Adult functions for the Drosophila DHR78 nuclear receptor

Abstract

Background

The Testicular Receptors 2 and 4 (TR2, TR4) comprise a small subfamily of orphan nuclear receptors. Genetic studies in mouse models have identified roles for TR4 in developmental progression, fertility, brain development, and metabolism, as well as genetic redundancy with TR2. Here we study the adult functions of the single Drosophila member of this subfamily, DHR78, with the goal of defining its ancestral functions in the absence of genetic redundancy.

Results

We show that DHR78 mutants have a shortened lifespan, reduced motility, and mated DHR78 mutant females display a reduced feeding rate. Transcriptional profiling reveals a major role for DHR78 in promoting the expression of genes that are expressed in the midgut, suggesting that it contributes to nutrient uptake. We also identify roles for DHR78 in maintaining the expression of genes in the ecdysone and Notch signaling pathways.

Conclusions

This study provides a new context for linking the molecular activity of the TR orphan nuclear receptors with their complex roles in adult physiology and lifespan.

Introduction

Nuclear receptors are ligand-regulated transcription factors that play central roles in growth, development, and metabolism (Chawla et al., 2001; Mullican et al., 2013). They are defined by a highly conserved zinc finger DNA binding domain and a C-terminal ligand binding domain, which can accommodate small lipophilic chemical compounds. Although the transcriptional activity of many nuclear receptors is regulated by ligand binding, a large number remain classified as orphan receptors, with no known ligand. In this study, we focus on a small subfamily of orphan nuclear receptors, the Testicular Receptors 2 and 4 (TR2, TR4; NR2C1/2). Although these receptors were named based on their abundant expression in the testes and brain, this receptor subfamily is also expressed in the kidneys, liver, adipose tissues, and muscle, suggesting that they may have broad functions (Lin et al., 2014).

Initial phenotypic studies of TR4 null mutant mice revealed clear postnatal phenotypes, with significant growth retardation and a shortened lifespan (Collins et al., 2004). TR4 mutants also display reproductive defects consistent with their abundant expression in the testes and ovary. These include defects in spermatogenesis and reduced male fertility, as well as reduced female fertility, smaller ovaries, and defects in follicle development (Collins et al., 2004; Mu et al., 2004; Chen et al., 2008a). TR4 mutants also display defects in motor coordination and abnormal development of the cerebellum, consistent with its abundant expression in neuronal tissues (Chen et al., 2008b). Several more recent studies have extended the roles of TR4 to metabolism and mitochondrial function. These include smaller adipocytes with reduced triglyceride accumulation, protection against hepatic steatosis, and resistance to diet-induced obesity in TR4 mutant mice (Kang et al., 2011). In contrast to these broad roles for TR4 in development and metabolism, TR2 mutant mice were reported as having no detectable phenotypes (Shyr et al., 2002), although a recent paper identified defects in vision and retinal development (Olivares et al., 2017). TR2, TR4 double-knockout mice, however, are early embryonic lethal, indicating that these paralogs have redundant functions (Shyr et al., 2009). Moreover, most studies of these widely expressed nuclear receptors have been restricted to whole body knockouts, leaving us with a poor understanding of their tissue-specific functions. In contrast, the fruit fly Drosophila has only a single copy of the TR2/4 receptor subfamily, DHR78. Functional studies of DHR78 in Drosophila thus provides an ideal opportunity to define the mechanisms by which the TRs regulate development and metabolism without the complications of genetic redundancy.

DHR78 was originally identified in molecular screens for new nuclear receptors (Fisk and Thummel, 1995; Zelhof et al., 1995). Two EMS-induced point mutations were generated that behave as genetic null alleles (Fisk and Thummel, 1998). DHR78¹ carries a G-to-A transition at the splice acceptor site in the second intron of DHR78, while DHR78³ contains a C-to-T transition that changes an arginine codon in the ligand-binding domain coding region to a stop codon (Fisk and Thummel, 1998). DHR78 mutants die as larvae with reduced size and developmental asynchrony. Their tracheal system appears highly disrupted, with defects in gas filling and remnants of tracheal cuticle from previous molts (Fisk and Thummel, 1998). All of these phenotypes, including the larval lethality, can be rescued by tracheal-specific expression of wild-type DHR78 in a mutant background (Astle et al., 2003). In addition, tracheal-specific DHR78 RNAi recapitulates the major defects seen in the whole animal mutant. Taken together, these studies indicate that DHR78 is necessary and sufficient for proper tracheal development during larval stages. This is consistent with the expression pattern of DHR78 protein, which is abundant in the trachea and salivary glands, along with expression in the imaginal discs, intestine, and Malpighian tubules (Fisk and Thummel, 1998).

