well-studied translation inhibition and RNP
granule biogenesis response ( 17 ). We observed
a significant increase in PB and stress granule
abundance in wild-type cells after NaAsO 2
treatment for 60 min (Fig. 4, A and B). PB
abundance also increased significantly after
NaAsO 2 treatment under two cisternal ER-
promoting conditions,RTN4depletion (Fig. 4,
A and B) or mCh-CLIMP63 overexpression
(Fig. 4, C and D). Furthermore, a closer inspec-
tion of PBs induced by NaAsO 2 revealed a
localization to ER domains of high membrane
curvature, such as tubules, fenestrae, and edges
of cisternae (Fig. 4C, insets). Thus, we mea-
sured the effect of NaAsO 2 treatment on the
percentage of PBs that are bound to ER tu-
bules. ER-PB contact was tracked in live cells
expressing GFP-Dcp2 and mCh-KDEL. Trans-
lation inhibition induced by NaAsO 2 treat-
ment doubled the percentage of PBs that
were tightly associated with the ER in these
cells compared with untreated cells (Fig. 4, E
and F).
We aimed to complement NaAsO 2 -induced
RNP granule studies with puromycin to di-
rectly target the translation machinery in wild-
type andRTN4KO cells. Puromycin is an
antibiotic that inhibits global translation by
disrupting peptide transfer, leading to the
release of ribosomes from mRNA. We used
a short 15-min treatment and high 200mM
concentration of puromycin that can clear
ribosomes off ER membranes ( 36 ). The short
puromycin treatment also allowed us to mini-
mize any downstream effects that accompany
a global translation block.
We observed a significant increase in PB
numbers in response to puromycin-induced
release of ribosomes from mRNA and ER
membranes in both wild-type andRTN4KO
cells (fig. S3). As expected ( 22 ), puromycin
did not induce stress granule formation. Thus,
the relationship between PB biogenesis and
ER shape is dependent on the translational
capacity of the ER because inhibiting mRNA
translation with NaAsO 2 or puromycin could
induce PB biogenesis even when cisternal ER
is abundant.ER stress and the unfolded protein response
induce PB disassembly
Cisternal ER and ER translational capacity will
also increase when the unfolded protein re-
sponse (UPR) is evoked during ER stress. The
UPR can relieve the burden of misfolded
protein accumulation by attenuating general
translation and by selectively up-regulating
and translating genes encoding for ER resident
proteins such as chaperones ( 37 ). Thus, we
further tested the relationship between ER
shape, ER translational capacity, and PB bio-
genesis by inducing the UPR. Cells were treated
with tunicamycin (Tm), which triggers the
UPR by blocking N-linked glycosylation ofLeeet al.,Science 367 , eaay7108 (2020) 31 January 2020 5of10
0s 60s 120sUntreatedNaAsO2GFP-Dcp2 / mCh-KDELF020406080100Percent of P-bodies(ER contact)UntreatedNaAsO 22 min on ER< 20s on ER
2130s 60s 120s(^1) < 20s on ER 2 3 2 min on ER
3
2
2
3
3
3
1
Dcp1b / G3BP
immunofluorescence
A
Untreated NaAsO 2
RTN4
KO
Wild-type
Endogenous P-bodies/Cell
0
5
10
15
20
Untreated
NaAsO 2
Wild-type RTN4 KO
NaAsO 2
Inset
mCh-Sec61β mCh-CLIMP63
Edc3 immunofluorescence
0
3
6
9
12
15
Untreated Untreated NaAsO 2
mCh-
Sec61β
mCh-
CLIMP63
B
C D
n (cells) = 90 90 9291
n (cells) = 58 60 6061
E
n (cells) = 30 30
Untreated
NaAsO 2
Endogenous P-bodies/Cell
Fig. 4. The relationship between PB biogenesis, ER shape, and oxidative stress–induced mRNA
translation inhibition.(A) Representative images of Dcp1b (PB marker, green) and G3BP (stress granule
marker, red) IF studies were performed in wild-type andRTN4KO U-2 OS cells treated with 0.5 mM
NaAsO 2 for 1 hour. Blue indicates the nucleus stained by Hoechst. (B) The mean numbers of endogenous
PBs were quantified from three biological replicates. (C) Representative images of Edc3 IF studies
performed in U-2 OS cells transiently transfected with mCh-Sec61bor mCh-CLIMP63 treated with
0.5 mM NaAsO 2 for 1 hour. Arrows within insets highlight PBs that overlap with domains of high ER
membrane curvature. (D) The mean numbers of endogenous PBs were quantified from three biological
replicates in U-2 OS cells exogenously expressing mCh-Sec61bor mCh-CLIMP63 that were either
untreated or treated with NaAsO 2 .(EandF) Representative merged images of U-2 OS cells exogenously
expressing GFP-Dcp2 and mCh-KDEL to label PBs and the ER in live cells, respectively (E). Insets show
movement of the two organelles through space andtime over a 2-min time-lapse movie with frames
captured every 5 s. PBs were binned as in Fig. 1C. Cells were imaged for each condition from three
biological replicates and quantified for the degree of association between the ER and PBs (F). In (B) and
(D), statistical significance was determined by one-way ANOVA with multiple comparisons, and error
bars indicate SEM; ****P< 0.0001. In (A), (C), and (E), scale bars are 5 and 2mminfullcellandinset
images, respectively.
RESEARCH | RESEARCH ARTICLE