Posttranslational Modification of Glycosylphosphatidylinositol (GPI)-Specific Phospholipase D and Its Activity in Cleavage of GPI Anchors

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.

251, 737–743 (1998)

RC989542

Posttranslational Modification of Glycosylphosphatidylinositol (GPI)-Specific Phospholipase D and Its Activity in Cleavage of GPI Anchors Hiroshi Tujioka, Yoshio Misumi, Noboru Takami,* and Yukio Ikehara1 Department of Biochemistry and *Radioisotope Laboratory, Fukuoka University School of Medicine, Jonan-ku, Fukuoka 814-0180, Japan

Received September 9, 1998

Human glycosylphosphatidylinositol-specific phospholipase D (GPI-PLD) was exogenously expressed in mammalian CHO cells and in insect H5 cells. GPI-PLD was initially synthesized as a 105-kDa form and then secreted as a mature 115-kDa form from the CHO cells, whereas it was secreted as an immature 98-kDa form from the H5 cells. The difference of the molecular forms was caused by its oligosaccharide processing in the two cell lines. These forms showed a different reactivity to anti-C-terminal peptide of GPI-PLD; the 105-kDa and 98-kDa forms were directly recognized by the antibodies, whereas the 115-kDa form was immunoreactive only after being denatured. In an in vitro assay, the 98-kDa form but not the 115-kDa form was able to release a significant amount of GPI-anchored proteins from intact membranes, although the two forms had the same GPI-anchor cleavage activity in the presence of detergents. In addition, a GPIanchored protein, when coexpressed in CHO cells, was intracellularly cleaved by GPI-PLD in the secretory pathway. Taken together, these results suggest that GPI-PLD undergoes a conformational change by posttranslational modification, which affects its immunoreactive and enzymatic properties. © 1998 Academic Press

Glycosylphosphatidylinositol (GPI)-anchored proteins are widely distributed in nature from yeast to mammals. All the proteins that are proteolytically processed and transferred to the GPI immediately after their synthesis in the endoplasmic reticulum (ER) are transported to the cell surface (1–3). It is well known 1

To whom correspondence and reprint requests should be addressed. Fax: 81-92-864-3865. E-mail: [email protected]. Abbreviations used: Endo H, endo-b-N-acetylglucosaminidase H; ER, endoplasmic reticulum; GPI, glycosylphosphatidylinositol; GPIPLD, GPI-specific phospholipase D; PI, phosphatidyl-inositol; PIPLC, PI-specific phospholipase C; PLAP, placental alkaline phosphatase; PAGE, polyacrylamide gel electrophoresis.

that the GPI-anchored proteins are easily released from the cell surface by phosphatidylinositol (PI)specific phospholipase C (PI-PLC) purified from bacteria (4, 5), which has been used for identification and characterization of the GPI-anchored proteins, although the enzyme is not specific for GPI. GPI-specific PLC was isolated from trypanosomes and characterized in detail (6 –9). Although other GPI-hydrolyzing phospholipase C activities were described in rat liver (10) and mouse brain (11), the enzymes responsible for these activities have not been characterized in detail. In mammals, the only purified and well-characterized GPI-specific phospholipase is a D-type phospholipase (GPI-PLD). GPI-PLD, a 115-kDa protein, is present in large amounts in mammalian plasma and is capable of cleaving the inositol phosphate linkage of GPI-anchored proteins but has little or no activity toward PI or phosphatidylcholine (12–15). It is of interest to note that GPI-PLD readily hydrolyzes detergentsolubilized GPI-anchored proteins but is not active towards these substrates anchored to intact membranes (16, 17). On the other hand, it was demonstrated that when GPI-anchored proteins were co-expressed with GPI-PLD in COS-1 cells by transfection, the GPIanchored proteins were intracellularly hydrolyzed by GPI-PLD (18, 19). The observation suggests that an intracellular compartment may provide a better environment for cleavage of the GPI-anchor by the newly synthesized enzyme. Since the membrane environments encountered by GPI-PLD and its potential substrates early in the secretory pathway are believed to be different from those on the cell surface, it is possible that GPI-anchor cleavage only takes place at an intracellular location on proteins which are transit to the cell surface. In the present study we have examined the biosynthesis and post-translational processing of GPI-PLD that was expressed in mammalian or insect cells, demonstrating that an immature form of the enzyme is

737

0006-291X/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.