Here we exploit our ability to rescue the larval lethality of DHR78 mutants to study the functions of this nuclear receptor in adults. We show that mutants rescued to adulthood with tracheal-specific expression of wild-type DHR78 appear normal but display reduced motility and a shortened lifespan. Stored energy in the forms of glycogen and triglycerides, as well as glucose, are unaffected. Mated DHR78 mutant females, however, feed less than controls. Transcriptional profiling by RNA-seq reveals a major effect on genes expressed in the intestine, including reduced expression of endopeptidases, transmembrane transporters, and components of the peritrophic matrix. Taken together, our results parallel those seen in studies of TR4 mutant mice and provide a molecular framework for understanding the roles of DHR78 in larval and adult development.

Results

DHR78 mutants die as larvae with tracheal defects

The DHR78¹ and DHR78³ mutants were recovered from our lab stocks and outcrossed to w¹¹¹⁸ for several generations. As expected, the resulting DHR78¹/DHR78³ transheterozygous null mutants display the range of phenotypes reported in our earlier studies (Fisk and Thummel, 1998; Astle et al., 2003). These include larval lethality and developmental asynchrony, with most mutants arresting their development during the second and third larval instars (Fig. 1A,B). These animals display defects in tracheal molting and reduced size (Fig. 1B,C). We also noted that mutant larvae were often on the sides of the vials, displaying a classic anoxic escape response consistent with their disrupted tracheal system (Wingrove and O'Farrell, 1999). Metabolite measurements in second instar larvae revealed low levels of glycogen and triglycerides, suggesting that these small wandering animals survive on stored nutrients (Fig. 1D).

Fig. 1. DHR78 mutants display larval lethality and tracheal defects.

(A) The percent of surviving w¹¹¹⁸ controls and DHR78^1/3 mutants are shown at different time points representing the stage of control animals as either first instar larvae (L1), second instars (L2), early or late third instars (L3), prepupae (PP), or pupae. Most DHR78 mutant larvae are dead when controls pupariate. DHR78 mutants progress normally into the second instar, when they begin to display roving behavior, developmental asynchrony, and lethality as second or early third instar larvae (Fisk and Thummel, 1998). (B) DHR78 mutants that reach the third instar are significantly smaller than w¹¹¹⁸ controls and (C) display tracheal molting defects. (D) Late second instar w¹¹¹⁸ controls (blue) and DHR78^1/3 mutants (red) display reduced glycogen and triglyceride stores, likely due to a starvation response from their roving behavior. Data was normalized to protein levels to account for size differences between controls and mutants, then graphed relative to control levels. Data are graphically represented as box plots. * p<0.05, ** p<0.005

DHR78 mutant adults display reduced motility and a shortened lifespan

The early lethality of DHR78 null mutants prevents a broader understanding of its regulatory roles outside of tracheal development. In addition, their wandering behavior complicates an assessment of possible metabolic functions for this nuclear receptor. To overcome these problems, we used tracheal-specific expression of wild-type DHR78 in a DHR78¹/DHR78³ transheterozygous mutant background to rescue their lethality through to adulthood (Astle et al., 2003). We generated five stocks for our studies of DHR78 function during adult stages, all in a common w¹¹¹⁸ genetic background: (1) a w¹¹¹⁸ control stock, (2) w¹¹¹⁸; btl-GAL4, (3) w¹¹¹⁸; UAS-DHR78, (4) w¹¹¹⁸; btl-GAL4; UAS-DHR78, and (5) these two transgenes in DHR78¹/DHR78³ mutants. We refer to the latter stock (w¹¹¹⁸; btl-GAL4/+; UAS-DHR78 DHR78³/DHR78¹) as "rescued mutants" throughout this study.

