Amides are excellent mimics of phosphate internucleoside linkages and are well tolerated in short interfering RNAs

Article abstract of DOI:10.1093/nar/gku235

RNA interference (RNAi) has become an important tool in functional genomics and has an intriguing therapeutic potential. However, the current design of short interfering RNAs (siRNAs) is not optimal for in vivo applications. Non-ionic phosphate backbone m

Full text of DOI:10.1093/nar/gku235

6542–6551 Nucleic Acids Research, 2014, Vol. 42, No. 10
Published online 09 May 2014
doi: 10.1093/nar/gku235
Amides are excellent mimics of phosphate
internucleoside linkages and are well tolerated in
short interfering RNAs
Daniel Mutisya1, Chelliah Selvam1, Benjamin D. Lunstad2, Pradeep S. Pallan3,
Amanda Haas2, Devin Leake2, Martin Egli3,* and Eriks Rozners1,*
1Department of Chemistry, Binghamton University, The State University of New York, Binghamton, NY 13902, USA,
2Global Research and Development in Molecular Biology, Thermo Fisher Scientific Bioscience Division, Lafayette, CO
80026, USA and 3Department of Biochemistry, School of Medicine, Vanderbilt University, Nashville, TN 37232, USA
Received December 20, 2013; Revised February 20, 2014; Accepted March 8, 2014
ABSTRACT
technologies (1,2). Chemical modifications have been in-
strumental in improving the stability of oligonucleotides in
biological media. However, difficulties in targeted delivery,
unfavorable pharmacokinetics and poor cellular uptake re-
main major obstacles for in vivo applications. These diffi-
culties are in large part due to the negatively charged and
polar phosphodiester backbone. Although replacement of
the non-bridging oxygen with sulfur has showed promising
improvement of antisense oligonucleotide properties (3),
more dramatic modifications of the phosphodiester back-
bone have been little explored (4,5).
Replacement of DNA phosphodiesters with non-ionic
linkages to improve the enzymatic stability has been studied
for antisense oligonucleotides (4,5). Among such linkages,
amides (Figure 1) emerged as favorites because they were
relatively easy to make by peptide-type couplings. More-
over, early results indicated that short DNA sequences with
isolated amide linkages formed stable duplexes with com-
plementary RNAs. Dimers AM3–AM5 (Figure 1), the first
amides studied in DNA, were found to destabilize DNA–
RNA heteroduplexes by –1 to –4◦C per modification (de-
crease in duplex melting temperature, tm) depending on the
sequence context (6–8).
Among various isomeric linkages AM1 and AM2 were
later identified as the best substituents for phosphodiester
linkages in DNA (4). Depending on the sequence context,
AM1 and AM2 modification of the DNA strand led to
slight stabilization or destabilization (+0.9 to –0.8◦C per
modification) of the DNA–RNA heteroduplexes (9). Both
modifications destabilized all-DNA duplexes by about –
4◦C per modification (9). AM1, first described indepen-
dently by Just et al. (10) and De Mesmaeker et al. (11,12)
in 1993–94, has become the most studied amide modifi-
cation in DNA. While preliminary nuclear magnetic reso-
nance (NMR) (13) and molecular modeling (14,15) studies
suggested that AM1 linkage adopted an A-like conforma-
RNA interference (RNAi) has become an important
tool in functional genomics and has an intriguing
therapeutic potential. However, the current design
of short interfering RNAs (siRNAs) is not optimal
for in vivo applications. Non-ionic phosphate back-
bone modifications may have the potential to im-
prove the properties of siRNAs, but are little ex-
plored in RNAi technologies. Using X-ray crystallog-
raphy and RNAi activity assays, the present study
demonstrates that 3´-CH2-CO-NH-5´ amides are ex-
cellent replacements for phosphodiester internucle-
oside linkages in RNA. The crystal structure shows
that amide-modified RNA forms a typical A-form du-
plex. The amide carbonyl group points into the major
groove and assumes an orientation that is similar
to the P–OP2 bond in the phosphate linkage. Amide
linkages are well hydrated by tandem waters link-
ing the carbonyl group and adjacent phosphate oxy-
gens. Amides are tolerated at internal positions of
both the guide and passenger strand of siRNAs and
may increase the silencing activity when placed near
the 5´-end of the passenger strand. As a result, an
siRNA containing eight amide linkages is more ac-
tive than the unmodified control. The results suggest
that RNAi may tolerate even more extensive amide
modification, which may be useful for optimization
of siRNAs for in vivo applications.
INTRODUCTION
Interest in synthetic chemistry of nucleic acids has been
driven by the need for modified oligonucleotides for in vivo
applications in antisense and RNA interference (RNAi)
*To whom correspondence should be addressed. Tel: +1 607 777 2441; Fax: +1 607 777 4478; Email: erozners@binghamton.edu
Correspondence may also be addressed to Martin Egli. Tel: +1 615 343 8070; Fax: +1 615 322 7122; Email: martin.egli@vanderbilt.edu
ꢀ
C
The Author(s) 2014. Published by Oxford University Press on behalf of Nucleic Acids Research.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0/), which
permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

Nucleic Acids Research, 2014, Vol. 42, No. 10 6543
changes caused by the amide linkage were easily accom-
modated by small adjustments in RNA structure suggest-
ing that amides may be excellent mimics of phosphate link-
ages in RNA and promising modifications to optimize short
interfering RNAs (siRNAs). Herein we extend these stud-
ies and present a crystal structure that provides detailed
conformational information on how the amide is accom-
modated in the RNA duplex and further illustrates the ex-
cellent hydration and conformational adaptability of amide
linkages in RNA.
While our studies on amide-modified RNA were in
progress, Iwase et al. showed that modification of the 3´-
overhangs of an siRNA with two AM1 (Figure 1) link-
ages increased the enzymatic stability but did not decrease
RNAi activity (24,25). However, this was not unexpected
because the 3´-overhangs in general tolerate modifications
much better than the internal positions of siRNAs. Gong
and Desaulniers studied siRNAs containing an amide link-
age derived from insertion of a PNA monomer (AM-PNA,
Figure 1) (26). The PNA-derived amide linkage was toler-
ated in the 3´-overhang of the passenger strand. However,
internal modification of the guide strand led to significant
loss of silencing activity. Potenza et al. also reported that
replacement of the phosphates in 3´-overhangs with two
PNA linkages increased the enzymatic stability of siRNAs
but did not affect their RNAi activity (27). Herein we show
that amide linkages are not only tolerated at internal posi-
tions of both guide and passenger strands of siRNAs but
may increase the silencing activity when placed near the 5´-
end of the passenger strand. These findings are unexpected
and raise the possibility that RNAi may tolerate and bene-
fit from even more substantial modifications than the ones
tried so far.
