Research Article 1759 PEX14 is required for microtubule-based motility in human cells

Pratima Bharti1,*, Wolfgang Schliebs1,*, Tanja Schievelbusch1, Alexander Neuhaus1, Christine David1, Klaus Kock2, Christian Herrmann2, Helmut E. Meyer3, Sebastian Wiese3,4, Bettina Warscheid3,4, Carsten Theiss5 and Ralf Erdmann1,‡ 1Institute for Physiological Chemistry, Department of Systems Biology, Faculty of Medicine, Ruhr University of Bochum, 44780 Bochum, Germany 2Department of Physical Chemistry 1, Faculty of Chemistry and Biochemistry, Ruhr University of Bochum, 44780 Bochum, Germany 3Medizinisches Proteom-Center, Faculty of Medicine, Ruhr University of Bochum, 44780 Bochum, Germany 4Faculty of Biology and BIOSS Centre for Biological Signalling Systems, University of Freiburg, 79104 Freiburg, Germany 5Institute of Anatomy and Molecular Embryology, Faculty of Medicine, Ruhr University of Bochum, 44780 Bochum, Germany *These authors contributed equally to this paper ‡Author for correspondence ([email protected])

Accepted 14 January 2011 Journal of Cell Science 124, 1759-1768 © 2011. Published by The Company of Biologists Ltd doi:10.1242/jcs.079368

Summary We have established a procedure for isolating native peroxisomal membrane complexes from cultured human cells. Protein- A-tagged 14 (PEX14), a central component of the peroxisomal protein translocation machinery was genomically expressed in Flp-In-293 cells and purified from digitonin-solubilized membranes. Size-exclusion chromatography revealed the existence of distinct multimeric PEX14 assemblies at the peroxisomal membrane. Using mass spectrometric analysis, almost all known human involved in protein import were identified as constituents of the PEX14 complexes. Unexpectedly, tubulin was discovered to be the major PEX14-associated protein, and direct binding of the was demonstrated. Accordingly, peroxisomal remnants in PEX14- deficient cells have lost their ability to move along microtubules. In vivo and in vitro analyses indicate that the physical binding to tubulin is mediated by the conserved N-terminal domain of PEX14. Thus, human PEX14 is a multi-tasking protein that not only facilitates peroxisomal protein import but is also required for peroxisome motility by serving as membrane anchor for microtubules.

Key words: Peroxisome, Microtubule, PEX14, Protein complex isolation

Introduction translocation takes place is still not known; however, the existence

Journal of Cell Science are ubiquitous, single membrane-bound cell organelles of a transient peroxisomal pore with properties expected for a protein- with a large variety of metabolic functions of which some are conducting channel has recently been identified (Meinecke et al., essential for humans. This is indicated by several inborn diseases 2010). Notably, membrane-embedded PEX5 and PEX14 are major caused by defects in peroxisome function. The most severe constituents of the translocation pore. After dissociation of receptor– peroxisomal diseases known as the ‘spectrum of Zellweger cargo complex and release of the cargo protein into the peroxisomal syndrome’ are characterized by a defect in the biogenesis of lumen, the receptors are recycled for another round of import, which peroxisomes and are lethal within the first months of life. So far, 12 requires release of the receptors from the membrane (Lanyon-Hogg complementation groups of peroxisome biogenesis disorders (PBDs) et al., 2010). The export of receptors is a complex process with have been identified and studied in detail (Steinberg et al., 2006). ubiquitylation and ATP-dependent dislocation by “ATPases Notably, with the exception of a coding for a protein of the associated with diverse cellular activities” (AAA – peroxins) as key peroxisomal fission machinery (Waterham et al., 2007), all steps (Fujiki et al., 2008). affected in these patients encode proteins required for the import of Other peroxins, including human PEX11 and Dlp1, play a role matrix and membrane proteins into peroxisomes. Proteins that are in peroxisome fission, proliferation and inheritance (Thoms and required for peroxisome biogenesis are collectively called peroxins Erdmann, 2005). In fact, peroxisomes are remarkably dynamic in (PEX), and three of them, PEX3, PEX16 and PEX19, are necessary terms of their intracellular movements, which are mediated by for the targeting and insertion of peroxisomal membrane proteins either microtubules or actin filaments, depending on the species. (Ma and Subramani, 2009). Seven peroxins that are defective in Mammalian peroxisomes are moved along microtubule tracks by Zellweger patients belong to the peroxisomal import machinery for dynein–dynactin and kinesin motors (Kural et al., 2005; Schrader matrix proteins. The matrix proteins contain either a peroxisomal et al., 2000). targeting signal of type 1 (PTS1) at the extreme C-terminus or a Most peroxins were identified by functional complementation of PTS2 close to the N-terminus, which are specifically recognized by yeast or Chinese hamster ovary (CHO) cell mutants affected in the import receptors PEX5 and PEX7, respectively. The receptors peroxisome biogenesis (Erdmann et al., 1989; Tsukamoto et al., bind their cargo in the cytosol and the cargo-loaded PTS1-receptor 1990) and the human orthologs were then identified by sequence alone or in conjunction with the PEX7–cargo complex targets to the similarity. However, current data indicate that the list of peroxins peroxisomal membrane where it associates with the docking complex identified is far from being complete. This is due to the fact that: of the peroxisomal protein import machinery. PEX14 and PEX13 (1) some peroxins are present in functionally overlapping forms, are known constituents of the human docking complex. How cargo like the ubiquitin-conjugating enzymes of the peroxisomal protein 1760 Journal of Cell Science 124 (10)