Initial analysis of DHR78 rescued mutants revealed that they all survive to adulthood and form normal-appearing animals, as reported in our earlier study (Astle et al., 2003). These mutant adults, however, die significantly more rapidly than controls, with males displaying a more prolonged lifespan than females (Fig. 2A,B). About one-third of mutant males survive to 45 days, when almost all mutant females have died. We also noticed that rescued DHR78 mutant males and females have reduced motility. This was quantified by tapping btl-GAL4; UAS-DHR78 controls and btl-GAL4; UAS-DHR78; DHR78¹/DHR78³ rescued mutants to the bottom of a vial and measuring the time it takes for them to climb 3 centimeters. This study revealed that the mutants display significantly reduced negative geotaxis (Fig. 2C,D). Many mutant animals fail to climb at all, remaining motile at the bottom of the vial. Interestingly, these defects are similar in both males and females, as well as across all three time points examined (Fig. 2C,D).

Fig. 2. Rescued DHR78 mutants display shortened lifespans and motility defects.

Lifespan was measured in four control genotypes (black, grey, grey dotted line, and blue) and rescued DHR78 mutant (red) males (A) and females (B). n > 55 flies per genotype. p<0.001 for both lifespan experiments, comparing btl>DHR78 controls (blue) to rescued DHR78 mutants (red). (C,D) btl>DHR78 controls (blue) and rescued DHR78 mutants (btl>DHR78; DHR78^1/3, red) as either males (C) or females (D) display reduced motility when tested with a negative geotaxis assay. Data are graphically represented as box plots. n = 15–35 flies per genotype. ***p<0.0005

DHR78 mutant females display a reduced feeding rate

In an effort to assess the metabolic health of DHR78 rescued mutants, we measured stored energy reserves in the forms of glycogen and triglycerides (TAG) as well as glucose in one week old males or mated females maintained on a normal diet (Fig. 3A–F). No significant differences were observed in any of these metabolites, although glucose levels appear to be more variable in males (Fig. 3A). We also assayed for feeding rates in rescued DHR78 mutant males and females to determine if this might contribute to their overall nutrient levels. Both virgin females and males displayed no significant changes in feeding rate over a two hour period (Fig. 3G). Mated females, however, consume approximately 50% less than controls. Taken together, these results suggest that mated DHR78 mutant females can compensate for their reduced feeding rate to maintain their stored energy reserves on a normal diet.

Fig. 3. DHR78 mutant mated females display a reduced feeding rate.

Samples were collected from four control genotypes (black, dark grey, light grey, and blue) and rescued DHR78 mutants (btl>DHR78; DHR78^1/3; red) as either males (A–C) or mated females (D–F), and assayed for glucose (A,D), glycogen (B,E), or triglycerides (C,F; TAG). n = 11–24 samples for each group. Data are graphically represented as box plots, relative to the metabolite level in btl>DHR78 controls (blue). (G) btl>DHR78 controls (blue) and rescued DHR78 mutants (btl>DHR78; DHR78^1/3, red) were assayed for feeding rate by measuring the amount of radioactive food ingested after an overnight fast. Data is combined from two separate experiments for each condition. *** p<0.0005

DHR78 regulates gene expression in the adult intestine

Defining the regulatory functions of DHR78 requires an understanding of its downstream target genes. Toward this goal, we analyzed the transcriptional profile of rescued DHR78 mutants using RNA-seq. We selected one week old female w¹¹¹⁸; btl-GAL4; UAS-DHR78 animals as controls and w¹¹¹⁸; btl-GAL4; UAS-DHR78 DHR78¹/DHR78³ rescued mutants. We used mated females because of their apparent defects in feeding (Fig. 3G). A total of 510 genes are differentially expressed in DHR78 mutants (≥1.5-fold change, 1% FDR), with 283 genes showing reduced abundance and 227 genes showing increased abundance (Supp. Table S1). Gene ontology (GO) analysis revealed that many genes encoding enzymes with predicted oxidoreductase activity are expressed at increased levels in DHR78 mutants (Table 1). Most of these are cytochrome P450 enzymes that could represent a detoxification response. Other major up-regulated gene categories correspond to defense response pathways suggesting that DHR78 mutants have an active stress response. GO analysis of the down-regulated dataset revealed a predominance of endopeptidases along with transmembrane transporters and components of the peritrophic matrix (Table 1). Interestingly, many of these genes are normally abundantly expressed in the Drosophila intestine. To examine this more closely, we analyzed the expression levels of DHR78-regulated genes in wild-type tissues as reported in FlyAtlas (Robinson et al., 2013). By depicting these results in a heat map, it became apparent that the genes that are up-regulated in DHR78 mutants are expressed in a range of different tissues in wild-type animals while many down-regulated genes are normally expressed in the midgut (Fig. 4). This includes many genes that are abundantly and specifically expressed in this tissue, such as the eleven Jonah peptidase genes that are down-regulated in DHR78 mutants (Supp. Table S1) (Carlson and Hogness, 1985). In contrast, comparisons of our data with published datasets revealed no significant overlaps with genes regulated by the DHR38, DHR96, or dHNF4 nuclear receptors (Sieber and Thummel, 2009; Ruaud et al., 2011; Barry and Thummel, 2016). Taken together, this data suggests that DHR78 plays a major role in regulating genes in the adult midgut, including enzymes that promote nutrient uptake. They also have an active stress response, consistent with their reduced survival as adults.

