Cloning and functional expression of the first eukaryotic Na⁺–tryptophan symporter, AgNAT6

The aromatic amino acids – phenylalanine, tyrosine and tryptophan– are critical for animal health and development. They serve as the metabolic precursors of proteins, signaling peptides, monoamine neurotransmitters and hormones. However, animal metabolism tends to be deficient in essential aromatic amino acids that are not synthesized de novo. The exception is tyrosine, which can be produced in animal cells from the essential precursor phenylalanine. In addition, aromatic amino acids comprise the most under-represented group of ecological nutrient chains [on average phenylalanine accounts for 3.00%, tyrosine for 3.06% and tryptophan for 1.02% of the amino acid residues in proteins(Brooks et al., 2002)]. In meeting metabolic demands, metazoan cells actively absorb aromatic amino acids despite low membrane permeability and unfavorable concentration gradients(Chakrabarti, 1994; Chakrabarti and Deamer, 1994). However, there is a serious knowledge gap regarding the molecular mechanisms of absorption. For example, no presently known metazoan transporter is able to transport adequate amounts of tryptophan due to insufficient selectivity,spatial distribution or energy-coupling mechanisms. In contrast after a highly selective tryptophan transporter was found in the prokaryotic neurotransmitter sodium symporter family (NSS, also known as solute carrier family 6, SLC6 by HUGO nomenclature). The tnaT gene encodes a unique Na⁺-dependent mechanism with strong tryptophan selectivity. It has been cloned from Symbiobacterium thermophilum(Androutsellis-Theotokis et al.,2003), a microorganism whose growth depends on commensalism and whose genome includes several transporters for peptides and amino acids(Ueda et al., 2004), but only one SLC6 member. The tnaT transporters mediate Na⁺ gradient-coupled uptake of tryptophan and act in synergy with other members of the tnaT operon of S. thermophilum that grow in nutrient-enriched media (Ueda et al.,2004). The SLC6 family expands dramatically in Metazoa (∼20 members per genome in mammals and insects compared with 0–4 members per genome in prokaryotes) with segregation into two large subfamilies:neurotransmitter transporters (NTTs) and nutrient amino acid transporters(NATs) (Boudko et al., 2005b). NTTs absorb two signaling amino acids, Gly and GABA, and the essential aromatic amino acid precursors of the signaling monoamines: dopamine,serotonin, norepinephrine (noradrenaline) and octopamine. All characterized NATs mediate [Na⁺ or K⁺]:[neutral amino acid] symport. The metazoan NATs are functional counterparts and evolutionary progeny of prokaryotic SLC6 members such as tnaT. Ironically, S. thermophilum, a compost bacterium, possesses a single SNF member per genome that mediates the absorption of energy-rich tryptophan for catabolic fuel, but no active tryptophan transporter has been identified in metazoan organisms, even though it is their most essential nutrient.

Previously, we characterized the first of the seven NAT-SLC6 members in the Anopheles gambiae genome, AgNAT8(Meleshkevitch et al., 2006). It mediates Na⁺ (or K⁺)-coupled symport of l-phenylalanine, l-tyrosine and 3,4-dihydroxy-l-phenylalanine (l-DOPA). In the absence of these substrates, AgNAT8 can absorb l-tryptophan and 5-hydroxytryptophan (5-HTP); however, it cannot acquire the indole-branched substrates in the presence of physiological concentrations of phenyl-branched substrates. Now, we have cloned and characterized AgNAT6, a phylogenetically close relative of AgNAT8. It absorbs tryptophan and, with less efficiency,other aromatic substrates. The identified indole- and phenol-branch-specific AgNATs have surprisingly little variation in their substrate-binding pockets,exhibiting only three amino acid differences that are likely to be responsible for their unique selectivity profiles. The AgNAT6 expression pattern correlates with tissues that are involved in nutrient absorption and neuronal functions.