import machinery (Grou et al., 2008) and (2) other peroxins have Purification, identification and subcellular distribution of different cellular functions in addition to their peroxisome-specific PEX14-associated proteins roles and thus, inactivation of the corresponding genes lead to For isolation of human PEX14-associated proteins, we established more generalized cellular defects and severe human diseases that a protocol that in principle was based on peroxin complex isolation are not immediately recognized as PBDs. These processes include described for yeast (Agne et al., 2003). A crucial step in the peroxisome fission, regulated degradation, inheritance and motility, isolation of membrane protein complexes is the choice of detergent which require proteins that are also necessary for the maintenance for solubilization. It was reported that a human peroxin complex of other organelles (Fagarasanu et al., 2009; Farre et al., 2009; including PEX14, PEX5 and the RING finger peroxins remains Waterham et al., 2007). Thus, because of the functional redundancy stable upon extraction with 1% (w/v) digitonin solution (Reguenga of peroxins and the sharing of peroxins by different organelles, et al., 2001). Accordingly, for purification of PEX14 complexes, other strategies have to be employed to identify such peroxins with membranes were prepared from 25g Flp-In-293[PEX14–TEV–PA] overlapping function. To fill this gap, proteomic approaches have cells (wet weight) and 1% (w/v) digitonin was applied for PEX14 been applied; however, so far the complete analyses of whole complex extraction. Solubilized PEX14 complexes were affinity- peroxisomal proteomes did not result in the identification of new purified with human IgG-Sepharose and eluted from the column biogenesis factors (Saleem et al., 2006; Wiese et al., 2007). This by TEV protease cleavage. Isolated proteins were further separated might be explained by the low abundance of such proteins or the by gradient gel electrophoresis and stained by colloidal Coomassie fact that they might be mistaken as contaminants. Blue (supplementary material Fig. S1). Non-transfected Flp-In- Here, we identified tubulin as novel binding partner for PEX14, 293 cells were treated in the same way and served as a negative the key protein of the protein import machinery, and assign a new control. The isolated and stained PEX14-associated proteins were function for the protein in peroxisome motility. We established a subjected to in-gel tryptic digestion and peptides were identified new procedure for isolation of subcomplexes from the human by nano-HPLC on-line coupled to electrospray ionization tandem peroxisomal membrane for mass spectrometric analyses. For this mass spectrometry (ESI-MS/MS). purpose, we generated stable human cell lines expressing In three independent analyses of PEX14 complexes, 22 known genomically tagged peroxins that are functional and suitable for peroxisomal proteins were identified (supplementary material Table affinity purification of protein complexes of the peroxisomal protein S1). Among these, were the known importomer constituents PEX5, import machinery, a strategy which has been used successfully for ubiquitin, PEX13, and the RING-finger peroxins PEX10, PEX12 yeast (Agne et al., 2003). Analysis of PEX14-associated proteins and the AAA-peroxin PEX1. In addition, peroxins involved in revealed tubulin as a major interacting partner of PEX14. Both in membrane protein targeting as PEX19, PEX3 and PEX16 could be vivo and in vitro experiments strongly suggest that PEX14, in identified in association with PEX14. Thus, almost all known addition to its well-established role in peroxisomal protein import, binding partners of human PEX14 were detected, thereby serves as the peroxisomal membrane anchor for microtubules and, demonstrating the efficiency of the purification procedure. thereby, enables peroxisome motility in human cells. In order to obtain information on the relative abundance of proteins identified in both PEX14-containing complexes and the Results respective mock preparations (control), we calculated their

Journal of Cell Science Establishment of a human cell line expressing Protein-A- individual spectral counts according to Liu et al. (Liu et al., 2004). tagged PEX14 Fig. 2A shows the relative abundance of proteins associated with In order to stably express Protein-A-tagged PEX14 in human cells, peroxisomes, microtubules and the cytoplasm as well as of proteins the fusion gene PEX14-TEV-Protein A was cloned into the modified with other subcellular locations in the two PEX14 complexes and expression vector pcDNA5/FRT (Invitrogen). To ensure that the the respective mock preparations. In comparison with the controls, tag comprising a tobacco etch virus protease cleavage site (TEV) proteins associated with peroxisomes and microtubules were and Protein A does not affect the functionality of PEX14 during significantly enriched in PEX14-containing complexes. Complete matrix protein import, the expression plasmid together with another lists of peroxisomal and microtubular proteins identified in PEX14 plasmid encoding either a PTS1- or a PTS2-marker protein were complexes, including information on accession numbers, gene co-transfected into a Zellweger patient fibroblast cell line deficient names and spectral counts are provided in supplementary material in PEX14. Functional complementation of the matrix protein import Tables S1 and S2. Interestingly, microtubular proteins were up to defect was visualized by fluorescence microscopy (Fig. 1A,B). 27.5-fold enriched (preparation #2) and accounted for the most Expression of the plasmid-encoded tagged PEX14 did result in abundant class of proteins associated with PEX14. The microtubular the restoration of peroxisomal protein import shown by the fraction comprised various isoforms of tubulin (predominantly - reappearance of a punctate fluorescence pattern for the peroxisomal tubulin), motor and adaptor proteins like dynein, kinesin and matrix markers (EGFP–PTS1 and PTS2–YFP) which co-localized dynactin (DynC1H1), and subunits of the T-complex (CCT with the fluorescence pattern for peroxisomal membrane marker complex) proposed to act as a chaperone for tubulin polymerization PMP70 as well as PEX14. The data show that the C-terminal (Brackley and Grantham, 2009) (supplementary material Table tagging of PEX14 did not interfere with the function of the protein S2). in peroxisomal protein import. In order to investigate whether the high enrichment of tubulin After genomic integration of the fusion gene into Flp-In-293 and tubulin-associated proteins indicates a functionally relevant cells, clones were selected by hygromycin resistance, Zeocine- association, PEX14-containing complexes were purified in the sensitivity, lack of -galactosidase activity and protein expression presence of 100 nM nocodazole, a potent microtubule- (data not shown). Neither the C-terminal tagging nor the ectopic depolymerizing agent (Vasquez et al., 1997). Relative quantitative expression affected the peroxisomal localization of PEX14–TEV– mass spectrometric analysis revealed that under these conditions Protein A, as indicated by the observed co-localization of PEX14 -tubulin (TUBBx and FLJ52712), dynein (DYNC1H1), dynactin and the peroxisomal marker EGFP–PTS1 (Fig. 1C). (DCTN1) and two subunits of the T-complex (CCT5 and CCT6A) PEX14-dependent peroxisome motility 1761

were the only proteins with a microtubular localization that co- PEX14 with the microtubules is mediated by direct binding to - purified with PEX14 (Fig. 2B). The low abundance or absence of tubulin. In line with this assumption, T-complex proteins as -tubulin and MAPs in this preparation confirms the dissociation chaperones for tubular frameworks were abundant binding partners of the microtubular assemblies and suggests that association of of PEX14. Journal of Cell Science

Fig. 1. PEX14–TEV–Protein A localizes to peroxisomes and is functional. (A,B)To test functionality of PEX14–TEV–Protein A during peroxisomal matrix protein import, plasmid-encoded EGFP–PTS1 (A) or PTS2–EYFP (B) was transiently expressed in PEX14-deficient (PEX14T) and normal fibroblast (GM5756T) cells. Normal fibroblast cells exhibit a punctate fluorescence pattern indicating peroxisomal targeting of both PTS1- and PTS2- proteins, whereas all peroxisomal matrix proteins are mislocalized in the cytosol in PEX14-deficient cells. Functional complementation of the import defect was achieved by expression of plasmid encoded PEX14–TEV–Protein A as indicated by a punctate fluorescence pattern, which is congruent with the immunofluorescence pattern obtained for the peroxisomal membrane proteins PMP70 and PEX14. (C)Validation of the peroxisomal localization of the Protein-A-tagged genomic copy of PEX14. Cell lines Flp-In-293 cells and Flp-In-293 [PEX14–TEV– Protein A] were transiently transfected with a plasmid encoding the peroxisomal marker EGFP–PTS1. Specific immunodetection of the fusion protein PEX14–TEV–Protein A was carried out with anti-Protein A antibodies, the endogenous and the fusion protein were detected by using antibodies against PEX14. Scale bars: 10m. 1762 Journal of Cell Science 124 (10)

Fig. 2. PEX14 associates with microtubules. (A) PEX14 complexes (#1 and #2) and the respective mock preparations (control #1 and #2) were analyzed by tandem mass spectrometry, and spectral counts were calculated in order to obtain information on the relative abundance of proteins. Pie charts display the summed percentage abundance of proteins that are located at peroxisomes, at microtubules or in the cytoplasm, as well as proteins with other subcellular locations.

