Synthesis and properties of oligoribonucleotide analogs having formacetal internucleoside linkages

Full text of DOI:10.1021/jo962080o

1846
J . Org. Chem. 1997, 62, 1846-1850
O-5′, and thioformacetal9 3′-SCH2O-5′ linkages (for recent
reviews, see ref 2).
Syn th esis a n d P r op er ties of
Oligor ibon u cleotid e An a logs Ha vin g
There are only a few reports on such modifications in
oligoribonucleotides. Damha et al.10 have found that
oligodeoxynucleotides containing sulfide-linked (3′-CH2-
CH2S-5′) ribofuranosylthymine dimers exhibit higher
affinity to the RNA target than their counterparts having
the corresponding deoxy dimers. Cao and Matteucci11
have incorporated a diribonucleoside thioformacetal ana-
log in an oligodeoxyribonucleotide. If compared with the
full phosphodiester DNA or the deoxythioformacetal
modification, this substitution results in slightly de-
creased affinity toward the complementary RNA. Uni-
formly modified oligoribonucleotides with dimethylene
sulfone linkages (3′-CH2SO2CH2-5′) have been prepared,
but the strong self-association of this analog prevents
hybridization with both complementary RNA and DNA.12
Thus, the effect of various dephosphono linkages on the
stability of modified RNA-RNA duplexes still remains
largely unexplored.
In this paper, we report the synthesis of a formacetal
linked uridine-uridine dimer and its incorporation in
oligoribonucleotides. The stability of modified RNA-
RNA duplexes having formacetal linkages replacing
natural phosphodiesters at selected positions was studied
by UV melting experiments. The formacetal substitution
caused a slight increase in duplex stability (∆tm ) +0.2
to +0.8 °C per modification).
F or m a ceta l In ter n u cleosid e Lin k a ges
Eriks Rozners*,† and Roger Stro¨mberg†,‡
Department of Organic Chemistry, Arrhenius Laboratory,
Stockholm University, S-10691 Stockholm, Sweden, and
Laboratory of Organic and Bioorganic Chemistry,
Department of Medical Biochemistry and Biophysics,
Karolinska Institutet, S-17177 Stockholm, Sweden
Received November 6, 1996
In tr od u ction
The possibility to use short oligonucleotides as thera-
peutic agents has recently received much attention.1,2 To
be efficient, such antisense agents should exhibit high
enzymatic stability while retaining binding affinity and
specificity to the intracellular target, usually a messenger
RNA. Unmodified oligonucleotides are readily degraded
by nucleases and, thus, are poor antisense compounds.
There are many modifications (for recent reviews, see refs
1 and 2) that increase the enzymatic stability of oligo-
nucleotides. However, in most instances the binding
affinity decreases. Most of these studies were done on
oligodeoxynucleotides or 2′-O-alkyl analogs.
The use of oligoribonucleotide analogs would have the
advantages of higher stability of RNA-RNA duplexes
than DNA-RNA hybrids, ready access to monomeric
synthons (e.g., can be often readily prepared from modi-
fied sugars and heterocyclic bases), and less costly
starting materials. Unfortunately, a free 2′-OH function
causes cleavage of the phosphodiester linkage at a far
too high rate under physiological conditions. Some 2′-
O-alkylated analogs (2′-O-Me, 2′-O-allyl, 2′-O-CH2CH2-
OCH3) show both increased enzymatic stability and good
affinity to complementary RNA, but the starting materi-
als are less attractive because of higher price or elaborate
synthesis.1,2 However, oligoribonucleotide analogs with
dephosphono internucleoside linkages can contain free
2′-hydroxyls and still have high enzymatic and chemical
stability.
Resu lts a n d Discu ssion
The synthesis of oligoribonucleotides containing inter-
nucleoside formacetal linkages at defined positions is
most conveniently done with suitably protected dimeric
building blocks that can be used in standard oligonucle-
otide synthesis. Preparation of such ribonucleoside
dimers requires synthesis of monomeric units (donor and
acceptor) and coupling of them to form the formacetal
linkage. Final protecting group manipulation is neces-
sary to make the building block suitable for oligonucle-
otide synthesis.
For the formation of internucleosidic formacetal link-
ages we employed (methylthio)methylene (MTM) acetals
that previously have been used in the DNA series.6-9 The
syntheses of MTM-donors are shown in Scheme 1. As
starting material in the first experiments we used 2′-O-
(tert-butyldimethylsilyl)-5′-O-(monomethoxytrityl)uri-
dine13 (1). Treatment of 1 with dimethyl sulfide/benzoyl
Dephosphono linkages in oligodeoxyribonucleotides
that are reported to give stable duplexes are those having
amide3 3′-CH2CONH-5′ and 3′-CH2NHCO-5′, methylene-
(methylimino)4 3′-CH2N(CH3)O-5′, methylene(dimethyl-
hydrazo)5 3′-CH2N(CH3)N(CH3)-5′, formacetal6-9 3′-OCH2-
* To whom correspondence should be addressed. Phone: 46-8-
162486. Fax: 46-8-154908. E-mail: er@lab12.organ.su.se.
† Stockholm University.
‡ Karolinska Institutet.
(1) (a) Milligan, J . F.; Matteucci, M. D.; Martin, J . C. J . Med. Chem.
1993, 36, 1923-1937. (b) De Mesmaeker, A.; Ha¨ner, R.; Martin, P.;
Moser, H. E. Acc. Chem. Res. 1995, 28, 366-374.
