Octopamine receptors in Drosophila larvae: expression, development, and behavior

Octopamine is involved in a variety of physiological and behavioral mechanisms throughout the Drosophila melanogaster life cycle. Octopaminergic neurons in both the central and the peripheral nervous system target a multitude of neurons and even non-neuronal tissues, making it challenging to analyze individual mechanisms of octopamine function. One approach to deconstructing this complex system is to examine the postsynaptic components of signal transmission, the six distinct G-protein-coupled octopamine receptors. Here, we compiled and validated a complete set of potential mutant lines of all known octopamine receptors (Oamb, Octα2R, Octβ1R, Octβ2R, Octβ3R, and Oct-TyrR) that were generated using the same genetic tool: the recently established Trojan Exon system. It integrates the Gal4/UAS binary expression strategy while simultaneously impairing receptor function. We generated a comprehensive anatomical map of receptor expression in the larva and, at the same time, analyzed the function of individual octopamine receptors during larval development, chemosensory perception, and locomotion. All octopamine receptors are expressed in the central and peripheral nervous system. Octβ2R stands out with a pronounced expression in the somatic muscles. We also observed a previously undescribed role of Octβ1R, Octβ3R, and Oct-TyrR in larval hatching and in the survival of larvae and pupae. Combined with tissue- and stage-specific gene expression data, this validated set of transgenic lines provides a basis for further functional studies of the octopaminergic system.

Octopamine (OA) is a biogenic amine that acts as a neuromodulator, neurotransmitter, and neurohormone in invertebrates and plays a vital role in regulating a variety of behavioral and physiological processes. Extensive research has been conducted on its functions in a range of species, including Drosophila melanogaster, Apis mellifera, and Caenorhabditis elegans, as well as on its structural analog norepinephrine in vertebrates (reviewed in David and Coulon, 1985; Roeder, 2005; Farooqui, 2012; Roeder, 2020; Mezheritskiy et al., 2024).

In Drosophila, octopaminergic cells that produce OA via Tyramine β-hydroxylase (Tβh), are found not only in the central and peripheral nervous system of both adults (Monastirioti et al., 1995; Monastirioti, 2003; Zhou et al., 2008; Busch et al., 2009; Busch and Tanimoto, 2010; Schneider et al., 2012; Pauls et al., 2018; Babski et al., 2024; Selcho, 2024) and larvae (Monastirioti et al., 1995; Monastirioti et al., 1996; Koon et al., 2011; Selcho et al., 2012; Selcho et al., 2014), but also in the adult female reproductive system (Monastirioti et al., 1996; Monastirioti, 2003; Rezaval et al., 2014). OA is versatile in its function, being involved in activity, arousal, aversion to harmful bacteria, aversion to repellents, the circadian clock, climbing activity, courtship behavior, decision-making, female fertility, ethanol preference and tolerance, feeding, flight activity, grooming, gut-brain axis, lipid metabolism, male aggression, mushroom body development, memory formation, movement endurance, muscle development, sleep, starvation response (Figure 1; Monastirioti et al., 1996; Kutsukake et al., 2000; Scholz et al., 2000; Baier et al., 2002; Lee et al., 2003; Monastirioti, 2003; Schwaerzel et al., 2003; Cole et al., 2005; Brembs et al., 2007; Certel et al., 2007; Crocker and Sehgal, 2008; Hoyer et al., 2008; Zhou et al., 2008; Lee et al., 2009; Certel et al., 2010; Crocker et al., 2010; Kula-Eversole et al., 2010; Sitaraman et al., 2010; Yarali and Gerber, 2010; Avila et al., 2012; Burke et al., 2012; Erion et al., 2012; Schneider et al., 2012; Suver et al., 2012; Zhou et al., 2012; Huang et al., 2013; Kim et al., 2013; Andrews et al., 2014; Lim et al., 2014; Luo et al., 2014; Rezaval et al., 2014; van Breugel et al., 2014; Deady and Sun, 2015; Huetteroth et al., 2015; Li et al., 2015; Yang et al., 2015; Li et al., 2016; Wang et al., 2016; Yang et al., 2016; Yu et al., 2016; Sujkowski et al., 2017; Watanabe et al., 2017; Classen and Scholz, 2018; Damrau et al., 2018; Schretter et al., 2018; Sujkowski and Wessells, 2018; Youn et al., 2018; Deng et al., 2019; Manjila et al., 2019; Sabandal et al., 2020; Yoshinari et al., 2020; Pauls et al., 2021; Zhao et al., 2021; El-Kholy et al., 2022; Nakagawa et al., 2022; Berger et al., 2024; Rohrbach et al., 2024; Bragina et al., 2025; Wu et al., 2025; Reyes et al., 2026; Wang et al., 2026). In larvae, OA plays a role in chemotaxis, feeding, learning and memory, locomotion, nociceptive escape behavior, sleep, startle response, starvation response (Saraswati et al., 2004; Fox et al., 2006; Koon et al., 2011; Koon and Budnik, 2012; Selcho et al., 2012; Zhang et al., 2013; Selcho et al., 2014; Pauls et al., 2015; Ma et al., 2016; Branch et al., 2017; Szuperak et al., 2018; Schützler et al., 2019; Li et al., 2023; Bakshinska et al., 2025; Blaum et al., 2025; Boivin et al., 2026; Franke et al., 2026). In contrast to behavior, the involvement of OA in early development, ranging from the embryo to the pupa, has been scarcely investigated.

OA can bind to six distinct G-protein-coupled receptors (GPCRs) in Drosophila, which are classified into three groups: (1) α-adrenergic-like receptors, including Oamb and Octα2R; (2) β-adrenergic-like receptors, consisting of Octβ1R, Octβ2R, and Octβ3R; and (3) a tyramine (TA) type 1 receptor, known as Oct-TyrR (Table 1). Each receptor class differs in its structure, with its classification derived from the vertebrate adrenergic receptors. Upon OA binding, different G-proteins interact with the receptors, inducing different downstream signals (Table 1). Oamb activation increases Ca 2+ levels via the Gαq-mediated phospholipase C pathway (Han et al., 1998; Balfanz et al., 2005). Interestingly, depending on the ligand concentration, the Ca 2+ response can be a single event that slowly declines (high concentration) or oscillatory (low concentration; Balfanz et al., 2005). A Ca 2+ /Calmodulin-dependent protein kinase II activity (CaMKII) in the oviduct epithelium has also been reported (Lee et al., 2009). The second α-adrenergic-like receptor, Octα2R, inhibits the adenylyl cyclase pathway via Gαi/o, leading to a decrease in cyclic adenosine monophosphate (cAMP) level (Qi et al., 2017). All three β-adrenergic-like receptors have been found to interact with the Gαs protein, which leads to an activation of the adenylyl cyclase pathway and an increase in cAMP (Octβ1R: Maqueira et al., 2005; Balfanz et al., 2005; Octβ2R and Octβ3R: Maqueira et al., 2005). Yet for Octβ1R, conflicting evidence suggests that it may reduce cAMP levels by coupling to the inhibitory Gαi/o (Koon and Budnik, 2012). The Oct-Tyr receptor can be activated by two ligands, OA and TA, resulting in two possible downstream signaling cascades: either Gαq activity leads to a Ca 2+ level increase (Robb et al., 1994) or Gαi/o leads to a cAMP level decrease (Saudou et al., 1990; Robb et al., 1994; Chatwin et al., 2003). OA is more potent to evoke the Gαq-mediated downstream pathway, while TA is more potent to evoke the Gαi/o mediated downstream pathway. The broad diversity of OA receptors (OARs) and their different downstream signaling pathways highlights the intricate nature of OA’s involvement in regulating numerous physiological processes and behaviors. This complexity also complicates efforts to link individual receptor types to specific functions, making a comprehensive analysis of OARs challenging.

Previous data on OA and OARs focused mostly on only OA, individual OARs, or incomplete sets of OARs. Here, we compiled a complete collection of transgenic lines to apply the same assays for OA and each OAR. Not only is this the first complete set for all known OARs, but they were also all generated using the same genetic tool—the Trojan Exon developed by Diao et al. (2015). The Trojan Exon lines were generated from the MiMIC collection (Minos- mediated integration cassette; Venken et al., 2011) by inserting the Trojan Exon cassette at the MiMIC docking site via recombination-mediated cassette exchange. The Trojan Exon is described as a “coding intron” since it is inserted into an intronic region and contains a splice acceptor and a splice donor, thereby masking it as an exon. Moreover, the cassette encodes a Gal4 for gene-specific targeting, which is translated separately from the endogenous protein. This integration event produces a truncated protein, the length of which is determined by the insertion site within the gene. Such truncation can compromise the functionality of the endogenous protein, while concurrently enabling precise visualization of its native expression pattern via a Gal4-dependent reporter.

