Pex18p is constitutively degraded during peroxisome biogenesis.

Pex18p and Pex21p are structurally related yeast peroxins (proteins required for peroxisome biogenesis) that are partially redundant in function. One or the other is essential for the import into peroxisomes of proteins with type 2 peroxisomal targeting sequences (PTS2). These sequences bind to the soluble PTS2 receptor, Pex7p, which in turn binds to Pex18p (or Pex21p or possibly both). Here we show that Pex18p is constitutively degraded with a half-time of less than 10 min in wild-type Saccharomyces cerevisiae. This degradation probably occurs in proteasomes, because it requires the related ubiquitin-conjugating enzymes Ubc4p and Ubc5p and occurs normally in a mutant lacking the Pep4p vacuolar protease. The turnover of Pex18p stops, and Pex18p accumulates to a much higher than normal abundance in pex mutants in which the import of all peroxisomal matrix proteins is blocked. This includes mutants that lack peroxins involved in receptor docking at the membrane (Deltapex13 or Deltapex14), a mutant that lacks the peroxisomal member of the E2 family of ubiquitin-conjugating enzymes (Deltapex4), and others (Deltapex1). This stabilization in a variety of pex mutants indicates that Pex18p turnover is associated with its normal function. A Pex18p-Pex7p complex is detected by immunoprecipitation in wild type cells, and its abundance increases considerably in the Deltapex14 peroxisome biogenesis mutant. Cells that lack Pex7p fail to stabilize and accumulate Pex18p, indicating an important role for complex formation in the stabilization. Mono- and diubiquitinated forms of Pex18p are detected in wild-type cells, and there is no Pex18p turnover in a yeast doa4 mutant in which ubiquitin homeostasis is defective. These data represent, to the best of our knowledge, the first instance of an organelle biogenesis factor that is degraded constitutively and rapidly.

Peroxisome biogenesis proceeds via a complex, branched pathway, in which a cellular machinery consisting of more than 20 proteins (peroxins) effects the recognition, targeting, and import of proteins containing peroxisomal targeting sequences (PTSs) 1 (reviewed in Refs. [1][2][3][4][5]. Multiple classes of PTS exist, of which one of the best characterized is the PTS2 family of NH 2 -terminal oligopeptides (6 -9) utilized by thiolase and sev-eral other peroxisomal proteins, which are imported into peroxisomes via interaction with Pex7p, the PTS2 receptor (10 -15). The importance of this pathway is highlighted by the observation that it is evolutionarily conserved between yeast and humans and by the fact that loss of the PTS2 branch of peroxisomal biogenesis through PEX7 mutation causes the lethal disorder rhizomelic chondrodysplasia punctata in humans (12)(13)(14), and inviability on oleic acid as carbon source in yeast.
We recently identified a novel pair of yeast peroxins, Pex18p and Pex21p, which interact with Pex7p and are essential for PTS2 targeting (16). Pex18p and Pex21p are weakly homologous to each other and appear to demonstrate partial functional redundancy, in that full loss of PTS2 targeting requires the absence of them both. Thus, although thiolase is the only known Saccharomyces cerevisiae PTS2 protein, at least three peroxins are involved in its delivery to peroxisomes. Several reports concerning different species place Pex7p within peroxisomes and/or within the cytosol, raising the intriguing possibility that Pex7p acts as a mobile receptor, shuttling between the cytosol and peroxisomal surface and/or lumen (10,11,12,15). The initial characterization of Pex18p/Pex21p identified these proteins as potentially key players in this proposed "mobile receptor" mechanism.
