized in Table 5. All of
the reactions were completed within 5 min. The diastereo-
selectivity of the reaction was little dependent on the counter-
anion of the activators, probably because of their lower
nucleophilicity. On the other hand, the structure of the N-
(cyanomethyl)amines, which play an important role in the
protonation of the N3 of the monomer, was found to affect the
(13) Iyer, R. P.; Egan, W.; Regan, J. B.; Beaucage, S. L. J. Am. Chem. Soc.
1990, 112, 1253-1254.
(14) 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-
5166.
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8311

A R T I C L E S
Oka et al.
Scheme 1. Synthesis of (Rp)- and (Sp)-Dithymidine
Phosphorothioates (Rp)-12a and (Sp)-12a
Table 6. Solid-Phase Synthesis of Dinucleoside
Phosphorothioates 12a-da
concn
of 13
(M)
reaction
time
(min)
solid
support
1
entry
13
B
(R )-12:(S )-12
p
p
1
2
3
CPG
HCP
HCP
HCP
HCP
HCP
HCP
HCP
HCP
HCP
a
a
a
a
a
b
c
d
a
a
Th
0.1
0.1b
0.2c
0.2
0.2
0.2
0.2
0.2
0.2
0.2
5
5
5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
64:36
97:3
98:2
99:1
2:98
97:3
99:1
98:2
99:1
96:4
Th
Th
Th
Th
4
5d
6
Adbz
Cybz
Guce,ibu
Th
7e
8
9f
10g
Th
a Reagents and conditions: (i) 1f, CH3CN, rt; (ii) Ac2O, N-methylimid-
azole, THF, rt, 30 s; (iii) Beaucage reagent, CH3CN, rt, 60 s; (iv) 3% TCA,
CH2Cl2, rt; (v) 25% NH3(aq)-pyridine (9:1, v/v), 55 °C. b A 10 µmol
sample of (Sp)-13a and 50 µmol of 1f were used for the condensation with
14 (0.5 µmol). c A 20 µmol sample of (Sp)-13a and 50 µmol of 1f were
used for the condensation with 14 (0.5 µmol). d (Rp)-13a (trans:cis ) 99:
1) was used as a monomer unit. e A 1.0 M solution of 1f in CH3CN was
used for the condensation since in the case of using 0.5 M 1f the coupling
efficiency was relatively low (ca. 90%). f After the condensation step, the
resultant phosphite triester intermediate was treated with a 0.5 M solution
of 1f in CH3CN for 5 min. g (Sp)-13a was treated with a 0.5 M solution of
1f in CH3CN for 5 min prior to being added to 14.
analysis.15 Diastereopure (Rp)-12a could also be obtained from
(Sp)-5a according to the same procedure. The exclusive produc-
tion of (Rp)-12a and (Sp)-12a from (Sp)-5a and (Rp)-5a,
respectively, in the presence of 1f strongly indicates that the
condensation of 5a with 7 in the presence of 1f proceeds with
complete inversion of configuration at the phosphorus atom,
upon assuming that the sulfurization by the Beaucage reagent
and the removal of the chiral auxiliary by DBU treatment
proceed with retention of configuration.
Solid-Phase Synthesis. The results of the solution-phase
synthesis indicate that the present method would be sufficiently
applicable to the solid-phase synthesis of PS-ODNs since the
condensation proceeded rapidly with excellent diastereoselec-
tivity. Then, we tried the solid-phase synthesis of PS-ODNs by
the present method. Diastereopure 5′-O-(DMTr)nucleoside 3′-
O-oxazaphospholidine monomers 13a-d could be synthesized
in 62-75% isolated yields in a manner similar to the procedure
for the preparation of 5; the phosphitylation reactions giving
13a-d were compeleted within 30 min with good diastereo-
selectivity in a range of 93:7 to 96:4.7,8,16 At first, dithymidine
phosphorothioate was synthesized on the conventional solid
supports, a controlled pore glass (CPG) and a highly cross-linked
polystyrene (HCP),17 using (Sp)-13a and 1f (Table 6). The
condensation on the HCP proceeded with excellent diastereo-
selectivity, whereas the diastereoselectivity of the condensation
on the CPG was extremely low (Table 6, entries 1 and 2).18
Thus, the HCP was selected as a solid support. However, the
diastereoselectivity must be further improved for the efficient
synthesis of PS-ODNs. We then considered that the loss of the
diastereopurity during the synthesis would be attributed to the
epimerization of the monomer unit (Sp)-13a by acidic 1f prior
to the condensation, because a large excess amount of 1f was
used in the solid-phase synthesis. On the basis of this consid-
eration, we carried out the condensation upon increasing the
concentration of (Sp)-13a to accelerate the condensation and
shortening the condensation time to prevent the epimerization.
In fact, the diastereoselectivity was improved upon these
modifications (entries 3 and 4). Under the optimized conditions
shown in entry 4, the dinucleoside phosphorothioates could be
synthesized with excellent diastereoselectivity (98:2 to 99:1)
and average coupling yields (98 to >99%, entries 4-8).
Our assumption that the epimerization of the monomer unit
(Sp)-13a is the main factor for the deterioration of the dia-
stereopurity was also confirmed by the following reactions.
