O-selectivity and utility of phosphorylation mediated by phosphite triester intermediates in the N-unprotected phosphoramidite method

Article abstract of DOI:10.1021/ja048125h

Previously, O-selective phosphorylation on polymer supports in the N-unprotected phosphoramidite method could not be carried out because the amino groups of dA and dC have high reactivity toward tervalent phosphorus(III)-type phosphitylating reagents. In

Full text of DOI:10.1021/ja048125h

Published on Web 08/17/2004
O-Selectivity and Utility of Phosphorylation Mediated by
Phosphite Triester Intermediates in the N-Unprotected
Phosphoramidite Method
Akihiro Ohkubo,†,§ Yusuke Ezawa, Kohji Seio, and Mitsuo Sekine*,†,§
†
‡,§
Contribution from the Department of Life Science, Tokyo Institute of Technology,
DiVision of Bioscience and Biotechnology, Frontier CollaboratiVe Research Center, and
CREST, Japan Science and Technology Agency (JST), 4259 Nagatsuta, Midoriku,
Yokohama 226-8501, Japan
Received April 1, 2004; E-mail: msekine@bio.titech.ac.jp
Abstract: Previously, O-selective phosphorylation on polymer supports in the N-unprotected phosphora-
midite method could not be carried out because the amino groups of dA and dC have high reactivity toward
tervalent phosphorus(III)-type phosphitylating reagents. In this paper, we developed a new coupling strategy
named the “activated phosphite method” in which the phosphitylation is mediated by phosphite triester
intermediates 1. Application of 1-hydroxybenzotriazole as the promoter to the solid-phase synthesis resulted
in excellent O-selectivity of more than 99.7%. This O-selectivity was explained by the frontier molecular
orbital interactions between the reactive intermediates and the nucleophiles such as the amino or hydroxyl
groups of nucleosides. Furthermore, longer oligonucleotides were synthesized not only by a manual
operation but also by a DNA synthesizer. The utility of our new method was demonstrated by the successful
synthesis of a base-labile modified oligodeoxyribonucleotide having 4-N-acetyldeoxycytidine residues. Finally,
DNA 20-mers containing dA or dC could be synthesized in good yields by use of a combined reagent of
6-trifluoromethyl-1-hydroxybenzotriazole and benzimidazolium triflate.
phosphotriester6 and H-phosphonate methods,1,7 much less
Introduction
interest has been focused on such P-O bond containing
In the oligodeoxyribonucleotide synthesis reported to date,1
the rapid alcoholysis of the P-N bonds has been a key reaction
to achieve facile and high-yield formation of the internucleotidic
intermediates, especially in the phosphoramidite method.8
In this paper, we report the intriguing properties of the
phosphite intermediates 1a-e which have been discovered by
us in our continuing studies to develop new methods for the
2
linkage. In the current phosphoramidite approach, P-N bond
containing species which are generated by the activation of
synthesis of oligodeoxyribonucleotides in the N-unprotected
2,3
4
9-13
phosphoramidite units with various tetrazole, triazole, and
approach.
These intermediates have a P-O bond activated
5
imidazole derivatives act as reactive intermediates. On the other
by substituted benzotriazol-1-yl or 2,4-dinitrophenyl groups and
hand, although the alcoholysis to give an internucleotidic bond
by the use of reactive intermediates having P-O bonds has also
been demonstrated by several authors in development of the
(
6) (a) Reese, B. C.; Zhuo, Z. P. J. Chem. Soc., Perkin Trans. 1 1993, 2291-
301. (b) Marugg, J. E.; van der Bargh, C.; Tromp, M.; van der Marel, G.
A.; van Zoest, W. J.; van Boom, J. H. Nucleic Acids Res. 1984, 12, 9095-
2
9
110. (c) de Vroom, E.; Fidder, A.; Marugg, J. E.; van der Marel, G. A.;
van Boom, J. H. Nucleic Acids Res. 1986, 14, 5885-5900.
(
7) Froehler, B. C. In Protocols for oligonucleotides and analogs; Agrawal,
S., Khorana, H. G., Eds.; Humana Press: Totowa, NJ, 1994; Vol. 26, pp
†
Tokyo Institute of Technology.
Frontier Collaborative Research Center.
JST.
‡
63-80 and references therein.
§
(8) (a) van der Marel, G.; van Boeckel, C. A. A.; Wille, G.; van Boom, J. H.
Tetrahedron Lett. 1981, 22, 3887-3890. (b) van der Marel, G. A.; van
Boeckel, C. A. A.; Wille, G.; van Boom, J. H. Nucleic Acids Res. 1982,
10, 2337-2351. (c) Dabkowski, W.; Twrowska, I.; Michalski, J.; Cramer,
F. Tetrahedron Lett. 2000, 41, 7535-7539.
(9) (a) Adamiak, R. W.; Biala, E.; Grzeskowiak, K.; Kierzek, R.; Kraszewski,
A.; Markiewicz, W. T.; Okupniak, J. Stawinski, J. Wiewiorowski, M.
Nucleic Acids Res. 1978, 5, 1889-1905. (b) Finnan, J. L.; Varshney, A.;
Letsinger, R. L. Nuceic Acids Res. Symp. Ser. 1980, 7, 133-145. (c)
Ogilvie, K. K.; Theriault, N.; Sadana, L. J. Am. Chem. Soc. 1977, 99, 7741-
7743. (d) Ogilvie, K. K.; Nemer, M. J. Tetrahedron Lett. 1981, 22, 2531-
2532. (e) Fourrey, J.-L.; Varenne, J. Tetrahedron Lett. 1985, 26, 2663-
2666. (f) Hayakawa, Y.; Nobori, T.; Noyori, R.; Imai, J. Tetrahedron Lett.
1987, 28, 2623-2626.
(10) (a) Gryaznov, S. M.; Letsinger, R. L. J. Am. Chem. Soc. 1991, 113, 5876-
5877. (b) Gryaznov, S. M.; Letsinger, R. L. Nucleic Acids Res. 1992, 20,
1879-1882.
(11) (a) Wada, T.; Sato, Y.; Honda, F.; Kawahara, S.; Sekine, M. J. Am. Chem.
Soc. 1997, 119, 12710-12721. (b) Wada, T.; Mochizuki, A.; Sato, Y.
Sekine, M. Tetrahedron Lett. 1998, 39, 5593-5596. (c) Wada, T.;
Mochizuki, A.; Sato, Y.; Sekine, M. Tetrahedron Lett. 1998, 39, 7123-
7126.
(
1) Current Protocols in Nucleic Acid Chemistry; Beaucage, S. L., Bergstrom,
D. E., Glick, G. D., Jones, R. A., Eds.; John Wiley & Sons: New York,
2
000; and references therein.
(
2) (a) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859-
1862. (b) Matteucci, M. D.; Caruthers, M. H. J. Am. Chem. Soc. 1981,
103, 3185-3191. (c) McBride, L. J.; Caruthers, M. H. Tetrahedron Lett.
1983, 24, 245-248. (d) Beaucage, S. L.; Iyer, R. P. Tetrahedron 1992,
48, 2223-2311.
(
3) (a) Sproat, B.; Colonna, F.; Mullah, B.; Tsou, D.; Androus, A.; Hampel,
A.; Vinayak, R. Nucleosides Nucleotides 1995, 14, 255-273. (b) Froeheler,
B. C.; Matteucci, M. D. Tetrahedron Lett. 1983, 24, 3171-3174. (c) Wolter,
A.; Biernat, J.; K o¨ ster, H. Nucleosides Nucleotides 1986, 5, 65-77. (d)
Pon, R. T. Tetrahedron Lett. 1987, 28, 3643-3646. (e) Moriguchi, T.;
Yanagi, T.; Wada, T.; Sekine, M. Tetrahedron Lett. 1998, 39, 3725-3728.
4) Nurminen, E. J.; Mattinen, J. K.; L o¨ nnberg, H. HelV. Chim. Acta 2003,
(
(
8
6, 2005-2008.
5) (a) Beaucage, S. L. Tetrahedron Lett. 1984, 25, 375-378. (b) Vargreese,
C.; Carter, J.; Yegg, J.; Krivjahsky, S.; Settle, A.; Kropp, E.; Peterson, K.;
Pieken, W. Nucleic Acids Res. 1998, 26, 1046-1050. (c) Hayakawa, H.;
Kawai, R.; Hirata, A.; Sugimoto, U.; Kataoka, M.; Sakakura, A.; Hirose,
M.; Noyori, R. J. Am. Chem. Soc. 2001, 123, 8165-8176.
10884
9
J. AM. CHEM. SOC. 2004, 126, 10884-10896
10.1021/ja048125h CCC: $27.50 © 2004 American Chemical Society

