1. Genetics and Genomics
  2. Neuroscience

Thirst interneurons that promote water seeking and limit feeding behavior in Drosophila

  1. Dan Landayan
  2. Brian P Wang
  3. Jennifer Zhou
  4. Fred W Wolf  Is a corresponding author
  1. Quantitative and Systems Biology Graduate Program, UC, United States
  2. Department of Molecular and Cell Biology, UC, United States
https://doi.org/10.7554/eLife.66286
Share this article
Cite this article
  1. Dan Landayan
  2. Brian P Wang
  3. Jennifer Zhou
  4. Fred W Wolf
(2021)
Thirst interneurons that promote water seeking and limit feeding behavior in Drosophila
eLife 10:e66286.
https://doi.org/10.7554/eLife.66286
  2. Download BibTeX
  3. Download .RIS

Abstract

Thirst is a motivational state that drives behaviors to obtain water for fluid homeostasis. We identified two types of central brain interneurons that regulate thirsty water seeking in Drosophila, that we term the Janu neurons. Janu-GABA, a local interneuron in the subesophageal zone, is activated by water deprivation and is specific to thirsty seeking. Janu-AstA projects from the subesophageal zone to the superior medial protocerebrum, a higher order processing area. Janu-AstA signals with the neuropeptide Allatostatin A to promote water seeking and to inhibit feeding behavior. NPF (Drosophila NPY) neurons are postsynaptic to Janu-AstA for water seeking and feeding through the AstA-R2 galanin-like receptor. NPF neurons use NPF to regulate thirst and hunger behaviors. Flies choose Janu neuron activation, suggesting that thirsty seeking up a humidity gradient is rewarding. These findings identify novel central brain circuit elements that coordinate internal state drives to selectively control motivated seeking behavior.

Introduction

Thirst is the neural representation of the internal state caused by dehydration or decreased blood volume. Thirst motivates a coordinated series of water seeking and intake behaviors across diverse animal phyla. Terrestrial animals with unpredictable access to patchily distributed sources of water, such as the fly Drosophila melanogaster, have evolved robust brain circuitry to rapidly detect water loss to drive goal-directed seeking and ingestion behaviors.

In flies, water ingestion and humidity preference studies led to the discovery of sensory and central brain circuit elements that control behavioral and hormonal aspects of thirst. Flies rapidly detect humidity gradients via humid-sensing (Ir68a receptor) and dry-sensing (Ir40a receptor) hygrosensory neurons that are housed in the sacculus on the antennae (Enjin et al., 2016; Ji and Zhu, 2015; Knecht et al., 2016; Knecht et al., 2017). The hygrosensory neurons project to dedicated glomeruli in the antennal lobe, where they synapse onto second order projection neurons to send information to the mushroom bodies, lateral horn, and broadly to other areas of the protocerebrum (Frank et al., 2017; Marin et al., 2020). Water taste is facilitated by water-sensing (PPK28 mechanoreceptor) gustatory neurons; taste is critical for water consumption in thirsty flies (Cameron et al., 2010; Lau et al., 2017). A pair of interoceptive SEZ neurons (ISN) that detect osmolarity are located in the suboesophageal ganglion zone (SEZ); they promote water consumption and reciprocally inhibit food intake when inhibited by high interstitial fluid osmolarity (Jourjine et al., 2016). The ISNs are regulated by the mechanoreceptor Nanchung that responds to changes in hemolymph osmolarity to signal water balance status, and by adipokinetic hormone that signals aspects of energy store status to the brain.

These sensory and interoceptive inputs regulate central brain circuitry to facilitate ingestive behavior and initiate the state of thirst. How the thirsty state is represented in the brain remains to be characterized. Functional mapping of water reward memories led to the discovery of leucokinin peptidergic neurons, whose activity is regulated by thirst, and that regulate memories of thirst through dopaminergic connections into the mushroom body learning and memory circuitry (Ji and Zhu, 2015; Lin et al., 2014; Senapati et al., 2019). Thirsty attraction to humidity may also route through some of the same dopamine neurons (Lin et al., 2014). Hunger, which drives a pattern of ingestive behaviors similar to that of thirst, is also conveyed through dopamine neurons and the mushroom body circuitry (Tsao et al., 2018). However, the role of hunger circuitry in water seeking is not yet known. Finally, the pars intercerebralis, a neuroendocrine brain region, contains ion transport peptide (ITP) hormone expressing neurons (Gáliková et al., 2018; Nässel and Vanden Broeck, 2016). Ion transport peptide bidirectionally controls thirst and hunger behaviors, and it regulates water balance in the body as a hormone.

Understanding how fundamental motivational states are related at the circuit level will help uncover the hierarchy of behavioral choice and its underlying neural mechanisms. Thirst in mammals has complex hierarchical interactions with other behavioral states, including hunger and sleep (Sutton and Krashes, 2020). Thirst and hunger are particularly tightly interconnected. For example, thirst can override hunger drive (dehydration-anorexia), and feeding induces thirst (prandial thirst) (Watts and Boyle, 2010). Manipulation of either cortical or hindbrain neurons in mice can simultaneously impact both hunger and thirst driven ingestive behavior (Eiselt et al., 2021; Gong et al., 2020). Moreover, thirst-driven water anticipation and subsequent intake changes the activity of a remarkably broad set of neurons throughout the mouse brain, in a manner that is time-locked to specific behavioral steps (Allen et al., 2019). This reveals the complexity of even an apparently simple motivational state, and it predicts interactions with other brain functions. Flies, exhibiting hunger:thirst and hunger:sleep interdependencies with apparently equivalent ethological relationships, will be useful for elaborating circuit motifs for behavioral state interactions, particularly at the level of individual neurons (Melnattur and Shaw, 2019).

To identify central brain neurons critical for thirst, we used an open-field behavioral assay that allows neural manipulation of freely moving animals as they evaluate the presentation of a point source of water or food. An unbiased neuroanatomical activation screen identified two bilateral pairs of neurons, Janu-GABA and Janu-AstA, whose activity is necessary and sufficient for promoting thirsty water seeking, but that have limited effect on fluid consumption. Flies choose to continuously activate the Janu neurons, suggesting that water seeking up a humidity gradient is rewarding. The Janu-GABA local interneurons are specific to thirsty water seeking. In contrast, the Janu-AstA neurons, that release the neuropeptide Allatostatin A (AstA), also inhibit feeding behavior. Neuropeptide F (NPF) neurons are a postsynaptic target of Janu-AstA neurons through the AstA-R2 receptor in the superior medial protocerebrum. NPF reciprocally regulates water seeking and feeding behavior. Thus, we uncovered central brain interneurons that encode water seeking and that simultaneously suppress food seeking, indicating that thirst and hunger are coordinated homeostatic states while animals are actively seeking.

Results

Anatomical activation screen identifies water-seeking neurons

Thirsty flies are motivated to seek a palatable source of water and to drink until they reach repletion (Figure 1A). We developed an open-field assay to assess appetitive behaviors, including thirst (Figure 1B). Flies are placed into an open arena, allowed to acclimate, and then presented with a small receptacle containing a source of water or food. A water source rapidly sets up a stable 3–6% humidity gradient from the edge to the source at the center of the chamber. Flies may perform a complete series of steps of thirst-driven behavior with an accessible water source, from seeking to achieving repletion, or they can be limited to water seeking by placing a mesh grid over the source. The longer flies are water deprived, the more quickly they find and occupy the water source (Figure 1C,D). Water deprivation also increases locomotor speed (Figure 1E). Given a choice, thirsty flies seek water over food, and hungry flies seek food over water, indicating that deprivation state-specific seeking can be readily measured in the open field assay (Figure 1F). A combination of hygrosensory and taste cues guide thirsty flies to water. Blocking access to the water source with the mesh grid had no effect on seeking, whereas removal of the third antennal segment that harbors hygrosensory neurons in the sacculus (antennectomy) reduced seeking (Figure 1G). Combining antennectomy with the mesh grid completely blocked seeking. Antennectomy also removes the aristae that sense temperature; ablation of the aristae alone had no effect on seeking to the inaccessible water source (Budelli et al., 2019). Water deprivation increased water intake: removing the third antennal segment had no effect, whereas, as expected, blocking access with the mesh grid fully blocked intake (Figure 1H). Thirsty flies were not attracted to an empty receptacle (not shown). Thus, thirsty flies use hygrosensory and taste environmental cues to locate and drink water, and we can isolate the hygrotactic water seeking step from other thirst-related behaviors.

Figure 1 with 1 supplement see all
Download asset Open asset
Thirst behavior in an open field; a forward screen uncovers Durstig thirst neurons.

(A) Thirst-induced sequence of behavior. (B) Two-chambered open field assay with a vertical divider at the center. Thirsty flies avidly seek and occupy a discrete water source (dyed blue for clarity). (C) Occupancy of an open water source over time by a group of 20 flies, n = 8 groups. (D) Occupancy increases with increasing water deprivation. One-way ANOVA/Dunnett’s compared to 2 hr water deprivation. Each dot represents one group of 20 flies. (E) Water deprivation increases locomotor activity. One-way ANOVA/Dunnett’s compared to 0 hr. (F) Two choice preference for water (1) and dry sucrose (−1) depends on deprivation state. One-sample t-test, compared to 0 (no preference). (G) Sensory systems in thirsty water seeking. A mesh grid atop the water source preserves humidity sensing and blocks water contact, taste, and ingestion. Antennectomy, removal of the third antennal segment and aristae, removes humidity and temperature sensors. One-way ANOVA/Dunnett’s compared to open/unoperated. (H) Sensory input in water intake. Kruskal-Wallis/Dunn’s compared to water replete (gray bar). (I) Screen for neurons that affect the occupancy of food, using a library of InSITE enhancer-Gal4 strains driving expression of the TrpA1 heat activated cation channel. Seeking index is enhancer-Gal4>TrpA1 minus enhancer-Gal4>+. Positive index indicates greater occupancy. Gray extends to two standard deviations from the mean. (J) Durstig neuron activation drives occupancy of food and water. (K) Durstig activation promotes water occupancy through hygrotaxis and contact-dependent mechanisms. Statistics are one-way ANOVA/Tukey’s or Kruskal-Wallis/Dunn’s, unless indicated otherwise.

To identify neurons that promote seeking behavior, we acutely activated neurons in 154 different InSITE enhancer-Gal4 patterns in water replete, well-fed (satiated) flies (Gohl et al., 2011). We initially screened for feeding behavior by presenting flies with a discrete source of standard fly food. The screen revealed patterns of neurons that, when activated by the heat-sensitive TrpA1 cation channel, promote or inhibit foraging (Figure 1I). One activation pattern, Durstig-Gal4 (the German for ‘thirst’), resulted in markedly high attraction to food. We asked what component of the food source the Durstig>TrpA1 flies were seeking. Well fed, water replete Durstig>TrpA1 flies avidly occupied standard fly food, wet sucrose or yeast alone in 1% agar, dry sucrose, 1% agar alone, or water alone, but they were not attracted to empty receptacles (Figure 1J). Hygrotaxis toward inaccessible water was strongly potentiated by Durstig activation (Figure 1K). Thus, Durstig contains neurons that promote water seeking. We chose to focus on water seeking behavior because it is a specific step in the thirst behavioral sequence that is strongly motivated.

We asked if neurons previously implicated in intake (food and water), water reward memory, and water sensing increase water seeking when activated. Activation of dopamine neurons for reward and aversion had either no effect or suppressed water seeking (Figure 1—figure supplement 1A). Moreover, dopaminergic neurons previously implicated in humidity preference and water reward, in the R48B04 pattern, are different from the Durstig water seeking neurons (Figure 1—figure supplement 1B). Of neurons implicated in feeding, activation of only two patterns increased water seeking: the NP883 pattern containing the Fdg neurons that promote proboscis extension and liquid food intake, and the R65D05 pattern that contains AstA neurons that inhibit food intake (Figure 1—figure supplement 1C; Chen et al., 2016; Flood et al., 2013; Hentze et al., 2015; Hergarden et al., 2012; Pool et al., 2014). Activation of internal osmosensory ISNs (R34G02), hygrosensory neurons (Ir40a, Ir68a), and water taste neurons (ppk28) did not increase water seeking (Figure 1—figure supplement 1D; Enjin et al., 2016; Jourjine et al., 2016; Knecht et al., 2016). These findings reveal segregation of water seeking from many of the previously identified neurons involved in steps of feeding and intake. Durstig-Gal4-driven GFP revealed that a complex pattern of neurons is labeled in the brain (Figure 1—figure supplement 1E). Because of this, we sought to isolate the individual neurons responsible for water seeking.