When the dithymidine phosphite intermediate, obtained by the
condensation of (Sp)-13a with 14, was treated with a 0.5 M
solution of 1f in CH3CN for 5 min, no deterioration of the
diastereoselectivity of 12a was observed (entry 9). On the
contrary, when (Sp)-13a was preactivated with a 0.5 M solution
of 1f in CH3CN for 5 min and then allowed to condense with
14, the reaction afforded 12a with a little lower diastereo-
selectivity (96:4, entry 10). These results strongly indicate that
(18) The relatively low coupling efficiency and the extremely low diastereo-
selectivity in the case of using the CPG may be attributed to the hydrophilic
nature of the CPG due to the silanol groups on the surface. The hydrophilic
CPG tends to retain water which hydrolyzes the monomer unit to reduce
the concentration of the monomer unit during the coupling step.17
Nucleophilic silanols also reduce the concentration of the monomer unit.17
A sufficiently high concentration of the monomer unit is necessary for an
excellent diastereoselectivity as well as for an excellent coupling yield as
described below.
(15) (Rp)-TPST eluted faster than (Sp)-TPST in RP-HPLC. Wada, T.; Kobayashi,
N.; Mori, T.; Sekine, M. Nucleosides Nucleotides 1998, 17 (1-3), 351-
364.
(16) B1 ) Th (13a); B1 ) Adbz (13b); B1 ) Cybz (13c); B1 ) Guce,ibu (13d).
(17) McCollum, C.; Andrus, A. Tetradedron Lett. 1991, 32, 4069-4072.
9
8312 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Stereocontrolled Synthesis of PS-ODNs
A R T I C L E S
[TPS]9T (18) were synthesized manually on an HCP resin. The
average coupling yields were estimated to be 98% in the cases
of tetramers and 97% in the case of a decamer by DMTr+ assay.
RP-HPLC profiles of the crude 15-18 and purified 18 are
shown in Figures 6 and 7. The yields and the average
diastereoselectivities of 15-17 were determined on the basis
of the RP-HPLC profiles to be 80% (15), 78% (16), and 75%
(17) (yields) and >97% (15), >95% (16), and >94% (17)
(selectivities) (Figure 6). 18 was isolated by RP-HPLC in 35%
yield (Figure 7B), and its diastereopurity was estimated to be
>96% on the basis of the main peak area of the RP-HPLC
chromatogram of the products produced by the incubation of
the purified 18 with nuclease P1 for 1 h.19
15, 17, and purified 18 were completely digested by incuba-
tion with snake venom phosphodiesterase for 1 h at 37 °C,20
and crude 16 was completely digested by incubation with
nuclease P1 for 1 h at 37 °C. These results indicate that the
present method gave fully P-stereoregulated PS-ODNs as the
main products.
Conclusion
We have developed a new efficient method for the highly
diastereoselective formation of internucleotidic phosphorothioate
linkages. The present new oxazaphospholidine method using a
new activator (1f) enabled us to produce stereoregulated PS-
ODNs. This method is expected to be applied to an automated
solid-phase synthesis directly, since the steps of this method
are identical to those used in the conventional automated
phosphoramidite method. This is a great advantage from a
viewpoint of the large-scale production of stereoregulated PS-
ODNs. In addition, this method has possibilities to produce other
stereoregulated backbone-modified DNA analogues. This strat-
egy is also applicable to the synthesis of chimeric PO/PS-ODNs
containing stereodefined phosphorothioate linkages upon switch-
ing the oxidation/sulfurization steps.
Additionally, new mechanistic aspects of diastereoselective
formation of 2-alkoxyoxazaphospholidine derivatives could be
proposed on the basis of ab initio MO calculations. The
information would facilitate the stereoselective synthesis of a
wide variety of chiral phosphorus compounds, which have been
reported to be useful for many purposes21 as well as the
precursors of the DNA analogues.5b,6,7
Experimental Section
General Information. 1H NMR spectra were obtained at 300 MHz
on a Varian MERCURY 300 spectrometer with tetramethylsilane (TMS)
as an internal standard in CDCl3, with TMS as an external standard in
CD3CN, and with 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt
(DSS) as an external standard in D2O. 13C NMR spectra were obtained
at 75.45 MHz on a Varian MERCURY 300 spectrometer with CDCl3
as an internal standard (δ 77.0) in CDCl3, with TMS as an external
Figure 6. Reversed-phase HPLC profiles of the crude tetramers: (A) 15;
(B) 16; (C) 17. RP-HPLC was performed with a linear gradient of 0-20%
acetonitrile in 0.1 M ammonium acetate buffer (pH 7.0) at 50 °C for 60
min at a rate of 0.5 mL/min.
standard in CD3CN, and with DSS as an external standard in D2O. 31
P
NMR spectra were obtained at 121.5 MHz on a Varian MERCURY
300 spectrometer using 85% H3PO4 as an external standard. IR spectra
were recorded on a JASCO FT/IR-480 Plus spectrophotometer. UV/
vis spectra were recorded on a JASCO V-550 UV/vis spectrophotom-
eter. Melting points were measured using a YAMATO MP-21 and are
(Sp)-13a is partially epimerized by acidic 1f prior to the
condensation with 14; the condensation should be carried out
in as a short time as possible with a high concentration of (Sp)-
13a for achieving excellent diastereoselectivity.
Under the optimized conditions, all-(Rp)-[TPS]3T (15), all-
(Sp)-[TPS]3T (16), all-(Rp)-d[CPSAPSGPS]T (17), and all-(Rp)-
(19) Potter, B. V. L.; Connolly, B. A.; Eckstein, F. Biochemistry 1983, 22, 1369-
1377.
