Improving RNA interference in mammalian cells by 4′-thio-modified small interfering RNA (siRNA): Effect on siRNA activity and nuclease stability when used in combination with 2′-O-alkyl modifications

Article abstract of DOI:10.1021/jm050822c

A systematic structure-activity relationship study of 4′-thioribose containing small interfering RNAs (siRNAs) has led to the identification of highly potent and stable antisense constructs. To enable this optimization effort for both in vitro and in vivo

Full text of DOI:10.1021/jm050822c

1624
J. Med. Chem. 2006, 49, 1624-1634
Improving RNA Interference in Mammalian Cells by 4′-Thio-Modified Small Interfering RNA
(siRNA): Effect on siRNA Activity and Nuclease Stability When Used in Combination with
2′-O-Alkyl Modifications
Prasad Dande, Thazha P. Prakash, Namir Sioufi, Hans Gaus, Russell Jarres, Andreas Berdeja, Eric E. Swayze,
Richard H. Griffey, and Balkrishen Bhat*
Department of Medicinal Chemistry and Antisense Core Research, Isis Pharmaceuticals Inc., Carlsbad, California 92008
ReceiVed August 17, 2005
A systematic structure-activity relationship study of 4′-thioribose containing small interfering RNAs (siRNAs)
has led to the identification of highly potent and stable antisense constructs. To enable this optimization
effort for both in vitro and in vivo applications, we have significantly improved the yields of
4′-thioribonucleosides by using a chirally pure (R)-sulfoxide precursor. siRNA duplexes containing
strategically placed regions of 4′-thio-RNA were synthesized and evaluated for RNA interference activity
and plasma stability. Stretches of 4′-thio-RNA were well tolerated in both the antisense and sense strands.
However, optimization of both the number and placement of 4′-thioribonucleosides was necessary for maximal
potency. These optimized siRNAs were generally equipotent or superior to native siRNAs and exhibited
increased thermal and plasma stability. Furthermore, significant improvements in siRNA activity and plasma
stability were achieved by judicious combination of 4′-thioribose with 2′-O-methyl and 2′-O-methoxyethyl
modifications. These optimized 4′-thio-siRNAs may be valuable for developing stable siRNAs for therapeutic
applications.
Introduction
Chemical modifications have long been used to effectively
improve the serum stability of ASOs,16-18,22-24 and several
groups have sought to develop a similar strategy for
siRNAs.6,15,19-21,25-29 Czauderna et. al. have reported increased
duration of siRNA activity in mammalian cells with increasing
nuclease stability of siRNAs.26 Although several investigators
have incorporated nuclease-resistant 2′-O-alkyl modifications
and locked nucleic acids into siRNAs, these modifications have
generally led to reduced potency when compared with parent
unmodified siRNAs.6,15,19-21,25-29 However, thorough optimiza-
tion of the placement of modifications in each strand can lead
to derivatives with both improved potency and improved
stability, as exemplified by our recent report on fully modified
siRNA duplexes having improved activity.30
Small interfering RNAs (siRNAs) are short 19-21-nucleotide
RNA duplexes that regulate gene expression via the RNA
interference (RNAi) pathway.1-7 Over the past few years several
groups have reported the use of chemically synthesized siRNAs
to control gene expression in cultured mammalian cells.1,2,8
siRNAs introduced into cells using cationic lipid transfection
agents have shown significant reduction in target mRNA
levels.2,9,10 Recent reports also provide preliminary evidence of
in vivo siRNA activity in mouse models using high-pressure
tail vein injections.11,12 Although detectable levels of siRNAs
were found in liver tissue in these mice, this method of
administration is unsuitable for clinical applications. Sorensen
et. al. have reported gene silencing by a cationic liposome based
intravenous co-injection of siRNAs and a reporter plasmid.13
More recently, Soutschek et. al. have demonstrated endogenous
gene silencing by a systemically administered siRNA-
cholesterol conjugate.14 These reports offer preliminary evidence
of the therapeutic potential of siRNAs and warrant further
improvements in siRNA design with a view toward improving
its potency and pharmacokinetic properties.
The biodistribution of phosphodiester and phosphorothioate
siRNAs in mice has been reported to be similar to that for
antisense oligonucleotides (ASOs) following intravenous or
intraperitoneal injection.15 Resistance to degradation by nu-
cleases is one of the key factors which determine the bioavail-
ability, in vivo potency, and efficacy of ASOs.10,16-18 Further-
more, since degradation in both tissues and plasma is observed
via 3′-exo-, 5′-exo-, and endonucleases, stabilizing all parts of
the ASO drug will likely be required to achieve the maximum
therapeutic benefit. Since there is growing evidence that
unmodified siRNAs are rapidly degraded in serum,15,19-21
identifying highly stabilized siRNA constructs is crucial to the
development of siRNA therapeutics.
We have also recently reported on the positional effects of
well-known RNA-stabilizing 2′-ribose modifications, namely,
2′-O-methyl (2′-O-Me), 2′-R-fluoro (2′-F), and 2′-O-methoxy-
ethyl (2′-O-MOE), on siRNA activity.31 These studies clearly
demonstrated that 2′-substitution at the 5′-end of the antisense
strand tended to reduce potency, and that the reduction correlated
with the steric bulk of the substituent. The exceptions to this
general rule are the alternating 2′-O-Me motifs with either
RNA26 or 2′-F30 which improve both stability and potency in
certain cases. However, these observations have not been
generalized to broad tolerance for other stabilizing modifications
at the 5′-end of the antisense strand. Since nuclease resistance
generally correlates positively with both the number and bulk
of 2′-substituents, additional methods to stabilize the 5′-end of
antisense strand siRNAs are needed.
Another way to enhance nuclease resistance and plasma-
protein binding of oligonucleotides is to introduce sulfur into
the molecule by using phosphorothioate oligonucleotides.16,18,22
These have been used extensively in antisense therapeutics to
generate serum-stable oligonucleotides with increased plasma-
protein binding and plasma residence time (as compared to those
of phosphodiester oligonucleotides). However, a similar strategy
* To whom correspondence should be addressed. Phone: (760) 603-
3886. Fax: (760) 603-2446. E-mail: bbhat@isisph.com.
10.1021/jm050822c CCC: $33.50 © 2006 American Chemical Society
Published on Web 02/08/2006

ImproVing RNA Interference in Mammalian Cells
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5 1625
bovine liver adenosine kinase and do not show substrate
inhibition of the enzyme.37 Thus, kinase resistance, metabolic
stability, and H-bonding abilities make 4′-thioribonucleosides
interesting candidates for incorporation into siRNA duplexes
to enhance siRNA serum stability and possibly protein binding
properties.
The synthesis and characterization of 4′-thio-RNA was first
38-41
reported by Imbach et. al
and more recently by Hoshika
et al.42-44 They showed that phosphodiester 4′-thio-RNA
oligomers were significantly more resistant to cleavage by serum
nucleases than unmodified phosphodiester RNA. Furthermore,
4′-thio-RNA hybridizes with complementary RNA sequences
to form stable duplexes with standard A-form helical geometry.
These properties make 4′-thio-RNA a structurally similar, yet
nuclease-resistant analogue of wild-type RNA and hence
attractive for siRNA applications.
In a recent report by Hoshika et. al. on the siRNA activity of
4′-thio-RNA in cultured NIH/3T3 cells against photinus lu-
ciferase,43 it was shown that 4′-thio-RNA is tolerated predomi-
nantly in the sense strand, and that incorporation into the
antisense strand in large numbers leads to greater than 10-fold
reductions in potency. The best construct tested consisted of
modifications on the sense strand alone, and showed similar
potency. In our preliminary communications we independently
reported utility of 4′-thioribose-modified siRNA in down-
regulating the mRNA of an endogenous gene (PTEN).45,46
However, the siRNAs with a minimum number of 4′-thioribose-
modified residues resulted in good activity but a small benefit
in stability toward nucleases. Because of our previous success
in optimizing the placement of chemical modifications in
siRNAs,30,31 we set out to carry out a detailed analysis of the
structure-activity relationship (SAR) of 4′-thio-RNA-containing
siRNAs to identify 4′-thioribose-modified motifs that will lead
to both improved potency and a significant enhancement in
nuclease stability.
Figure 1. (a) 4′-Thio-RNA and 2′-O-modified RNA. Base ) Ade,
Gua, Cyt, or Ura. (b) 4′-Thioribonucleoside 3′-phosphoramidites used
in 4′-thio-RNA synthesis.
when applied to siRNAs resulted in reduced activity as
compared to that of native phosphodiester siRNAs.31 Incorpora-
tion of a thioether into the 2′-substitutent of single-stranded
oligonucleotides has resulted in improved binding to human
serum albumin.32 As the binding of oligonucleotides to plasma
proteins has been positively correlated to in vivo tissue
distribution,18 the introduction of sulfur is expected to improve
the pharmacokinetic properties of siRNA by both increasing
the stability toward degradation and improving the distribution
to tissues. However, this 2′-modified sulfur-containing modifica-
tion is not ideal for siRNA, as it is expected to decrease activity
when incorporated into the antisense strand due to its large steric
bulk as is evident from our previous results with 2′-O-MOE
modifications.31
In contrast, a sulfur-containing modification seemingly ideal
for use in siRNAs is 4′-thio-RNA, synthesized from 4′-
thioribonucleosides that have a sulfur atom at the 4-position of
the thiofuranose ring (Figure 1a). Although 4′-thioribonucleo-
sides share the C3′-endo sugar pucker and anti conformation
with ribonucleosides, they exhibit some key structural and
electronic differences that make them attractive for siRNA
applications. Sulfur is larger and less electronegative than
oxygen, and as such the C-S bond length is longer (1.82 Å),
the S-C-N anomeric effect is weaker, and the endocyclic
C-S-C angle is more acute in 4′-thioribonucleosides.33-35 The
lower electronegativity of the sulfur in 4′-thioribonucleosides
also causes a perturbation in the C-N glycosidic bond, altering
its chemical reactivity when compared to the C-N bond in
ribonucleosides.36 These properties presumably contribute to
their resistance to nucleoside phosphorylases and phosphatases.
