Cloning and Characterization of a Potassium-Dependent Sodium/Calcium Exchanger in Drosophila

A cDNA clone was constructed, corresponding to the bovine rod photoreceptor NCKX1 that encoded a fusion between the highly conserved TM1–TM5 and TM6–TM11 regions. The clone was ³²P-radiolabeled using the Prime-It II Random Primer Labeling Kit (Stratagene). We used this ³²P-radiolabeled DNA probe to screen a Drosophila genomic library (a gift of K. Moses, Emory University, Atlanta, GA, and G.M. Rubin, University of California, Berkeley, Berkeley, CA). The genomic library was screened according to the methods described in Sambrook et al. 1989. We obtained several genomic clones. A 1.3-kb EcoRI and XbaI genomic fragment was subcloned into pBluescriptII KS (Stratagene), ³²P-radiolabeled, and used to screen a Drosophila eye-enriched lambda ZAPII cDNA library (gift of C.S. Zuker, University of California, San Diego, San Diego, CA) (Shieh et al. 1989). A set of overlapping cDNAs encompassing the entire Nckx30C transcript was isolated and sequenced. No other cDNAs were identified. The full-length Nckx30C cDNA was constructed and the DNA sequence was determined by fluorescent-based sequencing methods (Applied Biosystems). The BLAST search program (Altschul et al. 1990) (http://www.ncbi.nlm.nih.gov/BLAST) and the CLUSTAL W multiple sequence alignment program were used for sequence similarity analysis (Thompson et al. 1994) (http://pbil.ibcp.fr/NPSA/npsa_clustalw.html). The sequence data have been submitted to the DDBJ/EMBL/GenBank databases under accession number AF190455.

The full-length Nckx30C cDNA was cloned into the novel insect cell vector pEIA as an EcoRI fragment, and this was used for the stable transfection of High Five insect cells as described (Farrell et al. 1998). High Five cells (BTI-TN-5B1-4) derived from Trichoplusia ni egg cell homogenates were purchased from Invitrogen. These cells do not display endogenous exchanger activity.

Potassium-dependent sodium/calcium exchange activity was measured in High Five cells transformed with Nckx30C cDNA. The cells were loaded with sodium via the sodium-potassium ionophore, monensin, in a medium containing high sodium, 150 mM NaCl, 80 mM sucrose, 0.05 mM EDTA, and 20 mM Hepes, pH 7.4, according to the methods described for rod outer segments (Schnetkamp et al. 1995). The ionophore was removed and the sodium-loaded cells were washed with and resuspended in 150 mM LiCl, 80 mM sucrose, 0.05 mM EDTA, and 20 mM Hepes, pH 7.4. The cells were diluted 10-fold in media containing 80 mM sucrose, 20 mM Hepes, pH 7.4, and either 150 mM KCl, 150 mM NaCl, or 150 mM LiCl. ⁴⁵Ca uptake was initiated by addition of 35 μM CaCl₂ and 1 μCi ⁴⁵Ca. ⁴⁵Ca uptake in High Five cells was measured with a rapid filtration method with the use of borosilicate glass fibers over a time-course of 5 min as described previously (Schnetkamp et al. 1991b). The ice-cold washing medium contained 140 mM KCl, 80 mM sucrose, 5 mM MgCl ₂, 1 mM EGTA, and 20 mM Hepes, pH 7.4.

