Proton-block strategy for the synthesis of oligodeoxynucleotides without base protection, capping reaction, and P-N bond cleavage reaction

Article abstract of DOI:10.1021/jo034204k

A new N-unprotected phosphoramidite method called the "proton-block" approach was developed for the chemical synthesis of oligodeoxynucleotides based on the hitherto simplest and rational principle of acid-base reactions. This concept involves protection

Full text of DOI:10.1021/jo034204k

P r oton -Block Str a tegy for th e Syn th esis of Oligod eoxyn u cleotid es
w ith ou t Ba se P r otection , Ca p p in g Rea ction , a n d P -N Bon d
Clea va ge Rea ction
Mitsuo Sekine,*,†,‡ Akihiro Ohkubo,† and Kohji Seio†,‡
Department of Life Science, Tokyo Institute of Technology, CREST,
J ST (J apan Science and Technology Corporation), Nagatsuta, Midoriku, Yokohama 226-8501, J apan
msekine@bio.titech.ac.jp
Received February 17, 2003
A new N-unprotected phosphoramidite method called the “proton-block” approach was developed
for the chemical synthesis of oligodeoxynucleotides based on the hitherto simplest and rational
principle of acid-base reactions. This concept involves protection of the nucleobases of deoxycytidine
and deoxyadenosine with “protons” to convert them to unreactive protonated bases during
condensation by use of promoters having pKa values lower than 2.8. This strategy was applied to
the synthesis of d[CpT] and d[ApT] to check the side reactions associated with the base residues.
In this “proton-block” method, 5-nitrobenzimidazolium triflate (NBT) was found to be the best
promoter, and THF was superior to CH3CN as the solvent so that the concomitant detritylation
due to the inherent acidity of the promoter could be greatly suppressed. Application of this strategy
to the solid-phase synthesis gave d[CpT], d[ApT], d[ApA], d[CpC], and d[GpT] as almost single
peaks in HPLC analysis. Similarly, d[ApApApT] and d[CpCpCpT] were successfully synthesized
without significant side reactions. Finally, d[CpCpCpCpCpCpT] and d[ApApApApApApT] were
obtained as the main products. In the case of a longer oligomer, d[CpApGpTpCpApGpTpCpApGpT],
a mixed solvent of CH3CN-N-methylpyrrolidone (1:1, v/v) was superior to THF so that the desired
oligodeoxynucleotide could be isolated in a satisfactory yield. These results suggest that DNA
synthesis can be carried out simply by using the protonated bases at the oligomer level not only
without base protection but also without the capping reaction and the posttreatment of branched
chains with MeOH-benzimidazolium triflate that previously was requisite. It is concluded that
most of the reactions and solvent effects involved in this strategy can be explained in terms of
simple acid-base reactions. Some problems associated with the previous posttreatment are also
discussed with our own results.
In tr od u ction
us to synthesize sufficiently pure oligodeoxynucleotides
so that custom-made supplies are now available for
laboratory use of DNA fragments. However, the principle
of these approaches has been based on the use of
appropriate protecting groups on the base residues
having amino groups.
Forty-seven years have already passed since Michelson
and Todd succeeded in synthesizing a TpT dimer by a
chemical approach.1 After that, Khorana developed the
phosphodiester method by use of DCC and 2,4,6-triiso-
propylbenzenesulfonyl chloride (TPS).2 In 1967, Letsinger
reported the phosphotriester approach3 and later the
phosphite approach.4 Ultimately, Beaucage and Caruth-
ers devised and developed extensively a practical method
called the phosphoramidite approach.5 Garreg also pro-
posed the H-phosphonate approach.6 Matteucci estab-
lished this approach using pivaloyl chloride as the
condensing reagent.7 These basic studies have enabled
Adamiak reported the synthesis of an RNA heptamer,
CpCpCpt6ApA, of an initiator tRNA anticodon loop in a
synthetic strategy without using the protecting group of
the adenine residue in the phosphotriester method.8
Letsinger showed at an early stage of development of the
(5) (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.
† Tokyo Institute of Technology.
‡ CREST, J ST.
* Corresponding author.
(1) Michelson, M.; Todd, A. R. J . Chem. Soc. 1955, 2632-2638.
(2) (a) Gilham, P. T.; Khorana, H. G. J . Am. Chem. Soc. 1958, 80,
6212-6222. (b) Khorana, H. G. Pure Appl. Chem. 1968, 17, 349-381.
(3) (a) Letsinger, R. L.; Ogilvie, K. K. J . Am. Chem. Soc. 1967, 89,
4801-4803. (b) Letsinger, R. L.; Ogilvie, K. K. J . Am. Chem. Soc. 1970,
91, 3350-3355.
(4) (a) Letsinger, R. L.; Finman, J . L.; Heavner, G. A.; Lunsford, W.
B. J . Am. Chem. Soc. 1975, 97, 3278-3279. (b) Letsinger, R. L.;
Lunsford, W. B. J . Am. Chem. Soc. 1976, 98, 3655-3661. (c) Letsinger,
R. L.; Groody, E. P.; Tanaka, T. J . Am. Chem. Soc. 1982, 104, 6805-
6806.
(6) (a) Garegg, P. J .; Regberg, T.; Stawinski, J .; Stro¨mberg, R. Chem.
Scr. 1985, 25, 280-282. (b) Garegg, P. J .; Lindh, I.; Regberg, T.;
Stawinski, J .; Stro¨mberg, R.; Henrichson, C. Tetrahedron Lett. 1986,
27, 4051-4054. (c) Garegg, P. J .; Lindh, I.; Regbert, T.; Stawinski, J .;
Stro¨mberg, R.; Henrichson, C. Tetrahedron Lett. 1986, 27, 4055-4058.
(d) Stro¨mberg, R.; Stawinski, J . In Current Protocols in Nucleic Acid
Chemistry; Beaucage, S. L., Bergstrom, D. E., Glick, G. D., J ones, R.
A., Eds.; J ohn Wiley & Sons: New York, 2000; p 3.4.1-3.4.11.
(7) (a) Froehler, B. C.; Matteucci, M. D. Tetrahedron Lett. 1986, 27,
469-472. (b) Froehler, B. C.; Ng, P. G.; Matteucci, M. D. Nucleic Acids
Res. 1986, 14, 5399-5407.
10.1021/jo034204k CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/13/2003
5478
J . Org. Chem. 2003, 68, 5478-5492

Synthesis of Oligodeoxynucleotides
phosphite approach that N-unprotected deoxyguanosine
and deoxyadenosine building blocks could be used with-
out protection of their amino groups.9 Several papers
have also reported the synthesis of short (2mer-4mer)
RNA oligomers by use of unprotected adenosine deriva-
tives.10 Actually, later, Gryaznov and Letsinger first
demonstrated the DNA synthesis without using base-
protecting groups and showed insight into the generality
of this possibility.11 After this breakthrough study, J ones
reported the synthesis of a DNA oligomer by the H-
phosphonate method without using base-protecting
groups.12 Later, in 1997 we also showed a successful
synthesis of a DNA 12mer as the main product without
base protection by use of the H-phosphonate approach
in combination with a phosphonium-type of condensing
reagent.13 In this N-unprotected H-phosphonate ap-
proach, we made strenuous efforts to realize selective
internucleotidic bond formation to avoid competitive
phosphitylation on the base residue, especially from a
mechanistic point of view based on HOMO-LUMO inter-
action between the activated phosphorus intermediates
and alcohols. Consequently, it was found that the use of
2-(benzotriazol-1-yloxy)-1,1-dimethyl-2-pyrrolidin-1-yl-
1,3,2-diazaphospholidinium hexafluorophosphate (BOMP)
resulted in rapid and selective internucleotide bond
formation.13a However, we also disclosed that the base-
unprotected H-phosphonate method was accompanied by
a side reaction due to intramolecular attack of the 5′-
hydroxyl group of a 5′-terminal deoxynucleoside on the
3′-phosphate group to produce one base-lacking oligode-
oxynucleotide at every step of the condensation.13d Our
further efforts to improve this approach failed to elimi-
nate this serious side reaction. A year later, Hayakawa
reported a practical procedure of DNA fragments based
on the phosphoramidite method without base protection
by use of an effective promoter, imidazolium triflate
(IMT), for activation of deoxynucleoside-3′-phosphora-
midite building blocks on Tenta gel supports.14a Hayaka-
wa’s method seemed to reach a certain goal in the
synthesis of DNA oligomers without base protection14 but
still required extensive workup of the simultaneously
formed branched chains by treatment with a more
powerful promoter, benzimidazolium triflate (BIT),15 in
methanol at the end of each elongation cycle.14a
In our own studies, however, application of their
method to the solid-phase synthesis of DNA fragments
failed when more widely accessible highly cross-linked
polystyrene16 (HCP) or CPG gel17 was employed as the
polymer support by use of both manual procedures and
automated synthesizers, as discussed in the text.
In this paper, we wish to report a new and simple
principle designed for the DNA synthesis based on an
original approach called the “proton-block“ approach that
allowed elimination of not only base-protecting groups
but also the capping reaction with an acylating reagent
and posttreatment with MeOH throughout the protocol.
To our best knowledge, this simple “proton-block” method
has not been realized to date.18 The aim of our present
study is to synthesize directly DNA fragments without
pre- and posttreatment of the base amino function,
namely, to find 100% selective internucleotidic bond
formation from the beginning. We believe such basic
studies without resort to stereotypical concepts would be
needed for further development of new methods for the
synthesis of artificial DNA molecules having functional
groups involving base-labile groups. In this paper, we
report the details of our basic studies on the proton-block
method directed toward this end.
Resu lts a n d Discu ssion
Sid e Rea ction d u r in g th e P osttr ea tm en t w ith
MeOH-BIT. Since the method described by Hayakawa
and Kataoka15a seemed to be the best in the phosphora-
midite approach without base protection, we extensively
studied it to clarify the reason a similar solid-phase
synthesis by us using the above-mentioned MeOH-BIT
system and HCP resin failed. From these studies, it was
finally concluded that our unexpected result was at-
tributed to the ester exchange of internucleotidic phos-
phite intermediates during the posttreatment with
MeOH-BIT at every step.15a
This side reaction was disclosed as follows: When a
trivalent species of DMTrTp(III)(OCE)T synthesized on an
HCP resin was treated with a 0.5 M solution of BIT in
MeOH at room temperature for 2 min,15a surprisingly,
the oxidation of the resulting mixture with I2 followed
by the successive treatments with 3% trichloroacetic acid
and ammonia resulted in a mixture of T and TpT in a
1:2 peak ratio in HPLC, as shown in Figure 2B. This
result suggested that the internucleotidic bond was
cleaved to a degree of ca. 50%. Contrary to this result,
the usual workup of the same resin without using the
posttreatment gave an almost single peak of TpT in
HPLC, as shown in Figure 2A.
(8) Adamiak, R. W.; Biala, E.; Grzeskowiak, K.; Kierzek, R.; Krasze-
wski, A.; Markiewicz, W. T.; Okupniak, J .; Stawinski, J .; Wiewiorowski,
M. Nucleic Acids Res. 1978, 5, 1889-1905.
(9) Finnan, J . L.; Varshney, A.; Letsinger, R. L. Nucleic Acids Res.
Symp. Ser. 1980, 7, 133-145.
(10) (a) Ogilvie, K. K.; Theriault, N.; Sadana, L. J . Am. Chem. Soc.
1977, 99, 7741-7743. (b) Ogilvie, K. K.; Nemer, M. J . Tetrahedron
Lett. 1981, 22, 2531-2532. (c) Fourrey, J .-L.; Varenne, J . Tetrahedron
Lett. 1985, 26, 2663-2666. (d) Hayakawa, Y.; Nobori, T.; Noyori, R.;
Imai, J . Tetrahedron Lett. 1987, 28, 2623-2626.
(11) (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.
(12) Kung, P.-P.; J ones, R. A. Tetrahedron Lett. 1992, 33, 5869-
5872.
(13) (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. (d) Wada, T.; Honda, F.; Sato, Y.; Sekine, M. Tetrahedron
Lett. 1999, 40, 915-918.
These results implied that serious internucleotidic
bond cleavage apparently occurred during the posttreat-
ment, as evidenced by considerable appearance of thy-
midine at 3 min. It turned out that the phosphite
intermediate underwent easy transesterification with
(16) (a) McCollum, C.; Andrus, A. Tetrahedron Lett. 1991, 32, 4069-
4072. (b) Sproat, B.; Colonna, F.; Mullah, B.; Tsou, D.; Andrus, A.;
Hampel, A.; Vinayak, R. Nucleosides Nucleotides 1995, 14, 255-273.
(17) Ko¨ster, H.; Stumpe, A.; Wolter, A. Tetrahedron Lett. 1983, 24,
747-750.
(18) In connection with our study, Efimov proposed the possibility
of the phosphoramidite approach without base protection: Efimov, V.
A.; Kalinkina, A. L.; Chakhmakhcheva, O. G. Bioorg. Khim. 1996, 22,
149-152.
(14) (a) Hayakawa Y.; Kataoka, M. J . Am. Chem. Soc. 1998, 120,
12395-12401. (b) Hayakawa, Y.; Kawai, R.; Kataoka, M. Eur. J .
Pharm. Sci. 2001, 13, 5-16.
(15) Hayakawa, Y.; Kataoka, M.; Noyori, R. J . Org. Chem. 1996,
61, 7996-7997.
J . Org. Chem, Vol. 68, No. 14, 2003 5479