However, the electronic environment of the nucleobases is not
significantly perturbed, and neither are their hydrogen-bonding
properties. 4′-Thioribonucleosides are also poor substrates for
Recent crystallography studies47,48 showing that the 5′-end
of the antisense strand interacts strongly with the RNase H like
RISC endonuclease highlight the importance of minimal modi-
fication at the 5′-end of the antisense strand. These observations
and the known intolerance for bulky substituents at the 5′-end
of the antisense strand, combined with the nuclease stability
benefit provided by 4′-thio-RNA, suggest that 4′-thio-RNA
could be a uniquely suited modification for improving the
pharmaceutical properties of siRNAs. Our experimental designs
were further based on reports suggesting that the RISC complex
is able to differentially select one of the two strands depending
on the duplex stability of the first several nucleotides.49,50 We
hypothesized that a similar bias should be possible using the
4′-thioribose modification in combination with bulkier 2′-O-
Me and especially 2′-O-MOE (Figure 1a) on the sense strand
in such a way that would lead to preferential loading of the
antisense strand.31 On the basis of these criteria, we have
designed, prepared, and evaluated a series of 4′-thio-RNA-
containing siRNAs for potency in reducing the target message
as well as nuclease stability.
While it remains to be seen with subsequent in vivo
pharmacokinetic and pharmacology studies, the incorporation
of significant numbers of 4′-thio-RNA residues into the potent
and stable constructs identified herein may also contribute to
improved tissue distribution required to achieve a robust
therapeutic effect in animals.
Results and Discussion
Synthesis of 4′-Thioribose Nucleosides and Phosphora-
midites. The synthesis of 4′-thioribonucleosides (Scheme 1) and

1626 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5
Dande et al.
Scheme 1a
Scheme 2a
a Reagents and conditions: (a) diethyl L-tartrate, titanium(IV) isopro-
poxide, tert-butyl hydroperoxide, CH2Cl2, -25 °C; (b) for 3, trimethylsilyl
triflate, triethylamine, 1:1 toluene/CH2Cl2, rt; for 4, trimethylsilyl triflate,
triethylamine, 1:1 1,2-dichloroethane/CH3CN, reflux; (c) (i) TBAF, AcOH,
THF; (ii) aqueous ammonia.
their phosphoramidites (Figure 1b) has been described previ-
ously.40,42,43,51,52 However, significantly improved yields were
obtained by making a key improvement in the synthetic
methodology, which is detailed below.
Arguably, the key step in the synthetic route is the oxidative
conversion of 1 to 2 (Scheme 1). The reported procedures for
converting 1 to 2 gave acceptable yields (60-80%) as a mixture
of (R)- and (S)-sulfoxides (R:S ) 2.7:1 with mCPBA, 16:1 by
ozonolysis). Unfortunately, only the R isomer undergoes
productive glycosylation under Pummerer conditions while the
S isomer undergoes elimination side reactions.51 Although
ozonolysis gave an improved ratio of R to S isomers, in our
hands this procedure gave consistently low yields (60%) and
was exceedingly difficult to scale up. We therefore investigated
alternate oxidation methods that could give us chirally pure (R)-2
and which could be rapidly and consistently scaled up. By using
a Di Furia-Modena oxidation53-56 (a mixture of tert-butyl
hydroperoxide, diethyl L-tartrate, and titanium(IV) isopro-
poxide), we were able to control the stereochemistry of the
oxidation product. Only the desired (R)-sulfoxide was obtained
in 90% isolated yield with no detectable S isomer (as indicated
a Reagents and conditions: (a) (TBS)Cl, imidazole, 4-(dimethylami-
no)pyridine, DMF; (b) 1,2,4-triazole, triethylamine, POCl3, CH3CN; then
NH4OH, 1,4-dioxane; (c) benzoic anhydride, pyridine; (d) triethylamine
trihydrofluoride, triethylamine, THF; (e) (TBS)Cl, AgNO3, pyridine, THF;
(f) 1% triethylamine in ethanol containing 0.1% imidazole; (g) 2-cyanoethyl
N,N,N′,N′-tetraisopropylphosphorodiamidite, 1H-tetrazole, N-methylimida-
zole, DMF.
and THF in the presence of silver nitrate to yield a mixture of
2′-and 3′-silyl derivatives (12/13). The two regioisomers were
separated using flash silica gel column chromatography. The
2′-O-TBS derivative 13 was converted to 3′-phosphoramidite
14 according to the reported procedure.59 Although this chemical
transformation has been reported for ribonucleosides,58 this is
the first report of this method being applied to 4′-thioribo-
nucleosides.
Target Gene and Sequence. siRNA duplexes targeting the
mature human PTEN messenger RNA (mRNA) were identified
by screening a series of siRNA constructs targeting the gene.
Unmodified siRNA-2 (Figure 2) is a 20-base-pair RNA duplex
which reduces endogenous PTEN mRNA levels with an IC50
≈ 0.9 nM when transfected into HeLa cells. siRNA-12 (Figure
3) is a 19-base-pair duplex which targets the same site on the
PTEN messenger RNA with an IC50 ≈ 0.4 nM.
1
by H NMR). Importantly, this procedure has been scaled up
to a 500 g scale with 92% isolated yields and without
chromatographic purification. When we used chirally pure (R)-2
as the sugar donor, the Pummerer thioglycosylation51 gave yields
of up to 97% for pyrimidines and up to 70% for purines. Thus,
stereoselective sulfur oxidation directly results in increased
overall synthetic yields and represents an important synthetic
improvement.
Compounds 5 and 6 (Scheme 1) were converted into their
corresponding phosphoramidites (Figure 1b, 7 and 8) via
standard procedures.52,57 The synthesis of 4′-thio-N4-benzoyl-
cytidine 3′-phosphoramidite was accomplished as described in
Scheme 2. Compound 9, a byproduct obtained40 during the
synthesis of 5′-O-(4,4′-dimethoxytrityl)-2′-O-(tert-butyldimeth-
ylsilyl)-4′-thiouridine, was silylated at the 3′-position with tert-
butyldimethylsilyl chloride [(TBS)Cl], imidazole, and pyridine
at room temperature to yield 10. Compound 10 was converted
to the triazolide derivative with 1,2,4-triazole, POCl3, and
triethylamine.58 It is important to note that the formation of the
triazolide is significantly slower (12 h) in the case of 4′-
thioribonucleosides as compared to ribonucleosides (1 h).
Subsequent treatment with aqueous ammonia/dioxane (1:1) at
ambient temperature for 18 h yielded a cytidine analogue. The
exocyclic amino group was protected with a benzoyl group,
followed by removal of the silyl group to give 11 (96%).
Compound 11 was treated with (TBS)Cl in a mixture of pyridine
siRNA Design. There is some evidence that different regions
of the siRNA play distinct roles in target recognition, cleavage,
and product release.50 Thus, the nature of the chemical
modifications as well as their placement within the siRNA may
influence RISC loading as well as subsequent processes in the
RNAi mechanism.2,4,6,7,27,28,50 We have recently reported on the
positional effects of 2′-F, 2′-O-Me, and 2′-O-MOE modifications
on siRNA-mediated target reduction in mammalian cells.31 Here
we have used a similar strategy to evaluate the positional effect
of 4′-thio modifications on siRNA activity. The 4′-thioribose,
2′-O-Me, or 2′-O-MOE nucleoside residues were systematically
introduced into the sense and/or antisense strands, and their

ImproVing RNA Interference in Mammalian Cells
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5 1627
Figure 3. Positional effects of 4′-thioribonucleoside residues in the
antisense strand on siRNA activity. Dose-dependent reduction of PTEN
mRNA in HeLa cells by siRNA gapmer duplexes containing 4′-
thioribonucleoside residues at the 3′- and 5′-termini of the antisense
strand. The 4′-thioribonucleoside residues are indicated in red. UTC
) untreated control.
in the 5′ to 3′ direction on the antisense strand (siRNA-3 to
siRNA-9, Figure 2). RNAi activity in reducing PTEN mRNA
levels was measured in HeLa cells (Figure 2). As a positive
control, siRNA-1 having a phosphorothioate antisense strand
paired with an unmodified sense strand was tested in parallel.
siRNA-1 is typically less potent than its phosphodiester
equivalent siRNA-2.
4′-Thioribose modifications are well tolerated at both the 3′-
and 5′-ends of the antisense strand (siRNA-3 and siRNA-9,
Figure 2). The 3′-end of the antisense strand has previously been
modified with a variety of chemical modifiers without any
observed loss in siRNA potency.20,26,31 Since the 3′-end of the
antisense strand is thought to play a minor role in siRNA
recognition by the RISC complex,7,50 it is not surprising that
4′-thioribose modifications as well as other chemical modifiers
are well tolerated in this region. However, the 5′-end of the
antisense strand was more sensitive to chemical modifica-
tions.19,25-29,31 The LNA, 2′-O-MOE, 2′-O-Me, and 2′-O-allyl
modifications when placed at the 5′-end of the antisense strand
resulted in a drop in siRNA activity. However, no such negative
effect on siRNA potency was observed with 4′-thioribose
residues at the 5′-end of the antisense strand (siRNA-9, Figure
2).
