Article
Translational Tolerance of Mitochondrial Genes to
Metabolic Energy Stress Involves TISU and eIF1eIF4GI Cooperation in Start Codon Selection
Graphical Abstract
Authors
Hadar Sinvani, Ora Haimov, ...,
Benoit Viollet, Rivka Dikstein
Correspondence
[email protected]
In Brief
The regulatory element TISU is a
translation initiator of short 50 UTR
mRNAs. Sinvani et al. report that TISU is
enriched in mitochondrial genes and
confers resistance to translation
inhibition mediated by AMPK and energy
deprivation, allowing cells to cope with
the stress. TISU-mediated initiation
involves cap complex eviction and eIF1eIF4G1 cooperation.
Highlights
d
TISU, translation initiator of short 50 UTR, is enriched among
mitochondrial genes
d
TISU-mediated initiation involves eIF1-eIF4G1 cooperation
and cap complex eviction
d
Low energy induces a stress response that includes inhibition
of global translation
d
TISU-directed translation enables continuous translation
under energy stress
Sinvani et al., 2015, Cell Metabolism 21, 479–492
March 3, 2015 ª2015 Elsevier Inc.
http://dx.doi.org/10.1016/j.cmet.2015.02.010
Cell Metabolism
Article
Translational Tolerance of Mitochondrial Genes
to Metabolic Energy Stress Involves TISU and
eIF1-eIF4GI Cooperation in Start Codon Selection
Hadar Sinvani,1,4 Ora Haimov,1,4 Yuri Svitkin,2 Nahum Sonenberg,2 Ana Tamarkin-Ben-Harush,1 Benoit Viollet,3
and Rivka Dikstein1,*
1Department
of Biological Chemistry, The Weizmann Institute of Science, Rehovot 76100, Israel
of Biochemistry and Goodman Cancer Centre, McGill University, Montreal, QC H3A 1A3, Canada
3University Paris Descartes, Institut Cochin, 75014 Paris, France
4Co-first author
*Correspondence: [email protected]
http://dx.doi.org/10.1016/j.cmet.2015.02.010
2Department
SUMMARY
Protein synthesis is a major energy-consuming process, which is rapidly repressed upon energy stress
by AMPK. How energy deficiency affects translation
of mRNAs that cope with the stress response is
poorly understood. We found that mitochondrial
genes remain translationally active upon energy
deprivation. Surprisingly, inhibition of translation is
partially retained in AMPKa1/AMPKa2 knockout
cells. Mitochondrial mRNAs are enriched with TISU,
a translation initiator of short 50 UTR, which confers
resistance specifically to energy stress. Purified 48S
preinitiation complex is sufficient for initiation via
TISU AUG, when preceded by a short 50 UTR.
eIF1 stimulates TISU but inhibits non-TISU-directed
initiation. Remarkably, eIF4GI shares this activity
and also interacts with eIF1. Furthermore, eIF4F is
released upon 48S formation on TISU. These findings
describe a specialized translation tolerance mechanism enabling continuous translation of TISU genes
under energy stress and reveal that a key step in start
codon selection of short 50 UTR is eIF4F release.
INTRODUCTION
Intracellular energy status is critical for most biological activities
and therefore is under tight control. Perturbations in energy
metabolism are associated with disease including cancer and
the metabolic syndrome (Dazert and Hall, 2011; Laplante and
Sabatini, 2012). Protein synthesis is one of the most energyconsuming cellular processes and metabolic stresses that
reduce intracellular energy availability induce rapid inhibition of
cap-dependent translation in order to save energy consumption.
This response is mediated by the AMP-activated protein kinase
(AMPK), a highly conserved sensor of the cellular energy status
and an important target for drugs used against type 2 diabetes
(Zhang et al., 2009). AMPK stimulates ATP production pathways,
mostly by enhancing mitochondrial biogenesis (Cantó et al.,
2010; Jäger et al., 2007; Lin et al., 2005; O’Neill et al., 2011;
Winder et al., 2000). In parallel, AMPK switches off anabolic
pathways that consume ATP such as protein synthesis by inhibiting cap-dependent translation (Bolster et al., 2002; Dubbelhuis
and Meijer, 2002; Krause et al., 2002; Reiter et al., 2005). In
contrast, translation of mRNAs of proteins that are required for
the energy stress response must be maintained. Presently, the
features of cellular mRNAs and the mechanism underlying translational resistance to energy stress are largely unknown.
An important element that controls translation initiation efficiency is the AUG context (Hinnebusch, 2011). Also, translation
initiation is often regulated by the 50 UTR, in particular the length
(Kozak, 1991a, 1991b, 1991c), the presence of stem-loop structures (Kozak, 1986; Pelletier and Sonenberg, 1985), and the existence of AUG(s) upstream of the main ORF (uAUGs) (Calvo
et al., 2009; Kozak, 1984). Efficient recognition of the AUG initiation codon requires 50 UTR length of at least 20 nucleotides (Kozak, 1991b). Shorter 50 UTRs generally exhibit leaky translation
initiation (Kozak, 1991a, 1991b; Sedman et al., 1990). Previously
we identified and characterized an element called TISU (translation initiator of short 50 UTR), located between the transcription
start site up to position +30, which, remarkably, controls both
transcription and translation (Elfakess and Dikstein, 2008).
TISU is present in 4% of all genes, preferentially in those with
‘‘housekeeping’’ functions. TISU controls transcription and is
bound by the YY1 transcription factor. The core of TISU consists
of an invariable ATG sequence, which serves also as the exclusive translation initiation codon in 80% of TISU genes. In the
rest of TISU genes the AUG of the element initiates an uORF.
This initiating AUG is preceded by an unusually short 50 UTR
with a median length of 12 nucleotides (Elfakess and Dikstein,
2008). Detailed analysis of TISU established it as an element
optimized to direct efficient translation initiation from mRNAs
with a 50 UTR as short as 5 nucleotides (Elfakess and Dikstein,
2008; Elfakess et al., 2011). Besides the 3 A and the +4 G, additional flanking sequences, including positions 2 and 1 and the
nucleotides in positions +5 to +8, are unique to TISU and cooperate to direct accurate and efficient translation initiation from
very short 50 UTRs. Importantly, TISU activity is cap dependent
(Elfakess et al., 2011). These findings suggest a translation initiation mode that is cap dependent but independent of scanning
(Elfakess et al., 2011). Considering its unique features, it is
Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc. 479
conceivable that TISU is differentially regulated under specific
physiological settings.
The process of translation initiation occurs in several mechanistic steps (for review, see Hinnebusch, 2011; Jackson et al.,
2010; Sonenberg and Hinnebusch, 2009) that include the formation of a 43S preinitiation complex comprising the 40S subunit,
eIF1, eIF1A, eIF3 eIF2-GTP-Met-tRNAMeti, and probably eIF5.
This complex is then recruited to the mRNA through the capbound eIF4F. Subsequently the 43S complex scans the 50 UTR
as an ‘‘open’’ complex until an initiation codon is encountered.
Initiation complex formation involves a switch of the scanning
conformation to a ‘‘closed’’ one followed by eIF5-mediated
release of eIF2-bound Pi along with eIF1 (Cheung et al., 2007;
Maag et al., 2005; Martin-Marcos et al., 2014). Initiation is then
completed by the joining of the 60S subunit to the 48S complex
and concomitant displacement of eIF2-GDP and other factors
(eIF3, eIF4B, eIF4F, and eIF5). Studies in yeast unraveled the critical roles that eIF1 plays in ternary complex recruitment and in
discrimination against near-cognate AUG start codons recognition (Cheung et al., 2007; Martin-Marcos et al., 2011, 2013;
Mitchell and Lorsch, 2008; Nanda et al., 2009). Consistent with
the high conservation of eIF1, it was shown to inhibit initiation
from suboptimal AUG contexts in higher eukaryotes as well (Elfakess et al., 2011; Ivanov et al., 2010; Pestova and Kolupaeva,
2002). Moreover, canonical cap-dependent translation imposes
a seemingly strong constraint on TISU-mediated translation by
positioning eIF4F and the AUG-48S initiation complex in an overlapping location. To what extent the mechanistic events associated with the initiation process directed by TISU resemble or
differ those of the canonical pathway is presently unknown.
Here we investigated the molecular mechanism and the physiological importance of TISU-mediated translation initiation. Our
findings revealed that eIF1 and eIF4GI cooperate to facilitate
TISU-mediated translation initiation. We also demonstrate that
a key step in start codon selection of short 50 UTR mRNAs is
eIF4F release. The unique mechanism directed by TISU operates
in energy-producing genes and confers resistance to inhibition
of cap-dependent translation in response to energy stress.
RESULTS
Sequence-Specific Recognition of TISU by the Basal
Translation Apparatus
When an initiator AUG is preceded by a very short 50 UTR,
it is frequently bypassed by the 48S ribosome, resulting in
leaky translation initiation (Elfakess and Dikstein, 2008; Kozak,
1991b). An exceptional paradigm is an AUG in the context of
the TISU sequence (Elfakess and Dikstein, 2008; Elfakess
et al., 2011). We analyzed translation initiation from an AUG in
either a canonical Kozak context or the TISU element, preceded
by a very short 50 UTR. Several cell lines were transfected with
in vitro synthesized GFP mRNA (Figures 1A and S1) in which
translation from the upstream AUG yields a 30 kDa protein, while
leaky translation initiation from an in-frame downstream AUG
yields a 27 kDa protein. Translation from the canonical upstream
AUG varied significantly between different cell types, constituting approximately 45%, 20%, and 15% of total translation in
293T, HeLa, and MEFs cells, respectively, while being undetectable in HepG2 cells (Figure 1A, right lanes and graph). In stark
480 Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc.
contrast, TISU-mediated translation started almost exclusively
from the upstream AUG in all the cell lines (Figure 1A, left lanes
and graph). These findings suggest that leaky translation initiation is highly variable and has a potential to generate alternative
protein products in a cell-type-dependent manner. On the other
hand, accurate translation initiation directed by TISU is robust
and is unaffected by differences in the cellular milieu.
The ability of TISU to confer translation initiation fidelity in
different cellular environments raises the question of whether
the basal translation machinery has intrinsic capability to initiate
from an AUG in a TISU context. To address this issue, we examined whether TISU-mediated initiation can be reconstituted
in vitro only with basal components of the translation apparatus.
We used a translation initiation system containing initiator MettRNAi; purified 40S ribosomal subunit; purified eIFs that include
eIF1, eIF1A, eIF2, eIF3, eIF4F, and ATP; and GMP-PNP, a nonhydrolysable GTP analog that stalls the 48S ribosomal complex
on the initiation site. The formation of the 48S complex at the initiator AUG was detected by inhibition of primer extension with
reverse transcriptase due to ribosome binding (toe-printing) (Kozak, 1998). In the presence of all the components, the short 50
UTR mRNA bearing the TISU element generated a strong toeprint 15 nt downstream of the AUG, which corresponds to
the position of the leading edge of the small ribosomal subunit
(Figure 1B, compare lane 1 to lane 2). In contrast, ribosome binding directed by the canonical AUG element was very weak and
could be observed only after a prolonged exposure (Figure 1B,
lanes 7 and 8). This result demonstrates that the TISU sequence
is sufficient for efficient AUG recognition by the translation machinery and that no additional cellular factor is required for the
basal activity of TISU.
