Two Functional but Noncomplementing Drosophila Tyrosine Decarboxylase Genes

The trace biogenic amine tyramine is present in the nervous systems of animals ranging in complexity from nematodes to mammals. Tyramine is synthesized from tyrosine by the enzyme tyrosine decarboxylase (TDC), a member of the aromatic amino acid family, but this enzyme has not been identified in Drosophila or in higher animals. To further clarify the roles of tyramine and its metabolite octopamine, we have cloned two TDC genes from Drosophila melanogaster, dTdc1 and dTdc2. Although both gene products have TDC activity in vivo, dTdc1 is expressed nonneurally, whereas dTdc2 is expressed neurally. Flies with a mutation in dTdc2 lack neural tyramine and octopamine and are female sterile due to egg retention. Although other Drosophila mutants that lack octopamine retain eggs completely within the ovaries, dTdc2 mutants release eggs into the oviducts but are unable to deposit them. This specific sterility phenotype can be partially rescued by driving the expression of dTdc2 in a dTdc2-specific pattern, whereas driving the expression of dTdc1 in the same pattern results in a complete rescue. The disparity in rescue efficiencies between the ectopically expressed Tdc genes may reflect the differential activities of these gene products. The egg retention phenotype of the dTdc2 mutant and the phenotypes associated with ectopic dTdc expression contribute to a model in which octopamine and tyramine have distinct and separable neural activities.

mals. Despite their wide distribution, the biological roles of trace amines are poorly understood. In vertebrates, trace amines are present in low levels and can displace classical biogenic amines from their stores and stimulate outward neurotransmitter efflux from biogenic amine transporters (1). Although there are currently no data to suggest that there are dedicated synapses for trace amines in the brain, the recent discovery of G-protein-coupled receptors that are selectively activated by trace amines (2,3), along with the presence of trace amine binding sites in the brain (reviewed in Ref. 4), indicates that they may function independently of classical neurotransmitters. A large body of clinical literature suggests a role for tyramine in the etiology of several psychiatric disorders (reviewed in Ref. 4), and the expression pattern of TA 1 , a trace amine receptor potently activated by tyramine (2,3), supports such a role.
In invertebrates, little is known about the functional role of tyramine (reviewed in Ref. 5), although distinct tyraminergic patterns of expression are thought to exist in locusts (6) and C. elegans, 1 and receptors with preferential response to tyramine have been identified in many insects (reviewed in Ref. 5). Additionally, tyramine affects chloride permeability in the Drosophila Malpighian tubule (7), relaxation of muscle tone in the locust oviduct (8), and behavioral responses to cocaine in Drosophila (9).
Octopamine in invertebrates is often considered to be the functional homolog of vertebrate norepinephrine and is involved in a diverse range of physiological processes (reviewed in Ref. 10), including Drosophila female fertility. Flies that lack a functional octopamine receptor or cannot synthesize octopamine due to a null allele of tyramine ␤-hydroxylase (T␤h) 2 show a complete egg retention phenotype (11)(12)(13). The sterility of T␤h null females is due to a defect in ovulation, which can be rescued by octopamine feeding or by driving the ectopic expression of a T␤h transgene in several driver lines that show overlapping expression in the thoracic tip of the CNS (12,13).
The first step in octopamine biosynthesis is catalyzed by tyrosine decarboxylase (TDC) (Fig. 1), but this enzyme has not been identified in Drosophila or in higher animals. Based upon sequence similarity to plant TDCs, we report here the cloning of two Drosophila melanogaster TDC genes, dTdc1 and dTdc2.
Although both gene products have TDC activity in vivo, dTdc1 is expressed nonneurally, whereas dTdc2 is expressed neurally. Flies with a mutation in dTdc2 lack neural tyramine and octopamine and are female sterile due to egg retention; however, unlike T␤h mutants, the sterility of dTdc2 mutant females is not due to a defect in ovulation, since eggs are often found in the oviducts. Transgenic rescue of this phenotype can be accomplished by expression of either dTdc1 or dTdc2 with a dTdc2-specific driver. The specific egg retention phenotype of the dTdc2 mutant and the phenotypes associated with ectopic dTdc expression contribute to a model in which neurally derived octopamine and tyramine have distinct functions.

Fly Stocks and Growth Conditions
Drosophila lines were maintained on standard cornmeal-agar medium and kept on a 12 h/12 h light/dark cycle at 25°C. The UAS-CD8-GFP (BL-5137), UAS-GFP.nls (BL-4775), and the second chromosome deficiency kit were obtained from the Bloomington Stock Center. The cn bw and RO54, cn bw/CyO lines were a gift from Trudy Schupbach (14).

