Synthesis of a novel conformationally locked carbocyclic nucleoside ring system.

Article abstract of DOI:10.1021/ol034326z

[reaction: see text] A fast and efficient synthetic route to novel Northern locked carbocyclic nucleosides (as precursors of carbocyclic locked nucleic acids or cLNAs) is described. The target nucleoside with a oxabicyclo[2.2.1]heptane ring system was pre

Full text of DOI:10.1021/ol034326z

ORGANIC
LETTERS
2003
Vol. 5, No. 10
1665-1668
Synthesis of a Novel Conformationally
Locked Carbocyclic Nucleoside Ring
System
,†
Hak Sung Kim†,‡ and Kenneth A. Jacobson*
Molecular Recognition Section, Laboratory of Bioorganic Chemistry, National Institute
of Diabetes & DigestiVe & Kidney Diseases, National Institutes of Health, Bethesda,
Maryland 20892, and College of Pharmacy and Medicinal Resources Research Center,
Wonkwang UniVersity, Iksan, 570-749, Chonbuk, South Korea
kajacobs@helix.nih.goV
Received February 24, 2003
ABSTRACT
A fast and efficient synthetic route to novel Northern locked carbocyclic nucleosides (as precursors of carbocyclic locked nucleic acids or
cLNAs) is described. The target nucleoside with a oxabicyclo[2.2.1]heptane ring system was prepared from a simple starting material, diethyl
malonate. Ring closure by intramolecular O-alkylation provided the target ring system as the major isomer over the [3.2.0] oxetane system.
The adenine moiety was introduced through a reactive triflate after inversion of the stereochemistry of the corresponding alcohol by oxidation
and reduction.
In general, the ribose rings of nucleosides and nucleotides
may adopt a range of conformations; this has been described
as a pseudorotational cycle.1 The Northern ((N), 2′-exo, 3′-
endo) and Southern ((S), 2′-endo, 3′-exo) conformations are
the most relevant to the biological activities observed for
nucleosides and nucleotides in association with DNA, RNA,
and various enzymes.1-6
analogues.2,3 A methanocarba approach introduced by Mar-
quez and co-workers4,5 was used to constrain the pseudosugar
(cyclopentane) ring of the nucleoside to either an (N) or (S)
conformation. In this approach, a cyclopropane moiety is
fused to the cyclopentane ring at one of two positions. Such
analogues helped to define the role of sugar puckering in
stabilizing the active receptor-bound conformation and
thereby allowed identification of a favored isomer, although
the possible influence of the cyclopropyl methylene group
on bioactivity distinct from the conformational effect has
been discussed.2 A preference for the (N) conformation of
ribose both at the adenosine A3 receptor and at the P2Y1
nucleotide receptor was defined by using methanocarba
analogues.2,3 For example, the bisphosphate derivative 1,
MRS 2279, which is locked in a (N) conformation by the
bicyclo[3.1.0]hexane ring system, is the most potent known
antagonist of the P2Y1 receptor.6 Subsequently, we discov-
ered that P2Y2, P2Y4, and P2Y11 receptors can recognize
(N)-methanocarba nucleotide triphosphates 2 and 3 roughly
as potently as the corresponding ribosides.3 However, in the
case of P2Y6 receptors, the (N)-methanocarba equivalent 4
In studies of the ribose ring conformation while searching
for improved ligands for binding to G protein-coupled
receptors, we synthesized ring-constrained carbocyclic-type
† National Institutes of Health.
‡ Wonkwang University.
(1) Altona, C.; Sundaralingham, M. J. Am. Chem. Soc. 1972, 94, 8205.
(2) Jacobson, K. A.; Ji, X.-d.; Li, A. H.; Melman, N.; Siddiqi, M. A.;
Shin, K. J.; Marquez, V. E.; Ravi, R. G. J. Med. Chem. 2000, 43, 2196.
(3) Kim, H. S.; Ravi, R. G.; Marquez, V. E.; Maddileti, S.; Wihlborg,
A.-K.; Erlinge, D.; Malmsjo¨, M.; Harden, T. K.; Boyer, J. L.; Jacobson, K.
