Phylogeny and cloning of ion transporters in mosquitoes

Blood-feeding insects have evolved highly specialized mechanisms to regulate fluid and ionic concentrations that would otherwise overwhelm the organism. These regulatory aspects are much appreciated in adult female insects such as Aedes spp. and Rhodnius prolixus and have been the focus of classical work in insect physiology by Ramsay and Wigglesworth. In recent decades, the cellular basis for such regulation has been elucidated in general terms. However, little is known about the molecular mechanisms involved in ion and fluid secretion in hematophagous insects.

Ion regulation is governed by the concerted actions of several membrane transport proteins, intracellular signaling molecules and brain-derived factors that regulate these proteins. Unlike vertebrates, in which ion homeostasis can be achieved by ultrafiltration, insects operate at low blood pressure and thus require hormonal signaling from the brain to regulate active ion transport. Brain-derived factors or hormones provide a means of regulating transporter activities. Several peptides that control diuresis and antidiuresis have been identified in insect Malpighian tubules(Coast, 1996). Signaling molecules involved in these opposing actions of the peptides also differ, in that cAMP stimulates fluid secretion and cGMP promotes antidiuresis.

Natriuretic peptides that stimulate secretion by Malpighian tubules have been isolated from mosquitoes (Petzel et al., 1987). However, the molecular targets of these peptides in stimulating ion and fluid transport can only be gleaned from studies that use pharmacological agents that act on known targets. Of special interest are processes affected by application of bafilomycin A₁, ouabain,bumetanide and amiloride, which respectively inhibit V-type H⁺-ATPase, Na⁺/K⁺-ATPase,Na⁺/K⁺/Cl^--cotransporters (CCCs) and sodium/proton exchangers (NHEs).

The cytoplasmic face of insect epithelia possesses characteristic stalked structures (Gupta and Berridge,1966), which were later called `portasomes' and have been implicated in K<sup>+</sup> transport in the lepidopteran midgut(Harvey, 1992; Harvey et al., 1983a,b). Harvey and colleagues first demonstrated that midgut membranes contain a K<sup>+</sup>-ATPase, which was also known as the `K⁺ pump'. A major accomplishment of these workers was their ability to isolate different membrane domains from the midgut as clean fractions(Cioffi and Harvey, 1981; Cioffi and Wolfersberger,1983). The K⁺ pump was identified in the goblet cell apical membrane (GCAM) fraction using the portasome assay(Cioffi and Harvey, 1981). Once the membrane fraction containing the pump was purified, the K⁺-transporting component was solubilized and identified as a V-ATPase (Schweikl et al.,1989; Wieczorek et al.,1989).

The V-ATPase is a large oligomeric complex consisting of two distinct sectors: the membrane-bound V₀ sector and the intracellular V₁ sector (Forgac,1999; Nishi and Forgac,2002). Protons generated from cellular respiration appear to be the major substrate for V-ATPase since mitochondrial poisoning with dinitrophenol abolishes measurable V-ATPase activity and, subsequently, ion transport. Furthermore, V-ATPases are present in close physical proximity to mitochondria, enabling protons to be channeled efficiently through the proton pump (also see Harvey and Wieczorek,1997). The role of V-ATPase in ion transport is further supported by its inhibition by bafilomycin A₁, which in parallel inhibits transport in insect epithelia. Consistent with this role, immunohistochemical analyses have demonstrated the presence of subunits of V-ATPase in K⁺-transporting midgut membranes(Klein, 1992; Sumner et al., 1995; Zhuang et al., 1999) and mosquito Malpighian tubules (Filippova et al., 1998; Weng et al.,2003; Pullikuth et al., submitted). Nearly all subunits of V-ATPase have been cloned, and the expression patterns of some have been examined (Filippova et al.,1998; Gill et al.,1998; Graf et al., 1994a,b; Merzendorfer et al., 2000; Pietrantonio and Gill, 1995),showing that they colocalize with the portasomes(Zhuang et al., 1999). Structural analysis now implies that portasomes are actually the V₁sector of V-ATPase (Radermacher et al.,1999; Rizzo et al.,2003).

