Interaction of Phospholipase A/Acyltransferase-3 with Pex19p

Background: The overexpression of phospholipase A/acyltransferase-3 (PLA/AT-3) in mammalian cells causes specific disappearance of peroxisomes. Results: PLA/AT-3 bound to Pex19p, one of peroxins, and inhibited the binding of Pex19p to peroxisomal membrane proteins. Conclusion: The interaction between Pex19p and PLA/AT-3 may be related to the down-regulation of peroxisomes by PLA/AT-3. Significance: PLA/AT-3 may be involved in a novel regulatory mechanism for peroxisome biogenesis. Phospholipase A/acyltransferase (PLA/AT)-3 (also known as H-rev107 or AdPLA) was originally isolated as a tumor suppressor and was later shown to have phospholipase A1/A2 activity. We have also found that the overexpression of PLA/AT-3 in mammalian cells results in specific disappearance of peroxisomes. However, its molecular mechanism remained unclear. In the present study, we first established a HEK293 cell line, which stably expresses a fluorescent peroxisome marker protein (DsRed2-Peroxi) and expresses PLA/AT-3 in a tetracycline-dependent manner. The treatment with tetracycline, as expected, caused disappearance of peroxisomes within 24 h, as revealed by diffuse signals of DsRed2-Peroxi and a remarkable decrease in a peroxisomal membrane protein, PMP70. A time-dependent decrease in ether-type lipid levels was also seen. Because the activation of LC3, a marker of autophagy, was not observed, the involvement of autophagy was unlikely. Among various peroxins responsible for peroxisome biogenesis, Pex19p functions as a chaperone protein for the transportation of peroxisomal membrane proteins. Immunoprecipitation analysis showed that PLA/AT-3 binds to Pex19p through its N-terminal proline-rich and C-terminal hydrophobic domains. The protein level and enzyme activity of PLA/AT-3 were increased by its coexpression with Pex19p. Moreover, PLA/AT-3 inhibited the binding of Pex19 to peroxisomal membrane proteins, such as Pex3p and Pex11βp. A catalytically inactive point mutant of PLA/AT-3 could bind to Pex19p but did not inhibit the chaperone activity of Pex19p. Altogether, these results suggest a novel regulatory mechanism for peroxisome biogenesis through the interaction between Pex19p and PLA/AT-3.


Phospholipase A/acyltransferase (PLA/AT)-3 (also known as H-rev107 or AdPLA) was originally isolated as a tumor suppressor and was later shown to have phospholipase A 1 /A 2 activity.
We have also found that the overexpression of PLA/AT-3 in mammalian cells results in specific disappearance of peroxisomes. However, its molecular mechanism remained unclear. In the present study, we first established a HEK293 cell line, which stably expresses a fluorescent peroxisome marker protein (DsRed2-Peroxi) and expresses PLA/AT-3 in a tetracycline-dependent manner. The treatment with tetracycline, as expected, caused disappearance of peroxisomes within 24 h, as revealed by diffuse signals of DsRed2-Peroxi and a remarkable decrease in a peroxisomal membrane protein, PMP70. A time-dependent decrease in ether-type lipid levels was also seen. Because the activation of LC3, a marker of autophagy, was not observed, the involvement of autophagy was unlikely. Among various peroxins responsible for peroxisome biogenesis, Pex19p functions as a chaperone protein for the transportation of peroxisomal membrane proteins. Immunoprecipitation analysis showed that PLA/AT-3 binds to Pex19p through its N-terminal proline-rich and C-terminal hydrophobic domains. The protein level and enzyme activity of PLA/AT-3 were increased by its coexpression with Pex19p. Moreover, PLA/AT-3 inhibited the binding of Pex19 to peroxisomal membrane proteins, such as Pex3p and Pex11␤p. A catalytically inactive point mutant of PLA/AT-3 could bind to Pex19p but did not inhibit the chaperone activity of Pex19p. Altogether, these results suggest a novel regulatory mechanism for peroxisome biogenesis through the interaction between Pex19p and PLA/AT-3.
Peroxisomes are single membrane-surrounded organelles present in virtually all eukaryotic cells and are involved in a wide variety of metabolic events, including scavenging of peroxides and reactive oxygen species, the catabolism of very long chain fatty acids and branched chain fatty acids, and the biosynthesis of all ether-type lipid precursors (1)(2)(3)(4)(5). The physiological importance of peroxisomes has been confirmed by the analysis of a number of inherited diseases exhibiting peroxisome dysfunctions (6,7). Peroxins, encoded by the PEX genes, are a series of proteins responsible for the biogenesis of peroxisomes, and their defects lead to the dysfunction of peroxisomes. So far, 31 peroxins have been reported, and they are involved in the generation and division of peroxisomes as well as the import of peroxisomal proteins (4,8). In humans, 14 peroxins have been identified and shown to link to peroxisome biogenesis disorders.
As a machinery of peroxisome biogenesis, it is well known that nascent peroxisomal matrix proteins are transported into peroxisomes with the aid of several peroxins, which recognize peroxisomal targeting signals (PTSs), 2 PTS1 and PTS2, of peroxisomal matrix proteins (2,9). However, it is poorly understood how peroxisome membrane structure is formed and how peroxisomal membrane proteins (PMPs) are transported into peroxisomes. In addition, because peroxisomes lack the phospholipid-synthesizing enzymes necessary for the formation of * This work was supported by Grants-in-Aid for Scientific Research (C) peroxisome membrane structure, phospholipids must be trafficked and supplied to peroxisomes from other organelles, such as the ER. Pex3p, Pex16p, and Pex19p have been identified as peroxins indispensable for peroxisome membrane assembly and PMP transport, and the cells deficient in these proteins are devoid of peroxisome structure itself (8,9). Pex19p is predominantly localized to cytoplasm and binds to various PMPs, whereas Pex3p and Pex16p are associated with peroxisomal membrane and function as the membrane-anchoring site for Pex19p⅐PMP complexes and as the receptor for Pex3p⅐Pex19p complex, respectively (10).