(2) (a) Sanghvi, Y. S.; Cook, P. D. In Carbohydrate Modifications in
Antisense Research; Sanghvi, Y. S., Cook, P. D., Eds.; American
Chemical Society: Washington, DC, 1994; pp 1-22. (b) De Mesmaeker,
A.; Altman, K.-H.; Waldner, A.; Wendeborn, S. Curr. Opin. Struct. Biol.
1995, 5, 343-355.
(3) (a) De Mesmaeker, A.; Waldner, A.; Lebreton, J .; Fritsch, V.;
Wolf, R. M. In Carbohydrate Modifications in Antisense Research;
Sanghvi, Y. S., Cook, P. D., Eds.; American Chemical Society: Wash-
ington, DC, 1994; pp 24-39. (b) Lebreton, J .; Waldner, A.; Fritsch, V.;
Wolf, R. M.; De Mesmaeker, A. Tetrahedron Lett. 1994, 35, 5225-5228.
(4) (a) Debart, F.; Vasseur, J .-J .; Sanghvi, Y. S.; Cook, P. D. Bioorg.
Med. Chem. Lett. 1992, 2, 1479-1482. (b) Vasseur, J .-J .; Debart, F.;
Sanghvi, Y. S.; Cook, P. D. J . Am. Chem. Soc. 1992, 114, 4006-4007.
(5) Sanghvi, Y. S.; Vasseur, J .-J .; Debart, F.; Cook, P. D. Collect.
Czech. Chem. Commun. Special Issue 1993, 58, 158-162.
(6) (a) Veeneman, G. H.; van der Marel, G. A.; van den Elst, H.;
van Boom, J . H. Tetrahedron 1991, 47 1547-1562. (b) Quaedflieg, P.
J . L. M.; Timmers, C. M.; van der Marel, G. A.; Kuyl-Yeheskiely, E.;
van Boom, J . H. Synthesis 1993, 627-633.
(7) (a) Quaedflieg, P. J . L. M.; Pikkemaat, J . A.; van der Marel, G.
A.; Kuyl-Yeheskiely, E.; Altona, C.; van Boom, J . H. Recl. Trav. Chim.
Pays-Bas 1993, 112, 15-21. (b) Veeneman, G. H.; van der Marel, G.
A.; van den Elst, H.; van Boom, J .H. Recl. Trav. Chim. Pays-Bas 1990,
109, 449-451. (c) Quaedflieg, P. J . L. M.; van der Marel, G. A.; Kuyl-
Yeheskiely, E.; van Boom, J . H. Recl. Trav. Chim. Pays-Bas 1991, 110,
435-436.
(8) Matteucci, M. Tetrahedron Lett. 1990, 31, 2385-2388.
(9) (a) J ones, R. J .; Lin, K.-Y.; Milligan, J . F.; Wadwani, S.;
Matteucci, M. D. J . Org. Chem. 1993, 58, 2983-2991. (b) Lin, K.-Y.;
Pudlo, J . S.; J ones, R. J .; Bischofberger, N.; Matteucci, M. D.; Froehler,
B. C. Bioorg. Med. Chem. Lett. 1994, 4, 1061-1064. (c) Matteucci, M.
D.; Lin, K.-Y.; Butcher, S.; Moulds, C. J . Am. Chem. Soc. 1991, 113,
7767-7768.
(10) (a) Damha, M. J .; Meng, B.; Wang, D.; Yannopoulos, C. G.; J ust,
G. Nucleic Acids Res. 1995, 23, 3967-3973. (b) Meng, B.; Kawai, S.
H.; Wang, D.; J ust, G.; Giannaris, P. A.; Damha, M. J . Angew. Chem.,
Int. Ed. Engl. 1993, 32, 729-731.
(11) Cao, X.; Matteucci, M. D. Bioorg. Med. Chem. Lett. 1994, 4,
807-810.
(12) Richert, C.; Roughton, A. L.; Benner, S. A. J . Am. Chem. Soc.
1996, 118, 4518-4531.
S0022-3263(96)02080-4 CCC: $14.00 © 1997 American Chemical Society

Notes
J . Org. Chem., Vol. 62, No. 6, 1997 1847
Sch em e 1a
5 in 20 and 27% yield, respectively. In both cases
unidentified more polar and more lipophilic by products
were formed.
The selective 2′-benzoylation of 5′-O-dimethoxytri-
tyl nucleosides has been previously reported.16 The same
selectivity could be expected with 5′-O-TBDPS uridine
3. However, migration of the o-ClBz group to the 3′-OH
in 4 could occur during the synthesis of 5, thus yielding
the 2′-O-MTM isomer. The correct structure of 5 was
unambiguously assigned as the correct 3′-O-MTM isomer
by a 2D long-range 13C-1H correlation NMR experiment
(strong cross-peak signals corresponding to correlation
from C3′ to OCH2S protons and from OCH2S carbon to
H3′). Coupling of donor 5 with acceptor 6 using NIS/
TfOH6a,7a activation gave dimer 7 in 55% isolated yield.