High Performance Liquid Chromatography (HPLC) Conditions and Sample Preparation
The HPLC system consisted of a Jasco model PU-2080Plus isocratic pump (Jasco Inc., Easton, MD), a Jasco model AS-950 Intelligent Autosampler with a 200-l sample loop, and an Antec Leyden INTRO electrochemical detector (Antec Leyden, The Netherlands). The VT-03 electrochemical flow cell contained a 2-mm glassy carbon working electrode mounted with a 50-m spacer and a salt bridge Ag/AgCl reference electrode (Antec Leyden). The cell potential was set to ϩ800 mV, and the range was 0.5 nA/V inside a Faraday-shielded compartment set to 27.5°C. Data from the detector were relayed to a computerized data acquisition system and analyzed using Jasco ChromPass chromatography software.
Reverse phase chromatography was performed with a Hypersil 100 C 18 column, 100 ϫ 4.6-mm inner diameter, with a 5-m particle size and 100-Å pore size (Thermo Electron Corp.). A Brownlee Validated C 18 Newguard column, 15 ϫ 3.2 mm (PerkinElmer Life Sciences), was placed between the injector and the analytical column. The mobile phase was delivered at a 1.0 ml/min flow rate. The mobile phase was carefully designed to maximize the separation and detection sensitivity of the biogenic amines octopamine, dopamine, tyramine, and serotonin. Because there is at least one peak that co-migrates with octopamine in all flies with pigmented eyes (wϩ), conditions were first tested in wild-type flies containing the wϩ gene and then in mutant strains in order to achieve optimal separation. The co-migrating peak could be separated from octopamine using the following mobile phase: 50 mM citrate/acetate, pH 4.5, 20% acetonitrile, and 10 mM decanesulfonic acid.
For brain samples, 10 adult female Drosophila brains were dissected directly into 100 l of ice-cold mobile phase and quickly homogenized with a Teflon pestle. Homogenates were run through a 0.22-m spin filter (Millipore Corp., Bedford, MA) and kept at Ϫ80°C until HPLC analysis. The typical injection volume was 10 l. Standard mixes of octopamine, tyramine, dopamine, and serotonin (Sigma) were injected at a concentration of 10 ng/ml, and a calibration curve was generated based on injections containing 1, 5, 10, 50, and 100 pg.

DNA Constructs and Transgenic Lines
GAL4 Constructs-To make the GAL4 constructs, a 3.3-kb fragment from pGaTB (15) containing the coding section of the yeast GAL4 gene and an hsp3Ј poly(A) segment was inserted into a genomic segment of DNA, which included either the dTdc1 or dTdc2 gene. In both cases, the GAL4 was inserted immediately before the coding start, and the entire construct (genomic segments interrupted by GAL4) was inserted into the polylinker of pCaSpeR4 (16). Genomic DNA was amplified with the Expand Long Template PCR system (Roche Applied Science), which contains a proofreading enzyme to ensure maximum accuracy. For dTdc1-GAL4, the genomic segment contained a region from Ϫ2938 to ϩ3940 bp relative to the coding start site. For dTdc2-GAL4, the genomic segment contained a region from Ϫ3459 to ϩ4530 bp. All constructs were thoroughly sequenced prior to injection into w 1118 embryos. Drosophila germ line transformation was performed by a standard procedure (17) using the plasmid p 25.7wc (18) as a source of transposase. Genetic Services, Inc. (Sudbury, MA) performed Drosophila germ line transformation for dTdc2-GAL4.
UAS Constructs-The coding fragments of dTdc1 and dTdc2 were amplified from cDNA using Expand Long Template PCR system (Roche Applied Science) and inserted into the polylinker of pUAST (15) to generate UAS-dTdc1 and UAS-dTdc2. All constructs were sequenced prior to injection into w 1118 embryos. Drosophila germ line transformation was performed in house by standard procedure (17) using the plasmid p 25.7wc (18) as a source of transposase.

Real Time PCR
Total RNA was isolated from adult female Drosophila tissues using the RNAqueous®-4PCR Kit (Ambion), and cDNA was synthesized using the SuperScript First-Strand Synthesis System for reverse transcription-PCR (Invitrogen) with oligo(dT) primers. Platinum® TaqDNA polymerase (Invitrogen) was used for real time cDNA amplification on the Smart Cycler System (Cepheid), and SYBR Green® (19) was used for quantification. PCR amplification conditions were as follows: 95°C for 2 min and 40 cycles of 95°C, 10 s; 60°C, 15 s; and 68°C, 30 s. The relative expression of each gene was calculated by the Smart Cycler software version 1.2c (Cepheid) using second derivative analysis of the cycle threshold (Ct). Analysis of the melt curve by the negative first derivative was used for product identification and to ensure product purity. Only those reactions that contained single products free of primer-dimers were used in the analysis. Primers included dTdc1 (5Ј-TTGAGGTTCGCAACGATGTTC-3Ј and 5Ј-AAGCACTTTATCTGGGT-CCAAGC-3Ј), dTdc2 (5Ј-ACGCATTGGCAGCATCCTC-3Ј and 5Ј-TGG-CAGCAAGCATCGTGAC-3Ј), and rp49 (5Ј-AAGATCGTGAAGAAGCG-CACCAA-3Ј and 5Ј-CTGTTGTCGATACCCTTGGGCTT-3Ј).

