A subset of brain neurons controls regurgitation in adult Drosophila melanogaster

One of the fundamental questions in the field of neuroscience is how the brain responds to different sensory inputs and mediates appropriate behaviors. To address this fundamental question, many have taken advantage of the vinegar fly, Drosophila melanogaster, as a neurogenetic model organism. With a numerically simpler nervous system compared with that in mammals, flies nevertheless exhibit complex behaviors. Importantly, fundamental principles of sensory coding and neuronal circuit function for processing sensory inputs and driving behaviors are often conserved across species. Therefore, Drosophila is a powerful model for functional dissection of neuronal circuits underlying behaviors.

The gustatory system, which influences selection of food, egg deposition sites and mates, among others, is an appealing sensory system to address such questions. The identification of chemosensory receptor genes (Clyne et al., 2000; Scott et al., 2001) and the development of methods to assess feeding behaviors (Ja et al., 2007; Deshpande et al., 2014; Itskov et al., 2014; Ro et al., 2014; Murphy et al., 2017; Shell et al., 2018; Park et al., 2018; Moreira et al., 2019; Yapici et al., 2016; Shiraiwa and Carlson, 2007; Diegelmann et al., 2017) provided a foundation for dissecting the functions of peripheral taste neurons with precise molecular genetic tools. Much is now known about how peripheral taste neurons detect various chemicals (Ling et al., 2014; Weiss et al., 2011; Chen and Dahanukar, 2017; Ledue et al., 2015; He et al., 2019; Raad et al., 2016; Steck et al., 2018; Jaeger et al., 2018), but higher-order gustatory processing in the central brain remains poorly understood. A number of recent studies have utilized powerful genetic screens for higher-order neurons in the brain that process taste information and control feeding behaviors. For example, a number of interneurons and motor neurons have been found to selectively respond to sugars (Miyazaki et al., 2015; Kain and Dahanukar, 2015; Flood et al., 2013; Yapici et al., 2016; Gordon and Scott, 2009) or bitter compounds (Bohra et al., 2018; Kim et al., 2017) and mediate innate feeding responses such as proboscis extension and food ingestion as well as learned taste aversion. In addition, several neuromodulatory interneurons, which modulate taste responses to sugars and bitter compounds, have also been described (Ledue et al., 2016; Youn et al., 2018; Inagaki et al., 2014b; Inagaki et al., 2012). In this study, we aimed to identify candidate higher-order brain neurons involved in processing taste information and mediating feeding behaviors.

We used both VT-GAL4 and Janelia-GAL4 transgenic fly lines to access different subsets of neurons in the adult fly brain (Kvon et al., 2014; Jenett et al., 2012) and asked which if any can induce proboscis extension when activated. We expressed dTrpA1, a heat-activated ion channel (Kang et al., 2011), under the control of a UAS promoter in subsets of neurons labeled by the selected VT-GAL4 and Janelia-GAL4 lines and examined heat-activated proboscis extension responses (PERs) (Shiraiwa and Carlson, 2007). We identified one candidate line (VT041723-GAL4), which labels a neuronal population that mediates regurgitation. Activation of VT041723-GAL4-labeled neurons induces prolonged proboscis extension (proboscis holding) for as long as 7 min without retraction. Similar results were observed by optogenetic activation of these neurons. Pre-feeding of flies with sucrose or water prior to neuronal activation leads to regurgitation, suggesting an aversive response for this prolonged proboscis extension. Using the GFP reconstitution across synaptic partners (GRASP) technique, we found that the VT041723-GAL4-labeled neurons have synaptic connections with peripheral taste neurons in the pharynx. Altogether, our results identify a subset of brain neurons labeled by VT041723-GAL4 that control regurgitation. Our behavioral data also suggest that proboscis extension, a commonly used acceptance feeding behavior readout, might not be a reliable indicator of appetitive feeding behavior.

Flies were reared on standard cornmeal-dextrose-agar food at 25°C and 60–70% relative humidity under a 12 h:12 h dark:light cycle. The following fly strains were used in this study: VT041723-GAL4 (Vienna Drosophila Resource Center) (Kvon et al., 2014), Gr43a-LexA (Miyamoto and Amrein, 2014), Ir76b-LexA (Ganguly et al., 2017), Poxn^ΔM22-B5 (Boll and Noll, 2002), Poxn⁷⁰ (Awasaki and Kimura, 1997), UAS-mCD8-GFP (Weiss et al., 2011), UAS-Syt-GFP, UAS-DenMark (BDSC 33064), UAS-dTrpA1 (BDSC 26263), UAS-CsChrimson (BDSC 55135), UAS-spGFP1-10::Nrx (Fan et al., 2013), LexAop-spGFP11::CD4 (Gordon and Scott, 2009) and LexAop2-6XmCherry-HA (BDSC 52271, 52272).

