Current Biology 22, 743–752, May 8, 2012 ª2012 Elsevier Ltd All rights reservedDOI 10.1016/j.cub.2012.02.066
Optogenetic Analysis of a Nociceptor Neuron
and Network Reveals Ion Channels
Acting Downstream of Primary Sensors
Steven J. Husson,1,2Wagner Steuer Costa,1,2
Sebastian Wabnig,1,2Jeffrey N. Stirman,3
Joseph D. Watson,4W. Clay Spencer,4,5Jasper Akerboom,6
Loren L. Looger,6Millet Treinin,7David M. Miller, III,4,5
Hang Lu,3and Alexander Gottschalk1,2,*
1Buchmann Institute for Molecular Life Sciences,
Goethe-University Frankfurt, Max von Laue Str. 15,
60438 Frankfurt, Germany
2Institute of Biochemistry, Goethe-University,
Max von Laue Str. 9, 60438 Frankfurt, Germany
3Interdisciplinary Bioengineering Program, School of
Chemical & Biomolecular Engineering, 311 Ferst Drive,
Georgia Institute of Technology, Atlanta, GA 30332-0100, USA
4Department of Cell and Developmental Biology 3120 MRB III,
465 21stAvenue South
5Program in Neuroscience
Vanderbilt University, Nashville, TN 37323-8240, USA
6Howard Hughes Medical Institute, Janelia Farm Research
Campus, Ashburn, VA 20147, USA
7Department of Medical Neurobiology, IMRIC, Hebrew
University-Hadassah Medical School, Jerusalem 91120, Israel
Background: Nociception generally evokes rapid withdrawal
behavior in order to protect the tissue from harmful insults.
Most nociceptive neurons responding to mechanical insults
display highly branched dendrites, an anatomy shared by
Caenorhabditis elegans FLP and PVD neurons, which mediate
harsh touch responses. Although several primary molecular
nociceptive sensors have been characterized, less is known
about modulation and amplification of noxious signals within
nociceptor neurons. First, we analyzed the FLP/PVD network
by optogenetics and studied integration of signals from these
cells in downstream interneurons. Second, we investigated
which genes modulate PVD function, based on prior single-
neuron mRNA profiling of PVD.
Results: Selectively photoactivating PVD, FLP, and down-
stream interneurons via Channelrhodopsin-2 (ChR2) enabled
the functional dissection of this nociceptive network, without
interfering signals by other mechanoreceptors. Forward or
reverse escape behaviors were determined by PVD and
FLP, via integration by command interneurons. To identify
mediators of PVD function, acting downstream of primary
nocisensor molecules, we knocked down PVD-specific tran-
scripts by RNAi and quantified light-evoked PVD-dependent
voltage-gated Ca2+channels (VGCCs) showed that PVD sig-
nals chemically to command interneurons. Knocking down
the DEG/ENaC channel ASIC-1 and the TRPM channel GTL-1
indicated that ASIC-1 may extend PVD’s dynamic range and
that GTL-1 may amplify its signals. These channels act cell
autonomously in PVD, downstream of primary mechanosen-
Conclusions: Our work implicates TRPM channels in modi-
fying excitability of and DEG/ENaCs in potentiating signal
output from a mechano-nociceptor neuron. ASIC-1 and
GTL-1 homologs, if functionally conserved, may denote valid
targets for novel analgesics.
Several protein classes are implicated in the primary sensory
signal transduction machinery in nociceptive cells, e.g., tran-
sient receptor potential (TRP) channels, degenerins/epithelial
(K2P) [1–4]. Nociceptor neuron signaling is facilitated or
modulated by further molecules like neurotransmitters and
neuropeptides, eicosanoids, neurotrophins, cyto- and chemo-
kines, voltage-gated Na+, K+, and Ca2+channels, opioid and
purinergic receptors, and TRP channels . Dissection of
neural circuits and identification of nociceptor modulators is
challenging in higher animals, which possess myriads of
contributing neurons. In contrast, only a few neurons mediate
nociception in C. elegans, which responds to aversive stimuli
like touch, heat, odorants, toxins, and nonphysiological osmo-
larity or pH . Six touch receptor neurons (TRN) detect
‘‘gentle touch’’ (e.g., by an eyelash) to anterior or posterior
regions, evoking forward or reverse withdrawal reflexes,
respectively . Nociceptive ‘‘harsh touch,’’ e.g., prodding
with a platinum wire, is sensed by four additional neurons,
pairs of FLP and PVD. These cover head (FLP) or body (PVD)
with extensive dendritic arbors (Figure 1A) and also respond
to acute heat or cold, respectively [7–10]. The DEG/ENaCs
MEC-10 and DEGT-1 were suggested as putative components
of the nociceptive mechanotransduction channel in PVD
(though MEC-10’s contribution was questioned [3, 11]),
whereas TRPA-1 confers cold responses, as indicated by
Ca2+imaging . Additional neurons putatively contributing
to harsh touch sensation were identified by laser ablation
and scoring fractions of animals responding to harsh touch
insults to different body regions .
PVD directs equal numbers of synapses to the command
interneurons PVC and AVA, which control forward and back-
ward movement, respectively, and FLP mainly synapses to
the backward-command neurons AVA and AVD  (Figure S1
available online). Harsh mechanical stimuli agitate large
portions of the body, probably coactivating FLP and PVD.
Animals must integrate mechanosensory input to predict
optimal escape behavior (e.g., from predators). Thus the rela-
tive activation of various sensory and command neurons
probably determines locomotor output.
Analyzing behavioral effects of hitting a worm with a wire in
a cell-specific and quantifiable manner is difficult; thus it is
unknown how these multidendritic neurons orchestrate the
prevailing behavioral response. Unlike prodding, optogenetic
tools such as the light-gated cation channel ChR2 allow selec-
tive noninvasive photostimulation of neurons of interest
without perturbing others, particularly in the thin, transparent
worm. Also, photostimulation can be performed more control-
lably and consistently. This approach enabled the triggering of
sensory neuron activity to phenocopy endogenous behaviors,
e.g., in C. elegans TRNs [13, 14], putative harsh touch cells ,
aversive chemosensory neurons , and in multidendritic
neurons in Drosophila larvae . Here, we optogenetically
dissect an entire harsh touch neuronal network at the single-
neuron level. Because ChR2 directly depolarizes PVD, thus
bypassing the primary mechanotransduction channels, we
could uncover genes required for nociceptor function within
PVD, downstream of primary sensory molecules. PVD evokes
behavior across just three synaptic layers, including the NMJ.