The V-ATPase transports H⁺ but not K⁺, yet K⁺ is the only ion that crosses the entire midgut in lepidopterans(Cioffi and Harvey, 1981). Further biochemical characterization led to the hypothesis that the GCAM contains a K⁺/H⁺ antiporter that uses the proton-motive force generated by V-ATPase in GCAM to drive transport(Harvey, 1992; Harvey et al., 1983a; Wieczorek et al., 1991). This transports K⁺ out to the lumen in an electrogenic exchange for luminal protons (Azuma et al.,1995). Thus, current models for insect ion transport incorporate V-ATPase in transporting cells as one of the major determinants of transport that favors activation of secondary transport processes for ion and fluid effluxes (Wieczorek, 1992;Wieczorek et al., 1991, 1999). However, the Na⁺/K⁺-ATPase may also be involved in fluid and ion transport, for example in insect Malpighian tubules (M. J. Patrick, K. Aimanova, H. R. Sanders and S. S. Gill, submitted).

The lepidopteran apical proton pump is apparently electrogenic, exchanging 2H⁺ for every cation efflux(Azuma et al., 1995; Grinstein and Wieczorek, 1994; Lepier et al., 1994). In lepidopteran goblet membrane preparations, vesicle acidification by V-ATPase was inhibited by bafilomycin A₁, whereas outward-directed K⁺ or Na⁺ transport was inhibited by amiloride and harmaline; features that indicate involvement of a K⁺/H⁺exchanger and Na⁺-transport mechanisms. V-ATPase could acidify independently of extracellular K⁺ when K⁺ efflux was blocked by amiloride, indicating that V-ATPase and K⁺-transporting molecules are different (Wieczorek et al.,1991). Furthermore, antibodies that blocked V-ATPase activity and ATP-dependent H⁺ transport had no effect on K⁺/H⁺ exchange. Together, these results led to the hypothesis that the lepidopteran K⁺/nH⁺exchanger is involved in electrogenic K⁺ efflux aided by the proton-motive force generated by the electrogenic V-ATPase(Grinstein and Wieczorek,1994; Harvey and Wieczorek,1997; Wieczorek et al.,1991).

Despite the accepted role of V-ATPase in energizing epithelial transport,little is known about the molecular identity of transporters that operate in parallel with V-ATPase. However, pharmacological approaches with known inhibitors of NHEs, such as amiloride and its related compounds, suggest that a proton/cation exchanger is intricately involved in ion transport in mosquito Malpighian tubules. In addition, involvement of other transporters, including those similar to bumetanide-sensitive sodium/chloride cotransporters and ouabain-sensitive Na⁺/K⁺-ATPase, contributes in varying degrees to ion transport mechanisms in many insects (reviewed in Pannabecker, 1995). Clearly,molecular identification of these transporters is a necessary step in conclusively assigning the involvement of such transporters to regulating ion transport in mosquitoes.

Whole-tubule assays and short-circuit and other electrophysiological analyses have helped to understand ion fluxes and thus fluid secretion in Malpighian tubules (Beyenbach,1995; Beyenbach et al.,2000). Acute activation of ion and fluid transport appears to be correlated with cAMP-dependent mechanisms. Fluid secretion by isolated Malpighian tubules can be stimulated by peritubular application of cAMP analogues or by activating processes that sustain higher cellular cAMP such as active adenylate cyclase or inhibition of phosphodiesterases(Petzel et al., 1987). Similar properties could be mimicked by the application of mosquito natriuretic peptides (MNPs). These brain-derived peptides possess diuretic properties, and their release is possibly controlled by activation of stretch receptors following ingestion of a blood meal(Petzel et al., 1987; Fig. 1).

