Synthesis of small interfering RNAs containing acetal-type nucleoside analogs at their 3′-ends and analysis of their silencing activity and their ability to bind to the Argonaute2 PAZ domain

Article abstract of DOI:10.1016/j.ejmech.2015.09.011

In this study, we aimed to create small interfering RNAs (siRNAs) with increased silencing activities and nuclease resistance properties. Therefore, we designed and synthesized five types of siRNA containing acetal-type nucleoside analogs at their 3′-dang

Full text of DOI:10.1016/j.ejmech.2015.09.011

European Journal of Medicinal Chemistry 103 (2015) 460e472
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Synthesis of small interfering RNAs containing acetal-type nucleoside
analogs at their 30-ends and analysis of their silencing activity and
their ability to bind to the Argonaute2 PAZ domain
, b, *
Natsumi Inada a, Kosuke Nakamoto b, Takashi Yokogawa c, Yoshihito Ueno a
a Course of Applied Life Science, Faculty of Applied Biological Sciences, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan
b United Graduate School of Agricultural Science, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan
c Department of Bio-molecular Science, Faculty of Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
In this study, we aimed to create small interfering RNAs (siRNAs) with increased silencing activities and
nuclease resistance properties. Therefore, we designed and synthesized five types of siRNA containing
acetal-type nucleoside analogs at their 30-dangling ends. We found that the siRNA containing 1-O-(2,2,2-
Received 13 March 2015
Received in revised form
20 August 2015
Accepted 6 September 2015
Available online 10 September 2015
trifluoroethyl)-
siRNAs and showed more resistance to nucleolytic degradation by a 30 exonuclease than a natural RNA
did. Thus, modification of siRNAs by addition of 1-O-(2,2,2-trifluoroethyl)- -D-ribofuranose may hold
promise as a means of improving the silencing activity and nuclease resistance of siRNAs.
© 2015 Elsevier Masson SAS. All rights reserved.
b
-D-ribofuranose at the 30-dangling end was the most potent among the synthesized
b
Keywords:
RNAi
siRNA
Acetal linkage
PAZ domain
Nuclease-resistant
1. Introduction
Piwi/Argonaute/Zwille (PAZ), middle (MID), and P-element-
induced wimpy testis (PIWI) domains [11ꢀ13]. It is known that the
50-OH of the guide strand is phosphorylated by cellular kinases in
cells [14], and in the RISC, the 50-phosphate of the guide strand is
recognized by the MID domain and tightly anchored in the Ago
protein [15,16]. It has also been reported that the 30-end of the
guide strand is recognized by the PAZ domain and that the nucle-
otides of the 30-dangling end are accommodated in the hydro-
phobic pocket in the domain [17ꢀ20]. However, it was suggested by
recent studies that the nucleotides are not always anchored in the
PAZ domain and released from the domain when the guide strand
forms a base pair with the target mRNA [11ꢀ13,21]. The nucleotides
of the 30-dangling end of the guide strand repeat binding to and
dissociation from the PAZ domain during RNAi.
Small interfering RNAs (siRNAs), which are composed of 21 or
22 nucleotides and contain a 2-nucleotide overhang at the 30-end,
are a type of small RNA. siRNAs are key molecules in RNA inter-
ference (RNAi), a cellular pathway for the control of gene expres-
sion [1ꢀ3]. During RNAi, a guide strand (an antisense strand) of
siRNA forms an RNA-induced silencing complex (RISC) with the
Argonaute (Ago) protein. The guide strand hybridizes with a target
mRNA in the RISC, and then the mRNA is cleaved by the slicer ac-
tivity of the RISC, leading to inhibition of protein expression of the
target mRNA. If the sequences of disease-causing genes are known,
the siRNA can be rationally designed and synthesized [2,4]. Thus,
siRNAs have attracted a great deal of attention as promising can-
didates for use as therapeutic agents for “unmet medical needs.” To
date, a variety of chemically modified siRNAs have been synthe-
sized to improve nuclease resistance and silencing abilities [5ꢀ10].
Ago proteins are composed of four domains: the N-terminal (N),
Recently, we reported the synthesis of a photo-cross-linking
microRNA (miRNA) probe consisting of the nucleoside analog 1-
O-[4-(3-trifluoromethyl-3H-diazirin-3-yl)]benzyl-b-D-ribofur-
anose (1), in which the natural base was replaced with an aryl
residue via an acetal linkage (Fig. 1) [22]. We considered that sub-
stituents of different sizes could be
1-position of -ribofuranose via an acetal linkage by using 1-O-
acetyl-2,3,5-tri-O-benzoyl- -ribofuranose as a starting material.
Based on these reports and results, we designed and synthesized
b-selectively introduced at the
D
* Corresponding author. Course of Applied Life Science, Faculty of Applied Bio-
logical Sciences, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan
E-mail address: uenoy@gifu-u.ac.jp (Y. Ueno).
b-D
http://dx.doi.org/10.1016/j.ejmech.2015.09.011
0223-5234/© 2015 Elsevier Masson SAS. All rights reserved.

N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
461
phosphitylated by a standard procedure to give the corresponding
phosphoramidite 24 with a 64% yield. In a similar manner, the
phosphoramidites 35e38, which had propyl, 2-methylpropyl, 2,2-
dimethylpropyl, and 2,2,2-trifluoroethyl groups, respectively,
were synthesized from the protected D-ribofuranose derivative 13
with final yields of 21%, 17%, 20%, and 13%, respectively.
To introduce the analog 6 compound at the 30-end of an RNA
strand, the 5-O-DMTre3-O-TBDMS derivative 32 was converted to
the corresponding succinate, which was then linked to controlled
pore glass (CPG) to generate the solid support 39 linked to com-
pound 32 (40 mmol/g) (Scheme 2). In a similar manner, the solid
support 40 linked to ethylene glycol was also synthesized.
Oligonucleotide (ON) sequences used in this study are listed in
Tables 1 and 2. In order to synthesize siRNAs with analogs at their
30-dangling ends, ONs 5e16, which had analogs 2e6 at their 30-end
moieties corresponding to the 30-dangling ends in the siRNAs, were
synthesized using solid support 40 as a universal solid support. All
ONs containing the analogs were synthesized using the phos-
phoramidites 34e38 and the solid support 39 or 40. After purifi-
cation via polyacrylamide gel electrophoresis (PAGE), the ONs were
analyzed by a reversed-phase C-18 high-performance liquid chro-
matography (HPLC). As shown in Fig. S1, a single peak was observed
in the profiles for ONs 11e16, while several peaks thought to be
attributable to the hydrolysis of the glycosidic bonds of the analogs
were observed in the profiles for ONs 7e10. These results suggested
that under the acidic conditions used in the ON synthesis, the
glycosidic bonds of the analogs with small substituents (e.g., ethyl
or propyl groups) were more unstable than the analogs with bulky
substituents (e.g., 2,2-dimethylpropyl group) or electron with-
drawing groups (e.g., 2,2,2-trifluoroethyl group). ONs 7e10 were
further purified by the HPLC.
Fig. 1. Structures of nucleoside analogs.
The ONs were analyzed by matrix-assisted laser desorption/
ionization time-of-flight mass spectrometry (MALDI-TOF/MS), and
the observed molecular weights from these analyses were in
agreement with the structures of the ONs.
five types of siRNA containing acetal-type nucleoside analogs at
their 30-dangling ends to investigate the influence of the 30-
dangling-end modifications on recognition of the siRNAs by the
human Argonaute2 PAZ domain protein.
In this paper, we report the synthesis of five types of siRNA
containing the acetal-type nucleoside analogs 2e6 at their 30-
dangling ends (Fig. 1). We assessed the properties of the modified
siRNAs by a thermal denaturation study, a dual-luciferase reporter
assay, an electrophoretic mobility shift assay (EMSA), an enzyme-
linked immunosorbent assay (ELISA), and a partial digestion us-
ing a 30 exonuclease.
2.2. Thermal stability of siRNAs
First, the stabilities of the RNA duplexes containing the analogs
at the 30-dangling ends were evaluated by thermal denaturation in
a buffer comprising 10 mM sodium phosphate (pH 7.0) and
100 mM NaCl (Fig. S2). Melting temperatures (Tms) and DTms (Tm
of each duplex e Tm of siRNA 1) are listed in Table 1. It was re-
ported that a base at a dangling end enhances a thermal stability of
an RNAeRNA duplex [23]. The Tm of the unmodified RNA duplex
(siRNA 1) was 78.3 ꢁC. The Tm of siRNA 2, which had thymidines
instead of uridines at the 30-dangling ends, was slightly lower than
2. Results
2.1. Synthesis
that of siRNA
1
(
D
Tm
¼
ꢀ0.5 ꢁC). Addition of the 2-
The synthetic route used to synthesize phosphoramidites
34e38 of analogs 2e6 is shown in Scheme 1. We considered that
hydroxylethylphosphate moiety at the 30-ends of the duplex
(siRNA 3) barely affected the thermal stability of the duplex. On
the other hand, incorporation of analog 2, which has an ethyl
group at the 30-dangling end, reduced the thermal stability of the
various substituents could be
position of -ribofuranose via neighboring group participation of
an acyl-type protective group at the 2-O-position. Glycosylation of
ethanol (8) with 1-O-acetyl-2,3,5-tri-O-benzoyl- -D-ribofuranose
b-selectively introduced at the 1-
D
duplex (
D
Tm ¼ ꢀ1.4 ꢁC). However, we found that the duplexes
b
became more thermally stable as the size of the side chains of the
analogs increased. The Tm value (78.2 ꢁC) of siRNA 7, incorporating
analog 5 (with the 2,2-dimethypropyl group), was almost the
same as that of siRNA 1 (with uridines) (Tm ¼ 78.3 ꢁC). Intrigu-
ingly, the Tm (77.9 ꢁC) of siRNA 8, which had a 2,2,2-trifluoroethyl
group, was greater than those of siRNA 4 (Tm ¼ 76.9 ꢁC), siRNA 5
(Tm ¼ 77.1 ꢁC), and siRNA 6 (Tm ¼ 77.0 ꢁC), which had ethyl,
propyl, and 2-methylpropyl groups, respectively. These results
suggest that dipoleedipole interactions as well as hydrophobic
interactions among the alkyl side chains and the adjacent bases in
the dangling-end positions are important for enhancing the
thermal stability of the duplexes.
(13), which is commercially available, in the presence of TMSOTf
at ꢀ30 ꢁC in CH2Cl2 quantitatively afforded the
b-anomer 14. Ste-
reochemistry of the anomeric carbon was determined from the
coupling constant between the protons at the 1-position and 2-
position. Subsequently, this compound was debenzoylated in the
presence of a catalytic amount of NaOCH3 in CH3OH, and the pri-
mary hydroxy group of 2 was protected by a 4,40-dimethoxytrityl
(DMTr) group to produce the 5-O-DMTr derivative 19 with a 77%
yield. Compound 19 was treated with TBDMSCl to afford the 3-O-
TBDMS and 2-O-TBDMS derivatives 24 and 25 with 42% and 33%
yields, respectively. The 2-O-TBDMS derivative 25 was

462
N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
Scheme 1. The experimental procedure is described in Experimental Section.
2.3. Silencing activities of siRNAs
the size of the side chains of the analogs increased. Intriguingly, the
silencing activity of siRNA 8 (21.3 and 57.4 at 1.0 and 0.1 nM,
respectively), which had a 2,2,2-trifluoroethyl group, was greater
than that of siRNA 3, which had an ethyl group, and comparable to
that of siRNA 1, which had uridines at the dangling ends, despite the
size of the 2,2,2-trifluoroethyl group being similar to that of the
ethyl group.
Next, we examined the ability of the siRNAs containing the
analogs to suppress gene expression by using a dual-luciferase re-
porter assay. The psiCHECK-2 vector, which encodes Renilla lucif-
erase and firefly luciferase, was used as the reporter. All siRNAs
were designed to target the Renilla luciferase gene. The percent
Renilla: firefly luciferase activities compared to transfection
without siRNA are shown in Table 1. The luciferase activity ratios for
an unmodified siRNA (siRNA 1) with uridines at the dangling ends
were 24.1 and 53.4. Replacement of the uridines at the dangling
ends by thymidines slightly reduced the silencing activity of the
siRNA (siRNA 2; 28.9 and 68.0 at 1.0 and 0.1 nM, respectively).
Addition of the 2-hydroxylethylphosphate at the 30-ends of the
siRNA also slightly decreased the silencing activity of the siRNA
(siRNA 3; 34.3 and 69.1 at 1.0 and 0.1 nM, respectively). Although
the silencing activities of siRNAs 4e7, which had analogs at the
dangling ends, were not largely different from each other, the siR-
NAs, except for siRNAs 5 and 8, tended to become more potent as
2.4. Ability of the human Ago2 PAZ domain to bind to modified
siRNAs
Next, we evaluated the ability of a recombinant human Ago2
PAZ domain protein to bind to the modified siRNAs by performing
an EMSA. The sequences of the siRNAs used in the EMSA are listed
in Table 2. AllsiRNAshad dangling endsat the 30-ends of the sense
strands, whereas the 30-ends of the antisense strands had blunt-
end structures. The 50-OH of the sense strand (ON 17) of siRNA 9
was labeled with fluorescein. We inferred that the PAZ domain
protein would recognize and bind to the dangling-end moieties of
Scheme 2. The experimental procedure is described in Experimental Section.