Sekine et al.
F IGURE 1. Previous N-unprotected phosphoramidite approaches reported by Letsinger11 and Hayakawa.15a These methods require
posttreatment of branched DNAs on the resin with pyridinium hydrochloride and benzimidazolium triflate in the presence of
nucleophiles.
F IGURE 2. Ion-exchange HPLC profiles of the mixtures obtained by condensation of a T-loaded HCP resin with the thymidine
phosphoramidite unit 1 in the presence of NBT. Panel A: The mixture obtained when the posttreatment was not carried out
after the condensation. Panel B: The mixture obtained when the posttreatment with a 0.5 M solution of BIT in MeOH at room
temperature for 2 min was carried out after the condensation. Panel C: The mixture obtained when the posttreatment with a
0.05 M solution of BIT in MeOH at room temperature for 2 min was carried out after the condensation. Solvent system C was
used for HPLC.
SCHEME 1. In ter n u cleotid ic Bon d Clea va ge d u r in g P osttr ea tm en t w ith MeOH-BIT
MeOH under the above conditions. When a 0.05 M
solution of BIT in MeOH was used, the formation of T
was considerably suppressed but an appreciable amount
of thymidine was still formed, as shown in Figure 2C.
The internucleotidic bond cleavage occurred to a degree
of ca. 10% in this case. Therefore, it seems to be difficult
to reproduce the previous results,15a at least when the
HCP resin is used.
A plausible mechanism for the internucleotidic bond
cleavage giving rise to thymidine is shown in Scheme 1.
There are two possible routes to thymidine, as shown in
Paths A and B. It is likely that BIT serves as a catalyst
for activation of the phosphite triester intermediate. On
the basis of these unexpected results, we searched for a
better method in the phosphoramidite approach without
base protection.
5480 J . Org. Chem., Vol. 68, No. 14, 2003

Synthesis of Oligodeoxynucleotides
SCHEME 2a
P r oton -Block Meth od . It is well-known that amino
groups are more reactive than hydroxyl groups toward
electrophilic reagents, but once protonated, their reactiv-
ity is decreased.19 Therefore, we took into account this
simple principle in organic chemistry for regulation of
the reactivity of the nucleobase amino functions by
protonation in the DNA synthesis. The pKa values of the
protonated forms of the four common deoxynucleosides
of T, dC, dA, and dG were reported to be -0.5, 4.3, 3.8,
and 2.5,20 as shown in Figure 3 (see Supporting Informa-
tion). Therefore, the basicity of the bases is ranked in
the following order: dC, dA, dG, and T. Thymidine and
deoxyguanosine exist in the neutral form under the usual
conditions prescribed for the coupling reaction and show
no reactivity in the phosphoramidite approach.5a,11,15a
Compared with these facts, it has been recognized that
the base residues of deoxycytidine and deoxyadenosine
exhibit considerable reactivity toward reactive species of
phosphoramidite reagents activated by promoters in the
phosphoramidite approach.11,15,21
It is expected that deoxycytidine being more basic than
deoxyadenosine is more effectively accessible to proto-
nation so that the nucleophilicity of this protonated
amino group would substantially decrease when a suf-
ficiently acidic activator can be used in the N-unprotected
phosphoramidite approach. Therefore, it was expected
that less basic (∆pKa ) 0.5) deoxyadenosine involved
more risk of side reactions than deoxycytidine since the
possibility of existence of the dissociated (free) adenine
residue is higher than that of the dissociated cytosine
residue in our approach.
The promoters used in the previous methods, such as
5-(p-nitrophenyl)-1H-tetrazole (pKa ) 3.7),22 5-ethylthio-
1H-tetrazole (pKa ) 4.28),23 benzimidazolium triflate
(BIT) (pKa ) 4.5),15a,16 1H-tetrazole (pKa ) 4.8),5a pyri-
dinium hydrochloride (pKa ) 5.1),11,24 N,N-dimethyla-
nilinium hydrochloride (pKa ) 5.15),5a 4,5-dicyanoimida-
zole (pKa ) 5.2),25 and imidazolium triflate (IMT) (pKa )
6.9),15a cannot protonate the base residues of dC and dA.
The use of these promoters has long been rationalized
in terms of avoidance of detritylation of the 5′-O-DMTr
group.26,27 This stereotype has been strongly held in this
field so that promoters having pKa values lower than 3.6
have not been examined to date.
a
Reagents and conditions: (a) activator, THF or CH3CN, rt, 5
min; (b) I2, pyridine-H2O (9:1, v/v), rt, 5 min.
having pKa values lower than 3.6. In our approach, these
new promoters have been designed to protonate inco-
mining N-unprotected deoxyribonucleoside phosphora-
midites without deprotonating the nucleobases of the
growing, solid-phase-linked, DNA oligonucleotide. It is
expected that the unprotected nucleobases of the oligo-
nucleotide are protonated through TCA treatment at
every detrityolation step, and that the activators will not
cause deprotonation like those activators exhibiting pKa
as greater than 4.3. It was ultimately found that, among
all the activators tested, the three reagents of 5-nitroben-
zimidazolium triflate (NBT, pKa ) 2.76), triazolium
triflate (TRT, pKa ) 2.85), and 1-hydroxy-4-nitro-6-
(trifluoromethyl)benzotriazole (NT-HOBt, pKa ) 2.70)28
gave satisfactory results in this new strategy. The former
two are new compounds that were synthesized as non-
hygroscopic fine crystals in this study. When these
promoters are used in the unprotected phosphoramidite
method, the thymine base remains unprotonated, the
guanine base may be partially protonated, and the
adenine and cytosine bases can be selectively protonated,
i.e., protected by the simplest proton, as depicted in the
boxes of Figure 3, which shows the original overall
concept we propose in this paper.
To test the possibility of the proton-block method, we
extensively searched for more than 70 kinds of promoters
(19) Challis, B.; Butler, A. R. In The Chemistry of the Amino Group;
Patai, S., Ed.; Wiley: London, UK, 1968.
(20) (a) Hall, R. H. In The Modified Nucleosides in Nucleic Acids;
Columbia University Press: New York, 1971. (b) Fox, J . J .; Shugar,
D. Biochim. Biophys. Acta 1952, 9. 369-384. (c) Shugar, D.; Fox, J . J .
Biochim. Biophys. Acta 1952, 9, 199-218.
(21) Wada, T.; Moriguchi, T.; Sekine, M. J . Am. Chem. Soc. 1994,
116, 9901-9911.
(22) Froehler, B. C.; Matteucci, M. D. Tetrahedron Lett. 1983, 24,
3171-3174. Pon R. T. Tetrahedron Lett. 1987, 28, 3643-3646.
(23) Wincott, F.; DiRenzo, A.; Shaffer, C.; Grimm, S.; Tracz, D.;
Workman, C.; Sweedler, D.; Gonzalez, C.; Scaringe, S.; Usman, N.
Nucleic Acids Res. 1995, 23, 2677-2684.
(24) Beier, M.; Pfleiderer, W. Helv. Chim. Acta 1999, 82 879-887.
(25) Vargreese, C.; Carter, J .; Yegge, J .; Krivjansky, S.; Settle, A.;
Kropp, E.; Peterson, K.; Pieken, W. Nucleic Acids Res. 1998, 26, 1046-
1050.
To evaluate the efficiency of these promoters, we used
the condensation of 1.2 equiv of the thymidine phos-
phoramidite unit (1) with 3′-O-tBDMS-deoxycytidine (2)29
in acetonitrile or THF in the presence of 2.2 equiv of an
activator for 5 min at room temperature, as shown in
Scheme 2. After the coupling, oxidation was carried out
by using I2 in pyridine-water30 to give the coupling
product 3.
(26) Beaucage, S. L. In Protocols for Oligonucleotides and Analogues;
Agrawal, S., Ed.; Humana: Totowa, NJ , 1993: pp 33-61. Beaucage,
S. L.; Caruthers, M. H. In Current Protocol in Nucleic Acid Chemistry;
J ohn Wiley & Sons Inc.: New York, 2000; p 3.31-3.3.20.
(27) Krotz, A. H.; Klopchin, P.; Walker, K. L.; Srivatsa, G. S.; Cole,
D. L.; Ravikumar, V. T. Tetrahedron Lett. 1997, 38, 3875-3878.
(28) Reese, C. B.; Zhuo, Z. P. J . Chem. Soc., Perkin. Trans. 1 1993,
2291-2301.
(29) Ogilvie, K. K. Can. J . Chem. 1973, 51, 3799-3807.
(30) Usman, N.; Pon, R. T.; Ogilvie, K. K. Tetrahedron Lett. 1985,
26, 4567-4570.
J . Org. Chem, Vol. 68, No. 14, 2003 5481

