Convenient syntheses of 3′-amino-2′,3′- dideoxynucleosides, their 5′-monophosphates, and 3′-aminoterminal oligodeoxynucleotide primers

Article abstract of DOI:10.1021/jo8018889

(Chemical Equation Presented) 5′-Protected 3′-amino-2′, 3′-dideoxynucleosides containing any of the four canonical nucleobases (A/C/G/T) were prepared via azides in five to six steps, starting from deoxynucleosides. For pyrimidines, the synthetic route in

Full text of DOI:10.1021/jo8018889

Convenient Syntheses of 3′-Amino-2′,3′-dideoxynucleosides, Their
5′-Monophosphates, and 3′-Aminoterminal Oligodeoxynucleotide
Primers
Ralf Eisenhuth and Clemens Richert*
Institute for Organic Chemistry, UniVersity of Karlsruhe (TH), 76131 Karlsruhe, Germany
cr@rrg.uka.de
ReceiVed September 5, 2008
5′-Protected 3′-amino-2′,3′-dideoxynucleosides containing any of the four canonical nucleobases (A/C/
G/T) were prepared via azides in five to six steps, starting from deoxynucleosides. For pyrimidines, the
synthetic route involved nucleophilic opening of anhydronucleosides. For purines, an in situ oxidation/
reduction sequence, followed by a Mitsunobu reaction with diphenyl-2-pyridylphosphine and sodium
azide, provided the 3′-azidonucleosides in high yield and purity. For solid-phase synthesis of aminoterminal
oligonucleotides, aminonucleosides were linked to controlled pore glass through a novel hexafluoroglutaric
acid linker. These supports gave 3′-aminoterminal primers in high yield and purity via conventional
DNA chain assembly and one-step deprotection/release with aqueous ammonia. Primers thus prepared
were successfully tested in enzyme-free chemical primer extension, an inexpensive methodology for
genotyping and labeling. Protected 5′-monophosphates of 3′-amino-2′,3′-dideoxynucleosides were also
prepared, providing starting materials for the preparation of labeled or photolably protected monomers
for chemical primer extension.
Introduction
Structurally modified 2′-deoxynucleosides are important as
building blocks for the synthesis of oligonucleotides with novel
properties and as antiviral and antitumor agents.1,2 The 3′-
position is frequently chosen for the modification, as removal
or replacement of its hydroxy group renders the corresponding
nucleoside triphosphate unsuitable for formation of phosphodi-
ester linkages, preventing elongation of DNA chains during
replication or reverse transcription after their incorporation.
FIGURE 1. Selected 3′-deoxynucleoside derivatives.
* Corresponding author. Current address: Institute for Organic Chemistry,
University of Stuttgart, 70569 Stuttgart, Germany. Fax: 49 (0) 711 608 643
21.
(1) AntiViral Nucleosides Chiral Synthesis and Chemotherapy; Chu, C. K.,
Ed.; Elsevier B.V.: Amsterdam, 2003.
(2) Hayden, F. G. AntiViral agents in Goodman & Gilman’s The Pharma-
cological Basis of Therapeutics, 9th ed.; Hardman, J. G., Limbird, L. E., Eds.;
McGraw-Hill: New York, 1996; pp 1191-1223.
Some well-known 3′-modified nucleosides with biological
activity are shown in Figure 1. Among them is 3′-azido-3′-
deoxythymidine (zidovudine or AZT, 1),3 a first-generation anti-
HIV drug. Other members of this class of antiviral compounds
(3) (a) Mitsuya, H.; Weinhold, K. J.; Furman, P. A.; St. Clair, M. H.;
Nusinoff-Lehrman, S.; Gallo, R. C.; Bolognesi, D.; Barry, D. W.; Broder, S.
Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 7096–7100. (b) Larder, B. A.; Kemp,
S. D.; Harrigan, P. R. Science 1995, 269, 696–699.
(4) (a) Vince, R.; Hua, M.; Brownell, J.; Daluge, S.; Lee, F. C.; Shannon,
W. M.; Lavelle, G. C.; Qualls, J.; Weislow, O. S.; Kiser, R. Biochem. Biophys.
Res. Commun. 1988, 156, 1046–1053. (b) Vince, R.; Hua, M. J. Med. Chem.
1990, 33, 17–21.
26 J. Org. Chem. 2009, 74, 26–37
10.1021/jo8018889 CCC: $40.75 2009 American Chemical Society
Published on Web 11/16/2008

3′-Amino-2′,3′-dideoxynucleosides and Primers
SCHEME 1. Template-Directed Primer Extensiona
include 2′,3′-didehydro-2′,3′-dideoxyguanosine (carbovir, 2),4
2′,3′-didehydro-3′-deoxythymidine (stavudine),5 2′,3′-dideoxy-
3′-fluoro-4-thiothymidine,6 and 2′,3′-dideoxy-3′-hydroxymeth-
ylcytidine.7 A more distantly related, ribonucleoside antibiotic
with an acylamido substituent replacing the 3′-OH group is
puromycin (3). This dimethyladenosine derivative has also been
used to set up in vitro selection of new proteins through
synthetically modified mRNA-peptide fusions.8
As mentioned above, access to nucleosides with a non-natural
functional group and/or the stereochemical configuration at the
3′-position is also important for the synthesis of novel oligo-
nucleotides. This includes aminonucleosides, such as 4, em-
ployed as building block for the solid-phase synthesis of cationic
deoxynucleic oligomers. Other aminonucleosides serve as
building blocks for the assembly of phosphoramidate-linked
oligonucleotides,9 which are important candidate structures for
sequence-specific inhibition of gene expression.10 For some
applications, amino groups are attractive because aliphatic
amines possess high nucleophilicity in aqueous solution, al-
lowing for labeling or bioconjugation.11 A conventional way
of introducing aliphatic amino groups to oligonucleotides
involves appending aminoalkyl substituents via phosphodiester
linkages.12 This approach has the advantage of producing
sterically well accessible amino groups. It is unsuitable,
however, for the generation of oligonucleotides with phospho-
ramidate linkages, which are isoelectronic to the natural
phosphodiesters of DNA and RNA.
a An amino-terminal primer (5) reacts with an active ester of a 3′-
amino-2′,3′-dideoxynucleoside monophosphate (6, PG ) protecting
group), as directed by a template strand to generate phosphoramidate-
linked extended primer 8. Both 5 and 6 are retrosynthetically linked to
common precursor 7.
Chemical primer extension avoids the need for polymerases
in the sequence-determining step and employs inexpensive
monophosphate substrates, rather than expensive triphosphates.
It is therefore expected to be considerably less expensive than
current, enzyme-based methods, making it a candidate tech-
nology for meeting the $1000 genome challenge.17 Routine use
of chemical primer extension requires streamlined syntheses of
the aminonucleosides and automated, massively parallel syn-
theses of aminoterminal primers. These requirements motivated
us to develop convenient syntheses of 3′-amino-2′,3′-dideoxy-
nuclesides and -nucleotides that are rugged, require inexpensive
starting materials, involve as few chromatographic steps as
possible, and allow for standard methodology and reagents to
be used for chain assembly of aminoterminal primers on solid
support. In the process of developing such syntheses, we hoped
to also gain convenient access to xylodeoxynucleosides and
azidodideoxynucleosides. Either may serve as intermediates in
the preparation of other functionalized nucleic acids. For
example, 3′-azidonucleosides may be used in copper-catalyzed
1,3-dipolar cycloaddition reactions.18
Chemical primer extension is the template-directed incorpora-
tion of mononucleotides in the absence of enzymes.13 This
reaction occurs in high yield and base selectivity when the 3′-
terminal nucleoside of the primer (5, Scheme 1) is reacted with
an oxyazabenzotriazolide (OAt) of a nucleoside 5′-monophos-
phate (6). Either can be traced back to azido building block 7.
Chemical primer extension generates oligodeoxyribonucleotides
with a phosphoramidate linkage (8). It is suitable for genotyping
single nucleotide polymorphisms14 and is being developed as a
technology for enzyme-free resequencing of short stretches of
DNA15 or possibly RNA.16
Results and Discussion
Two routes to 3′-amino-terminal primer were tested. The first
employs 5′- or “reversed” phosphoramidites of 2′-deoxynucleo-
sides19 for chain assembly on solid support and phthalimidate
10 in the last coupling step. Compound 10 was obtained through
phosphitylation of 920 in 72% yield and an overall yield of 14%
from thymidine (Scheme 2). The deprotection of oligonucleo-
tides thus prepared proved difficult, yielding no more than 20%
of model oligodeoxyribonucleotide 5′-CTGT*-3′, where T*
denotes the 3′-aminothymidine residue, when deprotecting with
a mixture of 40% methylamine/ammonium hydroxide (1:1)21
at elevated temperature. MALDI analysis suggested that
breaking the second, amidic bond of the phthalimido group is
a slow process. Ammonium hydroxide alone did not produce
(5) (a) Lin, T.-S.; Schinazi, R. F.; Prusoff, W. H. Biochem. Pharmacol. 1987,
36, 2713–2718. (b) Chu, C. K.; Schinazi, R. F.; Arnold, B. H.; Cannon, D. L.;
Doboszewski, B.; Bhadti, V. B.; Gu, Z. Biochem. Pharmacol. 1988, 37, 3543–
3548. (c) Balzarini, J.; Van Aeschot, A.; Herdewijn, P.; De Clercq, E. Biochem.
Pharmacol. 1989, 38, 869–874.
(6) Matthes, E.; Lehmann, C.; Von Janta-Lipinski, M.; Scholz, D. Biochem.
Biophys. Res. Commun. 1989, 165, 488–495.
(7) (a) Sterzycki, R. Z.; Martin, J. C.; Wittman, M.; Brankovan, V.; Yang,
H.; Hitchcock, M. J.; Mansuri, M. M. Nucleosides Nucleotides 1991, 10, 291–
294. (b) Svansson, L.; Kvarnstrom, I.; Classon, B.; Samuelsson, B. J. Org. Chem.
1991, 56, 2993–2997. (c) Bottiger, D.; Johansson, N.-G.; Samuelsson, B.; Zhang,
H.; Putkonen, P.; Vrang, L.; Oberg, B. AIDS 1997, 11, 157–162.
(8) Roberts, R. W.; Szostak, J. W. Proc. Natl. Acad. Sci. U.S.A. 1997, 94,
12297–12302.
(9) Linkletter, B. A.; Szabo, I. E.; Bruice, T. C. J. Am. Chem. Soc. 1999,
121, 3888–3896.
(10) Gryaznov, S. M. Biochim. Biophys. Acta 1999, 1489, 131–140.
(11) Hermanson, G. T. Bioconjugate Techniques; Academic Press: San Diego,
1996.
(12) (a) Agrawal, S. In Methods in Molecular Biology Vol. 26: Protocols
for Oligonucleotide Conjugates; Agrawal, S., Ed.; Humana: Totowa, 1994; pp
93-120. (b) Gamper, H. B.; Reed, M. W.; Cox, T.; Virosco, J. S.; Adams, A. D.;
Gall, A. A.; Scholler, J. K.; Meyer, R. B., Jr. Nucleic Acids Res. 1993, 21, 145–
150.
(16) Plutowski, U.; Vogel, S. R.; Bauer, M.; Deck, C.; Pankratz, M.; Richert,
C. Org. Lett. 2007, 9, 2187–2190.
(17) Service, R. F. Science 2006, 311, 1544.
(18) (a) Tornoe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002,
66, 3057–3064. (b) Demko, Z. P.; Sharpless, K. B. A. Angew. Chem., Int. Ed.
2002, 41, 2110–2113.
(19) (a) Uhlmann, E.; Engels, J. Tetrahedron Lett. 1986, 27, 1023–1026. (b)
Sande, v. d. J. H.; Ramsing, N. B.; Germann, M. W.; Elhorst, W.; Kalisch, B. W.;
Kitzing, v. E.; Pon, R. T.; Clegg, R. C.; Jovin, T. M. Science 1988, 241, 551–
557. (c) Koga, M.; Moore, M. F.; Beaucage, S. L. J. Org. Chem. 1991, 56,
3757–3759.
(13) Rojas Stu¨tz, J. A.; Kervio, E.; Deck, C.; Richert, C. Chem. BiodiV. 2007,
4, 784–802.
(14) Hagenbuch, P.; Kervio, E.; Hochgesand, A.; Plutowski, U.; Richert, C.
Angew. Chem. 2005, 117, 6746-6750; Angew. Chem., Int. Ed. 2005, 44, 6588-
6592.
(15) Griesang, N.; Giessler, K.; Lommel, T.; Richert, C. Angew. Chem. 2006,
118, 6290-6294; Angew. Chem., Int. Ed. 2006, 45, 6144-6148.
(20) Miller, N.; Fox, J. J. J. Org. Chem. 1964, 29, 1772–6.
(21) Reddy, M. P.; Hanna, N. B.; Farooqui, F. Tetrahedron Lett. 1994, 35,
4311–4314.
J. Org. Chem. Vol. 74, No. 1, 2009 27

