Universal solid supports for the synthesis of oligonucleotides via a transesterification of H-phosphonate diester linkage

Article abstract of DOI:10.1021/jo051172n

Three universal solid supports exhibiting an hydroxyl function were prepared. The introduction of a first H-phosphonate diester linkage which was kept throughout the elongation allowed the release of 3′-hydroxyl oligonucleotides by a transesterification m

Full text of DOI:10.1021/jo051172n

Universal Solid Supports for the Synthesis of Oligonucleotides via
a Transesterification of H-phosphonate Diester Linkage
Fernando Ferreira, Albert Meyer, Jean-Jacques Vasseur, and Franc¸ois Morvan*
Laboratoire de Chimie Organique Biomole´culaire de Synthe`se, UMR 5625 CNRS-UM II,
ERT “Oligonucleotides: Methodologie Valorisation”, Universite´ de Montpellier II, CC008,
Place E. Bataillon, 34095 Montpellier Cedex 5, France
morvan@univ-montp2.fr
Received June 10, 2005
Three universal solid supports exhibiting an hydroxyl function were prepared. The introduction of
a first H-phosphonate diester linkage which was kept throughout the elongation allowed the release
of 3′-hydroxyl oligonucleotides by a transesterification mechanism. The transesterification was
performed in a few minutes with either amino alcohols or K2CO3/methanol. Starting from a hydroxyl
solid support, tandem oligonucleotides were synthesized and the solid support was easily recyclable.
This strategy was extended to the release of an oligonucleotide from the solid support by a nonbasic
treatment opening the way to the synthesis of base-sensitive oligonucleotides thanks to the selective
deprotection of a hydroxyl in â of the H-phosphonate diester linkage.
Introduction
phonate diesters are also subject to transesterification
reactions by amino alcohols.6 In the case of oligonucle-
otides (oligos), we have employed this behavior to release
them from a solid support under different conditions:
from basic, mild basic, or neutral according to the design
of the solid support.
Herein, we present three universal solid supports, with
a hydroxyl function, and among them a reusable one,
used in combination with a H-phosphonate diester link-
age between the oligonucleotide and the solid support.
The H-phosphonate diester linkage was kept throughout
the elongation process. Different conditions were applied
to release the oligonucleotide in solution, such as etha-
nolamine, K2CO3/methanol, or Et3N‚3HF according to the
structure of the solid support.
H-Phosphonate diesters are unstable under basic
conditions.1 Furthermore, H-phosphonate diesters are
subject to intra-transesterification reactions by the vici-
nal hydroxyl of a ribonucleoside2 or cis-diol3,4 but also
propyl to pentyl diol.5 The consequence is the removal of
the phosphorus group by an intramolecular attack of the
free alcohol function on the phosphorus atom with the
formation of a cyclic H-phosphonate. Moreover, H-phos-
* Corresponding author. Tel: +33/467 144 961. Fax: +33/467 042
029.
(1) Westheimer, F. H.; Huang, S.; Covitz, F. J. Am. Chem. Soc. 1988,
110, 181-185.
(2) Huss, S.; Gosselin, G.; Stawinski, J.; Stromberg, R.; Imbach, J.
L. Nucleosides Nucleotides 1988, 7, 321-337.
(3) Reese, C. B.; Visintin, C. Tetrahedron Lett. 1999, 40, 6477-6480.
(4) Wang, W. M.; Song, Q. L.; Jones, R. A. Tetrahedron Lett. 1999,
40, 8971-8974.
(5) Wada, T.; Honda, F.; Sato, Y.; Sekine, M. Tetrahedron Lett. 1999,
40, 915-918.
Two solid supports 1a,b share a structure with a
protected diol and triol, and the last (1c) is a hexanol
(6) Sobkowski, M.; Stawinski, J.; Sobkowska, A.; Kraszewski, A. J.
Chem. Soc., Perkin Trans. 1 1994, 1803-1808.
10.1021/jo051172n CCC: $30.25 © 2005 American Chemical Society
Published on Web 10/15/2005
9198
J. Org. Chem. 2005, 70, 9198-9206

Transesterification of H-Phosphonate Diester Linkage
further cyanoethyl phosphite triesters was performed
with peroxides (tert-butyl hydroperoxide9 or butanone
peroxide10) or with the Beaucage reagent11 to end up with
phosphodiester or phosphorothioate oligonucleotides,
respectively. Indeed, since H-phosphonate diester link-
ages do not have an electron pair on the phosphorus
atom, they do not react with these reagents.12-15
A first universal solid support (1a) with a dimethoxy-
trityl-O-ethylene glycol-O,O′-hydroquinone diacetic linker
(EG-Q-linker) was synthesized (Scheme 2). Ethylene
glycol was monotritylated using dimethoxytrityl chloride
in anhydrous pyridine and then converted to the di-
methoxytrityl-O-ethylene glycol-O,O′-hydroquinone O′-
FIGURE 1. Structure of the solid supports, EG-Q 1a, TB-
DMS-glycerol 1b, and hexanol 1c.
7
hemiester using hydroquinone-O,O′-diacetic acid and
DEC/DMAP. This hemiester was loaded on a long-chain
alkylamino (LCAA) CPG using DEC/DMAP. After cap-
ping with acetic anhydride in the presence of N-methyl
imidazole, a loading of 49 µmol per gram was determined
by trityl assay.
