Nonionic analogs of RNA with dimethylene sulfone bridges

Article abstract of DOI:10.1021/ja952322m

Analogs of RNA have been synthesized where each of the phosphodiester linking groups is replaced by dimethylene sulfone units (sulfone-linked nucleic acid analogs of RNA, or 'rSNAs'). These are the first fully nonionic analogs of RNA to be prepared as oligomers. Sequences leading to the octamer 5'-r(A(SO₂)U(SO₂)G(SO₂)G(SO₂)U(SO₂)C(SO₂)A(SO₂)U)-3' have been prepared from 3'5'-bishomo-β-ribonucleoside derivatives as building blocks prepared from diacetone D-glucose, and their chemistry has been explored. Coupling was performed in solution via S(N)2 reactions between a thiol from one fragment and a bromide from the other, oxidation of the resulting thioether to the sulfone, and deprotection of a terminal primary hydroxyl group and regioselective conversion of it - in the presence of secondary hydroxyl groups - to an active group (thiol or bromide) to yield another fragment for coupling. Base-labile protecting groups were used for the nucleobases, and one-step full deprotection was achieved using 1 M NaOH. The target octamer and each isolated intermediate were characterized by NMR, UV spectroscopy, and mass spectrometry. While chemical reactions involving longer rSNAs were in several cases retarded relative to analogous reactions with monomers, some rates were enhanced. In water, the rSNA octamer displayed a thermal transition in the UV spectrum above 65°C with a large hyperchronicity. The behaviors of rSNAs suggest roles for the polyanionic backbone in DNA and RNA beyond its role in conferring aqueous solubility. The repeating anionic charges in natural oligonucleotides evidently also control the potent molecular recognition properties of these richly functionalized molecules, direct strand-strand interactions to the part of the biopolymer distant from the backbone (the Watson-Crick edge of the nucleobases), cause the polymer to favor an extended conformation, and ensure that the physical properties of the oligonucleotide are largely independent of its sequence. This suggests structural features that must be built into nonionic oligonucleotide analogs generally.

Full text of DOI:10.1021/ja952322m

4518
J. Am. Chem. Soc. 1996, 118, 4518-4531
Nonionic Analogs of RNA with Dimethylene Sulfone Bridges
Clemens Richert,†,‡ Andrew L. Roughton,‡ and Steven A. Benner*,‡,§
Contribution from the Department of Chemistry, Swiss Federal Institute of Technology,
UniVersita¨tstrasse 16, CH-8092 Zu¨rich, Switzerland, and Department of Chemistry, UniVersity of
Florida, GainesVille, Florida 32611
ReceiVed July 14, 1995X
Abstract: Analogs of RNA have been synthesized where each of the phosphodiester linking groups is replaced by
dimethylene sulfone units (sulfone-linked nucleic acid analogs of RNA, or “rSNAs”). These are the first fully nonionic
analogs of RNA to be prepared as oligomers. Sequences leading to the octamer 5′-r(ASO USO GSO GSO USO CSO -
2
2
2
2
2
2
ASO U)-3′ have been prepared from 3′,5′-bishomo-â-ribonucleoside derivatives as building blocks prepared from
2
diacetone D-glucose, and their chemistry has been explored. Coupling was performed in solution via SN2 reactions
between a thiol from one fragment and a bromide from the other, oxidation of the resulting thioether to the sulfone,
and deprotection of a terminal primary hydroxyl group and regioselective conversion of itsin the presence of secondary
hydroxyl groupssto an active group (thiol or bromide) to yield another fragment for coupling. Base-labile protecting
groups were used for the nucleobases, and one-step full deprotection was achieved using 1 M NaOH. The target
octamer and each isolated intermediate were characterized by NMR, UV spectroscopy, and mass spectrometry. While
chemical reactions involving longer rSNAs were in several cases retarded relative to analogous reactions with
monomers, some rates were enhanced. In water, the rSNA octamer displayed a thermal transition in the UV spectrum
above 65 °C with a large hyperchromicity. The behaviors of rSNAs suggest roles for the polyanionic backbone in
DNA and RNA beyond its role in conferring aqueous solubility. The repeating anionic charges in natural
oligonucleotides evidently also control the potent molecular recognition properties of these richly functionalized
molecules, direct strand-strand interactions to the part of the biopolymer distant from the backbone (the Watson-
Crick edge of the nucleobases), cause the polymer to favor an extended conformation, and ensure that the physical
properties of the oligonucleotide are largely independent of its sequence. This suggests structural features that must
be built into nonionic oligonucleotide analogs generally.
Introduction
of natural oligonucleotides might also be replaced by a
dimethylene sulfone unit to yield a hydrolytically stable
oligonucleotide analog lacking the diastereoisomerism associated
with the linkage in several other analogs, referred to here as
sulfone-linked nucleic acid analogs (SNAs).10,11 Because the
-SO2- and -PO2-- units are largely isosteric and isoelec-
tronic, the structural perturbation introduced by the sulfone-
for-phosphate substitution should be relatively small. Further,
DNA and RNA are composed of a backbone consisting of
repeating phosphate diester units that join sugars bearing a
nonrepeating nucleobase. In the standard model of duplex
structure proposed by Watson and Crick four decades ago,1 the
nucleobases play the central role in molecular recognition, while
the sugar and phosphate linkers play secondary roles as
scaffolding for the nucleobases, perhaps with some preorgani-
zation. Accordingly, it is widely believed that the scaffolding
can be modified to improve the physical properties of oligo-
nucleotides while retaining their molecular recognition proper-
ties, provided that the modification continues to meet preorga-
nizational requirements. This model has underlain efforts by
many groups, academic and industrial, to seek oligonucleotide
analogs with improved biological properties by modification
of the backbone.2
One class of backbone-modified oligonucleotide analogs
replaces the anionic phosphate groups with neutral linkers.
Among the most prominent examples of these are the meth-
ylphosphonate,3 sulfonate,4 sulfonamide,5 amide,6 and formacetal
and thioformacetal7 analogs of DNA,8 as well as the more
distantly related peptide-linked nucleic acid analogs (PNAs).9
Some time ago, we noted that the bridging phosphate groups
(3) (a) Miller, P. S.; McFarland, K. B.; Jayaraman, K.; Ts’o, P. O. P.
Biochemistry 1981, 20, 1878-1880. (b) Miller, P. S.; Cushman, C. D.; Levis,
J. T. In Oligonucleotides and Analogues. A Practical Approach; Eckstein,
F., Ed.; IRL Press: Oxford, 1991; pp 137-154. (c) Wozniak, L. A.;
Pyzowski, J.; Wieczorek, M.; Stec, W. J. J. Org. Chem. 1994, 59, 5843-
5846. (d) Lesnikowski, Z. J.; Jaworska, M.; Stec, W. J. Nucleic Acids Res.
1990, 18, 2112-2114.
(4) (a) Huang, J.; McElroy, E. B.; Widlanski, T. S. J. Org. Chem. 1994,
59, 3520-3521. (b) Musicki, B.; Widlanksi, T. S. Tetrahedron Lett. 1991,
32, 1267-1270.
(5) (a) McElroy, E. B.; Bandaru, R.; Huang, J. X.; Widlanski, T. S.
Bioorg. Med. Chem. Lett. 1994, 4, 1071-1076. (b) Perrin, K. A.; Huang,
J.; McElroy, E. B.; Iams, K. P.; Widlanski, T. S. J. Am. Chem. Soc. 1994,
116, 7427-7428.
(6) (a) De Mesmaeker, A.; Waldner, A.; Fritsch, V.; Lebreton, J.; Wolf,
R. M. Bull. Soc. Chim. Belg. 1994, 103, 705-717. (b) Demesmaeker, A.;
Lebreton, J.; Waldner, A.; Fritsch, V.; Wolf, R. M.; Freier, S. Synlett 1993,
733-736. (c) Chur, A.; Holst, B.; Dahl, O.; Valentinhansen, P.; Pedersen,
E. B. Nucleic Acids Res. 1993, 21, 5179-5183.
(7) Jones, R. J.; Lin, K. Y.; Milligan, J. F.; Wadwani, S.; Matteucci, M.
D. J. Org. Chem. 1993, 58, 2983-2991.
(8) For uncharged dimeric sulfide RNA analogs see: (a) Meng, B.;
Kawai, S. H.; Wang, D.; Just, G.; Giannaris, P. A.; Damha, M. J. Angew.
Chem. 1993, 105, 733-735; Angew. Chem., Int. Ed. Engl. 1993, 32, 729-
731. (b) Kawai, S. H.; Just, G. Nucleosides Nucleotides 1991, 10, 1485-
1489.
(9) (a) Egholm, M.; Burchardt, O.; Nielsen, P. E.; Berg R. H. J. Am.
Chem. Soc. 1992, 114, 1895-1897. (b) Nielsen, P. E.; Egholm, M.;
Burchardt, O. Bioconjugate Chem. 1994, 5, 3-7 and references therein.
(10) (a) Huang, Z.; Schneider, K. C.; Benner, S. A. J. Org. Chem. 1991,
56, 3869-3882 (b) Huang, Z.; Benner, S. A. Synlett 1993, 83-84.
* To whom correspondence should be addressed at the Swiss Federal
Institute of Technology.
† Present address: Department of Chemistry, Tufts University, Medford,
MA 02155.
‡ ETH Zu¨rich.
§ University of Florida.
X Abstract published in AdVance ACS Abstracts, March 15, 1996.
(1) (a) Watson, J. D.; Crick, F. H. Nature 1953, 171, 737-738. (b)
Watson, J. D.; Crick, F. H. Nature 1953, 171, 964-967.
(2) Hunziker, J.; Leumann, C. In Modern Synthetic Methods 1995; Ernst,
B., Leumann, C., Eds.; Verlag Helv. Chimica Acta: Basel, 1995; pp 331-
418.
S0002-7863(95)02322-5 CCC: $12.00 © 1996 American Chemical Society

Nonionic Analogs of RNA with Dimethylene Sulfone Bridges
J. Am. Chem. Soc., Vol. 118, No. 19, 1996 4519
Table 1. 13C NMR Chemical Shifts of Nucleobase Carbon Atoms
for Selected Purine 3′,5′-Bishomonucleosides (CDCl3)
Scheme 1
chemical shift (ppm)
compd
C2
C4
C5
C6
C8
11 (N9)
7b (N9)
10b (N9)
6 (N9)
147.9/148.1a
151.2/152.2a
147.3/147.4a
152.8
120.3
118.6
122.2
123.8
123.1
108.2
110.9
114.2
155.7
160.7
155.5
151.3
150.8
163.8
157.5
148.5
137.5
140.7
137.8
141.8
141.5
144.0
140.6
142.6
149.7
149.2
13b (N9)b
7a (N7)
10a (N7)
13a (N7)b
152.1
152.8/155.7a
147.2/152.6a
156.9
141.8
a C2 and C4 resonances of guanosine residues were not unequivocally
assigned. b The N7 and N9 isomers 13a and 13b, formed by hydrolysis
of 9a and 9b, were resolved by HPLC chromatography.
(6) in 54% yield as the sole product.17 The yield was 60%
when 2′,3′-diacetate 818 was used as the glycosyl donor to
furnish 9b.
because the SdO bond of sulfones retains part of the dipolar
character of the phosphorus-oxygen bond in the phosphate
monoanion, SNAs might retain some of the aqueous solubility
of natural DNA, but be soluble in organic solvents as well, and
retain some of the interstrand recognition properties of natural
oligonucleotides.
This is the case with a dimeric SNA comprised of ribose-
type bishomo sugars (where rSNA denotes the exclusive
inclusion of 2′-hydroxyl or 2′-O-acyl groups), as shown by X-ray
crystallography.12 In crystals grown at room temperature, the
Analogous Vorbru¨ggen-type glycosylation with N2-isobu-
tyrylguanine gave ca. 1:1 mixtures of the N7 (10a) and N9 (10b)
regioisomers. Protection of O6 as the 2-(p-nitrophenyl)ethyl
ether,19 and the use of 1.2 equiv of TMSTf in CH3CN at 50 °C
for 38 h, gave a 1:1.8 ratio of N7 (7a) to N9 (7b) isomers. The
isomers were separated by chromatography and isolated in good
combined yield (82%). The N7 isomer was then rearranged in
80% yield to the N9 isomer using MSTFA and 5 mol % TMSTf
at 50 °C over 23 h. Recently, glycosylation involving presi-
lylation with MSTFA and TMSCl, followed by removal of
excess reagents under vacuum, and the use of 0.5 equiv of
TMSTf at 60 °C for 95 h gave a more favorable 1:7 ratio of N7
to N9 isomers in 86% chemical yield.20 Thus, all four
bishomonucleosides (4, 5, 7b, 9b) are accessible as pure â,N1
or N9 isomers in at least 60% yield. An X-ray analysis
established the structure of the partially deprotected guanosine
analog 11 (from 10b; see Scheme 2), and underlay a correlation
between 13C NMR chemical shifts and the position of attachment
of the purine bases for these molecules.21 Table 1 gives 13C
resonances for 11, its protected precursors, and selected ad-
enosine analogs.
self-complementary r(GSO C) sequence exhibited a duplex
2
structure with Watson-Crick base pairing characteristics similar
to those found in RNA. The r(ASO U) dimer, however,
2
crystallized at high temperature and was found to be single
stranded.13 We report here the synthesis of rSNAs up to eight
bases in length (the ribose-type “r” descriptor is hereafter
assumed for all specific species in this context, e.g., GSO C).
2
These were prepared in mixed base sequences including all four
natural nucleobases and characterized.
Synthesis
Monomers. The nucleoside analog building blocks were
constructed from diacetone D-glucose. Homoallose derivative
110a (Scheme 1) was obtained in nine steps and 20% overall
yield following optimization14 of a route developed in these
laboratories.10 Silylation of the 6-position of 1 yielded 2.
Benzoylation of the remaining 2-hydroxyl group gave glycosyl
donor 3.
The glycosyl donor was reacted with (protected) nucleobases
under modified Vorbru¨ggen conditions15 to yield bishomo-
nucleosides 4-716 (Scheme 1). Pyrimidines were efficiently
introduced with a stoichiometric excess of Lewis acid in
CH3CN at 40-60 °C. Under similar conditions, N6-benzoy-
ladenine gave mixtures of N7 and N9 isomers. Therefore,
reaction conditions were optimized with respect to the catalyst,
solvent, temperature, and reagent stoichiometry to allow isomer-
ization during the reaction of the N7 derivative to the favored
N9 isomer. Incubation for two days at 80 °C with 0.5 equiv of
TMSTf in dichloroethane yielded the N9 derivative of adenosine
The octameric heterosequence 6′-A
U
G
G
U
C
-
SO2 SO2 SO2 SO2 SO2 SO2
ASO U-3′′ was then prepared to provide a test of the oligomer-
2
ization methodology. This octamer was prepared convergently
in solution by coupling four dimers (ASO U, GSO G, USO C,
2
2
2
and ASO U) to yield two tetramers (ASO USO GSO G and
2
2
2
2
USO CSO ASO U), and coupling of the tetramers to yield the
2
2
2
octamer (see Schemes 3-5).
To prepare monomers functionalized at the 3′′-end as
protected thiols, the 2′- and 3′′-benzoyl groups in the adenosine,
uridine, and guanosine analogs were removed with 0.1 M LiOH
in methanol/THF/water. Compounds 4, 9, and 7b were
converted to diols 12, 13b, and 14 in 80-93% yield (Scheme
2). Reaction times were optimized to minimize a slow
N6-debenzoylation of adenine. Conversion of diols 12, 13b,
and 14 to the corresponding 3′′-thioesters 15-17 by a Mitsunobu
(11) Roughton, A. L.; Portmann, S.; Benner, S. A.; Egli, M. J. Am. Chem.
Soc. 1995, 117, 7249-7250.
(16) Hyrup, B.; Richert, C.; Benner S. A. Unpublished results.
(17) Obtained from 2 by hydrolysis with aqueous NaOH and subsequent
acetylation with acetic anhydride in 70% yield. Huang, Z. Dissertation No.
10429, ETH Zu¨rich, 1993.
(18) (a) Himmelsbach, F.; Schulz, B. S.; Trichtinger, T.; Charubala, T.;
Pfleiderer W. Tetrahedron 1984, 40, 59. (b) Jenny, T. F.; Benner, S. A.
Tetrahedron Lett. 1992, 33, 6619.
(19) Ko¨nig, M.; Benner, S. A. Unpublished results.
(20) Reference 10a reports a correlation based on NMR data of natural
nucleosides.
(12) Hyrup, B.; Richert, C.; Schulte-Herbrueggen, T.; Benner, S. A.; Egli,
M. Nucleic Acids Res. 1995, 23, 2427-2433.
(13) Richert, C. Dissertation No. 10895, ETH Zu¨rich, 1994.
(14) (a) Vorbru¨ggen, H.; Bennua, B. Chem. Ber. 1981, 114, 1279-1286.
(b) Vorbru¨ggen, H.; Krolikiewicz, K.; Bennua, B. Chem. Ber. 1981, 114,
1234-1255.
(15) Methylene groups attached to the 3′ and the 5′ carbon atoms of
natural nucleosides were designated as the 3′′ and 6′ positions in 3,5-
bishomonucleosides; see Scheme 1.

