Dinucleotides of 4′-C-vinyl- and 5′-C-allylthymidine as substrates for ring-closing metathesis reactions

Article abstract of DOI:10.1016/S0040-4039(03)01584-3

Thymidine derivatives containing a 4′-C-vinyl group or a 5′-C-allyl group are synthesized and used as building blocks for three different dinucleotides. These are evaluated as substrates for ring-closing metathesis cyclisations, and a protected 5′-C-allyl

Full text of DOI:10.1016/S0040-4039(03)01584-3

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 6475–6478
Dinucleotides of 4%-C-vinyl- and 5%-C-allylthymidine as substrates
for ring-closing metathesis reactions
Claus Kirchhoff and Poul Nielsen*
Nucleic Acid Center,† Department of Chemistry, University of Southern Denmark, 5230 Odense M, Denmark
Received 20 May 2003; revised 13 June 2003; accepted 20 June 2003
Abstract—Thymidine derivatives containing a 4%-C-vinyl group or a 5%-C-allyl group are synthesized and used as building blocks
for three different dinucleotides. These are evaluated as substrates for ring-closing metathesis cyclisations, and a protected
5%-C-allylthymidine homo-dimer is found to be the most reactive. A protected precursor for a conformationally restricted cyclic
dinucleotide with a four carbon 5%-C to 5%-C connection is hereby efficiently obtained, whereas a corresponding three carbon 4%-C
to 5%-C connection is obtained in a lower yield.
© 2003 Elsevier Ltd. All rights reserved.
The construction of conformationally restricted nucleic
acid fragments is a powerful tool towards potential
therapeutics and diagnostics.1,2 Thus, oligonucleotides
containing nucleoside monomers with restricted carbo-
hydrate moieties have demonstrated high affinity for
complementary nucleic acids.2,3 On the other hand,
conformationally restricted models of the secondary or
tertiary structural elements found especially in RNA4
has gained much less attention, though a few restricted
dinucleotides have been synthesised and investigated,5,6
e.g. as a model of the anticodon loop in a bacterial
tRNA.7
high basic lability.12 The latter problem, however, has
been significantly reduced by saturation of the allylic
moiety using a tandem RCM-hydrogenation protocol.12
In order to obtain even more stable and convenient
molecules, cyclic dinucleotides with an intact, charged
and achiral, phosphate internucleoside linkage were
approached. Thus, we decided to perform dinucleotide
couplings of nucleoside monomers on which terminal
double bonds are placed on the 4%-C and/or on the 5%-C
positions, and subsequently use these dinucleotides as
substrates for RCM-reactions. Thus,
a
4%-C-
vinylthymidine derivative should be conveniently
obtained,13–15 and in contrary to 5%-C-vinylthymidine
which we applied in a former study,10 the 5%(S)-epimer
of a 5%-C-allylthymidine derivative can be obtained by a
completely stereoselective method.16 Herein we present
the preparation of the appropriately protected
nucleoside monomers, their coupling into dinucleotides
as well as the properties of these as substrates for RCM
cyclisation reactions.
We have recently applied the ring-closing metathesis
(RCM) methodology8 using the ruthenium based pre-
catalyst A (Scheme 1) developed by Grubbs and co-
workers9 in the construction of conformationally
restricted dinucleotide10–12 and trinucleotide struc-
tures.11 Thus, nucleoside monomers with an allyl sub-
stituent either at the 5-position of a uracil or at a
3%-O-phosphoramidite moiety,11,12 or with a 5%-C-vinyl
substituent10 have been coupled to afford di- or trinu-
cleotides. Subsequently, medium to large unsaturated
rings have been formed by RCM in medium or high
yields.10–12 Nevertheless, the cyclic di- or trinucleotides
formed involves phosphotriester internucleoside
linkages10–12 revealing two problems; (1) the chirality of
the phosphorus leads to several isomeric products and
(2) the allylic phosphotriester moiety demonstrates a
The 4%-C-hydroxymethyl thymidine derivative 1 was
obtained from thymidine in five steps as published
before.17 Among the two primary hydroxyl groups the
pro-(S) position has been shown to be the more
reactive17,18 and after tritylation of this, the other
hydroxyl group was silylated to give after detritylation
either 219 or 320 as the major products in reasonable
yields (Scheme 1). Oxidation of the alcohol in 2 or 3
and subsequent Wittig methylenation21 afforded the
two 4%-C-vinyl thymidine derivatives 4 and 514 in 51
†
Nucleic Acid Center is funded by the Danish National Research
Foundation for studies on nucleic acid chemical biology.
