Compartmentalization of neuronal and peripheral serotonin synthesis in Drosophila melanogaster


*W. S. Neckameyer, Department of Pharmacological and Physiological Science, 1402 South Grand Boulevard, St Louis, University School of Medicine, MO 63104, USA. E-mail:


In Drosophila, one enzyme (Drosophila tryptophan-phenylalanine hydroxylase, DTPHu) hydroxylates both tryptophan to yield 5-hydroxytryptophan, the first step in serotonin synthesis, and phenylalanine, to generate tyrosine. Analysis of the sequenced Drosophila genome identified an additional enzyme with extensive homology to mammalian tryptophan hydroxylase (TPH), which we have termed DTRHn. We have shown that DTRHn can hydroxylate tryptophan in vitro but displays differential activity relative to DTPHu when using tryptophan as a substrate. Recent studies in mice identified the presence of two TPH genes, Tph1 and Tph2, from distinct genetic loci. Tph1 represents the non-neuronal TPH gene, and Tph2 is expressed exclusively in the brain. In this article, we show that DTRHn is neuronal in expression and function and thus represents the Drosophila homologue of Tph2. Using a DTRHn-null mutation, we show that diminished neuronal serotonin affects locomotor, olfactory and feeding behaviors, as well as heart rate. We also show that DTPHu functions in vivo as a phenylalanine hydroxylase in addition to its role as the peripheral TPH in Drosophila, and is critical for non-neuronal developmental events.

Biogenic amines are ancient, evolutionarily conserved molecules, which function in many diverse physiological contexts in both vertebrate and invertebrate species, including modulation of behavior, fertility, reproduction and the development of neuronal and non-neuronal tissues (reviewed by Monastirioti 1999). Perturbation of biogenic amine signaling pathways has been correlated with serious neurological diseases in humans, such as depression, obsessive-compulsive disorder (Mockus & Vrana 1998) and Parkinson’s disease (Cao et al. 1995).

The biogenic amines are formed by the actions of the aromatic amino acid hydroxylases: tryptophan hydroxylase (TPH) (EC, tyrosine hydroxylase (TH) (EC and phenylalanine hydroxylase (PAH) (EC TPH catalyzes the synthesis of 5–hydroxytryptophan from tryptophan, the first and rate-limiting step in serotonin (5–hydroxytryptamine, 5-HT) biosynthesis; similarly, TH catalyzes the formation of L-dihydroxyphenylalanine from tyrosine, the first and rate-limiting step in dopamine biosynthesis. PAH is a metabolic enzyme that catalyzes the conversion of phenylalanine to tyrosine, a rate-limiting step in phenylalanine catabolism, and protein and neurotransmitter biosynthesis. The genes encoding these enzymes have likely derived from an ancestral aromatic amino acid hydroxylase, as there is extensive sequence and biochemical conservation between them.

It has recently been discovered that in mammals, there are two distinct and largely nonoverlapping TPH enzymes. The first, encoded by Tph1, is expressed in the periphery, and the second, Tph2, is expressed within the central nervous system (CNS) (Walther et al. 2003; Zhang et al. 2004). In mice, loss of Tph1 expression results in cardiac dysfunction (Cote et al. 2003). Polymorphisms within the Tph2 locus have been associated with major depressive disorder (Harvey et al. 2004), and a loss of function mutation within Tph2 has been found within a significant proportion of patients with unipolar major depression (Zhang et al. 2005). While these data provide strong evidence for compartmentalization of expression and function of the two 5-HT-synthesizing enzymes, there are limited genetic tools available in mammalian systems to fully assess their individual contributions to 5-HT synthesis.

5–hydroxytryptamine is the final product of the actions of the TPH enzymes. In mammals, 5-HT regulates appetite, sleep, learning and memory, temperature regulation and sexual behavior (Wilkinson & Dourish 1991). There is an overwhelming body of evidence suggesting that 5-HT signaling is a major modulator of emotional behavior including anxiety, impulsivity and aggression, as well as integrated complex brain functions such as cognition, sensory processing and motor activity (Westenberg 1996); serotonin has also been implicated in the actions of numerous drugs of abuse (Lucki & Wieland 1990). Recent studies in mammals show that Tph2 is responsible for the rhythmic fluctuations in 5-HT synthesis (Malek et al. 2005), which modulates entrainment of the circadian clock. 5–hydroxytryptamine also serves as a hormone, cytokine, biological modifier, growth factor and regulator of vascular tone and intestinal activity (Hoyer et al. 2002; Kroeze et al. 2002; Seuwen & Pouyssegur 1990; Talley 2001). It has been suggested that 5-HT orchestrates the activity and interaction of several other transmitter systems, which can account for the diversity of these roles. In Drosophila and other insects, 5-HT also modulates circadian rhythms (Yuan et al. 2005), reproduction (Barreteau et al. 1991), feeding (Novak & Rowley 1994; Novak et al. 1995), heart rate (Dasari & Cooper 2006) and locomotion (Kamyshev et al. 1983), evidence for conservation of these numerous modulatory roles for 5-HT. This is consistent with the observation that there are only ∼100 serotonergic neurons within the Drosophila CNS but, as in higher vertebrates, these neurons send projections to most parts of the nervous system (Valles & White 1988).

In mammals, 5-HT is first detected in the early stages of gastrulation where it appears to regulate cell division and migration (Lauder et al. 1981). It is one of the first biogenic amines to appear in the developing CNS (Lauder & Bloom 1974); high levels of 5-HT are detected within the cells before and during early neurite outgrowth, suggesting that it might serve a trophic role during development prior to assuming its adult functions of neurotransmission and neuromodulation (Azmitia & Whitaker-Azmitia 1991). In Drosophila, 5-HT plays a vital role in early embryonic development, before the appearance of neuroblasts (Colas et al. 1999a,b), and studies in other insects, primarily Manduca sexta, have shown a requirement for 5-HT in the development of olfactory glomeruli (Mercer et al. 1995, 1996; Oland et al. 1995), suggesting the neuronal and non-neuronal developmental roles for 5-HT have also been conserved.

The predicted size of mammalian Tph1 is 51 kDa, which is similar to that of Drosophila tryptophan-phenylalanine hydroxylase (DTPHu) (50 kDa); the predicted size of mammalian Tph2 is 57 kDa, similar to the size of DTRHn (61 kDa); DTRHn shares slightly more sequence homology with Tph2 than with Tph1 (Coleman & Neckameyer 2005). Like mammalian Tph2, a larger N-terminal domain accounts for the increased size of DTRHn relative to Tph1 and DTPHu. Both DTPHu and DTRHn are subject to substrate inhibition by tryptophan, although not by phenylalanine (Coleman & Neckameyer 2004, 2005), which is similar to Tph1 and Tph2 (McKinney et al. 2004, 2005). Both Tph1 and Tph2 are phosphorylated by cyclic adenosine 5′-phosphate-dependent protein kinase; whereas DTPHu TPH activity is increased after phosphorylation by this kinase, that of DTRHn remains unchanged (Coleman & Neckameyer 2004, 2005), which is what has been observed for recombinant Tph2 (McKinney et al. 2005). These data suggest that mechanisms for biochemical regulation of the two enzymes have been conserved.

