| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 41, 39960-39968, October 10, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Pharmacology, School of Medicine, University of North Carolina, Chapel Hill, North Carolina 27599-7365
Received for publication, July 21, 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
,-2
, and -2
). Although all three enzyme forms were expressed heterologously to similar levels in Chinese hamster ovary cells clone K-1 (CHO-K1) cells, only the 495-amino acid protein (NTPDase2 exhibited ecto-ATPase activity. Immunolocalization studies demonstrated that NTPDase2 is fully processed and trafficked to the plasma membrane, whereas the NTPDase2 and -2 splice variants were retained in not fully glycosylated forms in the endoplasmic reticulum. The potential roles of two highly conserved residues, Cys399 and Asn443, in the activity and cellular trafficking of the ecto-ATPase were examined. Mutation of Cys399, which is absent in NTPDase2 and -2, produced a protein completely devoid of nucleotidase activity, while mutation of Asn443 to Asp resulted in substantial loss of activity. Neither the Cys399 nor Asn443 mutants were fully glycosylated, and both were retained in the endoplasmic reticulum. These results indicate that the lack of ecto-nucleotidase activity exhibited by NTPDase2 and -2 and the C399S mutant, as well as the large reduction of activity in the N443D mutant are due to alterations in the folding/maturation of these proteins. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The ecto-nucleoside triphosphate diphosphohydrolases (E-NTPDases)1 are a family of ecto-nucleotidases, including ecto-ATPases and ecto-ATPDases (apyrases), that hydrolyze nucleoside 5'-tri- and diphosphates with isozyme-dependent nucleotide selectivities and biochemical properties. The E-NTPDase family includes membrane-associated enzymes (NTPDase1 to NTPDase4) and soluble enzymes (NTPDase5 and NTPDase6) (see Zimmermann (10) for a review). NTPDase2 (also known as ecto-ATPase or CD39L1) exhibits a strong preference for nucleoside triphosphates (NTPs) over diphosphates (nucleoside diphosphates), hydrolyzing nucleoside diphosphates at 3–5% of the rate of NTPs (10–14). NTPDase2 contains two transmembrane domains with very short cytoplasmic amino and carboxyl termini and a large extracellular loop comprising the catalytic site. Five apyrase conserved regions (ACRs) are common to all members of the E-NTPDase family and presumably are important for enzymatic activity (15–17). NTPDase2 requires millimolar concentrations of divalent cations (Ca2+ or Mg2+) and hydrolyzes all naturally occurring nucleoside 5'-triphosphates. Genes encoding NTPDase2 have been cloned from chicken gizzard (13), rat brain (11), mouse hepatoma cells (18), and a human bladder epithelial cell line (14). Sequence analysis of the predicted protein suggests that the human NTPDase2 contains six potential N-glycosylation sites and 10 highly conserved cysteine residues that may be involved in formation of disulfide bonds important for structure and enzymatic activity (17, 19, 20).
A human ecto-ATPase sequence named CD39L1 was reported previously (21). Although enzyme activity was not measured, this sequence exhibited a high degree of homology with NTPDase2 (57–83%) and NTPDase1 (40–50%) from different species. Sequence comparisons of this protein from different species indicated that CD39L1 lacked a 23-amino acid sequence in the putative extracellular loop between ACR4 and ACR5. The full-length human ecto-ATPase was cloned and its activity determined (14). Thus, the possibility of the existence of splice variants for the human ecto-ATPase was established (14). A novel rat ecto-ATPase mRNA was identified that contains an additional 193-bp exon arising from alternative splicing in the 3'-end of the gene (22). Splice variants also were reported for the related ecto-nucleotide pyrophosphatase/phosphodiesterase family (E-NPP), where human phosphodiesterase-I (PD-I) and autotaxin likely represent splice variants of the NPP2 gene (23–26)
In this study we delineate the genomic organization of the human ecto-ATPase gene and describe three splice variants of the human NTPDase2 gene encoding proteins of 495, 472, and 450 amino acids, respectively. Additionally, we demonstrate the importance of Cys399 for correct processing and trafficking to plasma membrane of the human NTPDase2. Absence of this residue is likely responsible for the lack of activity exhibited by the shorter splice variants of the enzyme, probably due to an impaired disulfide bond necessary for maintaining structure of the enzyme.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-D-maltoside was from Dojindo Labs (Gaithersburg, MD). LipofectAMINE2000 reagent was purchased from Invitrogen. KH2PO4 standard solution was from Sigma. Flavobacterium meningosepticium peptide:N-glycosidase F (PNGase F) was obtained from New England Biolabs, Inc. (Beverly, MA). Polyclonal antibodies against the trans-Golgi network (TGN38) were kindly provided by Dr. Sharon Milgram (Department of Cellular and Molecular Physiology at University of North Carolina-Chapel Hill). All tissue culture reagents were from the Lineberger Comprehensive Cancer Center tissue culture facility at the University of North Carolina. Sources of other products were reported previously (27). Cell Culture—ECV304 human bladder tumor epithelial cells were grown in 199 medium supplemented with 10% fetal calf serum, 50 units/ml penicillin, and 50 µg/ml streptomycin. Chinese hamster ovary cells clone K-1 (CHO-K1) were grown in Ham's F-12 nutrient mixture supplemented with 10% fetal calf serum. NIH-3T3 mouse fibroblasts were grown in DMEM supplemented with 10% bovine calf serum and 4.5 g/liter of glucose. All cells were maintained at 37 °C in a humidified atmosphere of 5% CO2, 95% air.