Eisenhuth and Richert
the reaction mixture of the ring opening to compound 1726
lowered the yield by favoring opening at the anomeric position
with concomitant loss of benzoylcytosine.
SCHEME 2. Phosphitylation of Phthalimidodeoxythymidine 9
While 5′-monophosphates were generated without protecting
group change at the 5′-position (vide infra), the aminonucleo-
sides subsequently used for solid-phase DNA synthesis were
provided with a 5′-DMT group to allow for chain assembly with
standard instrumentation and reagents. The installation of the
DMT group proceeded to give compound 19 in 83% yield in
pyridine. The reduction of the azido group occurred in 93%
yield when hydrogen and Pd/C were used. Amine 20 was thus
obtained in 40% overall yield over five steps in high purity and
with no more than two chromatographic purifications. The route
proved convenient, rugged, and inexpensive and gives a slightly
higher overall yield than literature-known protocols,27 which
also start from more costly starting materials.
We then turned to the purines. The route to the 3′-amino-
2′,3′-dideoxynucleosides of guanine and adenine is shown in
Scheme 4. Attempts to prepare the 3′-azides of the purines via
the N3,3′-anhydronucleosides failed, due to their low reactivity
in the ring opening reaction. Transglycosylation reactions with
AZT as initial starting material were considered,28 but not
actively pursued due to the expected low yield and time-
consuming chromatographic isolation of the desired ꢀ-nucleo-
sides from the anomeric mixtures that makes the approach
unattractive for large scale syntheses. A route via the 2′-
deoxyxylonucleosides, conversion of the 3′-hydroxy group to a
leaving group, and SN2 displacement with azide appeared more
attractive, but published versions of this general route had
significant drawbacks. The route of Robins and co-workers gives
access to xylonucleosides,29,30 starting from 2′-tosylated ribo-
nucleosides, but it involves pyrophoric reagents and unconven-
tional steps. The best known route, established for both
deoxyadenosine and deoxyguanosine,31,32 is that of Herdewijn
and co-workers, involving intramolecular attack on a 3′-mesylate
from the 5′-position to generate the xylo configuration at the
3′-position. The approach has also been studied for the deoxy-
cytidine derivative.33 After initial experiments, we decided not
to pursue this route further, since the acyl migration ac-
companying the intramolecular attack leads to a mixture of
isomers that limits the yield and requires an expensive and time-
consuming chromatographic purification. Instead, we opted for
inversion of configuration at C3′ via an oxidation/reduction
sequence (Scheme 4). Related approaches are known for
ribonucleosides.34
the free amine, even after heating to 55 °C for 4 d. An Fmoc-
or trifluoroactyl-protected version of 10 would have led to
aminoterminal oligomers without a trityl or DMT group on the
full-length product, a fact that was expected to complicate
purification. This and the high cost of the reversed phosphora-
midites prompted us to pursue the second route, which involves
inexpensive, commercial 3′-phosphoramidites and solid support-
bound 3′-aminonucleosides.
Aminonucleosides. Starting from thymidine, 5′-protected
aminodeoxythymidine 11 (Scheme 3) was prepared in five steps
via AZT, using literature-known transformations including
reduction of the azido group with NaBH4.22,23 This route
(Scheme S1, Supporting Information) proceeded in 51% overall
yield and required no more than a single chromatographic step.
The synthesis of 5′-protected 3′-amino-2′,3′-dideoxycytidine 20
started from commercial N4-benzoyl-2′-deoxycytidine (12,
Scheme 3). A double-Mitsunobu reaction induced the cyclization
and protected the 5′-hydroxy function.24 The resulting anhy-
dronucleoside was opened with azide, followed by exchange
of the 5′-protecting group and reduction to the amine.
Attempts to induce these transformations without protecting
the nucleobase failed. A series of 5′-acyl groups were tested to
achieve high yields in the first part of the route, leading up to
the 3′-azide. Anhydronucleoside 13 with an unsubstituted
benzoyl group was too soluble for precipitation from ether, and
chromatographic separation from the reagents proved difficult.
A p-methoxybenzoyl moiety, the preferred protecting group in
the AZT synthesis of Valery and co-workers,22 gave no more
than 25% overall yield for azide 18, due to partial loss of the
nucleobase protecting group during 5′-deprotection. A p-
nitrobenzoyl group gave the anhydronucleoside 14 in 75% yield
but unacceptably low yields during the subsequent nucleophilic
ring opening with azide, most likely due to the high lability of
the protecting group.25 Protection with the p-chlorobenzoyl
group led to low yields in the Mitsunobu reaction to anhydro-
nucleoside 15 (40%). A p-bromobenzoyl group gave satisfactory
results for all three steps (cyclization/protecting, ring opening,
and hydrolytic 5′-deprotection), with yields of 80%, 67%, and
95%, respectively. With this protecting group, anhydronucleo-
side 16 could be readily separated from triphenylphosphine
oxide and other contaminations through precipitation from
diethylether. Interestingly, the addition of lithium chloride to
As expected, attempts to isolate the 3′-ketones of deoxy-
adenosine35 and deoxyguanosine36 were complicated by their
lability toward depurination, which can lead to full decomposi-
(27) Chen, J. K.; Schultz, R. G.; Lloyd, D. H.; Gryaznov, S. M. Nucleic
Acids Res. 1995, 23, 2661–2668.
(28) Imazawa, M.; Eckstein, F. J. Org. Chem. 1978, 43, 3044–3048.
(29) (a) Hansske, F.; Robins, M. J. J. Am. Chem. Soc. 1983, 105, 6736–
6737. (b) Robins, M. J.; Nowak, I.; Wnuk, S. F.; Hansske, F.; Madej, D. J. Org.
Chem. 2007, 72, 8216–8221.
(30) Timoshchuk, V. A.; Hogrefe, R. I.; Vaghefi, M. M. Nucleosides
Nucleotides Nucl. Acids 2004, 23, 171–181.
(31) Herdewijn, P. A. M. J. Org. Chem. 1988, 53, 5050–5053.
(32) Herdewijn, P.; Van Aerschot, A. Tetrahedron Lett. 1989, 30, 855–858.
(33) Seela, F.; Woerner, K.; Rosemeyer, H. HelV. Chim. Acta 1994, 77, 883–
896.
(34) (a) Kim, B.-T.; Kim, S.-K. L.; Seung-Jae, H. K.-J. Bull. Chem. Soc.
Jpn. 2004, 25, 243–248. (b) Pfitzner, K. E.; Moffatt, J. G. J. Am. Chem. Soc.
1965, 87, 5661–5670. (c) Cook, A. F.; Moffatt, J. G. J. Am. Chem. Soc. 1967,
89, 2697–2705. (d) Hansske, F.; Robins, M. J. Tetrahedron Lett. 1983, 24, 1589–
1592.
(22) Czernecki, S.; Valery, J. M. Synthesis 1991, 239–240.
(23) Celewicz, L.; Urjasz, W.; Golankiewicz, K. Nucleosides Nucleotides
1993, 12, 951–956.
(24) Czernecki, S.; Valery, J. M. Chem. Commun. 1990, 801–802.
(25) Gomez-Vidal, J. A.; Forrester, M. T.; Silverman, R. B. Org. Lett. 2001,
3, 2477–2479.
(26) Mikhailopulo, I. A.; Zaitseva, G. V.; Vaaks, E. V.; Rosemeyer, H.; Seela,
F. Nucleosides Nucleotides 1992, 11, 273–238.
(35) Samano, V.; Robins, M. J. J. Org. Chem. 1990, 55, 5186–5188.
28 J. Org. Chem. Vol. 74, No. 1, 2009

3′-Amino-2′,3′-dideoxynucleosides and Primers
SCHEME 3. Synthesis of 3′-Amino-3′-deoxythymidine and 3′-Amino-2′,3′-dideoxycytidine
tion, even at room temperature. To avoid this, a one-pot
oxidation/reduction approach with a low temperature regime
was developed. We chose a TBDPS protecting group for the
5′-hydroxy group to ensure high regioselectivity and to prevent
partial deprotection during the oxidation step, which was to be
expected for trityl-based protecting groups. Oxidation at the 3′-
position with the Dess-Martin reagent and immediate reduction
of the ketone with NaBH4 in ethanol at 0 °C gave a diastereo-
meric mixture of 4:1 (xylo to ribonucleoside). The initial yield
for the two-step conversion of 21 to 22 was low (22%) due to
partial depurination and cumbersome chromatographic separa-
tion of the polar mixture. Protecting groups for the nucleobase
were introduced to address these issues. A dimethylformamidino
group37 at N6 worsened the depurination problem, so a more
conventional benzoyl protection was used. Silylation of the 5′-
hydroxy group of N6-benzoyl-2′-deoxyadenosine was higher
yielding with TBS-Cl than with TBDPS-Cl and the separation
of the epimeric mixture was easier, so the less sterically
demanding reagent was used. Optimization of all three steps of
the protection/oxidation/reduction sequence led to 70% overall
yield for A. Key features of the optimized procedure are
imidazole-free pyridine for silylation, addition of the solid
nucleoside to a solution of Dess-Martin reagent at 0 °C,
quenching of excess oxidation reagent with 2-propanol, which
also serves as solvent for the reduction step, and lowering of
the temperature for the latter step to -60 °C. The low
temperature reduced the extent of depurination; it also improved
the diastereoselectivity of the reduction step to 19:1 in favor of
the desired xylonucleoside.
The oxidation/reduction route was also employed for generat-
ing the xylo derivative of 2′-deoxyguanosine. Amidine protect-
ing groups for the exocyclic amino group of the nucleobase
(dimethylformamidino or dibutylforamidino groups)37 again
fared more poorly than acyl protection (iBu group), due to the
resulting lability and/or more complicated chromatographic
separation of the epimers. For the N2-isobutyryl derivative, 5′-
silylation proceeded in 84% yield and the two-step epimerization
of 24g to 25g gave 77% overall yield.
Different approaches for generating the 3′-azides from the
xylopurines were tested. Conversion of the 3′-hydroxy group
to the corresponding triflate31 followed by SN2 displacement
with NaN3 in DMF gave no more than 30% yield, most probably
due to competing sulfonation at the nucleobase protecting group
(36) Kralikova, S.; Budesinsky, M.; Masojidkova, M.; Rosenberg, I. Nucleo-
sides, Nucleotides Nucl. Acids 2000, 19, 1159–1183.
(37) (a) Zemlicka, J.; Chladek, S.; Holy, A.; Smrt, J. Collect. Czech. Chem.
Commun. 1966, 31, 3198–3212. (b) Froehler, B. C.; Matteucci, M. D. Nucleic
Acids Res. 1983, 11, 8031–8036. (c) McBride, L. J.; Kierzek, R.; Beaucage,
S. L.; Caruthers, M. H. J. Am. Chem. Soc. 1986, 108, 2040–2048.
J. Org. Chem. Vol. 74, No. 1, 2009 29

