RNA synthesis via dimer and trimer phosphoramidite block coupling

Article abstract of DOI:10.1016/j.tetlet.2011.03.042

The solid-phase synthesis of oligoribonucleotides using dimer and trimer phosphoramidite blocks is described. This method significantly reduces the total number of steps required in the synthesis of a target RNA sequence, provides more material, and simpl

Full text of DOI:10.1016/j.tetlet.2011.03.042

Tetrahedron Letters 52 (2011) 2575–2578
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
RNA synthesis via dimer and trimer phosphoramidite block coupling
⇑
Matthew Hassler, Yi Qiao Wu, N. Mallikarjuna Reddy, Tak Hang Chan, Masad J. Damha
Department of Chemistry, McGill University, Montreal, QC, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
The solid-phase synthesis of oligoribonucleotides using dimer and trimer phosphoramidite blocks is
described. This method significantly reduces the total number of steps required in the synthesis of a tar-
get RNA sequence, provides more material, and simplifies separation of the product from shorter failure
sequences. The procedure is illustrated by the synthesis of UpU, ApA, and UpUpU phosphoramidite blocks
and their use in the rapid synthesis of oligoribonucleotides on a solid support. Dimer and trimer amidite
blocks will likely find use in the large scale solution (or solid)-phase synthesis of siRNA drugs.
Ó 2011 Elsevier Ltd. All rights reserved.
Received 17 January 2011
Revised 4 March 2011
Accepted 8 March 2011
Available online 15 March 2011
Keywords:
RNA amidite
Dimer amidite
Trimer amidite
Block coupling
1. Introduction
describing the synthesis RNA phosphoramidite block-mers for
use in solid-phase synthesis.9 Herein we describe a facile method
Current methods for RNA synthesis rely on stepwise addition of
monomeric phosphoramidite units on solid supports. Careful chro-
matographic purification of the final product is required after the
oligomer is detached from the solid support and deprotected be-
cause of the inevitable presence of nꢀ1 mers arising from incom-
plete coupling. An alternative approach is to assemble
oligonucleotide chains more rapidly through ‘block’ condensation
reactions, as exemplified by the early work of Khorana and co-
workers in the synthesis of a DNA gene fragment from phosphodi-
ester intermediates.1,2 This has not been applied to solid phase
RNA synthesis because of the difficulty of obtaining the block-
mer phosphoramidites and the perceived inefficiency of block con-
densation on solid-phase supports.
The generation of oligoribonucleotide blocks is made difficult
by the presence of the 20-hydroxyl group and the protection it re-
quires. Ogilvie and co-workers described the synthesis of 50-O-
MMTr-20-O-TBDMS-30-O-levulinyl ribonucleoside monomers and
their use in the assembly of a hexadecauridylic acid via the phos-
phodichloridite procedure in 51% yield.3 This and other4,5 reports
utilized earlier coupling methods, which generally provided far
too low yields to be considered effective synthetic strategy espe-
cially for large scale. Since then, the phosphoramidite method
has proven to be the coupling method of choice,6 offering higher
yields and faster reactions. Although block-wise solid-phase syn-
thesis of DNA using phosphoramidite block-mers has been demon-
strated,7,8 there has been only one conference proceeding
to construct dimer/trimer RNA phosphoramidites that can be read-
ily applied to current synthesis approaches.
Our method builds upon the contributions of Ogilvie and co-
workers,3 who introduced the levulinyl (Lev) ester group as a tran-
sient and orthogonal protection for the 30-position. Initially, the
synthesis of UpU amidite was attempted by coupling commercially
available 50-O-DMTr-20-O-TBDMS-uridine-30-O-cyanoethyl phos-
phoramidite with 20-O-TBDMS 30-O-levulinyl-uridine 1. While the
desired product could be obtained, we observed up to 15% decya-
noethylation during column chromatography caused by the basic
eluent system necessary to avoid activation of the amidite moiety.
This led us to switch to the less base labile methyl protecting group
(Scheme 1).10–13
In addition, it was found that the removal of the 30 levulinyl pro-
tecting group caused migration of the 20-TBDMS group to the 30-
position resulting in a mixture of regioisomers 4 and 5 (Scheme 1).
The isomerization was confirmed by preparing the 30-silyl regio-
isomer 5 independently and comparing its 1H and 31P NMR chem-
ical shifts to compounds in the isolated mixture. We also
established that these regioisomers could not be resolved by TLC
which may explain why previous reports3 did not detect TBDMS
isomerization under these conditions.
The synthetic standard 5 allowed the evaluation of various con-
ditions for the removal of the levulinyl group for the degree of
isomerization (31P NMR detection limit: 0.4 mol %). Variation of
pH, concentration, mole equivalence of hydrazine, workup condi-
tions, solvent and temperature all led to similar results: 5–10% for-
mation of the inseparable 30-isomer 5 as assessed by 31P NMR. This
result came as a surprise as the monomeric nucleoside 1 did not
isomerize under identical conditions (and even longer reaction
⇑
Corresponding author.
E-mail address: masad.damha@mcgill.ca (M.J. Damha).
0040-4039/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2011.03.042

