Peroxisome Biogenesis and Selective Degradation Converge at Pex 14 p *

We have analyzed the function of Hansenula polymorpha Pex14p in selective peroxisome degradation. Previously, we showed that Pex14p was involved in peroxisome biogenesis and functions in peroxisome matrix protein import. Evidence for the additional function of HpPex14p in selective peroxisome degradation (pexophagy) came from cells defective in HpPex14p synthesis. The suggestion that the absence of HpPex14p interfered with pexophagy was further analyzed by mutational analysis. These studies indicated that deletions at the C terminus of up to 124 amino acids of HpPex14p did not affect peroxisome degradation. Conversely, short deletions of the N terminus (31 and 64 amino acids, respectively) of the protein fully impaired pexophagy. Peroxisomes present in these cells remained intact for at least 6 h of incubation in the presence of excess glucose, conditions that led to the rapid turnover of the organelles in wild-type control cells. We conclude that the N terminus of HpPex14p contains essential information to control pexophagy in H. polymorpha and thus, that organelle development and turnover converge at Pex14p.

Hansenula polymorpha is a methylotrophic yeast that is used as a model organism in contemporary peroxisome research. Methylotrophic yeast species, also including Candida boidinii and Pichia pastoris, have the advantage that the ultrastructural changes accompanying peroxisome development and degradation are much more pronounced, relative to Saccharomyces cerevisiae. Also, enhanced numbers of growth substrate-dependent peroxisome functions can be induced compared with bakers' yeast.
In H. polymorpha highest peroxisome induction rates are observed during growth of cells on methanol. Under these conditions the organelles may occupy up to 80% of the cytoplasmic volume and are essential for growth as they contain the key enzymes of methanol metabolism, alcohol oxidase (AO), 1 cata-lase, and dihydroxyacetone synthase. These enzymes all possess a C-terminal targeting signal and require the function of the PTS1 import machinery for sorting to peroxisomes (1,2). In H. polymorpha, Pex14p is involved in matrix protein import and functions (probably together with Pex13p and Pex17p) as the putative docking site for PTS1 receptor-cargo complexes at the peroxisomal membrane (3). Recent data, however, indicate that the protein is not essential for import but most likely enhances the efficiency of the process (4).
Peroxisome degradation may occur aselectively during general autophagy (5) or in a selective way, when the organelles have become dysfunctional or, alternatively, redundant for growth. In H. polymorpha the selective degradation process is morphologically characterized by three subsequent steps, namely (i) tagging, followed by sequestration of the organelle to be degraded by multiple membranous layers, (ii) heterotypical fusion of the sequestering membranes with the vacuolar membrane, and (iii) hydrolysis of the organelle contents in the vacuole (6). This process, designated macropexophagy, is also observed in another methylotrophic yeast species, P. pastoris (7,8). In this yeast, but not in H. polymorpha, a second mode of selective degradation is described, specifically induced by glucose, which involves uptake of clusters of peroxisomes by engulfment by the vacuole (micropexophagy). Selective peroxisome degradation has also been described in the yeasts S. cerevisiae and Yarrowia lipolytica (9); however, the exact mode of degradation is as yet unknown.
Various mutants affected in pexophagy have been isolated from P. pastoris (gsa and pag mutants) (10 -13) and H. polymorpha (pdd mutants) (14,15). The analysis of the corresponding genes has shown that pexophagy has several components in common with non-selective autophagy, vacuolar protein sorting, endocytosis, and the cytosol-to-vacuole transport pathway in bakers' yeast (16,17).
In H. polymorpha pex mutants, peroxisomal remnants are normally susceptible to glucose-induced selective degradation, except for those present in ⌬pex14 cells (18). This raised the question whether Pex14p, besides being involved in matrix protein import, could have additional functions in the control of the susceptibility of individual organelles for selective degradation. This aspect was analyzed in cells of constructed mutants that contained peroxisomes that either completely lacked Pex14p or contained truncated forms of this peroxin. The results of this work are included in this paper. * 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 (4), and ⌬pex14 strains expressing full-length or truncated forms of PEX14 under control of the PEX14 promoter (P PEX14 ) (this study).
For analysis of peroxisome degradation, cells were extensively precultivated to the mid-exponential phase on mineral medium with glucose and then shifted to mineral medium containing glycerol ϩ methanol for a period of 16 h to induce peroxisome biogenesis. Subsequently 0.5% glucose was than added to induce selective peroxisome degradation. Escherichia coli DH5␣ (27) was used for plasmid construction and was grown on LB medium supplemented with the appropriate antibiotics.
