A Novel Phospholipase A1 with Sequence Homology to a Mammalian Sec23p-interacting Protein, p125*

p125, a mammalian Sec23p-interacting protein, exhibits sequence homology with bovine testis phosphatidic acid-preferring phospholipase A1. In this study, we identified and characterized a new homologue of p125, KIAA0725p. KIAA0725p exhibited remarkable sequence similarity with p125 throughout the entire sequence determined but lacked an N-terminal proline-rich, Sec23p-interacting region. In vitro binding analysis showed that KIAA0725p does not bind to Sec23p. KIAA0725p possessed phospholipase A1 activity preferentially for phosphatidic acid. We examined the effects of overexpression of KIAA0725p on the morphology of organelles. Overexpression of KIAA0725p, like that of p125, caused dispersion of the endoplasmic reticulum-Golgi intermediate compartment and Golgi apparatus. Different from the case of p125, overexpression of KIAA0725p resulted in dispersion of tethering proteins located in the Golgi region and caused aggregation of the endoplasmic reticulum. Our results indicate that KIAA0725p is a new member of the phosphatidic acid-preferring phospholipase A1protein family and suggest that the cellular function of KIAA0725p is different from that of p125.

Phosphatidic acid (PA) 1 plays a key role in the regulation of several cellular processes. In response to hormones, growth factors, and cytokines, the PA level increases rapidly (1). PA activates a variety of proteins involved in signal transduction pathways (2)(3)(4)(5)(6)(7)(8). In addition to the regulatory function, PA is known to cause actin polymerization (9,10).
Several lines of evidence suggest that PA is also involved in vesicle-mediated transport and Golgi organization (11)(12)(13)(14). Recent elegant studies demonstrated that the conversion of lyso-PA, an inverted cone-shaped lipid, to PA, a cone-shaped lipid, induces fission of vesicles from the plasma membrane (15) or tubulation of the Golgi apparatus (16). This change in lipid structure most likely induces the high curvature required for the fission of vesicles (15,16). In addition to this structural contribution, PA may play a regulatory role by interacting with proteins involved in vesicular transport processes. It binds to ADP-ribosylation factor, N-ethylmaleimide-sensitive factor, and kinesin (17), all of which are components necessary for vesicular transport. Thus, the analysis of PA metabolism may provide insight into the mechanisms underlying vesicle-mediated transport and Golgi organization.
We previously identified p125 as a protein that interacts with mammalian Sec23p via its N-terminal proline-rich region (18,19). Sec23p is a component of the COPII coat that functions in the production of vesicles from the ER (20). p125 is localized in the ERGIC and/or cis-Golgi, and its overexpression causes the dispersion of these membrane compartments, suggesting its involvement in the early secretory pathway (18). Interestingly, the central and C-terminal regions of p125 exhibit significant sequence homology with PA-PLA 1 . PA-PLA 1 preferentially cleaves PA and is predominantly expressed in testis (21).
In the present study, we searched data bases using the amino acid sequence of p125 and identified a homologous protein that is encoded by human expressed sequence tag clone KIAA0725. This protein exhibits PLA 1 activity preferentially for PA. Consistent with the fact that KIAA0725p does not possess an N-terminal proline-rich region, it does not bind to Sec23p. Overexpression of KIAA0725p affects the morphology of various organelles including the ER.

EXPERIMENTAL PROCEDURES
cDNA Cloning of KIAA0725p-The KIAA0725 clone was obtained from the Kazusa DNA Research Institute, Japan. To amplify the region upstream of KIAA0725, 5Ј-RACE was carried out using a Marathon TM cDNA amplification kit (CLONTECH) according to the manufacturer's instructions. Marathon-Ready TM cDNA from human placenta and a synthetic oligonucleotide (5Ј-CCTGCTCCATGGGTGTTGAACCCCA) were used as the template and antisense primer for PCR, respectively. The antisense primer was complementary to nucleotides 128 -152 of the reported KIAA0725 sequence.