Hence, quantifying escape velocity of knockdown or knockout
lines provided an accurate readout for PVD functionality. We
demonstrate that the TRPM channel GTL-1 probably amplifies
PVD signals, whereas the DEG/ENaC ASIC-1 facilitates signal
output from PVD and determines the promptness of the
Photoactivation of PVD Evokes Rapid Forward Escape
To study the PVD-associated neural network, we sought
to stimulate PVD without concomitantly activating other
mechanosensors. We coexpressed ChR2(H134R)::mCherry
and GFP via the F49H12.4 promoter  (zxIs12; Figure 1A)
in lite-1(ce314) mutants  to avoid photophobic responses.
When raised in the presence of the ChR2 chromophore all-
trans retinal (ATR), zxIs12 animals illuminated with blue-light
pulses of 0.2, 1, or 5 s (Figure 1B; Movie S1) showed rapid
forward escape responses. In contrast, a previous study
reported reversals upon midbody harsh touch . This indi-
cates differencesin optical(i.e., PVD-specific) versus mechan-
ical harsh touch stimulus perception (probably involving
additional mechanoreceptive neurons). The evoked behavior
Figure 1. Photostimulation of PVD Results in
Forward Escape Behavior
(A) Confocal stacks of an adult animal carrying
pF49H12.4::GFP], expressing GFP and ChR2::
mCherry in PVD (cell body and branched arbors),
in addition to neuron AQR and one unidentified
(B) lite-1(ce314); zxIs12 animals were challenged
with blue light pulses of 0.2, 1, and 5 s (indicated
by blue shading). Resulting escape behavior was
quantified with a video analysis tool ; mean
speeds and SEMs are shown. Animals cultivated
without the ChR2 cofactor all-trans-retinal (ATR)
showed no responses.
(C) Fractions of animals reacting to increasing
light intensities are shown for different develop-
mental stages (larval stages L2–L4 and adult;
n > 30 each).
depended on developmental stage,
(increasing until adulthood) and branch-
ing (Figure 1C). Of note, PVD neurons
are born in the L2 larval stage, 1?longi-
tudinal branches extend during L3, and
complete branching is achieved by the
end of L4 [9, 19].
In addition to PVD, the F49H12.4 pro-
moter also expresses in a head neuron
(identified as AQR) and a tail neuron
(Figure 1A). To exclude contributions of these cells to the
observed behavior, we illuminated predefined body segments
of freely moving animals that were simultaneously tracked
(Figures 2A–2C; Movie S2). In a recent report, ablation of
AQR reduced responses to anterior harsh touch, and concom-
itantphotoactivation ofAQR, SDQR,andBDUneurons evoked
reversals . In contrast, we observed no escape behavior
when selectively photoactivating AQR (indicating that BDU
and SDQR are responsible for photoevoked behaviors re-
ported by Li et al. ), whereas illuminating the region contain-
ing the PVD soma robustly evoked acceleration. We observed
a result of concomitant illumination of a small posterior area of
PVD. Accelerations were also evoked by illuminating different
small areas of PVD’s anterior dendrites (Figure S2). Thus, the
forward escape we observed is specifically evoked by PVD
Less Habituation to Harsh than Gentle Touch or
Gentle touch is subject to substantial habituation in C. elegans
. However, as nociceptors detect potential threats and
evoke withdrawal to avoid tissue damage, we wondered
whether PVD showed any pronounced habituation to re-
peating or continuing noxious insults. To test this, we used a
‘‘slow’’ ChR2 variant (C128S mutant with much slower off-
kinetics) [21, 22]. A 0.2 s light pulse resulted in continuous
forward locomotion for minutes (Movie S3). For repeated
stimuli, we again used ChR2(H134R), applied 20 light
to each stimulus (Figure 3A). Compared to withdrawal
responses of animals expressing ChR2 in TRNs via Pmec-4::
ChR2(H134R), PVD-evoked escape responses habituated
Current Biology Vol 22 No 9
less. This was confirmed by mechanical touch assays (Fig-
ure 3B), showing that nociceptive harsh touch is less suscep-
tible to habituation than gentle touch.
Optogenetic Dissection of the Harsh Touch Nociceptive
We mainly observed forward locomotion when PVD was
photostimulated, whereas midbody harsh touch reportedly
causes reversals , potentially through recruitment of addi-
tional mechanosensory neurons. We therefore asked whether
reversals may be triggered by concomitant photoactivation of
such candidate cells. When PVD somata were photostimu-
lated with a fixed light intensity, coactivation of anterior
TRNs (ALM, AVM) with increasing light intensities gradually
shifted escape from forward to reverse (Figures 4A and S3A–
S3I). Moreover, reversals prevailed when PVD soma stimula-
tion was reduced, whereas TRN stimulation was kept maximal
(i.e., mimicking a more anteriorly presented harsh touch).
Thus, relative stimulus strengths that different mechanosen-
sory neurons receive determine the probability of forward or
backward escape responses, probably via integration by
synapses equally onto forward and backward command
neurons PVC and AVA (28 and 27 synapses, respectively)
 (Figures 4B and S1). The exact nature of these synapses
is unknown, i.e., whether they are all excitatory or if some are
inhibitory. Mechanical harsh touch was shown to induce
Ca2+increases in PVC, which were higher than those evoked
by gentle touch . First, we directly examined whether PVD
activates PVC. We phototriggered PVD and measured Ca2+
responses inPVCbyusing anovel,red fluorescent,genetically
encoded Ca2+sensor, RCaMP (J.A. and L.L.L., unpublished
data), which requires yellow-light excitation and thus can be
used without unwantedcoactivation
Figure 2. Selective Illumination of PVD Cell Body and of
Other Cells Expressing ChR2
Selective illumination of predefined body segments of
freely moving wild-type (N2); zxIs12 animals was used
as indicated in pictograms to restrict light to the cell
bodies of PVD (A), AQR (B), or a tail neuron expressing
ChR2 (C). Mean velocity traces and SEMs are shown.
