TrpA1 Regulates Thermal Nociception in Drosophila

Pain is a significant medical concern and represents a major unmet clinical need. The ability to perceive and react to tissue-damaging stimuli is essential in order to maintain bodily integrity in the face of environmental danger. To prevent damage the systems that detect noxious stimuli are therefore under strict evolutionary pressure. We developed a high-throughput behavioral method to identify genes contributing to thermal nociception in the fruit fly and have reported a large-scale screen that identified the Ca2+ channel straightjacket (stj) as a conserved regulator of thermal nociception. Here we present the minimal anatomical and neuronal requirements for Drosophila to avoid noxious heat in our novel behavioral paradigm. Bioinformatics analysis of our whole genome data set revealed 23 genes implicated in Ca2+ signaling that are required for noxious heat avoidance. One of these genes, the conserved thermoreceptor TrpA1, was confirmed as a bona fide “pain” gene in both adult and larval fly nociception paradigms. The nociceptive function of TrpA1 required expression within the Drosophila nervous system, specifically within nociceptive multi-dendritic (MD) sensory neurons. Therefore, our analysis identifies the channel TRPA1 as a conserved regulator of nociception.


Introduction
Acute and chronic pain will affect most people at some stage in their lives [1]. Chronic pain in particular represents an unmet clinical need [2]. Nociception, the neuronal sensory and processing apparatus that relays the perception of acute pain, allows an organism to avoid potential tissue damage and death, and many genes regulating this process are conserved across phyla [3]. Transient Receptor Potential (TRP) channels are a family of sensory ion channels that were first identified in Drosophila melanogaster [4,5], and have subsequently been identified as critical mediators of nociception in mammals [6]. The TRP family channel painless was identified in Drosophila using a larval heat probe assay [7]. While painless has no mammalian orthologue [8,9], it is possible that other components of the Drosophila nociception apparatus are indeed conserved from flies to mammals.
To interrogate Drosophila for conserved genes that regulate nociception, we developed a high-throughput screening procedure [3]. This behavioral system utilizes the robust ability of adult fruit flies to rapidly avoid noxious heat. This system has led to the identification of hundreds of candidate fly ''pain'' genes, for example straightjacket (stj), as regulators of nociception behavior in Drosophila [3]. Here we show this innate avoidance behavior is independent of other sensory modalities known to promote avoidance responses, such as vision, olfaction, CO 2 perception, hearing, and taste and requires intact antennae and proboscis for a full response. Importantly, painless-expressing neurons, but not the mushroom body which is required for sub noxious thermopreference [10], are a necessity for this behavior. We further provide genetic evidence that one of the candidate pain genes, TrpA1, is a bona fide mediator of thermal nociception in the fly. Tissue-specific RNAi knockdown of TrpA1 revealed that TrpA1 functions in nociceptive multi-dendritic (MD) sensory neurons. Thus, TRPA1 regulates the Drosophila behavioral response to a noxious thermal insult. Combined with TRPA1's role in chemical nociception, our results identify TRPA1 as an evolutionary conserved regulator of polymodal nociception.

Set-up of a high-throughput system to screen for nociception in Drosophila
We recently developed a high-throughput assay to perform an in vivo genome-wide pain screen in Drosophila [3]. Here we report the detailed set-up and anatomical/neuronal requirements for this novel behavioral paradigm, data we believe are essential for the field and future use of this system. In preliminary pilot studies to address nociceptive responses, we found the Drosophila response to noxious heat exposure more reliable and robust compared to mechanical pain paradigms (not shown). Furthermore, while the commonly used Drosophila larval nociception paradigm has proven suitable for identifying genes required for nociceptive behavior [7], it is labor intensive and requires evaluations of larval responses not compatible with large-scale applications.