To obtain the full-length cDNA of KIAA0725p, a DNA fragment comprising from the putative initiation codon to the SalI site of KIAA0725 was amplified by PCR, and the resultant fragment was replaced with the DNA fragment comprising from the 5Ј-end to the SalI site of the KIAA0725 clone.
Northern Blot Analysis-The cDNA fragment (the AflII-FbaI fragment) corresponding to amino acid residues 586 -706 of KIAA0725p was used as a probe. Human multiple tissues blots of poly(A) ϩ RNA (CLONTECH) were incubated with the 32 P-labeled probe in an Ex-pressHyb hybridization solution (CLONTECH) overnight at 68°C. The blots were washed for 40 min in 2ϫ SSC containing 0.05% SDS at room temperature and then for 40 min in 0.1ϫ SSC containing 0.1% SDS at 50°C. Radioactivity was detected with a Fuji Bioimage analyzer BAS2000.
Antibodies-The full-length KIAA0725p with the N-terminal histidine tag was expressed in Escherichia coli cells using expression vector pQE30 (Qiagen). Since the expressed protein was located in insoluble inclusion bodies, it was isolated as follows. E. coli cells expressing the histidine-tagged KIAA0725p were suspended in phosphate-buffered sa-* This work was supported in part by Ministry of Education, Science, Sports and Culture of Japan Grants-in-aid 13680792, 10215205, and 11480183. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom all correspondence should be addressed. line containing protease inhibitors and then passed through a French pressure cell (Aminco, SLM Instruments Inc.) three times at 8,000 p.s.i. The resulting suspension was centrifuged at 27,000 ϫ g for 10 min, and the insoluble fraction obtained was subjected to SDS-PAGE. After staining with 0.05% Coomassie Brilliant Blue R-250, the portion of the gel corresponding to the histidine-tagged KIAA0725p band was excised. The gel fragment was homogenized in phosphate-buffered saline with a Teflon homogenizer and then directly injected into rabbits. The antibody was affinity-purified by using the antigen immobilized on nitrocellulose filters.
The monoclonal anti-ERGIC53 antibody was a generous gift from Dr. H.-P. Hauri of the University of Basel. The polyclonal anti-␤-COP antibody was prepared in this laboratory. The anti-FLAG antibody and anti-GST antibody were obtained from Sigma and Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), respectively.
Plasmid Construction, Transfection, and in Vitro Binding Assay-Mammalian expression plasmid pFLAG-CMV-2 (Eastman Kodak Co.) was used to express proteins fused with the N-terminal FLAG epitope. The cDNA fragment encoding full-length KIAA0725p was inserted into pFLAG-CMV-2. Mammalian expression plasmids for GST-Sec23p and FLAG-tagged p125 were prepared as described previously (18).
For the expression of fusion proteins, 293T cells plated on 35-mm dishes were transfected with 1-2 g of expression plasmids using the LipofectAMINE PLUS reagent (Invitrogen) according to the manufacturer's instructions. At 24 h after transfection, the cells were lysed in lysis buffer (0.35 ml/dish) consisting of 25 mM Hepes-KOH, pH 7.2, 1% Triton X-100, 150 mM KCl, 0.5 g/ml leupeptin, 2 M pepstatin, 2 g/ml aprotinin, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol. The lysate was clarified by centrifugation for 10 min at 17,000 ϫ g. The supernatant was used for the in vitro binding assay as described previously (18).
Site-directed Mutagenesis-The PstI fragment encoding amino acid residues 286 -470 of KIAA0725p was isolated from pFLAG-CMV-2 carrying KIAA0725p cDNA and then inserted into the PstI site of pKF19k DNA (TaKaRa). Oligonucleotide-directed mutagenesis was carried out using Mutan-Super Express K m (TaKaRa) according to the manufacturer's instructions. For the replacement of Ser 351 with alanine, a synthetic oligonucleotide, 5Ј-TGCGGTCATGCTTTAGGTTCG, was used. The replacement was confirmed by DNA sequencing. The PstI fragment carrying the mutation was replaced with the corresponding fragment of pFLAG-CMV-2 carrying KIAA0725p cDNA.
