RNA Helicase A Interacts with RISC
in Human Cells and Functions in RISC Loading
G. Brett Robb1and Tariq M. Rana1,*
1Department of Biochemistry and Molecular Pharmacology, University of Massachusetts Medical School, Worcester,
MA 01605, USA
RNA interference is a conserved pathway of
sequence-specific gene silencing that depends
on small guide RNAs and the action of proteins
assembledin the RNA-induced
complex (RISC). Minimally, the action of RISC
requires the endonucleolytic slicer activity of
Argonaute2 (Ago2) directed to RNA targets
whose sequences are complementary to RISC-
incorporated small RNA. To identify RISC
components in human cells, we developed an
affinity-purification strategy to isolate siRNA-
programmed RISC. Here we report the identifi-
cation of RNA helicase A (RHA) as a human
RISC-associated factor. We show that RHA
interacts in human cells with siRNA, Ago2,
TRBP, and Dicer and functions in the RNAi
pathway. In RHA-depleted cells, RNAi was re-
duced as a consequence of decreased intracel-
lular concentration of active RISC assembled
with the guide-strand RNA and Ago2. Our re-
sults identify RHA as a RISC component and
demonstrate that RHA functions in RISC as an
RNA interference (RNAi), the sequence-specific silencing
of gene expression, is directed by small RNAs acting in
concert with several proteins in the RNA-induced silenc-
ing complex (RISC). Although human RISC activity can
be reconstituted in vitro using single-stranded siRNA
and recombinant Argonaute2 (Ago2) (Rivas et al., 2005),
in vitro formation of human RISC using double-stranded
RNAs requires additional factors such as Dicer and trans-
activation-responsive RNA binding protein (TRBP) (Chen-
drimada et al., 2005). RISC fractionates in large molecular
mass complexes containing multiple proteins, many of
which have been identified by immunoaffinity purification
of Ago2 from cell extracts. Some of these proteins appear
to play a functional role in RNAi. For example, depleting
MOV10 (Moloney leukemia virus 10, homolog) or TNRC6B
from cells prevented silencing of microRNA-targeted re-
porter gene expression (Meister et al., 2005). Knockdown
of Dicer, TRBP, or protein activator of the interferon-
induced protein kinase (PACT) reduced the siRNA-
induced silencing of the reporter gene (Chendrimada
et al., 2005; Haase et al., 2005; Lee et al., 2006). Thus,
while siRNA and Ago2 constitute the core endonucleolytic
components of the RISC, other proteins clearly modify
and enhance their function in gene silencing.
RHA (also known as DHX9 and NDHII) is a member of
the DEAH-containing family of RNA helicases (Abdelha-
leem et al., 2003). Originally purified and characterized in
HeLa cells (Lee and Hurwitz, 1992), human RHA was later
identified as the ortholog of Drosophila maleless protein
(Lee and Hurwitz, 1993). RHA enzymatically unwinds
RNA-RNA duplexes as well as RNA-DNA hybrids in a 30
to 50direction, provided that the RNA has unpaired 30
ture helicase core domain, RHA also contains three RNA
binding domains that include two N-terminal double-
stranded RNA binding domains and a C-terminal RGG
box domain, allowing RHA to bind double- and single-
stranded RNA as well as single-stranded DNA (Lee and
The Drosophila RHA ortholog, maleless, functions in
transcriptional upregulation of the male X chromosome
1998). Likewise, RHA plays many roles in transcriptional
regulation of human genes. RHA occupies the promoters
of the p16INK4a (Myohanen and Baylin, 2001) and the
MDR1 (Zhong and Safa, 2004) genes, from which it en-
hances transcription. In addition, RHA interacts directly
with the p65 subunit of nuclear factor-kB and stimulates
its transcriptional regulatory activity (Tetsuka et al.,
2004). RHA also acts as a bridging factor, coupling RNA
polymerase II (RNAPII) to the transcriptional coactivators
BRCA1 and CBP/p300 (Anderson et al., 1998; Nakajima
factor between the SMN (survival of motor neurons) com-
plex and RNAPII, linking RHA to splicing/cotranscriptional
mRNA processing (Pellizzoni et al., 2001).
RHA localizes to both the nucleus and the cytoplasm of
human cells and shuttles between the two compartments
by virtue of a bidirectional nuclear transport domain at the
protein’s C terminus (Tang et al., 1999). RHA binds the
constitutive transport element (CTE) of type D retroviruses
to facilitate expression of CTE-dependent genes and
Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc. 523
plays a role in HIV-1 RRE-dependent gene expression (Li
et al., 1999; Tang et al., 1999). RHA was recently shown to
be required for the translation of JUND and a posttran-
scriptional control element-containing viral mRNA (Hart-
man etal.,2006).Translation of PCEmRNAwasproposed
to be stimulated by RHA remodeling of complex protein-
RNA 50untranslated region structures that inhibit efficient
ribosomal scanning of these transcripts.
Since RHA appears to regulate many aspects of gene
expression, it is not surprising that murine RHA is required
for development. RHA null embryos fail to undergo normal
gastrulation, displaying massive apoptosis of the primitive
ectoderm and resulting in early embryonic lethality (Lee
et al., 1998). In worms, rha-1 null animals are defective
in germline transcriptional control and display reduced
germ cell mitosis and defects in meiosis (Walstrom et al.,
In addition to the siRNA and Ago2 in human RISC,
a number of other proteins such as Dicer, TRBP, PACT,
and MOV10 clearly modify and enhance RISC function in
gene silencing (Rana, 2007). To identify RISC-associated
proteins, we used an approach based on affinity purifica-
tion of the guide siRNA strand incorporated into RISC in
human cells. Here we report the identification of RHA
that specifically interacts with active RISC. Depleting
RHA in cells before programming with siRNA or short hair-
pin RNA (shRNA) reduced gene silencing due to de-
creased RISC formation. RHA depletion also reduced
the recruitment of siRNA to intracellular Ago2, supporting
a role for RHA in RISC loading.
RHA Interacts with Active RISC
To identify proteins that are components of active RISC
containing the guide strand of siRNA, we developed an
affinity-purification strategy (summarized in Figure 1A).
Previously described affinity-purification strategies have
been based on coimmunoprecipitating unidentified pro-
teins with affinity-tagged RISC protein components. Our
approach used 30-LC biotinylated siRNA to isolate RISC.
To verify that this approach didn’t compromise the func-
tional integrity of RISC, we tested the ability of purified
RISC to specifically cleave radiolabeled target RNA
in vitro. Using this scheme, we purified RISC from cells
transfected with siRNA duplexes containing 30-LC biotin
on the antisense or guide strand and found that RISC spe-
cifically andpotently cleaved target RNA(Figure 1A).RISC
purified in parallel from cells programmed with siRNAs of
identical sequence but lacking 30-LC biotin on the anti-
sense strand did not cleave its target RNA. Thus, active
RISC was associated with streptavidin-coated magnetic
beads, and RISC associated with the magnetic beads
was specificallyisolated bythebiotin-streptavidin interac-
tion rather than by nonspecific interactions.
Toanalyze other components ofthe RISC, themagnetic
beads were treated with RNase. The resultant eluate con-
tained several proteins specific to extracts from cells pro-
grammed with biotinylated siRNA (Figure 1B). Mass spec-
trometric analysis of protein components eluted from the
isolated RISC-identified RHA as two bands with approxi-
mate molecular masses of ?145 and ?125 kDa. Consis-
tent with our findings, full-length RHA has a predicted
and observed molecular mass of ?142 kDa. It is currently
unclear whether the smaller RHA species we recovered
reflects posttranslational modification of the protein or
its proteolytic cleavage during preparation and purifica-
tion of the complexes. Our mass spectrometric analysis
also identified HSP90b, a protein previously identified as
an Argonaute-associated protein (Liu et al., 2004, 2005;
Maniataki and Mourelatos, 2005).
