Two Drosophila members of this superfamily have been isolated and characterized by cDNA cloning: a serotonin transporter (dSERT) (Corey et al., 1994; Demchyshyn et al., 1994) and a dopamine transporter (dDAT) (Porzgen et al., 2001). In addition, isolation of partial cDNA clones encoding a transporter with high-sequence homology to mammalian GABA transporters has been reported (Neckameyer and Cooper, 1998). Two other members of this superfamily were isolated by a forward genetic approach. The mutant pheno-types suggest that one of these functions primarily in the nervous system (rosA/ine: Burg et al., 1996; Soehnge et al., 1996), whereas the primary effect of the other occurs outside the nervous system (blot: Johnson et al., 1999). These are the only members of the neuro-transmitter transporter superfamilies for which mutants have been isolated in the genes encoding the transporters.
Isolation of Drosophila members of this superfamily of transporters is of interest, in part because, in humans, the transporters for serotonin, dopamine, and norepineph-rine are the sites of action of antidepressants and such psychostimulants as cocaine and amphetamines. Most of the reinforcing properties and abuse potential of psychostimulants are thought to arise from the blockade of the dopamine transporter (Ritz et al., 1987). Indeed, the two Drosophila members of this family isolated by cDNA cloning, discussed below, have been shown to be antidepressant- and cocaine-sensitive.
Drosophila Serotonin Transporter (dSERT). cDNA clones encoding a dSERT were isolated either by screening a Drosophila head cDNA library with a) a fragment of the human DA transporter cDNA (Demchyshyn et al., 1994) or b) a Drosophila cDNA fragment obtained by PCR amplification primed by degenerate oligonucleotides corresponding to conserved portions of the known Na+/Cl- transporters (Corey et al., 1994).
The predicted proteins reported by the two groups differed slightly in length: 581 (Corey et al., 1994) vs. 622 amino acids (Demchyshyn et al., 1994). However, except for the N-terminal 30-70 residues, the amino acid sequences were identical, which suggests that these represented products of the same gene. Hydropathy analysis showed that the proteins had 12 putative TMDs with a large extracellular loop between the third (TMD3) and the fourth (TMD4), which is consistent with the membrane topology proposed for all other known Na+/Cl-dependent transporters. The proteins showed highest overall amino acid homology with rat (Blakely et al., 1991; Hoffman et al., 1991) and human (Ra-mamoorthy et al., 1993; Lesch et al., 1993); 5-HT transporters; 52 and 53% identities, respectively, Corey et al., 1994; 51% identity, Demchyshyn et al., 1994) and slightly lower homology with human norepinephrine (Pacholczyk et al., 1991; 51% identities, Corey et al., 1994; 47% identities, Demchyshyn et al. 1994), human dopamine (Giros et al., 1992; Vandenbergh et al., 1992; Pristupa, 1994; 47% identities, Demchyshyn et al., 1994), and rat/bovine dopamine transporters (Shimada et al., 1991; Kilty et al., 1991; Usdin et al., 1991; Giros et al., 1991; 48% identities, Corey et al., 1994).
The chromosomal in situ hybridization showed that the dSERT gene was located at 60C of the right arm of the second chromosome (Demchyshyn, 1994). Northern blot analysis revealed a strongly labeled ~3.3 kb transcript from adult heads but no detectable signals from adult bodies or embryos (Corey et al., 1994). In situ hybridization to embryos and larvae showed the accumulation of dSERT mRNA in a restricted number of cells in the ventral ganglia arranged in a stereotypic pattern (Demchyshyn et al., 1994) consistent with the pattern of distribution of 5-HT-containing neurons (Valles and White, 1988), which strongly suggests that dSERT functions as a SERT.