Egg Laying Assays
One 2-4-day-old female and one male were placed in individual vials, and the number of eggs was counted at 24-h intervals for 6 -7 days. 15 females per genotype were analyzed.

Drug Feeding Assays
3-5 females of the genotype Tdc2 RO54 /Tdc2 RO54 or Tdc2 RO54 / Df(2R)42 were collected within 1 day of eclosion and placed on small apple-agar Petri plates with an equal number of males. Yeast paste containing either no drug, tyrosine, tryamine, or octopamine (12 or 25 mg/ml concentrations for each drug) was added to each plate. Plates were examined each day for eggs, and the total number of eggs laid was determined after approximately 1 week. At least 13 females were tested in a minimum of three replicate experiments. Drugs were also fed to cn bw parental females and did not have an adverse effect on egg laying (data not shown).

Staining and Analysis of Tissue
Ovaries and associated oviducts were dissected and fixed in 4% paraformaldehyde for 10 min and stained with 4Ј,6-diamidino-2-phenylindole, as described in Ref. 20, and rhodamine-conjugated phalloidin (according to the protocol outlined by Molecular Probes, Inc. (Eugene, OR)). Stained reproductive tissues were visualized with a Bio-Rad Radiance 2100 confocal microscope. Confocal images were stacked and manipulated with Image J software. Color images depict GFP as green and TRITC as magenta (overlap white) to make the images color blindaccessible. Adult brain, nerve cord, and abdominal tissues were dissected and fixed in 4% paraformaldehyde for 10 min, and GFP was visualized with a Zeiss Axioskop microscope with a CARV spinning disk confocal imaging unit. Images were stacked and manipulated with OpenLab software. For oocyte distribution in mutant and control females, reproductive tracts including ovaries were dissected in PEM buffer (0.1 M Pipes (Sigma), 2 mM EGTA, 1 mM magnesium sulfate), fixed in 4% paraformaldehyde, and then mounted on slides for viewing in differential interference contrast optics on a Vanox 2 Olympus research microscope.

RESULTS
Two Candidate Tyrosine Decarboxylase Genes in the Drosophila Genome-Based on sequence similarity to plant TDCs, we identified two potential TDC genes in the D. melanogaster genome, CG30445 and CG30446. Due to the evidence in this report, we name these genes dTdc1 and dTdc2, respectively. The genes are adjacent, separated by 4.6 kb, and directly repeated on chromosome 2R at cytological position 42C3 (Fig. 2b). Alignment of the two coding sequences indicates a 52% identity at the nucleotide level and a 47% identity at the amino acid level. Both proteins contain the pyridoxal-dependent decarboxylase conserved domain that is common to all group II decarboxylase proteins. Two predicted Anopheles gambiae proteins have the strongest similarity to the predicted Drosophila proteins: predicted A. gambiae protein EAA03914.2 and dTDC2 are 76% identical, whereas predicted A. gambiae protein EAA03915.1 and dTDC1 are 36% similar. Further analysis of the A. gambiae genome revealed that the two genes encoding these predicted proteins are also adjacent and directly repeated, separated by 10.2 kb on chromosome 2L. Given their sequence similarity and clustering in the genome, it is likely that these genes are the mosquito homologs of dTdc1 and -2.
Bootstrap analyses of all established TDC proteins and of other decarboxylases with strong similarity to the candidate Drosophila TDCs verify that the A. gambiae EAA003915.1 and EAA003914.2 are likely homologs of dTDC1 and -2 (Supplemental Fig. 1). The only other confirmed TDC in a metazoan animal, C. elegans TDC-1, 1 is also closely related to the Drosophila TDCs, as is an unidentified gene product in Caenorhabditis briggsae. The most closely related vertebrate proteins are the human, mouse, guinea pig, and rat AADCs, but interestingly, no additional metazoan TDC-like genes were identified in our analyses.
To determine whether the TDCs contain any identifying amino acids, we aligned this same group of decarboxylases and searched for conserved residues. We were able to identify five amino acids that are conserved among the known and suspected TDCs, but generally not among other closely related decarboxylases (Supplemental Fig. 2). Only leucine 328 is absolutely reliable as an identifier of TDCs. Lysine 386 appears to be consistent among plant and animal TDCs, but it is not conserved among the bacterial TDCs, whereas the three other conserved residues predict very closely related AADCs as well. Four out of five of the conserved amino acids are contained within the pyridoxal-dependent decarboxylase conserved domain, but none is within the region known to contain the highly conserved pyridoxal phosphate attachment site at residues 298 -319. The final residue is located near the C-terminal end of the protein in a region where similarity between the decarboxylases declines. After identifying these residues, we reanalyzed the available vertebrate genomes to determine whether any TDCs could be identified, but none was found.