Sekine et al.
SCHEME 3a
a
Reagents and conditions: (a) t-BuOOH in CH3CN, (b) H2O-CH3CN (1:9, v/v), rt, 1 h, (c) I2 in pyridine-H2O (9:1, v/v), rt, 5 min, (d)
1H-tetrazole in CH3CN, rt, 1 h.
To analyze these reactions by 31P NMR, we synthesized
in advance reference materials 5-8 which were expected
to be formed as byproducts, as shown in Scheme 3. The
phosphoramidate derivative 531 and the hydrolyzed prod-
uct 626 were obtained quantitatively by oxidation and
hydrolysis, respectively, of 1. The N-phosphorylated
species 7 was synthesized in 80% yield by reaction of 1
with 3′,5′-O-bis(tert-butyldimethylsilyl)deoxycytosine 929
in the presence of 1H-tetrazole followed by I2 oxidation.
The phosphodiester derivative 8 was also obtained.
Figure 4-A1 shows that, when 5-(ethylthio)-1H-tetra-
zole (EtSTet)17b,23 was employed as a representative
promoter in the reaction of 1 with 2, the desired product
3 appeared at -1.8 ppm as the main peak in the 31P NMR
spectrum of the mixture. Significant amounts of the
oxidized product 5 and the phosphodiester product 8 were
simultaneously formed. These results reflect the incom-
pleteness of the coupling reaction in the case of EtSTet.
Neither the hydrolyzed product 6 nor the N-phosphoryl-
ated product 4 was detected. It is likely that the once-
hydrolyzed product 6 was completely converted to the
phosphodiester derivative 8 during the I2 treatment.
Benzimidazolium triflate (BIT, pKa ) 4.5)15a,16 has been
reported as the hitherto most powerful activator and is
used as the reagent for the posttreatment of trivalent
phosphoramidite species attached to the nucleobases in
the previous N-unprotected phosphoramidite approach,
as shown in Figure 1.15a As expected, the use of BIT
resulted in complete condensation showing two diaster-
eomeric resonance signals at ca. -2 ppm, as shown in
Figure 4-B1, and the dimer 3 was isolated in 92% yield.
When a new promoter, NBT, was used, the 31P NMR
spectrum (Figure 4-C1) of the mixture showed a set of
diastereomers of 3 along with a peak at -2.1 ppm which
seemed to be the phosphodiester 8. After purification of
the dimer 3, it was obtained in 94% yield. In the case of
NT-HOBt, the 31P NMR spectrum (Figure 4-D1) of the
mixture showed the diastereomeric resonance peaks of
3 with two minor peaks between -1.3 and -1.8 ppm.
When TRT was used, the 31P NMR spectrum (Figure
4-E1) of the crude mixture showed a main peak at -2
ppm and a minor peak at -2.2 ppm. The former seemed
to be an unresolved signal of the diastereomers of 3, and
the latter seemed to be compound 8. After purification,
the dimer was isolated in 95% yield. As shown in these
examples, resolution and chemical shifts of the peaks of
3 and 8 tended to change in each case, suggesting that
the residues of the promoters delicately affect the whole
resonance system. In all cases, however, the isolated
product clearly exhibited only the diastereomeric reso-
nance signals of 3, as shown in Figure 4 (A2-E2). These
reactions were done in THF except for the use of TRT
since TRT was found to be sparingly soluble in THF.
These results suggested that, in the liquid-phase
synthesis of 3 using a small excess amount of 1, the base
modification giving rise to 4 essentially did not occur. It
(31) Wada, T.; Kato, R.; Hata, T. J . Org. Chem. 1991, 56, 1243-
1250.
5482 J . Org. Chem., Vol. 68, No. 14, 2003

Synthesis of Oligodeoxynucleotides
F IGURE 4. 31P NMR spectra of the crude mixture obtained after the reaction of 1 and 2 in the presence of various promoters
and the purified dimer 3: Panel A, 5-(ethylthio)-1H-tetrazole (EtSTet); Panel B, benzimidazolium triflate (BIT); Panel C,
4-nitrobenzimidazolium triflate (NBT), Panel D, 4-nitro-6-(trifluoromethyl)benzotriazole-1-ol (NT-HOBt). Numbers 1 and 2 after
bold alphabets refer to the spectra of the crude mixture and the purified dimer 3, respectively. The figure in parentheses is the
isolated yield.
TABLE 1. P r otocol for th e Syn th esis of DNA Oligom er s
on P olym er Su p p or ts
was found to be the best reagent. Other salts such as
the corresponding chloride, hexafluorophosphate (PF6-),
and nitrate were less effective.
time,
entry manipulation
reagent
min
Syn th esis of Dideoxyn u cleotides on P olym er Su p-
p or ts. Compared with the solution-phase synthesis, the
solid-phase method requires excess deoxynucleoside-3′-
phosphoramidite reagents for condensation. Therefore,
a greater risk of side reactions on the base residues of
dC and dA was expected. Actually, Hayakawa reported
in his phosphoramidite approach without base protection
that the use of 25 equiv of monomer units in the presence
of imidazolium triflate (IBT) resulted in the occurrence
of the base modification of dC and dA to a degree of 8%
and 4%, respectively.15a It was also expected that the loss
of the DMTr group during the condensation becomes
more visible because of the excess use of the phosphora-
midites. Therefore, to examine the side reaction on the
base parts and the detritylation in our proton-block meth-
od, we chose the sequence of d[ApT], d[CpT], d[ApA],
d[CpC], and d[GpT] to synthesize these dimers on a high-
ly cross-linked polystyrene resin.17 The T-, dA-, and dC-
loaded resins were prepared according to the method
described in our previous paper.13b For the synthesis of
these dimers we used the typical protocol involving the
conditions as described in Table 1. This protocol is very
simple so that the total time required for the three
reactions (detritylation, condensation, and oxidation) is
1
2
3
4
washing
detritylation 3% TCA/CH2Cl2
washing CH2Cl2‚THF
condensation phosphoramidite (20 equiv)
promoter (40 equiv for dC or dA)
(20 equiv for dG or T)
CH2Cl2
0.5
1
5
6
7
washing
oxidation
washing
CH3CN
I2/pyridine-H2O (9:1, v/v)
CH3CN
2
is also noted that under the conditions used for the
coupling reaction the elimination of the DMTr group was
not detected on TLC.
Among the promoters tested, several reagents such as
ethyl nicotinate hydrochloride (pKa ) 3.2) and 4-bro-
mobenzoic acid (pKa ) 3.6) showed nonhygroscopic
property and sufficient solubility in organic solvents.
However, no reaction occurred when these reagents were
used as promoters in THF. These results strongly sug-
gested that the nucleophilic property should be involved
in the promoter after the P-N bond activation by an
acidic substance, and some intermediates such as a
phosphorus benzimidazolide species must be formed to
enable the smooth esterification. As the promoter, NBT
J . Org. Chem, Vol. 68, No. 14, 2003 5483