To confirm our identification of RHA as a component of
active RISC, we used an alternative approach in which
293T cells were cotransfected with biotinylated siRNA
and plasmids encoding HA-tagged RHA and myc-tagged
Ago2. Immunoblot analysis of streptavidin pullouts from
cell extracts showed that RHA was present with the guide
strand (Figure 2A). In addition, Ago2 and Dicer were pres-
ent in the RISC pulled out with the biotin-containing guide
strand. Taken together, these results show that Ago2,
Dicer, and RHA are components of active human RISC
programmed with siRNA.
To determine if siRNA was associated with RHA in hu-
man cells, we transfected 293T cells with an HA-tagged
RHA vector and immunoprecipitated the protein com-
plexes. RNA was extracted from the precipitated com-
plexes, and siRNA was detected by streptavidin-alkaline
phosphatase chemiluminescence (see Figure S1 in the
Supplemental Data available with this article online). Anti-
sense strand 30-LC biotin-conjugated siRNA was immu-
noprecipitated only from cells transfected with HA-RHA
and not with HA alone. Furthermore, in extracts of cells
transfected with HA-RHA and with nonbiotinylated siRNA,
no streptavidin binding RNA species were detected.
These results are consistent with our other findings
(Figure 2A) that RHA is a RISC-associated protein.
To determine whether or not RHA interacts with the
RISC components Dicer, TRBP, and Ago2 in the absence
ofbiotinylated siRNA,wetransfected293T cellswithFlag-
Dicer, Flag-TRBP, and Flag-Ago2. As a specificity control,
293T cells were transfected with Flag-Hexim1. Hexim1 is
an RNA binding protein that interacts with a double-
stranded region of 7SK small nuclear RNA (snRNA) and
with cyclin T1 in the inactive P-TEFb ribonucleoprotein
complex (reviewed in Zhou and Yik ). From these
transfected cells, we prepared total cell extracts (TCE)
and immunoprecipitated Flag-tagged RISC components.
Immunopurified complexes were assayed for the pres-
ence of RHA using rabbit antibodies against endogenous
RHA (Figure 2B). Immunopurified complexes from cells
transfected with Flag-Hexim1 did not contain detectable
RHA, whereas RHA was coimmunoprecipitated with
Flag-Dicer, -TRBP, and -Ago2. To verify that the observed
interactions of RHA with Dicer and Ago2 were not due to
overexpression of RHA, Dicer, or Ago2, we immunopre-
cipitated endogenous RHA from nontransfected cells.
524 Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc.
RNA Helicase A in RISC
Figure 1. RNA Helicase A Is a RISC-Associated Protein
(A) Overview of the strategy used to purify biotinylated siGFP-RISC. RISC activity bound to streptavidin beads was assayed in vitro using a 124 nt
32P-cap-labeled synthetic RNA target. Site-specific endonucleolytic cleavage produced a 55 nt32P-cap-labeled product. Lane 1 contains strepta-
from HeLa cells transfected with GFP siRNA modified at the 30end of the antisense (guide) strand with LC-biotin.
(B) RHA and HSP90b are associated with RISC. Streptavidin beads (B) were extensively washed and treated with RNase to release RISC-associated
proteins. Supernatants (S) were resolved on 8% denaturing polyacrylamide gel and silver stained. Markers indicate molecular mass positions.
Proteins were excised from gels, eluted, and identified by mass spectrometry (MALDI QIT MS/MS).
Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc. 525
RNA Helicase A in RISC
Figure 2. RNA Helicase A Is a Component of Human RISC
(A) RHA, along with Ago2 and Dicer, is a component of human RISC. 293T cells were transfected with biotinylated siRNA, HA-tagged, and myc-
tagged constructs, as indicated, and siRNA-programmed RISC were purified by streptavidin pullout. Isolated proteins were resolved on 8% dena-
turing SDS-polyacrylamide gels, electroblotted, and analyzed for HA-RNA helicase A, myc-Ago2, and Dicer, as indicated.
(B) RHA interacts with RISC components Ago2, Dicer, and TRBP. 293T cells were transfected with Flag-tagged protein expression constructs as
indicated, and cell extracts were prepared. TCE (10% of input) and anti-Flag immunoprecipitates were resolved on denaturing SDS-polyacrylamide
gels and blotted, and endogenous RHA was detected. Flag-tagged Hexim1, Dicer, TRBP, and Ago2 were resolved on denaturing 4%–20%
SDS-polyacrylamide gels and detected by using anti-Flag antibody.
526 Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc.
RNA Helicase A in RISC
Examination of these immunoprecipitated complexes
revealed that both Ago2 and Dicer were coimmunopreci-
pitated with endogenous RHA (Figure 2C). In similar
experiments, myc-Ago1 coimmunoprecipitated with en-
dogenous Dicer and RHA (Figure S2). We are unable to
determine whether the interaction of Ago1 with RHA is
direct or is a consequence of the interaction of Ago1
with Ago2. In addition, FPLC fractionation analysis of
RISC revealed that Ago2, Dicer, RHA, and the guide-
strand RNA coeluted in a high molecular mass complex
(Figure S3). Together, these results show that RHA specif-
ically interacts with the RISC components Ago2, Ago1,
Dicer, and TRBP and provide further evidence that RHA
is a component of RISC.
We next asked whether the interaction between RHA
and RISC components depended on helicase activity or
the N-terminal double-stranded binding domains. To this
end, we cotransfected 293T cells with plasmids encoding
wild-type and mutant RHA tagged with the HA epitope,
Flag-tagged TRBP, and myc-tagged Ago2. TCE were pre-
pared and RHA was immunoprecipitated using anti-HA
agarose beads. Precipitated complexes were assayed
for the presence of endogenous Dicer (Figure 2D), Flag-
In extracts of cells transfected with empty vector (HA),
immunopurified complexes did not contain Dicer, myc-
Ago2, or Flag-TRBP. As expected, however, in extracts
of cells transfected with wild-type HA-tagged RHA (HA
RHA WT), immunoprecipitates contained endogenous
Dicer, myc-Ago2, and Flag-TRBP.
Mutating RHA residue 417 from lysine to arginine
disrupts the ATP binding domain and abolishes RHA heli-
case activity (Nakajima et al., 1997). Complexes contain-
ing this K417R mutant RHA coimmunoprecipitated with
lower amounts of Dicer, Flag-TRBP, and myc-Ago2 than
did complexes with wild-type RHA (Figures 2D–2F, lanes
tant but not required for RHA to interact with RISC. Re-
moving the N-terminal 272 amino acids of RHA disrupts
its interaction with CBP/p300 (Nakajima et al., 1997) and
removes two dsRNA binding domains. This mutant RHA
did not coimmunoprecipitate with Dicer, myc-Ago2, and
Flag-TRBP (Figures 2D–2F, lane 4), demonstrating the
requirement of these two dsRNA binding domains for
interaction with RISC. Finally, the importance of the RHA
transactivation domain in RHA interactions with RISC
components was tested by transfecting cells with HA-
tagged RHA whose tryptophan 339 was mutated to ala-
with RNAPII nor stimulates CREB-dependent transcrip-
tion (Aratani et al., 2001). Immunoprecipitated HA RHA
W339A complexes contained endogenous Dicer, Flag-
TRBP, and myc-Ago2 (Figures 2D–2F, lane 5), arguing
against a role for the transactivation domain in RHA inter-
actions with RISC. Together, these results highlight the
importance of the dsRNA binding domains of RHA in
mediating its interaction with RISC.
