Conformationally constrained 2′-N,4′-C-ethylene-bridged thymidine (Aza-ENA-T): Synthesis, structure, physical, and biochemical studies of Aza-ENA-T-modified oligonucleotides

Article abstract of DOI:10.1021/ja0634977

The 2′-deoxy-2′-N,4′-C-ethylene-bridged thymidine (aza-ENA-T) has been synthesized using a key cyclization step involving 2′-ara-trifluoromethylsufonyl-4′-cyanomethylene 11 to give a pair of 3′,5′-bis-OBn-protected diastereomerically pure aza-ENA-Ts (12a

Full text of DOI:10.1021/ja0634977

Published on Web 11/08/2006
Conformationally Constrained 2′-N,4′-C-Ethylene-Bridged
Thymidine (Aza-ENA-T): Synthesis, Structure, Physical, and
Biochemical Studies of Aza-ENA-T-Modified Oligonucleotides
Oommen P. Varghese, Jharna Barman, Wimal Pathmasiri, Oleksandr Plashkevych,
Dmytro Honcharenko, and Jyoti Chattopadhyaya*
Contribution from the Department of Bioorganic Chemistry, Box 581, Biomedical Center,
Uppsala UniVersity, SE-75123 Uppsala, Sweden
Received May 25, 2006; E-mail: jyoti@boc.uu.se
Abstract: The 2′-deoxy-2′-N,4′-C-ethylene-bridged thymidine (aza-ENA-T) has been synthesized using a
key cyclization step involving 2′-ara-trifluoromethylsufonyl-4′-cyanomethylene 11 to give a pair of 3′,5′-bis-
OBn-protected diastereomerically pure aza-ENA-Ts (12a and 12b) with the fused piperidino skeleton in
the chair conformation, whereas the pentofuranosyl moiety is locked in the North-type conformation (7° <
P < 27°, 44° < φm < 52°). The origin of the chirality of two diastereomerically pure aza-ENA-Ts was found
to be due to the endocyclic chiral 2′-nitrogen, which has axial N-H in 12b and equatorial N-H in 12a. The
latter is thermodynamically preferred, while the former is kinetically preferred with Ea ) 25.4 kcal mol-1
,
which is thus far the highest observed inversion barrier at pyramidal N-H in the bicyclic amines. The
5′-O-DMTr-aza-ENA-T-3′-phosphoramidite was employed for solid-phase synthesis to give four different
singly modified 15-mer antisense oligonucleotides (AONs). Their AON/RNA duplexes showed a Tm increase
of 2.5-4 °C per modification, depending upon the modification site in the AON. The relative rates of the
RNase H1 cleavage of the aza-ENA-T-modified AON/RNA heteroduplexes were very comparable to that
of the native counterpart, but the RNA cleavage sites of the modified AON/RNA were found to be very
different. The aza-ENA-T modifications also made the AONs very resistant to 3′ degradation (stable over
48 h) in the blood serum compared to the unmodified AON (fully degraded in 4 h). Thus, the aza-ENA-T
modification in the AON fulfilled three important antisense criteria, compared to the native: (i) improved
RNA target affinity, (ii) comparable RNase H cleavage rate, and (iii) higher blood serum stability.
Introduction
expression utilizing a naturally occurring mechanism called
RNA interference (RNAi).
Modified oligonucleotides have successfully been used as
valuable tools to inhibit gene expression by utilizing various
mechanisms of action.1-7 The most matured method is the
antisense technology,6,8 which exploits the ability of a single-
stranded DNA oligonucleotide to bind to the target messenger
RNA (mRNA) via Watson-Crick base pairing in a sequence-
specific manner. Once bound to the target RNA, the antisense
agent either sterically blocks the synthesis of ribosomal proteins
or induces RNase-H-mediated degradation of the target mRNA.
The modified oligonucleotides have also been found useful as
short double-stranded RNA (siRNA),7,9-11 which silence gene
For in ViVo applications12,13 of the oligonucleotide-based
approaches, appropriate chemical modifications14,15 are war-
ranted to enhance target affinity, specificity, and stability toward
the endo- and exonucleases, as well as tissue-specific delivery
to improve the overall pharmacokinetic properties. It is a
challenge to come up with optimally modified monomer blocks
with a natural phosphate backbone that can successfully address
all of the above issues related to the improvement of the
pharmacokinetic properties.14,16
Among various sugar, phosphate, and nucleobase modifica-
tions reported,14,15 synthetic oligonucleotides having conforma-
tionally constrained furanose-fused bi- and tricyclic carbohydrate
moieties17-31 (Figure S1 and Discussion S1 in the Supporting
Information) or a modified pyranose derivative16,32,33 in the
(1) Helene, C. Eur. J. Cancer 1991, 27, 1466.
(2) Vasquez, K. M.; Wilson, J. Trends Biochem. Sci. 1998, 23, 4.
(3) Chech, T. R.; Zaug, A. J.; Grabowski, P. J. Cell 1981, 27, 487.
(4) Kruger, K.; Grabowski, P. J.; Zaug, A. J.; Sands, J.; Gottschling, D. E.;
Chech, T. R. Cell 1982, 31, 147.
(5) Schubert, S.; Gu¨l, D. C.; Grunert, H. P.; Grunert, H.; Erdmann, V. A.;
Kurreck, J. Nucl. Acids Res. 2003, 31, 5982.
(12) Stephenson, M. L.; Zamecnik, P. C. Proc. Natl. Acad. Sci. U.S.A. 1978,
75, 285.
(13) Zamecnik, P. C.; Stephenson, M. L. Proc. Natl. Acad. Sci. U.S.A. 1978,
75, 280.
(14) Kurreck, J. Eur. J. Biochem. 2003, 270, 1628.
(15) Freier, S. M.; Altmann, K. H. Nucleic Acids Res. 1997, 25, 4429.
(16) Herdewijn, P. Biochim. Biophys. Acta 1999, 1489, 167.
(17) Wengel, J. Acc. Chem. Res 1999, 32, 301.
(18) Steffens, R.; Leumann, C. J. J. Am. Chem. Soc. 1997, 119, 11548.
(19) Meldgaard, M.; Wengel, J. J. Chem. Soc., Perkin Trans. 1 2000, 3539.
(20) Leumann, C. J. Bioorg. Med. Chem. 2002, 10, 841.
(6) Crooke, S. T. Annu. ReV. Med. 2004, 55, 61.
(7) Fire, A.; Xu, S.; Motegomery, M. K.; Kostas, S. A.; Driver, S. E.; Mello,
C. C. Nature 1998, 391, 806.
(8) Brenda, F. B.; Brett, P. M. Biochem. Biophys. Acta 1999, 1489, 3.
(9) Hutvagner, G.; Zamore, P. D. Science 2002, 297, 2056.
(10) Novina, C. D.; Sharp, P. A. Nature 2004, 430, 161.
(11) Xia, J.; Noronha, A.; Toudjarska, I.; Li, F.; Akinc, A.; Braich, R.; Frank-
Kamenetsky, M.; Rajeev, G. K.; Egli, M.; Manoharan, M. Chem. Biol.
2006, 1, 176.
9
10.1021/ja0634977 CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 15173-15187
15173

A R T I C L E S
Varghese et al.
monomer nucleotide units have been found to be promising in
terms of target RNA binding, accessibility, and nuclease
resistance. The enhanced target-binding ability of oligonucleo-
tides modified with North-conformationally constrained sugar
units23,34-40 can be attributed to the conformational pre-
organization by improved stacking between the nearest neigh-
bors,41 thereby minimizing the entropic energy penalty in the
free energy of stabilization for the duplex formation with RNA.
Koizumi et al.42 have shown that the 2′-O,4′-C-ethylene-
bridged nucleoside22,42 (ENA) (structure I in Figure S1 in the
Supporting Information) modified antisense oligonucleotides
(AONs) have approximately 55 times higher stability toward
3′-exonuclease compared to the locked nucleic acid (LNA)
analogue.37 The ability of ENA for efficient target binding and
high nuclease resistance has been well-exploited to evaluate its
antisense,37,43-46 antigene,47 and RNAi37,45-48 properties.
conformationally constrained 2′-N,4′-C-ethylene-bridged ENA-
thymidine (aza-ENA-T, structure F in Scheme 1) and incorpo-
rate it into the AONs to explore their potential for effective
gene-directed therapeutics and diagnostics. The aza-ENA-based
AONs may have three clear advantages over the corresponding
ENA-containing counterparts in a similar manner as the
2′-amino-LNA-modified AONs have over the LNA-modified
counterparts.25,49 First, the endocyclic amino functionality of
the aza-ENA analogue could be utilized as a well-defined
conjugation site,50 and thereby, we can control the hydrophilic,
hydrophobic, and steric requirements of a minor groove of the
duplex. Second, the amine-derivatized AONs have displayed
increased thermal affinities51,52 toward complementary RNA,
possibly because of the presence of positively charged moieties
at physiological pH, and thus could influence partial neutraliza-
tion of the negatively charged phosphates in the duplexes.25,31,53
Third, introduction of a fluorescence probe connected to this
nitrogen moiety will enable us for real-time in ViVo imaging of
RNA and can therefore be used for specific detection of nucleic
acids while maintaining their hybridization properties.49,54
The ENA-modified AONs have been shown to have unique
properties, such as high target RNA affinity (+3.5 to +5.2 °C
per modification),42 sequence selectivity toward ssRNA/dsDNA
targets,47 and high nuclease resistance (in ViVo and in
Vitro),22,37,46 in ViVo RNase recruitment,37 and triplex forming47
properties. This prompted us to design and synthesize the
We report here synthesis of aza-ENA-T, physicochemical,
nuclear magnetic resonance (NMR), and computational struc-
tural studies of aza-ENA-T and biochemical studies of aza-
ENA-T containing oligonucleotides. The NMR and computa-
tional structural studies of aza-ENA-T showed that the piperidino
moiety of the aza-ENA-T is indeed locked in the chair
conformation (with the nitrogen lone pair in the axial position
and the N-H proton in the equatorial position), whereas the
fused sugar is constrained to a North-type conformation similar
to that of the 2′-O,4′-C-ethylene-bridged ENA analogue.22,42
Finally, the aza-ENA-T nucleotides have been incorporated in
15-mer AONs as a single modification at four different sites to
give four mono-aza-ENA-T-substituted AONs 2-5 (sequences
shown in Table 1). These AON/RNA duplexes have shown an
increase in the thermal stability by +2.5 to +4 °C per
modification toward complementary RNA, depending upon the
substitution site. We also show that the relative rates of the
RNase-H1-promoted cleavage of the aza-ENA-T-modified
AON/RNA heteroduplexes are comparable to that of the native
counterpart, and quite interestingly, the aza-ENA-T modifica-
tions also result in a significant increase of the AON resistance
to 3′-exonuclease degradation in the blood serum compared to
the native counterpart. No blood serum stability or the RNase
H recruitment capability study has thus far been reported for
the 2′-amino-LNA-incorporated AONs, and hence, no direct
comparison is yet possible with the aza-ENA-T-substituted
AONs.
(21) Koshkin, A. A.; Singh, S. K.; Nielson, P.; Rajwanshi, V. K.; Kumar, R.;
Meldgaard, M.; Wengel, J. Tetrahedron 1998, 54, 3607.
(22) Morita, K.; Takagi, M.; Hasegawa, C.; Kaneko, M.; Tsutsumi, S.; Sone,
J.; Ishikawa, T.; Imanishi, T.; Koizumi, M. Bioorg. Med. Chem. 2003, 11,
2211.
(23) Obika, S.; Nanbu, D.; Hari, Y.; Morio, K.-I.; In, Y.; Ishida, T.; Imanishi,
T. Tetrahedron Lett. 1997, 38, 8735.
(24) Kumar, R.; Singh, S. K.; Koshkin, A. A.; Rajwanshi, V. K.; Meldgaard,
M.; Wengel, J. Bioorg. Med. Chem. Lett. 1998, 8, 2219.
(25) Singh, S. K.; Kumar, R.; Wengel, J. J. Org. Chem. 1998, 63, 10035.
(26) Rajwanshi, V. K.; Hakansson, A. E.; Sorensen, M. D.; Pitsch, S.; Singh,
S. K.; Kumar, R.; Nielsen, P.; Wengel, J. Angew. Chem., Int. Ed. 2000,
39, 1656.
(27) Sorensen, M. D.; Kvaerno, L.; Bryld, T.; Hakansson, A. E.; Verbeure, B.;
Gaubert, G.; Herdewijn, P.; Wengel, J. J. Am. Chem. Soc. 2002, 124, 2164.
(28) Wang, G.; Gunic, E.; Girardet, J. L.; Stoisavljevic, V. Bioorg. Med. Chem.
Lett. 1999, 9, 1147.
