A base-labile group for 2′-OH protection of ribonucleosides: A major challenge for RNA synthesis

Article abstract of DOI:10.1002/chem.200801392

A base-labile group for 2'-OH protection of ribonucleosides was investigated. The solid support was dried by blowing argon through a DNA synthesizer and was first treated with 10% anhydrous piperidine in CH ₃CN at room temperature for 15 minutes to eliminate cyanoethyl groups from phosphates. The piperidine solution was removed from the column and the solid support was washed with CH₃CN. The three ammoniacal eluates were collected in a screw-capped glass vial and were left at room temperature for a further 1.5 hours to completely deprotect nucleobases and 2'-hydroxyl groups. The fully deprotected oligonucleotide was transferred to a 50 mL round-bottomed flask and isopropylamine was added to the solution before evaporation to dryness. It was observed that PivOM method provides highly pure RNA without any additional desalting step.

Full text of DOI:10.1002/chem.200801392

COMMUNICATION
DOI: 10.1002/chem.200801392
A Base-Labile Group for 2’-OHProtection of Ribonucleosides: A Major
Challenge for RNA Synthesis
Thomas Lavergne,[a] Jean-RØmi Bertrand,[b] Jean-Jacques Vasseur,[a] and
FranÅoise Debart*[a]
A great interest in the chemical synthesis of RNA has
grown up since the advent of RNA interference (RNAi)[1]
with the crucial need of a large number of short RNA mole-
cules for biological research and therapeutic applications.[2]
Compared to DNA synthesis, RNA production is more com-
plex. In well-established DNA synthesis, all the nucleophilic
functions are protected with base-labile protecting groups
that are removed at the end of the elongation process with a
base treatment. Beside lower coupling yields in chain assem-
bly compared to DNA synthesis, the main difficulty of RNA
chemistry results from the instability of RNA in basic
media. It is generally admitted that the 2’-OH protection
must not be base-labile to avoid the 2’-OH nucleophilic
attack on the phosphorus atom of the internucleoside link-
ages resulting in 3’–5’ to 2’–5’ isomerisation or 3’–5’ cleavage
of the linkages under deprotection conditions.[3] With regard
to the use of acyl (acetyl or benzoyl) protection for 2’-OH
reported in an early work, this resulted in very poor yields.[4]
The tert-butyldimethylsilyl (TBDMS) group is certainly the
most commonly utilized group for 2’-OH protection.[5] Sev-
eral protecting groups[6,7] have been proposed in place of
TBDMS, such as triisopropylsilyloxymethyl (TOM),[8] bis(2-
acetoxyethyloxy)methyl (ACE),[9] tert-butyldithiomethyl
(DTM),[10] 1-(2-cyanoethoxy)ethyl (CEE),[11] 2-cyanoethoxy-
(TEM),[14] and 2-cyanoethyl[15]; most of them, like TBDMS,
are removed by fluoride ions. This deprotection is a major
hurdle because it requires a desalting step by precipitation
or cartridge purification leading to additional time-consum-
ing workup procedures to obtain pure oligoribonucleotides.
In our search for a radically improved synthetic method
to obtain RNA efficiently, rapidly, and in high purity, we
now report a new RNA synthesis strategy based on protect-
ing the 2’-OH with a base-labile pivaloyloxymethyl (PivOM)
group compatible with standard protection for 5’-OH
(DMTr), phosphates (2-cyanoethyl), and nucleobases (acyl
groups). The main advantage of this RNA strategy with an
acetal ester group is a straightforward two-step all-base de-
protection in a short period of time (3 h total) at room tem-
perature. It consists of 1) the selective removal of the phos-
phate protecting group by b-elimination induced by a non-
nucleophilic strong organic base (1,8-diazabicyclo-
A
neous liberation of the nucleobases, the 2’-OH, and the rup-
ture of the succinyl or the Q-linker by an ammonia treat-
ment without migration or chain rupture. This approach,
which makes exclusive use of base-labile protecting groups,
is challenging because of the well-known RNA instability in
basic media.[3] To illustrate it, we report on the use of 2’-O-
PivOM phosphoramidites to synthesize RNA oligomers up
to 21 nucleotides in length and we demonstrate that they
were obtained in high yield and high purity without chain
rupture or migration.