Flies aged 4–8 days were anesthetized on ice, and brain tissues were dissected in 1× PBST (PBS with 0.3% Triton X-100) followed by fixing with 4% paraformaldehyde in 1× PBST for 30 min at room temperature. After three washes with 1× PBST, samples were blocked with 5% normal goat serum (Sigma-Aldrich, G9023) in 1× PBST. Tissues were incubated in primary antibody solutions for 3 days at 4°C. Primary antibodies were: chicken anti-GFP (1:5000; Abcam, ab13970), rabbit anti-DsRed (1:200; Clontech, 632496) and mouse anti-nc82 (1:20; Developmental Studies Hybridoma Bank). Secondary antibodies (1:400; Invitrogen) were: goat anti-chicken Alexa Fluor 488, goat anti-rabbit Alexa Fluor 546, goat anti-mouse Alexa Fluor 568 and goat anti-mouse Alexa Fluor 647. Samples were mounted in Vectashield antifade mounting medium (Vector Laboratories, H-1000) and stored at 4°C. Fluorescent images were acquired using a Leica SP5 confocal microscope with 400 Hz scan speed in 512×512 or 1024×1024 pixel format. Image stacks were acquired at 1 µm optical sections. All images are presented as maximum projections of the z-stack generated using Leica LAS AF software.

Immunofluorescence staining procedures were similar to those described above with the following minor modifications for detecting GRASP signals between Ir76b-LexA-labeled peripheral taste neurons and VT041723-GAL4-labeled central neurons in the brain. To detect native reconstituted GFP signal, only the primary antibody of mouse anti-nc82 (1:20; Developmental Studies Hybridoma Bank) was used for staining neuropil. The two transgene controls were stained together with experimental genotypes at the same time and imaged with the same settings using a Leica SP5 confocal microscope. Image stacks were acquired at 1 µm optical sections. All images are presented as maximum projections of the z-stack generated using Leica LAS AF software.

Flies of both sexes, aged 4–8 days, were immobilized on glass coverslips with drops of clear, non-toxic nail polish and then allowed to acclimate for 30–60 min in a humidified chamber prepared by filling a pipette tip box with water and placing damp Kimwipes (Kimberly-Clark Kimtech) on top. One by one, each coverslip containing an individual fly was placed on a 31°C heat block and proboscis extensions were observed under a light microscope. In the initial screening of 194 VT-GAL4 and Janelia-GAL4 lines (Fig. 1A), we scored flies showing full proboscis extension as an indication of food acceptance. In subsequent experiments focusing on the VT041723-GAL4 line, we recorded trial number, sex, proboscis extension and extension duration for each experimental trial. Proboscis holding was scored when flies fully extended their proboscis for more than 10 s without retraction. For the experiments examining regurgitation phenotype, flies were starved for 24 h on either water-saturated tissues, and then pre-fed 0.5 µl of 100 mmol l⁻¹ sucrose (Sigma-Aldrich, S7903) (Fig. 4B,C), or dry tissues, and then pre-fed 0.5 µl of distilled water (Fig. 4D). Flies that did not consume the pre-feeding tastant solutions in their entirety were excluded from the analysis. Flies that consumed all of the pre-feeding tastant solutions were transferred to a 31°C heat block and the number of flies showing regurgitation was recorded. Regurgitation was defined by the presence of a liquid bubble at the tip of the proboscis (Fig. 4A). In all experiments, we tested both GAL4 and UAS controls together with experimental flies in parallel, in random order, and experimenters were bind to genotype. Among all control flies, we did not observe any that showed proboscis holding or regurgitation behaviors.