N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
463
Table 1
Sequences of siRNAs, their Tm values and their abilities to suppress gene expression.
Abbreviation of siRNA
Abbreviation of ON
Sequencea
Tm (ꢁC)
D
Tm (ꢁC)b
Upper: 1 nMc
lower:0.1 nM
Control
siRNA 1
(Buffer)
ON 1
e
e
e
e
100 1.2
24.1 1.6
53.4 4.6
Sense (passenger) strand
78.3
50-GGCCUUUCACUACUCCUACUU-30
30-UUCCGGAAAGUGAUGAGGAUG-50
Antisense (guide) strand
ON 2
siRNA 2
siRNA 3
siRNA 4
siRNA 5
siRNA 6
siRNA 7
siRNA 8
ON 3
ON 4
ON 5
ON 6
ON 7
ON 8
ON 9
ON 10
ON 11
ON 12
ON 13
ON 14
ON 15
ON 16
50-GGCCUUUCACUACUCCUACtt-30
30-ttCCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUACtt7-30
30-7ttCCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC227-30
30-722CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC337-30
30-733CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC447-30
30-744CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC557-30
30-755CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC667-30
30-766CCGGAAAGUGAUGAGGAUG-50
77.8
77.8
76.9
77.1
77.0
78.2
77.9
ꢀ0.5
ꢀ0.5
ꢀ1.4
ꢀ1.2
ꢀ1.3
ꢀ0.1
ꢀ0.4
28.9 3.3
68.0 1.7
34.3 1.9
69.1 1.8
35.5 2.9
71.4 3.1
24.4 1.2
60.0 5.5
31.7 3.5
61.9 1.0
27.8 0.7
59.4 3.7
21.3 1.9
57.4 0.9
a
Small letters indicate 20-deoxyribonucleosides.
b
c
DTm ¼ Tm (each siRNA) e Tm (siRNA 1).
Activities of firefly and Renilla luciferases in cell lysates were determined with a dual-luciferase assay system (Promega) according to a manufacturer's protocol. The results
were confirmed by at least three independent transfection experiments with two cultures each and are expressed as the average from four experiments as mean SD.
the siRNAs. First, the labeled siRNA 9 was incubated with various
concentrations of the recombinant human Ago2 PAZ domain pro-
tein at 37 ꢁC in a buffer of 20 mM TriseHCl (pH 8.0), 0.1 M NaCl, and
1 mM dithiothreitol (DTT), and the mixtures were analyzed by
nondenaturing polyacrylamide gel electrophoresis (PAGE) at 4 ꢁC.
As shown in Fig. 2, lower-mobility bands corresponding to the
siRNA-protein complexes were observed. The dissociation constant
which had thymidines at the 30-dangling end, on the formation of
the labeled siRNA-protein complex was significantly weaker than
that of siRNA 10, which had uridines at the 30-dangling end. This
result is consistent with previous reports [20,24] and implies that
the 20-OH groups or the sugar conformation of the nucleosides at
the dangling end is important for recognition of the siRNA by the
PAZ domain protein. The inhibitory ability of siRNA 12, which had
2-hydroxylethylphosphate at the 30-OH of the 30-dangling end, was
not largely different from that of siRNA 11. This result indicates that
the 2-hydroxylethylphosphate moiety does not significantly influ-
ence the binding of the PAZ domain protein to the siRNA. The
inhibitory effects of siRNAs 13e16, with their respective analogs, on
the formation of the labeled siRNA-protein complex were similar to
that of siRNA 12, which had thymidines at the 30-dangling end.
Although the inhibitory activities of siRNAs 13e16 were not largely
different from each other, the siRNAs tended to become more
(Kd) was estimated to be about 20 mM.
Next, we carried out the EMSA in the presence of siRNAs 10e17
as competitors. As shown in Fig. 3, formation of the siRNA-protein
complex between the labeled siRNA 9 and the PAZ domain protein
was severely inhibited by adding the unlabeled siRNA 10. The ratio
of the siRNA-protein complex between the labeled siRNA 9 and the
PAZ domain protein decreased from 53% to 28% or 14% in the
presence of a 10- or 100-fold molar excess, respectively, of the
unlabeled siRNA 10. In contrast, the inhibitory effect of siRNA 11,
Table 2
Sequences of siRNAs used for EMSA.
Abbreviation of siRNA
Abbreviation of ON
Sequencea
siRNA 9
ON 17
Sense (passenger) strand
F-50-GGCCUUUCACUACUCCUACUU-30
30-CCGGAAAGUGAUGAGGAUG-50
Antisense (Guide) strand
ON 18
siRNA 10
siRNA 11
siRNA 12
siRNA 13
siRNA 14
siRNA 15
siRNA 16
siRNA 17
ON 1
ON 18
ON 3
ON 18
ON 5
ON 18
ON 7
50-GGCCUUUCACUACUCCUACUU-30
30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUACtt-30
30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUACtt7-30
30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC227-30
30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC337-30
30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC447-30
30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC557-30
30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC667-30
30-CCGGAAAGUGAUGAGGAUG-50
ON 18
ON 9
ON 18
ON 11
ON 18
ON 13
ON 18
ON 15
ON 18
a
Small letters indicate 20-deoxyribonucleosides. F denotes a fluorescin residue.

464
N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
protein to siRNA 21, which had an ethyl group, was slightly weaker
than that to siRNA 24, which had a 2,2-dimethylpropyl group, and
was similar to that to siRNA 25, which had a 2,2,2-trifluoroethyl
group.
2.6. Nuclease resistance properties of modified RNAs
As siRNA is readily hydrolyzed by ubiquitous nucleases existing
inside and outside cells, improving the nuclease resistance of siRNA
is important for therapeutic applications of synthetic siRNA.
Therefore, we next investigated the nuclease resistance properties
of the modified ONs by using snake venom phosphodiesterase
(SVPD), a 30 exonuclease. Unmodified ONs 17 and 20 and modified
ONs 21e24, which were labeled with fluorescein at the 50-ends,
were incubated with SVPD. The reactions were analyzed using
PAGE under denaturing conditions. As shown in Fig. 5, unmodified
ONs 17 and 20 were randomly hydrolyzed after 1 min of incubation,
whereas modified ONs 21e24 were resistant to the enzyme. The
half-lives (t1/2) of unmodified ONs 17 and 20 were <1 min, while
those of modified ONs 21e24 were 14, 12, 12, and 11 min, respec-
tively. On the other hand, the stability of the siRNA modified with
analog 5 at the 30-dangling end in 40% bovine serum was the same
as that of the unmodified siRNA (data not shown).
Fig. 2. EMSA to detect binding of siRNA 9 to various concentrations of the recombinant
human Ago2 PAZ domain protein. Experimental conditions are described in the
Experimental Section.
potent as the size of the side chains of the analogs increased. The
inhibitory activity of siRNA 17, which had the 2,2,2-trifluoroethyl
group, was similar to that of siRNA 13, which had an ethyl group.
2.5. Enzyme-linked immunosorbent assay
Furthermore, we performed ELISA to examine the ability of the
human Ago2 PAZ domain to bind to modified siRNAs, the sequences
of which are listed in Table 3. All siRNAs had biotin residues at the
30-end of the antisense strands, and they were immobilized on the
streptavidin-coated 96-well microplate. The human Ago2 PAZ
domain protein was then added to each well. The protein, which
bound to the siRNAs, was detected by an anti-His tag monoclonal
antibody conjugated with horseradish peroxidase (HRP). The ac-
tivity of HRP was evaluated by an oxidation reaction using 3,30,5,50-
tetramethylbenzidine (TMB) as a substrate. The absorbance of each
well at 450 nm is shown in Fig. 4. A trend similar to that noted in
the EMSA was observed. The binding ability of the human Ago2 PAZ
domain protein to siRNA 19, which had thymidines at the 30-
dangling end of the sense strand, was weaker than that to siRNA 18,
which had uridines at the 30-dangling end. Binding ability of the
3. Discussion
In order to create novel siRNAs that are more potent and
nuclease resistant than natural unmodified siRNAs, we designed
and synthesized five types of siRNA containing the acetal-type
nucleoside analogs 2e6 at their 30-dangling ends. We assessed
the properties of the modified siRNAs by Tm measurement, a dual-
luciferase reporter assay, an EMSA, an ELISA, and a partial digestion
using SVPD. The binding affinities of the human Ago2 PAZ domain
protein for the modified siRNAs tended to increase with increase in
the size of the side chains of the analogs, although they were
apparently weaker than that for the siRNA with uridines at its 30-
Fig. 3. EMSA in the presence of competitors (siRNAs 10e17). Experimental conditions are described in the Experimental Section.