From the section above, it is apparent that a clear understanding of hormones involved in mosquito diuresis and activation is essential to correlate their roles in activating specific intracellular signaling culminating in functioning of membrane transport proteins that mobilize ions and fluid. By analogy to well-characterized systems in vertebrate fluid transport, some important functional properties are expected to be conserved in mosquitoes. These processes might broadly involve factors that recognize ion and fluid changes, volume sensing and hormones released in response to these changes. Collectively, these parameters may influence distinct steps in maintaining the overall ion and fluid composition to physiologically relevant limits. Ion/proton exchangers (e.g. NHEs), anion/cation cotransporters/exchangers (e.g. CCCs), regulators of pH (e.g. carbonic anhydrase, NHEs, HCO₃ transporters/exchangers,H⁺/K⁺-ATPases and V-ATPases) and aquaporins are likely candidates that act downstream of hormonal cues in regulating several aspects of ion and fluid secretion in mosquitoes. Our current knowledge on the molecular identity of insect members of two families, namely the NHEs (solute carrier family 9) and cation-coupled chloride cotransporters (CCCs, solute carrier family 12), is provided below to appreciate their potential roles in insect ion homeostasis, where they operate together with V-ATPase and Na⁺/K⁺-ATPase.

Members of solute carrier family 9 are involved in exchanging protons for cations (Na⁺ or K⁺), thereby regulating cellular and systemic pH and ionic concentrations. In mammals, NHEs function in a variety of cells and intracellular locations, indicating that members of this family perform ubiquitous housekeeping functions as well as more specialized physiological roles (Burckhardt et al.,2002; Counillon and Pouyssegur, 2000; Orlowski and Grinstein, 1997). Cation transport dictated by the transmembrane voltage established by apical V-ATPase in insect epithelium could be mediated through members of the NHE family. Support for this notion comes from the pharmacological properties of K⁺ and Na⁺ transport in insect epithelia. Peritubular application of amiloride inhibits basal fluid secretion in Malpighian tubules of Aedes aegypti(Hegarty et al., 1991), and serotonin (5-HT) stimulates secretion in Rhodnius prolixus(Maddrell and O'Donnell,1992). This effect probably does not involve basal membrane Na⁺ channels, as no detectable effect on basal membrane voltages was observed (Hegarty et al.,1991). Amiloride is also an Na⁺ channel antagonist,which confounds the above issues. However, D. melanogaster Malpighian tubules were shown not to express any known Na⁺ channels(Giannakou and Dow, 2001). The amiloride derivative EIPA, which is more potent towards NHEs, does indeed affect tubular pH and recovery from acid load, lending support to the involvement of NHE-like molecules in apical membrane transport(Petzel, 2000).

In mammals, eight distinct NHEs have been cloned. These exchangers exhibit differential tissue distribution, cellular localization and transport properties, indicating their specialized functions in particular tissues. A detailed review of mammalian NHEs can be found in Orlowski and Grinstein(1997), Counillon and Pouyssegur (2000) and Burckhardt et al. (2002) and references cited therein. Mammalian NHEs are predicted to form 10-12 transmembrane (TM) helices with two distinct domains. The N-terminal half of the protein contains all the TMs and is sufficient to transport H⁺and cations. The current working model of membrane topology for NHEs(Fig. 2) is largely based on NHE1 (Wakabayashi et al.,2000). Future structure-function studies and approaches similar to those described above would be necessary to determine whether this model is applicable to all other NHEs as well.

The cytoplasmic carboxy region modulates the overall function in response to pHi and other regulatory factors. In the absence of the carboxy region,high proton concentration (low pHi) is required for NHE1 activation. Thus,intracellular protons bind to the sensor site, and the cytoplasmic region integrates the `pH set point value' for NHE1 activation(Orlowski and Grinstein,1997). Several regulatory domains have been identified in mammalian NHEs through which their activation is altered by several interacting factors (Burckhardt et al.,2002).

