Synthesis and properties of 2'-O-methyl-2-thiouridine and oligoribonucleotides containing 2'-O-methyl-2-thiouridine

Article abstract of DOI:10.1016/S0960-894X(00)00342-5

A new method for the synthesis of 2'-O-methyl-2-thiouridine (s²Um) found in thermophilic bacterial tRNA was developed. Structural properties of s²Um and s²Um(p)U were studied by using ¹H NMR spectroscopy. A modified nonaribonucleotide (RNA*: 5'-CGUUs²UmUUGC-3') was synthesized to study the base-recognition ability of s²Um in formation of RNA-RNA and RNA-DNA duplexes. The UV melting experiments revealed that RNA*-RNA and RNA*-DNA duplexes having an s²U-A base pair are more stable than those having a U-A base pair. On the contrary, the thermal stability of RNA*-RNA and RNA*-DNA duplexes having an s²U-G wobble base pair was much lower than that of the unmodified duplexes having a natural U-G base pair. It is concluded that s²Um has higher selectivity toward A over G than unmodified U. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00342-5

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1795±1798
Synthesis and Properties of 20-O-Methyl-2-thiouridine and
Oligoribonucleotides Containing 20-O-Methyl-2-thiouridine
Koh-ichiroh Shohda, Itaru Okamoto, Takeshi Wada, Kohji Seio and Mitsuo Sekine*
Department of Life Science, Faculty of Bioscience and Biotechnology, Tokyo Institute of Technology,
Nagatsuta, Midoriku, Yokohama 226-8501, Japan
Received 19 April 2000; accepted 7 June 2000
AbstractÐA new method for the synthesis of 20-O-methyl-2-thiouridine (s2Um) found in thermophilic bacterial tRNA was developed.
Structural properties of s2Um and s2UmpU were studied by using 1H NMR spectroscopy. A modi®ed nonaribonucleotide (RNA*:
50-CGUUs2UmUUGC-30) was synthesized to study the base-recognition ability of s2Um in formation of RNA±RNA and RNA±
DNA duplexes. The UV melting experiments revealed that RNA*±RNA and RNA*±DNA duplexes having an s2U-A base pair are
more stable than those having a U-A base pair. On the contrary, the thermal stability of RNA*±RNA and RNA*±DNA duplexes
having an s2U-G wobble base pair was much lower than that of the unmodi®ed duplexes having a natural U-G base pair. It is
concluded that s2Um has higher selectivity toward A over G than unmodi®ed U. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
In addition to this expectation, oligonucleotides con-
taining s2Um are thought to form double helixes more
selectively than those containing uridines. The word
``selectively'' means that the 2-thiocarbonyl group
destabilizes an s2U-G wobble base pair because of the
weaker hydrogen bonding nature of the thiocarbonyl
group and the steric repulsion derived from the CˆS
bond that is longer than the CˆO bond. Although the
U-G wobble base pair is a well-known base mismatch, it
does not decrease the Tm value so much as other base
mismatches like the C-A base pair. This is because the
U-G pair can form two hydrogen bonds (N3 of UÁ Á ÁO6
of G, and O2 of UÁ Á ÁN1 of G).9 Quite recently, the
thermodynamic parameters (in 1 M NaCl) of RNA
duplexes having an s2U-G wobble base pair were
reported by Testa et al.10 They described that the melt-
ing temperatures of RNA duplexes having the s2U-G
wobble base pair were considerably lower than those
obtained in the case of the matched s2U-A base pair.
