INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ION CHANNELS PLAY a crucial role in many diseases, most notably in cystic fibrosis, where a genetic defect of the cystic fibrosis transmembrane conductance regulator (CFTR) is responsible for the disturbed ion transport (3, 19, 23, 25, 26, 30). CFTR is a multifunctional transport protein that functions not only as an epithelial Cl channel but also as a regulator of other ion channels and cellular pathways (10, 29, 31). Several studies have indicated that, in addition to CFTR, a Ca²⁺-activated secretory pathway for Cl may play an important role in modulating the disease severity in various tissues of cystic fibrosis patients and CFTR knockout mice (1, 4, 18, 20, 27, 33, 35). However, little is known about the molecular basis of the channels involved.

A new family of proteins has recently been discovered that mediate a Ca²⁺-activated Cl conductance in a variety of tissues. Four members of this family have been identified, including bovine lung endothelial cell adhesion molecule-1 (Lu-ECAM-1) (8, 40), bovine Ca²⁺-activated Cl channel (CaCC or bCLCA1) (6), murine CLCA1 (mCLCA1) (9), and human CLCA1 (hCLCA1) (11). Patch-clamp studies with transfected human embryonic kidney (HEK-293) cells have shown that bCLCA1, mCLCA1, and hCLCA1 mediate a Ca²⁺-activated Cl conductance that can be inhibited by the anion channel blocker DIDS and the reducing agent dithiothreitol (DTT). The protein size, structure, and processing seem to be similar among different CLCA family members and have been studied in most detail for Lu-ECAM-1 (8). The Lu-ECAM-1 open reading frame (ORF) encodes a precursor glycoprotein of 130 kDa that is processed to a 90-kDa amino-terminal cleavage product and a group of 30- to 40-kDa glycoproteins that are glycosylation variants of a single polypeptide derived from its carboxy terminus. Both subunits are associated with the outer cell surface, but only the 90-kDa subunit is thought to be anchored to the cell membrane via four transmembrane domains (8). Based on hydrophobicity analyses, analogous structural models have been suggested for mCLCA1 and hCLCA1 (9, 11). Although the protein processing and function appear to be conserved among CLCA homologs, significant differences exist in their tissue expression patterns. For example, bovine Lu-ECAM-1 is expressed primarily in vascular endothelia (40), bCLCA1 is exclusively detected in the trachea (6), and hCLCA1 is selectively expressed in a subset of human intestinal epithelial cells (11). Thus the emerging picture is that of a multigene family with members that are highly tissue specific, similar to the ClC family of voltage-gated Cl channels (14). A role for CLCA homologs in the complex ion-trafficking disorder of cystic fibrosis has been speculated, based on observations that the cellular expression patterns of bCLCA1, mCLCA1, and hCLCA1 overlap with that of CFTR in the respective tissues (6, 9, 11, 12). Before now no Ca²⁺-activated Cl channel had been cloned from human lung, the most severely affected organ in cystic fibrosis.

Here we describe the cloning of hCLCA2, the second hCLCA family member, and provide a detailed account of the membrane topology of this new Cl channel, using glycosylation site scanning and protease protection assays (28, 38). hCLCA2 is selectively expressed in lung, trachea, and mammary gland. Its transient expression in HEK-293 cells reveals a Ca²⁺-activated Cl conductance, which is similar to that of previously cloned CLCA family members.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Isolation and cloning of the hCLCA2 cDNA. A human lung cDNA library (Clontech) was screened as described (11) using [-³²P]dCTP nick-labeled Lu-ECAM-1 cDNA as probe. For hybridization and washing, low-stringency conditions were applied [2 washes with 2× standard sodium citrate (SSC), 0.1% SDS at 55°C for 20 min, followed by 2 washes with 0.4× SSC, 0.1% SDS at 40°C for 10 min]. Positive phage clones were amplified, cloned into pBluescript (Stratagene), and sequenced. Automated sequencing with initial plasmid-derived primers followed by internal gene-specific primers was performed by the Cornell University DNA Sequencing Facility using dRhodamine terminator cycle sequencing on an ABI prism 377 DNA sequencer (PE Applied Biosystems). Missing 5' and 3' ends of the isolated cDNA species were completed using the rapid amplification of cDNA ends technique (RACE; Life Technologies). For expression studies, the 2,832-bp hCLCA2 ORF was PCR amplified from human trachea poly(A)⁺ RNA (Clontech) following reverse transcription with Superscript RNase H⁺ RT (Life Technologies) and random hexamer priming. PCR was performed with Pwo DNA polymerase [Boehringer; initial denaturation at 94°C for 3 min, 35 cycles at 94°C for 50 s, 58°C for 30 s, and 72°C for 2 min, with a time increment of 3 s/cycle for each extension step (72°C), followed by a final extension step of 72°C for 8 min]. Primer sequences were 5'-TACAACATGACCCAAAGGAGC-3' (upstream) and 5'-GACACTTTGGATATTTATTTATAATAATTTTGTTC-3 (downstream), with Not I linkers underlined. PCR products were gel purified, incubated with Not I, and cloned into the expression vector pcDNA3.1 (Invitrogen). Four different full-length PCR products were sequenced to control for potential PCR-induced sequence errors.