RHA Functions in the Gene-Silencing Pathway
gene silencing, we depleted RHA in HeLa cells and trans-
fected them with either siRNA or shRNA targeting a re-
porter RNA. mRNA levels were assayed 18 hr later by
quantitative real-time RT-PCR (qRT-PCR), or protein ex-
tracts were assayed for specific in vitro mRNA cleavage
by RISC (Figure 3A). siRNA-mediated depletion of endog-
enous RHA was verified by quantifying RHA mRNA 48 hr
tion with RHA siRNA reduced RHA mRNA levels to 0.18 ±
0.02 of control duplex-transfected cells. Similarly, RHA
protein was shown by immunoblot analysis to be potently
with RHA-targeting siRNA, RHA was effectively depleted
To ensure that RHA depletion did not affect the effi-
ciency of subsequent transfections, cells were first
depleted of RHA or Ago2 and transfected the next day
with siRNA containing Cy3 at the 30end of the AS strand.
cal nuclease to degrade any RNA not internalized by the
cells. The cells were then washed extensively and lysed.
Equal amounts of protein from TCE were assayed for
Cy3 fluorescence. Under these conditions, whether RHA
or Ago2 was first depleted did not significantly affect
Cy3 fluorescence. Relative to control, the Cy3 fluores-
cence signals in the RHA- and Ago2-depleted cells were
0.97 ± 0.03 and 1.05 ± 0.05, respectively (Figure S5).
These results indicate that depleting RHA or Ago2 did
notaffect theefficiency ofsubsequentsiRNA transfection.
To examine the effect of RHA depletion on gene silenc-
ing in human cells, we first transfected HeLa cells stably
expressing GFP with siRNA targeting RHA, transfected
them 24 hr later with siRNA or shRNA targeting GFP,
and quantified levels of GFPmRNA. The target sequences
of siRNA used to knock down GFP and CDK9 have been
described (Chiu and Rana, 2002, 2003). In control cells
(C) Endogenous RHA interacts with endogenous Ago2 and Dicer. Cell extracts were prepared from nontransfected 293T cells and pretreated with
normal rabbit serum before immunoprecipitation with anti-RHA serum. TCE (10% of input) and precipitated complexes were resolved as indicated
by SDS-PAGE. Dicer and Ago2 were detected using rabbit polyclonal antibodies as indicated.
(D) RHA interaction with Dicer depends on RHA amino acids 1–272. 293T cells were transfected with wild-type or mutant HA-tagged RHA as indi-
cated. Following anti-HA immunoprecipitation, TCE (10% of input) and precipitates were resolved by 6% SDS-PAGE. Endogenous Dicer was
detected using anti-Dicer antibodies.
(E) RHA interaction with TRBP depends on amino acids 1–272. 293T cells were cotransfected with wild-type or mutant HA-tagged RHA and Flag-
tagged TRBP and processed as in (D). Precipitated complexes were analyzed for TRBP using anti-Flag antibodies.
(F) RHA interaction with Ago2 depends on amino acids 1–272. 293T cells were cotransfected as indicated with HA-RHA and myc-tagged Ago2.
Anti-HA immunoprecipitates were analyzed for myc-Ago2 by western blot using anti-myc antibodies.
Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc. 527
RNA Helicase A in RISC
Figure 3. RHA Depletion Inhibits siRNA- and shRNA-Mediated Gene Silencing In Vivo
(A) Overview of experiments analyzing RHA involvement in the RNAi pathway.
(B) Depletion of RHA reduces mRNA knockdown in response to exogenous siRNA or shRNA. GFP mRNA was quantified by qRT-PCR of randomly
primed cDNA from HeLa cells after consecutive transfections with RHA siRNA and with GFP siRNA or shRNA. GFP mRNA levels are expressed in
arbitrary units ± SEM relative to that in control siRNA-transfected cells, which was set to 1. For each sample, GFP mRNA levels were normalized
to cyclophilin A mRNA levels in triplicate experiments. Comparisons represent the mean ± SD from three independent experiments. RNA duplexes
are shown as the following: C, control; si, siRNA; sh, shRNA.
528 Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc.
RNA Helicase A in RISC
that were first transfected with control mismatch se-
quence, subsequent siRNA and shRNA transfection re-
duced GFP mRNA levels to 0.15 ± 0.04 and 0.08 ± 0.02
of control levels, respectively. In RHA-depleted cells, sub-
sequent siRNA and shRNA transfection reduced GFP
mRNA levels to 0.31 ± 0.08 and 0.20 ± 0.06 of control
levels, respectively, for siRNA targeting RHA mRNA nt
544 to 564, and 0.38 ± 0.01 and 0.37 ± 0.07 of control
levels for siRNA targeting RHA mRNA nt 2408–2428
(Figure 3B). As a positive control for inhibition of gene
silencing by previous RHA knockdown, cells were treated
with siRNA targeting Ago2. Knockdown of RHA before
programming cells with GFP siRNA or shRNA dere-
pressed GFP mRNA accumulation 2.42-fold ± 0.56 and
2.82-fold ± 0.96, respectively, over that of control du-
plex-transfected cells for siRNA targeting RHA mRNA nt
544–564 (Figure 3C). When RHA was depleted by siRNA
targeting RHA mRNA nt 2408–2428, GFP mRNA accumu-
lation was induced 3.02-fold ± 1.3 and 3.14-fold ± 0.36
with siRNA or shRNA, respectively (Figure 3C).
We next asked if gene silencing could also be reduced
for an endogenous mRNA target. To answer this question,
duplexes and then with siRNA or shRNA targeting CDK9
mRNA (Figure 3A). We found that in cells first treated
with control siRNA duplexes, subsequent treatment with
CDK9-targeted siRNA or shRNAs reduced CDK9 mRNA
levels to 0.25 ± 0.02 and 0.17 ± 0.02 of control levels, re-
spectively (Figure 3D). However, levels of CDK9 mRNA
were not reduced to the same extent in cells that were first
siRNA or shRNAs. In cells first depleted of RHA by an
siRNA targeting RHA mRNA nt 544–564, subsequent pro-
gramming with CDK9 siRNA and shRNA reduced CDK9
mRNA to 0.48 ± 0.05 and 0.36 ± 0.2, respectively, of that
in cells first depleted of RHA but subsequently transfected
levels corresponded to a 2.06-fold ± 0.27 and 2.22-fold ±
0.42derepression of CDK9 mRNA accumulation triggered
Similar results were observed in cells depleted of RHA by
siRNA targeting RHA mRNA nt 2408–2428. CDK9 mRNA
levels were reduced to 0.38 ± 0.01 and 0.36 ± 0.07 of con-
trol levels in response to siRNA or shRNA, which corre-
sponded to 1.63-fold ± 0.02 and 1.8-fold ± 0.12 of control
for siRNA and shRNA, respectively (Figure 3E). Taken
together, these results indicate that depletion of RHA
reduced RNAi-induced gene silencing ?2-fold. Thus,
RHA plays a functional role in RNAi-mediated reduction
in levels of target mRNA.
