Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
Transcript of Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
-
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
1/11
984 VOLUME 14 | NUMBER 8 | AUGUST 2011 nature neurOSCIenCe
a r t I C l e S
Temperature is an unavoidable somatic stimulus that ultimately
affects survival and reproduction. The impact of temperature hasbeen clearly demonstrated by studies showing that change of core
body temperature by means of changing the temperature of thehypothalamus alters the lifespan of transgenic mice1. To confront
environmental temperature change, animals must evolve a behavioralstrategy for ensuring the temperature homeostasis in their bodies.
This evolution is exemplified by poikilotherms, whose body tempera-ture is determined by temperature exchange with the surrounding
environment owing to the large surface area to volume ratios of their
bodies
2
. Various poikilotherms have acquired the ability to memorizecomfortable temperatures and to move to places that minimize tem-perature deviation from the memorized temperatures.
One such successful strategy of poikilotherms is observable inthermotaxis in the nematode C. elegans3,4: most worms that are cul-
tivated at a particular temperature between 15 C and 25 C for a fewhours with a food source migrate to the cultivation temperature when
placed on a thermal gradient for an hour5,6. In addition, worms thatare conditioned to migrate to a certain temperature are capable of
migrating to a new cultivation temperature a few hours after shiftingto the new cultivation temperature3,5. Previous studies provided a
neural model for thermotaxis, in which temperature is sensed, pro-cessed and remembered through a neural circuit that is composed of
thermosensory neurons and interneurons4,79.
Here we show that systemic temperature signaling induces modula-tion of the thermotactic neural circuit in C. elegans. Genome-wideprofiling upon behavioral conditioning indicated that perception of
a change in normal cultivation temperatures by heat-shock transcrip-
tion factor HSF-1, known for its responsiveness to transient heat-shock stimuli (between approximately 30 C and 35 C, where the
range of normal cultivation temperatures in C. elegans is betweenapproximately 15 C and 25 C), drives the expression dynamics of
direct or indirect downstream genes not found among the classical
heat-shock response hsp genes. Expressions of HSF-1 in body wallmuscles or intestine restored the thermotactic defect ofhsf-1 mutants,
suggesting that non-neuronal cells can induce HSF-1-dependentdownstream signaling in accordance with cultivation temperature
changes. Genetic epistasis analysis and calcium imaging showed thatthis systemic signaling regulates non-cell autonomously the thermo-
sensory neurons through the estrogen signaling pathway, leading tobehavioral change.
RESULTS
Genome-wide profiling highlights HSF-1
To gain detailed molecular insight into thermotactic behavior, we
performed a genome-wide microarray analysis during behavioralconditioning and compared the transcriptional profile of worms con-
ditioned to migrate to the new temperature of 17 C with worms con-ditioned to migrate to the previous temperature of 23 C (Fi. 1a andSuppary Fis. 1 and 2). We detected 79 candidate genes, whichare likely to include genes that are involved in thermotactic behavior
in addition to genes that are differentially expressed simply as a resultof temperature shift. Examining the candidate genes, we noticed sub-
stantial changes in expression of several hsp genes, which are regulatedby HSF-1 (ref. 10). HSF-1 is highly conserved from nematodes to mam-
mals11. Previous structural and functional characterizations of HSF-1
revealed that the transient heat-shock stimuli enable HSF-1 to changestructurally from an autoinhibited monomer to a trimer and to trans-locate to the nucleus, thereby activating gene transcription10,1216. Thus,
HSF-1 is well known to directly perceive transient heat-shock stimuli.
Notably, previous in vitro studies have shown that normal growth tem-peratures, as well as transient heat-shock stimuli, can induce the reversi-
ble structural transition of HSF-1 and its DNA binding17,18. Consideringthese previous observations and our own (Suppary Fi. 2),
1Group o Molecular Neurobiology, Graduate School o Science, Nagoya University, Nagoya, Japan. 2CREST, Japan Science and Technology Agency, Tokyo, Japan.3Institute or Advanced Research, Nagoya University, Nagoya, Japan. 4Present address: Department o Molecular Engineering, Graduate School o Engineering, Kyoto
University, Kyoto, Japan. Correspondence should be addressed to I.M. ([email protected]).
Received 1 April; accepted 3 May; published online 26 June 2011; doi:10.1038/nn.2854
Regulation of behavioral plasticity by systemictemperature signaling in Caenorhabditis elegans
Takuma Sugi1,4, Yukuo Nishida1 & Ikue Mori13
Animals cope with environmental changes by altering behavioral strategy. Environmental information is generally received by
sensory neurons in the neural circuit that generates behavior. However, although environmental temperature inevitably influences
an animals entire body, the mechanism of systemic temperature perception remains largely unknown. We show here that
systemic temperature signaling induces a change in a memory-based behavior in C. elegans. During behavioral conditioning,
non-neuronal cells as well as neuronal cells respond to cultivation temperature through a heat-shock transcription factor that
drives newly identified gene expression dynamics. This systemic temperature signaling regulates thermosensory neurons
non-cell-autonomously through the estrogen signaling pathway, producing thermotactic behavior. We provide a link betweensystemic environmental recognition and behavioral plasticity in the nervous system.
http://www.nature.com/doifinder/10.1038/nn.2854http://www.nature.com/doifinder/10.1038/nn.2854 -
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
2/11
nature neurOSCIenCe VOLUME 14 | NUMBER 8 | AUGUST 2011 98 5
a r t I C l e S
we wished to address the new biological importance of the role ofHSF-1 in responding to normal cultivation temperatures.
We confirmed in vivo whether a temperature shift within a range of
ambient cultivation temperatures (between approximately 15 C and25 C) activates C. elegans HSF-1 by observing the green fluorescentprotein (GFP) fluorescence of wild-type strains expressing hsp-16.2
promoter::gfp reporter, whose promoter contains two heat-shock tran-scription factor binding elements (HSEs) (Fi. 1). We observed large
GFP fluorescence changes that were dependent on two HSEs at 2 hafter shifting the temperature from 15 C to 23 C or to 25 C (Fi. 1fh
and Suppary Fi. 3a), which were comparable with thoseobserved at 2 h after the transient heat-shock treatment (Fi. 1i).
As HSF-1 activity responded to ambient cultivation temperatures,we addressed whether HSF-1 is required for thermotactic behavior.
hsf-1(sy441) mutants19, carrying a reduction-of-function mutation(see Suppary Srais a pasis), migrated to a cultiva-
tion temperature to which they had previously been habituated when
cultivated at 17 C (Suppary Fi. 4a,b,), whereas mutantscultivated at 23 C did not migrate to the cultivation temperature(Suppary Fi. 4a,,). In a temperature shift assay, at 5 h after
shifting the cultivation temperature from 17 C to 23 C, hsf-1(sy441)
mutants failed to migrate toward higher temperatures (23 C) andinstead migrated to colder temperatures (Fi. 2a). At 5 h after shift-
ing the cultivation temperature from 23 C to 17 C, they showed atendency to migrate toward the new cultivation temperature of 17 C
(Fi. 2b). Worms carrying the hsf-1(ok600) null mutation as either ahomozygote or heterozygote showed a behavioral defect similar to
that of the hsf-1(sy441) mutants (Fi. 2).We next addressed whether the abnormal thermotactic migration
ofhsf-1(sy441) mutants is merely a reflection of slower temperature
memory acquisition: might they eventually become conditionedto the higher temperatures if cultivated at them longer? This was,
however, not the case, since hsf-1(sy441) mutants also showed the
behavioral defect at 10 h after shifting the temperature from 17 Cto 23 C (Fi. 2).Excess expression of native HSF-1 in wild-type worms10 also
induced partially abnormal migration, which was partly reminiscentof the defect ofhsf-1(sy441) mutants (Fi. 2a,b). However, hsf-1(sy441)
mutants grown at 23 C showed cryophilic behavior of greater sever-ity than hsf-1 overexpression lines grown at 23 C, which probably
leads to the behavioral difference observed after shifting the cultiva-tion temperature to 17 C. When cultivated for a longer time after
shifting temperature from 17 C to 23 C, hsf-1(sy441) mutants alsoshowed a strong tendency to migrate toward colder regions than hsf-1
overexpression lines, although that difference was not statisticallysignificant (Fi. 2). These observations suggest that hsf-1(sy441)
mutants and hsf-1 overexpression lines both show an abnormal cryo-
philic behavior, but that their tendency to do so seems to be different.These results suggest that the balanced activity of HSF-1 is importantfor proper thermotactic response.