To determine the minimal set of translation initiation factors,
which are required for TISU-containing mRNAs, we performed
toe-print reactions in the absence of individual factors (Figure 1B). As expected, TISU-mediated translation initiation is
stimulated by eIF2 (lane 16), initiator tRNA (lane 17), eIF3 (lane
15), and eIF4F (lane 3). Reduction of the 48S complex formation
by eIF4F and eIF3 omission was not complete, most probably
due to cross-contamination of these factors (Méthot et al.,
1996). Interestingly, removal of eIF1 decreased initiation from
the TISU AUG, which was unexpected, considering the previous
studies demonstrating that eIF1 inhibits translation from an AUG
preceded by a short 50 UTR (Cheung et al., 2007; Mitchell and
Lorsch, 2008; Nanda et al., 2009; Pestova et al., 1998; Pestova
and Kolupaeva, 2002). Exclusion of both eIF1 and eIF1A resulted
in loss of translation initiation (Figure 1B, lanes 6 and 12). This is
consistent with a previous report on the redundant activities of
the two factors in vitro (Pestova et al., 1998). Interestingly, omission of both eIF4F and eIF1 had much greater effect on 48S
formation than omission of eIF1A and eIF4F (Figure 1C).
eIF1 and eIF4GI Cooperate to Control Translation
Initiation Fidelity and Ribosomal Scanning
The differential effects of eIFs on TISU and non-TISU AUG in vitro
prompted us to examine the relevance of these findings to translation in vivo. We first knocked down eIF1 using siRNA, and 48 hr
later cells were transfected with GFP reporter constructs containing either a TISU or a canonical AUG context preceded by
a short 50 UTR (US AUG) followed by an in-frame downstream
TISU/Canonical
A
m7G
5 nt
AUG
AUG
US
TISU
Figure 1. TISU-Directed Translation Initiation in Cells and In Vitro
GFP
DS
Canonical
US AUG
DS AUG
MEFs
US AUG
HepG2
1.2
US AUG
1
DS AUG
0.8
DS AUG
0.6
US AUG
0.4
293T
DS AUG
0.2
HeLa
US AUG
0
DS AUG
B
TISU
Canonical
Canonical
TISU
TISU
TCAG
AUG48S-
-48S
1
2
3
4
5 6
7
8
13 14 15 16 17
9 10 11 12
-48S
C
Long exposure
(A) TISU is refractory to cellular variations in translation initiation fidelity. Short 50 UTR mRNAs with
either TISU or a canonical AUG context (ACCATGG),
as shown schematically on the top, were in vitro
transcribed and capped and then transfected into
the indicated cell lines. The quality of the mRNA
used for transfection and their relative levels within
the cell are shown in Figures S1A and S1B. Translation was assayed by western blot analysis with
GFP antibodies. Quantified results of translation
from the upstream (US) AUG and downstream (DS)
AUG are shown.
(B) TISU-directed initiation can be reconstituted by
purified components of the 48S ribosome and is
stimulated by eIF1. Toe-printing analysis of 48S
complexes assembled on AUG either in a TISU
(lanes 1–6 and 13–17) or canonical (lanes 7–12)
contexts preceded by 5 nt 50 UTR, with purified
components. Reactions were carried out without
translation factors (lanes 1, 7, and 13), with all
translation initiation factors (lanes 2, 8, and 14),
without eIF4F (lanes 3 and 9), without eIF1 (lanes 4
and 10), without eIF1A (lanes 5 and 11), without
eIF1 and eIF1A (lanes 6 and 12), without eIF3 (lane
15), without eIF2 (lane 16), and without initiator
tRNA (lane 17). The positions of the 48S and the
AUG initiation codon are marked.
(C) Cooperation of eIF1 and eIF4F in TISU-mediated initiation. Toe-printing analysis as in (B) with
TISU mRNA with all initiation factors (lane 1),
without eIF4F (lane 2), without eIF4F and eIF1 (lane
3), and without eIF4F and eIF1A (lane 4). The
control lane ‘‘All’’ was spliced together with lane
‘‘G.’’
T C A G
-48S
1
2
3
4
AUG (DS AUG), or by a long unstructured 50 UTR as shown schematically in Figure S2 (available online). With a knockdown (KD)
efficiency of 60% of eIF1 protein, TISU US-AUG level was
reduced by 35%, consistent with the in vitro results, while
the canonical US-AUG was slightly increased (20%) (Figure S2B). Translation of the mRNA containing the long 50 UTR
was marginally affected, which is unexpected, considering that
eIF1 promotes ribosomal scanning (Hinnebusch, 2011; Mitchell
and Lorsch, 2008). We found in the human genome database
(http://genome.ucsc.edu/) that eIF1 has a highly similar paralog
called eIF1B that is expressed as abundantly as eIF1. We therefore knocked down either eIF1B alone or the two human paralogs of eIF1 together, and 48 hr later expressed the GFP reporter
genes described above. With the eIF1B siRNA treatment, eIF1
levels were barely changed and this was nevertheless associated with a modest differential effect on TISU and canonical
AUG (Figure S2C), as seen with the KD of eIF1. Upon eIF1 and
eIF1B double siRNA transfection, the KD efficiency was
increased and led to a clear and more dramatic differential effect
(Figure 2A). The ratio of the canonical USAUG to the DS-AUG was increased, suggesting that leaky scanning was inhibited.
In addition, translation from the long 50
UTR which is dependent on ribosomal
scanning was reduced, consistent with
the role of eIF1 in scanning (Pestova and Kolupaeva, 2002). In
contrast to the canonical US-AUG, TISU US-AUG levels were
greatly diminished, without concomitant increase of leaky translation. These findings bolster the idea that eIF1 facilitates TISUmediated translation and is a major player in its potency.
We also examined the roles of eIF4GI and eIF3c, subunits of
eIF4F and eIF3, respectively, on translation initiation. Strikingly,
eIF4GI depletion caused the same phenotype as eIF1 KD with
respect to the different AUG contexts and 50 UTR lengths (Figure 2B). Specifically, in eIF4GI-deficient cells, the relative translation of the canonical US-AUG to the DS-AUG was increased
while TISU US-AUG was almost abolished. Similar to eIF1,
eIF4GI depletion caused the reduction of translation from the unstructured long 50 UTR, indicating its role in promoting ribosomal
scanning. To examine whether the similarity of the KD effects of
eIF4GI and eIF1 is associated with indirect effect on eIF1 expression, we analyzed eIF1 levels in eIF4GI KD cells and found that its
expression was unaffected (Figure 2B). Downregulation of eIF3c,
on the other hand, diminished translation irrespective of AUG
Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc. 481
A
B
Figure 2. Effects of eIF1/eIF1B and eIF4GI
KD on Translation Initiation Fidelity
The GFP reporter genes used for the experiments
in (A) and (B) are shown schematically in Figure S2A and include two short 50 UTRs with an
AUG in a context of either TISU or canonical AUG
that is in-frame with the downstream GFP AUG;
the third consists of a long unstructured 50 UTR
preceding the native GFP AUG.
(A and B) HEK293T cells were transfected with a
mixture of eIF1 and eIF1B siRNA (A) or siRNA
against eIF4GI (B). As a control, a nontargeting
siRNA was transfected in each of the experiments.
After 48 hr cells were transfected again with the
GFP reporter genes described above. Cells were
harvested 24 hr after the second transfection and
analyzed by western blot with GFP antibody.
Representative immunoblots are shown. US AUG
and DS AUG denote upstream and downstream
initiation site, respectively. The KD efficiency is
shown on the right panel of each section. The
effect of eIF4GI KD on eIF1 levels is also shown
(B). The graphs represent the average ± SEM of
densitometric measurements of GFP levels of
three experiments. The translation in control cells
was set to 1. Light and dark gray bars represent
the relative translation from the US AUG and DS
AUG, respectively. All differences are statistically
significant p < 0.05.
(C) eIF4GI interacts with eIF1. HEK293T cells were
transfected with plasmids directing expression of
either HA-tagged eIF1 or HA-eIF1A. After 48 hr
cells were lysed and subjected to immunoprecipitation assays using either HA antibody, control
antibody. The immune complexes were then run
on SDS-PAGE followed by western blot with the
indicated antibodies.
C
context and 50 UTR length (Figure S2D). These findings raise the
possibility that eIF1 and eIF4GI functionally cooperate to facilitate translation initiation fidelity of TISU and non-TISU mRNAs
as well as ribosomal scanning. This possibility is supported by
the observation that omission of both eIF4F and eIF1 cooperate
for efficient TISU-mediated translation in vitro (Figure 1C).
To assess the effects of eIF1/eIF1B, eIF4GI, and eIF3c KD on
global translation, we subjected cell lysates from siRNA-treated
cells to polysome profiling. While eIF1/eIF1B double KD did not
cause visible effect on global translation, depletion of either eIF4GI
or eIF3c resulted in accumulation of 80S monoribosome and a
decrease in polysomes, indicating for global inhibition of translation initiation (Figure S3). To examine the influence of initiation
factor depletion on translation of specific genes, we pooled the
482 Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc.
fractions from the free, the light, and the
heavy polysomes; extracted their RNA;
and performed RT-qPCR analysis of six
non-TISU and eight TISU genes. KD of
eIF1/eIF1B and eIF3c caused modest
reduction of translation of most TISU and
non-TISU genes (Figure S3). The mild
effects of eIF1 and eIF3c depletion may
be a consequence of insufficient KD or
compensation by other factors with redundant activities such as eIF1A (see Figure 1B). Downregulation of
eIF4GI had a more notable effect on translation of all the examined
genes, consistent with its general role in translation (Figure S3).
To investigate the underlying basis for the shared activities of
eIF1 and eIF4GI on initiation site selection, we examined whether
they can interact with each other as shown for the yeast homologs (Asano et al., 2000; He et al., 2003; Valásek et al., 2002).
Initial coimmunoprecipitation (coIP) experiments with endogenous proteins failed to reveal an association between the two
proteins. Considering that such an interaction may be transient,
we carried out coIP following overexpression of HA-tagged
eIF1 or, as a control, HA-eIF1A. The results revealed efficient
coprecipitation of endogenous eIF4GI with HA-eIF1, but not
with HA-eIF1A or control antibodies (Figure 2C).
A
Figure 3. Dissociation of eIF4F from the 48S
Complex upon AUG Recognition
(A) Capped mRNAs, bearing either short 50 UTR
and TISU or long 50 UTR with a strong AUG
context, were incubated with RRL in the presence
of GMP-PNP or cycloheximide to stall the 48S
(left) or 80S (right), respectively. Ribosomal-mRNA
complexes were then subjected to 8%–32% sucrose gradient sedimentation, and the collected
fractions were recorded (top panel) and analyzed
by western blot for the distribution of eIFs and ribosomal subunits as indicated (bottom panels). A
similar experiment with short 50 UTR and canonical
AUG is shown in Figure S5.
(B) A model describing TISU-mediated translation
initiation. The 43S complex is recruited by the
cap-binding complex and is followed by an immediate encounter with an AUG and the subsequent joining of the 60S subunit. This process
involves the detachment of eIF4F from the cap to
allow proper engagement of the AUG at the P site.
Initiation is suggested to be facilitated by eIF1eIF4GI interaction.