Pattern refinement reveals a small group of neurons that promote thirsty water seeking

We used genetic intersectional techniques to isolate the water seeking neurons embedded in the Durstig pattern. First, subtracting R65D05 neurons from Durstig reduced activation-dependent water seeking, indicating that the water seeking neurons in Durstig and R65D05 are the same (Figure 2A). Activation of central brain R65D05 neurons promoted water seeking, whereas their inactivation decreased seeking, indicating that central brain neurons in the R65D05 pattern (and not ventral nervous system neurons or other non-neuronal cells) are both necessary and sufficient for water seeking (Figure 2B,C; Figure 2—figure supplement 1A–E). Feeding behavior toward dry sucrose was increased with R65D05 inactivation, as expected from previous reports (Figure 2—figure supplement 1F; Chen et al., 2016; Hentze et al., 2015; Hergarden et al., 2012). Second, we expressed the GAL4 inhibitor GAL80 in various neurotransmitter classes of neurons and assessed Durstig activation-driven water seeking. Subtracting out neurons with VGlut-Gal80 (vesicular glutamate transporter enhancer) specifically suppressed Durstig water seeking (Figure 2—figure supplement 2A,B). Next, we activated neurons captured by enhancer-Gal4 transgenes created with enhancer fragments taken from the VGlut locus, and we identified R52A01-Gal4 neurons as being able to promote water seeking (Figure 2—figure supplement 2C,D). Subtracting R52A01 neurons from the Durstig pattern blocked Durstig activation from increasing water seeking, indicating that Durstig water seeking neurons were also found in R52A01 (Figure 2D). Like R65D05, R52A01 activation increased and R52A01 inhibition decreased water seeking (Figure 2E,F). Subtracting out neurons with VGlut-Gal80 decreased thirsty water seeking in both TrpA1-activated R52A01 and R65D05, suggesting that the water seeking neurons were shared between these two Gal4 driver transgenes (Figure 2—figure supplement 2E,F). To test if R65D05 and R52A01 labeled the same water-seeking neurons, we assembled a split-Gal4 (spGal4) derived from these two enhancer fragments that we named Janu (the Estonian for thirst) (Figure 2G). Janu-spGal4 is R52A01-Gal4.AD;R65D05-Gal4.DBD. Janu-spGal4 is expressed in four bilaterally symmetric neurons in the central brain (Figure 2H). Janu neuron postsynaptic inputs are confined to the subesophageal zone (SEZ), and their presynaptic outputs are distributed between the SEZ, the crepine region (CRE), and the superior medial protocerebrum (SMP) (Figure 2I). TrpA1 activation of Janu neurons increased water seeking, albeit to a lesser extent than either parental Gal4 driver alone (Figure 2J). Conversely, acute inactivation of Janu neurons decreased thirsty water seeking (Figure 2K). The behavioral effects of both Janu activation and inactivation on water seeking mapped to the central brain, similar to the parental enhancer-Gal4s for Janu-spGal4 (Figure 2—figure supplement 3A–C). Interestingly, feeding behavior was increased when Janu neurons were acutely inactivated, the opposite effect as compared to water seeking (Figure 2—figure supplement 3D,E). Thus, the Janu neurons promote water seeking and inhibit feeding behavior.

Figure 2 with 4 supplements see all
Download asset Open asset
Isolation of Janu thirst neurons from the Durstig pattern.

(A) Subtraction of R65D05 neurons (R65D05>Gal80) from the Durstig pattern decreases Durstig activation water seeking. One-way ANOVA/Dunnett's compared to Durstig>TrpA1. (B) Activation of R65D05 neurons specifically in the central brain promotes water seeking. otd-nlsFLPo expresses Flippase in the central brain and not the ventral nervous system, to remove the STOP cassette from UAS-STOP-TrpA1, thus restricting TrpA1 expression to central brain R65D05 neurons. (C) Inactivation of R65D05 neurons specifically in the central brain inhibits thirsty water seeking. (D) Subtraction of R52A01 neurons (R52A01>Gal80) from the Durstig pattern decreases Durstig activation (Durstig>TrpA1) water seeking. One-way ANOVA/Dunnett's compared to Durstig>TrpA1. (E) R52A01 neuron activation increases water seeking. (F) R52A01 inactivation decreases thirsty water seeking. (G) A split-Gal4 (spGal4) with R52A01-AD and R65D05-DBD forms functional GAL4 (Janu-spGal4) selectively in the R52A01/R65D05 overlap. (H) Expression pattern of Janu-spGal4 in the brain, counterstained with the Brp synaptic marker. (I) Janu neuron polarity, revealed by presynaptic synaptotagmin-GFP (UAS-sytGFP, green) and dendrite localized UAS-DenMark (magenta). (J) Janu neuron activation increases water seeking. (K) Janu neuron inactivation decreases thirsty water seeking. Statistics are one-way ANOVA/Tukey’s or Kruskal-Wallis/Dunn’s, unless indicated otherwise.

GABAergic local interneurons in the SEZ promote water seeking

To identify the specific neurons that control water seeking and feeding behavior, we characterized the individual Janu neurons. First, we used a stochastic labeling technique to determine the anatomy of each Janu neuron (Figure 2—figure supplement 4). Two of the four central brain Janu neurons project locally in the SEZ, and the other two neurons project from the SEZ to the CRE and/or the SMP. Second, we determined Janu neurotransmitter/neuromodulator identity using immunohistochemistry with antibodies for specific neurotransmitter classes of neurons and for neuropeptides. We found that the two local SEZ neurons were GABAergic, expressing the VGAT vesicular transporter for GABA; we named them Janu-GABA1 and Janu-GABA2 (Figure 3A–H). VGAT precisely localized with nearly all Janu-GABA axonal varicosities, and did not localize with other Janu neurons. Knockdown of VGAT with either of two distinct RNAi transgenes in the Janu neurons decreased water seeking in thirsty animals (Figure 3I; Figure 3—figure supplement 1A). Knockdown of Gad1, encoding the GABA biosynthetic enzyme Glutamic acid decarboxylase 1, in Janu neurons also decreased thirsty water seeking (Figure 3J). To ask if Janu-GABA neurons were responsible for promoting water seeking when Janu neurons were activated in water replete flies, we knocked down Gad1 while simultaneously activating the Janu neurons with TrpA1. This manipulation blocked activation-dependent water seeking (Figure 3K). Thus, the Janu-GABA neurons promote thirsty water seeking and are capable of driving water seeking in water replete animals. To ask if Janu-GABA neurons are specific to the seeking phase of thirst or if they drive a sequence of thirst behaviors, we tested water intake. We presented thirsty flies with a drop of water and measured total water consumption time, the length of the first intake bout, and pharyngeal pumping rate. Gad1 knockdown in Janu neurons had no effect on total water consumption time, but it reduced the length of the first intake bout (Figure 3L,M). Pharyngeal pumping rate was unchanged in thirsty Janu neuron Gad1 knockdown flies, indicating that total intake and motor behaviors were unaffected (Figure 3—figure supplement 1B,C). Finally, feeding behavior was unaltered in the Janu Gad1 knockdown flies (Figure 3N). Thus, the Janu-GABA neurons are thirst-driven water seeking inhibitory local interneurons in the SEZ.

Figure 3 with 1 supplement see all
Download asset Open asset
Janu-GABA SEZ local interneurons specifically promote water-seeking behaviors.

(A–D) Janu-GABA1 interneuron morphology in the SEZ of a Janu>multicolor flip-out brain, compressed z-stack. Arrowhead in (A) points to the cell body. Boxed area is enlarged in (B–D), showing colocalization of the Janu-GABA1 neuron (anti-FLAG) with immunoreactivity for VGAT (arrowheads in B–D). Single 1 μm confocal section. (E–H) Janu-GABA2 interneuron morphology in the SEZ of a Janu>multicolor flip-out brain, compressed z-stack. Arrowhead in (E) points to the cell body. Boxed area is enlarged in (F–H), showing colocalization of the Janu-GABA2 neuron (anti-FLAG) with immunoreactivity for VGAT (arrowheads in F–H). Single 1 μm confocal section. (I) VGAT knockdown in Janu neurons decreases thirsty water seeking. (J) Gad1 knockdown in Janu neurons decreases thirsty water seeking. (K) Simultaneous TrpA1 neuronal activation and Gad1 knockdown in R65D05 neurons blocked R65D05>TrpA1 increased water seeking in replete flies. (L) No effect on water consumption time with Gad1 knockdown in Janu neurons in thirsty flies. (M) Decreased time to termination of first drinking bout with Gad1 knockdown in Janu neurons in thirsty flies. (N) No effect on feeding behavior with Gad1 knockdown in Janu neurons. Statistics are one-way ANOVA/Tukey’s or Kruskal-Wallis/Dunn’s.

AstA neurons reciprocally regulate water seeking and feeding behavior

The Janu-AstA neuron is a bilaterally symmetric interneuron that projects ipsilaterally from the SEZ to the ventral medial SMP, and that expresses the neuropeptide AstA (Figure 4A–G; Figure 4—figure supplement 1A, B). AstA knockdown in Janu-spGal4 neurons with either of two independent RNAi transgenes reduced thirsty water seeking (Figure 4H,I; Figure 4—figure supplement 1C–E). Moreover, loss-of-function AstA mutant flies showed markedly reduced water seeking, both as a homozygote and when in trans to a small chromosomal deletion at the AstA locus (Figure 4J,K). To ask if Janu-AstA neurons were responsible for promoting water seeking when Janu neurons were activated in water replete flies, we knocked down AstA while simultaneously activating the Janu neurons with TrpA1. This manipulation blocked activation-dependent water seeking (Figure 4L; Figure 4—figure supplement 1F). Thus, like Janu-GABA neurons, the Janu-AstA neurons promote thirsty water seeking and are capable of driving seeking in water replete flies. This finding suggests that there may exist a functional relationship between the Janu-GABA and the Janu-AstA neurons. However, there are key functional differences between the Janu-GABA and Janu-AstA neurons. AstA knockdown in Janu neurons did not affect total water consumption time, the length of the first intake bout, pharyngeal pumping rate, or intake measured in the open-field assay (Figure 4M,N; Figure 4—figure supplement 1G–I). Thus, Janu-AstA neurons are required for thirsty water seeking, and they promote seeking when activated in water replete flies, but they have no role in water intake. Interestingly, AstA in the Janu neurons inhibits feeding behavior in both well-fed and hungry flies (Figure 4O; Figure 4—figure supplement 1J). Thus, the Janu-AstA neurons oppositely regulate water seeking and feeding behavior. This is in contrast to the Janu-GABA neurons that are specific to water-seeking behavior.

Figure 4 with 1 supplement see all
Download asset Open asset
The Janu-AstA neuron reciprocally regulates thirst and feeding behaviors.

(A–G) Janu-AstA neuron morphology and AstA expression. (A) Janu-AstA neuron morphology obtained from a multicolor flip-out brain. Green: anti-FLAG; magenta: anti-VGAT. Arrowhead points to cell body. Process at bottom left is from a co-labeled Janu-GABA neuron. (B–D) The Janu-AstA neuron cell body is immunoreactive for AstA. Arrowheads points to the cell body. (E–G) Colocalization of AstA immunoreactivity with Janu-AstA presynaptic endings in the medial SMP. Arrowheads point to a subset of overlapping expression. Single 1 μm confocal section. (H, I) AstA knockdown in Janu neurons reduced thirsty water seeking and open water occupancy. (J, K) Flies lacking AstA showed reduced thirsty water seeking and open water occupancy. (K) t-test. (L) Simultaneous TrpA1 neuronal activation and AstA knockdown in Janu neurons blocked increased water seeking in replete flies. (M) No effect on water consumption time with AstA knockdown in Janu neurons in thirsty flies. (N) No effect on time to termination of first drinking bout with AstA knockdown in Janu neurons in thirsty flies. (O) AstA knockdown in Janu neurons increased food occupancy in well-fed and hungry flies. Statistics are one-way ANOVA/Tukey’s or Kruskal-Wallis/Dunn’s, unless indicated otherwise.

Neural activity and motivational properties of Janu neurons

To test if thirst and hunger regulate the activity of the Janu neurons, we used the TRIC genetic calcium indicator that accumulates GFP in active neurons (Gao et al., 2015). The Janu-GABA1 neurons were selectively activated in thirsty flies; we detected no thirst-dependent activity in any other Janu neurons (Figure 5A,B). Moreover, hunger did not change the activity of any Janu neuron (Figure 5C). These data indicate that the Janu-GABA1 neurons selectively increase activity with increasing water deprivation.

Figure 5 with 1 supplement see all
Download asset Open asset
Thirst regulation of Janu neuron activity, and positive valuation of optogenetic activation of Janu.

(A) Examples of TRIC detection of neuronal activity (green) and the pattern of cells expressing the calcium indicator (magenta) under water replete and thirsty conditions in Janu neurons, detected with anti-GFP (green) and dsRed (magenta) antibodies. (B, C) Relative intensity of TRIC signal (green in A) to pattern (magenta in A) for the individual Janu neurons in replete and thirsty (B), and well-fed and hungry (C) flies. Multiple adjusted t-test, with FDR = 1%. (D) Positional preference of Janu>Chrimson (red) and Janu>+ (blue) water replete and well-fed flies that were fed all-trans retinal (ATR). Positive values indicate attraction to LED-ON half of the rectangular arena, controlled for arena side bias. (E) Quantification of D. Activation effect is the average positional preference for a 1 min bin at 4 min (activation lights off) subtracted from 9 min (activating LEDs on) (see also Figure 5—figure supplement 1A). One-way ANOVA/Tukey's.

Finally, we asked if Janu neuron activity imparts appetitive or aversive value to the fly. We measured positional preference in an open-field arena where the light-gated ion channel Chrimson, expressed in Janu neurons, could be selectively activated in either half of an open-field arena. Janu>Chrimson flies that were well fed and water replete were attracted to the activation side of the arena (Figure 5D,E; Figure 5—figure supplement 1A). Flies showed a similar preference for Chrimson activation of the R65D05 pattern that also contains the Janu water seeking neurons (Figure 5—figure supplement 1B,C). These findings suggest that water seeking driven by Janu neurons is rewarding. Moving up a humidity gradient when thirsty may imply impending repletion to the fly, and an end to the negative valence state of thirst (Betley et al., 2015).