Isolation of Splice Variants of the Human Ecto-ATPase—All DNA and RNA manipulations were carried out according to standard procedures (Sambrook et al. (44). The cDNAs encoding the splice variants of the human ecto-ATPase were obtained by reverse transcriptase-PCR amplification of mRNA from ECV304 cells. Messenger RNA was isolated from ECV304 cultures using the Fast Track 2.0 kit (Invitrogen). First-strand cDNA synthesis was performed using 0.1 µg of mRNA as template and oligo(dT) as primer for the reverse transcriptase (reverse transcriptase-PCR kit, Stratagene, La Jolla, CA). The PCR reaction (100 µl) was developed in the presence of 5.0 Units of Pfu DNA polymerase (Stratagene) and contained 5 µl of the cDNA obtained in the reverse transcriptase reaction as the template and 0.4 µM of each of the amplification primers. The PCR primers were designed to amplify the complete coding sequence of the human ecto-ATPase CD39L1 (GenBankTM accession number U91510 [GenBank] ), and the amplified PCR products were ligated into the retroviral expression vector pLXPIH as we described previously (14). The PCR amplification products were absolutely dependent on the presence of reverse transcriptase, indicating that the amplified sequences originated from mRNA. The amplified sequences contained open reading frames encoding predicted proteins of 495, 472, and 450 amino acids for the products obtained with the primers for CD39L1.
Site-directed Mutagenesis of Human Ecto-ATPase—Mutagenesis of the full-length human ecto-ATPase (NTPDase2) was performed by PCR. The oligonucleotides used for mutagenesis were as follows: Cys399 
Ser mutation: sense, 5'-CGC CTG GCC GAC TAC TCC GCC GGG GCC ATG TTC-3'; antisense, 5'-GAA CAT GGC CCC GGC GGA GTA GTC GGC CAG GCG-3' and N443D mutation: sense, 5'-CTC GGC TAC ATG CTG GAC CTG ACC AAC CTG ATC-3'; antisense, 5'-GAT CAG GTT GGT CAG GTC CAG CAT GTA GCC GAG-3' (the mutated codons are underlined).
Cloning of Human Ecto-ATPase Isoforms and Mutant cDNAs into pcDNA4/Myc-His Expression Vector—cDNAs encoding the three splice variants of the human ecto-ATPase and the two mutants were subcloned into the mammalian expression vector pcDNA4/Myc-His that allows immunodetection of the recombinant proteins with antibodies against Myc or hexa-His epitopes localized at the carboxyl terminus of each of the fusion proteins. To subclone the cDNAs into pcDNA4/Myc-His we took advantage of the EcoRI (5') and XbaI (3') sites of the multicloning site. An XbaI site was generated by PCR in the 3'-end of the coding sequence of the five cDNAs prior to the stop codon and in frame with the Myc and His6 tags in pcDNA4/Myc-His. The cDNAs cloned into the retoviral vector pLXPIH (14) were utilized as the template for the PCR reactions. The sense primer internal to the coding sequences was 5'-TAC CTG GGA GCC ACA GCG GGT A-3', and the antisense primer containing the new XbaI site at the 3'-end of the sequence was 5'-GAG ATC TAG AAT GGT GCT TGG CAG CTT GGC G-3' (the XbaI is underlined). The PCR amplification products of each DNA were digested with BglII (unique restriction site present in the five sequences) and XbaI and ligated with the smaller product of digestion (EcoRI plus BglII) of pLXPIH-ecto-ATPase into pcDNA4/Myc-His previously digested with EcoRI and XbaI. The resulting DNA constructs were amplified and purified (Qiagen). DNAs were sequenced at the University of North Carolina-Chapel Hill Automated DNA Sequencing Facility on a model 377 DNA sequencer (PerkinElmer Life Sciences, Applied Biosystems Division) using the ABI PRISMTM dye terminator cycle sequencing ready reaction kit with AmpliTaq DNA polymerase (PerkinElmer Life Sciences, Applied Biosystems Division).