Total RNA was prepared from the heads and bodies of 0–7-d-old Drosophila w1118 and eya lines using the Ultraspec RNA isolation system (Biotecx). PolyA⁺ RNA was isolated by affinity chromatography on oligo(dT) cellulose columns using the FastTrack 2.0 System (Invitrogen, Inc.). PolyA⁺ RNA from third instar larvae and 0–24-h embryos (Canton S strain) was purchased from Clontech. PolyA⁺ RNA (10 μg) from each sample was run on a denaturing 0.9% agarose gel for 3 h 130V/cm. The gel was stained with ethidium bromide and photographed on a UV transilluminator. The mRNA was transferred overnight by capillary action in 20× saline sodium citrate (SSC) to a positively charged nylon membrane (Nytran Plus), and fixed by UV cross-linking. The membrane was probed with [³²P]UTP antisense riboprobes (Maxiscript In Vitro Transcription kit; Ambion, Inc.). Four subclones of Nckx30C were constructed and used for riboprobe production; 1.3 kb BamHI-BamHI (5′ end), 500 bp BamHI-ClaI (TM1–TM5), ClaI-ClaI (loop region), and ClaI-EcoRV (TM6–TM11) (see Fig. 4 A). The antisense and sense riboprobes were incubated with the membranes using the NorthernMax hybridization buffer (Ambion, Inc.), the membranes were washed with 0.2× SSC, 1% SDS at 65°C, and exposed to Kodak X-OMAT X-ray film (Eastman Kodak). All four antisense probes displayed the same labeling pattern. No signal was detected with the sense probes. A 250-bp mouse β-actin antisense riboprobe was used as a control for the amount of RNA loaded (Maxiscript In Vitro Transcription kit; Ambion, Inc.).

Polytene chromosome squashes (w m f strain) were prepared as described (Lim 1993). A 1.3-kb EcoRI and XbaI genomic fragment was nick-translated using biotin-16-UTP (Boehringer Mannheim Corp.). Hybridization with the biotinylated probe was carried out in 600 mM sodium chloride, 1× Denhardt's, 50 mM magnesium chloride, and 50 mM sodium phosphate, pH 6.8, at 37°C overnight. Samples were washed with 2× SSC, 0.1% Triton X-100, and PBS, pH 7.2 (Lim 1993). Hybrids were detected using Vectastain ABC (Vector Laboratories, Inc.). Labeling was identified at a single cytological position on chromosome 2L, 30C5-7.

Heads from the w1118 strain of flies were embedded in Tissue-Tek OCT Compound (Miles, Inc.) and frozen on dry ice. Eyes were fixed in 4% paraformaldehyde, infiltrated with 2.3 M sucrose, and frozen in liquid nitrogen according to methods described previously (Colley et al. 1991). In situ hybridizations were carried out on 14-μm head sections and 0.5-μm eye sections in 50% formamide, 5× SSC, 0.5 mg/ml sheared herring sperm DNA, 0.05 mg/ml heparin, 0.25% CHAPS, 0.1% Tween 20, 1 mg/ml yeast torula RNA, 1× Denhardt's, at 55°C overnight. Embryo and imaginal disc in situ hybridizations were carried out essentially as described in Panganiban et al. 1994. In brief, embryos (Canton S strain) were collected, dechorionated, fixed, and processed for in situ hybridization as described by Panganiban et al. 1994. Proteinase K was not used. Embryos were hybridized in 50% formamide, 5× SSC, 500 μg/ml sheared salmon sperm DNA, 0.1% Tween 20, 1 mg/ml glycogen. Hybridization was carried out at 55°C for 24–30 h. Probe concentrations were 90 ng/300 μl. The larval imaginal discs were dissected from crawling third-instar larvae, but left attached to the cuticle. The discs were fixed, permeabilized with proteinase K at 25 μg/ml for 3 min at room temperature, and postfixed. Probe concentrations were 120 ng/300 μl.

Digoxigenin-labeled sense and antisense riboprobes were generated by in vitro transciption of five DNA templates as recommended by the digoxigenin/UTP supplier (Boehringer Mannheim Corp.). Five Nckx30C cDNA probes were used: 1.3 kb BamHI-BamHI (5′ end), 500 bp BamHI-ClaI (TM1–TM5), ClaI-ClaI (loop region), ClaI-EcoRV (TM6–TM11), and the full-length Nckx30C cDNA clone (see Fig. 4 A). Calx cDNA was provided by E. Schwarz (Columbia University, New York, NY) (Schwarz and Benzer 1997). Chaoptin cDNA was provided by D. Van Vactor (University of California, Berkeley) and S.L. Zipursky (University of California, Los Angeles, Los Angeles, CA) (Reinke et al. 1988). After transcription, the reaction mix was treated with DNase and the riboprobes were hydrolyzed in carbonate buffer for 2 min at 80°C. All riboprobes were quantified by analysis in denaturing 0.8% agarose gels and by dot blot analysis. All detection steps were as described in Tautz and Pfeifle 1989. All Nckx30C antisense probes showed the same labeling pattern. No signal was detected with the sense probes.

