| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Veterinary Biomedical Sciences, University of Saskatchewan, Saskatoon, Saskatchewan, Canada
ABSTRACT I. INTRODUCTION TO CLCA PROTEINS II. CLONING THE CLCA GENES A. Bovine CLCAs B. Human and Mouse CLCAs C. Porcine CLCAs D. Miscellaneous III. CLCA PROTEIN STRUCTURE A. Human CLCAs 1. hCLCA1 2. hCLCA2 3. hCLCA3 4. hCLCA4 B. Mouse CLCAs 1. mCLCA1/2 2. mCLCA3 3. mCLCA4 4. mCLCA5 and mCLCA6 C. Bovine CLCAs 1. bCLCA1 2. bCLCA2 D. Porcine pCLCA1 IV. ENDOGENOUS TISSUE EXPRESSION SITES A. Human CLCAs 1. hCLCA1 2. hCLCA2 3. hCLCA3 4. hCLCA4 B. Murine CLCAs 1. mCLCA1 2. mCLCA2 3. mCLCA3 4. mCLCA4 5. mCLCA5 and mCLCA6 C. Bovine CLCAs 1. bCLCA1 2. bCLCA2 D. Porcine pCLCA1 E. Other CLCAs F. Correlating Tissue Expression Sites to CLCA Groups V. FUNCTIONAL RESPONSES TO EXPRESSION OF CLCA PROTEINS A. Human CLCAs 1. hCLCA1 2. hCLCA2 3. hCLCA3 and hCLCA4 B. Mouse CLCAs 1. mCLCA1 2. mCLCA2 3. mCLCA3 4. mCLCA4 5. mCLCA5 and 6 C. Bovine CLCAs 1. bCLCA1 D. Porcine pCLCA1 E. Other CLCAs VI. PATHOPHYSIOLOGICAL CONNECTIONS TO CLCA EXPRESSION A. CLCA in Asthma 1. Asthma and the genesis of its mediators 2. Th2 cytokines mediate CLCA expression B. CLCA Function in Cystic Fibrosis C. CLCA in the Normal Epithelium D. CLCA in CF Epithelium VII. CLCA PROTEINS IN ONCOLOGY A. Cell Cycle Control 1. Effects of CLCA overexpression on the cell cycle 2. CLCA suppression of tumor cell cycling 3. Apoptosis in CLCA tumor suppression 4. A mechanism for CLCA proapoptotic effects 5. Loss of CLCA in tumorigenic cell lines B. Cell Adhesion and Tumor Metastases 1. CLCA and cell adhesion 2. CLCA binds lung colonizing cell lines in vitro 3. {beta}4-Integrin expression, CLCA binding, and tumor metastasis 4. CLCA binding domains 5. Novel {beta}4-integrin binding domain 6. Response to CLCA/{beta}4-integrin binding 7. A novel {beta}4 mitogenic signaling pathway 8. Tumor invasion VIII. CLCA, VASCULAR TONE, AND HYPERTENSION A. Hypertension IX. CLCA, BESTROPHINS, AND RETINOPATHY X. SUMMARY REFERENCES
 |
ABSTRACT |
|---|
 Top NextReferences |
|---|
100 kDa. Three of the four human loci code for the synthesis of membrane-associated proteins. CLCA proteins affect chloride conductance, epithelial secretion, cell-cell adhesion, apoptosis, cell cycle control, mucus production in asthma, and blood pressure. There is a structural and probable functional divergence between CLCA isoforms containing or not containing 
4-integrin binding domains. Cell cycle control and tumor metastasis are affected by isoforms with the binding domains. These isoforms are expressed prominently in smooth muscle, in some endothelial cells, in the central nervous system, and also in secretory epithelial cells. The isoform with disrupted 4-integrin binding (hCLCA1, pCLCA1, mCLCA3) alters epithelial mucus secretion and ion transport processes. It is preferentially expressed in secretory epithelial tissues including trachea and small intestine. Chloride conductance is affected by the expression of several CLCA proteins. However, the dependence of the resulting electrical signature on the expression system rather than the CLCA protein suggests that these proteins are not independent Ca2+-dependent chloride channels, but may contribute to the activity of chloride channels formed by, or in conjunction with, other proteins. |
I. INTRODUCTION TO CLCA PROTEINS |
|---|
|
TopPreviousNextReferences |
|---|
The predominant chloride conductance of most secretory tissues occurs via the cAMP-dependent cystic fibrosis transmembrane regulator protein (CFTR). However, many tissues also have an independent chloride conductance that can be activated by elevating the concentration of intracellular Ca2+. Activity of such a protein has been functionally linked to fluid secretion in the trachea and in the small intestine. Upon their discovery in 1995, the members of the CLCA protein family were thought to be this elusive calcium-dependent chloride conductance. Nearly a decade later, research into CLCA protein biology is proceeding in interesting and unexpected directions.
Attempts to understand CLCA protein function have been complicated by uncertainty about a unifying molecular basis for the actions of family members. Different CLCA proteins have several apparently unrelated, and sometimes conflicting, functional roles. The first discovered member of the family was implicated with chloride conductance in secretory epithelial tissues (39). However, another original CLCA protein, LuECAM-1, functions in tumor cell adhesion and metastasis (1–3, 50). This role for LuECAM-1 in tumor adhesion appears to conflict with significant tumor suppressor activity reported for other CLCA proteins (49, 79, 112). CLCA proteins have also been connected recently to the regulation of mucus production in the asthmatic airway (88, 135), and to ion transport in the airway of cystic fibrosis patients (83, 84). Other roles include CLCA involvement in chloride transport affecting vascular smooth muscle tone and blood pressure, and in retinal functions that are disrupted in Best's vitelliform macular dystrophy (9, 73, 94, 124).
