Nature | Vol 584 | 20 August 2020 | 413together with the other overexpressed ECRs, stops extracellular protein
aggregation.
We next investigated whether an improvement of extracellular pro-
teostasis could have a beneficial effect on lifespan. Notably, the over-
expression of F56B6.6 and LYS-3 extended C. elegans lifespan, whereas
animals that lacked coelomocytes were shorter lived (Extended Data
Fig. 5a, b, Extended Data Table 1). LBP-2 or secreted GFP overexpression
alone had no effect on lifespan (Extended Data Fig. 5c). ECRs expression
was differently regulated during ageing (Extended Data Fig. 5d) and
notably dod-21 (−24 fold) and F11E6.3 (−57 fold) were strongly down-
regulated with age, probably contributing to the collapse in extracel-
lular proteome integrity with age. These results show that secreted
ECRs can delay ageing and may affect other life history traits.
Bioinformatic analysis revealed a highly significant
over-representation of ECRs among genes upregulated in response
to diverse pathogens (Supplementary Table 2). This raises the pos-
sibility that enhancing extracellular proteostasis is part of the host
response to pathogens, similar to expanding intracellular proteostasis
capacity^6 ,^8 such as inducing the ER unfolded protein response^9 ,^16. To
investigate this, we exposed animals that overexpress LBP-2 to the
virulence factor Cry5B, which forms pores in the intestinal plasma
membrane and induces an extensive innate immune response^17 ,^18.
Intoxication with Cry5B led to a strong reduction in LBP-2 aggregation
at different concentrations (Fig. 3a) without reducing the transcrip-
tion of lbp-2 (Extended Data Fig. 6a). Two other non-lethal pathogens,
Microbacterium nematophilum (which targets the anal cuticle) and
Bacillus atrophaeus, also caused suppression of LBP-2 aggregation
(Extended Data Fig. 6b). We examined whether extracellular proteosta-
sis is upregulated, and found that three hours of exposure to Cry5B was
sufficient to increase the expression of four out of eight ECRs (Fig. 3b).
The JNK-like MAP kinase KGB-1, which is essential for survival on Cry5B,
controls more than half of the genes induced^18. Accordingly, we found
that KGB-1 regulates ECR induction, as kgb-1 mutants induced only one
out of four ECRs upregulated in response to Cry5B, without strongly
affecting basal levels of expression (Fig. 3c, Extended Data Fig. 6c).
By contrast, the inhibition of VHP-1, a MAP kinase phosphatase that
negatively regulates both KGB-1 and PMK-1^19 ,^20 , induced three out of
eight ECRs (Extended Data Fig. 6d) and strongly reduced LBP-2 aggre-
gation in a KGB-1-dependent manner (Fig. 3d). We also examined the
role of another stress-activated MAP kinase, JNK-1. Although JNK-1 is
not essential for survival on Cry5B^21 , jnk-1(gk7) mutants were signifi-
cantly shorter lived when exposed to Cry5B (Extended Data Fig. 6e).
Notably, during Cry5B intoxication, loss-of-function of jnk-1 or the
upstream regulator jkk-1 restored the aggregation of LBP-2 (Fig. 3e).efab100Survival on Cry5BLBP-2::tagRFP aggregation qRT–PCRdLBP-2::tagRFP aggregation LBP-2::tagRFP aggregationWT kgb-1(km21)cqRT–PCRCry5B (%) 50 5Time (days on Cry5B) Time (days on Cry5B)02468101214020406080100Time (days on Cry5B)F56B6.6 OE
+Cry5B
Control
+Cry5BCLEC-1 OE
+Cry5B
Control
+Cry5BLYS-3 OE
+Cry5B
Control
+Cry5B
P = 0.0052 P = 2.6 × 10–8Ctrlvhp-1Ctrlvhp-1lys-3
F54E2.1tag-196C36C5.5F56B6.6clec-1
F11E6.3dod-2102468101214 0246810121402468102004006008000246810lys-3
F54E2.1tag-196C36C5.5Fold induction (Cry5B/ctrl)Foldinduction(Cry5B/ctrl)kgb-1(km21)
backgroundWT backgroundRNAiAnimals withaggregates (%) 020406080100n= 1114669P < 0.0001P = 0.0047Animals withaggregates (%) 020406080100n=4 3371–10 aggregates >10 aggregates0 Cry5B (%)0Animals withaggregates (%) 020406080100n=9 0835896P < 0.0001P = 0.33
>10 aggregates1–10 aggregates020406080100Animals withaggregates (%)n = 90 42 78 42 81 53
Cry5B
WT jnk-1
(gk7)jkk-1
(km2)- ++– +–
P = 0.51
P = 0.56P < 0.0001P < 0.0001P = 0.0006P < 0.0001
P = 0.025P = 0.59P = 0.95P = 0.95P = 0.48 P = 0.49P = 0.14P = 0.0018P = 0.31Survival (%) Survival (%)P = 4.7 × 10–7020406080100020406080100Survival (%)Fig. 3 | Pathogenic attack stimulates extracellular proteostasis through
stress-activated MAP kinase signalling. a, Quantification of LBP-2::tagRFP
aggregation during exposure to the Cry5B toxin (day 4, left graph, and day 5,
right graph) (n = 2 independent experiments). b, c, Expression levels of ECRs,
after 3 h exposure to 100% Cry5B, determined by qRT–PCR in the wild-type
background (n = 6 biologically independent samples) and in the kgb-1(km21)
loss-of-function mutant background (n = 5 biologically independent samples).
Data are means ± s.e.m. d, Quantification of LBP-2::tagRFP aggregation at day 6
in wild-type and kgb-1(km21) mutant worms treated with empty vector or vhp-1
RNAi. e, Quantification of LBP-2::tagRFP aggregation at day 6 in wild-type and
jnk-1(gk7) or jk k-1(km2) loss-of-function mutants exposed to 5% Cry5B toxin. In
d and e, n = 2 independent experiments. f, Survival analysis of F56B6.6-, CLEC-1-
and LYS-3-overexpressing animals compared to non-overexpressing siblings
subjected to 50% Cry5B exposure (n = 2 independent experiments, for detailed
values see Extended Data Table 2). P values determined by chi-square test (a,
left, and d), two-sided Fisher’s exact test (a, right, and e), two-sided unpaired
t-test with Welch’s correction (b, c) or log-rank test (f).