Two to four days after eclosion, flies were transferred to standard cornmeal-dextrose-agar food supplemented with 1 mmol l⁻¹ all-trans-retinal (ATR; Sigma-Aldrich, R2500), and placed in aluminium foil-wrapped vials at 25°C for 2–3 days. Sibling flies were transferred to the same food vials without ATR to serve as controls. Flies were prepared as for the thermogenetically activated PER assay described above, with the exception that they were prepared under low-light conditions, in which the intensity of room lights was too low to activate CsChrimson. Flies were then stimulated with 626 nm LED light (Super Bright LEDs Inc.), and the number of flies showing proboscis holding was recorded. In all experiments, we performed tests on both control and experimental flies on each day, in random order, and experimenters were blind to fly genotype and rearing conditions.

All data are presented as median±interquartile range. Statistical tests were conducted using Prism 8 (GraphPad software). Differences between means of different groups were evaluated for statistical significance with unpaired t-tests. All control and experimental genotypes were always tested in parallel, and experimenters were blind to all genotypes and rearing conditions. All independent trials were performed over 2 days.

To identify higher-order brain neurons involved in feeding behaviors, we took advantage of available transgenic resources in the Vienna Tiles GAL4 (VT-GAL4) Library at the Vienna Drosophila Resource Center (VDRC) and the Janelia-GAL4 collection at the Janelia Farm Research Campus. Transgenic GAL4 lines created with different promoter DNA sequences show different labeling patterns that can be visualized with different reporters, such as UAS-GFP. The expression patterns of VT-GAL4 and Janelia-GAL4 lines in the adult Drosophila brain have been well documented (Pfeiffer et al., 2008; Jenett et al., 2012; Kvon et al., 2014). Using the Virtual Fly Brain online database (www.virtualflybrain.org) (Milyaev et al., 2012), we first did a preliminary image-based screen for neurons that arborize in and around the subesophageal zone (SEZ), the primary taste center in the fly brain, and selected several candidate lines for further analysis. Among these, GAL4 lines that showed sparse labeling in the adult brain were prioritized for subsequent behavioral screening. To determine whether any of the selected GAL4 lines labeled neurons involved in feeding behaviors, we expressed the Drosophila transient receptor potential channel, subfamily A, member 1 (dTrpA1), a heat-activated cation channel (Kang et al., 2011), using the GAL4/UAS binary expression system (Brand and Perrimon, 1993). By elevating the ambient temperature to 31°C, we could thermogenetically activate these neurons and record the PER, in which the fly protrudes its mouthpart (proboscis), as a readout of feeding behavior (Shiraiwa and Carlson, 2007). From a preliminary screen of 194 GAL4 lines (155 VT-GAL4 lines and 39 Janelia-GAL4 lines) (Table S1), we found five lines (VT062245-GAL4, VT040416-GAL4, VT041723-GAL4, VT038168-GAL4 and R77B08-GAL4) that exhibited more than 40% PER (Fig. 1A). Closer examination of the expression patterns of the five lines excluded three (VT062245-GAL4, VT038168-GAL4 and R77B08-GAL4) based on expression in peripheral taste neurons that project to the SEZ (Fig. 1B) (Kwon et al., 2014). Interestingly, PER activated by the VT041723-GAL4 line was unique in that the flies did not retract the proboscis after extension, but rather maintained it in the extended position at length (Fig. 1C; Movie 1). We termed this unusual PER response ‘proboscis holding’ and selected the VT041723-GAL4 line for further analysis.

To determine whether both males and females exhibited proboscis holding upon activation of VT041723-GAL4 neurons, we performed the heat-activated PER assay with mated male and female flies of both experimental and control genotypes (Fig. 1D,E). The proboscis holding phenotype was recorded on an all-or-nothing basis. If a fly extended its proboscis for 10 s or longer upon heat activation, it was considered to have proboscis holding. If the fly did not extend its proboscis, or if the duration of proboscis extension was less than 10 s, it was considered to have no proboscis holding. As expected, both male and female control flies with either VT041723-GAL4 or UAS-dTrpA1 transgenes did not show any proboscis holding under any test conditions. The experimental VT041723-GAL4>UAS-dTrpA1 flies demonstrated varying levels of proboscis holding between sexes. We found that 10.7% of male flies (N=56) and 54.5% of mated female flies (N=66) showed the proboscis holding response (Fig. 1D,E). As starvation increases the PER response in flies (Dethier, 1976), we next assessed whether VT041723-GAL4 neuron-activated proboscis holding behavior is modulated by starvation. We tested flies that were starved for 24 h (N=63 for males and N=74 for females) and found that similar fractions of fed and starved flies exhibited proboscis holding (Fig. 1D,E).