AgNAT6 was cloned from a cDNA collection of An. gambiae larval midgut into pGEM-T and pXOON vectors as described elsewhere(Meleshkevitch et al., 2006). Exact primers with or without BamHI (5′-cgG GAT CCc gAT GGC GAC CAG TAA TCC GGC CTT T-3′) and EcoRI(5′-cgG AAT TCcg TCA ACT GAA CAC GTT GTC GAA CAT-3′)flanking sequences (underlined) were used to obtain PCR products which were inserted into the BamHI/EcoRI-digested expression vector pXOON (Jespersen et al., 2002)or the pGEM-T cloning vector, respectively. cRNA for oocyte injections was obtained by in vitro transcription of PmeI-linearized AgNAT6-pXOON plasmids using mMessage mMachine®, a high-yield, capped RNA transcription kit (Ambion, Austin, TX, USA). Surgically isolated and collagenase-treated stage V–VI Xenopus laevis oocytes (Nasco,Fort Atkinson, WI, USA) were injected with ∼40 ng of AgNAT6 cRNA and incubated for 2–6 days at 17°C in sterile N98-oocyte medium(Boudko et al., 2005a). AgNAT6-injected oocytes were analyzed in a small volume perfusion chamber using 2-electrode voltage clamp techniques identical to those used to characterize AgNAT8 (Meleshkevitch et al.,2006).

AgNAT6 sequence fragments were assembled using SeqManII (DNASTAR, Madison,WI, USA). Homologous sequences were derived from non-redundant protein databases (NCBI). The phylogenetic analysis included 63 SLC6 members from selected dipteran genomes plus a few reference sequences of cloned NAT-SLC6 members. The protein alignments were generated using ClustalX(Thompson et al., 1997) and visualized using GeneDoc software (Nicholas et al., 1977). A phylogenetic tree was constructed using Mega 4 software(Tamura et al., 2007). Fine-tuning of the alignment was performed manually by considering sequence and 3D structure alignments of selected NATs with LeuT (PDB ID 2a65)(Yamashita et al., 2005). Homology modeling was performed by satisfaction of spatial restraints(Sali and Blundell, 1993)using Modeler (Marti-Renom et al.,2000).

RNA isolation, cDNA synthesis and qPCR analysis were performed as described previously (Meleshkevitch et al.,2006). Primers were designed with the following sequences:5′-GGC AAC ACC AGT CGA ACC A-3′ and 5′-GCT GCA CCT TGT GGA TGT TCT-3′. Expression of AgNAT6 in each tissue was normalized to that found in whole larvae, which was set to a value of one. Data represent three averaged replicates of three independent experiments. Excel software(Microsoft Inc., Redmond, WA, USA) was used to collect and analyze data and SigmaPlot 10 (Systat Software, Inc., San Jose, CA, USA) was used to generate final graphs.

Uptake assays were performed 4 days after RNA injection on oocytes using a previously described protocol(Meleshkevitch et al., 2006). Briefly, distilled water (DW)- and AgNAT6 RNA-injected oocytes were exposed in a 100 mmol l^–1 NaCl solution supplemented with 10 μCi ml^–1l-[5-³H]tryptophan or l-[2,3,4,5,6-³H]phenylalanine (specific activity 25 and 125 μCi mmol l^–1, respectively, Amersham Biosciences,Piscataway, NJ, USA), brought to a final concentration of 1 mmol l^–1 with unlabeled tryptophan or phenylalanine. Uptake was terminated after 10 min by washing oocytes with a cold 100 mmol l^–1 choline chloride solution. Pairs of oocytes were placed in scintillation vials with 200 μl of 10% SDS; 4 ml of scintillation fluid was added, and the radioactivity was counted using a Beckman-Coulter LS 6500 scintillation counter (Beckman-Coulter, Inc., Fullerton, CA, USA). The quantity and ratio of substrate uptake were calculated as described earlier(Meleshkevitch et al., 2006). The measured ratio values were normalized vs the value for radiolabeled l-tryptophan uptake in AgNAT6 transcript-injected oocytes.

