Commercially available 5'-DMT phosphoramidites as reagents for the synthesis of vinylphosphonate-linked oligonucleic acids.

Article abstract of DOI:10.1021/ol0166045

[reaction: see text]. Commercially available cyanoethyl phosphoramidites derived from T, d(C), d(A), and d(G) were hydrolyzed (1H-tetrazole, MeCN/H2O) to give the corresponding H-phosphonates in excellent yields. Palladium(0)-catalyzed cross-coupling of e

Full text of DOI:10.1021/ol0166045

ORGANIC
LETTERS
2001
Vol. 3, No. 21
3365-3367
Commercially Available 5′-DMT
Phosphoramidites as Reagents for the
Synthesis of Vinylphosphonate-Linked
Oligonucleic Acids
Sahar Abbas, Richard D. Bertram, and Christopher J. Hayes*
The School of Chemistry, The UniVersity of Nottingham,
UniVersity Park, Nottingham NG7 2RD, U.K.
chris.hayes@nottingham.ac.uk
Received August 17, 2001
ABSTRACT
Commercially available cyanoethyl phosphoramidites derived from T, d(C), d(A), and d(G) were hydrolyzed (1H-tetrazole, MeCN/H O) to give
2
the corresponding H-phosphonates in excellent yields. Palladium(0)-catalyzed cross-coupling of each of these with the thymidine-derived
vinylbromide 2 afforded the corresponding vinylphosphonate-linked dimers T*T, d(C)*T, d(A)*T, and d(G)*T in modest to good yields. The T*T
dimer was further elaborated to give a 5′-DMT-T*T-3′-CEP building block, and this was used in the automated synthesis of the TpT*TpT
tetramer.
The ability of exogenous synthetic oligonucleotides to
modulate the function of specific genes in vitro and in vivo
has stimulated a considerable amount of research into their
use as therapeutic agents and diagnostic tools.1 A large
amount of this work has focused upon chemical modification
of the internucleotide linkage in order to address problems
associated with nuclease degradation of the naturally occur-
ring phosphodiester moiety. Recently we reported a method
for the synthesis of modified T*T and d(A)*T dimers
containing a hydrolytically stable vinylphosphonate inter-
nucleotide linkage 3. The key step utilizes a palladium(0)-
catalyzed coupling reaction between the thymidine-derived
1-bromo-1-alkene 2 and nucleotide-derived H-phosphonates
(e.g., 1) to yield the corresponding dimers 3 in moderate to
high yield (Scheme 1).2 We found that under our optimized
conditions (Pd(OAc)2, dppf, propylene oxide,3 THF) a range
of functional groups were tolerated, including the base-labile
cyanoethyl group. We now wish to report further results in
this area, which have led to the synthesis of building blocks
suitable for oligonucleotide construction. Much of our early
development work was performed on materials possessing
silicon-based protecting groups on the 5′- and 3′-hydroxyls,
but we were aware of the need to manipulate these termini
into functionality that would be more compatible with
(1) (a) Uhlmann, E.; Peyman, A. Chem. ReV. 1990, 90, 543. (b)
Szymkowski, D. E. Drug DiscoVery Today 1996, 1, 415. (c) Englisch, U.;
Gauss, D. H. Angew. Chem., Int. Ed. Engl. 1991, 30, 613. (d) Uhlmann,
E.; Peyman, A.; Breipohl, G.; Will, D. W. Angew. Chem., Int. Ed. 1998,
37, 2797.
(2) (a) Abbas, S.; Hayes, C. J. Tetrahedron Lett. 2000, 41, 4513. (b)
Abbas, S.; Hayes, C. J., Synlett 1999, 1124. For an alternative route to
vinylphosphonate-linked nucleic acids, see also: (c) Zhao, Z.; Caruthers,
M. H. Tetrahedron Lett. 1996, 37, 6239.
(3) Lei, H.; Stoakes, M. S.; Schwabacher, A. W. Synthesis 1992, 1255.
10.1021/ol0166045 CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/25/2001

