Oligodeoxyribonucleotide analogues with bridging dimethylene sulfide, sulfoxide, and sulfone groups. Toward a second-generation model of nucleic acid structure

Article abstract of DOI:10.1021/jo0003910

Short DNA analogues with bridging dimethylene sulfide, sulfoxide, and sulfone groups replacing the phosphate diesters (S-DNAs) were synthesized from building blocks prepared via two routes, both starting from D-glucose. Building blocks for RNA analogues were prepared by stereoselective introduction ofnucleobase into a 2′-acylated ribose analogue. The ribose analogues were converted to deoxyribose analogues by replacement of a 3″-OH group by a thioacetyl unit, followed by photolytic deoxygenation or radical-based 2′-deoxygenation. DNA analogues joined via CH₂-S-CH₂ units were prepared by S_N2 displacement of a 6′-mesyl group on one building block using a thiolate nucleophile of another. 4,4′-Dimethoxytrityl protection and deprotection schemes were established for both the thiol and hydroxyl groups. The corresponding sulfoxide DNA analogues were obtained by oxidation with hydrogen peroxide. Sulfone DNA analogues were obtained by oxidation of the sulfide DNA with persulfate or hydrogen peroxide in the presence of a titanium silicate catalyst. The physical properties of several representative oligonucleotide analogues were examined, and interpreted in light of a "second-generation" model for DNA strand-strand recognition, a model that emphasizes the role of the polyanionic backbone in diminishing unwanted tendencies of highly functionalized molecules to form "structure" in solution. Even short sulfide-linked DNA analogues displayed association properties different from those displayed by standard DNA molecules. Complex formation observed with sulfide-linked tetramers by HPLC study in different solvents suggested that the complex is formed using hydrogen bonding. Sulfone-linked dinucleotides display Watson-Crick behavior; the tetramer, however, displayed self-structure. Self-structure and self-aggregation become more prominent as the length of the oligonucleotide analogues increases. The tendency to self-aggregate can be decreased by adding a charged sulfonate group to the 3″-end of the DNA analogue. Features of the second-generation model are important for many areas of nucleic acid chemistry, from the design of nucleic acid therapeutic agents to the search for life on other planets.

Full text of DOI:10.1021/jo0003910

3996
J . Org. Chem. 2002, 67, 3996-4013
Oligod eoxyr ibon u cleotid e An a logu es w ith Br id gin g Dim eth ylen e
Su lfid e, Su lfoxid e, a n d Su lfon e Gr ou p s. Tow a r d a
Secon d -Gen er a tion Mod el of Nu cleic Acid Str u ctu r e†
Zhen Huang‡ and Steven A. Benner*,§
Departments of Chemistry and Anatomy and Cell Biology, University of Florida,
Gainesville, Florida 32611, and Department of Chemistry, Brooklyn College,
Brooklyn, New York 11210-2889
benner@chem.ufl.edu
Received March 17, 2000
Short DNA analogues with bridging dimethylene sulfide, sulfoxide, and sulfone groups replacing
the phosphate diesters (S-DNAs) were synthesized from building blocks prepared via two routes,
both starting from D-glucose. Building blocks for RNA analogues were prepared by stereoselective
introduction of nucleobase into a 2′-acylated ribose analogue. The ribose analogues were converted
to deoxyribose analogues by replacement of a 3′′-OH group by a thioacetyl unit, followed by photolytic
deoxygenation or radical-based 2′-deoxygenation. DNA analogues joined via CH2-S-CH2 units were
prepared by SN2 displacement of a 6′-mesyl group on one building block using a thiolate nucleophile
of another. 4,4′-Dimethoxytrityl protection and deprotection schemes were established for both the
thiol and hydroxyl groups. The corresponding sulfoxide DNA analogues were obtained by oxidation
with hydrogen peroxide. Sulfone DNA analogues were obtained by oxidation of the sulfide DNA
with persulfate or hydrogen peroxide in the presence of a titanium silicate catalyst. The physical
properties of several representative oligonucleotide analogues were examined, and interpreted in
light of a “second-generation” model for DNA strand-strand recognition, a model that emphasizes
the role of the polyanionic backbone in diminishing unwanted tendencies of highly functionalized
molecules to form “structure” in solution. Even short sulfide-linked DNA analogues displayed
association properties different from those displayed by standard DNA molecules. Complex formation
observed with sulfide-linked tetramers by HPLC study in different solvents suggested that the
complex is formed using hydrogen bonding. Sulfone-linked dinucleotides display Watson-Crick
behavior; the tetramer, however, displayed self-structure. Self-structure and self-aggregation become
more prominent as the length of the oligonucleotide analogues increases. The tendency to self-
aggregate can be decreased by adding a charged sulfonate group to the 3′′-end of the DNA analogue.
Features of the second-generation model are important for many areas of nucleic acid chemistry,
from the design of nucleic acid therapeutic agents to the search for life on other planets.
In tr od u ction
molecular recognition systems generally attempts to
obtain a rigid “lock-and-key” relationship between binder
and ligand, minimizing the amount of flexibility in the
recognition elements.1
The structure of DNA and RNA combines three ele-
ments: nucleobases, sugars, and a backbone consisting
of repeating phosphate diester units. This structure
displays molecular recognition following simple rules (A
pairs with T; G pairs with C) largely without analogy in
other polymeric organic molecular systems.
As common as these facts are, paradoxes remain within
the structure of DNA as a molecular recognition system.
These are especially evident when DNA is compared with
artificial systems being designed by organic chemists to
do “biomimetic” molecule recognition. Three are particu-
larly important.
(b) Molecular recognition in natural DNA involves a
polyanion binding to another polyanion. In contrast, most
artificial molecular recognition systems designed by
organic chemists exploit charge complementarity wher-
ever possible; chemists would normally design a cation,
rather than another anion, to bind to an anion.2
(c) The specificity of molecular recognition in natural
DNA involves hydrogen bonding, and occurs in water.
Water has abundant opportunities for hydrogen bonding.
Indeed, a nonaqueous medium appears to be a prereq-
uisite for most artificial systems using hydrogen bonding
as a key element of molecular recognition.3
(a) DNA is a flexible molecule. The backbone contains
five single bonds per nucleotide capable of largely free
rotation. By contrast, the design by chemists of artificial
(1) Konig B. J . Prakt. Chem. 1995, 337, 339-346.
(2) (a) Zepik, H.; Benner, S. A. J . Org. Chem. 1999, 64, 8080-8083
(b) Paris T.; Vigneron, J . P.; Lehn J . M.; Cesario, M.; Guilhem, J .;
Pascard, C. J . Inclusion Phenom. Macrocyclic Chem. 1999, 33, 191-
202.
(3) See, for example: Salameh, A. S.; Ghaddar, T.; Isied, S. S. J .
Phys. Org. Chem. 1999, 12, 247-254.
* To whom correspondence should be addressed. Phone: (352) 392
7773. Fax: (352) 392 7918.
† Dedicated to Frank H. Westheimer on the occasion of his 90th
birthday.
‡ Brooklyn College.
§ University of Florida.
10.1021/jo0003910 CCC: $22.00 © 2002 American Chemical Society
Published on Web 05/17/2002

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 3997
The “first-generation” model of duplex structure in
DNA was proposed by Watson and Crick four decades
ago.4 In this model, the nucleobases play the central role
in molecular recognition, with both size complementarity
(big purines pair with small pyrimidines) and hydrogen-
bonding complementarity underlying the rule-based mo-
lecular recognition properties that are at the core of
genetics. In the first-generation model, the sugar and
phosphate linkers play secondary roles, serving simply
to hold the nucleobases together, perhaps offering as well
some “preorganization” of the molecular recognition
units.
The first-generation model for DNA structure has
performed well in molecular biology, generating reliable
semiquantitative predictions concerning the behavior of
standard DNA and RNA sequences.5 The model has even
served well to support an increase in the number of
letters in the genetic alphabet, artificially designed to
maintain a Watson-Crick geometry and Watson-Crick-
type hydrogen-bonding pattern.6
The first-generation model has proven surprisingly
inadequate, however, to guide the redesign of the part
of the nucleic acid structure that is proposed to be the
least important to the molecular recognition phenom-
enon: the backbone. Several hundred attempts have now
been made to alter the structure of the backbone in an
attempt to obtain backbone-modified DNA and RNA
analogues that retain the rule-based molecular recogni-
tion displayed by standard nucleic acids.7a These include
DNA analogues where the sugars are linked by units that
include the following groups: methylphosphonate,7b
methylene(methylimino),7c thioformacetal,7d amide,7e
sulfonamide,7f sulfonate,7g ester,7h carbamate,7i 1,3-
propanediol,7j diisopropylsilyl,7k alkene,7l alkyne,7m morpho-
lino,7n and others.
The expectation, given the peripheral role assigned to
the backbone by the first-generation theory, is that such
analogues should display molecular recognition proper-
ties similar to those found in natural DNA. Indeed, some
biotechnology companies were founded on an expectation
that such analogues should be preparable with improved
suitability for practical applications (as an antisense
therapeutic, for example) without losing the predictable
recognition properties displayed by DNA and RNA.
Surprisingly in this context, the majority of the back-
bone-modified nucleic acids damage the rule-based mo-
lecular recognition;7,8 some destroy it entirely. Some,
unpredictably, alter the nature of the rules.9 Most
remarkably, a few enhance DNA-type molecular recogni-
tion phenomena in scaffolding that are quite different
from those in natural scaffolding. The best known of these
are polyamide-linked nucleic acid (PNA) analogues,10
which display enhanced, rule-based molecular recogni-
tion, even though they have little structural resemblance
to natural DNA.
Some time ago, we noted from the first experiments of
backbone-modified nucleic acid analogues11 that the
nature of the backbone was more important to the
molecular recognition properties of DNA and RNA than
the first-generation model suggested. Together, these
results suggest the need for a “second-generation” model
for nucleic acid structure. Elaborating such a model will
require the synthesis of many backbone-modified DNA
and RNA analogues, coupled with full thermodynamic
and structural studies of their behavior.
The first nonionic analogues of RNA (as opposed to
DNA) to be prepared involved the replacement of the
phosphate diester linker by dimethylene sulfide (-CH2-S-
CH2-), sulfoxide (-CH2-SO-CH2-), and sulfone (-CH2-
-
SO2-CH2-) linkages.12 In the sulfone-linked RNA ana-
logues, the replacement is approximately isosteric and
isoelectronic. Short oligonucleotide analogues with a
dimethylene sulfone linkage proved to be soluble in
aqueous and organic solvents, stable to hydrolytic cleav-
age, and resistant to nuclease digestion.13 They even
displayed Watson-Crick-based pairing in their dinucle-
otide version, a fact confirmed by X-ray crystallography
of a self-complementary sulfone dimer (Gso2C).14 Wat-
son-Crick pairing was lost in longer oligosulfone ana-
logues, however, which had an increasing tendency to
self-associate, fold, and display other behaviors not
characteristic of nucleic acids, including a propensity to
catalyze reactions.12
Similar behavior has been suggested for PNA, the
other nonionic nucleic acid analogue shown to retain
Watson-Crick molecular recognition. PNA displays rule-
based molecular recognition up to a certain length,
provided that the sequences are relatively poor in gua-
nine. This has proven to be useful in a variety of
applications. Longer PNA molecules, however, suffer
aggregation and other physical behaviors that interfere
with their ability to recognize complementary DNA.15
We report here the synthesis of building blocks for
S-DNA with sulfide, sulfoxide, and sulfone linkages of
(4) (a) Watson, J . D.; Crick, F. H. Nature 1953, 171, 737-738. (b)
Watson, J . D.; Crick, F. H. Nature 1953, 171, 964-967.
(5) SantaLucia, J . Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 1460-
1465.
(6) (a) Switzer, C. Y.; Moroney, S. E.; Benner, S. A. J . Am. Chem.
Soc. 1989, 111, 8322-8323. (b) Piccirilli, J . A.; Krauch, T.; Moroney,
S. E.; Benner, S. A. Nature 1990, 343, 33-37.
(8) Epple, C.; Leumann, C. Chem. Biol. 1998, 5, 209-216.
(9) Groebke, K.; Hunziker, J .; Fraser, W.; Peng, L.; Diederichsen,
U.; Zimmermann, K.; Holzner, A.; Leumann, C.; Eschenmoser, A. Helv.
Chim. Acta 1998, 81, 375-474.
(7) (a) Freier, S. M.; Altmann, K. H. Nucleic Acids Res. 1997, 25,
4429-4443. (b) Miller, P. S.; McFarland, K. B.; J ayaraman, K.; Ts’o,
P. O. P. Biochemistry 1981, 20, 1878-1880. (c) Morvan, F.; Sanghvi,
Y. S.; Perbost, M.; Vasseur, J . J .; Bellon, L. J . Am. Chem. Soc. 1996,
118, 255-256. (d) Rice, J . S.; Gao, X. Biochemistry 1997, 36, 399-
411. (e) De Mesmaeker, E. B.; Waldner, A.; Fritsch, V.; Lebreton, J .;
Wolf, R. M. Bull. Soc. Chim. Belg. 1994, 4, 1071-1076. (f) Perrin, K.
A.; Huang, J .; McElroy, E. B.; Iams, K. P.; Widlanksi, T. S. J . Am.
Chem. Soc. 1994, 116, 7427-7428. (g) Huang, J .; McElroy, E. B.;
Wildlanksi, T. S. J . Org. Chem. 1994, 59, 3520-3521. (h) J ones, A. S.;
MacCoss, M.; Walker, R. T. Biochem. Biophys. Acta 1973, 294, 365-
377 (i) Mungall, W. S. Kaiser, J . K. J . Org. Chem. 1977, 42, 703-706.
(j) Seela, F.; Kaiser, K. Nucleic Acids Res. 1987, 15, 3113-3129. (k)
Cormier, J . F.; Ogilvie, K. K. Nucleic Acids Res. 1988, 16, 4583-4594.
(l) Wendeborn, S.; Wolf, R. M.; De Mesmaeker, A. Tetrahedron Lett.
1995, 36, 6879-6882. (m) Wendeborn, S.; J ouanno, C.; Wolf, R. M.;
De Mesmaeker, A. Tetrahedron Lett. 1996, 37, 5511-5514. (n) Sum-
merton, J .; Stein, D.; Huang, S. B.; Matthews, P.; Weller, D.; Partridge,
M. Antisense Nucleic Acid Drug Dev. 1997, 7, 63-70.
(10) (a) Egholm, M.; Buchardt, O.; Nielsen, P. E.; Berg, R. H. J . Am.
Chem. Soc. 1992, 114, 1895-1897. (b) Knudsen, H.; Nielsen, P. E.
Nucleic Acids Res. 1996, 24, 494-500.
(11) Schneider, K. C.; Benner, S. A. J . Am. Chem. Soc. 1990, 112,
453-455.
(12) Richert, C.; Roughton, A. L.; Benner, S. A. J . Am. Chem. Soc.
1996, 118, 4518-4531.
(13) (a) Huang, Z.; Schneider, K. C.; Benner, S. A. J . Org. Chem.
1991, 56, 3869-3882. (b) Huang, Z. Dissertation No. 10429, ETH
Zurich, Switzerland, 1993. (c) Hyrup, B.; Richert, C.; Schulte-Her-
bruggen T.; Benner, S. A.; Egli, M. Nucleic Acids Res. 1995, 23, 2427-
2433.
(14) Roughton, A. L.; Portmann, S.; Benner, S. A.; Egli, M. J . Am.
Chem. Soc. 1995, 117, 7249-7250.
(15) (a) Bergmann, F.; Bannwarth, W.; Tam, S. Tetrahedron Lett.
1995, 36, 6823-6826. (b) Gildea, B. D.; Casey, S.; MacNeill, J .; Perry-
O’Keefe, H.; Sorensen, D.; Coull, J . M. Tetrahedron Lett. 1998, 39,
7255-7258.

3998 J . Org. Chem., Vol. 67, No. 12, 2002
Huang and Benner
Sch em e 1
two, four, six, and eight units. We also present some
physiochemical and biological studies of the sulfone
analogues that shed light on the features to be expected
from nonionic DNA analogues.
the N-9 isomer, affording the N-9 isomer 10a in 67% yield
after 2 days. This suggests that the N-9 isomer is
thermodynamically more stable than the N-7 isomer in
this system, as it is in many others.17 The amount of
silylating reagent MSTFA also influenced the relative
amounts of various isomers. Less MSTFA gave a higher
proportion of the N-9 isomer. Optimized, the conditions
for incorporating N6-benzoyladenine required base (2.0
equiv to 4), MSTFA (2.0 equiv), 4 (0.1 M in acetonitrile
solution), and TMSOTf (1.0 equiv), incubating for 2 days
at 60 °C. The unprotected A derivative 11 was prepared
by transient trimethylsilyl protection of adenine.18 The
optimized conditions required adenine (2.0 equiv to 4),
MSTFA (2.0 equiv), 4 (0.1 M in acetonitrile solution), and
TMSOTf (2.5 equiv). The mixture was incubated for 10
min at 60 °C, giving 11 in ca. 70% yield.
Syn th esis
Mon om er s. Two routes starting from diacetone-D-
glucose 1 (Scheme 1) were developed to prepare the
building blocks for S-DNA.13a,b Both shared a common
precursor, the allofuranose derivative 2, which was
prepared from 1 in nine steps in 28% overall yield.
Because hydrophobicity was believed to encourage ag-
gregation of the S-DNA oligomers, the less hydrophobic
uridine nucleobase generally replaced thymidine; other-
wise, the standard DNA nucleobases were used.
In the first route, nucleoside analogues 5-12 were
prepared by modified Vorbru¨ggen conditions from inter-
mediate dioxabicyclo[3.2.1]octane 4, prepared from 2 by
cyclization under acidic conditions and acylation to yield
the 2′-m-trifluoromethylbenzoyl derivative.13a The ben-
zoyl group controlled the stereochemistry of heterocycle
introduction, and was reductively removed by photolysis
in a subsequent step.
The reaction of N2-isobutyrylguanine with 4 in the
presence of SnCl4 was low-yielding, and gave more of the
N-9 isomer 12a than the N-7 isomer 12b. The N-7 isomer
predominated at 60 °C, even after prolonged incubation.
To convert the N-7 isomer to the N-9 isomer, the isolated
N-7 isomer was heated at 60 °C in the presence of
MSTFA (4 equiv) and TMSOTf (1.5 equiv). Little change
was observed after 24 h, implying that the N-7 isomer is
thermodynamically preferred in this case. Therefore, the
guanosine derivatives 12a and 12b were prepared with
MSTFA (6 equiv) and TMSOTf (1.0 equiv) at rt in 65%
yield (N-9:N-7 ) ca. 1:1), and separated by column
chromatography. Structures of the N-9 and N-7 isomers
of the adenosine and guanosine analogues were assigned
by comparing the 1H and 13C NMR spectra of natural
adenosine19 and guanosine17c,19b,c and their N-7 isomers
with the spectra of completely deprotected N-9 and N-7
isomers following NaOH hydrolysis (Table 1).
The standard Vorbru¨ggen procedures16 were adapted
to attach pyrimidine and purine bases to 4 (Scheme 1).
N-Methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) in
acetonitrile was used as a silylating reagent. For intro-
ducing pyrimidines, SnCl4 was found to be the best
catalyst. In optimized procedures, U, T, and C (with or
without protection for the exocyclic amino group) (2.0
equiv to 4) were treated with MSTFA (6.0 equiv), 4 (0.1
M in acetonitrile solution), and SnCl4 (1.5 equiv). The
reactions were complete in 2 h (>90%). The unprotected
C derivative 9 was prepared (70% yield) by transient
protection of cytosine as the trimethylsilyl derivative.
Analysis by NMR showed that the â-anomers of 5-12
were formed exclusively.
Both SnCl4 and trimethylsilyl triflate (TMSOTf) were
used as catalysts to append N6-benzoyladenine to 4. At
room temperature (rt) with TMSOTf, slightly more of the
desired N-9 isomer 10a was formed than N-7 isomer 10b.
With SnCl4, more of the undesired N-7 isomer was
formed. The N-9 and N-7 isomers were separable by silica
gel chromatography. At high temperature (60 °C) in the
presence of TMSOTf, the N-7 isomer was converted to
The 2′-benzoyloxy group of uridine analogue 5 was
directly removed by photolysis using the procedure of
(17) (a) Shimizu, B.; Miyaki, M. Chem. Pharm. Bull. 1970, 18, 570-
580, 732-740. (b) Miyaki, M.; Saito, A.; Shimizu, B. Chem. Pharm.
Bull. 1970, 18, 1446-1456, 2459-2468. (c) Garner, P.; Ramakanth,
S. J . Org. Chem. 1988, 53, 1294-1298. (d) Visser, G. M.; Keemink, R.;
Schattenkerk, C.; Kraal, B.; von Boom, J . H. Nucleosides Nucleotides
1984, 3, 277-286.
(18) (a) Ti, G. S.; Gaffney, B. L.; J ones, R. A. J . Am. Chem. Soc.
1982, 104, 1316-1319. (b) McGee, D. P. C.; Martin, J . C.; Webb, A. S.
Synthesis 1983, 540-541. (c) Bhat, V.; Ugarkar, B. G.; Sayeed, V. A.;
Grimm, K.; Kosora, N.; Domenico, P. A.; Stocker, E. Nucleosides
Nucleotides 1989, 8, 179-183.
(19) (a) Chenon, M.-T.; Pugmire, R. J .; Grant, D. M.; Panzica, R. P.;
Townsend, L. B. J . Am. Chem. Soc. 1975, 97, 4627-4636. (b) Ogilvie,
K. K.; Nguyen-Ba N.; Hamilton, R. G. Can. J . Chem. 1984, 62, 1622-
1627. (c) Kjellberg, J .; J ohansson, N. G. Tetrahedron 1986, 42, 6541-
6544.
(16) (a) Niedballa, U.; Vorbru¨ggen, H. J . Org. Chem. 1974, 39, 3654-
3673. (b) Vorbru¨ggen, H.; Bennua, B. Chem. Ber. 1981, 114, 12234-
1255, 1279-1286.