Preparation of 14 C-Labeled Substrates-1-[1-14 C]Palmitoyl-2-oleoyl-PC was prepared from 1-[1-14 C]palmitoyl-lyso-PC as described previously (23) with a slight modification. A solution comprising 0.18 mol of 1-[1-14 C]palmitoyl-lyso-PC (PerkinElmer Life Sciences) and 0.27 mol of oleic acid was dried by evaporation, and then the residual materials were suspended in 100 mM potassium phosphate buffer, pH 7.4. To this suspension were added 5 mol of ATP, 5 mol of MgCl 2 , 0.15 mol of coenzyme A, and 100 g of bovine liver microsomal proteins. The total reaction volume was 480 l. After a 1-h incubation at 37°C with shaking, lipids were extracted from the reaction mixture twice by the method of Bligh and Dyer (24). The extract was spotted in a line onto a preparative thin layer chromatography plate, and then the plate was developed in 50:30:8:4 (v/v) chloroform/methanol/acetic acid/water. The 14 C-labeled PC band was scraped off from the plate, and 14 C-labeled PC was extracted from the silica gels.
1-[1-14 C]Palmitoyl-2-oleoyl-PA, 1-[1-14 C]palmitoyl-2-oleoyl-PS, and 1-[1-14 C]palmitoyl-2-oleoyl-PE were prepared by transphosphatidylation catalyzed by Actinomadura phospholipase D as described previously (25,26). For the preparation of PS or PA, 14 C-labeled PC was suspended in 500 l of diethylether, and then 500 l of a solution comprising 100 mM sodium acetate, pH 5.5, 100 mM CaCl 2 , and 5 M L-Ser was added. The reaction was started by the addition of 2.5 units of Actinomadura phospholipase D (Meito Sangyo, Tokyo, Japan), and 2.5 units of phospholipase D was added every 15 min. After 1 h at 45°C with vigorous stirring, the products were extracted twice by the method of Bligh and Dyer (24). 14 C-Labeled PS and PA were separated by thin layer chromatography in 65:25:8.9:1.1 (v/v/v/v) chloroform, methanol, formic acid, water. For the preparation of PE and PA, 14 C-labeled PC was treated as described above except that the reaction mixture comprised 100 mM sodium acetate, pH 5.5, 100 mM CaCl 2 , and 200 mM ethanolamine. 14 C-Labeled PE and PA were separated by thin layer chromatography in 50:40:3: Preparation of Cell Lysates for the Phospholipase Activity Assay-293T cells plated on 60-mm dishes were transfected with 3 g of expression plasmids using the LipofectAMINE PLUS reagent. At 24 h after transfection, the cells were washed twice with phosphate-buffered saline, and then lysis buffer (200 l) consisting of 25 mM Hepes-KOH, pH 7.2, 250 mM sucrose, 0.5 g/ml leupeptin, 2 M pepstatin, 2 g/ml aprotinin, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol was added to the plate. The cells were scraped off with a cell scraper, and the resulting suspension was collected. This procedure was repeated again, and the two suspensions were pooled. The cells were homogenized with 10 strokes in a stainless steel homogenizer. The homogenate was centrifuged at 1,000 ϫ g for 5 min to obtain a supernatant. The protein concentration was determined using the BCA protein assay reagent (Pierce). The final protein concentration of the samples was adjusted to 3 mg/ml with lysis buffer, and then the samples were used for the phospholipase A assay.