Northern blot and RT-PCR analyses. Human multiple tissue Northern blots (Clontech) contained 2 µg poly(A)⁺ RNA per lane of heart, brain, placenta, lung, liver, skeletal muscle, kidney, pancreas, spleen, thymus, prostate, testis, ovary, small intestine, colon mucosa, peripheral blood leukocytes, stomach, thyroid, spinal cord, lymph node, trachea, adrenal gland, or bone marrow. In addition, total RNA was isolated (Trizol method, Life Technologies) from MCF-10A cells at 80% confluency (American Type Culture Collection). MCF-10A total RNA (20 µg/lane) and human mammary gland poly(A)⁺ RNA (2 µg/lane, Clontech) were resolved on a formaldehyde gel, blotted onto nitrocellulose, and hybridized with the [-³²P]dCTP nick-labeled (RTS RadPrime, Life Technologies) hCLCA2 cDNA as described (11). For the study on the hCLCA2 expression pattern, highly stringent washing conditions were employed (2 washes with 2× SSC, 0.1% SDS at 55°C for 20 min followed by 2 washes with 0.1× SSC, 0.1% SDS at 65°C for 20 min). Autoradiographs were exposed using an intensifying screen at 70°C for up to 8 days. Stripping of the blots and rehybridization with a probe for the housekeeping gene elongation factor-1 (EF-1) were performed to control for RNA quality and loading amounts as described previously (13). RT-PCR was performed using the above-mentioned conditions and primers to detect hCLCA2 expression in poly(A)⁺ RNA samples from human lung, trachea, mammary gland, intestine, and spleen (Clontech) and in total RNA isolated from MCF-10A cells. PCR products were gel purified (QIAquick gel extraction kit; Qiagen), cloned into the pGem-T vector (Promega), and sequenced. In all RT-PCR assays, water, substituting for RNA in the reverse transcription, served as a negative control. A cDNA fragment of EF-1 was amplified to control for conditions of reverse transcription and PCR (13).

In vitro translation, construction of glycosylation mutants, and protease protection assay. The hCLCA2 ORF cloned into the expression vector pcDNA3.1 was transcribed and translated with the TNT T7 coupled reticulocyte lysate system (Promega) in the presence of L-[³⁵S]methionine (Amersham). Reactions were carried out at 30°C for 90 min with and without canine pancreatic microsomal membranes (Promega). Samples were analyzed by 8% SDS-PAGE, followed by drying of the gel and exposure to film for 8 h.

Six cDNA constructs were generated with an asparagine (AAT or AAC) to glutamine (CAA) mutation that abolishes a consensus glycosylation site (NxS/TQxS/T). These mutations (N150Q, N292Q, N522Q, N637Q, N822Q, and N938Q) were positioned between adjacent potential transmembrane domains as determined by hydrophobicity analysis (Kyte-Doolittle algorithm, 19 amino acid residues per window). The constructs were generated by overlap extension PCR as described (11) and cloned into the Not I site of pcDNA3.1, using Not I site-incorporated PCR primers. Correct sequences of the products were verified by sequencing. The six constructs were in vitro translated in the absence or presence of microsomal membranes and resolved on an 8% ProSieve 50 Tris/glycine SDS-polyacrylamide gel (FMC Bioproducts) that enabled optimal resolution in the high-molecular-mass range. Molecular masses were estimated using molecular mass standards and a digital image analysis system (AlphaImager; AlphaInnotech). In addition, protease protection assays were performed as described (26). Briefly, in the presence of microsomal membranes in vitro translated and ³⁵S-labeled wild-type hCLCA2 was incubated with proteinase K (Sigma; 100 µg/ml) for 60 min on ice with or without detergent present (0.5% Nonidet P-40). The reaction was stopped by adding 5 mM phenylmethylsulfonyl fluoride, and the products were analyzed by 12% SDS-PAGE, drying of the gel, and exposure to film.

