N- and C-terminal Upf1 phosphorylations create
binding platforms for SMG-6 and SMG-5:SMG-7
Yukiko Okada-Katsuhata1,2, Akio Yamashita1,3,4,5,*, Kei Kutsuzawa1, Natsuko Izumi1,
Fumiki Hirahara2and Shigeo Ohno1,5,*
1Department of Molecular Biology,2Department of Obstetrics, Gynecology and Molecular Reproductive
Science,3Department of Microbiology and Molecular Biodefense Research, Yokohama City University,
School of Medicine, 3–9, Fuku-ura, Kanazawa-ku, Yokohama 236-0004,4Precursory Research for Embryonic
Science and Technology, Japan Science and Technology Agency, 4-1-8, Honcho, Kawaguchi 332-0012 and
5Advanced Medical Research Center, Yokohama City University, Yokohama 236-0004, Japan
Received August 4, 2011; Revised September 7, 2011; Accepted September 8, 2011
degrades mRNAs containing premature termination
codons (PTCs). SMG-1-mediated Upf1 phosphor-
ylation takes place in the decay inducing complex
factors, Upf1, SMG-1, an exon junction complex
(EJC) and a PTC-mRNA. However, the significance
and the consequence of Upf1 phosphorylation
remain to be clarified. Here, we demonstrate that
SMG-6 binds to a newly identified phosphorylation
site in Upf1 at N-terminal threonine 28, whereas the
SMG-5:SMG-7 complex binds to phosphorylated
serine 1096 of Upf1. In addition, the binding of the
SMG-5:SMG-7 complex to Upf1 resulted in the dis-
sociation of the ribosome and release factors from
the DECID complex. Importantly, the simultaneous
binding of both the SMG-5:SMG-7 complex and
SMG-6 to phospho-Upf1 are required for both
NMD and Upf1 dissociation from mRNA. Thus, the
SMG-1-mediated phosphorylation of Upf1 creates a
binding platforms for the SMG-5:SMG-7 complex
and for SMG-6, and triggers sequential remodeling
of the mRNA surveillance complex for NMD induc-
tion and recycling of the ribosome, release factors
and NMD factors.
Eukaryotes have a conserved RNA surveillance mech-
anismto help maintaincorrectgene expression.
Nonsense-mediated mRNA decay (NMD) is an mRNA
(PTCs) to eliminate potentially harmful C-terminally
truncated proteins (1–3). NMD also targets many
physiological mRNAs to regulate abundance, including
mRNAs encoding selenocysteine-containing proteins and
mRNA-like non-coding RNAs (4–6). If C-terminally
truncated proteins retain some of their function and/or
PTC-read through produces functional proteins, NMD
suppression leads to the phenotypic rescue of certain
PTC-related mutations (7–9). In addition, NMD suppres-
sion can augment un-natural polypeptides, which are
putative tumor-specific antigens encoded by frame-shift
mutations on PTC-mRNAs (10). Thus, clarification of
the mechanism of NMD is critical for the development
of pharmacological reagents for genetic diseases and
The current model of mammalian PTC recognition
junction complex (EJC) components, 20–24nt upstream
of an exon–exon junction (13) and deposition of nine
conserved trans-acting factors of NMD (1,14,15). In the
initial round of translation (16,17), if a translation termin-
ation codon is located upstream of an EJC, the SMG-1
Upf1 RNA helicase and translation termination factors
form the SMG-1:Upf1:eRF1:eRF3 complex (SURF),
which associates with the ribosome on the messenger
ribonucleoprotein (mRNP) (18–21). The association of
ribosome:SURF with the EJC forms the decay inducing
complex (DECID), which can distinguish a PTC from a
SMG-1-mediated Upf1 phosphorylation, a rate-limiting
The DECID induces
*To whom correspondence should be addressed. Tel: +81 45 787 2596; Fax: +81 45 785 4140; Email: firstname.lastname@example.org
Correspondence may also be addressed to Akio Yamashita. Tel: +81 45 787 2597; Fax: +81 45 785 4140; Email: email@example.com
Published online 29 September 2011 Nucleic Acids Research, 2012, Vol. 40, No. 3 1251–1266
? The Author(s) 2011. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/
by-nc/3.0), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
step of NMD (19–25). Recognition of PTCs, independ-
ently of the downstream splicing junction, has also been
demonstrated in mammals, and the molecular mechanism
for this process is under debate (26–29). In contrast to the
events before PTC recognition, those after PTC recogni-
tion are not well understood. For instance, whereas a
ribosome is present in the DECID complex during PTC
discrimination and PTC mRNA decay is expected to
occur after dissociation of the mRNA from ribosomes,
the timing and mechanism of ribosome dissociation from
mRNA is unclear.
As noted above, a critical consequence of PTC recogni-
tion is the phosphorylation of Upf1 by SMG-1 (22,25).
SMG-1 phosphorylates Upf1 in vitro at several serine/
C-terminal regions (22). Among them, S1078, S1096 and
S1116 are phosphorylated in vivo in mammals (22,30,31).
However, the functional importance of these phosphoryl-
ation sites remains to be clarified. In addition to phos-
phorylation, dephosphorylation is also necessary for
NMD (30,32,33). SMG-5, SMG-6 and SMG-7 are
involved in the dephosphorylation of Upf1, probably
through the recruitment of protein phosphatase 2A
(PP2A) (19,30,32–35). SMG-5, SMG-6 and SMG-7 are
evolutionally conserved related proteins, but each is
required for NMD (32,36). The majority of SMG-5 and
SMG-7 forms a complex (the SMG-5:SMG-7 complex)
(phospho-Upf1) in vivo, whereas phospho-independent
bindingof the SMG-5:SMG-7
N-terminal region of Upf1 is observed during NMD
(30). The binding of the SMG-5:SMG-7 complex to
phospho-Upf1 induces Upf2 dissociation from Upf1,
and this step is involved in NMD (30). SMG-5 and
SMG-7 share the 14-3-3-like domain. The SMG-7
14-3-3-like domain can directly bind in vitro to a Upf1
complex on Upf1 in vivo remain to be clarified. SMG-7
is considered as mRNA decay mediator since it is
tethering at either 30- or 50-UTR of mRNA induce Dcp2
(decapping enzyme) and Xrn1 (50-30-exonuclease) depend-
ent mRNA decay (38). SMG-6 also shares the 14-3-3-like
domain, which has been proposed to compete with the
SMG-5:SMG-7 complex for binding to phospho-Upf1
(2,3), but association of SMG-6 with phospho-Upf1 has
not been determined (39). SMG-5 and SMG-6 have a
C-terminal PilTN-terminus (PIN) domain. The PIN
domain of SMG-6 has endonuclease activity in vitro and
catalytically inactive SMG-6 fails to support NMD in
mammalian cells (40,41). While SMG-5, SMG-6 and
SMG-7 are required for NMD, their mechanisms of
action remain to be clarified.
phosphorylation of T28 and S1096 of Upf1 create
binding platforms for SMG-6 and the SMG-5:SMG-7
complex, respectively. SMG-6 associates in vivo with
phosphorylated Upf1 through its 14-3-3-like domain. We
also show that the phospho-specific binding of SMG-6 and
the SMG-5:SMG-7 complex to Upf1 is required for NMD.
Furthermore, we provide
in the N- and
to phosphorylated Upf1
involvement of the SMG-5:SMG-7 complex in the dissoci-
ation of the ribosome from DECID after Upf1 phosphor-
ylation. In addition, we suggest that the phospho-specific
binding of SMG-6 is required for Upf1 dissociation from
MATERIALS AND METHODS
Plasmids, antibodies and siRNAs
Expression vectors for wild-type Flag-HA-streptavidin
Upf1-mutants [-dCT (amino acids 6–1027), -dNCT
(amino acids 64–1027), -S1078A, -S1096A, -S1116A, -
T28A, -2SA (SS1078/1096AA), -4SA (SSSS1073/1078/
1116AAAAA)] were constructed in the mammalian ex-
pEF_Flag-HA-SBP or pSR-HA, following standard pro-
cedures. The wild-type Flag-HA-SBP-SMG-6 and SMG-6
mutants were mutated at coding sequence nucleotides to
confer siRNA SMG-6 resistance by site-directed mutagen-
esis. HA-SMG-5, HA-SMG-5dCT and HA-Upf1-4SA
plasmids were previously described (19,22,30).
The following siRNA target sequences were used:
SMG-5, GAAGGAAATTGGTTGATAC; SMG-6, GG
GTCACAGTGCTGAAGTA; SMG-7, CAGCACAGTC
TACAAGCCA; non-silencing (NS), All Star Negative
Control siRNA (Qiagen).
