Automated High-Throughput RNAi Screening in Human
Cells Combined with Reporter mRNA Transfection to
Identify Novel Regulators of Translation
Claudia M. Casanova1,2, Peter Sehr2, Kerstin Putzker2, Matthias W. Hentze2, Beate Neumann2,
Kent E. Duncan3*, Christian Thoma1,2*
1Department of Medicine II, University Hospital of Freiburg, Freiburg, Germany, 2EMBL, Heidelberg, Germany, 3Center for Molecular Neurobiology, University Medical
Center Hamburg-Eppendorf, Hamburg, Germany
Proteins that promote angiogenesis, such as vascular endothelial growth factor (VEGF), are major targets for cancer therapy.
Accordingly, proteins that specifically activate expression of factors like VEGF are potential alternative therapeutic targets
and may help to combat evasive resistance to angiogenesis inhibitors. VEGF mRNA contains two internal ribosome entry
sites (IRESs) that enable selective activation of VEGF protein synthesis under hypoxic conditions that trigger angiogenesis.
To identify novel regulators of VEGF IRES-driven translation in human cells, we have developed a high-throughput screening
approach that combines siRNA treatment with transfection of a VEGF-IRES reporter mRNA. We identified the kinase MAPK3
as a novel positive regulator of VEGF IRES-driven translation and have validated its regulatory effect on endogenous VEGF.
Our automated method is scalable and readily adapted for use with other mRNA regulatory elements. Consequently, it
should be a generally useful approach for high-throughput identification of novel regulators of mRNA translation.
Citation: Casanova CM, Sehr P, Putzker K, Hentze MW, Neumann B, et al. (2012) Automated High-Throughput RNAi Screening in Human Cells Combined with
Reporter mRNA Transfection to Identify Novel Regulators of Translation. PLoS ONE 7(9): e45943. doi:10.1371/journal.pone.0045943
Editor: Sung Key Jang, Pohang University of Science and Technology, Republic of Korea
Received June 25, 2012; Accepted August 23, 2012; Published September 27, 2012
Copyright: ? 2012 Casanova et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by a grant from the Deutsche Forschungsgemeinschaft to C.T. (TH788/3-1). C.T. is a recipient of a Heisenberg-Fellowship from
the Deutsche Forschungsgemeinschaft (TH788/2-1 and TH788/2-2). The funders had no role in study design, data collection and analysis, decision to publish, or
preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: email@example.com (CT); firstname.lastname@example.org (KD)
mRNA translation by the ribosome is the ultimate step in the
expression of the ,20,000 human genes that encode proteins.
Regulation of this event- ‘translational control’- ensures that the
right amount of each protein is synthesized in the right place
within an organism or cell at the right time. Translational control
of gene expression plays a crucial role in adaptive cellular
responses to external stimuli  and failure to properly regulate
protein synthesis is a common feature of many diseases, including
cancer . Under normal physiological conditions, translation
initiates via a ‘cap-dependent’ mode, in which recruitment of the
small ribosomal subunit to the mRNA involves the 7-methyl-
guanosine (‘cap’) structure, located at the 59 end of cellular
mRNAs [3,4]. This interaction is mediated by the cytoplasmic
cap-binding complex eIF4F, which enables recruitment of other
translation initiation factors, scanning to the start codon, and
joining of the large ribosomal subunit for translational elongation
and protein synthesis [3,4,5]. Cap-dependent initiation appears to
be the dominant mode for most cellular mRNAs under most
conditions, and is the target of a wide variety of regulatory
mechanisms . However, certain viral RNAs and some
mammalian mRNAs can use alternatives to cap-dependent
initiation and thus are translated efficiently under conditions in
which cap-dependent translation is repressed, such as apoptosis,
mitosis, hypoxia, and cellular stress [6,7]. In these cases, trans-
lation initiation occurs efficiently independent of the cap structure
and many of the associated translation initiation factors .
The subset of cellular mRNAs (,3%) that apparently can be
translated efficiently when cap-dependent translation is generally
compromised  includes cell growth regulators that are critical in
cancer [2,9,10]. One important example of major clinical
relevance is the mRNA encoding vascular endothelial growth
factor-A (VEGF-A, henceforth referred to as ‘VEGF’). As a key
regulator of tumor angiogenesis, VEGF plays a crucial role in
cancer progression for essentially all solid tumors [11,12] and
consequently is a major oncology drug target. VEGF is also
important for development and maintenance of the nervous
system and both VEGF and regulators of VEGF signaling are of
great therapeutic interest in neurodegenerative disease and acute
neurological disorders, including cerebral ischemia/stroke [13,14].
Cap-independent translation of the VEGF mRNA is mediated by
two internal ribosome entry sequences (VEGF IRES-A and –B,
respectively) located in the 59 untranslated region (UTR) .
Cancer-relevant cellular stress conditions, such as hypoxia, can
activate cap-independent translation mediated by these and other
IRESs, while simultaneously repressing cap-dependent translation
[16,17,18]. Thus, in many cancers tumorigenesis involves a switch
enabling more cap-independent translation, which appears to be
important for tumor progression . Accordingly, ‘druggable’
specific positive regulators of the translational activity of VEGF
IRES (and perhaps other cellular IRESs) could potentially be
PLOS ONE | www.plosone.org1September 2012 | Volume 7 | Issue 9 | e45943
attractive additional therapeutic targets in oncology. However, to
our knowledge no cellular factors that specifically modulate VEGF
IRES activity have yet been identified.