(20) Bryant, F. R.; Benkovic, S. J. Biochemistry 1979, 18, 2825-2828.
(21) For a review, see: Kolodiazhnyi, O. I. Tetrahedron: Asymmetry 1998, 9,
1279-1332.
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8313

A R T I C L E S
Oka et al.
Figure 7. Reversed-phase HPLC profiles: (A) crude 18; (B) 18 purified by reversed-phase HPLC. RP-HPLC was performed with a linear gradient of
0-20% acetonitrile in 0.1 M ammonium acetate buffer (pH 7.0) at 50 °C for 60 min at a rate of 0.5 mL/min.
1
uncorrected. MALDI-TOF-MS spectrometry was carried out on an
Applied Biosystems Voyager-DE STR spectrometer. Thin-layer chro-
matography (TLC) was performed on TLC plates coated with silica
gel 60 F254 (Merck, No. 5715). Silica gel column chromatography was
carried out using silica gel 60N (63-210 µm). Reversed-phase HPLC
was carried out using a µBondasphere 5 µm C18 column (100 Å, 3.9
mm × 150 mm) (Waters). Organic solvents were purified and dried
by the appropriate procedure. Aminomethylated highly cross-linked
polystyrene was purchased from Perkin-Elmer Applied Biosystems.17
Manual solid-phase synthesis was performed by using a glass filter
(10 mm × 50 mm) with a stopper at the top and a stopcock at the
bottom as a reaction vessel. Snake venom phoshodiesterase (svPDE)
was purchased from Sigma, and nuclease P1 was purchased from
Yamasa. Stereorandom PS-ODNs were purchased from Amersham
Biosciences.
121.5 (q, JCF ) 320 Hz), 112.6, 56.2, 42.3, 23.9. Anal. Calcd for
C7H11F3N2O3S: C, 32.31; H, 4.26; N, 10.76. Found: C, 32.31; H, 4.22;
N, 10.81.
1a-e,g-n were synthesized in a similar manner. The experimental
details and characterization data for 1a-e,g-n are given in the
Supporting Information.
Preparation of (5R)-2-Chloro-3-methyl-5-phenyl-1,3,2-oxaza-
phospholidine [(5R)-3a]. A Typical Procedure for the Synthesis of
3a-f. (R)-2a (756 mg, 5.0 mmol) was dried by repeated coevaporations
with dry toluene and dissolved in dry toluene (2.5 mL). N-Methyl-
morpholine (1.10 mL, 10 mmol) was added to the solution, and the
mixture was added dropwise to a stirred solution of phosphorus
trichloride (436 µL, 5.0 mmol) in dry toluene (2.5 mL) at 0 °C under
argon. The mixture was allowed to warm to rt and stirred for 30 min.
The resultant salt was removed by filtration at -78 °C under argon,
and the filtrate was concentrated under reduced pressure to afford crude
(5R)-3a (1.07 g, 5.0 mmol) as a pale yellow oil. This material was
used without further purification for the phosphitylation of nucleosides.
13C NMR (75 MHz, CDCl3): δ 138.1, 128.5, 128.5, 126.5, 85.4 (br),
Ab Initio Calculations. Ab initio molecular orbital calculations were
carried out using the Gaussian 9822 and Spartan’0223,24 programs on a
Silicon Graphics Inc. OCTANE workstation. Geometry optimizations
were carried out at the HF/6-31G* level. In some cases, single-point
energy calculations were carried out at the B3LYP/6-31G* level.
N-(Cyanomethyl)pyrrolidinium Trifluoromethanesulfonate (1f).
Trifluoromethanesulfonic acid (1.77 mL, 20 mmol) was added dropwise
to a stirred solution of N-(cyanomethyl)pyrrolidine (2.20 g, 20 mmol)
in dry CH2Cl2 (20 mL) at 0 °C. The mixture was then allowed to warm
to rt, and diluted with dry Et2O (40 mL). The resultant precipitate was
collected by filtration, washed with dry Et2O (3 × 5 mL), and dried
under vacuum to afford 1f (5.20 g, 20 mmol) as a white crystalline
solid. Mp: 67.0-67.5 °C. IR (KBr): νmax 2996, 2841, 2651, 2477,
2347, 2282, 1637, 1462, 1437, 1269, 1228, 1168, 1033, 985, 911, 849,
2
2
56.3 (d, JPC ) 6.6 Hz), 31.0 (d, JPC ) 13.0 Hz). Crude (5S)-3a and
3b-f were obtained according to the same procedure. Further purifica-
tion conditions and the characterization data for (5S)-3a and 3b-f are
given below. Crude (5R)-3a could be distilled under reduced pressure
to afford (5R)-3a (2.10 g, 9.7 mmol) as a colorless liquid (bp 81-82
°C, 0.2 mmHg) from 2.27 g (15 mmol) of (R)-2a as a starting material.
The H and 31P NMR spectra of crude and distilled (5R)-3a are given
in the Supporting Information.
1
(5S)-2-Chloro-3-methyl-5-phenyl-1,3,2-oxazaphospholidine [(5S)-
3a]. Crude (5S)-3a was also obtained quantitatively from (S)-2a and
could be used for the phosphitylation of nucleosides without further
purification. Crude (5S)-3a could also be distilled under reduced
pressure to afford (5S)-3a (2.22 g, 12 mmol) as a colorless liquid, using
3.02 g (20 mmol) of (S)-2a as a starting material. The physical data
were identical to those of (5R)-3a.
