Transcriptional Down-regulation of Peroxisome Numbers Affects Selective Peroxisome Degradation in Hansenula polymorpha*

We have isolated and characterized a novel transcription factor of Hansenula polymorpha that is involved in the regulation of peroxisomal protein levels. This protein, designated Mpp1p, belongs to the family of Zn(II)2Cys6 proteins. In cells deleted for the function of Mpp1p the levels of various proteins involved in peroxisome biogenesis (peroxins) and function (enzymes) are reduced compared with wild type or, in the case of the matrix protein dihydroxyacetone synthase, fully absent. Also, upon induction of mpp1 cells on methanol, the number of peroxisomes was strongly reduced relative to wild type cells and generally amounted to one organelle per cell. Remarkably, this single organelle was not susceptible to selective peroxisome degradation (pexophagy) and remained unaffected during exposure of methanol-induced cells to excess glucose conditions. We show that this mechanism is a general phenomenon in H. polymorpha in the case of cells that contain only a single peroxisome.

Eukaryotic cells are thought to have evolved ϳ1.5 billion years ago. The development of cell organelles allowed primitive eukaryotes to compartmentalize specific cellular functions. Concomitantly, genetic mechanisms that control the biogenesis and function of these compartments had to be developed. Obviously, the separate classes of organelles are characterized by their specific function, as in energy metabolism (mitochondrion), degradation processes (vacuole, lysosome), or protein transport (Golgi system, endoplasmatic reticulum).
Among the organelles, peroxisomes are remarkable because of their highly versatile functions most of which are related to specific metabolic pathways in the organism in which they occur. This functional flexibility is not reflected in their morphology. The organelles are invariably very simple of construction and consist of a proteinaceous matrix, surrounded by a single membrane. Nevertheless, their function varies from the oxidation of very long chain fatty acids in man, germination of oil-bearing seed and photorespiration in green plants, to the metabolism of unusual carbon and/or nitrogen sources in fungi (1). In the methylotrophic yeast Hansenula polymorpha peroxisomes are essential to support growth of cells on media containing methanol as the sole source of carbon and energy. Under these conditions many organelles that contain the key enzymes involved in methanol metabolism, alcohol oxidase (AO), 1 dihydroxyacetone synthase (DHAS), and catalase, develop in the cells. Conversely, when methanol-grown wild type (WT) cells are shifted to conditions in which the organelles are redundant for growth (e.g. glucose), they are rapidly and sequentially degraded by a process designated pexophagy (reviewed in Ref. 2). Morphological data suggest that in each cell generally a single (or few) small peroxisome(s) escape(s) the degradation process. The resistance of these organelles to degradation is thought to be of physiological advantage in that it allows the cells to quickly adapt to new environments that require new peroxisome functions.
In the present work, we report the identification of a novel H. polymorpha transcription factor, Mpp1p, which is involved in the regulation of peroxisomal proteins. In H. polymorpha mpp1 cells, various peroxisomal matrix enzymes involved in methanol metabolism and proteins essential for peroxisome biogenesis (peroxins) are present at reduced levels. As a result, mpp1 cells cannot grow on methanol as the sole source of carbon and energy. Interestingly, methanol-induced mpp1 cells generally contained a single enlarged peroxisome. Remarkably, these single organelles are protected from selective degradation upon exposure of cells to excess glucose.
Gene Tagging Mutagenesis and Isolation of Mut Ϫ Mutants-The RALF (random integration of linear DNA fragments) method (4) was used to generate yeast mutants. H. polymorpha HF246 was transformed with BamHI-linearized pREMI-Z plasmid (Table II) in the presence of 1 unit of BamHI restriction enzyme. Transformants were initially selected on YPD plates supplemented with zeocin and subsequently replica-plated to YND and YNM plates, respectively. Colonies unable to grow on YNM plates (methanol utilization-defective, Mut Ϫ colonies) were further analyzed. Two Mut Ϫ mutants, designated mpp1-1 and mpp1-2 (previously designated ARJ-59; see Ref. 4), were studied further.