O-Selectivity and Utility of Phosphorylation
A R T I C L E S
Scheme 1
Table 1. Selectivity of Condensation in the Activated Phosphite and Previous Methods
can be readily generated by activation of the conventional
phosphoramidite 2 with 1-hydroxybenzotriazole (HOBt) deriva-
tives or phenol derivatives. Interestingly, the tervalent phosphite
intermediates had highly chemoselective reactivity toward the
synthesis of TpT on polymer supports, we screened 25 alcohol-
type reagents (see Supporting Information). Among these phenol
8a,b
and HOBt derivatives, only five compounds [HOBt,
6-tri-
tf
6c
n
6b,c
fluoromethyl-HOBt (HO Bt), 6-nitro-HOBt (HO Bt), 4-ni-
nt
6a
5′-hydroxyl group of deoxyribonucleosides over the amino group
tro-6-trifluoromethyl-HOBt (HO Bt), and 2,4-dinitrophenol
8
c
of unprotected nucleobases and allowed chain elongation without
using any protecting groups for adenine, guanine, cytosine, and
thymine. Moreover, it should be noted that, when HOBt was
used as a promoter, the chemoselectivity was much higher than
(DNP) ] showed coupling yields of more than 95%. Although
studies on the internucleotidic bond formation by use of HOBt
and DNP have been previously reported in the conventional
N-protected approach, their properties in the N-unprotected
DNA synthesis have not been demonstrated. In Table 1, the
14
those of the previously reported N-unprotected DNA synthesis.
1
2
Here, we report the interesting properties of the phosphite
intermediates 1a-e in detail in terms of reactivity, selectivity,
and applicability to the automated synthesis of natural DNA.
These unique and excellent results can be explained by ab initio
molecular orbital calculations of the phosphite intermediates.
Finally, the potential utility of our “activated phosphite method”
was demonstrated by the successful synthesis of a modified
DNA having base-labile substituents.
results obtained by use of imidazolium triflate (IMT) and
1
3
4-nitrobenzimidazolium triflate (NBT) are also shown as
references that exhibited relatively high O-selectivity via the
formation of reactive intermediates having a P-N bond.
O-Selective Phosphorylation Promoted by HOBt in Solu-
tion-Phase Synthesis. The efficiency of HOBt as the promoter
for the O-selective phosphorylation was demonstrated by the
condensation of 1.2 equiv of the thymidine 3′-phosphoramidite
unit 4 with 3′-O-(tert-butyldimethylsilyl)deoxyadenosine [3′-
O-tBDMS-deoxyadenosine (5)] and 3′-O-tBDMS-deoxycytidine
Results and Discussion
Selection of Activators. The activators tested in this study
are summarized in Table 1. To find suitable activators for the
(6) in the presence of 2.2 equiv of HOBt in acetonitrile for 5
min at room temperature (RT), as shown in Scheme 2. The
amino groups of the adenine and cytosine moieties were not
blocked by any protecting groups. After the condensation,
oxidation was carried out by use of I2 in pyridine-water for 2
min at room temperature to give the desired d(TpA) and d(TpC)
derivatives 7 and 8. The N-phosphorylated byproducts could
(
(
(
12) (a) Hayakawa, Y.; Kataoka, M. J. Am. Chem. Soc. 1998, 120, 12395-
1
2401. (b) Hayakawa, Y.; Kawai, R.; Kataoka, M. Eur. J. Pharm. Sci.
001, 13, 5-16.
2
13) (a) Ohkubo, A.; Seio, K.; Sekine, M. Nucleic Acids Res. Suppl. 2001, 1,
7
5
7-78. (b) Sekine, M.; Ohkubo, A.; Seio, K. J. Org. Chem. 2003, 68,
478-5492.
14) (a) Sekine, M.; Ohkubo, A.; Seio, K. Nucleic Acids Res. Suppl. 2002, 2,
2
3
9-30. (b) Ohkubo, A.; Sekine, M.; Seio, K. Tetrahedron Lett. 2003, 45,
63-366.
31
not be detected at all by P NMR in these reactions. Compounds
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10885

A R T I C L E S
Ohkubo et al.
Scheme 2 a
a
Reagents and conditions: (a) HOBt, CH3CN, RT, 5 min; (b) I2, pyridine-H2O (9:1, v/v), RT, 5 min.
Scheme 3
7
and 8 were isolated in 91 and 92% yields, respectively, by
the phosphoramidite unit 9a or 9c (Figure 1e,f). These results
clearly showed that the phosphite intermediate 1a generated
from the phosphoramidite building blocks has excellent chemo-
selectivity toward the 5′-hydroxyl group on the resin under these
solid-phase conditions.
silica gel column chromatography. These results indicate that
the condensation by using HOBt has both sufficient coupling
efficiency and high O-selectivity at least under the solution-
phase conditions.
O-Selective Phosphorylation Promoted by HOBt in Solid-
Phase Synthesis. Next, we examined the O-selective inter-
nucleotidic bond formation under solid-phase conditions ac-
cording to a manual operation shown in Scheme 3. Although
Since the T-loaded resin was used in these reactions, no free
nucleophilic amino groups were present on the HCP resin. The
phosphoramidite units 9a and 9c have free amino groups, so it
is somewhat difficult to estimate the contribution of the
N-phosphorylation to the overall side reactions. Therefore, to
clarify the reactivity of the phosphite intemediate 1a generated
by HOBt toward the unprotected amino groups of the DNA
chain on the resin, we carried out the synthesis of d[TpApT]
and d[TpCpT] by the condensation of d[Ap(ce)T]-HCP and
d[Cp(ce)T]-HCP, which were synthesized in advance on the
HCP resin by use of HOBt as the activator, with a large excess
amount of thymidine 3′-phosphoramidite 4 in the presence of
HOBt or IMT. These results are shown in Figure 1c,d,g,h. When
IMT was used as the promoter, considerable side reactions were
detected, as shown in Figure 1c,d. Particularly, the reaction of
1
5
CPG resins have been generally used in the synthesis of
oligodeoxynucleosides by use of DNA synthesizers, we selected
commercially available highly cross-linked polystyrene (HCP)
resins16 since the coupling on the HCP resin was very high and
most suitable for the manual operation in our previous studies.
In Scheme 3, the final deprotection and release of oligomers
from the HCP resin were performed by treatment with concen-
trated NH3 aq for 12 h because it took more prolonged time for
removal of the cyanoethyl groups from the phosphoramidate
residues formed on the nucleobases than those present in the
internucleotidic phosphate (O-P(V)) bonds.
Initially, d[ApT] was synthesized on the T-loaded HCP resin
by use of IMT as an activator. In this case, selectively
O-phosphorylated derivatives (d[ApT] and d[CpT]) were formed
with selectivity of only 77.0 and 82.9%, respectively, as shown
in Figure 1a,b.
In contrast, the use of HOBt in place of IMT improved the
O-selectivity remarkably, and the desired dimers d[ApT] and
d[CpT] were formed with selectivity of 99.7 and 99.9%,
respectively, even in the presence of a large excess amount of
d[Cp(ce)T]-HCP with 4 gave extremely poor O-selectivity of
NHpT
9.7% and the N-phosphorylated product (TpC
pT) was
obtained as the major product, as shown in Figure 1d. On the
other hand, when HOBt was used as the promoter, the amounts
of such N-phosphorylated species became closer to the baseline
level, as shown in HPLC profiles of Figure 1g,h, so that the
O-selectivity in these cases was calculated to be more than
99.9%. Thus, d[TpApT] and d[TpCpT] were isolated in 89 and
83% yields, respectively.
Similarly, the O-selectivity of the HOBt-promoted phospho-
rylation was studied in the d[GpT] and d[TpGpT] synthesis.
As the result, N-phosphorylation of the guanine residue could
(
15) K o¨ ster, H.; Biernat, J.; McManus, J.; Wolter, A.; Stumpe, A.; Narang, C.
K.; Sinha, N. D. Tetrahedron 1984, 40, 103-112.
(
16) McCollum, C.; Andrus, A. Tetrahedron Lett. 1991, 32, 4069-4072.
10886 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

