Streamlined process for the chemical synthesis of RNA using 2-O-thionocarbamate-protected nucleoside phosphoramidites in the solid phase

Article abstract of DOI:10.1021/ja201561z

An improved method for the chemical synthesis of RNA was developed utilizing a streamlined method for the preparation of phosphoramidite monomers and a single-step deprotection of the resulting oligoribonucleotide product using 1,2-diamines under anhydrous conditions. The process is compatible with most standard heterobase protection and employs a 2′-O-(1,1-dioxo- 1λ⁶-thiomorpholine-4-carbothioate) as a unique 2′-hydroxyl protective group. Using this approach, it was demonstrated that the chemical synthesis of RNA can be as simple and robust as the chemical synthesis of DNA.

Full text of DOI:10.1021/ja201561z

ARTICLE
pubs.acs.org/JACS
Streamlined Process for the Chemical Synthesis of RNA Using
20-O-Thionocarbamate-Protected Nucleoside Phosphoramidites
in the Solid Phase
Douglas J. Dellinger,*,† Zoltꢀan Timꢀar,†,‡ Joel Myerson,†,‡ Agnieszka B. Sierzchala,§,‡ John Turner,†,z
Fernando Ferreira,† Zoltꢀan Kupihꢀar,#,^ Geraldine Dellinger,† Kenneth W. Hill,§ James A. Powell,§
Jeffrey R. Sampson,† and Marvin H. Caruthers#
†Agilent Laboratories, Agilent Technologies, Inc., 5301 Stevens Creek Boulevard, Santa Clara, California 95051, United States
#Department of Chemistry and Biochemistry, University of Colorado, Boulder, Colorado 80309, United States
§Agilent Nucleic Acids Solutions Division, 5555 Airport Boulevard, Suite 100, Boulder, Colorado 80301, United States
ABSTRACT: An improved method for the chemical synthesis of RNA was
developed utilizing a streamlined method for the preparation of phosphor-
amidite monomers and a single-step deprotection of the resulting oligoribo-
nucleotide product using 1,2-diamines under anhydrous conditions. The
process is compatible with most standard heterobase protection and employs
a 20-O-(1,1-dioxo-1λ6-thiomorpholine-4-carbothioate) as a unique 20-hydroxyl
protective group. Using this approach, it was demonstrated that the chemical
synthesis of RNA can be as simple and robust as the chemical synthesis of DNA.
’ INTRODUCTION
postsynthetic processing needed to deprotect the RNA pro-
ducts can be unwieldy, time-consuming, difficult to scale, and
expensive.
Achieving high-yield synthesis of phosphoramidite monomers
necessitates that protection of each of the reactive groups on the
carbohydrate residues be accomplished regiospecifically in nearly
quantitative yield. For DNA synthesis monomers, this is illu-
strated by the preparation of 50-(4,40-dimethoxytrityl)-N-acetyl-
20-deoxycytidine 30-(2-cyanoethyl)-N,N-diisopropylphosphor-
amidite (Scheme 1).
Recent revelations about the large and expanding roles that
noncoding RNAs have in the biology of cell differentiation and
maintenance have inspired a significant expansion in both
research and drug discovery on small RNAs.1ꢀ4 In turn, there
has been a renewed interest in finding a cheaper and less
cumbersome method for the chemical synthesis of small RNAs
and complex RNA sequences containing a variety of chemical
modifications.5,6 Several new protective groups, synthesis proto-
cols, and deprotection strategies have been proposed and
demonstrated.7ꢀ14 Akin to previous protocols, none of these
recent developments combine streamlined, high-yield, and in-
expensive monomer synthesis with simple and inexpensive
isolation of the desired RNA products.
For many years, it had been acknowledged that the most
effective method for making sequence-defined DNA oligo-
nucleotides is based upon the solid-phase phosphoramidite
method.15,16 Modifications to the original protocols gave rise
to streamlined, high-yield, inexpensive monomer synthesis,
followed by rapid, high-yielding, solid-phase oligonucleotide
synthesis, and finally simple, inexpensive deprotection and iso-
lation of the desired DNA products.17 By comparison, applica-
tion of the phosphoramidite method to RNA synthesis has
historically reported monomer preparations that are cumber-
some, low-yielding, and expensive. The lower coupling efficiency
of typical RNA monomers also severely limits the length of RNA
attainable by solid-phase oligonucleotide synthesis, and the
20-Deoxynucleosides are converted in three steps to the
desired phosphoramidite monomer: (1) The heterobase amine
is regiospecifically converted to an amide using transient silyla-
tion of the hydroxyl residues with trimethylsilyl chloride and
acylation. (2) The primary hydroxyl (50-OH) is regioselectively
protected with 4,40-dimethoxytriphenylmethyl chloride (DMT-
Cl). (3) The remaining secondary hydroxyl is phosphitylated
using (2-cyanoethyl)-N,N-diisopropylchlorophosphoramidite or
(2-cyanoethyl)-N,N,N0,N0-tetraisopropylphosphordiamidite.
The overall scheme gives ∼70% isolated yield of the desired
DNA synthesis monomer. This high-yield regiospecific scheme
results in low-cost, high-purity DNA monomers.
The typical deprotection and isolation protocol for DNA pro-
ducts is also simple and efficient. Once the DNA oligonucleotide
Received: February 18, 2011
Published: June 20, 2011
r
2011 American Chemical Society
11540
dx.doi.org/10.1021/ja201561z J. Am. Chem. Soc. 2011, 133, 11540–11556
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Three-Step Procedure for the Preparation of DNA Synthesis Monomers
has been prepared in the solid phase, the phosphorus protective
groups and solid-support linkage are cleaved concurrently with
the heterobase protective groups using ammonium hydroxide.
For large-scale synthesis or to obtain higher integrity DNA
products, the phosphorus protective group is removed first with
a hindered primary or secondary amine such as tert-butylamine
or diethylamine.18 The amine traps the acrylonitrile byproduct of
the β-cyanoethyl protective group19 and prevents modification
to the heterobases. The solid-support linkage is then cleaved
along with the heterobase amide protective groups using a more
nucleophilic amine, typically ammonia or methylamine under
aqueous conditions. The DNA dissolves in this aqueous amine
solution and is isolated by filtration from the solid support. The
aqueous amine solution containing the DNA is finally evaporated
and the product purified by reverse-phase or ion-exchange
chromatography.
Monomers for RNA synthesis have conventionally been much
more difficult to prepare than DNA monomers, the additional
hydroxyl in the 20-position compelling the presence of an extra
protective group. For a streamlined synthesis, this protective
group needs to be efficiently placed on the nucleoside in a
regiospecific manner. The vicinal cis-diol motif of the 20- and 30-
hydroxyls on the RNA nucleoside is notoriously difficult to
regiospecifically protect due to protective group migration
between the two hydroxyls. Typical esters are eliminated as a
potential choice for 20-hydroxyl protection because the rate of
acyl migration is such that it is difficult to isolate the 20- from the
30-isomer.20 A number of quite elegant attempts at producing
RNA synthesis monomers have been reported, but often the
reaction of the protective group reagent with the secondary
hydroxyl proceeds in low yield, and the reagent itself is difficult or
expensive to synthesize.21 In several examples, low reactivity of
the protective group reagent necessitates conversion of the 20-
hydroxyl to an alkoxide or metal complex,22 which invariably
results in low yields due to side reactions with the heterobases.23
The most frequently used 20-protective group for RNA
synthesis is the fluoride-labile tert-butyldimethylsilyl (TBDMS)
ether.24 This protective group is commonly placed upon the
nucleoside non-regiospecifically, and chromatographic separa-
tion is used to isolate the 20-protected nucleoside intermediate
from the 30-protected regioisomer, giving low yields of the final
RNA monomers.25 The use of silicon-containing protective
groups for the 20-hydroxyl, removed from the RNA product with
fluoride ion, has produced robust methods for the chemical
synthesis of RNA but necessitates removal of the synthetic RNA
from glass surfaces prior to fluoride ion exposure. This technique
is not practical in applications such as beads or microarrays26 and,
together with the extra processing and purification required to
remove the fluoride reagent, greatly increases the complexity of
performing high-throughput RNA synthesis.