Expression of Myc-tagged protein constructs in HEK-293 cells. Two Myc-tagged hCLCA2 constructs were generated by inserting a partial sequence of the human c-Myc protein (EQKLISEEDL) (5) near the amino or carboxy terminus of hCLCA2 (between amino acids 33 and 34 or 725 and 726), using overlap extension PCR as described (11), and cloned into pcDNA3.1. Correct sequences were verified by sequencing. DNA constructs were transfected into 70% confluent HEK-293 cells via the Lipofectamine Plus method (Life Technologies) using 20 µl Plus reagent, 30 µl lipid, and 4 µg DNA/100-mm dish in a 3-h incubation. Cells were lysed 48 h later in the presence of protease inhibitors (1% aprotinin, 1 µM leupeptin, 2 mM phenylmethylsulfonyl fluoride). Lysates were resolved via 10% SDS-PAGE, blotted, and probed with mouse anti-human c-Myc antibody 9E10 (Calbiochem), followed by enhanced chemiluminescence detection (ECL; Amersham). Surface expression of the proteins was shown by surface biotinylation (Biotin NHS, Vector; 100 µg/ml) of transfected nonpermeabilized HEK-293 cells 48 h after transfection (20 min at 37°C, followed by extensive washing with PBS). Immunoprecipitation with antibody 9E10, incubation with protein G beads, boiling of the beads in SDS loading buffer, and SDS-PAGE were followed by blotting and detection with peroxidase-conjugated streptavidin and ECL. To estimate the extent of hCLCA2 glycosylation, 9E10 immunoprecipitates were incubated with N-glycanase (0.3 U/40 µl sample; Genzyme) at 37°C for 18 h before sample denaturation and loading of the gel.

Electrophysiology. The hCLCA2 cDNA was transiently transfected into HEK-293 cells, which lack an endogenous Ca²⁺-activated Cl conductance (9, 11, 41). To control for transfection efficiency and to identify transfected cells, the same cells were cotransfected with a green fluorescent protein reporter vector (EGFP, Clontech). Parental HEK-293 cells were cultured in DMEM with 10% fetal bovine serum in the absence of antibiotics. Cells grown on collagen-coated glass coverslips placed in the bottom of 35-mm dishes were transfected with 5 µl Lipofectamine, 0.5 µg of hCLCA2 cloned into pcDNA3.1, and 0.5 µg of EGFP during a 2- to 3-h incubation period (Life Technologies). After transfection, cells were allowed to recover for 24 h before patch-clamp recording. In all cases, the cells that were cotransfected with hCLCA2 and EGFP and that fluoresced green were also expressing a Ca²⁺-sensitive Cl conductance, whereas no currents were observed in mock-transfected (EGFP alone) or untransfected cells (see also Refs. 9 and 11). To record channel activities under whole cell conditions, cells were superfused at 1-2 ml/min with bath solution (in mM: 112 N-methyl-D-glucamine chloride, 30 sucrose, 2 CaCl₂, 2 MgCl₂, and 5 HEPES, pH 7.4). Borosilicate glass electrodes (tip resistance 6-9 M) were filled with an identical solution plus 5 mM ATP. In some experiments designed to examine the effect of 4 µM ionomycin, the pipette solution also contained 1 mM EGTA and 0.366 mM CaCl₂. The free Ca²⁺ concentration under these conditions was calculated to be ~25 nM. After seal formation (>1 G) and establishment of the whole cell recording configuration, cells were clamped at +20 mV and currents recorded at room temperature using an Axopatch 200A (Axon Instruments, Foster City, CA) connected to a personal computer through a TL1 interface (Axon) with 12-bit resolution. The records were sampled at 5-10 kHz and filtered at 1-2 kHz with a four-pole Bessel filter. The current-voltage relationship of hCLCA2 was determined using 300-ms voltage steps from a holding potential of +20 mV to potentials from 100 to +100 mV at 10-mV intervals. To normalize membrane currents for differences in cell size, the capacitative current transiently recorded in response to a 10-mV hyperpolarizing pulse was integrated and divided by the given voltage to give total membrane capacitance for each cell.