O-Selectivity and Utility of Phosphorylation
A R T I C L E S
Figure 1. Anion-exchange HPLC profiles of the crude mixture from synthesis with unprotected nucleobases. Solvent system A was used for HPLC. (a)
d[ApT], IMT; (b) d[CpT], IMT; (c) d[TpApT], IMT; (d) d[TpCpT], IMT; (e) d[ApT], HOBt; (f) d[CpT], HOBt; (g) d[TpApT], HOBt; (h) d[TpCpT], HOBt.
Scheme 4
not be essentially detected. This observation is in agreement
with the previous results of low reactivity of free guanine
residues.
rylation among the HOBt derivatives tested in this study
followed by HO Bt. It should be also noted that the “proton-
tf
1
0,12,13
block strategy” utilizing NBT also gave satisfactory results
tf
O-Selectivity of the Phosphorylation Promoted by Other
comparable to HO Bt. The isolated yields in the synthesis of
Activators. Similar dimer and trimer syntheses were examined
d[ApT], d[CpT], d[GpT], d[TpApT], d[TpCpT], and d[TpGpT]
by use of the best activator HOBt were 92, 94, 90, 89, 83, and
84%, respectively.
tf
n
by use of other promoters such as NBT, HO Bt, HO BT,
nt
HO Bt, and DNP, and the results are listed in Table 1 together
with the results of IMT and HOBt described above. The results
are summarized as follows.
Effects of Phosphate Protecting Groups on the O-Selectiv-
ity and Coupling Efficiency. The steric hindrance around the
phosphorus atom generally affects the coupling efficiency in
In the synthesis of d[ApT] or d[CpT], it was found that the
activated phosphite method using alcohol-type activators dra-
matically increased the O-selectivity of the phosphorylation
compared with that of IMT. (See the first and second rows of
Table 1.) Particularly, in the synthesis of d[ApT] and d[CpT],
HOBt suppressed N-phosphorylation about 77 and 170 times
as much as IMT. It was also shown that HOBt was superior to
NBT, used in the “proton-block” method previously reported
by us,1 in terms of O-selectivity. With an increase of the acidity
of the promoter, the O-selectivity gradually decreased in the
condensation mediated by HOBt derivatives in the order HOBt
1
7
the phosphoramidite method. The smaller the phosphate
protecting group is, the higher the coupling efficiency is.1
Therefore, we examined whether the steric hindrance around
the phosphorus atom affected not only the reactivity but also
the O-selectivity in the activated phosphite method. Thus, we
examined the O-selectivity of phosphorylation mediated by
HOBt and the phosphoramidite unit 10 having a sterically
compact methyl group instead of the more bulky 2-cyanoethyl
group. The protocol is shown in Scheme 4. For the deprotection
of the phosphate group, we used the conditions of thiophenol-
triethylamine for 1 h at room temperature.
7a
3
tf
n
nt
>
HO Bt > HO Bt > HO BT. A similar trend was also
observed for the synthesis of d[TpApT] and d[TpCpT]. Espe-
cially when highly acidic HO Bt was used as the promoter in
the synthesis of TpCpT, the O-selectivity was very poor among
the promoters used in this study. However, the value is still ca.
The O-selectivity in the synthesis of d[ApT], d[CpT], and
d[GpT] showed 99.2, 98.9, and >99.9%, respectively, as
calculated by the HPLC profiles shown in Figure 2. These results
indicate that the O-selectivity of the activated phosphite method
promoted by HOBt decreases slightly due to the steric hindrance
of the phosphate protecting group and is as high as that obtained
nt
6
.6 times higher than that of IMT. The condensations mediated
by DNP constantly showed higher O-selectivity than IMT and
nt
HO Bt, while the rates are somewhat low compared with those
tf
by HO Bt, the second best activator in this study. Thus, the
activated phosphite method allows the general use of N-
of HOBt in all cases.
In the synthesis of d[GpT] (the third row in Table 1) by use
of any promoter, there were also essentially no N-phosphorylated
byproducts. On the basis of the data in Table 1, it was concluded
that HOBt was the best in terms of the O-selective phospho-
(17) (a) Dahl, B. H.; Nielsen, J.; Dahl, O. Nucleic Acids Res. 1987, 15, 1729-
1
743. (b) Hirata, A.; Sugimoto, J.; Kawai, R.; Kataoka, M.; Hayakawa, Y.
Nucleic Acids Res. Suppl. 2001, 1, 215-216. (c) Pitsch, S.; Weiss, P. A.;
Jenny, L.; Stutz, A.; Wu, X. HelV. Chim. Acta 2001, 84, 3773-3795.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10887

A R T I C L E S
Ohkubo et al.
Figure 2. Anion-exchange HPLC profiles of the crude mixtures from synthesis by use of HOBt and N-unprotected phosphoramidite units 14a,c,g. Solvent
system A was used for HPLC. (a) d(ApT); (b) d(CpT); (c) d(GpT).
3
1
i
i
Figure 3. P NMR spectra of (MeO)2PN Pr2 15 (a) and crude mixture obtained after (MeO)2PN Pr2 15 was activated by 0.25 M solution of HOBt in
CD3CN-NMP (9:1, v/v) at room temperature for 5 min (b) and for 30 min (c).
Scheme 5
unprotected phosphoramidite units with higher coupling ef-
ficiency.
Involvement of Pentavalent Intermediates in the HOBt-
Mediated Phosphorylation. In the previous study, we found
the conversion of the tervalent phosphite benzotriazol-1-yl ester
into the corresponding pentavalent dialkyl phosphorobenzot-
reagent.19 On the other hand, when the pentavalent product 18
generated via the rearrangement reacts with the 5′-hydroxyl
group, the undesired product 16 having a phosphate backbone
(P-O) should be generated (Figure 4). Consequently, the ratio
of the P-S compound 12 to the P-O compound 16 reflects
the contribution of the pentavalent rearrangement products to
the coupling reaction.
1
1a,18
riazolidate derivative.
(
Actually, the phosphoramidite unit
MeO)2PN Pr2 11 ( P NMR: 150.2 ppm) was immediately
converted to the corresponding phosphite compound (MeO)2POBt
i
31
For example, the condensation by use of benzimidazolium
triflate (BIT), which generates only the tervalent azolide
intermediate, followed by the sulfurization resulted in the
formation (98.6%) of the desired P-S compound 15 along with
1.4% of the P-O product, as determined from the integrated
areas of these HPLC peaks (Figure 5). On the other hand, the
use of HOBt as the promoter increased the P-O product to
4.9%. Although this result suggests that the intermediates using
the HOBt system involve not only the phosphite benzotriazol-
1-yl ester 17a but also the pentavalent phosphorus compound
18a, more than 95% of the hydroxyl groups on the HCP resin
reacted with the trivalent phosphite intermediate in the HOBt-
promoted phosphitylation.
3
1
1
2 ( P NMR: 146.6 ppm) by treatment with 5 equiv of HOBt
as a promoter. The resulting phosphite (MeO)2POBt 12 rapidly
changed to the phosphotriester compound (MeO)2P(O)OBt 14
3
1
(
(
P NMR: 0.8 ppm) via the phosphoramidate compound
MeO)2P(O)Bt 12, as shown in Scheme 5 and Figure 3.
Since there was a possibility of the phosphorylation via the
pentavalent phosphorus intermediates in the activated phosphite
method, we tried to quantify the contribution of the pentavalent
phosphorus intermediate to the overall condensation. For this
purpose, we planned the synthesis of the Tp(s)T phosphorothio-
ate dimer 15, as shown in Scheme 6. When the phosphite
intermediate reacts with the 5′-hydroxyl group of thymidine on
HCP, the desired product 15 having a phosphorothioate inter-
nucleotidic bond (P-S) should be generated by Beaucage’s
tf
Interestingly, the ratio in the condensation by use of HO Bt
remarkably increased up to 10%. This increase can be explained
in terms of two mechanisms. One might be based on the
(
18) Ohkubo, A.; Aoki, K.; Sekine, M.; Seio, K. Tetrahedron Lett. 2003, 45,
(19) Iyer, R. P.; Egan, W.; Reagan, J. B.; Beaucage, S. L. J. Am. Chem. Soc.
1990, 112, 1253-1254.
9
79-982.
10888 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

