5′-C-Malonyl RNA: Small Interfering RNAs Modified with 5′-Monophosphate Bioisostere Demonstrate Gene Silencing Activity

Article abstract of DOI:10.1021/acschembio.5b00654

5′-Phosphorylation is a critical step in the cascade of events that leads to loading of small interfering RNAs (siRNAs) into the RNA-induced silencing complex (RISC) to elicit gene silencing. 5′-Phosphorylation of exogenous siRNAs is generally accomplished by a cytosolic Clp1 kinase, and in most cases, the presence of a 5′-monophosphate on synthetic siRNAs is not a prerequisite for activity. Chemically introduced, metabolically stable 5′-phosphate mimics can lead to higher metabolic stability, increased RISC loading, and higher gene silencing activities of chemically modified siRNAs. In this study, we report the synthesis of 5′-C-malonyl RNA, a 5′-monophosphate bioisostere. A 5′-C-malonyl-modified nucleotide was incorporated at the 5′-terminus of chemically modified RNA oligonucleotides using solid-phase synthesis. In vitro silencing activity, in vitro metabolic stability, and in vitro RISC loading of 5′-C-malonyl siRNA was compared to corresponding 5′-phosphorylated and 5′-nonphosphorylated siRNAs. The 5′-C-malonyl siRNAs showed sustained or improved in vitro gene silencing and high levels of Ago2 loading and conferred dramatically improved metabolic stability to the antisense strand of the siRNA duplexes. In silico modeling studies indicate a favorable fit of the 5′-C-malonyl group within the 5′-phosphate binding pocket of human Ago2MID domain.

Full text of DOI:10.1021/acschembio.5b00654

Articles
pubs.acs.org/acschemicalbiology
5′‑C‑Malonyl RNA: Small Interfering RNAs Modified with 5′-
Monophosphate Bioisostere Demonstrate Gene Silencing Activity
Ivan Zlatev,† Donald J. Foster,† Jingxuan Liu,† Klaus Charisse,† Benjamin Brigham,† Rubina G. Parmar,†
Vasant Jadhav,† Martin A. Maier,† Kallanthottathil G. Rajeev,† Martin Egli,‡ and Muthiah Manoharan*,†
†Alnylam Pharmaceuticals, 300 Third Street, Cambridge, Massachusetts 02142, United States
‡Department of Biochemistry, Vanderbilt University, Nashville, Tennessee 37232, United States
S
* Supporting Information
ABSTRACT: 5′-Phosphorylation is a critical step in the cascade of events that leads to
loading of small interfering RNAs (siRNAs) into the RNA-induced silencing complex
(RISC) to elicit gene silencing. 5′-Phosphorylation of exogenous siRNAs is generally
accomplished by a cytosolic Clp1 kinase, and in most cases, the presence of a 5′-
monophosphate on synthetic siRNAs is not a prerequisite for activity. Chemically
introduced, metabolically stable 5′-phosphate mimics can lead to higher metabolic stability,
increased RISC loading, and higher gene silencing activities of chemically modified siRNAs.
In this study, we report the synthesis of 5′-C-malonyl RNA, a 5′-monophosphate
bioisostere. A 5′-C-malonyl-modified nucleotide was incorporated at the 5′-terminus of
chemically modified RNA oligonucleotides using solid-phase synthesis. In vitro silencing
activity, in vitro metabolic stability, and in vitro RISC loading of 5′-C-malonyl siRNA was
compared to corresponding 5′-phosphorylated and 5′-nonphosphorylated siRNAs. The 5′-
C-malonyl siRNAs showed sustained or improved in vitro gene silencing and high levels of
Ago2 loading and conferred dramatically improved metabolic stability to the antisense
strand of the siRNA duplexes. In silico modeling studies indicate a favorable fit of the 5′-C-malonyl group within the 5′-phosphate
binding pocket of human Ago2MID domain.
ntroduced by Harris Friedman in 1950,1 the concept of
(RNAi) pathway. RNAi is driven by Dicer-processed
I“bioisosteres” defines structurally different compounds that endogenous small interfering RNAs (siRNAs), which are 21−
23 nucleotide-long RNA duplexes with 5′-monophos-
phates.23,24 siRNAs are loaded into the polyprotein RNA-
induced silencing complex (RISC), where the two RNA strands
are dissociated, and the active antisense (guide) strand is loaded
into the Argonaute-2 (Ago2) endonuclease. Upon Watson−
Crick base pairing with target mRNA, Ago2 can cleave the
mRNA target.25−27 Synthetic siRNAs added exogenously
operate through the same mechanism. A 5′-phosphate group
on the antisense strand of the siRNA duplex is a requisite for
optimal loading on to the RISC machinery;28,29 after loading,
the terminal phosphate is buried within the basic 5′-phosphate-
binding pocket of the Ago2 MID domain. The 5′-terminal
phosphate contributes to the antisense strand selectivity and its
correct positioning for accurate mRNA cleavage.30,31 Synthetic
siRNAs bearing a 5′-hydroxyl (OH) group mediate RNAi
function, as the siRNA is phosphorylated by an endogenous
Clp1 kinase.29 When phosphorylation at the 5′ end of the
antisense strand is blocked, gene silencing is hampered.9,31,32 A
recent study from Kenski et al. have described the in vivo
evaluation of liposome-formulated siRNAs bearing a synthetic
5′-phosphate compared to the nonphosphorylated versions.33
induce similar biological effects and has been widely exploited
in drug design.2,3 Bioisosteres involve the introduction of
changes directed toward modulating or optimizing properties of
drugs such as potency, selectivity, toxicity, and/or metabolic
profile.2 The concept of bioisosterism also provides a starting
point for exploration of novel structural elements that offer an
extensive palette in biomimicry-induced functionalities.2,3
As nucleotide and nucleic acid analogs have found an
increasingly important place in modern drug discovery,4−8
various bioisosteres of phosphate have been developed in order
to mitigate the undesirable pharmacological features of the
phosphate group.2,9,10 The isosteres that have been explored
include 5′-phosphonates,9,11−13 squaric acid mono- and
diamides designed as electronic mimetics,14,15 and different
acetal linkages probed as shape mimetics.16−18 Recently,
Herdewijn et al. developed a series of unique pyrophosphate
mimics with electronic and topological isosterism with natural
dNTPs.19,20 Among these structures, several are efficiently
incorporated into DNA by HIV-RT-1.21 Smietana et al.