The abundance of peroxisomes is remarkably affected by the nutritional environment and specific conditions (11). For example, peroxisomes can be induced and proliferated in yeast cultured in methanol as a sole carbon source and in rodents treated with peroxisome proliferators. Increased peroxisomes are rapidly decreased and degraded by altering the environment or by withdrawing peroxisome proliferators (12). These results indicate that peroxisome levels are reciprocally regulated by the balance between their biogenesis and degradation. Recently, it has been reported that the selective autophagy of peroxisomes, pexophagy, contributes to the maintenance of quality and quantity of peroxisomes (11,13).
The HRAS-like suppressor (HRASLS) family, consisting of five members (HRASLS1 to 5), was originally isolated as tumor suppressors negatively regulating the oncogene H-Ras (14,15). It has been reported that these proteins are related to various diseases, such as cancers (16,17), obesity (18,19), and Poland syndrome, a rare disorder characterized by hypoplasia/aplasia of the pectoralis major muscle (20). We and others have demonstrated that all of these members function as enzymes with phospholipase A 1/2 (PLA 1/2 ) and phospholipid acyltransferase activities (21)(22)(23)(24)(25)(26). Thus, we proposed to rename HRASLS1 to 5 as phospholipase/acyltransferase-1 to -5 (PLA/AT-1 to -5), respectively (26). Unexpectedly, we found that the overexpression of PLA/AT-3 (HRASLS3, H-rev107, or AdPLA) or PLA/ AT-2 (HRASLS2) in mammalian cells results in the disappearance of peroxisome membrane structure and the dysfunction of peroxisomes, as revealed by a remarkable decrease in the intracellular levels of ether-type lipids (27,28). The disappearance of peroxisomes was dependent on the enzyme activity of PLA/ AT-3 because the mutant devoid of enzyme activity failed to affect peroxisome functions (27). However, the mechanism by which PLA/AT-3 and -2 regulate the abundance of peroxisomes remained unclear.
Because the phenotypes appearing in cells by the overexpression of PLA/AT-3 or -2 were similar to those in Pex3p-, Pex16p-, or Pex19p-deficient cells, we presumed that the negative regulation of peroxisomes by these PLA/AT proteins is mediated by these peroxins. In this study, we show that PLA/ AT-3 is a novel Pex19p-binding protein and suppresses the binding of Pex19p to PMPs. The enzyme activity of PLA/AT-3 was unrelated to its binding to Pex19p but was required to suppress the interaction of Pex19p with PMPs. These results suggest the presence of a novel regulatory mechanism for peroxisome biogenesis through the interaction between Pex19p and PLA/AT-3.
Cell Culture, Transfection, and RNA Interference-HEK293, HEK293T, and COS-7 cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal calf serum at 37°C in humidified air containing 5% CO 2 . Expression vectors or siRNAs were introduced into cells with Lipofectamine 2000 or Lipofectamine RNAiMAX according to the manufacturer's instructions, respectively. The final concentration of siRNA was 20 nM. Forty-eight hours after transfection, the cells were used for RT-PCR, PLA 1/2 assay, and metabolic labeling with Establishment of FL-PLA/AT-3-Tet-on Cells Expressing DsRed2-Peroxi-HEK293 cells were transfected with pcDNA6/TR vector, expressing tetracycline repressor, by the use of Lipofectamine 2000. Cells were selected in the medium containing 10 g/ml blasticidin. Clonal cell lines were isolated by colony lifting and maintained in the blasticidin-containing medium. The cell line expressing the highest level of tetracycline repressor was transfected with DsRed2-Peroxi vector by the use of Lipofectamine 2000. Cells were selected in the medium containing 1 mg/ml Geneticin and 6 g/ml blasticidin. Clonal cell lines were isolated by colony lifting and maintained in the medium containing Geneticin and blasticidin. One of the obtained clones, named DsRed2-PTS1 cell, was transfected with pcDNA5/TO vector, expressing FL-PLA/AT-3, by the use of Lipofectamine 2000. Cells were selected in the medium containing 200 g/ml hygromycin, 600 g/ml Geneticin, and 6 g/ml blasticidin. Clonal cell lines were isolated by colony lifting and maintained in the medium containing hygromycin, Geneticin, and blasticidin. One of the obtained cell lines was named FL-PLA/AT-3-Tet-on and used throughout this study.
To measure PLA 1/2 activity, the harvested cells were suspended in homogenization buffer (0.25 M sucrose, 1 mM EDTA, and 20 mM Tris-HCl, pH 7.4) and sonicated three times each for 3 s. The cell homogenates (30 g of protein) were incubated with 200 M 1,2-[1-14 C]dipalmitoyl-PC (45,000 cpm) in 100 l of 50 mM Tris-HCl (pH 8.0), 2 mM DTT, and 0.1% Nonidet P-40 at 37°C for 30 min. The reaction was terminated with the addition of 320 l of a mixture of chloroform/methanol (2:1, v/v) containing 5 mM 3(2)-t-butyl-4-hydroxyanisole. After centrifugation, 100 l of the lower fraction was spotted on a silica gel thin layer plate (10-cm height) and developed at 4°C for 25 min in chloroform/methanol/H 2 O (65:25:4, v/v/v). The distribution of radioactivity on the plate was quantified using a BAS1500 bioimaging analyzer (FUJIX Ltd., Tokyo, Japan). The protein concentration was determined by the method of Bradford with bovine serum albumin as a standard.