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 3999
Ta ble 1. Com p a r ison of Selected 13C NMR Da ta of
Ad en osin e a n d Gu a n osin e a n d Th eir An a logu es
in THF in the presence of 1,8-diazabicyclo[5.4.0]undec-
7-ene (DBU); no cyclization was observed. With this
problem not evident, using the 6′-position as the electro-
philic center in the coupling reaction was preferred given
the well-known generalization that SN2 reactions are
slowed by substitution at the position adjacent to the
electrophilic center. Thus, carbon 6′ is less hindered as
an electrophilic center than the 3′′-carbon.
Both approaches were examined experimentally. Cou-
pling between a 3′′-thiol and a 6′-mesylate was found to
be approximately 10-fold faster than coupling between a
6′-thiol and a 3′′-mesylate. The building blocks with a
3′′-thiol therefore became the focus for an optimized
synthetic strategy. The 6′-hydroxyl group of deoxyuridine
13 was protected in quantitative yield by treatment with
4,4′-dimethoxytrityl chloride [(DMTr)Cl] in triethylamine
(TEA)/THF to yield 15. After hydrolysis of the 3′′-
benzoate ester yielded alcohol 17 (MeOH/aq NaOH), a
thioacetate group was introduced by a Mitsunobu reac-
tion with PPh3, diisopropyl azodicarboxylate (DIAD), and
HSAc in THF to give protected building block 19.23 The
6′-DMTr group was found to be stable under Mitsunobu
conditions. The analogous procedure generated the 2′-
deoxygenated building block containing adenine (20) after
reprotection of the 6-amino group on adenine following
hydrolysis.24
C2
C4
C5
C6
C8
C2′
adenosine19a
152.6 149.2 119.5 156.3 140.2 88.2
(N7)-adenosine19a 152.8 160.7 110.2 151.7 144.6 89.4
10a a
152.4 148.6 120.2 156.5 139.5 90.1
152.3 161.6 110.8 150.2 144.1 91.8
155.3 151.6 121.8 158.6 142.7 91.5
153.7 151.4 116.7 156.9 135.7 86.4
160.7 107.8 154.5 142.5 89.2
154.0 150.6 115.8 156.3 134.8 89.3
152.3 160.5 108.8 155.8 140.8 92.6
10ba
11
guanosine17c
(N7)-guanosine17c 153
12a a
12ba
a
These molecules were in their fully deprotected forms and
purified by HPLC. All signals are reported relative to that of
Me4Si.
Saito et al.,20 to give the 2′-deoxyuridine analogue 13 in
85% yield (Scheme 2). Attempts to photodeoxygenate the
benzoylcytidine analogue 7 and benzoyladenosine ana-
logue 10a failed. Instead, compounds 9 and 11, respec-
tively, having lost only the benzoyl protecting groups,
were isolated in ca. 60% yield. This observation was
interesting. An initial hypothesis proposed that the
benzoyl protecting groups on the bases yielded a new
chromophore that generated heterocycle-focused photo-
chemistry, leading to the loss of the benzoyl groups upon
irradiation. To test this hypothesis, 8, where the exocyclic
amino group of the cytosine was protected with an
isobutyryl group instead of a benzoyl group, was subject
to photolysis. Again, 9 was isolated in approximately 50%
yield. These results suggest that an amide functionality
to the heterocyclic bases is sufficient to generate hetero-
cycle-focused photochemistry.
To test this, the unprotected adenosine analogue 11
was prepared and irradiated. The 2′-deoxyadenosine
analogue 14 was isolated in 81% yield, corresponding to
reports in the literature with adenosine.20a-c However,
the unprotected cytidine analogue 9, upon photolysis,
gave a complicated mixture under the same conditions.
In the presence of Mg(ClO4)2, which is believed to
enhance electron transfer from N-methylcarbazole (MCZ)
to trifluoromethylbenzoate,21 the 2′-deoxycytidine ana-
logue was isolated in 10% yield. The photodeoxygenation
of guanosine analogue 12a gave only a complex mixture
under a variety of conditions.
To synthesize building blocks for S-DNA, a thiol
function must be introduced. As a sulfur can be placed
at either position 6′ or 3′′ in the units to be coupled (with
activation to create an electrophilic center at the 3′′- or
6′-position, correspondingly), two pairs of coupling units
are conceivable. Several considerations influenced the
choice of coupling units. In natural nucleosides, the
2-oxygen on a cytosine or the 3-nitrogen on adenosine
can displace a 5′-tosylate in an intramolecular cycliza-
tion.22 There was a worry that an electrophilic 6′-position
might analogously react with the 2-oxygen on pyrimidine
during the coupling reaction. In a model reaction, the 6′-
mesylate, prepared by mesylation of 13, was incubated
As the photolytic deoxygenation of the 2′-m-trifluo-
romethylbenzoyl group was not universally applicable in
this synthesis, Barton deoxygenation25 was also examined
to remove the 2′-oxygen from the nucleoside analogues.
The 2′-m-trifluoromethylbenzoyl group of uridine ana-
logue 21, prepared from 5 by treatment with (DMTr)Cl,
was removed in NaHCO3(satd)/MeOH to afford 25 (86%).
These conditions are mild enough to leave the benzoyl
group largely unaffected, as trifluoromethylbenzoate
esters hydrolyze ca. 2 orders of magnitude faster than
benzoate esters.26 The 2′-m-trifluoromethylbenzoyl group
was also removed from the adenosine analogue 22 in this
way to give 26 (91%). Selective hydrolysis of 23 (benzoyl
C) was problematic, however, as a mixture of compounds
with benzoyl groups partially remaining was obtained.
For the Barton reaction, uridine analogue 25 was
converted to thiocarbamide 27 (Scheme 2) (90%, 5 g scale)
by treatment with 1,1′-thiocarbonyldiimidazole (TCD) in
THF for 44 h. Heating a solution of 27 in toluene at reflux
with Bu3SnH in the presence of initiator azobisisobuty-
ronitrile (AIBN) reduced 27 to 15 (identical to the same
compound prepared by the photolysis route by IR and
1H and 13C NMR). Perhaps surprisingly, the 6′-DMTr
group is stable under Barton radical deoxygenation
conditions. A Mitsunobu reaction gave the U-containing
building block 19 as before. The A-containing intermedi-
ate 26 was likewise converted to thiocarbamide 28 (89%)
in toluene/THF (1:19), and 16 was prepared (82%) by the
Barton reaction. Products prepared in both routes were
identical to the same compound prepared by the pho-
1
tolysis route by IR and H and 13C NMR.
(20) (a) Saito, I.; Ito, S.; Shinmura, T.; Matsuura, T. Tetrahedron
Lett. 1980, 21, 2813-2816. (b) Ikehira, H.; Matsuura, T.; Saito, I.
Tetrahedron Lett. 1985, 26, 1743-1746. (c) Saito, I.; Ikehira, H.;
Kasatani, R.; Watanabe, M.; Matsuura, T. J . Am. Chem. Soc. 1986,
108, 3115-3117. (d) Shaw, A. A.; Shetlar, M. D. J . Am. Chem. Soc.
1990, 112, 7736-7742. (e) Skalski, B.; Steer, R. P.; Verrall, R. E. J .
Am. Chem. Soc. 1991, 113, 1756-1762.
(23) Mitsunobu, O. Synthesis 1981, 1-34.
(24) Ti, G. S.; Gaffney, B. L.; J ones, R. A. J . Am. Chem. Soc. 1982,
104, 1316-1319.
(25) (a) Barton, D. H. R.; McCombie, S. W. J . Chem. Soc., Perkin
Trans. 1 1975, 1574-1585. (b) Barton, D. H. R.; Subramanian, R. J .
Chem. Soc., Perkin Trans. 1 1977, 1718-1723. (c) Pankievicz, K.;
Matsuda, A. Watanabe, K. A. J . Org. Chem. 1982, 47, 485-488. (d)
Robins, M. J .; Wilson, J . S.; Hansske, F. J . Am. Chem. Soc. 1983, 105,
4059-4065.
(21) Mizuno, K.; Ichnose, N.; Toshuyi, Y. Chem. Lett. 1985, 455-
458.
(22) Andersen, W.; Hayes, D. H.; Michelson, A. M.; Todd, A. R. J .
Chem. Soc. 1954, 1882-1887.
(26) J ones, R. A. Y. Physical and Mechanistic Organic Chemistry;
Cambridge University Press: Cambridge, 1984; pp 38-68.

4000 J . Org. Chem., Vol. 67, No. 12, 2002
Huang and Benner
Sch em e 2
1
The Barton reaction removes oxygen from the second-
ary carbon, but not the primary carbon, of a cyclic
thiocarbonate, presumably because of the different sta-
bilities of the primary and secondary radicals.27 To
explore this approach, diol 29 was prepared by hydrolyz-
ing both benzoyl groups from 21 at once (Scheme 2).
Synthesis of the 2′,3′′-cyclic thiocarbonate was not suc-
cessful.
tions. Analysis of this compound by IR and H and 13C
NMR revealed that it was identical to 19 prepared by
the photolytic and Barton routes.
To obtain diol 30a , adenosine analogue 22 was treated
with dilute NaOH, yielding diol 30a (52%), along with
fully deprotected 30b (43%). It proved difficult to find
conditions that removed the benzoate esters at the 2′-
and 3′′-positions without disturbing the protection on the
base. A thioacetate unit was then introduced to yield 32,
which was deoxygenated to give building block 20.
This approach based on selective treatment of a diol
followed by deoxygenation generated a more efficient
approach to the building blocks (Scheme 3). In this route,
the 6′-hydroxyl group of 2 was blocked with the tert-
butyldiphenylsilyl (TBDPS) group in quantitative yield
with (TBDPS)Cl/DMF in the presence of imidazole
(Scheme 3).28 After removal of the 3′′-benzoyl group, the
2′- and 3′′-hydroxyl groups were protected as acetate
esters, which were easier to hydrolyze than benzoyl esters
under basic conditions without disturbing the functional-
ity on the nucleobases. The 2′-acetyl group in the key
intermediate 35 also directed the formation of the
â-isomer in the reaction that introduced the nucleobases.
As examples, thymine and N4-benzoylcytosine (each 2
equiv) were reacted with intermediate 35, MSTFA (6.0
equiv), and SnCl4 (1.5 equiv) in acetonitrile to give 36
and 37 (Scheme 3). The two acetyl groups were removed
from the thymine derivative with NaOH during the
workup, giving thymine-containing diol 36 in ca. 90%
yield over two steps. The crude product containing N4-
benzoylcytosine was converted to diol 37 (88% over two
steps) by treatment with NaOH (aq, 5 equiv) in MeOH
(45 s). This removed selectively the acetate groups,
leaving the N-benzoyl group on the base unaffected. The
thymine-containing building block 42 and the cytosine-
containing building block 43 were prepared in ca. 60%
Nevertheless, the availability of diol 29 and the success
of the Mitsunobu reaction for the introduction of the
thioacetate functionality into intermediate 17 encouraged
us to learn whether a primary hydroxyl group might be
replaced by a thioacetate in the presence of a secondary
hydroxyl group using Mitsunobu conditions. Because of
the size of the phosphonium intermediate in the Mit-
sunobu reaction, primary hydroxyl groups are expected
to react more rapidly than secondary hydroxyl groups.
Therefore, assuming that the steps preceding the break-
down of the phosphorane are not in rapid equilibrium,
primary hydroxyl groups are expected to be converted to
thioacetates more rapidly than secondary hydroxyl groups.
Consistent with this hypothesis, diol 29 was transformed
to monothioacetate 31 exclusively (89%) via a modified
Mitsunobu procedure (Scheme 2). This regiospecific
synthesis fills a gap in the synthetic literature, as there
appears to be no other satisfactory method for transform-
ing a primary hydroxyl group to a thioacetate in the
presence of a secondary hydroxyl group.
Incorporation of a thioacetyl group in a single step
avoids the need to protect the 3′′-hydroxyl group during
deoxygenation. Thus, deoxyuridine building block 19 was
prepared from 29 by successive Mitsunobu reaction,
thiocarbamidation (88%), and deoxygenation (86%). It is
interesting that both the thioacetate and DMTr ether
moieties are stable under radical deoxygenation condi-
(27) (a) Fox, J . J .; Wempen, I. Tetrahedron Lett. 1966, 7, 643-646.
(b) Mubarak, A. M.; Brown, D. M. Tetrahedron Lett. 1981, 22, 683-
684. (c) Hartwig, W. Tetrahedron 1983, 39, 2609-2611.
(28) Fieser & Fieser. Reagents for Organic Synthesis; Vol. 6, p 81.

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 4001
Sch em e 3
Sch em e 4
yield (over three steps) from diols 36 and 37 using the
thioacylation-deoxygenation sequence described previ-
ously. Both adenine-containing and guanine-containing
building blocks were also prepared by an analogous route
on a large scale.
coupling reaction via mesylation with MsCl in THF/TEA,
giving 47 in quantitative yield (Scheme 4). When the
trifluoromethanesulfonyl ester was used as the leaving
group, more rapid coupling was achieved.30 However, the
trifluoromesylate intermediate proved to be difficult to
purify. Activation of the hydroxyl group by tosylation31
was also examined, but the tosylate coupling reaction was
slow. Several solvents (THF, dioxane, CH3CN, DMF, and
MeOH) were investigated. THF was ultimately chosen
because it was easy to degas, gave smooth and clean
reactions, and made the workup easy. DBU deprotonated
the thiol of the monomer to yield a salt that is soluble in
THF, and was found to be the best base for the coupling
reaction.
During the coupling reaction, a disulfide was formed
as a side product due to oxidation of the thioate by
oxygen. To compensate for the loss of the thioate, excess
thiol was used; the optimal ratio of thiol and mesyl units
was 1.2:1. The thioether product and the disulfide of the
building blocks often had similar mobilities on thin-layer
chromatography (TLC), making the two difficult to
separate. Therefore, crude products were routinely treated
with 1,4-dithioerythritol (DTE) after the reaction to
reduce the disulfide to the thiol and thus facilitate the
isolation of the pure thioether. DTE was generally
removed by washing with water, since most of the
Oligom er s. Sulfide-linked oligonucleotide analogues
were assembled in the solution phase by nucleophilic
substitution (SN2) under basic conditions (Scheme 4),
with a thioate anion on one unit serving as the nucleo-
phile to displace a leaving group (e.g., MsO-, TsO-, or
Br-) on the second unit. The sulfone oligonucleotide
analogues were synthesized by oxidizing the sulfide
analogues. A convergent approach to synthesize the
sulfide oligonucleotide analogues was employed, which
gave a higher overall yield because it involved fewer
steps. A protection strategy was developed for both the
6′-hydroxyl group and the 3′′-thiol group using DMTr
groups, which are stable under coupling conditions, but
which are selectively removable (80% HOAc and AgNO3,
respectively)29 in quantitative yield under conditions that
do not cause the loss of protecting groups of the bases,
depurination, or cleavage of the backbone. The DMTr
group also proved to diminish aggregation of the longer
nucleotide analogues during the course of the convergent
synthesis (see below).
The 6′-hydroxyl group of 46, prepared from 19 by
protecting the 3′′-thiol group as the DMTr thioether and
removing the 6′-DMTr group, was activated for the
(30) Blanc-Muesser, M.; Defaye, J . Driguez, H. J . Chem. Soc., Perkin
Trans. 1 1982, 15-18.
(31) Bandara, B. M. R.; Bernofsky, C. Nucleosides Nucleotides 1990,
9, 189-195.
(29) Huang, Z.; Benner, S. A. Synlett 1993, 1, 83-84.

4002 J . Org. Chem., Vol. 67, No. 12, 2002
Huang and Benner
Sch em e 5
Sch em e 6
oligonucleotide analogues with bridging thioethers are
insoluble in that solvent. In optimized conditions, thiol
44 (1.2 equiv) and mesylate 47 reacted in the presence
of DBU (2 equiv) in THF under Ar to give 51 in 96% yield.
The thioate can also be generated in situ from thiol
acetate 19 in the presence of 10% MeOH; the reaction
yield was also high.
days of incubation. Sulfide dimer 60 was also directly
oxidized to dimeric sulfone 66 with Oxone(aq) in EtOH/
EtOAc (ca. 3:1) in high yield. The DMTr group of 66 was
retained to facilitate workup, and removed at the final
step to yield 67.
The DMTr group on the sulfur of dimer 51 was
oxidatively removed when the dimer was treated with
Oxone (2KHSO5-KHSO4-K2SO4); the product was a
sulfonate.32 When a sulfonate is not desired, the DMTr
groups are replaced by a blocking group that is stable
under oxidative conditions. A simple alkyl group was
found to be satisfactory for this purpose (Scheme 5). Thiol
57 was prepared in ca. 90% yield by treating 51 with
AgNO3 in H2O/MeOH followed by DTE. 57 was then
alkylated with either benzyl bromide in TEA/THF or
ethyl methanesulfonate in DBU/THF to give dimer 58
or 60, respectively. NMR spectra suggested that a second
ethyl group was incorporated into the product in the
second procedure, perhaps by ethylation of a heterocyclic
base. In contrast, no dibenzylated products were ob-
served, and the reaction yield was essentially quantita-
tive. Later in the convergent synthesis, tetramer and
octamer thiols were benzylated in TEA/THF, and no
multiply benzylated products were observed.
Sulfides may be oxidized to sulfone-linked oligonucle-
otide analogues directly with a strong oxidant (e.g.,
Oxone).33 Evidence emerged that such harsh treatment
caused degradation. It was reported that sulfones could
be obtained from sulfides by treatment with H2O2 in the
presence of a titanium silicate catalyst (TS-2).34 In a
model reaction, diethyl sulfide was converted to diethyl
sulfone in methanol using H2O2 in the presence of TS-2.
The catalyst increased the rate of the reaction ca. 300-
fold at rt. In the preferred procedure, dimer sulfide 59
was first oxidized to sulfoxide dimer 64 in MeOH with
H2O2, and then further oxidized to sulfone dimer 65 in
the presence of TS-2 (Scheme 5). The reaction was
monitored by HPLC, which showed that the sulfide was
converted first to diastereomeric sulfoxides, then to
sulfide-sulfone chimeras, and then to the full sulfone.
Sulfone dimer 65 was isolated in 91% yield after 2-3
Dimers 52, 55, 59, and 61 were prepared by removing
the DMTr groups from the oxygen atoms of 51, 54, 58,
and 60 with 80% HOAc. Mesylates 53, 56, and 62 were
prepared in 69%, 87%, and 90% yield, respectively, by
treating 52, 55, and 61 with MsCl in TEA/THF. When
pyridine was used instead of TEA, the product mixture
contained ca. 10% of a species containing two mesyl
groups (as judged by NMR), perhaps resulting from
mesylation on the pyrimidine. After being incubated with
Na2CO3(aq) in MeOH for several hours, the doubly
mesylated species were evidently converted to the mo-
nomesylated dimer. Tetramers 68 and 73 were prepared
(Scheme 6) by coupling dimer thiol 57 (1.2 equiv) and
dimer mesylate 53 or 56 in the presence of DBU (2 equiv)
in THF under Ar for 18 h (>90% yield). Since the
tetramer nucleotide analogue and the disulfide have
nearly the same mobility on silica gel TLC, the crude
product was treated with DTE to reduce the disulfide.
The coupling yield of the two dimer units was almost as
high as the coupling yield (96%) of the two monomer
units. The two lipophilic DMTr groups on the tetramer
facilitated the workup.
Tetramer 68 or 73 was treated with silver nitrate to
remove the DMTr group from the sulfur, yielding 71 or
75 in ca. 70% yield. The (DMTr)OMe and (DMTr)OH
formed in the reaction were removed by washing the
pellet of the silver thioate salt (obtained by centrifuga-
tion) with MeOH/THF. Various amounts of silver nitrate
(5, 10, and 20 equiv) were examined. With the largest
excess of silver nitrate, some tetramer was lost by
coprecipitation with the silver hydroxide formed by
hydrolysis of silver nitrate in the basic aqueous solution.
With the smallest excess of silver nitrate (5 equiv),
deprotection was not complete. A 10 equiv sample of
AgNO3 proved to be optimal. Under these conditions,
5-10% protected starting material was still found. The
diprotected tetramer influenced neither the coupling
reaction nor the benzylation. DTE was removed again
by washing with water.
(32) (a) Capozzi, G.; Modena, G. The Chemistry of the Thiol Group;
J ohn Wiley & Sons: London, New York, Sydney, and Toronto, 1974;
pp 785-839. (b) Gilbert, E. E. Sulfonation and Related Phenomena;
New York, 1965; pp 217-239.
(33) Trost, B. M.; Braslau, R. J . Org. Chem. 1988, 53, 532-536.
(34) (a) Reddy, J . S.; Sivasanker, S.; Ratnasamy, P. J . Mol. Catal.
1991, 69, 383-392. (b) Reddy, R. S.; Reddy, J . S.; Kumar, R.; Kumar,
P. J . Chem. Soc., Chem. Commun. 1992, 84-85.