A. J. Med. Chem. 2002, 45, 208.
(4) Marquez, V. E.; Siddiqui, M. A.; Ezzitouni, A.; Russ, P.; Wang, J.;
Wagner, R. W.; Matteucci, M. D. J. Med. Chem. 1996, 39, 3739.
(5) Shin, K. J.; Moon, H. R.; George C.; Marquez V. E. J. Org. Chem.
2000, 65, 2172.
(6) Boyer, J. L.; Adams, M.; Ravi, R. G.; Jacobson, K. A.; Harden, T.
K. Br. J. Pharmacol. 2002, 135, 2004.
10.1021/ol034326z This article not subject to U.S. Copyright. Published 2003 by the American Chemical Society
Published on Web 04/18/2003

of the natural nucleotide activator, UDP, was inactive.3 This
prompted us to search for other (N)-like ring systems, which
may adopt a slightly different conformation according to the
pseudorotational cycle (Figure 1).
Scheme 1
key starting material, diol 9, was prepared according to a
ring closure reaction with Grubbs catalyst10 and a known
procedure11 from a commercially available starting material,
diethyl diallyl malonate 6. During the process, ring opening
of a meso-epoxide 8 by the chiral base (1S,2R)-norephedrine
provided an allyl alcohol 9 containing a chiral quaternary
carbon, which is usually elusive synthetically. Protection of
diol 9 with the tert-butyldiphenylsilyl group and successive
treatment with sequential dihydroxylation12 and debenzyla-
tion gave an intermediate triol, which was converted into
mesyl carbonate 10 by sequential treatment with triphosgene
and mesyl chloride. Deprotection of the cyclic carbonate
group and intramolecular O-alkylation could be accomplished
in one pot in basic conditions (K2CO3 in methanol), yielding
two isomers 11and 12 in a ratio of 2:1, which were separated
by column chromatography.
Figure 1.
Following their introduction several years ago, the locked
nucleic acids 5 (LNAs)7-9 have demonstrated excellent
mismatch discrimination toward complementary RNA and
have been shown to form more thermally stable complexes.
The derivatization of LNAs is fully compatible with con-
ventional DNA/RNA chemistry. Several synthetic routes to
ring oxygen-containing LNAs were developed.7-9
Here we present the first synthetic route to nucleoside
monomers of carbocyclic LNAs (cLNAs), in which the ring
oxygen is replaced with a methylene group. As shown in
the synthetic Scheme 2, the coupling of pseudosugar and
base moieties could be accomplished at a late stage within
this route. This would permit a common route to lead
efficiently to the preparation of many nucleosides and
nucleotides. Also, cLNAs would be more stable than oxygen
LNAs because of their nonglycosidic nature, as has been
discussed for methanocarba ribose in comparison with natural
ribose.2 These potential advantages encouraged us to devise
a synthetic route to novel cLNA monomers of the (N)
conformation. We anticipate that this synthesis will be
applicable to nucleoside and nucleotide chemistry, leading
to the development of new G protein-coupled receptor
ligands, new antisense oligonucleotide variations, and other
RNA targeting strategies.
At the stage of isolation of the two isomers, 1D and 2D
proton NMR and NOE spectra were collected for 11 and
12, and all protons were carefully analyzed and assigned.13
In the spectrum of isomer 12, there was axial-axial coupling
between H-3 and H-4 and between H-3 and the axial proton
of two H-2s, which was consistent with isomer 12 being in
(10) (a) Nguyen, S. T.; Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. J.
Am. Chem. Soc. 1992, 114, 3974. (b) Grubbs, R. H.; Miller, S. J. Acc. Chem.
Res. 1995, 28, 446. (c) Schmalz, H.-G. Angew. Chem., Int. Ed. Engl. 1995,
34, 1833. (d) Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413.
(11) Hodgson, D. M.; Gibbs, A. R.; Drew, M. G. B. J. Chem. Soc., Perkin
Trans. 1 1999, 24, 3579.
