ARTICLE IN PRESS. Review. Dorothee Lay, Karin Gorgas, Wilhelm W. Just

Transcription

1 BBAMCR-15485; No. of pages: 10; 4C: 6 + model Biochimica et Biophysica Acta xx (2006) xxx xxx Review Peroxisome biogenesis: Where Arf and coatomer might be involved Dorothee Lay, Karin Gorgas, Wilhelm W. Just Biochemie-Zentrum der Universität Heidelberg (BZH), Im Neuenheimer Feld 328D Heidelberg, Germany Received 19 May 2006; received in revised form 12 August 2006; accepted 23 August 2006 Abstract The present review summarizes recent observations on binding of Arf and COPI coat to isolated rat liver peroxisomes. The general structural and functional features of both Arf and coatomer were considered along with the requirements and dependencies of peroxisomal Arf and coatomer recruitment. Studies on the expression of mammalian Pex11 proteins, mainly Pex11α and Pex11β, intimately related to the process of peroxisome proliferation, revealed a sequence of individual steps including organelle elongation/tubulation, formation of membrane and matrix protein patches segregating distinct proteins from each other, development of membrane constrictions and final membrane fission. Based on the similarities of the processes leading to cargo selection and concentration on Golgi membranes on the one hand and to the formation of peroxisomal protein patches on the other hand, an implication of Arf and COPI in distinct processes of peroxisomal proliferation is hypothesized. Alternatively, peroxisomal Arf/COPI might facilitate the formation of COPI-coated peroxisomal vesicles functioning in cargo transport and retrieval from peroxisomes to the ER. Recent observations suggesting transport of Pex3 and Pex19 during early steps of peroxisome biogenesis from the ER to peroxisomes inevitably propose such a retrieval mechanism, provided the ER to peroxisome pathway is based on transporting vesicles Published by Elsevier B.V. Keywords: Peroxisome; Biogenesis; Arf; Small GTPase; Coatomer; COPI 1. Introduction Recent work in the yeast S. cerevisiae provided evidence that the peroxisomal membrane is derived from the endoplasmic reticulum (ER) [1]. Cells that do not express a functional Pex3 protein and hence do not contain peroxisomes formed new peroxisomes upon expression of a functional PEX3 gene. Formation of new peroxisomes has been suggested to require targeting of Pex3 to the ER and Pex19 supported packaging of Pex3 into budding vesicles that sequestered from the ER for maturation. A similar mechanism involving an ER-derived Abbreviations: Arf, ADP ribosylation factor; BFA, brefeldin A; COP, coat protein; ER, endoplasmic reticulum; GAP, GTPase activating protein; GEF, guanine nucleotide exchange factor; GDPβS, Guanosine 5 -[β-thio]diphosphate; GMP-PNP, Guanosine 5 -[β,γ-imido]triphosphate; PtdOH, phosphatidic acid; Pex, peroxin; PH, pleckstrin homology; PLD, phospholipase D; PMP, peroxisomal membrane protein; PId, phosphoinositide; PtdIns(4,5)P 2, phosphatidylinositol(4,5)-bisphosphate; PPAR, peroxisome proliferator-activated receptor Corresponding author. Tel.: ; fax: address: wilhelm.just@bzh.uni-heidelberg.de (W.W. Just). structure containing the 70 kda peroxisomal membrane protein (PMP70) and Pex13 was proposed to occur in mouse dendritic cells [2,3]. Studies in COS-7 cells expressing GFP-tagged versions of Pex16 further supported an ER to peroxisome transport [4]. Whereas these observations suggest the ER to be implicated in the formation of new peroxisomes, other studies in yeast and mammalian systems indicated the ability of peroxisomes to divide autonomously without involvement of the ER [5 8]. Studies supporting this concept were based on the expression of Pex11α or Pex11β leading to peroxisome proliferation [9 11]. Pex11-mediated division of peroxisomes so far has been observed in a variety of organisms ranging from yeast to man (reviewed in: [12]). All these observations suggest that the formation of a new peroxisome may follow two different pathways including either vesiculation of the ER or budding from preexisting peroxisomes. The present review summarizes recent observations on binding of ADP-ribosylation factor (Arf) and the COPI coat (coatomer) to peroxisomal membranes in vitro that may shed some light on these processes of peroxisome biogenesis /$ - see front matter 2006 Published by Elsevier B.V.

2 2 D. Lay et al. / Biochimica et Biophysica Acta xx (2006) xxx xxx 2. Arf regulators of membrane traffic This section reviews the functions of Arf1 and coatomer focusing on selected aspects necessary to discuss the potential role Arf/coatomer might play in peroxisome biogenesis Functions of Arf Arf molecules belong to the Ras superfamily of low molecular weight GTPases that by themselves represent a conserved family of proteins. In mammalian systems the Arffamily consists of six proteins that have been divided into three classes based on sequence homologies [13]. Accordingly, human Arf1 and Arf3 belong to class I, Arf4 and Arf5 to class II and Arf6 to class III. Arf2, belonging to class I, has been lost in humans but not in other mammals. Yeast contains three family members that are assigned to class I (ScArf1 and ScArf2) and class III (ScArf3). Arf molecules are cotranslationally modified by N-myristoylation that is essential for their membrane contacts and biological activities [14,15]. Common to all low molecular weight GTPases including Arf is their cycling between the active GTP-bound and the inactive GDPbound state. Exchange of GTP for GDP occurring during Arf cycling is supported by guanine nucleotide exchange factors (GEFs), while GTPase-activating proteins (GAPs) pivotally stimulate GTP hydrolysis on the Arfs (reviewed in [16,17]). Numerous GEFs and GAPs have been identified and are believed to confer intracellular specificity to the different Arfs. Arf1 has been shown to reversibly bind to target membranes in its GTP-bound form and upon GTP hydrolysis is released into the cytosol. The GTP-bound Arf1 recruits the COPI complex onto cis-golgi structures and the adaptor protein (AP) complexes AP1, AP3, AP4, and GGA (Golgi-localized, γ-earcontaining, Arf-binding proteins) onto structures of the trans- Golgi network and endosomal membranes [18 20]. Arf6 does not appear to act on the Golgi but rather appears to be localized to the plasma membrane; to some extent it also occurs intracellularly on endosomal membranes regulating intracellular traffic and plasma membrane actin [21 24]. Some of the effects of Arf6 may be related to its activity to modulate the membrane concentration of acidic phospholipids including phosphatidylinositol (4,5)-bisphosphate (PtdIns(4,5) P 2 ) and phosphatidic acid (PtdOH) [25,26]. The ability to interact with phosphoinositide metabolism is also attributed to Arf1 that, for example, has been shown to recruit phosphatidylinositol 4-kinase β (PI4Kβ) to the Golgi and also to regulate PtdIns4P 5-kinase α [27 29]. Another