Molecular Techniques-Standard recombinant DNA techniques were carried out essentially according to Sambrook et al. (27). Transformation of H. polymorpha was performed by electroporation as described previously (28). Restriction enzymes and biochemicals were obtained from Roche Molecular Biochemicals and used as detailed by the manufacturer. Protein sequences were aligned with the program CLUST-ALX (29).
Construction of Plasmids-To express WT and truncated forms of PEX14 in ⌬pex14 cells under control of its endogenous promoter, we first constructed the expression vector pHIPX10, a derivative of pHIPX4-B (20). The PEX14 promoter was isolated by PCR using the P PEX14 primer (5Ј GGG GAT CCG GTG AGG AAG AAA AAG AG 3Ј), the M13/pUC universal primer, and plasmid pBS3.2P14 (20) as template. The PCR fragment was digested with EcoRV and BamHI and inserted between the NotI (blunted) and BamHI sites of plasmid pHIPX4-B thus replacing the alcohol oxidase promoter. To obtain pX10-PEX14-WT we inserted the 1.5-kilobase BamHI-EcoRV fragment of pP AOX PEX14 (20) between the BamHI and SacI sites of pHIPX10. Plasmids expressing truncated versions of PEX14 were constructed as follows: pX10-PEX14-⌬C58 was constructed by filling in the EcoRI site in the PEX14 coding region followed by self-ligation. Similarly, pX10-PEX14-⌬C124 was obtained by blunting the PstI site in the PEX14 coding region followed by self-ligation. Plasmids pX10-PEX14-⌬N31 and pX10-PEX14-⌬N64 were constructed by PCR using the N31 primer (5Ј AGA GGA TCC ATG GCC AAA AAG GTC GAA TTT C 3Ј) or the N64 primer (5Ј AA GGA TCC ATG TCA CAG CAG TCC GTT GTA 3Ј) in combination with the M13/pUC sequencing primer and pBS1.3P14 as template. Subsequently, for pX10-PEX14-⌬N31, a BamHI ϩ HindIII-digested PCR fragment was cloned between the BamHI and HindIII sites of pHIPX10. For pX10-PEX14-⌬N64 a BamHI ϩ EcoRV-digested PCR fragment was cloned between the BamHI and SmaI sites of pHIPX10. Plasmids were transformed into H. polymorpha ⌬pex14 (leu1.1) (20) and produced Pex14p levels similar to those observed in WT cells.
Biochemical Methods-Crude extracts using trichloroacetic acid-precipitated H. polymorpha cells were prepared as described (30). SDSpolyacrylamide gel electrophoresis and Western blotting were performed by established procedures. Proteins on Western blots were detected using a chemiluminescent Western blotting kit (Roche Molecular Biochemicals) after decoration with polyclonal antibodies against various H. polymorpha proteins.
In vivo phosphorylation of HpPex14p was determined as described (31); cells were pregrown on YPD, depleted for phosphate, harvested, and subsequently suspended in the same volume of phosphate-depleted glycerol/methanol medium containing [ 32 P]orthophosphate for 16 h. 32 P-labeled H. polymorpha Pex14p was recovered from crude extracts by immunoprecipitation, separated by SDS-polyacrylamide gel electrophoresis, and detected by autoradiography.
Electron Microscopy-Whole cells were fixed and prepared for electron microscopy and immunocytochemistry as detailed before (32). Immunolabeling was performed on ultrathin section of unicryl-embedded cells, using specific antibodies against various proteins and gold-conjugated goat anti-rabbit antibodies according to the instructions of the manufacturer (Amersham Pharmacia Biotech).

RESULTS
In an initial series of experiments we have analyzed the fate of peroxisomes that had developed in cells of a pex14 deletion strain that overproduced Pex5p (⌬pex14::P AOX .PEX5 mc ), (4), after a shift of cells from methanol to glucose-excess conditions. To this end cells were pre-grown on a mixture of glycerol and methanol until an optical density (OD 663 ) of 1.5 before excess glucose (final concentration 0.5%) was added. As described before, glycerol-/methanol-grown ⌬pex14::P AOX .PEX5 mc cells contained several well developed peroxisomes (Fig. 1A) that lacked Pex14p but contained the bulk of two key enzymes in  methanol utilization, AO and dihydroxyacetone synthase (data not shown), in conjunction with minor amounts of these proteins in the cytosol (4).