In addition to the PTS2-specific peroxins (Pex7p, Pex18p, and Pex21p), many other peroxins are required for import of peroxisomal proteins, including, but not limited to, those with a PTS2. These peroxins are believed to include components of a common translocation machinery acting downstream of the point at which the various branches of peroxisomal protein import converge. Examples include the peroxisomal membrane peroxins Pex13p and Pex14p, which have been reported to interact with both Pex7p and the PTS1 receptor, Pex5p, consistent with roles in a common docking site for the PTS1 and PTS2 branches (18 -24). Still other peroxins, including Pex1p, Pex4p, and Pex6p, are required for both PTS1 and PTS2 packaging but not for peroxisomal membrane assembly; their molecular roles have yet to be established. Pex1p and Pex6p are AAA ATPases localized in various yeasts and mammals to either cytosol, peroxisomes, or vesicles, depending upon the species under investigation, but not yet firmly localized in S. cerevisiae (25)(26)(27)(28)(29)(30)(31)(32)(33). Pex4p is a ubiquitin-conjugating enzyme for which substrates have not been identified, localized to the peroxisome membrane (34,35).
In this paper, we report that the functioning of Pex18p and Pex21p is accompanied by their rapid proteolytic turnover and that this pathway of peroxin degradation is obligatorily connected to ongoing peroxisome assembly. This represents, to the best of our knowledge, the first instance of an organelle biogenesis factor that is constitutively degraded during its normal function.
(prepared as slices excised from SDS-PAGE gels) corresponding to the NH 2 -terminal 156 residues of Pex7p and the COOH-terminal 200 residues of Pex18p were purified as His 6 -tagged proteins from overexpressing bacterial clones as described previously (16). The Pex7p antigen was prepared in the presence of 6 M urea to overcome its low solubility. The anti-Pex18p serum was affinity-purified on amino-link columns (Pierce) coupled with His 6 -tagged antigen, eluted with 0.1 M glycine, pH 2.5, and adjusted to pH 8.0 with Tris base. 10 ml of sera yielded 0.8 ml of purified antibody.
Yeast Strains, Media, and Growth Conditions-All of the pex mutants described in this paper were generated by targeted disruption from wild-type haploid W303 of the corresponding PEX genes. W303⌬pex5, W303⌬pex7, and W303⌬pex18 and have been described previously (16). W303⌬pex1, in which the entire PEX1 coding region is replaced with the LEU2 marker gene, was a kind gift of Dr. Marek Skoneczny. W303⌬pex4 was generated by transformation of W303 with the HindIII-NotI insert of plasmid pPEX4::TRP1 (see below) and selection of tryptophan auxotrophs. All double knockout strains were generated by successive transformations with the appropriate disruption constructs. W303⌬pep4 was generated by transformation of W303 with the EcoRI-XbaI insert of plasmid pBJ3287 (a kind gift of Dr. Beth Jones, Carnegie Mellon University) and selection of histidine auxotrophs. All new disruption strains were checked by polymerase chain reaction and/or Western blotting. MHY623, in which DOA4 has been disrupted, and the congenic wild-type strain MHY501 (36) were kind gifts of Dr. Marc Hochstrasser (University of Chicago). Y0238, in which UBC4 and UBC5 have been disrupted, and the congenic wild-type strain Y0002 were kind gifts of Dr. Stefan Jentsch (University of Heidelberg).
For induction of peroxisomes, cells were precultured to a density of 5 ϫ 10 7 /ml in YPD (1% yeast extract, 2% peptone, 2% dextrose) and then inoculated at a density of 5 ϫ 10 6 /ml into YPEO (1% yeast extract, 2% peptone, 2% ethanol, 0.1% oleic acid, 0.25% Tween 40) and grown for a further 18 h. Where appropriate, cycloheximide was then added to a concentration of 10 g/ml (0.2 l/ml of a 50 mg/ml stock solution in ethanol), and growth was continued for up to 30 min, prior to protein extraction. For expression of hemagglutinin epitope-tagged ubiquitin (HA-Ub) from plasmid pKR81, galactose was added to the YPEO at a concentration of 0.01%. For strains harboring plasmids, preculturing was in SCD (0.67% yeast nitrogen base, 2% dextrose supplemented with all amino acids, adenine, and uracil, except those omitted to maintain plasmid selection). For some experiments, YPEO was replaced with YPE (1% yeast extract, 2% peptone, 2% ethanol). For assessment of Pex18p levels in YPD (see Fig. 1B), aliquots of YPD precultures (2 ϫ 10 7 cells/ml) were removed and processed for protein extraction (see below). For the pulse labeling shown in Fig. 2B, peroxisome induction was in SCEO lacking methionine (0.67% yeast nitrogen base, 2% ethanol, 0.1% oleic acid, 0.25% Tween 40 supplemented with adenine, uracil, and all amino acids except methionine) for 18 h, after which cells were diluted into fresh SCEO lacking methionine, supplemented with 100 Ci/ml [ 35 S]methionine, to a density of 5 ϫ 10 7 cells/ml, and incubated for a further 5 min prior to protein extraction.
