A comprehensive anatomical map of the peripheral octopaminergic/tyraminergic system of Drosophila melanogaster

The modulation of an animal’s behavior through external sensory stimuli, previous experience and its internal state is crucial to survive in a constantly changing environment. In most insects, octopamine (OA) and its precursor tyramine (TA) modulate a variety of physiological processes and behaviors by shifting the organism from a relaxed or dormant condition to a responsive, excited and alerted state. Even though OA/TA neurons of the central brain are described on single cell level in Drosophila melanogaster, the periphery was largely omitted from anatomical studies. Given that OA/TA is involved in behaviors like feeding, flying and locomotion, which highly depend on a variety of peripheral organs, it is necessary to study the peripheral connections of these neurons to get a complete picture of the OA/TA circuitry. We here describe the anatomy of this aminergic system in relation to peripheral tissues of the entire fly. OA/TA neurons arborize onto skeletal muscles all over the body and innervate reproductive organs, the heart, the corpora allata, and sensory organs in the antennae, legs, wings and halteres underlining their relevance in modulating complex behaviors.

body and AKH cells, OA is shown to modulate the heart, trachea and air sacs, gut, hemocytes, salivary glands, Malpighian tubules and ovaries in insects, mainly to induce a general stress or arousal state. However, in total OA seems to modulate a vast number of behaviors, which are not necessarily coupled to stress responses. The OA/ TA system is shown to also act on inter alia (i.a.) learning and memory, sleep, feeding, flight, locomotion, and aggression 5,8,10,12,13,15-40 . As mentioned above, OA and TA act as neurotransmitters and neuromodulators, allowing them to act in a paracrine, endocrine or autocrine fashion. In the fruit fly Drosophila, huge efforts were made to describe OA/ TA neurons (OANs/TANs) in the brain and ventral nervous system (VNS) down to the single cell level 8,17,[41][42][43][44][45] . Nevertheless, although our knowledge about physiological processes and behaviors modulated by the OA/TA system in the brain is rich, less is known about how OA and TA reach all its target organs and tissues in the periphery (exceptions: reproductive organs 41,[45][46][47][48] and muscles 8,[49][50][51][52]. Here we use the genetically tractable fruit fly Drosophila melanogaster to describe the arborizations of Tdc2-Gal4-positive, and therefore OANs and TANs in the periphery, as the Drosophila Tdc2 gene is expressed neurally 45 . We found that OANs/TANs are widespread distributed throughout the fly's body with innervations in the skeletal muscles, reproductive organs, corpora allata, antennae, legs, wings, halteres and the heart. This diverse innervation pattern reflects the modulatory role of OA/TA in many different behaviors and physiological processes. Our results provide, for the very first time, a complete and comprehensive map of the OA/TA circuitry in the entire insect body. This map allows assumptions about the type of OA/TA signaling (paracrine or endocrine) to a specific organ and, at the same time, it provides a deeper understanding to what extend the OA/ TA-dependent activity of peripheral organs is altered, for example by genetically manipulating Tdc2-Gal4-positive neurons in the brain and VNS.