With this set of Trojan Exon OA and OAR lines, we observed labeling of each OAR in the central (CNS) and the peripheral nervous system (PNS), and Octβ2R-specific signal in somatic muscles. Bulk RNA-sequencing analysis allowed us to identify additional, previously unknown expression locations, including larval lymph gland, pupal brain, eye, and salivary gland, and adult glia, olfactory projection neurons, and neuronal subsets in the protocerebral bridge. For the first time, we generated an anatomical map of Octα2R in the larva. We observed large overlapping expression patterns of different OARs in various tissues and cell populations throughout the body, from larva to adult. We found that larval hatching ability and larval survival are impaired in Octβ1R, Octβ3R, and Oct-TyrR Trojan lines. Pupal mortality is increased in Trojan Oct-TyrR. These results suggest that the involvement of OA during development is greater than previously thought. We compared our results with findings from previous studies and, together with a molecular evaluation of the Trojan Exon OA and OAR lines, provide a complete and validated set of lines for analyzing the OA system beyond this study.

Fly strains

Fly strains were reared on standardized cornmeal medium at 25 °C, 65 % relative humidity, and in a 14:10 h light:dark cycle. The animals were consistently staged by transferring adults regularly to fresh vials, and we controlled the number of parent animals and offspring per vial to minimize variability due to crowding or nutritional differences. Trojan Exon octopamine and octopamine receptor lines were obtained from the BDSC: Oamb (RRID:BDSC_67506), Octα2R (RRID:BDSC_67636), Octβ1R (RRID:BDSC_76677), Octβ2R (RRID:BDSC_67511), Octβ3R (RRID:BDSC_76680), Oct-TyrR (RRID:BDSC_77735) and Tβh (RRID:BDSC_76655). The first three lines carried the TM3, Sb, 1 Ser 1 balancer chromosome with no visible marker in larvae; accordingly they were re-balanced over TM6B, Hu, e 1, Tb 1 using the double balancer w –;Bl/CyO, dp lvl, pr 1, cn 2; TM2, Ubx, e S /TM6B, Hu, e 1, Tb 1 obtained from the Galizia lab (see also Gonzalez et al., 2016). Control lines were CantonS (WT) and yellow white (y 1 w*). Only homozygous individuals were used for experiments. Each experiment was performed at 25 °C.

Oviposition

Freshly hatched or max. 1 day old female flies of each genotype were paired with WT male flies in a 1:1 ratio to ensure a high copulation probability and maximize the yield of laid eggs. After a mating period of 72 h, individual female flies were placed on oviposition plates (Sarstedt, 82.1135.500, Ø 35 mm) filled with 2.5% agarose (w/v) (VWR-Life Science, CAS 9012-36-6), 3% sucrose (w/v) (Roth, CAS 57-50-1) and 30% store-bought apple juice (Sachsen Obst, 100% fruit). The oviposition plates were topped with 2% activated baker’s yeast solution. 24 h later, the number of laid eggs was counted. The sample size is 16 per genotype (one female per n).

Viability

The oviposition assay was repeated with female and male flies of the same genotype and up to 4 female flies per oviposition plate. The increased number of females ensured a proper number of eggs and later larvae to assess larval and pupal mortality. Female flies were allowed to lay eggs for 24 h, and after another 24 h, the number of hatched larvae was counted. From these plates, 30 hatched larvae were then transferred to standard food vials and monitored closely. The number of pupae and emerging flies was quantified three times a day over a period of 16 days. We did not distinguish between the sexes of the animals, as we are not aware of any sex-specific differences in octopamine and octopamine receptor expression at the larval stage. Hatching rate describes the number of hatched larvae per total number of eggs. Larval mortality is the percentage of larvae that did not enter the pupal stage. Pupal mortality is the percentage of flies that did not hatch from the pupa. Sample size is 16 per genotype and 30 larvae per n, resulting in 480 larvae per genotype, similar to previous studies (Schou, 2013; Lee et al., 2015; Orsted et al., 2019).

Locomotion assay

The protocol used was modified from Bhatt and Neckameyer (2013). Third-instar foraging-stage larvae were washed with tap water to remove residual food and transferred to a Petri dish (Sarstedt, 82.1472, Ø 92 mm) filled with 2.5% agarose (w/v). After 30 s of acclimatization, single larvae were filmed for 75 s at a rate of 20 fps with an area scan camera (Basler, 106909, lens: 25 mm) and the software Pylon viewer (v 6.3.0; Basler). For a total duration of 1 min, each peristaltic wave was counted for 15 larvae per genotype. This experiment was performed blindly.

Feeding assay

Feeding behavior was assessed via the rate of mouth hook contractions of third-instar foraging-stage larvae on a Petri dish filled with 2.5% agarose (w/v) and topped with 5 ml of 2% activated baker’s yeast solution. Yeast solution prevented larvae from moving around, ensuring that the recorded mouth hook contractions were a result of feeding behavior. Before the experiment, larvae were washed with tap water to remove residual food. Larvae were allowed to adjust to the new surroundings for 30 s and afterwards were filmed for 75 s with the same setup and software as for the locomotion assay. Each mouth hook contraction over a period of 1 min was counted for 15 individual larvae per genotype. This experiment was performed blindly.

Olfactory and gustatory preference tests

Experiments were performed as described in Apostolopoulou et al. (2013). For the gustatory preference test, Petri dishes were filled with a thin layer of pure 2.5 % agarose (w/v) on one half and 2.5 % agarose (w/v) mixed with the appetitive stimulus [fructose (Roth, 57-48-7; 2 M) or arabinose (Sigma Aldrich, 10323-20-3; 2 M)] on the other half. The middle of the Petri dish was marked as a 10 mm-wide stripe separating the sugar and pure agarose sides. In this area, 30 larvae were initially placed and were given 5 min of free motion before being counted. Three groups emerged: larvae on the sugar side (#sugar), larvae on the pure agarose side (#agarose), and larvae in the middle. The total number of larvae resulted from the number of larvae in each group (#total). The preference index was calculated as follows: #sugar minus #agarose divided by #total.

Olfactory preference tests were performed on a Petri dish filled with 2.5% agarose (w/v) and one empty custom-made Teflon container (4.5 mm diameter) opposite a second container filled with 10 μl of the odors amyl acetate [Sigma Aldrich, 628-63-7; diluted 1:50 in paraffin oil (Sigma Aldrich, 8012-95-1)] or 1-octanol (Sigma Aldrich, 111-87-5; undiluted). The odor containers were closed with a perforated lid. 30 larvae were transferred to the middle of the Petri dish and were given 5 min of free motion before being counted. Again, three groups emerged: larvae located on the odor side (#odor), larvae on the side with an empty container (#empty container), and larvae in the middle (10 mm wide stripe in the middle of the Petri dish). The total number of larvae resulted from the number of larvae in each group (#total). Preference index is calculated as follows: #odor minus #empty container divided by #total. Positive preference indices represent an attraction toward a gustatory or olfactory stimulus. Sample size is 16 per genotype and 30 larvae per n, resulting in 480 larvae per genotype.

Anatomical expression analysis

Trojan Exon lines were crossed with 20xUAS-6xmCherry/CyO; nsyb-LexA, LexAop-GFP to investigate native fluorescence signal, omitting the need for immunolabeling.

First-instar larvae were collected in phosphate-buffered saline (PBS; Sigma Aldrich, P4417) and starved for 1 h. Afterwards, they were placed in 4 % bleach for 10 min, rinsed twice with H 2 O and incubated for another 5 min in H 2 O. Larvae were fixed in 4% paraformaldehyde in PBS (PFA; Sigma Aldrich, F8775) for 15 min at room temperature and 15 min on ice. A washing step included incubation in 3 % Triton-X in PBS (PBT; Sigma Aldrich, X100) for 15 min, repeated three times. Finally, larvae were washed three times in PBS for 15 min. Specimens were embedded in VECTASHIELD ® PLUS (Vector Laboratories, H-2000).

Third-instar larvae were collected in PBS and starved for 2 h. Larvae were then heat-fixed by immersion in 70 °C water for 1 min and subsequent cooling in cold water. After drying, samples were mounted in Vectashield and imaged immediately.

Brains of third-instar wandering-stage larvae were dissected in PBS on ice and fixed in 4% PFA for 25 min under vacuum at room temperature. Afterwards, brains were washed six times in 3% PBT and six times in PBS. Brains were mounted on poly-L-lysine (Sigma Aldrich, P1524-25MG) coated coverslips and embedded in Vectashield. Specimens were scanned at a Confocal laser scanning microscope LSM800 by Zeiss with ZEN 2.3 software and processed for publication with ImageJ (National Institutes of Health, Bethesda, Maryland). 1 Co-labeling was assessed by counting overlapping signal in cell bodies in the calyx and innervated glomeruli of the antennal lobe of the OAR (mCherry) and reference channel (GFP). The mean was calculated from six hemispheres per genotype, and the standard error of the mean (SEM) was calculated as standard deviation divided by the square root of the sample size.