Molecular characterization of insect NHEs is now beginning. A cDNA with high sequence similarity to NHE3 has been isolated from Aedes aegypti(Table 1; Pullikuth et al.,submitted; Hart et al., 2002). Unlike vertebrate NHE3, Aedes NHE3 possesses a larger cytoplasmic tail, accounting for nearly 56% of the protein. Aedes NHE3 lacks potential N-glycosylation sites but does contain sites for potential cAMP- and cGMP-dependent protein kinase phosphorylation. Protein kinase C and casein kinase II substrate sites are also present in this isoform (Pullikuth et al., submitted). Aedes NHE3 is 53% and 50% similar to Drosophila NHE2 (Giannakou and Dow, 2001) and vertebrate NHE3(Brant et al., 1995),respectively. We also identified NHE homologues in the draft genome sequence of Anopheles gambiae. The genes for A. gambiae NHEs vary significantly in size and exon usage (Table 1). Further searches revealed newer isoforms from Drosophila in addition to the three reported by Giannakou and Dow(2001). High-throughput cDNA sequencing of Drosophila cDNA libraries by Celera Genomics(Rockville, MD, USA) indicated the presence of at least five distinct NHEs. Consistent with this, our annotation of the Anopheles draft genome also uncovered five distinct NHE-encoding genes. It is reasonable to assume that insects might utilize five NHEs to effect cation/proton exchange. However, it should be cautioned that NHE9 and NHE10 are at present only assigned to the NHE family through their relatedness to NHE signature sequence. These isoforms diverge considerably from the rest of the family and are more closely related to bacterial NHEs, suggesting a possible horizontal transfer event in evolution. Furthermore, they have a poorly conserved amiloride-binding pocket, which is unusual among vertebrate members. Alternatively, these isoforms could be expected to be insensitive to pharmacological inhibition by agents conventionally used to distinguish specific isoforms in vertebrates. A detailed knowledge on the expression pattern and functional properties of these forms would be needed to conclusively assign them to the invertebrate family of NHEs. The presence of similar forms in vertebrates indicates the conservation of these molecules and their function in evolution. Immunohistochemical studies of AedesNHE3 indicate that this isoform is predominantly expressed in the basolateral membrane of midgut, Malpighian tubules and gastric ceca (Pullikuth et al.,submitted). Distinct segments of the tubule exhibit apical, intracellular and both basal and apical localization patterns for NHE3. This is consistent with Petzel's suggestion that the basolateral Na⁺/H⁺ exchange resembles the pharmacology of vertebrate NHE3(Petzel, 2000). However,differential localization patterns might indicate its functional diversity at specific segments of the tubule.

In Drosophila, three NHEs (NHE1-3) were identified by Giannakou and Dow (2001). Based on their evolutionary relationship to well-characterized members of the vertebrate family, we propose these Drosophila NHEs be assigned to NHE8, NHE3 and NHE6, respectively (Table 1). In addition, Drosophila cDNAs with high sequence similarity to NHE9 and NHE10 have also been deposited in the database through high-throughput cDNA sequencing efforts. It is important to note that both Drosophila and Anopheles genomes do not contain members that are closely related to mammalian NHE1, NHE2, NHE4, NHE5 and NHE7. The physiological relevance of this finding is not clear, but it might indicate functional redundancy by insect NHEs. Another possibility is that distinct spliced isoforms may exist for insect NHEs, which might fulfill roles not accomplished by the above forms. Drosophila, Caenorhabditis elegansand Aedes NHEs have been shown to exist as spliced variants(Giannakou and Dow, 2001; Hart et al., 2002; Nehrke and Melvin, 2002). At present, the specific roles of these variants are not known. However, it does support the idea that functional diversity could be achieved by molecular variants of a single gene in the NHE family in insects.

Anopheles NHE3 also possesses a large carboxy cytoplasmic region containing several potential regulatory sites. Among the cloned NHEs, insect NHE3s appear to be the largest with about 1200 amino acids. Potential splice variants that lack the majority of the cytoplasmic tail, which contains several potential phosphorylation sites, have been suggested to exist in Aedes NHE3 (Hart et al.,2002). However, the truncated isoform does retain a critical phosphorylation site after the last predicted TM that appears to be analogous to vertebrate Ser605, which is a target for cAMP-dependent inhibition of NHE3(Kurashima et al., 1997). Conversely, in trout, βNHE adjacent serines (Ser659 and Ser664) are involved in cAMP-mediated activation of the exchanger(Malapert et al., 1997). Anopheles NHE3 also contains dual serines at similar positions. It would be interesting to determine the role of these residues in cAMP-dependent processes in mosquito tubule function. As phosphorylation in the carboxy region is one means of regulating the number of exchangers in the plasma membrane, kinasing residues closer to the last TM would be expected to either increase the number of exchangers on the plasma membrane (similar to troutβNHE; Malapert et al.,1997) or increase its endocytosis, thereby inactivating the exchanger (similar to mammalian NHE3; Kurashima et al., 1998). Among the insect NHEs, NHE8 appears to contain a sequence that is predicted to be most sensitive to amiloride and its analogs (A. K. Pullikuth, K. Aimanova, W. Kang'ethe and S. S. Gill, unpublished). These parameters render NHE8 a likely candidate for the amiloride-sensitive exchanger in the apical membrane of mosquito Malpighian tubules. To understand the function of this isoform, we have recently cloned the NHE8 from Aedes and are currently undertaking molecular studies to ratify this prediction (A. K. Pullikuth, K. Aimanova, W. Kang'ethe and S. S. Gill, unpublished).