Eisenhuth and Richert
SCHEME 4. Synthesis of 3′-Amino-2′,3′-dideoxyguanosine and 3′-Amino-2′,3′-dideoxyadenosine
and subsequent loss of the benzoyl group. Mesylation at O3′
with conventional DNA synthesizers. Desilylation with TBAF,
followed by reprotection with DMT, and reduction of the azido
functionality with H2, Pd/C proceeded uneventfully in overall
yields of 65-79%. Thus, the conversion of N6-benzoyl-2′-
deoxyadenosine (23a) to amine 29a requires six steps and
proceeds in 44% overall yield. For the conversion of 2′-deoxy-
N2-isobutyrylguanosine (23g) to the corresponding amine 29g,
the route requires six steps and gives 37% overall yield. These
yields are 7% and 17%, higher, respectively, than those of the
highest yielding route known previously.27 We found the routes
for either of the four nucleosidic amines easy-to-scale up to a
gram scale. They have been successfully performed by under-
graduate students in a teaching laboratory.
Aminoterminal Primers. We then developed a linker for
immobilizing the aminonucleosides on controlled pore glass
(cpg) such that standard DNA synthesis with conventional 3′-
phosphoramidites generates aminoterminal primers of any given
sequence without a single nonstandard reagent. This was to
ensure that aminoterminal primers may be generated in auto-
mated and possibly massively parallel fashion. Supports for the
solid-phase synthesis of aminoalkyloxyphosphoryl-terminated
oligonucleotides based on phthalimido45 or acetoxymethylben-
zoyl46 linkers are known from the literature. These were not
pursued in the present case to avoid loss or migration of
protecting groups. Immobilization of aminonucleosides has also
been achieved47 using linkers derived from the photolabile 2-(2-
nitrophenyl)ethoxycarbonyl protecting group48 or from phos-
phoramidate linkages.49 The former leads to low loading and
the latter to a two-stage chemical deprotection.
proved feasible,38 but the SN2 displacement with sodium azide
was low yielding. Treatment of deoxyxyloguanosine 25g with
32,39
CBr4, PPh3, and NaN3
gave azide 26g in 56% yield, but
the subsequent removal of the triphenylphosphine proved more
than tedious. The same problem plagued initial attempts to use
a Mitsunobu reaction with PPh3, DEAD, and NaN3 for the
conversion of the xylopurine to the azidonucleoside via the
Hirschbein procedure.40 tert-Butylphosphine as reagent41 did
not give discernible conversion, and gas evolution upon addition
of this reagent suggested that a Staudinger reduction of the azide
occurred instead. Reactions with solid-phase bound triph-
enylphosphine42 were also unsuccessful. The purification prob-
lem encountered with PPh3 was overcome by employing
diphenyl-2-pyridylphosphine. While removal of the correspond-
ing phosphine oxide in an acidic extraction step was doable,43
chromatographic removal proved more reliable, particularly
when DIAD, rather than DEAD was used. The availability of
pure azides (27a and 27g) allowed us to avoid extensive
purification on the stage of the free amine. With the optimized
protocol, azides 27a and 27g were obtained in isolated yields
of 81% and 86%, respectively (Scheme 4).
Though oligonucleotide syntheses involving 5′-silyl protecting
groups are known,44 we opted for DMT protection of the 5′-
hydroxy groups of our building blocks to ensure compatibility
(38) Herdewijn, P.; Balzarini, J.; De Clercq, E.; Pauwels, R.; Baba, M.;
Broder, S.; Vanderhaeghe, H. J. Med. Chem. 1987, 30, 1270–1278.
(39) Herdewijn, P.; Balzarini, J.; Baba, M.; Pauwels, R.; Van Aerschot, A.;
Janssen, G.; De Clercq, E. J. Med. Chem. 1988, 31, 2040–2048.
(40) Nelson, J. S.; Fearon, K. L.; Nguyen, M. Q.; McCurdy, S. N.; Frediani,
J. E.; Foy, M. F.; Hirschbein, B. L. J. Org. Chem. 1997, 62, 7278–7287.
(41) Von Itzstein, M.; Jenkins, I. D. Aust. J. Chem. 1983, 36, 557–563.
(42) Amos, R. A.; Emblidge, R. W.; Havens, N. J. Org. Chem. 1983, 48,
3598–3600.
(45) Petrie, C. R.; Reed, M. W.; Adams, A. D.; Meyer, R. B., Jr. Bioconjugate
Chem. 1992, 3, 85–87.
(46) Leuck, M.; Giare, R.; Paul, M.; Zien, N.e; Wolter, A. Tetrahedron Lett.
2004, 45, 317–320.
(47) Avino, A.; Garcia, R. G.; Albericio, F.; Mann, Ms.; Wilm, M.; Neubauer,
G.; Eritja, R. Bioorg. Med. Chem. 1996, 4, 1649–1658.
(43) Camp, D.; Jenkins, I. D. Aust. J. Chem. 1988, 41, 1835–1839.
(44) Scaringe, S. A.; Wincott, F. E.; Caruthers, M. H. J. Am. Chem. Soc.
1998, 120, 11820–11821.
30 J. Org. Chem. Vol. 74, No. 1, 2009

3′-Amino-2′,3′-dideoxynucleosides and Primers
SCHEME 5. (a) Preparation of Solid Supports Loaded with Aminonucleosides and (b) Synthesis of Amino-Terminal Primer
Earlier work on lysyl-terminated oligonucleotides had shown
that substantial strand cleavage can occur during deprotection
in the presence of free aliphatic amino groups.50,51 A trifluo-
roacetyl group protecting the ꢀ-amino group of conjugated lysine
residues gives oligonucleotides in high yields and crudes of
excellent purity, probably because amino groups are released
more slowly than the internucleosidic phosphodiester anions,
which are much more resistant toward nucleophilic attacks than
the initial phosphotriesters. This prompted us to test fluorinated
derivatives of the succinyl linker employed in conventional DNA
syntheses.52 Hexafluorogutaric anhydride is an affordable,
commercial reagent that creates sufficiently labile amide bonds.
Using this anhydride, amide 30t was obtained as diisopropyl-
ethylammonium salt (Scheme 5a). All attempts to couple 30t
to long-chain alkylamine controlled pore glass (LCAA-cpg, 31)
or simple model amines by using conventional amide-forming
reagents, such as DCC, DIC, EDC ·HCl, HATU53 (HOAt),
HBTU (HOBt), and TFFH54 failed, probably because of the
low reactivity of the carboxylate. Triphenylphosphine/I2, an
activation mixture developed for peptide synthesis,55 was
successful in model syntheses. For oligonucleotides, the highest
yields of aminoterminal primers were obtained when this reagent
was used for linking the aminonucleosides to the LCAA-cpg
previously treated with hexafluoroglutaric anhydride, rather than
coupling the previously acylated nucleoside (compare Scheme
5b).
conditions for the I2/PPh3 mixture.55 This was traced to partial
loss of the DMT groups, on which the quantitation was based,
probably through local release of HI during the coupling
reaction. Using more equivalents of base and switching to
pyridine as solvent did not eliminate this problem fully, but
loadings in the range of 17-28 µmol/g were measured after
changing the order of addition of the reagents (iodine was added
last). This level of loading, which is comparable to that of
commercial cpgs with unmodified nucleosides, was achieved
with no more than 3 equiv of the respective aminonucleoside
in the coupling reaction.
The average overall yield of aminoterminal primers 33a-t
was 40%, as determined by HPLC. This value is twice as high
as that obtained using 5′-phosphoramidites and incorporation
of an aminodeoxynucleoside in the last coupling cycle (21%).
The purity of crude aminoterminal primers prepared from any
of the four cpgs (32a-t) was determined by HPLC and MALDI-
TOF MS and was found to be identical to that of unmodified
control oligonucleotides prepared under the same conditions
(compare parts a and b of Figure 2 for representative spectra).
Chromatographically and spectrometrically pure primers were
also obtained after simple cartridge purification. This was
considered an important feature for routine, automated, and
potentially parallel syntheses of primers for practical applications.
The ability of the aminoterminal primers to undergo chemical
primer extension was also tested. Figure 3 shows representative
A low apparent loading (10 and 14 µmol/g support) was
measured after coupling the aminonucleosides under literature
(48) Himmelsbach, F.; Schulz, B. S.; Trichtinger, T.; Charubala, R.; Pfleiderer,
W. Tetrahedron 1984, 40, 59–72.
(49) (a) Gryaznov, S. M.; Letsinger, R. L. Nucleic Acids Res. 1992, 20, 3403–
3409. (b) Gryaznov, S. M.; Letsinger, R. L. Tetrahedron Lett. 1993, 34, 1261–
1264. (c) Rojas-Stuetz, J. A.; Richert, C. J. Am. Chem. Soc. 2001, 123, 12718–
12719.
(50) Schwope, I.; Bleczinski, C. F.; Richert, C. J. Org. Chem. 1999, 64,
4749–4761.
(51) Gruenefeld, P.; Richert, C. Synlett 2007, 1–18.
(52) Matteucci, M. D.; Caruthers, M. H. J. Am. Chem. Soc. 1981, 103, 3185–
3191.
(53) Carpino, L. A.; El-Faham, A.; Albericio, F. J. Org. Chem. 1995, 60,
3561–3564.
(54) Carpino, L. A.; El-Faham, A. J. Am. Chem. Soc. 1995, 117, 5401–5402.
(55) Zhou, Q. Z.; Chen, Z. C. Synth. Commun. 2000, 30, 3189–3194.
FIGURE 2. MALDI-TOF mass spectrum (lower part) and HPLC trace
(insert) of (a) crude aminoterminal primer 33a or (b) unmodified control
oligonucleotide 34, prepared using the same protocol and synthesizer,
starting from either 32a or conventional cpg.
J. Org. Chem. Vol. 74, No. 1, 2009 31