Results
The OANs/TANs of the brain and ventral nervous system (VNS) are described in detail even on single cell level in Drosophila 10,[42][43][44]49 . In contrast little is known about their peripheral arborizations. We used the well-characterized Tdc2-Gal4 line to deepen our knowledge about the OA/TA system in the entire body of Drosophila [43][44][45][46]53 . Nearly all Tdc2-Gal4-positive cells in the brain are stained by a Tdc2 antibody 53 . In the VNS all of the Tdc2-Gal4-positive cells were labeled by a TβH antibody and therefore have to be Tdc2-positive 44 . We here expressed myristoylated GFP, enhanced by GFP antibody staining, to label the membranes of Tdc2-Gal4-positive neurons from the soma to its fine endings in the periphery. The peripheral organs, tissues and cells are visualized by fluorescent markers for cell bodies (DAPI binds to DNA), muscles (Phalloidin binds F-actin) and antibodies against the synaptic protein Bruchpilot and the cell adhesion molecule Fasciclin 2 (Fig. 1). Tdc2-Gal4-positive arborizations in the head. Tdc2-Gal4-positive neurons (Tdc2Ns) project through all peripheral nerves of the brain: antennal nerve (AN), ocellar nerve (OCN), pharyngeal nerve (PhN) and accessory PhN (APhN), maxillary-labial nerve (MxLbN), corpora cardiaca nerve (NCC), and the cervical connective (CV) (Fig. 2). Tdc2Ns in the antennal nerve are connected to the antennal lobe and the antennal mechanosensory and motor center (AMMC), respectively, and give rise to staining in the pedicel-the Johnston's organ (JO)-, funiculus and arista of the antenna ( Fig. 2A-D). While no cell bodies are visible in the third to fifth segment of the antennae, the JO contains stained cell bodies indicating that the Tdc2-Gal4 line includes mechanosensory neurons (Fig. 2C).
The Tdc2-Gal4-positive efferent nerves of the subesophageal zone (SEZ) arborize in the rostrum of the proboscis mainly onto muscles ( Fig. 2E-M). Cells leaving the brain via the PhN innervate muscles 9, 10, 11 and 12 (nomenclature after 54 ; Fig. 2F,G). Cells projecting towards the APhN build bouton like structures beside muscle 11 ventral to the pharynx (ph; Fig. 2H). Cells of the MxLbN arborize along muscles 1 and 2 ( Fig. 2J-L). It seems as if also muscles 7 and 8 of the haustellum are innervated (Fig. 2M). We only observed this staining in two different specimens. Due to our cutting technique we probably lost these parts of the haustellum frequently. In addition to the innervation of the proboscis muscles we observed arborizations in the ventrolateral head arising from the MxLbN (arrowhead Fig. 2K). The ocellar nerve, which connects the ocellar ganglion with the brain, contains fibers arising from the brain (Fig. 2N). The central brain and optic lobes were shown to contain a dense network of OANs/TANs 42,43 . In addition, we identified arborizations in the distal part of the lamina by Tdc2Ns (Fig. 2O). Muscle 16, which is located dorsal to the esophagus, is innervated via ascending Tdc2Ns from the thorax (Fig. 2I).