Plasmids-For FLAG tagging, the complete coding sequence of the PEX18 gene was excised as a BamHI fragment from clone pYcp-PEX18 (16) and cloned into the BglII site of pESC-LEU (Stratagene, La Jolla, CA). This produced a sequence encoding an in-frame fusion between the FLAG tag and the amino terminus of PEX18, which was then excised as an SpeI fragment and cloned into the XbaI site of Yep351-POT1, which contains the POT1 (thiolase) promoter in the episomal vector Yep351. This produced a clone encoding FLAG-Pex18p under the regulation of the oleate-responsive POT1 promoter. To produce a similar clone lacking the FLAG tag, the PEX18 gene was excised as an SpeI fragment from clone pYcp-PEX18 (16) and cloned directly into the into the XbaI site of Yep351-POT1. Plasmid pKR81, encoding HA-Ub (37) under the regulation of the GAL1 promoter, was a kind gift of Dr. K. Robzyk (Memorial Sloan Kettering Cancer Center, New York). For construction of plasmid pPEX4::TRP1, which was used for targeted disruption of PEX4 from W303 wild-type yeast, PEX4 was polymerase chain reaction-amplified from W303 genomic DNA using primers 5Ј-CCTGTT-GCTTTACACCATTGA-3Ј (maps to 593 bases upstream of PEX4 coding region) and 5Ј-GATACTGCAGTCAATGGTTGTTGATCCGCT-3Ј (maps to the 3Ј end of PEX4 coding sequence). This polymerase chain reaction product was cloned into pCR-TOPO2.1 (Invitrogen, Carlsbad, CA), linearized with StuI, and ligated to a SmaI-StuI fragment containing the TRP1 gene. A fragment containing the TRP1 gene flanked by PEX4 sequences was excised from the resultant construct and used for transformation of W303. PEX4 knockouts were verified by polymerase chain reaction.
RNA and Protein Methods-Preparation of total cellular RNA by the hot acidic phenol method and Northern blotting were as described (38). 32 P-Labeled probes were prepared by random priming from the coding regions of yeast PEX18 and actin genes, the latter of which were a kind gift of Drs. Igor Karpichev and Gillian Small (Mount Sinai School of Medicine, New York). Radioactive bands on the washed blots were visualized by autoradiography or with a PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA).
Preparation by glass bead homogenization of total cellular protein extracts for immunoblotting was as described previously (16). Preparation by glass bead homogenization of total cellular protein extracts for immunoprecipitation with anti-Pex18p (Figs. 2B and 4B) was by this same method but using IP buffer (50 mM Tris-HCl, pH 7.4, 1 mM EDTA, 150 mM NaCl, 0.1% Triton X-100, supplemented as usual with protease inhibitor mixture (39)) in place of our normal protein extraction buffer. Immunoprecipitations, with affinity-purified anti-Pex18p, were as described previously (16).