B
To determine whether the eIF1-eIF4GI interaction is direct, we
carried out GST pull-down assays, in which eIF1 and eIF1A were
fused to GST, expressed in bacteria, purified, and subjected to
in vitro binding reaction with recombinant His-tagged eIF4GI
fragments (Figure S4A). The results revealed an interaction
between eIF1 and the middle domain of eIF4GI (amino acids
674–1,129) and a weak interaction with the most C-terminal region (amino acids 1,130–1,599) (Figure S4A, left). These interactions are highly specific, as similar amounts of GST or GST-eIF1A
(Figure S4A, right) failed to interact with any of the eIF4GI fragments. We also carried the reciprocal GST pull-down assay (Figure S4B), which confirmed that the central region, but not to the
C-terminal domain, is the major eIF1-binding site. These results
suggest that mammalian eIF4GI specifically interacts with eIF1.
Release of eIF4F Following Initiation Complex Assembly
on TISU
Considering the short length of the 50 UTR of TISU genes, both
the cap and the 50 UTR nucleotides are expected to occupy
the exit channel of the ribosome. On the
other hand, translation directed by TISU
is highly eIF4F and cap dependent, suggesting that the positions of eIF4F and
the ribosome on the mRNA are overlapping. To elucidate this incompatibility,
we considered the previous observation
that in yeast eIF1 is released upon AUG
recognition (Cheung et al., 2007; MartinMarcos et al., 2011, 2013; Mitchell and
Lorsch, 2008; Nanda et al., 2009) and
raised the hypothesis that upon assembly
of the ribosome on TISU’s AUG the cap
complex dissociates from the mRNA
together with eIF1 to allow the positioning
of the short 50 UTR in the ribosomal
exit channel. To test this possibility, we
analyzed the 48S and 80S complexes with two mRNAs, one
bearing a 5 nt long 50 UTR under the control of TISU and the other
with a long and unstructured 50 UTR with a canonical AUG
context. These mRNAs were generated without poly(A) to prevent retention of eIFs on the mRNA through PABP. RibosomalmRNA complexes were formed with rabbit reticulocyte lysates
(RRLs) in the presence of either nonhydrolysable GTP analog
(GMP-PNP) or cycloheximide, which cause accumulation
of either 48S or 80S ribosomal complexes, respectively. The
complexes were then resolved on sucrose gradients, and the
collected fractions were monitored and then analyzed by western blot. With both mRNAs formation of the 48S complex is
evident from the pattern of the gradient and the presence of
the small ribosomal subunit (Rps5) in the peak fractions 5 and
6 (Figure 3A, left). Likewise, eIF2a is associated with the 48S
fractions 5 and 6 in both mRNAs. Interestingly, with TISU
mRNA the patterns of eIF4GI and eIF4E are highly similar,
showing a sharp reduction in the 48S peak fraction 6 and enrichment in the top fraction 2. Likewise, eIF1 is diminished from 48S
Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc. 483
fraction 6 and is detected in top fraction 2 with TISU mRNA. With
the long 50 UTR mRNA, the amounts of eIF4GI, eIF4E, and eIF1 in
the top fraction 2 are much lower. For both mRNAs eIF3d is
concentrated in 48S fractions 5–7 (Figure 3A, left). A similar analysis of a short 50 UTR mRNA with a canonical AUG context
resulted in eIF4F release, reminiscent of TISU (Figure S5A), supporting the notion that initiation from an AUG close to the 50 end
of an mRNA is associated with ejection of the cap complex (Figure 3B). The analysis of the 80S complex (Figure 3A, right) revealed the expected presence of the small and large ribosomal
subunits in the peak fractions 6–8. On the other hand, for both
mRNAs, all the initiation factors are shifted to the top of the
gradient, suggesting their release upon 80S formation (Figures
3A and 3B).
TISU Confers Resistance to Translation Inhibition
Imposed by AMPK Agonists
The mechanistic studies of TISU revealed a unique mode of
translation initiation directed by the basal machinery that is
strictly dependent on the AUG context. We next wished to
explore the possible physiological importance of this mechanism. As a first step, we analyzed the functional classes associated with the TISU genes in which their ORF begins from the
element’s AUG. One of the most enriched functional classes is
the mitochondrion (Elfakess and Dikstein, 2008). Within this
class we found a significant enrichment of genes involved in
mitochondrial activities and energy metabolism (Table S1, Table
S2, and Table S3). As translation of exogenous and endogenous
genes governed by TISU is highly eIF4GI dependent (Figures 2
and S3), it is expected that eIF4GI depletion would affect energy
homeostasis. To test this possibility, we measured cellular ATP
levels in cells treated with control or eIF4GI siRNA, using luminescent assay. The data, after normalization to cell number,
revealed a significant reduction in endogenous ATP levels
upon eIF4GI KD (Figure 4A). With eIF1/eIF1B KD no change in
ATP levels was observed (data not shown), which is in line with
its modest effect on translation of endogenous TISU genes (Figure S3). A possible explanation is the functional redundancy of
eIF1 and eIF1A as shown in Figure 1B.
Among TISU genes are the two catalytic subunits of AMPK,
a highly conserved sensor of the cellular energy status (Zhang
et al., 2009). AMPK stimulates ATP production pathways, mostly
by enhancing oxidative metabolism and mitochondrial biogenesis (Cantó et al., 2010; Jäger et al., 2007; Lin et al., 2005; O’Neill
et al., 2011; Winder et al., 2000), while switching off anabolic
pathways that consume ATP, in particular protein synthesis
through interference with cap-dependent translation (Bolster
et al., 2002; Dubbelhuis and Meijer, 2002; Krause et al., 2002;
Reiter et al., 2005). Taking into account that many TISU genes
are required for the proper response to energy stress and that
TISU-mediated translation occurs without ribosomal scanning
(Elfakess et al., 2011), an ATP-consuming process, we examined
TISU activity under conditions of low energy availability. As a first
step we treated HepG2 cells with the AMPK agonist 5-aminoimidazole-4-carboxamide-1-b-D-ribofuranoside (AICAR) and then
analyzed global translation using polysome profiling. The results
confirmed that the energy stress mimic exerted by AICAR
induced AMPK phosphorylation and phosphorylation of its
downstream substrate acetyl-CoA carboxylase (ACC) (Figure 4B)
484 Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc.
and inhibited global translation, as evident from the reduction in
the relative amount of polysomes and the increase of the 80S
monoribosome (Figures 4C and S5B). To examine the impact
of AICAR on translation of specific genes, fractions from the
free, the light, and the heavy polysomes were pooled; RNA
was prepared and subjected to RT-qPCR analysis of six nonTISU and 19 TISU genes (Figure 4D). Consistent with the global
inhibitory effect of AICAR on translation, we found that translation of most non-TISU mRNAs was inhibited, as the amount of
mRNA was shifted from the polysomal fractions to the free fractions. In contrast, the free and the polysomal association of all
tested TISU mRNAs, including the AMPK catalytic subunits,
were unaffected by AICAR. We examined the effect of AICAR
on genes involved in fatty acid oxidation (FAO) in the mitochondria, which are also involved in energy production but are
non-TISU. We found that their translation is unaffected as well
(Figure 4D).
The resistance to AICAR treatment observed for TISU mRNAs
can be mediated either by the TISU element or by another unknown feature present in the endogenous TISU mRNAs tested
above. To examine the importance of the TISU sequence itself
in the translational activity under conditions of induced AMPK,
we transfected into HepG2 cells in vitro synthesized GFP
mRNA with a 5 nt long 50 UTR under the control of TISU (Figure 5A). Cells were treated with two AMPK agonists AICAR
and metformin and the AMPK antagonist Compound C. As a
mRNA with 5 nt long 50 UTR followed by canonical AUG context
was inactive in these cells (Figure 1A), we used as a control a
GFP reporter under the control of the native Rpl18 sequences,
which also have short 50 UTR and a canonical AUG context.
TISU-directed translation was either enhanced or unchanged
by AICAR and metformin, respectively, while Compound C
diminished TISU activity in the presence or absence of these
drugs (Figure 5A). In contrast, both AICAR and metformin
decreased translation directed by Rpl18 sequences while Compound C partially restored AICAR-mediated inhibition (Figure 5A).
We validated that both AICAR and metformin induced the
phosphorylation of AMPK (Figure 5B), thus causing its activation.
We also confirmed that the translational activity of TISU under
these conditions remains cap dependent by transfecting into
cells TISU mRNAs containing an unmethylated cap analog,
which was inactive in directing translation in the absence or
presence of AICAR (Figure 5C).
TISU mRNAs Remain Translationally Active under
Metabolic Energy Deficiency
Having shown that TISU-mediated translation initiation is resistant to the inhibition conferred by AMPK inducers, we next
wished to investigate the direct effect of the cellular energy status on its activity. For these experiments we used differentiated
mouse C2 myotubes that represent muscle cells (Figure S6) and
mouse embryonic fibroblasts (MEFs) (Figure 6). Cells were
grown in the presence or absence of glucose and then subjected
to analysis of global translation using polysome profiling. The results confirmed that low availability of metabolic energy induced
AMPK and ACC phosphorylation (Figures 6C and S6C) and
significantly reduced the relative amount of polysomes while
increasing the 80S monoribosomes (Figures 6A and S6B). We
next examined the effect of this stress on a subset of non-TISU
A
B
C
D
Figure 4. The Effect of eIF4GI Depletion on Intracellular Metabolic Energy and the Translational Response to AMPK Activation
(A) HEK293T cells in 96-well plate were transfected with nontargeting or eIF4GI siRNA. After 72 hr, cellular ATP levels were measured by luminescent assay. The
graph represents the average ± SEM of three independent experiments.
(B) Western blot analysis of phosphorylated and total AMPK and ACC (p-AMPK/AMPK and p-ACC/ACC, respectively) following treatment of 2 mM AICAR for 3 hr.
(C) Polysomal profiling of HepG2 cells treated (red) or untreated (black) with AICAR for 3 hr.
(D) RT-qPCR analysis of the relative levels of the indicated mRNAs (non TISU, TISU, and fatty acid oxidation) in the polysome-free fractions of the gradient (F), light
polysomal fractions (L), and heavy polysomal fractions (H).
and TISU mouse genes by RT-qPCR as described above. With
the non-TISU mRNAs (Rps18, Rpl18, Gapdh, and Eef2) we
observed an increase in the free fractions and a concomitant
decrease in the polysomal fractions upon glucose starvation in
the two cell types (Figures 6D and S6D). Remarkably, translation
of all the 15 analyzed TISU mRNAs was unaffected by the
glucose starvation, both in the myotubes (Figure S6D) and in
MEFs (Figure 6D), indicating that their resistance to this stress
in not cell type specific. Likewise, translation of fatty acid oxidation genes, which are non-TISU, was unchanged by this stress
(Figure S6E).
Inhibition of translation upon energy stress is thought to be
mediated by AMPK (Bolster et al., 2002; Dubbelhuis and Meijer,
2002; Krause et al., 2002; Reiter et al., 2005), which is activated
by this stress (Figures 6C and S6C) and induces 4E-BP dephosphorylation (Figure 7E). We therefore examined the direct
involvement of AMPK by performing a similar experiment in
MEFs derived from a mouse deficient for the two catalytic subunits of AMPK (AMPKa1 and AMPKa2, AMPK DKO). Surprisingly, we found that glucose deprivation halted translation in
the DKO cells as well (Figure 6B). Furthermore, even in the
absence of AMPK, the effect on 4EBP dephosphrylation was retained (Figure 7E). We determined the level of inhibition by calculating the ratio between the RNA in the polysomal and the free
fractions (P/F). While the energy stress decreased this ratio by
27.5% ± 0.5% in the WT MEFs, the decrease in the AMPK
DKO MEFs is 8% ± 1.7%. These findings suggest that the contribution of AMPK to inhibition of translation by energy stress is significant, but part of the translational response to this stress is
clearly AMPK independent. Analysis of translation of specific
TISU and non-TISU genes revealed that the effect of the glucose
deprivation is quite similar to the WT MEFs, and TISU genes are
refractory to this stress in the AMPK DKO cells as well
(Figure S7).