2576
M. Hassler et al. / Tetrahedron Letters 52 (2011) 2575–2578
The method outlined in Scheme 2 proved to be successful only
when 20-O-triisopropylsilyl (TIPS) protection was used, as the pres-
ence of DCI in solution caused the terminal 20-O-TBDMS group to
isomerize during coupling and/or work up. Furthermore, the 20-
TIPS protecting group increases the steric bulk around the 30-hy-
droxyl group significantly so that the regiospecific coupling pro-
ceeded with greater efficiency. Also, unlike TBDMS, the TIPS
group is uniquely resistant to silyl isomerization during the
delevulination conditions. The increased stability of the TIPS pro-
tected dimer over the TBDMS dimer was confirmed by NMR. The
TBDMS dimer 3 (Fig. 1A) and TIPS dimer 11a (Fig. 1B) were sub-
jected to 5 equiv of hydrazine (0.045 M) and their conversion into
4 and 8a, respectively, monitored by 31P NMR over time.
Under these conditions, addition of hydrazine resulted in com-
plete cleavage of the levulinyl group from 3 and 11a within 2.5 and
10 min, respectively. Detectable isomerization of the TBDMS group
was observed within 40–80 min. This is in contrast to the TIPS di-
mer 11a, which was resistant to silyl isomerization for at least 68 h
under the same conditions (31P NMR detection limit: 0.4 mol %).
Neither TIPS nor TBDMS protected monomers underwent isomeri-
zation (up to 10 h) under these conditions. As shown in Figure 1,
the emergence of demethylated products became apparent after
40 min in the case of the 20-TBDMS protected dimer (ꢀ0.84 and
ꢀ1.1 ppm) and 2.3 h in the case of the 20-TIPS protected dimer (sin-
gle resonance at ꢀ1.00 ppm). The presence of a single peak at
around ꢀ1.00 ppm for the TIPS demethylated dimer also confirmed
that no isomerization occurred during the delevulination step.
These experiments confirmed that the use of TIPS protection is sig-
nificantly more likely to produce isomerically pure dimers within
the time required for delevulination. Since removal of the levulinyl
group of TIPS dimer is complete within 10 min, there is little
chance of either isomerization or demethylation during the con-
version of 11a–8a and neither was observed.
The dimer and trimer amidites 9a and 14 were synthesized
according to Schemes 3 and 4, respectively. Coupling of 2a with
20-O-TIPS 30-O-Lev nucleoside 10a in the presence of DCI followed
by in situ oxidation with 6 M tert-butyl hydroperoxide afforded di-
mer 11a in excellent yields (Scheme 3). After purification, dimer
11a (1 mmol/ml of ACN) was treated with a 0.5 M solution of
hydrazine hydrate in pyridine/acetic acid (3:2 v/v) cleaving the lev-
ulinyl group quantitatively within 10 min. The reaction was cooled
to 0 °C and the excess hydrazine was quenched by addition of 2,4-
Scheme 1. Dimer phosphoramidite synthesis (TBDMS protection).
times), as also reported by Ogilvie and co-workers.3 To directly
compare the extent of isomerization between monomer and dimer
units, the delevulination reaction was performed in an NMR tube
and the isomerization was monitored over time (see Supplemen-
tary data). Under dilute conditions, isomerisation of 50-O-DMTr
20-O-TBDMS uridine was detectable after 2.3 h whereas isomerisa-
tion of dimer 4 was observed after 80 min (Fig. 1A). The dimer is
more susceptible to isomerization than the monomer nucleoside
under identical conditions [see Table 2, Supplementary data].
An alternative approach to the synthesis of dimer nucleotides
was to circumvent the use of 30-protection completely and rely
on the greater steric bulk around the secondary hydroxyl to reduce
its reactivity, allowing selective coupling with a phosphoramidite
2a at the less unhindered 50-hydroxyl (Scheme 2). This was
achieved by pre-activating the phosphoramidite with 4,5-dicyano-
imidazole (DCI) and adding that solution dropwise to a stirred
solution of either nucleoside 6 or 7 in ACN at 0 °C, and then allow-
ing the mixture to warm up to room temperature. The solution was
then oxidized with tert-butyl hydroperoxide in situ affording the
desired dinucleoside phosphotriesters. The use of iodine/water
for the oxidation should be avoided as premature demethylation
via an Arbuzov-like reaction can occur under these conditions.14
Figure 1. 31P NMR analyses of (A) UpU TBDMS protected dimer (3), (B) UpU TIPS protected dimer (11a) under delevulination conditions: 5 equiv of hydrazine hydrate
(0.045 M) in pyridine/acetic acid (3:2, v/v) in CDCl3.