reported the synthesis of a full series of 5′-borono-DNA
nucleotides against human TMP kinase, carefully designed
using semiempirical calculations as “nearly perfect mimic” of
the natural 5′-monophosphates.22
Received: August 12, 2015
Accepted: December 17, 2015
We focus here on the use of 5′-monophosphate bioisosteres
in RNA molecules that operate through the RNA interference
© XXXX American Chemical Society
A
DOI: 10.1021/acschembio.5b00654
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
These siRNAs have similar in vivo activity, indicating that
phosphorylation of the siRNA is not a rate-limiting step in
vivo.33 The addition of synthetic 5′-phosphate is necessary,
however, in certain cases of extensively chemically modified
siRNA duplexes that are not recognized by endogenous
kinase.29 It has been also shown that the presence of the 5′-
phosphate group is of critical importance for reduction in
mRNA levels triggered by single-stranded siRNAs.27,32,34−36
Lima et al. analyzed the effects of several 5′-terminal
modifications on the ss-siRNA activity in vitro.9 The loss of
natural 5′-phosphate in vivo due to phosphatase activity, the
introduction of metabolically stable, phosphonate bioisosteres
was required for efficient in vivo silencing.9 Synthesis of such
modified phosphate mimics is not trivial and requires multistep
syntheses.
Use of unformulated, chemically modified siRNAs con-
jugated to N-acetylgalactosamine (GalNAc), which mediates
uptake into hepatocytes, elicited RNAi mediated degradation of
targeted mRNA in the hepatocytes following subcutaneous
administration.38 This approach for delivery of siRNAs has
been recognized as therapeutically relevant, and the screening
of novel chemical modifications on siRNA-GalNAc conjugates
that increase RISC loading, gene silencing, and metabolic
stability has become an important aspect in current and future
siRNA drug design. Here, we describe the design and synthesis
of a novel metabolically stable 5′-monophosphate bioisostere in
siRNAs, the 5′-C-malonyl group. Malonyl phosphate mimetics
have been successfully used in the small-molecule field, mainly
as nonhydrolyzable phosphotyrosyl (pTyr) analogs.39−41 To
the best of our knowledge, use of these mimics has not yet been
reported in the fields of nucleotide or oligonucleotide
therapeutics.
phosphatases and 5′-exonucleases. Like the 5′-phosphate, the
5′-C-malonyl is a dianion at physiological pH due to two acid
dissociation constants, with pKa values of the malonic acid
being 2.83 and 5.69.43 The malonate group is also known for its
excellent chelating properties, and its mimicry of a
pyrophosphate group.20 Furthermore, free energy perturbation
(FEP) calculations predict that 5′-C-malonyl adenosine
requires more energy to desolvate but has more favorable
interactions with the residues in the phosphate binding site of
fructose 1,6-bisphosphatase when compared with adenosine 5′-
monophosphate.44 For the evaluation of 5′-C-malonyl RNA in
the context of siRNAs, we chose to introduce the malonate-
modified nucleotide at the 5′-extremity of the fully modified
antisense strand of siRNAs targeted to silence expression of the
human phosphatase and tensin homologue (PTEN) and the
apolipoprotein B (ApoB) genes. The siRNA chemistry
architecture is based on analysis of chemically modified
siRNA in the context of GalNAc-siRNAs previously reported
by our group.38
Chemical Synthesis of 5′-C-Malonyl RNA. For the
introduction of the 5′-C-malonyl group, we followed a
procedure similar to the one previously described by Fiandor
et al.,45 namely an SN2 reaction between a 5′-deoxy-5′-halo-
nucleoside precursor and dialkyl malonate. The synthetic route
is depicted in Scheme 1. The N3-benzyloxymethyl (BOM)
protected uridine derivative 2 was obtained in quantitative yield
from the commercially available 2′-O-methyl-3′-O-tert-butyldi-
methylsilyl-uridine (1) upon treatment with BOM-chloride.46,47
The BOM protection on N3 in 2 was preferred over the
previously used benzyl45 as it can be more readily removed
under hydrogenolysis conditions. The N3-BOM protected
nucleoside 2 was converted to the corresponding 5′-deoxy-5′-
iodo nucleoside 3 in one step in 92% yield. The 5′-iodide 3 was
then efficiently substituted with the malonyl anion generated in
situ from dimethyl malonate and sodium methoxide (Scheme
1), resulting in a fully protected 5′-deoxy-5′-C-malonyl
nucleoside 4 in 92% yield. Subsequent deprotection of the
N3-BOM group on 4 using 10% Pd/C under a hydrogen
atmosphere in the presence of catalytic formic acid in
isopropanol/water,48 followed by removal of the 3′-O-TBS
group, afforded the nucleoside precursor 6. Finally, phosphi-
tylation of the 3′-hydroxyl group of 6 under standard
conditions afforded the phosphoramidite 7 as the monomer
building block for oligonucleotide synthesis. In addition, we
also prepared the fully deprotected 5′-deoxy-5′-C-malonyl
nucleoside 8 using the conditions applied for 5′-C-malonyl
oligonucleotide deprotection: treatment by 1 M aqueous
piperidine, followed by aqueous ammonia/ethanol treatment
at RT. Using these conditions, complete deprotection of the
malonate methyl esters was achieved, and no decarboxylation of
the free diacid was observed. HRMS analysis confirmed
complete hydrolysis of the methyl esters on 6 to the
corresponding diacid 8 under this condition without any
aminolysis.
RESULTS AND DISCUSSION
■
Choice of the 5′-Monophosphate Bioisostere and
siRNA Design. We have identified a number of non-
hydrolyzable phosphate mimics.42 The 5′-C-malonyl group
(Figure 1) is a promising 5′-phosphate bioisostere for several
reasons. Due to the nonhydrolyzable C-C linkage to the
terminal 5′-nucleotide of RNA, the 5′-C-malonyl should confer
resistance to nuclease degradation, in particular to 5′-
Oligonucleotide synthesis was performed on an oligonucleo-
tide synthesizer. The 5′-deoxy-5′-C-(dimethylmalonyl)-2′-O-
methyluridine-3′-O-phosphoramidite (7) was coupled as the 5′-
terminal nucleotide to yield protected, full-length oligonucleo-
tides. For the 5′-C-malonyl oligonucleotides, the two-step
deprotection procedure involved initial treatment with 1 M
aqueous piperidine for 24 h, followed by evaporation of
piperidine and a subsequent treatment with aqueous ammonia/
ethanol for 36 h. It was critical to perform these deprotections
Figure 1. Chemical structures and calculated total surface charge
density for (A) a 5′-monophosphate nucleoside and (B) a 5′-deoxy-5′-
C-malonyl nucleoside. Generated using MM2 calculation for energy
minimization and total surface charge density calculation functions on
ChemBio 3D. Nucleobase and chemical modifications are not
represented for simplicity.