Western Blotting and Immunoprecipitation-For Western blotting unrelated to the immunoprecipitation assay, cells were homogenized in homogenization buffer by being passed through a 25-gauge syringe (for COS-7 cells) or a 27-gauge syringe (for HEK293 cells) (27), and nuclear fractions were prepared from the homogenates by centrifugation at 800 ϫ g for 10 min at 4°C. The resultant postnuclear supernatant (PNS) fractions were then centrifuged at 105,000 ϫ g for 30 min at 4°C to separate the cytosol (supernatant fractions) from cellular organelles (particulate fractions). For the immunoprecipitation assay, cells were lysed in ice-cold lysis buffer (0.25 M sucrose, 1 mM EDTA, 0.5% Triton X-100, and 20 mM Tris-HCl, pH 7.4) containing protease inhibitors for 1 h at 4°C. After centrifugation at 18,000 ϫ g for 10 min at 4°C, the supernatant was used as cell lysate. The cell lysates were immunoprecipitated with the aid of anti-FLAG M2 antibody-conjugated agarose or antibody-conjugated Protein A-Sepharose CL-4B. Samples were separated by SDS-PAGE and electrotransferred to a hydrophobic polyvinylidene difluoride membrane (Hybond P). The membrane was blocked with PBS containing 5% dried milk and 0.1% Tween 20 (buffer A) and then incubated with primary antibodies in buffer A at room temperature for 1 h, followed by incubation with horseradish peroxidase-labeled secondary antibodies in buffer A at room temperature for 1 h. Proteins were finally treated with an ECL Plus kit and visualized with the aid of a LAS1000plus lumino-imaging analyzer (FUJIX Ltd.).
RT-PCR-Total RNAs were isolated from cells using an RNeasy Mini Kit. cDNAs were prepared from total RNAs using a PrimeScript RT reagent kit and subjected to PCR amplification by Ex Taq DNA polymerase. The primers used are shown in supplemental Table 2. The PCR conditions used were as follows: denaturation at 96°C for 20 s, annealing at 60°C for 20 s, and extension at 72°C for 20 s (24 cycles for GAPDH, 28 cycles for others).
Metabolic Labeling-Cells were grown at 37°C to 80% confluence in a 6-well plate containing Dulbecco's modified Eagle's medium with 10% fetal calf serum and were labeled with [1-14 C]hexadecanol (1.6 Ci) for 18 h. Cells were then harvested and washed twice with PBS. The total lipids were extracted by the method of Bligh and Dyer (29), spotted on a silica gel thin layer plate (20-cm height), and developed at 4°C for 1 h in hexane/diethyl ether/acetic acid (75:25:1, v/v/v). The distribution of radioactivity on the plate was quantified using a BAS1500 bioimaging analyzer.
LC-ESI/MS Analysis-The lipids of cultured cells were extracted by the method of Bligh and Dyer. Mass spectrometry (MS) analysis was performed as described previously by using a Quattro Micro tandem quadrupole mass spectrometer (Waters, Milford, MA) equipped with an electrospray ionization (ESI) source (27). Phospholipid molecular species were separated and analyzed using normal phase capillary-liquid chromatography (LC)-ESI/MS with a Deverosil Si60 silica column (150 ϫ 0.3-mm inner diameter, 5-m particle size; Nomura Chemicals, Nagoya, Japan). The samples were provided by an UltiMate high performance liquid chromatography system (Dionex Corp.) into the electrospray interface at a flow rate of 4 l/min. For both positive and negative ionization, the mobile phase consisted of acetonitrile/methanol (2:1) containing 0.1% ammonium formate (pH 6.4) (solvent A) and methanol/water (2:1) containing 0.1% ammonium formate (pH 6.4) (solvent B). The mobile phase was started with 100% acetonitrile. After 5 min, the mobile phase was changed from 100% acetonitrile to a mixture of solvent A and B (90:10). The linear gradient reached 40% solvent A at 20 min. The mobile phase was then held constant for 15 min. Nitrogen was used as drying gas. Desolvation gas flow was 523 liters/h, and cone gas flow was maintained at 42 liters/h. Desolvation temperature was 150°C, and source temperature was 90°C. The capillary and cone voltages were set at 3.4 -3.7 kV and 36 -40 V, respectively. Argon was used as collision gas for collision-associated dissociation experiments at a pressure of 5 ϫ 10 Ϫ4 millibars. The collision energy was set to 40 eV.
For immunocytochemistry, cells were cultured on 18-mm glass coverslips containing Dulbecco's modified Eagle's medium with 10% fetal calf serum. The cells were then fixed with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 15 min. The fixed cells were rinsed in PBS and treated with 0.25% NH 4 Cl in PBS for 10 min to quench free aldehyde groups. The cells were then washed with PBS and permeabilized with 0.15% Triton X-100 in PBS for 10 min. After treatment with a blocking buffer (5% normal goat serum in PBS) for 1 h, the cells were incubated with anti-FLAG antibody (1:1,000 dilution) or anti-Myc antibody (1:500 dilution) in 1% normal goat serum plus PBS for 1 h at room temperature. The cells were washed with PBS and labeled with Alexa 488-conjugated anti-mouse IgG (1:1,000 dilution) or Alexa 594-conjugated anti-rabbit IgG (1:1,000 dilution) in 1% normal goat serum plus PBS for 1 h. The specimen coverslips were mounted on glass slides using a mounting medium Permafluor and were observed with the LSM 700 confocal laser microscope.