In contrast, RNA monomers that do not require fluoride ion to
remove the 20-hydroxyl protective group can utilize the con-
venient regiospecific synthesis methods originally developed by
Markiewicz (Scheme 2).27 This approach was shown to be an
effective way to produce 20-orthoester-protected RNA mono-
mers in high yield.28 A similar strategy has been described to
produce TBDMS-protected RNA monomers in higher yields by
eliminating the need for chromatographic separation of the 20-
and 30-isomers using 30ꢀ50-cyclic di-tert-butylsilane-protected
nucleosides.29 The TBDMS group is regiospecifically attached to
the 20-position, and then the 30ꢀ50-cyclic silane is orthogonally
removed using HF/pyridine solutions in dichloromethane
(DCM) at low temperature.30
It has been well accepted in the literature that the protective
group for the 20-hydroxyl should be removed from the final RNA
product under neutral to mildly acidic conditions due to the
lability of the RNA internucleotide linkage at pH extremes31
(Scheme 3). In both aqueous acid and aqueous base, the 30ꢀ50-
phosphodiester linkage of the RNA molecule entropically favors
the formation of a nucleotide 20ꢀ30-cyclic phosphate, resulting in
strand cleavage. The ease with which these reactions can occur
has limited the choice of functional groups appropriate for
protection of the 20-hydroxyl and significantly constrains the
reaction conditions suitable for removal. Under acidic condi-
tions, the internucleotide bond of RNA can cleave or isomerize,
giving rise to 20ꢀ50-linked oligonucleotides, while under basic
conditions RNA readily undergoes cleavage via a transesterifica-
tion reaction involving the deprotonated 20-oxyanion group.32
Under basic conditions, 20f30 migration does not occur, appar-
ently because the lifetime of the dianionic phosphorane inter-
mediate is too short for pseudorotation to occur.33 The pKa of
the RNA 20-hydroxyl in aqueous solution can vary depending on
salt concentration and base sequence, but it is typically reported
to be <13.34,35 The pKa of (protonated) ammonia is ∼9.2; the
concentrated aqueous ammonium hydroxide solution most often
used for removing protective groups from synthetically prepared
oligonucleotides thus has a pH > 12. At this high pH, the 20-
hydroxyl will be at least 10% ionized, and base-catalyzed trans-
esterification results in backbone cleavage.
Commonly used nucleobase protecting groups for DNA
synthesis are removed with aqueous solutions of amines such
as ammonia or methylamine. For such groups to be used in RNA
synthesis, it is essential that the 20-hydroxyl protection remain
intact during any aqueous amine treatment, hence the common
11541
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Scheme 2. Regiospecific Placement of a 20-Hydroxyl Protective Group (PG) Using the Markiewicz Method
Scheme 3. RNA Internucleotide Bond Cleavage under
Acidic or Basic Conditions
Scheme 4. Effect of Solvent on Charge Stabilization and
Acid/Base Equilibrium Constants
cleavage is observed during extended treatment at elevated
temperatures of amide nucleobase-protected RNA oligonucleo-
tides, due to hydroxide-mediated cleavage of the 20-TBDMS
siliconꢀoxygen bond,36 and can be minimized by using aqueous
methylamine at ambient temperature for short exposure times.37
Under basic conditions, the rate of RNA 20ꢀ30-cyclophos-
phate formation and internucleotide bond cleavage is known to
be first-order in 20-oxyanion concentration. From pH 9 to ∼13,
specific base catalysis generates a 10-fold increase in the number
of reactive 20-oxyanion groups for each unit increase in pH, and a
10-fold increase in the rate of cleavage.38 However, it is also
known that the ionization constants of weak acids and bases may
be substantially altered in the presence of organic solvents.39
Under certain nonaqueous conditions, amine bases will be less
prone to deprotonate alcohols due to the properties of acid/base
equilibria in such systems. In general, for a neutral compound
ionizing to a charged anionic species, such as a hydroxyl group
ionizing to an alkoxide, decreasing the dielectric of the solvent
results in a decrease in the acid equilibrium constant (increase in
pKa) for the equilibrium shown in Scheme 4a.
use of the amine-stable, fluoride-labile TBDMS 20-hydroxyl
protecting group. If the 20-hydroxyl is unintentionally uncovered
during the aqueous amine nucleobase deprotection, the result is
hydroxyl deprotonation and RNA backbone cleavage. This
This effect is due to the less polar solvent being less able to
stabilize the two new charges formed, resulting in a shift of the
11542
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Scheme 5. Depyrimidination of Uridine Nucleosides with
Peroxyanions
Scheme 6. Cleavage of Thiocarbonate Protective Groups
from Uridine Nucleosides with TBAF in DMSO
Figure 1. Carbonyl and thionyl 20-hydroxyl protective group motifs
shown to be resistant to 20f30 migration.
equilibrium to the left. For a charged compound dissociating to a
charged plus a neutral compound (Scheme 4b), such as the
dissociation of a protonated amine, decreasing the dielectric of
the solventingeneral results inonly relativelysmall changes in pKa.
The dissociation results in no net change in charge, and the less
polar solvent destabilizes both sides of the equilibria, with the
resulting changes in pKa tending to be significantly less. This
observationisdemonstrated by measuringthe changeinpKa of the
acid/base pair of phenol and ammonia in water versus acetonitrile
(ACN). The pKa of phenol is reported to be ∼10 in water
(dielectric constant, 78), while in ACN (dielectric constant, 36)
phenol is a much weaker acid, with pKa ≈ 27 (ΔpKa ≈ 17).40
Ammonia does become a stronger base in ACN; the pKa of the
conjugate acid increases from 9.2 to 16.5 when moving from water
to ACN, but the ΔpKa is only ∼7;41 the acid/base pair of phenol
and ammonia, which in water has a pKa difference of <1 pKa unit,
has a pKa difference of ∼10 pKa units in ACN.
mediated removal of the siloxane. We found that rapid 20f30
migration limited our choices to non-aryl thiocarbonates, tertiary
carbonates, carbamates, and thionocarbamates (Figure 1); all
other carbonyl or thionyl protective groups examined quickly
isomerized from the 20- to the 30-hydroxyl during removal of the
disiloxane, using a range of HF/triethylamine (TEA) or HF/
pyridine compositions, under conditions of varying concentra-
tions, pH, and temperatures.
Tertiary carbonates were previously described as 20-hydroxyl
protective groups for RNA synthesis by Losse et al.42 Although
stable to migration, this protective group motif was eliminated
from our study due to our inability to find a high-yield method for
its removal with nucleophiles.
This phenomenon could be used advantageously in a depro-
tection scheme for synthetic oligoribonucleotides, wherein a
hydroxyl becomes significantly less acidic while a conjugate
amine base becomes only slightly more basic. The acid/base
pair of such an oligoribonucleotide 20-hydroxyl and an anhydrous
amine should show a trend similar to that demonstrated with
phenol/ammonia. Based upon reported pKa values, it should be
possible to find an amine whose nucleophilic properties could be
utilized to simultaneously remove the heterobase protective
group, the 20-hydroxyl protective group, and the solid-support
attachment while not giving base-catalyzed backbone cleavage.
This approach would avoid the complications introduced by the
use of fluoride and completely avert any possible acid-catalyzed
isomerization of the internucleotide bond, resulting in higher
integrity RNA synthesis products obtained under simplified
conditions analogous to those of DNA synthesis.
We then evaluated removal of the remaining carbonyl and
thionyl protective groups using the mildly basic, highly nucleo-
philic deprotection conditions containing peroxyanions as devel-
oped by Sierzchala et al.43 Under these conditions, both
thionocarbamates and thiocarbonates were removed effectively
while carbamates were left intact.
During our studies with peroxyanions, we discovered that
uridine heterobase, unlike thymine, is susceptible to cleavage,
resulting in a uridine byproduct with a loss of 52 amu. The
formation of this compound is consistent with the conversion of
the heterobase to urea (Scheme 5) and appears to be analogous
to the described depyrimidination of 20-fluorouridine with alkyl
amines.44 Thus, peroxyanion-containing solutions were aban-
doned as potential cleavage reagents.
Thiocarbonates could be readily removed from the 20-hydr-
oxyl using tetrabutylammonium fluoride (TBAF) in dimethyl
sulfoxide (DMSO) under conditions that should leave the
internucleotide bond intact (Scheme 6). Although effective,
the use of fluoride ion for the removal of the 20-hydroxyl
protective group would not allow for a one-step final deprotec-
tion and as a result this approach was also abandoned.
Use of strong nucleophiles such as ammonia or methylamine
demonstrated that most 20-carbamates remained completely
’ RESULTS
The desire for an efficient monomer synthesis led us to initiate
our investigation of nucleophile-labile 20-protective groups by
exhaustively screening possible carbonyl and thionyl groups for
two qualities: first, their ability to react with a 30ꢀ50-tetraisoprop-
yldisiloxane-protected nucleoside to produce the desired 20-
hydroxyl-protected products in high yield, and second, their
ability to resist protective group migration during fluoride-
11543
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Scheme 7. Complete Conversion of Thiocarbonates to
Carbamates by Exposure to Nucleophilic Amines with
No Detectable Formation of Ureas or Cyclic Ureas
Scheme 9. Elimination of Primary Thionocarbamates To
Form Isothiocyanates
products (Scheme 8a), and when protected nucleosides were
treated with 1,2- or 1,3-diamines at room temperature, along with
product it was possible to detect formation of the cyclic thiourea
products by LC/MS (Scheme 8b). However, due to rapid 20f30
migration, we were unable to isolate the 20-thionocarbonate-
protected nucleosides free of a significant contamination of the
30-isomer.