Given its numerous critical and vital functions, 5-HT levels must be appropriately regulated to correctly modulate these multifunctional roles. A mammalian knockout of Tph2 is not yet available, and assessing behavioral roles for the Tph1 mouse knockout is problematic because these mice die shortly after birth from cardiac malfunction. The genetic tractability of the Drosophila model system, which permits targeted in vivo manipulation of 5-HT synthesis in specific tissues, circumvents these issues. In this article, we show that, like mammals, Drosophila has compartmentalized the neuronal and peripheral regulation of serotonin by having two distinct enzymes, each specific for neuronal or peripheral synthesis, suggesting that this is an ancient and highly conserved mechanism.

Materials and methods

Fly culture

Flies were maintained in pint bottles containing standard agar–cornmeal–yeast food at 25°C on a 12-h light–dark cycle. Staged embryos were collected from a population cage containing ∼100 adult flies maintained at 25°C on a 12-h light–dark cycle. Females were allowed to lay eggs for 4 h on apple juice–agar plates; staged larvae were collected by maintaining plates with newly deposited eggs at 25°C for 20 h and removing the hatched first instars by migration of the animals onto yeast paste in the center of the agar plate. Second and third-instar larvae were collected by allowing first instars to age for the appropriate times. Newly eclosed male and female adults were collected after light anesthesia under CO2 and maintained separately in food vials at 25°C. For 5-HT reuptake studies, larvae were aged for 2 days before exposure for 24 h to yeast paste (0.5 g yeast/1 ml H2O) containing 5 mg/ml of the selective serotonin reuptake inhibitor fluoxetine-HCl (Ivax, Miami, FL, USA).

Fly strains

All strains were maintained as described above.

Canton S (CS): the wild-type laboratory strain used for enzymatic and developmental expression analyses.

w1118: parental strain for insertional mutations and the RNA interference (RNAi) transgenic lines. These flies have white eyes and are isogenic for the second and third chromosomes.

pBac{PB}CG9122c01440: insertional mutation in the DTRHn gene generated by Excelexis, Inc. and obtained from the Bloomington Stock Center (Bloomington, IN, USA), first described by Thibault et al. (2004). We refer to this stock as pBacTRH.

pBac{WH}plef01945: insertional mutation in the gene encoding TH (pale) generated by Excelexis, Inc., first described by Thibault et al. (2004), and obtained from the Bloomington Stock Center. We refer to this stock as pBacple. This stock is homozygous lethal, and the insertion is maintained over the TM6B, Tb1third chromosome balancer; only one functional copy of pale is required for wild-type expression of TH levels, so these flies have normal TH levels.

w*; P{GAL4-da.G32}UH1: GAL4 line carrying the daughterless (da) promoter on chromosome 3.

y*, w*; P{w+[mc]-UAS-2XEYFP : carries a upstream activator sequence enhanced yellow fluorescent protein construct on chromosome 2 for visualization of tissues and organs under a narrow green excitation wavelength.

P[w, CgGAL4]: GAL4 line carrying sequences that drive expression within larval and adult fat body; expression is initiated during mid-first instar. The gift of Dr Deborah Hoshizaki.

Generation of transgenic RNAi lines

A complementary DNA encoding the full-length DTPHu protein (Neckameyer & White 1992) was subcloned downstream of yeast UAS in the SympUAST vector (Giordano et al. 2002). Using this vector, in which the DTPHu coding region is symmetrically transcribed in both the sense and antisense directions, repression of DTPHu expression was accomplished with only a single copy of both the transgene and promoter. This construct was injected into 20-min w1118embryos with pπ25.7wc as a source of transposase using standard germ line transformation protocols (Robertson et al. 1988). Seven independent transformant lines were obtained and genetically mapped to the X, second or third chromosomes. One line, DTPHu G, mapped to chromosome 2, was maintained over UAS-2XEYFP and crossed to daGAL4 to direct panembryonic expression. An additional line, DTPHu F, was mapped to the third chromosome, and maintained over the TM6 balancer, which is marked with Tubby (Tb), a larval marker. These flies were crossed to the CgGAL4 driver to direct expression in fat body.

Enzymatic analyses

Protein extracts from developmentally staged Drosophila tissues were added to the assay reaction composed of 50 mm N-(2-Hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid) (HEPES) (pH 7.0), 50 μm substrate (L-tryptophan or L-phenylalanine), 5 mm dithiothreitol, 10 μm Fe(NH4)2(SO4)2, 50 μm BH4(the tetrahydrobiopterin cofactor), 0.1 mg/ml catalase and 1 μCi 3H–tryptophan or phenylalanine (∼30 Ci/mmol; Amersham, Piscataway, NJ, USA). Four hundred micrograms of 0- to 3-h embryos, 300 μg 10- to 14-h embryos, 400 μg second-instar larvae, 200 μg white prepupae, 200 μg 2-day pupae, 150 μg adult heads and 250 μg adult bodies were determined as optimal concentrations with which to assay. The assay was performed as described by Coleman & Neckameyer (2005). Each point was performed in duplicate; the data represent the average value of each point. Each assay was repeated four to six times. No activity was observed without the addition of iron, BH4 or catalase.

Similar assays were also performed to assess the decrease in activity after knockdown of DTPHu by RNAi. Two hundred micrograms of protein extracts derived from first-instar larvae from knockdown (da-RNAi or CgGal4-RNAi) and control siblings (da-FP or CgGal4-Tb) were analyzed for tryptophan or phenylalanine hydroxylation.

Western immunoblot analysis

Relative DTRHn and DTPHu protein levels were assessed by extracting crude protein from staged animals. The samples were ground with a Teflon pestle in phosphate-buffered saline (PBS) with protease inhibitors, and centrifuged to remove cuticular debris. Protein samples (400 μg) were subjected to Western immunoblot analysis using standard protocols (Harlow & Lane 1988), and were done in duplicate to ensure reproducibility. The filters were incubated with preabsorbed antibody, and the signals visualized using the electrochemiluminescence detection system (Amersham Pharmacia). Preabsorption of the antibody was accomplished by adding the appropriate amount of each antibody to 70 μg of either DTRHn or DTPHu crude bacterial supernatant in 100 μl PBS and incubated for 2 h at 37°C followed by an overnight incubation at 4°C. The preabsorbed antibody was spun for 15 min at 15 000 g at 4°C, and the pellet was resuspended in ‘blotto’ (5% nonfat dry milk/0.2% Tween-20, 1× PBS) at a final dilution of 1:5000. Characterization of these antibodies has been previously described (Coleman & Neckameyer 2004, 2005).