The history of CLCA cloning will be surveyed, followed by physiological effects observed upon heterologous CLCA isoform expression. Predictions for protein structure and function will connect some basic biophysical results of CLCA isoform expression with the pathophysiology of associated disease processes. The targeting of CLCA proteins for pharmacological modulation of related disease conditions will also be discussed.
|
II. CLONING THE CLCA GENES |
|---|
|
TopPreviousNextReferences |
|---|
|
The first CLCA cDNA sequence (bCLCA1) was reported in 1995 (39). The authors were pursuing the gene coding for a 140-kDa bovine tracheal protein that had been identified with by Ca2+/calmodulin-dependent protein kinase-regulated chloride conductance. The 140-kDa protein was thought to be a homotetramer of identical 38-kDa subunits, linked together by disulfide bonds. Reduction of the 140-kDa protein produced a 36- to 38-kDa polypeptide lacking chloride channel activity in lipid bilayers (39, 155). A candidate cDNA was identified by screening a bovine tracheal cDNA library with polyclonal antibody to this polypeptide. The identified cDNA was described as coding for a calcium-activated chloride channel (CaCC), but is now called bCLCA1 (39, 59, 60, 146).
A second bovine CLCA sequence (Lu ECAM-1or bCLCA2) was identified by screening a gene expression library with monoclonal antibody that blocked an adhesion-receptor/ligand pair interaction between lung-metastatic melanoma cells and lung matrix-modulated bovine aortic endothelial cells (50, 211–214). Expression of the Lu ECAM-1 cDNA produced a protein with properties of a lung endothelial cell adhesion molecule(50). Lu ECAM-1 is now termed bCLCA2 (59, 146).
Clones similar to bCLCA1 and -2 were reported in 1998 from human genome and mouse lung library screening. The human CLCA1 gene and its putative promoter region span 31,902 bp of chromosome 1 at p22-p31 (75). A promoter region containing a TATA box precedes the predicted site of transcriptional initiation by 22 nucleotides. Fifteen exons from 90 to 604 bp are interspersed with 14 introns of 170 to 5,651 bp. The hCLCA1 cDNA was found independently in 1999 by screening an EST database with bCLCA1 sequence. The EST clone, named HCaCC1, is identical to hCLCA1 (4). The 2,745 bp open reading frame (ORF) of hCLCA1 was amplified in polymerase chain reaction (PCR) to clone a full-length cDNA for functional expression.
A bCLCA2 probe was used to identify mCLCA1 in a mouse lung cDNA library (67). mCLCA1 was also identified in a EST database using a probe based on bovine bCLCA1 sequence (160). This group localized mCLCA1 to mouse chromosome 3 at the H2-H3 band. mCLCA1 is the murine ortholog of truncated hCLCA3 (Fig. 1) (158).
Several CLCA isoforms were reported in 1999. mCLCA2 was cloned from mouse mammary gland in a suppression subtractive hybridization on the involuting mammary gland (107). mCLCA2 maps closely to mCLCA1 in the mouse genome and may even be a splice variant of mCLCA1 (158). The hCLCA2 cDNA was cloned from a human lung library using a bCLCA1 cDNA probe (78). As with mCLCA2, hCLCA2 was expressed at high levels in the mammary gland, but the murine counterpart of hCLCA2 is the hypothetical mouse protein 4732440A06 (158), now called mCLCA5 (54). hCaCC2 and hCaCC3 (now hCLCA4 and hCLCA3) were discovered independently at the same time (4).
hCLCA3 is a short CLCA protein with a premature stop codon (78). It is most similar to mCLCA1/2 and to bCLCA1 and bCLCA2, although the mouse and bovine isoforms are full-length membrane-associated proteins. hCLCA3 was cloned from human spleen cDNA and has been found in several tissues.
Gob-5 or mCLCA3, the mouse ortholog to hCLCA1, was cloned from mouse gut cDNA (103). This mCLCA3 ortholog, as well as mCLCA1/2, mCLCA4, and related mouse ESTs 4732440A06, AI747448 and AI504701 (mCLCA5, 6, and ?) colocalize on mouse chromosome 3 H2-H3 (108, 158, 160).
The human hCLCA2, hCLCA1, hCLCA4, and hCLCA3 genes align consecutively in the same orientation, encompassing 232 kb of chromosome 1 with no other genes interspersed (76). Gene clustering can be a mechanism for controlling temporal gene expression by internal intron or bidirectional promoters (95, 208). An area 1,617 bp upstream of the first intron codes for 24 different transcription factor binding sites.
A monoclonal antibody selected to inhibit chloride conductance in ileal brush-border vesicles was used to identify pCLCA1 cDNA in a porcine ileal gene expression library (61, 63). Cloning and expression of the full-length cDNA was reported in 2000 (68). The screening strategy used to identify this isoform connects pCLCA1 to chloride conductance.
Reports of a calcium-activated chloride conductance in smooth muscle (85, 137) led to the discovery of mCLCA4 (48). mCLCA4 is most similar to the prematurely truncated hCLCA3, but mCLCA4 is a full-length integral membrane protein.
Incomplete segments of CLCA cDNA have been cloned from other species. A portion of a rat ortholog rCLCA1 has been cloned from rat pancreas cDNA (179). A partial canine CLCA, cCLCA1, cloned from dog retinal pigment epithelium with primers specific for the porcine pCLCA1 has also been reported, and expression levels have been investigated (117). An hCLCA5 cDNA has been reported recently (1). However, little information on its cloning or function is available.
|
III. CLCA PROTEIN STRUCTURE |
|---|
|
TopPreviousNextReferences |
|---|
The cDNA for hCLCA1 encodes a 914-amino acid protein with a calculated molecular mass of 100.9 kDa (75). A signal sequence and four putative transmembrane domains are predicted from hydropathy data. There are 9 potential sites for asparagine-linked glycosylation, 13 consensus sites for protein kinase C (PKC) phosphorylation, and 3 consensus sites for phosphorylation by Ca2+/calmodulin-dependent kinase II. There are no PKA or tyrosine phosphorylation consensus sequences (75). In vitro translation of the human CLCA1 resulted in a 100-kDa product that increased in size to 125 kDa after glycosylation.