Arf1 effector is phospholipase D1 (PLD1) enhancing the level of PtdOH and hence triggering the synthesis of PtdIns(4,5)P 2. PLD1 has been localized to intracellular vesicles and may not be localized to the Golgi [30]. Arf molecules are myristoylated at the N-terminus. Structural studies established that the lipid anchor together with the amphipathic N-terminal α-helix exposed in the GDP-bound conformation mediates weak membrane association [31,32]. Actually, it has been claimed that Arf1-GDP first interacts with a p23 oligomer before nucleotide exchange takes place [33].To the early Golgi Arf1 is recruited by membrin, an ER-Golgi SNARE protein [34]. All Arf-GEFs identified so far possess an about 200 amino acid Sec7 domain that is sufficient for GEF activity [16,35]. Arf-GEFs are usually divided into two classes the high (> 100 kda) and the low (45 50 kda) molecular weight GEFs. The Sec7 domain is target of the fungal metabolite brefeldin A (BFA) that thereby reacts with Arf-GDP-GEF, stabilizing a reaction intermediate and thus interrupting the Arf cycle [36,37]. Most of the high molecular weight GEFs are affected in this way, human GBF1 (Golgi BFA resistant factor 1) may be an exception [38 40]. Interestingly, although containing the Sec7 domain, the low molecular weight Arf-GEFs are largely resistant to the drug. A 35-amino acid region within the Sec7 domains of BFA-sensitive and BFA-resistant Arf-GEFs exhibit high sequence variations proposed to be responsible for the differences in BFA sensitivity [36, 41]. Most of the low molecular weight Arf-GEFs have a common domain structure including a N-terminal coiled-coil domain mediating homodimerization, the central Sec7 domain followed by a pleckstrinhomology (PH) domain [16,42,43]. The PH domain is known to mediate membrane association by binding to phosphoinositides leading to a remarkable stimulation of GTP-GDP nucleotide exchange on Arf [44]. Similar to the Arf-GEFs, a large number of Arf-GAPs have been identified so far. Some of them, e.g. ARAP1, PAP1 or GIT2 short, also contain PH domains mediating interaction with phosphoinositides particularly PtdInsP 3 and PtdInsP 2 facilitating membrane recruitment [17,45,46] Structure and functions of the COPI coat Two COP coats, COPI and COPII, are known in eukaryotic cells. Despite the similarity in nomenclature of these COP coats, COPI exhibits strong homology to the clathrin-ap coat rather than to COPII suggesting a common ancestral origin [47 51]. COPI is present in the cytosol as a large heptameric complex composed of the α, β, β, γ, δ, ε and ζ subunits. Its recruitment to the target membrane is mediated by Arf1-GTP. Two subcomplexes of COPI, the F- and B-subcomplex, form a functional unit upon Arf1-mediated association with the membrane promoting polymerization of the coat, induction of membrane curvature, cargo selection (membrane protein and lipid exclusion) and membrane budding [51]. Structural studies revealed striking topological similarities both between the COPI F-subcomplex, consisting of the β, γ, ζ and δ subunits, and the AP complex and between the COPI B- subcomplex, consisting of the α, β and ε subunits, and its likely functional equivalent clathrin [52]. A common functional feature of COPI and the clathrin-ap complexes, except the clathrin AP2 complex, is their requirement for Arf1 to recruit to target membranes. By binding to membranes, the interaction of the subcomplexes favors COPI oligomerization and the direct or indirect association with cargo molecules. These interactions are mediated by conserved β-propeller domains in α- and β - COP formed by WD40 domains and also formed in the clathrin molecule (reviewed in [52]). The coat/cargo association

3 D. Lay et al. / Biochimica et Biophysica Acta xx (2006) xxx xxx 3 concentrates cargo and enables coat polymerization, two processes that are intimately dependent on each other [53,54]. Coat assembly might initiate membrane curvature either directly through the nature of the coat complex and/or indirectly through accessory proteins. Subdomains, for example, in β- and γ-cop, called appendage domains, might function in recruiting accessory proteins required for budding. Indication for this is derived from studies on mutagenized γ-cop that has lost its ability to interact with Arf-GAP [55]. In this context it might be interesting to note that the finding of additional γ- and ζ-cop subunits (γ2-, ζ2-cop [56, 57]) suggested the existence of three isotypes of coatomer in mammalian cells. Each isotype is defined by a distinct γ/ζ subcomplex consisting of γ1/ζ1-, γ1/ ζ2-, or γ2/ζ1-cop, and occurring in a ratio of about 2:1:2 [58]. Thus, coatomer might exist in form of three isotypes serving diverse functions and exhibiting different intracellular locations. Considering accessory proteins and lipid regulators proposed to be involved in constitutive membrane traffic, models emerged elucidating the complex processes regulating vesicle budding. These models take into account basal, negative or positive feedback mechanisms controlling membrane-bound Arf. In the basal state and the state where a negative feedback loop is established, Arf cycles, however, the rates of GTP exchange and hydrolysis are balanced. Arf, although getting activated, is subsequently released and returns to the inactive state [59]. The question remains to be answered whether this seemingly useless cycling of Arf/coatomer is of physiological relevance. Although coated vesicles may not be formed, cargo concentration may proceed. In case that a positive feedback loop is initiated, GEF activation of Arf results in the production of regulatory phosphoinositides that act synergistically with Arf. The concentrations of both Arf and coatomer are raised above a critical level leading to vesicle budding. Other factors and enzymes, such as phosphoinositide phosphatases, PtdOH hydrolase or PI-kinase kinase in addition might influence this process and adopt it to the cellular requirements [46,59]. 