* Corresponding author. E-mail: pon@chem.sdu.dk
0040-4039/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)01584-3

6476
C. Kirchhoff, P. Nielsen / Tetrahedron Letters 44 (2003) 6475–6478
and 85% yield, respectively. The high yield of 5 com-
pared to 4 might be ascribed to the higher basic stabil-
ity of a primary TBDPS-ether.
low solubility of 12 in acetonitrile. In an appropriate
mixture of acetonitrile and dichloromethane, the cou-
pling reaction succeeded though never completely and
14 was isolated in 25% yield as an approximate 2:1
ratio of epimers as indicated from 31P NMR. The
dinucleotide 14 was finally subjected to 5 mol% of the
precatalyst A in dichloromethane at 40°C, and this time
a product was observed. After treatment with an addi-
tional amount of A the RCM cyclisation was completed
and the product 15 isolated in good yield as a mixture
of four isomeric products as clearly demonstrated from
the 31P NMR spectrum.27 Thus, both E- and Z-isomers
of both epimeric phosphotriesters were found, but the
precise elucidation of the observed approx. 10:5:4:2
ratio was not performed. The RCM reaction did even-
tually succeed also with a lower 2×2 mol% loading of
A.
In order to obtain the two appropriately mono-pro-
tected 4%-C-vinyl nucleosides, selective desilylations of 4
and 5 were approached. Thus, the primary silyl ether of
the bis-TBDMS protected compound 4 was selectively
hydrolysed under acidic conditions to give, conve-
niently, 6 in high yield. Also, the secondary silyl ether
of 5 could be cleaved selectively applying TMS-triflate22
to afford 7. However, as the Wittig reaction leading to
5 was the more efficient, we attempted also to obtain
the 3%-O-protected building block 6 from 5 using meth-
ods published to cleave selectively TBDPS-ethers over
TBDMS-ethers.23 Nevertheless, this conversion was not
possible in our hands, and in conclusion, 6 was most
conveniently produced via 4. Finally, 7 was also
obtained from 4 using full deprotection24 and subse-
quent selective TBDPS-protection of the primary 5%-
hydroxyl functionality.
Phosphitylation of 6 using dicyanoimidazole as the
activating reagent25 gave in high yield the phospho-
ramidite 8 as a mixture of two epimers as judged from
31P NMR. A standard nucleotide coupling reaction26 of
the alcohol 7 and the phosphoramidite 8 afforded the
dinucleotide 9 after oxidation.27 Due to the chiral phos-
photriester moiety, the dinucleotide contains two
epimers in almost equimolar yields as judged from 31P
NMR. Finally, 9 was investigated as a substrate for an
RCM cyclisation. Thus, 9 was treated with 5 mol% of
the precatalyst A in dichloromethane at 40°C. How-
ever, even though additional amounts of the precatalyst
were added, no reaction was detected and after 48 h,
the starting material 9, completely unchanged as shown
by MS and NMR, was almost quantitatively recovered.
Similar attempts using solutions in dichloroethane at
80°C or in toluene at 100°C gave the same disappoint-
ing result. Steric hindrance of the vinyl moieties due to
their ‘neopentylic’ positions might explain the low reac-
tivity with the catalyst, and therefore, we turned our
attention towards
analogue.
a
5%-C-substituted nucleoside
The introduction of a 5%-C-allyl moiety on thymidine
derivatives has been accomplished from a 3%-O-pro-
tected 5%-aldehyde derivative by a conventional Grig-
nard reaction affording both possible epimers28 as well
as by a completely stereoselective allylation using
allyltrimethylsilane and BF3·OEt2 as a Lewis acid cata-
lyst.16 We decided to follow the latter strategy and
obtained the 5%(S)-isomer 10 in five steps from
thymidine.16 In order to obtain an appropriately 5%-O-
protected derivative, 10 was esterified to give 11 which
after conventional desilylation afforded the 5%-O-benz-
oyl protected compound 12 (Scheme 2).