For preparation of total protein extracts prior to immunoprecipitation of Pex18p-ubiquitin conjugates with anti-FLAG, a modified extraction procedure was employed in order to minimize the risk of proteolytic degradation of these conjugates. Cells were grown in 25 ml of YPEO and washed and collected as usual. All subsequent steps were performed at room temperature. Pellets were resuspended in 0.5 ml of 40% trichloroacetic acid and centrifuged at 14,000 rpm. The supernatant was discarded, and the pellets were briefly frozen in liquid nitrogen and then thawed and resuspended in 0.5 ml of 20% trichloroacetic acid. Glass beads were added, and the samples were then vortexed vigorously for 2 min, after which the extracts were decanted from the beads and centrifuged at 3000 rpm for 10 min. The resultant pellets were resuspended in 200 l of SDS-PAGE sample buffer, and 50 l of unbuffered 2 M Tris base was added (to adjust the pH). Samples were boiled for 3 min and centrifuged at 3000 rpm for 10 min. 200-l aliquots of the blue supernatant were then recovered and added to 770 l of fresh IP buffer (supplemented with 0.5 g/ml bovine serum albumin) and 30 l of anti-FLAG M2-agarose affinity gel (Sigma). Samples were rotated at 4°C for 1 h, after which the anti-FLAG M2-agarose affinity gel was recovered by centrifugation at 1000 rpm for 1 min and washed twice with 1 ml of IP buffer. Bound proteins were eluted with 40 l of 0.2 mg/ml FLAG peptide (Sigma) in IP buffer. The final eluate was processed as usual for SDS-PAGE.
SDS-PAGE and immunoblotting were as described previously (16), except that prior to blotting with anti-ubiquitin, the nitrocellulose membrane was autoclaved for 20 min. This greatly enhances the signal from both free ubiquitin and ubiquitin conjugates (data not shown). Antibody dilutions used were as follows: 1:100 affinity-purified anti-Pex18p, 1:500 anti-Pex7p serum, and 1:100 anti-ubiquitin (Sigma). For the pulse-labeling experiment shown in Fig. 2B, the SDS-PAGE gel was incubated in 1 M sodium salicylate for 1 h prior to drying and fluorography.

Pex18p Abundance Varies Markedly between Wild Type and pex Mutant
Strains-A polyclonal antiserum raised against bacterially expressed Pex18p (see "Experimental Procedures") was used to monitor Pex18p expression levels in wild-type and various pex mutant yeast strains. Pex18p could be detected in extracts from oleate-induced wild-type yeast as a doublet with an estimated size of about 36 kDa (see Fig. 1), compared with the predicted size, based on the sequence of the PEX18 gene, of ϳ32 kDa (16). The reason for this apparent size difference is unclear, but it was also observed for bacterially expressed Pex18p (which likewise runs as a doublet), indicating that its mobility in SDS-PAGE may be an intrinsic feature of the protein rather than due to covalent modifications. Treatment of extracts with iodoacetamide, phosphatases, or hydroxylamine did not alter the Pex18p gel migration pattern (data not shown). Both bands of the doublet were absent in extracts from W303⌬pex18 yeast, confirming that they correspond to Pex18p (Fig. 1A).
Remarkably, the abundance of Pex18p was greatly increased in strains with generalized defects in the import of peroxisomal matrix proteins (W303⌬pex1, W303⌬pex4, W303⌬pex13, and W303⌬pex14) (Figs. 1 and 3B). On the other hand, Pex18p abundance was close to normal in W303⌬pex7 and W303⌬pex5, which display partial defects in this process.
In common with many peroxisomal proteins and peroxins, Pex18p expression was oleate-inducible, as shown by the apparent absence or near absence of immunoreactive protein in strains grown in YPD or YPE (Fig. 1B).
The Variation among Strains in Pex18p Abundance Is Not Primarily Due to Variations in PEX18 mRNA Levels or Pex18p Synthetic Rates-To investigate the mechanism of the variation in Pex18p levels, RNA was extracted from oleate-induced strains and analyzed by Northern blotting. PEX18 mRNA levels were similar to the wild-type level in several pex mutants, including W303⌬pex5, W303⌬pex7, W303⌬pex13, and W303⌬pex14 (shown for W303⌬pex14 in Fig. 2A), suggesting that the variation in Pex18p levels is not primarily due to differences in abundance of PEX18 message. In agreement with the immunoblot data shown in Fig. 1B, PEX18 mRNA levels were much lower in the absence of oleate. The rate of Pex18p synthesis was tested directly by pulse-labeling oleategrown cells with [ 35 S]methionine and then analyzing incorporation into Pex18p by immunoprecipitation with anti-Pex18p, followed by SDS-PAGE and fluorography (Fig. 2B). Little difference was observed between wild type and W303⌬pex14, suggesting that Pex18p is made at similar rates in these strains.