Wild-type females were allowed to lay eggs for either 2 or 4 h, which were aged until the appropriate time for harvesting. Embryos were collected, washed, dechorionated, fixed (in 4% formaldehyde, 1× PBS, 50 mm ethyleneglycoltetraacetic acid plus an equal volume of heptane) and then devitillinized. The embryos were washed with methanol and stored at −20°C. An aliquot of the settled volume of embryos was transferred into new tubes and treated with 1.5% hydrogen peroxide in methanol, rinsed in methanol and washed in PBT (1× PBS, 0.1% protease-free bovine serum albumin, 0.1% Triton-X-100). The embryos were blocked with 0.2% goat serum (Vector Laboratories, Burlingame, CA, USA), and preabsorbed primary antibody was added at a final dilution of 1:1000 for αDTPHu and 1:2000 for αDTRHn. The embryos were extensively washed in PBT after removal of the primary antibody and then blocked in the same manner as before with PBT + 2% goat serum. After incubation of the staged embryos with preabsorbed secondary antibody, the embryos were washed and stained using VectaStain Elite ABC (Vector Laboratories). Preabsorption of the secondary antibody was accomplished by fixing a 0- to 16-h collection of wild-type embryos and incubating them with 1.5% normal goat serum and horse radish peroxidase-conjugated secondary antibody (Vector Laboratories). The 3, 3′-diaminobenzidine (DAB) solution was added to the embryos, which were gently swirled and allowed to stand until color development. The embryos were then rinsed with 1× PBS, dehydrated through a sequential ethanol series (30%, 50% and 70%, 100%), rinsed in methyl salicylate and mounted in Canada balsam. The slides were placed at 42°C to dry.

Larval brains were dissected from wild-type Drosophila larvae in 1× PBS and fixed for 1 h in 4% formaldehyde, followed by extensive washing in PBT. The brains were incubated in preabsorbed primary antibody to a final dilution of 1:1000. After extensive washing in PBT, the brains were incubated with secondary antibody (used at 1:350 dilution; Alexa 488 or Alexa 568 anti-rabbit immunoglobulin G, Molecular Probes, Carlsbad, CA, USA) followed by PBT washes. The brains were then incubated in 4 mm sodium carbonate (pH 9) and mounted in 4% n-propyl gallate in 20 mm sodium carbonate (pH 9.5), 80% glycerol. For colocalization studies, brains were incubated with preabsorbed DTPHu antibody and with a 1:30 dilution of antiserotonin antibody (Spring Bioscience, Fremont, CA, USA), followed by the Alexa 568 anti-rabbit and Alexa 488 anti-mouse antibodies. The brains were then incubated with the Alexa 568 (anti-mouse) and Alexa 488 (anti-mouse) secondary antibodies. For colocalization of DTPH and Drosophila tyrosine hydroxylase (DTH), the brains were incubated with DTH antibody at a dilution of 1:9000 followed by the Alexa 488 anti-rabbit secondary antibody. After extensive washing, the brains were incubated with preabsorbed DTPHu antibody followed by the Alexa 568 anti-rabbit secondary. The DTH antibody was raised in rabbits against bacterially generated enzyme and was first described by Neckameyer et al. (2000).

Developmental analyses

Statistical analyses were accomplished by one-way analysis of variance (anova) using Tukey’s post-test or by Student’s t-test.

Embryonic survival

Population cages were established containing w1118(the parental stock), or the insertional mutations pBacTRH or pBacple, and flies were allowed to lay eggs for 4 h. Individual embryos were then aligned on egg-laying plates and maintained at 25°C. The number of hatched eggs was determined after 24 h.

Larval and pupal development

Population cages were established containing w1118, pBacTRH or pBacple, and embryos were collected overnight. Newly hatched larvae were collected after 24 h and allowed to age for 48 h, after which 25 larvae were placed in food vials and maintained at 25°C until eclosion from the pupal case.

Adult survival

Ten flies from each of the above genotypes were placed in food vials and kept at 25°C. Separate vials were maintained for female and male flies. The number of dead flies in each vial was determined every 2 days for a 20-day observation period.

Larval behaviors

Statistical analyses were accomplished by one-way anova using Tukey’s post-test or by Student’s ttest.

Measurement of body wall and mouth hook contractions

A single second-instar larva was placed on a 2% agar substrate and allowed to acclimate for 30 seconds. The larva was then observed as it crawled over the substrate for a period of 1 min, and each anterior to posterior contractile wave was counted. To assay mouth hook contractions, a larva was placed on the agar substrate containing a thin layer of 2% yeast solution. In this medium, larvae engage in feeding activity by contracting their mouth hooks via pharyngeal muscles to shovel yeast solution. Following a pre-equilibration period of 30 seconds, mouth hook contractions were counted over a 1-min period.

Heart rate

The heart is visible through the cuticle of the white prepupa, which was placed on a stage and viewed using a stereozoom microscope (Olympus, Hitschfel Instruments, St Louis, MO, USA) at ×40 magnification in a temperature-controlled room. The number of beats was counted over 15 seconds, and the average of five separate observation periods, with 15-second intervals in between, was used to assess heart rate.

Odor preference

Approximately 50 larvae were placed in the center of a 2% agar plate in which two 1.5-cm filter discs were positioned at opposite ends of the plate. For the experimental odor preference assay, one filter disc was saturated with 7 μl water and the other with 7 μl odorant (nonanol or heptanol); for control assays, both discs were saturated with water. The orientation of the sides (+/- odorant) was randomly changed to control for environmental factors during the assay. Response indices for aversion (RIs) were obtained by counting the number of larvae present on the half-side of the plate with the water-saturated filter disc divided by the total number of larvae that migrated throughout both halves of the dish. Larvae remaining in the center of the dish were not counted. Control assays should yield an RI value of approximately 0.5, indicating an equal distribution.

Adult behaviors

Newly eclosed adults were separated by sex and maintained under standard conditions (30/vial) for 5 days before assaying for behavior. Statistical analyses were accomplished by one-way anova using Tukey’s post-test or by Student’s t-test.

Locomotor assays

General activity for adult animals was assessed using a simple locomotor paradigm (Neckameyer 1998). A single fly was aspirated into a 60-mm Petri dish marked with a grid of 1-cm squares, allowed to recover for 30 seconds and locomotor activity was observed for the first 2 and last 2 min of a 15-min period (there is a trend for locomotor activity to diminish over time, Neckameyer 1998). The number of grid lines crossed during each observation period was recorded.

Feeding experiments

Animals were placed in a vial containing 0.5 ml 2% yeast/5% sucrose/0.5% blue food coloring (FD&C #1) placed on a filter paper disc. A total of 25 flies from each treatment were collected after 24 h, frozen in liquid nitrogen and homogenized in 0.1 m phosphate buffer (pH 7.2). The homogenates were spun at 13 000 g for 2 min, and an aliquot was transferred to a new tube and recentrifuged. An aliquot from the second spin was analyzed spectrophotometrically by reading absorbance at 625 nm (Edgecomb et al. 1994). Background absorbance was corrected by using homogenates from 25 flies maintained on yeast/sucrose without food coloring. These analyses were repeated five times.