Gene models for Trojan Exon lines

Gene sequences, transcripts, and annotations for Tβh and octopamine receptor genes were obtained via the FlyBase Sequence Downloader (Öztürk-Çolak et al., 2024; release: FB2019_05). Conserved domain predictions were performed with the NCBI Conserved Domain Database (Supplementary Table 1) using the nucleotide sequence of each isoform of the genes. The Trojan Exon Cassette Sequence was acquired from Supplementary Information from Diao et al. (2015). To annotate the sequence of the Trojan Exon, we followed Diao et al. (2015), supplementing this with data from previous publications (Venken et al., 2011; Diao and White, 2012; Diao et al., 2015). The locations of the transgenic insertions can be found in the respective FlyBase entries. Insertion and visualization of the Trojan Exon construct sequence into the receptor gene sequences were done using Benchling (Supplementary Table 1).

Molecular evaluation—PCR and qualitative RT-PCR

Genomic DNA was isolated using the kit NucleoSpin ® Tissue; DNA, RNA, and protein purification; Genomic DNA from Tissue (Macherey-Nagel, 740952.250) following the manufacturer’s protocol. RNA was isolated using TRIzol Reagent (Invitrogen, 15596026) following the manufacturer’s protocol. The second step comprised DNase digestion (Roche, 04 716 728 001) followed by RNA purification using the RNeasy ® MinElute ® Cleanup kit (Qiagen, 74204), which was performed as stated in the manual. The RevertAid First Strand cDNA Synthesis kit (Thermo Scientific, K1621) was used for cDNA synthesis (performed as in the manufacturer’s protocol).

Taq PCR Master Mix (QIAGEN, 201443) was used to amplify target regions according to the manufacturer’s protocol. Primer sequences can be found in Supplementary Table 1. PCR products were separated during gel electrophoresis [1% agarose gel (Sigma Aldrich, A9539), 1xTBE buffer, MassRuler DNA Ladder Mix (Thermo Scientific, SM0403)]. Gel extraction of amplified products was performed with QIAquick Gel Extraction Kit (QIAGEN, 28706) and was sequenced by Microsynth Seqlab ®. Online tools and websites used are listed in Supplementary Table 1.

Quantitative real-time RT-PCR

Total RNA from third-instar larvae was isolated using QIAzol (Qiagen, 79306) followed by DNA digestion using RNase-Free DNase Set (Qiagen, 79256) and purification using RNeasy Mini Kit (Qiagen, 74104); all steps were performed according to the manufacturer’s protocol. 250 ng RNA was reverse transcribed using SuperScript™ III First-Strand Synthesis SuperMix for qRT-PCR (Invitrogen, 11752050). Gene expression was measured by quantitative real-time RT-PCR and the TaqMan methodology (Dm02139710_g1 for Octβ2R and Dm01825538_m1 for Oct-TyrR). Fluorescence was detected on an ABI PRISM 7500 sequence detector (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s instructions (Applied Biosystems). In short, we used the standard detection method. The conditions were as follows: 50.0 °C for 2 min, 95.0 °C for 10 min, and 40 cycles of 95.0 °C for 15 s and 60.0 °C for 1 min. Data were collected at the 60.0 °C step in each of the 40 cycles. The total reaction volume was 10 μl, and a standard curve was used on each plate. mRNA expression was calculated relative to the mRNA expression of Rpl32 (Applied Biosystems; assay Dm02151827_g1 for Rpl32) and to y 1 w* control larvae.

Bulk RNA-seq analysis

Raw sequencing reads were obtained from the European Nucleotide Archive. 2 The RNA-seq runs used in this study are listed in Supplementary Table 2. Quality control of raw FASTQ files was performed using FastQC (v0.11.9), and reads were pseudo-aligned to the reference transcriptome (Drosophila melanogaster, genome assembly BDGP6.54, Ensembl release 113) using Salmon (v1.10.2) in quasi-mapping mode with sequence-specific and GC bias correction enabled (–seqBias, –gcBias). Transcript-level expression estimates were imported into R (v4.3.3) using the importIsoformExpression function from the IsoformSwitchAnalyzeR package (v2.2.0). Gene-level expression values were derived by summing transcript-level TPMs for each gene. Genes with low expression were filtered before downstream analyses. Specifically, genes were retained only if they showed a gene-level expression of ≥ 5 TPM in at least 3 samples. For comparative and visualization purposes, gene expression values were summarized as median TPM per tissue and developmental stage. TPM values were log-transformed where indicated for improved visualization. Gene expression levels were compared between individual tissues and all remaining tissues within the same developmental stage. Expression differences are reported as log2 fold change (log2FC), where positive values indicate higher expression in the focal tissue. Statistical significance was assessed using non-parametric Wilcoxon tests. P -values were adjusted for multiple testing using the Benjamini–Hochberg procedure, and false discovery rate (FDR) values are reported (Supplementary Table 3). FDR < 0.05 was considered statistically significant.

Statistical analysis

GraphPad Prism 8.0 (GraphPad Software Inc.) was used for statistical analysis and visualization. First, the normal distribution of data points was assessed with the Shapiro-Wilk test. If a normal distribution was assumed, a one-way ANOVA followed by a Tukey honestly significant difference post hoc test for pairwise comparison between more than two groups was applied. If assumptions of normal distribution were not met, data were analyzed with a Kruskal-Wallis test and Dunn’s multiple pairwise comparison for more than two groups. Experimental groups were compared to the genetic background line, y* w*. The Wilcoxon signed-rank test was chosen to compare groups against the chance level. Significance level was set at 0.05. Box plots are shown with 25 and 75% quantiles as box boundaries, while whiskers display the complete data set (min to max values). The median is indicated as the middle line, and each dot represents a data point. Bar plots show the mean and standard deviation of the data set. Each dot represents a data point. The following statistical tests were applied: Figure 2: For A-D, data were analyzed with a Kruskal-Wallis test and Dunn’s multiple pairwise comparison. Figure 3: For A, B, C, and F, data were analyzed with a one-way ANOVA followed by a Tukey honestly significant difference post-hoc test. For D and E, data were analyzed with a Kruskal-Wallis test and Dunn’s multiple pairwise comparison. For A, B, E, and F, the Wilcoxon signed-rank test was applied to compare groups against the chance level.

In order to use the Trojan Exon OA and OAR lines for the analysis of the OA system, we investigated whether known phenotypes could be reproduced. We tested for I) the prominent role of Tβh, Oamb, and Octβ2R in female sterility (Monastirioti et al., 1996; Lee et al., 2003; Monastirioti, 2003; Lim et al., 2014; Rezaval et al., 2014; Deady and Sun, 2015; Li et al., 2015), II) the role of Octβ3R in pupation formation (Ohhara et al., 2015) and III) disrupted larval locomotion in Tβh, Octβ1R, Octβ2R, and Oct-TyrR larvae (Saraswati et al., 2004; Fox et al., 2006; Koon and Budnik, 2012; Selcho et al., 2012). We integrated these tests in the (A) developmental assays, comprising female fecundity, larval hatching, as well as larval and pupal survival (Figure 2); (B) behavioral assays, comprising chemosensation, locomotion, and feeding behavior (Figure 3); and (C) anatomical mapping, comprising the central and peripheral nervous system as well as non-neuronal tissue (Figures 4–6). A molecular analysis supported the evaluation of the Trojan Exon OA and OAR lines (Figure 7). In addition to the developmental, behavioral, and anatomical experiments with OA and OAR lines, transcript profiling based on bulk RNA-sequencing data revealed previously unknown expression locations and demonstrated quantitative differences in expression across distinct developmental stages, tissues, and cell types (Figure 8).