Discussion
Here we show a comprehensive description of the innervation of OANs/TANs in the periphery of Drosophila. For this, we used the Tdc2-Gal4 line, allowing Gal4 expression under the control of a regulatory sequence of the tyrosine decarboxylase enzyme 45 . As this enzyme is essential for the synthesis of TA from tyrosine, the Tdc2-Gal4-line labels both TANs and OANs. Within the Drosophila brain, Tdc2-Gal4 labels in total about 137 cells, while additional 39 cells are located in the VNS 43,44 . The small number of Tdc2Ns lead to arborizations in large parts of the central brain, optic lobes and the thoracic and abdominal ganglion [42][43][44][45] . Based on the profound innervation of Tdc2Ns in the brain and VNS, the variety of behaviors modulated by the OA/TA system including learning and memory, feeding, vision, and sleep, are not surprising. Beyond the brain and VNS, OANs and TANs massively innervate regions within the periphery of the fly. Here, we described arborizations on most skeletal muscles, the antennae, wings, halteres and reproductive system and parts of the circulatory system and stomodaeal ganglion ( Fig. 6; Table 1).
Our findings are in line with previous reports focusing on the expression of different OA and TA receptors in the fly 56,57 . Accordingly, the OA receptor OAMB is expressed in reproductive organs (in both male and female flies) and muscles, which are directly innervated by Tdc2Ns. Additionally, the midgut and trachea contain OA and TA receptors 56,57 , but do not seem to be innervated by Tdc2Ns, even though axons run in close vicinity to these organs. Likewise, the OA receptor Octß2R is expressed in the fat body, salivary glands and Malpighian tubules, tissues that seem not to be innervated by Tdc2Ns, while the expression of Octß1R and Octß3R is more specific 56,57 . The three tyramine receptors TyrR, TyrRII and TyrRIII show a broad expression in the periphery, also in tissues not innervated by Tdc2Ns 56 . The lack of direct innervation of these peripheral tissues might argue for volume transmission or hemolymph released OA/TA from Tdc2Ns. Alternatively, Drosophila TDC1, the product of the non-neurally expressed Tdc gene, is expressed in the gut musculature, rectal papillae, Malpighian tubules and two small clusters in the thoracic nervous system and might be the source for peripheral TA 45 . Interestingly, TyrR seems to be the only receptor expressed in the heart, suggesting that only TA modulates heart function 56 . Contrary, OA has a modulatory effect on the heart of other insect species including honeybees, olive fruit flies and cockroaches 58 . This is also in line with a previous report providing evidence that OA modulates the heart rate of the Drosophila fly and pupa, but not the larva 59 .
OA-dependent modulation of organs and tissues is mainly elicited through muscle action, especially in terms of its impact on the "fight or flight" response. In line with this, we observed Tdc2-Gal4-positive arborizations on nearly all skeletal muscles and many visceral muscles (Table 1). In both Drosophila and desert locusts OA and TA is expressed in type II terminals of skeletal muscles 50 . OA has an excitatory effect on Drosophila flight muscles, while TA was shown to inhibit excitatory junction potentials, and thereby reduce muscle contractions and locomotion at least in the larva 6,8,60,61 . In addition, flies lacking OA show severe deficits in flight initiation and maintenance 10,50 . Interestingly, in an antagonistic effect to serotonin, OA reduces crop muscle activity presumably via Octß1R, suggesting that OA has different effects on muscle activity dependent on the type of muscle 56,62 . However, our data do not provide any convincing evidence of a direct innervation of Tdc2Ns of the crop, even though many fibers run in close vicinity, suggesting that OA might target the crop by volume transmission. Furthermore, OA modulates ovulation and fertilization in insects [63][64][65][66][67] . Flies lacking OA display a severe egg-laying phenotype. Remarkably, within the female reproductive organ two different OA receptors, OAMB and Octß2R, are necessary. Again, OA has a strong impact on muscle activity within the reproductive system. Octß2R is expressed in the visceral oviduct muscle and elicits muscle relaxation through an increase of intracellular cAMP levels 63 . Such an OA-dependent modulation appears to be conserved as OA is found in dorsal unpaired median neurons of locusts innervating oviduct muscles through the oviducal nerve 68 . However, our data suggest that OA-positive fibers not only innervate oviduct muscles, but also enter the organs themselves. The OAMB receptor is expressed in epithelial cells inducing fluid secretion through increasing intracellular Ca 2+ levels 63 . Thus, OA affects different processes within the female reproductive organ due to the expression of different receptors and their coupled signaling pathways, which may be a general mechanism of the OA/TA system to fulfill an extensive modulatory function 69 .
OA does not exclusively modulate muscle activity, but also sensory neurons of external tissues like the antennae, halteres and wings. OA has also been shown to increase the spontaneous activity of olfactory receptor neurons (ORN) 70,71 . The modulation of ORNs allows OA to modulate the innate response to attractive stimuli like fruit odors or pheromones 72,73 . Further, this modulation helps nestmate recognition in ants 74 . In addition to Tdc2-Gal4-positive arborizations in the funiculus, we found Tdc2-Gal4-positive sensory neurons in the JO, a chordotonal organ sensitive to mechanosensory stimuli and thus important for hearing in insects. In mosquitos, OA modulates auditory frequency tuning and thereby affects mating behavior 75 . In locusts, OA similarly modulates the response of chordotonal neurons in the legs to encode proprioceptive information 76 . Our data suggest that chordotonal neurons in the legs, wings, halteres and thorax are included in the Tdc2-Gal4 line suggesting a conserved modulatory role of OA/TA for insect proprioception.
Taken together, our study suggest that the OA/TA system massively modulates various organs and tissues in the periphery of Drosophila. Through distinct receptors and coupled signaling pathways OANs/TANs mainly induce "fight or flight" responses by modulating muscle activity, proprioception, and heart rate. As a result, the innervation pattern in the periphery supports the idea that the OA/TA system is crucial for insects to switch from a dormant to an excited state, by a positive modulation of muscle activity, heart rate and energy supply, and a simultaneous negative modulation of physiological processes like sleep.