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 4003
Sch em e 7
The product thiols 71 and 75 were alkylated with BnBr
in TEA/THF, and the product was treated with 80%
HOAc to afford 72 and 76, respectively. Tetramers 68
and 73 were treated with 80% HOAc and then mesylated
to give 70 and 74 in more than 80% yield, respectively.
In both cases, multiple mesylation products were ob-
served. Hexamer 81 was prepared in 20% yield by
coupling dimer mesylate 62 with thiol tetramer 71 in
THF (Scheme 6). The low yield probably resulted from
the aggregation and low solubility of the educts (see
below).
Tetramers 72 and 76 were oxidized to a diastereomeric
mixture of sulfoxides (represented as a single structure,
77 or 78) with H2O2 in THF/MeOH (1:1). While such
mixtures are normally avoided in synthesis, in this case,
the mixture appeared to have improved solubility in
water, where subsequent oxidation with stronger oxidiz-
ing reagents and catalysts was used to generate sulfone
links. THF was used to increase the solubility of 72 or
76 in the reaction mixture. To avoid the formation of
peroxides, the reaction was allowed to run for only
several minutes. After the solvents were evaporated
under reduced pressure, sulfoxide 77 or 78 was redis-
solved in MeOH, and the oxidation continued in the
presence of TS-2. UV spectroscopy showed that the
concentration of the nucleotide analogues in solution
decreased after the catalyst was added, suggesting that
the tetramers were adsorbed onto the solid phase. The
incubation was continued for 6 days. The catalyst was
then extracted with CH3CN/H2O (1:1) to yield a solution
of 79 or 80 (after removal of the benzoyl protecting
group), which was recovered in 30-40% overall yield by
HPLC (CN column, 23% CH3CN/H2O as eluent). Faster
moving compounds (structures not assigned) were also
recovered. Because the longer oligonucleotides adhere
strongly to this catalyst, the hexameric and octameric
sulfone-linked analogues must be prepared with Oxone.
Hexameric sulfide 81 was treated with 80% HOAc and
oxidized first with H2O2 to the sulfoxides, and then to
82 with Oxone(aq) at neutral pH (Scheme 6).
The increasing difficulty associated with handling
longer sulfone-linked oligonucleotide analogues became
a theme for the synthesis of larger molecules, at the levels
of both strategy and tactics. For example, the thioate salt
of tetramer 71 did not dissolve well in pure THF. To form
octamers by coupling two tetrameric units (Scheme 7),
MeOH was therefore used to increase the solubility of
the reactants. Mesylate 70 and thiol 71 or mesylate 74
and 75 were incubated in 10% MeOH/THF containing
DBU to yield 83 or 87 (50-70% yield). Thiol 84, obtained
by treating octamer 83 with AgNO3 (32% yield), was
treated with BnBr in TEA/MeOH/THF and with 80%
HOAc to yield 85. Both 83 and 87 were treated with 80%
HOAc to give 86 and 88, respectively.
Ta ble 2. Com p a r ison of NMR Sign a ls of Dim er ,
Tetr a m er , Hexa m er , a n d Octa m er Su lfid es
sulfide CH3O H-5 H-2 m-ar-H C4′′ C5′ C2′ C5 C6
51
52
57
63
68
81
83
12H
6H
6H
0H
12H
6H
1 + 1
1 + 1
1 + 1
1 + 1
1 + 3
1 + 5
1 + 7
2
2
2
2
4
6
7
8
4
4
0
8
4
8
2× 2× 2× 2× 2×
2× 2× 2× 2× 2×
2× 2× 2× 2× 2×
2× 2× 2× 2× 2×
4× 4× 4× 4× 4×
12H
7) of Oxone/KHCO3 in CH3CN/H2O (1:1) to yield the
sulfone-sulfonate octamer 90 or 91 (after removal of the
benzoyl groups by KOH treatment). This oxidation was
followed by HPLC (CN column, 0-20% CH3CN/H2O as
eluent), which showed that significant amounts of sec-
ondary products eluting faster than the sulfonate were
also observed, and increased with longer incubation
times, implying that they arose from overoxidation. After
10 min, the reaction was quenched with H2O2, and
octamer sulfonate 90 or 91 was recovered by HPLC in
20-30% yield.
Ch a r a cter iza tion
The oligonucleotide analogues were characterized by
1
IR, UV, MS, 400 MHz H NMR, and 100 MHz 13C NMR.
1
The H and 13C NMR signals in the spectra of dimers,
tetramers, and octamers were assigned by systematically
comparing the spectra with those of the monomers and
all intermediates. Since each DMTr group gives a signal
at 6.8 ppm integrating to 4 protons, this was used as an
internal standard for the integration. The integral of H-2′
and H-5′ protons was a useful indicator of the number of
building units in each oligonucleotide analogue (Table 2).
The signal from a single H-5 proton was shifted upfield
Attempts to oxidize sulfide-linked octamers to sulfones
using the TS-2/H2O2/MeOH system failed to yield recov-
erable products, presumably because of adsorption of the
octamers on the TS-2 catalyst. Sulfide 85 was therefore
oxidized first to the corresponding sulfoxide with H2O2,
which was then oxidized with Oxone/KHCO3 at ca. pH 7
in CH3CN/H2O (1:1). After the reaction was quenched
with H2O2, sulfone 89 was purified by HPLC (C18
column, 40% CH3CN/H2O as eluent). A sulfonate group
was introduced at the 3′′-terminal of the sulfone octamer
by direct oxidation with Oxone. Sulfide octamer 86 or 88
in CH3CN/H2O (1:1) was treated with a solution (pH ca.
1
relative to the other(s) in the H NMR spectra of dimer
51 (1:1 ratio), tetramer 68 (1:3), hexamer 81 (1:5), and
octamer 83 (1:7). We speculated that this H-5 proton is
likely to be the one on the 6′-end unit, because of
shielding by the 6′-DMTr group. In their 13C NMR
spectra, groups of signals with nearly the same intensity
were also observed, arising from groups of similar
carbons, for example, dimer 51 with two lines in a group,

4004 J . Org. Chem., Vol. 67, No. 12, 2002
Huang and Benner
Ta ble 3. Solven t Effect on th e Aggr ega tion of
and tetramer 68 with four lines in a group. In the FAB-
MS spectra of dimer 63 and tetramer 72, m/e 549 ([M +
23]+) and 1125 (M+) were observed, respectively. The
structure of the sulfide octamer was also confirmed by
TOF-MS data obtained from the two sulfone octamers.
Tetr a m er sa
tetramer retention time (min) while
eluted with MeOH (%) in EtOAc
sample in solvent system
2%
3%
4%
5%
Upon oxidation to the sulfone, the chemical shift of CH2
of the 3′′-benzyl group was shifted from 3.6 to 4.4 ppm.
Characteristic IR signals, 1314 cm-1 (OdSdO, asym) and
1137 cm-1 (OdSdO, sym), were recorded. Each group of
13C signals was composed of lines with nearly the same
intensity, each contributed by one unit in the oligonucle-
otide analogues. In the FAB-MS spectrum of dimer 65,
m/e 703 ([M + 23]+) and 681 ([M + 1]+) were observed.
TOF-MS of both tetrameric sulfones 79 and 80 showed
masses of 1254, which corresponds to the molecular mass
(M+ ) 1252, without isotopic substitution), and 1408
(1254 + mass of the matrix [154]). Interestingly, the
tetrameric sulfones adhered to both plastic filters and
plastic syringes. TOF-MS analysis was consistent with
the proposed structure for 89 (6′-HO-USO2USO2USO2-
USO2USO2USO2USO2USO2-Bn-3′′), where the major
peak had a mass of 2411, corresponding to the molecular
mass (M+ ) 2390, without isotopic substitution) plus a
molecule of water. The TOF-MS of 90 (6′-HO-USO2-
USO2USO2USO2USO2USO2USO2U-SO3K-3′′), in its acid
form, was consistent with the proposed structure. The
major peak had a mass of 2340, corresponding to the
molecular mass (M+ ) 2322, without isotopic substitu-
tion) plus a molecule of water. The oligonucleotide
analogues containing cytosine were also proved to be
68 in THF
71 in THF
71 in 20% MeOH/THF
70 in THF
70 in 30% MeOH/THF
4.3, 6.2
3.1
8.0
16
6.1, 8.0
14
5.7, 7.3
5.7
5.7
4.1
4.1
7.5
a
Analytical column: Econosil silica, 10 m, 250 × 4.6 mm, eluent
2-5% MeOH/EtOAc, fluid rate 2 mL/min, pump pressure 50 atm,
wavelength 260 nm, each injection 25 mL, dead time of the column
1.5 min.
forms in equilibrium, where the rate of interconversion
of the forms was slow enough to allow their separation
by HPLC.
(3) The relative areas of the two peaks was a function
of the concentration of tetrameric species. The ratio
depends on the concentration squared, indicating that
the slower running substance is most likely a duplex of
the faster running substance. The disassociation con-
stants for the duplex-monoplex equilibrium were mea-
sured in several solvent mixtures (Table 4).
(4) The relative amounts of the two peaks varied
according to the proportion of MeOH in the solution. The
higher the proportion of MeOH, the higher the proportion
of the faster running peak, consistent with the assign-
ment of this peak to the monoplex. When THF was used
as the solvent, the duplex predominated (Table 3). When
THF/MeOH containing a high proportion of MeOH was
used as solvent, the monoplex predominated.
(5) Varying the proportion of MeOH in the eluent
(MeOH/EtOAc) influences only the retention time, not
the ratio of the two peaks, suggesting that a low percent-
age of MeOH (2-5%) in EtOAc does not strongly affect
the duplex-monoplex equilibrium.
1
satisfactory by H and 13C NMR and MS. The TOF-MS
of 91 (6′-HO-USO2USO2USO2CSO2USO2USO2USO2C-
SO3K-3′′) showed a major peak at m/e 2342.
P h ysioch em ica l a n d Biologica l P r op er ties of
S-DNA
(6) The 6′-DMTr and 3′′-DMTr diprotected tetramers
aggregate less. This is consistent with what was observed
in the synthesis of the octameric sulfide nucleotides,
when different thiol protection groups were used. The
mesylate coupling unit containing a 3′′-thiol with DMTr
protection gave a higher yield than that with benzyl
protection. In both THF and a MeOH/THF mixture, the
Kdiss for variously protected tetramers was in the order
6′-mesylate/3′′-DMTr thiol < 6′-(DMTr)O/3′′-thiol < 6′-
(DMTr)O/3′′-DMTr thiol.
These results are consistent with the hypothesis that
aggregation arises by hydrogen bonding between uracil
groups. Methanol is expected to disrupt this type of
aggregate by offering hydrogen-bonding opportunities as
a protic solvent. Further, sterically large groups (such
as the DMTr group) at the ends of the analogues might
be expected to interfere with aggregation.
Aggregation of sulfone analogues of tetranucleotides
containing uracil was studied in a variety of solvents. In
water, sulfone DNA tetramer HO-USO2USO2USO2-
USO2-Bn (79) displayed a thermal transition at 46 °C (5
µM) or 50 °C (10 µM) as indicated by the UV spectrum,
with a large hyperchromaticity (69% increase, compared
with 20% for typical DNA molecules). This suggested that
the uridine sulfone in solution is structured at room
temperature. Concentration studies implied that these
molecules formed duplexes.
One of the most readily observed physical properties
of oligomeric sulfide-linked analogues of oligonucleotides
is their insolubility in many solvent mixtures. Further,
problems with solubility increase as the oligonucleotide
analogues become longer. This had both practical and
theoretical implications. Experimentally, the insolubility
of these analogues causes problems in workup, isolation,
and further reaction of the oligomers. Theoretically,
aggregation of analogous DNA and RNA sequences is
virtually never observed. Thus, this is a different and
interesting property of the oligosulfides and oligosulfones.
Even in solvents where outright precipitation was not
observed, evidence for aggregation of sulfide-linked oli-
gouridine analogues could be obtained. For example,
when a variety of tetrameric oligonucleotide analogues
(diprotected tetramer 68, mesylate 70, and thiol 71) were
individually dissolved in THF in the presence of various
amounts of MeOH, allowed to stand for 30 min, injected
onto an HPLC column (Econosil Silica), and then eluted
using 2%, 3%, 4%, or 5% MeOH/EtOAc, the following
observations were made.
(1) Two major peaks were observed in HPLC analysis
in each case (Table 3).
(2) When material eluting in each peak was recovered,
the solvents were removed, the residue was redissolved,
and the HPLC analysis was repeated, two peaks were
again observed in the same ratio. This suggested that
the two peaks correspond to the tetrameric species in two
The oligosulfone DNA analogues dissolved in water
more easily than the oligosulfide DNA analogues. Even

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 4005
Ta ble 4. Com p osition of th e Aggr ega te of 68 a n d Disa ssocia tion Con sta n t of th e Tetr a m er ic Su lfid e Nu cleotid e
An a logu es 68, 71, a n d 70 (Kd iss ) [A]n /[B], or log [B] ) -(log Kd iss) + n log [A])
68
71
70
MeOH (%) in THF
composition number (n)
Kdiss (1 × 10-6 M)
0
2
150
10
0
10
12
20
30
10
20
14
30
40
1010
1.2
420
1600
1
110
150
Ta ble 5. In h ibition of Exp r ession of â-Ga la ctosid a se in
Tr a n sgen ic Dr osoph ila Cells
hibition, even at high concentration (160 µM). The sulfone
octamer 90 showed strong inhibition at 16 µM, and
completely inhibited gene expression of â-galactosidase
at 160 µM. In contrast, sulfone 91 produced no inhibition
at the same concentrations.
â-galactosidase expression (%)
sulfone concn
U2
U4
UUUC
U8
(UUUC)2
(µM)
(65)
(79)
(80)
(90)
(91)
The mechanism for the effects observed remains un-
certain. It is possible that sulfone analogues form a
complex with the polyadenylate tails of the mRNA for
the â-galactosidase gene, which reduced the protein
translation. If true, this would imply that the nonionic
sulfone DNA analogues were able to enter the cell. This
would be consistent with recent animal studies by feeding
mouse with tritium-labeled sulfone analogues; the sul-
fones were found in liver, brain, and other tissues.35
However, it is well-known in the antisense field that
many experiments are required to establish mechanisms
for biological phenomena created by DNA and DNA
analogues. More work is needed here as well.
1.6
16
100
100
50
100
100
90
100
100
100
100
50
0
100
100
90
160
only ecdysone
100
these molecules showed aggregation. In perhaps the
clearest demonstration of the size-dependent behavior of
these molecules, solubility decreases with increasing
chain length. Sulfone dimer 65 dissolves in water up to
ca. 10 mM at rt. The solubility of tetramer 79 or 80 is
lower, however, ca. 100 µM at saturation. The solubility
of sulfone octamer 89 is only ca. 1 µM in water after
heating and storage at rt. The octamer precipitates from
aqueous solution when allowed to stand for 1-2 days at
rt. The precipitate is nearly insoluble, even upon heating.
The structure of the precipitate is not known, but it is
not believed that the molecule has undergone constitu-
tional changes during the precipitation.
Discu ssion
The two synthetic routes described here yield all of the
building blocks needed for the synthesis of the sulfide-
and sulfone-linked DNA analogues. These were used in
a convergent synthetic strategy to prepare deoxyoligo-
nucleotide analogues with different sequences. These, in
turn, supported the examination of the physical and
chemical properties of the nonionic and terminal-sul-
fonated oligonucleotide analogues, as well as preliminary
biological studies. The most striking feature of these
nonionic DNA analogues is the propensity with which
they form self-structure that cannot involve Watson-
Crick pairing. This propensity increases with increasing
length in the oligonucleotide. By crystallography, the
dinucleotide GSO2C forms a nearly standard Watson-
Crick duplex.14 Tetrameric U-rich DNA analogues, se-
quences that cannot form Watson-Crick pairs, dimerize
in nonaqueous solvents, however. Octamer DNA ana-
logues frequently form so much self-structure that they
do not pair successfully with complementary DNA.
This behavior is analogous to the behavior observed
in sulfone-linked RNA analogues.12 Non-Watson-Crick
interactions in a dinucleotide sulfone have now been
shown in solution by NMR spectrometry.36 The tetramer
sulfone 79 displayed a melting temperature as high as
46 °C, and octamer sulfone 90 shows a large hypochro-
maticity increase over a broad temperature range. As
with sulfur-linked RNA analogues,12 the behavior of
S-DNA analogues can change quite dramatically with
small changes in sequence.
Taking a cue from the PNA analogues, where a charged
group is frequently introduced at the end to overcome
problems with insolubility, an effort was made to increase
the solubility of the sulfone DNA analogues by appending
a charged group to one end. This was easily done through
the oxidative cleavage of the S-DMTr group by Oxone.
This generated a sulfonate at one end of the oligomer.
As is observed with PNAs, the aqueous solubility of the
sulfone octamer with a single negative charge (potassium
sulfonate octamer 90) introduced by oxidation is much
improved over that of 89 (up to 3 mM at rt). Sulfonates
90 and 91 also dissolve in MeOH and EtOH.
Octamer 90 in water displayed a very broad transition
in the UV as a function of temperature. The hyperchro-
micity continued to increase even when the temperature
reached 96 °C, a behavior having no precedent in natural
polyuridine or polythymidine. In melting experiments
where the sulfone tetramer 79 or octamer 90 was mixed
with d(A8) at 10 µM total concentration, evidence for the
formation of a weak complex was observed. Mixing
studies at 20 °C (10 mM phosphate, pH 7.0, 100 mM
NaCl, 0.1 mM EDTA, total concentration of nucleotides
15 µM) suggested a possible 2:1 complex with the
octamer. These results were difficult to interpret, how-
ever, due to evident aggregation to give higher order
complexes.
An intriguing set of preliminary biological experiments
suggested that these deoxyoligouridine sulfone DNA
analogues might be biologically active when added ex-
tracellularly to transgenic cells of Drosophila melano-
gaster in culture. In these cells, the change in the level
of expression of a â-galactosidase gene after induction
with ecdysone was monitored (Table 5). Inhibition of
Similar behavior is observed with PNA, another non-
ionic DNA analogue, whose practical application is
hampered by limited solubility in aqueous systems and
pronounced self-organization.37 In PNA, addition of a
(35) Eschgfaeller, B.; Koenig, M.; Boess, F.; Boelsterli, U. A.; Benner,
S. A. J . Med. Chem. 1998, 41, 276-283.
(36) Steinbeck, C.; Richert, C. J . Am. Chem. Soc. 1998, 120, 11576-
11580.
(37) . Gangamani, B. P.; Kumar, V. A.; Ganesh, K. N. Biochem.
Biophys. Res. Commun. 1997, 240, 778-782.
expression was higher with dimer 6′-HO-USO2USO2-
-
Bn-3′′ (65) than with tetramer 79. The sequence 6′-HO-
USO2USO2USO2CSO2-Bn-3′′ (80) did not show any in-