Scheme 1. Reagents and conditions: (a) (i) DMT-Cl, DMAP,
pyr, rt, (ii) TBDMS-Cl, imidazole, DMF, rt, (iii) TFA,
CH2Cl2, CHCl3, Et3SiH, rt, 55% (three steps); (b) (Ref. 20) (i)
DMT-Cl, DMAP, pyr, rt, (ii) TBDPS-Cl, imidazole, DMF,
rt, (iii) 80% aq. AcOH, THF, rt, 63% (three steps); (c) (i)
Dess–Martin periodinane, CH2Cl2, rt, (ii) Ph3PCH3Br, n-
BuLi, hexane, THF, 0°C, 51% 4 or 85% 5 (two steps); (d) 80%
aq. AcOH, rt, 90%; (e) TMS-OTf, CH2Cl2, −30°C, 73%; (f) (i)
90% aq. TFA, rt, (ii) TBDPS-Cl, pyr, rt, 82% (two steps); (g)
(iPr2N)2PO(CH2)2CN, 4,5-dicyanoimidazole, CH2Cl2, rt,
98%; (h) 1H-tetrazole, CH3CN, rt, then t-BuOOH, toluene,
0°C, 68%; (j) A, see text, 0%. DMT=dimethoxytrityl,
Phosphitylation of 10 applying the same conditions as
before with 6 afforded the phosphoramidite 13, and
subsequently, a standard nucleotide coupling of 12 and
13 followed by oxidation afforded the dinucleotide 14.27
This coupling reaction, however, was hampered by the
TBDMS=tert-butyldimethylsilyl,
TBDPS=tert-butyl-
diphenylsilyl, CE=cyanoethyl, T=thymin-1-yl.

C. Kirchhoff, P. Nielsen / Tetrahedron Letters 44 (2003) 6475–6478
6477
treated with 5 mol% of the precatalyst A in dichloro-
methane at 40°C but no reaction was detected. When
the reaction was run in dichloroethane at 80°C with
several consecutive additions of A, up to a total loading
of 15 mol%, an RCM cyclisation did proceed to give
after 5 days, and a chromatographic purification, 38%
of the product 17 and 50% of recovered starting mate-
rial 16. The product contains, as indicated from NMR,
one major product as well as smaller amounts of the
other three possible isomers. This might be due to one
of the two phosphotriester epimers being more
favourably cyclised than the other, in combination with
the fact that the E-configuration of the double bond
might be much less favourable than the Z-configuration
in the nine-membered ring. The latter suggestion seems
probable from simple modelling.
In summary, the efficiency of the three dinucleotides as
substrates for RCM cyclisation falls in the order 14>
16>9. Thus, the 5%-C-allyl group is much more reactive
towards the catalyst than the more sterically hindered
4%-C-vinyl group, which in 9 does not react at all. On
the other hand, the 4%-C-vinyl group can, in fact, react
with the catalyst in RCM-reactions as reported recently
for a nucleoside substrate giving efficiently a six-mem-
bered ring15 and now with 16 affording a larger ring.
Apparently, the catalyst reacts first with the 5%-C-allyl
group in 16 (or alternatively a 3%-O-allylgroup)15 and
next in an intramolecular reaction with the hindered
4%-C-vinyl group.
Scheme 2. Reagents and conditions: (a) BzCl, pyr, rt, 71%; (b)
TBAF, THF, rt, 63%; (c) (iPr2N)2PO(CH2)2CN, 4,5-
dicyanoimidazole, CH2Cl2, rt, 77%; (d) 1H-tetrazole,
CH3CN, CH2Cl2, rt, then t-BuOOH, toluene, 0°C, 25%; (e) 4
mol% A, CH2Cl2, 40°C, 79%. TBDPS=tert-butyldiphenyl-
silyl, CE=cyanoethyl, T=thymin-1-yl.