Octopamine receptors contribute to the animal’s viability—from egg to adult

First, we performed a general assessment of the animal’s viability over the different developmental stages. Embryonic mortality was examined based on the ability of the first-instar larva to hatch from the egg. Larval and pupal mortality were used to evaluate survival rates. Additionally, female sterility was analyzed by applying an egg-laying test, in which the number of eggs laid by individual flies within 24 h was counted. To exclude potential effects arising from male reproductive system defects (as has been reported for Oamb; Zhou et al., 2012; Luo et al., 2014), Trojan OAR females were mated with WT males. A significantly reduced number of eggs was observed for Trojan Oamb and Octβ2R compared to the genetic control. In contrast, no reduction was seen for Trojan Tβh, Octβ1R, Octβ3R, and Oct-TyrR (Figure 2A). Due to the limited number of experimental animals resulting from the small number of eggs laid (Figure 2A) and the low percentage of homozygous animals (Supplementary Table 4), further analysis of larval development in Trojan Oamb, Octα2R, and Octβ2R was not possible. For all other groups, we assessed the ability of larvae to hatch from the egg and calculated a hatching rate, which is defined as the number of hatched larvae divided by the total number of eggs. This time, females were mated with males of the same genotype to analyze only homozygous larvae. The hatching rate was found to be reduced in Trojan Octβ1R, Octβ3R, and Oct-TyrR compared to control animals, while no reduction was observed in Tβh (Figure 2B). Subsequently, the survival of larvae and pupae was monitored. Larval mortality was assessed as the percentage of larvae that did not enter the pupal stage. A significant increase was observed in Trojan Octβ1R, Octβ3R, and Oct-TyrR (Figure 2C). Pupal mortality, which indicates the percentage of animals unable to hatch from the pupa, showed that only Trojan Oct-TyrR had a reduced number of hatched flies, reflected in an increased pupal mortality (Figure 2D). Raw data and statistical analysis for all data presented in Figure 2 can be found in Supplementary Table 5. So, on the one hand, the Trojan Oamb and Octβ2R lines reproduced the well-known female sterility phenotype; on the other hand, for the first time, it was shown that Octβ1R and Oct-TyrR are involved in larval hatching and larval survival. Even the survival of pupae was impaired in Trojan Oct-TyrR. We also observed a decreased survival rate of larvae in Trojan Octβ3R, which could be related to the reported impairment of pupation (Ohhara et al., 2015). Nevertheless, for the first time, we showed that Octβ3R is involved in larval hatching.

Analysis of larval foraging behavior

In addition to physiological mechanisms, changes in behavior can also explain the decreased viability of Trojan OAR lines observed in Figure 2. Therefore, we aimed to analyze the role of the OARs in one such behavior—foraging. During the larval stage, energy requirements are high due to a body mass increase of about 200 times (Bakker, 1959). Consequently, foraging plays an essential role in the life of larvae, ensuring their survival and progression into metamorphosis (Bakker, 1961; Ohnishi, 1979). In this study, we subdivide the larval foraging behavior into (I) the perception of olfactory cues of a potential food source (Figures 3A,B), (II) locomotion on pure agarose (Figure 3C), (III) the ability of food uptake (Figure 3D), and (IV) the gustatory preference for food sources that are generally indicative of high energy supply (Figures 3E,F). Olfactory cues can provide a first indication of a potential food source and help the larva to navigate in their environment and initiate and/or maintain contact with a substrate (reviewed in Cobb, 1999; shown in Gomez-Marin et al., 2011; Gershow et al., 2012). Here, we tested the preference for the two odors, amyl acetate and 1-octanol, which can be associated with a food source in an olfactory associative learning and memory assay (reviewed in Gerber and Stocker, 2007; shown in Selcho et al., 2009; Pauls et al., 2010; Schleyer et al., 2011; Apostolopoulou et al., 2013; Weber et al., 2023a; Weber et al., 2023b). We used the same experimental setup and odor concentrations as in Apostolopoulou et al. (2013). Larvae from each Trojan Exon line showed a preference for the two odors, similar to control animals (Figures 3A,B). It is important to note that due to the small number of laid eggs, which resulted in a limited number of experimental animals, Trojan Oamb and Octβ2R could not be included in these mass assay experiments.

After perceiving an appetitive odor cue, the crawling behavior is directed more specifically toward a potential food source (Fishilevich et al., 2005; Louis et al., 2008). In a locomotion assay, we evaluated the crawling ability by quantifying the peristaltic waves (pw) as a read-out of rhythmic motor patterns (Gjorgjieva et al., 2013; Liu et al., 2023). Tested animals were neither exposed to external stimuli nor treated in any other way. While Trojan Octβ1R exhibited an increased number of pw/min, Trojan Octβ2R showed significantly less pw/min than the control animals (Figure 3C). Given that these assays analyze individual larvae, it was possible to include the relatively small number of surviving animals of Trojan Oamb and Octβ2R.

The next assay focused on assessing the feeding behavior of the larvae. In contrast to crawling behavior, feeding primarily relies on contractions of the cephalopharyngeal muscles, which are separate from the somatic segment muscles in the abdomen (Schoofs et al., 2010). However, mouth hook contractions (mhc) are also required for crawling behavior. In the feeding assay, larvae were placed in a nutrient-rich environment that eliminated the need to search for food. Mouth hook contractions serve as a straightforward indicator of feeding behavior (Bhatt and Neckameyer, 2013). Most Trojan Exon lines showed no change in feeding rate; only Octβ2R larvae had a significantly reduced number of mhc per minute compared to control animals (Figure 3D).

Gustatory preferences can indicate whether a larva is able to identify beneficial sources of nutrition, such as sweet and high-calorie food (Rohwedder et al., 2012). In a choice assay, larvae were presented with sugar on one side and tasteless, non-nutritious agarose on the other side. We used two different sugars: (I) fructose, which is sweet and has a high caloric value, while (II) arabinose is only sweet but non-nutritious. All Trojan Exon lines preferred the sugars over the neutral agarose (Figures 3E,F). However, Trojan Oct-TyrR showed a significantly reduced preference for arabinose compared to the control. Please note that due to the small number of laid eggs, which resulted in a limited number of experimental animals, Trojan Oamb and Octβ2R were not included in these mass assay experiments. Raw data and statistical analysis for all data in Figure 3 can be found in Supplementary Table 6.

Expression pattern of Trojan Exon OAR lines in the larva

Utilizing the simultaneously driven Gal4 expression in the Trojan Exon lines, we anatomically mapped the expression of each OAR in neuronal and non-neuronal cells. Two protocols were applied, one to evaluate the respective expression throughout the body (Figure 4) and the other in the CNS (Figure 5) of third-instar larvae. For each receptor, the head (pseudocephalon and up to thoracic segment 3), abdominal segment 3 and/or 4, and abdominal segment A8/9 are shown (Figure 4). We map only unambiguous signals, mostly at the body surface, but the labeling of deep-seated tissues cannot be excluded. The first panel in Figure 5 shows an overview of the CNS. Close-ups of one hemisphere (HEMI), the subesophageal zone (SEZ), the antennal lobe (AL), the calyx (CA), and the mushroom body (MB) are shown from left to right. Below each panel is a grayscale image only of the channel showing receptor expression. For the CA and MB, additional panels show receptor expression and reference signal separated from each other. To evaluate co-labeling of OARs with the CA and the AL, we counted the number of cell bodies of intrinsic Kenyon Cells (KC) in the CA and glomeruli of the AL with OAR signal, utilizing six hemispheres per genotype (Supplementary Table 7).

Trojan Oamb showed labeling in the dorsal organ (DO), labial organ (LO), terminal organ (TO), and ventral organ (VO) in the pseudocephalon (PCE) (Figure 4). Each segment, from thoracic segment 1 (T1) to abdominal segment 8/9 (A8/9), featured signal in dendritic arborization neurons (DAs) as well as motor neurons (MNs). Additional labeling was found in bipolar dendrite neurons (BDs), chordotonal organs (ChOs), external sense neurons (ESs), and tracheal dendrite neurons (TDs) in the abdomen, as well as in the terminal sensory cones (TSC) in A8/9. Trojan Oamb was observed to be the only OAR expressed in TDs. The CNS showed labeling from the hemispheres (Figures 5AI,II) to the subesophageal zone (Figures 5AIII,IV). The larval AL is organized in 21 glomeruli (Ramaekers et al., 2005). Approximately 18 ± 1 glomeruli of the AL showed labeling (Figures 5AV,VI). The MB consists of approximately 600–650 intrinsic KC per hemisphere, whose axons form the lobes of the MB (Ito and Hotta, 1992; Ramaekers et al., 2005). About 34 ± 2 KC cell bodies in the CA showed double-labeling (Figures 5AVII–XIV).

In Trojan Octα2R, both TO and VO labeling can be found in the PCE (Figure 4). While the DO itself is not labeled, signals were detected in a group of sensory neurons belonging to the dorso-lateral group of the TO, which have their cell bodies within the associated ganglion (DOG). DAs, ESs, and MNs showed Trojan Octα2R expression in both thoracic and abdominal segments. ChOs were solely labeled in the abdomen. In terminal abdominal segments A8/A9, TSC, and spiracle sense organ (SSO) were labeled. Multiple cells in the CNS showed a signal. Only a few KC (approximately 16 ± 1), likely born embryonic (Pauls et al., 2010), were found to express Octα2R (Figures 5BVII–XIV).

Trojan Octβ1R showed expression in DO, TO, and VO in the PCE and in ChOs, DAs, ESs, and MNs in each thoracic and abdominal segment (Figure 4). Additionally, BDs were labeled in the abdominal segments, and TSCs and SSO were observed to be labeled. In the CNS, we found many cell bodies labeled, from the hemispheres to the VNC. About 20 ± 0 glomeruli of the AL and approx. 14 ± 1 KC cell bodies showed Trojan Octβ1R signal (Figures 5C,CI–XIV,CV).