Cloning of the MPP1 Gene-To identify the gene(s) disrupted by pREMI-Z in mutants mpp1-1 and mpp1-2, the chromosomal DNA of the cells was digested with either EcoRI or SphI, self-ligated, and transformed to E. coli, giving rise to the following plasmids: pANL7, pANL15, pREMI-59, and pANL22 (Table II). The genomic regions of pANL7 and pREMI-59 were initially sequenced using vector-based primers (4). Sequence analysis showed that the pREMI-Z vector had integrated in mutants mpp1-1 and mpp1-2 at two different locations in the same gene that was designated MPP1. Subsequently, the entire nucleotide sequence of the MPP1 gene was determined by primer walking on the rescued pREMI-Z plasmids. The nucleotide sequence of the MPP1 gene was deposited in GenBank TM (accession number AY190521).
To clone the MPP1 open reading frame, mutant mpp1-1 was transformed with a H. polymorpha genomic library constructed in the pYT3 vector (5). Leucine prototrophic transformants were screened on YNM plates for the ability to grow on methanol. From a complementing plasmid, a sub-clone containing a 3.4-kb PstI-NheI fragment with the entire MPP1 gene was obtained, designated pANL26, that was used for complementation studies.
Construction of an H. polymorpha MPP1 Null Mutant-A strain deleted for MPP1 was constructed by replacing the region of MPP1 comprising nucleotides ϩ1 to ϩ1042 by an auxotrophic marker. To this end, a deletion cassette was constructed as follows. First, two DNA fragments comprising the regions Ϫ816 to Ϫ1 and ϩ1042 to ϩ1841 of the MPP1 genomic region were obtained by PCR, using primers MPP1del-1 ϩ MPP1del-2 and MPP1del-3 ϩ MPP1del-4, respectively (see Table III). After restriction with NotI ϩ BglII and PstI ϩ Asp718I, respectively, the resulting fragments were inserted upstream and downstream of the H. polymorpha URA3 gene (6) in pBSK-URA3. From the resulting plasmid, designated pANL17, a 2650-bp BamHI-PvuI fragment was used to transform H. polymorpha NCYC495 leu1.1 ura3. Uracil prototrophic transformants were selected by their inability to grow on YNM plates. Proper deletion of MPP1 was confirmed by Southern blotting (data not shown). The resulting strain was designated mpp1.
To enable visualization of peroxisomes by fluorescence microscopy, the eGFP.SKL reporter gene was introduced in the resulting mpp1 strain. First, we constructed a H. polymorpha integrative plasmid containing the P AOX .eGFP.SKL cassette and the zeocin-resistance gene by inserting the 1.2-kb SphI-SalI fragment of pFEM34 (7) in pHIPZ4 (8). Subsequently, the resulting plasmid, designated pANL29, was linearized with SphI and integrated in the mpp1 genome. Zeocin-resistant transformants were analyzed for correct integration in the P AOX region by Southern blotting (data not shown). A strain containing a single copy of the integrated plasmid, designated mpp1.eGFP.SKL, was used for further studies.
Construction of a Strain Expressing an MPP1.eGFP Fusion Gene-To enable replacement of the genomic MPP1 gene by a MPP1.eGFP fusion gene, we first constructed plasmid pANL31. This plasmid is based on the pBluescript vector and contains an eGFP gene, lacking its start codon, and the zeocin resistance cassette. Subsequently, a 749bp Hin-dIII-BamHI PCR fragment containing the 3Ј end of the MPP1 gene lacking its stop codon (comprising the region ϩ1306 to ϩ2052 of MPP1), obtained using primers MPP1-Ori2 and MPP1w/oStop (Table III), was inserted in HindIII/BglII-digested pANL31. The resulting plasmid, designated pANL32, was linearized with EcoRI in the MPP1 region and transformed to WT H. polymorpha NCYC495 leu 1.1. Zeocin-resistant colonies were analyzed by Southern blotting to confirm correct integration in the MPP1 region (data not shown).
Miscellaneous DNA Techniques-Plasmids and primers used in this study are listed in Tables II and III, respectively. All DNA manipulations were carried out according to standard techniques (9). H. polymorpha cells were transformed by electroporation (10). DNA modifying enzymes were used as recommended by the supplier (Roche Applied Science). Pwo polymerase was used for preparative PCR. The ECL direct nucleic acid labeling and detection system (Amersham Biosciences) was used for Southern blot analysis. Oligonucleotides were synthesized by Invitrogen. DNA sequencing reactions were performed at BaseClear (Leiden, The Netherlands) using a LiCor automated DNA sequencer and dye primer chemistry (LiCor, Lincoln, NB). For DNA sequence analysis, the Clone Manager 5 program (Scientific and Educational Software, Durham, NC) was used. The BLASTP algorithm (11) was used to screen databases at the National Center for Biotechnology Information (Bethesda, MD). The ClustalX program was used to align protein sequences (12), the GeneDoc program was used to display the aligned protein sequences, and the ScanProsite program (13) was used to scan protein sequences for profiles and patterns of the PROSITE data base (14).