B
DOI: 10.1021/acschembio.5b00654
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
a
Scheme 1. Synthesis of 5′-Deoxy-5′-C-malonyl Nucleoside Monomer
a
Reagents and conditions: (a) BOM-chloride, DBU, DMF, 30 min, 0 °C, quant.;46,47 (b) methyltriphenoxyphosphonium iodide, DMF, 15 min, rt,
92%;49 (c) sodium methoxide, dimethyl malonate, 1,2-DME, 24 h, reflux, 92%; (d) 10% Pd/C, H2 atm, i-PrOH/H2O (10:1, v/v), 0.05 equiv formic
acid, 12 h, rt, 98%;48 (e) NEt3-3HF, THF, 48 h, rt, 88%; (f) 2-cyanoethyl N,N-diisopropylchlorophosphoramidite, DIEA, DCM, 18 h, rt, 56%; (g) 1
M aq. piperidine, 24 h, rt; then 30% aq. ammonia/ethanol (3:1, v/v), 36 h, rt, quant., Z+ = piperidinium.
Table 1. IC50 Values for 5′-C-Malonyl, 5′-Phosphate, and 5′-OH siRNAs in PTEN and ApoB Silencing in Cell-based Assays
a
a
b
siRNA target
5′-modification
sense strand (5′-3′)
antisense strand (5′-3′)
IC50 (nM)
1 ApoB
2 ApoB
3 ApoB
4 ApoB
5 ApoB
6 ApoB
7 PTEN
8 PTEN
9 PTEN
OH
C•c•UgGaCaUUCaGaAcAaGaA-GalNAc
C•c•UgGaCaUUCaGaAcAaGaA-GalNAc
U•g•UgAcAaAUAuGgGcAuCaA-GalNAc
U•g•UgAcAaAUAuGgGcAuCaA-GalNAc
C•c•UgGaCaUUCaGaAcAaGaA-GalNAc
U•g•UgAcAaAUAuGgGcAuCaA-GalNAc
AaGuAaGgAcCaGaGaCaAdT•dT
u•U•cUuGuUcUgaaUgUcCaGg•g•u
Pu•U•cUuGuUcUgaaUgUcCaGg•g•u
u•U•gAuGcCcAuauUuGuCaCa•a•a
Pu•U•gAuGcCcAuauUuGuCaCa•a•a
Mu•U•cUuGuUcUgaaUgUcCaGg•g•u
Mu•U•gAuGcCcAuauUuGuCaCa•a•a
uUgUcUcUgGuCcUuAcUudT•dT
PuUgUcUcUgGuCcUuAcUudT•dT
MuUgUcUcUgGuCcUuAcUudT•dT
1.03 0.52
0.13 0.03
0.50 0.21
0.13 0.03
0.71 0.41
0.40 0.22
0.71 0.19
0.24 0.01
0.22 0.02
phosphate
OH
phosphate
malonate
malonate
OH
phosphate
malonate
AaGuAaGgAcCaGaGaCaAdT•dT
AaGuAaGgAcCaGaGaCaAdT•dT
a
P indicates 5′-monophosphate; M indicates 5′-malonate; italicized upper case and normal lower case letters indicate 2′-deoxy-2′-fluoro (2′-F) and
2′-O-methyl (2′-OMe) sugar modifications, respectively; • indicates phosphorothioate (PS) linkage; dT indicates 2′-deoxythymidine nucleotide;
GalNAc indicates hydroxyprolynyl trivalent N-acetyl-galactosamine ligand.38 Half maximal inhibitory concentration (IC50) for gene silencing in
b
primary mouse hepatocytes. All values are from triplicate experiments. Mean values SEM are reported.
at RT (∼25 °C) in order to avoid decarboxylation of the
malonic acid.
(Table 1). We observed a strong discrimination between 5′-
phosphorylated and 5′-nonphosphorylated siRNA duplexes.
For each of the two ApoB sequences evaluated, the 5′-
phosphate siRNAs had 5- to 10-fold higher potency than the
nonphosphorylated siRNA. The 5′-C-malonyl phosphate mimic
siRNAs showed levels of ApoB gene silencing somewhat
inferior to that observed with the 5′-phosphate siRNAs (Table
1) but similar to that of the nonphosphorylated siRNAs. The
5′-malonyl and 5′-phosphate PTEN targeting siRNAs conferred
similar advantages relative to the nonphosphorylated siRNA.
Both were approximately 3-fold more potent than the 5′-OH
PTEN siRNA (Table 1 and Figure 2). The difference in
potency between 5′-nonphosphorylated and 5′-phosphorylated
(or phosphate mimic) siRNAs is possibly sequence dependent
and is likely due to differences in 5′-phosphorylation
In Vitro Gene Silencing. The 5′-C-malonyl siRNAs were
evaluated in vitro against two different targets, ApoB and PTEN
(Table 1). To investigate the effect of chemical phosphor-
ylation/phosphate mimicry at the 5′-end of siRNAs antisense
strands, siRNA duplexes with 5′-C-malonyl, 5′-phosphate, and
5′-OH termini were synthesized. The siRNAs were fully
chemically modified with alternating 2′-deoxy-2′-fluoro (2′-F)
and 2′-O-methyl (2′-OMe) sugar-modified nucleotides. In the
case of siRNA sequences targeting ApoB (siRNAs 1−6 ApoB),
we prepared GalNAc-siRNA conjugates as previously de-
scribed.38 The ApoB-targeting GalNAc-siRNAs were delivered
into primary mouse hepatocytes by transfection, and gene
silencing was measured as a function of siRNA concentration
C
DOI: 10.1021/acschembio.5b00654
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 2. Dose−response curves for 5′-OH, 5′-C-malonyl, and 5′-phosphate PTEN siRNAs in primary mouse hepatocytes. All values are from
triplicate experiments.
efficiencies in the cells.29 In the case of the PTEN siRNA, the
5′-C-malonyl phosphate mimic siRNAs performed better than
its 5′-nonphosphorylated counterpart.
as very little of the full-length 9 PTEN sense strand was
detected after 4 h of incubation with tritosomes (Figure 3).