First, our investigations began with the synthesis of the
four 3’-phosphoramidite ribonucleosides (Scheme 1). Previ-
ously we described the site-specific introduction of the
PivOM protecting group in which the Markiewicz reagent
(TIPSiCl2)[16] simultaneously blocks 5’-OH and 3’-OH and
leaves the 2’-OH free to accept the PivOM. This method
avoids separation of different isomers, but it requires six
steps for the whole uridine phosphoramidite synthesis.[17] Al-
though all these steps proceed in high yields, they are time
consuming and, moreover, TIPSiCl2 is an expensive reagent.
Alternatively we developed a convenient method with only
methyl
(CEM),[12,13]
2-(4-toluylsulfonyl)ethoxymethyl
[a] T. Lavergne, Dr. J.-J. Vasseur, Dr. F. Debart
Equipe OligonuclØotides ModifiØs
Institut des BiomolØcules Max Mousseron (IBMM)
UMR 5247 CNRS—UniversitØ Montpellier 1—
UniversitØ Montpellier 2
Place Eug›ne Bataillon, 34095 Montpellier Cedex 05 (France)
Fax : (+33)467-042-029
E-mail: debart@univ-montp2.fr
[b] Dr. J.-R. Bertrand
UMR 8121 CNRS—UniversitØ Paris-Sud
Institut Gustave Roussy, 39, Rue Camille Desmoulins
94805 Villejuif Cedex (France)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801392.
Chem. Eur. J. 2008, 14, 9135 – 9138
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9135

to 5’-O-DMTr-dT or 5’-O-DMTr-dC through a 3’-O-succinyl
linker or Q-linker. We first synthesized U12dC (ON-1; ON=
oligoribonucleotide) as a model RNA oligonucleotide to
test the efficacy of PivOM as 2’-OH protection. Syntheses
were performed on a 1 mmol scale with 180 s coupling time
and 5-benzylmercaptotetrazole (BMT) as activator. The
average stepwise yield was 99.7%, which is comparable with
DNA or other efficient RNA amidites. Upon completion of
chain assembly, ON-1 anchored to CPG solid support was
first treated with DBU in dry THF to remove phosphate
protecting groups (2-cyanoethyl) over a period of 45 min,
after having proved that such conditions are completely
inert to PivOM groups in monomers. Then, simultaneous
cleavage from the solid support and 2’-OH deprotection
with concentrated ammonia at room temperature for 3 h led
to a very pure U12dC without any cleavage products. In the
first step of this process, DBU was used to deprotect the in-
ternucleoside linkages, because phosphodiesters are less
prone to nucleophilic attack than phosphotriesters. On the
other hand, ON-2 was treated with piperidine, instead of
DBU, in acetonitrile for 15 min to eliminate cyanoethyl
groups with similar efficacy. After ammonia treatment, RP-
HPLC profile of the crude U19TT (Figure 1, top) reveals ef-
ficient elongation and deprotection without chain rupture.
The proposed mechanism for base-mediated hydrolysis of
the acetal ester PivOM group consists of cleavage of the
ester function by ammonia with formation of a formalde-
hyde hemiacetal. This intermediate is stable enough to
ensure protection of oligoribonucleotides in aqueous ammo-
nia and upon evaporation, when pH decreases, the hemiace-
tal undergoes fragmentation to the 2’-OH ribonucleoside
and formaldehyde (see scheme in the Supporting Informa-
tion).
Scheme 1. Synthesis of 2’-O-PivOM protected phosphoramidites 5a–d.