These results led us to focus on thionocarbamates, which
when treated with strong nucleophilic amines proved to be more
labile than the corresponding carbamates. We evaluated the rate
of removal of potential thionocarbamate 20-hydroxyl protective
groups with anhydrous gaseous ammonia by preparing a variety
of uridine phosphoramidite monomers, each with a different 20-
hydroxyl thionocarbamate protective group, and synthesizing the
corresponding protected UT dimers on controlled pore glass
(CPG). Ammonia deprotection was performed in a pressure
vessel at 80 psi, conditions previously used for deprotection and
cleavage of oligonucleotides from a solid-phase resin.45 While it
has been reported that thionocarbamates formed from primary
amines are readily cleaved by an E1cb mechanism through base-
catalyzed isothiocyanate formation,46 we found that, under
anhydrous gaseous ammonia conditions, thionocarbamates de-
rived from primary amines were cleaved quite slowly. Even the
aniline thionocarbamate was slow to cleave, which via elimina-
tion should favorably form phenylisothiocyanate (Scheme 9).
The relative acidity of the conjugate acids of the primary
amines was not consistently correlated with the rate of deprotec-
tion of the thionocarbamate, and most notably the thionocarb-
amate of morpholine, a secondary amine, was significantly more
rapidly removed by gaseous ammonia than any of the tested
primary amine thionocarbamates (Table 1).
Scheme 8. Complete Cleavage of Thionocarbonates by
Exposure to Nucleophilic Amines with Detectable
Formation of Ureas or Cyclic Ureas
We then screened cleavage rates with gaseous ammonia for a
variety of thionocarbamates made with secondary amines that
spanned a range of pKa values. Thiomorpholine-1,1-dioxide in
particular became the leading candidate, as the cleavage rate of
this group was comparable to the reported rate of deprotection
of heterobase-protected N-2-isobutyrylguanosine by gaseous
ammonia45 (Table 2).
With this encouraging result, the uridine phosphoramidite
monomer, 20-hydroxyl-protected as the thionocarbamate of
thiomorpholine-1,1-dioxide, was used for solid-phase synthesis
of U4T. Deprotection with gaseous ammonia at room tempera-
ture for 6 h gave RNA products without detectable amounts of
internucleotide bond cleavage, as analyzed by ion-exchange
HPLC or LC/MS (Figure 2).
However, for the longer U9T oligoribonucleotide, 6 h of
exposure to ammonia gas at room temperature gave a typical
internucleotide bond cleavage profile (Figure 3) and diminished
amounts of full-length product. The use of gaseous methylamine
for a shorter time45 gave even more internucleotide bond
intact after several hours of treatment at elevated temperatures,
and when 20-thiocarbonates were treated with alkylamines, they
were shown to undergo facile transformation into highly stable
20-carbamate products (Scheme 7a). The use of 1,2- or 1,3-
diamines gave only the monosubstituted product and did not
result in any detectable formation of the cyclic urea and free 20-
hydroxyl products at temperatures <100 ꢀC (Scheme 7b).
When the sulfur atom of the thiocarbonate linkage was moved
to the double bond, forming the thionocarbonate, the reaction
with ammonia gas gave exclusively the desired 20-hydroxyl
11544
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Table 1. Cleavage Rates of Primary Amine 20-O-Thionocarbamate Protective Groups from UT Dimers on CPG Using Ammonia
Gas in a Pressure Vessel at 80 psi
* Calculated using Advanced Chemistry Development Software V11.02 (1994ꢀ2011, ACD/Labs).
Table 2. Cleavage Rates of Secondary Amine 20-O-Thionocarbamate Protective Groups from UT Dimers on CPG Using
Ammonia Gas in a Pressure Vessel at 80 psi
* Calculated using Advanced Chemistry Development Software V11.02 (1994ꢀ2011, ACD/Labs).
cleavage. Apparently the rate of removal of the thiomorpholine-
1,1-dioxide thionocarbamate group by the nucleophilic gaseous
amine was too slow to avoid the gaseous amine base-catalyzed
internucleotide bond cleavage.
As discussed above, we believed that it should be possible to
find anhydrous amine treatment conditions that do not cause
cleavage of the RNA internucleotide bond. This was supported
by a report by Goldsborough, in which full-length native RNA
11545
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Figure 2. Ion-exchange HPLC and LC/MS evaluation of U4T synthesized on thymidine containing CPG using 20-O-TC-protected uridine
phosphoramidite and deprotected for 6 h with gaseous ammonia.
Ethylenediamine was then examined for its ability to remove a
variety of thionocarbamate protecting groups from CPG-bound
UT15 oligonucleotides. These model oligoribonucleotides were
synthesized by reacting CPG-bound T15 with our previously
prepared uridine phosphoramidites containing a series of differ-
ent thionocarbamates in the 20-position. The degree of thiono-
carbamate protective group removal from the products after a 2 h
exposure to neat, anhydrous EDA was evaluated by HPLC and
LC/MS. The 20-thionocarbamate of thiomorpholine-1,1-dioxide
was shown to be completely removed after 2 h, which was faster
than for any of the other tested thionocarbamates (Table 3).
Importantly, this rate was once again similar to that reported for
removal of the N-2-isobutyryl group from guanosine using
EDA,48 suggesting that all protective groups could be removed
simultaneously from an oligoribonucleotide without increasing
reaction exposure time or resorting to alternative heterobase
protective groups.
Figure 3. Ion-exchange HPLC evaluation of U9T synthesized on
thymidine containing CPG using 20-O-TC-protected uridine phosphor-
amidite and deprotected for 6 h with gaseous ammonia.
All four typical nucleoside phosphoramidites were then pre-
pared with the 20-hydroxyls protected as the thionocarbamate of
thiomorpholine-1,1-dioxide, forming 20-O-(1,1-dioxo-1λ6-thio-
morpholine-4-carbothioate), the TC protective group (Figure 5).
A mixed-sequence oligoribonucleotide was synthesized on an
ABI 394 DNA/RNA synthesizer using a thymidine containing
CPG and the standard 1.0 μmol RNA synthesis cycle.
After treatment of the resin with EDA for 2 h, the amine was
removed by filtration, and the CPG was washed with ACN. The
oligonucleotide remained adsorbed to the CPG during this
process, after which it was eluted from the resin with water. This
crude product was compared to the same oligoribonucleotide
synthesized from commercially obtained 20-TBDMS-protected
was recovered from 20-hydroxyl acetylated transcripts after
exposure to alkylamines under conditions containing minimal
amounts of water.47 To explore the stability of RNA under
conditions pertinent for removal of a thionocarbamate protect-
ing group, a synthetic, fully deprotected, mixed-sequence 21-mer
oligoribonucleotide was treated with a variety of neat, anhydrous
alkyl amines. From this study, ethylenediamine (EDA) was
shown to be particularly advantageous, since no significant
internucleotide cleavage was detected after 17 h at room
temperature (Figure 4).
11546
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Figure 4. LC/MS of mixed-sequence RNA 21-mer adsorbed to CPG, treated with neat EDA for 17 h, washed with ACN to remove EDA, dried under
vacuum, eluted from CPG with water, and injected into the LC/MS instrument.
Table 3. Cleavage Rates of Primary and Secondary Amine
20-O-Thionocarbamate Protective Groups from UT15 on
CPG Using Neat EDA for 2 h at Room Temperature
monomers was of a quality comparable to that of the product
made with TBDMS-protected monomers.
To further demonstrate the utility of this approach, we
prepared a single-strand 54-nucleotide (nt) minimal ribozyme
hammerhead50 (Figure 8) on 30-O-acetylguanosine containing
CPG. The oligoribonucleotide should self-cleave in the presence
of divalent metal ions to produce a 46-nt fragment and an 8-nt
fragment with a 20ꢀ30 cyclic phosphate.
The oligoribonucleotide was deprotected using neat EDA for 2 h
at room temperature, the CPG was washed with ACN and dried
with a stream of argon, the desired RNA was washed from the
column using 100 mM ethylenediaminetetraacetic acid (EDTA),
pH 8.0, and the crude product was analyzed by ion-exchange
HPLC (Figure 9a). The EDTA was removed using Sepadex G-25
in 15 mM Tris buffer, pH 7.5, and the product was analyzed by ion-
exchange HPLC, showing partial cleavage of the oligonucleotide
resulting in three distinct products (Figure 9b), and by LC/MS,
confirming the identity of the three products as the full-length 54-
mer, a 46-mer, and an 8-mer cyclic phosphate (Figures 10 and 11).
Magnesium chloride was added to the desalted oligoribonucleotide
to a final concentration of 15 mM, and the mixture was allowed
to sit at room temperature for 24 h, resulting in complete conver-
sion of the material to the 46-mer and 8-mer cyclic phosphate
(Figure 9c). The 46-nt cleavage product was synthesized as an ion-
exchange HPLC standard using the same method, confirming the
ion-exchange HLPC retention time (Figure 9d).