The synthesis of the selectively protected ribonucleo-
side dimer suitable for incorporation in oligonucleotides
is complicated because of the need of selective protection
of the terminal 2′-OH. This could be done by synthesiz-
ing a monomeric acceptor having different protecting
groups at 2′, and 3′-OH and selectively removing the 3′-
O-protection after the synthesis of the dimer. However,
this approach would require more synthetic steps and
additional protecting groups to achieve selective protec-
tion and deprotection. Instead we used a more straight-
forward protecting group manipulation based again on
the selective 2′-o-chlorobenzoylatoin (Scheme 2).
a
Key: (i) o-Chlorobenzoyl chloride, 1.1 equiv, -78 °C (85%);
(ii) dimethyl sulfide, benzoyl peroxide, 2,4,6-collidine (20%); (iii)
DMSO, acetic anhydride, acetic acid (27%).
peroxide6 gave the 3′-O-MTM derivative 2 in only 20%
yield after silica gel column chromatography. The major
product isolated (in 31% yield) was the 3′-keto derivative.
To remove traces of non-nucleosidic byproducts, 2 was
also crystallized from hexane-ether. Coupling of 2 with
2′,3′-O-bis(tert-butyldimethylsilyl)uridine14 6 using either
N-bromosuccinimide6a,8 (NBS) or N-iodosuccinimide6a,7b
(NIS) as activators was not successful. Varying yields
of the corresponding couplings in the deoxy series have
been reported (NBS 45%8 and 11%,6a NIS 30%,7b and
32%6a). Possibly the steric bulkiness of the 2′-O-TBDMS
group complicated the synthesis further.
Recently, the activation with NIS in the presence of a
catalytic amount of trifluoromethanesulfonic acid (TfOH)15
was shown to be superior in the synthesis of formacetal-
linked 2′-deoxynucleosides.6a,7a,c Since this method would
not be compatible with the acid-labile MMT protection,
we developed a new protecting group strategy. The key
step was the selective 2′-o-chlorobenzoylation16,17 of 5′-
O-(tert-butyldiphenylsilyl) uridine18 (3) (tert-butyldiphe-
nylsilyl ) TBDPS) (Scheme 1). Reaction of 3 with 1.1
equiv of o-chlorobenzoyl chloride at -78 °C gave selec-
tively the 2′-acyl derivative 4 containing only traces of
the possible 3′-acyl isomer. Reactions of 4 with dimethyl
sulfide/benzoyl peroxide6 or DMSO/acetic anhydride/
acetic acid19 gave after silica gel column chromatography
and crystallization from hexane-ether the 3′-MTM donor
Removal of all silyl protecting groups with tetrabutyl-
ammonium fluoride gave dimer 8, which was protected
at the 5′-OH with a monomethoxytrityl (MMT) group to
give 9. Dimer 9 was reacted with 1.1 equiv of o-
chlorobenzoyl chloride at -78 °C and successively phos-
phonylated17 in the same reaction mixture using the PCl3/
imidazole reagent20 to give the dimeric building block 10
for use in oligonucleotide synthesis. The correct structure
of 10 was confirmed by 1H NMR spectroscopy.21 This
protecting group strategy can be used as a general
method for preparation of modified ribonucleoside dimers
suitable for incorporation in oligonucleotides.
The isomeric purity of the product isolated is a crucial
question when using the selective 2′-acylation of ribo-
nucleosides. When this approach was used to synthesize
protected ribonucleoside 3′-O-hydrogen phosphonates we
found that the target products can be efficiently sepa-
rated from the 2′-isomers using simple silica gel chro-
matography.17b This finding was important because even
a small contamination would produce a nonhomogeneous
product after multiple coupling steps. In the present
work we incorporated the formacetal linked dimer 10 in
oligoribonucleotides at selected sites. We used TLC and
NMR to check the isomeric purity of compounds 5 and
7-10. Potential contamination undetectable using these
methods should have little influence on the evaluation
of the properties of the modified oligoribonucleotides
synthesized.
(13) Hakimelahi, G. H.; Proba, Z. A.; Ogilvie, K. K. Can. J . Chem.
1982, 60, 1106-1113.
(14) Compound 6 was synthesized using the procedures published:
Ogilvie, K. K.; Sadana, K. L.; Thompson, E. A.; Quilliam, M. A.;
Westmore, J . B. Tetrahedron Lett. 1974, 2861-2863. Ogilvie, K. K.;
Thompson, E. A.; Quilliam, M. A.; Westmore, J . B. Tetrahedron Lett.
1974, 2865-2868.
(15) (a) Veeneman, G. H.; van Leeuwen, S. H.; van Boom, J . H.
Tetrahedron Lett. 1990, 31, 1331-1334. (b) Konradson, P.; Mootoo, D.
R.; McDevitt, R. E.; Fraser-Reid, B. J . Chem. Soc., Chem. Commun.
1990, 270-272.
Three model oligoribonucleotides (Table 1) having two
(11), three (12), and six (13) internucleoside formacetal
linkages (f) were synthesized using the automated solid-
(16) Kempe, T.; Chow, F.; Sundquist, W. I.; Nardi, T. J .; Paulson,
B.; Peterson, S. M. Nucl. Acids Res. 1982, 10, 6695-6714.
(17) (a) Rozners, E.; Stro¨mberg, R.; Bizdena, E. Nucleosides Nucle-
otides 1995, 14, 855-857. (b) Rozners, E.; Renhofa, R.; Petrova, M.;
Popelis, J .; Kumpins, V.; Bizdena, E. Nucleosides Nucleotides 1992,
11, 1579-1593.
(19) Pojer, P. M.; Angyal, S. J . Tetrahedron Lett. 1976, 3067-3068.