Sequence similarities and inhibitor sensitivity profiles of vertebrate NHEs have been used to classify members of this family (reviewed in Burckhardt et al., 2002). The relationship of vertebrate and invertebrate NHEs is presented in Fig. 3 based on deduced amino acid sequences. Most vertebrate NHEs fall into a large clade where distinct isoforms group in to separate branches of the phylogenetic tree. The organelle forms, NHE6 and NHE7, form a separate branch that includes AnophelesNHE6 and Drosophila NHE6 (previously assigned DmNHE3 by Giannakou and Dow, 2001). Recently identified NHE8s are distantly related to the rest of the family,whereas the novel forms NHE9 and NHE10 are more closely grouped with bacterial NHEs. NHE9 and NHE10 are only tentatively assigned to the NHE family; their conclusive inclusion as authentic members would require further functional characterization. Moreover, the putative inhibitor-binding region is poorly conserved in NHE9 and NHE10, raising the possibility that their function is insensitive to currently used inhibitors.

Electroneutral transport of Na⁺ or K⁺ coupled with Cl^- can be separated into three distinct subtypes based on ion specificity and pharmacology: Na⁺/K⁺/Cl^-cotransport, K⁺/Cl^- cotransport and Na⁺/Cl^- cotransport. Characterization of these processes in mammals facilitated the isolation of a group of membrane proteins responsible for these activities, collectively named cation-coupled chloride cotransporters (CCCs). Several excellent reviews on characterization of ion cotransport in mammals and molecular and pharmacological studies of CCCs in vertebrates have been published (Haas and Forbush, 2000; Mount et al.,1998).

Sequence and structure analyses of CCCs revealed that they form a protein family that is divergent from other ion transport proteins. Currently, eight members of this family have been identified and functionally characterized in vertebrates. Four isoforms are K⁺/Cl^- transporters(KCC1, KCC2, KCC3 and KCC4), two are Na⁺/K⁺/2Cl^- cotransporters (NKCC1 and NKCC2),one is an Na⁺/Cl^- cotransporter (NCC) and one is a recently identified membrane protein, CIP1 (cotransporter interacting protein 1), which interacts with NKCC but does not transport ions by itself. Theoretical hydrophobicity models and experimental data obtained by antibody accessibility and protease-sensitivity analyses collectively predict that CCCs possess 12 transmembrane segments, which are flanked by hydrophilic N- and C-terminal cytoplasmic domains(Gerelsaikhan and Turner,2000; Moore-Hoon and Turner,1998).

CCCs are subdivided based on their ion specificities and different sensitivities to loop diuretics. KCC isoforms are Na⁺ independent and mediate K⁺ and Cl^- cotransport. These cotransporters can be stimulated by N-ethylmaleimide (NEM) and by the protein kinase inhibitor staurosporine (Bize and Dunham,1994). Although there are no inhibitors that reliably differentiate between Na⁺-independent and Na⁺-dependent K⁺/Cl^- cotransporters, the alkaloid(dihydroindenyl)oxyalkanoic acid is a more potent inhibitor of KCC than of NKCC (Diecke and Beyer-Mears,1997). By contrast, the NCC protein provides K⁺-independent Na⁺/Cl^- cotransport and is sensitive to thiazide diuretics (Costanzo,1985). The two NKCC proteins transport ions in a 1Na⁺:1K⁺: 2Cl^- stoichiometry and are very sensitive to bumetanide and other loop diuretics(Haas and Forbush, 1998). Several reviews are available for detailed information on vertebrate CCCs (for example, see Haas and Forbush,2000; Isenring and Forbush,2001; Mount et al.,1998).