Journal of Cell Science Subcellular protein localizations are generally based on information provided by the SwissProt database. (B)List of microtubule-associated proteins and their relative abundance in PEX14-containing complexes that were isolated in the absence or presence of nocodazole as microtubule-depolymerizing agent. –, not detected; +, detected with spectral counts (SC) of 1–100; ++, 101–200; +++, 201–300; ++++, 301–400; +++++, 401–500. All identified peroxisomal and microtubules-associated proteins are listed in supplementary material Tables S1 and S2.

The concurrent presence of components of the peroxisomal normal fibroblast cells and PEX14-deficient Zellweger patient import machinery and the cytoskeleton raises the issue of whether cell lines. These mutant cell lines lack normal peroxisomes due the components are constituents of the same complex. To this end, to the import defect but still contain peroxisomal membrane the isolated PEX14-containing protein complexes were also ghosts. For in vivo labeling of peroxisomes, fibroblasts were analyzed by size-exclusion chromatography. PEX14 covered a transfected with an expression plasmid encoding the peroxisomal broad range of molecular assemblies, as indicated by its presence membrane protein EGFP–PEX26. In agreement with previous in fractions ranging in molecular size from larger than 1000 kDa literature (Rapp et al., 1996; Wiemer et al., 1997), two kinds of to smaller than 150 kDa (Fig. 3). One of the higher molecular peroxisomal movements were observed in normal fibroblasts weight subcomplexes might represent the so-called importomer (Fig. 4 and supplementary material Movie 1). Most peroxisomes that enables matrix protein import, as indicated by co-elution with (87–92%) oscillated in place, bearing a slow, random, vibrational PEX5. Another PEX14 subcomplex with a molecular size of about motion with a speed of 0.015±0.002 m/second whereas a 800 kDa was associated with PEX19, suggesting that this complex significant population of peroxisomes (8–13%) displayed fast represents PEX14 in transit. Additionally, PEX14 formed high directional, long range saltatory movement in the range of molecular weight complexes with -tubulin that only contained 0.1±0.02 m/second (Fig. 4B). The long-range movement of minor amounts of PEX5 and no PEX19. peroxisomes was abolished when transfected cells were preincubated with 10 g/ml colchicine, which inhibits microtubule PEX14 is required for the movement of peroxisomes along polymerization by binding to tubulin (data not shown). microtubules Accordingly, the long-range movement was not inhibited by 10 To investigate the role of PEX14 as a possible peroxisomal g/ml Cytochalasin D, which is a potent inhibitor of actin anchor for the microtubular network, the motile behavior of assembly (data not shown). These results confirm recent work peroxisomal membranes was studied using time-lapse imaging in indicating that peroxisomes in mammalian cells can either PEX14-dependent peroxisome motility 1763

Fig. 3. PEX14 is constituent of various subcomplexes at the peroxisomal membrane. Separation of PEX14-associated protein complexes by size- exclusion chromatography. PEX14 complexes were prepared by affinity chromatography as described in Materials and Methods and further separated by gel filtration on a calibrated Superose 6 column. Fractions of 80l were collected and subjected to SDS-PAGE followed by immunoblot analysis using antibodies raised against the indicated proteins. Void volume (V0) and molecular masses of marker proteins (kDa) are indicated.

associate to actin filaments or to the microtubular cytoskeleton to move for short or long distances, respectively (Schollenberger et al., 2010). Remarkably, PEX14-deficient cell lines were drastically affected in the long-range directional movement of peroxisomal structures (Fig. 4 and supplementary material Movie 2). The GFP-labeled remnant peroxisomes still showed some short-range motions, which could be completely blocked with Cytochalasin D demonstrating that the residual movement of peroxisomal structures in PEX14-deficient cells depends on the actin filaments (data not shown). To test whether the defect in microtubule- dependent motility is a feature of peroxisomal ghosts per se, as

Journal of Cell Science suggested previously (Nguyen et al., 2006), or whether it is specifically caused by the absence of PEX14, we conducted time- lapse imaging in mutant cell lines deficient in PEX5 or PEX1, other components of the peroxisomal import machinery. In both mutant cells, the peroxisomal ghosts displayed a vibrational and directional motion of the peroxisomal membranes similar to that Fig. 4. Peroxisomal remnants of PEX14-deficient cells are defective in in normal fibroblasts (Fig. 4 and supplementary material Movies long-range directional movement. Normal human fibroblast cells 3, 4). Statistical analysis revealed that PEX5- as well as PEX1- (GM5657T) and PBD cell lines deficient in PEX14 (PEX14T), PEX5 ( PEX5T) and PEX1 ( PEX1T) were transiently transfected with EGFP– deficient fibroblasts displayed wild-type-like motility with 8–   PEX26 expression plasmid. At 48 hours after transfection, live cell imaging 13% of the peroxisomes exhibiting a fast, directional movement was performed at 37°C. (A)Projections were made every 5 seconds for up to 5 and 87–92% of peroxisomal population exhibiting slow and minutes. For movies, see supplementary material Movies 1–4. In this part of microtubular-independent oscillatory movement with little or no the figure, single captures (0, 20, 40 seconds) representative for each cell line net displacement. By contrast, PEX14-deficient cell lines showed are shown. The arrowheads indicate peroxisomes moving quickly during no long-range saltatory movement of peroxisomes, but 100% of image capture and the arrows marks the starting position. Normal fibroblasts the peroxisomal population only exhibited slow vibrational and PEX1- and PEX5-deficient cells showed normal vibrational and movement (Fig. 4B). directional movement of peroxisomes, whereas PEX14-deficient cell lines did not show any directional movement. Scale bar: 10m. (B)For statistical Restoration of the peroxisomal motility of PEX14-deficient analysis of directional (dark gray) and vibrational (light gray) velocities for each cell line, ten peroxisomes of distinct cells were tracked. The values cells by complementation indicate means ± s.d. The peroxisome motility defect in PEX14-deficient cells could be rescued by expression of plasmid-encoded full-length PEX14 (Fig. 5 and supplementary material Movie 5). The plasmid was co- transfected together with the peroxisomal membrane marker EGFP– phenotype was indeed caused by the deficiency in PEX14 and not PEX26. The peroxisome motility mutant as well as complemented by a second site mutation. PEX14-deficient cell lines were assessed 48 hours after transfection To identify the minimal regions of PEX14 responsible for by live cell imaging. The observed restoration of peroxisome functional complementation, various truncated forms of PEX14 motility by the expression plasmid confirmed that the mutant were tested for their ability to complement the defect in peroxisome 1764 Journal of Cell Science 124 (10)

could be restored by expression of the fragment PEX14(1–140) (Fig. 5 and supplementary material Movie 7). On the other hand, an N-terminal truncated form, PEX14(79–376) containing transmembrane domain and coiled-coil region, could not restore the peroxisomal directional movement (Fig. 5 and supplementary material Movie 8). Taken together, the data show that the N- terminal region is essential for the PEX14-dependent movement of peroxisomes and that the minimal fragment to perform that function comprises the conserved N-terminal domain together with a peroxisomal transmembrane region.