The NCKX-type exchangers are a novel group of calcium extrusion proteins. NCKX1 function was initially described in the outer segments of retinal rod photoreceptor cells (Fig. 1) (Cervetto et al. 1989; Schnetkamp et al. 1989). Antibodies directed to NCKX1 were used to screen a bovine retinal expression library, and NCKX1 cDNA was cloned and sequenced (Reiländer et al. 1992). A second isoform, NCKX2, was cloned from rat brain, and was shown on the basis of in situ hybridization to be widely expressed in various regions of the brain (Tsoi et al. 1998). To isolate Drosophila Nckx30C, we constructed a cDNA probe from the bovine rod photoreceptor NCKX1 that encoded a fusion between the highly conserved TM1–TM5 and TM6–TM11 regions. We used this probe to screen the Drosophila genomic library and obtained a 1.3-kbp EcoRI-XbaI fragment that corresponds to an EcoRI site at the 5′ end of the gene. The XbaI site is located in the intron. This 1.3-kbp genomic fragment was used to screen the cDNA library. A set of overlapping cDNAs encompassing the entire Nckx30C transcript was isolated and sequenced. The composite cDNA has a single open reading frame that encodes a protein of 856 amino acids (Fig. 2). We compared the derived amino acid sequence of Nckx30C with the human NCKX1 and rat NCKX2 (Fig. 3). There is 66 and 71% identity in two clusters of amino acids that corresponds to the predicted TM domains of human rod photoreceptor cell NCKX1 and rat brain NCKX2, respectively. A partial restriction map for Nckx30C is shown in Fig. 4 A. The hydropathy analysis confirms that Drosophila NCKX contains 11 hydrophobic regions that correspond to the 11 predicted TM helices in NCXK1 and NCKX2 (Fig. 4 B). As with NCKX1 and NCKX2, there is a large cytoplasmic loop located between TM5 and TM6 in Nckx30C. This large cytoplasmic loop displays very little amino acid identity with the corresponding cytoplasmic loops of NCKX1 or NCKX2 (Fig. 3). The number of amino acids that make up the hydrophilic loops and the two sets of membrane-spanning segments are shown in Fig. 4 C. One feature distinguishes NCKX30C from other members of the NCKX superfamily. Hydropathy analysis revealed that NH₂-terminal to TM1 is a region that may contain two additional membrane-spanning segments not observed in either NCKX1 or NCKX2. The correct identity of this NH₂-terminal sequence for Nckx30C was confirmed by 5′ rapid amplification of cDNA ends (RACE) starting with a primer corresponding to sequence of TM1, and by PCR using first-strand cDNA.

We cytologically mapped Nckx to 2L at 30C5-7 by in situ hybridization to polytene chromosomes, and based on the chromosomal location have named the gene Nckx30C.