Nucleotide sequence accession number. The GenBank accession number for the hCLCA2 sequence is AF043977.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Identification and cloning of hCLCA2. A human lung cDNA library was screened with Lu-ECAM-1 cDNA as probe in an attempt to isolate a CLCA homolog from human lung. After the sequencing of the positive clones and completion of the 5' and 3' cDNA ends by the RACE technique, a single 3.6-kb cDNA species was identified and named hCLCA2. Sequence accuracy was verified by sequencing of four different full-length RT-PCR products from human trachea mRNA generated by the highly accurate Pwo DNA polymerase. The nucleotide sequence shared high degrees of identity with those of Lu-ECAM-1 (86%), bCLCA1 (85%), mCLCA1 (76%), and hCLCA1 (63%). Northern blot analyses under highly stringent conditions yielded bands of the expected size of 3.6 kb in trachea and mammary gland, whereas all other tissues tested were negative (Fig. 1). Although isolated from a lung cDNA library, hCLCA2 was not detected in the lung by Northern blot hybridization. However, the more sensitive RT-PCR revealed its expression in lung in addition to trachea and mammary gland, suggesting a significantly lower expression level in the lung. Because of RNA analyses from whole tissue extracts, the cell types expressing hCLCA2 could not be identified. However, hCLCA2 was also detected in the nonmalignant human mammary epithelial cell line MCF-10A using both Northern blot and RT-PCR analyses (Fig. 1), suggesting epithelial expression at least in the mammary gland. All PCR products were sequenced, and sequence identities with the cDNA isolated from the lung library together with the observed signal size of 3.6 kb on the RNA blots indicated that both the RNA blot and RT-PCR signals in fact represented hCLCA2.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Northern blot hybridization and RT-PCR analyses of hCLCA2 tissue expression pattern. Poly(A)+ RNA (2 µg/lane) from human tissues and mammary epithelial cell line MCF-10A was hybridized with a 32P-labeled probe representing the hCLCA2 open reading frame (ORF; top left). Absence of a signal for intestinal hCLCA1 indicates specificity for hCLCA2. Blots were stripped and rehybridized with a probe for elongation factor-1 (EF-1) as an internal control (bottom left). Exposure times were 48 h (hCLCA2) and 24 h (EF-1). Top right: RT-PCR analysis of hCLCA2 expression in human tissues and cell line MCF-10A with primers flanking entire ORF (2,838 bp). A 219-bp fragment of EF-1 was amplified to control for RNA quality and RT-PCR conditions (bottom right). Negative controls starting from reverse transcription included sterile water instead of RNA.

Characterization of the hCLCA2 protein. The ORF of the hCLCA2 cDNA encodes a 943-amino acid polypeptide with high levels of amino acid sequence identity with Lu-ECAM-1 (76%), bCLCA1 (76%), mCLCA1 (69%), and hCLCA1 (51%; Fig. 2). The polypeptide is preceded by a canonical signal sequence with a predicted signal peptidase cleavage site between amino acids 31 and 32 (32). The predicted size of the full-length protein (104 kDa) is consistent with the results of an in vitro translation assay yielding a primary translation product of ~105 kDa (Fig. 3A). In the presence of microsomal membranes, the protein was glycosylated in vitro to a 120-kDa glycoprotein. To ascertain whether the hCLCA2 protein is cleaved into two subunits in mammalian cells as reported for other CLCA homologs (8, 9, 11), two cDNA constructs were generated with a c-Myc tag within the amino or carboxy terminus, respectively (constructs "m1" and "m2") and transfected into HEK-293 cells. In fact, immunoblots of cell lysates probed with an anti-Myc antibody identified an 86-kDa protein when the tag was inserted near the amino terminus (m1) and a 34-kDa protein when the tag was situated near its carboxy terminus (m2; Fig. 3B), confirming a similar cleavage in hCLCA2. The presence of protease inhibitors in the lysis buffer suggests that the observed cleavage did not occur after lysis of the cells. To study the extent of glycosylation of each subunit, immunoprecipitates of both Myc-tagged constructs from transfected HEK-293 cells were deglycosylated by N-glycanase treatment. The 86- and 34-kDa glycoproteins were reduced in size by 11 and 2.5 kDa, proposing approximately four and one glycosylation sites, respectively (Fig. 3B). Detection of the two Myc-tagged constructs in anti-Myc antibody immunoprecipitates from surface-biotinylated, nonpermeabilized HEK-293 cells suggests that both the 86- and 34-kDa proteins are expressed on the surface of the transfected cells (Fig. 3C).

View larger version (101K): [in this window] [in a new window]   Fig. 2.   Predicted hCLCA2 amino acid sequence aligned (Clustal method) with all known CLCA homologs. Identical amino acids are indicated by dots; dashes represent gaps. Major hydrophobic regions spanning 19 or more amino acid residues are overlined (S, signal sequence; 1-5, transmembrane domains). Cysteine residues conserved in amino-terminal extracellular domain are bold and enlarged. Consensus sites are marked for N-linked glycosylation (*), phosphorylation by protein kinase C (PKC) (:), and phosphorylation by cAMP-dependent protein kinase (). Arrow, conserved consensus site for monobasic proteolytic cleavage; ::, 2 adjacent PKC sites. GenBank accession numbers are AF039400 (hCLCA1), U36455 (bCLCA1 or CaCC), AF001261 (Lu-ECAM-1), and AF047838 (mCLCA1). Lu-ECAM-1, lung endothelial cell adhesion molecule-1; h, human; b, bovine; m, murine.