3. Arf and COPI coat on peroxisomes Peroxisome proliferators, such as hypolipidemic drugs, particularly in rodents cause proliferation of peroxisomes mainly in liver and kidney [60,61]. Proliferation is raised by peroxisome proliferator-activated receptor (PPAR) α-mediated induction of genes coding for distinct peroxisomal matrix and membrane proteins. PMP70, a member of the family of peroxisomal ABC-transporters, and Pex11α, one of the three known isoforms of Pex11, are two integral PMPs that most significantly raise in concentration within the peroxisomal membrane upon drug treatment [62,63]. Whereas induction of PMP70 might be related to the enhanced transport of acyl-coas into peroxisomes, induction of Pex11α rather might be concerned with biogenetic processes. This view is mainly supported by observations correlating peroxisome abundance with expression levels of Pex11 in a variety of species including mammals (reviewed in: [12]). Mammals, trypanosomes, plants and yeast all express three Pex11 isoforms (see: [12]). The family of mammalian Pex11 proteins consists of Pex11α, Pex11β and Pex11γ. Pex11γ like Pex11β is of low abundance, constitutively expressed in liver and not induced by peroxisome proliferators [64]. The primary sequence of Pex11α, but not Pex11β and Pex11γ, contains a C-terminal dilysine-based retrieval motif of the type KXKXX known to mediate retrograde Golgi to ER transport [9,62]. Functional dilysine motifs so far have been found in a number of p24 family proteins that specifically recruit the COPI B-subcomplex [65 67]. There are other p24 members containing additional double phenylalanine motifs mediating recruitment of the COPI F-subcomplex [52,66] (see above). Whereas these p24 family proteins are ER/Golgi-resident type I transmembrane proteins with a large N-terminal domain facing the lumen and a short cytoplasmic tail, Pex11α is a type II transmembrane protein exposing its N- and C-terminus toward the cytoplasm [9,62]. However, common to all of these proteins is the short extramembranous C-terminal tail consisting of 8 10 amino acid residues bearing the dilysine motif and facing the cytoplasmic side of the organelle membrane. As the C-terminal tails of the ER/Golgi-localized proteins bearing the dilysine motif have clearly been shown to be involved in the recruitment of COPI, presence of the dilysine motif in Pex11α consequently suggested Arf/COPI coat binding to peroxisomes. Subsequent in vitro experiments on isolated rat liver peroxisomes demonstrated Arf and all the seven subunits of the COPI coat to be recruited to the organelles from rat liver cytosol [9,68]. Using isolated rat liver peroxisomes and rat liver cytosol various aspects of peroxisomal Arf/coatomer binding were investigated. In a series of experiments, the tail peptides of rat Pex11α and trypanosome Pex11 were found to recruit coatomer from cytosol of rat liver, bovine brain, trypanosomes and S. cerevisiae [69]. Binding was observed neither with mutated trypanosomal peptides having exchanged the lysine for serine residues nor with HsPex11β and ScPex11 that both do not contain a consensus motif. However, overexpression in trypanosomes of a mutated full length version of TbPex11, corresponding to the tail peptide that no longer bound coatomer, resulted in proliferation and clustering of peroxisomes. In S. cerevisiae, this mutant TbPex11 complemented the ScPex11 deletion [69]. Although these data are difficult to interpret, since the Pex11-related trypanosomal proteins GIM5A and GIM5B [70] were not considered at that time and none of the Pex11 family members in S. cerevisiae does contain a dilysine motif [71 73], they suggest that Pex11α-mediated peroxisome proliferation does not require a functional dilysine motif Factors affecting peroxisomal Arf/coatomer recruitment Analyzing PPARα-mediated induction of peroxisome proliferation on peroxisomal Arf/coatomer recruitment revealed highest amounts of Arf and coatomer in incubations containing peroxisomes and cytosol from stimulated livers (Fig. 1A, [74]). Whereas more coatomer was recruited from stimulated cytosol, more Arf was recruited to stimulated peroxisomes. The data show that Arf/ coatomer binding is significantly affected by the induction of peroxisome proliferation and suggest the involvement in this process of two factors, a peroxisomal and a cytosolic one.

4 4 D. Lay et al. / Biochimica et Biophysica Acta xx (2006) xxx xxx Fig. 1. Effects of peroxisome proliferation and BFA on Arf and coatomer binding to isolated peroxisomes. (A) Cytosol and highly purified rat liver peroxisomes were isolated from normal (U, unstimulated) and clofibrate-treated (S, stimulated) rats. Highest concentrations of Arf and coatomer were recruited when stimulated peroxisomes were incubated with stimulated cytosol (lane 1). Higher concentrations of Arf were bound to stimulated peroxisomes (lane 2), whereas a higher amount of coatomer represented by the β COP subunit was recruited from stimulated cytosol (lane 3). Arf/coatomer recruitment was dependent on GTP and did not occur in the presence of GDPβS (lane 5). The Pex11α signal reflects the expression level of Pex11α induced by the peroxisome proliferator. Expression of PMP22 is not affected by the drug treatment. (B) Binding of Arf1 from rat liver cytosol to isolated peroxisomes and Golgi membranes in the presence (+) and absence ( ) of BFA. Note that presence of BFA in the incubations affects Arf1 recruitment to the Golgi but not to peroxisomes. The Pex11α and p23 signals reflect the amount of peroxisomes and Golgi membranes loaded onto the gel, respectively. Cytosol fractions obtained by gel chromatography were used to study the activity of cytosolic components and to answer questions, such as, is Arf binding necessary for peroxisomal coatomer recruitment and if so which Arf subtype mediates this coatomer recruitment. The results demonstrated that preceding Arf binding is necessary for coatomer recruitment to occur. Mass spectroscopic analysis of the Arfs bound to peroxisomes identified peptides specific for Arf1 and Arf6 suggesting that these Arfs interact with the peroxisomal membrane. Subsequent experiments using recombinant Arf proteins further revealed that although both Arf subtypes were bound to peroxisomes in a GTP-dependent manner, it was only Arf1 that supported peroxisomal coatomer recruitment [74]. Arf/coatomer binding to peroxisomes was regulated by both ATP and a cytosolic activity. In the presence of ATP both Arf1 and coatomer binding were reduced. The concentration of bound Arf6 was even enhanced. The effect of ATP on Arf1 required the hydrolysis of ATP and occurred irrespective of the presence of a cytosolic pool fraction void of both Arf and coatomer. The cytosolic activity that triggered recruitment of Arf and coatomer was localized to a cytosolic fraction eluting by gel chromatography between the coatomer- and the Arfcontaining fractions. This cytosolic pool fraction significantly reduced recruitment of Arf1 but not Arf6 and increased the concentration of coatomer on the peroxisomal membrane. From these results it was concluded that ATP and a cytosolic factor independently from each other affect Arf/coatomer on peroxisomes. The ATP-dependent activity is removed from the membrane by carbonate treatment, but can be regained from the cytosolic fraction free of Arf and coatomer. How can all these findings be interpreted? Although there are several ATPases confined to the peroxisomal membrane including the families of peroxisomal ABC transporters and AAA proteins, these ATPases are implicated in metabolic functions and late steps of peroxisomal matrix protein import, respectively [75 78]. Therefore, one might assume the activity in question is a kinase activity being recruited from cytosol onto peroxisomes. Interestingly, a mechanism