Figure 1. Structure of amide-modified DNA and RNA dimers.
tion in the DNA strand, a more detailed structure of an
amide-modified oligodeoxynucleotide has not been deter-
mined to date. De Mesmaeker et al. (15) briefly explored
AM1 dimers derived from RNA (R = OH) and 2´-O-Me
RNA (R = OCH3). They found that the 2´-O-Me modi-
fication increased thermal stability of duplexes, especially
when placed on the sugar that has the amide modification
attached to the C5´ (15,16). In contrast, the 2´-OH next to
the amide linkage (on the sugar that has the amide modi-
fication attached to the C3´) decreased the thermal stabil-
ity (15). However, rationalization of these observations was
complicated by the fact that the amide modifications with
RNA-like sugars were placed in the DNA strand of DNA–
RNA heteroduplexes.
Early studies explored binding of modified DNA to RNA
targets (formation of DNA–RNA heteroduplexes) because
of the focus on potential antisense applications of the mod-
ified oligodeoxyribonucleotides (4). More recently, the dis-
covery of RNAi revitalized the interest in chemical modi-
fications of RNA (1,2). Robins et al. reported synthesis of
AM1 linked RNA dimers (17–19) and pentamers (20), but
did not study the biophysical properties of these analogues.
Rozners et al. found that both AM1 and AM2 dimers with
either 2´-OH or 2´-O-Me sugars were well accommodated
in all-RNA duplexes (21,22). While the effect of AM1 on
thermal stability was relatively small, AM2 significantly in-
creased stability of RNA duplexes (22). More recently, de-
tailed thermodynamic and NMR structural studies by our
group showed that AM1 amides had surprisingly little ef-
fect on the A-type conformation, thermal stability and hy-
dration of RNA duplexes (23). The local conformational
MATERIALS AND METHODS
Synthesis and purification of amide-modified RNA
Amide-modified oligoribonucleotides were prepared on a 1
␮mol scale using the standard 2´-O-TOM RNA phospho-
ramidite (Glen Research) synthesis protocol on an Expe-
dite 8909 Nucleic Acid Synthesis System. A standard cou-
pling time was used for the dimeric phosphoramidites 11a
and 11b. Cleavage of oligoribonucleotides from solid sup-
port and deprotection of the heterocyclic amino groups was
done by treating the solid support with a mixture of ethano-
lic methylamine and aqueous methylamine (1:1) solution
at room temperature for 24 h. The cleavage solution was
freeze-dried and the residue was dissolved in dimethyl sul-
foxide (100 ␮l). Triethylamine trihydrofluoride (125 ␮l) was
added and the reaction mixture was left for 24 h at room
temperature to remove the 2´-O-TBS and 2´-O-TOM pro-
tecting groups. The reaction mixtures were diluted with wa-
ter (1275 ␮l) and desalted on a Sephadex C25 NAP col-
umn in accordance with the manufacturer’s recommenda-
tions. The amide-modified oligoribonucleotides were ana-
lyzed by reverse-phase high-performance liquid chromatog-
raphy (HPLC) with an XBridgeTM C18 column (4.6 ×
150 mm, 3.5 ␮m, 1 ml/min) and purified by reverse-phase
HPLC with an XBridgeTM C18 column (10 × 150 mm,
5 ␮m, 2 ml/min) using a linear gradient (2–40%) of ace-
tonitrile in 0.1 M triethylammonium acetate buffer, pH 7.0

6544 Nucleic Acids Research, 2014, Vol. 42, No. 10
Figure 2. (A) Example of the quality of the final Fourier (2Foo-Fcc) sum electron density (∼1.2 ␴ threshold). (B) Overall view of RNA13 duplexes in the
unit cell. Residues in the four strands are numbered from 101–113, 201–213, 301–313 and 401–413, respectively. Carbon atoms of strands 1, 2, 3 and 4 are
colored in beige, pink, salmon and gray, respectively, and red, blue and orange mark oxygen, nitrogen and phosphorus atoms, respectively. Strontium ions
are shown as orange spheres and smaller spheres in cyan represent water molecules. The carbon atoms in the amide carbonyls and the adjacent linking
carbon atoms are colored in green and the C-NH-CO-C linkages are highlighted in ball and stick mode.
(for HPLC chromatograms, see Supplementary Figures S1
and S12–S27). The fractions containing the target mate-
rial were freeze-dried. The residue was dissolved in water (5
ml) and freeze-dried again. To remove the bulk of triethy-
lammonium salts, the latter operation (dissolve and freeze-
dry) was repeated one or two times. To completely convert
the amide-modified RNAs into the sodium salt form, the
samples were dissolved in phosphate buffer (0.5 ml of 20
mM sodium phosphate, 80 mM NaCl, 50 ␮M EDTA, pH
6.3) and desalted on a Sephadex C25 NAP column. The
residue was dialyzed against pure water using a Float-A-
Lyzer G2 (MWCO: 100–500 D, Spectrum Laboratories) for
10 h and changing the water after 2 and 6 h. Quantifica-
tion of amide-modified RNAs was done using the nearest-
neighbor approximation (28). The identity of the siRNA
oligonucleotides was confirmed by electrospray ionization
mass spectrometry (for details, see Supplementary Data).
(45 mM), magnesium chloride (10 mM), spermine tetrahy-
drochloride (6 mM), and 5% MPD (v/v), that were equili-
brated against a reservoir of MPD (1 ml, 35%). All crys-
tals were mounted in nylon loops without further cryo-
protection and frozen in liquid nitrogen. Diffraction data
were collected on the 21-ID-D beam line of the Life Sciences
Collaborative Access Team (LS-CAT) at the Advanced
Photon Source (APS), located at Argonne National Lab-
oratory (Argonne, IL). The wavelength was tuned to 0.76
A for crystals containing Sr2+ (for Ba2+ containing crystals
˚
˚
data were collected at 1.77 A) and data were collected at
110 K using a MARCCD 300 detector. Diffraction data
were integrated, scaled and merged with the XDS pack-
age (30). Selected crystal data and data collection param-
eters are listed in Supplementary Table S1. The structure of
RNA13 contains two duplexes in space group P1 and was
phased by single wavelength anomalous dispersion (SAD)
of the Sr2+ data using the program SHELXD in the SAD
mode, followed by density modification using SHELXE
(31). Building of the models was performed with COOT
(32). The initial orientations of duplexes were optimized by
several rounds of rigid body refinement in Refmac5 (33).