Reagents and conditions: 1) DMTrCl, pyridine, RT; 2) Bu2SnO, Bu4NBr,
1a–d, 2,2-dichloroethane, ClCH2OCOC(CH3)3, microwave, 758C, 2.5 h
G
(4a 36%, 4b 38%, 4c 34%, 4d 49% from 2a–d); 3)
(iPr2N)PCl(OCH2CH2CN), iPr2NEt, CH2Cl2, RT, 3 h (5a–d 75-82% from
4a–d). DMTrCl : 4,4’-dimethoxytrityl chloride, Bp: base-protected, Ac:
acetyl, Pac: phenoxyacetyl, tBuPac: tert-butylphenoxyacetyl.
four steps affording amidites 5a–d with the same overall
yield. The first step consists in the protection of the exocy-
clic amines of nucleobases with fast labile groups, such as
acetyl (Ac) for cytosine,[18] phenoxyacetyl (Pac) for ade-
nine,[19] and tert-butyl phenoxyacetyl (tBuPac) for guanine.[20]
Starting from uridine 1a or base-protected nucleosides 1b–
d, the 5’-OH group was blocked with DMTr. Then the 2’-
OH group was derivatized with PivOM via a 2’,3’-O-dibutyl-
stannylidene intermediate, which was treated with the cheap
commercial alkylating agent pivaloyloxymethyl chloride to
This promising data prompted us to synthesize hetero-
A
N
give a mixture of the 3’-O-PivOM
A
the capping step in which phenoxyacetic anhydride was used
instead of acetic anhydride to prevent replacement of
tBuPac group by Ac group of guanine residues.[19]
ON-3 was initially deprotected similarly to U19TT. How-
ever, formation of adducts was evidenced by MALDI-TOF
MS analysis (peaks at +41). These by-products resulted
from a side reaction of guanines as nucleophiles (at N1 and
NH2 in position 2, see Supporting Information) with formal-
dehyde generated by liberation of PivOM in the presence of
ammonia.[21] We noted that guanosine adducts were unstable
ACHTREUNG
microwave irradiation at 758C. The desired 2’-O-PivOM
compounds 4a–d were obtained in 34–49% yield from 2a–d
after silica gel chromatography. For all ribonucleosides, the
fast eluting 2’-isomers were isolated with higher yield than
the undesired 3’-isomers, and for guanosine the 2’- to 3’-
isomer ratio was the highest, which explains a better yield
(49%). The tritylated 2’-O-PivOM compounds 4a–d were
converted to the corresponding amidites 5a–d with 75–82%
yield by using 2-cyanoethyl N,N-diisopropylchloro-phos-
phoramidite. All amidites were
and disappeared after several hours in water or in triethyl-
A
lyophilized and were complete-
Table 1. Data for synthesized oligoribonucleotides.
ly stable during long-term stor-
age at À208C.
5’-sequence-3’
CT[a]
OY[b]
AY[c]
99.7
Crude material[d]
n.d.[e]
ON-1
ON-2
ON-3
ON-4
ON-5
U12dC
U19TT
180
180
180
180
180
96.5
94.299.7
91.1
82.1
83.8
With phosphoramidite mono-
mers 5a–d in hand, various oli-
140
CCC GUA GCU GTT
UGG AUC CUC GAU GGU AAC GdCT
CGU UAC CAU CGA GGA UCC AdAT
99.1
99.0
99.1
86
130
125
goribonucleotides
(Table 1)
were prepared on an automated
DNA synthesizer by using com-
mercially available controlled-
[a] CT=coupling time [s] in automated synthesis cycle. [b] OY=overall coupling yield [%]. [c] AY=average
stepwise coupling yield [%]. [d] Overall crude material O.D units measured at 260 nm UV absorption.
pore glass (LCAA-CPG) linked [e] n.d.=not determined.