Trojan Octβ2R signals can be found in somatic muscles of the body wall across each thoracic and abdominal segment (Figure 4). In the PCE, two intersegmental muscles showed labeling: muscle 12, which runs from dorsal T1 to ventral PCE, and muscle 17, which extends from laterofrontal PCE to dorsal T2. In the prothorax (T1), we identified dorsal longitudinal muscle 36 and ventral longitudinal muscles 31–35, which are associated with the mouthparts (Hooper, 1986). In the mesothorax (T2), signal was observed in dorsal longitudinal (muscle 1–4, 9, 10), dorsal oblique (muscle 5 and 19; muscle 20 is probably covered by muscle 19), lateral transverse (muscle 8, 18, 21, 22, 24; muscle 23 is probably covered by muscle 22 and 24) and ventral longitudinal muscle (muscle 14). Only the dorsal oblique muscle 30 appears to be not labeled. In the metathorax (T3), we identified dorsal longitudinal (muscle 1–4, 9, 10), dorsal oblique (muscle 5, 19, 20), lateral transverse (muscle 8, 18, 21, 22, 24; muscle 23 is probably covered by muscle 22 and 24), ventral longitudinal (muscle 6, 7, 12, 13, 14) and ventral oblique muscles (muscle 15). Only ventral acute muscle 26 and dorsal oblique muscle 30 showed no signal. Each muscle of the body-wall in abdominal segments A1-7 showed expression in the Trojan line and can be summarized as follows since the muscle pattern is identical for each segment: dorsal longitudinal (muscle 1–4, 9, 10), dorsal oblique (muscle 5, 11, 19, 20), lateral transverse (muscle 8, 18, 21–25), ventral longitudinal (muscle 6, 7, 12–14, 30), ventral oblique (muscle 15–17, 28) and ventral acute (muscle 26, 27, 29). Due to their position, the labeling of muscles 6, 7, 15, 16, 17, and 29 is shown in Supplementary Figure 1A. Muscle 25 is the only muscle not present in A1 (Zarin et al., 2019) and hence was not found during the expression analysis. In segment A8/9, dorsal longitudinal muscle 40, dorsal transverse muscle 48 (Supplementary Figure 1B), lateral transverse muscles 46, 47, 49, and 50, ventral longitudinal muscles 41 and 45, as well as ventral acute muscle 42, were identified. These muscles can be grouped into I) associated with the terminal spiracle (muscles 40, 41, and 42) and II) associated with the anal division (muscles 45, 46, 47, 49, and 50), depending on their position (Wipfler et al., 2013). The ring-shaped muscle 43, which surrounds the terminal spiracles, showed no labeling. According to Wipfler et al. (2013), we found a muscle labeled as number 5, although it was not described in detail. Due to the uncertainty of their position, the identity of two additional muscles (44 and 51) was not determined. Muscles 44 and 51 have been previously described but not visualized (Wipfler et al., 2013). Muscle identity was determined based on the work of Wipfler et al. (2013) for PCE and A8/9, Hooper (1986) for T1-3, and Zarin et al. (2019) for A1-7. Additionally, we identified a signal in at least one sensory neuron in the TO (Supplementary Figure 1C). While additional expression may also occur in other sensory neurons, motor neurons, or non-neuronal cells, the strong signal in the somatic muscles could be overshadowing these other signals. The CNS of Trojan Octβ2R showed signal in the VNC and the hemispheres (Figures 5D,DI,II). We found single cell bodies around the AL but no innervation (Figures 5DV,VI), and a cell cluster lateral to the vertical lobe of the MB, but no labeled KC (Figures 5DXI–XIV).

In Trojan Octβ3R, signals were observed in both the DO and TO within the PCE (Figure 4). Expression was detected in the thoracic and abdominal segments in certain neuron types, including BDs, ChOs, DAs, ESs (only thorax), and MNs. In segment A8/9, ChOs and MNs, as well as SSO and TSCs, were labeled. In the CNS, the expression signal encompassed all segmental ganglia within the VNC and the hemispheres (Figures 5E,EI,II). The SEZ showed cell bodies along the midline and laterally toward the VNC (Figures 5EIII,IV). While we observed innervation of the AL (about 20 ± 0 glomeruli; Figures 5EV,VI), many KCs in the CA and the lobes of the MB were distinctively labeled (approx. 190 ± 13; Figures 5EVII–XIV). Additional cell bodies in the hemispheres, around the MB, expressed Trojan Octβ3R (Figures 5EI,II,XI–XIV).

In thoracic and abdominal segments, we observed ChOs, DAs, ESs, and MNs in Trojan Oct-TyrR larvae (Figure 4). Signal in the dorsal multidendritic neuron 1 (DMD1) was observed in the abdomen, and other OARs were not expressed in this particular neuron type. Similar to most OARs, Oct-TyrR showed expression in TSCs in segment A8/9. The CNS showed the most specific expression pattern among the Trojan Exon OAR lines. Cell bodies in the abdominal segments were positioned laterally, while the thoracic segments had additional cell bodies along the midline (Figure 5F). In the SEZ, we found cell bodies specifically near the esophagus (Figures 5FIII,IV). The hemispheres showed only little Oct-TyrR signal and no distinct labeling of the AL or MB (Figures 5FI,II,V–XIV).

A summary of the expression locations throughout the entire body and the CNS can be found in Figure 6. Additional whole-body scans of first-instar larvae show similar expression patterns compared to the third-instar larvae (Supplementary Figures 2A–F, including a summary in Supplementary Figure 3). For ChOs, TSCs, and somatic muscles, the same cellular composition could be found. For DAs, DOs, ESs, MNs, SSO, and TOs, the first-instar larval organizational pattern is maintained in third-instar larvae, though supplemented by additional Trojan OAR lines (DAs: Octβ1R, Oct-TyrR; DO: Octβ1R; ESs: Oct-TyrR; MNs: Octα2R, Octβ1R, Oct-TyrR; SSO: Octβ1R; TOs: Octβ2R). It remains unclear whether these differences represent developmental changes in OAR expression or technical limitations.

Molecular analysis reveals the absence of Trojan Exon in the Tβh line

Tβh lines are known to exhibit severe egg laying impairments as well as larval locomotor deficits (Monastirioti, 2003; Saraswati et al., 2004; Selcho et al., 2012), prominent behavioral phenotypes that were not observed in Trojan Tβh. Moreover, during expression pattern analysis, no signal was observed in the brain, which contradicted an earlier study that found labeling throughout the CNS, applying immunohistochemistry (Selcho et al., 2012). In Trojan Tβh, the Trojan Exon is located in the first half of the Tβh gene locus (based on FlyBase entry ID FBti0195387), resulting in a premature truncation of the protein (Figure 7A). The conserved domains DOMON (dopamine beta-monooxygenase N-terminal; yellow) and the Copper type II ascorbate-dependent monooxygenase (blue) were predicted to be located downstream of the Trojan Exon insertion site (Supplementary Table 8) and are crucial for the enzymatic function of Tβh (Figure 7A). It can therefore be assumed that the truncated Trojan Tβh protein is non-functional, which is why we expected to find previously known phenotypes of Tβh lines. Thus, we performed targeted PCR to analyze the Trojan Exon on the genomic level and found that the Trojan Tβh line lost the cassette. According to the published insertion site of the Trojan Exon in the Tβh gene locus, MI04010, we used a primer pair that surrounded the Tβh MI04010 docking site. In the case of the presence of the Trojan Exon, we expected an amplified product of around 4 kb. But, as shown in Figure 7AI, the size of the amplicon was similar to that of the WT and the y 1 w* controls. Sequencing of the amplified PCR product revealed no differences between Trojan Tβh and the WT Tβh gene locus of the reference genome acquired from FlyBase (Supplementary Figure 4). We repeated the targeted PCR with a different copy of the Trojan Tβh fly line and received identical results (data not shown).