We examined the functional properties of NCKX30C in High Five cells. The strategy for these experiments takes advantage of properties of both NCX and NCKX. First, NCX and NCKX are bidirectional in their ability to mediate both calcium efflux (forward exchange) and calcium influx (reverse exchange). The direction of transport is dictated by the direction of the TM sodium gradient. Normally, the inward sodium gradient removes calcium from the cell (forward exchange). However, when external sodium is removed, the outward sodium gradient drives calcium into the cell (reverse exchange). Therefore, we measured the properties of reverse exchange of NCKX30C by measuring ⁴⁵Ca uptake in sodium-loaded cells. We tested for NCKX activity by using three different manipulations of the cation gradient that are known to prevent NCKX activity. The cells were loaded with sodium via the sodium-potassium ionophore, monensin, in a medium containing high sodium, according to the methods described for rod outer segments (Schnetkamp et al. 1995). The ionophore was removed and the sodium-loaded cells were washed with and resuspended in low sodium buffer as described in the Materials and Methods. The cells were diluted 10-fold in media containing 80 mM sucrose, and 20 mM Hepes, pH 7.4, and either 150 mM KCl, 150 mM NaCl, or 150 mM LiCl. ⁴⁵Ca uptake was initiated by addition of 35 μM CaCl₂ and 1 μCi ⁴⁵Ca. The first condition causes the release of internal sodium by the action of monensin, a sodium(-potassium) selective ionophore. Fig. 5 A shows that ⁴⁵Ca uptake by NCKX30C was strongly inhibited by the addition of monensin, thus demonstrating a requirement for intracellular sodium for calcium transport.

A second property of both NCK and NCKX is that calcium uptake by reverse exchange is inhibited by high extracellular sodium (Schnetkamp et al. 1991a). This is thought to occur because sodium and calcium compete for a common binding site (Schnetkamp et al. 1991a). We demonstrate that ⁴⁵Ca uptake by NCKX30C is strongly inhibited by high extracellular sodium, consistent with its identity as an NCX (Fig. 5 B).

A critical distinction between NCKX and NCX is that calcium influx via NCKX requires the presence of potassium, and lithium cannot substitute for potassium (Tsoi et al. 1998). In contrast, calcium influx via reverse exchange in NCX is the same in media containing potassium or lithium (Tsoi et al. 1998; Cooper et al. 1999). ⁴⁵Ca uptake by NCKX30C requires potassium and was not observed in media containing lithium (Fig. 5 C). This result demonstrates that NCKX30C functions as an NCKX.

As negative controls, we subjected nontransformed High Five cells to the same ⁴⁵Ca uptake experimental protocols described above: ⁴⁵Ca uptake in media containing potassium was indistinguishable from that in media containing either sodium or lithium, showing that there is no endogenous NCX or NCKX activity in the cells (data not shown). The above results demonstrate that NCKX30C mediates potassium-dependent sodium/calcium exchange similar to that observed for NCKX1 and NCKX2 (Tsoi et al. 1998; Cooper et al. 1999).

Northern blot analysis revealed an 8-kb transcript that is expressed in embryos and larvae (Fig. 6). In the adult, two transcripts of 8 kb and 10 kb were detected. Expression was detected predominately in the head. Transcripts present in the body may represent the thoracic ganglia or other components of the nervous system (Demerec 1994). The 8-kb and 10-kb transcripts were also detected in eya-1 mutant flies, which lack eyes, indicating that expression is not restricted to the eyes (Fig. 6).

Previous work has identified a Drosophila NCX, Calx, which encodes a functional ortholog of the NCX-type exchangers (Hryshko et al. 1996; Ruknudin et al. 1997; Schwarz and Benzer 1997). Expression patterns for Calx were reported in Schwarz and Benzer 1997. In situ hybridization analysis reveals that both Nckx30C and Calx are expressed in the adult nervous system. Fig. 7 A shows that Nckx30C is expressed in the photoreceptor cells as well as in the lamina, medulla, and optic lobes of the brain. Fig. 7 C shows that Calx is also expressed in the photoreceptors as well as in the brain. Analysis of many images did not reveal a convincing difference in expression patterns between Nckx30C and Calx. In situ analysis at higher resolution, on 0.5-μm cryosections, confirmed that expression is detected in the photoreceptor cells (data not shown). However, this does not exclude the possibility that the subcellular distribution of the proteins could be very different. No signal was detected with the sense probes (Fig. 7B and Fig. D).