O-Selectivity and Utility of Phosphorylation
A R T I C L E S
Figure 4. Reaction mechanism of the phosphorylation in the activated phosphite method.
tf
Figure 5. Reversed-phase HPLC profiles of the crude mixtures from synthesis of TpT having a P-S bond by use of BIT, HOBt, or HO Bt. Solvent system
tf
was used for HPLC. (a) BIT; (b) HOBt; (c) HO Bt.
Scheme 6
assumption that the rearrangement of the tervalent phosphite
intermediate 17b proceeded much faster than that of 17a. The
other might be explained by the assumption that the reactivity
of the pentavalent phosphorus compound 18b is much higher
than that of 18a. Our early study suggested that the rearrange-
ment of a tervalent phosphite generated by HO Bt is about 6
times as fast as that by use of HOBt.18 This result strongly
supports the former mechanism.
(FMO) interaction in our previous N-unprotected H-phosphonate
1
1a
method. Similarly, a plausible reaction mechanism for the
O-selective phosphorylation via the phosphite intermediate can
be considered as follows. In the first step of the phosphitylation,
the lone pair of the phosphorus atom interacts with the hydrogen
atom of the hydroxyl group. Since this interaction is expected
to increase both the nucleophilicity of the hydroxyl oxygen and
the electrophilicity of the phosphorus center, the successive
nucleophilic attack of the hydroxyl group toward the phosphorus
atom proceeded more readily. In contrast to the hydroxyl group,
such an acid-base interaction must be less prominent for the
amino group because of the lower acidity of the N-H proton.
As a result, the amino groups become less reactive to the
tf
Theoretical Studies of the Mechanism of the
O-Selective Phosphorylation
Previously, we proposed the reaction mechanism for the
O-selective condensation based on frontier molecular orbital
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10889

A R T I C L E S
Ohkubo et al.
Figure 6. Optimized geometries and MOs of the reaction intermediates at the HF/6-31G* level: (A) phosphite intermediate 12; (B) protonated imidazolide
intermediate 19. The MOs are visualized on the threshold of 0.030.
phosphite intermediates. (See the Supporting Information re-
garding the molecular orbitals (MOs) of the nucleobases and
MeOH.)
To evaluate this hypothesis, the calculation of the MO of the
phosphite intermediate model 12 at the HF/6-31G* level was
carried out. It was found that there was a large unoccupied
orbital on the phosphorus center in the LUMO+1 of the
phosphite intermediate 12, as shown in Figure 6A. Considering
the high energy of this LUMO+1 (97.99 kcal/mol), compound
12 seems less reactive toward the nucleophilic hydroxyl and
10890 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

O-Selectivity and Utility of Phosphorylation
A R T I C L E S
Figure 7. Reaction mechanism of phosphitylation mediated by the phosphite intermediate.
amino groups of deoxyribonucleoside derivatives. However, the
intermediate 12 also has an occupied orbital on the phosphorus
center which is included in the HOMO that can interact at the
hydrogen atom with the LUMO of alcoholic nucleophiles, as
shown in Figure 6. It is likely that such two simultaneous FMO
interactions between the tervalent phosphorus intermediate and
the alcoholic OH group account for the high O-selectivity in
our method.
Similarly, it was suggested that the protonated imidazolide
intermediate 19 has a very large unoccupied orbital on the phos-
phorus atom, as shown in Figure 6B. Considering the energy
level of the LUMO of the imidazolide (-14.42 kcal/mol), the
compound seems to be reactive enough to interact with the
HOMO or HOMO-1 on the oxygen atoms of the hydroxyl
groups and the nitrogen atoms of the amino groups. Furthermore,
the protonation on the phosphorus atom of imidazolide inter-
mediate 19 by the hydroxyl group as shown in Figure 7 seems
less prominent because of the low energy level (-336.03 kcal/
mol) of the HOMO-1 of the intermediate 19 which corresponds
to the lone pair on the phosphorus center. Therefore, the use of
such an acid-azole complex salt as the promoter must result
in poor O-selectivity in the phosphorylation.
Utility of the Activated Phosphite Method for the Auto-
mated Synthesis of Longer and Modified Oligodeoxyribo-
nucleotides. To demonstrate the superiority of our activated
phosphite method, we previously reported the synthesis of
d[AAAAAAT], d[CCCCCCT], d[GGTGGTGGT], and d[CAGT-
CAGTCAGT] by a manual operation, and the N-phosphoryla-
tion side reactions were checked by anion-exchange HPLC (see
Supporting Information). Two consecutive couplings at each
cycle were employed to increase the coupling efficiency of the
condensation to as high as possible. In the synthesis of
d[AAAAAAT], the desired product was obtained as the main
peak. Likewise, there was no detectable peak arising from
N-phosphorylation in the synthesis of d[CCCCCCT]. These
results suggested that the N-phosphorylations were almost
completely suppressed even in the successive sequence of dA
or dC. In these cases, the coupling efficiency was more than
the practical usefulness of the method. Unexpectedly, it was
found that the coupling efficiency by using HOBt as the
promoter was not high enough when it was applied to an ABI
392 DNA synthesizer. The coupling efficiency was 97%,
although two consecutive couplings at each cycle were em-
ployed. In our previous study, it was found that the use of more
acidic promoters in the phosphoramidite chemistry could
2
0
increase the coupling efficiency. Therefore, we chose a more
tf
acidic activator, HO BT. Consequently, it was found that this
system gave the best result of high O-selectivity comparable
with HOBt in the manual operation, as shown in Table 1. We
tf
used a 0.2 M solution of HO Bt in CH3CN-N-methyl-2-
pyrrolidinone (NMP) (15:1, v/v) instead of a 0.45 M CH3CN
solution of 1H-tetrazole used in the protocol of the current DNA
synthesis. The synthetic protocol was the same as the manual
operation described above.
We reexamined which resin was the most suitable for the
tf
coupling by use of HO Bt in the automated DNA synthesis.
We synthesized TpT on the synthesizer in a single-coupling
mode on HCP (ABI), primer support, TentaGel, ArgoPore, and
tf
CPG by use of a 0.2 M solution of HO Bt in CH3CN-NMP
(15:1, v/v). As a result, the coupling yields were found to be
99.0, 96.4, 98.7, 88.0, and 97.8%, respectively. These results
showed HCP was the choice of reagent in the automated
activated phosphite method.
On the basis of these results, the automated synthesis of
d[CAGTCAGTCAGT] on HCP was carried out by use of
tf
HO Bt. The HPLC profile of the mixture obtained was similar
to that of the 12-mer in the manual operation, as shown in Figure
8a. It was also confirmed that the N-phosphorylated byproducts
are rarely observed. The isolated yield was 32%. This result
indicates that the activated phosphite method with the use of
tf
HO Bt can be applied to the DNA synthesizer system.
In the same protocol we also synthesized an R-oligodeoxy-
ribonucleotide 12-mer d[TC*TTC*C*TTC*TTT] having 5-
methylcytosine base (C*), which was frequently used in
antisense or antigene method because the nucleobase has a
stacking effect and is more basic than a cytosine base.21 Since
the 3-nitrogen atom of this nucleobase is protonated more easily
than that of cytosine, a Hoogsteen base pair of this nucleobase
with a guanine base is easily formed in DNA triple helix. It is
also well-known that R-oligodeoxyribonucleotides have high
9
9% (trityl cation assay). d[AAAAAAT] and d[CCCCCCT]
were satisfactorily isolated in 48 and 67% yields, respectively.
The amounts of the (n - 1)-mers lacking a nucleotide
increased to some extent in the synthesis of G-rich sequence
d[GGTGGTGGT] because of the somewhat lower average
coupling yield, 98.1%. However, no N-phosphorylated byprod-
uct was observed as in the cases of d[GpT] and d[TpGpT]. The
excellent O-selective phosphorylation was also observed in the
synthesis of d[CAGTCAGTCAGT]. Although (n - 1)-mer side
products were observed because of the lower coupling efficiency
at the dG positions, they could be separated by HPLC. The
2
2
resistance to nucleases.
In the synthesis of the R-DNA 12-mer, the desired product
was obtained as the main peak, as shown in Figure 8b. The
N-phosphorylations were almost completely suppressed even
in the presence of 5-methyl-dC. In these cases, the coupling
efficiency was more than 99% (trityl cation assay). The R-DNA
(
20) Moriguchi, T.; Yanagi, T.; Kunimori, M.; Wada, T.; Sekine, M. J. Org.
Chem. 2000, 65, 8229-8338.
1
2-mer was isolated in 36% yield. These oligomers were well
(
21) (a) Kemal, O.; Brown, T.; Burgess, S.; Bishop, J.-D.; Leigh-Brown, A. J.
Nucleosides Nucleotides 1991, 10, 555-561. (b) Morvan, F.; Rayner, B.;
Imbach, J.-L. Anti-Cancer Drug Des. 1991, 6, 521-529.
characterized by enzyme digestion and MALDI-TOF mass
spectroscopy.
Next, we tried to apply the activated phosphite method to
the manipulation using a DNA synthesizer in order to elucidate
(
22) Chaix, C.; Toulme, J.-J.; Morvan, F.; Rayner, B.; Imbach, J.-L. In Antisense
Research and Applications; Crooke, S. T., Lebleu, B., Eds.; CRC Press:
Boca Raton, FL, 1993; Chapter 12, pp 223-234, and references therein.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10891