Vol. 251, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

able to cleave the GPI anchors on membranes. We also examined the expression site of GPI-PLD by Northern blot analysis of human tissues. MATERIALS AND METHODS Materials. [35S]Methionine and a-[32P]dCTP were purchased from Amersham Corp. (Tokyo, Japan); endo-b-N-acetylhexosaminidase H (Endo H) from Seikagaku Kogyo (Tokyo); neuraminidase (Arthrobacter ureafaciens) from Nakarai Chemicals (Kyoto); tunicamycin from Sigma Co. (St. Louis, MO); Lipofectin reagent and G418 from GIBCO-BRL (Gaithersburg, MD); PI-specific phospholipase C (PI-PLC) from Funakoshi (Tokyo). Anti-human placental alkaline phosphatase (PLAP) IgG was raised in rabbits and purified as described (1, 2). Isolation and sequencing of human GPI-PLD cDNA. Two cDNA fragments of bovine GPI-PLD (18) were prepared and used as probes for screening of a human liver cDNA library in lZAPII bacteriophage. Screening of 1 x 106 independent clones yielded 8 positive clones. These clones were subcloned into pBluescript SK2 plasmid vector by automatic excision process (20). A cDNA clone with the longest insert (4.5 kbp) was subcloned and subjected to nucleotide sequence determination by the dideoxynucleotide chain termination method (21). Expression of GPI-PLD in mammalian cells. An EcoRI-EcoRI fragment (3.2 kbp) of GPI-PLD cDNA containing the entire coding region of the enzyme was inserted into the EcoRI site of pSG5 expression vector and the resulting plasmid pSG5/GPI-PLD was transfected into COS-1 cells for transient expression of GPI-PLD (19). The 3.2-kbp EcoRI-EcoRI fragment was inserted into pBluescript KS2 plasmid and re-excised at XhoI-NotI sites. The XhoI-NotI fragment containing the GPI-PLD cDNA was inserted into the XhoINotI sites of BCMGSneo (22). The resulting plasmid BCMGSneo/ GPI-PLD was transfected into CHO cells, which were maintained in a G418-supplemented medium at 37°C for isolation of a cell line stably expressing GPI-PLD (CHO/GPI-PLD). An expression plasmid of PLAP (pSG5/PLAP) was also prepared and transfected into COS-1 cells as described previously (23). Expression of GPI-PLD in insect cells. The 3.2-kbp EcoRI-EcoRI fragment of GPI-PLD cDNA was inserted into an EcoRI site of the transfer vector pVL1393 (pVL1393/GPI-PLD). Spodoptera frugiperda Sf 21 cells were maintained at 27°C in a TNM-FH insect medium with 10% fetal bovine serum. The plasmid pVL1393/GPIPLD (1 mg/dish) was cotransfected by the Lipofectin method (24) with 10 ng of BaculoGold DNA (PharMingen), which was employed as an Autographa californica nuclear polyhedrosis viral genome, into 2 x 106 Sf 21 cells in a 60-mm dish. Recombinant viruses in media from several dishes were collected 4 days after infection. The generated recombinant viruses were purified by a single round of plaque assay and amplified to .107 plaque-forming units (pfu)/ml. High-Five (H5) cells were infected with the recombinant viruses at multiplicity of infection (moi) of 10 in SF900II-SFM serum-free medium. The cells cultured for 24 h after infection (H5/GPI-PLD) were used for analysis of GPI-PLD. For partial purification of the expressed GPI-PLD, 5 x 108 H5 cells in 20 flasks of 175-cm2 were infected with 5 x 108 pfu of the recombinant viruses (moi of 10) and media collected 4 days after infection were used. Production of antibodies to the human GPI-PLD C-terminus. The synthetic peptide GCSL-GARLSGALHVYSLGSD, corresponding to the C-terminal sequence of human GPI-PLD (positions 823– 840), was used for raising antibodies in rabbits. The peptide (2 mg) was conjugated to keyhole limpet haemocyanin (5 mg) (23) and the conjugates (300 mg/rabbit) were injected 4 –5 times into two rabbits every 2 weeks. Antisera obtained from the rabbits were used as anti-GPI-PLD antibodies.