4520 J. Am. Chem. Soc., Vol. 118, No. 19, 1996
Richert et al.
Scheme 2
Scheme 3
reaction22 employing thioacetic acid in THF or acetonitrile
proceeded quickly at room temperature, and was fully selective
for the primary hydroxyl group.
lower and side products were observed when mesylating
monomers and oligomers containing uracil.
Oligomers. Liberation of the thiol from 15-17 and genera-
tion of the thiolate for the SN2 coupling depended on the nature
of the nucleobase and its protecting group. For example, the
acetate was removed smoothly from 17 by ammonolysis in
methanol at 0 °C, and the resulting 24 used in Cs2CO3-
mediated27 coupling either as a crude substance or as a
chromatographically isolated substance (Scheme 3). With 15,
however, substantial amounts of (presumed) disulfide were
obtained as an undesired byproduct. The benzoyl group in 16
was relatively labile under these conditions. Thioacetates 15
and 16 were therefore deprotected in situ, by migration of the
acetyl group from sulfur to the neighboring 2′-hydroxyl group
(yielding 25 and 26) in the coupling reaction. This in situ
deprotection reaction simultaneously protected the 2′-hydroxyl
group, and yielded coupled products without the formation of
Monomers carrying an electrophilic 6′-carbon were prepared
by desilylation of 4, 5, and 7b with TBAF in THF (Scheme 2).
Partial removal of the O6-[2-(p-nitrophenyl)ethyl] protection
from the guanine in 7b was always observed, however. Hence,
O6 of 7b was first liberated by treatment with DBU in pyridine19a
(see Scheme 1), and the resulting 10b desilylated to give 20.
Treatment of the alcohols 18-20 with CBr4/PPh323,24 gave the
bromides 21-23. Generally, CH3CN was used as the solvent.
The low solubility of 19 necessitated the use of a mixture of
CH3CN and dichloroethane and slightly elevated temperatures.
Bromination was regioselective for primary hydroxyl groups.
Intramolecular displacement of the bromides by O2 in pyrim-
idines, a reaction seen in natural nucleosides bearing a leaving
group at the 5′-position,25 was never observed with these
molecules. While a mesylate was an alternative to the bromide
as a leaving group,8,26 it was not used here because yields were
(25) (a) Andersen, W.; Hayes, D. H.; Michelson A. M.; Todd, A. R. J.
Chem. Soc. 1954, 1882-1887. (b) Verheyden, J. P. H., Moffat, J. G. J.
Org. Chem. 1970, 35, 2868-2877.
(22) Mitsunobu, O. Synthesis 1981, 1.
(23) Ramirez, F.; Desai, N. B.; McKelvie, N. J. Am. Chem. Soc. 1962,
84, 1745.
(26) Kawai, S. H.; Wang, D.; Just, G. Can. J. Chem. 1992, 70, 1573-
1580.
(27) (a) Buter, J.; Kellogg, R. M. J. Chem. Soc., Chem. Commun. 1980,
466. (b) Buter, J.; Kellogg R. M. J. Org. Chem. 1981, 46, 4481.
(24) Verheyden, J. P. H.; Moffat, J. G. J. Org. Chem. 1972, 37, 2289.

Nonionic Analogs of RNA with Dimethylene Sulfone Bridges
J. Am. Chem. Soc., Vol. 118, No. 19, 1996 4521
Scheme 4
detectable amounts of disulfide. The in situ acetyl migration
and coupling of guanine-containing compound 17 were less
efficient.
quantitative yield. Sulfoxide intermediates, formed almost
immediately, could be recovered by early termination of a
reaction. These were characterized by proton NMR (“doubling”
of signals resulting from the diastereomeric mixture of sulfox-
ides) and mass spectrometry. The oxidation was followed by
thin layer chromatography for the dimers, as the sulfoxides
exhibited higher polarity than the sulfones, to which they were
slowly but completely transformed over 1-2 h. No other
products were detectable, even with longer oligomers. To
improve solubility of the educts during the reaction, THF was
added as required. Using CHCl3 as an additive, a homogeneous
reaction mixture could be obtained, although oxidations under
these conditions were considerably slower.
Migration of the acetyl groups to generate 3′′-thiol(ate)s 25
and 26 was ca. 10-fold faster than the actual coupling with 21
and 22 (both 30 mM) to give USC (27, 54%) and ASU (28,
57%) dimers. Approximately 20% of the dimers isolated from
these reactions lacked the 2′-O-acetyl group [products 29 (23%)
and 30 (23%)], presumably because of an intermolecular
migration of the acetyl group from the 2′-O-position. Since
this acetyl group was removed two steps later (see below), no
effort was made to reduce the formation of this useful byproduct.
The combined yields of isolated dimers were 77% (USC) and
80% (ASU) with Cs2CO3 in a homogeneous CH3CN/H2O
mixture over 16-24 h at 50 °C. Yields calculated on the basis
of recovered bromide starting material (e10%) were cor-
respondingly higher. The coupling of guanine-containing
compounds 23 and 24 in a suspension of Cs2CO3 in THF over
2.5 h at 35 °C gave GSG dimer 31, isolated in 92% yield, as
well as traces of bromide starting material and (presumed)
disulfide. DBU as the base in the coupling reaction10,28 led to
considerable amounts of symmetrical disulfide and accordingly
lower yields with thiols generated by hydrolytic deacetylation
of 15 and 16.
The synthetic cycle of selective deprotection, functionaliza-
tion, coupling, and oxidation was repeated starting from dimeric
sulfones 32-34. Deprotection of the 2′- and 3′′-alcohol groups
with LiOH was lower yielding with USO C (32 to 35, 54%)
2
than with ASO U (33 to 36, 78%). Loss of the protection on
2
the base was shown to account for the lower yields. Some
chemical shifts of the proton signals of the parent dimer 32
exhibited a concentration dependence that may point to molec-
ular associates, which may in turn influence the rate with which
the N-benzoate or O-benzoate groups are cleaved. Further,
signals in the 1H NMR spectrum of 35 suggested that the USO2C
dimer, notably in contrast to the ASO U dimer 36, is involved
Thioethers were oxidized to sulfones after each coupling step,
a precaution taken because full oxidation of a DNA-type
thioether octamer had required forced conditions and was
frequently low yielding.28 Further, sulfones were shown on the
basis of UV hyperchromicity studies (see below) to aggregate
less than their parent thioethers. Thus, oxidation proved to be
an important element in the strategy to manage the solubility
properties of synthetic intermediates.
2
in intermolecular association in solution (see the Discussion).
Mitsunobu thioacetylation (35 to 37 and 36 to 38) proceeded
in 82% and 92% yield for the USO C and ASO U dimers.
2
2
Desilylation of the ASO U and GSO G dimers 33 and 34 used
2
2
TBAF/THF and HF/pyridine, respectively. The latter was a
mild alternative to TBAF compatible with the NPE protecting
group. The 6′-alcohols were converted to bromides with CBr4/
PPh3 in CH3CN (39) and in 1,2-dichloroethane (40) to give 41
and 42 in high yields.
Treatment of dimer thioethers 27, 28, and 31 with Oxone in
methanol/water29 gave sulfones 32-34 (Scheme 4) in almost
(28) Huang, Z. Dissertation No. 10429, ETH Zu¨rich, 1993.
(29) Trost, B. M.; Curran D. P. Tetrahedron Lett. 1981, 22, 1287.

4522 J. Am. Chem. Soc., Vol. 118, No. 19, 1996
Richert et al.
Scheme 5

Nonionic Analogs of RNA with Dimethylene Sulfone Bridges
J. Am. Chem. Soc., Vol. 118, No. 19, 1996 4523
Coupling of the two sets of dimers to yield tetramers 43 and
44 proceeded in ca. 70% yield with Cs2CO3 in THF (Scheme
5). Again, ca. 25% of the products were isolated as tetramers
lacking the acetyl groups on the nucleoside unit bearing the
3′′-thioether, but no symmetrical disulfides were isolated. More
unreacted dimer was isolated after these couplings compared
to that recovered in the synthesis of the dimers from monomer
building blocks. This suggested that the coupling reaction was
slower with dimers. The thioethers were oxidized to yield 45
and 46.
and the former to protected sulfone 55 (Scheme 5). The sulfonic
acid and residual bromide were removed easily by HPLC,
yielding 55 in 53% yield for the two steps (70% based on
recovered bromide 54).
Conditions for deprotecting rSNAs varied according to the
protecting groups. USO CSO ASO U tetramer 49 was depro-
2
2
2
tected in CH3OH/THF treated with saturated aqueous NH3,
smoothly yielding fully deprotected tetramer USO CSO ASO2U
2
2
56 (structural formula not shown) at room temperature in 20 h.
With tetramer 45, which contains each type of protecting group
present in 55, the TBDPS and the NPE groups were removed
only slowly (reaction times g24 h) under the conditions
normally used for deprotecting synthetic DNA and RNA
oligomers (NH4OH, e60°C). They were quantitatively re-
moved, however, together with all acyl groups, by treatment
with 1 M NaOH in a mixture containing CH3OH and THF. The
reaction was complete within a few hours at 40 °C, or after
one day at room temperature. No side products were detected,
demonstrating the high stability of the sulfones to strongly basic
conditions. Further, all of the rSNAs studied here were soluble
in aqueous media at high pH, which facilitated their handling.
Precipitation/extraction yielded an oligomer pure by NMR,
MALDI-TOF, and reversed phase HPLC. With this protocol,
45 and 46 were converted to the ASO USO GSO G tetramer
For the third synthetic cycle, the 2′- and 3′′-hydroxyl groups
of tetramer 45 were deprotected to yield 47 in 67% yield, with
ca. 10% of the A-N6-debenzoylated product 48. Interestingly,
when dissolved in a CH2Cl2/CH3OH mixture (4 µM), 47 was
found after one week at room temperature to have spontaneously
transformed itself into 48 (see below). Desilylation of
USO CSO ASO U tetramer 46 with HF/pyridine furnished 49
2
2
2
in 77% yield.
Activation of the tetramers was most successful when the
intermolecular association of the synthetic intermediates was
managed. For example, synthesis of 50 from 47 was best in a
mixed solvent system (dioxane/1,2-dichloroethane) at slightly
elevated temperature, providing product in 60% yield (72%
based on recovered starting material). Synthesis of a bromide
from 49 was problematic under the conditions used for the
dimers. Thus, 46 was 2′-O-acetylated to give 51, desilylated
(TBAF, THF) to 52, and converted to bromide 53 in CH3CN/
dichloroethane in 52% yield. Alternatively, 49 was mixed with
monomer alcohol 18, which improved its solubility, and both
alcohols were brominated together in CH3CN/dichloroethane
to give 54 in 33% yield (83% based on recovered starting
material). Tetramer 45 could also be desilylated (HF, pyridine)
and converted to the 6′-bromide (CBr4/PPh3/pyridine) in 36%
yield over two steps.
To understand better the lower reactivity of the longer rSNA
oligomers, partially protected oligomers 47 and 49 were
examined by UV spectroscopy in mixtures of CH2Cl2 and
CH3OH (see the supporting information). At micromolar
concentrations, strong hyperchromicity at 260 nm was observed
upon addition of CH3OH. The amount of CH3OH necessary
for maximum hyperchromicity was higher for tetramers contain-
ing primary alcohols than for the protected parent compounds,
and lower for sulfones than for their parent thioethers. Proton
NMR signals of the tetramers were broad in CDCl3 (at
millimolar concentrations) and sharpened with CD3OD. The
addition of D2O further sharpened the signals. These observa-
tions are consistent with intermolecular aggregation via hydro-
gen bonding.
2
2
2
57 (63%) and the USO CSO ASO U tetramer 56 (60%),
2
2
2
respectively (structural formulas not shown). Under the same
conditions, octamer 55 was converted to 58 (Scheme 5), which
after precipitation and a series of extraction and washing steps
was shown to be pure by spectroscopic techniques.
Characterization
1
Oligomeric rSNAs were characterized by H and 13C NMR
and UV spectroscopy and mass spectrometry, and occasionally
by combustion analysis and X-ray structure determination. Mass
spectrometric analysis is invaluable to identify longer depro-
tected oligonucleotide analogs.32 Unsatisfactory signal to noise
ratios were obtained for longer sulfone oligonucleotides in
MALDI-TOF spectra recorded under conditions recommended
for natural oligonucleotides.33 The matrix 2-[(4-hydroxyphe-
nyl)azo]benzoic acid,34 often used for proteins, yielded sufficient
sensitivity and good resolution for sulfone oligomers with data
accumulation in the positive mode with repeated low power
laser shots (see the supporting information). About 10 pmol
of oligomer per matrix preparation was used. Monosodium
adducts produced the most prominent MALDI MS peaks, as
was shown independently by electrospray ionization mass
spectrometry of octamer 58 (see the supporting information).
The conformations of rSNAs were explored by proton NMR;
1H and 13C assignments are reported in the supporting informa-
tion. The ring pucker of the bishomo ribose rings, estimated
from the J1′,2′ couplings using an approximation [2′-endo (%)
) J1′,2′/(J1′,2′ + J3′,4′) × 100],35 was predominantly 3′-endo
(“north”) for most protected and all unprotected rSNAs, similar
to that observed in natural RNA. A north conformation of the
ribose ring in solution might be expected on the basis of the
conformations observed in the single crystal X-ray structure
Coupling of the tetramers 50 and 54 (Scheme 5) was slower
than with shorter fragments, yielding for the first time sym-
metrical disulfide as a side product of the intramolecular acetyl
transfer method. Similar influence of the length of the synthetic
fragment on the success of a coupling is not uniformly observed
in the synthesis of RNA itself,30 but may have been encountered
in the solution phase synthesis of other nonionic oligo-
nucleotides.4a Low yields are reported in esterification reactions
with largely unprotected carbamate-linked thymidine trimers and
tetramers.31
(32) (a) Buck, H. M.; Koole, L. H.; van Genderen, M. H. P.; Smit, L.;
Geelen, L. M. C.; Jurriaans, S.; Goudsmit, J. Science 1990, 248, 208-212.
(b) Moody, H. M.; Quaedflieg, P. J. L. M.; Koole, L. H.; van Genderen,
M. H. P.; Buck, H. M.; Smit, L.; Jurriaans, S.; Geelen, J. L. M. C.; Goudsmit,
J. Science 1990, 250, 125-126.
(33) Pieles, U.; Zu¨rcher, W.; Scha¨r, M.; Moser, H. E. Nucleic Acids Res.
1993, 21, 3191-3196.
(34) Juhasz, P.; Costello, C. E.; Biemann, K. J. Am. Soc. Mass Spectrom.
1993, 4, 399.
Directly treating the crude product mixture containing desired
octamer sulfide and the octamer disulfide side product with
Oxone converted the latter to the corresponding sulfonic acid
(30) Van Boom, J. H.; Wreesman, C. T. J. In Oligonucleotide Synthesis,
A Practical Approach; Gait, M. J., Ed.; IRL Press: Oxford, 1984; pp 153-
183.
(31) Habus, I.; Temsamani, J.; Agrawal, S. Bioorg. Med. Chem. Lett.
1994, 4, 1065-1070.
(35) (a) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1973, 95,
2333-2344. (b) Davies, D. B. Prog. NMR Spectrosc. 1978, 12, 135-225.