Pex18p Is Rapidly Turned Over in Wild-type Cells but Not in Certain pex Mutants-Since neither the levels of PEX18 mRNA nor the rates of Pex18p synthesis vary substantially between strains with grossly different levels of Pex18p, we next addressed the question of Pex18p turnover. Cells were grown and induced as usual, and then growth was continued in the presence or absence of the protein synthesis inhibitor, cycloheximide. As shown in Fig. 3A, Pex18p rapidly disappeared from wild-type cells treated with cycloheximide, being barely detectable within 20 min. In contrast, the high level of Pex18p in W303⌬pex14 cells was maintained after 20 min in cycloheximide (Fig. 3A) and showed no discernible decline even 5 h after administration of cycloheximide (not shown). The turnover of Pex18p in W303⌬pex7 cells, which maintain Pex18p at a level similar to that seen in wild-type cells, is similar to the turnover rate in wild-type cells (Fig. 3A).
Analysis of additional pex mutants revealed a consistent pattern. Mutants that have a generalized defect in peroxisome biogenesis accumulated high levels of Pex18p, which turned over slowly, if at all, whereas W303⌬pex7 and W303⌬pex5, which are defective only for PTS2 and PTS1 targeting, respectively, behaved similarly to wild type (Fig. 3B). Thus, in mutants W303⌬pex13 and W303⌬pex14, where Pex18p is unable to fulfill its role of PTS2 protein delivery to peroxisomes, Pex18p is stabilized. Notably in W303⌬pex4, which lacks the peroxisomal member of the ubiquitin-conjugating enzyme family, Pex18p is likewise stabilized. Stabilization also occurs in W303⌬pex1. In W303⌬pex5, where PTS2 packaging is functional, Pex18p is turned over as in wild-type cells. Pex18p is known to interact with Pex7p (16). The instability of Pex18p in W303⌬pex7 suggests that Pex18p may be stabilized by its interaction with Pex7p.
Pex21p Also Turns Over Rapidly-Pex21p, whose function can partially compensate for an absence of Pex18p (16), also turns over rapidly in wild-type yeast (data not shown). In cells defective in peroxisome biogenesis (W303⌬pex13 or W303⌬pex14), Pex21p is stabilized little, if at all, unless Pex18p is missing, in which case there is a striking stabilization and accumulation of Pex21p. As for Pex18p, this stabilization requires Pex7p to be present.
Pex7p Forms a Complex with Pex18p and Is Required for Its Stabilization-In order to investigate further the possible involvement of Pex7p in Pex18p stabilization, we disrupted the PEX7 gene from several pex mutants and assessed the consequences for Pex18p accumulation and turnover. As shown in Fig. 4A, the absence of Pex7p completely abolishes Pex18p stabilization in the W303⌬pex13 background. W303⌬pex13⌬pex7 appeared identical to W303⌬pex7 with respect to the abundance of Pex18p and its turnover. Similar Pex18p instability was observed in W303⌬pex14⌬pex7 (Fig.  4A) and W303⌬pex4⌬pex7 (not shown). These data support the idea that Pex18p becomes stabilized in pex mutants through persistence of its interaction with Pex7p. FIG. 1. Pex18p abundance varies enormously between wildtype and pex mutant yeast strains. A, 50-g aliquots of total cellular protein extracted from YPEO-grown yeast were analyzed by immunoblotting with anti-Pex18p antisera. Pex18p runs as a characteristic doublet with an apparent molecular mass of ϳ36 kDa, with the higher mobility form being the more abundant. Both bands are absent from extracts of W303⌬pex18 cells. A third cross-reacting band (*), with an apparent molecular mass of ϳ40 kDa, is unrelated to Pex18p, being present in equal amounts in wild-type and ⌬pex18 extracts. B, samples were prepared and analyzed as above, except growth was in YPEO, YPE, or YPD.