’ DISCUSSION
The synthesis of RNA phosphoramidite monomers contain-
ing a 20-O-(1,1-dioxo-1λ6-thiomorpholine-4-carbothioate) (TC)
protective group can be accomplished regiospecifically in very
high overall yield with inexpensive, commercially available
reagents. These monomers exploit the standard 50-DMT pro-
tection as well as the amide heterobase and cyanoethylꢀ
phosphorus protections found in routine DNA synthesis. This
simple, streamlined monomer synthesis is illustrated in
Scheme 10, resulting in a 60ꢀ70% overall isolated yield of the
phosphoramidite monomer from the tetraisopropyldisiloxane-
protected nucleoside for all four common nucleosides.
The observation that RNA with a free 20-hydroxyl is remark-
ably resistant to degradation during treatment with certain amine
bases under nonaqueous conditions has allowed us to employ 20-
TC-protected RNA monomers for the high-yield synthesis and
isolation of mixed-sequence 21-mer oligoribonucleotides, which
RNA phosphoramidites using the recommended conditions
for synthesis, deprotection, and isolation.49 Analysis by ion-
exchange HPLC (Figure 6) and LC/MS (Figure 7) indicated
that the product made with the new RNA phosphoramidite
11547
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Figure 5. 20-O-TC-protected phosphoramidite monomers used for mixed-sequence RNA synthesis.
Figure 6. Comparison of ion-exchange chromatograms of crude RNA using TBDMS- and TC-protected monomers synthesized on thymidine
containing CPG (50-GUG UCA GUA CAG AUG AGG CCT-30): TBDMS, 40.8% full-length; TC, 45.9% full-length by peak integration.
Figure 7. LC/MS evaluation of crude RNA using TC-protected monomers synthesized on thymidine containing CPG (50-GUG UCA GUA CAG AUG
AGG CCT-30).
11548
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
are deprotected in a single step with neat EDA. The process flow
for the use of neat EDA is further advantageous in that the
cleaved and deprotected RNA 21-mer remains insoluble and
adsorbed to the solid-phase matrix. To obtain the fully
deprotected RNA, the solid support is simply washed with a few
drops of water or an aqueous buffer. For syntheses of oligo-
ribonucleotides smaller than ∼10 residues, recovery of the desired
RNA product is improved if toluene is used as a cosolvent during
the deprotection, since very short oligoribonucleotides can be
partially soluble in neat EDA. The rate of deprotection appears to
be proportional to the amount of EDA present: a 50% solution of
EDA in toluene gives complete deprotection in 4 h rather than
2 h, but with no detectable decrease in RNA quality. In our
screening of various amines, only anhydrous 1,2-diamines were
found to provide effective RNA deprotection without detectable
internucleotide bond cleavage.
amine under nonaqueous conditions is shown in Scheme 11. Amine
R-NH2 attacks the thiocarbonyl I, and the desired 20-hydroxyl II is
liberated along with the side product thiourea III. This reaction
likely goes through transition state IV and may be concerted or go
through tetrahedral intermediate V.51,52
In addition to the desired reaction is the possible formation of the
transcarbamoylation products VI and VII, a process that may be
favored by electron-withdrawing substituents on the amine leaving
group. LC/MS analysis of the reaction of anhydrous monoamines
with thionocarbamate I showed varying amounts of the transcarba-
moylated product VI, along with the desired product II. Compound
VI was shown to be slowly converted to the desired product II,
presumably by further reaction with amine and formation of thiourea
VIII as a byproduct. The time required for conversion to the desired
product by this process can potentially result in undesirable amounts
of RNA degradation, due to the extended exposure of the depro-
tected RNA to amine base, even in nonaqueous systems.
1,2-Diamines proved to be particularly advantageous nucleo-
philes for the deprotection of the thionocarbamates. Ethylenedi-
amine, when used together with the TC protective group formed
from thiomorpholine-1,1-dioxide, gave complete deprotection and
formed only trace amounts of transcarbamoyated product
(Scheme 12). TC-protected RNA IX gives the desired fully
deprotected product II, and thiourea X that is formed may
decompose to the cyclic ethylenethiourea XI and thiomorpho-
line-1,1-dioxide XII. In addition, the transcarbomylation product
XIII is rapidly converted to the RNA product II, presumably via
direct intramolecular formation of the cyclic thiourea XI, although a
reversible addition/elimination mechanism going through the
intermediate XIV to form thiourea XV has not been ruled out.
Successful nucleophilic deprotection of the 20-TC-hydroxyl using
an amine requires that the thionocarbamate be sufficiently electro-
philic and sterically accessible such that the rate of deprotection is
fast relative to any subsequent base-catalyzed internucleotide bond
cleavage. Thus, it proved necessary to design a thionocarbamate that
was slightly electron-withdrawing. The desired reaction of a sec-
ondary amine thionocarbamate when treated with a nucleophilic
’ CONCLUSION
In all cases, the RNA produced by the currently described
method gave products that were comparable or superior to
control oligoribonucleotides produced using TBDMS-protected
Figure 8. Single-strand 54-nt minimal ribozyme hammerhead.
Figure 9. Ion-exchange HPLC of (a) crude 54-mer washed from CPG with 100 mM EDTA, (b) crude 54-mer desalted on G-25, (c) crude 54-mer
incubated with 15 mM of MgCl2 for 24 h at room temperature, and (d) crude 46-mer.
11549
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Figure 10. LC/MS of partially cleaved, desalted, 54-nt minimal hammerhead ribozyme showing the total ion chromatogram (TIC), UV absorbance
chromatogram (A260), and molecular-specific ion chromatograms for the 54-mer, 46-mer, and 8-mer cyclic phosphate.
monomers when evaluated by ion-exchange HPLC and LC/MS
for percent full-length product and total amount of deprotected
RNA recovered postsynthesis. This method appears to be
broadly useful for the rapid and simplified construction of
high-quality RNA sequences.
from Dionex. NMR data were recorded on a Bruker 400 MHz spectro-
meter. Tetramethylsilane was used as an internal reference for 1H and
13C NMR. An external capillary containing 85% H3PO4 was used as a
reference for 31P NMR. Downfield chemical shifts were recorded as
positive values for 31P NMR. LC/MS analyses were performed on
Agilent Technologies ESI-TOF and ESI-ion trap mass spectrometers
attached to Agilent Technologies 1100 or 1200 HPLC systems.
General Procedure for the Preparation of 20-O-Thiono-
carbamate-Protected Nucleoside Phosphoramidites 6aꢀp.
Step 1. 30,50-O-(Tetraisopropyldisiloxane-1,3-diyl) nucleosides of uri-
dine, adenosine, cytidine, and guanosine (1, N-protected as shown in
Figure 5) were dissolved in anhydrous ACN at concentrations of
0.2ꢀ0.3 M in a round-bottom flask. Next, 1.1 equiv of 1,10-thiocarbon-
yldiimidazole and 0.1 equiv of 4-(dimethylamino)pyridine (DMAP)
were added to the flask. The reaction was stirred or shaken for 2ꢀ3 h and
the reaction progress monitored by TLC, showing spot-to-spot conver-
sion. If necessary, the flask was heated to keep the reaction soluble. The
reaction products (2), the 30,50-O-(tetraisopropyldisiloxane-1,3-diyl)-20-
O-thiocarbonylimidazole nucleosides of uridine and cytidine, usually
crystallized after the reaction mixture was left to sit at room temperature
and were collected by filtration, while the nucleoside products (2) of
adenosine and guanosine typically remained in solution.
’ EXPERIMENTAL SECTION
General Procedures. Unless otherwise noted, materials were ob-
tained from commercial suppliers and used without further purification. All
solvents were obtained from Sigma-Aldrich. Thiomorpholine-1,1-dioxide
was obtained from TCI America. 1,10-Thiocarbonyldiimidazole was
obtained g95.0% pure from Sigma-Aldrich. (2-Cyanoethyl)-N,N-diiso-
propylchlorophosphoramidite was obtained from either Sigma-Aldrich
or ChemGenes. 30,50-O-(Tetraisopropyldisiloxane-1,3-diyl)-N2-acetyl-
cytidine, 30,50-O-(tetraisopropyldisiloxane-1,3-diyl)-uridine, 30,50-O-(tetra-
isopropyldisiloxane-1,3-diyl)-N6-benzoyl-adenosine, and 30,50-O-(tetra-
isopropyldisiloxane-1,3-diyl)-N2-isobutyryl-guanosine were obtained
from ChemGenes. All other DNA synthesis reagents were obtained
from Glen Research. Medium-pressure, preparative column chromatog-
raphy was performed using 230ꢀ400 mesh silica gel from Sorbent
Technologies. Thin-layer chromatography (TLC) was performed on
aluminum-backed silica 60 F254 plates from EMD Chemicals. HPLC
chromatography was performed on an Agilent Technologies 1200
HPLC instrument, reverse-phase using Zorbax C-18 columns from
Agilent Technologies and ion-exchange using DNAPac-100 columns
Step 2. Nucleoside products 2 that remained soluble in ACN were
used in solution without isolation or purification. Crystalline nucleoside
products 2 were dissolved at a concentration of 0.2 M in ACN with
heating and used for the next reaction while still warm to keep the
11550
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Figure 11. Deconvoluted LC/MS of partially cleaved, desalted, 54-nt minimal hammerhead ribozyme showing the calculated and observed exact
masses for the 54-mer, 46-mer, and 8-mer cyclic phosphate products.