DNA Genomic Library Screening and Sequencing—To isolate the gene of the human ecto-ATPase we screened a Lambda FIX II genomic library from human placenta (Stratagene). Total genomic DNA was isolated from ECV304 cells (QIAamp blood kit, Qiagen) and used as template for PCR amplification to generate a [32P]dCTP radiolabeled oligonucleotide probe of 
1,300 bp. Primers corresponding to the coding region of CD39L1 were utilized: sense primer, 5'-CAT CGT CCT GGA TGC TGG TTC TTC ACA CAC GTC-3'; antisense primer, 5'-GGA CTG GTT GTC TCA AAA GTG ATC TGG GTA GAG GCA CC-3'. Positive clones were purified from the plates, and phage DNA was amplified by infecting TAP-90 cells. Following cell lysis, phage DNA was obtained using the Lambda DNA Kit (Qiagen). Purified Lambda DNA was sequenced as described above.
Transient and Stable Transfection of CHO-K1 Cells—CHO-K1 cells were plated at 75,000 cells/cm2. The cells were transfected by the LipofectAMINE method 24 h later with empty vector or with pcDNA4/Myc-His plasmid-DNA harboring the cDNAs corresponding to the ecto-ATPase mutants or the three spliced forms of the enzyme. Transfections were according to manufacturer's indications using 0.4 µg of DNA and 1.1 µl of LipofectAMINE2000 per cm2 of tissue culture area. Stable cell lines were isolated after growth of cells in the presence of zeocin (200 µg/ml).
Membrane Preparation—CHO-K1 cells stably expressing each DNA construct were grown on 10-cm culture dishes. Confluent cell monolayers were washed twice with ice-cold PBS, incubated for 20 min at 4 °C in lysis buffer containing protease inhibitors (50 mM NaCl, 20 mM Tris, pH 8.0, 2 µg/ml aprotinin, 2 µg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 1 mM benzamidine), and harvested by scraping in the same lysis buffer. The cells were collected, homogenized in a Potter homogenizer, and centrifuged at 500 x g for 10 min at 4 °C. The supernatant was recovered for further centrifugation, and the pellet was resuspended in lysis buffer, homogenized, and centrifuged. The supernatants were combined and centrifuged at 45,000 x g for 30 min at 4 °C. The resulting pellet was resuspended in lysis buffer and centrifuged under the same conditions. The final pellet was resuspended in 1 ml of lysis buffer supplemented with 250 mM sucrose and protease inhibitors.
Assay of Ecto-ATPase Activity in Intact Cells and Membrane Preparations—Intact CHO-K1 cells stably or transiently expressing the two mutants, the three isoforms of human ecto-ATPase, and vector control were assayed in cells growing in 48-well plates 24 h after transfection for transient expression or after cells reached confluence for stable cell lines. Briefly, the cells were washed once with 1 ml of phosphate-free saline solution (125 mM NaCl, 5.2 mM KCl, 20 mM HEPES, pH 7.4, 2 mM CaCl2, 1.2 mM MgCl2, and 5 mM D-glucose), and incubated at 37 °C in 250 µl final volume of the same medium containing the indicated concentrations of nucleotide. The incubation was terminated by transferring the cell-free supernatants of each well to a new plate containing 50 µl/well of 60 mM EDTA (10 mM final concentration) at 4 °C. The cell supernatants were microwaved for 20 s to inactivate any released nucleotidase activity and kept on ice until assayed for inorganic phosphate content. Enzyme activity in total cell membrane preparations was determined at 37 °C in 200 µl of assay buffer (20 mM Tris, pH 7.4, 2 mM CaCl2, and 1.2 mM MgCl2) containing the indicated amount of nucleotide as the substrate and 0.3–1.0 µg of protein. Reactions were stopped by the addition of 200 µl of ice-cold 20 mM EDTA. Nucleotide hydrolysis was assayed by the release of inorganic phosphate.
Phosphate Assay—Inorganic phosphate was determined colorimetrically using a modification of the malachite green-based assay (28). Supernatants (30 µl) from cell lysates or 100 µl of the membrane fraction were combined with 100 µl of malachite green reagent and the absorbance determined at 595 nm in a plate reader (PerkinElmer Life Sciences, model HTS 7000 bioassay reader). The inorganic phosphate content was determined by comparison against a standard curve constructed with known amounts of phosphate.
Western Blot Analysis—SDS-polyacrylamide gel electrophoreis was performed according to standard procedures (29). Membrane proteins, usually 5–10 µg per lane, were dissolved in reducing Laemmli's sample buffer (2% SDS, 10% glycerol, 0.02% bromphenol blue, 65 mM Tris, pH 6.8, and 5% 2-mercaptoethanol), separated on 5–8% polyacrylamide gels, and electrophoretically transferred to a nitrocellulose membrane (Bio-Rad). The membrane was probed with mouse monoclonal antibody against Myc or hexa-His epitopes in 5% milk in TBS (20 mM Tris, pH 7.2, 120 mM NaCl, 0.1% Tween 20). Bands were visualized by chemiluminescence after incubation with horseradish peroxidase-conjugated goat anti-mouse IgG as the secondary antibody (Pierce).