Anti-eIF4A3, anti-SMG-5 and anti-SMG-6 antibodies
were generated against recombinant human eIF4A3
(amino acids 1–48), SMG-5 (amino acids 416–541) or
SMG-6 (aminoacids 58–181)
glutathione S-transferase (GST) (anti-eIF4A3 and -
SMG-5) or maltose-binding protein (MBP) (anti-SMG-
6), respectively. Affinity purification of the antibody was
performed following standard procedures. An anti-P-T28-
Upf1 antibody was generated as described previously (30)
using a keyhole limpet hemocyanin (KLH)-conjugated
antigen. The anti-SMG-1, Upf1 (clone 5C3), Upf2,
SMG-7, SMG-9 and P-1078/1096-Upf1 (clone 7H1)
(19,20,22,30). Antibodies to eRF1 (Abcam), rpL7a (Cell
signaling technology), rpS16 (Abcam), Magoh (Abcam),
Y14 (Abcam), eIF4E (Santa Cruz), PABPC1 (Abcam),
PABPC4 (Bethyl), CBP80 (Bethyl), b-actin (Sigma),
GAPDH (Abcam) and HA (3F10) (Roche) were used.
Upf1 wild-type and
Transfection with plasmids and siRNAs
HeLa TetOff cells (Clontech) and HEK 293 cells were
grown in Dulbecco’s modified Eagle’s medium (DMEM)
supplemented with 10% fetal bovine serum, 100U/ml
penicillin and 100mg/ml streptomycin. Plasmid transfec-
tions were performed in six-well plates, 10 or 15cm
dishes using Polyfect (Qiagen), Transfectin (BioRad) or
LipofectaminLTX (Invitrogen) according to the manufac-
turer’s recommendations. For affinity purification of
SBP-tagged proteins, cells were harvested 48h after trans-
fection. siRNA transfections were performed in six-well
1252Nucleic Acids Research, 2012,Vol.40, No. 3
plates or in 15cm dishes, using siLentFect (BIORAD) or
LipofectaminRNAiMAX (Invitrogen) according to the
Immunoprecipitation, SBP purification and western blot
HeLa TetOff cells were transfected using Lipofectamine
LTX (Invitrogen) or Polyfect (Qiagen) and incubated for
30min at 4?C in a lysis buffer containing 50mM Tris–HCl
at pH 7.4, 50mM NaCl, 0.05% Tween-20, 100nM
lysates were added to 200mg/ml RNaseA (Qiagen), and
then incubated with appropriate antibodies for 1h with
gentle rotation. Subsequently, lysates with antibodies
were incubated with 30ml of protein G-Sepharose (GE
Biotech) for an additional 1h at 4?C with gentle
rotation. To immunoprecipitate HA-tagged proteins,
25ml of anti-HA-affinity matrix (Roche), pre-adsorbed
with antibody, was used instead of protein G-sepharose.
Sepharose (GE Biotech) was used and lyastes were
incubated for 2h at 4?C. The immunocomplexes (or
SBP–streptavidin complex) were washed with washing
buffer containing 50mM Tris–HCl at pH 7.4, 50mM
NaCl, 0.05% Tween-20 and 100nM okadaic acid, boiled
in 100ml of standard 1? SDS sample buffer and then
analyzed by western blotting. In experiments, where
streptavidin-coated sepharose was used, proteins re-
covered on the matrix were eluted by incubation at 37?C
for 15min (for HA-affinity matrix and Upf1-monoclonal
antibody) or at 4?C for 30min (for streptavidin-coated
sepharose) in the presence of 50ml washing buffer contain-
ing 1mg/ml HA-peptide, 1mg/ml Upf1 peptide or 2mM
biotin, boiled with 50ml of 2?sample buffer and then
analyzed by western blotting. Except for Figure 3,
anti-Upf1, P-T28 and P-S1078/1096 antibodies were
used as probes after 8% polyacrylamide gel electrophor-
esis. In Figure 3, samples were separated using a 8–16%
gradient polyacrylamide gel, which did not provide the
resolution required for the Upf1 mobility shift.
All western blot experiments were detected with the
ECL western blotting detection kit (GE Biotech) or with
Lumi-Light western blot substrate (Roche) and quantified
with a LuminoImager, LAS-3000 and Science Lab 2001
Image Gauge software (Fuji Photo Film). All experiments
were performed at least three times, and typical results are
NMD analysis in mammalian cells
NMD analysis was performed in essentially the same
manner as described elsewhere (42), except that HeLa
TetOff cells (1.8?106) were transfected with siRNAs
and plasmid vectors, which are indicated in figures, in
combination with 1.5mg of pTRE-BGG-WT or pTRE-
BGG-PTC. Briefly, 44h after transfection, doxycyclin
was added at time zero to inhibit de novo transcription.
Cells were then harvested at the indicated time points, and
total RNAs were analyzed by northern blotting using
either a b-globin or GAPDH (control) probe. The
quantities of BGG mRNA were normalized to GAPDH
signals and then graphically plotted. Values in figures rep-
resented the mean±SE for three to four independent
SMG-6 binds to phosphorylated Upf1 through its
To investigate the processes of NMD subsequent to Upf1
SMG-6, involved in the dephosphorylation of Upf1,
binds to phospho-Upf1. Immunoprecipitation experi-
ments using HeLa TetOff cell lysates and an anti-SMG-
6 antibody revealed that SMG-6 coprecipitated with Upf1
in the presence of RNaseA (Figure 1A). Note that the
coprecipitated Upf1 produced a doublet band, suggesting
the presence of phospho-Upf1 in the SMG-6 complex
(Figure 1A). To confirm this notion, we transfected cells
with HA-SBP-tagged SMG-6 and affinity purified the
SMG-6 complex using streptavidin-conjugated sepharose
beads. The binding between endogenous Upf1 and
exogenous wild-type SMG-6 is not strong; therefore, we
also analyzed the nuclease-inactive SMG-6-mtPIN and
-dCT mutants (Figure 1B). These mutants were designed
based on observations of the corresponding SMG-5
mutants, SMG-5-D860A and -dCT, which accumulate
the phospho-Upf1:SMG-5 complex in cells (30). As
shown in Figure 1C, the amount of endogenous Upf1
coprecipitated with SBP-tagged SMG-6 mutants was
greater than that with wild-type SMG-6. Intriguingly,
both mutants coprecipitated a doublet band of Upf1
similar to the endogenous SMG-6 precipitate. Highly
mobility shifted band, indicated by arrowhead, was
phosphorylated at least at serine (S) 1078 and 1096
residues, as detected by our phospho-Upf1 antibody,
which recognizes phosphorylated S1078/S1096 (anti-P-
S1078/S1096-Upf1) (Supplementary Figure S1) (20,30).
Note that coprecipitated phospho-Upf1 with SMG-6
is mostly in the mobility shifted (hyper-phosphorylated)
form, although the anti-P-S1078/S1096-Upf1 can detect a
doublet in total cell extract (Supplementary Figure S1).
These results indicate that, in vivo, SMG-6 preferentially
binds to hyper-phosphorylated-Upf1, as the SMG-5:
SMG-7 complex does (30).
The 14-3-3-like domain of SMG-7 can bind directly with
a phosphorylated amino acid residue (37). The presence of
a 14-3-3-like domain in SMG-6 suggests that SMG-6 can
also bind to phospho-Upf1 through this domain. To
evaluate this possibility, we made a mutant of SMG-6 con-
taining the amino acid changes, R656E, R737E and
Y738H, which are all located in the 14-3-3-like domain.
Note that the corresponding amino acid residues of the
SMG-7 14-3-3-like domain are required for direct
expected, both SMG-6-mt1433/PIN and -mt1433/dCT
double mutants showed greatly reduced coprecipitation
Nucleic Acids Research,2012, Vol.40, No. 31253
of phospho-Upf1 (Figure 1C). These results support the
notion that, in vivo, SMG-6 binds to phospho-Upf1 via its
14-3-3-like domain. These
coprecipitation of phospho-Upf1, but non-phospho-Upf1
was still precipitated at similar levels as wild-type, suggest-
ing the existence of additional phospho-independent
binding of SMG-6 to Upf1 (Figure 1C).
The coprecipitation of endogenous EJC components,
eIF4A3 and Magoh were also observed in endogenous
SMG-6 immunoprecipitates (Figure 1A). The N-terminal
region of SMG-6 contains two sequences homologous
to the EJC binding region of Upf3. To test the involve-
ment of the SMG-6 Upf3-homology region in EJC
binding, we tagged SMG-6-mtEJC with SBP and per-
formed affinity purification from cell extracts transfected
with SBP-tagged wild-type SMG-6 or SMG-6-mtEJC
(Supplementary Figure S2A and S2B). As expected,
Magoh, whereas wild-type SMG-6 did coprecipitate
eIF4A3 and Magoh (Supplementary Figure S2C). These
results suggest that SMG-6 binds to EJC components via
the Upf3-homology region, which is consistent with a
recent report (39).