20-O-Methyl-2-thiouridine is a doubly modi®ed nucleoside
discovered from tRNAs of extremely thermophilic
archaebacteria (Sulfolobus solfataricus, Thermoproteus
neutrophilus, and Pyrodictium occultum) by Edmonds et
al.1 Both the ribose residues of 2-thiouridine (s2U)2 and
20-O-methyluridine (Um)3 have a C30-endo conformation,
since the 2-thiocarbonyl and 20-O-methyl groups are more
sterically hindered than the 2-carbonyl and 20-hydroxyl
groups, respectively.4 Therefore, the ribose puckering of
s2Um is predicted to have preferentially the C30-endo
form because of the additive e€ect of the 2-thiolation and
20-O-methylation. The C30-endo predominance of nucleo-
sides is entropically favorable for stable A-type RNA
duplex formation, because all the ribose conformations of
A-type RNA duplexes are C30-endo.5 It was reported
that RNA duplexes containing s2U or Um were thermo-
dynamically more stable than unmodi®ed natural RNA
duplexes.6,7 Therefore, an oligoribonucleotide bearing
s2Um, which was predicted to have a rigid C30-endo form,
was expected to form more stable A-type RNA duplexes.
Moreover, the hydrophobic nature of the 2-thiocarbonyl
and 20-O-methyl groups may enhance permeability to the
cell membrane, and the 20-O-methylated oligoribonucleo-
tides acquire resistance against nucleases.7b These proper-
ties of s2Um are expected to be signi®cantly useful for
the antisense strategy.8
In order to examine the e€ects of the 2-thiocarbonyl
group on formation of a base pair, we performed ab
initio calculation of U-G wobble base pairs (Fig. 1).11
The hydrogen bonding energy of the base pair formed
between 1-methyl-2-thiouracil and 9-methylguanine was
10.4 kcal/mol, while that of the natural U-G base was
calculated to be 14.1kcal/mol.12 These results apparently
suggest the weakened hydrogen bonding property of the
2-thiocarbonyl group as expected above. In addition, the
optimized geometry of the 1-methyl-2-thiouracil and 9-
methylguanine base pair was simultaneously calculated.
*Corresponding author. Tel.:+81-45-924-5706; fax:+84-45-924-5772;
e-mail: msekine@bio.titech.ac.jp
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00342-5

1796
K. Shohda et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1795±1798
transient protection of the 20-hydroxyl group with the
trimethylsilyl group.15 The 20-O-methylation of the N-
benzoylated uridine derivative 3 was examined by using
CH3I±NaH or CH3I±Ag2O.16 When we used the former
conditions, 20-O-methylation of 3 was accomplished, but
the latter caused complex side reactions because Ag2O
activated simultaneously the 2-thiocarbonyl group of 3.
Finally, both the TIPDS and benzoyl groups of the 20-
O-methylated product 4 were removed in the usual
manner. Thus, the desired product 5 could be success-
fully synthesized in an overall yield of 49% from 1.17
The 50-O-dimethoxytritylation of 5 followed by 30-O-
phosphitylation gave the building block 7 required for
incorporation of s2Um into oligonucleotides.18 By using
this building block, the fully protected dimer 8 bearing
s2Um as a 50-upstream nucleoside was synthesized
(Scheme 1). Since it was reported that the thiocarbonyl
group of 2-thiothymidine reacted with I2,19 the phos-
phite intermediate was oxidized with t-butyl hydroper-
oxide in a manner similar to that described by Kumar et
al.6,20 The successive treatments of 8 with ammonium
hydroxide and 80% AcOH gave the dinucleoside mono-
phosphate 9 (s2UmpU) after reversed-phase HPLC puri-
®cation. In order to investigate the conformational
Figure 1. Match (left) and wobble (right) base pairs containing s2U.
The distance between the two methyl groups of 1-
methyl-2-thiouracil and 9-methylguanine was longer by
ca. 1 A than that of the corresponding U-G base pair.
This implies that the steric repulsion also contributes to
the duplex destabilization by s2U.
In this study, we developed a new route to s2Um and a
building block for incorporation of s2Um into oligonu-
cleotides. Furthermore, structural and thermodynamic
properties of s2Um and oligoribonucleotides having
1
s2Um were studied by means of H NMR spectroscopy
and Tm experiments.