Table 1.

Top gene ontology categories for DHR78-regulated genes

Down-regulated genes (283)
GO category	Number of genes (total)	P value	# additional similar categories
Endopeptidase activity	25 (398)	1e-10	9
Urea cycle intermediate	3 (4)	1.7e-3	1
Transmembrane transporter activity	11 (150)	2.4e-3	9
Peritrophic matrix	5 (24)	2.7e-3
Lysozymes	4 (14)	4.1e-3	1
Hydrolase acting on glycosyl bonds	8 (102)	1.2e-2	1
Arginine metabolism	2(3)	3.0e-2

Up-regulated genes (227)
GO category	Number of genes (total)	P value	# additional similar categories
Oxidoreductase activity	24 (541)	3.9e-10
Defense response	16 (164)	1.3e-8
Electron carrier (P450s)	13 (166)	3.9e-6	4
Defense response to bacteria	7 (64)	3.8e-4	8
Response to stimulus	25 (982)	1.0e-3
Peptidase activity	17 (557)	1.1e-3	5

Determined using GOstat: http://gostat.wehi.edu.au/cgi-bin/goStat.pl

Fig. 4. Many intestinal genes are expressed at reduced levels in DHR78 mutants.

Heat maps are depicted for genes that change their expression level in DHR78 mutant females. Each row of the heat map represents a specific tissue, and each column represents an individual gene either (A) up-regulated or (B) down-regulated in mutants. The color corresponds to the expression level for that gene in wild-type tissues as reported by FlyAtlas (Robinson et al., 2013), with black representing little or no expression and green representing higher expression. More broadly expressed genes are shown on the right side of the heat map. Many genes that are abundantly expressed in the midgut of wild-type flies are down-regulated in DHR78 mutants.

Discussion

DHR78 represents the single Drosophila ortholog of the TR2/TR4 orphan nuclear receptor subfamily, providing an ideal simple model for functional analysis. In this study, we use genetic rescue to overcome the larval lethality of DHR78 null mutants, allowing us to perform phenotypic characterization during the adult stage. We show that DHR78 mutants have a shortened lifespan and reduced motility. Interestingly, these phenotypes resemble those seen in TR4 mutant mice, which have a shortened lifespan and defects in motor coordination (Collins et al., 2004; Chen et al., 2008b). In contrast, TR4 mutant mice display a slight but significant increase in feeding rate, the opposite of what we report here in mated Drosophila females (Kang et al., 2011). In addition, a number of genes encoding digestive enzymes are expressed at reduced levels in DHR78 mutant females, providing another potential level for reduced nutrient uptake. It is also interesting to speculate that the reduced feeding behavior could arise from an overall defect in neuronal function, which could also explain the reduced motility. This motility defect likely occurs prior to adult development since there is no change with age as is normally seen in progressive forms of neuronal or muscle degeneration in Drosophila. A defect in neuronal development would be consistent with studies of TR4 in mouse models (Chen et al., 2008b). In addition, a number of genes involved in Drosophila neuronal function are expressed at reduced levels in DHR78 mutants (Supp. Table S1). These include zydeco, which is involved in glial-neuronal interactions, Dscam4, which regulates dendritic targeting, and several genes involved in early neuronal development, including miranda, ventral nervous system defective, and robo2 (Melom and Littleton, 2013; Tadros et al., 2016; Gramates et al., 2017). Future studies of the tissue-specific functions of DHR78 could address this possible association between neuronal development and the reduced motility and feeding rate in mutants.