dTdc2 Is Expressed throughout the CNS and Provides Extensive Innervation to the Female Reproductive Tract, whereas dTdc1 Is Expressed Nonneurally-In Drosophila, as in other animals, each decarboxylase is typically encoded by one gene, although distinct promoters and/or alternative splicing may direct expression of neural and nonneural products, as is the case for the Drosophila DOPA decarboxylase and mammalian AADCs (21)(22)(23). Since there appear to be two TDCs in the Drosophila genome, we examined their expression patterns to determine whether they had unique or overlapping distributions. Using quantitative real time PCR, we analyzed the expression of dTdc1 and -2 in various Drosophila tissues and found that the two genes show a largely nonoverlapping pattern of expression (Fig. 3). dTdc1 is most abundant in the abdomen and in whole flies, but dTdc2 RNA is ϳ100-fold enriched in the brain relative to dTdc1. The indicated deficiencies were crossed to RO54 females and scored for their ability to complement the RO54 sterility phenotype. Complementing (boldface type) and noncomplementing deficiencies (lightface type) are indicated, and an asterisk denotes the position of dTdc2. Df(2R)cn87e is reported to cover 42B4-C1 through 43F1-44A1 on the cytological map, but recent data indicate that it does not disrupt Adf1 (54), which is located in 42C3 within 20 kbp of the dTdc genes. B, the RO54 candidate region at higher resolution, assembled from the Flybase genomic map. Predicted genes and mRNA sequences are shown above and below the line representing the genetic region. Quantitative real time PCR demonstrated no changes in mRNA abundance of any transcripts in this region in homozygous RO54 flies (data not shown). C, dTdc2 gene structure and the mutation in RO54. dTdc2 encodes a decarboxylase-like gene with a pyridoxal-dependent conserved domain. DNA sequence analysis determined that a single point mutation in RO54 results in a glutamic acid to lysine substitution within this domain; hereafter, this mutation will be designated Tdc2 RO54 . The gene structure prediction is from Flybase release 3.1.
To further examine the expression patterns of dTdc1 and dTdc2, we took advantage of the binary GAL4-UAS system (15). We developed GAL4 drivers for each gene and examined their in vivo expression patterns by using two different UAS-GFP reporter lines: UAS-CD8-GFP, a marker reported to be highly concentrated in neural processes including axons and dendrites (24), and UAS-GFP.nls, which contains a nuclear localization signal. Previous studies have detected octopamine and TBH in distinct clusters of cells throughout the central brain and along the ventral midline of the adult thoracic nerve cord (12,25), and recent studies indicate that TBH-expressing neurons send projections through the abdominal ganglion to the female reproductive tract, where they innervate the ovaries and oviduct (12). Based on these results and our real time PCR evidence, we predicted that dTdc2 would be expressed in a similar pattern as that seen for TBH and octopamine.
As expected, the dTdc2-GAL4 driver shows expression in discrete clusters of neurons throughout the adult brain and nerve cord and provides extensive innervation to the female reproductive tract (Fig. 4), but we were not able to detect dTdc2-GAL4 expression in any nonneural peripheral tissues. The dTdc2-GAL4 pattern of expression in the brain and nerve cord is strikingly similar to that seen with TBH immunoreactivity (25), although some of the TBH neurons reported to be located in the dorsal protocerebrum could not be detected. Furthermore, several clusters of cells not known to contain TBH in the central brain are detected with dTdc2-GAL4 (Fig.  4A, arrows). These cells could either be tyramine-specific neurons or an artifact of the GAL4 driver. As described for TBH (12), dTdc2-GAL4 neurons are located along the ventral midline of the abdominal segment of the thoracic CNS and send axons into the abdominal nerve, where they project over the posterior region of the ovaries and provide heavy innervation to the calyx and to the lateral and common oviducts (Fig. 4, B-D). Interestingly, the Drosophila octopamine receptor OAMB has been detected in the epithelial cell layer of the oviducts (11), but labeling of the epithelium in addition to the musculature shows that the dTdc2 -GAL4 projections are in the same focal plane as the musculature and ϳ10m above the epithelial layer (Fig. 4D). These results provide evidence that dTdc2containing neurons innervate the lateral and common oviducts and offer further support to the hypothesis that neural tyramine and octopamine are involved in the control of egg laying.
In contrast, dTdc1-GAL4 expression was absent from the brain and detected in only two small clusters of neurons in the thoracic nerve cord. As expected, dTdc1 is highly expressed throughout the body, including the gut musculature, rectal papillae, and Malpighian tubules (Fig. 5), but was completely absent from the female reproductive tract (data not shown). Although the presence of tyramine and octopamine in the Drosophila CNS is well established (25), tyramine was only recently shown to modulate transepithelial chloride conductance in the Drosophila Malpighian tubule (7). Since we detect dTdc1 expression and tyramine, but not octopamine, in these nonneural tissues (data not shown), we predict that dTdc1 is responsible for this nonneural tyramine production.