Janu-AstA neurons are anatomically and functionally upstream of AstA-R2-expressing NPF neurons

NPF promotes feeding behaviors in Drosophila (Shao et al., 2017; Shohat-Ophir et al., 2012; Wu et al., 2003). Janu-AstA projects to the SMP, where NPF neurons elaborate synaptic endings. Janu-AstA presynaptic outputs partially overlapped with NPF in the SMP (Figure 6A,B). Moreover, synaptic GRASP (GFP Reconstitution Across Synaptic Partners) revealed that Janu-AstA neurons are presynaptic to NPF neurons (Figure 6C; Macpherson et al., 2015). To test for signaling to NPF neurons by AstA, we reduced expression of the two AstA receptors AstA-R1 and AstA-R2 in NPF neurons using previously characterized RNAi transgenes (Yamagata et al., 2016). AstA-R2, but not AstA-R1, knockdown decreased thirsty water seeking, and also increased feeding behavior, similar to AstA function in the Janu-AstA neurons (Figure 6D; Figure 6—figure supplement 1A). AstA-R2 knockdown in NPF neurons, however, reduced both total water consumption time and the duration of the first intake bout, indicating that AstA signaling to NPF neurons affects both water seeking and water ingestion (Figure 6E,F; Figure 6—figure supplement 1B). Consistent with AstA-R2 functioning as an inhibitory receptor in NPF neurons, we observed increased water seeking in flies completely lacking NPF and in flies with NPF knocked down in NPF neurons (Figure 6G,H). Measurement of neuronal activity in NPF neurons indicated that only two NPF neurons, the P1 neurons and the L1-L neurons, showed any activity, and they both increased activity in thirsty flies (Figure 6I; Figure 6—figure supplement 1C). These findings indicate that NPF neurons respond to thirst and that NPF neurons receive AstA signals via AstA-R2 to reciprocally regulate thirst and hunger behaviors. Figure 7 summarizes our current understanding of the thirst circuitry and its relation to feeding behavior.

Figure 6 with 1 supplement see all
Download asset Open asset
NPF neurons are postsynaptic to Janu-AstA neurons via the AstA-R2 receptor, and NPF inhibits water seeking.

(A) Janu-AstA presynaptic endings in boxed area imaged for immunohistochemistry. (B) Janu-AstA presynaptic endings partially overlap with NPF immunolabeling. Single 1 μm confocal section. Yellow arrowheads point to overlap. (C) Synaptic GRASP (syb:GRASP) between R65D05-Gal4>UAS-syb:spGFP1–10 and NPF-LexA>LexAOP-spGFP11 in thirsty flies. Single 1 μm confocal section. Yellow arrowheads point to GRASP GFP signal that overlaps with AstA immunoreactivity in the superior medial protocerebrum. (D) Behavioral effects of AstA-R2 RNAi in NPF neurons. (E) AstA-R2 knockdown in NPF neurons decreased total water consumption time. (F) AstA-R2 knockdown in NPF neurons decreased the length of the first water intake bout. (G) Flies mutant for NPF show increased water seeking when mildly water deprived (4 hr). Mann-Whitney t-test. (H) NPF knockdown in NPF neurons increased water seeking in mildly water deprived flies (8 hr). (I) NPF neuron activity in thirsty flies with TRIC. Activity was detected only in P1 and L1-l NPF neurons. t-test. Statistics are one-way ANOVA/Tukey’s or Kruskal/Wallis/Dunn's, unless indicated otherwise.

Figure 7
Download asset Open asset
Thirst interneuron circuitry in Drosophila.

(A) Functional relationship for thirst and feeding behavior. The dotted lines represent putative connectivity between interoceptive sensory neurons (ISN) and targets neuropeptide F (NPF) and Janu neurons. (B) Functions of GABA, AstA, and NPF in their respective thirst interneurons.

Discussion

Thirst is an internal state encoded by discrete neural circuits that generate innate goal-directed behavioral actions. We used an unbiased screen followed by genetic techniques to discover central brain interneurons that regulate specific aspects of thirsty water seeking, and that contribute to a specific motivational state. While sensory and interoceptive inputs are known, there is little understanding of how the brain forms the representation of thirst and coordinates sequences of behaviors to achieve water repletion. The Janu-GABA water seeking and Janu-AstA to NPF water seeking and feeding behavior circuit elements define specific functional steps of thirst representation, and they provide a means to help understand the interactions of different need states.

The Janu-GABA neurons are GABAergic local interneurons whose activity is increased by thirst and that promote thirsty water seeking. GABA in Janu-GABA is sufficient for driving water seeking in the water replete state, and it is necessary for thirsty seeking. The role of the Janu-GABA neurons appears to be limited to the seeking step of thirst behavior, since thirsty water consumption is largely unaffected by their inhibition. The specific role for Janu-GABA in promoting persistence during the first thirsty intake bout may reflect the end of the seeking component of the sequence of thirst behaviors that may overlap with the beginning of the intake phase. Moreover, the Janu-GABA neurons did not affect feeding behavior and are not activated by hunger. Thus, the Janu-GABA neurons increase thirsty seeking to sources of water. The Janu-GABA neurons are functionally distinct from the Fdg local interneurons that are also located in the SEZ (Flood et al., 2013; Pool et al., 2014). The Fdg neurons increase intake of liquid food and promote proboscis extension, behavioral functions we did not observe for the Janu-GABA neurons. We showed that the Fdg neurons may also promote water seeking. However, we did not identify the specific neurons in the NP883-Gal4 pattern that are responsible for water seeking. Defining the source of the signal to drive Janu-GABA dependent thirsty seeking and the mechanism of thirsty seeking require further elaboration of the circuitry for thirst and other behaviors in the SEZ region of the Drosophila brain. We speculate that the Janu-GABA neurons may function upstream or downstream of the osmosensory ISN neurons that, when inhibited, promote water intake (Jourjine et al., 2016). Janu-GABA neuron activation in the SEZ is predicted to lead to inhibition of SEZ neurons, perhaps to repress competing behavioral states represented in the SEZ. Alternatively, the Janu-GABA neurons may repress repressor neurons of thirsty seeking, that remain to be identified.

Janu-AstA is a bilaterally symmetric AstAergic projection interneuron with inputs in the SEZ and outputs in the medial SMP. AstA in the Janu-AstA neuron promotes thirsty water seeking, and it inhibits feeding behavior in a hunger-independent manner. For thirst, AstA in the Janu-AstA neuron was specific to thirsty seeking with no impact on water intake, indicating a functional role similar to Janu-GABA for thirst. However, we did not detect changes in Janu-AstA neuronal activity with either thirst or hunger. This lack of response may indicate a tonic role for AstA in feeding and thirst. Tonic release of neuropeptides can tune neural circuit function, including the peptidergic modulation of hunger (Fu et al., 2004; Tu et al., 2005). Peptidergic signaling can also be tuned by changing the amount of the receptor in the receiving neuron, as was demonstrated for sensory tuning by the short-NPF (sNPF) neuropeptide (Root et al., 2011). Interestingly, knockdown of either AstA or GABA in the Janu neurons blocked water seeking evoked by Janu activation. Janu-GABA and Janu-AstA may work in series in the same circuit, or in parallel in separate circuits that converge, to control thirsty-seeking behaviors.

Whereas Janu-GABA and Janu-AstA may work together for thirsty seeking, Janu-AstA also regulates feeding behavior, indicating branched functionality in water seeking circuitry. Janu-AstA transmits thirst and hunger information to NPF neurons through the AstA-R2 receptor on NPF neurons, where NPF coordinately regulates thirsty seeking and feeding behavior. NPF is well studied in Drosophila and in vertebrates as a promoter of feeding behavior – our work indicates a reciprocal role for NPF in thirst behavior (Chen et al., 2019; Luquet et al., 2005; Pool and Scott, 2014; Shen and Cai, 2001). Coordinated regulation of thirst and hunger is most clearly manifested behaviorally as prandial thirst (thirst induced by feeding) and as dehydration anorexia (thirst suppression of feeding) (Fitzsimons, 1972; Zimmerman et al., 2017). Janu-AstA and AstA-R2-positive NPF neurons join a small list of neurons that oppositely regulate thirst and hunger behaviors: the osmosensory ISN neurons oppositely regulate food and water intake, and Ion Transport Peptide (ITP) neurons promote thirst and inhibit food intake (Gáliková et al., 2018; Pool et al., 2014). It remains unclear if all these neurons support water seeking, or if they function specifically during ingestion.

Interestingly, two NPF neurons, the P1 and L1-l neurons, showed increased activity in thirsty flies. Hunger also selectively increases the activity of these two NPF neurons (Beshel and Zhong, 2013). This might indicate that the P1 and L1-l represent an internal state imbalance rather than a specific internal state like hunger or thirst. Additionally, the direction of their regulation – increased activity when thirsty – is opposite of what is expected for NPF in these neurons to inhibit thirsty seeking. Thus, we suggest that Janu-AstA may synapse onto different NPF neurons that, like Janu-AstA, do not exhibit state-dependent changes in neuronal activity, but that promote thirst and inhibits feeding behavior. A potential mechanism may be through NPF-NPF neuronal synapses: both the P1 and the L1-l neurons express the NPF receptor NPFR (Shao et al., 2017). Moreover, it is likely that one or more of the water-seeking circuit elements we discovered feeds into navigational decision making circuitry to direct hygrotaxis toward a water source and to suppress other goal directed behavior in a dynamic fashion (Hulse et al., 2020; Lyu et al., 2020).

Optogenetic activation of the Janu or the R65D05 neurons induces place preference, indicating that animals ‘like’ activation of these neurons. Thirst and hunger are deprivation states that have a negative valence, therefore the Janu neurons likely do not encode the state of deprivation (Burnett et al., 2016; Leib et al., 2017; Sutton and Krashes, 2020). We propose that, instead, they encode the anticipation of terminating a thirsty state. To a thirsty fly, the presence of a humidity gradient likely predicts impending water intake and repletion, which may have rewarding value.

Materials and methods

Strains and culturing

Request a detailed protocol

All strains were outcrossed for five generations to the Berlin genetic background. Flies were raised on standard food containing agar (1.2% w/v), cornmeal (6.75% w/v), molasses (9% v/v), and yeast (1.7% w/v) at 25°C and 60% humidity in a 16:8 light:dark cycle. For experiments with UAS-Shibirets and UAS-TrpA1, flies were reared and held at 18–23°C prior to testing. For split-Gal4, RNAi, and UAS-STOP-effector/FLP experiments, adult F1 progeny were held at 29°C for 2–3 days prior to testing to enhance transgene activity. Supplementary file 1 provides a list of the InSITE-Gal4 strains screened by neuronal activation for feeding behavior, and their seeking index (fraction on food, InSITE-Gal4>TrpA1 – InSITE-Gal4>+) (Gohl et al., 2011). Durstig-Gal4 is an InSITE-Gal4 enhancer trap inserted into the first intron of CG4502 on chromosome 2. All strains used in this study are described in Supplementary file 2.

Behavioral measurements

Request a detailed protocol

Genotype and treatment were blinded for all behavioral counts. Groups of approximately 20 adult males were collected 1–2 days prior to the experiment. One group is an n = 1. Matched experimental and control groups were tested within day, and experiments were performed across days, to account for between-day variability. For water deprivation, flies were placed in empty culture vials containing 1 mL 5% sucrose that was dried onto a 2.5 × 15 cm strip of Whatman 3 MM filter paper. For food deprivation, flies were placed into empty culture vials containing water saturated 3 MM filter paper. Dry, rough cut cane sugar (La Perruche Pure Cane Rough Cut Cubes) was used in all food seeking experiments, except for the original screen that used standard fly food. Sources of food or water were presented to flies either on a small square of Parafilm or in the cap of a 0.65 mL microcentrifuge tube. Water was made inaccessible by gluing a 210–300 μm mesh grid onto a cap. Flies were deprived of food or water for the times indicated in Supplementary file 3. Thin-walled Plexiglas behavioral chambers were designed with two side-by-side arenas, each arena measuring 45 × 75×10 mm. Chambers were designed and built by IO Rodeo (Pasadena, CA); design files are available from the company. The chambers and sources were acclimated to the testing temperature in a Peltier incubator, and flies were acclimated into the chambers for 10 min prior to introducing the source (IN45, Torrey Pines Scientific). For TrpA1 and Shibire thermogenetics, flies were temperature acclimated for 20 min before being introduced into the chamber. Humidity was measured using an EK-H4 multiplexer with SHT71 sensors (Sensirion). The incubator maintained 30–40% relative humidity. For thermogenetic activation experiments with UAS-STOP-TrpA1-mCherry, flies were pre-acclimated for 20 min in food or deprivation vials, then 10 min in the behavioral arena, as previously reported (Asahina et al., 2014). Flies were filmed from above with Logitech C510 web cameras at 10 fps with the arena placed on white light LED panel (Edmund Optics). The number of flies on the source was divided by total number of flies to determine the fraction of flies on the source. In binary choice experiments, dry sucrose and water were placed in closely apposed receptacles just prior to performing the experiment.

Locomotor speed was calculated per group as the total distance traveled divided by the total time each individual was tracked for 10 s intervals every 5 min, and then averaged across the 10 min and 15 min time points, allowing for acclimation to the arena (Wolf et al., 2002).