Scheme 1a
Scheme 2
a Reagents: (a) Pd(OAc)2, propylene oxide, THF, 70 °C (78%);
(b) HF (48% aq), MeCN, rt (73%).
oligonucleotide synthesis (i.e., 5′-DMT and 3′-phosphor-
amidite). To this end, we quickly found that the 5′-TBS group
could be removed cleanly from 3 using HF (48% aqueous)
in MeCN, thus revealing the free 5′-hydroxyl group 4. We
hoped to protect 4 as the corresponding 5′-DMT derivative
5, but after much effort we were unable to find suitable
conditions to achieve this transformation (Scheme 1). Initially
we found this result very surprising, especially since very
similar protections have been reported in the literature, but
further experimentation showed the 5′-hydroxyl of 4 to be
resistant to reaction with a range of other electrophiles.4 The
reason for this lack of reactivity was not particularly clear,
but comparison to successful DMT-protections on analogous
substrates seemed to indicate that the 3′-TBDPS group was
somehow responsible.
Coupling of each of these H-phosphonates (1.4 equiv) with
the thymidine-derived vinyl bromide 2, using our previously
developed conditions (Pd(OAc)2 (0.2 equiv), dppf (0.4
equiv), propylene oxide (10 equiv), THF), provided the
corresponding vinylphosphonate-linked dimers (T*T 14,
d(C)*T 15, d(A)*T 16, and d(G)*T 17)7 in modest to good
yields (Scheme 2). The all pyrimidine-containing dimers 14
Scheme 3a
To overcome the frustration of these initial results, we
decided to examine a much more direct solution to the “5′-
DMT problem”. We wondered whether H-phosphonates that
already contained the 5′-DMT moiety could be used as
coupling partners in the cross-coupling reaction, as this would
avoid the troublesome protecting group swap. This was
particularly attractive as the corresponding cyanoethyl-
protected phosphoramidite precursors were commercially
available, and the use of these materials would remove two
unnecessary steps from the synthetic sequence. We were
pleased to find that hydrolysis (1H-tetrazole, MeCN/H2O)
of each of the four commercially available5 5′-DMT-
cyanoethyl phosphoramidites derived from thymidine 6, 2′-
deoxycytidine 7, 2′-deoxyadenosine 8, and 2′-deoxygua-
nosine 9, provided the corresponding H-phosphonates 10,
11, 12, and 13, respectively, in uniformly high yield
(∼100%). As a result of the efficiency of this hydrolysis
reaction, purification of the H-phosphonates was not neces-
sary and these materials could be used in crude form for the
subsequent cross-coupling reactions (Scheme 2).6
a Reagents: (a) TBAF, THF (99%); (b) Pd(OAc)2 (0.2 equiv),
dppf (0.4 equiv), propylene oxide (10 equiv), THF, 70 °C (59%);
(c) 1H-tetrazole, 20, CH2Cl2 (70%).
(4) Caruthers and Zhao (ref 2c) have reported the 5′ -DMT protection
of a vinylphosphonate-linked nucleic acid. Other attempts were made to
acylate (Ac2O or AcCl, pyridine, DMAP) and TBS protect (TBS-Cl,
imidazole, DMF, and TBS-OTf, Pr2NEt, CH2Cl2) the 5′-OH, but these
efforts were unsuccessful.
(5) Materials purchased from Cruachem Ltd. and used without further
purification.
and 15 were produced in very pleasing 73% and 61% yields,
respectively. Gratifyingly, coupling of the adenosine-derived
H-phosphonate 12 with 2 afforded the mixed purine-
pyrimidine vinylphosphonate-linked dimer 16 in an excellent
85% yield. As we anticipated at the outset of this work, the
i
3366
Org. Lett., Vol. 3, No. 21, 2001