4006 J . Org. Chem., Vol. 67, No. 12, 2002
Huang and Benner
negative charge to one end of the molecule diminishes
aggregation and improves solubility. This was also
observed in the S-DNA analogues, where adding a
sulfonate at the 3′′-end diminished aggregation and
improved solubility in these species. Nevertheless, it
remains a fact needing explanation that Watson-Crick-
type molecular interactions survive in PNA molecules
past the dinucleotide level of oligomerization, and con-
tinue to survive in oligomers substantially larger than
those where the S-DNA analogues lose these interactions.
These observations have implications for the nature
of genetic molecules generally. We have recently sug-
gested that, to serve a role as a genetic molecule, a
molecule must be capable of suffering mutation indepen-
dent of concern that it might lose physical and chemical
properties that are essential for replication.38 This prop-
erty lends itself to a convenient acronym (COSMIC-
LOPER, capable of suffering mutation independent of
concern of loss of properties essential for replication). A
genetic molecule must be able to undergo mutation
without large changes in physical properties. If a genetic
molecule, for example, precipitated when a mutation
occurred, it would not support a Darwinian process that
requires natural selection to operate on a wide range of
functional sequences for their ability to contribute to
fitness, rather than on the narrow range of genetic
sequences that remain soluble.
Solubility is not, of course, the only feature that a
molecule must have to be COSMIC-LOPER. Key to the
ability to serve a role in genetics is the ability to act as
a template. This, in turn, requires that an oligomer not
fold. A polyelectrolyte backbone has been known for
perhaps 50 years to be a feature of a molecule that
discourages folding.42 Even in DNA, it is known that
single-strand folding, or “structure”, is an important
contributor to overall duplex stability.43 Thus, in a simple
melting experiment, a duplex may appear to be less
stable not because it is intrinsically so, but because it is
in competition with a single-strand fold that is more
stable.
The foundations for a second-generation model for
nucleic acid structure must accommodate all of these
factors. It appears as if the polyanionic nature of the DNA
backbone is critical for its ability to perform extensive
genetic function. First, it makes the molecule soluble in
water. Second, it discourages self-folding. Third, it directs
strand-strand interactions to the Watson-Crick “edge”
of the nucleobases, as far from the backbone as possible.
Last, it creates COSMIC-LOPER behavior. The repeating
charge dominates the physical properties of the molecule
to such an extent that dipolar features of the molecule
can be greatly changed without altering the overall
behavior of the molecule.
This model explains why nonionic analogues of DNA
are difficult to generate without losing rule-based mo-
lecular recognition. It further sheds light on at least some
of the paradoxes of DNA structure outlined in the
Introduction. It may also suggest that a polyelectrolyte
structural feature will be general for all genetic mol-
ecules, including those that may be encountered in
hypothetical life forms on planets other than Earth. We
also suggest that this model be considered when PNA or
other nonionic molecules are proposed as genetic sub-
stances in early forms of life on Earth itself.
In most organic molecules, small changes in constitu-
tion frequently can generate large changes in properties,
including solubility. DNA and RNA, in contrast, can
suffer major changes in sequence without substantial
changes in behavior. S-DNA analogues resemble “nor-
mal” organic molecules more than DNA. PNA, in con-
trast, displays intrinsic behavior lying more centrally
between the behavior of normal organic molecules and
that of DNA. Nevertheless, PNA is still far less COSMIC-
LOPER than DNA.
A second-generation model for nucleic acid structure
is now emerging.39 In this model, a compelling case is
made that the polyanionic nature of the DNA and RNA
backbones is the single most important structural feature
creating their COSMIC-LOPER behavior. Even here, the
COSMIC-LOPER property is not complete. If, for ex-
ample, an RNA molecule is rich in guanine, self-structure
of the molecule can render it incapable of serving as a
template. This is paralleled in PNAs, which have prob-
lematic physical properties when they become G-rich.37
Why do oligosulfones lose rule-based behavior sooner
than PNAs in a series of homologues with increasing
length? We do not know. Perhaps some of the answer
lies in the remarkable interaction that PNA displays with
solvent water.40 It will be interesting to see how this
changes with the PNA analogues that introduce an ether
oxygen into the backbone,41 or that carry side chains on
the backbone that may disrupt interaction with water.
Exp er im en ta l Section
Gen er a l P r oced u r es. Most solvents and reagents (pa)
were used without purification unless mentioned otherwise.
Solvents (EtOAc, hexane, and CH2Cl2) for flash chromatogra-
phy were distilled. Dry THF was prepared by distillation from
sodium/benzophenone under an argon atmosphere immedi-
ately before use. TEA was dried over KOH(s) and distilled
under argon. DBU was distilled under reduced pressure and
kept under argon. When necessary, solid reagents were dried
under a high vacuum. Reactions with compounds sensitive to
air or moisture were performed under argon. Solvent mixtures
are indicated as volume/volume ratios. Melting points are
uncorrected. TLC was run on silica gel 60 F254 plates (0.25 mm
thick; Rf values in the text are for the title products), and
visualized under UV light or by a Ce-Mo staining solution
(phosphomolybdate, 25 g; Ce(SO4)2‚4H2O, 10 g; H2SO4, 60 mL,
concentrated; H2O, 940 mL) with heating. Preparative TLC
was performed using silica gel 60 F254 precoated plates (2 mm
thick). Flash chromatography was performed using silica gel
60 (mesh size 0.040-0.063 mm) using a silica gel:crude
compound weight ratio of ca. 30:1. Analytical GC was per-
formed using an instrument with a mass spectral detector.
Specific rotations were measured at rt. IR spectra were
(38) Benner, S. A.; Switzer, C. Y. Simplicity and Complexity in
Proteins and Nucleic Acids; Dahlem Workshop Report; Frauenfelder,
H., Deisenhofer, J ., Wolynes, P. G., Eds.; Dahlem University Press:
Berlin, 1999; pp 335-359.
1
recorded in a range of 4000-600 cm-1. H NMR spectra were
(39) Benner, S. A.; Battersby, T. R.; Kodra, J .; Switzer, C. Y.;
Moroney, S. E.; Voegel, J .; MacPherson, L.; von Krosigk, U.; Hammer,
C.; Richert, C.; Huang, Z.; Horlacher, J .; Schneider, K. C.; Koenig, M.;
Blaettler, M.; Arslan, T.; Hyrup, B.; Egli, M.; Baeschlin, D.; Mueller,
E.; Schmidt, J .; Piccirilli, J .; Roughton, A.; Held, H.; Lutz, M.;
Eschgfaeller, B.; J urczyk, S.; Lutz, S.; Hutter, D.; Nambiar, K. P.;
Stackhouse, J . Pure Appl. Chem. 1998, 70, 263-266.
(41) Kuwahara M.; Arimitsu, M.; Sisido, M. Tetrahedron 1999, 55,
10067-10078.
(42) Flory, P. Principles of Polymer Chemistry; Cornell University
Press: Ithaca, NY, 1953.
(43) Vesnaver, G.; Breslauer, K. J . Proc. Natl. Acad. Sci. U.S.A.
1991, 88, 3569-3573.
(40) Rasmussen, H.; Sandholm, J . Nat. Struct. Biol. 1997, 4, 98-
101.

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 4007
recorded using 500, 400, 300, or 200 MHz. 13C NMR spectra
were recorded using 100 or 75 MHz using DEPT and APT for
determination of hydrogen substitution; all chemical shifts (δ)
are in parts per million relative to that of tetramethylsilane,
and all coupling constants (J ) are in hertz. MS spectra were
recorded in EI-MS, FAB-MS, or MALDI-TOF {2-[(4-hydroxy-
phenyl)azo]benzoic acid (HABA) as matrix} mode for the sul-
fone DNAs. Throughout this Experimental Section, the atom
numbering in the designated compounds follows IUPAC rules.
1-{(2R,3R,4R,5R)-4-Ben zoyloxym et h yl-5-(2-h yd r oxy-
eth yl)-3-[(m-tr iflu or om eth yl)ben zoyloxy]tetr ah ydr ofu r an -
2-yl}u r a cil (5). A mixture of 4 (4.91 g, 11.2 mmol),13a uracil
(2.52 g, 2 equiv), and MSTFA (12.4 mL, 6 equiv) in acetonitrile
(112 mL) was stirred at rt until a clear solution was formed
(ca. 30 min). To this solution was added SnCl4 (2.0 mL, 1.5
equiv), and the mixture was stirred for 14 h at rt (monitored
by TLC, 10% MeOH/CH2Cl2, Rf ) 0.41). The reaction mixture
was cooled to 0 °C, and Na2CO3 (40 mL, 30% aqueous solution,
8 equiv) was added. The mixture was then extracted three
times with EtOAc. The combined organic layers were washed
with NaCl(satd), dried over MgSO4(s), and evaporated under
reduced pressure. The residue was chromatographed on silica
gel (2% MeOH/CH2Cl2) to give 5 (5.92 g, 96%) as a colorless
foam.
and the mixture was stirred for 18 h at rt (TLC, 7.5% MeOH/
CH2Cl2, Rf ) 0.42). The reaction mixture was cooled to 0 °C,
and saturated NaHCO3 (8 mL) was added. The mixture was
then extracted three times with EtOAc. The combined organic
layers were washed with NaCl(satd), dried over MgSO4(s), and
evaporated under reduced pressure. The residue was chro-
matographed on silica gel (2% MeOH/CH2Cl2) to give 7 (709
mg, 95%) as a colorless foam. This was crystallized from
EtOAc/hexane.
1H NMR (CDCl3): δ 2.02-2.30 (2m, 2H, HOCH2CH2), 2.53
(br, 1H, HO), 3.05 (m, 1H, H-4′), 3.93 (t, J ) 4.9, 2H, HOCH2),
4.65 (md, J ) 6.6, 3H, 4′-CH2OBz, H-5′), 5.91 (d, J ) 1.5, 1H,
H-2′), 6.11 (dd, J ) 1.5, 6.2, 1H, H-3′), 7.35 (m, 2H), 7.42-
7.65 (m, 6H), 7.82 (m, 1H), 7.94 (m, 5H), [ar-H, H-5], 8.18 (d,
J ) 8.1, 1H, H-6), 8.22 (s, 1H), [ar-H], 9.03 (s, 1H, NH).
13C NMR (CDCl3): δ 36.16 (HOCH2CH2), 45.00 (C4′), 59.54
(C4′′), 60.28 (HOCH2), 78.47 (C3′), 81.09 (C5′), 93.85 (C2′), 96.93
(C5), 123.42 (q, J ) 270.8, CF3), 127.73 (m, CHCHCHCCF3),
128.42, 128.97 (CHCHCPh), 129.25, 129.58, 129.94, 130.10,
131.49, 133.02, 133.21 (C, ar-CH), 145.18 (C6), 154.50 (C2),
162.8 (CONH), 164.38 (ArCO), 166.21 (PhCO), 166.7 (C4).
MS (m/e, relative intensity): 634 ([M - 17]+, 1), 564 (2),
556 (1), 541 (1), 528 (1), 488 (4), 473 (3), 460 (4), 384 (1), 356
(5), 296 (6), 268 (2), 255 (7), 228 (80), 190 (88), 173 (100), 145
(87), 122 (60), 111 (30), 105 (97), 95 (40), 77 (95), 51 (67).
Anal. Calcd for C33H28N3O8F3 (651.59): C 60.83, H 4.33, N
6.45. Found: C 60.58, H 4.27, N 6.56.
1H NMR (CDCl3): δ 1.98-2.26 (m, 3H, HOCH2CH2, OH),
3.07 (m, 1H, H-4′), 3.93 (m, 2H, HOCH2), 4.40-4.65 (m, 3H,
4′-CH2OBz, H-5′), 5.76-5.88 (m, 3H, H-2′, H-3′, H-5), 7.32-
7.44 (m, 3H), 7.50-7.60 (m, 2H), 7.83 (m, 1H) 7.94 (m, 2H),
8.18 (m, 1H), 8.24 (m, 1H), [H-6, ar-H], 9.03 (s, 1H, NH).
13C NMR (CDCl3): δ 36.16 (HOCH2CH2), 45.06 (C4′), 59.41
(C4′′), 60.39 (HOCH2), 78.17 (C5′), 80.23 (C3′), 92.55 (C2′), 102.95
(C5), 123.42 (q, J ) 270.8, CF3), 126.73 (m, CHCHCHCCF3),
128.50 (CHCHCPh), 129.25 (C-Ph), 129.38 (CHCHCHCCF3),
129,57 (CHCPh), 129.65 (CCHCCF3), 130.28 (d, J ) 3.4,
CCHCCF3), 131.37 (q, J ) 33.1, CCF3), 132.98 (CHCHCH-
CCF3), 133.38 (CHCHCHCPh), 140.97 (C6), 150.50 (C2), 163.17
(C4), 164.38 (ArCO), 166.21 (PhCO). IR (KBr): 3430, 3120,
3060, 2950, 1720, 1690, 1618, 1600, 1582, 1450, 1250, 1130,
N6-Ben zoyl-9- a n d -7-{(2R,3R,4R,5R)-4-b en zoyloxy-
m et h yl-5-(2-h yd r oxyet h yl)-3-[(m -t r iflu or om et h yl)b en -
zoyloxy]tetr a h yd r ofu r a n -2-yl}a d en in e (10a a n d 10b).
Met h od A. A mixture of 4 (500 mg, 1.147 mmol), N6-ben-
zoyladenine (555 mg, 2 equiv), and MSTFA (455 µL, 2 equiv)
in acetonitrile (11 mL) was stirred at rt for 15 min. To this
mixture was added TMSOTf (205 µL, 1.0 equiv), and the
mixture was stirred for 20 h at rt (monitored by TLC, 10%
MeOH/CH2Cl2, Rf ) 0.46, 0.35; the N-9 isomer moves faster
than the N-7 isomer). The reaction mixture was cooled to 0
°C, saturated NaHCO3 (5 mL) was added, and the suspension
was then extracted with EtOAc. The combined organic layers
were dried over MgSO4(s) and evaporated under reduced
pressure. The residue was chromatographed on silica gel (2%
MeOH/CH2Cl2) to give 10a (340 mg, 44% yield) and 10b (326
mg, 42%) as colorless foams.
815, 755, 714, 693 cm-1
.
MS (m/e, relative intensity): 530 ([M - 18]+, 3) 462 (11),
437 (21), 427 (35), 357 (14), 253 (37), 247 (60), 228 (36), 190
(10), 173 (100), 145 (24), 125 (72), 111 (31), 105 (49), 95 (40),
77 (58).
Anal. Calcd for C26H23N2O8F3 (548.47): C 56.94, H 4.23, N,
5.11. Found: C 56.78, H 4.27, N 5.18.
Meth od B. A mixture of 4 (500 mg, 1.147 mmol), N6-
benzoyladenine (555 mg, 2 equiv), and MSTFA (455 µL, 2
equiv) in acetonitrile (11 mL) was stirred at rt for 15 min. To
this mixture was added TMSOTf (205 µL, 1.0 equiv), and the
mixture was stirred for 45 h at 60 °C (monitored by TLC, 10%
MeOH/CH2Cl2, Rf ) 0.46, only the N-9 isomer). The reaction
mixture was cooled to 0 °C, saturated NaHCO3 (5 mL) was
added, and the suspension was then extracted with EtOAc.
The combined organic layers were dried over MgSO4(s) and
evaporated under reduced pressure. The residue was chro-
matographed on silica gel (3% MeOH/CH2Cl2) to give 10a (520
mg, 67%) as a colorless foam; the N-9 isomer was the only
product isolated.
1-{(2R,3R,4R,5R)-4-Ben zoyloxym et h yl-5-(2-h yd r oxy-
eth yl)-3-[(m-tr iflu or om eth yl)ben zoyloxy]tetr ah ydr ofu r an -
2-yl}th ym in e (6). A mixture of 4 (429 mg, 1.01 mmol),
thymine (250 mg, 2.0 equiv), and MSTFA (1.18 mL, 6 equiv)
in acetonitrile (10.1 mL) was stirred until a clear solution was
formed (ca. 30 min). To this solution was added SnCl4 (162
µL, 1.5 equiv), and the mixture was stirred for 14 h at rt
(monitored by TLC, 7.5% MeOH/CH2Cl2, Rf ) 0.37). The
reaction mixture was cooled to 0 °C, saturated NaHCO3 (5 mL)
was added, and the mixture was then extracted three times
with EtOAc. The combined organic layers were dried over
MgSO4(s), the solvent was evaporated under reduced pressure,
and the residue was chromatographed on silica gel (EtOAc,
Rf ) 0.42) to give 6 (507 mg, 92%) as a colorless foam.
1H NMR (CDCl3): δ 1.96 (s, 3H, CH3), 2.00-2.26 (2m, 2H,
HOCH2CH2), 3.14 (m, 1H, H-4′), 3.93 (m, 2H, HOCH2), 4.43
(td, J ) 3.3, 9.6, 1H, H-5′), 4.56 (dd, J ) 5.7, 11.6, 1H, 4′-CH2-
OBz), 4.60 (dd, J ) 7.4, 11.6, 1H, 4′-CH2OBz), 5.79 (d, J )
2.6, 1H, H-2′), 5.85 (dd, J ) 2.7, 7.1, 1H, H-3′), 7.12 (s, 1H,
H-6), 7.40 (t, 2H), 7.56 (t, 2H) 7.82 (d, 1H), 7.93 (d, 2H), 8.28
(d, 2H), [ar-H], 8.24 (s, 1H, NH).
1
Da ta for th e N-9 Isom er 10a . H NMR (CDCl3): δ 2.08-
2.28 (m, 2H, HOCH2CH2), 2.85 (br, 1H, HO), 3.60 (m, 1H, H-4′),
3.92 (t, J ) 4.9, 2H, HOCH2), 4.65 (m, 3H, 4′-CH2OBz, H-5′),
6.21 (dd, J ) 1.9, 6.3, 1H, H-3′), 6.25 (d, J ) 2.0, 1H, H-2′),
7.39 (m, 2H), 7.46-7.63 (m, 5H), 7.85 (m, 1H), 7.94 (m, 2H),
8.04 (m, 2H), 8.19 (m, 2H), 8.27 (s, 1H, H-8), 8.80 (s, 1H, H-2),
[ar-H], 9.38 (s, 1H, NH).
13C NMR (CDCl3): δ 36.21 (HOCH2CH2), 45.18 (C4′), 59.39
(C4′′), 60.40 (HOCH2), 78.78 (C3′), 80.96 (C5′), 89.92 (C2′), 123.73
(C5), 127.01 (q, J ) 273.5, CF3), 126.76 (m, CHCHCHCCF3),
127.93, 128.51 (CHCHCPh), 128.87, 129.29, 129.42, 129.55,
130.36, 132.84, 132.94, 133.40, 133.59 (C, ar-CH), 141.98 (C8),
149.86 (C4), 151.24 (C6), 152.90 (C2), 164.44 (CONH), 164.67
(ArCO), 166.19 (PhCO), 166.7 (C4).
Anal. Calcd for C27H25N2O8F3 (562.49): C 57.65, H 4.48, N,
4.98. Found: C 57.48, H 4.66, N 5.05.
N4-Ben zoyl-1-{(2R,3R,4R,5R)-4-ben zoyloxym eth yl-5-(2-
h yd r oxye t h yl)-3-[(m -t r iflu or om e t h yl)b e n zoyloxy]t e t -
r a h yd r ofu r a n -2-yl}cytosin e (7). MSTFA (1.42 mL, 6 equiv)
was added to a mixture of 4 (500 mg, 1.147 mmol) and N4-
benzoylcytosine (493 mg, 2.0 equiv) in acetonitrile (11.5 mL).
The mixture was stirred until a clear solution was formed (ca.
30 min). To this solution was added SnCl4 (203 µL, 1.5 equiv),
MS (m/e, relative intensity): 676 ([M + 1]+, 3), 675 (M+,
10), 659 (1), 609 (8), 570 (3), 554 (3), 541 (18), 486 (3), 473 (3),
437 (8), 419 (2), 308 (2), 297 (7), 270 (2), 215 (2), 190 (6), 173
(100), 164 (20), 145 (40), 135 (20), 105 (90), 95 (20), 77 (31), 51
(15).