Binding of SMG-6 to phospho-Upf1 is required for
NMD and dephosphorylation of Upf1
We next investigated the significance of the binding of
SMG-6 to phosho-Upf1. For this purpose, we expressed
containing b-globin mRNA reporter in HeLa TetOff
cells (Figure 2A) and analyzed the dominant-negative
effect of SMG-6-mt1433 on NMD. The cells were
treated with doxycycline to repress the transcription of
the reporter gene, and mRNAs extracted at indicated
time points were subjected to northern blot analysis to
estimate the half-life of the b-globin reporter mRNA.
Overexpression of wild-type SMG-6 had no effect on the
half-life of PTC-containing b-globin mRNA. In contrast,
overexpressed SMG-6-mt1433 prolonged the half-life of
PTC-containing b-globin mRNA without affecting that
of wild-type b-globin mRNA (Figure 2C and D, data
not shown). These results support the notion that
binding ofSMG-6 tophospho-Upf1
b-globin mRNA decay.
Genetic analysis in Caenorbabditis elegans revealed that
SMG-6 is required for the dephosphorylation of Upf1
(32). Therefore, we next investigated the physiological
significance of the binding of SMG-6 to phosho-Upf1
for dephosphorylation of Upf1. The transfection of
SMG-6-targeted siRNA induced the accumulation of
S1096-Upf1 antibody in total cell extracts (Figure 2E).
This phenotype was rescued by the expression of
siRNA-insensitive wild-type SMG-6 but not by the ex-
pression of SMG-6-mt1433 (Figure 2E). These results es-
tablish that the binding of SMG-6 to phospho-Upf1
SMG-6 or SMG-6-mt1433anda PTC-
by the anti-P-S1078/
anti HA (SMG-6)
AP: SBP tag
Figure 1. SMG-6 interacts with phosphorylated Upf1 via its 14-3-3-like domain. (A) Interactions between SMG-6 and components of the NMD
machinery. HeLa TetOff cell lysates were immunoprecipitated with anti-SMG-6 antibody in the presence of RNaseA. Immunoprecipitates (IP) were
analyzed by western blotting with the indicated antibody probes. (B) Schematic representation of SMG-6 mutants. The 14-3-3-like domain and PIN
domain are depicted as gray boxes. In SMG-6-mt1433, two arginines and one tyrosine in the 14-3-3-like domain were replaced by glutamic acid and
histidine, respectively (R656E, R737E and Y738H). In SMG-6-mtPIN, an aspartic acid in the PIN domain is replaced by alanine (D1251A).
SMG-6-mt1433/PIN has both mt1433 and mtPIN mutations. In SMG-6-dCT, the C-terminal 11 amino acids are deleted. SMG-6-mt1433/dCT
has both mt1433 and dCT mutations. (C) HeLa TetOff cell lysates were affinity purified with streptavidin sepharose in the presence of RNaseA. The
cell lysates (input) and affinity-purified (AP) fractions were analyzed by western blotting with the indicated antibody probes. Arrow head indicates
the hyper-phosphorylated isoform of Upf1 detected by our Upf1 and phospho-Upf1 antibodies.
1254Nucleic Acids Research, 2012,Vol.40, No. 3
codon104 codon105codon146codon30 codon30codon1
polyA addition signal
pTRE BGG PTC
mRNA remaining (%)
Figure 2. Phospho-dependent binding of SMG-6 to Upf1 is required for Upf1 dephosphorylation and PTC-containing b-globin mRNA decay.
(A) Schematic representation of the Tet-inducible human b-globin gene (BGG) reporter construct containing a PTC in exon 2. The open reading
frame (ORF) is represented by boxes and introns and UTRs by lines. (B and C) NMD is inhibited by overexpression of SMG-6 mutants. HeLa
TetOff cells were transfected with the plasmids shown above (B) or to the left of the panel (C). Total cell lysates were analyzed by western blotting
with indicated antibody probes (B). HeLa TetOff cells were co-transfected with the indicated plasmids and the Tet-inducible BGG-PTC plasmid.
After the addition of doxycycline to repress the transcription of the reporter plasmids, total RNAs, prepared at the times indicated, were analyzed by
northern blotting (C). (D) The quantities of BGG mRNA, normalized to GAPDH signals, were plotted. The means±SD from three independent
experiments are shown. (E) The 14-3-3-like domain of SMG-6 is required for Upf1 dephosphorylation. HeLa TetOff cells were transfected with
siRNA targeting SMG-6, together with the indicated siRNA-resistant SMG-6 plasmid vectors. The total cell lysates were analyzed by western
blotting with the indicated antibody probes. Valosin-containing protein (VCP) was used as a loading control.
Nucleic Acids Research,2012, Vol.40, No. 31255
through its 14-3-3-like domain is required for the
dephosphorylation of Upf1.
We also assessed the effect of SMG-6-mtPIN, -dCT and
-mtEJC mutants on NMD. The results indicate that
overexpression of mutants, SMG-6-mtPIN and -dCT but
not of -mtEJC, suppressed the decay of PTC-containing
b-globin mRNA (Figure 2C and D). Similarly, rescue ex-
periments showed that the -dCT mutant failed to rescue
phospho-Upf1, supporting the above notion (Figure 2E).
SMG-1-mediated phosphorylation of Upf1 at T28 is
required for association with SMG-6
The preferential binding of SMG-6 to phospho-Upf1
prompted us to identify the phosphorylated amino
acid residue(s) of Upf1 responsible for SMG-6 binding.
Both the N- and C-terminal regions of Upf1 have
phosphorylated by SMG-1 in vitro (Supplementary
Figure S3) (22). Based on these observations, we created
phospho-resistant mutants of SBP-tagged Upf1, which do
not contain the putative phosphorylation sites. The first
mutant, Upf1-dCT, lacked the C-terminal 90 amino acids,
residues 1028–1118. The second mutant, Upf1-dNCT,
lacked the N-terminal63
C-terminal 90 amino acids, and the third mutant,
Upf1-5S/TA, harbored alanine (A) substitutions at T28,
S1073, S1078,S1096 and
Immunoprecipitation analysis of endogenous SMG-6 in
the presence of RNaseA from HeLa TetOff cell extracts
transfected with HA-tagged wild-type Upf1, or the Upf1
mutants, -dCT, -dNCT or -5S/TA showed that endogen-
ous SMG-6 precipitated exogenously expressed wild-type
Upf1 and the Upf1-dCT mutant. On the other hand,
SMG-6 precipitated only small amounts of Upf1-dNCT
conserved S/TQresidues,which are
aminoacids and the
Figure 3. Threonine residue 28 of Upf1 is necessary for the binding of SMG-6. (A) Schematic representation of Upf1 mutants. The N-terminal
conserved region (NCR), cystein rich region, helicase domains and SQ-rich region of Upf1 are depicted as gray boxes. In Upf1-dCT, the C-terminal
90 amino acids are deleted. In Upf1-dNCT, the N-terminal 63 amino acids and the C-terminal 90 amino acids are deleted. In Upf1-5S/TA, T28,
S1073, S1078, S1096 and S1116 are replaced by A. In Upf1-T28A, T28 is substituted for A. In Upf1-4SA, S1073, S1078, S1096 and S1116 are
replaced by A. (B and C) HeLa TetOff cells were transfected with the plasmids shown above. The cell extracts were immunoprecipitated with
anti-SMG-6 antibody or with normal rabbit IgG (NRIgG) in the presence of RNaseA. Immunoprecipitated fractions were analyzed by western
blotting with the indicated antibody probes. ‘Asterisks’ indicates unexpected degradation product of HA-tagged-Upf1. ‘Hash’ indicates an
uncharacterized band detected by the anti-HA antibody in Upf1-dNCT mutant-expressed extract.
1256 Nucleic Acids Research, 2012,Vol.40, No. 3
or -5S/TA mutants compared with wild-type (Figure 3B).
Similar experiments using two additional mutants of
Upf1, Upf1-T28A, in which T28 was substituted by A,
and Upf1-4SA, in which S1073, S1078, S1096 and S1116
were substituted by A, revealed that endogenous SMG-6
precipitated only a faint amount of Upf1-T28A, whereas it
precipitated similar amounts of Upf1-4SA and wild-type
Upf1 (Figure 3C). These results demonstrate that T28 of
Upf1 is needed for SMG-6 binding.