Figure 3. Both Mechanical and Optical Nociceptive
Harsh Touch Stimuli Elicit Less Habituation than Gentle
(A) Fractions of wild-type (N2); zxIs12 animals (black) and
animals expressing ChR2 in the TRNs (Pmec-4::ChR2,
gray) were assessed for escape responses to 20 succes-
sive light pulses of 0.5 s with 9.5 s interstimulus interval
(n R 25).
(B) Wild-type animals were challenged with harsh or
gentle touch every 10 s, and fractions of animals reacting
were quantified (n = 20).
Regression analysis shows different slopes for the fitted
lines in both optical and mechanical touch experiments
(*p < 0.05).
increases in PVC (Figure 4C; Movie S4).
deg-1(d) mutant background in which PVC
neurons degenerate, thus limiting PVD output
to AVA. Upon photostimulation, animals now
moved backward (Figures 4D and 4E; Movie S5), as confirmed
by mechanical touch assays. Thus, PVD-AVA synapses are
functional and (net) excitatory; however, in sum, PVD-PVC
synapses must be more excitatory than PVD-AVA synapses.
Interestingly, mec-4(d) mutants lacking TRNs responded
with accelerations when we mechanically stimulated the tail
(probably because of PVD activation) and reversed when ante-
additional mechanosensory neurons; possible candidates are
mainly synapse to AVA. Indeed, selective FLP photoactivation
resulted in instantaneous reversals (Figure 4F; Movie S6).
Command interneurons must integrate PVD and FLP harsh
touch signals, so behaviors evoked by the sensory neurons
may be mimicked via photoactivation of command cells. In
animals expressing Pglr-1::ChR2::mCherry (in command
neurons and other cells), selective tail illumination, activating
only PVC, triggered forward movement, whereas head illumi-
nation, activating AVD, AVA, and AVB (and other) neurons,
evoked robust reversals. Illuminating all command neurons
caused mild reversals (Figures S3J–S3L; Movie S7). In sum,
PVDs and FLPs constitute a nociceptive network responding
to harsh mechanical insults and orchestrate forward versus
backward escape reflexes via the command neurons.
An RNAi Screen with Optogenetic Readout Identifies
Genes Required for PVD Function
Optical stimulation, together with genetic tools in C. elegans,
allowed us to unravel factors required for PVD function, acting
downstream of the primary sensory proteins; Figure 5A
describes our strategy. Previously, FLAG-tagged poly(A)
binding protein was expressed specifically in PVD (and OLL)
neurons and used to pull down mRNAs, which were identified
by microarray profiling . We chose candidate genes
TRP & DEG/ENaC Channels Modulate Nociceptor Output
(Table S1) from this data set and knocked them down system-
ically by feeding bacteria containing dsRNA targeting the
respective gene  to animals sensitized for RNAi in neurons
(nre-1; lin-15b; zxIs12) . To test gene knockdown effi-
ciency, we monitored mRNA levels of several targeted genes
by RT-PCR and further confirmed feasibility of RNAi in PVD
by effective GFP knockdown (Figure S4). Adverse effects of
gene knockdown on PVD function were monitored by calcu-
lating the fraction of animals responding to photostimulation
mec-3 or unc-86 (encoding transcription factors) knockdown
animals, PVD neurons lacked all but the primary branches
(Figure S4B) [19, 25]. This loss also impaired phototriggered
escape responses (Figure 5B). Thus, only full branching of
PVD may permit efficient ChR2-mediated behavior, e.g., as
more PVD plasma membrane may accommodate more ChR2.
Alternatively, MEC-3 andUNC-86 mayregulate genesrequired
for PVD signaling. The EFF-1 fusogen is implicated in ordered
branching of PVD and is required for mechanical harsh touch
responses . Yet, eff-1 knockdown did not reduce photo-
evoked PVD-dependent escape behavior, although these
animals had malformed dendrites (Figure S4B). This suggests
of PVD is crucial for its sensory function but not for its ability to
Figure 4. Functional Analysis of the PVD and FLP Harsh Touch Nociceptor Network
(A) PVD was costimulated with anterior TRNs ALM and AVM by selective illumination with variable light intensities as represented by the blue arrows below
the graph (see also Figures S3A–S3I). Fractions of tested animals responding with forward or backward escape are shown.
(B) Simplified neuronal wiring diagram based on Figure S1, displaying synaptic connections from PVD and FLP sensory neurons (triangles) and their inte-
gration into the command neuron (hexagons) and motor neuron circuits (circles) . Numbers of synapses, which could be excitatory, inhibitory, or both,
(C) Ca2+imaging in PVC (via RCaMP) before and after a 1 s photostimulus to PVD (blue bar, bottom). Fluorescent signals (mean DF/F 6 SEM) in PVC are
compared to red fluorescence of ChR2::mCherry expressed from the pF49H12.4 promoter in a tail neuron next to PVC (ROIs indicated in upper panel;
see also Movie S4).
(D) Fractions of animals that moved forward, backward, or did not respond upon photoactivation of PVD (zxIs12) in the wild-type (N2) or in deg-1(d) mutants
(PVC forward command neurons degenerated). Similarly, behaviors evoked by anterior or posterior mechanical harsh touch are indicated for deg-1(d),
mec-4(d) (lacking TRNs), and wild-type (N2) animals; n R 120; ***p < 0.0001, **p < 0.005, n.s. not significant (chi-square tests).
(E) Selective illumination and video tracking was used to record behaviors induced by PVD photoactivation in wild-type (N2); zxIs12 animals or deg-1(d);
(F) Illuminating the head of animals expressing Pegl-46::ChR2::YFP, photoactivating FLP neurons, evokes backward movement.
In (E) and (F), mean velocity 6 SEM are shown.
Current Biology Vol 22 No 9
PVD Uses Chemical Transmission that Requires VGCC
Subunits UNC-2 and CCB-1
Although PVD has no known electrical synapses in the ventral
nerve cord , the innexin mRNA inx-19 was highly enriched
the branches. However, the inx-19(ky634) mutation did not
affect PVD-dependent behaviors and neither did knockdown
of INX-19 or UNC-7 innexins (Figure 5B), despite clear reduc-
tion of their transcripts (Figure S4A). In contrast, PVD-specific
disruption of chemical synaptic transmission by expressing
tetanus toxin light-chain (TeTx) abolished PVD-mediated
escape behavior (Figures 6A and 6B). Importantly, expression
of TeTx in TRNs did not affect PVD-dependent behavior.