To further investigate the nature of proboscis holding in VT041723>dTrpA1 flies, we recorded the duration of proboscis holding in fed and starved flies that showed this behavior. For feasibility, we capped measurement of proboscis holding time at 7 min. Our results showed that the average proboscis holding duration was not significantly different between fed and starved flies of the same sex (unpaired t-test for males and Mann–Whitney test for females, P>0.05). However, mated female flies showed significantly longer times of proboscis holding compared with males in both fed and starved conditions (unpaired t-tests, P<0.05) (Fig. 1F). In fact, many female flies held the proboscis in the extended position for the maximum recording time (7 min) (Movie 1). Together, our results show that activation of VT041723-GAL4 neurons induces proboscis holding in a sexually dimorphic manner, with females exhibiting proboscis holding at a higher frequency and for a longer duration.

We next verified the role of VT041723-GAL4 neurons in proboscis holding in an independent optogenetic activation paradigm using a red-shifted channelrhodopsin, CsChrimson (Klapoetke et al., 2014). Experimental flies were transferred to food supplemented with ATR for 2–3 days in the dark and tested for behavioral responses with 626 nm red LED stimulation. Consistent with the results of thermogenetic activation experiments (Fig. 1), optogenetic activation of VT041723-GAL4 neurons resulted in proboscis holding (Fig. 2A; Movie 2). We noted, however, that in most cases the proboscis was not fully extended (partial proboscis holding) by optogenetic activation. Nonetheless, these flies also maintained the partial proboscis holding for up to 7 min under continuous red LED exposure, at which point the trial was completed (see Movie 2). Further, the partial proboscis holding responses were sexually dimorphic; 4.3% of male flies (N=47) and 39.1% of mated female flies (N=69) exhibited the phenotype (Fig. 2B). Control flies that were not given ATR food (−ATR) showed little if any proboscis holding upon light stimulation (N=36 for males and N=71 for females).

We next examined the expression pattern of VT041723-GAL4 in the brain using UAS-GFP. Similar to the expression pattern described previously (Kvon et al., 2014), we found labeling in neurons that showed dense innervation in the antennal mechanosensory and motor center (AMMC), and some labeled neurites traveling across the midline between the SEZ and the pars intercerebralis regions (Fig. 3A). Some weakly labeled cell bodies were observed within the SEZ. Notably, one pair of neurons in the dorsolateral protocerebrum was strongly labeled, and their anatomical characteristics were reminiscent of previously reported Gr43a⁺ fructose-sensing neurons in the brain (Miyamoto et al., 2012). To confirm whether VT041723-GAL4 labeled Gr43a⁺ neurons, we performed double-labeling experiments with two fluorescent reporters driven by VT041723-GAL4 and Gr43a-LexA, respectively (Fig. 3B). We found no overlap between expression of the two reporters, indicating that VT041723-GAL4 labels a different set of neurons in the brain.

To characterize the neuroanatomy of VT041723-GAL4 neurons in more detail, we expressed the presynaptic marker Syt-GFP (Zhang et al., 2002) and the postsynaptic marker DenMark (Nicolai et al., 2010) and examined their distribution in the brain (Fig. 3C). We found the Syt-GFP signal was located medially relative to DenMark in the protocerebrum region. Both Syt-GFP and DenMark signals were observed in the AMMC and the SEZ. In the AMMC, DenMark was distributed across the whole neuropil whereas Syt-GFP was confined to the lateral AMMC region. In summary, the VT041723-GAL4 line labels neurons in the anterior SEZ as well as the dorsolateral protocerebrum of the fly brain, consistent with a role in controlling proboscis extension and holding.

We next aimed to determine whether the VT041723-GAL4-activated proboscis holding phenotype is modulated by prior feeding experience. To test this possibility, we starved the VT041723-GAL4>UAS-dTrpA1 flies for 24 h and then pre-fed the flies with a fixed amount of 100 mmol l⁻¹ sucrose (0.5 µl) immediately before transferring them to the 31°C heat block for thermogenic activation (Fig. 4A; Movie 3). Surprisingly, we found that half of male (49.1%) and more than half of mated female (76.3%) flies exhibited regurgitation (Fig. 4B), which was apparent by the formation of a liquid bubble at the tip of the proboscis (Fig. 4A). In addition, about 10% of the flies showed proboscis holding without regurgitation. These results suggest that activation of VT041723-GAL4 neurons conveys an aversive signal that causes regurgitation of an ingested meal.