Refinement was then continued in Refmac5, using the max-
imum likelihood residual method. Subsequent refinement
cycles were carried out using the program SHELXL (31),
keeping aside 5% of the reflections to compute the R-free
(34). Following a few cycles of refinement in SHELXL,
phosphate groups were replaced by amide linkages and the
dictionary was adapted to account for the altered backbone
chemistry. Ions and water molecules were placed at sites
of peaks in the Fourier 2Fo-Fc sum and Fo-Fc difference
X-ray crystallography
Crystals of 5´-r(UpGpApGpCpUaUpCpGpGpCpUpC)-3´
(RNA13) were grown by the hanging-drop vapor diffusion
technique using the Nucleic Acid Miniscreen (Hampton
Research, Aliso Viejo, CA) (29) and later optimized by vary-
ing the concentration of strontium chloride. The 13-mer
amide RNA crystallized from two conditions: 1) droplets
(2 ␮l) containing oligonucleotide (0.6 mM), sodium ca-
codylate (20 mM, pH 6.0), sodium chloride (40 mM),
barium chloride (10 mM), spermine tetrahydrochloride (6
mM), and 2-methyl-2,4-pentanediol (MPD; 5% (v/v)) and
2) droplets (2 ␮l) containing oligonucleotide (0.6 mM),
sodium cacodylate (20 mM, pH 6.0), strontium chloride

Nucleic Acids Research, 2014, Vol. 42, No. 10 6545
Scheme 1. Synthesis of dimeric r(UAM1U) and r(UAM1A) phosphoramidites. Steps: (a) TFA, tetrahydrofuran, H2O 0◦C, 4 h, 92%; (b) p-methoxytrityl
chloride, pyridine, 0◦C to rt, 14 h, 85%; (c) OsO4, 4-methylmorpholine N-oxide, dioxane, rt, 10 h, then NaIO4 in water, rt, 12 h, 95%; (d) NaClO4, NaH2PO4,
2-methylbut-2-ene, t-BuOH, THF water, rt, 1 h, 82%; (e) DIEA, Bu2SnCl2, dichloroethane, rt, 1 h, then add TOM-Cl, 80◦C, 45 min, 39%; (f) H2S, pyridine,
water, rt, 14 h 93%; (g) HBTU, HOBt, DIEA, CH2Cl2, rt, 12 h, 95% 11a, 86% 11b; (h) DIEA, ClP(OCH2CH2CN)N(iPr)2, CH2Cl2, rt, 7 h, 71% 11a, 60%
11b.
electron density maps, and accepted on the basis of stan-
dard distance criteria. Final refinement parameters and de-
viations from ideal geometries are listed in Supplementary
Table S1. An example of the quality of the final electron
density is depicted in Figure 2A and an overall view of the
two duplexes is shown in Figure 2B. While phasing by Sr2+-
SAD was thus successful in the case of RNA13 crystals; we
have so far been unable to phase the Ba2+ crystal form. Fi-
nal coordinates and structure factors for RNA13 have been
deposited in the Protein Data Bank (http://www.rcsb.org).
The PDB ID code is 4O41.
5´-phosphorylated unmodified controls were prepared and
used and the conclusions were based on the relative activi-
ties. The results presented in Figures 5–9 are expressed as ra-
tios of target Cyclophilin B mRNA expression (PPIB) and
housekeeping mRNA expression (GAPDH) and are aver-
ages of three experiments (biological triplicates from three
separate wells of cells). The standard deviations are shown
in the length of the error bars. UV melting experiments on
siRNA duplexes were carried out in 25 mM NaCl, 15 mM
sodium citrate, 1 mM EDTA at pH 7.2 using a Thermo Sci-
entific Evolution 3000.
siRNA activity test
HeLa cells were plated in 96-well plates (1 × 104 cells/well)
for 24 h before transfection. On the day of transfec-
tion, RNA-lipid complexes were introduced into each well
of cells (0.1–100 nM RNA for HeLa cells 0.2 ml/well
DharmaFECT 1). siGENOME Non-Targeting siRNA #1
(r(UAGCGACUAAACACAUCAAUU), Thermo Fisher,
catalogue # D-001210–01) was used as a non-target con-
trol (NTC). NTC and siRNAs targeting the Cyclophilin B
(PPIB) gene were both titrated separately at the same con-
centrations to show that they behave in the same way and
do not cause toxicity (cell viability was determined using a
Resazurin assay, Supplementary Figures S4–S10). Twenty-
four hours post-transfection, the level of target knockdown
was assessed using a branched DNA (bDNA) assay spe-
cific for the targets of interest according to the manufac-
turer’s instructions (QuantiGene branched DNA signal am-
plification kit; Panomics, Fremont, CA). Cells were lysed
and directly added to the bDNA assay plate without nu-
cleic acid purification or cDNA synthesis steps that can in-
troduce bias. Except for sequences having amide modifi-
cations at the 3´-end, modified guide strands were chem-
ically 5´-phosphorylated because of concern that amide
modifications close to the 5´-end may interfere with en-
zymatic phosphorylation normally observed in cells. The
modified passenger strands were prepared in 5´-OH and
5´-phosphorylated form. This was done to study different
potential phosphorylation states. Appropriate 5´-OH and
RESULTS
Synthesis of amide-modified RNA
Synthesis of amide-linked dimers started with optimization
of the previous procedures (22) for preparation of the uri-
dine carboxylic acid 5 (Scheme 1). The 5´-O-TBS group
was selectively cleaved using trifluoroacetic acid (TFA)
and replaced with the 5´-O-methoxytrityl (MMT) group
in 3. Two-step oxidative degradation of the alkene gave
the carboxylic acid 5, which had the more stable 2´-O-
TBS instead of the previously used 2´-O-Ac protection (23).
The new route was also one step shorter. The uridine 8
and adenosine 9 amines were prepared as previously de-
scribed (23). N,N,Nꢁ,Nꢁ-Tetramethyl-O-(1H-benzotriazol-
1-yl)uronium hexafluorophosphate (HBTU) mediated cou-
pling of 5 with aminouridine 8 or aminoadenosine 9 gave
the r(UAM1U) and r(UAM1A) dimers 10a and 10b, respec-
tively (Scheme 1), which were converted into 11a and 11b in
one standard step of phosphoramidite synthesis (for exper-
imental details, see Supplementary Data). We chose modi-
fication of U and A nucleosides because of more straight-
forward chemical synthesis than that of C and G nucleo-
sides. Dimers 11 were used together with common 2´-O-
TOM protected ribonucleoside monomers (Glen Research)
to synthesize a series of amide-modified RNAs according to
standard phosphoramidite chemistry on an Expedite 8909
instrument.