(29) Pradeepkumar, P. I.; Cheruku, P.; Plashkevych, O.; Acharya, P.; Gohil,
S.; Chattopadhyaya, J. J. Am. Chem. Soc. 2004, 126, 11484.
(30) Pradeepkumar, P. I.; Chattopadhyaya, J. J. Chem. Soc., Perkin Trans. 2
2001, 2074.
(31) Honcharenko, D.; Varghese, O. P.; Plashkevych, O.; Barman, J.; Chatto-
padhyaya, J. J. Org. Chem. 2006, 71, 299.
(32) Kozlov, I. A.; de Bouvere, B.; van Aerschot, A.; Herdewijn, P.; Orgel, L.
E. J. Am. Chem. Soc. 1999, 121, 5856.
(33) Wang, J.; Verbeure, B.; Luyten, I.; Lescrinier, E.; Froeyen, M.; Hendrix,
C.; Rosemeyer, H.; Seela, F.; van Aerschot, A.; Herdewijn, P. J. Am. Chem.
Soc. 2000, 122, 8595.
(34) Tarkoy, M.; Bolli, M.; Schweizer, B.; Leumann, C. HelV. Chim. Acta 1993,
76, 481.
(35) Siddiqui, M. A.; Ford, H.; George, C.; Marquez, V. E. Nucleosides
Nucleotides 1996, 15, 235.
(36) Wang, G.; Giradet, J. L.; Gunic, E. Tetrahedron 1999, 55, 7707.
(37) Morita, K.; Yamate, K.; Kurakata, S.; Abe, K.; Imanishi, T.; Koizumi, M.
Nucleic Acids Res. Sup. 2002, 2, 99.
(38) Kvaerno, L.; Wengel, J. J. Org. Chem. 2001, 66, 5498.
(39) Babu, B. R.; Raunak; Poopeiko, N. E.; Juhl, M.; Bond, A. D.; Parmar, V.
S.; Wengel, J. Eur. J. Org. Chem. 2005, 11, 2297.
(40) Kumar, T. S.; Madsen, A. S.; Wengel, J.; Hrdlicka, P. J. J. Org. Chem.
2006, 71, 4188-4201.
(41) Isaksson, J.; Acharya, S.; Barman, J.; Cheruku, P.; Chattopadhyaya, J.
Biochemistry 2004, 43, 15996.
(42) Morita, K.; Hasegawa, C.; Kaneko, M.; Tsutsumi, S.; Sone, J.; Ishikawa,
T.; Imanishi, T.; Koizumi, M. Bioorg. Med. Chem. Lett. 2002, 12, 73.
(43) Yagi, M.; Takeshima, Y.; Surono, A.; Takagi, M.; Koizumi, M.; Matsuo,
M. Oligonucleotides 2004, 14, 33.
(44) Surono, A.; van Khanh, T.; Takeshima, Y.; Wada, H.; Yagi, M.; Takagi,
M.; Koizumi, M.; Matsuo, M. Hum. Gene Ther. 2004, 15, 749.
(45) Morita, K.; Yamate, K.; Kurakata, S.-I.; Watanabe, K.; Imanishi, T.;
Koizumi, M. Nucleosides, Nucleotides Nucleic Acids 2003, 22, 1619.
(46) Takagi, M.; Morita, K.; Nakai, D.; Nakagomi, R.; Tokui, T.; Koizumi, M.
Biochemistry 2004, 43, 4501.
Results and Discussion
Synthesis of all 2′,4′-bridged nucleosides reported21,22 thus
far involve a nucleophilic displacement reaction with the
nucleophile positioned at C2′ and a leaving group at the
extended arm of C4′ (general structure “A1” in Scheme 1). We
(49) Hrdlicka, P. J.; Babu, B. R.; Sorensen, M. D.; Harrit, N.; Wengel, J. J.
Am. Chem. Soc. 2005, 127, 13293.
(50) Maag, H.; Schmidt, B.; Rose, S. J. Tetrahedron Lett. 1994, 35, 6449.
(51) Cuenoud, B.; Casset, F.; Hu¨sken, D.; Natt, F.; Wolf, R. M.; Altmann, K.
H.; Martin, P.; Moser, H. E. Angew. Chem., Int. Ed. 1998, 37, 1288.
(52) Prakash, T. P.; Ask, P.; Lesnik, E.; Mohan, V.; Tereshko, V.; Egli, M.;
Manoharan, M. Org. Lett. 2004, 6, 1971.
(47) Koizumi, M.; Morita, K.; Daigo, M.; Tsutsumi, S.; Abe, K.; Obika, S.;
Imanishi, T. Nucleic Acids Res. 2003, 31, 3267.
(48) Hamada, M.; Ohtsuka, T.; Kawaida, R.; Koizumi, M.; Morita, K.; Furukawa,
H.; Imanishi, T.; Miyagishi, M.; Taira, K. Antisense Nucleic Acid Drug
DeV. 2002, 12, 301.
(53) Singh, S. K.; Kumar, R.; Wengel, J. J. Org. Chem. 1998, 63, 6078.
(54) Hrdlicka, P. J.; Kumar, T. S.; Wengel, J. Chem. Comm. 2005, 34, 4279.
9
15174 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Conformationally Constrained Aza-ENA-T
A R T I C L E S
Scheme 1. Structure (A1)a
a General strategy for 2′,4′-cyclization thus far reported in the literature involves the engineering of a nucleophile (Nu) at C2′ and a leaving group (LG)
at the extended arm of C4′. Structure (A2), on the other hand, shows a reverse strategy in which the leaving group LG is placed in the sugar moiety, whereas
the nucleophile Nu is engineered at the sidearm. Structure (B): Our strategy involving intermediate A2 involves a ring-closure reaction (path A) to give
cyanomethylene-bridged carbocyclic nucleoside (C), through a R-carbanion, or alternatively involving path B, where CN is first reduced to putative (D)
followed by instantaneous cyclization to give 2′,4′-conformationally constrained 2′-N,4′-C-ethylene-bridged aza-ENA-T (structure E). Both path A and B
however involve a nucleophile at C4′ (Nu) and a departing group at C2′ (LG).
Table 1. Aza-ENA-T-Modified AONs and the Thermal Denaturation Studies of Their Duplexes with Complementary RNA or DNA Targetsa
Tm
Tm
MALDI
−
MS of AON 1
−
H]+
5:
AON
AON sequences synthesized
(°
C with RNA)
∆Tm
(°
C with DNA)
∆Tm
*
found/calcd [M
+
1
2
3
4
5
3′-d(CTTCTTTTTTACTTC)-5′
3′-d(CTTCTTTTTTACTTC)
3′-d(CTTCTTTTTTACTTC)
3′-d(CTTCTTTTTTACTTC)
3′-d(CTTCTTTTTTACTTC)
44
48
46.5
47.5
48
45
4448.6/4448.7
4489.7/4491.1b
4489.7/4490.7
4489.7/4490.7
4489.7/4490.8
+4
44.5
42.5
42
-0.5
-2.5
-3
+2.5
+3.5
+4
42
-3
a Tm values measured as the maximum of the first derivative of the melting curve (A260 versus temperature) recorded in medium salt buffer (60 mM
Tris-HCl at pH 7.5, 60 mM KCl, 0.8 mM MgCl2, and 2 mM DTT) with a temperature range of 20-70 °C using 1 µM concentrations of the two complementary
strands. ∆Tm ) Tm relative to RNA compliment. ∆Tm* ) Tm relative to DNA compliment. T ) aza-ENA-T monomer. b [M + 2H]+.
considered an alternative strategy (general structure “A2” in
Scheme 1) for nucleophilic ring closure to give desired
constrained nucleosides. An intermediate such as B in Scheme
1 can potentially lead us to two types of products, depending
upon whether we generate a carbanion at the R carbon to the
cyano group (path A in Scheme 1) or successfully reduce the
cyano to the primary amino group without generating a R
carbanion next to the -CN (path B in Scheme 1). We argued
that, in the former case, we might be able to engineer the
construction of a cyanomethylene-bridged carbocyclic [2.1.1]
system (C) and that, in the latter case, it could lead us to the
2′-deoxy-2′-N,4′-C-ethylene-bridged nucleoside (aza-ENA-T)
having a 2-aza-6-oxabicyclo[3.2.1]octane skeleton.
Thus, upon treatment of the intermediate (B) (Scheme 1) with
NaHMDS in anhydrous tetrahydrofuran (THF), we found the
formation of a stable fused product (C) (Scheme 1) with a 35%
yield as a mixture of two diastereomers (R/S, 8:2) because of
chiral C6′ (Scheme 1). The stereochemistry at C6′ in compound
C has been proven by 2D correlation spectroscopy (COSY) and
1H-homodecoupling experiments (parts a-c of Figure S13 and
parts a and b of S14 in the Supporting Information). In the major
diastereomer, the H3′ appears as a double doublet (dd, J ) 7.7
and 1.5 Hz) because of the vicinal3JH3′,H2′ coupling (7.7 Hz)
with H2′ and, most importantly, a4JHH W coupling with H6′
(1.5 Hz), which is only possible when C6′ is in the R
configuration (Figure S49 in the Supporting Information). As a
result of the R configuration at C6′, the dihedral angle between
H6′ and H2′, φ(H6′-C6′-C2′-H2′), becomes very close to 90°
(see the energy-minimized models in Figure S49 in the
Supporting Information), which is why no vicinal three-bond
coupling between H6′ and H2′ (3JH2′,H6′) has been observed for
the major diastereomer. This means that the C6′ in the minor
diastereomer is in the S stereochemistry (for a model, see Figure
S49 in the Supporting Information). The observed significant
upfield 13C chemical shift at the C1′ (63.6 ppm) in the fused
product (C) is probably due to a change from the more
electronegative 2′-OTf to the 2′-deoxy-2′-C (carbocyclic) con-
jugate, resulting in an increase of the anomeric effect in the
highly constrained North-fused nucleoside as well as a +I effect
owing to the carbon substituent at C2′. Finally, the 1D difference
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15175

A R T I C L E S
Varghese et al.
Scheme 2. Reagents and Conditionsa
a (i) NaH, BnBr, CH3CN, -5 °C to room temperature, overnight; (ii) Tf2O, pyridine, CH2Cl2, 0 °C, 3 h; (iii) LiCN, DMF, room temperature, 3 days; (iv)
acetic acid, Ac2O, triflic acid, room temperature, 3 h; (v) persilylated thymine, TMSOTf, CH3CN, 80 °C, overnight; (vi) NaOMe, methanol, 3 h; (vii) MsCl,
pyridine, 0 °C, 6 h; (viii) DBU, CH3CN, room temperature, 1 h; (ix) 0.1 M H2SO4, acetone, reflux, overnight; (x) Tf2O, pyridine, CH2Cl2, DMAP, 0 °C, 2.5
h; (xi) NaBH4, trifluoroacetic acid, THF, room temperature, overnight; (xii) Pd(OH)2, ammonium formate, methanol, reflux, overnight, followed by 1 M
BCl3 in CH2Cl2, -78 °C, 3 h; (xii) phenoxyacetyl chloride, pyridine, room temperature, 3 h; (xiv) DMTr-Cl, pyridine, room temperature, 7 h (overnight for
19); (xv) NC(CH2)2OP(Cl)N(iPr)2, DIPEA, THF, room temperature, 3 h (overnight for 20). Abbreviations: Bn, benzyl; Ac, acetyl; Tf, trifluoromethylsulfonyl;
PAC, phenoxyacetyl; THF, tetrahydrofuran; DBU, 1,8-diazabicyclo[5.4.0]undec-7-en; DMTr, 4,4′-dimethoxytrityl; DIPEA, diisopropylethylamine; TFA,
trifluoroacetyl.
nuclear Overhauser effect (NOE) experiment is consistent with
the North-type sugar conformation (see ii in Figure S2 and
Figure S10 in the Supporting Information).
These initial unsuccessful efforts to proceed through path A
in Scheme 1 have made us explore the feasibility of structure
B (path B in Scheme 1), which has a masked amino function
in the form of a -CN group in the side arm as well as a -CH2-
for one carbon homologation (D in Scheme 1), both of which
are necessary for intramolecular cyclization to yield a novel
aza-ENA-T nucleoside (compound E, Scheme 1). The complete
synthetic strategy, which leads us to the successful synthesis
of aza-ENA-T, is shown in Scheme 2. It is also noteworthy
that the aliphatic nitrile (pKa ∼ 29-31)55 derivatives (4-11,
Scheme 2) were fully compatible with the synthetic strategy
containing strongly acidic and basic conditions as described in
Scheme 2.