1
761, 641 cm-1. H NMR (300 MHz, CD3CN): δ 8.14 (br, 1H), 4.28
(s, 2H), 3.48 (br, 4H), 2.09 (m, 4H). 13C NMR (75 MHz, CD3CN): δ
(22) Gaussian 98, revision A.11.2: Frisch, M. J.; Trucks, G. W.; Schlegel, H.
B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.;
Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.;
Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.;
Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.;
Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui,
Q.; Morokuma, K.; Rega, N.; Salvador, P.; Dannenberg, J. J.; Malick, D.
K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A., Gaussian, Inc., Pittsburgh, PA,
2001.
(23) Spartan’02, version 115a: Deppmeier, B. J.; Driessen, A. J.; Hehre, T. S.;
Hehre, W. J.; Johnson, J. A.; Klunzinger, P. E.; Leonard, J. M.; Pham, I.
N.; Pietro, W. J.; Yu, J., Wavefunction, Inc., Irvine, CA, 2002.
(24) Kong, J.; White, CA.; Krylov, A. I.; Sherrill, C. D.; Adamson, R. D.; Furlani,
T. R.; Lee, MS.; Lee, A. M.; Gwaltney, S. R.; Adams, T. R.; Ochsenfeld,
C.; Gilbert, A. T. B.; Kedziora, G. S.; Rassolov, V. A.; Maurice, D. R.;
Nair, N.; Shao, Y.; Besley, N. A.; Maslen, P. E.; Dombroski, J. P.; Dachsel,
H.; Zhang, W. M.; Korambath, P. P.; Baker, J.; Byrd, E. F. C.; Van Voorhis,
T.; Oumi, M.; Hirata, S.; Hsu, C. P.; Ishikawa, N.; Florian, J.; Warshel,
A.; Johnson, B. G.; Gill, P. M. W.; Head-Gordon, M.; Pople, J. A. J.
Comput. Chem. 2000, 21, 1532.
(4S,5R)-2-Chloro-3,4-dimethyl-5-phenyl-1,3,2-oxazaphospholi-
dine (3b).6a 3b (356 mg, 1.6 mmol) was obtained as a colorless liquid
from the crude product by Kugelrhor bulb-to-bulb distillation (170 °C,
0.4 mmHg), using (1R,2S)-ephedrine (2b) (496 mg, 3.0 mmol) as a
starting material. 1H NMR (300 MHz, CDCl3): δ 7.40-7.23 (m, 5H),
5.85 (d, 3JHH ) 6.3 Hz, 1H), 3.72-3.61 (m, 1H), 2.72 (d, 3JPH ) 15.6
Hz, 3H), 0.72 (d, 3JHH ) 4.2 Hz, 3H). 13C NMR (75 MHz, CDCl3): δ
136.4 (d, 3JPC ) 2.6 Hz), 128.3, 128.3, 126.9, 87.7 (d, 2JPC ) 8.9 Hz),
57.7 (br), 29.0 (d, JPC ) 13.8 Hz), 14.4. 31P NMR (121 MHz,
CDCl3): δ 171.5.
2
(4S,5R)-2-Chloro-3-methyl-4,5-diphenyl-1,3,2-oxazaphospholi-
dine (3c). Crude 3c (3.17 g, 10 mmol) was obtained as a pale yellow
solid from (1R,2S)-2-methylamino-1,2-diphenylethanol (2c) (2.27 g,
10 mmol). 1H NMR (300 MHz, CDCl3): δ 7.08-7.05 (m, 6H), 6.91-
3
3
6.81 (m, 4H), 6.15 (d, JHH ) 8.3 Hz, 1H), 4.64 (dd, JHH ) 8.3 Hz,
9
8314 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Stereocontrolled Synthesis of PS-ODNs
A R T I C L E S
3JPH ) 4.2 Hz, 1H), 2.64 (d, 3JPH ) 15.3 Hz, 3H). 13C NMR (75 MHz,
5′-O-(tert-Butyldiphenylsilyl)-3′-O-[(2R,4S,5R)-3-methyl-4,5-di-
phenyl-1,3,2-oxazaphospholidin-2-yl]thymidine (5c). 4 (961 mg, 2.0
mmol) was dried by repeated coevaporations with dry pyridine and
dry toluene, and then dissolved in dry THF (10 mL). i-Pr2NEt (1.70
mL, 10 mmol) was added, and the mixture was cooled to -78 °C. A
0.22 M solution of 3c in dry THF (10 mL, 2.2 mmol) was added
dropwise via a syringe, and then the mixture was allowed to warm to
rt, and then refluxed for 1 h. The additional heating did not affect the
diastereomer ratio of the crude product. The mixture was then cooled
to rt, saturated NaHCO3 aqueous solution (75 mL) and CHCl3 (75 mL)
were added to the mixture, and the organic layer was separated and
washed with saturated NaHCO3 aqueous solution (2 × 75 mL). The
combined aqueous layers were back-extracted with CHCl3 (2 × 75 mL).
The combined organic layers were then dried over Na2SO4, filtered,
and concentrated to dryness to give the crude product. 31P NMR (121
MHz, CDCl3): δ 150.6 (3%, cis-5c), 144.2 (97%, trans-5c). Purification
of the crude product by silica gel column chromatography (2 × 40
cm, 50 g of silica gel, hexanes-ethyl acetate-triethylamine, 67:33:2
and then 0:100:2, v/v/v) afforded 5c (1.41 g, 1.9 mmol) as a colorless
foam. 31P NMR (121 MHz, CDCl3): δ 142.3. FAB-HRMS: m/z calcd
for C41H47N3O6PSi+ (M + H+) 736.2972, found 736.2975.