(12) (a) Alex, G.; Kruger, A. W.; Meyers, A. I. Tetrahedron Lett. 2001,
42, 4305. (b) Arrington, M. P.; Meyers, A. I. Chem. Commun. 1999, 15,
1371.
(13) Data for 11: 1H NMR (CD3OD) δ 7.34-7.71 (m, 20H, aromatic
protons on TBDPS groups), 4.36 (s, 1H, H-7), 3.86 (d, 1H, J ) 10.4 Hz,
one of CH2-OTBDPS), 3.85 (ddd, 1H, J ) 1.7, 2.8, 7.7 Hz, H-6), 3.72 (d,
1H, J ) 1.7 Hz, H-1), 3.61 (d, 1H, J ) 10.4 Hz, one of CH2-OTBDPS),
3.61 (dd, 1H, J ) 2.8, 6.3 Hz, one of two H-3s), 3.18 (d, 1H, J ) 6.3 Hz,
one of two H-3s), 1.84 (ddd, 1H, J ) 2.7, 2.8, 13.2 Hz, one of two H-5s),
1.66 (dd, 1H, J ) 7.7, 13.2 Hz, onf of two H-5s), 1.06 (s, 18H, tert-butyl
of TBDPS). Data for 12: 1H NMR (CD3OD) δ 7.26-7.83 (m, 20H,
aromatic protons on TBDPS groups), 4.66 (ddd, 1H, J ) 6.6, 8.0, 9.9 Hz,
H-3), 4.60 (d, 1H, J ) 4.1 Hz, H-5), 4.41 (dd, 1H, J ) 1.1, 6.0 Hz, one of
two H-7s), 3.79 (dd, 1H, J ) 4.1, 8.0 Hz, H-4), 3.76 (d, 1H, J ) 6.0 Hz,
one of two H-7s), 3.71 (d, 1H, J ) 10.4 Hz, one of CH2-OTBDPS), 3.55
(d, 1H, J ) 10.4 Hz, CH2-OTBDPS), 1.62 (ddd, 1H, J ) 1.1, 9.9, 12.9 Hz,
one of two H-2s), 1.44 (dd, 1H, J ) 6.9, 12.9 Hz, one of two H-2s), 1.09
(s, 9H, t-Bu on TBDPS), 1.05 (s, 9H, t-Bu on TBDPS).
Initially, our primary interest was the formation of the
bicyclo[2.2.1]heptane system. As shown in Scheme 1, the
(7) (a) Obika, S.; Nanbu, D.; Fari, Y.; Morio, K.-I.; In, Y.; Ishida, T.;
Imanishi, T. Tetrahedron Lett. 1997, 38, 8735. (b) Obika, S.; Hari,
Yoshiyuki; Morio, K.-I.; Imanishi, T. Tetrahedron Lett. 2000, 41, 215.
(8) (a) Petersen, M.; Bondensgaard, K.; Wengel, J.; Jacobsen, J. P. J.
Am. Chem. Soc. 2002, 124, 5974. (b) Petersen, M.; Wengel, J. Trends
Biotechnol. 2003, 21, 74.
(9) Koshkin, A. A.; Fensholdt, J.; Pfundheller, H. M.; Lomholt, C. J.
Org. Chem. 2001, 66, 8504.
1666
Org. Lett., Vol. 5, No. 10, 2003

the (S) conformation.14 In contrast, in the spectrum of the
desired oxabicyclo[2.2.1]heptane system 11, H-7 (pseudo-
axial) was found as singlet, and H-1 (pseudoequatorial) and
H-6 (pseudoequatorial) showed two small coupling constants
(J1,6 ) 1.7 Hz, J5,6 ) 2.8 Hz). These observations are
consistent with the assigned (N) conformation of isomer 11.
Also in the NOESY spectra of 11 and 12, NOE was detected
between H-1 and H-6 in 11 but there was no NOE between
H-3 and H-4 in 12. This also supported a (S) conformation
of 12, in which H-3 and H-4 would be pseudoaxial. We also
noted that following the acetylation of 11 and 12, H-7 in 11
and H-4 in 12 were shifted downfield by ∼1 ppm,15 which
further supported our conformational analysis and proton
assignment.