A dTdc2 Mutation Causes Female Sterility-Flies containing a null mutation in the Drosophila T␤h gene (T␤h nM18 ) completely lack octopamine and are female sterile due to egg retention (13). Eggs are retained in the ovaries of T␤h nM18 females as a result of defective ovulation, which is defined as the passage of eggs from any one of the posterior ovarioles to the lateral oviducts (12,13). Since T␤h nM18 females do not ovulate, it is not known whether they have a coexisting deficit in oviposition, the extrusion of eggs from the uterus. Regardless, one would expect that flies containing a mutation in a Tdc gene that is expressed in cells involved in egg laying would also be female sterile (see Fig. 1). An ethyl methane sulfonate mutagenesis screen for recessive female sterile mutations on the second chromosome revealed one mutant, named RO54 (14), which had a severe egg retention phenotype similar to that seen in T␤h nM18 mutants and whose inability to lay eggs could be partially rescued by octopamine and tyramine feeding (Table I). By the following molecular and genetic analysis, we identified the RO54 mutation as a lesion within the dTdc2 gene and have renamed the mutant allele as Tdc2 RO54 . In contrast to T␤h nM18 females, mature eggs or their collapsed remnants were found in the calyx, lateral, or common oviducts in 83% of Tdc2 RO54 females, indicating that oocytes are ovulated but cannot be oviposited; by contrast, only one of 18 cn bw control females had an egg in these structures (Supplemental Table I).
We mapped the mutation to the general region of dTdc1 and -2 using a series of second chromosome deletions ( Fig. 2A), scoring for their ability to complement the sterility phenotype. The phenotype was not complemented by deficiency Df(2R)42, which deletes the cytological region 42C2-42D2-3, but was complemented by the adjacent deficiencies Df(2R)cn87e and Df(2R)nap1 and the inversion In(2R)pk87s. Using this information, we were able to localize the mutation within the Df(2R)42 deficiency but to the left of Df(2R)cn87e, which narrows the mutation to the small region of 42C2-42C3 containing dTdc1 and -2 (Fig. 2B).
Since it was possible that the Tdc2 RO54 mutation altered mRNA abundance in the affected gene, we used quantitative real time PCR to screen genes in the dTdc1 and -2 region. We did not find altered RNA levels in any of the genes in this immediate region between homozygous Tdc2 RO54 and wildtype flies (data not shown). However, sequencing the genomic DNA of dTdc1 and -2 revealed a single point mutation in the third exon of dTdc2 that results in a glutamic acid to lysine substitution at amino acid 51, within the highly conserved pyridoxal-dependent decarboxylase conserved domain (Fig.  2C). An alignment of the predicted dTDC2 amino acid sequence with the 10 most similar decarboxylases shows that each of the FIG. 3. Differential expression of dTdc1 and dTdc2 RNA. RNA from brain, abdomen, or whole fly was used for quantitative real time PCR analysis. dTdc2 RNA is selectively enriched in the brain although dTdc1 is most abundant in abdomen and whole flies. Cycle threshold (Ct) was determined using second derivative analysis, and the difference in Ct between dTdc1 or dTdc2 versus rp49 was used to determine relative expression levels in each tissue. The error bars represent S.E. between at least four independent samples.
aligned proteins contains either a glutamic acid or aspartic acid at the location of the Tdc2 RO54 mutation (Fig. 6A).
Tdc2 RO54 Mutants Lack Neural Tyramine and Octopamine-To assess levels of tyramine and octopamine in Tdc2 RO54 brains, ex-tracts from Tdc2 RO54 females were compared with their cn bw parental line using HPLC with electrochemical detection (Fig. 6B). We were unable to detect any octopamine or tyramine in Tdc2 RO54 brains, whereas normal levels are present in the cn bw controls. We did not observe a significant difference in the amount of octopamine or tyramine in Tdc2 RO54 heterozygote brains versus control lines. Repeated analyses also failed to show significant differences in levels of dopamine and serotonin between Tdc2 RO54 and cn bw brains.
Expression of either dTdc1 or -2 in a dTdc2-specific Pattern Rescues the Tdc2 RO54 Sterility Phenotype-If dTdc1 and 2 both encode TDC enzymes, then the expression of either gene in the brains of Tdc2 RO54 females should rescue the sterility phenotype. To test this hypothesis, we used the dTdc2-GAL4 driver to specifically express either UAS-dTdc1 or UAS-dTdc2 in Tdc2 RO54 flies. Somewhat paradoxically, expression of UAS-dTdc1 provided a complete rescue of the egg retention phenotype, whereas expression of UAS-dTdc2 gave only a partial rescue (Fig. 7A). Intriguingly, however, HPLC analyses of brain extracts showed that mutant females expressing UAS-dTdc2 had significantly higher levels of both tyramine and octopamine than females expressing UAS-dTdc1 (Fig. 7B). In fact, we could not detect tyramine in the brains of dTdc1expressing females, but dTdc2-expressing females had tyramine levels well above those seen in wild-type females (see Fig.  6B). Although these results seem somewhat contradictory, the varying degrees of rescue achieved by the ectopic expression of dTdc1 or dTdc2 indicate that both genes express TDC activity in vivo.