Monitoring the Condensation of (Rp)-5a with 7 in the Presence
of 1d by 31P NMR Spectroscopy. A Typical Procedure for Monitor-
ing the Condensation of 5a-d with 7 in the Presence of 1d, and
(Rp)-5a with 7 in the Presence of 1a-n. (Rp)-5a (33.0 mg, 50 µmol)
and 7 (17.8 mg, 50 µmol) were dried over P2O5 under high vacuum
for 12 h in an NMR sample tube. A 0.25 M solution of 1d (400 µL,
100 µmol) in CH3CN, dried over MS 3A for 8 h, and dry CD3CN (100
µL) were added to the tube under argon. After 3 min, the data
accumulation was started. The diastereomer ratio was estimated on the
basis of the integration of the resonance signals.
3
3
CDCl3): δ 135.8 (d, JPC ) 3.2 Hz), 134.5 (d, JPC ) 5.4 Hz), 128.2,
2
127.9, 127.7, 127.6, 127.5, 126.5, 88.0 (d, JPC ) 8.6 Hz), 68.1 (d,
2JPC ) 6.3 Hz), 29.6 (d, 2JPC ) 14.1 Hz). 31P NMR (121 MHz, CDCl3):
δ 171.7.
(5R)-2-Chloro-3-isopropyl-5-phenyl-1,3,2-oxazaphospholidine (3d).
Crude 3d (1.22 g, 5.0 mmol) was obtained as a pale yellow liquid
from (R)-2-isopropylamino-1-phenylethanol (2d) (896 mg, 5.0 mmol).
1H NMR (300 MHz, CDCl3): δ 7.45-7.34 (m, 5H), 5.94-5.24 (br,
2
1H), 3.62-3.47 (m, 1H), 3.18 (dd, JHH ) 18.3 Hz,3JHH ) 9.3 Hz,
3
1H), 1.35 (d, JHH ) 6.6 Hz, 6H). 13C NMR (75 MHz, CDCl3): δ
2
138.5 (br), 128.6, 128.5, 126.5 (br), 84.7 (br), 51.3 (br), 47.3 (d, JPC
) 10.6 Hz), 22.0 (d, 3JPC ) 8.9 Hz), 21.8 (d, 3JPC ) 9.5 Hz). 31P NMR
(121 MHz, CDCl3): δ 170.6 (br).
(5R)-2-Chloro-3-methyl-5-isopropyl-1,3,2-oxazaphospholidine (3e).
3e (1.09 g, 6.0 mmol) was obtained as a colorless liquid from the crude
product by bulb-to-bulb distillation (80 °C, 0.3 mmHg), using (R)-1-
methylamino-3-methyl-2-butanol (2e) (1.17 g, 10 mmol) as a starting
material. 1H NMR (300 MHz, CDCl3): δ 4.68, 4.18 (br, br, 1H), 3.16,
2.95 (m, m, 2H), 2.72 (d, 3JPH ) 15.0 Hz, 3H), 2.01, 1.90 (br, br, 1H),
1.06-0.99 (m, 3H), 0.94 (d, 3JHH ) 6.6 Hz, 3H). 13C NMR (75 MHz,
CDCl3): δ 90.7, 88.7 (br, br), 53.0, 51.8 (br, br), 32.9 (br), 31.1 (d,
2JPC ) 14.3 Hz), 18.3, 18.2 (br, br). 31P NMR (121 MHz, CDCl3): δ
173.7, 172.5 (br, br).
(4S)-2-Chloro-3-methyl-4-isopropyl-1,3,2-oxazaphospholidine (3f).
3f (1.27 g, 7.0 mmol) was obtained as a colorless liquid from the crude
product by bulb-to-bulb distillation (80 °C, 0.4 mmHg), using (S)-2-
methylamino-3-methyl-1-butanol (2f) (1.17 g, 10 mmol) as a starting
material. 1H NMR (300 MHz, CDCl3): δ 4.57 (d, 2JHH ) 8.7 Hz, 3JHH
3
) 8.7 Hz, 1H), 4.21 (m, 1H), 3.25 (m, 1H), 2.68 (d, JPH ) 16.2 Hz,
3
3
3H), 2.03 (m, 1H), 0.92 (d, JHH ) 7.2 Hz, 3H), 0.84 (d, JHH ) 6.6
Hz, 3H). 13C NMR (75 MHz, CDCl3): δ 71.5 (d, 2JPC ) 9.2 Hz), 63.5
(Sp)-1,8-Diazabicyclo[5.4.0]undec-7-enium 5′-O-(tert-Butyldi-
phenylsilyl)thymidin-3′-yl 3′-O-(tert-Butyldimethylsilyl)thymidin-5′-
yl Phosphorothioate [(Sp)-11]. A Typical Procedure for the Synthesis
of (Rp)-11 and (Sp)-11. (Rp)-5a (39.6 mg, 60 µmol) and 7 (17.8 mg,
50 µmol) were dried over P2O5 for 12 h in an NMR sample tube, and
a 0.2 M solution of 1f in CH3CN (500 µL, 100 µmol), dried over MS
3A for 8 h, was added at rt under argon. After 5 min, dry pyridine
(43.0 µL, 500 µmol) and Ac2O (14.9 µL, 100 µmol) were added. After
an additional 30 s, 3H-1,2-benzodithiol-3-one 1,1-dioxide (12.0 mg,
60 µmol) was added. After an additional 3 min, 1,8-diazabicyclo[5.4.0]-
undec-7-ene (74.6 µL, 500 µmol) was added, and the reaction mixture
was heated for 30 min at 50 °C. The mixture was then allowed to cool
to rt, diluted with CHCl3 (3 mL), and washed with 0.2 M phosphate
buffer (pH 7.0, 3 mL). The aqueous layers were back-extracted with
CHCl3 (2 × 3 mL). The combined organic layers were dried over Na2-
SO4, filtered, and concentrated to dryness under reduced pressure. The
residue was purified by PTLC [CH2Cl2-MeOH-Et3N (99:1:0.5, v/v/
v, × 2, 96:4:0.5, v/v/v, × 2)]. The product was dissolved in CHCl3 (3
mL), washed with 0.2 M 1,8-diazabicyclo[5.4.0]undec-7-enium bicar-
bonate buffer (3 mL), and back-extracted with CHCl3 (2 × 3 mL).