View larger version (52K): [in this window] [in a new window]   Fig. 3.   Biochemical analysis of the hCLCA2 protein. A: in vitro translation without (M) and with (+M) processing and glycosylation by microsomal membranes. Glycosylation site scanning was performed by in vitro translation of 6 glycosylation knockouts [N150Q to N938Q; wild type (wt)]. Unglycosylated constructs uniformly ran at 105 kDa (M), whereas glycosylated proteins were 120 kDa (wt, N292Q, N637Q, N938Q) or 118 kDa (N150Q, N522Q, N822Q) in size (+M; 8% gel). In a protease protection assay, three fragments of 18, 21, and 30 kDa were protected from degradation by proteinase K in absence of detergent (D), and the polypeptide was fully degraded in presence of detergent (+D; 12% gel). In all experiments, L-[35S]methionine-labeled proteins were detected by SDS-PAGE, drying of the gel, and exposure to film for 8 h. B: immunoblot detection of Myc-tagged hCLCA2 constructs overexpressed in HEK-293 cells. Tags were placed near the amino terminus (m1) or within the carboxy-terminal cleavage product (m2; see Fig. 5). The 120-kDa precursor protein was processed into 2 proteins of 86 kDa (m1) and 34 kDa (m2). The somewhat weaker band below the 34-kDa band (2nd lane) probably represents an incomplete glycosylation variant of this protein, which may not be visible in the less abundant 120-kDa precursor due to lower resolution in the high-molecular-mass range of this blot. Bands at 65 and 67 kDa represent endogenous c-Myc. Cell lysates including protease inhibitors were resolved by 10% SDS-PAGE 48 h after transfection, transferred to a membrane, and probed with anti-Myc antibody 9E10. Deglycosylation with N-glycanase (+G) reduced the molecular masses from 86 to 75 kDa (m1) and from 33.8 to 31.4 kDa (m2). C: analysis of surface expression of Myc-tagged hCLCA2 constructs m1 and m2. Transfected HEK-293 cells were surface biotinylated, washed extensively with PBS, and lysed in the presence of protease inhibitors, followed by immunoprecipitation with anti-Myc antibody 9E10, SDS-PAGE, and probing with horseradish peroxidase-conjugated streptavidin. Both the 86- and 34-kDa proteins were biotinylated and therefore associated with the apical cell membrane.

To elucidate the transmembrane topology of hCLCA2, glycosylation site scanning and protease protection assays were performed as described for other channel proteins (28, 38). Potential transmembrane regions were determined using a hydrophobicity analysis (Fig. 4). Based on this prediction, six glycosylation knockout cDNA constructs were generated, each having a single asparagine-to-glutamine (NQ) mutation that abolishes a consensus glycosylation site between two adjacent potential transmembrane domains (N150Q, N292Q, N522Q, N637Q, N822Q, and N938Q). Following in vitro translation and glycosylation, the products were analyzed using a high-resolution PAGE. Mutation of three of the six sites (N150Q, N522Q, and N822Q) led to a reduction in size of the respective glycoproteins by ~2 kDa, indicating usage of these sites and therefore extracellular location (Fig. 3A). However, lack of size reduction of the remaining constructs did not necessarily prove intracellular location of their mutated sites, because an extracellular consensus glycosylation site may not have been used due to sterical hindrance. Therefore, a protease protection assay was performed to determine the sizes of the extracellular domains, complementing the information derived from the glycosylation site scanning. In principle, extracellular loops are protected from proteolysis due to their translocation into the lumen of lipid microsomes, functionally representing the endoplasmic reticulum. The calculated sizes of each possible extracellular domain of hCLCA2 are given in Fig. 4 with ~2 kDa to be added per adjacent transmembrane domain. Wild-type hCLCA2 was in vitro translated in the presence of microsomal membranes and digested with proteinase K in the absence or presence of detergent. Three degradation products of 18, 21, and 30 kDa were protected from proteolysis in the absence of detergent (Fig. 3A), indicating their extension into the microsomes. In the presence of detergent, the protein was fully degraded. Both the glycosylation data and the sizes of protected extracellular domains are consistent with a five-transmembrane topology (Fig. 5). In the proposed model, the fragments protected from proteolysis correspond in size to the first extracellular domain (30 kDa, resulting from 27.7 kDa plus one transmembrane segment, tm1), the second extracellular domain (18 kDa, resulting from 13.7 kDa plus two transmembrane segments, tm2 and tm3), and the third extracellular domain (21 kDa, resulting from 16.8 kDa plus two transmembrane segments, tm4 and tm5). The size of the 21-kDa fragment also indicates degradation and therefore intracellular location of the carboxy-terminal tail of ~2 kDa. Given the number and locations of consensus glycosylation sites of the primary hCLCA2 polypeptide (Fig. 2), this transmembrane model is also consistent with the extent of glycosylation of the two 86- and 34-kDa subunits as detected by N-glycanase treatment (Fig. 3B). Accordingly, the 86-kDa subunit contains three glycosylation sites within the first (N74, N97, and N150) and one within the second extracellular loop (N522), whereas only one site (N822) is present within the extracellular loop of the 34-kDa cleavage product (Fig. 5).