TABLE I
Rescue of Tdc2 RO54 mutant female fecundity Between three and five females per vial were mated and then fed the indicated drugs at 12 mg/ml or 25 mg/ml concentrations in at least three replicate experiments. The total number of eggs laid is given. Octopamine rescues egg laying with a significant effect of concentration seen for both genotypes (two-factor analysis of variance, p ϭ 0.0006). There is a small but significant effect of tyramine concentration on rescue in the Tdc2 RO54 /Tdc2 RO54 genotype (linear regression, r 2 ϭ 0.21, df ϭ 20, p ϭ 0.036). Our results show that Drosophila contain two functional and highly related TDC genes, dTdc1 and dTdc2. dTdc1 is primarily expressed in nonneural abdominal organs, whereas dTdc2 is expressed in the CNS and innervates the female reproductive tract. dTdc2 RO54 , a point mutation in dTdc2, results in a loss of neural tyramine and octopamine and leads to female sterility. The ectopic expression of a dTdc1 transgene results in a complete rescue of the sterility phenotype, whereas the expression of a dTdc2 transgene leads to high levels of brain tyramine and octopamine but only partial rescue of the sterility phenotype. The difference in rescue efficiency between the genes shows that the level of expression of a TDC enzyme is as critical as the pattern of expression.
Comparative Enzymology of Aromatic Amino Acid Family Decarboxylases-Tryptophan hydroxylase and tyrosine hydroxylase catalyze the first and rate-limiting steps in the serotonin and dopamine synthesis pathways (reviewed in Ref. 26). In the CNS, these enzymes are highly regulated, which prevents depletion of the essential amino acids tyrosine or tryptophan, respectively (26). AADCs are responsible for the final enzymatic step in the pathways and convert 5-hydroxytryptophan to serotonin and L-DOPA to dopamine. AADCs are not ratelimiting, and although they are thought to be relatively nonselective, they are subject to short term regulatory mechanisms (27). TDCs are closely related to the AADC family of enzymes; however, unlike the AADCs, they act directly on tyrosine and must be highly regulated in order to maintain sufficient concentrations of neural tyrosine.
This requirement for acute regulation of TDCs complicates attempts to rescue the dTdc2 RO54 mutant, which is more readily rescued by ectopic expression of dTdc1 than by dTdc2. The simplest explanation is that the GAL4-UAS rescue paradigm more effectively controls the cellular pattern of expression than the transcriptional level of expression within partic-ular neurons. The level of expression depends upon several factors, including the insertion site of the transgene and the number of UAS sequences within the target transgene. To control for insertion site variation, we used two separate insertions for each of the UAS lines; in each case, the two insertions give comparable results (data not shown). However, because the UAS target vectors contain five tandem repeats of the GAL4 binding sequences to generate high levels of activation (15), we are expressing dTdc1 and -2 at potentially much higher levels than they would normally be expressed. In this situation, an unregulated Tdc could lead to depletion of the essential amino acid tyrosine, which could have adverse functional consequences for the cell. We suspect that this may be the case for the ectopic expression of dTdc2, which results in abnormally high levels of brain tyramine and octopamine but only partial functional rescue. Alternately, increased levels of tyramine in these neurons could have an inhibitory effect on egg laying, as seen in C. elegans. 1 In fact, the abnormally high levels of neural tyramine in T␤h nM18 females (13) could contribute to their complete lack of ovulation, as discussed in greater detail below.
In contrast, the dTdc2-GAL4 driven expression of dTdc1 results in an undetectable amount of tyramine and low levels of octopamine but fully rescues the sterility phenotype. We therefore predict that dTdc1 is a much less active enzyme. In any case, the varying degrees of rescue achieved by the ectopic expression of dTdc1 or dTdc2 indicate that both genes express TDC activity in vivo. Additionally, these results show that precise control of this rate-limiting enzyme activity is critical for normal egg laying and may be critical for proper cellular function. Consistent with this, broader ectopic expression of dTdc2 but not dTdc1 can result in embryonic lethality (data not shown).