Results and Discussion
1
properties of s2Um, the H NMR spectra of 5 and 9
Edmonds et al. ®rst reported that s2Um was obtained
by methylation of O2,50-anhydrouridine followed by a
ring-opening reaction with H2S.1 However, no details of
the 20-O-regioselectivity of the methylation and the yield
at each step were given. It seemed that it was dicult to
separate the 20-O-and 30-O-methyl-O2,50-anhydrouridine
derivatives unless HPLC was used. Therefore, we searched
for a more practical route to s2Um capable of large-
scale synthesis of this material required for the antisense
strategy. Consequently, all the reactive functional
groups of s2U were protected except for the 20-hydroxyl
group, and the successive 20-O-methylation was per-
formed, as depicted in Scheme 1.
were measured in sodium phosphate bu€er (pH 7.0).
The ratio of the C30-endo and C20-endo conformers of
s2Um was calculated by using the coupling constant
between the vicinal protons of 5 and 9, and these results
are listed in Table 1.21
The %N (C30-endo) value of 5 was calculated to be
70%, while that of 2-thiouridine (s2U) reported by
Sierzputowska±Gracz was 78%.2 In spite of the pre-
sence of the bulky 20-O-methyl group, the predominance
of the C30-endo conformer of 5 was unexpectedly almost
equal to that of s2U. On the other hand, the ribose
puckering of s2Um in the dimer 9 was almost C30-endo
only (3J1 H-2 H<1 Hz), although the predominance of
the C30-endo conformer of the 50-upstream nucleoside in
s2UpU was at most 90%.22 These results indicated that
the steric repulsion between the 2-thiocarbonyl and 20-
O-methyl groups of s2Um was more e€ective in the
0
0
First, the 30- and 50-hydroxyl groups of 2-thiouridine13
(1) were protected by use of the 1,1,3,3-tetraisopropyl-
disiloxane-1,3-diyl (TIPDS) group.14 Subsequently, the
base moiety of the resulting product 2 was acylated via
Scheme 1. Synthesis of s2Um and s2UmpU. Reagents and conditions: (a) TIPDSCl2 in pyridine, 99%; (b) i. HMDS in CH3CN, ii. BzCl, Bu4NBr in
CH2Cl2/Na2CO3 aq, iii. TFA in MeOH-CH2Cl2 (1:1, v/v), 76%; (c) NaH, CH3I in DMF, 79%; (d) i. Bu4NF, AcOH in THF, ii. Concd NH3±
pyridine (1:9, v/v), 83%; (e) DMTrCl in pyridine, 85%; (f) Chloro(2-cyanoethoxy)(N, N-diisopropylamino)phosphine, diisopropylethylamine in
CH2Cl2, 81% (purity 89%); (g) i. N3,20,30-O-Tribenzoyluridine, 1H-tetrazole in CH3CN, ii) t-BuOOH, 38%; (h) i. Concd NH3±pyridine (9:1, v/v), ii.
80% AcOH, 67%.

K. Shohda et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1795±1798
1797
Table 1. Vicinal coupling constants (3J11 H-22 H) and C30-endo preference of nucleosides
0
0
s2UmpU
s2UpU
UpU
s2Um
s2U
2.5a
U
50-Nu
Nu-30
50-Nu
Nu-30
50-Nu
Nu-30
3
4.6a
48
41.0
100
2.8
74
1.6b
91
3.5b
64
4.4b
51
3.7b
61
0
0
J1 H-2 H (Hz)
%N (C30-endo)c
3.1
70
78
aThese coupling constants were reported in ref. 2.
bThese data were taken from ref. 22.