Biochemical Assays-Crude cell extracts were prepared as described (15). SDS-PAGE (16) and Western blot analysis (17) were performed by established methods. The degradation of peroxisomes in batch cultures of H. polymorpha was determined as described (18). Relative AO levels were determined by densitometric scanning of Western blots decorated with specific antibodies against AO. The decrease in AO levels during peroxisome degradation is expressed as a percentage of the initial value, which is arbitrarily set to 100%. ␤-Lactamase activities were assayed by established methods (19).
Morphological Analysis-Intact cells were prepared for electron microscopy and immunocytochemistry as described previously (19). Fluorescence microscopy studies were performed using a Zeiss Axioskop microscope (Carl Zeiss, Göttingen, Germany). Nuclear staining was performed as follows: 20 -40 OD 660 units of cells were harvested by centrifugation, resuspended in 1 ml of fresh medium supplemented with 40 l of a 2.7 mg/ml Hoechst 33258 stock solution, and incubated at 37°C for 45 min prior to analysis.

Mpp1p Is a Member of the Zn(II) 2 Cys 6 Cluster Protein Family and Is Induced during Growth of H. polymorpha on Methanol-
From a collection of 5,000 RALF transformants, mutants were selected that were impaired in growth on methanol as sole carbon source (Mut Ϫ phenotype). Two mutants, designated mpp1-1 and mpp1-2, were identified of identical morphological phenotype. Characteristically, upon growth in glycerol/methanol-containing media, conditions that lead to a strong peroxisome development in WT cells, generally a single peroxisome was observed in the vast majority of the mpp1-1 and mpp1-2 cells (Fig. 1). Sequencing of the plasmids recovered from these mutants revealed that the flanking regions were overlapping. Hence, both were apparently disturbed in the function of the same gene, termed MPP1 (methylotrophic peroxisomal protein regulator 1). Further sequencing of the plasmids recovered from mutants mpp1-1 and mpp1-2 allowed the determination of the remainder of the MPP1 open reading frame (ORF). Sequence analysis indicated that in case of mutant mpp1-1, the integration of the pREMI-Z plasmid had occurred 456 bp downstream from the initiation of the MPP1 ORF. In mutant mpp1-2, pREMI-Z had inserted at the promoter region of the same ORF (nucleotide Ϫ195). The sequence upstream the 5Ј region of this ORF comprised the 3Ј end of the H. polymorpha dihydroxyacetone synthase gene (DAS; see Fig. 2A).
The MPP1 gene encodes a protein of 684 amino acids. A BLASTP search revealed that the N-terminal region of Mpp1p is similar to that of many DNA-binding proteins. However, a true homologue was not found in the available databases. Further analysis using the ScanPROSITE program revealed that Mpp1p is a putative member of the Zn(II) 2 Cys 6 family of transcription factors. These transcription regulators are exclusively detected in fungi and contain a well conserved DNA binding domain (20). This domain consists of a cystein-rich motif (CysX 2 CysX 6 CysX 6 CysX 2 CysX 6 Cys) that complexes two Zn 2ϩ ions and in most cases recognizes a pair of 5Ј-CGG-3Ј triplets in the promoters of target genes (20,21). A primary sequence comparison of the region comprising the DNA binding domain of characteristic Zn(II) 2 Cys 6 proteins of Saccharomyces cerevisiae and the putative DNA binding domain of Mpp1p is depicted in Fig. 2B. The subcellular location of Mpp1p in H. polymorpha was studied in a strain in which the endogenous MPP1 gene was replaced by a MPP1.eGFP fusion gene. The presence of the C-terminal tag most likely did not interfere with Mpp1p function, because the resulting strain, MPP1.eGFP, grew normally on glucose and methanol (data not shown). In MPP1.eGFP cells grown on glucose, GFP fluorescence was invariably undetectable, independent of the growth stage. However, upon a shift of glucose-grown cell to fresh methanol-containing media, GFP fluorescence was readily observed as a single dot at all growth phases. This spot was observed in the same region of the cells as Hoechst 33258, which is a nuclear stain. These findings therefore indicate that Mpp1p is associated with the nucleus (Fig. 3).