starting material in solution. To each 20-O-thiocarbonylimidazole nucleo-
side 2 in solution was added 1.2 equiv of a primary or secondary amine, and
the reaction mixture was stirred or shaken and monitored by TLC for
completion. Most 30,50-O-(tetraisopropyldisiloxane-1,3-diyl)-20-O-thiono-
carbamate nucleoside products (3) spontaneously crystallized after this
step, and any remaining reactants were removed by filtration. Those which
did not readily crystallize at this stage were purified by medium-pressure
silica gel chromatography. For non-aniline derivatives, the reactions were
typically shown to be complete by TLC in 2 h without further heating. To
prepare aniline derivatives of cytidine, adenosine, and guanosine, the
thiocarbonylimidazole nucleosides 2 were heated to reflux in the presence
of a catalytic amount of DMAP. The reaction was typically complete in 3 h,
and the products (3) were isolated by filtration or silica gel chromatog-
raphy. Under these reaction conditions, the thiocarbonylimidazole nucleo-
side (2) derived from uridine was preferentially converted to the 2,20-
anhydro product. The 20-O-thionocarbamate aniline derivatives of uridine
(3) were successfully prepared by displacement of the 20-O-thiocarbon-
ylimidazole in the presence of p-toluenesulfonic acid monohydrate
(pTsOH). To the dissolved uridine nucleoside 2 in ACN was added 2.0
equiv of the aniline compound and 1.0 equiv of pTsOH. The reaction was
heated to reflux and allowed to proceed for 12 h, after which the reaction
was ∼50% complete as assayed by TLC, and products were purified by
silica gel chromatography.
proceed for 2 h at room temperature. The 2-MeTHF solution was
extracted twice with water and then dried over anhydrous sodium
sulfate. The sodium sulfate was removed by filtration and the solvent
evaporated under vacuum on a rotary evaporator, keeping the water bath
temperature <40 ꢀC.
Step 4. The crude desilylated 20-O-thionocarbamate nucleoside
product 4 was dissolved in DCM at a concentration of 0.1 M, and
1.05 equiv of N-methylmorpholine (NMM) was added, followed by 1.0
equiv of DMT-Cl. The reaction was followed by TLC, and if after 30 min
the reaction was not shown to be complete, additional portions of 0.1
equiv each of NMM and DMT-Cl were added sequentially until
complete consumption of the starting material.
Step 5. An additional 1.05 equiv of NMM was then added to the
reaction mixture, and the flask was cooled to 0 ꢀC in an ice/water bath.
To the unisolated 50-O-(4,40-dimethoxytrityl)-20-O-thionocarbamate
nucleoside product (5) was added (2-cyanoethyl)-N,N-diisopropyl-
chlorophosphoramidite (1.0 equiv) dropwise with stirring, and the
reaction was allowed to proceed for 2 h. The reaction was followed by
TLC, and if after 2 h the reaction was not shown to be complete,
additional portions of 0.1 equiv each of NMM and (2-cyanoethyl)-N,N-
diisopropylchlorophosphoramidite were added sequentially until com-
plete consumption of the starting material. The reaction mixture was
then extracted twice with a saturated aqueous solution of sodium
bicarbonate. Bulk water was removed from the organic layer by extrac-
tion with brine, and the DCM solution was dried over anhydrous sodium
sulfate. The solvent was evaporated on a rotary evaporator, and the
resulting 50-O-(4,40-dimethoxytrityl)-20-O-thionocarbamate nucleoside-
30-O-(β-cyanoethyl)-N,N-diisopropylphosphoramidite product (6) was
purified by silica gel chromatography in hexanes using the appropriate
ethyl acetate gradients from 50% to 100% or acetone gradients from 10%
Step 3. All isolated 30,50-O-(tetraisopropyldisiloxane-1,3-diyl)-20-O-
thionocarbamate nucleoside products (3) were dissolved in 2-methyl-
tetrahydrofuran (2-MeTHF) at a concentration of 0.2 M, and 6.0 equiv of
anhydrous pyridine was added. The mixture was cooled to 0 ꢀC in an
ice/water bath, and 12.0 equiv of hydrogen fluoride was added dropwise
as a solution of HF/pyridine (70% HF). After addition, the flask was
removed from the ice/water bath, and the reaction was allowed to
11551
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Scheme 10. Streamlined Synthesis of 20-O-TC-Protected Phosphoramidite Monomers
to 50%, determined by product resolution on TLC. The identity of each
compound was determined by 31P and 1H NMR and mass spectrometry.
Important Experimental Notes. It is important to crystallize or silica
gel purify the protected nucleoside products prior to removing the 30,50-
O-tetraisopropyldisiloxane group with HF/pyridine in order to produce
high-quality final phosphoramidite products; this removes reaction
products from the first two steps of the reaction that can contaminate
the final products.
It is important to remove residual HF/pyridine, pyridine, and ACN
from the protected nucleosides prior to reacting with DMT-Cl; this is
done by extraction with water followed by evaporation under high
vacuum. Failure to remove these reagents will result in inhibition of the
tritylation and phosphitylation reactions.
the 20-thiocarbonylimidazole product (2a). The crystallized product was
dissolved in ACN (100 mL) by heating with a heat gun, and while still
hot, thiomorpholine-1,1-dioxide (2.90 g, 22 mmol, 1.1 equiv) was added
to the mixture and the reaction stirred for 2 h at 30 ꢀC. If the reaction
mixture crystallized before the first hour of stirring, it was dissolved by
additional heating and allowed to react for the final hour. The reaction
was evaporated to a solid on a rotary evaporator, and the product (3a)
was dissolved in 2-MeTHF (100 mL) with 10 mL of anhydrous pyridine
(6.0 equiv). The mixture was then cooled to 0 ꢀC in an ice/water bath,
and HF/pyridine (3.1 mL, 12 equiv) was added dropwise with stirring.
After addition, the flask was removed from the ice/water bath, and the
reaction was allowed to proceed for 2 h at room temperature. The
2-MeTHF solution was extracted with water twice (50 mL), and the
aqueous layers were combined and extracted twice with 25 mL portions
of 2-MeTHF. The organic layers were combined and dried over
anhydrous sodium sulfate. The sodium sulfate was removed by filtration
and the solvent evaporated under vacuum on a rotary evaporator,
keeping the water bath temperature <40 ꢀC. The resulting foam was
placed under high vacuum overnight to remove any residual solvent.
This crude product (4a) was then dissolved in 200 mL of DCM (0.1 M)
along with 1.6 mL of NMM (14.7 mmol). DMT-Cl (4.75 g, 14 mmol)
was added to the solution with stirring. The reaction progress was
followed by HPLC on an Agilent Eclipse XDB-C8 column (3.5 μm,
3.0 ꢁ 100 mm column) using a water/ACN gradient from 35% to 95%
ACN. After 1 h, the reaction was shown to be ∼75% complete, and
another 0.6 mL of NMM was added to the reaction mixture, followed by
1.5 g (4.5 mmol) of DMT-Cl. The reaction was stirred for 2 h, and
analysis by HPLC demonstrated that the reaction was complete. To the
crude reaction mixture containing 5a, in the same flask was then added
2.1 mL of NMM (19.5 mmol), followed by 4.1 mL (18.5 mmol) of
It is important not to heat the protected nucleoside products above
40 ꢀC during the evaporation process after removing the 30,50-O-
tetraisopropyldisiloxane group with HF/pyridine; excess heating or
exposure to acids or bases can result in detectable 20f30 migration of
the thionocarbamate protective group.
Optimized Protocols for the Preparation of 20-O-(1,1-
Dioxo-1λ6-thiomorpholine-4-carbothioate)-Protected Nu-
cleoside Phosphoramidites (6aꢀd). Synthesis of 20-O-(1,1-Di-
oxo-1λ6-thiomorpholine-4-carbothioate)-50-O-(4,40-dimethoxytrityl)-
uridine-30-O-(β-cyanoethyl)-N,N-diisopropylphosphoramidite (6a).
30,50-O-(Tetraisopropyldisiloxane-1,3-diyl)-uridine (1a, 10 g, 20.5
mmol) was dissolved in ACN (100 mL). 1,10-Thiocarbonyldiimidazole
(4.13 g, 22.6 mmol, 1.1 equiv) and a catalytic amount of DMAP were
added, and the mixture was briefly heated with a heat gun to dissolve.