coupling of the 2′-deoxyguanosine-derived H-phosphonate
13 proved to be the most challenging. After a reaction time
of 7 h we were able to isolate 31% (67% based on recovered
2) of the desired coupled product 17. We are currently
examining the use of alternative catalyst systems and
protecting group strategies in order to improve this 2′-
deoxyguanosine coupling, and these results will be reported
in due course. Having demonstrated that the 5′-DMT
H-phosphonates could be utilized as coupling partners in
palladium(0)-catalyzed P-C couplings, we were anxious to
see whether the resulting dimers could be used as building
blocks in oligonucleic acid synthesis as originally intended.
We therefore decided to examine the construction of the
suitably functionalized 5′-DMT-T*T-3′-CEP building block
19.8 In anticipation of encountering problems with removal
niques.11 As an illustration, a 10 µmol synthesis of the
TpT*TpT tetramer 21 was performed (Figure 1). To ensure
good incorporation of the dimer 19, the coupling time for
this step was extended from 60 to 180 s, but the rest of the
automated synthesis remained as standard. Cleavage from
the solid support and deprotection of the cyanoethyl protect-
ing groups was achieved by treatment with NH4OH,12 and
purification of the crude material by reverse phase HPLC
provided the desired tetramer 21 as the major product.13
At the outset of this project we were concerned that the
vinylphosphonate may not withstand the NH4OH deprotec-
tion conditions, as both Michael additions and cleavage of
the 3′-O-P phosphate bond are possible. The identity of
21 was firmly established by NMR spectroscopy, with the
1
1D H NMR spectrum of 21 clearly showing all four H1′
protons and both of the olefinic (H5′ (6.56 ppm) and H6′
(6.11 ppm)) protons associated with the vinylphosphonate
internucleotide linkage. The proton decoupled 31P NMR
spectrum of 21 shows three distinct peaks, two characteristic
of phosphodiesters (δ 0.01 and 0.13 ppm) and one of the
vinylphosphonate (δ 13.64 ppm).14
In summary, we have shown that commercially available
phosphoramidites can serve as precursors to a range of
vinylphosphonate-containing dimers. We have shown that
both acid- and base-sensitive functional groups are tolerated
in the palladium-catalyzed coupling reaction when propylene
oxide is used as the HBr scavenger. Furthermore, we have
shown that the T*T dimer 19 synthesized using this
methodology can be used successfully in the automated
synthesis of oligonucleic acids. Work is currently in progress
to extend this work to the incorporation of modified purine-
containing dimers and the synthesis of longer oligonucleic
acid sequences of biological importance.
Figure 1.
Acknowledgment. The Authors thank the Nuffield
Foundation (NUF-NAL), AstraZeneca, Pfizer Central Re-
search, and the School of Chemistry, University of Notting-
ham, for financial support and the BBSRC (R.D.B.) and
University of Nottingham (S.A.) for the provision of Ph.D.
studentships. The authors also thank Mr. John Keyte for help
with the solid-phase oligonucleotide synthesis.
of the 3′-TBDPS protecting group from the acid- and base-
sensitive dimer 14, we decided to deprotect the vinyl bromide
2 and couple the resulting free alcohol 18 with the H-
phosphonate 10 (Scheme 3). Thus, treatment of 2 with TBAF
in THF cleanly gave the desired product 18 in excellent yield.
Pleasingly, coupling of 18 and 10, using our standard
conditions (vide supra), provided the desired 3′-OH dimer
in 59% isolated yield. Tetrazole-mediated reaction of this
material with 2-cyanoethyl-tetraisopropylphosphoramidite
209 cleanly afforded the desired phosphoramidite building
block 19 as a 1:1:1:1 mixture of diastereoisomers.10 With
access to this material we were now in a position to examine
the incorporation of the modified T*T dimer into an
oligonucleic acid using standard solid-phase synthesis tech-
Supporting Information Available: Experimental pro-
cedures and characterization data for all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL0166045
(11) Caruthers, M. H. Science 1985, 230, 281.
(12) Caruthers and Zhao have also reported the solid-phase synthesis of
vinylphosphonate linked oligonucleic acids (see ref 2c). In their case,
however, they used an o-chlorophenyl group for protection of the phos-
phonate internucleotide linkage and an extra synthetic step was needed for
its deprotection.
(6) Purification by SiO2 gel column chromatography lead to substantial
decomposition of the H-phosphonates and provided only poor yields of the
desired coupling partners.
(7) The asterisk (*) signifies the position of the vinylphosphonate
internucleotide linkage.
(8) McBride, L. J.; Caruthers, M. H. Tetrahedron Lett. 1983, 24, 245.
(9) Bannwarth, W.; Trzeciak, A. HelV. Chim. Acta 1987, 70, 175.
(10) As judged by proton-decoupled 31P NMR.
(13) The only other product to be isolated from the automated synthesis
was thymidine, which results from the incomplete coupling of the
phosphoramidite 20 to the 5′-OH of the initial CPG-supported thymidine
(corresponding to T4 in 22 (Figure 2)).
(14) For copies of relevant spectra see Supporting Information.
Org. Lett., Vol. 3, No. 21, 2001
3367