The thermotaxis defect in hsf-1(sy441) mutants (hereafter referred
to as hsf-1 mutants) might simply reflect their temperature-sensitivedebility19, thus causing limited migration ability for all sensory cues
at 23 C. We therefore addressed whether other sensory responses,such as odorant chemotaxis20,21 ofhsf-1 mutants cultivated at 23 C,
are more defective than those cultivated at 17 C (Suppary
Fi. 3). The hsf-1 mutants cultivated at 23 C did not exhibit
olfactory defects of greater severity than the 17 C-cultivated mutants.Moreover, hsf-1 mutants cultivated at 23 C moved freely like wild-
type worms on the thermotaxis plate without a thermal gradient
a
b
e f
i
c
Cultivated
at 23 C
4.00
Linear TTX assay
Microarray
Linear TTX assay
Microarray
0 1 2 3
Time (h)
4 5 6
Shift
to17
C
at0h
4hafter
shifti
ngto
17C
3.002.001.00
TTXi
ndex
0
14
Intensityat1h(a.u.)
Intensityat2h(a.u.)
Intensity at 0 h (a.u.)
1.002.003.004.00
12
10
8
6
4
2
14
12
10
8
6
4
2
2 4
HSF-1 binding element
hsp-16.2p
(500 bp)
6
40 m
40 m g h40 m 40 m
gfp
8
2 h after transient heat-
shock treatment
10 12 14
Intensity at 0 h (a.u.)
Cultivation at 15 C 2 h after shifting to 23 C 2 h after shifting to 25 C
2 4 6 8 10 12 14
d
Intensityat4h(a.u.) 14
12
10
8
6
4
2
Intensity at 0 h (a.u.)
2 4 6 8 10 12 14
Figure 1 The essential role o HSF-1 in a
memory-based behavior revealed by genome-
wide microarray analyses. (a) Procedure or
microarray analysis o the thermotactic (TTX)
behavior conditioning process. C. elegans
were cultivated at 23 C or 2.5 d and were
then shited to a new temperature o 17 C.
Immediately ater start o the assay, we isolated
mRNA rom worms that were not used or the
behavioral assays. Error bars represent s.e.m.
(bd) Overall transition o gene expression
change during behavioral conditioning.
Scatter plot o intensity values (log2) or a
representative hybridization o RNA isolated
rom worms at 1 h (b), 2 h (c) or 4 h (d) ater
shiting the cultivation temperature rom
23 C to 17 C compared with that isolated
rom worms just beore shiting the cultivation
temperature. A.u., arbitrary units. (ei) The
HSF-1 activity within a range o cultivation
temperatures in vivo. The hsp-16.2p::gfp
reporter gene, including the two HSF-1 binding
elements (HSEs), was expressed in a wild-type
background (e). Each worm was cultivated
with ood at 15 C or 7 d. The GFPobservations were perormed beore shiting
the temperature (f) and at 2 h ater shiting
to the new temperatures (which were within
the range o normal cultivation temperatures);
23 C (g) or 25 C (h). For the transient heat-
shock treatment (i), each worm was incubated at 35 C or 15 min and then allowed to recover at 15 C or 2 h beore the
GFP observation. The GFP observations were conducted at least our times in each experimental condition.
-
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
3/11
986 VOLUME 14 | NUMBER 8 | AUGUST 2011 nature neurOSCIenCe
a r t I C l e S
(Suppary Fi. 5bf). These results suggest that the abnormal
thermotactic behavior ofhsf-1 mutants does not simply arise fromtemperature-sensitive debility but instead reflects a defect in the
mechanism specific for thermotaxis.To examine whether HSF-1 acts after a temperature shift from 17 C
to 23 C, we constructed worms expressing in a wild-type backgroundan HSF-1 dominant negative form (HSF-1DN)22 under the hsp-16.2
promoter and induced expression of HSF-1DN by heat-shock treat-
ment after worms had reached the adult stage. As the activity of thehsp-16.2 promoter itself is regulated by HSF-1, we co-injected hsp-
16.2p::gfp to confirm whether expression of HSF-1DN appropriately
inhibits the endogenous HSF-1 activity by monitoring the GFP fluo-rescence from this reporter gene. Just before heat-shock treatment,
we observed hardly any fluorescence from worms carrying only thereporter gene and from those carrying hsp-16.2p::hsf-1dn and the
reporter gene, and showed that both groups appropriately migratedtoward lower temperature regions of around 17 C on the thermal
gradient (data not shown). The wild-type worms carrying only thereporter gene and those carrying both hsp-16.2p::hsf-1dn and the
reporter gene were then subjected to the transient heat-shock treat-ment just before a temperature shift from 17 C to 23 C. At 3 h after
a temperature shift from 17 C to 23 following the transient heat-
shock treatment, wild-type worms carrying both hsp-16.2::hsf-1dnand the reporter gene showed slightly less fluorescence than wild-typeworms carrying only the reporter gene (Suppary Fi. 6), sug-
gesting that HSF-1DN inhibited the activity of endogenous HSF-1.
Moreover, wild-type worms carrying both hsp-16.2::hsf-1dn and thereporter gene showed abnormal cryophilic behavior compared with
wild-type worms without both hsp16.2::hsf-1dn and the reporter gene(data not shown) and compared to those carrying only the reporter
gene (Fi. 2). These results suggest that HSF-1 functions after shift-ing the temperature from 17 C to 23 C. The thermotactic defects
characteristic of hsf-1 mutants suggest that HSF-1 is required forresponding to higher temperature, for memorizing higher tempera-
ture or for both processes.
Importance of HSF-1 downstream signaling in thermotaxis
Notably, thermotactic behavior was barely affected by the tran-sient heat-shock treatment, although as expected markedly affected
by shifting the cultivation temperature (Fi. 3a). To identify thegenes involved in the HSF-1 downstream signaling that is involved
in regulating thermotaxis, we classified the genes that were dif-ferentially expressed under microarray analyses into functional
categories (Suppary Fi. 1b) and then explored whether
expression changes of representative genes in each category areHSF-1-dependent.After shifting the temperature from 15 C to 25 C, or vice
versa, amounts of mRNA for hsp genes (Suppary Fi. 3)were substantially changed in wild-type worms, whereas changes
in mRNA amounts were smaller in hsf-1 mutants. Similarly, theHSF-1-dependent expression changes were observed for genes
encoding proteins that potentially have catalytic ac tivity, such asDHS-4 (hydroxysteroid 17 dehydrogenase 11) (Fi. 3b); genes
encoding ion and peptide transporters, such as PGP-9 (ABC trans-porter superfamily) (Fi. 3); and other gene functions (Fi. 3).
Thus, cultivation temperaturedependent changes in expression ofgenes with certain funct ions seem to occur in an HSF-1-dependent
manner. Consistent with the importance of HSF-1 in thermotaxis,
mutants for kgb-2 (Jun N-terminal kinase), dhs-4, kqt-2 (KvQLT-family potassium channel),pgp-9 and max-1 (a regulator of axonguidance) showed abnormal thermotactic migration (Fi. 3). By
contrast, the amounts of mRNA for genes such as cgt-1 (ceramideglucosyltransferase), pmp-1 (long-chain acyl-CoA transporter)and dmd-7 (the transcription factor doublesex) changed inde-
pendently of HSF-1 (Fi. 3b). Mutants for these genes showedalmost normal thermotaxis (Fi. 3). Thus, quantitative polymer-
ase chain reaction (PCR) analysis and behavioral analysis sug-gest that expression of several genes, such as dhs-4, is directly
or indirectly regulated by HSF-1, and that HSF-1-dependentchanges in downstream signaling are likely to be important
for thermotaxis.
Figure 2 Characterization o thermotaxis
behavior o hsf-1 mutants. (a,b) Population
thermotaxis (TTX) temperature shit assay
o wild-type worms, hsf-1(sy441)mutants19
and hsf-1-overexpressing worms10.
Cultivation temperature was shited rom
17 C to 23 C (a) and vice versa (b). The
population TTX assay was conducted just
beore shiting the cultivation temperature
(0 h) and at every 1 h ater shiting
the temperature. (c) Thermotaxis o
hsf-1(ok600)null mutants. Population
TTX assays were conducted at 2 h ater
shiting the temperature rom 17 C to
23 C. **P< 0.01 (t-tests, n = 3).
(d) Evaluation o the eect o long-
term cultivation at 23 C ater shiting
temperature, on the thermotaxis o hsf-1
mutants and hsf-1-overexpressing worms.
