Drosophila melanogaster were reared in standard 68 ml plastic vials with foam stoppers on a yeast-cornmeal-syrup-agar diet at 25°C and 55% humidity with a 12-h light/dark regimen. Fly food contained the following ingredients (per 1000 ml): 116 g glucose (Carl Roth, Germany, HN06.04), 55 g cornmeal (Rapunzel, Germany, Demeter standard 420505), 11 g agar (Roth 5210.5), 29 g active dry yeast (Huber Mühle, Germany), 0.6 g ascorbic acid (Roth 3525.3) and 12.2 ml of 10% tegosept (Apex BioResearch Products 20–258) in 100% ethanol. All experiments were carried out with 2–3 days old adult male flies. To drive expression of UAS-transgenes in glial cells we used repo-GAL4 (w¹¹¹⁸; P{GAL4}repo/TM3, Sb¹; Bloomington # 7415). To knock down Naz we used a UAS-naz-RNAi flystrain (VDRC RNAi fly stock center # 107974). To enhance RNAi knock down efficacy we included an extra copy of dicer 2 (UAS-Dcr2, Bloomington Stock 24650, Dietzl et al., 2007). To drive expression of UAS transgenes in muscle cells we used Mef2a-GAL4 (Myocyte enhancer factor 2; y w[*]; P{w[+mC] = GAL4-Mef2.R}3; Bloomington # 27390).

Two to three days old male flies were immobilized by cooling for 30 s in an empty 68 ml plastic vial on ice and then immediately transferred onto a cold plate at 2°C. Then animals were glued (clear glass adhesive (Duro; Pacer Technology, Rancho Cucamonga, CA, USA)) with head and thorax to a triangle-shaped copper hook (0.02 mm diameter). Adhesion was achieved by exposure to UV light for 10 s. The animals were then immediately returned to room temperature and kept individually in small chambers containing a few grains of sucrose on filter paper until testing (2–3 h). Flies, glued to copper hooks were attached to the experimental setup via a clamp to accomplish stationary flight. For observation, the fly was illuminated from behind and above 150 W (15 V; Schott, Elmsford, NY, USA) and fixed in front of a paper panel with horizontal white and black stripe patterns. Tarsal contact with a bead of polystyrene prevented flight initiation before the experiment started. For flight initiation, the polystyrene bead was removed, and the fly was gently aspirated by mouth to produce an air current of about 0.8 m per second. The time until the fly first stopped flying was measured. After each flight bout the fly was again gently aspirated as a stimulation to fly. The time was recorded for each flight bout. The experiment was completed when no flight reaction was initiated in response to three consecutive air stimuli. The person recording the flight times was unaware of the genotype of the animal.

Brains of 2–3 day-old male adult Drosophila were dissected according to Wu and Luo (2006). For SDS-PAGE Western blot 10 brains per lane (40 ml Hoefer Western blot chamber with 1.5 mm spacer and comb with 15 × 100 μl pockets) were dissected and mechanically homogenized with a clean micro-pestle in 85 μl ice cold 2× concentrated sample buffer (7.5 ml 4× Tris-HCl/SDS, pH 6.8, 6 ml glycerol, 1.5 g SDS, 0.3 g dithiothreitol (DTT), ~1 mg bromophenol blue in 30 ml with ddH₂O). 4× Tris-HCl was prepared with 40 ml ddH₂O, 6.05 g Tris base, 0.4 g SDS, pH was adjusted to 6.8 with HCl, then filled up to 100 ml with ddH₂O and filtered through a 0.45 μm sterile filter. Homogenized samples were boiled for 3 min and then loaded into the gel chambers. We used a 5% polyacrylamide stacking gel (10 ml, 1.7 ml of 30% bis-acrylamide, 1.25 ml 4× Tris/SDS pH 8.8, 6.8 ml ddH₂O, 100 μl 10% ammoniumpersulfate (APS), 10 μl TEMED) and let the samples run through a 10% polyacrylamide running gel (30 ml, 13.3 ml 30% bis-acrylamide, 10 ml 4× Tris/SDS pH 8.8, 16 ml ddH₂O, 400 μl 10% APS, 16 μl TEMED. APS and TEMED were added just prior to pouring the gel. Samples were run through the stacking gel at 20 mA and then through the running gel at 30 mA constant current at room temperature. As protein marker 20 μl Precision Plus Protein^™ WesternC^™ Blotting Standard 10–250 kDa was used (Bio Rad #161-0376). Proteins were then transferred onto nitrocellulose membrane (Bio Rad) overnight at 4°C at 35 V constant voltage in a 5 L 42E Hoefer blotting chamber using transfer buffer (18.2 g Tris base, 86.5 g glycine, 900 ml methanol, fill up to 6 L with ddH₂O). Nitrocellulose membrane was incubated in blocking solution (10% milk powder in electrophoresis buffer with Tween-20 (TBST) for 2 h at room temperature. Before antibody incubation, nitrocellulose membrane was cut horizontally because both primary antibodies were made in rabbit. Primary antibodies were prepared in 2.5% milk powder in TBST and the nitrocellulose membrane incubated overnight: rabbit α-hsp90 (New England Biolabs or Cell Signaling Technology, #4872S) as loading control 1:1000 and rabbit α-Naz 1:250 (von Hilchen et al., 2013) were used. Secondary antibody incubation was done together for both pieces of the nitrocellulose membrane. Secondary antibody was applied for 2 h at RT: goat α-rabbit (H + L) HRP conjugate 1:5000 (Millipore #AP106P) was used. Protein bands were detected with luminol-based Immobilon Western Chemiluminescence detection kit (Millipore #WBKLS01 00). Hsp90 band was expected at 84 kDa, Naz (CG31235) was expected at 46 kDa. Pictures were taken and saved as .tiff image files.