6546 Nucleic Acids Research, 2014, Vol. 42, No. 10
(Figure 2B). One is coordinated to N7 and O6 of G from
a G:C pair flanking the central stretch of mismatch pairs.
The other is coordinated to O2´ and O3´ of the 3´-terminal
C of the first strand, reminiscent of the coordination modes
of platinum and osmium in crystals of phenylalanine tRNA
(35).
The RNA13 duplex adopts an A-like conformation with
modulated groove widths as a result of the UoG wobble and
U*C pairs in the central part. Thus, the minor groove width
along the duplex as measured by phosphate–phosphate dis-
˚
tances varies between 15 and 18 A, whereby the section with
mismatch base pairs is notably wider. The conformation of
RNA13 differs significantly from that displayed by the na-
tive RNA 13mer (Figure 3). The native duplex is of the so-
called A´-form with 12 base pairs per turn compared to 11
base pairs in the canonical A-form duplex (PDB ID 413D)
(36). Moreover, the base pairs in the native RNA exhibit
smaller inclination angles and the major groove is consid-
erably wider (and the minor groove therefore contracted)
relative to the RNA13 duplex. Substantial differences be-
tween the two duplexes are also illustrated by the fact that
attempts to solve the RNA13 structure with molecular re-
placement and using the native RNA duplex as the search
model failed consistently.
The reasons for this switch from the A´- to the A-
conformation are complex. It is likely that relatively sub-
tle changes such as the replacement of a phosphate by an
amide moiety, altered packing forces or ionic strength, or
the presence of a ligand can trigger the conversion from one
form to the other. Interestingly, the native RNA duplex is
located on a dyad in space group C2, whereas RNA13 du-
plexes are located in general positions in P1. The different
packing modes go along with a marked reduction in the
volume per base pair of about 25% in the case of RNA13
duplexes. The two independent RNA13 duplexes in space
group P1 also exhibit subtle differences in their overall con-
formations (Supplementary Figure S3).
Figure 3. Comparison between the conformation of the amide-modified
RNA13 duplex (backbone ribbons of strands 1 and 2 colored in beige and
pink, respectively) and that of the native 13mer RNA duplex (PDB ID
413D; backbone ribbons colored in yellow). The image was generated by
overlaying terminal base pairs at one end of the duplexes. The view is across
the major and minor grooves, strontium ions are orange spheres and base
pairs are depicted as blue slabs.
X-ray crystal structure of amide-modified RNA
Out of several self-complementary RNAs with a single iso-
lated amide linkage we synthesized for crystallography stud-
ies, RNA 13mer rUGAGCUAM1UCGGCUC (abbreviated
RNA13) gave well diffracting crystals. The melting temper-
atures of RNA13 (tm = 78◦C) and the unmodified RNA (tm
= 79◦C) were similar. The circular dichroism spectra had
significant differences (Supplementary Figure S2) but were
not expected to be similar because of different conforma-
tions of the respective crystal structures (see Figure 3). This
13mer features non-canonical UoG wobble and U*C base
pairs in the middle of the duplex and 5´-terminal U over-
hangs. The space group was P1 with two 13mer duplexes in
the unit cell. The structure was phased by strontium single
wavelength anomalous dispersion (Sr-SAD) using a wave-
Amide moieties in all four strands assume similar confor-
mations (Figure 4A). The backbone torsion angles fall into
the following ranges: ac- (⑀), sc- (␨), ap (␣), ac- (␤), ap (␥)
and sc+ (␦). By comparison, the standard A-form RNA tor-
sion angle ranges are ap (⑀), sc- (␨), sc- (␣), ap (␤), sc+ (␥)
and sc+ (␦). The structure reveals that the amide carbonyl
group is rotated into the major groove and thus assumes an
orientation that is similar to that of the P–OP2 bond (Fig-
ure 4A). In the case of the UAM1U step, this orientation of
the amide C = O bond results in a relatively short contact
between the amide oxygen and uracil C6–H6 (average dis-
˚
tance 3.4 A) that is consistent with formation of a C–H...O
hydrogen bond.
Analysis of the water structure around the backbones of
RNA13 duplexes shows that phosphate and amide carbonyl
oxygens as well as 2´-hydroxyl groups are well hydrated
(Figure 4B and C). Compared to 2´-OH, the amide N–H
function appears to be a less attractive donor for H-bonding
interactions. Along the rim of the major groove OP2 phos-
phate oxygens are bridged by single water molecules (Figure
4B), a pattern long ago noticed in A-form DNA and RNA
duplexes (37,38). Compared with the distances between ad-
˚
length of 0.76 A for data collection on the LS-CAT 21-ID-D
beamline at the APS (Argonne, IL). Two partially occupied
strontium ions were retrieved and the experimental density
allowed building of complete models of the two duplexes
without the dangling Us. Their structures along with 269
water molecules were refined to an R-factor of 17.1% (R-
˚
free 23.5%) at a resolution of 1.2 A. The 5´-terminal uridines
are disordered in all four strands and were omitted from the
refinement. An electron density map is depicted in Figure
2A. Both strontium ions are associated with a single duplex
˚
jacent OP2 atoms (as short as 5 A), the distances between

Nucleic Acids Research, 2014, Vol. 42, No. 10 6547
Figure 4. (A) Conformation of the amide moiety. The image depicts a superimposition of the four independent amide linkages with average values of
torsion angles indicated on the right. (B and C) Major and minor groove hydration in the central portion of the RNA13 duplex. Two views of the overlay
of four strands (only residues 5–8 are shown) depict (B) major groove hydration and (C) minor groove hydration (obtained after rotation by 90? around
the vertical). Carbon atoms of strands 1, 2, 3 and 4 are colored in beige, pink, salmon and gray, respectively. The small spheres represent water molecules
with colors matching their respective strands and dashed lines indicate hydrogen bonds. The amide carbonyls and the adjacent linking carbon atoms are
colored in green and C-NH-CO-C linkages are highlighted in ball and stick mode.
ing the Cyclophilin B (PPIB) gene. The modified siRNAs
(combined with an unmodified passenger strand) and ap-
propriate unmodified controls were transfected into HeLa
cells using DharmaFECT 1 (Thermo Fisher) for 24 h. PPIB
expression was assayed by branched DNA (bDNA Quanti-
gene) relative to GAPDH (housekeeping gene) expression
at different siRNA concentrations (100–0.1 nM). Consis-
tent with Iwase’s results (24,25), insertion of one amide at
the 3´-end of siRNA 3 had little effect on silencing activ-
ity (Figure 5). One modification in the middle of the guide
strand (siRNA 2) was also relatively well tolerated, which
was in contrast to Gong and Desaulniers’ report (26) that
the AM-PNA linkage in the middle of the guide strand
significantly decreased RNAi activity. Increasing the num-
ber of amide modifications in the guide strand siRNA 1
decreased the silencing activity. It is very likely that the
reduced RNAi activity was due to inefficient loading of
the amide-modified guide strand in RNA-induced silencing
Figure 5. RNAi activity of siRNAs with amide modifications in the guide
complex (RISC), as discussed below.