Beginning at embryonic stage 13–14, Nckx30C transcripts were detected in a discrete pattern of unpaired cells at the ventral midline of the central nervous system (Fig. 8 E). Nckx30C transcripts were not detected in the preblastoderm or blastoderm stages (Fig. 8A and Fig. C). Unlike Nckx30C, Calx expression is robust in the preblastoderm before and during nuclear migration from the interior to the cortex (Fig. 8 B). Since this is before activation of zygotic transcription, Calx transcripts are likely to be maternally inherited. Calx transcripts disappear in the cellular blastoderm (Fig. 8 D) and then reappear in the stage 11 embryo (Fig. 8 F). By stage 12, Calx expression is noted in two cells per hemisegment, one on either side of the ventral midline (Fig. 8 H). In embryonic stage 15, Nckx30C expression was detected in several neurons within the ventral nerve cord (Fig. 8 G). By stage 16, when many neurons within the ventral nerve cord and the embryonic brain express Nckx30C (Fig. 8 I), Calx expression is restricted to a much smaller subset of neurons in the ventral nerve cord (Fig. 8 J). Outside the central nervous system, Nckx30C and Calx transcripts were observed in some cells, which may represent parts of the peripheral nervous system (Fig. 8I and Fig. J).

Drosophila appendages develop from imaginal discs in a highly orchestrated series of events that divide the discs into distinct subregions during patterning. Retinal differentiation begins at the posterior end of the eye disc, which coincides with the dorsal–ventral midline (Treisman and Rubin 1995; Heberlein et al. 1998; Treisman and Heberlein 1998), and proceeds as a wave across the eye disc from posterior to anterior (Ready et al. 1976; Tomlinson and Ready 1987; Tomlinson 1988). The morphogenetic furrow is a dorsoventral indentation in the eye disc, with the region anterior to the furrow comprising actively dividing and unpatterned cells, and the region posterior to the furrow containing differentiating photoreceptor cells assembling into ommatidia. (Tomlinson 1988; Ready 1989; Banerjee and Zipursky 1990). Fig. 9 A shows expression of chaoptin, which encodes a photoreceptor cell surface glycoprotein (Reinke et al. 1988), in photoreceptor cells located posterior to the furrow. Calx expression was not detected in any of the imaginal discs. Nckx30C transcripts were detected both anterior and posterior to the morphogenetic furrow in a dorsal–ventral pattern with no labeling in the midline (Fig. 9B and Fig. C). This expression pattern is similar to but broader than that observed for wingless (wg), which is known to play an important role in dorsal–ventral patterning in the eye disc (Heberlein and Moses 1995; Ma and Moses 1995; Treisman and Rubin 1995; Heberlein et al. 1998; Treisman and Heberlein 1998). In vertebrates, one of the two apparently distinct Wnt pathways may act through a G protein–mediated phosphatidylinositol signaling cascade that results in an increase in intracellular calcium via inositol 1,4,5 trisphosphate (InsP₃)-sensitive stores (Moon et al. 1997; Slusarski et al. 1997a,Slusarski et al. 1997b). It is intriguing to hypothesize that NCKX30C may be playing a role in modulating calcium in this patterning pathway.

Nckx30C is expressed strongly throughout most of the wing disc and throughout most of the haltere disc (Fig. 9, D–F). In the leg disc, Nckx30C is expressed in a ventral wedge in the cells that will give rise to the ventral and/or proximal structures (Fig. 9 G). This pattern of expression is similar to that observed for wingless in the leg discs (Baker 1988; Couso et al. 1993; Theisen et al. 1994). Given what we know about the role of NCKX in retinal rod and cone photoreceptors, it is tempting to speculate that cells that express Nckx30C may be experiencing large and sustained rises in cytosolic calcium.

We have cloned Nckx30C from Drosophila and have shown that it is expressed in the adult nervous system and also during development in the embryo and the imaginal discs. Heterologous expression in cultured insect High Five cells demonstrates that NCKX30C functions as an NCKX. Calx, a member of the NCX gene family, is distantly related to Nckx30C (Hryshko et al. 1996; Ruknudin et al. 1997; Schwarz and Benzer 1997; Dyck et al. 1998; Omelchenko et al. 1998). Both NCX and NCKX function in sodium–calcium exchange in cells experiencing large calcium fluxes. In addition, both Nckx30C and Calx are expressed in Drosophila photoreceptors as well as in other cells, indicating that there may be multiple mechanisms for calcium efflux in these cells.