Membrane topology predictions have been based on detection of the c-myc epitope (EQKLISEEDL) inserted at different regions within the protein. However, epitope insertion may be affecting protein conformation, as CLCA proteins are cleaved by monobasic proteases with conformation-dependent cleavage specificity, and several of the tagged constructs were not processed normally (44, 164). Insertion of the epitope between amino acids 366/367 or 492/493 resulted in a 90-kDa polypeptide, apparently interfering with the ability of the smaller cleaved 40-kDa product to interact with the larger 90-kDa subunit. Other epitope placements blocked proteolytic cleavage (75). A model based on the c-myc epitope insertion and cell surface biotinylation has an extracellular NH2 and COOH terminus and four transmembrane domains (80, 146) (Fig. 2).
|
2-integrin collagen receptor domain, the CLCA family members are suggested to consist of a central (extracellular) von Willebrand factor domain A (VWA domain), leaving only a single transmembrane domain near the COOH terminus (197). The significant disparity between the alternate models involves predictions about the folding and topography of the NH2-terminal half of the protein. This region of the protein is predicted to contain either two helical transmembrane domains (TM1 and TM2), or to be extracellular with secondary and tertiary structure forming a VWA domain. The TM3 and TM4 domains and the cytoplasmic loop between these domains containing protein kinase consensus sites are not excluded by either conformation of the NH2-terminal folding.
VWA domains are recognized for involvement in protein-protein interactions. The voltage-gated calcium channel is a precedent for VWA domain involvement in ion channel subunit interactions. The properties of this channel are modulated by interaction of the pore-forming 1-subunit with the 2 VWA domain-containing subunit complex (86, 197). There is parallel evidence for CLCA interaction with other proteins. Hetero-oligomeric CLCA interactions affect cell binding and mitogenic signaling via integrins (1–3), as well as ion channel activity of the 1-subunit of the large-conductance potassium channel (73). CLCA expression also modulates chloride conductance, leading to predictions that some chloride conductance channels may be hetero-oligomeric structures.
A metal ion-dependent adhesion site (MIDAS) motif is a noncontiguous amino acid sequence that organizes into a divalent cation binding structure through normal VWA domain folding. MIDAS motifs bind magnesium or calcium and participate in stabilizing protein-protein interactions. CLCA folding, based on sequence homology to a collagen receptor, could form a VWA domain with MIDAS function (197). The MIDAS motif could be a site for calcium binding in CLCA proteins to interact with and alter the conductivity of chloride ion channels. However, MIDAS motif precedents in other proteins involve extracellular polypeptide structure, while effects of Ca2+ ionophores on chloride conductance associated with CLCA expression invoke changes in intracellular Ca2+ concentration as the basis for CLCA regulation of chloride conductance. Without structural data, there is no direct evidence for VWA domain formation and function in CLCA proteins in spite of significant amino acid sequence identity and modeling compatibility (see Fig. 3).
|
The hCLCA2 ortholog is predicted to form a 943-amino acid polypeptide (80). It contains a canonical signal sequence with a peptidase cleavage site. A primary translation product of 105 kDa from an in vitro translation assay increases to 120 kDa after glycosylation. Proteolytic processing at the monobasic proteolytic 673/674 cleavage site leaves an NH2-terminal c-myc epitope tag in an 86-kDa protein and a 34-kDa COOH terminus. Deglycosylation reduced the size corresponding to four glycosylation sites on the 86-kDa fragment and one on the smaller 34-kDa fragment. Loss of glycosylation in mutants N150Q and N522Q indicated that these regions are extracellular (80). However, these results are compatible with either multiple transmembrane domains or an extracellular NH2-terminal VWA domain. A N822Q mutation indicated an extracellular loop in the smaller cleaved product between transmembrane domains 4 and 5.
Findings from a microsomal protein protection assay digesting all internal loops gave results in agreement with the interpreted five transmembrane domain structure of hCLCA2. It should also be noted that both cleavage products were detected in surface-biotinylated, nonpermeablized HEK293 cells, suggesting that both the larger and the smaller products localize to the cell surface. In this model the internal loops contained seven PKC phosphorylation sites, but there were no consensus sites for Ca2+/calmodulin protein kinase II or cAMP-dependent protein kinase (80). hCLCA2, as well as several other CLCA proteins, contain a 4-integrin binding domain (1) whose significance is discussed below.
The hCLCA3 cDNA contains two internal stop codons (78). Sequential ORFs code for the NH2-terminal 262 amino acids, and a second polypeptide from amino acid 266 to 461. Together, the two ORFs code for approximately the first one-third of the full-length CLCA protein. The NH2-terminal polypeptide terminates before potential transmembrane domains, and the second fragment lacks a signal sequence, although its amino acid sequence includes the first two predicted transmembrane domains of other CLCA family members. In vitro synthesis produced 30- and 22-kDa products corresponding to calculated size for ORF1 and ORF2, respectively, but only the NH2-terminal polypeptide was detected in transfected HEK293 cells. The secreted ORF1 product was also detected upon immunoblotting the supernatant from transfected cells (78). Mouse homologs mCLCA1/2 and mCLCA4 lack the internal stop codons, implying that there may be no function for the truncated, secreted hCLCA3 product.
There are no current structural data for the hCLCA4 clone.
The in vitro translated size of 100 kDa for this 902-amino acid protein increases to 125 or 130 kDa with glycosylation. Posttranslational processing gives NH2-terminal 90-kDa and COOH-terminal 38/32-kDa components. A general four transmembrane domain structure is proposed for this protein, leaving the COOH terminus without a transmembrane domain (67). There is little structural data available for mCLCA2. The mCLCA2 protein may be a splice variant of mCLCA1 (158), sharing significant cDNA sequence identity with mCLCA1 and hCLCA3. It is reported to produce similar processed products as found for mCLCA1 when expressed in HEK293 cells (49).
This mouse CLCA is the murine counterpart of hCLCA1. The mCLCA3 cDNA codes for a 100-kDa protein in an in vitro translation assay, and increases to 110 kDa after glycosylation with microsomal membranes (109). Microsomal protease treatment produced a 35-kDa product corresponding to an extracellular NH2 terminus. An NH2-terminal 90-kDa product was identified in transiently transfected HEK293 and COS-7 cells and in the mouse large and small intestine, with the expression being more intense in the small intestine. An unexplained 45-kDa product which may be a fragment of a truncated portion of the amino terminus of the protein with similarities to the secreted 37 kDa amino terminus of hCLCA3 was also detected in the large, but not small intestine (109).
The primary structure of mCLCA4 is similar to mCLCA1 and 2. It is 909 amino acids in length, with a conserved cleaved NH2-terminal signal and a second monobasic cleavage site which results in a 90- and 30- to 40-kDa product (48). The primary structure also contains most of the calcium/calmodulin kinase II (CaMKII) sites found in mCLCA1 and mCLCA2.