Metabolic labeling, immunoprecipitation, and SDS–PAGE. CHO/ GPI-PLD cells (5 x 106 cells/60-mm dish) were incubated at 37°C for 30 min with [35S]methionine (4 MBq/dish) in 2 ml of minimum essential medium lacking methionine and chased in 2 ml of complete medium. H5/GPI-PLD cells (5 x 106 cells/60-mm dish) were labeled at 27°C for 30 min with [35S]methionine (4 MBq/dish) in 2 ml of methionine-free insect medium and chased in 2 ml of the complete insect medium. When indicated, cells were incubated with 5 mg/ml of tunicamycin for 4 h followed by labeling with [35S]methionine (4 MBq/dish) (25). At the indicated times of chase, cell lysates and culture medium were prepared in a solution containing 1% Triton X-100, 1% sodium deoxycholate, 0.01% or 1.0% SDS and a protease inhibitor mixture (1) and subjected to immunoprecipitation with anti-GPI-PLD or anti-PLAP antibodies in combination with Protein A-Sepharose (26). The immunocomplexes were boiled in Laemmli’s sample buffer and analyzed by SDS–PAGE (7.5 % gel) and fluorography (27). When indicated, the immunocomplex was incubated with Endo H (0.2 U/ml) or neuraminidase (0.2 U/ml) in 50 mM acetate buffer (pH 5.5 or 5.0, respectively) at 37°C for 18 h (1, 25) before being analyzed by SDS–PAGE. Apparent molecular masses were determined as described previously (1, 25). In vitro assay of GPI-anchor cleavage activity of recombinant GPIPLD. GPI-PLD secreted by CHO or H5 cells was partially purified by affinity chromatography through a concanavalin A-Sepharose column. A substrate for GPI-PLD was prepared as follows. COS-1 cells expressing GPI-PLD were pulse-labeled with [35S]methionine for 30 min and chased for 4 h. A postnuclear fraction prepared from the cells were separated by centrifugation at 105,000 x g for 1 h into a pellet (membranes) and a supernatant (cytosol). The membranes were resuspended in 20 mM Tris–HCl (pH 7.4) containing 0.1 mM CaCl2. Equal amounts of the membrane suspension were incubated at 37°C for 2 h with the indicated GPI-PLD preparation with or without 1% Triton-X 100. The samples were recentrifuged at 105,000 x g for 1 h, and the resulting supernatants were subjected to immunoprecipitation with anti-PLAP antibodies and analyzed by SDS– PAGE. Quantitative analysis of released PLAP was performed by NIH Image software (28). GPI-anchor cleavage activity of GPI-PLD was expressed as percentages of that exhibited by PI-PLC. Immunoblotting. Proteins separated by SDS–PAGE were transferred onto a polyvinylidene difluoride membrane (Millipore), followed by incubation with anti GPI-PLD antibodies (1:1000). Peroxidase-conjugated anti-rabbit IgG antibodies (1:2000) were used as secondary antibodies. The immunoreactive proteins were visualized using the Enhanced Chemiluminescence kit (Amersham Corp.). Northern blot analysis. Poly (A)1 RNAs from various human tissues were obtained from CLONTECH Laboratories (Palo Alto, CA). An EcoRI-HindIII fragment (831 bp) and a HindIII-PstI fragment (766 bp) excised from pSG5/GPI-PLD were labeled with a-32PdCTP by the random primed DNA labeling method and used as probes. Poly(A)1 RNAs (2 mg/lane) were separated by SDS-PAGE and transferred to a membrane, followed by hybridization with the mixture of 32P-labeled probes as described previously (26).