DTRHn and DTPHu enzymatic activities and expression are spatially distinct throughout development

Developmental comparisons of DTPHu and DTRHn enzyme activities and protein levels were undertaken to determine whether they were subject to differential regulation. Analysis of TPH and PAH activities throughout development showed that phenylalanine hydroxylation is far more robust than tryptophan hydroxylation, except during early embryogenesis (Fig. 1a–f); the 0- to 2-h embryonic stage (Fig. 1a) reflects maternal deposition of protein rather than zygotically expressed enzyme. However, the activity of both enzymes changed throughout development. Similarly, Western immunoblot analyses of protein extracts from developmentally staged animals showed a dynamic pattern for DTPHu, but DTRHn protein expression was undetectable (Fig. 1g; the blot visualized with anti-DTRHn was stripped and incubated with anti-DTPHu to ensure equal loading of the blots, data not shown).

Figure 1.

Developmental expression of PAH and TPH activities. Protein extracts from (a) 0- to 3-h embryos, (b) second-instar larvae, (c) white prepupae, (d) 2-day pupae and (e) 1-day-old heads and (f) bodies were analyzed for (a–f) TPH and PAH activities and (g) by Western immunoblot analysis. eE, 0- to 3-h embryos; lE, 10- to 14-h embryos; 2°L, second-instar larvae; wpp, white prepupae; 2dp, 2-day pupae; H, adult heads; B, adult bodies. Protein was visualized using polyclonal antibodies raised against either DTPHu (preincubated with DTRHn protein) or DTRHn (preincubated with DTPHu protein); preincubation was necessary to eliminate cross-reactivity arising from shared epitopes. Generation and characterization of these antibodies have been previously described (Coleman & Neckameyer 2004 2005). Standard error of the mean is denoted by the line above each bar. Phe, phenylalanine hydroxylation; Trp, tryptophan hydroxylation.

Levels of DTPHu protein do not correlate with enzymatic activity for PAH: second-instar larvae have low DTPHu protein levels, but high enzyme activity; the highest DTPHu protein level is observed in white prepupae, but the activity at that stage is less than that in second-instar larvae. Head and body tissue have approximately equal amounts of DTPHu protein, but there is greater enzymatic activity in heads. The relative contribution of DTRHn to tryptophan hydroxylation is small because flies lacking DTRHn protein and neuronal 5-HT show no change in tryptophan hydroxylation activity (see Fig. 4i,j). These data suggest that the activity of both enzymes is differentially regulated in vivo.

Figure 4.

The pBacTRH mutation lacks neuronal DTRHn and displays significantly diminished serotonin immunoreactivity. The CNSs were dissected from wandering third (a, b) or early second-instar larvae (c, d) from the pBacTRH mutation as well as from the parental line (w1118) and incubated with antibodies raised against DTRHn or serotonin. The DTRHn immunofluorescence is not observed in the pBacTRH brain (a), but the expected pattern is seen in the parental line, w1118(b). The 5-HT immunoreactivity is significantly decreased in the pBacTRH brain (c) relative to the parental stock (d). All samples were incubated in parallel, and photomicrographs were taken at the same exposure. (e–h) Fluoxetine treatment does not affect 5-HT immunoreactivity in w1118brains (compare w1118brain e, with fluoxetine treated, f), but this treatment does decrease 5-HT immunoreactivity in pBacTRH brains (compare control, g, with treated, h). First-instar larvae were fed 5 mg/ml fluoxetine-HCl for 24 h. The w1118brains were photographed using a 50-millisecond exposure; the pBacTRH brains required a 400-millisecond exposure so that the immunoreactivity could be observed. 0.2 cm = 20 μm. (i) The DTPHu in head tissue can function as a TPH. TPH activity levels in protein extracts from w1118, pBacple and pBacTRH heads. Specific activity is defined as picomoles 5-hydroxytryptophan formed per minute per milligram protein. Standard error of the mean is denoted by the line above each bar. (n) Western immunoblot analysis of DTPHu protein from w1118, pBacple and pBacTRH heads.

We then determined the temporal and spatial expression of DTPHu and DTRHn to determine whether there was any overlap, which would provide evidence for redundant function. Because of the high degree of similarity between the two proteins, antiserum raised against DTPHu was preabsorbed with DTRHn protein (and vice versa) to eliminate cross-reactivity from shared epitopes. Immunohistochemical analyses showed that DTPHu is ubiquitously expressed in 0- to 2-h embryos (Fig. 2a), the result of a maternal contribution. The DTPHu is also expressed in cells surrounding invaginating tissues in 2- to 6-h embryos (Fig. 2d); later expression is observed in fat body (Fig. 2g), which carries out many of the activities carried out by the human liver (Hoshizaki 2005). Fat body constitutes a substantial proportion of the embryo, and the high degree of DTPHu expression in this tissue may obscure other patterns. The DTPHu expression is also found within a small subset of developing neuroblasts (Fig. 2j) in a pattern highly reminiscent of Drosophila TH (Neckameyer 1996), suggesting that these cells may be dopaminergic neuron precursors. This is consistent with previous observations that DTPHu is expressed in dopaminergic neurons within the larval CNS (Neckameyer & White 1992; Fig. 3).

Figure 2.

Embryonic expression patterns for DTPHu and DTRHn proteins are discrete and nonoverlapping. Staged wild-type (CS) embryos were subjected to immunohistochemistry to determine expression patterns for DTPHu (a, d, g, j,) and DTRHn (b, e, h, k). a–c, 0- to 2-h embryos; d–f, 2- to 6-h embryos; g–k, 14- to 18-h embryos. c, f, i, control embryos (no primary antibody). For a–i, 0.2 cm = 40 μm. For j–k, 0.3 cm = 10 μm.

Figure 3.

DTRHn expression is limited to serotonergic neurons and DTPHu expression is limited to dopaminergic neurons. Brains dissected from wandering third-instar larvae were subjected to immunohistochemistry to determine expression patterns for DTRHn and DTPHu. (a) DTRH antibody absorbed with DTPH protein and visualized using the Alexa 568 anti-rabbit secondary antibody. (b) Antiserotonin antibody visualized with the Alexa 488 anti-mouse secondary antibody. (c) Merge. Only the ventral ganglion, highlighting the paired symmetrical serotonergic neurons, is shown. (d, g) DTPH antibody absorbed with DTRH protein and visualized using the Alexa 568 anti-rabbit secondary antibody. (e, h) The same brain treated sequentially with DTH antibody and the Alexa 488 anti-rabbit secondary antibody. (f, i) Merge. (d–f) Dorsolateral neurons. (g–i) Medial neurons. 0.5 cm = 50 μm.