Physical interaction between the N-terminal conserved domain of PEX14 and microtubules Tubulin was purified from porcine brain (supplementary material Fig. S2) and incubated with glutathione-Sepharose beads coated with GST–PEX14(1–78) or GST alone. Most of tubulin fraction (more than 60%) was retained by GST–PEX14(1–78) whereas only minor amounts (less than 5%) were bound by GST alone (Fig. 6A). The result of this in vitro binding assay demonstrates that the N-terminal domain of PEX14 directly interacts with tubulin. Remarkably, PEX5 as a known binding partner of PEX14(1–78) can compete with tubulin for binding to PEX14. For the competition analysis, a synthetic peptide representing the first N-terminal Wxxx[F/Y] motif (WAQEF) of the human PEX5 sequence was used. It is well established that the di-aromatic pentapeptide motif docks with high affinity to a groove at the surface of PEX14(1–78) (Neufeld et al., 2009). As shown in Fig. 6B (left panel), concomitant incubation of the GST–PEX14(1– 78)-coated-beads with tubulin and an excess of a single Wxxx[F/Y] peptide (PEX5–W1 peptide) decreased the amount of bound tubulin to the level of the negative control. Moreover, incubation of a preformed GST–PEX14(1–78)–tubulin complex with PEX5–W1 peptide led to dissociation of the complex (Fig. 6B, right panel). These data demonstrate that PEX14(1–78) binds

Journal of Cell Science tubulin and PEX5 directly and that both proteins compete for PEX14 binding.

Discussion Fig. 5. Microtubular movement of peroxisomes can be restored in PEX14- Here we present a novel function for the human peroxin PEX14. deficient cells by expression of PEX14 and fragments thereof. Besides its well-established role as a central constituent of (A)Representation of the full-length PEX14(1–377), and its various truncated translocation machinery for matrix proteins, we discovered that forms PEX14(1–260), PEX14(1–140) and PEX14(79–377). The conserved N- PEX14 also serves as a membrane anchor for microtubules and is terminal domain and transmembrane (TM) region are shown in green. required for peroxisome motility in human cells. The existence of (B)Human PEX14-deficient fibroblast cells were transiently co-transfected a peroxisomal protein that might anchor peroxisomes to with expression plasmids encoding full-length PEX14(1–377) or one of the microtubules had been proposed earlier (Schrader et al., 1996) and truncated forms of PEX14 and with pEGFP–PEX26 to label peroxisomal it is known that peroxisomes can move along microtubular tracks remnants. At 48 hours after transfection, live cell imaging was performed at in mammalian cells (Rapp et al., 1996; Schrader et al., 1996; 37°C. The arrowheads indicate peroxisomes moving quickly during image Wiemer et al., 1997). However, the mechanistic details of this capture (0, 20, 40 seconds) and the arrows marks the starting position. Whereas expression of full-length and N-terminal fragments PEX14(1–140) movement, especially the nature of the peroxisomal interface and and PEX14(1–260) restored directional movement of peroxisomes, the C- its functional relevance, remained unclear. Some light on the role terminal fragment PEX14(79–377) did not rescue the mobility defect. Scale of peroxisome–microtubule association was shed by the application bar: 10m. For movies, refer to supplementary material Movies 4–8. of microtubule-disrupting drugs, which led to clustering and non- mobility of mature peroxisomes (Wiemer et al., 1997), suggesting that microtubules are required for the spatial organization of motility. A C-terminal truncated version, PEX14(1–260) including peroxisomes. the conserved N-terminal region and the transmembrane domain Microtubular association also was reported to be crucial for the (Fig. 5A), could restore the peroxisomal directional movement formation of peroxisomal precursors (Brocard et al., 2005) and the (Fig. 5 and supplementary material Movie 6). However, a predicted regulated degradation of peroxisomes under starvation conditions coiled-coil region ranging from amino acid 141 to amino acid 180 (Hara-Kuge and Fujiki, 2008). In plant cells, the close association was also shown to be dispensable for complementation because of microtubules and peroxisomes is thought to improve the the directional movement of peroxisomes along microtubular tracks efficiency of import of several matrix enzymes with microtubule- PEX14-dependent peroxisome motility 1765

Fig. 6. Physical binding of the N-terminal domain of PEX14 to tubulin and PEX5. (A)GST or GST–PEX14(1–78) was bound to glutathione Sepharose 4B and incubated with purified porcine brain tubulin. The samples total (T), flow-through (FT), wash (W) and SDS- eluates (E) were analyzed by SDS-PAGE and Coomassie staining (left panel). The relative amount of tubulin bound to GST and GST– PEX14(1–78) was determined by densitometric quantification using the Odyssey software (right panel). Percentages indicate the ratio of tubulin and GST/GST–PEX14(1–78) amounts in the flow-through, wash and eluate fractions compared with the total amount in the load fraction. The values indicate means ± s.d. obtained from three independent binding experiments. (B)For competition analysis, GST–PEX14(1–78) bound to glutathione Sepharose 4B was incubated with purified porcine brain tubulin in the absence and presence of an overexcess of competing PEX5– W1 peptide or PEX5–WQ peptide as control (left panel). Alternatively, purified tubulin was bound to immobilized GST–PEX14(1–78) and then eluted with PEX5-W1 peptide (right panel).

binding domains (Chuong et al., 2005). The application of reported (Stelter et al., 2007). In this respect, it is noteworthy that microtubule-disrupting drugs has the disadvantage of provoking pexophagy in yeast involves the actin filaments rather than general defects that might interfere with peroxisome morphology microtubules (Reggiori et al., 2005). PEX14 is not required for non- and abundance. The finding that PEX14 is required for microtubule- selective autophagic peroxisome degradation (micropexophagy) association of peroxisomes now allows study of the phenotypic (de Vries et al., 2006), a process that does not necessarily involve effects of the non-motility of peroxisomes without harming other microtubules or microfilament attachment of peroxisomes (Kirisako cytoskeleton-dependent cellular processes. et al., 1999; Kunz et al., 2004). The existence of peroxisomal ghosts in PEX14-deficient cells Our data are clear in that PEX14 is required for peroxisome apparently contradicts the previous assumption that the long-range motility and that the protein serves as the connecting