However, significant stabilization of the 5′ N-1 sense strand was
observed in the case of the stabilized 9 PTEN 5′-C-malonyl
siRNA duplex, even after 24 h of incubation (Figure S1,
Supporting Information). This observation demonstrates that a
metabolically stable antisense strand did not fully protect the
sense strand by stabilizing the entire duplex siRNA. However,
after cleavage of the 5′-terminal nucleotide of the sense strand,
the 5′-C-malonyl antisense strand, by its inherently higher
stability toward 5′-exonuclease degradation, was able to
stabilize the N-1 sense strand containing duplex, possibly by
maintaining the daughter duplex integrity and hence exerting
more resistance to nucleolytic attacks.
It is evident from our results that in vitro potency of
extensively modified siRNAs, currently being used in the
context of GalNAc-siRNA conjugates,38 depends on the
presence or absence of the 5′-phosphate. The introduction of
5′-phosphate bioisosteres that preserve the intrinsic RNAi gene
silencing potency and confer higher metabolic stability would
be of particular interest for siRNAs used clinically. siRNAs
chemically synthesized with a natural 5′-phosphate are not
suitable for in vivo application, as the terminal 5′-phosphate is
swiftly removed in vivo by the actions of phosphatases.9
Tritosome Stability of 5′-C-Malonyl siRNAs. To evaluate
the stability of 5′-C-malonyl siRNAs, we used a rat liver
tritosome assay. Liver tritosomes are used as a model for
lysosomes.50 siRNA duplexes were incubated with rat liver
tritosomes for 4 and 24 h, and levels of individual strands of the
siRNA duplex quantified as full-length single strands were
evaluated using HPLC/MS (Supporting Information). The 5′-
C-malonyl isostere conferred a dramatic increase in metabolic
stability relative to the 5′-phosphate- and 5′-OH-containing
antisense strands (Figure 3). The antisense strand of siRNA 9
PTEN was intact after 24 h, whereas the full-length antisense
strands of the 5′-OH (7 PTEN) or the 5′-phosphate (8 PTEN)
siRNAs remained undetectable level after 4 h of incubation
(Figure 3). The increased stability of the 5′-C-malonyl
antisense strand did not stabilize the sense strand significantly,
RNA taken up by hepatocytes is degraded in lysosomes.51,52
An acid endoribonuclease, a nonspecific acid nucleotidase
(phosphatase), and an acid 5′-OH-specific phosphodiesterase
(5′-exonuclease) are the primary actors of RNA degradation in
lysosomes.52 It is likely that a 5′-phosphorylated siRNA duplex
will be rapidly dephosphorylated, resulting in the 5′-OH
siRNA, which becomes a target for the lysosomal 5′-OH-
specific phosphodiesterase. The 5′-deoxy-5′-C-malonyl group
exerts protection against both the phosphatase and the 5′-
exonuclease. The 5′-C-malonyl group thus appears to be an
excellent 5′-phosphate bioisostere by significantly enhancing
the metabolic stability of the antisense strand of siRNAs and by
retaining gene silencing activity.
In Vitro Ago2 Loading of 5′-C-Malonyl siRNAs. To
understand the mechanistic features that enable recognition of
the 5′-C-malonyl bioisostere by the RISC machinery, we
analyzed the interaction of 5′-C-malonyl-modified siRNA with
Ago2 by immunoprecipitation and molecular modeling. The 5′-
C-malonyl antisense strand was efficiently loaded on RISC
confirmed by Ago2 immunoprecipitation from primary mouse
hepatocytes transfected with siRNAs. Primary hepatocytes were
transfected with 5′-C-malonyl, 5′-phosphate, and 5′-OH PTEN-
targeting siRNAs. Ago2 protein was immunoprecipitated from
cell lysate after 24 h, and the amount of single strands
coimmunoprecipitated were determined by RT-PCR. As shown
in Figure 4, each of the siRNAs was loaded on RISC. The 5′-C-
malonyl siRNA antisense strand was present at 3.8-fold higher
levels than the 5′-OH and 5′-phosphate siRNAs. Interestingly,
the in vitro gene silencing potencies of the 5′-C-malonyl and 5′-
phosphate PTEN siRNAs were similar (Table 1 and Figure 2).
A plausible explanation of not showing improved potency by
the modified siRNA despite higher RISC loading is poor
catalytic efficiency of the malonylated antisense strand loaded
Figure 3. Stabilities of 5′-C-malonyl, 5′-phosphate, and 5′-OH siRNAs
incubated in rat liver tritosomes. siRNAs were incubated at 0.4 mg
mL−1 (ca. 5 μM) concentration for 4 and 24 h in the presence of
tritosomes. Percent full-length strand was determined by HPLC. Data
were normalized to untreated controls.
D
DOI: 10.1021/acschembio.5b00654
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 5. Superimposition of the modeled complex between 5′-deoxy-
5′-C-malonyluridine (carbon atoms colored in green) and the
hAgo2MID domain (gray carbon atoms) and the crystal structure of
UMP (magenta carbons) complexed with the MID (light blue
carbons) complex.53 The Matchmaker option in UCSF Chimera54 was
used to generate the overlay. Key residues are labeled. H-bonds
stabilizing the 5′-malonyl moiety are dashed black lines with distances
in Å indicated, and H-bonds stabilizing the 5′-phosphate are thin solid
lines. Hydrogen atoms were omitted from the modeled complex for
clarity.
Figure 4. RISC loaded antisense strands 5′-C-malonyl, 5′-phosphate,
and 5′-OH siRNAs determined by immunoprecipitation of Ago2 from
primary mouse hepatocytes and by RT-PCR amplification of the
Ago2-loaded single strands. siRNAs 7, 8, and 9 were transfected into
cells at 10 nM. Levels of antisense strands are based on incubated nM
siRNA per milligram of cell lysate. Levels of endogenous miR122 were
determined as controls.