(20) (a) Garegg, P. J .; Regberg, T.; Stawinski, J .; Stro¨mberg, R.
Chem. Script. 1986, 26, 59-62. (b) Stawinski, J .; Stro¨mberg, R.; Thelin,
M.; Westman, E. Nucleic Acids Res. 1988, 16, 9285-9298.
(21) The signal of H3′ in the 5′-yl uridine unit of 10 showed typical
coupling pattern (pentet, 1:2:2:2:1) due to spin-spin interactions with
H2′ and H4′ (J ) ca. 5 Hz) and with P (J ) ca. 10 Hz), thus confirming
the correct structure of 10 as the 3′-O-hydrogenphosphonate.
(18) Compound 3 was synthesized using the published procedure:
Hanessian, S.; Lavallee, P. Can. J . Chem. 1975, 53, 2975-2977.

1848 J . Org. Chem., Vol. 62, No. 6, 1997
Notes
Sch em e 2a
a
Key: (i) NIS, TfOH (cat.) (55%); (ii) tetrabutylammonium fluoride (84%); (iii) 4-monomethoxytrityl chloride (65%); (iv) o-chlorobenzoyl
chloride, 1.1 equiv, -78 °C; (v) PCl3, imidazole, NEt3, -78 °C, triethylammonium bicarbonate (aqueous) (71%).
Ta ble 1. Oligon u cleotid e Sequ en ces a n d Affin ity
Tow a r d Com p lem en ta r y RNA (tm
The effect of formacetal replacement in our models was
an increase of duplex stability of +0.2 to +0.8 °C per
modification (Table 1). This is in contrast to incorpora-
tion of formacetal linkages in oligodeoxynucleotides that
is reported to decrease the stability of the DNA-RNA
duplexes by -0.7 °C per modification (ref 9a, seven
modifications in a 15mer). A decrease is also reported
for DNA-DNA duplexes: ∆tm from -1.4 to -2.4 °C (ref
7a, one to three modifications in a 10mer).
)
no.
oligonucleotidesa
full ribo
modifdb
11
12
13
AAGCGAUfUUfUGACACU
ACAUfUCGUfUGUfUCGA
(UfU)6U
56.7
53.4
13.8
57.2 (+0.3)
53.9 (+0.2)
18.8 (+0.8)
a
b
f denotes position of the formacetal linkage. ∆tm are given
in parentheses.
Replacement of the phosphodiester moiety with a
formacetal linkage should weaken the O4′-C4′-C3′-O3′
gauche effect,24 which drives the ribose conformation
toward 2′-endo. This in turn should stabilize the RNA-
RNA duplex (A type, 3′-endo sugars), whereas DNA-
DNA duplexes (B type, 2′-endo sugars) should be desta-
bilized.25 The effects in DNA-RNA duplexes where the
DNA strand has modified internucleoside linkages and
may adopt intermediate equilibria or other conformations
are more difficult to predict. Destabilization of such
hybrids is reported.8,9a
The reduction of the unfavorable O4′-C4′-C3′-O3′
gauche effect apparently compensated for changes in
bond length when the phosphodiester linkages were
replaced by formacetals (C-O compared to P-O). Thus,
a slight stabilization was observed in our models. The
CD spectra of RNA-RNA duplexes containing formacetal
linkages were practically identical with those of unmodi-
fied ones and consistent with an A form helix (see the
Supporting Information, Figure 2). This is in good
agreement with NMR studies26 and molecular dynamics
simulations27 showing that the perturbation of the DNA-
DNA duplex caused by a single formacetal linkage is
minimal.
F igu r e 1. UV melting curves of duplexes formed by 11 (s)
and full ribo (- - -) and full deoxy (-‚-) versions of 11 with
complementary RNA.
phase H-phosphonate method for RNA synthesis22 and
2′-O-o-ClBz protection.17 Dimer 10 was used with stan-
dard coupling conditions.22b Full ribo versions of se-
quences 11-13 and the complementary oligoribonucle-
otides were also synthesized. The stability of the corre-
sponding duplexes were characterized by UV melting
experiments (see Table 1 and Figure 1). The correct
nucleoside residue composition of all oligonucleotides was
confirmed by enzymatic degradation followed by RP
HPLC analysis. The molecular weight of 11 was also
confirmed by matrix-assisted laser desorption ionization
time-of-flight MS.23
(23) Calcd 4966.2, found 4974.3. MS experiments were performed
as described: Pieles, U.; Zurcher, W.; Scha¨r, M.; Moser, H. E. Nucleic
Acids Res. 1993, 21, 3191-3196.
(24) (a) Plavec, J .; Thibaudeau, C.; Viswanadham, G.; Sund, C.;
Chattopadhyaya, J . J . Chem. Soc., Chem. Commun. 1994, 781-783.
(b) Plavec, J .; Thibaudeau, C.; Viswanadham, G.; Sund, C.; Sandstro¨m,
A.; Chattopadhyaya, J . Tetrahedron 1995, 51, 11775-11792.
(25) (a) Plavec, J .; Tong, W.; Chattopadhyaya, J . J . Am. Chem. Soc.
1993, 115, 9734-9746. (b) Griffey, R. H.; Lesnik, E.; Freier, S.; Sanghvi,
Y. S.; Teng, K.; Kawasaki, A.; Wheeler, P.; Mohan, V.; Cook, P. D. In
Carbohydrate Modifications in Antisense Research; Sanghvi, Y. S.,
Cook, P. D., Eds.; American Chemical Society: Washington, DC, 1994;
pp 212-224.