9136
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9135 – 9138

RNA Chemistry with Base-Labile Protection
COMMUNICATION
Figure 2. Analysis after RT-PCR quantification of Ret/PTC1 expression
inhibition by cytofectine transfected siRNA in vitro. siRNA AS: duplex
purchased, same sequences than ON-4/ON-5; siRNA Ct: control siRNA
purchased.
methodology that disrupts the dogma “RNA synthesis is in-
compatible with a whole base labile strategy”.[3] The key
feature is the 2’-O-pivaloyloxymethyl protecting group of
the nucleotide building blocks that guarantees very high
average coupling yields >99% by using benzylmercaptote-
trazole as activator. As for TOM phosphoramidites, these
excellent yields are possibly the consequence of the low hin-
drance of the PivOM group, such that the base-labile pivalo-
yl ester is bound to the nucleoside through a small formal-
dehyde spacer. Furthermore PivOM is easy to introduce,
stable during each synthetic step, and easily removed under
basic conditions. In our strategy, RNA was protected with
exclusively base-labile groups, which were completely re-
moved in less than 3 h at room temperature in a straightfor-
ward all-base two-step procedure, first with piperidine in
acetonitrile followed by aqueous ammonia without concomi-
tant degradation of the RNA. A further strength of the
PivOM method is that it readily provides highly pure RNA
without any additional desalting step. The PivOM strategy is
a powerful and efficient method for siRNA synthesis.
Figure 1. RP-HPLC analysis of crude U19TT (top) and crude ON-4
(bottom).
ammonium acetate buffer. However, to avoid the formation
of adducts that occurred during ammonia concentration
under reduced pressure, isopropylamine was added just
before evaporation, affording the desired ON-3. No chain
cleavage was detected.
Then, the fully deprotected 21-mers ON-4 and ON-5 were
obtained following the optimized protocol: first piperidine
in acetonitrile for 15 min, secondly concentrated ammonia
at room temperature for 3 h and finally addition of isopro-
pylamine just before evaporation. The HPLC profile of un-
purified ON-4 (Figure 1, bottom) illustrates the efficiency of
the chain assembly and of the deprotection. For comparison,
ON-4 was prepared according standard conditions with
TBDMS chemistry. In our hands, the average stepwise cou-
pling yield was 96.8% providing ON-4 with a lower purity
than that obtained with our new methodology.
The digestion of the purified ON-4 and ON-5 by nuclease
P1 and alkaline phosphatase gave the four natural ribonu-
cleosides indicating the absence of any nucleobase modifica-
tion or unnatural internucleotide linkages. Hybridization of
21-mer ON-4 with its complementary ON-5 resulted in
duplex formation with a single cooperative transition (Tm =
798C).
The activity of siRNA duplex 4/5 was evaluated in an
RNAi assay that targets the Ret/PTC1 junction oncogene in-
volved in papillary thyroid carcinoma (Figure 2). The
siRNA duplex 4/5 obtained with the PivOM method had a
similar gene silencing activity (60% inhibition) to a pur-
chased siRNA duplex (siRNA AS) with the same sequence
(40% inhibition). This result confirms the integrity and
purity of the synthesized RNA by the PivOM method.
In conclusion, in the need for more robust RNA routine-
synthesis strategies to prepare siRNA, we developed a novel
Experimental Section
After chain assembly of oligoribonucleotides on an ABI model 381 A
DNA synthesizer, the solid support still in the column was dried by blow-
ing argon through it and was first treated with 10% anhydrous piperidine
in CH3CN at room temperature for 15 min to eliminate cyanoethyl
groups from phosphates. Then the piperidine solution was removed from
the column and the solid support was washed with CH3CN (3”2mL).
Secondly, a 28% aqueous ammonia solution was applied to the column
in three batches (1.5 mL, 1 mL, 0.5 mL) for 30 min each. The three am-
moniacal eluates were collected in a screw-capped glass vial and were
left at room temperature for a further 1.5 h to completely deprotect nu-
cleobases and 2’-hydroxyl groups. The fully deprotected oligonucleotide
was transferred to a 50 mL round-bottomed flask and isopropylamine
(15% of total volume: 0.45 mL) was added to the solution before evapo-
ration to dryness.