Results
Peroxisomal Proteins Are Abrogated in PLA/AT-3-expressing Cells-We first examined the effect of PLA/AT-3 expression on the expression levels and intracellular localization of a peroxisomal membrane protein (PMP70) and three peroxisomal matrix proteins (catalase, ADHAPS), and peroxisomal 3-ketoacyl-CoA thiolase (thiolase)). As for PTS of these proteins, PMP70 possesses mPTS, catalase possesses PTS1, and ADHAPS and thiolase possess PTS2 (2,9). In the assays shown in Fig. 1, control HEK293 cells and HEK293 cells stably expressing N-terminally FLAG-tagged PLA/AT-3 (FL-PLA/AT-3 cells) were homogenized and centrifuged at 800 ϫ g. The resultant PNS was then fractionated into particulate and soluble fractions by ultracentrifugation at 105,000 ϫ g. As analyzed by Western blotting, endogenous PMP70 and catalase in the control cells were mostly detected in the particulate fractions, where peroxisomes were expected to exist. By contrast, in FL-PLA/AT-3 cells, faint signals of PMP70 were seen in both particulate and soluble fractions, and catalase was mostly detected in soluble rather than particulate fractions. These effects of PLA/AT-3 expression were in agreement with our previous results (27). Moreover, when Myctagged ADHAPS or Myc-tagged thiolase was transiently expressed, the Myc signals of both proteins were detected primarily in the particulate fractions of the control cells and were much fainter in the FL-PLA/AT-3 cells. These results suggested that the overexpression of PLA/AT-3 affects the abundance or localization of the peroxisomal proteins with PTS2 as well as those with mPTS or PTS1.
Expression of PLA/AT-3 Rapidly Causes Disappearance of Peroxisomes-To cytochemically monitor peroxisomes, we established a HEK293 cell line named DsRed2-PTS1. In these cells, peroxisomes were fluorescently labeled with DsRed2-Peroxi (a fluorescent peroxisome marker protein possessing a PTS1 sequence at the C terminus) (30). Using DsRed2-PTS1 cells, we further established a Tet-on HEK293 cell line (FL-PLA/AT-3-Tet-on cells), which can express FL-PLA/AT-3 in a tetracycline-dependent manner (31). When FL-PLA/AT-3-Tet-on cells were observed by fluorescent confocal microscopy, peroxisomes, visualized by DsRed2-Peroxi, were discernible as punctate signals ( Fig. 2A). Before the addition of the tetracycline analog doxycycline (DOX) to the cells, FL-PLA/AT-3 signals were not detected. At 12 h after the addition of DOX,   FL-PLA/AT-3 signals appeared in cytoplasm, and at 24 h, the signals were further enhanced. This localization was consistent with the previous observation that PLA/AT-3 has a hydrophobic domain at the C terminus, yet the overexpressed PLA/AT-3 is distributed throughout cells, including cytoplasm (22). Consistently, as analyzed by RT-PCR and Western blotting, DOX induced mRNA and protein of FL-PLA/AT-3 (Fig. 2, B and C). The PLA 1/2 activity of PLA/AT-3 in the cell homogenates was also enhanced from 0.07 to 1.4 nmol/min/mg protein at 24 h, and this level was maintained at least until 96 h (Fig. 2D).
Regarding time-dependent changes of peroxisomal proteins, DsRed2-Peroxi signals became diffuse at 24 h ( Fig. 2A), and the level of endogenous PMP70 protein began to decrease at 12 h and became hardly detectable at 48 h (Fig. 2C). To determine whether the peroxisomal loss effect of PLA/AT-3 is reversible, the medium was returned to DOX-free medium after FL-PLA/ AT-3-Tet-on cells were cultured in the presence of DOX for 48 h. As analyzed by Western blotting, endogenous PMP70 protein became detectable 7 days after the removal of DOX. This result suggests that the effect of PLA/AT-3 on peroxisomes is reversible (data not shown).
When FL-PLA/AT-3-Tet-on cells were treated with DOX for different time periods, followed by metabolic labeling with [ 14 C]hexadecanol as a precursor of MADAG, the intracellular levels of [ 14 C]MADAG were time-dependently decreased concomitant with the accumulation of [ 14 C]hexadecanol in the cells (Fig. 2E). Similarly, LC-ESI/MS analyses showed that the levels of major plasmenylethanolamine species (e32:2, e34:2, e36:5, e38:5, and e40:7) were decreased in a DOX-dependent manner without obvious changes in the levels of diacyl-type phosphatidylethanolamine species (Fig. 3). To establish the structural assignments of plasmenylethanolamine species, we also performed LC-ESI-MS/MS for [M Ϫ H] Ϫ ions (supplemental Fig. 1). Plasmanylcholine levels were also reduced (data not shown). These results indicated that the dysfunction of peroxisomes rapidly occurs, according to the increasing expression level of PLA/AT-3.