The resistance of all tested endogenous TISU genes to the
strong translational inhibition caused by energy stress may
involve either the TISU element or other regulatory elements
associated with these genes. To examine the role of TISU
more directly, we used GFP reporters, either as mRNAs that
were in vitro synthesized or a plasmid. In these reporters the
Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc. 485
A
B
C
Figure 5. TISU-Mediated Initiation Is Refractory to AMPK-Mediated Translation Inhibition
(A) HepG2 cells were transfected with in vitro-synthesized mRNA containing TISU or a construct bearing the Rpl18 sequences (promoter, 50 UTR, and initiating
AUG) upstream to GFP. Cells were then incubated with 2 mM AICAR or 2 mM metformin (MTF) with or without 20 mM Compound C (CC) for 6 hr. Translation of
GFP was analyzed by western blot. Quantified results of four to six experiments (average ± SEM) and representative blots are shown. *p < 0.05, ***p < 0.0005.
(B) Western blot analysis of phosphorylated and total AMPK.
(C) mRNAs bearing TISU were in vitro transcribed in the presence of m7GpppG cap analog or the unmethylated cap analog ApppG. The mRNAs were transfected
into HepG2 cells and then treated with AICAR as described. Translation of GFP was analyzed by western blot using GFP antibodies.
GFP is driven by different 50 UTRs and AUG contexts as schematically shown (Figure 7A). These include a long and unstructured 50 UTR (mRNA), a short 50 UTR and AUG of the Rpl18
gene (plasmid), and a 5 nt long 50 UTR that is followed by an
AUG either in canonical or TISU contexts (mRNAs) which are followed by an in-frame downstream AUG. The reporter mRNAs or
plasmid was transfected into the MEFs, which were then subjected to glucose starvation. Translation directed by the long 50
UTR, the short 50 UTR with the Rpl18 sequences, and the canonical US-AUG was inhibited by reduced energy availability in the
WT MEFs (Figure 7B). In contrast, the reporter driven by TISU
continued to be translated in spite of the energy deficiency.
This result indicates that the TISU AUG context rather than the
50 UTR length confers the energy stress tolerance. The effect
of glucose starvation on translation is largely mediated by the en486 Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc.
ergy sensor AMPK (Figures 6A and 6B), which is activated by this
stress (Figure 6C). Therefore we carried out similar experiments
in AMPK DKO MEFs. In these cells the inhibitory effect of energy
stress on translation of all the reporters was reduced (Figure 7C),
indicating that AMPK mediates the translational inhibition of the
different reporters. We examined the effect of the energy stress
on 4EBP dephosphrylation in the WT and the AMPK DKO MEFs
and found that the effect exerted by glucose starvation of 4EBP
is comparable, suggesting that the stronger translational inhibition observed in the WT MEFs involves an additional, unknown
component.
We also determined the effect of rapamycin on translation
directed by the different reporters and found that it inhibited
translation directed by all short 50 UTR reporters including
Rpl18, canonical AUG, and TISU (Figure 7D). The reporter
A
B
C
D
Figure 6. Translational Response of WT and AMPKa1/AMPKa2 DKO-MEFs to Energy Stress
(A and B) Ribosomal profiling of WT (A) and AMPK DKO (B) MEFs in response to 8 hr glucose starvation (GS, red).
(C) Western blot analysis of phosphorylated and total AMPK and ACC under glucose starvation (GS) in WT MEFs. The bottom panel shows an AMPK western blot
of cell lysate from WT and AMPK DKO MEFs as indicated.
(D) Real-time qPCR analysis of the levels (average ± SEM) of the indicated non-TISU and TISU mRNAs from WT MEFs in the (F) polysome-free fractions of the
gradient, (L) light polysomes fractions, and (H) heavy polysome fractions. A similar analysis in the AMPK DKO MEFs is shown in Figure S7.
driven by the long and unstructured 50 UTR was refractory to
20 nM rapamycin (Figure 7D) as well as 100 nM (data not
shown). Interestingly, the effect of rapamycin on 4EBP dephosphorylation is more pronounced than that conferred by the
energy stress (Figure 7E). These findings suggest that short 50
UTR mRNAs are highly sensitive to severe mTOR inhibition.
On the other hand, unstructured 50 UTR mRNAs are resistant
to mTOR inhibition. Taken together, the results with the different
reporters unravel the importance of the 50 UTR length and AUG
contexts for the differential translational response to various
stresses.
DISCUSSION
Translation is one of the most energy-consuming cellular processes and therefore is a major target for inhibition in response
Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc. 487
A
E
B
C
D
Figure 7. The Differential Translational Response to Energy Stress and Rapamycin Involves AUG Context and 50 UTR Length, Respectively
(A) A scheme of the reporter mRNAs used for transfection.
(B and C) Wild-type MEFs (B) or AMPKa1/2 DKO-MEFs (C) were transfected with each of the mRNAs shown in (A) and incubated overnight with (CT) or without
glucose (GS). Translation of GFP was analyzed by western blot. Quantified results of five to six experiments (average ± SEM) and representative blots are shown.
(D) WT MEFs were transfected with each of the constructs shown in (A) and incubated overnight with 20 nM rapamycin. Quantified results of four experiments
(average ± SEM) and representative blots are shown. In all graphs, *p < 0.05, **p < 0.005, and ***p < 0.0005.
(E) Western blot analysis of 4EBP under glucose starvation (GS) and rapamycin in WT MEFs and GS in AMPK DKO MEFs.
488 Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc.
to various cellular stresses (Sonenberg and Hinnebusch, 2009).
However, under conditions in which global protein synthesis is
impaired, some proteins must continue to be synthesized as
part of the adaptation to the stress. In many cases the molecular details underlying the mechanism used by cellular mRNAs
to bypass the inhibition of translation imposed by various
stresses are largely unknown. In the present study we discover
a tolerance mechanism that selectively operates under the
stress caused by reduced energy levels. Our findings reveal
that genes and pathways that are specifically needed for the
cellular response to energy stress such as mitochondrial enzymes, factors involved in ATP production and fatty acid metabolism, remain translationally active when global translation is
inhibited. The strategy adopted by a subset of these genes to
counteract the energy-stress-mediated inhibition involves their
translational control by the TISU element. We demonstrate
that initiation directed by this sequence is sufficient to confer
resistance to inhibition of translation mediated by AMPK and
energy stress. This tolerance mechanism is specific to this
stress, as the inhibition by other stresses such as rapamycin
and amino acid deprivation is maintained (this study and Elfakess et al. [2011]). Consistent with that, we also examined the
data from a genome-wide analysis of translational response
to metformin, an inducer of AMPK (Larsson et al., 2012), and
found that the vast majority of TISU genes were unaffected
by this drug.
There are several mechanisms for suppression of the general
translation initiation during stress. These include the well-characterized impairment of initiation factors eIF2 and eIF4E. While
eIF2 does not respond to energy deprivation (data not shown),
eIF4E activity is suppressed through 4EBP activation in a
manner dependent and independent of AMPK (Figure 7E), in
accordance with previous studies (Inoki et al., 2003; Kalender
et al., 2010). Considering that TISU activity is cap dependent,
we suggest that it has a capability to utilize eIF4E despite the
reduction in its availability (Figure 5C). This is possible because
the extent of eIF4E inhibition by the energy stress is smaller than
the effect exerted by rapamycin (Figure 7E). This feature is specific to TISU, as the non-TISU AUGs are clearly sensitive to this
level of eIF4E inhibition. On the other hand, the translational
repression exerted by glucose deprivation is comparable to
that of rapamycin, suggesting that this stress impacts another,
yet-unknown translation initiation factor, to which TISU is also
resistant. The activity of TISU under these conditions is also explained by its ability to strongly interact with the basal translation machinery and its unique reliance on eIF1 and eIF4GI, all
of which are TISU-sequence specific. Hence the biochemical
features of TISU are well correlated with its activity during energy stress. It is noteworthy that during the course of the energy
stress carried out here we did not observe changes in the relative amounts of eIFs. We did, however, see an increase in the
levels of initiation factors in the ribosomal free fractions, as expected from inhibition of initiation (Figure S5B). Thus it appears
that, by and large, the intrinsic properties of TISU, rather than
alteration in translation initiation factors, are central to this tolerance mechanism. Taken together, our findings revealed that
genes that encode proteins which function to balance the
cellular energy needs have evolved mechanisms to promote
enhanced mRNA binding to the 48S ribosomal preinitiation
complex to overcome 4E-BP activation, thereby ensuring
continuous protein synthesis.
The results of this study also uncover a pivotal role of eIF4GI
in translation initiation fidelity in mammalian cells. eIF4GI is a
large scaffolding subunit of the eIF4F complex, containing binding sites for many components of the translation apparatus
including eIF4E, eIF4A, and eIF3 (Imataka et al., 1998; Imataka
and Sonenberg, 1997; Korneeva et al., 2000; Mader et al., 1995;
Marcotrigiano et al., 2001; Villa et al., 2013; Yanagiya et al.,
2009). By the use of a set of reporters with different 50 UTR
lengths and AUG contexts along with eIF4GI KD experiments,
we found that eIF4GI promotes leaky translation initiation
when the AUG is too close to the cap, while facilitating TISUmediated translation as well as ribosomal scanning. All these
features are common to those attributed to eIF1. Importantly,
these factors can directly interact with each other, suggesting
that their physical association facilitates translation initiation
fidelity.
How does eIF1-eIF4GI interaction promote initiation by TISU?
Our previous findings, as well as the data presented here,
strongly argue for the cap-dependent nature of initiation
directed by TISU. Taking into account the short length of the
50 UTR of TISU mRNAs, it is expected that both the cap and
the 50 UTR nucleotides will be within the ribosomal exit channel
upon 80S formation. This situation is incompatible with
continued association of eIF4F with the mRNA through the
cap. The analysis of the fate of translation initiation factors in
the 48S and 80S TISU ribosomal complexes reported here supports a model (Figure 3B) in which the cap-binding complex is
evicted from the mRNA 50 end upon 48S initiation complex
formation. We propose that eIF4GI is dissociated along with a
fraction of eIF1, as a consequence of their physical association
allowing proper initiation (Figure 3B).
In addition to its role in translation TISU is also a critical
element for transcription (Elfakess and Dikstein, 2008) raising
the possibility that the transcriptional and the translational activities of TISU are coordinated. TISU binds the YY1 transcription
factor (Elfakess and Dikstein, 2008). Our attempts to address
the role of YY1 for mitochondrial gene expression in HepG2
cells and in MEFs were unsuccessful. However, it was recently
reported that targeted knockout of YY1 in skeletal muscle resulted in reduced expression of mitochondrial genes including
many TISU genes (Blättler et al., 2012). In addition, YY1 is
required for the inhibition of these genes by rapamycin at the
mRNA level, through disruption of its interaction with the coactivator PGC-1a (Blättler et al., 2012). As PGC-1a is specifically
expressed in cells and tissues with high mitochondrial content,
it is possible that the activity of YY1 through TISU may be
particularly important in PGC-1a-expressing cells while in
other cell types YY1 activity through TISU in transcription is
redundant.