Phospholipase A Assay-Phospholipase A activity was measured as described previously (27,28). 14 C-Labeled phospholipid was dried and suspended in water at a concentration of 200 M. The assay was started by the addition of 50 l of a lipid solution to 200 l of assay buffer. The assay mixture comprised 40 M phospholipid (2,000 -10,000 cpm) and cell lysate (120 g as protein) in 100 mM Tris-HCl, pH 7.5, and 1 mM EDTA (PA as substrate) or in 100 mM Tris-HCl (pH 7.5) and 4 mM CaCl 2 (PS, PC, or PE as substrate). In some experiments, 330 M Triton X-100 was included in the assay mixture. After 1 h at 37°C, the reaction was stopped by the addition of 1.25 ml of Dole's reagent (78:20: . The liberated fatty acid was extracted, and its radioactivity was measured with a scintillation counter. Lipase (Triacylglycerol Lipase) Assay-Lipase activity was measured by the method of Groener and Knauer (29) with a slight modification. Briefly, 250 nmol of 14 C-labeled triolein (trioleoyl glycerol) was dried under a gentle stream of nitrogen and then suspended with a probetype sonicator in 100 l of water containing 0.25 mg of gum arabic. The reaction was started by the addition of this solution to assay buffer containing an appropriate amount of enzyme. The assay mixture (250 l) comprised 100 mM Tris-maleate, pH 8.5, 1% bovine serum albumin, and 1 mM triolein. The reaction was performed at 37°C for 1 h and stopped by the addition of 1 ml of a solution (37:18:44 (v/v/v) benzene, chloroform, methanol) containing 0.1 mM oleic acid, followed by the addition of 125 l of 0.6 N NaOH. The mixture was then vortexed vigorously for 4 -5 min, and briefly centrifuged. The resulting upper phase was taken for the determination of released 14 C-oleic acid with a scintillation counter.
Immunofluorescence Microscopy-Immunofluorescence microscopy was performed as described previously (30). Briefly, cells plated on coverslips were fixed with 4% paraformaldehyde, followed by sequential incubation with primary antibodies and a fluorescein isothiocyanateconjugated or Texas Red-conjugated secondary antibody. Confocal microscopic analysis was performed with an Olympus Fluoview 300 laserscanning microscope.

Identification of KIAA0725p
Homologous to p125-The expressed sequence tag data base was searched with the BLAST program using the amino acid sequence of p125. The search revealed the presence of a human expressed sequence tag clone, KIAA0725 (3,911 base pairs), that codes for a homologous protein (KIAA0725p). Clone KIAA0725 encodes 573 amino acid residues but appears to lack nucleotides corresponding to the N-terminal region. To determine the nucleotide sequence 5Ј-upstream of KIAA0725, the 5Ј-RACE method was used. We found in RACE products that a putative initiation codon resides 413 nucleotides upstream of the 5Ј-end of the reported sequence. Although the sequence surrounding this methionine codon did not match the Kozak consensus sequence, one stop codon was found 30 nucleotides upstream of this methionine codon. Therefore, we assumed that this codon is the initiation codon of KIAA0725p.
By combining the sequence of the KIAA0725 clone with the 5Ј-upstream sequence obtained in the 5Ј-RACE experiments, we identified an open reading frame that encodes a protein of 711 amino acids with a calculated molecular weight of 81,003 (Fig. 1A). As shown in Fig. 1B, KIAA0725p exhibits strong sequence similarity with p125 throughout the entire sequence FIG. 1. KIAA0725p is homologous to p125. A, amino acid sequence of KIAA0725p. The sequence was deduced from the cDNA. The boxed area represents the region whose structure was determined on RACE analysis. The consensus sequence of lipases is indicated in white against a black background with the active site Ser 351 in boldface type. B, sequence comparison between KIAA0725p and p125. The amino acid sequences of KIAA0725p and p125 were analyzed using the FASTA program. The two proteins exhibit 52.6% identity in the 770-amino acid overlap. Two dots and one dot represent identical and similar amino acid residues, respectively. determined (52.6% identity in the 770-amino acid overlap). KIAA0725p is more similar to p125 than PA-PLA 1 . KIAA0725p has a consensus sequence for most lipases, Gly-X-Ser-X-Gly, where X represents any amino acid (31). Other than this consensus sequence, no significant homology was found with other phospholipases or triacylglycerol lipases. Therefore, it seemed that KIAA0725p is a new member of the PA-PLA 1 protein family.