We reasoned that the reduced efficiency of RNA silenc-
ing in RHA-depleted cells could result from reduced
cellular RISC activity or from reduced degradation of tran-
scripts after RISC-mediated endonucleolytic cleavage. To
test the hypothesis that RHA depletion reduced cellular
RISC activity, we used in vitro target mRNA cleavage as-
of cell extracts from RHA-depleted cells. Cells were trans-
fected with a control siRNA duplex or withsiRNA targeting
RHA mRNA. Twenty-four hours later, cells were trans-
fected with siRNA or shRNA targeting GFP or CDK9
mRNA. After a further 24 hr, cell extracts were prepared,
and protein content was determined. Equivalent amounts
of extracts were incubated in vitro with32P-cap-labeled
RNAs containing target sequences perfectly complemen-
tary to the guide strand of siRNAs or shRNAs for GFP or
CDK9. We found that, in RHA-depleted cell extracts,
cleavage of GFP (Figure 4A) and CDK9 (Figure 4B) target
mRNA was decreased. Relative to extracts from cells
with intact RHA, the cleavage efficiency in extracts from
cells depleted of RHA and subsequently programmed
with GFP siRNA or shRNA was reduced to 48.5% ±
9.6%and 38.1%±11.2%, respectively(Figure 4A).Inpar-
allel experiments, where extracts from control-treated
cells were compared with those from cells depleted of
RHA and subsequently programmed with CDK9 siRNA or
shRNA, target mRNA cleavage was reduced to 34.5% ±
11.8% and 27.5% ± 14.9%, respectively (Figure 4B).
These results support the hypothesis that depletion of
RHA decreases cellular RISC activity but are inconsistent
with the hypothesis that RHA depletion reduces RNA
degradation after RISC-mediated endonucleolytic target
RHA Is Involved in Active RISC Formation
Since RHA plays a functional role in RISC-mediated gene
silencing, we next asked if RHA functions in RISC-
mediated mRNA cleavage. To examine this possibility,
we prepared cell extracts from control and RHA-depleted
cells programmed with siRNA targeting GFP and with
siRNA or shRNA targeting CDK9. Cell extracts were incu-
bated with increasing amounts of 20-O-methyl oligonucle-
otides with sequences complementary to theguide strand
incorporated into RISC. The effective RISC concentra-
tions in these cell extracts were calculated by quantifying
their cleavage products, thus allowing comparison of the
cleavage efficiency of equimolar amounts of target-
specific RISC programmed in RHA-depleted cells or con-
trolcells withintactRHA.The amountof RISC produced in
RHA-depleted cells programmed with GFP siRNA was
(C) Derepression of GFP mRNA in RHA- and Ago2-depleted HeLa cells. Fold increase in mRNA is expressed relative to that in control duplex-trans-
fected cells (set to 1). Comparisons represent the mean ± SD from three independent experiments.
(D) Depletion of RHA reduces mRNA knockdown in response to exogenous siRNA or shRNA. CDK9 mRNA was quantified by qRT-PCR of randomly
primed cDNA from HeLa cells transfected on consecutive days with the indicated siRNAs or shRNAs. CDK9 mRNA levels are expressed in arbitrary
units ± SEM relative to that in control siRNA-transfected cells (set to 1). For each sample, triplicate determinations of CDK9 mRNA levels were nor-
malized to cyclophilin A mRNA levels. Comparisons represent the mean ± SD from three independent experiments.
(E) Derepression of CDK9 mRNA in RHA- and Ago2-depleted HeLa cells. Fold increase in CDK9 mRNA is expressed relative to that in control duplex-
transfected cells (set to 1). Comparisons represent the mean ± SD from three independent experiments.
Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc. 529
RNA Helicase A in RISC
0.66 ± 0.06 times that produced in control duplex-
with CDK9 siRNA was 0.68 ± 0.14 times that produced in
mock-treated control cells (Figure 5B). Furthermore, the
relative concentration of RISC formed in RHA-depleted
cells programmed with CDK9 shRNA was 0.65 ± 0.16
indicate that RHA depletion reduced RISC formation in re-
sponse to siRNA and shRNA triggers.
RHA Depletion Reduces siRNA Association
To understandtheroleof RHAinRISC formation, westud-
ied the effect of depleting RHA on the incorporation of
siRNA into RISC containing Ago2. As outlined in
Figure 6A, HeLa cells were first transfected with siRNA
to deplete RHA; 24 hr later, they were transfected with
GFP siRNA duplexes whose antisense strand was 30
labeled with LC biotin, and 24 hr later their cell extracts
were immunoprecipitated for Ago2-containing com-
plexes. RNA from the Ago2 immunoprecipitates was
(Figure 6B). We found that depletion of RHA significantly
reduced the association of guide-strand siRNA (eGFP
AS) with Ago2. Identical results were obtained when the
biotinylated guide-strand siRNA was analyzed using
streptavidin-conjugated alkaline phosphatase and chem-
iluminescence detection. No biotinylated RNA species
were detectable in Ago2 immunoprecipitates from cells
Figure 4. RHA Depletion Reduces RISC-Mediated mRNA Cleavage In Vitro
(A) Depletion of RHA reduces cleavage of GFP target mRNA triggered by GFP siRNA. Following knockdown of RHA, HeLacells were transfected with
siRNAor shRNA targeting GFP mRNA (si and sh,respectively) or ansiRNAduplex containing amismatch (mm) at theGFP mRNA cleavagesite.Equal
amounts of cytoplasmic extracts were incubated with a 124 nt32P-cap-labeled synthetic RNA target. Site-specific endonucleolytic cleavage pro-
duces a 55 nt cap-labeled product. Cleavage efficiency was determined by densitometric quantification of full-length target and cleavage product
for eachcondition. Relative cleavage efficiency is expressed as apercentof the maximal activity measured incontrolduplex-transfected cell extracts
(set to 100). Comparisons represent the mean ± SEM from three independent experiments.
(B) Depletion of RHA reduces cleavage of CDK9 target mRNA triggered by CDK9 siRNA. Following knockdown of RHA, cells were transfected with
siRNA or shRNA targeting CDK9 mRNA (si and sh, respectively) or an siRNA duplex containing a mismatch (mm) at the CDK9 mRNA cleavage site.
Equal amounts of cytoplasmic extracts were incubated with a 150 nt32P-cap-labeled synthetic mRNA target. Site-specific endonucleolytic cleavage
three independent experiments.
530 Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc.
RNA Helicase A in RISC
transfected with GFP siRNA duplexes containing un-
labeled guide strand (data not shown). When signal in-
tensities of northern blots were quantified, we observed
a 56% ± 11% decrease in siRNA loading. A minor amount
of passenger-strand siRNA (eGFP SS), which could not
be explained by incomplete stripping, was detectable on
the same membranes by northern blot using32P-labeled
guide-strand probes. This signal most likely reflected
passenger-strand incorporation into RISC, consistent
with the currently understood influence of siRNA duplex
internal stability parameters on RISC loading (Khvorova
et al., 2003; Reynolds et al., 2004; Tomari et al., 2004b).
The amount of passenger-strand siRNA associated with
Ago2 was also reduced under conditions of RHA deple-
tion. As a control, the amount of GFP siRNA associated
with Ago2 was examined in extracts from cells previously
transfected with siRNA targeting Dicer. Under conditions
of Dicer depletion, GFP siRNA associated with Ago2 de-
suggesting that Dicer plays a role in siRNA recruitment to
Ago2 (Chendrimada et al., 2005; Lee et al., 2006). How-
ever, we found that siRNA association with Ago2 was
affected more by depleting RHA than by depleting Dicer.