Analysis of the RNA-seq data revealed that a number of genes in the ecdysone signaling pathway are expressed at reduced levels in DHR78 mutant females (Supp. Table S1). These include disembodied, which encodes a key ecdysone biosynthetic enzyme, SMRTER, which acts as a corepressor for EcR, broad, which encodes an ecdysone-inducible zinc finger transcription factor, and Eip71CD (DiBello et al., 1991; Tsai et al., 1999; Chavez et al., 2000). Interestingly, our original study of DHR78 mutants centered on its role in coordinating ecdysone-regulated gene expression in larvae (Fisk and Thummel, 1998). Future studies could address the possibility that ecdysone levels are reduced in DHR78 mutant females. In this regard, we also note that the ovaries of mated DHR78 mutant females are reduced in size – another phenotype associated with both TR4 mutant mice and ecdysone signaling defects (Chen et al., 2008a; Sieber and Spradling, 2015).

Interestingly, several genes in the Notch signaling pathway are also expressed at reduced levels in DHR78 mutants (Supp. Table S1). These include Notch itself, pigs, which is a direct target of Notch signaling that restricts Notch activation, Rhabconnectin (Rbcn-3A), and uninflatable (uif) (Siebel and Lendahl, 2017). Notably, pigs mutants display reduced motility, suggesting that the reduced expression of this gene may contribute to this DHR78 mutant phenotype (Pines et al., 2010). Similarly, uif mutants display defects in tracheal development, tracheal growth, and tracheal molting, with lethality during larval development – phenotypes that resemble those of DHR78 mutants (Zhang and Ward, 2009). Taken together, these observations raise the interesting possibility that DHR78 acts to support proper levels of Notch signaling and that the reduced expression of uif may contribute to the lethal phenotypes seen in mutant larvae.

We also note that Notch lies next to another gene that is underexpressed in DHR78 mutants, kirre. Many other genes that are underexpressed in DHR78 mutants lie adjacent to each other. These include the LysC, LysD, and LysE gene cluster, the CG6295 and CG6296 predicted lipases, and the Jon25Bi, Jon25Bii, and Jon25Biii gene cluster. In contrast, only a few of the up-regulated genes lie next to one another. This observation suggests that DHR78 is binding directly to these loci and acting as a transcriptional activator, similar to its mouse ortholog TR4 (Lin et al., 2014). Molecular studies of the regulatory functions of DHR78 are required to allow a direct test of this model.

In conclusion, by moving our studies of DHR78 mutants into adults we have provided a new context for understanding the molecular and developmental roles of this orphan receptor. Further studies of DHR78 mutants will complement the characterization of TR2 and TR4 in mice, and will provide more insights into the molecular pathways that lie downstream from these orphan nuclear receptors – linking their transcriptional regulatory functions to the complex developmental and metabolic phenotypes in these animals.

Experimental Procedures

Drosophila Strains and Media

Flies were raised at 25°C on media containing 8% yeast, 6% glucose, 3% sucrose and 1% agar for all adult studies. Larvae were raised on grape juice media containing yeast paste at 25°C. All genetic studies used a transheterozygous combination of DHR78¹ and DHR78³ null alleles generated previously by EMS-induced mutation (Fisk and Thummel, 1998). Both strains were outcrossed for three generations to w¹¹¹⁸ to establish a uniform genetic background. The UAS-DHR78 and btl-GAL4 transgenic lines (Astle et al., 2003) were also outcrossed for three generations to w¹¹¹⁸, and then crossed to the DHR78³ and DHR78¹ mutant lines, respectively. The DHR78 stocks used in this study are: (1) w¹¹¹⁸; btl-GAL4; +/+, (2) w¹¹¹⁸; +/+; UAS-DHR78, (3) w¹¹¹⁸; btl-GAL4; DHR78¹/TM3 Dfd-GMR-nvYFP3 Sb¹, (4) w¹¹¹⁸; +/+; UAS-DHR78 DHR78³/TM3 Dfd-GMR-nvYFP3 Sb¹. Stocks 1 and 2 were crossed to generate the w¹¹¹⁸; btl-GAL4/+; UAS-DHR78/+ control line, and stocks 3 and 4 were crossed to generate the w¹¹¹⁸; btl-GAL4/+; DHR78¹/ UAS-DHR78 DHR78³ rescued mutants.

Lifespan Studies

The lethality of DHR78 mutants was analyzed by transferring 20–30 newly-hatched first instar larvae to fresh media with yeast paste and scoring daily for survival. At least three replicates were analyzed per condition. For adult lifespan studies, newly-eclosed flies were sorted based on sex and 20–25 flies were transferred to vials containing 8% yeast, 9% sugar for a total of 55–80 flies per condition. Five control males were added to each vial containing female flies. Flies were transferred daily to fresh food and scored for lethality.