The discovery that RNAi works effectively in mammalian cells
 paved the way for high-throughput RNAi screening to
address such questions in a global manner. Here, we present
a novel, automated, easily scalable high-throughput screening
approach that enables the identification of regulatory proteins that
modulate the translational activity of a specific mRNA element
contained in a transfected reporter mRNA. We took advantage of
the well established protocol for solid-phase reverse transfection of
cells on small interfering siRNA transfection mixes in coated 96-
well plate format . This robust approach routinely achieves
strong and specific knockdowns, but has so far been used
predominantly for high-content microscopy applications. Here
we significantly expand its range of use by optimizing its
compatibility with automated transfection of a reporter mRNA
of interest. Specifically, we present a robust method to: 1)
quantitatively analyze IRES-dependent translation and 2) com-
pare it with conventional cap-dependent translation activity using
monocistronic reporter mRNAs. Our approach addresses the
numerous caveats for studying IRES activity that are inherent to
the commonly used DNA transfection method with bicistronic
reporters, which has led to much confusion and contention in the
IRES translation field . For this reason, we expect the method
presented here to be a more sensitive screening tool with a much
lower false positive rate.
In a proof-of-concept screen we assessed the potential role of
702 human kinases and 298 human phosphatases in VEGF IRES-
driven translation. 91 genes that qualified as hits in the primary
screen were selected for secondary validation assays, in which
specificity for IRES-driven translation was examined by transfect-
ing the original VEGF IRES-driven reporter mRNA and a cap-
driven reporter mRNA. Ultimately, we identified three kinases
specifically involved in IRES-, but not in cap-dependent trans-
lational regulation. For one of these, MAPK3, we show here that it
is a bona fide novel, positive-acting, post-transcriptional regulator
of VEGF production. In our view this provides a striking
demonstration of the power of this method for identifying novel
regulators of mRNA translation. Importantly, the method can be
easily scaled up for genome-wide screens and could also be readily
adapted for use with other mRNA elements of (clinical) interest.
Combining RNAi with mRNA Transfection Enables High-
throughput Screening for Human Kinases and
Phosphatases that Regulate mRNA Translation
To identify potential regulators of VEGF IRES-dependent
translation we developed a general RNAi screening approach
integrated with mRNA transfections. In the first step siRNA
transfection complexes are ‘solid-phase reverse transfected’ (i.e.
cells are added to plates coated with lyophilized siRNAs trans-
fection mixes) . Subsequently, a second RNA transfection
introduces a reporter mRNA whose structure can be modified
according to the interest of the investigator. Transfection mixes
containing siRNAs of interest are distributed to 96-well plates
using automated liquid handling (Fig. 1). Lyophilization of the
plates allows long-term storage and guarantees ‘‘ready to trans-
fect’’ plates with similar transfection efficiency for up to 15 months
. We have chosen to perform the screen in the HeLa human
cell line because: I) the established siRNA treatment procedure is
optimized for these cells; II) the VEGF IRES is active (Fig. S1A)
and III) and responds to low oxygen tension (hypoxia, 0.7% O2)
with an approximately 3-fold activation (Fig. S1B). After cell
seeding, siRNA solid-phase reverse transfection is performed for
48 hours under normoxic conditions (Fig. 1). Subsequently, cells
are transfected with the reporter mRNA of interest, in this case
a firefly luciferase (FLuc) reporter containing the VEGF IRES
elements a and b and a non-physiological adenosine cap structure
(‘A-cap’). The A-cap maintains stability of the mRNA, but is not
recognized by the cytoplasmic cap binding complex [22,23] and
thereby ensures that translation of this reporter mRNA is driven in
a cap-independent mode via the VEGF IRES contained in the
mRNA 59 UTR (Fig. 1). For this screen cells are incubated under
hypoxic conditions (0.7% oxygen tension) after reporter mRNA
transfection for 6 hours prior to lysis and measurement of Fluc
activity (Fig. 1), since this mimics the physiological condition under
which the VEGF IRES is maximally active. Importantly, direct
introduction of a reporter mRNA in the second step is designed to
significantly limit the number of hits to those that affect
translation, as opposed to other steps in gene expression.
Identification of 91 Potential Novel Regulators of VEGF
We screened 702 human kinases (Table S1) and 298 human
phosphatases (Table S2). Each gene was targeted with three
different siRNA sequences (Tables S1 and S2), each in an
independent reaction. Pilot experiments revealed low well-to-well
variability in different wells on different plates (Fig. S2). Neverthe-
less, to ensure assay quality, each reaction was performed in
triplicate on separate plates. In other words, each gene was
analyzed in 9 independent reactions in total. Each value was
normalized to the median value of four scrambled siRNA negative
control reactions (100) on the same plate. The median value of the
triplicate was obtained from the normalized individual values.
siRNAs targeting firefly luciferase (FLuc) were used as positive
controls. In principle, silencing of a defined kinase or phosphatase
can result in either a higher (‘Up Hit’) or a lower (‘Down Hit’)
production of luciferase, due to the corresponding modulation of
VEGF IRES activity. We used the frequency distribution of the
normalized median value of all samples to define cut-off values of
250 and 80 for Up and Down hits, respectively in order to give
a hit rate ,10% (Fig. 2A). From the total number of siRNAs
tested, siRNAs targeting 457 kinases and 188 phosphatases
affected VEGF IRES activity. Among them only 69 kinases and
22 phosphatases had a similar effect on VEGF IRES activity when
silenced with at least 2 different siRNA sequences in independent
reactions: 64 genes behaved as negative regulators and 27 as
positive regulators of VEGF IRES function (Fig. 2B). Genes
identified with two or more siRNAs in individual reactions are
most likely true hits rather than false positives due to off target
effects. The use of silencer select siRNAs also reduced the
likelihood of off target effects . We re-tested the 91 genes (69
kinases and 22 phosphatases) that were identified with at least two
siRNAs (Fig 2B). These genes have been grouped on the basis of
their relevance in particular biological processes, according to
Gene Ontology (GO) annotation (Fig. 2C). This classification
revealed that: (1) 36% of the genes participate in cell signaling
pathways, (2) 18% are genes with unknown function, and (3) 2%
are previously linked to tumorigenesis/angiogenesis.