Many enzymes that are expressed in both neural and nonneural patterns are subject to differential regulation, so the FIG. 6. Tdc2 RO54 mutants contain a single point mutation in the highly conserved pyridoxal-dependent decarboxylase domain of dTDC2. A, a portion of the amino acid sequence of dTDC2 is aligned with the 10 most similar decarboxylase proteins as predicted by a BLAST similarity search (ClustalW alignment; MacVector). The asterisk denotes the amino acid change in Tdc2 that results in a glutamic acid to lysine substitution. The pyridoxal-dependent decarboxylase domain (CD search; NCBI) extends from amino acid 35 to 414, and the pyridoxal-phosphate attachment site extends from amino acid 298 to 319 with the co-factor activation site at Lys-305. The numbers above the sequence correspond to the amino acid sequence of dTDC2. B, Tdc2 RO54 mutants have no detectable brain tyramine or octopamine. HPLC analyses of brain homogenates from the background line cn bw (dashed line) and Tdc2 RO54 (solid line) females show that Tdc2 RO54 females completely lack octopamine; the small base-line increase in the region of tyramine in Tdc2 RO54 does not specifically co-migrate with tyramine standards and is most likely the result of another interfering peak. Repeated analyses indicate that levels of dopamine and serotonin are not significantly different between cn bw and Tdc2 RO54 .
idea that dTDC1 and -2 are differentially regulated is not unanticipated. Although most of these enzymes are encoded by one gene, a recently discovered exception is mammalian tryptophan hydroxylase; TPH1 is expressed in the periphery, whereas TPH2 is expressed in the brain (28). In contrast, mammalian AADC and Drosophila DOPA decarboxylase and tyrosine hydroxylase use unique promoters and/or alternative splicing to direct neural and nonneural expression of distinct gene products from the same gene (29 -31). Taken together, our results, along with existing evidence, suggest that enzymes responsible for both central and peripheral biogenic amine synthesis are subject to distinct regulatory mechanisms.
Distinct Phenotypes of dTdc2 RO54 and T␤h Female Sterility: Distinct Roles for Neurally Derived Tyramine and Octopamine-Octopamine is a well established neurotransmitter, neuromodulator, and neurohormone in invertebrates (reviewed in Ref. 10). In the fly, octopamine has been implicated in complex behavioral processes and in physiological processes such as ovulation and egg laying (11,12). Our study supports previous work demonstrating the importance of neural octopamine in egg laying but suggests that tyramine may have an important function in this process as well. Interestingly, dTdc2 RO54 flies retain a functional dTdc1 gene that expresses at high levels in nonneural tissues. Based on our inability to detect octopamine by HPLC in these nonneural tissues (data not shown) and the almost exclusively neural expression of TBH (13), the end product of this nonneural pathway is prob-ably tyramine. Because this nonneural tyramine does not enter the CNS and/or ovaries, where it would be converted into octopamine, we predict that barriers must exist to prevent tyramine diffusion.
A previous study concluded that a small population of TBHpositive ventral ganglion neurons was necessary for ovulation (12). This conclusion was based on driving the expression of a T␤h transgene in enhancer trap lines known to have various patterns of CNS expression but without considering potential nonneural expression. If such nonneural expression exists in these lines, the expression of a T␤h transgene could lead to the synthesis of octopamine outside of the CNS. Selective diffusion of octopamine into the CNS or ovaries could not only explain the achieved rescue but could also explain the enhanced ability of octopamine versus tyramine feeding to suppress the dTdc2 RO54 sterility phenotype (Table I).
In other insects, octopamine relaxes ovarian muscles to allow the release of mature eggs into the oviducts (32,33). In the locust Locusta migratoria, this occurs via activation of octopamine-2B receptors, which leads to a rise in cAMP (34). In the stable fly Stomoxys calcitrans, isolated oviducts respond to octopamine with a decrease in spontaneous contractions (35) and an increase in myogenic and neurogenic contractions of spermathecal muscles, which moved stored sperm (36). The coordination of oviduct muscle tone with spermathecal contractions is necessary for successful fertilization; relaxation of oviduct contractions, presumably by octopamine, allows eggs to progress toward the uterus, where fertilization can occur (37,38). Our observation that dTdc2 neurons in the ventral ganglion project directly to the posterior ovariole, and oviduct musculature supports a similar role for octopamine in Drosophila. In the ovariole, relaxation of muscle tone by octopamine could allow the release of mature eggs into the lateral oviducts, whereas relaxation of muscle tone in the lateral and common oviducts could allow the passage of eggs toward the uterus.
In many insects, octopamine also plays a role in the regulation of juvenile hormone (JH) biosynthesis. JH, a sesquiterpenoid hormone, is synthesized and released from the corpora allata of adult insects and is tightly controlled at the levels of synthesis, release, and degradation (reviewed in Ref. 39). JH is involved in every major developmental transition in the insect and is essential for many aspects of reproductive function. The involvement of octopamine in JH metabolism has been clearly demonstrated in many insects; in L. migratoria and the honeybee Apis mellifera, octopamine stimulates JH biosynthesis, whereas in the cricket Gryllus bimaculatus and the cockroach Diploptera punctata, it inhibits JH production (reviewed in Ref. 39). In adult A. mellifera, very fine octopamine immunoreactive fibers with varicose terminals surround each of the gland cells in the corpora allata tissue (40). Furthermore, in the silkworm Bombyx mori and the flour beetle Tribolium freemani, there is in vitro evidence that secretion of octopamine increases the activity of JH esterase prior to the onset of pupation (41).