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Hydrophobicity plot of hCLCA2 amino acid sequence (Kyte-Doolittle algorithm, 19 amino acid residues/window). Hydrophobic domains are given as positive values (SS, signal sequence; tm1 to tm5, transmembrane regions). Numbers and bars above curve indicate calculated molecular masses between transmembrane domains, including 2 kDa per potential glycosylation site for interpretation of results of protease protection assay. Names of glycosylation knockout mutants (NxQ) indicate locations of their abolished glycosylation sites. Sites where a Myc tag was placed in constructs m1 and m2 are indicated with arrows under curve. Arrowhead, consensus site for monobasic proteolytic cleavage. Units are in kcal/mol (vertical axis) and first amino acid residue/window (horizontal axis).

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Proposed transmembrane topology of hCLCA2 with transmembrane domains numbered 1-5. Dashed lines, locations of c-Myc tag in constructs m1 and m2. Sites of asparagine-linked glycosylation (Nx) are indicated by treelike drawing. Cx, cysteine residues conserved among all known CLCA homologs; Px, consensus sites for phosphorylation by protein kinase C, with sites conserved among all known homologs underlined; arrow, conserved consensus site for monobasic proteolytic cleavage.

When compared with the amino acid sequences of previously cloned homologs, hCLCA2 shares a number of interesting sequence motifs. For example, the pattern of cysteine residues present within the large amino-terminal, extracellular domain of all previously cloned homologs is conserved in hCLCA2 (Fig. 2). Also conserved is the consensus site for monobasic proteolytic cleavage following arginine residue 675 (7), the location of which is consistent with the sizes of the cleavage products of 86 and 34 kDa. Analyses of the predicted intracellular domains of hCLCA2 revealed seven consensus phosphorylation sites for protein kinase C (PKC; Fig. 5) but none for Ca²⁺/calmodulin protein kinase II or cAMP-dependent protein kinase.

Electrophysiological characterization of hCLCA2. Electrophysiological studies for hCLCA2 were conducted in a manner analogous to those reported for bCLCA1, mCLCA1, and hCLCA1, which have all been shown to be associated with activation of a Ca²⁺-sensitive Cl conductance. When transiently expressed in HEK-293 cells, hCLCA2 was associated with an outwardly rectified current that was activated by ionomycin. In contrast, nontransfected cells, cells transfected with the EGFP vector alone (mock-transfected), or transfected cells in the absence of ionomycin were not associated with any significant current (Fig. 6). The outwardly rectifying current-voltage relationship exhibited by hCLCA2-transfected HEK-293 cells exposed to 2 mM Ca²⁺ in the pipette was absent from vector alone transfected cells (Fig. 7) (see also Refs. 9, 11, 39). This current was sensitive to DIDS (300 µM), DTT (2 mM), niflumic acid (NFA; 100 µM), and tamoxifen (10 µM) (Fig. 8). The average current recorded at +100 mV was 9.60 ± 2.87 pA/pF and was reduced to 0.15 ± 1.60 (SE) pA/pF (n = 5) in the presence of DIDS. Exposure to DTT reduced the mean current from 9.70 ± 6.42 to 1.96 ± 2.09 pA/pF (n = 6). Similarly, both NFA and tamoxifen reduced the current from a mean of 6.24 ± 4.75 to 0.64 ± 0.96 pA/pF (n = 6) and from 12.05 ± 3.85 to 1.02 ± 1.68 pA/pF (n = 5), respectively (Fig. 9). In contrast, no significant current was recorded from cells that were either untransfected or transfected with the EGFP vector alone. In the case of untransfected cells, the average current recorded in the presence of 2 mM Ca²⁺ in the pipette was 1.57 ± 0.72 pA/pF (n = 8), whereas, in mock-transfected cells, the current in the presence of Ca²⁺ was 0.97 ± 0.39 pA/pF (n = 10). When the pipette solution contained low Ca²⁺ (~25 nM) with 2 mM Ca²⁺ in the bath, perfusion of the Ca²⁺ ionophore ionomycin through the bath also activated the current (Figs. 6 and 9). Under these conditions, average currents in vector alone transfected and untransfected cells in the presence of ionomycin were 1.52 ± 1.83 (n = 5) and 0.22 ± 1.02 pA/pF (n = 8), respectively. In hCLCA2-transfected cells, addition of ionomycin increased the current from 1.7 ± 1.04 to 10.77 ± 3.8 pA/pF (n = 7, P < 0.001). These results suggest that expression of hCLCA2 in HEK-293 cells is associated with the appearance of a Ca²⁺-sensitive Cl conductance.