The combined organic layers were dried over Na2SO4, filtered, and
concentrated to dryness to afford (Sp)-11 (50.4 mg, 47 µmol) as a
colorless foam. 1H and 31P NMR spectra were identical to those of the
authentic sample synthesized by the conventional H-phosphonate
(br), 29.1 (d, 2JPC ) 14.1 Hz), 26.7 (d, 3JPC ) 4.9 Hz), 19.3, 14.5. 31
NMR (121 MHz, CDCl3): δ 176.8.
P
Preparation of 5′-O-(tert-Butyldiphenylsilyl)-3′-O-[(2R,5R)-3-
methyl-5-phenyl-1,3,2-oxazaphospholidin-2-yl]thymidine [(Rp)-5a].
A Typical Procedure for the Synthesis of 5a,b,d-f and 13a-d. 4
(721 mg, 1.5 mmol) was dried by repeated coevaporations with dry
pyridine and dry toluene, and then dissolved in dry THF (7.5 mL).
Et3N (1.05 mL, 7.5 mmol) was added, and the mixture was cooled to
-78 °C. A 0.22 M solution of (5R)-3a in dry THF (7.50 mL, 1.65
mmol) was added dropwise via a syringe, and then the mixture was
allowed to warm to rt and stirred for 30 min under argon. Saturated
NaHCO3 aqueous solution (75 mL) and CHCl3 (75 mL) were added to
the mixture, and the organic layer was separated and washed with
saturated NaHCO3 aqueous solution (2 × 75 mL). The combined
aqueous layers were back-extracted with CHCl3 (2 × 75 mL). The
combined organic layers were then dried over Na2SO4, filtered, and
concentrated to dryness to give the crude product. 31P NMR (121 MHz,
CDCl3): δ 144.1 (95%, trans-5a), 143.2 (5%, cis-5a). The diastereomer
ratio was essentially the same as that obtained by the reaction performed
for 5 min at -78 °C. The crude product was applied to a column of
silica gel (2.5 × 14 cm, 40 g of silica gel). The product was eluted
with hexanes-ethyl acetate-triethylamine (50:50:2, v/v/v). The frac-
tions containing (Rp)-5a were combined, washed with saturated NaHCO3
aqueous solution (100 mL), dried over Na2SO4, filtered, and concen-
trated to dryness to afford (Rp)-5a (765 mg, 1.2 mmol) as a colorless
foam. 31P NMR (121 MHz, CDCl3): δ 144.6. FAB-HRMS: m/z calcd
for C35H43N3O6PSi+ (M + H+) 660.2659, found 660.2658. The Rf value
[hexanes-ethyl acetate (50:50, v/v)] of (Rp)-5a on the silica gel TLC
plate was 0.51. The minor cis-isomer was eluted a little faster than
(Rp)-5a. Since the amount of the cis-isomer was very small, it could
be easily removed by silica gel column chromatography, though the Rf
values were close to each other. The purification conditions and
characterization data for 5a,b,d-f and 13a-d are given in the
Supporting Information.
1
method.25 The H and 31P NMR spectra of (Sp)-11 are given in the
Supporting Information.
(Rp)-1,8-Diazabicyclo[5.4.0]undec-7-enium 5′-O-(tert-Butyldi-
phenylsilyl)thymidin-3′-yl 3′-O-(tert-Butyldimethylsilyl)thymidin-5′-
yl Phosphorothioate [(Rp)-11]. (Rp)-11 (52.1 mg, 49 µmol) was
obtained as a colorless foam, using (Sp)-5a (39.6 mg, 60 µmol) as a
starting material. The 1H and 31P NMR spectra were identical to those
of the authentic sample synthesized by the conventional H-phosphonate
(25) Kers, A.; Kers, I.; Kraszewski, A.; Sobkowski, M.; Szabo´, T.; Thelin, M.;
Zain, R.; Stawin´ski, J. Nucleosides Nucleotides 1996, 15 (1-3), 361-378.
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8315

A R T I C L E S
Oka et al.