The structural integrity of this conformationally constrained
product was also proven by a 1H-13C NMR correlation (HMBC)
experiment (see i in Figure S2 in the Supporting Information),
which showed C2′/H6′ (2JCH), through bond correlation. The
mass-spectral data by matrix-assisted laser desorption ioniza-
tion-time of flight (MALDI-TOF) have also provided evidence
for the structural integrity of the highly strained carbocyclic
nucleoside (see the Experimental Section for full characteriza-
tion). However, our efforts to remove the benzyl group from
compound C using Pd(OH)2/HCOONH4 and BCl3 (Scheme 1)
has not met with success thus far. The other problem that
remains is to find a mild reducing agent, which can convert the
-CN group in compound C (Scheme 1) to a relatively poorer
electron-withdrawing group (for example, to an amino function),
to assess its potential in the strategies concerning gene-
directed drug development [reduction by (CF3CO2)3BH was
unsuccessful].
1. Synthesis of Aza-ENA-T. The synthesis of the aza-ENA-T
derivative was started with a known sugar precursor21 1, which
was converted to 3,5-di-O-benzyl-4-C-hydroxymethyl-1,2-O-
isopropylidene-R-D-ribofuranose 2.21 Compound 2 was then
converted to the 4-triflyloxymethylene derivative 3 using triflic
(55) Fleming, F. F.; Shook, B. C. Tetrahedron 2002, 58, 1.
9
15176 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Conformationally Constrained Aza-ENA-T
A R T I C L E S
Figure 1. (A and B) 1H-13C HMBC spectra, showing the long-range through-bond connectivities between C7′/H2′ and C2′/H7e′ for the two diastereomers
12a and 12b. (C and D) Energy-minimized stereochemical representations, showing the 1H-13C HMBC connectivities of the two diastereomers 12a and 12b;
R ) OCH2Ph (not shown).
anhydride in a dichloromethane/pyridine mixture (3:1, v/v) at
0 °C. The crude product obtained after aqueous workup was
subsequently treated with 3 equiv of LiCN in N,N-dimethyl-
formamide (DMF) and stirred at room temperature for 3 days,
which afforded the cyano-sugar 4 in an overall yield of 56%
from 2, along with some unidentified minor compounds.
Compound 4 was converted to diacetate 5 (1,2-di-O-acetyl-3,5-
di-O-benzyl-4-C-cyanomethyl-D-ribofuranose) using a mixture
of acetic acid, acetic anhydride, and triflic acid by stirring for
3 h at room temperature. The crude product 5 (4 f 5 was almost
quantitative) was subjected to the modified Vorbruggen reac-
tion31,56 using in situ silylation of thymine and subsequent
trimethylsilyl triflate mediated coupling to give the â-configured
thymine nucleoside 6 in 80% yield. The â configuration of 6
was confirmed by a 1D differential NOE experiment, which
showed 8% NOE enhancement of H6 upon irradiation of H2′
(dH6-H2′ ≈ 2.7 Å for â anomer, and dH6-H2′ ≈ 4 Å for R
anomer). Deacetylation of 6 at C2′ was carried out using sodium
methoxide in methanol at room temperature, and the product
(7) was isolated as a crude material [single spot on thin-layer
chromatography (TLC); for 13C NMR, see Figure S17 in the
Supporting Information], which was directly mesylated using
mesyl chloride in pyridine at room temperature to afford 8 with
an overall yield of 95% in two steps from 6. The compound 8
was converted to the 2,2′-anhydro product 9 using 1.05 equiv
of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in acetonitrile in
91% yield. It should be noted that the use of excess base or
stronger base such as NaHMDS resulted in instantaneous
depyrimidation. Opening of the 2,2′-anhydro ring in 9 went
smoothly by refluxing with a mixture of 0.1 M aqueous sulfuric
acid/acetone (1:1, v/v) to give the arabino product (10)
quantitatively.
Treatment of 10 with triflic anhydride, pyridine, 4-dimethyl-
aminopyridine (DMAP), and anhydrous CH2Cl2 at 0 °C gave
the desired triflate nucleoside 11 in 84% yield. Reduction of
(56) Vorbru¨ggen, H.; Ho¨fle, G. Chem. Ber. 1981, 114, 1256.
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15177

A R T I C L E S
Varghese et al.
Figure 2. One-dimensional selective NOESY spectra of 12a and 12b. (A) Irradiation at H1′ in 12a shows enhancements at H7a′ (3.9%), H2′ (1.5%), and
H6 (1.8%). (B) Irradiation at H1′ in 12b shows enhancements at H7a′ (8.0%), H2′ (5.5%), and H6 (5.9%). (C and D) NOE contacts in the two diastereomers
12a and 12b, respectively (R ) OCH2Ph and are not shown for clarity of the picture).
the cyano group using trifluoroacetoxy borohydride,57 prepared
in situ from NaBH4 and trifluoroacetic acid, gave the primary
amine, which spontaneously resulted in intramolecular cycliza-
tion to give a mixture of two diastereomeric aza-ENA-T 12a
and 12b isomers, isolated in 40 and 5% yield, respectively.
These diastereomers showed identical masses by MALDI-TOF
mass spectroscopy (see the Experimental Section). The fact that
the intramolecular ring-closure reaction has indeed taken place
to give the 2′-deoxy-2′-N,4′-C-ethylene-bridged nucleoside (aza-
ENA-T) having a 2-aza-6-oxabicyclo[3.2.1]octane skeleton
fused with a North-conformationally constrained pentofuranosyl
moiety in 12a and 12b was unequivocally proven by long-range
1H-13C NMR correlation (HMBC, Figure 1) and NOE experi-
ments (Figure 2). The benzyl groups of the diasteromeric 12a/
12b were deprotected for characterization using Pd(OH)2/
ammonium formate in methanol and subsequently BCl3 in
dichloromethane at -78 °C to aza-ENA-T (13) in 60% yield
(Scheme 2).
(TOCSY),59 and 13C NMR experiments, including distortionless
enhancement by polarization transfer (DEPT),60 as well as long-
1
61
range H-13C HMBC correlation (2JH,C and3JH,C
)
and a one-
bond heteronuclear multiple-quantum coherence (HMQC) ex-
periment.62
2.1. Assignment of 1H and 13C Chemical Shifts and
Evidence for Ring Closure in 12a, 12b, and 13. The COSY
1
and HMQC experiments allowed us to assign the H and 13C
chemical shifts for 12a, 12b, and 13 (see Figures S23, S25,
S28, and S30 in the Supporting Information). The HMBC
spectrum for 12a, 12b, and 13 showed that only H7e′ (equatorial
H7′) has a correlation with C2′ but none between H7a′ (axial
H7′) and C2′. This suggested that the dihedral angle between
H7e′ and C2′, φ[H7e′-C7′-N-C2′], is close to 180°, whereas
the dihedral angle between H7a′ and C2′, φ[H7a′-C7′-N-C2′],
is about 90°. The presence of a long-range HMBC correlation
of H7e′ with C2′ also unequivocally showed that the six-
membered piperidino[3.2.1] ring fused with the pentofuranose
ring has indeed been formed in the ring-closure reaction (11 f
2. NMR Characterization of 12a, 12b, and 13. The
characterization and conformational analysis of 12a, 12b, and
13 have been performed using NMR data (at 500 and 600 MHz
in CDCl3/DMSO-d6) obtained by double homodecoupling
experiments, 1D nuclear Overhauser effect spectrometry
(NOESY),58,59 1D selective total correlation spectroscopy
(58) Kessler, H.; Oschkinat, H.; Griesinger, C.; Bermel, W. J. Magn. Reson.
1986, 70, 106.
(59) Stonehouse, J.; Adell, P.; Keeler, J.; Shaka, A. J. J. Am. Chem. Soc. 1994,
116, 6037.
(60) Doddrell, D. M.; Pegg, D. T.; Bendall, M. R. J. Magn. Reson. 1982, 48,
323.
(61) Bax, A.; Summers, M. J. Am. Chem. Soc. 1986, 108, 2093.
(62) Bax, A.; Geriffey, R.; Hawkins, B. J. Magn. Reson. 1983, 55, 301.
(57) Gruenefeld, P.; Richert, C., J. Org. Chem. 2004, 69, 7543.
9
15178 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Conformationally Constrained Aza-ENA-T
A R T I C L E S
Figure 4. Determination of activation energy (Ea) for the nitrogen inversion
of 12b to 12a using an Arrhenius plot of ln k versus 1/T.
using the equation of the unimolecular first-order rate kinetics.63
The populations of 12a and 12b at different time intervals were
obtained from the peak integrals of H6 for the two isomers.
The Arrhenius plot of ln k versus 1/T shows a linear
Figure 3. Nonreversible conversion of 12b (with NHa) to 12a (with NHe)
in pyridine-d5 at 298 K.
correlation with R ) 0.98. The slope gave Ea ) 25.4 kcal mol-1
,
12a/12b, Scheme 2). The deprotected 13 showed very similar
NMR spectra as that of the major isomer 12a (see Figure S37
in the Supporting Information). The 1D selective TOCSY
experiment confirmed the spin system for H7a′, H7e′, and H6a′,
for 12a and 13, whereas an additional correlation for H7a′, H7e′,
H6a′, and H6e′, with the NH proton at δ 4.55 ppm was also
found for 12b. The presence of the NH proton was also
confirmed by D2O exchange and NMR simulation experiments.
The H7a′ has three vicinal couplings (3JH7a′,H6a′ ) 11.9 Hz,
3JH7a′,H6e′ ) 5.2 Hz, and 3JH7a′,NH ) 11.9 Hz) besides the geminal
coupling (2JH7a′,H7e′ ) 14.2 Hz). This suggests that the NH proton
in 12b is in an axial orientation (NHa) and that the NH proton
in 12a is in the equatorial configuration. The vicinal coupling
of the NH proton in 12a with neither H7a′ nor H7e′ has however
been observable by NMR, probably because of the fast exchange
with the bulk solvent. Details of the assignments of proton
resonances and stereochemistry for 12a, 12b, and 13 can be
found in Discussion S3 in the Supporting Information.
3. Conversion of Minor 2′-NHa-Diastereomer (12b) to the
Major 2′-NHe-Diastereomer (12a): Kinetics and Confor-
mational Studies of Nitrogen Inversion. Diastereomers 12a
and 12b have been isolated in pure form and fully characterized
by NMR and mass spectroscopy, suggesting that these isomers
were fairly stable in CH2Cl2 containing ca. 5-10% methanol
or in CHCl3 solution. The piperidino moiety in the major isomer
(12a) takes up the chair conformation with N-H equatorial
(NHe) (see Discussion S3 in the Supporting Information) to
reduce the 1,3-diaxial interaction, whereas the minor isomer
(12b), with the piperidino ring also in the chair conformation
but with axial N-H (NHa) (see Discussion S3 in the Supporting
Information), is relatively unstable because of the unfavorable
1,3-diaxial interaction (see below).
whereas the intercept showed the frequency of the collision
factor A ) 1.190 × 1014 s-1 (Figure 4). The free energy of
activation was also calculated,64 which was found to be ∆Gq )
23.4 kcal mol-1 at 298 K in pyridine-d5. In CDCl3, the two
diastereomers 12a and 12b have very slowly (in 30 days)
reached an equilibrium (40:60 of 12a/12b) with the equilibrium
constant63 Kc ) 0.67. The ∆Gq was found to be 25.4 kcal mol-1
at 298 K (Figure S43 in the Supporting Information). The
complete conversion in pyridine-d5 with a 2 kcal mol-1 lower
∆Gq suggests that the conversion of 12b to 12a is base-
catalyzed.
The factors that influence inversion at pyramidal nitrogen in
bicylic amines have been discussed for several decades.65 Usual
values of nitrogen inversion barriers for alicyclic amines lie in
the 5-9 kcal mol-1 range.65,66 However, abnormal barriers (>13
kcal mol-1) were found for azanorbornanes, which has been
implied to nitrogen inversion C-N rotation (NIR), the “bicyclic
effect”.65,67,68 Despite several attempts, the mechanistic reason
for this bicyclic effect could not be satisfactory explained in
terms of steric interactions or ring strain.69 Only in crowded
systems such as N-t-Bu azanorbornanes could steric interactions
be sufficiently strong enough to play an important factor. The
nitrogen inversion rotation barrier determined among a set of
various categories of bicyclic amines using dynamic NMR and
MP2/6-31G* is found to vary from 6.4 to 13 kcal mol-1 68
.
Several interesting conclusions have thus far emerged from
various works66,68 on nitrogen inversion and bicyclic effect in
cyclic amines: (1) The NIR barriers increase with a decrease
of the ring size in azabicycles. (2) The presence of a five-
(63) Atkins, P.; de Paula, J., Atkin’s Physical Chemistry, 7th ed.; Oxford
University Press: New York, 2002.