Sekine et al.
3.5 min. A 20-equiv sample of a promoter was used in
the condensation with 20 equiv of the dG or T phosphor-
amidite unit, while 40 equiv of a promoter was used in
the case of the dC or dA phosphoramidite unit (20 equiv).
This is because it is not necessary to protonate the base
residues of dG and T and the basic base residues of dC
and dA should be simultaneously protonated during the
condensation when the P-N bond is activated by the
promoter.
In our proton-block method, it is expected that the
reaction of an adenine base is a little more serious than
that of the cytosine base as far as the N-phosphitylation
is concerned since the former is weaker by a 0.5 pKa unit
than the latter. Therefore, we checked condensation of a
T-loaded resin with the deoxyadenosine phosphoramidite
unit 11 by the action of various promoters.
The crude mixtures obtained after successive treat-
ments with trichloroacetic acid (TCA) and ammonia were
analyzed by reverse-phase HPLC. 1H-Tetrazole is a
popular promoter in the phosphoramidite approach.
Therefore, we examined how significantly the side reac-
tions occurred when this azole was used. Indeed, this
promoter gave byproducts X and X′ at 27 and 29.5 min
in HPLC as shown in Figure 5A. Since the DMTr group
was not eliminated in this case at all, these byproducts
were ascribed to the phosphorylation on the adenine base.
This tendency was also observed in the case of BIT, as
shown in Figure 5B. Considerable amounts of such
byproducts were formed. Even in the case of a mild
activator, IMT, which was used for the N-unprotected
phosphoramidite approach reported by Hayakawa,15a
a
considerable amount of the same byproduct X was
formed, as shown in Figure 5C. Actually, the O-selectivity
of the phosphorylation was estimated to be 77% from the
area ratio of the peaks. To determine the structure of
the byproduct X, this peak was analyzed by MALDI TOF
mass spectroscopy. As a result, this analysis suggested
that X has the molecular weight (M + H, calcd 869.21,
found 869.21) corresponding to a branched trinucleotide
d(ApNHApT). A minor peak of X′ at 34 min in Panel C was
also observed at the region of tetranucleotides in the
anion-exchange HPLC. This peak might be a branched
tetranucleotide d[ApNHApApT] or d[ApApNHApT] although
this product was not characterized.
Contrary to these facts, the use of NT-HOBt having
the pKa value of 2.70 gave a somewhat improved HPLC
profile, as shown in Figure 5D. The formation of a
branched trinucleotide was greatly suppressed. TRT also
exhibited quite selective formation of d[ApT] in acetoni-
trile with generation of a lesser amount of the branched
trinucleotide (1.4% estimated from the HPLC peaks), but
a minor peak (3.6%) at 23.5 min was observed, as shown
in Figure 5E. It was confirmed by comparison with the
authentic sample that this minor peak was d[ApApT],
which was formed owing to the partial detritylation
during the condensation followed by further condensation
of a newly generated 5′-free hydroxyl group with the
excess phosphoramidite unit. In acetonitrile, NBT showed
the same byproduct (3.9%) in Figure 5F. More interest-
ingly, it was found that this reagent in THF underwent
smooth reaction to give the best result essentially without
formation of the base-modified product and the detrityl-
ated species, as shown in Figure 5G. These results clearly
indicated that the NBT reagent was superior to the other
F IGURE 5. The reverse-phase HPLC profiles of crude d[ApT]
obtained after the condensation of a T-loaded polystyrene resin
with deoxyadenosine 3′-O-phosphoramidite building block 10
in various promoters: In each panel, the activator and the
solvent used for the condensation are depicted. System B was
used for HPLC. In the case of Panel G, the isolated yield of
d[ApT] is shown in parentheses. Panel H is an expanded figure
of Panel G.
two. The N-phosphorylated product having 0.6% area and
d[ApApT] having 0.7% area were detected, as shown in
Figure 5H. The dimer of d[ApT] was isolated in 75% and
was well characterized by ESI mass and enzyme analysis
(Experimental Section).
In the proton-block method there is a possibility that
undesired depurination and detritylation might occur
during the condensation. To check this possibility, we
examined the stability of d[ApT] on the polymer support.
After d[ApT] without base protection was synthesized on
HCP by use of NBT as a promoter, the resin was treated
with a 0.2 M solution of NBT in THF for a prolonged time
of 2 h at room temperature. Subsequently, the d[ApT]
mixture was released by treatment with ammonia and
the mixture obtained was analyzed by RP-HPLC (Figure
5484 J . Org. Chem., Vol. 68, No. 14, 2003

Synthesis of Oligodeoxynucleotides
F IGURE 6. Reverse-phase HPLC profiles of the mixtures obtained by treatment of a d[ApT]- or d[ABzT]-loaded HCP resin withNBT
or TCA. Panel A: The mixture obtained when the d[ApT]-loaded HCP resin having dA without the base protecting group was
treated with a 0.2 M solution of NBT in THF for 2 h at room temperature. Panel B: The mixture obtained when the d[ApT]-
loaded HCP resin having dA without the base protecting group was treated with 3% TCA in CH2Cl2 for 2 h at room temperature.
Panel C: The mixture obtained when the d[ABzpT]-loaded HCP resin having N-benzoyl dA was treated with 3% TCA in CH2Cl2
for 2 h at room temperature. Solvent system A was used for HPLC.
SCHEME 4
6A). The area ratio of depurinated byproducts under
these conditions was less than 1%. This fact suggested
that depurination rarely occurred in the proton-block
strategy.
This result was also confirmed by an independent
experiment in the solution phase, using deoxyadenosine
which was treated with a 0.2 M solution of NBT in THF
at room temperature. It was found that dA was es-
sentially stable in 10 h without decomposition.
To demonstrate the higher stability of unprotected dA
than that of N-benzoylated dA on the polymer support,
the stability of d[ApT]-HCP and d[ABzpT]-HCP in a 3%
TCA solution in CH2Cl2 was checked. Consequently, it
turned out serious decomposition occurred to give com-
F IGURE 7. Reverse-phase HPLC profiles of the mixtures
obtained after keeping a DMTr-T10-loaded CPG or HCP resin
plex HPLC peaks when d[ABzpT]-HCP was treated with
3% TCA for 2 h at room temperature, as shown in Figure
6C. On the other hand, it was confirmed that d[ApT]-
HCP slightly (1.6%) decomposed under the same condi-
tions, as shown in Figure 6B. Apparently, the superiority
of the proton-block strategy over the N-protected method
is confirmed from these results, as far as the suppression
of the depurination is concerned.
Since there is a risk of detritylation during the
condensation with NBT, we studied in detail the stability
of the 5′-terminal DMTr group in a growing chain on CPG
and HCP resin in a simpler system which is similar to
that of the proton-block method (Scheme 4). Panels A and
C in Figure 7 show DMTr-T10, which was synthesized on
CPG and HCP resins, respectively, and released by
treatment with ammonia. The detritylated byproduct T10
was observed at 13 min in reverse-phase HPLC. When
DMTr-T10-CPG was treated with a 1:2:1 mixture of 5′-
O-DMTr deoxyadenosine, NBT, and diisopropylamine in
THF for 10 min at room temperature, the loss of the
DMTr group to give T10 was estimated to be 10%, as
shown in Figure 7B. Under similar conditions, DMTr-
T10-HCP was found to be more stable, showing the loss
under conditions similar to those used in the proton-block
method. Panel A: The mixture obtained when DMTr-T10 was
synthesized on a CPG resin. Panel B: The mixture obtained
when the DMTr-T10-loaded CPG resin was treated with NBT
(40 equiv), diisopropylamine (20 equiv), and 5′-O-DMTr-
deoxyadenosine (20 equiv) in THF for 10 min at room tem-
perature. Panel C: The mixture obtained when DMTr-T10 was
synthesized on a HCP resin. Panel D: The mixture obtained
when the DMTr-T10-loaded HCP resin was treated with NBT
(40 equiv), diisopropylamine (20 equiv), and 5′-O-DMTr-
deoxyadenosine (20 equiv) in THF for 10 min at room tem-
peature. Solvent system A was used for HPLC.
of the DMTr group in less than 1%, as shown in Figure
7D. These results suggested that the HCP resin was
more suitable than the CPG resin in the proton-block
method.
Next, the dimer d[CpT] was similarly synthesized by
condensation of a T-loaded resin with deoxycytidine-3′-
phosphoramidite unit 10 in the presence of IMT, NBT,
NT-HOBt, or TRT in THF or CH3CN. When IMT was
used as the activator, d[CpT] appeared as the main peak
at 8 min in the reverse-HPLC chart but a considerable
J . Org. Chem, Vol. 68, No. 14, 2003 5485

Sekine et al.
F IGURE 9. The anion-exchange HPLC profiles of crude
d[TpApT] or d[TpCpT] obtained after condensation of a d[ApT]-
or d[CpT]-loaded polystyrene resin with the thymidine 3′-O-
phosphoramidite building block 1 for 1 min in THF or CH3CN.
Panels A and B show the mixtures obtained in the synthesis
of d[TpApT] by use of IMT and NBT, respectively. Panels C
and D show the mixtures obtained in the synthesis of d[TpCpT]
by use of IMT and NBT, respectively. In each panel, the
activator and the solvent used are depicted. System C was used
for HPLC.
F IGURE 8. The reverse-phase HPLC profiles of crude d[CpT]
obtained after condensation of a T-loaded polystyrene resin
with the deoxycytidine 3′-O-phosphoramidite building block
10 in THF or CH3CN. In each panel, the activator and the
solvent used are depicted. The isolated yield of the dimer after
purification by HPLC is shown in parentheses in the case of
D. System A was used for HPLC.
Interestingly, the use of NBT resulted in more than
99.9% of O-selectivity, as shown in Figure 9B.
amount of byproduct Y was formed, as shown in Figure
8A
Next, the O-selectivity was similarly examined in the
case of d[TCpT]-HCP. The O-selectivity was surprisingly
poor (less than 10%), as shown in Figure 9C when IMT
was used, while with the use of NBT it was excellent
(99.8%), as shown in Figure 9D. These results strongly
suggested the usefulness of our new method.
According to the analysis with anion-exchange HPLC,
the O-selectivity of the phosphorylation was calculated
to be 83% (data not shown). This peak Y was also
analyzed by ESI mass spectroscopy. Consequently, it was
determined to be an N-phosphorylated byproduct,
d[[(dCp)NHCpT]. Among the activators tested, NBT gave the
most promising result so that an almost single peak was
observed in the HPLC chart, as shown in Panels B-D
in Figure 8.
These successful results encouraged us to study other
sequences using NBT. Therefore, we focused on this
reagent for the synthesis of d[ApA] and d[CpC] having
two basic deoxynucleosides. These results are shown in
Figure 10A,B. In each case, an almost single peak was
observed. When this system was applied to the synthesis
of d[GpT], a similar result was obtained, as shown in
Figure 10C. Thus, d[ApA], d[CpC], and d[GpT] could be
isolated in 70%, 78%, and 82% yields, respectively.
Since at the dimer and trimer levels we succeeded in
synthesizing satisfactorily several di- and tri-deoxynucle-
otides, further studies were done by using the sequences
of d[ApApApT] and d[CpCpCpT]. These results are
depicted in Figure 11A,B. In each case, the main peak
was found to be the desired tetradeoxynucleotide. These
oligomers were isolated in satisfactory yields and char-
acterized by MALDI-TOF mass and enzymatic degrada-
tion.
Syn th esis of Lon ger Oligod eoxyn u cleotid es in
th e P r oton -Block Ap p r oa ch . Since with an increase
of the length of the oligodeoxynucleotide chain the pos-
sibility of occurrence of side reactions should be amplified,
we evaluated our proton-block approach at the 7mer le-
vel. For this purpose, the sequences of d[CpCpCpCpCp-
CpT] and d[ApApApApApApT] as the target oligomers
were chosen. In the synthesis of d[CpCpCpCpCpCpT], the
desired product was obtained as the main peak as shown
in Figure 12A. However, many minor peaks in the region
In this case, the O-selectivity of the phosphorylation
was 99.0% and detritylation was not essentially observed,
as shown in Figure 8D, since the cytosine base is more
basic than the adenine base. This dimer d[CpT] was
isolated in 71% yield and was well characterized by ESI
mass and enzyme analysis. Under the conditions used
for the condensation, no byproduct such as d[CpCpT],
which would be derived from the partial detritylation
during the condensation and should appear at 9 min in
reverse-phase HPLC, were detected.
Next, we examined the detailed side reaction under the
conditions where a more straightforward chemoselectiv-
ity of the base amino group vs the 5′-hydroxyl group,
which are involved in an oligonucleotide growing on the
HCP resin, can be estimated. Thus, d[ApT] on an HCP
resin was synthesized by the use of NBT in the proton-
block method. The thymidine 3′-phosporamidite unit was
condensed with the d[ApT]-HCP resin for the synthesis
of d[TpApT] by use of two promoters, i.e., IMT in the
previous method or NBT in the proton-block method.
Consequently, 9% of the expected undesired byproduct
d[TpTpNHApT] was detected in the previous method, as
shown in Figure 9A, so that the O-selectivity was 91%.
5486 J . Org. Chem., Vol. 68, No. 14, 2003