Anopheles gambiae larvae (G3 strain) were hatched from eggs supplied by MR4 (The Malaria Research and Reference Reagents Resource Center,Atlanta, GA, USA) and raised as described earlier(Okech et al., 2008b). A purified pGEM-T AgNAT6 plasmid was linearized with NcoI or NotI restriction enzymes to obtain full-length, run-off transcripts using SP6 and T7 promoters for anti-sense and sense probes, respectively. DIG-labeled probes were transcribed in vitro using a DIG RNA labeling kit (Roche Diagnostics, Mannheim, Germany). Fourth instar An. gambiaelarvae were immobilized in ice-cold phosphate-buffered saline (PBS, Fisher Scientific, Itasca, IL, USA) opened by a lateral incision, and fixed in 4%paraformaldehyde/PBS overnight. Preparations were dehydrated/rehydrated by passing through a PBS/methanol gradient set (100%PBS–3:1–1:1–1:3–100% methanol, then in reverse order),10 min for each mixture, pre-treated with proteinase K/detergent solution(0.1% Tween-20 in PBS supplemented with 10 μgml^–1proteinase K) for 30–40 min. The preparation was pre-hybridized for 6–8 h at 50°C in hybridization solution (50% formamide, 5 mmol l^–1 EDTA, 5× SSC, 1× Denhardt's solution, 0.1%Tween-20, 0.5 mg ml^–1 yeast tRNA), then hybridized by incubation with approximately 1 μg of DIG-labeled RNA probe per ml of hybridization solution at 50°C overnight. The hybridized preparations were labeled with alkaline phosphatase-conjugated, anti-DIG antibodies according to the manufacturer's protocol. Hybridization patterns were visualized in a NBT/BCIP alkaline buffer solution (Boehringer Mannheim, Inc., GmbH, Mannheim,Germany). Labeled preparations were embedded in 3:1 glycerol:PBS on glass slides and photographed using an Olympus SZX 12 stereo microscope (Olympus America, Center Valley, PA, USA) and Pixera CCD camera (Pixera Corp., Los Gatos, CA, USA).

Values depicted in graphs represent the mean ± s.d. from at least three different experiments using at least three different oocytes. The electrical current amplitudes were normalized relative to a maximum current in each data set. Kinetic profiles and constants were derived by fitting normalized data sets with a three-parameter sigmoidal Hill function y=ax^b/(c^b+x^b);where: a=ymax, the derived maximum current; b=η, the order of the transport process; c=x_50%; (x,y)=K_0.5,_[S], the substrate concentration at 50% of the transport velocity.

AgNAT6 was cloned from a midgut cDNA collection of 4th instar mosquito larvae (GenBank CDS AJ626713 and PID CAF25029). It has a 1953 bp ORF encoding a 650 amino acid-long protein with an estimated molecular mass of 71.5 kDa and pI of 7.31. It is the closest phylogenetic relative to the earlier characterized Na⁺–phenylalanine, phenyl-branched substrate symporter AgNAT8 (Fig. 1;pairwise similarity 52.9%, identical sites 343, 51.7%) and a newly characterized member of the NATs subfamily of the SLC6 family. The AgNAT6 and AgNAT8 genes result from the most recent duplication in the AgNAT-SLC6 cluster(Fig. 2). They are the first upstream and the last downstream genes, respectively, of a seven-gene NAT cluster that is mapped on the positive strand of the 3L chromosome between 12.09 and 12.16 Mb positions(Meleshkevitch et al., 2006),based on the genome annotation (Holt et al., 2002).