cThese values were calculated according to the equation:%N (C30-endo)={(7.9 J1 H-2 H (Hz))/6.9}Â100.21
3
0
0
presence of the 30-phosphate group because of its steric
hindrance. Therefore, it was expected that the ribose
puckering of s2Um in oligonucleotides would also be
C30-endo. Interestingly, the C30-endo preference of s2Um
in¯uenced the sugar conformation of the 30-downstream
nucleoside, which became C30-endo. The %N (C30-endo)
value of the 30-nucleoside in the s2UmpU calculated was
74%, i.e., greater by 10% than that of s2UpU which
showed 64% C30-endo predominance.
matched duplex, but the ÁTm value became larger than
those of the RNA±RNA series (ÁTm=+6.9 ꢀC (s2Um),
and +9.4 ꢀC (s2U)). Although oligonucleotides modi®ed
with s2U or Um form thermodynamically stable A-type
RNA duplexes,6,7,10 the Tm values of both RNA±RNA
and RNA±DNA duplexes modi®ed by s2Um were
somewhat lower than those of the duplexes modi®ed by
s2U, and no additive e€ect of the double modi®cation
was observed.
Since the C30-endo preference of s2Um was clearly con-
®rmed, the stability of duplexes containing s2Um was
evaluated by the melting temperature (Tm) measurement.
An oligoribonucleotide nonamer (50-CGUUs2UmUU
GC-30) was synthesized by the usual phosphoramidite
method, except for the oxidation, similar to that descri-
bed in the synthesis of dimer 8. Subsequent deprotection
was performed by the routine procedure, and the result-
ing mixture was separated by anion exchange HPLC.23
For the control experiments, an oligoribonucleotide
having s2U instead of s2Um was obtained by Kumar's
method.6
The selectivity toward the target sequence is also an
important property of antisense agents. Particularly, the
discrimination of the U-G wobble base pair from the U-
A Watson±Crick base pair is essential for a highly spe-
ci®c antisense method. Therefore, the thermal stability
of RNA±RNA duplexes containing an s2Um-G wobble
pair was examined.
In order to evaluate the selectivity toward the target
sequence, di€erence in the Tm value between the duplex
having a s2Um-G and the duplex having a s2Um-A base
pair (=ÁTm*, see Table 2) was calculated. These
ÁTm* values were found to be 15.7 and 14.7 ꢀC, for
s2Um and s2U, respectively. However, the ÁTm* value
between the unmodi®ed duplexes was only 3.9 ꢀC.
Testa et al. also reported a similar trend for U-G and
s2U-G, but the absolute values of ÁTm* are smaller
than those in this study.10 Similar experiments were
performed in a series of RNA±DNA duplexes, and the
ÁTm* values were determined to be 20.4, 18.5, and
11.2 ꢀC for the duplexes having s2Um, s2U, and U,
respectively. In both the RNA±RNA and RNA±DNA
UV melting curves of RNA±RNA and RNA±DNA
duplexes were measured in 10 mM sodium phosphate
bu€er (pH 7.0) containing 150 mM NaCl, and these
pro®les and Tm values are shown in Fig. 2 and Table 2,
respectively. In the case of RNA±RNA duplexes, the
Tm value of the duplex containing either s2Um or s2U
was higher than that of the wild-type matched duplex
(ÁTm=+5.5 ꢀC (s2Um), and +6.9 ꢀC (s2U)). Such a
tendency was also found in the case of an RNA±DNA
Figure 2. Melting curves of CGUUU*UUGC/GCAAAAACG (RNA±RNA series, left) and CGUUU*UUGC/d(GCAAAAACG) (RNA±DNA
series, right).