Journal of Cell Science peroxisome–microtubule interaction plays a crucial role in the link for anchoring peroxisomes to microtubules. Analysis of the earliest stages of peroxisome biogenesis, namely the PEX16- molecular basis of the PEX14-association with microtubules dependent de novo synthesis of peroxisomal membranes (Brocard revealed two surprising findings. First, the interaction between et al., 2005). It might still be possible that early states of tubulin and PEX14 turned out to be direct and, second, the C- peroxisomes, so-called pre-peroxisomes, possess additional PEX14- terminal region of PEX14, which is required for its function in independent microtubule-binding sites; however, our data are clear peroxisomal protein import, is dispensable for tubulin binding and in that mature peroxisomal membranes rely on the PEX14-mediated its role in peroxisome motility. In fact, the 140 amino acid segment association with microtubules. that maintained the activity does comprise the N-terminal tubulin- There are an increasing number of findings in support of a role binding fragment (amino acids 1–78) and the peroxisomal for PEX14 in the autophagic degradation of peroxisomes. The membrane anchor sequence. This region also contains the binding importance of PEX14 for pexophagy has been demonstrated in sites for PEX5 and PEX19, raising the question whether they all yeast (Bellu et al., 2001; Zutphen et al., 2008) and indirect evidence exist in the same complex. An answer to this question is provided was presented for an involvement of PEX14 in peroxisome-specific by Fig. 3, which indicates that PEX14 forms three distinct autophagic processes in mammals. In fact, PEX14 from CHO cells complexes with its binding partners. For PEX19 and PEX5, it has has been shown to interact with the microtubule-associated protein already been shown that they both occupy the same binding site of I light chain 3 (LC3), which is required for pexophagy (Hara-Kuge PEX14 (Neufeld et al., 2009). The finding that PEX5 competes and Fujiki, 2008). All Hansenula polymorpha pex mutants, with with tubulin for binding to PEX14 (Fig. 6) provides a molecular the exception of pex14, exhibit sensitivity to selective degradation rationale for the existence of distinct functional subpopulations of of peroxisomal remnants (macropexophagy) (Zutphen et al., 2008). PEX14 at the peroxisomal membrane. We identified the N-terminal conserved region of PEX14 as being Interestingly, our findings might also explain the previously responsible for tubulin binding, and Bellu et al. identified the same observed perturbed microtubule association of peroxisomal region as being essential for regulated degradation of yeast remnants in PEX13-deficient cells (Nguyen et al., 2006) because peroxisomes (Bellu et al., 2001). In Pichia pastoris, PEX14 PEX13 was shown to be required for correct targeting of PEX14 interacts directly or indirectly with ATG30, which is a key player to the peroxisomal membrane (Fransen et al., 2004; Itoh and Fujiki, in the selection of peroxisomes for their delivery to the autophagy 2006). machinery (Farre et al., 2009). In Saccharomyces cerevisiae, an The fact that PEX5 and tubulin are not in the same complex interaction between PEX14 and the actin binding domain LC8 was might reflect a possible mechanism of regulation. When the high- 1766 Journal of Cell Science 124 (10)

affinity binding partner PEX5 is present in its cargo-loaded form, were blue dextran (V0void volume, larger than 2000 kDa), thyroglobulin (669 matrix protein import will occur and microtubule binding might be kDa), ferritin (440 kDa), (232 kDa) and aldolase (158 kDa). Fractions of 80 l were collected and subjected to immunoblot analysis using monoclonal anti- diminished. By contrast, upon starvation, less cargo-loaded receptor PEX19 antibodies (BD Biosciences, Heidelberg), monoclonal anti-Hs-tubulin might bind to PEX14, thus opening the way for tubulin association (Sigma-Aldrich) and anti-HsPEX14 (Will et al., 1999). Polyclonal rabbit anti- and regulated degradation. HsPEX5 antibodies were raised against recombinant His-tagged PEX5p, which was expressed in E.coli and purified as described previously (Schliebs et al., 1999). It has been suggested that cytosolic linker proteins (CLIPs), which attach organelle membranes to microtubules, might also Mass spectrometry play a role in connecting peroxisomes to microtubules (Rickard For mass spectrometric analyses, samples were suspended in LDS sample buffer (Invitrogen, Carlsbad, CA) and loaded onto NuPAGE Novex 4–12% acyrylamide and Kreis, 1996). CLIPs are components of a general mechanism Bis-Tris gels (Invitrogen) or 12% acrylamide Tris-glycine gels. Gel electrophoresis that promotes microtubule growth and organizes the dynein– was carried out according to the manufacturer’s protocol and subsequent staining dynactin complex (Galjart, 2005). Because these motor proteins was performed with colloidal Coomassie Brilliant Blue G-250. and other CLIP-interacting proteins like IQGAP1 were found as Processing of gels, in-gel tryptic digestion and online reversed-phase capillary HPLC/ESI-MS/MS analyses of peptide samples were performed as described constituents of the PEX14 complex (see supplementary material previously (Wiese et al., 2007). In brief, mass spectrometric analyses were performed Table S2), general steps in movements along microtubule tracks on a Bruker Daltonics HCT plus ion trap instrument (Bremen, Germany). The three might be used jointly by the peroxisomes and other organelles. most abundant ions were selected for fragmentation by low-energy collision-induced dissociation. Exclusion limits were automatically placed on previously selected mass-to-charge (m/z) ratios for 1.2 minutes. MS spectra were the sum of seven Materials and Methods individual scans ranging from m/z 300–1500 with a scanning speed of 8,100 Plasmids (m/z)/second; and MS/MS spectra were the sum of four scans ranging from m/z 100– Expression/integration plasmid pPEX14-TEV-PA encoding human PEX14 fused to 2200 at a scan rate of 26,000 (m/z)/second. For peptide identification, peaklists were a TEV protease cleavable Protein A tag was constructed in three subsequent steps. correlated with the human IPI database (http://www.ebi.ac.uk) and a duplicate of the First, the DNA-sequence encoding the TEV-protease cleavable Protein A was same database, in which the amino acid sequence of each protein entry was randomly PCR-amplified from plasmid pYM8 (Knop et al., 1999) using primer pair 5Ј- shuffled (Stephan et al., 2006), using MASCOT (version 2.2) (Perkins et al., 1999). TACACGATATCCGTACGCTGCAGGTCG-3Ј, 5Ј-CATCCGCGGCCGCCTAA - All searches were performed with tryptic specificity allowing two missed cleavages. AGAGCCGCGGAATT-3 Ј (EcoRV/NotI sites are underlined) and subcloned into Oxidation of methionine was considered as variable modification, as well as mass pcDNA5/FRT (Invitrogen). Second, an adaptor generated by oligonucleotides 5Ј- tolerances of 1.2 Da and 0.4 Da for MS and MS2 experiments, respectively. Proteins CTAGCGATATCATCGATAAGCTT-3Ј, 5Ј-AAGCTTATCGATGATATCG-3Ј was were assembled on the basis of peptide identifications using the ProteinExtractor subcloned into NheI/EcoRV digested vector introducing new EcoRV and HindIII Tool in ProteinScape (version 1.3 S.R. 2; Bruker Daltonics, Bremen, Germany). cloning sites. Third, PEX14 was PCR-amplified from pcDNA3-HsPEX14 (Will et ProteinExtractor was used with a MASCOT score threshold of 15. Subsequent to the al., 1999) using primer pair 5Ј-TAGATATCATGGCGTCCTCGGAGCAG-3Ј, 5Ј- assembly of proteins, hits up to an accumulated false positive rate of 5% were ACAAGCTTGTCCCGCTCACTCTCGTT-3Ј (EcoRV/HindIII sites are underlined) considered as true-positive protein identifications. For the relative quantification of and subcloned. For complementation assays, truncated fragments of PEX14 were proteins, spectral counts were calculated using in-house developed software (Liu et PCR-amplified and subcloned into pcDNA3.1 expression vector either digested by al., 2004). KpnI/XbaI (PEX14(79–377) or HindIII/NotI [(PEX14(1–140) and PEX14(1–260)]. The full-length PEX14 expression plasmid pcDNA3-HsPEX14 (Will et al., 1999), Fluorescence microscopy E.coli GST–PEX14(1–78) expression plasmid (Schliebs et al., 1999), as well as the Human fibroblast cells were cultured as described previously (Stanley et al., 2006). human GFP–PEX26 expression plasmid (Halbach et al., 2006) and pEGFP–PTS1 PEX14-deficient fibroblast cell line PEX14T (synonymous K-01 T) (Neufeld et al., (Stanley et al., 2006) have been described previously. The expression plasmid 2009), the PEX1-deficient cell line PBDG-01, kindly provided by Ronald Wanders encoding PTS2–EYFP was kindly provided by Gabriele Dodt (University of (AMC, The Netherlands) and the PEX5-deficient cell line PBD005 (Dodt et al., Tübingen, Germany). 1995) were SV40-transfected for immortalization resulting in PEX1T and PEX5T,