1
method.25 The H and 31P NMR spectra of (Rp)-11 are given in the
all-(Rp)-[TPS]3T (15). According to the typical procedure described
above, 5′-O-(DMTr)thymidine 3′-O-succinate bound to HCP (18.5 mg,
0.5 µmol) gave 15 as a crude product. The average yield per cycle was
estimated to be 98% by DMTr+ assay. The generation of [TPS]3T was
confirmed by MALDI-TOF-MS analysis [m/z 1201 (M - H+)-]. RP-
HPLC analysis showed that the retention time of the main product was
identical to that of the fastest eluted diastereomer of random-[TPS]3T,
which indicated that the main product was 15.8,26 The P-configurations
of 15 were confirmed by svPDE digestion; the main product was
completely digested with svPDE (the experimental procedure is
described below).8 The yield of 15 determined by RP-HPLC was 80%.
The average diastereoselectivity was estimated on the basis of the RP-
HPLC profiles to be >97%.
all-(Sp)-[TPS]3T (16). According to the typical procedure described
above, 5′-O-(DMTr)thymidine 3′-O-succinate bound to HCP (18.5 mg,
0.5 µmol) gave 16 as a crude product. The average yield per cycle was
estimated to be 98% by DMTr+ assay. The generation of [TPS]3T was
confirmed by MALDI-TOF-MS analysis [m/z 1201 (M - H+)-]. RP-
HPLC analysis showed that the retention time of the main product was
identical to that of the slowest eluted diastereomer of random-[TPS]3T,
which indicated that the main product was 16.8,26 The P-configurations
of 16 were confirmed by nuclease P1 digestion; the main product was
completely digested with nuclease P1 (the experimental procedure is
described below).8 The yield of 16 determined by RP-HPLC was 78%.
The average diastereoselectivity was estimated on the basis of the RP-
HPLC profiles to be >95%.
all-(Rp)-d[CPSAPSGPS]T (17). According to the typical procedure
described above, 5′-O-(DMTr)thymidine 3′-O-succinate bound to HCP
(18.5 mg, 0.5 µmol) gave 17 as a crude product. The average yield per
cycle was estimated to be 98% by DMTr+ assay. The generation of
d[CPSAPSGPS]T was confirmed by MALDI-TOF-MS analysis [m/z 1220
(M - H+)-]. RP-HPLC analysis showed that the retention time of the
main product was identical to that of the fastest eluted diastereomer of
random-d[CPSAPSGPS]T, which indicated that the main product was
17.8,26 The P-configurations of 17 were confirmed by svPDE digestion;
the main product was completely digested with svPDE (the experimental
procedure is described below).8 The yield of 17 determined by RP-
HPLC was 75%. The average diastereoselectivity was estimated on
the basis of the RP-HPLC profiles to be >94%.
Supporting Information.
1,8-Diazabicyclo[5.4.0]undec-7-enium 5′-O-(tert-Butyldiphenyl-
silyl)thymidin-3′-yl 3′-O-(tert-Butyldimethylsilyl)thymidin-5′-yl Phos-
phorothioate (11). 11 was synthesized according to the conventional
H-phosphonate method25 as a mixture of diastereomers [(Rp)-11:(Sp)-
11 ) 46:54]. The 1H and 31P NMR spectra of 11 are given in the
Supporting Information.
(Sp)-Ammonium Thymidin-3′-yl Thymidin-5′-yl Phosphoro-
thioate [(Sp)-12a]. A Typical Procedure for the Synthesis of (Rp)-
12a and (Sp)-12a. (Sp)-11 (50.4 mg, 47 µmol) was dried by repeated
coevaporations with dry pyridine and dry toluene, and then dissolved
in triethylamine trihydrofluoride (500 µL). The mixture was stirred for
15 h at rt. A 0.1 M ammonium acetate buffer (3 mL) was then added
to the mixture, and the mixture was washed with Et2O (3 × 3 mL).
The combined organic layers were back-extracted with 0.1 M am-
monium acetate buffer (3 mL). The combined aqueous layers were then
concentrated to dryness under reduced pressure, and the residue was
purified by reversed-phase column chromatography [a linear gradient
of acetonitrile (0-10%) in 0.1 M ammonium acetate buffer (pH 7.0)]
to afford (Sp)-12a (20.5 mg, 35.4 µmol) as a colorless foam. Rp:Sp )
0.5:>99.5 [determined by reversed-phase HPLC performed with a linear
gradient of 0-10% (30 min) and then 10% (10 min) acetonitrile in 0.1
M ammonium acetate buffer (pH 7.0) at 50 °C at a rate of 1.0 mL/
min]. 1H and 31P NMR spectra and the RP-HPLC profile were identical
to those of the authentic sample synthesized by the conventional
H-phosphonate method.25 The H and 31P NMR spectra and the RP-
1
HPLC profile of (Sp)-12a are given in the Supporting Information.
(Rp)-Ammonium Thymidin-3′-yl Thymidin-5′-yl Phosphoro-
thioate [(Rp)-12a]. (Rp)-12a (20.6 mg, 36 µmol) was obtained as a
colorless foam, using (Rp)-11 (52.1 mg, 49 µmol) as a starting material.
Rp:Sp ) >99.5:0.5 [determined by reversed-phase HPLC performed
with a linear gradient of 0-10% (30 min) and then 10% (10 min)
acetonitrile in 0.1 M ammonium acetate buffer (pH 7.0) at 50 °C at a
rate of 1.0 mL/min]. 1H and 31P NMR spectra and the RP-HPLC profile
were identical to those of the authentic sample synthesized by the
conventional H-phosphonate method.25 The H and 31P NMR spectra
and the RP-HPLC profile of (Rp)-12a are given in the Supporting
Information.