(22) (a) Garegg, P. J .; Lindh, I.; Regberg, T.; Stawinski, J .; Stro¨m-
berg, R.; Henrichson, C. Tetrahedron Lett. 1986, 27, 4055-4059. (b)
Westman, E.; Sigurdsson, S.; Stawinski, J .; Stro¨mberg, R. Nucleic Acids
Res. Symp. Ser. No. 31 1994, 25-26.

Notes
Although there are many different sides to developing
antisense compounds, favorable duplex formation is likely
to be of major importance. Because RNA-RNA duplexes
are generally more stable than DNA-RNA hybrids,
oligoribonucleotide analogs having chemically and enzy-
matically stable internucleoside linkages may be better
second generation antisense compounds than their deoxy
counterparts. For instance, the duplex formed by the full
J . Org. Chem., Vol. 62, No. 6, 1997 1849
270 MHz) δ: 8.92 (s, 1H, NH), 7.92 (d, J ) 7.7 Hz, 1H, o-ClBz),
7.82 (d, J ) 8.1 Hz, 1H, H6), 7.68 (m, 4H, Ar), 7.44 (m, 9H, Ar),
6.32 (d, J ) 4.4 Hz, 1H, H1′), 5.54 (m, 1H, H2′), 5.43 (d, 1H,
H5), 4.67 (m, 1H, H3′), 4.17 (m, 1H, H4′), 4.12 and 3.90 (ABX
system, J H5′-H5′′ ) 12.0 Hz, J H5′-H4′ ) 1.2 Hz, 2H, H5′), 1.12 (s,
9H, CH3). 13C NMR (CDCl3, 67.9 MHz) δ: 164.80, 163.08 (C4,
CdO in o-ClBz), 150.26 (C2), 139.85 (C6), 135.77, 135.53, 132.80,
132.18, 130.38, 130.30, 128.19 (TBDPS), 134.05, 133.56, 132.43,
131.37, 128.65, 127.06 (o-ClBz), 102.88 (C5), 86.67 (C1′), 84.75
(C4′), 77.21 (C2′), 69.72 (C3′), 63.14 (C5′), 27.15 (CH3), 19.49
(quaternary C in t-Bu).
deoxy version of 11 with complementary RNA had tm
)
46.9 °C (Figure 1), i.e., 9.8 °C lower than the duplex
formed by the full ribo oligomer. The replacement of
selected phosphodiester linkages with formacetal groups
further increased the affinity of the oligoribonucleotides
towards their RNA complements. Thus, oligonucleotides
containing ribonucleoside formacetal units are of poten-
tial use as second-generation antisense compounds. The
protecting group manipulations reported herein should
be also useful for synthesis of ribonucleoside dimers
having other modified linkages and thus allow synthesis
and more extensive studies of oligonucleotides carrying
such linkages.
2′-O-(o-ch lor oben zoyl)-5′-O-(ter t-bu tyld ip h en ylsilyl)-3′-
O-[(m eth ylth io)m eth yl]u r id in e (5). 2′-O-(o-Chlorobenzoyl)-
5′-O-(tert-butyldiphenylsilyl)uridine (4) (2.30 g 3.7 mmol) was
dissolved in a freshly prepared mixture of DMSO (12.5 mL),
acetic anhydride (7.5 mL), and acetic acid (2.5 mL). The mixture
was stirred at room temperature for 24 h (TLC, solvent B) and
evaporated. The residue was dissolved in toluene (100 mL) and
extracted with saturated NaHCO3 (aqueous) (3 × 100 mL) and
saturated NaCl (aqueous) (100 mL). The organic layer was
separated, dried over Na2SO4, evaporated, and purified by silica
gel column chromatography (0-40% of ethylacetate in toluene).
The fractions containing product were pooled, evaporated and
crystallized from hexane-ether to give 5 as a white powder.
Yield: 0.68 g, 27% (unidentified trace contamination was
1
detected by H NMR). Rf ) 0.62 (solvent B). 1H NMR (CDCl3,
270 MHz) δ: 9.06 (s, 1H, NH), 7.91 (d, J ) 7.0 Hz, 1H, o-ClBz),
7.80 (d, J ) 8.0 Hz, 1H, H6), 7.74-7.65 and 7.46-7.30 (m, 13H,
Ar), 6.33 (d, J ) 4.0 Hz, 1H, H1′), 5.49 (t, 1H, H2′), 5.35 (dd,
J H5-NH ) 2.0 Hz, 1H, H5), 4.72 (m, 1H, H3′), 4.68 and 4.60 (AB
system, J ) 11.7 Hz, 2H, OCH2S), 4.20 (m, 1H, H4′), 3.94 and
3.70 (ABX system, J H5′-H5′′ ) 11.7 Hz, J H5′-H4′ ) 1.9 Hz, 2H, H5′),
2.02 (s, 3H, SCH3), 1.14 (s, 9H, CH3). 13C NMR (CDCl3, 67.9
MHz) δ: 164.30, 163.11 (C4, CdO in o-ClBz), 150.25 (C2), 139.65
(C6), 135.84, 135.55, 132.93, 132.14, 130.39, 130.28, 128.22,
128.17 (TBDPS), 134.36, 133.36, 132.14, 131.37, 128.82, 126.93
(o-ClBz), 103.01 (C5), 87.12 (C1′), 83.42 (C4′), 75.93 (OCH2S),
75.50 (C2′), 73.04 (C3′), 62.60 (C5′), 27.21 (CH3), 19.49 (quater-
nary C in t-Bu), 14.23 (SCH3). Positive FAB MS (mNBA matrix)
m/ z: 703 (M + Na), 681 (M + H), 633 (M - SCH3), 623 (M -
t-Bu), 569 (M - Ura).