Population TTX assays were conducted at
7 h and 10 h ater shiting temperature
rom 17 C to 23 C. *P< 0.05,
**P< 0.01 (t-tests, n = 3). (e) Assessment
o the dominant negative eect o HSF-1 on
thermotaxis in the adult stage o wild-typeworms. HSF-1DN was expressed in wild-type
worms under hsp-16.2promoter (hsp-16.2p::hsf-1dn). Population TTX assays were perormed at 3 h ater heat-shock treatment ollowed
by the temperature shit rom 17 C to 23 C. **P< 0.01 (t-tests, n = 3). NS, not signiicant. Error bars represent s.e.m.
3.00
4.00
3.00
2.00
1.00
0
2.00
1.00
0
TTXi
ndex
TTXi
ndex
TTXi
ndex
TTXi
ndex
TTXi
ndex
3.00
1.00
4.00
3.00
2.00
1.00
0
7 h after shifting
to 23 C
10 h after shifting
to 23 C
1.00
3.00
2.00
1.00
0
1.00
12
Time (h)
Time (h)
Wild typehsf-1(sy441) mutanthsf-1 overexpression
Wild type
hsf-1(sy441)mutant
hsf-1 overexpression
3 4 5
1 2 3 4 5
2.00
1.00
3.00
2.00
1.00
0
3.00
2.00
1.00
NS NS NS
**
**
*
**
**
****
a
c d e
b
hsf-1
(ok6
00)
hsf-1
(sy4
41)
Wild
type
hsf-1
over-
expre
ssion
hsf-1
(sy4
41)
Wild
type
hsf-1
over-
expre
ssion
hsf-1
(sy4
41)
Wild
type
Wild
type
hsf-1
dn
-
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
4/11
-
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
5/11
988 VOLUME 14 | NUMBER 8 | AUGUST 2011 nature neurOSCIenCe
a r t I C l e S
calcium transients of AFD and AWC in the hsf-1 mutants by calcium
imaging with a genetically encoded calcium indicator, cameleon.When temperature was raised and then lowered, the fluorescence
resonance energy transfer (FRET) ratios in AFD changed less in hsf-1mutants than in wild-type worms (Fi. 6a). By contrast, AWC in hsf-1
mutants was more responsive to the transient temperature changefrom 17 C to 23 C than AWC in wild-type worms (Fi. 6b). These
results suggest that the HSF-1 signaling is involved in modulation ofthe responsiveness of AFD and AWC to the transient temperature
change. Consistent with these observations, FRET ratios of the AIY
interneuron, postsynaptic to both AFD and AWC, in hsf-1 mutantswere substantially less than those in wild-type worms, probablyowing to the decreased and increased responsiveness of AFD and
AWC to transient temperature changes in hsf-1 mutants, respectively
(Suppary Fi. 9). Expression ofhsf-1 in body wall musclesrestores the temperature responsiveness of AFD, AWC and AIY inhsf-1 mutants (Fi. 6a,b and Suppary Fi. 9b). These resultsare consistent with the results from genetic epistasis analysis, which
found that the HSF-1 signaling regulated AFD and AWC (Fi. 5).Previous calcium imaging experiments have shown that, when tem-
perature is raised from 15 C to 25 C in a step-wise manner, AFDand AWC neurons respond to every warming above a threshold tem-
perature that is set by the previous cultivation temperature, indicating
a
300 m
e f
g
i
h
50 m
Nerve ringb
30 m
Body wall musclec
40 m
Intestined
3.00
Wildtype
hsf-1
unc-14p
Wildtype
Pan-neuronal
hsf-1
unc-14p
***
3.00
2.00
2.00
1.00
1 ng l1
10 ng l1
TTXindex
1.00
0
3.00
Wildtype
hsf-1
myo-3p
Wildtype
Body wall
muscles
hsf-1
myo-3p
Wildtype
hsf-1
ges-1p
Wildtype
hsf-1
ges-1p
Wildtype
hsf-1
ges-1p::gfp
** ** **
3.00
2.00
2.00
1.00
1 ng l1
10 ng l1
Intestine Marker
gene
1 ng l1
10 ng l1
50 ng l1
TTXindex
1.00
0
3.00
Wildtype
hsf-1
gcy-8p,ceh-36p,
AIZp,g
lr-3p
Thermotactic
neural circuit
**
3.00
2.00
2.00
1.00
TTXindex
1.00
0
Wildtype
hsf-1
ttx-3p
AIY AIZ RIA
3.00
3.00
2.00
2.00
1.00
TTXindex
1.00
0
* ***
1 ng l1
Wildtype
hsf-1
ttx-3p
10 ng l1
10 ng l1
10 ng l1
Wildtype
hsf-1
lin-11p
Wildtype
hsf-1
lin-11p
1 ng l1
Wildtype
hsf-1
glr-3p
Wildtype
hsf-1
glr-3p
1 ng l1
3.00
Wildtype
hsf-1
gcy-8p
Wildtype
hsf-1
gcy-8p
Wildtype
hsf-1
ceh-36p
Wildtype
hsf-1
ceh-36p
Wildtype
hsf-1
sra-6p
** **
3.00
2.00
2.00
1.00
TTXindex
1.00
0
AFD AWC ASH
1 ng l1
10 ng l1
1 ng l1
10 ng l1
1 ng l1
Wildtype
hsf-1
sra-6p
Wildtype
hsf-1
gcy-5p
gcy-7p
Wildtype
hsf-1
gcy-5p
gcy-7p
10 ng l1
ASE
1 ng l1
10 ng l1
Figure 4 Cell-speciic rescue o deective
thermotactic behavior in hsf-1 mutants.
(ad) Expression pattern o HSF-1 in wild-type
worms. To drive the GFP luorescence under the
hsf-1 promoter, hsf-1 promoter::gfpreporter
gene was injected into the wild-type worms.
The GFP luorescence was detected in nearly
all cells (a): or example, in the nerve ring (b),
body wall muscles (c) and intestine (d).
(ei) Cell-speciic rescue experiments.
Thermotactic behavior was examined at 2 h
ater shiting temperature rom 17 C to 23 C
by population thermotaxis (TTX) assays.
hsf-1 cDNA was expressed as a transgene
(1 ng l1 or 10 ng l1) in almost all neurons
(pan-neuronal; e or in the body wall muscles
or intestine (f) o hsf-1 mutants using individual
cell-speciic promoters with ges-1p::gfp
(50 ng l1) as an injection marker.
The transgene was also coordinately expressed
in AFD neurons at 1 ng l1, AWC neurons
at 10 ng l1 and AIZ and RIA neurons at
1 ng l1 (g) and individually expressed in the
sensory neurons AFD, AWC, ASH or ASE (h) or
in the interneurons AIY, AIZ or RIA ( i). At leasttwo lines were assayed or each transgenic
strain. Experiments or each transgenic strain
were carried out at least three times. The
signiicant dierences o thermotactic plasticity
between each transgenic strain and hsf-1(sy441)
mutants were determined by Fishers protected
least signiicant dierence multiple-comparison
test. *P< 0.05, **P< 0.01. Error bars
represent s.e.m.
genetic epistasis analysis suggests that the HSF-1-mediated non-cell-autonomous signaling regulates AFD and AWC thermosensory
neurons, AIY interneurons or both the thermosensory neurons and
interneurons, possibly through its effects downstream of the cGMP-dependent pathway (Suppary Fi. 7).Although recent studies addressed whether the HSF-1 activation is
dependent on the AFD neuron16 or not27, our quantitative PCR analy-sis suggests that expression changes of the hsp genes in response to
both heat-shock stimuli and cultivation temperature changes were notsubstantially different between wild-type worms andgcy-23 gcy-8 gcy-18 triple mutants, or ttx-3 mutants (Suppary Fi. 8). We alsoshowed that the GFP fluorescence level of non-neuronal cells, such as
intestine ingcy-23 gcy-8 gcy-18 mutants or ttx-3 mutants expressinghsp-16.2 reporter constructs, is comparable to the fluorescence of
those in wild-type worms. The results in this study were consistentwith a recently reported result suggesting that the HSF-1 activity is
independent of the ttx-1 mutation that is known to specifically disrupt
AFD structure27, but partly contradicted a previous study16 that foundthe AFD and AIY neurons to be important for the expression of a fewhsp genes. This contradiction might arise from the subtle difference in
growth conditions. In particular, humidity has been known to affecttemperature sensation and thermotaxis, although little is known about
the mechanism dictating the relationship between hygrosensation andthermosensation. Taken together, at least in our growth condition,
expression of a few HSF-1 downstream genes such as hsp-70 seems to
be independent ofgcy-23 gcy-8 gcy-18 and ttx-3 mutations.