Cells of the MPP1 Deletion Strain Generally Contain a Single Peroxisome-For construction of a MPP1 deletion strain, we replaced a 1042-bp fragment from the MPP1 open reading frame by the H. polymorpha URA3 gene ( Fig. 2A). Cells of the mpp1 strain and the original RALF mutants, mpp1-1 and mpp1-2, showed identical phenotypes. Mpp1 cells were unable to grow on methanol as sole carbon source and, when grown on glycerol/methanol mixtures, characteristically contained a single peroxisome (Fig. 4B). Reintroduction of the MPP1 gene from an autonomously replicating plasmid (pANL26) restored normal growth of mpp1 cells on methanol, as well as normal peroxisome proliferation (Fig. 4C).
In crude extracts of glycerol/methanol-grown WT and mpp1 cells, the levels of the peroxisomal matrix protein AO were strongly reduced, whereas DHAS, another peroxisomal matrix enzyme, could not be detected (Fig. 5A). However, catalase was present at approximately WT levels. Also the levels of malate synthase, a peroxisomal key enzyme of glyoxylate (C 2 ) metabolism (22), did not change significantly under the conditions tested (Fig. 5A). In addition, the level of various peroxins that play a role in peroxisome formation was analyzed. All were present at significantly reduced levels (Pex3p, Pex5p, Pex10p) with the exception of Pex14p (Fig. 5B). Cytosolic alcohol dehydrogenase levels were normal.
To test whether the observed lower protein levels were because of decreased expression or increased protein degradation, we analyzed the promoter activities of the AOX (P AOX ) and PEX3 (P PEX3 ) genes in WT and mpp1 cells using ␤-lactamase as a reporter (19,23). As shown in Fig. 5C, the activities of both promoters are significantly reduced in mpp1 cells compared with WT controls.
Despite the low peroxisome numbers we did not observe a defect in matrix protein import in mpp1 cells. GFP.SKL was solely observed in spots (compare Fig. 4B), and immunocytochemistry revealed that anti-alcohol oxidase and anti-catalasedependent-specific labeling was confined to the peroxisomal profiles (Fig. 4D), indicating that these proteins were incorporated in their correct target organelle. As expected, using antidihydroxyacetone synthase antibodies, specific labeling was not detected (not shown).
Single Peroxisomes in H. polymorpha Cells Are Not Susceptible to Glucose-induced Selective Peroxisome Degradation (Pexophagy)-In methanol-grown WT H. polymorpha cells peroxisomes are rapidly degraded when they have become redun-  dant for growth (24). To analyze the fate of the single organelles in mpp1 cells, glycerol/methanol-grown cells of this strain were exposed to excess glucose conditions. In the first 2 h after the shift of cells, no significant AO protein degradation had occurred in mpp1 cells, as judged from Western blots (Fig.  6, A and B). In mpp1.eGFP.SKL cells, fluorescence microscopy studies failed to demonstrate the uptake of the fluorescent reporter protein in vacuoles, a phenomenon that was readily observed in WT controls (Fig. 6C). Also, electron microscopically we were never able to detect peroxisome sequestration, the typical initial event of pexophagy in H. polymorpha (not shown).
To further substantiate the possibility that also in WT cells specific organelles escape the pexophagy process, we have grown H. polymorpha WT cells on methanol to a phase that they generally contain only one or two organelles per cell (early exponential growth; see Ref. 25) and studied pexophagy relative to cells that were in the mid-exponential growth phase and contained several peroxisomes. The results clearly show that the single peroxisome present in early exponential WT cells is not degraded within a period of 4 h after exposure of the cells to excess glucose (Fig. 7C). In contrast, in mid-exponential cells a rapid reduction of peroxisome numbers was observed. Nevertheless, in these cells generally a single fluorescent spot  1A) and mpp1-2 (Fig. 1B) cells that express P AOX .eGFP.SKL. Cells were grown for 15 h in glycerol/methanol-containing media. In WT cells clusters of fluorescent spots are evident whereas in mpp1-2 cells only a single spot can be observed.
remained indicating that not all peroxisomes were degraded (Fig. 7C). Remarkably, the morphological phenotypes of early and mid-exponential WT cells after 4 h of exposure of cells to glucose was indistinguishable in that they generally contained a single fluorescent spot. Most likely, solely small organelles remain unaffected as was evident after careful electron microscopical observations (Fig. 4E).