In crystal structures of complexes between oligonu-
cleotides and Ago or Ago domains, the protein recognizes
strand. The results are expressed as ratios of target mRNA expression
(PPIB) and housekeeping mRNA expression (GAPDH) and are averages
of three replicates; the standard deviations are shown in the length of the
the overall shape of siRNA by hydrogen bonding to back-
bone phosphates (39–43). Except for the preferential recog-
nition of U and A at the 5´-end of guide strand by the MID
domain (39,40,44), Ago does not make specific interactions
to nucleobases. Since the interactions of Ago with siRNA
should have little sequence specificity, it is conceivable that
the effect of amide modification would depend mostly on
its location along the siRNA duplex and not on the spe-
cific sequence of siRNA. To test this notion, we selected
two guide strands targeting the Cyclophilin B gene, PPIB-
HFS438 (siRNA 4) and PPIBHFS542 (siRNA 5) that al-
lowed for an amide modification at the same position, be-
tween nucleosides 12 and 13 (Figure 6). Similar to siRNA
2, this position is close to the mRNA cleavage site and may
be sensitive to modifications. As expected, amide-modified
error bars. Control P and Control OH are unmodified siRNAs with 5´-
phosphate or 5´-OH, respectively. NTC is a non-targeting control siRNA
used as a negative control; tm values are the melting temperatures of the
siRNA duplexes.
amide carbonyl oxygen and OP2s from 5´- and 3´-adjacent
˚
residues are slightly longer (ca. 6 A), resulting in tandem
waters linking the carbonyl group and phosphate oxygens
(Figure 4B).
RNAi activity of amide-modified siRNAs
Using the dimeric r(UAM1U) and r(UAM1A) phospho-
ramidites, we first introduced up to three amide modifica-
tions in the guide strand of siRNA (PPIBHFS308) target-

6548 Nucleic Acids Research, 2014, Vol. 42, No. 10
creased RNAi activity (Figure 7). The most favorable effect
came from the modifications at the 5´-end because siRNA
7 with three modifications in the middle and at the 3´-end
of the siRNA had activity similar to unmodified control.
A single amide in siRNA 9 appeared to have relatively lit-
tle effect, while the 3´-end modification in siRNA 8 slightly
reduced the activity.
Testing individual 5´-end modifications revealed that an
amide between nucleosides 1 and 2 was the most beneficial
(Figure 8 and Supplementary Figure S11), at least in the
5´-OH series. These experiments also revealed an interest-
ing effect of 5´-phosphorylation on the activity of amide-
modified passenger strands. The guide strand requires 5´-
phosphorylation to be loaded in RISC. Conversely, 5´-
phosphorylation of the passenger strand is not desirable
as it may decrease the activity by enhancing loading of
the passenger strand instead of the guide strand. Chemi-
cal modifications of the passenger strand that prevent 5´-
phosphorylation have been used to improve RNAi activ-
ity and suppress off-target effects (45). As expected, the
non-phosphorylated siRNA 11 was somewhat more active
than the phosphorylated siRNA 10 (Figure 8). Interestingly,
phosphorylation had little effect on the heavily modified
siRNA 6 and siRNA 12, indicating that the amide modifi-
cations are more important for directing the RISC loading
than phosphorylation.
Figure 6. Comparison of the effect of amide modification at the same po-
sition in two different guide strands PPIBHFS438 (siRNA 4) and PPIB-
HFS542 (siRNA 5). Control 4 and Control 5 are unmodified sequences of
siRNA 4 and siRNA 5, respectively. For other notes, see Figure 5.
Consistent with our earlier biophysical studies (22,23),
the AM1 modification had relatively little impact on ther-
mal stability (melting temperatures, tm) of siRNA duplexes
(Figures 5 and 7). The lack of correlation between tm and
silencing activity suggested that the observed changes in ac-
tivity were likely due to specific siRNA–Ago interactions
and not due to modulation of the thermodynamic stability
of siRNA duplexes.
Thermodynamic strand bias (TSB, Figure 9), a passenger
strand having two 2´-OMe groups at the 5´-end, is a mod-
ification for increasing activity and selectivity by enforcing
the loading of the guide strand in RISC (46). Both TSB and
amide-modified passenger strand improved siRNA activity
in a similar way. Interestingly, the amide-modified passen-
ger siRNA 6 was more efficient in improving the silencing
activity of the amide-modified guide siRNA 1 (c.f., Figure
9, siRNA 6/1 and TSB/siRNA 1). These data show that suit-
able amide modifications in the passenger strand are capa-
ble of forcing effective loading of the guide strand in RISC.
Another striking result was the ability of amide modifica-
tions in the passenger strand siRNA 6 to more than com-
pensate for RNAi activity decrease upon modification of
the guide strand siRNA 1 (Figure 9). The highly modified
siRNA 6/1 (eight amide linkages) was more active than the
unmodified control.
Figure 7. RNAi activity of siRNAs with amide modifications in the pas-
senger strand. For other notes, see Figure 5.
siRNA 4 and siRNA 5 had somewhat decreased RNAi ac-
tivity compared to unmodified controls. Most importantly,
the change in RNAi activity upon amide modification was
about the same for both sequences. Control 5 was somewhat
less active than Control 4 and this difference was reflected
in siRNA 4 and siRNA 5 as well. These results provided
preliminary evidence that amide modifications are likely to
have position-specific, but not sequence-dependent effects
on RNAi activity. In other words, amide modification may
be expected to have a similar effect at the same position in
different siRNAs.