N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
465
Table 3
Sequences of siRNAs used for ELISA.
Abbreviation of siRNA
Abbreviation of ON
Sequencea
siRNA 18
ON 1
Sense (passenger) strand
50-GGCCUUUCACUACUCCUACUU-30
Bio-30-CCGGAAAGUGAUGAGGAUG-50
Antisense (Guide) strand
ON 19
siRNA 19
siRNA 20
siRNA 21
siRNA 22
siRNA 23
siRNA 24
siRNA 25
ON 3
ON 19
ON 5
ON 19
ON 7
ON 19
ON 9
ON 19
ON 11
ON 19
ON 13
ON 19
ON 15
ON 19
50-GGCCUUUCACUACUCCUACtt-30
Bio-30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUACtt7-30
Bio-30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC227-30
Bio-30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC337-30
Bio-30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC447-30
Bio-30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC557-30
Bio-30-CCGGAAAGUGAUGAGGAUG-50
50-GGCCUUUCACUACUCCUAC667-30
Bio-30-CCGGAAAGUGAUGAGGAUG-50
a
Bio denotes a biotin residue.
dangling end. On the other hand, for siRNAs 4, 6, and 7, the silencing
activities of the siRNAs increased with increase in the size of the
side chains of the analogs, but siRNAs 5 and 8 were more potent
than siRNAs 4, 6, and 7, despite the sizes of the side chains of an-
alogs 3 and 6 being smaller than those of analogs 4 and 5. These
results suggest that although the 30-dangling-end moiety is
important for enhancing the silencing activity of the siRNA, a
higher binding affinity of the 30-dangling-end moiety for the PAZ
domain protein is not always positively correlated with a higher
silencing efficacy of the siRNA. It was suggested by recent studies
that the nucleotides of the 30-end of the guide strand are not always
anchored in the PAZ domain in the Ago protein and released from
the domain when the guide strand forms a base pair with the target
mRNA [11e13,21]. Thus, enhancing the dissociation of the 30-
dangling-end nucleosides from the PAZ domain might also be
important for improving the silencing activity of siRNA.
We found that the silencing activity of siRNA 8, which had a
2,2,2-trifluoroethyl group, was greater than that of siRNA 3, which
had an ethyl group, and comparable to that of siRNA 1, which had
uridines at the dangling ends, despite the size of the 2,2,2-
trifluoroethyl group being similar to that of the ethyl group. It
was previously reported that thermodynamic features of siRNAs
influence the silencing activity of siRNAs [25ꢀ29]. The Tm value of
siRNA 8, which had a 2,2,2-trifluoroethyl group, was greater than
those of siRNAs 4e6, which had ethyl, propyl, and 2-methylpropyl
groups, respectively. Thus, this high thermal stability of siRNA 8
might be advantageous for eliciting a high silencing activity.
Furthermore, ON 24, which had an analog containing the 2,2,2-
trifluoroethyl group, was more resistant to SVPD, a 30 exonu-
clease, than unmodified ONs.
4. Conclusion
We have demonstrated the synthesis of siRNAs containing
acetal-type nucleoside analogs at their 30-ends. Several groups
including our group have previously reported the synthesis of
siRNAs containing aromatic compounds at their 30-ends
[7,8,30ꢀ32]. Bulky compounds such as pyrene and phenanthrene
decreased the silencing activities of the siRNAs owing to steric
hindrance or a strong hydrophobic interaction with the PAZ domain
[7,30]. Introduction of fluorine atoms into the aromatic compounds
enhanced the silencing activities of the siRNAs [30,32]. The results
indicate that the hydrogen bonds existing between the fluorine
atoms and the amino acids in the hydrophobic pocket of the PAZ
domain are favorable for eliciting RNA interference. In this study,
we also found that the silencing activity of siRNA with an analog
containing the 2,2,2-trifluoroethyl group was greater than that of
siRNAs with analogs containing other groups and comparable to
that of siRNAs with natural uridines at their dangling ends,
although a significant increase in the binding affinity of siRNA
containing the 2,2,2-trifluoroethyl-modified nucleoside analog to
the PAZ domain was not observed by EMSA and ELISA. Further-
more, siRNA with an analog containing the 2,2,2-trifluoroethyl
group at its 30-end was more resistant to hydrolytic degradation
by a 30 exonuclease than a natural RNA was. Thus, modification of
siRNAs by a nucleoside analog containing a 2,2,2-trifluoroethyl
group may hold promise as a means of improving the silencing
activity and nuclease resistance of siRNAs.
4.1. Experimental Section
4.1.1. General remarks
CDCl3 (CIL) or DMSO-d6 (CIL) was used as a solvent for obtaining
NMR spectra. Chemical shifts (d) are given in parts per million
(ppm) downfield from (CH3)4Si (d
0.00 for 1H NMR in CDCl3), or a
Fig. 4. ELISA to detect binding of siRNAs to the recombinant human Ago2 PAZ domain
protein. Bio denotes a biotin residue. Experimental conditions are described in the
Experimental Section.
solvent (for 13C NMR and 1H NMR in DMSO-d6) as an internal
reference with coupling constants (J) in Hz. The abbreviations s, d,

466
N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
Fig. 5. 20% PAGE of ONs hydrolyzed by SVPD. F denotes a fluorescin. The sequence of the RNA part is 50-GGCCUUUCACUACUCCUAC-30. Experimental conditions are described in the
Experimental Section.
and q signify singlet, doublet, and quartet, respectively.
4.1.3. 1-O-Ethyl-b-D-ribofuranose (2)
To a stirred solution of 14 (3.05 g, 6.22 mmol) in MeOH (30 mL)
was added a catalytic amount of a 28% MeOH solution of sodium
methoxide at room temperature. The mixture was stirred at room
temperature for 12 h and quenched by using aqueous NH4Cl
(saturated) (1 mL). The solvent was evaporated in vacuo, and the
resulting residue was purified by column chromatography (SiO2, 9%
MeOH in CHCl3) to give 2 (1.11 g, 6.22 mmol, 100%): 1H NMR
(400 MHz, CDCl3) 1.20 (t, 3H, J ¼ 7.1), 2.31 (s, 1H), 3.12 (s, 1H), 3.52
(dq,1H, J ¼ 7.4 and 8.3), 3.67 (dd,1H, J ¼ 5.0 and 11.4), 3.73e3.79 (m,
3H), 4.03e4.06 (m, 2H), 4.30 (d, 1H, J ¼ 5.5) 4.95 (s, 1H); 13C NMR
4.1.2. 2,3,5-O-Tribenzoyl-1-O-ethyl-
b
-
D
-ribofuranose (14)
To a solution of 1-O-acethyl-2,3,5-O-tribenzoyl-
b
-D-ribofur-
anose (3.00 g, 5.95 mmol) in CH2Cl2 (30 mL) was added dropwise
TMSOTf (1.29 mL, 7.14 mmol) and ethanol (0.35 mL, 5.95 mmol) at
e30 ꢁC. The mixture was stirred at ꢀ30 ꢁC for 1.5 h and quenched
by using aqueous NaHCO3 (saturated). The mixture was partitioned
between CHCl3 and H2O. The organic layer was washed with
aqueous NaHCO3 (saturated) and brine, dried (Na2SO4), and
concentrated. The residue was purified by column chromatography
(SiO2, 20% EtOAc in hexane) to give 14 (2.91 g, 5.94 mmol, 100%): 1H
(101 MHz, CDCl3)
d 15.2, 63.3, 64.3, 71.4, 75.7, 84.3, 107.4. Anal.
Calcd for C7H14O5$3/5H2O: C, 44.49; H, 8.11. Found: C, 44.56; H,
8.02.
NMR (400 MHz, CDCl3)
d
1.18 (t, 3H, J ¼ 7.1), 3.54 (dq, 1H, J ¼ 6.8 and
8.5), 3.84 (dq, 1H, J ¼ 6.9 and 11.9), 4.52 (dd, 1H, J ¼ 6.4 and 12.8),
4.66e4.74 (m, 2H), 5.26 (s, 1H), 5.67 (d, 1H, J ¼ 4.6), 5.88 (dd, 1H,
J ¼ 5.0 and 6.4), 7.25e8.11 (m, 15H); 13C NMR (126 MHz, CDCl3)
4.1.4. 5-O-(4,40-Dimethoxytrityl)-1-O-ethyl-
b-D-ribofuranose (19)
To a solution of 2 (0.97 g, 5.43 mmol) in pyridine (10 mL) was
added DMTrCl (2.21 g, 6.52 mmol) at room temperature. The
mixture was stirred at room temperature for 5 h and partitioned
d
15.0, 64.0, 65.0, 72.7, 75.8, 78.9, 105.4, 128.4, 128.6, 129.1, 129.4,
129.8, 129.9, 133.2, 133.4, 133.5, 165.4, 165.5, 166.3. Anal. Calcd for
28H26O8$7/10H2O: C, 66.84; H, 5.49. Found: C, 66.86; H, 5.19.
C