4524 J. Am. Chem. Soc., Vol. 118, No. 19, 1996
Richert et al.
analyses of the guanosine analog 11 (see the supporting
acetyl transfer should be slower. Thus, lower yields in
migration/coupling reactions observed with GNPE,iBu-3′′-SAc/
2′-OH monomer 17, where NMR spectra suggest that more of
the south conformation is present, can be rationalized. Similarly,
the thioacetylation of 3′′-alcohol 14, also showing a significant
south contribution, gave almost exclusively the 3′′-SAc/2′-OH
product 17. However, the GiBu diol 11, with more north
conformation, gave a 1.9:1 ratio of 3′′-SAc/2′-OH to 3′′-SH/
2′-OAc under the same conditions.
information) and the GSO C and ASO U dimers.12,13
2
2
Remarkably, (p-nitrophenyl)ethyl-protected guanosine analogs
having a free 2′-hydroxyl group had significantly increased
3
values for J1′,2′ and reciprocal changes in the multiplicities of
H3′ and H4′ signals, suggesting an increased contribution of
2′-endo-type (“south”) states. The GiBuSGiBu dimer, where G1
(residues take their letter plus numerical position from the 5′-
to 3′-end) has a free 2′-hydroxyl group but no longer bears a
NPE group at O6, also displayed an increased contribution of
south states in G1 (J1′,2′ ) 5.3 Hz). These results indicate that
the conformational equilibrium of the tetraydrofuran ring is
sensitive to small changes in structure elsewhere in the molecule.
While this may be the result of a “through bond” effect, it is
more likely due to an intermolecular association of rSNAs
containing G.
In pure water, fully deprotected sulfone oligomers show
moderate solubility. The solubility of octamer 58 was found
to be g5 µM in water and >1 mM in DMSO and DMF; higher
solubilities were often observed, but not reproducibly. Dis-
solution in organic solvents and water is slow, and requires
gentle heating. The rSNAs without guanine dissolve consider-
ably faster than rSNAs containing G. Those rSNAs containing
guanine could, however, be efficiently solubilized with 0.1 M
NaOH.
Octamer 58 displayed a transition in UV absorbance (258
nm) as a function of temperature. Remarkably, this began at
ca. 65 °C, and was not complete at 95 °C. The first derivative
of the curve, extracted from plots obtained at three different
concentrations (ca. 1, 0.5, and 0.33 µM), yield in each case a
“melting point” of 83 °C.36 The hyperchromicity was large
compared to that observed in a model RNA with the same
sequence. To rule out the possibility that the hyperchromicity
reflects simple, reversible solution/dissolution of the octamer,
the melting experiment was repeated at pH 12.5, where 58
dissolves well in water. The hyperchromicity was lower at this
higher pH, but the transition did not disappear. The transition
did nearly completely disappear, however, when 20% DMF was
added to the neutral aqueous solution.
Detailed analysis of the NMR spectra of several intermediates
suggested that intermolecular association could also be an
important factor determining reactivity. For example, in contrast
to other monomeric and dimeric rSNAs bearing two or more
free hydroxyl groups, all signals of the more easily N-
Bz
debenzoylated U C
SO2
triol 35 were sharp in neat CDCl3,
consistent with the molecule forming a single defined structure.
Signals for N3-H of uracil and both 2′-OH protons of the dimer
are ca. 1-1.5 ppm downfield of the corresponding signals in
other compounds, suggesting that one of the base NHs and at
least two of the ribose OHs are involved in hydrogen bonding.
Further, each of the base hydrogens gives at least one NOESY
cross peak to a proton of the other pyrimidine, indicating base
stacking. Strong interresidue base-to-ribose NOE cross peaks
(e.g., NH(U) to H-2′(C)) indicate that 35 forms at least a
bimolecular associate. Formation of this associate is sensitive
to small structural changes. For example, functionalization of
the 3′′-hydroxyl group as a thioester (35 to 37) seems to block
association.
The yields of coupling reactions correlated with increasing
chain length. This effect could be explained by an increase in
intermolecular association through collection of a critical mass
of hydrogen bonding groups in the intermediates. Evidence for
this comes from the solvent effects on yield in the coupling.
Yields of monomer couplings to produce dimers (e.g., 15 and
22 to give 27) were higher in THF and lower in DMF; the
opposite was observed in the tetramer to octamer coupling (50
and 54 to give the precursor of 55), where disrupting the
hydrogen bonding with the polar solvent may have been
essential.
From a synthetic point of view, these effects are simply
technological annoyances. As such, they were managed by
adjusting the polarity of the solvents, manipulating the sulfide-
sulfone oxidation states of intermediates, and changing purifica-
tion strategies. With optimized conditions, coupling, oxidation,
and either 6′- or 3′′-deprotection without chromatography
proceed in approximately 70% overall yield for the three steps
at the dimer and trimer levels.37
Discussion
Our synthetic strategy makes available rSNAs having any
sequence, and therefore allows examination of the physical and
biological properties of these nonionic oligonucleotide analogs
to begin. Preliminary results show that rSNAs have many
characteristics that deserve detailed exploration. These char-
acteristics offer insights into why natural oligoribonucleotides
have phosphate bridging groups, and how nonionic oligonucle-
otides might be designed to exhibit altered physical properties
while retaining Watson-Crick recognition.
Physical and Catalytic Properties of rSNAs. Some inter-
esting reactivities were observed at the outset in the partially
protected rSNAs used as synthetic intermediates. For example,
the rates and yields of the coupling reactions were sensitive to
the sequences being coupled. Since a key step in the coupling
is the rate of migration of the acetate from the 3′′-thiol to the
2′-hydroxyl group (e.g., 15 to 25, Scheme 3), this reaction should
be influenced by the conformation of the sugar ring. When
the sugar is in a 3′-endo-type conformation, the transfer can
proceed via a six-membered ring transition state in a chair-type
conformation that is free from significant steric clashes. In
contrast, if the sugar ring adopts the 2′-endo-type conformation,
At a scientific level, however, these effects are interesting
statements about the conformation and reactivity of nonionic
oligonucleotide analogs. For example, not all effects observed
in rSNAs with higher molecular weight involved slower
chemical reactions. In some cases, longer oligomers had
increased rates of reactions. For example, as a solution (4 µM)
in CH2Cl2/CH3OH, 47 was observed to transform itself spon-
taneously into 48 after one week at room temperature. In more
detailed studies at 20 µM concentration, 47 showed a half-life
for the debenzoylation of ca. 30 h at 25 °C. The analogous
reaction was also observed in dimer 36 (t1/2 ≈ 12 h), but was
much slower in the 2′,3′′-deprotected adenosine monomer 13b
(t1/2 g 10 days).38 Comparison of the debenzoylation rates of
the quickest self-debenzoylating SNA 36 with that of N6-
(37) Baeschlin, D.; Richert, C.; Benner, S. A. Unpublished results.
(38) The structure of products was proven by NMR and mass spectrom-
etry; controls suggested that the reaction did not occur through intervention
of adventitious silica, but rather arose from “catalysis” by the rSNA itself.
(36) Melting temperatures were determined as described in the follow-
ing: Egholm, M.; Nielsen, P.; Buchardt, O.; Berg, R. H. J. Am. Chem.
Soc. 1992, 114, 9677-9678.

Nonionic Analogs of RNA with Dimethylene Sulfone Bridges
J. Am. Chem. Soc., Vol. 118, No. 19, 1996 4525
benzoyl-2′-deoxyadenosine as a reference showed a rate en-
hancement of g30-fold.
of the backbone has long been believed to account for the
difficulty with which oligonucleotides cross the lipid phase of
membranes to enter cells.
Because 36 and 47 are not recovered unchanged from the
reaction mixture, it is incorrect to state that they catalyze their
own debenzoylations. However, if the reaction occurs Via an
aggregate, then the rSNA molecules might be acting as a true
catalyst in the debenzoylation reaction. The structural origin
of the “catalytic” activity of these rSNAs is unknown. The
phenomenon deserves further exploration, especially as it is
reminiscent of the self-transformations catalyzed by natural
RNA.39
This model has underlain many efforts to generate backbone-
modified oligonucleotides with altered solubility properties. Best
known are efforts to develop nuclease-resistant and membrane-
permeable DNA analogs for “antisense” applications41 by
replacing the repeating phosphate polyanion.42-44 In a very
early study, nonionic but randomly polymeric oligonucleotide
analogs were prepared.45
The rich molecular association properties extended to the fully
deprotected rSNAs. First, rSNAs can participate in Watson-
The behaviors of rSNAs documented here underscore the
incompleteness of the standard model in describing the role of
the backbone. Much of the current disenchantment46 with
oligonucleotide analogs as therapeutic agents may be due to
the failure of the standard model in this respect. These results
also suggest at least three mechanisms by which the phosphate
groups define the molecular recognition properties of oligo-
nucleotides in ways incidental to their effect on solubility in
water.
Crick pairing. For example, the GSO C dimer forms in the
2
crystal a duplex quite similar in overall structure to the analogous
RNA (GPC) dimer.12 Non-Watson-Crick pairing presumably
is involved in the molecular aggregation of the GSO G dimer,
2
however, as it is noticeably less soluble in water than the GSO2C
dimer (data not shown). Self-association in G-rich RNA and
DNA is well known.40
However, it is clear that non-Watson-Crick interactions play
a far more important role in rSNAs than in RNAs. The
molecular association of the octameric rSNA 58 in water proved
to be especially interesting. This sequence is nearly self-
complementary. If Watson-Crick rules are followed, two of
these molecules might form an antiparallel duplex joined by
six Watson-Crick base pairs and two G-U wobble base pairs.
An RNA molecule with the identical sequence indeed displayed
a melting transition consistent with this structure. With this
natural RNA strand at 4 and 1.4 µM concentrations, the melting
transitions were at ca. 16 and 12 °C, respectively.
In contrast, the “melting temperature” of octameric rSNA 58
displayed a transition at much higher temperature. The “melting
temperature” was difficult to obtain precisely, as the end point
was >95 °C. The transition temperature did not change notably
with a 3-fold change in concentration of the octamer, suggesting
that the transition reflects a unimolecular change in the
conformation rather than a disassociation of two octamer strands.
Further, the relatively weak dependence of the transition on pH
suggests that the structure being melted is formed by interaction
between hydrogen bonding groups that are not easily ionizable
(the 2′-OH group of the backbone sugars and the exocyclic NH2
groups of the bases, for example). One conformation consistent
with these data is a tight “hairpin”. These suggestions must be
qualified, however, in light of the extremely high transition
temperature, making it impossible to establish a high-temper-
ature baseline, to determine accurately the midpoint of the
transition, and thereby to assess accurately the stoichiometry
of the transition. The self-association of octameric 58 appears
to preclude intermolecular association of 58 with complementary
RNA and DNA. The melting transition of 58 was the same in
the presence of both complementary RNA and DNA as in their
absence.
First, the phosphate groups evidently force the preferred
interaction surface between oligonucleotide strands to the parts
of the molecule that are as distant from the polyanionic backbone
as possible. This is, of course, the Watson-Crick “edge” of
the nucleobases. In the absence of interstrand phosphate-
phosphate repulsion, sugar-sugar interstrand interactions, sugar-
backbone interstrand interactions, interactions between the sugar
and backbone groups of one strand and the Hoogsteen edge of
the nucleobases on the other, Hoogsteen-Hoogsteen interstrand
interactions, and Watson-Crick-Hoogsteen interstrand interac-
tions all become more important. In rSNAs, the absence of
the repeating negative charges in the backbone allows the
interactions between the two strands to “move in” from the
Watson-Crick edge of the nucleobases to involve other parts
of the strand.
Second, the repeating polyanionic backbone in natural DNA
and RNA makes the properties of the oligonucleotide largely
independent of its sequence. Interactions between monopoles
(charges) dominate dipolar interactions, which in turn dominate
quadrapolar interactions (and so on). The presence of a
repeating monopole in oligonucleotides implies that dipolar
interactions (hydrogen bonding, for example) are secondary in
defining the properties of an oligonucleotide. This means that
the behavior of a natural oligonucleotide will, to a first
approximation, be independent of its sequence. This property
is important for an encoding molecule. With rSNAs, virtually
every behavior is strongly dependent on sequence and length,
including solubility, conformation, and reactivity, especially
when compared with natural oligonucleotides. In this respect,
Comparison with Phosphate-Linked Oligonucleotides.
The “standard model” of nucleic acid structure, proposed by
Watson and Crick in 1953,1 assigns to the nucleobases the
predominant role for the strand-strand recognition displayed
by oligonucleotides. Accordingly, the backbone is often viewed
as being largely irrelevant to their molecular recognition
properties, even as it is recognized as being a key to the
solubility properties of oligonucleotides in aqueous and non-
aqueous environments. In particular, the polyanionic character
(41) (a) Zamecnik, P. C.; Stephenson, M. L. Proc. Natl. Acad. Sci. U.S.A.
1978, 75, 280-284. (b) Stephenson, M. L.; Zamecnik, P. C. Proc. Natl.
Acad. Sci. U.S.A. 1978, 75, 285-288.
(42) (a) Uhlmann, E.; Peyman, A. Chem. ReV. 1990, 90, 544-584. (b)
Milligan, J. F.; Matteucci, M. D.; Martin, J. C. J. Med. Chem. 1993, 36,
1923-1937.
(43) Hanvey, J. C.; Peffer, N. J.; Bisi, J. E.; Thomson, S. A.; Cadilla,
R.; Josey, J. A.; Ricca D. J.; Hassmann, C. F.; Bonham, M. A.; Au K. G.;
Carter S. G.; Bruckenstein, D. A.; Boyd, A. L.; Noble, S. A.; Babiss, L. E.
Science 1992, 258, 1481-1485.
(44) Wagner, R. W.; Matteucci, M. D.; Lewis, J. G.; Gutierrez, A. J.;
Moulds, C.; Froehler, B. C. Science 1993, 260, 1510-1513.
(45) Pitha, J.; Pitha, M.; Ts’o, P. O. P. Biochem. Biophys. Acta 1970,
204, 39-48.
(39) (a) Cech, T. R.; Bass, B. L. Annu. ReV. Biochem. 1986, 55, 599-
629. (b) Pyle, A. M. Science 1993, 261, 703-714. (c) Symons, R. H. Curr.
Opin. Struct. Biol. 1994, 4, 322-330.
(40) Lipsett, M. N. J. Biol. Chem. 1964, 239, 1250-1255
(46) Gura, T. Science 1995, 270, 575-577.