Chem. Eur. J. 2008, 14, 9135 – 9138
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9137

F. Debart et al.
[9] S. A. Scaringe, F. E. Wincott, M. H. Caruthers, J. Am. Chem. Soc.
1998, 120, 11820–11821.
Acknowledgements
[10] A. Semenyuk, A. Fçldesi, T. Johansson, C. Estmer-Nilsson, P.
Blomgren, M. Brännvall, L. A. Kirsebom, M. Kwiatkowski, J. Am.
Chem. Soc. 2006, 128, 12356–12357.
This work was supported by the Association pour la Recherche contre le
Cancer (ARC) and Minist›re de lꢁEducation Nationale de la Recherche
et Technologie.
[11] T. Umemoto, T. Wada, TetrahedronLett. 2004, 45, 9529–9531.
[12] T. Ohgi, Y. Masutomi, K. Ishiyama, H. Kitagawa, Y. Shiba, J. Yano,
Org. Lett. 2005, 7, 3477–3480.
[13] Y. Shiba, H. Masuda, N. Watanabe, T. Ego, K. Takagaki, K. Ishiya-
ma, T. Ohgi, J. Yano, Nucleic Acids Res. 2007, 35, 3287–3296.
[14] C. Zhou, D. Honcharenko, J. Chattopadhyaya, Org. Biomol. Chem.
2007, 5, 333–343.
Keywords: microwave irradiation
protecting groups · RNA · solid-phase synthesis
·
oligonucleotides
·
[15] H. Saneyoshi, K. Ando, K. Seio, M. Sekine, Tetrahedron 2007, 63,
11195–11203.
[16] W. T. Markiewicz, J. Chem. Res. Synop. 1979, 24–25.
[17] N. Parey, C. Baraguey, J.-J. Vasseur, F. Debart, Org. Lett. 2006, 8,
3869–3872.
[1] A. Fire, S. Xu, M. K. Montgomery, S. A. Kostos, S. E. Driver, C.
Mello, Nature 1998, 391, 806–811.
[2] D. Bumcrot, M. Manoharan, V. Koteliansky, D. W. Y. Sah, Nat.
Chem. Biol. 2006, 2, 711–719.
[3] C. B. Reese, Org. Biomol. Chem. 2005, 3, 3851–3868.
[4] T. Kempe, F. Chow, W. I. Sundquist, T. J. Nardi, B. Paulson, S. M.
Peterson, Nucleic Acids Res. 1982, 10, 6695–6714.
[5] K. K. Ogilvie, M. J. Nemer, TetrahedronLett. 1985, 26, 4567–4570.
[6] S. L. Beaucage, Curr. Opin. Drug Discovery Dev. 2008, 11, 203–216.
[7] R. Micura, Angew. Chem. 2002, 114, 2369–2373; Angew. Chem. Int.
Ed. 2002, 41, 2265–2269.
[18] V. Bhat, B. G. Ugarkar, V. A. Sayeed, K. Grimm, N. Kosora, P. A.
Domenico, E. Stocker, Nucleosides Nucleotides 1989, 8, 179–183.
[19] C. Chaix, D. Molko, R. Teoule, TetrahedronLett. 1989, 30, 71–74.
[20] N. D. Sinha, P. Davis, N. Usman, J. Perez, R. Hodge, J. Kremsky, R.
Casale, Biochimie 1993, 75, 13–23.
[21] H. Takahashi, Y. Hashimoto, Bioorg. Med. Chem. Lett. 2001, 11,
729–731.
[8] S. Pitsch, P. A. Weiss, L. Jenny, A. Stutz, J.-C. Wu, Helv. Chim. Acta
2001, 84, 3773–3795.
Received: July 9, 2008
Published online: September 2, 2008
9138
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9135 – 9138