Journal of Cell Science respectively. PEX1T cells harbor the two heterozygous and severe mutations c.1960 Generation and selection of stable cell lines expressing PEX14–TEV– 1961dupCAGTGTGGA (p.T651_W653dup) and c.1108_1109insA (p.I370NfsX2) ProteinA (Ebberink et al., 2011). PEX14T cells are homozygous for a nonsense mutation in Flp-In system (Invitrogen, Carlsbad, CA) was used for the generation of a stable a putative coiled-coil region of PEX14, c.553C>T (p.Q185X) (Shimozawa et al., PEX14–TEV–Protein A expression cell line Flp-In-293[PEX14–TEV–PA]. Flp-In- 2004). The absence of the endogenous full-length or truncated PEX14 was confirmed 293 cells (Invitrogen) were co-transfected with expression vector pPEX14–TEV–PA by immunoblot analysis (data not shown). and the FLP-recombinase vector (pOG44) resulting in a stable integration of the For functionality and localization analyses of plasmid-encoded PEX14-TEV– gene of interest at the ‘FLP Recombination Target’ (FRT)-site in the genome ProteinA, PEX14-deficient fibroblast cells were double-transfected with expression (O’Gorman et al., 1991). For the selective growth test, individual cell colonies were vector pPEX14-TEV–PA alone or together with either EGFP–PTS1 (synonymous grown in duplicate on six-well plates and exposed to a confluence of 25%. The EGFP–SCP2) (Stanley et al., 2006) or PTS2–YFP using the Amaxa Nucleofector- culture medium was supplemented with hygromycin (200 g /ml) or Zeocin (100 Kit (Lonza, Switzerland). At 48 hours after transfection, cells were subjected to g/ml). The growth of possible integrants and Flp-In-293 cells as a control was fluorescence and immunofluorescence microscopy using antibodies against PEX14 monitored over a period of 7 days. Hygromycin-resistant and Zeocin-sensitive cells (Will et al., 1999), PMP70 (rabbit; Zymed/Invitrogen) and Protein A (Sigma-Aldrich). were further assayed for the lack of -galactosidase activity. The expression of PEX14–TEV–ProteinA was verified by immunoblot analysis using anti-PEX14 Live cell imaging antiserum (Will et al., 1999). To detect peroxisomes, cells were transfected with plasmid encoding GFP–PEX26 (Halbach et al., 2006). Fibroblast cells were observed 48 hours after Amaxa Isolation of PEX14 complex transfection in DMEM medium. Cells were observed by confocal laser scanning For complex isolation, cells were cultured on triple flasks (Nunc 500 cm2 each) on microscopy (Zeiss LSM510; Meta, Germany). To maintain the incubation settings at DMEM medium supplemented with hygromycin. Approximately 25 g cells (wet 37°C and 5% CO2 on the microscope stage, a CTI controller 3700 digital, O2 weight) were disrupted by vortexing with three volumina glass beads (diameter 0.5 controller, Tempcontrol 37-2 digital, and the Incubator Soxygen together with the mm) in 20 mM HEPES pH 7.5, 100 mM potassium acetate, 5 mM magnesium heating insert P (Zeiss) was used. Additionally, the immersion oil objective was acetate and selected protease inhibitors. Glass beads and cell debris were removed heated with a Tempcontrol Minisystem (Zeiss). by centrifugation. Cellular membranes were collected by centrifugation at 100,000 For time-lapse imaging, cells were imaged at low magnification (zoom level 1, g for one hour. Membrane proteins (5 mg/ml) were solubilized with 1% (w/v) 1024ϫ1024 pixels) and laser power of 5%. Images were taken every 5 seconds. At digitonin at a detergent/protein ratio of 2 for one hour. Insoluble material was each time point, optical sections were captured in 2-m steps, covering the thickness removed by centrifugation at 100,000 g for one hour. The supernatant was incubated of the cell, and reconstructions were obtained using the Zeiss image and analysis overnight with IgG Sepharose beads. Beads were concentrated by centrifugation, software (Zeiss LSM image browser version 4.2) (Theiss et al., 2005). thoroughly washed and incubated with TEV protease in lysis buffer containing 0.1% For disassembly of the actin and microtubule network, cells were incubated with (w/v) digitonin at 16°C for 2 hours. After centrifugation, the beads were washed with 10 g/ml Cytochalasin D for at least 30 minutes or 10 g/ml colchicine for 1 hour, lysis buffer containing 200 mM potassium acetate and 0.1% (w/v) digitonin. Eluates respectively. were combined and further analyzed by size-exclusion chromatography or mass spectrometry. In vitro binding For size-exclusion chromatography, TEV–protease eluate was separated on a Recombinant glutathione S-transferase (GST) or GST–PEX14 (1-78) (Schliebs et calibrated Superose 6 PC 3.2/30 column (GE Healthcare). Molecular mass markers al., 1999) was expressed in E. coli and bound to glutathione Sepharose 4B (GE PEX14-dependent peroxisome motility 1767