4526 J. Am. Chem. Soc., Vol. 118, No. 19, 1996
Richert et al.
rSNAs behave more like peptides than oligonucleotides,47 which
also self-associate, form conformations strongly dependent on
sequence, and display low levels of catalytic activity.48,49 By
amplifying interactions that are only second-order in RNA,
sRNAs should reveal second-order factors influencing behavior
in natural oligonucleotides, factors that may have biological
significance.50
Third, the statistical mechanical theory of polymers suggests
that the polyanionic backbone will cause natural oligonucleotides
to adopt an extended structure.51 This deters the formation of
hairpins and other folded unimolecular structures that bring the
phosphate groups together. An extended structure can, of
course, be viewed as “preorganization” to prepare a single strand
for binding to a complementary strand. Conversely, it can be
viewed as an effect disfavoring unimolecular conformations in
an oligonucleotide that might compete with its binding to
complementary oligonucleotides. Thus, the polyanionic RNA
sequence APUPGPGPUPCPAPU forms duplexes. The self-
structure of the corresponding rSNA 58 precludes association
with complementary oligonucleotides.
As nonionic oligonucleotide backbones can no longer direct
the interaction between strands toward the extremities of the
nucleobases (the “Watson-Crick edges”), or force the molecule
to adopt an extended structure, nonionic oligonucleotide analogs
must incorporate other features if they are to consistently achieve
binding to complementary oligonucleotides following simple
rules. For example, functionality not directly involved in
Watson-Crick pairing (nitrogens 3 and 7 of the purines, the
furanose ring oxygen, the 2′-hydroxyl group) might be removed.
Further, because the properties of nonionic oligonucleotides
depend strongly on their sequence, the sequence of a nonionic
analog to be used in a biological system (for example, as an
antisense agent) cannot be selected simply to make it comple-
mentary to a target sequence, but must also be mindful of the
physicochemical properties of the antisense sequence itself.
(FAB, 3-nitrobenzyl alcohol (NOBA) matrix), or Bruker Reflex
(MALDI-TOF, 2-[(4-hydroxyphenyl)azo]benzoic acid (HABA) matrix)
spectrometer. Electrospray ionization spectra: VG Biotech BIO Q
spectrometer, acquired by MScan Corp., Geneva. Optical rotations:
Perkin-Elmer 241 or Jasco DIP-370 polarimeter. Microanalyses:
Microanalytical Laboratory, ETH, Zu¨rich. The following experimental
part gives representative procedures and data only; additional procedures
and assigned (for oligonucleotides, residues take their letter, or letter
plus numerical position from the 5′- to 3′-end, e.g., A or G2)
spectroscopic data for all compounds are reported in the supporting
information.
1-[2′-Benzoyl-3′-[(benzoyloxy)methyl]-6′-(tert-butyldiphenylsilyl)-
3′,5′-dideoxy-â-allofuranosyl]uracil (4). Uracil (1.3 g, 11 mmol) was
dried under high vacuum at 60 °C and suspended in CH3CN (20 mL).
N-Methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA) (8.8 g, 8.1 mL,
44 mmol) was added under stirring, resulting in a clear solution after
15 min. Glycosyl donor 3 (3.5 g, 5.5 mmol), which had been
coevaporated with toluene and dried under high vacuum at 50 °C, was
dissolved in CH3CN (40 mL) and added to the silylated base. The
resulting solution was cooled (0 °C), and trimethylsilyl triflate (TMSTf)
(1.8 g, 1.5 mL, 8.3 mmol) was injected. The solution was warmed to
40 °C over 30 min and then stirred for 6 h before being cooled to 0 °C
and poured into a mixture of saturated Na2CO3 (200 mL), ice (100 g),
and CH2Cl2 (300 mL). The mixture was filtered through Celite, the
organic layer separated, and the aqueous layer reextracted 2× with
CH2Cl2. The combined organic phases were washed with brine, which
was reextracted 2× with CH2Cl2. The organic layer was dried (MgSO4),
filtered, and concentrated in Vacuo, and the residue chromatographed
on silica (120 g) with CH2Cl2/EtOAc (4:1) and a MeOH gradient (1.5%
to 6%). Removal of residual MSTFA under high vacuum yielded
bishomouridine 4 (3.4 g, 4.7 mmol, 85%) as a colorless glass. 1H NMR
(300 MHz, CDCl3): δ 1.07 (s, 9H, CH3); 1.98, 2.16 (2m, 2H, H-5′);
2.95 (m, 1H, H-3′); 3.93 (m, 2H, H-6′); 4.39 (m, 1H, H-4′); 4.54 (m,
2H, H-3′′); 5.74 (d, 1H, J ) 8.1 Hz, H-5); 5.81 (m, 2H, H-1′, H-2′);
7.24 (d, 1H, J ) 8.1 Hz, H-6); 7.40 (m, 10H, m-Bz, m,p-Ph); 7.55 (2
AA′BB′C systems, 2 C parts, 2H, p-Bz); 7.68 (m, 4H, o-Ph); 7.99 (2
AA′BB′C systems, 2 AA′ parts, 4H, o-Bz); 9.38 (br s, 1H, NH).
FABMS: m/z 741 (M + Na+). Anal. Calcd for C41H42N2O8Si: C,
68.50; H, 5.89; N, 3.90. Found: C, 68.57; H, 5.74; N, 3.99.
N 4-Benzoyl-1-[2′-benzoyl-3′-[(benzoyloxy)methyl]-6′-(tert-butyl-
diphenylsilyl)-3′,5′-dideoxy-â-allofuranosyl]cytosine (5). Following
the procedure for preparing 4, 2 equiv of TMSTf was stirred with N4-
benzoylcytosine52 (1.23 g, 5.73 mmol) and glycosyl donor 3 (2.92 g,
4.58 mmol) (1 h, 60 °C). The residue was chromatographed using
EtOAc/petroleum ether/chloroform/EtOH (100:100:60:1) to yield 5
(2.29 g, 2.80 mmol, 61%). 1H NMR (300 MHz, CDCl3): δ 1.09 (s,
9H, CH3); 2.00, 2.20 (2m, 2H, H-5′); 2.79 (m, 1H, H-3′); 3.97 (m, 2H,
H-6′); 4.19 (d, 2H, J ) 6.9 Hz, H-3′′); 4.54 (dt, 1H, J ) 10, 2.7 Hz,
H-4′); 5.93 (d, 1H, J ) 1.4 Hz, H-1′); 6.06 (d, 1H, J ) 4.5 Hz, H-2′);
7.32-7.69 (m, 16H, H-5, m,p-Bz, m,p-Ph); 7.75 (m, 4H, o-Ph); 7.77
(d, 1H, J ) 6.7 Hz, H-6); 7.98 (3 AA′BB′C systems, 3 AA′ parts, 6H,
o-Bz); 9.4 (br s, 1H, NH). FABMS: m/z 822 (M + H+). Anal. Calcd
for C48H47N3O8Si: C, 70.14; H, 5.76; N, 5.11. Found: C, 69.70; H,
5.83; N, 5.23.
N2-Isobutyryl-O6-[2-(4-nitrophenyl)ethyl]-9-[2′-benzoyl-3′-[(ben-
zoyloxy)methyl]-6′-(tert-butyldiphenylsilyl)-3′,5′-dideoxy-â-allofura-
nosyl]guanine (7b). N2-Isobutyryl-O6-[2-(4-nitrophenyl)ethyl]guanine19
(3.54 g, 9.55 mmol, dried under HV, 2 h) was suspended in CH3CN
(60 mL) at room temperature (RT). MSTFA (3.53 mL, 19.09 mmol)
was added to the thick white paste and stirring continued at RT for ca.
2 h, giving a faintly yellow solution containing a small amount of fine,
suspended white material. A solution of 3 (4.10 g, 6.36 mmol) in CH3-
CN (30 mL) was added at RT, followed by TMSTf (1.39 mL, 7.64
mmol). The mixture was heated to 50 °C for 38 h and then cooled to
0 °C and quenched with saturated aqueous NaHCO3 (30 mL) under
vigorous stirring. The suspension was filtered through cotton/sand,
and the pellet was washed with CH2Cl2 (150 mL). The filtrate was
shaken with NaHCO3 (50 mL), the aqueous phase was extracted with
CH2Cl2 (2 × 30 mL), and the combined organic phases were washed
Experimental Section
General Techniques. Reactions were carried out under an argon
atmosphere with freshly distilled (THF and dioxane from Na) or
commercially dried (DMF, pyridine, CH3CN, and Cl(CH2)2Cl; puriss
quality, absolute, stored over molecular sieves) solvents (Fluka) under
anhydrous conditions, unless otherwise noted. Reagents: highest
commercial quality (Fluka), used without further purification. Acetate
buffer: HOAc (3 M), NaOAc (1 M), and deionized water. Analytical
thin layer chromatography: E. Merck 60 F254 precoated silica gel plates,
compounds visualized by UV light or staining with cerium(IV) sulfate/
phosphomolybdic acid/H2SO4 (concentrated) and subsequent heating.
Flash chromatography: E. Merck silica gel (60, 0.040-0.063 mm
mesh), solvents distilled from anhydrous CaSO4 (Sikkon, Fluka),
stepwise (3-5 steps) eluant gradients unless otherwise noted. NMR
spectra: Bruker AMX 500, AMX 600, or Varian Gemini 300, 200
spectrometer. Ultraviolet (UV) spectra: Shimadzu UV/VIS 240 or
Perkin-Elmer Lambda 2 spectrophotometer, the latter being equipped
with a thermoelectric cuvette holder, thermoelement, and Perkin-Elmer
Digital Controller 570-071 for the collection of UV data as a function
of temperature. Infrared spectra: Perkin-Elmer 781 spectrophotometer.
Low-resolution mass spectra (MS): VG Tribrid (EI), VG ZAB2-SEQ
(47) Hirschmann, R.; Nutt, R. F.; Veber, D. F.; Vitali, R. A.; Varga, S.
L.; Jacob, T. A.; Holly, F. W.; Denkewalter, R. G. J. Am. Chem. Soc. 1969,
91, 507-508 and preceding papers.
(48) Johnsson, K.; Allemann, R. K.; Widmer, H.; Benner, S. A. Nature
1993, 365, 530-532.
(49) Barbier, B.; Brack, A. J. Am. Chem. Soc. 1992, 114, 3511-3515.
(50) (a) Gorin, A. A.; Zhurkin, V. B.; Olson, W. K. J. Mol. Biol. 1995,
247, 34-48. (b) Yang, Y; Westcott, T. P.; Pedersen, S. C.; Tobias, I.; Olson,
W. K. Trends Biochem. Sci. 1995, 20, 313-319. (c) Sindon, R. R. DNA
Structure and Function; Academic Press: New York, 1994.
(51) Flory, P. J. Principles of Polymer Chemistry; Cornell University
Press: Ithaca, 1953.
(52) Brown, D. M.; Todd, Sir A.; Varadarajan, S. J. Chem. Soc. 1956,
2384-2387.

Nonionic Analogs of RNA with Dimethylene Sulfone Bridges
J. Am. Chem. Soc., Vol. 118, No. 19, 1996 4527
with saturated aqueous NaCl (150 mL), dried (MgSO4), filtered, and
concentrated in Vacuo to a yellowish foam. Chromatography on silica
gel (325 g, petroleum ether/EtOAc (2:3)) followed by removal of
N-methyltrifluoroacetamide under high vacuum yielded regioisomers
7a (N7; 1.79 g, 29%) and 7b (N9; 3.28 g, 53%), isolated as white foams.
For the data of 7a, see the supporting information. 7b 1H NMR (400
MHz, CDCl3): δ 1.045 (s, 9H, C(CH3)3); 1.182 (d, J ) 5.7 Hz, 3H,
CH(CH3)2); 1.199 (d, J ) 5.7 Hz, 3H, CH(CH3)2); 1.952-2.042 (m,
1H, H-5′); 2.153-2.232 (m, 1H, H-5′); 2.92-3.05 (m, 1H, CH(CH3)2);
3.311 (t, J ) 6.7 Hz, 2H, OCH2CH2Ar); 3.62-3.73 (m, 1H, H-3′);
3.872 (d, J ) 5.4 Hz, 1H, H-6′); 3.894 (d, J ) 4.9 Hz, 1H, H-6′);
4.547-4.688 (m, 3H, H-4′, 2 × H-3′′); 4.811 (t, J ) 6.7 Hz, 2H,
OCH2CH2Ar); 5.963 (d, J ) 1.6 Hz, 1H, H-1′); 6.108 (dd, J ) 1.6, 6.0
Hz, 1H, H-2′); 7.296-7.452 (m, 10H, 4 × m-Bz, 2 × p-Ph, 4 × m-Ph);
7.488-7.522 (AA′BB′ system, AA′ part, 2H, 2 × o-PhNO2); 7.530-
7.661 (m, 6H, 2 × p-Bz, 4 × o-Ph); 7.861 (s, 1H, N2-H); 7.867 (s, 1H,
H-8); 7.988-8.022 (2 AA′BB′C systems, 2 AA′ parts, 4H, 4 × o-Bz);
8.145-8.179 (AA′BB′ system, BB′ part, 2H, 2 × m-PhNO2).
FABMS: m/z 978 (M + H+). Anal. Calcd for C54H56N6O10Si: C,
66.38; H, 5.78; N, 8.60. Found: C, 65.76; H, 5.86; N, 8.50.
N6-Benzoyl-9-[2′-acetyl-3′-[(acetyloxy)methyl]-6′-(tert-butyldiphe-
nylsilyl)-3′,5′-dideoxy-â-allofuranosyl]adenine (9b). N6-Benzoylad-
enine53 (770 mg, 3.2 mmol, recrystallized from EtOH, dried over P4O10)
was suspended in Cl(CH2)2Cl and silylated with MSTFA (1.5 g, 1.4
mL, 7.6 mmol) over 30 min at RT. Glycosyl donor 818 (1.0 g, 2.0
mmol) in Cl(CH2)2Cl (5 mL) was added to the clear solution, and
TMSTf (220 mg, 180 µL, 1 mmol) injected. The mixture was stirred
(80 °C, 50 h), cooled (0 °C), and diluted with CH2Cl2 (25 mL). Under
slow stirring, saturated Na2CO3 solution was added dropwise and the
resulting mixture filtered through sand. The organic layer was separated
and the aqueous layer reextracted 3× with CH2Cl2. The sand was
washed (CH2Cl2/MeOH (99:1), 300 mL), and the washings were
combined with the organic extracts. These were washed with brine,
dried (MgSO4), filtered, and dried in Vacuo. The residue was
chromatographed on silica (120 g, EtOAc/CH2Cl2/MeOH (60:40:1))
to yield 9b (870 mg, 1.2 mmol, 60%) as a colorless glass. 1H NMR
(300 MHz, CDCl3): δ 1.06 (s, 9H, CH3-tBu); 1.96, 2.03 (2m, 2H, H-5′);
2.06, 2.15 (2s, 2 × 3H, COCH3); 3.16 (m, 1H, H-3′); 3.84 (m, 2H,
H-6′); 4.18 (AA′X system ) q, 1H, J ) 5.8, 11.4 Hz, H-3′′); 4.33 (m,
2H, H-3′′, H-4′); 5.84 (dd, 1H, J ) 1.7, 6.2 Hz; H-2′); 5.96 (d, 1H, J
) 1.8 Hz, H-1′); 7.38 (m, 8H, m-Bz, m,p-Ph); 7.55 (AA′BB′C system,
C part, 1H, p-Bz); 7.64 (AA′BB′C system, AA′ part, 4H, o-Ph); 8.01
(AA′BB′C system, AA′ part ) d, 2H, J ) 7 Hz, o-Bz); 8.03 (s, 1H,
H-8); 8.75 (s, 1H, H-2); 9.31 (br s, 1H, NH). FABMS: m/z 722 (M
+ H+). Anal. Calcd for C39H43N5O7Si: C, 64.89; H, 6.00; N, 9.70.
Found: C, 64.46; H, 6.11; N, 9.68.
N2-Isobutyryl-9-[2′-benzoyl-3′-[(benzoyloxy)methyl]-6′-(tert-bu-
tyldiphenylsilyl)-3′,5′-dideoxy-â-allofuranosyl]guanine (10b). A so-
lution of N9 isomer 7b (2.0 g, 2.1 mmol) in absolute pyridine (21 mL)
was cooled (0 °C) and treated with 1,8-diazabicyclo[5.4.0]undec-7-
ene (DBU; 1.53 mL, 10.3 mmol). The ice bath was removed and the
solution stirred for 2.75 h before recooling the solution to 0 °C, adding
glacial HOAc (586 µL, 10.3 mmol), rotoevaporating the solvents to
near dryness, and coevaporating the residue with toluene (3 × 30 mL).
The oil was chromatographed (silica gel, 60 g, petroleum ether/EtOAc
(2:3)). This afforded a mixture of p-vinylnitrobenzene and 10b in a
ca. 4:1 ratio, as determined by 1H NMR (1.69 g, ca. 80% yield of 10b).
A pure analytical sample (data below) was obtained after a second
column under similar conditions. 1H NMR (400 MHz, CDCl3): δ 1.055
(s, 9H, C(CH3)3); 1.235 (d, J ) 6.9 Hz, 3H, CH(CH3)2); 1.241 (d, J )
6.9 Hz, 3H, CH(CH3)2); 1.845-1.946 (m, 1H, H-5′); 2.096-2.175 (m,
1H, H-5′); 2.604 (hept, J ) 6.9 Hz, 1H, CH(CH3)2); 3.476-3.549 (m,
1H, H-3′); 3.880 (d, J ) 5.1 Hz, 1H, H-6′); 3.898 (d, J ) 5.0 Hz, 1H,
H-6′); 4.480-4.592 (m, 3H, H-4′, 2 × H-3′′); 5.858 (d, J ) 1.2 Hz,
1H, H-1′); 6.117 (dd, J ) 1.1, 5.5 Hz, 1H, H-2′); 7.317-7.462 (m,
10H, 4 × m-Bz, 2 × p-Ph, 4 × m-Ph); 7.531-7.664 (m, 6H, 2 ×
p-Bz, 4 × o-Ph); 7.689 (s, 1H, H-8); 7.906-7.930 (AA′BB′C system,
AA′ part, 2H, 2 × o-Bz); 8.002-8.025 (AA′BB′C system, AA′ part,
2H, 2 × o-Bz); 8.700 (s, 1H, N2-H); 12.00 (br s, 1H, N1-H). FABMS:
m/z 829 (M + H+). Anal. Calcd for C46H49N5O8Si: C, 66.73; H, 5.96;
N, 8.46. Found: C, 66.46; H, 5.93; N, 8.31.
N6-Benzoyl-9-[6′-(tert-butyldiphenylsilyl)-2′,3′-dideoxy-3′-(hy-
droxymethyl)-â-allofuranosyl]adenine (13b). Diacetate 9b (3.6 g,
5 mmol) in a mixture of MeOH (60 mL) and THF (50 mL) was treated
with 0.2 N LiOH (62.5 mL, 12.5 mmol) with stirring at RT. The
hydrolysis was stopped after 12 min by addition of acetate buffer (6.25
mL). CH2Cl2 (150 mL) was added, the organic layer was separated,
and the aqueous layer was reextracted four times with CH2Cl2. The
combined organic layers were washed with brine, dried (MgSO4),
filtered, and dried in Vacuo. The residue was chromatographed on silica
(200 g, EtOAc/CH2Cl2/MeOH (6:4:1)) to yield 13b (2.55 g, 4 mmol,
80%) as a colorless foam. 1H NMR (300 MHz, CDCl3): δ 1.06 (s,
9H, CH3); 1.91 (m, 2H, H-5′); 2.34 (m, 1H, H-3′); 3.83 (m, 3H, H-6′,
H-3′′); 3.94 (m, 1H, H-3′′); 4.1 (br s, 1H, OH); 4.38 (dt, 1H, J ) 3.8,
8.2 Hz, H-4′); 4.90 (dd, 1H, J ) 2.4, 6.2 Hz, H-2′); 5.95 (d, 1H, J )
2.7 Hz, H-1′); 6.0 (br s, 1H, OH); 7.39 (m, 8H, m-Bz, m,p-Ph); 7.49
(AA′BB′C system, C part, 1H, m-Bz); 7.67 (AA′BB′C system, AA′
part, 4H, o-Ph); 7.95 (AA′BB′C system, AA′ part ) d, 2H, J ) 7.5
Hz, o-Bz); 8.02 (s, 1H, H-8); 8.52 (s, 1H, H-2); 9.47 (br s, 1H, NH).
FABMS: m/z 660 (M + Na+). Anal. Calcd for C35H39N5O5Si: C,
65.91; H, 6.16; N, 10.98. Found: C, 65.87; H, 6.49; N, 10.90. For
the related conversions of 4, 7b, and 10b, see the supporting
information.
1-[3′-[(Acetylthio)methyl]-6′-(tert-butyldiphenylsilyl)-3′,5′-dideoxy-
â-allofuranosyl]uracil (15). Diol 12 (3.4 g, 6.67 mmol) was coevapo-
rated from toluene, dried under high vacuum, and dissolved in THF
(60 mL). PPh3 (2.62 g, 10 mmol) in THF (60 mL) was cooled to 0 °C
and treated with diisopropyl azodicarboxylate (DIAD; 2.02 g, 1.95 mL,
10 mmol). Within 5 min, a white precipitate formed. After 30 min,
the solution of 12 and thioacetic acid (0.76 g, 0.71 mL, 10 mmol) were
added concurrently. The precipitate dissolved, and the reaction was
allowed to warm to RT. After overnight stirring, MeOH (0.5 mL) was
added and the solvents were removed by rotary evaporation. The
residue was chromatographed twice on silica (EtOAc/CH2Cl2, first run
3:1, second run 1:1) to yield thioester 15 (2.43 g, 4.3 mmol, 64%).
The acetyl group had migrated to the 2′-position in ca. 20% of this
product. A reaction with 70 mg of 12 gave a 75% yield of pure 2′-
OH product after one chromatographic separation. 1H NMR (300 MHz,
CDCl3): δ 1.06 (s, 9H, CH3-tBu); 1.82 (m, 2H, H-5′); 2.17 (m, 1H,
H-3′); 2.33 (s, 3H, SCOCH3); 3.05 (m, 2H, H-3′′); 3.93 (m, 2H, H-6′);
4.29 (m, 2H, H-4′, OH); 5.04 (d, 1H, J ) 3 Hz, H-2′); 5.67 (s, 1H,
H-1′); 5.69 (d, 1H, J ) 8.1 Hz, H-5); 7.38 (d, 1H, J ) 8.1 Hz, H-6);
7.43 (m, 6H, m,p-Ph); 7.70 (AA′BB′C system, AA′ part, 4H, o-Ph);
10.54 (br s, 1H, NH). FABMS: m/z 591 (M + Na+). Anal. Calcd
for C29H36N2O6SSi: C, 61.24; H, 6.38; N, 4.93. Found: C, 61.05; H,
6.42; N, 4.90. For the related conversions of 13 and 14, see the
supporting information.
N4-Benzoyl-1-[2′-benzoyl-3′-[(benzoyloxy)methyl]-3′,5′-dideoxy-
â-allofuranosyl]cytosine (19). Compound 5 (361 mg, 0.44 mmol) was
dissolved in THF (10 mL) under stirring and treated with tetrabuty-
lammonium fluoride (TBAF; 153 mg, 0.484 mmol) in THF (10 mL)
at RT. After 4 h, additional TBAF (25 mg, 0.08 mmol) was added. 30
min later, the solvent was removed by rotary evaporation at 30 °C and
the residue was directly applied to a silica column (20 g). The product
was eluted using CH2Cl2/EtOAc (1:4, 100 mL) followed by pure EtOAc
(50 mL) and CH2Cl2/MeOH (9:1, 100 mL) to yield pure 19 (240 mg,
0.41 mmol, 93%) as a colorless foam. 1H NMR (300 MHz, CDCl3/
CD3OD (4:1)): δ 1.93, 2.06 (2m, 2H, H-5′); 2.78 (m, 1H, H-3′); 3.75
(m, 2H, H-6′), 4.39 (m, 3H, H-4′, H-3′′); 5.81 (s, 1H, H-1′); 5.82 (d,
1H, J ) 4.2 Hz, H-2′); 7.22-7.58 (m, 10H, H-5, m,p-Bz); 7.78-7.92
(m, 7H, H-6, o-Bz). FABMS: m/z 606 (M + Na+). Anal. Calcd for
C32H29N3O8: C, 65.86; H, 5.01; N, 7.20. Found: C, 65.79; H, 5.31;
N, 6.99. For the related conversions of 4 and 10b, see the supporting
information.
N2-Isobutyryl-9-[2′-benzoyl-3′-[(benzoyloxy)methyl]-6′-bromo-
3′,5′,6′-trideoxy-â-allofuranosyl]guanine (23). Compound 20 (765
mg, 1.30 mmol) and PPh3 (511 mg, 1.95 mmol) were dried under high
vacuum and then dissolved in CH3CN (12 mL). A solution of CBr4
(535 mg, 1.61 mmol) in CH3CN (3 mL) was added and stirring
continued for 45 min. The crude foam resulting from concentration
of the solution in Vacuo was first chromatographed on silica gel (120
(53) Baizer, M. M.; Clark, J. R.; Dub, M.; Loter, A. J. Org. Chem. 1956,
21, 1276-1277.