Immunocytochemistry Analysis—The recombinant proteins were localized by indirect immunofluorescence using the following protocol. CHO-K1 cells stably expressing the Myc- and His6-tagged mutants and splice variants of the human ecto-ATPase were grown in glass coverslips for 24 h, washed with PBS, and fixed for 5–10 min in 4% paraformaldehyde at room temperature. The cells were washed with PBS and permeabilized for 4 min in 100% ethanol at –20 °C and subsequently blocked in blocking solution (12.5% (w/v) normal goat serum in PBS) for 1–2 h at room temperature. Anti-Myc antibody (dilution, 1:500) alone or in combination with TGN38 (dilution 1:800), a polyclonal antibody against vesicles of the trans-Golgi network (30), were added in blocking solution and incubated for 1–2 h at room temperature. After washing with PBS, the cells were incubated for 1 h with fluorescently labeled secondary antibodies: anti-mouse conjugated to Texas Red (dilution 1:200) or anti-rabbit conjugated to fluorescein isothiocyanate (dilution 1:200). In some experiments the cells were co-stained with fluorescein-labeled phalloidin (dilution 1:50) or alexa-concanavalin A (1:300). Both reagents were added together with the secondary antibody. After washing with PBS, the coverslips were mounted in Vectashield medium containing 4,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories Inc., Burlingame, CA). Photomicrographs were taken using a Leica TCS 4D confocal microscope and analyzed using Metamorph software (Universal Image Co.).
Protein Assay—Protein concentrations were determined according to the Bradford method (31) using the Bio-Rad protein assay dye reagent (Bio-Rad). Bovine serum albumin was used as the standard.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
To confirm that the isolated sequences correspond to splice variants of the enzyme, the genomic organization of the ecto-ATPase gene was examined. A human genomic library was screened, and a 9-kb clone encoding the entire coding sequence for the 495-amino acid ecto-ATPase was isolated. This clone was sequenced, and a sequence spanning 6 kb containing nine exons interrupted by eight introns was delineated that encoded the full-length 495-amino acid ecto-ATPase (Fig. 1B). Sequence analysis revealed that the two cDNA sequences encoding the 472 and 450 amino acid proteins correspond to true alternative spliced forms of the human ecto-ATPase gene. That is, we identified the three alternative exon-intron flanking sequences (Fig. 1B) that produced the full-length (human NTPDase2) and the two alternative spliced forms, NTPDase2 and -2. We previously reported that the 495-amino acid form is a functional ecto-ATPase (14). Thus, the cDNA of NTPDase2 (495 amino acids) contains the entire eighth exon (135 bp), NTPDase2 (472 amino acids) contains only the 66 bp at the 3'-end of the eighth exon, and the eighth exon is completely spliced out of NTPDase2 (450 amino acids) (Fig. 1B). NTPDase2 corresponds to the DNA-sequence named CD39L1 reported by Chadwick and Frischauf (21). Splicing of the gene at the intron/exon boundaries indicated in Fig. 1 does not change the open reading frame of any of the three cDNAs reported here.
Expression of Myc- and Hexahistidine-tagged Splice Variants of the Human NTPDase2—To study the expression and enzymatic activity of the splice variants of the human ecto-ATPase (NTPDase2, -2, and -2), we constructed expression vectors that incorporated Myc and hexa-His tags at the COOH terminus of the proteins (see "Experimental Procedures"). Western blot analyses indicated that all proteins were expressed at similar levels after stable and transient expression in CHO-K1 cells (Fig. 2A). Anti-Myc and anti-polyhistidine antibodies detected major bands of 66–70 kDa and a broad and very diffuse band migrating at 75–80 kDa for the full-length recombinant human ecto-ATPase (Fig. 2A). Essentially identical results were obtained in a Western blot of the untagged full-length ecto-ATPase blotted with anti-peptide polyclonal antibodies against residues 387–401 (RVPGQRARLADYCAG) or against the carboxyl terminus (LRQVHSAKLPSTI-COOH) of the human ecto-ATPase. Thus, the epitope tags had no effect on the expression of the enzyme or on its mobility in SDS-PAGE (data not shown). The electrophoretic mobilities of the shorter splice variants of the ecto-ATPase (2 and 2) were reduced as predicted by their protein sequences (23 and 45 amino acids shorter, respectively); interestingly, neither of these splice variants exhibited the more diffuse upper species present in the full-length enzyme (Figs. 2 and 3). Glycosylation apparently accounts for the larger apparent size than that predicted for the core non-glycosylated enzyme (
54 kDa for the full-length non-tagged and 57 kDa for the Myc-His-tagged protein). Consistent with this conclusion, treatment of the three isoforms of the human ecto-ATPase with F. meningosepticium PNGase F resulted in single bands for the three proteins that migrated at the expected size for the non-glycosylated amino acid backbones (Fig. 2B). Taken together, these data suggest that the human ecto-ATPase is initially synthesized as a core-glycosylated precursor that matures into a diffusely migrating fully glycosylated polypeptide, which is known to be a common characteristic of most secretory and plasma membrane glycoproteins (32).