4008 J . Org. Chem., Vol. 67, No. 12, 2002
Huang and Benner
Anal. Calcd for C34H28N5O7F3 (675.62): C 60.44, H 4.18, N
10.36. Found: C 60.23, H 4.32, N 10.28.
Da ta for th e N-7 Isom er 10b. H NMR (CDCl3): δ 2.22-
2.44 (m, 2H, HOCH2CH2), 2.90 (br m, 2H, H-4′, HO), 4.10 (m,
2H, HOCH2), 4.44 (dd, J ) 7.4, 12.6, 1H, 4′-CH2OBz), 4.56 (dd,
J ) 3.0, 11.3, 1H, 4′-CH2OBz), 4.76 (t, J ) 8.1, 1H, H-5′), 6.08
(d, J ) 4.9, 1H, H-2′), 6.84 (m, 2H), 7.02 (m, 1H, H-3′), 7.46-
7.63 (2m, 5H), 7.85 (m, 1H), 7.84, 7.90 (2m, 6H), [ar-H], 8.33
(s, 1H, H-8), 8.67 (s, 1H, H-2), 15.38 (br, 1H, NH).
bottom flask containing 5 (5.975 g, 10.9 mmol) and dimethoxy-
triphenylmethyl chloride (7.35 g, 2 equiv) was added TEA (12.1
mL, 8 equiv) in THF (109 mL) at 0 °C under argon. The
mixture was allowed to warm to rt, and stirred for 4 h (the
reaction was followed by TLC, 5% MeOH/CH2Cl2, Rf ) 0.46).
MeOH (5 mL) was then added, and the mixture was stirred
for 5 min. The solvents were removed under reduced pressure,
and the residue was partitioned between water and EtOAc
(each ca. 100 mL). The organic layers were separated, and the
aqueous phase was extracted with EtOAc three times. The
combined organic layers were washed with saturated NaCl,
dried over MgSO4(s), and evaporated under a vacuum. The
residue was chromatographed on silica gel (1.5% MeOH/CH2-
Cl2, Rf ) 0.18) to give 21 (8.783 g, 95%) as a white foam.
1H NMR (CDCl3): δ 1.98-2.22 (2m, 2H, (DMTr)OCH2CH2),
2.92 (m, 1H, H-4′), 3.34 (m, 2H, (DMTr)OCH2), 3.75 (s, 6H,
2CH3O), 4.34-4.51 (m, 1H, H-5′), 4.43-4.57 (m, 2H, 4′-CH2-
OBz), 5.72 (dd; d, J ) 3.1, 9.6; 8.2, 2H, H-3′, H-5), 5.87 (d, J )
3.1, 1H, H-2′), 6.80 (m, 4H, ar-H), 7.18 (d, J ) 8.1, 1H, H-6),
7.20-7.45 (m, 11H, ar-H), 7.53 (m, 2H), 7.79 (m, 1H), 7.90 (m,
2H), 8.13 (m, 1H), 8.20 (m, 1H), (ar-H), 8.47 (s, 1H, NH).
13C NMR (CDCl3): δ 34.53 ((DMTr)OCH2CH2), 44.87 (C4′),
55.20 (CH3O), 59.72 ((DMTr)OCH2), 60.42 (C4′′), 77.75 (C3′),
79.25 (C5′), 86.48 (ar3-CO), 90.66 (C2′), 102.89 (C5), 113.15
(CHCHCOCH3 in ar) 123.5 (q, J ) 270, CF3), 126.74 (d, J )
3.8, CHCHCHCCF3), 126.82, 127.86, 128.06,128.48, 129.28,
129.32, 129.57, 129.66, 129.96,130.20 (ar-C, CH), 131.51 (d, J
) 33, CCF3), 132.94 (CHCHCHCCF3), 133.32 (CHCHCHCPh
in Bz), 136.11 (CCHCHCOCH3 in ar), 139.94 (C6), 144.87 (C
in Ph of DMTr), 149.68 (C2), 158.50 (CH3OC), 162.53 (C4),
164.19 (ArCO), 166.06 (PhCO).
1
13C NMR (CDCl3): δ 36.68 (HOCH2CH2), 44.95 (C4′), 59.58
(C4′′, HOCH2), 78.96 (C3′), 81.37 (C5′), 91.68 (C2′), 114.52 (C5),
123.36 (q, J ) 273.0, CF3), 126.41 (m, CHCHCHCCF3), 127.63,
128.40 (CHCHCPh), 128.69 129.07, 129.52, 129.93, 130.88,
131.32, 131.69, 132.76, 133.32 (C, ar-CH), 136.51 (C4), 142.41
(C8), 143.24 (C2), 150.27 (C6), 163.01 (COPh), 166.06 (ArCO),
166.06 (CONH).
Anal. Calcd for C34H28N5O7F3 (675.62): C 60.44, H 4.18, N
10.36. Found: C 60.28, H 4.22, N 10.49.
N2-Isobu tyr yl-9- a n d -7-{(2R,3R,4R,5R)-4-ben zoyloxy-
m et h yl-5-(2-h yd r oxyet h yl)-3-[(m -t r iflu or om et h yl)b en -
zoyloxy]tetr a h yd r ofu r a n -2-yl}gu a n in e (12a a n d 12b). A
mixture of 4 (345 mg, 0.791 mmol), N2-isobutyrylguanine
(349.7 mg, 2.0 equiv), and MSTFA (0.88 mL, 6 equiv) in
acetonitrile (7.9 mL) was stirred until a clear solution appeared
(ca. 30 min). TMSOTf (141 µL, 1.0 equiv) was then added, and
the mixture stirred for 15 h at rt (monitored by TLC, 7.5%
MeOH/CH2Cl2, Rf ) 0.48, 0.41; N-9 isomer 12a moves slower
than N-7 isomer 12b). Saturated aqueous NaHCO3 (12 mL)
was added to the mixture, and the mixture was then extracted
three times with EtOAc. The combined organic layers were
dried over MgSO4(s) and evaporated in vacuo. The residue was
chromatographed on silica gel (1-3% MeOH/CH2Cl2) to give
12a (161 mg, 31%, N-9 isomer) and 12b (177 mg, 34%, N-7
isomer) as white foams.
NOE: irradiation at δ 2.92 (H-4′) gives NOE at 2.05
((DMTr)OCH2CH2, 1.0%), 4.42 (H-5′, 2.1%), 4.50 (4′-CH2OBz,
3.2%), 5.75 (H-3′, 9.8%), 7.18 (H-6, 1.1%); irradiation at δ 4.42
(H-5′) gives NOE at 2.10 ((DMTr)OCH2CH2, 2.2%), 2.95 (H-
4′, 1.9%), 3.37 ((DMTr)OCH2, 2.0%), 4.50 (4′-CH2OBz, 2.4%),
5.89 (H-2′, 7.7%); irradiation at δ 5.89 (H-2′) gives NOE at 4.42
(H-5′, 2.3%), 5.75 (H-3′, 3.7%), 7.18 (H-6, 5.5%), 7.92, 8.1 (ar-
H-m-CF3, 1.6%).
1
Da ta for th e N-9 Isom er 12a . H NMR (CDCl3): δ 1.24,
1.25 (2d, J ) 6.9, 6.8, 6H, 2CH3), 2.10-2.33 (m, 2H,
HOCH2CH2), 2.78 (sept, J ) 6.7, 1H, CH3CHCH3), 3.02 (br,
1H, OH), 3.64 (m, 1H, H-4′), 3.95 (m, 2H, HOCH2), 4.65-4.82
(m, 3H, 4′-CH2OBz, H-5′), 5.87 (d, J ) 1.3, 1H, H-2′), 5.98 (dd,
J ) 1.2, 5.9, 1H, H-3′), 7.17 (m, 2H), 7.26, 7.43 (2m, 2H), 7.70-
7.89 (m, 3H), 7.95 (s, 1H, H-8), 8.27 (m, 2H), [ar-H, H-5], 9.48,
11.9 (2s, 2H, 2NH).
FAB-MS (m/e, relative intensity): 850 (M+, 5), 773 ([M -
77]+, 3), 743 ([M - 107]+, 3), 621 (1), 509 (2), 303 (100), 173
(18), 154 (30), 105 (25), 77 (12).
13C NMR (CDCl3): δ 18.95 (CH3), 35.81 (HOCH2CH2), 35.25
(CH3CHCH3), 44.42 (C4′), 59.06 (C4′′), 60.77 (HOCH2), 78.30
(C5′), 80.73 (C3′), 89.52 (C2′), 122.57 (C5), 123.87 (q, J ) 274.8,
CF3), 125.71 (m, CHCHCHCCF3), 127.29 (CHCHCPh), 129.44
(CPh), 129.59 (CHCHCHCCF3), 129.68 (CHCPh), 129.79 (CCH-
CCF3), 131.10 (CCHCCF3), 131.57 (q, J ) 33.2, CCF3), 133.09
(CHCHCHCCF3), 133.04 (CHCHCHCPh), 138.73 (C8), 147.60
(C2), 147.41 (C4), 155.13 (C6), 165.03 (ArCO), 164.21 (PhCO),
179.41 (CO-ipr).
Anal. Calcd for C47H41N2O10F3 (850.84): C 66.35, H 4.86, N
3.29. Found: C 66.23, H 4.77, N 3.38.
1-{(2R,3R,4R,5R)-5-[2-(4,4′-Dim et h oxyt r ip h en ylm et h -
oxy)eth yl]-3-h yd r oxy-4-h yd r oxym eth yltetr a h yd r ofu r a n -
2-yl}u r a cil (29). NaOH (5.55 mL, 1 M, aq, 4 equiv) was added
to 21 (1.18 g, 1.39 mmol) in MeOH (20 mL), and the mixture
was stirred for 15 min (monitored by TLC, 10% MeOH/CH2-
Cl2, Rf ) 0.48). The pH of the solution was then adjusted to
6-7 with dilute HOAc. The MeOH was removed under reduced
pressure, and the residue was extracted with EtOAc three
times. The combined organic layers were washed with NaCl-
(satd), dried over MgSO4(s), and evaporated. The residue was
chromatographed on silica gel (2% MeOH/CH2Cl2) to give 29
(770 mg, 96% yield) as a white foam.
Anal. Calcd for C31H30N5O8F3 (657.59): C 56.62, H 4.59, N
10.65. Found: C 56.38, H 4.37, N 10.58.
1
Da ta for th e N-7 Isom er 12b. H NMR (CDCl3): δ 1.22,
1.28 (2d, J ) 6.9, 6.9, 6H, 2CH3), 2.14-2.38 (m, 2H,
HOCH2CH2), 2.76 (sept, J ) 6.9, 1H, CH3CHCH3), 2.97 (br,
1H, OH), 3.46 (m, 1H, H-4′), 4.06 (m, 2H, HOCH2), 4.55-4.70
(m, 3H, 4′-CH2OBz, H-5′), 6.39 (d, J ) 1.1, 1H, H-2′), 6.62 (dd,
J ) 1.0, 5.8, 1H, H-3′), 7.27 (m, 2H), 7.46, 7.56 (2m, 2H), 7.78-
7.92 (m, 3H), 8.19 (s, 1H, H-8), 8.23 (m, 2H), [ar-H, H-5], 9.68,
12.5 (2s, 2H, 2NH).
1H NMR (CDCl3): δ 1.88-2.10 (m, 3H, H-4′, (DMTr)-
OCH2CH2), 3.34 (t, 2 H, J ) 6.0, (DMTr)OCH2), 3.80 (s, 6H,
2CH3O), 3.70-3.90 (m, 2H, 4′-CH2OH), 4.33-4.50 (m, 2H, H-3′,
H-5′), 5.05 (br, 2H, 2OH), 5.62 (s, 1H, H-2′), 5.67 (d, J ) 6.7,
H-5), 6.80 (m, 4H, ar-H), 7.20-7.45 (m, 10H, ar-H, H-6), 10.7
(br, 1H, NH).
13C NMR (CDCl3): δ 18.84 (CH3), 36.29 (HOCH2CH2), 36.36
(CH3CHCH3), 44.36 (C4′), 59.69 (C4′′), 60.08 (HOCH2), 80.24
(C5′), 81.48 (C3′), 91.71 (C2′), 110.87 (C5), 123.46 (q, J ) 273.5,
CF3), 126.78 (m, CHCHCHCCF3), 128.23 (CHCHCPh), 129.36
(CPh), 129.58 (CHCHCHCCF3), 129.87 (CHCPh), 130.29 (CCH-
CCF3), 131.15 (CCHCCF3), 131.36 (q, J ) 32.3, CCF3), 132.99
(CHCHCHCCF3), 133.14 (CHCHCHCPh), 141.17 (C8), 147.48
(C2), 152.63 (C4), 157.59 (C6), 164.48 (ArCO), 166.26 (PhCO),
179.60 (CO-ipr).
Anal. Calcd for C31H30N5O8F3 (657.59): C 56.62, H 4.59, N
10.65. Found: C 56.46, H 4.45, N 10.76.
1-{(2R,3R,4R,5R)-4-Ben zoyloxym eth yl-5-[2-(4,4′-dim eth -
oxytr ip h en ylm eth oxy)eth yl]-3-[(m -tr iflu or om eth yl)ben -
zoyloxy]tetr a h yd r ofu r a n -2-yl}u r a cil (21). To a round-
13C NMR (CDCl3): δ 34.9 ((DMTr)OCH2CH2), 48.3 (C4′), 55.2
(CH3O), 58.1 ((DMTr)OCH2), 60.4 (C4′′), 77.2 (C3′), 79.3 (C5′),
86.3 (ar3-CO), 92.8 (C2′), 101.7 (C5), 113.1 (CHCHCOCH3 in
ar), 126.4 (CHCHCHC- in Ph of DMTr), 127.8 (CHC- in Ph),
128.1 (CHCHC- in Ph), 130.0 (CHCOCH3 in ar), 136.2
(CCHCHCOCH3 in ar), 139.2 (C6), 144.8 (C in Ph), 151.2 (C2),
158.4 (CH3OC), 163.7 (C4). FAB-MS (m/e, relative intensity):
597 ([M + 23]+, 5), 574 (M+, 6), 497 (2), 467 (2), 399 (1), 303
(100), 289 (7), 154 (37), 136 (30), 107 (9), 91 (7), 77 (10).
Anal. Calcd for C32H34N2O8F3 (574.63): C 66.88, H 5.96, N
4.88. Found: C 66.55, H 5.84, N 4.97.
1-{(2R,3R,4R,5R)-4-Acetylthiomethyl-5-[2-(4,4′-dimethoxy-
tr iph en ylm eth oxy)eth yl]-3-h ydr oxytetr ah ydr ofu r an -2-yl}-