Triacetylation with acetic anhydride and debenzylation
followed by mesylation of the primary alcohol afforded the
triacetyl mesylate 14, which was a desired substrate for
formation of the bicyclo[2.2.1]heptane ring system. The
acetyl group was selected for alcohol protection because of
a low yield in the formation of the cyclic carbonate using
triphosgene (i.e., formation of 10 in Scheme 1) and a side
reaction upon debenzylation with cyclic carbonate. The
debenzylation step was sluggish, and a side reaction of
desilylation was unavoidable with Pd/C; however, Pd black
in formic acid17 at 75 °C provided the optimal yield.
The oxabicyclo[2.2.1]heptane system was constructed by
simultaneous deacetylation and intramolecular O-alkylation
of 14. The ring-closure step resulted in a mixture of two
isomers containing the [3.2.0] oxetane ring system 15 and
oxabicyclo[2.2.1]heptane 16, similar to Scheme 1. The
structural analysis of the mixture of two isomers was based
on NMR chemical shifts and a unique coupling pattern
observed in the spectra of 11 and 12 in Scheme 1. The ratio
obtained was ∼1.76:1 in favor of the desired oxabicyclo-
[2.2.1]heptane system 16 and was dependent on the reaction
temperature. Although a low temperature increased the
fraction of the major isomer 16 over the minor oxetane 15,
the reaction did not reach completion at 0 °C. Therefore,
the reaction was carried out at room temperature to achieve
completion. We hope to optimize this route for exclusive
formation of the major isomer, and we noted a possible lead
for this improvement.7a
To obtain the adenine adduct of the oxabicyclo[2.2.1]-
heptane system, an alternate protection scheme was employed
(Scheme 2). Selective protection16 of the allylic alcohol 9
Scheme 2 a
Benzoylation of a mixture of two isomers, 15 and 16, made
separation possible and gave 17 and 18. The ratio of isolated
isomers indicated in Scheme 2 was consistent with that
1
calculated by H NMR at the stage of a mixture of 15 and
16. The major benzoylated isomer 18 was deprotected by
treatment with TBAF in THF to yield dibenzoyl alcohol 19.18
Inversion of the C-6 stereochemistry was accomplished
successfully in two steps from 19 (oxidation with PDC,
reduction with NaBH4). The resulting inverted alcohol 20
was converted to a triflate 21 by treatment with triflic
anhydride and DMAP. Although mesylate was initially
chosen as the leaving group instead of triflate, the reactivity
of mesylate was insufficient for the substitution reaction. We
surmised that this was mainly because of steric hindrance
to attack by a nucleophile, due to the proximal quaternary
carbon and the unique ring system. This problem was
overcome by the introduction of a more reactive leaving
a Reaction conditions: (a) 1 equiv of TBDPSCl, TEA, cat.
DMAP, CH2Cl2, 0 °C, overnight, 76%. (b) cat. OsO4, NMO,
acetone/water (15/1), 0 °C, overnight, 94%. (c) Ac2O, TEA, cat.
DMAP, CH2Cl2, 5 °C, overnight, 95%. (d) Pd Black, HCO2H,
MeOH, 75 °C, 4 h, 87%. (e) MsCl, TEA, CH2Cl2, 0 °C, 30 min.
(f) K2CO3, MeOH, 0 °C, overnight, and then rt, 1 h, 89% two steps
yield. (g) BzCl, TEA, CH2Cl2, 0 °C, overnight, and then separation
of 17 and 18. (h) TBAF, THF, 10 °C, 2 h, 83%. (i) PDC, CH2Cl2,
reflux, 4 h. (j) NaBH4, EtOH, -20 °C, 1 h and then 0 °C, 1 h,
87% two steps yield. (k) Tf2O, DMAP, CH2Cl2, 0 °C, 5 min, 83%.