Xenopus oocytes injected with dSERT mRNA showed fivefold higher transport of 5-HT than that of GABA, glutamate, histamine, dopamine, or norepinephrine (Corey et al., 1994). The transport of 5-HT occurred with high affinity and in a concentration-dependent and saturable manner in both injected oocytes (Corey et al., 1994) and transfected HeLa cells (Demchyshyn, 1994). The reported Km of 490 nM ± 35 nM for HeLa cells and 637 ± 100 nM for oocytes compared favorably with the Km values determined for the cloned human (463 nM) (Ramamoorthy et al., 1993) and rat SERTs (320-529 nM) (Blakely et al., 1991; Hoffman et al., 1991). dSERT-mediated transport of 5-HT displayed an absolute requirement for Na+ in both expression systems. However, unlike in mammalian SERTs, the requirement for Cl- was not absolute. Substitution of Cl- with nitrate, acetate, or gluconate inhibited 5-HT uptake by 85-91% in oocytes (Corey et al., 1994) and only by ~50% in HeLa cells (Demchyshyn et al., 1994).
The pharmacological profile of dSERT-mediated transport had some similarity, but was not identical to that of mammalian SERT-mediated transport (Corey et al., 1994; Demchyshyn et al., 1994). Corey et al. (1994) noted that, with respect to a certain group of antidepressants, dSERT had a pharmacological profile with closer similarity to the mammalian catecholamine transporters than to the mammalian 5-HT transporters. Thus, for example, fluoxetine and clomipramine are 6- to 71-fold lower in potency at dSERT than at the mammalian SERTs, whereas mazindol and nomifensine, strong inhibitors of the mammalian catecholamine transporters but not as effective inhibitors of the mammalian SERTs, were potent inhibitors of dSERT. Thus, the pharmacological profile of dSERT represents a mixture of those of the mammalian catecholamine and 5-HT transporters. dSERT is sensitive to cocaine, although the degree of sensitivity compared with human SERT reported by the two groups differs somewhat. Both groups agree that the specificity of dSERT for 5-HT is very high. For example, unlabeled 5-HT showed significant inhibition of [3H]5-HT uptake by dilution, but the other potential substrates tested— epinephrine, norepinephrine, dopamine, tyramine, octopamine, and histamine—had little or no effect. These findings taken together established that dSERT indeed is a 5-HT transporter.
Drosophila Dopamine Transporter (dDAT). Because all monoamine transporters cloned contain a characteristic aspartate residue in the first TMD, Porzgen et al. (2001) utilized the sequence information of the first TMD of the human dopamine transporter (hDAT) to identify in the BDGP database a gene encoding a putative monoamine transporter. Primers were designed from the sequence in the database to amplify a DNA fragment encompassing the entire coding region from a Drosophila cDNA pool.
The isolated cDNA encoded a protein of 631 amino acids with highest homology to the human norepinephrine, dopamine, and serotonin transporters and the C. elegans dopamine transporter (52, 49, 45, and 51% identity, respectively). Hydropathy analysis predicted 12 TMDs with intracellularly located N- and C-terminals and a large extracellular loop between TMD3 and TMD4 (EL2). Found in TMD2 and TMD5 of dDAT were leucine zipper motifs, observed in TMD2 of a number of neurotransmitter transporters (Amara and Kuhar, 1993), including dSERT (Demchyshyn et al., 1994).
The gene was localized to 53C7-14 on the right arm of the second chromosome. The developmental profile of mRNA expression, determined by semiquantitative RT-PCR (Porzgen et al., 2001), showed dDAT expression in late embryos and larval stages. Expression decreased during the pupal stage but increased again in adults, with the strongest expression in adult heads. In situ hybridization to whole-mount third instar larval CNS (Porzgen et al., 2001) showed that the dDAT expressing cells were distributed in a characteristic pattern consistent with the distribution pattern of dopamine-containing cells (Budnik and White, 1988).
The functional properties of dDAT were examined in a Madin-Darby canine kidney cell line stably expressing dDAT (Porzgen et al., 2001). The results showed that dopamine had the highest maximal uptake velocity (Vmax) and the highest apparent transport affinity of any substrate tested, and the rank order of Vmax for biogenic amines was dopamine > norepinephrine > tyramine > epinephrine. Dopamine uptake into dDAT-expressing cells showed an absolute dependence on extracellular Na+ but a more relaxed dependence on Cl-, as was the case of dSERT (Corey et al., 1994; Demchyshyn et al., 1994).