While DTRHn is also ubiquitously expressed in 0- to 2-h embryos (Fig. 2b), again suggesting maternal deposition, no other expression for this protein can be detected (Fig. 2e) until late in embryogenesis where it appears limited to a discrete subsets of cells within the CNS (Fig. 2h,k).

The patterns of dopaminergic and serotonergic neurons are distinct and nonoverlapping and have been described numerous times in the literature. Immunohistochemical analyses of wild-type CNS hand-dissected from larval instars showed that DTRHn antiserum, after preincubation with DTPHu protein, is limited to serotonergic neurons (Fig. 3a); this was verified by colocalization with an antibody raised against 5-HT (Fig. 3b,c). This is consistent with observations that Tph2 is expressed solely in the CNS (Zill et al. 2004). The DTPHu CNS expression is restricted to dopaminergic neurons (Fig. 3d–i). The DTPHu is expressed in the dorsal lateral (Fig. 3d) and the medial (Fig. 3g) dopaminergic neurons, and colocalizes with TH, the rate-limiting step in the synthesis of dopamine (Fig. 3e,f,h,i). It was previously believed that DTPHu was expressed in both dopaminergic and serotonergic neurons (Neckameyer & White 1992); however, these experiments show that the expression in serotonergic neurons was because of cross-reactivity with DTRHn.

A mutation in the DTRHn locus is viable and has diminished 5-HT immunoreactivity in the CNS

An insertional mutation residing within the DTRHn locus at 61F was generated by Excelexis (Thibault et al. 2004) and obtained from the Bloomington Stock Center (pBac{PB}CG9122c01440, referred to here as pBacTRH). Immunohistochemical analyses of brains from pBacTRH larval instars showed no staining for DTRHn protein (Fig. 3a,b) and diminished fluorescence for 5-HT (Fig. 3c). Brains from the parental stock (w1118) analyzed in parallel using antibodies raised against DTRHn (Fig. 3b) or 5-HT (Fig. 3d) displayed the expected pattern of serotonergic neurons. To determine whether the 5-HT immunoreactivity observed in the pBacTRH brains was the result of uptake from circulating 5-HT synthesized by DTPHu, 2-day old larvae from both the w1118and pBacTRH strains were fed 5 mg/ml of the selective serotonin reuptake inhibitor, fluoxetine, for 24 h, and their brains were dissected and analyzed for 5-HT immunoreactivity. This drug has been shown to act selectively in vitro on the Drosophila serotonin transporter, which shares a high degree of conservation with its mammalian counterparts (Demchyshyn et al. 1994). No change in fluorescent intensity was observed in 5-HT immunoreactivity between w1118control (Fig. 3e) and fluoxetine-treated (Fig. 3f) brains (photomicrograph exposed for 50 milliseconds). However, there was a clearly observable drop in 5-HT immunoreactivity between pBacTRH control (Fig. 3g) and fluoxetine-treated (Fig. 3h) animals (exposed for 400 milliseconds), strong evidence that the limited 5-HT immunoreactivity observed in pBacTRH brains was due to reuptake from circulating 5-HT generated via DTPHu.

The DTPHu tryptophan hydroxylating activity in head tissues is unaffected by the DTRHn mutation because there is no difference in tryptophan hydroxylation activity in protein extracts from head tissues from w1118, pBacple and pBacTRH flies (pBacple flies carry a single copy of an insertion disrupting the TH gene maintained over a third chromosome balancer, and which is in the same w1118background) (Fig. 3i). The DTPHu protein levels are also unaffected by the mutation in DTRHn (Fig. 3j). This is not surprising because there is a substantial amount of fat body surrounding the brain, as well as cuticular tissue, and DTPHu is strongly expressed in both tissues (Neckameyer & White 1992; Fig. 2). The results of these experiments, in addition to those described in Figs 2 and 3, show that DTRHn is required for production of neuronal serotonin and is therefore the Drosophila orthologue of mammalian Tph2.

The DTRHn-null mutation displays behavioral aberrations

5–Hydroxytryptamine modulates feeding behaviors in mammals (reviewed by Lucki 1998); in Drosophila, 5-HT innervates the feeding apparatus (Monastirioti et al. 1999), suggesting a role for 5-HT in feeding in this species as well. We therefore assayed DTRHn mutants for defects in this behavior at the larval and adult stages (Figs 5a and 6b). As expected, larval feeding (measured by the number of mouth hook contractions in a yeast solution) is significantly affected by the decrease in neuronal 5-HT (Fig. 5a); this is unlikely to be due to aberrant motor control because body wall contractions (measured on an agar substrate) were unaffected (Fig. 5b).

Figure 5.

Neuronal serotonin is required for specific larval behaviors. Staged second-instar larvae from the w1118, pBacTRH and pBacple lines were analyzed for larval behaviors. (a) Feeding. pBacTRH animals display significantly reduced numbers of contractions (P < 0.0001). (b) Body wall contractions. There is no significant difference in locomotor behavior in the larva. (d) Odor response. nonanol. w1118animals were neither repulsed nor attracted by this odorant (P > 0.05), while pBacTRH animals were highly aversive to the odor (P < 0.01). n = the number in parentheses above each bar. (c) White prepupae from the same genotypes were assessed for heart rate. n = 20 for each genotype. *P < 0.05, **P < 0.01, ***P < 0.0001, a–c, one-way anova; d, Student’s t-test. Standard error of the mean is denoted by the line above each bar. c. Heart rate is significantly reduced in pBacTRH prepupae (P < 0.0001).

Figure 6.

Neuronal serotonin is required for specific adult behaviors. (a) Locomotion. Females lacking neuronal serotonin display significantly reduced locomotor activity. Animals were assayed for the (a) first 2 min or (b) last 2 min of a 15-min observation period. n = 40 and 41, for pBacTRH females and pBacple females, respectively. (b) Feeding. Males and females lacking DTRHn show a significantly reduced feeding ability. n = 4 independent assays. Standard error of the mean is denoted by the line above each bar. *P < 0.05, **P < 0.01, ***P < 0.001, Student’s t-tests.

The larval heart is myogenic in origin, but circulating transmitters, including 5-HT, affect its function because exogenously applied 5-HT increases the Drosophila heart rate at all stages (Zornik et al. 1999). Therefore, we assessed the heart rate of white prepupa from the pBacTRH mutant line as well as from w1118and pBacple. Heart rate was assessed in white prepupae for the following reasons: (1) it is an extremely tight developmental window, lasting only 20 min, and is easily and clearly identifiable; (2) the larval CNS has not begun to histolyze; (3) the animal is immobile and translucent, facilitating quantitation of heart rate; and (4) this is a common stage for analysis of heart rate. Heart rate was significantly reduced in pBacTRH animals relative to the parental strain (w1118) as well as to pBacple (Fig. 5c), suggesting that the neuronal actions of 5-HT are critical for normal heart function at this stage.