Because peroxins, such as Pex3p, Pex11␤p, Pex16p, and Pex19p, are indispensable for the biogenesis of peroxisome, we analyzed a possible DOX-dependent translocation of these proteins in FL-PLA/AT-3-Tet-on cells. In the absence of DOX, Pex3p-EGFP and Pex16p-EGFP were clearly detected as punctate signals and mainly colocalized with DsRed2-Peroxi signals (Fig. 4, A and B). However, after the treatment with DOX for 24 h, the EGFP signals of both proteins were detected in nonperoxisomal structures. These results were similar to previous reports that the deficiency of peroxisomes resulted in the mislocalization of Pex3p and Pex16p to non-peroxisomal structures (32)(33)(34). Another peroxisomal membrane protein, Pex11␤p, was similarly translocated (Fig. 4C). By contrast, Pex19p signals were detected throughout the cytoplasm irrespective of the DOX treatment (Fig. 4D). Importantly, the diffuse localization of DsRed2-Peroxi signals induced by DOX was not restored to their original localization by the overexpression of these peroxins. However, we could not rule out the possibility that the expression levels of these proteins were not high enough to overcome the effect of PLA/AT-3 on peroxisomal structure.
Our finding might result from the enhanced clearance of peroxisomes by the selective autophagy "pexophagy" (11,13,35). We therefore examined whether PLA/AT-3 stimulates the conversion of LC3-I protein to its phosphatidylethanolamineconjugated form, LC3-II, which is known to occur during autophagosome formation (36). However, the DOX-dependent expression of PLA/AT-3 did not stimulate LC3-II formation (Fig. 2C). Furthermore, we analyzed a possible translocation of EGFP-LC3 in FL-PLA/AT-3-Tet-on cells in a DOX-dependent manner. As reported previously (37), the diffuse distribution of EGFP-LC3 was changed to punctate autophagosome signals by culturing the cells under the starved conditions (Fig. 2F). However, such translocation of EGFP-LC3 did not occur by the treatment with DOX, suggesting that autophagy is not involved in the peroxisome disappearance caused by PLA/AT-3.
Interaction of PLA/AT-3 with Pex19p-Because the functional abnormality of Pex3p, Pex16p, or Pex19p results in the deficiency of peroxisomes due to the defects in peroxisome membrane biogenesis (1, 6, 38), we examined whether PLA/ AT-3 interacts with these peroxins and regulates their functions. To this end, we used COS-7 cells, which were more suitable for transient expression, instead of HEK293 cells, which were suitable for stable expression and were used in the experiments of Figs. 1-4, and carried out the immunoprecipitation assay between PLA/AT-3 and these peroxins. The C-terminally FLAG-tagged PLA/AT-3 (PLA/AT-3-FL) was transiently coexpressed with either HA-tagged Pex3p (Pex3p-HA), V5-tagged Pex16p (Pex16p-V5), or Myc-tagged Pex19p (Myc-Pex19p) in COS-7 cells, and cell lysates were subjected to immunoprecipitation using anti-FLAG antibody-conjugated agarose. The immunoprecipitate, which we used for Western blotting, corresponded to 13 times the volume of the lysate. Because the remaining FLAG-tagged PLA/AT-3 was not detected in the lysate after the immunoprecipitation with anti-FLAG antibody, the efficiency of the immunoprecipitation was thought to be high enough.
When analyzed by Western blotting, Pex3p-HA and Myc-Pex19p, but not Pex16p-V5, were coimmunoprecipitated with PLA/AT-3-FL (Fig. 5A). The doublet bands of Myc-Pex19p were presumably attributed to its partial farnesylation (39). We also examined whether the immunoprecipitation occurs in the other directions. The cell lysates were subjected to immunoprecipitation using anti-HA antibody for Pex3p-HA, anti-V5 antibody for Pex16p-V5, and anti-Myc antibody for Myc-Pex19p. Consistent with the aforementioned results, Western blot analysis with anti-FLAG antibody showed that PLA/AT-3-FL was coimmunoprecipitated with Pex3p-HA and Myc-Pex19p but not with Pex16p-V5 (data not shown). The amount of PLA/ AT-3 immunoprecipitated decreased when Pex3p-HA or Pex16-V5 was coexpressed (Fig. 5A). Because the expression Because Pex19p is a chaperone protein binding to and regulating multiple PMPs (33,40) and because both Pex19p (Fig.  4D) (33) and PLA/AT-3 ( Fig. 2A) are mainly present in the cytoplasm, it was likely that the interaction between PLA/AT-3 and Pex19p is related to the effect of PLA/AT-3 on peroxisomes. We therefore examined which domain of PLA/AT-3 is required for this interaction. Because the primary structure of  (Fig. 5B). These mutants were previously reported to be catalytically inactive (24). When these PLA/ AT-3 mutants were co-expressed with Myc-Pex19p and treated with anti-FLAG antibody-conjugated agarose gel, Myc-Pex19p was coimmunoprecipitated with PLA/AT-3-C113S-FL as well as the wild-type PLA/AT-3, whereas the intensity of the Myc-Pex19p signal was weak or hardly detectable in the presence of PLA/AT-3-⌬N10-FL or PLA/AT-3-⌬C27-FL (Fig. 5C). These results suggested that the proline-rich domain and hydrophobic domain of PLA/AT-3 play important roles in its binding to Pex19p. Judging from relative intensities of the doublet Myc-Pex19p bands in the cell lysate, the expression of PLA/AT-3 did not appear to affect the farnesylation of Pex19p (Fig. 5C). It was previously reported that Pex19p increases the level of Pex3p by stabilizing this protein (34). We could reproduce this result by co-expressing Pex3p and Pex19p (Fig. 5D). Similarly, the expression of Pex19p increased the amount of PLA/AT-3 protein in the cell lysate, whereas the PLA/AT-3 mRNA level was unaltered (Fig. 5E). Consistent with this observation, the PLA 1/2 activity derived from PLA/AT-3 in the cell lysate was 1.5-fold increased by co-expression of Pex19p (Fig. 5E). Pex19p itself did not have PLA 1/2 activity. Furthermore, Pex19p did not stimulate the PLA 1/2 activity of PLA/AT-3 when both proteins were separately prepared and mixed upon enzyme assay (Fig.   5E). These results suggest that Pex19p binds to PLA/AT-3 and enhances the intracellular level of the catalytically active PLA/AT-3. We next examined whether the interaction between PLA/ AT-3 and Pex19p occurs in a cell-free system. The lysate expressing PLA/AT-3-FL and that expressing Myc-Pex19p were prepared separately and mixed together. When the mixture was subjected to immunoprecipitation using anti-FLAG antibody, Myc-Pex19p was coimmunoprecipitated with PLA/ AT-3-FL (Fig. 5F). Thus, the interaction between PLA/AT-3 and Pex19p was reproduced in a cell-free system.