4528 J. Am. Chem. Soc., Vol. 118, No. 19, 1996
Richert et al.
g, EtOAc/CH2Cl2/MeOH (80:20:1)). Two fractions were collected, one
of pure bromide and one containing bromide/P(O)Ph3. The latter
yielded pure bromide after a second column (120 g silica) using EtOAc/
CH2Cl2 (3:1) and a MeOH gradient (1% to 2% to 4%). The combined
pure fractions of bromide 23 were concentrated in Vacuo to a white
foam (635 mg, 75%). 1H NMR (400 MHz, CDCl3): δ 1.227 (d, J )
6.9 Hz, 3H, CH(CH3)2); 1.254 (d, J ) 6.9 Hz, 3H, CH(CH3)2); 2.291-
2.427 (m, 2H, 2 × H-5′); 2.698 (hept, J ) 6.9 Hz, 1H, CH(CH3)2);
3.475-3.553 (m, 1H, H-6′); 3.592-3.645 (m, 1H, H-6′); 3.699-3.772
(m, 1H, H-3′); 4.471(dd, J ) 11.5, 6.7 Hz, 1H, H-3′′); 4.486-4.530
(m, 1H, H-4′); 4.574 (dd, J ) 11.5, 7.1 Hz, 1H, H-3′′); 5.933 (d, J )
1.5 Hz, 1H, H-1′); 6.147 (dd, J ) 1.4, 5.9 Hz, 1H, H-2′); 7.373-7.455
(m, 4H, 4 × m-Bz); 7.522-7.618 (m, 2H, 2 × p-Bz); 7.750 (s, 1H,
H-8); 7.764-7.928 (AA′BB′C system, AA′ part, 2H, 2 × o-Bz); 7.990-
8.018 (AA′BB′C system, AA′ part, 2H, 2 × o-Bz); 9.470 (br s, 1H,
N2-H); 12.120 (br s, 1H, N1-H). FABMS: m/z 652 (M + H+). Anal.
Calcd for C30H30N5O7Br: C, 55.22; H, 4.63; N, 10.73; Br, 12.25.
Found: C, 55.03; H, 4.79; N, 10.42; Br, 12.45. For the related
conversions of 18 and 19, see the supporting information.
N2-Isobutyryl-O6-[2-(4-nitrophenyl)ethyl]-9-[3′-(mercaptomethyl)-
6′-(tert-butyldiphenylsilyl)-3′,5′-dideoxy-â-allofuranosyl]guanine (24).
Compound 17 (400 mg, 485 µmol) was dissolved in Ar-saturated MeOH
(36 mL) at 0 °C. Gaseous NH3 was gently bubbled into the stirred
solution for 10 min. The lines were removed, and the solution was
stirred for 65 min at 0 °C before the solvent and NH3 were removed
by rotary evaporation (P ) 700-25 Torr, Tbath ) ca. 10 °C). Residual
NH3 and acetamide were removed by high vacuum (14 h) to yield the
glass 24, which was used in the subsequent coupling reaction (to give
31) without further purification. An analytically pure sample of 24
(14 mg, 87%; data below) was obtained by the ammonolysis of 17 mg
of 17 in a similar manner, followed by chromatographic removal of
traces of presumed disulfide and/or C2′-OAc side products (the latter
resulting from migration of the acetyl group from C3′′-sulfur) using
silica gel (1.3 g) and petroleum ether/EtOAc (2:3) as the eluant. 1H
NMR (400 MHz, CDCl3): δ 1.064 (s, 9H, C(CH3)3); 1.279 (d, J ) 6.9
Hz, 3H, CH(CH3)2); 1.283 (d, J ) 6.9 Hz, 3H, CH(CH3)2); 1.622 (t, J
) 8.5 Hz, 1H, SΗ); 1.728-1.838 (m, 1H, H-5′); 1.853-1.920 (m,
1H, H-5′); 2.403-2.468 (m, 1H, H-3′); 2.581 (hept, J ) 6.9 Hz, 1H,
CH(CH3)2); 2.644 (ddd, J ) 13.7, 8.6, ca. 8.4, 1H, H-3′′); 3.115 (ddd,
J ) 13.8, 8.2, 5.5 Hz, 1H, H-3′′); 3.309 (t, J ) 6.8 Hz, 2H, OCH2-
CH2Ar); 3.7616-3.860 (m, 2H, 2 × H-6′); 4.540-4.584 (m ) pent,
1H, H-4′); 4.706 (ddd, J ) 7.8, 4.8, 3.2 Hz, 1H, H-2′); 4.734-4.826
(m ) dt, 2H, OCH2CH2Ar); 5.736 (d, J ) 5.1 Hz, 1H, H-1′); 6.385 (br
s, 1H, C-2′-OH); 7.356-7.452 (m, 6H, 4 × m-Ph, 2 × p-Ph); 7.496
(AA′BB′ system, AA′ part, 2H, 2 × o-PhNO2); 7.635-7.702 (m, 4H,
4 × o-Ph); 7.863 (d, J ) 0.2 Hz, 1H, H-8); 7.883 (s, 1H, N2-H); 8.159-
8.194 (AA′BB′ system, BB′ part, 2H, 2 × m-PhNO2). FABMS: m/z
785 (M + H+).
Dimer Thioether 27. A mixture of 22 (599 mg, 929 µmol), 15
(527 mg, 930 µmol), and Cs2CO3 (1.21 g, 3.72 mmol) was dissolved
in CH3CN (25 mL) and water (2.9 mL) under Ar. The solution was
stirred (50 °C, 24 h). The solution was concentrated (rotary evapora-
tion) by 75% and neutralized with acetate buffer (1.9 mL). CH2Cl2
(30 mL) and 9/10 saturated brine (10 mL) were added, the organic
layer was separated, and the aqueous phase was extracted 2× with
CH2Cl2. The combined organic phases were concentrated in Vacuo,
and the residue was chromatographed (120 g of silica, CH2Cl2/EtOAc
(2:1) and a MeOH gradient (1% to 5%)) to give unreacted bromide
(52.9 mg, 80 µmol), 2′,3′′-diacetate of the uridine thiol (80.2 mg, 130
µmol), 27 (572 mg, 505 µmol), and 29 (228 mg, 209 µmol). The
(combined) yield of dimer was 77% (84% based on recovered bromide).
The following are data for 27. 1H NMR (300 MHz, CDCl3): δ 1.05
(s, 9H, CH3-tBu); 1.80 (m, 1H, H-5′-U); 2.0-2.25 (m, 3H, 2 × H-5′-
C, H-5′-U); 2.13 (s, 3H, COCH3); 2.50-2.70 (m, 3H, H-6′-C, H-3′-
U); 2.72, 2.83 (2m, 2H, H-3′′-U); 3.06 (m, 1H, H-3′-C); 3.84 (m, 2H,
H-6′-U); 4.04 (m, 1H, H-4′-U); 4.39 (m, 1H, H-4′-C); 4.57 (AA′B
system, AA′ part, 2H, H-3′′-C); 5.39 (d, 1H, J ) 1.9 Hz, H-1′-U);
5.49 (dd ) d, 1H, J ) 3.6 Hz, H-2′-U); 5.67 (d, 1H, J ) 8.2 Hz,
H-5-U); 5.83 (d, 1H, J ) 1.9 Hz, H-1′-C); 5.96 (dd, 1H, J ) 1.6, 6.8
Hz, H-2′-C); 7.13 (d, 1H, J ) 8.2 Hz, H-6-U); 7.33-7.70 (m, 20H,
o,m,p-Ph, m,p-Bz, H-5-C); 7.77 (d, 1H, J ) 7 Hz, H-6-C); 7.90-8.06
(m, 6H, o-Bz); 8.56 (br s, 1H, NH); 8.85 (br s, 1H, NH). FABMS:
m/z 1157 (M + Na+). For the related conversions of 16/21 and 17/24,
see the supporting information.
Dimer Sulfone 34. Compound 31 (563 mg, 0.42 mmol) was
dissolved in MeOH (76 mL) and THF (11 mL). A freshly prepared
solution of Oxone (2 KHSO5‚KHSO4‚K2SO4; 1.02 g, 1.66 mmol) and
NaOAc (450 mg, 5.50 mmol) in water (15 mL) was added under
vigorous stirring. The white slurry was stirred (2 h, RT). Ap-
proximately half of the organic solvent was evaporated, and saturated
aqueous Na2S2O3 (30 mL) and CH2Cl2 (120 mL) were added. The
organic layer was separated and the aqueous layer extracted twice with
CH2Cl2. The combined organic phases were washed with brine, which
was again extracted twice with CH2Cl2. The combined organic layers
were concentrated in Vacuo, and the crude foam was chromatographed
on silica gel (95 g, stepwise gradient CH2Cl2/MeOH/acetone (96:3:1
to 93:5:2 to 91:7:3)). Compound 34 was isolated as a white foam (555
mg, 96%). 1H NMR (500 MHz, CDCl3): δ 1.041 (s, 9H, C(CH3)3);
1.151-1.207 (m ) 4 × d, 12H, 2 × CH(CH3)2); 1.763-1.829 (m,
1H, H-5′-G1); 1.958-2.022 (m, 1H, H-5′-G1); 2.482-2.530 (m, 1H,
H-5′-G2); 2.594-2.621 (m, 1H, CH(CH3)2-G1); 2.649-2.772 (m, 2H,
CH(CH3)2-G2, H-5′-G2); 2.980-3.001 (m, 2H, H-3′′-G1, H-3′-G1);
3.310 (t, J ) 6.8 Hz, 2H, OCH2CH2Ar); 3.374-3.452 (m, 2H, H-6′-
G2, H-3′-G2); 3.588-3.647 (m, 1H, H-6′-G2); 3.823-3.848 (m, 2H,
2 × H-6′-G1); 3.909 (dd, J ) 16.0, 9.2, 1H, H-3′′-G1); 4.385-4.423
(m, 1H, H-4′-G1); 4.509 (dd, J ) 11.6, 6.3, 1H, H-3′′-G2); 4.532-
4.571 (m, 1H, H-4′-G2); 4.648 (dd, J ) 11.6, 6.4, 1H, H-3′′-G2);
4.697-4.726 (m ) ddd, 1H, H-2′-G1); 4.726-4.837 (m, 2H, OCH2CH2-
Ar); 5.617 (d, J ) 3.4 Hz, 1H, H-1′-G1); 5.994 (s, 1H, H-1′-G2); 6.000
(dd, J ) 3.7, 8.7 Hz, 1H, H-2′-G2); 6.402 (br s ) d, 1H, OH); 7.287-
7.419 (m, 10H, 4 × m-Bz, 2 × p-Ph, 4 × m-Ph); 7.450-7.482 (m,
1H, p-Bz); 7.482-7.510 (m, 2H, AA′BB′ system, AA′ part, 2H, 2 ×
o-PhNO2); 7.552-7.588 (m, 1H, p-Bz); 7.626-7.666 (m, 4H, 4 ×
o-Ph); 7.716 (s, 1H, H-8-G2); 7.801 (s, 1H, H-8-G1); 7.872-7.892
(AA′BB′C system, AA′ part, 2H, 2 × o-Bz); 7.935-7.955 (AA′BB′C
system, AA′ part, 2H, 2 × o-Bz); 8.133-8.168 (m, 3H, N2-H-G1,
AA′BB′ system, BB′ part, 2H, 2 × m-PhNO2); 9.528 (br s, 1H, N2-
H-G2); 12.190 (br s, 1H, N1-H-G2). FABMS: m/z 1389 (M + H+).
For the related conversions of 27 and 28, see the supporting information.
Dimer Triol 35. A solution of 32 (73.7 mg, 63.3 µmol) in MeOH
(20 mL) and THF (7 mL) was cooled to 0 °C, and a 0.2 M aqueous
LiOH solution (1.55 mL, 309 µmol) was added. The solution was
stirred (90 min, 0 °C). Acetate buffer (2 mL) and CH2Cl2 (120 mL)
were added, and the aqueous layer was separated and extracted five
times with CH2Cl2. The combined organic phases were concentrated
in Vacuo and the residue dried under high vacuum. Silica gel
chromatography (18 g of silica, CH2Cl2 with an EtOH/MeOH (1:1)
gradient of 4% to 15% and a water gradient of 0.1% to 1%) yielded
35 (31.3 mg, 34.2 µmol, 54%) as a glass. 1H NMR (300 MHz,
CDCl3): δ 1.00 (s, 9H, CH3-tBu); 1.11 (m, 1H, H-5′-U); 1.83 (m, 1H,
H-5′-U); 2.02 (m, 1H, H-3′-C); 2.23 (m, 2H, H-5′-C); 2.41 (m, 1H,
H-3′-U); 2.93 (m, 1H, H-3′′-U); 3.33 (m, 1H, H-6′-C); 3.50, 3.57 (2m,
2H, H-6′-U); 3.74 (m, 1H, H-3′′-C); 3.87 (m, 1H, H-3′′-C); 3.94 (m,
1H, H-6′-C); 3.99 (dd, 1H, H-3′′-U); 4.20 (t, 1H, H-4′-U); 4.29 (t, 1H,
H-4′-C); 4.42 (br s, 1H, H-2′-C); 4.59 (d, 1H, J ) 8.6 Hz, OH 3′′-C);
4.96 (t, 1H, H-2′-U); 5.22 (dd, 1H, J ) 8, 2 Hz, H-5-U); 5.67, 5.69
(2s, 2H, H-1′-C and H-1′-U); 6.32 (d, 1H, J ) 4.3 Hz, OH 2′-U); 6.65
(br s, 1H, OH 2′-C); 6.97 (d, 1H, J ) 8.1 Hz, H-6-U); 7.16 (d, 1H, J
) 7.3 Hz, H-5-C); 7.25-7.5 (m, 9H, m,p-Bz, m,p-Ph); 7.58 (AA′BB′C
system, AA′ part, 4H, o-Ph); 7.81 (d, 1H, J ) 7.5 Hz, H-6-C); 8.01
(AA′BB′C system, AA′ part, 2H, o-Bz); 9.18 (br s, 1H, NH); 11.72
(br s, 1H, NH). FABMS: m/z 938 (M + Na+). For the related
conversion of 33, see the supporting information.
Dimer Thioester 38. A CH2Cl2 solution of 36 (446 mg, 475 µmol)
was coevaporated with toluene, dried (14 h, 35 °C, high vacuum), and
dissolved in CH3CN (5 mL) and THF (3 mL). PPh3 (249 mg, 950
µmol) was dissolved in CH3CN (2 mL), cooled quickly to 0 °C, and
treated with DIAD (178 µL, 183 mg, 902 µmol). After 15 min, the
solution of 36 was added, immediately followed by thioacetic acid (67.8
µL, 72.6 mg, 950 µmol). The solution was stirred (45 min at RT),
concentrated by rotary evaporation, and dried (high vacuum). The
residue was chromatographed on silica (135 g; CH2Cl2/2-propanol (95:
5) and a MeOH gradient of 2% to 12%) to yield 38 (436 mg, 437
µmol, 92%) as a foam. 1H NMR (300 MHz, CDCl3/CD3OD/D2O