Thus, in pyridine-d5 (Figure 3), the minor isomer (12b) was
found to have converted almost completely (>99%) to the major
isomer (12a) in 33 h at 298 K and no reverse isomerism was
observed starting from the major isomer (12a) under our
experimental conditions. We have subsequently determined the
rate of inversion at 293 K (k ) 1.0 × 10-5 s-1), 298 K (k )
4.4 × 10-5 s-1), 303 K (k ) 8.0 × 10-5 s-1), 308 K (k ) 1.0
× 10-4 s-1), and 318 K (k ) 4.0 × 10-4 s-1), respectively,
(64) Inezedy, J.; Lengyel, T.; Ure, A. M. IUPAC Compendium of Analytical
Nomenclature, 3rd ed.; Blackwell Science, Oxford, U.K., 1997; p 15.
(65) Lehn, J. M. Chem. Forsch. 1970, 15, 311.
(66) Belostotskii, A. M.; Gottlieb, H. E.; Hassner, A. J. Am. Chem. Soc. 1996,
118, 7783.
(67) Nelsen, S. F.; Ippoliti, J. T.; Frigo, T. B.; Petillo, P. A. J. Am. Chem. Soc.
1989, 111, 1776.
(68) Belostotskii, A. M.; Gottlieb, H. E.; Shokhen, M. J. Org. Chem. 2002, 67,
9257.
(69) Davies, J. W.; Malpass, J. R.; Moss, R. E. Tetrahedron Lett. 1985, 26,
4533.
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15179

A R T I C L E S
Varghese et al.
membered ring as a component of a rigid nitrogen-bridged
bicyclic skeleton increases the NIR barrier by ca. 3 kcal mol-1
per ring, thereby showing a relationship between azabicycle
geometry and the NIR barrier. (3) The flattening of the nitrogen
pyramid, for example, through the introduction of a double bond
in the six-membered ring as a component of a rigid nitrogen-
bridged bicyclic skeleton, decreases the NIR barrier of the ring
inversion. (4) It has been evidenced by dynamic NMR at ∼180
K that the ring inversion in the rigid nitrogen-bridged bicyclic
skeleton involves interconversion of conformers with equatorial
and axial N-alkyl substituents. (5) Although an earlier suggestion
that high strain, which develops during NIR for the endocyclic
CNC angle change from N-pyramid, is responsible for the
bicyclic effect, no satisfactory correlation has been however
found between NIR barriers and the CNC angle for different
bicyclic amines. No experimental evidence has been found thus
far to support the suggestion that the observed high NIR barriers
in the constrained amines is caused by the delocalization of the
N lone pair.
In conclusion, the high ∆Gq found 23.4 kcal mol-1 at 298 K
for the conversion of axial N-H containing isomer 12b to the
energetically stable equatorial N-H containing isomer 12a is
unique in the long history65,67,68 of abnormally high activation
barriers for nitrogen inversion. This also constitutes the first
example in which both the piperidine isomers with the axial
and equatorial lone-pair orientation have been isolated in the
pure form and fully characterized by NMR and mass spectros-
copy.
rotational phase angle P ) 14 ( 7° for 12a and 12b and 19 (
8° for 13, sugar puckering amplitude φm ) 48 ( 4° for 12a,
12b, and 13; Table S7 in the Supporting Information), and the
sugar pucker in aza-ENA-T is close to that of the ENA and
LNA (P ) 12-19°),22 however, with the sugar puckering
amplitude, φm, lower by ∼10° (φm ≈ 46°) compared to that of
the LNA (φm ≈ 56°).22,23 ENA-T and aza-ENA-T have also
showed very similar conformational dynamics with 5-8° of
variation of the sugar torsions along the MD trajectories (Table
S7 in the Supporting Information), and the sugar atom position
root-mean-square deviation (RMSD) is less than 0.1 Å (Figure
S46 in the Supporting Information).
4.2. Conformation of the Piperidino Ring. Ab initio and
MD simulations, a 1D NOESY experiment (Figure 2), as well
as dihedral angles obtained using the generalized Karplus
equation70,71 (shown in Figure S45 in the Supporting Informa-
tion) point to the chair conformation for the piperidino
heterocycle. Both the sugar and piperidino rings of aza-ENA-T
show exceptional rigidity, with rmsd values of the sugar and
piperidino exocylic heavy (C, N, and O) atoms being less than
0.09 Å (Figure S46 in the Supporting Information) along the
MD trajectories. Higher dynamics have been observed for the
base (RMSD of about 0.7 Å) of aza-ENA-T (12a, 12b, and
13), while flanking OBn groups in the 3′,5′-bis-OBn-protected
aza-ENA-T compounds 12a and 12b (RMSD of 1.3-1.7 Å;
Figure S46 in the Supporting Information) are expectedly found
to be the most dynamic parts of these protected compounds.
5. Preparation of Aza-ENA-T Phosphoramidite for AON
Synthesis. We have then prepared both the N-phenoxyacetyl
(PAC)- and N-trifluoroacetyl-protected aza-ENA-T phosphor-
amidite blocks 17 and 20 from 12a/12b to synthesize aza-ENA-
T-incorporated AONs (Scheme 2).
4. Molecular Structure of the Aza-ENA-T Derivatives 12a,
12b, and 13 Based on NMR, ab initio, and Molecular
Dynamics (MD) Calculations. The experimental coupling
constants from 600 MHz spectra of the 3′,5′-bis-OBn-protected
(12a and 12b) and fully deprotected (13) aza-ENA-T compounds
have been further analyzed to build up their molecular structures
using the following protocol: (i) Derive initial dihedral angles
from the observed3 JHH using the Haasnoot-de Leeuw-Altona
generalized Karplus equation70,71 (Table S6 in the Supporting
Information). (ii) Perform NMR-constrained MD simulation (0.5
ns, 10 steps) simulated annealing (SA) followed by 0.5 ns of
NMR-constrained simulations at 298 K using the NMR-derived
torsional constraints from step (i) to yield NMR-defined
molecular structures of 3′,5′-bis-OBn-protected (12a and 12b)
and deprotected aza-ENA-T (13) (for details of theoretical
simulations, see the Supporting Information). Our conclusions
based on detailed MD, SA, and ab initio simulations are as
follows (for full details, see Tables S6 and S7 and Figures S44,
S45, and S46 in the Supporting Information).
The aza-ENA-T analogue 12a/12b was N-protected using
PAC-Cl in pyridine, affording 14 (70%) as a mixture of rotamers
because of the restricted rotation of the amide bond of PAC.
Debenzylation of 14 was achieved by doing a successive
reaction with Pd(OH)2/ammonium formate in methanol and then
with BCl3 in anhydrous CH2Cl2 for 3 h to give 15 in 75% yield.
Dimethoxytrytilation of 5′-OH in 15 using DMTr-Cl and
pyridine followed by phosphytilation of 3′-OH using standard
conditions31 afforded 16 and 17 in 90 and 86%, respectively
[31P NMR of 17 (CDCl3) δ: 150.7, 150.3, 149.2, and 148.1;
see Figure S29 in the Supporting Information].
The phosphoramidite 17 was successfully incorporated into
the mixed 15-mer sequence (Table 1), but to our surprise, the
PAC-protecting group was very stable and could not be
deprotected even in 33% aqueous ammonia and AMA [33%
aqueous ammonia/methylamine, 1:1 (v/v)] at 65 °C for 2 days,
which was clear from the mass measurement using MALDI-
TOF mass spectroscopy (expected mass with PAC protection,
m/z 4624.7; and observed, 4624.9). The PAC-protected nucleo-
side 15, on the other hand, could be deprotected with aqueous
ammonia at 55 °C overnight.
3
4.1. Sugar Pucker Conformation. Non-observable JH1′,H2′
and low 3JH2′,H3′ (Tables S3 and S6 in the Supporting Informa-
tion) experimental coupling indicate that, similar to that in the
ENA,22 LNA,22,23 and 2′-amino LNA,25 the piperidino modifica-
tion of the sugar moiety in aza-ENA-T restricts sugar pucker
to the North-type conformation (for details, see Tables S6 and
S7 and Figure S44 in the Supporting Information). Ab initio
and MD simulations (Tables S6 and S7 in the Supporting
Information) have shown that the sugar moiety is indeed
conformationally restricted to the North conformation (pseudo-
Because the PAC protection of the amino function of aza-
ENA-T did not work satisfactorily in our hands, we incorporated
a trifluoroacetyl-protecting group (Scheme 2), where the benzyls
were first deprotected using the same condition as for 15
(Scheme 2). The deprotected nucleoside 13 was directly treated
with an excess of ethyl trifluoroacetate in methanol at room
temperature overnight to give the N-trifluoroacetyl-protected
(70) Haasnoot, C. A. G.; de Leeuw, F. A. A. M.; Altona, C. Tetrahedron 1980,
36, 2783.
(71) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1972, 94, 8205.
9
15180 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Conformationally Constrained Aza-ENA-T
A R T I C L E S
nucleoside 18 in 45% yield after 3 steps from 12 as a mixure
of rotamers. Dimethoxytritylation and phosphitylation were
carried out using the same conditions as for 16 and 17 in Scheme
2 to afford 19 in 81% yield and 20 in 61% yield as a mixture
of four isomers [31P NMR (CDCl3) δ: 150.1, 149.9, 149.8, and
149.2; Figure S36 in the Supporting Information]. Phosphor-
amidite 20 was successfully incorporated at different positions
of mixed 15-mer AON sequences (Table 1), using a standard
phosphoramidite approach,31 with a 10 min coupling time and
a coupling efficiency of ∼95%. Deprotection of all of the base-
labile protecting groups went smoothly, with 33% aqueous
ammonia at 55 °C as confirmed by MALDI-TOF mass
spectroscopy (Table 1). All oligos were purified by polyacryl-
amide gel electrophoresis (PAGE) (20% polyacrylamide/7 M
urea), extracted with 0.3 M NaOAc, and desalted with a C18
reverse-phase cartridge to give AONs in >99% purity.
6. Thermal Denaturation Studies of Aza-ENA-T-Modified
AONs. The thermal stability of duplexes involving aza-ENA-
Ts was determined by the complementary RNA and DNA as
shown in Table 1. Single modifications were incorporated one
at a time at different positions (Table 1) of the 15-nucleotide
AON sequence, 3′-d(CTTCTTTTTTACTTC)-5′, to determine
the sequence dependency in the target affinity. The results reveal
(Table 1) that the single aza-ENA-T modification enhances the
target affinity significantly with complementary RNA (∆Tm )
+2.5 to +4 °C), depending upon the site of the modification in
the AON strand. This can be attributed to the site dependency
of the variable conformational pre-organization imparted by the
North-fused sugar moiety on the single-stranded AON. Even
though the thermal stabilities of duplexes containing single ENA
modifications were not reported,22,42 we presume that aza-
ENA-T will have slightly more favorable target affinity toward
RNA than the isosteric ENA counterpart (double modification
gave +3.5 °C per modification).22 The 2′-amino function of
aza-ENA-T is almost 50% protonated at the physiological pH
considering its pKa of 6.66 ( 0.03 compared to that of the 2′-
amino-LNA counterpart (pKa of 6.17 ( 0.03), determined using
pH-dependent 1H NMR measurements72 (see the Experimental
Methods and Figure S48 in the Supporting Information for
details). This means that the aza-ENA-T- or 2′-amino-LNA-T-
incorporated AONs can have an electrostatic interaction with
the neighboring phosphate at the physiological pH, which favors
efficient duplex formation as observed in azetidine-modified
AONs.31 On the other hand, with complementary DNA, there
was a significant drop in duplex melting. This can be due to
the 2′,4′-ethylene bridge, which causes a steric clash in the minor
groove of the AON-DNA duplex.
model system. Four different aza-ENA-T-containing AON
mixtures, each obtained by incorporating single aza-ENA-T
modification at one of the four different positions (AONs 2-5
in Table 1), of identical DNA sequence, when formed duplex
with the complementary RNA, were found to be excellent
substrates for RNase H1 but with varying RNA cleavage sites
depending upon the site of modification on the AON strand.
We have previously reported the RNase H1 digestion properties
of oxetane-modified AON/RNA hybrid duplexes in an identical
sequence.74 The RNA cleavage patterns of all aza-ENA-T-
modified AONs (AON 2-5 in Table 1) were found to be
uniquely different from those of the isosequential oxetane-
modified AONs (AON 6-9; Figure 6). AON 2 showed only
one prominent cleavage site at the A8 position of the comple-
mentary RNA (Figure 5), unlike the oxetane-modified AON 6,
which showed cleavages at A7, A8, A10, and U11, with no
clear preferences (Figure 6). A comparison of AON 4 versus
AON 8 and AON 5 versus AON 9 clearly shows the absence
of a single RNA cleavage site in aza-ENA-T-modified AONs/
RNA duplexes compared to those of the oxetane-modified
counterparts (A7 of AON 8 and A9 of AON 9). AON 3 and its
isosequential oxetane analogue AON 7, on the other hand,
showed an identical cleavage footprint pattern of a 5-nucleotide
gap. This shows that the local structures of all aza-ENA-T-
modified AONs/RNA duplexes are not the same. The RNase
H enzyme indeed can finely discriminate these local variations
of the stereochemical properties of the microstructure brought
about by various types and incorporation sites of the North-
type modification (aza-ENA-T versus oxetane modifications)
in the AON. These local structure variations were however not
observable by circular dichroism (CD) spectra (see Figure S3
and Discussion S4 in the Supporting Information).