Methods
Fly strains and fly rearing. All flies were cultured according to standard methods. In short, vials were kept under constant conditions with 25 °C and 60% humidity in a 12:12 light:dark cycle. Flies carrying the Tdc2-Gal4 45 (Bloomington Stock Center) and 10xUAS-IVS-myrGFP 77 , (Bloomington Stock Center) constructs were used for immunohistochemistry. To control for an unspecific expression of the UAS construct, we stained 10xUAS-IVS-myrGFP alone. No GFP staining was detected. Additionally, we performed anti-Tdc2 (pab0822-P; Covalab; 1:200) staining experiments to confirm our Tdc2-Gal4 expression pattern. Indeed, we observed Tdc2 staining in all Tdc2-Gal4-positive peripheral nerves (data not shown).

Immunocytochemistry.
To visualize the arborizations of Tdc2-Gal4-positive neurons in the periphery whole body sections, as well as sections of the head, thorax and abdomen, were performed, respectively (see 78 ). In short, the cuticle of 4 to 7 days old flies was opened in phosphate buffered saline (PBS, 0.1 M) to ensure that the fixative can penetrate into the tissue. Whole flies were fixed with 4% paraformaldehyde in PBS for two hours at room temperature and afterwards washed three times with PBS. Subsequently flies were embedded in hot 7% Agarose low EEO (A2114; AppliChem). After hardening, the flies were cut with a vibratome (Leica VT1000S) into 80-100 µm sections. Staining of the sections was continued after washing in PBS containing 0.3% Triton-X100 (PBT) and blocking in 5% normal goat serum in PBT. Rabbit anti-GFP (A6455, Molecular Probes; 1:1000) in combination with mouse anti-Synapsin (3C11; 79 ; 1:50) or mouse anti-Bruchpilot (nc82; 80 ; 1:50) and mouse anti-Fasciclin 2 (1D4; DSHB; 1:100) were used as primary antibodies. After one night at 4 °C the specimens were washed six times in PBT and incubated in secondary antibody solution for a subsequent night at 4 °C. As secondary antibodies goat anti-rabbit Alexa488 (Molecular Probes; 1:200) and goat anti-mouse DyLight649 (Jackson ImmunoResearch; 1:200) were used. 4′,6-Diamidino-2-phenylindol Dihydrochlorid (DAPI; Sigma-Aldrich; 1:1000) and Alexa Fluor 633 Phalloidin (Molecular Probes; 1:400) were used to visualize DNA and actin, respectively.
Confocal microscopy and data processing. Confocal images were taken with a Leica TCS SP8 microscope (Leica Microsystems, Germany) with a 20x high aperture objective. Labelled specimens were scanned with a step size of 1.0 μm to 1.5 μm. Image processing and stitching was performed using Fiji 81 and Adobe Photoshop CS6 (Adobe Systems, USA). Adobe Illustrator CS6 was used to draw the scheme of the fly and its peripheral organs and muscles (after 54,82 ). Table 1 summarizes the arborization pattern of the Tdc2Ns. The innervation of organs, tissues and muscles is confirmed by at least two male and two female samples (except for reproductive organs and MOL).

Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author.