We next asked whether starvation time affects the regurgitation phenotype. For this purpose, we performed mild starvation (4 h) before pre-feeding flies with 0.5 µl of 100 mmol l⁻¹ sucrose. Similar to the results obtained with 24 h starvation, we found more than half of the male (72.3%) and half of mated female (50%) flies exhibited regurgitation upon activation of VT041723-GAL4 neurons (Fig. 4C). In addition, regurgitation behavior was also observed when flies were pre-fed with 0.5 µl of water after starvation on dry tissue paper, suggesting that the observed behavioral response is independent of tastants in the pre-fed meal (Fig. 4D). Thus, VT041723-GAL4-induced regurgitation of a meal appears to be independent of starvation state and meal quality.

We next investigated the possibility that VT041723-GAL4 neurons may be part of taste circuits by performing GRASP experiments (Fan et al., 2013). We first examined the expression of both VT041723-GAL4 and Ir76b-LexA in the fly brain. Ir76b-LexA labels some olfactory neuronal projections in the antennal lobes as well as projections in the SEZ from many gustatory receptor neurons (GRNs) from different taste organs, including those in labellar taste hairs, labellar taste pegs, pharynx and legs (Ganguly et al., 2017; Zhang et al., 2013; Hussain et al., 2016; Chen and Dahanukar, 2017; Steck et al., 2018; Jaeger et al., 2018; Chen and Amrein, 2017; Ahn et al., 2017). We found that neurites of VT041723-GAL4 neurons and Ir76b⁺ pharyngeal GRNs appeared to be in close proximity to each other in the SEZ (Fig. 5A). We then performed a GRASP experiment by expressing split GFP1-10 fused with a transmembrane protein involved in synapse formation (Knight et al., 2011), neurexin, in VT041723-GAL4 neurons, and split GFP11 fused with CD4 in Ir76b⁺ neurons. We stained the neuropil using anti-nc82 and visualized direct GFP fluorescence. Controls lacking either VT041723-GAL4 or Ir76b-LexA did not show any GFP signal. A different candidate line from our screen (Fig. 1A,B), VT040416-GAL4, that labels extensive neurite arborization in the SEZ (Fig. S1), also did not show any positive GRASP signal with Ir76b⁺ GRNs. Notably, we observed reconstitution of GFP fluorescence in the SEZ when VT041723-GAL4 and Ir76b-LexA were used to express the two split GFP components (Fig. 5B), suggesting that termini of VT041723-GAL4 neurons are in close proximity with those of Ir76b-LexA GRNs, and may receive taste input from Ir76b⁺ neurons.

One previous study showed that thermogenetic activation of Gr66a-expressing taste neurons in the mouthpart caused regurgitation (Kang et al., 2011), which raised the possibility that VT041723-GAL4 neurons receive input from pharyngeal Gr66a⁺ GRNs. To test this possibility, we used Pox-neuro (Poxn) mutants in which all external taste hairs are transformed into mechanosensory hairs, leaving pharyngeal taste neurons intact (Chen et al., 2018; Chen and Dahanukar, 2017; Ledue et al., 2015). Consistent with our previous report (Chen and Dahanukar, 2017), Poxn mutants retained Ir76b⁺ projections from the pharynx and a few taste pegs, while lacking projections from all external taste organs. GRASP experiments in a Poxn mutant background revealed a positive GRASP signal between VT041723-GAL4 and Ir76b-LexA GRNs in the SEZ (Fig. 5D). The results support the idea that VT041723-GAL4 neurons receive taste input from pharyngeal GRNs and regulate regurgitation.

Knowledge about how neural circuits are wired in the brain is crucial for understanding how sensory information is translated into behavior. In Drosophila, higher-order brain regions that process olfactory information, such as the lateral horn and mushroom body, have been described in detail (Dolan et al., 2019; Jefferis et al., 2007; Marin et al., 2002; Wong et al., 2002), but much less is known about processing of gustatory information after the first relay in the SEZ, with reports of only a few central neurons that have been anatomically or functionally characterized (Bohra et al., 2018; Kim et al., 2017; Yapici et al., 2016; Miyazaki et al., 2015; Kain and Dahanukar, 2015; Flood et al., 2013). In this study, we identified that activation of VT041723-GAL4-labeled neurons causes proboscis holding and regurgitation behavior in adult Drosophila.