Metabolic Assays

Larval metabolic studies shown in Figure 1D were conducted using 25 staged late second instar larvae from at least three independent collections. Metabolite levels were normalized to protein levels and combined to determine average concentration and standard error of the mean, relative to the metabolite levels found in w¹¹¹⁸ controls. Triglyceride (TAG) measurements were performed using a coupled colorimetric assay (Sigma T2449) as described (Palanker et al., 2009; Tennessen et al., 2014). Glycogen and glucose concentrations were determined using the Hexokinase (HK) and/or Glucose Oxidase (GO) assay kits (Sigma GAHK20, GAGO20) as described (Tennessen et al., 2014). Total protein levels were determined in parallel by Bradford assay. For adults, samples of five flies per replicate were collected at one week of age and homogenized in 100 µl of 1× PBS. Each assay was repeated at least three times for a combined total of 11–24 replicates per genotype, per sex. Metabolite levels were normalized to protein levels and combined to determine average concentration and standard error of the mean, relative to the metabolite levels found in btl-GAL4; UAS-DHR78 controls. Both control and tracheal-rescued mutant females were maintained with w¹¹¹⁸ males to ensure that mating took place.

Motility Assay

Flies were anesthetized at eclosion and sorted based on sex into fresh vials of food for a total of 10–15 flies per vial, with four vials per experiment. Motility was assayed no earlier than 24 hours post anesthetization and flies were not anesthetized again after the initial sorting. The motility assay was performed by transferring flies to an empty vial placed in front of graph paper and a ruler. The vial was tapped down until all flies fell to the bottom. The time it took for individual flies to climb 3 cm was recorded by video and later analyzed. This experiment was repeated three times with similar results.

Feeding Rate Assay

Control and mutant flies were sorted by sex at eclosion and placed into separate vials, with several w¹¹¹⁸ males added to the vials with females. Flies were collected at one week of age, fasted overnight on 1% agar, and then allowed to feed for two hours on radioactive media containing ~5,000 cpm/µl α-³²P- dCTP in 8% yeast, 9% sugar in 1% agar (Tennessen et al., 2014). Animals were then transferred to unlabeled food for 30 minutes and sorted into samples of five flies on ice. A scintillation counter was used to measure the radioactivity in each sample, and this value was used to determine the relative volume of media consumed, as described (Tennessen et al., 2014). Concentrations were normalized to mated control females. This experiment was repeated 2–3 times with a total of 8–10 flies per genotype, per condition.

RNA-Seq

One week old female DHR78 rescued mutants and controls (btl>DHR78) were raised on a 8% yeast, 9% sugar diet. Eight biological replicates per genotype (n=12 females per sample) were collected. RNA was extracted using Trizol (ThermoFisher 15596026) and pairs of biological replicates were pooled to obtain four biological replicates for further purification using an RNeasy Mini kit (Qiagen 74104). RNA quality was analyzed using an Agilent Bioanalyzer RNA 6000. Library generation (Illumina TruSeq RNA Sample Preparation Kit v2 with oligo dT selection) and sequencing (HiSeq 50 Cycle Single Read Sequencing v3) were performed by the High-Throughput Genomics core facility at the University of Utah. Standard replicate RNA-seq analysis was performed using USeq and DESeq analysis packages with alignment to the Drosophila melanogaster dm3 genome assembly. Transcripts meeting a cutoff of ≥1.5 fold difference in mRNA abundance and 1% FDR were considered as differentially expressed genes. Heat maps were generated in R using ggplots2 combined with tissue-specific expression data from FlyAtlas (Robinson et al., 2013). Genes were separated into lists based on whether they were up- or down-regulated in DHR78 rescued mutants and analyzed based on the abundance of gene expression in wild-type dissected tissues. RNA-seq data from this study can be accessed at NCBI GEO (accession number: GSE106866).

Statistics

Data and statistical analysis was performed using GraphPad PRISM 7 software. Pairwise comparisons were calculated using a two-tailed Student’s t-test comparing w¹¹¹⁸ controls and DHR78 mutants (Fig. 1D) or btl-rescued DHR78 mutants with genetically-matched btl>DHR78 controls (Figs. 2C, 3). Error bars are ±1 SEM unless otherwise noted. Box plots display the upper and lower quartiles and the median, with the bars representing the 5–95% confidence intervals. Feeding data is presented as a dot plot showing individual data points and the median.