Ultrastructural and biochemical analyses revealed that, upon exposure of such cells to glucose-excess conditions, selec-tive peroxisome degradation (pexophagy) was inhibited. Electron microscopy revealed that after the addition of glucose the initial event in peroxisome degradation, namely sequestration of the organelle to be degraded, was never observed (data not shown). Also, immunocytochemistry failed to demonstrate any AO protein in the vacuole, a typical morphological characteristic of pexophagy, in the same time interval (Fig. 1B). In WT control cells both phenomena were frequently observed (data not shown; see Ref. 6). Biochemical experiments showed that the amount of Pex10p, an integral component of the peroxisomal membrane, had slightly increased 4 h after addition of glucose to ⌬pex14::P AOX .PEX5 mc cells whereas in WT controls this marker protein markedly decreased during this time period (Fig. 2). Taken together these data suggest that the peroxisomes present in ⌬pex14::P AOX .PEX5 mc cells were not susceptible to glucose-induced pexophagy.
Mutational Analysis of Pex14p-To delineate the region of Pex14p controlling peroxisome turnover, we constructed mutant genes that encoded various truncated Pex14ps and transformed these into a PEX14 deletion strain (⌬pex14). These proteins lacked either the initial 31 or 64 N-terminal amino acids (designated ⌬N31 and ⌬N64, respectively) or the extreme 58 or 124 C-terminal amino acids (designated ⌬C58 and ⌬C124, respectively). A ⌬pex14 strain expressing full-length PEX14 was taken as control (designated WT). Cells of these strains were subsequently analyzed for growth on methanol, Pex14p synthesis and location, peroxisome development, and the susceptibility of these organelles to selective degradation.
The Mutant Pex14ps Are Normally Synthesized and Sorted to Peroxisomes-Cells of the various constructed strains were analyzed for their capacity to grow on methanol as the sole source of carbon. As shown in Table I only the WT control and mutant ⌬C58 grew normally on methanol at WT rates, whereas the other strains showed no or severely retarded growth. The rea-son for this became clear from electron microscopy, which revealed that WT and ⌬C58 cells displayed normal peroxisomes that were the sole site of AO protein, judged from immunocytochemistry (Fig. 3, A and B). The three other strains contained several peroxisomes of smaller size (Table I). This most likely reflects the observation that only a portion of AO protein was present in peroxisomes, whereas the remaining portion resided in the cytosol (Fig. 3, C and D). However, in these three mutants the amount of AO imported into peroxisomes was substantially higher compared with the residual import in peroxisomal remnants in ⌬pex14 cells. Because AO is a PTS1 protein, these data indicate that the long (124 amino acids) C-terminal deletion, as well as N-terminal deletions of HpPex14p, affect Pex5p-dependent protein import. We showed before that a minor amount of active cytosolic AO prevents growth of H. polymorpha cells on methanol (19). Therefore, the cytosolic portion of AO protein in ⌬C124, ⌬N31, and ⌬N64 cells most likely explains the failure of the cells to grow on methanol.
Subsequently, all constructed strains were analyzed for the presence of Pex14p by Western blotting, using ␣-Pex14p antibodies and crude extracts prepared from glycerol-/methanolgrown cells. These experiments revealed that all truncated Pex14ps were normally synthesized and were of the expected mass, judged from their migration pattern in the gel (Fig. 4A). In these blots Pex14 protein is observed as a double band at ϳ47 kDa, of which the upper band represents the phosphorylated state of the protein (20). As evident in Fig. 4B, normal phosphorylation of Pex14p can also be observed in ⌬N64 and ⌬C58 cells but not in ⌬C124 cells, in which phosphorylation of Pex14p was hardly detectable. This suggests that the putative phosphorylation site(s) of the protein are present within the region of amino acids 228 to 293.
The subcellular location of these Pex14p variants was subsequently analyzed biochemically and immunocytochemically. Upon differential centrifugation of homogenized protoplasts, prepared from methanol-induced cells of the various strains, WT and mutant Pex14ps were predominantly found in the 30,000 ϫ g organellar pellets (Fig. 5). This suggests that the proteins are indeed organelle-bound. Immunocytochemically, using ␣-Pex14p antibodies, the specific labeling was exclusively localized on the peroxisomal membrane (Fig. 6). From this we concluded that all truncated Pex14ps (WT, ⌬C58, ⌬C124, ⌬N31, and ⌬N64) were normally synthesized and sorted to the correct target membrane (summarized in Table I).