To measure intake in the open-field arena, water was made to 0.2% w/v erioglaucine and presented to flies for 20 min in the behavioral chamber. Consumption was determined spectrophotometrically in cleared fly extracts at 628 nm by a BioTek Epoch two microplate reader, background subtracted with flies fed undyed water, and was normalized to the enhancer-Gal4>+ control strain.

Intake time and pumping rate were measured essentially as previously described (Manzo et al., 2012). Flies were glued to slides with clear nail polish while cold anesthetized and allowed to recover at 60% humidity and 25°C for 1–2 hr. Consumption time was measured by repeatedly presenting the proboscis with a water drop extruded from the end of a 5 μL calibrated micropipet (Avantor), until flies failed to extend the proboscis to four sequential presentations. The first bout length was the time from first engagement to first complete disengagement with the water drop. Pumping rate was determined with manual counts of video recordings of pharyngeal contractions during the initial 10 s of engagement, or shorter if the flies disengaged earlier.

We used a custom-built device for determining the positional preference of optogenetically stimulated flies (Flidea). Groups of about 10 flies were kept in the dark on standard food containing 35 mM all-trans retinal for 2–4 days (Sigma). They were placed in an 80 × 40 × 3 mm arena with a clear roof and white acrylic floor. Optogenetic LEDs (617 nm) were arrayed 1 cm below the arena and were switched to allow illumination of either 40 × 40 mm side of the arena. Optogenetic stimulation was at 40 Hz for 200 ms every second. Flies were allowed to acclimate to the arena, and then were given a 5 min off, 5 min on, 5 min off pattern of stimulation, with the ON side randomly assigned. Filming was done at 2 fps in dim ambient light. Object recognition and position in the arena was determined using custom developed DTrack software (Ro et al., 2014). Fly positional preference was calculated as the number of flies on a side divided by the total number of flies. Experiments where flies showed a strong side bias prior to optogenetic stimulation were excluded from further analysis. The resulting positional information was averaged into 20 s bins. The activation effect is the positional preference averaged across minute nine minus across minute 4, such that a positive value indicates increased occupancy of the LED-on side when the LEDs are on. The experiment was repeated with flies from 5 to 8 separate crosses across different days.

Statistical measurements were made with Prism 8.0 (GraphPad). One-way ANOVA followed by Tukey’s post-hoc comparisons (or Bonferroni post-hoc planned comparison) were used when data did not show unequal variance by the Brown-Forsythe test, otherwise the Kruskal-Wallis test followed with Dunn’s post-hoc was used. t-tests were two-tailed (Supplementary file 3). Error bars are the SEM. Dots overlaid on the bar graphs are the value measured for individual groups. No statistical methods were used to predetermine sample sizes. Sample sizes are similar to those used in other Drosophila behavioral experimental paradigms, and they were held constant with rare exceptions. No data was excluded using statistical methods.

Immunohistochemistry

Request a detailed protocol

Adult fly brains were dissected in PBS with 0.1% Triton X-100 (PBT), fixed overnight at 4°C in PBT with 2% paraformaldehyde, blocked in PBT with 5% normal goat serum and 0.5% bovine serum albumin, and immunostained as described previously (Kong et al., 2010). Antibodies were rabbit anti-GFP (1:1000, Life Technologies), chicken anti-GFP (1:1000, AbCam), mouse anti-GFP for GRASP (1:100, Sigma), rabbit anti-dsRed (1:500, Takara Bio), mouse anti-Bruchpilot (1:25, Developmental Studies Hybridoma Bank, Iowa), rabbit anti-VGAT (1:200, gift from David Krantz, UCLA), rabbit anti-AstA (1:2000, Jena Bioscience), rabbit anti-NPF (1:100, RayBiotech), and rat anti-FLAG (1:200, Novus Biologicals). Samples were mounted in Vectashield (Vector Laboratories) and imaged on a Zeiss LSM-880 confocal microscope. Image stacks were processed in Fiji, and brightness and contrast were adjusted in Photoshop CC 2021 (Adobe).

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files.

References

Decision letter

  1. Mani Ramaswami
    Reviewing Editor; Trinity College Dublin, Ireland
  2. K VijayRaghavan
    Senior Editor; National Centre for Biological Sciences, Tata Institute of Fundamental Research, India

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

The work identifies multiple types of "Janu" neurons in Drosophila with specific anatomical and neurochemical properties that control the thirst reponse as well as other unique (but related) behaviors. The Janu-AstA neuron is argued to synapse with and signal to NPF neurons, and evidence is provided supporting the role of this neuronal pathway in water-seeking. These interesting set of neurons constitute a valuable resource for further understanding of thirst-induced behaviours.

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting your work entitled "Allatostatin A thirst interneurons promote water seeking and limit feeding behavior in Drosophila" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by Mani Ramaswami as Reviewing Editor and K. VIjayRaghavan as Senior Editor.

All three reviewers expressed strong interest and enthusiasm for the work. However, as you will read, there was also a consensus that the considerable amount of extra work on presentation and content necessary before potential acceptance would require more than the two months allowed for revisions to a paper in eLife. For this and other reasons listed below, after extensive discussion and consultation among the reviewers, we regret to inform you that this paper will not be considered further for publication in eLife. The reviewers however, are supportive enough to suggest that, should you choose to do so, then eLife will remain willing to examine a carefully revised and resubmitted manuscript as a new submission to the journal.

A single consensus review is provide below, in addition to the individual reviews, and we hope you will find these to be useful.

Summary:

Landayan, Zhou, and Wolf present a new behavioral assay to test different aspects of water/food seeking behaviors in Drosophila. From an initial activation screen of 154 Gal4 enhancer trap flies, the work identifies several that mediate food seeking, in particular Durstig-Gal4, where activation of marked neurons causes shows the highest attraction and occupancy level toward food. Significantly, Durstig-Gal4 marked neurons are involved in modulating both food and humidity seeking behaviors. From this starting point, using a vast array of neural activation/inhibition experiments, neuromodulator RNAi, and some neuroanatomy and Split Gal4s, they identify a line called Janu, which labels 4 neurons in the central brain, 3 interneurons in the ventral nervous system and 2 ascending neurons projecting from the ventral nervous system to the central brain. These Janu neurons are either GABAergic or produce allatostatin and may function as a multi-modal integrator required for water seeking (but not water intake). Thus, this subset of inhibitory and Allatostatin A-positive neurons promote thirst-driven behavior. Using synaptic GRASP, they show that Janu-Ast neurons connect to NPF expressing neurons.

Previous work on the interplay between hunger and thirst has focused on sensory circuits, so identifying central neurons mediating thirst would represent a significant contribution to understanding this process. Going from an unbiased behavior screen to a subset of neurons docked into a circuit is admirable. The work is careful, thorough, and on an important topic.

Essential revisions:

1. A major rewrite is required. While the authors may be commended for wrestling with a complex story and for being open about all of the data even when it is hard to reconcile with a simple model, in the end the insights need to be distilled more for public consumption. In current form, the paper is extraordinarily difficult to read and difficult to follow, in part due to the complexity of the methods. There is a large number of different types of behavioral assays (with flies in different internal states), and many different driver lines are presented as the authors home in on those that form the central novel findings (i.e., Janu neurons). The huge amount of data masks the main messages of the paper. There is a lot of chronological and extraneous data included that could be moved to supplementary figures to streamline and we'd strongly suggest the authors prune irrelevant experiments. For example, it jumps from characterization of the new hygrotactic assay (though not the first of this type of assay – – see minor comment below) to a screen using a completely different assay, which gives the impression the authors wanted to show the results of the screen rather than present a logical flow of experiments for the reader. In addition, inconsistent use of nomenclature in text/figures/legends renders it difficult to follow, e.g., Janu neurons are sometimes to referred by name, but often in the behavior data panels only the codes for the split-Gal4s are shown. In general, the data on thirst is stronger than that on NPF and feeding, so the thirst conclusions should be emphasized. One suggestion is to describe the original identification of Durstig neurons as thirst-promoting, then how a subset within that was identified as critical and Allatostatin-A-positive, and then demonstrate that this subset promotes thirst as a drive, suppressing the hunger drive. With a serious organizational revision and streamlining of extra data into supplementary figures, the findings should be clearer and much easier to convey. Some summary schematics could help significantly, particular accompanying the Discussion sectionDiscussion section. For example: one showing the proposed circuit, and a another showing the nested/Venn diagram of Durstig, Janu, AllatostatinA+, and GABA+ neurons, (maybe including their respective GAL4 targeting reagents).

2. One problem than runs through the manuscript is that it remains formally unclear whether the phenotypic differences observed upon activation/inhibition of populations of neurons is due to the difference in their individual roles or to the difference in the magnitude of activation/inhibition. (One might be activating a lot of neurons that have the same role in water seeking and thus the phenotype is stronger than activating only a subpopulation of those neurons). Similarly, many RNAi experiments are presented but, while several do point towards a consistent conclusion, there is no evidence that particular knock-downs work or do not work, and single mutant alleles are shown without rescue or obvious control of genetic background (for both Npf and AstA mutants, a w strain is used as the control). Some of these issues need to be addressed experimentally and certainly in the text.

3. Many of the issues above could be addressed by physiological experiments (calcium imaging or electrophysiology), which would potentially make it easier to appreciate the function of individual neurons. For example, a simple question is whether Janu-AstA and Janu-GABA neurons respond to water vapor and/or water contact? We appreciate such experiments are probably beyond the scope of a reasonable revision. But in the absence of this, extra care should be paid to both establishing the specificity of the genetic perturbations, the resulting phenotypes and in interpretation of the results. As the quantity and complexity of the work makes extraction of a compelling central discovery of broad general interest difficult to appreciate, major revisions should concentrate on marshalling and presenting all the best evidence to arguing for and documenting the identification of thirst-promoting central neurons.

4. The title and the summary are misleading, as the Janu neurons described here are not only allatostatin producing but also GABAergic. Perturbations specific to Allostatin/Janu neurons form only a relatively small part of the entire analysis, which ignores that other Janu neurons are GABAergic. This should be explicitly clarified. Similarly, in the abstract the phrase: "allowed us to uncover the neural basis.……" is a bit of an over-statement (perhaps "a neural basis").

Reviewer #1:

Landayan et al. et al. combine a simple behavioral test with an extensive use of genetic tools to identify neurons which promote water seeking and limit feeding behavior in Drosophila. Starting from an initial screen of "a" library of 154 Gal4 enhancer trap flies, they select one strain, Durstig, which shows the highest attraction and occupancy level toward food. Starting with this strain, they show that Durstig is involved in modulating both food and humidity seeking behaviors. From this starting point, they analyze which population of neurons are involved in activating Durstig neurons and which neurons are downstream. This demonstration makes an extensive use of various genetic constructions, following a logic that looks solid and supported by behavioral observations, however lacking functional or pharmacological observations. The reasoning is explicitly outlined in the manuscript but the number of constructions used make it difficult to follow to someone which is exterior to this research field.

The authors first verify that the activation of Durstig neurons involves humidity and antennae. Other stimuli related to food seeking or from other receptors are not checked. Then, they proceed to find neurons upstream and downstream of Durstig neurons. They first look at different populations of neurons known to be involved in humidity preference or in the control of liquid or food intake, to see if activating them with TrpA1 (and heat) would induce similar water seeking behaviors.

Since Durstig-gal4 neurons are numerous, the authors tried then to narrow down their analysis using intersectional approaches to find which of these neurons are involved in water-seeking, looking at cholinergic and glutaminergic neurotransmission. This analysis led them to find an enhancer, R52A01, which marks neurons using glutaminergic transmission within the Durstig population of neurons, as well as R65D05 neurons which are involved in inhibiting food intake.

Using a split-Gal4 approach, they could tag neurons (called Janu), namely 4 neurons in the central brain, 3 interneurons in the ventral nervous system and 2 ascending neurons projecting from the ventral nervous system to the central brain.

Using antibodies directed against neurotransmitters and neuropeptides, they show that Janu neurons are either GABAergic or produce allatostatin and that these neurons promote water seeking (but does not change water consumption).

Lastly, using synaptic GRASP, they shown that Janu-Ast neurons connect to NPF neurons.

All in all, this paper sums out an impressive amount of work which leads to the identification and description of neurons involved in water-seeking behavior. To this respect, it constitutes a beautiful analysis. It is a paper difficult to read for me because of the complexity of the methods. I find the title and the summary a bit misleading, as the Janu neurons described here (unless I am mistaken) are not only allatostatin producing but also GABAergic. From the summary and the title, I was also expecting functional studies related to allatostatin.

Reviewer #2:

This work presents a new behavioral assay to test different aspects of water/food seeking behaviors in Drosophila, and presents a vast array of neural activation/inhibition experiments, neuromodulator RNAi, and some neuroanatomy to characterize several classes of central interneurons that have different roles in state-dependent water-seeking behavior; these include the Janu-AstA neurons, which seem to function as a multi-modal integrator required for water seeking (but not water intake). The topic is certainly interesting and there are several novel findings. However, the paper lacks any accompanying neurophysiological analyses or careful control that the neuronal/molecular manipulations have the intended/expected effects, which limits the strength of the conclusions. Moreover, the sheer quantity and complexity of the work makes extraction of a compelling central discovery of broad general interest difficult to appreciate at this stage.