Synthesis of Oligodeoxynucleotides
F IGURE 12. The anion-exchange HPLC profiles of the mix-
tures obtained by use of NBT: Panel A, d[CpCpCpCpCpCpT];
and Panel B, d[ApApApApApApT]. System D was used for
HPLC.
System for Better Con d en sa tion in Oligom er Syn -
th esis. Finally, we applied our proton-block approach to
the synthesis of a 12mer, d(CpApGpTpCpApGpTpCpA-
pGpT). During this synthesis, we encountered a problem
in using the solvent, THF. In this solvent, we could not
obtain satisfactory yields, particularly in the synthesis
of more than 9mer. Therefore, we further searched for
better solvent systems. Consequently, at the oligomer
level, we found that a mixture of CH3CN and N-
methylpyrrolidone (NMP) is much superior to THF. NMP
has a pKa value of -0.71, which is somewhat stronger in
basicity than THF (pKa ca. -2), as shown in Figure 3.
The pKa value of CH3CN is -10, as mentioned before.
Therefore, we can prepare appropriate solvent systems
having suitable basicity within the range from pKa -0.71
to -10 by changing the ratio of the two solvents. It was
ultimately found that a 1:1 mixture of CH3CN and NMP
was effective in this study. Consequently, we were able
to obtain d(CpApGpTpCpApGpTpCpApGpT) in 18% iso-
lated yield. The crude HPLC chart is shown in Figure
13, which shows a main peak corresponding to the
desired 12mer.
F IGURE 10. The reverse-phase HPLC profiles of the mix-
tures obtained after the condensation of dC-, dA-, and T-loaded
polystyrene resins with the deoxyadenosine 3′-O-phosphora-
midite building blocks 10, 11, and 13 respectively, in various
promoters: Panel A, d[ApA], NBT, THF; Panel B, d[CpC],
NBT, THF; and Panel C, d[GpT], NBT, THF. The isolated yield
of each dimer after purification by HPLC is shown in paren-
theses. System A was used for HPLC.
It should be emphasized that without using base
protection, capping reaction, and posttreatment of the
cleavage of concomitant byproducts having the P-N
bond, we were able to obtain the 12mer as the highest
peak in HPLC. To demonstrate the significance of this
result, we also did an independent experiment to syn-
thesize the same 12mer. As mentioned before, Hayakawa
reported his successful DNA 60mer synthesis.15a In this
method, the condensation was carried out by use of
imidazolium triflate (IMT). Therefore, we examined the
synthesis of d[CpApGpTpCpApGpTpCpApGpT] by using
the hitherto best activator, IMT,15a as far as the DNA
synthesis without base protection is concerned. We
checked what happened when the posttreatment with
MeOH-BIT was eliminated in his procedure. Conse-
quently, it turned out that, when the posttreatment was
eliminated, a complex mixture was obtained, as shown
in Figure 14. As seen in Figure 14, a cluster of peaks
such as a mountain was observed.
F IGURE 11. The reverse-phase HPLC profiles of the crude
mixture obtained by use of NBT in THF. The isolated yield of
each tetramer after purification by HPLC is shown in paren-
theses. System B was used for HPLC.
of retention time of more than 20 min were observed.
These complex peaks might be ascribed to the base
modification that might occur at any place with the
consecutive sequence of dC. The isolated yield of this
7mer was 21%. In the synthesis of d[ApApApApApApT],
we were able to obtain a little better yield of 25%, as
shown in Figure 12B. These oligomers were characterized
by enzyme digestion and MALDI-TOF mass spectroscopy.
Since we eliminated both capping and post-P-N bond
cleavage reactions from the current DNA protocol, such
side reactions arose on a visible scale at the level of longer
DNA fragments. Nevertheless, it should be emphasized
that it is possible to obtain these oligomers as main
products without any posttreatment.
This result implied that multicomplex branched DNA
chains were apparently formed, as evidenced by the
presence of many peaks around retention times of 27-
50 min. This is also supported by an early experiment
Syn th esis of d [CAGTCAGTCAGT] in th e P r oton -
Block Ap p r oa ch a n d th e Ch oice of a Mixed Solven t
J . Org. Chem, Vol. 68, No. 14, 2003 5487

Sekine et al.
compared. However, an increase of branched chains on
the 12mer increases simultaneously the number of phos-
phodiester linkages on the branched chains so that the
13-fold difference sufficiently explains the outlooks of
Figures 13 and 14, as evidenced by our own theoretical
simulation (data not shown). These results indicated that
our new strategy can suppress considerably N-phospho-
rylation compared with the previous methods in the
synthesis of oligodeoxynucluotides with mixed sequen-
ces.
Sid e Rea ction s Associa ted w ith th e P r oton -Block
Ap p r oa ch . It was a matter of concern that the use of
more acidic activators would cause depurination of
deoxyadenosine and deoxyguanosine.26a However, we
could not observe this kind of problem during this study.
It is known that the depurination rate of N-unprotected
deoxyadenosine is slower than that of N-benzoyldeoxy-
adenosine.32,33 Therefore, it is likely that, since depuri-
nation occurs via a diprotonated species, it is more
difficult to protonate the position 7 of the protonated
deoxyadenosine form than that of N-benzoyldeoxyad-
enosine by the action of NBT.
F IGURE 13. The anion-exchange HPLC profiles of the crude
d[CAGTCAGTCAGT] obtained by use of NBT in CH3CN-NMP
(1:1, v/v). System D was used for HPLC.
Since our new activators ranked between deoxygua-
nosine (pKa ) 2.5) and deoxyadenosine (pKa ) 3.8) in the
pKa list, as shown in Figure 3, deoxyguanosine is partially
protonated theoretically by NBT, while deoxyadenosine
is mostly protonated by these activators and deoxycyti-
dine is completely protected. However, the growing chain
of oligodeoxynucleotides on the polymer was treated by
the action of TCA for removal of the DMTr group. The
TCA-treated resin was washed with dichloromethane.
Therefore, all three basic deoxynucleosides of dG, dA, and
dC formed salts with TCA (pKa ) 0.52) after the TCA
treatment. As a matter of fact, this protonated resin was
allowed to react with phosphoramidite reagents in the
presence of NBT. This manipulation resulted in a com-
plex mixture in the reaction system. It is expected that
the TCA salts of dG, dA, and dC come to acquire the
ability as activators of the P-N bond of the phosphora-
midites but the efficiency of the activation of such TCA
salts has yet to be evaluated in this study. This situation
becomes more complex with an increase of the length of
oligodeoxynucleotides since the relative ratio of the TCA
salts to NBT would increase. When a DNA 10mer
d[(Xp)9T] containing dA, and/or dC as X is synthesized
in the proton-block method, it is expected that, at the
last coupling, nine TCA salts are generated on the
polymer. Usually, we used 40 equiv of an activator and
20 equiv of a phosphoramidite building unit. Probably,
this side effect is one reason the HPLC pattern becomes
complicated at the oligomer level.
F IGURE 14. The anion-exchange HPLC profiles of the
mixture obtained by use of IMT as the activator in CH3CN in
an attempt to obtain d[CAGTCAGTCAGT]. Solvent D was
used for HPLC.
F IGURE 15. 31P NMR spectrum of crude d[CAGT]3. Panel
A: The profile of d[CAGT]3 by use of IMT; Panel B: The profile
of d[CAGT]3 by use of NBT.
at the dimer synthesis, as shown in Figure 5C. Therefore,
as shown in Figure 13, our new reagent, NBT, has proved
to work very well as an effective activator in the proton-
block method.
In our approach, the nucleobases are initially proto-
nated by treatment with TCA and these protonated bases
are preserved during the condensation, since the diiso-
propylamine generated by condensation is effectively
neutralized by triflic acid and, simultaneously, the free,
less basic activator (5-nitrobenzimidazole pKa ) 2.76)
that cannot deprotonate the nucleobases is generated. On
the other hand, the previous method15a with IMT pro-
duces imidazole (pKa ) 6.9), which is more basic than
The O-selectivity of Figure 13 or Figure 14 was
evaluated by 31P NMR. Figure 15A shows the NMR
profile of the crude mixture involving d(CAGT)3, which
was obtained by use of IMT, and Figure 15B shows that
obtained by use of NBT. There were broad signals of
N-phosphorylated byproducts at -4 to -6 ppm in both
cases, but the O-selectivity in the proton-block method
was 13-fold over that calculated in the previous method.
At a glance, it seems that this difference is not significant
when the dramatic change between Figures 13 and 14 is
(32) York J . L. J . Org. Chem. 1981, 46, 2171-2173.
(33) Schaller, H.; Khorana, H. G. J . Am. Chem. Soc. 1963, 85, 3828-
3835. Moon, M. W.; Nishimura, S.; Khorana, H. G. Biochemistry 1966,
5, 937-945.
5488 J . Org. Chem., Vol. 68, No. 14, 2003