RISC compared to the 5′-phosphorylated siRNA. Although
correlation between RISC loading and in vitro potency or in
vitro stability cannot be clearly established using this analysis,
the data demonstrate that the 5′-C-malonyl isostere-modified
siRNA was recognized in vitro by Ago2.
gene silencing. In this study, we described the synthesis of a
novel metabolically stable 5′-monophosphate bioisostere, the
5′-C-malonyl group. The modification was incorporated at the
5′ terminus of the antisense strand of double stranded siRNAs.
These 5′-C-malonyl siRNAs silenced expression of two
different genes, PTEN and ApoB, in primary mouse hepatocytes
as effectively as or more effectively than corresponding 5′-
phosphate or 5′-OH siRNAs. The 5′-malonate phosphate
mimic conferred significant increase in the metabolic stability of
siRNA duplexes in lysosomal matrices. siRNAs with a malonyl
moiety at the 5′-end of antisense strand were recognized by the
MID domain, the phosphate-binding pocket of Ago2, as
evident from efficient loading of the modified antisense strand
into RISC in vitro, confirmed by Ago2 immunoprecipitation.
The in silico molecular modeling studies showed a favorable fit
of the 5′-C-malonyl group within the 5′-phosphate binding site
of the MID domain of human Ago2. When taken together,
these results indicate that the 5′-C-malonyl, a metabolically
stable 5′-phosphate bioisostere, has excellent biomimicry
properties and high potential for use in therapeutic siRNAs.
We anticipate that use of the 5′-malonyl phosphate mimic will
significantly enhance the therapeutic value of future RNAi-
based drugs.
In Silico Model of the Complex between 5′-Deoxy-5′-
C-malonyluridine and the hAgo2MID Domain. The
interactions of a 5′-deoxy-5′-C-malonyluridine at the NMP
binding site of the hAgo2MID domain53 were modeled using a
molecular mechanics approach (Amber force field in UCSF
Chimera54). The 5′-phosphate of UMP in the crystal structure
of the MID domain complex makes hydrogen-bond-mediated
contacts with Tyr529, Lys533, Gln545, Cys546, and Lys570
(Figure 5). The malonyl carboxylates in the modeled complex
are contacted by Gln545, Lys566, and Lys570. Thus, the
malonyl moiety displays a slight shift relative to the standard
phosphate and likely makes fewer contacts with the protein.
The malonyl is farther from Tyr529 (4.05 Å between the
tyrosine hydroxyl oxygen and one of the carboxylate oxygens)
and Lys533 (4.85 Å between Nζ and the same malonyl oxygen)
than the phosphate. Interestingly, in the complex between
hAgo2 and miR-20a, Lys566 also contributes to the recognition
of the 5′-phosphate moiety (see Figure 5B in cited ref 55); in
our 5′-malonyl model, this interaction is observed. The 5′-
malonyl group was successfully accommodated within the
hAgo2MID binding pocket; no steric or electronic-clash
interactions between enzyme and phosphate mimic were
observed. Hydrogen bonds with two of the five residues that
interact with the 5′ phosphate of UMP likely also form between
5′-malonyl and the protein. The more limited interaction mode
of 5′-malonyl compared to 5′-phosphate is in line with the
slight reduction of in vitro potency observed between some 5′-
phosphate and 5′-malonate siRNAs (Table 1).
METHODS
■
Cell Culture and Transfection. Primary mouse hepatocytes were
obtained from Life Technologies and cultured in Williams E Medium
with 10% fetal bovine serum (FBS). Transfection was carried out by
adding 4.9 μL of Opti-MEM plus 0.1 μL of Lipofectamine RNAiMax
(Invitrogen) per well to 5 μL of each siRNA duplex at the desired
concentration to an individual well in a 384-well plate. The mixture
was incubated at RT for 20 min, and 40 μL of complete growth media
containing 5000 cells was added to the siRNA mixture. Cells were
incubated for 24 h prior to RNA isolation. A similar procedure was
followed for the transfection of 10 000 000 cells and scaled
accordingly. Dose response experiments were done using eight 6-
fold serial dilutions over the range of 20 nM to 75 pM or 50 nM to
187.5 pM.
CONCLUSIONS
■
Various phosphate bioisostere groups have been synthesized,
but few have been evaluated in therapeutic RNA molecules. 5′-
Phosphorylation of siRNAs is a critical step for RNAi-mediated
E
DOI: 10.1021/acschembio.5b00654
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
RNA Isolation. RNA was isolated using a Dynabeads mRNA
Isolation Kit (Invitrogen). Cells were lysed in 75 μL of Lysis/Binding
Buffer containing 3 μL of beads per well and mixed for 10 min on an
electrostatic shaker. Buffers were prepared according to the
manufacturer’s protocol. The washing steps were automated on a
Biotek EL406 using a magnetic plate support. Beads were washed (90
μL) once in buffer A, once in buffer B, and twice in buffer E, with
aspiration steps between washes.
cDNA Synthesis. cDNA synthesis was accomplished with the ABI
High capacity cDNA reverse transcription kit (Applied Biosystems). A
mixture of 1 μL of 10× buffer, 0.4 μL of 25× dNTPs, 1 μL of random
primers, 0.5 μL of reverse transcriptase, 0.5 μL of RNase inhibitor, and
6.6 μL of water per reaction were added per well. Plates were sealed,
agitated for 10 min on an electrostatic shaker, and then incubated at 37
°C for 2 h. Following this, the plates were agitated at 80 °C for 8 min.
Real-time PCR. cDNA (2 μL) was added to a master mix
containing 0.5 μL mouse GAPDH TaqMan Probe (Applied
Biosystems, Cat.# 4308313), 0.5 μL of mouse ApoB or PTEN
TaqMan probes (Applied Biosystems, Cat.# Mm01545156_m1 and
Mm01212532_m1, respectively), and 5 μL of Lightcycler 480 probe
master mix (Roche) per well in a 384-well 50 plate (Roche). Real-time
PCR was done in an ABI 7900HT RT-PCR system (Applied
Biosystems) using the ΔΔCt (RQ) assay. Each duplex and
concentration was tested in four biological replicates. To calculate
relative fold change, real time data were analyzed using the ΔΔCt
method and normalized to assays performed with cells transfected with
10 nM nonspecific siRNA. IC50 values were calculated using a four-
parameter fit model using XLFit.