Eisenhuth and Richert
SCHEME 6. Synthesis of 3′-Aminonucleoside
5′-Monophosphates
FIGURE 3. Chemical primer extension: (a) sequences of aminoterminal
primer 33c, DNA template 35, and DNA “helper strand” GAC-
CTAAAGGAGTCG (36) reacting with activated mononucleotide to
form the extended primer 37; (b) MALDI-TOF mass spectrum showing
full primer extension.
of nucleobase-protected 3′-azidonucleosides, followed by oxida-
tion and stepwise deprotection, was also considered, but not
pursued due to the large number of steps required. Instead,
azidonucleosides 1, 18, and 27a/g were converted to triesters
38a-t by reacting them with dibenzylphosphoryl chloride,
which is commercially available. The benzyl groups of the
resulting (uncharged) triester were hydrogenolytically removed
without reduction of the nucleobases,59,60 with concomitant
reduction of the azido groups to the free amines (Scheme 6).
The reaction of azidonucleosides 1, 18, and 27a/g with
dibenzylphosphoryl chloride in pyridine at -30 °C gave yields
of 69-79%. For 38g, near-identical Rf values for educt and
product led to low yields. Unless a more suitable eluant and/or
stationary phase for chromatography can be identified, this route
will probably remain problematic for this compound. Explor-
atory work shows that a satisfactory yield for the phosphory-
lation step can be achieved when position O6 is protected, as
shown in Scheme 7. Substrate 41 may be obtained from 26g
through a combination of a Mitsunobu reaction with benzyl
alcohol and desilylation. The former of the two steps has yet to
be optimized. All three benzyl groups of 42 should be removable
in a single hydrogenation step.61
The hydrogenation of 38a-t proceeded smoothly when water
was added after the initial phase of the reaction, which was run
in ethanolic bicarbonate solution. This ensured that poorly
soluble intermediates were brought back into solution and
prevented incomplete conversion to the free aminonucleotides.
The hydrogenation step thus produced nucleobase-protected
amines 39a and 39c, in yields of 81 and 83%. The corresponding
guanosine derivative 39g also reacted smoothly in exploratory
reactions. Aminothymidine monophosphate 39t was obtained
MALDI spectra and the sequences of the strands involved. Full
primer extension was observed in every case, confirming that
the 3′-termini react in the expected, template-directed fashion.
Additional data can be found in the Supporting Information.
Monophosphates of Aminonucleosides. Chemical primer
extension by more than one nucleotide requires nucleosidic
building blocks containing both a reactive nucleophilic group
and a 5′-phosphate.15 Aminonucleoside monophosphates are also
of interest as starting materials for the synthesis of fluorophore-
labeled monomers. Such fluorescent monomers allow for
genotyping single nucleotide polymorphisms with optical detec-
tion of alleles.56 Further, other forms of labeling nucleotides,
e.g., with affinity ligands become feasible through the high
reactivity of the aliphatic 3′-amino group of aminonucleotides
toward active esters of carboxylic esters in aqueous solution.
We sought a short, convenient route for the 5′-monophospates
of any of the four canonical aminonucleosides (A/C/G/T) that
would avoid ionic intermediates, which are difficult to purify
by conventional column chromatography. Further, the route
should ideally provide intermediates with free 3′-amino groups
but protected nucleobases to prevent side reactions during
acylation, even under harsh reaction conditions. A route to 3′-
amino-3′-deoxythymidine 5′-monophosphate (39t) is known
which involves DCC-mediated coupling of the pyridinium salt
of cyanoethylphosphate57 to AZT.58 The pyridinium salt of the
phosphorylating reagent used is not commercially available
though, and a purification of ionic intermediates cannot be
avoided. Phosphorylation with POCl3 suffers from modest yields
and also requires ion-exchange chromatography or similar
methods for the purification of intermediates. Phosphitylation
(59) Michelson, A. M.; Todd, A. R. J. Chem. Soc. 1949, 2476–2486.
(60) Sim, M. M.; Kondo, H.; Wong, C. H. J. Am. Chem. Soc. 1993, 115,
2260–2267.
(61) Note added in proof: The hydrogenation has been performed successfully,
with liberation of the free aminonucleotide in high yield.
(56) Griesang, N.; Kervio, E.; Richert, C. Synthesis 2005, 2327–2334.
(57) Tener, G. M. J. Am. Chem. Soc. 1961, 83, 159–168.
(58) Glinski, R. P.; Khan, M. S.; Kalamas, R. L.; Sporn, M. B. J. Org. Chem.
1973, 38, 4299–4305.
32 J. Org. Chem. Vol. 74, No. 1, 2009

3′-Amino-2′,3′-dideoxynucleosides and Primers
SCHEME 7. Synthesis of Protected Azidodideoxyguanosine-5′-monophosphate 42
in 78% yield. For A, C, and T, filtration and lyophilization alone
gave crude products that are pure by NMR, avoiding compli-
cated purification steps of these zwitterionic compounds (see
the Supporting Information for spectra). The overall yield from
the commercial, nucleobase-protected (A and C) or unprotected
(T) deoxynucleosides to the 5′-monophosphates of the amino-
nucleoside was 28-38%. Exploratory experiments with pho-
tolably protected aminonucleoside monophosphates led to
doubly extended aminoterminal primers in high yield.62
The usefulness of aminodeoxynucleosides and oligonucle-
otides containing them is not limited to primer extension
reactions. The high reactivity of the aliphatic amino group at
the 3′-position can also be used to label nucleic acids with
fluorescent residues.15,56 Labeling63 and immobilization on
surfaces64 may also be performed via azidonucleosides undergo-
ing “Click”-type 1,3-dipolar cycloaddition18 with underlying
ligation.65 Thus, azidonucleosides synthesized by the method-
ology described here may become valuable intermediates in the
corresponding syntheses. Other intermediates, namely the xy-
lonucleosides 25a and 25g, may also lead to other 3′-modified
deoxynucleosides66 that are pharmaceutically relevant.
Conclusions
The syntheses described here give access to aminodideoxy-
nucleosides 11, 20, and 29a/g, in five to six steps and 37-50%
overall yield, starting from inexpensive deoxynucleosides. The
route to the pyrimidines avoids HMPA as solvent and produces
DMT-protected nucleosides compatible with trityl monitoring
on conventional synthesizers. The route for the purines proceeds
in 7-17% higher overall yield than that reported previously.27
Further, the xylonucleosides were obtained in two rather than
four steps and in 25-28% higher yield. Either route avoids
difficult chromatographic steps, and either route proved easy
to scale up.
Controlled pore glass loaded with aminonucleosides using
the hexafluoroglutaric acid linker gives access to aminoterminal
primers of any sequence via inexpensive, conventional DNA
syntheses. The loading reaction via a seldom used redox
reaction55 requires no more than 3 equiv of the 5′- and
nucleobase-protected aminonucleosides and generates supports
with 17 and 28 µmol of nuclosides per gram of cpg. The
aminoterminal primers obtained are of unusually high purity,
suggesting that side reactions common for oligonucleotides with
aliphatic amino groups51 are successfully suppressed. The
aminoterminal primers also passed all functional tests, involving
template-directed chemical primer extension assays.
Experimental Section
3′-Phthalimido-3′-deoxythymidine-5′-O-(2-cyanoethyl)diiso-
propylphosphoramidite (10). A slurry of 3′-phthalimido-3′-
deoxythymidine20 (9, 35 mg, 0.09 mmol) in acetonitrile (2 mL)
and NEt3 (45 µL, 0.27 mmol, 3 equiv) was treated with 2-cyano-
ethyl-N,N-diisopropylchlorophosphoramidite (32 µL, 0.14 mmol,
1.5 equiv). After 2 h, the solution was poured into CH2Cl2, and the
organic solution was washed with satd aqueous NaHCO3, dried
over Na2SO4, filtered, and concentrated in vacuo to a volume of 1
mL. The solution was poured into pentanes (20 mL) and the
precipitate collected to give the title compound in 81% (47 mg,
0.08 mmol) yield: TLC (1% NEt3 in toluene/acetone, 2/1, v/v) Rf
) 0.52; 31P NMR (101 MHz, CD3CN) δ 150.3, 149.7; MS
(MALDI-TOF) 594 (M + Na)+.
O2,3′-Anhydro-N4-benzoyl-5′-O-(p-bromobenzoyl)-2′-deoxy-
cytidine (16). To a solution of N4-benzoyl-2′-deoxycytidine (12,
2 g, 6.04 mmol) in DMF (15 mL) were added PPh3 (2.38 g, 9.06
mmol, 1.5 equiv) and p-bromobenzoic acid (1.82 g, 9.06 mmol,
1.5 equiv), followed by addition of DIAD (1.76 mL, 9.06 mmol,
1.5 equiv). The resulting yellow solution was stirred at rt for 30
min. A solution of PPh3 (2.38 g, 9.06 mmol, 1.5 equiv) and DIAD
(1.76 mL, 9.06 mmol, 1.5 equiv) in DMF (3 mL) was added
dropwise, followed by stirring at rt for an additional 90 min. The
Finally, the results from our exploratory work on nucleobase-
protected aminodideoxynucleoside-5′-monophosphates suggest
that such charged and richly functionalized monomers may also
be generated in convenient syntheses via our methodology. The
monophosphates were obtained in analytically pure form after
hydrogenation, again avoiding any unconventional or time-
consuming purification steps.
(63) Deo, T. S.; Li, Z.; Ruparel, H.; Ju, J. C. J. Org. Chem. 2003, 68, 609–
612.
(64) Seo, T. S.; Bai, X.; Ruparel, H.; Li, Z.; Turro, N. J.; Ju, J. Proc. Natl.
Acad. Sci. U.S.A. 2004, 101, 5488–5493.
(65) Kumar, R.; El-Sagheer, A.; Tumpane, J.; Lincoln, P.; Wilhelmsson,
L. M.; Brown, T. J. Am. Chem. Soc. 2007, 129, 6859–6864.
(66) (a) Chiba, J.; Tanaka, K.; Ohshiro, Y.; Miyake, R.; Hiraoka, S.; Shiro,
M.; Shionoya, M. J. Org. Chem. 2003, 68, 331–338. (b) Chidgivadze, Z. G.;
Bibilashvilli, R. S.; Kraevskii, A. A.; Kukhanova, M. K. Biochim. Biophys. Acta
1986, 868, 145–152. (c) Herrlein, M. K.; Konrad, R. E.; Engels, J. W.; Holletz,
T.; Cech, D. HelV. Chim. Acta 1994, 77, 586–596. (d) Yuzhakov, A. A.;
Chidgeavadze, Z. G.; Bibilashvilli, R. Sh. FEBS Lett. 1992, 306, 185–188.
(62) Roethlingshoefer, M.; Kervio, E.; Lommel, T.; Plutowski, U.; Hoch-
gesand, A.; Richert, C. Angew. Chem., Int. Ed. 2008, 47, 6065–6068.
J. Org. Chem. Vol. 74, No. 1, 2009 33