To develop a high-throughput screening system in adult Drosophila, we first defined acute noxious heat thresholds. By subjecting flies to increasing gradients of noxious heat exposure, we found that flies rapidly became incapacitated following exposure to temperatures above 40uC ( Figure 1A) and continued exposure to such temperature (.2 minutes) was lethal (not shown). Temperatures below 39uC did not incapacitate flies within the 10 minute time course. This thermal tolerance profile was very similar to that reported by others, indicating a common thermal tolerance threshold across Drosophila strains and experimental paradigms [11]. Since nociception is the sense an animal employs to detect and avoid potential harm, and because exposure to temperatures about 40uC were acutely harmful to Drosophila, we exploited this rapid incapacitation as a means of large-scale screening for nociception behavior. We therefore developed an experimental test chamber where the temperature of the base surface of the chamber could be controlled and rapidly increased when required, giving flies a choice between a hot and potentially lethal surface and a surface that remained close to room temperature. Using this paradigm, we found that wild-type Canton-S flies rapidly avoid all noxious temperatures tested. Flies mutant for the classical painless (pain 1 ) gene could avoid surfaces heated up to 42uC, but failed avoiding the surface if the temperature is $42uC ( Figure 1B). These differences between wild-type and pain 1 flies were greatest at 46uC. To assess a potential influence of the genetic background, we assayed three different laboratory Drosophila melanogaster strains. Canton-S, Canton-S w1118 , and Oregon-R strains all rapidly and reproducibly avoided the heated surface during the course of the experiment ( Figure 1C). Our final experimental apparatus involves an inverted petri dish with ,20 flies, sealed with tape, and floated on a 46uC water bath ( Figure 1D). The chamber is 35 mm wide with a 10 mm distance between the hot and warm surfaces. The bottom heated surface reaches 46uC within 15 seconds of the experiment, while the internal top and middle surfaces reach 31uC and 33uC by the end of the 4 minute test period. The maximum internal air temperature recorded during the experiment was 31uC. Using this system we can generate a % avoidance value for each genotype tested ( Figure 1D). Thus, adult Drosophila exhibit a robust and highly reproducible innate avoidance response to noxious heat, and in the fly this response is dependent on the painless gene.

Mapping anatomical structures critical for high temperature nociception in Drosophila
Little is known about the anatomy regulating the behavioral response to an acute noxious insult in adult Drosophila [12]. We therefore assayed whether noxious heat avoidance requires sensory and higher order neurons implicated in other avoidance behaviors. Blocking synaptic transmission in neurons by driving UAS-Shi ts1 (Shibire ts1 ; a temperature sensitive dynamin mutant [13]) in subsets of neurons controlling olfaction (OR83b-Gal4), hearing and hygrosensation (nan-Gal4), and vision (GMR-GAL4) did not affect nociceptive behaviors, indicating that avoidance of noxious heat is independent of these sensory modalities ( Figure 2A). Furthermore, thermal nociception did not require neurons previously implicated in other aversive behaviors such as the Gr21a-expressing CO2sensitive olfactory neurons [14], or neurons expressing neuropeptide F (NPF-Gal4.UAS-Shi ts1 ) or its receptor (NPFR-Gal4.UAS-Shi ts1 ) which have been implicated in bitter taste avoidance [15]. Importantly, thermal nociception was dependent on synaptic transmission in painless-expressing neurons ( Figure 2A). Taken together, these findings indicate that thermal nociception in adult Drosophila requires painless expressing neurons but does not rely on sensors or cells previously implicated in vision, olfaction, hearing, CO2 sensing, hygrosensation, or bitter avoidance.
To map the anatomy of thermal nociception in adult fly, we surgically removed the Drosophila wings, proboscis and antennae to determine if these structures are involved in the response to avoid noxious heat. We found flies lacking wings respond normally to noxious heat, indicating that wings are not required for thermal nociception ( Figure 2B). Previous studies have implicated the third-antennal segment as a component of the sensory apparatus required for rapid avoidance of elevated temperatures between 25uC and ,33uC [16,17]. We observed a modest requirement for antenna in avoidance of noxious heat, which appeared to cooperate with the proboscis in the observed avoidance response ( Figure 2C). Thus, the antenna and proboscis are candidate organs for sensing noxious heat.