The two proteins have predicted mass ratios of 103.6 and 101.9 kDa, respectively (54). Uncleaved glycosylated proteins expressed in tsA201 cells were 125 kDa. Processing by monobasic cleavage produced COOH-terminal glycosylated fragments of 35 kDa after subtraction of the EGFP tag (54). Several consensus acceptor sites for protein kinase A (PKA), PKC and CaMKII occur in the amino acid sequence, but the significance of these sites is uncertain without more information on the transmembrane orientation of the expressed proteins. Conserved 4-integrin binding motifs on NH2- and COOH-fragments of mCLCA5 are altered so as to be nonfunctional in mCLCA6.
Cloned bovine isoforms are homologs of hCLCA3 and mCLCA1/2. The bCLCA1 cDNA codes for 903 amino acids, giving a predicted product of 100 kDa which shifts to 140 kDa upon glycosylation (39). Motif analysis predicts 15 PKC acceptor sites, 10 Ca2+/calmodulin-dependent protein kinase sites, and 3 tyrosine kinase sites. Four transmembrane domains and an NH2-terminal signal sequence are predicted from hydropathy plots. Polyclonal antibodies generated against an NH2-terminal fragment of the bCLCA1 clone recognized a reduced 36/38-kDa bovine tracheal lysate product similar to the antigen identified by the original antibodies that were used to clone bCLCA1 (39, 155). The original purified protein of 140 kDa was thought to consist of four identical 38-kDa subunits linked together by disulfide bonds (155). However, the sequence data and the identification of the 100-kDa unglycosylated form of bCLCA1 indicate that the 38-kDa protein is a posttranslational cleavage product that remains associated with its larger counterpart (39). This model has now been promoted for the other cloned isoforms and orthologs.
The cDNA sequence codes for a predicted hydrophobic NH2-terminal signal sequence and cleavage site for membrane insertion. In vitro translation produced a 101-kDa protein that increased to 120 kDa with glycosylation. Expression of bCLCA2 in HEK293 cells produced similar 90- and 38-kDa products. Native expression in bovine endothelial cells produced 130- and 32-kDa glycoforms (50). As for bCLCA1, a hydropathy plot predicted four transmembrane domains. Deglycosylation of native bCLCA2 with N-glycosidase F from endothelial cells reduced the 38- and 32-kDa polypeptides to 22 kDa, and the 90-kDa band to 77 kDa, giving the predicted deglycosylated size of bCLCA2 (50). There is good evidence that the primary 101-kDa protein is cleaved into two nonidentical proteins (50), and earlier suggestions of a homotetrameric structure for this protein (155) are incorrect.
The pCLCA1 protein is the porcine homolog of hCLCA1 and mCLCA3. The cDNA sequence contains 3,079 bp and a 2,751-base ORF that is predicted to encode a 917-amino acid protein with a molecular mass of 100.7 kDa (68). It shares 78% amino acid sequence identity with hCLCA1, its closest ortholog. Similarities of amino acid sequence and hydrophobicity between pCLCA1 and hCLCA1 suggested a similar transmembrane topology for the two proteins. Modeling this structure according to the suggestions of Gruber et al. (80) would give an extracellular NH2 terminus followed by four transmembrane domains and a hydrophobic COOH terminus (Fig. 2) (68).
The predicted pCLCA1 protein shares a signal sequence for membrane targeting and potential proteolytic cleavage sites with hCLCA1. There is an amidation cleavage site at residue 140 and monobasic proteolytic cleavage sites at R662 and K720. This could account for the multiple sizes of protein (130, 90, and 60 kDa) detected by Western blotting or immunoprecipitation of brush-border vesicle protein with an inhibitory antibody (63, 154).
Expression of exogenous pCLCA1 in transfected Caco-2 cells produced a 60-kDa product upon Western blotting with polyclonal antibody raised to an NH2-terminal peptide (C250-K266) of pCLCA1. The processing of CLCA isoforms is known to vary with expression in different tissues (60, 146). Examples include an 60-kDa bCLCA1 fragment and 60-kDa pCLCA1 product expressed in epithelial cells, as well as a preliminary report of a 60-kDa protein contributing to anion conductance in the rabbit ileum (147, 155).
Similar to other CLCA clones, pCLCA1 amino acid sequence from V307 to Q462 has extensive homology with sequences that form a VWA domain (Fig. 3). There is a perfect noncontiguous consensus sequence for a MIDAS motif (D313-x-S315-x-S317... T384... D416) within the putative pCLCA1 VWA domain, a motif shared with hCLCA1. Structural modeling of the pCLCA1 MIDAS site reveals that appropriate folding is permitted, although the -SH of C387 replaces -OH of T384 as one of the divalent cation ligands (Fig. 4). Loewen et al. (113) have presented evidence for a direct regulatory role of calcium in the modulation of chloride conductance by pCLCA1, but there is no evidence connecting regulation of pCLCA1 to Ca2+ binding to the MIDAS motif.
|
|
IV. ENDOGENOUS TISSUE EXPRESSION SITES |
|---|
|
TopPreviousNextReferences |
|---|
The strongest expression of hCLCA1 occurs in mucus-secreting cells of large and small intestine (75). With Northern blotting the highest expression was in the small intestine, appendix, and colon, with much lower expression levels in the uterus, stomach, testis, and kidney (4). Another extensive study using highly stringent in situ hybridization conditions only detected expression in the small intestine and colon (75). The majority of the mRNA signal was in cells at the base of the crypts, with the goblet cells having the highest expression. Induction of hCLCA1 expression in the airways has been reported more recently under pathophysiological conditions (83, 88, 135, 181, 209).
By Northern blotting the highest expression of hCLCA2 mRNA occurs in the trachea and mammary gland (80). This result was confirmed independently, and the confirming group also reported expression in the testis, prostate, and uterus (4). It has also been found in nasal epithelium (122). Strong expression occurs in most basal cells in stratified epithelia (35). Although isolated from a lung cDNA library, hCLCA2 was not detected in the lung by Northern blot hybridization, and only a weak RT-PCR signal for hCLCA2 was obtained from the lung. Others could not confirm expression in the lung (4).