To confirm the results so far obtained with COS-7 cells, we next used HEK293T as host cells. When co-expressed in HEK293T cells, Myc-Pex19p was coimmunoprecipitated with wild-type PLA/AT-3-FL and PLA/AT-3-C113S-FL but not with PLA/AT-3-⌬C27-FL (Fig. 6A). In this assay, the expression of Myc-Pex19p increased the amounts of wild-type PLA/ AT-3 and C113S mutant, but not ⌬C27 mutant, in the cell lysates. In contrast, mRNA levels of these PLA/AT-3s were hardly affected by the co-expression of Pex19p (Fig. 6B). The PLA 1/2 activity in the homogenate of wild-type PLA/AT-3-expressing cells was significantly increased by co-expression of Pex19p (Fig. 6C). The C113S and ⌬C27 mutants were catalytically inactive as expected.
In the preceding and following experiments concerning Pex19p, we detected recombinant Pex19p with a tag. Here, we also examined the effect of endogenous Pex19p on recombinant FL-PLA/AT-3 in HEK293 cells by using FL-PLA/AT-3-Tet-on cells. The expression of endogenous Pex19p was successfully suppressed by two different siRNAs against Pex19p mRNA, as examined by RT-PCR (Fig. 7A). This suppression significantly reduced the PLA 1/2 activity in the cell homogenates (Fig. 7B). In another assay, after FL-PLA/AT-3 was DOX-dependently expressed in FL-PLA/AT-3-Tet-on cells, FL-PLA/AT-3 was immunoprecipitated with anti-FLAG antibody from the cell lysates and visualized as a ϳ18.9 kDa band by silver staining (Fig. 7C). When these immunoprecipitates were immunostained with anti-Pex19p antibody, endogenous Pex19p was detected only in the DOX-treated cells (Fig. 7D), indicating that PLA/AT-3 interacts with endogenous Pex19p as well as recombinant Pex19p. PLA/AT-3 Interacts with Pex19p in Vivo-To further confirm that PLA/AT-3 binds to Pex19p in living cells, we employed the nuclear mislocalization assay (32). For this purpose, we constructed Pex19p-NLS-Myc, which has two tandem copies of NLS and a Myc tag. PLA/AT-3-FL was co-expressed with Myc-Pex19p or Pex19p-NLS-Myc in COS-7 cells, and the cell homogenates were fractionated into nuclear and PNS fractions. When analyzed by Western blotting, Myc-Pex19p was mostly localized in the PNS fraction, whereas Pex19p-NLS-Myc was seen in both nuclear and PNS fractions (Fig. 8A). As for the localization of PLA/AT-3-FL, its signal in the nuclear fraction of Pex19p-NLS-Myc-expressing cells was more intense than that in the nuclear fraction of Myc-Pex19pexpressing cells (Fig. 8B). Immunofluorescence microscopy also revealed that PLA/AT-3-FL accumulates in the nucleoplasm when Pex19p-NLS-Myc, but not Myc-Pex19p, was co-expressed (Fig. 8C). These results showed that PLA/AT-3 at least partly binds to Pex19p in living cells.
PLA/AT-3 Inhibits the Interaction of Pex19p with PMPs in a Catalytic Activity-dependent Manner-We examined whether the coexistence of PLA/AT-3 affects the ability of Pex19p to bind to PMPs. When Myc-Pex19p and Pex3p-HA were co-expressed in COS-7 cells, both proteins in the cell lysate were precipitated by anti-HA antibody (Fig. 9A), as reported previously (34,41). The additional expression of wild-type PLA/AT-3-FL abolished the binding of Pex19p to Pex3p despite the unaltered protein levels of Pex19p and Pex3p in the cell lysate. Interestingly, the C113S mutant of PLA/AT-3 did not interfere with the binding of Pex19p to Pex3p. Similarly, when HA-Pex19p and Myc-Pex11␤p were co-expressed, the additional expression of PLA/AT-3-FL, but not its C113S mutant, potently inhibited the co-precipitation of Pex19p and Pex11␤p (Fig. 9B). Notably, the protein level of Pex11␤p in the lysate, but not its mRNA level, was remarkably increased by the simultaneous expression of Pex19p (Fig. 9B). This result suggested that Pex19p stabilized Pex11␤p. Moreover, the overexpression of PLA/AT-3 reduced the levels of recombinant PMP70 in an enzyme activity-dependent manner (Fig. 9C). Similar results were obtained with the cells in which Pex19p was additionally expressed (Fig. 9C). These results suggested that PLA/AT-3 interferes with the binding of both Pex3p and Pex11␤p to Pex19p and destabilizes PMP70. Ineffectiveness of the catalytically inactive C113S mutant suggested an important role of the enzyme activity in this action of PLA/AT-3.