involving casein kinase 1δ has been suggested to regulate Arf activity at the Golgi membrane [79]. Activation of Arf-GAP1 by phosphorylation may accelerate GTP hydrolysis on Arf1 resulting in its membrane release. Alternatively, the relevant kinase might be a phosphoinositide kinase, as on Golgi membranes Arf1 directly associates in a phosphoinositide-dependent manner [80]. Actually, recent work in different laboratories including ours provided strong evidence for the synthesis of phosphoinositides on the peroxisomal membrane [81 83] Arf-dependent peroxisome proliferation in S. cerevisiae The putative function of Arf in oleate-induced peroxisome proliferation in vivo was investigated in the yeast S. cerevisiae. Cells carrying deletions (ScArf2 and ScArf3) and/or temperature sensitive alleles (ScArf1) were investigated [74]. The experiments revealed that ScArf1 was essential for peroxisome proliferation whereas deletion of ScArf3, the ortholog of mammalian Arf6, significantly enhanced proliferation (Fig. 2). The data indicate that ScArf1 and ScArf3 are implicated in biogenetic processes of yeast peroxisomes in vivo and favor a model in which ScArf1 and ScArf3 regulate peroxisome division in a positive and negative way, respectively. Such a dual regulation of one and the same process does not represent redundant activities but rather is an efficient means of triggering and amplifying important cellular processes Analysis of Pex11 functions Summarizing these data related to the interaction of Arf/ coatomer with peroxisomes we find observations that argue both for and against such an interaction. The hypothesis is favored, for example, by the fact that the dilysine motif is conserved throughout mammalian Pex11α proteins and that overexpression of Pex11α promotes proliferation of peroxisomes [9,11,62,84]. Also in line with these are observations showing that the induction of PPARα-mediated peroxisome proliferation known to stimulate expression of Pex11α distinctly affects binding of Arf and coatomer (Fig. 1A). However, the increase in Arf/coatomer binding is not in a stoichiometric relation to the remarkable increase in concentration of Pex11α, suggesting Pex11α not to be a rate-limiting

5 D. Lay et al. / Biochimica et Biophysica Acta xx (2006) xxx xxx 5 Fig. 2. Involvement of ScArf1 and ScArf3 in oleate-induced peroxisome proliferation in the yeast S. cerevisiae. Cells of the indicated strains were grown on either glucose (SD medium) at 25 C or for 6 h on oleate at the permissive (25 C) or nonpermissive (35 C) temperature. Cells were analyzed for the number of GFP-labeled peroxisomes (Pox) by confocal laser scanning microscopy. The increase in number of peroxisomes per cell is given in percent. Note that arf1 and arf3 exert a positive and negative effect on peroxisome proliferation, respectively. The mutant strain sec23-ts is blocked in protein export from the ER [93] and served as control regarding a general impairment of the secretory pathway (for details, see [74]). factor in this process. Furthermore, a conditionally lethal mutant strain of CHO cells expressing a temperature-sensitive allele of ε-cop exhibited various phenotype alterations related to a disrupted ER-Golgi transport including rapid degradation of low density lipoprotein receptors and disintegration of the Golgi apparatus [85]. In addition, these cells, when transfected with RnPex11α and kept at the nonpermissive temperature, frequently generated clusters of significantly elongated tubular peroxisomes suggesting proliferation but impaired division of the organelle [9]. On the other hand, several arguments have been raised against the involvement of Arf/coatomer in biogenetic processes of peroxisomes. Some of these were related to the activity of BFA [86]. BFA and mutant versions (T39N or H79G) of the small GTPase Sar1 regulating COPII recruitment to the ER were analyzed along with import studies of Pex3 and several other peroxins involved in early steps of peroxisome biogenesis [87,88]. These inhibitors bi-directionally blocked ER-Golgi transport, however, neither directly nor indirectly affected peroxisome biogenesis, a conclusion that has also been suggested by studies in S. cerevisiae demonstrating that inhibition of vesicular transport in a temperature-sensitive sec23 strain does not influence oleate-induced peroxisome proliferation (Fig. 2 and [74]). These results strongly argue against a role of Arf1/coatomer or Sar1/COPII coat in ER to peroxisome transport of membranes. As these inhibitors relate to processes involved in ER-Golgi traffic, they are not in conflict with observations considering Arf/coatomer recruitment to peroxisomal membranes. This recruitment has been investigated in detail for its BFA sensitivity and it has been shown that different to Golgi membranes it was resistant to BFA (Fig. 1B) suggesting distinct Arf-GEFs to be active on Golgi and peroxisomal membranes. Experiments dealing with the interaction of Arf1/coatomer with peroxisomes indicated a potential interrelation between Arf1/coatomer and Pex11. Consequently, focusing on functions of Pex11 proteins might be a suitable means to further elucidate the role Arf1/coatomer might play on peroxisomes. Of the three Pex11 proteins (Pex11α, Pex11β, and Pex11γ) expressed in rat liver, only Pex11α is strongly induced by peroxisome proliferators [9,62]. Pex11α knockout mice showed a normal peroxisome proliferation response to classical peroxisome proliferators, such as WY-14,643 or ciprofibrate that both activate PPARα. However, they no longer responded to 4-phenylbutyrate that induces the expression of Pex11α, but different to the classical proliferators acts independently of PPARα [89]. Thus, Pex11α is required for 4-phenylbutyrate-stimulated peroxisome proliferation. Using proliferators, such as clofibrate and/or thyroxin [63] a moderate induction of Pex11β is also observed (Lay et al., unpublished). Interestingly, in rat hepatocytes these proliferators induced the frequent appearance of multiply constricted tubular peroxisomes resembling intermediate states of proliferation prior to fission, i.e. increase in the absolute number of organelles (Fig. 3G, H). These constricted but still interconnected peroxisomal segments were highly variable in size making predictions difficult at which state fission actually might occur. Overexpression of Pex11α and Pex11β but not Pex11γ induced peroxisome proliferation in different cell types, Pex11β being more efficient than Pex11α. This difference in the activity