SCHEME 1. Formation of a H-Phosphonate
Diester Linkage from a tert-Butyl Phosphite
Triester by an Arbuzov-Type Reaction during the
Detritylation Step
Applying the cycle described in Table 1 (see the
Supporting Information), without the capping step, a T6
exhibiting a 3′-H-phosphonate (HP) diester linkage (A)
was synthesized on this new solid support (Figure 2). For
comparison, a T6 (B) using a standard elongation with
cyanoethyl phosphoramidite was also performed on this
support (Figure 2), and then several basic treatments
were performed.
Treatment of supported oligonucleotide B, bearing only
CNE phosphotriester linkages, with ammonia at room
temperature for 2 h or at 55 °C for 2 h yielded a mix-
ture of T6-OH and T6pOCH2CH2OH in a 4:6 ratio
(characterized by MALDI-TOF MS and quantified by
HPLC) showing that the elimination of the 3′-phos-
phorus group by an intramolecular attack of the vicinal
hydroxyl function with the formation of a cyclic phos-
phate was not efficient owing to the flexibility of the ethyl
linker.16 Indeed, it has been shown that the removal of
3′-phosphate is more efficient when the alcohol is con-
strained.17,18
Supported oligonucleotide A, bearing a 3′-HP diester
linkage, was treated with concentrated ammonia or with
an ethanol/ammonia solution (3:1, v/v) for 2 h at room
temperature. HPLC profiles showed the formation of T6
with a 3′-HP monoester (tR 15.06 min) in 24% and 14%
yield, respectively, and the expected T6 (tR 15.28 min)
being produced in yields of 76% and 86% (Figure 3a,b).
Each compound was isolated and characterized by MALDI-
TOF MS. This result could be explained by an attack of
solid support (Figure 1). The first solid support 1a with
a Q-linker7 allows a fast release of oligonucleotides. The
second solid support 1b was designed to release oligo-
nucleotides by a nonbasic Et3N‚3HF treatment. The
reusable third solid support 1c was used to synthesize
several times the same oligonucleotide or even different
oligonucleotides. Furthermore, multiple oligonucleotides
were synthesized in a row on the same support.
Results and Discussion
The common principle of the three solid supports is to
introduce a H-phosphonate diester linkage between the
hydroxyl solid support and the oligonucleotide. To do so,
two paths are possible. On one hand, classic H-phospho-
nate chemistry with commercial H-phosphonate nucleo-
tide and acyl chloride activator could be used. This
implies combination of phosphoramidite and H-phospho-
nate chemistries and specific reagents on the DNA
synthesizer (i.e., acyl chloride, pyridine/acetonitrile wash).
On the other hand, we applied a “fully phosphoramidite
chemistry”. A tert-butyl phosphoramidite building block
was used to generate a H-phosphonate diester for the
first coupling on the solid suppport. In this case, the
coupling step was performed with benzylmercaptotetra-
zole (BMT) owing to a lower reactivity of tert-butyl
phosphoramidite building blocks in comparison of the
standard cyanoethyl ones. Then the oxidation step was
omitted and the H-phosphonate diester was formed
during the detritylation step by an Arbuzov-type reaction
(Scheme 1) as described before with benzyl phosphite
triesters.8 Standard cyanoethyl phosphoramidites were
then used for the elongation process of the oligonucleo-
tide.
(9) Hayakawa, Y.; Uchiyama, M.; Noyori, R. Tetrahedron Lett. 1986,
27, 4191-4194.
(10) Kataoka, M.; Hattori, A.; Okino, S.; Hyodo, M.; Asano, M.; R.,
K.; Hayakawa, Y. Org. Lett. 2001, 3, 815-818 and 2939.
(11) Iyer, R. P.; Phillips, L. R.; Egan, W.; Regan, J. B.; Beaucage, S.
L. J. Org. Chem. 1990, 55, 4693-4699.
(12) Stawinski, J.; Thelin, M. J. Org. Chem. 1991, 56, 5169-5175.
(13) Jung, P. M.; Histand, G.; Letsinger, R. L. Nucleosides Nucle-
otides 1994, 13, 1597-1605.
(14) Bologna, J. C.; Morvan, F.; Imbach, J. L. Eur. J. Org. Chem.
1999, 2353-2358.
(15) Stawinski, J.; Kraszewski, A. Acc. Chem. Res. 2002, 35, 952-
960.
To avoid oxidation of the H-phosphonate linkage dur-
ing the elongation process, the oxidation step of the
(16) Laurent, A.; de Lambert, B.; Charreyre, M. T.; Mandrand, B.;
Chaix, C. Tetrahedron Lett. 2004, 45, 8883-8887.
(17) Kumar, P.; Dhawan, G.; Chandra, R.; Gupta, K. C. Nucleic Acids
Res. 2002, 30, e130.
(7) Pon, R. T.; Yu, S. Y. Nucleic Acids Res. 1997, 25, 3629-3635.
(8) Meyer, A.; Morvan, F.; Vasseur, J.-J. Tetrahedron Lett. 2004,
45, 3745-3748.
(18) Guzaev, A. P.; Manoharan, M. J. Am. Chem. Soc. 2003, 125,
2380-1.