1
all-(Rp)-[TPS]9T (18). According to the typical procedure described
above, 5′-O-(DMTr)thymidine 3′-O-succinate bound to HCP (18.5 mg,
0.5 µmol) gave 18 [1.52 A260 units, 17.7 nmol (35%) based on the
assumption of 7% hypochoromicity: UV (H2O) λmax 267 nm, λmin 236
nm] after purification of 1/10 of the crude product by RP-HPLC with
a linear gradient of 0-20% acetonitrile in 0.1 M ammonium acetate
buffer (pH 7.0) for 60 min at 50 °C at a rate of 0.5 mL/min. MALDI-
TOF-MS: m/z 3121 (M - H+)-. 18 was completely digested by
incubation with svPDE for 1 h at 37 °C; <4% of 18 was digested by
incubation with nuclease P1 for 1 h at 37 °C (the experimental
procedure is described below).8
Digestion with svPDE. A 10 µL sample of the crude 15, 17
dissolved in 500 µL of H2O (<10 nmol), or purified 18 (5 nmol) was
diluted with 50 mM Tris-HCl buffer (pH 8.0, 100 µL). Snake venom
phosphodiesterase (20 units) was added, and the mixture was incubated
for 1 h at 37 °C. The enzyme was inactivated upon heating (100 °C, 1
min), and then the mixture was analyzed by RP-HPLC at 260 nm.
Digestion with Nuclease P1. A 10 µL sample of the crude 16
dissolved in 500 µL of H2O (<10 nmol) or purified 18 (5 nmol) was
diluted with 50 mM sodium acetate buffer (pH 5.3, 100 µL). Nuclease
P1 (20 units) was added, and the mixture was incubated for 1 h at 37
°C. The enzyme was inactivated upon heating (100 °C, 1 min), and
then the mixture was analyzed by RP-HPLC at 260 nm.
Ammonium Thymidin-3′-yl Thymidin-5′-yl Phosphorothioate
(12a). 12a was synthesized according to the conventional H-phos-
phonate method25 as a mixture of diastereomers [(Rp)-12a:(Sp)-12a )
41:59]. 1H and 31P NMR spectra and the RP-HPLC profile of 12a are
given in the Supporting Information.
Typical Procedure for Manual Solid-Phase Synthesis. Each cycle
of the chain elongation consisted of detritylation (3% TCA in CH2Cl2;
3 × 5 s), washing (CH2Cl2 followed by CH3CN), coupling (0.2 M
monomer 13 and 1.0 M 1f in CH3CN; 90 s), washing (CH3CN), capping
[Ac2O-N-methylimidazole-THF (1:2:7, v/v/v); 30 s], washing (CH3CN),
sulfurizing (0.5 M Beaucage reagent in CH3CN; 60 s), and washing
(CH3CN). The average yield per cycle was estimated to be 97-98%
by DMTr+ assay. After the chain elongation, the DMTr group was
removed by treatment with 3% TCA in CH2Cl2 (3 × 5 s), and washed
with CH2Cl2. The oligomer on the HCP resin was then treated with
25% NH3(aq)-pyridine (9:1, v/v) for 30 min (in the cases of 15, 16,
and 18) or 15 h (in the case of 17) at 55 °C to remove the chiral
auxiliaries and the protecting groups of the nucleobases and also to
release the oligomer from the HCP resin. The HCP resin was removed
by filtration and washed with H2O. The filtrate was concentrated to
dryness. The residue was dissolved in H2O (1 mL) and washed with
Et2O (3 × 1 mL), and the combined washings were back-extracted
with H2O (1 mL). The combined aqueous layers were concentrated to
dryness. The resulting crude product was analyzed and/or purified by
reversed-phase HPLC with a linear gradient of 0-20% acetonitrile in
0.1 M ammonium acetate buffer (pH 7.0) for 60 min at 50 °C at a rate
of 0.5 mL/min.
(26) (a) Wilk, A.; Stec, W. J. Nucleic Acids Res. 1995, 23, 530-534. (b) Tamura,
Y.; Miyoshi, H.; Yokota, T.; Makino, K.; Murakami, A. Nucleosides
Nucleotides 1998, 17 (1-3), 269-282.
9
8316 J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003

Stereocontrolled Synthesis of PS-ODNs
A R T I C L E S
1
Acknowledgment. We thank Professor Jacek Stawin´ski
(Stockholm University) for useful suggestions. We also thank
Professor Masaaki Ueki (Tokyo University of Science) for
obtaining HRMS spectra. This work was supported by Grants
from the Ministry of Education, Culture, Sports, Science and
Technology, Japan.
n, 2a-f, 5a-f, 6b,d, and 13a-d, H and 31P NMR spectra of
(Rp)-11, (Sp)-11, and 11 (mixture of diastereomers), 1H and 31
P
NMR spectra and the RP-HPLC profile of (Rp)-12a, (Sp)-12a,
and 12a (mixture of diastereomers), RP-HPLC profiles of the
mixtures obtained by the digestion of crude 15 with svPDE,
crude 17 with svPDE, 18 with svPDE, crude 16 with nuclease
P1, and 18 with nuclease P1, and RP-HPLC profiles of random-
[TPS]3T and random-d[CPSAPSGPS]T (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
Supporting Information Available: 1H and 31P NMR spectra
1
of crude (5R)-3a and distilled (5R)-3a, H spectrum of crude
3d, experimental details for the configurational assignment of
5, experimental details and characterization data for 1a-e,g-
JA034502Z
9
J. AM. CHEM. SOC. VOL. 125, NO. 27, 2003 8317