A R T I C L E S
Ohkubo et al.
tf
tf
Figure 8. Anion-exchange HPLC profiles of the crude mixtures obtained by using a 0.2 M HO Bt solution (method A) or a 0.2 M HO Bt + BIT solution
method B) under the condition used for ABI 392 DNA synthesizer. Solvent system C or D was used for HPLC. (a) d(CAGTCAGTCAGT), method A,
(
ac
ac ac
solvent system C; (b) R-d(TC*TTC*C*TTC*TTT), method A, solvent system D; (c) d(GA ATCAGC C TCAT, method A, solvent system C; (d)
d(CCCCCTTTTTCTCTCTCTCT), method A, solvent system D; (e) d(CCCCCTTTTTCTCTCTCTCT), method B, solvent system D; (f) d(TTAAAAAT-
TATTAAATTATT), method B, solvent system D.
Figure 9. Oligonucleotides and Tm values of their duplexes or triplexes. (a) Experiments were carried out at 2 µM duplex concentration in a buffer containing
1
0 mM sodium phosphate, pH 7, 100 mM sodium chloride in the presence of 0.1 mM EDTA. (b) Experiments were carried out at 2 µM triplex concentration
in a buffer containing 10 mM sodium cacodylate, 100 mM sodium chloride.
12-mer R-d[TC*TTC*C*TTC*TTT] was satisfactorily isolated
in 61% yield. This oligomer was characterized by MALDI-TOF
mass spectroscopy. This R-DNA 12-mer had a high ability of
duplex formation (parallel strand duplex) or triplex formation
toward ssDNA or dsDNA having the compensate sequence
compared with the â-DNA 12-mer d[TTTCTTCCTTCT], as
shown in Figure 9.
Furthermore, we demonstrated the synthesis of a modified
ac
ac ac
DNA 13-mer d[GC ATCAGC C TCAT] having three 4-N-
acetyldeoxycytidine residues23 (dCac). We previously reported
Figure 10. New original T-loaded HCP resin having a silyl-type linker.
that acetylation of the 4-N-amino group of deoxycytidine
stabilizes DNA duplexes, and the base recognition ability of
this modified base is very similar to that of deoxycytidine.
However, this N-acetyl group is sensitive to basic conditions
that are required for removal of N-protecting group in the
conventional DNA synthesis. Because the deacylation of the
acetyl group occurs in aqueous ammonia, we used our new
(TBAF) without using aqueous ammonia, as shown in Figure
10. (Detailed information on this linker will be reported in
another paper.) The protocol of the synthesis was the same as
that of the synthesis on the HCP resin containing a succinate
linker described above. After the final condensation, the
cyanoethyl groups were removed by treatment with a 10%
solution of 1,8-diazabicyclo[5,4,0]-7-undecene (DBU) in CH3-
CN at room temperature for 1 min. The crude oligodeoxyribo-
nucleotide was released from the resin by treatment with a 1 M
TBAF-AcOH solution in THF, and analyzed by HPLC (Figure
2
4
original T-loaded HCP resin having a silyl-type linker which
could be cleaved by treatment with tetrabutylammonium fluoride
(
23) (a) Wada, T.; Kobori, A.; Kawahara, S.; Sekine, M. Tetrahedron Lett. 1998,
3
9, 6907. (b) Wada, T.; Kobori, A.; Kawahara, S.; Sekine, M. Eur. J. Org.
Chem. 2001, 4583-4593.
8
c). The panel 8c shows both excellent O-selectivity and high
(24) Kobori, A.; Miyata, K.; Ushioda, M.; Seio, K.; Sekine, M. Chem. Lett.
2
002, 16-17.
coupling efficiency. The modified 13-mer was isolated in 33%
10892 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