Figure 2. Positional effects of 4′-thioribonucleoside residues on siRNA
activity. Dose-dependent reduction of the PTEN mRNA level in HeLa
cells by siRNA duplexes containing 4′-thioribonucleoside residues in
the antisense strand of siRNA duplexes. The 4′-thioribonucleoside
residues are indicated in red. The underlined sequence has a phospho-
rothioate backbone. UTC ) untreated control.
effect on siRNA activity was studied as a function of their
position within the siRNA duplex. In initial experiments the
5′- end of the modified antisense strand was phosphorylated
chemically to generate a 5′-phosphorylated siRNA. This was
done to minimize the effect of chemical modification on siRNA
phosphorylation by cellular kinases (reported to be essential for
siRNA activity).1 However, we did not observe any difference
in siRNA activity when using 5′-phosphorylated or 5′-nonphos-
phorylated antisense strands, and subsequent studies were carried
out without 5′-phosphorylation.
Antisense Strand SAR. Although 4′-thio-RNA is globally
similar in structure to unmodified RNA, the substitution of
oxygen with sulfur32,33 may result in localized structural
perturbations. The endocyclic oxygen to sulfur substitution may
also influence the ability of siRNA duplexes to serve as
substrates for enzymes involved in eliciting siRNA activity.
These effects are likely to depend on not only the number of
modifications but also their position within the siRNA duplex.
Hence, the positional effect of 4′-thio modifications on siRNA
activity was studied. Optimal positioning of these residues within
the antisense strand of the siRNA was studied by walking them
The RISC complex has been shown to recognize the ends of
an siRNA duplex with a distinct 5′-antisense directionality,50
and the drop in siRNA potency observed when 5′-termini are
modified with 2′-ribose modifications has been attributed to
steric interference with either duplex recognition or loading by
the RISC complex.31 Crystallography studies of model antisense
strands interacting with proteins homologous to mammalian
Ago2 (the putative nuclease of the RISC complex) support this
model tolerating minimal modification at the 2′-position of 5′-

1628 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5
Dande et al.
terminal residues of the antisense strand.47,48 Furthermore, these
structural studies indicate a substantial conformational distortion
of the 5′-terminal end of the antisense strand must occur to
accommodate binding to the C-terminal domain of the Ago2
protein which anchors the strand. This explains, in part, the
reduced activity observed in certain circumstances, with multiple
5′-modifications such as 2′-F and LNA, which have a high
degree of conformational rigidity and would not be expected
to easily adopt the distorted conformation seen in the crystal
structures. Since siRNAs with 4′-thioribose modifications at the
5′-end of the antisense strand show no decrease in siRNA
activity, one infers that this modification is able to interact
productively with the nuclease component of mammalian RISC.
When 4′-thioribose residues were placed toward the interior
of the siRNA duplex, a slightly different picture emerged.
siRNA-5 (IC50 ) 0.48 nM, Figure 2) retains siRNA activity
and is 2-fold more potent than the control (siRNA-2, IC50
)
0.95 nM, Figure 2). This is in contrast to siRNA-4, which
contains only a one-base shift of the block of three 4′-thioribose
modifications and suffers a 15-fold loss in potency. siRNA-8
(IC50 ) 41.5 nM, Figure 2), which has residues 7-9 from the
5′-end of the antisense strand replaced with 4′-thioribonucleoside
residues, leads to substantial loss (40-fold) in activity.
Elbashir et al. have shown that the primary site of RISC-
mediated cleavage on the target mRNA was opposite the 10th
and 11th residues on the antisense strand.27 For the expected
A-form RNA helix that is formed between the siRNA antisense
strand and the target mRNA, the positions where 4′-thioribose
inhibits potency are not directly across the major or minor
groove from the expected cleavage site. Furthermore, they are
not on the side of the duplex that makes obvious direct contacts
with the Ago2 protein in a model derived from the crystal
structure of the bound antisense strand.48,62 It is therefore not
readily apparent why modifying positions 7-9 or 13-15 on
the antisense strand inhibits activity. It is conceivable that 4′-
thioribonucleotides might have a local geometry different from
that of their 4′-oxo counterparts in an siRNA duplex in a manner
not evident from the structural data available to date. However,
we did not observe any gross deviations from A-form helical
geometry (CD spectroscopy, data not shown), so it would have
to be a subtle effect. Also, Haeberli et al.63 have recently
described the crystal structure of an RNA duplex containing
4′-thiocytidine residues without noting any significant differ-
ences in helix conformation around the 4′-thioribose residues.
Also, there were no significant changes in the hydration of the
RNA backbone and the minor groove.
Figure 4. Positional effects of 4′-thioribonucleoside residues in the
sense and antisense strands on siRNA activity. Dose-dependent
reduction of PTEN mRNA in HeLa cells by siRNA duplexes containing
4′-thioribonucleoside residues in the sense strand. Underlined sequences
have a phosphorothioate backbone. The 4′-thioribonucloeside residues
are indicated in red. UTC ) untreated control.
of 4′-thioribose modifications can result in decreased siRNA
activity as has been reported previously,42 while optimized
placement can actually enhance siRNA potency as seen for
siRNA-5, siRNA-7, and siRNA-9.
Sense Strand SAR. siRNA duplexes containing 4′-thioribo-
nucleoside residues in the sense strand (siRNA-14 to siRNA-
19, Figure 4) show good activity in reducing PTEN mRNA
levels in HeLa cells. Dispersing these residues in the interior
of the sense strand (siRNA-14) has a slight negative effect on
siRNA potency (IC50 ) 1.1 nM vs IC50 ) 0.4 nM). Pairing
Next we placed 4′-thioribose residues at both the 3′- and 5′-
ends of the antisense strand. This is referred to as the “gapmer”
design represented by siRNA-10 and siRNA-13 (Figure 3). The
gapmer design was chosen as it is used extensively in antisense
therapeutics to protect the ends of the oligonucleotide from
exonuclease cleavage.64 siRNA-10 and siRNA-13 (IC50 ) 5.9
and 1.2 nM, respectively, Figure 3) exhibit respectable siRNA
activity. However, siRNA-10, which has four 4′-thioribonucleo-
side residues at each end of the antisense strand, is about 6-fold
less potent than the parent siRNA-2 (IC50 ) 0.95 nM, Figure
3). siRNA-13, which has only two 4′-thioribonucleoside residues
at both ends, is only about 3-fold less potent than the unmodified
siRNA-12 (IC50 ) 0.4 nM), indicating a definite inverse
relationship between the number of 4′-thioribose modifications
in the antisense strand and siRNA activity. Taken together, these
results suggest that not only the number but also the placement
of 4′-thioribose modifications within the antisense strand has
implications for siRNA activity. Simply increasing the number
with a phosphorothioate antisense strand (siRNA-16, IC50
)
2.4 nM) causes a further loss in potency as observed previously
(Figure 2, siRNA-1, IC50 ) 1.6 nM, vs siRNA-2, IC50 ) 0.95
nM). Activity is regained when the dispersed motif is replaced
with the gapmer motif (siRNA-15, IC50 ) 0.28 nM), once again
highlighting the utility of an optimized gapmer design. This
trend is confirmed when one compares siRNA-16 (IC50 ) 2.4
nM) with siRNA-17 (IC50 ) 1.4 nM). siRNA-19, in which both
the 3′- and 5′-termini of the sense and the antisense strands are
made up of 4′-thioribonucleoside residues (gapmer siRNA), is
one of the most potent siRNAs in this study, having an IC50 of
0.21 nM. We attribute this to the stabilization of the ends of
the siRNA toward exonuclease degradation.
Combined Effect of 4′-Thioribose and 2′-Ribose Modifica-
tions on siRNA Activity. During the course of designing 4′-

ImproVing RNA Interference in Mammalian Cells
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5 1629
thioribose-modified motifs, we soon realized that we would not
achieve the desired degree of serum stability by modifying only
one strand at a time. We believe the construct designs reported
recently43 that have only one strand modified will face similar
limitations in therapeutic application. Our preliminary experi-
ments showed that the best motif for 4′-thioribose modification
is the one that will have both strands end modified (Figure 4).
As both our data and the literature report43 indicate that there
is a threshold of the number of 4′-thioribose residues in an
siRNA duplex, we envisaged that combining other nuclease-
resistant RNA modifications with 4′-thio-RNA should lead to
superior siRNA designs. The 2′-modified RNAs are known to
be resistant to nucleases and as such have been candidates for
siRNA applications. We have recently reported an extensive
SAR of 2′-ribose-modified siRNAs containing nuclease-resistant
2′-O-Me and 2′-O-MOE residues in the sense and/or antisense
strands, where we showed that both modifications were well
tolerated at the 3′- and 5′-ends of the sense strand of siRNA.31
In the same report, we also demonstrated that both these
modifications were not well tolerated at the 5′-end of the
antisense strand while the 2′-O-Me modification was better
tolerated at the 3′-end of the antisense strand. These data suggest
that using bulkier 2′-substituted ribose residues in the sense
strand or 3′-end of the antisense strand should not interfere with
siRNA activity. This design should also lead to a favorable bias
for loading the guide antisense strand into the RISC.