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 4009
u r a cil (31). Diisopropyl azodicarboxylate (90-95%, 759 µL,
1.5 equiv) and triphenylphosphine (910 mg, 1.5 equiv) were
dissolved in THF (45 mL), and the mixture was stirred in an
ice-salt bath (-20 °C). After a white precipitate appeared,
29 (1.33 g, 2.31 mmol) in THF (20 mL) and thioacetic acid (247
µL, 1.5 equiv) in THF (2.2 mL) were added simultaneously
and separately. The temperature was allowed to rise to rt, the
mixture was stirred at rt for 5 h (monitored by TLC, 7.5%
MeOH/CH2Cl2, Rf ) 0.50), and TEA (1 mL) was added. The
solvent was removed under reduced pressure, and the residue
was partitioned between water and EtOAc (each 10 mL). The
organic layer was separated, and the aqueous phase was
extracted three times with EtOAc. The combined organic
layers were washed with NaCl(satd), dried, and evaporated
under a vacuum. The residue was carefully chromatographed
on silica gel twice (gradient, 50-80% EtOAc/hexane and 1.5-
3% MeOH/CH2Cl2) to give 31 (1.31 g, 89%) as a white foam.
1H NMR (CDCl3): δ 1.82-2.28 (2m, 3H, (DMTr)OCH2CH2,
H-4′), 2.34 (s, 3H, CH3COS), 2.95-3.10 (m, 2H, 4′-CH2S), 3.48
(t, J ) 6.3, 2H, (DMTr)OCH2), 3.83 (s, 6H, 2CH3O), 4.23-4.40
(m, 2H, HO, H-3′), 5.03 (m, 1H, H-5′), 5.68 (dd, J ) 2.1, 8.22,
2H, H-2′, H-5), 6.87 (m, 4H, ar-H), 7.20-7.60 (m, 10H, ar-H),
10.4 (s, 1 H, NH).
CH2Cl2, Rf ) 0.36), and the toluene was evaporated under
reduced pressure. The residue was chromatographed on silica
gel (0.5-2% MeOH/CH2Cl2) to give 19 (511 mg, 86%).
1H NMR (CDCl3): δ 1.80-1.90 (m, 1H, (DMTr)OCH2CH2),
2.05-2.20 (m, 4H, (DMTr)OCH2CH2, H-3′, H-4′), 2.84 (dd, J
) 7.1, 13.7, 1H, 4′-CH2S), 3.04 (dd, J ) 4.1, 13.7, 1H, 4′-CH2S),
3.79 (s, 6H, 2CH3O), 3.88 (m, 1 H, H-5′), 5.65 (d, J ) 8.1, 1H,
H-5), 5.94 (dd, J ) 4.1, 6.3, 1H, H-2′), 6.72 (m, 4H, ar-H), 7.18-
7.35 (m, 8H, ar-H, H-6), 7.44 (m, 2H, ar-H), 8.55 (s, 1H, NH).
13C NMR (CDCl3): δ 29.98 (C4′′), 30.62 (CH3CO), 34.57
((DMTr)OCH2CH2), 38.55 (C3′), 42.54 (C4′), 55.26 (CH3O), 60.23
((DMTr)OCH2), 82.36 (C5′), 85.27 (C2′), 86.39 (ar3-CO), 101.97
(C5), 113.14 (CHCHCOCH3 in ar), 126.81 (CHCHCHC- in Ph
of DMTr), 127.86 (CHC- in Ph of DMTr), 128.13 (CHCHC-
in Ph of DMTr), 130.03 (CHCOCH3 in ar), 136.26 (CCHCH-
COCH3 in ar), 139.32 (C6), 145.05 (C in Ph of DMTr), 149.85
(C2), 158.52 (CH3OC), 162.70 (C4), 194.78 (COS).
FAB-MS (m/e, relative intensity): 616 (M+, 2), 539 ([M -
77]+, 2), 509 ([M - 107]+, 1), 414 (2), 337 (2), 303 (100), 288
(11), 273 (10), 227 (15), 215 (10), 195 (10), 152 (11), 135 (20),
112 (17), 107 (4), 105 (6), 95 (15), 91 (5), 77 (11).
IR (KBr): 3430, 3200, 3060, 2940, 1690, 1610, 1560, 1510,
1465, 1450, 1370, 1250, 1170, 1115, 1075, 1035, 830, 765, 705,
13C NMR (CDCl3): δ 24.4 (C4′′), 30.8 (CH3COS), 34.7
((DMTr)OCH2CH2), 47.7 (C4′), 55.6 (CH3O), 60.7 ((DMTr)-
OCH2), 77.2 (C3′), 82.1 (C5′), 86.8 (ar3-CO), 93.9 (C2′), 102.2 (C5),
113.3 (CHCHCOCH3 in ar), 127.1 (CHCHCHC- in Ph), 127.9
(CHC- in Ph), 128.1 (CHCHC- in Ph), 130.2 (CHCOCH3 in
ar), 136.6 (CCHCHCOCH3 in ar), 139.6 (C6), 145.4 (C in Ph),
150.9 (C2), 158.8 (CH3OC), 164.3 (C4), 196.7 (MeCOS).
FAB-MS (m/e, relative intensity): 655 ([M + 23]+, 1), 633-
([M + 1]+, 1), 632 (M+, 2), 555 (2), 525 (2), 515 (3), 454 (1), 319
(4), 303 (100), 289 (4), 279 (9), 273 (3), 213 (3), 165 (3), 154
(6), 136 (6), 107 (3), 105 (3), 91 (3), 77 (5).
630, 585 cm-1
.
UV (CH2Cl2): λmax ) 260, 235 nm.
Optical rotation: [R]D ) +60.5° (c ) 1.5, MeOH). Anal.
Calcd for C34H36N2O7S (616.73): C 66.21, H 5.88, N 4.54.
Found: C 66.03, H 5.97, N 4.49. This compound is identical
to 19 prepared by photolytic deoxygenation, as indicated by
1H and 13C NMR.
Met h yl-3,5-d id eoxy-2-a cet yl-3-a cet oxym et h yl-6-ter t-
bu tyld ip h en ylsilyl-1-O-r- or -â-D-a llofu r a n osid e (35). Into
a solution of 2 (18.23 g, 61.59 mmol) and imidazole (10.07 g,
2.4 equiv) in dry DMF (123 mL) was injected tert-butyldiphen-
ylchlorosilane (18.86 mL, 1.2 equiv) under argon. The reaction
was monitored by TLC (EtOAc:hexane ) 3:7, Rf ) 0.35). After
40 min, MeOH (10 mL) was added and the mixture stirred
for 10 min. Most of the solvents were removed by rotary
evaporation under a high vacuum at 60 °C (water bath). The
residue was then dissolved in MeOH (300 mL), and NaOH (4.9
g, 2.0 equiv) in H2O (30 mL) was added. The reaction was
monitored by TLC (EtOAc:hexane ) 4:6, Rf ) 0.17). After 20
min, dilute HOAc was added to adjust the pH to 6-7. Most of
the MeOH was removed by evaporation, and the residue was
extracted with EtOAc three times. The combined organic
layers were washed with NaCl(satd), dried over MgSO4(s), and
evaporated by rotary evaporation. The crude product was dried
under a high vacuum overnight. An analytical sample of the
major anomer was prepared by chromatographic purification
on silica gel (EtOAc:hexane ) 1:1). Into a solution of the crude
product (ca. 60 mmol), 4-dimethylaminopyridine (100 mg), and
triethylamine (85.6, 10 equiv) in dry THF (123 mL) was
injected acetic anhydride (29.1 mL, 5 equiv) in an ice bath
under argon. The reaction was monitored by TLC (EtOAc:
hexane ) 1:4, Rf ) 0.27). After 1 h, MeOH (10 mL) was added
and the mixture stirred for 10 min. Solvents were evaporated
in vacuo, EtOAc/hexane (3:7, 200 mL) was added to dissolve
the residue, and the mixture was filtered to remove salts. The
filtrate was concentrated by rotary evaporation, and the crude
product was purified by chromatography on silica gel (EtOAc:
hexane ) 1:4) to give 35 (21.90 g, 69% from 2) as a colorless
oil. Only one anomer was obtained; the other (ca. 10%) was
lost during the chromatographic purification.
IR: 3400, 3200, 3060, 2950, 1705, 1690, 1610, 1510, 1460,
1380, 1250, 1170, 1090, 1040, 960 cm-1
.
Anal. Calcd for C34H36N2O8S (632.73): C 64.54, H 5.74, N
4.43. Found: C 64.23, H 5.87, N 4.50.
1-{(2R,3R,4R,5R)-4-Acetylth iom eth yl-5-[2-(4,4-d im eth -
oxytr ip h en ylm eth oxy)eth yl]-3-(1-im id a zoth ioca r bon yl-
oxy)tetr a h yd r ofu r a n -2-yl}u r a cil (33). Compound 31 (950
mg, 1.50 mmol) and 1,1′-thiocarbonyldiimidazole (535 mg, 2.0
equiv) were dissolved in dry THF (15 mL). The mixture was
stirred at rt for 45 h (monitored by TLC, 5% MeOH/CH2Cl2,
Rf ) 0.21), the THF was removed under reduced pressure, and
the residue was chromatographed on silica gel (1-3% MeOH/
CH2Cl2) to give 33 (981 mg, 88%) as a foam.
1H NMR (CDCl3): δ 1.85-2.20 (2m, 2H, (DMTr)OCH2CH2),
2.28 (s, 3H, CH3COS), 2.75-2.90 (m, 1H, H-4′), 2.98 (dd, J )
3.5, 12.1, 1H, 4′-CH2S), 3.12 (dd, J ) 4.7, 13.1, 1H, 4′-CH2S),
3.33 (t, J ) 5.0, 2H, (DMTr)OCH2), 3.75 (s, 6H, 2CH3O), 4.05-
4.20 (m, 1H, H-5′), 5.67 (dd, J ) 2.2, 5.4, 2H, H-2′, H-5), 6.02
(dd, J ) 2.2, 6.5, 1H, H-3′), 6.81 (m, 4H, ar-H), 7.05 (m, 1H,
HIm-5), 7.20-7.60 (m, 11H, ar-H, HIm-4), 8.33 (s, 1H, HIm-2),
9.94 (2br, 1 H, NH).
13C NMR (CDCl3) δ: 24.4 (C4′′), 30.8 (CH3COS), 34.5
((DMTr)OCH2CH2), 46.2 (C4′), 55.5 (CH3O), 60.3 ((DMTr)-
OCH2), 81.2 (C5′), 86.2 (C3′), 86.8 (ar3-CO), 91.8 (C2′), 103.5 (C5),
113.3 (CHCHCOCH3 in ar), 127.1 (CHCHCHC- in Ph), 128.3,
128.9 (ImC5, ImC4), 127.9 (CHC- in Ph), 128.1 (CHCHC- in
Ph), 130.2 (CHCOCH3 in ar), 131.7 (ImC2), 136.6 (CCHCH-
COCH3 in ar), 141.2 (C6), 145.4 (C in Ph), 150.2 (C2), 158.8
(CH3OC), 163.8 (C4), 182.7 (ImCSO), 195.3 (MeCOS).
FAB-MS (m/e, relative intensity): 743 ([M + 1]+, 1), 742
(M+, 2), 557 (13), 414 (2), 303 (50), 289 (4), 279 (100), 201 (7),
154 (15), 136 (12), 107 (5), 89 (6), 77 (8).
Da ta for Meth yl-3,5-d id eoxy-6-ter t-bu tyld ip h en ylsilyl-
3-h yd r oxym eth yl-1-O-r- a n d -â-D-a llofu r a n osid e. 1H NMR
(CDCl3): δ 1.05 (s, 9H, 3CH3), 1.75-1.90 (m, 2H, H-5), 2.25
(m, 1H, H-3), 2.6-3.4 (br, 2H, 2OH), 3.22 (s, 3H, OCH3), 3.77-
3.92 (m, 4H, H-2, H-4, 2H-6), 4.25 (m, 2H, CH2OH), 4.73 (s,
1H, H-1), 7.33-7.46 (m, 6H, m-ar- and p-ar-H), 7.64-7.75 (m,
4H, o-ar-H).
13C NMR (CDCl3): δ 19.3 (SiCCH3), 27.2 (3CH3), 39.8
(HOCH2CH2), 48.3 (CH3O), 54.3 (C3), 60.1 (HOCH2), 61.7 (C3′),
77.7 (C4), 78.1 (C2), 109.2 (C1), 128.1 (o-ar-CH), 130.1 (p-ar-
CH), 134.0 (ar-C), 136.2 (m-ar-CH).
Anal. Calcd for C38H38N4O8S2 (742.87): C 61.44, H 5.16, N
7.54. Found: C 61.21, H 5.32, N 7.63.
1-{(2R,4R,5R)-4-Acetylth iom eth yl-5-[2-(4,4′-d im eth oxy-
tr iph en ylm eth oxy)eth yl]tetr ah ydr ofu r an -2-yl}u r acil (19).
Compound 33 (720 mg, 0.97 mmol), Bu3SnH (1.03 mL, 4
equiv), and AIBN (64 mg, 0.4 equiv) were dissolved in toluene
(9.7 mL). The mixture was slowly injected into refluxing dry
toluene (5 mL) under argon. This process lasted 3 min. The
solution was stirred for 2 min (monitored by TLC, 5% MeOH/

4010 J . Org. Chem., Vol. 67, No. 12, 2002
Huang and Benner
FAB-MS (m/e, relative intensity): 453 ([M + 23]+, 7), 399
(2), 341 (3), 303 (4), 273 (10), 255 (14), 213 (18), 199 (77), 197
(43), 183 (23), 153 (15), 135 (100), 105 (30), 91 (32), 74 (26).
Da ta for 35. 1H NMR (CDCl3): δ 1.05 (s, 9H, 3CH3), 1.70-
2.02 (m, 2H, H-5), 2.03 (s, 3H, CH3CO), 2.11 (s, 3H, CH3CO),
2.64 (m, 1H, H-3), 3.23 (s, 3H, OCH3), 3.88 (dd, J ) 5.8, 6.3,
2H, H-6), 4.08-4.31 (m, 3H, CH2COCH3, H-4), 4.77 (s, 1H,
H-1), 5.19 (d, J ) 4.8, H-2), 7.33-7.46 (m, 6H, m-ar- and p-ar-
H), 7.64-7.75 (m, 4H, o-ar-H).
2H, CH2O(DPTBS)), 4.04 (t, J ) 11.1, 1H, 4′-CH2OH), 4.32-
4.42 (m, 3H, H-3′, H-5′, HO)), 5.79 (s, 1H, H-2′), 5.85 (br, 1H,
HO), 7.48-7.55 (m, 8H, ar-H, H-5), 7.62 (m, 2H, ar-H), 7.70
(m, 4H, ar-H), 7.82 (d, J ) 7.5, 1H, H-6), 7.96 (m, 2H, ar-H).
13C NMR (100 MHz, CDCl3): δ 19.18 (SiCCH3), 26.90
(3CH3), 37.56 (HOCH2CH2), 49.16 (C4′), 58.61 (HOCH2), 60.90
(C4′′), 77.90 (C5′), 79.40 (C3′), 95.07 (C2′), 96.67 (C5), 127.76 (Si-
o-ar-CH), 129.02 (Si-p-ar-CH), 129.80, 129.84 (COPh-o-CH,
COPh-m-CH), 132.75 (Si-ar-C), 133.26 (COPh-p-CH), 133.58
(COPhC), 135.62 (Si-m-ar-CH), 143.49 (C6), 155.99 (C4), 162.71
(C2), 166.33 (PhCO).
13C NMR (CDCl3): δ 19.3 (SiCCH3), 21.0 (CH3CO), 27.2
(3CH3), 40.0 (HOCH2CH2), 45.1 (CH3O), 54.5 (C3), 60.6 (HOCH2),
61.2 (C3′), 77.5 (C4), 79.1 (C2), 107.2 (C1), 128.1 (o-ar-CH), 130.1
(p-ar-CH), 134.0 (ar-C), 136.2 (m-ar-CH), 170.2 (CH3CO), 171.0
(CH3CO).
FAB-MS (m/e, relative intensity): 636 ([M + 23]+, 2), 614
([M + 1]+, 9), 556 (3), 501 (2), 307 (3), 273 (2), 216 (100), 199
(15), 183 (4), 154 (16), 135 (25), 112 (35), 105 (43), 91 (9), 77
(15).
MS (m/e, relative intensity): 514 (M+, 0.4), 513 ([M - 1]+,
0.7), 483 ([M - 31]+, 1), 457 (45), 425 (63), 365 (30), 337 (60),
323 (30), 305 (60), 277 (10), 255 (19), 213 (50), 199 (100), 197
(30), 181 (57), 153 (15), 139 (50), 135 (30), 107 (14), 91 (10), 43
(50).
Anal. Calcd for C34H39N3O6Si (613.78): C 66.53, H 6.41, N
6.85. Found: C 66.38, H 6.30, N 6.93.
N4-Ben zoyl-1-[(2R,3R,4R,5R)-4-a cet ylt h iom et h yl-5-(2-
ter t-b u t yld ip h en ylsilyloxyet h yl)-3-h yd r oxyt et r a h yd r o-
fu r a n -2-yl]cytosin e (39). Diisopropyl azodicarboxylate (2.44
mL, 90-95%, 1.5 equiv) and triphenylphosphine (2.9 g, 1.5
equiv) were dissolved in dry THF (55 mL), and the mixture
was stirred at -20 °C (ice-salt bath) under argon. After a
yellow precipitate appeared, 37 (4.536 g, 7.4 mmol, dissolved
in dry THF, 15 mL) and thioacetate (790 µL, 1.5 equiv) in THF
(5 mL) were separately injected simultaneously. The temper-
ature was allowed to rise to rt, the mixture was stirred at rt
for 1 h (TLC, 7.5% MeOH/CH2Cl2, Rf ) 0.42), and the solvents
were removed under reduced pressure. The residue was
chromatographed on silica gel (from 2% to 3% MeOH/CH2Cl2)
to give 39 (5.14 g, high yield, containing triphenylphosphine
oxide) as a white foam.
Anal. Calcd for C28H38O7Si (514.69): C 65.34, H 7.44.
Found: C 65.13, H 7.37.
N4-Ben zoyl-1-[(2R,3R,4R,5R)-5-(2-ter t-bu tyld ip h en ylsi-
lyloxyeth yl)-3-h ydr oxy-4-h ydr oxym eth yltetr ah ydr ofu r an -
2-yl]cytosin e (37). A mixture of 35 (5.20 g, 10.1 mmol), N4-
isobutyrylcytosine (3.26 g, 1.5 equiv), and MSTFA (18.7 mL,
10 equiv) in acetonitrile (101 mL) was stirred until a clear
solution was formed (ca. 30 min). To this solution was added
SnCl4 (1.8 mL, 1.5 equiv), and the solution was stirred for 0.5
h at rt (monitored by TLC, 7.5% MeOH/CH2Cl2, Rf ) 0.40 for
product; EtOAc:hexane ) 2:8, Rf ) 0.32 for starting material).
The reaction mixture was cooled to 0 °C, and a NaHCO3
solution (10.3 g in 100 mL of water, 1.5 equiv to SnCl4) was
carefully added. The suspension was filtered through a layer
of sand, and the layer was then washed with EtOAc. The
filtrate was then extracted three times with EtOAc. The
combined organic layers were washed with NaCl(satd), dried
over MgSO4(s), and evaporated in vacuo. The crude product
was used directly for the next reaction. An analytical sample
was prepared by chromatography on silica gel (gradient from
EtOAc:hexane ) 3:7 to EtOAc). NaOH (2.0 g in 10 mL of water,
5 equiv) was added to the crude product (10.1 mmol) in MeOH
(101 mL). The mixture was stirred for 45 s at rt. The pH was
then adjusted to 6-7 with dilute HOAc (quickly, otherwise the
benzoyl amide will also be hydrolyzed). Monitoring by TLC
showed that the hydrolysis was complete (7.5% MeOH/CH2-
Cl2, Rf ) 0.31). MeOH was removed under reduced pressure,
and the residue was extracted with EtOAc (300 mL). The
organic phase was washed with aqueous NaCl(satd), dried over
MgSO4(s), and evaporated by rotary evaporation. The residue
was chromatographed on silica gel (gradient from 2% to 5%
MeOH/CH2Cl2) to give 37 (5.432 g, 88% yield over two steps)
as a white foam.
1H NMR (400 MHz, CDCl3): δ 1.08 (s, 9H, 3CH3), 1.84 (m,
1H, H-4′), 2.03 (s, 3H, CH3COS), 2.05-2.15 (m, 2H, CH2CH2O-
(DPTBS)), 3.91 (t, J ) 6.6, 2H, CH2O(DPTBS)), 4.14 (dd, J )
6.1, 11.4, 1H, 4′-CH3COSCH2), 4.32-4.52 (m, 3H, H-3′, H-5′,
4′-CH3COSCH2), 4.90 (br, 1H, HO), 5.71 (s, 1H, H-2′), 7.48-
7.55 (m, 8H, ar-H, H-5), 7.62 (m, 2H, ar-H), 7.70 (m, 4H, ar-
H), 7.80 (d, J ) 7.5, 1H, H-6), 7.96 (m, 2H, ar-H), 9.18 (br,
1H, NH).
13C NMR (CDCl3): δ 19.2 (SiCCH3), 20.8 (CH3COS), 26.9
(3CH3), 37.5 (HOCH2CH2), 45.8 (C4′), 60.4 (HOCH2), 60.9 (C4′′),
76.7 (C5′), 80.4 (C3′), 94.5 (C2′), 96.3 (C5), 127.6 (Si-o-ar-CH),
129.1 (Si-p-ar-CH), 129.7, 129.4 (COPh-o-CH, COPh-m-CH),
132.9 (Si-ar-C), 133.5 (COPh-p-CH), 133.3 (COPhC), 135.7 (Si-
m-ar-CH), 143.4 (C6), 155.8 (C4), 162.6 (C2), 166.3 (PhCO),
170.8 (COS).
FAB-MS (m/e, relative intensity): 656 ([M - 31]+, 24), 598
(6), 460 (3), 399 (4), 351 (2), 307 (30), 289 (15), 279 (32), 216
(86), 199 (17), 183 (5), 154 (100), 131 (81), 105 (48), 91 (20), 77
(30).
Anal. Calcd for C36H41N3O6SSi (671.89): C 64.35, H 6.16,
N, 6.25. Found: C 64.18, H 6.23, N 6.21.
Da ta for N4-Ben zoyl-1-[(2R,3R,4R,5R)-3-a cetoxy-4-a c-
et oxym et h yl-5-(2-ter t-b u t yld ip h en ylsilyloxyet h yl)t et -
r a h yd r ofu r a n -2-yl]cytosin e. 1H NMR (CDCl3): δ 1.05 (s,
9H, 3CH3), 1.80-2.12 (m, 2H, CH2CH2O(DPTBS)), 2.03 (s, 3H,
CH3CO), 2.14 (s, 3H, CH3CO), 2.47 (m, 1H, H-4′), 3.88 (t, J )
6.6, 2H, CH2O(DPTBS)), 4.05 (dd, J ) 5.8, 11.4, 1H, 4′-CH3-
COOCH2), 4.23 (dd, J ) 8.0, 11.2, 1H, 4′-CH3COOCH2), 4.18
(m, 1H, H-5′), 5.60 (dd, J ) 1.6, 6.1, 1H, H-3′), 5.78 (d, J )
1.5, 1H, H-2′), 7.33-7.70 (m, 13H, ar-H, H-5, H-6), 7.92 (m,
2H, ar-H), 8.9 (br, 1H, NH).
N4-Ben zoyl-1-[(2R,3R,4R,5R)-4-a cet ylt h iom et h yl-5-(2-
ter t-bu tyld ip h en ylsilyloxyeth yl)-3-(1-im id a zoth ioca r bo-
n yloxy)tetr a h yd r ofu r a n -2-yl]cytosin e (41). Compound 39
(ca. 6.6 mmol from the above reaction) and 1,1′-thiocarbonyl
diimidazole (1.77 g, 1.5 equiv) were dissolved in dry DMF (66
mL). The solution was stirred at rt for 1.5 h (TLC, 7.5% MeOH/
CH2Cl2, Rf ) 0.51), and the DMF was removed under 0.5 Torr
of pressure at 55 °C. The residue was chromatographed on
silica gel (gradient from EtOAc:hexane ) 8:2 to EtOAc:hexane
) 9:1) to give 41 (5.3 g, 83% based on 1H NMR analysis,
containing 20% triphenylphosphine oxide and imidazole) as a
foam. A sample for NMR analysis was prepared by preparative
TLC.
13C NMR (CDCl3): δ 19.3 (SiCCH3), 20.9 (2CH3CO), 27.0
(3CH3), 37.6 (HOCH2CH2), 45.2 (C4′), 59.7 (HOCH2), 60.8 (C4′′),
76.7 (C5′), 79.8 (C3′), 92.5 (C2′), 97.4 (C5), 128.1 (Si-m-ar-CH),
129.2 (Si-p-ar-CH), 130.0, 130.3 (CO-o-PhCH, CO-m-PhCH),
133.6 (CO-p-PhCH), 133.9 (Si-ar-C), 135.9 (Si-o-ar-CH), 144.3
(C6), 163.0 (PhCO), 169.5 (CH3CO), 170.8 (CH3CO).
FAB-MS (m/e, relative intensity): 698 ([M + 1]+, 52), 640
([M - 57]+, 37), 483 (37), 363 (7), 302 (12), 241 (23), 216 (73),
199 (45), 197 (53), 181 (17), 165 (15), 135 (100), 105 (94), 91
(16), 77 (26).
1H NMR (CDCl3): δ 1.08 (s, 9H, 3CH3), 1.80 (m, 2H, CH2-
CH2O(DPTBS)), 2.18 (s, 3H, CH3COS), 2.79 (m, 1H, H-4′), 2.94
(m, 2H, 4′-CH3COSCH2), 3.88 (t, J ) 4.7, 2H, CH2O(DPTBS)),
4.30 (m, 2H, H-3′, H-5′), 5.73 (s, 1H, H-2′), 7.48-7.55 (m, 9H,
ar-H, H-5, HIm-5), 7.62 (m, 2H, ar-H), 7.70 (m, 6H, ar-H, HIm
-
2,4), 7.74 (d, J ) 7.5, 1H, H-6), 7.96 (m, 2H, ar-H), 9.48 (br,
Da ta for 37. 1H NMR (400 MHz, CDCl3): δ 1.08 (s, 9H,
3CH3), 1.84 (m, 1H, H-4′), 1.90-2.05 (m, 2H, CH2CH2O-
(DPTBS)), 3.70 (br d, J ) 8.3, 1H, 4′-CH2OH), 3.91 (t, J ) 8.3,
1H, NH).
13C NMR (CDCl3): δ 19.2 (SiCCH3), 23.8 (C4′′), 26.7 (3CH3),
30.2 (CH3COS), 36.8 (HOCH2CH2), 47.2 (C4′), 60.7 (HOCH2),

