Gαq and Phospholipase Cβ signaling regulate nociceptor sensitivity in Drosophila melanogaster larvae

Drosophila stocks and genetics

All larvae were reared on cornmeal-molasses medium (Nutri-Fly M; Genesee Scientific, El Cajon, CA, USA). Flies and larvae in all experiments were grown at 25 °C and ∼50% humidity under a 12 h light: 12 h dark cycle. Larvae for all experiments were obtained by washing wandering 3rd instar larvae from the walls of vials using distilled water. The driver stocks w; ppk1.9-GAL4; UAS-dicer2 and w; ppk1.9-GAL4, UAS-mCD8::GFP; UAS-dicer2 were obtained as gifts from the laboratory of the Dr. Dan Tracey as was the w¹¹¹⁸ control stock and the UAS-para-RNAi line. All other RNAi lines were developed as part of the Drosophila Transgenic RNAi Project (TRiP) and obtained from the Bloomington Drosophila Stock Center (BDSC). GL01048 (BDSC# 36820), JF02390 (BDSC# 36775), and JF02464 (BDSC# 33765) were used as UAS-Gαq-RNAi lines. JF01713 (BDSC# 31197) and JF01585 (BDSC #31113) were used as UAS-norpA-RNAi lines. The y¹ v¹; {Py[+t7.7]=CaryPattP}2 (BDSC# 36303) line was used in driver-only control crosses for RNAi experiments. w^∗; {Pw[+mC]=UAS-Galphaq.R}2 (BDSC# 30734) flies were used for UAS-Gαq overexpression experiments. The norpA³⁶ mutant was obtained from BDSC# 9048.

With one exception, the RNAi lines used in this study have been previously validated, as described in the Drosophila RNAi Stock Validation & Phenotypes (RSVP) database (Perkins et al., 2015). All three UAS-Gαq-RNAi lines used in this study have positive validation data for knockdown efficacy, showing embryonic lethality phenotypes following broad knockdown (Yan et al., 2014; Zeng et al., 2015). One UAS-norpA-RNAi line (JF01585) used in this study has positive validation data for knockdown efficacy, showing larval thermotaxis defects following nervous system knockdown (Shen et al., 2011). The second UAS-norpA-RNAi line (JF01713) does not have positive validation data per the RSVP database.

Microarray analysis of mdIV gene expression

To confirm expression of Gαq and norpA in the mdIV neurons, an existing microarray dataset (ArrayExpress Accession # E-MTAB-3863) was analyzed. This dataset was created using the Affymetrix GeneChip Drosophila Genome 2.0 Array and contains data from four biological replicates, each consisting of 40–50 mdIV neuron cell bodies collected via laser capture microdissection (Honjo et al., 2016; Mauthner et al., 2014). Normalized detection values for Gαq and norpA transcripts were compared to those for the trio, dTrpA1, painless, and Gr28b transcripts, all of which are known to be expressed in the mdIV neurons (Iyer et al., 2012; Tracey et al., 2003; Xiang et al., 2010; Zhong et al., 2012).

Thermal nociception assays

Thermal nociception assays were conducted as previously described (Tracey et al., 2003). Wandering 3rd instar larvae were washed from vials into glass petri dishes with distilled water. A small amount (<10 mg) of dry baker’s yeast was added to each petri dish to disrupt surface tension of the water, and enough water was removed from the dish to leave only a thin film of water covering the surface. Larvae were then stimulated along their lateral surface with a custom-built thermal probe consisting of a soldering iron with a tip filed into a ∼6 mm chisel shape plugged into a Variac Variable Transformer (Part No. ST3PN1210B) (ISE, Inc., Cleveland, OH) to control probe temperature. The probe temperature was monitored using an IT-23 thermistor and BAT-12 digital thermometer (Physitemp, Clifton, NJ). The thermistor and thermometer provided temperature information to a tenth of a degree. During experiments, the probe tip temperature was monitored before and after application of the stimulus. Trials in which the probe temperature deviated from the desired temperature by ±0.5 °C were discarded during analysis. Behavioral responses were recorded at 30 frames per second using a digital video camera mounted on a dissecting microscope. Adobe Premier Pro was used to determine the time of probe contact with the larva and execution of NEL for each trial in order to calculate response latency. Larvae that did not respond to the probe within ten seconds were scored as 11 for all subsequent analysis. At least 40 larvae were tested for each genotype in each experiment. Experimenters were blinded to larval genotype during stimulation and scoring. The Wilcoxon Rank-Sum Test was used to determine statistical significance of latency differences between control and experimental genotypes. Bonferroni correction was used to correct the α value for multiple comparisons in experiments with more than one experimental group compared to the control.