RESULTS We isolated the cDNA for human GPI-PLD from a liver cDNA library. The open reading frame encodes an 840-amino-acid protein of 92-kDa which has 81.6% homology to the bovine enzyme (13) and contains 10 potential sites for N-linked glycosylation. Immunoreactivity of GPI-PLD expressed in CHO cells and H5 insect cells. We used the cDNA for expression of GPI-PLD in mammalian CHO cells and in insect H5 cells, establishing a CHO cell line with stable

738

Vol. 251, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

results suggest that the matured 115-kDa form which is secreted from the CHO/GPI-PLD cells is immunoreactive with our anti-C-peptide antibodies only after the protein is denatured, in contrast to that secreted from the insect cells.

FIG. 1. Immunodetection of secreted forms of GPI-PLD. CHO/ GPI-PLD cells (A) and H5/GPI-PLD insect cells (B) were pulselabeled with [35S]methionine for 30 min. At the indicated times of chase, cells and medium were separated and subjected to immunoprecipitation with anti-GPI-PLD in the absence of SDS. The immunoprecipitates were analyzed by SDS-PAGE (7.5% gels) and fluorography. (C) Immunoblotting with anti-GPI-PLD. Media from a 24-h culture with wild-type CHO cells (lane 1) and CHO/GPI-PLD cells (lane 2) were analyzed by immunoblotting (20 mg protein/lane). (D) A medium obtained after 35S-labeled CHO/GPI-PLD cells were chased for 5 h was subjected to immunoprecipitation with anti-GPI-PLD after the following treatments; lane 1, no treatment; lane 2, 0.1% SDS; lane 3, 1% SDS/boiling followed by addition of 2% Triton-X 100. The immunoprecipitates were analyzed as described in A and B. Apparent molecular masses of GPI-PLD forms (in kDa) were determined as described (1, 25).

expression of GPI-PLD (CHO/GPI-PLD) and H5 cells with transient expression of the enzyme by infection with recombinant baculoviruses (H5/GPI-PLD). The biosynthesis and secretion of GPI-PLD were analyzed by pulse-chase experiments. In the insect H5/GPI-PLD cells, GPI-PLD was initially synthesized as a 105-kDa form and then processed into a 98-kDa form, which was secreted into the medium (Fig. 1B). In contrast, no form of GPI-PLD was detectable in the medium of CHO/GPI-PLD cells, although the same 105-kDa form as in the H5 cells was immunoprecipitated from CHO/ GPI-PLD cell lysates (Fig. 1A). A 92-kDa form may be an unglycosylated form of GPI-PLD, as explained later. Immunoblot analysis, however, demonstrated that a 115-kDa molecule, corresponding to the mature form of the enzyme, was present in the medium collected from a 24-h culture of CHO/GPI-PLD cells (Fig. 1C, lane 2). In addition, when the medium sample obtained at 5 h of chase was boiled with 1% SDS and then used for immunoprecipitation, the 35S-labeled 115-kDa form was identified in the medium (Fig. 1D, lane 3). These

Posttranslational processing of GPI-PLD. Based on the above findings, pulse-chase experiments were carried out again to see a difference in the posttranslational modification of GPI-PLD between CHO and H5 cells. Under the new conditions for immunoprecipitation, it was found that the 105-kDa form synthesized in CHO cells was converted to the 115-kDa form, which was secreted into the medium (Fig. 2A, left panel). When the samples were digested with Endo H, the 105-kDa form was converted to a form of 92-kDa that migrated to the same position of GPI-PLD prepared from tunicamycin-treated cells (TM), whereas the 115kDa form did not change its mobility (Fig. 2B, left panel). The results indicate that the 105-kDa form contains Endo H-sensitive high-mannose type oligosaccharides, which are processed to complex-type sugar chains resistant to Endo H during intracellular transport, resulting in the production of the mature 115-kDa form. When being expressed in insect H5 cells, however, the newly synthesized GPI-PLD did not increase the molecular mass during the chase times; the 105-kDa form was converted to the 98-kDa form by processing of its oligosaccharides from the high-mannose type to the complex type, as judged from the response to Endo H (Fig. 2A and B, right panels). In addition, a different response to neuraminidase was observed between the three forms (Fig. 2C). Although the 105-kDa form in both the CHO and H5 cells was insensitive to neuraminidase, the 115-kDa mature form secreted from the CHO cells was decreased its molecular mass to 107kDa, whereas the 98-kDa form secreted from the H5 cells was not affected by treatment with neuraminidase, indicating that the latter form contains no sialic acid in the oligosaccharides. The molecular mass of the secreted form in H5 cells is still smaller than that of the desialylated one from CHO cells, suggesting that the oligosaccharides of the 98-kDa form are not modified by not only sialylation but also other terminal glycosylation. Thus, it is evident that in the insect cells GPI-PLD is secreted as an immature form, although it is resistant to Endo H, clearly different from the 105kDa form. Both the immature forms of 105 kDa and 98 kDa are directly recognized by anti-C-peptide antibodies, whereas the mature 115-kDa form is immunoreactive with the antibodies only after being denatured (see Fig. 1). In vitro GPI-anchor cleavage activity of recombinant GPI-PLD. We then examined the GPI-anchor cleavage activity of GPI-PLD, for which the recombinant GPI-PLDs secreted from the CHO and H5 cells were