1798
K. Shohda et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1795±1798
Table 2. Melting temperature (Tm Value, ꢀC) of RNA±RNA and RNA±DNA duplexes containing an s2Um-A (match) or s2Um-G (wobble) base
pair
GCAAXAACG
d(GCAAXAACG)
CGUUU*UUGC
X=A (match Tm value
(ÁTm*)a
X=G (wobble) Tm value
(ÁTm*)b
X=A (match) Tm Value
(ÁTm)a
X=G (wobble) Tm value
(ÁTm*)b
U*=U (control)
U*=s2Um
U*=s2U
36.7
42.2 (+5.5)
43.6 (+6.9)
32.8 ( 3.9)
26.5 ( 15.7)
28.9 ( 14.7)
18.6
25.5 (+6.9)
28.0 (+9.4)
7.4 ( 11.2)
5.1 ( 20.4)
9.5 ( 18.5)
aÁTm=(Tm value of a modi®ed duplex) (that of a wild-type duplex).
bÁTm*=(Tm value of a duplex having a U*-G wobble base pair) (that of a duplex having a matched U*-A base pair in place of a U*-G base pair).
duplexes having s2U the ÁTm* values of the ones
7. (a) Lesnik, E. A.; Guinosso, C. J.; Kawasaki, A. M.; Sas-
mor, H.; Zounes, M.; Cummins, L. L.; Ecker, D. J.; Cook, P.
modi®ed by s2U derivatives were always larger than
those of the unmodi®ed, and the ÁTm* value of s2Um-
D.; Freier, S. M. Biochemistry 1993, 32, 7832. (b) Cummins, L.
L.; Owen, S. R.; Risen, L. M.; Lesnik, E. A.; Freier, S. M.;
containing duplexes was fortunately somewhat larger
McGee, D.; Guinosso, C. J.; Cook, P. D. Nucleic Acids Res.
1995, 23, 2019.
than those having s2U.
8. (a) Crooke, S. T., Lebleu, B., Eds., Antisense Research and
Applications; CRC Press: Boca Raton, 1993. (b) Crooke, S. T.,
Ed.; Antisense Research and Application; Springer-Verlag:
Berlin, 1998.
9. Sugimoto, N.; Kierzek, R.; Freier, S. M.; Turner, D. H.
Biochemistry 1986, 25, 5755.
10. Testa, S. M.; Disney, M. D.; Turner, D. H.; Kierzek, R.
Biochemistry 1999, 38, 16655.
11. (a) Sponer, J.; Leszczynski, J.; Hobza, P. J. Biomol.
Struct. Dyn. 1996, 14, 117. (b) Hobza, P.; Sponer, J. Chem.
Rev. 1999, 99, 3247.
12. The hydrogen bonding energy was estimated as the di€er-
ence between the energy of complex (base pair) and energies of
isolated nucleobases. The geometries and energies of all the
systems were calculated at the HF/6-31G* level.
13. (a) Vorbruggen, H.; Strehlke, P. Chem. Ber. 1973, 106,
3039. (b) Niedballa, U.; Vorbruggen, H. J. Org. Chem., 1974,
39, 3654.
The duplex destabilization by the s2Um (or s2U)-G base
pair was a consequence of both the weaker hydrogen
bonding nature and the steric hindrance of the 2-thio-
carbonyl group, as expected by ab initio calculation.
However, the reason why the ÁTm* values of the duplexes
containing s2Um were larger than those of the duplex
containing s2U is not clear. Since the e€ect of the uri-
dine-2-thiolation on the stabilization and selectivity
enhancement of RNA duplexes is excellent, we propose
that several or all uracil residues of the antisense agents
should be replaced by 2-thiouracil, if such replacement
is chemically and economically possible.24 Therefore,
antisense agents having 2-thiouracil residues instead of
uracil will be one of the second-generation antisense
drugs. Further studies are under way in this direction.
14. Markiewics, W. T. J. Chem. Research (S) 1979, 24.; J.
Chem. Research (M) 1979, 0181.
Acknowledgements
15. Sekine, M. J. Org. Chem. 1989, 54, 2321.
16. (a) Kamimura, T.; Masegi, T.; Hata, T. Chem. Lett. 1982,
965. (b) Inoue, H.; Hayase, Y.; Imura, A.; Iwai, S.; Miura, K.,
Ohtsuka, E. Nucleic Acids Res. 1987, 6131
This work was supported by a Grant from ``Research
for the Future'' Program of the Japan Society for the
Promotion of Science (JSPS-RFTF97I00301) and a
Grant-in-Aid for Scienti®c Research from the Ministry
of Education, Science, Sports and Culture, Japan.