Mechanical nociception assays

Mechanical nociception assays were conducted as previously described (Zhong, Hwang & Tracey, 2010). Larvae were prepared in petri dishes as described for thermal nociception assays. Mechanical stimulation was delivered with a custom-built Von Frey filament consisting of a 10 mm length of 8 lb. test nylon fishing line (Stren Original Monofilament 8 lb. line, Part #1304152, Pure Fishing, Inc., Columbia, SC) affixed to a Pasteur pipette. This Von Frey filament delivers a force of ∼50 millinewtons. Each larva was stimulated with the Von Frey filament and NEL was scored as a binary variable in order to calculate the proportion of larvae that responded to the stimulus. Each larva was tested three times, but only the first stimulus was used for subsequent analysis. At least 100 larvae were tested per genotype in each experiment except for the experiment shown in Fig. S1B, in which >50 larvae were tested per genotype. Experimenters were blinded to larval genotype during stimulation and scoring. The Chi-square test was used to test for statistically significant differences in proportion between control and experimental genotypes. Bonferroni correction was used to correct the α value for multiple comparisons in experiments with more than one experimental group compared to the control.

Confocal microscopy

Wandering 3rd instar larvae were immobilized by circumferential ligation by a hair around segment A3. This manipulation paralyzed all body segments posterior to the ligation. Immobilized larvae were then mounted in glycerol between two glass coverslips and subjected to confocal microscopy using a Zeiss LSM 880 microscope with a 488 nm laser line. Tiled z-stacks were obtained in order to capture the full arborization of ddaC mdIV neurons from body segments posterior to the ligation. Neurons from the A8 body segments were not captured or analyzed, as these appeared to have qualitatively different arborization from more anterior segments. Dendrites were analyzed using the NeuronJ plugin for ImageJ (Meijering et al., 2004). In brief, dendrites were traced manually in order to measure the total length of individual neurons’ arborizations and to count the total number of branches present. Experimenters were blind to larval genotype during tracing and analysis. Student’s t-test was used to test for statistically significant differences between control and experimental conditions. Experimenters were blinded to genotype during this analysis to avoid bias in dendrite tracing.

Gαq is required for normal sensitivity to noxious thermal and mechanical stimuli

In order to test the role of Gαq signaling in larval nociception, we tested larvae with nociceptor-specific Gαq knockdown for defects in thermal and mechanical nociception. Before knocking down Gαq with RNAi, we confirmed that Gαq transcript is present in the nociceptors under normal conditions. We analyzed an existing Affymetrix microarray dataset created using mdIV cell bodies collected using laser capture microdissection(Honjo et al., 2016; Mauthner et al., 2014). We found that Gαq transcript was detected at a similar or higher level to trio, dTrpA1, painless, and Gr28b transcripts, all of which are known to be expressed in the mdIV neurons (Fig. S1) (Iyer et al., 2012; Tracey et al., 2003; Xiang et al., 2010; Zhong et al., 2012).

In our first behavioral experiment, we used the ppk-GAL4; UAS-dicer2 line to drive nociceptor-specific expression of the GL01048 UAS-Gαq-RNAi transgene and tested for defects in thermal nociception. We found that Gαq RNAi larvae had significantly longer latency to respond to a 46 °C probe than controls with no RNAi transgene (Fig. 1A). Gαq RNAi larvae displayed a mean latency of 3.0 s, compared to a mean latency of 1.9 s for GAL4-only controls. During this experiment and all subsequent experiments, para knockdown larvae were used as a positive control and found to show nearly complete insensitivity to noxious thermal stimuli (Fig. 1A).

To confirm the apparent hyposensitive thermal nociception phenotype of Gαq knockdown larvae, we tested two additional UAS-Gαq-RNAi lines (carrying the JF02390 and JF02464 transgenes) for defects in thermal nociception (Fig. 1B). We found that larvae with ppk-GAL4-driven expression of the JF02390 RNAi transgene had a significantly longer latency to respond to a 46 °C probe than GAL4-only controls (2.2 versus 1.7 s). Larvae with ppk-GAL4-driven expression of the JF02464 RNAi transgene did not differ significantly from GAL4-only control larvae in their response latency (1.9 versus 1.7 s). We also compared ppk-GAL4-driven expression of UAS-Gαq-RNAi to UAS-RNAi-only controls (Fig. 1C). We found that larvae with ppk-GAL4-driven expression of the GL01048 and JF02390 RNAi transgenes had significantly longer response latencies than their respective UAS-RNAi-only controls (4.1 versus 2.3 s for GL01048; 2.4 versus 1.8 s for JF02390). Larvae with ppk-GAL4-driven expression of the JF02464 RNAi transgene did not differ significantly from UAS-RNAi-only control larvae in their response latency (2.9 versus 2.6 s). Taken together, these results suggest that Gαq knockdown produces a hyposensitive thermal nociception phenotype.