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 4011
76.3 (C5′), 80.9 (C3′), 94.4 (C2′), 96.7 (C5), 127.8, 128.6 (Si-o-ar-
CH, ImC5), 129.3 (Si-p-ar-CH), 129.6, 129.9 (COPh-o-CH,
COPh-m-CH), 131.2, 132.9, 133.4 (Si-ar-C, ImC4, ImC2), 134.9
(COPh-p-CH), 135.7 (Si-m-ar-CH), 143.7 (C6), 155.6 (C4), 162.4
(C2), 166.3 (PhCO), 170.8 (COS), 196.0 (OCSIm).
An Alter n a tive Ap p r oa ch . Compounds 47 (100 mg, 0.153
mmol) and 19 (113 mg, 1.2 equiv) were dissolved in MeOH
(150 µL, degassed) and THF (1.5 mL, degassed) under argon.
DBU (45 µL, 2 equiv) was added, and the solution was allowed
to stand at rt overnight (TLC, 7.5% MeOH/CH2Cl2, Rf ) 0.45).
THF/MeOH (1 mL, 1:1) and DTE (7 mg, 0.3 equiv) were added,
and the mixture was stirred for 5 min. The pH of the mixture
was adjusted with dilute aqueous HOAc to ca. pH 7, the
solvents were removed under reduced pressure, and the resi-
due was washed with water (5 × 2 mL) to remove DTE and
salts. The crude product was dried under a vacuum, dissolved
in CH2Cl2, and purified by chromatography on silica gel
(gradient 0.5-2% MeOH/CH2Cl2 containing 0.5% TEA) to give
dimer 51 (161 mg, 93%, cannot be reduced with DTE) as a
white foam.
(DMTr )O-USU-SH (57). A solution of AgNO3 (186 mg, 5
equiv) in water/MeOH (1:2, 1.5 mL) was added with rapid
stirring to 51 (250 mg, 0.221 mmol) in THF/MeOH (8 mL, 1:1)
containing NaOAc (750 µL, 3 M). After 1 min, the mixture was
centrifuged, and the pellet was washed twice with MeOH (10
mL) and recovered each time by centrifugation. The pellet was
then suspended in THF/MeOH (8 mL, 1:1) with NaOAc (750
µL, 3 M), and DTE (339 mg, 2 equiv to Ag+) was added. The
mixture was stirred for 5 min (TLC, 7.5% MeOH/CH2Cl2, Rf
) 0.33). The suspension was filtered through a sand-silica
layer to remove the precipitate ((DTE)Ag), the sand-silica layer
was washed with THF/MeOH (1:1) and the combined filtrates
were evaporated under reduced pressure. The residue was
washed with water (5 × 5 mL) to remove DTE and salts and
dried under a high vacuum to give 57 (177 mg, 97% yield; the
purity was higher than 95%) as a white foam. A sample for
NMR analysis was prepared by TLC purification (silica gel,
10% MeOH/CH2Cl2).
FAB-MS (m/e, relative intensity): 672 ([M - 127]+, 16), 628
(4), 614 (6), 557 (6), 469 (2), 414 (3), 379 (2), 321 (3), 279 (98),
216 (100), 199 (32), 183 (15), 154 (22), 13 (55), 105 (68), 91
(16), 77 (26).
Anal. Calcd for C40H43N5O6S2Si (782.03): C 61.43, H 5.55,
N 8.96. Found: C 61.19, H 5.51, N 8.88.
N4-Ben zoyl-1-[(2R,4R,5R)-4-a cetylth iom eth yl-5-(2-ter t-
b u t yld ip h e n ylsilyloxye t h yl)t e t r a h yd r ofu r a n -2-yl]cy-
tosin e (43). Compound 41 (124 mg, 0.155 mmol), Bu3SnH (164
µL, 4 equiv), and AIBN (10 mg, 0.4 equiv, 10%) were dissolved
in toluene (1.5 mL). The mixture was slowly injected into
refluxing dry toluene (1.5 mL) under argon. The solution was
stirred for 2 min (TLC, 5% MeOH/CH2Cl2, Rf ) 0.32) and the
toluene evaporated in vacuo. The crude product was purified
on preparative TLC (silica gel, 5% MeOH/CH2Cl2) to give 43
(89 mg, 86%) as a white foam.
1H NMR (CDCl3): δ 1.08 (s, 9H, 3CH3), 1.80-2.0 (m, 2H,
CH2CH2O(DPTBS)), 2.05 (s, 3H, CH3COS), 2.18-2.50 (m, 3H,
H-3′, H-4′), 3.88 (dd, J ) 6.8, 13.8, 2H, CH2O(DPTBS)), 4.10
(m, 3H, 4′-CH3COSCH2, H-5′), 5.98 (dd, J ) 2.8, 6.9, 1H, H-2′),
7.38-7.55 (m, 8H, ar-H, H-5), 7.62 (m, 2H, ar-H), 7.70 (m, 4H,
ar-H), 7.74 (d, J ) 7.5, 1H, H-6), 7.96 (m, 2 H, ar-H), 8.75 (br,
1H, NH).
13C NMR (CDCl3): δ 19.19 (SiCCH3), 20.74 (CH3COS), 26.88
(3CH3), 37.04 (C3′), 37.55 (HOCH2CH2), 41.43 (C4′), 60.79
(HOCH2), 63.77 (C4′′), 77.22 (C5′), 80.93 (C2′), 87.40 (C5), 128.44
(Si-o-ar-CH), 129.10 (Si-p-ar-CH), 129.4, 129.9 (COPh-o-CH,
COPh-m-CH), 133.22 (Si-ar-C), 133.58 (COPh-p-CH), 135.57
(Si-m-ar-CH), 143.7 (C6), 155.9 (C4), 162.2 (C2), 170.5 (COS).
Anal. Calcd for C36H41N3O5SSi (655.89): C 65.92, H 6.30,
N 6.41. Found: C 65.79, H 6.38, N 6.50.
(DMTr )O-USU-S(DMTr ) (51). DBU (144 µL, 2 equiv)
was added to a solution of 47 (315 mg, 0.483 mmol) and 44
(332 mg, 0.579 mmol, 1.2 equiv) in THF (degassed, 4.8 mL)
under an argon atmosphere. The mixture was allowed to stand
overnight (TLC, 7.5% MeOH/CH2Cl2, Rf ) 0.45; the dimer and
the disulfide are indistinguishable in TLC), THF/MeOH (5 mL,
1:1) and DTE (35 mg, 0.3 eq) were added, and the mixture
was stirred for 5 min. The pH of the solution was adjusted
with dilute aqueous HOAc to ca. 7, all solvents were removed
under reduced pressure, and the residue was washed with
water (5 × 5 mL) to remove DTE and salts. After being dried
under a high vacuum, the crude product dissolved in CH2Cl2
was purified by chromatography on silica gel (gradient, from
0.5% to 2% MeOH/CH2Cl2 containing 0.5% TEA) to give 51
(524 mg, 96%) as a white foam.
1H NMR (400 MHz, CDCl3): δ 1.73-1.92 (m, 2H, CH2CH2S),
1.95-2.08 (m, 2H, CH2CH2ODMTr), 2.12-2.30 (m, 6H, 4H-
3′, 2H-4′), 2.48-2.67 (m, 6H, 2(4′-CH2S), CH2CH2S), 3.28 (t, J
) 6.3, 2H, CH2O(DMTr)), 3.75 (s, 6H, 2CH3), 3.84 (td, J ) 2.9,
9.1, 1H, H-5′), 3.93 (m, 1H, H-5′), 5.66, 5.74 (d, d, J ) 8.1, 8.1,
2H, H-5), 5.97, 6.02 (dd, dd, J ) 3.8, 6.9; 4.4, 5.6, 2H, H-2′),
6.82 (d, J ) 8.9, 4H, 4CHCOCH3), 7.22-7.32 (m, 8H, ar-H,
H-6), 7.34 (d, J ) 8.1, 1H, H-6), 7.42 (m, 2H, ar-H), 8.8 (br,
2H, NH).
13C NMR (100 MHz, CDCl3): δ 25.95 (CH2CH2S), 29.87
((DMTr)OCH2CH2), 34.47, 34.53 (C3′), 34.87 (CH2CH2S), 38.10,
38.79 (C4′′), 42.74, 45.57 (C4′), 55.26 (CH3O), 60.28 ((DMTr)-
OCH2), 82.56, 82.72 (C5′), 85.25, 85.41 (C2′), 86.39 (ar3-CO),
102.07, 102.44 (C5), 113.14 (CHCOCH3 in ar), 126.83 (CHCH-
CHC- in Ph), 127.84 (CHC- in Ph), 128.11 (CHCHC- in Ph),
130.00 (CHCHCOCH3 in ar), 136.18 (CCHCHCOCH3 in ar),
139.30, 139.36 (C6), 144.97 (C in Ph), 150.05, 150.10 (C2),
158.51 (CH3OC), 162.91, 163.03 (C4).
1H NMR (400 MHz, CDCl3): δ 1.55-1.67, 1.73-1.90 (2m,
4H, CH2CH2O(DMTr), CH2CH2S), 1.98-2.32 (m, 8H, 4H-3′,
2H-4′, CH2CH2S), 2.42-2.55 (m, 2H, 4′-CH2S), 2.60-2.68 (m,
2H, 4′-CH2S), 3.28 (t, J ) 6.2, 2H, CH2O(DMTr)), 3.58 (td, J
) 2.8, 9.2, 1H, H-5′), 3.77 (s, 12H, 4CH3), 3.87 (m, 1H, H-5′),
5.67, 5.72 (d, d, J ) 8.1, 8.1, 2H, H-5), 5.91, 5.95 (dd, dd, J )
4.0, 6.4; 4.0, 6.2, 2H, H-2′), 6.80, 6.81 (d, d, J ) 9.2, 9.2, 8H,
8CHCOCH3), 7.15 (d, J ) 8.2, 1H, H-6), 7.20-7.35 (m, 15H,
ar-H), 7.40 (m, 3H, ar-H), 7.42 (d, J ) 8.2, 1H, H-6), 8.3 (br,
2H, NH).
Anal. Calcd for C43H48N4O9S2 (828.99): C 62.29, H 5.84, N
6.76. Found: C 62.06, H 5.68, N 6.92.
HO-USO2USO2-Bn (65). H2O2 (30%, 198 µL, 10 equiv
calculated for converting the sulfide to the sulfone) was added
to 59 (27 mg, 44 mmol) in MeOH (4.4 mL), and the mixture
stirred for 10 min to allow the sulfoxides to form. TS-2 (54
mg) was then added to catalyze the oxidation of the sulfoxide
groups to sulfone groups. The reaction was monitored by HPLC
(C4 column, 6.5 µm, 250 × 4.6 mm, CH3CN:H2O ) 20:80). After
72 h, essentially all of the sulfoxide groups had been oxidized
to sulfones; only a single peak was detectable by HPLC. In
the elution from the HPLC column, sulfoxide eluted first
followed by the two sulfoxide-sulfone intermediates (unas-
signed) followed by the disulfone. The catalyst was removed
by filtration and washed with MeOH (3 × 2 mL), the combined
filtrates were evaporated to remove MeOH, and the hydrogen
peroxide was removed under a high vacuum to give 65 (25 mg,
84%) as a foam, which dissolves in acetone, MeOH, and water.
1H NMR (CD3OD:CD3COCD3 ) 25:75): δ 1.90-2.15 (m,
CH2CH2SO2, CH2CH2OH, overlapped with CD3COCHD2), 2.30-
2.50 (m, 4H, 4H-3′), 2.52-2.72 (m, 2H, 2H-4′), 3.10-3.50 (2m,
6H, CH2CH2SO2, 2(4′-CH2SO2)), 3.60-3.80 (m, 2H, CH2OH),
3.90-4.05 (m, H-5′, partially overlapped with CD3OH), 4.48
(s, 2H, SO2CH2Ph), 5.63, 5.66 (d, d, J ) 8.1, 8.1, 2H, H-5),
13C NMR (100 MHz, CDCl3): δ 29.84 (CH2CH2S), 33.44
((DMTr)OCH2CH2), 34.20, 34.45 (C3′), 34.83 (CH2CH2S), 38.80,
38.88 (C4′′), 42.24, 42.72 (C4′), 55.24, 55.28 (4CH3O), 60.30
((DMTr)OCH2), 66.32 (ar3-CS), 82.55, 83.48 (C5′), 84.89, 85.36
(C2′), 86.37 (ar3-CO), 102.02, 102.27 (C5), 113.12, 113.27 (d,
CHCOCH3 in ar), 126.81 (CHCHCHC- in Ph), 127.83, 128.00
(CHC- in Ph), 128.09, 129.33 (d, CHCHC- in Ph), 129.98,
130.63 (CHCHC-OCH3 in ar), 136.18, 136.68, 136.71 (CCH-
CHCOCH3 in ar), 139.24, 139.29 (C6), 145.00 (C in Ph), 150.02,
150.07 (C2), 158.23, 158.49 (CH3OC), 162.99, 163.10 (C4). FAB
-MS (m/e, relative intensity): 1154 ([M + 23 + 1]+, 0.2), 1129
([M - 1]+, 0.1) 577 (0.3), 523 (0.2), 473 (0.5), 454 (1), 409 (1),
369 (1), 303 (100), 289 (8), 273 (6), 195 (7), 176 (9), 155 (52),
136 (58), 105 (38), 91 (30), 77 (40).