Synthesis of Oligodeoxynucleotides
SCHEME 5
of THF enabled us to use more acidic activators such as
NBT for activation of phosphoramidite building units.
The difference in pKa between NBT and the 5′-trityl
ether is expected to be ca. 5 units. This difference allows
the concomitant detritylation in CH3CN as exemplified
by Figure 5F, but selective internucleotidic bond forma-
tion can be done in a more basic solvent of THF, as shown
in Figure 5G (THF).
In addition, we checked the stability of the DMTr group
on a DMTrT-HCP resin (0.25 µmol) in a 0.1 M solution
of NBT (50 µmol) in CH3CN, THF, and CH3CN-NMP
(1:1, v/v) at 25 °C to evaluate the solvent effect. Conse-
quently, it was found that the amounts of the DMTr
group released from the resins after 5 min were 40%,
26%, and 2% for CH3CN, THF, and CH3CN-NMP (1:1,
v/v), respectively. These results clearly suggested that
the detritylation can be suppressed more effectively by
use of a more basic solvent system. It should also be noted
that under the actual conditions used for the proton-block
method, the actual loss of the DMTr group can be
expected to be far less than those obtained in the above
experiments since essentially basic phosphoramidite
units were not added.
the adenine and cytosine bases having pKa values of 3.8
and 4.2, respectively, so that the side reactions on the
base moieties cannot be avoided. As a matter of course,
the use of IMT resulted in considerable amounts of
N-phosphorylated byproducts, as evidenced by Figures
5C, 8A, 9A, and 9C as well as Figure 14.
In the whole reaction system of our proton-block
method, the nitrogen of the phosphoramidite seems to
be the most basic site. However, this basicity is not
known, because the phosphoramidite building unit is too
unstable to allow examination of this fundamental phys-
icochemical property. Since the phosphoramidite bond
can be activated by IMT (pKa ) 6.9), the pKa value would
be at least more than 7.
Su p er ior ity of THF or CH3CN-NMP a s th e Sol-
ven t in Con d en sa tion . In our study we found that THF
is superior to CH3CN as the solvent as far as avoidance
of the loss of the DMTr group during condensation is
concerned. Previous activators, such as 1H-tetrazole, BIT,
and 5-(ethylthio)-1H-tetrazole, usually work well in CH3-
CN. We also observed that these reagents having pKa
values higher than 3.7 become less effective in THF than
in CH3CN as far as the reaction rate is concerned. The
reason there were significant differences both in detri-
tylation during the condensation and in the reaction rate
of condensation between CH3CN and THF is apparently
explained in terms of the basicity of the solvents used.
CH3CN and THF have pKa values34 of ca. -10 and ca.
-2, respectively, showing there is a difference of 108 in
the dissociation constant between the protonated forms
of CH3CN and THF. The following data35 are available
for discussion about the evaluation of the basicity of the
5′-oxygen: CH3C(O)OH, pKa ) 4.79; p-MeO-C6H4-CH2C-
(O)OH, pKa ) 4.45; and C6H5-CH2C(O)OH, pKa ) 4.30.
From these data, the substituents of phenyl and p-
methoxyphenyl are generally recognized as electron-
withdrawing groups. In addition, it is also expected that
the sugar residue serves as an electron-withdrawing
group since there are â- and γ-oxygens at the 4′- and 3′
-positions that can abstract electrons. However, once this
ethereal 5′-oxigen is protonated, the ether bond cleavage
more easily occurs than that of the protonated THF since
a much more stable trityl cation is generated. Therefore,
a more basic solvent, THF, would serve as a buffer site
more accessible to protonation than the 5′-ether site so
that the risk of direct protonation of the DMTr ether bond
causing the detritylation would decrease. Thus, the choice
Therefore, it might be ideal if we can use a little more
basic solvent than THF as far as the avoidance of the
detritylation is concerned. In this manner, the 1:1
mixture of CH3CN-NMP used for the synthesis of the
DNA 12mer is the better choice of the solvent system.
Con clu sion
The present method evolved from an idea that most of
the important items (the P-N bond activation, solvent
effects, detritylation, and related side reactions) involved
in the DNA synthesis can all be recognized as very simple
“acid-base reactions”. This new concept would stimulate
a new possibility of DNA synthesis without base protec-
tion. Although visible side reactions were observed at the
level of longer DNA fragments, this proton-block method
would provide new insight into the synthesis of artifi-
cially modified DNA oligomers and extremely base-
sensitive functional groups. When the proton-block method
is employed for the synthesis of such fragile modified
DNA oligomers, it is necessary to use a linker that can
be removed under neutral conditions in place of the
succinate linker. Quite recently, we also developed a
useful silyl-type linker capable of removal by treatment
with Bu4NF under neutral conditions.36 Therefore, the
combined use of the present proton-block method and this
linker would provide a new tool for the synthesis of base-
sensitive DNA oligomers.
We also reported previously the preparation of HCP
resin-linked 3′-terminal deoxynucleotide derivatives, which
must be required for the general synthesis of oligodeoxy-
nucleotides having dA, dG, or dC at the 3′-terminal site.
These polymer-linked deoxynucleoside derivatives have
two DMTr groups at the 5′- and N-positions. Therefore,
it will be possible to synthesize any sequences without
basic conditions such as ammonia by use of the proton-
block method, the silyl-type linker, and these polymers.
However, to improve the proton-block approach, ex-
tensive work to find more effective activators and solvents
(34) The list of the basicity of common solvents is now available from
http://mail.chor.unipd.it/wisor/lecture-notes/acidbase.pdf (Dr. A. Bag-
no’s lab, Centro CNR, Padova, Italy).
(35) The pKa values were calculated by use of Advanced Chemistry
Development (ACD) Software Solaris V4.76 (copyright 1994-2003
ACD) in SciFinder 2002.
(36) Kobori, A.; Miyata, K.; Ushioda, M.; Seio, K.; Sekine, M. Chem.
Lett. 2002, 16-17.
J . Org. Chem, Vol. 68, No. 14, 2003 5489