In both Bombyx mori and M. sexta, 5-HT stimulates neurite extension and branching in cultured olfactory neurons (Mercer et al. 1995b; Park et al. 2003), suggesting a role for 5-HT in olfaction, so this behavior was assayed in larvae. In olfaction paradigms, pBacTRH mutants displayed an unusually strong aversion to the odorant nonanol (Fig. 5d); w1118larvae were not sensitive to this odor. Wild-type CS larvae also showed no apparent choice when presented with nonanol (data not shown). However, larvae from both w1118and pBacTRH genotypes were strongly and equally attracted to the odorant heptanol (data not shown), suggesting that the sensitivity to nonanol may reflect a specific change in odorant processing in animals with decreased neuronal 5-HT.

5–Hydroxytryptamine has also been shown to modulate locomotion in mammals (Lucki 1988), so this behavior was assessed in adults. Adult motor behavior in DTRHn mutants was significantly depressed in females, showing a requirement for 5-HT in the modulation of locomotion (Fig. 6a). These flies were compared with pBacple individuals because the parental strain is completely blind (which would confound the results in this paradigm). Consistent with what was observed in larvae, adult feeding behavior was also significantly decreased relative to the parental control stock in both males and females (Fig. 6b).

Geotactic ability and the ability to right if placed in a prone position are behaviors modulated by γ-aminobutyric acid (GABA) (Leal & Neckameyer 2002), and were unaffected in the DTRHn mutant (data not shown), showing specificity for the assessed behaviors.

DTRHn is not required for embryonic or adult viability

Given the known role for 5-HT in early, non-neuronal developmental events (e.g. Colas et al. 1999a), we assessed viability of the DTRHn mutant at embryonic and other stages. Embryonic survival of DTRHn mutants is actually enhanced relative to the parental strain and pBacple (Fig. 7a). Therefore, DTRHn activity cannot be vital during embryogenesis. The number of eclosing animals is decreased in the DTRHn mutant relative to controls (Fig. 7b), suggesting that DTRHn activity is critical, but not vital, during the larval and/or pupal stages. Adult survival does not require DTRHn activity for either female (Fig. 7c) or male (Fig. 7d) flies.

Figure 7.

Lack of neuronal serotonin negatively affects larval but not embryonic or adult survival. The pBacTRH line was compared with both the parental line (w1118) and pBacple (a similar insertional mutant at a different locus). (a) Number of eggs hatched 30 h after egg laying at 24°C (out of 10 embryos). n = 18, 25 and 34 sets of 10 for w1118, pBacTRH and pBacple, respectively. The pBacTRH animals display significantly greater hatching than either of the controls (P < 0.0001, one-way anova, Tukey’s post-test). (b) Number of adults eclosing from vials containing 25 staged second-instar larvae maintained at 24°C. n = 18, 16 and 15 vials for w1118, pBacTRH and pBacple, respectively. The pBacTRH animals display significantly fewer eclosing adults than either of the controls (P < 0.01, one-way anova, Tukey’s post-test). (c, d) Survival curves for adult females (c) and males (d). Newly eclosed adult flies from the w1118, pBacple or pBacTRH were maintained in vials at 24°C. Number of dead animals was counted every 2 days for a 20-day observation period. n = 19, 10 and 17 vials for w1118, pBacTRH and pBacple females, and 20, 10 and 17 for males. There is no significant difference in adult survival through 20 days of age for either sex. Standard error of the mean is denoted by the line above each bar.

DTPHu is vitally required for early embryogenesis and functions as a PAH in fat body.

We used a transgenic GAL4 strain carrying the da promoter, which drives ubiquitous embryonic expression within the developing embryo (Wodarz et al. 1995), to knockdown DTPHu levels at this developmental stage. Because a deletion spanning the DTPHu locus, which phenocopies the 5-HT2Dro receptor mutation, is an embryonic lethal (Colas et al. 1999a), we anticipated that knockdown of DTPHu in the embryo would result in significant lethality. The da is a class I helix-loop-helix transcription factor required for numerous developmental events in the Drosophila embryo. Most critically for our purposes, it is required for both mesoderm differentiation (Gonzalez-Crespo & Levine 1993) and establishment of the proneural field (Caudy et al. 1988). We generated crosses using a daGAL4 driver (w; P[w+, pda-gal]4 =da.G32) driving a second chromosome UAS-RNAi transgene for DTPHu maintained over a UAS-enhanced yellow fluorescent protein (y*, w*; P{w+[mc]-UAS-2XEYFP}). This permitted easy identification of control (da/2XEYFP, and thus fluorescent) and experimental (da/RNAi, not fluorescent) embryos, generated in the same cross and thus exposed to identical environmental conditions. Knockdown of expression was confirmed by Western immunoblot analysis for DTPHu (Fig. 8a) as well as by assessment of enzymatic activity (Fig. 8b). Enzymatic activity of TPH in the knockdown transgenic flies is clearly decreased, although it is not statistically significant. This was not unexpected because the animals were assayed in the larval stage, where DTPHu PAH is enhanced via post-translational modifications (see Fig. 1). Significant lethality was observed in the DTPHu knockdown embryos (Fig. 8c). However, later development was unaffected because adult eclosion was unchanged (Fig. 8d). These data show that although control and knockdown embryos are laid in essentially equivalent amounts, those with a knockdown of DTPHu hatch at a significantly diminished rate, and postembryonic viability is unaffected, consistent with a vital role for 5-HT and DTPHu activity in embryogenesis.

Figure 8.

Knockdown of DTPHu expression in embryos. Transgenic animals carrying an integrated DTPHu RNAi construct, maintained over a UAS-enhanced yellow fluorescent protein, were crossed to a daGAL4 driver. Knockdown was assessed by Western immunoblot analysis of DTPHu levels (a) and by TPH enzymatic activity (b) of larval protein extracts from da/FP and da/DTPHu RNAi animals .The progeny were analyzed for (c) embryonic hatching and (d) pupal eclosion. Control, da/UAS-FP, n = 463 embryos; RNAi, da/DTPHu RNAi, n = 559 embryos. P < 0.0001, Fisher’s exact test. For eclosion studies, 10 independent vials of 25 larvae/vial were analyzed. Standard error of the mean is denoted by the line above each bar.