This shows that the certain specific flexibility of the AON/
RNA duplex and accessibility of the RNA strand in the
heteroduplex is required for RNase H binding and cleavage.
The RNase H1 recruitment capability for LNA-modified AONs
were reported earlier,73 which showed a minimum gap of 7-8
DNA monomers to induce full cleavage activity. This difference
in RNase H activity between the LNA-, oxetane-, and aza-ENA-
T-modified AONs/RNA duplexes can be due to the difference
in the conformational flexibility as well as the resulting hydra-
tion pattern in the minor groove imparted by the fused four-,
five-, and six-membered rings locked to the pentofuranose ring.
Finally, to evaluate the cleavage rate, quantification of gels
was performed densitometrically and the uncleaved RNA
fraction was plotted as a function of the incubation time (Figure
5B). The enzyme digestion experiment was performed at a lower
enzyme concentration (0.08 unit) to observe the cleavage rate
(see Figure S4 in the Supporting Information).
7. RNase H Digestion Studies of Aza-ENA-T-Modified
AON/RNA Heteroduplexes. Escherichia coli RNase H1 has
been used in this work because of two reasons: first, it is
commercially available in a pure form, and second, its cleavage
properties are not very different from those of the mammalian
enzyme.73 Hence, the antisense properties of aza-ENA-T-
modified oligonucleotides in duplex with the complementary
RNA were compared with the native as well as with the identical
oxetane-modified counterparts,74 with E. coli RNase H1 as a
Reaction rates were determined by fitting to single-exponen-
tial decay functions. Recently, Kurreck et al.73 have shown that
the RNase H cleavage efficiency of AON can be correlated with
its affinity toward target RNA. The relative cleavage rates with
aza-ENA-T-modified AON/RNA duplex were however quite
comparable to that of the native counterpart (Table S1 in the
Supporting Information and parts B and D in Figure 5).
(72) Acharya, S.; Barman, J.; Cheruku, P.; Chatterjee, S.; Acharya, P.; Isaksson,
J.; Chattopadhyaya, J. J. Am. Chem. Soc. 2004, 126, 8674.
(73) Kurreck, J.; Wyszko, E.; Gillen, C.; Erdmann, V. A. Nucleic Acids Res.
2002, 30, 1911.
(74) Pradeepkumar, P. I.; Zamaratski, E.; Foldesi, A.; Chattopadhyaya, J.
Tetrahedron Lett. 2000, 41, 8601.
8. Stability of Aza-ENA-T/DNA Chimeras in Human
Serum. The stability of AON in cells toward various exo- and
endonucleases is warranted to fulfill the requirements for an
ideal antisense agent.6 The stabilities of aza-ENA-T-modified
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15181

A R T I C L E S
Varghese et al.
Figure 5. (A) Autoradiograms of 20% denaturing PAGE, showing the cleavage kinetics of 5′-32P-labeled target RNA by E. coli RNase H1 in native AON
1/RNA (lane 5) and the aza-ENA-T-modified AONs (2-5)/RNA hybrid duplexes (lanes 1-4) after 2, 5, 15, 35, and 60 min of incubation. Conditions of
cleavage reactions: RNA (0.8 µM) and AONs (4 µM) in buffer containing 20 mM Tris-HCl (pH 8.0), 20 mM KCl, 10 mM MgCl2, and 0.1 mM DTT at
21 °C; 0.08 unit of RNase H. Total reaction volume 30 µL. (B) Kinetics of RNase H cleavage. The target RNA remaining is densitometrically evaluated and
plotted as a function of time. The inset in the right corner shows the initial cleavage rates for clarity. (C) Pictorial representation of RNase H1 cleavage
pattern of AONs 1-5/RNA hybrid duplexes. Vertical arrows show the RNase H cleavage sites, with the relative length of an arrow showing the relative
extension of cleavage at that site, and dotted arrows show the partial cleavage at the initial reaction time. The relative percentage of cleavage is indicated
above the arrow, which is taken at a 15 min time point from the gel shown above. (D) Quantitative evaluation of the gel picture (shown in the Supporting
Information) of the remaining full-length [32P]RNA at 1 h as obtained by the densitometer.
AONs were tested against human serum, which is mainly
comprised of 3′-exonucleases (Figure 7). When compared to
the native counterpart, which completely degraded after 9 h,
AON 3, 4, and 5 (Table 1) were still remaining to a certain
extent (8, 15, and 20%, respectively). It is noteworthy that all
modified AONs were cleaved by 3′-exonucleases in the blood
serum at the phosphodiester, which is one nucleotide before
the aza-ENA-T modification site toward the 3′ end, and the
residual sequences were found to be stable in human serum for
48 h at 21 °C (Figure S5 in the Supporting Information). This
is a surprising result in view of the fact that identical AON
sequences with North-constrained oxetane75 modification are
cleaved at the phosphodiester immediately before the modifica-
tion site under an identical condition. This suggests that the
conformational effect of the aza-ENA-T modification in the
AON is transmitted toward the 3′ end and recognized by 3′-
exonucleases, just as in the RNase H cleavage of the AON/
RNA duplex, which recognizes the local RNA/RNA-type duplex
structure and leaves a footprint at the 5′ end because of the
modulation of the structure of the complementary RNA strand
by the North-type constrained aza-ENA-T in the AON strand.
The stability studies in human serum with a single LNA73
nucleotide at the 3′ and 5′ ends showed complete degradation
in 24 h at 37 °C, while our residual sequences from the aza-
ENA-T modifications were stable over 48 h under our experi-
mental condition (Figure S5 in the Supporting Information).
Even though a direct comparison could not be made, these
results clearly show that aza-ENA-T modification can certainly
give substantial stability in human serum, which is probably
more than that of LNA.
(75) Pradeepkumar, P. I.; Amirkhanov, N. V.; Chattopadhyaya, J. Org. Biomol.
Chem. 2003, 1, 81.
9
15182 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Conformationally Constrained Aza-ENA-T
A R T I C L E S
Figure 6. (A and B) are autoradiograms of 20% denaturing PAGE, showing the cleavage kinetics of 5′-32P-labeled target RNA by E. coli RNase H1 in
native AON 1/RNA and the oxetane-modified AONs (6-9)/RNA hybrid duplexes after 30 min and 1 and 2 h of incubation. Conditions of cleavage
reactions: RNA (0.8 µM) and AONs (4 µM) in buffer containing 20 mM Tris-HCl (pH 8.0), 20 mM KCl, 10 mM MgCl2, and 0.1 mM DTT at 21 °C; 0.08
unit of RNase H. Total reaction volume of 30 µL. (C) Pictorial representation of the RNase H1 cleavage pattern of AONs (6-9)/RNA hybrid duplexes.
Vertical arrows show the RNase H cleavage sites, with the relative length of an arrow showing the relative extension of cleavage at that site, and dotted
arrows show the cleavage at the initial 30 min of the reaction.
Figure 7. Autoradiograms of 20% denaturing PAGE, showing the degradation pattern of 5′-32P-labeled AONs 1-5 in human serum. Time points are taken
after 0, 15, and 30 min and 1, 2, and 9 h of incubation. The percentage of AON remaining after 9 h of incubation: 0% of AON 1, 0% of AON 2, 8% of
AON 3, 15% of AON 4, and 20% of AON 5.
9. 3′-Exonuclease Stability Assay [Snake Venom Phospho-
diesterase (SVPDE)]. The stability of aza-ENA-T-modified
AONs 1-5 toward 3′-exonuclease was investigated using
SVPDE over a period of 24 h at 21 °C. Time points were taken
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15183

A R T I C L E S
Varghese et al.
while hybridization with the complementary DNA sequence lead
to substantial destabilization of the duplexes (Tm drop of -0.5
to -3 °C per modification). (4) All of the aza-ENA-T-modified
AON/RNA hybrid duplexes have been found to be good
substrates for the E. coli RNase H1. In these AON/RNA hybrids,
except for one case (AON 2, Figure 8), a region of 5-6
nucleotides in the RNA strand in the 3′-end direction from the
site opposite to the aza-ENA-T modification was found to be
insensitive toward RNase H cleavage, presumably owing to the
local structural perturbations brought about by the conforma-
tional constrain. These cleavage patterns of the aza-ENA-T-
modified AON/RNA hybrids are uniquely different from that
of the oxetane-modified AONs, which had shown a gap of 5
nucleotide units. (5) All of the aza-ENA-T-modified AONs
offered greater protection toward 3′-exonucleases compared to
the native sequence (stable for over 48 h in human serum). (6)
This study provides valuable information regarding the optimal
design of AONs having a completely natural phosphodiester
backbone for the therapeutic applications that will not only show
high target affinity but also favorable RNase H recruitment as
well as high stability toward nucleases in ViVo.
Figure 8. Denaturing PAGE analysis of the SVPDE degradation pattern
of 5′-32P-labeled AONs 1-5. Time points are taken after 0, 1, 2, and 24 h
of incubation with the enzyme. The percentage of AON remaining after 24
h of incubation: 0% of AON 1, 18% of AON 2, 20% of AON 3, 14% of
AON 4, and 10% of AON 5.
Experimental Section
Compound C in Scheme 1: [(1R,3R,4R,6S)-1-Benzyloxymethyl-
6-benzyloxy-5-carbonitrile-3-(thymin-1-yl)-2-oxabicyclo[2.1.1]-
hexane]. The nucleoside 11 (2.5 g, 4.1 mmol) was dissolved in 45 mL
of dry THF and cooled in an ice bath, and 1 M NaHMDS (8.2 mL)
was added dropwise. The reaction was warmed slowly to room
temperature and stirred for 3 h under nitrogen, which was then quenched
by adding water and extracted with CH2Cl2 (3 times). The organic phase
was dried over anhydrous MgSO4, filtered, and concentrated under
reduced pressure. The crude residue was purified by silica gel column
chromatography (0-3% methanol in dichloromethane, v/v), which
afforded C in Scheme 1 with traces of another diastereomer as shown
by NMR (655 mg, 1.4 mmol, 35%). Rf ) 0.40 [96:4 CH2Cl2/CH3OH
at 0, 1, 2, and 24 h to examine the cleavage pattern (Figure 8).
The enzyme digestion pattern was similar to that obtained from
the digestion with human serum, except the fact that these aza-
ENA-T-modified AONs offered resistance to degradation even
after 24 h. Note that native AON 1 was completely degraded
in 1 h, whereas the full-length AONs 2-5 were 18, 20, 14, and
10% left undegraded, respectively, after 24 h. It is however
noteworthy that, similar to the blood serum digestion, all
modified AONs were cleaved by SVPDE at the phosphodiester,
which is one nucleotide before the aza-ENA-T modification site
toward the 3′ end, and the residual sequences were found to be
fully stable for 24 h at 21 °C (Figure 8). This is another proof
for the future design of aza-ENA-T-modified AONs, where only
a single modification at the second position from the 3′ end
will offer significant stability toward 3′-exonucleases.
1
(v/v)]. MALDI-TOF m/z: [M + H]+ found, 460.9; calcd, 459.1. H
NMR (500 MHz, major isomer, CDCl3) δ: 8.93 (s, 1H, NH, thymine),
7.45-7.12 (m, 10H, Bn), 6.92 (q, J ) 1.2 Hz, 1H, H6), 5.81 (s, 1H,
H6′; H6′ appears as t with a resolution enhancement, W coupling, JH6′,H3′
) 1.5 H and JH6′,H5′ ) 1.5 Hz), 5.08 (d, JH1′,H2′ ) 2.5 Hz, 1H, H1′),
4.58 (d, Jgem ) 13.4 Hz, 1H, CH2Ph), 4.50-4.46 (m, 5H, CH2Ph, H5′,
H5′′), 3.96 (dd, JH2′,H3′ ) 7.7, JH3′,H6′ ) 1.5 Hz, 1H, H3′), 3.25 (dd,
JH1′,H2′ ) 7.7, JH2′,H3′ ) 2.5 Hz, 1H, H2′), 1.50 (d, J ) 1.2 Hz, CH3,
thymine). 13C NMR (125.7 MHz, CDCl3) δ: 163.2 (C4), 159.9 (C4′),
150.2 (C2), 136.6, 135.7, 134.4 (C6), 128.6, 128.5, 128.4, 128.2, 128.0,
127.8, 115.1 (CN), 110.3 (C5), 99.3 (C6′), 73.2 (CH2Ph), 71.7 (CH2Ph),
70.7 (C5′), 68.3 (C3′), 63.6 (C1′), 58.8 (C2′), 11.9 (CH3, thymine). 1H
NMR (500 MHz, minor isomer, CDCl3) δ: 7.39-7.15 (m, 10H, Bn),
7.07 (q, J ) 1.5 Hz, 1H, H6), 5.09 (d, JH1′,H2′ ) 2.5 Hz, 1H, H1′),
4.65-4.40 (m, 6H, 2×CH2Ph, H5′, H5′′), 4.06 (d, JH2′,H3′ ) 8.2 Hz,
1H, H3′), 3.35 (dd, JH2′,H3′ ) 8.2 Hz, JH1′,H2′ ) 2.5 Hz, 1H, H2′), 1.78
(d, J ) 1.5 Hz, CH3, thymine). 13C NMR (125.7 MHz, CDCl3) δ: 163.4,
160.0, 150.0, 137.0, 136.4, 135.3, 134.6, 128.9-127.6, 117.6, 115.5,
114.8, 112.9, 99.7, 73.8, 71.8, 71.3, 68.1, 64.9, 57.7, 12.2.