Furthermore, depleting RHA or Dicer did not affect total
cellular levels of Ago2 or siRNA, and depleting Dicer did
not change the amount of endogenous RHA that coimmu-
noprecipitated with Ago2. Depletion of RHA similarly had
no effect on TRBP mRNA levels (data not shown). To nor-
malize proteinand RNAin ourimmunoprecipitation exper-
iments, we used western blots for precipitated Ago2 and
northern blots for precipitated let-7a since mature
miRNAs appear to be highly stable (Haase et al., 2005;
Lee et al., 2003, 2006; Lund et al., 2004). Consistent with
previous reports (Haase et al., 2005; Lee et al., 2003,
2006; Lund et al., 2004), our results did not show gross
changes in the amount of let-7a miRNA associated with
Next, we performed similar experiments to examine the
role of RHA in assembly of RISC programmed with Dicer
substrate shRNA (Figure 6C). GFP guide-strand siRNA
(eGFP AS) was significantly reduced in Ago2 immunopre-
cipitates from RHA-depleted cells, as was the minor
amount of loaded passenger strand (eGFP SS). Quantifi-
cation of the guide-strand RNA associated with Ago2
under conditions of RHA depletion revealed a 67% ±
9.6% reduction as compared to control immunoprecipi-
tates. These results are consistent with our observations
for the effect of RHA on siRNA loading. Depletion of Dicer
reduced the amount of processed GFP siRNA associated
with Ago2 to 64% ± 6.3%, as expected, since shRNA
requires Dicer processing to be detected at the appropri-
ate size after denaturing gels in this assay. Interestingly,
the amount of GFP siRNA associated with Ago2 in
Dicer-depleted cells was slightly greater than that ob-
served in RHA-depleted cells. We attribute this observa-
tion to the loading of processed shRNA resulting from re-
sidual cellular Dicer activity. Taken together, these data
show that RHA depletion decreased the amount of guide
strand incorporated into RISC and support a role for
RHA in recruiting siRNA into RISC.
In this study, we used a purification scheme that isolates
tify RHA as a component of human RISC. We found that
RHA functions in the RNAi pathway by promoting the for-
mation of active RISC. Current models of human RISC as-
proteins that bind double-stranded RNA, namely Dicer,
TRBP, and PACT (Chendrimada et al., 2005; Gregory
et al., 2005; Haase et al., 2005; Lee et al., 2006). TRBP
and PACT are structurally related proteins that contain
three double-stranded RNA binding domains. Interest-
ingly, both proteins have been identified as regulators of
PKR function (Gupta et al., 2003; Patel and Sen, 1998; Pe-
ters et al., 2001), and both interact with Dicer to promote
the formation of active RISC (Chendrimada et al., 2005;
Gregory et al., 2005; Haase et al., 2005; Lee et al., 2006).
Depletion of RHA decreased mRNA knockdown and
formation of RISC in human cells programmed with 21 nt
dsRNA or with shRNAs. The amount of guide-strand
siRNA associated with Ago2 decreased markedly in cells
depleted of RHA. Our data support a model (Figure 7) in
which RHA plays a role consistent with current models
of human RISC formation and functions in siRNA loading
onto Ago2. Pre-miRNA stem-loop structures are pro-
cessed by Dicer and other interacting proteins into
double-stranded siRNAs. Alternatively, exogenous siRNA
enters the pathway at this point, where it is recognized by
the TRBP/PACT/Ago2 and Dicer complex, although its
loading into RISC is not as efficient as that of hairpin
pre-miRNAs or 27-mer Dicer substrate RNA duplexes
(Elbashir et al., 2001; Gregory et al., 2005; Kim et al.,
2005; Siolas et al., 2005). Ago2 by itself cannot form
a complex with dsRNA, and its recruitment to dsRNA in
purified in vitro systems depends on TRBP and Dicer
(Chendrimada et al., 2005). The data presented here dem-
onstrate that RHA facilitates the formation of active RISC
in human cells by promoting the interaction of siRNA
and processed Dicer substrate shRNA with Ago2, most
likely by virtue of RHA’s ability to bind siRNA and its inter-
action with Ago2, TRBP, and Dicer.
catalyzes its cleavage (Rana, 2007). RISC must then
dissociate from the cleaved target RNA before it can be
recycled to mediate multiple rounds of cleavage in an
ATP-independent manner (Gregory et al., 2005). Our
results indicate that RHA is not involved in the recycling
process, since RISC formed in the absence of RHA did
not exhibit poorer cleavage activity under multiple-
turnover conditions than under single-turnover conditions
(data not shown).
Mature RISC formation requires the removal of passen-
but the details of this RISC activation are currently
Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc. 531
RNA Helicase A in RISC
Figure 5. Depletion of RHA Reduces Formation of Active RISC
(A) Active RISC decreases in RHA-depleted cells programmed with GFP siRNA. Concentration of active GFP RISC was determined by 20-O-methyl
inhibition ofinvitromRNA cleavage reactions. Increasingamountsof 20-O-methyl inhibitor wereincubated withcytoplasmic extracts from cells trans-
fected with control, RHA, or Ago2 siRNAs and subsequently transfected with siRNA targeting GFP. Representative gels are shown. Percent mRNA
cleavage was monitored as in Figure 4, and IC50 values were calculated for each condition by fitting the data to sigmoidal curves. The concentration
of active RISC in extracts from RHA-depleted cells was compared to that in extracts from cells transfected with control duplexes. Arrows indicate
the migration positions of input RNA and specific cleavage products (Product). Comparisons represent the mean ± SD from three independent
532 Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc.
RNA Helicase A in RISC
unknown in human cells. In Drosophila cells, cleavage of
the passenger strand of the siRNA duplex facilitates the
formation of RISC programmed with siRNA (Matranga
et al., 2005; Rand et al., 2005). However, miRNA-pro-
grammed RISC in Drosophila forgoes this requirement
et al., 2005).
The precise stage of human RISC formation at which
passenger-strand removal or unwinding is required and
which proteins are responsible for this removal remain
open questions. RHA is the bona fide human RISC-
associated helicase protein to be reported. Depletion of
the putative helicase MOV10, which was identified by
immunoaffinity purification of Ago2 (Meister et al., 2005),
blunted silencing of a reporter gene bearing a perfect
match to endogenous miR-21, indicating that MOV10
also functions in the RNAi pathway (Meister et al., 2005).
Whether or not MOV10 has helicase activity and can
unwind siRNA-like structures is currently unknown, and
no data have yet provided insight into the defect in RNA
silencing when MOV10 is depleted in human cells. The
Drosophila homolog of MOV10 is armitage (armi), and
armi mutant ovary lysates failed to support the formation
of active RISC (Tomari et al., 2004a). Our studies show
(C) Active RISC decreases in RHA-depleted cells programmed with CDK9 shRNA. Concentration of active RISC formed after transfection with CDK9
shRNA was determined by 20-O-methyl inhibition of in vitro cleavage reactions as in (A). Representative gels are shown. Comparisons represent the
mean ± SD from three independent experiments.
Figure 6. RHA Depletion Reduces the
Association of siRNA with Ago2
(A) Outline of experiments analyzing the role of
RHA in siRNA incorporation into RISC.
(B) Depleting RHA reduces the association of
siRNA with Ago2. HeLa cells were transfected
with RHA siRNA or a control duplex and trans-
fected 24 hr later with a GFP siRNA duplex. En-
dogenous Ago2 was immunoprecipitated 24 hr
later from cell extracts. RNA was detected by
northern blot hybridization in TCE and in Ago2
immunoprecipitates. Proteins were detected
in TCE (10% of input) and in Ago2 precipitates
by western blot analysis with the indicated
antibodies. For siRNA levels in TCE, RNA
from 0.5% of immunoprecipitation input was
(C) Depleting RHA reduces the association of
Dicer substrate siRNA with Ago2. HeLa cells
were transfected with a control duplex siRNA
or siRNAs targeting RHA or Dicer and trans-
fected 24 hr later with shRNA. TCE were immu-
and both TCE and Ago2 immunoprecipitates
were analyzed for RNA by northern blot hybrid-
ization. Proteins were detected by western blot
with the indicated antibodies.
Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc. 533
RNA Helicase A in RISC
of guide-strand siRNA associated with Ago2, suggesting
that RHA functions in the human RNAi pathway by pro-
moting the association of guide-strand siRNA with RISC.
Is RHA the helicase that unwinds siRNA duplexes to
promote RISC activation? The data presented here dem-
onstratethatRHAbinds tosiRNA,TRBP,Dicer, andAgo2,
In addition, we demonstrate that, in RHA-depleted cells,
the concentration of specific activeRISC isreduced, while
overall levels of Ago2 and transfected siRNA remain con-
stant. In vitro characterization of purified RHA indicated
that this enzyme can unwind RNA duplexes, but it ineffi-
ciently unwound siRNA-like structures with short 30over-
hangs (Lee and Hurwitz, 1992). It should be noted, how-
ever, that protein-protein interactions influence the
helicase specificity and activity of RHA (Friedemann
et al., 2005). In our experiments, depleting RHA in cells re-
duced the amount of both guide- and passenger-strand
siRNA associated with Ago2. If RHA functioned as a heli-
case that unwinds siRNA duplexes, we would have ex-
pected that guide-strand siRNA associated with Ago2
would remain unchanged while passenger-strand siRNA
associated with Ago2 would increase in RHA-depleted
extracts. Further studies will differentiate whether this is
solely a result of recruitment or a consequence of re-
cruited siRNA not being unwound, since the fate of siRNA
that remains unwound or unloaded in a cell is currently
An alternative paradigm for the DExD/H family of pro-
teins is the remodeling or ordered assembly of RNP com-
plexes (for recent review, see Linder ). For example,
the DEAD box protein Prp28p is required for the first step
of pre-mRNA splicing and promotes the exchange of U1
and U6 snRNAs at the 50splice site. The mechanism for
this function was later found to be that Prp28p counter-
acts the stabilizing effect of the U1-C protein on the U1
and 50splice site RNA-RNA interaction, thus promoting
remodeling of the spliceosome, which allows the splicing
reaction to proceed (Chen et al., 2001). In an additional
example, the ATPase function of the DEAD box protein
Mss116 is required for intron splicing, while its helicase
activity is dispensable, indicating that Mss116 does not
function in splicing by unwinding RNA structures.
Mss116 has been proposed to function by promoting
the ordered assembly of the RNP complex responsible
Figure 7. Model for the Role of RHA in
Human RISC Formation and Function
Dicer substrate RNAs are recognized by TRBP
in complex with Dicer and other proteins. Dicer
mediates the cleavage of substrate RNAs into
short double-stranded RNAs. Alternatively,
siRNA can be introduced into cells. RISC con-
strand selection result in the formation of ma-
motes active RISC assembly by promoting the
association of siRNA with Ago2. Activated
RISC associates with cognate target RNA and
catalyzes its cleavage. RISC is then recycled
to mediate multiple rounds of target cleavage.
For clarity, not all RISC-associated factors are
534 Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc.
RNA Helicase A in RISC
(Solem et al., 2006). Thus, RHA might function in RISC
loading by inducing the remodeling of the human RISC
loading complex that facilitates loading of siRNA onto
Ago2; alternatively, RHA might promote the ordered as-
sembly of the RISC during siRNA loading.
The enzymatic activity of RHA is supported by all eight
common nucleoside triphosphates (Lee and Hurwitz,
1992). Therefore, it is of interest that, although human
RISC activity is independent of ATP, ATP and other nucle-
otides were recently shown to stimulate RISC activity
(Gregory et al., 2005).
In our RHA depletion experiments, we observed signif-
icant deficits in RNA silencing and RISC formation. How-
ever, neither process was completely disrupted. One
possible explanation for these observations is that RHA
was not completely depleted from cells and that residual
RHA was sufficient for the residual (?50%) silencing
efficiency and RISC formation. Alternatively, other RISC-
interacting factors such as MOV10 may have overlapping,
partially shared, and/or redundant functions with RHA in
RISC formation and RNA silencing. In this regard, TRBP
and PACT appear to have overlapping roles in human
RISC loading (Lee et al., 2006). Regardless of the incom-
plete disruption of RNA silencing and RISC formation,
when taken together, our observations implicate RHA as
a RISC-loading factor and potential unwinding factor for
RISC activation. Future in vitro reconstitution experiments
are required to confirm this specific function of RHA.
Cell Culture, Transfections, and Plasmids
HeLa and HEK 293T cells were obtained from ATCC (Manassas, VA)
10% FCS. HeLa cells were transfected with lipofectamine (Invitrogen,
Carlsbad, CA), and HEK 293T cells were transfected with Lipofect-
amine 2000 (Invitrogen), as recommended by the manufacturer.
W339A mutant RHA and expression plasmids for HA-tagged RHA
were generously provided by Toshihiro Nakajima (Aratani et al.,
2001). The RHA K417R mutant was provided by Marc Montminy
(Nakajima et al., 1997) Plasmids encoding myc-tagged Ago1, Ago2,
and mutants of Ago2 were provided by Gregory Hannon (Liu et al.,
(Meister et al., 2004). Plasmids encoding Flag-Dicer and Flag-TRBP
were provided by Ramin Shiekhattar (Chendrimada et al., 2005).To
verify transfection efficiency in sequential transfection experiments,
cells were transfected a second time with siRNA duplexes Cy3 labeled
on the 30end of the antisense strand. Micrococcal nuclease (USB,
Cleveland, OH) was added 18 hr later to the culture medium (5 mg/ml
final concentration), and cells were incubated for 30 min. Cells were
trypsinized, washed five times with PBS, lysed in 1 3 passive lysis
buffer (Stratagene, La Jolla, CA), and the protein concentration of
cell extracts determined by the Dc Protein assay (Bio-Rad, Hercules,
CA). Equal amounts of protein were analyzed for Cy3 fluorescence
using a Safire2 monochromator-based microplate detection system
(Tecan, Zurich, Switzerland). For details of oligonucleotide sequences,
see the Supplemental Data.
Antibodies, Immunoprecipitation, Immunoblots,
and Northern Blots
Mouse anti-myc, mouse anti-tubulin, and mouse anti-HA antibodies
and goat anti-myc and anti-HA conjugated to agarose beads were
from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse anti-RHA
antibodies were from Abnova (Taiwan, R.O.C.), and rabbit anti-RHA
antibodies were from Vaxron (Rockaway, NJ). Rabbit anti-Ago2 anti-
bodies were described previously (Chu and Rana, 2006). Anti Flag
M2 antibody and antibody-conjugated agarose beads were from
Sigma (St. Louis, MO). Immunoprecipitations were carried out using
500–1000 mg of protein and 1–2 mg of antibody in 1 3 lysis buffer for
16 hr at 4?C. Immunoprecipitates were washed four times for 10 min
in 1 ml of wash buffer (10 mM Tris [pH 8], 150 mM NaCl, 1% v/v Triton
X-100) and resuspended in SDS protein loading buffer. Proteins were
resolved on SDS-polyacrylamide gels under denaturing conditions,
electroblotted onto PVDF membranes (Bio-Rad), and probed with an-
tibodies as indicated. For IP RNA analysis, immunoprecipitated RNA
was extracted from the beads or from 50 mg of cell extract using Trizol
and resolved on 20% polyacrylamide-urea gels (National Diagnostics,
Atlanta, GA). RNA was electoblotted onto HyBond N+ membrane (GE
Healthcare, Piscataway, NJ). Biotinylated siRNA was detected using
detection were collected below saturation using a LAS-3000 lumines-
cent image analyzer (Fujifilm Medical Systems USA, Inc.). For northern
blot hybridizations, membranes were blocked for 1 hr and hybridized
overnight in Church buffer (0.5 M Na2HPO4, 0.1 M EDTA, 7% SDS).