DISCUSSION
siRNAs have been modified mostly in the ribose moiety and
to lesser extent in the heterocyclic bases (1,2). Apart from
phosphorothioates (47–49), phosphorodithioates (50) and
boranophosphates (51), chemical modification of the phos-
phate backbone in siRNAs has been little explored. Our
early work suggested that AM1 amides are structurally fit as
replacements of phosphodiesters in A-form RNA duplexes
The effect of amide modifications in the passenger strand
on silencing activity was studied using the same siRNA
(PPIBHFS308) as for the guide strand. The most striking
result was that five amide modifications in siRNA 6 in-

Nucleic Acids Research, 2014, Vol. 42, No. 10 6549
showed that even three consecutive amide linkages had rela-
tively little effect on structure and hydration of a short RNA
duplex (52). The alignment of the amide carbonyl group in
an orientation that is similar to that of the P–OP2 bond is
noteworthy and suggests that the amide C = O may be able
to mimic non-bridging P–O in RNA–protein interactions.
Recent crystal structures of siRNAs bound to Argonaute
2 (Ago2), the key protein of the RISC, show that most of
the phosphates of the guide (antisense) strand make hydro-
gen bonds to amino acid side chains of Ago2 (39–43). In
contrast, the passenger (sense) strand is partially solvent ex-
posed and phosphates 1, 2, 5, and 13–17 do not interact
with the protein (43). The phosphates implicated in RNA–
protein interactions are circled in RNA sequences in Figure
9. Accordingly, one would expect that replacement of the
guide strand phosphates with amides might significantly de-
crease RNAi activity, while similar modification of the pas-
senger strand might be tolerated at certain positions. This
would be in line with previous results on other modifications
that showed the guide strand to be more sensitive to back-
bone modifications than the passenger strand (26,51). From
this perspective, the results in Figure 5 were encouraging
because all three phosphates in the modified guide strand
siRNA 1 form hydrogen bonds to Piwi and Paz domains of
Ago in crystal structures (39,42,43), yet the loss of activity
was less than that observed for a single AM-PNA linkage in
the middle of the guide strand (26). These results suggest a
hypothesis that amide C = O may be able to interact with
amino acid side chains of Ago in a manner similar to the
non-bridging P–O of unmodified siRNAs.
Figure 8. Comparison of the effect of 5´-phosphorylation on silencing ac-
tivity of amide siRNAs. For other notes, see Figure 5.
The above hypothesis was also supported by the lack of
strong correlation between RNAi activities of modified pas-
senger strands and whether the phosphate modified was or
was not shown to hydrogen bond with Ago. Instead, the re-
sults suggested that modification of the 5´-end of the pas-
senger strand might have an unexpected beneficial effect.
While future studies will be needed to confirm the general-
ity of these results, the observation that siRNAs having the
heavily modified passenger strand siRNA 6 were among the
most active was surprising and encouraging. The increase in
RNAi activity of siRNA 6 is most likely caused by favorable
loading of the associated guide strand into the RISC.
Perhaps the most significant result of the present study
is that the combination of highly modified siRNA 6 and
siRNA 1 (an siRNA with eight amide linkages) was more ac-
tive than the unmodified control. This result leads us to hy-
pothesize that RNAi may tolerate significant guide strand
modifications when combined with properly modified pas-
senger strands that enhance the RISC loading. It is con-
ceivable that optimization of siRNAs for in vivo applica-
tions will require more extensive amide modification than
in the present study. Future work will explore introduction
of consecutive amide linkages in siRNAs. Toward this goal
we recently reported that introduction of three consecutive
amide linkages between the four uridines at the 5´-end of
siRNA PPIBHFS308 had relatively small impact on RNAi
activity (52). It should be noted that the three consecutive
amide linkages did not include the 5´-UaA modification (as
in siRNA 10 and 11) that may be the most beneficial for
RNAi activity. Taken together, our results suggest that op-
timized combination of amide-modified guide and passen-
Figure 9. Comparison of the abilities of 2´-OMe (TSB) and amide-
modified (siRNA 6) passenger strands to enhance RNAi activity of amide-
modified guide strand siRNA 1. Phosphates implicated in RNA–protein
interactions are circled in RNA sequences; for other notes, see Figure 5.
(21,22). More recently, we used NMR and osmotic stress
to show that AM1 linkages did not disturb the conforma-
tion and hydration of double-stranded RNA (23). Before
the present study, a crystal structure of an AM1 modified
RNA duplex that would confirm these notions and provide
a detailed structure of amide hydration had not been solved.
The present crystal structure provides unique insights
into the conformation and hydration of the amide link-
age and confirms our earlier observations that AM1 amide
is surprisingly efficient in mimicking the phosphate back-
bone in RNA (22,23). Consistent with our previous osmotic
stress data (23), amide linkages are well hydrated by tandem
waters linking the carbonyl group and phosphate oxygens
(Figure 4B). The amide conformation in the present struc-
ture is similar to that observed in our earlier NMR struc-
ture (23). Our most recent NMR and osmotic stress study

6550 Nucleic Acids Research, 2014, Vol. 42, No. 10
9. Lebreton,J., Waldner,A., Fritsch,V., Wolf,R.M. and De
Mesmaeker,A. (1994) Comparison of two amides as backbone
replacement of the phosphodiester linkage in oligodeoxynucleotides.
Tetrahedron Lett., 35, 5225–5228.
ger strands may have potential to improve the properties of
siRNAs critical for in vivo applications.
10. Idziak,I., Just,G., Damha,M.J. and Giannaris,P.A. (1993) Synthesis
and hybridization properties of amide-linked thymidine dimers
incorporated into oligodeoxynucleotides. Tetrahedron Lett., 34,
5417–5420.
11. De Mesmaeker,A., Waldner,A., Lebreton,J., Hoffmann,P., Fritsch,V.,
Wolf,R.M. and Freier,S.M. (1994) Amides as a new type of backbone
modifications in oligonucleotides. Angew. Chem., Int. Ed. Engl., 33,
226–229.
CONCLUSION
Our crystallographic and RNAi activity studies are consis-
tent with earlier biophysical and NMR results (23) and,
taken together, strongly suggest that amides are excellent
structural mimics of the phosphate backbone in RNA and
may have the potential to improve the properties of siR-
NAs. The crystal structure of an amide-modified RNA il-
lustrates how the amide linkage is accommodated in an A-
form duplex. Consistent with previous osmotic stress results
(23), amides are well hydrated. The amide linkages are not
only tolerated at internal positions of both guide and pas-
senger strands of siRNAs, but may also increase the silenc-
ing activity when placed near the 5´-end of the passenger
strand. As a result, an siRNA containing eight amide link-
ages was more active than the unmodified control. Taken to-
gether, these results are encouraging for further exploration
of amides and other non-ionic backbone modifications as
means of improving the properties of siRNAs.