In the adult fly, it has been reported that painless is expressed in the wing, proboscis, leg, and in central brain neurons that include the mushroom bodies ( [18] and Figure 2D). Mushroom bodies have been implicated in thermal preference in the more long-term (15 minutes) behavioral response to sub-noxious temperatures [10]. To silence mushroom body neurons, we used the mushroom body drivers OK107-Gal4 [19,20] and MB247-Gal4, two of the same Gal4 driver implicated in sub-noxious thermal preference [10], both of which express broadly throughout the mushroom bodies [21]. In addition we used the mushroom body associated DPM driver c316-Gal4 [22]. Using the Gal4/UAS system to express Shi ts1 , we again found that synaptic silencing of pain-Gal4 neurons was sufficient to abolish the heat avoidance response. However, silencing of the mushroom body itself, or mushroom body associated neurons, had no effect on the avoidance response to noxious heat ( Figure 2E). Flies with UAS-Shi ts1 , or pain-Gal4 alone showed wild-type avoidance, indicating the avoidance defects observed in pain-Gal4;UAS-shi ts1 are due to the silencing of painexpressing neurons. Importantly, silencing of painless-expressing neurons did not result in general coordination defects, as assessed by negative geotactic response ( Figure 2F). Thus, pain-Gal4 expressing cells outside of the mushroom bodies are required for adult Drosophila to sense noxious heat.
Genes implicated in calcium signaling are overrepresented among candidate heat nociception genes We have previously reported 580 candidate genes involved in nociception in Drosophila [3]. Based on these primary hits, we performed hypergeometric enrichment analysis using KEGG pathways and Broad Institute C2 gene sets to identify groups of genes that were over-represented in our genome wide screen for nociception. One prominent gene category identified by this analysis was calcium signaling ( Figure 3A, Table S1 for full data set). For instance, Calphotin (Cpn; CG4795) is a calcium binding molecule implicated in rhabdomere development. Moreover, we hit the calcium channel subunit straightjacket as we have reported in detail elsewhere [3]. We also found the fly ortholog of DREAM (Calsenilin, KChip3), a calcium-regulated central pain gene in mice [23], the calcium regulated cell adhesion molecule CG7100 (CadN), a calcium-dependent EF hand protein serine/threonine phosphatase (CG6571, rdgC), CAMKII, a gene that has also been linked to the modulation of TRPV1 channel function [24], as well as the CAMKII activator Caki (CG6703, Camguk), a member of the MAGUK family of proteins that contains a CaMKII-like domain and participates in regulation of calcium channel function in other species [25]. Finally, our approach identified dTrpA1 (TrpA1), the Drosophila ortholog of the chemical and cold sensing human TRPA1, a Ca 2+ -permeable non-selective cation channel implicated in acute chemical pain and cold hypersensitivity in rodents [26,27] and infrared sensation in snakes [28].

TRPA1 is required for thermal nociception in both larvae and adult Drosophila
TrpA1 is a member of the TRPA subfamily of TRP channels. Drosophila TRPA1 is the fly ortholog of human TRPA1 and the two proteins share ,33% overall sequence identity, whereas other insect TRPA family members such as painless and pyrexia belong to a distinct subclass of TRPAs lost during vertebrate evolution [8]. Drosophila TRPA1 has been found to act as a warmth-activated ion channel required for thermotaxis at non-noxious temperatures and to act as a receptor mediating avoidance of reactive electrophilic chemicals [8,29,30]. To definitively demonstrate that fly TrpA1 is essential for noxious heat avoidance, we tested whether classical mutants for TrpA1 exhibit impaired avoidance of noxious heat. Consistent with the RNAi knockdown results, animals homozygous for a loss-of-function mutation in TrpA1 (TrpA1 ins ) [30] failed to avoid noxious temperature to a level similar to painless mutants ( Figure 3B). To determine if the observed TrpA1 adult pain phenotype was the result of increased temperature-induced paralysis at 46uC, we exposed TrpA1 and control flies to a chamber set to 46uC and recorded