17. Spectral data of 5: 1H NMR (400 MHz, D2O) d 3.60 (3H, s,
20-O-CH3), 3.83 (1H, dd, J4 H-5 H=3.8Hz, J5 H-5 H=13.1 Hz, 50-
0
00
0
00
H), 3.96 (1H, dd, J4 H-5 H=2.5 Hz, 500-H), 4.06 (1H, dd, J1 H-
0
00
0
2 H=3.100 Hz, J2 H-3 H=5.000 Hz, 20-H), 4.12 (1H, m, 40-H), 4.30
0
0
0
(1H, dd, J3 H-4 H=6.9 Hz, 30-H), 6.15 (1H, d, J5H-6H=8.1 Hz, 5-
0
0
H), 6.76 (1H, d, 10-H), 8.08 (1H, d, 6-H).
18. (a) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett.
1981, 22, 1859. (b) Sinha, N. D.; Biernat, J.; Koster, H. Tet-
rahedron Lett. 1983, 24, 5843.
References and Notes
1. Edmonds, C. G.; Crain, P. F.; Hashizume, T.; Gupta, R.;
Stetter, K. O.; McCloskey, J. A. J. Chem. Soc., Chem. Com-
mun. 1987, 909.
2. Sierzputowska-Gracz, H.; Sochacka, E.; Malkiewicz, A.;
Kuo, K.; Gehrke, C. W.; Agris, P. F. J. Am. Chem. Soc. 1987,
109, 7171.
3. Davies, D. B. In Progress in NMR Spectroscopy; Emsley, J.
W.; Feenly, J.; Sutcli€e, L. H., Eds.; Pergamon: New York,
1978; pp 135±225.
4. Sakamoto, K.; Kawai, G.; Watanabe, S.; Niimi, T.; Haya-
shi, N.; Muto, Y.; Watanabe, K.; Satoh, T.; Sekine, M.;
Yokoyama, S. Biochemisty 1996, 35, 6533.
19. Kuimelis, R. G.; Nambiar, K. P. Tetrahedron Lett. 1993,
34, 3813.
20. Kumar, R. K.; Davis, D. R. J. Org. Chem. 1995, 60, 7726.
21. (a) de Leeuw, F. A. A. M.; Altona, C. J. Chem. Soc.,
Perkin Trans. 2 1982, 375. (b) Altona, C. Recl. Trav. Chim.
Pays-Bas, 1982, 101, 413. (c) Rousse, B.; Puri, N.; Viswanad-
ham, P.; Agback, P.; Glemarec, C.; Sandstrom, A; Sund, C.;
Chattopadhyaya, J. Tetrahedron 1994, 50, 1777.
22. Smith, W.; Sierzputowska-Gracz, H.; Sochacka, E.;
Malkiewicz, A.; Agris, P. F. J. Am. Chem. Soc. 1992, 114,
7989.
23. The RNA nonamer (50-CGUUs2UmUUGC-30) was char-
acterized by MALDI-TOF-Mass spectroscopy (calcd. for
C84H105N26O65P8S (M H ): 2797.33; Found: 2797.20).
24. Lohse, J.; Dahl, O.; Nielsen, P. E. Proc. Natl. Acad. Sci.
USA 1999, 96, 11804.
5. Seanger, W. In Principles of Nucleic Acid Structure
Springer-Verlag: Berlin, 1981; Chapter 10.
6. (a) Kumar, R. K.; Davis, D. R. Nucleic Acids Res. 1997, 25,
1272. (b) Idem. Nucleosides Nucleotides 1997, 16, 1469.