M. Hassler et al. / Tetrahedron Letters 52 (2011) 2575–2578
2577
diisopropylamino)methoxyphosphine reagent was required for
this step as the use of the corresponding phosphonamidic chloride
led to incomplete reaction and silyl migration.
The trimer amidite could be prepared by treatment of 11a with
3% TFA in DCM to deprotect the 50-hydroxyl (95%), followed by
coupling (2a) and in situ oxidation to afford 13 in good yield
(88%). Delevulination and phosphitylation afforded the desired tri-
mer amidite in 80% yield (over two steps) without any detectable
20/30 isomerization (Scheme 4; 31P NMR detection limit: 0.4 mol %).
Of the two routes used to prepare dimer amidite 9a (Schemes 2
and 3), use of levulinyl protection was preferred as the coupling
step is more straightforward to carry out and provides higher yield
of the desired product (75% vs 70%). The ApA dimer phosphorami-
dite 9b was synthesized in a similar manner in good yields
(Scheme 3), demonstrating the general applicability of our
procedures.
Scheme 2. Regioselective dimer phosphoramidite synthesis.
With dimer and trimer amidites in hand, their coupling effi-
ciency were assessed on a solid support through the synthesis of
50-(rU)18dT-30 using the following conditions: 1
lmol scale; dimer
and trimer (0.15 M) blocks in MeCN, 0.25 M DCI as an activator,
and 20 min coupling cycles (Table 1) (see Supplementary data).
Oxidation of the phosphite triester intermediates was achieved
using a 3 M solution of tert-butyl hydroperoxide in toluene.13,15
Following chain assembly, the fully deprotected oligomers were
obtained by: (1) treatment with 5-tert-butyl-2-methylthiophenol
(rt, 120 min) to deblock the phosphate methyl groups; (2) washing
the solid-support (ACN, 10 ml); (3) ammonolysis with NH4OH/
EtOH (3:1) (rt, 60 h) to release the oligomers from the support,
and (4) fluoride treatment (1 ml of 1 M TBAF in THF, rt, 48 h) to de-
block the TBDMS and TIPS protecting groups. The three oligomers
were identical as shown by HPLC and MALDI-TOF (Calcd 5753.3,
found: 5752.9 m/z).
Scheme 3. Dimer phosphoramidite synthesis (TIPS protection).
Yields of oligomers were calculated from the HPLC traces (Sup-
plementary data), from which coupling efficiencies of the amidites
were estimated at 98.7%, 98.3%, and 97.2% for monomer, dimer,
and trimer amidite, respectively (Table 1; entries II, IV, VI). The
overall yield of 50-(rU)18dT-30 prepared from trimer couplings
(6ꢁ at 97.2% efficiency) was 84.7% (Fig. 2), which was superior to
the 80% overall yield obtained by coupling rU monomers (18ꢁ at
98.7% efficiency) (Table 1; entry II vs VI). Reducing the trimer ami-
dite concentration to 0.10 M and the coupling time by 50% reduced
the coupling efficiency by 11% (Table 1; entry V). These conditions
afforded, as expected, a mixture of the full length product and a
series of well resolved nꢀ3 ‘failure’ sequences (Fig. 3). A second
poly-py/pu sequence, 50-(rAAUU)4dTdT-30, was assembled from di-
mers 9a and 9b in 88.8% overall yield (98.5% per coupling). This
contrasts the 72.5% overall yield obtained when this sequence
was assembled via monomer coupling (98.0% per coupling; Table 1,
entry VII vs VIII). Thus in this case, the block coupling strategy
yielded ca. 16% more product than the standard method.
Scheme 4. Trimer phosphoramidite synthesis.
pentanedione. Chromatography was performed, avoiding the use
of base, and the dimer was phosphitylated with bis(N,N-diisopro-
pylamino)methoxyphosphine and N,N-diisopropylammonium tet-
razolide, with sonication to speed up the reaction, yielding dimer
30-phosphoramidite 9a in isomerically pure form. The bis(N,N-
While we have not yet fully optimized deprotection conditions,
we have noted, in some instances, the presence of very small
Table 1
Synthesis of oligonucleotides via monomer and block coupling
Entry
Oligomer 50-to-30
Amidite
Concd (M)
# of couplings
Time (min)
Coupling efficiency (%)
Yielda (%)
I
II
(rU)18dT
(rU)18dT
(rU)18dT
(rU)18dT
rU (2a)
rU (2a)
rUU (9a)
rUU (9a)
rUUU (14a)
rUUU (14a)
rU (2a), rA (2b)
rUU (9a), rAA (9b)
0.10
0.15
0.10
0.15
0.10
0.15
0.15
0.15
18
18
9
9
6
6
16
8
10
20
10
20
10
20
20
20
98.5
98.7
97.2
98.3
86.5
97.2
98.0
98.5
76.5
80.1
77.8
85.9
41.8
84.7
72.5
88.8
III
IV
V
VI
VII
VIII
(rU)18dT
(rAAUU)4dTdT
(rAAUU)4dTdT
(rAAUU)4dTdT
a
Yield % of oligomer in crude material (HPLC). Solid supports: dT or dTT functionalized succinyl-LLAA-CPG. 50-(rAAUU)4dTdT-30 was deprotected by a modification of fast
deprotection;18 50-(rU)18dT-30 was deprotected as described in this manuscript.