Preferential binding of SMG-6 to phospho-Upf1
(Figure 1) and the requirement of Upf1 T28 for this
binding (Figure 3), in addition to the observation that a
GST-fused peptide derived from the region of Upf1 T28 is
efficiently phosphorylated by SMG-1 in vitro (22), suggest
that, in vivo, SMG-1 phosphorylates T28 of Upf1 to create
an SMG-6 binding site. To confirm this, we generated a
phosphorylated-T28 (P-T28-Upf1) and analyzed the re-
activity to wild-type Upf1 and to Upf1 mutants -T28A
and -2SA, in which S1078 and S1096 were substituted
by A (Figure 4A). Anti-P-T28-Upf1 antibody recognized
Upf1-2SA, whereas it failed to detect the mutant,
Upf1-T28A (Figure 4B). In addition, anti-P-T28-Upf1
detected an endogenous Upf1 phosphorylation signal
(Figure 4C). Furthermore, SMG-1-targeted siRNA treat-
ment resulted in a decrease in anti-P-T28-Upf1 signal
(Figure 4C). These results indicate that SMG-1 phosphor-
ylates T28 of Upf1 in vivo, which was recognized by our
anti-P-T28-Upf1 antibody. SMG-5, SMG-6 and SMG-7
are involved in the PP2A-mediated dephosphorylation of
Upf1 at S1078 and S1096 in mammals (22,30,32). To
dephosphorylation of the newly identified Upf1 phosphor-
ylation site, T28, we analyzed P-T28-Upf1 signals in
response to treatment with okadaic acid (OA), a PP2A
inhibitor, knockdown of SMG-5, SMG-6 or SMG-7, or
the overexpression of SBP-tagged SMG-6 mutants. We
observed an increasein
anti-P-T28-Upf1 signals after OA treatment (Figure 4D)
and knockdown of SMG-5, SMG-6 or SMG-7 (Figure
4E), similar to the P-S1078/S1096 signals. Note that the
amount of SMG-5 is slightly decreased in SMG-7
knocked-down and showed an unexpected mobility shift
to the lower molecular weight side (Figure 4E). The over-
production of SMG-6-mt1433, -mtPIN and -dCT, but not
of wild-type or -mtEJC, resulted in the accumulation of
P-T28-Upf1 signals (Figure 4F). These results suggest
that, similar to residues S1078/S1096, T28 of Upf1 is
dephosphorylated by PP2A in an SMG-5, SMG-6 and
SMG-1-mediated phosphorylation of Upf1 at S1096 is
required for binding to SMG-5 and SMG-7
Although the SMG-5:SMG-7 complex also preferentially
binds to phospho-Upf1, the exact binding site(s) of them
are unknown (30). To identify the Upf1 residue(s) respon-
sible for SMG-5:SMG-7 complex binding, we made three
SBP-tagged Upf1 mutants, Upf1-S1078A, Upf1-S1096A
and Upf1-S1116A, which harbored alanine substitutions
at S1078, S1096 or S1116, respectively (Figure 5A). These
mutants were chosen for the following two reasons:
(i) SMG-7 bindingtophospho-Upf1
decreased in the Upf1-4SA mutant (19) and (ii) S1078,
S1096 and S1116 residues of Upf1 are phosphorylated
in vivo (22,30,31). Affinity purification of SBP-tagged
Upf1 in the presence of RNaseA from cell extracts
revealed that the S1096A mutant failed to coprecipitate
endogenous SMG-5 or SMG-7, whereas mutants S1078
and S1116 coprecipitated equivalent amounts of SMG-5
and SMG-7 compared with wild-type Upf1 (Figure 5B). A
similar experiment preformed using an SBP-tagged
Upf1-T28A mutant showed that no apparent effect on
SMG-5 or SMG-7 coprecipitation was observed for the
Upf1-T28A mutant (Figure 5C). These results indicate
that phosphorylated S1096 of Upf1 is integral to the
binding site of the SMG-5:SMG-7 complex. Although
both of SMG-5 and SMG-7 have a 14-3-3-like domain,
only one serine residue, S1096, is responsible for the
SMG-5:SMG-7 complex binding. To investigate which
immunoprecipitated endogenous Upf1 from cell extracts
transfected with siRNA targeted against SMG-5 or
SMG-7 in the presence of RNaseA. As shown in
Figure 5D, depletion of SMG-7 decreased the SMG-5
coprecipitation compared with a non-silencing control.
On the other hand, SMG-5 knockdown do not signifi-
cantly decrease the coprecipitation of SMG-7. This
suggests that SMG-7 links the SMG-5:SMG-7 complex
to phospho-S1096 of Upf1.
inthis binding, we
Phosphorylation at both T28 and S1096 of Upf1 is
required for NMD
As described above, T28 and S1096 of Upf1 are
required for SMG-6 and the SMG-5:SMG-7 complex, re-
spectively, to bind to Upf1. If the binding of either of
these factors is necessary for NMD to proceed in vivo,
overproduction of mutant Upf1-T28A or -S1096A might
inhibit NMD in a dominant-negative manner. To analyze
this, we expressed wild-type Upf1 or the Upf1 mutants,
-T28A, -S1078A, -S1096A or -S1116A in HeLa TetOff
analysis, as described in Figure 2. Overexpression of
wild-type Upf1 or Upf1 mutants -S1078A or -S1116A
had no affect on the half-lives of PTC-containing
b-globin mRNA (Figure 6B and C). This result indicates
that phosphorylation of both T28 and S1096 of Upf1 is
required for PTC-dependent degradation of b-globin
Distinct phospho-Upf1 complexes are formed in SMG-5
or SMG-6 inactivated cell extracts
Depletion of NMD components or expression of specific
NMD component mutants enabled us to ‘freeze’ the tran-
sient protein complex formed during the process of PTC
recognition and allowed us to suggest the sequential
Nucleic Acids Research,2012, Vol.40, No. 3 1257
Figure 4. SMG-1-mediated phosphorylation of Upf1 threonine residue 28. (A) Schematic representation of Upf1 mutants. The NCR, cystein rich
region, helicase domains and SQ-rich region are depicted by gray boxes. In Upf1-T28A, T28 is substituted for A. In Upf1-2SA, S1078 and S1096 are
replaced by A. (B) The anti-P-T28-Upf1 antibody recognizes T28 phosphorylation of Upf1 in vivo. HeLa TetOff cells were transfected with the
plasmids indicated above the blot. Total cell extracts were probed with the antibodies indicated. (C) The anti-P-T28-Upf1 antibody detects
SMG-1-mediated endogenous-Upf1 phosphorylation. Cells were transfected with the siRNAs shown above. Total cell extracts were probed with
the antibodies indicated. (D) OA-sensitive phosphatase-mediated dephosphorylation of Upf1 P-T28. Total extracts of cells treated for 4h with 50nM
OA were analyzed as above. (E and F) Knockdown of SMG-5, SMG-6 or SMG-7, or overexpression of SMG-6 mutants accumulates T28
phosphorylated Upf1. Cells were transfected with the siRNAs or SMG-6 plasmids indicated above the blot. Total cell extracts were probed with
the antibodies indicated. To estimate the knockdown efficiency, 100, 50, 25 and 12.5% (B) or 100, 33 and 11% (E) of NS control samples were
1258Nucleic Acids Research, 2012,Vol.40, No. 3
mRNP remodeling of the mRNA surveillance complex
SMG-5:SMG-7 complex and SMG-6 to phospho-Upf1
and their essential role for NMD prompted us to
capture the post-Upf1 phosphorylation process of NMD
by the inactivation of these proteins. For this, we
immunoprecipitated endogenous Upf1 from cell extracts
transfected with siRNA targeted against SMG-5 or
SMG-6 in the presence of RNaseA and cycloheximide,
which stabilizes 80S ribosomes on mRNAs in vitro and
prevents their dissociation during immunoprecipitation
(45) during cell lysis. The Upf1 immune complexes
eluted with the Upf1-peptide antigen contained eRF1,
EJC components (Upf3b, Y14 and Magoh), rpS16 and
SMG-7. Compared with a non-silencing control, depletion
of either SMG-5 or SMG-6 caused an accumulation of
phospho-Upf1 in the total cell extracts and in the Upf1
immune complex (Figure 4E; data not shown). Depletion
of SMG-5 enhanced the amounts of eRF1, Upf3b
and EJCcomponents (eIF4A3,
coprecipitated with Upf1, whereas the amounts of
coprecipitated rpS16 and SMG-7 were not apparently
altered. In contrast, SMG-6 depletion decreased the
coprecipitation of rpS16 and eRF1 with Upf1, although
it increased coprecipitation of SMG-5, SMG-7 and EJC
components (eIF4A3, Y14 and Magoh) (Figure 7A). We
failed to detect SMG-6 coprecipitation in this experimen-
tal condition, most likely because of the low binding
affinity of the anti-SMG-6 antibody.