1. The manuscript is very difficult to follow: there is a large number of different types of behavioral assays (with flies in different internal states), and many different driver lines are presented as the authors home in on those that form the central novel findings (i.e., Janu neurons). The huge amount of data masks the main messages of the paper, and we'd strongly suggest the authors prune irrelevant experiments. For example, it jumps from characterization of the new hygrotactic assay (though not the first of this type of assay – – see minor comment below) to a screen using a completely different assay, which gives the impression the authors wanted to show the results of the screen rather than present a logical flow of experiments for the reader. In addition, inconsistent use of nomenclature in text/figures/legends renders it difficult to follow, e.g., Janu neurons are sometimes to referred by name, but often in the behavior data panels only the codes for the split-Gal4s are shown. Some summary schematics could help significantly, particular accompanying the Discussion sectionDiscussion section.

2. The authors infer roles for specific neurons and neural signaling molecules using a variety of genetic tools, but there is no molecular, cellular or physiological level validation that links these manipulations to changes in neuronal function. This is important, because it remains unclear whether the phenotypic differences observed upon activation/inhibition of a population of neurons is due to the difference in their individual roles or to the difference in the magnitude of activation/inhibition. (One might be activating a lot of neurons that have the same role in water seeking and thus the phenotype is stronger than activating only a subpopulation of those neurons). Many RNAi experiments are presented but, while several do point towards a consistent conclusion, there is no evidence that particular knock-downs work or do not work, and single mutant alleles are shown without rescue or obvious control of genetic background (for both Npf and AstA mutants, a w strain is used as the control).

3. No physiological experiments are presented (calcium imaging or electrophysiology), making it difficult to appreciate the function of individual neurons. For example, a simple question is whether Janu-AstA and Janu-GABA neurons respond to water vapor and/or water contact? We appreciate such experiments are probably beyond the scope of a reasonable revision but are part of the expected standard in the field now to make claims that the "neural basis" of a particular process is understood.

Reviewer #3:

Landayan, Zhou, and Wolf have identified a population of neurons that affect the relative drive strengths of thirst and hunger, and thus the choice between consummatory behaviors. Subsets of inhibitory and Allatostatin A-positive neurons seem to promote thirst.

Previous work on the interplay between hunger and thirst has focused on sensory circuits, so implicating central neurons, as proposed here, would be a significant contribution to understanding this process. Going from an unbiased behavior screen to a subset of neurons docked into a circuit is admirable. The work is careful, thorough, and on an important topic, so overall, I support publication, but the presentation needs some revisions to make the most supportable findings clearer to the reader.

The data order in the text and the figures do not align well, making it difficult to extract the conclusions and the evidence for them. There is a lot of chronological and extraneous data included that could be moved to supplementary figures to streamline. In general, the data on thirst is stronger than that on NPF and feeding, so the thirst conclusions should be emphasized.

The paper would benefit from a few schematics: one showing the proposed circuit, and a another showing the nested/Venn diagram of Durstig, Janu, AllatostatinA+, and GABA+ neurons, (maybe including their respective GAL4 targeting reagents).

The authors should be commended for wrestling with a complex story and for being open about all of the data even when it is hard to reconcile with a simple model, but in the end the insights need to be distilled more for public consumption.

One suggestion is to describe the original identification of Durstig neurons as thirst-promoting, then how a subset within that was identified as critical and Allatostatin-A-positive, and then demonstrate that this subset promotes thirst as a drive, suppressing the hunger drive. With a serious organizational revision and streamlining of extra data into supplementary figures, the findings should be clearer and much easier to convey.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Thirst interneurons that promote water seeking and limit feeding behavior in Drosophila" for further consideration by eLife. Your revised article has been evaluated by a Senior Editor and a Reviewing Editor.

All of reviewers appreciated the quality of the data, the potential interest of the work as well as the amount of work undertaken by your group, but they also concurred with regret that the data remained confusing and that conclusions reached would need to be supported by significantly more experimental work, some of which would not likely not be feasible with current tools. Their individual comments are detailed below.

Reviewer #1:

I remain ambivalent about this manuscript. On the one hand, the authors have responded to most of our requests on the original version with additional experiments and reformatting. On the other, it is still a complex story to navigate, and is still not very conclusive. I sense the authors have gone as far as is practically possible in this project with the resources available. While they have identified a potentially very interesting set of neurons, thirst seeking is clearly an intricate behavior to dissect using neurogenetics. The work opens a number of interesting questions, but I feel the level of evidence here is lower than similar types of papers in this field here. In addition, because I don't think they can make extremely strong claims about the biological role of these neurons and the circuit in which they are embedded, the breadth of interest beyond the immediate field is still relatively limited.

The authors have identified multiple types of "Janu" neurons with specific anatomical and neurochemical properties (e.g., GABA, AstA). These neurons all control thirst seeking, and also other unique (but related) behaviors. The Janu-AstA neuron is argued to synapse with and signal to NPF neurons. Focusing in on the key data in support of these main claims, I am only mildly convinced. The strongest phenotypes they observe are always with the broad drivers (e.g. Durstig-Gal4, which has widespread expression through the brain), but once they home in on a split-Gal4 combination with very selective labeling, the phenotype effects of activation or suppression are very small (e.g. Figure 2J-K or the new optogenetic activation experiments in Figure 5E). For some experiments, they revert to using the broader driver, such as the GRASP with NPF neurons using R65D05-Gal4. Unfortunately, this Gal4 line is expressed in a mass of neurons in this region of the brain (Figure 2S1A), so it's not clear to me that a positive GRASP signal really reflects synaptic contact between Janu-AstA neurons (somewhere within the R65D05 pattern) and NPF neurons. As mentioned in the previous round of reviews, the baseline phenotypes of controls is variable (e.g. in Figure 3I and Figure 3S1A VGAT RNAi experiments, there are many fewer flies on H2O compared to those with Gad1 RNAi in Figure 3J). The authors somewhat dismissed this concern by arguing that this may reflect variable environmental conditions and water deprivation treatments of flies of experiments on different days (and they only compared assays run in parallel), but this remains a concern for me. The n of several assays seems rather low (6-7), so it's not clear if they did all experiments of one type on a single day. In any case, questions linger as to how robust some of their results are.

In response to the original request to examine the physiological response of Janu neurons to water deprivation, the authors have used the TRIC sensor, a measure of long-term changes in neuronal calcium. This is not yet an extremely widely-used reporter, and I suspect requires careful controls to be able to interpret properly. They observe that one of the Janu-GABA neurons, but not any of the other three Janu neurons, displays increased TRIC signal upon water deprivation. This is interesting, but I find it puzzling that it should so selective, given that all of these neurons are behaviorally implicated in thirst seeking. I'd be interested to know what TRIC signal would be observed with a broad driver (e.g. Durstig-Gal4); is the single Janu-GABA neuron the only one that would have a significant increase in TRIC fluorescence? Water deprivation presumably induces widespread physiological changes in the brain, some of which may be direct sensors, and others show indirect activity. Ultimately, I don't find the new data so compelling in revealing this neuron to be a special sensor of thirst. (I appreciate more acute measurements of physiological activity are beyond the scope of the current work).

Although the authors have streamlined the figures, the reader is still directed to the many ancillary experiments in supplementary figures, now often with little explanation (e.g. Figure 1S1A-B, are only obliquely referred to in the text as "Neuronal activation and inactivation experiments uncovered additional behavioral roles for neurons in the Durstig pattern", and I don't fully understand the results in these panels). There is still an awkward transition at line 99 from the behavioral characterization of thirst-seeking to the screening for neuron populations promoting "seeking behavior"; ultimately, I find the Durstig-Gal4 neuron data almost impossible to interpret given its widespread expression pattern. On line 127 they say that "subtracting R65D05 neurons from Durstig reduced activation-dependent water-seeking, indicating that the water seeking neurons in Durstig and R65D05 are the same"; this is true, but there is still substantial water-seeking behavior implying there are other seeking-promoting neurons in the Durstig pattern. Are these Janu-GABA neurons or others not characterized here?

As this is a review of a revised manuscript, and as I believe the authors have gone as far as they can in the current revision, I do not propose further experimentation, as I believe a decision should be made based solely on the data now available.

Reviewer #2:

In this paper, Landayan et al. characterize two pairs of Drosophila neurons (called Janu) involved in orientation behaviors toward water following water deprivation (called thirst here). One pair of these neurons is GABAergic (Janu-GABA), while the other is producing allatostatin A (Janu-AstA). The anatomy, projections and role in the behavior of these neurons is dissected in detail using a host of approaches and genetic intersectional approaches. By manipulating the activity of either of these neurons, the authors show that Janu neurons motivate flies to seek humidity but that they are not promoting drinking. They also confirm earlier studies showing that motivation to search for humidity are antagonist to food searching. These results are new, interesting, and also puzzling because they indicate that the activation of only a few interneurons has a profound influence on complex behaviors, the authors suggesting that these neurons are linked to motivational states (like thirst).

While the experiments and the results reported here seem sound, the paper is difficult to read for those unfamiliar to the field. The starting point of this study is the use of a simple behavioral test to evaluate the aggregation of flies next to food or water to screen a library of Gal enhancer trap flies (InSite), while activating the corresponding cells (using TrpA1 and heat). They find one enhancer-GAL4 mutant called Durstig (thirst in German), which is particularly attracted by food and by humidity when the corresponding cells are activated. From the description of this behavior (ms lines 99-110), it is not clear if Durstig neurons are only involved in the attraction toward humidity or also toward food odors.

In a first step of their research, the authors further used this behavioral test to evaluate if other genetic constructions could reproduce the behavioral pattern observed in Durstig> TrpA1 flies. Only two of these constructions reproduced the water-seeking pattern of Durstig flies. One is described as the "NP883 pattern", but this pattern has not been further compared to Durstig neurons. The second is called "R65D05 pattern" which was shown here as including Durstig neurons responsible for humidity searching.

In order to narrow down the Durstig neurons responsible for thirst-related attraction behavior, they used Gal80 constructions to subtract out neurons with various neurotransmitters. Finally, the authors were able to narrow down the neurons responsible to the attraction to 2 pairs of neurons characterized / defined by a split-Gal4 construction called Janu (again meaning thirst but this time in Estonian). According to figure 2G, this construction combines R52A01-AD and D65D05-DBD. It is thus not directly related to Durstig, except that activating the Janu cells induces the same behavioral phenotype related to humidity attraction and thirst.

What is next described in the paper is a thorough analysis of the Janu neurons, ie their anatomical location (in the adults), and the finding that two are GABAergic, while the two others are expressing allatostatin A. They also characterize their function and role in the behaviors studied here.

As compared to the first version, the ms is much improved and easier to understand. Besides the factual results shown here, it represents a very interesting work illustrating top-notch genetic strategies to identify neurons responsible for specific behaviors.

A general sketch of the Janu neurons, summarizing their role in thirst, and their relation to feeding, etc, might be useful.

Reviewer #3:

The authors use an unbiased screen for neurons that contribute to feeding behavior to identify new circuits that affect thirst. Some of these neurons are GABAergic, while others release the neuropeptide Allatostatin, and they connect to the NPF-positive neurons previously implicated in hunger sensation. The interplay between thirst and hunger is complex and a better understanding of how these neurons interact with sensory and motor components will help illuminate how flies balance these conflicting motivational drives.

The authors have clearly made a good faith effort to address the issues raised in the initial review of this manuscript.

There is an enormous amount of work presented, and there is the potential that it reveals neurons that contribute to appropriate behavioral responses to hunger and thirst. Understanding how neural circuits function to regulate choices between essential and competing drives is a really important area of research.

But I remain very confused. There is still no summary diagram for either the (known and new) neuronal circuitry or the interaction between the motivational drives. The brief flow diagram of the response to thirst helps a bit but does not really clarify. Some neurons seem to be able to induce a motivational state (or the behavioral response associated with it) while others are partially but not completely required for an appropriate choice between eating and drinking? I think the Janu-GABA and Janu-allatostatin neurons are different populations, but I am actually not sure.

The figures are very dense and intermix feeding and drinking behaviors, activation and inactivation experiments, and different sets of neurons. The combination thirst response and place preference in figure 5, for example, makes no sense at all.

The anti-VGAT staining seems to label the entire neuropil, which makes the co-localization with some of the Janu neurons very difficult to interpret. The allatostatin colocalization is similarly confusing.

This manuscript still needs a major overhaul to make its contributions to thirst circuitry interpretable.

https://doi.org/10.7554/eLife.66286.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Essential revisions:

1. (A) A major rewrite is required. While the authors may be commended for wrestling with a complex story and for being open about all of the data even when it is hard to reconcile with a simple model, in the end the insights need to be distilled more for public consumption. In current form, the paper is extraordinarily difficult to read and difficult to follow, in part due to the complexity of the methods. There is a large number of different types of behavioral assays (with flies in different internal states), and many different driver lines are presented as the authors home in on those that form the central novel findings (i.e., Janu neurons). (B) The huge amount of data masks the main messages of the paper. There is a lot of chronological and extraneous data included that could be moved to supplementary figures to streamline and we'd strongly suggest the authors prune irrelevant experiments. For example, it jumps from characterization of the new hygrotactic assay (though not the first of this type of assay – – see minor comment below) to a screen using a completely different assay, which gives the impression the authors wanted to show the results of the screen rather than present a logical flow of experiments for the reader. In addition, inconsistent use of nomenclature in text/figures/legends renders it difficult to follow, e.g., Janu neurons are sometimes to referred by name, but often in the behavior data panels only the codes for the split-Gal4s are shown. (C) In general, the data on thirst is stronger than that on NPF and feeding, so the thirst conclusions should be emphasized. One suggestion is to describe the original identification of Durstig neurons as thirst-promoting, then how a subset within that was identified as critical and Allatostatin-A-positive, and then demonstrate that this subset promotes thirst as a drive, suppressing the hunger drive. With a serious organizational revision and streamlining of extra data into supplementary figures, the findings should be clearer and much easier to convey. Some summary schematics could help significantly, particular accompanying the Discussion section. For example: one showing the proposed circuit, and a another showing the nested/Venn diagram of Durstig, Janu, AllatostatinA+, and GABA+ neurons, (maybe including their respective GAL4 targeting reagents).