Following these observations, we conclude that the
dinucleotide of 5%-C-allylnucleosides is the best choice
for the future construction of conformationally
restricted cyclic dinucleotides and subsequent oligonu-
cleotides. However, the dinucleotide 14 was, unfortu-
nately, synthesised from the less efficient coupling
reaction. Nevertheless, we believe that these problems
are due to the choice of a benzoyl ester as the 5%-O-pro-
tecting group, and for future examinations, we plan to
use alternatively protected analogues of 14. The
efficient RCM reaction should not be compromised by
the use of other protecting groups. Also, the tandem
RCM-hydrogenation protocol, to give in situ, a satu-
rated ring is expected from our former investigations12
to be straightforward.
In order to investigate the possibility of an RCM
cyclisation of a mixed dinucleotide containing the 4%-C-
vinyl nucleoside monomer, the alcohol 7 was coupled to
the phosphoramidite 13 using the same reaction condi-
tions as before (Scheme 3). The following dinucleotide
16 was easily obtained in high yield and an approxi-
mate 3:1 mixture of epimers.27 The dinucleotide was
In conclusion, we have proved that nucleoside
monomers containing 4%-C-vinyl and 5%-C-allyl groups
can be readily obtained and used in the preparation of
dinucleotides. Furthermore, we have found that a dinu-
cleotide containing two 5%-C-allyl groups is an efficient
substrate for RCM reactions, whereas dinucleotides
containing one or two 4%-C-vinyl groups are less
efficient. Further investigations towards cyclic dinucle-
otides with a four-carbon 5%-C to 5%-C connection and
oligonucleotides containing this conformationally
restricting nucleic acid fragment is in progress in our
laboratory. We expect this and other cyclic dinucleotide
structures to be attractive building blocks towards
nucleic acid secondary and tertiary structural mimics,
and we expect the present RCM strategy to be an
important tool in nucleic acid chemical biology and
drug development.
Scheme 3. Reagents and conditions: (a) 1H-tetrazole, CH3CN,
CH2Cl2, rt, then t-BuOOH, toluene, 0°C, 89%; (b) 15 mol%
A, ClCH2CH2Cl, 80°C, 38%. TBDPS=tert-butyldiphenyl-
silyl, CE=cyanoethyl, T=thymin-1-yl.

6478
C. Kirchhoff, P. Nielsen / Tetrahedron Letters 44 (2003) 6475–6478
Acknowledgements
15. Recently,
3%-O-allyl-5%-O-TBDPS-4%-C-vinylthymidine
has been prepared and used as a substrate for an RCM-
reaction affording a 3%-O,4%-C-bicyclic nucleoside deriva-
tive: Montembault, M.; Bourgougnon, N.; Lebreton, J.
Tetrahedron Lett. 2002, 43, 8091.
We thank the Danish National Research Foundation
for financial support.
16. Banuls, V.; Escudier, J.-M. Tetrahedron 1999, 55, 5831.
17. (a) O-Yang, C.; Wu, H. Y.; Fraser-Smith, E. B.; Walker,
K. A. M. Tetrahedron Lett. 1992, 33, 37; (b) Thrane, H.;
Fensholdt, J.; Regner, M.; Wengel, J. Tetrahedron 1995,
37, 10389.
18. O-Yang, C.; Kurz, W.; Eugui, E. M.; McRoberts, M. J.;
Verheyden, J. P. H.; Kurz, L. J.; Walker, K. A. M.
Tetrahedron Lett. 1992, 33, 41.
References
1. Kool, E. T. Chem. Rev. 1997, 97, 1473.
2. Leumann, C. J. Bioorg. Med. Chem. 2002, 10, 841.
3. Meldgaard, M.; Wengel, J. J. Chem. Soc., Perkin Trans.
1 2000, 3539.
19. Compound 2 has been obtained before by different pro-
tecting group manipulations: Wang, G.; Seifert, W. E.
Tetrahedron Lett. 1996, 37, 6515.
20. Marx, A.; Erdmann, P.; Senn, M.; Ko¨rner, S.; Jungo, T.;
Petretta, M.; Imwinkelried, P.; Dussy, A.; Kulicke, K. J.;
Macko, L.; Zehnder, M.; Giese, B. Helv. Chim. Acta
1996, 79, 1980.