A puzzling finding is that the photoreceptors in Drosophila express both NCKX30C and Calx exchangers. The NCKX exchangers have several unique features that make them well-suited for calcium extrusion during phototransduction. Illumination of Drosophila photoreceptors activates the phototransduction cascade leading to the opening of the cation-selective channels and an increase in the intracellular calcium (Fig. 1). In both flies and vertebrate photoreceptor cells, sodium influx contributes substantially to the inward current, potentially leading to elevated [Na⁺]_inside (Coles and Orkand 1985; Nakatani and Yau 1988b; Hardie 1996b; Gerster 1997). As the cytosolic sodium concentration increases and reduces the TM sodium gradient, NCX exchangers, like Calx, are expected to reverse direction at much lower cytosolic sodium concentrations when compared with the potassium-dependent NCKX exchangers (Schnetkamp et al. 1991c). Therefore, the potassium-dependent exchangers will be better able to function in phototransduction under these conditions. The apparent redundancy could also be explained if the subcellular distribution of NCKX30C and Calx are very different, with each fulfilling distinct functions. A combination of immunocytochemistry with mutant analysis will assist in resolving these questions.

To date, the central focus of efforts in development of Drosophila has been on the identification of regulatory genes that control development, with the potential contribution of calcium signaling remaining largely unexplored. Notable exceptions include systems responsible for dorsal–ventral positional information, such as the dorsal protein in dorsal–ventral pattern information in the early embryo and the wingless pathway. It has been demonstrated that increased calcium can act as a second messenger in the signal transduction pathway leading to the nuclear localization of dorsal, a maternally inherited transcription factor (Kubota et al. 1993). Another signaling pathway that may utilize calcium is wingless. wingless, a member of the Wnt gene family, encodes a secreted glycoprotein that is involved in a complex variety of signaling events (Nusse and Varmus 1992; Cadigan and Nusse 1997). In vertebrate embryos, one of the two apparently distinct Wnt pathways can act through a G protein–mediated phosphatidylinositol signaling cascade, resulting in a release of intracellular calcium from inositol 1,4,5 trisphosphate (InsP₃ )-sensitive stores (Moon et al. 1997; Slusarski et al. 1997a,Slusarski et al. 1997b). Here, we demonstrate that Nckx30C is expressed in a pattern similar to, but broader than that observed for wingless. Nckx30C expression would indicate that, most likely, a substantial calcium flux is occurring in these cells. In this study, we show that exchangers are expressed during development, indicating that they may not only function in the removal of calcium and maintenance of calcium homeostasis during signaling in the adult, but may also play critical roles in signaling events during embryogenesis and patterning of imaginal discs. The isolation of Nckx30C in Drosophila permits a genetic analysis of the in vivo role of calcium in modulating signaling pathways in Drosophila.

We thank P.J. Bauer, D. Bownds, C. Doe, V.E. Foe, G. Halder, R. Hardie, B. Minke, K. Moses, R. Moon, G. Panganiban, R. Payne, A. Polans, P. Robinson, E. Schwarz, and C.S. Zuker for valuable discussions and comments on the manuscript. We also thank E. Corse, T. Hoening, T. James, S. Park, and J. Riley for their technical assistance; and R.A. Kreber and J. Lim for expert assistance with the chromosome in situs. E. Schwarz and S. Benzer generously provided Calx and unpublished data. We are grateful to E. Nielsen and J. Loeffelholz for their excellent assistance with the computer graphics. S. Carroll and S. Paddock generously provided their expertise and microscopes.

This work was supported by the Medical Research Council of Canada (P.P.M. Schnetkamp), and National Institutes of Health grant EY08768, Retina Research Foundation, Howard Hughes Medical Institute, Foundation Fighting Blindness, Fight-For-Sight, and the Research to Prevent Blindness Foundation (N.J. Colley).