J. Org. Chem, Vol. 70, No. 23, 2005 9199

Ferreira et al.
SCHEME 2a
a Reagents: (i) DMTrCl, anhydrous pyridine; (ii) O,O′-hydroquinone diacetic acid, DMAP, DEC, anhydrous pyridine; (iii) activated
LCAA-CPG, DMAP, DEC, Et3N, anhydrous pyridine.
FIGURE 2. Schematic structure of T6 on the EG-Q-solid support 1a: (A) with a 3′-H-phosphonate diester linkage and (B) with
a standard 3′-cyanoethylphosphotriester linkage.
FIGURE 3. HPLC profiles of supported oligonucleotide A treated with (a) NH4OH, (b) NH4OH/MeOH 1:3, v/v, and (c) K2CO3/
MeOH.
SCHEME 3. Competitive Formation of T6 and
T6-3′-HP
It is noteworthy that treatment with 50 mM K2CO3 in
methanol hydrolyzes the Q linker within 2 min.7 Here,
when the supported oligonucleotide A was treated with
a solution of K2CO3 in methanol (10 mM or 5 mM) for 30
min, the T6 was obtained (Figure 3c) as characterized by
MALDI-TOF MS.
Then a pentamer d(TCAGA), bearing the four nucleo-
bases of DNA, was synthesized on this solid support
according to the elongation cycle described in Table 1
without the capping step and using the tert-butyl phos-
phoramidite of deoxyadenosine. A few milligrams of
support were treated with a 5 mM K2CO3/ methanol
solution, and the release of the oligonucleotide from the
solid support was monitored by UV at 260 nm. A plateau
was reached after 2 min indicating that the release from
the solid support is fast even with a low concentration of
K2CO3. MALDI-TOF MS analysis showed that cyanoethyl
groups were removed but the benzoyl and isobutyryl
nucleobase protections were partially removed, so an
ammonia treatment was applied to complete the depro-
tection (5 h, 55 °C). For the synthesis of these short
oligonucleotides no capping step was applied. Since acetic
anhydride could give rise to a side reaction with a
H-phosphonate diester linkage, we used a phosphor-
the base on the phosphorus atom and the release of
T6-3′-HP monoester or T6 (Scheme 3: paths a and b). This
does not preclude the nucleophilic attack of the base
on the Q linker with the formation of the T6-3′HP-
CH2CH2OH followed by an intramolecular attack of
the hydroxyl function on the phosphorus atom releasing
the cyclic H-phosphonate and the expected T6 (Scheme
3, path c). In such a case, reducing the basic conditions
will favor the hydrolysis of the Q-linker and will give
more T6.
9200 J. Org. Chem., Vol. 70, No. 23, 2005

Transesterification of H-Phosphonate Diester Linkage
FIGURE 4. HPLC profiles of d(TGACAGTGTCATGTACGA): (left) crude, (right) pure.
SCHEME 4a
a Reagents: (i) DMTrCl, pyridine; (ii) TBDMSCl, pyridine; (ii) succinic anhydride; (iv) LCAA-CPG, DMAP, EDC, pyridine.
SCHEME 5. Release of Protected Oligonucleotide
from Solid Support 1b and Deprotection
amidite as capping agent in the following syntheses. We
chose the dicyanoethyl N,N-diisopropyl phosphoramidite
since it will yield more hydrophilic shorter mers bearing
a 5′-phosphate monoester. Hence, shorter mers will be
first eluted on a reversed-phase column and could be
eliminated more easily.
An 18-mer d(TGACAGTGTCATGTACGA) was synthe-
sized by applying the elongation cycle reported in Table
1. After elongation, the solid support was treated with a
5 mM K2CO3 methanol solution for 30 min. MS analysis
showed the total removal of cyanoethyl groups and
partial removal of the benzoyl and isobutyryl groups.
Thus, the supernatant was treated with concentrated
ammonia for 5 h at 55 °C. After evaporation, the crude
(86 OD260nm) was analyzed and purified by HPLC (Figure
4), and characterized by MALDI-TOF MS (m/z calcd
5537.70, found 5537.46).
Finally, to obtain easily a fully deprotected oligonu-
cleotide, a decamer d(TCATGTACGT) was synthesized
using UltraMild phosphoramidites (i.e., N4-Ac-dC,
N6-pac-dA, and N2-isopro-pac-dG).19,20 After elongation,
a single treatment with 50 mM K2CO3/methanol for 20
min at 50 °C led to the fully deprotected oligonucleotide
(38 OD260nm) as confirmed by MALDI-TOF MS (negative
mode m/z calcd 3017.04, found 3016.69) with 83% spec-
troscopic purity at 260 nm (see the Supporting Informa-
tion).
This universal solid support 1a, with an EG-Q linker
was easily synthesized with a high loading. The 3′-OH
oligonucleotides were rapidly released from the solid
support by a treatment with 5 mM K2CO3 in methanol.
The second solid support 1b was synthesized from
glycerol.21 The primary alcohols were respectively pro-
tected with a dimethoxytrityl and a tert-butyldimethyl-
silyl group, and then the secondary alcohol reacted with
succinic anhydride to form the hemiester which was
finally loaded on LCAA-CPG using EDC DMAP (loading
59 µmol/g) (Scheme 4).