While mitochondrial and energy-producing genes are highly
enriched with the TISU element, there are other genes in these
classes that are not regulated by TISU. An example is the
set of genes involved in fatty acid oxidation in the mitochondria, which we found to be also resistant translation inhibition
caused by metabolic energy stress. These genes most likely
utilize other, yet-unknown mechanisms to maintain their
translation ongoing during energy stress. Additional studies
Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc. 489
are required to elucidate other tolerance mechanism and
whether and how these are deregulated in disease states.
Another unexpected discovery emerging from this study is
that cells harbor an additional, AMPK-independent mechanism
of inhibition of translation initiation in response to energy deprivation. The molecular basis of this mechanism awaits further
investigation.
Knockdown of eIFs, measurement of ATP levels, plasmid construction,
coIP, GST pull-down assay, and global and gene-specific translation analysis
are described in the Supplemental Experimental Procedures.
SUPPLEMENTAL INFORMATION
Supplemental Information includes seven figures, three tables, and Supplemental Experimental Procedures and can be found with this article at http://
dx.doi.org/10.1016/j.cmet.2015.02.010.
EXPERIMENTAL PROCEDURES
AUTHOR CONTRIBUTIONS
Preparation of mRNA for In Vitro and In Vivo Translation Assays
All constructs used for mRNA preparation were described previously (Elfakess and Dikstein, 2008; Elfakess et al., 2011). For synthesis of capped
mRNA, the constructs containing the T7 promoter were linearized and
used with the RiboMAXTM Large Scale RNA Production Systems T7 (Promega) supplemented with Ribo m7G Cap Analog (New England Biolabs).
The reaction was stopped by the addition of RQ1 RNase-Free DNaseI
(Promega) and the mRNA extracted with phenol:chloroform and precipitated
with ethanol. The capped mRNAs were denatured at 65 C for 10 min
and then placed on ice for 2 min. The concentration of the synthesized
mRNAs was determined, and their integrity was confirmed by agarose gel
electrophoresis.
Cells, In Vivo Translation Assays, and Antibodies
HepG2, HeLa, and HEK293T cell lines were maintained in DMEM supplemented with 10% fetal calf serum, 100 units/ml penicillin, and 100 mg/ml streptomycin. MEFs were maintained as above in addition to 1 mM pyruvate. C2
myoblasts were maintained as above but with 20% serum. To differentiate
the C2 cells to myotubes, when the cells reached 80% confluency they were
grown in DMEM supplemented with 2% horse serum for 48 hr (Yaffe and
Saxel, 1977). AICAR and metformin treatments of HepG2 cells were done
in the absence of serum. For the in vivo translation assays, 0.5–1 mg of the
in vitro-transcribed mRNA was transfected into cells that had been previously
seeded on 12-well plates, using DharmaFECT1 transfection reagent (Dharmacon, Thermo Scientific). Two (HepG2) or four (MEFs) hours after transfection,
the cells were treated with the different drugs or media as indicated in the
figure legends. Cells were harvested and cell lysates were subjected to 12%
SDS-polyacrylamide gel electrophoresis. The levels of GFP protein were
determined using the commercial mouse monoclonal-anti-GFP antibody
(Abcam). Antibodies against AMPK, phospho AMPK, 4EBP, ACC, and phospho ACC are from Cell Signaling. RPS5 (ab58345) and eIF4GI (ab47649)
antibodies are from Abcam. eIF3c (sc74507) and eIF3d (sc28856) are from
SantaCruz, eIF1 antibodies are a kind gift from Ariel Stanhill (Technion, Haifa),
and anti-His are from QIAGEN.
Assembly of Ribosomal Complexes and Toe-Printing Assays
The reconstitution of 48S ribosomal complexes from purified components was
performed as described earlier (Pisarev et al., 2007). Primer extension inhibition (toe-printing) analysis of 48S initiation ribosomal complex formation was
performed with the use of AMV reverse transcriptase and the primer 50 TGAACTTGTGGCCGTTTACGT-30 . cDNA was analyzed by denaturing 6%
PAGE alongside sequencing reactions.
Assembly of 48S and 80S complexes was performed as described (Pisarev et al., 2007). Briefly, 50 ml rabbit reticulocyte lysate (Promega, Rabbit
Reticulocyte Lysate System L4960), 1.4 ml amino acid mix, 4 mM GMP-PNP
(Sigma) or 2 mg/ml cycloheximide (Sigma), 50 units RNase inhibitor (Promega), and 8 mg capped mRNA were mixed and incubated for 10 min at 30 C
(total volume of 70 ml). For 48S complex, the mixture was preincubated for
5 min at 30 C before adding the mRNA. Samples were chilled on ice and
loaded onto 8%–32% sucrose gradient made with buffer containing
20 mM TRIS (pH 8), 100 mM KAc, 2 mM DTT, 2.5 mM MgAc (for 48S),
or 5 mM MgCl2 and 0.1 mg/ml cycloheximide (for 80S). Gradients were
centrifuged at 40,000 RPM in a SW41 rotor for 3.5 hr (for 48S) or 3 hr (for
80S) at 4 C. Gradients were fractionated and the optical density at
254 nm was continuously recorded using ISCO absorbance detector
UA-6. Proteins from each fraction were precipitated by TCA and subjected
to western blot.
490 Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc.
R.D., H.S., and O.H. conceived and designed the study, analyzed the data, and
wrote the paper. H.S. and O.H. carried out most of the experiments. The toeprinting assays following 48S assembly from purified factors were carried out
by Y.S. and H.S in N.S.’s lab. A.T.-B.-H. conducted the bioinformatics analyses. B.V. contributed the AMPKa1/AMPKa2 knockout cells.
ACKNOWLEDGMENTS
We are grateful to Rafi Emmanuel (Weizmann Institute) for his help in setting
the RNA transfection assays and for helpful suggestions, Prof. Alan Hinnebusch (NIH) for critical reading and suggestions, Prof. David Yaffe and
Dr. Ora Fuchs (WIS) for the C2 myoblasts, and Adi Jacob (WIS) for her assistance in constructing recombinant eIF4GI fragments. This work was supported
by grants from the Minerva Foundation (#711124) (R.D.), Israel Science Foundation (#1168/13) (R.D.), Canadian Institute of Health Research (N.S.), and the
Norman Zavalkoff Family Foundation Travel and Research Fund for Cancer
Research (N.S. and H.S.). R.D. is the incumbent of the Ruth and Leonard
Simon Chair of Cancer Research.
Received: June 16, 2014
Revised: October 22, 2014
Accepted: February 9, 2015
Published: March 3, 2015
REFERENCES
Asano, K., Clayton, J., Shalev, A., and Hinnebusch, A.G. (2000). A multifactor
complex of eukaryotic initiation factors, eIF1, eIF2, eIF3, eIF5, and initiator
tRNA(Met) is an important translation initiation intermediate in vivo. Genes
Dev. 14, 2534–2546.
Blättler, S.M., Verdeguer, F., Liesa, M., Cunningham, J.T., Vogel, R.O., Chim,
H., Liu, H., Romanino, K., Shirihai, O.S., Vazquez, F., et al. (2012). Defective
mitochondrial morphology and bioenergetic function in mice lacking the transcription factor Yin Yang 1 in skeletal muscle. Mol. Cell. Biol. 32, 3333–3346.
Bolster, D.R., Crozier, S.J., Kimball, S.R., and Jefferson, L.S. (2002). AMP-activated protein kinase suppresses protein synthesis in rat skeletal muscle
through down-regulated mammalian target of rapamycin (mTOR) signaling.
J. Biol. Chem. 277, 23977–23980.
Calvo, S.E., Pagliarini, D.J., and Mootha, V.K. (2009). Upstream open reading
frames cause widespread reduction of protein expression and are polymorphic among humans. Proc. Natl. Acad. Sci. USA 106, 7507–7512.
Cantó, C., Jiang, L.Q., Deshmukh, A.S., Mataki, C., Coste, A., Lagouge, M.,
Zierath, J.R., and Auwerx, J. (2010). Interdependence of AMPK and SIRT1
for metabolic adaptation to fasting and exercise in skeletal muscle. Cell
Metab. 11, 213–219.
Cheung, Y.N., Maag, D., Mitchell, S.F., Fekete, C.A., Algire, M.A., Takacs, J.E.,
Shirokikh, N., Pestova, T., Lorsch, J.R., and Hinnebusch, A.G. (2007).
Dissociation of eIF1 from the 40S ribosomal subunit is a key step in start codon
selection in vivo. Genes Dev. 21, 1217–1230.
Dazert, E., and Hall, M.N. (2011). mTOR signaling in disease. Curr. Opin. Cell
Biol. 23, 744–755.
Dubbelhuis, P.F., and Meijer, A.J. (2002). Hepatic amino acid-dependent
signaling is under the control of AMP-dependent protein kinase. FEBS Lett.
521, 39–42.
Elfakess, R., and Dikstein, R. (2008). A translation initiation element specific to
mRNAs with very short 5’UTR that also regulates transcription. PLoS ONE 3,
e3094.
Elfakess, R., Sinvani, H., Haimov, O., Svitkin, Y., Sonenberg, N., and Dikstein,
R. (2011). Unique translation initiation of mRNAs-containing TISU element.
Nucleic Acids Res. 39, 7598–7609.
He, H., von der Haar, T., Singh, C.R., Ii, M., Li, B., Hinnebusch, A.G., McCarthy,
J.E., and Asano, K. (2003). The yeast eukaryotic initiation factor 4G (eIF4G)
HEAT domain interacts with eIF1 and eIF5 and is involved in stringent AUG selection. Mol. Cell. Biol. 23, 5431–5445.
Hinnebusch, A.G. (2011). Molecular mechanism of scanning and start codon
selection in eukaryotes. Microbiol. Mol. Biol. Rev. 75, 434–467.
Imataka, H., and Sonenberg, N. (1997). Human eukaryotic translation initiation
factor 4G (eIF4G) possesses two separate and independent binding sites for
eIF4A. Mol. Cell. Biol. 17, 6940–6947.
Imataka, H., Gradi, A., and Sonenberg, N. (1998). A newly identified N-terminal
amino acid sequence of human eIF4G binds poly(A)-binding protein and functions in poly(A)-dependent translation. EMBO J. 17, 7480–7489.
Inoki, K., Zhu, T., and Guan, K.L. (2003). TSC2 mediates cellular energy
response to control cell growth and survival. Cell 115, 577–590.
Ivanov, I.P., Loughran, G., Sachs, M.S., and Atkins, J.F. (2010). Initiation
context modulates autoregulation of eukaryotic translation initiation factor 1
(eIF1). Proc. Natl. Acad. Sci. USA 107, 18056–18060.
Jackson, R.J., Hellen, C.U., and Pestova, T.V. (2010). The mechanism of eukaryotic translation initiation and principles of its regulation. Nat. Rev. Mol.
Cell Biol. 11, 113–127.
Jäger, S., Handschin, C., St-Pierre, J., and Spiegelman, B.M. (2007). AMPactivated protein kinase (AMPK) action in skeletal muscle via direct phosphorylation of PGC-1alpha. Proc. Natl. Acad. Sci. USA 104, 12017–12022.
Kalender, A., Selvaraj, A., Kim, S.Y., Gulati, P., Brûlé, S., Viollet, B., Kemp, B.E.,
Bardeesy, N., Dennis, P., Schlager, J.J., et al. (2010). Metformin, independent
of AMPK, inhibits mTORC1 in a rag GTPase-dependent manner. Cell Metab.