The placement of bulky residues in combination with the 4′-
thioribose modification was optimized in the designs shown in
Figure 5. As expected, siRNA-21 through siRNA-25 generally
showed potent siRNA activity, demonstrating the compatibility
of the 4′-thioribose with 2′-ribose modifications. siRNA-21 (IC50
) 0.66 nM) and siRNA-22 (IC50 ) 0.78 nM) show slightly
reduced potency relative to unmodified siRNA-12 (IC50 ) 0.4
nM). Next, by optimizing the placement of bulky 2′-modified
residues in combination with the 4′-thioribose modification, we
introduced minimal steric bulk at the 5′-end of the antisense
strand while increasing it at the 3′-end (siRNA-23 to siRNA-
25, Figure 5). These designs showed improved potency over
the parent siRNA (siRNA-12), exhibiting a distinct additive
effect of these designs on siRNA activity. siRNA-24 showed
reduced potency (IC50 ) 1.9 nM), which we attribute to a
combination of reduced nuclease resistance resulting from the
exposed unmodified ends of the sense strand. Thus, the
4′-thioribose modification is generally well tolerated in both
the sense and antisense strands of siRNA duplexes, and is also
compatible with other nuclease-stabilizing 2′-modifications.
Importantly, by using optimized chemically modified siRNA
constructs (siRNA-23 and siRNA-25, Figure 5), we can perhaps
achieve a degree of selectivity in the RISC strand-loading
process which results in enhanced siRNA potency.
Figure 5. Effects of 4′-thioribonucleoside residues in combination with
2′-modifications on siRNA activity. Dose-dependent reduction of PTEN
mRNA in HeLa cells by siRNA duplexes containing 4′-thioribose
residues in combination with 2′-O-Me- or 2′-O-MOE-modified residues.
The 4′-thioribose, 2′-O-Me-modified, and 2′-O-MOE-modified nucleo-
side residues are indicated in red, blue (italics), and orange (italics),
respectively. UTC ) untreated control.
2′-O-MOE-modified sense strand at the 3′- and 5′-termini
exhibited improved thermal stability (Tm ) 74.3 °C, Table 1)
relative to the unmodified siRNA-12 (Tm ) 72.8 °C, Table 1).
Interesting enough, the thermal stability was further enhanced
when 4′-thioribonucleoside substitutions were combined with
2′-substitutions, as in siRNA-21 and siRNA-25 (Tm ) 75.5 and
75.4 °C, Table 1). Thus, the thermal stabilizing effect of the
4′-thioribonucleoside residues works additively with 2′-
modifications. It is especially interesting in light of the literature,
which reports increased activity of duplexes having a destabi-
lized 5′-antisense strand end, to note that our duplexes having
increased overall Tm do not show decreased potency. We made
no attempt to intentionally destabilize one end of the duplex,
as we felt that added thermal stability may also be an important
factor in the stability of the duplex to degradation by nucleases,
as the duplex appears to aid in protection from degradation by
single-stranded nucleases. It is reasonable to predict that
increased rates of “breathing” by the ends of the duplex would
increase degradation as the single-strand region of RNA is
exposed.
Thermal Stability of 4′-Thioribose-Modified siRNA Du-
plexes. Because of the known increase in the thermal stability
of 4′-thio-RNA-containing duplexes, and the literature reports
suggesting that decreasing the stability of the end of the duplex
containing the 5′-end of the antisense strand increases its
propensity to be loaded into the RISC complex,49,50 we evaluated
the thermal and plasma stability of selected siRNA duplexes
(Table 1). The 4′-thioribonucleoside residues caused a slight
increase in the thermal stability of siRNA duplexes (0.3 °C per
residue) when present either in the sense or antisense strands.
This is in agreement with the previously reported thermody-
namic stability of 4′-thio-RNA duplexes.38-42 The 2′-O-MOE
modification is known to provide a significant enhancement in
thermal stability to oligonucleotides. Indeed, siRNA-20 with a
Plasma Stability of 4′-Thioribose-Modified siRNAs. As
expected, we found that chemical optimization led to significant
improvement of the plasma stability profiles for this series
(Table 1) of siRNAs. On the basis of their respective half-lives

1630 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5
Dande et al.
Table 1. Thermal and Plasma Stability and siRNA Activity of Selected
siRNA Duplexes Containing 4′-Thioribose (Red) Residues in
Combination with 2′-O-Me-Modified (Blue Italics) or
2′-O-MOE-Modified (Orange Italics) Residuesa
siRNAs containing 4′-thioribonucleoside substitutions were
highly effective in reducing target PTEN mRNA in HeLa cells.
Moreover, we showed that, by carefully positioning and/or
combining 4′-thioribonucleosides with 2′-ribose modifications,
siRNA potency could be increased beyond that of unmodified
siRNAs (siRNA-7, siRNA-9, siRNA-19, siRNA-23, and siRNA-
25 are 2-4-fold more potent than the corresponding unmodified
siRNAs). Importantly, the improved potency of siRNA-23 and
siRNA-25 demonstrates that modified chemistries with dif-
ferential tolerances for substitutions at key positions in the
duplex such as the 5′-end of the antisense strand can be
employed in consideration of the biased strand-loading hypoth-
esis reported for the RISC complex49,50 to rationally design
potent siRNAs.
Our results also demonstrate the success of our positional
SAR strategy in identifying plasma-stable siRNAs. We have
confirmed that 4′-thioribose modifications when not optimally
placed lead to a moderate increase in plasma stability (siRNA-
18). Using a rational design strategy (gapmer design, siRNA-
19) and a combination of modified chemistries (siRNA-21 and
siRNA-25), we have identified constructs that are simultaneously
2-4-fold more potent and 13-18-fold more stable than
unmodified siRNAs. This observed plasma stability and potent
siRNA activity of our modified siRNA constructs will be
valuable when elucidating an SAR for systemically administered
siRNAs in vivo, where unmodified siRNAs are likely to be too
unstable to be therapeutically viable. The potency, plasma
stability, and compatibility of the 4′-thioribose modification with
other chemical modifications demonstrated herein provide
opportunities to address these questions in animal studies to
further optimize siRNAs for therapeutic applications.
Importantly, the 4′-thioribose modification adequately com-
pensates for the inability to utilize highly stabilizing 2′-ribose
modifications at the 5′-end of the antisense strand. Furthermore,
the incorporation of sulfur into the siRNA duplex without
incurring the loss of potency observed with phosphorothioate
siRNAs in a non-phosphorothioate context raises the possibility
of influencing tissue distribution and in vivo potency in much
the same way as sulfur-containing antisense oligonucleotides.
Further studies to fully exploit these potential benefits of 4′-
thio-siRNAs are currently under way in our laboratory, and the
results will be published in due time.
a Notes: *, siRNA duplexes, sense strand 5′ to 3′, antisense strand 3′ to
5′; a, t1/2 ) plasma half-life; thermal melting and plasma (25% heparinized
mouse plasma) stability studies were carried out as described in the
Experimental Section.
in 25% heparinized mouse plasma, siRNA-18, siRNA-19,
siRNA-21, and siRNA-25 were approximately 2.5-, 14-, 13-,
and 18-fold more resistant to degradation than unmodified
siRNA-12. This observation is in agreement with the known
nuclease stability of 4′-thioribonucleotides.38-42 The need to
protect the ends of individual strands is obvious when one
compares siRNA-18 and siRNA-19. siRNA-18, in which the
unmodified 3′- and 5′-ends of the sense strand are exposed to
serum nucleases, has a half-life of 2.5 h. When these exposed
ends are blocked by incorporation of 4′-thioribose residues
(siRNA-19), the plasma half-life increases to 14 h. It should
however be noted that siRNA-18 is still 2.5 times more stable
in mouse plasma than unmodified siRNA-12 probably due to
stabilization of the duplex toward endonucleolytic cleavage.
These differences in plasma half-life between siRNA-19 and
siRNA-18 indicate that exonucleases and not endonucleases are
primarily responsible for siRNA degradation in mouse plasma.
In addition to the benefit provided by 4′-thioribose “gapmers”
in each strand, the strategic placement of 4′-thioribose and 2′-
O-Me and 2′-O-MOE modifications also greatly enhanced the
plasma stability of siRNA. These modifications, when used in
combination at the ends of the duplex (siRNA-21 and siRNA-
25), show significant gains in plasma stability over unmodified
control siRNA-12. These data support the hypothesis that
degradation in plasma is occurring by nucleases which attack
the ends of the construct most likely when they are exposed as
single strands due to breathing of the duplex. It is remarkable
that as few as two 4′-thioribonucleoside residues at each end
of both strands provide greatly enhanced protection from
nucleolytic degradation without compromising the intrinsic
potency of the siRNA mechanism.
Experimental Section
General Procedures. Solvents used were of anhydrous grade
and were stored under nitrogen at all times. The RNA phosphora-
midites with a 2′-O-(tert-butyldimethylsilyl) protecting group and
reagents were procured form Glen Research Inc., Virginia. All other
starting materials and reagents were purchased from Aldrich
Chemical Co. and were used without further purification. Thin-
layer chromatography was performed on precoated plates (silica
gel 60 F254, EM Science, New Jersey) and visualized with UV
light and spraying with a solution of p-anisaldehyde (6 mL), H2-
SO4 (8.3 mL), and CH3COOH (2.5 mL) in C2H5OH (227 mL)
followed by charring.1H NMR spectra were referenced using
internal standard (CH3)4Si and 31P NMR spectra using external
standard 85% H3PO4. Mass spectra were recorded by Mass
Consortium, San Diego, CA, and the College of Chemistry,
University of California, Berkeley, CA.