O-Selectivity and Utility of Phosphorylation
A R T I C L E S
Figure 11. Oligonucleotides and Tm values of their duplexes. Experiments were carried out at 2 µM duplex concentration in a buffer containing 10 mM
sodium phosphate, pH 7, 100 mM sodium chloride in the presence of 0.1 mM EDTA.
Figure 12. Possiblility of cleavage of N-P(III) bonds in activated phosphite.
yield. This oligomer was characterized by MALDI-TOF mass
spectroscopy.
This 13-mer proved to have strong affinity for ssDNA having
the complementary sequence compared with DNA 13-mer
d[GCATCAGCCTCAT] without an N-acetyl group, as shown
in Figure 11.
The above-mentioned successful synthesis of the base-labile
modified deoxyoligoribonucleotide strongly demonstrates the
usefulness of the “activated phosphite” method for the synthesis
of base-labile functional DNA molecules.
was the best in our strategy as described above, there was a
possibility that, even if the branched chain species 22 is formed
by reaction of a dimer 20 with an activated phosphoramidite
derivative 21, it might be converted to the parent material 20
and the activated phosphite triester 23 due to the reaction of
N-P(III) bond cleavage by HOBt, as shown in Figure 12.
Therefore, we carried out the experiment as shown in Scheme
7 to examine whether cleavage of N-P(III) bonds occurred in
the coupling reaction. Three consecutive condensations of a
Cp(OCE)T-loaded HCP resin 24 with the thymidine phosphora-
midite unit 4 in the presence of IMT as a promoter were carried
out to generate an N-P(III) bond on the cytosine base. A
mixture 24 and 25 of the resin having the N-P(III) bond was
divided into three portions. The first portion was oxidized and
deprotected (method A). The second was treated with HOBt
before oxidation (method B). The last was treated with the
phosphoramidite unit 4 in the present of HOBt before oxidation
(Scheme 7).
Further Improvement of Activated Phosphite Approach.
Furthermore, a protocol for the automated DNA synthesis using
tf
a 0.2 M solution of HO Bt was applied to the synthesis of a
DNA 20-mer d[CCCCCTTTTTCTCTCTCTCT], as shown in
Figure 8d. The amounts of the (n - 1)-mers lacking a nucleotide
increased to some extent in the HCP profile because the coupling
efficiency was not complete.
We have also made strenuous efforts to increase the coupling
efficiency in our activated phosphite approach. As a result, it
turned out that the coupling efficiency significantly increased
when the condensation was carried out in the presence of
Method A gave a 2:3 mixture of CNH pT 27 and CNHpTpT
2
26, as shown in Figure 13a. However, it turned out that
posttreatment with 20 equiv of HOBt in CH3CN for 1 min
resulted in cleavage of the N-P(III) bond of 25 on polymer
supports, as shown in Figure 13b. The efficiency of the cleavage
was determined to be 85% from the integrated areas of HPLC
tf
benzimidazolium triflate (BIT) in addition to HO Bt. The
desired oligomer was obtained as a main peak with very high
coupling efficiency and O-selectivity, as shown in Figure 8e.
Actually, the 20-mer could be isolated in 81% yield. This result
strongly suggests that the activated phosphite method could
make the synthesis of antigene molecules having only pyrimi-
dine sequences very facile and immediate. Under similar
NH
NHpT
peaks of C
2
pT 27 and C
pT 26 compared with Figure 13a.
On the other hand, posttreatment of a mixture of 25 and 24
with the T-phosphoramidite unit and HOBt gave a result similar
to that of method B, as shown in Figure 13c. These results
showed that HOBt in the activated phosphite approach could
promote not only O-selective phosphitylation but also cleavage
of the N-P(III) bond.
tf
conditions using a combined reagent of HO Bt + BIT, a DNA
2
0-mer d[TTAAAAATTATTAAATTATT] was also synthe-
sized (Figure 8f). Although both the coupling efficiency and
the O-selectivity were somewhat lower, we could synthesize
the desired compound as a major product without P-N bond
cleavage and capping treatment. This oligomer was isolated in
In conclusion, even if the N-P(III) bond is formed on the
cytosine residue, it can be removed by the action of HOBt
regardless of the presence or absence of phosphoramidite units
to regenerate the parent N-unprotected oligonucleotides with
the release of an activated phosphite triester like 23. This is the
essential reason the condensation of the present approach by
24% yield and characterized by enzyme digestion and MALDI-
TOF mass spectroscopy.
Possibility of Cleavage of N-P(III) Bonds in Activated
Phosphite Approach. Although the condensation in the pres-
tf
use of BIT + HO Bt as promoters also retained high O-
tf
ence of benzimidazolium triflate (BIT) in addition to HO Bt
selectivity in the internucleotidic bond formation.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10893

A R T I C L E S
Ohkubo et al.
Figure 13. Anion-exchange HPLC profiles of the crude mixtures obtained by three condensations of a CpT derivative loaded HCP resin with the thymidine
phosphoramidite unit in the presence of IMT(40 equiv). (a) The mixture obtained when the posttreatment was not carried out after the condensation. (b) The
mixture obtained when the posttreatment with HOBt (20 equiv) in CH3CN at room temperature for 1 min was carried out after the condensation. (c) The
mixture obtained when the posttreatment with HOBt (40 equiv) and T-phosphoramidite (20 equiv) in CH3CN at room temperature for 1 min was carried out
after the condensation.
Scheme 7
Experimental Section
3 4
(0-30%) of solvent III (CH CN) in solvent IV (0.1 M NH Osc buffer
(
pH 7.0)) was used at 50 °C at a flow rate of 1.0 mL/min for 30 min.
General Remarks. H, 13C, and 31P NMR spectra were recorded at
1
System C: Anion-exchange HPLC was done on a Shimadzu LC-10
AD VP with a Shimadzu 3D UV detector and a Gen-PakTM FAX
column (Waters, 4.6 × 100 mm). A linear gradient (0-60%) of solvent
2
70, 68, and 109 MHz, respectively. The chemical shifts were measured
from tetramethylsilane for H NMR spectra, CDCl (77 ppm) for C
3
1
13
3
1
NMR spectra, and 85% phosphoric acid (0 ppm) for P NMR spectra.
UV spectra were recorded on a U-2000 spectrometer. Column chro-
matography was performed with silica gel C-200 purchased from Wako
Co. Ltd., and a minipump for a goldfish bowl was conveniently used
to attain sufficient pressure for rapid chromatographic separation. HPLC
was performed using the following systems. System A: Anion-exchange
HPLC was done on an Waters Aliance system with a Waters 3D UV
detector and a Waters Spherisorb S5 SAX analytical column (4.6 ×
V (1 M NaCl in 25 mM phosphate buffer (pH 6.0)-10% CH
3
CN) in
solvent VI (25 mM phosphate buffer (pH 6.0)-10% CH
3
CN) was used
at 50 °C at a flow rate of 1.0 mL/min for 45 min. System D: Anion-
exchange HPLC was done on a Shimadzu LC-10 AD VP with a
Shimadzu 3D UV detector and a Gen-PakTM FAX column (Waters,
4
2
.6 × 100 mm). A linear gradient (0-80%) solvent V (1 M NaCl in
5 mM phosphate buffer (pH 6.0)-10% CH CN) in solvent VI (25
CN) was used at 50 °C at a
3
mM phosphate buffer (pH 6.0)-10% CH
3
2
50 mm). A linear gradient (0-30%) of solvent I (20% CH
M potassium phosphate) in solvent II (20% CH CN in 0.005 M
potassium phosphate) was used at 50 °C at a rate of 1.0 mL/min for
0 min. System B: Reversed-phase HPLC was done on a Waters
Aliance system with a Shimadzu 3D UV detector and a Gen-PakTM
FAX column (Waters, 4.6 × 100 mm) Waters 3D UV detector and a
µBondasphere 5µ C18 100 Å column (3.9 × 150 mm). A linear gradient
3
CN in 0.5
flow rate of 1.0 mL/min for 45 min. ESI mass spectrometry was
performed by use of a Mariner (PerSeptive Biosystems Inc.). MALDI-
TOF mass was performed by use of a Voyager RP (PerSeptive
Biosystems Inc.). Highly cross-linked polystyrene was purchased from
Perkin-Elmer, ABI. The thymidine-3′-O-phosphoramidite was prepared
according to the standard method. The deoxycytidine-, deoxyadenosine-,
3
3
10894 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