AgNAT6 has a typical SLC6 structure with 12 transmembrane domains and intracellular C and N termini (Fig. 1). Twelve coordinating organic substrate residues are identifiable based on the sequence-structure alignment of AgNAT6, AeAAT1(Boudko et al., 2005a), AgNAT8(Meleshkevitch et al., 2006)and TnaT (Androutsellis-Theotokis et al.,2003) with crystallized LeuT_Aa(Yamashita et al., 2005)(Fig. 1; Table 1). The substrate-interacting moieties comprise a surprisingly conserved pattern, with nine absolutely conserved sites between AgNAT6 and AgNAT8(Table 1). Three different amino acids in transmembrane domains 1, 6 and 8 could be responsible for the tryptophan selectivity of AgNAT6. Specifically they include: L91, because this site is different from a corresponding M95 site in AgNAT8; T333,which is different from AgNAT8 S339 and identical to TnaT T234, and G437 which is unique to AgNAT6 (d, i and u indicate corresponding sites in Table 1). These three positions correspond to N21, S356 and A358 of LeuT_Aa. All three substitutions aid in the reduction of the side-chain volume and corresponding increases in the substrate-binding envelope volume, which correlates with a capacity to accept larger sized indole- vs phenol-branched substrates(Fig. 3A, inset).

The expression of all seven AgNATs, nine AeNATs and six DmNATs was confirmed by molecular cloning from specific cDNA collections (all 22 clone sequences are available in the NCBI database, see Fig. 1 for NCBI accession numbers), but presently only two AgNATs have been characterized by heterologous expression; previously, AgNAT8(Meleshkevitch et al., 2006)and, now, AgNAT6. The application of aromatic amino acids produced a barely detectable current in naive and DW-injected X. laevis oocytes (<5 nA at 1 mmol l^–1 concentration of aromatic substrate in ND98 at –50 mV holding transmembrane potential, Hp; data not shown). AgNAT6 expression increased a background leak current from 7±5 nA to 82±24 nA, N>100 at –30 mV Hp) and mediated large aromatic amino acid-induced currents of 20–150 nA(Fig. 3A). In contrast to phenylalanine-preferring AgNAT8, AgNAT6 responded with notably larger currents upon tryptophan and 5-HTP application (Fig. 3B), suggesting a higher transport velocity for indole- than phenol-branched substrates. Among Phe-derived metabolites only Trp (addition of a hydroxyl group to the end of the 6-carbon aromatic ring with corresponding increase in molecular volume) produced currents that were similar to the indole-branch substrate-induced currents(Fig. 3A). Replacing Na⁺ with Li⁺ or K⁺ abolished tryptophan-induced currents at –30 mV Hp(Fig. 3D). However, the extended current–voltage (I–V) graph revealed that AgNAT6 can generate large substrate-coupled K⁺ currents at transmembrane voltages more negative than –40 mV(Fig. 3E). Isotope-uptake experiments confirmed AgNAT6-coupled organic substrate uptake with notably higher uptake ratios for isotope-labeled tryptophan than for phenylalanine(Fig. 3F).

The AgNAT6 mechanism has saturable kinetics(Fig. 4A). Tryptophan has a remarkably higher apparent affinity than all other tested substrates(K₅₀^Trp=1.3 μmol l^–1compared with K_0.5^5-HTP=270<K_0.5^Tyr<K_0.5^Phe<K_0.5^DOPA<K_0.5^Leu=890μmol l^–1). The orders of the organic substrate translocation reaction (Hill constant η) determined at 98 mmol l^–1 Na⁺ concentrations were all similar and approached 1 (Fig. 4B).

AgNAT6 tested at two different concentrations of tryptophan showed little increase in K₅₀^Na+ at 3 mmol l^–1 Trp vs 0.3 mmol l^–1 Trp(Fig. 4C). Unexpectedly, we have identified a capacity of AgNAT6 to exhibit Hill coefficients of 2 and 1 Na⁺ for tryptophan translocations at 0.3 and 3 mmol l^–1 concentrations, respectively(Fig. 4C). This `slip' may suggest a switch of the preferred transport stoichiometry in the AgNAT6 mechanism: 1 Na⁺:1 amino acid at high concentrations of organic substrates and 2 Na⁺:1 amino acid at low concentrations of organic substrates. AgNAT6 showed non-linear pH dependency with relatively high transport turnover at moderately acidic (pH 6) and strongly alkaline (pH 8–9) vs neutral (pH 7) external solutions.