12. Lebreton,J., Waldner,A., Lesueur,C. and De Mesmaeker,A. (1994)
Antisense oligonucleotides with alternating
phosphodiester-”amide-3” linkages. Synlett, 137–140.
13. Blommers,M.J.J., Pieles,U. and De Mesmaeker,A. (1994) An
approach to the structure determination of nucleic acid analogs
hybridized to RNA. NMR studies of a duplex between 2’-OMe RNA
and an oligonucleotide containing a single amide backbone
modification. Nucleic Acids Res., 22, 4187–4194.
14. Nina,M., Fonne-Pfister,R., Beaudegnies,R., Chekatt,H., Jung,P.M.J.,
Murphy-Kessabi,F., De Mesmaeker,A. and Wendeborn,S. (2005)
Recognition of RNA by amide modified backbone nucleic acids:
molecular dynamics simulations of DNA-RNA hybrids in aqueous
solution. J. Am. Chem. Soc., 127, 6027–6038.
15. De Mesmaeker,A., Lesueur,C., Bevierre,M.O., Waldner,A.,
Fritsch,V. and Wolf,R.M. (1996) Amide backbones with
conformationally restricted furanose rings: Highly improved affinity
of the modified oligonucleotides for their RNA complements. Angew.
Chem., Int. Ed., 35, 2790–2794.
16. De Mesmaeker,A., Lebreton,J., Jouanno,C., Fritsch,V., Wolf,R.M.
and Wendeborn,S. (1997) Amide-modified oligonucleotides with
preorganized backbone and furanose rings. Highly increased
thermodynamic stability of the duplexes formed with their RNA and
DNA complements. Synlett, 1287–1290.
17. Robins,M.J., Sarker,S., Xie,M., Zhang,W. and Peterson,M.A. (1996)
Nucleic acid related compounds. 90. Synthesis of 2’,3’-fused (3.3.0)
␥-butyrolactone-nucleosides and coupling with amino-nucleosides to
give amide-linked nucleotide-dimer analogs. Tetrahedron Lett., 37,
3921–3924.
18. Robins,M.J., Zou,R., Guo,Z. and Wnuk,S.F. (1996) Nucleic acid
related compounds. 93. A solution for the historic problem of
regioselective sugar-base coupling to produce 9-glycosylguanines or
7-glycosylguanines. J. Org. Chem., 61, 9207–9212.
19. Peterson,M.A., Nilsson,B.L., Sarker,S., Doboszewski,B., Zhang,W.
and Robins,M.J. (1999) Amide-linked ribonucleoside dimers derived
from 5’-amino-5’-deoxy- and 3’-(carboxymethyl)-3’-deoxynucleoside
precursors. J. Org. Chem., 64, 8183–8192.
20. Robins,M.J., Doboszewski,B., Nilsson,B.L. and Peterson,M.A.
(2000) Nucleic acid related compounds. 112. Synthesis of
amide-linked [(3’)CH2CO-NH(5’)] nucleoside analogs of small
oligonucleotides. Nucleosides Nucleotides Nucleic Acids, 19, 69–86.
21. Rozners,E. and Stro¨mberg,R. (1997) Synthesis and properties of
oligoribonucleotide analogs having amide (3’-CH2-CO-NH-5’)
internucleoside linkages. Nucleosides Nucleotides, 16, 967–970.
22. Rozners,E., Katkevica,D., Bizdena,E. and Stro¨mberg,R. (2003)
Synthesis and properties of RNA analogs having amides as
interuridyl linkages at selected positions. J. Am. Chem. Soc., 125,
12125–12136.
23. Selvam,C., Thomas,S., Abbott,J., Kennedy,S.D. and Rozners,E.
(2011) Amides as excellent mimics of phosphate linkages in RNA.
Angew. Chem., Int. Ed., 50, 2068–2070.
24. Iwase,R., Toyama,T. and Nishimori,K. (2007) Solid-phase synthesis
of modified RNAs containing amide-linked oligoribonucleosides at
their 3’-end and their application to siRNA. Nucleosides Nucleotides
Nucleic Acids, 26, 1451–1454.
ACCESSION NUMBER
PDB ID 4O41
SUPPLEMENTARY DATA
Supplementary Data are available at NAR Online.
FUNDING
National Institutes of Health [R01 GM55237 to M.E., R01
GM71461 to E.R.]; National Science Foundation [CHE-
0922815 to E.R.].
Conflict of interest statement. None declared.
REFERENCES
1. Watts,J.K. and Corey,D.R. (2012) Silencing disease genes in the
laboratory and the clinic. J. Pathol., 226, 365–379.
2. Deleavey,G.F. and Damha,M.J. (2012) Designing chemically modified
oligonucleotides for targeted gene silencing. Chem. Biol., 19, 937–954.
3. Levin,A.A. (1999) A review of issues in the pharmacokinetics and
toxicology of phosphorothioate antisense oligonucleotides. Biochim.
Biophys. Acta, 1489, 69–84.
4. Freier,S.M. and Altmann,K.H. (1997) The ups and downs of nucleic
acid duplex stability: structure-stability studies on
chemically-modified DNA:RNA duplexes. Nucleic Acids Res., 25,
4429–4443.
5. De Mesmaeker,A., Altmann,K.-H., Waldner,A. and Wendeborn,S.
(1995) Backbone modifications in oligonucleotides and peptide
nucleic acid systems. Curr. Opin. Struct. Biol., 5, 343–355.
6. Lebreton,J., De Mesmaeker,A., Waldner,A., Fritsch,V., Wolf,R.M.
and Freier,S.M. (1993) Synthesis of thymidine dimer derivatives
containing an amide linkage and their incorporation into
oligodeoxyribonucleotides. Tetrahedron Lett., 34, 6383–6386.
7. De Mesmaeker,A., Lebreton,J., Waldner,A., Fritsch,V., Wolf,R.M.
and Freier,S.M. (1993) Amides as substitute for the phosphodiester
linkage in antisense oligonucleotides. Synlett, 733–736.
8. De Mesmaeker,A., Lebreton,J., Waldner,A., Fritsch,V. and
Wolf,R.M. (1994) Replacement of the phosphodiester linkage in
oligonucleotides: comparison of two structural amide isomers.
Bioorg. Med. Chem. Lett., 4, 873–878.