(R)-1,4-Anhydro-2-O-(2,4-dimethoxybenzoyl)-3,5-O-(1,1,3,3-
tetraisopropyldisiloxane-1,3-diyl)-4-sulfinyl-D-ribitol (2). Diethyl
L-tartrate (60 mL, 200.16 mmol) was added to a vigorously stirring
solution of titanium(IV) isopropoxide (30 mL, 50.2 mmol) in dry
dichloromethane (200 mL), resulting in a straw-yellow-colored
solution. After being stirred at room temperature, the solution was
cooled to -20 °C, and tert-butyl hydroperoxide (40 mL of a 6.0
Conclusion
We have evaluated the effect of the 4′-thioribonucleoside
modification on the RNAi activity of siRNAs, individually and
in combination with 2′-O-Me and 2′-O-MOE modifications. The

ImproVing RNA Interference in Mammalian Cells
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5 1631
M solution in decane) was added. After a few minutes a solution
of 1 (70 g, 126 mmol) in dry dichloromethane (200 mL) was added.
Stirring was continued at the same temperature for 18-24 h till
TLC showed no presence of 1. The reaction was quenched by the
addition of water and allowed to come to room temperature. The
resulting white precipitate was filtered through a Celite plug, and
the cake was washed exhaustively with pentanes. The filtrate was
transferred to a separating funnel, washed three times with water
and once with brine, and dried over anhydrous sodium sulfate.
Solvents were removed under reduced pressure, and the residue
was purified by flash chromatography (1:1 ethyl acetate/hexanes)
to give 2 (66.4 g, 116 mmol) in 92% isolated yield as a single
isomer: 1H NMR (200 MHz, CDCl3) δ 7.92 (d, 1 H, J ) 8.3 Hz),
6.49 (m, 2 H), 5.79 (dd, 1 H, J1 ) 5.4 Hz, J2 ) 3.6 Hz), 4.59 (d,
1 H, J ) 12.8 Hz), 4.22 (dd, 1 H, J1 ) 2.8 Hz, J2 ) 12.8 Hz) 4.12
(dd, 1 H, J1 ) 3.6 Hz, J2 ) 12.0 Hz), 3.88 (s, 3 H), 3.86 (s, 3 H),
3.57 (dd, 1 H, J1 ) 5.4 Hz, J2 ) 15.5 Hz), 3.49 (dd, 1 H, J1 )
12.0 Hz, J2 ) 2.8 Hz), 2.89 (d, 1 H, J ) 15.5 Hz), 1.11-0.94 (m,
28 H); 13C NMR (50 MHz, CDCl3) δ 164.5, 164.1, 161.6, 134.0,
111.7, 104.5, 98.8, 77.2, 72.5, 72.4, 67.9, 55.8, 55.4, 54.3, 17.2,
17.1, 17.1,17.0, 17.0, 16.8, 16.8, 13.28, 13.0, 12.6, 12.5; HRMS
phosphorus pentoxide. The residue was then dissolved in anhydrous
tetrahydrofuran (100 mL). Triethylamine (10 mL) and triethylamine
trihydrofluoride (20 mL) were added, and the reaction mixture was
stirred overnight at room temperature. The reaction mixture was
diluted with ethyl acetate (300 mL) and transferred to a separating
funnel. The ethyl acetate layer was washed with saturated sodium
bicarbonate solution (3 × 100 mL) and brine (10 mL) and dried
over anhydrous sodium sulfate. The residue obtained was purified
by flash silica gel column chromatography (5% methanol, 0.1%
triethylamine in dichloromethane) to yield 11 (13 g, 96% yield):
1H NMR spectra in agreement with published literature values;40
HRMS (FAB) m/z calcd for C37H36N3O7S+ 666.2196, found
666.2204.
N4-Benzyol-1-[2-O-(tert-butyldimethylsilyl)-5-O-(4,4′-dimethox-
ytrityl)-4-thio-â-D-ribofuranosyl)]cytosine (13). Compound 11 (10
g, 15 mmol) was dissolved in anhydrous tetrahydrofuran (150 mL)
containing pyridine (3 mL). Silver nitrate (3.2 g, 18.75 mmol) was
added and the reaction mixture stirred for 30 min at room
temperature under an inert atmosphere. tert-Butyldimethylsilyl
chloride (2.83 g, 18.75 mmol) was added, and the reaction mixture
was stirred overnight at room temperature. The reaction mixture
was filtered through a Celite plug and partitioned between ethyl
acetate (300 mL) and saturated sodium bicarbonate solution (100
mL). The ethyl acetate layer was washed with brine (10 mL) and
dried over anhydrous sodium sulfate. Solvent was removed under
reduced pressure, and the residue was purified by flash silica gel
column chromatography (25% ethyl acetate, 0.1% triethylamine in
dichloromethane) to yield 12 (5.5 g, 47%) and 13 (6.0 g, 51.3%).
The isolated 12 was converted to 13 by being stirred in dry ethanol
(50 mL) containing 1% triethylamine and 0.1% imidazole to give
a 1:1 mixture of 12 and 13. The desired isomer 13 was separated
by flash silica gel column chromatography as before, and the
procedure was repeated till almost all of 12 was converted to 13
(9.95 g) with a final isolated yield of 85%: 1H NMR spectra in
agreement with published literature values;41,44 HRMS (FAB) m/z
calcd for C43H50N3O7SSi+ 780.3060, found 780.3143 [M + H]+.
N4-Benzyol-1-[2-O-(tert-butyldimethylsilyl)-3-O-[(2-cyano-
ethyl)(N,N-diisopropylamino)phosphino]-5-O-(4,4′-dimethoxy-
trityl)-4-thio-â-D-ribofuranosyl)]cytosine (14). Compound 14 was
synthesized using a reported procedure.59 Compound 13 (5.4 g, 6.93
mmol) and tetrazole (294 mg, 1.35 mmol) were dissolved in
anhydrous DMF (40 mL) under an inert atmosphere. N-Methylimi-
dazole (107 µL, 1.35 mmol) and 2-cyanoethyl N,N,N′,N′-tetraiso-
propylphosphorodiamidite (3.3 mL, 10.4 mmol) were added. The
reaction mixture was stirred overnight at room temperature and
partitioned between ethyl acetate (200 mL) and saturated sodium
bicarbonate solution (50 mL). The ethyl acetate layer was washed
with water (2 × 300 mL) and brine (10 mL) and dried over
anhydrous sodium sulfate. The solvent was removed under reduced
pressure and the residue purified by flash silica gel column
chromatography (25% ethyl acetate, 0.1% triethylamine in dichlo-
romethane) to yield 14 (6.4 g, 94%): 1H and 31P NMR spectra
were agreement with published literature values;40,43 HRMS (FAB)
m/z calcd for C52H67N5O8PSSi+ 980.4139, found 980.4199.
RNA Synthesis. RNA oligonucleotides were synthesized on a
solid-phase DNA/RNA synthesizer using the 2′-O-TBS-RNA
phosphoramidites (TBS ) tert-butyldimethylsilyl) according to the
reported protocols.38 The 4′-thio, 2′-O-Me, and 2′-O-MOE phos-
phoramidites with exocyclic amino groups protected with benzoyl
(Bz for A and C) or isobutyryl (ibu for G) were used for the
synthesis of the RNA chimera. A 0.12 M solution of the phos-
phoramidites in anhydrous acetonitrile was used for the synthesis.
Oxidation of the internucleosidic phosphite to the phosphate was
carried out using tert-butyl hydroperoxide/acetonitrile/water (10:
87:3) with a 10 min oxidation time. 3H-1,2-Benzodithiol-3-one 1,1-
dioxide57 (Beaucage reagent, 0.15 M solution in anhydrous
acetonitrile) was used as the sulfur-transfer agent for the synthesis
of oligoribonucleotide phosphorothioates. Chemical phosphorylation
reagent procured from Glen Research Inc. was used to phospho-
rylate the 5′-terminus of modified oligonucleotides. Samples of 12
equiv of the phosphoramidite solutions were delivered in two
+
(FAB) m/z calcd for C26H45O8SSi2 572.2295, found 573. 2296.
1-[2,3-O-Bis(tert-butyldimethylsilyl)-5-O-(4,4′-dimethoxytri-
tyl)-4-thio-â-D-ribofuranosyl)]uracil (10). 941 (9.2 g, 13.6 mmol)
and imidazole (6 g, 81.6 mmol) were dissolved in anhydrous
pyridine (30 mL) and the resulting solution cooled in an ice bath.
(TBS)Cl (4.1 g, 27.2 mmol) was added and the reaction mixture
stirred at room temperature for 24 h, at which point no more starting
material was detected by TLC (SiO2, 25% ethyl acetate in
dichloromethane). The reaction was quenched by addition of ice
and partitioned between ethyl acetate (200 mL) and water (200
mL). The ethyl acetate layer was washed with saturated sodium
bicarbonate solution (3 × 100 mL), ice-cold 10% citric acid solution
(100 mL), and brine (20 mL). After the layer was dried over
anhydrous sodium sulfate, the solvent was removed under reduced
pressure and the residue was dried to a white foam in vacuo over
phosphorus pentoxide. This material was used as such in the next
step without further purification.