Secondary Screen for Hit Verification and Determination
of Specificity for VEGF IRES Translation
In secondary validation experiments we aimed to: 1) identify
genes specifically showing an effect on IRES, but not on cap-
dependent translation, and 2) exclude genes affecting cell viability
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org2 September 2012 | Volume 7 | Issue 9 | e45943
that might be indirectly affecting protein synthesis. Validation
experiments were performed according to the workflow depicted
in Fig. 3A. To test specificity, the original VEGF IRES-driven
reporter mRNA and a reporter mRNA containing a physiological
cap structure, but lacking VEGF IRES sequences were trans-
fected. Confirmed hits that showed no effect on the cap-driven
reporter mRNA were considered IRES-specific. Hits that specif-
ically affected IRES-, but not cap-dependent translation were then
assessed for their effect on cell viability through ATP measure-
ment. Following this workflow, we identified three kinases that
specifically affected IRES-, but not cap-dependent translation, and
that had no effect on cell viability when targeted by siRNAs
(Fig. 3B and Table S3).
Novel Role for MAPK3 as a Positive Regulator of VEGF
IRES Activity and VEGF Protein Levels
One of the three confirmed IRES-specific regulators with no
effect on cell viability was mitogen-activated protein kinase 3
(MAPK3), also known as ERK1, which we selected for further
validation experiments. siRNA-mediated reduction of MAPK3
levels by approximately 70% (Figure 4D) led to a 50% decrease in
VEGF IRES Fluc reporter mRNA activity (Fig. 4A, left panel), but
had no effect on cap-dependent Fluc reporter mRNA activity
(Fig. 4A, right panel). Mock and scrambled siRNA reverse
transfection (negative controls) did not affect translation of either
reporter. siRNAs targeting the Fluc coding sequence or Polo-Like
Kinase 1‘‘ (PLK1) served as positive controls for RNAi efficacy.
Both led to a reduction of Fluc levels for both IRES- and cap-
dependent reporter mRNAs, the former due to RNAi and the
latter as an indirect consequence of impaired growth (silencing of
PLK1 leads to a prometaphase arrest followed by apoptosis .
Importantly, the reduction in reporter expression in MAPK3
siRNA-treated cells is not due to changes in mRNA abundance, as
shown in Fig. 4C. To rule out that the reduction in VEGF IRES
translation due to MAPK3 siRNA action might result from an
indirect effect on cell growth, we assessed cell viability through
ATP measurement. As a positive control we silenced PLK1, which
led to a commensurate reduction of ATP counts to 25%, as
expected (Fig. 4B). In contrast, silencing of MAPK3 had almost no
effect on cell viability (Fig. 4B). Taken together, these observations
strongly suggest that MAPK3 is a novel specific activator of VEGF
Figure 1. Workflow of the high-throughput siRNA screening approach to identify novel regulators of mRNA translation. The strategy
features automated generation of lyophilized siRNA-coated 96 well plates, followed by two independent RNA transfections. Plates are first coated
with siRNAs transfection mixes, cells are then seeded onto siRNA-coated plates (solid-phase reverse transfection) and incubated for two days to allow
reduction of the protein levels of the targeted genes. Subsequently, a second RNA transfection is performed to introduce the reporter mRNA of
interest. Finally, the effects of each siRNA knock-down are measured.
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org3 September 2012 | Volume 7 | Issue 9 | e45943
If translation driven by the VEGF IRES is contributing
significantly to VEGF protein levels under hypoxia, then our
results predict that MAPK3 knockdown would lead to a reduction
in endogenous VEGF protein levels. To test this prediction we
wanted to use as reproducible and quantitative an assay as
possible, since we expected on the basis of our mRNA reporter
assays to see a relatively modest reduction in VEGF levels due to
MAPK3 knockdown. We therefore analyzed secreted VEGF levels
by ELISA, which is typically superior to immunoblotting for
accurate quantification of mild changes and is a standard method
for monitoring changes in endogenous VEGF levels . HeLa
cells were seeded on siRNA-coated 96 well plates and grown for 24
Figure 2. Identification of kinases and phosphatases modulating VEGF IRES activity. (a) Frequency distribution of the normalized median
values of all samples (kinases and phosphatases). Numbers of normalized median values in each bin are shown on the y-axis, bins are shown on the x-
axis. The cut-off values for up and down hits are indicated. (b) 702 kinases and 298 phosphatases were screened with three different siRNA sequences
in three independent reactions, each performed in triplicates. Among them 69 kinases and 22 phosphatases had a similar effect on VEGF IRES activity
when silenced with at least 2 different siRNA sequences in independent reactions. These 91 genes have been considered for further validation in
secondary screening experiments. (c) Schematic classification of the 91 genes selected for validation studies according to their described function in
GO annotation. * Two of the specific hits belong to the ‘‘Unknown’’ group and one to the ‘‘Cell signaling’’ group.
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org4 September 2012 | Volume 7 | Issue 9 | e45943
hours under normoxic conditions, followed by incubation under
hypoxic conditions for 24 hours. Thereafter, ELISA analysis of
VEGF secretion was performed, which revealed that MAPK3
depletion causes a 25% reduction of VEGF (Fig. 5A). Statistical
analysis was performed as described in ‘‘Methods’’ and achieved
p=0.00063. To determine whether these effects are due to an
effect on endogenous VEGF mRNA stability, RT-qPCR experi-
ments were performed: down-regulation of MAPK3 by RNAi led
to a significant reduction of endogenous MAPK3 mRNA levels to
25% (Fig. 5B), demonstrating that RNAi was effective. However,
endogenous VEGF mRNA levels were not affected (Fig. 5B).