A. The manuscript is completely rewritten and completely reorganized. In addition, extraneous data is removed and new data is added as detailed below. Briefly, we de-emphasize the process of isolating the individual thirst water seeking neurons to a single figure, and we now focus on their characterization. Moreover, we completed additional experiments to make each thirst neuron characterization follow the same logical path and to be more complete. The main story presented in the original manuscript remains valid, however new data allows us to make stronger and clearer conclusions about the role of each interneuron in thirst and hunger.

B. We removed extraneous experiments, moved many others into supplemental figures, and added explanatory diagrams to help guide the reader. We think that the main figures focus clearly on thirsty water seeking. Moreover, the emphasis is not on finding the neurons but instead their characterization.

C. We emphasize the conclusions on Janu-GABA, Janu-AstA, and the relation of NPF to Janu-AstA. Our findings for NPF are now clearer with the addition of new data.

2. (A) One problem than runs through the manuscript is that it remains formally unclear whether the phenotypic differences observed upon activation/inhibition of populations of neurons is due to the difference in their individual roles or to the difference in the magnitude of activation/inhibition. (One might be activating a lot of neurons that have the same role in water seeking and thus the phenotype is stronger than activating only a subpopulation of those neurons). (B) Similarly, many RNAi experiments are presented but, while several do point towards a consistent conclusion, there is no evidence that particular knock-downs work or do not work, (C) and single mutant alleles are shown without rescue or obvious control of genetic background (for both Npf and AstA mutants, a w strain is used as the control). Some of these issues need to be addressed experimentally and certainly in the text.

A. This issue is largely resolved by focusing our characterization on individual neurons (Janu-GABA, Janu-AstA, and AstA receptor expressing NPF neurons). It is also resolved by focusing on water seeking almost exclusively, and only testing for feeding behavior at strategic points.

B. RNAi experiments were done with previously characterized RNAi transgenes (for example, AstA receptor, but true for almost every RNAi transgene used in this study, as documented in the Results and in Table S2), with distinct RNAi transgenes targeting the same gene (for example AstA), and targeting different genes that are expected to affect the process in the same way (for example Gad1 and VGAT to decrease GABA synthesis).

C. In new data, we placed the mutants (all previously characterized in the literature) in trans to small deficiencies and validated the loss-of-function behavioral effects. The mutant alleles were outcrossed to the genetic background strain prior to behavioral testing.

3. Many of the issues above could be addressed by physiological experiments (calcium imaging or electrophysiology), which would potentially make it easier to appreciate the function of individual neurons. For example, a simple question is whether Janu-AstA and Janu-GABA neurons respond to water vapor and/or water contact? We appreciate such experiments are probably beyond the scope of a reasonable revision. But in the absence of this, extra care should be paid to both establishing the specificity of the genetic perturbations, the resulting phenotypes and in interpretation of the results. As the quantity and complexity of the work makes extraction of a compelling central discovery of broad general interest difficult to appreciate, major revisions should concentrate on marshalling and presenting all the best evidence to arguing for and documenting the identification of thirst-promoting central neurons.

We used TRIC to demonstrate that both the Janu-GABA neuron and NPF neurons are activated by thirst. Neither Janu-GABA nor Janu-AstA are activated by hunger. NPF neurons are previously shown to be activated by hunger. Nor did TRIC identify any effect of water vapor or water intake on Janu neuron activity in thirsty flies. However, it is possible that TRIC may not have the appropriate sensitivity to detect these responses, and therefore we excluded this negative data. We also collaborated with the lab of Tom Clandinin at Stanford to test for GCaMP6 responses in Janu neurons with optogenetic activation of Ir68a humidity sensing neurons, but we were unable to detect any signal in either water replete or thirsty flies.

We wish to emphasize the current complexity of measuring thirst-related neuronal activity in flies, due to the effects of hemolymph osmolarity on thirst circuitry, as shown by Kristin Scott’s lab at UC Berkeley: The dissections needed for imaging impact interstitial fluid composition and so thirst neuron activity. The level of expertise needed to perform these experiments is quite high, and it is beyond the means of my lab (at an R2 university that is only 15 years old, three neurobiology labs total at the university, a graduate student-driven lab, and this project is a new direction for my lab that was funded off of my start-up funds). Through collaboration we will be able to conclusively address neuronal activity using in vivoin vivo real time activity sensors with 3-photon microscopy or similar technology, where imaging can be done with intact animals. However, at the moment this is out of our reach.

Finally, we had an optogenetic device built expressly to address if flies prefer activation of Janu neurons, and we now report that they do. This adds a new functional dimension to our findings and provides a more accurate description of Janu neuron function.

4. The title and the summary are misleading, as the Janu neurons described here are not only allatostatin producing but also GABAergic. Perturbations specific to Allostatin/Janu neurons form only a relatively small part of the entire analysis, which ignores that other Janu neurons are GABAergic. This should be explicitly clarified. Similarly, in the abstract the phrase: "allowed us to uncover the neural basis.……" is a bit of an over-statement (perhaps "a neural basis").

The title is updated to reflect our main findings. The abstract, like the rest of the manuscript, is completely rewritten.

Author response table 1 describes the most consequential of the new experiments in the manuscript.

Author response table 1
FigureQuestionApproachResultInterpretation
3KAre Janu-GABA neurons important for neural activation driven seeking?We co-expressed a GABA neurotransmitter RNAi with TrpA1 activation.Water seeking is suppressed.Janu-GABA neuron activation drives replete water seeking.
3L,M
3S1B,C		Are Janu-GABA neurons important
for water ingestion?		We used well-established behavioral assays to specifically address ingestive behaviors.Water ingestion does not change. But, the first bout of water intake decreases. Pumping rate is unchaged.The first bout of water intake is mediated by Janu-GABA neurons. Other downstream neurons may be responsible for the maintenance of persistent ingestive behaviors.
3NAre Janu-GABA neurons important for fed and hungry feeding behavior?We expressed a GABA neurotransmitter RNAi and assessed dry sucrose occupancy in fed and hungry flies.Fed and hungry dry sucrose occupancy is unaltered.Janu-GABA neurons are
specific for thirst and do not contribute to feeding behavior.
4LAre Janu-AstA neurons important for neural activation driven seeking?We co-expressed an AstA neurotransmitter RNAi with TrpA1 activation.Water seeking is suppressedJanu-AstA neuron activation drives replete water seeking.
4M,NAre Janu-AstA neurons important
for water ingestion?		We used well-established behavioral assays to specifically address ingestive behaviors.Water ingestion and first bout of water intake does not change. Pumping rate is unchanged.Janu-AstA neurons are dispensable for thirsty water ingestion.
5A,B,CAre Janu-GABA and Janu-AstA activated by states of thirst and hunger?We used TRIC, a calciumdependent fluorescent reporter, to assess neuronal activity.Only the Janu-GABA1 subtype was highly active in thirsty flies. Janu-AstA was not active in thirsty or hungry states.The Janu-GABA1 neuron is thirst activated. It may be regulated by the osmosensory ISNs.
5D,E
5S1A,B,C		Does the activity of Janu and R65D05 neurons encode positive or negative valence?We optogenetically activated Janu and R65D05 neurons to assess positional preference.Flies exhibit positional preference for activation of Janu or R65D05 neurons.Janu and R65D05 neurons encode positive valence. This is distinct from mammalian hunger and thirst neurons which encode negative valence.
6EF
6S1B		Is AstA-R2 signaling on NPF neurons important for thirsty water ingestion?We expressed an AstA-R2 RNAi in NPF neurons. We used well-established behavioral assays to specifically address ingestive behaviors.Total water consumption and first bout water consumption are decreased. Pumping rate is unaffectedAstA signaling onto NPF neurons is important for water intake.
6HIs NPF signaling important for thirsty water seeking?We expressed a NPF RNAi in thirsty flies to assess water seeking.Water seeking is increased.NPF released by NPF neurons suppresses water seeking. Thus, NPF can reciprocally modulate hunger and thirst.
6I 6S1CAre NPF neurons activated by thirst?We used TRIC, a calcium-dependent fluorescent reporter, to assess neuronal activity.Dorsomedial P1 and lateral L1 NPF neurons both are significantly activated in thirsty states.NPF neurons encode thirst and are a critical convergence node for the integration of thirst and hunger signals.

[Editors’ note: what follows is the authors’ response to the second round of review.]

Reviewer #1:

I remain ambivalent about this manuscript. On the one hand, the authors have responded to most of our requests on the original version with additional experiments and reformatting. On the other, it is still a complex story to navigate, and is still not very conclusive. I sense the authors have gone as far as is practically possible in this project with the resources available. While they have identified a potentially very interesting set of neurons, thirst seeking is clearly an intricate behavior to dissect using neurogenetics.

A comprehensive description of the thirst circuitry is an impossible task for any lab, given its multiphasic character and its complex relation to other brain functions. Moreover, there is hardly anything known about the central circuitry for thirst; our work identifies components of it. We went from zero to two well-described neurons, and two neurons that have not had any function ascribed to them previously, using unbiased techniques.

Neurogenetics is a remarkable tool for dissecting thirst behavior – our work and that of many other labs is evidence of that. Plus, we extensively used behavior, anatomy, functional characterization, and some physiological tests of neuronal activity.

The work opens a number of interesting questions, but I feel the level of evidence here is lower than similar types of papers in this field here.

Two examples of similar types of papers:

1. Perhaps the best example is characterization of ITP as an endocrine regulator of water homeostasis from the Nässel lab in PLoS Genetics (https://doi.org/10.1371/journal.pgen.1007618). Here, the authors characterized the role of ITP in a variety of behaviors related to thirst and hunger. How they differ from us is they started with a peptide previously implicated in ionic balance, and the majority of the manuscript uses a pan-organismal driver to knock down ITP and assess phenotypes. Thus, we do not yet know if ITP functions in the same or different cells (or which cells, neuronal or otherwise) for water vs. food ingestion. By contrast, we de novo identified individual neurons that regulate specific steps of thirst representation through specific transmitters and modulators. We show functional connectivity between the Janu-AstA and NPF neurons for reciprocal regulation of water seeking and feeding behavior.

2. A second example is the regulation of thirst and hunger through the ISN neurons from the Scott lab in Cell (http://dx.doi.org/10.1016/j.cell.2016.06.046). Here the authors performed a neuroanatomical screen to identify the ISN neurons that they show promote food intake and inhibit water intake in an osmosensitive manner. This paper is an exceptionally well-done characterization of a single set of neurons, yet they describe no circuit connectivity, no neurotransmitter/neuromodulator for the ISNs, and test a very narrow range of behavioral characterization (food and water intake). Thus, while it is difficult to find fault with the research presented, it is also very limited in the scope for understanding their function.

At its core, our work is an example of how classic neurogenetic manipulations can provide a solid foothold for the scientific community, here to understand how thirst is encoded in the central brain, and its interactions with hunger.

To summarize, we showed:

1. that AstA in Janu-AstA is necessary and sufficient for water seeking, that it is specific for the seeking step of the thirst behavioral sequence, that it reciprocally suppresses feeding behavior, that it synapses onto NPF neurons, that the AstA-R2 receptor functions in NPF neurons for thirsty seeking and reciprocally feeding behavior, that NPF in NPF neurons regulates thirsty seeking, and that thirst increases the activity of specific NPF neurons.

2. that Janu-GABA neurons are activated by thirst, and that GABA in Janu-GABA is highly specific to thirsty water seeking.

We note that the neuromodulators AstA and NPF were previously known to regulate feeding behaviors but not thirst.

We make a guess that the reviewer is noting the lack of state-of-the-art electrophysiology and in vivo GCaMP imaging in awake behaving animals. However, we believe thirst circuitry physiology, in particular, awaits better establishment of non-invasive imaging techniques like 3-photon, due to the unique need to preserve physiological osmolarity.

In addition, because I don't think they can make extremely strong claims about the biological role of these neurons and the circuit in which they are embedded, the breadth of interest beyond the immediate field is still relatively limited.

This work is foundational: we identify new neurons for thirsty seeking of water. This is in contrast to providing new characterization of previously described neurons that is more common in the literature of the circuitry of behavior (for example, the mushroom body circuitry or the P1 neurons in courtship and aggression). Indeed, even those well-characterized circuits still give up new information for their function, so completeness is relative (see also limits of ITP and ISN characterization in the previous comment). Our work provides a clear basis for new studies on thirst and its relationships to other internal states, and the specific behavioral steps that compose it.

Most surprisingly, the Janu neurons we characterize can simultaneously antagonize competing states like hunger through interactions with NPF neurons. In the mammalian literature, thirst can suppress feeding behaviors. The neural mechanism is executed by glutamatergic neurons in the subfornical organ and are critical for this thirst-dependent hunger suppression (Zimmerman et al., Nature. 2017, Figure S7 CD), however the hierarchical wiring between thirst and hunger neurons is only beginning to be constructed and is an active area of research (Gong et al., Cell. 2020). We hope that our work will shed light on the fundamental characteristics of one type of thirst neuron (Janu neurons) and provide an entry point into the discovery of even more thirst neurons that may work in a parallel or redundant fashion, in concordance with homeostatic neuron functionality (Betley et al., Cell. 2013).