KIAA0725p lacks a signal sequence or transmembrane domain, suggesting that it is a cytosolic or peripheral membrane protein. It is noteworthy that KIAA0725p, unlike p125, does not possess a proline-rich region. In addition, KIAA0725p does not have a coiled-coil region, which also exists in the C-terminal region of p125.
Tissue and Subcellular Distribution of KIAA0725p-As shown in Fig. 2A, Northern blot analysis with the KIAA0725p probe revealed a 5.4-kb transcript in all tissues examined, suggesting the ubiquitous expression of KIAA0725p. A 2.8-kb transcript was also observed. At present, we do not know whether this smaller transcript is a degradation product derived from the larger one or an alternative splicing product.
To examine the expression of KIAA0725p in cells, a rabbit polyclonal antibody against bacterially expressed KIAA0725p was prepared and affinity-purified. On immunoblot analysis of various tissue extracts, the affinity-purified antibody detected a protein of 90 kDa (Fig. 2B). This value is in fair agreement with the calculated mass of KIAA0725p. Indeed, this protein exhibited the same mobility on gels as an overexpressed protein in 293T cells that had been transfected with a plasmid encoding the full-length KIAA0725p (data not shown). In addition to this 90-kDa band, an 85-kDa band was observed on the blot. This might represent a degradation product derived from the 90-kDa protein or an alternative translation product in which Met-31 is used as the initiation codon.
To clarify the subcellular localization of KIAA0725p, homogenates of 293T cells were fractionated by differential centrif-ugation. Centrifugation of homogenates at 1,000 ϫ g for 5 min gave PNS and nuclear fractions. The PNS was then centrifuged at 9,000 ϫ g for 10 min, giving rise to supernatant and mitochondrial fractions. From the supernatant fraction, microsomal and cytosolic fractions were obtained by centrifugation at 105,000 ϫ g for 1 h. The proteins in each fraction were separated by SDS-PAGE and then analyzed by immunoblotting with an anti-KIAA0725p antibody (Fig. 2C). The results showed that KIAA0725p is predominantly localized in the cytosolic fraction in 293T cells, but some was observed in the microsomal fraction. Essentially the same results were obtained for rat liver (data not shown).
KIAA0725p Does Not Bind to Sec23p-p125 interacts with mouse Sec23p via its N-terminal proline-rich region (18). Despite its high sequence similarity with p125, KIAA0725p does not have such a proline-rich region. To determine whether or not KIAA0725p interacts with Sec23p, a pull-down assay was conducted. GST-Sec23p and FLAG-tagged KIAA0725p or FLAG-tagged p125 were transiently expressed in 293T cells. A cell lysate was prepared from each transfectant and then incubated with glutathione beads. The proteins bound to the beads were analyzed by immunoblotting with an anti-FLAG antibody. As shown in Fig. 3A, FLAG-tagged KIAA0725p was not pulled down with GST-Sec23p (lane 2). Under the same conditions, FLAG-tagged p125 was pulled down with GST-Sec23p (lane 1), as observed previously (18). These results suggest that KIAA0725p, unlike p125, does not bind to Sec23p. This observation is consistent with the finding that the proline-rich region, but not the phospholipase-like domain, of p125 interacts with Sec23p. KIAA0725p Possesses PLA 1 Activity-We examined whether or not KIAA0725p possesses PLA 1 activity. FLAG-tagged KIAA0725p was transiently expressed in 293T cells, and then PLA 1 activity in the PNS of the transfected cells was measured essentially as described previously (27). As shown in Fig. 4A, the PNS of KIAA0725p-expressed cells showed much higher PLA 1 activity compared with the PNS of vector-transfected cells. Among the phospholipids examined, PA and PE were effectively hydrolyzed, whereas PS was weakly hydrolyzed and PC was very weakly hydrolyzed.