Probes for sense and antisense GFP siRNA were 20-O-methyl oligonu-
cleotides that were hybridized and washed three times in 2 3 SSC at
65?C. The probe for human let-7a was a deoxyoligonucleotide that
was hybridized and washed at 37?C. Between hybridizations, mem-
branes were stripped three times using boiling 0.1% SDS, and com-
pletion of probe removal was monitored by analyzing blot lanes
containing sense or antisense siRNA. Membranes were exposed to
imaging screens overnight and scanned using a FLA-5000 fluoroim-
ager (Fuji). Image analysis was performed using Image Gauge V4.22
Identification of RISC-Interacting Proteins
Biotin-containing siRNA duplexes were transfected into HeLa cells
using lipofectamine (Invitrogen), and cytoplasmic extracts were
prepared as described earlier (Robb et al., 2005). AS 30-LC biotin-
associated proteins were isolated by incubating 75 ml (approximately
375 mg) of extract with 100 ml of streptavidin-conjugated magnetic
beads (Dynabeads M-280, Dynal) in a final volumeof 500 ml of 1 3 lysis
buffer overnight at 4?C with rotation. Beads were washed two times
bead pellet was resuspended in 50 ml of lysis buffer without sucrose.
An aliquot (10 ml) was assayed for RISC activity specific for the LC-
biotin-conjugated siRNA. The remaining suspension was mixed with
1 ml RNase 1 (Ambion, Austin, TX) and incubated at 37?C for 30 min
to elute protein complexes from the beads. Supernatants from the
RNase reaction were mixed with 4 3 SDS PAGE loading buffer. Beads
from the RNase reaction were resuspended in 50 ml of lysis buffer lack-
ing sucrose and mixed with 4 3 SDS PAGE loading buffer. Samples
were heated to 100?C for 3 min. Magnetic beads were removed, and
proteins were resolved on 8% denaturing polyacrylamide gels. Gels
were silver stained using the Silver Stain Plus kit (Bio-Rad). Bands
specific to the LC-biotin eluate condition were excised and protein
sequence was determined by MALDI-QIT-TOF MS and MS/MS at
the UMass Proteomic Mass Spectrometry facility.
In Vitro Target mRNA Cleavage
In vitro target mRNA cleavage reactions were performed as previously
described (Brown et al., 2005; Robb et al., 2005) except that protein
content or amount of RISC was equalized between conditions. To
quantify specific RISC concentration in cell extracts, in vitro cleavage
reactions were performed in the presence of 20-O-methyl oligonucleo-
tide inhibitors with sequences corresponding to the target RNAs, as
described (Brown et al., 2005).
Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc. 535
RNA Helicase A in RISC
Supplemental Data include supplemental text and five figures and can
be found with this article online at http://www.molecule.org/cgi/
We thank members of the Rana laboratory for helpful discussions and
support and Chia-ying Chu for her assistance in illustrations. We thank
Gregory Hannon, Marc Montminy, Toshihiro Nakajima, Ramin Shie-
khattar, and Thomas Tuschl for kindly providing reagents and Dr.
John Leszyk at the University of Massachusetts Medical School Pro-
teomic Mass Spectrometry Facility for identification of proteins.
G.B.R. is the recipient of a postdoctoral fellowship award from the On-
tario HIV Treatment Network. This work was supported by grants from
the National Institutes of Health (T.M.R).
Received: September 8, 2006
Revised: February 16, 2007
Accepted: April 23, 2007
Published: May 24, 2007
DHX gene families of putative RNA helicases. Genomics 81, 618–622.
Anderson, S.F., Schlegel, B.P., Nakajima, T., Wolpin, E.S., and Parvin,
J.D. (1998). BRCA1 protein is linked to the RNA polymerase II holoen-
zyme complex via RNA helicase A. Nat. Genet. 19, 254–256.
Aratani, S., Fujii, R., Oishi, T., Fujita, H., Amano, T., Ohshima, T., Hagi-
wara, M., Fukamizu, A., and Nakajima, T. (2001). Dual roles of RNA
helicase A in CREB-dependent transcription. Mol. Cell. Biol. 21,
Brown, K.M., Chu, C.Y., and Rana, T.M. (2005). Target accessibility
dictates the potency of human RISC. Nat. Struct. Mol. Biol. 12, 469–
Chen, J.Y., Stands, L., Staley, J.P., Jackups, R.R., Jr., Latus, L.J., and
clear RNA can eliminate the requirement of Prp28p, an essential DEAD
box splicing factor. Mol. Cell 7, 227–232.
Chendrimada, T.P., Gregory, R.I., Kumaraswamy, E., Norman, J.,
Cooch, N., Nishikura, K., and Shiekhattar, R. (2005). TRBP recruits
the Dicer complex to Ago2 for microRNA processing and gene silenc-
ing. Nature 436, 740–744.
Chiu, Y.L.,and Rana, T.M. (2002).RNAiinhuman cells: basic structural
and functional features of small interfering RNA. Mol. Cell 10, 549–561.
Chiu, Y.L., and Rana, T.M. (2003). siRNA function in RNAi: a chemical
modification analysis. RNA 9, 1034–1048.
Chu, C.Y., and Rana, T.M. (2006). Translation repression in human
cells by microRNA-induced gene silencing requires RCK/p54. PLoS
Biol. 4, e210. 10.1371/journal.pbio.0040210.
Elbashir, S.M., Harborth, J., Lendeckel, W., Yalcin, A., Weber, K., and
Tuschl, T. (2001). Duplexes of 21-nucleotide RNAs mediate RNA inter-
ference in cultured mammalian cells. Nature 411, 494–498.
Friedemann, J., Grosse, F., and Zhang, S. (2005). Nuclear DNA heli-
case II (RNA helicase A) interacts with Werner syndrome helicase
and stimulates its exonuclease activity. J. Biol. Chem. 280, 31303–
Gregory, R.I., Chendrimada, T.P., Cooch, N., and Shiekhattar, R.
(2005).HumanRISC couples microRNA biogenesis and posttranscrip-
tional gene silencing. Cell 123, 631–640.
Gupta, V., Huang, X., and Patel, R.C. (2003). The carboxy-terminal, M3
motifs of PACT and TRBP have opposite effects on PKR activity. Virol-
ogy 315, 283–291.
Haase, A.D., Jaskiewicz, L., Zhang, H., Laine, S., Sack, R., Gatignol,
A., and Filipowicz, W. (2005). TRBP, a regulator of cellular PKR and
HIV-1 virus expression, interacts with Dicer and functions in RNA
silencing. EMBO Rep. 6, 961–967.
Hartman, T.R., Qian, S., Bolinger, C., Fernandez, S., Schoenberg,
D.R., and Boris-Lawrie, K. (2006). RNA helicase A is necessary for
translation of selected messenger RNAs. Nat. Struct. Mol. Biol. 13,
Khvorova, A., Reynolds, A., and Jayasena, S.D. (2003). Functional
siRNAs and miRNAs exhibit strand bias. Cell 115, 209–216.
Kim, D.H., Behlke, M.A., Rose, S.D., Chang, M.S., Choi, S., and Rossi,
J.J. (2005). Synthetic dsRNA Dicer substrates enhance RNAi potency
and efficacy. Nat. Biotechnol. 23, 222–226.
Lee, C.G., and Hurwitz, J. (1992). A new RNA helicase isolated from
HeLa cells that catalytically translocates in the 30to 50direction.