Among PVD-enriched transcripts possibly acting in neuro-
transmission were the VGCC a-subunits egl-19 and unc-2
and the b-subunit ccb-1. Knockdown of unc-2 and ccb-1
(but not egl-19) essentially eliminated photoevoked reactions
(Figure 5B). To assess whether these channel subunits act in
nociception cell autonomously in PVD, we expressed sense
and antisense RNAs  of unc-2 or ccb-1 from the
F49H12.4 promoter (Figures S5A–S5C). The resulting PVD-
specific RNAi strains responded significantly less to PVD
photoactivation (Figures 6C and 6D) yet remained responsive
to gentle touch (Figure S5D). Thus, CCB-1 and UNC-2 are
probably essential for PVD chemical transmission; alterna-
tively, they may enhance PVD excitability. Notably, our
approach allows the identification of genes acting in nocicep-
tion within PVD, downstream of primary sensor molecules.
The TRPM Channel GTL-1 and the DEG/ENaC Channel
ASIC-1 Act within PVD to Mediate Signal Amplification and
TRP and DEG/ENaC channels were previously implicated in
nociception in several organisms, either as sensors, signal
facilitators, or enhancing neurotransmission [1–4, 27, 28].
Among transcripts enriched in PVD/OLL were four DEG/
ENaCs (mec-10, degt-1, del-1, and asic-1) and three TRP
channels (trp-2, gon-2, and gtl-1). Light responses were unaf-
fected by mec-10, del-1, degt-1, and gon-2 knockdown or
introduction of the trp-2(gk298) allele (Figure 5B). gtl-1 knock-
down moderately reduced reactions to PVD photoactivation,
and these effects were significant in gtl-1(ok375) mutants.
Likewise, asic-1(ok415) mutants were significantly affected.
We investigated potential cell-autonomous roles of asic-1
Figure 5. Optogenetics-Assisted Functional RNAi Screen of PVD-Enriched Genes
(A)We previously isolated mRNAsfromPVDneurons,withthemRNA-tagging method asreportedinSmithetal..Epitope (FLAG)-tagged poly(A)binding
protein (PAB-1) was expressed in PVD (and OLL) neurons, with the ser-2prom3 promoter. PAB-1-mRNA was immunoprecipitated and compared to mRNA
isolated from all cells by microarray profiling. Candidate genes were knocked down by RNAi and photoevoked PVD-dependent escape behaviors were
(B) For each knockdown line (or genomic mutant), 25–30 young adults were assessed for blue light-evoked escape behavior and fractions of animals
reacting were compared to negative controls (animals fed with the empty L4440 RNAi vector). Each trial was repeated several times and normalized to
the respective negative control done on the same day. Bars are color-coded according to x-fold mRNA enrichment in PVD (and OLL neurons, compared
to all other cells, see Table S1). Error bars represent SEMs; ANOVA with Tukey’s post hoc test; ***p < 0.0001, **p < 0.005, *p < 0.05.
TRP & DEG/ENaC Channels Modulate Nociceptor Output
and gtl-1 in PVD-specific knockdown lines asic-1 RNAi(PVD)
and gtl-1 RNAi(PVD). We used different light intensities and
whole-body illumination, measuring fractions of animals re-
sponding (Figure 7A). Regression analysis of fitted Boltzmann
sigmoidals showed reduced maximal fractions of animals re-
acting for both PVD-specific RNAi lines, and reduced respon-
siveness at all light intensities was observed for gtl-1
RNAi(PVD) animals, reflected by a significantly right-shifted
We also analyzed how escape velocities evolved over time
as a sensitive readout for promptness and amplitude of PVD
output. When selectively illuminating PVD cell bodies and
tracking the animals, loss of ASIC-1 caused a w50% delay
to reach the maximal speed, whereas GTL-1 deficiency did
not affect the immediateness of escape (Figure 7B). Maximal
velocities for asic-1 RNAi(PVD) animals differed from wild-
type only for the highest light intensity used (Figures 7C
and S6). However, gtl-1 RNAi(PVD) animals showed reduced
maximal velocities over a much broader range of intensities.
Consistent with Figure 7A, this also caused an overall right-
shift of stimulus intensity versus fraction responding (Fig-
ures 7D and S6). Importantly, these effects were not due to
variations in ChR2::mCherry expression in PVD in knockdown
or knockout lines (Figure S5E). In sum, ASIC-1 may expand the
dynamic range of PVD signal output, particularly for strong
noxious stimuli where a boost of synaptic transmission could
allow faster escape, whereas GTL-1 may generally amplify
signals within PVD, i.e., at all stimulus intensities.
By using an optogenetic approach, we defined the function
of PVD and FLP harsh touch nociceptors and down-
stream interneurons in a network mediating escape behavior.
Furthermore, we identified factors required for transmitter
Figure 6. Inhibition of Synaptic Signaling and
Cell-Specific Knockdown of unc-2 and ccb-1
Impair PVD Function
(A and B) Animals with impaired synaptic sig-
naling were generated by injecting wild-type
(N2); zxIs12 animals (control) with the coding
sequence of tetanus toxin light chain via the
mec-4 promoter to achieve expression in TRNs
(TeTx(TRNs), 25 ng/ml) or via the PVD-specific
promoter F49H12.4 at two different concentra-
tions (TeTx(PVD), 5 and 25 ng/ml).
(A) Fractions of animals reacting for each strain
upon 0.2 mW/mm2blue light pulses.
(B) A 1 sblue light pulse was applied to the nema-
tode segment with PVD cell bodies and resulting
ities 6 SEM (gray) are shown.
(C and D) Cell-specific knockdown of unc-2 or
ccb-1 was achieved as described  (see
Figure S5A). Shown are (C) fractions of animals
reacting and (D) velocity diagrams.
All error bars are SEMs; paired t test; ***p <
output from PVD and/or enhanced
TRPM channel GTL-1 may amplify
signals within PVD, whereas the DEG/
ENaC ASIC-1 potentiates PVD output
at high stimulus regimes, thus expanding the dynamic range
of PVD and enabling faster nocifensive behaviors.