6 6 D. Lay et al. / Biochimica et Biophysica Acta xx (2006) xxx xxx Fig. 3. Peroxisome proliferation induced by overexpression of Pex11β and by peroxisome proliferators. (A E) Expression of N-terminally myc-tagged Pex11β in CHO cells causing peroxisome tubulation and segregation of peroxisomal membrane proteins. Peroxisomes were stained for Nmyc-Pex11β (A, D), PMP70 (B), and Pex11α (E). In panel C the staining for Nmyc-Pex11β and PMP70, in panel F the staining for Nmyc-Pex11β and Pex11α were merged. Note that Nmyc-Pex11β is distributed over the entire tubules, whereas PMP70 and Pex11α are concentrated in discrete patches. Scale bar represents 10 μm. (G, H) Electron micrographs of proliferating rat liver showing peroxisomal tubules of variable size (P or *) with multiple constrictions (arrowheads). Bar represents 500 nm. to proliferate peroxisomes correlates well with phenotypic alterations in mice carrying targeted deletions of PEX11α and/ or PEX11β. Whereas Pex11β knockout mice showed a number of Zellweger phenotypes, such as defects in neuronal migration, hypotonia and developmental delay, deletion of Pex11α did not alter the phenotype [11,90]. The expression of Pex11γ is highest in liver where it is constitutively expressed, and is not induced by peroxisome proliferators [64,89]. Different to

7 D. Lay et al. / Biochimica et Biophysica Acta xx (2006) xxx xxx 7 Pex11α and Pex11β, its overexpression did not induce peroxisome proliferation. However, peroxisome proliferation generated by overexpression of Pex11α and Pex11β showed distinct peculiarities dependent on the expression of the wild type, the myc-tagged or the GFP-tagged protein. The N- terminally myc-tagged form of Pex11β (Nmyc-Pex11β), for instance, resulted in the formation of tubular peroxisomes that revealed patches of both accumulated membrane (Pex11α, PMP70) and matrix (catalase, acyl-coa oxidase) proteins (Fig. 3A F and [10]). In contrast to that, Nmyc-Pex11β appeared to be distributed over the entire peroxisome. These data complement electron microscopic observations in proliferating rat liver mentioned above that showed extended multiply constricted peroxisomal tubules (Fig. 3G, H). Tagging Pex11β with the large egfp or CFP molecule also caused tubulation of peroxisomes and in addition resulted in the formation of large peroxisomal clusters suggesting interaction of Pex11β with cytoskeletal components, most likely microtubules (Lay et al., unpublished observations). A detailed study of Pex11 functions was recently reported in Arabidopsis plants [91]. These cells express five Pex11 homologs, three of which contain a dilysine motif. Two of these isoforms when overexpressed caused peroxisome elongation without subsequent fission. The C-terminal dilysine motif was not necessary for elongation. The third homolog caused peroxisome duplication. Deletion of the motif from this homolog, however, led to peroxisome elongation prior to fission. Thus, the motif in this homolog prevented elongation, and limits a putative site of Arf/coatomer action to a process between elongation and fission. What is presently discussed as main functions of Arf and coatomer in vesicular transport implicates coat formation, cargo selection and recruitment, changes in the lipid environment, and association with cytoskeletal elements particularly the actin cytoskeleton. It seems that Arf and coatomer are involved in establishing a suitable platform prior to fission of a COPIcoated vesicle. The observed segregation of peroxisomal membrane and matrix proteins involving Pex11β might represent a transient stage necessary for peroxisome division. Arf/coatomer might be required to reach this stage. The final division of peroxisomes both in yeast and mammalian cells clearly involves the dynamin-related proteins, such as Vps1 in yeast and DLP1 in mammals [5 8,92]. ER resident proteins that have to be relocated to the ER, a scenario well known for ER-Golgi trafficking (see: [93]). In this context, the observation in S. cerevisiae might be interesting that Emp24, an ER resident protein functioning in cargo protein selection and sorting, has recently been localized to peroxisomes [94]. Thus, retrieval of proteins from peroxisomes to ER via COPI-coated vesicles might represent another domain of peroxisomal Arf/coatomer function. The possible sites of Arf/ coatomer interference with peroxisomes are objected in a model shown in Fig. 4, considering two putative implications. One addresses the retrieval of ER proteins transiently localized to peroxisomes back to the ER. The other considers the formation of membrane and matrix protein patches as well as constrictions of tubular membranes, both processes that might precede peroxisome fission. A vesicular shuttle between the ER and peroxisomes may serve the exchange of both membrane proteins implicated in shuttling and lipids, such as phospholipids and ether lipid precursors. Recent work in plant cells also discussed shuttling between peroxisomes and ER. Various models were depicted [95] postulating ER to peroxisome intermediary structures, such as peroxisomal ER (per), preperoxisomal ER vesicles or preperoxisomal lamellae, structures that resemble preperoxisomal compartments previously described in mammalian systems [2, 96]. Interestingly, infecting tobacco Bright Yello-2 cells with Tomato bushy stunt virus (TBSV) causes alterations in the peroxisomal membrane and the formation of peroxisomal multivesicular bodies by membrane invagination [97]. These alterations were accompanied by the relocation of p33 from peroxisomes to per. p33 is a viral auxiliary replication protein initially accumulating in peroxisomes of infected cells. The observed sorting to per of p33 is completely abolished by the coexpression of a mutant version of Arf1 suggesting Arf1/ coatomer to be implicated in the formation of peroxisomal 3.4. Putative involvement of Arf/coatomer in peroxisome to ER retrieval Provided Pex3 and Pex19, that both have been proposed to take their route to the peroxisome via the ER, exit the ER in form of vesicles, a protein coat might be needed to promote vesiculation. Current discussions suggest that the coat involved in this process might not be COPI or COPII. Regardless of the nature of this coat complex subsequent fusion of the vesicles with the peroxisomal target membrane requires SNARE proteins, v-snares on the vesicles and t-snares on peroxisomes. Thus, constitutive transport of vesicles from the ER to peroxisomes leads to the accumulation in peroxisomes of Fig. 4. Model showing putative sites of action of Arf/coatomer on the peroxisomal system. Provided the recently suggested involvement of the ER in peroxisome biogenesis implicates vesicular transport from the ER to peroxisomes (1), a retrieval pathway is proposed (2) that transports ER-derived factors from peroxisomes back to the ER. Such a shuttle system could also be useful for lipid exchange between these two compartments, e.g., phospholipids and ether lipid precursors. Under conditions of proliferating peroxisomes peroxisomal elongation, formation of segregated protein patches and tubular constrictions may precede fission (3, 4). Cycles of Arf/coatomer recruitment and release are proposed to facilitate patch formation and or interaction with cytoskeletal elements.