The reaction was stirred for 4 h at 30 ꢀC and then placed in a freezer at
ꢀ20 ꢀC to cool. After 1 h the product had crystallized to a white solid
and was isolated by filtration and dried under vacuum, giving 11.5 g of
11552
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
Scheme 11. Cleavage of 20-Thionocarbamates with Alkyl Amines
Scheme 12. Cleavage of 20-Thionocarbamates with 1,2-Diamines
(2-cyanoethyl)-N,N-diisopropylchlorophosphoramidite. After 15 min, a
precipitate formed and was dissolved by the addition of 25 mL of DCM
with stirring. After 2 h, the reaction was shown to be complete by HPLC,
and 100 mL of DCM was added, followed by 100 mL of saturated
NaHCO3. The reaction was stirred until neutralization was complete
and then transferred to a separatory funnel. The DCM layer was iso-
lated, extracted with saturated NaCl, and finally transferred to an
Erlenmeyer flask containing anhydrous Na2SO4. The dried DCM layer
was filtered and evaporated on a rotary evaporator. The resulting oil
was purified by medium-pressure silica gel chromatography on a column
11553
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
pre-equilibrated with 10% acetone in hexanes and neutralized with 1%
TEA; the excess TEA was removed by flushing the column with a
volume of 10% acetone in hexanes. The evaporated product was
dissolved in a minimum volume of DCM and placed on top of the silica
gel column. The product 6a was eluted using a gradient of acetone from
10% to 30% and evaporated to a foam, giving 13.5 g of product, a 71%
1H NMR and mass spectrometry. 31P NMR (ACN-d3) δ (ppm):
150.08, 149.35 (diastereomers). H NMR (ACN-d3) δ: 8.88 (s, 1H,
1
NH), 8.09ꢀ8.02 (dd, 1H, C5 or 6), 7.5, 7.35, 6.89 (m, 13H, DMT), 7.1
(dd, 1H, C5 or 6), 6.2 (m, 1H, 10), 6.11, 6.08 (m, 1H, 20), 4.8, 4.71 (m,
1H, 30), 4.9, 4.65 (m, 2H, thiomorpholine), 4.4, 4.35 (m, 1H, 40), 4.05,
3.87 (m, 2H, thiomorpholine), 3.49 (m, 2H, iPr), 3.48, 3.41 (m, 2H,
50), 2.77 (m, 2H, thiomorpholine), 2.66, 2.53 (m, 2H, thiomorpho-
line), 2.18 (s, 6H, DMT), 2.14 (s, 3H, CH3), 1.25 (m, 12H, iPr). ESI-
MS: m/z 987.3179 [M+Na]+.
1
overall yield. The compound was identified by 31P and H NMR and
mass spectroscopy. 31P NMR (ACN-d3) δ (ppm): 150.31, 149.40
1
(diastereomers). H NMR (ACN-d3) δ: 9.2 (bs, 1H), 7.65 (dd, 1H),
7.5, 7.35, 6.9 (m, 13H), 6.18 (dd, 1H), 6.04, 5.97 (m, 1H), 5.47 (dd, 1H),
4.93 (m, 1H), 4.83 (m, 1H), 4.65 (m, 1H), 4.4, 4.32 (m, 1H), 4.1, 4.0
(m, 1H), 3.9 (m, 1H), 3.78 (d, 6H), 3.8ꢀ3.7 (m, 2H), 3.6 (m, 2H),
3.49ꢀ3.17 (m, 4H), 3.07 (m, 2H), 2.67, 2.52 (m, 2H), 1.17 (m, 12H).
ESI-MS: m/z 946.2911 [M+Na]+.
Synthesis of 20-O-(1,1-Dioxo-1λ6-thiomorpholine-4-carbothioate)-
50-O-(4,40-dimethoxytrityl)-N6-benzoyladenosine-30-O-(β-cyanoethyl)-
N,N-diisopropylphosphoramidite (6c). 30,50-O-(Tetraisopropyldisilox-
ane-1,3-diyl)-N6-benzoyladenosine (1c, 6.14 g, 10 mmol) was dissolved
in DCM (14 mL, 0.7 M). 1,10-Thiocarbonyldiimidazole (1.88 g, 10.5
mmol) was added with a catalytic amount of DMAP. The mixture was
briefly heated with a heat gun to dissolve and then stirred for 3.5 h at
ambient temperature. Thiomorpholine-1,1-dioxide (1.49 g, 11 mmol)
was added and stirred for 2 h at ambient temperature. An additional
30 mL of ACN was then added to the reaction mixture, followed by 5 mL
of anhydrous pyridine. The mixture was cooled to 0 ꢀC in an ice bath,
and HF/pyridine (3.1 mL, 119 mmol) was added dropwise with stirring.
After the addition was complete, the reaction mixture was stirred for 2 h
at ambient temperature. Water (350 mL) was added to the stirring
mixture, the resulting emulsion was transferred to a separatory funnel,
and a small organic layer was isolated upon standing. The cloudy
aqueous layer was then extracted with 2-MeTHF (250 mL) to remove
the suspended organic material, and the two organic portions were
combined. The organic layer was then extracted again with water
(300 mL) to remove residual HF and pyridine, followed by brine
(300 mL) to remove bulk water. The organics were finally dried over
anhydrous Na2SO4. The Na2SO4 was removed by filtration, and the
solvents were evaporated. The resulting oily residue was dissolved in
2-MeTHF (150 mL) and the solvent evaporated, yielding 5.6 g of a
crude solid (4c) after drying under high vacuum. The solid was
suspended in a mixture of 2-MeTHF/DCM (1:1, 200 mL, 0.05 M),
and then 1.2 mL of NMM (11 mmol, 1.1 equiv) and 3.38 g of DMT-Cl
(10 mmol, 1 equiv) were added in small portions over 30 min with
stirring. Once the reaction was demonstrated to be complete by HPLC,
an additional 1.4 mL of NMM (13 mmol, 1.3 equiv) was added to the
crude reaction mixture containing 5c, followed by 2.67 mL of (2-
cyanoethyl)-N,N-diisopropylchlorophosphoramidite (12 mmol, 1.2
equiv), and the reaction was stirred at ambient temperature for 3.5 h.
Saturated NaHCO3 (100 mL) was added with stirring, and the mixture
was transferred to a separatory funnel. The organic layer was removed
and the aqueous layer extracted with 100 mL of DCM. The organic
layers were combined and dried over anhydrous Na2SO4. The solvent
solution was separated from the Na2SO4 by filtration, and the filtrate was
precipitated into hexanes (400 mL). The mixture was placed in a freezer
at ꢀ20 ꢀC overnight, after which time the suspended precipitate had
settled and adhered to the walls of the glass flask. Purification was
performed by medium-pressure silica gel chromatography on a column
pre-equilibrated with 10% acetone in hexanes and neutralized with 1%
TEA; the excess TEA was removed by flushing the column with a
volume of 10% acetone in hexanes. The cold solvent was decanted, and
the crude product was immediately dissolved in dry DCM and loaded
onto the top of the silica gel column. The product 6c was eluted using a
gradient of acetone from 10% to 40% and evaporated to a foam, giving
6.72 g of product, a 64% overall yield. The compound was identified by
31P and 1H NMR and mass spectrometry. 31P NMR (ACN-d3) δ
(ppm): 149.98, 149.50 (diastereomers). 1H NMR (ACN-d3) δ (ppm):
9.52 (bs, 1H, NH), 8.62 (s, 1H, C2 or 8), 8.33 (s, 1H, C2 or 8), 7.98 (m,
2H, Bz), 7.66 (m, 1H, Bz), 7.53 (m, 2H, Bz), 7.43 (m, 2H, DMT),
7.33ꢀ7.15 (m, 6H, DMT), 6.85ꢀ6.8 (m, 4H, DMT), 6.43 (m, 1H, 10),
5.25 (m, 1H, 20), 4.43 (m, 1H, 30), 4.78 (m, 1H, thiomorpholine), 4.62
Synthesis of 20-O-(1,1-Dioxo-1λ6-thiomorpholine-4-carbothioate)-
50-O-(4,40-dimethoxytrityl)-N2-acetylcytidine-30-O-(β-cyanoethyl)-N,
N-diisopropylphosphoramidite (6b). 30,50-O-(Tetraisopropyldisilox-
ane-1,3-diyl)-N2-acetylcytidine (1b, 10 g, 19 mmol) was dissolved in
DCM (28 mL, 0.7M). 1,10-Thiocarbonyldiimidazole (3.74 g, 21 mmol,
1.1 equiv) was added with a catalytic amount of DMAP, and the
mixture was briefly heated with a heat gun to dissolve. The reaction was
stirred for 2.5 h at ambient temperature, after which time the reaction
mixture became cloudy from the formation of a white precipitate.