KIAA0725p has a stretch of sequence, Gly 349 -His-Ser-Leu-Gly 353 , that is common to most lipases including PA-PLA 1 (Ser-His-Ser-Leu-Gly). The seryl residue in this sequence is known to be essential for lipase activity (32). To determine whether or not Ser 351 of KIAA0725p is also required for its PLA 1 activity, we constructed a KIAA0725p mutant, S351A, in which Ser 351 was replaced with alanine to eliminate the hydroxyl group. The plasmid for the FLAG-tagged S351A mutant was transfected into 293T cells, and then PLA 1 activity in the PNS was measured. The PNS of cells expressing the S351A mutant contained very low PLA 1 activity toward PA, PE, PS, and PC, which was indistinguishable from that observed in the vector-transfected cells, although the expression levels of the mutant and wild type KIAA0725p were comparable (Fig. 4A). These results indicate that Ser 351 of KIAA0725p is important for enzyme activity and confirm that the increased PLA 1 activity observed in the PNS of KIAA0725p-transfected cells is due to the expression of KIAA0725p. KIAA0725p appeared to pref-erentially hydrolyze a certain fatty acyl residue of phospholipids. No lipase activity was detected (data not shown).
It was suggested that the enzyme activity of PA-PLA 1 is influenced by the assay conditions. PA-PLA 1 predominantly hydrolyzes PA in the presence of 330 M Triton X-100 and PE in its absence (33). Therefore, we examined the effect of Triton X-100 on the PLA 1 activity of KIAA0725p. In the presence of 330 M Triton X-100, the PNS of KIAA0725p-overexpressing cells hydrolyzed PA, PS, and PC, but not PE (Fig. 4B). It should be noted that in the absence of Triton X-100, KIAA0725p effectively hydrolyzes PE as well. These results suggest that the substrate specificity of KIAA0725p is influenced by the presence of detergents such as Triton X-100, and PA is the most preferred substrate for KIAA0725p irrespective of whether Triton X-100 is present or not.
Because of the sequence similarities among p125, KIAA0725p, and PA-PLA 1 , it was worth examining whether or not p125 has PLA 1 activity. FLAG-tagged p125 was transiently expressed in 293T cells, and the activity was measured as described for KIAA0725p. The PNS of p125-expressed cells did not exhibit significant PLA 1 activity toward PA, PE, PS, and PC compared with that of vector-transfected cells (data not shown). This was surprising for us, because p125 has the consensus sequence of lipases and is quite similar to KIAA0725. Further studies are necessary to explain this unexpected observation.
Overexpression of KIAA0725p Causes Dispersion of the ER-GIC and Golgi Apparatus-To gain more insight into the subcellular localization of KIAA0725p, we carried out immunofluorescence analysis. Since staining of the endogenous protein with our anti-KIAA0725p antibody was weak, FLAG-tagged KIAA0725p was transiently expressed in Vero cells, and then its subcellular distribution was analyzed with an anti-FLAG antibody. FLAG-tagged KIAA0725p, when expressed at low levels, was observed in dotlike structures and/or large aggregates in cells (Fig. 5, A, G, and M). The weak staining throughout the cells might reflect the presence of KIAA0725p in the cytosol. In a previous study (18), we showed that p125, when expressed at low levels, is colocalized with ERGIC53, a marker protein for the ERGIC (34), and/or ␤-COP, a subunit of COP I located in the cis-Golgi (35). Although the perinuclear staining of overexpressed FLAG-KIAA0725p partially overlapped with the ␤-COP or ERGIC53 staining, the dotlike structures positive for FLAG-KIAA0725p did not exactly overlap with the ␤-COPor ERGIC53-positive dots (Fig. 5, C and I). These results may suggest that the location of expressed KIAA0725p is similar but slightly different from that of p125.