Based on these results, we continued analyzing the Trojan Exon OAR lines to ensure the correct insertion of the Trojan Exon. This time, a different primer strategy was applied: based on the MiMIC docking sites (Oamb: MI12417, Octα2R: MI10227, Octβ1R: MI05807, Octβ2R: MI13416, Octβ3R: MI06217, Oct-TyrR: MI03485), one primer was located in the genomic region and one in the Trojan Exon (see schematic in Figure 7 for each OAR). Due to the size of the amplified products, sequencing results, and comparison with the reference genome acquired from FlyBase, we conclude that the location of the Trojan Exon in the Oamb, Octα2R, Octβ1R, Octβ2R, Octβ3R and Oct-Tyr receptor lines is identical with the corresponding MiMIC docking sites (Figures 7BI,CI,DI,EI,FI,GI; and Supplementary Figures 5–12). Only for Trojan Octβ2R, the amplicon size was larger than expected (expected: 1072 bp, observed: ∼ 3,000 bp; Figure 7EI). A second PCR at the 3′ end of the Trojan Exon in the Octβ2R gene locus showed a similar result (expected: 1366 bp, observed: ∼1500 bp; Figure 7EII). Sequencing revealed that the additional parts were not endogenous to the Octβ2R gene, and BLAST analysis showed no distinct origin (Supplementary Figures 10–12 and Supplementary Table 9, 10). We further analyzed if the additional sequences upstream of the 5′ end of the Trojan Exon were transcribed and could affect the formation of the Trojan Exon. A schematic (not to scale) shows the unknown additional parts (∼ 2,000 bp) right upstream of the Trojan Exon at the genomic level (Supplementary Figure 13A). In a qualitative RT-PCR, primers were designed to bind in the upstream Exon 7 and in the Trojan Exon. If the inserted sequences were excluded from transcription, an amplicon of 532 bp was expected. The observed PCR product, visualized below the schematic, migrated between 500 and 600 bp. Subsequent sequencing of this fragment confirmed the absence of the additional sequences in the Octβ2R mRNA of the Trojan Octβ2R line (Supplementary Figure 14). Moreover, quantitative real-time RT-PCR was performed to analyze Octβ2R mRNA levels. We applied the TaqMan methodology with a probe binding downstream of the inserted Trojan Exon (Supplementary Figure 13A). Only a fraction (∼6%) of Octβ2R mRNA was detected compared to the control, showing the decreased expression of this OAR in the Trojan Exon line (Supplementary Figure 13C). Therefore, although additional sequences are present in the genomic locus, our data suggest that they are spliced out and are unlikely to cause off-target effects at the transcript level.

In each tested line, the Trojan Exon is inserted into the first half of the respective OAR gene (Figure 7), suggesting that the Trojan Exon-induced protein truncation leads to the loss of major structural parts of the receptor. Conserved domain prediction analysis revealed that the conserved transmembrane domains of the GPCRs are affected (Figure 7 and Supplementary Table 8), suggesting that the OARs are non-functional and thus that the Trojan Exon OAR lines are suitable for investigating the OAR system. Only for Oct-TyrR, the insert location of the Trojan Exon is in an intron in the non-coding 5′ UTR. Since the Trojan Exon must be inserted into an intron between two coding exons, the intended effect of truncating the Oct-TyrR protein and thereby disrupting its functionality is questionable. But quantitative real-time RT-PCR showed that no Oct-TyrR transcript was detectable in the Trojan Exon line, indicating that the Oct-TyrR protein is absent (Supplementary Figures 13B,C). Further, due to the insertion location in an intron in the 5′ UTR, the expression pattern obtained during anatomical mapping may not correspond to the endogenous pattern of Oct-TyrR.

Overall, insertion sites were confirmed in most Trojan Exon OARs, making these lines suitable genetic tools for interfering with the OA system. Even though the insertion site of the Trojan Exon in Octβ2R is compromised, the line seems to be valid based on our mRNA analysis. For Trojan Oct-TyrR, we found that the targeted protein disruption takes place even if it may not have been in the intended manner. But due to the insertion location in an intron in the 5′ UTR, Gal4-induced expression may not correspond to the endogenous Oct-TyrR pattern. Only Trojan Tβh cannot be used for further studies because it lacks the Trojan Exon.

Gene expression data

All results obtained so far are based on experiments with Trojan Exon lines. In a different approach, we utilized bulk RNA-sequencing data retrieved from the European Nucleotide Archive to focus on OAR expression location and quantity in third-instar larvae, pupae, and adults, each in different tissues and cell populations (Figure 8 and Supplementary Table 3).

In third-instar larva, expression of Oamb (log2FC = 2.29, FDR < 0.0001), Octα2R (log2FC = 3.40, FDR < 0.0001), Octβ3R (log2FC = 2.88, FDR < 0.001), and Oct-TyrR (log2FC = 3.55, FDR < 0.0001) was more pronounced in the VNC than the hemispheres, in contrast to Octβ1R (log2FC = 0.00, FDR = 0.53) that was detected throughout the brain (hemispheres and VNC). Increased levels of Octβ1R (log2FC = 3.31, FDR = 0.09) and Octβ3R (log2FC = 2.63, FDR = 0.02) were observed in eye imaginal disc nuclei, while Octβ2R (log2FC = 0.63, FDR = 0.006) showed increased expression in third-instar larva muscle compared to the other OARs and other cells and tissues. A yet unknown expression location was found for Oct-TyrR (log2FC = 2.85, FDR < 0.0001) in the lymph gland.

In the pupa, every OAR was detected in the brain and in a neuronal subset, the fru P1-expressing neurons of the male. The pupal eye showed expression of each OAR except Octβ1R, and olfactory projection neurons (OPN) were found to express all OARs except Octβ2R at increased levels. In the salivary glands of the prepupa, only Octβ1R (log2FC = 4.90, FDR < 0.0001) and Oct-TyrR (log2FC = 1.08, FDR = 0.05) were observed. Throughout the body, Octα2R was the most abundant OAR.

In the adult, whole head samples showed expression of each OAR except Oct-TyrR (log2FC = 0.00, FDR = 0.62). But in the brain, Oct-TyrR (log2FC = 4.19, FDR < 0.0001) transcript levels were increased as well, indicating that head tissue surrounding the brain does not express Oct-TyrR as much as the other OARs. In some neuron populations, every OAR was observed at high abundance, namely: Δ7 neurons (D7) of the protocerebral bridge, dopaminergic neurons (DAN), glia, and olfactory projection neurons (OPN). Other neuron populations were found to express some OARs more than others, such as KC (Oamb, Octα2R, Octβ1R and Octβ3R); octopaminergic neurons (OAN, Octα2R, Octβ1R and Octβ2R); the anterior paired lateral neuron (APL; Oamb and Octβ2R) and the so-called E-PG neurons in the central complex (Octβ1R and Octβ2R). Similarly, non-neuronal tissue was found to express max. two OARs at elevated levels: labellum and spermatheca express the same OARs (Oamb and Octβ2R), and the accessory gland and ovary express only Oamb. In addition, due to the assignment of cell identity in RNA-sequencing data, we were able to find so far unknown expression locations in neuronal cells: APL (Oamb, Octβ2R), DAN (Octα2R, Octβ1R, Octβ2R, Octβ3R, and Oct-TyrR), Δ7 neurons (Oamb, Octα2R, Octβ1R, Octβ2R, Octβ3R and Oct-TyrR), E-PG neurons (Octβ1R and Octβ2R), glia (Oamb, Octα2R, Octβ1R, Octβ3R, and Oct-TyrR), KC (Octα2R), OPN (Oamb, Octα2R, Octβ1R, Octβ2R, Octβ3R, and Oct-TyrR), and in non-neuronal cells: labellum (Oamb, Octβ2R).

Interestingly, Octβ1R was found at exceptionally high transcript levels in adult OPN (median gene TPM: 535, log2FC = 7.4, FDR = 0.013) and in KC (median gene TPM: 225, log2FC = 6.1, FDR < 0.0001) compared to all other expression levels in the different tissues and cells, as well as for other OARs. Only Oamb in DAN (median gene TPM: 105, log2FC = 4.7, FDR < 0.0001) showed a similar high transcript level.

Overall, each OAR was expressed in fewer tissues and cells in third-instar larva compared to the pupa and adult. And although the median transcript abundance in general was low in the whole body in each stage (with Octα2R being the only exception in the pupa), expression was elevated in specific tissues and cell populations, varying from third-instar larva to adult.

The Trojan Exon enables analysis of a functionally disrupted target gene and localizes its gene product. Using this genetic tool, we were able to break down the analysis from the entire OA circuit to the receiver side, focusing on each receptor individually. We identified novel phenotypes of OARs in larval hatching and the survival of larvae and pupae. Figure 1 summarizes the published and novel functions of the OARs across the developmental stages. Moreover, we expanded the expression map of OARs to (I) cells and tissues that were not considered to be a potential place of action of OA, (II) to the first-instar larva, and (III) to Octα2R, which was not reported before. Bulk RNA-seq analysis allowed the identification of additional putative sites of expression across developmental stages, such as Oct-TyrR in larval lymph gland and adult glia. Although transcript abundance does not permit direct conclusions regarding protein concentration, quantitative mRNA analysis provides an alternative approach for estimating gene expression levels in cases where antibodies are unavailable. Supplementary Table 11 provides an overview of the expression patterns in the third-instar larva from this study combined with previous findings. The molecular evaluation of the Trojan Exon OARs provides a solid basis for future investigations of the OAR system. Likewise, the results show that Trojan Exon lines generally require molecular validation to ensure that the gene of interest is disrupted as intended.