The hCLCA2 mRNA was highly expressed in a nonmalignant transformed human mammary epithelial cell line MCF10A using both Northern blotting and RT-PCR. However, hCLCA2 expression could not be detected by Northern blotting in tumorigenic cell lines MDA-MB-231, MDA-MD-468, and MCF7. These findings were in agreement with an in situ hybridization staining of acini and small ducts in normal mammary tissue, but there was no staining in breast cancer samples (80).
The third human ortholog, hCLCA3, was originally cloned from spleen. The mRNA transcript is expressed in the lung, trachea, mammary gland, and thymus (78). Its presence has also been reported in nasal epithelium (122).
Unlike the other human orthologs, hCLCA4 is highly expressed in neural tissue (4). On Northern blot analysis it is found in the amygdala, caudate nucleus, cerebral cortex, frontal lobe, hippocampus, medulla oblongata, occipital lobe, putamen, substantia nigra, temporal lobe, thalamus, and accumbens expressed hCLCA4. The cerebellum and spinal cord did not show any evidence of hCLCA4 expression. The strongest signal from hCLCA4 came from the colon, with twice the intensity found in neural tissue. Expression also occurred in the bladder, uterus, prostate, stomach, testis, salivary gland, mammary gland, small intestine, appendix, and trachea (4).
The strongest expression of mCLCA1 was detected in the lung, aorta, spleen, and bone marrow using real-time RT-PCR quantitation normalized against expression of elongation factor 1a (110). High levels were also reported in the kidney and skin by Romio et al. (160). Lower levels occur in many tissues, but expression was not detected in the prostate or in the pregnant, lactating, and involuting mammary gland. However, mCLCA1 expression was found in virgin, pregnant, and lactating gland in an independent study (49). There is agreement that mCLCA1 expression is lost during mammary gland involution. There is a dynamic change in mCLCA1 expression over gestation (49, 110) pointing to an undefined role of mCLCA1 in cell maturation and tissue differentiation.
mCLCA1 is the only CLCA reported in the cerebellum and brain stem (110). This differs significantly from hCLCA4, which is expressed extensively in other areas of the nervous system (4). In the same study, mCLCA2 was undetectable in cerebellum and brain stem (77). These reports of temporal and tissue-specific expression patterns within the same physiological system may be hints of diverse roles for CLCA proteins in tissue development and functional differentiation.
Given the importance of a calcium-activated chloride conductance in renal epithelium, it was expected that CLCA cDNAs related to mCLCA1 and mCLCA2 would be expressed in a mouse inner medullary collecting tubule cell line (19). RT-PCR of mouse nephron RNA found mCLCA1 mRNA in the glomeruli and thick ascending limb, but not in the proximal tubule and cortical collecting duct (19). Interestingly, a study using a nonspecific mCLCA1/ mCLCA2 probe identified CLCA in the tubular epithelial cells but not in the glomeruli. The proximal tubule had the most intense staining with weaker staining in the loop of Henle and distal tubuli. The collecting duct was negative (77). These contradictory findings may reflect probe specificity and primers used in the studies, as much as they show a difference in the CLCA form expressed at different areas within the kidney.
mCLCA1 and mCLCA2 may be splice variants of the same gene locus (158). High mCLCA2 expression was reported from pregnancy to involution as determined in two studies by RT-PCR (49, 115), but the strongest mCLCA2 expression was detected in the involuting mammary gland (48, 107, 110). mCLCA1/2 expression was detected in both alveolar and ductal epithelial cells by in situ hybridization in normal murine mammary tissue (77). Lower mCLCA2 levels were detected in most tissues tested except for the brain stem, cerebellum, and prostate (110). In a real time quantitative RT-PCR study, the involuting mammary gland seemed to express purely mCLCA2. This was the only tissue screen with absolutely no mCLCA1 signal (110). The dynamic interplay between the mCLCA1 and mCLCA2 isoforms suggests that they are involved in different physiological roles. It is interesting to note that the highest expression of mCLCA2 seems to be in those tissues with the high rate of cell division and cell death (49). A possible connection to cell cycle control or an apoptotic role for mCLCA2 could be considered for these tissues.
Mucus-secreting cells are the primary site for expression of mCLCA3, the mouse homolog of hCLCA1 and pCLCA1 (103, 109). The mCLCA3 was originally cloned from, and found to have highest expression in, the crypts of the large intestine (103). An extensive immunohistochemical investigation of mCLCA3 showed exclusive expression in the digestive and respiratory tracts and uterus (109). The mCLCA3 antigen was not found in the gallbladder, kidney, pancreas, sublingual salivary glands, oviduct, mammary gland, or prostate (109). Concurrent staining for mucin and immunohistochemical localization of mCLCA3 antigen indicated that mCLCA3 was expressed only in mucin-producing cells, but not in all mucin-producing cells.
A dynamic expression pattern was seen in the crypt goblet cells in both the large and small intestine, with expression only in approximately the upper two-thirds of the crypt, whereas weak, or no, staining was observed in the basal third (109). It is interesting that the distribution pattern for gastrointestinal mCLCA3 correlates with the crypt-villus axis (Fig. 5). CLCA proteins have been promoted as proapoptotic (49), but it is the cell group directed toward apoptotic removal at the base of the crypts that loses mCLCA3 expression. However, mCLCA1/2 expression may increase as mCLCA3 levels decline and play some role in apoptosis at the base of the crypts.
|
Cloned from smooth muscle, mCLCA4 was detected in the gastrointestinal tract, uterus, lung, and heart in a multiple cDNA array using gene-specific primers. A large component of mCLCA4 cDNA expression was detected by RT-PCR in the smooth muscle of the dissected tunica muscularis of the bladder and stomach (49). In situ hybridization with a mCLCA4-specific probe was performed to determine cell-specific expression in these mixed organs. This probe produced a very strong signal in the pulmonary vein, aorta, and atrioventricular bundle with a much lower level in cardiac muscle, coronary artery, and endothelium. Both the bronchioles and blood vessels were labeled in the lung. In the gastrointestinal tract, labeling was associated more with the mucosa than the muscularis tunica. This mucosal labeling was most intense towards the villus tip (48). It should be noted that this is the first CLCA with a reported increase in expression towards the villus tip. mCLCA1 and possibly mCLCA2 are thought to be primarily expressed in the crypt (77), and mCLCA3 is associated with upper crypt and midshaft of the villus (109). These differing expression patterns of each isoform within a tissue apparently associate with major differences in cell physiology. Adipose and connective tissue were consistently negative (48).