Regulation of the thermosensory neurons by HSF-1 signaling
To address the possibility that the HSF-1 signaling affects AFD and
AWC sensory neurons, we monitored temperature stimulusevoked
-
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
6/11
nature neurOSCIenCe VOLUME 14 | NUMBER 8 | AUGUST 2011 98 9
a r t I C l e S
that these thermosensory neurons can, of themselves, memorize theperceived cultivation temperature9,28. To investigate the responsive
temperature range, we also measured the FRET ratio changes of AFDneurons of wild-type worms and hsf-1 mutants during the step-wise
warming (Fi. 6). Consistent with the results from the previous cal-cium imaging experiments28, the AFD neuron of wild-type worms
cultivated at 23 C showed a marked response when warmed to atemperature above 23 C (Fi. 6). By contrast, the AFD neuron ofhsf-1 mutants cultivated at 23 C showed a marked response when
warmed to a temperature above 21 C; that is, the response ofhsf-1mutants to warming occurred earlier than that of wild-type worms(Fi. 6,). There was, however, hardly any observable difference in
the cultivation-temperature dependency of AWC response betweenwild-type worms and hsf-1 mutants (Fi. 6). Furthermore, expres-
sion ofhsf-1 in body wall muscles sufficiently rescued an abnormalresponse of the AFD neuron ofhsf-1 mutants (Fi. 6,). These results
indicate that the HSF-1 signaling is important for the cultivation tem-peraturedependent response of the AFD neuron.
Estrogen signaling acts downstream of HSF-1 signaling
We also addressed how the HSF-1 signaling regulates the thermo-sensory neurons. Among genes likely to act downstream of HSF-1,
dhs-4 and cyp-37B1 genes encode hydroxysteroid 17 dehydroge-
nase 11 (Wormbase gene ID, WBGene00000968) and a cytochromeP450 CYP4/CYP19/CYP26 subfamily protein (WBGene00009226),respectively, both of which are believed to be involved in estrogen
synthesis. In a local a lignment search using the sequences ofHomosapiens estradiol 17 dehydrogenase 1 (HSD17B1) and Homo sapiens
cytochrome P450 19A1, known as aromatase (CYP19A1) as queriesagainst the C. elegans genome database, both DHS-4 (25% identity,expected value (E-value) = 2 1021) and CYP-37B1 (23% identity,E-value = 3 109) were ranked in the top five matches. Recentreports have shown that estrogen and its receptor are involved in
hippocampal synaptic plasticity and memory through the regulationof calcium influx2931. We therefore assessed the effect of estrogen
signaling on C. elegans thermotaxis.
We exogenously applied 17-estradiol (referred to as estradiol fromhere on), a main form of estrogen, to wild-type worms grown at 23 C
from their birth. An application of estradiol caused thermophilicmovement in a dose-dependent manner (Fi. 7a). Further, estradiol
application on 23 Ccultivated hsf-1 mutants and dhs-4 mutants par-tially restored their behavioral defects (Fi. 7b,). Notably, estradiol
application only during the behavioral conditioning process causedthermophilic movement of wild-type worms and partially rescued
the behavioral defect ofhsf-1 mutants and dhs-4 mutants (Fi. 7
and Suppary Fi. 1). Despite the clearly observable effectsof estradiol on wild-type worms as well as hsf-1 mutants and dhs-4mutants, estradiol application to ttx-3 mutants did not induce any
obvious effect, suggesting that ttx-3 is genetically epistatic to estrogensignaling as well as to the HSF-1 signaling (Fi. 5, and 7b).
On the basis of the results from the estradiol application experi-ments, we investigated the relationship between HSF-1 signaling and
estrogen signaling. The previous in vitro binding analysis showedthat the three nuclear receptors NHR-14, NHR-69 and NHR-121,
all of which were identified by their highest homologies to humanestrogen receptor, bound estradiol in a dose-dependent fashion32.
We analyzed thermotaxis of mutants impaired in the nhr-14, nhr-69or nhr-121 genes. nhr-69 mutants cultivated at 23 C showed a sub-
stantial tendency to migrate toward colder regions than wild-type
worms, whereas nhr-14 mutants and nhr-121 mutants cultivated at17 C or 23 C showed no obvious defect (Fi. 8a). We further inves-tigated the functional redundancy between NHR-69 and NHR-14 or
NHR-121, examining thermotaxis ofnhr-69;nhr-121 mutants, nhr-69;nhr-14 mutants and nhr-69;nhr-121;nhr-14 mutants. When cul-tivated at 23 C, the behavioral defects in nhr-69;nhr-121 mutants,nhr-69;nhr-14 mutants and nhr-69;nhr-121;nhr-14mutants were eachsimilar to that ofnhr-69 single mutants (Fi. 8a), suggesting that
there is no redundancy between nhr-69 and the other two nhrgenesunder the 23 C-cultivated condition. By contrast, nhr-69;nhr-121
mutants, nhr-69;nhr-14 mutants and nhr-69;nhr-121;nhr-14 mutantseach cultivated at 17 C exhibited a slightly more severe behavioral
defect than 17 C-cultivated nhr-69 mutants (Fi. 8a), suggesting that
b ca
d4.00
TTXi
ndex
3.00
2.00
1.00
0
1.00
2.00
3.00
NS
NS
Wild type hsf-1 odr-3
**
******
**
ttx-3 hsf-1;ttx-3
gcy-23
gcy-8
gcy-18
hsf-1;
gcy-23
gcy-8
gcy-18
hsf-1;gcy-23gcy-8
gcy-18Ex[myo-3p::
hsf-1]
hsf-1;
odr-3
0.50
0.40
0.30
0.20
0.10
0
4 3 2 1 +1 +2 +3 +4
hsf-1
hsf-1;gcy-23
gcy-8 gcy-18hsf-1;gcy-23 gcy-8 gcy-18
Ex[myo-3p::hsf-1]
gcy-23 gcy-8 gcy-18
Wild type
Region
Fractionofanimals
ineachregion
4 3 2 1 +1 +2 +3 +4
0.50
0.40
0.30
0.20
0.10
0
Wild type
ttx-3
hsf-1
hsf-1;ttx-3
Fractionofanimals
ineachregion
0.50
0.40
0.30
0.20
0.10
4 3 2 1 +1 +2 +3 +4
0
Wild type
odr-3
hsf-1
hsf-1;odr-3
Region
Fractionofanimals
ineachregion
Figure 5 Genetic interactions between HSF-1 signaling
and the genes that act in the thermotactic neural circuit.
(ac) Genetic epistasis analysis. Thermotactic plasticity
o wild-type worms and each mutant was examined using
population thermotaxis (TTX) assays at 2 h ater shiting
the temperature rom 17 C to 23 C. The TTX plate
was divided into eight regions, 4 (coldest) to +4
(warmest) (Supplementary Fig. 1a). Epistatic relations
between hsf-1 mutants and gcy-23 gcy-8 gcy-18
mutants (a), odr-3mutants (b) or ttx-3mutants (c) are
shown. The ttx-7;hsf-1 double mutants were also
constructed, but it was diicult to analyze their
thermotactic behaviors owing to their severe locomotory
deects. (d) Summary o genetic epistasis analysis or the relationship between systemic temperature perception mechanism and thermotactic neuralcircuit. Thermotactic plasticity was examined as above. The signiicance o dierences o thermotactic plasticity was determined using the Fishers
protected least signiicant dierence multiple-comparison test. **P< 0.01; NS, not signiicant. Error bars represent s.e.m.
-
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
7/11
990 VOLUME 14 | NUMBER 8 | AUGUST 2011 nature neurOSCIenCe
a r t I C l e S
NHR-69 functions redundantly with NHR-14 and NHR-121 under
the 17 C-cultivated condition. Taken together, these data suggest that
at least NHR-69 is required for thermotaxis.We then conducted a genetic epistasis analysis between nhr-69 andhsf-1. The behavior of nhr-69 hsf-1 double mutants was similar to
that ofnhr-69 mutants (Fi. 8b), indicating that nhr-69 is epistatic tohsf-1. This result suggests that NHR-69, which might mediate both
thermophilic and cryophilic movements, could act downstream of theHSF-1 signaling to be involved in the former movement. To further
address this possibility, we took a pharmacological approach to testwhether estradiol application affects thermotaxis ofnhr-69 mutants.
Estradiol application had no substantial effect on the thermotacticbehavior ofnhr-69 mutants (Fi. 8,). Considering the effect of estra-
diol application on hsf-1 mutants (Fi. 7b), this result is consistentwith the idea that NHR-69 acts downstream of the HSF-1 signaling.