Next, we examined the effects of overexpression of KIAA0725p on organelle morphology along the early secretory pathway. We previously demonstrated that overexpression of p125 causes the dispersion of proteins located in the ERGIC and cis-Golgi such as ERGIC53 and ␤-COP (18). On the other hand, p115 and GM130, both of which are also located in the ERGIC and play a role in vesicle tethering (36,37), are not dispersed upon the overexpression of p125 but rather are colocalized with expressed p125 (19). Overexpression of KIAA0725p caused the dispersion of ␤-COP and ERGIC53 (Fig.  5, B, E, and K), as observed in the case of overexpression of p125. Different from the case of p125, however, the overexpression of KIAA0725 also caused dispersion of p115 (Fig. 5, N and Q) and GM130 (data not shown). These results imply that overexpression of KIAA0725p and p125 has the same effect on the localization of ␤-COP and ERGIC53 but different effects on tethering proteins such as p115 and GM130. These different effects may be related to the different locations of p125 and KIAA0725p. PLA 1 Activity Is Required for ER Aggregation-The ER is a central organelle in the synthesis of phospholipids. Next, we analyzed the effect of overexpression of KIAA0725p on the ER. Overexpression of KIAA0725p caused aggregation of ER marker proteins, calnexin (Fig. 6B) and protein-disulfide isomerase (data not shown). It should be noted that KIAA0725p-positive aggregates in the perinuclear region were not colocalized with aggregated ER membranes (Fig. 6C). In contrast, overexpression of p125 did not change the staining pattern of calnexin (Fig. 6H).
To determine the correlation between PLA 1 activity and ER aggregation, we analyzed the morphology of the ER in cells overexpressing a PLA 1 activity-deficient mutant, S351A. As shown in Fig. 6, A and D, no significant differences were ob-served in the localization patterns between the wild-type and mutant proteins. Interestingly, the staining of calnexin (Fig.  6E) and protein-disulfide isomerase (data not shown) did not change as a consequence of overexpression of the PLA 1 activitydeficient mutant. On the other hand, overexpression of the PLA 1 activity-deficient mutant caused the dispersion of ␤-COP, ERGIC53, p115, and GM130, as observed for the wild-type protein (data not shown). These results suggest that the ER aggregation caused by the overexpression of KIAA0725p depends on its PLA 1 activity. DISCUSSION In this study, we identified a novel protein, KIAA0725p, which exhibits a sequence similarity with p125 (18) and bovine PA-PLA 1 (21). Except for the presence of the short consensus sequence of lipases, the overall primary structures of these proteins are remarkably different from those of conventional phospholipases and lipases. Therefore, these proteins appear to constitute a novel family of mammalian phospholipases.
A search of the human genome draft data base revealed that the genes of p125 and KIAA0725p are located on chromosomes 10 and 8, respectively. In addition, it revealed the presence of a putative human orthologue of PA-PLA 1 on chromosome 14. To identify other members of the PA-PLA 1 family, we searched the human genome data base using the BLAST program and the whole sequences of the three members of the PA-PLA 1 family. Two potential homologue genes were found on chromosomes 16 and 17, respectively. A mouse expressed sequence tag clone, AK008229, may correspond to a transcript of the gene on chromosome 16. However, the sequence encoded by the clone did not cover the whole coding region. It is therefore difficult to conclude whether the protein encoded by the AK008229 clone is a member of the PA-PLA 1 family or not.