Thus, the reduction of VEGF protein levels in MAPK3 siRNA-
treated cells is not due to changes in transcription or mRNA
abundance. These data demonstrate a novel functional role for
MAPK3 as a bona fide positive regulator of endogenous VEGF
protein levels in human cells. Taken together with our reporter
mRNA results described above, these data strongly suggest that
the mechanism by which MAPK3 affects VEGF protein levels is
through specific modulation of VEGF IRES-dependent trans-
lation. To our knowledge this is the first demonstration of
a signaling pathway kinase that stimulates VEGF IRES trans-
lational activity without affecting cap-dependent translation. We
interpret our ability to identify such a factor through the relatively
small screen described here as strong evidence of the power of our
Regulation of ‘gene expression’ has been viewed historically as
a primarily nuclear event controlled by variations in transcription
factor activity. More recent developments indicate that trans-
lational regulation of gene expression is not only common, but also
a major contributor to regulation of protein expression levels. This
paradigm shift has been driven in part by the discovery of
microRNAs as major cytoplasmic regulators of gene expression
[26,27,28], as well as the development of functional genomic and
proteomic approaches that enable large-scale quantitative analysis-
and therefore direct comparison- of the correspondence between
mRNA and protein levels under different cellular conditions .
At the same time, there has been increasing appreciation of the
major role of altered translational control in numerous human
diseases, particularly cancer, and a corresponding major interest in
targeting translational regulatory factors for drug discovery
[2,30,31,32,33,34,35]. In principle, cell-based RNAi screening
should be a powerful approach for identification of regulators of
mRNA translation. However, we are not aware of any publication
describing a screen for regulators of cellular translation in a high-
throughput format in mammalian cells. A major reason for this is
undoubtedly technical: translation is a ‘downstream step’ in the
pathway from gene to protein, and therefore the usual genome-
wide RNAi screening methodologies do not enable direct and
specific assessment of translation. For example, screening
Figure 3. Secondary screen of 91 candidate regulators of VEGF IRES-dependent translation. (a) Workflow of the secondary screen. Cells
were reverse-transfected with individual siRNAs designated as hits from the primary screen (Table S1 and S2). This was followed 48 hours later by
three assays which were designed, respectively, to assess: 1) specificity of the effects for IRES-driven translation and 2) to determine effects on cell
viability that might confound the other analyses. (b) Summary of secondary screening data.
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org5 September 2012 | Volume 7 | Issue 9 | e45943
approaches that rely on traditional DNA-based reporter genes are
sensitive to identification of proteins affecting other steps of gene
expression. This is likely to be particularly problematic for cellular
IRES element-mediated translation, where using common DNA-
based transfection approaches such as bicistronic reporters bear
the risk that few of the molecules identified will specifically affect
translation. The lower level of activity that is typical for cellular
IRESs in comparison to cap- or viral IRES-driven translation
makes it difficult to exclude that changes in activity observed with
DNA-based strategies result from effects on other steps in gene
expression, such as cryptic promoter or splicing activity .
While it may sometimes be possible through rigorous controls to
demonstrate IRES activity using DNA reporters, this is not easily
amenable to a high-throughput format.
Here we have presented a novel high-throughput RNAi
screening strategy that avoids by design all of the problems
associated with DNA-encoded reporters. Our approach combines
highly efficient siRNA solid-phase reverse transfection with
subsequent direct reporter mRNA transfection. We reasoned that
this combination would be more likely to identify bona fide
regulators of VEGF levels that act by specific translational
mechanisms. This idea is borne out by our successful identification
of three IRES-specific translational regulatory proteins in what
was primarily intended to be a proof-of-concept screen. In this
study we focused on the further analysis of MAPK3. The
additional kinases will be the subject of future studies.
Importantly, we have also demonstrated here that MAPK3 is
a bona fide regulator of endogenous VEGF expression under
hypoxia in human cells (Figure 5A). This effect is independent
of changes in VEGF mRNA levels (Fig. 5B), as expected from
an effect of MAPK3 on the VEGF IRES. We note that
reducing MAPK3 levels by 75% leads to a 25% reduction of
endogenous VEGF levels. While this might seem to be a modest
effect, we think it is unlikely that the effect on endogenous
Figure 4. MAPK3 specifically regulates VEGF IRES-dependent translation. (a) HeLa cells were reverse transfected with the indicated siRNAs.
This was followed 48 h later by reporter mRNA transfection for IRES-driven translation (left panel) and cap-driven translation (right panel). Luciferase
activity was analyzed and the corresponding relative translation rate is indicated. (b) Quantification of ATP content as an indicator of cell viability. (c)
Physical stability of the reporter VEGF IRES luciferase mRNA analyzed by Northern blotting. (d) Western blot of MAPK3 in siRNA-treated cells or control
cells treated with scrambled siRNA. Beta-actin serves as a control for loading/transfer efficiency.
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org6 September 2012 | Volume 7 | Issue 9 | e45943
VEGF would be much stronger for two reasons. First, the
knockdown is not complete and kinases can act catalytically.
Second, the effect on the VEGF IRES reporter upon MAPK3
was only ,50% reduction and presumably this result sets
a lower bound for the magnitude of any effect on endogenous
VEGF levels. In fact, the magnitude of endogenous VEGF
reduction in response to MAPK3 depletion is less than that
observed with our reporter (25% vs. 50% reduced, respectively).
Although we do not know the exact reason for this difference,
one possibility is that it relates to the large number of regulatory
controls that act to tune VEGF levels within a relatively narrow
range . Indeed, ample evidence suggests that induction of
tumor angiogenesis, a hallmark of cancer, is governed by
a complex biological rheostat that senses numerous positive and
negative inputs that converge through multiple mechanisms on
regulation of VEGF protein levels . Thus, there might be
a positive compensatory mechanism that would act home-
ostatically to raise levels of endogenous VEGF when IRES
activity is reduced by MAPK3 knockdown (e.g. increased
endogenous VEGF-A transcription or VEGF protein stability).