Effects of PLA/AT-1 and -2-We also examined whether other members of the PLA/AT family show this inhibitory effect. The FLAG-tagged PLA/AT-2 (PLA/AT-2-FL) inhibited the binding of Pex19p to Pex11␤p potently, whereas PLA/AT-1-FL hardly inhibited this binding (Fig. 9B). Both PLA/AT-1 and -2 reduced the PMP70 levels in the presence or absence of recombinant Pex19p (Fig. 9C).
Interaction of PLA/AT-3 with Pex3p-As shown in Fig. 4A, Pex3p was also coimmunoprecipitated with PLA/AT-3. In order to clarify which domain of PLA/AT-3 is necessary for this interaction, we examined a possible co-immunoprecipitation of Pex3p-HA with PLA/AT-3-C113S-FL, PLA/AT-3-⌬N10-FL, and PLA/AT-3-⌬C27-FL (Fig. 10). The results showed that Pex3p was immunoprecipitated with the C113S and ⌬N10 mutants as well as wild-type PLA/AT-3, whereas the ⌬C27 mutant was ineffective. These results suggested that PLA/AT-3 binds to Pex3p through the C-terminal hydrophobic domain in an enzyme activity-independent manner.

Discussion
We previously reported that the overexpression of recombinant PLA/AT-3 (also referred to as HRASLS3, H-rev107, or AdPLA) causes specific disappearance of peroxisomes in mammalian cells (27). In the present study, we first confirmed this finding (Figs. 1-4). In FL-PLA/AT-3-Tet-on cells, the induction of PLA/AT-3 with DOX readily decreased the abundance of peroxisomes labeled with DsRed2-Peroxi, and the peroxisomes almost completely disappeared within 24 h. The dysfunction of peroxisomes was also revealed by the decreased levels or abnormal distribution of PMPs (PMP70, Pex3p, Pex11␤p, and Pex16) and peroxisomal matrix proteins (catalase, ADHAPS, and thiolase) as well as large reductions in the levels of ether-type lipids, such as MADAG, plasmenylethanolamine, and plasmanylcholine. An obvious decrease in PMP70 level was observed even with a low concentration (0.01 g/ml) of DOX, and in the presence of this concentration of DOX, the PLA 1 /A 2 activity was 10-fold higher than that of DOX-free con-  C113S)), and the cell lysates were analyzed by Western blotting with the indicated antibodies. Total RNAs were also isolated and analyzed by RT-PCR using the indicated primers.
trol cells (data not shown). Thus, it appeared that the expression level of PLA/AT-3, increasing PLA 1 /A 2 activity 10-fold over the control cells, is sufficient to see the loss of peroxisomes. It has been reported that the quality and quantity of peroxisomes are maintained by peroxisome-specific autophagy, pexophagy (11,13). However, during the DOX-dependent reduction of peroxisomes, the activation of LC3, a marker of autophagy, was not observed. Thus, we could not show evidence that PLA/AT-3 stimulates the degradation of peroxisomes. Because it is thought that the level of peroxisomes is regulated by the balance between their biogenesis and degradation (11), we next focused on a possible inhibition of peroxisome biogenesis by PLA/AT-3.
Among peroxins responsible for peroxisome membrane biogenesis, Pex19p is mainly localized in cytoplasm, whereas Pex3p and Pex16p are mainly associated with peroxisomes (1). In the present study, immunoprecipitation assays revealed that PLA/AT-3 binds to both Pex19p and Pex3p but not to Pex16p. Compared with the expression of PLA/AT-3 alone, co-expression of PLA/AT-3 with Pex19p caused an increase in the protein level and PLA 1/2 activity of PLA/AT-3. Using different PLA/AT-3 mutants, we showed that PLA/AT-3 binds to Pex19p through its N-terminal proline-rich and C-terminal hydrophobic domains. PLA/AT-3-C113S, a single point mutant devoid of the enzyme activity, was still capable of binding to Pex19p, suggesting that the binding is independent of its enzyme activity. The nuclear mislocalization assay using Pex19p-NLS also revealed an interaction between PLA/AT-3 and Pex19p. Taken together, these findings strongly suggested that PLA/AT-3 is a novel Pex19p-binding protein. PLA/AT-3 also bound to Pex3p through its C-terminal hydrophobic domain. Because Pex3p is one of the Pex19p-binding proteins (41), it remains unclear whether this binding occurs directly or indirectly through the binding of PLA/AT-3 to Pex19p.