The authors have identified multiple types of "Janu" neurons with specific anatomical and neurochemical properties (e.g., GABA, AstA). These neurons all control thirst seeking, and also other unique (but related) behaviors. The Janu-AstA neuron is argued to synapse with and signal to NPF neurons. Focusing in on the key data in support of these main claims, I am only mildly convinced. The strongest phenotypes they observe are always with the broad drivers (e.g. Durstig-Gal4, which has widespread expression through the brain), but once they home in on a split-Gal4 combination with very selective labeling, the phenotype effects of activation or suppression are very small (e.g. Figure 2J-K or the new optogenetic activation experiments in Figure 5E).

Specific reduction of either GABA in Janu-GABA (Figure 3J) or AstA in Janu-AstA (Figure 4H) strongly inhibits thirsty seeking, indicating their critical role in the behavior. Thus the critique of “strongest behavior in the broadest driver” is limited to the activation experiments. Why the reduced activation (TrpA1) and inactivation (Shi) effects in 2J-K may be due to weaker driving of these dose dependent heat activated tools), however we note that the results are significant. The Janu-GABA and JanuAstA neurons are clearly critical for thirsty seeking.

Milder behavioral phenotypes are commonly observed with split-Gal4s. For example, activation of single glutamatergic mushroom body output neurons (MBONs) that tile the horizontal lobe will induce avoidance. However, manipulation of multiple groups of glutamatergic MBONs results in a stronger and synergistic avoidance phenotype (Aso et al., eLife. 2014. Figure 4). It is possible that JanuAstA is one type of AstA neuron (or non-AstA neurons in R65D05) that can promote water seeking. We also note that Janu is a weaker Gal4 driver for the Janu-AstAs than R65D05, assessed by myristoylated-GFP expression levels, so it may simply be Gal4 driver strength.

For some experiments, they revert to using the broader driver, such as the GRASP with NPF neurons using R65D05-Gal4. Unfortunately, this Gal4 line is expressed in a mass of neurons in this region of the brain (Figure 2S1A), so it's not clear to me that a positive GRASP signal really reflects synaptic contact between Janu-AstA neurons (somewhere within the R65D05 pattern) and NPF neurons.

GRASP between Janu and NPF was technically impossible because NPF-LexA and R52A01-AD reside in the same attP40 landing site, and the GRASP halves are located nearby in the genome on the second chromosome. We tried and failed in making triple transgenic recombinants with GRASP to make Janu/NPF GRASP possible.

We agree that positive GRASP signals between Janu-AstA and NPF neurons would be explicitly clear, however we have provided orthogonal lines of evidence that argue Janu-AstA is pre-synaptic to NPF neurons:

A. These are AstA-positive directional GRASP signals between R65D05 (presynaptic) and NPF (postsynaptic) neurons. While R65D05 does contain other AstA-positive neurons in the SMP, the location at the extreme medial aspect of the SMP is Janu-AstA exclusively. We can provide immunohistochemical evidence.

B. AstA in Janu-AstA neurons promotes thirsty seeking (Figure 4H).

C. AstA-R2 is required in NPF neurons for thirsty seeking (Figure 6D).

As mentioned in the previous round of reviews, the baseline phenotypes of controls is variable (e.g. in Figure 3I and Figure 3S1A VGAT RNAi experiments, there are many fewer flies on H2O compared to those with Gad1 RNAi in Figure 3J). The authors somewhat dismissed this concern by arguing that this may reflect variable environmental conditions and water deprivation treatments of flies of experiments on different days (and they only compared assays run in parallel), but this remains a concern for me.

We argue that our approach may be better, albeit more variable, than those in prior publications studying thirst in Drosophila. Thirst is a linearly graded state that is read out by the intensity of seeking (Figure 1D, see also Pool et al. for intake; https://doi.org/10.1016/j.neuron.2014.05.006), so the absolute amount of water deprivation can vary in a range without compromising the results. Other labs that studied thirst typically take the flies to a fraction of lethality to ensure similar deprivation between experiments. We chose a more naturalistic and intermediate deprivation because it is likely that flies invoke emergency pathways when nearing death, and we specifically wanted to study thirst.

Moreover, we provide genetic controls for all experiments to distinguish differences in water seeking within an experiment.

Finally, biological variability between experimental controls is not uncommon in these types of behavioral assays. For example, in a similar ‘seeking assay’ using 24 hour food deprived flies, ‘control food seeking’ can ranges from 40-80% (Tsao et al., eLife. 2018. Figure 2,3,4,6, 7, 8, 10, 11)

The n of several assays seems rather low (6-7), so it's not clear if they did all experiments of one type on a single day.

We will amend the Methods to reflect that all experiments were done across multiple days. This approach to animal behavior was instilled into the senior author (Wolf) through training in the behavioral mechanisms of alcohol inebriation in flies, where day-to-day variability is much greater and is routinely factored into the experiments.

In any case, questions linger as to how robust some of their results are.

In response to the original request to examine the physiological response of Janu neurons to water deprivation, the authors have used the TRIC sensor, a measure of long-term changes in neuronal calcium. This is not yet an extremely widely-used reporter, and I suspect requires careful controls to be able to interpret properly.

It is well-used and well-vetted in the literature, including but not limited to:

www.pnas.org/cgi/doi/10.1073/pnas.1706608114

https://doi.org/10.1038/s41593-019-0515-z

https://doi.org/10.1016/j.neuron.2018.07.001

https://doi.org/10.7554/eLife.54229

https://doi.org/10.1016/j.neuron.2019.04.009

http://dx.doi.org/10.7554/eLife.22441

Our use to measure internal state-dependent neural activity (longer time frame neuromodulatory events) is consistent with its prior uses.

We showed that the Janu-GABAs respond to thirst but not hunger, indicating that the reporter is not responding in some non-specific manner. An additional example, the NPF P1 neurons respond to both thirst and hunger, indicating TRIC can detect modulatory activity changes for distinct states in the same neuron.

They observe that one of the Janu-GABA neurons, but not any of the other three Janu neurons, displays increased TRIC signal upon water deprivation. This is interesting, but I find it puzzling that it should so selective, given that all of these neurons are behaviorally implicated in thirst seeking.

Yes, biology is richly complicated (thank goodness), and we captured aspects of that. We propose in the discussion that the Janu-AstA may be regulated by thirst below the detection limit of the tool. Alternatively, the regulation may be at the AstA-R2 receptors, with Janu-AstA providing a tonic signal. The latter mechanism is mechanistically analogous to sNPFR regulation by starvation to tune sensory input (https://doi.org/10.7554/eLife.08298).

I'd be interested to know what TRIC signal would be observed with a broad driver (e.g. Durstig-Gal4); is the single Janu-GABA neuron the only one that would have a significant increase in TRIC fluorescence? Water deprivation presumably induces widespread physiological changes in the brain, some of which may be direct sensors, and others show indirect activity. Ultimately, I don't find the new data so compelling in revealing this neuron to be a special sensor of thirst. (I appreciate more acute measurements of physiological activity are beyond the scope of the current work).

We propose that Janu-AstA promotes thirsty seeking of water and suppresses feeding behavior, and therefore is a critical neuron in the representation of thirst in the brain. That we can subtract nearly all water seeking from Durstig with R52A01 and R65D05, that we can subtract nearly all thirsty seeking from Janu with either Gad1 RNAi (Figure 3J) or AstA RNAi (Figure 4I) speaks to the importance of these neurons for thirst representation. AstA RNAi also strongly suppresses thirsty seeking from Durstig and R65D05 (Figure 4S1 DandE). The specificity for the seeking aspect of thirst strengthens the argument that it is not due to the creation of a competing drive (i.e. it is a thirst neuron).

The question of how many neurons change activity in thirst is fascinating: we predict widespread circuit activity, due to the multiphasic nature of thirst behavioral responses, and the interactions of thirst circuitry with other brain systems, as was recently demonstrated in mice (http://science.sciencemag.org/content/364/6437/eaav3932). Building tools to functionally capture and then manipulate these patterns will be crucial to future complete understanding of the parsing of internal and external sensory cues to select and carry out the appropriate behavior.

Our narrower scope approach is typical for behavioral assignment advances, while the scientific field designs functional approaches to determine the causal role of all the neurons (and glia) regulated by a state.

Although the authors have streamlined the figures, the reader is still directed to the many ancillary experiments in supplementary figures, now often with little explanation (e.g. Figure 1S1A-B, are only obliquely referred to in the text as "Neuronal activation and inactivation experiments uncovered additional behavioral roles for neurons in the Durstig pattern", and I don't fully understand the results in these panels).

This was an apparently bad decision by the senior author (Wolf) to leave in some of the earlier but confounding data from Durstig (1S1A-B). The data shows that, in a more complex expression pattern, additional neurons influence the measured behaviors; this is useful for future characterization of additional neurons, but we agree that it can be confusing for the readers. We will further streamline the presentation if allowed to do so, specifically eliminating 1S1A-B, and reassessing the value of others in the supplementals.

There is still an awkward transition at line 99 from the behavioral characterization of thirst-seeking to the screening for neuron populations promoting "seeking behavior"; ultimately, I find the Durstig-Gal4 neuron data almost impossible to interpret given its widespread expression pattern.

We agree that the use of “seeking behavior” to describe the TrpA1 activation screen is unclear.

The Durstig data should have been limited to our discovery that thirst neurons exist in the pattern. We will further streamline the presentation if allowed to do so, as noted in the previous reply. We intended for the reader to understand that we began the discovery of the thirst neurons with Durstig.

On line 127 they say that "subtracting R65D05 neurons from Durstig reduced activation-dependent water-seeking, indicating that the water seeking neurons in Durstig and R65D05 are the same"; this is true, but there is still substantial water-seeking behavior implying there are other seeking-promoting neurons in the Durstig pattern. Are these Janu-GABA neurons or others not characterized here?

The partial effects in Figure 2A (line 127) are likely due to the generally weaker activity of LexA drivers, here used to drive Gal80 to subtract R65D05 from Durstig.

Both the Janu-GABA and the Janu-AstA neurons are contained in Durstig, R65D05, and R52A01. Subtracting either R52A01 or R65D05 from Durstig decreases Durstig water seeking (Figure 2A, 2D), and thirsty seeking is readily blocked with inactivating the Janu split (Figures3J and 4H). We can add statements to ensure clarity. We included the subtraction experiments (and others) because they demonstrate that we isolated the Janu neurons from the Durstig pattern, and therefore are drivers of the original water seeking phenotype. There may be additional thirst neurons in the Durstig pattern. We demonstrated that neurons previously found to influence humidity preference (R48B04, likely a specific subset of PAM dopamine neurons) were not in Durstig.

As this is a review of a revised manuscript, and as I believe the authors have gone as far as they can in the current revision, I do not propose further experimentation, as I believe a decision should be made based solely on the data now available.

Thank you for your comments.

Reviewer #2:

In this paper, Landayan et al. characterize two pairs of Drosophila neurons (called Janu) involved in orientation behaviors toward water following water deprivation (called thirst here). One pair of these neurons is GABAergic (Janu-GABA), while the other is producing allatostatin A (Janu-AstA). The anatomy, projections and role in the behavior of these neurons is dissected in detail using a host of approaches and genetic intersectional approaches. By manipulating the activity of either of these neurons, the authors show that Janu neurons motivate flies to seek humidity but that they are not promoting drinking. They also confirm earlier studies showing that motivation to search for humidity are antagonist to food searching. These results are new, interesting, and also puzzling because they indicate that the activation of only a few interneurons has a profound influence on complex behaviors, the authors suggesting that these neurons are linked to motivational states (like thirst).

Thank you.

While the experiments and the results reported here seem sound, the paper is difficult to read for those unfamiliar to the field. The starting point of this study is the use of a simple behavioral test to evaluate the aggregation of flies next to food or water to screen a library of Gal enhancer trap flies (InSite), while activating the corresponding cells (using TrpA1 and heat). They find one enhancer-GAL4 mutant called Durstig (thirst in German), which is particularly attracted by food and by humidity when the corresponding cells are activated. From the description of this behavior (ms lines 99-110), it is not clear if Durstig neurons are only involved in the attraction toward humidity or also toward food odors.

We will clarify by focusing on Durstig as a starting point to find the thirst interneurons, dropping our additional characterization of the pattern. The Durstig pattern likely contains multiple neurons that influence a variety of behaviors, including but not limited to thirst and feeding behaviors.

In a first step of their research, the authors further used this behavioral test to evaluate if other genetic constructions could reproduce the behavioral pattern observed in Durstig> TrpA1 flies. Only two of these constructions reproduced the water-seeking pattern of Durstig flies. One is described as the "NP883 pattern", but this pattern has not been further compared to Durstig neurons. The second is called "R65D05 pattern" which was shown here as including Durstig neurons responsible for humidity searching.

In order to narrow down the Durstig neurons responsible for thirst-related attraction behavior, they used Gal80 constructions to subtract out neurons with various neurotransmitters. Finally, the authors were able to narrow down the neurons responsible to the attraction to 2 pairs of neurons characterized / defined by a split-Gal4 construction called Janu (again meaning thirst but this time in Estonian). According to figure 2G, this construction combines R52A01-AD and D65D05-DBD. It is thus not directly related to Durstig, except that activating the Janu cells induces the same behavioral phenotype related to humidity attraction and thirst.