In contrast to the substantial data on vertebrate CCCs, characterization of insect members is limited. Only two insect CCC protein cDNAs have been cloned so far, one in Manduca sexta(Reagan, 1995) and another in A. aegypti (Filippov et al.,2003), but no insect cotransporters have been functionally characterized. Complete sequencing and annotation of two insect genomes, Drosophila melanogaster and Anopheles gambiae, identified five new CCC members in each of these species(Adams et al., 2000; Holt et al., 2002). To evaluate relationships among the newly identified insect members of the CCC family and characterized human cotransporters, we performed a phylogenetic analysis of insect and human CCCs (Fig. 4). Protein sequences of eight known human CCCs were aligned with 12 insect CCCs identified in Drosophila, Manduca, Aedes and Anopheles together with the sequence of human glycine transporter type 2 (Evans et al., 1999),which was used as an outgroup member. This analysis showed that all insect CCC proteins, with the exception of Drosophila CG5594, form their separate branches, suggesting that they maintain higher levels of conservation within Insecta compared with their potential vertebrate orthologues.

The human CCC members form two major branches. One branch contains all human proteins involved in cotransport of Na⁺: NCC and two NKCCs. This branch also contains two clusters of insect proteins: the first one consists of Anopheles agCG57252 and its Drosophila homologue CG4357, while the second clusters the Manduca sexta NKCl cotransporter together with Drosophila CG31547 and two closely related proteins identified in Anopheles (agCG46536 and agCG46505). Clustering of these insect cotransporters in this branch might indicate that they are also sodium dependent in their ion cotransport across the cell membrane.

Another major branch of the phylogenetic tree has four human KCC isoforms and the CIP1 protein. Drosophila CG5594 protein clusters confidently with this branch and is the only insect protein used in the phylogenetic analysis that shows considerable amino acid similarity to human members of the CCC family (approximately 50% to KCC isoforms). The human CIP1 protein does not transport ions itself but interacts with NKCC1(Caron et al., 2000), forming its own cluster within the KCC branch, which is more distant from human KCCs than Drosophila CG5594. Two predicted insect proteins, Drosophila CG10413 and Anopheles agCG54315, fall into this branch. It would be interesting to determine whether these proteins are also unable to transport ions themselves. This will mean that these proteins are true orthologues of human CIP1 and that regulation by CCC heterodimerization is an evolutionarily conserved way of changing kinetic and pharmacological properties of ion cotransport.

Three insect proteins, Drosophila CG12773, Anopheles agCG52356 and Aedes AaeCG12773, form a branch that is equally distant from the two major branches of the tree. Since these CCCs are substantially divergent from mammalian members of the family, it is difficult to predict their cation requirements based on phylogenetic analysis. Uptake experiments using heterologous expression of the Aedes protein, which falls in this cluster, showed that its pharmacological properties are more related to the properties of human KCC isoforms than to other CCC members (V. Filippov and S. S. Gill, unpublished), suggesting that this cluster represents KCC proteins.

Transmembrane topology of insect CCC members predicted by several theoretical methods showed results broadly similar to those found during analysis of vertebrate cotransporters(Gamba et al., 1994; Gerelsaikhan and Turner,2000). As with mammalian CCCs, many prediction programs fail to identify the last four membrane-spanning segments properly in insect CCCs and instead predict two or three transmembrane domains in this region. Experimental data obtained from mammalian cotransporters showed that these CCC members have 12 transmembrane domains and their C- and N- ends are intracellular (Haas and Forbush,2000). Based on comparative analysis of predicted topology of mammalian and insect CCC members we can conclude that the latter also have 12 transmembrane domains. It is also interesting to note that predicted intracellular regions are enriched in phosphorylation sites. In the case of the Aedes cotransporter AaeCG12773, the longest intracellular loop between transmembrane domains 10 and 11 contains eight out of 12 predicted serine phosphorylation sites and four out of seven sites for tyrosine phosphorylation (Filippov et al.,2003).