The pharmacological profile of dDAT, obtained with a series of high-affinity nonsubstrate inhibitors for the monoamine transporters, showed a pattern distinct from that of the hDAT but closer in similarity to profiles reported for mammalian norepinephrine transporters (NERT), reminiscent of the inhibitor pharmacological profile reported for dSERT (Corey et al., 1994). Thus, dDAT showed high affinity for the tricyclic antidepressants, desipramine, imipramine, amitriptyline, and nisoxetine, all of which are highly selective for inhibiting hNERT over hDAT. Thus, echoing the findings for dSERT, dDAT is a unique catecholamine transporter with the substrate specificity of human DAT but with inhibitor sensitivity most closely resembling mammalian NETs. dDAT was moderately sensitive to cocaine. It is one of two cocaine-sensitive targets identified in flies to date.
Electrophysiological properties of dDAT were also distinct from those of hDAT. Both dDAT and hDAT mediate transport-associated currents when expressed in Xenopus oocytes (Porzgen et al., 2001). The mammalian monoamine carriers characteristically exhibit a pronounced, substrate- and inhibitor-sensitive constitutive leak conductance (Mager et al., 1994; Galli et al., 1995; Sonders et al., 1997). dDAT seemed to lack this leak conductance.
GABA Transporter. Neckameyer and Cooper (1998) have reported on cDNA clones containing partial open reading frames that predict polypeptides with high sequence homology with mammalian GABA transporters over a stretch of 120 amino acids. The presumptive Drosophila GABA transporter gene was mapped to 102BC of the fourth chromosome. However, the full sequence of the transporter has not been obtained, and no other information has been reported on the transporter.
Bloated tubules (Blot). The Blot protein is one of the two Drosophila members of the Na+/Cl-dependent transporter superfamily that were identified by a forward genetic approach; that is, isolation and genetic analysis of mutants preceded the identification of the protein. The blot gene was identified from enhancer trap lines that showed expres-
sion in the developing Malpighian tubules (Johnson et al., 1999). Initial search for such lines identified two, l(3)1658 (Spradling et al., 1995) and A434 (Bellen et al., 1989), with very similar patterns of reporter gene expression. Subsequently, two stronger alle-les, blotM51 and blotM55, were generated by remobilization of the P-element insert in the A434 line to induce imprecise excision of the P-insert (Johnson et al., 1999). blot1658 was mapped to 74B on the left arm of the third chromosome (Spradling 1995; Johnson et al., 1999).
Johnson et al. (1999) were interested in the Malpighian tubules as a model system for analyzing the development of epithelial tissues. The Malpighian tubules are an insect equivalent of the renal system that excrete toxic waste and adjust and maintain ionic and osmotic balance (Wigglesworth, 1939). The Malpighian tubules originate from evagina-tion of two pairs of primordia at the border between midgut and hindgut of the embryo. In larvae, they consist of two pairs of slender tubes with cells containing a large number of fluorescent inclusion materials. The lumen of the tubule is encircled by the brush border, which is covered with microvilli. Actin filaments are concentrated in a submembranous region beneath the apical (luminal) membrane, and they extend into the microvilli. In blot mutant larvae, the Malpighian tubules appear bloated (hence the name "bloated tubules") and show a dramatic reduction in the fluorescent inclusion material. The luminal side (apical side) of the epithelial cells has extensive folds protruding into the lumen, although actin-containing microvilli are still present. Animals homozygous for blot1658 are late larval lethal (Spradling, 1995), whereas those homozygous for blotM51 or blotM55 are second-instar larval lethal.
The blot gene was cloned by obtaining a genomic fragment adjacent to the P-insert in the A434 enhancer trap line by plasmid rescue and by using this fragment to carry out a chromosomal walk (Johnson et al., 1999). The open reading frame was identified from a composite cDNA generated from two partial cDNA clones isolated in cDNA library screening. The open reading frame predicted a protein of 1035 amino acids having 12 putative transmembrane domains and significant homology with members of the Na+/Cl--dependent transporter superfamily. The highest homology (28% identity in a stretch of 436 amino acids) was found with the rat orphan transporter Rb21A, for which no substrate has been identified (Smith et al., 1995). The membrane topology of Blot also more closely resembled the orphan transporters than other members of this transporter super-family in that, although most members of this superfamily have only one large extracellular loop between TMD3 and TMD4, Blot has two others between TMD5 and TMD6 and TMD7 and TMD8, in addition to the one between TMD3 and TMD4.