Molecular considerations regarding the OAR Trojan Exon lines

Prominent phenotypes of Tβh lines are female sterility (Monastirioti et al., 1996; Monastirioti, 2003; Rezaval et al., 2014) and impaired larval locomotion (Saraswati et al., 2004; Fox et al., 2006; Selcho et al., 2012), among others. We analyzed the performance of Trojan Tβh for these two selected traits and couldn’t find any impairments (Figures 2A, 3C). An anatomical analysis of Tβh in the Trojan Tβh line following crosses with several UAS reporters revealed no signal (data not shown). Molecular evaluation revealed that the Trojan Tβh line had lost the Trojan Exon (Figure 7AI). Reordering the line showed the same results. Therefore, the Trojan Tβh line is not appropriate for studying the function of OA in Drosophila.

Molecular evaluation of the Trojan Octβ2R line revealed additional sequences 5′ and 3′ of the Trojan Exon (Figures 7EI,II). However, we conclude that despite these genetic inconsistencies, the Trojan Octβ2R is a suitable tool to investigate this receptor. The additional sequences are not present at the mRNA level (Supplementary Figure 14), and the Trojan Exon leads to decreased Octβ2R transcript levels (Supplementary Figure 13C). Residual Octβ2R expression of approximately 6% may reflect low-level alternative splicing that bypasses the Trojan Exon. The utility of the Trojan Octβ2R line is supported both at the behavioral and anatomical levels. Behaviorally, it recapitulated previously reported phenotypes, including female fecundity and impaired larval locomotion (Figures 2A, 3C; Lim et al., 2014; Li et al., 2015; Koon and Budnik, 2012). Anatomically, Trojan Octβ2R expression showed larval muscle labeling as reported before (Figure 4; El-Kholy et al., 2015).

For the Trojan Oct-TyrR, no mRNA was detected (Supplementary Figure 13C). The absence of Oct-TyrR mRNA may therefore account for the observed developmental and behavioral phenotypes (Figures 2B–D, 3E), much like the previously reported chemotaxis defect (Ma et al., 2016). But due to the insertion of the Trojan Exon in the Oct-TyrR 5′ UTR (Figure 7G), the expression pattern shown in Figures 4, 5F might not resemble the endogenous one. Thus, the Trojan Oct-TyrR should be used with caution, despite showing CNS and MNs expression patterns comparable to those previously reported (Stowers, 2011).

Oamb

Given the established role of Oamb in the female reproductive system (Lee et al., 2003; Deady and Sun, 2015), the Trojan Oamb line’s egg-laying phenotype (Figure 2A), combined with the molecular validation (Figure 7), confirms its suitability as a tool for further studies, thereby substantiating our expression analysis data (Figures 4, 5A). We observed labeling throughout the larval CNS with Oamb being more abundant in the VNC than the hemispheres (Figure 8). The primary olfactory processing center, the AL, and the learning and memory center, the MB, were labeled, suggesting potential function in these specified structures (Figures 5A, 6). The PNS showed signal from the head to the tail: beginning in the PCE, gustatory (TO, VO), olfactory (DO), mechanosensory (DO, LO, TO, VO), thermosensory (DO), chemosensory (oxygen; TO), and osmosensory (TO) organs were identified (for further information on the function of the different organs, see Rist and Thum, 2017; Richter et al., 2023). Thoracic and abdominal segments were found to have signal in neurons and organs for proprioception (DAs; Ferreira Castro et al., 2020; He et al., 2019; ChOs; Caldwell et al., 2003), innocuous touch (DAs; Hu et al., 2017; Tsubouchi et al., 2012), as well as noxious stimuli such as cold, heat, touch, light, or acidity (DAs; Tsubouchi et al., 2012). Recently, a function for Oamb in DAs was found in nociceptive plasticity (Boivin et al., 2026). ESs in the abdomen indicate further involvement in the perception of mechanical and/or chemical stimuli (Jan and Jan, 1993). Labeled TDs that wrap tracheal branches and TSCs at A8/9 suggest a role in the respiratory system since these structures are associated with the perception of CO 2 and/or oxygen (Bodmer and Jan, 1987; Hückesfeld et al., 2021; Lu et al., 2024; Richter et al., 2023). In the neuromuscular system, BD neurons associated with somatic muscles (Bodmer and Jan, 1987) and motor neurons were labeled (also reported by Jetti et al., 2023; Bakshinska et al., 2025). Among the Trojan OAR lines, the Trojan Oamb is the most broadly distributed OAR in the larval periphery, suggesting an as- yet unknown function in the sensory system (for further discussion, see the last paragraph).

Octα2R

The Trojan Octα2R line allowed us to generate the first anatomical expression map in the larva (Figures 4, 5B, 6). Interestingly, a small number of KC was labeled, indicating a potential function in learning and memory (approx. 16 ± 1; Figures 5B, 6). In the PCE, Octα2R was observed in the DOG, TO, and VO, suggesting involvement in gustatory, mechanosensory, chemosensory, and/or osmosensory processes. Across the body, we found signals in DAs, ESs, and MNs, as well as in ChOs in the abdominal segments; thus, Octα2R may be associated with proprioception and/or nociception, mechano-/chemosensory, and the neuromuscular system. At the terminus, the TSCs and SSO were labeled. Richter et al. (2023) discussed the function of the TSCs and SSO in the positioning of the larva in a food medium to hide from predators and daylight, but in the same way, avoid suffocation from digging too deep (Kim et al., 2017). Although the exact role of OA in this survival mechanism remains to be elucidated, the labeling of the Trojan Octα2R line may serve as an initial indicator of its potential involvement.

Octβ1R

The Trojan Octβ1R line exhibited increased larval mortality (Figure 2C). However, the larval feeding rate remained unaltered (Figure 3D), making it unlikely that insufficient nutrient intake is the primary driver of mortality. Instead, the hyperactivity phenotype (Figure 3C) points toward a fatal energetic mismatch. Since increased locomotion in Drosophila significantly raises metabolic rate and accelerates the depletion of energy reserves (Sujkowski et al., 2017), the larvae likely succumb to metabolic exhaustion. Under these conditions, a normal feeding rate appears insufficient to compensate for the high metabolic costs of chronic hyperactivity, ultimately leading to the observed mortality. Beyond these metabolic constraints, the Trojan line labeled various cells in the third-instar larva, suggesting potential functional involvements that may contribute to survival. This labeling was observed across diverse sensory structures, including gustatory (TO, VO), olfactory (DO), mechanosensory (DO, TO, VO), thermosensory (DO), chemosensory (TO), and osmosensory (TO) organs (Figure 4). Larval survival could be affected by a potential function in proprioception (ChOs and DAs), nociception (DAs), mechano-/chemosensation (ESs), or the neuromuscular system (BDs and MNs). Defects within the TSCs and SSO might have resulted in the detrimental positioning of larvae within the food medium, subsequently leading to a misjudgment of oxygen availability. Nevertheless, the labeling in the CNS, particularly in specialized structures such as the AL and MB, remains a crucial factor when analyzing the role of Octβ1R in larval mortality (Figures 5C–CXIV). Specifically, the high abundance of this receptor throughout the CNS and in eye imaginal disc nuclei (compared to other OARs) suggests a functional role in the nervous system and during development (Figure 8). Additionally, earlier studies reported this receptor in the Malpighian tubules, the tracheal system (El-Kholy et al., 2015), the salivary glands (Ohhara et al., 2012), and the eye-antennal discs (Ohhara et al., 2012; Koon and Budnik, 2012). Finally, a potential involvement in the viability of first- and second-instar larvae, or during molting, cannot be excluded.

Bulk RNA-seq analysis revealed that most OARs are expressed at low levels across different tissues and developmental stages. In contrast, Octβ1R exhibited exceptionally high abundance, restricted to specific cell populations: KC and OPN in adults. While Octβ1R in the MB drives aversive learning, its localization in OPN plays a major role in appetitive learning (Sabandal et al., 2020). The exceptional abundance of Octβ1R stands out against the otherwise low expression background in different cells and developmental stages, as well as the other OARs, pointing to its specialized importance in KC and OPN.

Octβ2R

In the Trojan Octβ2R line, we observed labeling of all 29 somatic muscles of abdominal segment 1 and all 30 somatic muscles of abdominal segments 2–7 in the body wall (Figure 4). Since crawling requires the activation of all somatic body wall muscles (Zarin et al., 2019), this widespread labeling pattern accounts for the locomotion deficit of the Trojan Octβ2R line (Figure 3C). Moreover, head muscles (PCE and T1–3) exhibited labeling and are engaged in larval movements like turning or burrowing. In terms of feeding behavior, mouthpart-associated muscles in the prothorax were labeled (Figure 4; Hooper, 1986), and consistently, the Trojan Exon line displayed a reduced feeding rate (Figure 3D). Labeling of the somatic muscles flanking the terminal spiracles and anal division implies an involvement in respiratory and excretory functions (Figure 4). Although Octβ2R has been described in MNs (Koon et al., 2011; Koon and Budnik, 2012; Jetti et al., 2023), no MN labeling was observed in the Trojan Exon line. This discrepancy might arise because delicate neuronal structures were masked by the strong muscular signal. This rationale could also explain the absence of signal in the fat body, intestine, imaginal discs, Malpighian tubules, midgut, salivary glands, and tracheal system, despite expression being detected in earlier studies (Ohhara et al., 2012; El-Kholy et al., 2015). Aside from the muscles, the TO exhibited the only peripheral signal, presumably due to the absence of overlapping muscle fluorescence.