11, 390–401.
Korneeva, N.L., Lamphear, B.J., Hennigan, F.L., and Rhoads, R.E. (2000).
Mutually cooperative binding of eukaryotic translation initiation factor (eIF) 3
and eIF4A to human eIF4G-1. J. Biol. Chem. 275, 41369–41376.
Kozak, M. (1984). Selection of initiation sites by eucaryotic ribosomes: effect of
inserting AUG triplets upstream from the coding sequence for preproinsulin.
Nucleic Acids Res. 12, 3873–3893.
Kozak, M. (1986). Influences of mRNA secondary structure on initiation by eukaryotic ribosomes. Proc. Natl. Acad. Sci. USA 83, 2850–2854.
Kozak, M. (1991a). Effects of long 50 leader sequences on initiation by eukaryotic ribosomes in vitro. Gene Expr. 1, 117–125.
Kozak, M. (1991b). A short leader sequence impairs the fidelity of initiation by
eukaryotic ribosomes. Gene Expr. 1, 111–115.
Kozak, M. (1991c). Structural features in eukaryotic mRNAs that modulate the
initiation of translation. J. Biol. Chem. 266, 19867–19870.
Kozak, M. (1998). Primer extension analysis of eukaryotic ribosome-mRNA
complexes. Nucleic Acids Res. 26, 4853–4859.
Krause, U., Bertrand, L., and Hue, L. (2002). Control of p70 ribosomal protein
S6 kinase and acetyl-CoA carboxylase by AMP-activated protein kinase and
protein phosphatases in isolated hepatocytes. Eur. J. Biochem. 269, 3751–
3759.
Laplante, M., and Sabatini, D.M. (2012). mTOR signaling in growth control and
disease. Cell 149, 274–293.
Larsson, O., Morita, M., Topisirovic, I., Alain, T., Blouin, M.J., Pollak, M., and
Sonenberg, N. (2012). Distinct perturbation of the translatome by the antidiabetic drug metformin. Proc. Natl. Acad. Sci. USA 109, 8977–8982.
Mader, S., Lee, H., Pause, A., and Sonenberg, N. (1995). The translation initiation factor eIF-4E binds to a common motif shared by the translation factor
eIF-4 gamma and the translational repressors 4E-binding proteins. Mol. Cell.
Biol. 15, 4990–4997.
Marcotrigiano, J., Lomakin, I.B., Sonenberg, N., Pestova, T.V., Hellen, C.U.,
and Burley, S.K. (2001). A conserved HEAT domain within eIF4G directs
assembly of the translation initiation machinery. Mol. Cell 7, 193–203.
Martin-Marcos, P., Cheung, Y.N., and Hinnebusch, A.G. (2011). Functional elements in initiation factors 1, 1A, and 2b discriminate against poor AUG
context and non-AUG start codons. Mol. Cell. Biol. 31, 4814–4831.
Martin-Marcos, P., Nanda, J., Luna, R.E., Wagner, G., Lorsch, J.R., and
Hinnebusch, A.G. (2013). b-Hairpin loop of eukaryotic initiation factor 1
(eIF1) mediates 40 S ribosome binding to regulate initiator tRNA(Met) recruitment and accuracy of AUG selection in vivo. J. Biol. Chem. 288, 27546–27562.
Martin-Marcos, P., Nanda, J.S., Luna, R.E., Zhang, F., Saini, A.K., Cherkasova,
V.A., Wagner, G., Lorsch, J.R., and Hinnebusch, A.G. (2014). Enhanced eIF1
binding to the 40S ribosome impedes conformational rearrangements of the
preinitiation complex and elevates initiation accuracy. RNA 20, 150–167.
Méthot, N., Song, M.S., and Sonenberg, N. (1996). A region rich in aspartic
acid, arginine, tyrosine, and glycine (DRYG) mediates eukaryotic initiation
factor 4B (eIF4B) self-association and interaction with eIF3. Mol. Cell. Biol.
16, 5328–5334.
Mitchell, S.F., and Lorsch, J.R. (2008). Should I stay or should I go? Eukaryotic
translation initiation factors 1 and 1A control start codon recognition. J. Biol.
Chem. 283, 27345–27349.
Nanda, J.S., Cheung, Y.N., Takacs, J.E., Martin-Marcos, P., Saini, A.K.,
Hinnebusch, A.G., and Lorsch, J.R. (2009). eIF1 controls multiple steps in start
codon recognition during eukaryotic translation initiation. J. Mol. Biol. 394,
268–285.
O’Neill, H.M., Maarbjerg, S.J., Crane, J.D., Jeppesen, J., Jørgensen, S.B.,
Schertzer, J.D., Shyroka, O., Kiens, B., van Denderen, B.J., Tarnopolsky,
M.A., et al. (2011). AMP-activated protein kinase (AMPK) beta1beta2 muscle
null mice reveal an essential role for AMPK in maintaining mitochondrial
content and glucose uptake during exercise. Proc. Natl. Acad. Sci. USA
108, 16092–16097.
Pelletier, J., and Sonenberg, N. (1985). Insertion mutagenesis to increase secondary structure within the 50 noncoding region of a eukaryotic mRNA reduces
translational efficiency. Cell 40, 515–526.
Pestova, T.V., and Kolupaeva, V.G. (2002). The roles of individual eukaryotic
translation initiation factors in ribosomal scanning and initiation codon selection. Genes Dev. 16, 2906–2922.
Pestova, T.V., Borukhov, S.I., and Hellen, C.U. (1998). Eukaryotic ribosomes
require initiation factors 1 and 1A to locate initiation codons. Nature 394,
854–859.
Pisarev, A.V., Unbehaun, A., Hellen, C.U., and Pestova, T.V. (2007). Assembly
and analysis of eukaryotic translation initiation complexes. Methods Enzymol.
430, 147–177.
Reiter, A.K., Bolster, D.R., Crozier, S.J., Kimball, S.R., and Jefferson, L.S.
(2005). Repression of protein synthesis and mTOR signaling in rat liver mediated by the AMPK activator aminoimidazole carboxamide ribonucleoside.
Am. J. Physiol. Endocrinol. Metab. 288, E980–E988.
Sedman, S.A., Gelembiuk, G.W., and Mertz, J.E. (1990). Translation initiation
at a downstream AUG occurs with increased efficiency when the upstream
AUG is located very close to the 50 cap. J. Virol. 64, 453–457.
Sonenberg, N., and Hinnebusch, A.G. (2009). Regulation of translation initiation in eukaryotes: mechanisms and biological targets. Cell 136, 731–745.
Lin, J., Handschin, C., and Spiegelman, B.M. (2005). Metabolic control through
the PGC-1 family of transcription coactivators. Cell Metab. 1, 361–370.
Valásek, L., Nielsen, K.H., and Hinnebusch, A.G. (2002). Direct eIF2-eIF3 contact in the multifactor complex is important for translation initiation in vivo.
EMBO J. 21, 5886–5898.
Maag, D., Fekete, C.A., Gryczynski, Z., and Lorsch, J.R. (2005). A conformational change in the eukaryotic translation preinitiation complex and release
of eIF1 signal recognition of the start codon. Mol. Cell 17, 265–275.
Villa, N., Do, A., Hershey, J.W., and Fraser, C.S. (2013). Human eukaryotic initiation factor 4G (eIF4G) protein binds to eIF3c, -d, and -e to promote mRNA
recruitment to the ribosome. J. Biol. Chem. 288, 32932–32940.
Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc. 491
Winder, W.W., Holmes, B.F., Rubink, D.S., Jensen, E.B., Chen, M., and
Holloszy, J.O. (2000). Activation of AMP-activated protein kinase increases
mitochondrial enzymes in skeletal muscle. J. Appl. Physiol. 88, 2219–2226.
Yaffe, D., and Saxel, O. (1977). Serial passaging and differentiation of
myogenic cells isolated from dystrophic mouse muscle. Nature 270,
725–727.
492 Cell Metabolism 21, 479–492, March 3, 2015 ª2015 Elsevier Inc.
Yanagiya, A., Svitkin, Y.V., Shibata, S., Mikami, S., Imataka, H., and
Sonenberg, N. (2009). Requirement of RNA binding of mammalian eukaryotic
translation initiation factor 4GI (eIF4GI) for efficient interaction of eIF4E with the
mRNA cap. Mol. Cell. Biol. 29, 1661–1669.
Zhang, B.B., Zhou, G., and Li, C. (2009). AMPK: an emerging drug target for
diabetes and the metabolic syndrome. Cell Metab. 9, 407–416.
Cell Metabolism, Volume 21
Supplemental Information
Translational Tolerance of Mitochondrial Genes
to Metabolic Energy Stress Involves TISU and
eIF1-eIF4GI Cooperation in Start Codon Selection
Hadar Sinvani, Ora Haimov, Yuri Svitkin, Nahum Sonenberg,
Ana Tamarkin-Ben-Harush, Benoit Viollet, and Rivka Dikstein
Figure S1, related to Figure 1A
A.
B.
!"#$%&'$()*)(+$,-).$$
/.,0+1)2340$
(#%"
(#$"
("
!#'"
!#&"
!#%"
!#$"
!"
HepG2
HeLa
293T
MEF
$'#
Figure S2, related to Figure 2
TISU/ Canonical
AUG
CMV-P
!"#
TISU
Si-C Si-eIF1
USDS-
Canonical
Si-C Si-eIF1
!-GFP
GFP
TISU
Si-C Si-eIF1B
eIF1
Tub
!-GFP
*
Canonical
Si-C Si-eIF1B
Si-C Si-eIF1
Long 5’UTR
Si-C Si-eIF1
!-GFP
*
)'#
$"#
AUG
CMV-P
('#
GFP
AUG
*
#
&"#$!%##
!"#$!%#
Long 5’UTR
Si-C Si-eIF1B
Si-C
eIF1B
USDS!-GFP
!-GFP
&'#
TISU
Si-C Si-eIF3c
Tub
!-GFP
*
*
#
#
!"#$!%#
&"#$!%#
Canonical
Si-C Si-eIF3c
Long 5’UTR
Si-C Si-eIF3c
USDS-
Si-C
eIF3c
!-GFP
Si-eIF1B
!-GFP
!-GFP
!"#$!%##
&"#$!%##
Tub
Si-eIF3c
Figure S3, related to Figure 2
Si-C Si-eIF3C
Si-C Si-eIF4GI
Si-C Si-eIF1/1B
Tub
Tub
Tub
eIF1
eIF4GI
eIF3C
F!
F!
F!
L!
H!
Sucrose density (10%-50%)
"#$%&
"#$'()
%"#
OD 254nM
!"# %"#
$"#
OD 254nM
OD 254nM
%"#
H!
L!
!"#$"#
Sucrose density (10%-50%)
L!
H!