In contrast, the VEGF IRES luciferase reporter would be
immune to these compensatory mechanisms. Regardless of why
MAPK3 depletion decreases VEGF levels by only 25%, this
reduction would be expected to be functionally relevant, since it
has previously been demonstrated in mice that a 25–30%
reduction in VEGF levels can produce dramatic biological
consequences in the nervous system . We conclude that
control of VEGF IRES-dependent translation via MAPK3
contributes an important additional regulatory layer for proper
maintenance of VEGF protein levels within a biologically
optimal range. Accordingly, the IRES-mediated translational
control mechanism that we have described here might be useful
as a novel target for therapeutic strategies to modulate VEGF
How does MAPK3 promote
Previous work has highlighted a hypoxia-controlled switch from
cap-dependent to cap-independent translation in breast cancer
that results from an increase in eIF4E-Binding Protein 1 (4E-
BP1) activity . Thus, MAPK3 might conceivably regulate
VEGF IRES translation via effects on 4E-BP1 activity. We
consider this unlikely for several reasons. First, in contrast to
MAPK1, which has been shown to phosphorylate 4EBP1 in
vitro , we are not aware of any evidence that MAPK3 can
directly phosphorylate 4E-BP1. MAPK3 could conceivably
indirectly promote 4E-BP1 phosphorylation, since it can act
through Tsc1/2 as an upstream activator of mTOR, the main
kinase known to phosphorylate 4E-BP1 in vivo . However,
phosphorylation of 4E-BP1 leads to reduced eIF4E binding and
therefore reduced inhibition of eIF4E activity. Thus, regardless
of whether MAPK3 would act directly or indirectly to promote
phosphorylation of 4E-BP1, the prediction in either case would
be that silencing of MAPK3 would lead to less phosphorylation
of 4E-BP1, with resultant inhibition of cap-dependent trans-
lation and stimulation of IRES translation. However, this is not
what we observe: we found no effect on cap-dependent
translation and a decrease in VEGF IRES function in MAPK3
siRNA-treated cells. We therefore conclude that our screening
strategy has uncovered a novel IRES-specific translational
stimulatory signaling mechanism mediated by MAPK3. Future
work will explore the nature of the underlying signaling cascade
in this pathway and how it impinges on the translational
machinery to specifically promote VEGF production.
In summary, we report a powerful method to identify novel
regulators of mRNA translation. We have applied our screening
method here to the VEGF IRES and have thereby identified
a novel function for MAPK3 as a specific, positive regulator of
VEGF-IRES and VEGF expression. Our approach is readily
applicable to other mRNA regulatory elements and should also
be straightforward to adapt to other cell types. Indeed, we think
the applications of this approach are broad. Areas of clear
future interest would be to screen with additional IRESs or
other mRNA regulatory motifs implicated in disease, such as
upstream open reading frames, microRNA or regulatory protein
binding sites , as well as alternative polyadenylation sites
[38,39]. More challenging, but equally interesting applications
would be to use this method to identify specific translational
regulators in other cell types, for example other cancer cell
lines, primary cell cultures, stem cells or differentiated cell types
obtained from patient-derived induced pluripotent stem cells
Figure 5. MAPK3 specifically regulates endogenous VEGF
expression without affecting mRNA stability. (a) Determination
of endogenous VEGF by ELISA. Error bars represent standard deviations
calculated from 3 independent experiments, each performed at least in
duplicates. Statistical analysis of differences between controls and
experimental samples were performed with the Microsoft Excel
unpaired, type 2, Student’s t test. (b) Endogenous VEGF and MAPK3
mRNA levels measured by qRT-PCR. The expression ratio of the
indicated endogenous mRNAs (VEGF and MAPK3) in MAPK3 siRNA-
treated cells relative to scramble siRNA treated cells (%) is shown.
Statistical analysis of differences between controls and experimental
samples were performed with the Microsoft Excel unpaired, type 2,
Student’s t test. A p-value of .0.5 and ,0.05 was determined for VEGF
and MAPK3, respectively. Relative VEGF and MAPK3 mRNA levels,
standard deviations and statistical significance were calculated with
REST 2009 software as described in ‘‘Material and Methods’’.
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org7 September 2012 | Volume 7 | Issue 9 | e45943
Materials and Methods
Cell Culture, Cell Extracts, Antibodies, Plasmids and in
HeLa Kyoto cells used in this study  were grown, unless
elsewhere indicated in 4.5 g/L glucose DMEM, 10% FCS, 1%
penicillin/streptomycin, 1% glutamine. Hypoxic studies were
performed at 0.7% O2 and 5% CO2. All experiments were
performed with cells in the exponential growth phase at sub-
confluent (,70%) density. For Western Blot analysis, HeLa total
cell extracts were prepared in RIPA buffer (10 mM Tris-HCl
pH 8.0, 150 mM NaCl, 1 mM EDTA, 1% NP40, 0.1% SDS)
containing freshly added CompleteTMEDTA-free Protease In-
hibitor Cocktail (Roche). Mouse monoclonal antibodies against
beta-actin and MAPK3 were purchased from Sigma and Abcam,
respectively, and anti-mouse secondary antibody from Amersham.
The plasmid encoding firefly luciferase pT3luc(pA) has been
described previously . The VEGF lucA construct with an A98
tail containing the VEGF 59UTR (1048 nt) upstream of the firefly
luciferase coding sequence cloned in a pBluescript II KS vector
was kindly provided by Antje Ostareck-Lederer. The ‘‘no-IRES’’
control was described previously . VEGF lucA templates
linearized with Not1 for VEGF IRES mRNAs, no IRES control
templates linearized with ECL136II for no IRES control mRNAs
or pT3luc(pA) templates linearized with BamH1 for cap mRNAs
were used with the T3 MEGAscript kit (Applied Biosystems/
Ambion; Austin, TX) and either ApppG cap analogs (NEB;
Ipswich, MA) (IRES and no IRES control mRNAs) or 3-O-Me-
m7G(5)ppp(5)G (cap-mRNAs) to give 80% capping efficiency.