FIG. 2. Strains with vastly different levels of Pex18p protein
show little variation in PEX18 mRNA abundance or synthetic rate of Pex18p. A, total RNA was prepared from yeast grown in YPEO or YPE, and 20-g aliquots were analyzed by Northern blotting with probes against PEX18 and actin (as a loading control). B, yeast cells grown in SCEO without methionine were pulse-labeled with radiolabeled methionine. Pex18p was immunoprecipitated from total cellular protein extracts (prepared by glass bead homogenization); its radioactivity was visualized by SDS-PAGE and fluorography.
To test directly whether the elevated Pex18p is associated with Pex7p, anti-Pex18p immunoprecipitates were immunoblotted with anti-Pex7p antisera (Fig. 4B). Pex7p was co-immunoprecipitated with Pex18p from wild-type cells; the amount of immunoprecipitated Pex7p was considerably greater with W303⌬pex14 cells. In neither case was immunoprecipitation of Pex7p complete, suggesting that the cellular Pex7p is not saturated with Pex18p, even when the latter is accumulated to the levels seen in W303⌬pex14.
Pex18p Is Ubiquitinated-To investigate the mechanism of the rapid turnover of Pex18p in wild-type cells, we first analyzed Pex18p turnover in strains defective in various aspects of degradation of cellular proteins. W303⌬pep4, which lacks the vacuolar proteinase A, had wild-type levels of Pex18p abundance and turnover (Fig. 5). On the other hand, Y0238, which lacks two functionally related ubiquitin-conjugating enzymes, Ubc4p and Ubc5p, showed accumulation of high levels of Pex18p and impaired Pex18p turnover. A congenic wild type strain (Y0002; Fig. 5) and various other ubc mutants, including a strain lacking Ubc6p and Ubc7p (Y0241; not shown), demonstrated normal Pex18p levels and rates of turnover. Finally, we analyzed strain MHY623, in which ubiquitin homeostasis is severely impaired due to a lack of the deubiquitinating enzyme, Doa4p (36). Cells lacking Doa4p have decreased ubiquitin levels and display strongly reduced turnover of several proteins degraded via ubiquitination. As shown in Fig. 5, Pex18p turnover is severely deficient in ⌬doa4 yeast (but normal in a congenic wild-type strain).
Higher molecular weight forms of Pex18p (which would be consistent with the existence of ubiquitinated degradation intermediates) could not be detected in immunoprecipitations and immunoblots with anti-Pex18p serum (not shown). However, when FLAG-tagged Pex18p was immunoprecipitated directly from total cellular protein extracts (prepared by glass bead homogenization in the presence of trichloroacetic acid), two higher molecular weight species in addition to FLAG-Pex18p were discernible (Fig. 6A, left panel). The stronger of these bands had an estimated mass about 8 kDa greater than FLAG-Pex18p, consistent with the presence of a single 76amino acid ubiquitin moiety (or related peptide), and the fainter band was a further 8 kDa larger. Curiously, the larger, but not the smaller of these putative ubiquitin conjugates could be detected with anti-ubiquitin antiserum (Fig. 6A, right panel). Co-expression of ubiquitin tagged with the hemagglutinin epitope (HA-Ub, 85 amino acids) reduced the intensities of both of these bands and generated two slightly larger new bands (Fig. 6B, arrowheads), indicating the incorporation of one or two HA-Ub moieties into FLAG-Pex18p. Taken together, these results indicate that Pex18p is ubiquitinated. FIG. 3. The variation in Pex18p abundance between wild-type and pex mutant yeast strains is a result of differential degradation rates. A, yeasts were grown as usual in YPEO for 18 h, at which point 10 g/ml cycloheximide was added to the medium, and growth resumed for the indicated times (in minutes). Total protein was extracted, and immunoblot analysis was performed using anti-Pex18p. B, after 18 h of YPEO growth, cultures were divided, and growth was continued for 30 min in the presence (ϩC) or absence (ϪC) of 10 g/ml cycloheximide. Pex18p was detected as in A.