In order to determine whether the hyposensitive Gαq knockdown phenotype is specific to thermal nociception or present for other nociceptive modalities, we also tested nociceptor-specific Gαq-RNAi for defects in mechanical nociception. In our first experiment, we used the ppk-GAL4; UAS-dicer2 line to drive nociceptor-specific expression of the GL01048 UAS-Gαq-RNAi transgene and stimulated larvae with a 10 mm Von Frey filament to induce nociceptive responses. We found that larvae with nociceptor-specific Gαq knockdown showed a significantly lower frequency of nociceptive responses compared to a control with no RNAi transgene (50.8% of larvae responding to first stimulus versus 71.7% of larvae responding to first stimulus) (Fig. 1D). To confirm this hyposensitive mechanical nociception phenotype, we also tested the responses of the JF02390 and JF02464 UAS-Gαq-RNAi lines to the same noxious mechanical stimulus. We found that each of these RNAi lines also produced a defective mechanical nociception phenotype (44.2% of JF02390 larvae and 45.8% of JF02464 larvae responding to the first stimulus versus 74.2% of GAL4-only control larvae responding to the first stimulus) (Fig. 1E). We also compared ppk-GAL4-driven expression of UAS- Gαq-RNAi to UAS-RNAi-only controls (Figs. 1F and 1G). We found that larvae with ppk-GAL4-driven expression of the GL01048 and JF02464 RNAi transgenes responded at a lower frequency than their respective UAS-RNAi-only controls (69.4% versus 41.1% for GL01048; 54% versus 38% for JF02464). Larvae with ppk-GAL4-driven expression of the JF02390 RNAi transgene did not differ significantly from UAS-RNAi-only control larvae in their response rate (74% versus 63% responding).

Overexpression of Gαq results in hypersensitive thermal and mechanical nociception

Our observation that nociceptor-specific Gαq knockdown results in hyposensitive thermal and mechanical nociception phenotypes suggests that Gαq signaling normally acts to increase the basal sensitivity of the mdIV neurons to noxious stimuli. To test this hypothesis, we overexpressed Gαq in the mdIV neurons by driving expression of a UAS-Gαq transgene with the ppk-GAL4 driver and tested larval responses to noxious thermal and mechanical stimuli. We found that Gαq-overexpressing larvae responded significantly faster than UAS-only and GAL4-only control larvae to a 46 °C thermal stimulus (Fig. 2A). Overexpression larvae responded with an average latency of 2.2 s, while UAS-only and GAL4-only controls responded with latencies of 2.8 s. The latency for overexpression larvae was significantly shorter than both controls. These results are consistent with Gαq overexpression causing hypersensitivity to noxious thermal stimuli.

To determine whether Gαq overexpression also leads to increased sensitivity to mechanical stimuli, we tested overexpression larvae for mechanical nociception responses. We found that a significantly greater proportion of larvae with nociceptor-specific Gαq overexpression responded to a noxious mechanical stimulus than did UAS-Gαq-only control larvae (84.0 percent responding to the first stimulus versus 65.3 percent responding to the first stimulus) (Fig. 2B). However, the difference in proportion of Gαq overexpression larvae and GAL4-only control larvae that responded to the mechanical stimulus was not significantly different (84.0 percent versus 71.0 percent responding to the first mechanical stimulus). These results only partially support an effect of Gαq overexpression on mechanical nociception, as the overexpression genotype is significantly different from only one of two control conditions.