2578
M. Hassler et al. / Tetrahedron Letters 52 (2011) 2575–2578
In conclusion, we have developed a viable route for the synthe-
sis of regioisomerically pure dimer and trimer RNA phosphorami-
dites that couple with similar efficiency as monomeric
phosphoramidite units. The method increases the overall yield of
the target oligoribonucleotide sequence by decreasing the number
of coupling steps required for chain assembly and has the potential
of significantly simplifying the final purification of RNA sequences.
For instance, dimer and trimer synthons can be utilized either in
solution or solid-phase in conjunction with monomer synthons
in the final stages of chain assembly, affording nꢀ2 or nꢀ3 failure
sequences that are more readily resolved. Protected trinucleotides
may also find applications in codon and anticodon construction of
combinatorial libraries of mRNA and tRNA mimics.
Figure 2. Polypyrimidine, 50-(rU18)-dT-30 made from trimer amidite 14 under
optimized conditions (0.15 M, 1200 s coupling time).
Acknowledgments
M.J.D. and T.-H.C. both acknowledge financial support from the
Natural Science and Engineering Research Council of Canada and
ScinoPharm Taiwan.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.tetlet.2011.03.042.
References and notes
1. Kossel, H.; Buchi, H.; Khorana, H. G. J. Am. Chem. Soc. 1967, 89, 2185.
2. Ohtsuka, E.; Khorana, H. G. J. Am. Chem. Soc. 1967, 89, 2195.
3. Ogilvie, K. K.; Nemer, M. J. Can. J. Chem. 1980, 58, 1389.
4. Ohtsuka, E.; Tanaka, S.; Ikehara, M. J. Am. Chem. Soc. 1978, 100, 8210.
5. Werstiuk, E. S.; Neilson, T. Can. J. Chem. 1976, 54, 2689.
6. Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859.
7. Ono, A.; Matsuda, A.; Zhao, J.; Santi, D. V. Nucleic Acids Res. 1995, 23, 4677.
8. Chattopadhyaya, J. B.; Reese, C. B. Nucleic Acids Res. 1980, 8, 2039.
9. Kim, S. W.; Kayushin, A.; Kim, M.; Miroshnikov, A.; Jung, K. E. Nucleic Acid Symp.
Ser. 2008, 52, 403.
Figure 3. Polypyrimidine sequence 50-(rU18)-dT-30 made from (A) monomer
amidite 2a (18ꢁ couplings), (B) dimer amidite 9a (9ꢁ couplings) and (C) trimer
amidite 14 (6ꢁ couplings), under, purposefully, un-optimized conditions (0.1 M,
600 s coupling).
10. Daub, G. W.; Van Tamelen, E. E. J. Am. Chem. Soc. 1977, 99, 3526.
11. Smith, D. J. H.; Ogilvie, K. K.; Gillen, M. F. Tetrahedron Lett. 1980, 21, 861.
12. Ogilvie, K. K.; Theriault, N. Y.; Seifert, J.-M.; Pon, R. T.; Nemer, M. J. Can. J. Chem.
1980, 58, 2686.
13. Scaringe, S. A.; Wincott, F. E.; Caruthers, M. H. J. Am. Chem. Soc. 1998, 120,
11820.
14. Alul, R. H.; Singman, C. N.; Zhang, G.; Letsinger, R. L. Nucleic Acids Res. 1991, 19,
1527.
amounts of nꢀ1 and nꢀ2 sequences using block amidites. This
likely arises from premature silyl cleavage during ammonia treat-
ment, as first reported by Ogilvie and co-workers for a poly-rU se-
quence.16,17 Conditions that greatly minimize this problem have
been described16 and will be adopted when deprotecting se-
quences of mixed base composition (work in progress).
15. Scaringe, S. A. Methods 2001, 23, 206.
16. Wu, T.; Ogilvie, K. K.; Pon, R. T. Nucleic Acids Res. 1989, 17, 3501.
17. Scaringe, S. A.; Francklyn, C.; Usman, N. Nucleic Acids Res. 1990, 18, 5433.
18. Yamada, T. et al J. Org. Chem. 2011, 76, 1198.