N4-Benzoyl-1-[-5-O-(4,4′-dimethoxytrityl)-4-thio-â-D-ribofura-
nosyl)]cytosine (11). In an oven-dried reaction vessel, 1,2,4-triazole
(19.6 g, 283.36 mmol) was suspended in anhydrous acetonitrile
(200 mL) and the resulting suspension cooled to 0 °C in an ice
bath. Phosphorus oxychloride (7.55 mL, 80.96 mmol) was added
dropwise to the suspension followed by triethylamine (56.5 mL,
404.8 mmol) as a solution in dry acetonitrile (100 mL). The
resulting suspension was stirred at 0 °C for 30 min. Compound 10
(16 g, 20.24 mmol) was dissolved in anhydrous acetonitrile (100
mL) and the resulting solution added dropwise to the reaction
mixture. The resulting pale yellow suspension was stirred overnight
at room temperature under an inert atmosphere. The reaction
mixture was then partitioned between ethyl acetate (500 mL) and
saturated sodium bicarbonate solution (500 mL). The ethyl acetate
layer was washed with brine (20 mL) and dried over anhydrous
sodium sulfate. The solvents were removed under reduced pressure,
and the residue was dissolved in 1,4-dioxane (100 mL). Aqueous
ammonia (28-30 wt % solution, 50 mL) was added, and the
mixture was stirred in a sealed flask overnight. The reaction mixture
was partitioned between ethyl acetate (300 mL) and water (500
mL). The ethyl acetate layer was washed with brine (20 mL) and
dried over anhydrous sodium sulfate. The solvent was removed
under reduced pressure, and the residue was dried overnight in
vacuo over phosphorus pentoxide. To this was added anhydrous
pyridine (50 mL), followed by benzoic anhydride (9.15 g, 40.48
mmol), and the mixture was stirred under an inert atmosphere
overnight. The reaction mixture was partitioned between ethyl
acetate (300 mL) and saturated sodium bicarbonate solution (200
mL). The ethyl acetate layer was washed twice with saturated
sodium bicarbonate solution (2 × 200 mL), once with ice-cold 10%
citric acid solution (100 mL), and with brine (20 mL) and dried
over anhydrous sodium sulfate. The solvent was removed under
reduced pressure, and the residue was dried in vacuo over

1632 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5
Dande et al.
portions, each followed by a 6 min coupling wait time. All other
steps in the protocol supplied by the manufacturer were used without
modification. The stepwise coupling efficiencies were more than
97%. After completion of the synthesis, the solid support was treated
with triethylamine/acetonitrile (1:1) for 30 min. Then it was
suspended in a mixture of aqueous ammonium hydroxide (28-30
wt %)/ethanol (3:1) and the suspension heated at 55 °C for 6 h to
complete the removal of all protecting groups except the TBS group
at the 2′-position. The solid support was filtered, and the filtrate
was concentrated to dryness. The residue obtained was resuspended
in anhydrous triethylamine trihydrofluoride/triethylamine/1-methyl-
2-pyrrolidinone solution (0.75 mL of a solution of 1 mL of
triethylamine trihydofluoride, 750 µL of triethylamine, and 1.5 mL
of 1-methyl-2-pyrrolidine, to provide a 1.4 M HF concentration)
and the suspension heated at 65 °C for 1.5 h to remove the TBS
groups at the 2′-position.36 The reaction was quenched with 1.5 M
ammonium bicarbonate (0.75 mL), and the mixture was loaded onto
a Sephadex G-25 column (NAP Columns, Amersham Biosciences
Inc.). The oligonucleotides were eluted with water, and the fractions
containing the oligonucleotides were pooled together and purified
by high-performance liquid chromatography (HPLC) on a strong
anion exchange column (Mono Q, Pharmacia Biotech, 16/10, 20
mL, 10 µm, ionic capacity 0.27-0.37 mmol mL-1, solvent A )
100 mM ammonium acetate, 30% aqueous acetonitrile, solvent B
U92436.1), forward primer 5′-AATGGCTAAGTGAAGATGAC
AATCAT, reverse primer 5′-TGCACATATCAT TACACCAGT-
TCGT, and FAM/TAMRA probe 5′-TTGCAGCAATTCACTG-
TAAAG CTGGAAAGG. Total RNA for each well was measured
using RiboGreen (Molecular Probes, Eugene, OR), and these values
were used for sample to sample normalization. IC50 values were
calculated from the linear regression analysis from the plot of the
logarithmic value of the siRNA concentration versus the percentage
of the untreated control.
Thermal Stability Determination. The thermal stability of the
modified siRNA duplexes was studied by measuring the UV
absorbance versus temperature curves as described previously.65,66
Each sample contained 100 mM Na+, 10 mM phosphate, 0.1 mM
EDTA, and 4 µM sense and 4 µM antisense strands. Each Tm
reported was an average of values from two experiments.
Plasma Stability Determination. Investigated siRNA constructs
have been incubated at a concentration of 10 µM in 25% heparinized
mouse plasma at 37 °C (Charles River Lab, Wilmingten, MA).
Orotic acid was added as an internal standard (200 µM, Sigma, St.
Louis, MO). At the indicated time points 10 µL aliquots were
removed and quenched by treatment with proteinase K (Sigma) at
45 °C for 60 min. Samples were stored on ice until analysis by
CGE utilizing a P/ACE MDQ instrument with UV detection
(Beckman, Fullerton, CA). For CGE analysis the ssDNA 100-R
Kit (Beckman) was used with the exception that no urea was added
to the buffer. To increase sample loading, a sample-stacking
procedure was performed by injecting water for 10 s at 10 psi prior
to sample injection. Peak areas of the siRNA construct and orotic
acid have been determined for the different time points and
normalized to the peak area of orotic acid to the zero time point.
) 1.5 M NaBr in A, 0-40% B in 60 min, flow rate 1.5 mL min-1
,
λ ) 260 nm). Fractions containing full-length oligonucleotides were
pooled together (assessed by capillary gel electrophoresis (CGE)
analysis, >90%) and evaporated. The residue was dissolved in
sterile water (0.3 mL), absolute ethanol (1 mL) was added, the
resulting mixture was cooled to -20 °C for 1 h, and the precipitate
formed was pelleted out by centrifugation (NYCentrifuge 5415C,
Eppendorf, Westbury, NY) at 3000 rpm. The supernatant was
decanted, and the pellet was redissolved in 5 M ammonium acetate
(0.3 mL) solution. Ethanol (1 mL) was added, the resulting mixture
cooled to -20 °C for 1 h to get a precipitate, and the precipitate
pelleted out in a centrifuge (NYCentrifuge 5415C, Eppendorf) at
3000 rpm for 15 min. The pellet was collected by decanting the
supernatant. The pelleted oligonucleotides were redissolved in sterile
water (0.3 mL) and precipitated by adding ethanol (1 mL) and
cooling the mixture at -20 °C for 1 h. The precipitate formed was
pelleted out and collected as described above. The oligonucleotides
were characterized by ES MS, and the purity was assessed by
capillary gel electrophoresis and HPLC (Waters, C-18, 3.9 × 300
mm, solvent A ) 100 mM triethylammonium acetate, pH 7, solvent
B ) acetonitrile, 5-60% B in 40 min, flow rate 1.5 mL min-1, λ
) 260 nm).
Acknowledgment. We thank Dr. Susan Freier and Dr.
Charles Allerson for helpful discussions and Lisa Caralli for
CD analysis of modified siRNA duplexes.
Supporting Information Available: 31P NMR and HRMS
(FAB) mass spectra of compounds 7, 8, and 14 and ES MS and
CGE profiles of selected modified oligonucleotides. This material
is available free of charge via the Internet at http://pubs.acs.org.
References
(1) Elbashir, S. M.; Lendeckel, W.; Tuschl, T. RNA Interference is
Mediated by 21- and 22-nucleotide RNAs. Genes DeV. 2001, 188-
200.
(2) Elbashir, S. M.; Harborth, J.; Lendeckel, W.; Yalcin, A.; Weber, K.;
Tuschl, T. Duplexes of 21-Nucleotide RNAs Mediate RNA Interfer-
ence in Cultured Mammalian Cells. Nature 2001, pp 494-498.
(3) Montgomery, M. K.; Fire, A. Double-stranded RNA as a Mediator
in Sequence-Specific Genetic Silencing and Co-Suppression. Trends
Genet. 1998, 14, 255-258.
(4) Tuschl, T.; Zamore, P. D.; Lehmann, R.; Bartel, D. P.; Sharp, P. A.
Targeted mRNA Degradation by Double-Stranded RNA In Vitro.
Genes DeV. 1999, 3191-3197.
(5) Tuschl, T. RNA Interference and Small Interfering RNAs. Chem-
BioChem 2001, 239-245.
(6) Tuschl, T.; Elbashir, S.; Lendeckel, W. (Max-Planck-Gesellschaft
Zur Foerderung Der Wissenschaften E.V., Germany). Analysis of
Sequence and Structural Requirements for Small RNAs Mediating
RNA Interference and Therapeutic Uses of Interfering RNA. PCT
Int. Appl. WO 0244321, 2002; 89 pp.
(7) Zamore, P. D.; Tuschl, T.; Sharp, P. A.; Bartel, D. P. RNAi: Double-
Stranded RNA Directs the ATP-Dependent Cleavage of mRNA at
21 to 23 Nucleotide Intervals. Cell 2000, 25-33.
(8) Elbashir, S. M.; Harborth, J.; Weber, K.; Tuschl, T. Analysis of Gene
Function in Somatic Mammalian Cells Using Small Interfering RNAs.
Methods (San Diego, CA, U.S.) 2002, 26, 199-213.
(9) Wu, H.; Macleod, A. R.; Lima, W. F.; Crooke, S. T. Identification
and Partial Purification of Human Double Strand RNase Activity. A
Novel Terminating Mechanism for Oligoribonucleotide Antisense
Drugs. J. Biol. Chem. 1998, 273, 2532-2542.
(10) Vickers, T. A.; Koo, S.; Bennett, C. F.; Crooke, S. T.; Dean, N. M.;
Baker, B. F. Efficient Reduction of Target RNAs by Small Interfering
RNA and RNase H-dependent Antisense Agents. A Comparative
Analysis. J. Biol. Chem. 2003, 278, 7108-7118.