Sekine et al.
H2O (3 × 1 mL). The filtrate was evaporated to dryness under
reduced pressure. The residue was dissolved in water and
analyzed by anion-exchange HPLC.
has yet to be done. Further studies are now underway
in this direction.
5-Nitr oben zim id a zoliu m Tr ifla te (NBT). To a solution
of 5-Nitrobenzimidazole (7.1 g, 44 mmol) in MeOH-ether (1:
1, v/v, 50 mL) was added dropwise triflic acid (3.9 mL, 44
mmol) with vigorous stirring at 0 °C. After being stirred for
10 min, the mixture was poured into ether (200 mL). The
resulting precipitates were collected by filtration and washed
with ether (300 mL) to give the triflate (13.2 g, 96%): mp 138-
140 °C (ethyl acetate); 1H NMR (DMSO) δ 7.97 (d, 1H, J )
8.9 Hz), 8.30-8.34 (m, 1H), 8.63 (d, 1H, J ) 2.2 Hz), 9.51 (s,
1H); 13C NMR (DMSO) δ 111.3, 113.4, 115.4, 118.2, 120.7,
122.9, 127.6, 130.9, 135.0, 144.7, 145.0. Anal. Calcd for
C8H6F3N3O5S: C, 30.68; H, 1.93; N, 13.42; F, 18.20. Found:
C, 30.60; H, 1.91; N, 13.44; F, 18.28.
Tr ia zoliu m Tr ifla te (TRT). This compound was synthe-
sized by use of 1,2,4-triazole (3.0 g 44 mmol) in the same
method as described above. The yield was 9.4 g (98%): mp
162-164 °C (ethyl acetate-acetonitrile); 1H NMR (DMSO) δ
9.31 (s, 1H), 9.33 (s, 1H); 13C NMR (DMSO) δ 113.4, 118.1,
122.8, 127.6, 142.3. Anal. Calcd for C3H4F3N3O3S: C, 16.4; H,
1.84; N, 19.17; F, 26.01; S, 14.63. Found: C, 16.33; H, 1.75; N,
19.35; F, 25.82; S, 14.30.
Eth yl Nicotin a te Hyd r och lor id e. To a solution of ethyl
nicotinate (27.8 g, 183 mmol) in ether (100 mL) was added
dropwise hydrochloride (4.0 M solution in dioxane, 45.8 mL,
183 mmol) with vigorous stirring at 0 °C. After being stirred
for 10 min, the mixture was poured into ether (500 mL). The
resulting precipitates were collected by filtration and washed
with ether (800 mL) to give the triflate (30 g, 87%): mp 122-
124 °C (ethyl acetate); 1H NMR (DMSO) δ 1.34 (t, 3H, J ) 7.0
Hz), 4.37 (dd, 2H, J ) 4.9 Hz, J ) 14.0 Hz), 7.86 (dd, 1H, J )
5.3 Hz, J ) 8.0 Hz), 8.62 (dd, 1H, J ) 3.4 Hz, J ) 8.0 Hz),
8.97 (dd, 1H, J ) 1.6 Hz, J ) 5.4 Hz), 9.17 (d, 1H, J ) 1.6 Hz),
11.94 (br s, 1H); 13C NMR (CDCl3) δ 14.1, 62.3, 127.3 128.5,
143.1, 144.3, 146.2, 162.2. Anal. Calcd for C8H6F3N3O5S: C,
51.21; H, 5.37; N, 7.47. Found: C, 51.13; H, 5.37; N, 7.43.
Syn th esis of th e N-Un p r otected d [Tp C] Der iva tive 3.
A mixture of the thymidine 3′-O-phosphoramidite derivative29
1 (300 mg, 0.403 mmol) and 3′-O-(tert-butyldimethylsilyl)-
deoxycytidine 2 (116 mg, 0.336 mmol) was rendered anhydrous
by repeated coevaporation successively with dry pyridine (×3),
dry toluene (×3), and dry CH2Cl2 (×3), and finally dissolved
in dry THF (3 mL). To the mixture was added an appropriate
activator (0.672 mmol). After the mixture was stirred at room
temperature for 1 h, a 1 M solution of I2 (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 CHCl3 (50 mL) and aqueous 5% Na2S2O3 (30 mL).
The organic phase was collected, washed twice with aqueous
5% Na2S2O3 (30 mL), filtered, dried over Na2SO4, and evapo-
rated under reduced pressure. The residue was chromato-
graphed on a column of silica gel (10 g) with 1% pyridine
containing hexane-CHCl3 (50:50-0:100, v/v) and then 1%
pyridine containing CHCl3-MeOH (100:0-97:3, v/v) to give
the fractions containing 3. The fractions were collected,
evaporated under reduced pressure, and finally evaporated by
repeated coevaporation three times each with toluene and CH2-
Cl2 to remove the last traces of pyridine to give the title
compound 3 (NBI, 315 mg, 94%; NT-HOBt, 280 mg, 83%; TRT,
319 mg, 95%; BIT, 310 mg, 92%; 5-(ethylthio)-1H-tetrazole,
228 mg, 68%): 1H NMR (CDCl3) δ 0.06 (s, 6H), 0.86 (s, 9H),
1.39 (s, 3H), 2.10-2.81 (m, 6H), 3.32-3.58 (m, 2H), 3.78 (s,
6H), 3.82-4.02 (m, 1H), 4.10-4.38 (m, 5H), 5.10-5.22 (m, 2H),
6.15-6.34 (m 3H), 6.76-6.82 (m, 4H), 7.23-7.33 (m, 9H), 7.46
(s, 1H), 7.71 (d, 1H, J ) 6.8 Hz), 9.10-9.55 (br, s, 2H); 31P
NMR (CDCl3) δ -1.58, -1.66; 13C NMR (CDCl3) δ -4.8, -4.6,
11.8, 17.9, 19.6, 19.7, 19.8, 25.7, 39.0, 41.2, 55.3, 62.5, 62.6,
63.3, 70.4, 77.2, 79.9, 84.2, 94.9, 111.8, 113.2, 116.2, 116.4,
123.6, 127.1, 127.9, 128.0, 128.9, 129.9, 134.7, 134.8, 135.8,
140.3, 143.8, 149.6, 150.6, 150.7, 155.3, 158.6, 163.7, 163.8,
Exp er im en ta l Section
Gen er a l Rem a r k s. 1H, 13C, and 31P NMR spectra were
recorded at 270, 68, and 109 MHz, respectively. The chemical
shifts were measured from tetramethylsilane or DSS for 1H
NMR spectra, CDCl3 (77 ppm) or DSS (0 ppm) for 13C NMR
spectra, and 85% phosphoric acid (0 ppm) for 31P NMR spectra.
UV spectra were recorded on a U-2000 spectrometer. CD
spectra were recorded on a J -500 C spectrometer with a 0.5-
cm cell. TLC was performed by the use of Kieselgel 60-F-254
(0.25 mm). Column chromatography 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. Reversed-
phase column chromatography was performed by the use of
µBondapak C-18 silica gel (prep S-500, Waters). Reversed-
phase HPLC was performed with the following systems.
System A:, A Waters Aliance system was used with an Waters
3D UV detector and a µBondasphere 5 µm C18 100 Å column
(3.9 × 150 mm) at 50 °C with a linear gradient (0-10%) of
CH3CN in 0.1 M NH4OAc, pH 7.0, at a flow rate of 1.0 mL/
min for 30 min. System B: A linear gradient (0-30%) of CH3-
CN in 0.1 M NH4OAc was used. The other conditions are the
same as those of System A. System C: An anion-exchange
HPLC was done on a Waters Aliance system with a Waters
3D UV detector and a Waters Spherisorb S5 SAX analytical
column (4.6 × 250 mm). A linear gradient (0-30%) of solvent
I (20% CH3CN in 0.5 M potassium phosphate) in solvent II
(20% CH3CN in 0.005M potassium phosphate) was used at 50
°C at a rate of 1.0 mL/min for 30 min. System D: An 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
III (1 M NaCl in 25 mM phosphate buffer (pH 6.0)) in solvent
IV (25 mM phosphate buffer (pH 6.0)) was used at 50 °C at a
flow rate of 1.0 mL/min for 45 min. If the HPLC analysis after
45 min is necessary, a linear gradient (60-100%) of Solvent
III in Solvent IV was further used at 50 °C at a flow rate of
1.0 mL/min for an additional 7 min. Gel permeation was
performed by use of Sephadex G-10 or G-15 after swelling with
sterilized water, and elution was done with sterilized water.
Pyridine was distilled two times from p-toluenesulfonyl chlo-
ride and from calcium hydride and then stored over 4A
molecular sieves. MALDI-TOF mass was performed by use of
Voyager RP (PerSeptive Biosystems Inc.). Highly cross-linked
polystyrene was purchased from Perkin-Elmer, ABI. NT-HOBt
was synthesized according to the literature method.28 The
thymidine-3′-O-phosphoramidite was prepared according to
the standard method.37 The deoxycytidine-, dexyadenosine-,
and deoxyguanosine-O-3′-phosphoramidite derivatives were
synthesized by the method reported by Gryaznov and Letsing-
er.11 Elemental analyses were performed by the Microanalyti-
cal Laboratory, Tokyo Institute of Technology at Nagatsuta.
Th e Sta bility of In ter n u cleotid ic P h osp h ite Tr iester
Lin k a ge on HCP Du r in g P osttr ea tm en t w ith MeOH-
BIT. After condensation of a T-loaded HCP resin (1 µmol) with
thymidine phosphoramidite unit 1 in the general procedure,
the resin was treated with a 0.5 or 0.05 M solution of BIT in
MeOH (400 µL) for 2 min. After the subsequent oxidation with
I2 in pyridine-H2O (9:1, v/v, 500 µL) for 2 min, the dimer was
deprotected and released from the polymer support by treat-
ment with concentrated aq NH3 (500 µL) for 40 min. The
polymer support was removed by filtration and washed with
(37) Sinha, N. D.; Biernat, J .; McManus, J .; Ko¨ster, H. Nucleic Acids
Res. 1984, 12, 4539-4557. For the synthesis of 1, the phosphitylating
reagent of chloro(2-cyanoethoxy)(diisopropylamino)phosphine was pre-
pared by: Tanimura, H.; Maeda, M.; Fukazawa, T.; Sekine, M.; Hata,
T. Nucleic Acids Res. 1989, 17, 8135-8147.
5490 J . Org. Chem., Vol. 68, No. 14, 2003