Given that ttx-3 is epistatic to estrogen signaling (Fi. 7b),
we further attempted to identify a site of action of NHR-69. We
examined whether the thermotactic behavior ofnhr-69 hsf-1 doublemutants was restored to that ofhsf-1 single mutants when NHR-69is expressed in AFD, AWC or AIY neurons ofnhr-69 hsf-1 mutants.
The expression of nhr-69 cDNA in AWC or AIY neurons had noobvious effects on the behavior ofnhr-69 hsf-1 mutants, but strong
expression ofnhr-69 (3 ng l1) in the AFD neuron of nhr-69 hsf-1mutants caused a strong tendency to migrate toward colder regions
rather than a restoration to the phenotype of hsf-1 single mutants(Fi. 8b). These results suggest that NHR-69 acts in AFD, although
excess expression ofnhr-69 in AFD on the HSF-1 signalingreducedbackground leads to an increase in inactive NHR-69 in the AFD,
which might cause a stronger tendency for nhr-69 hsf-1 mutantsto migrate to colder regions. Indeed, weaker expression of nhr-69
a b c
e
d
24
Ratiochange(%)
AFD neuron
23 C
17 C 17 C
Wild typehsf-1(sy441)hsf-1(sy441);Ex
[myo-3p::hsf-1]18
12
6
0
6
12
0 30 60 90 120
Time (s)
150 180 210 240
Ratiochange(%)
AFD neuron
15 C17 C
19 C21 C
23 C25 C
30
0 30 60 90 120
Time (s)
150 180 210 240
20100
20100
20100
AWC neuron
15 C17 C
19 C21 C
23 C25 C
0 30 60 90 120
Time (s)
150 180 210 240
Ratiochange(%)
302010
02010
02010
0
Wild
type
hsf-1
(sy4
41)
hsf-1
(sy4
41);
Ex[m
yo-3p::hsf-
1]
Ratiochange(%)
12NS
** **
9
6
3
0
AWC neuron
23 C
17 C 17 C
12
Ratiochange(%)
8
4
0
4
8
0 30 60 90 120
Time (s)
150 180 210 240
Wild typehsf-1(sy441)hsf-1(sy441);Ex
[myo-3p::hsf-1]
Figure 6 Regulation o the thermosensory neurons by the HSF-1 signaling. (a,b) In vivocalcium
imaging o the thermosensory neurons in 23 C-conditioned wild-type worms, hsf-1 mutants
and hsf-1 mutants expressing hsf-1 in body wall muscles during transient up-and-down warming.
This method o warming was used to examine calcium transients in AFD (a) and AWC (b). The average ratio changes o
AFD and AWC were measured or wild-type worms (blue), hsf-1 mutants (red) and hsf-1 mutants expressing hsf-1 inbody wall muscles (green). Stimulus temperatures are shown at the bottom. n = 11 to 17 worms. (ce) In vivocalcium
imaging o the thermosensory neurons in 23 C-conditioned wild-type worms, hsf-1 mutants and hsf-1 rescue lines
during step-wise warming. Temperature was increased in a step-wise manner to examine the temperature memory o
the AFD (c) and AWC (d) thermosensory neurons, which was set by the previous cultivation temperature. The average
ratio changes in AFD (c) and AWC (d) were measured each or wild-type worms (blue), hsf-1 mutants (red) and hsf-1
mutants expressing hsf-1 in body wall muscles (green). Arrows indicate the irst marked ratio changes during s tep-
wise warming. The average ratio change o the AFD neurons in response to the temperature change rom 19 C to 21 C
during step-wise warming is shown in a bar graph (e). **P< 0.01; NS, not signiicant (t-tests, n = 12 to 20 worms).
Error bars represent s.e.m.
3.00a
2.00
** **
** **
1.00
TTXi
ndex
0
1.00Concentration (M)
108
107
106
105
104
103
4.00b3.00
3.00
2.00
2.00
***
**
1.00
1.00TTXi
ndex
0
4.00
OffWild type h sf -1 d hs -4
ttx-3
NS
On Off On Off On
Off On
4.00d3.002.00
2.00
****
**
1.00
1.00TTXi
ndex
0
3.00
Off
Wild type h sf -1 d hs -4
ttx-3
NS
NS
NSNSNS
On Off On Off OnOff On
Off
Wild type h sf -1 d hs -4
On Off On Off On1.00
c
2.003.00
1.00
TTXi
ndex 0
4.00
ttx-3
Off On
Figure 7 Eect o estradiol application on thermotaxis. (a) The dose-
dependent eect o estradiol application. Wild-type worms were cultivated
at 23 C with indicated concentration o estradiol. To clearly observe the
thermophilic movement o worms cultivated at 23 C, the center o the
thermotaxis (TTX) plate was adjusted to 23 C to establish a linear thermal
gradient ranging rom approximately 20 C to 26 C. However, other behavioral
assays were conducted by adjusting the center o the TTX plate to 20 C
to establish a thermal gradient ranging rom approximately 17 C to 23 C
(see Supplementary Fig. 1). (b,c) Eect o exogenously applied estrogen on
thermotactic behaviors o wild-type worms and several mutants. Each worm
was cultivated at 23 C (b) or 17 C (c), either with (on) or without (o) 10 M
estradiol. (d) Eect o estradiol on the behavioral conditioning process.
Assay was conducted as described in Supplementary Figure 1c. *P< 0.05,
**P< 0.01 (t-tests). Error bars represent s.e.m.
-
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
8/11
nature neurOSCIenCe VOLUME 14 | NUMBER 8 | AUGUST 2011 99 1
a r t I C l e S
(0.3 ng l1) in AFD restored the thermo-
tactic behavior ofnhr-69 hsf-1 double mutants
(Fi. 8b). To further examine whether theexpression ofnhr-69 in AFD affects the AFD
activity in 23 C-cultivated nhr-69 hsf-1mutants, we measured calcium transients
of the AFD in nhr-69 hsf-1 mutants and themutants expressing nhr-69 cDNA in the AFD
at 0.3 ng l1 or 3 ng l1. Consistent withthe results from behavioral assays (Fi. 8b),
expression of nhr-69 in the AFD neuron of nhr-69 hsf-1 mutants
affected the AFD activity in a dose-dependent manner (Fi. 8).Together with the results from the exogenous estrogen application,genetic and calcium imaging analyses, this suggests that HSF-1 sig-
naling acts at least in part through estrogen signaling, directly orindirectly, to regulate the thermosensory neuron AFD.