Next, we tested whether the phospho-Upf1 complexes
enriched in SMG-5 or SMG-6 depleted cell extracts are
formed on mRNPs. For this, we immunopurified en-
dogenous Upf1 from cells transfected with siRNAs target-
ing SMG-5 or SMG-6 in the presence or absence of
RNaseA. RNaseA sensitive coprecipitation of CBP80, a
nuclear cap binding protein and of PABPC1/4, cytoplas-
mic poly(A) binding proteins, was observed in the Upf1
immunoprecipitates from SMG-5- or SMG-6-depleted cell
extracts, suggesting that the phospho-Upf1 complexes are
formed on mRNPs (Figure 7B). Our model to explain
these observations is depicted in Figure 7C.
To further evaluate the distinct processes regulated by
SMG-5 or SMG-6, we took advantage of the dominant-
negative mutants of SMG-5, SMG-5-dCT and of SMG-6,
SMG-6-dCT (Figure 8A). SMG-5-dCT preferentially
binds to phospho-Upf1 and inhibits Upf1 dephosp-
horylation and NMD (30). SMG-6-dCT also showed a
similar phenotype (Figures 1 and 2). We immunopre-
cipitated endogenous Upf1 from cell extracts transfected
with wild-type HA-tagged-SMG-5, HA-tagged-SMG-
5-dCT, wild-type SMG-6 or HA-tagged-SMG-6-dCT in
the presence of RNaseA and cycloheximide. Compared
with the wild-type, expression of either SMG-5-dCT or
phospho-Upf1 in cell extracts and in the Upf1 immune
complex and showed greater levels of SMG-5dCT and
SMG-6-dCT coprecipitation with Upf1 (Figure 8B; data
not shown). Overexpression of the SMG-6-dCT mutant
increased the coprecipitation of endogenous SMG-5 and
SMG-7, and reduced the coprecipitation of ribosomal
proteins, rpL7a and rpS16, eRF1 and Upf3b with Upf1
Y14 and Magoh)
AP: SBP tag
AP: SBP tag
Figure 5. Upf1 serine residue 1096 is necessary for the binding of the
mutants. The NCR, cystein rich region, helicase domains and SQ-rich
region are depicted by gray boxes. In Upf1-S1078A, S1078 is
substituted for A. In Upf1-S1096, S1096 is replaced by A. In
Upf1-S1116A, S1116 is substituted for A. In, Upf1-T28A, T28 is
replaced by A. (B and C) HeLa TetOff cells were transfected with
the indicated plasmids or an empty vector (‘vector’). The cell extracts
were purified with streptavidin sephorose in the presence of RNaseA.
Purified fractions were analyzed by western blotting with the antibodies
indicated. (D) HeLa TetOff cells were transfected with the indicated
siRNAs. The cell extracts were immunoprecipitated with anti-Upf1
(5C3) antibody in the presence of RNaseA. The immunoprecipitates
(IP) were analyzed by western blotting with the antibodies shown on
the left of the panels.
Nucleic Acids Research,2012, Vol.40, No. 3 1259
compared with wild-type SMG-6 (Figure 8C). On the
other hand, overexpression of the SMG-5dCT mutant
did not alter the coprecipitation of these proteins except
for the increased coprecipitation of endogenous SMG-7
compared with wild-type SMG-5 (Figure 8C). No
apparent alteration was observed for the coprecipitation
of EJC components, Y14, Magoh and eIF4A3 in these
coprecipitated with Upf1
RNaseA, suggesting that the phospho-Upf1 complexes
with SMG-5dCT or SMG-6-dCT are formed on mRNPs
(Supplementary Figure S4). Taken together with the
knockdown experiment presented above, these results
indicate that the phospho-Upf1 complex containing ribo-
inactivated SMG-5, but not with the inactivated SMG-6.
The decrease of Upf3b in the Upf1 complex accumulated
with SMG-6-dCT might reflect competitive binding of
SMG-6 and Upf3b to phospho-Upf1. Our model is
depicted in Figure 8D.
Here, we demonstrated that the SMG-1-mediated phos-
phorylation of Upf1 creates at least two distinct
phospho-specific binding platforms for SMG-6 and
required for the mRNP remodeling to dissociate Upf1
from mRNA and NMD. Further, the binding of
pohospho-Upf1 leads the dissociation of ribosome and
release factors from DECID complex. Figure 9 illus-
trates our view of the simultaneous recruitments of the
but not SMG-6,to
mRNA remaining (%)
Figure 6. Phosphorylation of Upf1 at both T28 and S1096 is required for PTC-containing b-globin mRNA decay. (A and B) HeLa TetOff cells were
transfected with the plasmids shown above (A) or to the left of the panel (B). Total cell lysates were analyzed by western blotting with the indicated
antibody probes (A). NMD is inhibited by overexpression of Upf1-T28A and -S1096 mutants. HeLa TetOff cells were co-transfected with the
Tet-inducible b-globin (BGG)-PTC plasmid shown in Figure 2A and with Upf1 plasmids indicated on the left of the panel. After the addition of
doxycycline to repress the transcription of the BGG reporter plasmid, total RNAs prepared at the times indicated were analyzed by northern blotting
(B). (C) The quantities of BGG mRNA, normalized to GAPDH signals, were plotted. The means±SD from three independent experiments are
1260Nucleic Acids Research, 2012,Vol.40, No. 3
In the present study, we have identified T28 as a novel
SMG-1-mediated in vivo phosphorylation site of Upf1
(Figure 4). The sequence surrounding T28 is conserved
among higher eukaryotes, including plants, supporting
(Supplementary Figure S3). An okadaic acid sensitive
phosphatase, PP2A, dephosphorylates P-T28 of Upf1
according to SMG-5, SMG-6 and SMG-7 dependent
mechanisms (Figure 4E), which is similar to other
S1096 (Figure 4D and E). However, ATPase-deficient
Upf1 is highly phosphorylated (19) and binds efficiently
with SMG-5, SMG-6 and SMG-7 (44). These observations
suggest that the binding of SMG-5, SMG-6 and SMG-7
to phospho-Upf1 is not
dephosphorylation even though they bind to PP2A
(30,33,35). Future studies will reveal the mechanism of
What is the role of T28 phosphorylation? The following
observations provide evidence that T28 phosphorylation
induces the binding of SMG-6 via its 14-3-3-like domain
during NMD. First, the binding of SMG-6 to Upf1 was
strongly diminished in a Upf1-T28A mutant (Figure 3C).
that leadto the remodelingof
Second, the SMG-6 14-3-3-like domain mutation abol-
ished the coprecipitation of phospho-Upf1 (Figure 1C).
dominant-negative manner (Figure 2 and 6). Another
Upf1 phosphorylation site, at S1096, whose phosphoryl-
ation is mediated by SMG-1, formed the binding site for
the SMG-5:SMG-7 complex (Figure 5B). SMG-7 is pre-
sumably responsible for the binding of the SMG5:SMG7
complex to phospho-S1096 (Figure 5D). However, it
remains unclear why we failed to detect an increase of
SMG-7 coprecipitation with phospho-Upf1 in SMG-5
knockdown cells, whereas SMG-5 knockdown accumu-
lates the phosphorylation of S1096 in a similar manner
to SMG-6 knockdown that induces accumulation of the
SMG-5:SMG-7 complex with phospho-Upf1 (Figure 7)
Hetero dimer formation might be required for the full
binding activity of the SMG-5:SMG-7 complex. The
dominant-negative effect of the Upf1-S1096A mutant
suggested that this phospho-mediated binding of the
(Figure 6). The involvement of two different phosphoryl-
ation sites for the SMG-5:SMG-7 complex and SMG-6
also suggests that they can simultaneously bind to
phospho-Upf1, although they have been expected to
RNaseA: - + - + - + - +
Upf1 complex without SMG-5 Upf1 complex without SMG-6
Figure 7. SMG-5 or SMG-6 knockdown causes the accumulation on mRNP of phospho-Upf1 complexes containing different components. (A and
B) HeLa TetOff cells were transfected with the indicated siRNAs. The cell extracts were immunoprecipitated with anti-Upf1 (5C3) antibody in the
presence of RNaseA (A), or in the absence or presence of RNaseA (B). The cell extracts (Input) and immunoprecipitates (IP) were analyzed by
western blotting with the antibodies shown on the left of the panels. (C) Schematic presentation of Upf1 complexes following SMG-5 or SMG-6
Nucleic Acids Research,2012, Vol.40, No. 31261
SMG-5:SMG-7 complex accumulated in the phospho-
Upf1 complex together with the SMG-6dCT mutant
(Figure 8C). These results indicate that simultaneous
binding of the SMG-5:SMG-7 complex and SMG-6 to
phospho-Upf1 are required for NMD. This is consistent
with the observation that SMG-5, SMG-6 and SMG-7 are
(2,3). Consistentwiththis notion,the
Although SMG-5, SMG-6 and SMG-7 share 14-3-3-like
domains (30,37), different phospho-S/T binding properties
between the SMG-5:SMG-7 complex and SMG-6 exist;
while the SMG5:SMG-7 complex binds to P-S1096 but
not to P-T28 (Figure 5B and C), SMG-6 binds to P-T28
but not to P-S1096 (Figure 3C). Upf1-S1078 is dispensable
in vivo, although SMG-7 can bind phospho-S1078 in vitro.