25. Iwase,R., Kurokawa,R. and Ueno,J. (2009) Synthesis of modified
double stranded RNAs containing duplex regions between
amide-linked RNA and RNA at both ends and enhanced nuclease
resistance. Nucleic Acids Symp. Ser., 53, 119–120.
26. Gong,W. and Desaulniers,J.-P. (2012) Gene-silencing properties of
siRNAs that contain internal amide-bond linkages. Bioorg. Med.
Chem. Lett., 22, 6934–6937.

Nucleic Acids Research, 2014, Vol. 42, No. 10 6551
27. Potenza,N., Moggio,L., Milano,G., Salvatore,V., Di Blasio,B.,
Russo,A. and Messere,A. (2008) RNA interference in mammalia cells
by RNA-3’-PNA chimeras. Int. J. Mol. Sci., 9, 299–315.
28. Puglisi,J.D. and Tinoco,I. Jr. (1989) Absorbance melting curves of
RNA. Methods Enzymol., 180, 304–325.
29. Berger,I., Kang,C., Sinha,N., Wolters,M. and Rich,A. (1996) A
highly efficient 24-condition matrix for the crystallization of nucleic
acid fragments. Acta Cryst. D, 52, 465–468.
43. Sheng,G., Zhao,H., Wang,J., Rao,Y., Tian,W., Swarts,D.C., van der
Oost,J., Patel,D.J. and Wang,Y. (2014) Structure-based cleavage
mechanism of Thermus thermophilus Argonaute DNA guide
strand-mediated DNA target cleavage. Proc. Natl. Acad. Sci. U. S. A.,
111, 652–657.
44. Frank,F., Sonenberg,N. and Nagar,B. (2010) Structural basis for
5’-nucleotide base-specific recognition of guide RNA by human
AGO2. Nature, 465, 818–822.
30. Kabsch,W. (1993) Automatic processing of rotation diffraction data
from crystals of initially unknown symmetry and cell constants. J.
Appl. Cryst., 26, 795–800.
31. Sheldrick,G.M. and Schneider,T.R. (1997) SHELXL: high-resolution
refinement. Methods Enzymol., 277, 319–343.
32. Emsley,P. and Cowtan,K. (2004) Coot: model-building tools for
molecular graphics. Acta Cryst. D, 60, 2126–2132.
33. Murshudov,G.N., Vagin,A.A. and Dodson,E.J. (1997) Refinement of
macromolecular structures by the maximum-likelihood method. Acta
Cryst. D,53, 240–255.
34. Bruenger,A.T. (1992) Free R value: a novel statistical quantity for
assessing the accuracy of crystal structures. Nature, 355, 472–475.
35. Kim,S.H., Shin,W.C. and Warrant,R.W. (1985) Heavy metal
ion-nucleic acid interaction. Methods Enzymol., 114, 156–167.
36. Tanaka,Y., Fujii,S., Hiroaki,H., Sakata,T., Tanaka,T., Uesugi,S.,
Tomita,K.-i. and Kyogoku,Y. (1999) A’-form RNA double helix in
the single crystal structure of r(UGAGCUUCGGCUC). Nucleic
Acids Res., 27, 949–955.
37. Saenger,W. (1984) Principles of Nucleic Acid Structure. Springer, New
York.
38. Egli,M., Portmann,S. and Usman,N. (1996) RNA hydration: A
detailed look. Biochemistry, 35, 8489–8494.
39. Wang,Y., Juranek,S., Li,H., Sheng,G., Tuschl,T. and Patel,D.J. (2008)
Structure of an argonaute silencing complex with a seed-containing
guide DNA and target RNA duplex. Nature, 456, 921–926.
40. Ma,J.B., Yuan,Y.R., Meister,G., Pei,Y., Tuschl,T. and Patel,D.J.
(2005) Structural basis for 5’-end-specific recognition of guide RNA
by the A. fulgidus Piwi protein. Nature, 434, 666–670.
41. Parker,J.S., Roe,S.M. and Barford,D. (2005) Structural insights into
mRNA recognition from a PIWI domain-siRNA guide complex.
Nature, 434, 663–666.
45. Chen,P.Y., Weinmann,L., Gaidatzis,D., Pei,Y., Zavolan,M., Tuschl,T.
and Meister,G. (2008) Strand-specific 5’-O-methylation of siRNA
duplexes controls guide strand selection and targeting specificity.
RNA, 14, 263–274.
46. Jackson,A.L., Burchard,J., Leake,D., Reynolds,A., Schelter,J.,
Guo,J., Johnson,J.M., Lim,L., Karpilow,J., Nichols,K. et al. (2006)
Position-specific chemical modification of siRNAs reduces
“off-target” transcript silencing. RNA, 12, 1197–1205.
47. Braasch,D.A., Jensen,S., Liu,Y., Kaur,K., Arar,K., White,M.A. and
Corey,D.R. (2003) RNA interference in mammalian cells by
chemically-modified RNA. Biochemistry, 42, 7967–7975.
48. Amarzguioui,M., Holen,T., Babaie,E. and Prydz,H. (2003) Tolerance
for mutations and chemical modifications in a siRNA. Nucleic Acids
Res., 31, 589–595.
49. Harborth,J., Elbashir,S.M., Vandenburgh,K., Manninga,H.,
Scaringe,S.A., Weber,K. and Tuschl,T. (2003) Sequence, chemical,
and structural variation of small interfering RNAs and short hairpin
RNAs and the effect on mammalian gene silencing. Antisense Nucl.
Acid Drug Dev., 13, 83–105.
50. Yang,X., Sierant,M., Janicka,M., Peczek,L., Martinez,C., Hassell,T.,
Li,N., Li,X., Wang,T. and Nawrot,B. (2012) Gene silencing activity of
siRNA molecules containing phosphorodithioate substitutions. ACS
Chem. Biol., 7, 1214–1220.
51. Hall,A.H.S., Wan,J., Shaughnessy,E.E., Ramsay Shaw,B. and
Alexander,K.A. (2004) RNA interference using boranophosphate
siRNAs: structure-activity relationships. Nucleic Acids Res., 32,
5991–6000.
52. Tanui,P., Kennedy,S.D., Lunstad,B.D., Haas,A., Leake,D. and
Rozners,E. (2014) Synthesis, biophysical studies and RNA
interference activity of RNA having three consecutive amide linkages.
Org. Biomol. Chem., 12, 1207–1210.
42. Ma,J.B., Ye,K. and Patel,D.J. (2004) Structural basis for
overhang-specific small interfering RNA recognition by the PAZ
domain. Nature, 429, 318–322.