Northern blot analysis detected two transcripts of 4.0 and 4.3 kb throughout the embryonic development, with the shorter one appearing to be primarily supplied maternally. In situ hybridization experiments showed that blot RNA was expressed primarily in epithelial tissues of ectodermal origin, in the nervous system of the embryo and larva, and also in the freshly laid egg and developing oocyte. Embryos lacking the maternally derived component of blot died in early stages of development. At the syncytial blastoderm stage, filamentous actin in the cortex was almost completely absent, the structure of the apical cortex was impaired severely, and cellularization failed completely. Thus, the feature common to both the early developmental phenotype of embryos lacking the maternal component of blot and the Malpighian phenotype of larvae appeared to be a dramatic alteration in the organization of the actin-containing cortical cytoskeleton. The substrate for the Blot transporter has not been identified, and it is not known how the observed pheno-types are produced by a loss or reduction in the Blot activity. This transporter appears to be the only one of this superfamily identified in Drosophila to date that functions primarily outside the nervous system.
RosA/Ine Transporter. The receptor oscillation A (rosA) gene was first identified as an electroretinogram (ERG)-defective mutant in which an oscillation is superimposed on the light response of photoreceptors in both ERG and intracellular recordings (Wilcox and Pak, 1977; Wu and Wong, 1977). The oscillation occurs at frequencies ~50 Hz, and the cell undergoes hyperpolarization after the light stimulus (Burg et al., 1996). The rosA gene was genetically mapped to 24F-25A of the chromosome 2L by deficiency mapping (Burg et al., 1996). Once physically localized to a small region, P1 phage genomic DNA clones that covered this region were identified by using database resources of the Drosophila community (Flybase, 1999). The molecular identification of the rosA gene was accomplished through Northern analysis of wild-type and rosA-mutant mRNA by using genomic DNA fragments as probe. Several of the genomic fragments from the P1 clones identified a reduction in the transcript level in some, but not all, rosA mutant alle-les. Northern analysis also indicated two transcripts detectable in adult tissue (a major transcript of 3.7 kb and a minor transcript of 5.7 kb), both of which were reduced in amount in some mutant alleles. A corresponding full-length cDNA was isolated from a Drosophila head cDNA library by using the same genomic fragment as probe. cDNA sequence analysis indicated that the transcript encoded a Na+/Cl-dependent NTT homologue, having 36-41% amino acid identity with other transporters of this class. However, there was not sufficient similarity with any one class of transporters to predict the substrate for this transporter. The predicted RosA transporter is unique in that it contains a much larger amino terminus (~300 amino acids) than any other Na+/Cl-dependent NTT reported to date. However, other structural similarities, such as 12 predicted transmembrane domains and a large extracellular loop #2 between TMD3 and TMD4, allowed this transporter to be classified as a Na+/Cl-dependent NTT.
To demonstrate that this cDNA was the rosA transcript, the P-element germline mediated transformation technique (Spradling, 1986) was used to rescue the mutant pheno-type. The results clearly demonstrated that the isolated cDNA directly restored wild-type function in rosA mutants when introduced to the mutants (Burg et al., 1996). Tissue in situ hybridization experiments indicated that the rosA transcript was expressed in many tissue types in adults, including photoreceptors (Burg et al., 1996). The generalized expression pattern suggested that this particular transporter might serve a very basic function by transporting a molecule common to many cell types. Alternatively, the rosA gene might be generating splice variants to produce functionally distinct transporters. Some of these splice variants have been detected by Northern analysis of mRNA isolated from different developmental stages of Drosophila. For example, although adult head tissue yielded one major transcript of 3.7 kb, several transcripts were detected in the embryonic stage (12-24 h embryo), which ranged from 3.7 to 2.3 kb (Burg et al., 1996).