Syn th esis of Dim er 7. 2′,3′-O-Bis(tert-butyldimethylsilyl)-
uridine (6)14 (0.38 g, 0.81 mmol) was coupled with 2′-O-(o-
chlorobenzoyl)-5′-O-(tert-butyldiphenylsilyl)-3′-O-[(methylthio)-
methyl]uridine (5) (0.61 g, 0.9 mmol, 1.2 equiv) using the NIS/
TfOH activation procedure previously reported.6a,7a The product
was purified by silica gel column chromatography (0-5% of CH3-
OH in CHCl3). Yield: 0.49 g, 55% (based on 6, no contamination
could be detected by TLC and 1H NMR). Rf ) 0.50 (solvent B).
1H NMR (CDCl3, 270 MHz)28 δ: 9.58 and 9.48 (2s, 2H, NH), 7.91
(d, J ) 7.3 Hz, 1H, o-ClBz), 7.71-7.35 (m, 17H, Ar, H6*, H6),
6.27 (d, J ) 5.1 Hz, 1H, H1′), 5.66-5.43 (m, 4H, H1′*, H2′, H5,
H5*), 4.82 and 4.70 (AB system, J ) 6.6 Hz, 2H, OCH2O), 4.57
(m, 1H, H3′), 4.26-3.58 (m, 8H, H4′, H2′*, H3′*, H4′*, H5′, H5′*),
1.13 (s, 9H, CH3 in TBDPS), 0.88 and 0.87 (2s, 18H, t-Bu in
TBDMS), 0.09-0.04 (4s, 12H, SiCH3). 13C NMR (CDCl3, 67.9
MHz)28 δ: 164.46, 163.62, 163.32 (C4, C4*, CdO in o-ClBz),
150.41, 150.32 (C2, C2*), 140.30, 139.95 (C6, C6*), 135.78,
135.51, 132.70, 132.16, 130.41, 130.33, 128.19 (TBDPS), 134.27,
133.49, 132.08, 131.43, 128.68, 127.00 (o-ClBz), 103.11, 102.11
(C5, C5*), 96.60 (OCH2O), 90.49 (C1′*), 87.25 (C1′), 83.69, 82.33,
75.71, 75.36, 75.17, 70.98 (C2′, C2′*, C3′, C3′*, C4′, C4′*), 67.33,
63.36 (C5′, C5′*), 27.17, 25.90 (CH3), 19.44, 18.12 (quaternary
C in t-Bu), -4.23, -4.42, -4.72, -4.83 (SiCH3). Positive FAB
MS (mNBA matrix) m/ z: 1127 (M + Na), 1105 (M + H), 1047
(M - t-Bu), 993 (M - Ura), 633 (M - Urd,2 × TBDMS), 383
(Urd, o-ClBz).
Exp er im en ta l Section
Pyridine and acetonitrile were dried over 3 Å molecular sieves.
Methylene chloride was dried over 4 Å molecular sieves. Tri-
ethylamine and DMSO were dried by refluxing with CaH2
overnight followed by distillation. PCl3 was distilled. Triethyl-
ammonium bicarbonate buffer (pH ca. 7.5) was prepared by
passing CO2 (g) through a mixture of triethylamine and water
until saturation. Acetic acid (pa), acetic anhydride (pa), hexane
(pa), 32% aqueous NH3 (pa), and trifluoromethanesulfonic acid
were purchased from Merck, tetrabutylammonium fluoride,
o-chlorobenzoyl chloride, and 4-methoxytrityl chloride were
purchased from Aldrich, and imidazole, N-iodosuccinimide were
purchased from Fluka and used without further purification.
NMR spectra were recorded on a J EOL GSX-270 spectrometer
at 25 °C. Chemical shifts are given in ppm relative to tetram-
ethylsilane (1H), CDCl3 (δ 77.17 ppm, 13C) and 2% H3PO4 in D2O
(coaxial inner tube, 31P). Signals were assigned by H-1H and
1
13C-1H COSY. Long-range 13C-1H correlation was done using
J EOL pulse sequence and parameters for standard 13C-1H
COSY experiment except that the delay time was 60 ms. FAB
mass spectra were recorded on a J EOL SX-102 instrument with
m-nitrobenzyl alcohol (m-NBA) or glycerol as a matrix. TLC
was done on Merck silica gel 60 F254 precoated plates using
solvents A (CHCl3/methanol, 19:1, v/v), B (toluene/ethyl acetate,
1:1, v/v), C (CHCl3/methanol, 4:1, v/v), D (CHCl3/methanol, 9:1,
v/v), E (2-propanol/water/25% aqueous NH3, 85:5:10, v/v/v).
Silica gel (35-70 µm) from Amicon Europe was used for column
chromatography, and the columns were run in the flash mode;
chloroform was passed through basic Al2O3 prior to use.