Tritosome Stability Assay. Rat liver tritosomes (Xenotech,
custom product PR14044) were thawed to RT and diluted to 0.5 units
mL−1 in 20 mM sodium citrate buffer, pH 5.0. Samples were prepared
by mixing 100 μL of 0.5 units mL−1 acid phosphatase tritosomes with
25 μL of 0.4 mg mL−1 siRNA in a microcentrifuge tube. After
incubation for 4 or 24 h in an Eppendorf Thermomixer set to 37 °C
and 300 rpm, 300 μL of Phenomenex Lysis Loading Buffer and 12.5
μL of a 0.4 mg mL−1 internal standard siRNA were added to each
sample. Samples for time 0 were prepared by mixing 100 μL of 0.5
units mL−1 acid phosphatase Tritosomes with 25 μL of 0.4 mg mL−1
siRNA sample, 300 μL of Phenomenex Lysis Loading Buffer, and 12.5
μL of a 0.4 mg mL−1 internal standard siRNA. siRNA was extracted
from each time point sample (0 h, 4 h, 24 h) using a Phenomenex
Clarity OTX Starter Kit. The samples were then resuspended with 500
μL of nuclease free water, and 50 μL of sample was analyzed by LC/
MS. Analytical procedures and LC/MS plots are included as
Supporting Information.
RISC Immunoprecipitation and RT-PCR Assay. siRNA-trans-
fected primary mouse hepatocytes (10 000 000 cells) were lysed in
lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1% NP-40, 0.1%
SDS) with protease inhibitor (Sigma-Aldrich). Lysate concentration
was measured with a protein BCA kit (Thermo Scientific). For each
reaction, 2 mg of total lysate was used. Anti-Ago2 antibody was
purchased from Wako Chemicals (Clone No.: 2D4). Control mouse
IgG was from Santa Cruz Biotechnology (sc-2025). Dynabeads (Life
Technologies) were used to precipitate antibodies. Ago2-associated
siRNA and endogenous miR122 were measured by Stem-Loop RT
followed by TaqMan PCR analysis based on previously published
methods.56
retrieved from the Protein Data Bank (http://www.rcsb.org). Using
the program UCSF Chimera (version 1.5.3),54 all water molecules
from the crystal structure were deleted, and the 5′-phosphate group
was converted to the 5′-C-malonyl moiety. Hydrogen atoms were then
added and the geometry of 5′-deoxy-5′-C-malonyluridine and its
orientation and H-bonding/nonbonded interactions at the 5′-
phosphate pocket of the hAgo2MID domain optimized with the
Amber force field (ff12SB and Gasteiger charges for standard amino
acids and nonstandard residues, respectively), as implemented in
UCSF Chimera.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acschem-
bio.5b00654.
Detailed experimental procedures and analytical data for
compounds and oligonucleotides, NMR spectra of
compounds, and LC/MS data for tritosome assay
(PDF)
Accession Codes
PDB: 3LUJ, 4F3T.
AUTHOR INFORMATION
Corresponding Author
*Tel.: +1 617 551 8319. Fax: +1 617 551 8102. E-mail:
mmanoharan@alnylam.com.
■
Funding
Alnylam Pharmaceuticals.
Notes
The authors declare the following competing financial
interest(s): All Alnylam authors except B.B. are current
employees of the company.
ACKNOWLEDGMENTS
The authors thank F. Morvan (Universite
helpful discussions.
■
́
de Montpellier) for
REFERENCES
■
(1) Friedman, H. L. (1951) Influence of Isosteric Replacements upon
Biological Activity, vol 206, pp 295−358, National Academy of
Sciences-National Research Council Publication, Washington, DC.
(2) Meanwell, N. A. (2011) Synopsis of Some Recent Tactical
Application of Bioisosteres in Drug Design. J. Med. Chem. 54, 2529−
2591.
(3) Patani, G. A., and LaVoie, E. J. (1996) Bioisosterism: A Rational
Approach in Drug Design. Chem. Rev. 96, 3147−3176.
(4) De Clercq, E. (2009) The history of antiretrovirals: key
discoveries over the past 25 years. Rev. Med. Virol. 19, 287−299.
(5) Uhlmann, E., and Peyman, A. (1990) Antisense oligonucleotides:
a new therapeutic principle. Chem. Rev. 90, 543−584.
(6) de Fougerolles, A., Vornlocher, H.-P., Maraganore, J., and
Lieberman, J. (2007) Interfering with disease: a progress report on
siRNA-based therapeutics. Nat. Rev. Drug Discovery 6, 443−453.
(7) Burnett, J. C., Rossi, J. J., and Tiemann, K. (2011) Current
progress of siRNA/shRNA therapeutics in clinical trials. Biotechnol. J.
6, 1130−1146.
Computational Simulation of the Interaction between 5′-
Deoxy-5′-C-malonyluridine and the Human Ago2MID Domain.
The recognition modes of available crystal structures of complexes
between hAgo2MID (amino acids 432−578; residues 440−572 are
resolved in the electron density) and UMP (PDB ID code 3LUJ)53
and full-length hAgo2 and miR-20a (PDB ID code 4F3T)55 revealed
that the recognition modes of 5′-terminal phosphates are very similar.
The only difference between the two structures is that in the complex
with full-length Ago2, a residue from the PIWI domain (Arg812)
makes a contribution to the recognition of the 5′-phosphate. We
therefore used the UMP:MID complex as the basis for modeling the
interaction between 5′-malonyluridine and the hAgo2MID domain.
Three-dimensional coordinates of the UMP:MID complex were
(8) Bumcrot, D., Manoharan, M., Koteliansky, V., and Sah, D. W. Y.
(2006) RNAi therapeutics: a potential new class of pharmaceutical
drugs. Nat. Chem. Biol. 2, 711−719.
(9) Lima, W. F., Prakash, T. P., Murray, H. M., Kinberger, G. A., Li,
W., Chappell, A. E., Li, C. S., Murray, S. F., Gaus, H., Seth, P. P.,
Swayze, E. E., and Crooke, S. T. (2012) Single-Stranded siRNAs
Activate RNAi in Animals. Cell 150, 883−894.
F
DOI: 10.1021/acschembio.5b00654
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
(10) Elliott, T. S., Slowey, A., Ye, Y., and Conway, S. J. (2012) The
use of phosphate bioisosteres in medicinal chemistry and chemical
biology. MedChemComm 3, 735−751.