Mouse homologs of hCLCA2 and hCLCA4 have been designated as mCLCA5 and mCLCA6 (54). Like its human counterpart, mCLCA5 is widely expressed, with eye, spleen, and lactating mammary tissue notable for strong expression (14, 54). However, mCLCA6 is expressed in intestine and stomach, but not in whole brain. In this respect the human and mouse homologs appear to be functionally distinct.
The bCLCA1 ortholog is primarily expressed on the brush border of ciliated tracheal epithelial cells (50), the tissue from which it was originally cloned (39). This expression site is similar to that found for mCLCA1/2 (77). The expression in other tissues has not been reported.
The bCLCA2 ortholog is reported to be predominantly a luminal membrane protein of the venular endothelia of the lungs and the spleen. However, antibodies to bCLCA2 located a strong antigen signal in endothelia of small to medium size venules as well as in the respiratory epithelial of the bronchi and trachea (50).
Immunohistochemistry with the IgM monoclonal antibody used to clone pCLCA1 identified antigen on the enterocyte border of the villi. Labeling intensity was distributed over the mucosal surface, with the most intense staining in the mid to upper crypt region (154). In the trachea, pCLCA1 mRNA expression was localized to surface epithelium and the underlying submucosal glands, with the most intense staining found in a subset of the submucosal glands (68). These tissues were also positive for pCLCA1 expression when tested by RT-PCR. pCLCA1 expression was not detected in the colon. RT-PCR identified pCLCA1 mRNA in the parotid, sublingual, and submandibular salivary glands (68). Other tissues that express a variety of chloride channels including the exocrine pancreas, cardiac and skeletal muscle, liver, and kidney had no detectable pCLCA1 mRNA.
Antibodies against a pCLCA1 peptide were used in immunohistochemistry to identify a CLCA epitope in canine retinal pigment epithelial cells (117). Intense staining was seen on the apical secretory side of the epithelium. The Muller cells, which are reported to maintain appropriate extracellular environment for retinal neurons, had significant cCLCA1 expression. CLCA epitope was also prominent in the corneal epithelium and in the ciliary body (M. E. Loewen, B. H. Grahn, and G. W. Forsyth, unpublished data).
Antibodies generated to rCLCA1 reacted extensively in the rat pancreas. The subcellular location of the antigen was mainly on the zymogen granules (179). In an interesting aside, hCLCA1 expression was not detected by RT-PCR in a human pancreatic duct adenocarcinoma cell line that had an active calcium-dependent chloride conductance (56). This negative finding is consistent with observations by others that CLCA expression does not always correlate with the presence of calcium-activated chloride currents (114).
F. Correlating Tissue Expression Sites to CLCA Groups
Figure 1 illustrates a connection between two structurally related CLCA groups. The first "group" contains the cross species homologs, hCLCA1, mCLCA3, and pCLCA1. The divergent connecting branch contains hCLCA2 and hCLCA4 and now also mCLCA5 and -6 corresponding to these human forms (14, 54). The second group contains two subgroups. The first of these includes mCLCA1/2, which may be alternate splice products from a single locus (158) and mCLCA4. The second subgroup contains hCLCA3, bCLCA1, and bCLCA2. Functional similarities within these structural subgroups are probable, and similarities in cross-species tissue expression sites within CLCA groups support this hypothesis.
The hCLCA1, mCLCA3, and pCLCA1 forms have a similar distribution pattern in the three species (Table 1), concentrated in mucus-producing epithelium in the gastrointestinal and respiratory tracts. This group of CLCA proteins may participate in mucin secretion and modulation of chloride conductance in secretory epithelial tissues. The members are conspicuously absent from pancreas, smooth muscle, and endothelial cells.
|
The most striking property of the second cluster of structures, including mCLCA1/2, mCLCA4, hCLCA3, bCLCA1, and bCLCA2, is the wide distribution of all of these forms. Their presence has been reported in most tissues where their expression has been investigated, although there is limited data for hCLCA3, bCLCA1, and bCLCA2. Disparate functional roles for this third group include suggestions of smooth muscle chloride channel modulation (mCLCA4, and by extrapolation, the human isoform hCLCA3), and possible endothelial adhesion factors bCLCA2, hCLCA2, and mCLCA1. A third function is suggested by the surface epithelial expression of endothelial adhesion CLCAs in the gastrointestinal tract.
|
V. FUNCTIONAL RESPONSES TO EXPRESSION OF CLCA PROTEINS |
|---|
|
TopPreviousNextReferences |
|---|
Transient expression of hCLCA1 in HEK293 human embryonic kidney cells increased calcium-activated whole cell currents from 1.57 ± 0.72 pA/pF in control cells to 11.06 pA/pF in transfected cells, producing a time-independent outwardly rectifying current (75). This current was said to be "electrically isolated" for chloride, based on the blockage of cation current using large bulky cations, but the chloride dependence of the current was not determined. These putative chloride currents were inhibited by DIDS, dithiothreitol (DTT), and niflumic acid, although DIDS is the only one of these compounds known to be a specific blocker of chloride channels.
This study also presented the only single-channel patch-clamp study of the CLCA proteins. Single channels of hCLCA1-transfected cells had a slope conductance of 13.4 pS in cell-attached conformation after the addition of ionomycin in the presence of 1 mM Ca2+ (75). There was no mention in the report of endogenous HEK293 channels responsible for the background of nonselective and anion currents previously reported in these cells (206, 215).
Transient transfection with hCLCA2 produced a calcium-activated, non time-dependent, outwardly rectifying increase in whole cell current in HEK293 cells. The whole cell current increased from 1.52 ± 1.83 pA/pF in control to 10.77 ± 3.8 pA/pF in transfected cells. Once again, neither the background current in untransfected cells nor the chloride dependence of the current produced in transfected cells were reported. However, the "electrically isolated" anion current was blocked by DIDS, DTT, niflumic acid, and tamoxifen (80).