Heterodimer formation of the SMG-5:SMG-7 complex
Upf1 with SMG-5-dCT complexUpf1 with SMG-6-dCT complex
Figure 8. Overexpression of SMG-5-dCT or SMG-6-dCT causes the accumulation of phospho-Upf1 complexes containing different components.
(A) Schematic representation of SMG-5dCT and SMG-6dCT mutants. The 14-3-3-like domain and PIN domain are depicted in gray boxes.
SMG-5dCT and SMG-6dCT have deletions of the C-terminal 23 or 11 amino acids, respectively. (B) Overexpression of SMG-5 or SMG-6
mutants results in the accumulation of Upf1 phosphorylated at T28 and S1078/S1096. Cells were transfected with the SMG-5 or SMG-6
plasmids indicated above the blot. Total cell extracts were probed with the antibodies indicated. (C) HeLa TetOff cells were transfected with the
indicated plasmids. The cell extracts were immunoprecipitated with anti-Upf1 (5C3) antibody in the presence of RNaseA. (D) Schematic presentation
of Upf1 complexes following overexpression of SMG-5-dCT or SMG-6-dCT.
1262 Nucleic Acids Research, 2012,Vol.40, No. 3
might affect their specificity in vivo. Future studies, such as
structural analysis, willclarify their specificity
phospho-S/TQ motifs. SMG-6 can associate with the
phospho-Upf1, but also through its ability to directly
bind the EJC (39) (Supplementary Figure S2) and/or via
direct RNA binding (47). These associations might be
involved in EJC recycling, similar to PYM (48), or in
some specific mRNA degradation (39).
While S1078 and S1116 of Upf1 are phosphorylated
in vivo (22,30,31), they are likely to be dispensable for
NMD (Figure 6). It is possible that they are involved
with the other SQ-directed kinases, such as ATR, ATM
and DNA-PKcs, which mediate phosphorylation events
for various cellular functions of Upf1, such as histone
mRNA decay or genome stability (49–51).
Our results presented in this study revealed that inacti-
vation of either SMG-5 or SMG-6 induces accumulation
of distinct phospho-Upf1 complexes on mRNP together
with CBP80 (cap) and PABPC1/C4 [poly(A)] (Figures 7
and 8; Supplementary Figure S4). This suggests that
phospho-specific binding of the SMG-5:SMG-7 complex
and of SMG-6 to Upf1 are required for mRNP remodel-
ing to dissociate Upf1 from mRNA and to promote
On the other hand, the complexes accumulated by
SMG-5 or SMG-6inactivation
instance, SMG-5 knockdown and overexpression of
phospho-Upf1 complex containing ribosome, release
factor and EJC (Figures 7 and 8). Importantly, the
complex containing ribosomal
factor. These results indicate that SMG-5 binding
to phospho-Upf1 is required but not sufficient for the
dissociation of ribosome and release factor from the
DECID complex. The C-terminal PIN domain of
SMG-5 is required for this action and SMG-5 might be
proteins and release
SMG-1 phosphorylates Upf1 in DECID
The SMG-5:SMG-7 complex
binds to phospho-S1096 of Upf1
SMG-6 binds to
phospho-T28 of Upf1
Upf1 dissociation from PTC-mRNA
Figure 9. Model depicting NMD from post-Upf1 phosphorylation to mRNA decay. PTC recognition is established by the formation of mRNA
surveillance complexes called ‘DECIDs’ that contain a PTC-recognizing ribosome, eRF, Upf1, SMG1C (SMG-1:SMG-8; SMG-9 complex) and the
EJC during the initial round of translation. DECID formation induces SMG-1–mediated Upf1 phosphorylation. The SMG-5:SMG-7 complex binds
to phospho-S1096 of Upf1 to dissociate the ribosome and release factor from Upf1. SMG-6 binds to phospho-T28 of Upf1 to induce Upf1
dissociation from the mRNA. AUG, start codon; Ter, termination codon; eRF, eRF1:eRF3 complex; P, phosphate group; S5:7, the
SMG-5:SMG-7 complex, S6, SMG-6.
Nucleic Acids Research,2012, Vol.40, No. 31263
involved in ribosome dissociation by recruitment and/or
activation of ribosome dissociation factors, such as eEF2,
eIF3 and/or ABCE1 (52,53), through its PIN domain
In contrast to SMG-5, SMG-6 knockdown and
overexpression of SMG-6-dCT resulted in the accumula-
tion of phospho-Upf1 complex containing EJC and the
SMG-5:SMG-7 complex, but not ribosomal proteins or
release factors (Figures 7 and 8). These results suggested
that, SMG-6 is likely to be dispensable for ribosome and
release factor dissociation. However, SMG-6 and its PIN
domain are still required for Upf1 dephosphorylation,
NMD and Upf1 dissociation from mRNA (Figure 8 and
Supplementary Figure S4). These results suggest that
ribosome and release factor dissociation from the Upf1
complex precedes but is not sufficient to promote Upf1
dephosphorylation, NMD and probably Upf1 dissoci-
ation from mRNA. Because ATPase activity of Upf1 is
required for the dissociation of Upf1 from mRNA (44)
(Supplementary Figure S5), the SMG-6 PIN domain
might be involved in Upf1 ATPase activation. It is also
endocleavage of PTC-mRNA is a prerequisite for Upf1
ATPase activation. In addition, even though ATP
SMG-6, SMG-7 and mRNA decay enzymes (19,44,54),
they fail to efficiently promote mRNA decay (19,32,44).
Consistent with this finidng, the RNase A sensitive accu-
mulation of CBP80, an indicator of intact cap structure, is
observed in Upf1-KQ mutant precipitate (Supplementary
Figure S5). Taken together with the RNaseA sensitive
coprecipitation of CBP80 with Upf1 in SMG-5 or
SMG-6 inactivated conditions, these results suggested
that both decapping and endocleavage do not occur
before (i) SMG-5, SMG-6 and SMG-7 have all bound to
Upf1 and (ii) Upf1 ATPase activation. Upf1 ATPase
activity might regulate decapping enzyme and SMG-6
endonuclease activation (Figure 9). Further reconstitution
analysis is required for resolving these issues.
N-terminal conserved region of Upf1 containing residue
T28 in phospho-independent manner (30). Thus, it is
possible that the SMG-5:SMG-7 complex might either
inhibit the SMG-1-mediated phosphorylation of T28 of
Upf1 orinhibitthe binding
phosphorylated T28 residue. One possible explanation
that the phospho-spesific binding of SMG6 to Upf1
might occur after the SMG-5:SMG-7 complex are seques-
tered by the C-terminal
(Figure 9). Simultaneous binding of the SMG-5:SMG-7
complex and SMG-6 to N-terminal conserved region of
Upf1 is also possible since Upf1-4SA mutants, which lose
the pohspho-S1096 binding
complex can bind SMG-6 (Figure 3C).
EJC components that trigger Upf1 phosphorylation
NMD, although the mechanism is unknown, (27,29)
SMG-1 and its co-factors are also needed for the
binds with SMG-5,
simultaneous binding of the SMG-5:SMG-7 complex
and SMG-6 to phospho-Upf1 described in this study
seems to be generally involved in mammalian NMD.
Even though Upf1 sites phosphorylated by SMG-1 are
conserved, SMG-5, SMG-6 and SMG-7 are highly diverse
among higher eukaryotes. For instance, vertebrates and
nematodes have all these molecules, while flies do not
have SMG-7, and plants have two non-redundant
SMG-7s, but not SMG-5 and SMG-6 (30,33,36,55).
These differences might reflect the distinct mRNA decay
mechanisms among these species. In particular, flies
mainly use an SMG-6-mediated endocleavage pathway,
and mammals use both endonucleolytic cleavage and
deadenylation-initiated decapping pathways (40,41,56).