Eisenhuth and Richert
deep red solution was poured into ice-cold Et2O (150 mL), and the
resulting suspension was kept at 4 °C for 2 h. The precipitate was
collected, washed with ice-cold diethyl ether (5 × 20 mL), and
dried in vacuo to afford a 80% yield of O2,3′-anhydro-N4-benzoyl-
5′-O-(p-bromobenzoyl)-2′-deoxycytidine (16, 2.4 g, 4.83 mmol) as
dideoxycytidine (19, 300 mg, 0.46 mmol) in ethanol (10 mL) was
added saturated aqueous NaHCO3 solution (200 µL). The resulting
slurry was placed under an argon atmosphere, and Pd/C (90 mg)
was added. After the argon was replaced with a hydrogen
atmosphere, generated by the hydrostatic pressure of a 20 cm water
column, the slurry was stirred, with the hydrogen atmosphere being
replaced every 60 min, until TLC indicated complete conversion
(typically less than 6 h). The catalyst was removed by filtration
over a bed of Celite and washed with ethanol. The combined
solutions were concentrated in vacuo to give 3′-amino-N4-benzoyl-
5′-O-(dimethoxytrityl)-2′,3′-dideoxycytidine (20) in 93% yield (269
mg, 0.42 mmol). Spectroscopic data were in agreement with the
literature.27
N6-Benzoyl-5′-O-tert-butyldimethylsilyl-2′-deoxyadenosine (24a).
To a stirred slurry of N6-benzoyl-2′-deoxyadenosine (23a, 2.5 g,
7.04 mmol) in pyridine (15 mL) was added TBS-Cl (1.5 g, 9.86
mmol, 1.4 equiv). After 1 h, additional TBS-Cl (530 mg, 3.53 mmol,
0.5 equiv) was added. After 3 h, methanol (1 mL) was added, and
the reaction mixture was transferred into ethyl acetate. The solution
was washed with satd aqueous NaHCO3, H2O, and brine, dried over
Na2SO4, filtered, and concentrated in vacuo. The residue was
coevaporated twice from toluene and purified on silica (CH2Cl2/
MeOH, 93/7, v/v) to give 24a as a colorless foam in 88% yield
(2.91 g, 6.19 mmol). Spectroscopic data were in agreement with
the literature.67
N6-Benzoyl-5′-O-tert-butyldimethylsilyl-2′-deoxyxyloadeno-
sine (25a). To a chilled solution (0 °C) of Dess-Martin periodinane
in CH2Cl2 (10 wt %, 24 mL, 11.56 mmol, 2.25 equiv) was added
N6-benzoyl-5′-O-tert-butyldimethylsilyl-2′-deoxyadenosine (24a,
2.4 g, 5.13 mmol), and the resulting solution was stirred at 0 °C.
After 30 min, the cooling bath was removed, and stirring was
continued for another 90 min. After addition of 2-propanol (25 mL),
the resulting slurry was cooled to -60 °C, and freshly powdered
NaBH4 (607 mg, 10.26 mmol, 2 equiv) was added. After 2 h of
stirring at -60 °C, acetone (25 mL) was added, and the slurry was
allowed to warm to rt and poured into ethyl acetate (500 mL). The
solution was washed with satd aqueous NaHCO3 solution, H2O,
and brine, dried over Na2SO4, filtered, and concentrated in vacuo.
The resulting yellow foam was purified on silica (CH2Cl2/MeOH,
95:5, v/v) to give the title compound (1.85 g, 3.95 mmol, 77%
yield) as a colorless foam: TLC (CH2Cl2/MeOH, 95/5) Rf ) 0.26;
1H NMR (500 MHz, CDCl3) δ 9.32 (bs, 1H, N-H), 8.88 (s, 1H,
H8), 8.35 (s, 1H, H2), 8.04 (d, J ) 7.5 Hz, 2H, Ar-H), 7.62 (t, J
) 7.5 Hz, 1H, Ar-H), 7.53-7.50 (m, 2H, Ar-H), 6.32 (dd, J ) 2.1
Hz, J ) 9.1 Hz, 1H, H1′), 5.91 (d, J ) 6.9 Hz, 1H, O-H),
4.58-4-53 (bs, 1H, H3′), 4.13-3.99 (m, 3H, H4′, H5′, and H5′′),
2.94-2.85 (m, 1H, H2′′), 2.60-2.54 (m, 1H, H2′), 0.89 (s, 9 H,
Si-CH-CH3), 0.08 (s, 3H, Si-CH3), 0.07 (s, 3H, Si-CH3); 13C
NMR (125 MHz, CDCl3) δ 164.7, 152.0, 150.5, 149.9, 143.0, 133.5,
132.8, 128.8, 127.9, 123.7, 84.8, 84.2, 71.1, 62.1, 40.9, 25.8, 18.3,
-5.3; MS (FAB) m/z 495.1 [M + H]+.
1
a colorless powder: H NMR (400 MHz, DMSO-d6) δ 7.95-7.92
(m, 2H, Ar-H), 7.83-7.80 (m, 2H, Ar-H), 7.75-7.71 (m, 2H, Ar-
H), 7.66 (d, J ) 7.6 Hz, 1H, H6), 7.54-7.50 (m, 1H, Ar-H),
7.46-7.41 (m, 2H, Ar-H), 6.46 (d, J ) 7.6, 1H, H5), 6.00 (bd, J
) 3.8, 1H, H1′), 5.49-5.46 (m, 1H, H3′), 4.63-4.58 (m, 2H, H5′
and H5′′), 4.46-4.40 (m, 1H, H4′), 2.74-2.66 (m, 1H, H2′′),
2.60-2.54 (m, 1H, H2′); 13C NMR (100 MHz, DMSO-d6) δ 177.2,
164.6, 161.9, 152.8, 139.8, 136.5, 131.9, 131.5, 131.1, 128.8, 128.1,
128.0, 127.6, 105.7, 87.2, 81.8, 77.9, 62.3, 32.6; MS (FAB) m/z
496.2, 498.2 [M + H]+.
3′-Azido-N4-benzoyl-5′-O-(p-bromobenzoyl)-2′,3′-dideoxycy-
tidine (17). A slurry of O2,3′-anhydro-N4-benzoyl-5′-O-(p-bro-
mobenzoyl)-2′-deoxycytidine (16, 2.2 g, 4.44 mmol) and NaN3 (866
mg, 13.32 mmol, 3 equiv) in DMF (15 mL) was stirred at 130 °C.
After 2 h, the slurry was poured into satd aqueous NH4Cl solution
(100 mL), and the resulting yellow precipitate was extracted with
ethyl acetate. The organic phase was washed with H2O and brine,
dried over Na2SO4, filtered, and concentrated in vacuo. The residue
was purified on silica (0.1% H2O in hexanes/ethyl acetate, 1/5, v/v)
to afford the title compound in 67% yield (1.6 g, 2.98 mmol) as a
yellow foam: TLC (hexanes/ethyl acetate, 1/2) Rf ) 0.34; 1H NMR
(250 MHz, CDCl3) δ 7.98 (d, J ) 7.6 Hz, 1H, H6), 7.85-7.76 (m,
4H, Ar-H), 7.59-7.34 (m, 6H, Ar-H and H5), 6.05 (t, J ) 5.6 Hz,
1H, H1′), 4.60-4.55 (m, 2H, H4′ and H3′), 4.27-4.12 (m, 2H,
H5′ and H5′′), 2.81-2.68 (m, 1H, H2′′), 2.48-2.36 (m, 1H, H2′);
13C NMR (100 MHz, DMSO-d6) δ 165.3, 162.4, 143.7, 133.2,
132.8, 132.0, 131.0, 129.0, 127.9, 127.5, 87.6, 82.6, 63.6, 60.0,
38.7; MS (FAB) m/z 514.5, 516.5 [M + H]+.
3′-Azido-N4-benzoyl-2′,3′-dideoxycytidine (18). A stirred solu-
tion of 3′-azido-N4-benzoyl-5′-O-(p-bromobenzoyl)-2′,3′-dideoxy-
cytidine (17, 716 mg, 1.39 mmol) in THF (8 mL) was treated with
aqueous LiOH solution (1 M, 5.5 mL, 5.5 mmol, 4 equiv). After
40 min, the reaction mixture was added to CHCl3. The solution
was washed with satd aqueous NH4Cl solution and brine, dried
over Na2SO4, filtered, and concentrated in vacuo to give the title
compound in 93% yield (461 mg, 1.29 mmol) as a yellow solid,
which was used in the subsequent step without further purification:
TLC (CH2Cl2/MeOH, 95/5) Rf ) 0.3; 1H NMR (400 MHz, DMSO-
d6) δ 11.30 (s, 1H, N-H), 8.47 (d, J ) 7.5 Hz, 1H, H6), 8.09-8.05
(m, 1H, Ar-H), 7.71-7.66 (m, 1H, Ar-H), 7.60-7.55 (m, 2H, Ar-
H), 7.43 (d, J ) 7.5 Hz, 1H, H5), 6.13 (t, J ) 5.9 Hz, 1H, H1′),
5.36 (t, J ) 5.3 Hz, 1H, OH), 4.44 (q, J ) 6.1 Hz, 1H, H4′),
4.03-3.99 (m, 1H, H3′), 3.81-3.67 (m, 2H, H5′ and H5′′),
2.57-2.42 (m, 2H, H2′ and H2′′); 13C NMR (100 MHz, DMSO-
d6) δ 167.2, 163.1, 154.3, 144.9, 133.0, 132.7, 128.4, 95.9, 85.9,
84.8, 60.3, 59.4, 37.6; MS (FAB) m/z 357.3 [M + H]+.
3′-Azido-N6-benzoyl-5′-O-tert-butyldimethylsilyl-2′,3′-dideoxy-
adenosine (26a). A stirred slurry of N6-benzoyl-5′-O-tert-butyldi-
methylsilyl-2′-deoxyxyloadenosine (25a, 710 mg, 1.50 mmol) and
NaN3 (293 mg, 4.52 mmol, 3 equiv) in DMF (5 mL) was treated
with a solution of diphenyl-2-pyridylphosphine (590 mg, 2.26
mmol, 1.5 equiv) and DIAD (438 µL, 2.26 mmol, 1.5 equiv) in
DMF (1.5 mL). After 4 h, H2O (2 mL) was added, and the reaction
mixture was transferred into ethyl acetate. The solution was washed
with H2O and brine, dried over Na2SO4, filtered, and concentrated
in vacuo. The resulting brown oil was dried overnight in vacuo
and purified on silica (CH2Cl2/ethyl acetate, 3/2, V/V) to give 26a
in 81% yield (605 mg, 1.22 mmol): TLC (CH2Cl2/ethyl acetate,
3′-Azido-N4-benzoyl-5′-O-(dimethoxytrityl)-2′,3′-dideoxycy-
tidine (19). A stirred solution of 3′-azido-N4-benzoyl-2′,3′-dideoxy-
cytidine (18, 300 mg, 0.85 mmol) in pyridine (5 mL), containing
some molecular sieves (3 Å, 300 mg), was treated with DMTrCl
(245 mg, 0.72 mmol, 1.3 equiv). After 1 h, methanol (500 µL)
was added, and the reaction mixture was transferred into ethyl
acetate. The solution was washed with satd aqueous NaHCO3
solution, H2O, and brine, dried over Na2SO4, filtered, and concen-
trated in vacuo. The resulting yellow foam was purified on silica
(pretreated with 1% NEt3 in CH2Cl2, elution with CH2Cl2 and then
ethyl acetate) to give the title compound in 86% (476 mg, 0.73
mmol) yield. Spectroscopic data were in agreement with the
literature.47
1
3/2) Rf ) 0.4; H NMR (400 MHz, CDCl3) δ 8.76 (s, 1H, H8),
8.32 (s, 1H, H2), 8.02 (d, J ) 7.2 Hz, 2H, Ar-H), 7.59-7.55 (m,
1H, Ar-H), 7.50-7.46 (m, 2H, Ar-H), 6.43 (t, J ) 6.0 Hz, 1H,
H1′), 4.50 (q, J ) 5.5 Hz, 1H, H3′), 4.08 (q, J ) 3.9 Hz, 1H, H4′),
Reduction of Azides. 3′-Amino-N4-benzoyl-5′-O-(dimethoxy-
trityl)-2′,3′-dideoxycytidine (20) (General Protocol A). The
following protocol is representative for the conversion of azido-
nucleosides 19 and 28a/g to the aminonucleosides 20 and 29a/g.
To a solution of 3′-azido-N4-benzoyl-5′-O-(dimethoxytrityl)-2′,3′-
(67) Ogilvie, K. K. Can. J. Chem. 1973, 51, 3799–3807.
34 J. Org. Chem. Vol. 74, No. 1, 2009