Optogenetic Deconstruction of a Harsh Touch Nociceptive
The multidendritic FLP and PVD neurons were previously
reported to sense harsh touch. Recently, additional cells
(BDU, SDQR, AQR, ADE, PDE, PHA, and PHB) contributing
to harsh touch sensation were identified by laser ablation,
although it is unclear whether these cells are sensors or partic-
ipate indirectly . When we photoactivated AQR, we
observed no acute behavior. Similarly, photostimulation of
PDE failed to evoke an escape response ( and data not
shown). Here, we focused on escape behavior induced by
FLP and PVD and characterized their integration into the
downstream command interneuron network. Studying the
behavioral output of the mechanosensory modality of PVD or
FLP cells was previously complicated because harsh touch
agitates the whole animal, coevoking contributions from
TRNs or other mechanosensors. Therefore, we selectively
photoactivated PVD or FLP, causing forward or reverse
escape responses, respectively.
Equivalent numbers of synapses link PVD to forward and
backward command neurons, PVC and AVA . Forward
escape depended on PVC, in which Ca2+transients were
evoked by PVD photostimulation. Upon PVC ablation, PVD
photoactivation evoked backward escape via PVD-AVA
synapses. Direction-specific responses were mimicked by
selectivePVCphotoactivation, resultingin forward movement,
whereas illumination of backward command neurons AVA,
AVD, and AVE (and one of the two forward command neurons,
AVB) evoked reversals. When PVC wascoactivated,backward
responses prevailed (although decreased), suggesting that
behavioral ‘‘output’’ depends on the strength of synaptic
inputs (number, synaptic weight) that individual interneurons
Current Biology Vol 22 No 9
receive. Also, the relative input that different sensory neurons
receive (e.g., by touch at different positions, as mimicked
by coactivation of PVD and anterior TRNs by different light
intensities) determines the probability of forward versus back-
ward escape. In sum, PVD, FLP, and other mechanoreceptors
like TRNs signal to command neurons, which integrate this
information, coordinating the prevailing escape response.
Combining mRNA Profiling, RNAi, and Optogenetics
to Identify Nociceptor Genes
We combined quantifying optogenetically triggered behavior
and RNAi screening to identify and characterize genes critical
for PVD-mediated nociceptive behavior. Here, we focused on
ion channels and genes acting in neurotransmission, among
transcripts we found to be enriched in PVD by microarray
profiling . After RNAi (or in knockouts), we identified genes
assays showed cell-autonomous functions for several genes,
yielding information on the nature of PVD signaling defects.
Thus, we revealed genes acting downstream of the primary
PVD mechanosensory cation channels, which were bypassed
VGCC Subunits UNC-2 and CCB-1 Are Crucial for PVD
VGCCs mediate Ca2+influx after membrane depolarization,
triggering neurotransmission and/or causing further depolari-
zation. Two VGCC a-subunits (EGL-19, UNC-2) and one
b-subunit (CCB-1) are enriched in PVD. EGL-19 L-type VGCCs
are critical particularly for muscle function , yet knocking
down egl-19 did not reduce the frequency of animals respond-
ing to PVD stimulation. UNC-2, homologous to mammalian
CACNA1A, functions mainly in neurons . Phototriggered,
PVD-mediated escape behavior was almost completely sup-
strating cell-autonomous roles in PVD excitatory output. PVD
signals mainly, if not exclusively, by chemical transmission,
as indicated by the fact that PVD-specific expression of TeTx
completely abolished escape responses and that knockdown
of innexins expressed in PVD had no effect. CCB-1 and
Figure 7. ASIC-1 and GTL-1 Accentuate and Amplify ChR2-Evoked Signals in PVD
per trial) for each genotype or cell-specific knockdown (n = 8 for wild-type (N2); n = 4 each for asic-1 RNAi(PVD) and gtl-1 RNAi(PVD)). Error bars indicate
SEM. Boltzmann sigmoidals were fitted according to [% reacting worms = %max / (1+exp((I50-I)/Slope))], where I is light intensity, %max is the maximal
fraction of worms reacting, and I50indicates the light intensity at %max/2. Statistical differences for the three free parameters, %max, I50, and slope (indi-
cated in the table) were determined by regression analysis; *p < 0.05, **p < 0.005, n.s. not significant.
(B) The onset of escape behavior for a 1 s illumination of the nematode segment with PVD cell bodies is shown from the time of illumination (indicated as 0 s)
to0.5 safterillumination.Meanvelocitytraces forN2,asic-1 RNAi(PVD), gtl-1RNAi(PVD), asic-1(ok415),andgtl-1(ok375) canbe fitted byBoltzmann sigmoi-
dals until the maximal velocity is reached according to [velocity = max velocity / (1+exp((T50-T)/Slope))].
(C) PVD cell bodies of R20 animals were photoactivated for 1 s by selective illumination for different light intensities; see Figure S6 for all velocity traces.
Mean maximal velocities 6 SEM for all light intensities are plotted for each genotype; t test; *p < 0.05, **p < 0.005.
(D) Fractions of animals responding are displayed for each light intensity, corresponding Boltzmann sigmoidal fits and free parameters are indicated;
*p < 0.05.
TRP & DEG/ENaC Channels Modulate Nociceptor Output
UNC-2 specifically localize to presynaptic terminals ( and
K. Shen, personal communication), consistent with their
apparent function in chemical neurotransmission.
ASIC-1 and GTL-1 Act within PVD for Signal Output and
DEG/ENaC and TRP channels function in mechanosensation,
forming mechanotransduction channels, or by contributing
indirectly to mechanoreceptor potentials . Precise roles of
these channels often remain elusive. The DEG/ENaCs MEC-4
and MEC-10 form heteromeric mechanoelectrical transduc-
tion channels in TRNs . Impairing mec-10 and degt-1 elimi-
nated Ca2+transients in PVD upon harsh touch, and therefore
their gene products were suggested to form the mechano-
transduction channel in PVD . However, mec-10(tm1552)
mutants retained mechanosensitive currents in PVD , and
similar findings were made in TRNs . Possibly, additional
channels contribute to the primary sensation, as shown for
nose touch perception in ASH neurons . Knockdown of
response, emphasizing that they do not determine general
PVD physiology or excitability.