There are no current reports of functional ion transport data for hCLCA3 and hCLCA4.
mCLCA1 expression in HEK293 cells caused the appearance of a calcium-activated chloride current. The current was outwardly rectifying and not time dependent (67). The current increased from 2.05 ± 1.09 to 10.23 pA/pF upon addition of ionomycin. Background chloride currents were apparently insignificant in this study, but the anionic dependence of the current was not determined. Niflumic acid, DTT, and DIDS inhibited the inferred chloride current.
Romio et al. (160) concurrently cloned and expressed mCLCA1 in Xenopus oocytes. Injection of mCLCA1 cRNA into oocytes caused a significant increase in current without added calcium ionophore. At –80 mV, the current in mCLCA1 mRNA-injected oocytes was –398 ± 136 nA compared with –97 ± 12 nA for water-injected oocytes. The current in the mCLCA1-injected oocytes, but not in the water-injected oocytes, was chloride dependent. The lack of a reversal potential (Erev) shift when bath solution was changed to low NaCl indicated that the background oocyte current was a combination of both anion and cation currents. DIDS and niflumic acid inhibited chloride current in mCLCA1-injected oocytes and in water-injected controls. The application of a Ca2+ ionophore (ionomycin) caused a transient, statistically insignificant increase in current to 1,805 ± 473 nA in mCLCA1-expressing oocytes compared with 1,109 ± 389 nA in control oocytes (160). It was difficult to determine any effect of Ca2+ on the chloride conductance associated with CLCA expression due to the large background currents in Xenopus oocytes.
Generally mCLCA1 expression in HEK293 cells resulted in a mildly outwardly rectifed, time-independent current (25, 73). This current requires 2 mM Ca2+ in the pipette and does not increase above control values with 500 nM Ca2+. The stimulated mCLCA1 current was most permeable to SCN– > Cl– > isethionate (25), as were the native Ca2+-activated chloride channels in smooth muscle from which mCLCA1 was cloned. The anion dependence of the whole cell current in control HEK293 cells was not assessed.
The mCLCA1-dependent calcium-stimulated current changed significantly from time independent to time dependent, as well as increasing the total whole cell conductance when coexpressed with a potassium channel -subunit (73). Coexpression also resulted in a greater sensitivity to activation by calcium. Addition of 500 nM Ca2+ in the pipette solution could stimulate the anion current in -subunit mCLCA1 expressing cells. This difference in current and agonist sensitivity was shown by a mammalian two-hybrid system to be the result of a direct interaction between the potassium channel -subunit and mCLCA1. Unfortunately, the anionic dependence of the current was only assessed through the permeability ratios for SCN– and Cl–, which were consistent with those found in smooth muscle. Again, as with several other CLCA studies, the background currents were not reported (73). This study by Greenwood et al. was the first report suggesting that a CLCA protein requires coexpression with other proteins to produce functional chloride channels and that the currents induced on CLCA1 expression could be modified by an accessory protein.
There are no published electrophysiological data on expressed mCLCA2.
The effects of mCLCA3 or Gob 5 expression have been characterized briefly in HEK293 cells (203). mCLCA3 expression increased an outwardly rectifying current without agonist addition. Current densities at +60 mV increased from 59 ± 17 to 230 ± 47 pA/pF. Addition of 10 mM EGTA to the pipette solution significantly decreased these currents (203).
Treating transient mCLCA4-transfected HEK293 cells held at an undefined controlled voltage with ionomycin or methacholine increased intracellular calcium and evoked transient inward currents (48). This effect would be consistent with the inward movement of cations or the outward movement of anions. In this case, a strong anionic dependence to the stimulated current was shown by a shift in the current-voltage relationship to the equilibrium potential of chloride when bath solution was switched to low chloride after ionomycin addition. Unlike previously characterized isoforms, the current was found to be nonrectifying. The transient nature of the current was thought to be analogous to the Ca2+-independent inactivation of the smooth muscle Ca2+-activated chloride channel. This inward current was found to be spontaneous, similar to those in smooth muscle cells that are induced by a "calcium spark" and are involved in buffering the membrane potential, relaxation, and spontaneous rhythmic contraction of smooth muscle (92, 136). Current-voltage relationships in nontransfected cells were not examined, as voltage-clamped inward currents were not seen (48).
Expression of these isoforms caused an ionomycin-dependent increase in whole cell current in HEK293 cells (54). Chloride involvement was verified by reversal potential measurements. A requirement for 2 mM calcium in the pipette to induce these increases in chloride current was consistent with other reports that are difficult to reconcile with channel regulation at intracellular calcium concentrations. In contrast to normal mammary epithelial cells, metastatic mammary tumor cells express little or no mCLCA5 upon starvation or detachment. Transfection of a mammary tumor cell line with mCLCA5 caused a dramatic reduction in colony formation by transfected tumor cells (14).
Oocytes transfected with bCLCA1 cRNA had a significantly larger current than water-injected control oocytes (39). bCLCA1 expression apparently increased an endogenous outwardly rectifying current. This current was sensitive to DIDS and DTT but insensitive to niflumic acid. Addition of 1 µM ionomycin to oocytes resulted in the activation of endogenous chloride channels which were inhibited by niflumic acid. Unfortunately, data showing the effect of the ionophore on bCLCA1-injected oocytes were not presented or discussed (39).
A nonrectifying, Ca2+-activated, DTT-inhibitable current was produced when bCLCA1 was expressed in COS-7 cells (39). This linear current observed after bCLCA1 expression was unlike the outwardly rectified currents seen with expression of most of the other CLCA family members. However, no other family members have been characterized using COS-7 cells. It is interesting to note that the currents induced in the oocytes had a both time-dependent and rectifying quality, whereas those induced in the COS-7 cells had neither (39).