3′-Amino-2′,3′-dideoxynucleosides and Primers
3.86-4.82 (m, 2H, H5′ and H5′′), 2.93-2.86 (m, 1H, H2′′),
2.64-2.58 (m, 1H, H2′), 0.90 (s, 9 H, Si-CH-CH3), 0.10 (s, 3H,
Si-CH3), 0.09 (s, 3H, Si-CH3); 13C NMR (125 MHz, CDCl3) δ
164.9, 152.5, 151.1, 149.6, 141.4, 133.7, 132.7, 128.7, 127.9, 123.4,
85.0, 84.4, 62.7, 60.4, 38.0, 25.9, 18.4, -5.3, -5.4; MS (FAB)
m/z 495.1 [M + H]+.
to -60 °C, and freshly powdered NaBH4 (270 mg, 7.10 mmol, 2 equiv)
added. After 13 h of stirring at -60 °C, acetone (11 mL) was added,
and the slurry was allowed to warm to rt and poured into ethyl acetate
(400 mL). The solution was washed with satd aqueous NaHCO3, H2O,
and brine, dried over Na2SO4, filtered, and concentrated in vacuo. The
resulting yellow foam was purified on silica (CH2Cl2/MeOH, 93/7,
v/v) to give the title compound in 77% yield (1.23 g, 2.73 mmol) as
a colorless foam: TLC (CH2Cl2/MeOH, 9/1) Rf ) 0.33. 1H NMR (500
MHz, CDCl3) δ 12.17 (b, 1H, N-H), 8.23 (bs, 1H, H8), 6.14 (dd, J )
1.9 Hz, J ) 8.5 Hz, 1H, H1′), 4.99 (b, 1H, O-H), 4.61 (bs, 1H, H3′),
4.09-3.99 (m, 3H, H4′, H5′, and H5′′), 2.82 (sept., J ) 6.9 Hz, 1H,
CH-CH3, iBu), 2.73-2.66 (m, 1H, H2′′), 2.61-2.56 (m, 1H, H2′),
1.26-1.22 (m, 6H, CH-CH3, iBu), 0.89 (s, 9H, Si-C-CH3), 0.08 (s,
3H, Si-CH3), 0.06 (s, 3H, Si-CH3); 13C NMR (125 MHz, CDCl3)
178.8, 154.8, 147.2, 146.9, 138.3, 119.3, 83.7, 82.1, 69.7, 61.3, 40.2,
35.0, 24.8, 18.0, 17.2, -6.3, -6.4; MS (FAB) m/z 452.2 [M + H]+.
3′-Azido-5′-O-tert-butyldimethylsilyl-N2-isobutyryl-2′,3′-dide-
oxyguanosine (26g). To a stirred slurry of N2-isobutyryl-5′-O-tert-
butyldimethylsilyl-2′-deoxyxyloguanosine (25g, 580 mg, 1.28 mmol)
and NaN3 (250 mg, 3.84 mmol, 3 equiv) in DMF (6 mL) was added
a solution of diphenyl-2-pyridylphosphine (504 mg, 1.93 mmol, 1.5
equiv) and DIAD (374 µL, 1.93 mmol, 1.5 equiv) in DMF (1.5 mL).
After 4 h, H2O (2 mL) was added, and the reaction mixture was
transferred into ethyl acetate. The organic solution was washed with
H2O and brine, dried over Na2SO4, filtered, and concentrated in vacuo.
The resulting brown oil was dried in vacuo and purified on silica
(CH2Cl2/ethyl acetate, 2/3, v/v) to give the title compound in 86%
yield (522 mg, 1.10 mmol): TLC (CH2Cl2/MeOH, 9/1) Rf ) 0.5; 1H
NMR (400 MHz, CDCl3) δ 12.31 (bs, 1H, N-H), 10.56 (bs, 1H, N-H),
7.96 (bs, 1H, H8), 6.03 (t, J ) 6.0 Hz, 1H, H1′), 4.31 (q, J ) 5.7 Hz,
1H, H3′), 3.97-3.93 (m, 1H, H4′), 3.84-3.72 (m, 2H, H5′ and H5′′),
2.89 (sept., J ) 6.8 Hz, 1H, CH-CH3), 2.62-2.54 (m, 1H, H2′′),
2.42-2.34 (m, 1H, H2′), 1.20-1.15 (m, 6H, CH-CH3, iBu), 0.81 (s,
9H, Si-C-CH3), 0.01 (s, 3H, Si-CH3), 0.00 (s, 3H, Si-CH3); 13C
NMR (125 MHz, CDCl3) δ 180.0, 155.9, 148.2, 148.1, 136.9, 120.9,
84.9, 83.8, 62.8, 60.2, 38.3, 36.0, 25.8, 19.0, 18.9, 18.3, -5.4, -5.5;
MS (FAB) m/z 477.1 [M + H]+.
3′-Azido-N6-benzoyl-2′,3′-dideoxyadenosine (27a). A stirred
solution of 3′-azido-N6-benzoyl-5′-O-tert-butyldimethylsilyl-2′,3′-
dideoxyadenosine (26a, 600 mg, 1.25 mmol) in THF (5 mL) was
treated with TBAF in THF (1 M solution, 2.5 mL, 2.5 mmol, 2
equiv). After 90 min, methoxytrimethylsilane (120 µL) was added,
and stirring was continued. After 1 h, the solvent was removed in
vacuo. The resulting residue was purified on silica (CH2Cl2/MeOH,
95/5, v/v) to give the title compound in 92% yield (380 mg, 1.12
mmol). Spectroscopic data were in agreement with the literature.39
3′-Azido-N6-benzoyl-5′-O-dimethoxytrityl-2′,3′-dideoxyade-
nosine (28a). A stirred solution of 3′-azido-N6-benzoyl-5′-O-tert-
butyldimethylsilyl-2′,3′-dideoxyadenosine (27a, 261 mg, 0.69 mmol)
in pyridine (5 mL) was treated with DMTrCl (303 mg, 0.90 mmol,
1.3 equiv). After 4 h, methanol (1 mL) was added, and stirring
was continued for another 30 min. The reaction mixture was added
to CH2Cl2 and the solution washed with H2O and brine, dried over
Na2SO4, filtered, and concentrated in vacuo. The yellow foam was
purified on silica (pretreated with 1% NEt3 in CH2Cl2, eluted with
1% NEt3 in CH2Cl2 and CH2Cl2/MeOH, 95/5, v/v) to give 28a in
96% (450 mg, 0.66 mmol) yield: TLC (CH2Cl2/MeOH, 95/5) Rf )
0.6; 1H NMR (400 MHz, CD3CN) δ 8.58 (s, 1H, H8), 8.29 (s, 1H,
H2), 8.00-7.94 (m, 2H, Ar-H), 7.59-7.55 (m, 1H, Ar-H),
7.48-7.43 (m, 2H, Ar-H), 7.36-7.35 (m, 2H, Ar-H), 7.28-7.15
(m, 7H, Ar-H), 6.82-6.76 (m, 4H, Ar-H), 6.39 (dd, J ) 4.5, J )
4.7 Hz, 1H, H1′), 4.72 (q, J ) 6.4 Hz, 1H, H3′), 4.10-4.06 (m,
1H, H4′), 3,72 (s, 6H, O-CH3), 3.42-4.40 (m, 2, H5′ and H5′′),
3.15-3.08 (m, 1H, H2′′), 2.63-2.54 (m, 1H, H2′); 13C NMR (125
MHz, CD3CN) δ 166.0, 159.3, 159.3, 145.6, 143.3, 136.4, 136.3,
133.2, 130.6, 130.6, 129.2, 128.8, 128.6, 128.5, 127.5, 125.4, 118.1,
113.7, 86.9, 84.8, 84.3, 63.8, 61.3, 55.6, 36.8; MS (FAB) m/z 683.4
[M + H]+.
3′-Azido-N2-isobutyryl-2′,3′-dideoxyguanosine (27g). A stirred
solution of 3′-azido-5′-O-tert-butyldimethylsilyl-N2-isobutyryl-2′,3′-
dideoxyguanosine (26g, 200 mg, 0.42 mmol) in THF (2 mL) was
treated with a solution of TBAF in THF (1 M, 844 µL, 0.844 mmol,
2 equiv). After 2 h, the solvent was removed in vacuo and the
residue purified on silica (CH2Cl2/MeOH, 9/1, v/v) to give the title
compound in 87% yield (133 mg, 0.37 mmol) as a colorless solid.
Spectroscopic data were in agreement with the literature.30
3′-Azido-5′-O-dimethoxytrityl-N2-isobutyryl-2′,3′-dideoxygua-
nosine (28g). A stirred solution of 3′-azido-N2-isobutyryl-2′,3′-
dideoxyguanosine (27g, 132 mg, 0.37 mmol) in pyridine (3 mL)
was treated with DMTrCl (161 mg, 0.48 mmol, 1.3 equiv). After
3 h, MeOH (500 µL) was added, and stirring was continued. After
1 h, the reaction mixture was transferred into CH2Cl2. The solution
was washed with H2O and brine, dried over Na2SO4, filtered, and
concentrated in vacuo. The resulting yellow oil was purified on
silica (pretreated with 1% NEt3 in CH2Cl2, and eluted with 1%
NEt3 in CH2Cl2 to 1% NEt3 in CH2Cl2/MeOH 95/5, v/v) to give
the title compound in 79% yield (191 mg, 0.29 mmol) as a colorless
foam. Spectroscopic data were in agreement with the literature.30
3′-Amino-5′-O-dimethoxytrityl-N2-isobutyryl-2′,3′-dideoxy-
guanosine (29g). Compound 29g was prepared according to
general protocol A using 3′-azido-N2-isobutyryl-5′-O-dimeth-
oxytrityl-2′,3′-dideoxyguanosine (28g, 190 mg, 0.29 mmol), ethanol
(5 mL), satd aqueous NaHCO3 (100 µL), and Pd/C (60 mg,). A
yield of 95% of 29g (179 mg, 0.27 mmol) was obtained in the
form of a slightly yellow foam. Spectroscopic were data in
agreement with the literature.27
3′-Amino-N6-benzoyl-5′-O-dimethoxytrityl-2′,3′-dideoxyade-
nosine (29a). Compound 21a was prepared according to general
protocol A, using 3′-azido-N2-isobutyryl-5′-O-dimethoxytrityl-2′,3′-
dideoxyguanosine (28a, 300 mg, 0.30 mmol), ethanol (5 mL), satd
aqueous NaHCO3 (100 µL), and Pd/C (60 mg). Compound 29a
was obtained as a slightly yellow foam in 90% (260 mg, 0.40 mmol)
yield. Spectroscopic data were in agreement with the literature.27
5′-O-tert-Butyldimethylsilyl-N2-isobutyryl-2′-deoxyguanosine
(24g). A stirred solution of N2-isobutyryl-2′-deoxyguanosine (23g,
1.5 g, 4.45 mmol) in pyridine (10 mL) was treated with TBS-Cl (773
mg 4.89 mmol, 1.1 equiv). After 16 h, the reaction mixture was
transferred into ethyl acetate, and the solution was washed with H2O
and brine, dried over Na2SO4, filtered, and concentrated in vacuo to
give a yellow gum. After coevaporation from toluene, the resulting
foam was purified on silica (CH2Cl2/MeOH, 9/1, v/v) to give 24g in
84% yield (1.68 g 3.67 mmol) as a colorless foam: TLC (CH2Cl2/
MeOH, 9/1) Rf ) 0.31; 1H NMR (400 MHz, DMSO-d6) δ 8.13 (bs,
1H, H8), 6.19 (t, J ) 6.6 Hz, 1H, H1′), 5.34 (d, J ) 3.8 Hz, 1H,
O-H), 4.37-4.31 (m, 1H, H3′), 3.85-3.81 (m, 1H, H4′), 3.70-3.64
(m, 2H, H5′ and H5′′), 2.74 (sept, J ) 6.8 Hz, 1H, CH-CH3, iBu),
2.59-2.51 (m, 1H, H2′′), 2.35-2.25 (m, 1H, H2′), 1.10 (d, J ) 6.8
Hz, 6H, CH-CH3, iBu), 0.83 (s, 9H, Si-C-CH3), 0.00 (s, 6H,
Si-CH3); 13C NMR (100 MHz, DMSO-d6) δ 180.0, 154.7, 148.2,
148.0, 137.0, 120.2, 87.1, 82.9, 70.0, 63.1, 34.7, 25.7, 18.8, 18.7, 17.9,
-5.4, -5.5; MS (FAB) m/z 452.3 [M + H]+.
5′-O-tert-Butyldimethylsilyl-N2-isobutyryl-2′-deoxyxylogua-
nosine (25g). To a chilled solution (0 °C) of Dess-Martin periodinane
in CH2Cl2 (10 wt%, 11 mL, 5.32 mmol, 1.5 equiv) was added N2-
isobutyryl-5′-O-tert-butyldimethylsilyl-2′-deoxyguanosine (24g, 1.6 g,
3.55 mmol) and the resulting solution stirred at 0 °C. After 60 min,
the cooling bath was removed, and stirring was continued at rt. After
3 h, 2-propanol (11 mL) was added, the resulting white slurry cooled
Loading of Nucleosides on Solid Supports. Solid Support
32a (General Protocol B). The following protocol is representative
for the synthesis of the solid supports 32a-t. LCAA-cpg (31, 1.4 g,
loading 75 µmol/g) was coevaporated twice from toluene and dried
J. Org. Chem. Vol. 74, No. 1, 2009 35