N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
467
between EtOAc and aqueous NaHCO3 (saturated). The organic layer
was washed with brine, dried (Na2SO4), and concentrated. The
residue was purified by column chromatography (SiO2, 50% EtOAc
in hexane) to give 19 (2.01 g, 4.19 mmol, 77%): 1H NMR (400 MHz,
NMR (600 MHz, CDCl3)
d
0.90 (t, 3H, J ¼ 7.6), 1.55e1.59 (m, 2H), 3.42
(dd, 1H, J ¼ 6.9 and 15.8), 3.74 (dd, 1H, J ¼ 6.8 and 14.4), 4.52e4.54
(m, 1H), 4.69e4.72 (m, 2H), 5.25 (s, 1H), 5.68 (d, 1H, J ¼ 4.8), 5.87
(dd, 1H, J ¼ 2.8 and 5.0), 7.31e8.07 (m, 15H); 13C NMR (151 MHz,
CDCl3)
d
1.11 (t, 3H, J ¼ 6.9), 2.26 (d, 1H, J ¼ 5.5), 2.48 (d, 1H, J ¼ 3.2),
CDCl3) d 10.5, 14.2, 21.0, 22.6, 60.4, 65.1, 70.1, 72.7, 75.6, 78.7, 105.5,
3.27 (dd, 2H, J ¼ 3.2 and 5.5), 3.44 (dq, 1H, J ¼ 6.9 and 8.3), 3.73 (dq,
1H, J ¼ 7.3 and 8.2), 3.80 (s, 6H), 4.04e4.08 (m, 2H), 4.27 (dd, 1H,
J ¼ 5.28 and 11.2), 4.96 (s, 1H), 6.81e7.48 (m, 13H); 13C NMR
128.3, 128.5, 129.0, 129.3, 129.7, 129.8, 129.8, 133.1, 133.3, 133.4.
Anal. Calcd for C29H28O8: C, 69.04; H, 5.59. Found: C, 68.80; H, 5.58.
(151 MHz, CDCl3)
d
15.0, 55.2, 63.4, 65.0, 72.9, 75.4, 81.9, 86.1, 106.9,
4.1.8. 1-O-Propyl-b-D-ribofuranose (3)
113.1, 126.8, 127.8, 128.2, 130.0, 136.1, 144.8, 158.5. Anal. Calcd for
28H32O7$9/10H2O: C, 67.70; H, 6.86. Found: C, 67.50; H, 6.56.
To a stirred solution of 15 (2.87 g, 5.69 mmol) in MeOH (29 mL)
was added a catalytic amount of a 28% MeOH solution of sodium
methoxide at room temperature. The mixture was stirred at room
temperature for 12 h and quenched by using aqueous NH4Cl
(saturated) (1 mL). The solvent was evaporated in vacuo, and the
resulting residue was purified by column chromatography (SiO2, 9%
MeOH in CHCl3) to give 3 (1.09 g, 5.69 mmol, 100%): 1H NMR
C
4.1.5. 3-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-
O-ethyl- -ribofuranose (24) and 2-O-(tert-butyldimethylsilyl)-5-
O-(4,40-dimethoxytrityl)-1-O-ethyl-
-ribofuranose (25)
b-D
b-D
To a stirred solution of 19 (1.39 g, 2.89 mmol) and Et3N (1.21 mL,
8.67 mmol) in DMF (8 mL) was added dropwise TBDMSCl (0.87 g,
5.78 mmol) in DMF (6 mL) at room temperature. The mixture was
stirred at room temperature for 12 h and partitioned between
EtOAc and aqueous NaHCO3 (saturated). The organic layer was
washed with brine, dried (Na2SO4), and concentrated. The residue
was purified by column chromatography (SiO2, 20% EtOAc in hex-
ane) to give 24 (0.73 g, 1.22 mmol, 42%) and 25 (0.56 g, 0.94 mmol,
(600 MHz, CDCl3)
d
0.92 (t, 3H, J ¼ 7.6), 1.57e1.63 (m, 2H), 2.42 (dd,
1H, J ¼ 3.8 and 8.6), 2.77 (d,1H, J ¼ 6.2), 2.91 (d,1H, J ¼ 3.4), 3.42 (dt,
1H, J ¼ 6.9 and 9.6), 3.65e3.70 (m, 2H), 3.81 (dd, 1H, J ¼ 8.2 and
11.7), 4.07e4.09 (m, 2H), 4.39 (d, 1H, J ¼ 5.5), 4.96 (s, 1H); 13C NMR
(151 MHz, CDCl3) d 10.5, 22.8, 63.5, 70.2, 71.6, 75.5, 83.8, 107.4. Anal.
Calcd for C8H16O5$1/2H2O: C, 47.75; H, 8.52. Found: C, 47.66; H,
8.33.
33%). Physical data of 24: 1H NMR (400 MHz, CDCl3)
d
ꢀ0.13 (s,
3H), ꢀ0.00 (s, 3H), 0.81 (s, 9H), 1.15 (t, 3H, J ¼ 7.1), 2.73 (d, 1H,
J ¼ 2.3), 3.03 (dd,1H, J ¼ 5.3 and 10.3), 3.31 (d,1H, J ¼ 10.1), 3.53 (dd,
1H, J ¼ 7.4 and 16.5), 3.78e3.89 (m, 7H), 3.89 (s, 1H), 4.05 (s, 1H),
4.1.9. 5-O-(4,40-Dimethoxy)trityl-1-O-propyl-
b-D-ribofuranose (20)
To a solution of 3 (0.65 g, 3.38 mmol) in pyridine (7 mL) was
added DMTrCl (1.58 g, 4.06 mmol) at room temperature. The
mixture was stirred at room temperature for 1 h and partitioned
between EtOAc and aqueous NaHCO3 (saturated). The organic layer
was washed with brine, dried (Na2SO4), and concentrated. The
residue was purified by column chromatography (SiO2, 33% EtOAc
in hexane) to give 20 (1.35 g, 2.80 mmol, 83%): 1H NMR (500 MHz,
4.33 (dd, 1H, J ¼ 6.4 and 11.0), 5.03 (s, 1H), 6.80e7.51 (m, 13H); 13
C
NMR (126 MHz, CDCl3)
d
ꢀ4.85, ꢀ4.82, 15.2, 18.0, 25.7, 55.3, 63.6,
64.0, 72.5, 75.5, 82.6, 85.9, 107.0, 113.1, 126.7, 127.8, 128.4, 130.2,
136.3, 136.4, 145.0, 158.5. Anal. Calcd for C34H46O7Si$1/2H2O: C,
67.63; H, 7.85. Found: C, 67.43; H, 7.72. Physical data of 25: 1H NMR
(400 MHz, CDCl3)
d
0.14 (t, 6H, J ¼ 2.3), 0.93 (s, 9H), 1.15 (t, 3H,
CDCl3)
d
0.73 (t, 3H, J ¼ 7.5), 1.37e1.44 (m, 2H), 2.31 (d, 1H, J ¼ 5.7),
J ¼ 7.1), 2.49 (d, 1H, J ¼ 7.4), 3.10 (dd, 1H, J ¼ 5.0 and 9.9), 3.31 (dd,
1H, J ¼ 3.2 and 10.1), 3.48 (dq, 1H, J ¼ 6.8 and 11.9), 3.78e3.86 (m,
7H), 4.07e4.10 (m, 2H), 4.13 (d, 1H, J ¼ 1.8), 4.90 (s, 1H), 6.80e7.51
2.56 (d, 1H, J ¼ 3.5), 3.15 (dd, 1H, J ¼ 5.7 and 9.8), 3.19e3.26 (m, 2H),
3.53 (dt, 1H, J ¼ 6.9 and 9.8), 3.70 (s, 6H), 3.95e4.03 (m, 2H), 4.16
(dd, 1H, J ¼ 5.7 and 10.9), 4.86 (s, 1H), 6.73e7.39 (m, 13H); 13C NMR
(m, 13H); 13C NMR (126 MHz, CDCl3)
d
ꢀ4.9, ꢀ4.5, 0.1, 15.2, 18.2,
(126 MHz, CDCl3) d 10.6, 22.8, 55.3, 65.3, 69.7, 73.2, 75.4, 82.0, 86.2,
25.8, 55.3, 63.8, 64.9, 72.4, 76.7, 83.5, 86.0, 107.4, 113.1, 126.7, 127.8,
128.4, 130.2, 136.3, 136.4, 145.1, 158.5. Anal. Calcd for
107.2,113.2,126.9,127.9,128.2,130.1,130.2,136.1,136.2,144.9,158.6.
Anal. Calcd for C28H32O7$4/5H2O: C, 68.43; H, 7.05. Found: C, 68.14;
H, 6.83.
C
34H46O7Si$1.0H2O: C, 66.64; H, 7.90. Found: C, 66.61; H, 7.61.
4.1.6. 2-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-
O-ethyl-3-O-[(2-cyanoethoxy)(N,N-diisopropyamino)]phosphanyl-
4.1.10. 3-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-
O-propyl-
b
-
D
-ribofuranose (26) and 2-O-(tert-butyldimethylsilyl)-
b
-D-ribofuranose (34)
5-O-(4,40-dimethoxytrityl)-1-O-propyl-
b-D
-ribofuranose (27)
To a solution of 25 (0.74 g, 1.25 mmol) in THF (8 mL) was added
N,N-diisopropylethylamine (1.09 mL, 6.25 mmol) and chloro(2-
cyanoethoxy)(N,N-diisopropylamino)phosphine (0.56 mL,
2.50 mmol) at room temperature. The mixture was stirred at room
temperature for 1 h and partitioned between CHCl3 and aqueous
NaHCO3 (saturated). The organic layer was washed with brine,
dried (Na2SO4), and concentrated. The residue was purified by
column chromatography (a neutralized SiO2, 33% EtOAc in hexane)
to give 34 (0.64 g, 0.80 mmol, 64%): 31P NMR (162 MHz, CDCl3)
To a stirred solution of 20 (1.70 g, 3.44 mmol) and Et3N (1.40 mL,
10.32 mmol) in DMF (20 mL) was added dropwise TBDMSCl (1.04 g,
6.88 mmol) in DMF (15 mL) at room temperature. The mixture was
stirred at room temperature for 12 h and partitioned between
EtOAc and aqueous NaHCO3 (saturated). The organic layer was
washed with brine, dried (Na2SO4), and concentrated. The residue
was purified by column chromatography (SiO2, 20e25% EtOAc in
hexane) to give 26 (0.90 g, 1.48 mmol, 43%) and 27 (0.77 g,
1.27 mmol, 37%). Physical data of 26: 1H NMR (400 MHz, CDCl3)
d
149.1, 149.7.
d
ꢀ0.13 (s, 3H), ꢀ0.00 (s, 3H), 0.81 (s, 9H), 2.74 (d, 1H, J ¼ 2.3), 3.03
(dd, 1H, J ¼ 5.0 and 10.1), 3.29 (dd, 1H, J ¼ 3.2 and 10.1), 3.40 (dt, 1H,
J ¼ 6.4 and 9.2), 3.73 (dt, 1H, J ¼ 6.9 and 9.6), 3.78 (s, 6H), 3.90 (s,
1H), 4.06e4.07 (m, 1H), 4.31 (dd, 1H, J ¼ 4.6 and 6.4), 5.02 (s, 1H),
4.1.7. 2,3,5-O-Tribenzoyl-1-O-propyl-
b-D-ribofuranose (15)
To solution of 1-O-acethyl-2,3,5-O-tribenzoyl-b-D-ribofur-
a
anose (3.00 g, 5.95 mmol) in CH2Cl2 (30 mL) was added dropwise
TMSOTf (1.29 mL, 7.14 mmol) and 1-propanol (0.53 mL, 7.14 mmol)
at ꢀ30 ꢁC. The mixture was stirred at ꢀ30 ꢁC for 30 min and
quenched by using aqueous NaHCO3 (saturated). The mixture was
partitioned between CHCl3 and H2O. The organic layer was washed
with aqueous NaHCO3 (saturated) and brine, dried (Na2SO4), and
concentrated. The residue was purified by column chromatography
(SiO2, 20% EtOAc in hexane) to give 15 (2.94 g, 5.83 mmol, 98%): 1H
6.80e7.51 (m, 13H); 13C NMR (151 MHz, CDCl3)
d
ꢀ5.0, ꢀ4.9, 10.5,
17.9, 22.7, 25.6, 55.2, 64.1, 69.8, 72.5, 75.4, 82.6, 85.8, 107.2, 113.0,
126.6, 127.7, 128.3, 130.1, 136.2, 136.2, 144.9, 158.4. Anal. Calcd for
C
35H48O7Si$1/2H2O: C, 68.04; H, 7.99. Found: C, 67.82; H, 7.76.
Physical data of 27: 1H NMR (600 MHz, CDCl3)
d
0.12 (t, 6H, J ¼ 7.2),
0.85 (t, 3H, J ¼ 7.6), 0.91 (s, 9H), 1.49e1.55 (m, 2H), 2.50 (d, 1H,
J ¼ 6.9), 3.07 (dd,1H, J ¼ 4.8 and 10.3), 3.28 (dd,1H, J ¼ 3.4 and 10.3),
3.34 (dt, 1H, J ¼ 6.8 and 9.6), 3.69 (dt, 1H, J ¼ 6.2 and 8.9), 3.76 (s,