Nonionic Analogs of RNA with Dimethylene Sulfone Bridges
J. Am. Chem. Soc., Vol. 118, No. 19, 1996 4529
(4:1:saturated)): δ 0.99 (s, 9H, CH3-tBu); 1.86, 2.01 (2m, 2 × 2H,
H-5′); 2.24 (s, 3H, Ac); 2.37 (m, 1H, H-3′); 2.69 (m, 1H, H-3′); 2.91
(d, 2H, J ) 7.4 Hz, H-3′′); 3.18, 3.30 (2m, 2H, H-6′-U); 3.65 (m, 2H,
H-3′′); 3.82 (m, 2H, H-6′-A); 3.98 (dt, 1H, J ) 2.0, 10.0 Hz, H-4′);
4.18 (d, 1H, J ) 3.5 Hz, H-2′-U); 4.24 (dt, 1H, J ) 2.2, 10.0 Hz,
H-4′); 4.70 (dd, 1H, J ) 5.2 Hz, H-2′-A); 5.52 (s, 1H, H-1′-U); 5.60
(d, 1H, J ) 8.0 Hz, H-5-U); 6.00 (s, 1H, H-1′-A); 7.27-7.57 (m, 10H,
H-6-U, m,p-Ph, m,p-Bz); 7.58 (2 AA′BB′C systems, 2 AA′ parts, 4H,
o-Ph); 7.99 (AA′BB′C system, AA′ part, 2H, o-Bz); 8.04 (s, 1H, H-8-
A); 8.66 (s, 1H, H-2-A). FABMS: m/z 1020 (M + Na+). For the
related conversion of 35, see the supporting information.
Dimer Alcohol 40. Compound 34 (455 mg, 0.33 mmol) was
dissolved in pyridine (46 µL) in a 10 mL polypropylene tube under Ar
and treated with HF/pyridine (319 µL, 1.64 mmol of a 5.2 M solution)
for 14 h at RT. Methoxytrimethylsilane (0.45 mL, 3.28 mmol) was
added. The mixture was stirred (5 min), transferred (CH2Cl2 wash) to
a glass flask, and concentrated in Vacuo. The oily foam was
chromatographed (silica gel, 32 g, gradient of CH2Cl2/MeOH/acetone
(92:6:2 to 9:1:0)) to yield 40 as a foam (339 mg, 90%). 1H NMR
(500 MHz, CDCl3): δ 1.151-1.197 (m ) 4 × d, 12H, 2 × CH(CH3)2);
1.866-1.962 (m, 2H, 2 × H-5′-G1); 2.39-2.47 (m, 1H, H-5′-G2);
2.53-2.730 (m, 3H, H-5′-G2, CH(CH3)2-G2, CH(CH3)2-G1); 3.23-
3-31 (m, 5H, H-3′′-G1, H-3′-G1, OCH2CH2Ar, C-6′-OH); 3.33-3.464
(m, 2H, H-6′-G2, H-3′-G2); 3.58-3.66 (m, 1H, H-6′-G2); 3.70-3.811
(m, 3H, 2 × H-6′-G1, H-3′′-G1); 4.346-4.383 (m, 1H, H-4′-G1); 4.43-
4.491 (m, 2H, H-4′-G2, H-3′′-G2) 4.556-4.592 (m, 1H, H-3′′-G2);
4.730-4.797 (m, 2H, OCH2CH2Ar); 4.80-4.85 (m, 1H, H-2′-G1);
5.788 (d, J ) 2.7 Hz, 1H, H-1′-G1); 5.842 (br s, 1H, C-2′-G1-OH);
5.948 (d, J ) 2.7 Hz, 1H, H-1′-G2); 5.995 (dd, J ) 2.8, 6.3 Hz, 1H,
H-2′-G2); 7.28-7.41 (m, 4H, 4 × m-Bz); 7.42-7.50 (m, 3H, p-Bz,
AA′BB′ system, AA′ part, 2H, 2 × o-PhNO2); 7.52-7.57 (m, 1H,
p-Bz); 7.777 (s, 1H, H-8-G2); 7.84-7.94 (2 × AA′BB′C system, 2 ×
AA′ part, 4H, 4 × o-Bz); 7.963 (s, 1H, H-8-G1); 8.114-8.149 (m,
2H, AA′BB′ system, BB′ part, 2H, 2 × m-PhNO2); 8.351 (s, 1H, N2-
H-G1); 9.82 (br s, 1H, N2-H-G2); 12.21 (br s, 1H, N1-H-G2).
FABMS: m/z 1151 (M + H+). For the related conversion of 33, see
the supporting information.
Dimer Bromide 41. A mixture of 39 (50.8 mg, 54 µmol) and PPh3
(28.1 mg, 107 µmol) in CH3CN (3 mL) was mixed with a solution of
CBr4 (42.6 mg, 129 µmol) in CH3CN (0.5 mL) at RT. The solution
became yellow after a few minutes. After 90 min, the solution was
added to a mixture of saturated bicarbonate (15 mL), ice (10 g), and
CH2Cl2 (100 mL). The organic layer was separated and the aqueous
layer reextracted four times with CH2Cl2. The combined organic layers
were concentrated and dried (high vacuum). The residue was applied
to a silica (57 g) column. Elution with EtOAc/CH2Cl2/THF (75:25:
10, 300 mL) and then CH2Cl2/EtOAc (83:17, and a MeOH gradient of
4% to 8%) yielded 41 (47.0 mg, 46.4 µmol, 86%) as a colorless foam.
1H NMR (300 MHz, CDCl3): δ 2.19 (s, 3H, Ac); 2.31 (m, 3H, 2 ×
H-5′-A, H-5′-U); 2.58 (m, 1H, H-5′-U); 3.10 (dd, 1H, H-3′′-A); 3.12
(m, 1H, H-3′); 3.45 (m, 5H, H-3′, H-3′′-A, 3 × H-6′); 3.87 (m, 1H,
H-6′); 4.28 (quint, 1H, H-4′); 4.39 (m, 1H, H-4′); 4.52, 4.69 (2dd, 2H,
H-3′′-U); 5.41 (d, 1H, J ) 1.9 Hz, H-1′-U); 5.70 (dd, 1H, J ) 1.9, 7.9
Hz, H-5-U); 5.80 (dd, 1H, J ) 1.9, 7.2 Hz, H-2′-U); 5.85 (d, 1H, J )
5.3 Hz, H-2′-A); 6.05 (s, 1H, H-1′-A); 7.21 (d, 1H, J ) 8.1 Hz, H-6-
U); 7.38-7.63 (m, 9H, o,m-Bz) 7.98 (2 AA′BB′C systems, 2 AA′ parts,
4H, o-OBz); 8.09 (AA′BB′C system, AA′ part ) d, 2H, J ) 7.2 Hz,
o-NBz); 8.15 (s, 1H, H-8-A); 8.79 (s, 1H, H-2-A); 9.53 (br s, 1H, NH);
10.13 (br s, 1H, NH). FABMS: m/z 1038/1040 (M + Na+). For the
related conversions of 40, see the supporting information.
Tetramer Thioether 43. Compound 38 (83.3 mg, 84 µmol),
compound 42 (92 mg, 76 µmol), and Cs2CO3 (149 mg, 456 µmol) were
dried (high vacuum, 50 °C) and pulverized with a stirring bar. THF
(30 mL) was added under stirring, and after 11 h the solvent was
removed from the slurry in Vacuo. Acetate buffer (0.28 mL) and 2/3
saturated brine (9 mL) were added, and the mixture was extracted 5×
with CH2Cl2. The organic layers were concentrated and the residue
chromatographed (50 g of silica, CH2Cl2 and a 2-propanol gradient of
3% to 0% with a concurrent MeOH gradient of 3% to 12.5%) to yield
42 (5.5 mg, 4.5 µmol), AU-2′,3′′-diacetate (7.4 mg, 7 µmol), 43 (57
mg, 27 µmol, 36%), the U2-2′-alcohol tetramer (20.8 mg, 10 µmol,
13%), and a mixture of both tetrameric species (30.5 mg, 14 µmol,
18%). The combined yield of tetramers was 67% (71% based on
recovered bromide). A reaction on a 0.1-fold scale gave a 76%
combined yield of tetramers. 1H NMR (500 MHz, DMSO-d6): δ 0.95
(s, 9H, CH3-tBu); 1.06 (m, 12H, CH3-iBu); 1.93 (3H), 2.00 (1H), 2.14-
2.37 (4H), 2.44 (1H) (4m, 8 × H-5′, H-3′); 2.51-2.70 (m, 4H, 2 ×
H-6′-G3, 2 × H-3′′-U); 2.74, 2.85 (2m, 2H, CH-iBu); 2.97 (m, 2H,
H-3′-A, H-3′-G4); 3.26-3.39 (partly under H2O), 3.40-3.58 (6H) (2m,
CH2-NPE, 2 × H-6′-U, 2 × H-6′-G4, H-3′-G4, 2 × H-3′′-A, 2 × H3′′-
G3); 3.77 (m, 2H, 2 × H-6′-A); 3.90 (dt, 1H, J ) 2.2, 8.9 Hz, H-4′);
4.01 (dt, 1H, J ) 2.7, 8.9 Hz, H-4′); 4.16 (dt, 1H, J ) 2.1, 9.2 Hz,
H-4′); 4.34 (br s, 1H, OH); 4.51 (dt, 1H, J ) 2.8, 8.3 Hz, H-4′-G4);
4.57 (dd, 1H, H-3′′-G4); 4.67 (m, 2H, H-3′′-G4, H-2′-A); 4.78 (m, 3H,
H-2′-G3, CH2-NPE); 5.31 (dd, 1H, J ) 2.2, 6.3 Hz, H-2′-U); 5.63 (d,
1H, J ) 8.0 Hz, H-5-U); 5.65 (d, 1H, J ) 2.4 Hz, H-1′-U); 5.88 (d,
1H, J ) 2.0 Hz, H-1′-G3); 6.03 (s, 1H, H-1′-A); 6.04 (dd, partly hidden,
J ) 1.8 Hz, H-2′-G4); 6.14 (d, 1H, J ) 2.1 Hz, H-1′-G4); 6.23 (br s,
1H, OH); 7.27-7.52 (10H), 7.54-7.62 (7H); 7.63-7.68 (5H) (3m,
H-6-U, 3 × m,p-Bz, 2 × o,m,p-Ph, 2 × o-NPE); 7.91, 7.97 (2 AA′BB′C
systems, 2 AA′ parts, 2 × 2H, o-OBz); 8.05 (AA′BB′C system, AA′
part, 2H, o-Bz-A); 8.17 (AA′BB′ system, BB′ part, 2H, m-NPE); 8.26
(s, 1H, H-8-G); 8.34 (s, 1H, H-8-G); 8.55 (s, 1H, H-8-A); 8.70 (s, 1H,
H-2-A); 10.32 (br s, 1H, NH); 11.4 (br s, ca. 3H, NH); 12.4 (very br
s, 1H, NH). FABMS: m/z 2152 (M + Na+). For the related conversion
of 37/41, see the supporting information.
Tetramer Sulfone 46. Following the procedure and general workup
used for 34, 44 (86.0 mg, 45.5 µmol) was reacted for 3 h with Oxone
(112 mg, 192 µmol) and NaOAc (49.3 mg, 600 µmol) in a mixture of
MeOH (25 mL), THF (10 mL), and water (5 mL). The residue was
filtered over silica, and 46 (87.2 mg, 45 µmol, 99%) was isolated as a
colorless foam. 1H NMR (300 MHz, CDCl3/CD3OD/D2O (3:1:
saturated)): δ 0.94 (s, 9H, CH3-tBu); 1.69 (2H); 1.90 (2H), 2.09 (3H),
2.32 (3H) (4m, 8 × H-5′, 2 × H-3′); 2.00, 2.05 (2s, 2 × 3H, 2 ×
COCH3); 2.73 (d, 1H, J ) 12.8 Hz, H-3′′); 2.98-3.42 (9H), 3.43-
3.68 (3H), 3.69-3.81 (3H) (3m, 2 × H-3′, 5 × H-3′′, 8 × H-6′), 3.88
(m, 1H, H-4′); 4.04 (dt, 1H, J ) 2.5, 8.8 Hz, H-4′); 4.13 (dt, 1H, J )
2, 9 Hz, H-4′); 4.27 (m, 1H, H-4′); 4.37 (d, 1H, J ) 5.1 Hz, H-2′-U1);
4.41, 4.53 (2dd, 2H, H-3′′-U2); 5.28 (s, 1H, H-1′-Py); 5.49 (m, 3H,
H-1′-Py, H-2′-Py, H-5-U), 5.55 (s, 1H, H-1′-Py); 5.59 (d, 1H, J ) 8.0
Hz, H-5-U); 5.73 (m, 2H, H-2′-Py, H-2′-A); 5.96 (s, 1H, H.-1′-A); 7.20
(d, 1H, J ) 8.1 Hz, H-6-U); 7.23-7.48 (m, 20H, H-6-U, H-5-C, m,p-
Ph, m,p-Bz); 7.53 (2 AA′BB′C systems, 2 AA′ parts, 4H, o-Ph); 7.62
(d, 1H, J ) 7.6 Hz, H-6-C); 7.82 (m, 6H, 6 × o-Bz); 7.94 (AA′BB′C
system, AA′ part ) d, 2H, J ) 7.2 Hz, o-Bz-A); 8.18 (s, 1H, H-8-A);
8.63 (s, 1H, H-2-A). FABMS: m/z 1963 (M + Na+). For the related
conversion of 43, see the supporting information.
Tetramer Pentaol 47. Compound 45 (62.2 mg, 29.2 µmol) was
hydrolyzed with 0.2 M LiOH (1.1 mL, 220 µmol; 65 min). The mixture
was neutralized and extracted with a mixture of CH2Cl2 and EtOH (9:
1), the solvents were removed, and the residue was chromatographed
(17 g of silica; CH2Cl2/MeOH/EtOH/water (84:7.5:7.5:1), 200 mL, and
then EtOAc/MeOH/water (76:18:3), 150 mL). Pentaol 47 (36.8 mg,
19.3 µmol, 66%) and the debenzoylated adenine side product 48 (3.3
mg, 1.8 µmol) were obtained as colorless solids. The following data
are for 47. 1H NMR (500 MHz, CDCl3/CD3OD/D2O (3:1:saturated)):
δ 1.00 (s, 9H, CH3-tBu); 1.17 (m, 12H, CH3-iBu); 1.84 (m, 1H, H-5′-
A); 2.02 (4H), 2.20-2.42 (4H), (2m, 7 × H-5′, H-3′-U); 2.52, 2.60,
2.66 (3m, 3H, 2 × CH-iBu, H-3′-A); 2.78, 2.83 (2m, 2H, 2 × H-3′-
G); 2.92 (d, 1H, J ) 12.4 Hz, H-3′′); 2.99 (m, 2H, 2 × H-6′); 3.15
(dd, 1H, J ) 2.5, 14.5 H, H-3′′); 3.21 (2H), 3.33 (3H), 3.54 (2H), 3.72
(2H) (4m, 5 × H-3′′, 4 × H-6′); 3.21 (m, 2H, CH2-NPE); 3.83 (m, 3H,
2 × H-6′-A, H-3′′-G4); 3.97 (t, 1H, J ) 8.0; H-4′); 4.16 (m, 2H, 2 ×
H-4′); 4.24 (dt, 1H, J ) 2.3, 9.7 Hz, H-4′); 4.45 (d, 1H, J ) 5.0 Hz,
H-2′-U); 4.69 (m, 4H, 2 × H-2′, CH2-NPE); 4.81 (d, 1H, J ) 5.5 Hz,
H-2′-G3); 5.44 (s, 1H, H-1′-U); 5.54 (d, 1H, J ) 8.0 Hz, H-5-U); 5.69
(s, 1H, H-1′-G); 5.88 (s, 1H, H-1′-G); 6.01 (s, 1H, H-1′-A); 7.29-
7.39 (7H), 7.45 (4H), (2m, H-6-U, m,p-Ph, m-Bz, o-NPE); 7.54
(AA′BB′C system, C part, 1H, p-Bz); 7.61 (2 AA′BB′C systems, 2
AA′ parts, 4H, o-Ph); 7.71 (br s, 1H, H-8-G4); 7.99 (s, 1H, H-8-G3);
8.01 (AA′BB′C system, AA′ part ) d, 2H, J ) 7.2 Hz, o-Bz-A); 8.06
(s, 1H, H-8-A); 8.07 (AA′BB′ system, BB′ part, 2H, m-NPE); 8.58 (s,
1H, H-2-A). FABMS: m/z 1956 (M + 2Na+ - H+). For the related
conversion of 48, see the supporting information.