Eisenhuth and Richert
in vacuo for 12 h. Under an argon atmosphere, CH2Cl2 (12 mL)
was added, followed by the addition of DIEA (406 µL, 1.3 mmol,
12 equiv compared to the NH2-loading given by the manufacturer).
Hexafluoroglutaric anhydride (145 µL, 1.1 mmol, 10 equiv) was
added, and the slurry was shaken at rt for 12 h. The cpg was filtered
off, washed with DMF containing 1% NEt3, MeOH, and CH2Cl2
(5 × 10 mL each), and dried in vacuo. Aminonucleoside 29a (260
mg, 315 µmol, 3 equiv) was added, and the resulting mixture was
coevaporated twice from toluene, followed by drying in vacuo. After
addition of DMF (8 mL), PPh3 (173 mg, 665 µmol, 5 equiv), and
DIEA (406 µL, 1.3 mmol, 12 equiv) to the residue, a solution of
iodine (169 mg, 665 µmol, 5 equiv) in DMF (2 mL) was added
dropwise while gently swirling the round-bottom flask. The resulting
red slurry was swirled gently on a shaker. After 16 h, the cpg was
filtered off and the slurry was washed with DMF containing 1%
NEt3, MeOH, and CH2Cl2 (5 × 10 mL each) and dried in vacuo.
The resulting loading of the solid support 32a, determined by DMT-
cation release, was 18 µmol/g.
Solid Support 32c. Solid support 32c was synthesized by following
the general protocol B. The quantities used were LCAA-cpg (31,
1.32 g, loading 95 µmol/g), 3′-amino-N4-benzoyl-5′-O-dimethoxytrityl-
2′,3′-dideoxycytidine (20, 240 mg, 376 µmol, 3 equiv), PPh3 (162 mg,
627 µmol, 5 equiv), DIEA (406 µL, 1.5 mmol, 12 equiv), I2 (156 mg,
626 µmol, 5 equiv), and DMF (10 and 2 mL for the iodine solution).
The loading of the resulting solid support (32c) was determined to be
of 24 µmol/g by trityl essay.
Solid Support 32g. Solid support 32g was synthesized by
following the general protocol B. The quantities used were LCAA-
cpg (31, 1.1 g, loading 75 µmol/g), 3′-amino-5′-O-dimethoxytrityl-
N2-isobutyryl-2′,3′-dideoxyguanosine (29g, 164 mg, 247 µmol, 3
equiv), PPh3 (125 mg, 413 µmol, 5 equiv), DIEA (255 µL, 940
µmol, 12 equiv), I2 (120 mg, 413 µmol, 5 equiv), and DMF (8 and
2 mL for the iodine solution). The loading of the resulting solid
support (32g) was determined to be of 17 µmol/g by trityl essay.
Solid Support 32t. Solid support 32t was synthesized by
following the general protocol B. The quantities used were LCAA-
cpg (31, 1.3 g, loading 75 µmol/g), 3′-amino-5′-O-dimethoxytrityl-
3′-deoxythymidine (11, 160 mg, 293 µmol, 3 equiv), PPh3 (128
mg, 488 µmol, 5 equiv), DIEA (317 µL, 1.17 mmol, 12 equiv), I2
(126 mg, 488 µmol, 5 equiv), and DMF (10 and 2 mL for the iodine
solution). The loading of the resulting solid support (32t) was
determined to be of 28 µmol/g by trityl essay.
DNA Chain Assembly (General Procedure C). Starting from
the support loaded with the 3′-amino-2′,3′-dideoxynucleoside
(32a-t), the chain of the 3′-aminoterminal primers 33a-t was
generated by standard phosphoramidite-based DNA synthesis,
following the protocol recommended by the manufacturer of the
synthesizer (8909 Expedite DNA synthesizer, system software 2.01,
16 equiv of phosphoramidite per coupling cycle). The DNA chains
were then fully deprotected and cleaved from the solid support by
treatment with 30% aqueous ammonia at 55 °C for 20 h. Excess
ammonia was removed with a gentle stream of air and the solution
lyophilized to dryness. The resulting crudes were either analyzed
by anion-exchange HPLC (IE-HPLC) using a gradient of 1 M KCl
in 10 mM NaOH, starting at 0% KCl-containing solution for 5 min
to 20% of the KCl-containing solution in 40 min (retention times
are given as tR) or bulk purified using SepPak-cartridge (Waters,
Eschborn, Germany) with a stepwise gradient of CH3CN in 0.1 M
triethylammonium acetate, pH 7.
CGCACGT* (33t): tR ) 13.8 min; MALDI-TOF MS calcd for
[M - H]- 2078.4.4, found 2079.1; yield (IE-HPLC) 33%; yield
(cartridge) 20%.
CGCACGT (34): tR ) 13.9 min; MALDI-TOF MS calcd for
[M - H]- 2079.4, found 2082.2; yield 150 nmol (50%).
5′-Phosphorylation of Azidonucleosides. 3′-Azido-N6-benzoyl-
2′,3′-dideoxyadenosine-5′-O-dibenzylphosphate (38a) (General
Procedure D). The following protocol is representative for the
synthesis of the phosphotriesters 38a, 38c, and 38t. Azidonucleoside
29a (150 mg, 0.42 mmol) was coevaporated twice from pyridine.
The resulting oily residue was dissolved in pyridine (5 mL) and
the solution cooled to -30 °C. After addition of freshly prepared
dibenzylphosphoryl chloride68 (400 µL, 1.89 mmol, 4.5 equiv), the
solution was stirred at -30 °C for 5 h. Once TLC indicated
complete conversion, the reaction mixture was added to ethyl
acetate, and the resulting solution was washed with citric acid (5%
aqueous solution) and brine, dried over Na2SO4, filtered, and
concentrated in vacuo. The resulting crude was purified via column
chromatography on silica (CH2Cl2/ethyl acetate, 6/4, v/v) to give
3′-azido-N6-benzoyl-2′,3′-dideoxyadenosine-5′-O-dibenzylphos-
phate (38a) in 69% (180 mg, 0.41 mmol) yield: TLC (CH2Cl2/
1
ethyl acetate, 6/4) Rf ) 0.27; H NMR (500 MHz, CDCl3) δ 9.22
(b, 1H, N-H),8.76 (s, 1H, H8), 8.20 (s, 1H, H2), 8.04 (d, J ) 7.2
Hz, 2H, Ar-H), 7.64-7.59 (m, 1H, Ar-H), 7.55-7.51 (m, 2H, Ar-
H), 7.36-7-32 (m, 10H, Ar-H), 6.37 (t, J ) 6.4 Hz, 1H, H1′),
5.09-5.00 (m, 4, O-CH2-Ph), 4.39-4,35 (m, 1H, H3′), 4.27-4.07
(m, 3H, H4′, H5,′ and H5′′), 2.92-2.87 (m, 1H, H2′′), 2.56-2.60
(m, 1H, H2′); 13C NMR (125 MHz, CDCL3) δ 164.7, 152.6, 151.2,
149.7, 141.5, 135.5, 135.4, 135.4, 135.4, 133.6, 132.8, 128.8, 128.8,
128.7, 128.6, 128.6, 128.1, 128.1, 127.9, 123.6, 82.8, 82.7, 69.7,
69.7, 69.7, 69.7, 66.2, 66.1, 60.7, 37.1; 31P NMR (101 MHz, CDCl3)
δ -0.04 ppm; MS (MALDI-TOF) m/z 641.4 [M + H]+.
3′-Azido-N4-benzoyl-2′,3′-dideoxycytidine-5′-O-dibenzylphos-
phate (38c). Phosphotriester 38c was prepared according to general
protocol D, using 3′-azido-N4-benzoyl-2′,3′-dideoxycytidine (18,
140 mg, 0.39 mmol), pyridine (4 mL), and dibenzylphosphoryl
chloride (370 µL, 1,76 mmol, 4.6 equiv). The yield of 38c, after
column chromatography on silica (0.1% water in CH2Cl2/ethyl
acetate, 6/4, v/v), was 74% (180 mg, 0.41 mmol): TLC (CH2Cl2/
1
MeOH, 95/5) Rf ) 0.38; H NMR (500 MHz, CDCl3) δ 8.74 (b,
1H, N-H), 8.04 (d, J ) 7.5 Hz, 1H, H6), 7.96-7.91 (m, 2H, Ar-
H), 7.67-7.63 (m, 1H, Ar-H), 7.59-7.50 (m, 3H, Ar-H and H5),
7.43-7-35 (m, 10H, Ar-H), 6.13 (t, J ) 6.0 Hz, 1H, H1′),
5.13-5.03 (m, 4,O-CH2-Ph), 4.26-4-11 (m, 2H, H5′ and H5′′),
4.02-3.99 (m, 1H, H4′), 3.97-3.92 (m, 1H, H3′), 2.67-2.61 (m,
1H, H2′′), 2.20-2.14 (m, 1H, H2′); 13C NMR (125 MHz, CDCL3)
δ135.4, 135.4, 135.4, 135.3, 133.2, 129.1, 128.9, 128.8, 128.8,
128.3, 128.2, 86.9, 82.9, 82.9, 70.0, 69.9, 69.9, 69.9, 65.7, 65.7,
59.3, 38.6; 31P NMR (101 MHz, CDCl3) δ 0.35 ppm. MS (MALDI-
TOF): 616 [M+H]+.
3′-Azido-2′,3′-dideoxythymidine-5′-O-dibenzylphosphate (38t).
Phosphotriester 38t was prepared according to general protocol D,
using 3′-azido-3′-deoxythymidine (1, 120 mg, 0.52 mmol), pyridine
(3 mL), and dibenzylphosphoryl chloride (480 µL, 2.37 mmol, 4.6
equiv). The yield of 38t, after column chromatography on silica
(CH2Cl2/MeOH, 95/5, v/v), was 79% (186 mg, 0.41 mmol): TLC
1
(CH2Cl2/MeOH, 95/5) Rf ) 0.31; H NMR (500 MHz, CDCl3) δ
9.10 (b, 1H, N-H), 7.42-7-36 (m, 11H, H6 and Ar-H), 6.15 (t, J
) 6.0 Hz, 1H, H1′), 5.12-5.08 (m, 4, O-CH2-Ph), 4.30-4.18 (m,
3H, H3′, H5′, and H5′′), 4.02-3.98 (m, 1H, H4′), 2.39-2.25 (m,
2H, H2′ and H2′′) 1.79 (d, J ) 1.3 Hz, 3H, CH3); 13C NMR (125
MHz, CDCl3) δ 163.5, 150.1, 135.7, 135.6, 135.4, 128.2, 127.6,
110.2, 84.1, 81.5, 81.4, 69.0, 69.0, 68.9, 68.9, 66.1, 66.1, 59.6, 36.1,
11.2; 31P NMR (101 MHz, CDCl3) δ 0.46 ppm; MS (MALDI-
TOF) ) 538.3 [M + H]+.
Analytical Data for Oligonucleotides. CGCACGA* (33a): tR
) 13.0 min; MALDI-TOF MS calcd for [M - H]- 2087.4, found
2087.9; yield (IE-HPLC) 37%; yield (cartridge) 24%.
CGCACGC* (33c): tR ) 13.9 min; MALDI-TOF MS calcd for
[M - H]- 2063.4, found 2064.1; yield (IE-HPLC) 47%; yield
(cartridge) 15%.
CGCACGG* (33g): tR ) 14.6 min; MALDI-TOF MS calcd
for [M - H]- 2103.4, found 2105.2; yield (IE-HPLC) 43%; yield
(cartridge) 21%.
(68) Gao, F.; Yan, X.; Shakya, T.; Baettig, O. M.; Ait-Mohand-Brunet, S.;
Berghuis, A. M.; Wright, G. D.; Auclair, K. J. Med. Chem. 2006, 49, 5273–
5281.
36 J. Org. Chem. Vol. 74, No. 1, 2009