View larger version (36K): [in this window] [in a new window]   Fig. 6.   Representative data collected from individual cells illustrating Cl current (ICl) expression in hCLCA2-transfected HEK-293 cells. In presence of low (25 nM) internal Ca2+ and high (2 mM) external Ca2+, neither untransfected nor mock-transfected [green fluorescent protein reporter vector (EGFP) alone] HEK-293 cells showed significant current expression following exposure to ionomycin (IONO). In contrast, in hCLCA2-transfected cells, ionomycin exposure resulted in appearance of an outwardly rectified ICl.

View larger version (40K): [in this window] [in a new window]   Fig. 7.   Attenuation of hCLCA2 current expression by ICl inhibitors. In presence of 2 mM intracellular (pipette) and extracellular Ca2+ (bath), hCLCA2-transfected HEK-293 cells consistently expressed Cl currents with a maximum amplitude between 300 and 500 pA. Subsequent exposure to 300 mM DIDS, 2 mM dithiothreitol (DTT), 100 mM niflumic acid (NFA), or 10 mM tamoxifen resulted in near-complete inhibition of expressed current. Data are taken from individual cells.

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Current-voltage relationship of hCLCA2 expressed in HEK-293 cells. Representative data collected from individual cells showing effect of inhibitors on whole cell currents in presence of 2 mM intracellular Ca2+. Currents recorded from untransfected or EGFP alone transfected cells were negligible (see text). A: current expressed in hCLCA2-transfected cells displayed slight outward rectification with maximal activation of 12.06 pA/pF measured at +100 mV (). Subsequent perfusion with 300 µM DIDS resulted in inhibition of the current to 0.66 pA/pF (). B-D: as above, using inhibitors DTT (2 mM, B) yielding 10.12 pA/pF () and 2.41 pA/pF (), NFA (100 µM, C) yielding 14.98 pA/pF () and 1.78 pA/pF (), and tamoxifen (10 µM, D) yielding 14.95 pA/pF () and 1.96 pA/pF ().

View larger version (28K): [in this window] [in a new window]   Fig. 9.   Summary of effects of ionomycin and inhibitors on hCLCA2 current expression recorded at +100 mV. In presence of low cytosolic free Ca2+ (~25 nM), exposure of EGFP alone transfected and untransfected HEK-293 cells in presence of ionomycin yielded currents of 1.52 ± 1.83 (n = 5) and 0.22 ± 1.02 (SE) pA/pF (n = 8), respectively. Exposure of hCLCA2-transfected cells to ionomycin resulted in an increase in current from 1.69 ± 1.04 to 10.77 ± 3.8 pA/pF (n = 7; * P < 0.001). When internally perfused with solutions containing 2 mM Ca2+, transfected cells exhibited a maximal current of 9.27 ± 4.91 pA/pF (n = 22; * P < 0.001). Exposure to DIDS (300 µM), DTT (2 mM), NFA (100 µM), or tamoxifen (10 µM) inhibited the maximally activated currents to 0.152 ± 1.6, 1.95 ± 2.09, 0.64 ± 0.96, or 1.02 ± 1.68 pA/pF, respectively.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

A novel family of Ca²⁺-activated Cl channels has recently been introduced by our laboratories (6, 8, 9, 11, 12). The family members cloned thus far are the bovine Lu-ECAM-1 (8, 40), the bovine Ca²⁺-sensitive Cl channel (CaCC or bCLCA1) (6), the murine mCLCA1 (9), and a first human homolog, hCLCA1, which is exclusively expressed in the intestine (11). Here we have described a second human family member that is expressed in human lung, trachea, and mammary gland. hCLCA2 shares many of the structural and functional peculiarities of its homologs. For example, the sizes and extent of glycosylation of the primary in vitro translation products of CLCA homologs are conserved within the family. In mammalian cells, the hCLCA2 primary translation product was cleaved into 86-kDa amino-terminal and 34-kDa carboxy-terminal polypeptides that are both associated with the outer cell surface. The corresponding cleavage products are 90 and 38 kDa for Lu-ECAM-1 (8), 90 and 37-41 kDa for hCLCA1 (11), and 90 and 32-38 kDa for mCLCA1 (9). It is unclear whether there are any structural or functional relationships between the larger amino-terminal and the smaller carboxy-terminal polypeptides of the CLCA homologs, although studies with amino- and carboxy-terminal-truncated constructs of bCLCA1 have suggested that the carboxy-terminal cleavage product is dispensable for channel function (15). Also conserved among all homologs are a pattern of amino-terminal, extracellular cysteine residues, a consensus recognition site for monobasic proteolytic cleavage that is consistent with the sizes of the two cleavage products, and consensus sites for phosphorylation by PKC, although their respective functional significance remains to be established.