Adult flies were dissected in standard saline along the dorsal midline and the gut was removed to expose the ventral nerve cord (VNC) as described previously (Boerner and Duch, 2010). Next preparations were fixed for 60 min in 4% PFA in 0.1 M PBS, washed 6 × 20 min 0.1 M PBS and 6 × 30 min in PBS Triton-X 0.5%. This was followed by primary antibody incubation overnight at 4°C in 0.1 m PBS with 0.2% BSA and 0.3% Triton-X on a shaker. Rabbit α-Naz and guinea pig α-Repo (von Hilchen et al., 2013) were both used at 1:1000 in 0.1 M PBS Triton-X 0.1%. Specificity for both primary antibodies has previously been reported (von Hilchen et al., 2013). Specificity for α-Naz has been further confirmed in this study by the absence of immunopositive signal in Western blots after Naz-RNAi knock down.

Following primary AB incubation preparations were rinsed in 0.1 M PBS and washed in 0.1 M PBS 6 × 30 min. For detection of Naz immunolabel donkey α-rabbit secondary antibody coupled to Alexa 568 (Invitrogen A10042) was used. For detection of Repo immunolabel goat α-guinea pig secondary antibody coupled to Cy5 (Dianova 106-605-003) was used. Both secondary antibodies were incubated at a concentration of 1:500 in 0.1 M PBS overnight at 4°C on a shaker. Next preparations were rinsed for 6 × 30 min with PBS followed by an ascending ethanol series (50, 70, 90, 100%), and then mounted in methylsalicylate. Immunolabel was visualized using a Leica SP8 confocal laser-scanning microscope (Leica, Germany). Alexa 568 was excited with a 568 solid state laser and detection was set between 580 nm and 610 nm. Cy5 was excited with a 633 Helium Neon laser and detection was set between 640 nm and 690 nm. Images were processed with Amira 5.3.3 (Mercury Systems) and Corel X7 (Corel Corporation) software.

Thoracic neuromeres, lateral nerves and neuropil borders were labeled and defined as previously described (Boerner and Duch, 2010). The thoracic neuropils are characterized by the absence of any cell bodies and a glial lining that separates nervous tissue densely packed with synapses from a VNC cortex packed with cell bodies. The demarcation line can be visualized in single optical sections through the VNC.

All flight behavioral data are presented as box blots with median, 25, and 75 percentiles. Error bars represent the 10 and the 90 percentiles. Statistical differences were tested for with Kruskal Wallis ANOVA. Following a statistical significant Kruskal-Wallis p value (significance level was set <0.05) Dunn-Bonferroni post hoc method was used for between groups post hoc comparisons. For between groups comparisons * indicates p < 0.05 and ** indicates p < 0.01, whereas ns indicates p > 0.1. All statistical testing was conducted with SPSS Statistics 22 software.