468
N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
6H), 4.01e4.07 (m, 2H), 4.13 (dd, 1H, J ¼ 2.0 and 4.8), 4.88 (s, 1H),
6.78e7.49 (m, 13H); 13C NMR (151 MHz, CDCl3)
22.6, 27.3, 30.3, 59.7, 69.4, 74.5, 77.0, 81.1, 88.0, 90.5, 112.1, 117.5,
131.1, 132.2, 132.8, 134.6, 134.7, 140.7, 140.8, 149.5, 162.9. Anal. Calcd
for C35H48O7Si$4/5H2O: C, 67.45; H, 8.02. Found: C, 67.33; H, 7.80.
CDCl3)
d
: 0.77 (d, 3H, J ¼ 6.9), 0.81 (d, 3H, J ¼ 6.4),1.69e1.76 (m, 1H),
d
ꢀ0.5, ꢀ0.1, 15.1,
2.29 (d,1H, J ¼ 5.5), 2.49 (d,1H, J ¼ 3.7), 3.11 (dd,1H, J ¼ 6.9 and 8.7),
3.20 (dd, 1H, J ¼ 5.5 and 9.2), 3.32 (dd, 1H, J ¼ 5.5 and 9.6), 3.41 (dd,
1H, J ¼ 6.9 and 9.2), 3.79 (s, 6H), 4.05e4.10 (m, 2H), 4.23 (dd, 1H,
J ¼ 5.48 and 11.0), 4.93 (s, 1H), 6.81e7.47 (m, 13H); 13C NMR
(151 MHz, CDCl3)
d 19.2, 28.2, 55.2, 65.4, 73.3, 74.6, 75.3, 81.8, 86.2,
4.1.11. 2-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-
O-propyl-3-O-[(2-cyanoethoxy)(N,N-diisopropyamino)]
phosphanyl-b-D-ribofuranose (35)
107.2, 113.1, 126.8, 127.8, 128.1, 130.0, 130.1, 136.0, 136.1, 144.8, 158.5.
Anal. Calcd for C30H36O7$1/2H2O: C, 69.61; H, 7.21. Found: C, 69.39;
H, 7.19.
To a solution of 27 (0.62 g,1.03 mmol) in THF (6.5 mL) was added
N,N-diisopropylethylamine (0.90 mL, 5.15 mmol) and chloro(2-
4.1.15. 3-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-
cyanoethoxy)(N,N-diisopropylamino)phosphine
(0.46
mL,
O-(2-methylpropyl)-
butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-O-(2-
methylpropyl)- -ribofuranose (29)
b-D-ribofuranose (28) and 2-O-(tert-
2.06 mmol) at room temperature. The mixture was stirred at room
temperature for 1 h and partitioned between CHCl3 and aqueous
NaHCO3 (saturated). The organic layer was washed with brine,
dried (Na2SO4), and concentrated. The residue was purified by
column chromatography (a neutralized SiO2, 25% EtOAc in hexane)
to give 35 (0.59 g, 0.73 mmol, 71%): 31P NMR (162 MHz, CDCl3)
b-D
To a stirred solution of 21 (1.29 g, 2.54 mmol) and Et3N (1.10 mL,
7.62 mmol) in DMF (12 mL) was added dropwise TBDMSCl (0.77 g,
5.08 mmol) in DMF (3 mL) at room temperature. The mixture was
stirred at room temperature for 12 h and partitioned between
EtOAc and aqueous NaHCO3 (saturated). The organic layer was
washed with brine, dried (Na2SO4), and concentrated. The residue
was purified by column chromatography (SiO2, 11% EtOAc in hex-
ane) to give 28 (0.53 g, 0.85 mmol, 34%) and 29 (0.53 g, 0.85 mmol,
d
149.4, 149.8.
4.1.12. 2,3,5-O-Tribenzoyl-1-O-(2-methylpropyl)-
b
-D
-ribofuranose
-ribofur-
(16)
To a solution of 1-O-acethyl-2,3,5-O-tribenzoyl-
b
-D
34%). Physical data of 28: 1H NMR (400 MHz, CDCl3)
d
ꢀ0.12 (s, 3H),
anose (2.00 g, 3.97 mmol) in CH2Cl2 (20 mL) was added dropwise
TMSOTf (0.86 mL, 4.76 mmol) and 2-methyl-1-propanol (0.37 mL,
3.97 mmol) at ꢀ30 ꢁC. The mixture was stirred at e30 ꢁC for 2.5 h
and quenched by using aqueous NaHCO3 (saturated). The mixture
was partitioned between CHCl3 and H2O. The organic layer was
washed with aqueous NaHCO3 (saturated) and brine, dried
(Na2SO4), and concentrated. The residue was purified by column
chromatography (SiO2, 20% EtOAc in hexane) to give 16 (2.02 g,
0.00 (s, 3H), 0.82 (s, 9H), 1.74e1.85 (m, 1H), 2.73 (d, 1H, J ¼ 2.3), 3.04
(dd, 1H, J ¼ 5.5 and 10.1), 3.19 (dd, 1H, J ¼ 8.0 and 9.2), 3.27 (dd, 1H,
J ¼ 3.2 and 10.1), 3.53 (dd, 1H, J ¼ 6.9 and 9.2), 3.78 (s, 6H), 3.90 (dd,
1H, J ¼ 2.7 and 4.6), 4.05e4.09 (m, 1H), 4.26 (dd,1H, J ¼ 4.6 and 6.0),
4.99 (s, 1H), 6.80e7.51 (m, 13H); 13C NMR (126 MHz, CDCl3)
d
ꢀ4.9, ꢀ4.8, 18.0, 19.4, 19.4, 25.7, 28.3, 55.3, 64.4, 72.7, 74.9, 75.5,
82.7, 86.0, 107.5, 113.1, 126.7, 127.8, 128.4, 130.2, 136.3, 145.0, 158.5.
Anal. Calcd for C36H50O7$1/2H2O: C, 68.43; H, 8.14. Found: C, 68.46;
3.90 mmol, 98%): 1H NMR (400 MHz, CDCl3)
d
0.87e0.92 (m, 6H),
H, 7.99. Physical data of 29: 1H NMR (400 MHz, CDCl3)
d 0.03 (s, 3H),
1.80e1.87 (m, 1H), 3.21 (dd, 1H, J ¼ 6.9 and 9.2), 3.58 (dd, 1H, J ¼ 6.4
and 9.2), 4.53 (dd, 1H, J ¼ 5.3 and 11.2), 4.66e4.73 (m, 2H), 5.24 (s,
1H), 5.69 (d, 1H, J ¼ 4.6), 5.85 (dd, 1H, J ¼ 4.6 and 6.9), 7.25e8.12 (m,
0.04 (s, 3H), 0.72 (d, 3H, J ¼ 6.4), 0.77 (d, 3H, J ¼ 6.9), 0.82 (s, 9H),
1.68e1.71 (m, 1H), 2.43 (d, 1H, J ¼ 6.8), 2.98 (dd, 1H, J ¼ 5.0 and
10.1), 3.04 (dd, 1H, J ¼ 6.9 and 16.0), 3.19 (dd, 1H, J ¼ 3.7 and 10.1),
3.44 (dd, 1H, J ¼ 6.4 and 15.6), 3.67 (s, 6H), 3.91 (dd, 1H, J ¼ 4.8 and
11.7), 3.98 (d, 1H, J ¼ 4.1), 4.06 (s, 1H), 4.78 (s, 1H), 6.69e7.40 (m,
15H); 13C NMR (151 MHz, CDCl3)
d 19.2, 19.3, 28.3, 64.1, 65.2, 70.9,
71.9, 72.8, 75.1, 75.6, 78.7, 101.0, 105.7, 128.3, 128.3, 128.5, 129.0,
129.3, 129.7, 129.8, 133.1, 133.3, 133.4, 165.3, 165.4, 166.2. Anal. Calcd
for C30H30O8$7/10H2O: C, 67.84; H, 5.96. Found: C, 67.59; H, 5.73.
13H). 13C NMR (126 MHz, CDCl3)
d
ꢀ4.9, ꢀ4.5, 18.2, 19.4, 19.5, 25.8,
28.5, 55.3, 65.1, 72.7, 75.2, 76.7, 83.6, 86.1, 107.9, 113.1, 126.7, 127.8,
128.4, 130.2, 130.2, 136.3, 136.4, 145.1, 158.5. Anal. Calcd for
4.1.13. 1-O-(2-Methylpropyl)-
b
-D
-ribofuranose (4)
C36H50O7$1/2H2O: C, 68.43; H, 8.14. Found: C, 68.39; H, 8.29.
To a stirred solution of 16 (2.02 g, 3.90 mmol) in MeOH (15 mL)
was added a catalytic amount of a 28% MeOH solution of sodium
methoxide at room temperature. The mixture was stirred at room
temperature for 12 h and quenched by using aqueous NH4Cl
(saturated) (1 mL). The solvent was evaporated in vacuo, and the
resulting residue was purified by column chromatography (SiO2, 9%
MeOH in CHCl3) to give 4 (0.63 g, 3.06 mmol, 79%): 1H NMR
4.1.16. 2-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-
O-(2-methylpropyl)-3-O-[(2-cyanoethoxy)(N,N-diisopropyamino)]
phosphanyl-b-D-ribofuranose (36)
To a solution of 29 (0.53 g, 0.85 mmol) in THF (5.5 mL) was
added N,N-diisopropylethylamine (0.74 mL, 4.25 mmol) and
chloro(2-cyanoethoxy)(N,N-diisopropylamino)phosphine
(400 MHz, CDCl3)
d
: 0.89 (d, 6H, J ¼ 8.0), 1.79e1.90 (m, 1H), 2.24 (d,
(0.38 mL, 1.70 mmol) at room temperature. The mixture was stirred
at room temperature for 1 h and partitioned between CHCl3 and
aqueous NaHCO3 (saturated). The organic layer was washed with
brine, dried (Na2SO4), and concentrated. The residue was purified
by column chromatography (a neutralized SiO2, 33% EtOAc in
hexane) to give 36 (0.54 g, 0.66 mmol, 78%): 31P NMR (162 MHz,
1H, J ¼ 3.7), 2.48 (d, 1H, J ¼ 6.4), 2.67 (d, 1H, J ¼ 3.2), 3.21 (dd, 1H,
J ¼ 6.4 and 9.2), 3.51 (dd, 1H, J ¼ 6.8 and 9.2), 3.64e3.70 (m, 1H),
3.79e3.84 (m, 1H), 4.07e4.11 (m, 2H), 4.42 (d, 1H, J ¼ 5.5), 4.95 (s,
1H); 13C NMR (151 MHz, CDCl3)
d 19.2, 28.4, 62.9, 71.2, 75.5, 75.8,
84.4, 107.8. Anal. Calcd for C9H18O5: C, 52.41; H, 8.80. Found: C,
52.12; H, 8.78.
CDCl3)
d 149.4, 149.6.
4.1.14. 5-O-(4,40-Dimethoxytrityl)-1-O-(2-methylpropyl)-
b
-D
-
4.1.17. 2,3,5-O-Tribenzoyl-1-O-(2,2-dimethylpropyl)-b-D-
ribofuranose (21)
ribofuranose (17)
To a solution of 4 (0.23 g, 1.12 mmol) in pyridine (5 mL) was
added DMTrCl (0.38 g, 1.12 mmol) at room temperature. The
mixture was stirred at room temperature for 1 h and partitioned
between EtOAc and aqueous NaHCO3 (saturated). The organic layer
was washed with brine, dried (Na2SO4), and concentrated. The
residue was purified by column chromatography (SiO2, 33% EtOAc
in hexane) to give 21 (0.48 g, 0.94 mmol, 84%): 1H NMR (400 MHz,
To a solution of 1-O-acethyl-2,3,5-O-tribenzoyl-b-D-ribofur-
anose (3.00 g, 5.95 mmol) in CH2Cl2 (25 mL) was added dropwise
TMSOTf (1.29 mL, 7.14 mmol) and 2,2-dimethyl-1-propanol (0.52 g,
5.95 mmol) in CH2Cl2 (5 mL) at e30 ꢁC. The mixture was stirred
at ꢀ30 ꢁC for 3.5 h and quenched by using aqueous NaHCO3
(saturated). The mixture was partitioned between CHCl3 and H2O.
The organic layer was washed with aqueous NaHCO3 (saturated)