The 12-mer d(AGATTCCGTCAT) was synthesized
from the solid support 1b, according to the elongation
cycle described in Table 1 without capping step. The solid
supported 12-mer was then treated with a solution of
Et3N‚3HF for 2.5 h at room temperature. MALDI-TOF
MS analysis showed that the fully protected 3′-OH
oligonucleotide was obtained (positive mode m/z calcd
4969.99, found 4972.45).
The protecting groups (i.e., cyanoethyl, benzoyl, and
isobutyryl) were finally removed by an ammonia treat-
ment at 55 °C for 5 h (Scheme 5). The crude was analyzed
by HPLC (see the Supporting Information) and charac-
terized by MS (negative mode m/z calcd 3619.44, found
3618.86). MS analysis showed that the 12-mer was
slightly contaminated with oligonucleotide N-1 (m/z calcd
3307.24, found 3305.87).
When the same supported oligonucleotide was treated
with HF/pyridine instead of Et3N‚3HF, both 3′-OH and
3′-HP 12-mers were obtained in a 7:3 ratio. This solid
support with a protected â-hydroxyl in combination with
the introduction of a first H-phosphonate diester linkage
is particularly interesting because it opens the way to
the synthesis of base-sensitive oligonucleotides22,23 This
(19) Schulhof, J. C.; Molko, D.; Teoule, R. Nucleic Acids Res. 1987,
15, 397-416.
(20) Uznanski, B.; Grajkowski, A.; Wilk, A. Nucleic Acids Res. 1989,
17, 4863-4871.
(21) De Napoli, L.; Di Fabio, G.; Messere, A.; Montesarchio, D.;
Musumeci, D.; Piccialli, G. Tetrahedron 1999, 55, 9899-9914.
J. Org. Chem, Vol. 70, No. 23, 2005 9201

Ferreira et al.
SCHEME 6a
a Reagents: (i) succinic anhydride, DMAP, pyridine; (ii) EDC, N-hydroxysuccinimide, CH2Cl2, Et3N; (iii) 6-amino-1-hexanol, CH2Cl2.
A decamer d(TCATGTACGA) phosphodiester (PO) and
phosphorothioate (PS) were synthesized on the solid
support 1c starting with the deoxyadenosine tert-butyl
phosphoramidite according to the elongation cycle with
a capping step (Table 1) and using Beaucage’s reagent
as oxidizer for the latter oligonucleotide. The oligonucleo-
tides were released from the solid support by treatment
FIGURE 5. Schematic structure of a T6 on hexanol solid
support with a 3′-HP diester linkage.
with 10 mM K2CO3 methanol for 30 min, and after
solid support, using suitable protecting groups on nucleo-
evaporation the residues were treated with concentrated
bases,24,25 was used for the synthesis of prooligonucleo-
ammonia at 55 °C for 5 h to removed the isobutyryl and
tides,23 and their synthesis will be described elsewhere.
the remaining benzoyl groups. The oligonucleotides were
The third solid support 1c was designed to be recycled
analyzed by HPLC (Figure 6) and characterized by
by simple washings and dryings before the next synthe-
MALDI-TOF MS (m/z for PO oligonucleotide calcd 3026.06,
sis. It consists of an aminohexanol linker anchored on
found 3026.91; for PS oligonucleotide calcd 3170.64, found
the LCAA-CPG through a succinyl spacer. The resulting
3170.44).
amide function is stable under moderate basic conditions.
Thus, LCAA-CPG was succinylated, and the hemiester
was activated as an N-hydoxysuccinimide derivative that
finally reacted with 6-amino-1-hexanol to give the desired
No difference was found between the phosphotriester
and the phosphorothioate oligonucleotide synthesis in
term of efficacy to obtain the expected 3′-hydroxyl oligo-
nucleotide.
solid support 1c (Scheme 6). A loading of 51 µmol/g was
Recycling. Treatments with an amino alcohol or
K2CO3/methanol yielded to the transesterification of the
H-phosphonate diester with the formation of the expected
3′-OH oligonucleotide, the bis alkyl H-phosphonate di-
ester and the recovery of the starting hexanol solid
support 1c (Scheme 7). Hence, the solid support could
be reused for a new synthesis.
determined by trityl assay after a first coupling.
Starting from the solid support 1c, a T6 (C) was
synthesized according to the elongation cycle in Table 1
without the capping step (Figure 5).
Likewise, when oligonucleotide C was treated with
concentrated ammonia at room temperature or at 55 °C
for 2 h, the T6 and T6-3′-HP monoester were obtained in
82% and 18% yield, respectively. In contrast, when
oligonucleotide C was treated for 30 min with a 10%
solution of 2-aminoethanol or 3-amino-1-propanol in
pyridine the T6 was quantitatively obtained. The cleav-
age of the oligonucleotide from the solid support pro-
ceeds by a transesterification of the amino alcohols.6
Likewise, when a treatment with 5 mM K2CO3 methanol
was applied the T6 was quantitatively obtained. Since
this behavior was not reported in the literature, we
further investigated this cleavage with 5 mM K2CO3 in
methanol and in water. The release of T6 from the
support was monitored by UV at 260 nm. A plateau was
reached after 8 and 60 min with the methanol and water
solution, respectively. MALDI-TOF MS analysis showed
that all cyanoethyl groups were removed in the methanol
solution while a few of them remained in the aqueous
solution.