8 8 D. Lay et al. / Biochimica et Biophysica Acta xx (2006) xxx xxx vesicles carrying p33. A dibasic targeting signal at the N- terminus of p33 (i.e. K 5 R 6 K 11 K 12 ) was found to be involved in p33 targeting from peroxisomes to per [97]. The models appear to offer attractive possibilities, however, COPI-coated peroxisomal vesicles, although much less abundant than COPIcoated Golgi-derived vesicles and therefore difficult to analyze, have still to be convincingly demonstrated [9,68]. Moreover, the vesicles once formed should be destined for fusion. For both homotypic (peroxisome peroxisome) [98] and heterotypic (peroxisome ER) fusion, these vesicles should be equipped with v-snares required for specific membrane targeting. So far, however, no real peroxisomal SNARE has been identified (cf. [99]). 4. Conclusions Highly purified rat liver peroxisomes bind Arf1, Arf6 and all subunits of coatomer from a rat liver cytosol. The peroxisomal recruitment of ARF/coatomer reflects characteristics similar to those observed for Golgi membranes, for example dependence on both Arf1 and GTP. There were, however, conspicuous peculiarities observed with the peroxisomal system that were not reported for other compartments, such as effects of ATP and a distinct cytosolic factor, response to treatment with peroxisome proliferators or insensitivity to BFA. These observations attribute to this Arf/coatomer binding a peroxisomal specificity. Further indications for a peroxisomal specificity of Arf/coatomer recruitment were derived from studies on CHO cells expressing a temperature-sensitive allele of ε-cop. These cells strikingly change morphology of peroxisomes when forced to proliferate the organelle. Other observations made in vivo in cells of the yeast S. cerevisiae expressing a temperature-sensitive arf1 allele and/or containing arf2/arf3 deletions additionally indicate a significant contribution of ScArf1 and ScArf3 on oleate-induced peroxisome proliferation. Although the studies on Arf/coatomer interaction with peroxisomes were initiated by detecting a Golgi to ER retrieval motif at the C-terminus of the peroxisomal membrane protein Pex11α, it is not clear to date whether this motif plays a significant role in this interaction. The structural requirements for the p24 family of proteins where the motif first was described and for mammalian Pex11α reveal common features such as the membrane topology of the motif-bearing C-terminal tail, favoring implication. Whereas these structural data might indicate involvement of the dilysine motif, functional studies do not support this view, although they do not completely rule out some contribution. Expression of Pex11 proteins in various species indicated that the process of peroxisome proliferation might be subdivided into sequential steps including elongation/tubulation of the organelle, formation of protein patches of distinct membrane and matrix proteins, development of multiple membrane constrictions and membrane fission. While membrane fission might involve the GTPase DLP1 that has been shown to function in a similar way also in mitochondria, the mechanism how peroxisomal proteins are segregated from each other prior to forming constrictions is not known. It might, however, be reminiscent of the mechanisms involved in cargo selection and concentration on Golgi membranes suggesting that one possible site of action Arf/coatomer may target on peroxisomes is related to this segregation process. Another role Arf/coatomer might play on peroxisomes is the formation of COPI-coated vesicles. Such vesicles might be destined to retrieve cargo from peroxisomes back to the ER, a process being suggested by the recent observations in yeast indicating in early steps of peroxisome biogenesis the transport of Pex3 Pex19 complexes from the ER to the peroxisome. Provided this transport is mediated through carrier vesicles, a retrieval process might be inevitable. References [1] D. Hoepfner, D. Schildknegt, I. Braakman, P. Philippsen, H.F. Tabak, Contribution of the endoplasmic reticulum to peroxisome formation, Cell 122 (2005) [2] H.J. Geuze, J.L. Murk, A.K. Stroobants, J.M. Griffith, M.J. Kleijmeer, A.J. Koster, A.J. Verkleij, B. Distel, H.F. Tabak, Involvement of the endoplasmic reticulum in peroxisome formation, Mol. Biol. Cell 14 (2003) [3] H.F. Tabak, J.L. Murk, I. Braakman, H.J. Geuze, Peroxisomes start their life in the endoplasmic reticulum, Traffic 4 (2003) [4] P.K. Kim, R.T. Mullen, U. Schumann, J. Lippincott-Schwartz, The origin and maintenance of mammalian peroxisomes involves a de novo PEX16- dependent pathway from the ER, J. Cell Biol. 173 (2006) [5] D. Hoepfner, M. van den Berg, P. Philippsen, H.F. Tabak, E.H. Hettema, A role for Vps1p, actin, and the Myo2p motor in peroxisome abundance and inheritance in Saccharomyces cerevisiae, J. Cell Biol. 155 (2001) [6] A. Koch, G. Schneider, G.H. Luers, M. Schrader, Peroxisome elongation and