One of the most important functions of Pex19p is to capture PMPs in the cytoplasm or on the ER membrane and transport them to peroxisomes (9,42,43). Although it is still controversial where Pex19p binds to PMPs, a growing body of evidence has suggested that nascent PMPs existing on the ER membrane are assembled to form nascent peroxisomes (preperoxisomal vesicles) and that the nascent peroxisomes bud from ER and are transported to preexisting peroxisomes in a Pex19p-dependent manner (42, 43) (Fig. 11A). Moreover, a recent report demonstrated that Pex16p is transported to peroxisomes from a spe-cialized domain in ER membrane in a Sec16p-dependent manner (44). Therefore, if PLA/AT-3 attaches to the ER membrane as the PLA/AT-3⅐Pex19p complex, PLA/AT-3 may act as a PLA 1/2 and acyltransferase enzyme toward glycerophospholipids of the membrane and disrupt the membrane structure required for the formation and budding of nascent peroxisomes (Fig. 11B). Alternatively, PLA/AT-3 may be transported to peroxisomes by Pex19p and enzymatically decompose membrane phospholipids of existing peroxisomes (Fig. 11C). These two possibilities are supported by the previous reports that a certain population of PLA/AT-3 localizes to the ER (22) or to peroxisomes (27) and may explain why the C113S mutant fails to disturb peroxisomal functions despite its ability to bind to Pex19p. However, there is no direct evidence for these possibilities, and further experiments will be needed to clarify where and how PLA/AT-3 as an enzyme decomposes the peroxisome membrane structure.
Our results also indicated that PLA/AT-3 inhibits the binding of Pex19p to PMPs, such as Pex3p and Pex11␤p. This finding suggests that PLA/AT-3 causes mislocalization of PMPs. Interestingly, despite the ability to bind to Pex19p, the C113S mutant failed to inhibit the binding of Pex19p to Pex3p and Pex11␤p. However, it remains unclear why the enzyme activity of PLA/AT-3 is necessary for this inhibitory effect.
Previously, we suggested that the overexpression of PLA/ AT-2 (HRASLS2) also causes the dysfunction of peroxisomes (28), whereas PLA/AT-1 (HRASLS1, A-C1) did not appear to affect peroxisomal functions (45). In the present study, we observed that PLA/AT-2 inhibited the binding of Pex19p to Pex3p and Pex11␤p and decreased the protein level of PMP70. As for PLA/AT-1, this inhibitory effect was not clearly seen, and a moderate effect on the PMP70 level was observed. These results may be related to the fact that PLA/AT-3 is closer to PLA/AT-2 than PLA/AT-1 in terms of primary structures (26).
Peroxisomes are responsible for many metabolic events, including the synthesis of precursors of ether-type lipids, including MADAG. PLA/AT-3 is abundant in adipose tissues of mice, rats, and humans (22,24,25), and its expression is further induced in obesity (19). Furthermore, the expression level of PLA/AT-3 was highly enhanced through the differentiation of mouse adipogenic cell line 3T3-L1 cells into adipocyte-like cells (18,22). Interestingly, MADAG levels rapidly decrease as the adipogenesis proceeds in 3T3-L1 cells (46). Given our present results, the negative regulation of peroxisome biogenesis by the increasing PLA/AT-3 may be one of the reasons for the decrease in MADAG levels. Jaworski et al. (19) reported that the mice deficient in PLA/AT-3 (AdPLA) have a defect in the growth of lipid droplets in adipocytes, causing a lean phenotype. This finding was explained by high rate of lipolysis caused by the marked reduction in the signaling pathway of the prostaglandin E 2 -EP3 axis because the deficiency in PLA/AT-3, acting as arachidonic acid-releasing PLA 2 , results in a decrease of prostaglandin E 2 levels. The elimination of adipocyte peroxisomes by the adipocyte-specific disruption of the Pex5p gene in mice was reported to lead to increased fat mass and reduced lipolysis in white adipose tissue (47), suggesting the inverse correlation between peroxisome level and lipid droplet production in adipose tissues. Thus, the dysregulation of peroxisomes in adipose tissues may be another reason for the thinness of PLA/AT-3-deficient mice.

Down-regulation of Peroxisomes by PLA/AT-3
Because PLA/AT-3 was originally isolated as a tumor suppressor negatively regulating the oncogene H-Ras (14), it is of particular interest to examine how PLA/AT-3 regulates the oncogenic activity of H-Ras. H-Ras is modified by prenylation at the C-terminal cysteine residue in the CAAX motif and palmitoylation at two cysteine residues (48). These lipid modifications are required for the intracellular trafficking of H-Ras from the endomembrane system to the plasma membrane (49 -51). Recently, Wang et al. (52) reported that PLA/AT-3 binds to H-Ras and decreases the level of the GTP-loaded active form of H-Ras by inhibiting its palmitoylation in an enzyme activity-dependent manner. However, it remains unclear how PLA/ AT-3 enzymatically inhibits the palmitoylation. Based on our present results, PLA/AT-3 may affect the oncogenic activity and/or intracellular localization of H-Ras by altering the membrane structure through the reduction in the content of ether-type lipids. It was also reported that aggressive cancer cells possess significantly higher levels of ether-type lipids than less aggressive counterparts (53). Aggressive cancer cells highly expressed ADHAPS, an enzyme involved in the biosynthesis of the ether-type lipids, and the transformation of less aggressive cancer cells with H-Ras up-regulated ADHAPS expression, leading to an increase in the levels of ether-type lipids. This change also increased the levels of oncogenic signaling lipids, such as ether-type and ester-type lysophosphatidic acids and eicosanoids. Similarly, the tumor suppressing activity of PLA/AT-3 may be related to the decreased levels of ether-type lipids resulting from the dysfunction of peroxisomes. Further studies on PLA/AT-3 are required to elucidate its role in obesity and cancers.
In conclusion, our results showed for the first time that PLA/ AT-3 is a Pex19p-binding protein and inhibits the chaperone functions of Pex19p. This action may be responsible for the negative regulation of PLA/AT-3 on peroxisome biogenesis in mammalian cells.