2′-O-o-(Ch lor oben zoyl)-5′-O-(ter t-bu tyld ip h en ylsilyl)u r i-
d in e (4). 5′-O-(tert-Butyldiphenylsilyl)uridine (3)18 (2.41 g, 5
mmol) was coevaporated with dry pyridine (2 × 50 mL) and
dissolved in CH2Cl2/pyridine (19:1, 80 mL). The reaction
mixture was cooled to -78 °C (acetone-dry ice), a solution of
o-chlorobenzoyl chloride (0.70 mL, 5.5 mmol, 1.1 equiv) in CH2-
Cl2 (5 mL) was added during 15 min, and the mixture was stirred
for 1 h at -78 °C. CH3OH (0.5 mL) was added, and the mixture
was extracted with saturated NaHCO3 (aqueous) (100 mL). The
organic layer was separated, dried over Na2SO4, evaporated to
ca. 20 mL and precipitated in hexane (500 mL). The precipitate
was filtered and dried in vacuo, yield 2.42 g, 85%, Rf ) 0.46
(solvent A). Trace amounts of the possible 3′-O-(o-chlorobenzoyl)
isomer could be detected Rf ) 0.39 (solvent A). To avoid
migration of the o-chlorobenzoyl group this crude material was
used without further purification. The trace impurities were
removed during subsequent synthetic steps. 1H NMR (CDCl3,
Syn th esis of Dim er 8. Dimer 7 (1.33 g, 1.2 mmol) was
dissolved in dry acetonitrile (25 mL), and tetrabutylammonium
fluoride (3.12 g, 12 mmol, dried by coevaporation with acetoni-
trile, 2 × 100 mL) was added. The mixture was stirred at room
temperature for 3 h (TLC, solvents A and C), evaporated to ca.
10 mL, and extracted with CHCl3 (50 mL) and saturated
NaHCO3 (aqueous) (50 mL). The aqueous layer was extracted
with CHCl3 (50 mL), organic layers were separated, dried over
(26) Gao, X.; Brown, F. K.; J effs, P.; Bischofberger, N.; Lin, K-Y.;
Pipe, A. J .; Noble, S. A. Biochemistry 1992, 31, 6228-6236.
(27) Veal, J . M.; Gao, X.; Brown, F. K. J . Am. Chem. Soc. 1993, 115,
7139-7145.
(28) * indicates resonances from protons and carbons in the 5′-yl
unit of the dimer.

1850 J . Org. Chem., Vol. 62, No. 6, 1997
Notes
Na2SO4, and evaporated, and the residue was purified by silica
gel column chromatography (0-16% of CH3OH in CHCl3).
Yield: 0.65 g, 84%. Rf ) 0.46 (solvent C). Positive FAB MS
(glycerol matrix) m/ z: 639 (M + H), 527 (M - Ura), 395 (M -
Urd).
Syn th esis of Dim er 9. Dimer 8 (0.64 g, 1 mmol) was reacted
with 4-monomethoxytrityl chloride (0.34 g, 1.1 mmol) according
to the standard procedure,29 and the product was purified using
silica gel column chromatography (0-10% of CH3OH in CHCl3).
Yield: 0.59 g, 65%. Rf ) 0.37 (solvent D). Positive FAB MS
(mNBA matrix) m/ z: 933 (M + Na), 910 (M), 666 (M - Urd).
Syn th esis of Dim er 10. Dimer 9 (0.27 g, 0.3 mmol) was
coevaporated with dry pyridine (2 × 20 mL) and dissolved in
dry CH2Cl2/pyridine (19:1, 20 mL). The reaction mixture was
through a disposable syringe filter (Millex GV13, 0.22 µm), and
purified on a Dionex NucleoPac PA-100 (4 × 250 mm) column
using a linear gradient of 0-90 mM LiClO4 in 20 mM sodium
acetate (pH 6.5), 10% CH3CN for 45 min. The collected fractions
were lyophilized and further purified on a Supelcosil LC-18 (3
µm, 4.6 × 150 mm) column using a linear gradient of 0-25%
CH3CN in 50 mM triethylammonium acetate (pH 6.5) for 40 min.
The oligonucleotides were collected, lyophilized, dissolved in
water, and lyophilized again.
En zym a tic Clea va ge. Oligonucleotides (0.5 A260 units) were
dissolved in buffer (100 mM Tris-HCl, 12 mM MgCl2, 100 mM
NaCl, pH 8.6, 0.5 mL), snake venom phosphodiesterase (25 µL,
0.1 units/ml, Sigma) and calf intestine phosphatase (10 µL, 50
units/mL, Boehring Mannheim) were added, and reactions were
incubated at 37 °C for 5 h. Aliquots (100 µL) were withdrawn,
filtered (ultrafree-MC 10000 NMWL filter unit, Millipore), and
analyzed on a Supelcosil LC-18 (3 µm, 4.6 × 150 mm) column
using a linear gradient of 0-5% CH3CN in 50 mM triethylam-
monium acetate (pH 6.5) for 40 min. Deprotected (32% NH3/
EtOH 3:1 for 8 h at rt) dimer 8 was used as standard.