Our data provide indirect evidence on the behavioral level that Naz normally functions to reduce biologically active TA levels in the flight motor neuropil, because knock down of Naz causes similar flight behavioral changes as observed with genetically increased TA levels (Brembs et al., 2007). Similarly to a reduction in flight durations in adult flies, increased TA levels reduce Drosophila larval crawling distances (Saraswati et al., 2004; Fox et al., 2006), and this phenotype is also recapitulated by knock down of Naz (de Visser, 2016). The hypothesis that Naz functions in the TA degradation pathway is supported by the finding that motor behavioral phenotypes as induced by naz knock down can be rescued by feeding the TAR blocker YH. Therefore, our data provide first hints to the mechanisms that might remove TA from the synaptic cleft. The most parsimonious explanation is that TA is taken up into Repo positive glial cells which are located at the neuropil borders and extend extensive processes into the flight motor neuropil. As typical for monoamines, TA is then likely desaminated by a MAO, and thus converted to p-hydroxyphenyl-acetaldehyde. Naz might then convert p-hydroxyphenyl-acetaldehyde further into p-hydroxyphenyl acetic acid (de Visser, 2016). We speculate that naz knock down causes accumulation of p-hydroxyphenyl-acetaldehyde, which in turn slows desamination of TA, and thus, transport of TA into glial cells. The resulting consequence would be higher extracellular TA levels in the flight motor neuropil. This scenario is consistent with our findings that naz knock down causes similar phenotypes as feeding TA or genetic upregulation of TA signaling (Brembs et al., 2007), and that pharmacological blockade TARs rescues the behavioral phenotype. This would mean that Naz function is rate limiting for TA uptake into glial cells. It would be instructive to test whether overexpression of Naz causes opposite effects on flight performance as compared to naz-RNAi, but we have so far not succeeded to produce a UAS-naz fly strain. However, it is widely accepted that biologically active monoamine levels in the synaptic cleft can be increased by blocking intracellular degradation. In Parkinson’s disease, for instance, pharmacological interventions with MAO function are utilized to enhance dopamine signaling (Unzeta and Sanz, 2011; Pathak et al., 2016). However, we have no direct evidence for the proposed function of Naz, and the transporter for TA into glial cells has also not yet been identified.

It has long been known in multiple species that the biogenic amines OA and TA modulate insect motor behavior in an antagonistic manner (Roeder, 2005; Pflüger and Duch, 2011). In addition, numerous sites of OA and TA action have been identified, ranging from the modulation of central circuitry, neuromuscular transmission and muscle contraction properties, to sensory sensitivity, hormone release and muscle metabolism (Pflüger and Duch, 2011). However, at present the relative contributions of these different sites of action to the motor behavioral changes that are observed upon altered TA and/or OA signaling remain largely unknown. Our data indicate that central actions of TA might play a prominent role in the modulation of flight motor performance. We found prominent Naz localization in glial cells with arborizations in the central motor neuropils, but almost no Naz immunopositive cells in the periphery. Accordingly, knock down of Naz in glial cells restricts the manipulation mostly to the CNS. This has not been possible with genetic manipulations of the OA and TA synthesizing enzymes (TDC2 and Tβh), the OA and TA receptors, or with drug feeding approaches. Given that we find significant reductions of flight durations and flight bout numbers, we suggest that the tyraminergic modulation of flight motor behavior is mediated to a large extent by central TA actions. In a next step it will be important to identify the cellular targets of TA action in the CNS. Given the localization of Naz positive glial arbors in the flight motor neuropil we suggest that premotor flight interneurons, flight motoneurons, or synapses between these cells might be promising targets for TA action. Accordingly, we have previously reported spatial overlap between the central arborizations of TDC2 positive OA/TA containing neurons and flight motoneuron dendrites in the VNC flight motor neuropil (Boerner and Duch, 2010). Alternative TA might act on brain circuitry that regulates the motivation to fly, but this is currently unknown. However, in vertebrate spinal cord, motoneurons are direct targets of monoaminergic modulation. There motoneuron excitability is strongly increased in the course of fight or flight reactions by the OA analog norepinephrine (Heckman et al., 2003). It will be interesting to test whether invertebrate motoneurons are also direct cellular targets of aminergic modulation and whether OA and TA exert opposite effects on motoneuron membrane excitability. Please note that our study refers to the modulation of flight motor behavior, but it is well known from larger insects that walking is also under octopaminergic/tyraminergic control (Mentel et al., 2008; Rillich et al., 2013). Although flight and leg motoneuron dendrites cover spatially separate area of the thoracic motor neuropils in both larger insects (Ramirez and Pearson, 1988) and Drosophila (Baek and Mann, 2009; Brierley et al., 2012), Naz positive glial arbors and central arbors of OA/TA containing neurons are present in both neuropil areas. Therefore, from a sole anatomical point of view similar central actions of TA and OA are possible for the modulation of walking motor behavior.

Given that OA and TA modulate not only motor behavior, but also learning and memory (Burke et al., 2012; Waddell, 2013; Wu et al., 2013) as well as states of motivation, aggression and addiction (McClung and Hirsh, 1999; Scholz et al., 2005; Hoyer et al., 2008; Zhou et al., 2008), identification of the molecular mechanisms of degradation and the cellular sites of action are likely of broad interest.