Phylogenetic and topological analyses of insect CCCs have evidently been very useful for initial assortment of these new members within the family. Being evolutionarily distant from the vertebrate CCC-encoding genes, the insect CCC members preserve the main structural characteristics necessary to execute similar functions, and this is reflected in the phylogenetic analysis. Comparison of insect CCCs from four different species (Drosophila,Anopheles, Aedes and Manduca) also revealed that they form separate clusters consisting of closely related members. Protein similarities within these clusters are considerably high (>50%; Table 2). The main tendency is that each mosquito protein has its own closely related homologue in Drosophila, and similarity in the primary sequences is also reflected in nearly identical predicted topology. There are two exceptions to this rule. First, as shown in Fig. 4, the Drosophila CG5594 does not have a similar Anophelescounterpart in the tree. However, the absence of a similar Anophelesgene is probably because the Anopheles genome is apparently not completely sequenced and annotated. Second, the Anopheles genome contains two similar annotated genes agCG46536 and agCG46505, whereas Drosophila has only one gene that falls in the cluster with them. Similarly, the presence of these two genes in the current annotation of the Anopheles genome could be due to heterogeneity in the Anopheles genome sequenced. Further analysis will be needed to confirm these initial results.

Studies of expression profiles and localization of CCCs in tissues proved to be useful for determination of physiological roles of these cotransporters in ion transport (Mount et al.,1998; Mount and Gamba,2001). Sequence data now available for insect CCC members allow the initiation of these studies in mosquitoes. Using RT-PCR, the expression profile of the first cloned mosquito cotransporter gene, AaeCG12773,has been obtained (Filippov et al.,2003). It was found that mRNA levels of this gene are most abundant in the midgut and are also present in the hindgut, both in larvae and adult females. No transcription of this gene was found in Malpighian tubules(Filippov et al., 2003). Availability of antibodies against the AaeCG12773 protein allowed us not only to confirm data obtained by RT-PCR but also to determine the subcellular localization of the cotransporter. Immunostaining revealed that the AaeCG12773 protein is indeed predominantly localized in the upper part of the gut and its expression is higher in adults compared with the larval stages(Fig. 5). No specific immunoreactivity above background was found in the Malpighian tubules. Immunostaining also revealed that the protein is concentrated mainly in the basolateral membranes.

Table 2 summarizes the available data so far on insect CCC members. They are arranged, based on sequence similarity, in five groups. It is not obligatory that CCCs from different species within the group are true orthologues, meaning that they have identical physiological roles. However, it is likely that they do share common ion preferences as well as pharmacological properties. Drosophila is currently the only insect species in which transcription profiles of all five CCC members are known. The mRNA levels of these CCC-encoding genes differ considerably. One of these genes, CG31547, showed such low levels of expression at all developmental stages and in all tissues analyzed that its RT-PCR-specific band could not be identified even after 70 cycles of amplification. It is interesting to note that the Manduca cotransporter, which is in the same cluster as the CG31547 gene, was found to be abundant in the Malpighian tubules(Reagan, 1995). Other Drosophila genes showed much higher expression levels. One important finding is that high mRNA levels of four CCC-encoding genes were found during embryonic development, including early embryos (0-4 h after deposition). Expression of cotransporters at this stage definitely indicates the importance of this type of ion transport for proliferation and/or differentiation. Three CCC members, CG12773, CG4357 and CG5594, showed high levels of expression in the larval midgut, suggesting that they are engaged in regulation of ion transport during digestion. CG10413 showed high expression levels in the larval brain; however, its expression dropped substantially in adult heads. In contrast to CG10413, the CG5594 gene, which has high similarity to human KCC isoforms, is not transcribed in larval brains but its mRNA is abundant in adult heads. Analysis of gene expression profiles of Drosophila CCCs also showed that,compared with their mammalian homologues, none of them has high levels of expression in Malpighian tubules.

Current models for ion transport in mosquito Malpighian tubules are derived from experiments on whole-tubule assays using electrophysiological properties and sensitivity to pharmacologically defined reagents. These models are useful in providing a broad framework for understanding Malpighian tubule function. However, a deeper understanding of expression pattern of candidate transporters along the length of the tubule will be necessary to clearly delineate the roles and participation of these proteins in mosquito tubule function. Presently, considerable attention and emphasis is placed on the principal cell and its role in fluid secretion. Emerging evidence indicates paracellular movement of chloride ions, possibly through stellate cells, to be operating in parallel to transcellular ion and fluid transport(Beyenbach, 1995; Pannabecker, 1995; Pannabecker et al., 1993). Understanding the role of ion transporters in homeostasis would also require a better understanding of factors that interact with the transporters to activate or deactivate them by trafficking to or occluding from membrane domains relevant for their active states. Obviously, there appear to be several parallels in the manner with which vertebrate transporters can be activated or their roles modulated to those circumstances in mosquito ion regulation. However, without definite molecular characterization of these processes, these comparative aspects would remain largely correlative or merely speculative at best.