constriction but not fission can occur independently of dynamin-like protein 1, J. Cell Sci. 117 (2004) [7] A. Koch, M. Thiemann, M. Grabenbauer, Y. Yoon, M.A. McNiven, M. Schrader, Dynamin-like protein 1 is involved in peroxisomal fission, J. Biol. Chem. 278 (2003) [8] X. Li, S.J. Gould, The dynamin-like GTPase DLP1 is essential for peroxisome division and is recruited to peroxisomes in part by PEX11, J. Biol. Chem. 278 (2003) [9] M. Passreiter, M. Anton, D. Lay, R. Frank, C. Harter, F.T. Wieland, K. Gorgas, W.W. Just, Peroxisome biogenesis: involvement of ARF and coatomer, J. Cell Biol. 141 (1998) [10] M. Schrader, B.E. Reuber, J.C. Morrell, G. Jimenez-Sanchez, C. Obie, T.A. Stroh, D. Valle, T.A. Schroer, S.J. Gould, Expression of PEX11beta mediates peroxisome proliferation in the absence of extracellular stimuli, J. Biol. Chem. 273 (1998) [11] X. Li, S.J. Gould, PEX11 promotes peroxisome division independently of peroxisome metabolism, J. Cell Biol. 156 (2002) [12] S. Thoms, R. Erdmann, Dynamin-related proteins and Pex11 proteins in peroxisome division and proliferation, FEBS J. 272 (2005) [13] J.G. Donaldson, A. Honda, Localization and function of Arf family GTPases, Biochem. Soc. Trans. 33 (2005) [14] M. Franco, P. Chardin, M. Chabre, S. Paris, Myristoylation-facilitated binding of the G protein ARF1GDP to membrane phospholipids is required for its activation by a soluble nucleotide exchange factor, J. Biol. Chem. 271 (1996) [15] P.A. Randazzo, T. Terui, S. Sturch, H.M. Fales, A.G. Ferrige, R.A. Kahn, The myristoylated amino terminus of ADP-ribosylation factor 1 is a phospholipid- and GTP-sensitive switch, J. Biol. Chem. 270 (1995) [16] C.L. Jackson, J.E. Casanova, Turning on ARF: the Sec7 family of guanine nucleotide-exchange factors, Trends Cell Biol. 10 (2000) [17] P.A. Randazzo, D.S. Hirsch, Arf GAPs: multifunctional proteins that

9 D. Lay et al. / Biochimica et Biophysica Acta xx (2006) xxx xxx 9 regulate membrane traffic and actin remodelling, Cell. Signal 16 (2004) [18] J.G. Donaldson, A. Honda, R. Weigert, Multiple activities for Arf1 at the Golgi complex, Biochim. Biophys. Acta 1744 (2005) [19] E.C. Dell'Angelica, R. Puertollano, C. Mullins, R.C. Aguilar, J.D. Vargas, L.M. Hartnell, J.S. Bonifacino, GGAs: a family of ADP ribosylation factor-binding proteins related to adaptors and associated with the Golgi complex, J. Cell Biol. 149 (2000) [20] R. Puertollano, P.A. Randazzo, J.F. Presley, L.M. Hartnell, J.S. Bonifacino, The GGAs promote ARF-dependent recruitment of clathrin to the TGN, Cell 105 (2001) [21] M. Franco, P.J. Peters, J. Boretto, E. van Donselaar, A. Neri, C. D'Souza-Schorey, P. Chavrier, EFA6, a sec7 domain-containing exchange factor for ARF6, coordinates membrane recycling and actin cytoskeleton organization, EMBO J. 18 (1999) [22] E. Macia, F. Luton, M. Partisani, J. Cherfils, P. Chardin, M. Franco, The GDP-bound form of Arf6 is located at the plasma membrane, J. Cell Sci. 117 (2004) [23] H. Radhakrishna, J.G. Donaldson, ADP-ribosylation factor 6 regulates a novel plasma membrane recycling pathway, J. Cell Biol. 139 (1997) [24] D.A. Schafer, C. D'Souza-Schorey, J.A. Cooper, Actin assembly at membranes controlled by ARF6, Traffic 1 (2000) [25] Y. Aikawa, T.F. Martin, ARF6 regulates a plasma membrane pool of phosphatidylinositol(4,5)bisphosphate required for regulated exocytosis, J. Cell Biol. 162 (2003) [26] L.C. Santy, J.E. Casanova, Activation of ARF6 by ARNO stimulates epithelial cell migration through downstream activation of both Rac1 and phospholipase D, J. Cell Biol. 154 (2001) [27] A. Godi, P. Pertile, R. Meyers, P. Marra, G. Di Tullio, C. Iurisci, A. Luini, D. Corda, M.A. De Matteis, ARF mediates recruitment of PtdIns-4-OH kinase-beta and stimulates synthesis of PtdIns(4,5)P2 on the Golgi complex, Nat. Cell Biol. 1 (1999) [28] L.P. Haynes, G.M. Thomas, R.D. Burgoyne, Interaction of neuronal calcium sensor-1 and ADP-ribosylation factor 1 allows bidirectional control of phosphatidylinositol 4-kinase beta and trans-golgi networkplasma membrane traffic, J. Biol. Chem. 280 (2005) [29] D.H. Jones, J.B. Morris, C.P. Morgan, H. Kondo, R.F. Irvine, S. Cockcroft, Type I phosphatidylinositol 4-phosphate 5-kinase directly interacts with ADP-ribosylation factor 1 and is responsible for phosphatidylinositol 4,5-bisphosphate synthesis in the Golgi compartment, J. Biol. Chem. 275 (2000) [30] S. Cockcroft, Signalling roles of mammalian phospholipase D1 and D2, Cell. Mol. Life Sci. 58 (2001) [31] B. Antonny, S. Beraud-Dufour, P. Chardin, M. Chabre, N-terminal hydrophobic residues of the G-protein ADP-ribosylation factor-1 insert into membrane phospholipids upon GDP to GTP exchange, Biochemistry (Mosc) 36 (1997) [32] S. Beraud-Dufour, S. Paris, M. Chabre, B. Antonny, Dual interaction of ADP ribosylation factor 1 with Sec7 domain and with lipid membranes during catalysis of guanine nucleotide exchange, J. Biol. Chem. 274 (1999) [33] D.U. Gommel, A.R. Memon, A. Heiss, F. Lottspeich, J. Pfannstiel, J. Lechner, C. Reinhard, J.B. Helms, W. Nickel, F.T. Wieland, Recruitment to Golgi membranes of ADP-ribosylation factor 1 is mediated by the cytoplasmic domain of p23, EMBO J. 20 (2001) [34] A. Honda, O.S. Al-Awar, J.C. Hay, J.G. Donaldson, Targeting of Arf-1 to the early Golgi by membrin, an ER-Golgi SNARE, J. Cell Biol. 168 (2005) [35] J.G. Donaldson, C.L. Jackson, Regulators and effectors of the ARF GTPases, Curr. Opin. Cell Biol. 12 (2000) [36] A. Peyroche, B. Antonny, S. Robineau, J. Acker, J. Cherfils, C.L. Jackson, Brefeldin A acts to stabilize an abortive ARF-GDP-Sec7 domain protein complex: involvement of specific residues of the Sec7 domain, Mol. Cell 3 (1999) [37] M. Zeghouf, B. Guibert, J.C. Zeeh, J. Cherfils, Arf, Sec7 and Brefeldin A: a model towards the therapeutic inhibition of guanine nucleotide-exchange factors, Biochem. Soc. Trans. 33 (2005) [38] T.K. Niu, A.C. Pfeifer, J. Lippincott-Schwartz, C.L. Jackson, Dynamics of GBF1, a Brefeldin A-sensitive Arf1 exchange factor at the Golgi, Mol. Biol. Cell 16 (2005) [39] A. Claude, B.P. Zhao, C.E. Kuziemsky, S. Dahan, S.J. Berger, J.P. Yan, A.D. Armold, E.M. Sullivan, P. Melancon, GBF1: a novel Golgi-associated BFA-resistant guanine nucleotide exchange factor that displays specificity for ADP-ribosylation factor 5, J. Cell Biol. 146 (1999) [40] K. Kawamoto, Y. Yoshida, H. Tamaki, S. Torii, C. Shinotsuka, S. Yamashina, K. Nakayama, GBF1, a guanine nucleotide exchange factor for ADP-ribosylation factors, is localized to the cis-golgi and involved in membrane association of the COPI coat, Traffic 3 (2002) [41] M. Sata, J. Moss, M. Vaughan, Structural basis for the inhibitory effect of brefeldin A on guanine nucleotide-exchange proteins for ADP-ribosylation factors, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) [42] P.J. Cullen, P. Chardin, Membrane targeting: what a difference a G makes, Curr. Biol. 10 (2000) R876 R878. [43] H.W. Shin, K. Nakayama, Guanine nucleotide-exchange factors for arf GTPases: their diverse functions in membrane traffic, J. Biochem. 136 (2004) (Tokyo). [44] S. Paris, S. Beraud-Dufour, S. Robineau, J. Bigay, B. Antonny, M. Chabre, P. Chardin, Role of protein phospholipid interactions in the activation of ARF1 by the guanine nucleotide exchange factor ARNO, J. Biol. Chem. 272 (1997) [45] M. De Matteis, A. Godi, D. Corda, Phosphoinositides and the Golgi complex, Curr. Opin. Cell Biol. 14 (2002) [46] M.G. Roth, Phosphoinositides in constitutive membrane traffic, Physiol. Rev. 84 (2004) [47] R. Duden, G. Griffiths, R. Frank, P. Argos, T.E. Kreis, Beta-COP, a 110 kd protein associated with non-clathrin-coated vesicles and the Golgi complex, shows homology to beta-adaptin, Cell 64 (1991) [48] A. Eugster, G. Frigerio, M. Dale, R. Duden, COP I domains required for coatomer integrity, and novel interactions with ARF and ARF-GAP, EMBO J. 19 (2000) [49] K. Schledzewski, H. Brinkmann, R.R. Mendel, Phylogenetic analysis of components of the eukaryotic vesicle transport system reveals a common origin of adaptor protein complexes 1, 2, and 3 and the F subcomplex of the coatomer COPI, J. Mol. Evol. 48 (1999) [50] T. Serafini, G. Stenbeck, A. Brecht, F. Lottspeich, L. Orci, J.E. Rothman, F.T. Wieland, A coat subunit of Golgi-derived non-clathrin-coated vesicles with homology to the clathrin-coated vesicle coat protein beta-adaptin, Nature 349 (1991) [51] H. Takatsu, M. Futatsumori, K. Yoshino, Y. Yoshida, H.W. Shin, K. Nakayama, Similar subunit interactions contribute to assembly of clathrin adaptor complexes and COPI complex: analysis using yeast three-hybrid system, Biochem. Biophys. Res. Commun. 284 (2001) [52] H.T. McMahon, I.G. Mills, COP and clathrin-coated vesicle budding: different pathways, common approaches, Curr. Opin. Cell Biol. 16 (2004) [53] A. Eugster, G. Frigerio, M. Dale, R. Duden, The alpha- and beta -COP WD40 domains mediate cargo-selective interactions with distinct di-lysine motifs, Mol. Biol. Cell 15 (2004) [54] G. Stenbeck, C. Harter, A. Brecht, D. Herrmann, F. Lottspeich, L. Orci, F.T. Wieland, beta -COP, a novel subunit of coatomer, EMBO J. 12 (1993) [55] P.J. Watson, G. Frigerio, B.M. Collins, R. Duden, D.J. Owen, Gamma-COP appendage domain Structure and function, Traffic 5 (2004) [56] N. Blagitko, U. Schulz, A.A. Schinzel, H.H. Ropers, V.M. Kalscheuer, gamma2-cop, a novel imprinted gene on chromosome 7q32, defines a new imprinting cluster in the human genome, Hum. Mol. Genet. 8 (1999) [57] M. Futatsumori, K. Kasai, H. Takatsu, H.W. Shin, K. Nakayama, Identification and characterization of novel isoforms of COP I subunits, J. Biochem. 128 (2000) (Tokyo). [58] D. Wegmann, P. Hess, C. Baier, F.T. Wieland, C. Reinhard, Novel isotypic gamma/zeta subunits reveal three coatomer complexes in mammals, Mol. Cell. Biol. 24 (2004) [59] M.G. Roth, Lipid regulators of membrane traffic through the Golgi complex, Trends Cell Biol. 9 (1999)