(l) adenine, K2CO3, 18-Crown ether-6, DMF, 40-45 °C, 6 h, 76%.
(qm) K2CO3, MeOH, rt, 4 h, 91%.
(14) Obike, S.; Morio, K.-I.; Nanbu, D.; Hari, Y.; Itoh, H.; Imanishi. T.
Tetrahedron 2002, 58, 3039.
(15) By acetylation of 11 and 12, H-7 in 11 was shifted downfield by
0.97 ppm while H-4 in 12 shifted downfield by 0.99 ppm. For comparison,
the NMR spectra were obtained in the same deuterated solvent, CD3OD.
(16) The secondary alcohol was selectively protected over neopentyl
alcohol, which was confirmed by acetylation of the silylated compound.
Chemical shift (CDCl3) of methylene group was changed from 3.32 to 4.17
ppm after acetylation.
(17) Elamine, B.; Anatharomaiah, G. M.; Royer, G. P.; Means, G. E. J.
Org. Chem. 1979, 44, 3442.
(18) The proton NMR spectrum for 19 after desilylation showed all
protons separated and gave a better splitting pattern, which made interpreta-
tion easier than 18. 1H NMR (CDCl3) δ 7.34-8.08 (m, 10H), 5.46 (s, 1H),
4.58 (one of AB quartet, 1H, J ) 11.5 Hz), 4.53 (one of AB quartet, 1H,
J ) 11.8 Hz), 4.26 (s, 1H), 4.09-4.19 (m, 2H), 3.68 (d, 1H, J ) 7.1 Hz),
2.26 (dd, 1H, J ) 7.9, 13.4 Hz), 1.81 (ddd, 1H, J ) 2.8, 3.0, 13.5 Hz).
by treatment with tert-butyldiphenylsilyl chloride and a
catalytic amount of 4-(dimethylamino)pyridine and sequential
dihydroxylation with OsO4 at 0 °C gave triol 13 as an oil.
Org. Lett., Vol. 5, No. 10, 2003
1667

group, triflate, which would be a better leaving group and
might add SN1 character to the substitution reaction. Sub-
stitution of triflate 21 by adenine proceeded smoothly at 40
°C to give dibenzoyl cLNA 22. The benzoyl groups were
then removed with potassium carbonate to give the target
molecule, a novel locked carbocyclic nucleoside 23 (as
precursor to cLNAs). In the spectrum of 20, 21, and the
dibenzoyl cLNA monomer 22, one instance of characteristic
W-type long-range coupling for each compound was de-
tected. In addition to a unique long-range coupling pattern,
two coupling constants of J1,6 and J1,7 were zero in the
spectrum of dibenzoyl cLNA monomer 22. As mentioned
before, it is thought that the relationship between H-1 and
H-6 is equatorial-equatorial and that between H-1 and H-7
is equatorial-axial, which indicated an (N)-like conformation
of the oxabicyclo[2.2.1] ring system.
because the carbocyclic ribose equivalent can be coupled
with the base at a late stage in the synthetic pathway. This
efficiency, derived from the ability to prepare a large quantity
of the triflate substrate as a common intermediate, is an
important advantage in synthetic nucleoside chemistry.
Development of a more efficient synthetic route is under way
in our laboratory.
In conclusion, we have introduced a novel carbocyclic
nucleoside ring system, which promises to have broad
biological interest due to its defined conformational proper-
ties and would complement studies of LNAs.
Acknowledgment. We thank Dr. Victor E. Marquez
(NCI) and Dr. Thomas F. Spande (NIDDK) for helpful
discussions.
Generally, the formation of carbocyclic nucleosides re-
quires more synthetic steps than does the preparation of the
corresponding natural nucleosides. The synthetic route we
describe here would actually decrease the number of
synthetic steps when applied to the synthesis of many
nucleoside and nucleotide derivatives. This is primarily
Supporting Information Available: Experimental pro-
cedures and spectroscopic data for compounds 13-23. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL034326Z
1668
Org. Lett., Vol. 5, No. 10, 2003