A major difference among vertebrate and invertebrate NHE function lies in the substrate stoichiometry. Mammalian NHEs mediate the exchange of one proton for each cation translocated and thus are electroneutral and membrane potential insensitive (Aronson,1989; Demaurex et al.,1995; Post and Dawson,1994). Interestingly, prokaryotes exhibit an electrogenic transport of 1Na⁺/2H⁺(Padan and Schuldiner, 1994),whereas crustacean transporters operate with a stoichiometry of 2Na⁺/1H⁺ (Ahearn et al., 1994). Due to the occurrence of two cation-binding sites,crustacean exchangers are capable of transporting divalent cations(Ca²⁺, Zn²⁺ and Cd²⁺) in exchange for protons that may be required for sequestration and detoxification of heavy metals in the hepatopancreas (Ahearn et al.,2001). Although electroneutral Na⁺/H⁺exchange in invertebrates has been reported occasionally(Deitmer and Schlue, 1987; Schlue and Thomas, 1985),electrogenic exchange of cations for protons is more widely distributed among various groups of invertebrates (Grinstein and Wieczorek, 1994). Unlike vertebrate cells, where the surrounding pH, ionic composition and buffering capacity do not vary dramatically, bacteria and invertebrates experience wide ranges of pH and ionic composition. In such cases, the chemical component itself would not be sufficient for cellular homeostasis by driving electroneutral exchange, thus the electrogenicity of cation/proton exchange could be more versatile in being able to maintain the high luminal pH in the lepidopteran midgut (through K⁺/2H⁺ exchange; Lepier et al., 1994) and in the acidification of the gastric lumen in crustaceans (through 2Na⁺/1H⁺; Ahearn et al., 1990). The prevalence of electrogenic cation/proton exchange in invertebrates has thus been viewed as an ancestral mechanism, whereas the electroneutral mammalian exchange is considered an evolutionary adaptation(Ahearn et al., 2001; Grinstein and Wieczorek,1994). In spite of these variations, Aedes aegyptiMalpighian tubules exhibit electroneutral exchange of 1K⁺/1H⁺ or 1Na⁺/1H⁺(Weng et al., 2003). The tubule lumen is near neutral and thus would not require the electrogenic apical cation/H⁺ exchange that has been implicated in maintaining the alkaline lepidopteran or mosquito midgut lumen or the acidic lumen of crustaceans. However, the electrogenic properties of this exchange under stimulated conditions should be understood since cytoplasmic ionic concentrations vary rapidly in blood-fed insects, which could conceivably change the dynamics of apical transport.

This review focused on two classes of proteins that are potentially involved in regulating ion and fluid secretion in the mosquito midgut and Malpighian tubules. Molecular characterization of these proteins and the manner in which their activities are regulated are crucial aspects to gain deeper understanding of ion transport in the insect midgut and Malpighian tubules. The field appears favorably poised to exploit the emerging genomic information to expedite the molecular identification and thorough characterization of important factors regulating ionic homeostasis in mosquitoes. A thorough understanding of ion transport mechanisms in mosquitoes would not only be useful for appreciating the biology but would also lead to the identification and development of reagents that can potentially interfere with these crucial processes, leading to novel means of disrupting disease transmission by mosquito vectors. Combined with the rapidly progressing mosquito transgenic technology, a combinatorial approach to inactivate or manipulate genes with deleterious fate in mosquitoes should be feasible in the future.

Research in our laboratories is funded through grants from the National Institutes of Health (AI 32572 and 48049). We thank Dr William Harvey for useful discussion on historical perspectives on the identification of the ion pump and Dr Karlygash Aimanova for images on CCC localization.