In situ hybridization of the larval alimentary canal with an AgNAT6 antisense probe revealed a relatively high accumulation of AgNAT6 transcript in the heart, gastric caeca and posterior midgut(Fig. 5A). Weaker but detectable labeling was present in the anterior midgut(Fig. 5A). Strong in situ hybridization signals were detected in the neuronal plexus of the larval head associated with chemo-, visual- and mechano-sensory modalities and in the neuropile of the ventral nerve cord(Fig. 5B,C). A strong signal was detected in a few individual neurons upon low resolution imaging of whole-mount preparations (images not shown). A very strong signal was also detected in the 2nd and 3rd instar larval alimentary canal, especially when young larvae were exposed to limited nutrient conditions(Fig. 5D). qPCR confirmed a ubiquitous expression of AgNAT6 with very strong tissue-specific and developmental stage-specific variations of AgNAT6 transcript quantities(Fig. 5E).

Several transporters which are thought to translocate aromatic amino acids via the plasma membrane of metazoan cells have been characterized previously (Table 2). However,the molecular identity of an active mechanism for absorbing tryptophan remained enigmatic. Here we report the molecular cloning and characterization of AgNAT6, the first metazoan transporter with definite roles in the absorption and systemic redistribution of tryptophan and related indole-branched substrates.

AgNAT6 is a new member of the NAT-SLC6 subfamily(Fig. 1). All currently characterized NATs encode transporters with similar electrochemical mechanisms utilizing the most ubiquitous monovalent cation (primarily Na⁺)electrochemical gradient to symport (co-transport) neutral amino acids. The NATs population includes broad substrate spectrum, neutral amino acid transporters, which were identified in mammals(Broer et al., 2004) and insects [see figure 1s,B⁰AT1 and DmNAT1, in Miller et al.(Miller et al., 2008)]. In insects the NAT subfamily has an additional expansion, which is absent in mammals. This expansion includes seven genes in An. gambiae(Fig. 2). One of these genes,AgNAT8, was recently characterized as a Na⁺-dependent,voltage-driven (Harvey et al.,2009) mechanism that is adapted to transport phenylalanine and its related metabolites (Meleshkevitch et al.,2006). The present characterization of its closest phylogenetic relative, AgNAT6, reveals a different mechanism, specializing in the active absorption of tryptophan and its related metabolites. AgNAT6 and AgNAT8 are expressed in similar loci and may act in synergy to optimize the absorption of indole- or phenyl-branched substrates, respectively(Okech et al., 2008b). The aromatic AgNATs result from earlier gene fission in the NAT population of An. gambiae that may correspond with adaptations of mosquito larvae for development in habitats with a low level of essential nutrients. The identified properties of AgNAT6 fit with our previous hypothesis that NATs undergo rapid gene duplication which drives the specialization of individual transporters and adaptive plasticity of integrated NAT functions(Boudko et al., 2005a).

AgNAT6, like all other characterized NATs, is a rectifying, normally irreversible, transporter (Fig. 3E). Both aromatic AgNATs use Na⁺ electromotive forces;however, they could utilize K⁺ electromotive forces with appropriate inward directions (Fig. 3E). In this respect they represent a transitional phenotype between Na⁺-specific and K⁺-preferring NATs such as mammalian B⁰ATs (Broer et al.,2004; Broer et al.,2006) and caterpillar KAAT1(Castagna et al., 1998),reflecting adaptations of these organisms to high-sodium ion (in mammals) and trace-sodium ion/high-potassium ion (in caterpillars) environments. Correspondingly, the capacity of AgNATs to use Na⁺ and K⁺ can be explained as an adaptation to varied concentrations of these cations. For example, mosquito larvae are able to deal with low concentrations of both alkali metal cations in their freshwater habitat. A cation pool that is depleted by NATs activity in the absorptive region of the posterior midgut may be internally recycled in mosquito larvae viathe earlier proposed anterior midgut alkalinization pathways(Harvey et al., 2009; Okech et al., 2008a). However,considering the limited efficiency of a cation-recycling framework, the balance of the Na⁺/K⁺ flux would be dependent upon the environmental availability of these ions. The availability of Na⁺and K⁺ also varies over the life history of adult mosquitoes. The use of Na⁺ is expected to be very low in adults feeding on plant fluids; but it may increase dramatically after a blood meal.