KIAA0725p preferentially cleaved the ester linkage at the sn-1 position of PA. It also effectively hydrolyzed PE in the absence of Triton X-100. On the other hand, the cleavage efficiencies were low for PS and PC. Although we measured enzyme activity using homogenates of KIAA0725p-expressing cells, the detected activity was obviously derived from expressed KIAA0725p, because mutation of Ser 351 , a putative catalytic residue in the consensus sequence of lipases (31), resulted in elimination of PLA 1 activity in homogenates.
Our previous study demonstrated that overexpression of p125 causes dispersion of the ERGIC containing ERGIC53 and a cis-Golgi compartment containing ␤-COP but not compartments containing tethering proteins such as p115 and GM130 and that overexpressed p125 is colocalized with the tethering protein-containing compartments (18,19). In the present study, we showed that overexpression of KIAA0725p in cultured cells caused the dispersion of all proteins examined in the ER-Golgi region. The fact that the effects of overexpression of KIAA0725p and p125 on intracellular membranes are different may reflect their different physiological functions. The prolinerich region of p125 is critical for the interaction with Sec23p and localization in an ER-Golgi compartment (18). The lack of such a region by KIAA0725p might account for its slightly different localization from that of p125.
Accumulating bodies of evidence suggest that PA, which is a substrate of PA-PLA 1 , is involved in membrane trafficking and Golgi organization (11)(12)(13)(14). However, the morphological disturbance caused by the overexpression of KIAA0725p was not due to its enzymatic activity, because a lipase-inactive mutant also caused the dispersion of membrane compartments. Perhaps the binding of the overexpressed wild-type and mutant proteins to membrane lipids or membrane proteins might cause dispersion of the Golgi and ERGIC.
A KIAA0725p-specific effect on intracellular membranes is that its overexpression causes aggregation of the ER. This effect is dependent on its PLA 1 activity. One plausible model to explain the formation of the ER aggregates is that KIAA0725p promotes fusion of ER membranes by changing the shape of lipids in the cytoplasmic leaflet of lipid bilayer. KIAA0725p can convert cone-shaped PA and/or PE to lysophospholipids by cleaving the ester linkage at their sn-1 position. It has been suggested that cone-shaped lipids favor the formation of a stalk, an intermediate structure formed during membrane fusion, and inverted-cone shaped lipids promote fusion pore formation, the final stage of membrane fusion (38,39). The ER is a highly dynamic organelle; new membrane tubules are formed and fused continuously (40). Therefore, the consumption of cone-shaped lipids such as PA and/or PE may accelerate membrane fusion by facilitating fusion pore formation, leading to ER aggregation.
KIAA0725p is ubiquitously expressed, whereas the expression of PA-PLA 1 is relatively limited to tissues such as testis and brain (21,33). One possible explanation for this difference in expression is that KIAA0725p plays a role in nonspecialized cells, whereas PA-PLA 1 does so in specialized cells. A recent study suggested that PA-PLA 1 undergoes phosphorylation and dephosphorylation and is involved in spermatogenesis and neuronal interactions (41). However, it is possible that PA-PLA 1 is involved in other functions such as exocytosis in testis and brain. Proteins involved in exocytosis are highly expressed in these tissues and undergo regulation through phosphorylation and dephosphorylation.
One enigma, which was not solved in this study, is that p125 does not exhibit PLA 1 activity. This finding does not necessarily imply that p125 has no PLA 1 activity. p125 might possess very low activity that is under the detection limit of our assay system. Another possibility is that the activity of p125 is regulated by protein factors such as Sec23p. However, our preliminary results showed that deletion of the N-terminal, Sec23pinteracting region of p125 did not result in the emergence of PLA 1 activity (data not shown). Further studies are necessary to solve this problem.
In summary, we identified a new member of the PA-PLA 1 family, KIAA0725p. Our results indicate that the enzyme activity and/or cellular localization of KIAA0725p are somewhat different from those of PA-PLA 1 and/or p125. These results suggest that the three proteins have different physiological functions in cells.