739

Vol. 251, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 2. Biosynthesis and processing of human GPI-PLD. CHO cells and H5 insect cells, both expressing GPI-PLD, were pulse-labeled with [35S]methionine for 30 min, and chased. At the indicated times of chase, cell lysates (cell) and media (med) were prepared and immunoprecipitated with anti-GPI-PLD directly (in H5 cells) or after being denatured with 1% SDS (CHO cells). The immunoprecipitates before (A) and after treatment with Endo H (B) were analyzed by SDS–PAGE (7.5%) and fluorography. TM indicates a sample from tunicamycin-treated cells. (C) Cell lysates and media prepared from 5-h chased cells were immunoprecipitated with anti-GPI-PLD. The immunoprecipitates before (2) and after treatment with neuraminidase (1) were analyzed by SDS–PAGE (7.5 %) and fluorography.

used after being partially purified. Membranes containing placental alkaline phosphatase (PLAP), a wellcharacterized GPI-anchored protein (1, 2), were used as a substrate after being labeled with [35S]methionine. The membrane form (mf) of PLAP directly extracted with SDS (Fig. 3A, membrane) or with Triton X-100 (Fig. 3A, lane 7) has an apparent molecular mass of 66 kDa in SDS-PAGE, although PLAP was not so efficiently extracted from the membranes with Triton X-100. When the membranes were incubated with bacterial PI-PLC, more than 98% of 35S-labeled PLAP was released into the medium and the released soluble form (sf) was found to migrate slightly slower than the membrane form (Fig. 3A, lane 2), as previously reported (1, 2). Incubation of the membranes with the recombinant GPI-PLD from CHO cells (Fig. 3A, lane 3) as well as with buffer alone (lane 1) did not cause significant release of PLAP into the medium. In contrast, treatment with the recombinant enzyme from the insect cells released a significant amount of PLAP from the membranes (Fig. 3A, lane 4). On the other hand, when the reaction mixture was incubated in the presence of Triton X-100, PLAP was almost completely released from the membranes by the recombinant GPI-

PLD from both the CHO and H5 cells (Fig. 3A, lanes 5 and 6). The soluble form of PLAP released from the membranes was quantified in comparison with that released by PI-PLC (Fig. 3B). Intracellular release of PLAP by coexpressed GPIPLD. To assess intracellular GPI-PLD activity, we prepared CHO cells coexpressing GPI-PLD and PLAP in comparison with cells expressing PLAP alone. The expressed PLAP was analyzed by pulse-chase experiments with [35S]methionine. When PLAP was expressed alone, the newly synthesized PLAP was remained associated with the cells and not released into a medium (Fig. 4A). When PLAP was co-expressed with GPI-PLD, a substantial amount of PLAP was released as a mature form of 68 kDa into the medium (Fig. 4B). There is a possibility that GPI-PLD secreted from the cells may cleave the surface-expressed PLAP. To rule out this possibility, another pulse-chase experiment with cells expressing PLAP alone was carried out in the medium containing GPI-PLD secreted from the stably-expressing cells, demonstrating that no PLAP was released into the medium (Fig. 4C). These results suggest that the newly synthesized GPI-PLD is