DISCUSSION
In the current study, we found that systemic temperature signal-ing regulates the thermosensory neurons for thermotaxis through
heat-shock transcription factor and estrogen signaling. The heat-shock transcription factor has been well known as a thermosensor
that is highly conserved from organisms as distant as Saccharomyces
cerevisiae. By contrast, the AFD neuron is a specialized thermosensor
with sensitivity enough to respond to differences of 0.01 C on a
thermal gradient8. We therefore infer that C. elegans added its finetemperature sensitivity through the evolution of the AFD thermosen-sory system. Given that the HSF-1 protein responds to temperature
through its multimerization ability, it is unlikely that HSF-1 is as sen-
sitive as AFD. Alternatively, as a plausible purpose of the temperatureresponse, HSF-1 might function during behavioral conditioning after
temperature shifts from 17 C to 23 C in an incubator, as microarrayanalysis, quantitative PCR analysis and GFP observation indeed indi-
cated that expression of several genes downstream of HSF-1 changedduring this process (Fis. 1h and 3b and Suppary Fis. 2
and 3a,b,). Although the AFD neurons can respond to temperaturedifferences of 0.01 C for isothermal tracking, worms were broadly
distributed around the cultivation temperature in the temperature
gradient in population assays (Fi. 5a). It is probable that C. elegans
possesses multiple types of thermosensory systems that are not mutu-ally exclusive.Given the recent studies indicating that the sensory system
operating behavior is more diverse than previously thought33, it isreasonable to consider the possibility that animals can perceive envi-
ronmental changes not only through a specialized sensory systembut also through sensors distributed across the body. For example, a
small set of warmth-activated anterior cell neurons is located in theinternal portion of the Drosophila melanogasterbrain, whose func-
tion is important for preferred temperature selection34. In addition,D. melanogaster larvae perceive light by using neurons embedded
under the cuticle encasing their bodies35. This sensory system doesnot eliminate the importance of photoreceptors in eyes. In C. elegans,
the PVD neurons that are located on either side of the worm and
that envelop the body with multidendritic processes respond to coldtemperatures36. Thus, it is possible to suppose that an HSF-1-mediatedmechanism might be responding to a signal from other neurons,
such as PVD neurons. Although further work is needed to clarify
the relationship among the various temperature signaling systems,this study and others3336 suggest that sensory systems influencing a
behavior might be broadly required throughout an animals body.A principal view in neuroscience is that environmental temperature
is detected by sensory neurons in the nervous system. We demon-strate here that temperature signaling by non-neuronal cells triggers
behavioral changes, as thermosensory neurons do. Given the robustinfluence of temperature on animals bodies, it improves behavioral
performance to monitor environmental temperatures at the level of
b
c e
d
NS
NSNS
NS
NS
NS
NS
****
3 0.3
gcy-8p ceh-36p ttx-3p
ng l1
:
nhr-69 hsf-1 worms expressingpromoter::nhr-69 cDNA
0.03 3 3
**
4.00
3.00
2.00
1.00
0
1.00
2.00
60 AFD neuronWild type
nhr-69 hsf-1
nhr-69 hsf-1 wormsexpressinggcy-8p::nhr-69 (0.3 ng l
1)
nhr-69 hsf-1 worms expressing
gcy-8p::nhr-69 (3 ng l1
)
50
40
30
Ratiocha
nge(%)
20
10
0
0 30
17 C 17 C
23 C
60 90 120
Time (s)
150 180 210 240
4.00
Off On
Wild type
Wild type
nhr-69
nhr-69
Off On
Off On Off On
3.00
2.00
1.00
0
0
1.00
2.00
3.00
TTXi
ndex
TTXi
ndex
TTXi
ndex
a NS
NS
NS
NS
NS
** **
** ** **
****
3.00
2.00
1.00
0
0
1.00
2.00
3.00
TTXi
ndex
TTXi
ndex
Wild
type
nhr-6
9
hsf-1
nhr-6
9hs
f-1
nhr-1
4
nhr-6
9
nhr-1
21
nhr-69
;nhr
-14
nhr-6
9;nhr
-121
nhr-69
;nhr
-121
;
nhr-1
4
Figure 8 HSF-1 signaling acts through estrogen
signaling to regulate the AFD thermosensory
neurons. (a) The thermotactic (TTX) behavior
o the mutants impaired in the putative
estrogen receptors. nhr-14(tm1473)mutants,
nhr-69(tm2365)mutants, nhr-121(tm1797)
mutants and their double or triple mutants were
cultivated at 23 C or 17 C and then assayed.
(b) Genetic interactions between hsf-1 and
nhr-69. Thermotaxis o worms was examined
2 h ater shiting temperature rom 17 C
to 23 C. Concentrations o nhr-69cDNA
expressed in nhr-69 hsf-1 mutants under cell-
speciic promoter were indicated. (c,d) Eect
o estradiol on thermotaxis o nhr-69mutants.
Worms were grown at 23 C (c) or 17 C (d) with
or without 100 M o estradiol. (e) Calcium
imaging o AFD neurons in 23 C-conditioned
wild-type worms (n = 9), nhr-69 hsf-1 mutants
(n = 12) and nhr-69 hsf-1 mutants expressing
nhr-69in AFD neurons at the concentration
o 0.3 ng l1 (n = 15) or 3 ng l1 (n = 13).
**P< 0.01; NS, not signiicant (t-tests). Error
bars represent s.e.m.
-
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
9/11
992 VOLUME 14 | NUMBER 8 | AUGUST 2011 nature neurOSCIenCe
a r t I C l e S
the entire body. To ensure that this sort of strategy is retained throughevolution, heat-shock factor, a highly conserved protein, might adopt
a role in transcriptionally coupling environmental recognition to amemory-regulated behavior, such as thermotaxis in C. elegans.
METHODS
Methods and any associated references are available in the online
version of the paper at http://www.nature.com/natureneuroscience/.
Assi s. Gene Expression Omnibus: Microarray data have
been deposited with accession code GSE28856.
Note: Supplementary information is available on the Nature Neuroscience website.
AcknowledgmentS
We thank N. Hisamoto (Nagoya University), T. Mizuno (Nagoya University)and K. Matsumoto (Nagoya University) for sharing strains; A. Fire (StanfordUniversity School of Medicine) for pPD plasmids; Y. Iino (University of Tokyo)for thegcy-5 andgcy-7promoters; Caenorhabditis Genetic Center, C. elegansKnockout Consortium and S. Mitani at the National Bioresource Project, Japan,for strains; K. Terauchi and T. Kondo for kindly sharing the microarray apparatus;T. Inada for kindly sharing the quantitative PCR apparatus; C. Bargmann,S. Takagi, N. Hisamoto, A. Kuhara, P. Jurado, H. Sasakura, N. Ohnishi, T. Kimata,
M. Nonomura and members of the Mori laboratory for comments on thismanuscript and discussion. I.M. is a Scholar of the Institute for Advanced Researchin Nagoya University, Japan. This work was supported by the Core Research forEvolutional Science and Technology (CREST) Program of the Japan Science andTechnology (JST) agency (to I.M.).
AUtHoR contRIBUtIonS
T.S. designed the research, performed most experiments, analyzed data and wrotethe manuscript; Y.N. performed the quantitative PCR experiments and conductedgermline transformation to construct the C. elegans transgenic line; I.M. supervisedthe project, conducted initial identification of cells expressing the hsf-1 promoter::gfpreporter gene and wrote the manuscript.
comPetIng FInAncIAl InteReStS
The authors declare no competing financial interests.
Published online at http://www.nature.com/natureneuroscience/.Reprints and permissions information is available online at http://www.nature.com/
reprints/index.html.
1. Conti, B. et al. Transgenic mice with a reduced core body temperature have an
increased lie span. Science314, 825828 (2006).
2. Bear, M.F., Connors, B.W. & Paradiso, M.A.A. Neuroscience: Exploring the Brain
(Lippincott Williams & Wilkins, Maryland, USA, 2001).
3. Hedgecock, E.M. & Russell, R.L. Normal and mutant thermotaxis in the nematode
Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA72, 40614065 (1975).
4. Mori, I. & Ohshima, Y. Neural regulation o thermotaxis in Caenorhabditis elegans.
Nature376, 344348 (1995).
5. Mohri, A. et al. Genetic control o temperature preerence in the nematode
Caenorhabditis elegans. Genetics169, 14371450 (2005).
6. Ito, H., Inada, H. & Mori, I. Quantitative analysis o thermotaxis in the nematode
Caenorhabditis elegans. J. Neurosci. Methods154, 4552 (2006).
7. Biron, D. et al. A diacylglycerol kinase modulates long-term thermotactic behavioral
plasticity in C. elegans. Nat. Neurosci.9, 14991505 (2006).
8. Mori, I., Sasakura, H. & Kuhara, A. Worm thermotaxis: a model system or analyzingthermosensation and neural plasticity. Curr. Opin. Neurobiol.17, 712719 (2007).
9. Kuhara, A. et al. Temperature sensing by an olactory neuron in a circuit controlling
behavior o C. elegans. Science320, 803807 (2008).
10. Hsu, A.L., Murphy, C.T. & Kenyon, C. Regulation o aging and age-related disease
by DAF-16 and heat-shock actor. Science300, 11421145 (2003).
11. Akerelt, M., Morimoto, R.I. & Sistonen, L. Heat shock actors: Integrators o
cell stress, development and liespan. Nat. Rev. Mol. Cell Biol. 11, 545555
(2010).
12. Zimarino, V. & Wu, C. Induction o sequence-specic binding o Drosophilaheat
shock activator protein without protein synthesis. Nature327, 727730 (1987).
13. Mosser, D.D., Kotzbauer, P.T., Sarge, K.D. & Morimoto, R.I. In vitroactivation o
heat shock transcription actor DNA-binding by calcium and biochemical conditionsthat aect protein conormation. Proc. Natl. Acad. Sci. USA 87, 37483752
(1990).
14. Goodson, M.L. & Sarge, K.D. Heat-inducible DNA binding o puried heat shock
transcription actor 1. J. Biol. Chem.270, 24472450 (1995).
15. Morimoto, R.I. Regulation o the heat shock transcriptional response: cross talk
between a amily o heat shock actors, molecular chaperones, and negative
regulators. Genes Dev.12, 37883796 (1998).
16. Prahlad, V., Cornelius, T. & Morimoto, R.I. Regulation o the cellular heat shock
response in Caenorhabditis elegans by thermosensory neurons. Science 320,
811814 (2008).