To evaluate the recovery of the solid support, a hex-
amer d(TGAGAT) was synthesized on the solid support
hexanol 1c according to the elongation cycle reported in
Table 1 with a capping step. The supported hexamer was
treated with 10% ethanolamine in pyridine for 30 min,
and the supernatant after evaporation was treated with
ammonia. For recycling, the solid support was washed
with pyridine (2 mL), acetonitrile/water (1:1, v/v, 2 mL),
and acetonitrile (10 mL) and dried under vacuum. The
same hexamer was synthesized on the same solid support
three more times (named 1, ×1, ×2, ×3). The amount of
each oligonucleotide was 38, 41, 32 and 38 OD260nm
,
respectively. MALDI-TOF MS visualized only the ex-
pected hexamer (m/z calcd 1830.28, found 1830.50,
1830.18, 1829.97, 1830.48). HPLC profiles of the crude
hexamer were similar for all of them and similar to that
of the same oligonucleotide synthesized according to a
standard protocol on a commercial succinyl solid support
loaded with thymidine (see the Supporting Information).
During the release, the ethanolamine could give some
transamination reaction with N4-benzoyl dC to yield the
N4-hydroxyethyl dC.26 To avoid this side reaction, the
(22) Zhou, Y. Z.; Tso, P. O. P. Nucleic Acids Res. 1996, 24, 2652-
2659.
(23) Morvan, F.; Vasseur, J.-J.; Vive`s, E.; Rayner, B.; Imbach, J.-L.
In Pharmaceutical aspects of oligonucleotides; Couvreur, P., Malvy, C.,
Eds.; Taylor & Francis: London, 2000; Vol. 3, p 79-97.
(24) Guerlavais Dagland, T.; Meyer, A.; Imbach, J. L.; Morvan, F.
Eur. J. Org. Chem. 2003, 2327-2335.
(25) Ferreira, F.; Vasseur, J. J.; Morvan, F. Tetrahedron Lett. 2004,
45, 6287-6290.
(26) Polushin, N. N.; Morocho, A. M.; Chen, B.; Cohen, J. S. Nucleic
Acids Res. 1994, 22, 639-645.
9202 J. Org. Chem., Vol. 70, No. 23, 2005

Transesterification of H-Phosphonate Diester Linkage
FIGURE 6. HPLC profiles of crude d(TCATGTACGA) phosphodiester (left) and phosphorothioate (right) released from solid
support with 10 mM K2CO3 methanol for 30 min and then deprotected with NH4OH, 5 h, 55 °C.
SCHEME 7. Synthesis of Oligonucleotides on a
Universal and Reusable Solid Support 1c Released
by Transesterification by Amino Alcohol or K2CO3/
Methanol
HPLC and MS analyses showed profiles similar to those
obtained for d(TGACGAT). For X ) C: 42 OD260 m 93%
purity, for X ) G: 30 OD260nm 87% purity and for X ) A:
37 OD260nm 90% purity (see the Supporting Information).
All tert-butyl phosphoramidites showed good coupling
and no difference of transesterification of the correspond-
ing H-phosphonate diester linkage.
To avoid two treatments for the release and the
deprotection of the oligonucleotide, we performed the
synthesis of d(TCATGTACGT) using UltraMild phos-
phoramidites. After elongation a single treatment with
50 mM K2CO3 methanol for 20 min at 50 °C led to the
fully deprotected oligonucleotide (40 OD260nm) as con-
firmed by MALDI-TOF MS (m/z calcd 3017.04, found
3017.35) with a 82% spectroscopic purity at 260 nm (see
the Supporting Information).
With the both ethanolamine or K2CO3 methanol treat-
ments the amount of oligonucleotide produced was stable
from one synthesis to another, showing no decrease of
available hydroxyl functions on the solid support. Among
the both treatments, the latter was more convenient since
standard phosphoramidite with a benzoyl group on de-
oxycytidine could be used and it avoids high-boiling
solvents.
Note that the recycling of the support was rapid and
straight since only washings and a drying of the solid
support were required. Two articles report universal and
reusable solid supports, but unfortunately, in one case
the coupling of the first nucleoside derivative required,
besides its preparation, the use of specific reagents (i.e.,
HBTU) to form an ester linkage with the hydroxyl solid
support,27 and in the second case, recycling proceeded
with a tedious protocol in four steps with some noncom-
mercial reagents.28
Multiple Synthesis. Since many applications use
more than one oligonucleotides at a time, synthesis of
defined mixtures of oligonucleotides is a useful way to
improve synthesis productivity. Synthesis of two oligo-
nucleotides in a row is particularly interesting in the case
of primers for PCR since it will reduce the manipulations.
Two versatile methods were reported to produce several
oligonucleotides in a row. The first method uses a
phosphoramidite derivative comprising two 1,4-anhydro-
erythritols linked by a succinyl arm (Figure 8, left,
TOPS). However a dephosphorylation step under harsh
standard benzoyl protecting group should by replaced by
an acetyl, isobutyryl, or tert-butylphenoxyacetyl group.