|
|
Ecto-nucleotidase Activity of the Spliced Forms of the Human NTPDase2—We compared the enzymatic activity of the full-length human ecto-ATPase (NTPDase2) with that of the 472 (NTPDase2) and 450 (NTPDase2) amino acid splice variants. Intact cells and membrane preparations from CHO-K1 cells were analyzed for hydrolytic activity 24 h after transient expression of the proteins. CHO-K1 cells transfected with empty vector exhibited negligible ecto-nucleotidase activity during incubations up to 30 min in the presence of ATP (Fig. 3) or ADP (data not shown) as substrates. In contrast, expression of the full-length isozyme (NTPDase2) resulted in marked ecto-nucleotidase activity against nucleoside triphosphates and very little activity against nucleoside diphosphates, as expected for an NTPDase type 2 and as previously observed for the non-tagged enzyme (10, 14). These results indicate that tagging the protein in the COOH terminus does not affect its enzymatic activity or substrate selectivity. Under the conditions of this study, CHO-K1 cells transiently or stably expressing the full-length ecto-ATPase hydrolyzed more than half of the substrate ATP in a 5-min incubation period (85.8 ± 7.5 nmol of Pi/min/106 Pi/min/106 cells for ATP versus 4.7 ± 0.5 nmol of cells for ADP). Under the same conditions, measurable nucleotidase activity was not detected in cells expressing NTPDase2 and NTPDase2, although these proteins were expressed at the same level as the full-length protein (Fig. 3, inset). Identical results were observed after expression of the isozymes in NIH-3T3 mouse fibroblasts (data not shown). Taken together, these results suggest that the region of the human ecto-ATPase between residues 384 and 406 contains important structural/functional determinants essential for enzymatic activity.
Identification of Important Cysteine and Asparagine Residues in the Human Ecto-ATPase—The sequence of the full-length human ecto-ATPase was compared with other family members to reveal possible explanations for the lack of activity observed in the NTPDase2 and -2 splice variants. Both splice variants retain the five ACRs known to be important for catalytic activity. However, other possibilities for lack of activity in the splice variants were considered. The human NTPDase2 contains 10 extracellular cysteine residues (amino acids 75, 99, 242, 265, 284, 310, 323, 328, 377, and 399) that are invariantly conserved among the NTPDases type 1 (ecto-apyrases) and type 2 (ecto-ATPases) from different species and that likely are involved in formation of intra- or interchain disulfide bonds important for the structure and function of these enzymes (17, 19, 20). One of these conserved cysteines, Cys399, is absent in the 472 (NTPDase2) and 450 (NTPDase2) amino acid splice variants of the human ecto-ATPase (Fig. 4). To assess the functional significance of this highly conserved residue and to determine whether its absence is responsible for the lack of activity of the shorter splice forms of the human ecto-ATPase, we mutated the cysteine at the 399 position to a serine, generating the mutant C399S.
|
Sequence analysis of the human ecto-ATPase also revealed that only two (Asn64 and Asn443) out of the six potential N-glycosylation sites (Asn64, Asn88, Asn129, Asn294, Asn378, and Asn443) are conserved among all ecto-ATPases (NTPDase2) from different species (human, mouse, rat, and chicken). N-Linked glycosylation is a common co- and posttranslational modification in many plasma membrane proteins (32), and it likely is important for the function and oligomerization of E-type ATPases and apyrases (7, 33–35). Since the glycosylation pattern of the shorter splice variants clearly differs from the full-length protein (Figs. 2 and 3), and since the proximity of an aspartic acid (Asp65) to Asn64 makes this site improbable as a glycosylation site (36), Asn443 was subjected to site-directed mutagenesis to assess the role of N-glycosylation in ecto-ATPase activity (Fig. 4). Thus, the asparagine at position 443 in the full-length human ecto-ATPase was replaced with an aspartic acid, generating the mutant N443D.