Conclusions
(1) In the cyclization of 11 f 12a + 12b, two pure
diastereomers were isolated because of the axial and equatorial
orientation of the chiral piperidino-NH proton (12a, NHe; 12b,
NHa) for the first time. The conversion of 12b to 12a in pyridine-
d5 is nonreversible (∆Gq at 298 K ) 23.4 kcal mol-1). In CDCl3,
1
diastereomers 12a / 12b (60:40 by H NMR) were however
in dynamic equilibrium with Kc ) 0.67, with the ∆Gq ) 25.4
kcal mol-1. (2) The molecular structures of the aza-ENA-T
monomer units by high-field 1H NMR and theoretical ab initio
and MD simulations have shown that the piperidino-fused
furanose ring is indeed locked in the typical North-type
conformation, with the pseudorotational phase angle (P) and
puckering amplitude (φm) for the ab initio optimized geometries
(HF, 6-31G**) varying in the ranges 7° < P < 27° and 44° <
φm < 52°, respectively. (3) Aza-ENA-T-modified AONs have
shown high target affinity to the complementary RNA strand
(Tm increase of +2.5 to +4 °C per modification), depending
upon the substitution site, compared to the native counterpart,
3,5-Di-O-benzyl-4-C-hydroxymethyl-1,2-O-isopropylidene-R-D-ribo-
furanose (2). To a stirred suspension of 1 (12.3 g, 39.5 mmol) in
anhydrous acetonitrile (400 mL) at -5 °C was added NaH (1.81 g,
1.15 mmol) in four portions during 1.5 h. Benzyl bromide (5.4 mL,
1.15 mmol) was added dropwise and stirred at room temperature
overnight under nitrogen atmosphere. The reaction was quenched with
water and extracted with dichloromethane. The organic phase was dried
over anhydrous MgSO4 and evaporated under reduced pressure. The
crude product was purified by silica gel column chromatography (0-
9
15184 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Conformationally Constrained Aza-ENA-T
A R T I C L E S
20% ethyl acetate in cyclohexane, v/v), which afforded 2 (10.6 g, 26.5
mmol, 67%). All analytical data were identical to those previously
reported.21
partially evaporated under reduced pressure and extracted with di-
chloromethane. The combined organic phase was dried over anhydrous
MgSO4 and evaporated to give 7 (more than 90% pure by NMR) as a
white solid. The crude product was co-evaporated with dry pyridine 3
times to remove traces of moisture and dissolved in 60 mL of the same
solvent. The reaction was cooled in an ice bath, and methanesulfonyl
chloride (1.8 mL, 23.6 mmol) was added dropwise and stirred at 0 °C
for 6 h. The reaction was quenched with saturated aqueous NaHCO3
and extracted with dichloromethane. The organic phase was dried over
anhydrous MgSO4, evaporated under reduced pressure, and co-
evaporated 3 times with toluene and 3 times with dichloromethane.
The product was purified using silica gel column chromatography (0-
3% methanol in dichloromethane, v/v) to afford 8 (6.2 g, 11.2 mmol,
95%). Rf ) 0.41 [96:4 CH2Cl2/CH3OH (v/v)]. MALDI-TOF m/z: [M
+ H]+ found, 556.0; calcd, 555.1. 1H NMR (270 MHz, CDCl3) δ: 10.1
(br s, 1H, NH), 7.46 (s, H6), 7.33-7.2 (m, 10H, benzyl), 6.05 (d, JH1′,H2′
) 2.72 Hz, 1H, H1′), 5.39 (m, 1H, H2′), 4.84 (d, Jgem ) 11.38 Hz, 1H,
3,5-Di-O-benzyl-4-C-cyanomethyl-1,2-O-isopropylidene-R-D-ribo-
furanose (4). The sugar 2 (10.6 g, 26.5 mmol) was dissolved in an
anhydrous dichloromethane/pyridine mixture (250 mL, 3:1, v/v) and
cooled in an ice bath. To this solution, triflic anhydride (5.3 mL, 31.8
mmol) was added dropwise and stirred for 3 h under nitrogen
atmosphere. The reaction was quenched with cold saturated aqueous
NaHCO3 and extracted with dichloromethane. The organic phase was
dried over anhydrous MgSO4 and evaporated under reduced pressure
followed by co-evaporation with toluene 3 times and dichloromethane
3 times. The crude reaction product was dissolved in 150 mL of dry
DMF, and 80 mL of 1 M LiCN in DMF was added and stirred for 3
days at room temperature. The solvent was evaporated, and the residue
was dissolved in dichloromethane. Saturated aqueous NaHCO3 was
added and extracted with dichloromethane (3 times). The organic phase
was dried over anhydrous MgSO4 and evaporated under reduced
pressure. The crude product was purified by silica gel column
chromatography (0-20% ethylacetate in cyclohexane, v/v), which
afforded 4 (6.06 g, 14.8 mmol, 56%). Rf ) 0.61 [60:40 cyclohexane/
ethylacetate (v/v)]. MALDI-TOF m/z: [M + H]+ found, 410.0; calcd,
CH2Ph), 4.53-4.40 (m, 4H, H3′, CH2Ph, 2×CH2Ph), 3.88 (d, Jgem
)
10.27 Hz, 1H, H5′), 3.56 (d, 1H, H5′′), 3.15 (s, 3H, OMs), 2.92 (d,
Jgem ) 17.32 Hz, 1H, H6′), 2.71 (1H, H6′′), 1.45 (s, 3H, CH3, thymine).
13C NMR (67.9 MHz, CDCl3) δ: 163.7 (C4), 150.5 (C2), 136.3 (C6),
136.2, 135.3, 128.7, 128.4, 128.3, 128.1, 128.0, 116.1 (CN), 111.1 (q,
C5), 85.6 (C1′), 84.7 (q, C4′), 79.2 (C2′), 75.3 (C3′), 73.7 (CH2Ph),
73.4 (CH2Ph), 70.8 (C5′), 38.6 (OMs), 22.1 (C6′), 11.6 (CH3, thymine).
1
409.1. H NMR (270 MHz, CDCl3) δ: 7.32-7.2 (m, 10H, benzyl),
5.71 (d, JH1,H2 ) 3.71 Hz, 1H, H1), 4.74 (d, Jgem ) 12 Hz, 1H, CH2Ph),
4.59-4.52 (m, 4H, H2, CH2Ph, 2×CH2Ph), 4.06 (d, JH2,H3 ) 4.95 Hz,
2,2′-Anhydro-1[3,5-di-O-benzyl-4-C-cyanomethyl-â-D-ribofuran-
osyl]thymine (9). Nucleoside 8 (6.2 g, 11.2 mmol) was dissolved in
70 mL of anhydrous CH3CN, and DBU (1.75 mL, 11.76 mmol) was
added dropwise and stirred at room temperature for 1 h. The reaction
was quenched with water and extracted with dichloromethane. The
combined organic phase was dried over anhydrous MgSO4 and
evaporated under reduced pressure. The crude product was purified by
silica gel column chromatography (0-4% methanol in dichloromethane,
v/v) to afford 9 (4.7 g, 10.2 mmol, 91%). Rf ) 0.21 [96:4 CH2Cl2/
CH3OH (v/v)]. MALDI-TOF m/z: [M + H]+ found, 460.0; calcd,
1H, H3), 3.5 (ABq, Jgem ) 10.39 Hz, 2H, H5′, H5′′), 3.15 (d, Jgem
)
17.07 Hz, 1H, H6′), 2.86 (d, 1H, H6′′), 1.58 (s, 3H, CH3, isopropyl),
1.32 (s, 3H, CH3, isopropyl). 13C NMR (67.9 MHz, CDCl3) δ: 137.3,
137, 128.3, 128.2, 127.7, 127.5, 117.0 (CN), 113.4 (q, isopropyl), 103.9
(C1), 83.2 (q, C4), 78.4 (C2), 78 (C3), 73.6 (CH2Ph), 72.4 (CH2Ph,
C5), 26.5 (CH3, isopropyl), 25.6 (CH3, isopropyl), 22.1 (C6).
1-[2-O-Acetyl-3,5-di-O-benzyl-4-C-cyanomethyl-â-D-ribofurano-
syl]thymine (6). Triflic acid (0.065 mL, 0.74 mmol) was added
dropwise to a stirred solution of 4 (6.06 g, 14.8 mmol) in acetic acid
(89 mL) and acetic anhydride (16.7 mL, 177.6 mmol). The solution
was stirred for 3 h at room temperature and then poured into a cold
NaHCO3 solution. The mixture was extacted with dichloromethane,
and the organic phase was dried over anhydrous MgSO4 and evaporated
under reduced pressure. The crude product was co-evaporated several
times with dry toluene until the product solidifies to give 5 (more than
90% pure by NMR). The crude product was dissolved in 150 mL of
anhydrous CH3CN, and thymine (2.24 g, 17.7 mmol) and N,O-bis-
(trimethylsilyl)acetamide (10.2 mL, 41.4 mmol) were added and stirred
at 80 °C for 1 h under a nitrogen atmosphere. The reaction mixture
was cooled to room temperature, and TMSOTf (3.48 mL, 19.24 mmol)
was added, again warmed to 80 °C, and stirred overnight under a
nitrogen atmosphere. The reaction was quenched with saturated aqueous
NaHCO3 and extracted with dichloromethane. The organic phase was
dried over anhydrous MgSO4, evaporated under reduced pressure, and
purified by silica gel column chromatography (0-3% methanol in
dichloromethane, v/v) to afford 6 (6.14 g, 11.8 mmol, 80%). Rf ) 0.56
[96:4 CH2Cl2/CH3OH (v/v)]. MALDI-TOF m/z: [M + H]+ found,
1
459.1. H NMR (270 MHz, CDCl3) δ: 7.35-7.12 (m, 11H, benzyl,
H6), 6.25 (br d, JH1′,H2′ ) 3.8 Hz, 1H, H1′), 5.32 (br d, 1H, JH2′,H1′
)
3.6 Hz, H2′), 4.77 (d, Jgem ) 11.63 Hz, 1H, CH2Ph), 4.62 (d, 1H,
CH2Ph), 4.43-4.31 (m, 3H, H3′, 2×CH2Ph), 3.36 (d, Jgem ) 10.02
Hz, 1H, H5′), 3.23 (d, 1H, H5′′), 2.80 (d, Jgem ) 16.70 Hz, 1H, H6′),
2.7 (d, 1H, H6′′), 1.96 (s, 3H, CH3, thymine). 13C NMR (67.9 MHz,
CDCl3) δ: 171.9 (C4), 159.0 (C2), 136, 135.4, 129.7 (C6), 128.7, 128.6,
128.4, 128.1, 127.8, 119.4 (q, C5), 116.4 (CN), 89.6 (C1′), 88.1 (q,
C4′), 85.7 (C2′), 83.4 (C3′), 73.7 (CH2Ph), 73.3 (CH2Ph), 71.1 (C5′),
22.2 (C6′), 13.9 (CH3, thymine).
1-[3,5-Di-O-benzyl-4-C-cyanomethyl-â-D-arabinofuranosyl]thy-
mine (10). To a solution of 9 (4.7 g, 10.2 mmol) in 200 mL of acetone,
204 mL of 0.1 M H2SO4 was added and refluxed overnight with stirring.
The solvent was partially evaporated, and saturated aqueous NaHCO3
was added and extracted with dichloromethane. The organic phase was
dried over anhydrous MgSO4 and evaporated under reduced pressure
to give 10 quantitatively. Rf ) 0.35 [96:4 CH2Cl2/CH3OH (v/v)].