These observations were confirmed biochemically by Western blot analysis of samples taken from the same cultures using specific antibodies against AO (Fig. 7, A and B). When WT cells in the early exponential growth phase were subjected to pexophagy conditions, AO protein levels only slightly decreased during the first hour after the shift. However, at later time points the AO levels remained constant, indicating that the peroxisomes in these cells were not degraded. The initial decrease most likely is because of the presence of a minor fraction of the cells that contains more than one peroxisome at the time of the shift. Mid-exponentially grown cells showed the continuous decrease in AO protein characteristic for WT cells subjected to pexophagy (18). DISCUSSION We have cloned and characterized a novel member of the family of zinc cluster proteins of the yeast H. polymorpha, designated Mpp1p. Mpp1p plays a role in the control of peroxisomal protein synthesis (see below) and represents the third H. polymorpha zinc cluster protein reported so far. The other two proteins, Yna1p and Yna2p, activate genes involved in nitrate assimilation (26,27). Zinc cluster proteins are a class of transcription factors involved in the modulation of various cell activities in fungi, among which are transcriptional control of genes required for growth on various carbon sources, synthesis of specific metabolites (e.g. amino acids, pyrimidines), and the expression of ABC transporters that mediate multi-drug resistance (reviewed in Refs. 28 and 29). A well known example of these proteins is Gal4p (Fig. 2B), responsible for the activation of genes of galactose metabolism. Gal4p function involves binding of the N-terminal zinc cluster to CGG triplets present in the promoter of target genes and the formation of a dimer through one or more very short coiled-coils present C-terminal from the zinc cluster motif (20). Most members of this family bind to H. polymorpha mpp1 cells exhibit a WT peroxisome phenotype after introduction of a plasmid containing the entire MPP1 gene. Fluorescence pictures are shown of cells that express P AOX .eGFP.SKL in WT (Fig. 4A) in mpp1 cells (Fig. 4B) and cells of the complemented mpp1 strain (mpp1.pANL26; see C). D and E, electronmicroscopic observations. Immunocytochemistry on ultrathin sections of glycerol/methanol-grown mpp1 cells, showing the selective location of gold particles on a peroxisomal profile after incubation with anti-AO antibodies (Fig. 4D). A similar location was observed when anti-catalase antibodies were used (inset; Fig. 4D). E shows a section through a methanolgrown H. polymorpha WT cells, subjected to excess glucose for 2 h. Note that in both the developing bud and the mother cell a relatively small organelle is not degraded. A, degrading peroxisome; M, mitochondrion; N, nucleus; P, peroxisome; V, vacuole. The marker represents 0.5 m.
DNA as a homodimer, although formation of heterodimers or activation by only one monomer has also been reported (30 -32).
We can now add methylotrophic growth to the list of Zn(II) 2 Cys 6 transcription factor-modulating activities. Our data clearly demonstrate that Mpp1p is necessary to sustain growth of H. polymorpha cells on media containing methanol as the sole carbon source. Interestingly, this is yet the only growth condition where the role of Mpp1p is essential as mpp1 cells grow normally on several other carbon and nitrogen sources (e.g. glucose, ethanol, glycerol, dihydroxyacetone, and ethylamine; data not shown), some of which require peroxisome functions. These alternative functions are presumably not regulated by MPP1 but require different transcription factors to induce the required enzyme repertoire.
The Mut Ϫ phenotype of mpp1 cells is readily explained by the observation that the levels of the two major enzymes of methanol metabolism, AO and DHAS, are strongly reduced (AO) or absent (DHAS). Also, in mpp1 cells peroxisomes do not proliferate in response to the presence of methanol. This inhibitory effect is most probably not related to the reduced amounts of AO and DHAS protein. In fact, several earlier observations argue against a direct relation between matrix protein levels and peroxisome numbers in H. polymorpha. For instance, H. polymorpha pim mutants that are affected in matrix protein import contain enhanced numbers of small-sized peroxisomes relative to wild type organelles (33). On the other hand, overproduction of AO protein in glucose-grown cells resulted in enlarged organelles but not an increase in organelle numbers (34).