17. Clos, J., Rabindran, S., Wisniewski, J. & Wu, C. Induction temperature o human
heat shock actor is reprogrammed in a Drosophila cellenvironment. Nature364,
252255 (1993).
18. Devaney, E. Thermoregulation in the lie cycle o nematodes. Int. J. Parasitol.36,
641649 (2006).
19. Hajdu-Cronin, Y.M., Chen, W.J. & Sternberg, P.W. The L-type cyclin CYL-1 and the
heat-shock-actor HSF-1 are required or heat-shock-induced protein expression in
Caenorhabditis elegans. Genetics168, 19371949 (2004).
20. Bargmann, C.I., Hartwieg, E. & Horvitz, H.R. Odorant-selective genes and neuronsmediate olaction in C. elegans. Cell74, 515527 (1993).
21. Okochi, Y., Kimura, K.D., Ohta, A. & Mori, I. Diverse regulation o sensory signaling
by C. elegansnPKC-epsilon/eta TTX-4. EMBO J.24, 21272137 (2005).
22. Morley, J.F. & Morimoto, R.I. Regulation o longevity in Caenorhabditis elegansby
heat shock actor and molecular chaperones. Mol. Biol. Cell 15, 657664
(2004).
23. Biron, D., Wasserman, S., Thomas, J.H., Samuel, A.D. & Sengupta, P. An olactory
neuron responds stochastically to temperature and modulates Caenorhabditis elegans
thermotactic behavior. Proc. Natl. Acad. Sci. USA105, 1100211007 (2008).
24. Hobert, O. et al. Regulation o interneuron unction in the C. elegansthermoregulatory
pathway by the ttx-3LIM homeobox gene. Neuron19, 345357 (1997).
25. Altun-Gultekin, Z. et al. A regulatory cascade o three homeobox genes, ceh-10,
ttx-3and ceh-23, controls cell ate specication o a dened interneuron class in
C. elegans. Development128, 19511969 (2001).
26. Zariwala, H.A., Miller, A.C., Faumont, S. & Lockery, S.R. Step response analysis o
thermotaxis in Caenorhabditis elegans. J. Neurosci.23, 43694377 (2003).
27. Lee, S.J. & Kenyon, C. Regulation o the longevity response to temperature by
thermosensory neurons in Caenorhabditis elegans. Curr. Biol. 19, 715722(2009).
28. Kimura, K.D., Miyawaki, A., Matsumoto, K. & Mori, I. The C. elegansthermosensory
neuron AFD responds to warming. Curr. Biol.14, 12911295 (2004).
29. Liu, F. et al. Activation o estrogen receptor-beta regulates hippocampal synaptic
plasticity and improves memory. Nat. Neurosci.11, 334343 (2008).
30. Woolley, C.S. Acute eects o estrogen on neuronal physiology. Annu. Rev.
Pharmacol. Toxicol.47, 657680 (2007).
31. Wu, M.V. et al. Estrogen masculinizes neural pathways and sex-specic behaviors.
Cell139, 6172 (2009).
32. Mimoto, A. et al. Identication o an estrogenic hormone receptor in Caenorhabditis
elegans. Biochem. Biophys. Res. Commun.364, 883888 (2007).
33. Lumpkin, E.A. & Caterina, M.J. Mechanisms o sensory transduction in the skin.
Nature445, 858865 (2007).
34. Hamada, F.N. et al. An internal thermal sensor controlling temperature preerence
in Drosophila. Nature454, 217220 (2008).
35. Xiang, Y. et al. Light-avoidance-mediating photoreceptors tile the Drosophilalarval
body wall. Nature468, 921926 (2010).
36. Chatzigeorgiou, M. et al. Specic roles or DEG/ENaC and TRP channels in touch
and thermosensation in C. elegans nociceptors. Nat. Neurosci. 13, 861868(2010).
http://www.nature.com/natureneuroscience/http://www.nature.com/natureneuroscience/ -
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
10/11
nature neurOSCIenCedoi:10.1038/nn.2854
ONLINE METHODSgis. The hsf-1(sy441) mutation19 was outcrossed to the wild type six times.
Thesy441 mutant has a single G-to-A mutation in the seventh exon (5-AGT AAT
CAT AAT TGG GAT GAT TTT GGG AAT-3), which would convert residue
585 (tryptophan) to a stop codon, leading to the truncation of the last 86 amino
acids at the C terminus. Double, triple and quadruple mutants in this study were
constructed by genetic and molecular methods. hsf-1(ok600) is a null allele and
is kept as a balanced heterozygous strain.
gri rasfrai. Germline transformations were performed
as described37. For cell-specific rescue experiments, co-injection mixes
consisted of pKDK66ges-1p::NLS GFP (50 ng l1) and experimental DNA
at various concentrations (1 or 10 ng l1). To analyze the expression
pattern of HSF-1, pTAK16 hsf-1p::gfp (20 ng l1) were injected into wild-
type worms. For GFP observation, pTAN51 hsp-16.2p::gfp (50 ng l1)
and pTAK14 hsp-16.2p(hsemut)::gfp (50 ng l1) were co-injected with pNAS88
ges-1p::NLS TaqRFP(50 ng l1) and pTAN124ges-1p::NLS DsRed monomer
(50 ng l1), respectively.
RnA isai a apifiai. For synchronization of worm growth38,
worms were initially grown to starvation in a 60-mm plate of nematode growth
medium (NGM) at 23 C. Half of the agar in the plate was split into four pieces,
and two pieces were placed on a 150-mm plate. The resultant plates were incu-
bated at 23 C until most of the food was depleted. The worms were collectedfrom the plates using M9 buffer (22 mM KH2PO4, 42 mM Na2HPO4, 85 mM
NaCl and 1 mM MgSO4) and then resuspended in 75 ml bleach solution (15 ml
sodium hypochlorite (Kao), 3.75 ml 10 M NaOH and 56.25 ml water). Worms
were transferred to a 200-ml glass beaker with a stir bar and incubated for
5 min while stirring rapidly. When most adults had burst, the solution was passed
through a 53-m nylon mesh (Tanaka Sanjiro Co.) to separate intact embryos
from worm debris. Embryos were collected by centrifugation, resuspended in
2 ml M9 buffer and split equally over six 150-mm NGM plates. The plates were
incubated at 23 C for 2.5 d for microarray analysis. The hatched worms were
collected from the plates using S-basal buffer (5.7 mM K2HPO4, 44 mM KH2PO4,
100 mM NaCl and 5 g ml1 cholesterol), and a fraction of worms was used
for the temperature shift thermotaxis assay, whereas the remaining worms were
killed and dissolved in TRI reagent (Ambion) for microarray analysis. mRNA was
isolated before shifting the temperature from 23 C to 17 C (0 h at 23 C) and at
every 1 h after shift ing temperature to 17 C (1, 2, 3 or 4 h at 17 C) (Fi. 1a). TotalRNA was isolated from the lysate using the RiboPure kit (Ambion).
The high-fidelity linear amplification of 40 ng of isolated mRNA was accom-
plished using the WT-Ovation Pico RNA Amplification System (NuGEN)
according to the manufacturers protocol, with some modifications39. The ampli-
fied cDNA was fragmented and biotinylated using FL-Ovation cDNA Biotin
Module V2 (NuGEN).
mirarray. Fragmented, biotin-labeled cDNA (50l) was mixed with hybridi-
zation cocktail (Affymetrix) and hybridized to the Affymetrix C. elegans chip
for 18 h at 45 C, according to the Affymetrix Expression Analysis Technical
Manual. The hybridized arrays were washed and scanned using the GeneChip
Scanner 3000 TG system. The C. elegans chip was designed using the December
2000 genome sequence.
Microarray data were processed as described38,40,41. Primary intensity values
from each hybridization were scaled against a global average signal from the samearray and normalized with Robust Multichip Average analysis (RMA). To identify
differentially expressed transcripts during the worms temperature memory acqui-
sition process, normalized intensity values from gene expression data sets of worms
cultivated for 4 h after shifting the temperature to 17 C were compared with those
from reference gene expression data sets of worms isolated before shifting tempera-
ture, using Significance Analysis of Microarray software (SAM)38,40,41.