740

Vol. 251, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 4. Intracellular release of PLAP by GPI-PLD. pSG5/PLAP was transfected into wild-type CHO cells (A and C) and CHO/GPIPLD cells (B). The cells were pulse-labeled with [35S]methionine for 30 min and chased in a complete fresh medium (A and B), or chased in the medium that had been used for 24-h culture of CHO/GPI-PLD cells (C). At the indicated time of chase, cell lysates and media were prepared and used for immunoprecipitation with anti-PLAP. The immunoprecipitates were analyzed by SDS–PAGE (7.5% gels) and fluorography.

This is the first report for identification of the mRNA in brain, although its presence in liver was also reported by analysis of bovine tissues (18). DISCUSSION FIG. 3. In vitro PLAP-releasing activity of GPI-PLD. Secreted forms of GPI-PLD from CHO cells and H5 cells were partially purified and membranes containing 35S-labeled PLAP, which was used as a substrate, were prepared as described in the Methods. In (A), the 35 S-labeled membranes (substrates) were incubated at 37° C for 2 h with buffer alone (lane 1), 0.5 U/ml PI-PLC (lane 2), 0.5 U/ml recombinant GPI-PLD from CHO cells (lane 3), 0.5 U/ml recombinant GPI-PLD from H5 cells (lane 4), 0.5 U/ml recombinant GPI-PLD from CHO cells 1 1% Triton-X 100 (lane 5), 0.5 U/ml recombinant GPI-PLD from H5 cells 1 1% Triton-X 100 (lane 6), or 1% Triton-X 100 alone (lane 7). After the reaction mixtures were centrifuged at 105,000 x g for 1 h, the supernatants were subjected to immunoprecipitation with anti-PLAP antibodies and analyzed by SDS–PAGE (7.5% gels). sf and mf indicate a soluble form and a membrane form, respectively, of PLAP. The membrane form directly extracted from the membranes (membrane) is also shown at the left. (B) Relative amounts of the soluble form of PLAP were quantified from the data shown in (A). The amount of the soluble form obtained with PI-PLC was taken as 100% and compared with those obtained from the indicated treatments. Values are shown as the mean 6 SD (n53).

In the present study we demonstrated that the antibodies directed against the C-terminal sequence of

able to cleave the GPI anchor of PLAP only within the cells but not after being secreted. Northern blot analysis of GPI-PLD in human tissues. Despite the detection of GPI-PLD at a high level in serum from mammals, little is known about the tissue origin of its synthesis. We examined the GPI-PLD mRNA distribution in human tissues by Northern blot analysis (Fig. 5). In the tissues tested, the GPI-PLD mRNA of 6.7 kbp was detected only in liver and brain.

FIG. 5. Northern blot analysis of GPI-PLD in human tissues. Poly(A)1 RNAs (2 mg/lane) from the indicated human tissues were separated by SDS–PAGE, transferred to a membrane and hybridized with a mixture of 32P-labeled EcoRI-HindIII fragment (831 bp) and HindIII-PstI fragment (766 bp) of GPI-PLD cDNA. The membrane was subjected to autoradiography.