Another independently isolated mutant, ine (inebriated), was reported to enhance excitability of neural tissue, particularly when mutations in potassium channel genes were also present (Stern and Ganetzky, 1992). For example, when examined in a Sh mutant background (the Shaker gene encodes a K+ channel), the ine mutation caused many motor defects in larvae and developmental abnormalities in adults (Stern and Ganetzky, 1992). Simultaneous with the molecular cloning of the rosA locus, the ine locus was identified by using RFLP (Restriction Fragment Length Polymorphism) recombination mapping. Once the ine gene was mapped to a small region (24F) of the Drosophila genome, a deletion mutant (ine3) was generated to enable the physical identification of the ine locus in the genomic DNA. From this analysis, genomic clones were obtained (via "chromosome hopping"), which then were used to isolate cDNA clones from an embryonic cDNA library (12-24 h library). Both the clones and the corresponding genomic regions were mapped, and one mutant allele (ine3) was found to be a deletion of 2.4 kb in the central part of the coding region of the cDNA, whereas another mutant allele (ine4), was found to be a nonsense mutation (Soehnge et al., 1996). Sequence analysis of the ine cDNA identified the ine gene product as a Na+/Cl~-dependent neurotransmitter transporter, which was proposed initially to serve as a glutamate transporter. Tissue in situ hybridization experiments performed on Drosophila embryonic tissue demonstrated ine expression in the CNS and numerous other tissues in the developing embryo (Soehnge et al., 1996).
Because the rosA and ine genes encoded nearly identical proteins and were located in the same genomic region, it was possible that these two mutants, although derived through different mutagenesis screens, defined the same gene. To establish whether these two genes were identical, genetic complementation tests were done between mutant alle-les from each gene. In this genetic test, a homozygous mutant of rosA was crossed to a homozygous mutant of ine. The rosA/ine heterozygotes, the F1 progeny of this cross, were still mutant in the ERG phenotype (M. Burg, unpublished results), indicating that both rosA and ine mutations affected the same gene.
In addition to the genetic confirmation that rosA and ine were the same gene, all ine mutants exhibited the same rosA ERG mutant oscillation phenotype, and all rosA mutants showed both motor and developmental defects in a Sh mutant background, as is characteristic of ine (M. Burg, unpublished). The rosA cDNA isolated from an adult head cDNA library (Burg et al., 1996) and the ine cDNA isolated from an embryonic cDNA library (Soehnge et al., 1996) were very different in size; the former was 3.7 kb and the latter, only 2.4 kb. Sequence comparisons between the cDNAs showed that the open reading frames contained in the cDNAs were essentially identical, except that the sequence encoding the amino terminal 313 amino acids of the predicted rosa protein were replaced in the ine cDNA by a sequence encoding 27 different amino acids. This substitution occurred at an intron/exon boundary, with distinct exons encoding the different N-termini of the two proteins, which suggests that the two cDNAs represent splice variants. Indeed, Northern analysis and tissue localization of transcripts described above indicated that the rosA/ine transcript is very heterogeneous in expression pattern and that many splice variants are produced by this gene at different developmental stages (Burg et al., 1996).
One of the more useful methods for studying structure-function relationships in vivo is P-element mediated germline transformation by using promoters that express the gene of interest in a subset of cells. This method takes advantage of the availability of Drosophila mutants and the ease of making transgenic flies, which provides a unique advantage to Drosophila in the study of NTTs. The initial phenotypic rescue of the rosA photoreceptor mutant phenotype with this technique utilized the 3.7 kb "rosA" cDNA (Burg et al., 1996), subsequently named the rosA/ine-long transcript (Huang et al., 1999). This experiment was then repeated with the distinct 2.4 kb "ine" cDNA of embryonic origin (Soehnge et al., 1996), subsequently named the rosA/ine-short transcript (Huang et al., 1999; 2.3 kb), to determine whether the cDNA differences could represent functional divergence in transporter function in vivo. The results of this study indicated that, for all known phenotypes, the rosA/ine-long transcript was more effective in restoring RosA/Ine function than the rosA/ine-short transcript (Huang and Stern, 2000; Huang et al., 1999). The results suggested that the 313 aa N-terminus of this transporter (encoded only in the rosA/ine-long transcript) has an important function in localization or expression of the RosA/Ine transporter.