(For all of them, the corresponding phosphoramidites are
now commercially available.). Furthermore, removal of
ethanolamine (bp 170 °C) and pyridine (bp 113 °C) is not
convenient owing to their high boiling points. An alterna-
tive was to use the treatment with K2CO3/methanol to
do the transesterification of the H-phosphonate diester
linkage. The synthesis of a heptamer d(TGACGAT) was
performed on a solid support 1c according to the elonga-
tion cycle reported in Table 1 with a capping step. The
supported heptamer was treated with 10 mM K2CO3
methanol for 30 min, and the supernatant after evapora-
tion was treated with ammonia since MS analysis showed
that isobutyryl groups were only partially removed by
K2CO3 treatment. For recycling, the solid support was
washed with methanol/water (1:1, v/v, 2 mL), acetonitrile/
water (1:1, v/v, 2 mL), and acetonitrile (10 mL) and dried
under vacuum. The same oligonucleotide was synthesized
on the same solid support three more times (named 1′,
×1′, ×2′, ×3′). The amount of each oligonucleotide was
37, 38, 36, and 36 OD260nm, respectively. MS analysis
visualized only the expected heptamer (calcd 2119.46,
found 2120.00, 2120.07, 2118.16, 2121.10). HPLC profiles
were similar for all of them and similar to that of the
same heptamer synthesized according to a standard
protocol (Figure 7).
For comparison, the heptamers d(TGACGAX) with X
) C, G, or A were synthesized starting from the corre-
sponding tert-butyl phosphoramidites. After release by
treatment with 10 mM K2CO3 MeOH and then ammonia,
(27) Pon, R. T.; Yu, S. Y.; Guo, Z. Q.; Sanghvi, Y. S. Nucleic Acids
Res. 1999, 27, 1531-1538.
(28) Kumar, P.; Mahajan, S.; Gupta, K. C. J. Org. Chem. 2004, 69,
6482-6485.
J. Org. Chem, Vol. 70, No. 23, 2005 9203

Ferreira et al.
FIGURE 7. HPLC profiles of d(TGACGAT) synthesized four times on the same solid support and released by 10 mM K2CO3
MeOH and a standard synthesis: standard (offset: 2 min and 0.1 Abs for each) 1′, ×1′, ×2′, ×3′.
of deoxyadenosine to start each oligonucleotide and using
pivaloyl chloride as activator. After elongation, the sup-
ported tandem oligonucleotides were treated with a
solution of 10 mM K2CO3/methanol for 30 min and with
concentrated ammonia for 5 h at 55 °C. HPLC profiles
showed two main peaks corresponding to the 18-mer and
15-mer (Figure 9). Each peak was collected and charac-
terized by MALDI-TOF MS, and surprisingly, the longer
oligonucleotide was first eluted (18-mer tR 15.0 min; m/z
FIGURE 8. Reagents for tandem oligonucleotide synthesis.
calcd 5537.70, found 5537.09; 15-mer tR 16.7 min m/z
or prolonged deprotection conditions is required, and
calcd 4502.01, found 4503.08). However, for the synthesis
nonquantitative dephosphorylation leaves unwanted im-
using tert-butyl phosphoramidite, we also observed by
purities in the mixture.29 The second method uses a
HPLC a small peak at 17.6 min which was isolated and
3′-O-hydroquinone-O,O′-diacetic acid derivative of each
characterized as a 32-mer d(TAT TCC GTC ATC GC-TGA
nucleoside (Figure 8, right, 3′-Q-nucleoside) and requires
CAG TGT CAT GTA CGA) corresponding to an incom-
its coupling with specific reagents (i.e., HBTU/DMAP)
plete coupling of the second tert-butyl phosphoramidite.
usually not used on a DNA synthesizer to form the ester
Hence, the elongation of the 14 next cyanoethyl phos-
function with the former oligonucleotide on the solid
phoramidite derivatives occurred on the 18-mer. This
support.30
small peak was not visualized for the second synthesis
Since a H-phosphonate diester linkage could be easily
using H-phosphonate monoester since pivaloyl chloride
cleaved by transesterification using either ethanolamine
acts also as a capping reagent. Note that since the
or K2CO3 methanol, the synthesis of multiple oligonu-
extinction coefficient (epsilon) of the 32-mer is almost
cleotides could be easily performed on the solid support
twice as high as the two other oligonucleotides it was
1c by starting each synthesis of the oligonucleotide with
a tert-butyl phosphoramidite without oxidation or with
a commercial H-phosphonate monoester nucleoside using
pivaloyl chloride both will yield a H-phosphonate diester
detected by HPLC although its low amount. In a PCR,
this impurity will lead to some arrest of the polymeri-
zation of the strand complementary to the 15-mer;
however, since primer are used in large excess this should
linkage (Scheme 8). The limitation will be the size of the
not affect the final PCR products.
pores of the solid support (500 Å in our case) but for
This solid support 1c in combination with a H-phos-
multiple and long oligonucleotides a solid support with
phonate diester linkage displays is (1) easy to prepare,
2000 Å pore could be easily prepared. Moreover, the solid
(2) universal, and (3) reusable and (4) allows multiple
support is recyclable.
synthesis.