Detailed biochemical analyses on the structure of hCLCA2 revealed a five-transmembrane topology with three transmembrane domains within the 86-kDa amino-terminal and two within the 34-kDa carboxy-terminal cleavage products. This result is at variance with the models proposed for earlier cloned CLCA homologs, where, based on the much less reliable hydrophobicity analyses alone, four transmembrane regions have been suggested, all located within the larger amino-terminal cleavage product (6, 8, 9, 11). A consequence of this difference is the intracellular location of the predicted cleavage site between the two hCLCA2 subunits and extracellular locations for the proposed models for Lu-ECAM-1, mCLCA1, and hCLCA1. Therefore, the data on the CLCA homologs other than hCLCA2 need to be reevaluated, and analogous biochemical studies will have to be performed for each family member. Given similar functional characteristics among CLCA homologs, it is likely that the established hCLCA2 transmembrane topology will serve as the prototype for all members of this channel family.

Measurements on the Ca²⁺-activated Cl conductance of hCLCA2 were performed in transfected HEK-293 cells. This cell line was chosen because it is devoid of any intrinsic Ca²⁺-activated Cl conductance (41). Consistent with previous findings on CLCA homologs, the Ca²⁺ ionophore ionomycin elicited an increase in whole cell current in hCLCA2-expressing HEK-293 cells. This current was sensitive to standard inhibitors of Cl channels, such as DIDS, NFA, and tamoxifen. Although Ca²⁺ was present in the bath at unphysiological concentrations (2 mM), recent evidence suggests that ionomycin may release Ca²⁺ from the endoplasmic reticulum store exclusively, in which case its concentration only increases in the range of 200-500 nM (34). The data presented here do not unequivocally prove that hCLCA2 forms an anion channel itself but would also be consistent with a role of hCLCA2 as a regulator of an as-yet-unidentified, endogenous channel that by itself is not sensitive to Ca²⁺. However, it has been shown that the closely related bovine tracheal bCLCA1 forms a genuine channel protein when reconstituted into planar lipid bilayers (16, 24). Under the conditions used in the present study, we did not observe any time dependence of activation of the Ca²⁺-sensitive current. Although time-dependent activation of Ca²⁺-sensitive Cl currents has been previously reported in epithelial cells (2, 37), the lack of such a characteristic may reflect the use of a heterologous expression system and the loss of associated proteins that confer this property (17). Alternatively, hCLCA2 may not underlie the time-dependent Ca²⁺-sensitive current recorded from native airway cells, even though it is expressed in that tissue. The observation that the hCLCA2-associated current is also sensitive to the anti-estrogen tamoxifen is also consistent with a role for the expressed protein as an anion channel. Several other anion channels, notably a volume-regulated channel described in Ehrlich ascites tumor cells (22), a Ca²⁺-activated Cl current identified in arterial endothelial cells (21), and a member of the ClC family of Cl channels (39), were also shown to be sensitive to this compound. Considering the expression of hCLCA2 in human breast epithelium, it remains to be established whether its sensitivity to tamoxifen plays any role in the effectiveness of this drug against breast cancer.

The function of hCLCA2 as a mediator of a Ca²⁺-activated Cl current and its expression in human lung and trachea warrant future investigations aimed at its potential involvement in the complex ion-secretory disorder of cystic fibrosis. Especially intriguing is the question whether it may form a viable alternate Cl channel that could be exploited for pharmacological targeting to circumvent the defect of the CFTR Cl channel. Studies in other systems, including animal models of cystic fibrosis, have shown that a Ca²⁺-sensitive Cl conductance is present in cystic fibrosis cells and may even be upregulated. This observation is particularly relevant in the cystic fibrosis knockout mouse model, where expression of an as-yet-unidentified Ca²⁺- and DIDS-sensitive Cl conductance is thought to rescue the cystic fibrosis mouse from significant airway disease (10, 27, 36). In the same CFTR (/) mouse, lethal intestinal pathology is associated with absence of a Ca²⁺-activated pathway for Cl secretion, whereas expression of a Ca²⁺-sensitive Cl conductance in the murine intestine is thought to compensate for the lack of CFTR function and rescue the intestinal phenotype (4, 36). However, to what extent the Ca²⁺-sensitive Cl conductance may substitute for the defective CFTR in human cystic fibrosis is unclear, especially since data obtained in CFTR (/) mice may not be readily extrapolated to human cystic fibrosis due to their significantly different cystic fibrosis phenotypes (3, 4, 27, 30). A future challenge will be to establish whether differences in tissue-specific channels between species contribute to the differences observed between the phenotypes of cystic fibrosis patients and murine CFTR knockouts.