N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
469
and brine, dried (Na2SO4), and concentrated. The residue was pu-
9.8), 3.44 (d, 1H, J ¼ 8.7), 3.78 (s, 6H), 3.99 (dd, 1H, J ¼ 4.6 and 11.5),
4.10 (d, 1H, J ¼ 4.1), 4.17 (dd, 1H, J ¼ 2.3 and 5.0), 4.88 (s, 1H),
rified by column chromatography (SiO2, 20% EtOAc in hexane) to
give 17 (2.84 g, 5.33 mmol, 90%): 1H NMR (400 MHz, CDCl3)
d
0.92
6.80e7.51 (m, 13H); 13C NMR (151 MHz, CDCl3)
d
ꢀ6.8, ꢀ6.4, ꢀ5.4,
(s, 9H), 3.10 (d, 1H, J ¼ 9.2), 3.52 (d, 1H, J ¼ 9.2), 4.55 (dd, 1H, J ¼ 6.0
and 11.4), 4.65 (dd, J ¼ 4.6 and 11.4), 4.73 (dd, 1H, J ¼ 6.9 and 11.5),
5.24 (s, 1H), 5.71 (d, 1H, J ¼ 5.0), 5.82 (dd, 1H, J ¼ 5.0 and 6.8),
16.3, 23.8, 23.9, 24.8, 29.9, 53.3, 63.2, 70.9, 74.8, 77.0, 81.7, 84.1,
106.4, 111.2, 124.8, 125.9, 126.4, 128.3, 128.3, 134.4, 134.4, 143.2,
156.5. Anal. Calcd for C37H52O7Si$4/5H2O: C, 68.27; H, 8.30. Found:
C, 68.09; H, 8.35.
7.28e8.12 (m, 15H); 13C NMR (126 MHz, CDCl3)
19.4, 19.4, 25.7, 28.3, 55.3, 64.4, 72.7, 74.9, 75.5, 82.7, 86.0, 107.5,
113.1, 126.7, 127.8, 128.4, 130.2, 136.3, 145.0, 158.5. Anal. Calcd for
d
ꢀ4.9, ꢀ4.8, 18.0,
4.1.21. 2-O-(Tert-butyldimethyl)silyl-5-O-(4,40-dimethoxy)trityl-1-
O-(2,2-dimethyl)propyl-3-O-[(2-cyanoethoxy)(N,N-
C
31H32O8$1/2H2O: C, 68.75; H, 6.14. Found: C, 68.66; H, 5.89.
diisopropyamino)]phosphanyl-b-D-ribofuranose (37)
4.1.18. 1-O-(2,2-Dimethylpropyl)- -ribofuranose (5)
b
-
D
To a solution of 31 (0.57 g, 0.90 mmol) in THF (6.0 mL) was
added N,N-diisopropylethylamine (0.79 mL, 4.50 mmol) and
chloro(2-cyanoethoxy)(N,N-diisopropylamino)phosphine
(0.40 mL, 1.80 mmol) at room temperature. The mixture was stirred
at room temperature for 1 h and partitioned between CHCl3 and
aqueous NaHCO3 (saturated). The organic layer was washed with
brine, dried (Na2SO4), and concentrated. The residue was purified
by column chromatography (a neutralized SiO2, 25% EtOAc in
hexane) to give 37 (0.63 g, 0.75 mmol, 83%): 31P NMR (162 MHz,
To a stirred solution of 17 (2.25 g, 4.23 mmol) in MeOH (23 mL)
was added a catalytic amount of a 28% MeOH solution of sodium
methoxide at room temperature. The mixture was stirred at room
temperature for 12 h and quenched by using aqueous NH4Cl
(saturated) (1 mL). The solvent was evaporated in vacuo, and the
resulting residue was purified by column chromatography (SiO2, 9%
MeOH in CHCl3) to give 5 (0.70 g, 3.18 mmol, 75%): 1H NMR
(400 MHz, CDCl3) d 0.90 (s, 9H), 2.18 (s, 1H), 2.41 (s, 1H), 2.61 (s, 1H),
3.05 (d, 1H, J ¼ 7.8), 3.44 (d, 1H, J ¼ 8.7), 3.67 (m, 1H), 3.80e3.83 (m,
CDCl3) d 149.2, 149.5.
1H), 4.10 (m, 2H), 4.43 (s, 1H), 4.94 (s, 1H); 13C NMR (151 MHz,
DMSO-d6)
d
26.5, 31.3, 63.6, 71.3, 74.4, 76.7, 83.4, 107.2. Anal. Calcd
4.1.22. 2,3,5-Tri-O-benzoyl-1-O-(2,2,2-trifluoroethyl)-
ribofuranose (18)
b-
D-
for C10H20O5: C, 54.53; H, 9.15. Found: C, 54.42; H, 9.14.
To a solution of 1-O-acethyl-2,3,5-O-tribenzoyl-
b
-
D
-ribofur-
4.1.19. 5-O-(4,40-Dimethoxytrityl)-1-O-(2,2-dimethylpropyl)-
ribofuranose (22)
b-
D-
anose (2.00 g, 3.97 mmol) in CH2Cl2 (20 mL) was added dropwise
TMSOTf (0.86 mL, 4.76 mmol) and 2,2,2-trifluoroethanol (0.28 mL,
3.97 mmol) at ꢀ30 ꢁC. The mixture was stirred at ꢀ30 ꢁC for 1.5 h
and quenched by using aqueous NaHCO3 (saturated). The mixture
was partitioned between CHCl3 and H2O. The organic layer was
washed with aqueous NaHCO3 (saturated) and brine, dried
(Na2SO4), and concentrated. The residue was purified by column
chromatography (SiO2, 20% EtOAc in hexane) to give 18 (2.16 g,
To a solution of 5 (0.38 g, 1.74 mmol) in pyridine (5 mL) was
added DMTrCl (0.71 g, 2.09 mmol) at room temperature. The
mixture was stirred at room temperature for 1 h and partitioned
between EtOAc and aqueous NaHCO3 (saturated). The organic layer
was washed with brine, dried (Na2SO4), and concentrated. The
residue was purified by column chromatography (SiO2, 50% EtOAc
in hexane) to give 22 (0.90 g, 1.72 mmol, 99%): 1H NMR (400 MHz,
3.97 mmol, 100%): 1H NMR (400 MHz, CDCl3)
d 3.90e4.15 (m, 2H),
CDCl3)
d
0.77 (s, 9H), 2.27 (d, 1H, J ¼ 5.5), 2.50 (s, 1H), 2.93 (d, 1H,
4.56 (dd, 1H, J ¼ 5.5 and 11.9), 4.70 (dd, 1H, J ¼ 4.2 and 12.4), 4.78
(dd, 1H, J ¼ 5.5 and 11.0), 5.37 (s, 1H), 5.78 (d, 1H, J ¼ 5.0), 5.88 (dd,
1H, J ¼ 5.0 and 7.4), 7.30e8.07 (m, 15H); 13C NMR (126 MHz, CDCl3)
J ¼ 8.7), 3.17 (dd,1H, J ¼ 6.0 and 9.2), 3.30e3.35 (m, 2H), 3.78 (s, 6H),
4.06e4.12 (m, 2H), 4.19 (t, 1H, J ¼ 5.26), 4.91 (s, 1H), 6.81e7.47 (m,
13H); 13C NMR (151 MHz, CDCl3)
d
26.5, 31.5, 55.2, 65.7, 73.7, 75.3,
d 64.5, 64.6, 64.9, 71.9, 75.3, 79.9, 105.6, 128.5, 128.5, 128.6, 129.6,
78.1, 81.7, 86.2, 107.5, 113.1, 126.8, 127.8, 128.1, 130.0, 130.0, 136.0,
136.1, 144.8, 158.5. HRMS (ESI-TOF) m/z: Calcd for C31H38O7
[M þ Na]þ 545.2515; Found 545.2509.
129.8,129.8,129.9,133.4,133.6,133.7,165.2,165.4,166.2. Anal. Calcd
for C28H23F3O8$1/10H2O: C, 61.56; H, 4.28. Found: C, 61.38; H, 4.24.
4.1.23. 1-O-(2,2,2-Trifluoroethyl)-b-D-ribofuranose (6)
4.1.20. 3-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-
To a stirred solution of 18 (2.00 g, 3.68 mmol) in MeOH (20 mL)
was added a catalytic amount of a 28% MeOH solution of sodium
methoxide at room temperature. The mixture was stirred at room
temperature for 12 h and quenched by using aqueous NH4Cl
(saturated) (1 mL). The solvent was evaporated in vacuo, and the
resulting residue was purified by column chromatography (SiO2, 9%
MeOH in CHCl3) to give 6 (0.51 g, 2.19 mmol, 60%): 1H NMR
O-(2,2-dimethylpropyl)-
butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-O-(2,2-
dimethylpropyl)- -ribofuranose (31)
b-D-ribofuranose (30) and 2-O-(tert-
b-D
To a stirred solution of 22 (0.92 g, 1.76 mmol) and Et3N (0.74 mL,
5.28 mmol) in DMF (6 mL) was added dropwise TBDMSCl (0.53 g,
3.52 mmol) in DMF (4 mL) at room temperature. The mixture was
stirred at room temperature for 12 h and partitioned between
EtOAc and aqueous NaHCO3 (saturated). The organic layer was
washed with brine, dried (Na2SO4), and concentrated. The residue
was purified by column chromatography (SiO2, 17% EtOAc in hex-
ane) to give 30 (0.32 g, 0.51 mmol, 29%) and 31 (0.40 g, 0.63 mmol,
(400 MHz, CDCl3)
d
1.79 (dd,1H, J ¼ 4.6 and 7.8), 2.30 (d,1H, J ¼ 6.9),
2.60 (d, 1H, J ¼ 3.7), 3.69e3.75 (m, 1H), 3.81e3.94 (m, 2H),
3.97e4.03 (m, 1H), 4.10e4.18 (m, 1H), 4.18 (d, 1H, J ¼ 4.1), 4.40 (dd,
1H, J ¼ 6.4 and 11.4), 5.06 (s, 1H); 13C NMR (126 MHz, DMSO-D6)
d:
63.1, 63.3, 63.5, 71.0, 74.7, 84.7, 107.3. Anal. Calcd for C7H11F3O5: C,
36.22; H, 4.78. Found: C, 35.93; H, 4.69.
36%). Physical data of 30: 1H NMR (400 MHz, CDCl3)
d 0.10 (s, 3H),
0.01 (s, 3H), 0.82 (s, 9H), 0.82 (s, 9H), 2.73 (d, 1H, J ¼ 2.3), 3.02e3.06
(m, 2H), 3.23 (dd, 1H, J ¼ 3.2 and 10.1), 3.44 (d, 1H, J ¼ 9.2), 3.78 (s,
6H), 3.90 (s, 1H), 4.08 (d, 1H, J ¼ 3.2), 4.24 (dd, 1H, J ¼ 5.0 and 6.4),
4.97 (s, 1H), 6.79e7.50 (m, 13H); 13C NMR (151 MHz, CDCl3)
4.1.24. 5-O-(4,40-Dimethoxytrityl)-1-O-(2,2,2-trifluoroethyl)-
ribofuranose (23)
b-D-
To a solution of 6 (0.48 g, 2.07 mmol) in pyridine (13 mL) was
added DMTrCl (0.84 g, 2.48 mmol) at room temperature. The
mixture was stirred at room temperature for 1.5 h and partitioned
between EtOAc and aqueous NaHCO3 (saturated). The organic layer
was washed with brine, dried (Na2SO4), and concentrated. The
residue was purified by column chromatography (SiO2, 50% EtOAc
in hexane) to give 23 (1.03 g, 1.93 mmol, 93%): 1H NMR (400 MHz,
d
ꢀ5.0, ꢀ4.9, 17.9, 25.6, 26.6, 31.7, 55.2, 64.7, 72.8, 75.4, 78.4, 82.6,
85.9, 108.0, 113.0, 126.6, 127.7, 128.2, 130.1, 136.2, 144.9, 158.4. Anal.
Calcd for C37H52O7Si$1/2H2O: C, 68.80; H, 8.27. Found: C, 68.93; H,
8.20. Physical data of 31: 1H NMR (400 MHz, CDCl3)
d 0.14 (d, 6H,
J ¼ 3.2), 0.85 (s, 9H), 0.92 (d, 9H, J ¼ 3.2), 2.56 (d,1H, J ¼ 6.9), 3.01 (d,
1H, J ¼ 9.2), 3.09 (dd, 1H, J ¼ 5.0 and 10.1), 3.28 (dd, 1H, J ¼ 3.9 and