4530 J. Am. Chem. Soc., Vol. 118, No. 19, 1996
Richert et al.
Tetramer Diol 49. Compound 46 (29.0 mg, 15.0 µmol) in CH2-
Cl2/MeOH (9:1, 1.2 mL) was dried in an Eppendorf reaction vessel
under Ar stream and then high vacuum. The residue, in 0.25 mL of
pyridine, was treated with a solution of HF in pyridine (5 M, 250 µL,
1.25 mmol), and the mixture was stirred for 8 h at RT. Trimethyl-
methoxysilane (350 µL, 317 mg, 3.03 mmol) was added to quench
residual HF, yielding a colorless precipitate. After 30 min at RT, the
mixture was transferred to a flask and dried in Vacuo. The residue
was chromatographed on silica (14 g; CH2Cl2/MeOH/H2O (93:7:0.5
(80 mL) followed by 88:11:0.65 (75 mL) and then EtOAc/MeOH/H2O
(79:18:3)) to yield dialcohol 49 (19.7 mg, 11.6 µmol, 77%) as a
colorless solid. 1H NMR (500 MHz, DMSO-d6): δ 1.74 (m, 1H, H-5′-
U1); 1.94 (1H), 2.03 (3H), 2.23 (2H), 2.37 (2H), 2.45 (2H) (5m, 7 ×
H-5′, 3 × H-3′); 2.07 (s, 3H, COCH3); 2.13 (s, 3H, COCH3); 3.10 (m,
3H, H-3′-A, 2 × H-3′′); 3.23-3.66 (m, 12H, 8 × H-6′, 4 × H-3′′);
3.94 (m, 2H, 2 × H-4′); 4.16 (t, 1H, J ) 10 Hz, H-4′); 4.26 (m, 1H,
OH-6′-U1); 4.40 (dt, 1H, J ) 2.8, 8.8 Hz, H-4′); 4.52 (m, 2H, H-3′′-
U2, H-2′-U1); 4.61 (dd, 1H, H-3′′-U2); 5.45 (d, 1H, J ) 8.2 Hz, H-2′-
Py); 5.62 (d, 1H, J ) 8.0 Hz, H-5-U); 5.63 (d, 1H, J ) 1.6 Hz, H-1′-
Py); 5.68 (d, 1H, J ) 1.6 Hz, H-1′-Py); 5.71 (d, 1H, J ) 8.0 Hz, H-5-
U); 5.74 (d, 1H, J ) 7.2 Hz, H-5-C); 5.81 (dd, 1H, J ) 3.0, 7.5 Hz,
H-2′-Py); 5.94 (d, 1H, J ) 3.0 Hz, H-1′-Py); 6.01 (d, 1H, J ) 5.1 Hz,
H-2′-A); 6.24 (d, 1H, J ) 1.4 Hz, H-1′-A); 7.44 (m, 1H, H-6-U); 7.46-
7.62 (m, 12H, 4 × o,m-Bz); 7.78 (d, 1H, J ) 8.1 Hz, H-6-U); 7.90,
7.96, 8.00 (3 AA′BB′C systems, 3 AA′ parts, 3 × 2H, o-Bz); 8.05
(AA′BB′C system, AA′ part, 2H, o-Bz-A); 8.13 (d, 1H, J ) 7.3 Hz,
H-6-C); 8.66 (s, 1H, H-8-A); 8.75 (s, 1H, H-2-A); 11.33 (br s, 3H, 3
× NH). FABMS: m/z 1739 (M + K+). For the related conversion of
51, see the supporting information.
1H, H-4′); 4.35 (m, 1H, H-4′); 4.48 (m, 2H, H-3′′-U2, H-2′-U1); 4.64
(dd, 1H, H-3′′-U2); 5.20 (s, 1H, H-1′-Py); 5.46 (d, 1H, J ) 2.1 Hz,
H-1′-Py); 5.51 (d, 1H, J ) 5.3 Hz, H-2′-Py); 5.62 (d, 1H, J ) 8.0 Hz,
H-5-U); 5.65 (d, 1H, J ) 8.1 Hz, H-5-U); 5.66 (s, 1H, H-1′-Py); 5.81
(d, 2H, J ) 5.6 Hz, H-2′-Py, H-2′-A); 5.99 (s, 1H, H-1′-A); 7.35-
7.62 (m, 16H, m,p-Bz, 2 × H-6-U, H-5,6-C); 7.92 (m, 6H, o-Bz); 8.02
(AA′BB′C system, AA′ part, 2H, o-Bz-A); 8.17 (s, 1H, H-8-A); 8.70
(s, 1H, H-2-A). FABMS: m/z 1788 (M + Na+). For the related
conversion of 52, see the supporting information.
Octamer 55. Cs2CO3 (4.6 mg, 14 µmol) was dried at 100 °C under
high vacuum for 5 min. 50 (6.8 mg, 3.45 µmol), 54 (6.2 mg, 3.5 µmol),
and anhydrous DMF (0.45 mL) were added, and the slurry was stirred
(RT, 4 h, Ar). Acetate buffer (10 mL) was added, and the solvents
were evaporated in Vacuo. The residue was treated with 50% brine (2
mL) and extracted 5× with CH2Cl2/EtOH (4:1, 5 mL each). The
combined organic phases were concentrated in Vacuo to yield a colorless
solid (18 mg). This was dissolved in MeOH (4 mL), THF (3 mL),
and water (0.2 mL) under stirring and moderate heating. A solution
of Oxone (18.6 mg, 30 µmol) and anhydrous NaOAc (8.1 mg, 99 µmol)
in water (0.6 mL) was added at RT and the turbid suspension stirred
for 9 h. Saturated Na2S2O3 solution (0.2 mL) was added, resulting in
a clear organic phase. Water (2 mL) was added and the mixture
extracted five times with CH2Cl2/EtOH (9:1, 20 mL). The combined
organic phases were evaporated to dryness, and the residue was
dissolved in a mixture of CH3CN and water (3:1, 7 mL) under slight
heating. The solution was filtered and purified by HPLC (Lichrospher
RP18 column, 10 µm, 250 × 16 mm, guard column 40 × 16 mm; flow
9 mL/min; gradient: B ) 85% CH3CN in water, A ) water; 40% to
90% B in 53 min; elution of product after 48 min). Evaporation of
solvents yielded 55 (7.0 mg, 1.85 µmol, 53%; 70% based on recovered
bromide) as a white amorphous solid. In Vacuo concentration of
solutions containing compounds having shorter retention times yielded
the G4-2′-acetyl-3′′-sulfonic acid of 50 (5 min, 2.1 mg, 1.0 µmol), and
54 (24 min, 1.5 mg, 0.84 µmol). 1H NMR (500 MHz, DMSO-d6): δ
0.96 (s, 9H, CH3-tBu); 1.07 (m, 12H, CH3-iBu); 1.82 (2H), 2.05 (8H),
2.19 (2H), 2.28 (4H), 2.39 (2H), 2.47 (partly under DMSO-d5) (6m,
16 × H-5′, 4 × H-3′-Py); 2.07, 2.07, 2.12 (3s, 3 × 3H, 3 × COCH3);
2.86 (m, 2H, 2 × CH-iBu); 2.96 (m, 3H, 3 × H-3′-Pu); 3.02-3.68 (m,
partly under H2O, H-3′-Pu, 16 × H-3′′, CH2-NPE, 14 × H-6′); 3.77
(m, 2H, H-6′-A1); 3.92 (m, 3H, 3 × H-4′-Py); 4.02 (m, 2H, 2 × H-4′);
4.15 (m, 2H, 2 × 4′); 4.31 (d, 1H, J ) 5.4 Hz, H-2′-U); 4.34 (d, 1H,
J ) 5.0 Hz, H-2′-U); 4.39 (dt, 1H, J ) 2.3, 9.0 Hz, H-4′-A6); 4.52
(dd, 1H, J ) 5.8, 11.3 Hz, H-3′′-U8); 4.61 (dd, 1H, J ) 6.7, 11.3 Hz,
H-3′′-U8); 4.64 (d, 1H, J ) 4.8 Hz, H-2′-Pu); 4.77 (m, 3H, H-2′-Pu,
CH2-NPE); 5.01, 5.12 (2 br s, 2H, 2 × OH); 5.38 (d, 1H, J ) 7.2 Hz,
H-2′-Py); 5.55 (m, 1H, H-2′-Py); 5.60 (d, 1H, J ) 8.1 Hz, H-5-U);
5.62 (d, 1H, J ) 8.1 Hz, H-5-U); 5.66 (s, 4H, 4 × H-1′-Py); 5.69 (d,
1H, J ) 7.9 Hz, H-5-U); 5.73 (d, 1H, J ) 6.7 Hz, H-5-C); 5.81 (dd,
1H, J ) 3.0, 7.1 Hz, H-2′); 5.87 (m, 2H, H-2′-A6, H-1′-G); 5.93 (d,
1H, J ) 2.9 Hz, H-1′-G); 5.93 (s, 1H, H-1′-A1); 6.15 (br s, 2H, 2 ×
OH); 6.23 (s, 1H, H-1′-A6); 7.28 (d, 1H, J ) 7.1 Hz, H-6-C); 7.34-
7.66 (m, 29H, 2 × H-6-U, o,m,p-Ph, m,p-Bz, o-NPE); 7.78 (d, 1H, J
) 8.0 Hz, H-6-U); 7.90 (2H); 7.94 (2H), 7.99 (2H), (3 AA′BB′C
systems, 3 AA′ parts, 4 × o-Bz-U8, 2 × o-Bz-C); 7.96 (s, 1H, H-8-
G); 8.06 (m, 4H, 4 × o-Bz-A); 8.17 (AA′BB′ system, BB′ part, 2H,
m-NPE); 8.34 (s, 1H, H-8-G); 8.53, 8.62, 8.67, 8.71 (4s, 4H, 2 × H-8-
A, 2 × H-2-A); 10.35 (br s, ca. 2H, NH); 11.4 (br s, ca. 5H, NH). For
Tetramer Thioester 50. Pentaol 47 (19.7 mg, 10.3 µmol) was
coevaporated with pyridine, dried (high vacuum), and reacted with PPh3
(9.5 mg, 36.0 µmol), DIAD (6.1 µL, 6.3 mg, 31 µmol), and thioacetic
acid (2.5 µL, 2.8 mg, 36 µmol) in a mixture of dioxane (0.45 mL) and
Cl(CH2)2Cl (0.2 mL) for 90 min. The reagent was prepared first, in a
manner similar to that used for compound 38. Chromatography on
silica (10 g; CH2Cl2/MeOH/EtOH/H2O (100:7.5:2.5:0.1 (100 mL)
followed by 100:7.5:7.5:1 (50 mL))) furnished 50 (12.1 mg, 6.1 µmol,
60%, 72% based on recovered 47) and unreacted 47 (3.5 mg, 1.8 µmol)
as colorless solids. 1H NMR (500 MHz, DMSO-d6): δ 0.96 (s, 9H,
CH3-tBu); 1.12 (m, 12H, CH3-iBu); 1.93 (m, 1H, H-5′-A); 2.05 (2H),
2.12 (1H), 2.22 (2H), 2.32 (1H), 2.46 (1H), 2.52 (1H) (6m, 7 × H-5′,
H-3′-U2); 2.32 (s, 3H, COCH3); 2.77, 2.83 (2m, 2H, 2 × CH-iBu);
2.94 (m, 1H, H-3′-A); 3.02 (m, 2H, 2 × H-3′′-G4); 3.11 (m, 1H, H-3′-
G); 3.22 (dd, 1H, J ) 2.0, 11.1 Hz, H-3′′); 3.28-3.43 (m, partly under
HDO), 3.51 (3H) (2m, H-3′-G, 5 H-3′′, 6 × H-6′, CH2-NPE); 3.77 (t,
2H, J ) 6.8 Hz, 2 × H-6′-A); 3.91 (t, 1H, J ) 10.4 Hz; H-4′); 4.03
(m, 2H, 2 × H-4′); 4.24 (dt, 1H, J ) 1.6, 9.7 Hz, H-4′); 4.30 (m, 1H,
H-2′-U); 4.52 (m, 1H, H-2′); 4.69 (m, 1H, H-2′); 4.76 (m, 1H, H-2′);
4.78 (t, 2H, J ) 6.8 Hz, CH2-NPE); 5.61 (d, 1H, J ) 8.1 Hz, H-5-U);
5.65 (d, 1H, J ) 1.9 Hz, H-1′-U); 5.78 (d, 1H, J ) 2.6 Hz, H-1′-G);
5.91 (d, 1H, J ) 2.1 Hz, H-1′-G); 6.01 (d, 1H, J ) 5.4 Hz, OH); 6.03
(d, 1H, J ) 1.6 Hz, H-1′-A); 6.17 (d, 1H, J ) 4.6 Hz, OH); 6.22 (br
s, 1H, OH); 7.34-7.44 (6H), 7.52-7.66 (10H) (2m, H-6-U, o,m,p-Ph,
m,p-Bz, o-NPE); 8.04 (AA′BB′C system, AA′ part, 2H, o-Bz-A); 8.14
(s, 1H, H-8-G4); 8.17 (AA′BB′ system, BB′ part, 2H, m-NPE); 8.39
(s, 1H, H-8-G3); 8.5 (s, 1H, H-8-A); 8.70 (s, 1H, H-2-A), 10.31 (s,
1H, NH); 11.2 (br s, 1H, NH); 11.35 (br s, 2H, 2 × NH); 12.1 (br s,
1H, NH). FABMS: m/z 1993 (M + Na+).
Tetramer Bromide 54. Monomer 18 (25.2 mg, 52.5 µmol),
tetramer 49 (6.5 mg, 3.8 µmol), and PPh3 (22 mg, 84.5 µmol) in
CH3CN (3 mL) and Cl(CH2)2Cl (2 mL) were treated with CBr4 (23.3
mg, 70.4 µmol) in Cl(CH2)2Cl (0.5 mL) for 2 h. Chromatography (12
g of silica; CH2Cl2/EtOAc/MeOH/H2O (50:50:3:0.5) (50 mL) followed
by CH2Cl2/MeOH/H2O (92:7:0.5 (50 mL) and then 88:11:0.65 (40
mL))) gave 21 (14.9 mg, 27.4 µmol, 52%), 18 (9.5 mg, 19.8 µmol), 54
(2.2 mg, 1.25 µmol, 33%, 83% based on recovered starting material),
and 49 (3.9 mg, 2.3 µmol) as colorless glasses or foams. The following
are data for 54. 1H NMR (300 MHz, CDCl3/CD3OD/D2O (4:1:
saturated)): δ 1.95-2.28 (5H), 2.36 (4H), 2.55 (2H) (3m, 8 × H-5′, 3
× H-3′); 2.08 (s, 3H, COCH3-C); 2.13 (s, 3H, COCH3-A); 2.80 (m,
2H, H-3′-A, H-3′′); 3.15 (m, ca. 3H, partly under HDO), 3.21-3.80
(10H), (2m, 5 × H-3′′, 8 × H-6′); 4.08 (m, 2H, 2 × H-4′); 4.19 (m,
1
other 500 MHz H NMR (DMSO-d6/D2O (4:1) and CDCl3/CD3OD/
D2O (5:2:saturated), see the supporting information. FABMS: m/z
3708 (M + Na+).
Deprotected Tetramer 56. Compound 46 (16.7 mg, 8.9 µmol) in
MeOH (1.2 mL) and THF (0.9 mL) was treated with 1 M NaOH (1.2
mL, 1.2 mmol) with stirring (40 °C, 4 h). The mixture was cooled to
RT and diluted with acetate buffer (0.55 mL). A white precipitate
formed. A mixture of CH3CN and water (1:1, 13 mL) was added, and
the resulting solution filtered and injected onto an HPLC column
(Nucleosil 10-CN, 250 × 22.5 mm; flow 11 mL/min; B ) CH3CN
(85%) in water, A ) water; gradient 10% to 35% B in 30 min; retention
time of the product 19 min, no detectable side products) to yield 56
(6.3 mg, 5.3 µmol, 60%) as a colorless solid. Injection of larger
amounts of the crude product onto the HPLC column led to a peak
eluting after 16 min. This was assigned as an aggregate, since it had