3′-Amino-2′,3′-dideoxynucleosides and Primers
Hydrogenation of Phosphotriester Azides. 3′-Amino-N6-
benzoyl-2′,3′-dideoxyadenosine-5′-monophosphate (39a) (Gen-
eral Procedure E). The following protocol is representative for the
synthesis of the aminophosphates 39a, 39c, and 39t. A solution of
3′-azido-N6-benzoyl-2′,3′-dideoxyadenosine-5′-O-dibenzylphosphate
(38a, 125 mg, 0.20 mmol) in ethanol (5 mL) was treated with a satd
aqueous solution of NaHCO3 (100 µL). The resulting slurry was placed
under an argon atmosphere, and Pd/C (20 mg) was added. After
introducing a hydrogen atmosphere, the slurry was stirred at the
hydrostatic pressure generated by a 20 cm water column until TLC
(CH2Cl2/MeOH, 9/1, v/v) indicated complete conversion of the starting
material. Then, an equal volume of H2O (v/v) was added, and the
hydrogenation process was continued until only one spot was detectable
by TLC (2-propanol/water/NEt3, 7/2/1, v/v/v). The catalyst was
removed by filtration over a bed of Celite and washed with H2O. The
combined solutions were lyophilized to dryness to give the title
compound in 81% yield (66 mg, 0.16 mmol): TLC (2-propanol/water/
NEt3, 7/2/1, v/v/v) Rf ) 0.07-0.1; 1H NMR (600 MHz, D2O) δ 8.65
(s,1H, H8), 8.58 (s,1H, H2), 7.86-7.83 (m, 2H, Ar-H), 7.57-7.53
(m, 1H, Ar-H), 7.45-7.42 (m, 2H, Ar-H), 6.50-6.48 (m, 1H, H1′),
4.20-4.17 (m, 1H, H4′), 4.95-3.88 (m, 2H, H5′ and H5′′), 4.20-4.17
(m, 1H, H4′), 4.04-4.00 (m, 1H, H3′), 2.88-2.83 (m, 1H, H2′′),
2.61-2.56 (m, 1H, H2′); 13C NMR (125 MHz, D2O) δ 169.0, 151.7,
151.4, 148.9, 143.4, 133.3, 132.7, 128.7, 128.0, 123.8, 83.7, 63.5, 63.5,
50.9, 37.6; 31P NMR (101 MHz, D2O) δ 4.65; MS (MALDI-TOF)
433.2 (M - H)-.
3′-Amino-N4-benzoyl-2′,3′-dideoxycytidine-5′-monophos-
phate (39c). Compound 39c was prepared according to general
protocol E, using 3′-azido-N4-benzoyl-2′,3′-dideoxycytidine-5′-O-
dibenzylphosphate (38c, 170 mg, 0.27 mmol), ethanol (5 mL), satd
aqueous NaHCO3 (100 µL), and H2O (5 mL). Aminonucleotide
39c was obtained in 83% yield (102 mg, 0.22 mmol) as a white
powder: TLC (2-propanol/water/NEt3, 7/2/1, v/v/v) Rf ) 0.1-0.15.
1H NMR (400 MHz, D2O) δ 8.33 (d, J ) 7.6 Hz, 1H, H6),
7.78-7.75 (m, 2H, Ar-H), 7.58-7.53 (m, 1H, Ar-H), 7.46-7.41
(m, 2H, Ar-H), 7.38 (d, J ) 7.6 Hz, 1H, H5), 6.15 (t, J ) 6.1 Hz,
1H, H1′), 4.38-4.34 (m, 1H, H4′), 4.04-3.97 (m, 3H, H3′, H5′,
and H5′′), 2.72-2.52 (m, 1H, H2′ and H2′′); 13C NMR (100 MHz,
D2O) 169.7, 163.2, 156.7, 145.7, 133.4, 132.6, 128.8, 128.7, 127.9,
127.3, 98.5, 86.9, 50.1, 36.4; 31P NMR (101 MHz, D2O) δ 3.33
ppm; MS (MALDI-TOF) 409.2 (M - H)-.
a yellow foam: TLC (0.1% water in petroleum ether/ethyl acetate,
1.5/1) Rf ) 0.2; 1H NMR (300 MHz, CDCl3) δ 8.13 (s, 1H), 7.90
(bs, 1 h), 7.53-7.30 (m, 5H), 6.31 (t, J ) 6.0 Hz, 1H), 5.62 (s,
2H), 4.64-4.57 (m, 1H), 4.07-4.02 (m, 1H), 3.96-3.80 (m, 2H),
3.15-3.01 (b, 1H), 2.91-2.80 (m, 1H), 2.61-2.50 (m, 1H),
1.28-1.25 (m, 6H), 0.89 (s, 9H), 0.08 (d, 6H); 13C NMR (75 MHz,
CDCl3) δ 135.9, 128.5, 128.3, 128.2, 85.1, 84.2, 77.5, 77.0, 76.6,
68.8, 62.9, 60.3, 38.1, 25.9, 22.0, 19.3, 18.4, -5.4, -5.5; HRMS
(ESI-Tof) m/z calcd for C27H38N8NaO4Si [M + Na]+ 589.2677,
m/z obsd 589.2668.
3′-Azido-O6-benzyl-N2-isobutyryl-2′,3′-dideoxyguanosine (41).
A stirred solution of 3′-azido-O6-benzyl-5′-O-tert-butyldimethyl-
silyl-N2-isobutyryl-2′,3′-dideoxyguanosine (40, 110 mg, 0.19 mmol)
in THF (1 mL) was treated with TBAF in THF (1 M solution, 230
µL, 0.23 mmol, 1.2 equiv). After 2 h, the resulting yellow solution
was directly objected to purification on silica (CH2Cl2/MeOH, 96/
4, v/v) to give the title compound in 80% yield (69 mg, 0.15 mmol)
as a colorless foam: TLC (CH2Cl2/MeOH, 95/5) Rf ) 0.43; 1H NMR
(300 MHz, CDCl3) δ 7.96 (b, 1H), 7.89 (b, 1H), 7.43-7.37 (m,
2H), 7.31-7.22 (m, 3H), 6.10 (t, J ) 6.5 Hz, 1H), 5.51 (s, 2H),
4.93-4.86 (m, 1H), 4.00-3.96 (m, 1H), 3.90-3.84 (m, 1H),
3.73-3.66 (m, 1H), 2.98-2.87 (m, 1H), 2.84-2.73 (m, 1H),
2.41-2.31 (m, 1H), 1.22-1.17 (m, 6H); 13C NMR (100 MHz,
CDCl3) δ 174.5, 159.8, 150.7, 150.5, 140.3, 134.6, 127.5, 127.3,
118.4, 84.9, 84.7, 75.6, 67.9, 61.4, 60.1, 36.8, 35.2, 24.6, 20.9, 18.3,
17.0; HRMS (ESI-Tof) m/z calcd for C21H24N8NaO4 [M + Na]+
475.1813, m/z obsd 475.1806.
3′-Azido-O6-benzyl-N2-isobutyryl-2′,3′-dideoxyguanosine-5′-
O-dibenzylphosphate (42). Phosphotriester 42 was prepared accord-
ing to general protocol D, using 3′-azido-O6-benzyl-5′-O-tert-
butyldimethylsilyl-N2-isobutyryl-2′,3′-dideoxyguanosine (41, 69 mg,
0.15 mmol), pyridine (1.5 mL), and dibenzylphosphoryl chloride (145
µL, 0.69 mmol, 4.6 equiv). The yield of 42, after column chromatog-
raphy on silica (first 0.1% water in CH2Cl2/ethyl acetate, 4/1, v/v, then
ethyl acetate), was 66% (70 mg, 0.1 mmol): TLC (CH2Cl2/MeOH,
1
95/5) Rf ) 0.27; H NMR (300 MHz, CDCl3) δ 8.69 (b, 1H), 7.83
(1H), 7.71-7.63 (m, 2H), 7.55-7.42 (m, 4H), 7.34-7.22 (m, 9H),
6.15 (t, j ) 6.3 Hz, 1H), 5.63 (s, 2H), 5-06-4.95 (m, 4H), 4.46-4.37
(m, 1H), 4.17-4.05 (m, 2H), 3.09-2.98 (m, 1H), 2.95-2.83 (m, 1H),
2.47-2.37 (m, 1H) 1.27-1.23 (m, 4H); 13C NMR (75 MHz, CDCl3)
δ 175.6, 160.8, 152.0, 151.9, 140.8, 135.9, 135.6, 135.5, 135.4, 132.2,
132.0, 132.0, 131.9, 128.7, 128.6, 128.5, 128.4, 128.2, 128.1, 127.9,
119.1, 85.1, 83.0, 82.9, 69.7, 69.7, 69., 69.5, 68.8, 66.6, 66.5, 61.0,
36.6, 36.0, 19.4; 31P NMR (121 MHz, CDCl3) δ -0.78 ppm; HRMS
(ESI-Tof) m/z calcd for C35H37N8NaO7P [M + Na]+ 735.2412, m/z
obsd 735.2412.
3′-Amino-3′-deoxythymidine-5′-monophosphate (39t). Com-
pound 39t was prepared according to general protocol E, using 3′-
azido-3′-deoxythymidine-5′-O-dibenzylphosphate (38t, 93 mg, 0.18
mmol), ethanol (5 mL), satd aqueous NaHCO3 (100 µL), and H2O
(5 mL). Aminonucleotide 39t was obtained in 78% yield (47 mg,
0.14 mmol) as a white powder: TLC (2-propanol/water/NEt3, 7/2/
1, v/v/v) Rf ) 0.1-0.13; 1H NMR (500 MHz, D2O) δ 7.62 (s, 1H,
H6), 6.20, (t, J ) 6.6 Hz, 1H, H1′), 4.16-4.30 (m, 1H, H4′),
3.98-3.87 (m, 3H, H3′, H5′, and H5′′), 2.55-2.48 (m, 1H, H2′),
2.46-2.39 (m, 1H, H2′′); 13C NMR (125 MHz, D2O) δ 137.6, 84.7,
82.2, 82.2, 63.4, 63.4, 50.8, 35.6, 11.5; 31P NMR (101 MHz, D2O)
δ 4.41 ppm; MS (MALDI-TOF) 320.3 (M - H)-.
Acknowledgment. We are grateful to A. Hochgesand for
performing chemical primer extension reactions, E. Kervio for
helpful discussions, and C. Deck, A. Singh, and R. Haug for a
review of the manuscript. This work was supported by DFG
Grant No. RI 1063/1-4 to C.R. and the Center for Functional
Nanostructures at the University of Karlsruhe (CFN).
3′-Azido-O6-benzyl-5′-O-tert-butyldimethylsilyl-N2-isobutyryl-
2′,3′-dideoxyguanosine (40). To a stirred solution of 3′-azido-5′-
O-tert-butyldimethylsilyl-N2-isobutyryl-2′,3′-dideoxyguanosine
(26g, 440 mg, 0.92 mmol) and benzyl alcohol (143 µL, 1.38 mmol,
1.5 equiv) in DMF (5 mL) was added a solution of PPh3 (363 mg,
1.38 mmol, 1.5 equiv) and DIAD (197 µL, 1.38 mmol, 1.5 equiv)
in DMF (2 mL). After 4 h, H2O was added, and stirring was
continued for another 2 h. The reaction mixture was transferred
into ethyl acetate and the organic solution washed with H2O and
brine, dried over Na2SO4, filtered, and concentrated in vacuo. The
resulting yellow gum was purified on silica (column length 10 cm,
eluant 0.1% water in petroleum ether/ethyl acetate, 1.5/1, v/v) to
give the title compound 40 in 30% yield (167 mg, 0.29 mmol) as
Supporting Information Available: 1H NMR of compounds
1, 9-11, 14-22, 24a,g, 25a,g, 26a,g, 27a,g, 28a,g, 29a,g, 30t,
38a,c,t, 39a,c,t, and 40-42. 13C NMR spectra of compounds
16-18, 25a, 28a, 24g, 25g, 26g, 38a,c,t, 39a,c,t, and 40-42.
Procedures for the synthesis of compounds 11, 15, and 16.
MALDI time-of-flight spectra and HPLC trace-analysis chro-
matograms of 3′-aminoterminal primers 33a-t and primer
extension reactions involving them. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO8018889
J. Org. Chem. Vol. 74, No. 1, 2009 37