Healthcare) according to the manufacturer’s manual. After washing with TBS Hara-Kuge, S. and Fujiki, Y. (2008). The peroxin Pex14p is involved in LC3-dependent (50 mM Tris-Cl pH7.5, 150 mM NaCl), 10 l of Sepharose binding approximately degradation of mammalian peroxisomes. Exp. Cell Res. 314, 3531-3541. 100 g or 2.9 nmol of GST fusion protein were used for each experiment. The resin Itoh, R. and Fujiki, Y. (2006). Functional domains and dynamic assembly of the peroxin was incubated with 1 nmol porcine brain tubulin in 600 l TBS overnight at Pex14p, the entry site of matrix proteins. J. Biol. Chem. 281, 10196-10205. 4°C. Sepharose was isolated using MoBiCol spin column (MoBiTec, Göttingen, Kirisako, T., Baba, M., Ishihara, N., Miyazawa, K., Ohsumi, M., Yoshimori, T., Noda, Germany), followed by washing with 600 l TBS. Elution was performed using T. and Ohsumi, Y. (1999). Formation process of autophagosome is traced with 600 l SDS sample buffer. All purification steps were performed in the presence of Apg8/Aut7p in yeast. J. Cell Biol. 147, 435-446. 1 mM DTT. Proteins were analyzed by SDS-PAGE, stained with Coomassie R-250 Knop, M., Siegers, K., Pereira, G., Zachariae, W., Winsor, B., Nasmyth, K. and Schiebel, E. (1999). Epitope tagging of yeast genes using a PCR-based strategy: more and scanned using LI-COR Odyssey IR imaging system (LI-COR Biosciences, tags and improved practical routines. Yeast 15, 963-972. Lincoln, NE). Densitometric quantification of tubulin was performed using Odyssey Kunz, J. B., Schwarz, H. and Mayer, A. (2004). Determination of four sequential stages application software version 3.0.21. For competition analysis, 14.5 nmol synthetic during microautophagy in vitro. J. Biol. Chem. 279, 9987-9996. PEX5–W1 peptide representing the first N-terminal Wxxx[F/Y] motif of the human Kural, C., Kim, H., Syed, S., Goshima, G., Gelfand, V. I. and Selvin, P. R. (2005). PEX5 sequence (ALSENWAQEFLAAGDA, amino acids 113–128, Wxxx[F/Y] Kinesin and dynein move a peroxisome in vivo: a tug-of-war or coordinated movement? motif underlined), or a control peptide containing a WxxxQ motif Science 308, 1469-1472. (SDVDFWDKLQAELEEM, PEX5 amino acids 283– 298) (PEX5–WQ peptide), Lanyon-Hogg, T., Warriner, S. L. and Baker, A. (2010). Getting a camel through the was incubated together with tubulin or added after binding. eye of a needle: the import of folded proteins by peroxisomes. Biol. Cell 102, 245- Purification of porcine brain tubulin was carried out as described previously 263. (Waterman-Storer, 2001). Essentially, pig brain was homogenized and the supernatant Liu, H., Sadygov, R. G. and Yates, J. R., 3rd (2004). A model for random sampling and after centrifugation was enriched by tubulin through repetitive polymerization and estimation of relative protein abundance in shotgun proteomics. Anal. Chem. 76, 4193- depolymerization cycles and ultracentrifugation at 100,000 g. Polymerization occurred 4201. in 100 mM PIPES, 2 mM EGTA, 1 mM MgSO4, 2 mM DTE, pH 6.9. For Ma, C. and Subramani, S. (2009). Peroxisome matrix and membrane protein biogenesis. depolymerization, 8 M glycerol was added to this buffer. A final cation exchange IUBMB Life 61, 713-722. chromatography step yielded pure tubulin (supplementary material Fig. S2). Meinecke, M., Cizmowski, C., Schliebs, W., Kruger, V., Beck, S., Wagner, R. and Erdmann, R. (2010). The peroxisomal importomer constitutes a large and highly dynamic pore. Nat. Cell Biol. 12, 273-277. We thank Elisabeth Becker and Christiane Sprenger for technical Neufeld, C., Filipp, F. V., Simon, B., Neuhaus, A., Schuller, N., David, C., Kooshapur, assistance. This work was supported by the Deutsche H., Madl, T., Erdmann, R., Schliebs, W. et al. (2009). Structural basis for competitive Forschungsgemeinschaft (SFB642). interactions of Pex14 with the import receptors Pex5 and Pex19. EMBO J. 28, 745- 754. Supplementary material available online at Nguyen, T., Bjorkman, J., Paton, B. C. and Crane, D. I. (2006). Failure of microtubule- http://jcs.biologists.org/cgi/content/full/124/10/1759/DC1 mediated peroxisome division and trafficking in disorders with reduced peroxisome abundance. J. Cell Sci. 119, 636-645. O’Gorman, S., Fox, D. T. and Wahl, G. M. (1991). Recombinase-mediated gene References activation and site-specific integration in mammalian cells. Science 251, 1351-1355. Agne, B., Meindl, N. M., Niederhoff, K., Einwachter, H., Rehling, P., Sickmann, A., Perkins, D. N., Pappin, D. J., Creasy, D. M. and Cottrell, J. S. (1999). Probability-based Meyer, H. E., Girzalsky, W. and Kunau, W. H. (2003). Pex8p: an intraperoxisomal protein identification by searching sequence databases using mass spectrometry data. organizer of the peroxisomal import machinery. Mol. Cell 11, 635-646. Electrophoresis 20, 3551-3567. Bellu, A. R., Komori, M., van der Klei, I. J., Kiel, J. A. and Veenhuis, M. (2001). Rapp, S., Saffrich, R., Anton, M., Jakle, U., Ansorge, W., Gorgas, K. and Just, W. W. Peroxisome biogenesis and selective degradation converge at Pex14p. J. Biol. Chem. (1996). Microtubule-based peroxisome movement. J. Cell Sci. 109, 837-849. 276, 44570-44574. Reggiori, F., Monastyrska, I., Shintani, T. and Klionsky, D. J. (2005). The actin Brackley, K. I. and Grantham, J. (2009). Activities of the chaperonin containing TCP- cytoskeleton is required for selective types of autophagy, but not nonspecific autophagy, 1 (CCT): implications for cell cycle progression and cytoskeletal organisation. Cell in the yeast Saccharomyces cerevisiae. Mol. Biol. Cell 16, 5843-5856. Stress Chaperones 14, 23-31. Reguenga, C., Oliveira, M. E., Gouveia, A. M., Sa-Miranda, C. and Azevedo, J. E. Brocard, C. B., Boucher, K. K., Jedeszko, C., Kim, P. K. and Walton, P. A. (2005). (2001). Characterization of the mammalian peroxisomal import machinery: Pex2p, Requirement for microtubules and dynein motors in the earliest stages of peroxisome Pex5p, Pex12p, and Pex14p are subunits of the same protein assembly. J. Biol. Chem. biogenesis. Traffic 6, 386-395. 276, 29935-29942.