Acetonitrile (45 mL) was added to the reaction, followed by thiomor-
pholine-1,1-dioxide (2.84 g, 21 mmol, 1.1 equiv). The reaction was
heated with a heat gun to dissolve the precipitate, and then the solution
was stirred for 2 h at ambient temperature. The reaction mixture was
evaporated to dryness, dissolved in a minimum amount of ACN with
mild heating, and then refrigerated overnight at 5 ꢀC. The product
crystallized from this mixture and was isolated by filtration, giving 13.4
g of 3b after drying under vacuum. The isolated crystals were dissolved
in 95 mL of 2-MeTHF, and 9.2 mL of pyridine was added. The mixture
was cooled to 0 ꢀC in an ice bath, and HF/pyridine (3.0 mL, 114 mmol,
12 equiv) was added dropwise with stirring. The reaction was stirred at
0 ꢀC for 1 h, transferred to a separatory funnel, and extracted with two
40 mL volumes of water. The water was back-extracted with 2-MeTHF,
and the organic layers were combined and evaporated on a rotary
evaporator, keeping the water bath at 35 ꢀC, resulting in a glassy solid
of 20-O-(1,1-dioxo-1λ6-thiomorpholine-4-carbothioate)-N2-acetyl-
cytidine (4b). The solid was dissolved in 175 mL of DCM, and 1.6 mL
of NMM (15 mmol) was added to the solution, followed by 5.1 g of
4,40-dimethoxytrityl chloride (15 mmol). The reaction progress was
followed by HPLC. After 1 h, the starting material reaction was shown
to be 85% converted to product. A second addition of NMM (0.33 mL,
3 mmol) and dimethoxytrityl chloride (1 g, 3 mmol) resulted in
complete conversion of the starting material, as assayed by HPLC.
The crude reaction mixture containing 5b was then converted to the
final product by addition of 1.6 mL of NMM (14.6 mmol), followed by
3.1 mL of (2-cyanoethyl)-N,N-diisopropylchlorophosphoramidite
(13.3 mmol). The reaction mixture was stirred for 2 h at room
temperature, and then 100 mL of DCM was added, followed by
100 mL of saturated NaHCO3. The reaction was stirred until neu-
tralization was complete and then transferred to a separatory funnel.
The DCM layer was isolated, extracted with saturated NaCl, and finally
transferred to an Erlenmeyer flask containing anhydrous Na2SO4. The
dried DCM layer was filtered and evaporated on a rotary evaporator.
The resulting oil was purified by medium-pressure silica gel chroma-
tography on a column pre-equilibrated with 10% acetone in hexanes
and neutralized with 1% TEA; the excess TEA was removed by flushing
the column with a volume of 10% acetone in hexanes. The evaporated
product was dissolved in a minimum volume of DCM and placed on
top of the silica gel column. The product 6b was eluted using a gradient
of acetone from 10% to 30% and evaporated to a foam, giving 12.3 g of
product, a 67% overall yield. The compound was identified by 31P and
11554
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
(m, 2H, thiomorpholine), 4.44 (m, 1H, thiomorpholine), 4.26 (m, 1H,
40), 3.8, 3.67 (m, 2H, CE), 3.75 (s, 6H, DMT), 3.59 (m, 2H, iPr),
3.5ꢀ3.36 (m, 2H, 50), 3.14 (m, 4H, thiomorpholine), 2.67, 2.52 (m, 2H,
CE), 1.18 (m, 12H, iPr). ESI-MS: m/z 1051.3601 [M+H]+, 1073.3429
[M+Na]+.
CE), 2.45 (m, 1H, CE), 2.44 (m, 1H, iBu), 1.12 (m, 6H, iBu), 1.18, 1.05 (m,
12H, iPr). ESI-MS: m/z 1033.3741 [M+H]+, 1055.3546 [M+Na]+.
Synthesis and Deprotection of Oligonucleotides. Oligonu-
cleotide synthesis was performed on an ABI 394 automated DNA/RNA
synthesizer using 1.0 μmol columns obtained from Glen Research. The
standard 1.0 μmol RNA synthesis cycle (10 min coupling) was utilized
without modification. The TC phosphoramidites were dissolved in
anhydrous ACN at concentration of 0.1 M, and a 0.25 M solution of
5-(ethylthio)-1H-tetrazole in anhydrous ACN was used as an activator.
Post synthesis, the column containing the CPG was briefly dried with a
stream of argon gas, followed by filling the column with neat EDA and
allowing the column to sit at ambient temperature for 2 h. The EDA was
rinsed from the column by flowing 1 mL of anhydrous ACN through the
column, followed by a stream of dry argon. A syringe containing 0.5 mL
of water was then attached to the column, and the RNA was washed from
the column into a 1.5 mL Eppendorf tube. The RNA was evaporated in a
Speed-Vac concentrator and analyzed by ion-exchange HPLC and LC/
MS (Figures 6 and 7). It was discovered that the TC phosphoramidites
of U and C (6a and 6b) could precipitate from the ACN solution upon
standing for several days, especially if the laboratory temperature
dropped below 20 ꢀC. The phosphoramidites of U and C were then
demonstrated to be stable to precipitation in a 50:50 mixture of
anhydrous ACN and anhydrous toluene (v/v), and the use of this
solvent system was shown to have no deleterious effect on the phos-
phoramidite coupling or resulting RNA products.
Synthesis of 20-O-(1,1-Dioxo-1λ6-thiomorpholine-4-carbothioate)-
50-O-(4,40-dimethoxytrityl)-N2-isobutyrylguanosine-30-O-(β-cyanoethyl)-
N,N-diisopropylphosphoramidite (6d). 20-O-(1,1-dioxo-1λ6-thiomorpho-
line-4-carbothioate)-N2-isobutyrylguanosine (1d, 12.2 g, 20.5 mmol) was
dissolved in DCM (40 mL, 0.5 M). 1,10-Thiocarbonyldiimidazole (4.13 g,
22.6 mmol, 1.1 equiv) and a catalytic amount of DMAP were added, and the
mixture was stirred to dissolve. The reaction was stirred for 2 h at ambient
temperature. Thiomorpholine-1,1-dioxide (2.90 g, 22 mmol, 1.1 equiv) was
added to the mixture, along with 20 mL of ACN. The reaction flask was
fitted with a condenser, heated to 50 ꢀCfor30min, andthenstirredfor1.5h
at ambient temperature, during which time the reaction product crystallized.
The crystallized product was isolated by filtration and dried under vacuum at
room temperature, yielding 15.9 g of dried solid (3d). The product was
dissolved in 2-MeTHF (100 mL) with 10 mL of anhydrous pyridine (6.0
equiv). The mixture was then cooled to 0 ꢀC in an ice/water bath, and HF/
pyridine (3.1 mL, 12 equiv) was added dropwise with stirring. After
addition, the flask was removed from the ice/water bath, and the reaction
was allowed to proceed for 2 h at room temperature. The 2-MeTHF
solution was extracted with water twice (50 mL). The aqueous layers were
combined and extracted twice with 25 mL portions of 2-MeTHF. The
organic layers were combined and then dried over anhydrous sodium
sulfate. The sodium sulfate was removed by filtration and the solvent
evaporated under vacuum on a rotary evaporator, keeping the water bath
temperature <40 ꢀC. The resulting foam was placed under high vacuum
overnight to remove any residual solvent. This crude product (4d) was then
dissolved in 200 mL of DCM (0.1 M), along with 1.6 mL of NMM (14.7
mmol). DMT-Cl (4.75 g, 14 mmol) was added to the solution with stirring.
The reaction progress was followed by HPLC. After 1 h, the reaction was
shown to be ∼70% complete, and another 0.7 mL of NMM was added to
the reaction mixture, followed by 2.1 g (6.2 mmol) of DMT-Cl. The
reaction was stirred for 2 h, and analysis by HPLC demonstrated that the
reaction was complete. To the crude reaction mixture containing 5d in the
same flask was then added 2.4 mL of NMM (21.5 mmol), followed by
4.6 mL (20.5 mmol) of (2-cyanoethyl)-N,N-diisopropylchlorophosphor-
amidite, and the reaction was stirred for 2 h at ambient temperature. After
2 h, the reaction was shown to be complete by HPLC, and 100 mL of DCM
added, followed by 100 mL of saturated NaHCO3. The reaction was stirred
until neutralization was complete, and then it was transferred to a separatory
funnel. The DCM layer was isolated, extracted with saturated NaCl, and
finally transferred to an Erlenmeyer flask containing anhydrous Na2SO4.
The dried DCM layer was filtered and evaporated on a rotary evaporator.
The resulting oil was purified by medium-pressure silica gel chromatogra-
phy on a column pre-equilibrated with 10% acetone in hexanes and
neutralized with 1% TEA; the excess TEA was removed by flushing the
column with a volume of 10% acetone in hexanes. The evaporated product
was dissolved in a minimum volume of DCM and placed on top of the silica
gel column. The product 6d was eluted using a gradient of acetone from
10% to 30% and evaporated to a foam, giving 14.4 g of product, a 68%
’ AUTHOR INFORMATION
Corresponding Author
doug_dellinger@agilent.com
Present Addresses
^Department of Medical Chemistry, University of Szeged, Hungary
zNoxxon Pharma AG, Berlin, Germany
Author Contributions
‡These authors contributed equally to this work.
’ ACKNOWLEDGMENT
We thank Richard Shoemaker, Albert Meyer, Victor Molker,
Margaret G. Readio, Michael Jones, David Eidenmeuller, and
Steve Simon for technical assistance; Andrew Goldsborough,
Brian Sproat, and Douglas Picken for helpful discussions; Steven
Laderman, Neil Cook, Darlene Solomon, and Agilent Technol-
ogies for support.
’ REFERENCES
(1) The ENCODE Project Consortium. Nature 2007, 447, 799ꢀ
816.