the kinetics of temperatureinduced paralysis. In support of an adult pain phenotype, we observed no difference in temperature-induced paralysis between these lines (36.6+/21.4 seconds control, 33.2+/21.3 seconds TrpA1 flies, n = 12, not significant by t test). The avoidance defect was rescued by reintroduction of a TrpA1 minigene into the TrpA1 mutant background, confirming that the observed defect in thermal nociception was due to the disruption of TrpA1 ( Figure 3C). In contrast, we did not observe a noxious heat avoidance phenotype in a mutant for another Drosophila TRPA, pyrexia (pyx 3 ) ( Figure 3B), previously implicated in high temperature (40uC) thermal tolerance [31]. Importantly, painless and TRPA1 Figure 2. The Drosophila antenna, proboscis, and painless expressing neurons are required for avoidance of noxious heat. (A) Synaptic output from painless expressing neurons (painless-Gal4.UAS-Shi ts1 ), but not other sensory modalities tested, is required to avoid noxious heat. Sensory nerves mediating olfaction (OR83b-Gal4.UAS-Shi ts1 ), hearing and hygrosensation (Nan-Gal4.UAS-Shi ts1 ), the CO 2 sensing apparatus (GR21a-Gal4.UAS-Shi ts1 ), bitter avoidance (NPFR-Gal4.UAS-Shi ts1 ), and vision (GMR-Gal4.UAS-Shi ts1 ) are not required for thermal nociception. Control line is w 1118 6UAS-Shi ts1 . (B) Wings are not required for noxious thermal avoidance in Drosophila. Wings were dissected off and flies were tested for the ability to avoid noxious temperature. (C) Antenna and proboscis cooperate to promote avoidance of noxious temperature. One or both antenna and the proboscis were dissected from wild type flies and then tested for avoidance of noxious temperature. (D) pain-Gal4 driving UAS-CD8:GFP labels painless expressing neurons in the mushroom body and other regions of the adult fly brain. MB247-Gal4.UAS-CD8:GFP labels the mushroom body. (E) Synaptic output from painless expressing nerves (pain-Gal4.UAS-Shi ts1 ), but not neurons from the mushroom body (MB247-Gal4.UAS-Shi ts1 , OK107-Gal4.UAS-Shi ts1 ) or mushroom body associated DPM neurons (c316-Gal4.UAS-Shi ts1 ), is required for thermal nociception in adult Drosophila. Cs (Canton S). (F) Synaptic silencing of painless expressing nerves (pain-Gal4.UAS-Shi ts1 ) does not affect basic motor coordination as assayed by a negative geotactic assay. All experiments involving UAS-shibire TS1 were pre-incubated at 30uC for 1 hour to induce shibire-mediated synaptic silencing. Data are presented as mean values +/2 SEM. ,20 flies tested per group, in replicates of at least four cohorts. In all experiments adult flies were challenged with 46uC as outlined in Figure 1D. * P,0.05, ** P,0.01 (Student's t-test). doi:10.1371/journal.pone.0024343.g002 mutant flies exhibit normal coordination as assessed by a negative geotactic response ( Figure 3D). Thus, both painless and TrpA1 are required for avoidance of noxious heat in adult Drosophila.
As TrpA1 participates in thermal nociception in adult Drosophila, we also assayed whether TrpA1 is involved in thermal pain behavior in larvae (Figure 4). Larvae were gently touched with a 46uC probe and avoidance time was measured. While control larvae showed a rapid response to noxious heat, the TrpA1 mutants showed a significantly diminished capacity for thermal nociception (Figure 4). In contrast to adult flies, pyx 3 mutant larvae also showed an impairment in this larval nociception assay. Thus, we report a key role for Drosophila TrpA1 in larval and adult thermal nociception, i.e. TrpA1 is a bona fide pain gene in Drosophila.
Blocking neurotransmission from multi-dendritic (MD) sensory neurons reduces thermal nociception responses in Drosophila larvae [7]. To localize the site of TrpA1 function in thermal nociception we therefore employed UAS-TrpA1-IR flies. Driving TrpA1-RNAi in all neurons using elav-Gal4 (elav-Gal4.TrpA1-UAS-IR) resulted in significant impairment of the larval thermal nociception response ( Figure 5A). Driving TrpA1 RNAi in MD-sensory neurons alone (MD-Gal4.TrpA1-UAS-IR) also impaired the larval thermal nociception response ( Figure 5B). These results indicate that TRPA1 is, at least in part, acting in multi-dendritic sensory neurons.