J. Biol. Chem. 267, 4398–4407.
Lee, C.G., and Hurwitz, J. (1993). Human RNA helicase A is homolo-
gous to the maleless protein of Drosophila. J. Biol. Chem. 268,
Lee, C.G., da Costa Soares, V., Newberger, C., Manova, K., Lacy, E.,
and Hurwitz, J. (1998). RNA helicase A is essential for normal gastrula-
tion. Proc. Natl. Acad. Sci. USA 95, 13709–13713.
Lee, Y., Ahn, C., Han, J., Choi, H., Kim, J., Yim, J., Lee, J., Provost, P.,
Radmark, O., Kim, S., et al. (2003). The nuclear RNase III Drosha initi-
ates microRNA processing. Nature 425, 415–419.
Lee, Y., Hur, I., Park, S.Y., Kim, Y.K., Suh, M.R., and Kim, V.N. (2006).
The role of PACT in the RNA silencing pathway. EMBO J. 25, 522–532.
Li, J., Tang, H., Mullen, T.M., Westberg, C., Reddy, T.R., Rose, D.W.,
and Wong-Staal, F. (1999). A role for RNA helicase A in post-transcrip-
tional regulation of HIV type 1. Proc. Natl. Acad. Sci. USA 96, 709–714.
Linder, P. (2006). Dead-box proteins: a family affair—active and pas-
sive players in RNP-remodeling. Nucleic Acids Res. 34, 4168–4180.
Liu, J., Carmell, M.A., Rivas, F.V., Marsden, C.G., Thomson, J.M.,
Song, J.J., Hammond, S.M., Joshua-Tor, L., and Hannon, G.J.
(2004). Argonaute2 is the catalytic engine of mammalian RNAi. Sci-
ence 305, 1437–1441.
Liu, J., Valencia-Sanchez, M.A., Hannon, G.J., and Parker, R. (2005).
MicroRNA-dependent localization of targeted mRNAs to mammalian
P-bodies. Nat. Cell Biol. 7, 719–723.
Lucchesi, J.C. (1998). Dosage compensation in flies and worms: the
ups and downs of X-chromosome regulation. Curr. Opin. Genet.
Dev. 8, 179–184.
Lund, E., Guttinger, S., Calado, A., Dahlberg, J.E., and Kutay, U.
(2004). Nuclear export of microRNA precursors. Science 303, 95–98.
Maniataki, E., and Mourelatos, Z. (2005). A human, ATP-independent,
RISC assembly machine fueled by pre-miRNA. Genes Dev. 19, 2979–
Matranga, C., Tomari, Y., Shin, C., Bartel, D.P., and Zamore, P.D.
(2005). Passenger-strand cleavage facilitates assembly of siRNA into
Ago2-containing RNAi enzyme complexes. Cell 123, 607–620.
Meister, G., Landthaler, M., Patkaniowska, A., Dorsett, Y., Teng, G.,
and Tuschl, T. (2004). Human Argonaute2 mediates RNA cleavage tar-
geted by miRNAs and siRNAs. Mol. Cell 15, 185–197.
Meister, G., Landthaler, M., Peters, L., Chen, P.Y., Urlaub, H., Luhr-
mann, R., and Tuschl, T. (2005). Identification of novel argonaute-as-
sociated proteins. Curr. Biol. 15, 2149–2155.
Myohanen, S., and Baylin, S.B. (2001). Sequence-specific DNA bind-
ing activity of RNA helicase A to the p16INK4a promoter. J. Biol.
Chem. 276, 1634–1642.
J.D., and Montminy, M. (1997). RNA helicase A mediates association
of CBP with RNA polymerase II. Cell 90, 1107–1112.
536 Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc.
RNA Helicase A in RISC
Patel, R.C., and Sen, G.C. (1998). PACT, a protein activator of the
interferon-induced protein kinase, PKR. EMBO J. 17, 4379–4390.
Pellizzoni, L., Charroux, B., Rappsilber, J., Mann, M., and Dreyfuss, G.
(2001). A functional interaction between the survival motor neuron
complex and RNA polymerase II. J. Cell Biol. 152, 75–85.
Peters, G.A., Hartmann, R., Qin, J., and Sen, G.C. (2001). Modular
structure of PACT: distinct domains for binding and activating PKR.
Mol. Cell. Biol. 21, 1908–1920.
and function of small RNAs. Nat. Rev. Mol. Cell Biol. 8, 23–36.
Rand, T.A., Petersen, S., Du, F., and Wang, X. (2005). Argonaute2
cleaves the anti-guide strand of siRNA during RISC activation. Cell
Reynolds, A., Leake, D., Boese, Q., Scaringe, S., Marshall, W.S., and
Khvorova, A. (2004). Rational siRNA design for RNA interference.
Nat. Biotechnol. 22, 326–330.
Rivas, F.V., Tolia, N.H., Song, J.J., Aragon, J.P., Liu, J., Hannon, G.J.,
and Joshua-Tor, L. (2005). Purified Argonaute2 and an siRNA form
recombinant human RISC. Nat. Struct. Mol. Biol. 12, 340–349.
Robb, G.B., Brown, K.M., Khurana, J., and Rana, T.M. (2005). Specific
and potent RNAi in the nucleus of human cells. Nat. Struct. Mol. Biol.
Siolas, D., Lerner, C., Burchard, J., Ge, W., Linsley, P.S., Paddison,
P.J., Hannon, G.J., and Cleary, M.A. (2005). Synthetic shRNAs as
potent RNAi triggers. Nat. Biotechnol. 23, 227–231.
Solem, A., Zingler, N., and Pyle, A.M. (2006). A DEAD protein that acti-
Tang, H., McDonald, D., Middlesworth, T., Hope, T.J., and Wong-
Staal, F. (1999). The carboxyl terminus of RNA helicase A contains
a bidirectional nuclear transport domain. Mol. Cell. Biol. 19, 3540–
Tetsuka, T., Uranishi, H., Sanda, T., Asamitsu, K., Yang, J.P., Wong-
Staal, F., and Okamoto, T. (2004). RNA helicase A interacts with nu-
clear factor kappaB p65 and functions as a transcriptional coactivator.
Eur. J. Biochem. 271, 3741–3751.
Tomari, Y., and Zamore, P.D. (2005). Perspective: machines for RNAi.
Genes Dev. 19, 517–529.
Tomari, Y., Du, T., Haley, B., Schwarz, D.S., Bennett, R., Cook, H.A.,
Koppetsch, B.S., Theurkauf, W.E., and Zamore, P.D. (2004a). RISC
assembly defects in the Drosophila RNAi mutant armitage. Cell 116,
Tomari, Y., Matranga, C., Haley, B., Martinez, N., and Zamore, P.D.
(2004b). A protein sensor for siRNA asymmetry. Science 306, 1377–
Walstrom, K.M., Schmidt, D., Bean, C.J., and Kelly, W.G. (2005). RNA
helicase A is important for germline transcriptional control, prolifera-
tion, and meiosis in C. elegans. Mech. Dev. 122, 707–720.
Zhong, X., and Safa, A.R. (2004). RNA helicase A in the MEF1 tran-
scription factor complex up-regulates the MDR1 gene in multidrug-
resistant cancer cells. J. Biol. Chem. 279, 17134–17141.
Zhou, Q., and Yik, J.H. (2006). The Yin and Yang of P-TEFb regulation:
implications for human immunodeficiency virus gene expression and
global control of cell growth and differentiation. Microbiol. Mol. Biol.
Rev. 70, 646–659.
Molecular Cell 26, 523–537, May 25, 2007 ª2007 Elsevier Inc. 537
RNA Helicase A in RISC