A 18-mer d(TGA CAG TGT CAT GTA CGA) and then
Finally, starting from a commercial solid support, the
a 15-mer d(TAT TCC GTC ATC GCA) were synthesized
introduction of a H-phosphonate linkage could be used
in a row on the solid support 1c each of them starting
with the coupling of a tert-butyl phosphoramidite of
deoxyadenosine according to the cycle in Table 1. For
to release tandem synthesized oligonucleotides one after
the other. Thus a first oligonucleotide d(TCATGTACGA)
was synthesized with a standard protocol on a com-
comparison, the same tandem was performed on a solid
mercial succinyl solid support then a H-phosphonate
support 1c using commercial H-phosphonate monoester
linkage was introduced according to our protocol and a
second oligonucleotide T8 was synthesized. A first treat-
ment with K2CO3/methanol at 0 °C for 20 min cleaved
the H-phosphonate linkage, the T8 was isolated from the
supernatant, and then a standard ammonia treatment
(29) Hardy, P. M.; Holland, D.; Scott, S.; Garman, A. J.; Newton, C.
R.; McLean, M. J. Nucleic Acids Res. 1994, 22, 2998-3004.
(30) Pon, R. T.; Yu, S.; Sanghvi, Y. S. J Org Chem 2002, 67, 856-
64.
9204 J. Org. Chem., Vol. 70, No. 23, 2005

Transesterification of H-Phosphonate Diester Linkage
FIGURE 9. HPLC profiles of crude tandem synthesis of 18- and 15-mers using (left) tert-butyl phosphoramidite or (right)
H-phosphonate monoester derivatives. Gradient: 8% to 24% CH3CN in 50 mM TEAAc over 30 min.
SCHEME 8. Principle of Multiple Synthesis of Oligonucleotide on Universal and Reusable
Solid Support 1c
of the CPG beads released the d(TCATGTACGA) (Sup-
porting Information).
nucleotides per synthesis. The introduction of the H-
phosphonate could either be done by a “fully phosphor-
amidite chemistry”8 using tert-butyl phosphoramidite
derivatives or with commercial H-phosphonate monoester
nucleosides using H-phosphonate chemistry.
This approach could also be applied to the RNA
synthesis since the transesterification proceeded under
mild basic conditions unlike the standard universal solid
supports that required harsh basic treatments. Thus, the
tandem synthesis of RNA for a RNAi purpose is possible
and works along this line is in progress.
Conclusion
All of the solid supports presented herein in combina-
tion with a first incorporation of a H-phosphonate diester
linkage are universal. Among them, the second one 1b
opens the way for the synthesis of base-sensitive oligo-
nucleotides and the third one 1c exhibits the advantage
to be recyclable by a straight and easy protocol.
For all of them, the major point to highlight is that a
H-phosphonate diester linkage allows a clean and rapid
transesterification yielding 3′ and 5′ hydroxyl oligonucle-
otides. This transesterification of the H-phosphonate
diester linkage allowed the synthesis of multiple oligo-
Experimental Section
Solid-Support EG-Q-CPG. 2-O-(4,4′-Dimethoxytrityl)-
1-hydroxyethane. To dry ethylene glycol (1.90 g; 30.6 mmol)
J. Org. Chem, Vol. 70, No. 23, 2005 9205

Ferreira et al.
dissolved in 20 mL of dry pyridine was added 4,4′-dimethoxy-
trityl chloride (2.10 g; 6.2 mmol) at room temperature. The
mixture was stirred for 0.5 h, diluted with CH2Cl2 (50 mL),
and washed with NaHCO3 (2 × 150 mL). The organic layers
was dried over Na2SO4 and evaporated to dryness. The residue
was purified by flash column chromatography (silica gel;
gradient 20-100% dichloromethane/2% Et3N/in cyclohexane).
The appropriate fractions were combined and concentrated to
dryness to afford 1.6 g (70%) of the product as a yellow oil.
TLC (cyclohexane/ether/Et3N 1:8:1, v/v/v) Rf: 0.4. 1H NMR
(CDCl3) δ ppm: 2.1 (bs, 1 H), 3.25 (m, 2 H), 3.7 (m, 2 H), 3.8
(s, 6 H), 6.8-6.9 (m, 4 H), 7.2-7.6 (m, 9 H).
sizer using a cycle involving phosphoramidite chemistry.
Detritylation was performed with 3% DCA in CH2Cl2 for 60 s.
Coupling step: BMT (0.3 M in dry acetonitrile) was used as
activator; tert-butyl phosphoramidites (0.09M in CH3CN) were
introduced with a 45 s coupling time and the oxidation step
was omitted; commercially available phosphoramidites (0.09
M in CH3CN) were introduced with a 15 s coupling time. The
capping step was performed with biscyanoethyl diisopropyl
phosphoramidite (0.03 M in dry acetonitrile) + BMT for 15 s.
Oxidation were performed with 1.1 M tert-butyl hydroperoxide
in CH2Cl2 or 0.05 M 3H-1,2-benzodithiol-3-one 1,1-dioxide in
dry acetonitrile for 60 s.