Discussion
Our novel high-throughput system for assessing nociception behavior in adult Drosophila has allowed us to screen the entire Drosophila genome to examine the neural basis of thermal nociceptive behavior. Here we describe neuronal requirements and anatomical structures involved in this innate behavioral paradigm of thermal nociception. We also provide first experimental proof that TrpA1 functions in sensory neurons as a novel component of the Drosophila thermal nociception apparatus.
TRPA1 is recognized as a key component of the nociception apparatus in mammals. Not only do mice mutant for TRPA1 show defects in nociception, a mutation in TRPA1 underlies a human episodic pain syndrome [32]. At the molecular level, the role of the mammalian TRPA1 as a receptor for electrophilic chemicals and other irritants is well established [27], and the mechanism of how TRPA1 functions in chemical nociception is highly conserved between humans and flies [8]. Mammalian TRPA1 is also implicated in cold nociception, at least under pathological conditions [26], suggesting it contributes to multiple nociceptive sensory modalities. In Drosophila, TrpA1 has been shown to act as an internal thermosensor regulating temperature preference at non-noxious temperatures [30] and as a chemoreceptor for noxious electrophilic irritants [8]. That TRPA1 participates in both subnoxious thermal preference [30] and avoidance of noxious heat is an interesting observation. This could reflect distinctions between the mechanisms that control responses to steep versus shallow gradients, with thermal preference behavior involving long-term (.15 minute) exposure to shallow (,0.5uC/mm) temperature gradients [30], while the current assay assesses behavioral short-term (,4 minutes) responses to much steeper gradients (,4.5uC/mm by assay endpoint). These potential distinctions are currently being explored in our laboratories.
Our study expands the role of Drosophila TrpA1 signaling to responses to noxious heat. Combined with recent work demonstrating its role in chemical nociception [8], TrpA1 mediates both chemical and thermal nociception in Drosophila. Thus, while the temperature-responsiveness of TRPA1 has undergone significant diversification within the animal lineage, from being heatactivated in flies [33] and snakes [28] to potentially coldresponsiveness in mammals [26,34], a role for TRPA1 in polymodal nociception has been retained from flies to mammals.
In addition to TrpA1, painless and stj (and to a lesser extend pyrexia in larvae) are required for the pain response in Drosophila. Thus it appears that, similar to mammals, multiple cation channels are involved in pain-responses in flies. While painless and pyrexia encode insect-specific TRPA-channels, TrpA1 and stj are conserved in humans. Both TRPA1 and stj (CACNA2D3 in humans) have been implicated in human nociception suggesting core genetic regulators of nociception are strongly conserved from flies to humans [3].
The ability to perform high-throughput screening for mediators of nociception has until now been limited to in vitro models and not intact behaving animals. Our in vivo system expands the toolbox available for pain researchers. Coordinated use of this fly system could accelerate the identification of new compounds to be short listed for validation in mammalian models of pain. Using this novel paradigm, we have now described hundreds of new candidate ''pain'' genes, opening the field up to multiple new candidate analgesic targets. Among these genes, our RNAi-screen identified multiple genes that are predicted to play a role in Ca 2+ signaling such as TRPA1. Future analysis of these genes should provide valuable insight into the neural basis of nociceptive behavior and the role of calcium signaling in pain.