N-terminal Deletions of Pex14p Affect Selective Peroxisome Degradation-Cells of the various strains were grown on glycerol/methanol mixtures and subsequently exposed to glucoseexcess conditions. Electron microscopical analysis revealed that in the strain producing full-length Pex14p peroxisomes were normally degraded. A similar phenomenon was observed for ⌬C58 and ⌬C124 cells. 30 min after addition of glucose the first cells were observed that contained AO protein in the Gaps were introduced to maximize the similarity. Conserved residues have been shaded, and the intensity of the shading depending on the degree of conservation. vacuole (Fig. 7), a characteristic feature for selective peroxisome degradation under these conditions (6). It should be stressed that the cytosolic portion of AO is not subject to degradation under these conditions (21) and therefore cannot be the source of vacuolar AO protein.
The above observations were confirmed by biochemical data. After Western blotting of crude extracts, prepared from the various strains at different time points after the addition of glucose, a rapid decrease in the levels of Pex10p and mutant Pex14ps were observed (Fig. 8). Opposite results were obtained in identical experiments, using ⌬N31 and ⌬N64 cells. In these cells, electron microscopical analysis failed to resolve any sign of peroxisome turnover in the first 4 h after addition of glucose (Fig. 7, C and D), a time interval that is sufficient to remove the bulk of the peroxisomal population in WT control cells. Also, the levels of Pex10p and mutant Pex14p hardly diminished in this period (Fig. 8). From this we conclude that the N-terminal deletions of Pex14p prevent glucose-induced pexophagy in H. polymorpha.

DISCUSSION
In this paper we have provided evidence that the H. polymorpha PEX14 gene product (HpPex14p) plays a role in the selective degradation of peroxisomes (pexophagy). We showed that considerable portions of the C terminus of the protein could be deleted without affecting PTS1 protein import (⌬C58 cells) and pexophagy (⌬C58 and ⌬C124 cells). In contrast, deletion of the first 31 N-terminal amino acids affected both peroxisome biogenesis (by reducing the rate of PTS1 protein import) and selective peroxisome degradation (pexophagy). The general function of Pex14p in PTS1 and PTS2 matrix protein import as a component of the docking site for the cytosolic receptors of the PTS1 and PTS2 targeting signals was demonstrated before (3). Very recently, we showed, however, that in H. polymorpha Pex14p most likely is not essential for PTS1 protein import but enhances the efficiency of this process (4). Our current data imply that Pex14p may have multiple roles and functions in both the biogenesis and selective degradation of the organelle. This mode of controlling peroxisome biogenesis versus degradation (homeostasis) in one "switch" has obvious physiological advantages in that it enables the cell to rapidly adapt peroxisome numbers and function upon changes in the environment. This is particularly important in methanol-grown H. polymorpha cells, in which specific organelles have become dysfunctional through, e.g. chemically induced damage of their membrane (22) resulting in the leakage of matrix proteins into the cytosol. It has been shown before that the presence of only minor amounts of enzymatically active AO protein in the cytosol prevents growth of cells on methanol because of the major energetic disadvantages related to cytosolic hydrogen peroxide metabolism (19).
In H. polymorpha pexophagy is a remarkable fast process; individual organelles can be degraded within a time interval of 15 min. Thus, a molecular switch controlling organelle homeostasis (both biogenesis and degradation) allows taking rapid decisions and consequently, may be of great value as a component in the control of cell viability. In surface plasmon resonance analyses Schliebs and colleagues (23) demonstrated that the N terminus of human Pex14p could efficiently bind to the PTS1 receptor, Pex5p . This indicates that the N terminus of the protein may be a prime target for the PTS1 receptor-cargo complex. These data are in line with our observations that N-terminal deletions of both 31 or 64 amino acids resulted in a partial AO protein-import defect. However, the same stretch of N-terminal amino acids also appears to control pexophagy. The mechanisms on how the discrimination is made between these functions of HpPex14p are still largely speculative. One possible option is that this may be related to conformational changes of HpPex14p, for instance by oligomerization from monomers to dimers or even by alteration of the location of the N terminus of Pex14p from inside to outside the organellar matrix. On the topology of Pex14p in yeast nothing is known yet; only for human Pex14p has evidence been presented that the N terminus of the protein is (at least in part) exposed to the lumen of the organelle (24). Our experiments strongly suggest that modification of the protein by phosphorylation is most likely not discriminative in this respect.
Sequence comparison of the various Pex14ps currently available, reveals that the similarity predominantly resides in the N-terminal part of the protein (Fig. 9). This similarity becomes even stronger when only the proteins of the methylotrophic yeast species H. polymorpha and P. pastoris are compared. We have now initiated further studies to analyze the significance of the conserved residues in peroxisome homeostasis in depth.