The DTPHu is also expressed in fat body (Fig. 2g), which carries out several of the activities carried out by the human liver, where PAH is highly expressed. Mutations at the PAH locus are the major cause of phenylketonuria (Benit et al. 1999), and hyperactivity is a noted behavioral aberration of untreated phenylketonurics (Yannicelli & Ryan 1995). While there is a broad spectrum of phenotypes, including cognitive disorders, we chose to assess locomotor activity in transgenic animals with reduced DTPHu levels in fat body. We used the CgGAL4 driver, which is expressed in fat body and hemocytes starting from the middle of the first larval instar (Asha et al. 2003), to drive a DTPHu RNAi transgenic construct and selectively knockdown DTPHu expression within this tissue. A DTPHu RNAi transgene was localized to the third chromosome and maintained over the TM6 balancer, marked with the larval marker Tb. These flies were crossed with flies carrying the CgGAL4 transgene, so that sibling controls (CgGAL4; Tb) were generated in the same cross as the CgGAL4; DTPHu RNAi animals, and thus exposed to the same environmental conditions. Knockdown was confirmed by immunoblot analysis of DTPHu levels and enzymatic analysis of DTPHu phenylalanine hydroxylation in both control and knockdown larvae (Fig. 9a). Locomotion was significantly increased in knockdown (Tb+) larvae relative to sibling controls (Tb) at both the first (3.1) and second (3.2) days of the third-instar larval stage (Fig. 9b). While mouth hook contractions were unaffected in 3.1 larvae, they were significantly increased in the knockdown larvae, with continued expression of the DTPHu double-stranded sRNA in fat body (3.2, Fig. 9c), suggesting that a buildup of phenylalanine was responsible for the hyperactivity.

Figure 9.

Knockdown of DTPHu expression in fat body results in hyperactivity. Transgenic animals carrying an integrated DTPHu RNAi construct, maintained over the third chromosome balancer TM6, marked with the bristle mutation Tb, were crossed to the CgGAL4 driver and assessed for decreases in levels of DTPHu protein and phenylalanine hydroxylating activity (a). The progeny were analyzed for larval locomotion (b), and feeding (c) at the first (3.1) and second (3.2) days of the third larval instar, and for heart rate at the white prepupal stage. Knockdown of DTPHu significantly increased both locomotor activity and feeding but did not affect heart rate. Tb, CgGAL4 /Tb control animals; Tb+, CgGAL4/DTPHu RNAi animals. b, c, n = 23 for each genotype. d, n = 20. Standard error of the mean is denoted by the line above each bar. P < 0.0001, Student’s t-test.

We also assessed heart rate in white prepupae from these crosses to determine whether 5-HT, synthesized via the actions of DTPHu in fat body and released into the circulating hemolymph, would affect function of the dorsal vessel. No changes in heart rate were observed (Fig. 9d).


DTPHu and DTRHn are differentially temporally and spatially expressed throughout development

Enzymatic activity of DTPHu is clearly greater than that of DTRHn (Fig. 1a–f), consistent with its more widespread expression at all stages, which would be expected for a metabolic enzyme such as PAH. The undetectable expression of DTRHn by Western immunoblot analysis is consistent with a neuronal-specific localization. There are only 84 larval and ∼100 adult serotonin-immunoreactive neurons within the Drosophila nervous system (reviewed by Monastirioti 1999), whereas the dual-function DTPHu is expressed ubiquitously in the early embryo, as well as in fat body, cuticle and the CNS (Fig. 2; Neckameyer & White 1992). These data strongly suggest that not only are the expression patterns of the two enzymes distinct and nonoverlapping, but that regulation of enzymatic activity of DTPHu and DTRHn is differentially controlled. While the two enzymes share many of the same biochemical characteristics (activation via phosphorylation of specific serine residues, requirement for the same cofactor; discussed by Coleman & Neckameyer 2005), the local milieu may differ, and the proteins may also be differentially translated and/or degraded.

DTPHu, and not DTRHn, synthesizes serotonin required for early developmental, non-neuronal events

Although 5-HT is required for a vital role in early embryogenesis in several species (reviewed by Nebigil et al. 2001), it is unlikely to be synthesized by DTRHn because flies carrying a null mutation in this gene are viable. In fact, DTRHn-null embryos display enhanced survival relative to controls (Fig. 5a). This is consistent with our observations (Fig. 2) that embryonic DTRHn expression is limited to the CNS, as is expression of the Drosophila 5-HT1A, 5-HT1B and 5-HT7 receptors (Sadou et al. 1992). It is also consistent with the observation that the tph-1 mutation in Caenorhabditis elegans, which does not synthesize neuronal 5-HT, is viable (Sze et al. 2000). Because zygotic 5-HT synthesis is critical for early embryonic viability (Colas et al. 1999a), it cannot be generated by DTRHn.

The number of DTRHn-null adult flies eclosing from pupal cases is significantly less than those from either the parental (w1118) or pBacple line, a similar pBac line with normal 5-HT levels also used as a control (Fig. 7b). Because, in this experiment, the number of eclosing adults was almost identical to the number of pupal cases (data not shown), pupal development is unlikely to require neuronal 5-HT synthesis. Given that animals lacking DTRHn display a diminished feeding capacity (Fig. 5a), this increase in lethality likely reflects poorer nutritional status at the larval stage, but it may also reflect aberrant endocrine regulation and/or perturbed neuronal circuitry. Diminished neuronal 5-HT does not affect adult survival because this is unaffected in the DTRHn mutation (Fig. 7c,d).

Embryonic DTPHu expression in early (stages 4–8) embryos is found in cells surrounding invaginating areas (Fig. 2d). This is consistent with the known role for 5-HT in directing normal germband expression because invaginating ventral mesodermal and posterior endodermal cells form the ventral furrow and proctodeal invagination to mark the beginning of germband extension. A peak of TPH activity precedes both expression of the 5-HT2Dro receptor and transient 5-HT concentration levels (Colas et al. 1995), and both a mutation within the 5-HT2Dro receptor and a deletion spanning the DTPHu locus result in the same embryonic lethal phenotype with defects in germband extension (Colas et al. 1999b). This developmental role for 5-HT2 receptors is evolutionarily conserved (reviewed by Gaspar et al. 2003). The early requirement for 5-HT appears to be conserved in mammals (Choi et al. 1997; Roux et al. 1995; Yavarone et al. 1993). The DTPHu does not colocalize with the Drosophila 5-HT2 receptor, which is found in a pattern of stripes reminiscent of the pair-rule genes at the cellular blastoderm stage (Colas et al. 1995). Synthesis of serotonin at this stage has been shown to be under zygotic control for TPH activity and is not the result of a maternal contribution (Colas et al. 1999a). Because no 5-HT or 5-HT2Dro protein is found in embryos of stages 1–4, the ubiquitous DTPHu expression observed in similarly staged embryos likely reflects maternal deposition of the protein to function as a PAH in the developing embryo. The function of the maternally deposited DTRHn protein is unknown.