4. Belmont, P.; Constant, J.-F.; Demeunynck, M. Chem.
Soc. Rev. 2001, 30, 70.
5. (a) Sekine, M.; Kurasawa, O.; Shohda, K.; Seio, K.;
Wada, T. Eur. J. Org. Chem. 2001, 1989; (b) Sekine, M.;
Kurasawa, O.; Shohda, K.; Seio, K.; Wada, T. J. Org.
Chem. 2000, 65, 6515.
6. Le Cle´zio, I.; Escudier, J.-M.; Vigroux, A. Org. Lett.
21. A different Wittig olefination performed after oxidation
of 1 and 5%-O-tritylation has been reported: Maag, H.;
Schmidt, B.; Rose, S. J. Tetrahedron Lett. 1994, 35, 6449.
22. Bou, V.; Vilarrasa, J. Tetrahedron Lett. 1990, 31, 567.
23. Higashibayashi, S.; Shinko, K.; Ishizu, T.; Hashimoto,
K.; Shirahama, H.; Nakata, M. Synlett 2000, 9, 1306.
24. The unprotected 4%-C-vinylthymidine was fully character-
ized and the data were in accordance with data given for
the compound obtained by a different strategy; Ref. 13a.
25. Vargeese, C.; Carter, J.; Yegge, J.; Krivjansky, S.; Settle,
A.; Kropp, E.; Peterson, K.; Pieken, W. Nucleic Acid Res.
1998, 26, 1046.
2003, 5, 161.
7. Seio, K.; Wada, T.; Sekine, M. Helv. Chim. Acta 2000,
83, 162.
8. For a review, see: Fu¨rstner, A. Angew. Chem., Int. Ed.
2000, 39, 3012.
9. Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34,
18.
10. (a) Sørensen, A. M.; Nielsen, P. Org. Lett. 2000, 2, 4217;
(b) Sørensen, A. M.; Nielsen, K. E.; Vogg, B.; Jacobsen,
J. P.; Nielsen, P. Tetrahedron 2001, 57, 10191.
11. Børsting, P.; Sørensen, A. M.; Nielsen, P. Synthesis 2002,
797.
26. Caruthers, M. H. Acc. Chem. Res. 1991, 55, 5831.
27. All dinucleotides 9 and 14–17 were isolated as isomeric
12. Børsting, P.; Nielsen, P. Chem. Commun. 2002, 2140.
13. Preparation of 4%-C-vinylthymidine has been accom-
plished from a radical-cyclisation reaction with a 3%-O-
vinylsilyl tether: (a) Sugimoto, I.; Shuto, S.; Mori, S.;
Shigeta, S.; Matsuda, A. Bioorg. Med. Chem. Lett. 1999,
9, 385; (b) Sugimoto, I.; Shuto, S.; Matsuda, A. J. Org.
Chem. 1999, 64, 7153.
14. Recently, a preparation similar to ours of the 4%-C-
vinylthymidine derivative 5 was presented: (a) Summerer,
D.; Marx, A. Angew. Chem., Int. Ed. 2001, 40, 3693; (b)
Detmer, I.; Summerer, D.; Marx, A. Eur. J. Org. Chem.
2003, 1837.
1
mixtures and characterized by H NMR, 31P NMR and
HRMS. Selected data; 31P NMR lP (CDCl3, 121.5 MHz
with 85% H3PO4 as external standard): 9 −1.92, −1.64; 14
−1.81, −1.56; 15 −7.74, −6.07, −0.25, 0.69; 16 −1.79,
−1.22; 17 (major product) −0.48. HR MALDI FT-MS
m/z [M+Na+] (found/calcd): 9 (1026.3906, 1026.3876); 14
(1044.3672, 1044.3587); 15 (1016.3289, 1016.3274); 16
(1164.4368, 1164.4346); 17 (1136.4033, 1135.4067).
28. Wang, G.; Middleton, P. J. Tetrahedron Lett. 1996, 37,
2739 and Nucleosides Nucleotides 1998, 17, 1033.