(30) Jinek, M., and Doudna, J. A. (2009) A three-dimensional view of
the molecular machinery of RNA interference. Nature 457, 405−412.
(31) 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.
(11) Gallier, F., Alexandre, J. A. C., El Amri, C., Deville-Bonne, D.,
́
Peyrottes, S., and Perigaud, C. (2011) 5′,6′-Nucleoside Phosphonate
Analogues Architecture: Synthesis and Comparative Evaluation
towards Metabolic Enzymes. ChemMedChem 6, 1094−1106.
(12) Gallier, F., Lallemand, P., Meurillon, M., Jordheim, L. P.,
(32) Martinez, J., Patkaniowska, A., Urlaub, H., Luhrmann, R., and
̈
Tuschl, T. (2002) Single-Stranded Antisense siRNAs Guide Target
RNA Cleavage in RNAi. Cell 110, 563−574.
(33) Kenski, D. M., Willingham, A. T., Haringsma, H. J., Li, J. J., and
Flanagan, W. M. (2012) In Vivo Activity and Duration of Short
Interfering RNAs Containing a Synthetic 5′-Phosphate. Nucleic Acid
Ther. 22, 90−95.
(34) Haringsma, H. J., Li, J. J., Soriano, F., Kenski, D. M., Flanagan,
W. M., and Willingham, A. T. (2012) mRNA knockdown by single
strand RNA is improved by chemical modifications. Nucleic Acids Res.
40, 4125−36.
(35) Kini, H. K., and Walton, S. P. (2007) In vitro binding of single-
stranded RNA by human Dicer. FEBS Lett. 581, 5611−5616.
(36) Schwarz, D. S., Hutvagner, G., Haley, B., and Zamore, P. D.
(2002) Evidence that siRNAs function as guides, not primers, in the
Drosophila and human RNAi pathways. Mol. Cell 10, 537−548.
(37) Prakash, T. P., Lima, W. F., Murray, H. M., Elbashir, S., Cantley,
W., Foster, D., Jayaraman, M., Chappell, A. E., Manoharan, M.,
Swayze, E. E., and Crooke, S. T. (2013) Lipid Nanoparticles Improve
Activity of Single-Stranded siRNA and Gapmer Antisense Oligonu-
cleotides in Animals. ACS Chem. Biol. 8, 1402−1406.
(38) Nair, J. K., Willoughby, J. L. S., Chan, A., Charisse, K., Alam, M.
R., Wang, Q., Hoekstra, M., Kandasamy, P., Kel’in, A. V., Milstein, S.,
Taneja, N., O’Shea, J., Shaikh, S., Zhang, L., van der Sluis, R. J., Jung,
M. E., Akinc, A., Hutabarat, R., Kuchimanchi, S., Fitzgerald, K.,
Zimmermann, T., van Berkel, T. J. C., Maier, M. A., Rajeev, K. G., and
Manoharan, M. (2014) Multivalent N-Acetylgalactosamine-Conju-
gated siRNA Localizes in Hepatocytes and Elicits Robust RNAi-
Mediated Gene Silencing. J. Am. Chem. Soc. 136, 16958−16961.
(39) Shi, Z.-D., Wei, C.-Q., Lee, K., Liu, H., Zhang, M., Araki, T.,
Roberts, L. R., Worthy, K. M., Fisher, R. J., Neel, B. G., Kelley, J. A.,
Yang, D., and Burke, T. R. (2004) Macrocyclization in the Design of
Non-Phosphorus-Containing Grb2 SH2 Domain-Binding Ligands. J.
Med. Chem. 47, 2166−2169.
(40) Gao, Y., Luo, J., Yao, Z.-J., Guo, R., Zou, H., Kelley, J., Voigt, J.
H., Yang, D., and Burke, T. R. (2000) Inhibition of Grb2 SH2 Domain
Binding by Non-Phosphate-Containing Ligands. 2. 4-(2-Malonyl)-
phenylalanine as a Potent Phosphotyrosyl Mimetic. J. Med. Chem. 43,
911−920.
(41) Rye, C. S., and Baell, J. B. (2005) Phosphate Isosteres in
Medicinal Chemistry. Curr. Med. Chem. 12, 3127−3141.
(42) Manoharan, M., Rajeev, K., Zlatev, I., Swayze, E. E., Prakash, T.
P., Lima, W. (2011) Stabilizing 5′-cap structures for small RNAs for
therapeutic use, WO2011133871.
́
Dumontet, C., Perigaud, C., Lionne, C., Peyrottes, S., and Chaloin, L.
(2011) Structural Insights into the Inhibition of Cytosolic 5′-
Nucleotidase II (cN-II) by Ribonucleoside 5′-Monophosphate
Analogues. PLoS Comput. Biol. 7, e1002295.
(13) Pradere, U., Amblard, F., Coats, S. J., and Schinazi, R. F. (2012)
Synthesis of 5′-Methylene-Phosphonate Furanonucleoside Prodrugs:
Application to D-2′-Deoxy-2′-α-fluoro-2′-β-C-methyl Nucleosides.
Org. Lett. 14, 4426−4429.
(14) Sato, K., Seio, K., and Sekine, M. (2002) Squaryl Group as a
New Mimic of Phosphate Group in Modified Oligodeoxynucleotides:
Synthesis and Properties of New Oligodeoxynucleotide Analogues
Containing an Internucleotidic Squaryldiamide Linkage. J. Am. Chem.
Soc. 124, 12715−12724.
(15) Seio, K., Miyashita, T., Sato, K., and Sekine, M. (2005) Synthesis
and Properties of New Nucleotide Analogues Possessing Squaramide
Moieties as New Phosphate Isosters. Eur. J. Org. Chem. 2005, 5163−
5170.
(16) Pudlo, J. S., Cao, X., Swaminathan, S., and Matteucci, M. D.
(1994) Deoxyoligonucleotides containing 2′,5′ acetal linkages: Syn-
thesis and hybridization properties. Tetrahedron Lett. 35, 9315−9318.
(17) Lin, K.-Y., and Matteucci, M. D. (1996) The synthesis and
hybridization properties of an oligonucleotide containing hexafluor-
oacetone ketal internucleotide linkages. Tetrahedron Lett. 37, 8667−
8670.