470
N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
CDCl3)
d
2.23 (d, 1H, J ¼ 5.5), 2.56 (s, 1H), 3.29 (d, 2H, J ¼ 5.0),
dried in vacuo. The amount of the compound 32 loaded on the solid
support 39 was 40 mol/g from the calculation based on absor-
bance of dimethoxytrityl cation released by treating with a solution
of 70% HClO4:EtOH (3:2, v/v). In a similar manner, the solid support
3.74e3.89 (m, 8H), 4.10e4.14 (m, 2H), 4.30 (dd, 1H, J ¼ 5.5 and 11.0),
m
5.04 (s, 1H), 6.82e7.45 (m, 13H); 13C NMR (126 MHz, CDCl3)
d
55.3,
64.2, 64.5, 64.6, 72.4, 75.1, 82.8, 86.4, 107.1, 113.3, 122.7, 124.9, 127.0,
128.0, 128.2, 130.1, 136.0, 144.8, 158.7. Anal. Calcd for C28H23F3O8$1/
2H2O: C, 61.87; H, 5.56. Found: C, 61.72; H, 5.29.
40 was obtained in 35 mmol/g of loading amount.
4.1.28. Oligonucleotide (ON) synthesis
4.1.25. 3-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-
Synthesis was carried out with a DNA/RNA synthesizer by
phosphoramidite method according to the normal protocol.
Deprotection of bases and phosphates was performed in concen-
trated NH4OH:EtOH (3:1, v/v) at 55 ꢁC for 4 h. 20-TBDMS groups
O-(2,2,2-trifluoroethyl)-
butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-1-O-(2,2,2-
trifluoroethyl)- -ribofuranose (33)
b-D-ribofuranose (32) and 2-O-(tert-
b-D
To a stirred solution of 23 (1.23 g, 2.30 mmol) and Et3N (0.96 mL,
6.90 mmol) in DMF (12 mL) was added dropwise TBDMSCl (0.69 g,
4.60 mmol) in DMF (8 mL) at room temperature. The mixture was
stirred at room temperature for 12 h and partitioned between
EtOAc and aqueous NaHCO3 (saturated). The organic layer was
washed with brine, dried (Na2SO4), and concentrated. The residue
was purified by column chromatography (SiO2, 11% EtOAc in hex-
ane) to give 32 (0.59 g, 0.91 mmol, 40%) and 33 (0.43 g, 0.66 mmol,
were removed by Et3N$3HF (125 mL, Aldrich) in DMSO (100 mL,
Aldrich) at 65 ꢁC for 90 min. The reaction was quenched with 0.1 M
TEAA buffer (pH 7.0) and desalted on Sep-Pak C18 cartridge.
Deprotected ONs were purified by 20% PAGE containing 7 M urea
and/or reversed-phase C-18 HPLC using a gradient of solution A (5%
MeCN in 0.1 M TEAA buffer, pH 7.0) and solution B (50% MeCN in
0.1 M TEAA buffer, pH 7.0) to give the purified ON 1 (6), ON 2 (9), ON
3 (6), ON 4 (10), ON 5 (9), ON 6 (16), ON 7 (1), ON 8 (3), ON 9 (2), ON
10 (3), ON 11 (5), ON 12 (5), ON 13 (7), ON 14 (5), ON 15 (10), ON 16
(16), ON 17 (16), ON 18 (15), ON 20 (23), ON 21 (19), ON 22 (12), ON
23 (18), ON 24 (3). The yields are indicated in parentheses as OD
29%). Physical data of 32: 1H NMR (400 MHz, CDCl3)
d
ꢀ0.13 (s,
3H), ꢀ0.00 (s, 3H), 0.81 (s, 9H), 2.77 (d, 1H, J ¼ 1.4), 3.03 (dd, 1H,
J ¼ 5.0 and 10.1), 3.35 (dd, 1H, J ¼ 2.8 and 10.5), 3.79 (s, 6H),
3.82e3.95 (m, 2H), 3.99 (d,1H, J ¼ 2.8), 4.12 (d, 1H, J ¼ 6.4), 4.32 (dd,
1H, J ¼ 4.6 and 11.4), 5.11 (s, 1H), 6.81e7.48 (m, 13H); 13C NMR
units at 260 nm starting from 0.2 or 1.0 mmol scale.
(126 MHz, CDCl3)
d
ꢀ4.9, 18.0, 25.7, 55.3, 63.7, 64.4, 64.6, 72.1, 75.2,
4.1.29. MALDI-TOF/MS analysis of oligonucleotides (ONs)
83.5, 86.1, 107.4, 113.2, 126.9, 127.9, 128.3, 130.1, 130.1, 136.0, 136.1,
144.8, 158.6. Anal. Calcd for C34H43F3O7Si$1/2H2O: C, 62.08; H, 6.74.
Found: C, 61.80; H, 6.60. Physical data of 33: 1H NMR (400 MHz,
Spectra were obtained with a SHIMAZU/KRATOS time-of-flight
mass spectrometer equipped with a nitrogen laser (337 nm, 3-ns
pulse). A solution of 3-hydroxypicolinic acid (3-HPA) and dia-
mmonium hydrogen citrate in H2O was used as the matrix. Data of
synthetic ONs: ON 1 m/z ¼ 6505.86 (calcd for C194H245N65O150P20
CDCl3)
d
0.14 (s, 3H), 0.14 (s, 3H), 0.92 (s, 9H), 2.45 (d, 1H, J ¼ 7.4),
3.11 (dd, 1H, J ¼ 5.0 and 10.1), 3.34 (dd, 1H, J ¼ 3.2 and 10.6), 3.78 (s,
6H), 3.81e3.96 (m, 2H), 4.07e4.13 (m, 2H), 4.23 (dd, 1H, J ¼ 1.8 and
4.6), 4.99 (s, 1H), 6.81e7.48 (m, 13H); 13C NMR (126 MHz, CDCl3)
[MꢀH]e, 6505.87); ON
2
m/z
¼
6815.21 (calcd for
C
203H248N86O144P20 [MꢀH]e, 6815.15); ON 3 m/z ¼ 6501.81 (calcd
d
ꢀ5.0, ꢀ4.7, 18.2, 25.8, 55.3, 64.4, 64.5, 64.8, 72.1, 76.4, 84.5, 86.2,
for C196H249N65O148P20 [MꢀH]e, 6501.93); ON 4 m/z ¼ 6811.59
107.7, 113.2, 126.9, 127.9, 128.3, 130.2, 136.0, 136.1, 144.9, 158.6.
HRMS (ESI-TOF) m/z: Calcd for C34H43F3O7Si [M þ Na]þ 671.2628;
Found 671.2631.
(calcd for
C
205H252N86O142P20 [MꢀH]e, 6811.21); ON
5 m/
z ¼ 6625.21 (calcd for C198H254N65O152P21 [MꢀH]e, 6625.96); ON
6 m/z ¼ 6936.55 (calcd for C207H257N86O146P21 [MꢀH]e, 6935.24);
ON 7 m/z ¼ 6498.06 (calcd for C192H254N61O152P21 [MꢀH]e,
6497.87); ON 8 m/z ¼ 6807.11 (calcd for C201H257N82O146P21
4.1.26. 2-O-(Tert-butyldimethylsilyl)-5-O-(4,40-dimethoxytrityl)-3-
O-[(2-cyanoethoxy) (N,N-diisopropyamino)]phosphanyl-1-O-
[MꢀH]e, 6807.14); ON
9
m/z
¼
6526.12 (calcd for
(2,2,2-trifluoroethyl)-
b-
D-ribofuranose (38)
C
194H258N61O152P21 [MꢀH]e, 6525.92); ON 10 m/z ¼ 6835.36 (calcd
To a solution of 33 (0.25 g, 0.39 mmol) in THF (3 mL) was added
N,N-diisopropylethylamine (0.34 mL, 1.95 mmol) and chloro(2-
for C203H261N82O146P21 [MꢀH]e, 6835.20); ON 11 m/z ¼ 6553.76
(calcd for
C
196H262N61O152P21 [MꢀH]e, 6554.98); ON 12 m/
cyanoethoxy)(N,N-diisopropylamino)phosphine
(0.17
mL,
z ¼ 6863.47 (calcd for C205H265N82O146P21 [MꢀH]e, 6863.25); ON
13 m/z ¼ 6582.34 (calcd for C198H266N61O152P21 [MꢀH]e, 6582.03);
ON 14 m/z ¼ 6891.75 (calcd for C207H269N82O146P21 [MꢀH]e,
6891.31); ON 15 m/z ¼ 6605.26 (calcd for C192H248F6N61O152P21
0.78 mmol) at room temperature. The mixture was stirred at room
temperature for 45 min and partitioned between CHCl3 and
aqueous NaHCO3 (saturated). The organic layer was washed with
brine, dried (Na2SO4), and concentrated. The residue was purified
by column chromatography (a neutralized SiO2, 33% EtOAc in
hexane) to give 38 (0.28 g, 0.32 mmol, 83%): 31P NMR (162 MHz,
[MꢀH]e, 6605.81); ON 16 m/z
¼
6915.77 (calcd for
C
201H251F6N82O146P21 [MꢀH]e, 6915.09); ON 17 m/z ¼ 7043.06
(calcd for
C
221H269N66O159P21 [MꢀH]e, 7043.34); ON 18 m/
CDCl3)
d
149.7, 150.2.
z ¼ 6202.99 (calcd for C185H226N82O128P18 [MꢀH]e, 6202.82); ON
20 m/z ¼ 7039.43 (calcd for C223H273N66O157P21 [MꢀH]e, 7039.39);
ON 21 m/z ¼ 7163.68 (calcd for C225H278N66O161P22 [MꢀH]e,
7163.42); ON 22 m/z ¼ 7121.75 (calcd for C225H290N62O161P22
4.1.27. Solid support synthesis
A mixture of 32 (0.13 g, 0.20 mmol), succinic anhydride (80 mg,
0.8 mmol), and DMAP (50 mg, 0.4 mmol) in pyridine (2.5 mL) was
stirred at room temperature for 18.5 h. The mixture was partitioned
between EtOAc and aqueous NaHCO3 (saturated). The organic layer
was washed with brine, dried (Na2SO4), and concentrated to give
the corresponding succinate. A mixture of the succinate (0.18 g,
[MꢀH]e, 7119.49); ON 23 m/z
¼
7143.60 (calcd for
C
219H272F6N62O161P22 [MꢀH]e, 7143.27); ON 24 m/z ¼ 7018.93
(calcd for C217H267F6N62O157P21 [MꢀH]e, 7019.24).
4.1.30. Preparation of the human Ago2 PAZ domain
0.24 mmol), aminopropyl controlled pore glass (0.45 g, 60
m
mol),
A synthetic DNA encoding the human Ago2 PAZ domain (resi-
dues 225e369 in Protein Data Bank entry 4EI1) with an N-terminal
His-tag was purchased from Life Technologies and cloned into the
pETY_Blue vector [33] by the patch cloning method as described in
the literature [34]. The Escherichia coli strain harboring the
resulting plasmid, pETY_PAZ, was cultivated at 37 ꢁC in LB-
ampicillin medium. The expression of the PAZ domain was
and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochlo-
ride (50 mg, 0.24 mmol) in DMF (2.4 mL) was shaken at room
temperature for 4 days. After the resin was washed with pyridine, a
capping solution (15 mL, 0.1 M DMAP in pyridine:Ac2O ¼ 9:1, v/v)
was added and the whole mixture was shaken at room temperature
for 24 h. The resin was washed with pyridine, EtOH, and MeCN, and