1
MALDI-TOF m/z: [M + H]+ found, 478.1; calcd, 477.1. H NMR
1
520.0; calcd, 519.2. H NMR (270 MHz, CDCl3) δ: 8.8 (br s, 1H,
(270 MHz, CDCl3) δ: 11.1 (br s, 1H, NH), 7.38-7.23 (m, 11H, benzyl,
H6), 6.18 (d, JH1′,H2′ ) 3.09 Hz, 1H, H1′), 5.22 (d, J ) 4.45 Hz, 1H,
2′-OH), 4.89 (m, 1H, H2′), 4.66-4.52 (m, 3H, CH2Ph, 2×CH2Ph),
4.41 (d, Jgem ) 11.5 Hz, 1H, CH2Ph), 4.05 (s, 1H, H3′), 3.80 (ABq,
Jgem ) 9.5 Hz, 2H, H5′, H5′′), 2.90 (d, Jgem ) 16.70 Hz, 1H, H6′),
2.73 (d, 1H, H6′′), 1.62 (s, 3H, CH3, thymine). 13C NMR (67.9 MHz,
CDCl3) δ: 165.9 (C4), 150.3 (C2), 138.7 (C6), 137.1, 136.6, 128.3,
128.1, 127.9, 127.8, 127.5, 117.1 (CN), 107.6 (q, C5), 87.3 (C1′), 84.5
(q, C4′), 83.5 (C3′), 73.5 (CH2Ph), 72.9 (C2′), 71.9 (CH2Ph), 70.8 (C5′),
21.6 (C6′), 12.1 (CH3, thymine).
NH), 7.35-7.24 (m, 11H, benzyl, H6), 6.1 (d, JH1′,H2′ ) 4.8 Hz, 1H,
H1′), 5.50 (app t, J ) 5.32 Hz, 1H, H2′), 4.62 (d, Jgem ) 11.13 Hz,
CH2Ph), 4.53-4.44 (m, 4H, H3′, CH2Ph, 2×CH2Ph), 3.80 (d, Jgem
)
10.14 Hz, 1H, H5′), 3.63 (d, 1H, H5′′), 2.75 (ABq, Jgem ) 17.07 Hz,
2H, H6′, H6′′), 2.10 (s, 3H, OAc), 1.60 (s, 3H, CH3, thymine). 13C
NMR (67.9 MHz, CDCl3) δ: 169.7 (CO), 163.3 (C4), 150.1 (C2),
136.6, 135.8 (C6), 128.6, 128.4, 128.1, 128.0, 116.3 (CN), 111.5 (q,
C5), 87.9 (C1′), 84.7 (q, C4′), 77.1 (C3′), 74.7 (CH2Ph), 74.3 (C2′),
73.7 (C5′), 22.2 (C6′), 20.5 (CH3, OAc), 11.9 (CH3, thymine).
1-[3,5-Di-O-benzyl-4-C-cyanomethyl-2-O-methanesulfonyl-â-D-
ribofuranosyl]thymine (8). Nucleoside 6 (6.14 g, 11.8 mmol) was
dissolved in methanol (60 mL), and 18 mL of 1 M sodium methoxide
was added and stirred at room temperature for 3 h. The solvent was
1-[3,5-Di-O-benzyl-4-C-cyanomethyl-2-O-trifluoromethanesul-
fonyl-â-D-arabinofuranosyl]thymine (11). Nucleoside 10 (4.8 g, 10.2
mmol) was co-evaporated 3 times with dry pyridine to remove traces
of moisture and dissolved in a mixture of anhydrous dichloromethane
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15185

A R T I C L E S
Varghese et al.
(40 mL) and anhydrous pyridine (10 mL). To this solution, DMAP (5
g, 40.8 mmol) was added and cooled in an ice bath, and trifluoro-
methanesulfonic anhydride was added dropwise and stirred under a
nitrogen atmosphere for 2.5 h. The reaction was quenched with cold
saturated aqueous NaHCO3 and extracted with dichloromethane. The
organic phase was dried over anhydrous MgSO4, evaporated under
reduced pressure, and co-evaporated with toluene (3 times) and
dichloromethane (3 times). The crude product was purified by silica
gel column chromatography (0-2% methanol in dichloromethane, v/v)
to afford 11 (5.2 g, 8.5 mmol, 84%). Rf ) 0.58 [96:4 CH2Cl2/CH3OH
(v/v)]. MALDI-TOF m/z: [M + H]+ found, 609.99; calcd, 609.1. 1H
NMR (270 MHz, CDCl3) δ: 9.47 (br s, 1H, NH), 7.38-7.20 (m, 10H,
(CH2Ph), 72.0 (C3′), 69.5 (C5′), 64.2 (C2′), 45.9 (C7′), 26.8 (C6′),
11.9 (CH3, thymine).
(1R,5R,7R,8S)-8-Hydroxy-5-hydroxymethyl-7-(thymin-1-yl)-2-
aza-6-oxabicyclo[3.2.1]octane (13). Nucleoside 12a/12b (1.6 g, 3.4
mmol) was dissolved in 15 mL of methanol, and 20% Pd(OH)2 on
charcoal (615 mg) was added, followed by ammonium formate (2.5 g,
40 mmol), and refluxed for 12 h. The catalyst was filtered off through
a celite bed, and the filtrate evaporated under reduced pressure and
co-evaporated with dichloromethane to remove traces of methanol. The
crude material was dissolved in 20 mL of anhydrous dichloromethane
and cooled to -78 °C, and 1 M BCl3 (27 mL) was added and stirred
under nitrogen atmosphere for 3 h. The reaction was quenched by
adding methanol, and volatile materials were removed under reduced
pressure to give 13, which was purified for characterization using silica
gel column chromatography (0-20% methanol in dichloromethane, v/v)
to give 13 in 60% yield. Rf ) 0.18 [80:20 CH2Cl2/CH3OH (v/v)].
benzyl), 7.15 (d, JH6,H-CH ) 1.11 Hz, 1H, H6), 6.33 (d, JH1′,H2′ ) 3.59
3
Hz, 1H, H1′), 5.47 (br d, J ) 2.85 Hz, 1H, H2′), 4.81 (d, Jgem ) 11.63
Hz, 1H, CH2Ph), 4.59-4.48 (m, 3H, CH2Ph, 2×CH2Ph), 4.42 (s, 1H,
H3′), 3.73 (d, Jgem ) 9.65 Hz, 1H, H5′), 3.57 (d, 1H, H5′′), 2.90 (d,
Jgem ) 16.95 Hz, 1H, H6′), 2.76 (d, 1H, H6′′), 1.85 (d, JH6,H-CH3 ) 1.1
Hz, 3H, CH3, thymine). 13C NMR (67.9 MHz, CDCl3) δ: 163.2 (C4),
149.8 (C2), 136.4 (C6), 135.1, 134.9, 128.6, 128.5, 128.2, 128.1, 127.6,
118 (q, J ) 320.2 Hz, CF3), 116.3 (CN), 111.2 (C5), 85.4 (C2′), 83.9
(C1′), 83.6 (C4′), 82.2 (C3′), 73.66 (CH2Ph), 73.60 (CH2Ph), 70.7 (C5′),
21.7 (C6′), 12.2 (CH3, thymine).
1
MALDI-TOF m/z: [M + H]+ found, 284.2; calcd, 284.1. H NMR
(600 MHz, DMSO-d6) δ: 11.30 [s, 1H, NH (thymine)], 8.27 (s, 1H,
H6), 5.84 (s, 1H, H1′), 5.39 [t, J ) 5.0, 5.0 Hz, 1H, OH(5′)], 5.16 [br,
1H, OH(3′)], 3.99 (dd, J ) 3.9, 4.6 Hz, 1H, H3′), 3.57 (dd, J ) 12.2,
5.2 Hz, 1H, H5′), 3.50 (dd, J ) 12.2, 5.2 Hz, 1H, H5′′), 3.25 (d, J )
3.19 Hz, 1H, H2′), 2.95 (dt, J ) 12.8, 13.0, 4.8 Hz, 1H, H7a′), 2.87
(dd, J ) 12.8, 6.6 Hz, 1H, H7e′), 1.78 (dt, J ) 12.9, 13.0, 6.8 Hz, 1H,
H6a′), 1.78 (s, 3H, CH3, thymine), 1.17 (dd, J ) 12.9, 4.6 Hz, 1H,
H6e′). 13C NMR (600 MHz, DMSO-d6) δ: 164.9 (C4), 151.0 (C2),
137.0 (C6), 108.5 (C5), 86.3 (C4′), 86.2 (C1′), 64.5 (C3′), 62.3 (C5′),
62.0 (C2′), 38.9 (C7′), 27.2 (C6′), 13.3 (CH3, thymine).
(1R,5R,7R,8S)-5-Benzyloxymethyl-8-benzyloxy-7-(thymin-1-yl)-
2-aza-6-oxabicyclo[3.2.1]octane (12a and 12b). To a solution of 11
(5.2 g, 8.5 mmol) in 120 mL of dry THF, NaBH4 (965 mg, 25.5 mmol)
was added. To this suspension, trifluoroacetic acid (1.3 mL, 17 mmol)
was added dropwise over a period of 30 min under a nitrogen
atmosphere and stirred overnight at room temperature. After complete
conversion, excess NaBH4 was hydrolyzed carefully with water, stirred
at room temperature for 2 h, and extracted with dichloromethane. Note
that if the reaction is worked up after 30 min after hydrolyzing, NaBH4
gives a substantial amount of minor isomer, indicating that this isomer
is the kinetic product that converts to the major isomer in the presence
of NaOH formed during hydrolyzing NaBH4. The organic phase was
dried over anhydrous MgSO4 and evaporated under reduced pressure.
Purification by silica gel column chromatography (0-6% methanol in
dichloromethane, v/v) afforded 12a (1.6 g, 3.4 mmol, 40%) along with
the other diasteriomer 12b as a minor product (190 mg, 0.4 mmol,
5%). (Major diastereomer 12a). Rf ) 0.42 [90:10 CH2Cl2/CH3OH (v/
v)]. MALDI-TOF m/z: [M + H]+ found, 464.1; calcd, 463.2. 1H NMR
(500 MHz, CDCl3) δ: 7.97 (q, J) 1.3 Hz, H6, thymine), 7.37-7.24
(m, 10H, benzyl), 5.94 (s, 1H, H1′), 4.68 (d, J ) 11.8 Hz, 1H, CH2Ph),
4.57 (d, J ) 11.5 Hz, 1H, CH2Ph), 4.53 (d, J ) 11.8 Hz, 1H, CH2Ph),
4.51 (d, J ) 11.5 Hz, 1H, CH2Ph), 3.98 (d, J ) 3.9 Hz, 1H, H3′), 3.71
(d, J ) 10.8 Hz, 1H, H5′), 3.58 (d, J ) 10.8 Hz, 1H, H5′′), 3.52 (d, J
) 3.9 Hz, 1H, H2′), 3.13 (ddd, J ) 13.3, 11.6, 4.9 Hz, 1H, H7a′), 3.02
(dd, J ) 13.3, 6.5 Hz, 1H, H7e′), 2.03 (ddd, J ) 13.1, 11.6, 6.7 Hz,
1H, H6a′), 1.43 (d, J ) 1.3 Hz, 1H, CH3, thymine), 1.31 (dd, J ) 13.1,
4.8 Hz, 1H, H6e′). 13C NMR (125.7 MHz, CDCl3) δ: 163.9 (C4), 150.0
(C2), 137.4, 137.2, 135.7 (C6), 128.5, 128.3, 128.1, 128.0, 127.8, 109.4
(C5), 87.0 (C1′), 84.4 (C4′), 73.4 (CH2Ph), 71.8 (CH2Ph), 71.7 (C3′),
70.3 (C5′), 59.1 (C2′), 38.4 (C7′), 27.5 (C6′), 11.7 (CH3, thymine).
(1R,5R,7R,8S)-5-Benzyloxymethyl-8-benzyloxy-2-phenoxyacetyl-
7-(thymin-1-yl)-2-aza-6-oxabicyclo[3.2.1]octane (14). To a solution
of 12a (1.6 g, 3.4 mmol) in pyridine was added phenoxyacetyl chloride
(0.6 mL, 4.4 mmol), which was added dropwise and stirred under a
nitrogen atmosphere for 2 h. The reaction was quenched with cold
saturated aqueous NaHCO3 and extracted with dichloromethane. The
organic phase was dried over anhydrous MgSO4, evaporated under
reduced pressure, and co-evaporated with toluene (3 times) and
dichloromethane (3 times). The crude product was purified by silica
gel column chromatography (0-3% methanol in dichloromethane, v/v)
to afford 14 (note that 12b reacted very slowly because it converted to
12a first and then to 14 in 24 h) (1.4 g, 2.4 mmol, 70%). Rf ) 0.35
[96:4 CH2Cl2/CH3OH (v/v)]. MALDI-TOF m/z: [M + H]+ found,
598.2; calcd, 597.2. 13C NMR (67.9 MHz, CDCl3) δ: 168.6, 166.8,
163.6, 163.5, 157.9, 157.6, 150.0, 149.5, 137.0, 136.9, 135.1, 135.0,
129.5, 129.4, 128.5, 128.3, 128.2, 128.1, 128.0, 127.9, 127.7, 127.4,
121.6, 121.4, 114.6, 114.5, 110.0, 109.8, 87.1, 85.9, 84.7, 84.6, 73.5,
73.4, 72.5, 72.4, 72.5, 72.4, 71.0, 69.6, 69.4, 67.7, 59.5, 55.8, 53.3,
39.1, 36.6, 27.0, 25.6, 11.7.