RNAs were purified via RNeasy (QIAGEN). mRNA concentra-
tion and integrity were assessed by OD measurement and agarose
Preparation of 96 Well Plates for siRNA Solid-phase
21 nt RNA duplexes were obtained from Ambion Europe, Ltd,
all with the silencer select modification . The full list of siRNA
sequences used in this study is available in the Table S1 and Table
S2. Control silencer select siRNA sequences (scrambled, firefly and
PLK1) are available in Table S4.
White 96 well plates (NunclonTMDelta Surface, Nunc) were
coated with siRNA transfection solutions using a Microlab STAR
pipetting robot (Hamilton), similar as previously described . In
detail: siRNA stock solution were prepared by dissolving siRNAs
with milliQ water to a final concentration of 3 mM. 3 ml
OptiMEM, containing 0.4 M sucrose was transferred to each
well of a 384 well low volume plate. 1.75 ml water and 1.75 ml
Lipofectamine 2000 was added to each well followed by a 8 times
mixing step. 5 ml of the respective siRNA stock solution (3 mM)
was added to each well followed again by a 8 times mixing step.
After incubation of 30 min at RT 7.25 ml of a 0.2% (w/v) gelatin
solution was added and the final solution was mixed 8 times using
the slower mixing mode possible on the MICROLAB STAR from
Hamilton. Plates were lyophilized in Concentrator System from
Genevac called Mivac Quattro (purchased via Fisher Scientific)
and stored in plastic boxes containing drying orange heavy metal
free (Fluka, catalog number 94098).
Solid-phase siRNA Reverse Transfection, Reporter mRNA
Transfection and Luciferase Assays
HeLa cells resuspended in low glucose medium (1 g/L) were
seeded onto siRNA-coated plates (3000 cells/well of a 96 well plate
in a volume of 100 ml/well), using a FlexDropTMIV EXi bulk
dispenser (PerkinElmer). After 48 hours of culture under normoxic
conditions the medium was removed and the second transfection
with reporter mRNA performed using TransMessenger trans-
fection reagent (Qiagen), according to manufacturer’s instructions
in serum-free medium. Same molar amounts of reporter mRNA,
corresponding to 150 ng of VEGF IRES reporter mRNA and
80 ng of no-IRES reporter mRNA, were transfected per well, with
a RNA:enhancer 1:2 ratio and a RNA:Transmessenger 1:4 ratio.
The transfection-mix (42 ml/well) was transferred immediately
after removal of the old medium onto 96 well plates with
a EvolutionTMP3 pipetting robot (PerkinElmer). After 6 hours
incubation under hypoxic conditions firefly luciferase was quan-
tified with the BriteliteTMplus Reporter Gene Assay System
(PerkinElmer). 40 ml/well BriteLite reagent was added with the
FlexDropTMbulk dispenser and luminescence measured with an
EnVisionTMHTS plate reader (PerkinElmer).
Effect of siRNA Transfection on Cell Proliferation
Similarly to the reporter assay, cells were solid-phase reverse
transfected and incubated for 48 hours in siRNA-coated plates.
Next, medium was replaced with serum-free medium and plates
were incubated for an additional 6 hours under low oxygen
conditions to induce hypoxia. Cell viability was then monitored
using the ATPlite 1step Luminescence Assay System (PerkinEl-
mer). Briefly, 40 ml/well ATPlite reagent was added with the
FlexDropTMIV EXi (PerkinElmer) and the ATP-dependent
chemilumiscence was measured as indicator of cell viability read
with an EnVisionTMHTS plate reader.
Total RNA was extracted from siRNA treated HeLa cells in 96
well plate format using 50 ml/well TRIzol reagent (Invitrogen),
according to manufacturer’s instructions. Equivalent amounts of
total RNA were loaded on 1% agarose gels containing 1%
formaldehyde and this was verified by ethidium bromide staining.
RNA was transferred onto positively charged nylon membranes
(Roth). Hybridisation was done with a
obtained upon digestion of the firefly luciferase coding plasmid
previously reported  with restriction enzymes NcoI and EclII.
Genes selected in the first screening round were classified with
Gene Ontology term annotation. GO terms were assigned to the
first suitable of the following function groups: gene expression,
tumorigenesis/angiogenesis, cell signaling, cell cycle, other and
Studies on endogenous VEGF production and secretion into
medium were performed with siRNA reverse transfected cells in
a 96 well plate format. Cells were seeded as described above and
cultured for 24 hours under normoxic conditions. After a medium
exchange cells were further incubated for 24 hours under hypoxic
conditions before VEGF concentration in the medium was
determined with VEGF Quantikine Colorimetric Sandwich
ELISA (R&D Systems) according to the manufacturer’s instruc-
Reverse Transcription-quantitative PCR (RT-qPCR)
Cell lysis, RNA extraction, DNAse treatment and reverse
transcription were performed using the Power SYBR Green Cells-
to-CTKit (Applied Biosystems), according to the manufacturer’s
instructions. RT-qPCR analysis of endogenous VEGF mRNAs
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org8September 2012 | Volume 7 | Issue 9 | e45943
was performed using the Power SYBR Green Cells-to-CTKit in
a 7500 Real-Time PCR System (Applied Biosystems). Forward
and reverse primer sequences used to detect VEGF, MAPK3 and
actin b mRNAs are available in Table S5. Transcripts levels in
MAPK3 siRNA-treated cells were normalized to the expression
levels of beta-actin mRNA and measured relative to those in
scrambled siRNA-treated cells as previously described .