In mice, removal of brain Na+channel 1 (BNC1/ASIC2),
homologous to MEC-4/-10, reduces sensitivity of low-
threshold mechanoreceptors . Disruption of the related
DRASIC/ASIC3 sensitized gentle touch mechanoreceptors
but reduced sensitivity of nociceptors responding to noxious
pinch, heat, or acid . Thus, only modulatory effects were
observed for these vertebrate ASICs, suggesting roles as
signal facilitators. Similarly, PVD-specific knockdown of
ASIC-1 reducedChR2-mediatedescapebehavior anddelayed
animals in reaching maximal velocity, and therefore ASIC-1
may modulate PVD output. Acidification of the synaptic cleft,
resulting from synaptic vesicle fusion, may trigger thisputative
proton-gated Na+channel, increasing presynaptic depolariza-
tion and neurotransmission, as suggested for dopaminergic
neurons in which ASIC-1 facilitated associative learning .
Likewise, vesicular protons were proposed to enhance neuro-
transmission viaASIC tomodulatesynaptic plasticity inhippo-
campal neurons and to affect learning and memory . We
suggest that ASIC-1 might augment PVD depolarization,
thereby expanding its dynamic range and speeding up the
it is often unclear whether they directly transduce mechanical
stimuli or if they function in downstream signaling. In
Drosophila, the Painless TRP channel  was suggested to
amplify mechanically gated currents mediated by the Pick-
pocket DEG/ENaC . Similarly, the TRPN channel NompC
was assumed to act downstream of the TRPVs Nanchung
and Inactive as a potential amplifier [38, 39], while a study on
amplification of sound-evoked vibrations of the fly hearing
organ led to opposite conclusions . C. elegans contains
at least 24 TRP channels in each of the subdivisions identified
in mammals.TRP-4 was shown tobe amechanotransduc-
tion channel; mutations in the pore altered mechanoreceptor
current ion selectivity . TRPA-1 is required for nose-touch
responses  and for cold sensation in PVD . OSM-9/
OCR-2 TRPV channels function in the polymodal neuron ASH
to mediate different aversive stimuli including noxious mecha-
nosensation , yet they do not act as mechanoelectrical
transduction channels in ASH, and thus may function in signal
encoding or transmission . Our results suggest similar
functions for the TRPM channel GTL-1. PVD-specific GTL-1
knockdown rendered fewer animals responsive to PVD photo-
activation and resulted in significantly lower velocity increases
at all but the smallest stimulus amplitudes. As ChR2 bypasses
natural nociceptive stimuli, we suggest that GTL-1 may cell
autonomously enhance PVD depolarization as an amplifier,
potentially by increasing membrane Ca2+permeability.
In conclusion, we analyzed a nociceptive circuit at single-
neuron level and combined mRNA profiling of single neurons
with an RNAi screen by using optogenetics-induced nocifen-
sive behavior as readout for single neuron function. Our
approach may become a general workflow for similar analy-
ses, potentially uncovering valid targets for pharmacological
intervention in human (chronic) pain conditions.
Strains used in this work are described in the Supplemental Experimental
Procedures. C. elegans culture was according to standard procedures.
Transgene zxEx609 was generated by injecting pSH103(pF49H12.4::
ChR2::mCherry) and pF49H12.4::GFP into wild-type (N2) animals and inte-
grated by UV irradiation, yielding zxIs12. The resulting strain (ZX679) was
outcrossed to N2 (63) and further crossed into different genetic
Optogenetics and Video Analysis
Preparation of animals for optogenetic assays is described in the Supple-
mental Experimental Procedures, which also contain more extended
descriptions of optogenetic methods. Whole-field illumination was done
on a Leica MZ16F stereomicroscope equipped with a Sony XCD-SX90
camera (15 fps) and an external light source (filtered 450–490 nm) with
built-in shutter. The speed of individual animals was calculated over two
successive frames by a modified version of the parallel worm tracker 
(http://www.biochem.uni-frankfurt.de/index.php?id=236). Selective illumi-
nation and tracking was described previously . In brief, we equipped
an inverted microscope (Axiovert 35, Zeiss) with an LCD projector (Hitachi
CP-X605) to illuminate individual segments of a freely moving nematode
while tracking its position and quantifying evoked behavior, via custom-
RNAi-Optogenetics Screen and Cell-Specific RNAi
Systemic knockdown was described previously . RNAi efficiency was
assessed by RT-PCR for a number of genes; experimental details are ex-
plained in Supplemental Experimental Procedures. Cell-specific RNAi was
achieved by expressing sense and antisense constructs of genes of
interest, driven by the F49H12.4 promoter (also expressing in AQR and
one tail neuron) as described ; see Figure S5A for PCR fusion strategy.
Confocal and Ca2+Imaging
Adult animals were immobilized on agarose pads in M9 buffer with 30 mM
NaN3. Images in Figure 1 were obtained with a Marianas spinning-disk
confocal (SDC) system (Intelligent Imaging Innovations [3I]) equipped with
an Evolve EMCCD camera (Photometrics). All other pictures were obtained
on an LSM510 (Zeiss). To image Ca2+transients in PVC, we used RCaMP
(J.A. and L.L.L., unpublished data), a red-fluorescent GCaMP-like Ca2+indi-
cator employing a circularly permuted mRuby protein , with a 561 nm
laser (i.e., without concomitant photoactivation of ChR2) on an Andor
SDC system equipped with an iXon897 EMCCD camera. Z-stacks were ob-
tained at1Hz. PVDwas photostimulated for1s withblue light fromafiltered
HBO lamp (450–490 nm).
Supplemental Information includes Supplemental Experimental Proce-
dures, six figures, one table, and seven movies and can be found with this
article online at doi:10.1016/j.cub.2012.02.066.