Synthesis of Oligodeoxynucleotides
165.5; MS m/z calcd for M + H 1001.3883, found 1001.3887.
Anal. Calcd for C40H49N4O9P‚2H2O: C, 56.74; H, 6.31; N, 8.10.
Found: C, 56.95; H, 6.04; N, 7.90.
84.3, 84.5, 84.6, 85.4, 86.9, 87.7, 111.2, 113.1, 116.5, 126.9,
127.7, 129.8, 134.9, 135.0, 135.2, 140.2, 143.8, 143.9, 150.2,
150.3, 158.4, 159.9, 16.37; MS m/z calcd for M + H 1115.4747,
found 1115.4748. Anal. Calcd for C55H75N6O13PSi2: C, 59.33;
H, 6.78; N, 7.53. Found: C, 59.80; H, 7.00; N, 7.30.
Cya n oeth yl 5′-O-(4,4′-Dim eth oxytr ityl)th ym id in e-3′-yl
N-Diisop r op ylp h osp h or a m id a te (5). To a solution of the
thymidine phosphoramidite unit 1 (470 mg, 0.63 mmol) in CH3-
CN (5 mL) was added t-BuOOH (5 M solution in decane, 0.5
mL, 2.5 mmol). After being stirred at room temperature for
10 min, the mixture was diluted with CHCl3. The CHCl3 layer
was washed three times with brine, dried over Na2SO4,
filtered, and evaporated under reduced pressure. The residue
was chromatographed on a column of silica gel (10 g) with
hexane-CHCl3 (50:50-0:100, v/v) containing 1% pyridine and
then with MeOH-CHCl3 (0:100-3:97, v/v) containing 1%
pyridine to give the title compound (468 mg, 80%): 1H NMR
(CDCl3) δ 1.09-1.35 (m, 15H), 2.33-2.72 (m, 4H), 3.34-3.50
(m, 4H), 3.77 (s, 6H), 4.00-4.29 (m, 3H), 5.03 (br s, 1H), 6.43
(dd, 1H, J ) 5.4 Hz, J ) 8.9 Hz), 6.81 (d, 4H, J ) 8.9 Hz),
7.13-7.59 (m, 10H), 8.66 (d, 1H, J ) 8.6 Hz); 31P NMR (CDCl3)
δ 8.03, 8.33; 13C NMR (CDCl3) δ 11.7, 19.6, 39.4, 46.3, 46.4,
63.4, 77.3, 84.3, 84.4, 87.0, 87.1, 111.5, 111.6, 113.1, 113.2,
126.9, 127.1, 127.6, 127.9, 128.1, 128.9, 129.0, 130.0, 134.8,
134.9, 135.0, 135.3, 139.3, 143.8, 143.9, 150.2, 158.4, 158.6,
163.4; MS m/z calcd for M + H 761.3505, found 761.3317. Anal.
Calcd for C40H49N4O9P‚1/2H2O: C, 62.41; H, 6.55; N, 7.27.
Found: C, 62.66; H, 6.62; N, 7.14.
Cya n oeth yl 5′-O-(4,4′-Dim eth oxytr ityl)th ym id in e-3′-yl
P h osp h on a te (6). To a solution of the thymidine phosphora-
midite unit 1 (470 mg, 0.63 mmol) in CH3CN-H2O (9:1, v/v, 3
mL) was added 1-H-tetrazole (88 mg, 1.26 mmol). After being
stirred at room temperature for 10 min, the mixture was
diluted with CHCl3. The CHCl3 layer was washed three times
with brine, dried over Na2SO4, filtered, and evaporated under
reduced pressure. The residue was purified by flash chroma-
tography on silica gel to give the title compound (354 mg,
85%): 1H NMR (CDCl3) δ 1.34 (s, 3H), 2.38-2.70 (m, 4H),
3.30-3.48 (m, 2H), 3.69 (s, 6H), 4.10-4.21 (m, 3H), 5.17 (s,
1H), 6.36 (dd, 1H, J ) 5.4 Hz, J ) 8.1 Hz), 6.76 (d, 4H, J )
8.6 Hz), 6.82 (d, 1H, J ) 725.0 Hz), 7.06-7.45 (m, 10H), 9.77
(d, 1H, J ) 11.3 Hz); 31P NMR (CDCl3) δ 7.57, 7.60; 13C NMR
(CDCl3) δ 11.9, 19.8, 19.9, 20.1, 39.2, 55.4, 60.5, 60.6, 60.7,
63.1, 77.6, 84.3, 87.3, 111.7, 113.3, 116.5, 123.9, 125.3, 127.2,
128.0, 130.1, 135.0, 135.2, 149.2, 150.7, 150.8, 158.7, 164.1;
MS m/z calcd for M + Na 684.2086, found 684.2135. Anal.
Calcd for C34H36N3O9P: C, 61.72; H, 5.72; N, 6.29. Found: C,
61.73; H, 5.78; N, 6.35.
4-N-[5′-O-(4,4′-Dim eth oxytr ityl)th ym id in e-3′-yl](2-cya -
n oeth yl)p h osp h or yl-3′,5′-O-bis(ter t-bu tyld im eth ylsilyl)-
d eoxycytid in e (7). A mixture of the thymidine 3′-O-phos-
phoramidite unit 1 (500 mg, 0.67 mmol) and 3′,5′-O-bis(tert-
butyldimethylsilyl)deoxycytidine 9 (204 mg, 0.449 mmol) was
rendered anhydrous by repeated coevaporation successively
three times each with dry pyridine, dry toluene, and dry CH2-
Cl2 and finally dissolved in dry CH3CN (5 mL). To the solution
was added 1H-tetrazole (70 mg, 0.674 mmol). After the mixture
was stirred at room temperature for 1 h, a 1 M solution of I2
in pyridine-H2O (9:1, v/v, 3 mL) was added to the mixture.
After being stirred at room temperature for 2 min, the mixture
was diluted with CHCl3. The CHCl3 layer was washed three
times with aqueous 5% Na2S2O3, dried over Na2SO4, filtered,
and evaporated under reduced pressure. The residue was
chromatographed on a column of silica gel (10 g) with hexane-
CHCl3 (50:50-0:100, v/v) containing 1% pyridine and then
with MeOH-CHCl3 (0:100-3:97, v/v) containing 1% pyridine
to give the title compound (400 mg, 80%): 1H NMR (CDCl3) δ
0.07 (s, 6H), 0.10 (s, 6H), 0.79 (s, 9H), 0.82 (s, 9H), 1.39 (s,
3H), 2.10-2.81(m, 6H), 3.32-3.58 (m, 2H), 3.78 (s, 6H), 3.82-
4.02 (m, 1H), 4.10-4.38 (m, 5H), 5.10-5.22 (m, 2H), 6.15-
6.34 (m, 3H), 6.76-6.82 (m, 4H), 7.23-7.33 (m, 9H), 7.46 (s,
1H), 7.71 (m, 1H); 31P NMR (CDCl3) δ 5.28, 5.16; 13C NMR
(CDCl3) δ -5.5, -5.4, -4.9, -4.6, 11.6, 17.9, 18.3, 19.5, 19.6,
25.6, 25.8, 39.1, 41.8, 55.2, 60.9, 61.0, 62.1, 63.2, 70.8, 76.9,
2-Cya n oeth yl 5′-O-(4,4′-Dim eth oxytr ityl)th ym id in e-3′-
yl P h osp h a te (8). To a solution of compound 6 (260 mg, 0.39
mmol) in pyridine-H2O (9:1, v/v, 3 mL) was added a 0.1 M
solution of I2 in pyridine-H2O (9:1, v/v, 20 mL). After being
stirred at room temperature for 2 min, the mixture was diluted
with CHCl3. The CHCl3 layer was washed three times with
aqueous 5% Na2S2O3, dried over Na2SO4, filtered, and evapo-
rated under reduced pressure to give 8 (300 mg 99%): 1H NMR
(CDCl3) δ 1.28-1.45 (m, 12H), 2.17-2.67 (m, 4H), 3.04 (dd,
1H, J ) 7.29 Hz, J ) 14.6 Hz), 3.38 (d, 1H, J ) 8.91 Hz), 3.49
(d, 1H, J ) 7.83 Hz), 3.78 (s, 6H), 3.98 (s, 2H), 4.29 (s, 1H),
4.97 (s, 1H), 6.42 (s, 1H), 6.83 (d, 4H, J ) 8.91 Hz), 7.23-7.63
(m, 10H), 8.52 (br s, 1H), 12.23 (br s, 1H); 31P NMR (CDCl3) δ
-1.17; 13C NMR (CDCl3) δ 8.6, 11.6, 19.8, 19.9, 45.6, 55.1, 60.1,
77.3, 84.4, 86.8, 110.9, 112.8, 113.0, 17.7, 125.0, 126.9, 127.7,
128.0, 128.8, 129.9, 135.1, 135.2, 135.3, 135.4, 144.0, 158.4,
163.9. Anal. Calcd for C40H51N4O10P‚1H2O: C, 60.29; H, 6.71;
N, 7.01. Found: C, 60.39; H, 6.97; N, 6.52.
Typ ica l P r oced u r e for Solid -P h a se Syn th esis. Each
cycle of chain elongation consisted of detritylation (3% trichlo-
roacetic acid in CH2Cl2, 2 mL, 1 min), washing [CH2Cl2 (1 mL
× 3), THF (1 mL × 3)], coupling [an appropriate phosphora-
midite unit (20 µmol) in THF (200 µL), NBT (12.5 mg, 40 µmol)
in THF (200 µL), 1 min], and washing (THF (1 mL × 3)),
oxidation (0.1 M I2, pyridine-H2O (9/1, v/v), 2 min), and
washing [pyridine (1 mL × 3), CH3CN (1 mL × 3), CH2Cl2 (1
mL × 3)]. Generally, the average yield per cycle was estimated
to be 97-99% by the DMTr cation assay. After chain elonga-
tion, the DMTr group was removed by treatment with 3%
trichloroacetic acid in CH2Cl2 (2 mL) for 1 min, and the resin
was washed with CH2Cl2 (1 mL × 3) and CH3CN (1 mL × 3).
The oligomer was deprotected and released from the polymer
support by treatment with concentrated aq NH3 (500 µL) for
40 min. The polymer support was removed by filtration and
washed with CH3CN (3 × 1 mL). The filtrate was evaporated
and purified by reversed-phase HPLC or anion-exchange
HPLC.
Th e Sta bility of th e DMTr Gr ou p on P olym er Su p -
p or ts u n d er Acid ic Con d ition s Sim ila r to Th ose Used
for t h e Con d en sa t ion of t h e P r ot on -Block Met h od .
DMTr-ON T10 (1 µmol) on a HCP resin or CPG resin was
synthesized by use of an ABI 392 DNA/RNA synthesizer. The
resin was treated with a THF solution (400 µL) of NBT (12.4
mg, 40 µmol), diisopropylamine (2.8 µL, 20 µmol), and 5′-O-
DMTr-deoxyadenosine (11.0 mg, 20 µmol) for 10 min. The
oligomer was deprotected and released from the polymer
support by treatment with concentrated aq NH3 (500 µL) for
40 min. The polymer support was removed by filtration and
washed with H2O (1 mL × 3). The filtrate was evaporated to
dryness under reduced pressure. The residue was dissolved
in water and analyzed by reversed-phase HPLC.
Th e Sta bility of d [Ap T] on P olym er Su p p or ts u n d er
Acid ic Con d ition s Sim ila r to Th ose Used for Detr ityla -
tion in th e P r oton -Block Meth od . d[ApT] (1 µmol) without
base protection on HCP were synthesized by use of the proton-
block method. The resin was treated with a 0.2 M solution of
NBT in THF (200 µL) or 3% trichloroacetic acid (400 µL). The
dimer was deprotected and released from the polymer support
by treatment with concentrated aq NH3 (500 µL) for 40 min.
The polymer support was removed by filtration and washed
with H2O (3 × 1 mL). The filtrate was evaporated and purified
by anion-exchange HPLC.
Next [ApT] (1 µmol) with an N-benzoyl group on HCP were
synthesized by use of the general procedure. Then the resin
was treated with 3% trichloroacetic acid (400 µL). the dimer
was deprotected and released from the polymer supports by
treatment with concentrated aq NH3 (500 µL) for 40 min. The
J . Org. Chem, Vol. 68, No. 14, 2003 5491

Sekine et al.
polymer support was removed by filtration and washed with
H2O (3 × 1 mL). The filtrate was evaporated to dryness under
reduced pressure. The residue was dissolved in water and
analyzed by anion-exchange HPLC.
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. d(CpCpCpT): dC:T 3.00:1.12. d(ApA-
pApT): dA:T 3.00:1.09. d(CpCpCpCpCpCpT): dC:T 6.0 0:0.97.
d(ApApApApApApT): dA:T 6.00:1.11. d(CpApGpTpCpApGpT-
pCpApGpT): dA:dG:dC:T 1.01:1.21:1.00:0.95.
P olym er Su p p or t Syn th esis of Oligod eoxyn u cleotid es.
d(ApT), mass (M + H) calcd 556.1558, found 556.1532; d(CpT)
mass, (M + H) calcd.532.1445, found 532.1443; d(GpT), mass
(M + H) calcd 572.1507, found 572.1475; d(ApA) calcd 565.1673,
found 565.1618; d(CpC), mass (M + H) calcd 517.1449, found
517.1404; d(ApApApT), mass (M + H) calcd 1182.2710, found
1182.2849; d(CpCpCpT), mass (M + H) calcd 1110.2374, found
1110.2199; d(ApApApApApApT), mass (M - H) calcd 2119.43,
found 2119.71; d(CpCpCpCpCpCpT), mass (M - H) calcd
1975.36, found 1975.50.
Ack n ow led gm en t. This work was supported by a
Grant from “Research for the Future” Program of the
J apan Society for the Promotion of Science (J SPS-
RFTF97I00301) and by CREST of J ST (J apan Science
and Technology Corporation) for the Synthesis of DNA
Oligomers on Polymer Supports.
En zym e Assa y of Mod ified Oligon u cleotid es. The en-
zymatic 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 (pH 9.0) at 37 °C for 40 min.
After the enzymes were deactivated by heating at 100 °C for
Su p p or tin g In for m a tion Ava ila ble: Basicity of nucleo-
bases and promoters for activation of phosphoramidites in the
proton-block approach (Figure 3). This material is available
free of charge via the Internet at http://pubs.acs.org.
J O034204K
5492 J . Org. Chem., Vol. 68, No. 14, 2003