4012 J . Org. Chem., Vol. 67, No. 12, 2002
Huang and Benner
6.06, 6.09 (d, d, J ) 6.7, 6.7, 2H, H-2′), 7.35-7.50 (m, 5H, Ph),
7.63, 7.68 (d, d, J ) 8.1, 8.2, 2H, H-6).
85.15, 85.72 (C2′), 102.51, 102.67 (C5), 139.68, 139.76 (C6),
150.36, 150.57 (C2), 163.30 (C4).
FAB-MS (m/e, relative intensity): 549 ([M + 23]+, 7), 513
(1), 413 (25), 281 (7), 218 (13), 199 (8), 149 (40), 135 (40), 107
(35), 91 (45), 77 (35), 55 (100).
13C NMR (100 MHz, CD3OD:CD3COCD3 ) 25:75): δ 24.37
(CH2CH2SO2), 34.77 (HOCH2CH2), 35.33, 35.43 (C4′), 36.44,
36.96 (C3′), 48.84 (CH2Ph), 51.89, 52.72 (C4′′), 57.35, 58.17
(CH2CH2SO2, HOCH2), 80.57, 81.05 (C5′), 84.03, 84.04 (C2′),
100.44, 100.74 (C5), 116.88 (p-ar-CH), 127.38 (C in Ph), 127.54
(m-PhCH), 129.92 (o-PhCH), 139.60, 139.89 (C6), 149.34,
149.39 (C2), 162.66, 162.76 (C4).
MsO-USU-S(DMTr ) (53). Methanesulfonyl chloride (30
µL, 1.5 equiv) was added to a solution of 52 (211 mg, 0.255
mmol) and TEA (355 µL, 10 equiv) in THF (2.6 mL) at 0 °C
under argon. After 20 min (TLC, monitoring, 7.5% MeOH/CH2-
Cl2, Rf ) 0.30), MeOH (1 mL) was added, and stirring at rt
was continued for 2 h. The solvents were then removed in
vacuo, and the residue was dissolved in CH2Cl2 and resolved
by chromatography on silica gel (gradient from 0.5% to 3%
MeOH/CH2Cl2) to give 53 (208 mg, 90%) as a white foam.
1H NMR (400 MHz, CDCl3): δ 1.55-1.65 (m, 2H, CH2CH2-
OMs), 1.75-1.85, 1.95-2.10 (2m, 4H, CH2CH2S, 2H-4′), 2.23-
2.33 (m, 6H, 4H-3′, CH2CH2S), 2.47-2.57, 2.60-2.68 (2m, 4H,
2(4′-CH2S)), 3.02 (s, 3H, CH3), 3.59 (dt, J ) 2.6, 9.2, 1H, H-5′),
3.78 (s, 6H, 2CH3), 3.82 (m, 1H, H-5′), 4.30-4.45 (2m, 2H, CH2-
OMs), 5.73 (m, 2H, H-5), 5.92, 6.00 (dd, t, J ) 4.1, 6.6; 5.3,
2H, H-2′), 6.81 (d, J ) 9.0, 4H, 4CHCOCH3), 7.19 (d, J ) 8.2,
1H, H-6), 7.20-7.32 (m, 7H, ar-H), 7.35 (d, J ) 8.1, 1H, H-6),
7.40 (m, 2H, ar-H), 8.80 (br, 2H, NH).
FAB-MS (m/e, relative intensity): 703 ([M + 23]+, 2), 681
([M + 1]+, 4), 663 ([M - 18]+, 1), 569 (3), 551 (1), 522 (1), 399
(15), 375 (24), 329 (10), 276 (17), 107 (32), 91 (25), 77 (25).
Anal. Calcd for C29H36N4O11S2 (680.75): C 51.16, H 5.33, N
8.24. Found: C 51.05, H 5.48, N 8.16.
Solubility: in MeOH, acetone, poor in dichloromethane, in
pure water (maximum 40 mM).
UV (in water): λmax ) 262.8 nm.
HO-USU-S(DMTr ) (52). Compound 51 (303 mg, 0.268
mmol) was dissolved in HOAc (2 mL, 80%) and the mixture
stirred for 20 min (10% MeOH/CH2Cl2, Rf ) 0.38). Na2CO3-
(satd) was added to adjust the pH of the solution to 6-7. A
precipitate of 52 was formed. The suspension was then filtered
through a layer of sand and the sand washed with water (10
mL). The precipitate was eluted from the sand with THF, and
the solvents were removed from the eluate by rotary evapora-
tion. The residue was dried under a high vacuum, dissolved
in CH2Cl2, and resolved by chromatography on a short column
of silica gel to remove (DMTr)OH. The column was washed
with EtOAc/hexane (3:7) until no (DMTr)OH eluted. The
product was then eluted with EtOH/EtOAc (1:9). After evapo-
ration of the solvents, 52 (215 mg, 97%) was obtained as a
white foam.
1H NMR (400 MHz, CDCl3): δ 1.55-1.65 (m, 2H, CH2CH2-
OH), 1.73-1.90, 1.98-2.23 (2m, 5H, CH2CH2S, 2H-4′, OH),
2.17-2.33 (m, 6H, 4H-3′, CH2CH2S), 2.45-2.55 (m, 2H, 4′-
CH2S), 2.60-2.68 (m, 2H, 4′-CH2S), 3.59 (dt, J ) 2.6, 9.2, 1H,
H-5′), 3.78 (s, 6H, 2CH3), 3.80 (m, 2H, CH2OH), 3.85 (m, 1H,
H-5′), 5.73, 5.75 (d, d, J ) 8.2, 8.1, 2H, H-5), 5.92, 6.03(dd, t,
J ) 3.9 6.9; 5.0, 2H, H-2′), 6.81 (d, J ) 9.2, 4H, 4CHCOCH3),
7.17 (d, J ) 8.2, 1H, H-6), 7.20-7.32 (m, 7H, ar-H), 7.40 (m,
2H, ar-H), 7.42 (d, J ) 8.1, 1H, H-6), 9.27 (br, 2H, NH).
13C NMR (100 MHz, CDCl3): δ 29.62 (CH2CH2S), 33.64
(HOCH2CH2), 34.27, 34.34 (C3′), 36.79 (CH2CH2S), 38.61, 38.67
(C4′′), 42.37, 42.91 (C4′), 55.34 (CH3O), 60.09 (HOCH2), 66.41
(ar3-CS), 83.39, 83.49 (C5′), 85.05, 85.62 (C2′), 102.39, 102.41
(C5), 113.37 (CHCOCH3 in ar), 126.81 (CHCHCHC- in Ph),
128.03 (CHC- in Ph), 129.39 (CHCHC- in Ph), 130.69
(CHCHCOCH3 in ar), 136.81 (CCHCHCOCH3 in ar), 139.46,
139.50 (C6), 145.07 (C in Ph), 150.19, 150.21 (C2), 158.34
(CH3OC), 163.10 (C4).
Anal. Calcd for C44H50N4O11S3 (907.09): C 58.26, H 5.56, N
6.18. Found: C 58.08, H 5.51, N 6.23.
(DMTr )O-USUSUSU-S(DMTr ) (68). DBU (45 µL, 1.5
equiv) was added under argon to a solution of 53 (139 mg, 153
µmol) and 57 (152 mg, 1.2 equiv) in THF (1.5 mL, degassed).
The mixture was allowed to stand at rt for 24 h (TLC, 7.5%
MeOH/CH2Cl2, Rf ) 0.23; the tetramer and the disulfide are
not distinguishable in TLC). THF/MeOH (5 mL, 4:1) and DTE
(7 mg, 0.3 equiv) were added to reduce the disulfide, the
mixture was incubated for 5 min, and the pH of the solution
was adjusted with dilute HOAc to ca. 7. The solvents were
removed under reduced pressure, and the residue was washed
with water (4 × 5 mL) to remove DTE and salts. After being
dried under a high vacuum, the crude product was dissolved
in CH2Cl2 and purified by chromatography on silica gel
(gradient from 1% to 5% MeOH/CH2Cl2) to give 68 (231 mg,
92%, cannot be reduced by DTE) as a white foam.
1H NMR (400 MHz, CDCl3): δ 1.75-1.90 (m, 4H, CH2-
CH2O(DMTr), CH2CH2S), 1.95-2.18, 2.22-2.31 (m, 16H, 8H-
3′, 4H-4′, 2CH2CH2S), 2.45-2.88 (m, 14H, 4(4′-CH2S), 3CH2-
CH2S), 3.28 (t, J ) 6.2, 2H, CH2O(DMTr)), 3.58 (td, J ) 2.8,
9.2, 1H, H-5′), 3.76, 3.78 (2s, 12H, 4CH3), 3.73-3.95 (m, 3H,
H-5′), 5.78 (d, J ) 8.1, 1H, H-5), 5.72-5.78 (m, 3H, H-5), 5.91-
6.03 (m, 4H, H-2′), 6.80, 6.81 (d, d, J ) 9.2, 9.2, 8H,
8CHCOCH3), 7.14 (d, J ) 8.2, 1H, H-6), 7.18-7.24, 7.25-7.35
(3m, 21H, 18(ar-H)), 7.38-7.45 (m, 3H, 3H-6), 9.23, 9.40, 9.53
(3br, 4H, NH).
13C NMR (100 MHz, CDCl3): δ (intensity) 29.66 (0.72), 29.83
(1.29) (CH2CH2S), 33.53 (0.70) ((DMTr)OCH2CH2), 34.24 (2.21)
(C3′), 34.50 (1.70), 34.87 (0.60) (CH2CH2S), 38.35 (0.61), 38.42
(0.73), 38.64 (0.63), 38.93 (0.62) (C4′′), 42.48 (0.83), 42.67 (0.92),
42.72 (1.70) (C4′), 55.26, 55.32 (4CH3O), 60.32 (0.82) ((DMTr)-
OCH2), 67.98 (0.67) (ar3-CS), 82.74 (0.81), 83.26 (0.82), 83.32
(0.90), 83.39 (0.85) (C5′), 84.63 (0.74), 85.37 (0.75), 85.49 (0.81),
85.71 (0.77) (C2′), 86.36 (0.71) (ar3-CO), 102.00 (0.82), 102.39
(0.87), 102.53 (0.78), 102.65 (0.75) (C5), 113.12, 113.30 (d,
CHCOCH3 in ar), 126.80 (CHCHCHC- in Ph), 127.84, 128.02
(CHC- in Ph), 128.11, 129.34 (d, CHCHC- in Ph), 129.99,
130.65 (CHCHCOCH3 in ar), 135.81, 136.19, 136.72 (CCHCH-
COCH3 in ar), 139.34 (0.65), 139.52 (0.75), 139.57 (0.76), 139.67
(0.60) (C6), 144.99 (C in Ph), 150.22 (1.15), 150.40 (0.80), 150.44
(0.86) (C2), 158.23, 158.49 (CH3OC), 163.18 (1.01), 163.31
(0.78), 163.54 (0.67) (C4).
(DMTr )O-USUSUSU-SH (71). To a solution of 68 (60
mg, 36.8 µmol) in THF/MeOH (4 mL, 1:1) containing NaOAc
(244 µL, 3 M) was added AgNO3 (62 mg, 10 equiv) in water
(200 µL) and MeOH (1 mL). The mixture was stirred for 2 min,
and then centrifuged. The pellet was washed twice with MeOH
(5 mL) and recovered by centrifugation. The pellet was then
suspended in THF/MeOH (4 mL, 1:1), NaOAc (244 µL, 3 M)
and DTE (114 mg, 2 equiv to Ag+) were added, and the mixture
was stirred for 5 min (7.5% MeOH/CH2Cl2, Rf ) 0.20). The
Anal. Calcd for C43H48N4O9S2 (828.99): C 62.29, H 5.84, N
6.76. Found: C 62.05, H 5.92, N 6.67.
HO-USU-SH (63). To a solution of 52 (10 mg) in THF/
MeOH (1 mL, 1:1) containing NaOAc (100 µL, 3 M) was added
AgNO3 (5 mg, 4 equiv) in water (20 µL) and MeOH (1 mL).
After 2 min, the mixture was centrifuged and the pellet washed
three times with MeOH (2 mL). The pellet was suspended in
THF/MeOH (1 mL, 1:1) with NaOAc (100 mL, 3 M), DTE (10
mg, 2 equiv to Ag+) was added, and the mixture was stirred
for 5 min. The solid ((DTE)Ag) was removed by filtration
through a layer of sand-silica gel, and the sand was washed
with THF/MeOH (1:1, 4 mL). The washings were evaporated
in vacuo, and the residue was dissolved in 5% MeOH/CH2Cl2
and resolved by TLC (silica gel, 12.5% MeOH/CH2Cl2, Rf )
0.44) to give 63 (2 mg, 32%), which is water soluble.
1H NMR (400 MHz, CD3OD): δ 1.73-2.00 (m, 2H, CH2-
CH2S), 2.02-2.08 (m, 2H, CH2CH2OH), 2.22-2.32 (m, 4H, 4H-
3′), 2.48-2.67 (m, 8H, 2(4′-CH2S), CH2CH2S, 2H-4′), 3.72 (m,
2H, 2H-5′), 3.88 (m, 2H, CH2OH), 5.70, 5.72 (d, d, J ) 6.1, 6.1,
2H, H-5), 6.06 (m, 2H, H-2′), 7.65, 7.67 (d, d, J ) 6.2, 6.2, 2H,
H-6).
13C NMR (100 MHz, CDCl3): δ 32.64 (CH2CH2S), 34.84
(HOCH2CH2), 34.77, 34.84 (C3′), 37.09 (CH2CH2S), 38.81, 38.97
(C4′′), 42.56, 43.01 (C4′), 60.21 (HOCH2), 83.54, 83.79 (C5′),

DNA Analogues with Bridging Sulfur Groups
J . Org. Chem., Vol. 67, No. 12, 2002 4013
UV (in water): λmax ) 260.8 nm.
precipitate ((DTE)Ag) was removed by filtration through a
layer of sand and silica gel, and the sand and the silica gel
were then washed with THF/MeOH (1:1). The combined
washings were evaporated in vacuo. The residue was washed
with water several times to remove DTE and salts, and then
dried under a high vacuum to give 71 (42 mg, 85%).
UV (MeOH): λmax ) 262.8 nm.
MsO-USUSUSU-S(DMTr ) (70). Methanesulfonyl chlo-
ride (3.4 µL, 3 equiv) was added to the solution of 69 (ca. 15
µmol) and TEA (21 µL, 10 equiv) in THF (200 µL) at 0 °C under
argon. After 20 min (TLC, 7.5% MeOH/CH2Cl2, Rf ) 0.22),
MeOH (0.1 mL) was added, the mixture was stirred at rt for
2 h, and the solvents were removed in vacuo. The residue was
dissolved in CH2Cl2 and then resolved by preparative TLC
(silica gel, 10% MeOH/CH2Cl2) to give 70 (17 mg, 81% over
two steps) as a white foam.
1H NMR (CD3OD:CDCl3 ) 5:95): δ 1.75-1.90, 1.95-2.15,
2.20-2.40 (m, 20H, CH2CH2OMs, 8H-3′, 4H-4′, 3CH2CH2S),
2.45-2.83 (m, 14H, 4(4′-CH2S), 3CH2CH2S), 3.04 (s, 3H, CH3),
3.58 (m, 1H, H-5′), 3.77 (s, 6H, 2CH3), 3.80-3.92 (m, 3H, H-5′),
4.39 (m, 2H, CH2OMs), 5.72 (d, J ) 7.8, 4H, H-5), 5.92-6.08
(m, 4H, H-2′), 6.81 (d, J ) 9.2, 4H, 4CHCOCH3), 7.14 (d, J )
7.7, 1H, H-6), 7.20-7.49 (m, 12H, 9(ar-H), 3H-6).
1H NMR (CD3OD:CDCl3 ) 5:95): δ 1.80-2.40 (m, 20H, CH2-
CH2O(DMTr), 8H-3′, 4H-4′, 3CH2CH2S), 2.45-2.88 (m, 14H,
4(4′-CH2S), 3CH2CH2S), 3.28 (t, J ) 6.2, 2H, CH2O(DMTr)),
3.58 (m, 1H, H-5′), 3.77 (s, 6H, 2CH3), 3.80-3.95 (m, 3H, H-5′),
5.65 (d, J ) 8.1, 1H, H-5), 5.69-5.78 (m, 3H, H-5), 5.90-6.05
(m, 4H, H-2′), 6.81 (d, J ) 9.2, 4H, 4CHCOCH3), 7.14-7.49
(m, 13H, 9(ar-H), 4H-6).
HO-USUSUSU-SBn (72). A solution of 71 (40 mg, 30
µmol) in THF (600 µL) under argon was treated with TEA (84
µL, 20 equiv) and benzyl bromide (36 µL, 10 equiv) for 20 min.
The THF was removed in vacuo, water (1 mL) was added to
precipitate 72, and the unreacted BnBr was removed by
extraction with toluene (3 × 2 mL). The suspension was
filtered through cotton to collect the precipitate, which was
washed with water. The crude product was eluted from the
cotton with THF, the solvents were evaporated in vacuo, and
HOAc (1 mL, 80%) was added to the residue to remove the
DMTr group. After 20 min, Na2CO3(satd) was added to
precipitate the product and adjust the pH of the solution to
7-8 (pH paper), and the (DMTr)OH was extracted using ether/
toluene (1:1, 3 × 2 mL). The precipitate was recovered by
filtration through a layer of sand, the sand washed with water,
and the product eluted with THF. Following the evaporation
of THF, the crude product was dissolved in 5% MeOH/CH2Cl2
and purified by TLC (silica gel, 12.5% MeOH/CH2Cl2, Rf )
0.23) to give 72 (30 mg, 89% yield) as a white foam.
(DMTr )O-USUSUSUSUSUSUSU-S(DMTr ) (83). DBU
(2.9 µL, 2.0 equiv) was added under argon to a solution of 70
(9.6 mg, 6.7 µmol) and 71 (13 mg, 1.5 equiv) in 10% MeOH/
THF (200 µL, degassed). The mixture was allowed to stand at
rt for 24 h (TLC, 10% MeOH/CH2Cl2, Rf ) 0.32; the octamer
and the disulfide of the tetramer are not distinguishable in
TLC). THF/MeOH (1 mL, 4:1) and DTE (1 mg) were added,
the mixture was incubated for 5 min, and the pH of the
solution was adjusted with dilute HOAc to ca. 7. The solvents
were removed under reduced pressure, and the residue was
dissolved in 5% MeOH/CH2Cl2 and then resolved by TLC
(12.5% MeOH/CH2Cl2) to give 83 (10.1 mg, 55%).
1H NMR (CD3OD:CDCl3 ) 1:9): δ 1.75-1.90, 1.93-2.10,
2.13-2.30 (3m, 40H, CH2CH2O(DMTr), 16H-3′, 8H-4′, 7CH2-
CH2S), 2.45-2.90 (m, 28H, 8(4′-CH2S), 6CH2CH2S), 3.28 (t, J
) 6.2, 2H, CH2O(DMTr)), 3.62-3.88 (m, 8H, 8H-5′), 3.78 (s,
12H, 4CH3), 5.62 (d, J ) 7.9, 1H, H-5), 5.68-5.76 (m, 7H, H-5),
5.87-6.05 (m, 8H, H-2′), 6.84 (d, J ) 9.2, 8H, 8CHCOCH3),
7.14-7.40 (m, 26H, 18(ar-H), 8H-6).
1H NMR (400 MHz, CD3OD:CDCl3 ) 10:90): δ 1.65-1.88,
1.90-2.03, 2.05-2.15, 2.19 (4m, 20H, 8H-3′, 4H-4′, CH2CH2-
OH, 3CH2CH2S), 2.38 (dd, J ) 7.5, 12.9, 1H, 0.5(4′-CH2S)),
2.43-2.72 (m, 13H, 3.5(4′-CH2S), 3CH2CH2S), 3.63-3.74 (m,
8H, CH2OH, 4H-5′, CH2Ph), 5.66 (intensity, 10.46), 5.68 (6.18),
5.69 (5.97) (d, d, d, J ) 8.1, 8.0, 8.0, 4H, H-5), 5.89-5.96 (m,
4H, H-2′), 7.21 (d, J ) 8.1, 1H, H-6), 7.23-7.27 (m, 5H, Ph-
H), 7.34, 7.35, 7.43 (d, d, d, J ) 8.1, 8.2, 8.1, 3H, 3H-6).
FAB-MS (m/e, relative intensity): 1147 ([M + 23]+,0.2) 1125
([M + 1]+, 0.3) 790 (1), 677 (6), 550 (13), 399 (43), 329 (27).
UV (MeOH): λmax ) 262.8 nm.
H O -U S O 2U S O 2U S O 2U S O 2U S O 2U S O 2U S O 2U -S O 2-
O-K+ (90). To a solution (pH ca. 7) of Oxone (2KHSO5-KHS-
O
4-K2SO4, 12 mg, 3 equiv to sulfur) and K2CO3 (3.7 mg, 1.5
HO-USO2USO2USO2U-SO2Bn (79). H2O2 (30%, 6 µL, 3
equiv from the sulfide to the sulfoxides) was added to 72 (5.5
mg, 4.9 µmol) in MeOH/THF (5 mL, 1:1), and the mixture was
stirred for 5 min to allow conversion of the sulfides in 72 to
the sulfoxides. The solvents were evaporated in vacuo, and the
solid was redissolved in MeOH (50 mL, with heating). TS-2
(100 mg) and H2O2 (30%, 88 µL, 40 equiv from the sulfoxides
to the sulfone) were then added (HPLC monitoring, C4 column,
6.5 µm, 250 × 4.6 mm, CH3CN:H2O ) 20:80; the more highly
oxidized products eluted later). After 6 days, nucleotide
analogues were nearly undetectable in the MeOH solution;
they were absorbed onto the catalyst. The reaction mixture
was centrifuged, the MeOH solution was decanted, and the
catalyst was extracted with CH3CN/H2O (3 × 5 mL). The
combined extracts were evaporated in vacuo. The residue was
redissolved in CH3CN/H2O (1:1) and resolved by HPLC (CN
column, 23% CH3CN in H2O as eluent) to give 79 (2.3 mg,
38%), which dissolves in acetone/water or MeOH/water.
1H NMR (500 MHz, D2O:CD3COCD3 ) 1:1): δ 1.83-2.17 (m,
8H-3′, partially overlapped with CHD2COCD3), 2.28-2.48 (m,
8H, CH2CH2OH, 3CH2CH2SO2), 2.53-2.72 (m, 4H, 4H-4′),
3.20-3.50 (m, 10H), 3.55-3.75 (2m, 2H), 3.85-3.95 (m, 4H),
(4(4′-CH2SO2), 3CH2CH2SO2, CH2OH), 4.00-4.15 (m, H-5′,
overlapped with HDO), 4.47 (s, 2H, SO2CH2Ph), 5.69 (intensity,
11.77), 5.70 (7.05), 5.71 (6.70) (d, d, d, J ) 8.0, 8.1, 8.0, 4H,
H-5), 5.99-6.04 (m, 4H, H-2′), 7.33-7.44 (2m, 5H, Ph-H), 7.62
7.66, 7.67, 7.68 (d, d, d, d, J ) 8.1, 8.1, 8.1, 8.1, 4H, H-6).
TOF-MS (2,5-dihydroxybenzoic acid matrix, mass 154) (m/
e): 1408 ([M + matrix]+), 1254 (parent mass of 79 without
isotopic substitution 1252), 1155. Solubility:
equiv to Oxone) in water (1.5 mL) was added CH3CN (1.5 mL).
This solution was mixed with 86 (ca. 0.79 µmol) in CH3CN/
H2O (2 mL, 1:1), the mixture was stirred at rt for 10 min, and
H2O2 (30%, 10 µL) was added to consume the excess oxidant.
The solution was purified by HPLC (CN preparative column,
250 × 22.5 mm, 10 µm, from 0% to 20% CH3CN/H2O) to give
90 (450 µg, 24% yield), which dissolves in MeOH and water.
The product is very soluble in CH3CN/H2O (1:1); in pure water,
a solution with concentration up to 3 mM could be obtained.
UV (H2O:CH3CN ) 1:1): λmax ) 262 nm.
TOF-MS (2,5-dihydroxybenzoic acid matrix) (m/e): 2340 ([M
+ H2O]+, parent mass of 6′-HO-USO2USO2USO2USO2USO2-
USO2USO2U-SO3H-3′′ without isotopic substitution 2322).
H O-USO2USO2USO2USO2USO2USO2USO2U-SO2Bn -
3′′ (89). TOF-MS (m/e): 2411 ([M + H2O]+, parent mass
without isotopic substitution 2390).
Ack n ow led gm en t. We are grateful to Dr. Klaus
Christian Schneider, Dr. Ulrike von Krosigk, Dr. An-
drew Roughton, Frank Lang, Dr. Helmut Zepik, Dr.
Clemens Richert, Andre´ Mu¨ller, German Theler, J on-
athan Urbach, and Dr. Fred Baker (Sandoz AG) for
helping with the biological experiments. Z.H. was sup-
ported by fellowships from the Swiss Federal Institute
of Technology. Support is acknowledged from the Na-
tional Institutes of Health (Grant GM 54048) and The
NASA Astrobiology Program.
soluble in acetone, MeOH, in pure water only up to 50 mM
at rt.
J O0003910