Nature - USA (2020-10-15)

446 | Nature | Vol 586 | 15 October 2020

Article

increases in the ASI neurons of the progeny of PA14-trained moth-
ers^3. daf-7p::gfp^3 ,^4 —which exhibits fluorescence in ASJ neurons in
response to bacterial metabolites^4 and fluorescence in ASI neurons
in the progeny of PA14 lawn-trained mothers^3 —is induced by expo-
sure to PA14 sRNA solely in the ASI neurons of trained mothers,
resembling expression in the F 1 –F 4 generations after PA14 lawn
training^3 (Fig. 1k, Extended Data Fig. 1l). sRNA from ΔlasR PA14 did
not increase levels of daf-7 in ASI neurons (Extended Data Fig. 1m).
Worms with genetically ablated ASI neurons exhibit high levels of
naive avoidance of PA14, but exposure to sRNA does not induce

further avoidance (Fig. 1l); this suggests that the ASI neuron is

required for PA14 avoidance mediated by bacterial sRNA (Fig. 1k),

as it is required in the F 1 generation for transgenerational inherit-

ance of avoidance^3.

sRNA avoidance requires the RNAi pathway

We next wondered whether avoidance induced by bacterial sRNA

requires the RNAi pathway^10. Although neither the double-stranded

(ds)RNA transporter SID-2^11 nor the dsRNA endoribonuclease Dicer

e

j

24 h OP50 sRNA 24 h PA14 sRNA

f

k 24 h OP50 sRNA

24 h PA14 sRNA

l

–0.6

–0.4

–0.2

0

0.2

0.4

0.6

0.8

Choice index

**** ********

NS

****

NS

RNase

DNase

Bacteria

lawn sRNA

****

–0.8

–0.6

–0.4

–0.2

0

0.2

0.4

0.6

0.8

1.0

Choice index

N2 ASI caspase

***** NS

–0.6

–0.4

–0.2

0

0.2

0.4

0.6

0.8

1.0

Choice index

**** *******

NS

Bacteria

lawn To tal RNA

RNase

24 h OP50

24 h P24 h S. marcescensA14

–1.0

–0.8

–0.6

–0.4

–0.2

0

0.2

0.4

0.6

0.8

Choice index

NS****

NS

NS

NS

NS

sRNA

24 h OP50 24 h PA14

PA

14

prefere nce

OP50

prefer

ence

–0.6

–0.4

–0.2

0

0.2

0.4

0.6

0.8

1.0

Choice index

**** NS

Bacteria

lawn

Supernatant

PA

14

prefer

ence

OP50

prefere nce

–0.6

–0.4

–0.2

0

0.2

0.4

0.6

0.8

1.0

Choice index

********

NS

********

NS

LawnsRNA

0

0.2

0.4

0.6

0.8

Lear

ning index

**

sRNA

indepednent

sRNA

dependent

i

h

Bacteria lawn

sRNA

OP50 PA 14

S.

marcescensprefer

ence

OP50

prefer

ence

PA

14

prefer

ence

OP50

prefere nce

sRNA Metabolites immunityInnate

Pseudomonas exposure

ASI

daf-7p::gfp

ASJ

daf-7p::gfp

daf-7

independent

Avoidance behaviour

OP50

OP50

OP50

OP50

PA14

2 d 24 h

2 d 24 h

OP50-nucleic acid training

PA14-nucleic acid training

Aversive

learning

assay

Aversive

learning

assay

OP50

PA14

PA 14

g

daf-7p::gfp

PA14

OP50

Lawn sRNA

ASIASIASIASIASASIASIASIASIASASIASIASIASASAASASISSISIASJASJASJASJASJASJASJASJASJASJASJAASJASJASJASJASASJSJSJJJ

ASIASASIASIASIASIASISSISSISISISISSIII

24 h OP50 24 h PA14 24 h OP50 24 h PA 14

–0.4

–0.2

0

0.2

0.4

0.6

0.8

1.0

Choice index

heat-killed

OP50

**** ****

live

OP50

PA14

prefer

ence

OP50

prefer

ence

sRNA on sRNA on

Bacteria

lawn

Bacteria sRNA

lawn

24 h OP50

24 h P24 h S. marcescensA14

a b cd

Fig. 1 | sRNA of PA14 is suff icient to induce C. elegans pathogen avoidance.
a, Worms were trained on non-pathogenic OP50, PA14 lawns, or OP50 lawns
spiked with bacterial components. Choice assays to OP50 versus PA14 bacteria
were then performed. Choice index = (number of worms on OP50 − number of
worms on PA14)/(total number of worms). b, Worms exposed to a PA14 lawn (for
24 h) learn to avoid PA14, but PA14 supernatant does not elicit PA14 avoidance.
c, Training with purified PA14 total RNA confers PA14 avoidance; RNase
treatment abolishes this effect. d, Purified PA14 sRNA (<200 nt) induce PA14
avoidance; RNase treatment abolishes this effect, but DNase treatment does
not. e, Heat-killing bacteria does not abolish sRNA learning. f, Trained worms
were tested in an OP50 versus S. marcescens choice assay. The S. marcescens
lawn training induces avoidance of S. marcescens, but exposure to
S. marcescens sRNA does not. PA14 sRNA training does not affect OP50 versus
S. marcescens preference. g, PA14 lawn and PA14 sRNA training induce PA14
avoidance, whereas S. marcescens bacteria and sRNA training do not affect
PA14 preference. h, Unlike PA14, sRNA exposure does not cause illness. Scale
bar, 100 μm. i, P 0 learning index. sRNA-induced learning is about half of that of

PA14 lawns. Learning index = average PA14 choice index − average OP50 choice

index. Each data point represents the learning index from an independent

experiment containing about 7–10 choice assay plates with an average of 115

worms per plate. Unpaired, two-tailed Student’s t-test. j, Worms learn to avoid

PA14 through several independent mechanisms. k, da f-7p:: g f p expression in

ASI neurons increases upon PA14 lawn or sRNA exposure (white arrows);

expression in ASJ neurons is induced only in PA14 lawn-trained worms (grey

arrows). Scale bar, 50 μm. l, Genetic ablation of ASI neurons abolishes PA14

sRNA-induced learning. Each dot represents an individual choice assay plate

(average of 115 worms per plate) from all replicates. Biological replicates:

3 (c–h, k, m), 4 (b) or 6 (i). Box plots: centre line, median; box range, 25–75th

percentiles; whiskers denote minimum–maximum values. One-way (c, d) and

two-way (b, e, f, l) analysis of variance (ANOVA), Tukey’s multiple comparison

test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P < 0.0001, NS, not significant.

Estimation plots are provided in the Supplementary Information; see

Supplementary Table 4 for exact sample sizes (n) and P values.