NorpA is required for normal sensitivity to noxious thermal and mechanical stimuli

One of the principal effectors of Gαq is PLCβ, and one of the major PLCβ enzymes in Drosophila neurons is encoded by the norpA gene. norpA transcripts are detected in larval nociceptors by microarray at similar levels to other known mdIV transcripts (Fig. S1). Given the thermal and mechanical nociception phenotypes that we have observed in nociceptor-specific Gαq knockdown and overexpression animals and previous studies detailing roles for NorpA in thermosensation, we hypothesized that NorpA signaling may also be required in the mdIV nociceptors for normal sensitivity to thermal and mechanical stimuli. To test this hypothesis, we used the ppk-GAL4 driver to express norpA RNAi in the nociceptors and tested larvae for nociception defects. First, we crossed the ppk-GAL; UAS-dicer2 driver line with the JF01713 and JF01585 UAS-norpA-RNAi lines and tested responses of larval progeny to a 46 °C thermal probe (Fig. 3A). We found that larvae expressing the JF01713 UAS-norpA-RNAi transgene possessed a significantly longer latency to the noxious thermal stimulus than the GAL4-only controls (3.8 s versus 1.7 s). However, larvae expressing the JF01585 UAS-norpA-RNAi transgene were indistinguishable from GAL4-only controls in their latency to respond to a thermal stimulus (1.7 s for both genotypes). We also compared ppk-GAL4-driven expression of UAS-norpA-RNAi to UAS-RNAi-only controls (Fig. 3B). We found that larvae expressing the JF01713 UAS-norpA-RNAi transgene possessed a significantly longer latency to the noxious thermal stimulus than UAS-RNAi-only controls (4.1 s versus 2.2 s). However, larvae expressing the JF01585 UAS-norpA-RNAi transgene were indistinguishable from UAS-RNAi-only controls in their latency to respond to a thermal stimulus (2.0 s versus 2.3 s). These results are consistent with a modest effect of norpA knockdown on thermal nociception.

Because the two norpA RNAi lines that we tested did not produce a consistent thermal nociception phenotype, we also tested a norpA mutant for thermal nociception defects. The norpA³⁶ mutant is a 28 bp deletion in the norpA gene that produces a frameshift in the norpA transcript and a truncated protein (Pearn et al., 1996). We found that the norpA³⁶ mutant responded to a noxious thermal stimulus with a significantly longer latency than w¹¹¹⁸ control larvae (4.9 s versus 2.5 s) (Fig. 3C). Thus the norpA loss-of-function phenotype is consistent with the phenotype produced by nociceptor-specific knockdown using the JF01713 UAS-norpA-RNAi transgene.

To determine whether norpA knockdown larvae also have defects in mechanical nociception, we crossed the ppk-GAL; UAS-dicer2 driver line with the JF01713 and JF01585 UAS-norpA-RNAi lines and tested the larval progeny for responses to a 10 mm Von Frey filament. We found that a significantly smaller proportion of larvae expressing the JF01585 UAS-norpA-RNAi transgene responded to the noxious mechanical stimulus than GAL4-only control larvae (55 percent responding to the first stimulus versus 71.7 percent responding to the first stimulus) (Fig. 3D). However, the proportion of larvae expressing the JF01713 UAS-norpA-RNAi transgene that responded to the stimulus was not significantly different from GAL4-only controls (63.3 percent responding versus 71.7 percent responding). We also compared ppk-GAL4-driven expression of UAS-norpA-RNAi to UAS-RNAi-only controls (Fig. 3E). We found that larvae expressing the JF01713 and JF01585 UAS-norpA-RNAi transgenes both responded at significantly lower frequency than their respective UAS-RNAi-only controls (63% versus 43% for JF01713; 61% versus 40% for JF01585). These results are consistent with a mechanical nociception defect produced by norpA knockdown using either UAS-RNAi transgene.

Gαq and NorpA are not required for nociceptor dendrite morphogenesis

Nociception defects in Drosophila larvae may arise from defects in mdIV development and morphogenesis. In order to determine whether manipulation of Gαq and NorpA signaling causes nociception defects via influence on mdIV dendritic arborization, we analyzed the arborizations of mdIV neurons in Gαq and norpA knockdown larvae. We used the ppk-GAL4 driver to express mCD8::GFP and Gαq and norpA RNAi in the mdIV neurons and then imaged the dendritic arborizations of individual ddaC mdIV neurons in immobilized larvae (Figs. 4A–4C). We then counted the total number of dendrite branches and measured the total dendrite length of individual ddaC neurons in order to provide a quantitative measure of dendrite arborization. We found that Gαq and norpA knockdown (using the JF02390 and JF01713 RNAi lines respectively) did not significantly affect the total length or number of ddaC dendrites as compared to a no-RNAi control (Figs. 4D and 4E). These data suggest that Gαq and norpA knockdown does not produce gross morphological changes in the mdIV neurons.