N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
471
induced by adding isopropylthiogalactoside (IPTG) to a final con-
centration of 0.5 mM and by overnight cultivation at a lowered
temperature, i.e., 20 ꢁC. The cells expressing the PAZ domain were
disrupted by sonication in Buffer A (20 mM TriseHCl (pH 7.6), 1 mM
concentration: 40
amounts of competitor (final concentration: 100
the EMSA buffer (5
L) at 37 ꢁC for 10 min. The reaction mixture
(5 L) was mixed with the loading buffer (5 L) on ice. Each sample
m
M) was incubated in the presence of various
m
M or 1000 M) in
m
m
m
m
MgCl2, 200 mM KCl, 6 mM
b
-mercaptoethanol, 5% glycerin, and
was analyzed by 8% nondenaturing PAGE in buffer containing 5%
glycerin at 10 ꢁC for 1.5 h at 200 V. The gel was visualized by use of a
Luminescent Image analyzer LAS-4000 (Fujifilm).
400 M 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride
m
(AEBSF)), and the cell extract obtained by centrifugation was loaded
onto Ni-nitrilotriacetic acid (Ni-NTA) agarose resin (2 mL) at 4 ꢁC.
The resin was then washed with Buffer A containing 10 mM
imidazole, and proteins containing the PAZ domain were eluted
with Buffer A containing 250 mM imidazole. The buffer of the
eluate was exchanged for Buffer B (20 mM HEPES-KOH (pH 7.5),
4.1.34. Enzyme-linked immunosorbent assay (ELISA)
A solution of dsRNAs (100 nM, 50 mL), which were modified with
biotin at the 30-end of the passenger strand, were added to a
streptavidin-coated 96-well plate. After 15 min, each well was
10 mM MgCl2, and 6 mM
b-mercaptoethanol) and applied to a
washed three times with a solution (200
in PBS (pH 7.4) and then a solution (100
(pH 7.4) was added. After 2 h, each well was washed and then a
solution of the PAZ domain protein (100 nM, 50 L) in PBS buffer
m
L ꢂ 3) of 0.05% Tween 20
HiTrap Heparin HP column (1 mL). The PAZ domain was eluted with
a linear gradient (7 mL) from 0 to 500 mM KCl in Buffer B. Finally,
PAZ was concentrated by ultrafiltration and stored at ꢀ80 ꢁC in a
buffer containing 40 mM TriseHCl (pH 7.6), 2 mM MgCl2, 80 mM
KCl, and 2 mM DTT. The amount of the purified PAZ domain was
determined from the absorbance at 280 nm and by using the
ProtParam tool (http://web.expasy.org/protrapam/).
mL) of 3% BSA in PBS buffer
m
(pH 7.4) contained 1% BSA was added. The plate was incubated at
37 ꢁC for 10 min. Each well was washed and then a solution of the
anti-His tag antibody conjugated horseradish peroxidase (50
PBS containing 1% BSA was added. After 1 h, each well was washed
and then a peroxidase (POD) substrate (50 L) was added. After
10 min, a stop solution (1 M H2SO4, 50 L) was added and then an
mL) in
m
4.1.31. Thermal denaturation study
m
The solution containing the duplex in a buffer comprising
10 mM sodium phosphate (pH 7.0) and 0.1 M NaCl was heated at
95 ꢁC for 3 min, cooled gradually to an appropriate temperature,
and then used for the thermal denaturation study. The thermal-
induced transition of each mixture was monitored at 260 nm on
UVeVis spectrophotometer fitted with a temperature controller in
absorbance of each well at 450 nm was measured by a microplate
reader.
4.1.35. Partial digestion of ONs by SVPD
Each ON (600 pmol) labeled with fluorescein at the 50-end was
incubated with SVPD (0.075 unit) in a buffer (150
0.1 M TriseHCl (pH 8.0) and 20 mM MgCl2 at 37 ꢁC. After 0, 1, 5, 10,
30, 60,180, or 360 min, an aliquot (5 L) of the reaction mixture was
mixed with the loading buffer (15 L), comprising Tris-borate-EDTA
mL) comprising
quartz cuvettes with a path length of 1.0 cm and a 3.0 mM duplex
concentration in a buffer of 10 mM sodium phosphate (pH 7.0) and
m
0.1 M NaCl. The sample temperature was increased by 0.5 ꢁC/min.
m
(TBE) buffer and 20% glycerin, on ice. Each sample was analyzed by
20% denaturing PAGE at room temperature for 2 h at 20 mA. The gel
was visualized by use of a Luminescent Image analyzer LAS-4000
(Fujifilm).
4.1.32. Dual-luciferase reporter assay
HeLa cells were grown at 37 ꢁC in a humidified atmosphere of 5%
CO2 in air in Minimum Essential Medium (MEM) (Invitrogen)
supplemented with 10% bovine serum (BS). Twenty-four hours
before transfection, HeLa cells (4 ꢂ 104/mL) were transferred to a
Acknowledgments
96-well plate (100 mL/well). They were transfected using TransFast
(Promega). Cells in each well were transfected with a solution
(35 L) of psiCHECK-2 vector (20 ng), the indicated amounts of
siRNA, and TransFast (0.3 g) in Opti-MEM I Reduced-Serum Me-
dium (Invitrogen), and incubated at 37 ꢁC. Transfection without
siRNA was used as a control. After 1 h, MEM (100 L) containing 10%
This research was partially supported by the Ministry of Edu-
cation, Culture, Sports, Science, and Technology of Japan (Grant-in-
Aid for Exploratory Research, 24659045) and by a C19 Kiyomi
Yoshizaki research grant. We thank Professor Tomio Yabe (Gifu
University) for the technical advice of the determination of the
biding ability by ELISA.
m
m
m
BS was added to each well, and the whole was incubated at 37 ꢁC.
After 24 h, solution in each well was removed and the plate was left
unattended at ꢀ80 ꢁC. After more than 4 h, activity of firefly and
Renilla luciferases in cell were determined with a dual-luciferase
assay system (Promega) according to a manufacturer's protocol.
The results were confirmed by at least three independent trans-
fection experiments with two cultures each and are expressed as
the average from four experiments as mean SD.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2015.09.011.
References
[1] A. Fire, S. Xu, M.K. Montgomery, S.A. Kostas, S.E. Driver, C.C. Mello, Potent and
specific genetic interference by double-stranded RNA in Caenorhabditis ele-
gans, Nature 391 (1998) 806e811.
[2] S.M. Elbashir, J. Harborth, W. Lendeckel, A. Yalcin, K. Weber, T. Tuschl, Du-
plexes of 21-nucleotide RNAs mediate RNA interference in cultured
mammalian cells, Nature 411 (2001) 494e498.
[3] S.M. Elbashir, W. Lendeckel, T. Tuschl, RNA interference is mediated by 21e
and 22enucleotide RNAs, Genes Dev. 15 (2001) 188e200.
[4] J.C. Burnett, J. Rossi, RNA-based therapeutics: current progress and future
prospects, J. Chem. Biol. 19 (2012) 60e71.
[5] A. Gallas, C. Alexander, M.C. Davies, S. Puri, S. Allenm, Chemistry and formu-
lations for siRNA therapeutics, Chem. Soc. Rev. 42 (2013) 7983e7997.
[6] S. Shukla, C.S. Sumaria, P.I. Pradeepkumar, Exploring chemical modifications
for siRNA therapeutics: a structural and functional outlook, ChemMedChem 5
(2010) 328e349.
4.1.33. Electrophoretic mobility shift assay (EMSA)
siRNA 9 (final concentration: 10
and ON 18, was incubated with various amounts of the PAZ domain
protein (final concentration: 0 M, 10 M, 20 M, 40 M, or 80 M)
in an EMSA buffer (10 L) comprising 20 mM TriseHCl (pH 8.0),
0.1 M NaCl, and 1 mM DTT at 37 ꢁC for 10 min. An aliquot (5
L) of
the reaction mixture was mixed with a loading buffer (an EMSA
buffer containing 50% glycerin; 5 L) on ice. Each sample was
mM), comprising labeled ON 17
m
m
m
m
m
m
m
m
analyzed by 8% nondenaturing PAGE in buffer containing 5% glyc-
erin at 10 ꢁC for 1.5 h at 200 V. The gel was visualized by use of a
Luminescent Image analyzer LAS-4000 (Fujifilm). The competition
assay was carried out as follows: A mixture of siRNA 9 (final con-
[7] K. Yoshikawa, A. Ogata, C. Matsuda, M. Kohara, H. Iba, Y. Kitade, Y. Ueno,
Incorporation of biaryl units into the 5' and 3' ends of sense and antisense
centration: 10
mM) and the PAZ domain protein (final

472
N. Inada et al. / European Journal of Medicinal Chemistry 103 (2015) 460e472
strands of siRNA duplexes improves strand selectivity and nuclease resistance,
Bioconj. Chem. 22 (2011) 42e49.
[8] Y. Ueno, Y. Watanabe, A. Shibata, K. Yoshikawa, T. Takano, M. Kohara,
Y. Kitade, Synthesis of nuclease-resistant siRNAs possessing universal over-
hangs, Bioorg. Med. Chem. 17 (2009) 1974e1981.
[9] Y. Ueno, T. Inoue, M. Yoshida, K. Yoshikawa, A. Shibata, Y. Kitamura, Y. Kitade,
Synthesis of nuclease-resistant siRNAs possessing benzene-phosphate back-
bones in their 3'-overhang regions, Bioorg. Med. Chem. Lett. 18 (2008)
5194e5196.
[10] K. Taniho, R. Nakashima, M. Kandeel, Y. Kitamura, Y. Kitade, Synthesis and
biological properties of chemically modified siRNAs bearing 1-deoxy-D-
ribofuranose in their 3'-overhang region, Bioorg. Med. Chem. Lett. 22 (2012)
2518e2521.
Y. Kitamura, Y. Kitade, Bioenergetics and gene silencing approaches for
unraveling nucleotide recognition by the human EIF2C2/Ago2 PAZ domain,
PLoS One 9 (2014) e94538.
[22] K. Nakamoto, Y. Ueno, Diazirine-containing RNA photo-crosslinking probes
for capturing microRNA targets, J. Org. Chem. 79 (2014) 2463e2472.
[23] T. Ohmichi, S. Nakano, D. Miyoshi, N. Sugimoto, Long RNA dangling end has
large energetic contribution to duplex stability, J. Am. Chem. Soc. 124 (2002)
10367e10372.
[24] T. Takahashi, S. Zenno, O. Ishibashi, T. Takizawa, K. Saigo, K. Ui-Tei, In-
teractions between the non-seed region of siRNA and RNA-binding RLC/RISC
proteins, Ago and TRBP, in mammalian cells, Nucleic Acids Res. 42 (2014)
5256e5269.
[25] C.L. Noland, J.A. Doudna, Multiple sensors ensure guide strand selection in
human RNAi pathways, RNA 19 (2013) 639e648.
[11] N.T. Schirle, I.J. MacRae, The crystal structure of human Argonaute2, Science
336 (2012) 1037e1040.
[12] N.T. Schirle, J. Sheu-Gruttadauria, I.J. MacRae, Structural basis for microRNA
targeting, Science 346 (2014) 608e613.
[26] C.L. Noland, E. Ma, J.A. Doudna, siRNA repositioning for guide strand selection
by human Dicer complexes, Mol. Cell 43 (2011) 110e121.
ꢀ
[27] D.S. Schwarz, G. Hutvagner, T. Du, Z. Xu, N. Aronin, P.D. Zamore, Asymmetry in
[13] E. Elkayam, C.-D. Kuhn, A. Tocilj, A.D. Haase, E.M. Greene, G.J. Hannon,
L. Joshua-Tor, The structure of human argonaute-2 in complex with miR-20a,
Cell 150 (2012) 100e110.
the assembly of the RNAi enzyme complex, Cell 115 (2003) 199e208.
[28] A. Khvorova, A. Reynolds, S.D. Jayasena, Functional siRNAs and miRNAs
exhibit strand bias, Cell 115 (2003) 209e216.
[14] S. Weitzer, J. Martinez, The human RNA kinase hClp1 is active on 39 transfer
RNA exons and short interfering RNAs, Nature 447 (2007) 222e226.
[15] J.S. Parker, S.M. Roe, D. Barford, Structure insights into mRNA recognition from
a PIWI domain-siRNA guide complex, Nature 434 (2005) 663e666.
[16] J.-B. Ma, Y.-R. Yuan, G. Meister, Y. Pei, T. Tuschl, D.J. Patel, Structure basis for
50-end-specific recognition of guide RNA by the A. fulgidus Piwi protein, Na-
ture 434 (2005) 666e670.
[17] A. Lingel, B. Simon, E. Izaurralde, M. Sattler, Structure and nucleic-acid binding
of the Drosophila Argonaute 2 PAZ domain, Nature 426 (2003) 465e469.
[18] K.S. Yan, S. Yan, A. Farooq, A. Han, L. Zeng, M.-M. Zhou, Structure and
conserved RNA binding of the PAZ domain, Nature 426 (2003) 469e474.
[19] J.-J. Song, J. Liu, N.H. Tolia, J. Schneiderman, S.K. Smith, R.A. Martienssen,
G.J. Hannon, L. Joshua-Tor, The crystal structure of the Argonaute2 PAZ
domain reveals an RNA binding motif in RNAi effector complexes, Nat. Struct.
Biol. 12 (2003) 1026e1032.
[29] Y. Tomari, C. Matranga, B. Haley, N. Martinez, P.D. Zamore, A protein sensor
for siRNA asymmetry, Science 306 (2004) 1377e1380.
[30] A. Somoza, M. Terrazas, R. Eritja, Modified siRNAs for the study of the PAZ
ꢀ
domain, Chem. Commun. 46 (2010) 4270e4272.
[31] M. Gaglione, N. Potenza, G.D. Fabio, V. Romanucci, N. Mosca, A. Russo,
E. Novellino, S. Cosconati, A. Messere, Tuning RNA interference by enhancing
siRNA/PAZ recognition, ACS Med. Chem. Lett. 4 (2013) 75e78.
[32] L. Xu, X. Wang, H. He, J. Zhou, X. Li, H. Ma, Z. Li, Y. Zeng, R. Shao, S. Cen,
Y. Wang, Structure-based design of novel chemical modification of the 30-
overhang for optimization of short interfering RNA performance, Biochem-
istry 54 (2015) 1268e1277.
[33] Y. Nomura, Y. Onda, S. Ohno, H. Taniguchi, K. Ando, N. Oka, K. Nishikawa,
T. Yokogawa, Purification and comparison of native and recombinant tRNA-
guanine transglycosylases from Methanosarcina acetivorans, Protein Expr.
Purif. 88 (2013) 13e19.
[20] J.-B. Ma, K. Ye, D.J. Patel, Structure basis for overhangespecific small inter-
fering RNA recognition by the PAZ domain, Nature 429 (2004) 318e322.
[21] M. Kandeel, A. Al-Taher, R. Nakashima, T. Sakaguchi, A. Kandeel, Y. Nagaya,
[34] N. Taniguchi, S. Nakayama, T. Kawakami, H. Murakami, Patch cloning method
for multiple site-directed and saturation mutagenesis, BMC Biotechnol. 13
(2013) 91.