No endonucleolytic cleavage of PTC-mRNA has been
reported in plants, which is consistent with the absence
of SMG-6 endonuclease (57). Intriguingly, at least two
SMG-5/6/7 related molecules exist among higher eukary-
otes and the two phosphorylation sites of Upf1, which are
essential for binding with them are conserved among
higher eukaryotes. These SMG-5/6/7 related molecules
might have similar sequence and functional specificity to
that of mammals.
Supplementary Data are available at NAR Online:
Supplementary Figures S1–S5.
We thank Ms Yumi Bamba, Ms Reiko Muramatsu,
Dr Tetsuo Ohnishi and Dr Isao Kashima for their tech-
performed experiments, analyzed data and wrote the
article; K.K. and N.I. performed the experiments; F.H.
supervised the project; S.O. supervised the project and
participated in the preparation of the article.
The Japan Society for the Promotion of Science (to S.O.
and N.I.); the Japan Science and Technology Corporation
(to A.Y. and S.O.); the Ministry of Education, Culture,
Sports, Science and Technology of Japan [a Grant-in-Aid
for Scientific Research (A) (to S.O.), Young Scientists (A)
(to A.Y.), Scientific Research on Innovative Areas ‘RNA
regulation’ (to A.Y.) and
Innovative Areas ‘Functional machinery for non-coding
RNAs’ (to A.Y.)]; Takeda Science Foundation (to S.O.);
Mitsubishi Foundation (to S.O.), Uehara Memorial
Foundation (to S.O.) and the Yokohama Foundation
for Advancement of Medical Science (to A.Y.). Funding
Technology Corporation (to A.Y.)
Conflict of interest statement. None declared.
1264 Nucleic Acids Research, 2012,Vol.40, No. 3
1. Chang,Y.F., Imam,J.S. and Wilkinson,M.F. (2007) The
nonsense-mediated decay RNA surveillance pathway.
Annu. Rev. Biochem., 76, 51–74.
2. Nicholson,P., Yepiskoposyan,H., Metze,S., Zamudio Orozco,R.,
Kleinschmidt,N. and Muhlemann,O. (2010) Nonsense-mediated
mRNA decay in human cells: mechanistic insights, functions
beyond quality control and the double-life of NMD factors.
Cell. Mol. Life Sci., 67, 677–700.
3. Hwang,J. and Maquat,L.E. (2011) Nonsense-mediated mRNA
decay (NMD) in animal embryogenesis: to die or not to die, that
is the question. Curr. Opin. Genet. Dev, 21, 422–430.
4. Mendell,J.T., Sharifi,N.A., Meyers,J.L., Martinez-Murillo,F. and
Dietz,H.C. (2004) Nonsense surveillance regulates expression of
diverse classes of mammalian transcripts and mutes genomic
noise. Nat. Genet., 36, 1073–1078.
5. Weischenfeldt,J., Damgaard,I., Bryder,D., Theilgaard-Monch,K.,
Thoren,L.A., Nielsen,F.C., Jacobsen,S.E., Nerlov,C. and
Porse,B.T. (2008) NMD is essential for hematopoietic stem and
progenitor cells and for eliminating by-products of programmed
DNA rearrangements. Genes Dev., 22, 1381–1396.
6. Usuki,F., Yamashita,A. and Fujimura,M. (2011)
Post-transcriptional defects of antioxidant selenoenzymes cause
oxidative stress under methylmercury exposure. J. Biol. Chem.,
7. Welch,E.M., Barton,E.R., Zhuo,J., Tomizawa,Y., Friesen,W.J.,
Trifillis,P., Paushkin,S., Patel,M., Trotta,C.R., Hwang,S. et al.
(2007) PTC124 targets genetic disorders caused by nonsense
mutations. Nature, 447, 87–91.
8. Usuki,F., Yamashita,A., Higuchi,I., Ohnishi,T., Shiraishi,T.,
Osame,M. and Ohno,S. (2004) Inhibition of nonsense-mediated
mRNA decay rescues the phenotype in Ullrich’s disease.
Ann. Neurol., 55, 740–744.
9. Usuki,F., Yamashita,A., Kashima,I., Higuchi,I., Osame,M. and
Ohno,S. (2006) Specific inhibition of nonsense-mediated mRNA
decay components, SMG-1 or Upf1, rescues the phenotype of
ullrich disease fibroblasts. Mol. Ther., 14, 351–360.
10. Pastor,F., Kolonias,D., Giangrande,P.H. and Gilboa,E. (2010)
Induction of tumour immunity by targeted inhibition of
nonsense-mediated mRNA decay. Nature, 465, 227–230.
11. Linde,L. and Kerem,B. (2008) Introducing sense into nonsense in
treatments of human genetic diseases. Trends Genet., 24, 552–563.
12. Bhuvanagiri,M., Schlitter,A.M., Hentze,M.W. and Kulozik,A.E.
(2010) NMD: RNA biology meets human genetic medicine.
Biochem. J., 430, 365–377.
13. Le Hir,H., Izaurralde,E., Maquat,L.E. and Moore,M.J. (2000)
The spliceosome deposits multiple proteins 20-24 nucleotides
upstream of mRNA exon-exon junctions. EMBO J., 19,
14. Anderson,P. (2006) NMD in Caenorbabditis elegans.
In Maquat,L. (ed.), Nonsense-Mediated mRNA Decay. Landes
bioscience, Austin, TX, pp. 139–150.
15. Culbertson,M.R. and Leeds,P.F. (2003) Looking at mRNA
decay pathways through the window of molecular evolution.
Curr. Opin. Genet. Dev., 13, 207–214.
16. Ishigaki,Y., Li,X., Serin,G. and Maquat,L.E. (2001)
Evidence for a pioneer round of mRNA translation:
mRNAs subject to nonsense-mediated decay in
mammalian cells are bound by CBP80 and CBP20. Cell, 106,
17. Thermann,R., Neu-Yilik,G., Deters,A., Frede,U., Wehr,K.,
Hagemeier,C., Hentze,M.W. and Kulozik,A.E. (1998) Binary
specification of nonsense codons by splicing and cytoplasmic
translation. EMBO J., 17, 3484–3494.
18. Yamashita,A., Kashima,I. and Ohno,S. (2005) The role
of SMG-1 in nonsense-mediated mRNA decay. Biochim.
Biophys. Acta, 1754, 305–315.
19. Kashima,I., Yamashita,A., Izumi,N., Kataoka,N., Morishita,R.,
Hoshino,S., Ohno,M., Dreyfuss,G. and Ohno,S. (2006)
Binding of a novel SMG-1-Upf1-eRF1-eRF3 complex (SURF)
to the exon junction complex triggers Upf1
phosphorylation and nonsense-mediated mRNA decay. Genes
Dev., 20, 355–367.
20. Yamashita,A., Izumi,N., Kashima,I., Ohnishi,T., Saari,B.,
Katsuhata,Y., Muramatsu,R., Morita,T., Iwamatsu,A., Hachiya,T.
et al. (2009) SMG-8 and SMG-9, two novel subunits of the
SMG-1 complex, regulate remodeling of the mRNA surveillance
complex during nonsense-mediated mRNA decay. Genes Dev., 23,
21. Izumi,N., Yamashita,A., Iwamatsu,A., Kurata,R., Nakamura,H.,
Saari,B., Hirano,H., Anderson,P. and Ohno,S. (2010) AAA+
proteins RUVBL1 and RUVBL2 coordinate PIKK activity and
function in nonsense-mediated mRNA decay. Sci. Signal, 3, ra27.
22. Yamashita,A., Ohnishi,T., Kashima,I., Taya,Y. and Ohno,S.
(2001) Human SMG-1, a novel phosphatidylinositol
3-kinase-related protein kinase, associates with components of
the mRNA surveillance complex and is involved in the
regulation of nonsense-mediated mRNA decay. Genes Dev., 15,
23. Fernandez,I.S., Yamashita,A., Arias-Palomo,E., Bamba,Y.,
Bartolome,R.A., Canales,M.A., Teixido,J., Ohno,S. and Llorca,O.
(2011) Characterization of SMG-9, an essential component of
the nonsense-mediated mRNA decay SMG1C complex.
Nucleic Acids Res., 39, 347–358.
24. Arias-Palomo,E., Yamashita,A., Fernandez,I.S., Nunez-
Ramirez,R., Bamba,Y., Izumi,N., Ohno,S. and Llorca,O. (2011)
The nonsense-mediated mRNA decay SMG-1 kinase is regulated
by large-scale conformational changes controlled by SMG-8.