We thank the Caenorhabditis Genetics Centre (CGC; supported by the NIH,
fulto K.Preckel andB.Rummel forexpert technical assistance, toH. Hutter,
Current Biology Vol 22 No 9
C. Bargmann, and W. Schafer for sharing reagents, to M. Goodman for
communicating results prior to publication, and to M. Goodman, W. Scha-
fer, and M. Zhen for comments on the manuscript. S.J.H. was supported
by the Human Frontiers Science Program Organization (HFSPO) and the
Research Fund Flanders (FWO-Vlaanderen). This work was supported by
grants from the US-Israel Binational Science Foundation Grant 2005036
(M.T. and D.M.M.), NIH R01 NS26115 and R21 NS06882 (D.M.M.), NIH T32
MH64913 and F31 NS49743 (J.D.W.), NIH P30 CA68485, P60 DK20593,
P30 DK58404, HD15052, P30 EY08126, and PO1 HL6744 to Vanderbilt
University, as well as by grants from the DFG (SFB807, FOR1279, Cluster
of Excellence Frankfurt [CEF-MC]; EXC115/1) and the Schram foundation
Received: September 15, 2011
Revised: January 12, 2012
Accepted: February 22, 2012
Published online: April 5, 2012
1. Basbaum, A.I., Bautista, D.M., Scherrer, G., and Julius, D. (2009).
Cellular and molecular mechanisms of pain. Cell 139, 267–284.
2. Kang, L., Gao, J., Schafer, W.R., Xie, Z., and Xu, X.Z. (2010). C. elegans
TRP family protein TRP-4 is a pore-forming subunit of a native mecha-
notransduction channel. Neuron 67, 381–391.
3. Li, W., Kang, L., Piggott, B.J., Feng, Z., and Xu, X.Z. (2011). The neural
circuits and sensory channels mediating harsh touch sensation in
Caenorhabditis elegans. Nat. Commun. 2, 315.
4. O’Hagan, R., Chalfie, M., and Goodman, M.B. (2005). The MEC-4 DEG/
ENaC channel of Caenorhabditis elegans touch receptor neurons trans-
duces mechanical signals. Nat. Neurosci. 8, 43–50.
5. Tobin, D.M., and Bargmann, C.I. (2004). Invertebrate nociception:
behaviors, neurons and molecules. J. Neurobiol. 61, 161–174.
6. Chalfie, M., Sulston,J.E., White, J.G.,Southgate, E.,Thomson,J.N.,and
Brenner, S. (1985). The neural circuit for touch sensitivity in
Caenorhabditis elegans. J. Neurosci. 5, 956–964.
7. Chatzigeorgiou, M., Yoo, S., Watson, J.D., Lee, W.H., Spencer, W.C.,
Kindt, K.S., Hwang, S.W., Miller, D.M., 3rd, Treinin, M., Driscoll, M.,
and Schafer, W.R. (2010). Specific roles for DEG/ENaC and TRP chan-
nels in touch and thermosensation in C. elegans nociceptors. Nat.
Neurosci. 13, 861–868.
8. Chatzigeorgiou, M., and Schafer, W.R. (2011). Lateral facilitation
between primary mechanosensory neurons controls nose touch
perception in C. elegans. Neuron 70, 299–309.
9. Oren-Suissa, M., Hall, D.H., Treinin, M., Shemer, G., and Podbilewicz, B.
(2010). The fusogen EFF-1 controls sculpting of mechanosensory
dendrites. Science 328, 1285–1288.
10. Way, J.C., and Chalfie, M. (1989). The mec-3 gene of Caenorhabditis
elegans requires its own product for maintained expression and is ex-
pressed in three neuronal cell types. Genes Dev. 3 (12A), 1823–1833.
11. Arnado ´ttir, J., O’Hagan, R., Chen, Y., Goodman, M.B., and Chalfie, M.
(2011). The DEG/ENaC protein MEC-10 regulates the transduction
channel complex in Caenorhabditis elegans touch receptor neurons.
J. Neurosci. 31, 12695–12704.
12. White, J.G., Southgate, E., Thomson, J.N., and Brenner, S. (1986).
The structure of the nervous system of the nematode Caenorhabditis
elegans. Philos. Trans. R. Soc. Lond. B Biol. Sci. 314, 1–340.
13. Nagel, G., Brauner, M., Liewald, J.F., Adeishvili, N., Bamberg, E., and
Gottschalk, A. (2005). Light activation of channelrhodopsin-2 in excit-
able cells of Caenorhabditis elegans triggers rapid behavioral
responses. Curr. Biol. 15, 2279–2284.
14. Stirman, J.N., Crane, M.M., Husson, S.J., Wabnig, S., Schultheis, C.,
Gottschalk, A., and Lu, H. (2011). Real-time multimodal optical control
of neurons and muscles in freely behaving Caenorhabditis elegans.
Nat. Methods 8, 153–158.
15. Ezcurra, M., Tanizawa, Y., Swoboda, P., and Schafer, W.R. (2011). Food
sensitizes C. elegans avoidance behaviours through acute dopamine
signalling. EMBO J. 30, 1110–1122.
16. Hwang, R.Y., Zhong, L., Xu, Y., Johnson, T., Zhang, F., Deisseroth, K.,
and Tracey, W.D. (2007). Nociceptive neurons protect Drosophila larvae
from parasitoid wasps. Curr. Biol. 17, 2105–2116.
17. Watson, J.D., Wang, S., Von Stetina, S.E., Spencer, W.C., Levy, S.,
Dexheimer, P.J., Kurn, N., Heath, J.D., and Miller, D.M., 3rd. (2008).
Complementary RNA amplification methods enhance microarray
BMC Genomics 9, 84.
18. Edwards, S.L., Charlie, N.K., Milfort, M.C., Brown, B.S., Gravlin, C.N.,
Knecht, J.E., and Miller,K.G. (2008). A novel molecular solution for ultra-
violet light detection in Caenorhabditis elegans. PLoS Biol. 6, e198.
19. Smith, C.J., Watson, J.D., Spencer, W.C., O’Brien, T., Cha, B., Albeg, A.,
Treinin, M., and Miller, D.M., 3rd. (2010). Time-lapse imaging and cell-
specific expression profiling reveal dynamic branching and molecular
determinants of a multi-dendritic nociceptor in C. elegans. Dev. Biol.
20. Rose, J.K., and Rankin, C.H. (2001). Analyses of habituation in
Caenorhabditis elegans. Learn. Mem. 8, 63–69.
21. Berndt, A., Yizhar, O., Gunaydin, L.A., Hegemann, P., and Deisseroth, K.
(2009). Bi-stable neural state switches. Nat. Neurosci. 12, 229–234.