(1R,5R,7R,8S)-8-Hydroxy-5-hydroxymethyl-2-phenoxyacetyl-7-
(thymin-1-yl)-2-aza-6-oxabicyclo[3.2.1]octane (15). Nucleoside 14
(1.4 g, 2.4 mmol) was dissolved in 15 mL of methanol, and 20%
Pd(OH)2 on charcoal (430 mg) was added, followed by ammonium
formate (1.81 g, 28.8 mmol), and refluxed for 12 h. The catalyst was
filtered off through a celite bed, and the filtrate evaporated under
reduced pressure and co-evaporated with dichloromethane to remove
traces of methanol. The crude material was dissolved in 25 mL of
anhydrous dichloromethane and cooled to -78 °C, and 1 M BCl3 (19
mL) was added and stirred under a nitrogen atmosphere for 3 h. Solvent
and volatile materials were removed under reduced pressure and co-
evaporated with methanol. The residue was purified by silica gel column
chromatography (0-4% methanol in dichloromethane, v/v) to give 15
(750 mg, 1.8 mmol, 75%). Rf ) 0.38 [90:10 CH2Cl2/CH3OH (v/v)].
MALDI-TOF m/z: [M + H]+ found, 418.2; calcd, 417.1. 13C NMR
(67.9 MHz, CD3OD) δ: 170.7, 166.9, 159.8, 152.4, 137.9, 137.7, 130.7,
130.6, 122.7, 122.5, 116.2, 116.0, 110.6, 87.8, 87.5, 87.3, 67.9, 66.9,
65.7, 65.4, 63.1, 62.9, 60.4, 37.8, 27.2, 26.5, 12.9.
(Minor diastereomer 12b). Rf ) 0.71 [92:8 CH2Cl2/CH3OH (v/v)].
1
MALDI-TOF m/z: [M + H]+ found, 464.0; calcd, 463.2. H NMR
(600 MHz, CDCl3) δ: 7.95 (s, 1H, NH, thymine), 7.77 (q, J ) 1.2 Hz,
1H, H6, thymine), 7.44-7.24 (m, 10 H, Bn), 6.28 (s, 1H, H1′), 4.64
(d, J ) 11.6 Hz, 1H, CH2Ph), 4.58 (d, J ) 11.5 Hz, 1H, CH2Ph), 4.56
(d, J ) 11.6 Hz, 1H, CH2Ph), 4.53 (dd, J ) 11.92, 3.91 Hz, 1H, NH),
4.50 (d, J ) 11.5 Hz, 1H, CH2Ph), 4.33 (d, J ) 4.1 Hz, 1H, H3′), 3.73
(d, J ) 11.0 Hz, 1H, H5′), 3.58 (d, J ) 4.1 Hz, 1H, H2′), 3.52 (d, J )
11.0 Hz, 1H, H5′′), 3.26 (ddd, J ) 14.2, 6.4, 3.9 Hz, 1H, H7e′), 2.98
(dtd, J ) 14.2, 11.9, 11.9, 5.2 Hz, 1H, H7a′), 2.04 (ddd, J ) 13.4,
11.9, 6.4 Hz, 1H, H6a′), 1.53 (dd, J ) 13.4, 5.2 Hz, 1H, H6e′), 1.50 (d,
J ) 1.2 Hz, CH3, thymine). 13C NMR (125.7 MHz, CDCl3) δ: 163.5
(C4), 149.1 (C2), 136.8, 136.0, 135.6 (C6), 128.6, 128.3, 128.2, 128.0,
127.9, 110.3 (C5), 82.6 (C1′), 82.5 (C4′), 73.68 (CH2Ph), 73.64
(1R,5R,7R,8S)-5-(4,4′-Dimethoxytrityloxymethyl)-8-hydroxy-2-
phenoxyacetyl-7-(thymin-1-yl)-2-aza-6-oxabicyclo[3.2.1]octane (16).
9
15186 J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006

Conformationally Constrained Aza-ENA-T
A R T I C L E S
Nucleoside 15 (750 mg, 1.8 mmol) was co-evaporated with anhydrous
pyridine to remove traces of water and dissolved in 15 mL of the same
solvent, and 4,4′-dimethoxytrityl chloride was added and stirred at room
temperature for 7 h. The reaction was quenched using cold saturated
aqueous NaHCO3 and extracted with dichloromethane. The organic
phase was dried over anhydrous MgSO4, evaporated under reduced
pressure, and co-evaporated with toluene (2 times) to remove pyridine
partially. The crude product was purified by silica gel column
chromatography [0-3% methanol in dichloromethane (v/v), containing
1% pyridine] to afford 16 (1.16 g, 1.6 mmol, 90%). Rf ) 0.30 [96:4
CH2Cl2/CH3OH (v/v)]. MALDI-TOF m/z: [M + H]+ found, 720.1;
calcd, 719.2. 13C NMR (67.9 MHz, CDCl3 plus DABCO) δ: 168.4,
163.6, 158.6, 157.1, 150.0, 144.1, 135.2, 135.0, 134.6, 129.9, 129.7,
129.5, 128.9, 128.0, 127.7, 127.0, 121.9, 121.6, 114.6, 114.4, 113.2,
110.5, 86.6, 86.0, 67.4, 65.3, 62.2, 62.2, 55.1, 46.7, 36.4, 25.6, 11.9.
(1R,5R,7R,8S)-8-(2-(Cyanoethoxy(diisopropylamino)-phosphinoxy)-
5-(4,4′-dimethoxytrityloxymethyl)-2-phenoxyacetyl-7-(thymin-1-yl)-
2-aza-6-oxabicyclo[3.2.1]octane (17). Compound 16 (1.16 g, 1.6
mmol) was dissolved in 15 mL of dry THF, and diisopropylethylamine
(1.4 mL, 8 mmol) was added at 0 °C followed by 2-cyanoethyl-N,N-
diisopropylphosphoramidochloridite (0.71 mL, 3.2 mmol). After 30 min,
the reaction was warmed to room temperature and stirred overnight.
MeOH (0.5 mL) was added, and stirring was continued for 5 min;
thereafter, saturated aqueous NaHCO3 was added and extracted with
freshly distilled CH2Cl2 (3 times). The organic phase was dried over
MgSO4, filtered, and concentrated under reduced pressure. The crude
residue was purified by silica gel column chromatography (40-70%
CH2Cl2 in cyclohexane containing 1% Et3N), which afforded 17 (1.26
g, 1.37 mmol, 86%) as a mixture of four isomers. Rf ) 0.40 [96:4
CH2Cl2/CH3OH (v/v)]. MALDI-TOF m/z: [M + H]+ found, 920.2;
calcd, 919.3. 31P NMR (67.9 MHz, CDCl3) δ: 150.7, 150.3, 149.2,
148.1.
partially. The crude product was purified by silica gel column
chromatography [0-3% methanol in dichloromethane (v/v), containing
1% pyridine] to afford 19 (827 mg, 1.2 mmol, 81%). Rf ) 0.31 [96:4
CH2Cl2/CH3OH (v/v)]. MALDI-TOF m/z: [M + Na]+ found, 704.2;
calcd, 681.2. 13C NMR (67.9 MHz, CDCl3 plus DABCO) δ: 164.02,
163.5, 158.6, 150.0, 149.7, 144.0, 143.9, 134.5, 129.9, 127.9, 127.2,
127.0, 113.3, 110.6, 86.7, 85.4, 85.3, 65.8, 65.4, 63.0, 62.9, 60.1, 55.1,
46.0, 40.1, 37.3, 29.5, 27.0, 26.1, 11.9, 11.8.
(1R,5R,7R,8S)-8-(2-(Cyanoethoxy(diisopropylamino)-phosphinoxy)-
5-(4,4′-dimethoxy-trityloxymethyl)-2-trifluoroacetyl-7-(thymin-1-yl)-
2-aza-6-oxabicyclo[3.2.1]octane (20). Compound 19 (827 mg, 1.2
mmol) was dissolved in 12 mL of dry THF, and diisopropylethylamine
(1.05 mL, 6 mmol) was added at 0 °C followed by 2-cyanoethyl-N,N-
diisopropylphosphoramidochloridite (0.53 mL, 2.4 mmol). After 30 min,
the reaction was warmed to room temperature and stirred overnight.
MeOH (0.5 mL) was added, and stirring was continued for 5 min;
thereafter, saturated aqueous NaHCO3 was added and extracted with
freshly distilled CH2Cl2 (3 times). The organic phase was dried over
anhydrous MgSO4, filtered, and concentrated under reduced pressure.
The crude residue was purified by silica gel column chromatography
(40-100% CH2Cl2 in cyclohexane containing 1% Et3N), which afforded
20 (645 mg, 0.73 mmol, 61%) as a mixture of four isomers. Rf ) 0.44
[96:4 CH2Cl2/CH3OH (v/v)]. MALDI-TOF m/z: [M + H]+ found,
882.2; calcd, 881.3. 31P NMR (109.4 MHz, CDCl3) δ: 150.1, 149.9,
149.8, 149.2.
Acknowledgment. Generous financial support from the
Swedish Natural Science Research Council (Vetenskapsrådet),
the Swedish Foundation for Strategic Research (Stiftelsen fo¨r
Strategisk Forskning), and the EU-FP6-funded RIGHT project
(project number LSHB-CT-2004-005276) is gratefully acknowl-
edged.
(1R,5R,7R,8S)-8-Hydroxy-5-hydroxymethyl-2-trifluoroacetyl-7-
(thymin-1-yl)-2-aza-6-oxabicyclo[3.2.1]octane (18). The nucleoside
12a (1.6 g, 3.4 mmol) was deprotected as it was done for 13, and the
crude material was dissolved in methanol (20 mL). DMAP (415 mg,
3.4mmol) and ethyl trifluoroacetate (4 mL, 34 mmol) were added and
stirred at room temperature overnight. The solvent was removed under
reduced pressure and purified by silica gel column chromatography
(0-6% methanol in dichloromethane, v/v) to give 18 (580 mg, 1.5
mmol, 45%). Rf ) 0.43 [90:10 CH2Cl2/CH3OH (v/v)]. MALDI-TOF
m/z: [M + H]+ found, 380.1; calcd, 379.0. 13C NMR (67.9 MHz,
CD3OD) δ: 166.8, 152.1, 138.4, 137.7, 110.8, 110.7, 87.3, 87.2, 86.8,
65.5, 65.2, 62.6, 62.5, 61.8, 41.5, 39.9, 39.0, 27.4, 26.7, 12.9.
(1R,5R,7R,8S)-5-(4,4′-Dimethoxytrityloxymethyl)-8-hydroxy-2-
trifluoroacetyl-7-(thymin-1-yl)-2-aza-6-oxabicyclo[3.2.1]octane (19).
Nucleoside 18 (580 mg, 1.5 mmol) was co-evaporated with anhydrous
pyridine to remove traces of water and dissolved in 15 mL of the same
solvent, and 4,4′-dimethoxytrityl chloride was added and stirred at room
temperature overnight. The reaction was quenched using cold saturated
aqueous NaHCO3 and extracted with dichloromethane. The organic
phase was dried over anhydrous MgSO4, evaporated under reduced
pressure, and co-evaporated with toluene (2 times) to remove pyridine
Supporting Information Available: 13C NMR spectra of
compounds 4, 6-11, 14-16, 18, and 19; 31P NMR spectra of
compounds 17 and 20; 13C, 1D NOE, HSQC, HMBC NMR
spectra of compound D (Scheme 1); 13C, HSQC, HMBC, and
DQF-COSY NMR spectra of compounds 12a, 12b, and 13;
RNase H digestion profile of AONs at 0.08 unit enzyme
concentration; denaturing PAGE picture in human serum, tables
of 1H and 13C chemical shifts, and coupling constants of
compounds 12a, 12b, and 13; introduction, discussions on
assignment, stereochemistry,3JHH simulation, and NOE studies
of 12a, 12b, and 13; analysis of CD of AON/RNA and AON/
DNA duplexes; cleavage rates of the AON/RNA duplexes by
RNase H; experimental methods and details of theoretical
calculations; generalized Karplus parametrization; and details
of pKa determination of 2′-amino-LNA-T and aza-ENA-T. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA0634977
9
J. AM. CHEM. SOC. VOL. 128, NO. 47, 2006 15187