Genes Dev., 25, 153–164.
25. Grimson,A., O’Connor,S., Newman,C.L. and Anderson,P. (2004)
SMG-1 is a phosphatidylinositol kinase-related protein kinase
required for nonsense-mediated mRNA Decay in Caenorhabditis
elegans. Mol. Cell. Biol., 24, 7483–7490.
26. Hogg,J.R. and Goff,S.P. (2010) Upf1 senses 30UTR length to
potentiate mRNA decay. Cell, 143, 379–389.
27. Singh,G., Rebbapragada,I. and Lykke-Andersen,J. (2008) A
competition between stimulators and antagonists of Upf complex
recruitment governs human nonsense-mediated mRNA decay.
PLoS Biol., 6, e111.
28. Eberle,A.B., Stalder,L., Mathys,H., Orozco,R.Z. and
Muhlemann,O. (2008) Posttranscriptional gene regulation by
spatial rearrangement of the 30untranslated region. PLoS Biol.,
29. Matsuda,D., Hosoda,N., Kim,Y.K. and Maquat,L.E. (2007)
Failsafe nonsense-mediated mRNA decay does not detectably
target eIF4E-bound mRNA. Nat. Struct. Mol. Biol., 14, 974–979.
30. Ohnishi,T., Yamashita,A., Kashima,I., Schell,T., Anders,K.R.,
Grimson,A., Hachiya,T., Hentze,M.W., Anderson,P. and Ohno,S.
(2003) Phosphorylation of hUPF1 induces formation of mRNA
surveillance complexes containing hSMG-5 and hSMG-7.
Mol. Cell, 12, 1187–1200.
31. Matsuoka,S., Ballif,B.A., Smogorzewska,A., McDonald,E.R. III,
Hurov,K.E., Luo,J., Bakalarski,C.E., Zhao,Z., Solimini,N.,
Lerenthal,Y. et al. (2007) ATM and ATR substrate analysis
reveals extensive protein networks responsive to DNA damage.
Science, 316, 1160–1166.
32. Page,M.F., Carr,B., Anders,K.R., Grimson,A. and Anderson,P.
(1999) SMG-2 is a phosphorylated protein required for mRNA
surveillance in Caenorhabditis elegans and related to Upf1p of
yeast. Mol. Cell. Biol., 19, 5943–5951.
33. Anders,K.R., Grimson,A. and Anderson,P. (2003) SMG-5,
required for C.elegans nonsense-mediated mRNA decay,
associates with SMG-2 and protein phosphatase 2A. EMBO J.,
34. Wittmann,J., Hol,E.M. and Jack,H.M. (2006) hUPF2 silencing
identifies physiologic substrates of mammalian nonsense-mediated
mRNA decay. Mol. Cell. Biol., 26, 1272–1287.
35. Chiu,S.Y., Serin,G., Ohara,O. and Maquat,L.E. (2003)
Characterization of human Smg5/7a: a protein with similarities to
Caenorhabditis elegans SMG5 and SMG7 that functions in the
dephosphorylation of Upf1. Rna, 9, 77–87.
36. Gatfield,D., Unterholzner,L., Ciccarelli,F.D., Bork,P. and
Izaurralde,E. (2003) Nonsense-mediated mRNA decay in
Drosophila: at the intersection of the yeast and mammalian
pathways. EMBO J., 22, 3960–3970.
37. Fukuhara,N., Ebert,J., Unterholzner,L., Lindner,D., Izaurralde,E.
and Conti,E. (2005) SMG7 is a 14-3-3-like adaptor in the
Nucleic Acids Research,2012, Vol.40, No. 3 1265
nonsense-mediated mRNA decay pathway. Mol. Cell, 17, Download full-text
38. Unterholzner,L. and Izaurralde,E. (2004) SMG7 acts as a
molecular link between mRNA surveillance and mRNA decay.
Mol. Cell, 16, 587–596.
39. Kashima,I., Jonas,S., Jayachandran,U., Buchwald,G., Conti,E.,
Lupas,A.N. and Izaurralde,E. (2010) SMG6 interacts with the
exon junction complex via two conserved EJC-binding motifs
(EBMs) required for nonsense-mediated mRNA decay.
Genes Dev., 24, 2440–2450.
40. Huntzinger,E., Kashima,I., Fauser,M., Sauliere,J. and
Izaurralde,E. (2008) SMG6 is the catalytic endonuclease that
cleaves mRNAs containing nonsense codons in metazoan. Rna,
41. Eberle,A.B., Lykke-Andersen,S., Muhlemann,O. and Jensen,T.H.
(2009) SMG6 promotes endonucleolytic cleavage of
nonsense mRNA in human cells. Nat. Struct. Mol. Biol., 16,
42. Yamashita,A. and Ohno,S. (2010) Analysis of nonsense-mediated
mRNA decay by monitoring mRNA half-lives in mammalian
cells. Cold Spring Harb. Protoc., 2010, pdb. prot5386.
43. Johns,L., Grimson,A., Kuchma,S.L., Newman,C.L. and
Anderson,P. (2007) Caenorhabditis elegans SMG-2 selectively
marks mRNAs containing premature translation termination
codons. Mol. Cell. Biol., 27, 5630–5638.
44. Franks,T.M., Singh,G. and Lykke-Andersen,J. (2010) Upf1
ATPase-dependent mRNP disassembly is required for completion
of nonsense-mediated mRNA decay. Cell, 143, 938–950.
45. Baliga,B.S., Pronczuk,A.W. and Munro,H.N. (1969)
Mechanism of cycloheximide inhibition of protein synthesis in a
cell-free system prepared from rat liver. J. Biol. Chem., 244,
46. Pulak,R. and Anderson,P. (1993) mRNA surveillance
by the Caenorhabditis elegans smg genes. Genes Dev., 7,
47. Redon,S., Reichenbach,P. and Lingner,J. (2007) Protein RNA
and protein protein interactions mediate association of human
EST1A/SMG6 with telomerase. Nucleic Acids Res., 35,
48. Gehring,N.H., Lamprinaki,S., Kulozik,A.E. and Hentze,M.W.
(2009) Disassembly of exon junction complexes by PYM. Cell,
49. Muller,B., Blackburn,J., Feijoo,C., Zhao,X. and Smythe,C. (2007)
DNA-activated protein kinase functions in a newly observed S
phase checkpoint that links histone mRNA abundance with DNA
replication. J. Cell. Biol., 179, 1385–1398.
50. Kaygun,H. and Marzluff,W.F. (2005) Regulated degradation of
replication-dependent histone mRNAs requires both ATR and
Upf1. Nat. Struct. Mol. Biol., 12, 794–800.
51. Azzalin,C.M. and Lingner,J. (2006) The human RNA surveillance
factor UPF1 is required for S phase progression and genome
stability. Curr. Biol., 16, 433–439.
52. Demeshkina,N., Hirokawa,G., Kaji,A. and Kaji,H. (2007) Novel
activity of eukaryotic translocase, eEF2: dissociation of the 80S
ribosome into subunits with ATP but not with GTP. Nucleic
Acids Res., 35, 4597–4607.
53. Pisareva,V.P., Skabkin,M.A., Hellen,C.U., Pestova,T.V. and
Pisarev,A.V. (2011) Dissociation by Pelota, Hbs1 and ABCE1 of
mammalian vacant 80S ribosomes and stalled elongation
complexes. EMBO J., 30, 1804–1817.
54. Isken,O., Kim,Y.K., Hosoda,N., Mayeur,G.L., Hershey,J.W. and
Maquat,L.E. (2008) Upf1 phosphorylation triggers translational
repression during nonsense-mediated mRNA decay. Cell, 133,
55. Benkovics,A.H., Nyiko,T., Merai,Z., Silhavy,D. and Bisztray,G.D.
(2011) Functional analysis of the grapevine paralogs of the
SMG7 NMD factor using a heterolog VIGS-based gene
depletion-complementation system. Plant Mol. Biol., 75, 277–290.
56. Yamashita,A., Chang,T.C., Yamashita,Y., Zhu,W., Zhong,Z.,
Chen,C.Y. and Shyu,A.B. (2005) Concerted action of poly(A)
nucleases and decapping enzyme in mammalian mRNA turnover.
Nat. Struct. Mol. Biol., 12, 1054–1063.
57. Kerenyi,Z., Merai,Z., Hiripi,L., Benkovics,A., Gyula,P.,
Lacomme,C., Barta,E., Nagy,F. and Silhavy,D. (2008)
Inter-kingdom conservation of mechanism of nonsense-mediated
mRNA decay. EMBO J., 27, 1585–1595.
1266 Nucleic Acids Research, 2012,Vol.40, No. 3