22. Schultheis, C., Liewald, J.F., Bamberg, E., Nagel, G., and Gottschalk, A.
(2011). Optogenetic long-term manipulation of behavior and animal
development. PLoS ONE 6, e18766.
23. Kamath, R.S., Fraser, A.G., Dong, Y., Poulin, G., Durbin, R., Gotta, M.,
Kanapin, A., Le Bot, N., Moreno, S., Sohrmann, M., et al. (2003).
Systematic functional analysis of the Caenorhabditis elegans genome
using RNAi. Nature 421, 231–237.
24. Schmitz, C., Kinge, P., and Hutter, H. (2007). Axon guidance genes
identified in a large-scale RNAi screen using the RNAi-hypersensitive
Caenorhabditis elegans strain nre-1(hd20) lin-15b(hd126). Proc. Natl.
Acad. Sci. USA 104, 834–839.
25. Tsalik, E.L., Niacaris, T., Wenick, A.S., Pau, K., Avery, L., and Hobert, O.
(2003). LIM homeobox gene-dependent expression of biogenic amine
receptors in restricted regions of the C. elegans nervous system. Dev.
Biol. 263, 81–102.
26. Esposito, G.,DiSchiavi, E., Bergamasco, C., andBazzicalupo, P.(2007).
Efficient and cell specific knock-down of gene function in targeted
C. elegans neurons. Gene 395, 170–176.
27. Go ¨pfert, M.C., Albert, J.T., Nadrowski, B., and Kamikouchi, A. (2006).
Specification of auditory sensitivity by Drosophila TRP channels. Nat.
Neurosci. 9, 999–1000.
28. Voglis, G., and Tavernarakis, N. (2008). A synaptic DEG/ENaC ion
channel mediates learning in C. elegans by facilitating dopamine signal-
ling. EMBO J. 27, 3288–3299.
29. Lee, R.Y., Lobel, L., Hengartner, M., Horvitz, H.R., and Avery, L. (1997).
Mutations in the alpha1 subunit of an L-type voltage-activated Ca2+
channel cause myotonia in Caenorhabditis elegans. EMBO J. 16,
30. Schafer, W.R., and Kenyon, C.J. (1995). A calcium-channel homologue
required for adaptation to dopamine and serotonin in Caenorhabditis
elegans. Nature 375, 73–78.
31. Saheki, Y., and Bargmann, C.I. (2009). Presynaptic CaV2 calcium
channel traffic requires CALF-1 and the alpha(2)delta subunit UNC-36.
Nat. Neurosci. 12, 1257–1265.
32. Geffeney, S.L., Cueva, J.G., Glauser, D.A., Doll, J.C., Lee, T.H.J.-C.,
Montoya, M., Karania, S., Garakani, A.M., Pruitt, B.L., and Goodman,
trical transduction channels in a C. elegans nociceptor. Neuron 71, 1–13.
33. Price, M.P., Lewin, G.R., McIlwrath, S.L., Cheng, C., Xie, J., Heppenstall,
P.A., Stucky, C.L., Mannsfeldt, A.G., Brennan, T.J., Drummond, H.A.,
et al. (2000). The mammalian sodium channel BNC1 is required for
normal touch sensation. Nature 407, 1007–1011.
34. Price, M.P.,McIlwrath, S.L., Xie, J.,Cheng, C., Qiao, J.,Tarr, D.E., Sluka,
K.A., Brennan, T.J., Lewin, G.R., and Welsh, M.J. (2001). The DRASIC
cation channel contributes to the detection of cutaneous touch and
acid stimuli in mice. Neuron 32, 1071–1083.
35. Wemmie, J.A., Chen, J., Askwith, C.C., Hruska-Hageman, A.M., Price,
M.P., Nolan, B.C., Yoder, P.G., Lamani, E., Hoshi, T., Freeman, J.H.,
Jr., and Welsh, M.J. (2002). The acid-activated ion channel ASIC
contributes to synaptic plasticity, learning, and memory. Neuron 34,
36. Tracey, W.D., Jr., Wilson, R.I., Laurent, G., and Benzer, S. (2003).
37. Zhong, L., Hwang, R.Y., and Tracey, W.D. (2010). Pickpocket is a DEG/
ENaC protein required for mechanical nociception in Drosophila larvae.
Curr. Biol. 20, 429–434.
38. Gong, Z., Son, W., Chung, Y.D., Kim, J., Shin, D.W., McClung, C.A., Lee,
Y., Lee, H.W., Chang, D.J., Kaang, B.K., et al. (2004). Two interdepen-
dent TRPV channel subunits, inactive and Nanchung, mediate hearing
in Drosophila. J. Neurosci. 24, 9059–9066.
TRP & DEG/ENaC Channels Modulate Nociceptor Output
39. Kim, J., Chung, Y.D., Park, D.Y., Choi, S., Shin, D.W., Soh, H., Lee, H.W.,
Son, W., Yim, J., Park, C.S., et al. (2003). A TRPV family ion channel
required for hearing in Drosophila. Nature 424, 81–84.
40. Goodman, M.B., and Schwarz, E.M. (2003). Transducing touch in
Caenorhabditis elegans. Annu. Rev. Physiol. 65, 429–452.
41. Kindt, K.S., Viswanath, V., Macpherson, L., Quast, K., Hu, H.,
Patapoutian, A., and Schafer, W.R. (2007). Caenorhabditis elegans
TRPA-1 functions in mechanosensation. Nat. Neurosci. 10, 568–577.
42. Colbert, H.A., Smith, T.L., and Bargmann, C.I. (1997). OSM-9, a novel
protein with structural similarity to channels, is required for olfaction,
elegans. J. Neurosci. 17, 8259–8269.
43. Ramot, D., Johnson, B.E., Berry, T.L., Jr., Carnell, L., and Goodman,
M.B. (2008). The Parallel Worm Tracker: a platform for measuring
average speed and drug-induced paralysis in nematodes. PLoS ONE
44. Kredel, S., Oswald, F., Nienhaus, K., Deuschle, K., Ro ¨cker, C., Wolff, M.,
Heilker, R., Nienhaus, G.U.,and Wiedenmann, J. (2009). mRuby,abright
monomeric redfluorescentprotein forlabeling ofsubcellularstructures.
PLoS ONE 4, e4391.
Current Biology Vol 22 No 9