Behavior experiments
For adult avoidance of noxious heat, ,20 four day old flies were placed into a behavioral chamber (35 mm610 mm Petri dish; Nunclon) and the chamber was sealed with scotch tape. Flies were rested for at least 30 minutes in the dark. The chambers were then floated on a 46uC water bath for 4 minutes. The bottom of the chamber was heated to 46uC over 15 seconds by floating on a water bath while the sub-noxious zone was measured to be 31uC (inner top of chamber) and 33uC (middle edge of the chamber) at the end of the 4 minute experiment. Chamber temperature was monitored using an electronic thermometer (Testo 925, Germany) coated in heat sink gel (RS components, UK). Chambers were then removed from the water and immobilized flies were counted as ''incapacitated''. In addition, total fly number was recorded. Percentage Avoidance was calculated by determining the percentage of flies that avoid the noxious temperature compared to the total number of flies in the chamber. All tests were performed under low red light. Of note, this assay is an absolute measurement where flies that avoid the bottom heated surface were considered capable of noxious thermal avoidance, independent of how far they move away from the 46uC surface, though the vast majority of flies avoiding the hot surface were found on the top of the chamber. For experiments involving UAS-Shibire ts1 , flies were transferred to the experimental chambers followed by a temperature shift to 30uC for 60 minutes. Larval pain assays were performed as described [7]. For assessing noxious temperature-induced paralysis, wild type flies were placed in 5 ml polystyrene round bottom tubes (BD Falcon, Germany) and exposed to temperatures ranging from 37-46uC with 1u increments) or only 46uC (for control vs TrpA1 flies) and the temperature at which 100% of flies were paralyzed was recorded. General coordination was assessed by tapping the test chamber on the bench and observing activity as flies move away from the site of impact [35]. This response was quantified using a geotactic repulsion assay where flies were knocked to the bottom of a 15 ml polystyrene tube and the geotactic response (number of flies climbing up the tube / total number of flies) was recorded.

Confocal microscopy
Drosophila brains were dissected in PBS, fixed in 4% paraformaldehyde in PBS for 30 min at room temperature (RT), washed three times for 10 min in PBS containing 0.3%Triton X-100, blocked for 1 hr at RT in PBT containing 5% normal goat serum, and incubated with primary anti-GFP (Sigma) and NC82 (Iowa Hybridoma Bank) counterstain antibodies in blocking solution overnight at 4 C. Samples were washed three times for 10 min in PBT at RT, and secondary antibodies were applied in blocking solution for 2 hr at RT. After washing three times for 10 min in PBS, samples were mounted in Vectashield (Vector Labs). Confocal images were captured on a Zeiss LSM510 Meta, Axiovert 200 M, and processed with LSM510 Image Examiner. Hypergeometric enrichment test A hypergeometric test, similar to the test used for GO enrichment analysis, was used to identify over-represented gene lists (C2 from Msigdb, BROAD Institute) and pathways (KEGG) amongst the pain hits. The hypergeometric test considers only the percentage representation of genes corresponding to a biological pathway in the pre-computed heart function gene list. This analysis was performed on the gene list identified as adult pain hits (Z-score.1.65) in Drosophila and their corresponding mouse or human orthologs.

Statistics
For analysis of adult Drosophila avoidance responses a Student's t test was performed. For analysis of larval pain behavior we have performed the Kruskal-wallis non-parametric test for median comparison followed by the Dunn's post-hoc test. For presentation of screening data, a Z-score was generated from (mean control avoidance 2 mean test avoidance)/standard deviation control) and P values were generated from total Z-score distributions. Unless otherwise indicated, data are represented as mean values 6 SEM.

Supporting Information
Table S1 Ca 2+ signalling involved in thermal nociception responses. Ca 2+ signaling components (elav-Gal4.UAS-IR fly lines) that exhibit defects in thermal nociception are listed from the strongest to weakest phenotype for noxious heat avoidance. An avoidance Z-score +/2 SEM and repetitions are included. Lethality was scored for each cross (0 = lethal, 0.5 = semi-lethal, 1 = viable). A mean score of #0.6666 was considered lethal. A qualitative coordination score (0 = uncoordinated, 1 = coordinated) +/2 SEM and number of repetitions are indicated for the rescreened lines. The CG numbers according to flybase annotation version 4.3 (FB4.3) and flybase annotation 5.7 (FB5.7) are included, as is the S19 score for RNAi off targeting effects (OTEs) and number of can repeats. A p-value is also included. Drosophila gene symbols and predicted human and mouse orthologs are shown. (XLSX)