FIG. 4. Pex7p forms a complex with
Pex18p and is required for Pex18p stabilization. A, epistatic analysis of Pex18p stabilization and turnover. The indicated single and double pex knockout strains were analyzed as described in the legend to Fig. 3B. B, immunoprecipitation of Pex7p-Pex18p complexes. 200-g aliquots of total cellular protein extracts from YPEO-grown yeast strains were subjected to immunoprecipitation with anti-Pex18p. The immunoprecipitates were analyzed alongside 50-g aliquots of the extracts by immunoblotting with anti-Pex18p and anti-Pex7p antisera.

DISCUSSION
In this paper, we report that Pex18p, a peroxin previously shown to be essential for peroxisome biogenesis via the PTS2 pathway, turns over rapidly during normal peroxisome biogenesis in wild-type cells. This is, to the best of our knowledge, the first report of an organelle biogenesis factor that is constitutively degraded in the course of its normal function. In pex mutants in which peroxisomal matrix protein import is completely abolished, Pex18p undergoes a remarkable increase in stability and abundance. Stabilization of Pex18p is entirely dependent upon the presence of Pex7p, the PTS2 receptor, with which it forms a complex both in vivo and in vitro. Pex18p turnover is impaired in cells lacking the ubiquitin-conjugating enzymes Ubc4p and Ubc5p and is deficient in ⌬doa4 cells, which display a generalized defect in ubiquitin homeostasis. Finally, mono-and diubiquitinated forms of Pex18p can be detected in wild-type cells during unperturbed peroxisome biogenesis.
The finding that Pex18p is rapidly turned over in wild-type cells growing on oleic acid was unexpected. Such turnover has not been described for mitochondrial or chloroplast biogenesis, nor for other organelles, so far as we are aware. Various peroxins, including Pex5p and several peroxisomal membrane proteins, have been shown to display altered abundance in certain pex mutants (40 -43), but rapid turnover of any peroxin during normal peroxisome biogenesis has not been reported. The observed rate of Pex18p degradation is as fast as that of some transcriptional regulators, for example the ␣2 homeodomain protein encoded by the MAT␣ locus (44).
Pex18p is stabilized in pex mutants with diverse defects. These include W303⌬pex13 and W303⌬pex14, which lack peroxisomal membrane proteins implicated in Pex7p docking at the peroxisomal surface, and mutants lacking either Pex4p, a peroxisomal member of the E2 family of ubiquitin-conjugating enzymes (35), or Pex1p, a peroxin whose function is uncertain. The common property of these mutants is a defect in the assembly of all peroxisomal matrix proteins. In contrast, in cells lacking PTS1 import, but with functional PTS2 import (W303⌬pex5), Pex18p abundance, and turnover are indistinguishable from wild type. Our interpretation of these data is that Pex18p turnover accompanies its participation in peroxisomal biogenesis.
Double mutant analysis demonstrated a requirement for PEX7 for the stabilization of Pex18p. Consistent with this, Pex7p and Pex18p form a complex in cells as determined by co-immunoprecipitation with anti-Pex18p. The abundance of Pex7p does not change when PEX18, PEX13, PEX14, PEX1, or PEX4 is knocked out (data not shown). A small amount of Pex7p is associated with Pex18p in wild-type cells. As the abundance of Pex18p increases in pex mutants such as W303⌬pex14, so does the abundance of the Pex7p-Pex18p complex (Fig. 4B), but it is noteworthy that much of the cellular Pex7p remains unprecipitated by anti-Pex18p even under these conditions. We do not know the absolute abundance of Pex7p or of Pex18p in g/cell. The data suggest that Pex7p may be present in excess relative to Pex18p. It is possible, but not yet testable, that all of the Pex18p is normally complexed to Pex7p.
In cells lacking Pex7p but with all other peroxins intact (W303⌬pex7), Pex18p shows approximately wild-type levels of FIG. 5. Pex18p turnover is impeded in strains defective in the ubiquitin pathway. Strain Y0238 (⌬ubc4⌬ubc5) and its congenic wild type, Y0002, were grown and treated with (ϩC) or without (ϪC) cycloheximide, as described in the legend to Fig. 3B (top panel). Total cellular protein extracts were separated by SDS-PAGE and immunoblotted with anti-Pex18p. Similar comparisons were made between wild-type W303, W303⌬pex14, and W303⌬pep4 (which lacks the vacuolar proteinase A) (middle panel) and between MHY623, which is defective in the deubiquitinating enzyme Doa4p, and the congenic wild-type strain, MHY501 (bottom panel).