Assembly of Modified siRNA Duplexes. Equimolar amounts
of sense and antisense strands were mixed together and annealed
by heating the mixture at 90 °C for 1 min and subsequently
incubating it at 37 °C for 1 h. Successful duplex formation was
confirmed by CGE (Beckman, MDQ CE).
Cell Culture and Transfection of Cells. HeLa cells (American
Type Tissue Culture Collection, Manassas, VA) were cultured in
a culture flask in Dulbecco’s modified Eagle’s medium (DMEM;
Invitrogen, California), liquid (high glucose) supplemented with
10% heat-inactivated fetal bovine serum (FBS). The cells were not
allowed to exceed 75-80% confluency. Prior to treatment (24 h
before treatment), the cells were detached from the flask using
trypsin (Invitrogen) and plated in 96-well plates at a density of
5000 cells/well. The cells were transfected with siRNAs complexed
with 6 mg mL-1 Lipofectin (Invitrogen) in serum-free Opti-MEM
I reduced-serum medium. The cells were incubated in the trans-
fection medium for 4 h, the transfection medium was removed from
the cells and replaced with fresh DMEM, 10% fetal calf serum,
and the cells were incubated at 37 °C and 5% CO2 for 16 h.
RNA Expression Analysis. Total RNA was harvested after 16
h using the RNeasy process from Qiagen (Valencia, California)
according to the manufacturer’s protocol. Gene expression was
determined via real time quantitative RT-PCR on the ABI Prism
7900 system (Applied Biosystems, Foster City, CA). The following
primer probe set (Qiagen) was used: hu PTEN (accession no.

ImproVing RNA Interference in Mammalian Cells
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5 1633
(11) Zender, L.; Huetker, S.; Liedtke, C.; Tillmann, H. L.; Zender, S.;
Mundt, B.; Waltemathe, M.; Goesling, T.; Flemming, P.; Malek, N.
P.; Trautwein, C.; Manns, M. P.; Kuehnel, F.; Kubicka, S. Caspase
8 Small Interfering RNA Prevents Acute Liver Failure in Mice. Proc.
Natl. Acad. Sci. U.S.A. 2003, 100, 7797-7802.
(12) Song, E.; Lee, S. K.; Wang, J.; Ince, N.; Ouyang, N.; Min, J.; Chen,
J.; Shankar, P.; Lieberman, J. RNA Interference Targeting Fas
Protects Mice From Fulminant Hepatitis. Nat. Med. 2003, 9, 347-
351.
(32) Prakash, T. P.; Manoharan, M.; Kawasaki, A. M.; Fraser, A. S.;
Lesnik, E. A.; Sioufi, N.; Leeds, J. M.; Teplova, M.; Egli, M. 2′-
O-[2-(Methylthio)Ethyl]-Modified Oligonucleotide: An Analogue of
2′-O-[2-(Methoxy)-Ethyl]-Modified Oligonucleotide With Improved
Protein Binding Properties and High Binding Affinity to Target RNA.
Biochemistry 2002, 41, 11642-11648.
(33) Crnugelj, M.; Dukhan, D.; Barascut, J.-L.; Imbach, J.-L.; Plavec, J.
How S-C-N Anomeric Effects and Energetic Preference Across [S-C-
C-O] Fragments Steer Conformational Equilibria in 4′-Thionucleo-
1
sides. H NMR and Ab Initio MO Study. Perkin 2 2000, 255-262.
(13) Sorensen, D. R.; Leirdal, M.; Sioud, M. Gene Silencing by Systemic
Delivery of Synthetic siRNAs in Adult Mice. J. Mol. Biol. 2003,
327, 761-766.
(34) Crnugelj, M.; Plavec, J. NMR Conformational Study Reveals That
S-C-N Anomeric Effect in Thionucleosides is Weaker Than O-C-N
Anomeric Effect in Natural Nucleosides. Croat. Chem. Acta 2001,
381-398.
(14) Soutschek, J.; Akinc, A.; Bramlage, B.; Charisse, K.; Constien, R.;
Donoghue, M.; Elbashir, S.; Geick, A.; Hadwiger, P.; Harborth, J.;
John, M.; Kesavan, V.; Lavine, G.; Pandey, R. K.; Racie, T.; Rajeev,
K. G.; Rohl, I.; Toudjarska, I.; Wang, G.; Wuschko, S.; Bumcrot,
D.; Koteliansky, V.; Limmer, S.; Manoharan, M.; Vornlocher, H. P.
Therapeutic Silencing of an Endogenous Gene by Systemic Admin-
istration of Modified siRNAs. Nature 2004, 432, 173-178.
(15) Braasch, D. A.; Paroo, Z.; Constantinescu, A.; Ren, G.; Oz, O. K.;
Mason, R. P.; Corey, D. R. Biodistribution of Phosphodiester and
Phosphorothioate siRNA. Bioorg. Med. Chem. Lett. 2004, 14, 1139-
1143.
(16) Crooke, S. T. Molecular Mechanisms of Action of Antisense Drugs.
Biochim. Biophys. Acta 1999, 1489, 31-44.
(17) Geary, R. S.; Watanabe, T. A.; Truong, L.; Freier, S. M.; Lesnik, E.
A.; Sioufi, N. B.; Sasmor, H.; Manoharaon, M.; Levin, A. A.
Pharmacokinetic Properties of 2′-O-(2-methoxyethyl)-Modified Oli-
gonucleotide Analogs in Rats. J. Pharmacol. Exp. Ther. 2001, 296,
890-897.
(35) Fernandez-Bolanos, J. G.; Al-Masoudi, N. A. L.; Maya, I. Sugar
Derivatives Having Sulfur in The Ring. AdV. Carbohydr. Chem.
Biochem. 2001, 57, 21-98.
(36) Yokoyama, M. Synthesis and Biological Activity of Thionucleosides.
Synthesis 2000, 1637-1655.
(37) Leydier, C.; Bellon, L.; Barascut, J.-L.; Deydier, J.; Maury, G.;
Pelicano, H.; Abdelaziz El Alaoui, M.; Imbach, J. L. A New Synthesis
of Some 4′-Thio-D-Ribonucleosides and Preliminary Enzymatic
Evaluation. Nucleosides Nucleotides 1994, 13, 2035-2050.
(38) Bellon, L.; Barascut, J. L.; Maury, G.; Divita, G.; Goody, R.; Imbach,
J. L. 4′-Thio-Oligo-â-D-Ribonucleotides: Synthesis of Beta-4′-
Thiooligouridylates, Nuclease Resistance, Base Pairing Properties,
and Interaction with HIV-1 Reverse Transcriptase. Nucleic Acids Res.
1993, 1587-1593.
(39) Bellon, L.; Leydier, C.; Barascut, J. L.; Maury, G.; Imbach, J. L.
4′-Thio-RNA: A Novel Class of Sugar-Modified â-RNA. ACS Symp.
Ser. 1994, 580, 68-79.
(40) Leydier, C.; Bellon, L.; Barascut, J. L.; Imbach, J. L. 4′-Thio-â-D-
Oligoribonucleotides: Nuclease Resistance and Hydrogen Bonding
Properties. Nucleosides Nucleotides 1995, 14, 1027-1030.
(41) Leydier, C.; Bellon, L.; Barascut, J. L.; Morvan, F.; Rayner, B.;
Imbach, J.-L. 4′-Thio-RNA: Synthesis of Mixed Base 4′-Thio-
Oligoribonucleotides, Nuclease Resistance, and Base Pairing Proper-
ties With Complementary Single and Double Strand. Antisense Res.
DeV. 1995, 5, 167-174.
(42) Hoshika, S.; Inoue, N.; Minakawa, N.; Matsuda, A. Investigation of
Physical and Physiological Properties of 4′-Thioribonucleotide (4′-
thioRNA). Nucleic Acids Res. Suppl. 2003, 3, 209-210.
(43) Hoshika, S.; Minakawa, N.; Kamiya, H.; Harashima, H.; Matsuda,
A. RNA Interference Induced by siRNAs Modified With 4′-
Thioribonucleosides in Cultured Mammalian Cells. FEBS Lett. 2005,
579, 3115-3118.
(18) Geary, R. S.; Yu, R. Z.; Levin, A. A. Pharmacokinetics of
Phosphorothioate Antisense Oligodeoxynucleotides. Curr. Opin.
InVest. Drugs 2001, 2, 562-573.
(19) Braasch, D. A.; Jensen, S.; Liu, Y.; Kaur, K.; Arar, K.; White, M.
A.; Corey, D. R. RNA Interference in Mammalian Cells by
Chemically-Modified RNA. Biochemistry 2003, 7967-7975.
(20) Chiu, Y. L.; Rana, T. M. siRNA function in RNAi: A Chemical
Modification Analysis. RNA 2003, 9, 1034-1048.
(21) Chiu, Y.-L.; Rana, T. M. RNAi in Human Cells: Basic Structural
and Functional Features of Small Interfering RNA. Mol. Cell 2002,
10, 549-561.
(22) Kurreck, J. Antisense technologies. Improvement Through Novel
Chemical Modifications. Eur. J. Biochem. 2003, 1628-1644.
(23) Cummins, L. L.; Owens, S. R.; Risen, L. M.; Lesnik, E. A.; Freier,
S. M.; McGee, D.; Guinosso, C. J.; Cook, P. D. Characterization of
fully 2′-Modified Oligoribonucleotide Hetero- and Homoduplex
Hybridization and Nuclease Sensitivity. Nucleic Acids Res. 1995, 23,
2019-2024.