Lipid bilayers constructed of membranes from oocytes that had been injected with bCLCA1 cRNA had a unit channel conductance of 21 pS (39). The open probability increased from 0.41 ± 0.07 to 0.60 ± 0.08 in the presence of Ca2+ added only to the side opposite to that where DIDS was added. The channel was inhibited by DIDS and DTT, but insensitive to niflumic acid. The channel had an 8:1 anion to cation selectivity ratio and 3:1 selectivity for iodide over chloride under bionic conditions. The biophysical properties were similar to those found for a purified chloride channel from bovine trachea. However, some bilayers had a Ca2+-activated chloride conductance that was inhibited by niflumic acid. This was said to be the endogenous Ca2+-activated chloride channel of the oocyte. Others using bCLCA1 as a control for mCLCA1 characterization found the bCLCA1-induced current was inhibited by 78% with niflumic acid. Subsequent addition of DIDS (100 µM) caused a further 33% inhibition of the anion current (160). The basis for these differences is not clear.
pCLCA1 expression in NIH/3T3 cells increased a calcium-activated 36Cl– efflux (68, 115, 117). Agonists for calcium-dependent protein kinase (PKC) or cAMP-dependent protein kinase (PKA) had no effect. However, application of 10 µM ionomycin significantly increased 36Cl–efflux from transfected cells. Inhibition of this effect by membrane-permeable Ca2+ chelator, combined with the lack of inhibition by the CaMKII inhibitor KN-93, suggested that Ca2+ may directly regulate a pCLCA1 effect on chloride transport. The pCLCA1 effect on chloride efflux was inhibited by 5-nitro-2-(3-phenylpropyl-amino)-benzoate (NPPB), glibenclamide, diphenylamine carboxylate (DPC), and -phenylcinnamate (-PC). NPPB and DPC inhibited the Ca2+-activated chloride efflux at much lower concentrations than were required for inhibition of CFTR. NPPB was effective at a concentration similar to that which was required to inhibit Ca2+-activated chloride conductance in Xenopus oocytes (115).
In whole cell patch clamp, pCLCA1 increased an endogenous Ca2+-activated chloride conductance, making the current more anion dependent (115). The pCLCA1 chloride current was outwardly rectifying and time dependent. This time dependence of chloride currents in NIH/3T3 fibroblasts has not been reported for other CLCA proteins, suggesting contributions by the expression system to the chloride current. As with 36Cl– efflux, DIDS did not inhibit whole cell chloride current, but NPPB, DPC, and -PC were inhibitory (115). Both DTT and DIDS were reported to inhibit chloride current activity in previous studies with different CLCA proteins and different expression systems (39, 67, 75), but did not inhibit effects of pCLCA1 in 3T3 cells. In contrast to the fibroblast expression system (113), transfection of an epithelial Caco-2 human colon carcinoma cells with pCLCA1 produced a cAMP-activated chloride conductance, possibly through effects on CFTR (113). The pCLCA1 and PKA-dependent currents were anion dependent, time independent, and nonrectifying. The effects of pCLCA1 expression on PKA-dependent chloride conductance were evident in freshly passaged and in differentiated epithelial cells, both of which have endogenous CFTR-mediated chloride conductance. This modulation of chloride conductance by pCLCA1 expression was also seen in nonepithelial NIH/3T3 fibroblasts coexpressing CFTR and pCLCA1 (117).
Expression of pCLCA1 changed the chloride conductance activated by calcium ionophore or PKA agonists in Caco-2 epithelial cells to a time-dependent outwardly rectifying current (114). However, the endogenous Ca2+-activated chloride conductance is lost as Caco-2 cells mature, and the ability of pCLCA1 to activate Ca2+-dependent chloride conductance also disappears (114). Again, the endogenous chloride channel activity in the expression system influences the properties of the chloride channels observed upon CLCA expression.
A rCLCA1 isoform indirectly implicated in bicarbonate transport and vesicle exocytosis in rat pancreas and a canine cCLCA1 isoform (117) produced in secretory retinal pigment epithelium have not been fully cloned or functionally characterized.
|
VI. PATHOPHYSIOLOGICAL CONNECTIONS TO CLCA EXPRESSION |
|---|
|
TopPreviousNextReferences |
|---|
Orthologs of the CLCA gene family are overexpressed in bronchial allergic asthmatic responses that overproduce mucus (88, 135, 181). There is significant evidence that the CLCA proteins have a causal role in this condition, suggesting that these proteins may have some interest as targets for pharmacological intervention in allergic asthma.
1. Asthma and the genesis of its mediators
Bronchial allergic asthma is a serious inflammatory condition of the airways resulting in bronchial narrowing, constriction, and overproduction of mucus (27, 181). Simplistically, the uncontrolled inflammatory response, accompanied by the overexpression of CLCA protein, is thought to occur through a dysregulation between type 1 helper T (Th1) and type 2 helper T (Th2) cell immune responses to allergens (10, 27, 47, 90, 134, 159, 192). The Th1-immune response involving interleukin (IL)-2 and interferon (IFN)-
cytokines is seen as a cell-mediated response associated with disease in situations of poor hygiene or infection with Mycobacterium tuberculosis, measles virus, or hepatitis A virus (21, 127, 130, 167, 171). The Th2 disease response involving a different set of cytokines (IL-4, -5, -6, -9, and -13) is considered to be humoral, involving the stimulation of B cells and the production of IgE. Released IgE binds to IgE receptors on mast cells, lymphocytes, eosinophils, platelets, and macrophages. Binding of allergen to the IgE results in the release of inflammatory mediators from mast cells and activation and potentiation of an inflammatory response (27).
Proper priming of the immune system to different immunological challenges is essential to develop a normal Th1 response. The Th1 response then has a role in modulating the Th2 response. The asthma disease state is dominated by an uncontrolled Th2 response, and blockade of the Th2 cytokine pathways significantly dampens the asthmatic response (26, 57).
2. Th2 cytokines mediate CLCA expression
The Th2 response and the associated cytokines IL-9, IL-4 and IL-13 appear to be directly linked to CLCA overexpression in the asthmatic patient, and in mouse models of asthma (209). However, instilled Th1 cytokine INF- did not induce expression of mCLCA3. mCLCA3 synthesis was induced in the lung and bronchi in an allergic asthma mouse model of ovalbumin sensitization, and transfection with an adenoviral mCLCA3 antisense construct prevented airway hyperresponsiveness (AHR) and mucus production after ovalbumin challenge (135). Application of a sense mCLCA3 mRNA construct to this same model produced severe AHR and mucus secretion (135). In vitro overexpression of mCLCA3 or its closest human ortholog hCLCA1 in human mucoepidermoid cells (NCI-H292) increased mucus secretion and overexpression of the major secretory mucin gene of asthma, MUC5AC (88). An increase in hCLCA1 correlated with an increase in mucus and IL-9 receptor in asthmatic patients (Fig. 6). This relationship suggests that hCLCA1 might be a therapeutic target in asthma patients.