Journal of Cell Science Chuong, S. D., Park, N. I., Freeman, M. C., Mullen, R. T. and Muench, D. G. (2005). Rickard, J. E. and Kreis, T. E. (1996). CLIPs for organelle-microtubule interactions. The peroxisomal multifunctional protein interacts with cortical microtubules in plant Trends Cell Biol. 6, 178-183. cells. BMC Cell Biol. 6, 40. Saleem, R. A., Smith, J. J. and Aitchison, J. D. (2006). Proteomics of the peroxisome. de Vries, B., Todde, V., Stevens, P., Salomons, F., van der Klei, I. J. and Veenhuis, M. Biochim. Biophys. Acta 1763, 1541-1551. (2006). Pex14p is not required for N-starvation induced microautophagy and in catalytic Schliebs, W., Saidowsky, J., Agianian, B., Dodt, G., Herberg, F. W. and Kunau, W. H. amounts for macropexophagy in Hansenula polymorpha. Autophagy 2, 183-188. (1999). Recombinant human peroxisomal targeting signal receptor PEX5. Structural Dodt, G., Braverman, N., Wong, C., Moser, A., Moser, H. W., Watkins, P., Valle, D. and basis for interaction of PEX5 with PEX14. J. Biol. Chem. 274, 5666-5673. Gould, S. J. (1995). Mutations in the PTS1 receptor gene, PXR1, define complementation Schollenberger, L., Gronemeyer, T., Huber, C. M., Lay, D., Wiese, S., Meyer, H. E., group 2 of the peroxisome biogenesis disorders. Nat. Genet. 9, 115-125. Warscheid, B., Saffrich, R., Peranen, J., Gorgas, K. et al. (2010). RhoA regulates Ebberink, M. S., Mooijer, P. A., Gootjes, J., Koster, J., Wanders, R. J. and Waterham, peroxisome association to microtubules and the actin cytoskeleton. PLoS ONE 5, H. R. (2011). Genetic classification and mutational spectrum of more than 600 patients e13886. with a Zellweger syndrome spectrum disorder. Hum. Mutat. 32, 59-69. Schrader, M., Burkhardt, J. K., Baumgart, E., Luers, G., Spring, H., Volkl, A. and Erdmann, R., Veenhuis, M., Mertens, D. and Kunau, W.-H. (1989). Isolation of Fahimi, H. D. (1996). Interaction of microtubules with peroxisomes. Tubular and peroxisome-deficient mutants of Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA spherical peroxisomes in HepG2 cells and their alterations induced by microtubule- 86, 2432-2436. active drugs. Eur. J. Cell Biol. 69, 24-35. Fagarasanu, A., Mast, F. D., Knoblach, B., Jin, Y., Brunner, M. J., Logan, M. R., Schrader, M., King, S. J., Stroh, T. A. and Schroer, T. A. (2000). Real time imaging Glover, J. N., Eitzen, G. A., Aitchison, J. D., Weisman, L. S. et al. (2009). Myosin- reveals a peroxisomal reticulum in living cells. J. Cell Sci. 113, 3663-3671. driven peroxisome partitioning in S. cerevisiae. J. Cell Biol. 186, 541-554. Shimozawa, N., Tsukamoto, T., Nagase, T., Takemoto, Y., Koyama, N., Suzuki, Y., Farre, J. C., Krick, R., Subramani, S. and Thumm, M. (2009). Turnover of organelles Komori, M., Osumi, T., Jeannette, G., Wanders, R. J. et al. (2004). Identification of by autophagy in yeast. Curr. Opin. Cell Biol. 21, 522-530. a new complementation group of the peroxisome biogenesis disorders and PEX14 as Fransen, M., Vastiau, I., Brees, C., Brys, V., Mannaerts, G. P. and Van Veldhoven, P. the mutated gene. Hum. Mutat. 23, 552-558. P. (2004). Potential role for Pex19p in assembly of PTS-receptor docking complexes. Stanley, W. A., Filipp, F. V., Kursula, P., Schuller, N., Erdmann, R., Schliebs, W., J. Biol. Chem. 279, 12615-12624. Sattler, M. and Wilmanns, M. (2006). Recognition of a functional peroxisome type 1 Fujiki, Y., Miyata, N., Matsumoto, N. and Tamura, S. (2008). Dynamic and functional target by the dynamic import receptor Pex5p. Mol. Cell 24, 653-663. assembly of the AAA peroxins, Pex1p and Pex6p, and their membrane receptor Pex26p Steinberg, S. J., Dodt, G., Raymond, G. V., Braverman, N. E., Moser, A. B. and Moser, involved in shuttling of the PTS1 receptor Pex5p in peroxisome biogenesis. Biochem. H. W. (2006). Peroxisome biogenesis disorders. Biochim. Biophys. Acta 1763, 1733- Soc. Trans. 36, 109-113. 1748. Galjart, N. (2005). CLIPs and CLASPs and cellular dynamics. Nat. Rev. Mol. Cell Biol. Stelter, P., Kunze, R., Flemming, D., Hopfner, D., Diepholz, M., Philippsen, P., 6, 487-498. Bottcher, B. and Hurt, E. (2007). Molecular basis for the functional interaction of Grou, C. P., Carvalho, A. F., Pinto, M. P., Wiese, S., Piechura, H., Meyer, H. E., dynein light chain with the nuclear-pore complex. Nat. Cell Biol. 9, 788-796. Warscheid, B., Sa-Miranda, C. and Azevedo, J. E. (2008). Members of the E2D Stephan, C., Reidegeld, K. A., Hamacher, M., van Hall, A., Marcus, K., Taylor, C., (UbcH5) family mediate the ubiquitination of the conserved cysteine of Pex5p, the Jones, P., Muller, M., Apweiler, R., Martens, L. et al. (2006). Automated reprocessing peroxisomal import receptor. J. Biol. Chem. 283, 14190-14197. pipeline for searching heterogeneous mass spectrometric data of the HUPO Brain Halbach, A., Landgraf, C., Lorenzen, S., Rosenkranz, K., Volkmer-Engert, R., Proteome Project pilot phase. Proteomics 6, 5015-5029. Erdmann, R. and Rottensteiner, H. (2006). Targeting of the tail-anchored peroxisomal Theiss, C., Napirei, M. and Meller, K. (2005). Impairment of anterograde and retrograde membrane proteins PEX26 and PEX15 occurs through C-terminal PEX19-binding neurofilament transport after anti-kinesin and anti-dynein antibody microinjection in sites. J. Cell Sci. 119, 2508-2517. chicken dorsal root ganglia. Eur. J. Cell Biol. 84, 29-43. 1768 Journal of Cell Science 124 (10)

Thoms, S. and Erdmann, R. (2005). Dynamin-related proteins and Pex11 proteins in Wiemer, E. A., Wenzel, T., Deerinck, T. J., Ellisman, M. H. and Subramani, S. (1997). peroxisome division and proliferation. FEBS J. 272, 5169-5181. Visualization of the peroxisomal compartment in living mammalian cells: dynamic Tsukamoto, T., Yokota, S. and Fujiki, Y. (1990). Isolation and characterization of Chinese behavior and association with microtubules. J. Cell Biol. 136, 71-80. hamster ovary cell mutants defective in assembly of peroxisomes. J. Cell Biol. 110, Wiese, S., Gronemeyer, T., Ofman, R., Kunze, M., Grou, C. P., Almeida, J. A., 651-660. Eisenacher, M., Stephan, C., Hayen, H., Schollenberger, L. et al. (2007). Proteomics Vasquez, R. J., Howell, B., Yvon, A. M., Wadsworth, P. and Cassimeris, L. (1997). characterization of mouse kidney peroxisomes by tandem mass spectrometry and protein Nanomolar concentrations of nocodazole alter microtubule dynamic instability in vivo correlation profiling. Mol. Cell. Proteomics 6, 2045-2057. and in vitro. Mol. Biol. Cell 8, 973-985. Will, G. K., Soukupova, M., Hong, X., Erdmann, K. S., Kiel, J. A., Dodt, G., Kunau, Waterham, H. R., Koster, J., van Roermund, C. W., Mooyer, P. A., Wanders, R. J. W.-H. and Erdmann, R. (1999). Identification and characterization of the human and Leonard, J. V. (2007). A lethal defect of mitochondrial and peroxisomal fission. orthologue of yeast Pex14p. Mol. Cell. Biol. 19, 2265-2277. N. Engl. J. Med. 356, 1736-1741. Zutphen, T., Veenhuis, M. and van der Klei, I. J. (2008). Pex14 is the sole Waterman-Storer, C. M. (2001). Microtubule/organelle motility assays. Curr. Protoc. component of the peroxisomal translocon that is required for pexophagy. Autophagy Cell Biol. Chapter 13, Unit 13 1. 4, 63-66. Journal of Cell Science