In an attempt to identify substrate(s) for the RosA/Ine transporter, substrate uptake assays were performed on Drosophila S2 cell lines transformed with the rosA/ine 3.7 kb (long) or the rosA/ine 2.3 kb (short) cDNA. Transformation of S2 cell lines utilized the pRmHa-3 (Bieber, 1994) or the pMT/V5-His vector (Invitrogen Inc., Carlsbad, CA), both of which carried a metallothionine promoter to induce gene expression. Details of the procedure are presented in Appendix I. Unfortunately, none of the following compounds were taken up specifically by cells in which the expression of the RosA/Ine transporter was induced: GABA, dopamine, norepinephrine, glycine, serotonin, choline, taurine, proline, histamine, and creatine. Consequently, the transporter encoded by rosA/ine was classified as an "orphan" transporter.
Because the transport experiments failed to identify any candidate substrate for the RosA/Ine transporter, mutant phenotypes were examined further to obtain clues to the transporter function. An electrophysiological examination of the larval neuromuscular junction showed that rosA/ine mutants exhibited altered synaptic transmission (Burg at al., 1997). Another phenotype reported more recently is the sensitivity of rosA/ine mutants to high osmotic conditions. The rosA/ine mutants were found to be unable to tolerate a high-salt food, on which wild-type flies thrive (Huang et al, 1999). This phenotype indicated that there may be a component of osmotic regulation that involves the RosA/Ine transporter.
Recently, a rosA/ine homologue was cloned in Manduca sexta (Chiu et al., 2000). A cDNA fragment was generated by RT-PCR by using degenerate primers against conserved regions of mammalian NTTs to isolate a cDNA from an embryonic cDNA library. The cDNA encoded a Na+/Cl-dependent neurotransmitter transporter with 55% amino acid identity to RosA/Ine. Analysis of the transporter function by using the Xenopus het-erologous expression system failed to identify the substrate for this transporter, which was reminiscent of results with RosA/Ine. Interestingly, this transporter had a long N terminus, a characteristic thought to be unique to the RosA/Ine transporter, along with the other characteristics common to all Na+/Cl-dependent neurotransmitter transporters. Oocytes injected with mRNA of this transporter were found to generate a chloride current on exposure to hyperosmotic conditions. This response appeared to be mediated via a phospholipase (PLC)-C-mediated signaling mechanism, because U73122 (a permeable PLC inhibitor) blocked this effect (Chiu et al., 2000). Thus far, this osmotic-sensing response, coupled with a PLC transduction cascade (presumably causing an intracellular Ca++ release), is the only glimpse into the molecular mechanism of this transporter type. Because the Drosophila mutant rosA/ine is sensitive to hyperosmotic conditions, it is possible that the RosA/Ine transporter might be functioning in vivo by responding to osmotic changes and transporting small osmolytes.
Identification of new genes in Drosophila has been facilitated greatly by the recent sequencing of the complete euchromatic genome of Drosophila (www.fruitfly.org). Mutants, once mapped, can now be screened efficiently from a host of candidate genes in the region to which the gene has been localized to identify the corresponding genes. Additionally, targeted mutagenesis can be performed on genes whose similarity to known proteins exists but precise function is unknown. Of the 26 predicted transporter genes in the Drosophila genome, only 12 have been identified through expression cloning or forward genetics approaches (Flybase, 1999). Analysis and mutagenesis of the remaining transporters will most likely provide new insights into the role transporters play in both neuro-
transmitter uptake and maintenance of basic physiological functions, such as osmolarity and resting membrane potential. It is also clear that the forward genetic method can yield surprises, as in the case of the rosA/ine gene, providing new insights into the role these proteins may serve in cells.
Was this article helpful?