Expression of the C399S and N443D Mutants of the Human Ecto-ATPase in CHO-K1 Cells—The C399S and N443D mutants of the human ecto-ATPase were produced and transiently expressed in CHO-K1 cells. Equivalent levels of expression for wild-type and mutant proteins were observed (Fig. 5A). The C399S mutant exhibited an electrophoretic pattern similar to that of the wild-type enzyme with the exception that the more diffuse upper bands migrating at 75–80 kDa were not observed. Interestingly, the N443D mutant migrated with the same mobility as the smaller of the two major bands (doublet at 66–70 kDa) observed in the wild-type enzyme and the C399S mutant (Fig. 5A). As observed for C399S, the diffuse bands at 75–80 kDa were absent in the N443D mutant (Figs. 5 and 6). However, the apparent size of the N443D mutant was higher than the predicted size of the amino acid backbone, suggesting the presence of other glycosylated Asn residues. Consistent with this observation, PNGase F treatment of membrane preparations from cells expressing the wild-type enzyme or the mutant proteins (Fig. 5B), or incubation of intact cells with tunicamycin (data not shown), resulted in a unique band at the expected size for the non-glycosylated core proteins (57 kDa). These results indicate that Asn443 is a major N-glycosylation target and that this site is normally glycosylated in the two splice variants of the human ecto-ATPase.
|
|
Enzymatic Activity of the C399S and N443D Mutants of the Human Ecto-ATPase—Mutation of Cys399 to serine resulted in a protein completely devoid of ecto-ATPase activity (Fig. 6), although no substantial differences were observed in protein expression levels between the mutant and the fully active wild-type enzyme (Fig. 6, inset). The N443D mutation resulted in an enzyme with an ecto-ATPase activity that was 7% of that observed with the wild-type enzyme (5.8 ± 0.8 nmol of Pi/min/106 cells versus 88.1 ± 17.4 nmol of Pi/min/106 cells for the wild-type enzyme with ATP as the substrate). Expression levels of the N443D mutant were similar to those of the wild-type enzyme (Fig. 6, inset).
Subcellular Localization of Splice Variants and C399S and N443D Mutants of the Human Ecto-ATPase—Failure to target to the plasma membrane could explain the lack/reduction of activity of the splice variants (NTPDase2 and -2) and the C399S and N443D mutants of the human ecto-ATPase. An alternative explanation is that these proteins undergo structural modification or misfolding but are nonetheless normally targeted to the cell surface. To address these possibilities, the subcellular localization of the ecto-ATPases was determined after stable expression in CHO-K1 cells. The wild-type NTPDase2 exhibited a similar distribution as that of cortical actin (which delimits the plasma membrane) detected by phalloidin staining (Fig. 7). The wild-type enzyme also exhibited a dense perinuclear staining that co-localized with trans-Golgi vesicles (Fig. 8), and a low level signal for the NTPDase2 also overlapped with the endoplasmic reticulum (ER) marker concanavalin A (Fig. 8). Thus, most of the wild-type protein was present either at the plasma membrane or in trans-Golgi vesicles likely in transit to the plasma membrane. In contrast, neither of the shorter splice variants of the human ecto-ATPase, NTPDase2 and -2, were localized at the plasma membrane (Fig. 7). Both splice variants were entirely intracellular, primarily in the ER (Fig. 7), since no co-localization of NTPDase2 or -2 with trans-Golgi vesicles was observed (data not shown). These results suggest that NTPDase2 and -2 displayed impaired trafficking to the plasma membrane, most likely due to folding alterations that lead to retention in the ER. The C339S and N443D mutants also accumulated intracellularly (Fig. 8). Both mutant proteins exhibited no expression at the plasma membrane (Fig. 8), or in the trans-Golgi network (data not shown), and were found mostly in the ER (Fig. 8).
|
|
Enzyme Solubilization—The subcellular localization analyses suggested that the lack/reduction in activity of the shorter splice variants and mutants of the human ecto-ATPase may be due to failure of these proteins to reach the plasma membrane. For example, the most common mutation (
F508) of the cystic fibrosis transmembrane conductance Cl– channel regulator, results in a protein that retains activity although trapped in the ER due to structural misfolding (37). Little ecto-ATPase activity was observed in membranes of vector-transfected cells when compared with the activity of cells expressing the ecto-ATPase (data not shown). Detergent solubilization experiments were carried out to determine whether the forms of the human ecto-ATPase that did not target the plasma membrane are nonetheless active. We initially tested whether ecto-ATPase activity of the full-length wild-type enzyme was retained after solubilization of membranes with Triton X-100, Nonidet P-40, digitonin, CHAPS, or dodecyl maltoside (Table I). Only digitonin preserved the enzymatic activity when compared with non-solubilized membranes. Indeed, an increase of ecto-ATPase activity occurred in the presence of digitonin, suggesting that this detergent releases significant activity enclosed in membrane vesicles or that is not accessible to the substrate in intact cell assays. Extraction of equivalent amounts of protein after solubilization with digitonin (Fig. 9B) did not uncover significant enzymatic activity of the spliced forms NTPDase2 and -2 or of the ecto-ATPase mutants (Fig. 9A) or of vector-transfected control cells (data not shown). These results indicate that the lack/reduction in enzymatic activity of the shorter splice variants and mutants of the human ecto-ATPase is not only due to lack of presence at the cell surface but also due to structural alterations that result in loss of enzymatic activity.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
, 2, and 2). Our results show that the region missing in the shorter splice variants 2 and 2 is essential for correct folding, trafficking, and enzymatic activity of the human NTPDase2. Importantly, we demonstrate that Cys399, which is absent in the shorter isoforms of the human NTPDase2, accounts for the inactive intracellular phenotype shown by 2 and 2. Accordingly, mutation of this residue to a serine in the full-length active isoform (2) recapitulates the defects in processing and enzymatic activity exhibited by the shorter splice variants. Our results suggest that Cys399 is involved in an intra-chain disulfide bond that is essential for acquisition of tertiary structure and consequently for enzymatic activity of the human NTPDase2.