In Drosophila, a direct link between octopamine and JH biosynthesis is not as well defined, but indirect evidence exists for a functional relationship. T␤h nM18 females are reported to have a significantly increased rate of JH degradation compared with wild-type females (42,43), leading to speculation that JH metabolism is under the control of biogenic amines. Furthermore, under stressful conditions, levels of biogenic amines including octopamine are reported to increase, whereas JH degradation decreases (44 -48). Nutritional stress leads to even more severe consequences, including a delay in oocyte maturation, degradation of early vitellogenic egg chambers, accumulation of mature oocytes, and a 24-h oviposition arrest (49). Until JH titers are measured directly in these situations, links FIG. 7. A, driving expression of dTdc1 in a dTdc2-GAL4 pattern rescues the Tdc2 RO54 egg laying defect, whereas expression of dTdc2 in the same pattern partially rescues the phenotype. Tdc2 RO54 females cannot lay eggs, but there is no significant difference between the number of eggs laid by control females, Tdc2 RO54 /ϩ heterozygotes, and Tdc2 RO54 females expressing dTdc1 in a dTdc2-GAL4 pattern. The error bars represent S.E. *, significant difference between the indicated genotypes and control females by single factor analysis of variance (p Ͻ 1.85 ϫ 10 Ϫ7 for Tdc2 RO54 females expressing UAS-dTdc2 in a dTdc2-GAL4 pattern). B, HPLC analyses of brain homogenates from females expressing either dTdc1 (solid line) or dTdc2 (dashed line) in a dTdc2-GAL4 pattern. Females expressing dTdc1 have no detectable brain tyramine and low levels of octopamine, whereas females expressing dTdc2 have high levels of both amines. between biogenic amines, stress, and JH will remain speculative. We suggest that the loss of octopamine is the primary cause of the failure to oviposit in Tdc2 RO54 and T␤h nM18 mutants; however, it is possible that there are consequently higher JH titers in both strains that may exacerbate the phenotype.
Little is known about the functional role of tyramine, although a recent study shows that tyramine inhibits egg laying in C. elegans 1 and has specific effects on myogenic and neurogenic contractions in locust oviduct tissue (8). It is tempting to speculate that tyramine may also have an inhibitory effect on Drosophila egg laying. Tdc2 RO54 flies lack both octopamine and tyramine, whereas T␤h nM18 flies lack neural octopamine but contain ϳ10-fold elevated tyramine (13). The high levels of neural tyramine in T␤h nM18 could account for the fact that T␤h nM18 females are deficient in ovulation but Tdc2 RO54 females are not. In further support of an inhibitory role for tyramine, the ectopic expression of dTdc2 in Tdc2 RO54 females results in abnormally high levels of brain tyramine and octopamine but only partial functional rescue. This could be a result of tyrosine depletion, as discussed previously, but it is also possible that the high levels of tyramine could lead to a partial inhibition in egg laying due to decreased ovulation.
An additional Drosophila mutant inactive, which encodes a TRPV channel (50), has been reported to contain low levels of tyramine and octopamine and show aberrant responses to cocaine (9, 51), but we have not found altered tyramine or octopamine levels in inactive brains using the analytical methods reported in this paper (data not shown). Further investigation will be required to determine whether this change in biogenic amines occurs only under selected physiological conditions.
Source of Endogenous Tyramine in Vertebrates-Our data indicate that there are two Tdc genes in Drosophila and A. gambiae and one in C. elegans and C. briggsae, but we have not identified any potential Tdc candidate genes in the available vertebrate genomes. Based on this evidence, it seems unlikely that TDCs exist in higher vertebrates. Nevertheless, it is probable that tyramine is synthesized endogenously in the CNS due to the fact that very little tyramine crosses the blood-brain barrier from food sources (52). Mammalian AADC is capable of catalyzing the decarboxylation of a number of substrates in addition to L-DOPA and 5-hydroxytryptophan, although not nearly as efficiently; in addition, AADC inhibitors result in reduced tyramine production (53). These results suggest that the primary route of tyramine synthesis in the vertebrate brain is via decarboxylation of tyrosine by AADC. Other potential tyramine synthesis pathways have been proposed, but there are no recent studies to confirm the existence of these pathways.
Although the existence of trace amines in the brain is well known, physiological roles for these amines have been difficult to establish. The identification of genes involved in tyramine and octopamine biosynthesis in Drosophila will provide the means to develop new approaches for defining functional roles for these compounds.