(44) Hoshika, S.; Minakawa, N.; Matsuda, A. Synthesis and Physical and
Physiological Properties of 4′-ThioRNA: Application to Post-
Modification of RNA Aptamer Toward NF-kappaB. Nucleic Acids
Res. 2004, 32, 3815-3825.
(24) Kawasaki, A. M.; Casper, M. D.; Freier, S. M.; Lesnik, E. A.; Zounes,
M. C.; Cummins, L. L.; Gonzalez, C.; Cook, P. D. Uniformly
Modified 2′-deoxy-2′-Fluoro Phosphorothioate Oligonucleotides as
Nuclease-Resistant Antisense Compounds with High Affinity and
Specificity for RNA Targets. J. Med. Chem. 1993, 36, 831-841.
(25) Amarzguioui, M.; Holen, T.; Babaie, E.; Prydz, H. Tolerance for
Mutations and Chemical Modifications in a siRNA. Nucleic Acids
Res. 2003, 31, 589-595.
(26) Czauderna, F.; Fechtner, M.; Dames, S.; Aygun, H.; Klippel, A.;
Pronk, G. J.; Giese, K.; Kaufmann, J. Structural Variations and
Stabilizing Modifications of Synthetic siRNAs in mammalian cells.
Nucleic Acids Res. 2003, 31, 2705-2716.
(27) Elbashir, S. M.; Martinez, J.; Patkaniowska, A.; Lendeckel, W.;
Tuschl, T. Functional Anatomy of siRNAs for Mediating Efficient
RNAi in Drosophila Melanogaster Embryo Lysate. EMBO J. 2001,
6877-6888.
(28) Harborth, J.; Elbashir, S. M.; Vandenburgh, K.; Manninga, H.;
Scaringe, S. A.; Weber, K.; Tuschl, T. Sequence, Chemical, and
Structural Variation of Small Interfering RNAs and Short Hairpin
RNAs and the Effect on Mammalian Gene Silencing. Antisense
Nucleic Acid Drug DeV. 2003, 13, 83-105.
(29) Ohmichi, T.; Sugimoto, N. Chemistry of RNAi (RNA interferon).
Kagaku (Kyoto) 2002, 43-47.
(45) Dande, P.; Prakash, T. P.; Baker, B.; Vickers, T.; Namir, S.; Jarres,
R.; Griffey, R. H.; Swayze, E. E.; Bhat, B. 4′-Thio-siRNA: Synthesis
and In Vitro RNA Interference by 4′-Thio-Modified RNA. Book of
Abstracts, 227th ACS National Meeting; American Chemical Soci-
ety: Washington, DC, 2004.
(46) Bhat, B.; Allerson, C. A.; Prakash, T. P.; Dande, P.; Sioufi, N.; Jarres,
R.; Griffey, R. H.; Swayze, E. E. Effect of Placement of Chemical
Modification on siRNA activity. XVI International Roundtable;
International Society for Nucleoside, Nucleotides & Nucleic Acids:
Minneapolis, MN, 2004.
(47) Ma, J. B.; Yuan, Y. R.; Meister, G.; Pei, Y.; Tuschl, T.; Patel, D. J.
Structural Basis for 5′-End-Specific Recognition of Guide RNA by
the A. Fulgidus Piwi Protein. Nature 2005, 434, 666-670.
(48) Parker, J. S.; Roe, S. M.; Barford, D. Structural Insights into mRNA
Recognition From a PIWI Domain-siRNA Guide Complex. Nature
2005, 434, 663-666.
(49) Khvorova, A.; Reynolds, A.; Jayasena, S. D. Functional siRNAs and
miRNAs Exhibit Strand Bias. Cell 2003, 115, 209-216.
(50) Schwarz, D. S.; Hutvagner, G.; Du, T.; Xu, Z.; Aronin, N.; Zamore,
P. D. Asymmetry in the Assembly of the RNAi Enzyme Complex.
Cell 2003, 115, 199-208.
(51) Naka, T.; Minakawa, N.; Abe, H.; Kaga, D.; Matsuda, A. The
Stereoselective Synthesis of 4′-â-Thioribonucleosides via the Pum-
merer Reaction. J. Am. Chem. Soc. 2000, 122, 7233-7243.
(52) Bellon, L.; Barascut, J. L.; Imbach, J. L. Efficient Synthesis of 4-Thio-
D-Ribofuranose and Some 4′-Thioribonucleosides. Nucleosides Nu-
cleotides 1992, 1467-1479.
(53) Conte, V.; Di Furia, F.; Licini, G.; Modena, G.; Sbampato, G.
Titanium-Promoted Enantioselective Oxidation of Thioethers and
Synthetic Applications. Stud. Surf. Sci. Catal. 1991, 66, 385-394.
(54) Di Furia, F.; Licini, G.; Modena, G.; Valle, G. Enantioselective
Oxidation of Thioethers. A route to The Resolution of [1,1′-
Binaphthalene]-2,2′-Dithiol. Bull. Soc. Chim. Fr. 1990, 734-744.
(30) Allerson, C. A.; Sioufi, N. B.; Jarres, R.; Prakash, T. P.; Naik, N.;
Berdeja, A.; Wanders, L.; Griffey, R. H.; Swayze, E. E.; Bhat, B.
Fully 2′-Modified Oligonucleotide Duplexes With Improved In Vitro
Potency and Stability Compared to Unmodified Small Interfering
RNA. J. Med. Chem. 2005, 48, 901-904.
(31) Prakash, T. P.; Allerson, C. A.; Dande, P.; Vickers, T. A.; Sioufi,
N.; Jarres, R.; Baker, B. F.; Swayze, E. E.; Griffey, R. H.; Bhat, B.
Positional Effect of Chemical Modifications on siRNA Activity in
mammalian cells. J. Med. Chem. 2005, 48, 4247-4253.

1634 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 5
Dande et al.
(55) Caputo, R.; Guaragna, A.; Palumbo, G.; Pedatella, S. Asymmetric
Synthesis of 1,3-Dithiolane Nucleoside Analogues. Eur. J. Org.
Chem. 2003, 346-350.
(62) Parker, J. S.; Roe, S. M.; Barford, D. Crystal Structure of a PIWI
Protein Suggests Mechanisms for siRNA Recognition and Slicer
Activity. EMBO J. 2004, 23, 4727-4737.
(63) Haeberli, P.; Berger, I.; Pallan, P. S.; Egli, M. Synthesis of
4′-Thioribonucleosides and Thermodynamic Stability and Crystal
Structure of RNA Oligomers with Incorporated 4′-Thiocytosine.
Nucleic Acids Res. 2005, 33, 3965-3975.
(64) Manoharan, M. 2′-Carbohydrate Modifications in Antisense Oligo-
nucleotide Therapy: Importance of Conformation, Configuration and
Conjugation. Biochim. Biophys. Acta 1999, 1489, 117-130.
(65) Lesnik, E. A.; Freier, S. M. What Affects the Effect of 2′-alkoxy
Modifications? 1. Stabilization Effect of 2′-Methoxy Substitutions
in Uniformly Modified DNA Oligonucleotides. Biochemistry 1998,
37, 6991-6997.
(56) Caputo, R.; Guaragna, A.; Palumbo, G.; Pedatella, S. Asymmetric
Synthesis of Optically Active 1,3-oxathiolane Nucleoside Analogs.
Nucleosides Nucleotides 1998, 1739-1745.
(57) Gait, M. J., Ed. Oligonucleotide Synthesis: A Practical Approach;
IRL Press: Oxford and Washington, DC, 1984; 217 pp.
(58) Divakar, K. J., Reese, C. B. 4-(1,2,4-Triazol-1-yl)-and 4-(3-Nitro-
1,2,4-Triazol-1-yl)-1-(â-D-2,3,5-tri-O-Acetylarabinofuranosyl) Py-
rimidin-2(1H)-ones. Valuable Intermediates in The Synthesis of
Derivatives of 1-(b-D-Arabinofuranosyl)Cytosine (ara-C). J. Chem.
Soc., Perkin Trans. 1 1982, 1171-1176.
(59) Song, Q., Ross, B. Process for Manufacturing Purified Phospho-
rodiamidite. Patent Application WO 03/087130, 2003.
(66) Lesnik, E. A.; Guinosso, C. J.; Kawasaki, A. M.; Sasmor, H.; Zounes,
M.; Cummins, L. L.; Ecker, D. J.; Cook, P. D.; Freier, S. M.
Oligodeoxynucleotides Containing 2′-O-Modified Adenosine: Syn-
thesis and Effects on Stability of DNA:RNA Duplexes. What Affects
the Effect of 2′-alkoxy Modifications? 1. Stabilization Effect of 2′-
Methoxy Substitutions in Uniformly Modified DNA Oligonucleotides.
Biochemistry 1993, 32, 7832-7838.
(60) Zamore, P. D.; Tuschl, T.; Sharp, P. A.; Bartel, D. P. RNAi: Double-
stranded RNA Directs the ATP-Dependent Cleavage of mRNA at
21 to 23 Nucleotide Intervals. Cell 2000, 25-33.
(61) Hamada, M.; Ohtsuka, T.; Kawaida, R.; Koizumi, M.; Morita, K.;
Furukawa, H.; Imanishi, T.; Miyagishi, M.; Taira, K. Effects on RNA
Interference in Gene Expression (RNAi) in Cultured Mammalian
Cells of Mismatches and the Introduction of Chemical Modifications
at The 3′-Ends of siRNAs. Antisense Nucleic Acid Drug DeV. 2002,
12, 301-309.
JM050822C