Standard deviations and statistic analysis was performed with
REST? 2009 Software .
stimulation by hypoxia. (a) VEGF IRES is functional in HeLa
cells. HeLa cells were transfected with A-capped reporter mRNAs
with or without the VEGF IRES element. 6 hours later FLuc
protein levels were measured. (b) HeLa cells transfected with
VEGF IRES reporter mRNA were incubated either under
normoxic or hypoxic conditions for 6 hours before assaying Fluc
VEGF IRES activity in HeLa cells and its
plate format displays very low well-to-well variability. HeLa cells
were seeded in four different random positions on three different
96 well plates and reverse transfected with scrambled siRNAs
(negative control). 48 hours later they were transfected with the
VEGF IRES Fluc reporter mRNA and 6 hours later FLuc
reporter expression was measured. Very similar VEGF IRES
activity was observed in different wells (indicated on the x axis) and
on different plates (standard deviation).
Robustness of Pilot screen. Pilot screen in 96 well
antisense siRNA sequences and normalized luciferase values for
the screened 702 kinases.
Kinase library. Gene symbol, RefSeq, sense- and
and antisense siRNA sequences and normalized luciferase values
for the screened 298 phosphatases.
Phosphatase library. Gene symbol, RefSeq, sense-
RefSeq, sense- and antisense siRNA sequences for the 3 IRES
Confirmed IRES specific hits. Gene symbol,
siRNA control sequences.
Primer sequences used for qRT-PCR experi-
We are grateful to Antje Ostareck-Lederer for providing the VEGF lucA
plasmid. We thank Vladimir Benes (Gene Core Facility, EMBL), Joe Lewis
(Chemical Core Facility, EMBL) and the team of the Advanced Light
Microscopy Core Facility, EMBL, for advice and discussions, Petra
Binninger for excellent technical assistance, and the members of the
Thoma lab for discussions.
Conceived and designed the experiments: CC KP PS BN KD CT.
Performed the experiments: CC KP PS CT. Analyzed the data: CC PS
MH BN KD CT. Wrote the paper: CC KD CT.
1. Gebauer F, Hentze MW (2004) Molecular mechanisms of translational control.
Nat Rev Mol Cell Biol 5: 827–835.
2. Silvera D, Formenti SC, Schneider RJ. (2010) Translational control in cancer.
Nat Rev Cancer. 10: 254–66.
3. Jackson RJ, Hellen CU, Pestova TV (2010) The mechanism of eukaryotic
translation initiation and principles of its regulation. Nat Rev Mol Cell Biol 11:
4. Sonenberg N, Hinnebusch AG (2009) Regulation of translation initiation in
eukaryotes: mechanisms and biological targets. Cell 136: 731–745.
5. Pestova TV, Lomakin IB, Lee JH, Choi SK, Dever TE, et al. (2000) The joining
of ribosomal subunits in eukaryotes requires eIF5B. Nature 403: 332–335.
6. Holcik M, Sonenberg N (2005) Translational control in stress and apoptosis. Nat
Rev Mol Cell Biol 6: 318–327.
7. Gilbert WV, Zhou K, Butler TK, Doudna JA (2007) Cap-independent
translation is required for starvation-induced differentiation in yeast. Science
8. Johannes G, Carter MS, Eisen MB, Brown PO, Sarnow P (1999) Identification
of eukaryotic mRNAs that are translated at reduced cap binding complex eIF4F
concentrations using a cDNA microarray. Proc Natl Acad Sci U S A 96: 13118–
9. Holcik M (2004) Targeting translation for treatment of cancer–a novel role for
IRES? Current cancer drug targets 4: 299–311.
10. Stoneley M, Willis AE (2004) Cellular internal ribosome entry segments:
structures, trans-acting factors and regulation of gene expression. Oncogene 23:
11. Hanahan D, Weinberg RA (2011) Hallmarks of cancer: the next generation. Cell
12. Harris AL (2002) Hypoxia–a key regulatory factor in tumour growth. Nature
reviews 2: 38–47.
13. de Almodovar CR, Lambrechts D, Mazzone M and Peter Carmeliet (2009) Role
and Therapeutic Potential of VEGF in the Nervous System. Physiol Rev 89:
14. Mackenzie F, Ruhrberg C (2012) Diverse roles for VEGF-A in the nervous
system. Development 139: 1371–80.
15. Huez I, Creancier L, Audigier S, Gensac MC, Prats AC, et al. (1998) Two
independent internal ribosome entry sites are involved in translation initiation of
vascular endothelial growth factor mRNA. Mol Cell Biol 18: 6178–6190.
16. Akiri G, Nahari D, Finkelstein Y, Le SY, Elroy-Stein O, et al. (1998) Regulation
of vascular endothelial growth factor (VEGF) expression is mediated by internal
initiation of translation and alternative initiation of transcription. Oncogene 17:
17. Stein I, Itin A, Einat P, Skaliter R, Grossman Z, et al. (1998) Translation of
vascular endothelial growth factor mRNA by internal ribosome entry:
implications for translation under hypoxia. Mol Cell Biol 18: 3112–3119.
18. Braunstein S, Karpisheva K, Pola C, Goldberg J, Hochman T, et al. (2007) A
hypoxia-controlled cap-dependent to cap-independent translation switch in
breast cancer. Molecular Cell 28: 501–512.
19. Elbashir SM, Harborth J, Weber K, Tuschl T (2002) Analysis of gene function in
somatic mammalian cells using small interfering RNAs. Methods 26: 199–213.
20. Erfle H, Neumann B, Rogers P, Bulkescher J, Ellenberg J, et al. (2008) Work
flow for multiplexing siRNA assays by solid-phase reverse transfection in
multiwell plates. J Biomol Screen 13: 575–580.