(2) Cheng, J.; et al. Science 2005, 308, 1149–1154.
(3) Carninci, P.; et al. Nat. Genet. 2006, 38, 626–635.
(4) Meister, G.; Tuschl, T. Nature 2004, 431, 343–349.
(5) Beaucage, S. L. Curr. Opin. Drug Discovery Dev. 2008, 11, 203–
216.
(6) Watts, J. K.; Deleavey, G. F.; Damha, M. J. Drug Discovery Today
2008, 13, 842–855.
(7) Cieslak, J.; Kauffman, J. S.; Kolodziejski, M. J.; Lloyd, J. R.;
Beaucage, S. L. Org. Lett. 2007, 9, 671–674.
(8) Lavergne, T.; Bertrand, J. R.; Vasseur, J. J.; Debart, F. Chemistry
1
overall yield. The compound was identified by 31P and H NMR and
mass spectrometry. 31P NMR (ACN-d3) δ (ppm): 150.96, 149.29
(diastereomers). 1H NMR (ACN-d3) δ (ppm): 9.03 (bs, 1H, NH), 7.91
(s, 1H, C8), 7.51 (m, 1H, DMT), 7.42ꢀ7.20 (m, 8H, DMT), 6.87ꢀ6.75
(m, 4H, DMT), 6.39, 6.31 (t, 1H, J = 6 Hz, 20), 6.22 (q, 1H, J = 5 Hz, 10),
4.98ꢀ4.88 (m, 1H, 30), 4.88ꢀ4.71 (m, 1H, thiomorpholine), 4.58ꢀ4.47
(m, 1H, thiomorpholine), 4.45, 4.38 (m, 1H, 40), 4.20 (m, 1H,
thiomorpholine), 4.00 (m, 1H, thiomorpholine), 3.87ꢀ3.65 (m, 1H,
CE), 3.76 (m, 6H, MeO), 3.59 (m, 2H, iPr), 3.59 (m, 1H, CE), 3.47, 3.4
(m, 2H, 50), 3.18ꢀ2.99 (m, 4H, thiomorpholine), 2.67 (t, 1H, J = 6 Hz,
2008, 14, 9135–9138.
(9) Umemoto, T.; Wada, T. Tetrahedron Lett. 2004, 45, 9529–9531.
(10) Ohgi, T.; Masutomi, Y.; Ishiyama, K.; Kitagawa, H.; Shiba, Y.;
Yano, J. Org. Lett. 2005, 7, 3477–3480.
11555
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556

Journal of the American Chemical Society
ARTICLE
(11) Karwowski, B.; Seio, K.; Sekine, M. Nucleosides Nucleotides
Nucleic Acids 2005, 24, 1111–1114.
(12) Zhou, C.; Honcharenko, D.; Chattopadhyaya, J. Org. Biomol.
Chem. 2007, 5, 333–343.
(46) Humeres, E.; Sanchez, M. D. M.; Lobato, C. M. L.; Debacher,
N. A.; de Souza, E. P. Can. J. Chem. 2005, 83, 1483–1491.
(47) Goldsborough, A. S. Stabilisation of nucleic acids. U.S. Patent
Appl. 20060147918, 2006.
(13) Semenyuk, A.; Foldesi, A.; Johansson, T.; Estmer-Nilsson, C.;
Blomgren, P.; Brannvall, M.; Kirsebom, L. A.; Kwiatkowski, M. J. Am.
Chem. Soc. 2006, 128, 12356–12357.
(14) Lackey, J. G.; Sabatino, D.; Damha, M. J. Org. Lett. 2007, 9,
789–792.
(48) Hogrefe, R. I.; Vaghefi, M. M.; Reynolds, M. A.; Young, K. M.;
Arnold, L. J., Jr. Nucleic Acids Res. 1993, 21, 2031–2038.
(49) Vargeese, C. Deprotection and purification of oligonucleotides
and their derivatives. U.S. Patent 6989442, 2006.
(50) Dahm, S. C.; Uhlenbeck, O. C. Biochemistry 1991, 30, 9464–
9469.
(15) Matteucci, M. D.; Caruthers, M. H. J. Am. Chem. Soc. 1981,
103, 3185–3191.
(51) Oh, H. K.; Oh, J. Y.; Sung, D. D.; Lee, I. J. Org. Chem. 2005,
(16) Beaucage, S. L.; Iyer, R. P. Tetrahedron 1992, 48, 2223–2311.
(17) Caruthers, M. H. Acc. Chem. Res. 1991, 24, 278–284.
(18) Hsiung, H.; Inouye, S.; West, J.; Sturm, B.; Inouye, M. Nucleic
Acids Res. 1983, 11, 3227–3239.
70, 5624–5629.
(52) Oh, H. K.; Oh, J. Y. Bull. Kor. Chem. Soc. 2006, 27, 143–146.
(19) Butskus, P. F. Russ. Chem. Rev. 1961, 30, 583–598.
(20) Reese, C. B.; Trentham, D. R. Tetrahedron Lett. 1965, 6, 2467.
(21) Somoza, A. Chem. Soc. Rev. 2008, 37, 2668–2675.
(22) Reginato, G.; Ricci, A.; Roelens, S.; Scapecchi, S. J. Org. Chem.
1990, 55, 5132–5139.
(23) Pitsch, S.; Weiss, P. A.; Jenny, L.; Stutz, A.; Wu, X. Helv. Chim.
Acta 2001, 84, 3773–3795.
(24) Ogilvie, K. K.; Sadana, K. L.; Thompson, E. A.; Quilliam, M. A.;
Westmore, J. B. Tetrahedron Lett. 1974, 15, 2861–2863.
(25) Scaringe, S. A.; Francklyn, C.; Usman, N. Nucleic Acids Res.
1990, 18, 5433–5441.
(26) Lackey, J. G.; Mitra, D.; Somoza, M. M.; Cerrina, F.; Damha,
M. J. J. Am. Chem. Soc. 2009, 131, 8496–8502.
(27) Markiewicz, W. T. J. Chem. Res. (S) 1979, 24–25.
(28) Scaringe, S. A.; Wincott, F. E.; Caruthers, M. H. J. Am. Chem.
Soc. 1998, 120, 11820–11821.
(29) Furusawa, K.; Ueno, K.; Katsura, T. Chem. Lett. 1990, 97–100.
(30) Serebryany, V.; Beigelman, L. Tetrahedron Lett. 2002, 43,
1983–1985.
(31) Reese, C. B. Protection of 20-hydroxy functions of ribonucleo-
sides. Current Protocols in Nucleic Acid Chemistry; Wiley: New York, May
2001; Chapter 2, Unit 2.2 (doi: 10.1002/0471142700.nc0202s00).
(32) Oivanen, M.; Kuusela, S.; Loennberg, H. Chem. Rev. 1998,
98, 961–990.
(33) Lonnberg, H.; Stromberg, R.; Williams, A. Org. Biomol. Chem.
2004, 2, 2165–2167.
(34) Levene, P. A.; Simms, H. S.; Bass, L. W. J. Biol. Chem. 1926,
70, 243–262.
(35) Acharya, S.; Foldesi, A.; Chattopadhyaya, J. J. Org. Chem. 2003,
68, 1906–1910.
(36) Wu, T.; Ogilvie, K. K.; Pon, R. T. Nucleic Acids Res. 1989,
17, 3501–3517.
(37) Reddy, M. P.; Farooqui, F.; Hanna, N. B. Tetrahedron Lett.
1995, 36, 8929–8932.
(38) Li, Y. F.; Breaker, R. R. J. Am. Chem. Soc. 1999, 121, 5364–5372.
(39) Sarmini, K.; Kenndler, E. J. Biochem. Biophys. Methods 1999,
38, 123–137.
(40) Coetzee, J. F.; Padmanabhan, G. R. J. Phys. Chem. 1965,
69, 3193–3196.
(41) Coetzee, J. F.; Padmanabhan, G. R. J. Am. Chem. Soc. 1965,
87, 5005–5010.
(42) Losse, G.; Naumann, W.; Winkler, A.; Sueptitz, G. J. Prakt.
Chem./Chem.-Ztg. 1994, 336, 233–236.
(43) Sierzchala, A. B.; Dellinger, D. J.; Betley, J. R.; Wyrzykiewicz,
T. K.; Yamada, C. M.; Caruthers, M. H. J. Am. Chem. Soc. 2003,
125, 13427–13441.
(44) Hill, K. W. Characterization of impurities formed during the
preparation of RNA based therapeutics: Reactions of substituted uridine
and cytidine bases with alkyl amines. Abstracts of Papers, 239th ACS
National Meeting, San Francisco, CA, March 21ꢀ25, 2010; American
Chemical Society: Washington, DC, 2010; ORGN-422.
(45) Boal, J. H.; Wilk, A.; Harindranath, N.; Max, E. E.; Kempe, T.;
Beaucage, S. L. Nucleic Acids Res. 1996, 24, 3115–3117.
11556
dx.doi.org/10.1021/ja201561z |J. Am. Chem. Soc. 2011, 133, 11540–11556