!"#
$"#
Sucrose density (10%-50%)
Si-CT
Si-eIF
mRNA level
mRNA level
0.7
0.6
0.4
0.3
0.2
0.1
0
F
L
H
F
L
H
EGR1 (non-TISU)
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0
F
L
mRNA level
L
0
H
ACTIN (non-TISU)
0.8
0.2
0.3
0.2
0.1
0.1
F
L
H
F
L
H
HPRT1 (non-TISU)
0.6
0.5
0
F
L
H
F
L
RPL18 (non-TISU)
0.6
0.4
0.4
0.4
F
L
H
0
F
L
H
F
L
H
ACTIN (non-TISU)
0.7
0
0.6
0.5
0.5
0.4
0.4
0.4
0.3
0.3
0.1
0.1
F
L
H
F
L
H
HPRT1 (non-TISU)
0
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0
F
L
H
F
L
H
RPL18 (non-TISU)
0.1
0.1
0.1
0.1
0
0
0
0
0
H
PPP2R2A
F
L
L
H
F
L
H
F
L
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0
0
F
L
H
F
L
H
CAPZA2
0.7
F
L
H
F
L
H
ATP5I
H
ARFGAP
0.1
0.1
F
L
H
F
L
H
HMGCL
0.7
0
H
F
L
H
AMPKa2
F
L
H
F
L
H
CYC1
0.7
0
0
0
H
CAPZA2
0.5
0.4
0.4
0.4
0.3
0.3
0.3
0.2
0.2
0.2
0.1
0.1
0.1
0
F
L
H
F
L
H
HMGCL
0.8
0
0.3
0.2
0.2
0.1
0.1
0.1
0.1
0
0
0
0
F
L
H
CYC1
F
L
AMPKa1
H
F
L
H
F
L
H
F
L
H
PPP2R2A
H
F
L
H
F
L
H
F
L
H
F
L
H
F
L
H
F
L
H
F
L
H
AMPKa1
0.4
0.3
0.2
0.1
F
L
H
F
L
H
AMPKa2
0
F
L
H
ATP5I
0.7
0.6
0.5
0.4
0.3
0.2
0.1
F
L
H
F
L
H
CAPZA2
0
F
L
H
ARFGAP
0.6
0.5
0.4
0.3
0.2
0.1
F
L
H
F
L
H
HMGCL
0
F
L
H
CYC1
0.7
0.6
0.5
0.4
0.3
0.2
0.1
L
0
0.5
0.2
F
H
RPL18 (non-TISU)
0.7
0.6
0.3
0.2
L
0.1
0.4
0.3
F
0.2
0.5
0.4
0.2
0
H
0.3
0.6
0.5
0.3
H
H
0.6
0.4
L
L
0.7
0.3
F
F
L
0.4
0.7
0.6
0.5
H
H
0.5
0.6
L
L
0.6
0.7
F
F
0.5
0.4
H
H
0.6
0.4
L
L
ARFGAP
0.7
0.5
F
F
F
0.5
0.2
0
0
0.6
0.3
0.1
L
H
0.4
0.1
0.6
H
0
0.1
F
L
0.5
0.2
H
F
HPRT1 (non-TISU)
0.6
0.3
0.5
L
ATP5I
0.7
0.6
F
H
0.4
L
H
0.1
0.5
F
L
0.2
L
ACTIN (non-TISU)
0.9
0.1
F
0.3
F
H
0.2
0.4
H
L
0.3
0.5
L
F
0.4
0.6
F
H
0.5
0.7
0.6
0.7
0.2
0.2
L
0.2
0.3
0.3
F
0.3
L
AMPKa1
0.1
0.4
F
H
0.2
0.5
H
L
0.3
0.6
L
F
0.4
0.7
F
H
0.5
0.8
0.4
0.4
0
L
L
0.6
0.2
F
F
0.7
0.3
0.6
0.1
0.5
0.5
PPP2R2A
0.2
0.6
0.6
H
0.3
0
H
L
0.4
0.1
0.1
F
0.5
0.2
0.2
H
0.6
0.3
0.3
L
0.7
0.4
0.4
F
0.8
0.5
0.5
0
F
AMPKa1
0.6
0.6
H
0
0.8
0.4
0.1
L
H
0.5
0.2
F
L
EGR1 (non-TISU)
0.6
0.2
H
F
0.1
0.3
L
H
0.2
0.3
F
L
0.3
0.7
0.6
F
0.6
0.5
0.2
0.1
0.1
EGR1 (non-TISU)
0.2
0.2
0.2
0.1
H
0.3
0.3
0.2
L
0.4
0.5
0.3
DNAJB1 (non-TISU)
0.6
0.5
0.6
0.4
0.7
0.5
0.7
0.5
F
GAPDH (non-TISU)
0.8
0.6
0
H
DNAJB1 (non-TISU)
0.7
0.6
0.7
0.5
AMPKa2
mRNA level
F
0.4
0
mRNA level
H
0.6
0.3
GAPDH (non-TISU)
0.6
0.5
0.7
0.4
0.7
mRNA level
DNAJB1 (non-TISU)
0.6
0.5
0.5
0
mRNA level
GAPDH (non-TISU)
0.1
0Si-C
F
Si-eIF
L
H
F
L
H
0
F
L
H
Figure S4, related to Figure 2
A.
84
eIF4E
197
198
eIF3/eIF4A
1599
eIF4A
674
675
1129
1599
1130
Coomassie
staining
eIF1
GST
GST-4GI
675-1129!
Input
B.
GST-4GI
1130-1599!
Coomassie
staining
GST
GST-eIF1A
Input
4GI
4GI
675-1129 1130-1599
GST
GST-eIF1A
Input
M
4GI
198-674
GST
GST-eIF1A
Input
GST
GST-eIF1A
GST
GST-eIF1
Input
GST
GST-eIF1
GST
GST-eIF1
Input
GST
GST-eIF1
Input
Input
GST
GST-eIF1
M
Input
4GI
84-197
4GI
4GI
4GI
4GI
84-197! 198-674! 675-1129! 1130-1599!
GST
GST-eIF1A
eIF4GI
1 PABP
Figure S5, related to Figure 3
A.
Sucrose density (Top-! bottom)
!" #" $"
%" &" '" (" )" *" !+"
OD 254nM
%),"
-.,/"
012324516"
!"
#" $" %" &" '" (" )" *" !+"
-.,/"
7"89,&""
012324516"
7"89,&""
-.,/"
7":.;#<""
012324516"
7":.;#<""
-.,/"
7":.;$?""
012324516"
7":.;$?""
-.,/"
7":.;%=.""
012324516"
7":.;%=.""
-.,/"
7":.;%>""
012324516"
7":.;%>""
B.
CT
RPS 5
AICAR
RPS 5
CT
RPL 11
AICAR
RPL 11
F 40s 60s 80s
Polysomal fractions
)'(
F!
"#!
$%&'()*!)+,-.,/(0!1$23
%"#
Myotubes
!"# $"#
L!
P-ACC-
mRNA level
ACC-
0.6
0.6
0.5
0.5
0.7
0.5
0.4
0.3
0.3
0.3
0.2
0.2
0.2
0.1
0.1
0.1
0
0
0
L
H
F
L
H
ARFGAP2
0.6
0.5
0.4
F
L
H
F
L
ATP5O
0.8
0.6
0.4
0.5
0.4
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0
0
0
MRPL35
0.7
F
H
L
0.5
0.5
0.4
0.4
F
L
L
H
F
L
H
NDUFS6
1
0.3
0.3
F
H
NDUFA9
0.6
0.6
H
0.1
0.1
F
L
H
F
L
SUCLG2
0.6
F
H
L
0.4
F
L
0.3
0.4
0.3
0.4
0.2
0.2
0.1
0.6
0.5
0.5
0.4
L
H
F
L
DPF2
0.6
F
H
L
H
F
L
H
DOLPP1
0.8
0.6
0.4
0.4
0.3
0.2
0.2
0.1
0.1
0
0
F
L
H
F
L
H
F
L
H
F
L
H
L
H
L
H
F
L
H
L
H
SIRT1
0
F
L
H
F
ATP6V1B2
0.7
0.5
0.4
0.3
0.2
0.1
0
F
L
H
F
L
H
0.3
0.2
0.2
0.1
F
L
H
F
L
H
0.1
0
0
F
L
F
L
H
F
L
H
H
F
L
H
0.8
0.7
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
H
F
L
H
L
CT
GS
0
L
F
CPT1A (FAO)
0.8
F
0.6
GS
0.5
0.3
ALG1
CT
0.7
0.5
H
0
0
F
L
0.1
0.1
0
F
0.2
0.2
0.1
H
0.3
0.3
0.2
L
0.6
0.4
0.3
F
0.4
0.2
0
F
0.6
H
ABCF3
0.7
0.5
H
F
0
0.8
0
0
0
H
CYC1
0.5
0.2
0.2
L
0.7
0.5
0.1
L
CAPZA2
0.5
0.3
ACSL5 (FAO)
F
0.5
0.1
F
H
0.6
0.2
H
L
0.6
0.1
L
F
0.6
0.3
F
H
0.7
0.4
0.2
L
0.7
0.5
0.3
F
H
0.7
0.6
EEF2 (NON-TISU)
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0.6
0.4
0.4
GAPDH(NON-TISU)
ACAD9 (FAO)
0.8
0.5
0.1
mRNA level
mRNA level
0.7
F
mRNA level
0.6
0
RPL18 (NON-TISU)
ACAA2 (FAO)
ACAA1 (FAO)
0.8
AMPK-
Sucrose density (10%-50%)
RPS18 (NON-TISU)
mRNA level
!
Figure S6, related to Figure 6
0.7
*'(
mRNA level
,'(
P-AMPK-
H!
OD 254nm
Myoblasts
+'(
mRNA level
&'(
F
L
H
F
L
H
H
F
L
H
Figure S7, related to Figure 6
CT
GS
mRNA level
RPS18 (NON-TISU)
0.7
0.7
0.6
0.6
0.6
0.6
0.5
0.5
0.5
0.5
0.4
0.4
0.4
0.4
0.3
0.3
0.3
0.3
0.2
0.2
0.2
0.2
0.1
0.1
0.1
0.1
0
0
0
mRNA level
L
H
F
L
H
ARFGAP2 (TISU)
0.6
0.4
0.3
L
H
F
L
H
ATP5O (TISU)
0.8
0
F
L
H
F
L
CAPZA2 (TISU)
0.8
F
H
L
H
F
L
H
CYC1 (TISU)
0.7
0.6
0.6
0.4
0.4
0.2
0.2
0
0
0.5
0.4
0.3
0.2
0.1
0.2
0.1
F
mRNA level
F
EEF2 (NON-TISU)
0.6
0.5
0
L
H
F
L
H
MRPL35 (TISU)
0.7
F
0.7
L
H
F
L
H
NDUFA9 (TISU)
F
0.6
0.5
0.5
0.4
0.4
0.6
0.3
0.3
0.4
0.2
0.2
0.1
0.1
0
0
L
H
F
L
H
SUCLG2 (TISU)
0.8
0.4
0.4
0.2
0.2
0
L
H
F
L
H
DOLPP1 (TISU)
0.8
L
H
F
L
H
ABCF3 (TISU)
H
F
L
H
F
L
H
SIRT1 (TISU)
0.6
0
F
L
H
F
L
H
ALG1 (TISU)
0.6
L
H
F
L
H
DPF2 (TISU)
0.6
F
L
H
F
L
H
L
H
F
L
H
ATP6V1B2(TISU)
0.8
0.4
0.2
0
F
F
0.6
0.1
0
L
H
0.1
0.2
0.1
F
L
0.2
0.3
0.2
0
F
0.4
0.4
0.3
0.2
0
0.5
0.5
0.4
0.4
H
0
0.5
0.6
L
0.2
0
F
F
0.3
0.8
0.6
H
0.8
F
0.6
L
NDUFS6 (TISU)
1
0.6
F
mRNA level
GAPDH(NON-TISU)
0.7
F
mRNA level
RPL18 (NON-TISU)
0.7
0
F
L
H
F
L
H
F
L
H
F
L
H
Supplementary figure legends
Figure S1: A. Representative image of an agarose gel that was used to estimate the
integrity of the in vitro transcribed mRNAs used for in vivo and in vitro translation
assays. B. Analysis of the relative stability of in vivo translated mRNAs. In vitro
transcribed mRNAs were transfected into the indicated cell lines and total RNA was
prepared 24 hours after transfection. GFP mRNA was analyzed by real-time RTqPCR.