O-Selectivity and Utility of Phosphorylation
A R T I C L E S
Table 2. Reaction Cycle for the Synthesis of Oligodeoxyribonucleotides by Use of ABI 392 DNA Synthesizer
step
operation
reagent(s)
time, min
1
2
3
4
5
6
7
8
washing
detritylation
washing
coupling
coupling
washing
oxidation
washing
CH3CN
3% Cl3CCOOH/CH2Cl2
CH3CN
0.2
1.5
0.4
1.0
1.0
0.2
0.5
0.4
tf
0.1 M amidite + 0.2 M HO Bt in CH3CN-NMP (15:1, v/v)
tf
0.1 M amidite + 0.2 M HO Bt in CH3CN-NMP (15:1, v/v)
CH3CN
0.1 M I2 in Py-H2O-THF (v/v/v)
CH3CN
and deoxyguanosine-O-3′-phosphoramidite derivatives were synthesized
by the method reported by Gryaznov and Letsinger.
Ab Initio Calculations. All ab initio molecular orbital calculations
were carried out using the Gaussian 94 program on a Cray C-916/
steps: (1) detritylation (3% trichloroacetic acid in CH
min); (2) washing [CH Cl CN (1 mL × 3)]; (3) the
(1 mL × 3), CH
first coupling [an appropriate phosphoramidite unit (10 µmol) in
CH CN-NMP (150 µL, 9:1, v/v), HOBt (6.3 mg, 20 µmol) in CH CN-
NMP (150 µL, 9:1, v/v), 1 min]; (4) washing [CH
CN (1 mL × 3)];
(5) the second coupling [an appropriate phosphoramidite unit (10 µmol)
in CH CN-NMP (150 µL, 9:1, v/v), HOBt (6.3 mg, 20 µmol) in CH
CN-NMP (150 µL, 9:1, v/v), 1 min]; (6) washing [CH
CN (1 mL ×
3)]; (7) oxidation [0.1 M I , pyridine-H O (9/1, v/v), 2 min]; and (8)
Cl
2 2
Cl , 2 mL, 1
1
0
2
2
3
3
3
1
2256 supercomputer. Geometry optimizations were carried out at the
3
HF/6-31G* level and single-point energy calculations were carried out
at the MP2/6-31G* level involving electronic correlation to obtain
accurate energies and atomic charges.
Synthesis of the N-Unprotected d[TpA] Derivative. A mixture of
the thymidine 3′-O-phosphoramidite derivative 4 (267 mg, 0.36 mmol)
and 3′-O-(tert-butyldimethylsilyl)deoxyadenosine 5 (109 mg, 0.3 mmol)
was rendered anhydrous by repeated coevaporation successively with
3
3
-
3
2
2
washing [pyridine (1 mL × 3), CH
3
CN (1 mL × 3), CH
2
2
(1 mL ×
3)]. Generally, the average yield per cycle was estimated to be 98-
99% by the DMTr cation assay. After chain elongation was finished,
the DMTr group was removed by treatment with 3% trichloroacetic
dry pyridine (×3), dry toluene (×3), and dry CH
dissolved in dry CH CN (5 mL). To the mixture was added HOBt (97
mg, 0.72 mmol). After the mixture was stirred at room temperature
for 5 min, a 1 M solution of I in pyridine-water (9:1, v/v, 3 mL) was
added to the mixture. After being stirred at room temperature for 2
min, the mixture was partitioned between CHCl (50 mL) and aqueous
% Na (30 mL). The organic phase was collected, washed twice
with aqueous 5% Na (30 mL), filtered, dried over Na SO , and
2
Cl
2
(×3), and finally
3
acid in CH
(1 mL × 3) and CH
and released from the polymer support by treatment with concentrated
NH aq (500 µL) for 40 min or 12 h. The polymer support was removed
by filtration and washed with CH
Cl
2 2
(2 mL) for 1 min, and the resin was washed with CH
2 2
Cl
3
CN (1 mL × 3). The oligomer was deprotected
2
3
3
3
CN (1 mL × 3). The filtrate was
5
2
S O
2 3
evaporated and purified by reversed-phase HPLC or anion-exchange
HPLC.
S
2 2
O
3
2
4
evaporated under reduced pressure. The residue was chromatographed
on a column of silica gel (10 g) with 1% pyridine containing hexane-
On the other hand, in synthesis of the dimers d[ApT], d[CpT], and
[GpT] by use of the phosphoramidite units 14a,c,g having a methyl
group as the protecting group of the phosphate group, the DMTr group
CHCl
3
(50:50-0:100, v/v) and then 1% pyridine containing CHCl -
3
MeOH (100:0-97:3, v/v) to give the fractions containing 7. To remove
the last traces of pyridine, the fractions were collected, evaporated under
reduced pressure, and finally evaporated by repeated coevaporation three
was removed by treatment with 3% trichloroacetic acid in CH
mL) for 1 min, and the resin was washed with CH Cl
(1 mL × 3) and
CH
group was deprotected by treatment with the solution of thiophenol-
triethylamine-dioxane (1:1:2, v/v/v, 400 µL) for 1 h at room temper-
2 2
Cl (2
2
2
3
CN (1 mL × 3) after chain elongation was finished. The methyl
1
times each with toluene and CH
NMR (CDCl
(
(
2
Cl
2
to give 7 (280 mg, 91%):
H
3
) δ 0.11 (s, 6H), 0.90 (s, 6H), 1.40 (s, 3H), 2.30-2.95
m, 6H), 3.28-3.47 (m, 2H), 3.76 (s, 6H), 4.01-4.30 (m, 6H), 4.73
d, 1H, J ) 2.4 Hz), 5.11 (s, 1H), 6.33-6.52 (m, 4H), 6.80 (dd, 4H, J
3.0 Hz, J ) 9.0 Hz), 7.23-7.33 (m, 9H), 8.02 (d, 1H, J ) 2.7 Hz),
ature, and the resin was washed with CH
(1 mL × 3). The oligomer was deprotected and released from the
polymer support by treatment with concentrated NH aq (500 µL) for
40 min or 12 h. The polymer support was removed by filtration and
washed with CH
2
Cl
2
(1 mL × 3) and CH
3
CN
)
3
3
1
13
8
.36 (d, 1H, J ) 4.59 Hz); P NMR (CDCl
3
) δ -2.19, -1.84;
C
3
NMR (CDCl ) δ -4.7, -4.5, 11.7, 11.8, 17.9, 19.1, 19.4, 19.5, 19.6,
3
CN (1 mL × 3). The filtrate was evaporated and
2
7
1
1
1
1
5.7, 39.8, 39.9, 55.2, 62.0, 62.1, 62.3, 63.1, 71.5, 71.7, 77.2, 79.1,
purified by reversed-phase HPLC or anion-exchange HPLC.
B. Synthesis of Phosphorothioate Thymidine Dimer Tp(s)T 15.
A thymidine-loaded HCP (1 µmol, 28 µmol/g, succinate linker) was
used. Each cycle of chain elongation consisted of the following steps:
9.4, 83.8, 83.9, 84.2, 84.4, 85.1, 86.9, 87.0, 111.4, 111.5, 113.2, 115.9,
16.2, 119.4, 119.6, 125.1, 127.0, 127.8, 128.0, 128.1, 128.9, 129.9,
34.9, 139.3, 139.5, 143.8, 143.9, 148.9, 151.1, 151.3, 152.6, 155.7,
58.5, 164.5, 164.7. MS m/z calcd for M + H 1025.3994, found
025.3987.
(1) detritylation (3% trichloroacetic acid in CH
washing [CH Cl CN (1 mL × 3)]; (3) coupling
(1 mL × 3), CH
[thymidine phosphoramidite unit (14.8 mg, 20 µmol) in CH CN-NMP
(150 µL, 9:1, v/v), activator (20 µmol) in CH CN-NMP (150 µL, 9:1,
v/v), 1 min]; (4) washing [CH
CN (1 mL × 3)]; (5) sulfurization [0.2
M 3H-1,2-benzodithiol-3one-1,1-dioxaide, CH CN (400 µL), 2 min];
and (6) washing [pyridine (1 mL × 3), CH CN (1 mL × 3), CH Cl (1
mL × 3)]. After chain elongation was finished, the DMTr group was
removed by treatment with 3% trichloroacetic acid in CH Cl (2 mL)
for 1 min, and the resin was washed with CH Cl
(1 mL × 3) and
CH
CN (1 mL × 3). The oligomer was deprotected and released from
the polymer support by treatment with concentrated NH aq (500 mL)
for 40 min. The polymer support was removed by filtration and washed
with CH
2 2
Cl , 2 mL, 1 min); (2)
2
2
3
In a similar manner, the N-unprotected d[TpC] derivative 8 was
3
1
synthesized: H NMR (CDCl
3
) δ -0.01 (s, 6H), 0.86 (s, 9H), 1.29 (s,
3
3
9
1
1
9
H), 2.11-2.71 (m, 6H), 3.33 (d, 1H, J ) 8.6 Hz), 3.45 (d, 1H, J )
3
.7 Hz), 3.67 (s, 6H), 3.92 (s, 1H), 4.13-4.23 (m, 5H), 5.10-5.22 (m,
H), 5.92 (dd, 1H, J ) 7.1 Hz, J ) 17.6 Hz), 6.15 (m, 1H), 6.36 (dd,
H, J ) 5.1 Hz, J ) 8.9 Hz), 6.77 (d, 4H, J ) 8.9 Hz), 7.23-7.33 (m,
3
3
2
2
31
H), 7.47 (s, 1H), 7.61 (dd, 1H, J ) 4.3 Hz, J ) 7.3 Hz); P NMR
2
2
1
3
(CDCl
3
) δ -1.67, -1.58; C NMR (CDCl
3
) δ -4.8, -4.6, 11.8, 17.9,
2
2
1
7
1
1
9.6, 19.7, 19.8, 25.7, 39.0, 41.2, 55.3, 62.5, 62.6, 63.3, 70.0, 70.4,
7.2, 79.9, 84.2, 84.4, 86.1, 94.9, 111.8, 113.2, 116.2, 116.4, 123.6,
27.1, 127.9, 128.0, 128.9, 129.9, 134.8, 134.9, 135.8, 140.3, 143.8,
49.6, 150.6, 150.7, 155.3, 158.6, 163.7, 163.8, 165.5. MS m/z calcd
3
3
3
CN (1 mL × 3). The filtrate was evaporated and purified by
for M + H 1001.3882, found 1001.3876.
reversed-phase HPLC.
Typical Procedure for Solid-Phase Synthesis. A. Manual Opera-
tion. A thymidine-loaded HCP (1 µmol, 28 µmol/g, succinate linker)
was used. Each cycle of chain elongation consisted of the following
C. Automated Operation. The synthesis of oligodeoxyribonucle-
otides by use of ABI 392 DNA synthesizer was carried out according
to the reaction cycle shown in Table 2. In the synthesis of d(CAGT)
3
,
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 10895
9