741

Vol. 251, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

GPI-PLD could not recognize the mature 115-kDa form secreted from CHO cells, despite their capability to interact with the 98-kDa form secreted from H5 cells as well as the immature 105-kDa from from the both cells. The 98-kDa from was an incompletely processed molecule that had oligosaccharides lacking sialic acid and possibly galactose residues, as observed for other glycoproteins synthesized in insect cells (29). Thus, it is likely that the C-terminal epitope(s) of GPI-PLD is masked by terminal glycosylation of the oligosaccharides. The C-terminal sequence used as the antigen, however, contained no potential glycosylation site. In addition, the 115-kDa form became immunoreactive when being denatured with the detergent. These observations suggest that the terminal glycosylation may cause a conformational change of GPI-PLD in which the C-terminal epitope(s) cannot be recognized by the antibodies, as observed in other glycoproteins (30). The terminal glycosylation including sialylation occurs in the trans region of the Golgi complex, suggesting that GPI-PLD undergoes the possible conformational change in the Golgi during intracellular transport. GPI-PLD purified from serum effectively cleaves detergent-solubilized GPI-anchored proteins, but is unable to hydrolyze these substrates attached to membranes (14, 16, 17). This is mostly ascribed to a structural feature of intact membranes that would protect the lipid bilayer-anchored substrates from the enzyme attack. In an in vivo system, however, the newly synthesized GPI-anchored proteins are found to be intracellularly cleaved from the membranes by the coexpressed GPI-PLD (18, 19, and this study) and by the endogenous cell-associated enzyme (31). Our preliminary experiment showed that when cells expressing GPI-PLD and PLAP were treated with brefeldin A, which blocks protein transport from the ER to the Golgi (25), PLAP was accumulated as a soluble and Endo H-sensitive form in the cells, indicating that PLAP is released from the membrane by GPI-PLD retained in the ER. Thus, it is likely that the conformational change of GPI-PLD possibly caused by posttranslational modification is closely related with its requirement for detergents in GPI-anchor cleavage activity. The immature form localized in the ER is able to cleave the substrates without detergents, whereas the mature form cleaves the GPI-anchors only after they are solubilized. At present it is not clear whether the detergents are required for the enzyme, the substrates or both. Previous studies demonstrated that although GPI-PLD has an active site in the N-terminal region (32), C-terminal deletions of more than 5 amino acids resulted in a complete loss of GPI-PLD enzymatic activity (33), suggesting that the C-terminal region may play an important role in maintaining the biological activity of GPI-PLD. This may favor our proposal that the C-terminal region of GPI-PLD has a different configuration in the molecule between the mature and

immature forms, so that the mature form is not reactive with anti-C-peptide and requires detergents for enzymic activity. Despite the high amount of GPI-PLD present in mammalian serum, little is known about its tissue origin. Northern blot and in situ hybridization analysis of bovine tissues showed that GPI-PLD mRNA was detected in liver and lung but not in brain (18, 34). Of human tissues, pancreas islets and bone-marrow stromal cells were also demonstrated to contain the GPIPLD mRNA, for which the cDNA amplified by polymerase chain reaction was used (31, 35). In this study the GPI-PLD mRNA was detected not only in liver but also in brain of the human tissues examined, in contrast to the lack of the mRNA in bovine brain. It is not known at present whether such a different expression of GPIPLD is due to the species specificity. The high level of the mRNA detected in the liver of both human and bovine suggest that the liver is the major tissue for production of GPI-PLD that accounts for a large amount of the enzyme in serum. ACKNOWLEDGMENTS This work was supported in part by grants from the Ministry of Education, Science, Culture, and Sports of Japan; from the Japan Science and Technology Corporation (CREST); and from the Central Research Institute of Fukuoka University.

REFERENCES 1. Takami, N., Ogata, S., Oda, K., Misumi, Y., and Ikehara, Y. (1988) J. Biol. Chem. 263, 3016 –3021. 2. Ogata, S., Hayashi, Y., Takami, N., and Ikehara, Y. (1988) J. Biol. Chem. 263, 10489 –10494. 3. Takami, N., Misumi, Y., Kuroki, M., Matsuoka, Y., and Ikehara, Y. (1988) J. Biol. Chem. 263, 12716 –12720. 4. Ikezawa, H., and Taguchi, R. (1981) Methods Enzymol. 71, 731– 741. 5. Low, M. G. (1981) Methods Enzymol. 71, 741–746. 6. Bu¨low, R., and Overath, P. (1986) J. Biol. Chem. 261, 11918 – 11923. 7. Hereld, D., Krakow, J. L., Bangs, J. D., Hart, G. W., and England, P. T. (1986) J. Biol. Chem. 261, 13813–13819. 8. Fox, J. A., Duszenko, M., Ferguson, M. A. J., Low, M. G., and Cross, G. A. M. (1986) J. Biol. Chem. 261, 15767–15771. 9. Morris, J. C., Ping-Sheng, L., Shen, T.-Y., and Mensa-Wilmot, K. (1995) J. Biol. Chem. 270, 2517–2524. 10. Fox, J. A., Soliz, N. M., and Saltiel, A. R. (1987) Proc. Natl. Acad. Sci. USA 84, 2663–2667. 11. Fouchier, F., Baltz, T., and Rougon, G. (1990) Biochem. J. 269, 321–327. 12. Davitz, M. A., Hereld, D., Shak, S., Krakow, J., Englund, P. T., and Nussenzweig, V. (1987) Science 238, 81– 84. 13. Low, M. G., and Prasad, A. R. (1988) Proc. Natl. Acad. Sci. USA 85, 980 –984. 14. Davitz, M. A., Hom, J., and Schenkman, S. (1989) J. Biol. Chem. 264, 13760 –13764. 15. Huang, K. S., Li, S., Fung, W. J., Hulmes, J. D., Reik, L., Pan, Y. C., and Low, M. G. (1990) J. Biol. Chem. 265, 17738 –17745.