DMTr-EG-HQDA. 2-O-(4,4′-Dimethoxytrityl)-1-hydroxy-
ethane (0.71 g; 2 mmol), hydroquinone-O,O′-diacetic acid
(HQDA) (0.529 g; 2.4 mmol), 1-(3-dimethylaminopropyl)-3-
ethylcarbodiimide (DEC) (0.373 g, 2 mmol), and 4-(dimethyl-
amino)pyridine (DMAP) (0.024 g, 0.1 mmol) were stirred in
anhydrous pyridine at room temperature for 5 h. Then, the
solvent was distilled off and the residue diluted with 100 mL
of CH2Cl2. The organic layer was washed with water (2 × 100
mL), and the solvent was evaporated to dryness. The residue
was purified by flash column chromatography (silica gel;
gradient 1-15% MeOH/1% Et3N/in dichloromethane). The
appropriate fractions were combined and concentrated to
dryness to afford 0.48 g (42%) of the product as an oil. TLC
Release from solid supports and deprotection:
With EG-Q-Solid Support 1a. The column containing the
supported oligonucleotide was treated back and forth with 2
mL of 5 mM K2CO3 methanol for 30 min using two syringes.
Then beads were washed with methanol (2 mL) and water (2
mL). After evaporation, the residue was treated with ammonia
for 5 h at 55 °C.
With TBDMS-Glycerol Solid Support 1b. The beads
were transferred to an Eppendorf tube and treated with 200
µL of commercial Et3N‚3HF in 200 µL of THF solution for 2 h
30 at room temperature, and then 50 mM TEAAc pH 7 buffer
was added (1 mL). After evaporation, the residue was treated
with concentrated ammonia for 5 h at 55 °C.
With Hexanol Solid Support 1c. Protocol A. The column
containing the supported oligonucleotide was treated back and
forth with 2 mL of 10% ethanolamine in pyridine for 30 min
using two syringes. Then beads were washed with pyridine (2
mL), pyridine/water (2 mL, 1:1, v/v), water (2 mL). After
evaporation, the residue was treated with ammonia for 5 h at
55 °C.
Protocol B. The same treatment used for support 1a using
K2CO3 methanol was applied.
For Ultramild Protecting Groups. The supported oligonu-
cleotide was treated with 50 mM K2CO3/MeOH for 20 min at
50 °C.
Recycling of Solid Support 1c. After Treatment Ac-
cording to Protocol A. The solid support was washed with
pyridine (2 mL), acetonitrile/water (1:1, v/v, 2 mL), and dry
acetonitrile (10 mL) and dried under vacuum for 5 h under
P2O5.
After Treatment According to Protocol B. The solid
support was washed with methanol/water (1:1, v/v, 2 mL),
acetonitrile/water (1:1, v/v, 2 mL), and dry acetonitrile (10 mL)
and dried under vacuum for 5 h under P2O5.
1
(5% CH3OH/CH2Cl2) Rf: 0.3. H NMR (CDCl3) δ ppm: 3.23-
3.27 (m, 2 H), 3.75 (s, 6 H), 4.29-4.32 (m, 2 H), 4.42 (s, 2 H),
4.55 (s, 2 H), 6.74-6.79 (m, 8 H), 7.15-7.4 (m, 9 H).
Preparation of DMTr-EG-Q-CPG (LCAA-CPG 500 A,
80-120 mesh, 80-90 µmol/g). LCAA-CPG (1.00 g), EG-
HQDA (0.114 g, 0.2 mmol), EDC (0.191 g, 1 mmol), DMAP
(0.012 g, 0.1 mmol), and Et3N (0.1 mL) were shaken in
anhydrous pyridine (5 mL) at room temperature for 5 h. The
solid support was filtered off, washed with MeOH and
CH2Cl2, and dried. A capping step with standard Cap A and
Cap B solutions was applied for 1 h, and the solid support was
filtered off, washed with MeOH and CH2Cl2, and dried. The
trityl assay indicated a loading of 48 µmol/g.
Solid Support from Glycerol. Synthesized according De
Napoli et al.21 Loading 59 µmol/g.
Hexanol Solid Support. LCAA-CPG (1.00 g), succinic
anhydride (0.2 g, 2 mmol), and N,N-(dimethylamino)pyridine
(DMAP) (0.04 g, 0.33 mmol) were added in a flask with
anhydrous pyridine (6 mL) and shaken overnight at room
temperature. The succinylated support was filtered off, washed
with pyridine and CH2Cl2, and dried under vacuum. The
succinylated CPG, 3-ethyl 1-[3-(dimethylamino)propyl]carbo-
diimine (EDC) (0.191 g, 1 mmol), N-hydroxysuccinimide (0.06
g, 0.5 mmol), and Et3N (40 µL) were added in CH2Cl2 (5 mL),
and the flask was shaken at room temperature for 4 h. Then,
6-aminohexan-1-ol (6 mg, 0.05 mmol) was added, and shaking
was continued overnight. Finally, piperidine (5 mL) was added,
and after 15 min, hexanol-CPG was filtered, washed with
MeOH and CH2Cl2, and dried under vacuum. The trityl assay
determined after the coupling of one standard phosphoramidite
indicated a loading of 51 µmol/g.
Acknowledgment. F.F. thanks the Ministe`re de la
Recherche et de la Technologie for the award of a
research studentship.
Supporting Information Available: Preparation of tert-
butyl phosphoramidites. Elongation cycle. HPLC profiles of
oligonucleotides synthesized on solid support 1a-c. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Oligonucleotide Synthesis. Oligonucleotides (1 µmol
scale) were synthesized on an ABI 381A or 394 DNA synthe-
JO051172N
9206 J. Org. Chem., Vol. 70, No. 23, 2005