(18) Zou, R., and Matteucci, M. D. (1996) Synthesis and
hybridization properties of an oligonucleotide analog containing a
glucose-derived conformation-restricted ribose moiety and 2′, 5′
formacetal linkages. Tetrahedron Lett. 37, 941−944.
(19) Adelfinskaya, O., and Herdewijn, P. (2007) Amino Acid
Phosphoramidate Nucleotides as Alternative Substrates for HIV-1
Reverse Transcriptase. Angew. Chem., Int. Ed. 46, 4356−4358.
(20) Zlatev, I., Giraut, A., Morvan, F., Herdewijn, P., and Vasseur, J.-J.
(2009) δ-Di-carboxybutyl phosphoramidate of 2′-deoxycytidine-5′-
monophosphate as substrate for DNA polymerization by HIV-1
reverse transcriptase. Bioorg. Med. Chem. 17, 7008−7014.
̀
(21) Herdewijn, P., and Marliere, P. (2012) Redesigning the leaving
group in nucleic acid polymerization. FEBS Lett. 586, 2049−2056.
(22) El Amri, C., Martin, A. R., Vasseur, J.-J., and Smietana, M.
(2012) Borononucleotides as Substrates/Binders for Human NMP
Kinases: Enzymatic and Spectroscopic Evaluation. ChemBioChem 13,
1605−1612.
(23) Zamore, P. D., Tuschl, T., Sharp, P. A., and Bartel, D. P. (2000)
RNAi: Double-Stranded RNA Directs the ATP-Dependent Cleavage
of mRNA at 21 to 23 Nucleotide Intervals. Cell 101, 25−33.
(24) Elbashir, S. M., Harborth, J., Lendeckel, W., Yalcin, A., Weber,
K., and Tuschl, T. (2001) Duplexes of 21-nucleotide RNAs mediate
RNA interference in cultured mammalian cells. Nature 411, 494−498.
(25) Ketting, R. F. (2011) The Many Faces of RNAi. Dev. Cell 20,
148−161.
(43) Brown, H. C., McDaniel, D. H., Hafliger, O. Braude, E. A., and
Nachod, F. C. (1955) Determination of Organic Structures by Physical
Methods, Academic Press, New York.
(44) Reddy, M. R., and Erion, M. D. (2001) Calculation of Relative
Binding Free Energy Differences for Fructose 1,6-Bisphosphatase
Inhibitors Using the Thermodynamic Cycle Perturbation Approach. J.
Am. Chem. Soc. 123, 6246−6252.
(26) Filipowicz, W. (2005) RNAi: The Nuts and Bolts of the RISC
Machine. Cell 122, 17−20.
́
(45) Fiandor, J., García-Lopez, M. T., and De las Heras, F. G. (1989)
Synthesis of 5′-C-Chain-Extended Uridines by Reaction of 5′-
Halonucleosides with Malonic Acid Type Derivatives. Nucleosides
Nucleotides 8, 1325−1334.
(27) Lima, W. F., Murray, H., Nichols, J. G., Wu, H., Sun, H.,
Prakash, T. P., Berdeja, A. R., Gaus, H. J., and Crooke, S. T. (2009)
Human Dicer Binds Short Single-strand and Double-strand RNA with
High Affinity and Interacts with Different Regions of the Nucleic
Acids. J. Biol. Chem. 284, 2535−2548.
(46) Kurosu, M., and Li, K. (2009) Synthetic studies towards the
identification of novel capuramycin analogs with mycobactericidal
activity. Heterocycles 77, 217−225.
(28) Nykanen, A., Haley, B., and Zamore, P. D. (2001) ATP
(47) Kurosu, M., Li, K., and Crick, D. C. (2009) Concise Synthesis of
Capuramycin. Org. Lett. 11, 2393−2396.
(48) Aleiwi, B. A., and Kurosu, M. (2012) A reliable Pd-mediated
hydrogenolytic deprotection of BOM group of uridine ureido
nitrogen. Tetrahedron Lett. 53, 3758−3762.
̈
Requirements and Small Interfering RNA Structure in the RNA
Interference Pathway. Cell 107, 309−321.
(29) Manoharan, M. (2004) RNA interference and chemically
modified small interfering RNAs. Curr. Opin. Chem. Biol. 8, 570−579.
G
DOI: 10.1021/acschembio.5b00654
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
(49) Verheyden, J. P. H., and Moffatt, J. G. (1970) Halo sugar
nucleosides. I. Iodination of the primary hydroxyl groups of
nucleosides with methyltriphenoxyphosphonium iodide. J. Org.
Chem. 35, 2319−2326.
(50) Bagshaw, R. D., Pasternak, S. H., Mahuran, D. J., and Callahan, J.
W. (2003) Nicastrin is a resident lysosomal membrane protein.
Biochem. Biophys. Res. Commun. 300, 615−618.
(51) Fujiwara, Y., Furuta, A., Kikuchi, H., Aizawa, S., Hatanaka, Y.,
Konya, C., Uchida, K., Yoshimura, A., Tamai, Y., Wada, K., and
Kabuta, T. (2013) Discovery of a novel type of autophagy targeting
RNA. Autophagy 9, 403−409.
(52) Heydrick, S. J., Lardeux, B. R., and Mortimore, G. E. (1991)
Uptake and degradation of cytoplasmic RNA by hepatic lysosomes.
Quantitative relationship to RNA turnover. J. Biol. Chem. 266, 8790−
8796.
(53) 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.
(54) Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S.,
Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF
ChimeraA visualization system for exploratory research and analysis.
J. Comput. Chem. 25, 1605−1612.
(55) Elkayam, E., Kuhn, C.-D., Tocilj, A., Haase, A. D., Greene, E. M.,
Hannon, G. J., and Joshua-Tor, L. (2012) The Structure of Human
Argonaute-2 in Complex with miR-20a. Cell 150, 100−110.
(56) Chen, C., Ridzon, D. A., Broomer, A. J., Zhou, Z., Lee, D. H.,
Nguyen, J. T., Barbisin, M., Xu, N. L., Mahuvakar, V. R., Andersen, M.
R., Lao, K. Q., Livak, K. J., and Guegler, K. J. (2005) Real-time
quantification of microRNAs by stem−loop RT−PCR. Nucleic Acids
Res. 33, e179.
H
DOI: 10.1021/acschembio.5b00654
ACS Chem. Biol. XXXX, XXX, XXX−XXX