A R T I C L E S
Ohkubo et al.
the oligomer after chain elongation was deprotected and released from
B. Triplex. An appropriate oligonucleotide (2 µmol) and its
complementary hairpin DNA 34-mer (2 µmol) were dissolved in three
kinds of buffer consisting of 150 mM NaCl, 10 mM sodium phosphate,
and 0.1 mM EDTA adjusted to pH 7.0, 6.2, or 5.4. The solution was
kept at 80 °C for 10 min for complete dissociation of the triplex to
single strands, cooled at the rate of -0.5 °C/min, and kept at 0 °C for
the polymer support by treatment with concentrated NH
for 40 min. The polymer support was removed by filtration and washed
with CH
CN (1 mL × 3). The filtrate was evaporated and purified by
reversed-phase HPLC or anion-exchange HPLC.
3
aq (500 µL)
3
ac
ac ac
On the other hand, in the synthesis of d[GC ATCAGC C TCAT],
the oligomer after chain elongation was deprotected by treatment with
m
10 min. After that, the melting temperatures (T ) were determined at
a 10% DBU solution in CH
3
CN (500 µL) for 1 min. Then, the
260 nm using a UV spectrometer (Pharma Spec UV-1700, Shimadzu)
oligodeoxyribonucleotide mixture was released from the resin by
treatment with a 1 M TBAF-AcOH (1:1) solution in THF (500 µL)
for 1 h. The polymer support was removed by filtration and washed
by increasing the temperature at the rate of 0.5 °C/min.
Conclusion
with CH
3
CN (1 mL × 3). The filtrate was evaporated and purified by
reversed-phase HPLC or anion-exchange HPLC.
For the first time, we have showed a new method for
phosphorylation giving the hitherto highest O-selectivity and
high coupling yield by use of HOBt or 6-trifluoromethyl-HOBt.
This new approach enabled us to synthesize medium-sized
oligodeoxyribonucleotides without using base protecting groups
in both the solution phase and the automated solid phase. We
proposed the reaction mechanism of the high O-selectivity by
considering the FMO interactions between the reaction inter-
mediates and the nucleophiles on the basis of the quantum
chemical calculations. It is expected that more ideal tervalent
phosphorus intermediates can be designed on the basis of the
theoretical calculations. We also demonstrated the first suc-
cessful synthesis of an base-labile modified oligodeoxyribo-
nucleotide having three 4-N-acetyldeoxycytidines by use of the
O-selective phosphorylation in the activated phosphite method
without base protection. This method prompts us to study the
synthesis of base-labile DNAs such as oxidatively damaged
DNAs or, in the near future, functional DNAs and RNAs with
various alkali-labile modified structures. Further studies are now
underway in this direction.
In the synthesis of d[CCCCCTTTTTCTCTCTCTCT] and d[T-
TAAAAATTATTAAATTATT] a solution of HO Bt (0.2 M) + BIT
tf
(
3
0.2 M) in CH CN-NMP (15:1, v/v) was used as a promoter instead
tf
of a 0.2 M solution of HO Bt.
The mass spectral data of oligodeoxynucleotides are as follows:
d[ApT] mass (M + H) calcd 556.1558, found 556.1523; d[CpT] mass
(
5
M + H) calcd 532.1445, found 532.1443; d[GpT] mass (M + H) calcd
72.1507, found 572.1503; d[Tp(s)T] 15 mass (M - H) calcd 561.4794,
found 561.5063; d[TpT] 16 mass (M - H) calcd 545.4135, found
45.5045; [TpApT] mass (M + H) calcd 860.2017, found 572.1503;
TpCpT] mass (M + H) calcd 836.1905, found 836.1918; [TpGpT]
mass (M + H) calcd 876.1967, found 876.2162; d[AAAAAAT] mass
M - H) calcd 2119.43, found 2119.12; d[CCCCCCT],mass (M - H)
calcd 1975.36, found 1975.22; d[CAGTCAGTCAGT] mass (M - H)
5
[
(
ac
ac ac
calcd 3643.65, found 3642.62; d[GC ATCAGC C TCAT] mass (M
H) calcd 4017.72, found 4018.85; d[TC*TTC*C*TTC*TTT] mass
M + H) calcd 3584.41, found 3584.37; d[CCCCCTTTTCTCTCTCTCT]
-
(
mass (M + H) calcd 5868.23, found 5869.92; d[TTAAAAATTAT-
TAAATTATT] mass (M + Na) calcd 6130.31, found 6132.69.
Enzyme Assay of Oligonucleotides. The enzymatic digestion was
performed by using an appropriate oligodeoxynucleotide (0.5 OD),
snake venom phosphodiesterase (4 µL), and calf intestine alkaline
phosphatase (2 µL) in 50 µL of alkaline phosphatase buffer [500 mM
Acknowledgment. This work was supported by a Grant from
CREST of JST (Japan Science and Technology Agency) and a
Grant-in-Aid for Scientific Research from the Ministry of
Education, Culture, Sports, Science and Technology, Japan.
2
Tris‚HCl (pH 9.0), 10 mM MgCl ] at 37 °C for 40 min. After the
enzymes were deactivated by heating at 100 °C for 1 min, the solution
was diluted and filtered by a 0.45 µm filter (Millex-HV, Millipore).
The mixture was analyzed by reversed-phase HPLC. For d[CCCCCCT],
dC:T ) 6.00:1.03; for d[AAAAAAT], dA:T ) 6.00:0.97; for d[CAGT-
CAGTCAGT], dA:dG:dC:T ) 1.00:1.20:0.95:0.94.; for d[CCCCC-
TTTTCTCTCTCTCT], dC:T ) 1.00:0.99; for d[TTAAAAATTAT-
TAAATTATT], dA:T )1.00:0.94.
Supporting Information Available: List of alcohol-type
activators tested in this study; H, 32P, and 13C NMR data and
ESI mass spectra of 7, 8, 15, and 16; mass spectra of all
synthesized oligonucleotide derivatives; enzyme assay of oli-
1
m
T Measurement. A. Duplex. An appropriate oligonucleotide (2
µmol) and its complementary ssDNA 12-mer (2 µmol) were dissolved
in a buffer consisting of 150 mM NaCl, 10 mM sodium phosphate,
and 0.1 mM EDTA adjusted to pH 7.0. The solution was kept at 80 °C
for 10 min for complete dissociation of the duplex to single strands,
cooled at the rate of -1.0 °C/min, and kept at 15 °C for 10 min. After
gonucleotides; T data of modified DNA duplexes; molecular
orbitals of 9-methyladenine, 9-methylguanine, 1-methylcytosine,
and methanol; and program of DNA synthesizer (ABI 392)
m
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
m
that, the melting temperatures (T ) were determined at 260 nm using
a UV spectrometer (Pharma Spec UV-1700, Shimadzu) by increasing
the temperature at the rate of 1.0 °C/min.
JA048125H
10896 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004