Serotonin, generated via the actions of DTRHn, modulates specific behaviors in Drosophila

Drosophila with diminished neuronal 5-HT are still viable and fertile, suggesting that serotonin may modulate many behaviors, but is not the principal neurotransmitter for any; this is also consistent with what has been observed in mammals (reviewed by Lucki 1998). The gross morphology of the CNS is normal, and the stereotypical pattern of a related biogenic amine, dopamine, is unaffected in flies carrying the mutation in DTRHn (data not shown). Mammalian studies have also shown that although 5-HT modulates numerous neuronal developmental events, including neuronal migration and synaptogenesis, targeted knockouts in 5-HT receptors do not affect gross brain morphology (see Gaspar et al. 2003, for review). However, Pet-1 knockout mice (with ∼20% normal neuronal 5-HT levels), which have no obvious morphological changes in the CNS, also display distinct behavioral changes (Hendricks et al. 2003), and C. elegans carrying a mutation within neuronal TPH show no gross aberrations within the CNS but do display specific behavioral aberrations (Sze et al. 2000). The DTRHn-null mutant displays a significant decrease in feeding behavior relative to controls at both the larval and adult stages (Figs 5a and 6b), consistent with the observation that the pharyngeal muscles, the proventriculus and the midgut are putative targets of the larval serotonergic neurons (Valles & White 1988). The larval behavior is very similar to the defects described in pharyngeal pumping on a bacterial lawn for the tph-1 mutation in C. elegans (Sze et al. 2000). In studies from other species, acute reduction in 5-HT in adult animals increases feeding behavior, and increased 5-HT levels decrease feeding (e.g. Dacks et al. 2003). This may be more reflective of abnormal circuitry rather than acute modulation of these behaviors by 5-HT, which can be tested by selective and temporal knockout of DTRHn (W. Neckameyer, manuscript in preparation).

Significant reduction in peripheral 5-HT levels in Tph1 knockout mice results in cardiac dysfunction without obvious morphological aberrations (Cote et al. 2003), yet knockouts of the 5-HT2B receptor in mice causes embryonic lethality due to defects in heart development (Nebigil et al. 2000), demonstration that 5-HT plays a critical role in both development and function of cardiac tissues. Exogenously applied 5-HT increases the Drosophila heart rate at all stages (Zornik et al. 1999); consistent with this observation, our data show that diminished neuronal 5-HT significantly reduced heart rate at the white prepupal stage (Fig. 5c). While the larval heart is largely myogenic, recent studies have shown that there is also a direct neural input to the Drosophila larval aorta (Johnstone & Cooper 2006), so it is clearly possible that neuronal 5-HT could have this effect. Knockdown of DTPHu in fat body and hemocytes via the CgGAL4 driver, which should result in decreased circulating peripheral 5-HT levels, did not affect heart function (Fig. 9d). These results suggest that, at least at this stage, serotonergic innervation from the CNS modulates heart rate. It is also possible that circulating levels of 5-HT in the DTPHu knockdown is still sufficient for normal cardiac function.

The enhanced sensitivity of flies with diminished neuronal 5-HT to nonanol, but not to heptanol, suggests that serotonin is also required for modulation of specific chemosensory circuits in Drosophila. These studies are consistent with the observation that, in Drosophila, increasing the number of synapses enhances olfactory perception (Acebes & Ferrus 2001). Serotonergic modulation of chemosensory circuits has also been shown by the tph-1 mutation in nematodes (Chao et al. 2005); animals with decreased 5-HT signaling show decreased avoidance to the odorant octanol.

Adult flies with diminished neuronal 5-HT also display significantly reduced locomotor behavior (Fig. 6a), similar to what has been observed for the tph-1 mutation in C. elegans (Hardaker et al. 2001). It is not known whether these changes reflect acute responses to decreased neuronal 5-HT levels, or whether developmental events that require 5-HT for normal neuronal circuitry have been compromised. However, where acutely reduced 5-HT levels have been shown to increase both locomotion and appetitive behaviors such as feeding in mammals, it decreases these activities in Drosophila (Figs 5 and 6). Serotonergic varicosities are selectively decreased in second larval instar Drosophila larvae by addition of exogenous serotonin (Sykes & Condron 2005), and serotonin inhibits the motile activity of growth cones in the CNS of snails (Haydon et al. 1984). A lack of neuronal 5-HT would presumably disrupt neuronal circuitry by disrupting the normal check process during synaptogenesis.

DTPHu serves critical functions in the fat body, consistent with its role as a PAH

The breakdown of phenylalanine is an important step in glucose homeostasis; phenylalanine hydroxylation of phenylalanine to tyrosine serves as the major pathway for phenylalanine degradation. Mammalian PAH is found largely in the liver, with limited expression also detected within the kidney, pancreas and brain (Kappock & Caradonna 1996; Lichter-Konecki et al. 1999). This is consistent with the strong expression we have observed for DTPHu in the fat body, which is required for the maintenance of nutritional status and appropriate larval growth (Hoshizaki 2005). That DTPHu functions as a PAH is shown by the knockdown studies targeted to fat body because knockdown animals display the expected behavioral aberration (Kappock & Caradonna 1996), Of the three aromatic amino acid hydroxylases in Drosophila (DTH, DTPHu and DTRHn), only DTPHu is expressed in fat body. Therefore, the ability of DTPHu to act as a functional TPH in vivo may simply reflect that over the course of evolution, mammalian phenylalanine and peripheral tryptophan hydroxylation activities are no longer directed by a single enzyme, consistent with recent studies that have identified the presence of four aromatic amino acid hydroxylase family members in mammals, whereas Drosophila has only three.


The experiments described above show that compartmentalization of 5-HT synthesis is evolutionarily conserved and may reflect the complex requirements for appropriate regulation of 5-HT as a trophic factor in both neuronal and non-neuronal tissues. Knockdown (Fig. 8) or knockout (Colas et al. 1999a) of DTPHu results in diminished embryonic survival or embryonic lethality, respectively, whereas knockout of DTRHn results in enhanced embryonic survival (Fig. 7a). In the DTRHn mutant, a significant decrease in the number of flies eclosing is observed, attributable to increased larval lethality (Fig. 7b), but knockdown of DTPHu in fat body has no effect at this stage (Fig. 8d). Knockdown of DTPHu in this tissue also results in increased motor activity (Fig. 9b,c), but knockout of DTRHn has no effect on this behavior in larvae (Fig. 5b), and decreases locomotor activity in adults (Fig. 6a). Continued knockdown of DTPHu in fat body increases feeding behavior (Fig. 9c), whereas feeding is significantly depressed in DTRHn-null mutants (Fig. 5a). Clearly, the effects of decreases in DTRHn and DTPHu expression in their target tissues are nonoverlapping, evidence that the two enzymes play different functional roles in 5-HT synthesis. In mammals, knockout of tph1 does not result in overexpression of tph2 to compensate for loss of 5-HT synthesis in the relevant tissues (Cote et al. 2003), suggesting that the two genes are clearly under separate transcriptional and translational controls. Similarly, knockout of DTRHn does not result in synthesis of 5-HT by DTPHu in the CNS (although there is reuptake of peripheral 5-HT synthesized by DTPHu), and DTRHn does not synthesize 5-HT for early developmental events in the absence of DTPHu. The genetic tractability and evolutionary conservation of Drosophila melanogaster thus provide a unique model in which to elucidate the distinct mechanisms governing precise control of 5-HT synthesis.


This work was supported in part by NSF grant IOB-0616062 to W. S. N. S. F. G. and S. E. are supported by grants from the Biotechnology and Biological Sciences Research Council (BBSRC) and the Wellcome Trust.