It has been proposed that peroxisomes may divide (multiply) through a constitutive or a regulated mechanism (35,36). In this view the constitutive mechanism is responsible for the partitioning of peroxisomes between mother and bud in noninduced (glucose-grown) cells, a phenomenon that may be re- lated to the cell cycle. The regulated division occurs in induced cells, and this mechanism is responsible for the formation and partitioning of the characteristic peroxisome cluster during growth of cells on methanol. In line with this hypothesis we speculate that the absence of Mpp1p prevents the regulated peroxisome division in H. polymorpha during C 1 metabolism but not the constitutive mechanism as developing buds are generally administrated with a single organelle. In H. polymorpha mpp1 cells the effect on peroxisome proliferation is therefore likely to be either because of the down-regulation of a yet unknown initial (signaling) compound of the proliferation machinery or directly related to the strongly reduced level(s) of protein(s) (e.g. Pex3p) involved in biogenesis and/or fission. It has been shown before (37) that in H. polymorpha a direct relation exists between the levels of the peroxin Pex3p and the number of peroxisomes. Possible additional candidates besides Pex3p are H. polymorpha homologues of S. cerevisiae Vps1p and Pex11p (35,38,39). In bakers' yeast, oleate induced-peroxisome proliferation was shown to be dependent on Oaf1p and Pip2p, both Zn(II) 2 Cys 6 transcription factors (29,31).Extensive studies showed that Oaf1p is constitutively synthesized and is the receptor for the oleate signal, whereas in the absence of the inducer, Pip2p is produced at low levels, and its activity is inhibited by Oaf1p. The current model (40) proposes that upon oleate induction, Oaf1p becomes active thereby abolishing its inhibitory effect on Pip2p, which, in turn, becomes able to up-regulate is own synthesis. As a consequence, maximal amounts of the heterodimer Oaf1p-Pip2p are produced that efficiently bind to the oleate response element, thereby activating transcription of genes involved in oleate metabolism. Mpp1p is not the homologue of S. cerevisiae Pip2p or Oaf1p. First, H. polymorpha Mpp1p only shares similarity with these proteins in the Zn(II) 2 Cys 6 region. Moreover, a search in the H. polymorpha genome data base revealed that H. polymorpha does contain genes that encode proteins with significant overall similarity to S. cerevisiae Pip2p and Oaf1p. 2 It is possible that a mechanism similar to the S. cerevisiae Oaf1p-Pip2p activation complex functions during methylotrophic growth of H. polymorpha and that Mpp1p behaves analogous to Pip2p, because the protein was only detected in cells grown on methanol. The identification of a putative Mpp1p partner and the definition of a DNA-response element are issues of current investigations.
We observed that peroxisomal matrix proteins were not mislocalized in mpp1 cells; both AO and catalase, as well as GFP.SKL, were normally imported into the single organelles of mpp1 cells. Apparently, the strongly reduced levels of specific peroxins (Pex3p, Pex5p) in methanol-induced mpp1 cells still suffice to allow quantitative matrix protein import into peroxisomes. Hence, the efficient matrix protein import in mpp1 cells may be related to the normal levels of Pex14p that showed no decrease in methanol-induced cultures.
We were particularly interested in the fate of the single 2 G. Gellissen, personal communication. In mid-exponential cells AO protein gradually decreases during the first 4 h after addition of glucose. In early exponential cells AO levels initially slightly peroxisome of mpp1 cells upon the induction of pexophagy. We observed that these organelles were protected from degradation after a shift to conditions that led to a rapid degradation of peroxisomes in identically grown WT cells. Subsequent studies demonstrated that also the single organelles present in shortly induced WT cells were resistant to pexophagy. Possibly, these single organelles are protected from degradation thus enabling the cell to rapidly respond to changes in the environment that require new peroxisome (2). Previously, we demonstrated that in H. polymorpha peroxisomes are only temporarily competent to import matrix proteins by a yet unknown mechanism (41). An attractive possibility is that exclusively the peroxisomes that no longer import matrix proteins are degraded during pexophagy, whereas the few organelles that can import matrix proteins are not susceptible to degradation. This reasoning would imply that proteins involved in the biogenesis of peroxisomes may significantly contribute to pexophagy. Indeed, we demonstrated recently (42,43) that two membrane-associated peroxins, Pex3p and Pex14p, are also involved in pexophagy in H. polymorpha. Undoubtedly, the understanding on how the cell distinguishes between peroxisomes that should be destroyed or preserved promises to be an exciting part of the study of the peroxisome life cycle.