Proboscis extension has been characterized as an appetitive behavioral response and is widely used as a read-out of food acceptance (Shiraiwa and Carlson, 2007). Several previous reports have shown that activation of external sweet taste neurons via Gr5a-GAL4 causes proboscis extension (Inagaki et al., 2012; Inagaki et al., 2014a; Dawydow et al., 2014; Du et al., 2016; Kain and Dahanukar, 2015; Yapici et al., 2016; Keene and Masek, 2012). Under these conditions, flies usually exhibit proboscis extensions followed by quick retractions. As activation of VT041723-GAL4 neurons resulted in a single extension without retraction for the duration of the assay, we considered that it may not be indicative of an appetitive response but rather that it represented an aversive response. Consistent with this idea, post-consumption activation of VT041723-GAL4 neurons induced regurgitation, similar to that observed in flies with stimulation of deterrent taste neurons (Kang et al., 2011) or with overconsumption (Pool et al., 2014). However, VT041723-GAL4 neurons induced regurgitation that was often accompanied by proboscis holding, and sustained proboscis extension is typically observed only when the fly is actively ingesting. We cannot, therefore, exclude the possibility that proboscis holding and regurgitation are controlled by different subsets of VT041723-GAL4 neurons. Alternatively, proboscis holding may be a common feature of feeding and regurgitation behaviors.

In this study, we found that the frequency of proboscis holding behavior is strikingly higher in females than in males. In Drosophila, doublesex (dsx) and fruitless (fru) are known as sex-determining transcription factors that specify sexually dimorphic neuronal circuits and behaviors (Erdman and Burtis, 1993; Ito et al., 1996; Ryner et al., 1996; Auer and Benton, 2016; Asahina, 2018). Although we found no sexual dimorphism in the pattern of VT041723-GAL4 expression in the brain (data not shown), a closer look at the expression of sex-specific fru and dsx in VT041723-GAL4 neurons would provide further insight into possible mechanisms underlying sexual dimorphism. In addition, sex-specific differences in feeding responses to salt (Walker et al., 2015), yeast (Ribeiro and Dickson, 2010), amino acids (Ganguly et al., 2017) and sugars (Chandegra et al., 2017) have been reported. Given the possibility of functional connectivity between VT041723-GAL4 neurons and peripheral taste neurons, it will be of interest to determine whether specific gustatory input is involved in sex-dependent variation in the proboscis holding phenotype. Moreover, as the sexual difference is lost when flies are pre-fed with either water or sucrose and tested in thermogenetic activation experiments, it appears that prior feeding experience differentially influences the proboscis holding phenotype in males and females.

VT041723-GAL4 labels multiple neurons that can be largely separated into two anatomical groups, one near the dorsolateral protocerebrum and a second around the SEZ with extensive neurite arborization in the AMMC. Although our study did not identify which of the two populations is involved in regurgitation behavior, GRASP experiments implicate the latter, which are poised to receive input from pharyngeal Ir76b⁺ GRNs, which encompass Gr66a⁺ GRNs in the number 8 and 9 sensilla of the labral sense organ (LSO) (Chen and Dahanukar, 2017) that induce regurgitation (Kang et al., 2011). Gr66a is broadly expressed in many bitter taste neurons and mediates feeding avoidance of various aversive compounds (Weiss et al., 2011; Moon et al., 2006; Marella et al., 2006; Wang et al., 2004; Thorne et al., 2004). It is plausible, therefore, that pharyngeal Gr66a⁺ GRNs act as a final checkpoint for food consumption and sense cues that induce regurgitation of unsavory meals via activation of VT041723-GAL4 neurons.

PER requires precise coordination of various motor programs, including rostrum lifting, haustellum extension, labella extension and labella spreading. Recently, motoneurons controlling the individual motor sequence of the PER have been described at the single-cell level (Schwarz et al., 2017). However, motor circuits controlling regurgitation have not been explored and, consequently, little is known about whether PER and regurgitation share common motor programs. Based on our observations, we posit that VT041723-GAL4 neurons provide a good starting point to address such questions. Future experiments using genetic intersectional strategies may identify the minimum subset of VT041723-GAL4 neurons that are required for regurgitation behavior. Overall, our results lay the groundwork to analyze a simple behavior and the neuronal circuits and conditions that control it.

We would like to thank members of the Dahanukar laboratory for useful comments on the manuscript, and Sen Miao for her help with the initial thermogenetic screen.