Hybr id iza tion Th er m od yn a m ics. Absorbance vs temper-
ature profiles were measured at 260 nm on a Varian Cary 3
spectrophotometer in a buffer containing 100 mM NaCl, 10 mM
sodium phosphate (pH 7.2), 0.1 mM EDTA, and 2 µM oligo-
nucleotides. Extinction coefficients were calculated from nearest
neighbor approximation.31 The temperature was increased at
a rate of 0.2 °C. The melting temperatures were obtained by
fitting the melting profile to a two-state transition model, with
linear sloping lower and upper base lines and also using the
Varian Cary software (Version 2.5). Reported values are the
average of at least three experiments. CD spectra were recorded
on a J ASCO Model 700 spectropolarimeter at 20 or 4 °C (for 13
duplexes) using the same samples.
cooled to -78 °C (acetone-dry ice), and
a solution of o-
chlorobenzoyl chloride (43 µL, 0.33 mmol) in CH2Cl2 (2 mL) was
added during 15 min. The mixture was stirred at -78 °C for
30 min (TLC, solvent D) and then added dropwise to a stirred
and cooled (-78 °C) mixture of imidazole (0.24 g, 3.5 mmol), PCl3
(0.1 mL, 1.1 mmol), and triethylamine (0.5 mL, 3.6 mmol) in
dry CH2Cl2 (30 mL). The reaction mixture was stirred at -78
°C for 30 min and then poured onto and extracted with 1.0 M
TEAB (aqueous) pH 7.5 (60 mL). The organic layer was
separated, dried over Na2SO4, and evaporated, and the residue
was purified by silica gel column chromatography (0-15% of
CH3OH in CHCl3 containing 0.1% of triethylamine). Yield: 0.26
g, 71%. Rf ) 0.27 (solvent C), 0.65 (solvent E). 1H NMR (CDCl3,
270 MHz)28 δ: 7.94, 7.88 (2m, 2H, o-ClBz), 7.70, 7.47-7.22 (m,
20H, H6, H6*, Ar), 6.88 (d, J ) 630 Hz, 1H, PH), 6.85 (m, 2H,
MMT), 6.20 (d, J ) 4.4 Hz, 1H, H1′), 6.14 (d, J ) 5.1 Hz, 1H,
H1′*), 5.68 (bt, 1H, H2′), 5.60, 5.36 (2d, J ) 8.0 Hz, 2H, H5,
H5*), 5.43 (bt, 1H, H2′*), 4.99 (p, 1H, H3′*), 4.75 (m, 2H,
OCH2O), 4.67 (bt, 1H, H3′), 4.31 (m, 2H, H4′, H4′*), 3.84-3.50
(2m, 4H, H5′, H5′*), 3.77 (s, 3H, OCH3), 2.77 (q, J ) 7.3 Hz, 6H,
NCH2), 1.11 (t, 9H, CH3). 13C NMR (CDCl3, 67.9 MHz)28 δ:
164.40, 164.08, 163.54, 163.30 (C4, C4*, CdO in o-ClBz), 150.56
(C2, C2*), 158.92, 143.79, 143.57, 130.62, 128.62, 128.19, 127.46
(MMT), 140.28, 140.04 (C6, C6*), 134.67, 134.13 (MMT, o-ClBz),
133.40, 133.10, 132.45, 132.08,, 131.35, 131.05, 129.00, 128.76,
127.00, 126.87 (o-ClBz), 113.46 (MMT), 103.21, 102.96 (C5, C5*),
96.41 (OCH2O), 87.64, 87.37, 87.07 (C1′, C1′*, MMT), 82.97,
82.57 (C4′, C4′*), 75.55, 75.28, 75.04 (C2′, C2′*, C3′), 70.78 (C3′*),
67.89, 62.71 (C5′, C5′*), 55.37 (OCH3), 45.80 (NCH2), 9.69 (CH3).
31P NMR (CH3CN:pyridine, 3:1, 109.4 MHz) δ: 1.75. Negative
FAB MS (mNBA matrix) m/ z: 1111 (M - H).
Ack n ow led gm en t. This study was supported by the
Swedish Natural Science Research Council, the Swedish
Royal Academy of Sciences, Pharmacia Research Foun-
dation, and Magnus Bergvall Foundation. We thank
Dr. J . J ohansson (Karolinska Institutet) for help with
MALDI-TOF MS analysis, Prof. A. Gra¨slund and Dr.
M. Sarkar (Stockholm University) for help with CD
measurements, and the Wenner-Gren Center Founda-
tion for a postdoctoral fellowship to E.R.
Syn th esis a n d P u r ifica tion of Oligon u cleotid es. Solid-
phase synthesis22 and N-protecting30 and 2′-O-o-ClBz protecting
groups17 have been previously reported. Synthesized oligonucle-
otides were cleaved from the support and deprotected using 32%
NH3/EtOH 3:1 for 8 h at rt. The ammonia solutions were
lyophilized, and the residue was dissolved in 30% CH3CN, passed
through a disposable C-18 cartridge (Waters Sep Pac), filtered
Su p p or tin g In for m a tion Ava ila ble: Spectral data of 8
and 9 and CD spectra of oligonucleotide duplexes formed by
11-13 with their complementary RNA sequences (2 pages).
This material is contained in libraries on microfiche, im-
mediately follows this article in the microfilm version of the
journal, and can be ordered from the ACS; see any current
masthead page for ordering information.
J O962080O
(29) Connolly, B. A. In Oligonucleotides and Analogues: A Practical
Approach; Eckstein, F., Ed.; IRL Press: Oxford, 1991; pp 161-162.
(30) Westman, E.; Stawinski, J .; Stro¨mberg, R. Collect. Czech. Chem.
Commun. 1993, 58, 236-237.
(31) Puglisi, J . D.; Tinoco, I., J r. Methods Enzymol. 1989, 180, 304-
324.