FIG. 6. Pex18p is ubiquitinated. A, total cellular protein extracts from W303⌬pex18 expressing FLAG-tagged Pex18p (F-Pex18p, lanes 2 and 4) or, as a control, untagged Pex18p (lanes 1 and 3) were subjected to immunoprecipitation with anti-FLAG. The immunoprecipitates were then resolved by SDS-PAGE and blotted with antisera against Pex18p  (left panel, lanes 1 and 2) and ubiquitin (right panel, lanes 3 and 4). abundance and turnover. This agrees with the finding that Pex7p is required for stabilization of Pex18p in generalized pex mutants and also suggests that the mechanisms of turnover of Pex18p synthesized in the absence of Pex7p and of Pex18p released from Pex7p during PTS2 import may be similar.
The rapid rate of turnover of Pex18p suggests that it is probably degraded by proteasomes, which degrade many proteins with short half-lives that are marked by polyubiquitination (46,47). Consistent with this, a mutant strain lacking the ubiquitin-conjugating enzymes Ubc4p and Ubc5p, which are responsible for directing many proteins to proteasomal degradation (46), showed striking stabilization and accumulation of Pex18p. The alternative of vacuolar import and proteolysis appears unlikely; an import half-time of 30 -40 min has been reported for vacuole-degraded aminopeptidase I (48), which is much slower than the observed Pex18p degradation rate. Moreover, we saw no change in Pex18p stability upon deletion of the PEP4 gene encoding the vacuolar proteinase A. However, we do not yet have definitive evidence for proteasomal involvement in Pex18p degradation; mutants lacking functional proteasomal subunits grow extremely poorly under peroxisome induction conditions. Also, targeting to proteasomes for degradation usually requires a chain of at least four ubiquitin moieties (49), whereas we have only detected mono-and diubiquitinated forms of Pex18p. More highly ubiquitinated forms of Pex18p may exist but be too transient for detection in our immunoprecipitation assay.
It is also possible that the observed mono-and diubiquitination of Pex18p may represent a signal distinct from, and perhaps preceding, commitment to proteasomal degradation. Ubiquitin plays a role in a remarkable variety of cellular functions and does not always do so by forming long chains. Apart from its role in targeting proteins to the proteasome (50), it affects DNA repair, ribosome function, mitochondrial DNA inheritance, and the stress response (51). It also contributes to the down-regulation of cell surface receptors, transporters, and ion channels (52). In S. cerevisiae, monoubiquitination of ␣-factor receptor has been shown to be sufficient to trigger its internalization by endocytosis, leading to its degradation in the vacuole (17). Other cellular proteins, including histone H2B (37), appear to be monoubiquitinated but not polyubiquitinated.
An involvement of ubiquitination in peroxisome biogenesis has been suspected, since the original identification of Pex4p as a member of the E2 family of ubiquitin-conjugating enzymes (35). However, thus far no Pex4p substrates have been reported. It is noteworthy that ⌬pex4 cells also show stabilization of Pex18p, raising the interesting possibility that Pex18p may be among the long sought after substrates for Pex4p-dependent ubiquitination. Consistent with this, Pex4p is localized (through its interaction with the integral membrane peroxin Pex22p) to the outer face of the peroxisomal membrane (45), where it might have the opportunity to interact with the docked PTS2 import complex and participate in the ubiquitination of Pex18p. Indeed, recent experiments (not shown) indicate that deletion of PEX4 alters, but does not abolish, the pattern of ubiquitination of Pex18p. This topology raises the tantalizing possibility that limited ubiquitination of Pex18p might play a role in the regulation of PTS2 targeting. However, such a role for Pex4p remains hypothetical at this time; future experiments will be required to investigate whether this might be the case.