Both aromatic NATs are sensitive to extracellular Cl^–activity. When it is present outside Cl^– increases the net current that is carried via these transporters(Fig. 3D,E). This observation suggests that Cl^– may act as a NAT modulator rather than an electrophoretic carrier because adding negative charges to an inward cation:neutral amino acid symport will reduce rather than increase the net inward current (e.g. Fig. 3D;by convention in neurophysiology and insect epithelial transport physiology a positive ion moving into a cell or a negative ion moving out is said to carry an inward current). The data present another interesting possibility –that AgNAT6 function is facilitated by bidirectional (electroneutral)Cl^– transport. During outward movement Cl^–may energize changes in transporter conformation. A similar mechanism was described in SLC1 transporters that use K⁺ to reset conformations(reviewed by Kanai and Hediger,2004). The structural basis of Cl^– interaction with SLC6 members was the subject of a recent analysis suggesting a possible replacement of Cl^– by another negatively charged ion in chloride-independent members of SLC6(Forrest et al., 2007; Zomot et al., 2007). However,the exact role of this anion in SLC6 function remains to be clarified. AgNAT6–Cl^– interaction will require a more detailed analysis that would be well beyond the scope of an initial characterization of AgNAT6.

Despite its obvious similarity with AgNAT8, the electrochemical properties of AgNAT6 have some unique traits. A most notable difference is in organic substrate specificity with AgNAT6's strong and narrow preference to indole-branched substrates (Fig. 3A, Fig. 4A,B). Collected data suggest that AgNAT6 is specialized to transport tryptophan with minimal interference from phenylalanine and other neutral amino acids, the cumulative concentration of which is about two orders of magnitude greater than that of tryptophan in the larval and adult nutrient digests [inferred from the frequency of amino acids in protein sequences(Brooks et al., 2002)]. Comparison of AgNAT6 and AgNAT8(Meleshkevitch et al., 2006) in situ hybridization (Fig. 5) and immunolabeling (Okech et al., 2008b) profiles revealed broadly similar localizations,but with notable differences in the details. Specifically, AgNAT6 occupies a broader area of the larval gut (Fig. 5) with significant differences in the polar distribution in the gastric caeca and anterior midgut area relative to AgNAT8(Okech et al., 2008b). Aromatic AgNATs also possess different expression patterns in the central nervous system and sensory afferents (Fig. 5B,C) (Meleshkevitch et al.,2006). These differences in expression may both reduce morphological overlap of aromatic NATs activities and support specific functions with elevated requirements for aromatic substrates. In the midgut the differing distributions may reduce competition of aromatic AgNATs for the energy of transmembrane electrochemical gradients whereas in the CNS they may support cell-specific metabolic processes, e.g. synthesis of indoleamine and catecholamine neurotransmitters.

AgNAT6 generates a large Na⁺ leak current that was not observed upon heterologous expression of AgNAT8(Meleshkevitch et al., 2006)or any other mosquito NAT such as AeAAT1 from Aedes aegypti(Boudko et al., 2005a). In this respect AgNAT6 is similar to CAATCH1 from the caterpillar(Quick and Stevens, 2001),which may act as a cationic leak channel with some inhibition by weakly transported substrates (i.e. Fig. 3C for l-Phe and l-DOPA). However, that kind of response was observed only 6–8 days after injection of AgNAT6 RNA into oocytes. It equally may be an artifact of heterologous over-expression rather than a physiologically significant mechanism. Additional experiments will be necessary to validate this phenomenon and to understand its physiological implications.