Figure S2: The effects of eIF1, eIF1B and or eIF3c KD on translation efficiency and
accuracy. HEK293T cells were transfected with siRNA against eIF1 (B), eIF1B (C),
eIF3c (D) or with a non-targeting siRNA as a negative control as indicated in each
section. After 48h cells were transfected with the GFP reporter genes shown
schematically (A). Cells were harvested 24h after the second transfection and
analyzed for GFP expression by Western blot as described in Fig. 2. The graphs
represent average ± SE of 3-4 independent experiments.
Figure S3: The effects of eIF1/eIF1B, eIF4GI and eIF3c KD on translation.
HEK293T cells were transfected with siRNA against the indicated translation
initiation factors and 72h later subjected to ribosomal profiling and western blot with
the indicated antibodies (upper panels). The black and the red lines show the profile
of the control (CT) and the siRNA treated cells, respectively. The bottom panel shows
real-time qPCR analysis of the levels of the indicated non-TISU and TISU mRNAs in
the polysome-free fractions (F), light polysomes fractions (L) and in the heavy
polysomes fractions (H).
Figure S4: eIF4GI interacts directly with eIF1. A. Upper panel: a scheme showing
eIF4GI domains and eIF4GI protein fragments used for the in vitro binding assays
with eIF1 and eIF1A. The black boxes denote protein interaction sites. Lower panel,
GST and GST-eIF1 (left) or GST and GST-eIF1A (right) were coupled to glutathione
sepharose beads and incubated with four different fragments of eIF4GI, each with a
C-terminal 6XHis tag. The pulled-down complexes were washed and then run on
SDS-PAGE followed by western blot with anti-6XHis antibody. Inputs represent 1%
of the lysate used for the binding. The arrows indicate the position of the eIF4G
fragments in the input lanes. The GST and the GST fusion proteins used for the
binding reactions are shown in the Coomassie blue stained gel shown on the left side
of each panel. B. A reciprocal binding assay in which GST-4GI-675-1129 and GST4GI-1129-1599 fragments coupled to glutathione sepharose beads were incubated
with bacterially expressed His-eIF1. Input represents 5% of the amount used for
binding.
Figure S5: A. The fate of eIFs following 48S complex assembly on short 5’UTR
mRNAs. Capped mRNAs, bearing a short 5’UTR with an AUG in either TISU or
canonical AUG contexts, were incubated with RRL in the presence of GMP-PNP to
stall the 48S. The canonical AUG mRNA has also an in-frame downstream AUG as
similar to the TISU used in Fig. 3A. The TISU mRNA used in this experiment lacks a
downstream AUG. Ribosomal-mRNA complexes were then subjected to 8-32%
sucrose gradient sedimentation and the collected fractions were recorded (top panel)
and analyzed by western blot for the distribution of eIFs and ribosomal subunits as
indicated (bottom panels). B. Analysis of ribosomal proteins along the sucrose
gradients following energy stress. Western blot analysis of the distribution of RPS5
and RPL11 in the 11 fractions of the sucrose gradients in control and AICAR treated
HepG2 cells.
Figure S6: Translational response of C2 myotubes to energy stress. A. A picture
taken from C2 myoblasts (left) and C2 cells that were differentiated into myotubes
(right). B. C2 myotubes were glucose deprived for 8 hours and then subjected to
ribosomal profiling. The black and the red lines show the profile of the control (CT)
and glucose starved (GS) cells, respectively. C. Western blot analysis of
phosphorylated and total AMPK and ACC under glucose starvation (GS). D. RealTime qPCR analysis of the levels of the indicated non-TISU and TISU mRNAs in the
polysome-free fractions (F) of the gradient, light polysomes fractions (L) and in the
heavy polysomes fractions (H). E. The effect of energy stress on translation of fatty
acid oxidation genes (FAO). MEFs were glucose deprived for 8 hours and then
subjected to ribosomal profiling (see Fig. 6A). Real-Time qPCR analysis of the levels
of the indicated FAO mRNAs in the polysome-free fractions (F), light polysomes
fractions (L) and in the heavy polysomes fractions (H).
Figure S7: The translational response of non-TISU and TISU genes to energy stress
in AMPK deficient cells. AMPK DKO MEFs were glucose deprived for 8 hours and
then subjected to ribosomal profiling (see Fig. 6B). Real-Time qPCR analysis of the
levels of the indicated non-TISU and TISU mRNAs in the polysome-free fractions
(F), light polysomes fractions (L) and in the heavy polysomes fractions (H).
!
!"#$%&'()&*%$"+%,&+-&./01*%&2&
!"#$%
!"#$%&$'()"*+,)"-$.$/0,
;0.<")*#$)=,%&*"',
!"#$%&$'()"*+,ABC,.='#&0.".,
%$D<+0(,0+0%#)$',#)*'.<$)#,
E0++D+*),)0.<")*#"$',
FG"(*#"H0,<&$.<&$)=+*#"$',
7+0%#)$',#)*'.<$)#,
7'0)I=,(0)"H*#"$',-=,$G"(*#"$',
$J,$)I*'"%,%$/<$D'(.,
!"#$%&$'()"*+,+D/0',
!"#$%&$'()"*+,/*#)"G,
B)*'.+*#"$',
!"#$%&$'()"*+,"''0),/0/-)*'0,
!"#$%&$'()"*+,/0/-)*'0,
!"#$%&$'()"$',
%
&'#()*$"'+% 0%12.3"%
,-./%
456%
1234,
1239,
?36,
53657824,
936>782?,
@34@7824,
9365782:,
>315782@,
23221125,
63>,
43>,
439,
>34,
531?782?,
:3257812,
53227824,
93@17824,
93>@782@,
:3:47824,
53@@782:,
>366782:,
>3:,
>3:,
:3@,
:3:,
93:,
935,
43647811,
43647811,
@3:>7811,
131>782?,
13447824,
63?17816,
13247824,
13247824,
132>7824,
13>>782@,
53:2782:,
1351781:,
Tables
Table S1: Functional classification of the TISU genes in which the AUG of TISU
constitutes the beginning of their ORF, using the Database for Annotation,
Visualization and Integrated Discovery (DAVID). The table shows the functional
groups with the highest enrichment score, their enrichment fold, P-value and the false
discovery rate (FDR).
Table S2: List of human genes bearing TISU element with up to 2 mismatches. The
file also contains a venn diagram showing the number of genes that common in
human and mouse.
Table S3: List of mouse genes bearing TISU element with up to 2 mismatches.!
Supplemental Experimental Procedures
Knockdown of eIFs
HEK293T cells were seeded on a 6-well plate and transfected with 50 nM Dharmacon
siGENOME SMART pool siRNA purchased from Thermo Scientific (eIF1 M015804-01-0010; eIF1B M-019996-00, eIF4GI M-019474-01 and eIF3 M-00903601) using DharmaFECT1 transfection reagent. The Dharmacon ON-TARGETplus
Non-targeting siRNA #3 was used as a negative control. 48h after the initial
transfection, cells were transfected with the GFP reporter plasmids using the standard
CaPO4 method. Cells were harvested 24h after the second transfection.
Measurement of ATP levels
HEK293T cells in 96-well plate (7,000 cells/well) were transfected with control or
eIF4GI siRNA as described above. Then, 72h later cells were subjected to ATP levels
measurement using CellTiter-Glo Luminescent Assay (promega).
Plasmid construction
Four different fragments of Homo sapiens eIF4GI (4GI 84-197, 4GI 198-674, 4GI
675-1129, 4GI 1130-1599) were amplified by PCR and cloned into pET28a
expression vector in frame to its C-terminal 6XHis tag sequence. GST-eIF1 and GSTeIF1A were constructed by cloning PCR fragments encoding eIF1 and eIF1A into
pGEX6p1 expression plasmid in frame to its N-terminal GST sequence using the TPCR cloning method (Erijman et al., 2011).
Co-immunoprecipitation and GST pull down assay
HEK293T cells were transfected with plasmids directing expression of either HAtagged eIF1 or eIF1A. 24h later cells were lysed using IP buffer (20mM Tris pH8,125
mM NaCl, 10% Glycerol,0.5% NP-4, 0.2mM EDTA) to which fresh protease
inhibitor cocktail (Sigma, 1:100) and PMSF 200µM (1:100) were added. Protein
extract was subjected to immunoprecipitation assay using either HA antibody (Sigma)
or control antibody in IP buffer, at 4oC for 16h. Each reaction was then washed three
times with IP buffer. After the washes 30µl of protein sample buffer were added to
each sample. 2% of the input and 30% of each IP sample were then subjected to 8%
and 15% SDS-PAGE followed by western blot.
GST-eIF1, GST-eIF1A and various 4GI protein fragments were expressed in E. Coli
BL-21 DE3 strain. GST fusion proteins were then purified using gluthatione
sepharose beads (GE healthcare). GST pull-down reactions were performed in HEMG
buffer (20mM HEPES, 12.5mM MgCl2, 0.5mM EDTA, 0.1% NP-40, 10% Glycerol
and 100 mM KCl), at 4oC for 2h. Each reaction was then washed three times, after
which 30µl of protein sample buffer was added. 1% or 5% input amount and 30% of
each pull-down sample were then subjected to 12% SDS-PAGE followed by western
blot. In addition, 30% of each pull-down assay sample were also subjected to another
SDS-PAGE followed by coomassie blue staining. His-eIF1 was purified as previously
described(Elfakess et al., 2011).
Global and gene specific translation analysis
HepG2 and MEFs and C2 cells were cultured in 15 cm plates and the following day
were treated with 2mM AICAR for 3 hours in the absence of serum (HepG2) or were
incubated in glucose free DMEM supplemented with 10% dialyzed FBS or 20nM
rapamycin for 8 hours (MEFs) or 2% dialyzed horse serum (C2 myotubes). The cells
were then incubated with 100 µg/ml Cycloheximide for 5 minutes and then washed
twice with cold buffer containing 20mM Tris pH 8, 140mM KCl, 5mM MgCl2 and
100 µg/ml Cycloheximide. The cells were collected and lyzed with 500 µl of same
buffer that also contains 0.5% Triton, 0.5% DOC, 1.5mM DTT, 150 units RNAse
inhibitor and 5µl of protease inhibitor (Sigma). The lyzed samples were centrifuged at
12,000g at 4°C for 5 minutes. The cleared lysates were loaded onto 10-50% sucrose
gradient and centrifuged at 41,000 RPM in a SW41 rotor for 90 minutes at 4°C.
Gradients were fractionated and the optical density at 254 nm was continuously
recorded using ISCO absorbance detector UA-6. RNA was isolated using Trizol and
Direc-Zol RNA mini-prep kits (Zymo Research). cDNA was prepared from 500ng
RNA using High-capacity cDNA reverse transcription kit (Applied Biosystems).
Real-Time PCR was done with Power-SYBR green master mix (Applied Biosystems)
in 7300 Real-Time PCR System. Primers sequences can be provided upon request.
!