We recently described the isolation of a human cDNA encoding a predicted 495 amino acid protein (14). This sequence was found identical in length and highly homologous to ecto-ATPases previously reported from rat brain (11), mouse hepatoma cells (18), and chicken muscle (13). In the current work two other human cDNAs were isolated encoding predicted proteins of 472 and 450 amino acids. Sequence analysis of a human genomic DNA clone indicated that these three sequences correspond to alternative splice variants of the human ecto-ATPase gene (Fig. 1). The missing regions in NTPDase2 and -2 are the result of differential splicing of exon eight and correspond to the extracellular domain of the protein near the fifth ACR and the COOH terminus. Although the shorter splice variants of the human ecto-ATPase contain the five ACRs proposed to be essential for enzymatic activity (15–17), no ecto-ATPase activity was observed when NTPDase2 and -2 were heterologously expressed in CHO-K1 or NIH-3T3 cells to levels similar to that of the full-length active enzyme.
The physiological significance of multiple mRNAs for the human ecto-ATPase, some of which do not produce active enzyme when expressed alone, is unknown. Alternative splice variants also have been described for the rat NTPDase2 (22). Although it is not known whether alternative messages for the rat ecto-ATPase are translated into functional enzyme, the occurrence of alternative splicing in multiple species suggests it to be a common mechanism for regulation of expression of these enzymes. Since E-NTPDases form active oligomers (7, 19, 33–35), the shorter splice variants of the human NTPDase2 potentially form hetero-oligomers with the full-length active isozyme (2) generating biochemical characteristics different from those of the homo-oligomeric forms. For example, two splice variants of a human serotonin receptor 5-HT3A, referred to as truncated (h5-HT3AT) and long (h-5HT3AL), did not assemble into functional homomeric ion channels when expressed individually (38). However, co-expression of either splice variant with the active h5-HT3A subunit resulted in heteromeric assembly and in significantly modified 5-HT3A receptor function. Considering that the message for NTPDase2 is the most abundant in ECV304 cells (data not shown), its co-assembly with 2 and/or 2 might serve as a mechanism for modification of enzymatic activity relative to the oligomeric form composed exclusively of NTPDase2 subunits. Tissue-specific splicing patterns yielding hetero-oligomers with varying monomeric compositions could result in diversification in the termination of nucleotide signaling.
Analysis of the subcellular localization of the three splice variants using confocal microscopy indicated that only NTPDase2 is delivered to the plasma membrane, with NTPDase2 and -2 mainly localized in the ER. Correct folding and oligomerization of newly synthesized membrane proteins are prerequisites for export from the ER. In general, misfolded polypeptides or unassembled subunits of oligomeric proteins are retained in this organelle and ultimately degraded (39, 40). Our data showing lack of enzymatic activity of NTPDase2 and -2 in both intact and digitonin-solubilized cells and retention of these variants in the ER strongly suggest that the shorter splice variants lack a structural determinant important for correct folding, activity, and targeting to the plasma membrane of the human ecto-ATPase.