The pH dependency profile of AgNAT6 is different from that of AgNAT8,perhaps reflecting adaptations to the changing pH profile along the alimentary canal of mosquito larvae (Clements,1992; Ramsay,1950). The transport stoichiometry of AgNAT6 changes reversibly between 1 and 2 Na⁺ per amino acid transported. The physiological benefit of such an adaptation is obvious: 1:1 operation will conserve Na⁺ pool-coupled electrochemical energy, whereas 1:2 operation permits organic substrate translocation against a 2 times higher chemical gradient. The recent crystallographic structure of the bacterial NAT-SLC6 member LeuT_Aa suggests that a single sodium ion is sufficient for substrate coordination in the substrate-binding pocket of the transporter(Yamashita et al., 2005). The paralogous mammalian B⁰ATs operate with 1:1 substrate stoichiometry(Broer, 2008), whereas other members of the NTT-SLC6 subfamily may use 3, 2 or 1 Na⁺ for 1 neurotransmitter molecule (Chen et al.,2004). For example, GlyT2a and GlyT1b expressed in neuronal and glial cells respectively conduct 3 Na⁺:Cl^–:Gly and 2 Na⁺:Cl^–:Gly symport(Roux and Supplisson, 2000). So, switching of a `stoichiometry gear' – often referred to as a `slip'– is common in the history of SLC6 family members, even among recently diverged transporters. It is also likely that the slip in AgNAT6 stoichiometry may depend upon physiological conditions via a yet to be elucidated mechanism. Incidentally, the `low' tested concentration of the substrate is still 230-fold greater than that of the K₅₀^Trp,suggesting an organic substrate binding, kinetically uncoupled mechanism. Considering the possibility of a concentration-dependent shift of the dose–response curve, a more direct and precise technique of stoichiometry assays will be necessary to clarify this phenomenon.

The presently characterized aromatic amino acid-specific NATs include one bacterial (Androutsellis-Theotokis et al.,2003) and two mosquito transporters(Meleshkevitch et al., 2006; Okech et al., 2008b) (and data presented here). The usage of aromatic amino acids varies dramatically in different organisms, developmental stages, tissues and individual cells. Their availability also varies with ecological and nutrient conditions. For example,freshwater mosquito larvae filter-feeding on micro-organisms and organic debris, and mammals digesting protein-rich food may have a different concentration profile of essential amino acids in the alimentary canal and systemic circulations. Mosquito larvae require high quantities of aromatic amino acids for cuticle formation and tanning during successive ecdyses as well as protein accumulation between ecdysial events. A high-throughput transport of aromatic amino acids from the digestive system to the ovaries is critical during egg development and chorion formation. AgNAT6 is perfectly suited to acquire the most under-represented essential amino acids from the larval environment as well as to normalize the supply of such amino acids during specific metabolic processes such as oogenesis or neurotransmitter synthesis.

In summary, we have characterized the first eukaryotic Na⁺:tryptophan symporter, AgNAT6, which also represents the second characterized narrow substrate spectra transporter (the first being AgNAT8)cloned from a biomedically important model organism, the African malaria mosquito, An. gambiae. AgNAT6 plays a key role in the uphill delivery of essential tryptophan and, in synergy with phenyl-branched substrate-specific AgNAT8, mediates comprehensive absorption and systemic redistribution of aromatic substrates. These narrow selectivity NATs may have evolved under pressure of high demand vs low availability for aromatic amino acids. A likely reason for the additional expansion of NATs in invertebrates with narrow substrate spectra is the amplification of the transport network for the acquisition and redistribution of environmentally scarce substrates. It is reasonable to propose that metazoan organisms competing for an identical set of essential amino acids evolved distinct strategies and mechanisms for their absorption and redistribution. Future analysis of NAT populations is necessary for a better understanding of the structural, integrative and evolutionary aspects of SLC6 functions. The fact that NATs are lineage-specific and unique providers of essential for development substrates identifies them as targets for environmentally safe control of pests and pathogen-transmitting invertebrate organisms.