Nonionic Analogs of RNA with Dimethylene Sulfone Bridges
J. Am. Chem. Soc., Vol. 118, No. 19, 1996 4531
identical retention time upon reinjection at higher dilution. 1H NMR
(600 MHz, DMSO-d6/D2O (3.5:1), 300 K): δ 1.710 (m, 1H, H-5′-
U1); 1.903 (m, 1H, H-5′-U1); 1.950 (m, 1H, H-5′-C); 1.973 (m, 1H,
H-5′-U4); 2.009 (m, 1H, H-5′-A); 2.033 (m, 1H, H-3′-U4); 2.227 (m,
1H, H-5′-U4); 2.239 (m, 1H, H-5′-C); 2.287 (m, 1H, H-5′-A); 2.296
(m, 1H, H-3′-U1); 2.333 (m, 1H, H-3′-C); 3.025 (m, 1H, H-3′-A); 3.093
(dd, 1H, J ) 2.6, 11.3 Hz, H-3′′-U1); 3.106 (dd, 1H, J ) 3.0, 8.7 Hz,
H-3′′-C); 3.239 (m, 1H, H-6′-A); 3.240 (m, 1H, H-3′′-A); 3.247 (m,
1H, H-6′-A); 3.253 (m, 1H, H-6′-C); 3.276 (m, 1H, H-6′-C); 3.314
(m, 1H, H-6′-U4); 3.331 (m, 1H, H-6′-U4); 3.415 (m, 1H, H-3′′-C);
3.439 (m, 1H, H-3′′-U1); 3.460 (m, 1H, H-3′′-U4); 3.489 (m, 1H, H-6′-
U1); 3.504 (m, 1H, H-3′′-A); 3.568 (m, 1H, H-6′-U1); 3.676 (dd, 1H,
J ) 6.0, 11.0 Hz, H-3′′-U4); 3.922 (dt, 2H, J ) 2.7, 9.6 Hz, H-4′-U1,
H-4′-C); 3.996 (m, partly under HDO, H-4′-U4); 4.070 (dt, 1H, J )
2.6, 8.6 Hz, H-4′-A); 4.199 (d, 1H, J ) 6.7 Hz, H-2′-C); 4.220 (dd,
1H, J ) 2.5, 5.9 Hz, H-2′-U4); 4.228 (dd, 1H, J ) 2.6, 4.3 Hz, H-2′-
U1); 4.697 (dd, 1H, J ) 1.5, 5.7 Hz, H-2′-A); 5.575 (d, 1H, J ) 7.9
Hz, H-5-U); 5.581 (d, 1H, J ) 8.0 Hz, H-5-U); 5.587 (m, 3H, 3 ×
H-1′-Py); 5.815 (d, 1H, J ) 7.4 Hz, H-5-C); 5.899 (d, 1H, J ) 1.6 Hz,
H-1′-A); 7.422 (d, 1H, J ) 8.1 Hz, H-6-U4); 7.449 (d, 1H, J ) 8.0
Hz, H-6-U1); 7.517 (d, 1H, J ) 7.5 Hz, H-6-C); 8.112 (s, 1H, H-2-A);
25% to 35% B in 15 min; elution of 58 after 13.5 min). 1H NMR
(500 MHz, DMSO-d6): δ 1.78 (m, 1H, H-5′-A1); 1.93-2.10 (10H),
2.22-2.42 (8H), 2.55 (partly under DMSO-d5) (3m, 15 × H-5′, 4 ×
H-3′-Py); 2.87 (3H), 3.02 (1H) (2m, 4 × H-3′-Pu); 3.14-3.43 (partly
under H2O), 3.43-3.60 (12H) (2m, 15 × H-3′′, 16 × H-6′); 3.70 (dd,
1H, J ) 5.4, 10.5 Hz, H-3′′-U8); 3.79 (br s, 1H, OH); 3.92 (m, 3H, 3
× H-4′); 4.02 (m, 5H, 5 × H-4′); 4.22 (d, 1H, J ) 5.0 Hz, H-2′-Py);
4.27 (d, 1H, J ) 6.2 Hz, H-2′-Py); 4.32 (m, 2H, 2 × H-2′-Py); 4.57
(m, 2H, 2 × H-2′-Pu); 4.60 (br s, 1H, OH); 4.64 (d, 1H, J ) 5.3 Hz,
H-2′-Pu); 4.70 (d, 1H, J ) 5.1 Hz, H-2′-Pu); 5.59-5.68 (m, 7H, 3 ×
H-5-U, 4 × H-1′); 5.70, 5.72 (2s, 1H, 2 × H-1′); 5.76 (d, 1H, J ) 7.6
Hz, H-5-C); 5.91, 5.95 (2s, 2H, 2 × H-1′-A); 6.05 (1H), 6.18 (4H),
6.38 (1H) (3br s, 6 × OH); 6.56 (2H), 6.64 (2H) (2 br s, 2 × NH2-G);
7.16-7.29 (m, 6H, 2 × NH2-A, NH2-C); 7.39 (m, H-6-U); 7.55-7.64
(m, 3H, 2 × H-6U, H-6-C); 7.84, 7.85 (2s, 2H, 2 × H-8-G); 8.13,
8.14, 8.23, 8.30 (4s, 4H, 2 × H-8-A, 2 × H-2-A). MALDI-TOFMS
(linear, positive mode, 20 kV): m/z 2534.9 (M + Na+), calcd (average
mass) 2534.5. Electrospray MS: m/z 2511.4 (M+), calcd (average mass)
2511.5.
Acknowledgment. The authors are indebted to Marcel
Ko¨nig, Birgitte Hyrup, and Daniel Baeschlin for their contribu-
tion of experimental details and helpful discussions to this study,
and to Zhen Huang for communication of experimental details
for the synthesis of 1. The authors are grateful to Andre´ Mu¨ller
for excellent technical assistance, to Dr. Bernd Schweizer for
solving the crystal structure of compound 11, and to B.
Brandenberg, B. Jaun, I. Schlo¨nvogt, and T. Schulte-Herbru¨ggen
(NMR) and R. Ha¨fliger, H. U. Hediger, O. Greter, and Dr. W.
Amrein (MS) for acquisition of spectra. We are also indebted
to Professor Martin Egli for many interesting discussions. C.R.
was supported by a Kekule´ fellowship from the Stipendien
Fonds des Verbandes der Chemischen Industrie, Frankfurt,
Germany.
1
8.231 (s, 1H, H-8-A). For 500 MHz H and 125 MHz 13C NMR in
DMSO-d6, see the supporting information. MALDI-TOFMS (linear
positive mode, 20 kV): m/z 1222 (M + Na+).
Deprotected Tetramer 57. Compound 45 (23 mg, 10.6 µmol) was
treated with 1 M NaOH (1.3 mL), MeOH (1.3 mL), and THF (1 mL)
for 6 h at 40 °C. HPLC (gradient of CH3CN (85%) in water from
10% to 30% in 33 min, retention time 18 min) yielded 57 (8.5 mg, 6.6
µmol, 63%) as a colorless solid. 1H NMR (500 MHz, DMSO-d6): δ
1.77 (m, 1H, H-5′-A); 2.01 (3H), 2.23 (1H), 2.30 (2H), 2.43 (1H), 2.50
(partly under DMSO-d5) (5m, 8H, 7 × H-5′, H-3′-U); 2.86 (m, 2H, 2
× H-3′); 3.13-3.39 (partly under H2O, integrated after addition of D2O;
9H), 3.40-3.59 (7H) (2m, 8 × H-6′, 7 × H-3′′, H-3′); 3.75 (m, 1H,
H-3′′-G4); 3.91 (dt, 1H, J ) 2.0, 7.5 Hz, H-4′); 4.04 (m, 2H, 2 ×
H-4′); 4.09 (dt, 1H, J ) 3.1, 6.0 Hz, H-4′); 4.31 (d, 1H, J ) 6.2 Hz,
H-2′-U); 4.57 (m, 2H, 2 × H-2′); 4.66 (d, 1H, J ) 5.6 Hz, H-2′); 5.54
(d, 1H, J ) 7.8 Hz, H-5-U); 5.65 (s, 1H, H-1′); 5.66 (d, 1H, J ) 2.7
Hz, H-1′); 5.70 (d, 1H, J ) 1.9 Hz, H-1′); 5.92 (d, 1H, J ) 1.2 Hz,
H-1′-A); 6.25 (br s, 2H, 2 × OH); 6.88, 6.96 (2 br s, 2 × 2H, 2 ×
NH2-G); 7.28 (br s, 2H, NH2-A); 7.51 (d, 1H, J ) 8.1 Hz, H-6-U);
7.73, 7.77 (2s, 2H, 2 × H-8-G); 8.13 (s, 1H, H-2-A); 8.24 (s, 1H, H-8-
A). MALDI-TOFMS (linear, positive mode, 20 kV): m/z 1304 (M +
Na+).
Deprotected Octamer 58. Compound 55 (2.0 mg, 0.54 µmol) was
treated with 1 M NaOH (0.4 mL), MeOH (0.3 mL), and THF (0.2
mL) for 10 h at 40 °C. Acidification with acetate buffer (0.2 mL) and
concentration to 0.8 mL under an Ar stream led to a white precipitate.
After cooling to -20 °C for 5 min, the mixture was centrifuged for 5
min at 8000g and the supernatant was carefully aspired. The precipitate
was washed twice with water and four times with diethyl ether/
CH2Cl2 (9:1) and dried under high vacuum to yield deprotected octamer
58 (0.9 mg, 0.36 µmol, 72%) as an amorphous, colorless solid. This
compound was characterized by HPLC using a Lichrospher RP18
column (10 µm, 250 × 16 mm, guard column 40 × 16 mm, flow 9
mL/min; B ) CH3CN, A ) water, gradient 10% to 25% B in 5 min,
Supporting Information Available: Additional experimen-
tal procedures, assigned NMR and mass spectra (generally FAB)
for all compounds, crystallographic data (crystal data, positional
and thermal parameters, bond lengths and angles; also in CIF
format) and an Ortep plot of 11, UV spectra of 55 in CH2Cl2
with 0.5% and 4% methanol and methanol adjuvant-induced
hyperchromicity plots for 47, 49, and 55, MALDI-TOF spectra
of 57 and 58, an electrospray ionization spectrum of 58, and
UV absorbance versus temperature profile of 58 and of its
natural RNA counterpart (57 pages). This material is contained
in many librairies on microfiche, immediately follows this article
in the microfilm version of the journal, can be ordered from
the ACS, and can be downloaded from the Internet; see any
current masthead page for ordering information and Internet
access instructions.
JA952322M