Nature - USA (2020-08-20)

Nature | Vol 584 | 20 August 2020 | 411

levels of soluble LBP-2 decreased with age (Fig. 1j, Extended Data Fig. 1g).
Together, these observations show that secreted LBP-2 aggregates with
age in the extracellular space.
To identify factors that are responsible for regulating protein aggre-
gation in the extracellular space, we performed an RNA interference
(RNAi) screen targeting genes encoding the secreted proteome. After
retesting, 57 genes knocked-down by RNAi were confirmed to acceler-
ate LBP-2 aggregation (Extended Data Fig. 2a, Supplementary Table 1).
Secreted candidates identified in this screen are likely to act in the
extracellular space to modulate LBP-2 aggregation rather than during
its secretion, as knockdown of several endoplasmic reticulum (ER)
proteostasis components had no effect. Using Phyre2, we predicted
secondary structure and found 36 proteins with domains related to
sequences of known structure including several with enzymatic activity
(Supplementary Table 1). To validate the ECRs further, we focused on 13
candidates with the strongest effect (Fig. 1k, Supplementary Table 1).

Aggregation of LBP-2 was accelerated when subjecting animals to RNAi

targeting candidates during adulthood (Fig. 1k) or during develop-

ment (Extended Data Fig. 2b, c), without affecting total LBP-2 levels

(Extended Data Fig. 2d, e). Loss-of-function mutants of four candidates

confirmed our RNAi-based observations (Fig. 1l). Knockdown of ECRs

by RNAi did not significantly impair coelomocyte uptake of secreted

GFP compared to positive control dyn-1 knockdown^12 (Extended Data

Fig. 3a), which demonstrates that increased protein aggregation is prob-

ably not due to defective endocytosis. Together, the RNAi screening

approach targeting secreted proteins represents a valuable method to

discover new components of the extracellular proteostasis network.

Next, we tested whether ECRs modify the aggregation of another

secreted aggregation-prone protein LYS-7^13 with a predicted anti-

microbial lysozyme function^15 , which forms puncta with age in the

pseudocoelom (Extended Data Fig. 3b, c). Ten out of thirteen ECRs

targeted by RNAi repeatedly accelerated LYS-7 aggregation (Extended

h j

Fluo DIC/Fluo merge

Day 2 Day 8 Day 12

LBP-2::tagRFP aggregation

b

+–

Coelomocytes

i LBP-2::tagRFP aggregation

aLBP-2::tagRFP, day 2

LBP-2::tagRFP, day 8

c

f g

Day 2

d e

DayDDDDDDDDDDayDayDayDayDayDayDayDDayyy 222222222222222222222222222222222222

Day 8

Ctrl (–

)

Ctrl (+)dod-2

1

lys-3irld-44clec-1

F56B6.6C36C5.5 tag-196

T24A6.16cyp-33C5 C13C12.

2

F11E6.3

Y38A10A.2

F54E2.

1

Animals withaggregates (%)

k

P < 0.0001

LBP-2::tagRFP aggregation

Ctrl (–

)

P = 0.0013

1–10 aggregates >10 aggregates

0

20

40

60

80

100

tag-196(ok822

)

l LBP-2::tagRFP aggregation

clec-1

(tm1291)

WT WTlys-3 WT

(5

(^05)
(^2) m
t)dod-21(ok1569)
WT
n= 14015894147161138133152143155136160 142131137142
P < 0.0001
LBP-2::tagRFP levels
LBP-2::
tagRFP
kDa
60
50
40
2 10 2 10 2 10 Days
To talSoluble Insoluble
11 .1 10 .47 1 2.3 Fold change
Animals withaggregates (%)
0
20
40
60
80
100
n = 131 10093
P < 0.0001 P < 0.0001

10 aggregates
1–10 aggregates
Animals withaggregates (%)
0
20
40
60
80
100
n=9 9 100
10 aggregates
1–10 aggregates
P < 0.0001
1–10 aggregates >10 aggregates
Animals withaggregates (%
)
0
20
40
60
80
100
n=6 16089878281
Animals withaggregates (%
)
0
20
40
60
80
100
n =9 997
P = 0.0052
P = 0.012
P = 0.0001
P < 0.0001
Fig. 1 | Secreted LBP-2 aggregates in the extracellular space with age in
C. elegans. a, b, LBP-2::tagRFP in young animal (representative image of n = 16
wor ms) (a) and diffuse localization in head region (n = 24 worms) (b). c, LBP-
2::tagRFP in coelomocytes (dashed circles, n = 5 worms). d, e, LBP-2::tagRFP in
aged animal (n = 8 worms) (d) and puncta localized in head region (n = 27
wor ms) (e). Images in a and d are from identical exposures; arrowheads mark
coelomocytes. Images in b and e are maximum projections, laser intensity 10%
(b) and 8% (e). f, LBP-2::tagRFP puncta formation (magenta) between body-wall
muscle (top) and pharyngeal muscle (bottom) (F-actin, green) (day 4, n = 7
wor ms). g, Without coelomocytes, LBP-2::tagRFP accumulates as puncta
(magenta) surrounded by secreted GFP (green) adjacent to pharynx (GFP
pharyngeal reporter, lower outline) (day 6, n = 5 worms). Images in f and g are
from a single plane. Scale bars, 100 μm (a, d), 5 μm (c), 20 μm (b, e) or 10 μm
(f, g). h, i, Quantification of LBP-2::tagRFP aggregation with age (n = 2
independent experiments) (h) and in young animals without coelomocytes
(n = 3 independent experiments) (i). j, LBP-2::tagRFP levels detected by western
blot, with fold changes shown relative to corresponding levels at day 2 (n = 2
independent experiments). k, Quantification of LBP-2::tagRFP aggregation
with RNAi targeting top 13 ECR candidates (25° C, day 6, n = 1 independent
experiment). ‘Ctrl (−)’ denotes empty vector negative control; ‘ctrl (+)’ denotes
rab-5 RNAi positive control. l, Quantification of LBP-2::tagRFP aggregation in
tag-196(ok822), lys-3(tm2505) and do d-21(ok1569) mutants (day 6, left graph)
and in the cle c-1(tm1291) mutant (day 2 as mutation was lethal afterwards, right
graph) (n = 3 independent experiments). P values were determined by
chi-square test (h, l left graph), two-sided Fisher’s exact test (i, l right graph)
and ordinal logistic regression (k). For blot source image, see Supplementary
Fig. 1.