21. Gilbert WV (2010) Alternative ways to think about cellular internal ribosome
entry. J Biol Chem 285: 29033–29038.
22. Gebauer F, Corona DFV, Preiss T, Becker PB, Hentze MW (1999)
Translational control of dosage compensation in Drosophila by Sex-lethal:
cooperative silencing via the 59 and 39 UTRs of msl-2 mRNA is independent of
the poly(A) tail. EMBOJ 18: 6146–6154
23. Duncan KE, Strein C, Hentze MW (2009) The SXL-UNR corepressor complex
uses a PABP-mediated mechanism to inhibit ribosome recruitment to msl-2
mRNA. Mol. Cell 36: 571–582.
24. Puri N, Wang X, Varma R, Burnett C, Beauchamp L, et al. (2008) LNA
incorporated siRNAs exhibit lower off-target effects compared to 2’-OMethoxy
in cell phenotypic assays and microarray analysis. Nucleic acids symposium
series 52: 25–26.
25. Sumara I, Gimenez-Abian JF, Gerlich D, Hirota T, Kraft C, et al. (2004) Roles
of polo-like kinase 1 in the assembly of functional mitotic spindles. Curr Biol 14:
26. Bushati N, Cohen SM (2007) microRNA functions. Annu Rev Cell Dev Biol. 23:
27. Fabian MR, Sonenberg N (2012) The mechanics of miRNA-mediated gene
silencing: a look under the hood of miRISC. Nat Struct Mol Biol. 19: 586–93.
28. Pasquinelli AE (2012) MicroRNAs and their targets: recognition, regulation and
an emerging reciprocal relationship. Nat Rev Genet. 13: 271–82.
29. Schwanha ¨usser B, Busse D, Li N, Dittmar G, Schuchhardt J, et al. (2011) Global
quantification of mammalian gene expression control. Nature. 473: 337–42.
30. Stumpf CR, Ruggero D (2011) The cancerous translation apparatus. Curr Opin
Genet Dev. 21: 474–83.
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org9 September 2012 | Volume 7 | Issue 9 | e45943
31. Robert F, Pelletier J (2009) Translation initiation: a critical signalling node in Download full-text
cancer. Expert Opin Ther Targets. 13: 1279–93.
32. Lujambio A, Lowe SW (2012) The microcosmos of cancer. Nature. 482: 347–55.
33. Hsieh AC, Liu Y, Edlind MP, Ingolia NT, Janes MR, et al. (2012) The
translational landscape of mTOR signalling steers cancer initiation and
metastasis. Nature. 485: 55–61.
34. Malina A, Cencic R, Pelletier J. (2011) Targeting translation dependence in
cancer. Oncotarget. 2: 76–88.
35. Averous J, Proud CG (2006) When translation meets transformation: the mTOR
story. Oncogene. 25: 6423–35.
36. Gingras AC, Raught B, Gygi SP, Niedzwiecka A, Miron M, et al. (2001)
Hierarchical phosphorylation of the translation inhibitor 4E-BP1. Genes &
development 15: 2852–2864.
37. Ma L, Chen Z, Erdjument-Bromage H, Tempst P, Pandolfi PP (2005)
Phosphorylation and functional inactivation of TSC2 by Erk implications for
tuberous sclerosis and cancer pathogenesis. Cell. 121: 179–93.
38. Danckwardt S, Gantzert AS, Macher-Goeppinger S, Probst HC, Gentzel M, et
al. (2011) p38 MAPK controls prothrombin expression by regulated RNA 3’ end
processing. Molecular Cell 41(3): 298–310.
39. Mayr C, Bartel DP (2009) Widespread shortening of 3’UTRs by alternative
cleavage and polyadenylation activates oncogenes in cancer cells. Cell 138: 673–
40. Grskovic M, Javaherian A, Strulovici B, Daley GQ (2011) Induced pluripotent
stem cells–opportunities for disease modelling and drug discovery. Nat Rev Drug
Discov. 10: 915–29.
41. Ding L, Poser I, Paszkowski-Rogacz M, Buchholz F. (2012) From RNAi screens
to molecular function in embryonic stem cells. Stem Cell Rev. 8: 32–42.
42. Hu G, Kim J, Xu Q, Leng Y, Orkin SH, et al. (2009) A genome-wide RNAi
screen identifies a new transcriptional module required for self-renewal. Genes
Dev. 23: 837–48.
43. Luo J, Emanuele MJ, Li D, Creighton CJ, Schlabach MR, et al. (2009) A
genome-wide RNAi screen identifies multiple synthetic lethal interactions with
the Ras oncogene. Cell. 137: 835–48.
44. Neumann B, Held M, Liebel U, Erfle H, Rogers P, et al. (2006) High-
throughput RNAi screening by time-lapse imaging of live human cells. Nat
Methods. 3: 385–90.
45. Iizuka N, Najita L, Franzusoff A, Sarnow P (1994) Cap-dependent and cap-
independent translation by internal initiation of mRNAs in cell extracts prepared
from Saccharomyces cerevisiae. Mol Cell Biol 14: 7322–7330.
46. Thoma C, Bergamini G, Galy B, Hundsdoerfer P, Hentze MW (2004)
Enhancement of IRES-mediated translation of the c-myc and BiP mRNAs by
the poly(A) tail is independent of intact eIF4G and PABP. Molecular cell 15:
47. Pfaffl MW, Horgan GW, Dempfle L (2002) Relative expression software tool
(REST) for group-wise comparison and statistical analysis of relative expression
results in real-time PCR. Nucleic acids research 30: e36.
Cell-Based RNAi Screening with mRNA Transfection
PLOS ONE | www.plosone.org10 September 2012 | Volume 7 | Issue 9 | e45943