742

Vol. 251, No. 3, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

16. Low, M. G., and Huang, K. S. (1991) Biochem. J. 279, 483– 493. 17. Bergman, A. S., and Carlsson, S. R. (1994) Biochem. J. 298, 661– 668. 18. Scallon, B. J., Fung, W. J., Tsang, T. C., Li, S., Kado-Fong, H., Huang, K. S., and Kochan, J. P. (1991) Science 252, 446 – 448. 19. Bernasconi, E., Fasel, N., and Wittek, R. (1996) J. Cell Sci. 109, 1195–1201. 20. Misumi, Y., Sohda, M., Ohkubo, K., Takami, N., Oda, K., and Ikehara, Y. (1990) J. Biochem. 108, 230 –234. 21. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463–5467. 22. Karasuyama, H., and Melchers, F. (1988) Eur. J. Immunol. 18, 97–104. 23. Misumi, Y., Oda, K., Fujiwara, T., Takami, N., Tashiro, K., and Ikehara, Y. (1991) J. Biol. Chem. 266, 16954 –16959. 24. Felgner, P. L., Gadek, T. R., Holm, M., Roman, R., Chan, H. W., Wenz, M., Northrop, J. P., Ringold, G. M., and Danielsen, M. (1987) Proc. Natl. Acad. Sci. USA 84, 7413–7417. 25. Takami, N., Oda, K., Fujiwara, T., and Ikehara, Y. (1990) Eur. J. Biochem. 194, 805– 810.

26. Misumi, Y., Sohda, M., Yano, A., Fujiwara, T., and Ikehara, Y. (1997) J. Biol. Chem. 272, 23851–23858. 27. Laemmli, U. K. (1970) Nature 227, 680 – 685. 28. Haruta, T., Takami, N., Ohmura, M., Misumi, Y., and Ikehara, Y. (1997) Biochem. J. 325, 455– 463. 29. Jarvis, D. L., and Finn, E. E. (1995) Virology 212, 500 –511. 30. Cyster, J. G., Shotton, D. M., and Williams, A. F. (1991) EMBO J. 10, 983–902. 31. Metz, C. N., Brunner, G., Choi-Muira, N. H., Nguyen, H., Gabrilove, J., Caras, I. W., Altszuler, N., Rifkin, D. B., Wilson, E. L., and Davitz, M. A. (1994) EMBO J. 13, 1741–1751. 32. Heller, M., Butikofer, P., and Brodbeck, U. (1994) Eur. J. Biochem. 224, 823– 833. 33. Stadelmann, B., Butikofer, P., Konig, A., and Brodbeck, U. (1997) Biochim. Biophys. Acta. 1355, 107–113. 34. Stadelmann, B., Zurbriggen, A., and Brodbeck, U. (1993) Cell. Tissue Res. 274, 547–552. 35. Brunner, G., Metz, C. N., Nguyen, H., Gabrilove, J., Patel, S. R., Davitz, M. A., Rifkin, D. B., and Wilson, E. L. (1994) Blood 8399, 2115–2125.

743