Octβ3R

In third-instar larvae, Octβ3R labeling was observed in both internal and external sensory organs, responding primarily to chemo- and mechanosensory signals. This indicates that Octβ3R plays a functional role in various sensory modalities, including gustation (TO), olfaction (DO), mechanosensation (DO, ESs, TO), thermosensation (DO), chemosensation (ESs, TO), osmosensation (TO), and proprioception (ChOs, DAs). Additionally, it may be involved in peripheral neurons related to nociception (DAs) and muscle activity (BDs, MNs) (Figure 4). Trojan Octβ3R showed labeling in TSCs and SSO, which are believed to respond to changes in oxygen levels or to give mechanosensory feedback when the larvae submerge into the substrate. When this complex information is not properly processed, the larva’s ability to adjust to environmental conditions may be compromised, potentially impacting survival (Figure 2C). Additionally, the CNS expression pattern (Figures 5E, 6) and elevated transcript levels (Figure 8) warrant attention, as does the high abundance in the nuclei of eye imaginal discs (Figure 8). However, as summarized in Supplementary Table 11, previous studies have detected Octβ3R in additional tissues, including various glands, the Malpighian tubules, and the midgut.

During our assessment of larval mortality, we focused on counting the number of emerging pupae rather than quantifying deceased larvae. Previously, Octβ3R was reported to play a crucial role in inducing metamorphosis by regulating ecdysone biosynthesis in the prothoracic gland (Ohhara et al., 2015). However, the pupariation defect observed in our Trojan Octβ3R line was less pronounced than in that earlier report. This discrepancy may be attributed to technical differences or variations in the genetic tools employed; specifically, Ohhara et al. (2015) used phm-22-Gal4 to drive two independent RNAi lines targeting Octβ3R in the prothoracic gland, an approach whose strong pupariation phenotype might be overestimated. Alternatively, because the Trojan Exon is inserted further downstream in the gene, it could potentially allow for the production of a truncated but partially functional receptor. However, this same line provides a distinct learning phenotype that we recently described in detail (Franke et al., 2026), demonstrating its utility for larval olfactory learning experiments. Consequently, further investigations remain necessary to determine the precise temporal and spatial window of Octβ3R function affecting larval survival and pupariation.

Oct-TyrR

The Trojan Oct-TyrR line exhibited reduced survival rates during larval and pupal development (Figures 2C,D). The labeling in ChOs, DAs, DMD1, ESs, and TSCs suggests a potential involvement in proprioception (ChOs, DAs, and DMD1; Heckscher et al., 2015), as well as mechano- and chemosensation and oxygen perception (ESs and TSCs, Figure 4). These sensory roles align with the observed reduction in sugar preference (Figure 3E), a previously reported chemotaxis phenotype, and the altered startle response in honoka larvae, a different Oct-TyrR line (Ma et al., 2016). Moreover, labeling was observed throughout the body, including the CNS, dorsal aorta, lymph gland, Malpighian tubule, posterior spiracles, salivary gland, and the tracheal system (Stowers, 2011; El-Kholy et al., 2015; Figures 5F, 8). Whether a functional role in the sensory system, the different tissues, or both accounts for the increased larval mortality remains unclear. However, a critical function during early larval stages or during molting cannot be ruled out. Future experiments utilizing stage-specific lethality assays or tissue-specific knockdown will be necessary to delineate the distinct developmental functions of OARs.

Further, we identified a potential role for Oct-TyrR in pupal survival, based on the increased pupal mortality observed in Trojan Oct-TyrR (Figure 2D). Interestingly, bulk RNA-sequencing analysis revealed that Oct-TyrR transcript levels were elevated not only in the pupal CNS and eye but also in the salivary glands (Figure 8). Although other OARs exhibited similar expression profiles, it remains to be elucidated why certain OARs impact pupal survival while others do not. Specifically, the high abundance of Oct-TyrR transcripts in prepupal salivary glands, which is a major secretory organ that undergoes autophagy during the early phases of metamorphosis, suggests a critical developmental function for OA signaling.

A role for Oct-TyrR in locomotion was first described in honoka mutants, which exhibited reduced excitatory junctional potential (EJP) amplitudes in body-wall muscles (Kutsukake et al., 2000; Nagaya et al., 2002). Later, expression was detected in the CNS, MNs, and at the neuromuscular junction (NMJ; Stowers, 2011; Figure 4). Our analysis of locomotion behavior in Trojan Oct-TyrR revealed no impairment compared to the control (Figure 3C), consistent with previous data from honoka larvae (Schützler et al., 2019). Although we reported a hyperactivity phenotype in honoka (Selcho et al., 2012), a direct comparison between these two studies is difficult; this discrepancy arises not only from the use of different control lines but also from distinct tracking methodologies, specifically automated (Schützler et al., 2019) versus manual tracking (Selcho et al., 2012). Consequently, further research is needed to clarify the role of Oct-TyrR in larval locomotion.

The OAR system

OA plays a role in various mechanisms spanning from embryogenesis through larval development to adulthood. A substantial body of literature has already identified multiple OA receptors involved in the same biological processes (Figure 1). Here, we observed a phenotypic overlap among three OARs (Octβ1R, Octβ3R, and Oct-TyrR) involved in embryonic development and/or larval hatching. Embryogenesis comprises 17 stages, with organ differentiation beginning at stage 13 and continuing through stage 16. During this phase, elevated Octβ1R transcript levels have been reported in the CNS and PNS (Berkeley Drosophila Genome Project; Tomancak et al., 2002; Tomancak et al., 2007). Similarly, during stages 15–17, Ohhara et al. (2012) observed increased mRNA levels of both Octβ1R and Octβ3R in the CNS as well as the whole embryo. Unfortunately, embryonic transcript data for Oct-TyrR are currently unavailable. Although the specific roles of Octβ1R, Octβ3R, and Oct-TyrR during embryogenesis remain to be fully investigated, their elevated embryonic expression levels coupled with the reduced hatching rates observed here warrant further functional studies. Alternatively, maternal effects cannot be ruled out. Embryos developing from eggs produced by a mutant mother, which lack maternal contribution, could account for the observed offspring lethality. Future experiments utilizing heterozygous mothers will be necessary to distinguish between zygotic and maternal effects, or a combination of both, on embryonic development. In conclusion, our data demonstrate a novel role for Octβ1R, Octβ3R, and Oct-TyrR in ensuring overall developmental viability.

Furthermore, consistent with previous data, anatomical mapping revealed an overlap of OARs across the third-instar larva, pupa, and adult stages (Figures 4–6, 8; Supplementary Table 11). However, whether distinct OARs are co-expressed within identical cells remains to be confirmed. Implementing a combination of different protein tags, such as Venus (Kondo et al., 2020), could provide deeper insight into co-expression and subcellular receptor localization within individual cells.

One experimental approach to understanding the functional implications of overlapping OAR labeling was to utilize chemosensory assays. The sensory system in the larval PCE exhibited distinct labeling for OARs in the DO, TO, and VO; however, behavioral tests did not suggest a direct involvement of OA. The similarities in OAR labeling patterns within the same functional unit for olfactory or gustatory perception may indicate that other OARs compensate for the absence of receptors in single OAR lines. This redundancy is particularly plausible given that downstream signaling is conserved among GPCRs, and certain OARs exhibit similar G-protein coupling (Table 1). Since OA is involved in various mechanisms crucial for survival, a redundant system wherein one receptor can compensate for the loss of another would be less vulnerable to interference. For instance, ectopic expression of Octβ1R in the oviduct epithelium has been shown to substitute for Octβ2R and restore ovulation in Octβ2R f05679 females (Lim et al., 2014). However, whether such a compensatory effect occurs in the chemosensory system remains to be clarified, as does whether it operates at the level of the functional unit or within individual cells. Alternatively, the diverse downstream pathways of OARs suggest a potential modulatory mechanism. For instance, OA signaling via Oamb, Octβ1R, and Octβ2R contributes to synaptic plasticity at the NMJ in a modulatory manner (Koon et al., 2011; Koon and Budnik, 2012; Bakshinska et al., 2025; Blaum et al., 2025). Ultimately, it remains to be elucidated whether the OA system functions in a redundant or modulatory mode, or perhaps both, depending on the cell type. Future investigations targeting these roles could utilize double and triple mutants in developmental, behavioral, and physiological assays.