The human ecto-ATPase contains 10 extracellular cysteine residues that are absolutely conserved among all type 1 (apyrases) and type 2 (ecto-ATPases) E-NTPDases. The conserved cysteine residues in E-NTPDases are likely involved in intra- and/or interchain disulfide bonds based both on structure-function analysis (17, 19, 41) and on their stability to proteolysis and inhibition of enzymatic activity by dithiothreitol (13, 19). Additionally, the cysteine-rich region between ACR4 and ACR5 (eight conserved cysteines) in NTPDase1 and -2 recently was proposed to be important for conferring tertiary structure, and probably catalytic specificity, to these proteins (20). The region missing in the shorter splice variants NTPDase2 and -2 contains one of these conserved cysteines, Cys399, and replacement of this residue by an amino acid unable to form disulfide bonds produced a completely inactive protein (Figs. 6 and 9). Lack of activity was accompanied by loss of the more diffuse upper bands at around 75–80 kDa observed after SDS-PAGE and by a clear ER retention phenotype with no enzyme detected at the plasma membrane. These features of the C399S mutant resemble those found for the splice variants 2 and 2 and strongly suggest that Cys399 is the structural determinant absent in 2 and 2 that is essential for correct folding and trafficking to the plasma membrane. In general, correct folding of proteins in the ER requires the early formation of intra-chain disulfide bonds that help to stabilize domain structure, so that cysteine mutations are frequently damaging, and lack of native disulfide bonds might lead to protein retention (39, 40, 42). Thus, our results suggest that Cys399 is implicated in the formation of an intramolecular disulfide bond essential for protein structure and function of the human NTPDase2 and give explanation for the lack of activity found in the shorter splice variants NTPDase2 and -2.
In addition to the absence of Cys399 in the shorter splice variants of the human ecto-ATPase and the consequent structural/functional alterations, a different glycosylation pattern is evident in these forms when compared with the wild-type NTPDase2 (Figs. 2 and 3). Therefore, the effect of N-glycosylation on enzymatic activity of the human ecto-ATPase also was investigated, since the lack of activity observed in NTP-Dase2 and -2 might result from of an impaired glycosylation as a downstream consequence of an altered disulfide bond. Glycosylation contributes to correct folding of glycoproteins by stabilizing protein conformation and preventing the unfolded or partially folded protein molecules from aggregation (7, 32). N-Linked oligosaccharides attach to asparagines at Asn-X-Ser/Thr sequences co-translationally as the polypeptide chains are translocated across the ER membrane. Core-glycosylated proteins then are trimmed by glucosidases and mannosidases in the ER and Golgi and are further modified in the Golgi apparatus to full glycosylation (32, 36). Although E-NTPDases are heavily glycosylated enzymes, the role of N-glycans in enzymatic activity has remained unclear. Some studies support an importance of glycosylation for enzymatic activity of E-NTPDases (33–35), while others report little or no effect (17, 23).
The human NTPDase2 possesses six potential consensus sites for N-glycosylation (14), all of which are present in the three splice variants. Only two of these, Asn64 and Asn443, are invariantly conserved among all cloned NTPDases1 and -2. Since Asn64 (Asn-Asp-Thr) presents a less susceptible site for glycosylation than Asn443 (Asn-Leu-Thr) due to the presence of an Asp immediately after Asn64 (36), we studied the potential role of Asn443. Substitution of Asn443 with an aspartic acid yielded a protein that was expressed at similar levels to wild-type enzyme but that retained only 7% of the hydrolytic activity. We further demonstrated that Asn443 is an actual target for N-glycosylation in the human ecto-ATPase, since the N443D mutant lacked one of the major bands of the doublet at 66–70 kDa present in the wild-type enzyme (Figs. 5 and 6). Asn443 is normally glycosylated in NTPDase2 and -2, since the doublet was present in the shorter isoforms. The N443D mutant apparently was recognized as misfolded and was retained in the ER. However, despite impaired trafficking, at least partial activity (7%) was observed. Although undetectable by means of the immunofluorescence technique utilized here, we anticipate that this active protein reached the plasma membrane. Either the mutant protein partially "escapes" the quality control of the ER, or perhaps the N443D mutation causes a decrease in the efficiency of folding and maturation of the core-glycosylated primary translation product as has been shown for F508 cystic fibrosis transmembrane regulator (43). Taken together, these results indicate that full glycosylation at position 443 is essential for correct folding and trafficking of NTPDase2 to the plasma membrane.
In conclusion, we report here the importance of correct processing and enzymatic activity of two highly conserved residues in the human NTPDase2, Cys399 and Asn443, and give explanation for the lack of activity of the shorter splice variants of the enzyme. The physiological significance of multiple splice forms of the human ecto-ATPase is under investigation.
|
FOOTNOTES |
|---|
To whom correspondence should be addressed: Dept. of Pharmacology, CB #7365, Mary Ellen Jones Bldg., University of North Carolina, Chapel Hill, NC 27599-7365. Tel.: 919-966-4816; Fax: 919-966-5640; E-mail: tkh{at}med.unc.edu.
1 The abbreviations used are: E-NTPDase, ecto-nucleoside-triphosphate diphosphohydrolase; ACR, apyrase conserved regions; CD39, lymphoid cell activation antigen 39 (apyrase); CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; CHO-K1, Chinese hamster ovary cells clone K-1; ER, endoplasmic reticulum; PBS, phosphate-buffered saline; PNGase F, peptide:N-glycosidase F; TBS, Tris-buffered saline; TGN, trans-Golgi network.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