Aaysis f irarray aa. A two-class paired analysis of the data was per-
formed to identify genes that differed by1.8-fold from the reference at a false
discovery rate of 1.12% for expression data sets of RNA isolated at 4 h after shift-
ing temperature to 17 C. In this set of arrays, 46 upregulated and 33 downregu-
lated genes (Suppary Fi. 2) were found to be significantly regulated, with
6.813 median false significant genes. The 79 differently expressed genes were
classified according to their gene ontology annotations using GoMiner software
(http://discover.nci.nih.gov/gominer/)40. To validate microarray data, the expres-
sion of some of genes was tested by quantitative PCR.
gFP bsrvai. All observations were carried out using an Axioplan2 light
microscope (Carl Zeiss) except for those in Fiur 4a, in which the fluo-
rescence images were taken with a Fluoview FV1000 confocal laser scanning
microscope (Olympus). Worms expressing reporter genes, hsp-16.2p::gfp or
hsp-16.2p(hsemut)::gfp, were cultivated at 15 C. Adult worms were transferredto the new cultivation temperature of 20 C, 23 C or 25 C, or to the heat-
shock temperature of 30 C and 35 C, and their GFP fluorescence was moni-
tored at 2 h or at each time point as shown in Suppary Fiur 3. The
GFP intensity of each worm was calculated as described in the figure legend
ofSuppary Fiur 3 and was quantified using ImageJ software (US
National Institutes of Health).
thraxis assay. A radial temperature gradient assay was carried out using a
9-cm agar plate and a vial containing frozen acetic acid, according to previously
described methods4,5. A thermotaxis assay using a linear temperature gradient
was also performed as previously reported6. Equipment for establishing the linear
thermal gradient was used as previously described3. A stable, linear thermal
gradient was established on a thin, 60-cm long aluminum platform, one end of
which was placed in a water bath at 5 C, with the opposite end in a water bath
at 35 C. A thermotaxis plate (13.5 cm 6 cm, 1.8 cm height) containing 10 mlof thermotaxis medium (3 g l1 NaCl, 20 g l1 Bacto Agar (Becton Dickinson)
and 25 mM KPO4) was placed on the aluminum platform such that the tem-
perature gradient could be established along the agar surface of 13.5 cm. The
center of the 13.5-cm-long agar surface in the thermotaxis plate was adjusted
to 20 C and the thermotaxis plate was maintained for 15 min, until a linear
thermal gradient ranging from approximately 17 C to 23 C was established.
Worms were grown on a 6-cm plate containing 14 ml of NGM with agar, on
which E. coli OP50 was seeded; each worm and its progeny were cultured at
17 C or 23 C. Worms were collected with 1 ml of S-basal buffer kept at 20 C
and were washed once with autoclaved water at 20 C. These steps were carried
out within 7 min in a water bath with constantly maintained temperature of
20 C. Approximately 80200 worms were placed at the center of the thermo-
taxis plate. Excess water surrounding the worms was removed with tissue paper.
After 60 min, the worms were killed by chloroform gas, and the worms in each
of the eight regions were scored. The thermotaxis index was calculated from theformula shown in Suppary Fiur 1a.
For the temperature shift assay based on the population thermotaxis assay5,
wild-type worms, hsf-1 mutants, hsf-1-overexpressing worms and other mutants
were cultivated with food. On the first day, two wild-type worms or three mutants
were each deposited on the NGM plate and cultivated at 17 C. After 12 h , the
deposited postnatal day 0 (P0) worms were removed from each plate to segregate
F1 progeny. In the case of the temperature shift assay from 17 C to 23 C, F1
worms were shifted to a new temperature of 23 C after cultivation at 17 C for
4.5 d (after deposition at P0 on the NGM plate), and the population thermotaxis
assay was performed at each time point (Fi. 2a). In the case of the temperature
shift assay from 23 C to 17 C, F1 worms were initially cultivated at 17 C until
they passed the L1L2 larval stage and were then cultivated at 23 C because of
the L1L2 larval arrest phenotype inhsf-1 mutants at 25 C (ref. 19). After worms
grew to the adult stage, the temperature was shifted from 23 C to 17 C, and the
thermotaxis assays were conducted at each time point (Fi. 2b).To evaluate the effect of estrogen on the thermotactic behavior, 17-estradiol
(Sigma, E2275-1G) was applied to wild-type worms or several mutants. A 1,000-
fold concentrated solution of estradiol was prepared in dimethyl sulfoxide, and
the solution was added to each 6-cm plate containing 14 ml of agar and a bacterial
lawn to obtain the NGM plate containing estradiol at the final concentration.
Worms were cultivated on the NGM plate supplied with estradiol at 17 C or 23 C
during behavioral conditioning or until the adult stage; worms cultivated without
estradiol were cultivated with dimethyl sulfoxide as a vehicle control.
Quaiaiv PcR aaysis. Wild-type worms and hsf-1 mutants were individu-
ally synchronized as described above. In the case of the temperature shift from
15 C to 25 C, after cultivation at 15 C for 5 d, mRNA was isolated by RiboPure
kit (Ambion) just before shifting the cultivation temperature to a new temperature
http://discover.nci.nih.gov/gominer/http://discover.nci.nih.gov/gominer/ -
8/3/2019 Regulation of Behavioral Plasticity by Systemic Temperature Signaling in Caenorhabditis Elegans
11/11
nature neurOSCIenCe doi:10 1038/nn 2854
25 C and at 5 h after shifting temperature. In the case of the temperature shift
from 25 C to 15 C, both the wild-type worms andhsf-1 mutants were initially
cultivated at 17 C until they passed the L1L2 larval stage and then cultivated
at 25 C until the adult stage. The isolated mRNA (300 ng) was transcribed
and amplified using primers specific to the genes, SuperScript III Platinum
SYBR Green One-Step qRT-PCR Kit (Invitrogen) and a DNA Engine Peltier
Thermal Cycler-200 (Bio-Rad). To estimate the fold-change value (Fi. 3b and
Suppary Fis. 3 and 8i,j), the mRNA abundance of each gene isolated
from worms at 5 h after shifting the temperature from 15 C to 25 C or from25 C to 15 C was divided by that of the worms measured before shifting the
temperature. Expression changes of each gene were normalized to that of the
house-keeping gene lmn-1 as a reference.
caiu iai aaysis.In vivo calcium imaging was performed according to
previously reported methods9,28, but with some modifications. To monitor the
temperature-evoked response of each neuron, yellow cameleon 3.6 was expressed
in each worm. These worms were glued onto a 1.5% agar pad on glass, immersed
in M9 buffer and covered with a glass coverslip. The agar pad and M9 buffer were
kept at the initial imaging temperature. Sample preparation was completed within
2 min. The sample was then placed onto a Peltier-based thermocontroller (Tokai
Hit, Japan) on the stage of an Olympus BX61WI microscope at the initial imag-
ing temperature for 2 min, and fluorescence was introduced into a Dual-View
optics system (Molecular Devices). Cyan fluorescent protein (CFP; F480) and
yellow fluorescent protein (YFP; F535) images were simultaneously capturedby an electron-multiplying charge-coupled device camera, C9100-13 ImagEM
(Hamamatsu Photonics). For calcium imaging during transient up-and-down
warming and step-wise warming, images were taken with a 500-ms exposure
and 1 1 binning. For Fiur 8, to protect the probe from photobleaching
during measurement, each image was taken with a 100-ms excitation pulse at a
1-Hz frame rate using a pulse-generator-equipped imaging apparatus (SG-4115,
Iwatsu). The temperature on the agar pad was monitored by a thermometer sys-
tem, DCM-20 (Tokai Hit and Hamamatsu Photonics). Fluorescence intensities
of F535 and F480 were measured using the MetaMorph imaging analysis system
(Molecular Devices).
Saisia aaysis. The statistical analysis forFiurs 4i and6 was performed
using two-way analysis of variance followed bypost hoc analysis using the Fishers
protected least significant difference test, whereas that forFiurs 2, 3b,
5, 7a and 8a and Suppary Fiur 3 was conducted using t-tests.
Differences were considered to be significant when P< 0.05.
37. Mello, C.C., Kramer, J.M., Stinchcomb, D. & Ambros, V. Ecient gene transer in
C.elegans: extrachromosomal maintenance and integration o transorming
sequences. EMBO J.10, 39593970 (1991).
38. Von Stetina, S.E. et al. Cell-specic microarray proling experiments reveal a
comprehensive picture o gene expression in the C. elegans nervous system.
Genome Biol.8, R135 (2007).
39. Watson, J.D. et al. Complementary RNA amplication methods enhance microarray
identication o transcripts expressed in the C. elegans nervous system.
BMC Genomics9, 8497 (2008).
40. Irizarry, R.A. et al. Summaries o Aymetrix GeneChip probe level data. Nucleic
Acids Res.31, e15 (2003).
41. Murphy, C.T. et al. Genes that act downstream o DAF-16 to infuence the liespan
o Caenorhabditis elegans. Nature424, 277283 (2003).