Subtracting either R52A01 or R65D05 from Durstig blocks Durstig water seeking (Figure 2A, 2D), thus the Janu neurons are in the Durstig pattern.

What is next described in the paper is a thorough analysis of the Janu neurons, ie their anatomical location (in the adults), and the finding that two are GABAergic, while the two others are expressing allatostatin A. They also characterize their function and role in the behaviors studied here.

As compared to the first version, the ms is much improved and easier to understand. Besides the factual results shown here, it represents a very interesting work illustrating top-notch genetic strategies to identify neurons responsible for specific behaviors.

Thank you.

A general sketch of the Janu neurons, summarizing their role in thirst, and their relation to feeding, etc, might be useful.

Summary of thirst circuitry. The ISN senses hemolymph osmolarity to switch between thirst and hunger, likely setting in motion the full thirst behavioral program when the ISN is turned off by high osmolarity (Jourjine et al., 2016). Janu-AstA and Janu-GABA are independently necessary and sufficient for promoting the water seeking step of thirst. Janu-AstA also inhibits feeding behavior, releasing AstA that acts through AstA-R2-expressing NPF neurons to regulate NPF release. AstAR2 expressing NPF neurons also regulate water intake independently of the Janu-AStA neurons.

Reviewer #3:

The authors use an unbiased screen for neurons that contribute to feeding behavior to identify new circuits that affect thirst. Some of these neurons are GABAergic, while others release the neuropeptide Allatostatin, and they connect to the NPF-positive neurons previously implicated in hunger sensation. The interplay between thirst and hunger is complex and a better understanding of how these neurons interact with sensory and motor components will help illuminate how flies balance these conflicting motivational drives.

The authors have clearly made a good faith effort to address the issues raised in the initial review of this manuscript.

Thank you.

There is an enormous amount of work presented, and there is the potential that it reveals neurons that contribute to appropriate behavioral responses to hunger and thirst. Understanding how neural circuits function to regulate choices between essential and competing drives is a really important area of research.

But I remain very confused. There is still no summary diagram for either the (known and new) neuronal circuitry or the interaction between the motivational drives.

Please see the summary diagram which is included as Figure 6.

The brief flow diagram of the response to thirst helps a bit but does not really clarify. Some neurons seem to be able to induce a motivational state (or the behavioral response associated with it) while others are partially but not completely required for an appropriate choice between eating and drinking?

The Janu-GABA neurons are activated by thirst, and they specifically encode thirsty water seeking. We did not detect Janu-AstA activation by thirst, however they, like the Janu-GABAs, are required for thirsty seeking. The Janu-AstA neurons also reciprocally inhibit feeding behavior. These behavioral characterizations are done in detail in Figures 3 and 4. The summary diagram for Figure 6 will encapsulate these findings.

I think the Janu-GABA and Janu-allatostatin neurons are different populations, but I am actually not sure.

Yes, they are. That is shown anatomically in the flip-out experiments (Figure 2S4), and anatomically and functionally in Figures 3 and 4.

The figures are very dense and intermix feeding and drinking behaviors, activation and inactivation experiments, and different sets of neurons. The combination thirst response and place preference in figure 5, for example, makes no sense at all.

Figure 5 is a marriage of necessity – Both experiments are important for understanding thirst neuron function, and this is the location they belong in the story flow, yet they do not neatly fit into the other figures. Also, both address the Janu group as a whole, which is necessary at this point, right after we defined the function of the Janu-GABA and Janu-AstA neurons (Figures 3 and 4), and right before we provide evidence that the Janu-AstA neurons connect to NPF neurons (Figure 6). We hope that this is a minor concern.

Finally, we will color code thirst vs feeding behaviors, to help readers distinguish the experiments.

The anti-VGAT staining seems to label the entire neuropil, which makes the co-localization with some of the Janu neurons very difficult to interpret. The allatostatin colocalization is similarly confusing.

GABAergic neurotransmission is widespread throughout the adult Drosophila brain. The anti-VGAT staining is specific: there was no overlap of anti-VGAT staining in the Janu-AstA neurons, and the anti-VGAT overlap with the Janu-GABAs was limited to the larger varicosities in the Janu-GABA neurons, which typically are presynaptic regions of the neurons. Moreover, nearly all of the JanuGABA varicosities were VGAT-positive, which would not be expected for non-specific staining. We used anti-VGAT according to the conditions described by the David Krantz lab at UCLA that produced and characterized the antibody. We can elaborate this in the manuscript to help clarify.

The Allatostatin localization is in the cell bodies of the Janu-AstA neurons (Figure 4B-D) and in their presynaptic release sites (Figure 4E-G), as defined with neuronal polarity markers in Figure 2I.

Neuromodulator staining in both the cell bodies and the endings is typical (for example extensively documented in the NPF literature). The Janu-AstA neurons are the only AstA-positive neurons in the Janu expression pattern (4S1). AstA knockdown in Janu neurons blocks thirsty seeking (Figure 4H). Like GABA, there are additional AstA neurons outside of Janu – we are specifically studying those in the Janu pattern.

This manuscript still needs a major overhaul to make its contributions to thirst circuitry interpretable.

We will do our best to decrease confusion by providing additional guiding diagrams as suggested, and by eliminating some of the supplemental material that is not central to the main conclusions (also as suggested). Some of the genetic techniques are moderately difficult to understand, and we will try to help the readers further with explanatory diagrams.

https://doi.org/10.7554/eLife.66286.sa2

Article and author information

Author details

  1. Dan Landayan

    Quantitative and Systems Biology Graduate Program, UC, Merced, United States
    Present address
    Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, United States
    Contribution
    Conceptualization, Formal analysis, Investigation, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared

  2. Brian P Wang

    Quantitative and Systems Biology Graduate Program, UC, Merced, United States
    Contribution
    Software, Formal analysis, Investigation, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared

  3. Jennifer Zhou

    Department of Molecular and Cell Biology, UC, Merced, United States
    Contribution
    Investigation
    Competing interests
    No competing interests declared

  4. Fred W Wolf

    1. Quantitative and Systems Biology Graduate Program, UC, Merced, United States
    2. Department of Molecular and Cell Biology, UC, Merced, United States
    Contribution
    Conceptualization, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing
    For correspondence
    fwolf@ucmerced.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4230-194X

Funding

UC Merced (Startup funds)

  • Fred W Wolf

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Thomas Clandinin, Daryl Gohl, and Marion Silies for InSITE strains, Flidea for building the custom optogenetic device, Ashley Smart and Thomas Clandinin for imaging help, Anand Bala Subramaniam for access to the LSM-880 confocal microscope, and Donnoban Orozco, Nailah Fatima, Natalie Banh, Winnie Wu, Lawrence Fung, Kenneth Mata, Maurice Jackson, Van Thai, and Layne Hazam for help with the experiments. This work was supported by UC Merced startup funds to FWW.

Senior Editor

  1. K VijayRaghavan, National Centre for Biological Sciences, Tata Institute of Fundamental Research, India

Reviewing Editor

  1. Mani Ramaswami, Trinity College Dublin, Ireland

Publication history

  1. Received: January 8, 2021
  2. Accepted: April 30, 2021
  3. Version of Record published: May 21, 2021 (version 1)

Copyright

© 2021, Landayan et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,954
    Page views
  • 259
    Downloads
  • 9
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Dan Landayan
  2. Brian P Wang
  3. Jennifer Zhou
  4. Fred W Wolf
(2021)
Thirst interneurons that promote water seeking and limit feeding behavior in Drosophila
eLife 10:e66286.
https://doi.org/10.7554/eLife.66286

Categories and tags

Research organism

Further reading

    1. Evolutionary Biology
    2. Genetics and Genomics

    The lingering effects of Neanderthal introgression on human complex traits

    Xinzhu Wei, Christopher R Robles ... Sriram Sankararaman
    Research Article

    The genetic variants introduced into the ancestors of modern humans from interbreeding with Neanderthals have been suggested to contribute an unexpected extent to complex human traits. However, testing this hypothesis has been challenging due to the idiosyncratic population genetic properties of introgressed variants. We developed rigorous methods to assess the contribution of introgressed Neanderthal variants to heritable trait variation relative to that of modern human variants. We applied these methods to analyze 235,592 introgressed Neanderthal variants and 96 distinct phenotypes measured in about 300,000 unrelated white British individuals in the UK Biobank. Introgressed Neanderthal variants have a significant contribution to trait variation consistent with the polygenic architecture of complex phenotypes (contributing 0.12% of heritable variation averaged across phenotypes). However, the contribution of introgressed variants tends to be significantly depleted relative to modern human variants matched for allele frequency and linkage disequilibrium (about 59% depletion on average), consistent with purifying selection on introgressed variants. Different from previous studies (McArthur 2021), we find no evidence for elevated heritability across the phenotypes examined. We identified 348 independent significant associations of introgressed Neanderthal variants with 64 phenotypes . Previous work (Skov 2020) has suggested that a majority of such associations are likely driven by statistical association with nearby modern human variants that are the true causal variants. We therefore developed a customized statistical fine-mapping methodology for introgressed variants that led us to identify 112 regions (at a false discovery proportion of 16%) across 47 phenotypes containing 4,303 unique genetic variants where introgressed variants are highly likely to have a phenotypic effect. Examination of these variants reveal their substantial impact on genes that are important for the immune system, development, and metabolism. Our results provide the first rigorous basis for understanding how Neanderthal introgression modulates complex trait variation in present-day humans.

    1. Cell Biology
    2. Genetics and Genomics

    Identification of a weight loss-associated causal eQTL in MTIF3 and the effects of MTIF3 deficiency on human adipocyte function

    Mi Huang, Daniel Coral ... Sebastian Kalamajski
    Research Article Updated

    Genetic variation at the MTIF3 (Mitochondrial Translational Initiation Factor 3) locus has been robustly associated with obesity in humans, but the functional basis behind this association is not known. Here, we applied luciferase reporter assay to map potential functional variants in the haplotype block tagged by rs1885988 and used CRISPR-Cas9 to edit the potential functional variants to confirm the regulatory effects on MTIF3 expression. We further conducted functional studies on MTIF3-deficient differentiated human white adipocyte cell line (hWAs-iCas9), generated through inducible expression of CRISPR-Cas9 combined with delivery of synthetic MTIF3-targeting guide RNA. We demonstrate that rs67785913-centered DNA fragment (in LD with rs1885988, r2 > 0.8) enhances transcription in a luciferase reporter assay, and CRISPR-Cas9-edited rs67785913 CTCT cells show significantly higher MTIF3 expression than rs67785913 CT cells. Perturbed MTIF3 expression led to reduced mitochondrial respiration and endogenous fatty acid oxidation, as well as altered expression of mitochondrial DNA-encoded genes and proteins, and disturbed mitochondrial OXPHOS complex assembly. Furthermore, after glucose restriction, the MTIF3 knockout cells retained more triglycerides than control cells. This study demonstrates an adipocyte function-specific role of MTIF3, which originates in the maintenance of mitochondrial function, providing potential explanations for why MTIF3 genetic variation at rs67785913 is associated with body corpulence and response to weight loss interventions.

    1. Epidemiology and Global Health
    2. Genetics and Genomics

    Smoking, alcohol consumption, and 24 gastrointestinal diseases: Mendelian randomization analysis

    Shuai Yuan, Jie Chen ... Susanna C Larsson
    Research Article Updated

    Background:

    Whether the positive associations of smoking and alcohol consumption with gastrointestinal diseases are causal is uncertain. We conducted this Mendelian randomization (MR) to comprehensively examine associations of smoking and alcohol consumption with common gastrointestinal diseases.

    Methods:

    Genetic variants associated with smoking initiation and alcohol consumption at the genome-wide significance level were selected as instrumental variables. Genetic associations with 24 gastrointestinal diseases were obtained from the UK Biobank, FinnGen study, and other large consortia. Univariable and multivariable MR analyses were conducted to estimate the overall and independent MR associations after mutual adjustment for genetic liability to smoking and alcohol consumption.

    Results:

    Genetic predisposition to smoking initiation was associated with increased risk of 20 of 24 gastrointestinal diseases, including 7 upper gastrointestinal diseases (gastroesophageal reflux, esophageal cancer, gastric ulcer, duodenal ulcer, acute gastritis, chronic gastritis, and gastric cancer), 4 lower gastrointestinal diseases (irritable bowel syndrome, diverticular disease, Crohn’s disease, and ulcerative colitis), 8 hepatobiliary and pancreatic diseases (non-alcoholic fatty liver disease, alcoholic liver disease, cirrhosis, liver cancer, cholecystitis, cholelithiasis, and acute and chronic pancreatitis), and acute appendicitis. Fifteen out of 20 associations persisted after adjusting for genetically predicted alcohol consumption. Genetically predicted higher alcohol consumption was associated with increased risk of duodenal ulcer, alcoholic liver disease, cirrhosis, and chronic pancreatitis; however, the association for duodenal ulcer did not remain statistically significant after adjustment for genetic predisposition to smoking initiation.

    Conclusions:

    This study provides MR evidence supporting causal associations of smoking with a broad range of gastrointestinal diseases, whereas alcohol consumption was associated with only a few gastrointestinal diseases.

    Funding:

    The Natural Science Fund for Distinguished Young Scholars of Zhejiang Province; National Natural Science Foundation of China; Key Project of Research and Development Plan of Hunan Province; the Swedish Heart Lung Foundation; the Swedish Research Council; the Swedish Cancer Society.